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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
3a73b715-c057-45c6-a3ff-6e98ea0864ab | on-the-use-of-semantically-aligned-speech | 2210.05291 | null | https://arxiv.org/abs/2210.05291v1 | https://arxiv.org/pdf/2210.05291v1.pdf | On the Use of Semantically-Aligned Speech Representations for Spoken Language Understanding | In this paper we examine the use of semantically-aligned speech representations for end-to-end spoken language understanding (SLU). We employ the recently-introduced SAMU-XLSR model, which is designed to generate a single embedding that captures the semantics at the utterance level, semantically aligned across different languages. This model combines the acoustic frame-level speech representation learning model (XLS-R) with the Language Agnostic BERT Sentence Embedding (LaBSE) model. We show that the use of the SAMU-XLSR model instead of the initial XLS-R model improves significantly the performance in the framework of end-to-end SLU. Finally, we present the benefits of using this model towards language portability in SLU. | ['Yannick Estève', 'Themos Stafylakis', 'Mickaël Rouvier', 'Valentin Pelloin', 'Gaëlle Laperrière'] | 2022-10-11 | null | null | null | null | ['spoken-language-understanding', 'spoken-language-understanding'] | ['natural-language-processing', 'speech'] | [ 2.03948379e-01 4.05937195e-01 1.71517655e-01 -7.97809482e-01
-1.23769367e+00 -4.63454217e-01 7.53153086e-01 3.04356180e-02
-3.98966461e-01 2.75006801e-01 1.04977953e+00 -4.11992550e-01
1.47941336e-01 -3.82382005e-01 -5.88947713e-01 -4.70936522e-02
2.84958817e-02 3.88198465e-01 1.09572057e-02 -6.24645531e-01
-2.42117196e-01 -5.79039641e-02 -1.47636390e+00 5.44251442e-01
3.78052652e-01 7.79853284e-01 5.50229132e-01 9.88121986e-01
-3.84242862e-01 9.18029010e-01 -4.66833323e-01 2.98352957e-01
-1.00505084e-01 -7.32206345e-01 -1.02435637e+00 -1.13581188e-01
3.94427329e-01 -2.09456131e-01 -3.69190127e-01 4.45042640e-01
4.73627329e-01 4.90320385e-01 5.85725844e-01 -9.05601323e-01
-6.03329003e-01 9.61142540e-01 2.49121442e-01 7.37160593e-02
8.61661196e-01 -1.92902029e-01 1.29152751e+00 -1.11797881e+00
7.09307373e-01 1.74857807e+00 3.73808324e-01 9.36630070e-01
-1.20663571e+00 -1.77127838e-01 3.53169262e-01 5.67878038e-02
-1.37728322e+00 -8.69814813e-01 5.53893864e-01 3.58606279e-02
1.54028058e+00 3.23634773e-01 2.81710029e-02 1.15420008e+00
-1.15042605e-01 1.33611345e+00 1.07026672e+00 -8.68853331e-01
3.80259514e-01 2.07537934e-02 5.59797287e-01 5.20205915e-01
-5.29066026e-01 2.80564755e-01 -8.42530787e-01 7.48271719e-02
2.39073902e-01 -5.30428290e-01 -3.05217057e-01 -1.85226366e-01
-1.01778841e+00 8.93026590e-01 3.03595752e-01 5.11297584e-01
-5.68940900e-02 3.04914296e-01 8.40782106e-01 4.68802154e-01
6.50778949e-01 6.48011193e-02 -6.75647676e-01 -1.57033384e-01
-8.90474677e-01 9.49112251e-02 7.80272663e-01 9.13280785e-01
4.10455048e-01 5.51028609e-01 -1.27438188e-01 1.06798077e+00
7.15824723e-01 2.63911277e-01 1.07734287e+00 -8.84851515e-01
1.58531681e-01 -2.39060167e-02 -2.34465480e-01 1.49921421e-02
-2.79560298e-01 -1.63824573e-01 -1.17094800e-01 1.03717566e-01
-1.04513392e-01 -6.87875748e-02 -1.00356758e+00 2.06616855e+00
-2.77040154e-02 3.85865450e-01 1.12706006e+00 7.48389423e-01
1.18576992e+00 1.15587926e+00 3.24098438e-01 -4.31078412e-02
1.30889070e+00 -1.13158762e+00 -9.65595901e-01 -5.26697278e-01
8.87087345e-01 -4.63313967e-01 1.53752387e+00 6.93767844e-03
-1.20147467e+00 -7.28217006e-01 -1.21499515e+00 -4.56220925e-01
-5.61981142e-01 6.64619282e-02 6.76296055e-02 5.40418506e-01
-1.62026644e+00 1.30105212e-01 -5.34648120e-01 -5.49133122e-01
-4.23828125e-01 7.24690855e-02 -4.20287281e-01 2.48612221e-02
-1.58701873e+00 8.91942322e-01 6.64297342e-01 -3.39361846e-01
-8.94739270e-01 -4.40426797e-01 -1.50129509e+00 1.07794359e-01
2.95784384e-01 -4.10436332e-01 1.80669284e+00 -8.43436658e-01
-2.11988592e+00 7.96205521e-01 -6.10343158e-01 -7.41516411e-01
-5.60496701e-03 -4.52273190e-01 -2.28473783e-01 -1.49601549e-02
-1.86249211e-01 8.25372756e-01 5.18852890e-01 -1.37003219e+00
-3.42448950e-01 -2.53336787e-01 -1.65948600e-01 4.50646996e-01
9.59837809e-02 1.52394444e-01 -5.98827982e-03 -6.98880970e-01
1.77943006e-01 -5.61436176e-01 -3.81561294e-02 -6.03664696e-01
1.24331892e-01 -6.16110325e-01 9.73680973e-01 -9.08103406e-01
1.09059179e+00 -2.30486012e+00 5.66076815e-01 -1.03387155e-01
-4.62295145e-01 2.30871737e-01 -6.78682387e-01 8.44387591e-01
-2.99421728e-01 -2.32552420e-02 -4.25257474e-01 -9.89011884e-01
1.53390989e-01 6.86878562e-01 -6.85938478e-01 -8.86218399e-02
1.38782412e-01 8.89038563e-01 -8.99807215e-01 1.05728664e-01
6.55111253e-01 5.03849685e-01 -3.66197586e-01 6.95577383e-01
-5.97297430e-01 2.98145980e-01 -6.30747974e-02 1.58776805e-01
5.22061400e-02 4.81879234e-01 3.07975441e-01 5.64104654e-02
-1.17570452e-01 9.48575914e-01 -1.03641486e+00 2.19795060e+00
-1.12957549e+00 3.72229218e-01 3.68319117e-02 -6.20304406e-01
1.04512608e+00 8.19277287e-01 2.31372584e-02 -6.59492016e-01
8.83590728e-02 4.27837193e-01 -2.02592179e-01 -8.32088292e-02
5.33167720e-01 -4.24734533e-01 -4.82766539e-01 7.29421794e-01
7.35501349e-01 -3.98998499e-01 -2.83252835e-01 3.13386858e-01
7.19595313e-01 3.47918600e-01 7.40168333e-01 -3.19937974e-01
7.59655952e-01 -5.83063662e-01 -1.38343304e-01 6.49732113e-01
-1.81250691e-01 7.30786562e-01 -1.12974001e-02 -1.33390082e-02
-8.70598495e-01 -1.31030798e+00 5.07221976e-03 1.48505521e+00
-3.88678253e-01 -6.46121085e-01 -8.61852884e-01 -5.88380098e-01
-3.25767249e-01 1.55316234e+00 -2.18760625e-01 -2.38027707e-01
-4.86879528e-01 1.33942127e-01 6.77969396e-01 6.28861785e-01
1.95297487e-02 -1.24397612e+00 -2.58759677e-01 3.62968624e-01
-1.63690388e-01 -1.26824009e+00 -5.71841061e-01 1.85657129e-01
-4.19865936e-01 -3.43863398e-01 -5.90535939e-01 -8.79302561e-01
5.60187064e-02 3.98329422e-02 9.23363209e-01 -3.97566110e-01
2.03311101e-01 1.06441367e+00 -9.25755322e-01 -4.23650056e-01
-1.16862905e+00 -1.10107446e-02 2.23512143e-01 1.14391893e-02
2.49273717e-01 -1.77009046e-01 1.58266649e-01 -2.36556008e-01
-1.06749785e+00 6.23795623e-03 4.92255576e-02 6.90587223e-01
4.61488307e-01 -7.02018499e-01 9.38062370e-01 -4.81717378e-01
8.37200880e-01 -2.46919647e-01 -1.31535769e-01 3.29889894e-01
-1.87023476e-01 3.96036476e-01 6.54161155e-01 -1.48261040e-01
-9.60016012e-01 3.26325074e-02 -9.12329674e-01 -3.75297248e-01
-4.03162301e-01 6.23378456e-01 -4.16797310e-01 2.87087619e-01
3.51053685e-01 6.14737332e-01 2.41980284e-01 -6.68982565e-01
9.53772843e-01 1.15359354e+00 6.11677408e-01 -3.91903698e-01
1.26569510e-01 -5.22552729e-02 -6.13430560e-01 -1.41404450e+00
-8.75473917e-01 -6.58320189e-01 -5.73045611e-01 -5.11756502e-02
9.26437974e-01 -1.02019191e+00 -1.07294068e-01 1.64965332e-01
-1.44572568e+00 -4.78795975e-01 -6.67020023e-01 4.93745506e-01
-1.03631854e+00 6.19186580e-01 -5.73355496e-01 -1.22722995e+00
-4.64919835e-01 -1.11568546e+00 1.47318006e+00 -1.34488046e-01
-6.20130897e-01 -1.13031209e+00 2.04489097e-01 2.28872493e-01
3.54854465e-01 -3.15625608e-01 8.54998648e-01 -1.09001184e+00
-1.13004945e-01 1.23450831e-01 1.86699793e-01 8.30370665e-01
9.35253222e-03 -5.84621429e-01 -1.34220481e+00 -4.32982296e-01
1.32151693e-01 -5.96899569e-01 8.18196535e-01 3.09768230e-01
7.04068840e-01 -2.80913860e-01 3.97544205e-01 3.11544865e-01
1.22392893e+00 1.57579020e-01 4.87948149e-01 -3.86487469e-02
4.48618919e-01 7.72169709e-01 3.20783466e-01 1.35439411e-01
5.89721203e-01 7.97827065e-01 6.98349178e-02 1.24713384e-01
-5.22189736e-01 -4.81101394e-01 1.08808672e+00 1.42373455e+00
6.06276989e-01 -4.98193771e-01 -8.19529116e-01 7.45963871e-01
-1.84691370e+00 -4.61248457e-01 3.10613990e-01 2.11365318e+00
6.66334152e-01 -8.57074335e-02 -3.63172889e-02 -3.90051752e-02
8.93895552e-02 6.34847403e-01 2.35948749e-02 -1.04604280e+00
-1.83577731e-01 4.79576766e-01 2.72917151e-02 1.34136808e+00
-7.83294499e-01 1.48126018e+00 6.82827806e+00 7.02207208e-01
-8.58089089e-01 4.72693950e-01 2.22927436e-01 3.09250563e-01
-5.77045619e-01 -3.82157718e-03 -7.92839825e-01 -1.41093526e-02
1.73960531e+00 -1.96122095e-01 4.31786925e-01 7.68407762e-01
3.69376153e-01 1.74842566e-01 -1.22870421e+00 7.02311456e-01
5.79964757e-01 -1.15099645e+00 4.61031288e-01 -5.66523552e-01
1.99920237e-01 2.72605181e-01 -1.05876498e-01 4.69622135e-01
2.26559162e-01 -1.00251198e+00 9.34754431e-01 2.20869794e-01
9.51825321e-01 -6.71378314e-01 7.18794048e-01 3.53839844e-01
-1.28221726e+00 2.13005111e-01 -8.74869898e-02 1.61698721e-02
6.29490376e-01 -3.04895699e-01 -1.13984370e+00 7.29770720e-01
1.09526165e-01 6.03518844e-01 -5.52861318e-02 1.65664971e-01
-4.02793795e-01 7.78764725e-01 -2.98738927e-01 -5.38219549e-02
6.68651223e-01 -1.06027238e-01 9.76539969e-01 1.47000802e+00
1.86120868e-01 -1.36039089e-02 3.77252609e-01 5.52078724e-01
9.63462591e-02 3.60187352e-01 -8.08035672e-01 -1.68006092e-01
1.75856516e-01 5.15682817e-01 -1.59136847e-01 -3.27419251e-01
-5.52776158e-01 1.31341076e+00 2.87893713e-01 3.69141102e-01
-1.73059627e-01 -1.49435923e-02 9.35638130e-01 -1.97048187e-01
2.20103264e-01 -6.60098135e-01 9.45050716e-02 -9.69117403e-01
-2.40597337e-01 -8.42460096e-01 3.00700814e-01 -1.04199290e+00
-1.13593721e+00 9.74093556e-01 9.25894454e-02 -6.75462425e-01
-9.72365439e-01 -6.89361393e-01 -5.33275127e-01 1.06356716e+00
-1.61209619e+00 -1.40176439e+00 3.87322009e-01 2.46431470e-01
1.52578139e+00 -4.07826185e-01 1.43073988e+00 -2.87474662e-01
-1.53459534e-01 4.42928702e-01 1.10052995e-01 -1.61859654e-02
3.23839247e-01 -1.36038554e+00 8.38850617e-01 7.12457895e-01
7.65856624e-01 4.86106783e-01 9.27125633e-01 -3.90146792e-01
-1.06202388e+00 -9.91822302e-01 1.25381422e+00 -5.82172990e-01
6.79701865e-01 -6.76062405e-01 -1.11155152e+00 1.00391138e+00
5.35810351e-01 -3.83166999e-01 9.12107766e-01 6.10904321e-02
-3.13116014e-01 2.57586598e-01 -7.60006368e-01 6.35949433e-01
7.56583631e-01 -1.25773513e+00 -1.29359174e+00 1.04828328e-01
1.34925199e+00 -2.02888042e-01 -6.47382975e-01 3.16585660e-01
3.07813793e-01 -6.17447257e-01 9.76389766e-01 -8.81170332e-01
5.32446392e-02 -1.07324503e-01 -8.04609716e-01 -1.49997067e+00
1.38072774e-01 -5.41386604e-01 3.04416772e-02 1.04056752e+00
3.67844224e-01 -4.56261516e-01 1.72125742e-01 6.16621263e-02
-6.28046095e-01 -4.81459111e-01 -1.54115140e+00 -8.07370782e-01
2.10542917e-01 -9.06070471e-01 3.35846096e-01 4.56240535e-01
1.14441551e-01 8.82755876e-01 -2.89879382e-01 2.35346004e-01
2.54498214e-01 -2.49824032e-01 4.36483234e-01 -7.18648493e-01
-5.37194610e-01 -1.12159356e-01 -5.16121805e-01 -1.32117939e+00
8.30333412e-01 -1.10367084e+00 3.79382640e-01 -1.65068090e+00
-4.09728825e-01 6.12273738e-02 -4.70498592e-01 3.91135961e-01
-2.51799058e-02 -1.99294925e-01 4.73720789e-01 -2.02813089e-01
-3.36021185e-01 1.05747581e+00 5.47861814e-01 1.67380162e-02
-2.37091765e-01 -2.56455690e-01 -2.74290890e-01 6.14883125e-01
6.67298436e-01 -3.50851342e-02 -6.78188562e-01 -3.56875658e-01
-5.74820042e-01 1.78025052e-01 5.18108197e-02 -9.11758721e-01
-1.20899700e-01 2.74436206e-01 -3.54121029e-01 -5.66116273e-01
6.76975727e-01 -5.28444588e-01 -4.89519894e-01 3.59523743e-01
-6.12130463e-01 -1.52891159e-01 5.00429630e-01 3.97349447e-01
-5.93099117e-01 -4.60900664e-01 7.10868359e-01 -1.42338350e-01
-1.30955970e+00 -3.34305406e-01 -7.14050770e-01 -1.16264626e-01
7.68165946e-01 -1.31990045e-01 2.04319492e-01 -8.24494123e-01
-9.85409379e-01 1.58580154e-01 -4.78586331e-02 9.18551207e-01
1.02397692e+00 -1.04905009e+00 -9.36487257e-01 5.91044486e-01
4.30564642e-01 -3.60001624e-01 1.99837387e-01 1.45670205e-01
-1.17805943e-01 7.06530511e-01 2.66466737e-01 -4.77157682e-01
-1.41960347e+00 2.38236070e-01 4.18108910e-01 -7.36120120e-02
-6.20288014e-01 9.95854199e-01 3.62255633e-01 -1.09937954e+00
3.47015381e-01 -3.98090124e-01 -2.36548424e-01 -1.31213441e-01
6.30578458e-01 1.48222044e-01 -1.42428763e-02 -1.08742523e+00
-4.00349975e-01 1.51793316e-01 -7.73832500e-02 -8.80538464e-01
1.07273769e+00 -4.85843122e-01 2.46208295e-01 1.19887483e+00
1.23008311e+00 -8.39433633e-04 -8.43774915e-01 -2.67151088e-01
2.31379285e-01 1.81246921e-01 2.60024995e-01 -8.74283850e-01
-1.49649188e-01 9.80367243e-01 6.49969280e-01 -3.43246795e-02
6.68832541e-01 3.29183012e-01 9.67699826e-01 3.51169854e-01
4.42752451e-01 -1.01804888e+00 -9.32917744e-02 1.18073618e+00
1.26007080e+00 -8.46973002e-01 -5.89557052e-01 -2.69775391e-01
-9.94785368e-01 1.00100064e+00 2.69547641e-01 -6.89265803e-02
5.26111603e-01 1.82236031e-01 4.87117141e-01 1.61038741e-01
-7.78611600e-01 -5.59346557e-01 3.16818446e-01 4.74368572e-01
6.66039705e-01 3.03716511e-01 -2.57550120e-01 6.13518596e-01
-3.27351242e-01 -3.07321548e-01 5.22710502e-01 8.52277994e-01
-6.42418385e-01 -1.55125701e+00 -1.98707923e-01 -6.42931238e-02
9.76429135e-02 -2.68711627e-01 -6.79070294e-01 5.74119925e-01
-3.85329157e-01 1.04298306e+00 1.54389158e-01 -2.77805209e-01
5.22680998e-01 9.14058328e-01 3.12369943e-01 -1.11940479e+00
-4.30283904e-01 2.68483192e-01 4.48261738e-01 -6.60946190e-01
-2.26135507e-01 -6.13806605e-01 -1.67826736e+00 4.35961455e-01
2.59004515e-02 1.91305175e-01 7.02446759e-01 1.20164084e+00
2.65991002e-01 6.49278700e-01 6.63430393e-01 -7.03105867e-01
-7.48314559e-01 -1.18286812e+00 -6.38404369e-01 1.52244940e-01
5.42910337e-01 -2.73319870e-01 -3.56530368e-01 -5.03209792e-02] | [14.053627967834473, 6.999042987823486] |
c0a383ce-9e18-48e2-902f-6117ec464670 | lexical-normalization-of-user-generated | null | null | https://aclanthology.org/W19-3202 | https://aclanthology.org/W19-3202.pdf | Lexical Normalization of User-Generated Medical Text | In the medical domain, user-generated social media text is increasingly used as a valuable complementary knowledge source to scientific medical literature. The extraction of this knowledge is complicated by colloquial language use and misspellings. Yet, lexical normalization of such data has not been addressed properly. This paper presents an unsupervised, data-driven spelling correction module for medical social media. Our method outperforms state-of-the-art spelling correction and can detect mistakes with an F0.5 of 0.888. Additionally, we present a novel corpus for spelling mistake detection and correction on a medical patient forum. | ['Wessel Kraaij', 'Suzan Verberne', 'Anne Dirkson'] | 2019-08-01 | null | null | null | ws-2019-8 | ['mistake-detection', 'lexical-normalization'] | ['computer-vision', 'natural-language-processing'] | [ 4.71247643e-01 1.79426625e-01 -1.33959115e-01 -8.24051425e-02
-9.13574755e-01 -3.27899784e-01 2.46961519e-01 1.40375030e+00
-9.70744848e-01 8.38778973e-01 3.26042116e-01 -4.45591420e-01
4.86856923e-02 -5.77071249e-01 -2.35021785e-01 -3.24531227e-01
5.57109654e-01 4.26486462e-01 1.81610435e-01 -3.24255109e-01
7.72841632e-01 1.21943451e-01 -1.13325977e+00 6.09548867e-01
1.31621361e+00 6.83238322e-04 3.94934714e-01 7.21111357e-01
-7.83064067e-01 6.57400072e-01 -1.05893433e+00 -6.02409422e-01
-4.36381668e-01 -5.09397388e-01 -9.26708221e-01 -1.78380713e-01
1.26268759e-01 4.32172894e-01 1.80502176e-01 1.38035214e+00
6.43554628e-01 -4.04709220e-01 4.50101286e-01 -4.86731946e-01
-5.41444421e-01 7.99700737e-01 -4.54205543e-01 5.42000413e-01
5.11359870e-01 -4.52988178e-01 5.82464516e-01 -8.20338190e-01
9.76708353e-01 6.28901899e-01 7.10620940e-01 6.90737724e-01
-7.74071634e-01 -5.77479184e-01 -5.58796585e-01 -4.48384844e-02
-1.59326506e+00 -1.41669780e-01 2.99903393e-01 -6.10266268e-01
9.68565941e-01 3.91757399e-01 3.89148831e-01 9.00817871e-01
5.95835209e-01 4.98953402e-01 8.65178466e-01 -1.13830638e+00
1.04408488e-01 3.30824941e-01 3.93647969e-01 6.12937868e-01
8.24149311e-01 -6.06466234e-01 -6.78507686e-01 -3.99599373e-01
3.92442465e-01 1.10450447e-01 -2.04512224e-01 7.85086036e-01
-1.28711534e+00 5.27970076e-01 -4.18722332e-01 9.34452295e-01
-4.69575614e-01 -5.47364414e-01 5.78838706e-01 2.99841762e-01
7.89201498e-01 1.00330627e+00 -4.54464406e-01 -2.80667037e-01
-1.23053670e+00 1.60776809e-01 7.92984843e-01 8.53447258e-01
2.24272713e-01 -5.43024123e-01 -4.10251051e-01 9.05151367e-01
5.02135716e-02 6.51137710e-01 1.06679273e+00 -1.60825625e-01
2.96256363e-01 7.59062409e-01 1.33593097e-01 -1.20102799e+00
-6.59456968e-01 -6.82332218e-01 -7.26812422e-01 -5.58650732e-01
3.95539939e-01 -1.06864637e-02 -7.62003124e-01 1.02016747e+00
2.90374249e-01 -1.17162503e-01 2.40312651e-01 3.21583122e-01
1.09107649e+00 2.06236929e-01 2.49470964e-01 -6.38672829e-01
1.53882027e+00 -6.94846690e-01 -1.43226159e+00 5.82062565e-02
1.09770453e+00 -1.47352529e+00 7.74153471e-01 5.76577604e-01
-1.04506099e+00 -8.33693519e-02 -7.00047433e-01 -2.24189386e-02
-3.97183746e-01 1.29452790e-03 2.43464455e-01 8.53492379e-01
-6.18916571e-01 7.86425591e-01 -7.68154025e-01 -5.85306585e-01
4.88979012e-01 -9.56862271e-02 -2.45386675e-01 1.03138283e-01
-1.14372551e+00 1.08715022e+00 3.81278187e-01 -5.84986150e-01
1.37745947e-01 -1.11664367e+00 -5.91320872e-01 -5.06190717e-01
1.51738212e-01 -5.96890926e-01 1.43155980e+00 -9.48284864e-01
-1.18229938e+00 1.41649258e+00 -3.69275004e-01 -4.54232275e-01
5.35836041e-01 -2.36694753e-01 -1.20267129e+00 2.00132713e-01
4.26597118e-01 -2.00166821e-01 4.10732627e-01 -5.84201753e-01
-8.19204867e-01 -1.40737340e-01 -7.22865999e-01 -1.56868380e-02
-4.29861307e-01 5.97001076e-01 -2.40324914e-01 -1.22559059e+00
2.10977886e-02 -7.44640052e-01 -3.62489402e-01 -3.50189507e-01
-3.99018586e-01 -1.24485828e-01 -1.01217866e-01 -1.11773491e+00
2.05142474e+00 -2.10447907e+00 -3.46840501e-01 4.10147548e-01
3.89474899e-01 6.17585897e-01 3.12846512e-01 6.80616140e-01
-3.25300097e-02 6.57169998e-01 -3.38988304e-01 -3.01110983e-01
-4.18180615e-01 4.38226908e-02 -5.26569746e-02 5.04363298e-01
-1.67354703e-01 7.19471574e-01 -1.47672296e+00 -8.59463692e-01
-1.75006479e-01 2.51568556e-01 -3.63187283e-01 7.30356853e-03
1.05127513e-01 3.95907789e-01 -2.29157940e-01 6.38240516e-01
5.00479460e-01 -3.98161411e-01 1.81992218e-01 1.33307144e-01
-3.18264216e-01 6.46064579e-01 -7.11280286e-01 1.81974030e+00
-2.95320094e-01 3.86830240e-01 -6.04880810e-01 -1.61220908e-01
5.87793529e-01 3.43168825e-01 5.81838608e-01 -3.82522464e-01
2.06944451e-01 5.81883073e-01 -1.33121416e-01 -9.04057741e-01
7.41005719e-01 -3.71453702e-01 -9.81377587e-02 2.61114478e-01
2.24749278e-02 1.29718646e-01 5.61724782e-01 4.33631182e-01
1.16171432e+00 -4.11303014e-01 1.09821677e+00 -4.65976954e-01
6.01902008e-01 2.54270524e-01 5.13946176e-01 8.11519086e-01
7.63103291e-02 7.58908868e-01 8.32904279e-02 -9.78141874e-02
-7.85386205e-01 -5.02448559e-01 -3.09990108e-01 5.78331351e-01
-2.38902301e-01 -1.02493167e+00 -9.83573079e-01 -7.64344633e-01
-2.19292045e-02 9.06044066e-01 -6.30503833e-01 2.90166009e-02
-3.27134550e-01 -8.33108604e-01 8.63017380e-01 5.89054711e-02
-9.08008590e-03 -1.00909734e+00 -3.52811486e-01 4.84888136e-01
-3.89096290e-01 -1.00467706e+00 -8.16249907e-01 -2.22285748e-01
-7.06418812e-01 -1.31820154e+00 -8.01983118e-01 -8.02240789e-01
9.25014317e-01 8.74341205e-02 1.08123517e+00 4.59172159e-01
-5.65710008e-01 3.78502421e-02 -9.12167549e-01 -1.05591023e+00
-1.11654568e+00 2.06941575e-01 -1.82199813e-02 -2.94988990e-01
9.48519289e-01 -1.45722598e-01 -3.83687347e-01 -1.35668203e-01
-1.10087955e+00 1.06553100e-01 4.68485355e-01 6.38599396e-01
5.89849830e-01 -1.88282430e-01 3.98069322e-01 -1.68676615e+00
7.64189959e-01 -4.71312225e-01 -1.43577695e-01 2.45062634e-01
-8.49804223e-01 6.59295246e-02 4.78523612e-01 -1.78518414e-01
-8.41800272e-01 -2.72645414e-01 -5.32460392e-01 3.72886240e-01
-3.55022013e-01 8.47330272e-01 4.59923387e-01 -9.92318317e-02
1.12927794e+00 1.09450124e-01 1.04971416e-01 -7.83660769e-01
-2.09627554e-01 1.06588411e+00 5.07735908e-01 2.96911538e-01
4.45510268e-01 1.82318196e-01 -2.90743560e-01 -1.11068177e+00
-1.05294347e+00 -1.10526347e+00 -5.70281029e-01 6.45295680e-02
7.17080772e-01 -5.86302280e-01 -3.80304426e-01 5.18078327e-01
-1.37790453e+00 3.26124132e-01 -2.32810557e-01 6.30476415e-01
1.44733340e-01 7.67374158e-01 -5.91636360e-01 -5.50132751e-01
-6.59981430e-01 -4.91246760e-01 5.97297668e-01 8.26356485e-02
-9.35222328e-01 -8.65455806e-01 4.60238427e-01 4.24988240e-01
1.64507404e-01 2.54038006e-01 6.60122395e-01 -1.06267369e+00
6.00932300e-01 -4.67361778e-01 1.67366490e-01 8.26578811e-02
5.66664696e-01 -5.92578854e-03 -6.37793422e-01 2.37034168e-02
-5.24513163e-02 3.45447332e-01 6.87225461e-01 7.93803409e-02
1.11645734e+00 -5.22657096e-01 -4.30026442e-01 1.74217433e-01
1.09996617e+00 9.92334709e-02 6.36542976e-01 5.58583915e-01
6.66936100e-01 3.79268259e-01 5.42252243e-01 6.91795409e-01
2.46351108e-01 1.94527298e-01 -1.85697809e-01 6.57193661e-02
-3.04032043e-02 -2.15493366e-01 -1.15547054e-01 1.25834239e+00
-1.05201658e-02 -2.46948451e-01 -1.46479094e+00 6.29569292e-01
-1.47957480e+00 -8.96394730e-01 -7.41622686e-01 2.25683498e+00
1.45540714e+00 1.40363365e-01 -1.46999717e-01 4.38649774e-01
8.05887163e-01 -5.21796823e-01 4.91556861e-02 -3.83538574e-01
-1.60696015e-01 4.48113918e-01 1.00108802e+00 5.32397449e-01
-8.41005445e-01 8.89272153e-01 6.25480843e+00 9.49393153e-01
-9.98788059e-01 4.22685057e-01 -1.19144050e-03 1.69382840e-01
-3.13101739e-01 -4.69203681e-01 -8.83068323e-01 8.13471258e-01
9.54477966e-01 -4.37348276e-01 -2.24903762e-01 6.05547071e-01
2.68143594e-01 -3.09515059e-01 -6.53155506e-01 1.22842252e+00
4.63094741e-01 -1.56756282e+00 1.75496697e-01 -1.30917132e-01
6.94137573e-01 -2.44843841e-01 -3.35749954e-01 -9.91220772e-02
1.16199153e-02 -8.23467970e-01 5.39365590e-01 5.92919469e-01
8.32407057e-01 -7.02125967e-01 1.15281618e+00 4.03450727e-01
-2.98386991e-01 5.21014631e-01 -8.48013833e-02 1.20507263e-01
1.48084089e-01 1.19368029e+00 -1.17403662e+00 6.02776587e-01
4.85915899e-01 6.95012629e-01 -7.04826951e-01 1.22240269e+00
-2.88800627e-01 7.84179091e-01 -1.13343066e-02 -1.97064221e-01
-1.00289926e-01 4.30219829e-01 6.55782640e-01 1.97175157e+00
3.38743746e-01 1.54003710e-01 -8.13164935e-02 1.89737573e-01
-1.16257995e-01 1.02952456e+00 -1.64706960e-01 -4.88484174e-01
3.38684171e-01 9.89990711e-01 -9.97738600e-01 -4.78258461e-01
-2.10755289e-01 1.22780490e+00 -3.39230783e-02 -1.41163826e-01
-3.78548205e-01 -6.24911666e-01 3.86213690e-01 3.65047127e-01
-1.50384113e-01 5.73408455e-02 -5.30401886e-01 -1.24955034e+00
-1.39336213e-01 -9.92294908e-01 6.51975274e-01 -1.36178821e-01
-1.41452646e+00 5.27603030e-01 -4.98087794e-01 -1.48848546e+00
5.34431562e-02 -5.04571974e-01 -1.41131908e-01 7.28547633e-01
-1.52589810e+00 -6.39698327e-01 -4.57922183e-02 3.39472234e-01
4.78086859e-01 -3.38608265e-01 1.16794336e+00 5.98281085e-01
-3.33512962e-01 8.73888135e-01 1.97787941e-01 1.44315839e-01
1.31256986e+00 -1.25051379e+00 2.66111165e-01 7.74704039e-01
-5.57731502e-02 9.17076170e-01 1.15062106e+00 -1.20949769e+00
-7.82162786e-01 -1.18301582e+00 2.15553808e+00 -5.92778385e-01
6.86728656e-01 1.88112810e-01 -1.13942516e+00 1.19855218e-01
1.55481145e-01 -3.90223235e-01 1.57424688e+00 -1.18592799e-01
-2.13557675e-01 2.70045221e-01 -1.30024087e+00 5.48983276e-01
8.54797602e-01 -4.16554034e-01 -8.75779033e-01 6.78733885e-01
4.93813127e-01 -6.87551618e-01 -7.78836370e-01 -2.34232619e-02
3.09861839e-01 -3.35006833e-01 6.30574524e-01 -6.23865485e-01
6.17279947e-01 -2.36041620e-01 3.49492550e-01 -1.28491652e+00
-3.24678689e-01 -8.49732399e-01 2.62995541e-01 1.10585642e+00
7.56145835e-01 -4.20278966e-01 3.64666849e-01 3.99285197e-01
-2.85114318e-01 -1.97912842e-01 -6.61145091e-01 -5.28279364e-01
7.46109337e-03 -4.50282782e-01 2.58709967e-01 1.30688310e+00
8.67118776e-01 -5.50420098e-02 -2.43138164e-01 2.16008723e-02
3.26673537e-01 -3.68907064e-01 3.03130090e-01 -1.28770173e+00
-5.98838814e-02 -4.39339608e-01 -4.03371513e-01 -1.82013541e-01
-1.75623611e-01 -1.19566214e+00 -5.52406535e-02 -1.49985707e+00
3.21504682e-01 -5.18068224e-02 -3.57478857e-01 3.84745777e-01
-4.31196749e-01 4.04146314e-01 -1.70797080e-01 3.22096705e-01
-6.26583695e-01 -2.82253325e-01 9.28233802e-01 5.34216054e-02
-3.69131863e-01 -1.82735965e-01 -8.39065075e-01 9.99680996e-01
1.10927379e+00 -1.13776207e+00 2.10515067e-01 -1.52566746e-01
8.85318100e-01 -3.58113766e-01 -3.14542860e-01 -7.39605904e-01
3.63349587e-01 -4.55319375e-01 -1.85846835e-02 -4.67785090e-01
-7.23578393e-01 -4.09209311e-01 1.06473818e-01 5.81574261e-01
-4.83262360e-01 2.88110644e-01 3.04816723e-01 3.29297572e-01
-5.61447367e-02 -6.68849647e-01 6.41007006e-01 -4.70914841e-01
-3.60915542e-01 -1.36072189e-01 -1.10114813e+00 3.23894650e-01
7.88266897e-01 -1.40859142e-01 -1.38093457e-01 2.92439193e-01
-4.72098053e-01 -9.42320153e-02 6.19623482e-01 3.52724582e-01
4.72341537e-01 -9.46088016e-01 -9.80909765e-01 -9.76249948e-03
5.75849414e-01 -2.72367060e-01 3.18659395e-01 8.41016710e-01
-1.05727232e+00 2.62384504e-01 -8.05900469e-02 -1.57808304e-01
-1.84993660e+00 4.88456458e-01 -4.81991135e-02 -4.58507925e-01
-6.00968540e-01 7.95999348e-01 -4.79312420e-01 -1.29311517e-01
2.87846792e-02 -3.14169586e-01 -5.76890409e-01 3.60212743e-01
1.13522243e+00 4.20608342e-01 6.72475040e-01 -6.36608243e-01
-5.49395680e-01 2.36676455e-01 -3.65717739e-01 -1.35867447e-01
1.09902704e+00 -2.47516751e-01 -4.48418558e-01 6.97862089e-01
8.59367311e-01 6.16620302e-01 2.68371105e-01 -2.86444366e-01
5.04594147e-01 -4.16888863e-01 5.20547144e-02 -1.06462634e+00
-3.33995849e-01 2.70189315e-01 2.00073555e-01 2.83940006e-02
7.50883222e-01 -2.03923896e-01 7.93601871e-01 5.98602653e-01
1.09003045e-01 -1.61892116e+00 -2.56970823e-01 7.81704426e-01
7.55520642e-01 -1.28617370e+00 3.04045901e-02 -6.05135441e-01
-5.74422836e-01 1.18359244e+00 -3.73317599e-02 3.92741293e-01
8.04744899e-01 3.62393588e-01 4.02970463e-01 -1.79594815e-01
-1.04751207e-01 -7.79524222e-02 4.45283681e-01 3.93422872e-01
1.01528299e+00 1.81000363e-02 -1.35042202e+00 9.43348229e-01
-4.78040367e-01 3.57043028e-01 7.74503291e-01 1.23919725e+00
-3.27096522e-01 -1.31842303e+00 -4.90866244e-01 6.17221475e-01
-1.48511744e+00 -6.85794592e-01 -2.73804009e-01 1.70186490e-01
2.10787863e-01 1.01109827e+00 -3.33401471e-01 -9.41394567e-02
3.09830427e-01 1.29213944e-01 2.87213892e-01 -1.08521783e+00
-1.19228745e+00 -5.16658463e-02 1.60157725e-01 -2.26861551e-01
-6.04945719e-01 -8.58097732e-01 -1.38839614e+00 -2.55992830e-01
-2.80691355e-01 4.96568620e-01 6.60298765e-01 1.06215787e+00
4.43260700e-01 5.65415978e-01 1.13910221e-01 1.07795950e-02
4.45875991e-03 -9.92396295e-01 -4.17051166e-01 7.24234819e-01
4.02161062e-01 -6.66415617e-02 -2.23294079e-01 4.64735508e-01] | [8.63771915435791, 8.842308044433594] |
ac4c5cc4-b78c-4cad-8a3f-b61fa7f6faaf | identifying-protein-protein-interactions-in | null | null | https://aclanthology.org/I17-2041 | https://aclanthology.org/I17-2041.pdf | Identifying Protein-protein Interactions in Biomedical Literature using Recurrent Neural Networks with Long Short-Term Memory | In this paper, we propose a recurrent neural network model for identifying protein-protein interactions in biomedical literature. Experiments on two largest public benchmark datasets, AIMed and BioInfer, demonstrate that our approach significantly surpasses state-of-the-art methods with relative improvements of 10{\%} and 18{\%}, respectively. Cross-corpus evaluation also demonstrate that the proposed model remains robust despite using different training data. These results suggest that RNN can effectively capture semantic relationships among proteins as well as generalizes over different corpora, without any feature engineering. | ['Wen-Lian Hsu', 'Nai-Wen Chang', 'Yung-Chun Chang', 'Yu-Lun Hsieh'] | 2017-11-01 | identifying-protein-protein-interactions-in-1 | https://aclanthology.org/I17-2041 | https://aclanthology.org/I17-2041.pdf | ijcnlp-2017-11 | ['cross-corpus'] | ['computer-vision'] | [ 4.43171740e-01 2.18348298e-02 -1.72229409e-01 -5.31418204e-01
-7.56315291e-01 -2.03844845e-01 1.23899095e-01 3.74095291e-01
-4.88807797e-01 1.20601821e+00 1.51967006e-02 -3.62612426e-01
-1.96783125e-01 -4.15689111e-01 -9.22919810e-01 -6.80864573e-01
-2.41581593e-02 5.72398841e-01 -3.50980759e-02 -1.97364479e-01
-1.29022747e-02 4.10713196e-01 -1.17274332e+00 3.86109173e-01
8.17074299e-01 6.05154872e-01 1.10429995e-01 2.83964992e-01
-2.16562405e-01 7.70998240e-01 -8.86299193e-01 -5.20063460e-01
-3.90153557e-01 -3.22115421e-01 -1.03800249e+00 -4.82256740e-01
6.85597956e-02 2.98205465e-01 -3.94844264e-01 9.62965488e-01
6.27566278e-01 8.92530382e-02 6.31337702e-01 -6.02078140e-01
-1.19704926e+00 6.92010641e-01 -4.14337158e-01 2.99308658e-01
2.26919726e-01 -1.94490716e-01 1.06155419e+00 -9.49400842e-01
1.11424220e+00 1.14319158e+00 7.62805998e-01 8.90027940e-01
-1.28761101e+00 -9.37422037e-01 -5.91951888e-04 1.33925900e-01
-1.40278614e+00 -2.23398820e-01 5.46957850e-01 -8.54070187e-02
1.75556779e+00 1.91445291e-01 -2.12565243e-01 1.46563148e+00
3.15294653e-01 7.70424366e-01 7.18257070e-01 -2.47801661e-01
-1.04870595e-01 -5.29658981e-02 6.67318165e-01 4.90962893e-01
3.74178201e-01 -3.69447380e-01 -4.63711143e-01 -4.22104478e-01
3.85282099e-01 6.76018223e-02 -4.76010740e-01 -2.56334562e-02
-1.19437385e+00 6.42615438e-01 1.80640638e-01 6.99910283e-01
-5.40144801e-01 -1.68816477e-01 8.54210377e-01 2.14568734e-01
7.95160890e-01 5.65545201e-01 -8.37490559e-01 2.04249639e-02
-3.98295164e-01 2.79630926e-02 8.76417994e-01 1.06374300e+00
1.85162887e-01 -1.24574825e-01 -1.92829847e-01 1.31752849e+00
6.30461331e-03 3.27856779e-01 7.92533517e-01 -4.20058489e-01
4.66296852e-01 6.82925522e-01 -7.18999580e-02 -9.58654940e-01
-7.89457262e-01 -5.76626122e-01 -1.27028751e+00 -6.56971931e-01
1.14796735e-01 -9.69531015e-02 -8.33183885e-01 1.81822479e+00
4.27101217e-02 4.60586488e-01 6.66805029e-01 4.54152644e-01
1.59327126e+00 4.74431038e-01 5.82666457e-01 -3.05468321e-01
1.36968958e+00 -8.40862691e-01 -9.20079231e-01 -4.10491508e-03
7.47669339e-01 -6.41558230e-01 6.29087806e-01 1.89981446e-01
-9.95509088e-01 -4.23026204e-01 -9.95013475e-01 -1.61313221e-01
-7.09078074e-01 2.62715429e-01 8.64708602e-01 2.62564123e-01
-8.31609666e-01 1.00885046e+00 -6.94244266e-01 -5.42336285e-01
4.32939351e-01 6.58575058e-01 -5.30431986e-01 1.19440354e-01
-1.52140057e+00 8.53665292e-01 5.69577932e-01 2.53077954e-01
-2.69882143e-01 -9.00220633e-01 -6.65021658e-01 1.10499673e-01
6.96958788e-03 -6.77988589e-01 9.29403126e-01 -3.07560146e-01
-1.33692622e+00 9.20356989e-01 -4.12279487e-01 -7.49207079e-01
-1.76275820e-01 -3.58942777e-01 -7.06348896e-01 1.23963550e-01
1.00813054e-01 4.95217651e-01 -2.42621213e-01 -8.47187161e-01
-1.77254796e-01 -1.80689767e-01 -1.97368756e-01 -3.12466957e-02
-4.33599442e-01 3.05234522e-01 -4.87306356e-01 -7.99033701e-01
-1.25721917e-01 -9.08001423e-01 -3.85421544e-01 -5.27255595e-01
-6.10576332e-01 -7.71142840e-01 4.75988567e-01 -5.96852720e-01
1.02671373e+00 -1.69371867e+00 2.51583070e-01 -1.20047126e-02
3.44435304e-01 5.95110297e-01 -3.95210922e-01 4.79634315e-01
-6.67587101e-01 1.22622348e-01 -2.04374105e-01 -9.63408574e-02
-3.00319135e-01 2.68239021e-01 -1.11556664e-01 3.81881058e-01
3.71505380e-01 1.05157959e+00 -7.93002248e-01 -2.60461569e-01
-1.74492508e-01 9.28757191e-01 -1.11238211e-01 8.50447938e-02
-3.07698578e-01 3.66952270e-01 -7.17568994e-01 6.76984966e-01
6.05509877e-01 -1.10016024e+00 7.81973422e-01 -3.47716630e-01
4.40929353e-01 4.48695898e-01 -4.25743371e-01 1.77087116e+00
-1.09860867e-01 2.74785578e-01 -4.33512211e-01 -1.47640085e+00
1.19304132e+00 5.66364229e-01 6.21643007e-01 -5.32441199e-01
1.34043783e-01 2.02733710e-01 -1.16013549e-02 -6.08299315e-01
4.36359644e-02 1.28486753e-01 1.03054471e-01 3.25911671e-01
1.71296552e-01 7.94083416e-01 2.76269794e-01 4.78359237e-02
1.34391427e+00 9.62943807e-02 3.08747530e-01 -3.46728951e-01
8.84651899e-01 -2.99437106e-01 9.30494547e-01 5.98159015e-01
-6.83583366e-03 1.84273303e-01 4.09981996e-01 -6.12387598e-01
-6.80815816e-01 -6.50944293e-01 -2.05621079e-01 1.18786407e+00
1.09702751e-01 -4.18801785e-01 -8.81147861e-01 -9.03150499e-01
-4.91472334e-02 3.21628541e-01 -7.33726561e-01 -9.50203985e-02
-6.82493269e-01 -1.41752005e+00 8.07098746e-01 6.52538955e-01
3.45731378e-01 -1.23178184e+00 -7.39039034e-02 4.13386792e-01
-2.33030677e-01 -1.38262153e+00 -3.63476157e-01 3.53602588e-01
-8.37587357e-01 -1.24765015e+00 -7.49226570e-01 -1.07727087e+00
6.78138852e-01 2.95966893e-01 1.43278348e+00 1.12417452e-01
-3.71546209e-01 -3.42080683e-01 -3.56861085e-01 -3.66716862e-01
-4.01011854e-01 3.21236730e-01 2.22342350e-02 -5.31297266e-01
1.13315427e+00 -6.38461053e-01 -5.10590196e-01 2.90312529e-01
-7.04338551e-01 -1.99962750e-01 8.60839188e-01 1.32691014e+00
7.17411995e-01 -2.93227345e-01 1.14689982e+00 -1.48475790e+00
9.09391105e-01 -5.98611474e-01 -1.61698684e-01 5.26830733e-01
-7.45559156e-01 2.87636578e-01 8.26423883e-01 -3.99133205e-01
-9.67958391e-01 -8.30300674e-02 -1.72122687e-01 -2.22889572e-01
-2.31228024e-01 7.29658067e-01 -6.75327778e-02 1.23215936e-01
6.60027683e-01 1.95047185e-01 -9.82990041e-02 -7.13322639e-01
1.23192362e-01 7.04672813e-01 4.38037783e-01 -4.86693293e-01
1.09999776e-01 2.54269153e-01 -1.39421551e-02 -4.46109086e-01
-8.18610191e-01 -6.02452278e-01 -4.72786218e-01 6.27387464e-01
7.90515780e-01 -1.00113904e+00 -1.26156771e+00 7.20548630e-02
-1.43936121e+00 3.57808709e-01 3.60623360e-01 5.52509069e-01
-1.68410361e-01 4.22043413e-01 -1.22828829e+00 -3.26042533e-01
-1.12171865e+00 -1.08571649e+00 9.59459662e-01 1.41010523e-01
-5.07926226e-01 -9.86916840e-01 8.47994257e-03 5.45775950e-01
2.28575140e-01 2.67685115e-01 1.04959524e+00 -1.32385135e+00
-1.43615052e-01 -1.44427866e-01 -4.68046665e-01 1.38872907e-01
4.75024313e-01 -3.22184056e-01 -9.39434528e-01 -3.23202640e-01
-3.38265032e-01 -3.36344451e-01 9.83591735e-01 2.52380401e-01
1.48526847e+00 -2.18865439e-01 -8.72940183e-01 4.83835757e-01
1.07110691e+00 4.00088698e-01 4.73695844e-01 1.63600489e-01
6.05897069e-01 5.05692363e-01 4.18069631e-01 -4.45496179e-02
1.41884685e-01 3.70774776e-01 4.81597297e-02 -4.52982813e-01
1.53020397e-01 1.13799952e-01 5.92604764e-02 1.09545743e+00
-2.13703021e-01 -2.44829431e-01 -9.03095305e-01 3.09536964e-01
-2.16776919e+00 -8.18192661e-01 4.74034250e-02 1.64157343e+00
1.25323403e+00 -2.95973197e-02 -1.42439365e-01 -3.73588502e-01
7.60462642e-01 -2.56061465e-01 -8.45724583e-01 -3.68069977e-01
-3.70294392e-01 6.91046536e-01 5.48695982e-01 6.50514960e-02
-1.14653218e+00 1.16798139e+00 7.37718582e+00 8.73471081e-01
-8.43946636e-01 -4.53715213e-02 9.26831186e-01 1.31634295e-01
-3.29965353e-02 -6.47436917e-01 -9.39512014e-01 2.85320103e-01
1.29535544e+00 -2.39299208e-01 -2.56170779e-02 8.06921303e-01
7.98950046e-02 7.44291306e-01 -1.05369437e+00 7.79676974e-01
2.90232718e-01 -1.79784811e+00 1.65404841e-01 -1.97136417e-01
3.97357643e-01 3.78407747e-01 -1.77566856e-01 3.57068330e-01
4.89636958e-01 -1.39169943e+00 -4.32482898e-01 5.34866154e-01
5.81720769e-01 -8.17412019e-01 1.28930688e+00 4.94901575e-02
-1.07735848e+00 5.29259503e-01 -5.61756551e-01 3.26522768e-01
-3.24757516e-01 7.32251108e-01 -1.07725668e+00 1.03718555e+00
8.92279267e-01 1.23793983e+00 -5.59378803e-01 5.07862985e-01
1.17811142e-02 4.18020725e-01 9.22959670e-02 -4.81914401e-01
2.52978150e-02 -1.09071277e-01 1.02815688e-01 1.65355253e+00
-1.76101789e-01 2.24367082e-01 1.66809544e-01 8.70822668e-01
-6.22495055e-01 4.24175262e-01 -4.55652386e-01 -2.81921297e-01
5.02372026e-01 1.10082054e+00 -6.24549150e-01 -4.47333217e-01
-3.76354128e-01 9.38599706e-01 6.34711504e-01 6.82214677e-01
-8.29571903e-01 -7.60471106e-01 8.30034316e-01 -7.20874548e-01
1.87475666e-01 2.89472073e-01 1.32499680e-01 -9.93260324e-01
7.98775181e-02 -1.06261647e+00 4.52901512e-01 -4.81923968e-01
-1.86552763e+00 9.50040936e-01 -3.58490318e-01 -9.15313661e-01
-1.61143765e-01 -9.39907432e-01 -3.56950089e-02 9.42082405e-01
-1.46429455e+00 -1.18762052e+00 2.81634450e-01 1.95525780e-01
4.71318930e-01 -5.87145627e-01 1.46045148e+00 5.18942535e-01
-1.08095503e+00 8.69055033e-01 5.08408070e-01 2.92110473e-01
8.34654629e-01 -9.17735279e-01 7.29327083e-01 2.26290926e-01
-1.44431219e-01 1.38437307e+00 6.71089888e-01 -5.28296828e-01
-1.34733939e+00 -1.22786653e+00 1.17257965e+00 -2.53555626e-01
6.82581127e-01 -3.10785383e-01 -1.40147913e+00 6.36260569e-01
3.45241249e-01 -1.00078344e-01 1.25277555e+00 6.16618812e-01
-5.17597258e-01 1.86182931e-01 -1.13394833e+00 5.72793067e-01
1.27351403e+00 -5.35016775e-01 -4.82048720e-01 7.43693233e-01
1.04934967e+00 -4.42524612e-01 -1.27618682e+00 9.05097485e-01
4.85161394e-01 -5.27244985e-01 1.33957708e+00 -1.19580221e+00
3.74052078e-01 2.95524467e-02 1.22992441e-01 -9.51322854e-01
-4.79767889e-01 -5.76115489e-01 -2.22349375e-01 1.10038269e+00
7.24912107e-01 -7.94521749e-01 9.14332509e-01 3.70112538e-01
4.54539172e-02 -1.21129096e+00 -8.06222677e-01 -5.59419692e-01
4.35187519e-01 -5.03796376e-02 6.18235767e-01 1.36854219e+00
2.20457941e-01 8.09837699e-01 -2.71217048e-01 -5.05222473e-04
3.28553647e-01 9.40605327e-02 3.55576545e-01 -1.14471996e+00
-4.32553440e-01 -4.34220821e-01 -3.96000475e-01 -9.68690097e-01
9.55669641e-01 -1.08297515e+00 -1.96292847e-01 -1.42561650e+00
6.07474864e-01 -7.05268010e-02 -1.03322637e+00 8.38717163e-01
-4.55472618e-01 2.21447706e-01 -5.74887753e-01 5.06191514e-02
-6.03981614e-01 2.96845287e-01 8.40162992e-01 -5.32795310e-01
5.43343835e-03 -3.38290542e-01 -1.06958795e+00 6.15614772e-01
8.01192284e-01 -4.12834764e-01 -3.20064843e-01 -2.25075558e-01
6.96664909e-03 -2.27651477e-01 -1.31817639e-01 -4.70643044e-01
8.75419676e-02 1.53324958e-02 5.07854998e-01 -6.77495599e-01
2.29728401e-01 -3.08676898e-01 3.91681716e-02 4.52769727e-01
-8.09162557e-01 2.39182755e-01 5.77220380e-01 8.51463437e-01
-1.67184826e-02 -2.13960111e-02 5.06011248e-01 -6.68143257e-02
-2.16703326e-01 1.75541505e-01 -1.47181243e-01 1.15105687e-02
7.28367627e-01 1.62098646e-01 -5.14651835e-01 7.19301105e-02
-6.48418367e-01 3.48595858e-01 -9.63881612e-02 6.06994033e-01
5.66718876e-01 -9.86617267e-01 -8.11530054e-01 -8.32568482e-02
2.10031360e-01 -2.85534859e-01 4.58992720e-02 4.68434751e-01
-3.77438873e-01 1.02255630e+00 1.80925265e-01 -4.81877357e-01
-1.70098341e+00 5.57624698e-01 3.39489222e-01 -6.16262019e-01
-7.39743233e-01 1.01776838e+00 3.05760026e-01 -7.38871634e-01
3.73752773e-01 -1.30230233e-01 -5.35468400e-01 -4.00444657e-01
5.23888767e-01 -3.58132310e-02 1.57117829e-01 -5.39757252e-01
-6.22330368e-01 4.77165282e-01 -7.00664937e-01 7.13530302e-01
1.59381390e+00 2.58567780e-01 -4.66972977e-01 3.64963025e-01
1.48234773e+00 -4.89959359e-01 -4.07126725e-01 -4.96520340e-01
4.38541710e-01 1.68246388e-01 -4.91100013e-01 -1.10276425e+00
-1.09538758e+00 5.07910907e-01 5.26496291e-01 -2.87175685e-01
8.47104967e-01 1.47063732e-02 1.06018519e+00 1.04120052e+00
1.01563714e-01 -9.33846354e-01 -2.88455933e-01 4.92103308e-01
5.85810065e-01 -1.14897788e+00 1.62990406e-01 -6.91601515e-01
-4.26146686e-01 9.70332563e-01 5.00395715e-01 -5.02799079e-02
5.15759528e-01 1.80846304e-01 1.46148622e-01 -4.41083491e-01
-9.69703853e-01 1.01930268e-01 3.83218318e-01 3.53876948e-01
1.39672637e+00 -6.58070073e-02 -6.16551399e-01 7.35753715e-01
1.60498217e-01 2.44956672e-01 6.31736442e-02 7.92580426e-01
1.19083099e-01 -1.47400093e+00 1.53189763e-01 4.93817329e-01
-9.54397857e-01 -4.04643506e-01 -6.44738317e-01 7.20116377e-01
-1.99172810e-01 9.29371595e-01 -2.72624880e-01 -2.54897505e-01
4.29125518e-01 2.77704358e-01 2.19890341e-01 -5.51444829e-01
-8.29122245e-01 8.45811442e-02 4.14726436e-01 -5.69857895e-01
-9.21851516e-01 -3.28104973e-01 -1.64911389e+00 -8.27009901e-02
-4.40836430e-01 6.90108061e-01 2.28964344e-01 9.29717183e-01
1.12066782e+00 1.06194961e+00 1.86771721e-01 -2.35569566e-01
-4.09715176e-01 -1.29987073e+00 -2.54391968e-01 5.58861792e-01
-1.73661057e-02 -6.14169419e-01 5.65596856e-03 1.34324208e-01] | [4.775875568389893, 5.748755931854248] |
83ed9ccc-1ec8-42cf-98aa-f42029cc3165 | recent-advances-in-optimal-transport-for | 2306.16156 | null | https://arxiv.org/abs/2306.16156v1 | https://arxiv.org/pdf/2306.16156v1.pdf | Recent Advances in Optimal Transport for Machine Learning | Recently, Optimal Transport has been proposed as a probabilistic framework in Machine Learning for comparing and manipulating probability distributions. This is rooted in its rich history and theory, and has offered new solutions to different problems in machine learning, such as generative modeling and transfer learning. In this survey we explore contributions of Optimal Transport for Machine Learning over the period 2012 -- 2022, focusing on four sub-fields of Machine Learning: supervised, unsupervised, transfer and reinforcement learning. We further highlight the recent development in computational Optimal Transport, and its interplay with Machine Learning practice. | ['Antoine Souloumiac', 'Fred Ngolè Mboula', 'Eduardo Fernandes Montesuma'] | 2023-06-28 | null | null | null | null | ['transfer-learning'] | ['miscellaneous'] | [ 7.19665885e-02 -2.25734785e-01 -9.17534530e-01 -2.98777550e-01
-7.74743497e-01 -4.55956131e-01 1.04699230e+00 4.41218242e-02
-4.27029938e-01 1.24872911e+00 7.85360336e-02 -3.65082264e-01
-6.47098660e-01 -9.88816857e-01 -8.81254733e-01 -1.01680171e+00
-4.83504146e-01 6.62763953e-01 1.63999334e-01 5.63498493e-03
3.16563606e-01 4.90940660e-01 -1.44156826e+00 3.59488755e-01
7.35718310e-01 1.03359663e+00 1.27926990e-01 7.46168435e-01
-2.49927044e-01 7.52605259e-01 -1.95192188e-01 -4.03939039e-01
-3.78095567e-01 -5.07305145e-01 -9.19123232e-01 -3.02374125e-01
-3.16189863e-02 9.60148573e-02 -7.80027866e-01 7.89687216e-01
3.56694192e-01 2.88790464e-01 1.43820643e+00 -1.96038973e+00
-7.69227862e-01 7.31120646e-01 -2.03371532e-02 5.04761219e-01
1.41845837e-01 1.48420408e-01 9.48553085e-01 -1.13630772e-01
3.72642517e-01 1.35805261e+00 6.25614107e-01 6.54849350e-01
-1.33389676e+00 -3.32445830e-01 -2.11104184e-01 7.46779501e-01
-1.00699437e+00 -6.85642287e-02 5.57629168e-01 -7.11873710e-01
8.45487297e-01 1.62790753e-02 7.68952250e-01 1.32426655e+00
7.99599528e-01 1.26776969e+00 1.30636823e+00 -6.17547691e-01
6.16778910e-01 2.54116237e-01 -4.47305381e-01 5.30300021e-01
-2.97475666e-01 5.50626755e-01 -7.35192597e-01 -1.33641645e-01
4.97959346e-01 -3.49542141e-01 2.20627993e-01 -2.23870307e-01
-1.17864227e+00 1.30323470e+00 4.60247576e-01 1.02768838e-01
-1.87760487e-01 9.10672963e-01 3.59742016e-01 2.12007657e-01
7.55112529e-01 1.43317863e-01 -6.87706113e-01 -3.64410490e-01
-7.59954512e-01 5.03041923e-01 9.29232299e-01 8.90836239e-01
9.15702701e-01 5.44581525e-02 -2.75668234e-01 8.87712896e-01
7.05533981e-01 9.91989553e-01 1.44274220e-01 -1.16933882e+00
3.96085680e-01 -1.51450142e-01 4.61317897e-02 -3.42018217e-01
-1.09382816e-01 6.23601414e-02 -7.78071523e-01 1.08232416e-01
4.07922208e-01 -3.48368704e-01 -6.58495486e-01 1.46731436e+00
3.19650978e-01 4.18083727e-01 1.10989027e-01 3.53475958e-01
3.73642892e-01 8.36927712e-01 2.59635031e-01 5.64761683e-02
7.46738851e-01 -7.75138855e-01 -5.50581753e-01 1.27074808e-01
4.62555110e-01 -7.25093544e-01 5.55223048e-01 1.01543747e-01
-6.78290367e-01 -2.67208181e-02 -6.23186707e-01 3.62893045e-01
-8.62908900e-01 -4.77113426e-01 8.95510316e-01 9.32243764e-01
-8.75957131e-01 1.24326169e+00 -1.04167092e+00 -4.18039531e-01
9.45202947e-01 2.78293937e-01 5.57156466e-02 -1.68192402e-01
-1.38962770e+00 1.18590677e+00 3.15154493e-01 -2.24574909e-01
-1.16297972e+00 -9.52950954e-01 -8.83141518e-01 -5.50674736e-01
-5.64623252e-02 -8.09882462e-01 1.39393413e+00 -2.38012329e-01
-1.89349389e+00 3.84290755e-01 -9.13843960e-02 -6.10177338e-01
5.87546527e-01 -3.19923460e-01 -2.34617263e-01 -3.53266507e-01
5.49911186e-02 7.36785352e-01 7.75148869e-01 -9.22154546e-01
-9.16516185e-01 -7.22977445e-02 -1.23245053e-01 -2.39115745e-01
1.77125912e-02 7.21429067e-04 1.91423759e-01 -5.18416584e-01
-2.43839234e-01 -1.12476194e+00 -4.83174533e-01 1.91686656e-02
-2.68106759e-01 -7.02042580e-01 9.83482480e-01 -2.80897766e-01
5.21332264e-01 -1.51001096e+00 3.24423790e-01 1.37884542e-01
-3.22547376e-01 -2.90370256e-01 2.14322522e-01 6.48217618e-01
3.89252901e-01 1.58277407e-01 -6.46374583e-01 -4.43088472e-01
2.49204159e-01 6.90697134e-01 -2.72668600e-01 7.68025517e-01
8.95896181e-02 1.20677066e+00 -1.31047916e+00 -4.76217330e-01
7.79767871e-01 5.28120399e-01 -4.04826015e-01 9.54798907e-02
-5.87851703e-01 8.92983854e-01 -4.67817217e-01 6.32450223e-01
4.70869631e-01 2.18729377e-02 -1.77031040e-01 1.06722221e-01
-2.74975181e-01 3.23989660e-01 -7.79765427e-01 1.68013036e+00
-6.10557437e-01 9.49102581e-01 -4.06296313e-01 -1.26283669e+00
5.29778540e-01 2.17570752e-01 8.75988901e-01 -7.26335108e-01
-7.67135527e-03 2.37476006e-01 -2.25339428e-01 -7.25111306e-01
2.17743143e-01 -1.85331926e-01 -1.84751987e-01 6.48006737e-01
5.08238971e-01 -6.37298167e-01 1.63786098e-01 -2.49267310e-01
6.85190082e-01 5.68931997e-01 -4.71407212e-02 -5.77451646e-01
2.48662204e-01 8.72556791e-02 -8.61481577e-02 8.78320813e-01
-2.35997647e-01 6.29720837e-02 1.67574003e-01 -3.29537600e-01
-1.12037873e+00 -1.32034194e+00 -5.56250393e-01 1.25348437e+00
1.52919665e-01 -8.76853534e-04 -6.26761079e-01 -7.43437290e-01
4.91685480e-01 8.61194253e-01 -6.88872993e-01 -2.95807570e-01
-5.83101213e-01 -1.28214669e+00 7.43970692e-01 5.93913198e-01
3.13857973e-01 -1.26528275e+00 -9.29604769e-02 1.62763387e-01
-3.41812260e-02 -9.18933868e-01 -5.80848269e-02 2.93855339e-01
-1.02806497e+00 -1.01187980e+00 -6.50636256e-01 -2.44625703e-01
1.04437899e-02 -2.22423553e-01 1.13891256e+00 -4.68897402e-01
-3.15293252e-01 6.13958657e-01 -2.21620589e-01 -6.76800728e-01
-8.49574089e-01 4.96714503e-01 6.59975931e-02 -2.35825986e-01
6.55996427e-02 -6.11976564e-01 -4.22214121e-01 3.44344467e-01
-7.22671807e-01 -3.69084984e-01 4.93830532e-01 6.14183426e-01
3.63800347e-01 4.78417613e-02 5.04495680e-01 -5.56187570e-01
5.99446774e-01 -1.10266519e+00 -7.50455320e-01 1.30896196e-01
-7.41563261e-01 4.99160558e-01 2.17631131e-01 -1.44699961e-01
-1.02295852e+00 -3.87201160e-01 -3.20568144e-01 -1.91946462e-01
-4.83994275e-01 2.41197273e-01 4.88727987e-02 -1.72977820e-01
5.08943141e-01 2.43108824e-01 1.73496172e-01 -3.67505521e-01
7.95415282e-01 6.32125974e-01 4.36899476e-02 -1.06959164e+00
7.00620532e-01 6.79069221e-01 4.48732585e-01 -9.43318546e-01
-9.91234481e-01 -3.04899752e-01 -1.10380697e+00 -6.51567519e-01
9.05032635e-01 -3.07120144e-01 -6.51933134e-01 7.77004242e-01
-1.09110701e+00 -9.67574775e-01 -3.68146896e-01 8.30527663e-01
-1.05285203e+00 -1.40749678e-01 -3.82037669e-01 -1.00148416e+00
7.79310092e-02 -1.25387323e+00 1.02227139e+00 4.26311404e-01
-4.44784528e-03 -2.01724696e+00 5.46844423e-01 5.24153821e-02
7.35491395e-01 1.83834523e-01 1.06297314e+00 -2.01855868e-01
-6.28009737e-01 2.19562054e-01 2.74995416e-02 5.03057122e-01
1.76401287e-01 1.41924262e-01 -1.12200212e+00 2.41826978e-02
-6.97355747e-01 -2.32487306e-01 1.07902586e+00 7.03427732e-01
1.06421685e+00 -1.86885104e-01 -7.25638270e-01 4.68769580e-01
1.33186781e+00 3.33389416e-02 4.71553117e-01 3.00378889e-01
6.13134384e-01 6.56789243e-01 3.26649129e-01 2.22602293e-01
5.21857202e-01 4.76730287e-01 6.09319806e-01 3.21792185e-01
-1.57149076e-01 -3.44581336e-01 2.47985169e-01 6.07324719e-01
-2.07054719e-01 -3.40117633e-01 -8.98045838e-01 5.39815903e-01
-1.95856285e+00 -1.17609560e+00 -2.19910294e-01 2.00989914e+00
6.32589996e-01 -1.47878304e-01 1.09391436e-01 9.69011337e-02
7.73894966e-01 2.85916150e-01 -4.89695489e-01 -3.68661463e-01
-5.20736165e-02 1.93121314e-01 7.95451581e-01 4.55832958e-01
-1.29935443e+00 9.31033075e-01 8.38790607e+00 1.16952956e+00
-1.03167307e+00 3.17529261e-01 4.62572098e-01 3.44403207e-01
-3.25909823e-01 -1.79243088e-01 -8.33125651e-01 8.08160484e-01
1.28865182e+00 -3.16155814e-02 8.64316642e-01 6.79709613e-01
2.22281694e-01 -3.45146924e-01 -1.31968486e+00 5.43808043e-01
-1.52206466e-01 -1.68202090e+00 -2.50010751e-02 1.09081760e-01
1.07843494e+00 7.68278599e-01 2.69463360e-01 4.72022772e-01
7.12563634e-01 -1.22026825e+00 8.58658671e-01 7.83666909e-01
4.99862999e-01 -6.96809351e-01 7.27404177e-01 5.17851114e-01
-9.83836234e-01 -1.20525230e-02 -3.86740148e-01 2.02020019e-01
4.16633129e-01 6.18812680e-01 -6.26542687e-01 6.72294140e-01
8.24328601e-01 1.04863703e+00 -3.65174673e-02 1.36685479e+00
-2.20435232e-01 9.11166966e-01 -3.33091468e-01 -3.97722542e-01
4.19368446e-01 -2.66270369e-01 5.96311033e-01 1.50882709e+00
2.26580307e-01 -7.71637380e-01 1.04927137e-01 1.06655908e+00
-7.25829154e-02 -1.95549488e-01 -7.51994252e-01 6.92545110e-03
3.84095192e-01 7.96406448e-01 -5.30493438e-01 4.47567217e-02
-1.95883721e-01 5.42045057e-01 1.57053038e-01 2.54229546e-01
-1.14515901e+00 1.00932598e-01 7.68964589e-01 -5.89428619e-02
5.26682027e-02 -5.72857559e-01 -1.16478838e-01 -6.07326686e-01
-6.42447889e-01 -9.80464648e-03 5.89595020e-01 -5.10252595e-01
-1.67915452e+00 1.21495076e-01 8.44169796e-01 -8.16515088e-01
-3.31504285e-01 -9.78105426e-01 -5.69893479e-01 7.30974376e-01
-1.77034545e+00 -1.12562335e+00 9.92071629e-02 4.88792956e-01
5.43438852e-01 -3.71335655e-01 7.38245606e-01 3.16667050e-01
-2.51555741e-01 2.45646819e-01 7.40483344e-01 -1.28849283e-01
4.03155506e-01 -1.53037322e+00 3.68388444e-01 1.03911534e-01
2.37010807e-01 -3.59183513e-02 5.99522591e-01 -4.11114216e-01
-1.33252203e+00 -1.44430935e+00 5.51062167e-01 -6.50822520e-01
9.95910347e-01 -2.78952152e-01 -3.37860882e-01 5.44188678e-01
3.29129070e-01 1.50787458e-01 6.31255805e-01 1.78103652e-02
-5.58814928e-02 -4.56796922e-02 -1.22443831e+00 2.82909065e-01
8.60699594e-01 -6.98395073e-01 -4.63804044e-02 6.23524666e-01
4.25817639e-01 -3.14977169e-01 -1.12271535e+00 3.74469489e-01
6.73756242e-01 -7.45855331e-01 9.54430282e-01 -9.03249204e-01
2.70328134e-01 1.42346367e-01 -4.82451230e-01 -1.77010763e+00
5.78330345e-02 -8.93267334e-01 -2.55929619e-01 9.71809864e-01
5.93254387e-01 -7.23136604e-01 8.90039086e-01 1.48434117e-01
-1.82417110e-01 -1.02732921e+00 -1.38402748e+00 -9.46004927e-01
8.36757481e-01 -1.15724576e+00 6.31695449e-01 4.98792350e-01
-3.49104822e-01 -1.06533408e-01 -3.54866236e-01 -8.35520253e-02
9.21485424e-01 -1.78618774e-01 5.27552545e-01 -1.20849502e+00
-4.16808911e-02 -7.24346459e-01 -4.90819901e-01 -1.00675118e+00
5.34935474e-01 -1.27183723e+00 2.19094262e-01 -1.81060898e+00
7.25608543e-02 -9.03637528e-01 -3.63010824e-01 1.82501957e-01
2.51226991e-01 1.12135634e-01 -3.40782881e-01 3.22052062e-01
-6.16500556e-01 8.97467315e-01 1.29705679e+00 -3.39299828e-01
1.49734139e-01 5.50875425e-01 1.06846942e-02 6.63513839e-01
1.10133100e+00 -9.24931705e-01 -1.50886193e-01 -3.90272945e-01
9.32362676e-03 -3.64637136e-01 6.36166751e-01 -8.00556183e-01
1.11001797e-01 -6.10630453e-01 6.24543056e-02 -4.80085462e-01
2.72708893e-01 -5.09977818e-01 -1.98530987e-01 4.10818875e-01
-2.78337657e-01 2.05256399e-02 2.23462299e-01 6.81208014e-01
1.62936449e-01 -3.45404953e-01 9.26675916e-01 -2.02209741e-01
-6.97289169e-01 5.64938128e-01 -7.54037321e-01 4.05435175e-01
1.06380355e+00 2.44095653e-01 -4.29252148e-01 -1.18466616e-01
-8.46202374e-01 1.13082953e-01 -1.80976942e-01 5.36276937e-01
1.86151668e-01 -1.41084313e+00 -8.44496846e-01 -2.21904129e-01
-1.32256299e-01 -4.18538898e-01 2.08323210e-01 1.02412856e+00
-2.59570062e-01 6.52912676e-01 -5.29971793e-02 -1.12319362e+00
-4.76725370e-01 1.76584885e-01 6.16401017e-01 -1.26570746e-01
-4.28815424e-01 6.21130526e-01 -4.75550324e-01 -6.33882105e-01
7.33413324e-02 -2.67204076e-01 -5.26073463e-02 -1.56537816e-01
1.20497064e-03 1.09078538e+00 2.68421676e-02 -5.58172703e-01
-3.15690458e-01 4.40985292e-01 4.18334693e-01 -3.66207749e-01
1.31213546e+00 -3.70164700e-02 -2.63079494e-01 1.04972482e+00
1.33071923e+00 -5.34324229e-01 -1.30861378e+00 -7.12682828e-02
2.86618978e-01 -1.15091659e-01 6.31716996e-02 -6.70942724e-01
-8.60066116e-01 9.21624601e-01 5.82206249e-01 4.50176716e-01
1.80556327e-01 4.33626354e-01 6.78618729e-01 6.19467318e-01
6.27361715e-01 -1.17703378e+00 1.49741650e-01 6.43817604e-01
7.64167011e-01 -1.49103987e+00 -1.19915642e-01 -1.30135462e-01
-2.01729149e-01 9.58893597e-01 1.68756619e-01 -1.63696811e-01
1.40055621e+00 4.45268333e-01 -2.01052412e-01 8.50216299e-02
-5.84977448e-01 -2.06495449e-01 4.29307967e-01 1.03486478e+00
2.77081221e-01 1.90471426e-01 4.26359713e-01 -4.60653976e-02
-2.93798298e-01 -4.23975326e-02 1.27202734e-01 9.20119882e-01
-5.40139318e-01 -1.35724318e+00 -1.82836369e-01 5.91335118e-01
-2.65539557e-01 7.45890960e-02 1.74921662e-01 8.41116548e-01
3.29776347e-01 9.80380356e-01 2.06419632e-01 -1.07043028e-01
4.06120643e-02 2.65362225e-02 1.01352978e+00 -3.39841872e-01
5.62751070e-02 -4.52964723e-01 -4.76125069e-02 -3.67985874e-01
-8.05148125e-01 -8.29474986e-01 -1.00167453e+00 -5.42930722e-01
-3.58656555e-01 4.33633119e-01 1.11142755e+00 1.48994994e+00
-5.77619746e-02 6.74077690e-01 5.06846488e-01 -1.23510849e+00
-4.55547422e-01 -9.99837399e-01 -5.90617180e-01 -1.33591503e-01
3.18040371e-01 -1.21932220e+00 -5.69572270e-01 1.10138260e-01] | [6.831689834594727, 3.6767516136169434] |
219f8020-ffbb-4521-acc9-cd6f163997bd | do-response-selection-models-really-know-what | 2009.04703 | null | https://arxiv.org/abs/2009.04703v2 | https://arxiv.org/pdf/2009.04703v2.pdf | Do Response Selection Models Really Know What's Next? Utterance Manipulation Strategies for Multi-turn Response Selection | In this paper, we study the task of selecting the optimal response given a user and system utterance history in retrieval-based multi-turn dialog systems. Recently, pre-trained language models (e.g., BERT, RoBERTa, and ELECTRA) showed significant improvements in various natural language processing tasks. This and similar response selection tasks can also be solved using such language models by formulating the tasks as dialog--response binary classification tasks. Although existing works using this approach successfully obtained state-of-the-art results, we observe that language models trained in this manner tend to make predictions based on the relatedness of history and candidates, ignoring the sequential nature of multi-turn dialog systems. This suggests that the response selection task alone is insufficient for learning temporal dependencies between utterances. To this end, we propose utterance manipulation strategies (UMS) to address this problem. Specifically, UMS consist of several strategies (i.e., insertion, deletion, and search), which aid the response selection model towards maintaining dialog coherence. Further, UMS are self-supervised methods that do not require additional annotation and thus can be easily incorporated into existing approaches. Extensive evaluation across multiple languages and models shows that UMS are highly effective in teaching dialog consistency, which leads to models pushing the state-of-the-art with significant margins on multiple public benchmark datasets. | ['Dong-hun Lee', 'Taesun Whang', 'Dongsuk Oh', 'Kijong Han', 'Chanhee Lee', 'Saebyeok Lee', 'Dongyub Lee'] | 2020-09-10 | null | null | null | null | ['conversational-response-selection'] | ['natural-language-processing'] | [-3.76399904e-02 9.71222147e-02 -4.50553149e-01 -7.60895193e-01
-7.30973184e-01 -5.95142007e-01 9.09180045e-01 3.24718714e-01
-3.48124892e-01 6.54198349e-01 2.77872950e-01 -4.69721079e-01
8.95800665e-02 -4.11132157e-01 -1.67921066e-01 -3.05812627e-01
3.32024544e-01 7.12527096e-01 5.10667324e-01 -8.40272784e-01
4.53561425e-01 9.57338735e-02 -1.23663366e+00 5.44984937e-01
8.84963810e-01 5.44230640e-01 2.96279639e-01 7.10476100e-01
-3.72335106e-01 1.19183338e+00 -6.35975182e-01 -2.50588030e-01
-2.77457148e-01 -6.77749336e-01 -1.32307112e+00 -3.19312960e-02
8.88610780e-02 -6.13357008e-01 -1.33359432e-01 3.54243785e-01
5.79539001e-01 4.67546880e-01 5.51084280e-01 -1.07315779e+00
-5.09640813e-01 8.50217223e-01 -1.31162360e-01 2.95083560e-02
7.95252442e-01 2.19979987e-01 1.22244954e+00 -9.08149362e-01
4.95117396e-01 1.74370694e+00 2.18621626e-01 7.43575454e-01
-1.42544067e+00 -3.31907183e-01 3.76170754e-01 1.45767570e-01
-9.54731941e-01 -6.86435163e-01 6.35338604e-01 -4.51693386e-01
1.25073206e+00 4.74065363e-01 9.27907452e-02 1.07495725e+00
2.40392387e-01 1.13377047e+00 1.06976724e+00 -7.27209210e-01
3.63269560e-02 6.81828916e-01 6.72172964e-01 4.67463613e-01
-7.10053921e-01 -2.93562651e-01 -6.35285556e-01 -3.10970634e-01
2.43134409e-01 -1.89799219e-01 -1.48765534e-01 3.30893472e-02
-9.27204430e-01 1.06997299e+00 1.07789300e-01 3.93777460e-01
-1.95868760e-01 -3.48654181e-01 3.96453261e-01 7.93982267e-01
6.07610762e-01 6.44354761e-01 -4.52325583e-01 -1.87705047e-02
-5.13149202e-01 4.00893986e-01 1.08131313e+00 8.72008383e-01
5.93099296e-01 -2.54296511e-01 -6.38867378e-01 1.15141940e+00
3.71265531e-01 8.87604337e-03 6.34887040e-01 -8.59352171e-01
3.40397090e-01 7.47681141e-01 1.12375215e-01 -5.52290618e-01
-5.38234353e-01 3.15057278e-01 -3.56475323e-01 -2.91793555e-01
4.72274899e-01 -1.29952341e-01 -5.89284897e-01 1.88079655e+00
3.97998363e-01 -3.51836264e-01 1.67587310e-01 7.09568799e-01
9.48478818e-01 6.58927441e-01 3.17090094e-01 -3.88268471e-01
1.37617779e+00 -1.34339035e+00 -8.84719491e-01 -3.14849377e-01
8.27445328e-01 -8.96460950e-01 1.44086707e+00 1.92871928e-01
-1.06232429e+00 -5.70603549e-01 -6.56633556e-01 1.19040422e-02
-1.44411281e-01 1.07597075e-01 5.55641651e-01 4.91845846e-01
-1.07894099e+00 2.97350287e-01 -6.82210445e-01 -5.47024488e-01
-5.79217434e-01 3.06815356e-01 -1.66072268e-02 2.14999035e-01
-1.40676653e+00 1.31541908e+00 1.18406489e-01 -2.23943293e-01
-7.34967232e-01 -3.40299487e-01 -7.54718721e-01 9.76468846e-02
5.96182227e-01 -4.00648713e-01 1.91080797e+00 -6.54083133e-01
-2.08397245e+00 6.02285981e-01 -3.67677420e-01 -5.10321736e-01
2.54129320e-01 -2.80055374e-01 -8.03185776e-02 1.05011202e-02
-2.12514326e-01 8.04338336e-01 4.91107494e-01 -1.11747146e+00
-5.79774380e-01 -4.02344838e-02 4.30369347e-01 4.23968971e-01
-4.44347501e-01 4.38079566e-01 -3.62436861e-01 -2.18230188e-01
-4.69255932e-02 -1.24916589e+00 -3.11899960e-01 -6.50898159e-01
-4.12327886e-01 -1.00325561e+00 6.02013409e-01 -5.51601708e-01
1.62968802e+00 -1.83623087e+00 2.77393907e-01 -2.43389800e-01
3.80853936e-02 1.43809050e-01 -1.32203519e-01 9.00618494e-01
2.17063710e-01 2.34118387e-01 1.01870395e-01 -6.45638466e-01
5.14109246e-02 -1.72270723e-02 -4.81490433e-01 4.66667600e-02
1.39702827e-01 8.19598496e-01 -6.77558661e-01 -5.26127040e-01
3.27796757e-01 -9.98701602e-02 -6.18750453e-01 8.76996160e-01
-6.04203820e-01 7.86501706e-01 -4.18458551e-01 4.91358608e-01
2.06060950e-02 -3.75820160e-01 6.74619794e-01 1.86245203e-01
-1.21418335e-01 9.45903301e-01 -6.54995561e-01 1.47465277e+00
-7.54556656e-01 4.26218837e-01 4.89948168e-02 -7.28734314e-01
9.08361018e-01 5.84148765e-01 3.64002764e-01 -7.67148316e-01
5.45823090e-02 2.64125112e-02 3.33855718e-01 -5.69791317e-01
8.33598673e-01 6.95152730e-02 -3.28409344e-01 8.01090598e-01
1.04302019e-01 -2.75284797e-01 2.42418349e-01 5.23227811e-01
8.90493095e-01 -1.45083055e-01 5.28626025e-01 -4.76656221e-02
7.87052453e-01 -5.16883172e-02 3.33779722e-01 9.62825596e-01
-1.95794955e-01 1.95361108e-01 3.40515286e-01 -3.77470776e-02
-5.42091012e-01 -5.48992693e-01 5.05648963e-02 1.82111239e+00
1.92883566e-01 -5.65465271e-01 -6.16630971e-01 -6.87523186e-01
-1.15974769e-01 1.01374412e+00 -2.23974809e-01 -2.30204225e-01
-7.94184923e-01 -4.43079531e-01 4.06222969e-01 3.64148229e-01
1.99774280e-01 -1.21374547e+00 -5.39131999e-01 4.15722042e-01
-4.98689413e-01 -1.15109825e+00 -5.24580538e-01 1.54259488e-01
-5.90021133e-01 -7.11205959e-01 -2.94049382e-01 -6.50318861e-01
2.40147054e-01 4.24664140e-01 1.29243207e+00 3.78850192e-01
3.49797189e-01 6.36951923e-01 -5.93395948e-01 -1.98725849e-01
-8.62847924e-01 2.73336232e-01 1.83811590e-01 -2.41902784e-01
4.12933111e-01 -2.89107680e-01 -3.71637046e-01 7.66195536e-01
-6.61484182e-01 8.34485888e-02 4.08711284e-01 1.15290344e+00
9.47158970e-03 -5.41296303e-01 1.06961679e+00 -9.62699890e-01
1.25502217e+00 -5.43845475e-01 -3.76931906e-01 8.80584478e-01
-9.36255634e-01 9.26072747e-02 5.88638961e-01 -6.54745340e-01
-1.33630168e+00 -4.76206355e-02 -2.22545400e-01 -7.42741525e-02
-2.83811480e-01 6.20376706e-01 1.66849941e-01 8.10356662e-02
5.27387798e-01 1.75268203e-01 -2.75951046e-02 -4.23832536e-01
3.89109462e-01 1.01463783e+00 -2.39477586e-02 -7.75679052e-01
9.81194749e-02 -1.08444586e-01 -6.76648319e-01 -7.52784491e-01
-7.16253817e-01 -8.15956235e-01 -5.44304311e-01 -3.47974867e-01
8.84418845e-01 -7.67887831e-01 -9.27442670e-01 2.11280748e-01
-1.33317673e+00 -6.63064361e-01 2.60699481e-01 3.63193184e-01
-4.23698574e-01 3.18420321e-01 -1.06334615e+00 -1.37279356e+00
-3.31168175e-01 -1.38558483e+00 9.27340031e-01 3.02079737e-01
-7.78437316e-01 -1.07088339e+00 8.51339698e-02 5.87898254e-01
4.86712247e-01 -5.23857355e-01 1.18904090e+00 -1.17974627e+00
-5.63608706e-01 1.33263767e-01 2.85489619e-01 6.53538555e-02
1.49407148e-01 -9.87604186e-02 -9.53207791e-01 -2.89077967e-01
2.66403228e-01 -9.65035260e-01 6.28348291e-01 2.27506027e-01
7.33396053e-01 -4.54402238e-01 -3.32565576e-01 -2.95963138e-01
7.59661078e-01 5.35084188e-01 2.03630626e-01 1.70040071e-01
2.66499192e-01 1.17412043e+00 1.00425255e+00 3.94870371e-01
7.90913701e-01 1.12216520e+00 9.39103663e-02 1.66682988e-01
1.52239472e-01 -1.93326846e-01 5.32180250e-01 9.95932341e-01
5.99305928e-01 -6.81027472e-01 -1.00412321e+00 4.65241134e-01
-2.08977652e+00 -7.16851175e-01 -3.25540751e-02 2.02743983e+00
1.33591998e+00 1.28482625e-01 2.53734469e-01 -3.40268284e-01
3.91751289e-01 2.97103971e-01 -4.97760057e-01 -6.49793327e-01
1.46095693e-01 6.10904954e-02 5.28610079e-03 8.39605689e-01
-8.97714138e-01 1.33817768e+00 6.06338358e+00 5.49936533e-01
-1.16499364e+00 9.46118087e-02 7.11656630e-01 1.79950044e-01
-3.20276588e-01 2.31465682e-01 -1.09471667e+00 2.53424108e-01
1.22817886e+00 -1.96139529e-01 3.39852482e-01 6.94584608e-01
2.87351966e-01 -3.74763548e-01 -1.47350729e+00 4.18198287e-01
3.95925641e-02 -1.07427406e+00 -4.15050462e-02 -4.09514755e-01
3.83545578e-01 -3.85251909e-01 -5.12956344e-02 8.89322817e-01
7.34198511e-01 -9.40987527e-01 5.10268033e-01 1.85100451e-01
3.14278096e-01 -1.83486417e-01 5.64670324e-01 7.99241304e-01
-9.36617851e-01 -2.40569428e-01 -1.41356260e-01 -2.55826283e-02
4.10092115e-01 -2.91894693e-02 -1.39831018e+00 5.87666690e-01
3.76550078e-01 3.30041707e-01 -5.18114388e-01 4.03381377e-01
-2.32628390e-01 8.04767132e-01 -1.25960009e-02 -3.19883764e-01
1.87107652e-01 -1.36571392e-01 3.96403432e-01 1.20053184e+00
-1.94239423e-01 4.29371864e-01 7.48899341e-01 5.71406305e-01
2.41490453e-02 3.21713626e-01 -4.23207432e-01 2.08179634e-02
7.78444469e-01 1.02635324e+00 -4.69726622e-01 -3.01098138e-01
-4.85344499e-01 6.15422428e-01 4.61121976e-01 2.98744798e-01
-6.71742618e-01 8.47861096e-02 3.97863686e-01 -8.63790214e-02
-1.50465012e-01 -1.95086494e-01 -1.61738008e-01 -8.94457817e-01
-1.96924493e-01 -1.30718732e+00 4.22648728e-01 -3.56835455e-01
-1.22268140e+00 7.33183086e-01 3.03326428e-01 -1.09274828e+00
-8.32118213e-01 -2.13184148e-01 -6.35765493e-01 6.82944000e-01
-1.29204702e+00 -9.03896391e-01 4.69502099e-02 4.30582225e-01
1.21763098e+00 -1.29298300e-01 9.51882839e-01 1.27011970e-01
-6.66142702e-01 6.97149456e-01 -2.11815372e-01 -8.40623230e-02
1.31593919e+00 -1.24762738e+00 1.75766692e-01 4.60219175e-01
-2.48364061e-02 9.96316552e-01 7.02585161e-01 -5.91138899e-01
-1.25618792e+00 -6.20120525e-01 1.14630389e+00 -5.28596699e-01
6.41614735e-01 -4.05654311e-01 -1.15596759e+00 6.39020264e-01
6.82701528e-01 -7.82634795e-01 7.60857701e-01 4.13791925e-01
-1.32106915e-01 3.01720768e-01 -8.30993295e-01 6.58155024e-01
6.50766790e-01 -6.32220387e-01 -7.22757220e-01 7.04767883e-01
1.00778544e+00 -6.44663036e-01 -6.46913826e-01 3.89041513e-01
4.48367268e-01 -9.21425700e-01 8.82744610e-01 -8.87900591e-01
4.47883785e-01 2.88712382e-01 -7.58801922e-02 -1.25095582e+00
-6.01423234e-02 -6.19248211e-01 -1.11675598e-01 1.40783119e+00
5.33787906e-01 -4.68806624e-01 3.16709608e-01 1.06992054e+00
-1.40994936e-01 -8.81591797e-01 -6.73296213e-01 -6.05933726e-01
1.07043505e-01 1.01113133e-01 2.93620467e-01 9.81711328e-01
3.97891581e-01 1.08111882e+00 -6.98308825e-01 -2.05109105e-01
-1.70695186e-01 3.32271487e-01 9.68850195e-01 -1.11238325e+00
-3.99148464e-01 -4.34011668e-01 4.36573535e-01 -1.70560682e+00
7.04581201e-01 -6.70782804e-01 3.96745533e-01 -1.37214482e+00
8.70202258e-02 -7.29209721e-01 6.52132332e-02 5.74988782e-01
-5.60413003e-01 -6.28728986e-01 2.66931504e-01 5.81270218e-01
-6.77085459e-01 7.09302485e-01 9.15383756e-01 -2.04361603e-01
-8.32343698e-01 4.11444664e-01 -4.85433728e-01 5.31184554e-01
7.82906353e-01 -4.49242711e-01 -6.04917347e-01 -2.69630253e-01
-2.94159919e-01 7.42372274e-01 -8.44413862e-02 -3.70629400e-01
5.73547125e-01 -4.65465367e-01 -4.38794792e-01 -4.60355014e-01
6.39201403e-01 -4.80944604e-01 -3.19812208e-01 3.79760951e-01
-1.12843251e+00 2.85649568e-01 1.00998245e-01 5.42410672e-01
-4.24279928e-01 -4.17827994e-01 6.19706631e-01 -2.09947109e-01
-5.94548881e-01 -2.74079323e-01 -6.69544280e-01 -1.60517722e-01
7.94803679e-01 3.07000279e-01 -4.70526099e-01 -8.94949079e-01
-5.29562294e-01 6.88071191e-01 1.14034675e-01 6.81492567e-01
4.58050072e-01 -8.25315654e-01 -5.99091053e-01 -1.87162787e-01
1.23699076e-01 -2.05536813e-01 9.13826227e-02 9.71300662e-01
1.40019625e-01 8.60491812e-01 9.82049406e-02 -8.52665544e-01
-1.67639995e+00 2.32686549e-01 2.63780445e-01 -8.18930387e-01
-1.87540367e-01 8.39676440e-01 2.25170001e-01 -8.69240940e-01
4.25156206e-01 -1.23379804e-01 -6.02342010e-01 1.77785173e-01
3.33795398e-01 -1.00076303e-01 7.53092319e-02 -3.29175502e-01
-3.17981452e-01 -1.22865528e-01 -5.42488098e-01 -4.25580502e-01
9.17555809e-01 -4.20182586e-01 -1.20012425e-01 7.45214105e-01
7.82506943e-01 1.13506101e-01 -7.62705088e-01 -6.00681782e-01
4.87531424e-01 -1.91486076e-01 -4.16282266e-01 -9.11228836e-01
-3.40802222e-01 7.91309595e-01 2.67420381e-01 5.45631528e-01
8.78308535e-01 7.62295797e-02 6.71967685e-01 7.55873322e-01
3.44816357e-01 -1.09936845e+00 5.66588819e-01 9.72275972e-01
8.97109032e-01 -1.63677800e+00 -1.48442179e-01 -4.04597282e-01
-1.12648559e+00 1.00420630e+00 1.23606277e+00 4.52986389e-01
4.23261970e-01 2.92005367e-03 2.04046696e-01 -4.23559919e-02
-1.47654021e+00 -1.71407029e-01 2.65253484e-01 -4.26809676e-02
1.02002597e+00 -2.64277570e-02 -7.37617135e-01 4.63177264e-01
3.16842906e-02 -5.05096197e-01 4.51969087e-01 1.09218562e+00
-4.91039783e-01 -1.55634844e+00 -1.75937384e-01 3.07465971e-01
-1.78196535e-01 -1.52608484e-01 -9.28469121e-01 6.42124176e-01
-6.95564508e-01 1.61710036e+00 -3.20586324e-01 -6.61214054e-01
3.99215877e-01 5.74484885e-01 7.07315505e-02 -9.78577673e-01
-1.32737815e+00 9.28096771e-02 4.23786044e-01 -3.60067606e-01
-3.90199691e-01 -4.67630625e-01 -1.16317248e+00 -1.73453242e-01
-6.52185857e-01 4.35040474e-01 3.85870129e-01 1.11104238e+00
2.01826319e-01 4.33125049e-01 8.95874023e-01 -4.45894152e-01
-1.29031646e+00 -1.33490980e+00 1.22128725e-01 2.90007204e-01
9.97955129e-02 -6.07522070e-01 -1.47189766e-01 -1.22402646e-01] | [12.611032485961914, 7.762442588806152] |
ea88ac44-629e-4ff1-b12a-b7089a7293c8 | best-arm-identification-in-stochastic-bandits | 2301.03785 | null | https://arxiv.org/abs/2301.03785v2 | https://arxiv.org/pdf/2301.03785v2.pdf | Best Arm Identification in Stochastic Bandits: Beyond $β-$optimality | This paper investigates a hitherto unaddressed aspect of best arm identification (BAI) in stochastic multi-armed bandits in the fixed-confidence setting. Two key metrics for assessing bandit algorithms are computational efficiency and performance optimality (e.g., in sample complexity). In stochastic BAI literature, there have been advances in designing algorithms to achieve optimal performance, but they are generally computationally expensive to implement (e.g., optimization-based methods). There also exist approaches with high computational efficiency, but they have provable gaps to the optimal performance (e.g., the $\beta$-optimal approaches in top-two methods). This paper introduces a framework and an algorithm for BAI that achieves optimal performance with a computationally efficient set of decision rules. The central process that facilitates this is a routine for sequentially estimating the optimal allocations up to sufficient fidelity. Specifically, these estimates are accurate enough for identifying the best arm (hence, achieving optimality) but not overly accurate to an unnecessary extent that creates excessive computational complexity (hence, maintaining efficiency). Furthermore, the existing relevant literature focuses on the family of exponential distributions. This paper considers a more general setting of any arbitrary family of distributions parameterized by their mean values (under mild regularity conditions). The optimality is established analytically, and numerical evaluations are provided to assess the analytical guarantees and compare the performance with those of the existing ones. | ['Ali Tajer', 'Arpan Mukherjee'] | 2023-01-10 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [-4.60168608e-02 -2.31169954e-01 -9.52962101e-01 5.53385494e-03
-1.24170089e+00 -7.51163185e-01 1.52862981e-01 -1.91002600e-02
-2.81548768e-01 1.10293841e+00 -5.41282967e-02 -8.04774463e-01
-8.90162051e-01 -7.53369391e-01 -7.26220250e-01 -1.03574145e+00
-4.39459421e-02 5.39651453e-01 -2.15507776e-01 5.48656024e-02
3.17993104e-01 4.62632090e-01 -1.29037154e+00 -3.11065137e-01
1.04706514e+00 1.45303285e+00 -8.14085826e-02 4.73387390e-01
6.70800209e-02 4.30512339e-01 -3.95862162e-01 -8.31424654e-01
4.33083773e-01 -5.86945832e-01 -6.21459007e-01 7.51203448e-02
-1.35401741e-01 -4.54975188e-01 1.08999968e-01 1.19747066e+00
3.75349194e-01 1.69840809e-02 6.70144737e-01 -1.10818219e+00
-4.59524304e-01 8.29610825e-01 -9.04592574e-01 4.63602215e-01
-1.21304179e-02 -2.39722565e-01 1.34505904e+00 -4.85186040e-01
-2.65554845e-01 1.07319570e+00 5.44524372e-01 -2.48909257e-02
-1.26042914e+00 -6.34004116e-01 1.56633407e-01 -7.84168169e-02
-1.40030074e+00 -5.01930833e-01 5.99655330e-01 -4.31259513e-01
1.43146113e-01 6.07695580e-01 6.31401718e-01 7.47773051e-01
-2.06466466e-01 9.03952777e-01 1.14141035e+00 -7.07209468e-01
5.40416777e-01 3.07479978e-01 5.18827856e-01 3.62730712e-01
8.13452303e-01 3.79305273e-01 -2.96855956e-01 -5.77097297e-01
7.95760512e-01 -6.26837313e-02 -2.24614024e-01 -3.98258120e-01
-6.50836945e-01 1.16048396e+00 9.76064429e-02 2.44053558e-01
-7.56264985e-01 1.00698680e-01 1.20538481e-01 1.69142306e-01
5.83419383e-01 1.39708772e-01 -1.98501214e-01 -2.10180491e-01
-1.07637405e+00 3.99206400e-01 7.80127704e-01 1.03061259e+00
4.24944252e-01 5.35431094e-02 -5.47863901e-01 7.38891482e-01
1.83192611e-01 5.37511826e-01 9.77424979e-02 -6.11512721e-01
7.58324444e-01 -1.38615131e-01 9.10785973e-01 -9.62995529e-01
-2.51604557e-01 -9.75490749e-01 -8.22025537e-01 -2.53384024e-01
8.05268168e-01 -2.48945802e-01 -3.59919399e-01 1.77550888e+00
4.13361907e-01 2.97451690e-02 -1.66226998e-01 1.00114167e+00
-5.15164621e-02 5.72251022e-01 -1.15912266e-01 -8.07965219e-01
1.38563848e+00 -8.01980972e-01 -7.97014296e-01 -2.53899813e-01
3.81206781e-01 -6.02880061e-01 8.84874105e-01 4.53157187e-01
-1.28266883e+00 7.54501298e-02 -9.46134329e-01 8.33228230e-01
1.04193605e-01 1.60756111e-01 8.56756628e-01 1.33277452e+00
-6.76791966e-01 2.60806948e-01 -5.89439750e-01 5.30703962e-02
4.29953218e-01 2.69127548e-01 3.94418567e-01 1.07285112e-01
-9.28654432e-01 7.11278379e-01 3.53488028e-01 2.26630032e-01
-7.00986505e-01 -7.13315606e-01 -3.88354212e-01 3.50180835e-01
6.70499265e-01 -6.32291675e-01 1.63144863e+00 -1.15157616e+00
-1.56131983e+00 2.25477427e-01 -1.39600858e-01 -7.97243834e-01
8.36422980e-01 -1.88943490e-01 -4.50075045e-02 1.05213180e-01
1.55762821e-01 -2.84205467e-01 7.53174484e-01 -1.22885442e+00
-9.83889163e-01 -3.53757381e-01 4.14337426e-01 1.33034354e-02
-5.93432546e-01 1.88221455e-01 -1.22940198e-01 -9.27032888e-01
1.07590958e-01 -8.82894158e-01 -3.13162893e-01 -6.04310989e-01
-5.35512209e-01 -7.65734017e-02 -1.02600485e-01 -5.73666275e-01
1.79383755e+00 -1.95516407e+00 -1.97258279e-01 5.54285824e-01
-2.63896197e-01 -7.35303685e-02 2.16296896e-01 4.25949365e-01
5.34570403e-02 1.68059796e-01 -2.22295091e-01 -1.17398500e-01
1.63216099e-01 -8.03615972e-02 -4.27439302e-01 8.50659549e-01
-3.81085038e-01 6.37542725e-01 -7.25026727e-01 -3.96744907e-01
2.17420124e-02 -1.08673736e-01 -6.31468713e-01 -3.23443152e-02
-1.49116116e-02 2.09687442e-01 -8.76781464e-01 7.68591344e-01
5.90225041e-01 -4.95696694e-01 1.17198512e-01 8.21396858e-02
-3.28169644e-01 5.17991558e-02 -1.45646691e+00 8.38988721e-01
-5.24225354e-01 2.80026700e-02 2.23898470e-01 -1.65420508e+00
5.34872234e-01 3.42026442e-01 6.29015744e-01 -2.47574762e-01
2.40973294e-01 4.47403729e-01 7.22612534e-03 -1.91050410e-01
3.05923074e-01 -2.90846974e-01 -7.28904828e-02 6.65582359e-01
-4.24764335e-01 3.25160623e-01 3.04114968e-01 -2.02862129e-01
5.39388716e-01 -5.25108695e-01 6.50181115e-01 -7.11186171e-01
3.58631283e-01 -7.98088685e-02 3.96413505e-01 1.32298636e+00
-1.36007294e-01 1.52299851e-01 3.84205222e-01 -5.86457737e-02
-9.53725696e-01 -7.82490253e-01 -3.48301440e-01 1.14132321e+00
1.48414075e-01 1.88735098e-01 -7.47774899e-01 -1.77269638e-01
2.13977367e-01 8.22989225e-01 -6.84289336e-01 1.50666386e-01
-2.97706664e-01 -1.07076180e+00 1.83231518e-01 4.66780543e-01
4.90301549e-01 -3.11660379e-01 -7.25350738e-01 5.27012765e-01
-1.47160545e-01 -8.08579862e-01 -4.62456495e-01 7.21858367e-02
-8.21735978e-01 -9.29564357e-01 -1.06895173e+00 -3.08226466e-01
5.23269415e-01 5.53307772e-01 8.20522487e-01 -2.06276149e-01
1.99817717e-01 3.68948519e-01 -2.59635925e-01 -6.82922125e-01
-2.17110440e-02 5.00229187e-02 -3.03255226e-02 4.45511192e-01
1.17427528e-01 -3.55469137e-01 -7.32135177e-01 5.33859551e-01
-7.97158241e-01 -1.41308978e-01 7.11247444e-01 1.06894028e+00
5.84516287e-01 1.13632858e-01 8.93463969e-01 -5.90651631e-01
7.76039243e-01 -6.30972505e-01 -1.15147316e+00 4.78866428e-01
-6.52193606e-01 -7.58550465e-02 5.79158604e-01 -4.33339298e-01
-1.10271823e+00 -3.22307497e-01 1.20314769e-01 -1.35993451e-01
2.85874695e-01 8.06807101e-01 5.31626344e-02 6.17849790e-02
4.16518152e-01 2.53173828e-01 -9.43294168e-02 -4.94313955e-01
2.41553783e-01 7.97646403e-01 2.48627678e-01 -1.09774959e+00
4.61367428e-01 4.57327306e-01 1.49262503e-01 -5.13068080e-01
-1.43870938e+00 -4.49255556e-01 5.57708144e-02 -9.14844498e-03
1.18754268e-01 -6.03727818e-01 -9.80725348e-01 6.47990778e-02
-7.30294466e-01 4.48924936e-02 -8.90580714e-02 7.83422172e-01
-1.02690196e+00 3.05264801e-01 -4.40537304e-01 -1.74264801e+00
-3.05867583e-01 -1.03438294e+00 5.69375992e-01 3.43682945e-01
-3.31363641e-02 -7.61775494e-01 -1.53026477e-01 3.35277289e-01
2.80694425e-01 1.33862659e-01 8.84038925e-01 -5.92025459e-01
-6.04614079e-01 -2.34734356e-01 -2.39432111e-01 3.44351046e-02
2.54789013e-02 -2.29489073e-01 -5.68332493e-01 -4.39320296e-01
1.41851723e-01 1.44113779e-01 4.65852112e-01 1.13560843e+00
1.38600469e+00 -8.69392812e-01 -4.60782498e-01 3.25633675e-01
1.49376822e+00 6.83936834e-01 2.15412676e-01 7.45577633e-01
-2.41065606e-01 5.15083551e-01 8.35759938e-01 8.92980635e-01
6.04979955e-02 7.63445377e-01 3.74846041e-01 2.26566210e-01
6.31741047e-01 -3.50584202e-02 -1.11141376e-01 3.74791294e-01
-1.87143281e-01 -4.20943290e-01 -7.21175849e-01 8.26621056e-01
-2.17167783e+00 -1.09183621e+00 1.44586906e-01 2.73640347e+00
7.70913005e-01 1.16503902e-01 7.46164143e-01 3.84215862e-01
1.11242735e+00 -2.07317650e-01 -6.61684036e-01 -3.83111000e-01
-1.84191525e-01 1.44597679e-01 8.21196020e-01 4.01741147e-01
-1.07953072e+00 3.07233036e-01 6.49949646e+00 1.23174500e+00
-5.09363949e-01 2.12495685e-01 9.56654489e-01 -2.11631879e-01
-1.84648380e-01 -3.42285680e-03 -9.22933459e-01 7.68946230e-01
8.81251037e-01 -6.21192336e-01 5.23268878e-01 9.89301324e-01
4.31645572e-01 -2.04801768e-01 -8.51128101e-01 1.05750930e+00
-4.01781857e-01 -1.41977525e+00 -2.93669432e-01 2.94070572e-01
7.94316232e-01 -6.48782730e-01 3.68456572e-01 -3.87230217e-02
4.77395117e-01 -8.34784806e-01 9.74568129e-01 3.57870638e-01
6.25819564e-01 -1.30154133e+00 7.70401120e-01 5.11068523e-01
-8.27494919e-01 -6.93877220e-01 -2.51811028e-01 -5.70530966e-02
3.26076865e-01 8.76594603e-01 -1.43921092e-01 8.65217209e-01
6.70327723e-01 1.94897696e-01 1.58646107e-01 1.56975198e+00
7.07780570e-02 8.53759229e-01 -6.31837547e-01 -5.10786355e-01
5.88091433e-01 -5.76281965e-01 4.85819429e-01 1.00471497e+00
7.86071837e-01 3.12477112e-01 7.68260434e-02 5.12381673e-01
2.13072821e-01 3.21374744e-01 -2.51723289e-01 6.50964156e-02
8.26648235e-01 7.11366117e-01 -5.30873239e-01 -3.17794472e-01
-5.76224983e-01 1.50702894e-01 2.13701427e-01 4.36276883e-01
-9.67790186e-01 -1.42106533e-01 7.08303988e-01 1.01067469e-01
3.90527576e-01 1.12545051e-01 -7.03775942e-01 -6.32305324e-01
2.47033074e-01 -7.61558235e-01 8.82264853e-01 -8.55127126e-02
-1.41239953e+00 9.35565308e-02 4.45591420e-01 -1.07755923e+00
-4.28150415e-01 -3.44917774e-01 -2.38155797e-01 8.40072513e-01
-1.45627689e+00 -6.53589368e-01 2.19742030e-01 4.34404880e-01
2.11824730e-01 -4.15836498e-02 3.85686338e-01 4.18103278e-01
-7.82072902e-01 1.05938399e+00 8.86744857e-01 -2.70137638e-01
-5.23495413e-02 -8.36558223e-01 -4.92404342e-01 7.67411888e-01
-7.93239288e-03 5.85804164e-01 9.70257938e-01 -1.92333177e-01
-1.30226803e+00 -7.66205907e-01 3.62490147e-01 -3.99905853e-02
9.45783794e-01 6.31432459e-02 -4.05594379e-01 6.45254910e-01
-1.47624880e-01 -4.19922829e-01 6.96534634e-01 3.22210997e-01
4.90721613e-02 -2.82197803e-01 -1.08136749e+00 6.96388006e-01
9.82978702e-01 5.59407659e-02 -2.20906883e-01 4.65096325e-01
2.04685256e-01 -2.48257220e-01 -6.60906911e-01 3.83560419e-01
8.36430430e-01 -1.08081114e+00 1.03365147e+00 -7.43803084e-01
4.95276488e-02 7.78602716e-03 -3.51803035e-01 -1.09806764e+00
-3.26223999e-01 -1.10678542e+00 -3.68770659e-01 1.18548024e+00
3.40606600e-01 -1.15340197e+00 7.82307923e-01 5.78237295e-01
1.64608255e-01 -9.63328838e-01 -1.22704148e+00 -1.35653675e+00
2.96929508e-01 -4.64363635e-01 8.41383517e-01 5.64866126e-01
7.77409121e-04 -1.53146014e-01 -5.35952091e-01 2.02357978e-01
8.48438978e-01 6.19037032e-01 6.01806760e-01 -1.02368546e+00
-4.88315731e-01 -8.58740032e-01 3.68035063e-02 -1.45830786e+00
3.25394124e-02 -4.49248880e-01 -1.52517349e-01 -1.16345596e+00
4.50880736e-01 -8.42142284e-01 -6.18916810e-01 3.08041144e-02
-1.30575180e-01 -3.24224830e-01 9.01401322e-03 3.03896457e-01
-2.10969716e-01 3.42534840e-01 9.28234696e-01 -4.10885029e-02
-2.37358227e-01 7.04378307e-01 -1.21345878e+00 5.37495553e-01
9.66491580e-01 -5.28991163e-01 -3.20445150e-01 -7.37224519e-02
2.25398645e-01 5.75703502e-01 4.17518675e-01 -5.35469592e-01
4.18368801e-02 -5.13143301e-01 -1.19603641e-01 -7.47471690e-01
1.88311502e-01 -8.41995835e-01 3.06738347e-01 5.25076032e-01
-3.75721276e-01 -2.60143448e-02 -9.75376815e-02 7.82542169e-01
-6.66693971e-02 -6.80749655e-01 8.72944951e-01 1.65237069e-01
9.29974765e-02 3.01677316e-01 -2.44260862e-01 3.33715416e-02
1.17870045e+00 -1.42762423e-01 -1.61239445e-01 -7.68325388e-01
-4.95140493e-01 8.81336257e-02 -9.47176591e-02 -1.56277671e-01
-3.94692905e-02 -1.42419708e+00 -6.64987028e-01 -3.03308874e-01
-1.01913504e-01 -4.43354100e-01 3.13345715e-02 1.10263431e+00
-3.10919322e-02 9.60666478e-01 4.03684109e-01 -4.83026415e-01
-8.48280489e-01 7.14374721e-01 2.56868571e-01 -6.14209533e-01
-1.24768756e-01 7.12581217e-01 2.41568387e-01 3.61898154e-01
3.83734435e-01 -1.58164173e-01 5.71830459e-02 1.38205573e-01
6.76490784e-01 6.36427283e-01 8.30940083e-02 -2.15850651e-01
-2.55679250e-01 3.76305670e-01 1.69554815e-01 -3.22087258e-01
1.20552897e+00 -2.55078048e-01 9.96240526e-02 1.46942675e-01
8.24870467e-01 3.57070118e-02 -1.13828588e+00 -4.92644846e-01
2.39753619e-01 -7.72910416e-01 1.72398925e-01 -4.77813035e-01
-9.28689241e-01 5.33909142e-01 3.57712209e-01 7.91287124e-01
1.07776713e+00 -1.40022978e-01 4.38013256e-01 1.79835305e-01
7.23825753e-01 -1.15739679e+00 -3.20696622e-01 4.32543419e-02
6.96969450e-01 -1.02434933e+00 1.10179797e-01 -4.06294733e-01
-2.87280649e-01 8.28690886e-01 7.02987462e-02 -1.08575365e-02
7.78406143e-01 9.59749222e-02 -5.77794135e-01 1.95277750e-01
-2.90997654e-01 -3.88059169e-01 1.78332508e-01 3.82064730e-01
1.78253472e-01 4.31558311e-01 -9.19643760e-01 1.11476219e+00
-3.51769894e-01 -1.84271589e-01 2.53794551e-01 7.40560949e-01
-3.87869865e-01 -6.73024416e-01 -8.06560516e-01 8.37132573e-01
-9.31878090e-01 -6.65522143e-02 2.69220918e-01 9.12495375e-01
-6.61733568e-01 1.44599032e+00 4.79010120e-03 3.08753103e-01
3.00068468e-01 -1.68971136e-01 4.77741182e-01 -5.64940758e-02
-1.56304479e-01 3.25325668e-01 2.98580378e-01 -1.12541221e-01
-5.77882171e-01 -8.46838713e-01 -3.84384364e-01 -5.45432985e-01
-7.36585855e-01 6.26382709e-01 5.36327064e-01 9.33299661e-01
9.08519477e-02 1.20084912e-01 1.00250494e+00 -4.47560847e-01
-1.36494935e+00 -7.09894419e-01 -7.97224104e-01 -1.27399191e-01
2.63422459e-01 -9.54799175e-01 -4.38828677e-01 -4.98749375e-01] | [4.559055805206299, 3.2856345176696777] |
d2ec3956-ad97-4f06-b620-f7df18d6bf67 | amvnet-assertion-based-multi-view-fusion | 2012.04934 | null | https://arxiv.org/abs/2012.04934v1 | https://arxiv.org/pdf/2012.04934v1.pdf | AMVNet: Assertion-based Multi-View Fusion Network for LiDAR Semantic Segmentation | In this paper, we present an Assertion-based Multi-View Fusion network (AMVNet) for LiDAR semantic segmentation which aggregates the semantic features of individual projection-based networks using late fusion. Given class scores from different projection-based networks, we perform assertion-guided point sampling on score disagreements and pass a set of point-level features for each sampled point to a simple point head which refines the predictions. This modular-and-hierarchical late fusion approach provides the flexibility of having two independent networks with a minor overhead from a light-weight network. Such approaches are desirable for robotic systems, e.g. autonomous vehicles, for which the computational and memory resources are often limited. Extensive experiments show that AMVNet achieves state-of-the-art results in both the SemanticKITTI and nuScenes benchmark datasets and that our approach outperforms the baseline method of combining the class scores of the projection-based networks. | ['Zhuang Jie Chong', 'Dhananjai Sharma', 'Sergi Widjaja', 'Thi Ngoc Tho Nguyen', 'Venice Erin Liong'] | 2020-12-09 | null | null | null | null | ['lidar-semantic-segmentation'] | ['computer-vision'] | [ 1.86577857e-01 3.80561918e-01 -4.29303586e-01 -7.56792009e-01
-8.89636755e-01 -4.73173380e-01 6.80891573e-01 1.44192092e-02
-1.50881916e-01 4.74087983e-01 -2.73253262e-01 -9.63473786e-03
-1.49882600e-01 -9.46947038e-01 -7.11864531e-01 -4.78934318e-01
3.29168320e-01 1.16366720e+00 1.27176750e+00 -2.71321118e-01
9.66528654e-02 8.56055558e-01 -1.89868259e+00 4.35286164e-01
6.71103120e-01 1.41522491e+00 5.13294302e-02 4.62666601e-01
-7.09581152e-02 4.11276996e-01 -8.08337107e-02 -5.95760524e-01
5.09841919e-01 2.92149395e-01 -5.80748141e-01 -1.22731425e-01
1.01344252e+00 -1.30082116e-01 1.67475879e-01 1.24278426e+00
1.96612194e-01 3.17110360e-01 6.56379282e-01 -1.66041875e+00
3.22674930e-01 5.15521407e-01 -7.17381835e-01 -1.62106857e-01
7.68650621e-02 3.92099470e-02 1.28046405e+00 -1.34356320e+00
7.31982470e-01 1.56852424e+00 1.02956498e+00 2.86063343e-01
-1.38359165e+00 -8.46835017e-01 3.30538243e-01 4.53729749e-01
-1.34781671e+00 -3.60668063e-01 8.48838151e-01 -2.90076405e-01
8.77620757e-01 -2.17454433e-02 7.21393645e-01 6.36398971e-01
-2.04194989e-02 7.55869627e-01 8.61970067e-01 2.58304864e-01
2.87237436e-01 7.47552291e-02 1.51211560e-01 8.90346587e-01
1.92548931e-01 -4.38375585e-02 -8.00340116e-01 -3.49722028e-01
4.21531200e-01 1.04290046e-01 2.05765426e-01 -1.02589095e+00
-8.88588607e-01 9.40732658e-01 8.73847008e-01 -3.34848046e-01
-4.76723939e-01 2.46648297e-01 2.12528065e-01 -1.31392181e-01
5.88725328e-01 -1.13943391e-01 -5.71372271e-01 2.96646982e-01
-1.34246767e+00 2.29183927e-01 8.28251541e-01 9.26941216e-01
1.15029466e+00 -2.00188562e-01 1.13258876e-01 7.66593218e-01
5.79218686e-01 4.61205095e-01 -3.73007506e-01 -1.56265044e+00
5.36284387e-01 7.81066597e-01 -2.49358356e-01 -8.08367491e-01
-3.37078363e-01 -8.28028023e-02 -6.76308572e-01 9.17776525e-01
5.45742922e-02 4.53434251e-02 -1.38541460e+00 1.39494240e+00
6.05159044e-01 2.65175432e-01 1.24051377e-01 8.18976939e-01
9.26175535e-01 3.13176423e-01 -4.85746711e-02 2.28905573e-01
1.12142754e+00 -1.17320514e+00 1.40314817e-01 -2.89537340e-01
1.76028192e-01 -4.83442008e-01 5.17439604e-01 4.21035767e-01
-1.25671780e+00 -6.35621846e-01 -1.29971135e+00 -1.62064675e-02
-3.93401653e-01 -1.78805783e-01 4.55161810e-01 3.85609567e-01
-1.31935322e+00 8.07432115e-01 -9.59347427e-01 -3.41866761e-01
8.23475718e-01 6.46664500e-01 -4.30747658e-01 -1.12754636e-01
-5.17167270e-01 9.72750902e-01 5.19402862e-01 4.99164201e-02
-7.78858602e-01 -7.62085617e-01 -7.42442250e-01 -8.41196813e-03
5.84072351e-01 -8.12725365e-01 1.36853659e+00 -5.29490650e-01
-1.20850074e+00 6.33203149e-01 -2.73915052e-01 -5.75222850e-01
6.01605833e-01 -5.70510179e-02 9.13705230e-02 4.29131031e-01
3.58466178e-01 1.51803684e+00 6.43072844e-01 -1.57738590e+00
-1.09821510e+00 -7.55482614e-01 2.06215575e-01 5.19107401e-01
1.69572383e-01 -2.76760429e-01 -6.55153036e-01 9.60154459e-02
7.52867758e-01 -1.15397191e+00 -5.11337936e-01 6.57737553e-01
-4.48058635e-01 -6.94305241e-01 1.44730997e+00 -5.56248240e-02
2.26516873e-01 -2.07805490e+00 3.02481711e-01 5.25199831e-01
4.46884245e-01 9.21363086e-02 1.44320494e-02 -6.95961788e-02
1.45040542e-01 -8.11535418e-02 -2.85881609e-01 -6.30138695e-01
-2.93192953e-01 5.09977043e-01 -3.74811701e-02 3.03877115e-01
2.43417740e-01 7.82496035e-01 -6.91541135e-01 -7.84207582e-01
5.59058070e-01 3.55083883e-01 -5.43559551e-01 -1.33791253e-01
-5.62907577e-01 5.79154082e-02 -4.22677308e-01 7.50251055e-01
7.63624370e-01 -2.16459438e-01 -1.54915666e-02 -4.54756230e-01
5.70790656e-02 2.05373242e-01 -1.26733971e+00 1.92379177e+00
-2.02563792e-01 3.92437726e-01 2.56399453e-01 -6.95431948e-01
8.96647990e-01 -9.26400721e-03 5.16196907e-01 -4.39162813e-02
-8.29409212e-02 2.10550144e-01 -1.32686570e-01 1.98593184e-01
7.02249110e-01 1.30219432e-02 -4.62309308e-02 7.23749623e-02
3.26037139e-01 -7.76542664e-01 1.95214882e-01 4.21195060e-01
9.74439263e-01 3.68559986e-01 1.07037216e-01 -1.28390267e-01
5.79858601e-01 2.17593908e-01 7.83493519e-01 5.95039546e-01
-1.80084690e-01 9.00340140e-01 4.49703783e-01 -3.57914388e-01
-8.46843660e-01 -1.30282485e+00 -6.40595183e-02 1.03132689e+00
3.26899916e-01 -3.28244686e-01 -2.99128801e-01 -1.01485419e+00
2.59644330e-01 7.24411488e-01 -3.71688813e-01 1.97424904e-01
-5.41574299e-01 -1.08176462e-01 5.45627594e-01 1.07855761e+00
5.62121272e-01 -7.16452956e-01 -5.06194174e-01 4.46464308e-02
-3.77040282e-02 -1.31380773e+00 1.27776787e-01 5.18919528e-01
-1.02165163e+00 -1.06091046e+00 -9.19939280e-02 -5.53117156e-01
4.92318243e-01 4.03135985e-01 9.94037688e-01 -1.61066547e-01
2.85670072e-01 2.42747083e-01 9.53104999e-03 -6.28547370e-01
-1.24904886e-01 2.87834078e-01 -3.77312936e-02 -2.88768649e-01
2.87531793e-01 -8.87372136e-01 -3.11640769e-01 4.23860192e-01
-5.16885936e-01 1.33424968e-01 3.94238830e-01 3.50977689e-01
9.57075238e-01 -2.48143196e-01 3.37186456e-01 -5.97194672e-01
6.69452250e-02 -2.57337332e-01 -9.16376173e-01 5.97250648e-02
-6.74189448e-01 -3.25004607e-02 1.76217303e-01 -3.90703641e-02
-9.36138868e-01 4.76510257e-01 -9.07271460e-04 -8.83511543e-01
-1.72646299e-01 1.92185298e-01 -9.42026377e-02 -3.84121686e-01
5.35758674e-01 -3.24608624e-01 1.13975406e-01 -2.42051244e-01
7.34206975e-01 4.16126728e-01 7.24281013e-01 -5.19391716e-01
9.54590738e-01 7.93578029e-01 5.35187423e-01 -5.08370280e-01
-1.06339860e+00 -7.31311977e-01 -9.46584105e-01 -5.83507657e-01
9.32113171e-01 -1.10139930e+00 -6.46697700e-01 2.42683366e-01
-1.39526987e+00 1.18545808e-01 -6.18969321e-01 1.34321257e-01
-6.61407351e-01 9.59085822e-02 -2.69771814e-01 -6.02956831e-01
-1.88205615e-01 -1.34329855e+00 1.38220942e+00 3.89568895e-01
-4.77938615e-02 -6.51459157e-01 -6.48945794e-02 5.30071855e-01
-2.33185794e-02 9.32059214e-02 5.51890254e-01 -7.81673193e-01
-9.21020210e-01 -3.60730812e-02 -5.52727938e-01 4.54570621e-01
-3.90705258e-01 6.60942644e-02 -1.22484696e+00 7.18006790e-02
-3.26792806e-01 -5.96899986e-01 1.21885335e+00 3.64711672e-01
1.21183813e+00 2.42118999e-01 -5.54260552e-01 5.91186106e-01
1.30583048e+00 -2.48524562e-01 1.93732604e-01 -4.78580520e-02
8.88790190e-01 6.29458189e-01 5.51582754e-01 6.90580811e-03
6.41396523e-01 6.26860619e-01 1.01394415e+00 2.43162829e-02
-1.76067144e-01 -1.52456209e-01 7.34744314e-03 2.87220478e-01
-2.29163975e-01 -1.78146571e-01 -8.91626298e-01 6.44070148e-01
-2.19876742e+00 -7.79611945e-01 -2.73785681e-01 1.85929418e+00
2.71925151e-01 4.58067238e-01 1.06298484e-01 2.99034268e-02
6.02201283e-01 5.40313125e-01 -7.11770415e-01 -2.78607219e-01
-7.24871531e-02 1.62787810e-01 6.99658275e-01 3.58404845e-01
-1.02403498e+00 1.11842310e+00 5.64360332e+00 7.92712748e-01
-7.28065312e-01 3.85650277e-01 3.97069782e-01 -2.41302490e-01
-1.86002061e-01 1.66064441e-01 -1.11484611e+00 -8.44005942e-02
7.80104518e-01 2.55106211e-01 1.31714970e-01 1.14225137e+00
-1.11844100e-01 -3.90859514e-01 -1.21242332e+00 7.11250961e-01
1.01936765e-01 -1.57137299e+00 -5.28324172e-02 4.64251079e-02
7.69779682e-01 1.08755970e+00 -7.78774619e-02 -1.00478856e-02
7.98120201e-01 -5.81375301e-01 8.50528300e-01 2.64489323e-01
6.04545116e-01 -8.67231190e-01 7.40785301e-01 4.09163326e-01
-1.50026929e+00 -6.73901290e-02 -2.80838549e-01 2.64214158e-01
4.38770205e-01 6.87774658e-01 -9.98547196e-01 8.25576127e-01
8.46256435e-01 7.30861962e-01 -4.15126890e-01 8.11923087e-01
-3.05306792e-01 4.02216822e-01 -9.86619890e-01 1.85398221e-01
4.26414579e-01 -1.45149797e-01 8.32840800e-01 7.95062780e-01
2.28269875e-01 -1.16329782e-01 6.36949360e-01 9.29250658e-01
-1.77744880e-01 -4.18506920e-01 -7.22957313e-01 5.47883034e-01
4.39938724e-01 1.69594789e+00 -1.10479414e+00 -5.83161414e-01
-3.76983345e-01 5.56146026e-01 3.50611538e-01 -7.76345804e-02
-6.65383577e-01 2.11771727e-01 7.65558779e-01 9.56740081e-02
6.55454576e-01 -2.65507549e-01 -7.37159789e-01 -7.83124208e-01
1.71880394e-01 -1.99604973e-01 3.50210786e-01 -1.04944897e+00
-1.12462676e+00 4.92856562e-01 4.06393707e-01 -1.22693992e+00
-3.63431394e-01 -5.53125679e-01 -5.64648271e-01 6.98836148e-01
-1.45826304e+00 -1.49641061e+00 -3.76719892e-01 4.00904387e-01
7.44844556e-01 -8.65833572e-05 3.99127305e-01 -8.84759799e-02
-7.64664263e-02 4.53254804e-02 -6.27722085e-01 -1.36037439e-01
3.11846524e-01 -1.28363180e+00 3.30912650e-01 6.77514851e-01
2.57543087e-01 -2.79527325e-02 3.86430532e-01 -9.10107553e-01
-8.51480782e-01 -1.34601676e+00 6.23809576e-01 -5.80689788e-01
5.62110364e-01 -1.43703386e-01 -7.17337608e-01 5.67249477e-01
1.11637793e-01 3.37027013e-01 2.64873922e-01 2.60666728e-01
-4.11168694e-01 -2.16815323e-01 -1.31485188e+00 3.19899827e-01
1.21910179e+00 -3.70725840e-01 -6.49335980e-01 3.45291615e-01
7.73775935e-01 -5.87216139e-01 -6.67654216e-01 9.41426516e-01
6.96153343e-01 -1.18167114e+00 1.11471128e+00 -2.55078614e-01
2.86340743e-01 -6.36049867e-01 -5.03526986e-01 -1.07326376e+00
-2.56308854e-01 1.36911161e-02 3.97046991e-02 8.27961564e-01
6.39500737e-01 -5.11769116e-01 1.31132782e+00 3.12962800e-01
-3.70449662e-01 -8.34661543e-01 -1.19822955e+00 -5.76229036e-01
-2.94681430e-01 -8.74158919e-01 5.26781261e-01 3.58739555e-01
-6.74013317e-01 5.74803352e-01 2.59117544e-01 5.04953444e-01
8.66221905e-01 5.06060123e-02 7.40813315e-01 -2.03208447e+00
1.50999039e-01 -3.59839141e-01 -8.18473339e-01 -8.25911403e-01
3.71488452e-01 -1.05112720e+00 2.58455932e-01 -1.69960082e+00
2.12294355e-01 -4.53930229e-01 -1.79763049e-01 6.75923467e-01
2.15876654e-01 5.77501237e-01 4.28638995e-01 1.66837111e-01
-9.01529491e-01 5.37806273e-01 9.32888448e-01 -2.45449588e-01
-2.52520770e-01 2.99519956e-01 -3.56030285e-01 1.36628425e+00
6.09820604e-01 -6.14460588e-01 -6.07002079e-01 -2.31348619e-01
1.64532706e-01 7.28518292e-02 7.71723509e-01 -1.42044461e+00
6.41605437e-01 -1.18561722e-01 4.25734162e-01 -1.59778154e+00
1.08757186e+00 -1.03111148e+00 4.25406434e-02 2.79186040e-01
2.27405410e-02 6.35927618e-02 9.22048744e-03 8.31636012e-01
-7.10726380e-02 1.49297360e-02 8.06419849e-01 -1.67388961e-01
-7.21242547e-01 4.46288496e-01 9.10219401e-02 -3.95724833e-01
9.64069664e-01 -6.72671139e-01 -1.70383900e-01 -9.81319696e-02
-7.08882093e-01 5.44209361e-01 4.55601484e-01 5.27608097e-01
7.53344893e-01 -1.13177729e+00 -4.55424219e-01 1.53382152e-01
1.62797689e-01 8.65608871e-01 -2.84527354e-02 7.99764574e-01
-1.52027145e-01 1.70532733e-01 -1.30155712e-01 -1.35806990e+00
-1.38272262e+00 -1.05978355e-01 2.19614148e-01 -2.31725112e-01
-7.48692036e-01 1.08324647e+00 9.30510610e-02 -8.79470527e-01
7.09688440e-02 -3.97571653e-01 -2.99908638e-01 3.04391414e-01
-1.84422374e-01 5.86821437e-01 2.90970832e-01 -8.60396624e-01
-5.04768193e-01 5.53761601e-01 -1.02685168e-01 -5.71675420e-01
1.34072876e+00 1.16343074e-01 -1.55226409e-01 5.41162372e-01
8.23710978e-01 -4.27713603e-01 -1.31156147e+00 -3.17072511e-01
-1.91596597e-01 -4.71700877e-02 3.11607242e-01 -6.60232723e-01
-1.28122652e+00 7.50241697e-01 4.66541976e-01 -2.44578749e-01
7.93651998e-01 3.20247442e-01 5.24095833e-01 6.39676452e-01
6.97962880e-01 -1.14463675e+00 -1.00944854e-01 6.25947893e-01
7.15301037e-01 -1.10259461e+00 1.90679207e-01 -8.46482635e-01
-3.72612745e-01 1.01193523e+00 1.00953639e+00 -1.99892595e-01
7.88146019e-01 3.34254533e-01 -1.61281660e-01 -4.60410535e-01
-8.49544227e-01 -4.21616733e-01 3.28684509e-01 6.16818666e-01
-3.45837504e-01 2.11386457e-01 1.34050429e-01 9.48335752e-02
-3.44884008e-01 -2.21498422e-02 2.83811867e-01 9.03193831e-01
-7.50564873e-01 -8.66278410e-01 -3.68819326e-01 7.72483826e-01
-3.05237193e-02 2.30539083e-01 -5.55686295e-01 6.17416978e-01
3.69131953e-01 8.60915482e-01 2.13093773e-01 -6.02079093e-01
3.98566872e-01 2.44726837e-01 3.06940049e-01 -8.12726140e-01
-4.80540812e-01 -5.01414016e-02 3.26379538e-01 -9.76462245e-01
-8.02496910e-01 -8.96287143e-01 -1.38010085e+00 -3.88480052e-02
-5.34312844e-01 -2.14508265e-01 9.50031221e-01 1.10814655e+00
3.95004183e-01 3.64207327e-01 3.49125117e-01 -1.41686833e+00
-4.88788754e-01 -8.31991792e-01 -1.99240014e-01 -2.84160286e-01
1.07618332e-01 -9.28342700e-01 -2.24058032e-01 -4.13093448e-01] | [8.212959289550781, -2.9059951305389404] |
1418f5ce-a964-4923-acc7-3156a32bde98 | a-simple-method-for-unsupervised-bilingual | 2305.14012 | null | https://arxiv.org/abs/2305.14012v1 | https://arxiv.org/pdf/2305.14012v1.pdf | A Simple Method for Unsupervised Bilingual Lexicon Induction for Data-Imbalanced, Closely Related Language Pairs | Existing approaches for unsupervised bilingual lexicon induction (BLI) often depend on good quality static or contextual embeddings trained on large monolingual corpora for both languages. In reality, however, unsupervised BLI is most likely to be useful for dialects and languages that do not have abundant amounts of monolingual data. We introduce a simple and fast method for unsupervised BLI for low-resource languages with a related mid-to-high resource language, only requiring inference on the higher-resource language monolingual BERT. We work with two low-resource languages ($<5M$ monolingual tokens), Bhojpuri and Magahi, of the severely under-researched Indic dialect continuum, showing that state-of-the-art methods in the literature show near-zero performance in these settings, and that our simpler method gives much better results. We repeat our experiments on Marathi and Nepali, two higher-resource Indic languages, to compare approach performances by resource range. We release automatically created bilingual lexicons for the first time for five languages of the Indic dialect continuum. | ['Rachel Bawden', 'Benoît Sagot', 'Josef van Genabith', 'Cristina España-Bonet', 'Niyati Bafna'] | 2023-05-23 | null | null | null | null | ['bilingual-lexicon-induction'] | ['natural-language-processing'] | [-4.65571672e-01 -2.04537705e-01 -5.52193761e-01 -3.84980738e-01
-1.05647230e+00 -9.00604546e-01 7.67115116e-01 1.75637245e-01
-9.87540424e-01 1.06004298e+00 4.39882636e-01 -8.10607910e-01
1.29285395e-01 -6.00746810e-01 -5.07221043e-01 -5.07752419e-01
-2.37821952e-01 1.15464807e+00 -1.67293977e-02 -7.12123513e-01
-1.56000584e-01 3.42376500e-01 -1.05895853e+00 -1.94108352e-01
1.02868819e+00 -7.49175102e-02 3.26499492e-01 2.52813876e-01
-2.68782586e-01 6.46008730e-01 -2.93630093e-01 -5.08092940e-01
2.89280176e-01 -3.67676258e-01 -1.03341925e+00 -4.25377190e-01
4.35398072e-01 -5.55069521e-02 -1.56248778e-01 9.33952928e-01
4.96953011e-01 -3.96109164e-01 6.57639563e-01 -5.86174250e-01
-7.44366586e-01 1.34410119e+00 -6.37079418e-01 3.47753078e-01
3.74220282e-01 -1.25296444e-01 1.44609201e+00 -1.20306289e+00
1.11688483e+00 1.24036884e+00 7.36942947e-01 2.45693058e-01
-1.40381026e+00 -8.56142342e-01 7.55137950e-02 2.02386647e-01
-1.66842246e+00 -6.66374803e-01 5.87017357e-01 -3.68958026e-01
1.25580561e+00 -7.00223446e-02 5.60268760e-01 7.09451258e-01
-1.84041783e-01 7.46546388e-01 1.58960044e+00 -9.22194958e-01
-3.19868654e-01 4.93938327e-01 1.31059557e-01 4.83045757e-01
3.97333562e-01 3.10423765e-02 -5.32498896e-01 8.88727903e-02
4.38501000e-01 -4.89910156e-01 5.92348911e-02 -3.59793901e-02
-1.51455963e+00 1.02723658e+00 -4.92783487e-02 8.80209923e-01
-1.22015014e-01 -4.90363896e-01 5.94598711e-01 7.80428767e-01
7.48771787e-01 3.95358741e-01 -7.79762149e-01 -3.01711828e-01
-9.56471205e-01 6.27754331e-02 8.66117120e-01 1.20682049e+00
8.99648428e-01 9.17509869e-02 5.12695789e-01 1.50014567e+00
1.19349487e-01 6.84959352e-01 5.48650563e-01 -8.59836936e-02
5.81947803e-01 3.24855894e-01 -1.17581777e-01 -3.09062690e-01
-2.67316282e-01 3.28761637e-02 -5.04215300e-01 2.33056433e-02
7.05614030e-01 -8.91773701e-02 -7.91643262e-01 1.97132146e+00
1.74333259e-01 -7.14902997e-01 3.55444044e-01 6.20960295e-01
7.08155990e-01 6.39120638e-01 -1.76868625e-02 -1.91230699e-01
1.44461465e+00 -7.63709307e-01 -5.03377378e-01 -3.81790608e-01
7.97691107e-01 -1.12960565e+00 1.61144590e+00 1.54708624e-01
-1.02014172e+00 -4.27729815e-01 -1.10056412e+00 -2.01591179e-01
-7.32095361e-01 2.35385090e-01 1.01830316e+00 7.85087109e-01
-1.17554116e+00 -1.22831529e-02 -4.24873114e-01 -7.63814807e-01
3.15158069e-02 4.74443853e-01 -7.72215128e-01 -3.94346565e-01
-1.52424550e+00 1.21499264e+00 4.23172623e-01 -1.20817982e-01
-8.32723200e-01 -5.20191133e-01 -1.00057423e+00 -4.84396547e-01
8.25319588e-02 3.49096984e-01 8.42378378e-01 -1.00523353e+00
-1.31996512e+00 1.62840426e+00 1.82325855e-01 -2.07094476e-01
3.39818239e-01 -2.46360824e-01 -5.64448833e-01 -3.08143377e-01
4.73242640e-01 5.61548471e-01 2.41674945e-01 -1.00577891e+00
-6.70322001e-01 -1.89744115e-01 9.81050879e-02 3.85909677e-01
-3.09368312e-01 7.30567276e-01 -3.97882521e-01 -5.85899711e-01
3.12348641e-02 -9.74204063e-01 -8.20521712e-02 -7.83674419e-01
-1.96936250e-01 -2.53073722e-01 1.38755620e-01 -1.04451263e+00
1.09581280e+00 -1.95164943e+00 -1.12417527e-01 6.43574744e-02
-2.53074229e-01 1.64304718e-01 -2.27917090e-01 6.71519995e-01
-9.96434689e-02 5.47026731e-02 -8.59062895e-02 -2.47282702e-02
3.43515836e-02 5.70421755e-01 -6.95482641e-02 6.36067808e-01
2.74909437e-01 8.27439964e-01 -1.01474416e+00 -6.74614668e-01
4.24269065e-02 3.01824152e-01 -4.42018956e-01 6.02258854e-02
1.84794858e-01 4.17203128e-01 3.57860595e-01 8.68345916e-01
6.39508843e-01 5.49979806e-01 8.31140220e-01 1.84488356e-01
-7.08003640e-01 8.18628669e-01 -8.70296717e-01 1.65388620e+00
-8.93351257e-01 8.20102394e-01 1.33194372e-01 -8.62678885e-01
7.76627600e-01 2.05969349e-01 2.32238337e-01 -6.87627316e-01
-7.71031296e-03 8.81828368e-01 6.86177790e-01 -1.31353363e-01
6.11405730e-01 -3.60486686e-01 -5.88829398e-01 6.54984713e-01
4.43270177e-01 -1.94140613e-01 5.78348339e-01 3.94640453e-02
6.54377937e-01 2.06146449e-01 5.70657432e-01 -1.10885012e+00
5.17334580e-01 1.22685701e-01 5.95666587e-01 6.60384774e-01
-1.54198781e-01 2.76060224e-01 1.48556396e-01 -3.38979661e-01
-1.32629287e+00 -1.23715186e+00 -6.12994730e-01 1.62610114e+00
-1.83100581e-01 -4.13953722e-01 -4.67390031e-01 -5.65937400e-01
-3.34982008e-01 4.78685081e-01 -3.70137393e-01 5.32723665e-01
-1.01651442e+00 -1.19586766e+00 7.46749282e-01 3.25084716e-01
8.39830339e-02 -1.37270188e+00 1.06086992e-01 4.58817691e-01
-9.18642357e-02 -9.91351902e-01 -3.71803284e-01 4.74678904e-01
-3.56113702e-01 -7.94106066e-01 -8.27846825e-01 -1.45442355e+00
5.15876055e-01 -1.46627128e-01 1.58533955e+00 -1.95073888e-01
-1.28947377e-01 -1.53029710e-01 -1.85056999e-01 -2.81334341e-01
-6.49128675e-01 3.94499481e-01 4.75600183e-01 -4.38449800e-01
8.55097890e-01 -4.67846960e-01 1.10061266e-01 9.84077603e-02
-6.48472011e-01 -4.29103747e-02 4.05425280e-01 1.03827286e+00
6.06768727e-01 -3.08316410e-01 6.23415291e-01 -1.29762471e+00
3.11958402e-01 -5.80528140e-01 -6.86523438e-01 2.77452677e-01
-5.20671725e-01 1.19888149e-01 6.95650816e-01 -5.99752188e-01
-9.11356032e-01 -1.60386994e-01 -5.39387941e-01 5.15751839e-01
-7.24633411e-02 5.91311932e-01 -3.78820747e-01 1.05633542e-01
6.32605851e-01 1.23874113e-01 -4.86607671e-01 -6.95475876e-01
6.99640334e-01 9.33749378e-01 4.25649196e-01 -9.85029101e-01
6.80820644e-01 1.77839041e-01 -6.32385790e-01 -1.17364967e+00
-5.17860591e-01 -5.44926584e-01 -1.08304930e+00 1.23642005e-01
5.56848288e-01 -1.21260297e+00 -1.06798664e-01 4.29385424e-01
-8.82723033e-01 -6.05680287e-01 -3.03221285e-01 8.22537661e-01
-4.71205562e-01 4.91698757e-02 -9.32755411e-01 -4.90316182e-01
-2.50606716e-01 -1.25821853e+00 5.73443472e-01 -2.65552223e-01
-3.17517161e-01 -1.39024460e+00 6.39835000e-01 -1.65631384e-01
3.54194045e-01 -1.60832644e-01 1.29193044e+00 -9.01084960e-01
-2.94712037e-01 1.95146382e-01 -2.94094682e-01 4.33468491e-01
3.84210557e-01 -2.45675176e-01 -8.66826773e-01 -3.52197349e-01
-2.20585823e-01 -6.97593033e-01 5.30086875e-01 6.59111217e-02
8.56638029e-02 -1.22521043e-01 1.02324195e-01 5.69983721e-01
1.66585660e+00 8.60657468e-02 3.14314663e-01 3.96445423e-01
7.90357649e-01 6.74888909e-01 4.97196257e-01 -1.89021438e-01
7.56786227e-01 5.41755378e-01 -5.78109145e-01 -3.56079817e-01
-3.48736882e-01 -1.50036320e-01 8.55175257e-01 2.07683849e+00
6.67543262e-02 2.92808086e-01 -1.34600449e+00 1.50093412e+00
-1.29170358e+00 -5.95516503e-01 -4.87486459e-02 2.34416819e+00
1.74145699e+00 1.29632905e-01 2.94446498e-01 -5.75730503e-02
3.84275198e-01 8.17829072e-02 -9.20409486e-02 -6.55646920e-01
-5.54937661e-01 7.08216310e-01 6.31070554e-01 9.46232200e-01
-1.00013173e+00 1.74298906e+00 6.94378567e+00 9.93874669e-01
-1.00846994e+00 6.24705613e-01 3.58243048e-01 2.62743440e-02
-3.94416153e-01 3.25654358e-01 -1.10579538e+00 1.99282169e-01
1.22564518e+00 -1.03615902e-01 7.64380157e-01 5.66286147e-01
-2.09605694e-01 -1.65035829e-01 -1.13888347e+00 8.18187296e-01
8.81333798e-02 -8.56441438e-01 -2.21488521e-01 -2.56050434e-02
9.47105110e-01 7.86248505e-01 -2.04560563e-01 5.83773077e-01
7.84084082e-01 -1.01908779e+00 6.60523355e-01 -1.13333561e-01
1.47723567e+00 -9.59201753e-01 7.46814907e-01 6.25984073e-02
-1.22690058e+00 4.66505736e-01 -5.94500780e-01 -3.09394281e-02
7.43042454e-02 3.78189445e-01 -6.24611259e-01 2.88371056e-01
5.71167409e-01 5.66344023e-01 -6.65021598e-01 4.69057351e-01
-4.57742691e-01 1.03623891e+00 -5.76833963e-01 -2.55216230e-02
4.90462512e-01 -3.93643022e-01 2.96456754e-01 1.76242149e+00
1.14284493e-02 -4.83611047e-01 2.98692882e-01 1.86236277e-01
-1.36738405e-01 8.75627637e-01 -1.05321503e+00 -1.97313234e-01
2.82280147e-01 1.02226269e+00 -7.67708898e-01 -3.79997611e-01
-8.51368904e-01 8.05976987e-01 7.52848506e-01 1.70004070e-01
-5.14412522e-01 -6.54525578e-01 8.01785171e-01 -2.83935312e-02
4.92748320e-02 -5.65888703e-01 -1.61344692e-01 -1.33909559e+00
-1.51683807e-01 -1.21250451e+00 4.31819528e-01 -9.08909068e-02
-1.53722811e+00 9.04807985e-01 1.01082139e-01 -8.98862302e-01
-4.57393885e-01 -9.39247668e-01 -7.19942078e-02 1.21622205e+00
-1.56630707e+00 -1.52569616e+00 6.18752599e-01 6.38811886e-01
5.31121075e-01 -6.30375504e-01 1.18515694e+00 6.26430333e-01
-1.78925246e-01 9.40115392e-01 5.73781550e-01 3.77809197e-01
8.87267530e-01 -1.39035738e+00 6.55381918e-01 7.88697958e-01
7.06251800e-01 8.18139255e-01 4.31255966e-01 -5.29776931e-01
-1.19087851e+00 -7.23560333e-01 1.71056521e+00 -4.78544056e-01
1.23901200e+00 -1.05491185e+00 -6.15734875e-01 9.59079266e-01
5.35811365e-01 -4.09919083e-01 9.55205560e-01 9.40702140e-01
-5.26710391e-01 -6.16108403e-02 -9.05527949e-01 8.76828909e-01
1.08653426e+00 -1.03475320e+00 -6.57425523e-01 6.50429130e-01
3.29577833e-01 6.48204163e-02 -9.30813253e-01 -3.25762527e-03
5.51290870e-01 -6.33017957e-01 7.69479930e-01 -4.27835256e-01
9.86646190e-02 -1.57349005e-01 -3.46398711e-01 -1.55090320e+00
-6.43125325e-02 -6.22088373e-01 7.52257049e-01 1.46079862e+00
7.54734755e-01 -7.31777906e-01 1.81636482e-01 -2.78072000e-01
7.55948871e-02 -3.45372260e-01 -8.89604330e-01 -1.02137315e+00
7.37680018e-01 -6.22126698e-01 5.43542147e-01 1.23252642e+00
2.11602747e-01 6.43212020e-01 -3.14005584e-01 -2.39280492e-01
3.97637904e-01 2.07077682e-01 6.51866078e-01 -1.03113639e+00
-1.84834272e-01 -3.44386458e-01 -2.77399957e-01 -6.63313925e-01
5.25390744e-01 -1.43079638e+00 3.18178535e-01 -1.20101929e+00
3.47892493e-01 -1.04381716e+00 -3.61808777e-01 5.35649180e-01
-4.54041511e-02 7.76573658e-01 2.54271226e-03 2.48480365e-01
-2.20577970e-01 1.63865522e-01 7.07044244e-01 -1.99188709e-01
-3.50288659e-01 -5.67978024e-01 -7.50065565e-01 7.85348654e-01
6.66774392e-01 -6.81002080e-01 -2.74891078e-01 -7.50533283e-01
2.92649806e-01 -5.26558518e-01 -5.78251541e-01 -6.86645627e-01
-2.34924778e-01 -1.84414372e-01 4.01949398e-02 -3.89950812e-01
-9.22752079e-03 -4.63141739e-01 -1.61905006e-01 8.40384886e-02
-4.73019555e-02 4.21327442e-01 3.17388773e-01 -5.10197937e-01
-3.66729587e-01 -3.37060779e-01 8.75122011e-01 -2.05272332e-01
-7.16163158e-01 3.25613856e-01 -7.04124749e-01 7.26402819e-01
5.12523413e-01 1.88751042e-01 1.03776775e-01 -1.76453471e-01
-4.19573188e-01 -1.20380357e-01 6.84389710e-01 6.13344848e-01
9.09776315e-02 -1.42486191e+00 -1.19454288e+00 7.38560200e-01
5.50076187e-01 -5.49168766e-01 -4.65368807e-01 7.03284264e-01
-9.43549097e-01 4.18385774e-01 -4.18838084e-01 -3.31955820e-01
-8.59168708e-01 3.41413200e-01 -9.86591652e-02 -5.28653204e-01
-4.32317853e-01 7.71832883e-01 2.39461213e-01 -1.28695273e+00
-1.38710141e-01 -5.74470721e-02 -1.54088035e-01 3.40265930e-01
2.91276246e-01 -1.65295631e-01 -1.35733075e-02 -1.39430952e+00
-6.46795332e-01 4.49864805e-01 -3.63890439e-01 -6.96018577e-01
1.26644778e+00 -7.95339271e-02 -4.50040728e-01 8.90830934e-01
1.17633116e+00 9.80371952e-01 -4.20887470e-01 -3.29305321e-01
4.03250545e-01 -1.70752376e-01 -2.18829811e-01 -8.01148832e-01
-5.40938616e-01 8.68318975e-01 4.58818197e-01 -2.33066857e-01
8.54484856e-01 1.31476820e-01 5.43031156e-01 2.27482408e-01
7.70803750e-01 -1.38427067e+00 -7.11550593e-01 1.15684104e+00
5.13144493e-01 -1.18430281e+00 1.72185677e-03 9.30992216e-02
-5.56379735e-01 6.94493473e-01 2.73591638e-01 -1.21483609e-01
8.26584756e-01 4.29466099e-01 7.24569082e-01 -3.32414694e-02
-2.99608916e-01 -6.22234225e-01 3.31404842e-02 5.74062884e-01
9.71562326e-01 4.76114601e-01 -8.47515345e-01 4.19762284e-01
-7.30179608e-01 -6.79266274e-01 2.94489861e-01 7.68103302e-01
2.73001827e-02 -1.62530327e+00 -2.30673403e-01 3.09445411e-01
-7.72720397e-01 -8.87579858e-01 -2.70897597e-01 1.37793577e+00
4.18739945e-01 7.13215828e-01 1.99038669e-01 -8.38496685e-02
-3.75258625e-02 2.54148781e-01 7.46793568e-01 -7.42890000e-01
-7.15024412e-01 2.19173074e-01 3.72278988e-01 -1.06880121e-01
-3.80404085e-01 -8.89338434e-01 -9.27089691e-01 -3.61984760e-01
-1.88080177e-01 2.15349987e-01 6.47736549e-01 8.68497908e-01
-4.23132986e-01 -2.99245536e-01 4.96579945e-01 -7.53691912e-01
-3.07867564e-02 -1.32871473e+00 -8.40586126e-01 5.06380260e-01
-4.43088822e-02 -4.07091916e-01 -1.76860705e-01 -8.61376449e-02] | [10.835322380065918, 10.074652671813965] |
46d7b79b-6800-4fa9-9798-7516f00a7373 | clip3dstyler-language-guided-3d-arbitrary | 2305.15732 | null | https://arxiv.org/abs/2305.15732v2 | https://arxiv.org/pdf/2305.15732v2.pdf | CLIP3Dstyler: Language Guided 3D Arbitrary Neural Style Transfer | In this paper, we propose a novel language-guided 3D arbitrary neural style transfer method (CLIP3Dstyler). We aim at stylizing any 3D scene with an arbitrary style from a text description, and synthesizing the novel stylized view, which is more flexible than the image-conditioned style transfer. Compared with the previous 2D method CLIPStyler, we are able to stylize a 3D scene and generalize to novel scenes without re-train our model. A straightforward solution is to combine previous image-conditioned 3D style transfer and text-conditioned 2D style transfer \bigskip methods. However, such a solution cannot achieve our goal due to two main challenges. First, there is no multi-modal model matching point clouds and language at different feature scales (low-level, high-level). Second, we observe a style mixing issue when we stylize the content with different style conditions from text prompts. To address the first issue, we propose a 3D stylization framework to match the point cloud features with text features in local and global views. For the second issue, we propose an improved directional divergence loss to make arbitrary text styles more distinguishable as a complement to our framework. We conduct extensive experiments to show the effectiveness of our model on text-guided 3D scene style transfer. | ['Mingming Gong', 'Chenkai Zhao', 'Tingbo Hou', 'Yang Zhao', 'Yanwu Xu', 'Ming Gao'] | 2023-05-25 | null | null | null | null | ['style-transfer'] | ['computer-vision'] | [ 1.87541068e-01 -2.94623435e-01 2.69822329e-01 -3.37748528e-01
-6.40741110e-01 -8.35434854e-01 7.20191419e-01 -5.00748038e-01
-1.01757504e-01 4.07075793e-01 3.03961430e-02 -3.76252793e-02
2.80957431e-01 -7.00458765e-01 -8.11080277e-01 -5.29733360e-01
7.92153716e-01 6.59382164e-01 2.53334820e-01 -2.10849538e-01
2.85040677e-01 7.90394425e-01 -1.23185182e+00 2.87047505e-01
9.69349384e-01 8.19303751e-01 5.90230525e-01 4.43249345e-01
-7.26947486e-01 1.56347603e-01 -5.65422118e-01 -3.34984422e-01
5.22454262e-01 -5.99172652e-01 -7.06619501e-01 6.08363390e-01
9.03498769e-01 -3.79765868e-01 -1.11710601e-01 1.18384624e+00
4.24720615e-01 1.02628604e-01 8.46685469e-01 -1.17962384e+00
-9.62503374e-01 5.23272110e-03 -8.27159762e-01 -4.22494233e-01
5.15255928e-01 1.92517877e-01 5.37997186e-01 -1.07334089e+00
8.71157467e-01 1.77074325e+00 4.62507695e-01 8.12682569e-01
-1.48937917e+00 -6.48086667e-01 5.32397270e-01 -3.39415371e-01
-1.23033261e+00 -8.62142518e-02 1.18869221e+00 -5.52144289e-01
3.40996265e-01 1.83286145e-01 6.59863830e-01 1.38343573e+00
1.15976624e-01 8.82558227e-01 1.42098570e+00 -2.16221169e-01
1.36501640e-01 3.80753249e-01 -3.23859394e-01 5.49904168e-01
-2.19827414e-01 4.40766029e-02 -4.59050953e-01 9.98588130e-02
1.24637544e+00 2.56919742e-01 -2.85356134e-01 -6.66471601e-01
-1.41343343e+00 6.91561580e-01 4.45561826e-01 2.44956464e-01
-1.74044329e-03 2.36211531e-02 2.64960736e-01 5.30829787e-01
9.98435676e-01 3.97370994e-01 -3.94091785e-01 -2.74637025e-02
-9.47006643e-01 3.39702010e-01 4.88365918e-01 1.48765850e+00
9.88949239e-01 1.81389242e-01 -2.96779007e-01 6.51109695e-01
8.81059691e-02 9.38835502e-01 1.75301060e-01 -9.18538511e-01
3.71723413e-01 4.41832989e-01 1.85884610e-02 -9.17462468e-01
-7.17967302e-02 -1.85378492e-01 -9.75587547e-01 5.94251931e-01
2.53101915e-01 -1.75230820e-02 -9.71862316e-01 1.63795733e+00
3.10094833e-01 1.33856699e-01 -1.96934670e-01 1.03313851e+00
8.71873975e-01 6.91290259e-01 -1.96165919e-01 1.45894602e-01
1.18170929e+00 -8.90894771e-01 -5.46671748e-01 -1.98151898e-02
5.26465297e-01 -1.05146646e+00 1.65723944e+00 1.26044646e-01
-1.07928538e+00 -7.85034895e-01 -7.68888116e-01 -3.31239432e-01
-3.08175772e-01 -7.69878253e-02 2.70506591e-01 2.68559068e-01
-1.07230008e+00 4.31958050e-01 -5.90784013e-01 -7.16818929e-01
2.08463594e-01 -1.16029464e-01 -4.37813401e-01 -4.82676029e-02
-8.33233893e-01 7.71651447e-01 2.50586092e-01 -3.30752611e-01
-5.12197077e-01 -7.73260593e-01 -8.59761536e-01 -1.48071706e-01
2.47771993e-01 -1.25598955e+00 9.80993092e-01 -1.17572999e+00
-1.87710464e+00 1.27301192e+00 -4.35804754e-01 2.11090297e-01
8.35744083e-01 -3.31573039e-01 -3.16886194e-02 3.45413201e-02
2.18019918e-01 8.05301845e-01 1.11563003e+00 -1.87383127e+00
-4.47169036e-01 -3.15549284e-01 1.55582845e-01 6.04827344e-01
-9.19150189e-02 -2.19035193e-01 -7.17281640e-01 -1.06466973e+00
2.29295269e-01 -9.60149825e-01 -7.29543194e-02 4.04729873e-01
-3.55298758e-01 1.48296691e-02 1.13850570e+00 -2.85708964e-01
5.57558119e-01 -2.36639190e+00 5.72292745e-01 -3.03464457e-02
2.17848971e-01 1.15970455e-01 -3.35999280e-01 3.41370195e-01
4.16555488e-03 1.55728266e-01 -3.03088397e-01 -8.63548875e-01
2.55277716e-02 2.93300986e-01 -6.29104316e-01 -4.01089899e-02
3.61896813e-01 9.19241428e-01 -9.46649373e-01 -4.56844032e-01
4.65172142e-01 3.50993812e-01 -7.40326583e-01 4.45545197e-01
-4.82783407e-01 1.12582397e+00 -6.09499097e-01 4.12968636e-01
1.06167400e+00 -1.56779423e-01 -4.38937873e-01 -3.45258713e-01
-8.68162587e-02 -5.85368574e-02 -1.17931986e+00 2.42989993e+00
-7.65556157e-01 3.11428040e-01 1.10318132e-01 -7.80168772e-01
1.00324774e+00 1.09550148e-01 2.94736862e-01 -5.06750822e-01
3.00616841e-03 1.61862224e-01 -7.49684632e-01 -3.80931079e-01
5.46903491e-01 -5.37224948e-01 -6.63520172e-02 4.26270217e-01
2.00483456e-01 -1.01716375e+00 -3.40439916e-01 3.00114781e-01
5.83961785e-01 6.48493350e-01 -2.63662934e-01 -3.28293234e-01
4.73237187e-01 -2.71765113e-01 3.17490071e-01 5.94744205e-01
1.41298652e-01 1.13134575e+00 4.37099159e-01 -2.46191293e-01
-8.68589520e-01 -1.36201847e+00 1.07999509e-02 7.83365965e-01
4.27067608e-01 -3.00719649e-01 -7.16930032e-01 -9.99829888e-01
4.40582111e-02 7.88290381e-01 -6.28003001e-01 7.11863395e-03
-4.64946151e-01 -5.33448579e-03 1.85787708e-01 2.80766308e-01
7.14142561e-01 -8.65921795e-01 -9.81315225e-02 -7.42565021e-02
-2.31885895e-01 -1.27643430e+00 -1.05448878e+00 -4.62915897e-02
-9.15170729e-01 -4.97903556e-01 -1.20166993e+00 -8.35531771e-01
8.09924483e-01 5.60973346e-01 1.18361306e+00 -2.74494618e-01
2.48348266e-01 5.70323229e-01 -3.84999603e-01 -3.74584317e-01
-5.48690259e-01 4.47828099e-02 1.58454955e-01 1.93794206e-01
3.44778178e-03 -8.74950588e-01 -3.89468729e-01 3.08456182e-01
-1.23041821e+00 5.56166828e-01 5.30653894e-01 7.72510529e-01
6.22768104e-01 -3.46655548e-01 1.56385556e-01 -8.82257521e-01
5.23771226e-01 7.99189359e-02 -4.63172942e-01 1.72018692e-01
-1.71699777e-01 1.83243454e-01 7.39275753e-01 -6.90624535e-01
-1.11147487e+00 3.52147780e-03 -2.40505531e-01 -1.03468835e+00
-4.75084722e-01 1.12620324e-01 -4.59523201e-01 -1.36243269e-01
5.28556883e-01 4.06078875e-01 7.22523853e-02 -7.50062704e-01
7.63724625e-01 3.77077401e-01 4.11191642e-01 -9.64428484e-01
1.42472267e+00 7.41181493e-01 2.56447718e-02 -8.12408268e-01
-9.97296274e-01 -2.18155608e-01 -9.86661196e-01 -9.73594040e-02
1.09562767e+00 -8.19493711e-01 -4.45530027e-01 5.59861243e-01
-1.48289680e+00 -4.32798117e-01 -4.85031813e-01 2.98552901e-01
-8.22768688e-01 5.88635266e-01 -3.93054813e-01 -2.94143289e-01
-2.14650869e-01 -1.22805703e+00 1.69090331e+00 -1.09787703e-01
-2.12532133e-01 -1.03692329e+00 2.42669396e-02 3.33977789e-02
3.52132440e-01 3.78186703e-01 9.06433284e-01 -3.59616503e-02
-5.06714880e-01 2.17318833e-01 -5.28496981e-01 2.96993285e-01
5.44649363e-01 -1.92752197e-01 -9.22567308e-01 -3.60863447e-01
2.57570505e-01 -2.59233207e-01 6.22879148e-01 1.10168725e-01
1.19565547e+00 -4.36476171e-02 -1.26531333e-01 1.15912473e+00
1.28410888e+00 7.04898164e-02 2.91367531e-01 6.97294846e-02
1.16454268e+00 5.21406949e-01 4.79031980e-01 1.60086378e-01
2.04520926e-01 9.35274541e-01 1.87333480e-01 -4.40775603e-01
-3.25439185e-01 -5.73437572e-01 3.29067588e-01 9.46388960e-01
-4.87171300e-02 -1.95448771e-01 -5.20503879e-01 1.80842340e-01
-1.63260365e+00 -6.95524514e-01 4.65156548e-02 1.99862266e+00
6.69236839e-01 1.55774906e-01 -2.02131905e-02 -2.50935614e-01
4.97044206e-01 1.72825560e-01 -6.87702298e-01 -3.40348661e-01
-1.89383820e-01 1.59825757e-01 1.06877111e-01 6.28006279e-01
-8.47740114e-01 1.36494327e+00 5.20884943e+00 1.08152330e+00
-1.38537776e+00 1.96060631e-02 2.88161635e-01 -5.23628183e-02
-7.36884296e-01 -4.24598828e-02 -6.83155179e-01 4.77756530e-01
4.34882753e-02 -9.21480358e-02 4.49458480e-01 5.69881976e-01
2.13775754e-01 2.74626285e-01 -1.39043367e+00 1.33226562e+00
2.89036095e-01 -1.29237401e+00 7.21041739e-01 -6.25410452e-02
8.12292159e-01 -2.04744324e-01 2.64215827e-01 2.50068575e-01
2.29387432e-01 -6.70658052e-01 1.04402459e+00 6.38409257e-01
1.16918993e+00 -5.86314559e-01 1.12367518e-01 3.73769999e-01
-1.06088889e+00 5.12713850e-01 -2.57038295e-01 1.35527879e-01
4.22802329e-01 5.70437312e-01 -2.02984989e-01 9.05153036e-01
8.42761874e-01 9.03662205e-01 -5.63206136e-01 5.59001982e-01
-6.80198669e-02 5.74515946e-02 -2.46691272e-01 1.57390431e-01
4.00478750e-01 -5.82632065e-01 8.78166139e-01 9.94978368e-01
5.92987478e-01 3.47545594e-02 2.25017399e-01 1.36597586e+00
2.11949944e-02 7.82671198e-02 -1.08123624e+00 3.72363746e-01
5.24328742e-03 1.06388366e+00 -6.05760276e-01 -4.44038153e-01
-4.79513377e-01 1.75619853e+00 3.16987216e-01 6.81024730e-01
-6.36403859e-01 -3.34490508e-01 5.93996942e-01 1.31214142e-01
1.05496146e-01 -5.24694622e-01 -5.49533546e-01 -1.69218957e+00
5.06434664e-02 -7.88517416e-01 -1.98083401e-01 -1.21100819e+00
-1.72146976e+00 6.43148065e-01 9.73381028e-02 -1.74350429e+00
1.21103019e-01 -5.72757185e-01 -6.81916893e-01 1.18606830e+00
-1.32370377e+00 -1.53033173e+00 -4.11532700e-01 8.44026566e-01
7.89859414e-01 -2.98076347e-02 6.11549139e-01 1.91083699e-01
3.74760702e-02 5.79736352e-01 5.50846756e-02 -1.58641174e-01
1.23124290e+00 -1.35750437e+00 6.43787503e-01 6.26140535e-01
1.35193780e-01 5.06236732e-01 6.16065502e-01 -5.94776034e-01
-1.29928887e+00 -1.21310008e+00 5.37602067e-01 -8.27745616e-01
4.64139402e-01 -8.27296615e-01 -1.01943755e+00 7.05257952e-01
2.78088152e-01 -3.29052895e-01 3.03051978e-01 -1.25020981e-01
-3.02212179e-01 -1.04316190e-01 -1.04839933e+00 1.07882988e+00
1.47154605e+00 -6.11719072e-01 -5.67557871e-01 2.13854671e-01
1.14508033e+00 -5.08049428e-01 -7.23076344e-01 4.51363415e-01
2.38092631e-01 -8.62960517e-01 8.42089832e-01 -4.20372725e-01
4.50208366e-01 -4.50434268e-01 -2.66769409e-01 -1.68783295e+00
-3.99725437e-01 -7.62405336e-01 2.84627765e-01 1.44331825e+00
4.88412827e-02 -6.17729306e-01 6.93008304e-01 3.11856031e-01
-3.62019390e-01 -3.94359231e-01 -7.57881343e-01 -9.75450754e-01
5.82007647e-01 -3.34328532e-01 7.69468904e-01 1.08505857e+00
-4.32706147e-01 5.00351727e-01 -4.72980857e-01 9.10632610e-02
6.88862920e-01 5.34347713e-01 1.24379075e+00 -9.96739924e-01
-3.12919348e-01 -6.10163927e-01 -1.62420943e-01 -1.70188558e+00
2.72701919e-01 -9.35486078e-01 -1.27533779e-01 -1.37774611e+00
5.81734106e-02 -5.04620314e-01 1.36215091e-01 1.40160893e-03
-1.82329655e-01 1.63667783e-01 5.60575724e-01 2.24902689e-01
-1.68102950e-01 9.61129785e-01 2.01938701e+00 -2.07142100e-01
-2.34546348e-01 -9.23487470e-02 -6.15276337e-01 7.47981489e-01
5.71548939e-01 -1.62528768e-01 -5.00766814e-01 -8.72530520e-01
-4.10562009e-02 1.59920916e-01 4.69462246e-01 -8.15329969e-01
-1.76471576e-01 -4.24207926e-01 3.86644065e-01 -9.38381732e-01
4.30294126e-01 -9.69414830e-01 5.66135459e-02 1.35439411e-01
-2.52175838e-01 2.76212655e-02 2.73400694e-01 5.25010407e-01
-9.98735577e-02 1.31466910e-01 8.57766807e-01 -1.68902919e-01
-4.82326031e-01 7.57537544e-01 -1.72165737e-01 3.90844718e-02
7.74653792e-01 -2.62689143e-01 -4.66146804e-02 -3.78408283e-01
-7.56112278e-01 1.57718122e-01 1.08829224e+00 7.81145215e-01
6.75733864e-01 -1.72738385e+00 -6.52070761e-01 4.95407104e-01
3.75408649e-01 3.70552391e-01 2.86806852e-01 4.87171203e-01
-4.71029490e-01 1.25198841e-01 -1.96770161e-01 -9.26527917e-01
-1.00566816e+00 8.33169103e-01 2.72691011e-01 -1.26836717e-01
-8.53661120e-01 5.58068633e-01 1.16298068e+00 -8.95070374e-01
9.23539624e-02 -4.70591098e-01 2.37200215e-01 -3.44461769e-01
1.66172251e-01 -1.61449075e-01 -2.18184099e-01 -5.64286172e-01
1.73144471e-02 1.31230712e+00 6.09881878e-02 -3.91201437e-01
9.43183839e-01 -3.60700607e-01 -1.10953348e-02 8.38303387e-01
1.27251840e+00 1.36109382e-01 -1.45337045e+00 -5.20399511e-01
-4.91190374e-01 -6.34103715e-01 -2.98703969e-01 -4.55025107e-01
-1.00729585e+00 1.21816158e+00 5.62343359e-01 -7.35914707e-03
1.07991171e+00 1.58829857e-02 8.93001556e-01 2.50592083e-01
4.10105586e-01 -7.63588846e-01 3.84630173e-01 8.08717728e-01
1.35395014e+00 -1.18109536e+00 -3.49268258e-01 -4.59476858e-01
-6.60993099e-01 9.10900474e-01 8.58378291e-01 -1.84108034e-01
7.29070485e-01 -4.51796055e-02 2.94071823e-01 -1.63032308e-01
-3.78502816e-01 -1.70749739e-01 5.40025413e-01 7.30797648e-01
1.38450593e-01 -1.15692921e-01 8.63087475e-02 1.93705216e-01
-3.02968442e-01 -1.44734830e-01 2.22587451e-01 5.74213207e-01
-1.02220453e-01 -1.19272971e+00 -4.18232977e-01 6.67962134e-02
7.48896003e-02 -8.39073882e-02 -5.31823158e-01 6.63533866e-01
-7.86413532e-03 5.36597013e-01 1.31167024e-01 -5.14812231e-01
6.96526289e-01 2.63294857e-02 6.90093696e-01 -8.67610693e-01
-2.37392902e-01 3.68751854e-01 -3.85876536e-01 -4.99749869e-01
-3.89021128e-01 -4.40323979e-01 -8.34059477e-01 -4.14835900e-01
-7.92063847e-02 -1.28845155e-01 4.27007049e-01 8.13595355e-01
4.28947419e-01 4.03873175e-01 9.69022036e-01 -1.45394135e+00
-1.27979904e-01 -7.80930221e-01 -7.42182732e-01 9.07234669e-01
2.35123858e-01 -7.81511128e-01 -4.56363320e-01 2.40216970e-01] | [9.25784683227539, -3.282867670059204] |
fa112a7e-de67-4fe5-a757-c6504519a563 | semantic-aware-dynamic-retrospective | 2305.08059 | null | https://arxiv.org/abs/2305.08059v1 | https://arxiv.org/pdf/2305.08059v1.pdf | Semantic-aware Dynamic Retrospective-Prospective Reasoning for Event-level Video Question Answering | Event-Level Video Question Answering (EVQA) requires complex reasoning across video events to obtain the visual information needed to provide optimal answers. However, despite significant progress in model performance, few studies have focused on using the explicit semantic connections between the question and visual information especially at the event level. There is need for using such semantic connections to facilitate complex reasoning across video frames. Therefore, we propose a semantic-aware dynamic retrospective-prospective reasoning approach for video-based question answering. Specifically, we explicitly use the Semantic Role Labeling (SRL) structure of the question in the dynamic reasoning process where we decide to move to the next frame based on which part of the SRL structure (agent, verb, patient, etc.) of the question is being focused on. We conduct experiments on a benchmark EVQA dataset - TrafficQA. Results show that our proposed approach achieves superior performance compared to previous state-of-the-art models. Our code will be made publicly available for research use. | ['Jennifer Foster', 'Yvette Graham', 'Tianbo Ji', 'Chenyang Lyu'] | 2023-05-14 | null | null | null | null | ['video-question-answering', 'semantic-role-labeling'] | ['computer-vision', 'natural-language-processing'] | [ 1.22614786e-01 5.60135171e-02 -1.41159758e-01 -6.16182804e-01
-5.78278303e-01 -5.46213925e-01 3.80941004e-01 2.01900437e-01
-2.88777739e-01 6.57403111e-01 6.70646846e-01 -4.78186637e-01
-9.05124694e-02 -6.46104217e-01 -6.43077493e-01 -3.74582291e-01
2.33429074e-01 2.86816090e-01 8.94393325e-01 -2.46296257e-01
2.21892402e-01 3.25026482e-01 -1.44557345e+00 9.24239516e-01
3.97274256e-01 8.53990316e-01 3.02479714e-01 6.43795490e-01
-4.56988364e-01 1.64309216e+00 -5.81322908e-01 -3.87462318e-01
7.10266903e-02 -8.08892965e-01 -1.26501071e+00 2.89611369e-01
4.11370009e-01 -5.38799882e-01 -3.60092580e-01 8.31850111e-01
2.42659211e-01 3.90340835e-01 2.52270430e-01 -1.32410669e+00
-3.46853435e-01 2.75297970e-01 -3.51412207e-01 8.50950301e-01
9.33916926e-01 2.69875348e-01 1.08159637e+00 -6.69495225e-01
1.13834107e+00 1.35815084e+00 2.51460783e-02 6.25283837e-01
-4.97204632e-01 -4.16789234e-01 5.89024067e-01 1.05129063e+00
-8.70188117e-01 -3.84397238e-01 1.06901991e+00 -2.39294872e-01
8.14597309e-01 2.80150592e-01 8.85940015e-01 9.19785142e-01
2.98779979e-02 1.18364024e+00 9.24225748e-01 -2.71235585e-01
4.02519614e-01 -1.64305091e-01 1.68739095e-01 9.97772694e-01
-2.64507443e-01 -4.03672129e-01 -6.54183984e-01 -3.43933813e-02
7.09247589e-01 -6.64589647e-03 -2.42179886e-01 -4.36573118e-01
-1.19370067e+00 5.69661021e-01 4.47234035e-01 3.17568332e-01
-6.71669543e-01 3.87353122e-01 3.71746451e-01 2.09501356e-01
2.22907215e-02 9.37992781e-02 -3.62082601e-01 -3.48144770e-01
-7.08738029e-01 3.74588400e-01 5.99702716e-01 6.66801393e-01
3.39360803e-01 -2.36643046e-01 -5.60931146e-01 5.27902782e-01
4.09153789e-01 -9.57502350e-02 3.00130993e-02 -1.39752018e+00
6.55432105e-01 9.41003859e-01 2.28725135e-01 -9.65907276e-01
-3.89788121e-01 1.76901042e-01 -1.38856128e-01 1.05202481e-01
4.71510023e-01 -3.89720537e-02 -8.96322548e-01 1.39208162e+00
8.83444011e-01 4.07349020e-01 2.11795241e-01 1.18119645e+00
1.29747462e+00 9.15177286e-01 5.50529003e-01 -2.12995514e-01
1.91759074e+00 -1.20177197e+00 -1.02351534e+00 -8.08089674e-02
4.60319072e-01 -6.69824600e-01 1.13672709e+00 1.20068148e-01
-1.36639774e+00 -5.02696753e-01 -7.72596717e-01 -4.47801203e-01
-1.39031008e-01 -1.95927694e-01 4.57772404e-01 2.46343225e-01
-9.74821627e-01 -1.87141150e-01 -7.41523683e-01 -4.61559594e-01
4.97651458e-01 -4.59317788e-02 -3.82802963e-01 -3.91113818e-01
-1.27881086e+00 6.96939230e-01 3.08727533e-01 3.28726843e-02
-1.04163194e+00 -4.38061833e-01 -7.70625591e-01 1.72606856e-02
7.64410615e-01 -9.58742678e-01 1.44846952e+00 -1.12840271e+00
-1.25893629e+00 7.03732610e-01 -4.24960107e-01 -4.00506139e-01
3.87079537e-01 -1.99724078e-01 -3.90267462e-01 1.07085645e+00
-3.93029675e-02 9.12747860e-01 7.55629182e-01 -1.31975400e+00
-9.26026940e-01 -3.79200518e-01 1.06936955e+00 5.82327783e-01
1.11132294e-01 4.63929296e-01 -9.06537473e-01 -5.43417454e-01
5.16805537e-02 -6.51713252e-01 -3.00974082e-02 3.82692933e-01
1.30956575e-01 -6.03235006e-01 1.14131534e+00 -1.01650143e+00
1.28650630e+00 -2.07442594e+00 1.41106904e-01 -2.11706206e-01
7.86073506e-02 1.66046605e-01 3.17025408e-02 3.87235105e-01
-6.10691607e-02 -2.39755929e-01 -1.16573811e-01 1.10838927e-01
-4.14101899e-01 4.05534446e-01 -1.37328014e-01 2.71219343e-01
6.75616413e-02 9.50229824e-01 -1.05052769e+00 -1.02197504e+00
3.17680657e-01 3.94185871e-01 -7.06616342e-01 4.15200740e-01
-5.39840639e-01 4.88341838e-01 -9.15957928e-01 8.59525919e-01
3.21070164e-01 -5.13095975e-01 1.21750399e-01 -5.30604184e-01
1.51446432e-01 3.20703954e-01 -1.07106280e+00 1.79217982e+00
-1.50948122e-01 4.08352882e-01 1.15786135e-01 -9.11495209e-01
3.42312008e-01 5.23764551e-01 6.32656336e-01 -8.10458362e-01
3.58200222e-02 -3.60906720e-01 7.93166831e-02 -9.64106321e-01
2.91318446e-01 -4.64892425e-02 1.08824618e-01 2.54122078e-01
-9.06619653e-02 1.46942049e-01 5.41486561e-01 5.67198157e-01
1.04720104e+00 5.57204247e-01 3.75892878e-01 1.33305654e-01
9.25728321e-01 3.98485154e-01 4.81468111e-01 4.97436374e-01
-6.14275157e-01 4.23749328e-01 8.29003215e-01 -5.71063757e-01
-6.54922009e-01 -9.26004469e-01 3.27500373e-01 1.21086729e+00
5.57302475e-01 -4.10523862e-01 -7.99497604e-01 -9.73585844e-01
-5.10857642e-01 8.22962523e-01 -5.93808889e-01 1.99454248e-01
-9.22979355e-01 -2.26609796e-01 1.37864798e-01 5.88543355e-01
7.65212953e-01 -1.42683351e+00 -1.00134230e+00 2.51128703e-01
-6.70502543e-01 -1.39413679e+00 -3.31832439e-01 -5.85029006e-01
-8.70132744e-01 -1.52199054e+00 -3.98504227e-01 -7.02893972e-01
5.23577154e-01 3.57141346e-01 1.08716238e+00 2.85757661e-01
-9.85441580e-02 8.86257768e-01 -7.70593584e-01 -1.34817421e-01
-6.58999830e-02 -5.00255346e-01 -7.01694846e-01 2.54181802e-01
1.27499431e-01 -1.86749082e-02 -1.05225134e+00 3.07917953e-01
-1.00590360e+00 3.39285344e-01 1.51296690e-01 2.91339815e-01
6.43150330e-01 -4.99867927e-03 4.48918998e-01 -8.42546165e-01
4.63976502e-01 -4.78476942e-01 -3.14500630e-01 5.71966767e-01
1.64940074e-01 5.74340783e-02 3.02644759e-01 -1.03801288e-01
-1.63031495e+00 2.36382913e-02 -3.58941942e-01 -3.23751450e-01
-2.75299758e-01 2.99438655e-01 -1.11280568e-01 3.50148559e-01
1.69782832e-01 1.28800854e-01 -2.88696647e-01 -1.26940638e-01
3.90259296e-01 2.43358389e-01 3.97871941e-01 -5.68022251e-01
3.52928847e-01 8.32985699e-01 -1.45320082e-02 -4.35404360e-01
-1.06399226e+00 -7.07753599e-01 -2.42152035e-01 -7.67058432e-01
1.48483932e+00 -8.98681343e-01 -9.19913650e-01 -3.29395421e-02
-1.18202496e+00 -1.66288421e-01 -1.84205323e-01 1.87234059e-01
-6.99074268e-01 4.35232967e-01 -5.31837642e-01 -8.09183955e-01
-2.04618037e-01 -1.22407043e+00 9.09430087e-01 2.97598600e-01
-8.80384669e-02 -8.57166231e-01 -1.35634720e-01 1.12107813e+00
2.48711128e-02 2.66991198e-01 9.09599602e-01 -4.13515210e-01
-9.38932180e-01 4.12074216e-02 -3.77501845e-01 -1.24816686e-01
-1.07916623e-01 -4.11599427e-02 -4.62189496e-01 1.88166842e-01
2.31440946e-01 -1.26469210e-01 5.58518350e-01 2.96410561e-01
1.35373020e+00 -1.81808889e-01 -2.57074386e-01 1.25547260e-01
1.16191602e+00 6.33980572e-01 8.29336762e-01 3.10012817e-01
7.10175037e-01 7.14358509e-01 1.09894633e+00 3.13073516e-01
9.90706205e-01 6.93335295e-01 6.23372197e-01 5.51173873e-02
-5.14025092e-01 -2.55340695e-01 4.13313895e-01 4.24748421e-01
-2.20921814e-01 -3.39852601e-01 -8.20352316e-01 6.48236513e-01
-2.07169795e+00 -1.30669546e+00 -3.77548516e-01 1.51623082e+00
5.30969679e-01 7.15010287e-03 1.45588532e-01 1.05118468e-01
5.45561135e-01 4.09952432e-01 -3.75618428e-01 -3.91539067e-01
2.13698640e-01 -1.26427397e-01 -1.26233965e-01 3.01115066e-01
-9.46251750e-01 9.20040429e-01 5.94335985e+00 5.39015591e-01
-6.15203857e-01 3.65788102e-01 5.84584832e-01 -6.18631169e-02
-1.94435000e-01 1.21176280e-01 -6.36152506e-01 1.95194021e-01
6.33165121e-01 2.91581720e-01 1.32403105e-01 4.56375420e-01
4.89692807e-01 -6.58591092e-01 -9.87477899e-01 9.92661059e-01
2.26731956e-01 -1.41958523e+00 3.02612096e-01 -5.51630497e-01
4.70097035e-01 -5.46159208e-01 -3.81190270e-01 1.83181971e-01
1.23155378e-02 -5.55569291e-01 7.64963686e-01 6.50620222e-01
3.02718639e-01 -5.49056530e-01 5.55124342e-01 1.84105411e-01
-1.53492951e+00 -2.41792336e-01 -5.48263080e-02 1.17903098e-01
5.81337810e-01 4.91620153e-02 -7.35584259e-01 6.51313782e-01
9.84619915e-01 6.49314880e-01 -5.86375117e-01 8.24592054e-01
-4.23295200e-01 6.16875291e-01 6.71072230e-02 6.46720380e-02
3.15611571e-01 3.64256352e-02 3.95465136e-01 8.09052944e-01
1.47639923e-02 7.83395767e-01 1.57215238e-01 2.81469524e-01
3.33113819e-02 2.00528488e-01 -2.66046405e-01 -2.38695852e-02
5.39203249e-02 1.02550530e+00 -1.00575840e+00 -6.18678510e-01
-8.64984691e-01 9.97618437e-01 1.95201680e-01 4.46350485e-01
-1.05249035e+00 -3.72978207e-03 5.27569056e-01 4.02709365e-01
4.43055838e-01 -9.77870449e-02 1.47650033e-01 -1.13515639e+00
9.98817459e-02 -9.46663499e-01 1.22786939e+00 -1.50208759e+00
-8.50891650e-01 4.88857120e-01 3.30429316e-01 -1.03859913e+00
-9.98854041e-02 -4.59392965e-01 -3.97477031e-01 3.10211509e-01
-1.57532609e+00 -1.10187030e+00 -4.92173880e-01 1.09893703e+00
1.07320786e+00 2.65495718e-01 3.61165732e-01 2.92180330e-01
-1.37509048e-01 -4.81611378e-02 -7.90907681e-01 1.86376154e-01
4.37176049e-01 -8.99962068e-01 -1.08174257e-01 8.51974130e-01
2.19442710e-01 1.92338139e-01 6.87473893e-01 -7.03688025e-01
-1.39877880e+00 -7.21937358e-01 7.68155336e-01 -6.80635214e-01
5.89260817e-01 9.42739025e-02 -8.42144012e-01 6.55755401e-01
4.28871453e-01 8.10349286e-02 5.52149594e-01 -4.63182718e-01
-1.88536569e-01 -1.66233450e-01 -1.28757441e+00 6.78242207e-01
1.15100491e+00 -3.95721674e-01 -1.22933102e+00 2.66237020e-01
8.76004040e-01 -4.38434362e-01 -6.33449733e-01 4.20753330e-01
4.42471206e-01 -9.94059920e-01 1.21281779e+00 -7.40329802e-01
5.64881742e-01 -7.48817623e-01 -5.76036870e-02 -6.22035921e-01
-4.45159562e-02 -1.08356819e-01 -4.52911943e-01 9.04594481e-01
7.06325844e-02 -7.80752487e-03 7.54219413e-01 6.78014755e-01
1.97168402e-02 -6.71661079e-01 -8.53002906e-01 -6.44022822e-02
-6.51689708e-01 -4.01059806e-01 3.52429271e-01 7.79883862e-01
-7.31567293e-02 3.49444747e-01 -2.97951549e-01 2.79681653e-01
2.49316260e-01 3.40665251e-01 5.07438660e-01 -8.45850885e-01
-2.49521866e-01 -1.58059448e-01 -3.41778606e-01 -1.02835715e+00
9.17798281e-03 -4.84729588e-01 -1.18614577e-01 -2.33242607e+00
3.02603692e-01 1.28582090e-01 -2.42244437e-01 5.43110371e-01
-4.37160045e-01 1.44709185e-01 5.28748572e-01 -4.05782908e-02
-1.23085034e+00 2.98642427e-01 1.72035706e+00 -8.04208778e-03
-5.97041659e-02 -2.61869431e-01 -3.90361339e-01 7.85754979e-01
7.52220631e-01 -3.68612438e-01 -7.31575787e-01 -4.67332751e-01
3.39124531e-01 7.24094868e-01 3.85174572e-01 -9.10813272e-01
4.23236728e-01 -4.29312527e-01 2.91726410e-01 -6.99046195e-01
5.57503819e-01 -1.11200750e+00 1.32292449e-01 4.30317760e-01
-3.56914759e-01 3.54815364e-01 7.02063739e-02 6.34765863e-01
-5.18034041e-01 -2.89951205e-01 5.14535308e-01 -3.89962226e-01
-1.23424506e+00 2.95240998e-01 -4.79910582e-01 1.48909420e-01
1.43532562e+00 -1.60862997e-01 -1.91791937e-01 -5.80044627e-01
-8.61671150e-01 4.49707419e-01 1.76050112e-01 6.46859944e-01
9.32143807e-01 -1.05760717e+00 -4.95948374e-01 -2.54320741e-01
2.25984782e-01 -2.15471730e-01 6.92917347e-01 6.96921229e-01
-8.24292898e-01 3.34298700e-01 -2.66357392e-01 -3.65878701e-01
-1.56628644e+00 7.67154515e-01 2.38982141e-01 -2.99527049e-01
-6.78222060e-01 7.28690743e-01 3.66460562e-01 1.16315156e-01
1.95332751e-01 -2.66108513e-01 -7.57059813e-01 1.85888991e-01
7.50743091e-01 3.05345476e-01 -2.70047098e-01 -8.37979436e-01
-5.90181828e-01 4.11241293e-01 1.27494663e-01 -1.58342242e-01
1.03114259e+00 -3.43642771e-01 -1.12892598e-01 3.77742529e-01
8.29897285e-01 -2.23989829e-01 -1.34733796e+00 6.49057608e-03
-9.39661860e-02 -5.43108225e-01 -7.60185570e-02 -8.21960449e-01
-1.11107266e+00 7.77208567e-01 4.94666964e-01 1.53339431e-01
1.30053222e+00 3.29538763e-01 6.83561563e-01 2.00305447e-01
4.16134685e-01 -9.79449153e-01 4.96886224e-01 2.88303137e-01
9.67441797e-01 -1.15986443e+00 7.03579411e-02 -5.20769298e-01
-1.01913118e+00 9.65424657e-01 8.48138630e-01 -6.65464299e-03
5.82205594e-01 -2.23557666e-01 1.77196890e-01 -5.65864921e-01
-9.96331155e-01 -3.50860357e-01 3.55999291e-01 2.99632221e-01
2.87924528e-01 -2.31222197e-01 -5.46949983e-01 9.46579948e-02
3.99742901e-01 1.01033613e-01 5.01190603e-01 1.19309247e+00
-4.82725650e-01 -9.65186894e-01 -4.81070012e-01 2.56623685e-01
-6.25846624e-01 1.27927214e-01 -2.01855376e-01 7.41351604e-01
7.95422643e-02 1.19685793e+00 9.28521752e-02 2.59621441e-01
4.54184949e-01 3.24058443e-01 6.44016445e-01 -4.32494611e-01
-6.66131496e-01 -1.85738847e-01 3.16870123e-01 -9.88488853e-01
-1.15182602e+00 -7.63588965e-01 -1.70061100e+00 1.12714767e-01
3.41732621e-01 5.36995120e-02 4.00102228e-01 1.11175513e+00
1.94108129e-01 6.47237837e-01 3.62007581e-02 -2.87919641e-01
2.75934130e-01 -4.51330721e-01 2.97706630e-02 6.44069135e-01
2.34314114e-01 -7.42294610e-01 -4.27561626e-03 4.11547154e-01] | [10.43026351928711, 1.1045786142349243] |
24c92771-da3f-4a2f-88f9-75f88b0310c6 | dsfer-net-a-deep-supervision-and-feature | 2304.01101 | null | https://arxiv.org/abs/2304.01101v1 | https://arxiv.org/pdf/2304.01101v1.pdf | Dsfer-Net: A Deep Supervision and Feature Retrieval Network for Bitemporal Change Detection Using Modern Hopfield Networks | Change detection, as an important application for high-resolution remote sensing images, aims to monitor and analyze changes in the land surface over time. With the rapid growth in the quantity of high-resolution remote sensing data and the complexity of texture features, a number of quantitative deep learning-based methods have been proposed. Although these methods outperform traditional change detection methods by extracting deep features and combining spatial-temporal information, reasonable explanations about how deep features work on improving the detection performance are still lacking. In our investigations, we find that modern Hopfield network layers achieve considerable performance in semantic understandings. In this paper, we propose a Deep Supervision and FEature Retrieval network (Dsfer-Net) for bitemporal change detection. Specifically, the highly representative deep features of bitemporal images are jointly extracted through a fully convolutional Siamese network. Based on the sequential geo-information of the bitemporal images, we then design a feature retrieval module to retrieve the difference feature and leverage discriminative information in a deeply supervised manner. We also note that the deeply supervised feature retrieval module gives explainable proofs about the semantic understandings of the proposed network in its deep layers. Finally, this end-to-end network achieves a novel framework by aggregating the retrieved features and feature pairs from different layers. Experiments conducted on three public datasets (LEVIR-CD, WHU-CD, and CDD) confirm the superiority of the proposed Dsfer-Net over other state-of-the-art methods. Code will be available online (https://github.com/ShizhenChang/Dsfer-Net). | ['Pedram Ghamisi', 'Michael Kopp', 'Shizhen Chang'] | 2023-04-03 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [-4.42243144e-02 -4.93876129e-01 -5.86011000e-02 -7.44436741e-01
-7.03783810e-01 -3.04896981e-01 8.91194284e-01 -1.41345650e-01
-4.47295874e-01 2.85064310e-01 2.45524645e-01 -2.31270306e-02
-4.32825118e-01 -1.12641740e+00 -6.12070322e-01 -7.94190764e-01
-3.98873180e-01 7.81707242e-02 2.43817382e-02 -4.98088211e-01
8.28611031e-02 6.47753775e-01 -1.59789026e+00 1.78513199e-01
7.00761676e-01 1.23452175e+00 3.17502201e-01 3.47725064e-01
2.31926456e-01 4.69975084e-01 1.91945761e-01 -2.10702464e-01
3.72143000e-01 -9.14048925e-02 -7.49569416e-01 -1.99586764e-01
6.29462123e-01 -9.40343261e-01 -6.79336846e-01 1.29316270e+00
5.71890652e-01 -1.18703373e-01 5.55393577e-01 -9.83453870e-01
-7.97300398e-01 5.19139349e-01 -8.55310380e-01 6.10465407e-01
-3.64581645e-02 1.75832197e-01 1.28650129e+00 -1.06212759e+00
6.71791732e-01 1.46446657e+00 6.70382738e-01 7.02306554e-02
-8.53964329e-01 -7.39715099e-01 2.61809051e-01 4.77586180e-01
-1.45340383e+00 -4.22404766e-01 8.33673775e-01 -4.52658325e-01
8.70165169e-01 5.97471781e-02 7.29464293e-01 9.76545334e-01
2.83449262e-01 1.27983952e+00 8.89990568e-01 3.26402970e-02
1.05042286e-01 -4.16303903e-01 1.11897096e-01 9.46666360e-01
3.05374950e-01 3.58336270e-01 -3.57788444e-01 5.13190813e-02
7.01680064e-01 7.43684947e-01 -1.39892325e-01 -2.08025351e-01
-1.18364251e+00 9.47670102e-01 1.14691436e+00 5.02177596e-01
-5.90027213e-01 5.09595513e-01 2.60148257e-01 3.02202791e-01
8.73338819e-01 9.79127586e-02 -4.60283458e-01 2.49943390e-01
-1.13086712e+00 2.94473112e-01 3.61859649e-01 4.39558595e-01
1.04979670e+00 -3.46494585e-01 3.08851600e-02 6.76478386e-01
6.33404553e-01 1.01194036e+00 3.47565502e-01 -8.37194026e-01
3.92696738e-01 6.35786235e-01 -5.47020603e-03 -1.35499930e+00
-5.57000697e-01 -3.61480743e-01 -1.11004996e+00 1.91771388e-01
-3.82997352e-03 8.02870095e-02 -1.09338999e+00 1.58404863e+00
4.40776438e-01 -9.42656845e-02 -1.06382176e-01 1.15223169e+00
6.28242493e-01 6.70334339e-01 -1.09266639e-01 3.14567417e-01
1.22400904e+00 -8.05883408e-01 -5.75035453e-01 -9.25807506e-02
3.71078998e-01 -3.40066552e-01 8.73461068e-01 5.73273096e-03
-5.04207671e-01 -3.57697457e-01 -1.06057155e+00 -9.56451148e-02
-6.55131757e-01 2.64651984e-01 8.69960964e-01 -3.12561728e-02
-1.01890504e+00 8.11466515e-01 -1.28893304e+00 -5.73045373e-01
1.04686105e+00 9.27393809e-02 -3.79486889e-01 -2.83176422e-01
-1.37434256e+00 7.05753148e-01 -2.22120304e-02 8.42120409e-01
-1.21399653e+00 -4.90159333e-01 -9.47129071e-01 2.44944349e-01
5.73984999e-03 -5.26608288e-01 1.03364158e+00 -9.72959101e-01
-1.30737102e+00 7.69113541e-01 4.74389642e-02 -2.85521626e-01
6.40823424e-01 -4.83541936e-01 -3.08538020e-01 5.24711549e-01
3.84486675e-01 8.38554621e-01 8.15558374e-01 -9.16565716e-01
-7.75296450e-01 -7.80742705e-01 1.20677933e-01 9.67864618e-02
-1.51855096e-01 -1.22987449e-01 -1.73582911e-01 -6.36539400e-01
4.30319220e-01 -7.63344049e-01 -2.06930995e-01 5.09865820e-01
-1.98673964e-01 -2.66897112e-01 8.34978461e-01 -7.22487211e-01
8.67889941e-01 -2.29236364e+00 -5.38563095e-02 1.77608192e-01
3.76086116e-01 2.94950843e-01 -4.50602949e-01 3.93015504e-01
3.22687067e-02 1.78219751e-01 -6.64480746e-01 5.49523393e-03
1.92001179e-01 9.24735144e-02 -4.58534718e-01 8.46959114e-01
5.06529093e-01 1.14127815e+00 -1.16011429e+00 -2.15277970e-01
3.20418894e-01 6.08747661e-01 -4.75448936e-01 -2.13786270e-02
-1.79981902e-01 1.06878698e-01 -8.18021894e-01 1.00012314e+00
9.25343752e-01 -2.04649612e-01 -2.03836650e-01 -2.50938654e-01
-2.28534192e-01 1.57977343e-01 -8.24662507e-01 1.79492593e+00
-3.24672788e-01 7.38160789e-01 7.94406515e-03 -1.03380466e+00
6.93032622e-01 -5.81400469e-02 4.76080716e-01 -1.10863459e+00
-2.98577026e-02 3.61698657e-01 -2.03354537e-01 -7.63260841e-01
3.23105156e-01 -9.26736221e-02 7.81224072e-02 3.70955735e-01
-1.03787102e-01 -3.92799787e-02 -5.51583879e-02 1.00830570e-01
1.11042583e+00 1.63586468e-01 1.38375968e-01 -3.20619792e-01
2.56060779e-01 -1.80752110e-02 3.71577293e-01 6.35798395e-01
-4.46674466e-01 4.53109503e-01 1.89211845e-01 -8.12178195e-01
-7.20759392e-01 -9.34666634e-01 -4.39647794e-01 9.95264292e-01
2.76081264e-01 -6.01389706e-02 -4.25915629e-01 -5.39973855e-01
2.59736121e-01 1.96667433e-01 -1.02926183e+00 -1.58602029e-01
-4.28315431e-01 -8.77561450e-01 7.21161187e-01 4.35212106e-01
1.19287515e+00 -9.87740219e-01 -6.88033760e-01 2.32191131e-01
-9.63935554e-02 -6.42690837e-01 -8.37078765e-02 -1.04603928e-03
-8.91907930e-01 -1.04721653e+00 -6.35603666e-01 -5.87927043e-01
4.08692211e-01 6.50402129e-01 6.88310325e-01 4.81270961e-02
-4.71452951e-01 7.48039484e-02 -3.90584469e-01 -1.03688046e-01
1.56969279e-01 2.22746372e-01 -1.40245125e-01 6.07304573e-02
6.28771961e-01 -7.14873433e-01 -1.06478202e+00 1.09592319e-01
-1.26576030e+00 -1.34694412e-01 8.29393744e-01 8.51105213e-01
5.42429984e-01 4.30682823e-02 3.65266025e-01 -5.25242090e-01
2.48558536e-01 -6.98891103e-01 -5.81642807e-01 1.61447585e-01
-6.26643956e-01 2.22803891e-01 1.34621069e-01 1.33918092e-01
-1.06191266e+00 -4.81163859e-02 -2.83924818e-01 -1.55295342e-01
-1.59918591e-01 9.29445386e-01 -1.34114206e-01 1.94830701e-01
4.98629868e-01 1.98658347e-01 -4.91084307e-02 -7.24216878e-01
4.85731781e-01 8.04136336e-01 4.10593063e-01 -1.64997771e-01
8.03436756e-01 1.10396767e+00 -3.64803940e-01 -6.50047362e-01
-1.11259329e+00 -5.59405744e-01 -6.83060527e-01 6.27315938e-02
7.58498430e-01 -1.24994516e+00 -4.26341683e-01 1.05570471e+00
-9.54639196e-01 -3.49886775e-01 3.75048779e-02 5.14459014e-01
-3.25936645e-01 2.04216301e-01 -6.79019332e-01 -3.90441477e-01
-6.19407058e-01 -9.04718637e-01 1.51887178e+00 1.30973637e-01
2.45416388e-01 -9.30686653e-01 3.62022787e-01 1.40842274e-01
7.66337335e-01 3.65385741e-01 6.84883237e-01 -2.49291420e-01
-6.72928810e-01 -1.29119113e-01 -5.90187371e-01 1.96074083e-01
5.06415069e-01 2.13760156e-02 -1.03789878e+00 -3.04377764e-01
-1.12238988e-01 -2.75763065e-01 1.74191368e+00 5.52084446e-01
1.08092630e+00 -5.12520075e-01 -4.02151346e-01 8.60373199e-01
1.57804179e+00 -1.95034727e-01 6.26047373e-01 5.74200273e-01
7.63176143e-01 4.47585046e-01 5.35606146e-01 4.60204512e-01
6.36254072e-01 5.18014133e-01 7.70963311e-01 -3.20651591e-01
-1.12345286e-01 -1.91819802e-01 6.36867225e-01 5.05286336e-01
-4.45839055e-02 7.04222172e-02 -9.11766291e-01 8.27240348e-01
-1.89727807e+00 -1.36112928e+00 5.40740341e-02 1.65889823e+00
8.78108382e-01 -3.08672577e-01 -3.25292617e-01 -3.95910814e-02
6.16628647e-01 6.95463419e-01 -9.71356809e-01 2.10237697e-01
-2.56106555e-01 1.89020991e-01 4.94458199e-01 5.79860091e-01
-1.55167639e+00 1.24816144e+00 5.30971050e+00 6.63078547e-01
-1.45680499e+00 1.46032557e-01 3.48038852e-01 -6.70491606e-02
-3.76441807e-01 -1.88916460e-01 -6.41320288e-01 2.17559904e-01
5.78441203e-01 2.14344636e-01 3.82964820e-01 6.81442082e-01
3.56382787e-01 -1.06212221e-01 -9.44747031e-01 7.55688608e-01
-3.64589766e-02 -1.32188678e+00 2.70576686e-01 -9.15991887e-02
7.08297610e-01 9.28409755e-01 2.23576188e-01 8.23725909e-02
4.47782934e-01 -9.53443050e-01 7.70985901e-01 6.45279348e-01
6.50936425e-01 -4.23542023e-01 7.83942640e-01 -5.05494997e-02
-1.20981014e+00 -9.03447419e-02 -5.53390324e-01 -1.40637219e-01
-1.33170575e-01 9.78372455e-01 -4.08100456e-01 5.06056190e-01
1.08808792e+00 1.56154764e+00 -7.13637412e-01 7.65556574e-01
-4.50398386e-01 7.01004386e-01 -5.12349606e-01 2.64415175e-01
6.38639271e-01 -1.00738458e-01 4.17728990e-01 1.26792908e+00
2.59594798e-01 1.18084922e-01 -2.12350994e-01 1.12920809e+00
-8.77674446e-02 -3.03781122e-01 -6.78542793e-01 -1.07198566e-01
2.36543283e-01 1.42892671e+00 -4.22423482e-01 -2.95205832e-01
-2.57675141e-01 9.84704137e-01 3.19856852e-01 4.17066067e-01
-5.93319416e-01 -5.84313452e-01 1.06804359e+00 -9.97280478e-02
6.21248841e-01 -2.62975276e-01 2.82236189e-03 -1.43931007e+00
9.82339159e-02 -6.82606339e-01 3.02185327e-01 -6.41080439e-01
-1.38476801e+00 3.92596662e-01 -1.12317927e-01 -1.07332301e+00
4.83399294e-02 -5.17959893e-01 -4.17408615e-01 6.05884433e-01
-2.30861878e+00 -1.28036773e+00 -5.61411560e-01 7.27635384e-01
4.26665932e-01 8.16647038e-02 5.00326157e-01 4.23194498e-01
-5.83328485e-01 2.07526132e-01 4.65146720e-01 4.92753804e-01
4.34598774e-01 -9.37752008e-01 6.09503329e-01 9.48157251e-01
1.37236759e-01 5.38183808e-01 2.60464251e-01 -4.86563951e-01
-1.50626349e+00 -1.47001159e+00 6.38306320e-01 -9.64529216e-02
9.01551783e-01 -1.42673865e-01 -9.94714200e-01 5.84839761e-01
8.83878674e-03 2.32581854e-01 3.77890676e-01 -1.84463963e-01
-6.73371315e-01 -5.11792958e-01 -1.01973152e+00 2.60831416e-01
1.24774325e+00 -7.11707592e-01 -7.01907754e-01 2.24620357e-01
4.96685028e-01 -4.34353054e-02 -6.12089396e-01 5.04920125e-01
7.21426129e-01 -9.03449476e-01 7.87311256e-01 -5.45965731e-01
8.89962733e-01 -4.66469109e-01 -4.80087608e-01 -1.27157903e+00
-5.15638292e-01 5.85527625e-03 1.57819971e-01 7.66773999e-01
1.37595624e-01 -7.02496350e-01 2.56663084e-01 5.53753115e-02
1.03136338e-02 -6.16794109e-01 -8.70619476e-01 -6.19178116e-01
2.84273058e-01 -2.49127209e-01 6.84207559e-01 1.09027588e+00
-2.68086463e-01 2.03008667e-01 -1.87914088e-01 5.39681971e-01
6.24192894e-01 5.75233281e-01 3.46782714e-01 -1.36822641e+00
2.52739992e-02 -5.27339280e-01 -6.04543328e-01 -1.02021515e+00
1.65423408e-01 -1.15255928e+00 2.96636634e-02 -1.75951099e+00
5.37250578e-01 -4.45516884e-01 -4.26859438e-01 1.05716634e+00
-1.75065622e-01 3.20807725e-01 -1.07432023e-01 5.33920646e-01
-4.76966590e-01 1.01347172e+00 1.09428144e+00 -6.79388463e-01
-4.61226813e-02 -4.12570119e-01 -6.00838065e-01 7.88141727e-01
8.03417921e-01 -6.42647445e-01 6.69493824e-02 -9.92076278e-01
5.32188892e-01 -3.81057799e-01 8.42628062e-01 -7.36931086e-01
8.34173858e-02 -2.00954095e-01 3.95335436e-01 -6.59415066e-01
-1.29844591e-01 -8.04821730e-01 -7.09740072e-02 7.56357551e-01
-3.59694928e-01 -2.67667085e-01 -2.15820060e-03 6.33895636e-01
-5.07570386e-01 2.35372007e-01 6.51617765e-01 -1.20232724e-01
-9.17109966e-01 7.72885501e-01 -3.29096466e-01 -2.37532869e-01
5.75065613e-01 1.44202888e-01 -5.84504068e-01 -1.89986899e-01
-4.45532054e-01 4.43465203e-01 2.65440941e-01 6.13975406e-01
6.63580179e-01 -1.39584196e+00 -9.89793062e-01 1.85644090e-01
3.18994820e-01 1.40113264e-01 2.28088886e-01 9.30127800e-01
-7.59509563e-01 3.33990574e-01 -3.07866126e-01 -6.58368289e-01
-6.82089984e-01 1.48437113e-01 6.52840018e-01 -1.21623911e-01
-8.07707608e-01 6.51244640e-01 3.12275857e-01 -4.70381111e-01
-2.01426372e-01 -6.37362480e-01 -1.24126785e-01 3.24265540e-01
6.95800841e-01 2.69373357e-01 1.11874036e-01 -5.35193026e-01
-6.36858642e-01 7.73545086e-01 -1.29620075e-01 -2.13703625e-02
2.13558316e+00 -2.72940725e-01 -3.00701261e-01 1.69762656e-01
1.58668435e+00 -3.81966233e-01 -1.49753773e+00 -6.46811008e-01
3.26585174e-02 -4.29282933e-01 4.98416901e-01 -7.39401102e-01
-1.61313832e+00 1.22210729e+00 9.69618499e-01 -1.72187030e-01
1.09188664e+00 1.77066684e-01 8.12548697e-01 8.46246660e-01
2.81092227e-01 -9.55361545e-01 -4.10100296e-02 6.12886012e-01
1.16583741e+00 -1.69344211e+00 1.22969598e-02 8.79448131e-02
-2.35207021e-01 1.02846670e+00 2.63760120e-01 -2.87212342e-01
1.05008304e+00 -8.50422382e-02 1.82939231e-01 -7.90054917e-01
-4.83130395e-01 -5.22195041e-01 5.05603254e-02 2.34525397e-01
1.51004434e-01 2.58453012e-01 -1.70909688e-01 3.38744313e-01
-6.46099001e-02 1.13532528e-01 1.12850763e-01 8.43319952e-01
-6.81482136e-01 -6.25267148e-01 -1.12304036e-02 3.77802432e-01
-2.87280083e-01 -4.01928365e-01 -3.10126632e-01 6.56045854e-01
4.90322150e-03 7.77195334e-01 1.18318573e-01 -2.45427921e-01
3.93586121e-02 -3.57738495e-01 1.61537901e-01 -2.89270490e-01
-2.96533614e-01 -2.00141236e-01 -3.22231293e-01 -9.37473297e-01
-8.78018737e-01 -7.43342698e-01 -1.29163909e+00 -2.51576006e-01
-1.54261932e-01 -1.97034836e-01 6.25311494e-01 1.05324996e+00
5.44233620e-01 1.39670700e-01 9.16085780e-01 -9.46743488e-01
-7.19972610e-01 -1.14718580e+00 -7.20585465e-01 2.57978588e-01
7.58715510e-01 -7.44996071e-01 -4.83784914e-01 -4.48407605e-02] | [9.656076431274414, -1.2613346576690674] |
20aff5c0-5bda-4b13-9eb6-dee50ed40250 | fcdsn-dc-an-accurate-and-lightweight | 2209.06525 | null | https://arxiv.org/abs/2209.06525v1 | https://arxiv.org/pdf/2209.06525v1.pdf | FCDSN-DC: An Accurate and Lightweight Convolutional Neural Network for Stereo Estimation with Depth Completion | We propose an accurate and lightweight convolutional neural network for stereo estimation with depth completion. We name this method fully-convolutional deformable similarity network with depth completion (FCDSN-DC). This method extends FC-DCNN by improving the feature extractor, adding a network structure for training highly accurate similarity functions and a network structure for filling inconsistent disparity estimates. The whole method consists of three parts. The first part consists of fully-convolutional densely connected layers that computes expressive features of rectified image pairs. The second part of our network learns highly accurate similarity functions between this learned features. It consists of densely-connected convolution layers with a deformable convolution block at the end to further improve the accuracy of the results. After this step an initial disparity map is created and the left-right consistency check is performed in order to remove inconsistent points. The last part of the network then uses this input together with the corresponding left RGB image in order to train a network that fills in the missing measurements. Consistent depth estimations are gathered around invalid points and are parsed together with the RGB points into a shallow CNN network structure in order to recover the missing values. We evaluate our method on challenging real world indoor and outdoor scenes, in particular Middlebury, KITTI and ETH3D were it produces competitive results. We furthermore show that this method generalizes well and is well suited for many applications without the need of further training. The code of our full framework is available at: https://github.com/thedodo/FCDSN-DC | ['Friedrich Fraundorfer', 'Dominik Hirner'] | 2022-09-14 | null | null | null | null | ['depth-completion'] | ['computer-vision'] | [ 9.85026509e-02 1.10085629e-01 3.89737219e-01 -6.52501225e-01
-6.52006090e-01 -4.12802935e-01 3.50546688e-01 -2.00420916e-01
-7.72892356e-01 7.35213816e-01 1.04275189e-01 8.37229490e-02
1.05620779e-01 -8.70983720e-01 -1.10674775e+00 -6.80404365e-01
1.57723621e-01 5.44627547e-01 5.93353450e-01 -3.17238837e-01
2.22561777e-01 5.97406089e-01 -1.79903650e+00 3.32139283e-01
7.31729329e-01 1.33951664e+00 1.14219300e-01 7.04399168e-01
9.75366980e-02 7.23894656e-01 -1.52097106e-01 -2.18349919e-01
7.21325517e-01 1.96325835e-02 -8.59262347e-01 1.37107596e-02
9.13828850e-01 -9.29690599e-01 -3.78671736e-01 6.69842839e-01
5.97634494e-01 2.80104250e-01 1.01967365e-01 -1.06174052e+00
-4.47110720e-02 1.02889076e-01 -3.02760184e-01 -1.34355530e-01
3.51099491e-01 2.31548965e-01 6.75521255e-01 -1.07091892e+00
7.59531319e-01 1.05613577e+00 1.12926280e+00 5.60207665e-01
-1.11717772e+00 -4.63442266e-01 -1.12870924e-01 3.62617187e-02
-1.19426835e+00 -3.94027591e-01 5.65795302e-01 -3.54008704e-01
1.14338851e+00 6.35545179e-02 9.07647610e-01 8.56217325e-01
-2.01229796e-01 5.17700613e-01 9.12893891e-01 -1.83740631e-01
3.04306507e-01 -2.79682755e-01 4.85919826e-02 7.25544155e-01
6.88704997e-02 2.37324759e-01 -2.91190118e-01 1.71089038e-01
1.03329384e+00 3.21787894e-01 -3.48025471e-01 -6.49464190e-01
-1.07098365e+00 6.29902899e-01 1.17371678e+00 2.22653195e-01
-3.37209851e-01 5.08993447e-01 2.32410267e-01 2.33286008e-01
5.05799711e-01 -2.43165307e-02 -6.55563653e-01 -4.69284654e-02
-1.06541944e+00 4.19507354e-01 6.70813501e-01 7.41337895e-01
1.37230968e+00 -3.18441540e-01 2.00667843e-01 7.68889070e-01
3.72303464e-02 3.98904800e-01 2.43098259e-01 -1.48633170e+00
6.48132920e-01 6.98320031e-01 1.41023412e-01 -6.63260400e-01
-5.80442786e-01 -2.41503239e-01 -7.64789283e-01 7.06748247e-01
4.22757834e-01 -2.15959266e-01 -1.13837314e+00 1.49612617e+00
4.19603080e-01 2.80017972e-01 2.82407477e-02 1.08411920e+00
9.42040980e-01 3.08326095e-01 -5.13034105e-01 5.94204605e-01
9.58302617e-01 -9.67022538e-01 -1.32677451e-01 -3.01835477e-01
5.01827002e-01 -5.47799051e-01 7.85592854e-01 2.85475582e-01
-1.28731906e+00 -7.39451289e-01 -1.09038281e+00 -7.58497775e-01
-4.40756619e-01 7.39381984e-02 6.92220271e-01 1.15342982e-01
-1.58747685e+00 1.09292471e+00 -1.02267849e+00 -3.30785602e-01
6.74024522e-01 7.73675740e-01 -7.11595654e-01 -4.64862794e-01
-8.24423909e-01 6.43018305e-01 3.88405055e-01 3.79098624e-01
-6.85998142e-01 -6.51528358e-01 -1.36587179e+00 -7.88139328e-02
5.73958121e-02 -1.02932632e+00 1.11940575e+00 -1.13632965e+00
-1.49274886e+00 1.03516686e+00 -3.25121805e-02 -4.77343380e-01
8.17338109e-01 -3.44503671e-01 4.57931370e-01 1.74276065e-02
1.72492951e-01 1.19620669e+00 6.73750997e-01 -1.12958539e+00
-7.32225180e-01 -5.98927855e-01 2.04833016e-01 1.09697089e-01
1.43181428e-01 -5.52820027e-01 -7.87851393e-01 -3.99095654e-01
4.62348700e-01 -9.51167881e-01 -2.85063863e-01 2.44949043e-01
-2.80541718e-01 7.64451325e-02 5.35485625e-01 -6.08163238e-01
5.88623822e-01 -2.06608677e+00 2.59699464e-01 3.30957323e-01
2.83714086e-01 1.92642912e-01 -2.24297523e-01 2.28957623e-01
-9.16057080e-02 -4.72595692e-01 -5.15579402e-01 -1.08561707e+00
-1.80268750e-01 4.44443136e-01 7.02883955e-03 5.87652028e-01
3.07396173e-01 8.90113413e-01 -7.69507051e-01 2.37990487e-02
6.83554471e-01 8.72662783e-01 -8.00568402e-01 1.80465922e-01
-1.53989762e-01 5.82266450e-01 -2.41888911e-02 4.17040735e-01
1.21735513e+00 -1.19657833e-02 -2.67746747e-01 -1.69923216e-01
-3.82158905e-01 3.55170906e-01 -1.45425653e+00 2.25847650e+00
-5.63250721e-01 7.23347604e-01 8.91296417e-02 -7.82265723e-01
7.83180594e-01 -1.40215814e-01 4.31107581e-01 -8.45808864e-01
2.19083160e-01 5.03907144e-01 -4.46135610e-01 -4.58992481e-01
5.49006641e-01 1.65376216e-01 2.06975803e-01 1.44270390e-01
2.00470880e-01 -4.85651881e-01 5.21991588e-02 6.70873299e-02
1.14514303e+00 5.86230934e-01 -2.18339097e-02 -4.26135547e-02
8.21240664e-01 -1.28659144e-01 6.03402913e-01 2.65827566e-01
7.20850229e-02 1.10500240e+00 2.32420564e-01 -9.12446260e-01
-1.17278206e+00 -1.00495613e+00 -1.24318384e-01 5.86018085e-01
2.10646510e-01 -1.60228342e-01 -8.10862422e-01 -4.95163649e-01
2.29614615e-01 9.98279676e-02 -8.80308509e-01 1.41862497e-01
-7.36065984e-01 -7.71834850e-02 2.37154484e-01 9.32742834e-01
9.96732295e-01 -1.09309936e+00 -7.81024694e-01 6.02867231e-02
-2.10848525e-01 -1.21366835e+00 -2.15876535e-01 6.40448034e-01
-1.00556803e+00 -1.19289911e+00 -7.50014961e-01 -8.37525845e-01
7.65025318e-01 1.62065715e-01 1.07259035e+00 2.90612251e-01
-2.10515082e-01 1.63356870e-01 -5.86345568e-02 -1.30106434e-01
-3.18030939e-02 1.88339666e-01 -3.38074714e-01 -1.58533260e-01
2.11306527e-01 -7.63320267e-01 -9.09386098e-01 2.12627828e-01
-9.40462530e-01 6.14200830e-02 3.45787436e-01 6.99580491e-01
7.75297046e-01 -5.28858364e-01 -2.32253343e-01 -6.37311995e-01
-1.30359441e-01 -9.30137336e-02 -1.07588923e+00 -3.80225599e-01
-3.38840969e-02 3.41203719e-01 4.89155620e-01 8.27997774e-02
-7.38695860e-01 5.90237975e-01 -6.82912290e-01 -3.85534078e-01
-2.72922337e-01 -1.15022719e-01 1.18219154e-02 -2.99664259e-01
6.68014824e-01 -5.12256324e-02 -2.17310656e-02 -5.14176667e-01
2.12366655e-01 4.58756417e-01 8.62520456e-01 -2.02982664e-01
9.08026040e-01 9.45012987e-01 1.73779368e-01 -5.18144727e-01
-8.63703072e-01 -6.69315636e-01 -1.04932368e+00 -3.10504232e-02
9.25818384e-01 -1.03223813e+00 -7.32880235e-01 8.17725897e-01
-1.20904315e+00 -9.53287959e-01 -6.02212906e-01 4.96426284e-01
-8.98546934e-01 1.56922862e-01 -6.64659798e-01 -3.20945024e-01
-3.69722784e-01 -1.10619164e+00 1.48530591e+00 1.72797650e-01
1.66191593e-01 -1.02673781e+00 3.64229620e-01 3.50051761e-01
3.60913277e-01 4.74295855e-01 2.46070325e-01 -1.57205909e-01
-8.23238313e-01 -3.17243785e-02 -3.52133602e-01 8.29017580e-01
-8.61702189e-02 -1.79243818e-01 -1.32658148e+00 -1.61521330e-01
-9.57288966e-02 -5.31509280e-01 1.34814501e+00 5.82162976e-01
1.09675860e+00 1.14555091e-01 -9.71776769e-02 1.30836880e+00
1.72151208e+00 -4.34772044e-01 1.14537299e+00 6.20634556e-01
7.38581359e-01 5.82481802e-01 4.51211542e-01 3.17227364e-01
7.84841955e-01 6.73411250e-01 9.33027327e-01 -4.65886056e-01
-3.52432579e-01 4.02074270e-02 1.98383838e-01 2.95966774e-01
-4.86831754e-01 1.55703798e-01 -7.30024099e-01 4.66655463e-01
-1.88755035e+00 -8.92009556e-01 -2.05086485e-01 2.33597636e+00
5.80569208e-01 6.76020756e-02 1.17026187e-01 4.44421858e-01
3.07610810e-01 4.68987564e-04 -4.58566159e-01 -4.10492331e-01
-1.17905974e-01 6.71464920e-01 5.67529082e-01 8.61320376e-01
-1.09458792e+00 9.61951196e-01 4.37084866e+00 3.12164217e-01
-1.08930480e+00 -8.63966122e-02 5.05735159e-01 1.53606827e-03
-1.21147618e-01 -4.79197577e-02 -6.39186382e-01 1.88616425e-01
4.47534293e-01 7.37160802e-01 4.43083942e-01 7.62906432e-01
1.03284314e-01 -5.16716838e-01 -1.10103750e+00 1.17373562e+00
-2.53979843e-02 -1.34424126e+00 -3.39285403e-01 9.40217003e-02
8.15472245e-01 6.71332181e-01 -2.21723348e-01 4.56906110e-02
2.34594166e-01 -8.49838912e-01 6.78118348e-01 5.07819474e-01
8.81870747e-01 -7.81741798e-01 1.00839567e+00 1.82612881e-01
-1.12812495e+00 -3.76008376e-02 -6.52137041e-01 -2.74691731e-01
-1.08066343e-01 7.75211096e-01 -4.54121083e-01 5.37014902e-01
9.78480041e-01 1.09779227e+00 -5.29824913e-01 1.16248739e+00
-3.97101700e-01 -1.65024757e-01 -5.50680399e-01 5.31518579e-01
3.55636269e-01 -1.24272667e-01 6.67907149e-02 9.75766122e-01
3.43361229e-01 -2.00173885e-01 -1.79375902e-01 7.82964408e-01
-2.84439683e-01 -3.33758056e-01 -5.37137270e-01 8.86042297e-01
6.64786622e-03 1.37961018e+00 -6.61606491e-01 -3.03827196e-01
-2.65894443e-01 1.53212380e+00 7.36887574e-01 1.81445301e-01
-7.48679101e-01 -3.89884084e-01 8.54628563e-01 1.59706429e-01
6.49315834e-01 -2.17409611e-01 -2.41755053e-01 -1.29383922e+00
3.33032936e-01 -4.28288162e-01 3.32364917e-01 -8.38254094e-01
-9.64593947e-01 6.91561818e-01 -4.38611865e-01 -1.19434643e+00
-3.70166123e-01 -5.71473897e-01 -4.56661433e-01 9.15279388e-01
-1.99427664e+00 -1.10020792e+00 -1.26099443e+00 9.92187321e-01
4.50423211e-01 4.92810667e-01 7.27320552e-01 3.47350746e-01
-2.65338451e-01 2.72818387e-01 -8.32033306e-02 2.52391726e-01
6.55340612e-01 -1.43774414e+00 7.51645505e-01 7.20190167e-01
-2.54453003e-01 2.37820432e-01 2.26482809e-01 -4.16387677e-01
-8.87604415e-01 -1.10890889e+00 9.43898082e-01 -3.47623080e-01
1.46476910e-01 -5.31700015e-01 -7.39744604e-01 7.07204998e-01
-3.80720831e-02 7.55167067e-01 5.66126332e-02 -4.41264063e-01
-3.96117568e-01 -3.36090863e-01 -1.30730307e+00 1.06789462e-01
1.37157094e+00 -3.57251346e-01 -3.19474131e-01 2.57352054e-01
5.96695662e-01 -9.47895706e-01 -6.84131324e-01 3.64823669e-01
5.72754383e-01 -1.71895564e+00 1.06652677e+00 2.28853747e-02
6.03955388e-01 -3.54544342e-01 -2.25825533e-01 -1.07765961e+00
4.05784920e-02 -3.73840004e-01 1.94370762e-01 7.81454206e-01
1.08952768e-01 -7.70946741e-01 1.00379407e+00 5.08809984e-01
-4.00767386e-01 -7.53030121e-01 -9.48365152e-01 -6.68188989e-01
-2.23470852e-02 -5.69856882e-01 5.94184637e-01 5.66033423e-01
-5.58140934e-01 -1.38991047e-02 -2.32474841e-02 7.71266893e-02
7.58809686e-01 7.21262172e-02 1.09425092e+00 -1.16676831e+00
-2.21561745e-01 -1.03032105e-01 -7.15354502e-01 -1.24398065e+00
-6.77657574e-02 -7.47049332e-01 1.09925799e-01 -1.79940081e+00
-1.63172051e-01 -6.35492444e-01 2.17841670e-01 5.84776580e-01
3.08334120e-02 7.62821436e-01 7.85580054e-02 7.22790584e-02
-4.71076518e-01 3.93206477e-01 1.10796714e+00 1.88576981e-01
-4.25059617e-01 1.76316991e-01 -7.95911327e-02 8.61234725e-01
7.59108484e-01 -3.35375071e-01 -1.83311790e-01 -6.57612622e-01
1.82255611e-01 -2.25694254e-01 9.00843322e-01 -1.61736143e+00
1.64844111e-01 3.63057971e-01 8.03761840e-01 -7.33251750e-01
6.04657650e-01 -1.05149496e+00 1.24080710e-01 6.32245600e-01
-2.44583506e-02 1.17019974e-02 1.85176492e-01 1.22970715e-01
-1.31452262e-01 -6.37753587e-03 9.02770996e-01 -2.41405249e-01
-7.28370607e-01 4.85596687e-01 3.29590857e-01 -4.29811366e-02
9.31961179e-01 -5.15725434e-01 -1.86695367e-01 -3.54634166e-01
-7.68843710e-01 2.20596239e-01 9.06078100e-01 1.38778180e-01
7.69444585e-01 -1.25124145e+00 -5.03694832e-01 5.59811950e-01
2.94191130e-02 8.17358315e-01 2.69486934e-01 8.88388097e-01
-1.08600092e+00 2.64247566e-01 -3.78692210e-01 -7.66396999e-01
-1.06051672e+00 3.18229735e-01 7.08515644e-01 -2.48680234e-01
-7.72895992e-01 8.62199426e-01 5.12204915e-02 -6.25373960e-01
4.20713693e-01 -9.76081729e-01 -3.01239896e-03 -1.51971236e-01
4.61299866e-01 4.33413267e-01 4.71017718e-01 -6.45891905e-01
-4.72910553e-01 9.98451591e-01 3.84150416e-01 -6.23824783e-02
1.85811985e+00 -1.15910403e-01 -2.06331968e-01 2.33824942e-02
1.60903370e+00 -9.83568728e-02 -1.80894113e+00 -1.67669162e-01
-3.05140078e-01 -5.27961612e-01 -1.59163568e-02 -5.04177213e-01
-1.39403152e+00 9.60118175e-01 6.09090984e-01 -3.00711989e-01
1.22284877e+00 -6.27987459e-02 9.59017217e-01 3.26121539e-01
3.96570742e-01 -9.63331878e-01 1.31695494e-01 7.27393746e-01
8.53228509e-01 -1.49981642e+00 -8.20363760e-02 -2.91299462e-01
-1.89892381e-01 1.35725534e+00 5.69039285e-01 -5.44013321e-01
4.93114203e-01 4.02638286e-01 1.71328768e-01 -2.50634849e-01
-2.23295838e-01 -6.39583826e-01 1.84570849e-01 7.83348441e-01
1.31928399e-01 -4.10751104e-01 1.42685592e-01 7.35617951e-02
-3.39072376e-01 3.01352620e-01 3.06487024e-01 9.33536768e-01
-3.29362839e-01 -1.04779458e+00 -2.53919721e-01 2.07556225e-02
-9.50853527e-02 -1.43579960e-01 -3.26768696e-01 7.90760398e-01
5.47659457e-01 5.75551152e-01 4.52078849e-01 -4.56088305e-01
5.93840659e-01 -2.05077738e-01 5.51408052e-01 -5.91836214e-01
-8.66191089e-01 -2.62723833e-01 -1.09419851e-02 -1.36744034e+00
-6.43612981e-01 -5.03633022e-01 -1.60052395e+00 -4.93692487e-01
-4.30089533e-02 -4.11656022e-01 8.37935328e-01 7.55921185e-01
4.05054480e-01 3.23058099e-01 5.60336769e-01 -1.49583006e+00
-2.65799910e-02 -6.78271651e-01 -2.00960129e-01 5.68737686e-01
8.09353769e-01 -4.29656982e-01 -3.73299509e-01 -6.87162429e-02] | [8.748614311218262, -2.4585893154144287] |
a5e2b5e7-aa28-4985-9e27-fbc19b44cc13 | graph-based-global-robot-localization | 2303.02076 | null | https://arxiv.org/abs/2303.02076v1 | https://arxiv.org/pdf/2303.02076v1.pdf | Graph-based Global Robot Localization Informing Situational Graphs with Architectural Graphs | In this paper, we propose a solution for legged robot localization using architectural plans. Our specific contributions towards this goal are several. Firstly, we develop a method for converting the plan of a building into what we denote as an architectural graph (A-Graph). When the robot starts moving in an environment, we assume it has no knowledge about it, and it estimates an online situational graph representation (S-Graph) of its surroundings. We develop a novel graph-to-graph matching method, in order to relate the S-Graph estimated online from the robot sensors and the A-Graph extracted from the building plans. Note the challenge in this, as the S-Graph may show a partial view of the full A-Graph, their nodes are heterogeneous and their reference frames are different. After the matching, both graphs are aligned and merged, resulting in what we denote as an informed Situational Graph (iS-Graph), with which we achieve global robot localization and exploitation of prior knowledge from the building plans. Our experiments show that our pipeline shows a higher robustness and a significantly lower pose error than several LiDAR localization baselines. | ['Holger Voos', 'Javier Civera', 'Jose Luis Sanchez-Lopez', 'Hriday Bavle', 'Jose Andres Millan-Romera', 'Muhammad Shaheer'] | 2023-03-03 | null | null | null | null | ['graph-matching'] | ['graphs'] | [ 8.20667371e-02 7.13983119e-01 2.25007713e-01 -3.74151617e-01
-7.34773457e-01 -6.18182421e-01 4.62784380e-01 3.60567272e-01
-8.82019401e-02 4.29780453e-01 -7.75294453e-02 1.70116469e-01
-6.16034381e-02 -1.09657443e+00 -9.04296219e-01 -3.33798319e-01
-1.90535948e-01 9.62761521e-01 7.98325658e-01 -2.95149833e-01
6.35402948e-02 6.54221833e-01 -1.29218173e+00 -2.08305404e-01
4.47547168e-01 7.91162848e-01 7.45261908e-01 6.17986500e-01
1.39924690e-01 5.62927485e-01 -4.17970955e-01 -7.86002260e-03
2.93388158e-01 -1.25468329e-01 -8.69407773e-01 2.85052001e-01
5.40067017e-01 -2.44097441e-01 -4.11887676e-01 1.04672372e+00
2.45256647e-01 -1.43812567e-01 2.72537112e-01 -1.48709667e+00
1.69197753e-01 3.86814415e-01 -4.26160723e-01 -5.33697963e-01
7.99937308e-01 -9.03166905e-02 9.80965734e-01 -7.60111034e-01
1.07631755e+00 1.19434655e+00 9.02055442e-01 -1.48454204e-01
-1.32023501e+00 -2.13121653e-01 3.14867049e-01 2.67152369e-01
-1.81009495e+00 -3.47931296e-01 8.51321518e-01 -4.54577744e-01
8.15756619e-01 -8.90532136e-02 8.96792948e-01 4.91691709e-01
3.42987716e-01 3.01952749e-01 8.18961561e-01 -6.07585013e-01
5.09083927e-01 -1.90512016e-01 7.88457319e-02 1.08753717e+00
3.78201008e-01 -1.31085411e-01 -6.11823618e-01 1.29355090e-02
7.18665838e-01 -1.23133264e-01 -3.38881314e-01 -1.41422737e+00
-1.25596428e+00 4.21373278e-01 8.42309535e-01 7.05957413e-02
-5.64194620e-01 2.87595958e-01 -3.08849820e-04 -7.93726593e-02
-2.26504266e-01 -1.01573341e-01 -5.83978333e-02 2.91215450e-01
-7.14479268e-01 1.84884787e-01 8.98199141e-01 1.51648033e+00
1.53396642e+00 -3.25369775e-01 6.32635176e-01 2.61211693e-01
4.36548322e-01 6.39017224e-01 9.12537426e-02 -1.13211668e+00
6.30747318e-01 8.76279473e-01 2.29667380e-01 -1.26077247e+00
-8.50786507e-01 5.06587438e-02 -3.01827490e-01 1.58039033e-01
2.63856351e-01 2.20530629e-01 -6.88332498e-01 1.49136353e+00
4.96161520e-01 4.76360805e-02 -3.78378443e-02 6.74293280e-01
5.74578345e-01 1.96448132e-01 -3.19836229e-01 1.34602249e-01
1.40113056e+00 -7.48859465e-01 -4.22447383e-01 -1.09062588e+00
5.30866563e-01 -4.20381039e-01 7.12477028e-01 1.49177417e-01
-7.55022347e-01 -5.22927880e-01 -1.52577519e+00 -2.74713784e-02
-3.37996900e-01 1.57405704e-01 5.43759093e-02 2.13771015e-01
-1.34863830e+00 3.95986110e-01 -1.03699601e+00 -1.00869894e+00
-1.55769095e-01 3.10844809e-01 -8.64134967e-01 -3.80191863e-01
-5.91424048e-01 1.07176602e+00 8.33887815e-01 1.34260058e-01
-7.32857168e-01 1.18568122e-01 -1.31400764e+00 -1.00345336e-01
7.33428836e-01 -6.49148285e-01 1.23562503e+00 -6.38907775e-03
-1.13803279e+00 8.45314145e-01 -7.48481750e-02 -3.82199079e-01
3.05209190e-01 -2.09055454e-01 -3.27391736e-02 1.03084214e-01
3.90187502e-01 5.92366219e-01 5.10459483e-01 -1.65185678e+00
-7.94569314e-01 -8.01903963e-01 2.12499052e-01 3.04187596e-01
3.44953895e-01 -6.25130177e-01 -8.03082407e-01 2.41296971e-03
1.16685402e+00 -1.33071113e+00 -3.43529493e-01 -8.10176805e-02
-4.40390706e-01 1.69408500e-01 9.65376854e-01 -6.73429549e-01
9.02533531e-01 -1.97749674e+00 1.90953031e-01 4.75613832e-01
2.34350920e-01 -3.88978451e-01 7.59481415e-02 8.13051462e-01
2.75480151e-01 -4.24477756e-01 -4.73133385e-01 -3.62434775e-01
2.84592900e-02 6.87399209e-01 -2.16282457e-01 6.96015120e-01
-1.71708018e-01 7.75300384e-01 -1.14582694e+00 -6.41691148e-01
4.17320698e-01 2.71142215e-01 -2.02024788e-01 1.05074599e-01
-2.04841509e-01 3.65512967e-01 -4.93426740e-01 5.24292946e-01
9.10248935e-01 6.24900460e-02 8.14168274e-01 -3.74118686e-01
-3.49993318e-01 3.47790271e-01 -1.57359016e+00 2.16959238e+00
-3.13885361e-01 4.41087514e-01 3.66159797e-01 -8.28442156e-01
1.16199100e+00 -8.64357054e-02 3.74678493e-01 -5.19341052e-01
-9.06319842e-02 2.64881909e-01 -4.65110123e-01 -1.29727393e-01
7.12832689e-01 2.52120495e-01 -3.81980687e-01 1.31927297e-01
1.69117272e-01 -8.04580569e-01 1.01600431e-01 3.73347968e-01
1.32546556e+00 7.92646170e-01 7.63584673e-01 -1.54485881e-01
5.27660966e-01 2.61301279e-01 5.18760145e-01 6.78296208e-01
-9.38336700e-02 6.20832503e-01 2.33937487e-01 -3.70897055e-01
-6.98683500e-01 -1.23896277e+00 4.04271156e-01 4.54967439e-01
6.44899607e-01 -9.69247401e-01 -6.01764739e-01 -5.87438762e-01
7.15975836e-02 4.80163366e-01 -2.50688821e-01 5.61079904e-02
-1.06549501e+00 9.31936875e-02 -6.64121006e-03 4.99929249e-01
5.42284489e-01 -5.74749172e-01 -1.30397224e+00 3.61562282e-01
-3.43277454e-01 -1.49389935e+00 -5.19686900e-02 3.92584145e-01
-8.25041533e-01 -1.32407951e+00 1.79614611e-02 -8.08034062e-01
6.95540190e-01 5.40878236e-01 1.12196219e+00 -1.27824128e-01
-7.22537488e-02 7.94119298e-01 -2.90492952e-01 -2.90398926e-01
-2.47366950e-01 -1.95127726e-01 -1.16481043e-01 -2.63229847e-01
-5.11598922e-02 -7.89424300e-01 -3.45329493e-01 3.63239646e-01
-3.20651889e-01 6.45731110e-03 5.42636454e-01 2.92329609e-01
9.80697632e-01 1.68697879e-01 -2.59283811e-01 -5.87845743e-01
4.09732610e-02 -3.55886251e-01 -9.22290444e-01 1.72267914e-01
-4.16111499e-01 6.10676818e-02 2.92638272e-01 1.87991723e-01
-7.45533109e-01 8.67247343e-01 1.50797144e-01 -1.40633568e-01
-1.77579582e-01 6.25926733e-01 -5.48487246e-01 -6.57177940e-02
5.58407843e-01 -8.59751850e-02 -5.28347753e-02 -3.53709906e-01
7.56610513e-01 3.49160582e-01 9.99474227e-01 -4.33909982e-01
1.07215095e+00 8.01593661e-01 4.36930329e-01 -8.75366986e-01
-7.03253031e-01 -7.56822169e-01 -1.54055059e+00 -4.63063151e-01
7.69387066e-01 -1.00956476e+00 -3.29793423e-01 1.44420892e-01
-1.40375865e+00 -3.91541600e-01 -3.66419077e-01 4.34772879e-01
-9.62911248e-01 4.88421381e-01 -2.30438918e-01 -6.55236483e-01
1.84298426e-01 -1.22496068e+00 1.46547234e+00 -8.10715184e-02
-2.90225387e-01 -7.22023964e-01 3.24526489e-01 -1.40722198e-02
-1.48602903e-01 7.08960474e-01 4.33147162e-01 -3.68159920e-01
-9.31202829e-01 -3.61057758e-01 -6.67087883e-02 -2.72297412e-01
7.25131929e-02 -2.95264453e-01 -8.94246578e-01 -2.67304122e-01
6.91675618e-02 1.95243925e-01 2.58641958e-01 1.56944081e-01
1.79134369e-01 -4.78800237e-02 -8.95280242e-01 3.75569582e-01
1.61780965e+00 1.57721892e-01 5.70021152e-01 8.41157317e-01
7.14136481e-01 9.11840498e-01 1.11048937e+00 3.03678036e-01
8.58053505e-01 1.14780056e+00 9.65772390e-01 2.57861972e-01
-2.50579327e-01 -5.62720597e-01 4.23583388e-01 7.91331589e-01
1.80711478e-01 7.60235339e-02 -1.30598533e+00 5.32168686e-01
-2.17117786e+00 -4.71028239e-01 -5.21419585e-01 2.21885538e+00
-6.13278076e-02 1.36585429e-01 2.48279423e-02 -9.68072936e-03
8.12871039e-01 1.54788136e-01 -4.27134424e-01 3.40255611e-02
1.36439785e-01 -1.12591922e-01 8.03320229e-01 7.96991944e-01
-9.83925581e-01 1.03838730e+00 6.12324381e+00 1.28224075e-01
-6.80839896e-01 6.40187785e-02 -4.18438792e-01 5.37675083e-01
1.26849025e-01 5.68737268e-01 -7.78482556e-01 -1.16903679e-02
8.70842040e-01 -5.64783253e-02 5.73817715e-02 1.28054070e+00
-1.40351042e-01 -3.50230008e-01 -1.21807122e+00 9.22121823e-01
2.28549674e-01 -9.34920192e-01 -4.23519909e-01 3.28126460e-01
1.59305945e-01 2.32770383e-01 -6.95028901e-01 1.21136263e-01
5.65081537e-01 -3.86523217e-01 1.34328878e+00 2.89782822e-01
5.85480511e-01 -4.92638111e-01 7.17609525e-01 7.25760698e-01
-1.95831084e+00 -3.37882712e-02 -4.19473946e-01 -1.44833615e-02
6.10046983e-01 4.27143574e-01 -1.37679076e+00 1.34356833e+00
6.71454489e-01 6.81886494e-01 -6.05589032e-01 9.31535304e-01
-6.53488159e-01 -7.90266767e-02 -4.81472343e-01 3.44939113e-01
-2.48730201e-02 -6.15761042e-01 6.78247273e-01 7.81136215e-01
6.25890791e-01 2.66626459e-02 6.90184474e-01 6.58057511e-01
2.39603281e-01 -8.68432075e-02 -9.74635661e-01 4.84054029e-01
6.61566734e-01 1.24879348e+00 -9.73233759e-01 -2.33513564e-01
-3.28177214e-01 1.01949728e+00 5.08532703e-01 -9.23212096e-02
-6.30817771e-01 -1.79596201e-01 1.34712279e-01 2.77754456e-01
3.01024944e-01 -7.54556596e-01 1.07207581e-01 -7.61224985e-01
1.73334390e-01 -2.95806676e-01 4.01989013e-01 -1.26682293e+00
-5.00823438e-01 6.81551397e-01 2.78713286e-01 -1.44605231e+00
-3.88483018e-01 -4.99259055e-01 -2.08375260e-01 4.83577788e-01
-1.32667089e+00 -1.47743440e+00 -8.82886469e-01 3.47456992e-01
3.35202456e-01 3.24092776e-01 7.25567997e-01 -1.24684870e-01
-3.87806483e-02 -3.57838243e-01 -3.70031059e-01 -7.61341602e-02
5.13092160e-01 -1.30482697e+00 6.48044229e-01 1.03660619e+00
4.73112136e-01 3.60026568e-01 8.37758124e-01 -9.61899757e-01
-1.59524727e+00 -1.04697084e+00 7.87699223e-01 -7.78593659e-01
6.77439213e-01 -5.26594102e-01 -6.01048648e-01 1.37665117e+00
-2.40163743e-01 -7.53215551e-02 -1.04705423e-01 -1.86137617e-01
-1.35191202e-01 -1.24134801e-01 -1.01986074e+00 4.38922226e-01
1.36164677e+00 -4.84573036e-01 -7.80547023e-01 3.72646332e-01
6.96347177e-01 -7.71262527e-01 -8.24189782e-01 4.44269121e-01
4.55022633e-01 -1.16348016e+00 1.09297383e+00 3.97948742e-01
-4.02732849e-01 -7.53273547e-01 -5.49287915e-01 -1.27748191e+00
-2.64079511e-01 -4.20334518e-01 4.51863185e-02 1.07438397e+00
-1.77879393e-01 -6.37406588e-01 9.19909000e-01 2.87212580e-01
-4.60992366e-01 -1.66785598e-01 -1.11060643e+00 -8.78021240e-01
-5.94846487e-01 -4.58183527e-01 7.08901644e-01 7.37062812e-01
-5.80407446e-03 4.20262814e-01 -5.17469756e-02 7.54553974e-01
7.60112286e-01 2.57546067e-01 1.38748968e+00 -1.52235389e+00
2.07659230e-01 1.57902956e-01 -1.10006046e+00 -1.32968330e+00
1.18894517e-01 -6.65657163e-01 6.21034205e-01 -2.01700974e+00
-2.46560812e-01 -4.23363447e-01 2.40271002e-01 4.88628119e-01
3.85835022e-01 -8.55694115e-02 1.92450687e-01 3.47291291e-01
-6.05055749e-01 3.64305228e-01 5.83343387e-01 -3.55207473e-02
-2.17433617e-01 -2.91660309e-01 -5.93721643e-02 9.94666398e-01
4.37736005e-01 -5.57238817e-01 -2.77527511e-01 -2.84140766e-01
1.14424586e-01 2.31226131e-01 4.07172143e-01 -1.34837496e+00
5.84328473e-01 1.26684621e-01 -1.60372585e-01 -1.14495766e+00
7.63326824e-01 -1.29239798e+00 7.46456623e-01 6.87233925e-01
4.56881046e-01 4.12030727e-01 4.51785475e-02 9.10412848e-01
-1.43567398e-01 -3.41578364e-01 3.96055996e-01 -3.23725343e-01
-1.16570997e+00 8.19997862e-02 -1.45864906e-03 -4.74720597e-01
1.11009085e+00 -6.31644547e-01 -2.95244306e-01 -4.53727186e-01
-8.99231374e-01 2.56835520e-01 1.06755006e+00 1.87742338e-01
7.38940477e-01 -1.31748402e+00 -3.37781817e-01 1.03736676e-01
4.60304528e-01 5.42926311e-01 -4.48291749e-02 7.27614343e-01
-8.98460567e-01 3.93514663e-01 -5.94203174e-02 -1.04980326e+00
-1.27728343e+00 5.16993463e-01 9.12152827e-02 -2.95045555e-01
-1.12080920e+00 3.93080652e-01 2.33186215e-01 -8.22225094e-01
-8.18976089e-02 -3.72474730e-01 4.29052971e-02 -8.78809243e-02
7.90015161e-02 5.26030779e-01 3.39234561e-01 -1.33561897e+00
-6.54588878e-01 9.77685750e-01 5.60728312e-01 -2.51821399e-01
1.35455275e+00 -5.84044695e-01 -3.42585921e-01 5.39354086e-01
8.29038501e-01 3.89072746e-01 -1.27232254e+00 -2.31058538e-01
4.87652928e-01 -5.21013439e-01 -2.52834111e-01 -1.31616369e-01
-4.96519595e-01 5.21835804e-01 5.39535880e-01 3.49057525e-01
7.39753067e-01 3.56841236e-01 5.18605053e-01 5.37785172e-01
1.39974582e+00 -9.88346517e-01 -7.14201629e-02 4.96182084e-01
1.08966458e+00 -9.46039677e-01 3.69521171e-01 -9.08158779e-01
-4.31865692e-01 1.25852573e+00 2.72166878e-01 -8.22919384e-02
3.46083999e-01 2.98443168e-01 -4.05369550e-02 -5.60544252e-01
-2.32503757e-01 -4.32538360e-01 -3.58192474e-02 1.01928198e+00
-3.35860878e-01 5.25690876e-02 3.19653213e-01 4.16503027e-02
-7.61245728e-01 -3.30791801e-01 6.86190009e-01 1.35991096e+00
-8.71746540e-01 -1.15298653e+00 -7.19904065e-01 -2.20558107e-01
4.35049713e-01 4.24032867e-01 -5.52230299e-01 1.29627109e+00
1.33061454e-01 9.86235261e-01 3.79921421e-02 -4.52286780e-01
9.18886542e-01 6.17811717e-02 6.77905679e-01 -9.18393135e-01
1.30257323e-01 -7.63085578e-03 3.00334364e-01 -1.02242184e+00
-4.62621689e-01 -7.30203569e-01 -1.44996393e+00 1.82175949e-01
-3.95940006e-01 8.10052305e-02 9.47139978e-01 6.97491765e-01
3.29849780e-01 2.44378343e-01 3.02516282e-01 -1.34635472e+00
-5.41652478e-02 -5.63752830e-01 -5.55570066e-01 2.30090901e-01
2.21062616e-01 -8.93893719e-01 -1.60533249e-01 1.23701669e-01] | [7.32467794418335, -2.235128164291382] |
8fb967c4-77f7-4283-b868-daeb9c6c1e0a | multi-modal-music-information-retrieval | 2002.00251 | null | https://arxiv.org/abs/2002.00251v1 | https://arxiv.org/pdf/2002.00251v1.pdf | Multi-Modal Music Information Retrieval: Augmenting Audio-Analysis with Visual Computing for Improved Music Video Analysis | This thesis combines audio-analysis with computer vision to approach Music Information Retrieval (MIR) tasks from a multi-modal perspective. This thesis focuses on the information provided by the visual layer of music videos and how it can be harnessed to augment and improve tasks of the MIR research domain. The main hypothesis of this work is based on the observation that certain expressive categories such as genre or theme can be recognized on the basis of the visual content alone, without the sound being heard. This leads to the hypothesis that there exists a visual language that is used to express mood or genre. In a further consequence it can be concluded that this visual information is music related and thus should be beneficial for the corresponding MIR tasks such as music genre classification or mood recognition. A series of comprehensive experiments and evaluations are conducted which are focused on the extraction of visual information and its application in different MIR tasks. A custom dataset is created, suitable to develop and test visual features which are able to represent music related information. Evaluations range from low-level visual features to high-level concepts retrieved by means of Deep Convolutional Neural Networks. Additionally, new visual features are introduced capturing rhythmic visual patterns. In all of these experiments the audio-based results serve as benchmark for the visual and audio-visual approaches. The experiments are conducted for three MIR tasks Artist Identification, Music Genre Classification and Cross-Genre Classification. Experiments show that an audio-visual approach harnessing high-level semantic information gained from visual concept detection, outperforms audio-only genre-classification accuracy by 16.43%. | ['Alexander Schindler'] | 2020-02-01 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 2.07177922e-01 -4.20745760e-01 -5.92277013e-02 1.03511237e-01
-5.03090739e-01 -7.06980705e-01 7.85564482e-01 5.53319573e-01
-2.35698521e-01 2.35119104e-01 4.10132200e-01 2.59466082e-01
-4.87218201e-01 -5.96040964e-01 -2.85762221e-01 -6.79665208e-01
-6.57055005e-02 1.68588027e-01 -7.58777261e-02 -3.90859991e-01
5.13027072e-01 6.52929187e-01 -2.64150333e+00 1.02313650e+00
-1.65191397e-01 1.20763028e+00 3.20424646e-01 7.96095848e-01
-2.10936815e-01 6.95198357e-01 -8.73646617e-01 1.46647081e-01
9.86908283e-03 -5.33352196e-01 -7.06877887e-01 2.93922480e-02
4.66147721e-01 2.65438616e-01 2.05309674e-01 7.68208265e-01
5.62920511e-01 1.62269160e-01 8.76477599e-01 -1.16241741e+00
-1.82142332e-01 6.71492279e-01 -3.03544581e-01 6.85948133e-02
9.17494178e-01 -1.62876099e-02 1.11685789e+00 -8.49519312e-01
9.07590508e-01 1.18114412e+00 5.23661911e-01 2.03991652e-01
-1.13424754e+00 -5.41147113e-01 -2.51000106e-01 6.15921736e-01
-1.29288638e+00 -2.61614382e-01 1.05899823e+00 -9.41554010e-01
7.34798849e-01 5.94516397e-01 1.14911592e+00 1.23952329e+00
-1.80509493e-01 6.88814044e-01 1.35707092e+00 -9.07194555e-01
9.83677208e-02 5.36525607e-01 2.58887988e-02 2.10296214e-01
-6.79348633e-02 1.07788406e-01 -9.60646033e-01 2.05630109e-01
3.94777209e-01 -3.26251060e-01 -2.42440939e-01 -3.61527771e-01
-1.09996569e+00 7.75316060e-01 5.02443492e-01 1.25952911e+00
-4.48557287e-01 9.81033146e-02 9.43361938e-01 3.65768790e-01
1.67746037e-01 6.25903726e-01 -6.64448887e-02 -2.71991163e-01
-1.24199188e+00 2.44670466e-01 5.42043090e-01 2.64141291e-01
5.74820995e-01 4.03705478e-01 -3.69591236e-01 8.17172408e-01
4.86114442e-01 3.62227201e-01 7.21415043e-01 -6.70242131e-01
-1.13344125e-01 7.68469870e-01 -3.88100177e-01 -1.26993001e+00
-4.59059983e-01 -4.71637011e-01 -4.01508242e-01 6.64285719e-01
3.34447831e-01 4.77227122e-01 -4.71793652e-01 1.49178994e+00
1.56000927e-01 -1.17496818e-01 -3.07567660e-02 1.04980421e+00
1.34488559e+00 6.08981907e-01 -1.17117189e-01 -2.73551673e-01
1.69537687e+00 -3.28038782e-01 -6.80494308e-01 4.37726229e-01
2.53019422e-01 -1.07437456e+00 1.22129881e+00 8.18000019e-01
-9.35691416e-01 -1.06894267e+00 -1.31428540e+00 1.49780124e-01
-8.36844206e-01 5.75558603e-01 2.86726505e-01 4.96729136e-01
-8.01529944e-01 5.91556668e-01 -2.40061864e-01 -4.27207232e-01
1.60308868e-01 1.34106293e-01 -4.46773767e-01 4.24143940e-01
-9.68989134e-01 6.79996789e-01 6.58755422e-01 7.24320859e-03
-8.38652313e-01 -5.95739722e-01 -7.50290275e-01 1.81886598e-01
-7.70393386e-02 -4.37357426e-01 7.65158713e-01 -1.61122561e+00
-1.35850191e+00 1.17473507e+00 2.00089261e-01 -3.73639852e-01
4.48213875e-01 1.13838837e-01 -4.17985529e-01 7.55384505e-01
7.99581930e-02 5.47470450e-01 1.01393998e+00 -1.61614621e+00
-5.15173912e-01 -4.04081196e-01 -1.49222285e-01 9.56284627e-02
-6.92931771e-01 1.36877969e-01 -2.78764784e-01 -8.23109686e-01
-2.76007503e-01 -7.10025430e-01 6.79311037e-01 -3.81006867e-01
-3.29582959e-01 -1.98161706e-01 8.78336549e-01 -5.81044793e-01
1.24483562e+00 -2.16157627e+00 4.60476756e-01 4.10784900e-01
4.50480916e-02 3.27180356e-01 -9.49283987e-02 6.60443664e-01
-3.43929291e-01 -1.46842122e-01 1.51595846e-01 3.29697393e-02
1.91830620e-01 -1.53771669e-01 -1.31020442e-01 1.84823364e-01
-2.50919402e-01 5.29381096e-01 -5.33381939e-01 -5.72574794e-01
7.31426060e-01 7.30314314e-01 -1.29225746e-01 -7.64839724e-02
-1.51944896e-02 5.11407733e-01 -6.51344135e-02 4.81520683e-01
2.44997248e-01 2.51056969e-01 9.58341509e-02 -4.47488070e-01
-1.47184089e-01 -3.11580729e-02 -1.28543019e+00 1.82265770e+00
-6.72134876e-01 1.22107625e+00 -3.74021754e-02 -1.14158523e+00
1.20334828e+00 6.19166315e-01 6.25267506e-01 -7.66415417e-01
4.00474429e-01 2.48177677e-01 -1.65573955e-01 -6.90431178e-01
4.00087208e-01 -3.55412275e-01 6.28027990e-02 1.15380682e-01
1.96179032e-01 1.32642895e-01 4.68098819e-01 -2.80446142e-01
6.21696770e-01 3.31600338e-01 1.76864743e-01 -2.56097652e-02
1.12590969e+00 1.39975742e-01 -1.93325981e-01 3.46661061e-01
2.46501416e-01 4.33693826e-01 3.75952125e-01 -3.40724140e-01
-9.13675368e-01 -9.62922394e-01 -2.69358903e-01 1.34240472e+00
-2.28369340e-01 -6.74285769e-01 -5.89953840e-01 -1.80363804e-01
-7.33850971e-02 3.06855232e-01 -7.66507089e-01 4.80167083e-02
-1.69205874e-01 -2.42936730e-01 6.02931738e-01 1.72495231e-01
1.28844127e-01 -1.71211803e+00 -9.13522184e-01 2.77254190e-02
-1.03537031e-01 -6.63729846e-01 4.25367892e-01 6.26215413e-02
-6.49558723e-01 -1.12474918e+00 -8.00006807e-01 -7.00783491e-01
-1.61656648e-01 9.37970430e-02 1.14144182e+00 -1.39211476e-01
-9.25833166e-01 8.17738533e-01 -6.82259560e-01 -6.78596020e-01
-6.03887498e-01 -2.21747812e-02 -1.61868051e-01 2.23429069e-01
3.98447007e-01 -7.66602099e-01 -4.73679572e-01 -7.27896765e-02
-9.61419702e-01 -2.72234112e-01 4.78258193e-01 4.60946232e-01
5.55923283e-01 -2.37351075e-01 4.17543560e-01 -1.96241483e-01
6.85255229e-01 -3.07780746e-02 -1.44095242e-01 -3.39986831e-02
-3.45582783e-01 -3.23579550e-01 3.69349569e-01 -5.32187283e-01
-6.21057808e-01 -3.60026285e-02 -1.59768283e-01 -7.94480085e-01
-7.51489103e-01 4.44315821e-01 -1.62226826e-01 -1.41041391e-02
9.81019258e-01 1.35045961e-01 -6.56502992e-02 -7.18232393e-01
2.65662968e-01 9.68628883e-01 4.38932866e-01 -3.01825047e-01
3.82821858e-01 5.08418143e-01 3.66152078e-01 -1.33558714e+00
-5.64490438e-01 -9.13515031e-01 -5.86216450e-01 -9.75151896e-01
1.06931067e+00 -8.62907410e-01 -8.39259505e-01 6.72596246e-02
-8.84569764e-01 1.44083396e-01 -6.79227114e-01 6.09870076e-01
-8.69888544e-01 3.23914826e-01 -4.24165785e-01 -1.10544217e+00
-4.63208616e-01 -9.48102653e-01 1.21820128e+00 -2.10696738e-02
-5.04591465e-01 -7.03976214e-01 3.45430464e-01 4.94612664e-01
-4.71842252e-02 4.26411420e-01 8.40802431e-01 -4.77494568e-01
-1.07429922e-02 -2.64108211e-01 -6.34748908e-03 3.22545707e-01
-2.97850192e-01 9.87523608e-03 -1.54325330e+00 -1.08261682e-01
-3.90224457e-01 -4.15159166e-01 1.06697202e+00 3.50023657e-01
9.37746584e-01 1.34751663e-01 1.48852188e-02 3.36739451e-01
1.65488601e+00 9.69623253e-02 6.04311705e-01 7.11436033e-01
5.87267399e-01 8.02552938e-01 7.28261232e-01 5.15132487e-01
-3.81815851e-01 1.34236443e+00 6.34210348e-01 2.49603074e-02
-5.20366371e-01 -7.41591379e-02 5.09366512e-01 4.84854162e-01
-3.71385574e-01 6.48072660e-02 -7.34111309e-01 3.32949877e-01
-1.77468705e+00 -1.28108609e+00 -5.13754189e-01 2.17119074e+00
3.77516270e-01 -1.14766330e-01 5.81070125e-01 1.15051854e+00
6.72444880e-01 1.04120020e-02 1.94266751e-01 -7.96687245e-01
-2.40518942e-01 5.41185260e-01 -1.94305807e-01 4.20618504e-02
-1.22663903e+00 5.92419684e-01 5.13532019e+00 1.12746882e+00
-1.38127935e+00 2.79144403e-02 -1.61664754e-01 -2.45888695e-01
7.34441727e-02 -3.98667991e-01 -2.65659839e-01 1.12469934e-01
9.58227932e-01 3.50507408e-01 2.80402631e-01 7.34832227e-01
3.66304547e-01 -3.11345309e-02 -8.82065773e-01 1.47625613e+00
5.91658294e-01 -1.36984229e+00 4.49367911e-01 1.19672969e-01
3.26513350e-01 -4.80937898e-01 8.71863812e-02 1.30339086e-01
-8.11948597e-01 -1.02467775e+00 1.09063780e+00 7.89783716e-01
7.36309886e-01 -1.01054049e+00 7.57812917e-01 -1.32782012e-01
-1.56659305e+00 -3.37615043e-01 -8.49858746e-02 -1.38948545e-01
-1.64301813e-01 1.82478335e-02 -6.67484343e-01 9.25883412e-01
8.90578866e-01 1.06402349e+00 -6.52926922e-01 1.22181010e+00
-7.20388070e-02 3.86208177e-01 -2.26772539e-02 1.05334148e-02
1.04797348e-01 7.32381344e-02 9.41442668e-01 1.59379745e+00
4.46915030e-01 -8.87370527e-01 -3.45725305e-02 8.97057712e-01
5.04800916e-01 8.59237254e-01 -7.72334397e-01 -1.84715047e-01
-4.81038243e-02 1.61571336e+00 -1.02151322e+00 -3.04975897e-01
7.50510395e-02 7.53505766e-01 -3.40960503e-01 1.00051323e-02
-5.36859989e-01 -4.85157281e-01 4.86276507e-01 2.58129090e-01
5.28344572e-01 1.97165072e-01 -9.65395644e-02 -7.41955936e-01
-1.13980874e-01 -7.73578405e-01 3.74252617e-01 -1.25770068e+00
-9.78298247e-01 5.58637977e-01 -4.64000255e-02 -1.54651427e+00
-4.24975574e-01 -8.60812783e-01 -5.03042579e-01 6.39019310e-01
-1.02242148e+00 -1.46531785e+00 -3.51830304e-01 7.28002787e-01
4.96092469e-01 -5.31862319e-01 1.10959065e+00 3.24422032e-01
-1.27969950e-01 1.45848364e-01 -1.43711060e-01 1.87499478e-01
6.44391119e-01 -1.35231864e+00 -9.92854595e-01 2.81225652e-01
9.51161087e-01 5.75024039e-02 7.71503925e-01 -1.13708004e-01
-1.08484542e+00 -5.19733429e-01 7.13210940e-01 -2.01662213e-01
5.62797844e-01 -1.18798606e-01 -5.28963327e-01 -1.55461565e-01
3.13736349e-01 -5.45817196e-01 1.09626424e+00 1.20889489e-02
-4.78068948e-01 -1.74578264e-01 -6.93381548e-01 1.74267039e-01
4.60314035e-01 -7.34839857e-01 -7.72324741e-01 1.02072574e-01
1.01321554e-02 1.43998370e-01 -1.06044793e+00 1.42697573e-01
8.93737197e-01 -1.21963656e+00 1.34463429e+00 -5.51294625e-01
6.18279815e-01 -4.90262061e-01 -7.98346251e-02 -9.16454077e-01
1.19428776e-01 -1.42339006e-01 1.45432398e-01 1.37342262e+00
1.84574232e-01 3.03823352e-01 3.73364359e-01 -8.02654207e-01
-1.22619597e-02 -5.33872843e-02 -9.90623951e-01 -6.79133177e-01
-2.73670554e-01 -8.84618640e-01 4.75299917e-02 9.74402726e-01
1.91485316e-01 6.48062587e-01 -1.98998570e-01 -4.32383150e-01
1.22583553e-01 5.65863431e-01 8.87068748e-01 -1.67672372e+00
-3.61146748e-01 -1.01302922e+00 -1.19406331e+00 1.72262609e-01
1.75316930e-01 -1.16894257e+00 -3.18642676e-01 -1.64764738e+00
2.85041761e-02 1.82860032e-01 -5.95710158e-01 3.19905758e-01
5.67020714e-01 1.11593163e+00 7.50567853e-01 1.99897960e-01
-5.02080977e-01 1.35852635e-01 1.06675673e+00 -2.83256888e-01
-3.59171510e-01 -5.72467558e-02 -1.81032583e-01 7.72109091e-01
7.23388314e-01 -1.83786243e-01 -1.99569672e-01 4.60165292e-01
6.44264162e-01 -2.15490218e-02 7.10582793e-01 -1.41811550e+00
-3.65795374e-01 4.23179448e-01 6.31266117e-01 -6.22372150e-01
7.54833758e-01 -8.97771597e-01 1.42982289e-01 4.96205419e-01
-5.23008943e-01 -2.32693002e-01 3.84634167e-01 5.23028731e-01
-6.17615640e-01 -4.74126577e-01 6.84241354e-01 -1.69539750e-01
-7.96540022e-01 -4.75690782e-01 -7.05335081e-01 -2.49526486e-01
8.49741459e-01 -6.13655746e-01 2.65467554e-01 -4.71493453e-01
-1.36600459e+00 -5.99501014e-01 1.67920083e-01 5.90959132e-01
4.72842455e-01 -1.46844113e+00 -6.46287799e-01 -1.08305581e-01
6.71930254e-01 -9.44527864e-01 6.01828098e-01 9.15782750e-01
-5.04853368e-01 3.97825152e-01 -7.63407648e-01 -9.12122011e-01
-2.13650203e+00 8.94452155e-01 7.55247250e-02 4.06681979e-03
-6.25864923e-01 3.81217182e-01 -7.64025375e-02 1.68852895e-01
5.17699957e-01 -7.05979243e-02 -1.21392763e+00 9.89683270e-01
6.43042743e-01 3.88736784e-01 7.33395889e-02 -1.12716699e+00
-2.17113674e-01 1.05123103e+00 7.75838256e-01 -4.79954869e-01
1.31509387e+00 -1.48686273e-02 -1.00629069e-01 1.11699307e+00
1.11895788e+00 1.65413767e-01 -4.08139825e-01 2.26373211e-01
3.49665806e-02 -2.54249811e-01 2.21389249e-01 -9.55793083e-01
-9.66628015e-01 1.45597303e+00 1.13543665e+00 6.76455677e-01
1.26246202e+00 -8.29185322e-02 -2.87876688e-02 1.47175714e-01
-6.02652244e-02 -1.20932317e+00 3.23820889e-01 4.26057205e-02
1.22029126e+00 -8.09955835e-01 -4.47976291e-02 -1.11204155e-01
-5.00242770e-01 1.91020238e+00 -3.14692594e-02 1.98176354e-02
4.08758134e-01 -5.86715415e-02 5.22229895e-02 -4.25862610e-01
-2.71216810e-01 -9.51073170e-01 9.08371091e-01 7.74931669e-01
6.18955553e-01 1.83030993e-01 -4.25367445e-01 6.23760283e-01
-4.42610919e-01 -5.73096098e-03 2.76871800e-01 6.19495690e-01
-5.87340117e-01 -1.16988194e+00 -8.16481531e-01 -7.70682767e-02
-5.64422905e-01 2.76393861e-01 -9.06109691e-01 1.07885814e+00
6.75791562e-01 1.05285716e+00 2.03333292e-02 -4.61250335e-01
2.10491851e-01 4.22010422e-01 7.51734972e-01 -4.14053082e-01
-1.21267712e+00 5.21806896e-01 1.17984535e-02 -3.69195580e-01
-9.73949969e-01 -5.88855028e-01 -8.56254458e-01 7.99941272e-02
-1.55072962e-03 1.02629356e-01 1.00730693e+00 7.38909304e-01
-6.08848780e-03 1.08500576e+00 2.84484923e-01 -1.17678690e+00
2.45849758e-01 -1.11936378e+00 -6.88797235e-01 6.01339757e-01
7.44166747e-02 -8.54225099e-01 -3.44718874e-01 2.08717525e-01] | [15.739297866821289, 5.152281761169434] |
788e242e-f07c-4afd-8f0f-43b53060c936 | 3dv-3d-dynamic-voxel-for-action-recognition | 2005.05501 | null | https://arxiv.org/abs/2005.05501v1 | https://arxiv.org/pdf/2005.05501v1.pdf | 3DV: 3D Dynamic Voxel for Action Recognition in Depth Video | To facilitate depth-based 3D action recognition, 3D dynamic voxel (3DV) is proposed as a novel 3D motion representation. With 3D space voxelization, the key idea of 3DV is to encode 3D motion information within depth video into a regular voxel set (i.e., 3DV) compactly, via temporal rank pooling. Each available 3DV voxel intrinsically involves 3D spatial and motion feature jointly. 3DV is then abstracted as a point set and input into PointNet++ for 3D action recognition, in the end-to-end learning way. The intuition for transferring 3DV into the point set form is that, PointNet++ is lightweight and effective for deep feature learning towards point set. Since 3DV may lose appearance clue, a multi-stream 3D action recognition manner is also proposed to learn motion and appearance feature jointly. To extract richer temporal order information of actions, we also divide the depth video into temporal splits and encode this procedure in 3DV integrally. The extensive experiments on 4 well-established benchmark datasets demonstrate the superiority of our proposition. Impressively, we acquire the accuracy of 82.4% and 93.5% on NTU RGB+D 120 [13] with the cross-subject and crosssetup test setting respectively. 3DV's code is available at https://github.com/3huo/3DV-Action. | ['Joey Tianyi Zhou', 'Zhiguo Cao', 'Wenxiang Jiang', 'Yancheng Wang', 'Fu Xiong', 'Yang Xiao', 'Junsong Yuan'] | 2020-05-12 | 3dv-3d-dynamic-voxel-for-action-recognition-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Wang_3DV_3D_Dynamic_Voxel_for_Action_Recognition_in_Depth_Video_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_3DV_3D_Dynamic_Voxel_for_Action_Recognition_in_Depth_Video_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-human-action-recognition'] | ['computer-vision'] | [-1.14219241e-01 -3.65021139e-01 -6.89408839e-01 -2.98780918e-01
-4.64600325e-01 -2.03631490e-01 4.40640092e-01 -5.19715965e-01
-2.14578435e-01 2.25516930e-01 5.69351017e-01 1.08642787e-01
3.61440815e-02 -6.41727328e-01 -6.67159915e-01 -8.68719399e-01
-4.02620018e-01 6.01020493e-02 4.26036566e-01 1.23181522e-01
1.66967064e-01 7.40879476e-01 -1.32522631e+00 3.48496437e-01
4.41905230e-01 1.41668618e+00 -3.19777317e-02 4.87269074e-01
-3.03361088e-01 9.59572911e-01 -3.93156201e-01 2.34655678e-01
7.00037181e-01 -2.01645553e-01 -8.00483167e-01 5.46626925e-01
4.77442801e-01 -9.49325144e-01 -9.87493813e-01 6.19621396e-01
4.69020188e-01 3.08809876e-01 4.42903161e-01 -1.40103924e+00
-6.86269045e-01 1.03118502e-01 -1.05519915e+00 5.90509653e-01
5.27036130e-01 5.96136987e-01 7.54792273e-01 -1.05432165e+00
7.74034023e-01 1.54211307e+00 1.48485616e-01 6.35785520e-01
-7.54261434e-01 -6.67310596e-01 3.50702077e-01 4.15725142e-01
-1.23035192e+00 -2.21765786e-01 1.12255752e+00 -3.80216986e-01
1.16611552e+00 2.39374951e-01 1.14935589e+00 1.12564969e+00
2.80753464e-01 1.62237573e+00 7.87516534e-01 1.97720841e-01
-1.69542395e-02 -7.76769757e-01 2.78550517e-02 8.85167778e-01
-3.07254761e-01 1.95102736e-01 -9.04354811e-01 1.12357065e-01
1.37478077e+00 5.00058830e-01 -4.04489070e-01 -7.10270584e-01
-1.43940616e+00 5.46484828e-01 6.40635371e-01 -1.12932473e-02
-4.47272748e-01 7.07002699e-01 7.99890101e-01 2.63639688e-01
7.31783986e-01 -3.65210772e-01 -5.61375082e-01 -4.55207884e-01
-4.52081442e-01 2.32670173e-01 8.87258872e-02 9.27612722e-01
6.18274808e-01 1.60394311e-01 -3.41007948e-01 7.44670630e-01
5.38102627e-01 7.65545905e-01 5.91304660e-01 -1.29801118e+00
6.70933962e-01 9.16523933e-01 -2.25459069e-01 -1.07103372e+00
-1.80030510e-01 -1.33923545e-01 -9.72762346e-01 3.44157159e-01
1.70062661e-01 2.94274896e-01 -1.10267174e+00 1.44012237e+00
7.68742263e-01 6.11586392e-01 -1.07596114e-01 1.31097472e+00
1.12627792e+00 7.67311335e-01 -1.57012463e-01 -1.37331188e-01
1.13078582e+00 -1.08068883e+00 -4.45005298e-01 8.96106288e-02
7.48640656e-01 -2.58997113e-01 9.43424165e-01 3.94213498e-01
-1.12116802e+00 -7.71588624e-01 -8.42218995e-01 -3.73100042e-01
1.32883862e-01 -1.62663609e-01 8.68455291e-01 2.31619388e-01
-8.58342052e-01 3.87169927e-01 -1.20800984e+00 -2.23667249e-02
1.05057549e+00 3.60524684e-01 -6.34665430e-01 -3.96314234e-01
-1.05905747e+00 3.57765704e-01 1.33645162e-01 1.16221264e-01
-1.07764196e+00 -6.40077651e-01 -1.10347176e+00 -4.91688639e-01
2.28135347e-01 -7.70538390e-01 1.05642414e+00 -6.33404791e-01
-1.46530437e+00 1.04806554e+00 -1.50899902e-01 -1.92412212e-01
6.43841088e-01 -2.27743506e-01 -6.87240735e-02 5.02141595e-01
2.55869001e-01 7.46809721e-01 7.60232091e-01 -7.88191378e-01
-6.19552851e-01 -7.38260686e-01 1.23179846e-01 4.76586521e-01
-5.16659860e-03 -2.10654154e-01 -8.34657073e-01 -6.94811225e-01
7.10748076e-01 -7.23770738e-01 -5.66510968e-02 6.46783710e-01
-1.88285485e-01 -5.65460026e-01 9.36985373e-01 -5.04571259e-01
8.67991388e-01 -2.25492859e+00 4.00879890e-01 -1.33141366e-04
4.24910784e-01 9.84519795e-02 -1.00291029e-01 -7.02079758e-02
-6.55749440e-02 -2.45124977e-02 -3.00768673e-01 -4.23476070e-01
4.03969549e-02 2.96041697e-01 -6.00740090e-02 8.78947973e-01
2.63098300e-01 1.35198045e+00 -9.68511283e-01 -5.80239654e-01
7.24101782e-01 6.44345462e-01 -5.28732598e-01 7.80463964e-03
1.26025164e-02 5.82368314e-01 -1.12324166e+00 1.14396548e+00
8.30328524e-01 -1.28029510e-01 -6.81178808e-01 -1.76658630e-01
5.71546629e-02 -1.33899245e-02 -1.11029589e+00 2.42334414e+00
-1.26665905e-01 4.59292471e-01 -1.80508286e-01 -1.01241791e+00
8.70598435e-01 6.45673648e-02 1.10881937e+00 -1.10422397e+00
2.62099922e-01 3.01389880e-02 -5.02471387e-01 -6.73724234e-01
9.18324813e-02 1.84148207e-01 -1.70876518e-01 4.27803218e-01
-5.02828695e-02 8.48995522e-03 -3.63615394e-01 -2.43312251e-02
1.22270024e+00 5.74133515e-01 1.92667842e-01 1.26341134e-01
6.00297928e-01 -9.04012471e-02 7.65227437e-01 1.78146511e-01
-7.64940798e-01 7.19451725e-01 3.59724432e-01 -5.84700525e-01
-6.69306040e-01 -1.19782257e+00 -1.87088400e-02 5.50253391e-01
4.54172254e-01 -1.16632864e-01 -1.63019806e-01 -8.94357562e-01
2.83172071e-01 4.98909652e-02 -8.50639343e-01 -2.92338490e-01
-8.21503043e-01 -1.88956663e-01 2.73404062e-01 7.63661385e-01
9.11144853e-01 -1.03082097e+00 -8.74275863e-01 1.15732491e-01
-1.23353787e-01 -9.66073096e-01 -7.03327715e-01 -1.99388355e-01
-1.25924253e+00 -9.78574574e-01 -9.82949197e-01 -6.13422096e-01
3.73234391e-01 7.75459588e-01 7.33218670e-01 8.18035603e-02
-2.87825555e-01 6.48245275e-01 -7.34941542e-01 1.45148188e-01
2.35858291e-01 -4.56392050e-01 1.78370774e-01 1.79923639e-01
4.64407980e-01 -8.78028512e-01 -1.15483415e+00 2.94413656e-01
-8.52449358e-01 -7.20036924e-02 5.82694829e-01 5.56579769e-01
1.02239239e+00 -1.95052549e-01 -5.52929332e-03 -1.38090536e-01
-9.64543596e-02 -4.54069227e-01 -1.34840131e-01 -2.43647382e-01
7.80509263e-02 -2.82132149e-01 1.83991902e-02 -3.93473655e-01
-6.07659161e-01 1.95390478e-01 -9.86033678e-02 -1.24976170e+00
-1.63251430e-01 1.17908515e-01 -4.65112329e-01 -1.01074949e-01
1.87645808e-01 6.02675974e-01 2.06897691e-01 -4.61821169e-01
3.93461823e-01 4.77241367e-01 1.89908832e-01 -3.55164319e-01
7.19076574e-01 9.53873634e-01 1.28966970e-02 -7.23834336e-01
-5.14810085e-01 -6.08098269e-01 -8.76671016e-01 -5.49909711e-01
1.20311713e+00 -1.16327882e+00 -8.14745247e-01 8.22324276e-01
-1.03763461e+00 -5.88664114e-01 -4.07671362e-01 6.69545174e-01
-9.03618336e-01 6.89917088e-01 -5.71810067e-01 -5.25128841e-01
-3.63391131e-01 -1.16316080e+00 1.48913801e+00 1.30123342e-03
5.49983084e-02 -7.69571424e-01 -1.33346692e-02 5.35256863e-01
-2.04322249e-01 6.66620493e-01 4.21343923e-01 1.85256973e-02
-9.32542682e-01 -1.12719700e-01 -2.42318958e-01 4.75073427e-01
2.29349717e-01 -9.98127684e-02 -7.82785714e-01 -6.33084923e-02
1.09232262e-01 -3.82516414e-01 1.03166914e+00 6.46137893e-01
1.34360325e+00 2.65527070e-02 -2.31123790e-01 1.01208556e+00
1.12576413e+00 2.81616807e-01 6.59784615e-01 2.94930398e-01
1.20118845e+00 2.97983795e-01 9.62413847e-01 7.33735263e-01
3.95468652e-01 8.00240815e-01 6.79680884e-01 3.30436006e-02
-2.67364860e-01 -2.34280333e-01 5.26266277e-01 8.21123719e-01
-2.56009042e-01 1.26910627e-01 -7.80341268e-01 2.72145420e-01
-1.78164816e+00 -1.07557082e+00 -2.25849599e-01 1.91646636e+00
4.81461972e-01 1.24271952e-01 2.64859259e-01 2.43560612e-01
3.00252229e-01 7.21770525e-01 -8.27039063e-01 2.32241780e-01
-2.22506925e-01 3.23731676e-02 4.43675697e-01 2.90119261e-01
-1.18700802e+00 9.61610138e-01 4.24578381e+00 9.45733905e-01
-1.09557545e+00 9.43392962e-02 7.39339709e-01 -5.51463127e-01
-1.48768812e-01 -3.15419137e-01 -5.57874918e-01 4.77546751e-01
1.15225166e-01 -2.00424045e-02 -5.66700846e-02 7.02757239e-01
7.12736964e-01 -9.12341848e-02 -1.01359844e+00 1.49214637e+00
3.67546417e-02 -1.35040236e+00 2.99814820e-01 2.25319743e-01
5.54621279e-01 1.19540028e-01 4.45542745e-02 3.19530457e-01
-8.18799734e-02 -9.33028221e-01 8.36432517e-01 5.27623415e-01
8.13916862e-01 -8.57281446e-01 3.28225374e-01 2.13370398e-01
-1.56581318e+00 -8.15066397e-02 -4.31189030e-01 -7.13646188e-02
3.64672184e-01 3.61917675e-01 -6.73203543e-02 6.62420571e-01
9.70916331e-01 1.65867019e+00 -2.74826616e-01 9.98357952e-01
-9.60576162e-02 1.73241824e-01 -1.06677413e-01 1.64496049e-01
6.26596034e-01 -1.08557798e-01 6.36658669e-01 6.90679431e-01
2.89554060e-01 6.81746006e-01 2.76864290e-01 5.75444460e-01
1.45907104e-01 -5.86338565e-02 -7.27401674e-01 1.71024203e-01
8.36292431e-02 8.10579181e-01 -6.36656284e-01 -1.93501711e-01
-4.89671141e-01 1.53391230e+00 7.81750157e-02 3.32840681e-01
-9.00196493e-01 -2.68053580e-02 1.09296334e+00 -1.27620205e-01
5.78131557e-01 -5.41690469e-01 3.15782018e-02 -1.25505042e+00
3.26458156e-01 -6.57475054e-01 3.22924554e-01 -8.52404118e-01
-1.19621241e+00 2.09772170e-01 8.70183930e-02 -1.90211999e+00
1.54118642e-01 -7.37606764e-01 -4.54445034e-01 5.79351902e-01
-1.40199196e+00 -1.05726302e+00 -6.37576640e-01 1.03630793e+00
9.80360806e-01 9.86068249e-02 2.74849981e-01 2.64435500e-01
-4.73665237e-01 4.39753056e-01 -1.87032491e-01 3.59112769e-01
2.97638327e-01 -9.01823163e-01 6.12642884e-01 4.66891855e-01
1.00706466e-01 5.55168204e-02 -6.20885715e-02 -6.32840395e-01
-1.98946464e+00 -1.23475611e+00 3.67696345e-01 -7.94957638e-01
4.88273561e-01 -3.27039033e-01 -8.81663322e-01 4.59257513e-01
-3.39292288e-01 5.15952587e-01 5.19030392e-01 -5.49285054e-01
-3.45082343e-01 -1.72801524e-01 -1.00226164e+00 4.58314121e-01
1.91326237e+00 -4.42106247e-01 -5.26487708e-01 3.78311515e-01
1.11639535e+00 -7.15699613e-01 -1.11384857e+00 4.28313911e-01
5.11163116e-01 -1.05149555e+00 1.32068551e+00 -5.19878268e-01
6.20743990e-01 -3.43288034e-01 -4.99567568e-01 -7.90681481e-01
-3.06777477e-01 -4.65018243e-01 -6.23138011e-01 6.93270922e-01
-1.21866249e-01 -3.56797963e-01 1.10722792e+00 3.20539266e-01
-3.92021894e-01 -1.28871512e+00 -1.32104015e+00 -8.01539958e-01
-1.39658600e-02 -8.63708735e-01 4.89708006e-01 9.16772008e-01
-3.00481260e-01 -5.35439886e-02 -2.69113213e-01 5.51857725e-02
6.82080984e-01 1.59105614e-01 9.43158805e-01 -8.13786626e-01
-1.43765375e-01 -5.89056730e-01 -1.06362247e+00 -1.92514932e+00
1.78878501e-01 -9.77255046e-01 -2.88080335e-01 -1.63120914e+00
1.08216822e-01 -3.17792147e-01 -4.36028719e-01 4.06024039e-01
3.22013795e-02 3.26487184e-01 1.51055217e-01 3.81124735e-01
-6.32736921e-01 1.21242571e+00 1.97699416e+00 -3.35718244e-01
-3.47619355e-01 -8.82229879e-02 -8.94187391e-02 4.95795935e-01
4.80184376e-01 -1.01209924e-01 -6.62990630e-01 -6.62171423e-01
-5.39335370e-01 3.48897994e-01 6.96000397e-01 -9.74575222e-01
-8.95713493e-02 -3.18300307e-01 6.86130106e-01 -1.02122915e+00
7.45111227e-01 -7.18885481e-01 -1.87728554e-01 5.43116271e-01
-7.17395032e-03 -1.70966893e-01 5.93095496e-02 7.84165502e-01
-2.97752529e-01 5.44818282e-01 5.57785392e-01 -2.93492466e-01
-1.42903984e+00 1.35090005e+00 1.42851174e-01 1.33763209e-01
1.32877040e+00 -8.93795073e-01 2.24315748e-01 -1.29710063e-01
-7.00337112e-01 4.65332299e-01 4.56612825e-01 6.63973749e-01
1.25845778e+00 -1.88013530e+00 -7.12951481e-01 2.74719447e-01
1.60549507e-01 3.71744454e-01 8.63300145e-01 1.11865604e+00
-5.87756753e-01 2.83434510e-01 -3.64751041e-01 -1.13876879e+00
-1.20017600e+00 3.13279122e-01 5.23452818e-01 1.05169348e-01
-1.20294440e+00 1.16490781e+00 4.62196082e-01 -1.41183361e-01
4.06526029e-01 -6.13354921e-01 -1.19742803e-01 1.02563435e-02
6.91133797e-01 2.69814670e-01 -4.33297127e-01 -8.60721588e-01
-7.23698735e-01 1.07855427e+00 1.11349948e-01 -9.93612632e-02
1.33292592e+00 -1.63330391e-01 1.06444947e-01 4.96131122e-01
1.81046140e+00 -5.73248029e-01 -1.85789168e+00 -3.30164790e-01
-4.99832571e-01 -1.12653494e+00 6.30708337e-02 -5.50410859e-02
-1.52102232e+00 9.57820356e-01 7.10928380e-01 -3.28669757e-01
1.20947206e+00 2.54023910e-01 1.07766306e+00 1.09556496e-01
5.43375552e-01 -6.40202224e-01 5.41969538e-01 4.62613046e-01
1.04690695e+00 -1.28338683e+00 1.59434468e-01 -2.83731610e-01
-6.63864017e-01 8.80219221e-01 6.82510495e-01 -4.02347028e-01
7.56976306e-01 -3.04934442e-01 -2.57559633e-03 -4.92557228e-01
-5.18732369e-01 -5.52302711e-02 2.78243452e-01 6.47506475e-01
4.30526510e-02 -7.17896745e-02 -5.87045327e-02 3.97871494e-01
2.37052187e-01 9.47939530e-02 7.61489645e-02 1.12085664e+00
-3.05505544e-01 -6.81597114e-01 -2.02789865e-02 2.25665197e-01
-9.07635614e-02 2.40244076e-01 -3.13024312e-01 9.44421351e-01
2.06456974e-01 5.55502832e-01 2.57802367e-01 -7.31894255e-01
5.38117111e-01 -3.46558660e-01 6.04079485e-01 -5.26667833e-01
-7.27281272e-02 7.81778432e-03 -2.85871506e-01 -1.40390050e+00
-5.96278071e-01 -8.53028297e-01 -1.67900527e+00 -2.98607796e-01
3.30972672e-01 -3.71066183e-01 1.92605898e-01 7.36431122e-01
3.99083912e-01 3.55902225e-01 9.65722322e-01 -1.16945982e+00
-3.59010816e-01 -5.31940877e-01 -6.28668904e-01 4.16331083e-01
5.09244382e-01 -1.02174342e+00 -3.75207543e-01 -1.54043898e-01] | [8.26118278503418, 0.23761463165283203] |
d7e15527-f233-40a8-b3f8-b5749e0cd3e6 | reformulating-the-sir-model-in-terms-of-the | 2110.00364 | null | https://arxiv.org/abs/2110.00364v1 | https://arxiv.org/pdf/2110.00364v1.pdf | Reformulating the SIR model in terms of the number of COVID-19 detected cases: well-posedness of the observational model | Compartmental models are popular in the mathematics of epidemiology for their simplicity and wide range of applications. Although they are typically solved as initial value problems for a system of ordinary differential equations, the observed data is typically akin of a boundary value type problem: we observe some of the dependent variables at given times, but we do not know the initial conditions. In this paper, we reformulate the classical Susceptible-Infectious-Recovered system in terms of the number of detected positive infected cases at different times, we then prove the existence and uniqueness of a solution to the derived boundary value problem and then present a numerical algorithm to approximate the solution. | ['Anotida Madzvamuse', 'Duc-Lam Duong', 'James Van Yperen', 'Hayley Wragg', 'Eduard Campillo-Funollet'] | 2021-10-01 | null | null | null | null | ['epidemiology'] | ['medical'] | [ 2.30446503e-01 -2.71301568e-01 -8.71708393e-02 1.64403781e-01
5.86425737e-02 -4.68404382e-01 2.17859089e-01 3.34711522e-01
-6.16830111e-01 1.00432193e+00 -4.52795625e-01 -2.71953017e-01
-6.41778290e-01 -6.53389633e-01 -3.99470389e-01 -1.19255745e+00
-3.26273173e-01 6.75863504e-01 4.79456261e-02 -1.18411653e-01
-4.59376425e-02 4.97859538e-01 -9.54921305e-01 -4.76230830e-01
8.56885195e-01 5.61651051e-01 6.78382218e-02 8.20570052e-01
-3.23678995e-03 5.79662800e-01 -6.51601374e-01 1.27256140e-01
1.13998279e-02 -6.23919368e-01 -4.32174355e-01 2.50815660e-01
-7.63682067e-01 -3.38772625e-01 1.03031173e-01 1.00390291e+00
3.41759026e-01 1.62257954e-01 1.22014153e+00 -1.62909293e+00
-6.30144596e-01 -2.24743970e-02 -9.10065770e-01 4.68501478e-01
1.17419101e-01 -1.46473631e-01 1.91659719e-01 -4.39757496e-01
6.25075281e-01 9.29844320e-01 6.52661324e-01 5.44685006e-01
-1.50211370e+00 -3.90897654e-02 -1.08373184e-02 -1.81300610e-01
-1.49943209e+00 -2.14953385e-02 2.46163249e-01 -8.75293493e-01
4.37632680e-01 2.09770024e-01 6.31713152e-01 1.02569592e+00
4.83969212e-01 1.91542447e-01 9.83762622e-01 -1.50732234e-01
5.96337259e-01 1.93279028e-01 2.56863505e-01 2.34305084e-01
7.16636062e-01 -1.55559741e-02 3.51229876e-01 -6.69199765e-01
9.90086377e-01 8.40664625e-01 -3.12206298e-01 -3.46058518e-01
-8.96750152e-01 9.98359501e-01 7.11114556e-02 2.51117080e-01
-9.09018338e-01 -1.62433892e-01 -4.11794335e-03 3.27050686e-01
5.91948152e-01 3.04842591e-02 -3.00590992e-01 5.64442873e-01
-3.94287884e-01 5.36342978e-01 9.84553397e-01 4.93271470e-01
3.64466846e-01 -1.92247882e-01 -1.11312538e-01 5.54052293e-01
2.97437996e-01 9.37243700e-01 -2.37551630e-02 -8.42418611e-01
-1.69950560e-01 4.26138848e-01 5.82252681e-01 -6.67713702e-01
-2.14915693e-01 -2.60031402e-01 -1.18483233e+00 -1.89049065e-01
6.44042492e-01 -5.30033886e-01 -7.35900164e-01 1.72297227e+00
4.85217959e-01 3.99914593e-01 5.64599447e-02 6.62809730e-01
5.79252187e-03 1.05528820e+00 8.68513808e-02 -1.07734644e+00
9.81010735e-01 -2.34060600e-01 -9.30214822e-01 9.95750576e-02
4.78539646e-01 -3.41112733e-01 1.76184252e-02 4.19455729e-02
-9.66979861e-01 3.68644416e-01 -4.06024963e-01 5.73940516e-01
-3.94818395e-01 -4.46053982e-01 -5.05031236e-02 8.57793391e-02
-1.08235776e+00 3.46670866e-01 -6.73146188e-01 -7.72385716e-01
3.36757803e-04 3.34326625e-01 -7.54665351e-03 -3.75703746e-03
-1.09826970e+00 6.62124276e-01 -2.88225114e-01 4.36360389e-01
-7.36283600e-01 -8.37331831e-01 -5.36073685e-01 -1.46616865e-02
4.38062429e-01 -7.66022563e-01 9.67929244e-01 -5.10263562e-01
-8.64754140e-01 9.17834520e-01 -3.22172076e-01 -2.05630898e-01
6.63635969e-01 4.38407958e-01 -9.09291804e-02 2.43149232e-02
3.41650337e-01 -2.88823187e-01 3.01808566e-01 -1.26438594e+00
-3.21998000e-01 -4.62321699e-01 -8.31595287e-02 2.74370816e-02
1.50131300e-01 2.36226588e-01 1.95443351e-02 -5.12929261e-01
-1.24928050e-01 -8.73906076e-01 -7.85737514e-01 2.16199830e-01
-1.00397512e-01 1.55405656e-01 5.73567629e-01 -5.63944221e-01
7.91539490e-01 -2.15642786e+00 1.43432975e-01 1.10877037e-01
2.80244112e-01 2.22298294e-01 1.34286538e-01 8.81441712e-01
1.85326681e-01 1.01829320e-01 -7.84195185e-01 1.03761656e-02
-4.71262932e-01 3.03777754e-01 -1.47790477e-01 9.97968733e-01
2.50704855e-01 4.25387084e-01 -1.18412101e+00 -5.65728366e-01
2.76856311e-02 7.51122832e-01 -3.90157312e-01 2.93377370e-01
1.03574462e-01 7.45522678e-01 -9.67879772e-01 9.47712362e-02
9.65625167e-01 -3.44763815e-01 1.18329503e-01 7.79296100e-01
-3.54093671e-01 -6.20406270e-01 -1.19802570e+00 4.05540466e-01
-2.32799649e-01 2.54145712e-01 6.44759893e-01 -1.45705771e+00
4.13682818e-01 6.33094370e-01 7.82903492e-01 -1.38134062e-01
3.52963567e-01 9.39798504e-02 1.35360092e-01 -7.47130752e-01
-1.01821478e-02 -7.95217216e-01 1.18166424e-01 5.24139524e-01
-5.04561424e-01 1.85739085e-01 1.47987202e-01 1.36334032e-01
1.02788007e+00 -4.88348603e-01 5.42005658e-01 -4.74457473e-01
4.02153611e-01 2.95110881e-01 6.04481518e-01 5.85175514e-01
-2.87954539e-01 2.39172265e-01 9.61056232e-01 -1.01750374e-01
-9.32211816e-01 -1.07556367e+00 -6.40012383e-01 4.21237975e-01
3.21841627e-01 4.83548075e-01 -7.55822599e-01 2.62357056e-01
2.33303547e-01 1.49626911e-01 -9.43468213e-01 -5.87042123e-02
-5.29852211e-01 -1.36893773e+00 2.36322754e-03 -3.29666659e-02
2.07898811e-01 -9.14009988e-01 -6.45780981e-01 3.14849615e-01
4.71877567e-02 -7.92589188e-01 -2.22639084e-01 1.03164457e-01
-8.49540532e-01 -1.40464306e+00 -1.37859535e+00 -6.96166277e-01
1.11394536e+00 1.17491558e-01 4.15698498e-01 2.98013568e-01
-5.33900619e-01 3.63184631e-01 -1.22564556e-02 -4.13025558e-01
-4.87380981e-01 -5.76151729e-01 2.77490705e-01 3.63751799e-01
9.71021876e-02 -8.75891149e-02 -7.69462168e-01 2.33677849e-01
-1.26983058e+00 -5.28700352e-01 2.85950840e-01 5.70862353e-01
6.02059662e-01 1.38832182e-01 6.92193925e-01 -8.00363839e-01
8.68135631e-01 -9.99274254e-01 -6.49136484e-01 2.92511910e-01
-2.50858366e-01 -8.54188669e-03 6.06746137e-01 -8.07103038e-01
-9.32689488e-01 -2.11489886e-01 3.94792080e-01 -1.56209171e-01
1.52911386e-02 5.66174150e-01 2.85778731e-01 1.16186969e-01
1.79551288e-01 3.26544642e-01 4.56134826e-01 -5.27742386e-01
-2.71432042e-01 7.25323558e-01 2.64040053e-01 -3.53990585e-01
4.77667511e-01 5.64123750e-01 4.58902478e-01 -1.12113822e+00
-4.59917516e-01 -5.23384750e-01 -3.41852486e-01 -1.99520916e-01
7.31063008e-01 -4.39383864e-01 -1.04495370e+00 7.17450142e-01
-1.09731686e+00 -4.61905956e-01 -2.45690629e-01 4.46976542e-01
-4.23680425e-01 1.16951764e-01 -7.39840329e-01 -1.51685464e+00
9.62654278e-02 -8.88911426e-01 6.64381504e-01 1.74220681e-01
9.42972079e-02 -1.57602954e+00 7.21396625e-01 -3.37425292e-01
6.58638656e-01 7.52423048e-01 9.80953336e-01 -3.14570278e-01
-7.22717419e-02 -4.41579103e-01 -3.42325158e-02 -1.39065936e-01
1.76055461e-01 9.27339196e-02 -3.39541286e-01 -3.48009259e-01
7.29717314e-01 3.84433478e-01 3.49542320e-01 1.05653512e+00
5.96012652e-01 -5.64732373e-01 -8.95322561e-01 3.66287947e-01
1.68941212e+00 4.18321818e-01 1.58723131e-01 2.74105812e-04
1.21048301e-01 9.54037547e-01 3.06810766e-01 6.40546560e-01
8.15599412e-02 3.73353094e-01 1.86011761e-01 -1.96811944e-01
9.18323815e-01 2.00768724e-01 -1.37582198e-01 3.37593853e-01
-1.75147243e-02 -4.16224867e-01 -9.66836870e-01 8.41676712e-01
-1.94612896e+00 -1.00589633e+00 -6.92254663e-01 2.57822585e+00
7.03011572e-01 -6.13560617e-01 5.03939629e-01 -9.91473440e-03
1.17344308e+00 -4.29010808e-01 -5.49821019e-01 -4.22456503e-01
-7.29579851e-02 -1.22536853e-01 7.41455615e-01 6.61474764e-01
-4.23092991e-01 1.00023225e-01 7.80550480e+00 1.15042366e-01
-1.12197936e+00 2.27468744e-01 7.25509107e-01 6.13264665e-02
1.70630231e-01 -1.76579002e-02 -4.40861017e-01 5.76253474e-01
8.82149577e-01 -5.68918169e-01 2.52939492e-01 2.61094362e-01
6.25284076e-01 -5.76859534e-01 -8.38996708e-01 4.67084944e-01
-3.54863405e-01 -5.90515792e-01 -4.49528545e-01 6.71623409e-01
9.16028321e-01 -4.45528179e-01 1.34648681e-01 -1.83083177e-01
3.14997464e-01 -8.83969128e-01 6.95724487e-02 6.30139410e-01
4.87427324e-01 -4.53024447e-01 7.42470443e-01 7.52035081e-01
-8.99955988e-01 9.48706716e-02 -3.98728132e-01 -5.01359820e-01
6.12957239e-01 5.77135384e-01 -3.18139791e-01 2.88046859e-02
3.60743463e-01 2.76935726e-01 1.11264102e-01 1.50797737e+00
2.61894703e-01 3.29138577e-01 -5.50266743e-01 -1.41783908e-01
9.21730474e-02 -6.31965220e-01 4.79166776e-01 6.56971276e-01
3.31116199e-01 4.47691500e-01 -2.03260496e-01 9.92881238e-01
3.49670857e-01 1.25685051e-01 -7.96406686e-01 -7.48387305e-03
1.69510663e-01 7.73666978e-01 -8.52224767e-01 -3.69832963e-01
-2.50284106e-01 8.85561585e-01 3.83680761e-02 8.27239335e-01
-6.73507273e-01 -1.96328372e-01 7.22784579e-01 5.41239798e-01
1.52887374e-01 7.70416930e-02 1.17294811e-01 -1.14835644e+00
-2.01144964e-01 -8.23205039e-02 4.73601639e-01 -3.64984214e-01
-1.24806798e+00 2.50828058e-01 4.62911248e-01 -8.89346838e-01
-3.44266623e-01 -5.68781912e-01 -8.10600221e-01 1.15668309e+00
-1.01708913e+00 -1.90365866e-01 2.28498474e-01 6.26530051e-01
-1.51873501e-02 5.29176831e-01 4.59451646e-01 8.28471631e-02
-9.36011851e-01 -2.33786732e-01 1.04167593e+00 -1.05923392e-01
2.49532565e-01 -8.65166128e-01 -2.03300059e-01 7.55535424e-01
-1.14318728e+00 9.16133940e-01 1.03990257e+00 -9.04941678e-01
-1.29243553e+00 -9.02209401e-01 1.10552585e+00 -1.05933351e-02
9.53348219e-01 -3.80271912e-01 -9.58313525e-01 6.83615088e-01
2.29614899e-02 2.58235008e-01 2.34769851e-01 -4.53464001e-01
4.82511640e-01 2.51049042e-01 -1.60792506e+00 4.92000788e-01
5.27736127e-01 1.61215756e-02 -1.04125723e-01 4.54174906e-01
2.85067201e-01 8.95462111e-02 -7.45025158e-01 2.43669435e-01
3.29812229e-01 -4.85579848e-01 9.96053517e-01 -7.95140803e-01
7.22706988e-02 -1.22087881e-01 2.10528299e-01 -1.29731739e+00
-2.13964671e-01 -5.41507900e-01 3.27417254e-01 1.03068614e+00
1.61046848e-01 -1.15395594e+00 2.91941851e-01 4.85367775e-01
7.45144546e-01 -9.12857890e-01 -1.07288194e+00 -1.00199854e+00
3.57796073e-01 3.16622645e-01 2.35443145e-01 9.55385089e-01
3.27503048e-02 5.94455339e-02 -4.33692873e-01 3.65079612e-01
1.06302261e+00 -1.59120753e-01 1.30592942e-01 -1.35607302e+00
-1.76971525e-01 -2.40636826e-01 -1.40706003e-01 -4.13521767e-01
-2.54030470e-02 -2.51331717e-01 3.29548687e-01 -1.56611991e+00
6.54566526e-01 -5.27509868e-01 -8.35384727e-02 -2.17665032e-01
-3.71652424e-01 -4.61418368e-02 -2.59825259e-01 2.92855352e-01
-4.58316579e-02 1.59137577e-01 9.11619902e-01 1.31629482e-01
-4.41741467e-01 2.30449393e-01 -3.56189489e-01 3.91248196e-01
6.16725862e-01 -8.94176841e-01 -3.60283554e-01 -1.26656860e-01
2.44729221e-01 5.08942664e-01 5.44672132e-01 -2.89602399e-01
5.42417616e-02 -8.75697911e-01 -1.76349785e-02 -3.59121680e-01
1.72644302e-01 -7.47351766e-01 7.40817070e-01 1.14440680e+00
-5.27285263e-02 5.59747890e-02 -1.31542563e-01 6.87703609e-01
2.56717652e-01 -6.37464166e-01 9.62467611e-01 -2.00178698e-02
9.73770469e-02 4.82013553e-01 -9.84705865e-01 2.17170566e-01
1.51764464e+00 -5.39643839e-02 -3.93956184e-01 -3.88733178e-01
-8.46130371e-01 3.80822033e-01 6.75290227e-01 -2.76139438e-01
4.63208288e-01 -9.26142752e-01 -6.18572056e-01 2.10085213e-02
-2.05092564e-01 -6.67062327e-02 4.78150427e-01 1.29039299e+00
-6.11884892e-01 3.00332069e-01 -1.29966453e-01 -4.97149378e-01
-7.75319993e-01 1.06383026e+00 7.59238064e-01 -1.66082248e-01
-2.22754896e-01 3.28846514e-01 7.86753356e-01 -9.10833776e-02
-7.10824803e-02 8.56808349e-02 -1.89642057e-01 1.01687998e-01
6.71157420e-01 5.41720748e-01 -5.62513411e-01 -7.67966151e-01
-5.50701618e-01 6.79275572e-01 2.04928234e-01 -2.23470837e-01
1.49642074e+00 -2.91950136e-01 -4.43019539e-01 7.29556143e-01
1.29149568e+00 -4.09705341e-01 -9.90755916e-01 -3.86062682e-01
-2.57994860e-01 -1.15538903e-01 -2.91594356e-01 -2.22655848e-01
-1.00461400e+00 5.49337387e-01 4.35387433e-01 7.78095365e-01
9.04667974e-01 1.35259166e-01 4.17913944e-01 2.16064486e-03
1.45110026e-01 -4.75959003e-01 -4.74958092e-01 1.62259847e-01
6.62176013e-01 -1.02060926e+00 -1.50492698e-01 -5.56246161e-01
-2.40064502e-01 4.36893702e-01 1.15638360e-01 -5.69057822e-01
9.60243046e-01 7.34180629e-01 4.56015356e-02 -1.02907822e-01
-7.22607374e-01 -1.41585857e-01 -3.50102127e-01 9.46578383e-01
3.00158739e-01 -1.66312959e-02 -9.94972110e-01 1.88013613e-01
4.45859641e-01 1.77062944e-01 7.22608566e-01 1.14729846e+00
-2.63042569e-01 -8.31902087e-01 -8.18717062e-01 4.67356354e-01
-7.72103667e-01 1.01970971e-01 -3.88822317e-01 7.36532748e-01
-1.76067710e-01 9.80688512e-01 1.96491227e-01 5.14955997e-01
1.47094041e-01 -3.00446656e-02 4.69360918e-01 -4.86899912e-01
-1.74179271e-01 1.06929332e-01 -6.35379016e-01 1.12160198e-01
-3.77736062e-01 -8.74359012e-01 -1.23007667e+00 -7.28991270e-01
-3.62137377e-01 4.40500289e-01 4.21198279e-01 1.08958602e+00
9.12438929e-02 3.12701046e-01 8.69815230e-01 -5.02437830e-01
-4.53695565e-01 -8.95006537e-01 -1.17603683e+00 1.57525644e-01
9.00507033e-01 -4.67382818e-01 -9.46441412e-01 3.28790098e-02] | [5.966383457183838, 4.322572231292725] |
7ca197a3-e486-4b9d-aa42-8244c438bcca | adaptive-unsupervised-self-training-for | null | null | https://aclanthology.org/2022.coling-1.632 | https://aclanthology.org/2022.coling-1.632.pdf | Adaptive Unsupervised Self-training for Disfluency Detection | Supervised methods have achieved remarkable results in disfluency detection. However, in real-world scenarios, human-annotated data is difficult to obtain. Recent works try to handle disfluency detection with unsupervised self-training, which can exploit existing large-scale unlabeled data efficiently. However, their self-training-based methods suffer from the problems of selection bias and error accumulation. To tackle these problems, we propose an adaptive unsupervised self-training method for disfluency detection. Specifically, we re-weight the importance of each training example according to its grammatical feature and prediction confidence. Experiments on the Switchboard dataset show that our method improves 2.3 points over the current SOTA unsupervised method. Moreover, our method is competitive with the SOTA supervised method. | ['Wanxiang Che', 'Shaolei Wang', 'YiXuan Wang', 'Zhongyuan Wang'] | null | null | null | null | coling-2022-10 | ['selection-bias'] | ['natural-language-processing'] | [-3.41637507e-02 -9.65587571e-02 -5.54294467e-01 -5.20617127e-01
-6.91256583e-01 -5.36175191e-01 3.11128907e-02 2.26537481e-01
-4.40366179e-01 8.65067601e-01 1.96974218e-01 -2.17866778e-01
3.38179857e-01 -4.82865065e-01 -2.92891502e-01 -5.13274550e-01
2.76396573e-01 6.05608582e-01 2.58455634e-01 -1.57204568e-01
2.77520627e-01 -7.11959973e-02 -1.59966135e+00 2.30840936e-01
1.51322520e+00 7.50964284e-01 3.44222754e-01 3.25236060e-02
-1.84079111e-02 8.98630142e-01 -6.30592465e-01 1.63130626e-01
-1.90782666e-01 -6.03321493e-01 -8.17020953e-01 1.50273502e-01
2.99237788e-01 -3.03580705e-02 -1.11557171e-01 1.26761615e+00
3.36925566e-01 1.84891552e-01 3.14785600e-01 -7.54620790e-01
-4.42967862e-01 8.93978536e-01 -1.14103369e-01 7.09015667e-01
2.35410810e-01 -3.52791175e-02 9.10095215e-01 -8.39970589e-01
6.89164519e-01 7.62563229e-01 3.78182918e-01 6.68685079e-01
-1.00195932e+00 -7.26223528e-01 2.00445116e-01 2.71969974e-01
-1.35960782e+00 -4.66546834e-01 9.68588293e-01 -4.24478859e-01
1.34449363e+00 1.11640409e-01 3.83938789e-01 8.25550020e-01
-2.42327929e-01 8.29145730e-01 1.24704063e+00 -7.75996745e-01
1.80084184e-01 2.80534863e-01 6.50265038e-01 5.22690773e-01
6.78353012e-02 1.72615334e-01 -5.05526185e-01 3.44782807e-02
3.77657652e-01 -2.28614777e-01 -2.80843079e-01 4.96949144e-02
-8.63644958e-01 6.43381596e-01 6.73675388e-02 1.02695954e+00
-6.26108125e-02 -4.91525531e-01 4.27125514e-01 5.48634887e-01
1.03111410e+00 5.48713982e-01 -7.21238256e-01 -3.98198277e-01
-1.07994795e+00 -4.86359961e-04 4.81871754e-01 9.12051141e-01
6.43632591e-01 1.21604748e-01 -1.22052412e-02 9.84710872e-01
5.56925982e-02 3.04529160e-01 9.54373837e-01 -3.56390059e-01
6.83129728e-01 8.58787417e-01 -2.13286318e-02 -7.35800207e-01
-6.18264914e-01 -3.94960791e-01 -5.94989717e-01 -2.04823241e-01
3.29630524e-01 -6.98317140e-02 -9.69222605e-01 1.76356053e+00
7.29978085e-02 -9.04117376e-02 -2.97106579e-02 8.05911064e-01
6.32829607e-01 4.75967407e-01 3.04176211e-01 -9.78970408e-01
8.32016170e-01 -1.02619934e+00 -1.05008221e+00 -3.39065731e-01
1.11685872e+00 -5.69064379e-01 1.38157797e+00 6.00433052e-01
-8.96805406e-01 -6.16070271e-01 -8.52615654e-01 2.87935466e-01
-2.53953524e-02 1.84954077e-01 6.18053734e-01 8.15229535e-01
-5.93117476e-01 6.07169032e-01 -8.34433794e-01 -3.19536477e-01
2.04644889e-01 3.87099475e-01 -2.35792413e-01 2.24982083e-01
-1.47250211e+00 9.37142313e-01 9.70956802e-01 -5.85491881e-02
-4.87791836e-01 -3.58754903e-01 -8.77722979e-01 6.33190572e-02
5.56826711e-01 1.15102224e-01 1.01252592e+00 -1.21286714e+00
-1.46637702e+00 8.45633149e-01 -3.18919837e-01 -3.96036804e-01
1.17405534e-01 -3.47366005e-01 -9.31894302e-01 -1.69605121e-01
8.76711830e-02 2.33893827e-01 6.49118125e-01 -7.74541497e-01
-6.48661017e-01 -3.45865637e-01 -1.95547193e-01 1.60069883e-01
-9.66774762e-01 3.07003826e-01 -3.04782361e-01 -9.45492804e-01
2.67670959e-01 -7.70992637e-01 -9.62257534e-02 -1.03630900e+00
-4.48665708e-01 -7.56077945e-01 7.45045364e-01 -4.93995279e-01
2.01363134e+00 -2.42133451e+00 6.72603846e-02 1.01921946e-01
2.95325279e-01 7.29587793e-01 2.45483920e-01 -3.58363874e-02
-1.43576145e-01 1.92987546e-01 -2.71555454e-01 -2.82633543e-01
-3.89288753e-01 1.19379394e-01 -2.24875063e-02 2.18683943e-01
1.44511819e-01 5.28858304e-01 -1.17973757e+00 -7.66917944e-01
1.72127694e-01 -1.90293893e-01 -7.17817783e-01 4.31269556e-01
-3.29078972e-01 7.11094439e-01 -4.09170568e-01 6.65844083e-01
5.27691543e-01 -1.69470400e-01 6.61653280e-01 1.40993610e-01
-2.73154169e-01 7.94209659e-01 -7.85677910e-01 1.56894672e+00
-3.55137408e-01 3.90964717e-01 -6.04498446e-01 -1.20929372e+00
9.74320650e-01 4.73664910e-01 1.75968975e-01 -7.93908656e-01
1.60100058e-01 5.77526152e-01 1.25436544e-01 -7.15541720e-01
4.51436430e-01 -4.02971864e-01 -1.40470713e-01 3.46114725e-01
2.81419039e-01 1.56658083e-01 5.76057315e-01 5.32771014e-02
1.04869115e+00 9.96556431e-02 3.86282176e-01 -5.02118468e-01
5.01988411e-01 1.91927791e-01 9.56763923e-01 4.84325171e-01
-3.31327796e-01 4.25175518e-01 3.10141206e-01 -5.93563437e-01
-6.34318709e-01 -7.52622783e-01 -2.25048736e-01 9.12259698e-01
7.68830180e-02 -9.00405109e-01 -8.98335457e-01 -1.26642585e+00
-3.18267316e-01 8.38984132e-01 -3.35263580e-01 -3.02025676e-01
-9.70769286e-01 -8.88262272e-01 3.78346592e-01 5.85948110e-01
3.53178233e-01 -1.51661062e+00 -4.56738919e-01 4.73823637e-01
-5.51639616e-01 -9.87086713e-01 -5.14685631e-01 4.11558449e-01
-1.10910416e+00 -9.34866786e-01 -7.23953396e-02 -1.11265838e+00
7.80070543e-01 7.43229464e-02 1.32334423e+00 3.16072255e-01
8.62845108e-02 -3.93678874e-01 -7.15957701e-01 7.73454783e-03
-6.56585813e-01 3.29196751e-01 3.75441909e-01 -2.72654474e-01
8.07116687e-01 -6.33603811e-01 -1.69773072e-01 4.56503242e-01
-7.30708122e-01 -2.47309446e-01 2.41943285e-01 8.41016471e-01
7.34091282e-01 2.48496607e-01 8.62357020e-01 -1.31486273e+00
5.72466552e-01 -4.56792265e-01 -4.89903331e-01 2.96277195e-01
-1.04485583e+00 1.60822675e-01 7.91725039e-01 -4.64798480e-01
-1.15021956e+00 2.38155767e-01 -8.64812434e-02 -3.43420208e-01
-3.91599715e-01 5.80261946e-01 -3.14664304e-01 2.69844353e-01
6.83639467e-01 5.01663499e-02 -4.12324786e-01 -6.94909632e-01
7.61453807e-02 9.54794168e-01 3.61315101e-01 -3.13063771e-01
4.52058285e-01 -5.14905304e-02 -7.86501586e-01 -5.01588404e-01
-1.20202827e+00 -5.65737128e-01 -8.29349160e-01 -5.72956391e-02
5.98000348e-01 -7.97659636e-01 -2.28392616e-01 5.19264698e-01
-1.05984116e+00 -2.35717371e-01 -1.53900802e-01 6.03099167e-01
-4.54825699e-01 4.95642066e-01 -5.50400138e-01 -7.17828453e-01
-3.00090611e-01 -8.77439499e-01 4.40692693e-01 1.16771452e-01
-4.02047276e-01 -9.50025260e-01 3.40149134e-01 3.51325780e-01
3.03257287e-01 -1.66041136e-01 8.31173360e-01 -8.19983065e-01
-1.18847214e-01 -1.33849695e-01 2.78617054e-01 6.29214525e-01
4.48612869e-01 -3.95706385e-01 -7.80901849e-01 -8.61170515e-02
3.21835399e-01 -6.26632690e-01 9.11702931e-01 1.80734307e-01
1.16412830e+00 -4.77443822e-02 -2.96583235e-01 2.33335420e-01
1.15414870e+00 2.64198363e-01 4.59446907e-01 2.86476076e-01
7.34863698e-01 2.98648506e-01 8.99987936e-01 3.03636193e-01
2.24979371e-01 8.28782439e-01 8.40835795e-02 1.58011183e-01
-7.85669163e-02 -3.20854157e-01 4.75518316e-01 1.22068942e+00
-1.03825979e-01 -3.90590280e-01 -8.94347072e-01 6.58374012e-01
-1.89364707e+00 -8.84268045e-01 -1.63137138e-01 2.23917913e+00
1.18979084e+00 6.71806931e-01 2.03455642e-01 5.53501010e-01
9.11463261e-01 -1.52617425e-01 -3.42997104e-01 -2.02122658e-01
-3.79744381e-01 1.64146021e-01 4.13204059e-02 3.80416512e-01
-1.40481710e+00 1.61395967e+00 6.55465603e+00 9.34846878e-01
-9.46303129e-01 5.18122673e-01 5.32210052e-01 -4.85127904e-02
-1.89927205e-01 4.90817167e-02 -9.86012816e-01 6.94343448e-01
1.01761818e+00 -5.48587032e-02 3.10172379e-01 6.85278833e-01
2.16064870e-01 -2.56692022e-01 -8.01605642e-01 8.32908750e-01
3.26523364e-01 -9.09564435e-01 -2.62535036e-01 -2.93704987e-01
6.59347832e-01 -6.85669184e-02 -2.65447259e-01 4.85811293e-01
7.82089755e-02 -7.28549242e-01 6.93932891e-01 8.84688720e-02
7.65115917e-01 -8.17524374e-01 6.99736476e-01 5.97495496e-01
-8.99898469e-01 -4.19458859e-02 -2.84976065e-01 -3.73611242e-01
2.67403990e-01 1.05896223e+00 -5.29480040e-01 3.72852474e-01
6.90979958e-01 8.97378922e-01 -4.29525018e-01 8.01975965e-01
-6.04040444e-01 1.27420759e+00 -3.75556290e-01 7.50599653e-02
2.31835395e-02 -2.88161617e-02 5.10635853e-01 1.11266267e+00
-6.86872005e-03 1.56232879e-01 3.84657621e-01 7.28123724e-01
-4.37867232e-02 5.14487863e-01 -4.04189765e-01 -1.91404372e-01
3.68057758e-01 1.03332853e+00 -8.77206385e-01 -5.56080043e-01
-2.49536708e-01 1.05643094e+00 8.37155640e-01 -1.52539045e-01
-7.05435336e-01 -3.50696355e-01 2.74114072e-01 -3.18388157e-02
2.28516668e-01 -1.77796379e-01 -1.52705163e-01 -1.42284501e+00
1.71590507e-01 -8.15009058e-01 6.47540629e-01 -2.09984601e-01
-1.15076947e+00 1.06673086e+00 -1.26408443e-01 -1.45610940e+00
-4.06192988e-01 -2.99108535e-01 -4.77504581e-01 2.72879183e-01
-1.25428140e+00 -6.49730563e-01 8.96185189e-02 4.38890517e-01
7.91603386e-01 -1.72296628e-01 8.87689650e-01 5.83030224e-01
-8.02041113e-01 7.38416195e-01 1.39818611e-02 5.03951497e-02
8.63159001e-01 -1.14808309e+00 2.99879670e-01 9.13897812e-01
4.88289863e-01 5.70176780e-01 5.73014379e-01 -9.09689665e-01
-6.73693955e-01 -1.00487804e+00 1.57386708e+00 -4.86710489e-01
4.65264261e-01 -2.74921775e-01 -1.04095912e+00 6.14855945e-01
-8.99175927e-02 2.20260471e-01 6.99542105e-01 6.07911408e-01
-3.40041280e-01 1.47235617e-01 -1.09442759e+00 1.11458421e-01
1.41162193e+00 -3.31255227e-01 -8.67614031e-01 5.30425310e-01
6.56639874e-01 -4.27066445e-01 -6.07495427e-01 5.36538184e-01
-5.06458338e-03 -1.10500360e+00 3.30799520e-01 -7.02566981e-01
2.53422469e-01 -2.07867220e-01 1.34197995e-01 -1.52382576e+00
-3.94359499e-01 -4.90648746e-01 -2.11832762e-01 1.37758994e+00
5.83954751e-01 -4.50218141e-01 5.19658446e-01 1.83526203e-01
-3.25957566e-01 -3.55587393e-01 -9.20204043e-01 -1.11056125e+00
6.24322034e-02 -4.99774694e-01 3.46348137e-01 1.13414419e+00
8.60057712e-01 4.36209053e-01 -6.03160918e-01 -9.49342623e-02
3.29761624e-01 2.71793813e-01 2.25676987e-02 -1.31318474e+00
-1.52797669e-01 -1.28992140e-01 -3.93995851e-01 -8.67540061e-01
5.75868607e-01 -9.40487862e-01 2.93787807e-01 -8.20455551e-01
2.25760520e-01 -6.99652076e-01 -7.11823046e-01 6.56026423e-01
-6.77600622e-01 4.10449237e-01 -2.45788097e-02 4.31549877e-01
-9.65320170e-01 3.07274252e-01 7.80282974e-01 4.83317114e-02
-6.45505786e-01 -8.74598175e-02 -3.92128795e-01 8.32073152e-01
1.22872603e+00 -7.67089605e-01 -3.43994617e-01 -5.21410882e-01
8.65619034e-02 -6.09290749e-02 -2.35815316e-01 -1.12471211e+00
-8.75352602e-03 -9.31559950e-02 -1.67149752e-01 -5.12098610e-01
-3.95967394e-01 -4.46439803e-01 -2.68215388e-01 3.41149628e-01
-2.39955157e-01 -4.32825200e-02 1.65518045e-01 3.01575661e-01
-5.65890253e-01 -2.91337013e-01 7.34056115e-01 -1.48845226e-01
-7.69324839e-01 5.75922662e-03 -6.37779117e-01 4.78078783e-01
7.24440515e-01 1.42111421e-01 -3.29277734e-03 -8.07025190e-03
-8.88958633e-01 1.16382107e-01 3.37295175e-01 4.95254904e-01
2.90811390e-01 -1.25235486e+00 -5.70291877e-01 4.87080753e-01
3.69163752e-01 -2.24535108e-01 9.53774229e-02 8.38124573e-01
4.52628545e-02 5.37704051e-01 -3.65091190e-02 -3.20218652e-01
-1.12753987e+00 7.67769456e-01 -3.43397912e-03 -5.59514940e-01
-6.18915677e-01 7.44245827e-01 -5.56999668e-02 -2.08536968e-01
1.10133961e-01 -3.33906680e-01 -4.78049129e-01 -8.01952034e-02
5.19650161e-01 8.90131146e-02 4.25568134e-01 -6.03477180e-01
-5.16954243e-01 3.76054049e-01 -2.97638386e-01 1.72637463e-01
1.16557872e+00 -3.17878127e-02 -9.32698920e-02 5.44336319e-01
8.00748765e-01 1.67715818e-01 -7.27163613e-01 -3.67014498e-01
5.43654025e-01 -4.50411320e-01 1.35462224e-01 -7.88173079e-01
-1.23422766e+00 5.56068122e-01 5.06465554e-01 1.59320042e-01
1.28582561e+00 2.12962016e-01 7.56971002e-01 3.52397978e-01
4.82073694e-01 -1.34512806e+00 -1.77032396e-01 5.83655596e-01
3.32071036e-01 -1.39862561e+00 -4.61967260e-01 -7.69723654e-01
-5.93706489e-01 8.76491308e-01 9.07412350e-01 -1.15420699e-01
7.18116641e-01 1.81281790e-01 2.36073971e-01 -9.15496051e-02
-6.64393663e-01 -2.30485126e-01 1.05548710e-01 2.35148072e-01
7.27957010e-01 1.64541081e-01 -9.53624785e-01 8.55180800e-01
-9.81390383e-03 1.52037904e-01 2.20000580e-01 9.68047380e-01
-5.24509668e-01 -1.43591750e+00 -1.06218845e-01 4.26821917e-01
-4.34348017e-01 -4.63820159e-01 -2.52291530e-01 4.30033505e-01
2.79852122e-01 1.23810208e+00 5.88710345e-02 -5.32211423e-01
3.80394936e-01 4.01379019e-01 5.58045268e-01 -9.50905859e-01
-7.69883692e-01 3.16204280e-01 1.65701360e-01 -4.37545091e-01
-6.54376447e-01 -6.91372931e-01 -1.32084262e+00 2.58651644e-01
-8.21082175e-01 5.19741356e-01 3.28077823e-02 1.29876018e+00
2.44281322e-01 1.72410220e-01 8.27790499e-01 -3.03599328e-01
-3.71742308e-01 -1.24485350e+00 -6.56379402e-01 6.53611541e-01
3.39079648e-02 -8.30507636e-01 -6.37799025e-01 1.68252096e-01] | [9.65064811706543, 4.1355977058410645] |
52b9c429-c9f5-4f75-ae71-97c2b8cefe8f | multiview-learning-of-weighted-majority-vote | 1805.10212 | null | http://arxiv.org/abs/1805.10212v1 | http://arxiv.org/pdf/1805.10212v1.pdf | Multiview Learning of Weighted Majority Vote by Bregman Divergence Minimization | We tackle the issue of classifier combinations when observations have
multiple views. Our method jointly learns view-specific weighted majority vote
classifiers (i.e. for each view) over a set of base voters, and a second
weighted majority vote classifier over the set of these view-specific weighted
majority vote classifiers. We show that the empirical risk minimization of the
final majority vote given a multiview training set can be cast as the
minimization of Bregman divergences. This allows us to derive a parallel-update
optimization algorithm for learning our multiview model. We empirically study
our algorithm with a particular focus on the impact of the training set size on
the multiview learning results. The experiments show that our approach is able
to overcome the lack of labeled information. | ['Massih-Reza Amini', 'Emilie Morvant', 'Anil Goyal'] | 2018-05-25 | null | null | null | null | ['multiview-learning', 'multilingual-text-classification'] | ['computer-vision', 'miscellaneous'] | [ 6.16347566e-02 3.34871083e-01 -7.50601649e-01 -8.44177425e-01
-1.26461899e+00 -7.02222049e-01 9.47867692e-01 5.34327440e-02
-2.80863643e-01 5.94646335e-01 1.92263767e-01 -6.13698959e-02
6.75342008e-02 -5.45464635e-01 -5.14568985e-01 -7.49177814e-01
5.43629110e-01 5.58918953e-01 3.26455981e-02 1.64840221e-01
2.62677282e-01 5.85997365e-02 -1.79679859e+00 5.91347814e-01
5.32695532e-01 1.02514076e+00 -1.58459082e-01 6.50955856e-01
4.16445851e-01 7.63913214e-01 -2.59499997e-01 -8.88891995e-01
5.08057594e-01 -1.52069986e-01 -6.63063884e-01 5.08560538e-01
1.09552693e+00 -2.51761734e-01 2.94905394e-01 1.18437290e+00
2.83060789e-01 -3.71949449e-02 1.08782983e+00 -1.42268229e+00
-1.25625238e-01 6.05240881e-01 -7.25263000e-01 -7.93572068e-02
1.96598351e-01 -1.84006691e-01 1.45483780e+00 -1.28942692e+00
1.18187726e+00 1.23922563e+00 5.35527289e-01 4.16382402e-01
-1.36318362e+00 -3.95632565e-01 5.27977824e-01 2.47171208e-01
-1.12437308e+00 -4.28727120e-01 6.61254764e-01 -6.30743802e-01
2.01164618e-01 2.91390538e-01 6.09606028e-01 1.05433202e+00
4.18116510e-01 6.40438557e-01 1.52145934e+00 -5.07197201e-01
5.84686577e-01 6.82851076e-01 6.09183431e-01 9.00180817e-01
3.15272540e-01 2.36548875e-02 -5.81219494e-01 -6.41873002e-01
-4.25717840e-03 2.01297641e-01 -1.35296527e-02 -9.80776846e-01
-8.12645555e-01 1.28723621e+00 -3.24701406e-02 -1.98787764e-01
-1.23171724e-01 -1.00616932e-01 3.45174104e-01 5.78136086e-01
7.87064433e-01 -1.13537945e-01 -4.71177518e-01 7.83238173e-01
-7.97856092e-01 2.62471944e-01 9.82221782e-01 7.29680419e-01
9.75869298e-01 -2.49821335e-01 4.01524417e-02 8.58653903e-01
4.83351350e-01 8.27342868e-01 -1.01071589e-01 -1.33227897e+00
4.22446698e-01 4.93047863e-01 -1.69221461e-02 -7.63542354e-01
-1.76993340e-01 -7.41913170e-02 -6.35821283e-01 6.94370210e-01
4.05934811e-01 -3.75027597e-01 -5.30187726e-01 1.81947219e+00
7.03451991e-01 -4.70643252e-01 1.04046367e-01 5.99185228e-01
6.12094820e-01 2.47087166e-01 -1.78654656e-01 -6.55784369e-01
1.25598085e+00 -8.21317077e-01 -5.62727869e-01 -1.21745959e-01
5.83388269e-01 -4.69596565e-01 3.14475924e-01 4.08851504e-01
-9.93824422e-01 -2.62279332e-01 -1.22398531e+00 2.94566423e-01
-5.33711277e-02 1.88404784e-01 3.84993523e-01 7.27415144e-01
-7.93488979e-01 4.00009066e-01 -3.02320778e-01 -4.95712981e-02
3.85023147e-01 3.12897980e-01 -7.63973117e-01 -2.10122302e-01
-6.60124838e-01 1.25843883e+00 -2.63194996e-03 -2.83783019e-01
-1.02427840e+00 -6.26044691e-01 -8.45223248e-01 -2.75319278e-01
5.46378851e-01 -1.01768148e+00 1.27994525e+00 -1.41980755e+00
-1.11698258e+00 1.33300173e+00 -4.65674490e-01 -1.47744581e-01
7.69401014e-01 1.78539246e-01 1.33745288e-02 -1.55913994e-01
2.19779760e-01 2.30431035e-01 1.11571240e+00 -1.79658437e+00
-1.01729381e+00 -8.54663789e-01 1.75937608e-01 3.64861399e-01
-8.85281339e-03 -1.25075772e-01 -1.50823399e-01 -2.75711358e-01
6.54643700e-02 -1.17907226e+00 -3.79437715e-01 -5.16373478e-02
-2.50230610e-01 -1.99336246e-01 5.61899781e-01 -3.34263891e-01
9.35784221e-01 -1.86822629e+00 4.35478657e-01 4.19271857e-01
2.74343282e-01 -2.73988903e-01 2.15807825e-01 3.57940763e-01
-3.50167118e-02 5.31497709e-02 -1.86340541e-01 -8.21788073e-01
-2.90502667e-01 1.27607733e-01 -3.46402049e-01 8.24846089e-01
-4.64086771e-01 3.15136373e-01 -6.25295043e-01 -6.15098834e-01
1.75581232e-01 6.80625513e-02 -5.35936296e-01 1.24234885e-01
-1.05590254e-01 7.00722784e-02 -3.04017931e-01 3.99670273e-01
6.77857697e-01 -3.47364455e-01 8.63716781e-01 -2.10956395e-01
1.27648711e-01 -3.74152690e-01 -1.51628208e+00 1.30126655e+00
-5.28004527e-01 3.85844052e-01 3.95136058e-01 -7.18165934e-01
5.57851732e-01 3.08666646e-01 2.39572242e-01 -5.12184799e-02
3.01440358e-01 2.18832552e-01 -4.02648985e-01 -2.26538405e-01
1.97337583e-01 -6.03937328e-01 -1.45919070e-01 7.02549875e-01
4.57277298e-01 -9.35039669e-02 -2.73913234e-01 2.66881615e-01
4.76907581e-01 -2.79947460e-01 8.94676447e-01 -4.79359806e-01
5.17579854e-01 -1.89287797e-01 6.73122108e-01 7.65887499e-01
6.38597980e-02 6.98907495e-01 5.34694254e-01 -7.06507981e-01
-1.11741626e+00 -1.03273642e+00 -1.96778283e-01 1.05437529e+00
-7.67763704e-03 -4.63199288e-01 -6.03076041e-01 -1.34463072e+00
9.68444347e-02 6.42417431e-01 -1.01706028e+00 1.39973402e-01
-2.15646550e-01 -4.76029158e-01 -2.97994018e-01 3.22809279e-01
6.43134909e-03 -4.46501732e-01 -4.71974760e-01 -3.14135581e-01
-8.35307837e-02 -7.58920848e-01 -4.32239294e-01 2.33740196e-01
-9.67754006e-01 -1.52194071e+00 -4.84322101e-01 -5.06309152e-01
8.31071794e-01 3.32500726e-01 1.16353881e+00 -1.00746349e-01
1.76062495e-01 9.05296385e-01 -4.07782607e-02 -5.59631169e-01
-6.14151180e-01 1.13300076e-02 2.37757698e-01 4.09621447e-01
9.45757702e-02 -3.53466660e-01 -2.70118952e-01 3.14999312e-01
-5.19566059e-01 -6.17557429e-02 -1.86326057e-02 1.05146146e+00
8.00887942e-01 -6.88058019e-01 5.57425678e-01 -1.61546779e+00
3.20793390e-01 -5.09552181e-01 -9.70715344e-01 7.99565375e-01
-7.92620242e-01 2.12638125e-01 2.71468431e-01 -3.38823408e-01
-1.20148313e+00 2.56678283e-01 3.48525882e-01 -4.93487984e-01
1.53715402e-01 3.85714561e-01 -1.04975998e-01 -4.22113779e-04
6.30530357e-01 -4.06842172e-01 9.20861587e-02 -8.05065557e-02
7.53605902e-01 4.29757029e-01 6.04463667e-02 -5.03287613e-01
5.28671980e-01 8.46599221e-01 1.92032307e-01 -5.48318565e-01
-1.32392418e+00 -4.38083321e-01 -6.75836325e-01 -6.28187358e-01
1.10518885e+00 -1.12918329e+00 -6.77173018e-01 1.10806622e-01
-1.03908432e+00 1.49534270e-01 -4.83995259e-01 5.49913645e-01
-6.78191602e-01 4.29556847e-01 -1.85708046e-01 -9.86279845e-01
-1.12368554e-01 -1.16513979e+00 8.38165402e-01 -1.15827829e-01
-4.07849431e-01 -1.27716732e+00 4.90023583e-01 7.33664751e-01
-2.05699265e-01 2.12623700e-01 7.91075468e-01 -8.26312184e-01
-5.12531996e-01 -2.22654760e-01 2.23766401e-01 5.14885664e-01
-2.71638155e-01 1.13074966e-01 -1.38106048e+00 -3.69564682e-01
4.16645676e-01 -6.06426477e-01 1.13554370e+00 5.51267684e-01
7.65712738e-01 -2.26649284e-01 -3.03540617e-01 3.70287418e-01
1.61378288e+00 -2.09944040e-01 3.10602970e-02 -9.57015827e-02
5.86039245e-01 7.84696996e-01 6.38859510e-01 5.86627901e-01
5.54391623e-01 5.35195470e-01 5.83526313e-01 2.81326622e-01
2.45327756e-01 -5.91688864e-02 2.14601055e-01 7.66276062e-01
-1.45473272e-01 -1.30287901e-01 -5.83075404e-01 4.24081057e-01
-2.00978994e+00 -1.20870543e+00 -7.65369758e-02 2.46061563e+00
3.89282554e-01 -2.55868465e-01 8.11145082e-02 -7.79921040e-02
7.57710159e-01 4.67429370e-01 -3.50169480e-01 -4.91403043e-01
-1.30554408e-01 -4.88364287e-02 5.42253733e-01 1.06725240e+00
-1.30708802e+00 3.20338815e-01 7.11139917e+00 5.74831307e-01
-6.28912807e-01 3.56029898e-01 6.73936903e-01 -2.31440738e-01
-7.30540514e-01 2.77995616e-01 -1.06744266e+00 1.78824499e-01
5.40718853e-01 2.36058310e-02 9.60595608e-02 1.00077760e+00
-4.21109349e-01 -3.66210997e-01 -1.41959286e+00 1.00480485e+00
5.31013846e-01 -1.51845503e+00 -2.16305163e-02 3.67770433e-01
1.39938986e+00 9.23712477e-02 1.47566333e-01 4.33959924e-02
7.82485783e-01 -6.80070102e-01 8.32466364e-01 4.43730652e-01
6.32667720e-01 -7.40587831e-01 6.36675715e-01 6.16643965e-01
-9.24623430e-01 -4.31805432e-01 -1.68312073e-01 5.17240688e-02
-2.47725565e-02 7.39085734e-01 -4.18243468e-01 6.51673198e-01
2.50619739e-01 6.17523193e-01 -2.10308984e-01 4.80508476e-01
-1.86358064e-01 1.66759968e-01 -2.08061457e-01 3.79992455e-01
-1.38473108e-01 -3.47975641e-01 5.06274343e-01 7.15140164e-01
9.84475091e-02 5.75529821e-02 4.47891146e-01 3.03623348e-01
-3.06031495e-01 6.84856027e-02 -9.67969060e-01 1.00230157e+00
4.22771364e-01 1.34262216e+00 -3.10671121e-01 -6.31853640e-01
-6.55971825e-01 6.42013550e-01 6.61781907e-01 2.28973418e-01
-4.03157562e-01 8.15659642e-01 5.90654433e-01 -1.23274706e-01
3.65719527e-01 4.30297136e-01 -4.57701027e-01 -1.65518653e+00
1.24276496e-01 -1.20837891e+00 9.78159189e-01 -3.36424381e-01
-1.61444187e+00 1.25202060e-01 3.05729896e-01 -1.25982606e+00
-3.41984928e-01 -4.89419103e-01 -7.71753252e-01 6.82081342e-01
-1.07767272e+00 -1.18326128e+00 5.12384027e-02 5.31813920e-01
6.29241467e-01 -2.52790034e-01 8.34349990e-01 -2.17323348e-01
-3.06281179e-01 5.38705766e-01 7.27926418e-02 -3.92857671e-01
8.47422659e-01 -1.16749239e+00 -2.66138077e-01 4.85186994e-01
7.92235583e-02 2.97723472e-01 5.77811480e-01 -3.73830229e-01
-1.16527677e+00 -7.84086704e-01 9.99793231e-01 -1.00201035e+00
1.73006177e-01 -2.22523823e-01 -4.17341530e-01 1.06991231e+00
4.54402536e-01 1.04556099e-01 1.08356881e+00 6.13220692e-01
-7.24666953e-01 -2.21948862e-01 -1.27157187e+00 3.37379187e-01
5.89177132e-01 -7.12284923e-01 -4.69919860e-01 2.14924842e-01
-6.23899093e-03 -1.43016562e-01 -9.05861378e-01 4.87774283e-01
9.79957819e-01 -1.27576077e+00 7.72754848e-01 -7.86075652e-01
5.49302876e-01 -1.28217414e-01 -6.06852055e-01 -1.47679293e+00
-7.93743357e-02 9.21904221e-02 -6.51480556e-02 1.01931429e+00
6.77138269e-01 -5.79087794e-01 6.73062742e-01 6.82950974e-01
3.30327481e-01 -7.87169099e-01 -1.05938721e+00 -3.68238449e-01
1.07819274e-01 -2.67779797e-01 -1.14350967e-01 1.10667253e+00
-7.81197324e-02 5.75852334e-01 -7.07076371e-01 2.09039822e-01
1.13760161e+00 5.71489990e-01 8.07534516e-01 -1.49267817e+00
-4.46243197e-01 -3.55330892e-02 -2.32283711e-01 -3.67630810e-01
4.70254451e-01 -1.06595433e+00 -2.37430796e-01 -1.29115081e+00
9.04173970e-01 -1.29065081e-01 3.82870156e-03 6.28856122e-02
-2.93935418e-01 -5.85581921e-02 5.32750189e-01 1.47671580e-01
-7.47791648e-01 2.20344007e-01 9.96691644e-01 -3.07752132e-01
2.05007046e-01 3.28105539e-01 -5.35857379e-01 1.21780801e+00
5.00731587e-01 -7.48769999e-01 -5.10985732e-01 -1.95721760e-01
5.22909582e-01 4.00351197e-01 3.23476613e-01 -4.99980003e-01
1.35684505e-01 -2.21913770e-01 3.99046242e-01 -6.92780554e-01
4.04811174e-01 -9.42329049e-01 2.89070070e-01 3.42193753e-01
-5.45416296e-01 1.10599503e-01 -5.00316203e-01 9.23855841e-01
-1.76900744e-01 -4.49274391e-01 9.99408126e-01 -3.97142202e-01
-3.22845489e-01 2.04789653e-01 -4.47659433e-01 2.53035188e-01
1.21954572e+00 4.04877961e-02 -1.90309733e-01 -5.44317901e-01
-1.08626974e+00 5.47384501e-01 7.79741585e-01 1.90107062e-01
3.23465347e-01 -1.48592079e+00 -8.59504938e-01 5.32137901e-02
3.53714049e-01 -3.43503386e-01 4.28318113e-01 6.06866539e-01
2.12966666e-01 -2.20608696e-01 2.61099003e-02 -5.97574770e-01
-1.91704881e+00 6.26556695e-01 5.25981605e-01 -5.70303917e-01
-1.94215387e-01 6.48424745e-01 4.04944748e-01 -5.95149815e-01
1.57717377e-01 9.94947031e-02 -2.96118379e-01 7.59433270e-01
4.87107724e-01 6.87598646e-01 -1.69152573e-01 -5.58175623e-01
-4.02906924e-01 8.57499063e-01 -1.01362191e-01 -5.90154886e-01
1.04002488e+00 -1.96767703e-01 4.08840040e-03 1.01698256e+00
1.29351437e+00 1.85764045e-01 -1.13491380e+00 -3.07684749e-01
-3.40461969e-01 -3.56118113e-01 -1.20840617e-01 -5.66048801e-01
-1.04669940e+00 7.22203970e-01 5.69616735e-01 -1.07490994e-01
6.58749640e-01 1.97571546e-01 -9.86945853e-02 3.96608025e-01
5.28066933e-01 -1.23108721e+00 -2.10751653e-01 4.12370205e-01
7.81745374e-01 -1.78100288e+00 5.01582503e-01 -5.50686359e-01
-1.03347945e+00 8.87340546e-01 1.95855498e-01 -2.21230239e-01
1.10526550e+00 7.05152452e-02 3.15168500e-01 -3.40790361e-01
-1.37924099e+00 -6.26849802e-03 4.54425931e-01 4.89091575e-01
1.33748144e-01 3.04927468e-01 -3.03757012e-01 3.00432831e-01
2.23525569e-01 -3.09531689e-01 3.51368308e-01 7.20925927e-01
-3.33960921e-01 -1.21691012e+00 -1.92380667e-01 5.46399474e-01
-4.60792989e-01 3.10828954e-01 -6.34300470e-01 5.99791348e-01
2.65622944e-01 1.05653739e+00 -2.43046433e-01 -3.81561130e-01
2.32324913e-01 4.70087230e-01 6.69721067e-01 -6.54160023e-01
-8.45747948e-01 -1.97337866e-01 3.55119973e-01 -4.35581028e-01
-6.83222175e-01 -1.03646851e+00 -3.09297681e-01 -3.14592183e-01
-6.37249887e-01 -6.20637042e-03 5.90090990e-01 9.33311403e-01
-1.46064997e-01 -1.41360626e-01 1.09347630e+00 -6.91647589e-01
-1.18746054e+00 -6.14391923e-01 -6.79984868e-01 6.23450756e-01
3.94737959e-01 -4.05667663e-01 -7.00869918e-01 4.88175228e-02] | [8.55849838256836, 4.461905002593994] |
115e88ec-c037-46ed-b6ea-41388ae546b2 | enhancing-space-time-video-super-resolution | 2207.08960 | null | https://arxiv.org/abs/2207.08960v3 | https://arxiv.org/pdf/2207.08960v3.pdf | Enhancing Space-time Video Super-resolution via Spatial-temporal Feature Interaction | The target of space-time video super-resolution (STVSR) is to increase both the frame rate (also referred to as the temporal resolution) and the spatial resolution of a given video. Recent approaches solve STVSR using end-to-end deep neural networks. A popular solution is to first increase the frame rate of the video; then perform feature refinement among different frame features; and last increase the spatial resolutions of these features. The temporal correlation among features of different frames is carefully exploited in this process. The spatial correlation among features of different (spatial) resolutions, despite being also very important, is however not emphasized. In this paper, we propose a spatial-temporal feature interaction network to enhance STVSR by exploiting both spatial and temporal correlations among features of different frames and spatial resolutions. Specifically, the spatial-temporal frame interpolation module is introduced to interpolate low- and high-resolution intermediate frame features simultaneously and interactively. The spatial-temporal local and global refinement modules are respectively deployed afterwards to exploit the spatial-temporal correlation among different features for their refinement. Finally, a novel motion consistency loss is employed to enhance the motion continuity among reconstructed frames. We conduct experiments on three standard benchmarks, Vid4, Vimeo-90K and Adobe240, and the results demonstrate that our method improves the state of the art methods by a considerable margin. Our codes will be available at https://github.com/yuezijie/STINet-Space-time-Video-Super-resolution. | ['Miaojing Shi', 'Zijie Yue'] | 2022-07-18 | null | null | null | null | ['space-time-video-super-resolution', 'video-super-resolution'] | ['computer-vision', 'computer-vision'] | [-8.90900865e-02 -6.70317173e-01 -1.84504122e-01 -2.82635421e-01
-5.24572313e-01 -2.06361845e-01 3.10142308e-01 -1.64591178e-01
-7.24466562e-01 7.68438041e-01 1.69639826e-01 1.82001084e-01
-1.02170408e-01 -8.72499526e-01 -6.28496170e-01 -6.17409110e-01
-1.86746389e-01 -4.50927109e-01 7.56333530e-01 -2.07941622e-01
2.21750602e-01 5.59216857e-01 -1.65084398e+00 3.92597944e-01
7.52603889e-01 1.11790109e+00 5.70413530e-01 7.18361259e-01
-3.81314754e-02 8.32849026e-01 -2.84715354e-01 5.37149869e-02
2.69052297e-01 -3.47906411e-01 -8.00025463e-01 1.24878913e-01
3.37482035e-01 -5.82775235e-01 -4.50045377e-01 1.10352945e+00
3.69693577e-01 4.26112086e-01 -7.99731314e-02 -9.56032515e-01
-5.04655659e-01 4.13550466e-01 -1.09694874e+00 8.06209743e-01
2.05469117e-01 2.19433114e-01 6.64020121e-01 -1.00133181e+00
7.27050006e-01 1.15499830e+00 4.34175551e-01 3.55161011e-01
-9.95122075e-01 -7.77134955e-01 2.64735192e-01 6.35320604e-01
-1.62153077e+00 -3.08857769e-01 7.73796499e-01 -2.17447162e-01
6.83679461e-01 1.89613923e-01 5.60591638e-01 6.05856001e-01
1.86369494e-01 4.65380311e-01 7.36095190e-01 -5.51703908e-02
-1.41353635e-02 -3.68598491e-01 -8.89823288e-02 6.38760269e-01
-2.11266980e-01 2.15970740e-01 -8.62960577e-01 2.62855589e-01
1.53755963e+00 1.93978161e-01 -5.97030699e-01 5.22675775e-02
-1.39674854e+00 5.32059431e-01 6.62967205e-01 6.61012292e-01
-6.12081826e-01 9.12738964e-02 3.93692911e-01 1.11867093e-01
4.81916457e-01 1.11054406e-02 -4.51876283e-01 -1.26731589e-01
-1.14641821e+00 -3.91816087e-02 8.00852478e-02 8.90233934e-01
1.00068164e+00 1.19060919e-01 -4.39206928e-01 8.79323184e-01
6.64173663e-02 1.22105956e-01 4.85566854e-01 -1.46863699e+00
4.78702366e-01 2.51480639e-01 2.96413124e-01 -1.24143314e+00
-1.81715772e-01 -2.07248405e-01 -1.29632747e+00 4.51463073e-01
3.95781994e-01 -3.34842056e-02 -7.43310273e-01 1.61703753e+00
4.75445718e-01 6.98846698e-01 -8.80040750e-02 1.32037187e+00
8.18799019e-01 9.50043321e-01 1.76636413e-01 -5.43095648e-01
1.43976462e+00 -9.72623289e-01 -8.70110214e-01 9.30230916e-02
1.40367866e-01 -8.11560333e-01 9.22501862e-01 1.04036875e-01
-1.60641253e+00 -1.05945194e+00 -9.45321918e-01 -2.85104245e-01
3.79881859e-02 -3.14186923e-02 9.25512314e-02 -1.94713846e-01
-9.86335099e-01 9.28110003e-01 -9.37434554e-01 2.07794875e-01
3.61284167e-01 3.07777345e-01 -3.54179174e-01 -1.25482216e-01
-1.44836247e+00 4.09626722e-01 4.26423192e-01 8.63240883e-02
-5.31805813e-01 -9.15479064e-01 -8.06634307e-01 6.98241591e-02
3.40546906e-01 -6.32380664e-01 8.98440957e-01 -1.16774678e+00
-1.31690216e+00 3.56636763e-01 -4.16869968e-01 -3.05270314e-01
6.54183507e-01 -7.35295936e-02 -4.72179592e-01 4.26316828e-01
1.70070324e-02 6.94671810e-01 1.00162375e+00 -9.61485267e-01
-1.27660823e+00 -9.01623964e-02 1.71302721e-01 3.91246885e-01
-2.61192799e-01 2.19170481e-01 -9.93157804e-01 -7.97685742e-01
4.48666438e-02 -4.80778307e-01 -2.31537014e-01 1.70086190e-01
-4.40281443e-02 4.02497724e-02 9.43149567e-01 -9.20143068e-01
1.40184915e+00 -2.37388539e+00 2.40584254e-01 -1.05824433e-01
4.03305590e-01 3.57792944e-01 -1.35204107e-01 -2.33885229e-01
-1.47471860e-01 1.05055191e-01 -5.14580943e-02 -2.29533836e-01
-6.49425745e-01 -5.72629906e-02 7.56179467e-02 3.91667575e-01
1.24200337e-01 6.74447238e-01 -9.56375360e-01 -6.17427528e-01
6.48529768e-01 1.11608088e+00 -4.39828962e-01 1.05081335e-01
1.23412311e-01 8.14873278e-01 -4.06524897e-01 2.29273319e-01
9.30381060e-01 -3.79683942e-01 -1.54708460e-01 -5.91765404e-01
-5.83710611e-01 6.24478459e-02 -1.45066226e+00 1.67510438e+00
-5.39488614e-01 6.84128761e-01 8.59921798e-02 -4.10289347e-01
7.42559373e-01 2.94929445e-01 6.73722029e-01 -1.05004919e+00
-5.48592657e-02 1.15195543e-01 -2.37806365e-01 -2.70632237e-01
7.25154936e-01 2.72831798e-01 4.11817342e-01 -1.51731774e-01
-2.08243191e-01 5.50395906e-01 3.63639802e-01 8.50116089e-02
7.37787008e-01 1.21591918e-01 3.03555369e-01 -1.58242181e-01
9.19388831e-01 -3.13536227e-01 8.95686388e-01 3.16162646e-01
-2.00769410e-01 9.15622413e-01 1.84503019e-01 -6.21013939e-01
-1.31824577e+00 -9.47670877e-01 -5.96703216e-02 9.59031403e-01
6.53467715e-01 -4.50598180e-01 -5.49885571e-01 -2.78741777e-01
-3.83388102e-01 2.57466555e-01 -6.40826404e-01 1.52747512e-01
-9.26537335e-01 -2.83715755e-01 -7.04237586e-03 6.42022789e-01
9.97585952e-01 -1.09470403e+00 -8.86596262e-01 4.16469574e-01
-3.56552213e-01 -1.30023658e+00 -8.12135637e-01 -2.83740461e-01
-8.71806324e-01 -7.72729993e-01 -1.03545904e+00 -6.46007061e-01
4.43464994e-01 5.98624527e-01 1.03587449e+00 3.67528945e-01
-1.52192220e-01 -1.76019236e-01 -4.26344693e-01 4.53638494e-01
-1.66269600e-01 -7.04255626e-02 -1.73724398e-01 1.75580949e-01
-1.20440587e-01 -7.81087577e-01 -9.92015958e-01 5.71963906e-01
-1.13589227e+00 6.04342222e-01 2.74843603e-01 8.16577435e-01
9.36151028e-01 3.08989137e-01 1.20779030e-01 -4.69631076e-01
2.40684319e-02 -2.41259411e-01 -6.75086379e-01 2.18531527e-02
5.32713197e-02 -3.29124592e-02 7.48034894e-01 -5.87678254e-01
-1.06776118e+00 -8.39669928e-02 -2.12956026e-01 -8.19796920e-01
-7.59343430e-02 3.64763021e-01 -2.79236585e-01 5.94970435e-02
2.39376649e-01 3.59690636e-01 -2.77997255e-01 -3.41650784e-01
1.17716543e-01 2.07083240e-01 7.51671135e-01 -2.87330866e-01
8.65985632e-01 6.26668870e-01 -6.18992299e-02 -7.98126042e-01
-6.24668181e-01 -3.81744325e-01 -6.39961362e-01 -2.35715941e-01
1.10495651e+00 -1.11088479e+00 -6.16135538e-01 5.46425164e-01
-1.16146433e+00 -4.11726773e-01 -1.68493330e-01 5.32877624e-01
-3.80364954e-01 4.46595013e-01 -9.77346122e-01 -2.15218812e-01
-3.72819901e-01 -1.37653291e+00 8.66069615e-01 5.90241671e-01
9.12769809e-02 -8.64923298e-01 -2.71072060e-01 4.56776358e-02
5.49819708e-01 2.11565763e-01 5.10209262e-01 2.06686437e-01
-8.12202632e-01 3.61788303e-01 -6.54759288e-01 2.14215100e-01
3.46223235e-01 3.00494134e-01 -4.71778244e-01 -3.67127061e-01
-1.40350863e-01 2.91175812e-01 7.93164670e-01 6.70376420e-01
1.41957879e+00 -2.77204543e-01 5.03845289e-02 9.51276064e-01
1.53565538e+00 2.08837718e-01 9.37352419e-01 5.38977921e-01
8.75333548e-01 3.45639288e-01 8.61541331e-01 5.90510786e-01
3.60063612e-01 9.81347561e-01 3.12725902e-01 -2.92065531e-01
-3.22693795e-01 1.18976235e-01 2.94143438e-01 6.09028578e-01
-6.83858991e-01 3.80429775e-02 -5.43320417e-01 4.53906089e-01
-1.93703532e+00 -1.18402338e+00 -2.42625251e-01 2.17128515e+00
8.65686178e-01 -7.40661249e-02 1.31441146e-01 1.81129068e-01
1.09444427e+00 3.89675885e-01 -5.23413956e-01 -1.71410423e-02
-2.52134264e-01 -2.05979757e-02 4.39643681e-01 6.77465975e-01
-1.11787140e+00 7.92081356e-01 4.29701138e+00 1.11592388e+00
-1.16691053e+00 2.33632863e-01 9.04685259e-01 -3.59898418e-01
3.77966054e-02 -2.87467599e-01 -7.41123915e-01 6.62534058e-01
6.87649846e-01 -2.04640374e-01 5.87386847e-01 5.95648825e-01
4.91757810e-01 -3.48571688e-02 -8.08925331e-01 1.11470759e+00
-4.30931509e-01 -1.66901207e+00 -1.96816862e-01 -7.46774673e-02
7.99247324e-01 -1.99386910e-01 1.73376858e-01 -2.15037335e-02
2.01162994e-02 -7.88410604e-01 7.92133212e-01 4.03868049e-01
1.04466903e+00 -1.16729379e+00 5.28436542e-01 2.01616567e-02
-1.90241992e+00 -4.97175381e-03 -4.05721635e-01 2.97134936e-01
2.96419352e-01 5.82496285e-01 1.66113436e-01 6.77013040e-01
1.17310238e+00 9.77089345e-01 -3.93666804e-01 8.89211953e-01
-1.20809399e-01 6.96536675e-02 -1.28605813e-01 6.30197465e-01
2.33401179e-01 -1.69449951e-02 4.97118503e-01 1.15922797e+00
4.99444306e-01 4.02003139e-01 6.71076681e-03 7.16473639e-01
-1.13530746e-02 -4.50422540e-02 3.29370536e-02 5.51849365e-01
5.34174979e-01 1.25652277e+00 -6.16806746e-01 -4.71054703e-01
-6.37220263e-01 1.13989449e+00 1.10680729e-01 4.81783748e-01
-1.20639932e+00 -1.90192282e-01 8.62117350e-01 2.34786794e-01
4.94598985e-01 -3.08471203e-01 -1.30330622e-01 -1.20968366e+00
1.91131473e-01 -8.02468538e-01 3.76929194e-01 -7.89487600e-01
-8.90007198e-01 8.41069758e-01 -8.18817019e-02 -1.56910443e+00
-1.61163926e-01 7.48120770e-02 -3.91371787e-01 1.06989896e+00
-1.82110417e+00 -7.86573529e-01 -6.87892139e-01 1.07135999e+00
9.32996869e-01 1.45964503e-01 2.17104360e-01 6.31560326e-01
-5.57346165e-01 4.32512105e-01 1.25465519e-03 2.13667914e-01
6.06702447e-01 -7.06400275e-01 3.44083935e-01 1.05541849e+00
-2.55252570e-01 3.57220769e-01 5.33433259e-01 -5.38076162e-01
-1.03385532e+00 -1.30261528e+00 6.41459107e-01 2.23680258e-01
4.10426974e-01 5.30080572e-02 -1.34235418e+00 3.43311518e-01
1.57342330e-01 6.88736975e-01 5.46320938e-02 -4.53692287e-01
-2.71805644e-01 -1.66680664e-01 -1.03516722e+00 6.63476467e-01
9.31763768e-01 -4.07811254e-01 1.55406278e-02 -2.45198950e-01
9.29350138e-01 -5.96712053e-01 -1.03968322e+00 4.68125522e-01
5.58991671e-01 -1.35979402e+00 1.27596128e+00 -3.67909782e-02
8.55560720e-01 -7.27963448e-01 -1.42563790e-01 -9.97213960e-01
-7.05942512e-01 -5.03726780e-01 -2.40262806e-01 1.32045496e+00
1.87930670e-02 -2.50393987e-01 6.10666633e-01 6.05975688e-01
5.59403338e-02 -7.13092029e-01 -1.02651739e+00 -6.19241059e-01
-1.86111480e-01 -2.63552785e-01 5.99769473e-01 1.04175079e+00
-4.03904229e-01 -6.25436604e-02 -5.59217095e-01 4.23363835e-01
6.37849748e-01 2.79071927e-03 4.32774395e-01 -8.98996234e-01
-1.98641688e-01 -4.29732531e-01 -3.91464233e-01 -1.04515541e+00
-9.02974382e-02 -2.69853830e-01 -1.23795755e-01 -1.23800504e+00
2.37445593e-01 -2.06284702e-01 -4.90483016e-01 1.76150054e-01
-4.53145027e-01 4.55420285e-01 4.87863868e-01 2.67644018e-01
-6.82285845e-01 4.14219946e-01 1.57166886e+00 2.26012245e-01
-5.58244586e-01 -3.00255448e-01 -1.80640936e-01 7.87189841e-01
7.89679050e-01 -2.37133965e-01 -1.88358337e-01 -6.00339353e-01
-1.71420604e-01 5.49118757e-01 5.66200793e-01 -1.23476350e+00
3.67927790e-01 -2.80134529e-01 8.01959395e-01 -6.92845702e-01
4.20659006e-01 -8.04395795e-01 4.37655717e-01 3.47896814e-01
-3.66957694e-01 1.75954595e-01 2.07285449e-01 3.84352118e-01
-3.74656886e-01 2.10906535e-01 1.40541482e+00 -7.22402111e-02
-1.17620325e+00 7.74889827e-01 -1.93893582e-01 -1.17624596e-01
9.89569485e-01 -2.18868360e-01 -2.11424604e-01 -2.69708037e-01
-6.69125855e-01 9.74584371e-02 6.70154512e-01 4.65078950e-01
9.29347217e-01 -1.44213450e+00 -8.22180748e-01 1.28077179e-01
-2.32931212e-01 1.09824814e-01 9.89510179e-01 9.64467525e-01
-5.24822176e-01 -4.09814045e-02 -6.01338148e-01 -6.33932829e-01
-1.58755910e+00 6.86456025e-01 3.99106443e-01 -3.41448635e-01
-9.33190346e-01 7.15122640e-01 4.14163291e-01 5.26839256e-01
1.60598740e-01 -2.69294381e-01 -3.57449830e-01 -1.97525755e-01
1.14347589e+00 3.96726012e-01 -3.56087416e-01 -8.93208086e-01
-2.51746267e-01 7.81501770e-01 -1.02795802e-01 -6.66571110e-02
1.32839310e+00 -6.29870117e-01 -5.48041351e-02 1.76550388e-01
1.38656735e+00 -1.39299944e-01 -1.78846276e+00 -6.26048684e-01
-3.52299362e-01 -9.75211859e-01 2.04401314e-01 -2.57486552e-01
-1.72132075e+00 7.99678981e-01 8.10230196e-01 -1.86613514e-04
1.59645939e+00 -2.65931517e-01 1.04136121e+00 -4.24907684e-01
3.03822190e-01 -9.10534263e-01 1.45202428e-01 5.05366862e-01
7.82404721e-01 -1.16625857e+00 3.90125392e-03 -5.22526383e-01
-6.12793148e-01 1.18491733e+00 6.36516035e-01 -2.47788921e-01
3.88052404e-01 4.33922470e-01 -2.12804586e-01 2.42957011e-01
-6.47714913e-01 -2.01707914e-01 3.61469120e-01 4.07704353e-01
6.09447300e-01 -2.91222811e-01 -2.58816183e-01 3.56342047e-01
2.49979839e-01 1.16040908e-01 3.81608337e-01 5.37833750e-01
-2.74762720e-01 -8.95556092e-01 -4.22699034e-01 -5.16247414e-02
-6.90899134e-01 -1.73672885e-01 4.23607230e-01 6.52735233e-01
2.67033458e-01 9.32423890e-01 2.28304401e-01 -4.22458023e-01
1.64659455e-01 -5.80116391e-01 2.07762599e-01 -5.69106340e-02
-5.04710138e-01 4.57773089e-01 -1.63160667e-01 -1.01914549e+00
-6.85598493e-01 -4.77683812e-01 -1.50909281e+00 -6.58106387e-01
1.45930246e-01 -3.62346619e-02 3.26154888e-01 6.02277458e-01
4.17317837e-01 9.55405653e-01 7.69671261e-01 -1.26077187e+00
2.38822281e-01 -5.58376968e-01 -5.98242760e-01 3.93361330e-01
6.05771184e-01 -4.57099050e-01 -2.30410174e-01 3.39221716e-01] | [11.055828094482422, -1.8473185300827026] |
2f40dec3-20bc-4a2c-ae01-43988e471f44 | c-3po-cyclic-three-phase-optimization-for | 1909.11303 | null | https://arxiv.org/abs/1909.11303v3 | https://arxiv.org/pdf/1909.11303v3.pdf | C-3PO: Cyclic-Three-Phase Optimization for Human-Robot Motion Retargeting based on Reinforcement Learning | Motion retargeting between heterogeneous polymorphs with different sizes and kinematic configurations requires a comprehensive knowledge of (inverse) kinematics. Moreover, it is non-trivial to provide a kinematic independent general solution. In this study, we developed a cyclic three-phase optimization method based on deep reinforcement learning for human-robot motion retargeting. The motion retargeting learning is performed using refined data in a latent space by the cyclic and filtering paths of our method. In addition, the human-in-the-loop based three-phase approach provides a framework for the improvement of the motion retargeting policy by both quantitative and qualitative manners. Using the proposed C-3PO method, we were successfully able to learn the motion retargeting skill between the human skeleton and motion of the multiple robots such as NAO, Pepper, Baxter and C-3PO. | ['Joo-Haeng Lee', 'Taewoo Kim'] | 2019-09-25 | null | null | null | null | ['motion-retargeting'] | ['computer-vision'] | [-1.83449700e-01 2.47189283e-01 -3.31463993e-01 4.34890479e-01
-4.17235285e-01 -3.43764007e-01 4.44308430e-01 -6.83397502e-02
-7.37472594e-01 8.69856834e-01 1.94796864e-02 -1.13413341e-01
-3.97620261e-01 -6.25225008e-01 -8.44548285e-01 -9.71967876e-01
-1.22408949e-01 5.49694955e-01 2.68144250e-01 -4.81090963e-01
2.04954311e-01 6.52319670e-01 -1.24159384e+00 -2.17480436e-01
6.95158124e-01 2.47615367e-01 4.91369009e-01 1.00282753e+00
5.37979841e-01 6.39190316e-01 -8.24956317e-03 1.84539452e-01
4.02439356e-01 -2.91276634e-01 -8.11666012e-01 1.44722208e-01
-2.40452006e-01 -7.30432197e-02 -4.08418894e-01 6.62654400e-01
5.33169150e-01 5.48729658e-01 6.13564789e-01 -1.24892735e+00
-2.88507700e-01 6.60940468e-01 -6.46770179e-01 -1.63469374e-01
5.65284967e-01 6.56326234e-01 5.47243476e-01 -6.24503553e-01
9.36084807e-01 1.31894124e+00 5.16622186e-01 6.84239566e-01
-1.14041722e+00 -5.25852561e-01 -1.07019737e-01 2.19563290e-01
-1.33312404e+00 4.99937534e-02 5.07776797e-01 -6.05103374e-01
7.86873341e-01 -4.53285165e-02 1.04678392e+00 1.30525911e+00
9.45464253e-01 7.31745839e-01 8.08190584e-01 -3.80187213e-01
2.82750547e-01 -3.89895946e-01 -3.96407962e-01 7.84698486e-01
4.53809232e-01 5.07121325e-01 -3.50982934e-01 4.49323133e-02
1.14928925e+00 -1.20596513e-01 1.35615719e-02 -8.96707714e-01
-1.56219101e+00 8.08133066e-01 3.50484520e-01 3.26749235e-01
-5.98574042e-01 5.48136115e-01 2.79554039e-01 8.69557559e-02
-1.14143655e-01 8.06777775e-01 -7.32344270e-01 -2.22473711e-01
-6.44594789e-01 7.49337435e-01 7.39294350e-01 8.14770341e-01
9.38841045e-01 1.04684852e-01 -7.74682686e-02 -3.04022264e-02
4.91389632e-01 3.61827344e-01 5.97280562e-01 -1.43890774e+00
3.40021521e-01 4.26727265e-01 4.84513134e-01 -1.13937795e+00
-8.66179585e-01 -2.43960455e-01 -6.63162768e-01 5.75179875e-01
4.18802887e-01 -6.22239351e-01 -6.63314223e-01 1.64012694e+00
7.36782193e-01 -1.37591407e-01 3.05024385e-01 1.03224480e+00
9.95699316e-02 3.99562180e-01 1.48507774e-01 -2.97311634e-01
9.98219609e-01 -1.19541645e+00 -6.32368624e-01 -8.71765241e-03
1.06632459e+00 -6.79744542e-01 1.00036883e+00 1.91194937e-01
-1.03283083e+00 -6.43710792e-01 -1.04848123e+00 3.31016123e-01
-1.65831625e-01 7.49452487e-02 3.01135063e-01 2.87595212e-01
-8.44288707e-01 9.49243605e-01 -1.26389861e+00 -7.21566617e-01
-1.80937037e-01 5.92441440e-01 -6.44711018e-01 2.33191103e-01
-1.14372504e+00 1.19537449e+00 7.56478548e-01 3.99774723e-02
-1.03428137e+00 -5.07799089e-01 -9.35342073e-01 -3.42840344e-01
6.30868256e-01 -1.21242869e+00 1.20883298e+00 -7.90892184e-01
-2.30327630e+00 3.67597431e-01 5.12971640e-01 -3.75150740e-01
8.94407451e-01 -4.38499063e-01 3.44412029e-02 2.18618706e-01
2.47465312e-01 8.25123906e-01 8.03047121e-01 -1.12915492e+00
-5.92887044e-01 -1.18156001e-02 6.37919875e-03 4.83899862e-01
2.35482484e-01 -3.67513269e-01 -1.92877173e-01 -9.09685373e-01
-7.05990270e-02 -1.41806376e+00 -6.90796435e-01 -6.54710382e-02
-2.66907185e-01 1.17574476e-01 6.64672375e-01 -4.91471291e-01
7.93371141e-01 -1.85589981e+00 9.84567225e-01 8.84096511e-03
2.26560491e-03 6.90546557e-02 -9.68749896e-02 6.97025895e-01
8.88240263e-02 -1.66311279e-01 -3.11060119e-02 8.31933767e-02
-1.81876019e-01 3.54097784e-01 3.13208163e-01 7.56331980e-01
-3.85667346e-02 9.02582467e-01 -1.19039345e+00 -5.77449560e-01
3.44965786e-01 9.16406959e-02 -7.25995004e-01 2.59193778e-01
-3.11142594e-01 1.19520867e+00 -6.89961255e-01 3.27234685e-01
2.32738480e-01 1.68979540e-01 1.67805672e-01 -9.93249938e-02
-4.92903113e-01 -5.43768227e-01 -1.22668409e+00 2.06601596e+00
-2.76995003e-01 1.39869064e-01 2.23129094e-01 -8.77491653e-01
6.13018811e-01 3.70273530e-01 1.15465701e+00 -5.29563844e-01
3.65097225e-01 2.76870161e-01 5.12059331e-02 -8.53844702e-01
7.86665559e-01 -3.26618135e-01 -3.95346940e-01 4.57381248e-01
4.97954749e-02 -1.79567084e-01 3.14040869e-01 -2.11042568e-01
1.00381243e+00 6.53459013e-01 4.57103968e-01 -4.81437713e-01
5.48545718e-01 3.67436826e-01 4.81042027e-01 5.11172175e-01
-2.54880011e-01 2.90253878e-01 1.28162876e-01 -2.09518224e-01
-1.22521389e+00 -8.46758664e-01 4.09209549e-01 9.06756759e-01
2.73007363e-01 -7.16723725e-02 -8.46694350e-01 -3.95932347e-01
3.05915251e-02 5.47058105e-01 -6.37627244e-01 -5.48191845e-01
-9.21451867e-01 -3.14368635e-01 4.37244534e-01 5.83161354e-01
3.98021102e-01 -1.02151227e+00 -1.05449605e+00 5.64306378e-01
-1.03282770e-02 -8.88112366e-01 -4.89037931e-01 2.48978660e-01
-8.20426106e-01 -1.33719420e+00 -9.38525617e-01 -7.57582426e-01
3.95634621e-01 2.25326512e-02 2.88723499e-01 -7.68342018e-02
-2.17506602e-01 7.22148895e-01 -6.14951432e-01 -1.54598085e-02
-6.66390955e-01 2.89364189e-01 4.49580133e-01 -3.83760273e-01
-5.39227486e-01 -3.42434138e-01 -6.90288484e-01 7.05439091e-01
-7.13127375e-01 2.34836593e-01 8.77915740e-01 6.72039330e-01
5.73734999e-01 2.75378041e-02 3.27342838e-01 -3.23556304e-01
3.74744475e-01 -3.50572348e-01 -6.23901427e-01 -1.91898197e-02
-5.24583578e-01 3.31294060e-01 4.23970789e-01 -8.23127270e-01
-1.06953979e+00 5.96087337e-01 -1.13199264e-01 -3.40922207e-01
-1.37392610e-01 4.43422407e-01 -1.23875163e-01 -3.53605717e-01
6.90893948e-01 -1.05419673e-01 6.76097423e-02 -3.14740926e-01
9.35796857e-01 1.07988477e-01 6.23061955e-01 -7.49652565e-01
9.38193619e-01 4.73952264e-01 2.78126270e-01 -6.86018288e-01
-9.04541761e-02 -3.50422263e-01 -1.04539204e+00 -3.80043507e-01
1.30998087e+00 -6.46202922e-01 -1.14232075e+00 5.28768003e-01
-1.10732293e+00 -8.61106396e-01 -4.70459402e-01 9.52526391e-01
-1.26135528e+00 6.72954738e-01 -7.19849467e-01 -4.79371071e-01
-1.18224882e-01 -1.46419668e+00 9.34324443e-01 2.46953651e-01
-3.94748837e-01 -8.50674927e-01 6.72599554e-01 6.60587251e-02
1.09367102e-01 5.58325648e-01 7.62163758e-01 -2.44447187e-01
-4.81800109e-01 7.42422789e-03 4.52575117e-01 -1.09457627e-01
2.33939052e-01 3.84887904e-02 -1.75754532e-01 -6.51292026e-01
-7.50600547e-02 -1.18230119e-01 2.37426415e-01 4.37811822e-01
4.28591698e-01 -4.73262072e-01 -5.04452527e-01 3.69314760e-01
1.16478121e+00 1.03271417e-01 5.43290496e-01 9.35007930e-01
8.33813846e-01 6.82251215e-01 1.17343533e+00 5.08638799e-01
3.92172158e-01 8.46220732e-01 5.42606831e-01 7.82112107e-02
-9.87144932e-03 -2.24787831e-01 4.94808227e-01 7.19243169e-01
-2.35615700e-01 7.16295466e-03 -9.10475552e-01 4.85064328e-01
-2.37458134e+00 -6.81868434e-01 -1.71851546e-01 1.97117281e+00
6.83547258e-01 -1.18606277e-01 2.05241129e-01 -5.58945239e-02
8.01199794e-01 -3.78847539e-01 -4.27546531e-01 -1.12598442e-01
2.92129725e-01 -1.91680655e-01 8.05451035e-01 5.93936384e-01
-9.97397721e-01 1.15828180e+00 6.63945723e+00 6.36874318e-01
-1.05354190e+00 -1.34189174e-01 -1.40936002e-01 1.74247921e-01
2.72017308e-02 1.64881989e-01 -6.19108021e-01 3.19018334e-01
7.48804271e-01 -1.90331534e-01 2.85823762e-01 7.16475189e-01
5.87963343e-01 -4.02951211e-01 -1.00304365e+00 6.62309170e-01
-4.63222265e-01 -1.19541478e+00 -1.96301609e-01 5.29477671e-02
9.73346591e-01 -3.31150204e-01 -1.00600697e-01 1.40161842e-01
4.13288713e-01 -7.26639390e-01 1.05366302e+00 8.00271273e-01
2.34676123e-01 -7.24218607e-01 6.57957196e-01 5.99834502e-01
-1.15142250e+00 -2.50740230e-01 -5.86860999e-02 -8.58363807e-02
3.30355942e-01 1.34615293e-02 -8.09409380e-01 8.59698296e-01
5.01394331e-01 5.89521408e-01 -1.87063530e-01 9.72211957e-01
-3.17828357e-01 -2.79550720e-02 -1.48637086e-01 -2.14809719e-02
1.46722570e-01 -1.96552068e-01 9.30432081e-01 8.54800463e-01
2.56547868e-01 -1.68774128e-01 3.70861322e-01 5.43008924e-01
5.01902103e-01 1.35399148e-01 -5.87808788e-01 -6.06887527e-02
-1.64276212e-01 1.23267233e+00 -8.57221186e-01 1.14616878e-01
-2.88759843e-02 6.95720077e-01 1.01706088e-01 4.22791600e-01
-8.61202240e-01 -1.02962449e-01 5.71619332e-01 3.74479182e-02
3.11759442e-01 -7.65197754e-01 -3.58806327e-02 -9.99544799e-01
-4.71678227e-01 -7.47184813e-01 2.68402219e-01 -6.36963248e-01
-9.40856576e-01 8.94743353e-02 4.77135509e-01 -1.24409008e+00
-6.40960515e-01 -5.30091703e-01 -2.34655991e-01 3.24212760e-01
-9.66937900e-01 -1.20022881e+00 -1.72591865e-01 6.62818968e-01
4.06010300e-01 -2.30847910e-01 4.94261950e-01 1.08662911e-01
-5.58497012e-01 2.66554981e-01 9.21527222e-02 -3.69143814e-01
8.23585272e-01 -9.56840277e-01 5.60824089e-02 7.01837897e-01
-4.54814672e-01 4.01885897e-01 1.34496474e+00 -9.75660145e-01
-1.71146595e+00 -9.87623572e-01 9.32555571e-02 -2.70513386e-01
8.86843920e-01 1.97248861e-01 -6.13729894e-01 7.85294890e-01
3.26275229e-02 -4.14014190e-01 1.99458972e-01 -5.66557527e-01
5.23174286e-01 2.33966485e-01 -9.62328315e-01 9.61701572e-01
7.71517456e-01 2.35675126e-02 -7.43632019e-01 2.03636989e-01
9.70705807e-01 -6.60049140e-01 -9.46207166e-01 4.61899072e-01
6.89482808e-01 -3.12388718e-01 1.02033138e+00 -6.49073958e-01
1.30709663e-01 -5.31347215e-01 -1.68046385e-01 -1.33175552e+00
-5.04899383e-01 -9.90950167e-01 -1.69913784e-01 6.56070113e-01
3.25155519e-02 -4.44250166e-01 8.77317727e-01 1.19967036e-01
-2.23260760e-01 -5.63569605e-01 -9.27762032e-01 -8.16399157e-01
2.96894222e-01 -2.14617819e-01 2.24558979e-01 5.35195112e-01
6.64356276e-02 1.35285139e-01 -6.32290781e-01 1.91639572e-01
5.11906743e-01 -1.38384894e-01 1.10583735e+00 -8.17097485e-01
-5.55510879e-01 -2.51887560e-01 -3.22217554e-01 -7.23099232e-01
2.65056819e-01 -5.96070170e-01 3.97440374e-01 -1.52746654e+00
3.87782184e-03 -1.92416087e-01 7.10565820e-02 4.62609768e-01
2.43510064e-02 -2.53908545e-01 2.85610676e-01 3.81915331e-01
-4.83515590e-01 8.63487959e-01 1.79769087e+00 -7.29919132e-03
-5.21048903e-01 6.08876906e-02 -1.46138012e-01 7.11542845e-01
9.25109029e-01 -7.31601715e-01 -2.75485545e-01 -1.56155661e-01
3.47647339e-01 3.81653517e-01 3.45091969e-01 -1.03476393e+00
3.61510962e-01 -5.93763888e-01 -3.25623304e-02 -4.78818566e-01
7.79942945e-02 -8.94567966e-01 4.24290448e-01 1.17880940e+00
-1.63135424e-01 1.63528159e-01 2.30501890e-01 9.16565359e-01
1.92121819e-01 -1.60102069e-01 8.58464241e-01 -2.56213307e-01
-8.74138951e-01 1.12215549e-01 -9.79603887e-01 -2.54599512e-01
1.45081544e+00 -2.62548089e-01 2.99693970e-03 -1.42682731e-01
-9.10843432e-01 3.25581014e-01 6.76834524e-01 3.71982455e-01
3.82367551e-01 -1.32553291e+00 -3.84607553e-01 -1.19587615e-01
-6.64128959e-02 9.57248360e-03 3.75789583e-01 1.18816721e+00
-9.72466946e-01 4.31811154e-01 -8.47527623e-01 -6.04739547e-01
-8.40897381e-01 8.25387716e-01 3.32908899e-01 -3.39936048e-01
-6.56077802e-01 2.35416993e-01 2.55553842e-01 -6.46188259e-01
-1.26001179e-01 -3.15196157e-01 -1.93582699e-01 -3.09991539e-01
-2.24030599e-01 9.83225942e-01 -3.09714794e-01 -8.30348969e-01
-3.01421851e-01 8.67931843e-01 3.57008219e-01 -4.57589269e-01
1.15784168e+00 -3.94724429e-01 7.46502280e-02 3.18974733e-01
9.27408934e-01 -2.55905867e-01 -1.51337016e+00 3.39372844e-01
5.86903468e-02 -1.83858499e-01 -3.41683894e-01 -3.33165348e-01
-7.66696811e-01 3.17531914e-01 6.50210977e-01 -3.67649794e-01
6.99574471e-01 -1.01678364e-01 7.24346161e-01 6.14168823e-01
6.48965240e-01 -1.22541356e+00 5.80206215e-01 4.59619999e-01
1.00210309e+00 -8.69439423e-01 8.34420323e-02 -9.15552452e-02
-6.88653827e-01 1.23761284e+00 7.48839796e-01 -2.79196501e-01
4.51277733e-01 1.28005385e-01 6.20148666e-02 -1.37452379e-01
-4.22475696e-01 -9.59023312e-02 2.76033789e-01 7.80579150e-01
6.07763715e-02 1.06429525e-01 -6.07859910e-01 2.91422039e-01
1.14679582e-01 -6.90647960e-02 7.21837342e-01 1.28721857e+00
-4.91974920e-01 -1.33579850e+00 -5.04932106e-01 -3.52709234e-01
2.77051181e-02 7.18909860e-01 -5.79764806e-02 1.31246066e+00
6.87114075e-02 7.64965177e-01 -4.56261843e-01 -5.11072636e-01
4.91847396e-01 -3.77754509e-01 6.91138804e-01 -6.03963852e-01
-4.86552536e-01 2.74387002e-01 -9.56234932e-02 -9.18036342e-01
-4.99629140e-01 -6.89999402e-01 -1.61402297e+00 -1.42348662e-01
-2.25938633e-01 -2.01007146e-02 5.71301758e-01 1.05717289e+00
3.03269774e-01 5.92676699e-01 3.86319339e-01 -1.36492562e+00
-5.25329590e-01 -8.04743290e-01 -6.12897456e-01 1.45966291e-01
2.91854590e-01 -1.13888001e+00 1.59345120e-02 2.49792468e-02] | [4.831786155700684, 1.070684790611267] |
7467bd2b-17c4-430f-8f56-8a57405c63ac | copr-towards-accurate-visual-localization | 2304.07426 | null | https://arxiv.org/abs/2304.07426v1 | https://arxiv.org/pdf/2304.07426v1.pdf | CoPR: Towards Accurate Visual Localization With Continuous Place-descriptor Regression | Visual Place Recognition (VPR) is an image-based localization method that estimates the camera location of a query image by retrieving the most similar reference image from a map of geo-tagged reference images. In this work, we look into two fundamental bottlenecks for its localization accuracy: reference map sparseness and viewpoint invariance. Firstly, the reference images for VPR are only available at sparse poses in a map, which enforces an upper bound on the maximum achievable localization accuracy through VPR. We therefore propose Continuous Place-descriptor Regression (CoPR) to densify the map and improve localization accuracy. We study various interpolation and extrapolation models to regress additional VPR feature descriptors from only the existing references. Secondly, we compare different feature encoders and show that CoPR presents value for all of them. We evaluate our models on three existing public datasets and report on average around 30% improvement in VPR-based localization accuracy using CoPR, on top of the 15% increase by using a viewpoint-variant loss for the feature encoder. The complementary relation between CoPR and Relative Pose Estimation is also discussed. | ['Julian Francisco Pieter Kooij', 'Liangliang Nan', 'Mubariz Zaffar'] | 2023-04-14 | null | null | null | null | ['image-based-localization', 'visual-localization', 'visual-place-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.20832331e-01 -2.57261038e-01 -4.29056019e-01 -4.14212584e-01
-1.38157308e+00 -7.31079221e-01 8.91939282e-01 1.87996998e-01
-5.67328811e-01 5.40991843e-01 3.21163386e-01 1.28886640e-01
-1.11992732e-01 -6.19618773e-01 -1.11088681e+00 -3.26756895e-01
-2.30104979e-02 1.50628820e-01 4.19189423e-01 -1.48131475e-01
4.54255939e-01 9.82883453e-01 -1.48846507e+00 -5.24291098e-02
4.96544540e-01 1.17310703e+00 4.72604156e-01 6.27509236e-01
3.23315740e-01 9.44823682e-01 -4.46182042e-01 -2.84392357e-01
3.56390566e-01 4.60856147e-02 -5.94075561e-01 -5.68554811e-02
5.56989431e-01 -4.42433536e-01 -9.71381068e-01 8.60922635e-01
3.79552245e-01 2.11600780e-01 6.86976731e-01 -1.40827072e+00
-8.69657457e-01 1.43830776e-01 -7.72210717e-01 8.43719691e-02
7.29605079e-01 -4.02363598e-01 9.39156413e-01 -1.12455118e+00
8.82080913e-01 8.03224683e-01 9.60456908e-01 1.07982174e-01
-1.14436042e+00 -3.61456603e-01 1.10008167e-02 3.85546148e-01
-2.48215199e+00 -6.88176751e-01 5.98653615e-01 -2.30254918e-01
9.74559844e-01 3.13719839e-01 1.16709746e-01 7.61329293e-01
1.29839301e-01 4.41956639e-01 7.45983064e-01 -3.14527303e-01
2.25118920e-01 1.93006948e-01 -4.23692375e-01 5.59546351e-01
2.18735054e-01 -3.83357331e-02 -1.05617535e+00 -5.58474027e-02
8.59723449e-01 1.86830580e-01 -3.58960271e-01 -6.32297397e-01
-1.40671122e+00 7.24063396e-01 9.33714867e-01 2.02564716e-01
-3.64276081e-01 4.43549871e-01 -1.06958471e-01 -1.79513022e-02
2.90635735e-01 3.46147776e-01 -2.10847035e-01 -1.01875111e-01
-1.26745570e+00 -2.11987030e-02 5.09223402e-01 1.50389457e+00
1.23932517e+00 -3.39271933e-01 -1.58468679e-01 7.87935615e-01
3.49439770e-01 1.02322960e+00 2.80567229e-01 -1.17248964e+00
6.24909103e-01 2.66986310e-01 4.62991416e-01 -1.56910002e+00
-1.46941632e-01 -8.65569860e-02 -5.58772385e-01 -3.16838264e-01
1.74858198e-01 3.51256996e-01 -7.16217995e-01 1.56772876e+00
4.88697104e-02 1.73913449e-01 -1.57599896e-01 9.23949540e-01
6.28400564e-01 7.86116660e-01 -2.69749522e-01 2.98030555e-01
1.08790016e+00 -8.39104712e-01 -3.33836257e-01 -2.86656886e-01
4.87552047e-01 -6.83373034e-01 5.00163436e-01 -8.91039670e-02
-5.80891550e-01 -3.81547898e-01 -1.13237894e+00 -5.23926973e-01
-4.77959663e-01 5.05478263e-01 3.18522990e-01 4.30517972e-01
-1.51631892e+00 3.80588382e-01 -8.71391237e-01 -4.69170272e-01
3.91585901e-02 4.52767611e-01 -8.61316502e-01 -5.36332250e-01
-8.61573339e-01 1.00369334e+00 -5.04455827e-02 3.72653082e-02
-6.82413161e-01 -6.36170268e-01 -1.09743059e+00 -6.80310428e-02
-2.88721770e-02 -2.19212845e-01 8.27538729e-01 -4.16330189e-01
-1.08955419e+00 8.17742288e-01 -6.92211568e-01 -6.01240396e-01
4.50013906e-01 -5.58707565e-02 -3.81765187e-01 2.59503990e-01
5.69400132e-01 1.02449095e+00 5.42630017e-01 -1.24320579e+00
-8.12638342e-01 -2.35182285e-01 1.48604110e-01 2.69507945e-01
1.94119647e-01 -2.06459969e-01 -8.54683518e-01 -3.48463863e-01
4.75232303e-01 -9.71284628e-01 -1.80828243e-01 1.83992401e-01
-1.65174335e-01 1.42913237e-01 4.49316859e-01 -7.61390567e-01
9.30144846e-01 -2.31511879e+00 -1.98170617e-01 2.72616059e-01
1.73550844e-01 -4.23345327e-01 -2.51710624e-01 6.06161475e-01
2.30739340e-01 -9.39175710e-02 1.62455097e-01 -3.92003149e-01
8.00446142e-03 2.52256989e-01 -5.57150185e-01 9.91876960e-01
1.63285494e-01 9.92448807e-01 -8.90668511e-01 -4.10715908e-01
5.44230640e-01 9.42957401e-01 -5.72407663e-01 -3.79853584e-02
3.97609651e-01 5.39025843e-01 -2.02423707e-01 7.29507327e-01
9.40184057e-01 -2.12328643e-01 -1.71043381e-01 -3.27355117e-01
-4.89275455e-01 2.91209400e-01 -1.06743681e+00 2.01963353e+00
-6.13371730e-01 9.92770076e-01 -2.50341028e-01 -5.05387843e-01
1.08770573e+00 -6.19203299e-02 5.62428176e-01 -1.04279721e+00
-3.23380142e-01 3.73899490e-01 -6.87420368e-01 1.85965031e-01
1.14677250e+00 5.66557467e-01 -1.79408491e-01 -1.02979138e-01
2.69598722e-01 2.15126708e-01 -1.64233550e-01 1.37387350e-01
1.28439593e+00 1.59843996e-01 4.36223328e-01 -2.05207527e-01
6.38555825e-01 3.17240618e-02 2.91694939e-01 1.12030959e+00
-1.55329138e-01 1.06550634e+00 1.61856875e-01 -4.28741872e-01
-1.08295822e+00 -1.16003704e+00 -2.33461812e-01 9.22937751e-01
6.70952678e-01 -4.75315303e-01 -3.27145815e-01 -4.07685727e-01
9.11728665e-02 5.36995471e-01 -6.86114550e-01 1.09747723e-01
-5.38684666e-01 -3.07847500e-01 4.70464617e-01 4.99171585e-01
7.27575719e-01 -3.11375797e-01 -5.06584764e-01 -1.74952731e-01
-5.62876761e-01 -1.19041574e+00 -6.94643319e-01 1.98746398e-01
-3.33573967e-01 -7.66216993e-01 -1.02746797e+00 -6.01521611e-01
8.70181382e-01 8.80977273e-01 9.42155242e-01 -2.35182434e-01
1.10784762e-01 8.25665593e-01 -4.39312935e-01 1.62962914e-01
4.29873243e-02 3.32186818e-01 1.33273751e-01 5.17715476e-02
4.11002398e-01 -3.63210738e-01 -8.14543426e-01 6.30051732e-01
-4.12082404e-01 -1.76917553e-01 6.12859368e-01 5.98723948e-01
9.90255713e-01 -5.28981984e-01 1.35948911e-01 -2.12190121e-01
-1.37552857e-01 -5.85493267e-01 -8.94059420e-01 3.49834740e-01
-4.52934742e-01 1.79203659e-01 2.36003265e-01 -1.89816192e-01
-5.28385580e-01 6.79472089e-01 -1.78056076e-01 -6.03321731e-01
7.83579126e-02 1.61993638e-01 -3.53646162e-03 -7.10473716e-01
6.97085202e-01 6.99986160e-01 -3.70743185e-01 -1.73946455e-01
6.31818533e-01 6.48109496e-01 8.03329170e-01 -3.24478090e-01
1.03788793e+00 5.75175762e-01 8.62308443e-02 -8.81387234e-01
-5.55914402e-01 -8.92971277e-01 -8.01838696e-01 -8.07675198e-02
6.10646725e-01 -1.40687943e+00 -7.33785808e-01 -1.96581483e-01
-1.20982242e+00 4.21637371e-02 -1.67710379e-01 5.44950128e-01
-7.35113025e-01 1.86412930e-01 -2.08921641e-01 -5.68265796e-01
7.44267255e-02 -1.20896101e+00 1.66141641e+00 8.72975588e-02
1.30845368e-01 -6.71749890e-01 2.38685817e-01 -4.98679988e-02
5.50199747e-01 1.42870247e-01 -1.25226416e-02 -4.55057085e-01
-1.11253488e+00 -5.15274346e-01 -6.33567035e-01 -1.60751536e-01
-1.61415413e-02 -5.74249685e-01 -1.03884971e+00 -4.11891043e-01
-2.07904622e-01 1.75486982e-01 8.15684378e-01 2.88891584e-01
8.52778494e-01 -2.99484044e-01 -5.57564616e-01 9.11823928e-01
1.81944382e+00 5.95284440e-02 7.75950432e-01 4.75532979e-01
8.27304602e-01 -1.70395465e-03 7.95406759e-01 4.68304396e-01
8.12125444e-01 1.02409697e+00 3.26606661e-01 1.00116611e-01
-2.00407758e-01 -8.11620235e-01 3.29018295e-01 6.06970668e-01
-1.72574773e-01 -4.12649848e-02 -9.66712594e-01 7.37434089e-01
-1.75669312e+00 -7.48907387e-01 1.26464814e-01 2.58081102e+00
3.22937071e-01 -5.77727318e-01 -2.55054384e-01 -2.85883844e-01
7.55543172e-01 3.34778845e-01 -2.29787230e-01 1.48876816e-01
-7.85465911e-02 -1.93785608e-01 1.40221643e+00 6.75975621e-01
-1.24020934e+00 8.58901381e-01 6.63060999e+00 7.31993735e-01
-1.22911859e+00 2.30032608e-01 4.29144025e-01 1.77744433e-01
-1.29300635e-02 7.44659901e-02 -9.96620834e-01 4.42112595e-01
1.02019954e+00 -1.38222560e-01 7.07588553e-01 1.17547143e+00
-8.51747841e-02 -4.14628953e-01 -1.15943420e+00 1.53839409e+00
4.65981990e-01 -1.54488647e+00 -1.12408750e-01 3.11765462e-01
7.71357119e-01 6.02137268e-01 8.43785778e-02 2.29524001e-01
-6.15050904e-02 -8.99009943e-01 1.07687271e+00 4.42853302e-01
1.16176391e+00 -8.47826362e-01 7.36953557e-01 8.27622637e-02
-1.52023625e+00 5.28175086e-02 -7.56598234e-01 1.95109591e-01
-2.95830164e-02 1.13994427e-01 -1.02998090e+00 7.42864966e-01
5.79233885e-01 7.85870671e-01 -1.10991085e+00 1.11955142e+00
-1.77080542e-01 -1.37583157e-02 -5.70935071e-01 2.27604881e-01
1.63781121e-01 9.09000635e-02 2.56969839e-01 1.02695882e+00
6.13564253e-01 -2.10953489e-01 4.57724649e-03 7.53766000e-01
-2.01301709e-01 -1.18658639e-01 -9.39226985e-01 5.13847411e-01
1.04007113e+00 1.09048569e+00 -7.43063867e-01 1.29995376e-01
-5.14886141e-01 1.20630896e+00 3.60973954e-01 2.04771295e-01
-1.14678073e+00 -5.30266285e-01 4.85791147e-01 3.62739861e-01
5.82164824e-01 -5.70117891e-01 8.74329284e-02 -1.32113492e+00
7.34422728e-02 -2.63901621e-01 -1.20228700e-01 -9.56488371e-01
-8.27159524e-01 4.50640082e-01 -1.07754275e-01 -1.52538347e+00
-2.56722212e-01 -3.98209244e-01 1.19759649e-01 8.89388204e-01
-1.73023736e+00 -1.35524797e+00 -4.83734608e-01 5.95181227e-01
1.77708611e-01 1.45393208e-01 8.04835796e-01 6.63780451e-01
1.99290682e-02 7.25926280e-01 5.03189862e-01 3.57253313e-01
7.93130219e-01 -1.06196737e+00 5.95674634e-01 8.99129808e-01
6.90373898e-01 6.67722583e-01 4.33046967e-01 -4.45639580e-01
-1.74453926e+00 -1.33555281e+00 1.11078727e+00 -1.05583549e+00
5.00230193e-01 -5.55112123e-01 -4.47905988e-01 9.03337836e-01
-2.38305375e-01 4.78959233e-01 3.13129485e-01 -4.25483771e-02
-6.66714787e-01 -3.02436620e-01 -1.11487985e+00 5.39221346e-01
9.19659972e-01 -9.67337370e-01 -3.61947358e-01 1.45281121e-01
6.79212213e-01 -7.17459381e-01 -8.60596120e-01 6.53653592e-02
5.18873036e-01 -9.14029121e-01 1.37150335e+00 4.61984545e-01
2.36516476e-01 -7.76079535e-01 -9.09572184e-01 -8.79112840e-01
-3.47328216e-01 -2.01607540e-01 5.11808433e-02 1.34718835e+00
1.80461690e-01 -5.75094700e-01 5.22103846e-01 5.99016666e-01
2.45104924e-01 -3.86124104e-01 -1.30188930e+00 -9.03523207e-01
-3.77719909e-01 -3.63472909e-01 5.90769947e-01 7.18060553e-01
-2.86626160e-01 2.29552109e-02 -4.94904935e-01 6.42525136e-01
5.76023638e-01 -1.60220429e-01 7.81048596e-01 -6.80182278e-01
1.47595137e-01 1.95423901e-01 -1.17991245e+00 -1.47605455e+00
-2.65717097e-02 -8.99547994e-01 2.62926340e-01 -1.50268388e+00
1.72018230e-01 -4.00221407e-01 -3.86290342e-01 3.71166766e-01
4.12426740e-01 6.97498679e-01 1.93477094e-01 6.13750994e-01
-1.20938027e+00 4.21543419e-01 4.35284376e-01 -1.02331959e-01
-1.39814392e-01 -2.54175544e-01 -5.13959169e-01 3.80141199e-01
3.44127268e-01 -5.93944311e-01 -1.14711404e-01 -7.27344394e-01
2.65245408e-01 3.57277058e-02 6.68175757e-01 -1.28567433e+00
7.00852096e-01 6.83987215e-02 7.43718028e-01 -9.76179540e-01
5.89575291e-01 -9.57271516e-01 3.56061608e-01 9.77349132e-02
-3.21365625e-01 2.02661604e-01 -1.58246920e-01 8.95043075e-01
-2.78438419e-01 -1.73312426e-02 6.02164924e-01 1.33289352e-01
-1.06017578e+00 2.87877709e-01 -5.40591143e-02 -1.51087984e-01
8.88146639e-01 -2.10916579e-01 -3.10497940e-01 -5.28755903e-01
-2.55445689e-01 -3.30734737e-02 8.62932920e-01 4.92359519e-01
6.12974584e-01 -1.68867314e+00 -4.36588258e-01 1.30781040e-01
5.46498537e-01 -1.23168796e-01 4.51144688e-02 9.07180011e-01
-8.86841476e-01 8.28889966e-01 -8.27070996e-02 -8.23969662e-01
-9.15082574e-01 8.06626260e-01 1.66258857e-01 1.52995154e-01
-5.55024862e-01 4.56658542e-01 3.65880057e-02 -3.66771430e-01
1.09842956e-01 -1.28733277e-01 1.32330075e-01 -1.50836647e-01
7.27252424e-01 4.50909197e-01 1.55958474e-01 -1.40618706e+00
-8.56443822e-01 8.08652461e-01 1.65679827e-01 -3.00947666e-01
1.25337386e+00 -6.77537560e-01 3.29891127e-03 3.47572923e-01
1.46220875e+00 3.02477092e-01 -1.41132247e+00 -2.59543002e-01
1.89420432e-01 -8.53703678e-01 7.40945265e-02 -4.55327868e-01
-7.76725590e-01 5.63852549e-01 6.90391898e-01 -2.75213987e-01
8.28662157e-01 2.63154209e-01 2.53734469e-01 6.18633866e-01
1.21313596e+00 -7.82089591e-01 -4.17060345e-01 5.21396756e-01
8.85032058e-01 -1.35410440e+00 1.95595622e-01 -2.88235724e-01
-4.89345998e-01 8.76233041e-01 8.08433071e-02 -4.22005624e-01
5.49060643e-01 1.99459679e-02 -2.37494171e-01 1.34394050e-01
-1.98643401e-01 -2.16940120e-01 3.72159153e-01 9.57152247e-01
1.37872472e-01 -8.52861628e-02 1.42172515e-01 3.03039938e-01
-1.14517160e-01 -1.37915328e-01 2.75287271e-01 6.83056295e-01
-2.47051597e-01 -6.83538854e-01 -4.55181271e-01 -2.76771467e-02
-2.52072722e-01 -1.62853256e-01 -1.05547473e-01 7.59314954e-01
-4.52514738e-02 8.53576720e-01 2.17992574e-01 -4.68023777e-01
2.12691933e-01 -3.53667974e-01 4.51601297e-01 -2.27332219e-01
5.45596257e-02 -2.98312128e-01 -2.35830933e-01 -1.01873386e+00
-3.95370156e-01 -5.22240460e-01 -8.41767728e-01 -4.38040823e-01
-1.94870770e-01 5.51775172e-02 9.16779757e-01 5.79982042e-01
7.66625226e-01 8.71543214e-02 8.32159460e-01 -1.30793738e+00
-3.16897273e-01 -5.18585682e-01 -5.48004806e-01 -5.03790900e-02
4.55457658e-01 -7.46833563e-01 -3.78089130e-01 9.42282900e-02] | [7.659852504730225, -1.9807161092758179] |
c02562e0-ea61-4e85-878d-5b2d3683d5e6 | regula-sub-rosa-latent-backdoor-attacks-on | 1905.10447 | null | https://arxiv.org/abs/1905.10447v1 | https://arxiv.org/pdf/1905.10447v1.pdf | Regula Sub-rosa: Latent Backdoor Attacks on Deep Neural Networks | Recent work has proposed the concept of backdoor attacks on deep neural networks (DNNs), where misbehaviors are hidden inside "normal" models, only to be triggered by very specific inputs. In practice, however, these attacks are difficult to perform and highly constrained by sharing of models through transfer learning. Adversaries have a small window during which they must compromise the student model before it is deployed. In this paper, we describe a significantly more powerful variant of the backdoor attack, latent backdoors, where hidden rules can be embedded in a single "Teacher" model, and automatically inherited by all "Student" models through the transfer learning process. We show that latent backdoors can be quite effective in a variety of application contexts, and validate its practicality through real-world attacks against traffic sign recognition, iris identification of lab volunteers, and facial recognition of public figures (politicians). Finally, we evaluate 4 potential defenses, and find that only one is effective in disrupting latent backdoors, but might incur a cost in classification accuracy as tradeoff. | ['Hai-Tao Zheng', 'Yuanshun Yao', 'Ben Y. Zhao', 'Huiying Li'] | 2019-05-24 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 1.98916808e-01 3.26443106e-01 -2.96945989e-01 -2.06090599e-01
-3.40505809e-01 -1.30188465e+00 6.53558195e-01 -5.51078916e-01
-3.01845461e-01 7.13037193e-01 -6.15884066e-01 -1.01475370e+00
1.28248632e-02 -8.07480514e-01 -1.18394375e+00 -8.70477140e-01
-2.34404817e-01 4.03757431e-02 3.98194373e-01 -8.03164318e-02
-2.73002125e-03 9.22462404e-01 -1.17611706e+00 3.19658071e-01
3.56580615e-01 5.11298299e-01 -7.67511368e-01 7.31943369e-01
3.33258510e-01 7.69314051e-01 -1.13048387e+00 -7.31290340e-01
5.08799136e-01 -2.61346661e-02 -7.30998933e-01 -3.54412764e-01
8.11409712e-01 -6.04164600e-01 -6.06342316e-01 1.21806014e+00
2.33229995e-01 -2.12576568e-01 2.29348645e-01 -1.73393834e+00
-5.21907210e-01 5.92170060e-01 -2.09648371e-01 8.39017779e-02
-5.08936793e-02 4.00252253e-01 6.55809224e-01 -1.43731222e-01
2.84765035e-01 1.24242520e+00 5.94857395e-01 1.16891122e+00
-1.44265842e+00 -1.52659166e+00 1.78861856e-01 -1.07280247e-01
-9.83711958e-01 -6.58764124e-01 6.86326444e-01 -3.42607379e-01
8.44965518e-01 6.14399672e-01 4.19983923e-01 1.74418116e+00
3.11008751e-01 6.81120336e-01 1.13610125e+00 -3.42330128e-01
1.58685774e-01 4.43041950e-01 4.10029769e-01 9.91415441e-01
5.08552194e-01 6.36900485e-01 -4.39484477e-01 -8.84662688e-01
6.59827530e-01 -7.77700469e-02 -2.66729444e-01 -3.16147059e-01
-4.05483842e-01 9.52586472e-01 -3.24214483e-03 -8.31823125e-02
2.46709779e-01 3.60555977e-01 1.26518220e-01 6.51095033e-01
1.54014140e-01 4.38077778e-01 -5.73027611e-01 1.94467559e-01
-7.47204661e-01 1.85219780e-01 1.10922146e+00 5.55662751e-01
5.76813161e-01 2.74762958e-01 2.61418015e-01 1.52403623e-01
3.05977881e-01 3.63568068e-01 3.29635143e-01 -7.51512945e-01
2.45094433e-01 7.01778159e-02 5.19611202e-02 -8.95161450e-01
-3.19454335e-02 -4.10398096e-01 -3.12277853e-01 5.11942804e-01
5.30700028e-01 -6.40683413e-01 -1.08833611e+00 1.94991052e+00
3.65411222e-01 6.14038050e-01 1.94134507e-02 1.75189495e-01
3.01689982e-01 4.78300065e-01 -6.76068887e-02 1.70462623e-01
1.09438324e+00 -6.76775098e-01 -3.03877264e-01 -1.18134588e-01
6.50330305e-01 -4.54261243e-01 6.70096457e-01 6.82877719e-01
-9.03761089e-01 -6.56831935e-02 -1.23611462e+00 4.71033096e-01
-6.55970097e-01 -3.67461741e-01 7.55740345e-01 1.65152109e+00
-9.94253516e-01 5.49699187e-01 -1.02599633e+00 4.17086110e-02
6.77480638e-01 9.79722381e-01 -4.90254134e-01 3.08864474e-01
-1.14405692e+00 6.38500750e-01 -2.40723342e-02 -8.53212774e-02
-1.58018124e+00 -6.66477025e-01 -6.62987709e-01 1.13667911e-02
3.16321492e-01 -2.99110532e-01 1.13021064e+00 -9.86284494e-01
-1.60727274e+00 9.62584794e-01 2.30927974e-01 -8.92144620e-01
4.53723609e-01 -1.92448184e-01 -6.08867705e-01 1.12210400e-01
-4.04720873e-01 4.51460749e-01 1.24928439e+00 -1.24788094e+00
-2.00296119e-01 -3.27686667e-01 4.03511286e-01 -5.98345041e-01
-9.13503706e-01 4.04002190e-01 2.64276057e-01 -4.75347221e-01
-3.27620149e-01 -1.20403230e+00 -8.99714697e-03 -8.46222639e-02
-6.02143109e-01 -6.06320165e-02 1.46152365e+00 -2.33260721e-01
8.73449206e-01 -2.11414075e+00 -5.67003071e-01 5.38198054e-01
3.35158467e-01 9.29304719e-01 -1.12928145e-01 1.29356876e-01
-2.49941602e-01 5.87517023e-01 1.03998303e-01 -1.53438240e-01
4.29944173e-02 4.54455584e-01 -1.05967557e+00 6.10258877e-01
-1.69901803e-01 7.36124456e-01 -4.58911747e-01 1.21967532e-02
-1.56258956e-01 3.42272103e-01 -6.64563060e-01 -1.97459888e-02
-1.64135799e-01 1.61415309e-01 -4.56316888e-01 5.89961410e-01
6.20265603e-01 -9.11862031e-02 2.01106116e-01 4.54273939e-01
2.31752232e-01 3.37887764e-01 -7.81861722e-01 6.34799421e-01
-9.97437984e-02 7.83281744e-01 2.75498211e-01 -8.47885489e-01
7.74894476e-01 6.26554668e-01 6.72956482e-02 5.65478159e-03
2.67546415e-01 -5.09792231e-02 1.52308628e-01 -1.83362186e-01
8.71294644e-03 7.85278082e-02 3.36772874e-02 7.76532054e-01
1.16810881e-01 4.38987195e-01 -4.73310232e-01 6.22045249e-02
1.26872480e+00 -2.21462622e-01 -2.18893781e-01 -6.55162483e-02
9.86787230e-02 -2.33709574e-01 6.26236618e-01 1.15320134e+00
-4.96345758e-01 -2.42582429e-02 9.20148790e-01 -5.22384107e-01
-6.32868350e-01 -1.06520486e+00 7.97420591e-02 9.79355097e-01
-1.68651149e-01 -1.91896662e-01 -1.06273830e+00 -1.24627852e+00
2.26951659e-01 5.98660946e-01 -7.00502157e-01 -6.28713667e-01
-4.75303620e-01 -5.81564665e-01 1.67371249e+00 3.93187284e-01
5.09213328e-01 -6.91451490e-01 -5.98772883e-01 -1.27080634e-01
4.99912828e-01 -1.19719219e+00 -3.06563914e-01 3.23809505e-01
-9.05671418e-01 -1.30211425e+00 -2.94294953e-01 -4.96038646e-01
8.59135032e-01 2.10689634e-01 6.60675049e-01 3.78424138e-01
-2.75897175e-01 3.31110507e-01 1.54073864e-01 -6.35524571e-01
-6.46847427e-01 -4.43556458e-02 5.06895900e-01 1.44907728e-01
5.00379384e-01 -7.45779932e-01 -1.30505905e-01 4.20139760e-01
-9.19409990e-01 -5.11571646e-01 2.96303898e-01 8.46896112e-01
-2.87068188e-01 2.18416020e-01 2.69468069e-01 -1.28868210e+00
5.42969048e-01 -3.66348952e-01 -1.11266458e+00 3.69190782e-01
-4.61223751e-01 -2.86144048e-01 8.41098726e-01 -8.08993757e-01
-6.68349266e-01 -1.20002858e-01 4.10402939e-02 -9.24508512e-01
-3.25617880e-01 -2.39815801e-01 -3.38053137e-01 -8.21700275e-01
8.65101635e-01 9.70497802e-02 -3.34544219e-02 -4.21507299e-01
6.25779852e-02 4.01231050e-01 3.07162851e-01 -8.01341832e-01
1.36209881e+00 5.68260431e-01 8.80781636e-02 -8.54504883e-01
-5.90873539e-01 2.55168796e-01 -8.51059332e-02 4.74072807e-02
4.89943057e-01 -6.42435431e-01 -1.07679605e+00 7.31407821e-01
-1.20631218e+00 -4.45548028e-01 1.09988973e-01 2.45005429e-01
-1.29730806e-01 3.07074696e-01 -7.41644204e-01 -7.13374496e-01
-1.15058785e-02 -1.09451699e+00 2.61151165e-01 3.16892594e-01
-1.71196058e-01 -1.13244605e+00 -2.29740784e-01 4.29474115e-01
1.46887049e-01 3.05642158e-01 1.12801218e+00 -1.20979083e+00
-6.95082307e-01 -5.42360723e-01 3.48159641e-01 6.93091333e-01
4.11033966e-02 3.90264213e-01 -1.47734332e+00 -5.57298243e-01
-7.01556774e-03 -4.74594682e-01 7.09393978e-01 5.77396899e-03
1.42793250e+00 -7.43708014e-01 -6.16430998e-01 8.92824233e-01
9.01576638e-01 5.28376281e-01 5.70105851e-01 2.08305970e-01
5.79847932e-01 4.78704363e-01 -3.35607588e-01 -4.68535498e-02
-2.57163793e-01 2.22359195e-01 5.05924940e-01 -2.25342244e-01
4.34882075e-01 -4.54951376e-01 7.79632151e-01 1.03489384e-01
2.20824227e-01 -1.72716215e-01 -8.56007755e-01 3.02231491e-01
-1.36348915e+00 -1.01054037e+00 3.00737917e-01 2.15292311e+00
7.44868398e-01 5.02404690e-01 6.75310344e-02 1.11300267e-01
6.91873014e-01 5.56213371e-02 -4.51709479e-01 -7.63304055e-01
8.61878097e-02 6.26659811e-01 9.15982485e-01 6.28459513e-01
-1.18825722e+00 1.20875597e+00 7.55448532e+00 6.31695390e-01
-1.23652446e+00 2.10093126e-01 6.29797816e-01 -2.34015703e-01
-3.46180707e-01 3.09237033e-01 -1.11882341e+00 3.34749192e-01
9.95307803e-01 9.49072093e-02 3.96098793e-01 9.04770553e-01
-2.03557223e-01 4.22447950e-01 -1.29756033e+00 4.67216939e-01
-1.33680716e-01 -1.52672493e+00 1.07033357e-01 5.39943635e-01
8.34406018e-01 -1.11742169e-01 6.20120406e-01 3.07728142e-01
8.83220494e-01 -1.22156262e+00 2.11072251e-01 6.11035600e-02
6.24437213e-01 -1.15510821e+00 7.24132508e-02 5.77303350e-01
-3.81770581e-01 -3.19916189e-01 -3.21647704e-01 1.96286172e-01
-7.71365643e-01 1.43438280e-01 -8.57773304e-01 -4.95653450e-02
6.21697426e-01 -1.11542553e-01 -2.95888752e-01 5.53598225e-01
-5.73230147e-01 1.28607762e+00 -4.16050673e-01 6.80589825e-02
5.39230287e-01 2.86959529e-01 6.65039599e-01 8.80891681e-01
-1.24675557e-01 3.12028695e-02 5.71212955e-02 1.02784419e+00
-3.06138396e-01 -6.80535376e-01 -1.14972234e+00 -2.18920350e-01
4.64019209e-01 1.04594934e+00 -5.43392599e-01 -9.95825604e-02
-4.76493798e-02 6.59180045e-01 8.09209701e-03 5.55609286e-01
-9.07423615e-01 -5.57584703e-01 1.10201263e+00 -3.72003354e-02
1.93067700e-01 -2.40061820e-01 -8.47614780e-02 -1.11160803e+00
-1.53275147e-01 -1.23932946e+00 2.90780187e-01 -2.52154559e-01
-1.01170731e+00 4.43401933e-01 -1.02628469e-01 -8.20103765e-01
-1.59234539e-01 -8.22009802e-01 -9.21470821e-01 7.53439724e-01
-1.36959743e+00 -1.17626643e+00 3.53274554e-01 1.08266318e+00
-1.38612390e-01 -6.39669418e-01 1.16859138e+00 1.12556651e-01
-8.15902829e-01 1.40067387e+00 8.28620493e-02 6.78751469e-01
4.66295332e-01 -7.67465651e-01 6.55218422e-01 1.03914428e+00
4.23476011e-01 1.18957114e+00 5.99313796e-01 -5.49282908e-01
-1.26471758e+00 -9.92303014e-01 5.13173521e-01 -6.55492365e-01
5.99658430e-01 -7.48586416e-01 -8.29492629e-01 1.26575649e+00
-1.97751597e-02 2.73873001e-01 1.13375366e+00 1.79093584e-01
-9.20943797e-01 -1.16638606e-02 -1.34192705e+00 8.93554389e-01
6.47201896e-01 -8.60898674e-01 -2.97257096e-01 3.62241715e-01
7.43274152e-01 -2.07837284e-01 -4.44545031e-01 1.68703839e-01
7.18114436e-01 -7.55164862e-01 8.83063197e-01 -1.23541319e+00
-1.22690752e-01 -1.91332176e-02 4.52845134e-02 -9.74118471e-01
1.69367954e-01 -1.40587234e+00 -5.51794589e-01 1.12453020e+00
3.26742500e-01 -1.26627147e+00 1.34332383e+00 7.48535156e-01
3.68657500e-01 -4.97313857e-01 -1.08459496e+00 -1.13853729e+00
4.21850055e-01 -3.12869102e-01 7.81907499e-01 1.25988460e+00
-2.33084455e-01 -2.39757836e-01 -5.64981222e-01 7.41711497e-01
9.60262477e-01 -3.91564935e-01 9.51879501e-01 -1.06233394e+00
-6.17307901e-01 -2.24572405e-01 -4.51885104e-01 -8.49798083e-01
5.47514677e-01 -7.70824432e-01 -4.63687360e-01 -2.13729233e-01
-2.62747556e-01 -4.31152076e-01 -5.54801464e-01 1.05101740e+00
3.45919907e-01 2.12537646e-01 1.21888377e-01 1.11067027e-01
1.00249909e-02 1.37751058e-01 7.98622727e-01 -1.94089562e-01
-5.93678541e-02 4.20590818e-01 -7.82145798e-01 1.05947661e+00
8.48432660e-01 -1.00497139e+00 -4.96221602e-01 -3.16525072e-01
-1.90979734e-01 -8.78501907e-02 8.04090500e-01 -9.53971148e-01
4.17058796e-01 -1.58858314e-01 4.59603578e-01 -2.58833040e-02
3.58992487e-01 -9.24545407e-01 -1.61798403e-01 7.03096390e-01
-3.47204268e-01 -1.65763140e-01 3.90857637e-01 4.26692694e-01
3.13526481e-01 -7.67776296e-02 9.64802444e-01 -7.97657222e-02
-2.37315491e-01 3.96547049e-01 -5.64163864e-01 -2.07013443e-01
1.26726770e+00 -3.72142673e-01 -6.43068433e-01 -2.79250294e-01
-7.53063321e-01 5.47016449e-02 2.72724837e-01 2.80551940e-01
6.46191001e-01 -7.90425181e-01 -2.18888685e-01 8.79473805e-01
-3.37269366e-01 -4.98214871e-01 -1.52993619e-01 3.54647130e-01
-3.40306461e-01 3.54644418e-01 -2.44349718e-01 -4.09866303e-01
-1.87744009e+00 4.61895674e-01 7.97353685e-01 -3.29741500e-02
-4.29900557e-01 1.10117674e+00 1.72488198e-01 -4.45261717e-01
6.06148362e-01 5.25959730e-02 1.72733888e-01 -3.20926577e-01
5.31428039e-01 6.53845444e-02 -5.26812635e-02 -8.45863447e-02
-3.50848436e-01 1.47167012e-01 -6.09586418e-01 -9.17623490e-02
9.49424863e-01 6.89870059e-01 -8.39739069e-02 -1.07417926e-01
1.08258653e+00 1.93673074e-01 -1.26673746e+00 -1.83974385e-01
-1.29190207e-01 -4.73486692e-01 -6.98934942e-02 -6.98397756e-01
-1.04167271e+00 9.81328845e-01 4.94090289e-01 4.48567331e-01
8.10471475e-01 -4.50606018e-01 9.06946957e-01 8.92651796e-01
4.74039406e-01 -7.64483988e-01 -8.63434225e-02 3.99270594e-01
1.03042237e-01 -6.28480077e-01 -3.27479392e-01 -2.51011491e-01
-4.51717824e-02 1.03326035e+00 8.05731654e-01 -3.34690064e-01
8.62555385e-01 5.85415542e-01 1.92929313e-01 -1.18572637e-03
-1.01859820e+00 5.95764995e-01 -1.83641449e-01 6.86560869e-01
-2.72061974e-01 1.37486354e-01 5.60248315e-01 5.47844470e-01
-3.12860161e-01 -1.94570363e-01 7.09789693e-01 1.16766930e+00
-4.52418119e-01 -1.40115488e+00 -6.92317069e-01 9.35539976e-02
-9.98553693e-01 1.24831505e-01 -6.73729300e-01 7.87906170e-01
3.16067904e-01 1.03159690e+00 -1.10527381e-01 -7.79357135e-01
-7.12280571e-02 2.85529733e-01 5.22006512e-01 -6.42317772e-01
-9.00939584e-01 -3.86999726e-01 -3.24613973e-02 -6.09036922e-01
1.53274313e-01 -3.03962380e-01 -8.25706840e-01 -7.25183964e-01
-4.08415467e-01 2.74360716e-01 5.48006952e-01 9.46271718e-01
1.34885564e-01 2.34036386e-01 1.06499422e+00 -1.15398407e-01
-1.09376478e+00 -2.41621152e-01 -5.63000619e-01 -1.08075075e-01
6.20389700e-01 -3.90496910e-01 -5.99457920e-01 -9.10660699e-02] | [5.750176429748535, 7.655880928039551] |
bb3ce437-3b9e-4f54-8d44-c99f54607756 | model-free-motion-planning-of-autonomous | 2305.00561 | null | https://arxiv.org/abs/2305.00561v1 | https://arxiv.org/pdf/2305.00561v1.pdf | Model-free Motion Planning of Autonomous Agents for Complex Tasks in Partially Observable Environments | Motion planning of autonomous agents in partially known environments with incomplete information is a challenging problem, particularly for complex tasks. This paper proposes a model-free reinforcement learning approach to address this problem. We formulate motion planning as a probabilistic-labeled partially observable Markov decision process (PL-POMDP) problem and use linear temporal logic (LTL) to express the complex task. The LTL formula is then converted to a limit-deterministic generalized B\"uchi automaton (LDGBA). The problem is redefined as finding an optimal policy on the product of PL-POMDP with LDGBA based on model-checking techniques to satisfy the complex task. We implement deep Q learning with long short-term memory (LSTM) to process the observation history and task recognition. Our contributions include the proposed method, the utilization of LTL and LDGBA, and the LSTM-enhanced deep Q learning. We demonstrate the applicability of the proposed method by conducting simulations in various environments, including grid worlds, a virtual office, and a multi-agent warehouse. The simulation results demonstrate that our proposed method effectively addresses environment, action, and observation uncertainties. This indicates its potential for real-world applications, including the control of unmanned aerial vehicles (UAVs). | ['Shaoping Xiao', 'Zhen Kan', 'Mingyu Cai', 'Junchao Li'] | 2023-04-30 | null | null | null | null | ['q-learning', 'motion-planning'] | ['methodology', 'robots'] | [ 1.59165598e-02 1.91388682e-01 -3.43099207e-01 -7.59073421e-02
-7.30235040e-01 -4.09228295e-01 5.40660501e-01 -3.09948862e-01
-5.02939880e-01 9.54191148e-01 -1.29579008e-01 -7.24950671e-01
-3.41422051e-01 -8.44942331e-01 -6.72195256e-01 -7.92158246e-01
-4.90269989e-01 6.11122966e-01 3.22402596e-01 -8.61194357e-02
4.77491282e-02 4.34196860e-01 -1.10401869e+00 7.08265752e-02
3.96398664e-01 1.10917270e+00 8.41109931e-01 8.03286672e-01
1.05004705e-01 1.28237283e+00 -3.17714274e-01 3.10784370e-01
3.73313546e-01 2.63055693e-02 -1.05015087e+00 2.98006535e-01
-6.92824304e-01 -6.41160131e-01 -4.40430224e-01 9.44142103e-01
1.48854718e-01 5.33289075e-01 3.80168378e-01 -1.78038979e+00
-1.74344689e-01 3.42917770e-01 -1.96479768e-01 3.18569392e-02
3.19323063e-01 8.63962829e-01 6.89960182e-01 -7.41713420e-02
2.66975999e-01 1.68665385e+00 2.43734092e-01 6.12571180e-01
-6.39839172e-01 -2.97512323e-01 5.66055715e-01 5.04917860e-01
-1.08809161e+00 1.36172965e-01 2.49193132e-01 -3.51054162e-01
1.52745020e+00 -3.13552439e-01 7.50468791e-01 9.44524407e-01
1.26342595e+00 9.84632075e-01 1.38315952e+00 -3.07082713e-01
6.79292202e-01 -6.16981208e-01 -1.88154608e-01 8.64533424e-01
2.19517276e-01 7.67883062e-01 -9.25055295e-02 -2.02822089e-01
6.01451933e-01 1.41254619e-01 4.11214739e-01 -3.15296561e-01
-1.38182354e+00 6.62860155e-01 -1.17823899e-01 -3.53577048e-01
-8.16451192e-01 6.35758102e-01 2.72956222e-01 3.25637072e-01
-2.28066504e-01 2.94877887e-01 -5.82849264e-01 -1.40783980e-01
-3.48286033e-01 4.47689921e-01 9.29887950e-01 1.27146792e+00
4.33815271e-01 5.52405238e-01 -3.54928195e-01 -2.64013439e-01
7.98412859e-01 9.51024652e-01 3.33674639e-01 -1.56024289e+00
4.25679415e-01 1.61790252e-01 9.56320345e-01 -3.46275866e-01
-7.12683260e-01 2.00689256e-01 -5.56312084e-01 6.49227381e-01
1.02797017e-01 -6.14534318e-01 -1.04104733e+00 1.64972270e+00
2.54231572e-01 2.84943163e-01 6.99775577e-01 6.61739767e-01
-1.71913058e-02 1.07602477e+00 1.71937168e-01 -4.86500204e-01
1.18398714e+00 -1.04165351e+00 -1.12734592e+00 -4.51410085e-01
4.40087616e-01 -5.58803976e-02 6.61174476e-01 5.91917753e-01
-9.71354187e-01 -5.11859119e-01 -1.19772363e+00 5.02498925e-01
5.80467321e-02 -2.79987723e-01 5.29750168e-01 1.45804286e-01
-1.07804918e+00 5.02048790e-01 -1.44371343e+00 -1.33584872e-01
9.16887224e-02 7.02978849e-01 -2.86383238e-02 -1.90989688e-01
-1.32829690e+00 1.14727008e+00 7.88304806e-01 2.70912498e-01
-1.94740522e+00 4.85521525e-01 -1.15873373e+00 -5.63734584e-02
1.11051989e+00 -7.33176410e-01 1.96982753e+00 -2.95362711e-01
-1.96700776e+00 -2.32765190e-02 -7.66929314e-02 -9.36650515e-01
1.66328341e-01 -3.95316258e-02 -4.26982880e-01 -6.08292297e-02
6.05632253e-02 4.76868510e-01 8.57012808e-01 -9.27835763e-01
-1.05314612e+00 -3.56309414e-01 3.81248713e-01 4.55588102e-01
5.83697498e-01 -2.20387399e-01 8.21920931e-02 4.35974561e-02
7.11011812e-02 -1.41435134e+00 -7.95872331e-01 -5.38786650e-01
-1.27792746e-01 -4.17418569e-01 5.96582413e-01 -3.90204400e-01
8.45936418e-01 -1.76314628e+00 -6.61493763e-02 -8.35809186e-02
-3.01124185e-01 4.20403667e-02 -3.49629790e-01 7.30862319e-01
5.72510481e-01 -2.42465422e-01 -1.55802682e-01 -1.89002886e-01
2.93131739e-01 7.61999011e-01 -5.51650465e-01 3.54343802e-01
2.08695859e-01 9.94688690e-01 -1.11325896e+00 -4.81647849e-01
2.66844660e-01 -2.62630135e-01 1.02760181e-01 2.38517478e-01
-1.00710368e+00 6.25072062e-01 -9.15092230e-01 6.07405901e-01
3.65655094e-01 1.43105701e-01 4.41437781e-01 7.03489602e-01
-3.86059582e-01 4.91843037e-02 -1.36298680e+00 1.51359117e+00
-3.36569875e-01 2.31196925e-01 -3.64183150e-02 -6.39011621e-01
7.91900396e-01 5.71602404e-01 4.73474115e-01 -6.83961987e-01
1.94693998e-01 -4.83920015e-02 1.30152870e-02 -7.69328177e-01
5.99654257e-01 -3.44699025e-01 -4.47297513e-01 5.54649472e-01
-1.39276177e-01 -3.78637463e-01 -1.54141963e-01 -2.98589230e-01
1.33142304e+00 5.31072617e-01 5.32549679e-01 1.05180629e-01
3.40153098e-01 3.89625758e-01 9.35354710e-01 1.09941649e+00
-7.37384498e-01 -3.66635114e-01 4.67324853e-01 -6.09491646e-01
-6.78950667e-01 -7.51932800e-01 6.29302323e-01 8.38768244e-01
4.66537297e-01 1.05043024e-01 -4.37846214e-01 -5.04717469e-01
-1.29938826e-01 1.11390162e+00 -2.44714960e-01 -2.73891449e-01
-4.85690117e-01 -3.99894923e-01 2.52092421e-01 5.41074038e-01
7.66507626e-01 -1.51975763e+00 -1.32916117e+00 5.57713270e-01
-4.34684977e-02 -1.15879917e+00 -4.16722298e-02 3.35686654e-01
-8.17476809e-01 -8.91301990e-01 -5.84022142e-02 -7.27682292e-01
2.19753459e-01 3.69497031e-01 6.03830040e-01 -5.49704731e-01
3.99736583e-01 6.48180962e-01 -1.14447229e-01 -6.80784702e-01
-5.28430402e-01 -3.57204944e-01 4.21789229e-01 -2.45523587e-01
1.49170637e-01 1.20848291e-01 -2.06250578e-01 2.94520110e-01
-8.55489373e-01 2.12368499e-02 7.07930148e-01 9.32307720e-01
8.67128849e-01 8.16613555e-01 3.89159501e-01 -2.95687973e-01
8.26375067e-01 -3.10927123e-01 -1.28886533e+00 5.06501973e-01
-6.36260509e-01 4.62714195e-01 5.07492363e-01 -3.43844116e-01
-1.24901342e+00 4.49944586e-01 1.73956648e-01 -3.59454513e-01
-3.44814569e-01 7.02011526e-01 -4.05899286e-01 1.83391735e-01
-6.15916476e-02 4.19374466e-01 6.56723976e-02 1.86581224e-01
1.02170885e-01 5.08506775e-01 3.90712857e-01 -4.81791735e-01
3.49792123e-01 5.34146607e-01 5.02929330e-01 -3.76526028e-01
-4.59147453e-01 -1.98907673e-01 -4.52613860e-01 -3.21734935e-01
1.03369641e+00 -9.17136610e-01 -1.39428627e+00 6.03114963e-01
-1.23970377e+00 -9.73766029e-01 -2.83743292e-01 8.25772524e-01
-1.42435277e+00 3.27217668e-01 -5.63072443e-01 -1.53424096e+00
-3.61337028e-02 -1.48534489e+00 8.95201504e-01 2.12891605e-02
1.67094827e-01 -7.34467745e-01 2.90967226e-01 1.01195931e-01
-7.81521276e-02 3.04333985e-01 8.47887099e-01 -3.62635285e-01
-9.17218566e-01 -1.61022261e-01 3.73123586e-01 5.62266596e-02
-1.36245415e-01 -4.26598638e-01 -5.99483669e-01 -5.88960767e-01
3.73776436e-01 -3.37507844e-01 4.25884157e-01 6.60887718e-01
5.40018499e-01 -4.65652645e-01 -2.57038563e-01 3.14309634e-03
1.33101809e+00 1.04891634e+00 3.32321942e-01 6.67947650e-01
2.33290762e-01 5.05560637e-01 1.51920831e+00 8.64968240e-01
7.88386881e-01 4.24317807e-01 1.11841297e+00 6.18250251e-01
6.00445807e-01 -3.65383893e-01 9.85807419e-01 5.27529538e-01
9.86549556e-02 -4.47298646e-01 -1.16490102e+00 4.89169329e-01
-2.31475425e+00 -9.71308529e-01 6.18039146e-02 2.05812311e+00
2.25619569e-01 2.50138760e-01 -7.73594007e-02 -1.56072035e-01
4.42571372e-01 6.06097952e-02 -9.37928855e-01 -5.96671760e-01
1.89846516e-01 -2.57514209e-01 8.07859540e-01 7.56779909e-01
-1.29835451e+00 1.13708639e+00 5.62900114e+00 4.72006410e-01
-8.23596776e-01 2.40005642e-01 -1.96312405e-02 2.63900220e-01
2.31519997e-01 1.58762977e-01 -8.25296462e-01 2.72940904e-01
1.24600816e+00 -1.98572367e-01 6.80132449e-01 9.28098440e-01
6.47302330e-01 -5.96545160e-01 -9.26922143e-01 7.22524107e-01
-5.47094226e-01 -8.31736147e-01 -2.82804608e-01 7.73746371e-02
7.03479111e-01 2.34149501e-01 1.78858235e-01 7.69563556e-01
1.17565310e+00 -8.90714288e-01 1.08521986e+00 5.75100541e-01
4.46762562e-01 -8.67223978e-01 9.75103974e-01 1.02313864e+00
-1.21474683e+00 -8.22127044e-01 -5.25363386e-01 -4.32765335e-01
5.97392499e-01 -2.64230698e-01 -1.06948459e+00 6.34709477e-01
3.49371046e-01 3.72975498e-01 1.47019356e-01 8.33271503e-01
-4.80355620e-01 2.47475281e-01 -5.39425462e-02 -2.50470459e-01
7.71636248e-01 -1.80475116e-01 6.55690610e-01 5.61595500e-01
3.66359204e-01 3.84700507e-01 8.37746084e-01 5.66095531e-01
6.69853687e-01 -7.71270990e-01 -9.20259237e-01 -1.40530750e-01
2.86944449e-01 6.49666131e-01 -6.91221118e-01 -2.95595348e-01
-2.43254080e-01 7.14557171e-01 -3.66751067e-02 4.63782698e-01
-9.20117557e-01 -2.53268075e-03 4.81339872e-01 -4.92040068e-01
2.82414287e-01 -8.13542247e-01 8.34924132e-02 -7.11060166e-01
-1.98689327e-01 -8.17735374e-01 3.93407434e-01 -9.82991278e-01
-8.11239302e-01 6.93990409e-01 1.61880493e-01 -1.37076008e+00
-7.77283072e-01 -6.00622058e-01 -2.93709576e-01 6.63797379e-01
-1.62750745e+00 -1.08841491e+00 -1.59802735e-02 6.89830601e-01
7.91243494e-01 -1.88807085e-01 9.25122797e-01 -5.42159021e-01
-4.41543758e-01 -5.66413879e-01 1.11060739e-01 -3.39637130e-01
1.15828224e-01 -1.07939851e+00 5.94979644e-01 8.99024606e-01
-3.32725048e-01 2.85399277e-02 7.25903451e-01 -1.11251974e+00
-2.04500079e+00 -1.39571607e+00 3.99708301e-01 -3.84583026e-01
6.05719209e-01 1.48514330e-01 -4.38834339e-01 1.08004594e+00
3.33868712e-01 -2.17895523e-01 1.66065637e-02 -6.25456929e-01
3.54579419e-01 1.68263927e-01 -1.14798045e+00 5.57060182e-01
7.12294996e-01 -2.73871630e-01 -6.86610937e-01 3.37616622e-01
1.06011343e+00 -4.52783614e-01 -4.91290718e-01 5.63824534e-01
3.69435698e-01 -3.84723395e-01 3.54764491e-01 -1.02162373e+00
-2.27723062e-01 -5.75368226e-01 -2.97242880e-01 -1.32902956e+00
-5.68624675e-01 -8.43380332e-01 -3.78751844e-01 3.18816602e-01
9.63212401e-02 -5.90027094e-01 5.59009969e-01 7.65887499e-01
-3.87218386e-01 -5.56018651e-01 -1.16762912e+00 -1.21252513e+00
-3.27121228e-01 -6.81810319e-01 7.34401226e-01 3.13926667e-01
-1.01651080e-01 -8.01919252e-02 -6.25652194e-01 7.30121315e-01
5.13334870e-01 1.23922907e-01 5.48804939e-01 -7.15582132e-01
-3.58620077e-01 2.53313065e-01 -9.45351124e-02 -9.91427064e-01
6.02346599e-01 -3.84506822e-01 8.74023557e-01 -1.84902990e+00
-1.68676674e-01 -2.70212263e-01 -4.38122034e-01 5.41412115e-01
3.83299917e-01 -9.02897477e-01 5.11791408e-01 1.18539758e-01
-1.17729449e+00 8.34091067e-01 1.51119924e+00 -4.78023708e-01
-3.67113054e-01 6.27426803e-01 6.92678168e-02 7.35103250e-01
1.13060784e+00 -5.56080103e-01 -6.83640301e-01 -6.68193042e-01
1.23491637e-01 1.05191457e+00 1.98116273e-01 -1.19072890e+00
3.97220284e-01 -1.10821962e+00 -2.26450890e-01 -7.82724559e-01
5.61550915e-01 -1.05098498e+00 2.16942519e-01 1.14817691e+00
-3.10594440e-01 6.13947034e-01 4.03353274e-01 9.87233818e-01
-3.38913612e-02 -3.21235269e-01 4.11762685e-01 -3.77329320e-01
-1.44477260e+00 3.73630673e-01 -1.18727291e+00 -3.44319016e-01
1.44752777e+00 9.43826735e-02 -1.26672789e-01 -6.28904879e-01
-9.67857957e-01 7.65801668e-01 -1.82577055e-02 1.97108150e-01
9.69955623e-01 -9.64618623e-01 -2.56601870e-01 1.35366708e-01
-2.46236250e-01 2.21108034e-01 3.27580899e-01 6.96728051e-01
-3.10685456e-01 8.62877190e-01 -5.48467696e-01 -2.97869325e-01
-7.60430098e-01 9.25509334e-01 2.73283243e-01 -6.53365791e-01
-3.42604756e-01 2.25125715e-01 -1.02272019e-01 -4.86522824e-01
4.13800269e-01 -7.36865938e-01 -2.62791693e-01 -4.58119184e-01
3.36392820e-01 4.23373044e-01 -3.13149661e-01 -3.19138229e-01
-2.04994261e-01 4.47070831e-03 4.61816818e-01 -7.54318893e-01
1.08058047e+00 -4.18560475e-01 6.60587475e-02 6.43265188e-01
1.42752081e-01 -8.23773682e-01 -1.71432006e+00 -1.11397766e-01
3.83743644e-01 5.67298867e-02 3.42918150e-02 -6.54272795e-01
-3.65614176e-01 8.99860203e-01 6.36632800e-01 9.51670483e-02
6.95224404e-01 -4.05283839e-01 8.39539409e-01 1.04665971e+00
1.33179522e+00 -1.18519914e+00 5.89003526e-02 1.20527506e+00
6.00367188e-01 -1.14846408e+00 -3.26486319e-01 3.29456598e-01
-1.20594382e+00 8.60749125e-01 7.36638784e-01 3.65143991e-03
3.09940606e-01 3.79751295e-01 -5.95581196e-02 1.06117263e-01
-1.32338178e+00 -4.15496737e-01 -4.14490074e-01 7.37475574e-01
-7.46104658e-01 4.95162457e-01 3.13144624e-01 6.60034895e-01
2.39560738e-01 3.98112148e-01 7.46820331e-01 1.55918634e+00
-7.15989232e-01 -7.94934154e-01 -6.30277216e-01 1.28394544e-01
-1.13369338e-01 4.25866574e-01 5.62493801e-02 7.52514839e-01
1.11430166e-02 1.26199651e+00 6.29629120e-02 -4.49724257e-01
2.20426008e-01 1.15822859e-01 5.04560769e-01 -6.56727314e-01
-7.69941416e-03 2.03768015e-02 -4.96614389e-02 -7.72657275e-01
-3.80070210e-01 -7.28569269e-01 -1.69471729e+00 -2.12386008e-02
-7.90457353e-02 -6.27262518e-02 6.17209852e-01 1.13163769e+00
2.34855980e-01 6.52934551e-01 5.48181653e-01 -7.25418150e-01
-1.26341152e+00 -6.73886061e-01 -6.61974907e-01 -4.07822371e-01
5.24140477e-01 -8.50236714e-01 1.47707075e-01 -2.02760145e-01] | [4.282928943634033, 2.123770236968994] |
80633492-d9cb-491a-9591-f7516f8f5d22 | robust-deep-learning-based-protein-sequence | null | null | https://www.biorxiv.org/content/10.1101/2022.06.03.494563v1 | https://www.biorxiv.org/content/10.1101/2022.06.03.494563v1.full.pdf | Robust deep learning based protein sequence design using ProteinMPNN | While deep learning has revolutionized protein structure prediction, almost all experimentally characterized de novo protein designs have been generated using physically based approaches such as Rosetta. Here we describe a deep learning based protein sequence design method, ProteinMPNN, with outstanding performance in both in silico and experimental tests. The amino acid sequence at different positions can be coupled between single or multiple chains, enabling application to a wide range of current protein design challenges. On native protein backbones, ProteinMPNN has a sequence recovery of 52.4%, compared to 32.9% for Rosetta. Incorporation of noise during training improves sequence recovery on protein structure models, and produces sequences which more robustly encode their structures as assessed using structure prediction algorithms. We demonstrate the broad utility and high accuracy of ProteinMPNN using X-ray crystallography, cryoEM and functional studies by rescuing previously failed designs, made using Rosetta or AlphaFold, of protein monomers, cyclic homo-oligomers, tetrahedral nanoparticles, and target binding proteins. | ['D. Baker', 'N. P. King', 'A. K. Bera', 'B. Sankaran', 'A. Kang', 'H. Nguyen', 'B. Koepnick', 'F. Chan', 'D. Tischer', 'S. Pellock', 'T. F. Huddy', 'P. J. Y. Leung', 'N. Bethel', 'R. J. de Haas', 'A. Courbet', 'B. I. M. Wicky', 'L. F. Milles', 'R. J. Ragotte', 'H. Bai', 'N. Bennett', 'I. Anishchenko', 'J. Dauparas'] | 2022-06-04 | null | null | null | biorxiv-2022-6 | ['protein-design', 'protein-function-prediction'] | ['medical', 'medical'] | [ 1.41878039e-01 2.69381180e-02 -7.96702281e-02 -2.04674274e-01
-5.71855009e-01 -7.49430418e-01 -8.16885568e-03 1.48095489e-02
-4.24486399e-01 1.47132742e+00 1.32609069e-01 -7.82063723e-01
1.79004535e-01 -5.21016955e-01 -1.16293442e+00 -1.05340850e+00
-3.95635962e-02 8.50254297e-01 -1.39023125e-01 -3.17972332e-01
5.86727671e-02 7.92113960e-01 -1.07694459e+00 5.80561161e-01
7.18508184e-01 4.51810330e-01 4.88076448e-01 4.02619302e-01
-1.51031092e-01 1.86152712e-01 -2.55248576e-01 -1.63479000e-01
1.56605288e-01 -3.71748179e-01 -6.28447115e-01 -2.49999061e-01
2.26437464e-01 -9.05984789e-02 4.32405323e-02 3.41426760e-01
7.70363748e-01 -7.06219673e-02 5.64238966e-01 -1.49596065e-01
-1.11652029e+00 1.80225536e-01 -5.00852346e-01 -1.99504718e-01
4.84024793e-01 7.59372711e-01 1.07295728e+00 -1.01060307e+00
9.34752226e-01 1.15840220e+00 1.03938854e+00 7.47773707e-01
-2.00587344e+00 -5.59423208e-01 -4.58813608e-01 8.96914601e-02
-9.31952238e-01 -2.85389692e-01 3.03347498e-01 -5.42176485e-01
1.90040147e+00 -3.39872539e-01 6.85410023e-01 1.26780033e+00
8.77497613e-01 3.08944494e-01 1.00641859e+00 -2.35693783e-01
3.56674582e-01 -7.68368304e-01 -1.13119401e-01 6.83968663e-01
2.14858517e-01 2.50183195e-01 -3.41307461e-01 -7.60320187e-01
3.99942905e-01 5.50331064e-02 -1.58748984e-01 -8.82683992e-01
-9.09508407e-01 8.77412915e-01 2.62499362e-01 -1.13297373e-01
-8.22842062e-01 -3.25971469e-02 3.87203485e-01 9.45003480e-02
-1.99196294e-01 8.61764193e-01 -1.05267525e+00 -3.54263075e-02
-6.43175721e-01 7.22310245e-01 7.47493029e-01 4.51883912e-01
9.20240521e-01 6.82480037e-02 5.92162788e-01 7.00483859e-01
2.36213952e-01 5.05101621e-01 5.25056779e-01 -7.98154771e-01
-2.55542129e-01 4.19111431e-01 5.22001088e-01 -4.63678211e-01
-6.86702609e-01 1.31999582e-01 -5.40474415e-01 4.30212468e-01
3.52283090e-01 -2.52185047e-01 -1.20166206e+00 1.73342395e+00
3.94184351e-01 -3.38085085e-01 3.67507190e-01 9.17502999e-01
6.08809948e-01 8.29310775e-01 4.01690185e-01 -4.08531934e-01
1.30519116e+00 -5.75741172e-01 -3.25829089e-01 6.83760494e-02
6.78910851e-01 -7.05690265e-01 6.67559087e-01 5.75690031e-01
-1.12997055e+00 -4.06379610e-01 -1.25023270e+00 -1.66445822e-01
-1.81139022e-01 -1.05323337e-01 6.76276386e-01 4.74138081e-01
-8.03020477e-01 9.88136470e-01 -8.66456270e-01 -3.33998948e-01
2.38719881e-01 7.75446534e-01 -6.67340159e-01 1.05974168e-01
-1.00695014e+00 1.17191780e+00 7.06250727e-01 -5.31301975e-01
-8.54109764e-01 -8.49618673e-01 -4.66643244e-01 2.47949474e-02
-3.87006328e-02 -8.49564850e-01 1.16793227e+00 -6.56113505e-01
-1.48969674e+00 9.75176930e-01 -3.85463089e-01 -8.48894477e-01
-2.24639684e-01 -1.77077472e-01 -2.11220786e-01 -2.96652280e-02
-1.34014249e-01 9.61274028e-01 2.42284611e-01 -8.99469376e-01
-1.12930000e-01 -5.17858982e-01 -2.84394711e-01 3.82033885e-02
6.25819027e-01 7.74851292e-02 6.50091887e-01 -5.10998547e-01
8.46090019e-02 -1.07264745e+00 -6.73632562e-01 4.17920277e-02
-1.00616455e-01 -1.76866993e-01 5.99835277e-01 -6.38946235e-01
3.44441444e-01 -1.33292627e+00 6.23079479e-01 1.67908713e-01
4.00958121e-01 8.53779078e-01 -1.88510120e-01 9.36809421e-01
-5.10199547e-01 -1.01965770e-01 -2.80584633e-01 5.10729313e-01
-3.81317675e-01 3.18058044e-01 -6.01610467e-02 4.14485544e-01
1.98883355e-01 1.25433743e+00 -5.37506521e-01 2.02639952e-01
2.14665234e-01 6.28750861e-01 -6.21878624e-01 1.02660663e-01
-8.73287737e-01 6.26401126e-01 -1.53378010e-01 5.73405087e-01
8.34286988e-01 -5.68045974e-01 1.05384040e+00 -3.37545604e-01
-1.51445374e-01 4.08145398e-01 -2.39559799e-01 1.67160034e+00
4.54174317e-02 1.13281041e-01 -4.08041142e-02 -7.31092155e-01
1.14715481e+00 2.78777778e-01 8.58094037e-01 -6.49054289e-01
-2.13118553e-01 2.33240888e-01 6.65175140e-01 -3.69956344e-01
1.38447741e-02 -6.54784381e-01 3.54226768e-01 6.71787977e-01
1.64737076e-01 5.15787542e-01 1.13651901e-01 -1.17612079e-01
1.16960287e+00 7.08824813e-01 4.18703139e-01 -4.58756506e-01
3.34897608e-01 5.70479810e-01 6.33761048e-01 3.42095971e-01
-5.78130372e-02 6.19614065e-01 4.40904886e-01 -1.14125836e+00
-2.11182284e+00 -1.02768195e+00 -4.58757505e-02 1.26726067e+00
-3.72310698e-01 -4.92224097e-01 -8.21941376e-01 -2.23085061e-01
3.47659528e-01 2.80776709e-01 -2.96523631e-01 -3.37137014e-01
-9.77465570e-01 -1.25818872e+00 5.89202106e-01 2.85246432e-01
-2.95664489e-01 -1.33347929e+00 -4.62102056e-01 9.17399764e-01
3.20537508e-01 -3.98858249e-01 -1.30838707e-01 7.60584712e-01
-7.08076298e-01 -1.24366999e+00 -7.21434772e-01 -8.94788742e-01
7.35914856e-02 7.33005553e-02 1.09871554e+00 -1.58665136e-01
-4.62179095e-01 -4.12624270e-01 4.76467889e-03 -5.56122623e-02
-7.34243333e-01 -1.10596223e-02 6.11182272e-01 -5.78252554e-01
9.46070552e-01 -1.04283619e+00 -7.85778463e-01 2.70592928e-01
-5.67373335e-01 2.19939396e-01 9.53894258e-01 1.09880304e+00
9.28163886e-01 -8.28372896e-01 9.05183613e-01 -8.43281627e-01
7.32151330e-01 -3.15109998e-01 -4.04114455e-01 3.33556384e-01
-8.83564293e-01 6.80456161e-01 7.58655190e-01 -4.44627404e-01
-9.38725412e-01 4.82676685e-01 -5.99923193e-01 -1.32961541e-01
-2.64891475e-01 4.31457967e-01 -2.79359400e-01 -2.50185102e-01
1.11760283e+00 5.73873341e-01 6.97441578e-01 -6.79210246e-01
3.13636750e-01 -4.31556429e-04 3.32856059e-01 -9.62244630e-01
2.17161164e-01 1.29818842e-01 3.72573324e-02 -7.71955132e-01
-2.66634136e-01 -8.99906456e-02 -5.33263326e-01 4.65269715e-01
8.40908587e-01 -6.84585333e-01 -1.50217295e+00 3.20980132e-01
-1.27113771e+00 -1.29871875e-01 2.65011430e-01 2.75320321e-01
-8.18804801e-01 6.77795887e-01 -7.52306998e-01 -1.16054796e-01
-6.88492358e-01 -1.56358266e+00 8.36735070e-01 -1.27196625e-01
-5.58615625e-01 -6.60556078e-01 5.58715582e-01 3.69408041e-01
2.21607834e-01 3.71503592e-01 1.51137841e+00 -5.58605492e-01
-3.92951995e-01 3.68861794e-01 -1.29764751e-01 1.32848220e-02
2.30323642e-01 -1.27505362e-01 -3.69795471e-01 -3.91122311e-01
-4.75189656e-01 -8.69421959e-01 8.85728836e-01 5.11306584e-01
5.85572004e-01 -2.83505380e-01 -4.70816135e-01 5.24098575e-01
1.51946151e+00 7.26243615e-01 8.77945483e-01 4.40127075e-01
8.53541017e-01 4.82414484e-01 2.05013931e-01 2.29249179e-01
-3.06261063e-01 6.20007634e-01 2.45084435e-01 2.17762426e-01
1.67617872e-01 -2.16073543e-01 2.91568637e-01 3.14208359e-01
-2.08111614e-01 -2.93342501e-01 -1.13027489e+00 -9.23660696e-02
-1.86782956e+00 -9.01678681e-01 -3.75291780e-02 1.70168817e+00
1.37371516e+00 -6.42413795e-02 2.67244130e-01 -4.90523875e-01
5.43850839e-01 -2.71613955e-01 -1.29659581e+00 -5.66633284e-01
-5.26308000e-01 7.64614761e-01 5.71684241e-01 2.75454134e-01
-6.67740643e-01 8.42509925e-01 7.91806030e+00 5.32780230e-01
-1.16294551e+00 -2.28838727e-01 4.59832788e-01 -1.87936231e-01
-2.37959132e-01 2.27825046e-01 -6.99365199e-01 4.11623210e-01
1.21563101e+00 -4.12035687e-03 3.45955759e-01 7.53495812e-01
4.79341418e-01 2.53576130e-01 -1.09779406e+00 7.29507148e-01
-6.16791546e-01 -2.31174707e+00 1.74596965e-01 4.34262067e-01
7.62597322e-01 2.01187700e-01 -1.57926098e-01 -1.14300894e-02
6.14039063e-01 -1.44585490e+00 4.29766439e-02 4.26990539e-01
5.15308380e-01 -9.17111874e-01 3.30533981e-01 2.42284805e-01
-5.13514340e-01 2.39218846e-01 -6.59226894e-01 -8.55986997e-02
2.55035013e-01 4.64688659e-01 -1.13801086e+00 4.10511233e-02
4.23201680e-01 3.71794015e-01 1.39270946e-01 3.72535586e-01
3.16797793e-01 5.20727098e-01 -2.68925101e-01 -1.16571106e-01
2.38922790e-01 -6.03968680e-01 1.53325990e-01 1.00657475e+00
-2.69952804e-01 3.82103503e-01 3.64765376e-01 9.13053632e-01
-2.04620779e-01 1.30777851e-01 -3.52742314e-01 -2.64658034e-01
2.12124303e-01 8.71855199e-01 -1.59095094e-01 -1.48244634e-01
-4.30928737e-01 5.52946389e-01 5.17351329e-01 3.99671614e-01
-7.02221513e-01 -2.79718250e-01 1.28154325e+00 1.95699289e-01
4.61131513e-01 -2.65072167e-01 2.39881463e-02 -8.02795589e-01
-1.85091078e-01 -1.36560476e+00 -2.56344467e-01 -9.39364433e-01
-1.16156757e+00 2.87548639e-02 -5.60159147e-01 -3.30639780e-01
-1.82371363e-01 -1.09194481e+00 -1.95025668e-01 1.17218208e+00
-6.83231294e-01 -1.01690900e+00 5.60099244e-01 -3.06544125e-01
2.53131896e-01 -4.26988184e-01 1.01012266e+00 8.58131722e-02
-3.59832287e-01 4.67940032e-01 8.73290181e-01 -5.77129662e-01
7.78386772e-01 -1.05868018e+00 9.10870910e-01 1.52581424e-01
-5.81924677e-01 9.76414859e-01 1.12051117e+00 -1.00940192e+00
-1.59789014e+00 -9.24176991e-01 5.25679767e-01 -2.68865317e-01
3.84867281e-01 -1.85432062e-01 -1.33195567e+00 5.88737667e-01
2.57115722e-01 -6.86371744e-01 1.09720993e+00 -4.14093472e-02
-3.71995300e-01 5.39513409e-01 -1.31727874e+00 5.01143634e-01
9.87943828e-01 -2.53081322e-01 -5.61873436e-01 5.62175572e-01
8.15241218e-01 -2.45610923e-01 -1.34043908e+00 5.81534147e-01
9.80146766e-01 -9.40109789e-01 1.28886080e+00 -1.53324115e+00
4.33547974e-01 -2.49892056e-01 -1.19516023e-01 -9.36406732e-01
-8.38265121e-01 -6.36611104e-01 -4.00193900e-01 2.33649909e-01
6.89354718e-01 -4.90964472e-01 1.42640758e+00 5.01615882e-01
-3.00318360e-01 -1.18781674e+00 -6.87748194e-01 -5.96814811e-01
6.65467083e-01 3.72533143e-01 5.50968945e-01 7.11935401e-01
9.85373482e-02 6.42831564e-01 -5.37558258e-01 -4.05584931e-01
3.26789081e-01 3.41492593e-01 6.98100567e-01 -1.13766754e+00
-5.06679952e-01 4.33631055e-03 -2.28951350e-01 -1.26262796e+00
2.09670722e-01 -9.24955010e-01 -1.98924288e-01 -1.32930899e+00
3.66681516e-01 -7.32633173e-02 1.01595111e-01 6.50912583e-01
2.67056197e-01 -9.42383781e-02 -4.64962214e-01 3.63950014e-01
-2.92175025e-01 6.76108420e-01 1.11479950e+00 -5.02512883e-03
-1.73207968e-01 -3.53236467e-01 -5.58547497e-01 3.20765406e-01
1.02937961e+00 -4.18561250e-01 -1.22901045e-01 5.74041829e-02
4.03725564e-01 -5.68457730e-02 -2.17700452e-02 -7.96745718e-01
-3.96110982e-01 -4.52748269e-01 7.18916357e-01 -7.39791632e-01
4.04189378e-01 -4.69002545e-01 8.31260204e-01 7.82788336e-01
-3.12776566e-01 2.84754813e-01 3.25980186e-01 5.42563915e-01
5.40821671e-01 2.75025796e-02 1.18352044e+00 -4.84590232e-01
-4.11989808e-01 2.55147010e-01 -9.78425264e-01 -2.40533575e-01
8.84359300e-01 -4.82101530e-01 -2.33380392e-01 1.46153823e-01
-1.19117975e+00 -2.19092742e-02 1.04292274e+00 2.02688742e-02
6.21422827e-01 -1.01653063e+00 -4.49900955e-01 1.95402622e-01
6.89774007e-02 -3.16914648e-01 1.77041113e-01 4.02540147e-01
-1.39540410e+00 7.46614099e-01 -7.13242829e-01 -6.96892500e-01
-1.18991673e+00 1.02411544e+00 5.83811998e-01 -3.37581895e-02
-5.36214530e-01 3.43451321e-01 3.75401318e-01 -8.31332207e-01
-2.69793332e-01 1.60543621e-01 2.72087872e-01 -5.97507358e-01
4.24964339e-01 1.02113150e-01 1.90868065e-01 -5.99268258e-01
-2.19486356e-01 3.42334688e-01 -6.49272621e-01 4.97341245e-01
1.48520637e+00 2.35723719e-01 -2.07963899e-01 -2.61192530e-01
1.15654778e+00 -5.63759804e-01 -1.50588012e+00 -3.08576792e-01
1.65766582e-01 3.25254381e-01 -5.72568774e-01 -1.06792188e+00
-4.98607606e-01 4.18470323e-01 9.20131683e-01 -6.22920692e-01
4.59677458e-01 -3.23325880e-02 1.23760009e+00 1.18173146e+00
4.65610266e-01 -7.59245157e-01 -1.02465622e-01 5.62108934e-01
5.37685275e-01 -1.10991347e+00 1.33073777e-01 2.66124487e-01
-1.63101703e-01 1.06896627e+00 5.55552661e-01 -1.74801618e-01
2.78944999e-01 3.27331692e-01 2.24430673e-02 -3.87530029e-01
-1.06719339e+00 2.02938810e-01 -3.39663476e-01 9.06676888e-01
7.64682233e-01 3.28105576e-02 -4.25879568e-01 2.54909128e-01
4.84971590e-02 -2.76281852e-02 2.22605854e-01 1.21757925e+00
-9.52741861e-01 -1.60168207e+00 -3.37874621e-01 2.36936554e-01
-6.90551460e-01 -2.33043790e-01 -8.42244327e-01 4.10373867e-01
-1.09812796e-01 4.27465588e-01 -2.41570100e-01 -2.98658043e-01
7.50789493e-02 5.00601888e-01 8.04459453e-01 -3.90804410e-01
-5.85936725e-01 -1.99241703e-03 1.19316079e-01 -2.25424290e-01
-3.34836364e-01 -4.94226396e-01 -1.66422641e+00 -7.01899886e-01
-3.54784071e-01 3.80004883e-01 6.25740170e-01 7.83400953e-01
1.10072696e+00 8.27832147e-02 1.56028092e-01 -1.11920500e+00
-4.70547467e-01 -8.38784158e-01 -3.10919315e-01 2.81639934e-01
1.76485866e-01 -6.52254581e-01 3.18730712e-01 1.61150381e-01] | [4.729691505432129, 5.5946550369262695] |
584b26ef-0227-41f9-8b8f-bcf7a60ac405 | latent-variable-nested-set-transformers | 2104.00563 | null | https://arxiv.org/abs/2104.00563v3 | https://arxiv.org/pdf/2104.00563v3.pdf | Latent Variable Sequential Set Transformers For Joint Multi-Agent Motion Prediction | Robust multi-agent trajectory prediction is essential for the safe control of robotic systems. A major challenge is to efficiently learn a representation that approximates the true joint distribution of contextual, social, and temporal information to enable planning. We propose Latent Variable Sequential Set Transformers which are encoder-decoder architectures that generate scene-consistent multi-agent trajectories. We refer to these architectures as "AutoBots". The encoder is a stack of interleaved temporal and social multi-head self-attention (MHSA) modules which alternately perform equivariant processing across the temporal and social dimensions. The decoder employs learnable seed parameters in combination with temporal and social MHSA modules allowing it to perform inference over the entire future scene in a single forward pass efficiently. AutoBots can produce either the trajectory of one ego-agent or a distribution over the future trajectories for all agents in the scene. For the single-agent prediction case, our model achieves top results on the global nuScenes vehicle motion prediction leaderboard, and produces strong results on the Argoverse vehicle prediction challenge. In the multi-agent setting, we evaluate on the synthetic partition of TrajNet++ dataset to showcase the model's socially-consistent predictions. We also demonstrate our model on general sequences of sets and provide illustrative experiments modelling the sequential structure of the multiple strokes that make up symbols in the Omniglot data. A distinguishing feature of AutoBots is that all models are trainable on a single desktop GPU (1080 Ti) in under 48h. | ["Jim Aldon D'Souza", 'Christopher Pal', 'Felix Heide', 'Samira Ebrahimi Kahou', 'Martin Weiss', 'Felipe Codevilla', 'Florian Golemo', 'Roger Girgis'] | 2021-02-19 | latent-variable-sequential-set-transformers | https://openreview.net/forum?id=Dup_dDqkZC5 | https://openreview.net/pdf?id=Dup_dDqkZC5 | iclr-2022-4 | ['trajectory-modeling'] | ['time-series'] | [-2.02208802e-01 4.17401016e-01 -1.47124052e-01 -2.52900422e-01
-7.63714373e-01 -5.69886923e-01 1.11378801e+00 -1.54156595e-01
-3.75497758e-01 6.55146658e-01 4.76219863e-01 -2.99701542e-01
-7.39805326e-02 -5.87180078e-01 -1.14067447e+00 -6.92134738e-01
-4.00152296e-01 1.02309442e+00 3.57535660e-01 -3.69494557e-01
-1.73677102e-01 3.47012818e-01 -1.52908170e+00 3.37738365e-01
5.07465065e-01 4.96818036e-01 5.96181095e-01 1.09045613e+00
3.75182152e-01 1.34717512e+00 -2.93849200e-01 -2.14507878e-01
4.38126892e-01 -2.16193661e-01 -7.76635885e-01 1.40017405e-01
7.42861778e-02 -7.03977227e-01 -6.86360955e-01 6.10771954e-01
1.80048838e-01 3.36235821e-01 8.09849501e-01 -1.73694289e+00
-2.20602512e-01 5.97223699e-01 -3.02555144e-01 -9.03280526e-02
2.87818816e-02 7.90599704e-01 9.00945663e-01 -4.10227120e-01
9.86263692e-01 1.58226967e+00 4.81079757e-01 4.32239681e-01
-1.07067156e+00 -3.84337425e-01 3.30918640e-01 4.46236163e-01
-1.04423165e+00 -5.86599886e-01 2.45517135e-01 -6.57977521e-01
1.12377214e+00 -1.56495422e-02 6.02582693e-01 1.38986719e+00
6.21340156e-01 9.63748097e-01 5.52610695e-01 3.06893021e-01
3.44696701e-01 -2.37326533e-01 -1.20388970e-01 8.53280187e-01
-2.28254497e-01 4.40431565e-01 -4.22922313e-01 -2.71153659e-01
5.21169126e-01 -4.68254089e-03 3.59265506e-01 -3.37594599e-01
-1.52330017e+00 7.83426166e-01 3.54770452e-01 -4.22993600e-01
-7.05854535e-01 8.21576059e-01 3.28871608e-01 2.99198240e-01
3.06599170e-01 2.57516623e-01 -3.31983536e-01 -2.84198850e-01
-5.27388394e-01 8.41489136e-01 8.37204576e-01 1.39935446e+00
7.35020161e-01 5.72563261e-02 -3.94997299e-01 2.79127091e-01
2.55720049e-01 7.20900834e-01 2.73742452e-02 -1.78685772e+00
4.51110840e-01 1.61949083e-01 6.31108105e-01 -7.53644586e-01
-5.82992971e-01 -1.02664858e-01 -3.79808187e-01 4.61338788e-01
2.79475212e-01 -5.58001995e-01 -8.29811811e-01 1.80147958e+00
5.32414496e-01 5.51458120e-01 3.70210707e-01 9.47799504e-01
3.99180800e-01 9.89283562e-01 1.72015950e-01 7.78300315e-02
1.08819616e+00 -1.47058439e+00 -2.98528045e-01 -4.29969132e-01
7.85759091e-01 -4.69198167e-01 5.24795890e-01 -2.51625162e-02
-1.06701696e+00 -3.47664177e-01 -8.69752586e-01 -8.74048546e-02
-3.22737880e-02 -5.01838289e-02 6.47341847e-01 -2.07854673e-01
-1.11643624e+00 6.42455339e-01 -1.33067262e+00 -3.16391975e-01
3.24489772e-01 3.11873049e-01 -3.56537670e-01 -3.51541750e-02
-6.09525859e-01 1.18055105e+00 1.62577629e-01 -1.81594223e-01
-1.73950815e+00 -4.79814112e-01 -7.79481053e-01 -6.25471920e-02
4.05167222e-01 -8.91781986e-01 1.62367511e+00 -7.75155187e-01
-1.63924813e+00 2.94911027e-01 -3.57110769e-01 -1.02394795e+00
7.14830875e-01 -1.36937434e-02 -8.59931018e-03 6.65361434e-02
3.13973725e-01 1.09640419e+00 7.69877374e-01 -1.17524302e+00
-9.61591244e-01 4.43210304e-02 2.31602907e-01 5.43264687e-01
3.90953362e-01 4.76756971e-03 -3.95005226e-01 -1.01093903e-01
-3.96265268e-01 -1.47807121e+00 -6.05350077e-01 -9.33514014e-02
-3.73922795e-01 -3.16854745e-01 8.30370426e-01 -5.14874458e-01
3.83974791e-01 -2.06504679e+00 6.11781895e-01 -1.49874717e-01
1.10875733e-01 -2.17400581e-01 -5.36943495e-01 6.65235221e-01
2.83286244e-01 -2.96790868e-01 -6.14337847e-02 -9.21870053e-01
3.10610831e-01 5.06946623e-01 -4.59148854e-01 7.73451567e-01
6.45416602e-02 1.05454743e+00 -1.13228655e+00 -3.59656721e-01
2.55126178e-01 1.88154951e-01 -7.81182945e-01 3.15960228e-01
-8.76733184e-01 6.51037872e-01 -4.90160465e-01 1.09847030e-02
2.37865552e-01 -1.93806753e-01 1.28759801e-01 2.57127106e-01
-3.52627397e-01 2.39521056e-01 -7.76769817e-01 1.74727798e+00
-3.27674806e-01 8.00112665e-01 2.07447231e-01 -3.78532499e-01
4.14392501e-01 2.64645576e-01 6.32638514e-01 -4.92972344e-01
9.10805464e-02 -5.69241755e-02 -9.78217125e-02 -4.56308633e-01
8.69215310e-01 1.48363546e-01 -3.56446296e-01 6.56921566e-01
-4.79697250e-02 -1.54504597e-01 2.08977297e-01 5.33299863e-01
1.43496406e+00 4.25314277e-01 -2.30315760e-01 -1.11286104e-01
-1.68632850e-01 5.74294329e-01 4.64392811e-01 7.02581108e-01
-1.45122349e-01 3.18545103e-01 6.60666108e-01 -5.91882408e-01
-1.56818473e+00 -9.95000184e-01 5.74181616e-01 1.12853730e+00
4.12944049e-01 -4.81248051e-01 -5.56276202e-01 -4.32634979e-01
3.11468802e-02 1.16841805e+00 -5.57781935e-01 -6.57724738e-02
-6.52191758e-01 -4.72328782e-01 3.59155029e-01 3.29091012e-01
7.84623548e-02 -1.02417898e+00 -9.53569710e-01 3.63800645e-01
-2.01025456e-01 -1.34259498e+00 -4.92024064e-01 -2.22200826e-02
-1.78105354e-01 -9.42789078e-01 -1.81745037e-01 -5.55487216e-01
4.16513413e-01 3.13337386e-01 9.07588601e-01 -2.28245974e-01
5.59543632e-02 3.66824567e-01 -1.81142330e-01 -3.63630712e-01
-8.66508365e-01 5.08392639e-02 1.67131752e-01 -3.38193104e-02
-1.83073968e-01 -5.11628449e-01 -4.84156311e-01 3.75165731e-01
-3.15301806e-01 7.85198569e-01 2.79873341e-01 6.00713551e-01
4.63720977e-01 -2.74257064e-01 3.66638780e-01 -3.61151338e-01
2.76989698e-01 -1.08544564e+00 -8.73906016e-01 -1.74588814e-01
2.28323182e-03 1.82847857e-01 7.14523852e-01 -3.43266875e-01
-9.18360412e-01 2.29201302e-01 1.24929130e-01 -6.61715746e-01
-1.17905058e-01 2.70569593e-01 2.68965423e-01 4.22493577e-01
4.27031964e-01 1.93705961e-01 4.34154361e-01 -1.70350879e-01
5.60906172e-01 4.02155429e-01 5.87878525e-01 -5.85115433e-01
6.54697239e-01 7.38075614e-01 1.54087812e-01 -8.11537266e-01
-3.26481014e-01 -1.00991219e-01 -2.80219287e-01 -3.79241347e-01
1.02133381e+00 -1.20464885e+00 -1.18760002e+00 4.19970661e-01
-1.48569989e+00 -1.21538103e+00 -4.17066395e-01 4.82630253e-01
-1.20055735e+00 -1.13193758e-01 -5.85351348e-01 -7.66986728e-01
5.08250035e-02 -1.44501591e+00 1.25810218e+00 -1.08039387e-01
-2.99018949e-01 -7.71989524e-01 1.70618072e-01 6.16254807e-02
1.88295588e-01 2.39071488e-01 7.84829676e-01 -4.68363166e-01
-1.15584552e+00 8.46227556e-02 1.50440838e-02 -1.74910143e-01
-3.58547539e-01 -6.00559488e-02 -5.81762910e-01 -4.00594056e-01
-5.15651584e-01 -2.75456905e-01 7.46627927e-01 2.95802742e-01
6.82860911e-01 -5.92178822e-01 -6.28706872e-01 5.15805900e-01
9.81438994e-01 1.16771646e-01 3.74525130e-01 1.32059485e-01
7.24263549e-01 6.35228753e-01 7.86277533e-01 4.75488722e-01
1.11230147e+00 9.04885352e-01 8.40578973e-01 2.09732652e-01
-2.23558798e-01 -4.13613141e-01 7.88102210e-01 4.32663858e-01
5.31591140e-02 -5.48034370e-01 -9.22996223e-01 8.60048771e-01
-2.42997503e+00 -1.33712924e+00 -1.01384766e-01 1.74674475e+00
2.74699181e-01 -1.49244264e-01 3.63457918e-01 -7.14974999e-01
3.61979753e-01 2.56538332e-01 -7.24410295e-01 -1.85868964e-01
1.78373943e-03 -5.36295235e-01 8.75451803e-01 9.96206284e-01
-1.05247998e+00 1.27612627e+00 6.22851658e+00 6.47325873e-01
-7.76310980e-01 2.46977329e-01 5.13161719e-01 -6.11316919e-01
-3.59702617e-01 7.33956918e-02 -8.76747012e-01 5.17882347e-01
1.33788610e+00 -1.24407679e-01 7.68047690e-01 7.89822102e-01
5.97965479e-01 -2.94050835e-02 -1.32019949e+00 5.07466018e-01
-2.95531482e-01 -1.72307575e+00 -1.89871758e-01 2.39711061e-01
8.11156154e-01 9.66672242e-01 -5.25294963e-05 3.38889509e-01
1.27121186e+00 -1.04370773e+00 1.20753217e+00 5.59027016e-01
4.98695135e-01 -7.30573118e-01 3.81307691e-01 8.01548243e-01
-8.98582160e-01 -3.44633460e-01 -2.02236667e-01 -1.46129370e-01
6.07985735e-01 -3.37776661e-01 -1.26486349e+00 3.81804466e-01
4.38785017e-01 9.59541738e-01 -7.72424787e-02 7.34007418e-01
3.02732121e-02 3.82976890e-01 -6.58119977e-01 -1.00363577e-02
7.57674932e-01 -1.80731982e-01 1.07270181e+00 8.94076049e-01
3.98863107e-01 5.42520918e-02 4.62850004e-01 6.88771546e-01
2.72880197e-01 -6.13650858e-01 -7.21249342e-01 -2.45562065e-02
4.93431002e-01 1.01362729e+00 -4.07476246e-01 -3.25223118e-01
-1.83968410e-01 8.52650404e-01 5.22502482e-01 4.85319704e-01
-1.07605624e+00 3.38705570e-01 1.20752919e+00 -2.12619994e-02
4.84544337e-01 -6.48429751e-01 -1.06306225e-01 -8.18593204e-01
-3.37467134e-01 -7.09733427e-01 9.61460769e-02 -8.34175527e-01
-7.37363636e-01 5.84929347e-01 1.31806046e-01 -1.12484670e+00
-8.25730979e-01 -2.80040175e-01 -6.27169073e-01 6.28054917e-01
-1.29105639e+00 -1.52253342e+00 -1.08236335e-01 4.14586842e-01
7.30063558e-01 -4.40162212e-01 6.84576631e-01 -8.83909166e-02
-4.69037294e-01 4.78808954e-02 1.63524657e-01 -2.82305807e-01
3.18580389e-01 -1.03903270e+00 9.74054992e-01 7.02489257e-01
-1.56567365e-01 6.56780675e-02 1.16016924e+00 -7.47706354e-01
-1.72044218e+00 -1.52616799e+00 6.77828133e-01 -7.95615554e-01
7.82763898e-01 -2.67622858e-01 -3.59914601e-01 1.31922233e+00
4.21217591e-01 -1.46190867e-01 -1.00575173e-02 -2.33992249e-01
6.45651156e-03 2.04499155e-01 -7.12391853e-01 1.06682575e+00
1.13218307e+00 -1.56289726e-01 -1.67768553e-01 6.15129888e-01
1.05701375e+00 -6.58956528e-01 -5.14489591e-01 8.25769901e-02
4.82553154e-01 -8.20074022e-01 8.30840170e-01 -7.04850912e-01
6.26690447e-01 -4.63247657e-01 -2.41661638e-01 -1.45293760e+00
-3.55130970e-01 -9.95303452e-01 -2.41718262e-01 5.98057389e-01
4.17402625e-01 -4.30958629e-01 7.89878249e-01 6.16072536e-01
-6.02146983e-01 -6.98087275e-01 -9.88350987e-01 -6.35088623e-01
-7.31089115e-02 -5.75669825e-01 8.51294398e-01 2.98209935e-01
-9.79636461e-02 3.78817111e-01 -7.06434727e-01 4.04443532e-01
7.34869897e-01 1.54708967e-01 1.29006219e+00 -5.32871425e-01
-5.56035042e-01 -2.45335042e-01 -2.14247391e-01 -1.38818383e+00
6.11320734e-01 -8.41272533e-01 3.70370835e-01 -1.58232605e+00
1.85904384e-01 -5.36168098e-01 3.39087605e-01 4.66041118e-01
2.78058380e-01 -2.51178056e-01 4.52572107e-01 2.79189229e-01
-1.02624428e+00 9.08994079e-01 1.20427811e+00 -1.58975676e-01
-1.57852039e-01 -7.38037098e-03 -2.60389119e-01 7.58192778e-01
6.14374161e-01 -4.72408235e-01 -6.63158536e-01 -9.08677340e-01
8.01661238e-02 6.02943838e-01 7.55650103e-01 -9.75708783e-01
5.80290616e-01 -6.32005870e-01 -1.86335459e-01 -7.17812598e-01
9.12484586e-01 -7.22884834e-01 4.82531399e-01 4.42679375e-01
-3.61611694e-01 3.18677098e-01 1.37328655e-01 8.12901855e-01
2.91037410e-01 3.51199657e-01 5.80340326e-01 -1.14390902e-01
-8.36904645e-01 6.29198074e-01 -7.52649963e-01 -2.10834801e-01
1.47178233e+00 1.51323020e-01 -4.57409352e-01 -7.64597476e-01
-6.05799019e-01 8.85074735e-01 6.76413715e-01 6.53042018e-01
4.05250043e-01 -1.14789689e+00 -9.59510386e-01 2.84851994e-02
-5.22186346e-02 1.57436311e-01 3.90517294e-01 6.71930671e-01
-5.40392339e-01 2.55397588e-01 -2.23794669e-01 -6.72612011e-01
-1.05660808e+00 4.70010966e-01 3.96936208e-01 -1.10631667e-01
-8.90357077e-01 7.00771213e-01 2.64898121e-01 -4.25909996e-01
3.45314741e-02 -1.26449555e-01 -8.38695318e-02 -2.91527450e-01
4.91070628e-01 5.41277111e-01 -5.27653992e-01 -1.00285661e+00
-1.58685327e-01 -8.89227986e-02 5.09808362e-02 -4.94427770e-01
1.48906016e+00 -2.71785349e-01 4.86869551e-02 3.73784482e-01
1.08051908e+00 -4.34700817e-01 -2.15934372e+00 3.46410125e-02
-2.63946801e-01 -4.31579584e-03 -1.50253400e-01 -4.53642219e-01
-5.66815078e-01 5.34020185e-01 7.02737123e-02 -4.07751687e-02
2.68090934e-01 2.00918317e-01 9.07735825e-01 4.49849308e-01
7.66746104e-01 -9.67903018e-01 -9.08840001e-02 8.91439080e-01
9.68591452e-01 -1.06774867e+00 -2.73404568e-01 -7.59943798e-02
-1.05541205e+00 6.07272387e-01 5.88678837e-01 -4.08241332e-01
2.61689663e-01 4.50060457e-01 -2.79158682e-01 -1.88537300e-01
-1.59314704e+00 -3.45411412e-02 -1.53386950e-01 4.26113367e-01
-3.57845843e-01 3.47319007e-01 3.94606531e-01 3.81064922e-01
-4.11458552e-01 -1.66405499e-01 6.42526031e-01 6.61096454e-01
-3.50499302e-01 -7.71468222e-01 4.20809723e-02 3.24477494e-01
1.28715992e-01 3.44441414e-01 7.74044245e-02 5.34042656e-01
-5.48877269e-02 1.01261139e+00 4.20443803e-01 -5.98733604e-01
8.76264945e-02 -3.03149879e-01 3.94371778e-01 -4.92015600e-01
-4.37622011e-01 -1.98580116e-01 5.24795711e-01 -9.38616455e-01
-8.42988417e-02 -1.01318836e+00 -1.41275811e+00 -8.21845531e-01
4.03230458e-01 8.47020075e-02 6.09432101e-01 1.09782076e+00
7.97220945e-01 5.74723899e-01 3.99192005e-01 -1.44544625e+00
-5.65984428e-01 -8.24557364e-01 -1.18824996e-01 3.39368641e-01
6.97252870e-01 -5.36490023e-01 -1.07086692e-02 2.98556536e-02] | [5.843475818634033, 0.783833920955658] |
2a25aaa0-5465-4514-b7f0-b66017923b53 | loss-of-plasticity-in-continual-deep | 2303.07507 | null | https://arxiv.org/abs/2303.07507v1 | https://arxiv.org/pdf/2303.07507v1.pdf | Loss of Plasticity in Continual Deep Reinforcement Learning | The ability to learn continually is essential in a complex and changing world. In this paper, we characterize the behavior of canonical value-based deep reinforcement learning (RL) approaches under varying degrees of non-stationarity. In particular, we demonstrate that deep RL agents lose their ability to learn good policies when they cycle through a sequence of Atari 2600 games. This phenomenon is alluded to in prior work under various guises -- e.g., loss of plasticity, implicit under-parameterization, primacy bias, and capacity loss. We investigate this phenomenon closely at scale and analyze how the weights, gradients, and activations change over time in several experiments with varying dimensions (e.g., similarity between games, number of games, number of frames per game), with some experiments spanning 50 days and 2 billion environment interactions. Our analysis shows that the activation footprint of the network becomes sparser, contributing to the diminishing gradients. We investigate a remarkably simple mitigation strategy -- Concatenated ReLUs (CReLUs) activation function -- and demonstrate its effectiveness in facilitating continual learning in a changing environment. | ['Marlos C. Machado', 'Adam White', 'Joseph Modayil', 'Rosie Zhao', 'Zaheer Abbas'] | 2023-03-13 | null | null | null | null | ['atari-games'] | ['playing-games'] | [-1.72886997e-01 -1.52800933e-01 2.52517253e-01 9.69369859e-02
-2.18401596e-01 -7.19595015e-01 7.20046818e-01 -1.91526249e-01
-1.04761398e+00 1.19019902e+00 2.28943750e-01 -2.14665771e-01
-3.26897591e-01 -4.73158777e-01 -9.33822215e-01 -9.33290184e-01
-7.34742105e-01 2.02970952e-01 2.82620847e-01 -7.35111833e-01
3.56678933e-01 4.35109079e-01 -1.35382211e+00 -2.42179856e-01
4.51747715e-01 4.05287415e-01 1.29032284e-01 8.13358247e-01
4.15341675e-01 1.12701392e+00 -8.20555031e-01 -1.59744054e-01
6.21380210e-01 -4.28166956e-01 -4.46481705e-01 -1.63469017e-01
9.50754285e-02 -6.24435484e-01 -8.20518672e-01 8.19703341e-01
8.08398008e-01 5.69865704e-01 4.91017580e-01 -1.14181352e+00
-3.27781171e-01 7.66849935e-01 -7.61831284e-01 9.13402438e-01
-8.94051865e-02 6.65930986e-01 7.45837390e-01 -4.09663618e-01
5.55498004e-01 1.19155037e+00 7.22444952e-01 6.62737250e-01
-1.38204622e+00 -7.10206032e-01 2.69786298e-01 -7.07440376e-02
-1.19218338e+00 -5.71792424e-01 3.46530020e-01 -1.89642623e-01
1.21686125e+00 -3.04599404e-01 1.06988943e+00 1.19836891e+00
5.03501356e-01 4.78475899e-01 1.09433293e+00 -1.62507832e-01
5.30808389e-01 -3.06069851e-01 -3.43156129e-01 6.15008652e-01
3.63700271e-01 3.72039855e-01 -6.08023763e-01 8.12496170e-02
1.35326397e+00 -1.68674320e-01 -1.95281610e-01 -2.95404881e-01
-9.66785371e-01 6.36617959e-01 3.02783191e-01 2.71560222e-01
-4.82961059e-01 8.56658936e-01 3.82530898e-01 1.00321615e+00
1.45105958e-01 9.28863168e-01 -4.05777901e-01 -7.42095411e-01
-3.62562269e-01 5.09361207e-01 8.04747045e-01 6.52140558e-01
6.69730783e-01 6.07430100e-01 1.39831021e-01 6.90920711e-01
-2.21310928e-01 3.43021452e-01 8.25174630e-01 -1.53144968e+00
3.57396364e-01 -1.54014167e-04 2.34798476e-01 -9.43261027e-01
-6.46996915e-01 -7.02562690e-01 -6.22995377e-01 4.71621573e-01
7.48310208e-01 -8.29179406e-01 -4.94505703e-01 2.34091425e+00
-8.47419575e-02 2.47228980e-01 1.16002038e-01 8.13279212e-01
2.37464923e-02 3.67871910e-01 -4.07116562e-02 -2.43403599e-01
7.36030281e-01 -6.98996723e-01 -3.85786593e-01 -3.54159147e-01
3.76753598e-01 -6.94127753e-02 1.27598524e+00 4.29934204e-01
-1.38024712e+00 -2.47933388e-01 -9.51435864e-01 4.81314480e-01
-1.72568157e-01 -7.04378247e-01 5.48072398e-01 4.56585348e-01
-1.50426400e+00 9.44336712e-01 -7.17017174e-01 -4.80553716e-01
3.63295972e-01 4.79087591e-01 -1.91270471e-01 1.71436563e-01
-1.22502935e+00 8.68376315e-01 2.43085504e-01 -3.26687425e-01
-1.46780014e+00 -5.77243447e-01 -3.35206062e-01 2.72105902e-01
5.15018284e-01 -6.54406726e-01 1.35937738e+00 -1.28979433e+00
-1.59600377e+00 4.61137235e-01 4.93138760e-01 -6.86483562e-01
5.89427769e-01 -2.74298668e-01 1.62181761e-02 -4.71053980e-02
-2.38908485e-01 7.82079041e-01 9.00310218e-01 -1.22418582e+00
-3.46523970e-01 -3.34242553e-01 5.31872690e-01 9.45463777e-01
-4.19530302e-01 -2.70870149e-01 -2.56953031e-01 -7.62413442e-01
-3.92488122e-01 -1.03986096e+00 -3.62579077e-01 -3.13918531e-01
2.15776116e-01 1.26274318e-01 2.82174379e-01 -1.63286194e-01
1.02315855e+00 -2.24741125e+00 4.03891206e-01 2.11821735e-01
4.43096459e-01 -8.62445757e-02 -4.63887125e-01 5.42989910e-01
-4.24231291e-02 2.24600315e-01 -1.40948826e-02 3.28600407e-04
-8.51409212e-02 5.71025789e-01 -1.24447802e-02 5.11199594e-01
4.33560722e-02 9.90421355e-01 -1.02938664e+00 -1.00090005e-01
-2.09251404e-01 3.43504339e-01 -8.00611734e-01 1.03054687e-01
-1.91189289e-01 4.36157733e-01 -2.60536104e-01 2.06859633e-01
9.22066644e-02 -2.06233829e-01 3.51217926e-01 4.85907316e-01
-2.62715816e-01 3.04423310e-02 -9.84069943e-01 1.64154160e+00
-2.57309169e-01 7.74553180e-01 1.74400359e-01 -9.67385948e-01
5.70521832e-01 1.04861774e-01 5.66152453e-01 -1.24600112e+00
2.82202274e-01 4.10377346e-02 5.69852173e-01 -2.87391007e-01
4.83794272e-01 -8.34577754e-02 6.11993149e-02 7.53380954e-01
1.16994083e-01 -4.05255593e-02 4.01477844e-01 3.60086262e-01
1.62242162e+00 -1.81784227e-01 7.90910572e-02 -4.85782892e-01
-2.77613670e-01 -1.59974426e-01 3.23394507e-01 1.23636043e+00
-5.05727232e-01 9.62351337e-02 9.87409532e-01 -3.97506535e-01
-1.39374483e+00 -9.91448104e-01 3.68828207e-01 1.48244262e+00
5.36677279e-02 -1.45291150e-01 -7.45071709e-01 -1.20525889e-01
1.44642666e-01 4.52072710e-01 -9.00419831e-01 -5.77880383e-01
-7.42847085e-01 -8.87168944e-01 7.52663612e-01 2.96923906e-01
7.51108646e-01 -1.34570205e+00 -1.06092989e+00 3.58740479e-01
3.03034604e-01 -6.56975985e-01 -2.96692550e-01 4.62156564e-01
-9.75947142e-01 -8.47480237e-01 -8.42767060e-01 -5.21511257e-01
5.44849575e-01 3.38272631e-01 1.43227923e+00 2.92377561e-01
-1.22435242e-01 7.54718006e-01 -2.43477166e-01 -1.15194462e-01
-8.84167701e-02 8.96197483e-02 3.93922865e-01 -6.57308936e-01
-1.15773849e-01 -9.67352986e-01 -8.81819189e-01 2.00779542e-01
-9.36798096e-01 -3.46734852e-01 6.14853084e-01 9.46866989e-01
2.86614448e-01 2.47863010e-01 7.02605367e-01 -7.72448480e-01
1.22940075e+00 -8.12408924e-01 -5.34213483e-01 -1.24550678e-01
-6.17200315e-01 1.37222320e-01 6.59646332e-01 -6.51049733e-01
-9.30423915e-01 -4.20796365e-01 1.83651194e-01 -3.95517498e-01
1.54234558e-01 4.13288295e-01 4.33920801e-01 -1.41678885e-01
9.99141276e-01 3.05458456e-01 2.16231212e-01 -2.11253669e-02
4.61369932e-01 2.71469913e-03 3.57213885e-01 -8.20636034e-01
6.67437196e-01 1.76633880e-01 -2.97270507e-01 -6.39247239e-01
-2.39596665e-01 1.49131030e-01 -9.33252200e-02 -4.97836500e-01
3.46370041e-01 -1.16557550e+00 -8.91523004e-01 7.20910490e-01
-6.05483949e-01 -1.40658057e+00 -6.49214327e-01 3.99365991e-01
-8.02600980e-01 -8.15818757e-02 -8.72894704e-01 -5.84548950e-01
-1.21245319e-02 -8.24919343e-01 2.36593515e-01 5.57769895e-01
7.39003494e-02 -9.76016283e-01 5.86375237e-01 -2.47348085e-01
9.24250245e-01 -7.55118905e-03 6.69035733e-01 -3.37880701e-01
-2.86574721e-01 4.84700531e-01 -3.17298956e-02 -5.31901978e-03
-4.31003459e-02 -2.06688374e-01 -5.98537743e-01 -7.26496994e-01
-1.03078119e-01 -6.57991469e-01 7.92310953e-01 4.02907848e-01
1.01686382e+00 -4.67819363e-01 2.35724285e-01 5.17782867e-01
1.45454454e+00 3.61797273e-01 7.85329878e-01 6.05518341e-01
2.04924017e-01 3.22403669e-01 1.83053732e-01 8.54120553e-01
1.74729034e-01 3.82996291e-01 7.05353618e-01 9.97256767e-03
1.31363198e-01 -1.05562225e-01 5.44863105e-01 6.14559352e-01
-4.08163458e-01 -3.33691269e-01 -8.11899662e-01 3.65800351e-01
-1.83409703e+00 -1.19691622e+00 7.74350703e-01 2.07233191e+00
1.01176214e+00 4.15859699e-01 5.70932508e-01 -3.85462314e-01
4.76224393e-01 2.33172908e-01 -1.26809871e+00 -3.35367441e-01
-2.87538320e-01 1.36352116e-02 8.05313706e-01 4.07565922e-01
-5.21604061e-01 1.21998096e+00 7.63741016e+00 5.60353696e-01
-1.18517935e+00 1.05332583e-01 8.24595511e-01 -6.88710392e-01
-1.45034581e-01 -4.64092702e-01 -2.14148656e-01 4.74541873e-01
1.15281570e+00 -3.88252646e-01 1.30777693e+00 5.90673923e-01
2.31652185e-01 -3.38712305e-01 -6.61969304e-01 8.59788477e-01
2.68993387e-03 -1.17878115e+00 -4.15727228e-01 7.65741467e-02
8.17004144e-01 6.16553843e-01 4.45850909e-01 6.75948322e-01
1.18808699e+00 -1.05710125e+00 6.49095178e-01 4.55735236e-01
8.63484085e-01 -8.98195148e-01 2.65145510e-01 2.64446229e-01
-7.34942734e-01 -5.43009162e-01 -3.39838505e-01 -4.69274223e-01
-3.72309566e-01 1.03443740e-02 -3.40208173e-01 -2.45194241e-01
7.99451828e-01 4.15872574e-01 -6.09493732e-01 8.39746237e-01
5.22714332e-02 5.63564837e-01 -4.06341881e-01 -8.74107853e-02
3.77672255e-01 -1.69742391e-01 3.97406965e-01 8.38903964e-01
2.03092620e-01 3.39865327e-01 -2.11791918e-01 5.27536035e-01
-3.87116045e-01 -2.67863661e-01 -7.63300359e-01 -1.78021789e-01
5.69670498e-01 9.02105153e-01 -8.35093856e-01 -1.88598529e-01
-1.16167180e-01 7.76341498e-01 7.28151023e-01 6.70718133e-01
-8.54865313e-01 -5.27320430e-02 8.82072747e-01 7.78083727e-02
2.07391873e-01 -6.56294465e-01 1.46852583e-01 -1.09172809e+00
-3.42955798e-01 -1.16129804e+00 5.88600226e-02 -5.99845052e-01
-9.96027470e-01 2.83473402e-01 -1.28475189e-01 -6.87556446e-01
-3.53751630e-01 -1.73527956e-01 -6.09773576e-01 3.00595760e-01
-1.39999485e+00 -1.04586691e-01 -9.74708330e-03 9.80042696e-01
4.65554118e-01 -5.29810369e-01 3.58406126e-01 2.80195415e-01
-8.68735492e-01 6.50412321e-01 5.18034220e-01 -1.63674485e-02
6.23248041e-01 -1.25756669e+00 4.63911831e-01 5.81769764e-01
-3.33357379e-02 6.39931858e-01 8.13033044e-01 -3.73738140e-01
-1.58688068e+00 -7.60029137e-01 -2.08329201e-01 -1.75052941e-01
7.21798956e-01 -3.00292224e-01 -6.28322482e-01 6.05888605e-01
3.62812012e-01 -1.01169735e-01 5.01958191e-01 3.32839452e-02
-2.64178157e-01 -2.89469343e-02 -9.84410048e-01 1.10249281e+00
1.33011234e+00 -3.26798856e-01 -9.66705829e-02 -5.34802228e-02
6.58874571e-01 -3.09549868e-01 -6.19408786e-01 -1.30496114e-01
6.50401950e-01 -1.15153325e+00 8.75977576e-01 -8.54793251e-01
2.87145287e-01 4.37916666e-02 -1.26497909e-01 -1.73758900e+00
-5.71006000e-01 -9.37067389e-01 -1.36428267e-01 6.12574816e-01
3.12709183e-01 -6.63066387e-01 8.75366330e-01 4.01351511e-01
2.69556474e-02 -6.17717803e-01 -1.00686824e+00 -8.10593426e-01
4.26083177e-01 1.00313470e-01 2.62771428e-01 9.36099112e-01
1.72237068e-01 2.78697878e-01 -5.31226754e-01 -3.34272176e-01
2.53388226e-01 -6.68300211e-01 7.50408173e-01 -7.79924989e-01
-5.78547120e-01 -6.57714307e-01 -1.35002434e-01 -9.15502071e-01
-1.16276041e-01 -3.73720437e-01 -1.98983382e-02 -1.04405022e+00
1.18063919e-01 -5.36022723e-01 -6.25785232e-01 4.83195782e-01
1.58208422e-03 -1.17875841e-02 4.89961505e-01 3.86660516e-01
-1.05795145e+00 6.50135934e-01 1.30476344e+00 2.50230044e-01
-4.02953684e-01 -5.64659059e-01 -7.80274808e-01 4.53606576e-01
1.26706266e+00 -4.88964975e-01 -7.16675580e-01 -6.57114387e-01
7.62862504e-01 1.12780184e-01 8.20308030e-02 -1.41296792e+00
9.74583328e-02 -2.74201542e-01 3.43517125e-01 4.35053974e-01
2.81818032e-01 -6.91861868e-01 4.72313119e-03 7.70999074e-01
-5.90545177e-01 7.00597048e-01 4.60985094e-01 7.31195331e-01
3.40096444e-01 1.49948612e-01 7.57416189e-01 -4.65056360e-01
-6.99300289e-01 2.02493578e-01 -7.66608238e-01 6.90910876e-01
1.02620149e+00 -6.42228499e-02 -5.45209646e-01 -7.50695348e-01
-5.90014219e-01 4.00035560e-01 6.12596571e-01 2.14187995e-01
4.34678018e-01 -1.01453197e+00 -6.47053838e-01 5.08262329e-02
-3.85988861e-01 -3.05078834e-01 2.67208219e-01 7.29597569e-01
-6.20725095e-01 -1.31197661e-01 -8.27977717e-01 -1.66573822e-01
-7.40356565e-01 1.85962453e-01 8.84719670e-01 -3.90558749e-01
-5.04191279e-01 8.58879983e-01 1.46086410e-01 -2.60882992e-02
4.10410821e-01 7.32964799e-02 -2.16596276e-01 1.05258189e-01
4.06413436e-01 2.86111116e-01 -2.81485945e-01 -6.40798360e-02
3.85019630e-02 2.61648387e-01 -2.76697338e-01 -4.79144096e-01
1.43042850e+00 -1.93825036e-01 2.63933390e-01 4.18996096e-01
6.09260380e-01 -3.07510108e-01 -2.03102422e+00 -1.47002697e-01
-3.66304636e-01 -3.98203224e-01 -8.09005052e-02 -6.63841307e-01
-1.34281337e+00 6.71403646e-01 6.96843266e-01 3.35675240e-01
8.23254406e-01 -2.18424737e-01 3.18303943e-01 8.79609108e-01
5.34331143e-01 -1.39882827e+00 7.75034606e-01 9.11764324e-01
8.54318976e-01 -9.70994174e-01 -1.61435261e-01 7.57006884e-01
-7.31217980e-01 7.31557190e-01 8.38010728e-01 -4.53159183e-01
5.48683047e-01 4.66378927e-01 -1.81241393e-01 -2.55446464e-01
-1.24550605e+00 -1.92558587e-01 -9.50651526e-01 7.25286543e-01
1.31317765e-01 -7.42407814e-02 -1.04551174e-01 8.12767744e-02
-2.39824831e-01 -2.03314111e-01 8.43752861e-01 9.87934291e-01
-6.53109193e-01 -6.64556682e-01 -7.72214830e-02 4.36326832e-01
-4.25580442e-01 -1.68152198e-01 -2.80506939e-01 1.01301539e+00
-2.47777373e-01 6.57823324e-01 3.39127600e-01 -4.10929978e-01
2.48110771e-01 -1.68891758e-01 6.46341205e-01 -4.25139636e-01
-7.99472332e-01 -5.64266630e-02 -1.66560769e-01 -6.26238227e-01
-2.46001899e-01 -6.84419751e-01 -1.26082850e+00 -8.09845388e-01
5.09853549e-02 -1.27105117e-01 3.57388556e-01 6.20594025e-01
3.53628397e-01 7.61457562e-01 6.11744702e-01 -7.86862910e-01
-6.64712727e-01 -7.51627088e-01 -8.75685334e-01 4.64398623e-01
3.86756569e-01 -7.47256279e-01 -5.81911027e-01 -3.23680788e-01] | [3.994443655014038, 1.9861928224563599] |
1354b5dc-305c-4088-86f8-32ec54893b61 | busybot-learning-to-interact-reason-and-plan | 2207.08192 | null | https://arxiv.org/abs/2207.08192v2 | https://arxiv.org/pdf/2207.08192v2.pdf | BusyBot: Learning to Interact, Reason, and Plan in a BusyBoard Environment | We introduce BusyBoard, a toy-inspired robot learning environment that leverages a diverse set of articulated objects and inter-object functional relations to provide rich visual feedback for robot interactions. Based on this environment, we introduce a learning framework, BusyBot, which allows an agent to jointly acquire three fundamental capabilities (interaction, reasoning, and planning) in an integrated and self-supervised manner. With the rich sensory feedback provided by BusyBoard, BusyBot first learns a policy to efficiently interact with the environment; then with data collected using the policy, BusyBot reasons the inter-object functional relations through a causal discovery network; and finally by combining the learned interaction policy and relation reasoning skill, the agent is able to perform goal-conditioned manipulation tasks. We evaluate BusyBot in both simulated and real-world environments, and validate its generalizability to unseen objects and relations. Video is available at https://youtu.be/EJ98xBJZ9ek. | ['Shuran Song', 'Zhenjia Xu', 'Zeyi Liu'] | 2022-07-17 | null | null | null | null | ['scene-graph-generation', 'robot-manipulation', 'robot-task-planning'] | ['computer-vision', 'robots', 'robots'] | [-1.84533611e-01 4.82503086e-01 -2.56202459e-01 -1.19415864e-01
-6.50191903e-02 -5.50038636e-01 5.67765415e-01 2.39506718e-02
-9.33115631e-02 7.95403957e-01 1.34733692e-01 6.00519814e-02
-4.24258024e-01 -7.35484838e-01 -8.94685447e-01 -5.73683321e-01
-4.61021185e-01 7.72869945e-01 3.76779974e-01 -2.28431582e-01
-5.74853756e-02 5.94583690e-01 -1.59517872e+00 1.24608562e-03
4.58837628e-01 6.37231112e-01 8.14346910e-01 8.49744201e-01
3.61334503e-01 1.41958809e+00 -2.19082430e-01 2.83559203e-01
4.16216552e-01 -2.14870960e-01 -1.05392015e+00 6.66518733e-02
-3.17177594e-01 -6.90742970e-01 -6.86159849e-01 6.91415071e-01
1.93313211e-01 5.65769136e-01 6.41655266e-01 -1.66113830e+00
-5.27236521e-01 6.51728511e-01 -7.49239028e-02 7.59235620e-02
6.67220294e-01 8.60455036e-01 1.03025162e+00 -4.85880882e-01
9.02035713e-01 1.61296177e+00 -7.30060935e-02 6.26697123e-01
-1.22712302e+00 -6.03865266e-01 3.87652427e-01 2.83983439e-01
-6.78037047e-01 -4.35305417e-01 6.74667776e-01 -4.62398231e-01
9.54461575e-01 -2.15249330e-01 9.21189427e-01 1.39412808e+00
2.54095554e-01 1.02871990e+00 9.53719616e-01 -1.87236518e-01
3.95860404e-01 -3.47245812e-01 -6.71468377e-02 1.05996144e+00
1.28012806e-01 6.23638749e-01 -7.84540236e-01 1.18538290e-01
1.12985516e+00 1.21774986e-01 -1.17863201e-01 -9.31146681e-01
-1.45090246e+00 3.41172189e-01 8.44832599e-01 -1.98368534e-01
-6.57134056e-01 6.99667931e-01 3.18355411e-02 3.87616992e-01
-4.12980676e-01 8.31129432e-01 -4.97491479e-01 -1.68641090e-01
4.40833271e-01 3.19366336e-01 8.89877617e-01 1.22086155e+00
9.49413836e-01 -2.38982663e-02 -3.38408165e-02 5.40983677e-01
5.72614968e-01 8.18123400e-01 -9.28708091e-02 -1.66108489e+00
2.35404074e-01 6.60359621e-01 4.28704858e-01 -7.46468663e-01
-4.87329364e-01 2.77778834e-01 -2.20207796e-01 8.03369164e-01
2.60552824e-01 -2.76668578e-01 -8.64628017e-01 1.75022888e+00
6.10598326e-01 -5.23181930e-02 3.50834519e-01 1.06043911e+00
7.64496148e-01 4.09454584e-01 2.43261740e-01 1.72494456e-01
1.13292074e+00 -1.18645775e+00 -5.53335488e-01 -3.67713064e-01
3.83718342e-01 -6.68481290e-02 1.16156948e+00 3.89088392e-01
-9.24652696e-01 -4.56560880e-01 -8.89362276e-01 8.65356252e-03
-1.94049194e-01 -1.48296162e-01 8.99253368e-01 -5.31424344e-01
-9.10852849e-01 5.94187856e-01 -1.24004316e+00 -6.35572553e-01
5.74301243e-01 3.75649512e-01 -6.95060194e-01 -2.23002583e-02
-6.90830529e-01 1.04240942e+00 5.78889430e-01 -5.99418655e-02
-1.94269371e+00 -1.10171042e-01 -1.00022304e+00 -1.12134993e-01
1.10624623e+00 -9.21182334e-01 1.71603608e+00 -3.71027112e-01
-1.92681420e+00 4.47289765e-01 1.76352456e-01 -2.73939788e-01
2.92069465e-01 -6.58099592e-01 2.42435396e-01 3.25986981e-01
1.63015440e-01 8.20153058e-01 5.73409438e-01 -1.72889495e+00
-5.99728823e-01 -2.41130695e-01 4.86488491e-01 5.78625679e-01
2.27063894e-01 -2.61862844e-01 -4.56281751e-01 -1.68631032e-01
-4.50287014e-02 -9.84788060e-01 -2.60772943e-01 2.40624428e-01
-5.25225282e-01 -4.81916785e-01 8.28987718e-01 3.60582061e-02
2.45850965e-01 -1.95342493e+00 5.48431039e-01 8.22112635e-02
3.09095740e-01 -1.59927636e-01 -4.14784133e-01 8.72208834e-01
1.75017744e-01 -3.39702070e-01 2.39544213e-01 -2.10007310e-01
9.56158563e-02 6.55048192e-01 -2.18237028e-01 2.12393045e-01
1.23483308e-01 1.27915227e+00 -1.58479452e+00 -3.59023929e-01
4.71580535e-01 4.87037003e-02 -4.11553353e-01 8.95282507e-01
-7.61509001e-01 9.97649550e-01 -9.66134191e-01 7.48192787e-01
-1.89330503e-01 -2.90228873e-01 3.82592738e-01 2.61775404e-01
1.05303042e-01 2.00386927e-01 -9.01396155e-01 1.76726270e+00
-4.26985800e-01 4.67549354e-01 3.15850288e-01 -7.21055090e-01
7.87962139e-01 2.80753314e-01 4.53672618e-01 -5.53941011e-01
2.51018673e-01 -1.22106515e-01 -7.02357218e-02 -1.11735129e+00
5.10726571e-02 2.86907554e-01 -1.23166233e-01 4.88115191e-01
4.18468237e-01 -6.17107987e-01 9.64052081e-02 5.31437755e-01
1.59303522e+00 8.02396953e-01 3.46927375e-01 9.90465879e-02
-7.41283670e-02 2.87102759e-01 3.18132162e-01 9.20738399e-01
-2.93149322e-01 -1.76611662e-01 3.36141944e-01 -4.67259049e-01
-7.16065705e-01 -1.39609468e+00 6.25872374e-01 1.23773491e+00
7.10845709e-01 -2.85803884e-01 -1.98628426e-01 -6.82799518e-01
1.64717555e-01 7.70580530e-01 -6.77559435e-01 -3.01341236e-01
-5.11418462e-01 1.52666807e-01 -2.59002205e-02 6.43124580e-01
3.16899002e-01 -2.03401113e+00 -1.26074278e+00 1.16641931e-01
1.15035348e-01 -1.01696837e+00 1.06770590e-01 6.10340297e-01
-5.63344121e-01 -1.41236830e+00 4.44989465e-02 -7.37520635e-01
7.37383246e-01 3.58509302e-01 1.04565072e+00 1.77546367e-01
-2.90182918e-01 1.08415115e+00 -4.41716522e-01 -3.68241519e-01
-4.75816250e-01 -2.88398832e-01 2.82382756e-01 -5.72375298e-01
-1.39740065e-01 -8.72010350e-01 -4.30696458e-01 1.78712040e-01
-4.03699338e-01 4.34001625e-01 8.16095293e-01 8.38664174e-01
2.70434827e-01 4.43958864e-02 2.55100012e-01 -3.76122385e-01
3.70338142e-01 -4.68448400e-01 -7.41641462e-01 3.78027596e-02
-2.33449310e-01 2.23793238e-01 3.07151377e-01 -5.64721465e-01
-1.22538459e+00 3.36904377e-01 5.31022847e-01 -5.13139307e-01
-5.49435496e-01 2.55376011e-01 -2.68266231e-01 2.73215830e-01
6.72604024e-01 -6.67585582e-02 2.46529793e-03 -3.03083897e-01
7.18186855e-01 1.97035685e-01 8.83392334e-01 -9.59728003e-01
8.47468734e-01 5.84651470e-01 2.78390031e-02 -3.46891195e-01
-6.53881490e-01 -2.42297187e-01 -7.25848079e-01 -5.62125802e-01
9.79462445e-01 -8.53958130e-01 -1.67738795e+00 3.11326712e-01
-1.06354535e+00 -1.27529132e+00 -4.78786498e-01 5.26952922e-01
-1.11403346e+00 -1.88568741e-01 -6.10419750e-01 -9.77411568e-01
1.97898239e-01 -1.13318920e+00 9.98885930e-01 2.43181884e-01
-2.82593936e-01 -6.51653528e-01 1.40774334e-02 1.87219560e-01
-1.03650637e-01 3.60758364e-01 7.16934800e-01 -5.11373341e-01
-1.02747035e+00 1.93617016e-01 -3.45661715e-02 -1.76122203e-01
3.98467302e-01 -1.97753072e-01 -6.39905512e-01 -1.29669517e-01
-4.37957138e-01 -8.45125318e-01 5.09302318e-01 6.94741607e-02
7.19463110e-01 -2.13134080e-01 -7.42884040e-01 1.86222792e-01
9.07099843e-01 4.64800239e-01 1.75506666e-01 3.61995280e-01
6.43707752e-01 8.25966179e-01 9.70610142e-01 4.51756418e-01
5.93053758e-01 5.83417952e-01 1.06522286e+00 2.41691232e-01
-1.57201126e-01 -6.84321642e-01 3.50261599e-01 2.14307740e-01
-1.14710733e-01 -2.51847714e-01 -9.06142116e-01 3.74871939e-01
-2.34574103e+00 -8.48007798e-01 2.30066106e-01 1.63984323e+00
7.35413373e-01 1.38023973e-01 5.19754626e-02 -3.23479980e-01
2.49443874e-01 -1.98920846e-01 -9.48287666e-01 -4.98667732e-02
3.20838362e-01 -3.22019488e-01 2.96695948e-01 5.13240993e-01
-8.23997796e-01 1.23491371e+00 6.31316137e+00 9.26097780e-02
-7.19026625e-01 -1.47941947e-01 3.84766841e-04 -6.96766153e-02
1.21172227e-01 1.03341579e-01 -3.27370793e-01 -3.25328149e-02
3.32458854e-01 1.24249302e-01 8.78377974e-01 1.05083668e+00
2.39073187e-01 -5.52367568e-01 -1.44280696e+00 7.71281362e-01
-4.08468038e-01 -1.00847733e+00 -4.01714027e-01 -1.17891282e-02
3.32980454e-01 1.78369328e-01 -1.44600928e-01 5.03038108e-01
1.69620776e+00 -9.31793690e-01 8.77365291e-01 5.61734080e-01
3.24256092e-01 -2.02837035e-01 2.69223630e-01 6.78253472e-01
-1.09977329e+00 -5.33582270e-01 5.43079674e-02 -4.64570284e-01
2.28859186e-01 -7.88748041e-02 -1.14911747e+00 4.04022843e-01
9.43802297e-01 8.86424303e-01 -7.82785751e-03 7.43341386e-01
-1.10815001e+00 1.09815672e-01 -2.57026672e-01 -3.00125718e-01
-1.15884848e-01 -9.57708880e-02 7.80538917e-01 4.36037332e-01
-1.68763340e-01 4.99327600e-01 6.17011189e-01 9.28749263e-01
1.80771321e-01 -5.44835448e-01 -8.71305645e-01 2.38296445e-02
7.51032948e-01 1.15984404e+00 -6.65079415e-01 -4.26702023e-01
-3.65068838e-02 9.22775805e-01 8.48234355e-01 6.77652478e-01
-6.40537500e-01 1.62861019e-01 8.10962856e-01 -4.20301855e-01
2.57898808e-01 -5.32808244e-01 1.74823970e-01 -8.47629964e-01
-1.72581658e-01 -6.85922503e-01 1.75037324e-01 -1.35951555e+00
-9.78492737e-01 3.71472716e-01 3.39760602e-01 -9.47526872e-01
-1.87737137e-01 -6.50158525e-01 -3.52663249e-01 3.32339108e-01
-1.19479477e+00 -1.16341698e+00 -5.67886949e-01 5.98004043e-01
5.73112786e-01 -1.59826487e-01 8.62728119e-01 -5.49925923e-01
-5.03273427e-01 -2.88056284e-01 -4.96749878e-01 3.15465592e-02
4.86788690e-01 -1.26761377e+00 1.63037777e-01 3.77459764e-01
1.43908309e-02 4.87591594e-01 7.15658486e-01 -9.16879177e-01
-1.78350151e+00 -8.78062427e-01 -1.18844263e-01 -7.42614269e-01
7.15252697e-01 -4.77759212e-01 -4.85361218e-01 9.91287529e-01
9.75764692e-02 2.27218226e-01 1.12509482e-01 3.96334827e-02
-4.18448120e-01 1.31380916e-01 -1.00741065e+00 9.89376009e-01
1.62871039e+00 -2.33010620e-01 -9.93597984e-01 3.45347852e-01
9.46285486e-01 -7.58475006e-01 -5.82662761e-01 3.64701182e-01
5.91366589e-01 -7.91558266e-01 9.61908758e-01 -8.29848289e-01
5.51053107e-01 -5.08146226e-01 -1.07302725e-01 -1.49649608e+00
-4.67500269e-01 -6.41639769e-01 -4.14818913e-01 6.94533527e-01
1.40915737e-01 -6.62159801e-01 4.80479330e-01 4.75181997e-01
-1.74706921e-01 -7.01080024e-01 -4.66015607e-01 -8.76971960e-01
-5.61274171e-01 -4.40812409e-01 5.18650830e-01 4.30524856e-01
3.66084516e-01 4.89639133e-01 -3.28727484e-01 5.30328095e-01
4.67100739e-01 2.57011473e-01 1.32285500e+00 -1.11715865e+00
-5.60695529e-01 -1.45968929e-01 -2.29092732e-01 -1.30408216e+00
3.48906457e-01 -5.77644885e-01 7.26027966e-01 -1.83777738e+00
1.69440642e-01 -6.29511833e-01 -1.07726917e-01 1.11797333e+00
6.24579377e-02 -2.96556115e-01 4.82016504e-01 4.56978619e-01
-1.03599262e+00 6.99949920e-01 1.64876294e+00 1.89552456e-02
-3.95409018e-01 -2.38942519e-01 -3.08613241e-01 9.05473650e-01
7.36498058e-01 -3.02574396e-01 -6.29142284e-01 -5.49615443e-01
1.79265201e-01 3.75907898e-01 6.62488520e-01 -1.00156879e+00
2.56997138e-01 -7.38160729e-01 2.47630835e-01 -2.29889050e-01
7.08857179e-01 -1.02382135e+00 -9.25627071e-03 6.73030853e-01
-5.49079418e-01 -1.60098523e-01 7.65164942e-02 9.28468108e-01
3.64544392e-01 2.10509047e-01 2.69035190e-01 -2.99820244e-01
-1.13851559e+00 2.11518988e-01 -5.62916398e-01 -1.91233769e-01
1.38251507e+00 1.91736370e-02 -4.41148758e-01 -5.43423414e-01
-9.33270991e-01 8.97715032e-01 5.81610441e-01 6.61163151e-01
6.26990259e-01 -1.18625247e+00 -1.67259529e-01 2.92502856e-03
2.93941796e-01 5.08281767e-01 -8.07738453e-02 5.21233022e-01
-1.98796064e-01 4.25889045e-02 -4.56062824e-01 -6.54245794e-01
-1.08177662e+00 7.92985439e-01 2.17003390e-01 1.13852285e-01
-8.57427001e-01 9.73058760e-01 5.15802562e-01 -7.17236280e-01
4.00719851e-01 -3.52140695e-01 -2.05368251e-01 -6.31551802e-01
1.59504071e-01 2.83417612e-01 -6.77168608e-01 -2.81143963e-01
-2.72237390e-01 1.81478173e-01 3.81654024e-01 -3.82823467e-01
1.34902859e+00 -2.31219813e-01 5.51835261e-02 6.03992462e-01
4.83387291e-01 -2.07868144e-01 -1.96475410e+00 -2.54438758e-01
-2.19074991e-02 -4.59964037e-01 -4.01919395e-01 -8.93943727e-01
-6.91995680e-01 4.03045565e-01 4.82594930e-02 2.63238493e-02
8.38065147e-01 4.62874740e-01 3.26576948e-01 9.84430611e-01
9.58846867e-01 -8.64178658e-01 9.56780255e-01 5.59223771e-01
1.19636953e+00 -1.34083903e+00 5.15897535e-02 -4.12120312e-01
-7.41834998e-01 8.63828480e-01 1.06633770e+00 -2.84332156e-01
4.59299594e-01 5.09949505e-01 1.33172765e-01 -5.93439698e-01
-1.16794825e+00 -4.51071203e-01 -2.03953296e-01 9.81869161e-01
-4.29292589e-01 1.01631299e-01 5.56890070e-01 1.36112034e-01
-1.38092592e-01 2.58379914e-02 3.05505782e-01 1.28778911e+00
-6.01834118e-01 -8.15708995e-01 -3.05184573e-01 1.89275742e-01
3.82579029e-01 4.49599981e-01 -4.61734384e-01 7.60872126e-01
9.21216235e-02 1.11399436e+00 -4.25378121e-02 -3.49932849e-01
3.73839468e-01 -2.98423409e-01 6.40221715e-01 -9.86608565e-01
-6.72039716e-03 -2.45010495e-01 1.29052982e-01 -1.17935729e+00
-5.04438877e-01 -5.62730610e-01 -1.98547649e+00 1.39604971e-01
-2.21096333e-02 -4.39663231e-02 3.71501535e-01 9.43803906e-01
3.77189845e-01 8.80411446e-01 4.12709683e-01 -1.31110072e+00
-3.32004160e-01 -8.66849184e-01 -2.49401197e-01 3.84147942e-01
6.16525829e-01 -1.26729488e+00 -1.50761843e-01 8.79471824e-02] | [4.542417049407959, 0.8327354788780212] |
95a0134f-0f63-4fa4-903b-cb510e835494 | molecular-property-prediction-by-semantic | 2303.06902 | null | https://arxiv.org/abs/2303.06902v1 | https://arxiv.org/pdf/2303.06902v1.pdf | Molecular Property Prediction by Semantic-invariant Contrastive Learning | Contrastive learning have been widely used as pretext tasks for self-supervised pre-trained molecular representation learning models in AI-aided drug design and discovery. However, exiting methods that generate molecular views by noise-adding operations for contrastive learning may face the semantic inconsistency problem, which leads to false positive pairs and consequently poor prediction performance. To address this problem, in this paper we first propose a semantic-invariant view generation method by properly breaking molecular graphs into fragment pairs. Then, we develop a Fragment-based Semantic-Invariant Contrastive Learning (FraSICL) model based on this view generation method for molecular property prediction. The FraSICL model consists of two branches to generate representations of views for contrastive learning, meanwhile a multi-view fusion and an auxiliary similarity loss are introduced to make better use of the information contained in different fragment-pair views. Extensive experiments on various benchmark datasets show that with the least number of pre-training samples, FraSICL can achieve state-of-the-art performance, compared with major existing counterpart models. | ['Shuigeng Zhou', 'Jihong Guan', 'Ailin Xie', 'Ziqiao Zhang'] | 2023-03-13 | null | null | null | null | ['molecular-property-prediction'] | ['miscellaneous'] | [ 6.20595157e-01 -2.44640023e-01 -5.80723584e-01 -3.61879647e-01
-7.04027712e-01 -4.62414324e-01 6.64476752e-01 3.35140198e-01
2.09877819e-01 1.01229465e+00 8.14199299e-02 -2.39231944e-01
-1.11007884e-01 -7.42153764e-01 -7.90234804e-01 -9.70367849e-01
9.90032181e-02 2.53451705e-01 1.77317962e-01 -1.82296082e-01
2.86176383e-01 3.29133719e-01 -1.26494932e+00 6.43186271e-01
1.19443059e+00 7.24735320e-01 2.37965703e-01 1.99520923e-02
-1.76044703e-01 9.26159501e-01 -2.59595573e-01 -5.11496365e-01
1.45776197e-01 -7.89651752e-01 -6.37391567e-01 3.11532579e-02
3.66546959e-01 8.46356445e-04 -1.94963813e-02 1.20085478e+00
6.23108327e-01 9.53829363e-02 9.20476675e-01 -1.29260230e+00
-5.94721437e-01 3.04244936e-01 -6.18426442e-01 8.97861551e-03
4.92670804e-01 9.02388766e-02 9.14705575e-01 -1.11532211e+00
8.85012031e-01 1.44341159e+00 3.96067679e-01 6.25541031e-01
-1.29654825e+00 -9.34070051e-01 3.75812501e-01 6.09417856e-01
-1.12473679e+00 -3.98272991e-01 1.12923372e+00 -3.44609976e-01
8.10400784e-01 2.28730917e-01 4.95837063e-01 1.20055950e+00
4.19363409e-01 7.54762173e-01 1.38052857e+00 -5.63385077e-02
2.95722246e-01 4.61292046e-04 -4.46981415e-02 1.01479173e+00
3.44549298e-01 1.32427067e-01 -5.98874032e-01 -3.82111877e-01
5.17589211e-01 4.17287618e-01 -4.53734100e-01 -9.43132341e-01
-9.89008129e-01 8.79998624e-01 8.06989849e-01 -8.97400919e-03
-3.35767597e-01 -2.73656189e-01 5.48991501e-01 2.57817924e-01
4.91883665e-01 4.19815421e-01 -4.70242858e-01 4.61574167e-01
-5.35633683e-01 1.90973401e-01 4.24520969e-01 7.94600308e-01
5.78847170e-01 -2.05926280e-02 -2.63391584e-02 7.07321465e-01
3.26840341e-01 1.27595589e-01 4.70306426e-01 -3.20326477e-01
6.32551253e-01 9.05730247e-01 -3.87076348e-01 -9.04000819e-01
-2.43662640e-01 -6.78499758e-01 -1.16555083e+00 1.07926019e-01
-2.05722094e-01 3.09828013e-01 -1.15570819e+00 1.70874989e+00
5.11811614e-01 3.18843722e-01 3.60696644e-01 6.28693819e-01
9.87168014e-01 8.49533498e-01 3.98383111e-01 -7.09815919e-01
1.20037401e+00 -1.07031882e+00 -6.89756751e-01 2.11049452e-01
6.56372190e-01 -7.31668890e-01 6.36258662e-01 6.01154804e-01
-7.71111965e-01 -6.15648031e-01 -1.48261249e+00 1.32562980e-01
-4.33715254e-01 -1.75050810e-01 8.97778094e-01 4.69804376e-01
-3.11369151e-01 7.52601922e-01 -6.00596726e-01 2.18712658e-01
9.76715147e-01 2.37641394e-01 -8.07962298e-01 -4.27692384e-01
-1.06895554e+00 6.39467895e-01 7.25524902e-01 -1.58438280e-01
-9.52855647e-01 -9.98844743e-01 -9.21854317e-01 -1.62906975e-01
5.55270731e-01 -8.71410489e-01 6.13252580e-01 -1.05010164e+00
-1.27921867e+00 5.14667571e-01 -2.47746885e-01 -1.68660909e-01
4.14503485e-01 1.13935016e-01 -4.90510166e-01 1.74726486e-01
2.11364388e-01 5.11630654e-01 6.89227223e-01 -1.56084645e+00
-2.89260834e-01 -7.83503175e-01 -1.11590631e-01 5.36277115e-01
2.43768916e-02 -4.83533829e-01 -1.31358474e-01 -7.18380213e-01
4.12423462e-01 -6.67824745e-01 -4.61182117e-01 4.49050106e-02
-4.02073145e-01 -1.85836256e-01 8.14417481e-01 -4.68399942e-01
1.08073366e+00 -1.63561821e+00 5.91048419e-01 3.43470842e-01
4.95757878e-01 4.33192909e-01 -2.40380153e-01 4.56082791e-01
-6.56453073e-01 2.01341771e-02 -3.20808381e-01 1.32554933e-01
-6.08625054e-01 -2.30700541e-02 -1.18104860e-01 2.69878417e-01
2.06402078e-01 6.88790441e-01 -1.22976053e+00 -5.01150548e-01
1.71545044e-01 4.10260171e-01 -6.10779405e-01 2.03546941e-01
-3.87206525e-01 4.10410553e-01 -6.64188802e-01 8.90701890e-01
9.35223937e-01 -3.99499685e-01 4.27225322e-01 -5.95585823e-01
2.19844058e-01 3.54272909e-02 -9.13876235e-01 1.87943029e+00
-1.84317812e-01 -2.62016743e-01 -5.20170569e-01 -1.13415956e+00
8.84357512e-01 2.14787468e-01 4.63311106e-01 -7.37686574e-01
-1.86340183e-01 3.04160178e-01 2.78434884e-02 -2.99909681e-01
-1.28925934e-01 -5.35921872e-01 4.20141816e-01 -5.85614629e-02
1.99519828e-01 7.56270364e-02 -7.89584965e-02 1.58423945e-01
7.71113098e-01 4.35210288e-01 7.16060936e-01 -4.65470701e-02
9.54586685e-01 -3.97566669e-02 7.46740878e-01 8.35025087e-02
-1.97454207e-02 3.02643239e-01 4.25411165e-01 -7.25756764e-01
-7.65603781e-01 -9.18865323e-01 -7.54736811e-02 6.83500051e-01
3.23570460e-01 -5.84227979e-01 -6.08312309e-01 -1.31521380e+00
-3.14352423e-01 4.72590566e-01 -4.55716580e-01 -5.31389177e-01
-2.46965706e-01 -1.06297278e+00 -8.92670546e-03 3.93924683e-01
6.05885148e-01 -8.03559899e-01 1.51690185e-01 3.50717336e-01
6.54669777e-02 -6.55079067e-01 -3.29483151e-01 3.65722358e-01
-9.92030084e-01 -1.40750718e+00 -7.42151260e-01 -9.13125753e-01
7.77485907e-01 5.93436301e-01 8.47054660e-01 -5.36932759e-02
-1.63076162e-01 -3.00194889e-01 -4.01134819e-01 -3.51043850e-01
-3.86801392e-01 -1.97887570e-01 -1.24252833e-01 -1.70825366e-02
2.57856518e-01 -6.93857610e-01 -8.21499288e-01 2.18050942e-01
-9.76789117e-01 4.04805541e-01 6.67355120e-01 1.20464158e+00
1.01850498e+00 -1.60285786e-01 7.94192433e-01 -1.29094148e+00
2.54653960e-01 -5.52966654e-01 -4.32771802e-01 4.40991014e-01
-7.52664089e-01 3.23657751e-01 1.01170611e+00 -3.60107601e-01
-9.85202253e-01 1.33961439e-01 -1.30760550e-01 -6.02537632e-01
1.50272623e-01 6.65403306e-01 -7.37380385e-01 -2.48872101e-01
6.17545009e-01 5.49501777e-01 6.12257309e-02 -3.75864476e-01
2.64665008e-01 3.74330044e-01 1.06222987e-01 -3.69540513e-01
6.41652465e-01 3.12301695e-01 5.57208538e-01 -4.32098061e-01
-8.42030048e-01 -2.81327754e-01 -4.49492812e-01 1.16570681e-01
6.77640855e-01 -1.10221744e+00 -6.51398361e-01 3.80261183e-01
-9.58562255e-01 3.58290583e-01 1.44934714e-01 4.21219170e-01
-5.71463168e-01 7.02492952e-01 -2.09795356e-01 -4.08335388e-01
-5.25245368e-01 -1.43129909e+00 7.67782152e-01 5.93872108e-02
2.94039905e-01 -7.29788363e-01 1.80090070e-01 6.90234482e-01
-7.67471492e-02 4.13088590e-01 1.39657199e+00 -9.01100576e-01
-5.72899282e-01 1.36746749e-01 -1.34278893e-01 3.26000094e-01
5.44871688e-01 -4.31123227e-01 -9.35257435e-01 -3.41564298e-01
-6.56048805e-02 -4.04707730e-01 1.05831850e+00 8.47418159e-02
1.45160472e+00 -4.77635384e-01 -6.08398557e-01 7.56956160e-01
1.52649891e+00 7.23156750e-01 6.85147345e-01 4.97472174e-02
1.19474423e+00 4.24084663e-01 6.43497407e-01 3.21897566e-01
1.57736912e-01 6.60788000e-01 4.69636053e-01 -1.63294420e-01
-2.68317610e-02 -7.47275352e-01 1.72346488e-01 6.96376681e-01
-2.19492495e-01 -2.57739395e-01 -3.83519471e-01 -2.06022263e-02
-1.83186686e+00 -1.20741570e+00 1.23437062e-01 2.21941042e+00
9.92501497e-01 1.18463319e-02 6.17406750e-03 7.45661482e-02
6.25845909e-01 3.04329783e-01 -9.12345231e-01 -4.53263558e-02
-6.52352199e-02 3.51177871e-01 2.50456452e-01 2.97034800e-01
-1.15352297e+00 9.30674613e-01 4.90620327e+00 1.38116097e+00
-1.15878320e+00 -1.64213166e-01 1.00076485e+00 4.55233127e-01
-6.84989989e-01 1.29258722e-01 -5.96065044e-01 5.42582214e-01
5.59973955e-01 5.07301046e-03 8.67907628e-02 9.01358366e-01
9.13839787e-02 3.92597526e-01 -1.16259503e+00 1.15975428e+00
2.84725010e-01 -1.72520709e+00 9.63625491e-01 8.73300806e-02
8.24046671e-01 -6.26404643e-01 4.92469855e-02 8.27528685e-02
-1.23976268e-01 -1.11672843e+00 1.56860813e-01 4.63046640e-01
7.22993731e-01 -1.13290179e+00 6.34501636e-01 1.89144418e-01
-1.36278033e+00 2.42272809e-01 -3.87548774e-01 3.45205814e-01
-2.67375648e-01 3.14772218e-01 -6.96363926e-01 1.33788466e+00
5.01912385e-02 1.17067838e+00 -5.42635560e-01 9.04057562e-01
1.29639700e-01 2.82465756e-01 2.55113274e-01 6.71956837e-02
2.01756611e-01 -4.30291563e-01 3.62268418e-01 6.89136565e-01
-3.17901559e-02 1.85975760e-01 3.96800518e-01 6.75788760e-01
-2.79146552e-01 3.86523366e-01 -7.42008388e-01 6.75196424e-02
4.42352712e-01 9.86392677e-01 -5.21996856e-01 -4.91775274e-01
-4.22389507e-01 1.07066166e+00 2.89788127e-01 1.53355211e-01
-8.19375992e-01 -3.22780401e-01 3.36268425e-01 1.73090443e-01
1.42216295e-01 2.85236061e-01 2.50120834e-02 -1.26933944e+00
-6.64760619e-02 -1.34193158e+00 4.64586347e-01 -6.17102146e-01
-1.46028519e+00 4.91831332e-01 -3.39237191e-02 -1.56977344e+00
1.43624738e-01 -7.70870149e-01 -4.96981889e-01 6.43739343e-01
-1.72542596e+00 -1.43637383e+00 -1.32760003e-01 4.57920194e-01
8.36175561e-01 -4.48733240e-01 8.80691826e-01 1.96779579e-01
-6.27986848e-01 6.34833515e-01 1.50048852e-01 -4.03950840e-01
6.82494402e-01 -1.11676848e+00 -5.15493006e-02 3.38032991e-01
-1.80890039e-02 6.71550453e-01 4.04875219e-01 -8.05299282e-01
-1.45260131e+00 -1.29398179e+00 3.45848858e-01 -5.27907722e-02
1.64435416e-01 -2.74094474e-03 -1.00632286e+00 2.99655139e-01
-8.09614658e-02 1.84582293e-01 1.06748688e+00 -1.80683389e-01
-6.91399693e-01 -1.57693103e-01 -1.22464550e+00 5.83302557e-01
1.33005953e+00 -3.42282265e-01 -5.08006334e-01 5.31760693e-01
7.19796240e-01 -2.39156902e-01 -8.33402097e-01 8.93526137e-01
5.14167309e-01 -9.48279202e-01 1.32246518e+00 -9.21786606e-01
5.39140582e-01 -5.47263622e-01 -9.82784405e-02 -1.24611259e+00
-3.96650583e-01 -3.42746198e-01 1.87353194e-02 9.16110158e-01
4.47332442e-01 -5.80570281e-01 7.44885743e-01 -5.24388254e-02
-2.41969824e-01 -1.18821692e+00 -8.24254811e-01 -7.81353831e-01
1.43122543e-02 2.82045037e-01 5.22381663e-01 1.18113482e+00
-5.87912910e-02 1.02101159e+00 -3.61458600e-01 -6.76306188e-02
7.17839062e-01 3.88858199e-01 5.73870659e-01 -1.25170362e+00
-3.46785218e-01 -2.09781080e-01 -6.52907610e-01 -9.09920692e-01
5.12263216e-02 -1.14925814e+00 -3.91550004e-01 -1.29498351e+00
8.67856443e-01 -2.67396092e-01 -4.40102279e-01 4.19895172e-01
-4.93209630e-01 -5.98858185e-02 -2.75874678e-02 2.61620253e-01
-6.51302397e-01 9.81663227e-01 1.63259208e+00 -3.61646771e-01
-2.26153776e-01 -1.18611366e-01 -7.80304611e-01 6.60889506e-01
5.71021616e-01 -2.58855432e-01 -9.03900206e-01 3.36543888e-01
2.80927494e-02 2.23400235e-01 2.59568006e-01 -8.58049214e-01
-1.36006579e-01 -4.84802037e-01 7.82319844e-01 -7.59472728e-01
3.32073748e-01 -7.18024731e-01 3.85439336e-01 8.05159926e-01
-2.44694084e-01 -8.51178765e-02 -4.57354225e-02 1.00640261e+00
-3.10143799e-01 -9.60275531e-02 8.82760465e-01 -4.68114763e-01
-5.03027916e-01 6.48518801e-01 1.37914896e-01 -1.85892165e-01
1.02717602e+00 -2.39001513e-01 -4.17006850e-01 3.34912874e-02
-4.70165581e-01 4.71501835e-02 4.00656015e-01 2.50398308e-01
9.44610476e-01 -1.40018010e+00 -4.63878602e-01 3.18107247e-01
4.93535757e-01 2.91313184e-03 4.19688612e-01 4.15763438e-01
-4.49504822e-01 3.10436577e-01 -2.13687181e-01 -4.31136161e-01
-1.55373371e+00 9.64934647e-01 2.56815434e-01 -5.07233143e-01
-2.83015907e-01 6.66835725e-01 6.10735238e-01 -2.85141677e-01
-8.97827893e-02 3.97731103e-02 -3.62382621e-01 1.09260574e-01
4.46686953e-01 1.62778795e-01 1.42149717e-01 -5.26438653e-01
-4.29681361e-01 6.09208226e-01 -6.86494410e-01 4.42720383e-01
1.37230217e+00 4.31402236e-01 1.65776834e-02 1.07944541e-01
1.35691309e+00 -2.08554551e-01 -1.01342368e+00 -1.64075792e-01
-3.14846724e-01 -3.22883040e-01 -2.62295175e-02 -9.16117966e-01
-8.29613209e-01 9.11699533e-01 8.78403127e-01 -4.46753681e-01
1.02887070e+00 -3.15358281e-01 6.60959780e-01 4.62938547e-01
3.90093476e-01 -6.12279594e-01 3.58980626e-01 9.62041393e-02
8.95345509e-01 -1.39677930e+00 4.52040702e-01 -8.35679650e-01
-6.48527205e-01 1.12077546e+00 7.29900062e-01 1.54684559e-01
4.51848954e-01 -4.12066847e-01 -3.11115861e-01 -4.54596400e-01
-7.47573733e-01 -2.58727036e-02 5.05354583e-01 6.45440519e-01
5.66274941e-01 6.13901354e-02 -4.90650147e-01 4.91341144e-01
3.34457755e-01 -1.17816053e-01 3.61905061e-02 1.08180094e+00
-5.36667824e-01 -1.48416960e+00 -1.23499194e-02 4.86344457e-01
-3.40170026e-01 -3.52296352e-01 -6.01357460e-01 5.27432799e-01
2.22611398e-01 6.37194932e-01 -4.66600627e-01 -5.05188704e-01
1.73387736e-01 6.20773248e-02 7.95086563e-01 -6.78543210e-01
-3.01281124e-01 1.89603001e-01 -1.34675309e-01 -3.95567536e-01
-5.84607482e-01 -2.01620564e-01 -1.25670421e+00 -5.98883033e-02
-5.35726726e-01 2.72810489e-01 1.51516706e-01 8.43287230e-01
6.00817323e-01 5.18347323e-01 1.06810880e+00 -5.12568116e-01
-6.03976667e-01 -7.26161838e-01 -2.63605237e-01 4.91368830e-01
2.26504207e-01 -8.14845443e-01 -1.51714116e-01 8.23008642e-02] | [5.144710063934326, 5.894207000732422] |
4b94d7f2-e02f-426c-b990-1ea8e57326b9 | impact-of-ecg-dataset-diversity-on | null | null | https://doi.org/10.1109/ACCESS.2019.2927726 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8758818 | Impact of ECG Dataset Diversity on Generalization of CNN Model for Detecting QRS Complex | Detection of QRS complexes in electrocardiogram (ECG) signal is crucial for automated cardiac diagnosis. Automated QRS detection has been a research topic for over three decades and several of the traditional QRS detection methods show acceptable detection accuracy, however, the applicability of these methods beyond their study-specific databases was not explored. The non-stationary nature of ECG and signal variance of intra and inter-patient recordings impose significant challenges on single QRS detectors to achieve reasonable performance. In real life, a promising QRS detector may be expected to achieve acceptable accuracy over diverse ECG recordings and, thus, investigation of the model's generalization capability is crucial. This paper investigates the generalization capability of convolutional neural network (CNN) based-models from intra (subject wise leave-one-out and five-fold cross validation) and inter-database (training with single and multiple databases) points-of-view over three publicly available ECG databases, namely MIT-BIH Arrhythmia, INCART, and QT. Leave-one-out test accuracy reports 99.22%, 97.13%, and 96.25% for these databases accordingly and inter-database tests report more than 90% accuracy with the single exception of INCART. The performance variation reveals the fact that a CNN model's generalization capability does not increase simply by adding more training samples, rather the inclusion of samples from a diverse range of subjects is necessary for reasonable QRS detection accuracy. | ['Chandan Karmakar', 'John Yearwood', 'Ahsan Habib'] | 2019-07-10 | null | null | null | ieee-access-2019-7 | ['qrs-complex-detection', 'electrocardiography-ecg'] | ['medical', 'methodology'] | [ 1.64730474e-01 -2.98392028e-01 2.07425326e-01 -4.31939602e-01
-8.90094280e-01 -5.67373574e-01 -1.00410208e-01 4.27455693e-01
-5.28808892e-01 8.78866613e-01 -4.80865419e-01 -3.25667262e-01
-4.28274781e-01 -3.42053682e-01 -1.49621665e-01 -5.86892068e-01
-4.53359365e-01 7.63296932e-02 6.80492222e-02 -5.60030676e-02
-1.66933104e-01 7.11023927e-01 -1.05916715e+00 1.53654829e-01
7.67092764e-01 1.06718493e+00 -3.31567407e-01 9.79476392e-01
6.14269733e-01 1.33358598e-01 -1.07151830e+00 -1.01921000e-01
9.69243348e-02 -8.69263470e-01 -2.67360210e-01 -1.96591616e-01
7.15033859e-02 6.80339187e-02 -2.31802370e-02 5.40030479e-01
1.34310400e+00 -4.58463222e-01 4.66319889e-01 -9.68650043e-01
-6.81988522e-02 4.96600717e-01 -4.11585659e-01 8.57377887e-01
2.94061610e-03 3.33627015e-01 6.45325243e-01 -6.71388745e-01
4.42075878e-01 3.38924080e-01 1.26604557e+00 6.39929697e-02
-1.33518720e+00 -5.26829660e-01 -6.37264311e-01 -9.31529328e-02
-1.77717316e+00 2.06291340e-02 9.07821536e-01 -4.49902505e-01
9.56033111e-01 4.40337002e-01 8.83408666e-01 8.17979217e-01
6.31192863e-01 2.35155836e-01 1.06202316e+00 -3.22341919e-01
1.66936532e-01 9.82977748e-02 2.01724604e-01 9.71845984e-02
6.18426085e-01 1.31900355e-01 -3.68196934e-01 -3.88649076e-01
8.41267526e-01 -2.22157165e-01 -3.62684101e-01 -2.75843084e-01
-1.33391905e+00 7.03027129e-01 2.83362985e-01 6.72742367e-01
-4.34230477e-01 -1.76657304e-01 9.89185154e-01 5.06783366e-01
2.45661661e-01 6.11306608e-01 -7.25926816e-01 -3.37978333e-01
-1.28765953e+00 3.63435298e-01 5.82315028e-01 4.39751625e-01
2.15799719e-01 4.82100695e-01 -2.14852080e-01 6.92703962e-01
-1.57423288e-01 4.59831983e-01 6.54421806e-01 -4.44420516e-01
3.38840336e-01 4.37707067e-01 -8.74840468e-02 -1.06731474e+00
-9.23629761e-01 -1.34002566e+00 -1.25928056e+00 4.13785176e-03
3.85876030e-01 -4.21293378e-01 -7.26847351e-01 1.46860862e+00
-2.38622762e-02 -1.75928950e-01 1.89512730e-01 9.52207327e-01
9.45104122e-01 1.04842901e-01 1.65097281e-01 -5.28768539e-01
1.32845163e+00 8.16395953e-02 -6.83046162e-01 -3.64367999e-02
7.84543872e-01 -4.98124450e-01 6.50899410e-01 5.94923735e-01
-1.09000134e+00 -8.77267301e-01 -1.31093156e+00 5.78892231e-01
4.25170660e-02 5.42193294e-01 2.23352775e-01 8.73673856e-01
-7.51124740e-01 7.93378592e-01 -7.38160551e-01 -4.99842405e-01
6.73708618e-01 5.65987468e-01 -3.47532660e-01 2.18874559e-01
-1.37545300e+00 7.57579327e-01 2.96408832e-01 3.13137382e-01
-5.42472064e-01 -7.25121379e-01 -6.21627152e-01 -8.86498839e-02
-1.31080687e-01 -7.14938462e-01 8.73264253e-01 -7.44016588e-01
-7.75743127e-01 8.61329079e-01 1.14125103e-01 -6.95631921e-01
7.14079559e-01 -1.37790650e-01 -7.25098133e-01 1.24735683e-01
-3.11416909e-02 6.83152154e-02 6.83060765e-01 -7.14056432e-01
-2.83666402e-01 -5.72625399e-01 -4.44608152e-01 -6.87761307e-02
2.28626892e-01 1.89205315e-02 8.27693418e-02 -7.55510807e-01
4.13458437e-01 -8.34835112e-01 -1.62910208e-01 -3.70658129e-01
-3.11708152e-01 2.20142096e-01 4.30403858e-01 -5.43223202e-01
1.46018362e+00 -2.33687377e+00 -2.28062913e-01 3.80750597e-01
4.23310369e-01 4.82079566e-01 2.00318903e-01 4.13524956e-01
-3.51459712e-01 1.48275271e-01 -4.24089909e-01 2.99786717e-01
-4.77904737e-01 4.56511267e-02 2.60423899e-01 7.06507325e-01
3.69086385e-01 9.45407033e-01 -2.97881484e-01 -3.01979840e-01
2.15227127e-01 5.64980090e-01 -1.28175706e-01 -1.55882716e-01
5.62637329e-01 5.67849100e-01 -2.43467599e-01 4.25397426e-01
4.92648274e-01 -1.24459095e-01 1.78733990e-01 -4.75572079e-01
1.47690982e-01 -9.79122370e-02 -1.32773721e+00 1.43641198e+00
5.91986738e-02 6.71640396e-01 -3.36617380e-01 -1.16190577e+00
1.14610744e+00 7.01437950e-01 5.00072241e-01 -6.27197027e-01
3.05781990e-01 4.14440840e-01 9.22771811e-01 -4.64486659e-01
-1.84666947e-01 -2.49220505e-01 5.86773977e-02 1.33265942e-01
-1.24906167e-01 2.10199684e-01 1.10770464e-01 -1.59723982e-01
1.01896596e+00 -2.96840161e-01 5.99681377e-01 -4.98616606e-01
3.36188972e-01 -1.57756791e-01 7.71379948e-01 9.75763440e-01
-4.82304007e-01 1.01323605e+00 4.68604833e-01 -8.31190467e-01
-7.53075600e-01 -1.07647610e+00 -6.46458030e-01 2.38341317e-01
-3.73792350e-01 -1.99615180e-01 -5.64427018e-01 -3.75566661e-01
7.24181458e-02 1.69595554e-01 -6.45086467e-01 -2.93898761e-01
-6.57911539e-01 -1.11872482e+00 1.24996495e+00 8.49704206e-01
4.24624205e-01 -1.25011885e+00 -1.44172263e+00 6.09269619e-01
4.95605618e-02 -9.72007871e-01 1.40188277e-01 4.53320682e-01
-1.19070351e+00 -1.21116400e+00 -8.43838871e-01 -4.22775537e-01
6.17983863e-02 -5.00555038e-01 1.23544121e+00 8.80119950e-02
-9.09564793e-01 9.30308476e-02 -1.46879360e-01 -7.59073913e-01
-2.60060281e-01 1.33078530e-01 8.26777592e-02 -6.98129088e-02
3.18680078e-01 -5.75917542e-01 -8.91106963e-01 1.20400444e-01
-4.58793104e-01 -5.52826822e-01 7.04723239e-01 8.24586213e-01
4.98040438e-01 -1.94460958e-01 1.27114248e+00 -7.11818993e-01
9.14062917e-01 -1.19413346e-01 -1.43767387e-01 -8.31002146e-02
-7.58116961e-01 -7.05862403e-01 4.32672620e-01 -2.69428521e-01
-2.66088963e-01 -2.30981167e-02 -1.63867265e-01 -3.28896433e-01
-3.79742354e-01 6.40875757e-01 8.03610682e-02 1.76828206e-01
1.04071176e+00 2.11756691e-01 8.62411503e-03 2.39871256e-02
-2.97733098e-01 5.32760322e-01 5.39338529e-01 -1.62196159e-01
4.77037191e-01 1.61790907e-01 1.58634916e-01 -8.37044597e-01
-3.37422967e-01 -5.41714072e-01 -8.41915071e-01 -6.57246038e-02
9.25346673e-01 -9.23293531e-01 -5.70992470e-01 3.99282426e-01
-9.48205113e-01 6.26319721e-02 -4.05359745e-01 7.31000006e-01
-2.68323451e-01 2.15315685e-01 -4.90876973e-01 -9.70937073e-01
-1.02406585e+00 -1.12145829e+00 7.50520945e-01 -3.41538377e-02
-9.57246602e-01 -7.59017289e-01 -7.52609372e-02 -8.55711401e-02
5.34199119e-01 8.40933144e-01 7.41324782e-01 -1.05166280e+00
1.61748216e-01 -6.53930664e-01 -2.05515884e-02 4.82932180e-01
3.40374738e-01 -1.81101300e-02 -1.12663543e+00 -4.60221261e-01
5.07611129e-03 5.80217130e-02 5.32393336e-01 7.75589883e-01
7.05538690e-01 3.76389951e-01 -1.59970358e-01 3.20196986e-01
1.41668105e+00 5.43545783e-01 8.31470191e-01 1.46173701e-01
4.00741309e-01 1.91158459e-01 4.10428286e-01 5.14115453e-01
-7.38517568e-02 3.00237238e-01 2.66562961e-02 -5.41632771e-01
2.05035836e-01 5.37124395e-01 -2.16387480e-01 6.14107311e-01
-2.06005707e-01 5.42921834e-02 -1.15938199e+00 5.86050987e-01
-1.45023465e+00 -8.68346810e-01 -5.88506103e-01 2.40222239e+00
5.07836938e-01 5.11323214e-01 5.00525892e-01 9.42537665e-01
4.84947890e-01 -3.08151364e-01 -7.64925122e-01 -5.55041611e-01
-5.04992545e-01 4.63718683e-01 1.51512817e-01 -1.89217091e-01
-1.03437805e+00 -4.54065055e-02 6.16156435e+00 2.68578082e-01
-1.52542377e+00 -3.10436171e-02 9.05030370e-01 8.21327791e-02
5.67634940e-01 -3.23091477e-01 -5.09223104e-01 2.27230310e-01
1.22084987e+00 -5.70701286e-02 -3.54587615e-01 6.08882606e-01
4.18720424e-01 -9.18226875e-03 -1.03866017e+00 1.36661768e+00
2.86906306e-02 -1.24185574e+00 -4.78124589e-01 -1.64557368e-01
3.51041019e-01 1.59955040e-01 -1.46697789e-01 2.26929218e-01
-9.95274425e-01 -1.09372282e+00 4.82386619e-01 5.44994473e-01
1.07336926e+00 -8.32809567e-01 1.34140825e+00 3.34109724e-01
-1.12953305e+00 -1.34005949e-01 -9.12201703e-02 -6.53709751e-03
-1.05951563e-01 8.65947664e-01 -7.13383853e-01 6.38238490e-01
9.77567315e-01 6.00534916e-01 -8.04605424e-01 1.19818723e+00
2.73278147e-01 9.86022294e-01 -1.94884911e-01 1.57288998e-01
-2.20332220e-01 1.82304859e-01 4.77114499e-01 1.28487134e+00
2.22502455e-01 1.64885148e-01 1.87229544e-01 7.31694400e-01
2.58344769e-01 2.41367146e-01 -5.04064143e-01 1.09155580e-01
2.60157734e-01 1.10680830e+00 -8.06964934e-01 -3.44044924e-01
-2.99430698e-01 4.87704575e-01 -1.56455353e-01 3.37083489e-01
-8.62824440e-01 -7.59904265e-01 2.03282699e-01 3.58787119e-01
4.30689789e-02 1.06369622e-01 -9.24143910e-01 -5.49126089e-01
2.33100682e-01 -1.15050972e+00 5.11994958e-01 -5.21002889e-01
-1.06207573e+00 8.21300685e-01 -1.45419598e-01 -1.40291679e+00
-2.88881958e-01 -2.79319614e-01 -7.35003114e-01 1.25478530e+00
-1.00228965e+00 -7.23479450e-01 -2.97397673e-01 3.50040823e-01
2.60587871e-01 -1.33679315e-01 1.02117813e+00 4.75965083e-01
-4.01067048e-01 8.49348724e-01 -2.82268196e-01 3.79526228e-01
7.11754680e-01 -1.15932357e+00 2.03780994e-01 7.31475711e-01
-3.66393104e-03 7.48104453e-01 7.17472434e-01 -4.65197057e-01
-9.67780650e-01 -1.06343722e+00 8.92427266e-01 -3.38040382e-01
1.00256025e-03 7.84729607e-03 -1.08220184e+00 1.70819879e-01
5.58329821e-02 2.95129865e-01 8.29650104e-01 1.15495786e-01
5.06199300e-02 -4.27657962e-01 -1.09872782e+00 2.60671377e-01
5.15628636e-01 -3.61513048e-01 -4.21914279e-01 -1.74136117e-01
-2.77018249e-02 -5.07990003e-01 -1.26951122e+00 9.81619418e-01
9.02614772e-01 -1.15780044e+00 8.57553840e-01 -4.75006163e-01
9.18791294e-02 -2.48445913e-01 1.78230956e-01 -1.05686378e+00
-2.12886810e-01 -5.20522714e-01 1.74168870e-01 9.71878886e-01
5.56269169e-01 -8.17694426e-01 6.31666183e-01 2.79778391e-01
-2.17465892e-01 -1.23886931e+00 -8.72264385e-01 -7.67805696e-01
1.11466505e-01 -6.58119082e-01 2.03236401e-01 7.90054619e-01
-1.46969244e-01 3.37237984e-01 -1.14032574e-01 1.69488713e-01
4.53546971e-01 5.27052134e-02 5.56126475e-01 -1.52399540e+00
-2.14660406e-01 -2.29201227e-01 -7.96825171e-01 1.42386407e-01
-6.19244754e-01 -7.80304492e-01 -2.70588666e-01 -1.39521420e+00
6.18428178e-02 -3.39814514e-01 -7.00433731e-01 2.09589601e-01
-2.27208525e-01 6.14662290e-01 2.38568131e-02 5.39270043e-02
-1.67173877e-01 2.04333710e-03 9.79510427e-01 8.09499025e-02
-5.41588247e-01 2.34250158e-01 -3.98527175e-01 6.05233192e-01
1.03756881e+00 -6.38003588e-01 -5.44196963e-01 2.01671198e-01
5.97848967e-02 4.30664927e-01 5.56204677e-01 -1.48161554e+00
-5.52893728e-02 5.70280075e-01 9.32197869e-01 -7.38614559e-01
9.08646658e-02 -3.88528049e-01 4.71027523e-01 8.79070342e-01
-2.90912151e-01 5.07742584e-01 4.85581309e-01 2.82732993e-01
-3.18403810e-01 1.76292598e-01 8.82973075e-01 -5.81529364e-02
-1.59239858e-01 2.24306643e-01 -4.15856361e-01 4.22379732e-01
8.49926353e-01 -7.98310161e-01 3.11343640e-01 -9.62901935e-02
-1.17805505e+00 -2.40829408e-01 -2.59789169e-01 2.33203083e-01
6.81317329e-01 -1.03366351e+00 -1.02398980e+00 3.32838416e-01
4.65733498e-01 -2.27407321e-01 4.35790896e-01 1.28856361e+00
-6.29396856e-01 3.82094592e-01 -2.38795266e-01 -1.24905884e+00
-1.51376820e+00 1.13825768e-01 7.00203478e-01 -5.65636195e-02
-7.74080455e-01 6.26050115e-01 -1.00023039e-01 1.54128850e-01
1.05931588e-01 -6.00765824e-01 -2.12700024e-01 1.03950955e-01
3.71613950e-01 3.36522013e-01 6.09104097e-01 -4.74192351e-01
-6.08574510e-01 6.15254700e-01 8.27013776e-02 3.10078353e-01
1.01292324e+00 1.14386357e-01 2.92231321e-01 7.72142231e-01
9.67419267e-01 -3.13777477e-01 -4.73000586e-01 1.93726495e-01
-2.47255247e-02 -3.13009252e-03 -2.08255872e-01 -9.91832912e-01
-9.92699862e-01 1.22849679e+00 1.26679718e+00 1.99877128e-01
1.19074476e+00 -4.00254399e-01 4.99697506e-01 1.10578418e-01
2.58853197e-01 -7.79885232e-01 -2.59263247e-01 -8.32112879e-03
8.58900249e-01 -1.05866039e+00 -1.30713284e-02 -1.39627486e-01
-6.94593489e-01 1.20875502e+00 2.26146400e-01 -1.61703974e-01
7.43151307e-01 1.43424615e-01 4.60684001e-01 -4.10284281e-01
-3.89587909e-01 4.53186892e-02 3.44673544e-01 6.87645793e-01
9.06576037e-01 8.30241889e-02 -6.44595742e-01 9.03734207e-01
-1.47402165e-02 3.22097093e-01 4.63145465e-01 1.00143945e+00
-8.16029683e-02 -7.18529880e-01 -1.36354700e-01 7.96531916e-01
-9.13882315e-01 -5.03489450e-02 -2.00568751e-01 1.09637260e+00
2.40635082e-01 9.05374467e-01 -2.66603768e-01 -1.45645738e-01
5.95500886e-01 3.30074370e-01 3.41682047e-01 -5.29066920e-01
-1.18937016e+00 2.39094555e-01 5.63407466e-02 -1.65485457e-01
-2.95199931e-01 -6.82281137e-01 -1.38328922e+00 2.42298201e-01
-3.71614069e-01 1.27935648e-01 4.72577423e-01 9.17264700e-01
5.16001940e-01 9.59448397e-01 2.27787018e-01 -3.05883348e-01
-7.16445267e-01 -1.14506745e+00 -9.45206225e-01 3.81756514e-01
3.79099995e-01 -2.98266351e-01 -2.37709507e-01 1.57235339e-01] | [14.324152946472168, 3.298428773880005] |
92bf76ca-d9af-4a2e-adc6-0dab1fb7873f | multi-normal-estimation-via-pair-consistency | null | null | https://ieeexplore.ieee.org/document/8340177 | https://ieeexplore.ieee.org/document/8340177 | Multi-Normal Estimation via Pair Consistency Voting | The normals of feature points, i.e., the intersection points of multiple smooth surfaces, are ambiguous and undefined. This paper presents a unified definition for point cloud normals of feature and non-feature points, which allows feature points to possess multiple normals. This definition facilitates several succeeding operations, such as feature points extraction and point cloud filtering. We also develop a feature preserving normal estimation method which outputs multiple normals per feature point. The core of the method is a pair consistency voting scheme. All neighbor point pairs vote for the local tangent plane. Each vote takes the fitting residuals of the pair of points and their preliminary normal consistency into consideration. Thus the pairs from the same subspace and relatively far off features dominate the voting. An adaptive strategy is designed to overcome sampling anisotropy. In addition, we introduce an error measure compatible with traditional normal estimators, and present the first benchmark for normal estimation, composed of 152 synthesized data with various features and sampling densities, and 288 real scans with different noise levels. Comprehensive and quantitative experiments show that our method generates faithful feature preserving normals and outperforms previous cutting edge normal estimation methods, including the latest deep learning based method. | ['Ligang Liu', 'Bo Li', 'He Chen', 'Xiuping Liu', 'Junjie Cao', 'Jie Zhang'] | 2019-04-01 | null | null | null | null | ['surface-normals-estimation-from-point-clouds'] | ['computer-vision'] | [-5.52339070e-02 -1.77863911e-01 -3.44608761e-02 -3.42199087e-01
-7.92427838e-01 -8.12902749e-02 4.95400012e-01 -6.58361092e-02
-3.30638111e-01 2.50837028e-01 -2.46548817e-01 2.26150542e-01
-2.59882808e-01 -9.51325893e-01 -6.73924327e-01 -7.78102100e-01
1.20472714e-01 7.40693748e-01 5.02509773e-01 -2.84463257e-01
6.58274174e-01 1.00426364e+00 -1.54985869e+00 -3.81765235e-03
7.04338729e-01 1.20016527e+00 -1.73850790e-01 2.50102609e-01
-2.06254140e-01 -3.91582161e-01 -3.48491013e-01 -2.29425684e-01
6.26252532e-01 1.45773426e-01 -4.49091345e-01 1.09012328e-01
9.80532944e-01 -3.90630573e-01 -1.38326123e-01 1.24105680e+00
2.44175449e-01 1.12471759e-01 1.00551021e+00 -1.43996418e+00
-5.82565725e-01 -2.76385158e-01 -9.51198757e-01 -4.26318139e-01
2.50970006e-01 -1.32843569e-01 9.84926522e-01 -1.45885658e+00
6.76111042e-01 9.90514874e-01 1.06560934e+00 3.55325907e-01
-9.76366580e-01 -5.18351316e-01 -1.48275897e-01 2.83911288e-01
-1.59472084e+00 -2.69480139e-01 1.12458634e+00 -3.61103535e-01
5.18683493e-01 6.14027500e-01 7.32962370e-01 5.62277913e-01
4.64182645e-01 3.62357736e-01 5.39224505e-01 -5.31345129e-01
1.81872800e-01 -1.21762127e-01 3.29127133e-01 7.32341230e-01
4.99771923e-01 -9.40343514e-02 -2.79912412e-01 -5.81413150e-01
1.16942751e+00 4.14765954e-01 -4.63074356e-01 -9.37263489e-01
-1.27628815e+00 9.21790898e-01 3.55833858e-01 1.24720134e-01
-4.45921808e-01 -1.05061084e-01 4.16501015e-02 2.30431855e-01
2.66897857e-01 -2.24450184e-03 -4.63458508e-01 1.99400827e-01
-7.07589030e-01 2.61071146e-01 8.65584314e-01 1.01699531e+00
1.13903284e+00 -2.10121289e-01 1.21960133e-01 7.76459455e-01
4.76641119e-01 7.10173130e-01 2.01912642e-01 -1.21178377e+00
1.51870161e-01 9.09782350e-01 4.46175085e-03 -1.51120353e+00
-3.40077370e-01 4.30535711e-02 -1.05377793e+00 5.82734585e-01
3.20982277e-01 1.04342349e-01 -8.72346282e-01 1.12099135e+00
5.74866474e-01 2.21605942e-01 -1.79602250e-01 7.61098564e-01
7.72170126e-01 4.31430966e-01 -7.76510775e-01 -1.56178966e-01
1.17081690e+00 -5.12086689e-01 -4.10800159e-01 3.45037043e-01
1.89155787e-01 -1.05046403e+00 1.06145775e+00 6.20550394e-01
-1.00805521e+00 -3.50595832e-01 -9.88498390e-01 -3.68981361e-02
1.67860575e-02 1.91599354e-01 3.75041366e-01 2.01662719e-01
-8.47351730e-01 8.34621370e-01 -8.79222274e-01 1.72520094e-02
2.95224905e-01 4.84682322e-01 -5.06295562e-01 1.49115860e-01
-6.04957819e-01 6.52130067e-01 -3.87528539e-01 3.12685162e-01
5.49095832e-02 -7.94486821e-01 -7.48336732e-01 -8.28112885e-02
3.77763472e-02 -6.42642915e-01 9.06382680e-01 -5.72894096e-01
-1.30204511e+00 7.47207403e-01 -3.91623557e-01 1.43934011e-01
7.29360521e-01 -3.07027519e-01 -3.72553021e-01 1.24486081e-01
2.46585578e-01 3.44979823e-01 9.27480936e-01 -1.46200669e+00
-6.28006399e-01 -6.14369869e-01 -7.18907356e-01 1.45815417e-01
-2.12550119e-01 -3.16470563e-01 -5.07281303e-01 -3.45257044e-01
1.01021469e+00 -7.12104201e-01 -3.52041781e-01 5.72836161e-01
-5.81178427e-01 -4.25815076e-01 1.23142219e+00 -2.65461683e-01
8.34183931e-01 -2.29958010e+00 -2.06691921e-01 1.07157660e+00
4.77239311e-01 -1.66761413e-01 1.20262401e-02 2.09551215e-01
-3.31586227e-02 -1.99700013e-01 -2.70717442e-01 -1.56502128e-01
-1.54693305e-01 1.10057987e-01 -1.16175838e-01 8.99519682e-01
1.59185469e-01 5.28158903e-01 -4.61379200e-01 -4.27683204e-01
3.11026931e-01 6.59540117e-01 -5.19076705e-01 -2.56849408e-01
3.13535064e-01 8.96547139e-02 -5.18055201e-01 7.40712047e-01
1.25910509e+00 1.57568201e-01 -4.97224122e-01 -6.77130282e-01
-2.59408474e-01 -3.61952364e-01 -1.94906950e+00 1.32359624e+00
-5.99001721e-02 3.82225513e-01 9.70582888e-02 -6.39879286e-01
1.43867362e+00 7.20384568e-02 8.02178502e-01 -3.31880301e-01
3.80818583e-02 5.97932696e-01 -2.09804639e-01 -1.86523587e-01
5.22821069e-01 1.74240172e-01 1.23805784e-01 3.67762666e-04
-2.59686261e-01 -5.34697533e-01 -3.88477266e-01 -2.38894612e-01
7.52631247e-01 -2.29304954e-01 3.57332021e-01 -4.00807202e-01
7.09967792e-01 -2.42894683e-02 7.70385325e-01 3.96651715e-01
-6.60857977e-03 1.08808947e+00 4.10222530e-01 -7.52050161e-01
-1.07837260e+00 -1.20999444e+00 -7.49990582e-01 4.13797319e-01
6.21149898e-01 -3.11534941e-01 -6.76238179e-01 -4.81987327e-01
2.71113366e-01 3.13552976e-01 -4.81083751e-01 -9.33282822e-02
-9.04757917e-01 -3.64016324e-01 6.63384274e-02 4.40462559e-01
3.68000001e-01 -7.06800222e-01 -5.39306402e-01 -1.08981162e-01
1.76221132e-01 -6.21507704e-01 -5.94529867e-01 2.83335466e-02
-9.93351042e-01 -1.48284733e+00 -6.50142372e-01 -9.43043709e-01
1.04672623e+00 3.08863014e-01 8.66132140e-01 2.73294568e-01
-2.34527588e-01 2.94129074e-01 -3.84006351e-02 -8.86850953e-02
-6.31956235e-02 -2.40926579e-01 1.32203609e-01 5.71181625e-02
6.33283734e-01 -5.48990309e-01 -6.72628224e-01 6.68187618e-01
-4.93521035e-01 -2.07924977e-01 5.33408701e-01 9.40237939e-01
1.20606887e+00 -1.55894220e-01 2.66182851e-02 -5.55807829e-01
4.68243718e-01 -5.90771399e-02 -6.76752269e-01 6.85205311e-02
-3.30255598e-01 8.38850588e-02 4.68019009e-01 -2.89206088e-01
-7.51216590e-01 2.88354188e-01 -2.27010116e-01 -5.05711854e-01
-2.44312540e-01 -3.73986363e-02 -2.89925396e-01 -5.16209662e-01
7.25706339e-01 1.72406346e-01 2.72906274e-01 -5.25084436e-01
1.77967012e-01 5.78383625e-01 6.47571027e-01 -4.15053070e-01
9.64311779e-01 9.18123662e-01 5.30489564e-01 -1.25190091e+00
-1.12285711e-01 -7.74813533e-01 -9.29555237e-01 -1.43137753e-01
3.65982741e-01 -3.88373822e-01 -7.35554457e-01 6.13549292e-01
-1.51294589e+00 3.96128833e-01 -6.39494836e-01 5.01017153e-01
-5.57825744e-01 7.38368392e-01 -4.96049047e-01 -3.73501539e-01
-4.28541720e-01 -1.21086204e+00 1.26131928e+00 1.74326271e-01
-1.45666182e-01 -6.68265283e-01 3.34106952e-01 -2.52334297e-01
1.66167602e-01 3.49644691e-01 7.49432981e-01 -6.15636647e-01
-4.62393284e-01 -5.33789754e-01 -8.93604308e-02 4.39846873e-01
2.04546943e-01 5.69218636e-01 -7.21425712e-01 -1.88925296e-01
3.70754540e-01 2.67101705e-01 6.65332139e-01 5.97784698e-01
9.49108660e-01 -1.14401542e-01 -6.19828463e-01 9.59744275e-01
1.51737869e+00 -4.13317047e-02 6.92969680e-01 3.39398712e-01
8.38671923e-01 4.57362294e-01 5.31022906e-01 4.17476863e-01
4.03993018e-02 6.39769495e-01 6.23616278e-01 -1.53787449e-01
1.15597226e-01 1.32485554e-01 -7.99723994e-03 1.03862345e+00
-3.17399502e-01 3.68566364e-01 -7.40191460e-01 1.73389077e-01
-1.71133268e+00 -7.41490901e-01 -8.99246633e-01 2.44844651e+00
4.31380957e-01 5.64913116e-02 -2.00637773e-01 4.16997999e-01
8.76361012e-01 -1.36538759e-01 -6.03816152e-01 -2.51880527e-01
-1.23091154e-01 3.37713182e-01 4.67775255e-01 7.45500624e-01
-1.15818036e+00 4.48389888e-01 6.39040041e+00 8.42113972e-01
-9.60300386e-01 -2.57701457e-01 8.15340728e-02 2.75630653e-01
-5.81508756e-01 -7.87539110e-02 -9.69614744e-01 3.02382499e-01
8.45779330e-02 8.78152624e-03 -8.10125396e-02 1.15789402e+00
-1.72887668e-01 7.60123059e-02 -9.04505312e-01 1.03296530e+00
1.69056267e-01 -1.30212665e+00 2.48573050e-01 1.35750279e-01
6.97164357e-01 1.84816062e-01 -6.75061420e-02 -1.68199316e-01
-2.49033108e-01 -7.83798039e-01 5.18275976e-01 8.06120813e-01
7.79992700e-01 -8.92084420e-01 7.72823632e-01 3.15229148e-01
-1.34643447e+00 3.32236797e-01 -6.84407890e-01 1.18361734e-01
7.52979666e-02 8.31306040e-01 -4.73209441e-01 6.84226334e-01
5.96541047e-01 6.92095637e-01 -3.44397277e-01 1.21992695e+00
1.04229391e-01 -1.55198947e-01 -8.28282773e-01 6.38298914e-02
-5.44162281e-02 -6.60611570e-01 7.44456768e-01 8.76480281e-01
6.44817531e-01 -7.61868656e-02 1.32081658e-01 9.54708815e-01
1.89521044e-01 3.67459625e-01 -5.64896643e-01 1.02146041e+00
7.69330740e-01 1.40474904e+00 -7.25460470e-01 1.99594758e-02
-5.26197314e-01 9.14156914e-01 4.43695188e-02 2.29619056e-01
-6.06205404e-01 -7.20964015e-01 7.62995720e-01 4.01603073e-01
8.62785056e-02 -3.09496671e-01 -5.71048677e-01 -9.86787140e-01
5.22737622e-01 -4.63389069e-01 -2.15966464e-03 -3.52659762e-01
-1.45235288e+00 5.06190836e-01 -1.01728216e-02 -1.63859713e+00
-7.59597868e-02 -8.27838182e-01 -1.06867862e+00 7.84711421e-01
-1.30749834e+00 -8.90233099e-01 -6.51480973e-01 8.13377440e-01
3.02273542e-01 -1.37737706e-01 6.78965032e-01 2.01902062e-01
-1.24481671e-01 6.02231562e-01 2.65242875e-01 1.98547691e-02
5.09443879e-01 -1.04383993e+00 5.54260015e-01 6.79950118e-01
4.66898121e-02 6.09196305e-01 3.31590623e-01 -8.20961177e-01
-1.40829325e+00 -7.29581296e-01 5.47381759e-01 -2.94104636e-01
2.70316601e-01 -1.73052445e-01 -1.25822914e+00 5.92521131e-01
-3.94872695e-01 4.23223734e-01 1.54251710e-01 -1.43094093e-01
2.81730127e-02 -1.11778244e-01 -1.23956740e+00 5.44460952e-01
8.89192343e-01 -9.34549272e-02 -7.40409911e-01 1.99476957e-01
3.83283466e-01 -4.99731749e-01 -7.72647023e-01 8.99734914e-01
6.28989756e-01 -1.22721410e+00 1.20354450e+00 -1.24140292e-01
-6.81607574e-02 -5.00028014e-01 -2.12947100e-01 -1.03219795e+00
-5.27400315e-01 -5.10413826e-01 -5.67995869e-02 1.02066910e+00
4.27327082e-02 -7.34167993e-01 1.14113557e+00 3.54546189e-01
-5.50306022e-01 -1.15789127e+00 -1.37757111e+00 -7.37796783e-01
5.44454865e-02 -3.37977558e-01 6.86187863e-01 8.98593903e-01
-3.45229894e-01 -7.02868178e-02 1.71119347e-01 2.34949499e-01
8.24574590e-01 1.38231024e-01 8.52244854e-01 -1.90856791e+00
3.42132062e-01 -6.49718463e-01 -7.10794806e-01 -1.05422723e+00
1.41995745e-02 -6.70693278e-01 -1.52936196e-02 -1.39222693e+00
-2.15585321e-01 -6.18096054e-01 7.01342151e-02 2.61435866e-01
1.79259792e-01 1.95299372e-01 -3.72213334e-01 5.60666323e-01
6.00794004e-03 6.06840670e-01 1.35337830e+00 1.33902594e-01
-3.13032299e-01 2.08422676e-01 -2.80575931e-01 1.48356390e+00
6.65235579e-01 -3.38565618e-01 7.69329891e-02 -2.25630388e-01
-5.90589307e-02 -3.72931004e-01 4.19650376e-01 -1.08713031e+00
5.29398203e-01 -2.26384431e-01 6.31352842e-01 -9.87207472e-01
3.88590008e-01 -1.14970100e+00 2.06044167e-01 4.90432322e-01
2.28642330e-01 1.98058575e-01 -2.93674111e-01 4.44069296e-01
-1.72162205e-01 -4.13924396e-01 8.44173729e-01 7.41126612e-02
-5.91814995e-01 5.37669122e-01 1.73879310e-01 -1.27744973e-01
1.18717372e+00 -8.85345161e-01 -1.50964484e-01 1.94126107e-02
-6.02840185e-01 -8.70584045e-03 8.03771853e-01 1.38205960e-01
1.21000516e+00 -1.73785758e+00 -7.99188137e-01 9.50686336e-01
8.67499486e-02 4.81011719e-01 6.78996742e-02 7.96218753e-01
-8.12886655e-01 -9.37584043e-02 -2.44139433e-01 -1.04061651e+00
-1.18421626e+00 -7.66955391e-02 4.43340629e-01 4.27931637e-01
-9.64085460e-01 7.01319396e-01 -8.39054286e-02 -6.00010633e-01
1.62132367e-01 -6.41841829e-01 -1.16567075e-01 -9.47135016e-02
3.10734302e-01 8.26222658e-01 4.11397487e-01 -8.30089271e-01
-3.98397356e-01 1.37525213e+00 -3.62897590e-02 1.76056787e-01
1.17951441e+00 2.75553823e-01 -2.26754010e-01 2.52043962e-01
1.40904629e+00 4.53908116e-01 -1.19082069e+00 -1.34591937e-01
-4.96008098e-02 -7.11796701e-01 -3.25595774e-02 1.35493696e-01
-1.17546523e+00 5.07199705e-01 5.76600611e-01 1.15166761e-01
7.31179953e-01 3.35731618e-02 8.74863207e-01 4.95110780e-01
3.94393265e-01 -1.02619290e+00 -2.66575336e-01 5.12158215e-01
1.07222545e+00 -1.08314919e+00 1.83988616e-01 -8.18773329e-01
-1.87838167e-01 1.56497550e+00 6.83142304e-01 -7.12375939e-01
8.17214847e-01 3.00970346e-01 -7.04073980e-02 -5.04664123e-01
-1.63817525e-01 2.84259349e-01 5.87467015e-01 6.94561303e-01
1.21179402e-01 -9.82461348e-02 -3.61472189e-01 4.69682932e-01
-2.97873020e-01 -2.81078756e-01 3.00251007e-01 7.28752792e-01
-7.51489460e-01 -8.88146698e-01 -7.03492045e-01 6.90707982e-01
-8.18554312e-02 2.97601014e-01 -7.11786300e-02 1.07084763e+00
7.51275867e-02 2.90572077e-01 6.27924323e-01 -3.22359532e-01
7.31699646e-01 -2.65750229e-01 3.34138334e-01 -3.41992199e-01
-1.75792560e-01 1.15833558e-01 -3.87972504e-01 -6.67290747e-01
-9.94047299e-02 -8.58258307e-01 -1.52903152e+00 -3.24798912e-01
-6.17322087e-01 1.38095826e-01 5.99877059e-01 5.01533091e-01
2.21101314e-01 -4.95520756e-02 9.23347175e-01 -1.06263649e+00
-8.68078828e-01 -5.60741007e-01 -7.10032701e-01 5.24602056e-01
2.89678723e-01 -8.29859734e-01 -8.21948707e-01 -2.66795307e-01] | [7.911673545837402, -3.0505049228668213] |
8ded5ff5-2889-4073-a06f-4d23197a00bf | supertagging-with-ccg-primitives | null | null | https://aclanthology.org/2020.repl4nlp-1.23 | https://aclanthology.org/2020.repl4nlp-1.23.pdf | Supertagging with CCG primitives | In CCG and other highly lexicalized grammars, supertagging a sentence{'}s words with their lexical categories is a critical step for efficient parsing. Because of the high degree of lexicalization in these grammars, the lexical categories can be very complex. Existing approaches to supervised CCG supertagging treat the categories as atomic units, even when the categories are not simple; when they encounter words with categories unseen during training, their guesses are accordingly unsophisticated. In this paper, we make use of the primitives and operators that constitute the lexical categories of categorial grammars. Instead of opaque labels, we treat lexical categories themselves as linear sequences. We present an LSTM-based model that replaces standard word-level classification with prediction of a sequence of primitives, similarly to LSTM decoders. Our model obtains state-of-the-art word accuracy for single-task English CCG supertagging, increases parser coverage and F1, and is able to produce novel categories. Analysis shows a synergistic effect between this decomposed view and incorporation of prediction history. | ['Gerald Penn', 'Aditya Bhargava'] | 2020-07-01 | null | null | null | ws-2020-7 | ['ccg-supertagging'] | ['natural-language-processing'] | [ 3.20716172e-01 7.98101306e-01 -2.17981219e-01 -5.93420565e-01
-8.24556530e-01 -1.03323090e+00 6.47830069e-01 2.60227889e-01
-4.04747635e-01 5.51544189e-01 3.89948845e-01 -9.89286304e-01
5.66303790e-01 -1.03341949e+00 -6.08329237e-01 -4.47321355e-01
6.17284998e-02 6.76883817e-01 4.85206455e-01 -3.04084897e-01
2.85191946e-02 -1.60028681e-01 -1.58805430e+00 6.88246965e-01
7.44224012e-01 6.46708488e-01 6.16258323e-01 4.72010493e-01
-6.73921525e-01 6.80398285e-01 -4.55280274e-01 -6.48493826e-01
-9.75565165e-02 -4.83145207e-01 -1.07329881e+00 -3.03377248e-02
3.52086961e-01 -1.37942405e-02 2.43296489e-01 1.00757051e+00
8.86814371e-02 8.44403654e-02 4.91303533e-01 -4.74070042e-01
-5.88389277e-01 1.51463664e+00 7.42284581e-02 -4.32987045e-03
3.29313785e-01 -2.08554134e-01 1.67899621e+00 -8.91363323e-01
7.22118855e-01 1.40092707e+00 6.79607928e-01 9.98916328e-01
-1.44008946e+00 -5.14758706e-01 5.92945993e-01 -5.65959066e-02
-1.01401436e+00 -3.53923142e-01 3.68593484e-01 -3.98631513e-01
1.52677143e+00 5.33722527e-02 6.64391816e-01 8.30005825e-01
3.17693889e-01 6.92651570e-01 9.67112541e-01 -1.08145964e+00
2.73105502e-01 -6.82890564e-02 5.64972222e-01 6.28053725e-01
1.31073713e-01 1.05191745e-01 -4.78506118e-01 8.20479468e-02
3.13697249e-01 -4.03764069e-01 2.02538460e-01 -1.24620959e-01
-1.21407497e+00 1.11788344e+00 1.41513020e-01 7.17044234e-01
-1.75986409e-01 4.31836188e-01 5.25918782e-01 3.87083501e-01
6.79012060e-01 4.72292662e-01 -7.67493784e-01 -6.59891590e-02
-5.68285882e-01 3.59344520e-02 8.03951979e-01 1.25035954e+00
7.14412391e-01 2.32528552e-01 1.52199179e-01 9.25272107e-01
2.04204962e-01 2.12013170e-01 7.50321686e-01 -6.45652771e-01
2.94803351e-01 4.70536113e-01 -3.00406665e-01 -3.92737001e-01
-4.76571769e-01 -4.57264781e-01 -1.80333197e-01 -3.07262689e-01
5.19614398e-01 -6.34709671e-02 -1.27102029e+00 2.20977426e+00
1.80336267e-01 -4.54077303e-01 1.45538166e-01 3.19707572e-01
7.58770466e-01 6.66860461e-01 9.64949787e-01 -1.86132699e-01
1.50307071e+00 -5.08950889e-01 -7.12400079e-01 -8.22239161e-01
1.16033638e+00 -4.87388343e-01 1.35142565e+00 2.22310230e-01
-1.13248634e+00 -6.69193447e-01 -7.57667482e-01 -2.68783808e-01
-5.51922619e-01 -3.54582280e-01 9.81664240e-01 7.75427699e-01
-1.39665914e+00 7.02334344e-01 -8.42886806e-01 -6.20093524e-01
-6.79249540e-02 5.83131790e-01 -1.60507336e-01 2.23779157e-01
-1.34663641e+00 1.02945578e+00 8.47346246e-01 -1.84571430e-01
-5.18717706e-01 -3.08792889e-01 -1.02900875e+00 -1.07959528e-02
2.61206537e-01 -3.95584524e-01 1.68631172e+00 -8.71711016e-01
-1.17509317e+00 1.36211538e+00 -2.91438460e-01 -4.19196606e-01
-1.69464350e-01 6.12066425e-02 -2.99552470e-01 -3.18945229e-01
3.06714654e-01 9.07876253e-01 5.00164330e-01 -1.14930916e+00
-1.32877040e+00 -2.67463118e-01 -5.87720349e-02 2.26413056e-01
-7.41938427e-02 1.84662566e-01 4.30626646e-02 -6.18382752e-01
7.32169449e-01 -9.85933661e-01 -1.08410440e-01 -9.25489366e-01
-4.94654596e-01 -8.63316476e-01 1.71470746e-01 -6.87208474e-01
1.37828159e+00 -2.02527809e+00 2.27912769e-01 -2.73945555e-02
3.32277834e-01 -9.99833364e-03 1.13109380e-01 3.67603481e-01
-8.87944624e-02 5.96657932e-01 -1.08753823e-01 -4.49772835e-01
3.84212255e-01 5.88377714e-01 -4.93216217e-01 -6.53386116e-02
-4.07485962e-02 1.11537170e+00 -1.00801837e+00 -4.08875912e-01
2.23886862e-01 -1.63316995e-01 -5.83770633e-01 4.83858138e-02
-5.27142167e-01 3.95064205e-02 -1.71051428e-01 5.44002891e-01
1.33852288e-01 7.50323236e-02 8.31020415e-01 2.20120251e-01
-2.44397223e-01 1.26679945e+00 -6.48189306e-01 1.40134394e+00
-6.64544463e-01 2.10183725e-01 -4.46697086e-01 -9.16965723e-01
6.41257048e-01 3.54177743e-01 -3.61072272e-01 -4.39370036e-01
1.52333871e-01 6.67875826e-01 3.99259686e-01 1.41284376e-01
6.66041672e-01 -7.64624000e-01 -8.62805009e-01 4.25758302e-01
3.59289706e-01 -4.88872051e-01 1.72961354e-01 1.20125217e-02
1.06791258e+00 2.58927554e-01 6.42358780e-01 -5.88390827e-01
2.24314444e-02 1.65535465e-01 6.06041431e-01 9.12473917e-01
3.16111565e-01 2.60223329e-01 4.77929771e-01 -6.21631265e-01
-9.80091035e-01 -1.13264251e+00 -1.33061171e-01 1.88547897e+00
-3.20529699e-01 -3.93804997e-01 -8.22882235e-01 -7.79565632e-01
-2.12342501e-01 1.49135578e+00 -6.60164416e-01 -1.57396689e-01
-1.01825464e+00 -5.11410654e-01 3.81963104e-01 5.33866346e-01
-8.49412307e-02 -1.52529371e+00 -4.10847068e-01 9.44521308e-01
-2.58842885e-01 -1.03689146e+00 -2.17517853e-01 9.55335259e-01
-8.48842263e-01 -6.52583539e-01 2.06714004e-01 -1.34654319e+00
3.90571028e-01 -1.11964129e-01 1.58442056e+00 3.84152234e-01
4.09003019e-01 -2.27640063e-01 -6.60499454e-01 -4.22794938e-01
-1.05422390e+00 4.47700292e-01 -1.38924465e-01 -4.22785044e-01
9.70858991e-01 -5.34011602e-01 1.12399474e-01 -4.05660504e-03
-6.45059943e-01 2.60791808e-01 3.81557703e-01 9.49711502e-01
5.75987399e-01 -4.12314415e-01 4.42813396e-01 -1.65460289e+00
3.19187969e-01 -4.30893093e-01 -4.38473910e-01 1.26515806e-01
-2.70024329e-01 1.78933382e-01 8.47170234e-01 -4.76475000e-01
-1.04270327e+00 1.30444035e-01 -3.65263760e-01 4.31522012e-01
-1.94015235e-01 5.98015904e-01 -3.97076100e-01 2.95977622e-01
6.43728256e-01 2.92576492e-01 -2.81687051e-01 -5.78823626e-01
6.17003143e-01 6.18146002e-01 3.93601388e-01 -7.36819267e-01
5.68833172e-01 -1.33396879e-01 -4.29174572e-01 -7.24444211e-01
-1.02628267e+00 -2.57761776e-01 -9.41539884e-01 -2.25524287e-04
9.70025301e-01 -6.87628925e-01 -1.76176101e-01 3.62666011e-01
-1.41701174e+00 -6.92205727e-01 -4.62528259e-01 1.87862813e-01
-5.00852704e-01 3.56126726e-01 -9.86690223e-01 -4.71539885e-01
-1.27436429e-01 -8.60711992e-01 1.13649523e+00 -3.13078165e-01
-6.16435170e-01 -1.33526599e+00 -3.52028817e-01 -1.39989689e-01
3.21472347e-01 -2.32844353e-02 1.72071767e+00 -1.12409019e+00
-4.83170539e-01 -8.99153873e-02 1.41745806e-01 2.14307725e-01
2.12668911e-01 -3.95963371e-01 -9.59451854e-01 -1.42303720e-01
-7.85748437e-02 -1.87926158e-01 7.11502969e-01 3.29067409e-01
6.90633655e-01 -3.63324344e-01 -4.66853768e-01 4.26026821e-01
1.24674749e+00 5.80950975e-01 4.09867942e-01 1.88307926e-01
7.82574415e-01 9.10823762e-01 4.51911181e-01 -5.16866893e-02
6.16143584e-01 5.36738813e-01 2.66027212e-01 9.51441154e-02
-2.96921730e-01 -4.71822202e-01 3.50648940e-01 1.38953817e+00
1.52793393e-01 -4.72829640e-01 -1.10708940e+00 7.80324578e-01
-1.44669366e+00 -5.67636549e-01 -3.13206851e-01 2.18099952e+00
1.13225651e+00 4.70434010e-01 -1.68616906e-01 1.62254777e-02
9.90983903e-01 4.86802272e-02 -7.95038044e-02 -8.58279884e-01
-9.75118019e-03 6.33814692e-01 4.16285187e-01 7.47430384e-01
-1.18133628e+00 1.92653477e+00 6.74296808e+00 9.28178549e-01
-8.00075233e-01 4.77936268e-01 5.34462988e-01 4.19404685e-01
-7.71408379e-01 3.02124888e-01 -1.19092810e+00 4.13571388e-01
1.09520257e+00 -8.43684971e-02 2.90574908e-01 8.62294078e-01
-3.91819000e-01 5.97596839e-02 -1.44386756e+00 3.97544175e-01
-2.53514826e-01 -1.08599472e+00 2.43167847e-01 1.85826078e-01
2.22582161e-01 3.60705666e-02 -2.26566449e-01 6.84859931e-01
1.07860315e+00 -8.41318130e-01 1.29821384e+00 -1.74024150e-01
1.07023144e+00 -5.30897319e-01 7.14652896e-01 1.67570382e-01
-1.16827059e+00 3.07639632e-02 -6.46944642e-01 -3.80520105e-01
2.35048622e-01 1.57702774e-01 -9.50775266e-01 -2.85266247e-02
2.43135273e-01 4.27487582e-01 -5.86100578e-01 2.59713382e-01
-8.31192434e-01 1.08982742e+00 -3.26435119e-01 -4.51524019e-01
4.04931605e-01 4.62609194e-02 3.10654521e-01 1.43331325e+00
2.31107250e-01 2.70106405e-01 4.38277215e-01 4.36696112e-01
9.45590511e-02 2.84629673e-01 -6.74200833e-01 7.71695934e-03
7.96135187e-01 8.47948015e-01 -1.01270592e+00 -5.91797590e-01
-5.23176014e-01 5.17129064e-01 6.72793806e-01 -7.83975050e-02
-3.44965011e-01 -1.92133531e-01 5.46509564e-01 1.58740759e-01
2.45053351e-01 -3.04083586e-01 -4.80655283e-01 -9.71859574e-01
-1.70170173e-01 -8.03263545e-01 6.12346947e-01 -4.60914969e-01
-1.12932765e+00 9.17340636e-01 1.15778990e-01 -7.89989948e-01
-9.07114685e-01 -7.00769663e-01 -3.11810285e-01 7.07038939e-01
-1.09187233e+00 -1.58740091e+00 3.99165154e-01 1.17034391e-01
7.98595905e-01 -1.51462466e-01 1.18331718e+00 -6.34907186e-02
-2.39852984e-02 5.56245685e-01 -2.12346405e-01 1.62463680e-01
2.03033686e-01 -1.74127531e+00 1.25039279e+00 1.19289744e+00
2.63487518e-01 7.00752616e-01 7.95300245e-01 -8.34569693e-01
-8.09673548e-01 -1.09052527e+00 1.83277881e+00 -6.47826672e-01
9.51990247e-01 -1.02011299e+00 -1.04260921e+00 1.21124017e+00
7.55932480e-02 -4.07029599e-01 7.10497856e-01 6.50235176e-01
-3.42346817e-01 2.63093114e-01 -6.81215703e-01 6.67668343e-01
1.47364509e+00 -5.30670047e-01 -1.26370692e+00 4.90729719e-01
1.23597884e+00 -3.03705096e-01 -5.35233498e-01 2.80651804e-02
3.18701088e-01 -5.96104205e-01 4.81621265e-01 -7.44632423e-01
1.44547313e-01 8.78582373e-02 -4.46333379e-01 -1.56934285e+00
-6.43446326e-01 -4.77159590e-01 3.75007689e-01 1.13996756e+00
7.74349093e-01 -7.79018462e-01 6.90214932e-01 6.04789257e-01
-8.12889755e-01 -1.82662562e-01 -1.06926548e+00 -6.35885954e-01
5.05121887e-01 -6.76785827e-01 6.36747301e-01 7.97872543e-01
3.13756675e-01 5.85253954e-01 5.17873242e-02 -2.62375146e-01
5.89598775e-01 3.53393815e-02 1.82871893e-01 -1.45641303e+00
-3.90055358e-01 -2.21904159e-01 -3.63970071e-01 -8.82966220e-01
3.77011269e-01 -1.24042928e+00 5.85287035e-01 -1.38952935e+00
-2.52828360e-01 -8.72040927e-01 -9.83583927e-02 1.17309988e+00
-1.28775284e-01 2.73251712e-01 2.54316479e-01 -3.66848223e-02
-4.71613765e-01 3.66704725e-03 5.65725744e-01 1.96489125e-01
-2.40591109e-01 -1.71642095e-01 -8.36821556e-01 9.18977737e-01
8.62888873e-01 -7.02469707e-01 -1.56594962e-01 -7.07134128e-01
2.50794888e-01 2.46414132e-02 -2.05921844e-01 -7.65723407e-01
5.48488647e-02 -1.08750112e-01 -1.14908099e-01 -4.33451921e-01
1.24952383e-01 -2.40004659e-01 1.58199385e-01 5.60450077e-01
-4.19050723e-01 7.09330365e-02 1.35018140e-01 7.70463571e-02
-8.45416784e-02 -3.92474920e-01 7.67618418e-01 -6.33028924e-01
-1.14258182e+00 -7.69374222e-02 -9.39407110e-01 2.77600259e-01
4.77019727e-01 -4.49679047e-01 -1.85700580e-01 -7.00026527e-02
-1.06926596e+00 -1.84044346e-01 5.32756329e-01 4.27714348e-01
3.38241994e-01 -1.02907383e+00 -4.67861146e-01 3.41470569e-01
1.54492646e-01 1.76530145e-03 -1.77343488e-01 -5.75961545e-03
-4.12785560e-01 4.65814739e-01 -7.99012631e-02 -3.85622978e-01
-9.82598245e-01 6.87632263e-01 1.60304919e-01 -2.68432736e-01
-5.28892756e-01 1.18984997e+00 6.58075571e-01 -7.37016737e-01
-4.28034849e-02 -6.15824819e-01 -2.98725456e-01 2.51255900e-01
1.93056732e-01 -3.91213924e-01 4.08123076e-01 -8.20460320e-01
-3.63881588e-01 2.86471725e-01 -2.92664737e-01 2.54153684e-02
1.25830245e+00 -2.46274054e-01 -3.52464169e-01 7.08565831e-01
7.70857632e-01 1.18791066e-01 -7.58875310e-01 -3.54460269e-01
6.07245088e-01 3.15269940e-02 -1.80993393e-01 -8.02391827e-01
-4.60166931e-01 9.41665769e-01 6.46569803e-02 4.58761692e-01
7.48763263e-01 4.12273943e-01 8.28138888e-01 2.94158489e-01
9.35897708e-01 -1.02789378e+00 -3.36895019e-01 1.01132381e+00
4.67322320e-01 -7.79191554e-01 -5.86731076e-01 -8.81140888e-01
-3.60324413e-01 1.04283273e+00 5.02211154e-01 -8.01132172e-02
4.96927291e-01 1.46521151e-01 1.95178345e-01 -1.21549582e-02
-9.89698350e-01 -5.51691830e-01 1.43934023e-02 6.44174993e-01
6.73474669e-01 5.78290522e-01 -6.42661393e-01 7.93413520e-01
-7.76382387e-01 -7.43310273e-01 7.02234626e-01 8.48359227e-01
-9.03038383e-01 -1.52862215e+00 2.41902359e-02 6.81649148e-01
-7.27940500e-01 -7.49606788e-01 -1.05245344e-01 1.05143583e+00
3.47630620e-01 9.76344109e-01 3.15882951e-01 -5.07879853e-01
2.10385352e-01 6.61564171e-01 2.68510669e-01 -1.62320912e+00
-7.78915286e-01 2.52405833e-02 6.40766084e-01 -2.94521779e-01
-4.88586873e-02 -8.49973917e-01 -1.37365615e+00 -1.88520491e-01
-3.53603959e-01 3.99290711e-01 6.09660983e-01 1.18780935e+00
-6.04608171e-02 3.33055466e-01 2.62025386e-01 -6.66541576e-01
-4.92900968e-01 -1.08230960e+00 -7.42109179e-01 3.47885758e-01
-9.13873166e-02 -6.72763228e-01 -4.49190050e-01 9.18048173e-02] | [10.451216697692871, 9.576507568359375] |
08123195-2cf3-418c-a2cb-f24e124a5ca4 | setrank-a-setwise-bayesian-approach-for | 2002.09841 | null | https://arxiv.org/abs/2002.09841v1 | https://arxiv.org/pdf/2002.09841v1.pdf | SetRank: A Setwise Bayesian Approach for Collaborative Ranking from Implicit Feedback | The recent development of online recommender systems has a focus on collaborative ranking from implicit feedback, such as user clicks and purchases. Different from explicit ratings, which reflect graded user preferences, the implicit feedback only generates positive and unobserved labels. While considerable efforts have been made in this direction, the well-known pairwise and listwise approaches have still been limited by various challenges. Specifically, for the pairwise approaches, the assumption of independent pairwise preference is not always held in practice. Also, the listwise approaches cannot efficiently accommodate "ties" due to the precondition of the entire list permutation. To this end, in this paper, we propose a novel setwise Bayesian approach for collaborative ranking, namely SetRank, to inherently accommodate the characteristics of implicit feedback in recommender system. Specifically, SetRank aims at maximizing the posterior probability of novel setwise preference comparisons and can be implemented with matrix factorization and neural networks. Meanwhile, we also present the theoretical analysis of SetRank to show that the bound of excess risk can be proportional to $\sqrt{M/N}$, where $M$ and $N$ are the numbers of items and users, respectively. Finally, extensive experiments on four real-world datasets clearly validate the superiority of SetRank compared with various state-of-the-art baselines. | ['Chuan Qin', 'HengShu Zhu', 'Hui Xiong', 'Chen Zhu', 'Chao Wang'] | 2020-02-23 | null | null | null | null | ['collaborative-ranking'] | ['graphs'] | [ 4.79080677e-02 -2.62299478e-01 -2.86784410e-01 -7.66048849e-01
-6.68516099e-01 -5.47498643e-01 2.34320715e-01 -1.36867100e-02
-2.31477499e-01 7.27119446e-01 3.06139827e-01 -2.64887720e-01
-8.28674436e-01 -7.23417282e-01 -5.79479635e-01 -6.40408993e-01
-3.13902646e-01 4.88259614e-01 -6.45018592e-02 -2.35368729e-01
5.58955550e-01 -1.27316132e-01 -1.57875824e+00 3.98504704e-01
1.04287136e+00 1.09498000e+00 8.68452042e-02 1.86862126e-01
5.67912161e-02 5.64057112e-01 -1.58485413e-01 -7.09382355e-01
4.23626333e-01 -3.49077843e-02 -3.89038503e-01 -2.45749697e-01
4.46573526e-01 -6.92549407e-01 -4.09214735e-01 1.17719066e+00
5.00661373e-01 5.01023948e-01 6.18667781e-01 -1.20043552e+00
-1.03547931e+00 1.05459416e+00 -6.14846587e-01 6.80297017e-02
3.72249573e-01 -4.01095361e-01 1.64392722e+00 -1.21888626e+00
6.01834804e-02 1.20114326e+00 5.05863786e-01 1.15163863e-01
-9.49635029e-01 -9.60027337e-01 7.16597497e-01 2.45387420e-01
-1.36718798e+00 -4.29397300e-02 6.63997531e-01 -3.97849083e-01
3.01115751e-01 5.01690984e-01 3.14060122e-01 7.28725612e-01
-3.17145824e-01 1.03088129e+00 1.22429299e+00 -2.30029270e-01
9.46179926e-02 3.31574231e-01 5.75713158e-01 3.82718384e-01
3.88015687e-01 3.20515305e-01 -5.58269918e-01 -5.58851898e-01
6.40747368e-01 5.69500744e-01 -3.42736721e-01 -5.24651229e-01
-9.13013816e-01 8.74090850e-01 3.68410110e-01 -1.52045161e-01
-1.88426629e-01 -7.78916106e-02 1.12527952e-01 2.82996684e-01
3.71151924e-01 2.71861315e-01 -6.00733280e-01 1.31802246e-01
-9.24204588e-01 1.82215273e-01 7.64047861e-01 9.36270237e-01
6.22826576e-01 -4.50236708e-01 -2.56257951e-01 9.26372230e-01
7.67851293e-01 4.92095321e-01 2.86744028e-01 -7.78833985e-01
4.44484442e-01 4.67962444e-01 5.59386432e-01 -1.30508685e+00
-5.20062000e-02 -7.62065172e-01 -9.58747089e-01 -2.82392651e-01
3.20978940e-01 -2.26248318e-04 -4.51223582e-01 1.65827250e+00
1.15418166e-01 2.97380775e-01 -3.71623725e-01 9.52972949e-01
5.82151592e-01 5.48978865e-01 -2.57730216e-01 -3.97124797e-01
9.50143337e-01 -8.67899656e-01 -6.09443724e-01 1.73197806e-01
3.97548586e-01 -7.20443428e-01 1.01730430e+00 7.43595898e-01
-8.67379248e-01 -4.89836246e-01 -8.50028396e-01 2.82727718e-01
-2.30529103e-02 1.98034301e-01 9.09871340e-01 8.16827655e-01
-6.91794038e-01 5.79029739e-01 -2.69357473e-01 2.16168910e-01
1.13986619e-01 6.54510856e-01 5.27437739e-02 -1.70237213e-01
-1.42897320e+00 3.92685205e-01 9.05355345e-03 5.87873042e-01
-5.75599611e-01 -7.49402702e-01 -2.59255201e-01 2.78628021e-01
6.01963043e-01 -5.40636003e-01 1.34037626e+00 -7.47526228e-01
-1.56801224e+00 5.77515438e-02 -4.61536832e-02 -1.69410333e-01
3.68278027e-01 -7.57260144e-01 -4.61702079e-01 -4.59545016e-01
-9.27774534e-02 6.67277202e-02 4.45393562e-01 -1.24147666e+00
-8.09817851e-01 -3.14484596e-01 5.07778525e-01 3.39806437e-01
-6.22149289e-01 1.55095488e-01 -2.54926980e-01 -6.06672943e-01
2.09514722e-01 -9.77036893e-01 -4.96592581e-01 -2.12680116e-01
-1.16018042e-01 -4.11525935e-01 2.27755398e-01 -2.75900871e-01
1.48780000e+00 -2.17207503e+00 -1.74455628e-01 6.41826272e-01
6.66352138e-02 8.76943916e-02 -2.03997772e-02 5.57966173e-01
1.56567782e-01 7.76782632e-02 1.37296587e-01 -3.28276418e-02
2.81145155e-01 1.17205391e-02 -6.94132566e-01 3.78428847e-01
-4.23061222e-01 5.42621553e-01 -1.11646307e+00 -1.74351409e-01
-8.22781250e-02 3.08839530e-01 -8.22419524e-01 1.62175626e-01
-1.75476167e-02 5.69413649e-04 -4.82688904e-01 3.26516211e-01
8.29717338e-01 -5.29728889e-01 6.81881666e-01 -2.07437649e-01
-1.01045318e-01 5.73161840e-01 -1.54449630e+00 1.41298378e+00
-2.73230404e-01 -1.50806621e-01 -2.53793262e-02 -8.38906050e-01
8.98860216e-01 2.61879265e-01 4.36256170e-01 -5.33804715e-01
1.14183970e-01 2.29700655e-01 1.32584265e-02 -6.50747791e-02
6.86955750e-01 1.14237927e-01 2.23766062e-02 7.97961831e-01
-3.59394222e-01 6.36701882e-01 1.86531097e-01 4.38049316e-01
7.26687312e-01 3.64782810e-02 9.85675380e-02 -1.60424128e-01
4.61118460e-01 -4.60961789e-01 8.58762920e-01 1.08057487e+00
1.78712323e-01 4.53270048e-01 8.40469524e-02 -1.66271940e-01
-4.13257569e-01 -1.20294356e+00 2.51291431e-02 1.46934402e+00
4.08596605e-01 -5.01158416e-01 -1.42448276e-01 -8.64466548e-01
1.54690087e-01 5.02360463e-01 -5.39955676e-01 5.60367294e-02
-2.74634540e-01 -9.66906846e-01 2.78717862e-03 6.18607163e-01
9.72450897e-02 -6.86541080e-01 1.58374846e-01 1.32405698e-01
-1.49230868e-01 -3.70709807e-01 -7.86996603e-01 5.95611939e-03
-1.01587212e+00 -1.04370570e+00 -5.39975226e-01 -5.47825396e-01
8.66828442e-01 7.93400824e-01 1.03652692e+00 5.21966480e-02
2.15446591e-01 1.89250648e-01 -6.59392059e-01 -1.44218192e-01
4.25461531e-01 -1.50027871e-01 4.47583199e-01 3.26184392e-01
4.37165052e-01 -6.26146376e-01 -1.05952311e+00 8.22711408e-01
-6.97999477e-01 -3.03748995e-01 8.50191534e-01 9.53340769e-01
5.25099397e-01 2.42641479e-01 7.26449609e-01 -1.36989236e+00
7.52928376e-01 -7.79608667e-01 -4.70486313e-01 4.11820084e-01
-1.00065875e+00 -4.25842516e-02 5.04573584e-01 -4.44150746e-01
-1.16738927e+00 -2.31386125e-01 8.29912797e-02 -1.22046970e-01
2.61726320e-01 8.62944067e-01 -2.34813988e-01 2.85598755e-01
2.99869627e-01 4.88083102e-02 -3.06613147e-01 -6.93648815e-01
6.90148830e-01 8.53105009e-01 1.37463838e-01 -6.45848989e-01
7.62324810e-01 3.29673529e-01 -4.68407154e-01 -8.91233087e-02
-1.17196131e+00 -7.24808931e-01 -4.18220162e-01 -1.71612248e-01
7.67021179e-02 -9.01377976e-01 -9.22579348e-01 7.52683729e-02
-6.38118148e-01 1.18965134e-01 -7.00042443e-03 7.80617952e-01
-3.03725511e-01 6.16175950e-01 -5.02807260e-01 -1.18126559e+00
-3.94041985e-01 -8.48996639e-01 3.64110738e-01 2.15956420e-01
-1.39524683e-01 -7.41044581e-01 7.26792589e-02 4.33932304e-01
3.99966478e-01 -5.10425746e-01 9.13777709e-01 -8.16261113e-01
-7.16401875e-01 -2.80108452e-01 -3.61249745e-01 3.59840631e-01
1.89181522e-01 -1.67607054e-01 -6.13331854e-01 -4.02100474e-01
-2.49386042e-01 -3.07055116e-01 7.06497073e-01 2.89558470e-01
1.19958460e+00 -4.34395641e-01 -3.13808680e-01 1.06325172e-01
1.20494211e+00 2.30692297e-01 5.50767839e-01 -6.73389658e-02
4.40709352e-01 5.86080909e-01 1.08241796e+00 7.69726157e-01
3.65418971e-01 5.88407815e-01 3.91489655e-01 1.45299494e-01
4.79234010e-01 -4.15412486e-01 4.28702712e-01 1.22377324e+00
-1.27208069e-01 -3.13139886e-01 -1.62149653e-01 2.35320985e-01
-2.16940689e+00 -1.04788756e+00 -1.84078619e-01 2.72632146e+00
6.92521751e-01 1.00647859e-01 1.01904131e-01 2.07996875e-01
8.81938279e-01 -1.03029385e-01 -4.29595739e-01 -1.65787667e-01
8.03370327e-02 -9.24722385e-03 4.58844602e-01 2.91966110e-01
-8.80356789e-01 3.89817625e-01 6.37846565e+00 7.96520770e-01
-6.08209133e-01 1.13707259e-02 5.25486052e-01 -4.74936217e-01
-5.29928863e-01 1.20334238e-01 -9.13459420e-01 6.75864995e-01
5.56980431e-01 -1.92615822e-01 5.14884353e-01 8.40666234e-01
1.53325289e-01 7.60968029e-02 -1.26905417e+00 8.90595317e-01
-6.04713568e-03 -8.51864994e-01 3.58531401e-02 2.86004484e-01
9.60833609e-01 -1.06669262e-01 4.69711483e-01 5.34171402e-01
8.52988183e-01 -8.52189064e-01 4.94798839e-01 6.64028347e-01
4.88699585e-01 -8.45182538e-01 7.86672533e-01 4.53412473e-01
-1.04181349e+00 -4.71351922e-01 -6.24913454e-01 -4.02108461e-01
2.79597461e-01 1.00034475e+00 -4.66584891e-01 7.45244980e-01
8.33282590e-01 8.22244287e-01 -8.92328545e-02 1.32259226e+00
-2.15348482e-01 7.11488545e-01 -3.88693422e-01 -2.58240551e-01
2.64237821e-02 -4.22172248e-01 1.98900089e-01 8.24305415e-01
5.12838721e-01 3.94439101e-01 4.06110018e-01 4.38574046e-01
-2.86396623e-01 4.30710405e-01 -1.92159086e-01 1.86315402e-01
6.29907370e-01 1.30384362e+00 -2.65209347e-01 -1.13651328e-01
-5.12361646e-01 4.63678658e-01 3.84981215e-01 2.85457373e-01
-7.35613525e-01 -1.43763572e-01 5.27503729e-01 1.45083740e-01
4.58586693e-01 -1.97185367e-01 1.96071304e-02 -1.06135237e+00
3.32090929e-02 -7.38623321e-01 5.22821844e-01 -4.60145742e-01
-1.87279892e+00 1.19506650e-01 -4.50314991e-02 -1.34346628e+00
1.57152250e-01 -5.00486791e-01 -4.90861744e-01 7.24893570e-01
-1.30239701e+00 -6.01438344e-01 -4.33001481e-02 4.40543652e-01
3.31833512e-01 1.87427830e-02 6.84085429e-01 6.90755486e-01
-4.33416724e-01 9.62345064e-01 7.25982547e-01 -2.23738095e-03
1.01371801e+00 -1.13609517e+00 -8.33369642e-02 4.15751457e-01
2.42999375e-01 1.26117384e+00 5.69519579e-01 -5.21619201e-01
-1.49366820e+00 -8.50917816e-01 6.79509401e-01 -3.35750997e-01
6.37053907e-01 -2.44367555e-01 -6.24706149e-01 5.33148706e-01
-2.51250695e-02 -6.40515760e-02 1.25461769e+00 8.55972946e-01
-6.00029409e-01 -3.93651426e-01 -8.17881644e-01 5.42032897e-01
1.21224737e+00 -3.23479265e-01 -5.33614635e-01 3.02572876e-01
3.52784365e-01 1.35757148e-01 -9.40402746e-01 5.48468649e-01
1.09170306e+00 -1.03586578e+00 1.02477241e+00 -5.96730292e-01
3.17857146e-01 -4.93984193e-01 -3.94624531e-01 -1.16745317e+00
-8.62772703e-01 -3.81959110e-01 -2.02220455e-01 1.45813382e+00
4.50213253e-01 -6.41184449e-01 8.67427051e-01 7.48612165e-01
7.62258470e-02 -8.28302145e-01 -5.45851290e-01 -6.87343657e-01
-1.90741792e-01 -1.38100445e-01 6.26684248e-01 7.97575295e-01
8.78402889e-02 2.52497584e-01 -9.25505340e-01 2.41299987e-01
6.26177371e-01 5.10479867e-01 5.56103587e-01 -1.58472264e+00
-6.36600316e-01 -2.47818932e-01 1.06387973e-01 -1.73663211e+00
-1.98518068e-01 -8.27396452e-01 7.20071569e-02 -1.37748742e+00
5.46469331e-01 -1.13795269e+00 -1.15724385e+00 1.32826641e-01
-2.29091153e-01 1.91638783e-01 -2.92966310e-02 3.80408227e-01
-1.01246631e+00 6.47098720e-01 9.97352421e-01 -3.70548181e-02
-1.31204858e-01 2.43872002e-01 -1.15647495e+00 7.02395916e-01
6.34739041e-01 -5.61498761e-01 -7.35928476e-01 -3.77045095e-01
9.57330465e-01 5.96825927e-02 -9.33692753e-02 -4.49497372e-01
3.69383484e-01 -3.40113014e-01 1.47171170e-01 -9.47148919e-01
1.98589250e-01 -8.47887576e-01 1.67600706e-01 -4.02845722e-03
-6.62211895e-01 4.70880382e-02 -5.50184131e-01 1.14874566e+00
-1.78500876e-01 -2.30886191e-01 1.52034238e-01 -4.75633629e-02
-3.36259395e-01 4.95871216e-01 -1.03448108e-01 -1.71413496e-01
6.10037088e-01 8.62311050e-02 -2.50690281e-01 -3.33342403e-01
-5.27915061e-01 4.07760382e-01 5.83008416e-02 4.71805811e-01
5.55493653e-01 -1.30258560e+00 -7.27553487e-01 -1.00214414e-01
3.16301659e-02 -2.66106963e-01 4.18294430e-01 9.13977206e-01
2.50407636e-01 3.89391601e-01 1.98096827e-01 -2.55128920e-01
-1.24124193e+00 5.99298477e-01 -1.44047007e-01 -4.81100529e-01
-9.72162113e-02 9.04825926e-01 5.07517576e-01 -6.30976796e-01
4.13778633e-01 5.58467545e-02 -4.14247364e-01 8.54316875e-02
5.48690975e-01 5.12577057e-01 -1.38930008e-01 -1.71615660e-01
-2.24832073e-01 2.05340102e-01 -6.62617147e-01 1.16939098e-02
1.16966426e+00 -3.47388148e-01 -2.09522426e-01 3.96890283e-01
7.69156992e-01 3.24445069e-01 -9.72991824e-01 -5.69810271e-01
-4.81422432e-02 -7.95071006e-01 -5.85223064e-02 -1.01978469e+00
-1.02331424e+00 5.94400406e-01 4.98012036e-01 1.83364078e-01
8.46862674e-01 -4.12508309e-01 6.58866167e-01 5.62230229e-01
7.42563248e-01 -1.06157482e+00 -1.09777376e-01 4.97044623e-01
5.30713856e-01 -1.11854041e+00 2.98835278e-01 -5.65268040e-01
-4.29835886e-01 5.85298657e-01 5.21052241e-01 -2.90170878e-01
9.88100350e-01 -2.83159494e-01 -1.27116844e-01 1.07489072e-01
-8.29590857e-01 2.28270933e-01 4.67591882e-01 8.31321180e-02
7.56482542e-01 2.63300955e-01 -6.57087386e-01 1.11348903e+00
-4.08727266e-02 2.31044590e-01 1.93515211e-01 7.14500368e-01
-5.61998367e-01 -1.43323755e+00 -1.57601148e-01 8.94027293e-01
-4.66068983e-01 -3.67769986e-01 -1.66930914e-01 2.35893160e-01
-3.76928039e-03 1.18231201e+00 -2.37828359e-01 -5.80330133e-01
1.12356655e-01 -2.64862835e-01 3.80766630e-01 -6.91767454e-01
-6.39191628e-01 -8.66326615e-02 2.27546468e-02 -2.87405849e-01
-3.66515636e-01 -6.52047157e-01 -9.84182239e-01 -2.14172244e-01
-8.55833352e-01 7.50923991e-01 4.31634158e-01 8.20194483e-01
4.53654110e-01 1.04714006e-01 9.68801379e-01 -5.69780350e-01
-1.09623516e+00 -9.18721318e-01 -8.70818913e-01 3.14815611e-01
-3.03839356e-01 -1.06078076e+00 -5.11017621e-01 -3.64774466e-01] | [9.935080528259277, 5.591276168823242] |
b5de4649-dbe0-4fc9-99b7-e6880c4c4b65 | pain-detection-in-masked-faces-during | 2211.06694 | null | https://arxiv.org/abs/2211.06694v1 | https://arxiv.org/pdf/2211.06694v1.pdf | Pain Detection in Masked Faces during Procedural Sedation | Pain monitoring is essential to the quality of care for patients undergoing a medical procedure with sedation. An automated mechanism for detecting pain could improve sedation dose titration. Previous studies on facial pain detection have shown the viability of computer vision methods in detecting pain in unoccluded faces. However, the faces of patients undergoing procedures are often partially occluded by medical devices and face masks. A previous preliminary study on pain detection on artificially occluded faces has shown a feasible approach to detect pain from a narrow band around the eyes. This study has collected video data from masked faces of 14 patients undergoing procedures in an interventional radiology department and has trained a deep learning model using this dataset. The model was able to detect expressions of pain accurately and, after causal temporal smoothing, achieved an average precision (AP) of 0.72 and an area under the receiver operating characteristic curve (AUC) of 0.82. These results outperform baseline models and show viability of computer vision approaches for pain detection of masked faces during procedural sedation. Cross-dataset performance is also examined when a model is trained on a publicly available dataset and tested on the sedation videos. The ways in which pain expressions differ in the two datasets are qualitatively examined. | ['B. Taati', 'A. Conway', 'S. Mafeld', 'Y. Zarghami'] | 2022-11-12 | null | null | null | null | ['medical-procedure'] | ['medical'] | [ 1.18450791e-01 2.77413756e-01 -2.84851462e-01 -3.64667684e-01
-9.34400678e-01 -2.95084059e-01 -1.14826702e-01 2.07539842e-01
-7.81066895e-01 3.03664804e-01 2.17707351e-01 -1.46964699e-01
7.53745735e-02 -1.48809135e-01 -3.49455237e-01 -6.63066864e-01
-5.18097043e-01 -5.82788587e-02 -7.03577220e-01 2.37928554e-01
1.77879512e-01 6.21593118e-01 -1.29524815e+00 7.50399411e-01
-8.96750167e-02 1.36878943e+00 8.96191283e-04 5.92363894e-01
1.29142314e-01 7.89391577e-01 -6.99931860e-01 -2.10821852e-01
3.66043031e-01 -1.91819534e-01 -5.52526414e-01 -6.48388937e-02
5.33553541e-01 -6.93616092e-01 -1.80143043e-01 8.30525219e-01
8.55290711e-01 -2.84404308e-01 3.39376390e-01 -7.85449982e-01
-1.71069410e-02 3.85204926e-02 -1.00162840e+00 4.58099037e-01
6.77837968e-01 1.11659713e-01 1.58069789e-01 -5.68995118e-01
4.97782826e-01 1.09314013e+00 6.27365589e-01 7.16595769e-01
-1.31560206e+00 -9.45909083e-01 -2.52349615e-01 -1.17842406e-01
-1.38063383e+00 -3.09110552e-01 6.29149020e-01 -6.60975575e-01
8.77941847e-01 1.74417809e-01 8.33606601e-01 1.03928185e+00
8.96254182e-01 3.46171767e-01 1.73645794e+00 -2.72007495e-01
3.90450656e-01 2.52213478e-01 -6.48244247e-02 8.53354394e-01
3.01875416e-02 2.68244863e-01 -6.65197790e-01 -5.98001838e-01
7.11431086e-01 -2.95633171e-02 -5.60538650e-01 -6.62113130e-02
-5.33029974e-01 1.28350627e+00 4.26697969e-01 3.08967143e-01
-6.34321213e-01 1.94135085e-01 7.69723535e-01 1.36143848e-01
3.45180631e-01 4.15365249e-01 -3.65111887e-01 -3.81879598e-01
-9.58526313e-01 -3.30431372e-01 8.43467474e-01 3.38834554e-01
3.39170396e-01 -3.81304324e-01 -3.52605641e-01 7.26209044e-01
-2.93496381e-02 3.22876461e-02 6.43945038e-01 -1.29068244e+00
-2.06397012e-01 1.24236837e-01 7.66653046e-02 -8.30478847e-01
-9.57760632e-01 -4.19652462e-02 -3.66508335e-01 8.25505614e-01
9.18836147e-02 -5.08494616e-01 -1.06953788e+00 1.56025791e+00
2.53117625e-02 -1.62619293e-01 -8.63200724e-02 8.56780052e-01
9.96476293e-01 4.97477874e-02 5.01341820e-01 -7.52273381e-01
1.87209618e+00 -6.03110671e-01 -1.04565203e+00 -3.49652439e-01
7.59675860e-01 -9.02544737e-01 8.97094011e-01 9.09793913e-01
-1.06553864e+00 -3.28683972e-01 -9.18159008e-01 2.82594353e-01
1.61781952e-01 2.05458432e-01 5.97412825e-01 9.86088455e-01
-1.08998966e+00 5.58261156e-01 -7.40456223e-01 -1.39489576e-01
1.05572677e+00 7.84650326e-01 -7.75582314e-01 -1.22847080e-01
-4.90925908e-01 1.32338417e+00 -1.14879891e-01 2.32477114e-01
-4.71138865e-01 -5.79682171e-01 -8.62773240e-01 -4.98435088e-02
-6.09644726e-02 -7.74425387e-01 1.56554377e+00 -1.40765452e+00
-1.40241349e+00 1.28768420e+00 -1.05935648e-01 -3.38166714e-01
4.82719511e-01 -4.09295201e-01 -2.27075309e-01 7.79121161e-01
1.15581052e-02 9.44990575e-01 9.61183131e-01 -1.12185943e+00
-1.86598405e-01 -6.48922741e-01 -1.47308022e-01 2.76374310e-01
1.48119316e-01 2.92931795e-01 -8.65080282e-02 -3.43834817e-01
-2.43554458e-01 -1.08022130e+00 -3.69111896e-01 3.76445800e-01
4.03802872e-01 2.79148936e-01 9.57946777e-01 -8.42758536e-01
1.00011253e+00 -2.38359237e+00 -4.50763822e-01 4.05387342e-01
2.48236999e-01 3.98005098e-02 6.19178712e-02 2.71016099e-02
-3.38503748e-01 3.91140878e-02 -2.03009754e-01 -1.20913357e-01
-6.12372518e-01 3.30969661e-01 2.88928092e-01 8.43881488e-01
5.24420291e-02 6.92162514e-01 -5.54193556e-01 -3.92913967e-01
2.37790167e-01 6.74682617e-01 -5.07567763e-01 2.18594894e-01
4.29203629e-01 6.15174234e-01 -7.83353671e-03 4.94490415e-01
8.40607822e-01 2.67699718e-01 2.85329372e-02 -1.86909154e-01
1.36453420e-01 -8.37147906e-02 -4.86941755e-01 2.11444712e+00
-4.08836663e-01 4.60888147e-01 7.55595744e-01 -5.58143795e-01
8.10173988e-01 5.99731803e-01 9.25323367e-01 -7.22631216e-01
7.70520568e-01 9.22421813e-02 1.99889898e-01 -1.03885686e+00
-2.85859883e-01 -1.09914148e+00 2.35481426e-01 1.26067713e-01
-2.32669488e-01 -4.30280954e-01 -4.16152000e-01 -1.37960911e-01
1.32659340e+00 -2.11119682e-01 2.97242641e-01 -2.84063399e-01
9.34147090e-02 -8.83128196e-02 1.38080483e-02 5.61243832e-01
-7.11664498e-01 7.87321091e-01 6.95896447e-01 -6.07675791e-01
-1.14312232e-01 -7.15359211e-01 -5.92853546e-01 5.25464356e-01
-3.12006235e-01 -1.49177209e-01 -5.86567283e-01 -6.01105988e-01
9.46116745e-02 4.89987791e-01 -1.16809630e+00 -4.80692744e-01
-1.55734167e-01 -7.04283357e-01 3.09237510e-01 6.10515535e-01
-2.40693033e-01 -1.26964307e+00 -1.50767910e+00 2.26911038e-01
-1.16774663e-01 -9.49251175e-01 -4.76190038e-02 6.17663920e-01
-7.12424099e-01 -1.22271705e+00 -7.46734262e-01 -9.28594172e-01
5.41178405e-01 -4.07625847e-02 8.89030397e-01 -1.03249416e-01
-1.12374842e+00 9.18884397e-01 4.34596613e-02 -1.11547077e+00
-1.90750510e-01 -8.68584096e-01 2.03317162e-02 -2.79705256e-01
7.94974566e-01 -2.23745868e-01 -1.14647532e+00 -1.57081544e-01
-7.42295921e-01 -5.77726662e-01 6.08863950e-01 7.51787961e-01
4.96121168e-01 -4.38320249e-01 3.30380857e-01 -6.21961236e-01
9.86763895e-01 -4.36762929e-01 -2.31862485e-01 -5.13365984e-01
-3.21038663e-01 -4.72773373e-01 -1.14060625e-01 -3.65990222e-01
-8.56745362e-01 4.60508287e-01 -3.82833853e-02 -9.10465658e-01
-2.44652048e-01 5.82534730e-01 8.93814147e-01 -2.54677176e-01
9.35312271e-01 -7.16613770e-01 7.69867480e-01 -4.16759029e-02
-1.19329780e-01 7.34108031e-01 5.48339963e-01 -3.68812412e-01
-1.84926525e-01 7.77844608e-01 1.13890022e-01 -7.51426458e-01
-7.53765881e-01 -7.16418743e-01 -3.95411789e-01 -1.73357770e-01
1.06114292e+00 -1.34531605e+00 -9.07156169e-01 -1.05183676e-01
-9.92664814e-01 -1.56452805e-01 -2.08400965e-01 8.80723298e-01
-6.23533607e-01 2.88821757e-01 -6.62379146e-01 -1.10117602e+00
-6.44324005e-01 -1.20674074e+00 1.15095186e+00 -6.38337582e-02
-1.04552126e+00 -5.90377927e-01 3.87217462e-01 3.31248939e-01
2.07030639e-01 7.85659313e-01 7.84999549e-01 -6.25602454e-02
4.46741879e-01 -6.30882919e-01 -1.08055428e-01 4.97510284e-01
5.38238704e-01 -5.75117283e-02 -1.57120466e+00 -3.42890561e-01
7.80031383e-01 -6.50081098e-01 6.10262573e-01 1.03462851e+00
1.40605783e+00 -2.29319841e-01 -1.91669002e-01 3.33060294e-01
1.36643469e+00 5.23889482e-01 7.42279172e-01 1.44591361e-01
-8.27999860e-02 9.01246786e-01 4.22851056e-01 4.82807040e-01
-3.10854793e-01 2.28141189e-01 6.69464767e-01 -3.31423461e-01
2.51873434e-01 6.04161501e-01 1.89976692e-01 5.91942444e-02
-3.20157409e-02 6.68310225e-01 -7.48959541e-01 2.63662070e-01
-1.13267970e+00 -7.49557316e-01 1.02572544e-02 2.01472616e+00
5.84672689e-01 -1.49334893e-01 1.52961254e-01 -7.89641961e-03
4.19550151e-01 -2.99097240e-01 -2.78368086e-01 -1.28749788e+00
2.37196982e-01 1.14073193e+00 2.76399463e-01 3.69186074e-01
-1.18259811e+00 1.07118897e-01 7.15650225e+00 2.23528445e-01
-1.43427300e+00 -1.58129260e-03 9.92695093e-01 -5.81991434e-01
4.39240605e-01 -5.26348233e-01 1.87450230e-01 2.59893447e-01
7.62447298e-01 2.38882616e-01 2.30231270e-01 9.99999881e-01
5.80523312e-01 -8.87690544e-01 -1.14628935e+00 1.54822469e+00
1.89903423e-01 -1.06653106e+00 -6.16405010e-01 -5.46144545e-02
2.39587560e-01 -5.48636056e-02 3.23555231e-01 1.78291142e-01
-5.27221859e-01 -1.53965926e+00 3.43091816e-01 1.65282726e-01
1.14497888e+00 -7.02812731e-01 1.10730696e+00 -9.13648829e-02
-3.54399353e-01 -4.02512878e-01 5.51241450e-02 -3.75004858e-01
-2.37472653e-01 1.52108103e-01 -1.13750088e+00 -1.64493933e-01
1.17626178e+00 2.58996338e-01 -1.10320941e-01 1.09435618e+00
1.17709227e-01 2.24527195e-01 -3.30588102e-01 3.03222746e-01
4.27466065e-01 -2.62630656e-02 1.44979618e-02 1.31633437e+00
1.49431765e-01 3.64966989e-01 -8.03018436e-02 3.85236979e-01
1.69164285e-01 3.14176887e-01 -7.31257677e-01 2.58194268e-01
-4.08911228e-01 1.54208922e+00 -7.12211668e-01 4.62301709e-02
-8.09897363e-01 7.92337835e-01 1.36333808e-01 1.33327007e-01
-6.29732668e-01 2.15988159e-01 5.71494401e-01 2.77705163e-01
5.24860881e-02 3.17725062e-01 -5.62219143e-01 -4.22392637e-01
9.38556045e-02 -1.00513816e+00 7.25060940e-01 -9.04333591e-01
-1.13287532e+00 5.05085051e-01 -1.11424424e-01 -1.08592153e+00
2.89814949e-01 -8.51132095e-01 -5.69965601e-01 9.41291153e-01
-1.24982846e+00 -5.64951181e-01 -6.15368307e-01 6.16374135e-01
2.54326910e-01 4.63682860e-01 1.51019895e+00 -5.43605238e-02
-2.05716610e-01 1.67835250e-01 -4.32162732e-01 -3.23707191e-03
7.73483276e-01 -8.96202207e-01 -8.79040301e-01 8.67396942e-04
-2.83921331e-01 6.19210005e-01 5.06508350e-01 -4.27957326e-01
-1.18277740e+00 -5.73307216e-01 -4.67503257e-02 -3.07513714e-01
3.27834368e-01 3.30397785e-02 -7.78884172e-01 4.62359041e-01
5.46529114e-01 -6.39673769e-02 1.27154207e+00 -9.50386077e-02
-2.22572178e-01 1.78369194e-01 -1.51759708e+00 2.43634835e-01
2.53094733e-01 -4.71344560e-01 -6.72279596e-01 4.38230336e-01
-1.59909025e-01 -7.43056297e-01 -8.70510936e-01 7.14923441e-01
8.32151532e-01 -1.01176715e+00 4.84894425e-01 -5.20447314e-01
5.29856503e-01 1.91507265e-01 1.85345024e-01 -1.02288604e+00
-1.60539180e-01 -8.43623698e-01 8.03634003e-02 2.23440379e-02
1.07906111e-01 -2.95252830e-01 9.13490534e-01 9.89403844e-01
7.77628422e-02 -1.10484362e+00 -1.12976718e+00 -2.08977550e-01
8.27279910e-02 -3.89939785e-01 -4.97498780e-01 6.54452085e-01
5.72554290e-01 2.16770053e-01 5.93982376e-02 -1.00845218e-01
-1.98595271e-01 -1.62765428e-01 4.56748605e-01 -1.02078056e+00
-1.27071351e-01 -5.01349032e-01 -6.52876854e-01 4.46084179e-02
-6.21134751e-02 -6.46980762e-01 -4.44978140e-02 -1.58887303e+00
4.06023562e-01 -7.28567243e-02 -5.18825471e-01 6.84470534e-01
-1.14380475e-02 2.39686117e-01 -4.84498031e-02 -9.82556343e-02
2.56845057e-01 6.25072941e-02 1.22517109e+00 1.17536969e-01
-4.73951697e-01 -1.17303275e-01 -9.62100267e-01 7.56705761e-01
9.63544905e-01 -7.47959912e-01 -3.35238725e-01 -2.89148334e-02
-1.70357361e-01 1.99398279e-01 2.38895699e-01 -8.34968925e-01
-3.35794836e-01 3.49776596e-01 7.21724570e-01 -2.16535941e-01
7.59667099e-01 -1.07186663e+00 1.39559016e-01 1.07913530e+00
7.23683611e-02 2.31433704e-01 9.41105425e-01 5.54119110e-01
-2.79502869e-01 -1.17083535e-01 1.17184722e+00 -6.14223182e-01
-2.61241049e-01 -8.28909501e-02 -9.30972695e-01 -2.32610002e-01
1.24524307e+00 -1.92203939e-01 -2.10971348e-02 -6.43831432e-01
-1.21130621e+00 -2.81881332e-01 3.15206736e-01 2.93887556e-01
6.93832576e-01 -8.69760573e-01 -4.57257301e-01 2.81235844e-01
9.97758284e-02 -9.38667823e-03 3.77605617e-01 1.53484297e+00
-1.03851199e+00 8.15513209e-02 -3.53380352e-01 -8.87106359e-01
-1.67644882e+00 7.59849668e-01 4.31556731e-01 1.26259401e-01
-6.34421408e-01 1.09371579e+00 3.86022359e-01 3.70472580e-01
4.35791850e-01 -5.18876135e-01 -5.25858812e-03 1.88777551e-01
5.27941763e-01 -1.20562442e-01 3.82789850e-01 -5.93322158e-01
-5.42254984e-01 6.45636857e-01 -7.82100037e-02 1.58502944e-02
1.13277721e+00 3.20426524e-01 -1.46122992e-01 1.42257079e-01
1.55086565e+00 1.50212184e-01 -9.81338501e-01 5.68648934e-01
-3.06581587e-01 -3.03175837e-01 3.02441984e-01 -9.31357980e-01
-1.06740153e+00 9.18829083e-01 1.32645047e+00 -1.44834980e-01
1.36042583e+00 -1.43630961e-02 2.36830920e-01 2.62476318e-02
8.16444978e-02 -1.13785625e+00 3.42654616e-01 -2.22571373e-01
1.12934065e+00 -1.29700065e+00 1.34054914e-01 -6.24934852e-01
-1.17833316e+00 1.30633724e+00 4.19205099e-01 -1.79801330e-01
9.43668485e-01 5.91113806e-01 8.66046369e-01 -7.59674072e-01
-2.58927971e-01 7.31854811e-02 4.45480533e-02 3.97264630e-01
5.42037666e-01 1.96676910e-01 -5.20577788e-01 5.79168677e-01
-2.22981721e-01 4.50066149e-01 6.43233657e-01 1.51640332e+00
-2.44732589e-01 -4.92056727e-01 -4.67970639e-01 6.86424553e-01
-1.33264685e+00 1.10009741e-02 -4.48679805e-01 8.27265739e-01
1.16306238e-01 1.02619588e+00 -5.69140613e-02 8.98262709e-02
3.59132469e-01 4.19102728e-01 7.85967588e-01 -8.64397883e-01
-8.08869183e-01 5.52486002e-01 7.53795803e-02 -8.45627904e-01
-5.99386692e-01 -5.23383021e-01 -1.21383882e+00 4.81180638e-01
-7.76125118e-02 -2.55550863e-03 8.14483523e-01 3.54338616e-01
2.86415368e-01 3.28579932e-01 4.29317027e-01 -7.43374348e-01
-6.62305802e-02 -8.87865424e-01 -7.25935161e-01 5.07369280e-01
5.29711604e-01 -5.47173500e-01 -4.72014427e-01 -1.23173490e-01] | [13.612739562988281, 2.124228000640869] |
cf8eb524-5376-4a99-926b-38251b5e42d6 | occluded-person-re-identification-via | 2212.04712 | null | https://arxiv.org/abs/2212.04712v1 | https://arxiv.org/pdf/2212.04712v1.pdf | Occluded Person Re-Identification via Relational Adaptive Feature Correction Learning | Occluded person re-identification (Re-ID) in images captured by multiple cameras is challenging because the target person is occluded by pedestrians or objects, especially in crowded scenes. In addition to the processes performed during holistic person Re-ID, occluded person Re-ID involves the removal of obstacles and the detection of partially visible body parts. Most existing methods utilize the off-the-shelf pose or parsing networks as pseudo labels, which are prone to error. To address these issues, we propose a novel Occlusion Correction Network (OCNet) that corrects features through relational-weight learning and obtains diverse and representative features without using external networks. In addition, we present a simple concept of a center feature in order to provide an intuitive solution to pedestrian occlusion scenarios. Furthermore, we suggest the idea of Separation Loss (SL) for focusing on different parts between global features and part features. We conduct extensive experiments on five challenging benchmark datasets for occluded and holistic Re-ID tasks to demonstrate that our method achieves superior performance to state-of-the-art methods especially on occluded scene. | ['Sangyoun Lee', 'Suhwan Cho', 'Heansung Lee', 'MyeongAh Cho', 'Minjung Kim'] | 2022-12-09 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [-3.17885987e-02 -2.77709574e-01 8.36399645e-02 -4.08552319e-01
-3.39327127e-01 -8.84104967e-02 3.63890052e-01 -9.97500271e-02
-5.63796699e-01 7.00393200e-01 3.87315869e-01 4.13732886e-01
1.69680998e-01 -5.68614960e-01 -5.07560492e-01 -5.53520858e-01
3.23123366e-01 4.13134485e-01 2.91895419e-01 9.97715816e-02
-1.79123774e-01 4.78245616e-01 -1.61443186e+00 2.02300213e-02
8.27035129e-01 6.64037168e-01 -6.02422692e-02 3.33265483e-01
5.28398827e-02 3.54895592e-01 -7.18510509e-01 -8.19206834e-01
4.64548349e-01 -3.57856639e-02 -4.85794634e-01 5.42535186e-01
8.41514707e-01 -5.46118259e-01 -4.35719937e-01 1.09860945e+00
7.33638763e-01 3.60084563e-01 5.22161424e-01 -1.32682538e+00
-6.58960998e-01 7.64111802e-02 -9.08889413e-01 1.01946920e-01
6.16733372e-01 2.14921400e-01 4.71990883e-01 -1.04082370e+00
3.67177010e-01 1.64720607e+00 9.69060302e-01 7.02673435e-01
-1.04752314e+00 -6.48630023e-01 6.59190536e-01 2.99133509e-01
-1.51424265e+00 -4.68030751e-01 7.00504124e-01 -4.22745377e-01
4.86118823e-01 1.55080125e-01 6.77727342e-01 1.22077763e+00
-2.17081890e-01 8.49216580e-01 7.95632780e-01 -2.62784690e-01
-1.87947422e-01 2.25318626e-01 4.46613342e-01 6.91698015e-01
8.82415414e-01 -3.74812745e-02 -2.78230488e-01 -6.77390248e-02
6.31200254e-01 4.17184800e-01 -2.79434174e-01 -5.74570239e-01
-1.03441358e+00 4.14674163e-01 4.12349880e-01 -2.21183464e-01
-5.09477668e-02 1.01920422e-02 5.76635361e-01 -3.03434730e-01
2.74621546e-01 -2.52934426e-01 -2.37094939e-01 4.01448190e-01
-4.09336925e-01 3.53197098e-01 4.11613137e-01 9.35188532e-01
5.80056667e-01 -1.34805501e-01 -5.37838459e-01 1.07837081e+00
3.74006212e-01 5.03695786e-01 3.11103582e-01 -6.58885598e-01
7.00607717e-01 7.72583246e-01 4.78842169e-01 -9.73726869e-01
-5.79228401e-01 -7.36109018e-01 -1.05063915e+00 1.89886000e-02
7.09536791e-01 -9.70957279e-02 -8.66129994e-01 1.69842470e+00
6.83295310e-01 2.33439639e-01 -9.03370306e-02 1.20152378e+00
1.07579315e+00 1.19131915e-01 4.29524094e-01 5.22922352e-03
1.77523708e+00 -1.38301194e+00 -6.24562860e-01 -4.32911873e-01
1.53059408e-01 -5.43504655e-01 8.07814419e-01 1.67530537e-01
-7.73008764e-01 -1.01893985e+00 -8.33160281e-01 -2.16400757e-01
-3.62818778e-01 5.26336312e-01 4.20485467e-01 8.36330950e-01
-5.50087094e-01 3.27778786e-01 -4.29347485e-01 -6.82774842e-01
4.77666557e-01 3.93212795e-01 -6.96576953e-01 -2.10745811e-01
-9.56997156e-01 7.74855077e-01 1.55793831e-01 4.05275792e-01
-8.05245101e-01 -3.72265518e-01 -9.65365350e-01 4.06724028e-02
4.02923077e-01 -1.19634712e+00 6.70853853e-01 -7.93165267e-01
-1.25003684e+00 9.07073557e-01 -5.04023969e-01 -2.43900627e-01
8.97780120e-01 -6.46443129e-01 -4.25344229e-01 1.19875893e-01
2.58799493e-01 5.52538157e-01 8.27571929e-01 -1.41020811e+00
-7.61484861e-01 -6.11761987e-01 7.99285769e-02 3.59730095e-01
-4.20639783e-01 1.76956773e-01 -8.15356672e-01 -6.45533383e-01
8.23888704e-02 -8.85645628e-01 -1.88849047e-01 2.85155535e-01
-7.62017190e-01 -4.68607396e-01 7.49682844e-01 -9.50306296e-01
8.79650772e-01 -1.95226359e+00 2.47936212e-02 2.90301032e-02
3.28860790e-01 4.54036742e-01 -6.88923523e-02 -1.32833207e-02
-1.03626758e-01 -3.21567804e-02 1.46362036e-01 -9.27729547e-01
-6.80413395e-02 -7.31117576e-02 8.77712369e-02 6.59186780e-01
-4.34408225e-02 7.01079547e-01 -6.24843121e-01 -7.15803623e-01
4.43504572e-01 6.87623739e-01 -2.18993917e-01 2.27303952e-01
5.25549352e-01 6.55674219e-01 -3.39086145e-01 8.80508661e-01
1.01882064e+00 -5.25649637e-02 -1.85370833e-01 -4.47531819e-01
-2.47326065e-02 -1.71931118e-01 -1.47115493e+00 1.52223980e+00
5.33266831e-03 1.26388058e-01 -1.01997286e-01 -7.40365863e-01
7.57217884e-01 1.29003614e-01 2.84992754e-01 -3.78580064e-01
3.01366985e-01 -1.15732618e-01 -3.87237042e-01 -6.14802182e-01
3.98391783e-01 3.89435679e-01 7.55170509e-02 3.69978510e-02
-2.79926986e-01 9.22237635e-01 2.43144438e-01 -7.08687603e-02
6.04678810e-01 3.39993089e-01 3.27719122e-01 -1.27060458e-01
1.00596559e+00 -3.42116058e-01 1.05937016e+00 8.54780912e-01
-7.04156578e-01 8.34270537e-01 -1.76555663e-02 -8.25520873e-01
-8.17564368e-01 -1.06926036e+00 -1.06181398e-01 9.61690247e-01
7.81995237e-01 -3.15199703e-01 -8.91374886e-01 -7.25342214e-01
6.13222606e-02 1.87620074e-01 -6.03754163e-01 3.25948931e-02
-7.50440598e-01 -1.02822399e+00 3.94394487e-01 7.82414019e-01
9.79192793e-01 -6.22356653e-01 -2.29559973e-01 5.48699759e-02
-5.54604173e-01 -1.30351770e+00 -8.01074684e-01 -4.20975357e-01
-5.27400076e-01 -1.21765280e+00 -1.10689819e+00 -8.13507617e-01
1.02742302e+00 8.90485048e-01 7.37014890e-01 1.47675127e-01
-4.02105540e-01 4.05309826e-01 -2.19978020e-01 -2.11693332e-01
3.22780699e-01 -6.38432652e-02 4.56173122e-01 3.32389861e-01
5.69107771e-01 -3.29584777e-01 -1.03332269e+00 6.23544276e-01
-1.98650241e-01 6.48804381e-02 3.23860824e-01 6.87819779e-01
6.55444682e-01 3.56659554e-02 3.12383413e-01 -5.19949496e-01
2.57274747e-01 -1.04148023e-01 -3.92132670e-01 5.57580769e-01
-1.59074754e-01 -2.93650717e-01 5.48926115e-01 -6.38875842e-01
-1.23833847e+00 3.23705405e-01 5.12113422e-02 -2.99305230e-01
-4.18233901e-01 -3.95990223e-01 -6.33338630e-01 -2.07690448e-01
2.87469357e-01 1.03549078e-01 -3.33731741e-01 -7.51153469e-01
1.92176998e-01 5.80001175e-01 7.93113351e-01 -6.04660511e-01
9.34938073e-01 8.45513105e-01 -4.89841215e-02 -5.73104620e-01
-1.01715982e+00 -5.68060577e-01 -8.30032587e-01 -2.60674715e-01
1.01985586e+00 -1.28693080e+00 -1.01587844e+00 5.98841488e-01
-1.30699801e+00 3.39863151e-01 -1.91484019e-01 4.17665780e-01
5.61315566e-03 7.64024019e-01 -5.33061981e-01 -9.42698002e-01
-4.29662704e-01 -1.13365459e+00 1.12002671e+00 7.34296501e-01
6.19897209e-02 -4.83836442e-01 -2.85698265e-01 6.03551745e-01
4.35177423e-02 2.78714716e-01 2.78701276e-01 -4.79549468e-01
-5.86908042e-01 -4.12729293e-01 -5.99296629e-01 1.34547636e-01
1.33995131e-01 -3.10883075e-01 -1.07681358e+00 -4.78151232e-01
-3.65465075e-01 4.01586341e-03 9.76553857e-01 4.03475583e-01
1.17665482e+00 -7.18713254e-02 -7.46833742e-01 8.28374922e-01
1.16951370e+00 -7.36735314e-02 5.10571957e-01 5.25238872e-01
1.06168163e+00 8.51831675e-01 4.03646797e-01 4.72563565e-01
8.23759973e-01 8.72714341e-01 1.65364802e-01 -3.68132353e-01
-4.95688945e-01 -3.61970395e-01 1.62505791e-01 2.76903957e-01
-3.75678271e-01 -1.95741370e-01 -4.45308924e-01 4.33530986e-01
-2.00632644e+00 -9.28628922e-01 -2.83347130e-01 2.42360926e+00
2.51398951e-01 -2.73244660e-02 3.84162635e-01 -1.37299309e-02
1.42311835e+00 -1.15310915e-01 -5.34396172e-01 3.42382252e-01
-3.09467643e-01 -5.74400246e-01 5.49632490e-01 3.21827769e-01
-1.53174782e+00 8.45538437e-01 5.94846058e+00 6.79999590e-01
-5.20984590e-01 3.88934433e-01 5.80696762e-01 1.06230704e-02
2.99293280e-01 -1.62479609e-01 -1.43034959e+00 5.94243824e-01
5.11497818e-02 1.97038472e-01 2.15902194e-01 9.71580267e-01
1.30700454e-01 -1.36768639e-01 -9.68391895e-01 1.52908671e+00
3.95915151e-01 -6.13544881e-01 -3.19872126e-02 -1.06104434e-01
6.39394224e-01 -5.14237344e-01 -1.23840667e-01 1.49474397e-01
1.98401660e-01 -6.77639365e-01 6.84911191e-01 7.23890483e-01
5.58743298e-01 -6.30617321e-01 8.58172238e-01 1.29269632e-02
-1.73398983e+00 -2.66422391e-01 -7.64695644e-01 7.59227714e-03
3.90169978e-01 5.16162872e-01 -1.82258964e-01 6.36730552e-01
9.76945639e-01 6.81297362e-01 -9.34792042e-01 1.38304329e+00
-2.96602458e-01 -6.99363723e-02 -2.68210918e-01 4.64029491e-01
-4.53686893e-01 -4.35434654e-02 5.59462190e-01 1.21490002e+00
8.28648210e-02 -3.85776572e-02 5.21818995e-01 5.91420531e-01
3.56086828e-02 1.00705914e-01 -2.63731420e-01 8.36974263e-01
3.40869397e-01 1.28571475e+00 -5.76761842e-01 -4.01622415e-01
-6.22567236e-01 1.18144822e+00 4.84800339e-01 5.63264132e-01
-9.60564554e-01 -1.90595403e-01 6.99957192e-01 1.62555948e-01
2.74408311e-01 -1.42110854e-01 -8.15898776e-02 -1.53094006e+00
2.79114604e-01 -5.55834472e-01 4.92175132e-01 -5.61434269e-01
-1.63496614e+00 5.28778374e-01 4.86265868e-02 -1.39963210e+00
3.11646134e-01 -4.90852207e-01 -5.27686596e-01 8.13993394e-01
-1.51800919e+00 -1.48845196e+00 -9.62263763e-01 8.16941202e-01
5.09563863e-01 -2.03211889e-01 4.15118426e-01 6.77031457e-01
-1.21546733e+00 9.50418770e-01 -1.82866529e-01 4.43014473e-01
9.65668082e-01 -8.71119082e-01 3.42656374e-01 1.11006522e+00
-3.90988439e-01 7.22514927e-01 6.30135834e-01 -8.13321173e-01
-9.28540885e-01 -1.32989597e+00 7.82177806e-01 -4.75083262e-01
1.10281846e-02 -4.46930051e-01 -6.16120636e-01 7.07339585e-01
-2.33090654e-01 3.01535606e-01 6.42656028e-01 1.15800850e-01
-3.82343739e-01 -4.89171326e-01 -1.12578964e+00 6.47281349e-01
1.55656171e+00 -5.24275377e-02 -4.79260504e-01 4.63148981e-01
6.14091039e-01 -3.79519194e-01 -5.27879357e-01 4.52315688e-01
7.39012122e-01 -1.01400912e+00 1.51196790e+00 -4.53085124e-01
-1.83090121e-01 -7.40216076e-01 -4.99022715e-02 -8.09925854e-01
-5.87255538e-01 -2.54133373e-01 -2.63040829e-02 1.64333808e+00
-3.08933526e-01 -7.87977219e-01 7.20096648e-01 8.61716866e-01
1.09948054e-01 -3.51076782e-01 -8.45016420e-01 -8.26540112e-01
-4.55819935e-01 -1.08015463e-01 7.16559649e-01 3.91136318e-01
-5.31434655e-01 8.06966946e-02 -8.76675606e-01 5.04897475e-01
1.08011746e+00 -1.49071664e-02 1.22983253e+00 -1.43424654e+00
-4.36060783e-03 -1.24405295e-01 -5.73812485e-01 -1.16365695e+00
2.18279362e-01 -5.03789544e-01 -1.40992746e-01 -1.60219610e+00
7.29505181e-01 -4.05634820e-01 -2.43237749e-01 3.35009187e-01
-5.69659352e-01 3.88390809e-01 2.74946779e-01 4.71438855e-01
-9.58485663e-01 7.45067537e-01 1.10320175e+00 -3.59051853e-01
-1.22636929e-01 8.98067430e-02 -8.54929268e-01 9.73695517e-01
5.25880039e-01 -4.58503485e-01 -1.19740352e-01 -6.31626606e-01
-3.44333559e-01 -4.07573164e-01 8.97847056e-01 -1.30895770e+00
3.87341470e-01 2.69189849e-02 1.01522160e+00 -8.17797005e-01
6.02651715e-01 -8.16443086e-01 1.86511472e-01 3.82847577e-01
1.61244161e-02 4.59696762e-02 -1.56075880e-02 7.63439238e-01
1.14604980e-02 -2.32341513e-01 7.14701593e-01 -2.82439262e-01
-7.25873232e-01 5.22421181e-01 7.15881884e-02 -1.16942339e-01
1.09871697e+00 -4.77109641e-01 -5.31763077e-01 -1.23055503e-01
-6.72022939e-01 3.80881608e-01 4.58061457e-01 4.77852255e-01
6.20494604e-01 -1.36370850e+00 -7.40692854e-01 1.52001873e-01
1.48663998e-01 -1.05609596e-01 5.94333053e-01 7.60593414e-01
-3.50521117e-01 3.48729700e-01 -1.49442211e-01 -5.17187595e-01
-1.53497910e+00 6.88573599e-01 5.55123389e-01 -1.41358659e-01
-9.15645003e-01 7.23433375e-01 6.46625876e-01 -3.08533907e-01
4.77251768e-01 1.89854801e-01 -4.53114659e-01 -1.49724245e-01
8.47023845e-01 6.71824098e-01 -4.50363159e-01 -1.07875741e+00
-6.25595331e-01 1.06978226e+00 -1.68250352e-01 2.90610611e-01
8.06367218e-01 -6.40509188e-01 6.82728440e-02 -1.18911639e-01
8.08380425e-01 2.63411570e-02 -1.46085143e+00 -4.04725909e-01
-4.71127570e-01 -7.18819201e-01 -4.88102883e-01 -4.91891682e-01
-1.01994705e+00 7.71752298e-01 1.04875588e+00 -3.58087420e-01
9.03542817e-01 -1.94311082e-01 9.39907074e-01 3.58809829e-01
4.54688400e-01 -1.15896261e+00 4.10616212e-02 1.92247212e-01
6.54626369e-01 -1.50385880e+00 3.03813756e-01 -8.07554185e-01
-5.30680656e-01 9.15580094e-01 1.00872767e+00 -9.23161283e-02
3.59205961e-01 -2.08427265e-01 -4.64986973e-02 2.72690326e-01
-1.84401943e-04 -5.16536355e-01 2.99207270e-01 9.01023984e-01
-4.13948745e-02 2.28276625e-02 -2.57618785e-01 9.68295872e-01
6.90858811e-02 -1.42228141e-01 1.38942122e-01 5.45142531e-01
-2.97402680e-01 -1.01758695e+00 -7.86756754e-01 1.86806917e-01
-3.92250925e-01 2.67109960e-01 -1.54793397e-01 8.36521864e-01
7.45831907e-01 1.09729552e+00 -1.18902080e-01 -2.45885044e-01
3.92808765e-01 -1.66838542e-01 3.76689762e-01 -2.90409714e-01
-4.19007480e-01 -9.20656323e-03 -2.61103641e-02 -4.48292643e-01
-5.80601931e-01 -7.60662138e-01 -8.45032930e-01 -2.40752906e-01
-4.10120964e-01 -2.30014697e-01 3.54401469e-01 9.40155149e-01
2.71395087e-01 6.26267314e-01 3.23476553e-01 -1.07816470e+00
-1.98970318e-01 -9.33458030e-01 -3.69186342e-01 7.78071404e-01
1.04489483e-01 -1.10553193e+00 -2.17958897e-01 1.40103251e-01] | [14.677999496459961, 0.8942734599113464] |
6f46f64a-178e-4bea-aaa5-17086fc597f7 | a-shared-representation-for-photorealistic | 2112.05134 | null | https://arxiv.org/abs/2112.05134v1 | https://arxiv.org/pdf/2112.05134v1.pdf | A Shared Representation for Photorealistic Driving Simulators | A powerful simulator highly decreases the need for real-world tests when training and evaluating autonomous vehicles. Data-driven simulators flourished with the recent advancement of conditional Generative Adversarial Networks (cGANs), providing high-fidelity images. The main challenge is synthesizing photorealistic images while following given constraints. In this work, we propose to improve the quality of generated images by rethinking the discriminator architecture. The focus is on the class of problems where images are generated given semantic inputs, such as scene segmentation maps or human body poses. We build on successful cGAN models to propose a new semantically-aware discriminator that better guides the generator. We aim to learn a shared latent representation that encodes enough information to jointly do semantic segmentation, content reconstruction, along with a coarse-to-fine grained adversarial reasoning. The achieved improvements are generic and simple enough to be applied to any architecture of conditional image synthesis. We demonstrate the strength of our method on the scene, building, and human synthesis tasks across three different datasets. The code is available at https://github.com/vita-epfl/SemDisc. | ['Alexandre Alahi', 'Taylor Mordan', 'Siyuan Li', 'Saeed Saadatnejad'] | 2021-12-09 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 5.30985177e-01 5.29599726e-01 1.64471731e-01 -3.59870970e-01
-7.86942363e-01 -6.61842704e-01 9.50862348e-01 -5.64566016e-01
-1.18589848e-01 7.34489977e-01 -3.01806051e-02 -1.79629907e-01
3.81375223e-01 -1.08381987e+00 -1.31397724e+00 -5.93260527e-01
2.78061897e-01 6.32044256e-01 1.13126859e-01 -5.17902017e-01
-1.26747429e-01 5.70283294e-01 -1.72513986e+00 2.77518749e-01
1.00905311e+00 6.45403385e-01 2.34837145e-01 9.03585970e-01
3.38603258e-01 9.78713691e-01 -7.81278670e-01 -7.14100838e-01
4.69267517e-01 -8.05875599e-01 -5.63907027e-01 2.22922359e-02
4.54282403e-01 -4.86213326e-01 -4.61071938e-01 1.09803128e+00
3.02002579e-01 -1.00996969e-02 6.96385920e-01 -1.81802332e+00
-5.81823707e-01 4.26207870e-01 1.45095643e-02 -3.64293367e-01
2.70832211e-01 6.48561060e-01 6.47580981e-01 -6.07686102e-01
9.44277704e-01 1.41330576e+00 4.80740815e-01 1.08614528e+00
-1.25135326e+00 -6.95226014e-01 -7.78705180e-02 3.48564610e-02
-1.22981143e+00 -5.19580960e-01 7.08437085e-01 -4.82870221e-01
4.00857240e-01 2.64467120e-01 5.87824225e-01 1.80128968e+00
2.08082989e-01 8.30171764e-01 1.19780397e+00 -1.88719094e-01
3.48308921e-01 1.80790737e-01 -7.80869365e-01 7.72318125e-01
2.78532624e-01 5.49712479e-01 -2.32414082e-01 1.86864406e-01
8.90773177e-01 -1.36610657e-01 -8.81311297e-02 -5.02669334e-01
-9.85803068e-01 9.85379994e-01 5.87340832e-01 7.89531693e-02
-2.47972384e-01 5.51817954e-01 8.47823620e-02 1.52573198e-01
1.15253173e-01 6.18507147e-01 5.57401553e-02 1.20441787e-01
-9.12585855e-01 7.52322137e-01 8.82437050e-01 1.11656344e+00
7.62896240e-01 4.22632307e-01 -3.62726897e-01 3.91372353e-01
1.94111437e-01 8.04610670e-01 2.77087986e-01 -1.25997674e+00
2.76793897e-01 2.86877692e-01 9.14029628e-02 -9.77886438e-01
-3.90931666e-02 -3.51560563e-01 -7.50588238e-01 6.52955830e-01
2.69752353e-01 -3.28989327e-01 -1.27304661e+00 1.96325850e+00
2.99225956e-01 3.64542305e-01 3.00000787e-01 9.08541203e-01
7.60881007e-01 6.03424609e-01 3.14597189e-01 5.20860791e-01
1.07927668e+00 -1.08757269e+00 -5.03568113e-01 -5.00154734e-01
3.24986696e-01 -6.09068036e-01 1.00676000e+00 1.28418997e-01
-1.24901783e+00 -8.00515771e-01 -1.21789026e+00 -1.50370449e-01
-4.87154007e-01 -5.52698448e-02 5.35779953e-01 6.04559362e-01
-1.13060749e+00 6.26765549e-01 -7.75104821e-01 -1.51642010e-01
5.52974582e-01 1.72559693e-01 -2.93675900e-01 -1.39008850e-01
-1.33680356e+00 9.64997411e-01 4.31283623e-01 -6.15670858e-03
-1.63698494e+00 -6.50071979e-01 -1.02057958e+00 -2.19701260e-01
2.72424340e-01 -1.14957047e+00 1.29744577e+00 -1.29042757e+00
-1.71295142e+00 1.01203489e+00 2.64197856e-01 -7.67997026e-01
9.83492076e-01 5.79723194e-02 -1.56178027e-01 1.09594993e-01
2.26404443e-01 1.33299494e+00 1.09632838e+00 -1.73272753e+00
-3.62600386e-01 -8.89955908e-02 2.86906481e-01 9.59000066e-02
3.54723305e-01 -3.37654024e-01 -4.03869599e-01 -7.03922629e-01
-4.85712588e-01 -1.19548118e+00 -5.09313464e-01 -3.42294835e-02
-5.13961911e-01 3.05999964e-01 7.60219693e-01 -6.96381688e-01
3.32120597e-01 -1.90656257e+00 4.14000809e-01 -2.05826368e-02
1.06227100e-01 2.80553013e-01 -2.94726729e-01 4.69197154e-01
-6.87687984e-03 5.45375980e-02 -6.32395267e-01 -6.20623827e-01
3.20881724e-01 4.93357241e-01 -5.07883847e-01 3.50947827e-01
4.97506499e-01 1.38086462e+00 -9.32073593e-01 -3.33075851e-01
4.07718956e-01 6.50035679e-01 -6.59083486e-01 5.72944105e-01
-6.68812811e-01 9.61988509e-01 -3.83491606e-01 4.02658999e-01
6.22749209e-01 -2.93005221e-02 -1.26691133e-01 -9.84731689e-02
1.97368279e-01 -6.77204877e-02 -9.30655003e-01 1.97040057e+00
-6.37553453e-01 4.83460218e-01 7.15515837e-02 -8.87688398e-01
9.19739902e-01 7.97891021e-02 1.49529681e-01 -8.07294846e-01
3.48529279e-01 1.53781250e-01 -1.21450894e-01 -3.20205301e-01
4.77562398e-01 -2.91360408e-01 -4.63244647e-01 8.67940038e-02
2.42056549e-01 -8.34259689e-01 1.79836169e-01 2.20632896e-01
9.06635880e-01 5.89171588e-01 6.81987926e-02 -1.42883182e-01
4.17617738e-01 1.57559067e-01 3.63720208e-01 5.54045141e-01
2.92726681e-02 1.00126374e+00 5.05966485e-01 -1.03423774e-01
-1.40267766e+00 -1.13857889e+00 1.24975681e-01 7.85845041e-01
2.81755894e-01 -1.66995436e-01 -1.21484244e+00 -5.42028010e-01
-3.23271602e-02 1.09363270e+00 -8.84670615e-01 -4.85496044e-01
-5.58228195e-01 -2.94920594e-01 9.45055783e-01 4.32359636e-01
6.18742406e-01 -1.25521100e+00 -7.10161328e-01 4.63104583e-02
-1.79627001e-01 -1.39032459e+00 -1.85760781e-01 -1.10741496e-01
-5.52810967e-01 -9.04781222e-01 -8.03992271e-01 -6.66021883e-01
6.81413770e-01 -2.41579771e-01 1.27300715e+00 -2.85725780e-02
-2.92708099e-01 3.59154016e-01 -1.62591472e-01 -4.35922980e-01
-1.14037466e+00 -4.35051322e-02 -2.24664375e-01 9.49795246e-02
-3.14677119e-01 -6.68013036e-01 -6.63521051e-01 3.58378321e-01
-1.21757007e+00 5.35814524e-01 6.14566743e-01 7.22296894e-01
5.35427034e-01 -2.05672562e-01 3.65600348e-01 -1.07620227e+00
3.47426444e-01 -4.66259152e-01 -6.76920772e-01 7.36564724e-03
-2.30909973e-01 2.07875431e-01 9.03385043e-01 -2.92062104e-01
-1.06750154e+00 1.21191077e-01 -4.16538477e-01 -6.16597116e-01
-3.88263106e-01 -9.80017856e-02 -4.23794836e-01 -9.90250856e-02
8.97582769e-01 2.08390251e-01 -5.70794754e-02 -1.06670760e-01
7.63323367e-01 1.87964022e-01 8.91564369e-01 -7.54145920e-01
1.13046503e+00 6.38298571e-01 1.35136575e-01 -4.79668230e-01
-4.70504224e-01 2.73736268e-01 -3.44195962e-01 -1.57687500e-01
1.20105469e+00 -1.06559706e+00 -5.19524574e-01 4.64087039e-01
-1.12542987e+00 -9.55414236e-01 -7.11718202e-01 2.41224505e-02
-1.09394658e+00 -2.21453160e-02 -3.59999090e-01 -4.13541943e-01
-1.22574314e-01 -1.44040740e+00 1.29753518e+00 2.14020863e-01
-4.18530330e-02 -8.82523060e-01 1.98706985e-01 4.50503767e-01
4.21951890e-01 8.03023100e-01 3.13653708e-01 -1.89550340e-01
-9.68627155e-01 8.81014112e-03 -1.87659282e-02 4.89617467e-01
-1.90707535e-01 4.62416932e-02 -1.04321086e+00 -2.78914034e-01
-1.31449938e-01 -5.51313400e-01 6.88745916e-01 1.11293316e-01
1.19745815e+00 -3.17820221e-01 -1.78677380e-01 1.00083315e+00
1.33211219e+00 -1.13633703e-02 1.01635110e+00 9.95080620e-02
9.12444174e-01 6.39121115e-01 3.97973061e-01 1.30320027e-01
3.98654461e-01 8.16907883e-01 7.05113232e-01 -2.53367513e-01
-6.06688797e-01 -8.27978849e-01 3.40441853e-01 2.37111956e-01
3.51072988e-03 -5.58861732e-01 -8.38263929e-01 5.56989431e-01
-1.71977448e+00 -8.77439976e-01 1.03943892e-01 1.83016133e+00
7.27933943e-01 7.95181841e-02 -7.61454925e-02 -1.83366835e-01
6.20063066e-01 1.11008838e-01 -6.17072105e-01 -3.84053111e-01
-1.02745421e-01 4.66837823e-01 6.14705205e-01 6.21580422e-01
-1.03120172e+00 1.26100874e+00 5.45572090e+00 7.60934532e-01
-1.01840174e+00 1.03932329e-01 6.69381440e-01 2.50914305e-01
-6.03830040e-01 -1.67152807e-02 -6.87783897e-01 4.91308659e-01
1.09888303e+00 -8.84516016e-02 4.15965974e-01 1.02170217e+00
9.42812487e-02 1.20144367e-01 -1.13940966e+00 7.20372558e-01
1.52683303e-01 -1.50914598e+00 2.84300625e-01 -1.38868615e-01
1.15631151e+00 -3.27894837e-02 2.91019797e-01 4.24363285e-01
7.71811485e-01 -1.35151088e+00 1.03439558e+00 5.37844300e-01
1.05997109e+00 -6.73370838e-01 3.40492129e-01 2.40708455e-01
-7.53806055e-01 1.90265477e-01 -1.51524812e-01 2.18554780e-01
2.43995994e-01 9.81481597e-02 -8.96102846e-01 5.28405130e-01
4.03846651e-01 5.19555926e-01 -6.16154134e-01 6.34706676e-01
-6.88913047e-01 3.31564128e-01 -1.02750801e-01 3.11347544e-01
3.50233614e-01 -1.31440744e-01 4.98513341e-01 1.06608331e+00
2.47636512e-01 -1.06469356e-01 1.62598826e-02 1.36066175e+00
-2.24881113e-01 -2.78825432e-01 -9.97431695e-01 4.14047278e-02
1.08577430e-01 1.17392635e+00 -7.18038619e-01 -3.49900067e-01
2.16810219e-02 1.26721513e+00 1.26648635e-01 3.91559720e-01
-1.37691975e+00 -8.89033377e-02 6.75215185e-01 1.32959768e-01
3.25141013e-01 -1.55760244e-01 -3.72767240e-01 -1.11972260e+00
-1.84865892e-01 -1.19718838e+00 -1.04346246e-01 -9.01832521e-01
-9.30639148e-01 8.08430910e-01 6.69093803e-02 -1.20452142e+00
-7.01118827e-01 -3.85344207e-01 -5.38273036e-01 8.00068080e-01
-1.31420219e+00 -1.49686742e+00 -6.31752491e-01 7.69414783e-01
5.32934666e-01 -7.28374273e-02 8.00182641e-01 3.41230452e-01
-3.04597080e-01 6.58580601e-01 -2.80271530e-01 1.25801474e-01
5.20339191e-01 -1.20929599e+00 8.13651681e-01 1.10205209e+00
1.32506043e-02 1.44017056e-01 1.05799592e+00 -5.96990168e-01
-1.29498482e+00 -1.45405626e+00 3.46770614e-01 -5.96475184e-01
3.84670138e-01 -6.61596119e-01 -5.18432379e-01 7.71040499e-01
3.09877664e-01 8.14085528e-02 1.20235913e-01 -6.83180630e-01
-2.80576319e-01 8.46060552e-03 -1.37921035e+00 8.85991454e-01
1.23732042e+00 -4.14196849e-01 -2.72540301e-01 2.51194090e-01
8.97439182e-01 -7.30060995e-01 -6.34450257e-01 5.44443667e-01
2.61264741e-01 -1.03719115e+00 1.07540238e+00 -4.85086650e-01
8.41399491e-01 -3.91327679e-01 -2.25067422e-01 -1.37271023e+00
8.40764195e-02 -7.63129234e-01 1.38389483e-01 8.60661447e-01
3.95712197e-01 -5.43975830e-01 9.39461231e-01 5.21484375e-01
-3.90232712e-01 -3.97199482e-01 -5.89233935e-01 -6.86145425e-01
1.69958040e-01 -3.16739261e-01 7.65325367e-01 6.61004364e-01
-7.15745330e-01 2.19615430e-01 -5.10272920e-01 1.85622975e-01
7.28657544e-01 1.43991947e-01 1.29917228e+00 -7.13135839e-01
-6.12687111e-01 -2.76826710e-01 -5.53842545e-01 -8.97896349e-01
2.87504256e-01 -8.78636718e-01 3.05545002e-01 -1.41252458e+00
-1.96504250e-01 -4.32254016e-01 3.24020773e-01 3.62064928e-01
3.89156900e-02 6.43851340e-01 5.23038983e-01 -4.73098159e-02
-4.31410164e-01 8.23748529e-01 1.75972199e+00 -1.67109638e-01
1.22505121e-01 -1.70800462e-02 -6.50347590e-01 5.91334343e-01
1.01933146e+00 -4.67223138e-01 -6.96341991e-01 -4.20554817e-01
2.96784099e-02 1.70778424e-01 9.38462615e-01 -1.36959326e+00
3.95735539e-02 -1.26291931e-01 3.00759494e-01 -1.73583791e-01
6.13914549e-01 -6.71432257e-01 6.61617160e-01 5.85431695e-01
-2.96642363e-01 -1.96445227e-01 2.90065050e-01 4.01787400e-01
-2.65877575e-01 -2.97039766e-02 1.01168549e+00 -2.30075881e-01
-8.52164984e-01 4.24222648e-01 -3.59975025e-02 2.70175308e-01
1.24657571e+00 -5.33693247e-02 -2.48969465e-01 -5.26301861e-01
-7.13204682e-01 3.54703404e-02 7.97783315e-01 6.69496477e-01
5.63451052e-01 -1.34085917e+00 -8.98098409e-01 3.42584014e-01
1.24962457e-01 1.82609722e-01 3.74295145e-01 1.99558273e-01
-8.56290102e-01 2.77805120e-01 -6.62352324e-01 -6.24750257e-01
-7.87497461e-01 5.57876289e-01 6.23594820e-01 -1.63832009e-01
-5.78794658e-01 8.31352770e-01 6.01293862e-01 -5.37791193e-01
-3.96623127e-02 -2.15043902e-01 1.99469373e-01 -3.98362815e-01
1.29714042e-01 9.30122137e-02 -1.00249082e-01 -7.75543809e-01
-7.21800188e-03 3.28577578e-01 3.71363878e-01 -2.86743939e-01
1.09342778e+00 1.25797600e-01 2.46782884e-01 6.28698841e-02
1.16992640e+00 -2.11852312e-01 -1.75526190e+00 2.24450558e-01
-5.46701431e-01 -3.12121272e-01 -1.77028522e-01 -7.66186833e-01
-1.23149681e+00 7.57329702e-01 4.48111117e-01 -2.86290701e-02
8.85959744e-01 -2.07661046e-03 8.72829199e-01 -2.85176910e-03
5.69061279e-01 -6.20744169e-01 1.81264475e-01 1.70301318e-01
1.20817935e+00 -1.18849337e+00 -4.21362936e-01 -3.59488189e-01
-8.55779648e-01 6.73903465e-01 6.86035633e-01 -6.29209161e-01
2.85563648e-01 4.07825738e-01 -2.24530101e-02 -1.33148342e-01
-4.66196865e-01 -4.06696469e-01 2.77089179e-01 8.53821695e-01
-2.12837998e-02 3.10577929e-01 2.94254674e-03 3.04438174e-01
-5.87110817e-01 1.37672713e-03 5.44786334e-01 6.06931627e-01
-5.59795368e-03 -1.29525721e+00 -2.36271873e-01 -2.03953627e-02
-1.89049229e-01 1.59332797e-01 -3.00989360e-01 9.08892989e-01
4.21023190e-01 7.51779020e-01 -9.30633768e-02 -3.57975632e-01
3.67873669e-01 -2.04256296e-01 7.05399454e-01 -5.79164445e-01
-3.26895028e-01 -3.95666748e-01 5.29310927e-02 -8.86294127e-01
-3.19713056e-01 -5.25620341e-01 -1.13235986e+00 -4.69278425e-01
3.71854484e-01 -9.42076743e-03 7.51602650e-01 5.97754002e-01
3.81667167e-01 7.72000492e-01 4.97702211e-01 -1.13900352e+00
-3.49767476e-01 -7.07839310e-01 -1.32814556e-01 7.16919601e-01
2.42043883e-01 -6.69605017e-01 -2.41218299e-01 3.67876351e-01] | [11.409423828125, -0.29389166831970215] |
2d0703ba-5397-410b-acf0-224bd1439231 | dynamic-joint-variational-graph-autoencoders | 1910.01963 | null | https://arxiv.org/abs/1910.01963v1 | https://arxiv.org/pdf/1910.01963v1.pdf | Dynamic Joint Variational Graph Autoencoders | Learning network representations is a fundamental task for many graph applications such as link prediction, node classification, graph clustering, and graph visualization. Many real-world networks are interpreted as dynamic networks and evolve over time. Most existing graph embedding algorithms were developed for static graphs mainly and cannot capture the evolution of a large dynamic network. In this paper, we propose Dynamic joint Variational Graph Autoencoders (Dyn-VGAE) that can learn both local structures and temporal evolutionary patterns in a dynamic network. Dyn-VGAE provides a joint learning framework for computing temporal representations of all graph snapshots simultaneously. Each auto-encoder embeds a graph snapshot based on its local structure and can also learn temporal dependencies by collaborating with other autoencoders. We conduct experimental studies on dynamic real-world graph datasets and the results demonstrate the effectiveness of the proposed method. | ['Shima Khoshraftar', 'Sedigheh Mahdavi', 'Aijun An'] | 2019-10-04 | null | null | null | null | ['learning-network-representations'] | ['methodology'] | [-6.22347355e-01 3.74374725e-03 -1.23152629e-01 -4.08203080e-02
4.12388742e-01 -4.73626852e-01 4.99840140e-01 1.04073450e-01
5.48871830e-02 3.86091650e-01 2.06599936e-01 -3.19686502e-01
-3.32452685e-01 -1.16469920e+00 -4.98163700e-01 -6.88353240e-01
-6.85552537e-01 5.62054396e-01 1.90139994e-01 -1.31145731e-01
-4.08758163e-01 3.75732958e-01 -7.11843252e-01 -1.45534769e-01
4.85286683e-01 2.45062158e-01 2.55969930e-02 1.07603204e+00
-1.82443410e-01 8.44520688e-01 -5.39505899e-01 -4.92450953e-01
-1.50373712e-01 -4.26288843e-01 -4.46604222e-01 8.49959552e-02
-2.40161970e-01 -1.70533285e-01 -1.43680751e+00 1.08543515e+00
2.43979439e-01 3.86288077e-01 5.52606285e-01 -1.71545517e+00
-1.20260608e+00 8.19234788e-01 -3.98053020e-01 5.28930962e-01
-1.08158097e-01 2.13562921e-01 1.29684365e+00 -7.38377646e-02
7.54848123e-01 1.29177284e+00 6.00897074e-01 5.87411106e-01
-1.28087378e+00 -4.89684552e-01 6.79318190e-01 4.72386569e-01
-1.09123015e+00 1.73087567e-01 1.34649682e+00 -6.20324552e-01
1.03880143e+00 -4.06836607e-02 1.23485422e+00 1.27271044e+00
5.43683469e-01 7.19627440e-01 2.91234463e-01 1.11612588e-01
-5.96643835e-02 -2.72998005e-01 1.96795672e-01 1.01850080e+00
-5.14641553e-02 1.51451766e-01 -1.21470578e-01 -3.09893727e-01
9.41380858e-01 4.26666528e-01 -2.20254660e-01 -4.40612257e-01
-1.09064329e+00 1.06352937e+00 9.78147089e-01 5.82846820e-01
-4.47391450e-01 8.54417264e-01 6.54640257e-01 8.24956059e-01
5.34657180e-01 -1.52453631e-01 -2.16632232e-01 5.79331182e-02
-3.20214510e-01 -2.15861410e-01 8.64244640e-01 7.38112867e-01
7.31881678e-01 5.28114736e-01 1.37848377e-01 6.07499182e-01
4.40301210e-01 2.04352513e-01 6.06776714e-01 -5.20169377e-01
5.05348071e-02 9.49521244e-01 -4.73170817e-01 -1.65432322e+00
-3.03238809e-01 -3.62870991e-01 -1.15025771e+00 -1.60964578e-01
-1.82838038e-01 -3.81812006e-01 -7.56964028e-01 1.81765771e+00
3.91854137e-01 9.49454069e-01 6.36759773e-02 4.89519060e-01
8.44540656e-01 1.10507858e+00 -5.02181128e-02 -2.55317450e-01
6.54530346e-01 -8.59209061e-01 -9.21497226e-01 -1.39745595e-02
5.04127324e-01 7.75491521e-02 5.71530700e-01 -2.65475720e-01
-5.29282808e-01 -4.68750805e-01 -9.47996974e-01 2.05714434e-01
-4.04090881e-01 -2.56272435e-01 8.44107211e-01 -4.48092557e-02
-1.27906966e+00 7.42062449e-01 -1.38558948e+00 -5.09159088e-01
1.46226645e-01 3.08316290e-01 -3.90822113e-01 2.53076941e-01
-1.23508883e+00 1.45012647e-01 5.05778491e-01 3.74629170e-01
-9.91908431e-01 -1.87401429e-01 -9.58311379e-01 3.86765063e-01
3.56010824e-01 -6.74119771e-01 7.46859670e-01 -1.06469440e+00
-1.32805061e+00 4.15092111e-01 1.14445657e-01 -5.57978392e-01
1.57939047e-01 3.09389114e-01 -1.01284814e+00 -1.42908648e-01
-2.53016531e-01 -1.80611722e-02 8.18103969e-01 -7.92525589e-01
1.95800364e-02 -3.17918062e-01 4.66040485e-02 6.22046813e-02
-5.82907915e-01 -5.68380535e-01 -5.24007738e-01 -6.30978465e-01
1.45535553e-02 -1.12494493e+00 -3.19847494e-01 -2.00671125e-02
-3.65707546e-01 -4.55158532e-01 1.41211700e+00 -6.50985062e-01
1.57638609e+00 -2.11906409e+00 8.36572051e-01 1.71150818e-01
6.94196463e-01 2.45075226e-02 -2.10889831e-01 8.08904529e-01
-2.29017839e-01 2.09734991e-01 3.22581008e-02 7.46289343e-02
-1.16870195e-01 6.09073281e-01 -1.91739842e-01 4.74856347e-01
-1.14742920e-01 1.30234897e+00 -1.09447467e+00 -4.02222633e-01
1.84881181e-01 6.10725224e-01 -4.08565760e-01 3.57977599e-01
-3.99584264e-01 8.59420672e-02 -6.45020723e-01 1.52736232e-01
1.36601314e-01 -8.53734374e-01 8.19433987e-01 -2.10983455e-01
3.71347994e-01 -6.28778338e-01 -8.26720297e-01 1.55809593e+00
-2.57774562e-01 1.08600903e+00 -1.04355052e-01 -1.25440574e+00
8.21961999e-01 4.28861886e-01 6.74881399e-01 -3.39230269e-01
2.12047741e-01 -2.35787392e-01 3.90294582e-01 -4.79131937e-01
1.28791615e-01 7.48156086e-02 1.32598311e-01 6.65062010e-01
4.40932363e-01 3.85198295e-01 1.28796279e-01 7.61110723e-01
1.41891849e+00 -3.89479190e-01 1.53090432e-01 2.79997680e-02
3.98321658e-01 -4.12185222e-01 6.26211941e-01 1.77278265e-01
-4.01359051e-01 -1.43391922e-01 8.95346224e-01 -7.81875908e-01
-9.83504713e-01 -1.18056488e+00 5.48042953e-01 7.68180907e-01
5.25975637e-02 -6.85497165e-01 -2.53736168e-01 -8.85174990e-01
3.69296700e-01 2.00081408e-01 -1.00984836e+00 -6.26801968e-01
-5.45837224e-01 -5.94241381e-01 8.38973001e-02 5.55162370e-01
2.23150309e-02 -1.14138746e+00 3.31234559e-02 5.68244100e-01
1.53385341e-01 -8.13938320e-01 -6.14202857e-01 -2.16666743e-01
-9.16367054e-01 -1.36206234e+00 -4.38147336e-01 -9.96867359e-01
5.08243024e-01 1.73566684e-01 9.82750893e-01 5.76525092e-01
-3.50534439e-01 7.24855244e-01 -2.06576601e-01 1.71016619e-01
-5.32065868e-01 4.64294404e-02 6.97653145e-02 2.96307862e-01
-6.89808652e-02 -1.16243446e+00 -4.44089890e-01 -4.20273095e-02
-7.89266169e-01 -1.26076564e-01 3.46736103e-01 1.17910707e+00
4.62090492e-01 4.45327044e-01 4.17821944e-01 -1.15016007e+00
1.01388764e+00 -8.48292768e-01 -6.61501110e-01 5.10549545e-01
-7.13345706e-01 2.51124740e-01 6.85101092e-01 -8.28281820e-01
-5.72781146e-01 -4.07449573e-01 3.81414831e-01 -1.09595168e+00
4.58384305e-01 1.04141021e+00 6.65251762e-02 1.80400074e-01
3.68115723e-01 4.31186706e-01 2.19050646e-01 -1.50979757e-01
5.98771453e-01 1.05884112e-01 2.41238162e-01 -1.72292605e-01
9.08240676e-01 2.90665835e-01 -1.07683369e-03 -8.10065925e-01
-2.01793179e-01 -1.37637347e-01 -5.81440449e-01 -5.14632404e-01
7.36873627e-01 -6.87832355e-01 -9.11095440e-01 4.06470001e-01
-9.93733585e-01 -7.35433340e-01 -4.19538207e-02 3.60844910e-01
-2.24207968e-01 4.04634744e-01 -9.24632072e-01 -4.73761111e-01
-2.57447362e-01 -7.02997863e-01 5.63757360e-01 3.09237093e-01
2.84577757e-01 -1.95255005e+00 6.77779078e-01 -5.05862832e-01
3.72792125e-01 7.49072671e-01 1.11639678e+00 -4.24073249e-01
-5.49207807e-01 -4.22973484e-01 -6.59511015e-02 -3.97540927e-02
2.99358755e-01 6.34988725e-01 -2.46963337e-01 -7.06898808e-01
-4.69157338e-01 6.16968311e-02 7.27867365e-01 4.67310160e-01
1.11707139e+00 -6.26219273e-01 -7.77136624e-01 6.86618805e-01
1.59531975e+00 2.34274849e-01 1.18381388e-01 -2.96228975e-01
1.16447687e+00 2.91917741e-01 -1.97479740e-01 3.60593855e-01
2.72200555e-01 3.53600055e-01 5.58962524e-01 2.65394211e-01
3.95641737e-02 -4.89670873e-01 3.53172809e-01 1.66537368e+00
6.88598752e-02 -6.65098548e-01 -1.05464828e+00 7.46662855e-01
-2.27876163e+00 -1.10573328e+00 -8.77729580e-02 1.41732693e+00
1.34107798e-01 8.47520456e-02 1.42974943e-01 -2.52971888e-01
9.98366296e-01 7.81553388e-01 -9.58574176e-01 -2.63437927e-01
8.47912729e-02 -1.49838448e-01 8.60263221e-03 6.23725832e-01
-8.08925807e-01 1.03462863e+00 5.92823982e+00 1.18057027e-01
-1.27457762e+00 1.94211587e-01 4.12067652e-01 4.20612916e-02
-5.39307952e-01 2.05931515e-01 2.49913841e-01 4.95241970e-01
1.12021422e+00 -8.36684823e-01 7.70690620e-01 9.98069286e-01
-3.38235378e-01 6.66540444e-01 -9.58701253e-01 1.04640210e+00
-1.93753660e-01 -1.52291441e+00 1.52827874e-01 1.27935812e-01
7.36107349e-01 2.91759908e-01 1.00014418e-01 5.72557807e-01
9.85648453e-01 -1.00173974e+00 1.29320240e-02 7.02374995e-01
5.43238282e-01 -8.79116237e-01 4.85378146e-01 2.79560059e-01
-1.74378359e+00 -2.47682646e-01 -3.67659807e-01 6.75373971e-02
3.04278791e-01 5.42760730e-01 -6.52107179e-01 6.58000112e-01
4.92083937e-01 1.53711224e+00 -5.39681852e-01 4.86572683e-01
-1.98606193e-01 6.70144916e-01 -4.42068242e-02 -2.78392196e-01
2.96819985e-01 -6.04084969e-01 8.20088327e-01 7.67905056e-01
8.80070776e-02 1.13815375e-01 3.06559950e-01 8.98716748e-01
-2.40561768e-01 -1.21826045e-01 -9.97309744e-01 -1.02190840e+00
4.81034398e-01 1.12848055e+00 -7.22066224e-01 -1.97458848e-01
-4.75784481e-01 1.25127220e+00 7.41060495e-01 8.17126572e-01
-9.67708528e-01 -4.55251604e-01 7.93029070e-01 -2.64842063e-01
1.87895283e-01 -6.36054516e-01 7.08388090e-01 -1.48372221e+00
-1.79398060e-01 -4.49673772e-01 7.32765675e-01 -6.32287085e-01
-1.38613677e+00 8.49617064e-01 -3.12785029e-01 -9.31749046e-01
-2.96717525e-01 -4.65174019e-01 -1.16871762e+00 4.08450603e-01
-8.38365078e-01 -1.09636402e+00 -5.40331841e-01 9.91628230e-01
1.28955543e-01 -2.58658230e-01 7.11345315e-01 8.24888761e-04
-9.76922750e-01 3.85450423e-01 4.13069516e-01 6.99254930e-01
-5.54027446e-02 -1.43281031e+00 6.23843133e-01 8.10462892e-01
7.44282365e-01 4.89525825e-01 4.58238721e-01 -8.62990320e-01
-1.87391901e+00 -1.18768275e+00 2.45512292e-01 -4.49291151e-03
1.28929675e+00 -3.84346068e-01 -1.06140828e+00 1.27637279e+00
3.12700272e-01 5.90761364e-01 4.44971114e-01 4.50246423e-01
-2.15861350e-01 -1.71176463e-01 -6.54909551e-01 6.73693419e-01
1.13070512e+00 -1.05348122e+00 3.01219281e-02 5.69746137e-01
1.22101688e+00 -2.73821801e-01 -1.14817655e+00 5.56355948e-03
4.64783072e-01 -5.14109612e-01 7.14796662e-01 -1.08314466e+00
1.14559531e-01 -9.09813419e-02 1.72900066e-01 -1.67705679e+00
-7.06341445e-01 -6.00680828e-01 -8.74006093e-01 1.06293464e+00
4.04177420e-02 -6.95975959e-01 8.31907630e-01 1.46388784e-01
1.93460628e-01 -6.24871254e-01 -8.05800378e-01 -6.12889946e-01
-2.54500300e-01 7.08339810e-02 6.89924598e-01 1.38707030e+00
-1.60071895e-01 5.15711486e-01 -3.77409607e-01 2.85947382e-01
5.29679418e-01 3.22156578e-01 8.14700723e-01 -1.34623420e+00
-5.85857689e-01 -6.59379661e-01 -1.06656754e+00 -5.73806643e-01
5.82701445e-01 -1.17847908e+00 -4.92719561e-01 -1.53738010e+00
9.57471430e-02 1.78248924e-03 -4.81647462e-01 2.21695736e-01
-1.88382566e-01 -6.42365038e-01 -1.06861845e-01 1.92395609e-03
-4.50050861e-01 1.00350845e+00 1.01123393e+00 -4.52676535e-01
-2.52015799e-01 -2.65555233e-01 -1.24779038e-01 2.43801251e-01
6.16958499e-01 -3.86993408e-01 -9.58969593e-01 -5.49879789e-01
2.49267668e-01 4.72523421e-01 2.54245162e-01 -8.20493281e-01
3.74585837e-01 -2.38127440e-01 1.23198450e-01 -2.92171061e-01
-1.74349528e-02 -8.17837238e-01 7.74453044e-01 7.60767996e-01
-5.94581515e-02 5.12106895e-01 -5.78942597e-02 1.57977128e+00
-3.87599796e-01 5.02084017e-01 5.33928335e-01 -4.90587354e-02
-8.30622911e-01 1.18095601e+00 -4.21231017e-02 -8.05875957e-02
1.14375567e+00 1.55929685e-01 -1.95212975e-01 -6.56586945e-01
-1.25436664e+00 5.51428795e-01 4.08898801e-01 6.46474242e-01
8.01054657e-01 -1.66997516e+00 -5.10610700e-01 -2.93733459e-03
-3.43892612e-02 -9.96750295e-02 6.21211886e-01 5.16636670e-01
-5.38098037e-01 6.71860250e-03 -1.79942027e-01 -6.17140174e-01
-1.18709910e+00 9.80865180e-01 6.10749006e-01 -5.41402638e-01
-9.87897635e-01 8.68250370e-01 7.29445219e-02 -4.61025357e-01
7.02230260e-02 -9.09385681e-02 -3.03308219e-01 -5.93702914e-03
6.08093515e-02 1.31817281e-01 -5.75827360e-01 -5.04428744e-01
-1.95747867e-01 3.40669394e-01 -5.67365959e-02 5.24739884e-02
1.58196080e+00 1.27724409e-01 -3.61315489e-01 8.72387946e-01
1.54890454e+00 -5.26603818e-01 -1.15263724e+00 -5.26660740e-01
-5.22838049e-02 -2.10632935e-01 2.02942505e-01 3.50938477e-02
-1.77664006e+00 9.07787979e-01 4.60884690e-01 5.20208895e-01
7.79735327e-01 5.19653521e-02 7.66916871e-01 3.91175300e-01
1.09818213e-01 -7.28170455e-01 7.18685508e-01 3.72374982e-01
8.51328492e-01 -9.91235793e-01 -7.20016807e-02 8.74768794e-02
-5.64901888e-01 1.29669809e+00 7.36872017e-01 -5.81788599e-01
1.40095687e+00 -2.09197924e-01 -4.52426612e-01 -7.06732869e-01
-1.19851482e+00 5.24237081e-02 1.95999593e-01 4.49777007e-01
1.68404043e-01 4.13994879e-01 1.36798203e-01 4.37381268e-01
1.30555063e-01 -4.87357140e-01 4.83274460e-01 6.69528365e-01
-8.45745504e-02 -1.02972960e+00 5.82119942e-01 6.00984275e-01
1.34096459e-01 3.07903528e-01 -5.97168088e-01 8.50905180e-01
-5.76124191e-01 4.12018269e-01 2.88470894e-01 -9.06067550e-01
-9.19386372e-02 -2.65444051e-02 2.51402825e-01 -5.17383039e-01
-1.75139114e-01 -1.11131251e-01 -3.40530485e-01 -6.04797840e-01
-2.40273193e-01 -5.65568209e-01 -1.14149082e+00 -6.64683402e-01
-3.74765486e-01 3.32642764e-01 2.71752715e-01 4.59592193e-01
5.59798300e-01 1.04363942e+00 8.90839159e-01 -4.10315603e-01
-8.56115371e-02 -9.02446687e-01 -8.22678983e-01 4.68242615e-01
4.86656338e-01 -6.14676952e-01 -4.26879436e-01 -1.62843779e-01] | [7.207624912261963, 6.1372389793396] |
162c49f7-7e3a-470f-a168-79e21398fe2b | a-new-probabilistic-distance-metric-with | 2306.07309 | null | https://arxiv.org/abs/2306.07309v1 | https://arxiv.org/pdf/2306.07309v1.pdf | A New Probabilistic Distance Metric With Application In Gaussian Mixture Reduction | This paper presents a new distance metric to compare two continuous probability density functions. The main advantage of this metric is that, unlike other statistical measurements, it can provide an analytic, closed-form expression for a mixture of Gaussian distributions while satisfying all metric properties. These characteristics enable fast, stable, and efficient calculations, which are highly desirable in real-world signal processing applications. The application in mind is Gaussian Mixture Reduction (GMR), which is widely used in density estimation, recursive tracking, and belief propagation. To address this problem, we developed a novel algorithm dubbed the Optimization-based Greedy GMR (OGGMR), which employs our metric as a criterion to approximate a high-order Gaussian mixture with a lower order. Experimental results show that the OGGMR algorithm is significantly faster and more efficient than state-of-the-art GMR algorithms while retaining the geometric shape of the original mixture. | ['Konstantinos N. Plataniotis', 'Yuri A. Lawryshyn', 'Ahmad Sajedi'] | 2023-06-12 | null | null | null | null | ['density-estimation'] | ['methodology'] | [-3.86614740e-01 -4.75631267e-01 5.05605433e-03 -1.92334116e-01
-7.92843401e-01 -6.00319840e-02 4.86035079e-01 3.53912972e-02
-2.92943358e-01 6.72605097e-01 -1.98164493e-01 -2.29500145e-01
-3.89701933e-01 -5.57316363e-01 -9.38385725e-02 -9.86622274e-01
-3.03644836e-01 6.86475873e-01 3.39184254e-01 1.03386566e-01
1.51693687e-01 9.25108552e-01 -1.26602650e+00 -7.12078273e-01
9.65816498e-01 1.00023329e+00 3.68284613e-01 4.57181990e-01
1.73771426e-01 5.05131423e-01 -6.53534293e-01 -4.11300451e-01
-7.27313459e-02 -3.91824514e-01 -1.84278518e-01 2.17481256e-01
-2.15383455e-01 -1.60207897e-02 -5.11587739e-01 1.49557972e+00
5.14723480e-01 5.44631541e-01 1.21247900e+00 -1.41179693e+00
-5.63650608e-01 5.43577850e-01 -8.01035941e-01 1.33175269e-01
2.87914217e-01 -5.31362951e-01 5.19604146e-01 -7.66320944e-01
3.30414549e-02 1.41773486e+00 6.72184885e-01 2.73360580e-01
-1.04364729e+00 -5.58841527e-01 -8.10527336e-03 2.24643812e-01
-2.30930352e+00 -3.26605618e-01 6.49310172e-01 -3.90793085e-01
2.42855489e-01 3.94541681e-01 4.02178824e-01 2.23095328e-01
2.45487601e-01 7.74241030e-01 9.66845512e-01 -3.25829178e-01
4.25747514e-01 7.76543319e-02 9.75796767e-03 7.13096678e-01
4.00197238e-01 -1.43153071e-01 -1.03333764e-01 -5.91099501e-01
8.64804089e-01 7.21826544e-03 -5.32764494e-01 -4.73039120e-01
-1.01359761e+00 7.44120836e-01 -1.05192237e-01 5.77284873e-01
-3.92724991e-01 4.53750044e-02 -3.16946894e-01 -1.05942331e-01
5.13959050e-01 -1.18763410e-02 2.09631518e-01 -2.99131870e-01
-1.14362824e+00 1.82190090e-01 8.74236047e-01 9.29240763e-01
6.76241934e-01 3.18546891e-01 5.95754795e-02 1.04683638e+00
9.31010306e-01 1.07684076e+00 5.06262898e-01 -8.75165343e-01
-7.47579336e-02 2.13930443e-01 2.63445407e-01 -1.04915941e+00
-3.20434898e-01 -4.36675936e-01 -1.25163698e+00 2.02376723e-01
3.94709736e-01 1.01211414e-01 -6.04991734e-01 1.48190129e+00
3.97474557e-01 4.65929925e-01 -5.95546328e-02 9.37910676e-01
4.32294816e-01 7.40210116e-01 -3.55588675e-01 -4.56918418e-01
1.15265322e+00 -3.61488044e-01 -8.29815090e-01 1.87433541e-01
1.64875567e-01 -1.02829528e+00 3.93603384e-01 5.61353564e-01
-9.92818654e-01 -4.08222675e-01 -1.20332670e+00 5.80397069e-01
3.20786913e-03 6.87781023e-03 4.94845837e-01 1.11106086e+00
-1.08044660e+00 6.02975190e-01 -8.48578036e-01 -9.87631530e-02
-3.79256234e-02 1.43524215e-01 -1.83058515e-01 -1.13099657e-01
-8.11424136e-01 9.29127693e-01 3.04325283e-01 1.45707324e-01
-5.71586013e-01 -5.24065733e-01 -9.05706108e-01 6.84098452e-02
2.67158691e-02 -3.26941580e-01 1.17135310e+00 -1.05026454e-01
-1.82644570e+00 4.98486787e-01 -3.95591170e-01 -4.17938948e-01
3.71872604e-01 5.39425273e-05 -8.19687366e-01 1.11387394e-01
-1.31794423e-01 6.91328049e-02 1.25986779e+00 -1.17234075e+00
-6.77615881e-01 -3.82093191e-01 -7.27775276e-01 2.86358837e-02
4.61292677e-02 7.54293427e-02 -5.72240114e-01 -6.80321753e-01
5.62807083e-01 -7.89930105e-01 -3.86064023e-01 -3.10363591e-01
-3.61013383e-01 -1.00500613e-01 7.72433400e-01 -6.00896060e-01
1.46119785e+00 -2.37130427e+00 1.09559365e-01 6.80168986e-01
3.46905887e-01 1.39953434e-01 2.61392534e-01 1.95967078e-01
1.50136381e-01 -2.80513763e-01 -4.07213598e-01 -4.36766833e-01
1.08698234e-01 -2.08660752e-01 -1.28576294e-01 1.16719675e+00
-2.64291316e-01 3.21672976e-01 -1.00838017e+00 -5.92444956e-01
5.67193031e-01 7.18780458e-01 -2.23758340e-01 1.82105660e-01
3.12664956e-01 4.35463250e-01 -1.44367307e-01 7.19816148e-01
1.04546738e+00 4.92680259e-02 -4.07714956e-02 -3.65615875e-01
3.97275994e-03 -3.32123965e-01 -1.69762743e+00 1.08674395e+00
-1.26102552e-01 6.64022028e-01 3.12458962e-01 -8.69997263e-01
1.42648828e+00 2.87165761e-01 7.41955161e-01 -1.71185255e-01
4.10407037e-01 4.78238165e-01 3.86279039e-02 9.84529108e-02
6.12635911e-01 -3.12439024e-01 1.13778107e-01 2.97922999e-01
6.40397072e-02 -2.66352504e-01 3.99063677e-01 2.00813383e-01
7.24477708e-01 -3.83747160e-01 7.17976034e-01 -5.51268280e-01
8.39101613e-01 -6.15973830e-01 5.14202654e-01 7.48072207e-01
-3.30307335e-01 6.87213778e-01 -9.59328860e-02 2.13337302e-01
-7.01745868e-01 -1.40393031e+00 -2.47970060e-01 4.39908922e-01
5.29296219e-01 -2.05982640e-01 -6.62290812e-01 -2.77487841e-03
3.88699211e-02 8.38796258e-01 -4.31166068e-02 -1.44217014e-01
-4.54686016e-01 -1.07544553e+00 3.32940310e-01 2.46038899e-01
5.02000570e-01 -3.50939661e-01 -7.81905744e-03 4.25826937e-01
-3.13605756e-01 -1.27733982e+00 -6.60068989e-01 -1.56471565e-01
-9.13417041e-01 -9.53041673e-01 -1.22104263e+00 -5.81465542e-01
5.31444788e-01 4.72421855e-01 8.10403705e-01 -2.70255864e-01
-4.45661284e-02 6.48601055e-01 -2.60970771e-01 -1.96006864e-01
-6.94147527e-01 -3.03384930e-01 4.75674450e-01 4.72698033e-01
4.52132940e-01 -6.18821681e-01 -2.23941699e-01 6.64587677e-01
-6.42896473e-01 -4.81444538e-01 5.04151225e-01 3.23838323e-01
8.08407068e-01 6.92736089e-01 6.59422517e-01 -3.34813595e-01
9.12652314e-01 -4.75180626e-01 -8.16965520e-01 1.78452715e-01
-4.56118852e-01 3.62078995e-02 5.56997716e-01 -5.65237343e-01
-9.23769832e-01 -1.98488891e-01 -3.50655049e-01 -5.57886899e-01
-4.62759696e-02 3.15748215e-01 -4.18085843e-01 -3.57945055e-01
2.94012904e-01 6.13714039e-01 1.65459111e-01 -6.21418774e-01
4.14111197e-01 9.68240798e-01 9.41588223e-01 -5.30042231e-01
8.80642295e-01 5.87499022e-01 2.67248571e-01 -1.42867005e+00
-4.68020707e-01 -1.08105838e+00 -4.03945148e-01 -2.66518921e-01
5.12377620e-01 -6.01668358e-01 -8.12534988e-01 7.03400731e-01
-7.86115587e-01 1.73969954e-01 -7.12587833e-02 1.03692043e+00
-7.99372733e-01 8.09381425e-01 -3.30462247e-01 -1.46591866e+00
-2.55129993e-01 -1.00950348e+00 7.97018766e-01 2.91011125e-01
4.36806232e-02 -1.23448706e+00 1.74664348e-01 -2.83371001e-01
4.10277665e-01 -1.07894622e-01 6.58434093e-01 -6.46862566e-01
-2.97967106e-01 -4.04007137e-01 -1.60236567e-01 4.07052785e-01
3.04863155e-01 6.21067137e-02 -5.67931712e-01 -4.02064145e-01
4.30152535e-01 5.12869775e-01 5.24887264e-01 6.14149213e-01
9.30957615e-01 6.51500002e-02 -4.77668792e-01 6.47171736e-01
1.20889568e+00 4.72515792e-01 8.41413438e-01 8.89858156e-02
3.96105140e-01 -6.13088859e-03 6.64303482e-01 5.99757850e-01
1.70029566e-01 7.81943321e-01 1.76470265e-01 1.73000693e-01
-6.30703568e-02 5.87166660e-02 4.23511356e-01 1.12655115e+00
1.02461003e-01 -2.89506435e-01 -5.23263991e-01 2.55655289e-01
-1.90419972e+00 -1.01769817e+00 -1.03604145e-01 2.52095008e+00
4.86264974e-01 -8.64953771e-02 1.10593930e-01 3.04505944e-01
1.20283413e+00 -8.95709693e-02 -1.66051313e-01 -3.23068835e-02
-4.26312275e-02 -1.10550858e-01 5.07141769e-01 6.42096221e-01
-1.25791216e+00 5.01490116e-01 6.90775728e+00 1.39803100e+00
-8.48107815e-01 1.76099706e-02 2.47440189e-01 5.53663254e-01
5.39479554e-02 -4.65338558e-01 -1.00680339e+00 6.04465723e-01
8.02103579e-01 -5.23476720e-01 4.27515239e-01 8.18991005e-01
1.70471743e-01 -1.90008715e-01 -8.04855287e-01 1.61179876e+00
1.86107486e-01 -8.48390877e-01 -1.20699033e-01 1.14423327e-01
4.61574495e-01 -2.50990301e-01 1.62328809e-01 1.80924162e-01
2.81552643e-01 -8.17040503e-01 7.32034385e-01 6.81702375e-01
4.43319380e-01 -9.35975790e-01 7.82941341e-01 4.07303184e-01
-1.48027241e+00 2.12183401e-01 -7.41112292e-01 2.86677212e-01
7.12075889e-01 1.12447572e+00 -7.16449499e-01 6.02650225e-01
3.59501809e-01 3.59571218e-01 -2.16554880e-01 1.86323142e+00
7.45044500e-02 5.11843681e-01 -6.02461100e-01 -1.05831064e-01
1.28427579e-03 -6.94908321e-01 1.14530659e+00 1.29057860e+00
8.20020080e-01 -1.18341669e-01 1.52792826e-01 6.21793389e-01
1.27622753e-01 1.50440708e-01 -3.34505647e-01 -6.45476878e-02
9.00797725e-01 1.29438174e+00 -8.82877350e-01 -2.74742186e-01
-2.73296416e-01 6.85551167e-01 -7.09564239e-02 2.91621834e-01
-9.83133495e-01 -4.66917843e-01 6.55796766e-01 1.03697047e-01
2.68668592e-01 -5.59228480e-01 1.15009621e-01 -1.03333223e+00
-3.40904623e-01 -6.40526950e-01 2.48166516e-01 -4.53112543e-01
-1.54863930e+00 5.55384696e-01 3.00864547e-01 -1.40262055e+00
-3.86595279e-01 -5.76657116e-01 -4.88057077e-01 7.18191385e-01
-1.24322569e+00 -7.63795078e-01 -1.56438544e-01 6.73101485e-01
8.91115516e-02 -5.00217795e-01 6.80040181e-01 5.91146708e-01
-3.37603897e-01 5.21704197e-01 6.40877306e-01 -1.08570725e-01
4.77600276e-01 -1.29301488e+00 6.29261434e-02 1.05264628e+00
3.55257154e-01 6.10458016e-01 8.70636404e-01 -5.17490506e-01
-1.25347686e+00 -1.02450550e+00 3.34696293e-01 1.47028923e-01
7.76975572e-01 7.30887055e-02 -9.41334665e-01 3.51528674e-01
-2.77182430e-01 -2.59127319e-01 8.27760041e-01 -5.33120744e-02
-1.24884456e-01 -2.66766828e-02 -1.23749173e+00 5.97013831e-01
5.70678353e-01 -1.62705228e-01 -5.38547933e-01 1.62359811e-02
1.49351582e-01 -2.63021201e-01 -1.01110399e+00 4.67226774e-01
4.30782259e-01 -9.12996769e-01 1.13531935e+00 2.91846395e-01
-7.94931471e-01 -8.03662360e-01 -5.23779988e-01 -1.41151679e+00
-5.86207688e-01 -9.39340651e-01 -7.36065567e-01 1.08736444e+00
3.78659256e-02 -1.05189443e+00 4.38384235e-01 1.55414551e-01
-1.20424457e-01 -5.60838997e-01 -1.10812473e+00 -1.33716238e+00
-2.30872959e-01 -5.41153908e-01 6.51776612e-01 4.63530242e-01
3.57825607e-02 2.58484501e-02 -4.74246293e-01 4.53278542e-01
1.10594463e+00 1.12638682e-01 6.59180820e-01 -1.35891902e+00
-3.45323086e-01 -6.33104980e-01 -9.78648245e-01 -1.48701143e+00
3.96005362e-02 -6.91052258e-01 2.79256850e-01 -1.48165166e+00
7.35166743e-02 -5.13145268e-01 -3.01465273e-01 -4.09094363e-01
-1.46950871e-01 4.14353609e-01 1.31610751e-01 2.57861942e-01
-7.41720855e-01 7.08719134e-01 9.50242162e-01 -5.72326854e-02
-1.56309441e-01 5.89384973e-01 -6.72590375e-01 1.10368121e+00
7.33523786e-01 -3.79244298e-01 -2.14585960e-01 1.48044810e-01
-3.19686383e-01 -1.79477632e-01 1.55818462e-01 -1.36799109e+00
4.01628822e-01 5.01115024e-02 2.29115084e-01 -9.55330670e-01
5.78116238e-01 -9.11450028e-01 3.39861214e-01 2.49959111e-01
3.50689769e-01 -2.40102008e-01 -1.05046369e-01 5.84194362e-01
-4.09368247e-01 -6.02071285e-01 1.23845971e+00 2.79812604e-01
-7.34007537e-01 4.19065803e-01 -7.10488260e-01 -6.32774457e-02
9.90454853e-01 -1.21803097e-01 -4.90102917e-02 -8.10119689e-01
-7.74497688e-01 -1.69346213e-01 3.53119314e-01 -2.83572450e-02
7.86381185e-01 -1.56157124e+00 -7.16664433e-01 8.33630711e-02
-1.68787852e-01 -2.36764625e-01 7.86906928e-02 1.05055690e+00
-5.61895251e-01 2.86073476e-01 2.13223964e-01 -7.98094869e-01
-1.02394128e+00 5.64778626e-01 4.32332963e-01 -1.32770047e-01
-5.55708706e-01 5.71974814e-01 8.36425796e-02 -2.18857557e-01
2.27226645e-01 -9.26418751e-02 -1.49486691e-01 -3.35110128e-01
9.66292620e-01 9.78738964e-01 7.21528474e-03 -9.20875371e-01
-3.84940267e-01 5.91309607e-01 1.60887122e-01 -3.08658332e-01
7.77632415e-01 -4.69302624e-01 -2.20883012e-01 4.89580035e-01
1.18077707e+00 2.42226467e-01 -9.21968758e-01 -2.67719090e-01
-2.44132783e-02 -7.11815417e-01 1.51692018e-01 -1.09163933e-01
-8.77683222e-01 5.66657424e-01 6.03275776e-01 5.89206815e-01
9.54169333e-01 6.26445785e-02 7.10949898e-01 3.44687253e-01
6.98075414e-01 -8.91798496e-01 -2.06965655e-01 4.15939659e-01
7.03228951e-01 -7.37699926e-01 6.52903765e-02 -6.41856194e-01
-2.97678381e-01 9.42937315e-01 2.48567119e-01 -1.17577836e-01
1.11824608e+00 4.25814897e-01 -1.54011622e-01 2.11491287e-01
7.96267204e-03 -3.46160799e-01 4.79580611e-01 1.07548833e+00
2.70791143e-01 2.28244796e-01 -1.48761421e-01 5.11271775e-01
-1.98495507e-01 -3.06034982e-01 3.39394152e-01 6.35133624e-01
-8.67752969e-01 -8.26395094e-01 -8.31209898e-01 4.54144210e-01
-4.92317438e-01 9.20184776e-02 3.76241744e-01 6.46219850e-01
-3.59920532e-01 1.14309275e+00 -9.81918052e-02 -2.32331276e-01
1.02011293e-01 -2.77943879e-01 7.02254653e-01 -2.55657762e-01
4.21733379e-01 4.49701220e-01 -3.29077959e-01 -3.09602350e-01
-4.15462524e-01 -6.68945670e-01 -1.18462563e+00 -5.02078235e-01
-6.93728089e-01 5.40723443e-01 7.22183645e-01 9.40804303e-01
6.76979199e-02 3.29976439e-01 6.87011659e-01 -9.54485416e-01
-8.26670229e-01 -9.57588255e-01 -1.28960705e+00 1.96173280e-01
1.28194824e-01 -1.05346453e+00 -4.52967376e-01 -1.51050195e-01] | [7.322110652923584, 4.2317705154418945] |
c82e950b-6092-497b-b8a8-a2da496eadc1 | minimizing-bias-in-massive-multi-arm | null | null | https://bmcmedresmethodol.biomedcentral.com/articles/10.1186/s12874-021-01383-x | https://bmcmedresmethodol.biomedcentral.com/counter/pdf/10.1186/s12874-021-01383-x.pdf?pdf=button%20sticky | Minimizing bias in massive multi-arm observational studies with BCAUS: balancing covariates automatically using supervision | Observational studies are increasingly being used to provide supplementary evidence in addition to Randomized Control Trials (RCTs) because they provide a scale and diversity of participants and outcomes that would be infeasible in an RCT. Additionally, they more closely reflect the settings in which the studied interventions will be applied in the future. Well-established propensity-score-based methods exist to overcome the challenges of working with observational data to estimate causal effects. These methods also provide quality assurance diagnostics to evaluate the degree to which bias has been removed and the estimates can be trusted. In large medical datasets it is common to find the same underlying health condition being treated with a variety of distinct drugs or drug combinations. Conventional methods require a manual iterative workflow, making them scale poorly to studies with many intervention arms. In such situations, automated causal inference methods that are compatible with traditional propensity-score-based workflows are highly desirable. | ['Beau Norgeot', 'Will Stedden', 'Chinmay Belthangady'] | 2021-09-20 | null | null | null | bmc-medical-research-methodology-2021-9 | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [ 3.50288153e-01 -1.12501070e-01 -1.14593744e+00 -3.29412699e-01
-6.04024768e-01 -6.18972003e-01 6.40444458e-01 6.67525411e-01
-5.11299372e-01 8.05644333e-01 7.10272670e-01 -1.00128281e+00
-4.88171518e-01 -7.39677191e-01 -4.89649534e-01 -3.37569922e-01
-1.23803869e-01 5.76776981e-01 -1.49419785e-01 3.50980282e-01
1.93902448e-01 3.22394967e-01 -1.12576461e+00 1.95226640e-01
9.59315836e-01 2.68366486e-01 -2.03031421e-01 2.54644513e-01
-3.70169505e-02 4.56955254e-01 -2.02824518e-01 -2.54463285e-01
6.19555116e-02 -4.65391248e-01 -3.39317560e-01 -5.03688216e-01
3.50939006e-01 -5.19756258e-01 -6.82734400e-02 6.13411963e-01
5.99433005e-01 -2.39020392e-01 7.12443709e-01 -1.15849257e+00
-4.30915922e-01 5.22201538e-01 -5.18841267e-01 1.48946028e-02
8.05141628e-01 3.88112485e-01 7.17825651e-01 -5.09632170e-01
8.43310177e-01 1.30846584e+00 7.33810663e-01 1.05056599e-01
-1.49240577e+00 -9.96935129e-01 8.21350217e-02 -1.19536072e-01
-7.67103314e-01 -4.95246083e-01 7.36864284e-02 -7.35769689e-01
6.83975935e-01 3.73640209e-01 7.68525898e-01 9.58796501e-01
3.96405309e-01 8.71453360e-02 1.21939206e+00 -2.59228289e-01
4.16258901e-01 -9.55034867e-02 -2.69221753e-01 3.59164983e-01
8.09257388e-01 7.11878717e-01 -2.59138137e-01 -9.24696267e-01
6.21507406e-01 7.25583076e-01 -1.86766848e-01 -6.36777043e-01
-1.58818138e+00 1.05797637e+00 3.25843126e-01 7.25174919e-02
-7.21199572e-01 -3.04508526e-02 5.98719597e-01 -3.67729180e-02
1.09543070e-01 4.64151472e-01 -5.85908532e-01 -7.46872127e-02
-1.27903545e+00 6.42341077e-01 5.64330518e-01 6.70187831e-01
1.93759337e-01 -4.18939352e-01 -3.10518265e-01 3.40271205e-01
2.44289979e-01 7.22642183e-01 2.46248111e-01 -1.19020140e+00
4.13311988e-01 8.13619852e-01 4.59853917e-01 -6.51675463e-01
-3.89422655e-01 4.79536206e-02 -8.13047528e-01 4.42113250e-01
5.35870016e-01 -1.13058835e-01 -9.71976697e-01 1.67014837e+00
4.56008792e-01 -5.32193370e-02 -1.26489505e-01 7.31156111e-01
4.29039747e-01 5.38878329e-03 5.87970555e-01 -3.24005514e-01
1.58929098e+00 -1.65065944e-01 -8.32169116e-01 -8.35068375e-02
8.41915905e-01 -8.10355127e-01 1.05341423e+00 2.03962207e-01
-1.27429152e+00 1.53080791e-01 -7.11423874e-01 4.12916094e-01
-3.57054472e-01 -5.00032604e-01 8.05799484e-01 8.56081426e-01
-5.08532524e-01 7.55897462e-01 -8.14982951e-01 -4.77932900e-01
7.11017609e-01 2.50460804e-01 -4.34209406e-01 -3.51436049e-01
-9.71664846e-01 1.07218707e+00 9.37591195e-02 -9.94883403e-02
-7.22315669e-01 -1.60473955e+00 -6.63231909e-01 1.29498035e-01
3.65721881e-01 -1.06485534e+00 1.05396748e+00 -3.85492623e-01
-6.49903178e-01 6.13222480e-01 -3.65685880e-01 -1.83533633e-03
4.72932756e-01 1.53030112e-01 -3.27630550e-01 -1.03717290e-01
5.25293827e-01 2.18079180e-01 1.36426166e-01 -6.43522978e-01
-6.36320591e-01 -7.63432801e-01 -3.43210220e-01 9.32627544e-02
1.42892584e-01 6.84906065e-01 1.15788154e-01 -7.16492891e-01
-8.91494825e-02 -8.70656013e-01 -8.51828337e-01 2.84340996e-02
-8.97818133e-02 1.81061134e-01 4.35220420e-01 -5.01435637e-01
1.34397829e+00 -1.74262989e+00 -1.70031220e-01 2.60508098e-02
2.51854539e-01 1.43112004e-01 1.86347917e-01 5.84390402e-01
-2.70339847e-01 4.41428781e-01 -1.11362673e-01 3.73344719e-01
-2.83187509e-01 -1.20297037e-01 4.26489674e-02 6.15127981e-01
2.22697973e-01 9.17501688e-01 -1.00660551e+00 -5.52508593e-01
4.03019965e-01 9.01733637e-02 -6.86788797e-01 1.55324668e-01
1.96712345e-01 4.20328379e-01 -5.59443295e-01 5.93342483e-01
6.12929821e-01 -3.02546352e-01 7.78863251e-01 1.87180683e-01
-2.55515933e-01 6.87510729e-01 -1.25027764e+00 1.31509912e+00
-3.61023366e-01 1.21524878e-01 2.44108170e-01 -8.22696865e-01
3.89365494e-01 5.20016909e-01 6.19225204e-01 -4.44707572e-01
-9.40040573e-02 1.64776176e-01 3.33107412e-01 -5.80908716e-01
-8.58154371e-02 -6.90178335e-01 -4.74853404e-02 5.85621476e-01
-4.53409731e-01 1.05458787e-02 -1.21156937e-02 -4.43750396e-02
1.53767407e+00 -2.06130311e-01 7.54927993e-01 -4.24963146e-01
-1.58985734e-01 3.83596212e-01 1.06916189e+00 8.27143252e-01
-1.09842308e-01 5.40087283e-01 5.22738636e-01 -5.29479742e-01
-1.12635231e+00 -1.30980778e+00 -4.87592578e-01 3.05747718e-01
-5.83095849e-01 -1.50850683e-01 -1.88805148e-01 -5.43631077e-01
1.61571831e-01 6.02707803e-01 -6.31854355e-01 -1.37077898e-01
-1.20667957e-01 -1.03853798e+00 4.84611779e-01 4.53962237e-01
-1.19604625e-01 -8.28859508e-01 -8.22284579e-01 4.87443537e-01
8.70119333e-02 -5.93367696e-01 -2.83717215e-01 -1.61468107e-02
-1.19925213e+00 -1.62824464e+00 -8.13601553e-01 -2.58151740e-01
5.46207130e-01 2.17493311e-01 1.05997586e+00 -2.87618414e-02
-2.79825717e-01 -8.97266120e-02 -1.24887392e-01 -7.56031275e-01
-4.69462067e-01 -3.69241059e-01 1.89880028e-01 -6.34317935e-01
6.30938470e-01 -4.24679935e-01 -1.11874151e+00 2.74683625e-01
-7.98994184e-01 -2.47877091e-01 7.25029409e-01 8.49692047e-01
2.48176426e-01 -8.75366256e-02 7.53684521e-01 -1.16868031e+00
6.20383978e-01 -6.86239898e-01 -5.77955425e-01 4.25111860e-01
-1.25806940e+00 -3.23235542e-01 3.16331863e-01 -6.81053638e-01
-8.94638658e-01 -2.32964575e-01 4.75470811e-01 -5.25547303e-02
-3.46785367e-01 8.53332698e-01 -2.99675614e-02 3.80552560e-01
7.03404069e-01 -7.19334662e-01 3.67931783e-01 -3.22837412e-01
1.28373519e-01 7.74014235e-01 -1.62573278e-01 -3.96273255e-01
2.94702947e-01 5.18362164e-01 2.67114025e-02 3.39350365e-02
-3.02437574e-01 -3.85463387e-01 -1.31573573e-01 2.44936407e-01
7.65829206e-01 -1.04381132e+00 -9.22998726e-01 -1.34339407e-01
-6.94071770e-01 -5.12559354e-01 -2.40711704e-01 1.19930458e+00
-2.85003453e-01 -7.97491968e-02 -5.31815067e-02 -7.31640816e-01
-1.76296368e-01 -1.43554413e+00 7.01961040e-01 1.04684301e-01
-8.16875815e-01 -1.04773390e+00 4.02732044e-01 2.15779454e-01
3.32799762e-01 4.59490061e-01 1.14117730e+00 -3.98316234e-01
-1.34817734e-01 -5.04097044e-01 -2.56857902e-01 -4.66532826e-01
6.64101720e-01 2.07643256e-01 -3.82831722e-01 -2.16795370e-01
-1.38051152e-01 1.34719918e-02 2.76968241e-01 1.22859049e+00
8.44750524e-01 -5.08379400e-01 -7.41902173e-01 1.33245200e-01
1.28458524e+00 5.64570367e-01 6.03644311e-01 1.92481086e-01
4.12891448e-01 1.03756928e+00 6.37878180e-01 1.41067311e-01
4.07924443e-01 9.57709193e-01 -9.45749655e-02 -5.89673936e-01
1.95086420e-01 -3.43574703e-01 -1.38434758e-02 -1.28691703e-01
6.80984706e-02 1.47260696e-01 -1.12304521e+00 6.37891412e-01
-1.82165134e+00 -9.97032225e-01 -7.30645537e-01 2.81602573e+00
9.06658649e-01 1.18526720e-01 4.65261638e-01 -1.39018819e-01
7.46815801e-01 -4.42346185e-01 -4.18781817e-01 -7.34788120e-01
2.02610508e-01 1.56580091e-01 9.51227486e-01 1.26727730e-01
-5.03282368e-01 3.22753817e-01 8.34282017e+00 3.40960175e-01
-9.37882662e-01 1.36086226e-01 4.19199288e-01 -2.85840511e-01
-5.00485420e-01 5.95482051e-01 -2.19327077e-01 4.70742732e-01
1.19682705e+00 -1.77303612e-01 1.63684636e-01 2.27174833e-01
1.15260220e+00 -4.53943580e-01 -1.13357317e+00 3.15309823e-01
-6.87317252e-01 -1.47825682e+00 -3.31220895e-01 4.68259871e-01
6.60972297e-01 -9.25210863e-02 -7.26966932e-02 -2.45781004e-01
9.48034585e-01 -1.26537633e+00 3.46962780e-01 4.26824361e-01
1.16855788e+00 -1.68496788e-01 8.26755524e-01 -6.67012632e-02
-5.41429281e-01 -7.32601210e-02 -1.60692111e-01 -2.93306708e-01
6.29091442e-01 1.08569312e+00 -1.15991950e+00 5.91302931e-01
9.26163316e-01 4.14492518e-01 -2.13905051e-01 1.23688304e+00
-1.92995727e-01 8.14433575e-01 -8.23402107e-02 2.30793685e-01
-6.78566694e-02 -2.77427733e-01 2.63559103e-01 6.91500843e-01
4.07428294e-01 -8.52744505e-02 -2.80269563e-01 7.74679244e-01
2.38546833e-01 6.58823773e-02 -1.06855989e+00 -2.81963110e-01
5.65498590e-01 9.17318940e-01 -6.37925982e-01 -2.42306814e-01
-9.14539754e-01 1.07880503e-01 -1.19849645e-01 1.74225554e-01
-3.26253116e-01 9.96637866e-02 6.99797094e-01 5.00497341e-01
-8.92931819e-02 1.38826966e-01 -4.59113955e-01 -7.58737624e-01
-1.82614401e-01 -1.37955165e+00 8.59202921e-01 -5.93114793e-01
-1.21531510e+00 -3.22952867e-01 2.31601372e-01 -1.07086253e+00
-1.49061069e-01 -1.12853400e-01 -4.34455961e-01 1.34196591e+00
-9.54358459e-01 -7.90926337e-01 2.51331061e-01 1.95423841e-01
-6.49765208e-02 1.59490034e-01 1.06931198e+00 3.32983464e-01
-6.65626287e-01 2.13733912e-01 1.89026505e-01 -2.90806472e-01
1.05559134e+00 -1.10467112e+00 1.85872409e-02 4.75689203e-01
-5.01994193e-01 1.09129250e+00 7.11101234e-01 -1.30945182e+00
-1.05027354e+00 -6.21368766e-01 9.08326745e-01 -4.93211836e-01
8.00688922e-01 -1.23508386e-01 -8.05431902e-01 5.92449903e-01
-5.43311313e-02 -4.17825997e-01 8.98974657e-01 8.82744908e-01
-3.65273505e-01 -2.61503104e-02 -1.25317609e+00 8.10947061e-01
8.37212026e-01 -2.66713560e-01 -6.72524691e-01 2.17865080e-01
5.85460305e-01 -3.74131352e-02 -1.25376189e+00 5.32281816e-01
8.16044092e-01 -6.78737223e-01 1.00196445e+00 -1.21135187e+00
4.83076870e-01 -3.03642094e-01 2.90076792e-01 -1.15455866e+00
-2.52314091e-01 -3.43270719e-01 4.94773746e-01 9.85708833e-01
6.13362134e-01 -6.05134785e-01 7.57830977e-01 1.23467612e+00
3.34239960e-01 -3.23309392e-01 -8.35173547e-01 -5.86654902e-01
2.82505542e-01 -1.02738038e-01 9.72508550e-01 1.32719064e+00
2.06405848e-01 -1.70538351e-02 3.89934219e-02 4.87561077e-02
5.87028921e-01 1.62614062e-01 5.39988101e-01 -1.33062577e+00
6.53946623e-02 -5.41210413e-01 -2.14011624e-01 1.99162290e-01
-3.19652289e-01 -5.93271136e-01 -3.22061867e-01 -1.73471987e+00
7.63059974e-01 -9.77402568e-01 -2.01973543e-01 5.25887847e-01
-6.53260946e-01 -4.26581830e-01 -2.82611489e-01 1.12545900e-01
3.26186977e-02 8.96525010e-02 9.35579121e-01 6.42741621e-02
-5.23147166e-01 2.66472667e-01 -9.84693825e-01 5.19327462e-01
8.10598671e-01 -1.08576560e+00 -3.92855465e-01 -7.03719482e-02
3.47655594e-01 3.33734453e-01 5.32134593e-01 -3.34833235e-01
-2.76449583e-02 -8.22982371e-01 1.81487828e-01 -1.84691697e-01
-4.74292815e-01 -9.66570079e-01 9.89149213e-01 9.37424839e-01
-5.36216915e-01 3.57204705e-01 1.58837631e-01 5.14197290e-01
1.74273700e-02 -3.00009042e-01 3.16409081e-01 -8.80789533e-02
1.76473752e-01 1.17689639e-01 -4.63271350e-01 2.86660735e-02
8.84275794e-01 -6.51736110e-02 -3.72366816e-01 -2.94142663e-01
-3.93233657e-01 2.29765207e-01 8.32968831e-01 3.23709249e-01
1.83848038e-01 -1.32700598e+00 -8.13100159e-01 -4.30887341e-01
3.33516449e-01 -1.10661902e-01 3.61382306e-01 1.27917230e+00
-3.59202594e-01 3.65700543e-01 -1.72097623e-01 -1.27500534e-01
-1.11727142e+00 1.04036129e+00 6.01457618e-02 -4.74748939e-01
-7.05309153e-01 -1.65237308e-01 3.84184659e-01 -5.66504121e-01
-1.01847053e-01 -4.10972424e-02 1.14179313e-01 1.16637863e-01
7.56429195e-01 5.74790657e-01 1.21591955e-01 -6.06906302e-02
-6.69589043e-01 -1.94850285e-02 1.52156428e-01 -3.29414964e-01
1.57927310e+00 -1.29485026e-01 -2.40344867e-01 3.95091683e-01
8.07027519e-01 1.65180981e-01 -9.71247792e-01 1.64626777e-01
8.66748765e-02 -4.91121620e-01 1.97362736e-01 -9.44366813e-01
-6.30216777e-01 5.18745363e-01 6.27394438e-01 1.72071487e-01
9.61731732e-01 -2.10280776e-01 3.07385419e-02 -5.35976768e-01
2.14906201e-01 -8.42153609e-01 -7.56750941e-01 -6.27432048e-01
6.20544493e-01 -1.32585943e+00 4.83370602e-01 -2.21553713e-01
-4.08366501e-01 4.85114396e-01 -6.71673939e-02 2.76113719e-01
5.28248608e-01 2.42030263e-01 -1.09471254e-01 -4.54880625e-01
-7.46044636e-01 4.24904712e-02 1.34333938e-01 8.60096395e-01
6.43823683e-01 4.76357400e-01 -1.01759505e+00 4.41131681e-01
1.64683029e-01 5.05066276e-01 5.87733150e-01 9.93362665e-01
4.42376524e-01 -1.55641413e+00 -6.40736401e-01 1.04052770e+00
-6.95300698e-01 -2.15518668e-01 -2.74005532e-01 8.72953713e-01
-1.47594929e-01 1.24466014e+00 1.03008542e-02 2.87414998e-01
6.28019154e-01 -1.21082865e-01 1.71387896e-01 -6.31349206e-01
-6.70951784e-01 1.40133291e-01 3.60585153e-01 -7.17795610e-01
-6.86529934e-01 -1.18335664e+00 -9.79211152e-01 -6.26112759e-01
-2.95307428e-01 2.42007926e-01 6.43112123e-01 7.21862376e-01
5.79846025e-01 4.57724571e-01 4.21087652e-01 -2.30733365e-01
-3.11704189e-01 -5.61162829e-01 -3.59883547e-01 3.30093801e-01
1.18266411e-01 -8.38670969e-01 -3.12766016e-01 -3.01010370e-01] | [8.000696182250977, 5.493827819824219] |
2c6b24a3-d3f4-4f67-bdb5-145b9f2f8e62 | shrec-2022-protein-ligand-binding-site | 2206.06035 | null | https://arxiv.org/abs/2206.06035v4 | https://arxiv.org/pdf/2206.06035v4.pdf | SHREC 2022: Protein-ligand binding site recognition | This paper presents the methods that have participated in the SHREC 2022 contest on protein-ligand binding site recognition. The prediction of protein-ligand binding regions is an active research domain in computational biophysics and structural biology and plays a relevant role for molecular docking and drug design. The goal of the contest is to assess the effectiveness of computational methods in recognizing ligand binding sites in a protein based on its geometrical structure. Performances of the segmentation algorithms are analyzed according to two evaluation scores describing the capacity of a putative pocket to contact a ligand and to pinpoint the correct binding region. Despite some methods perform remarkably, we show that simple non-machine-learning approaches remain very competitive against data-driven algorithms. In general, the task of pocket detection remains a challenging learning problem which suffers of intrinsic difficulties due to the lack of negative examples (data imbalance problem). | ['Petros Daras', 'Stelios Mylonas', 'Apostolos Axenopoulos', 'Daisuke Kihara', 'Charles Christoffer', 'Xiao Wang', 'Yuanyuan Zhang', 'Yi Fang', 'Boulbaba Ben Amor', 'Hao Huang', 'Walter Rocchia', 'Silvia Biasotti', 'Ulderico Fugacci', 'Andrea Raffo', 'Luca Gagliardi'] | 2022-06-13 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 4.87400144e-01 4.87215482e-02 -4.83322024e-01 -3.84928226e-01
-9.86662924e-01 -5.51451862e-01 2.80264407e-01 3.46799582e-01
-4.82107371e-01 1.30341220e+00 -1.06643014e-01 -5.48061311e-01
-1.81376010e-01 -2.58545041e-01 -6.60902619e-01 -1.17234886e+00
1.54986950e-02 9.44587350e-01 4.71373260e-01 -1.76755235e-01
6.42652392e-01 9.14234757e-01 -1.17443836e+00 6.49810255e-01
6.83802962e-01 5.05008459e-01 1.86705485e-01 4.21789646e-01
-5.26058301e-02 2.69753605e-01 -2.47165963e-01 -1.19268335e-02
-1.17177077e-01 -5.48924088e-01 -1.06062961e+00 -4.19683665e-01
1.70876384e-01 3.98301721e-01 4.14361507e-01 6.99821532e-01
8.33225608e-01 -2.08071157e-01 1.13620555e+00 -5.23454845e-01
4.28330302e-01 -7.53675029e-02 -2.45652556e-01 2.72378266e-01
8.12174261e-01 -2.08859041e-01 1.12046075e+00 -1.35032320e+00
9.56905782e-01 7.86149144e-01 4.78189290e-01 7.06695437e-01
-1.58298922e+00 -4.04746741e-01 -2.78853685e-01 3.20692331e-01
-1.32758379e+00 -3.13567132e-01 4.05197173e-01 -8.67413104e-01
1.12572408e+00 6.36502922e-01 4.84641135e-01 8.06310594e-01
3.15102369e-01 4.03106779e-01 1.02768862e+00 -5.31593084e-01
6.56702518e-01 -3.82729806e-02 1.13663629e-01 5.36243379e-01
3.68237309e-02 1.92439854e-02 -7.32829452e-01 -7.83399522e-01
3.82488877e-01 -2.51644820e-01 -3.72603089e-01 -7.99516797e-01
-6.26051605e-01 9.50289190e-01 2.71974653e-01 4.77181405e-01
-1.83356971e-01 -3.52336407e-01 1.89867869e-01 -1.26019463e-01
1.67151541e-01 5.13545811e-01 -8.15566480e-01 -6.67762337e-03
-8.68820190e-01 5.76951921e-01 7.19038844e-01 6.36668727e-02
4.90266770e-01 -7.09045649e-01 5.21625485e-03 5.94138324e-01
4.41791266e-01 -1.55898809e-01 2.72030950e-01 -2.10726887e-01
6.55046403e-02 8.75452280e-01 2.75003552e-01 -6.68137789e-01
-8.39379132e-01 9.32571143e-02 -4.89972502e-01 4.34194595e-01
8.27162981e-01 1.41488865e-01 -6.70820296e-01 1.32254446e+00
3.08203638e-01 -3.84879351e-01 -1.35452300e-02 9.29428101e-01
8.13070416e-01 5.27748048e-01 4.24206942e-01 -5.22220612e-01
1.30091083e+00 -4.69436020e-01 -2.43491530e-01 7.81894252e-02
7.17255712e-01 -8.23412061e-01 7.09970415e-01 4.40373540e-01
-7.94859767e-01 -1.16639711e-01 -9.80540812e-01 2.40457132e-01
-4.00870562e-01 9.49736759e-02 5.82834482e-01 6.99516356e-01
-3.66479397e-01 6.54254556e-01 -5.12946546e-01 -2.55568832e-01
4.75623816e-01 9.42358673e-01 -6.97783470e-01 2.43224651e-01
-1.01637280e+00 1.04023850e+00 5.50104976e-01 -1.78545639e-01
-4.84503061e-01 -7.22031415e-01 -2.80589759e-01 -1.39820695e-01
2.03860268e-01 -9.87864286e-02 8.38797390e-01 -8.09475720e-01
-1.14128530e+00 1.34482920e+00 -3.89764577e-01 -5.13763726e-01
6.45507753e-01 1.80507034e-01 -9.12347808e-03 -7.95331001e-02
-6.89778253e-02 4.51306552e-01 1.89230904e-01 -8.89973938e-01
-5.56538761e-01 -7.69407570e-01 -2.95038313e-01 3.06355804e-01
4.19974893e-01 2.37325355e-01 -1.26385108e-01 -3.48161668e-01
3.87704581e-01 -7.73700118e-01 -6.47812426e-01 -2.99944758e-01
-2.93458074e-01 -4.11771119e-01 3.85685533e-01 -4.44758713e-01
9.05525565e-01 -1.76077199e+00 3.24964911e-01 6.54449642e-01
2.35900149e-01 3.15879852e-01 3.85525316e-01 7.43246019e-01
-4.38617170e-01 4.62568179e-02 -1.53538436e-01 7.35232711e-01
-3.66129547e-01 -1.40675485e-01 -1.77049488e-01 8.02643359e-01
-1.49067864e-01 6.57488883e-01 -5.21119654e-01 -3.70311081e-01
2.04625905e-01 3.49249721e-01 -4.95152891e-01 9.30241868e-02
-5.18960595e-01 7.48505890e-01 -8.42210114e-01 6.94628179e-01
6.58068180e-01 -2.57911474e-01 9.26372468e-01 -2.00507209e-01
-1.26927048e-01 4.67647105e-01 -9.35494959e-01 1.32089770e+00
3.84249300e-01 2.75341243e-01 -2.41661251e-01 -9.29727495e-01
9.08019125e-01 4.67172921e-01 7.55604565e-01 -6.66111231e-01
-7.91427270e-02 5.83702147e-01 2.85526544e-01 -2.56203413e-01
-4.82215323e-02 -3.48714679e-01 2.82777488e-01 8.00988600e-02
-1.12558916e-01 2.90850133e-01 6.99371547e-02 -1.72575533e-01
8.23003650e-01 4.08427387e-01 5.93657136e-01 -5.21545649e-01
8.15909445e-01 3.82886410e-01 6.61264122e-01 1.86247200e-01
-1.93141162e-01 4.39479947e-01 7.12013900e-01 -8.25617671e-01
-9.70988512e-01 -5.86344719e-01 -5.28306901e-01 1.32302034e+00
8.45515877e-02 -3.81369174e-01 -1.09163165e+00 -6.25537872e-01
-6.99724406e-02 -3.44031118e-02 -6.76266432e-01 -3.43743078e-02
-4.80976373e-01 -1.59531355e+00 2.20335022e-01 1.32559568e-01
-1.67690128e-01 -1.14046741e+00 -6.02046728e-01 4.83445287e-01
-9.19559679e-04 -4.26945895e-01 2.66348198e-02 8.36031735e-01
-7.54931450e-01 -1.75748920e+00 -8.89682531e-01 -9.26576674e-01
5.31456172e-01 -3.22481513e-01 9.56894815e-01 -8.90150815e-02
-6.45600259e-01 -3.81778449e-01 -6.98455349e-02 -4.24820215e-01
-2.95918763e-01 1.75498113e-01 -5.26771545e-02 -2.36860007e-01
8.75824511e-01 -3.56137753e-01 -8.03130865e-01 7.89541543e-01
-4.86940682e-01 -1.66376397e-01 3.54285806e-01 8.23683441e-01
1.05557072e+00 -4.12731260e-01 3.74390543e-01 -1.21364355e+00
5.57840347e-01 -1.07569816e-02 -7.43006527e-01 2.22945705e-01
-5.54831862e-01 5.60216680e-02 1.52827457e-01 -6.70325756e-03
-7.50788331e-01 9.41863298e-01 -4.06067103e-01 6.18059278e-01
-2.01463610e-01 3.28022718e-01 -5.75603545e-01 -6.35209382e-01
9.00754333e-01 8.97090361e-02 -1.14873745e-01 -6.97984397e-01
-2.45404437e-01 3.69336247e-01 3.22348140e-02 -5.39185882e-01
2.41238922e-01 2.51060873e-01 2.73606956e-01 -7.28701770e-01
-5.62891304e-01 -8.32812846e-01 -8.47690046e-01 9.76391062e-02
7.37912774e-01 -4.92875397e-01 -1.11816370e+00 1.37795642e-01
-1.06128287e+00 -9.76570472e-02 2.19190493e-01 2.97897041e-01
-6.76054716e-01 5.05518913e-01 -2.47186393e-01 -5.82047582e-01
-2.52361715e-01 -1.38863623e+00 8.74331653e-01 4.61733080e-02
-2.91210204e-01 -6.93490028e-01 6.43301070e-01 5.98207235e-01
-1.81552738e-01 3.66226494e-01 1.19560850e+00 -1.12120986e+00
-3.96639585e-01 -2.92326003e-01 1.10152267e-01 -2.33816236e-01
-1.66846648e-01 -2.46444549e-02 -1.04585147e+00 -1.62007958e-01
-2.50252575e-01 -3.29177141e-01 8.89264584e-01 5.03271401e-01
9.45356131e-01 -5.68209700e-02 -8.32767487e-01 2.14324325e-01
1.51809251e+00 5.30892909e-01 8.60950887e-01 4.34965789e-01
2.68854737e-01 9.02460754e-01 8.60764980e-01 2.39236385e-01
-5.58472097e-01 1.16120911e+00 5.77568293e-01 -3.75468999e-01
4.67350543e-01 1.07254229e-01 -1.09585188e-01 -3.62275839e-01
-5.36398530e-01 -5.68167046e-02 -1.19749594e+00 -1.44909220e-02
-1.80615985e+00 -8.16343963e-01 -5.77745497e-01 2.32134628e+00
1.05019772e+00 1.61417857e-01 6.28683746e-01 2.74875373e-01
5.05106330e-01 -4.34179336e-01 -4.71673936e-01 -3.70633692e-01
-3.48838121e-01 4.24450547e-01 5.20480692e-01 6.02422059e-01
-1.10656440e+00 8.33312988e-01 7.30819654e+00 1.18990004e+00
-9.96153116e-01 -3.26806754e-01 8.52921665e-01 3.75291824e-01
2.23185092e-01 -5.55432662e-02 -9.40963507e-01 3.59031975e-01
7.77208567e-01 2.93337256e-01 -9.18848291e-02 7.44766593e-01
2.97219723e-01 -4.47837949e-01 -1.14280581e+00 6.16496623e-01
-3.92368942e-01 -1.68604350e+00 1.28257364e-01 4.81120497e-01
3.75082552e-01 -1.35762542e-01 -6.97342828e-02 -3.94065499e-01
-3.97914797e-01 -1.57493734e+00 1.55965820e-01 4.52619821e-01
7.38583446e-01 -7.64149129e-01 7.04128683e-01 3.01971883e-01
-8.62089634e-01 3.12195927e-01 -4.81699198e-01 4.28559035e-02
-2.00136900e-01 3.24363261e-01 -1.12255287e+00 1.61628112e-01
3.33477527e-01 1.55108869e-01 -2.74295717e-01 1.21305120e+00
2.01780781e-01 3.96174312e-01 -4.87151705e-02 -1.10720299e-01
1.10300355e-01 -2.61131644e-01 2.97568798e-01 1.09341097e+00
-5.27393103e-01 2.87835658e-01 1.67940900e-01 6.24858856e-01
2.10306391e-01 8.13803017e-01 -1.67509228e-01 3.32218371e-02
-6.23238683e-02 1.00863147e+00 -8.48064482e-01 1.32706046e-01
-2.22245991e-01 8.13930094e-01 1.27684742e-01 1.36347234e-01
-5.10988533e-01 -9.61142704e-02 6.11752570e-01 5.39224446e-01
5.65042272e-02 2.61414409e-01 1.27276145e-02 -5.68919599e-01
-7.15855956e-02 -8.48033011e-01 5.09469867e-01 -3.12564909e-01
-9.29196537e-01 3.08321387e-01 -5.41898251e-01 -8.07247400e-01
-4.54424992e-02 -1.09968925e+00 -2.44566455e-01 1.09540749e+00
-1.12536299e+00 -8.57558608e-01 1.72483593e-01 2.90012151e-01
2.72751153e-01 -2.91436583e-01 1.17736387e+00 1.87836796e-01
-3.39503735e-01 3.98157507e-01 5.17317235e-01 -1.45911157e-01
7.25939333e-01 -1.41440737e+00 1.80033445e-02 -1.70605958e-01
-1.32366987e-02 3.64470780e-01 9.14107025e-01 -6.78557098e-01
-1.03837276e+00 -5.98014772e-01 1.09738231e+00 -5.87618649e-01
1.91082358e-01 -2.81871319e-01 -9.07426357e-01 8.80448669e-02
-4.31557000e-01 -6.62628263e-02 1.25812650e+00 -1.09548308e-02
-9.87683684e-02 4.02980775e-01 -1.25972974e+00 1.52082473e-01
5.78209996e-01 -3.71921480e-01 -3.89527529e-01 8.04980993e-01
-2.18299776e-01 -4.67111468e-01 -6.68869913e-01 5.51613986e-01
6.30542159e-01 -1.30003905e+00 1.12442410e+00 -1.14924300e+00
-7.25696683e-02 -2.02228636e-01 -1.08204335e-01 -4.06092227e-01
-2.20521167e-01 -4.15811419e-01 4.87652123e-01 4.00122613e-01
8.52057278e-01 -2.24503458e-01 1.63408089e+00 3.80603492e-01
4.82691020e-01 -1.13375962e+00 -1.20539570e+00 -3.73411983e-01
2.25226671e-01 2.58654375e-02 8.76166578e-03 7.45220721e-01
3.23510021e-01 6.25741959e-01 -1.01568080e-01 -2.68186033e-01
2.33021498e-01 2.57283330e-01 6.07762337e-01 -1.70625556e+00
-1.86477810e-01 -5.07648110e-01 -5.60496509e-01 -6.38739765e-01
-9.72971395e-02 -7.89691865e-01 -1.76639616e-01 -1.35139537e+00
5.71958601e-01 -2.39528522e-01 -2.45856225e-01 3.99651200e-01
2.09306225e-01 3.63522708e-01 -4.37260568e-01 4.13828164e-01
-4.53469753e-01 -7.59100318e-02 7.41894722e-01 -8.45541656e-02
-3.68441641e-01 4.59738165e-01 -1.89531282e-01 7.90281594e-01
6.43926501e-01 -5.52427769e-01 1.33329347e-01 8.20215166e-01
4.13465112e-01 -1.13892399e-01 2.23165706e-01 -5.22250533e-01
-2.11271688e-01 -3.74684721e-01 5.34713209e-01 -5.53861797e-01
2.83122867e-01 -8.46846104e-01 2.74354935e-01 9.44378078e-01
-1.52162775e-01 -4.14435089e-01 3.80600765e-02 6.19248867e-01
-1.21071994e-01 -3.46637934e-01 9.42706108e-01 -1.16328508e-01
-4.74722743e-01 9.88772511e-02 -6.38231874e-01 -1.10742904e-01
1.13857603e+00 -7.82520473e-01 2.71055192e-01 2.42697254e-01
-1.27033424e+00 -3.83208513e-01 6.60559297e-01 -1.79027081e-01
3.08647454e-01 -6.95204556e-01 -4.87137914e-01 -4.18852530e-02
3.81669879e-01 -3.66388112e-01 -1.10295162e-01 7.24174917e-01
-9.52331305e-01 9.16939318e-01 -2.25654021e-01 -5.75563014e-01
-1.87029564e+00 3.11212838e-01 1.04924786e+00 -3.85028303e-01
-1.87259898e-01 9.50374663e-01 3.89897019e-01 -2.50394464e-01
4.46511209e-01 1.78214058e-01 -4.72140998e-01 1.23801410e-01
4.65102851e-01 3.55439663e-01 6.13358498e-01 -7.19702959e-01
-8.01371813e-01 6.08212471e-01 -9.59845483e-02 3.99839520e-01
1.27156913e+00 5.49377739e-01 -2.14694172e-01 1.34563977e-02
9.53760624e-01 -5.36497757e-02 -8.61768961e-01 -5.52613242e-03
6.43937767e-01 -1.53013244e-01 -2.69666225e-01 -1.21698141e+00
-3.69117260e-01 6.90906167e-01 1.01029670e+00 -3.66801441e-01
5.83889604e-01 -1.35320285e-02 2.33504847e-01 4.37128454e-01
3.38622540e-01 -1.01394975e+00 -1.32248774e-01 2.01196164e-01
6.80692494e-01 -1.19898939e+00 8.58026743e-02 -7.10481703e-01
-3.77034575e-01 1.25409472e+00 4.11392778e-01 2.76191175e-01
4.06154335e-01 -4.19497080e-02 -1.11373283e-01 -4.59790617e-01
-4.59982485e-01 9.96148866e-03 6.84856355e-01 5.45131862e-01
8.81089687e-01 3.46280225e-02 -1.12206817e+00 6.50034666e-01
3.60305130e-01 -7.23341331e-02 -1.20033966e-02 8.55598629e-01
-8.24509978e-01 -1.68161964e+00 -3.64565313e-01 1.74570456e-01
-8.63415182e-01 -5.79827540e-02 -1.25187206e+00 6.96419656e-01
3.07855662e-02 5.28755724e-01 -4.97893810e-01 5.36345877e-02
2.27671027e-01 1.26521423e-01 6.39715254e-01 -5.55348873e-01
-6.41673625e-01 4.48811918e-01 1.23991137e-02 -4.21146572e-01
-5.25125027e-01 -4.38601464e-01 -1.64235115e+00 1.46651804e-01
-5.60902059e-01 8.90303493e-01 6.85827911e-01 8.97785723e-01
3.01647395e-01 -5.58834299e-02 3.02895159e-01 -7.02701330e-01
-2.69003391e-01 -5.14037907e-01 -8.01388741e-01 5.64971529e-02
1.47200432e-02 -7.27339685e-01 4.49200235e-02 -8.24687071e-03] | [4.806020259857178, 5.480987548828125] |
5263bae8-0418-4925-961b-7b7854ffe5fc | global-context-aware-progressive-aggregation | 2003.00651 | null | https://arxiv.org/abs/2003.00651v1 | https://arxiv.org/pdf/2003.00651v1.pdf | Global Context-Aware Progressive Aggregation Network for Salient Object Detection | Deep convolutional neural networks have achieved competitive performance in salient object detection, in which how to learn effective and comprehensive features plays a critical role. Most of the previous works mainly adopted multiple level feature integration yet ignored the gap between different features. Besides, there also exists a dilution process of high-level features as they passed on the top-down pathway. To remedy these issues, we propose a novel network named GCPANet to effectively integrate low-level appearance features, high-level semantic features, and global context features through some progressive context-aware Feature Interweaved Aggregation (FIA) modules and generate the saliency map in a supervised way. Moreover, a Head Attention (HA) module is used to reduce information redundancy and enhance the top layers features by leveraging the spatial and channel-wise attention, and the Self Refinement (SR) module is utilized to further refine and heighten the input features. Furthermore, we design the Global Context Flow (GCF) module to generate the global context information at different stages, which aims to learn the relationship among different salient regions and alleviate the dilution effect of high-level features. Experimental results on six benchmark datasets demonstrate that the proposed approach outperforms the state-of-the-art methods both quantitatively and qualitatively. | ['Qianqian Xu', 'Zuyao Chen', 'Runmin Cong', 'Qingming Huang'] | 2020-03-02 | null | null | null | null | ['dichotomous-image-segmentation'] | ['computer-vision'] | [ 1.41613230e-01 -2.32898951e-01 3.56099494e-02 -4.19154495e-01
-3.41782302e-01 1.77180484e-01 4.16426420e-01 3.51739854e-01
-5.03770709e-01 4.27141517e-01 4.44361269e-01 2.16636017e-01
-2.42390633e-01 -9.37759578e-01 -5.59441745e-01 -7.80849516e-01
8.86519626e-02 -5.22463739e-01 9.58670497e-01 -4.26331043e-01
3.90922189e-01 1.76261455e-01 -1.76668513e+00 2.59974778e-01
1.14634442e+00 1.18249273e+00 6.79116726e-01 -5.07861674e-02
-3.11868578e-01 7.81408429e-01 -2.55563051e-01 -4.82342206e-02
-4.72425483e-03 -3.30025911e-01 -5.80007434e-01 1.01659589e-01
1.29055813e-01 -4.03794229e-01 -1.54491693e-01 1.36937141e+00
4.79054868e-01 4.46143076e-02 1.57363802e-01 -1.11716306e+00
-7.00387597e-01 4.84205335e-01 -9.91830468e-01 5.79603374e-01
-4.93134931e-02 1.95046976e-01 9.45953488e-01 -1.21193349e+00
3.70058864e-01 1.33041942e+00 3.36440802e-01 9.80821177e-02
-7.70808876e-01 -7.02647507e-01 7.45207846e-01 4.89041626e-01
-1.36293173e+00 -9.30877402e-02 1.22366774e+00 -1.26009181e-01
5.13093591e-01 5.50393425e-02 8.21038783e-01 4.69652086e-01
6.31494820e-02 1.02438152e+00 8.16531241e-01 -1.95537195e-01
-8.19057003e-02 1.55915432e-02 2.09444746e-01 9.14264917e-01
2.62254596e-01 -1.10167399e-01 -5.19272864e-01 1.79033071e-01
9.01934087e-01 4.41242367e-01 -2.21960887e-01 -3.19474757e-01
-1.26743913e+00 8.17011118e-01 1.21576142e+00 5.30948818e-01
-6.60007298e-01 -1.35263994e-01 4.28697169e-01 -1.19891196e-01
3.07852924e-01 3.87710124e-01 -4.44866627e-01 4.78867292e-01
-7.91069508e-01 7.52666220e-02 -1.35485204e-02 7.89311826e-01
1.26979673e+00 -1.92157887e-02 -7.20754743e-01 8.17747593e-01
3.45522106e-01 -3.87504213e-02 6.64526701e-01 -6.29898190e-01
5.32248020e-01 1.17491150e+00 -3.66688631e-02 -1.41698635e+00
-5.15115917e-01 -1.00778174e+00 -8.30752134e-01 -1.69606134e-02
-3.11608482e-02 -3.71718779e-02 -9.90818322e-01 1.76329124e+00
5.33118784e-01 2.07094833e-01 -1.98347494e-01 1.37125409e+00
1.14035392e+00 6.28061593e-01 4.10710126e-01 -2.36348018e-01
1.42751789e+00 -1.33099020e+00 -5.90326548e-01 -3.24446321e-01
2.74635196e-01 -7.82532692e-01 1.10316980e+00 -1.85600579e-01
-1.09744501e+00 -9.13151801e-01 -1.11015987e+00 -3.30151498e-01
-4.51213896e-01 2.10005715e-01 7.76337266e-01 -1.23679079e-01
-8.49197030e-01 3.21163952e-01 -4.51454401e-01 -1.03462428e-01
8.69343877e-01 3.84862900e-01 -8.10846314e-02 -4.54860553e-02
-1.48022413e+00 6.01898432e-01 5.42950571e-01 4.22484487e-01
-7.68676817e-01 -7.48404026e-01 -1.01141334e+00 3.40361446e-01
5.18216550e-01 -6.64650321e-01 8.82048249e-01 -1.15921068e+00
-9.92038190e-01 3.63768131e-01 -1.75280645e-01 -4.84078117e-02
3.08250129e-01 -1.52372539e-01 -3.48294079e-01 2.74332911e-01
3.03559005e-01 9.60107565e-01 8.44940901e-01 -1.21619284e+00
-1.23960590e+00 -3.55315804e-01 1.69336140e-01 6.33549809e-01
-6.43575609e-01 1.08016431e-01 -6.14050925e-01 -6.71357989e-01
3.39003980e-01 -1.81755275e-01 -3.71595681e-01 3.16948108e-02
-2.88111866e-01 -2.39420176e-01 1.13219607e+00 -5.60653508e-01
1.39919782e+00 -2.20186067e+00 2.24475086e-01 6.44388273e-02
4.18747246e-01 4.81742978e-01 -2.19709948e-01 8.74336716e-03
2.84241829e-02 -1.03348732e-01 -2.26110280e-01 -7.13188723e-02
-3.59361261e-01 -6.90923631e-02 -1.50425762e-01 5.20391250e-03
6.38483703e-01 1.07926798e+00 -1.22104847e+00 -7.85202026e-01
3.75546634e-01 6.16134703e-01 -5.19418359e-01 -2.69452687e-02
-7.83664174e-03 3.45494151e-01 -8.56985211e-01 6.22177243e-01
8.02125692e-01 -3.29944611e-01 -4.37894106e-01 -3.79680753e-01
-4.31654185e-01 1.06180154e-01 -1.08725715e+00 1.72479498e+00
-2.55565137e-01 1.26674905e-01 5.83047718e-02 -8.05550933e-01
1.04324853e+00 -2.80388921e-01 2.89994687e-01 -8.87612224e-01
3.08036625e-01 9.46056694e-02 1.69582412e-01 -4.73263711e-01
6.60960615e-01 1.83144268e-02 1.66779961e-02 -7.16356635e-02
8.23212266e-02 4.18996900e-01 -1.02662161e-01 3.33366662e-01
6.46438420e-01 3.44214812e-02 2.09509894e-01 -4.28323656e-01
1.03772986e+00 -1.63857445e-01 9.03712153e-01 4.04139817e-01
-5.13640404e-01 5.64682007e-01 3.11722398e-01 -4.19504672e-01
-6.57939732e-01 -6.66429043e-01 3.78852375e-02 1.30492198e+00
9.72317040e-01 -3.67120236e-01 -6.30980492e-01 -7.04146206e-01
-4.90754321e-02 1.79799482e-01 -8.79522800e-01 -5.64008415e-01
-4.62706417e-01 -8.44470561e-01 -2.08260626e-01 7.35933423e-01
1.10744298e+00 -1.41867459e+00 -6.70387387e-01 4.51092809e-01
-7.75048286e-02 -8.32006156e-01 -5.77416778e-01 -2.11723298e-02
-7.21782744e-01 -7.88443327e-01 -7.31535673e-01 -1.07941949e+00
8.58877003e-01 8.54365289e-01 6.71059549e-01 5.00651658e-01
-3.47317606e-01 -2.60305613e-01 -5.00453293e-01 -4.46214765e-01
3.40885878e-01 2.49822736e-01 -4.38103646e-01 3.96682084e-01
2.51356840e-01 -4.97009933e-01 -1.07060897e+00 3.06322366e-01
-9.02453899e-01 3.95182401e-01 8.80990446e-01 1.04735887e+00
6.67625725e-01 2.06698164e-01 7.64859140e-01 -5.41777909e-01
3.98835987e-01 -4.46852356e-01 -2.80275404e-01 1.06222004e-01
-4.17920560e-01 -4.79727946e-02 7.21459508e-01 -2.24891335e-01
-1.28304160e+00 -4.28852998e-02 -7.61420354e-02 -3.68389189e-01
1.03798307e-01 4.87100631e-01 -5.17549634e-01 -1.34176522e-01
1.32459939e-01 5.07001936e-01 -3.07999015e-01 -4.19622123e-01
2.50374317e-01 3.37933600e-01 4.59129274e-01 -2.68647343e-01
7.57934809e-01 5.01622796e-01 -1.13138534e-01 -3.67641389e-01
-1.31284416e+00 -4.06792462e-01 -6.71141267e-01 -1.31243274e-01
8.49396169e-01 -1.10992670e+00 -2.86041290e-01 7.85346508e-01
-8.93372357e-01 4.21359092e-02 -3.27729046e-01 2.34044611e-01
-1.15954556e-01 1.94622457e-01 -5.53199947e-01 -5.37130952e-01
-5.42456985e-01 -1.18936551e+00 1.17697740e+00 1.08383334e+00
4.14230227e-01 -5.39187968e-01 -5.74477732e-01 -2.16308516e-02
5.69576085e-01 2.17352271e-01 7.30850279e-01 -1.68785706e-01
-8.63504827e-01 1.24606676e-01 -9.22286749e-01 8.06593671e-02
3.72227699e-01 2.80972160e-02 -7.74369001e-01 -2.09095895e-01
-2.22706139e-01 -1.29274577e-01 1.27008569e+00 3.69534016e-01
1.41961515e+00 -1.86142236e-01 -4.65133727e-01 5.88255346e-01
1.37020695e+00 -3.13924588e-02 4.46777791e-01 4.26700532e-01
1.10344982e+00 5.72744071e-01 8.91936183e-01 4.02909726e-01
6.38923287e-01 4.67623055e-01 6.59556150e-01 -5.78089297e-01
-2.10172325e-01 -4.67828304e-01 -5.89388125e-02 6.61338508e-01
7.55648389e-02 5.09145021e-01 -3.33718866e-01 7.77172625e-01
-2.02352357e+00 -8.86214972e-01 -2.03511864e-01 1.86681056e+00
8.53865981e-01 3.19674432e-01 1.71877801e-01 1.48416713e-01
1.05450833e+00 2.75236279e-01 -6.27936542e-01 4.51962240e-02
-2.45747834e-01 -1.42778456e-01 1.25828803e-01 2.10965991e-01
-1.32142782e+00 1.07064450e+00 4.95873833e+00 9.79092777e-01
-1.24774587e+00 1.46853432e-01 7.92604864e-01 1.03584044e-01
-4.03161824e-01 6.30786875e-03 -7.42880762e-01 7.28743255e-01
-2.84626894e-02 -1.12005174e-01 5.88483363e-02 8.81450593e-01
1.90064505e-01 -2.64398575e-01 -2.79498249e-01 6.73045933e-01
-7.89713785e-02 -1.25696969e+00 1.74417824e-01 -2.72273421e-01
8.75199676e-01 -7.67539740e-02 1.25424728e-01 4.13883269e-01
-1.18796401e-01 -5.39600432e-01 8.17568064e-01 5.80418885e-01
4.06068087e-01 -9.93414760e-01 8.29172254e-01 2.47440949e-01
-1.71597886e+00 -4.56593305e-01 -6.73263788e-01 -6.64199367e-02
-1.23170145e-01 7.02544570e-01 -2.63169974e-01 8.43266010e-01
8.40690970e-01 9.98346746e-01 -9.69635367e-01 1.32453668e+00
-3.54154140e-01 2.47850046e-01 -8.88224691e-02 -4.52655777e-02
5.98386467e-01 7.12016448e-02 4.68992710e-01 1.04845691e+00
8.76368135e-02 1.80967852e-01 2.03859359e-01 8.54635298e-01
7.42739215e-02 2.88663238e-01 1.92750972e-02 5.83099902e-01
5.02305925e-01 1.74082315e+00 -8.35685074e-01 -5.02825201e-01
-5.02299368e-01 8.16980660e-01 4.47594464e-01 2.39600316e-01
-6.66003108e-01 -6.80735409e-01 4.89536524e-01 1.23478182e-01
5.57495773e-01 1.50314242e-01 -2.84788400e-01 -1.12416053e+00
1.82676256e-01 -3.20143133e-01 5.48578799e-01 -7.56623447e-01
-1.05090666e+00 5.31738341e-01 -2.06467867e-01 -1.32872820e+00
4.36845869e-01 -9.90562737e-02 -9.80823874e-01 1.14816403e+00
-2.19282174e+00 -1.25984097e+00 -7.20943272e-01 7.07456529e-01
6.31230354e-01 8.49225968e-02 1.12854153e-01 3.53516936e-01
-7.89979696e-01 5.10858715e-01 -3.69923204e-01 6.11980222e-02
4.65574473e-01 -9.40945983e-01 -2.00362392e-02 1.05583096e+00
-3.95298302e-01 6.95084929e-01 3.32843930e-01 -5.79014361e-01
-9.41363156e-01 -1.44817662e+00 5.89570642e-01 1.91193223e-01
4.02794957e-01 -1.46620527e-01 -1.20506132e+00 1.76684484e-01
-4.56939526e-02 4.82454002e-01 1.21256694e-01 -1.77352950e-01
-2.20111132e-01 -3.81983548e-01 -9.40200329e-01 4.97002929e-01
1.08239019e+00 -2.24461406e-01 -7.13131130e-01 -1.89273342e-01
1.14251971e+00 -2.39919811e-01 -4.30773795e-01 7.44148910e-01
3.34496886e-01 -1.13708448e+00 8.70756030e-01 -2.30546892e-01
6.50710821e-01 -8.91861141e-01 1.94942832e-01 -1.09213567e+00
-8.08537900e-01 -3.37518692e-01 -1.09372726e-02 1.52894711e+00
1.03288025e-01 -4.76189137e-01 4.45935011e-01 1.79932728e-01
-4.00880158e-01 -1.15293610e+00 -5.50273478e-01 -3.01004559e-01
-2.80405670e-01 8.55100304e-02 9.68856633e-01 8.39056730e-01
-1.36406571e-01 4.82069492e-01 -2.19289929e-01 1.40373319e-01
3.84300143e-01 6.03103697e-01 3.07624727e-01 -1.21453369e+00
1.37159541e-01 -6.97428524e-01 -4.94922221e-01 -8.88469517e-01
-2.58564621e-01 -8.37368906e-01 4.75193672e-02 -1.60936344e+00
6.06832385e-01 -4.87194419e-01 -9.50561047e-01 5.43094099e-01
-1.12633002e+00 1.78903162e-01 3.16533446e-01 9.09828842e-02
-9.58646715e-01 9.56992805e-01 1.79896879e+00 -3.37712164e-03
-2.42344081e-01 -2.88213938e-01 -1.34216213e+00 6.91895485e-01
6.19230151e-01 -1.82964653e-01 -4.39117283e-01 -3.77778947e-01
-1.16135322e-01 -3.54670346e-01 6.12728179e-01 -1.06694102e+00
2.93571889e-01 -1.72743425e-01 8.40926647e-01 -8.08705986e-01
4.31267843e-02 -7.10094690e-01 -5.43147862e-01 3.78656298e-01
-2.51239508e-01 -2.52768546e-01 1.46996781e-01 4.76733357e-01
-4.77997363e-01 9.07384306e-02 8.62822175e-01 -1.65348023e-01
-1.12846255e+00 6.33829415e-01 8.10166821e-02 2.21356973e-02
1.05061853e+00 -3.78355533e-02 -4.34176505e-01 7.21578747e-02
-3.48753482e-01 6.69961810e-01 3.96950692e-01 7.13466465e-01
7.92561233e-01 -1.44227839e+00 -6.50036395e-01 4.97056097e-01
2.56051838e-01 2.78728068e-01 7.56091893e-01 8.01666021e-01
3.88438557e-03 1.55672595e-01 -5.24842381e-01 -5.64696670e-01
-9.17340636e-01 6.44101620e-01 1.98068112e-01 -2.26368919e-01
-5.30705273e-01 1.15008759e+00 7.21386015e-01 2.65452731e-02
-1.94793232e-02 -2.72660106e-01 -6.39601171e-01 2.20681489e-01
8.69427621e-01 7.53237754e-02 -6.67061210e-02 -9.04690862e-01
-5.11697590e-01 7.51740932e-01 -3.92782420e-01 3.76558930e-01
1.27574444e+00 -3.91570061e-01 -1.66423827e-01 -1.53179304e-03
1.13712835e+00 -2.42726043e-01 -1.63606954e+00 -5.94108522e-01
-2.85183251e-01 -5.58344305e-01 3.13940138e-01 -6.26225173e-01
-1.38538432e+00 9.78990674e-01 5.00191391e-01 7.50031844e-02
1.59412956e+00 -3.01914830e-02 9.58745778e-01 -2.24663123e-01
2.14685440e-01 -1.01152909e+00 2.51746207e-01 3.39035094e-01
8.16500366e-01 -1.17930579e+00 2.01277845e-02 -7.20451891e-01
-6.79544866e-01 7.70284414e-01 1.13356268e+00 -3.02169800e-01
6.33080065e-01 -3.21704075e-02 -1.98282018e-01 -1.53094858e-01
-5.58255911e-01 -6.98867917e-01 4.03742969e-01 3.69284660e-01
5.23089543e-02 -1.13572858e-01 -5.02220988e-01 1.11163402e+00
1.45314053e-01 -3.95234041e-02 1.74531013e-01 8.97214472e-01
-8.29007983e-01 -7.18734145e-01 -1.41364589e-01 4.90073323e-01
-2.60735959e-01 -3.07019114e-01 -6.09139651e-02 5.72632134e-01
6.01223648e-01 8.30629051e-01 4.07744758e-02 -5.57582378e-01
2.51202762e-01 -3.89079809e-01 1.42200179e-02 -4.97138917e-01
-6.82263434e-01 1.89406484e-01 -4.20234740e-01 -5.83045661e-01
-5.33356190e-01 -6.66787505e-01 -1.51064265e+00 3.26803714e-01
-4.57809836e-01 1.11908786e-01 1.49076551e-01 9.61910903e-01
5.30245245e-01 1.00796568e+00 9.49520767e-01 -9.58369493e-01
1.75105818e-02 -9.46631670e-01 -3.71389091e-01 3.57819468e-01
5.53571105e-01 -8.80674660e-01 -2.15056375e-01 -1.90544665e-01] | [9.739144325256348, -0.4727803170681] |
5d863a25-470c-40ba-84a7-6b63a7beda65 | rethinking-image-based-table-recognition | 2303.07641 | null | https://arxiv.org/abs/2303.07641v1 | https://arxiv.org/pdf/2303.07641v1.pdf | Rethinking Image-based Table Recognition Using Weakly Supervised Methods | Most of the previous methods for table recognition rely on training datasets containing many richly annotated table images. Detailed table image annotation, e.g., cell or text bounding box annotation, however, is costly and often subjective. In this paper, we propose a weakly supervised model named WSTabNet for table recognition that relies only on HTML (or LaTeX) code-level annotations of table images. The proposed model consists of three main parts: an encoder for feature extraction, a structure decoder for generating table structure, and a cell decoder for predicting the content of each cell in the table. Our system is trained end-to-end by stochastic gradient descent algorithms, requiring only table images and their ground-truth HTML (or LaTeX) representations. To facilitate table recognition with deep learning, we create and release WikiTableSet, the largest publicly available image-based table recognition dataset built from Wikipedia. WikiTableSet contains nearly 4 million English table images, 590K Japanese table images, and 640k French table images with corresponding HTML representation and cell bounding boxes. The extensive experiments on WikiTableSet and two large-scale datasets: FinTabNet and PubTabNet demonstrate that the proposed weakly supervised model achieves better, or similar accuracies compared to the state-of-the-art models on all benchmark datasets. | ['Hideaki Takeda', 'Phuc Nguyen', 'Atsuhiro Takasu', 'Nam Tuan Ly'] | 2023-03-14 | null | null | null | null | ['table-recognition'] | ['computer-vision'] | [ 2.37285167e-01 -8.27713162e-02 -3.86151612e-01 -5.00256419e-01
-1.44826615e+00 -7.78814256e-01 2.67339438e-01 3.18248600e-01
-1.66367620e-01 9.40740049e-01 3.23287636e-01 -6.00134917e-02
4.75268811e-01 -9.17281926e-01 -1.31838775e+00 -5.24468362e-01
2.04550683e-01 8.37396145e-01 1.72814861e-01 -7.10353628e-02
2.32405201e-01 1.95881296e-02 -1.34738564e+00 1.17002833e+00
7.91075766e-01 1.84491944e+00 6.11224100e-02 8.67711186e-01
-5.54896057e-01 1.61442232e+00 -5.47833085e-01 -1.06031954e+00
2.52318919e-01 -2.18384996e-01 -7.14797318e-01 2.61462450e-01
7.24458575e-01 -4.70723271e-01 -5.42567015e-01 8.67178142e-01
2.94380575e-01 -3.95187259e-01 6.16086960e-01 -1.09700131e+00
-9.40743387e-01 1.09529161e+00 -4.18564647e-01 -7.93034509e-02
4.77069229e-01 2.63562221e-02 1.01960921e+00 -9.87616956e-01
7.95057416e-01 1.01301539e+00 6.32593989e-01 1.39551923e-01
-1.01465118e+00 -6.89966798e-01 2.17068736e-02 1.57683969e-01
-1.64358866e+00 -6.95338488e-01 5.79763770e-01 -4.80692476e-01
1.05969167e+00 9.06951278e-02 2.32882366e-01 1.01840615e+00
2.71659106e-01 1.13693321e+00 8.43607605e-01 -1.67795479e-01
-1.03896186e-01 6.29818201e-01 -1.30477294e-01 1.13350034e+00
5.88104725e-01 -7.63587534e-01 -7.62336731e-01 1.43398494e-01
6.02611780e-01 -3.70946616e-01 -7.32313544e-02 -8.96193564e-01
-1.70720279e+00 3.86482954e-01 3.74168515e-01 -2.22839639e-01
-3.86732630e-02 6.87617064e-03 8.68575752e-01 2.47489233e-02
1.35356784e-01 1.34875961e-02 -6.71573877e-01 -2.81290263e-01
-4.56126750e-01 6.33654743e-03 1.08479178e+00 1.72362423e+00
7.48097837e-01 -1.53349295e-01 -3.94913405e-01 9.43948269e-01
8.17230567e-02 7.32456565e-01 9.12685096e-02 -4.28647995e-01
1.56885767e+00 1.06394994e+00 -3.15689668e-02 -1.12497079e+00
-1.15664989e-01 -1.26630422e-02 -1.21811116e+00 -4.08005416e-01
6.71317875e-01 1.25845477e-01 -8.11302543e-01 1.06881559e+00
1.09527722e-01 -6.17034435e-01 3.62081826e-01 5.52493691e-01
1.13589978e+00 8.39194775e-01 -2.41360411e-01 1.78582981e-01
1.50783491e+00 -1.21131611e+00 -9.45752442e-01 -2.36087173e-01
7.76286900e-01 -8.27803493e-01 1.43972969e+00 4.97184902e-01
-9.74466324e-01 -5.38430274e-01 -1.12110138e+00 -4.96758521e-01
-9.59840119e-01 8.35029960e-01 5.56085944e-01 6.24994814e-01
-4.89413530e-01 -3.29206586e-02 -4.94192600e-01 -3.10284644e-01
7.88018882e-01 3.03020984e-01 -6.22007191e-01 -1.65270388e-01
-8.04551244e-01 3.70240211e-01 5.58447897e-01 3.16723615e-01
-7.60796309e-01 -4.38044041e-01 -1.30566251e+00 1.80337653e-01
1.03023970e+00 -1.89091295e-01 9.67414618e-01 -4.47964191e-01
-1.06861377e+00 1.26330471e+00 -4.19298373e-03 -3.44287723e-01
5.71217954e-01 -3.02892122e-02 -3.21083635e-01 -5.66354021e-02
2.70549983e-01 5.82113028e-01 4.34245229e-01 -1.29493034e+00
-5.89653492e-01 -3.98509264e-01 -2.47497499e-01 1.06786042e-01
-2.41884843e-01 -1.43044591e-01 -1.40688705e+00 -6.98076606e-01
-1.50519535e-01 -5.80202699e-01 2.99644887e-01 8.07756111e-02
-1.08170617e+00 1.08828709e-01 3.31477880e-01 -8.93278062e-01
1.47990561e+00 -2.11598444e+00 -3.45013291e-01 4.38044742e-02
2.05389075e-02 -2.36228988e-01 1.46942809e-01 3.35031837e-01
2.80035675e-01 -4.27055359e-03 -3.99317056e-01 -1.85756348e-02
2.75637537e-01 5.46425246e-02 -3.14066827e-01 2.49147534e-01
1.41595095e-01 9.43361580e-01 -4.22948092e-01 -1.02715647e+00
-6.10504858e-02 2.17026621e-01 -4.88564253e-01 2.86172092e-01
-3.57719749e-01 -2.26146400e-01 -3.05833817e-01 1.18289101e+00
7.49889255e-01 -6.78755283e-01 4.27191496e-01 -5.76055229e-01
1.34642571e-01 3.90356630e-01 -1.47247267e+00 1.64281225e+00
-1.10118970e-01 3.11565280e-01 -4.74545024e-02 -6.99329793e-01
1.10828829e+00 -8.23144242e-02 1.91575065e-01 -8.66505682e-01
2.02234015e-01 4.29437369e-01 -2.65773922e-01 -3.27710092e-01
5.92311978e-01 7.56926358e-01 -6.86304331e-01 1.92554146e-01
-8.25134479e-03 1.29499491e-02 7.73445368e-01 3.98619264e-01
8.14635813e-01 3.40947300e-01 3.34085643e-01 -1.94048077e-01
1.06504381e+00 3.51895005e-01 4.59835440e-01 7.31178284e-01
-1.42729059e-02 7.17596292e-01 9.39075112e-01 -6.79822803e-01
-1.30542541e+00 -9.34991658e-01 -1.45892441e-01 1.00460923e+00
2.73674339e-01 -6.69559717e-01 -1.03721356e+00 -9.02372181e-01
3.01557988e-01 2.88264398e-02 -8.89040411e-01 1.23799324e-01
-7.40186930e-01 -5.65198183e-01 5.60217977e-01 9.71429050e-01
1.07840979e+00 -1.36978185e+00 9.73894745e-02 1.45708084e-01
-2.67428935e-01 -1.74051654e+00 -1.04686844e+00 5.52129447e-01
-3.78518790e-01 -1.24974728e+00 -2.81605184e-01 -1.23764145e+00
8.96804333e-01 -2.46920392e-01 1.41930020e+00 -7.29920045e-02
-2.52874672e-01 -8.57295990e-02 3.26404697e-03 -1.92112759e-01
-3.04469049e-01 3.69533062e-01 -3.00856620e-01 2.08886385e-01
3.21745694e-01 3.48809481e-01 -1.36889204e-01 5.14781296e-01
-7.24547446e-01 3.82337213e-01 7.92034447e-01 8.08641672e-01
1.21792769e+00 -2.10861284e-02 -5.17355604e-03 -1.53301370e+00
1.57662705e-01 -4.39825952e-02 -9.97703195e-01 6.58761203e-01
-5.30454636e-01 4.10811126e-01 1.06872046e+00 -2.15315036e-02
-1.11471164e+00 3.35569471e-01 -1.22146070e-01 4.98498976e-02
5.45258680e-03 4.14239109e-01 -8.33870649e-01 2.47965366e-01
3.93956006e-01 6.83425963e-01 -3.15133750e-01 -6.28834903e-01
1.63348928e-01 9.52816546e-01 9.12394524e-01 -7.11739182e-01
8.58783364e-01 5.01251817e-01 -4.53551710e-01 -1.98217228e-01
-1.10001004e+00 -8.49014968e-02 -1.02248073e+00 1.61648951e-02
1.04976988e+00 -1.15488148e+00 -9.76399124e-01 6.73043549e-01
-5.99364638e-01 -5.96633494e-01 6.96915165e-02 6.70064613e-02
-7.72066653e-01 -1.04917735e-01 -8.84131491e-01 -3.36492300e-01
-2.80854195e-01 -1.24274790e+00 1.20489192e+00 -1.01251945e-01
2.86072075e-01 -5.68245053e-01 -3.98478359e-01 8.79380286e-01
1.73340544e-01 1.45567432e-01 1.10908771e+00 -6.39178395e-01
-1.22105038e+00 -3.51999342e-01 -8.09427738e-01 1.21587627e-01
8.06600899e-02 5.09327985e-02 -7.02060342e-01 1.56325191e-01
-8.01984549e-01 -9.43828166e-01 8.23790491e-01 -1.17478460e-01
1.66298068e+00 -5.33320069e-01 -2.71226645e-01 1.03820682e+00
1.65650654e+00 4.01225537e-01 8.10494542e-01 7.12278247e-01
1.17531574e+00 4.37452286e-01 8.04380298e-01 5.61179876e-01
8.70188177e-01 5.49883425e-01 1.55278668e-01 6.04730397e-02
-1.73342451e-02 -6.89012766e-01 3.27359617e-01 8.01600695e-01
6.42745733e-01 -6.38784885e-01 -1.03794920e+00 2.85346836e-01
-1.69328094e+00 -6.78250730e-01 4.85742465e-03 2.02733159e+00
1.14128065e+00 7.19752371e-01 6.49184361e-02 2.27363646e-01
6.36219621e-01 2.36635685e-01 -7.43782341e-01 -3.77655596e-01
-3.92585546e-01 -4.27654296e-01 8.27591419e-01 1.38745502e-01
-1.44216716e+00 1.04508018e+00 5.52496147e+00 9.44810152e-01
-6.26535416e-01 -5.12312651e-01 1.30768204e+00 2.63651401e-01
-1.75591074e-02 -4.87743884e-01 -1.35560060e+00 3.76594692e-01
7.54475653e-01 1.08195163e-01 5.06144583e-01 1.17137468e+00
-4.72826898e-01 -1.74864978e-01 -1.14888489e+00 1.29905236e+00
5.24848938e-01 -1.83135855e+00 3.34427714e-01 -4.40983772e-02
5.72822869e-01 -4.06399131e-01 -8.60458538e-02 6.62613451e-01
8.77071992e-02 -9.45922434e-01 1.04792726e+00 2.22170755e-01
1.43740857e+00 -7.74064660e-01 8.91846657e-01 -8.45466852e-02
-1.69308829e+00 1.35633960e-01 -3.91083121e-01 3.36471915e-01
-5.32446682e-01 9.21915174e-02 -5.92898309e-01 1.66668192e-01
1.00315714e+00 8.17752957e-01 -1.00394177e+00 4.58391070e-01
2.42276136e-02 3.24529409e-01 -1.63364634e-01 -1.78765178e-01
1.26446754e-01 -7.88658410e-02 -3.83979261e-01 1.37428713e+00
-1.80074289e-01 -2.01650336e-01 3.51092786e-01 6.38155520e-01
-9.27242994e-01 4.17725533e-01 -5.64783514e-01 -3.39457065e-01
4.08289164e-01 1.29323876e+00 -1.30266201e+00 -7.95879602e-01
-6.89236462e-01 9.31867540e-01 4.91817236e-01 -2.87555065e-02
-9.45198894e-01 -7.81912684e-01 4.13517058e-01 -2.06955094e-02
8.79101396e-01 1.61134318e-01 -8.55154395e-01 -1.40156579e+00
5.41849554e-01 -1.23850071e+00 7.44634449e-01 -9.66046631e-01
-9.84188735e-01 9.20834064e-01 -3.23339701e-01 -1.08224070e+00
-7.04922248e-03 -9.81286108e-01 2.32761949e-01 5.76086700e-01
-1.39495480e+00 -1.06647587e+00 -6.67043626e-01 7.32889473e-01
5.07389307e-01 -3.29335898e-01 6.46559596e-01 5.07329404e-01
-9.67738986e-01 1.09216499e+00 3.27908576e-01 1.11282694e+00
9.14008200e-01 -1.64319694e+00 6.03050649e-01 6.69664621e-01
1.31197140e-01 3.21205646e-01 1.43640637e-01 -7.52644777e-01
-2.08931875e+00 -1.18548679e+00 9.00240004e-01 -7.25007951e-01
5.23253679e-01 -1.44235563e+00 -8.92049253e-01 8.71203840e-01
1.55740365e-01 3.59789401e-01 4.28719342e-01 -2.05914438e-01
-7.08662748e-01 -8.33726048e-01 -1.17323637e+00 5.29874086e-01
9.38373268e-01 -6.31252348e-01 2.78848619e-03 5.17854989e-01
5.97978592e-01 -9.63053048e-01 -9.92518306e-01 2.77006894e-01
7.60954440e-01 -8.00075889e-01 8.18530679e-01 -6.63512275e-02
6.92601264e-01 -4.25752550e-01 -4.32830691e-01 -4.21243727e-01
-4.93520312e-02 -3.39479983e-01 -3.43609542e-01 1.42207611e+00
1.00915873e+00 -1.27196806e-02 1.42097342e+00 4.69336063e-01
-1.36660680e-01 -1.00581408e+00 -1.95097879e-01 -5.93753338e-01
-3.26499730e-01 -2.94420809e-01 7.20036209e-01 6.47282004e-01
-1.92356825e-01 3.90605450e-01 -4.25956517e-01 -1.12654241e-02
5.18731236e-01 4.64368165e-01 1.28968406e+00 -6.92241132e-01
7.25267753e-02 -1.47176206e-01 -2.69066006e-01 -1.23783398e+00
-1.57454312e-01 -8.16850066e-01 3.51410985e-01 -1.65208662e+00
6.16913021e-01 -2.76719004e-01 -3.48061025e-02 6.00794137e-01
-2.07466334e-01 5.56898415e-01 3.18294056e-02 2.21386731e-01
-1.15955770e+00 2.65224189e-01 1.26807225e+00 -4.81718570e-01
1.49260655e-01 -7.08241642e-01 -8.31737101e-01 3.85433257e-01
4.66252446e-01 -4.30281013e-01 -1.73637331e-01 -4.62141424e-01
4.06265318e-01 1.90955862e-01 -3.76132637e-01 -1.07678127e+00
2.60174841e-01 -3.42726079e-03 1.03102219e+00 -1.31040871e+00
6.38181120e-02 -7.92538464e-01 -2.19527677e-01 3.00850093e-01
-6.28982484e-01 2.55035162e-01 3.42288285e-01 4.68460590e-01
-4.34307724e-01 2.09905878e-01 5.53801835e-01 -2.48880193e-01
-7.52805471e-01 5.96157193e-01 -2.69687206e-01 4.27108943e-01
8.06746781e-01 -3.46274823e-01 -3.61273855e-01 -1.02229275e-01
-3.08866292e-01 3.86793405e-01 4.98291701e-01 3.94793779e-01
5.84640682e-01 -1.39292204e+00 -6.39967442e-01 4.72337633e-01
8.20580363e-01 1.71636790e-01 -3.12270243e-02 4.67288047e-01
-1.10058606e+00 7.65377760e-01 -9.76372585e-02 -4.63419437e-01
-1.09319532e+00 8.19098353e-01 4.75875027e-02 -6.17280960e-01
-5.37032187e-01 9.71425712e-01 2.86364794e-01 -6.38234735e-01
7.88946509e-01 -7.49592960e-01 -6.27857149e-02 1.79584846e-02
5.16258121e-01 -2.80117035e-01 3.55340570e-01 -5.40122390e-01
-5.86752295e-01 4.37873185e-01 -4.77293581e-01 4.00130659e-01
1.02622628e+00 -2.71132179e-02 1.93909183e-02 5.17552018e-01
1.17630684e+00 2.30831176e-01 -1.36133730e+00 -1.72167391e-01
1.39093876e-01 -4.37566131e-01 -5.09322882e-01 -1.17913651e+00
-1.09978020e+00 5.88690460e-01 1.51448503e-01 -1.57989889e-01
1.08326125e+00 -1.62444219e-01 1.02878678e+00 9.20579791e-01
5.00950336e-01 -1.07566452e+00 7.23064840e-02 4.65171486e-01
7.71733403e-01 -1.72914338e+00 -3.68771888e-03 -7.04419494e-01
-1.01251316e+00 1.15043080e+00 1.13723695e+00 5.29523157e-02
4.38975900e-01 7.42164910e-01 2.85506576e-01 -1.29352147e-02
-1.00921893e+00 -8.00504237e-02 2.83565164e-01 4.51796621e-01
6.92160606e-01 -2.79785395e-01 2.01840699e-01 7.76927590e-01
-3.06579113e-01 -1.03400119e-01 5.44542074e-01 1.10288739e+00
-1.83681265e-01 -8.81045401e-01 -5.18573165e-01 8.08168113e-01
-4.96260852e-01 -2.82462537e-01 -5.74613035e-01 1.03395951e+00
-2.93973703e-02 6.54405117e-01 1.41673505e-01 -4.07815397e-01
3.57849121e-01 1.83816459e-02 4.04879034e-01 -5.73343039e-01
-5.77824593e-01 -8.08334351e-02 3.56523693e-01 -6.38871849e-01
6.20444119e-02 -4.67502981e-01 -1.21324503e+00 -5.38469672e-01
1.73131190e-02 1.84078246e-01 4.03347731e-01 5.38241684e-01
3.11450690e-01 5.44386506e-01 3.45049977e-01 -1.54353857e-01
-9.47314054e-02 -8.26698899e-01 -6.26839280e-01 4.50832456e-01
1.97303236e-01 -4.05925214e-01 9.35964435e-02 5.98030925e-01] | [11.70032024383545, 3.046144485473633] |
e1b6a50c-1183-4918-b6be-771abb618249 | inter-gps-interpretable-geometry-problem | 2105.04165 | null | https://arxiv.org/abs/2105.04165v3 | https://arxiv.org/pdf/2105.04165v3.pdf | Inter-GPS: Interpretable Geometry Problem Solving with Formal Language and Symbolic Reasoning | Geometry problem solving has attracted much attention in the NLP community recently. The task is challenging as it requires abstract problem understanding and symbolic reasoning with axiomatic knowledge. However, current datasets are either small in scale or not publicly available. Thus, we construct a new large-scale benchmark, Geometry3K, consisting of 3,002 geometry problems with dense annotation in formal language. We further propose a novel geometry solving approach with formal language and symbolic reasoning, called Interpretable Geometry Problem Solver (Inter-GPS). Inter-GPS first parses the problem text and diagram into formal language automatically via rule-based text parsing and neural object detecting, respectively. Unlike implicit learning in existing methods, Inter-GPS incorporates theorem knowledge as conditional rules and performs symbolic reasoning step by step. Also, a theorem predictor is designed to infer the theorem application sequence fed to the symbolic solver for the more efficient and reasonable searching path. Extensive experiments on the Geometry3K and GEOS datasets demonstrate that Inter-GPS achieves significant improvements over existing methods. The project with code and data is available at https://lupantech.github.io/inter-gps. | ['Song-Chun Zhu', 'Xiaodan Liang', 'Siyuan Huang', 'Liang Qiu', 'Shibiao Jiang', 'Ran Gong', 'Pan Lu'] | 2021-05-10 | null | https://aclanthology.org/2021.acl-long.528 | https://aclanthology.org/2021.acl-long.528.pdf | acl-2021-5 | ['scene-parsing', 'mathematical-question-answering', 'mathematical-reasoning', 'arithmetic-reasoning'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing', 'reasoning'] | [-1.94400996e-01 4.82167006e-01 -1.62447602e-01 -4.39569920e-01
-8.14079165e-01 -8.66469443e-01 1.95228934e-01 1.02294579e-01
3.10203850e-01 7.68302977e-01 -1.27788216e-01 -9.22193229e-01
-3.88289303e-01 -1.22140098e+00 -1.18511748e+00 -2.28410270e-02
-1.07263923e-01 5.17559350e-01 1.54634610e-01 1.27745524e-01
3.01205248e-01 2.24222630e-01 -1.24425721e+00 4.93782848e-01
1.29347026e+00 1.08020329e+00 -6.13412745e-02 7.63802588e-01
-2.64821023e-01 1.06815553e+00 -1.90073162e-01 -4.51082617e-01
3.03842694e-01 -1.18539706e-01 -1.53583968e+00 -3.34852070e-01
5.09840310e-01 -4.37044799e-01 -3.18949848e-01 1.12622738e+00
-1.88262761e-02 -9.53433514e-02 3.86319071e-01 -1.66315114e+00
-5.94035804e-01 1.07773876e+00 -2.72983283e-01 -3.58240813e-01
6.45155489e-01 6.84289709e-02 1.28282320e+00 -9.35668945e-01
5.31587005e-01 1.45830929e+00 7.01021671e-01 1.82513699e-01
-1.21480155e+00 -8.98533106e-01 2.69565165e-01 5.44522226e-01
-1.81407177e+00 -1.23187505e-01 8.24038804e-01 -4.13481861e-01
1.11830974e+00 3.04095596e-01 5.91814399e-01 4.88960236e-01
-1.88728556e-01 8.84286106e-01 5.59316337e-01 -2.79145390e-01
3.14908981e-01 -2.02088490e-01 2.26619869e-01 1.20904672e+00
7.24254996e-02 -2.47563213e-01 -2.96890467e-01 -2.71421611e-01
9.92210627e-01 -3.88276577e-01 -2.88399868e-02 -3.59908134e-01
-1.26397145e+00 7.49373138e-01 4.80741054e-01 -2.99945235e-01
4.38026227e-02 5.60546994e-01 4.19851959e-01 1.68137655e-01
1.02872021e-01 6.20702505e-01 -8.28385770e-01 -7.03007281e-02
-6.78990126e-01 8.96732211e-01 1.09027910e+00 1.36937606e+00
8.72329414e-01 -3.68581563e-01 2.63254613e-01 4.58836883e-01
2.83892274e-01 5.30969083e-01 -2.15609059e-01 -1.33799911e+00
8.88153553e-01 1.11489391e+00 -1.48781210e-01 -1.28744936e+00
-3.47997129e-01 4.32537384e-02 -5.83665848e-01 1.27382111e-02
6.29406095e-01 -2.63131950e-02 -3.45954835e-01 1.53403652e+00
6.52999043e-01 5.68169832e-01 1.14096873e-01 7.15819895e-01
9.08509791e-01 8.50920677e-01 -1.39714479e-01 3.07950795e-01
1.16829681e+00 -9.23842788e-01 -2.39805028e-01 1.20081663e-01
1.22472394e+00 -2.69210041e-01 1.17850900e+00 7.63923883e-01
-1.11367655e+00 -1.39975846e-01 -9.51683879e-01 -4.27578449e-01
-5.22086799e-01 1.47418350e-01 1.04780567e+00 1.06813863e-01
-6.68070853e-01 4.60765719e-01 -6.76069915e-01 7.86239002e-03
5.58956742e-01 4.21389788e-01 -3.85794006e-02 -2.37901092e-01
-1.31071401e+00 4.65442449e-01 8.46054971e-01 1.13803118e-01
-4.97342557e-01 -1.03369510e+00 -1.25273740e+00 -6.98562562e-02
1.02271914e+00 -7.22505152e-01 1.69865835e+00 -4.85367686e-01
-1.50135565e+00 6.10251665e-01 -9.05018151e-02 -3.99768859e-01
3.95485014e-01 -3.19292128e-01 -1.41954720e-01 2.70423681e-01
3.07452679e-01 5.91607153e-01 3.45096409e-01 -8.81832063e-01
-4.21532750e-01 -1.52714789e-01 6.40924394e-01 -1.03714146e-01
3.54085118e-01 -3.33257504e-02 -6.32803202e-01 -4.17420149e-01
3.21949303e-01 -6.60754740e-01 -6.43782765e-02 2.35097334e-01
-6.93391442e-01 -7.42895901e-01 7.11226165e-01 -6.41091883e-01
1.37194669e+00 -1.82048726e+00 1.93894684e-01 3.77506316e-01
4.34882760e-01 4.72963043e-02 3.29805702e-01 4.23153549e-01
-9.34024677e-02 2.81805992e-01 -3.38716209e-01 2.28920639e-01
3.12171042e-01 3.04342777e-01 -6.41424477e-01 2.05450267e-01
3.86768132e-01 1.22486031e+00 -9.84655857e-01 -9.86900210e-01
3.67209584e-01 -2.66270310e-01 -1.03071260e+00 -1.18758999e-01
-1.06594110e+00 1.12590298e-01 -9.67465818e-01 7.99780309e-01
8.10167253e-01 -6.57181203e-01 2.33608857e-01 2.49691345e-02
-1.14715151e-01 4.66353267e-01 -1.43235803e+00 2.18978071e+00
-4.24558043e-01 3.27205598e-01 -1.19577922e-01 -1.19106019e+00
7.05757856e-01 9.52355638e-02 1.02576211e-01 -6.29618406e-01
2.24781642e-03 2.76349872e-01 -2.99254835e-01 -6.11158907e-01
3.02001923e-01 3.34875762e-01 -4.96414065e-01 2.77320921e-01
-3.39913875e-01 -4.86690819e-01 3.29321712e-01 6.20833635e-01
1.11895990e+00 6.03404701e-01 4.26068217e-01 -3.04498702e-01
8.43226194e-01 6.81737900e-01 8.25787187e-01 5.50256550e-01
3.08796108e-01 2.15994507e-01 9.67567027e-01 -5.72045982e-01
-7.89045751e-01 -1.11269474e+00 -8.50088671e-02 5.76286197e-01
4.39306259e-01 -1.05320406e+00 -8.15200627e-01 -7.14380920e-01
2.63753384e-01 8.48135769e-01 -3.83458704e-01 7.00980891e-03
-8.05756509e-01 5.40853888e-02 9.42091525e-01 8.01414430e-01
7.16657996e-01 -8.94363582e-01 -1.59112468e-01 1.79616675e-01
-2.96545386e-01 -1.35353291e+00 -6.18856214e-02 -2.00143173e-01
-8.06616843e-01 -1.48299205e+00 6.98068663e-02 -7.63965607e-01
8.29020917e-01 -1.71266675e-01 1.18832588e+00 3.95992309e-01
-4.44974840e-01 1.40273273e-01 -2.56120145e-01 -3.30002815e-01
-1.12633467e-01 1.46134958e-01 -2.10249588e-01 -5.61862588e-01
8.97639692e-02 -5.53308785e-01 -2.34744385e-01 1.80050462e-01
-5.46059072e-01 8.13568890e-01 2.43690819e-01 4.55609232e-01
6.86018944e-01 3.83594543e-01 3.72467309e-01 -7.79859006e-01
2.06141070e-01 -3.80950153e-01 -1.15680349e+00 4.07397538e-01
-3.48397166e-01 2.34280899e-01 1.01199162e+00 5.79474866e-03
-9.88184392e-01 1.99862555e-01 1.85456723e-01 -3.68030667e-01
-2.64443159e-01 7.41555750e-01 -5.17651200e-01 8.91767815e-02
4.29734051e-01 1.71969920e-01 -5.26600122e-01 -3.22649986e-01
4.28537130e-01 3.61960143e-01 7.28318572e-01 -1.51672864e+00
1.17009604e+00 1.92514896e-01 2.16330990e-01 -5.93172967e-01
-1.03061080e+00 -8.78650323e-02 -7.43526280e-01 5.42836487e-02
6.62360966e-01 -8.23913753e-01 -1.18589449e+00 3.58183116e-01
-1.47354484e+00 -7.63663769e-01 -3.99971157e-02 2.30546638e-01
-6.72345400e-01 3.34279269e-01 -4.28654611e-01 -6.75462067e-01
-2.71318913e-01 -1.10277116e+00 1.13377810e+00 -7.33540803e-02
-3.94687504e-01 -9.02981758e-01 -2.16482610e-01 4.49827909e-01
-3.20923209e-01 4.61455107e-01 1.32588983e+00 -2.97563165e-01
-1.15345836e+00 -3.55257243e-02 -6.81252122e-01 6.22205772e-02
-2.10061729e-01 1.40687287e-01 -5.18297374e-01 4.99093235e-01
-6.70104325e-01 -6.41434968e-01 1.82970822e-01 1.60611287e-01
1.96966660e+00 -5.67624867e-01 -2.86954433e-01 8.63770962e-01
1.43168676e+00 -1.57804698e-01 4.43560958e-01 2.96711922e-01
8.14162254e-01 3.65982056e-01 7.68075883e-01 2.96506941e-01
9.42068636e-01 4.25279349e-01 2.98930377e-01 2.01353207e-01
2.34074309e-01 -6.22939229e-01 3.88832614e-02 5.32570004e-01
5.39854774e-03 1.46467984e-01 -1.52175248e+00 2.79463977e-01
-2.28678727e+00 -8.09273362e-01 -4.51447576e-01 1.56503391e+00
1.12851357e+00 5.76298758e-02 -6.80304989e-02 4.66773123e-01
3.50075066e-01 -2.97868073e-01 -6.62644744e-01 -3.43743503e-01
1.39464855e-01 1.90827236e-01 3.29187274e-01 6.72428668e-01
-1.05819190e+00 1.27235663e+00 5.28120279e+00 1.01380491e+00
-7.86975324e-01 -4.91210133e-01 4.18131590e-01 1.52332515e-01
-1.11575179e-01 3.68594378e-01 -7.03921616e-01 1.11868702e-01
5.62746108e-01 -5.39538264e-01 6.44457638e-01 1.24889529e+00
2.23901838e-01 -2.09183350e-01 -1.54966891e+00 1.03415787e+00
-3.23821634e-01 -1.73711514e+00 6.79115802e-02 -9.66171771e-02
4.47497457e-01 -4.01588142e-01 -3.41760337e-01 4.08229440e-01
5.74437976e-01 -1.26273775e+00 1.07991123e+00 4.62746710e-01
8.78320336e-01 -8.33348572e-01 1.60394356e-01 5.38988888e-01
-1.71907222e+00 1.40356183e-01 -1.23318784e-01 -5.26345372e-01
1.69238020e-02 3.23460668e-01 -7.78744221e-01 9.95926380e-01
6.43039942e-01 1.06841481e+00 -5.20958602e-01 6.68699443e-01
-7.21244335e-01 6.40182972e-01 -5.23228467e-01 -9.18623898e-03
3.20068628e-01 -2.11104155e-01 2.64033049e-01 9.47416246e-01
-1.12630084e-01 7.17463136e-01 4.81175363e-01 1.60102940e+00
-6.56588078e-02 -9.83429998e-02 -4.97840911e-01 -1.03910185e-01
5.54931819e-01 1.00247347e+00 -8.18102896e-01 -5.40799558e-01
-3.87502640e-01 6.04791760e-01 5.38581252e-01 3.95444036e-01
-1.21122038e+00 -6.66064680e-01 4.76907164e-01 5.15538007e-02
1.36645690e-01 -4.89562660e-01 -7.64852226e-01 -1.15091300e+00
2.38245293e-01 -7.63123214e-01 4.74960029e-01 -1.16047573e+00
-8.56207669e-01 -1.11186549e-01 5.31903207e-01 -6.71163142e-01
-1.43182250e-02 -8.93428922e-01 -6.01570845e-01 7.88710833e-01
-1.36338317e+00 -1.26456773e+00 -4.35893238e-01 6.34141624e-01
4.64345098e-01 3.09914857e-01 7.02551603e-01 1.53338730e-01
-7.38851666e-01 4.91888821e-01 -3.85322124e-01 5.26561797e-01
1.40465170e-01 -1.40803730e+00 5.48844874e-01 7.32984960e-01
-1.85959026e-01 6.70531034e-01 3.44426304e-01 -7.46735573e-01
-1.92624485e+00 -1.26905739e+00 7.56165206e-01 -5.49348295e-01
8.64889026e-01 -5.35023153e-01 -8.38376522e-01 9.62856829e-01
-3.47904235e-01 1.97481327e-02 3.22066069e-01 9.36836600e-02
-6.53200626e-01 1.02928832e-01 -9.46990550e-01 6.74986303e-01
1.47855067e+00 -4.24565583e-01 -7.45232344e-01 5.16710341e-01
1.16156447e+00 -8.72061312e-01 -8.31160426e-01 3.91438574e-01
3.59621674e-01 -5.01820683e-01 1.08311725e+00 -6.50596023e-01
9.87099767e-01 -5.34121513e-01 -2.30370820e-01 -6.25782430e-01
-8.97344351e-02 -5.38812757e-01 -4.35266316e-01 1.12095821e+00
4.54151630e-01 -6.65693700e-01 8.58743370e-01 9.05775487e-01
-2.74988979e-01 -1.12751818e+00 -4.84878987e-01 -8.28892648e-01
3.49045515e-01 -1.11372626e+00 1.10753024e+00 1.03380847e+00
3.32151085e-01 1.81955069e-01 1.59315288e-01 4.55882072e-01
6.62684679e-01 6.38090491e-01 1.06342518e+00 -1.18797052e+00
-3.00862402e-01 -3.27853888e-01 -2.04328448e-01 -1.20067608e+00
6.17681801e-01 -1.44940269e+00 -1.71430737e-01 -1.70060945e+00
-2.37762019e-01 -5.82958579e-01 4.41845536e-01 9.39873576e-01
1.46269456e-01 -3.75195265e-01 -2.03504384e-01 -2.13117346e-01
-9.67313468e-01 5.09359181e-01 1.22777581e+00 -1.32387921e-01
-1.04264438e-01 -4.02590454e-01 -6.57930434e-01 1.06980145e+00
8.76858175e-01 -2.41689563e-01 -4.75953609e-01 -6.47614300e-01
7.32401431e-01 3.17509204e-01 8.41318369e-01 -9.08288896e-01
4.91904587e-01 -6.33329093e-01 1.06378749e-01 -9.95079875e-01
-2.20720619e-02 -6.97616458e-01 -7.29278952e-04 2.97578931e-01
-2.52884626e-01 -3.86029780e-02 3.92889082e-01 3.76103401e-01
-3.16884257e-02 -4.91223559e-02 2.46978596e-01 -1.91112235e-01
-9.26022351e-01 4.35966939e-01 4.99846004e-02 4.96328473e-01
1.13473892e+00 -1.69978395e-01 -1.73282638e-01 2.99646109e-02
-4.29504126e-01 8.82210016e-01 3.08989495e-01 1.52708650e-01
6.76231563e-01 -1.21105015e+00 -4.85461175e-01 -1.09949723e-01
1.70355216e-01 1.05263448e+00 7.36370906e-02 6.99407339e-01
-1.07340395e+00 7.18488693e-01 3.49616766e-01 -4.37986761e-01
-1.10176063e+00 4.36701506e-01 3.58593255e-01 -1.59066528e-01
-9.29242432e-01 7.64324546e-01 2.71023452e-01 -9.00220156e-01
3.10796201e-01 -9.64642823e-01 3.93662214e-01 -7.96118736e-01
4.40734148e-01 4.63480741e-01 -1.85194746e-01 -1.59581289e-01
-5.47736347e-01 6.73566997e-01 9.25279781e-02 2.34040156e-01
1.30353224e+00 2.05581933e-01 -4.32695478e-01 2.41838515e-01
8.65749240e-01 -2.97659755e-01 -8.00980449e-01 -4.28123564e-01
1.72670662e-01 -2.79952735e-01 -2.02695429e-01 -7.75152266e-01
-6.37467265e-01 6.49703324e-01 -6.03340685e-01 -1.10428937e-01
1.01153910e+00 3.09997886e-01 7.46407628e-01 1.01772881e+00
5.11179328e-01 -1.01793778e+00 -1.40033647e-01 8.40755343e-01
9.97713923e-01 -9.93981600e-01 2.16365114e-01 -1.15341306e+00
-1.37713522e-01 1.24557078e+00 9.01226640e-01 -2.53038883e-01
5.93714178e-01 3.52574110e-01 -4.37451273e-01 -5.15664697e-01
-8.11095357e-01 2.70645678e-01 1.41147569e-01 2.41057366e-01
1.48160830e-01 2.39023790e-02 8.67572352e-02 1.05145001e+00
-9.12465870e-01 3.20327044e-01 1.84116483e-01 1.00141394e+00
-1.17811926e-01 -9.59279895e-01 -4.63093072e-01 2.40573451e-01
-3.37875709e-02 -1.89413995e-01 -5.17370403e-01 1.00219178e+00
2.06975117e-01 6.07293665e-01 -1.90763593e-01 -1.89507350e-01
1.26248479e-01 1.43126726e-01 6.30430579e-01 -7.55664289e-01
-1.82581171e-01 -6.03059053e-01 2.60198206e-01 -8.75643373e-01
6.56368285e-02 -5.78342199e-01 -2.10554790e+00 -5.55412173e-01
-1.46630317e-01 2.62999594e-01 2.71616280e-01 1.11587620e+00
3.16378683e-01 5.69993556e-01 1.37028456e-01 -4.53913033e-01
-3.13389271e-01 -2.62814522e-01 -3.09240639e-01 2.43379530e-02
-1.48658855e-02 -5.09265363e-01 -4.04258929e-02 2.85526514e-01] | [9.378515243530273, 7.431276798248291] |
df02f351-d91e-4dc7-9d01-8caaa78c5aab | training-strategies-for-neural-multilingual | null | null | https://aclanthology.org/2021.sigmorphon-1.26 | https://aclanthology.org/2021.sigmorphon-1.26.pdf | Training Strategies for Neural Multilingual Morphological Inflection | This paper presents the submission of team GUCLASP to SIGMORPHON 2021 Shared Task on Generalization in Morphological Inflection Generation. We develop a multilingual model for Morphological Inflection and primarily focus on improving the model by using various training strategies to improve accuracy and generalization across languages. | ['Jean-Philippe Bernardy', 'Adam Ek'] | null | null | null | null | acl-sigmorphon-2021-8 | ['morphological-inflection'] | ['natural-language-processing'] | [-1.94198802e-01 2.74519138e-02 -2.28342965e-01 -6.06963694e-01
-8.42595816e-01 -9.96919632e-01 4.82246667e-01 4.41863835e-01
-8.38428080e-01 6.52647853e-01 5.87033093e-01 -6.29060686e-01
1.11945346e-01 -5.82756579e-01 -4.91533905e-01 7.24214921e-03
-7.06969276e-02 5.77483296e-01 -1.74923331e-01 -6.18677378e-01
3.01912248e-01 2.84166276e-01 -7.62975752e-01 5.30413449e-01
9.55903471e-01 1.58750534e-01 1.64782524e-01 5.02126873e-01
-2.02864110e-01 2.82613218e-01 -1.01280892e+00 -8.15614343e-01
1.00739926e-01 -3.33894610e-01 -1.17865646e+00 -3.42665434e-01
8.48309994e-01 2.72198319e-01 3.93674940e-01 1.06783402e+00
5.46680868e-01 4.28415954e-01 7.49070227e-01 -6.88808858e-01
-1.34116161e+00 1.88794398e+00 -3.23967427e-01 8.64677608e-01
5.28651714e-01 8.28004852e-02 1.24749458e+00 -1.22476268e+00
8.31217647e-01 1.71700788e+00 9.75624323e-01 6.51739717e-01
-1.47036529e+00 -5.14369428e-01 4.97401804e-01 3.00744981e-01
-1.44447148e+00 -2.55079538e-01 5.29139400e-01 -2.96344995e-01
1.68852782e+00 2.46571749e-01 6.20768249e-01 9.28888500e-01
1.61923885e-01 8.53430629e-01 1.29677176e+00 -5.40972292e-01
-1.18034378e-01 -1.72832921e-01 4.44341719e-01 4.13449109e-01
8.30461085e-02 2.06528783e-01 -3.88723761e-01 3.31785902e-02
7.15840459e-01 -1.19009221e+00 -1.78467453e-01 5.82061648e-01
-1.33660865e+00 7.50253081e-01 5.55385351e-01 6.27836823e-01
-3.16913277e-01 -3.80874202e-02 5.90054870e-01 5.39155483e-01
5.49808443e-01 1.02011287e+00 -8.41486216e-01 2.87523091e-01
-9.61161375e-01 5.93151569e-01 5.37126899e-01 1.01374674e+00
4.33621734e-01 6.60799086e-01 -6.87263608e-02 1.30949914e+00
8.14039111e-02 3.65378559e-01 8.71994793e-01 -6.73928142e-01
5.13081074e-01 2.09567964e-01 -4.29116756e-01 -5.68953276e-01
-7.40135252e-01 -3.23695630e-01 -3.41156751e-01 -2.11986020e-01
3.11247468e-01 -1.41808897e-01 -1.19893157e+00 1.91549718e+00
3.69914435e-02 -6.44663334e-01 5.16387559e-02 4.73998457e-01
9.40112293e-01 6.67899311e-01 8.25250268e-01 -2.67572939e-01
1.05779278e+00 -7.76379704e-01 -8.24977696e-01 -3.15661043e-01
9.31489706e-01 -1.03264034e+00 1.40548575e+00 5.36840677e-01
-1.63113952e+00 -8.87938797e-01 -1.33656967e+00 -4.43644166e-01
-8.85341763e-01 -1.18486479e-01 8.05596471e-01 8.56540799e-01
-1.16259217e+00 4.49760705e-01 -6.16011202e-01 -4.03781444e-01
-1.54622406e-01 2.96435714e-01 -1.48415551e-01 3.60458404e-01
-1.55044317e+00 1.22454584e+00 1.14795053e+00 -8.98765028e-02
-3.20120901e-01 -1.20184350e+00 -1.09522605e+00 -7.51075208e-01
-4.42483783e-01 -7.66011178e-01 1.31547654e+00 -8.17151129e-01
-1.09592295e+00 1.16503298e+00 8.42006803e-02 -5.55024087e-01
1.47921899e-02 -1.59878537e-01 -7.10020304e-01 -4.57957685e-01
1.20096989e-01 1.22738540e+00 2.04799280e-01 -1.27313733e+00
-5.91025591e-01 -4.72955704e-01 -3.68490703e-02 3.16914141e-01
-1.62855253e-01 6.55645013e-01 4.75613713e-01 -1.38036740e+00
2.40069911e-01 -8.28353167e-01 -1.51113957e-01 -1.09262753e+00
-2.63158888e-01 -6.78988218e-01 2.28987992e-01 -1.09980869e+00
1.29535437e+00 -1.54497075e+00 3.71964097e-01 6.61561787e-02
-4.30997640e-01 -3.63076553e-02 -2.10830912e-01 1.99736863e-01
-4.06925231e-01 4.54997897e-01 -5.05623877e-01 -4.73567218e-01
1.34282365e-01 3.24376732e-01 -4.66421396e-01 8.05272013e-02
1.57866642e-01 1.35331857e+00 -9.41818833e-01 -5.02466023e-01
-2.58151838e-03 1.30583090e-03 -5.89841783e-01 -3.17620188e-01
-3.48641813e-01 4.61240172e-01 6.71957433e-01 5.94260454e-01
8.41544211e-01 8.86524677e-01 6.08038977e-02 -1.03029072e-01
-5.25600851e-01 8.04690480e-01 -7.94565439e-01 1.93014812e+00
-9.28029358e-01 5.22005446e-02 -1.96437255e-01 -2.53434569e-01
8.22990417e-01 2.67000973e-01 -1.77462384e-01 -3.22878331e-01
-4.26966101e-02 4.43066090e-01 3.26683700e-01 4.82397489e-02
1.10082316e+00 -6.38243079e-01 -5.82174361e-01 4.43878680e-01
4.94368911e-01 -9.91001129e-01 5.05977511e-01 -7.39225596e-02
1.72019929e-01 1.75877243e-01 6.21040881e-01 -1.05411816e+00
5.35440445e-01 1.07527755e-01 3.81768912e-01 3.27238739e-01
1.31472960e-01 3.72578174e-01 -3.96626115e-01 -4.56460834e-01
-7.81632006e-01 -1.69005585e+00 -5.11342347e-01 1.61851251e+00
-6.14759028e-01 -4.58314776e-01 -7.34776318e-01 -7.30337381e-01
-2.87975699e-01 1.60794675e+00 -5.48403323e-01 -1.00351505e-01
-1.19432569e+00 -1.26700902e+00 1.26246393e+00 8.61383080e-01
2.59662122e-01 -1.67455781e+00 7.46072456e-02 2.79057145e-01
-3.40532184e-01 -7.08939075e-01 -7.35489786e-01 1.34809732e-01
-9.23832595e-01 -4.18201059e-01 -3.56953412e-01 -1.37806940e+00
1.67317271e-01 -3.96181107e-01 1.49411118e+00 -1.36442170e-01
-1.83129475e-01 -8.66844505e-02 -2.54971117e-01 -8.65228891e-01
-7.29020536e-01 5.77897847e-01 1.84878126e-01 -7.14345872e-01
4.88878876e-01 -4.34980065e-01 -1.32562988e-03 -1.47243693e-01
-9.55852926e-01 -4.78277147e-01 9.68057513e-02 3.55428368e-01
6.32315278e-01 -4.17441070e-01 8.79079103e-01 -9.57588971e-01
1.14258146e+00 -5.16560733e-01 -2.27102757e-01 8.80292803e-02
-5.36378801e-01 -8.88799205e-02 5.41233420e-01 -3.87192607e-01
-1.30483544e+00 -2.10297912e-01 -7.54576087e-01 2.80848294e-01
-1.89065307e-01 7.51637340e-01 -3.63354117e-01 1.47862568e-01
8.96279395e-01 2.30789986e-02 -6.38469934e-01 -6.02730215e-01
1.03502893e+00 3.24728042e-01 8.65416884e-01 -9.88023520e-01
3.22148740e-01 -6.31259829e-02 -3.74549657e-01 -1.06963170e+00
-7.59945035e-01 -1.31464198e-01 -1.11162198e+00 1.44419864e-01
8.18237662e-01 -5.59025466e-01 1.74701199e-01 6.23091221e-01
-1.50608742e+00 -4.05345947e-01 -6.27776027e-01 2.36293495e-01
-7.22442150e-01 2.05929801e-01 -9.85544622e-01 -1.59210250e-01
-7.67448008e-01 -5.69255471e-01 8.05251956e-01 -1.74866289e-01
-8.58925104e-01 -1.69511819e+00 6.93432570e-01 -2.10274190e-01
5.33451915e-01 -3.74677666e-02 1.58334625e+00 -6.22341096e-01
8.73807222e-02 3.09025735e-01 1.85796276e-01 5.58112979e-01
2.88110852e-01 -1.35186911e-02 -4.92540479e-01 -2.03847453e-01
-2.06221398e-02 -4.69994605e-01 1.03426051e+00 4.57816690e-01
9.14383829e-01 -2.54264325e-01 -2.20716372e-02 9.84303057e-01
1.01700044e+00 -1.53744742e-01 4.66536105e-01 3.80932570e-01
7.96094418e-01 7.33655453e-01 4.16937947e-01 -5.26945516e-02
8.75742972e-01 4.44279253e-01 -1.27243698e-01 4.45967615e-02
-7.89791942e-01 -3.85141820e-01 4.84065920e-01 1.45871603e+00
6.65175635e-03 -1.47999465e-01 -1.10716617e+00 1.09156418e+00
-1.36696112e+00 -6.16421342e-01 -5.71689069e-01 2.00931716e+00
1.41337216e+00 -9.99303162e-02 3.25553626e-01 9.42255929e-03
6.41716123e-01 1.60040215e-01 7.99521133e-02 -1.23943603e+00
-7.95711517e-01 6.71341360e-01 2.96431452e-01 1.04876244e+00
-9.18019712e-01 1.89508843e+00 8.30146408e+00 7.81431556e-01
-7.44493902e-01 1.65209740e-01 1.48058206e-01 2.21209586e-01
-6.65933609e-01 -1.78954899e-01 -1.07912838e+00 -3.40814935e-03
1.13030243e+00 -5.88211596e-01 3.24141771e-01 2.08610922e-01
-1.81984082e-01 4.25458640e-01 -1.04120207e+00 4.00580198e-01
3.73581320e-01 -7.03930795e-01 6.63357496e-01 -3.90426755e-01
9.54473138e-01 1.26811832e-01 2.76321650e-01 4.87389833e-01
6.51788473e-01 -1.06882000e+00 8.19452107e-01 2.60544479e-01
7.49773085e-01 -1.12196219e+00 3.90450805e-01 -1.08198322e-01
-1.17096841e+00 2.78878957e-01 -5.90222299e-01 6.78558499e-02
4.78242785e-01 7.13960081e-02 -8.99356604e-01 5.16201794e-01
1.79154426e-01 5.17112017e-01 -1.09471881e+00 1.02516568e+00
-9.00670767e-01 8.02919984e-01 -1.75320134e-01 2.78269768e-01
1.25372157e-01 3.84235457e-02 8.52142870e-01 2.07457113e+00
1.24872401e-01 -2.12216333e-01 4.77915883e-01 6.46759868e-01
-2.21175067e-02 8.67414534e-01 -5.42756200e-01 3.16168427e-01
3.66227508e-01 1.03036904e+00 -5.00493526e-01 -3.54274660e-01
1.42353214e-02 9.51217055e-01 6.35779738e-01 6.80256858e-02
-6.06044590e-01 -3.10146302e-01 4.58437860e-01 -5.23304529e-02
-1.18396759e-01 -7.46714056e-01 -8.08603942e-01 -6.91904843e-01
-4.78284657e-01 -8.46080005e-01 8.71058345e-01 -4.41933393e-01
-1.70948637e+00 1.00500882e+00 4.83455926e-01 -5.31570673e-01
-4.08216149e-01 -7.61372089e-01 -4.51454222e-01 1.18726695e+00
-9.44116354e-01 -1.88235378e+00 4.94928062e-01 5.78791916e-01
7.99173415e-01 -2.36964785e-02 1.12809157e+00 1.51803233e-02
-2.63275653e-01 1.12161648e+00 -7.04151630e-01 -1.73184842e-01
9.13307190e-01 -1.94038749e+00 1.32984900e+00 9.92233932e-01
5.56014359e-01 8.79306793e-01 6.28038764e-01 -9.69185352e-01
-3.21994871e-01 -1.39217889e+00 1.42480958e+00 -7.22343445e-01
1.04350638e+00 -3.80488366e-01 -9.77115691e-01 1.24125266e+00
8.04201305e-01 -6.08694494e-01 7.99878716e-01 6.53605342e-01
-5.03063977e-01 6.10279024e-01 -1.12236285e+00 8.92479479e-01
1.40850568e+00 -3.66799712e-01 -1.51691473e+00 2.98522711e-01
1.12692785e+00 -3.14709753e-01 -9.16651845e-01 4.46837783e-01
1.30102292e-01 -1.94357932e-01 8.77439260e-01 -1.02566886e+00
1.57674968e-01 -1.38485640e-01 -2.24465236e-01 -2.30515122e+00
-6.04361713e-01 -5.93597472e-01 2.83550203e-01 1.33920372e+00
1.05371189e+00 -5.57477176e-01 2.24439114e-01 -6.10433482e-02
-6.78922296e-01 -2.97476590e-01 -9.60961223e-01 -1.13458085e+00
1.17631054e+00 -7.23219812e-01 6.43445373e-01 9.92578566e-01
3.74493986e-01 7.50391483e-01 3.67114782e-01 -1.09007746e-01
5.23845553e-01 -6.34430796e-02 -4.13180627e-02 -7.36156106e-01
-2.81788141e-01 -7.20244765e-01 -3.64065707e-01 -8.21144700e-01
6.92705393e-01 -1.79345357e+00 6.82643279e-02 -1.28257203e+00
-2.16261715e-01 -2.35142544e-01 -4.20722276e-01 6.82516515e-01
-4.75413889e-01 5.41074514e-01 1.94388002e-01 -2.33269528e-01
-1.81729883e-01 2.36950621e-01 8.67257893e-01 5.27697392e-02
-4.16105330e-01 -1.05136797e-01 -7.28219926e-01 8.28867435e-01
1.34201980e+00 -4.02092308e-01 -2.45910794e-01 -1.04145277e+00
2.87239432e-01 -5.18069625e-01 -3.62516195e-02 -7.29696453e-01
-2.32264549e-02 -3.13884020e-02 3.93304229e-01 -7.94881761e-01
1.36034891e-01 1.61415696e-01 -3.35639864e-01 3.66549879e-01
-3.93428206e-01 1.18145955e+00 9.14576173e-01 -2.32484236e-01
-8.20793584e-02 -4.26090628e-01 7.72313535e-01 -3.42506349e-01
-5.66708744e-01 8.62159729e-02 -6.06422663e-01 4.90999997e-01
5.54510295e-01 8.41652825e-02 -4.23193306e-01 2.18924239e-01
-1.05743611e+00 2.29120389e-01 4.90815908e-01 9.01854396e-01
6.11829579e-01 -1.37750852e+00 -1.39380527e+00 2.93654859e-01
1.77853689e-01 -6.98758721e-01 -2.85409093e-01 2.61450529e-01
-3.62864286e-01 4.50304717e-01 -4.16399747e-01 -4.79890406e-02
-1.08535397e+00 8.26517940e-01 4.28716272e-01 -1.10616505e-01
-5.55983707e-02 1.42018223e+00 -6.72926232e-02 -1.13169467e+00
-1.78505108e-01 -3.45901847e-01 -2.94361323e-01 2.01841921e-01
5.30852795e-01 6.88144743e-01 4.27118808e-01 -8.94375443e-01
-3.56158227e-01 4.68071908e-01 -4.67883855e-01 -8.19348633e-01
7.98672497e-01 -1.30982846e-01 -6.39015436e-01 9.73573327e-01
8.47057581e-01 5.31470716e-01 -2.70522207e-01 1.54769138e-01
7.19473064e-01 5.08695468e-02 4.92262421e-03 -1.33547795e+00
-6.01409137e-01 6.07811391e-01 1.03806019e-01 -1.15081064e-01
8.80210519e-01 1.64792284e-01 8.76520991e-01 1.76805943e-01
3.04073900e-01 -1.30390930e+00 -4.17522490e-01 1.23926270e+00
1.40566230e+00 -5.07348776e-01 -1.29524320e-01 -5.01207113e-01
-5.28222978e-01 1.06420290e+00 6.80858195e-01 -3.73577327e-01
6.47653759e-01 3.01342577e-01 4.64372396e-01 -2.01443285e-01
-5.73826432e-01 -2.16088429e-01 4.10745949e-01 8.22967172e-01
1.17811251e+00 5.65543234e-01 -1.14490902e+00 6.50640666e-01
-1.29530156e+00 -6.43600464e-01 3.69953990e-01 4.66547728e-01
-3.50469381e-01 -1.47861111e+00 -2.27762386e-01 1.58442661e-01
-6.66112363e-01 -9.07676697e-01 -8.38502824e-01 7.93176770e-01
5.76845288e-01 9.08750176e-01 1.25803754e-01 -6.42117411e-02
6.80361867e-01 3.79133105e-01 1.12023425e+00 -1.12771237e+00
-1.09687674e+00 -2.84401864e-01 4.98378754e-01 -1.63392171e-01
-6.71436861e-02 -1.19831669e+00 -1.41356170e+00 -1.10347141e-02
-1.09528285e-02 5.31866886e-02 6.32146358e-01 6.71182573e-01
-2.94055611e-01 4.65541720e-01 1.06098764e-01 -7.88753629e-01
-4.62169290e-01 -1.52771473e+00 -3.63424331e-01 2.38564387e-01
-1.44562036e-01 -4.87291589e-02 -3.92991513e-01 9.92819518e-02] | [10.731457710266113, 9.735614776611328] |
c101e9d7-1985-4728-890d-d4b66478d087 | augmenting-small-data-to-classify | null | null | https://aclanthology.org/2020.lrec-1.74 | https://aclanthology.org/2020.lrec-1.74.pdf | Augmenting Small Data to Classify Contextualized Dialogue Acts for Exploratory Visualization | Our goal is to develop an intelligent assistant to support users explore data via visualizations. We have collected a new corpus of conversations, CHICAGO-CRIME-VIS, geared towards supporting data visualization exploration, and we have annotated it for a variety of features, including contextualized dialogue acts. In this paper, we describe our strategies and their evaluation for dialogue act classification. We highlight how thinking aloud affects interpretation of dialogue acts in our setting and how to best capture that information. A key component of our strategy is data augmentation as applied to the training data, since our corpus is inherently small. We ran experiments with the Balanced Bagging Classifier (BAGC), Condiontal Random Field (CRF), and several Long Short Term Memory (LSTM) networks, and found that all of them improved compared to the baseline (e.g., without the data augmentation pipeline). CRF outperformed the other classification algorithms, with the LSTM networks showing modest improvement, even after obtaining a performance boost from domain-trained word embeddings. This result is of note because training a CRF is far less resource-intensive than training deep learning models, hence given a similar if not better performance, traditional methods may still be preferable in order to lower resource consumption. | ['Abhinav Kumar', 'Jillian Aurisano', 'Andrew Johnson', 'Barbara Di Eugenio'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['dialogue-act-classification'] | ['natural-language-processing'] | [ 1.22478463e-01 6.80703700e-01 1.44589208e-02 -5.77745199e-01
-3.22952569e-01 -5.24475932e-01 1.14512217e+00 3.96900952e-01
-7.61096597e-01 8.48892093e-01 9.77568388e-01 -7.33005345e-01
1.58398837e-01 -6.08552694e-01 5.03760669e-03 -4.00391847e-01
-1.43093139e-01 8.39741051e-01 -2.36374885e-01 -4.75238681e-01
4.36774492e-01 3.19912553e-01 -1.24228072e+00 3.73506397e-01
8.48681211e-01 5.98053813e-01 8.01899582e-02 9.13772285e-01
-5.14983594e-01 1.02747965e+00 -8.85797799e-01 -4.50851858e-01
-2.64541060e-01 -1.32522672e-01 -1.45840752e+00 -7.44222179e-02
2.51846701e-01 -6.04633033e-01 -1.66748136e-01 3.66194934e-01
3.46721381e-01 6.06974185e-01 6.43372178e-01 -9.96708453e-01
-6.20633125e-01 5.04064918e-01 1.53745282e-02 5.23718074e-02
5.54661691e-01 2.00087503e-01 1.00295174e+00 -8.07158530e-01
6.71081960e-01 1.53081214e+00 4.71442640e-01 9.45032656e-01
-1.47805667e+00 -1.64785892e-01 1.96196571e-01 -7.70883709e-02
-5.94866157e-01 -6.07039809e-01 4.92668033e-01 -4.04778361e-01
1.57289803e+00 3.17922503e-01 5.13320327e-01 1.50614238e+00
-2.72122353e-01 7.71766961e-01 1.27447486e+00 -8.66869032e-01
1.87119827e-01 6.04530334e-01 7.37387061e-01 6.46112204e-01
-3.34916890e-01 -8.33450779e-02 -5.46890080e-01 -4.22492981e-01
5.23405433e-01 -6.84641004e-02 -1.47689134e-01 5.81214540e-02
-9.20619190e-01 9.85847771e-01 3.38129073e-01 5.71720958e-01
-8.89626443e-02 -1.31403178e-01 5.64424098e-01 2.67779827e-01
8.20591450e-01 8.19401920e-01 -6.40818775e-01 -8.80827427e-01
-2.65612900e-01 2.47865364e-01 1.18580043e+00 7.57261574e-01
5.80399156e-01 -1.29198179e-01 -5.09838521e-01 1.16948533e+00
2.51261503e-01 -5.26873581e-02 5.05686522e-01 -8.45815599e-01
6.09216094e-01 6.59622729e-01 2.24629045e-01 -7.58235157e-01
-7.56713390e-01 1.28014907e-01 -7.66613781e-01 3.05899948e-01
8.84588659e-01 -4.26281124e-01 -7.81309247e-01 1.64078653e+00
3.78678776e-02 -4.08799857e-01 2.72098500e-02 5.52962601e-01
7.04017222e-01 5.43130398e-01 6.20179176e-01 -9.54156592e-02
1.42005050e+00 -7.93677092e-01 -1.08596385e+00 -4.92558628e-01
1.25199819e+00 -4.80474472e-01 1.44603539e+00 2.52135634e-01
-8.01490128e-01 -3.79020214e-01 -8.65421832e-01 -4.55648631e-01
-8.04508626e-01 -1.29712820e-01 8.47791076e-01 8.39901090e-01
-1.09184194e+00 6.63370669e-01 -7.23797858e-01 -5.90877473e-01
4.40833807e-01 2.08727464e-01 -4.09670800e-01 3.80456775e-01
-1.27552581e+00 1.51833415e+00 3.46760869e-01 -5.89611828e-02
-1.46881968e-01 -4.15394187e-01 -1.03668797e+00 1.23874255e-01
8.89497772e-02 -1.65260211e-01 1.48905504e+00 -4.64124233e-01
-1.63561988e+00 8.39751780e-01 -2.87653118e-01 -5.67076981e-01
2.93835253e-01 -4.41902280e-01 -1.54689431e-01 -2.41977960e-01
-3.07630152e-01 8.56058538e-01 3.10801357e-01 -1.09195364e+00
-4.50142205e-01 -3.44971061e-01 2.85583675e-01 3.12948614e-01
-7.88657725e-01 3.39904904e-01 -8.40304978e-03 -2.97199935e-01
-4.55104560e-01 -7.41961420e-01 -3.73690665e-01 -2.93191344e-01
-4.27023798e-01 -6.83077693e-01 8.13049912e-01 -9.33743298e-01
1.34982812e+00 -1.92318189e+00 -1.51419029e-01 -3.25053111e-02
3.57829422e-01 5.44377744e-01 1.25480413e-01 6.66756511e-01
3.62795517e-02 4.89016950e-01 -2.93670803e-01 -8.86031628e-01
9.46804509e-02 3.96864146e-01 -3.31751674e-01 9.14240181e-02
3.55449080e-01 7.85397351e-01 -8.01676333e-01 -5.51044106e-01
5.75320601e-01 4.97679204e-01 -4.60160017e-01 5.01132488e-01
-4.37023103e-01 5.33252299e-01 -8.35228711e-02 2.02228576e-01
5.09056747e-01 -2.11168990e-01 4.17079747e-01 2.13602126e-01
-2.58905143e-01 9.64992106e-01 -7.30055511e-01 1.51916814e+00
-9.14202034e-01 1.09570920e+00 3.28282826e-02 -7.73238659e-01
1.11354625e+00 4.50277478e-01 8.91968310e-02 -6.01660252e-01
-6.56696707e-02 -2.08258390e-01 2.28979811e-02 -6.50006235e-01
8.66218626e-01 2.12019365e-02 9.12531242e-02 8.64464343e-01
8.39632452e-02 1.46847323e-01 1.38439521e-01 3.99171978e-01
1.07942474e+00 3.27864103e-02 3.81343007e-01 -1.10852659e-01
1.66149348e-01 8.77310112e-02 1.28341824e-01 7.65592515e-01
-1.74063459e-01 2.33464196e-01 5.91545820e-01 -5.82934499e-01
-7.76474237e-01 -5.50891042e-01 -2.84277022e-01 1.64222252e+00
-4.51288491e-01 -7.02692747e-01 -7.00061858e-01 -9.26023245e-01
-2.42708087e-01 1.29941905e+00 -6.26587987e-01 1.90938786e-01
-9.30342913e-01 -7.25799203e-01 6.15545869e-01 7.19178081e-01
4.32129383e-01 -1.58086824e+00 -7.51797080e-01 2.14677989e-01
-1.25956371e-01 -8.02163720e-01 -3.81464399e-02 4.90938216e-01
-8.42414916e-01 -8.64908576e-01 -2.62098581e-01 -5.11945784e-01
4.65742052e-01 -1.02256350e-01 1.29593623e+00 2.84578532e-01
-6.49578348e-02 3.35268945e-01 -5.00403643e-01 -4.60067421e-01
-5.83463192e-01 1.69038624e-01 -2.29925990e-01 -5.04969895e-01
7.90504515e-01 -4.44395542e-01 -6.57183304e-02 -7.42393918e-03
-5.43039382e-01 2.86135882e-01 1.19899854e-01 1.11273801e+00
-5.94249785e-01 -5.10044336e-01 4.35569644e-01 -1.20899200e+00
1.27599943e+00 -3.10734212e-01 -4.23962995e-02 3.74631695e-02
-7.69274652e-01 6.95425794e-02 4.97931719e-01 -8.46959129e-02
-1.52354717e+00 -2.29634821e-01 -4.18164879e-01 3.09321374e-01
-7.18415201e-01 4.05918956e-01 -9.71146673e-02 4.67223883e-01
7.99404383e-01 -3.25637281e-01 1.73778713e-01 -9.16567981e-01
5.72046638e-01 1.16566956e+00 3.21620733e-01 -6.20913088e-01
2.04180673e-01 9.97079983e-02 -5.40315926e-01 -9.64615822e-01
-6.54552877e-01 -3.23841304e-01 -9.60321367e-01 -1.85378000e-01
1.08290946e+00 -4.00219202e-01 -1.01888192e+00 1.67089164e-01
-1.59170961e+00 -7.41526067e-01 -1.67081967e-01 2.06954747e-01
-4.20587748e-01 9.73553583e-02 -8.26666832e-01 -1.32689917e+00
-2.37429038e-01 -8.55394542e-01 6.98305488e-01 2.02217489e-01
-8.50664020e-01 -1.60607219e+00 1.13797598e-01 2.82457888e-01
6.70956075e-01 2.42515147e-01 1.14197290e+00 -1.37688315e+00
3.35557796e-02 6.12291917e-02 -3.39328110e-01 1.26486212e-01
2.25716367e-01 6.20145642e-04 -1.58660305e+00 -9.64659546e-03
-1.12612024e-02 -7.22580791e-01 7.67119408e-01 -6.42832220e-02
1.13730884e+00 -5.84320664e-01 -3.83262366e-01 1.18890461e-02
7.14658201e-01 4.38072979e-01 5.75426996e-01 4.80067372e-01
6.51385248e-01 1.06556892e+00 6.12039864e-01 5.05062938e-01
4.80045259e-01 8.58060122e-01 2.30020732e-01 -4.06783491e-01
9.86636505e-02 -2.73527950e-01 2.99541265e-01 3.33712816e-01
-2.05871183e-02 -3.79650891e-01 -1.21383560e+00 3.57801795e-01
-1.88392580e+00 -7.93803871e-01 -1.69739440e-01 1.91869223e+00
1.11254108e+00 1.85416609e-01 3.25131595e-01 1.16915867e-01
4.09925252e-01 4.22156721e-01 -1.04279846e-01 -1.06871450e+00
2.64178365e-01 3.45684767e-01 1.15405418e-01 9.15681541e-01
-1.12099957e+00 8.54146183e-01 6.69246960e+00 4.40227151e-01
-8.11068356e-01 8.72774869e-02 8.97334516e-01 1.27432197e-01
-1.55886188e-01 -1.59598008e-01 -6.91122949e-01 3.02111506e-01
1.33028018e+00 7.49614835e-02 3.06184947e-01 9.23684418e-01
2.88426697e-01 -2.39326090e-01 -1.38135302e+00 4.17850554e-01
-1.20900035e-01 -1.42965496e+00 -4.84114766e-01 1.80939168e-01
-1.19287565e-01 -8.97681639e-02 -2.44180903e-01 5.91380060e-01
7.75586903e-01 -1.39396203e+00 8.32596570e-02 3.33821476e-01
7.23302245e-01 -5.46356320e-01 6.76365077e-01 4.83155191e-01
-5.30328035e-01 1.49534987e-02 -1.05194822e-01 -5.30256212e-01
2.01644227e-01 2.52266318e-01 -1.42671466e+00 -1.51766688e-02
5.94969630e-01 6.59196019e-01 -6.65094018e-01 3.98848087e-01
-1.10395163e-01 7.00020015e-01 -2.04753429e-01 -3.04385662e-01
2.64289021e-01 -1.26723439e-01 3.65488440e-01 1.74799240e+00
-3.12823117e-01 1.44698888e-01 2.39338562e-01 7.32316673e-01
1.01639004e-02 -5.73383830e-03 -8.61272275e-01 -2.01017708e-01
5.22522092e-01 1.39742684e+00 -4.56593812e-01 -4.97559249e-01
-4.24602658e-01 5.97456872e-01 7.66884267e-01 1.75312147e-01
-3.23870033e-01 -2.65398264e-01 6.65747881e-01 1.45408334e-02
-2.63981104e-01 -4.14413750e-01 -8.00606966e-01 -8.58040929e-01
-2.90459096e-01 -7.97626317e-01 4.23617333e-01 -4.56104726e-01
-1.28598022e+00 8.43263030e-01 1.55093253e-01 -7.81301081e-01
-7.55804777e-01 -7.82971799e-01 -8.52382600e-01 1.09341776e+00
-1.03988945e+00 -1.04470873e+00 -2.94079989e-01 1.95510447e-01
7.10744262e-01 -8.11356306e-02 1.53698266e+00 1.99160688e-02
-5.03404617e-01 4.65319067e-01 -1.05273508e-01 4.08634484e-01
8.89934778e-01 -1.70825899e+00 6.44663334e-01 2.78090686e-01
1.96515486e-01 1.00785005e+00 8.75597715e-01 -5.58340549e-01
-7.65968800e-01 -6.64733529e-01 1.22694743e+00 -8.02362919e-01
6.05812192e-01 -7.80868411e-01 -1.20839286e+00 8.96088839e-01
9.46747899e-01 -4.89025384e-01 9.88346100e-01 9.04611826e-01
-2.47766793e-01 6.23409688e-01 -1.17929339e+00 6.26021445e-01
9.38229144e-01 -5.11878073e-01 -9.17701662e-01 4.89371777e-01
7.12562084e-01 -4.15530175e-01 -8.08196008e-01 -7.83619806e-02
3.90043825e-01 -1.13229299e+00 6.65262878e-01 -1.03959978e+00
4.87598926e-01 3.71620506e-01 1.70468897e-01 -1.55787516e+00
-1.12849616e-01 -6.93805635e-01 -8.22095051e-02 1.60121429e+00
5.23027778e-01 -7.42678940e-01 8.15596461e-01 1.56394315e+00
-2.21847013e-01 -6.60233259e-01 -7.13968039e-01 -3.72189164e-01
3.29463303e-01 -5.05052686e-01 4.08222735e-01 1.03673685e+00
7.82903552e-01 6.37928903e-01 -5.90423882e-01 -6.17355943e-01
-5.68286106e-02 -2.41044685e-01 9.36489105e-01 -1.24811792e+00
-1.67599216e-01 -3.61627787e-01 1.30422696e-01 -1.19390190e+00
3.50213945e-01 -7.44931400e-01 9.98573825e-02 -1.58575535e+00
-1.26496270e-01 -7.31804609e-01 1.30609706e-01 9.25562561e-01
-3.64599735e-01 -2.11980447e-01 2.96509027e-01 1.05372213e-01
-3.41833293e-01 5.87678313e-01 8.74833047e-01 -1.50716696e-02
-6.07277751e-01 7.92094022e-02 -7.47319400e-01 8.50510240e-01
9.59547639e-01 -2.44317740e-01 -1.64497569e-01 -3.01221013e-01
-3.57789963e-01 -5.44902273e-02 2.06847653e-01 -5.86681664e-01
2.29332045e-01 -9.00561139e-02 4.09627527e-01 -4.77788776e-01
6.75248325e-01 -5.83308458e-01 -6.19044602e-01 1.33848816e-01
-1.02139270e+00 5.31525090e-02 4.08567041e-01 1.99042127e-01
-2.53505744e-02 -3.35531861e-01 5.56857705e-01 -2.28144638e-02
-5.18210411e-01 -4.26706105e-01 -8.96018326e-01 1.08314879e-01
5.08370757e-01 -5.75699918e-02 -7.44584799e-01 -6.50581241e-01
-7.29506612e-01 2.77026027e-01 2.25114033e-01 5.04730463e-01
4.01199162e-01 -1.14210463e+00 -3.96052629e-01 3.94822687e-01
6.12710565e-02 -7.38260755e-03 -1.14782497e-01 5.58356345e-01
-2.23871410e-01 6.21207476e-01 -2.62158275e-01 -3.55589420e-01
-1.44361877e+00 2.05228597e-01 9.34322700e-02 -6.05636120e-01
-6.91229105e-01 8.51893187e-01 -1.22075930e-01 -8.48273039e-01
6.06269002e-01 -2.24337623e-01 -8.05925786e-01 2.62261420e-01
6.34944439e-01 6.03854120e-01 -1.66540537e-02 -4.12579805e-01
-2.39080444e-01 -2.30405703e-01 -4.39045370e-01 -3.74233782e-01
1.47461987e+00 -3.96999083e-02 -2.02865452e-02 6.63208127e-01
1.05339479e+00 -1.37273178e-01 -9.85720336e-01 -2.07952783e-01
4.99388814e-01 -5.76172709e-01 -1.08823493e-01 -1.02958143e+00
-3.80838156e-01 1.19488966e+00 3.82041276e-01 7.96267569e-01
5.92482090e-01 -1.41676739e-01 4.77245599e-01 7.18474746e-01
5.68614602e-02 -1.24359202e+00 -3.58901434e-02 8.70873690e-01
1.00215220e+00 -1.48410177e+00 -1.33183375e-01 -1.90600291e-01
-9.82118130e-01 1.42662847e+00 1.01966965e+00 1.79499313e-01
6.21608794e-01 3.79651666e-01 2.45313734e-01 -3.96682739e-01
-1.00728095e+00 -1.77984208e-01 -8.69205296e-02 8.26189578e-01
9.65409279e-01 -9.91894528e-02 -1.24684133e-01 3.02338511e-01
-3.25082093e-01 -1.19607911e-01 5.05164921e-01 9.10965323e-01
-3.56153816e-01 -1.38681459e+00 -2.04775453e-01 6.26446187e-01
-2.32642129e-01 -1.84886694e-01 -7.25742519e-01 1.00910103e+00
-2.26565197e-01 1.38660324e+00 3.97415221e-01 -4.74350154e-01
2.61647075e-01 5.11466324e-01 -6.47331402e-02 -9.15644765e-01
-9.63349164e-01 -2.66296089e-01 8.59009981e-01 -4.25999969e-01
-3.10139239e-01 -5.22030413e-01 -1.32839119e+00 -4.01885986e-01
-2.65562654e-01 2.69756138e-01 6.30209684e-01 1.15882790e+00
1.94298625e-01 3.91849279e-01 4.11901832e-01 -8.82350385e-01
-4.92657512e-01 -1.62925935e+00 -1.96980730e-01 5.02099752e-01
2.50086755e-01 -6.48840666e-01 -3.60892922e-01 -7.30497241e-02] | [12.726888656616211, 7.763911247253418] |
91367a14-84dc-458d-b57c-6cdb034f10c7 | deep-neural-networks-in-video-human-action | 2305.15692 | null | https://arxiv.org/abs/2305.15692v1 | https://arxiv.org/pdf/2305.15692v1.pdf | Deep Neural Networks in Video Human Action Recognition: A Review | Currently, video behavior recognition is one of the most foundational tasks of computer vision. The 2D neural networks of deep learning are built for recognizing pixel-level information such as images with RGB, RGB-D, or optical flow formats, with the current increasingly wide usage of surveillance video and more tasks related to human action recognition. There are increasing tasks requiring temporal information for frames dependency analysis. The researchers have widely studied video-based recognition rather than image-based(pixel-based) only to extract more informative elements from geometry tasks. Our current related research addresses multiple novel proposed research works and compares their advantages and disadvantages between the derived deep learning frameworks rather than machine learning frameworks. The comparison happened between existing frameworks and datasets, which are video format data only. Due to the specific properties of human actions and the increasingly wide usage of deep neural networks, we collected all research works within the last three years between 2020 to 2022. In our article, the performance of deep neural networks surpassed most of the techniques in the feature learning and extraction tasks, especially video action recognition. | ['Yifan Zheng', 'Zhi Liu', 'Yang Yang', 'Zihan Wang'] | 2023-05-25 | null | null | null | null | ['action-recognition-in-videos', 'action-recognition'] | ['computer-vision', 'computer-vision'] | [ 3.06795478e-01 -4.89025772e-01 -2.52785593e-01 -3.90196472e-01
6.51010722e-02 -1.84008479e-01 4.15677965e-01 -5.82653940e-01
-6.74183011e-01 6.11733317e-01 2.59610593e-01 -5.73015399e-02
-1.23016804e-01 -8.18089902e-01 -5.35695255e-01 -6.75042331e-01
-2.64807284e-01 -3.47478032e-01 3.37252825e-01 -9.45850685e-02
3.84770840e-01 7.66047895e-01 -1.94081247e+00 4.89149034e-01
1.31766155e-01 1.53480828e+00 -2.03538731e-01 9.94027674e-01
-1.11006416e-01 1.38492405e+00 -4.06762004e-01 -2.29580969e-01
5.90586364e-01 -1.90346777e-01 -5.85231781e-01 2.52501994e-01
7.58707464e-01 -1.02885890e+00 -1.13583958e+00 8.12266231e-01
2.86541045e-01 2.27078989e-01 3.84366333e-01 -1.45183909e+00
-7.63635814e-01 -4.97052334e-02 -4.57879931e-01 7.94529796e-01
3.72381538e-01 2.30895206e-01 3.10623944e-01 -6.90583408e-01
5.21756709e-01 1.25360382e+00 6.86015427e-01 5.27058005e-01
-1.45707518e-01 -4.45225298e-01 8.63114595e-02 1.00317252e+00
-9.70744550e-01 -4.09562290e-01 8.03359866e-01 -5.37245274e-01
1.30743575e+00 8.46598744e-02 9.87444520e-01 1.36667740e+00
2.83247620e-01 9.35563326e-01 8.31618428e-01 -1.87975869e-01
-9.00220573e-02 -2.28690684e-01 2.05539837e-01 7.98088372e-01
2.81954497e-01 2.43367761e-01 -5.82938075e-01 2.70195395e-01
1.09643579e+00 5.60877204e-01 -2.52844274e-01 -1.83271840e-01
-1.17316890e+00 4.72824305e-01 2.27388754e-01 4.91850615e-01
-4.46209818e-01 3.14752758e-01 6.30502641e-01 3.60206038e-01
1.38167858e-01 -1.08379856e-01 -5.68686187e-01 -8.31043959e-01
-9.88943756e-01 -1.10547118e-01 4.19969529e-01 9.71496820e-01
6.38225019e-01 4.21570063e-01 -2.93229166e-02 4.87143129e-01
1.42569542e-01 4.65086341e-01 7.45398819e-01 -1.32553983e+00
3.95873338e-01 8.25414956e-01 -6.11929148e-02 -1.42614210e+00
-5.28400898e-01 1.45725831e-02 -1.01735961e+00 2.41348863e-01
4.51275229e-01 -2.03899622e-01 -8.18711996e-01 1.26689982e+00
1.29831642e-01 3.92041177e-01 -1.90919579e-03 9.09603715e-01
9.42096949e-01 6.13165736e-01 -2.32020002e-02 -3.68071318e-01
1.03576195e+00 -9.59301174e-01 -7.88862586e-01 5.87745123e-02
6.16130531e-01 -3.56931925e-01 5.42604387e-01 5.60793936e-01
-8.45251501e-01 -9.81351018e-01 -1.01164436e+00 -1.96290433e-01
-5.86278379e-01 2.03199789e-01 7.95804203e-01 8.22846115e-01
-1.12744939e+00 5.96830010e-01 -1.01648891e+00 -4.82554317e-01
6.67465627e-01 3.48337203e-01 -6.67751610e-01 -5.72649017e-02
-1.23265123e+00 7.49589741e-01 4.73910481e-01 4.66730744e-01
-9.02435958e-01 -3.88668925e-01 -7.51095414e-01 -3.00308853e-01
3.44560534e-01 -4.30846900e-01 9.85501647e-01 -1.31342864e+00
-1.49515748e+00 6.52888954e-01 1.07841559e-01 -5.48853219e-01
5.50220370e-01 -5.01685977e-01 -4.35951293e-01 4.80696142e-01
-4.22017783e-01 5.67209184e-01 8.65098000e-01 -4.42865402e-01
-8.87421846e-01 -6.27179325e-01 4.17561322e-01 1.26513317e-01
-7.55947053e-01 3.00730523e-02 -4.10919458e-01 -3.23492974e-01
-1.38087362e-01 -8.14728618e-01 2.49570087e-02 2.65155375e-01
2.61620227e-02 -1.76621825e-01 1.49562454e+00 -7.89119065e-01
1.12726104e+00 -2.05784011e+00 6.27685487e-02 -3.80288452e-01
1.39157355e-01 6.42771184e-01 2.02734619e-02 1.22713841e-01
-2.01806445e-02 -1.11169904e-01 3.02416906e-02 2.15638950e-01
-2.94128686e-01 2.40484506e-01 8.41355696e-02 6.24710619e-01
6.36371002e-02 7.88956761e-01 -7.39453495e-01 -6.24557078e-01
8.70213628e-01 6.03078187e-01 -2.60541350e-01 8.39825645e-02
1.33055255e-01 3.45624954e-01 -7.69255459e-01 9.67847347e-01
6.59961104e-01 8.81946459e-02 -1.31536528e-01 -6.70414209e-01
-1.78875044e-01 -2.88159996e-01 -1.05077088e+00 1.79679692e+00
-2.44450212e-01 1.04716396e+00 -9.26831216e-02 -1.48892832e+00
7.14493155e-01 3.73209178e-01 1.23133361e+00 -8.86198759e-01
2.78753221e-01 -3.80389541e-02 -1.14465177e-01 -1.27330172e+00
3.90448540e-01 5.37696004e-01 3.16273749e-01 1.06032953e-01
1.40401527e-01 4.23424721e-01 2.11879089e-01 -2.43140846e-01
1.36305249e+00 6.68078303e-01 3.77206981e-01 6.87709004e-02
8.17050219e-01 1.65928323e-02 6.69749379e-01 4.71410304e-01
-9.73935723e-01 4.60734367e-01 2.45804742e-01 -1.25780356e+00
-9.93991256e-01 -6.06684744e-01 -3.34734768e-02 7.81181216e-01
1.08774707e-01 -1.70216307e-01 -6.75134420e-01 -6.73411489e-01
-1.12270519e-01 1.17078535e-01 -6.64997220e-01 -2.26179749e-01
-9.93139267e-01 -6.46043241e-01 8.22619855e-01 8.34511578e-01
1.35191095e+00 -1.20825291e+00 -1.26326001e+00 1.39899179e-01
7.44322687e-02 -1.48207080e+00 1.13281652e-01 -4.34523880e-01
-1.03231025e+00 -1.56159258e+00 -7.56794274e-01 -5.41064918e-01
2.38969892e-01 4.82485980e-01 8.30753028e-01 1.66905858e-02
-5.09247959e-01 8.82191718e-01 -6.34736240e-01 -1.73378080e-01
1.25039443e-01 -2.89416164e-01 1.84001520e-01 1.67454556e-01
8.68303001e-01 -5.19473076e-01 -8.08739603e-01 1.61681190e-01
-1.05226398e+00 -1.01032272e-01 6.83921576e-01 2.91352659e-01
1.77241787e-01 3.76062770e-03 7.87966400e-02 -1.89995915e-01
2.23109558e-01 -2.93582886e-01 -3.73725921e-01 3.01198661e-01
-4.08161357e-02 -2.18512088e-01 4.62774992e-01 -2.49275327e-01
-1.03549731e+00 1.75180167e-01 -2.74871252e-02 -6.70370758e-01
-6.16623044e-01 1.20701008e-01 -3.18684317e-02 -1.14444874e-01
3.62715274e-01 4.08898622e-01 9.83229876e-02 -1.72401667e-01
1.26514822e-01 8.56453240e-01 4.18049991e-01 -2.70692110e-01
3.08507144e-01 6.81919158e-01 2.26500884e-01 -1.22837198e+00
-5.31609654e-01 -2.37494126e-01 -1.08054125e+00 -8.34405601e-01
1.27698505e+00 -6.71383023e-01 -8.39002788e-01 1.14437985e+00
-1.23098564e+00 7.28154853e-02 -1.06486835e-01 7.89471209e-01
-6.18065596e-01 6.30577445e-01 -6.66074991e-01 -9.27454412e-01
-2.17838600e-01 -1.31448412e+00 9.21039402e-01 3.84106636e-01
2.49385059e-01 -8.98442209e-01 -2.42224671e-02 4.52295780e-01
4.58061367e-01 6.21277273e-01 4.20905143e-01 -3.17413896e-01
-7.18125284e-01 -2.37652227e-01 -3.29756230e-01 6.78934872e-01
3.43790799e-01 4.39770401e-01 -9.72195148e-01 1.11943498e-01
2.76835471e-01 -2.71036983e-01 8.44874322e-01 6.62110984e-01
1.43162632e+00 -3.71298604e-02 -1.13640174e-01 7.71290481e-01
1.33929217e+00 7.92116344e-01 1.03300178e+00 6.64326966e-01
9.43768561e-01 5.52351594e-01 4.78876293e-01 4.52964008e-01
2.54321098e-01 4.98938531e-01 5.96955895e-01 1.79817304e-01
-3.86093222e-02 2.35098958e-01 6.42742038e-01 6.96296453e-01
-7.39703119e-01 -7.50860125e-02 -8.81769717e-01 2.49706894e-01
-1.97236049e+00 -1.37141180e+00 -9.62150469e-02 1.88975906e+00
2.01291665e-01 -4.73664179e-02 8.01390931e-02 2.23183766e-01
7.01900363e-01 3.28839689e-01 -6.75891757e-01 -3.39924127e-01
-7.25101009e-02 -7.29463398e-02 6.24016166e-01 -1.76746055e-01
-1.56304061e+00 5.89460015e-01 6.45645428e+00 4.27426606e-01
-1.27966094e+00 -1.94102302e-01 7.13189781e-01 -1.56549752e-01
5.28757751e-01 -4.44491893e-01 -6.39875233e-01 4.24656749e-01
1.13098514e+00 1.76814929e-01 1.45261288e-01 9.70575154e-01
4.58522052e-01 -1.85995921e-01 -1.11052871e+00 1.45995057e+00
3.13719660e-01 -1.28135180e+00 2.05169410e-01 6.90699592e-02
5.14454424e-01 -1.25259370e-01 -7.63345659e-02 2.83294439e-01
-3.10915321e-01 -9.69565094e-01 4.89964753e-01 9.76202190e-01
5.57347178e-01 -5.99904656e-01 9.00494516e-01 9.77634639e-02
-1.12009442e+00 -4.57742929e-01 -4.44067687e-01 -3.66252750e-01
1.07345834e-01 4.18140858e-01 -6.47091568e-02 5.96556723e-01
1.21068740e+00 1.54192209e+00 -4.97844577e-01 8.84826958e-01
1.77752540e-01 3.92881691e-01 -1.90617383e-01 2.26148870e-03
4.77648288e-01 -3.32387865e-01 3.31772983e-01 1.07146060e+00
4.34935898e-01 1.27557650e-01 1.48671240e-01 3.52762669e-01
8.23595449e-02 -2.39597112e-01 -9.22451615e-01 -1.40551180e-01
-1.52642988e-02 1.18575823e+00 -4.89799768e-01 -4.73801285e-01
-9.41231072e-01 8.58534098e-01 -1.48953408e-01 2.55462468e-01
-1.07161880e+00 -1.28302008e-01 7.67671525e-01 -6.14009053e-02
3.12499046e-01 -6.78138494e-01 1.21187963e-01 -1.13114214e+00
1.17973812e-01 -7.46970177e-01 3.39944810e-01 -8.14706028e-01
-9.90887165e-01 5.08079648e-01 2.90692776e-01 -1.51068985e+00
-2.89111465e-01 -1.31061244e+00 -3.86792183e-01 1.72009990e-01
-1.30499864e+00 -9.56796646e-01 -6.11733556e-01 8.86157453e-01
8.26836765e-01 -5.34533143e-01 5.21142185e-01 5.59771359e-01
-8.65557909e-01 4.94001806e-02 5.65981567e-02 6.83976054e-01
3.88038367e-01 -7.28493929e-01 5.55960611e-02 9.93713915e-01
2.08179913e-02 2.37916857e-01 1.54252082e-01 -3.27969015e-01
-1.85339975e+00 -9.35100436e-01 2.13596553e-01 -3.51448119e-01
4.18276250e-01 4.56768870e-02 -4.49177533e-01 6.86712027e-01
2.30985656e-01 3.46705168e-01 5.17756462e-01 -5.69700420e-01
-6.89358562e-02 -4.77551877e-01 -1.01515615e+00 2.57959843e-01
1.27651048e+00 -4.21161979e-01 -5.16156137e-01 1.53093770e-01
3.60333025e-01 -2.16866881e-01 -8.28422189e-01 6.48387969e-01
8.40653300e-01 -1.47702754e+00 1.11811078e+00 -7.92734385e-01
5.74842632e-01 -2.72333145e-01 -4.23856199e-01 -6.31640792e-01
-1.83188975e-01 -1.95704907e-01 -5.88943243e-01 7.11401761e-01
-2.52981097e-01 -2.08182588e-01 9.88775313e-01 7.48230278e-01
-1.25381008e-01 -6.27570748e-01 -1.03655446e+00 -5.97375274e-01
-2.61529237e-01 -5.56234777e-01 2.21624449e-01 7.06776142e-01
-5.32571077e-01 -3.73290867e-01 -5.71821451e-01 -1.40539780e-01
4.45371509e-01 -1.67028919e-01 5.98060369e-01 -9.01074886e-01
-4.56076581e-03 -4.23778147e-01 -1.34254086e+00 -9.99186575e-01
-8.64601210e-02 -2.13675126e-01 -4.46853220e-01 -1.50137413e+00
4.77791019e-02 2.94025421e-01 -5.21963120e-01 3.66327792e-01
2.66915798e-01 2.02613652e-01 2.49629036e-01 1.33789062e-01
-7.61550963e-01 4.69246924e-01 1.30161142e+00 -3.45776558e-01
4.46836725e-02 -1.12801298e-01 -1.49649635e-01 9.72070098e-01
6.99824035e-01 1.07564561e-01 -5.17895997e-01 -6.95095778e-01
-1.00670122e-01 1.35194426e-02 5.54503858e-01 -1.54898727e+00
1.44481972e-01 -2.93675214e-01 8.14950407e-01 -6.32355809e-01
3.81209522e-01 -1.12137020e+00 4.47522663e-02 5.79153836e-01
-1.62194267e-01 2.70388633e-01 1.42375708e-01 4.53552186e-01
-4.78806376e-01 2.01467704e-02 4.99347627e-01 -4.97872740e-01
-1.69288743e+00 6.52446628e-01 -4.87859696e-01 -1.42796427e-01
1.37143219e+00 -9.38986123e-01 -2.21346617e-01 -3.73466879e-01
-6.28363490e-01 -4.17554319e-01 5.53334542e-02 7.08305895e-01
8.44373763e-01 -1.40960240e+00 -4.51466143e-01 1.53119609e-01
-2.50065207e-01 -2.05922380e-01 4.79581743e-01 9.03499901e-01
-9.01401401e-01 6.99488997e-01 -1.02559566e+00 -6.47980511e-01
-1.15726411e+00 6.76151574e-01 5.32069862e-01 4.87622283e-02
-6.21541739e-01 5.37377119e-01 1.41752344e-02 2.02077493e-01
3.40419650e-01 -5.21404862e-01 -5.85515201e-01 -4.81173582e-02
8.34939361e-01 8.44078422e-01 5.80041222e-02 -7.95536578e-01
-4.93269742e-01 6.53512120e-01 8.64971131e-02 2.89307505e-01
1.51476634e+00 -1.38902422e-02 5.22259530e-03 3.08422267e-01
1.42329454e+00 -7.87248850e-01 -1.58081400e+00 1.25271380e-01
-1.41067997e-01 -5.94563246e-01 2.09679589e-01 -1.96847171e-01
-1.61622214e+00 1.07659137e+00 9.76395726e-01 3.61639529e-01
1.38199329e+00 -6.49839401e-01 9.38122809e-01 6.10891223e-01
4.98229980e-01 -1.39709795e+00 1.84358925e-01 5.20248115e-01
6.86686635e-01 -1.34661329e+00 1.72227591e-01 -6.47133887e-02
-3.65927279e-01 1.73403931e+00 8.59005272e-01 -2.29060605e-01
8.16123664e-01 2.52653003e-01 -4.55635414e-02 -1.58477321e-01
-4.00562406e-01 -6.23843446e-02 1.88351244e-01 8.84236693e-01
5.56791127e-01 -3.31368774e-01 -8.20773318e-02 1.51823550e-01
2.68462300e-01 4.09597844e-01 4.06947970e-01 1.21934855e+00
-4.43162829e-01 -8.09904873e-01 -2.85231531e-01 5.03094375e-01
-3.62965584e-01 2.42032826e-01 -2.64029771e-01 9.87015724e-01
4.07947183e-01 8.15982223e-01 1.64763883e-01 -6.72101140e-01
3.45212996e-01 1.06347445e-02 5.90400219e-01 1.36274964e-01
-1.98643833e-01 -5.84328651e-01 -6.82079569e-02 -1.14128375e+00
-1.34133804e+00 -7.27276802e-01 -9.59289014e-01 -3.16963822e-01
2.52254725e-01 -5.73256671e-01 6.03726447e-01 1.11272395e+00
3.55952889e-01 4.54825431e-01 4.84636247e-01 -1.00450563e+00
-2.04636186e-01 -8.13960195e-01 -3.89614284e-01 6.39819801e-01
3.58393937e-01 -8.46671164e-01 -6.33735284e-02 4.00327325e-01] | [7.954170227050781, 0.5038995742797852] |
593540e9-8f0c-499c-b279-c5ddbb3a8d10 | modeling-sense-structure-in-word-usage-graphs | null | null | https://aclanthology.org/2021.starsem-1.23 | https://aclanthology.org/2021.starsem-1.23.pdf | Modeling Sense Structure in Word Usage Graphs with the Weighted Stochastic Block Model | We suggest to model human-annotated Word Usage Graphs capturing fine-grained semantic proximity distinctions between word uses with a Bayesian formulation of the Weighted Stochastic Block Model, a generative model for random graphs popular in biology, physics and social sciences. By providing a probabilistic model of graded word meaning we aim to approach the slippery and yet widely used notion of word sense in a novel way. The proposed framework enables us to rigorously compare models of word senses with respect to their fit to the data. We perform extensive experiments and select the empirically most adequate model. | ['Sabine Schulte im Walde', 'Jonas Kuhn', 'Enrique Castaneda', 'Dominik Schlechtweg'] | 2021-08-01 | null | null | null | joint-conference-on-lexical-and-computational-1 | ['stochastic-block-model'] | ['graphs'] | [ 1.43744692e-01 2.39163920e-01 -3.99904788e-01 -3.34709585e-01
4.45429049e-02 -6.67346895e-01 1.06491756e+00 5.76730311e-01
-8.35787237e-01 6.12708211e-01 6.06097043e-01 -4.95689541e-01
-4.44487303e-01 -9.82283175e-01 -5.60513176e-02 -4.55874026e-01
1.04219861e-01 3.14720303e-01 3.20463657e-01 -4.17833388e-01
5.74995041e-01 2.58565217e-01 -1.40884769e+00 -1.24977782e-01
7.78246820e-01 3.26077938e-01 3.06391209e-01 4.65227753e-01
-3.85851622e-01 3.46511573e-01 -4.12341624e-01 -5.88781476e-01
-1.62745729e-01 -2.54460543e-01 -1.04845989e+00 -3.32077086e-01
1.68619663e-01 6.46654963e-01 -1.54623345e-01 1.26775157e+00
2.12261751e-01 2.49812141e-01 1.20420051e+00 -9.64451432e-01
-9.15066540e-01 8.78230512e-01 -5.15025258e-01 4.70362604e-01
6.87746048e-01 -2.39706203e-01 1.43622720e+00 -6.76321983e-01
6.89866066e-01 1.60501289e+00 5.75957417e-01 3.47533613e-01
-1.36163270e+00 -2.13097438e-01 2.57886916e-01 2.33530253e-01
-1.71390808e+00 1.54795215e-01 6.31609142e-01 -4.83915538e-01
9.47073579e-01 2.60776341e-01 9.07573879e-01 1.31646979e+00
4.87267703e-01 3.80818903e-01 1.36920524e+00 -6.57942176e-01
4.78599459e-01 -2.23857716e-01 7.33038485e-01 6.84634626e-01
9.03793633e-01 -2.36580580e-01 -4.89961475e-01 -7.64119804e-01
3.73893797e-01 1.75190613e-01 -1.41396686e-01 -2.72236049e-01
-9.57803667e-01 9.87726629e-01 3.26131552e-01 9.67605412e-01
-4.81791228e-01 3.85844618e-01 1.14589199e-01 -1.39762759e-01
6.68769896e-01 3.06064874e-01 -3.04705828e-01 -6.45694963e-04
-8.66778195e-01 3.93074274e-01 8.78754914e-01 6.86647475e-01
8.09594810e-01 -2.51407593e-01 -7.15002865e-02 8.09528291e-01
9.50318813e-01 4.62579757e-01 5.87700844e-01 -2.23780230e-01
-2.66928405e-01 5.18816352e-01 3.71010415e-02 -1.25534046e+00
-3.72306347e-01 -2.48956844e-01 -6.27873600e-01 -3.02292615e-01
3.19416791e-01 3.53198618e-01 -8.72652471e-01 1.78704619e+00
1.90665200e-01 4.68528211e-01 -2.66502053e-01 4.35406953e-01
6.16068423e-01 1.53025135e-01 9.13308263e-01 -9.22353789e-02
1.66900361e+00 1.19261429e-01 -6.65932000e-01 -3.08410097e-02
5.04189789e-01 -6.24796450e-01 1.26204360e+00 1.39471099e-01
-5.70065022e-01 -1.70903921e-01 -9.50931251e-01 -3.15854289e-02
-6.99682534e-01 -6.17053092e-01 6.49613261e-01 9.25622106e-01
-1.23988903e+00 6.31616712e-01 -5.44392169e-01 -7.93702960e-01
3.34288985e-01 -1.43366784e-01 -2.41807014e-01 -1.06243253e-01
-1.50152683e+00 1.21337962e+00 7.22391725e-01 -6.64728731e-02
-4.20910865e-01 -4.45911407e-01 -1.00302148e+00 -7.53267258e-02
2.21463814e-01 -1.17217851e+00 9.39996243e-01 -5.23763239e-01
-1.01127088e+00 1.11882508e+00 -3.18923324e-01 -4.50143069e-01
2.20428351e-02 7.23646162e-03 -4.12173480e-01 -2.25398079e-01
1.14835538e-01 2.29949906e-01 4.68827784e-01 -1.29328823e+00
-2.80817568e-01 -4.36046541e-01 2.04650268e-01 -7.98371956e-02
-2.99699843e-01 -2.52291113e-01 2.34608889e-01 -9.14563715e-01
-5.37412837e-02 -7.59906173e-01 -5.54413140e-01 -4.49318677e-01
-3.93224537e-01 -8.41803670e-01 1.67337179e-01 -1.04494296e-01
1.59535027e+00 -1.78435493e+00 1.02157891e-01 5.95230103e-01
5.65667212e-01 -8.21482483e-03 -1.67387217e-01 8.14788401e-01
-4.77907360e-02 5.56809723e-01 -2.99828947e-01 -1.72528058e-01
2.95130938e-01 5.06385744e-01 -2.32245818e-01 5.88786900e-01
-2.61045754e-01 9.24535871e-01 -1.49251163e+00 -5.02281547e-01
1.80046946e-01 5.66459835e-01 -3.34913254e-01 -5.25381304e-02
-1.16306715e-01 -2.47128531e-01 -8.75225008e-01 3.02161008e-01
3.43249619e-01 -1.95192859e-01 4.45946068e-01 -8.20718408e-02
2.37406090e-01 4.17210728e-01 -1.01776099e+00 1.69532204e+00
-5.06372571e-01 3.99550915e-01 -5.27358592e-01 -6.97753370e-01
7.68024623e-01 9.09095183e-02 7.43539259e-02 -2.39518434e-01
2.05384463e-01 -4.44410965e-02 -3.86673734e-02 -3.03924233e-01
8.68393242e-01 -5.96556306e-01 -4.10026371e-01 3.79587978e-01
2.57340223e-01 -3.73099118e-01 1.66000530e-01 2.75821179e-01
1.03608012e+00 -7.51091763e-02 1.06497955e+00 -1.20584941e+00
2.73811519e-01 -2.02402219e-01 8.67924690e-02 8.62261176e-01
-3.77846620e-04 1.09022804e-01 2.44478837e-01 -2.97341734e-01
-6.65298402e-01 -1.32153511e+00 -3.04237902e-01 1.21777046e+00
2.78354853e-01 -7.90591717e-01 -6.23247504e-01 -4.14971143e-01
-3.33194807e-02 1.12707496e+00 -7.89750218e-01 -2.68931419e-01
-1.19113989e-01 -7.96458542e-01 5.89315712e-01 1.97400436e-01
-1.26548305e-01 -1.03508472e+00 -5.21676123e-01 1.07091069e-01
2.12410659e-01 -8.12695086e-01 -2.62641042e-01 1.21585466e-01
-5.31145692e-01 -1.02145195e+00 -6.88067973e-01 -3.26024264e-01
3.50808531e-01 2.39689872e-01 1.68859255e+00 1.96704909e-01
-3.63553464e-01 6.61184430e-01 -6.04471028e-01 -4.65388566e-01
-3.90575200e-01 4.20784913e-02 2.58560032e-01 -2.48071268e-01
8.43296826e-01 -6.54281080e-01 -6.00139141e-01 -8.92116651e-02
-1.38375604e+00 -4.85546559e-01 8.13585371e-02 6.74882770e-01
1.44949839e-01 -2.21486241e-01 2.82635480e-01 -1.15096438e+00
1.13407409e+00 -7.92661488e-01 -3.61757666e-01 3.11936170e-01
-1.03957641e+00 4.67804581e-01 1.07035927e-01 -3.71368825e-01
-8.29885960e-01 -4.95584905e-01 -2.55230516e-01 2.22393081e-01
-3.79505545e-01 7.52545476e-01 -8.93353671e-02 1.02015130e-01
7.81124592e-01 -4.95673083e-02 -6.02411389e-01 -2.48003557e-01
1.09070694e+00 6.85758948e-01 2.13374883e-01 -7.85264552e-01
5.03918827e-01 4.79282111e-01 1.44869447e-01 -1.07978320e+00
-9.34377670e-01 -6.74616814e-01 -6.41955495e-01 2.12447047e-02
9.41575587e-01 -4.42638457e-01 -7.24038601e-01 1.42329680e-02
-1.29169345e+00 7.29479939e-02 -1.94404095e-01 3.54558766e-01
-3.83377016e-01 7.28956342e-01 -3.35922450e-01 -1.01152170e+00
-7.81232268e-02 -5.96285403e-01 1.20312393e+00 1.82034627e-01
-9.45006788e-01 -1.70647562e+00 6.96258068e-01 -2.36179024e-01
3.01697731e-01 2.42923573e-01 1.11049867e+00 -8.09597373e-01
5.59611283e-02 7.54376501e-02 -4.65562753e-02 -8.02146718e-02
3.65789264e-01 6.11913577e-02 -8.68718803e-01 1.18985660e-01
-3.32634419e-01 2.70666063e-01 1.11214912e+00 2.79911220e-01
8.43967140e-01 -3.59981358e-02 -6.09056771e-01 -1.10549793e-01
1.70268440e+00 -9.62467045e-02 4.99668032e-01 4.27721776e-02
6.34173393e-01 5.33484697e-01 3.64508778e-01 4.30824220e-01
5.74642181e-01 4.04241979e-01 2.70214796e-01 2.81881154e-01
1.77978247e-01 -4.32483852e-01 -8.38187858e-02 8.47131908e-01
-1.55712053e-01 -6.42059982e-01 -1.21189237e+00 9.49752152e-01
-1.78005850e+00 -9.74168777e-01 -3.92554134e-01 2.07203698e+00
7.29438603e-01 2.78850377e-01 1.04074650e-01 5.71582802e-02
7.45267272e-01 5.21620810e-01 2.85331339e-01 -5.73659360e-01
4.57715727e-02 6.26612127e-01 6.34508431e-01 9.15385246e-01
-5.95806420e-01 1.11752701e+00 7.63699770e+00 1.12040412e+00
-5.07468402e-01 1.24776110e-01 1.74718902e-01 4.39761847e-01
-1.15574646e+00 1.53571680e-01 -4.17523086e-01 4.55740720e-01
1.02060294e+00 -4.68688071e-01 1.81964114e-01 4.38891381e-01
1.13858722e-01 -2.14684844e-01 -8.92683983e-01 8.89569700e-01
5.09990342e-02 -1.17886353e+00 4.40367728e-01 1.12896495e-01
4.45997804e-01 -8.69126469e-02 -3.34533006e-01 -1.42984778e-01
8.85502338e-01 -1.13658035e+00 6.57926857e-01 9.51960683e-01
3.17833364e-01 -5.39934754e-01 5.61049402e-01 2.54458368e-01
-1.13756049e+00 2.33961582e-01 -3.68161827e-01 -3.89308989e-01
1.68437138e-01 9.79761660e-01 -8.02677155e-01 6.15265667e-01
3.74079257e-01 6.00038946e-01 -3.80024701e-01 6.72294497e-01
-4.53700393e-01 7.54194617e-01 -2.47128814e-01 -6.68782651e-01
2.58074015e-01 -1.72382906e-01 5.37471473e-01 1.69600379e+00
4.19906914e-01 4.45306785e-02 1.55569762e-01 8.89082789e-01
2.23656967e-01 2.84746110e-01 -1.01712608e+00 -3.11823100e-01
6.26731575e-01 1.10190403e+00 -1.09976256e+00 -3.99038315e-01
-1.37612984e-01 7.96210706e-01 3.12652260e-01 2.79831558e-01
-4.13785517e-01 -8.03478509e-02 8.30928326e-01 1.85706034e-01
-9.51124206e-02 -3.98433656e-01 -2.17552885e-01 -1.01913500e+00
-2.60678619e-01 -1.85011610e-01 5.24848104e-01 -7.28676498e-01
-1.88347614e+00 6.90719068e-01 6.08532250e-01 -7.29564548e-01
-4.56058919e-01 -9.54050362e-01 -5.34485757e-01 1.01872385e+00
-1.35762405e+00 -1.05770195e+00 1.50167989e-02 4.33274537e-01
2.52962649e-01 1.81444436e-01 1.00883698e+00 -1.75171897e-01
9.24092010e-02 1.54044658e-01 -3.38304460e-01 -4.98105317e-01
3.03490311e-01 -1.62002015e+00 6.46721363e-01 4.80036527e-01
6.05898440e-01 1.09090996e+00 1.33981895e+00 -7.70228326e-01
-9.18624341e-01 -7.83245921e-01 1.37278426e+00 -6.74002111e-01
1.03289855e+00 -4.08586890e-01 -7.93047309e-01 3.53580683e-01
4.02674466e-01 -1.40541464e-01 1.05037642e+00 2.76962906e-01
-6.57199919e-01 4.47673053e-01 -1.14536500e+00 8.19772780e-01
1.33375907e+00 -7.77147293e-01 -1.33860457e+00 1.86450586e-01
7.13759065e-01 4.03075755e-01 -7.40732372e-01 2.66818572e-02
5.61868370e-01 -6.81678116e-01 8.24170172e-01 -7.57245064e-01
1.90306176e-02 -1.92271814e-01 -3.49067986e-01 -1.74609017e+00
-4.33246404e-01 -6.91162527e-01 2.31751591e-01 1.00592494e+00
2.26766020e-01 -7.98739135e-01 3.15284252e-01 2.01824188e-01
3.06529582e-01 -5.62425137e-01 -9.95349050e-01 -4.81182396e-01
4.11602974e-01 -6.73804402e-01 5.94413340e-01 1.07184601e+00
3.14936370e-01 4.76863325e-01 -3.65811177e-02 2.75194738e-02
5.72062910e-01 -1.93787739e-01 9.58490819e-02 -1.70851433e+00
-1.63022727e-01 -7.82561302e-01 -9.14261043e-01 -7.40428925e-01
4.94159669e-01 -9.55970228e-01 -2.60038264e-02 -1.63972723e+00
4.09431517e-01 -1.30782649e-01 -7.09057808e-01 9.05961841e-02
-5.32718241e-01 3.80455911e-01 4.71397229e-02 -2.84128726e-01
-5.90050101e-01 2.69155711e-01 6.61655724e-01 -6.59661787e-03
3.83543909e-01 -3.34585607e-01 -9.53306496e-01 9.67765749e-01
4.40483510e-01 -6.84336066e-01 -5.95883667e-01 -1.24367833e-01
9.63319778e-01 -3.50611538e-01 5.61263978e-01 -2.01070756e-01
-1.21325858e-01 -4.65973586e-01 -1.25844270e-01 -1.07266791e-01
-5.24561442e-02 -7.63719440e-01 1.16030313e-01 3.99447739e-01
-4.87101763e-01 2.97796100e-01 -7.58368224e-02 8.48348498e-01
2.55737398e-02 -4.88006413e-01 3.60372007e-01 -2.17404678e-01
-7.63015091e-01 9.41495895e-02 -5.95811367e-01 1.14988752e-01
6.86514795e-01 -1.60619900e-01 1.90853979e-02 -3.15996230e-01
-8.67544353e-01 -1.25824362e-01 5.26773989e-01 6.12536430e-01
3.47638369e-01 -1.24936557e+00 -6.95073783e-01 -3.87015164e-01
3.70367378e-01 -7.50226617e-01 1.21002257e-01 2.88886279e-01
-1.78261921e-01 2.63621777e-01 9.09911245e-02 -4.64692086e-01
-1.10617793e+00 7.30931818e-01 8.13560337e-02 -3.29046786e-01
-1.91912413e-01 6.86944664e-01 1.33665591e-01 -2.49770030e-01
-3.74540865e-01 -5.29145420e-01 -4.71994013e-01 8.99344832e-02
4.66738611e-01 4.62885231e-01 -1.71570376e-01 -8.44118237e-01
-6.09451890e-01 5.90044618e-01 2.95994997e-01 -2.30255097e-01
1.11971843e+00 -3.59707355e-01 -3.52944106e-01 1.01604688e+00
9.19390678e-01 2.85890728e-01 -3.79869998e-01 -2.12250203e-01
5.62769473e-01 -3.87036741e-01 -1.04078520e-02 -5.19874752e-01
-1.60056338e-01 4.96985018e-01 5.64023077e-01 9.16643023e-01
6.15621924e-01 3.37209493e-01 4.48886156e-01 1.89621016e-01
7.59749591e-01 -6.71076894e-01 -4.85247999e-01 5.59576631e-01
6.43558323e-01 -6.61345303e-01 1.56982392e-01 -6.18153274e-01
-1.84305757e-01 9.22802806e-01 -1.25266656e-01 -4.95755672e-01
1.08539069e+00 1.49174452e-01 -1.66720040e-02 -5.84692001e-01
-5.62706113e-01 -6.13802254e-01 6.73099518e-01 7.75125921e-01
9.88317907e-01 3.75838220e-01 -1.21900773e+00 5.18530905e-01
-3.39726985e-01 -1.16861977e-01 4.23008144e-01 6.84040904e-01
-4.42862839e-01 -1.31879401e+00 -1.26874000e-01 2.62982339e-01
-4.92190748e-01 -4.31031078e-01 -5.76999366e-01 6.40719414e-01
8.01198408e-02 1.02096379e+00 -7.08535239e-02 -1.87959403e-01
6.35243356e-02 1.52581796e-01 5.34459233e-01 -8.08568478e-01
-3.97360325e-01 -1.61678836e-01 1.28201157e-01 -3.89089316e-01
-7.14483798e-01 -3.88178617e-01 -1.13528168e+00 -1.55749559e-01
-2.08104521e-01 2.97935754e-01 5.65749764e-01 1.32098114e+00
1.77345976e-01 4.90136206e-01 1.60272360e-01 -5.99132836e-01
-4.14094508e-01 -1.24431527e+00 -8.54981720e-01 6.10923946e-01
-1.22820511e-01 -8.71936262e-01 -1.86931208e-01 -1.37480050e-01] | [10.267486572265625, 8.883526802062988] |
1f8a99c5-01e6-4fa3-9d25-a12f9cac7fbd | deceptive-ai-ecosystems-the-case-of-chatgpt | 2306.13671 | null | https://arxiv.org/abs/2306.13671v1 | https://arxiv.org/pdf/2306.13671v1.pdf | Deceptive AI Ecosystems: The Case of ChatGPT | ChatGPT, an AI chatbot, has gained popularity for its capability in generating human-like responses. However, this feature carries several risks, most notably due to its deceptive behaviour such as offering users misleading or fabricated information that could further cause ethical issues. To better understand the impact of ChatGPT on our social, cultural, economic, and political interactions, it is crucial to investigate how ChatGPT operates in the real world where various societal pressures influence its development and deployment. This paper emphasizes the need to study ChatGPT "in the wild", as part of the ecosystem it is embedded in, with a strong focus on user involvement. We examine the ethical challenges stemming from ChatGPT's deceptive human-like interactions and propose a roadmap for developing more transparent and trustworthy chatbots. Central to our approach is the importance of proactive risk assessment and user participation in shaping the future of chatbot technology. | ['Stefan Sarkadi', 'Yifan Xu', 'Xiao Zhan'] | 2023-06-18 | null | null | null | null | ['chatbot', 'chatbot'] | ['methodology', 'natural-language-processing'] | [-1.33274108e-01 5.41161478e-01 3.31420511e-01 -2.21689437e-02
-3.24074119e-01 -7.56692290e-01 6.42213225e-01 5.71902767e-02
-1.95528701e-01 7.52374291e-01 3.89945418e-01 -3.94809783e-01
1.92891985e-01 -4.21971142e-01 1.06082000e-01 -5.54107964e-01
3.98875207e-01 1.86853081e-01 6.19150437e-02 -5.26974082e-01
7.39464879e-01 2.34978586e-01 -8.21137130e-01 7.56090805e-02
8.33957076e-01 5.33507049e-01 -1.99456424e-01 5.98284900e-01
4.82427292e-02 1.47663915e+00 -1.18530238e+00 -1.32324588e+00
-3.00052874e-02 -5.78553796e-01 -1.14765334e+00 -3.55928510e-01
-1.56739086e-01 -5.73453546e-01 1.80946589e-01 9.55396652e-01
3.13381463e-01 -1.89276695e-01 3.34984273e-01 -1.69388163e+00
-7.18154192e-01 5.81351399e-01 -2.21620157e-01 -4.49477173e-02
3.73151124e-01 5.47663808e-01 8.40157688e-01 -2.50209063e-01
7.82861769e-01 1.18753481e+00 9.71140146e-01 6.82069361e-01
-1.16003704e+00 -7.33109236e-01 -5.15289605e-01 -2.57010758e-02
-8.52792621e-01 -4.51178104e-01 4.89316225e-01 -6.68535650e-01
7.13748813e-01 5.18701911e-01 7.82923579e-01 1.47948062e+00
5.00914156e-01 2.98368752e-01 1.13605809e+00 -3.91042262e-01
1.97058037e-01 7.40288079e-01 -1.92518830e-01 8.27733353e-02
4.19762492e-01 -2.67793387e-01 -4.73594218e-01 -7.35016048e-01
5.35421133e-01 -4.86775756e-01 1.91404670e-01 1.03284225e-01
-6.76097155e-01 9.34612989e-01 3.04245085e-01 7.73055792e-01
-5.29645681e-01 5.30878484e-01 6.38231516e-01 3.86141032e-01
5.42159438e-01 7.91204214e-01 7.01054186e-02 -1.16903663e+00
-2.17766508e-01 1.06942609e-01 1.14779854e+00 4.17343378e-01
2.97289431e-01 -3.62775564e-01 1.58203989e-01 6.15967691e-01
4.00577039e-01 3.72215100e-02 -2.40576416e-01 -1.36649728e+00
2.30456889e-01 7.04167724e-01 5.29788435e-01 -1.39649534e+00
1.51578709e-01 3.11773811e-02 -1.21932223e-01 3.49539071e-01
7.55550861e-01 -5.90814650e-01 1.36977047e-01 1.33701491e+00
2.68009335e-01 -4.77272034e-01 -2.57369906e-01 7.77337193e-01
3.61377090e-01 4.06498253e-01 3.16352397e-01 6.54383153e-02
1.21543610e+00 -1.27056599e-01 -7.51252174e-01 -1.59757778e-01
8.64727855e-01 -8.14260602e-01 1.11181056e+00 3.31263036e-01
-9.87710416e-01 7.65891910e-01 -5.67323267e-01 -6.56818077e-02
-1.57246411e-01 -2.92569309e-01 5.28005421e-01 1.41365528e+00
-7.87298858e-01 5.32280743e-01 -3.82206291e-01 -8.50721240e-01
5.55734992e-01 9.28434208e-02 -3.35991621e-01 1.65938213e-01
-1.10040295e+00 1.53245664e+00 -3.60598266e-01 2.49403790e-01
-1.16961665e-01 -3.50525111e-01 -3.35693240e-01 -2.15305239e-02
4.51281309e-01 -1.92866459e-01 1.49898505e+00 -1.02850926e+00
-1.55988443e+00 8.14532220e-01 4.27122742e-01 -2.43567288e-01
9.94646788e-01 -8.36031586e-02 1.91997886e-01 1.47157043e-01
1.55040145e-01 1.32941008e-01 4.20177042e-01 -1.18877459e+00
-2.12848321e-01 -2.02633440e-01 1.97747350e-01 -5.74474633e-02
-3.56098115e-01 9.54919934e-01 6.14246964e-01 -4.91869561e-02
-4.92107302e-01 -1.01607096e+00 -1.11666553e-01 -4.50862497e-02
-2.80384868e-01 -2.00554937e-01 1.02864802e+00 -6.76126540e-01
8.03583443e-01 -2.10741282e+00 -6.05759203e-01 5.33009768e-02
4.58769828e-01 4.53715056e-01 4.12108749e-01 1.27194595e+00
8.21024895e-01 8.19881260e-01 4.23150986e-01 -1.10057443e-01
1.48392007e-01 3.79134715e-02 -4.10147011e-02 4.05307114e-01
1.99785441e-01 7.48566449e-01 -1.12560105e+00 -4.34957147e-01
-1.12502821e-01 4.87880677e-01 -3.02283287e-01 1.66135803e-01
1.50980428e-01 5.72826028e-01 -5.50267398e-01 4.15927112e-01
5.06666183e-01 3.09029892e-02 5.51007807e-01 6.80235207e-01
-5.96088052e-01 5.31734347e-01 -1.28226519e-01 5.89909077e-01
-4.12843317e-01 1.00925720e+00 7.94744432e-01 -3.30225676e-01
7.48556018e-01 5.40367126e-01 1.40634447e-01 -3.68562758e-01
7.48244822e-01 3.52838039e-01 4.48515594e-01 -6.58744693e-01
7.17945337e-01 -3.62417758e-01 -3.90453339e-02 1.12905121e+00
-3.20806324e-01 -1.75400063e-01 -4.89363998e-01 5.11487722e-01
1.26566625e+00 -2.70202104e-02 1.43921956e-01 -2.99736708e-01
1.15595020e-01 1.70683891e-01 3.74257982e-01 5.29051721e-01
-9.41764176e-01 1.35429397e-01 1.24916494e+00 -3.13695729e-01
-7.93985963e-01 -3.99824739e-01 3.03630382e-01 9.57698703e-01
3.27221379e-02 -2.45089486e-01 -8.68022323e-01 -7.71430671e-01
-6.91017956e-02 1.03708494e+00 -5.81073761e-01 -3.65446985e-01
-2.69308567e-01 -4.43788767e-01 1.08688951e+00 -1.11724481e-01
6.18208349e-01 -1.27542305e+00 -1.20438254e+00 4.11603361e-01
-5.27388155e-01 -7.90412605e-01 -2.32118383e-01 -3.00570399e-01
-3.82338107e-01 -1.22295761e+00 -3.50095987e-01 -1.09998412e-01
4.05010134e-01 3.42807263e-01 8.36709201e-01 5.42438149e-01
-2.25117162e-01 3.89082909e-01 -7.65574813e-01 -6.52670324e-01
-1.16068542e+00 4.40531829e-03 -2.97080368e-01 -1.53311685e-01
2.89953053e-01 -8.53843510e-01 -4.75469619e-01 5.69038093e-01
-5.40175200e-01 -2.50657320e-01 2.32178614e-01 5.20657837e-01
-1.09837973e+00 -3.09207708e-01 1.02773619e+00 -1.09530449e+00
1.15408337e+00 -8.35476398e-01 -7.70886019e-02 1.40331268e-01
-4.27702069e-01 -7.54923820e-01 4.80117470e-01 -3.64951342e-01
-1.34720254e+00 -6.39307976e-01 5.62329311e-03 3.83644551e-01
-5.58295622e-02 1.13813587e-01 5.33374883e-02 -4.48092639e-01
8.75431716e-01 -5.31002820e-01 5.36713898e-01 -3.22710574e-01
7.98658580e-02 9.69203055e-01 -2.10519075e-01 -6.17470682e-01
7.25259662e-01 3.73024732e-01 -4.38610524e-01 -1.11889946e+00
-4.58990820e-02 -1.78009629e-01 -1.53879598e-01 -7.97691226e-01
6.91393614e-01 -2.32247561e-01 -1.29957986e+00 7.20658123e-01
-1.47838008e+00 -5.08848608e-01 2.28070229e-01 9.08945873e-02
-9.89642888e-02 4.59748685e-01 -8.78029466e-01 -1.56152427e+00
-3.53310794e-01 -7.38981247e-01 2.55036026e-01 1.65757030e-01
-9.99346912e-01 -8.85104656e-01 1.09566096e-02 1.14454222e+00
9.04914439e-01 3.80017906e-01 7.27652192e-01 -6.06514394e-01
-4.53984290e-01 -4.74286318e-01 -2.73926228e-01 4.17449385e-01
1.40392557e-01 2.79957354e-01 -8.14118385e-01 1.63945749e-01
3.01406324e-01 -7.28701293e-01 -3.80568177e-01 -4.65181768e-01
-5.87500893e-02 -8.79014552e-01 1.05699070e-01 -4.58439261e-01
1.07415807e+00 4.71149892e-01 8.05629075e-01 4.07998770e-01
3.17410141e-01 1.21620345e+00 6.31178200e-01 4.82679427e-01
3.12011153e-01 2.57766187e-01 6.61682129e-01 4.76068556e-01
5.28434694e-01 -3.17346960e-01 5.23987830e-01 4.74439293e-01
-2.42839888e-01 -1.21828422e-01 -9.49830711e-01 4.31646168e-01
-1.85540807e+00 -1.17275751e+00 -5.03220260e-01 1.86304331e+00
3.69743317e-01 2.31763273e-02 5.39087415e-01 1.14509203e-01
7.60777652e-01 -1.26497492e-01 -1.43483013e-01 -1.23503566e+00
4.51109171e-01 -9.64312628e-02 2.58497894e-01 1.98298007e-01
-2.10976690e-01 7.71351933e-01 6.15141106e+00 1.95379689e-01
-9.90710378e-01 4.36849862e-01 9.50228155e-01 1.42929122e-01
-2.01747835e-01 2.69702852e-01 1.15918525e-01 4.95419890e-01
8.46381366e-01 -4.36506182e-01 4.55798179e-01 6.85270488e-01
6.32757545e-01 -6.40703440e-01 -5.38615942e-01 2.07181737e-01
-6.63611740e-02 -1.10979855e+00 -9.52561080e-01 3.22888792e-01
2.49850750e-01 -2.08219722e-01 -1.29963234e-01 7.46360645e-02
6.43030405e-01 -9.89171684e-01 8.18033040e-01 -1.75644234e-01
2.49349505e-01 -6.48003280e-01 7.85945714e-01 5.77365160e-01
-2.14696750e-01 2.80581154e-02 4.99993972e-02 -9.09254491e-01
2.88670599e-01 1.46149486e-01 -1.15518367e+00 -9.08123255e-02
6.28460407e-01 7.08003268e-02 -2.64937878e-01 7.25295365e-01
-1.64749414e-01 8.46338749e-01 -1.83050618e-01 -6.71695352e-01
3.25307131e-01 -4.95682687e-01 5.61613619e-01 7.95245588e-01
9.89969596e-02 1.68751985e-01 -5.92464268e-01 1.14539194e+00
6.81791827e-02 1.90768614e-02 -8.68659437e-01 -6.02949083e-01
7.69017100e-01 1.42561400e+00 -8.51491094e-01 2.86301345e-01
-8.75703320e-02 8.49915802e-01 1.29790917e-01 -1.63609475e-01
-5.21592319e-01 -2.53962070e-01 8.07174265e-01 3.75298500e-01
-2.86041617e-01 -1.19676860e-02 -6.65123641e-01 -7.98586309e-01
8.01599622e-02 -9.75897193e-01 -1.49703786e-01 -6.42439425e-01
-1.39404547e+00 2.04848439e-01 -4.24657881e-01 -5.37472010e-01
-1.22961275e-01 -8.03800374e-02 -1.13787425e+00 7.06201136e-01
-4.36464161e-01 -1.33339238e+00 -2.59110369e-02 -1.38467282e-01
-2.29706168e-01 3.61152142e-01 5.64661503e-01 1.01010434e-01
-3.33863139e-01 6.18668795e-01 -1.21508643e-01 -3.07725705e-02
5.89832187e-01 -6.81845605e-01 4.98200834e-01 4.86029267e-01
-4.94853199e-01 7.81731665e-01 8.32817435e-01 -6.24935389e-01
-1.38915586e+00 -3.50022167e-01 1.23688018e+00 -1.08685219e+00
1.07652700e+00 -6.80874646e-01 -6.93583012e-01 6.45403087e-01
4.63798493e-01 -8.75776350e-01 6.45789206e-01 3.20240885e-01
-2.26972908e-01 3.10171723e-01 -1.91927671e+00 8.66146207e-01
1.01900840e+00 -7.86565840e-01 -3.90666068e-01 2.78779417e-01
4.25946206e-01 1.50928453e-01 -5.95764220e-01 -7.66539276e-01
9.03519392e-01 -1.48276055e+00 2.10119396e-01 -3.56864810e-01
6.32165372e-01 4.05324012e-01 5.49023271e-01 -1.08391309e+00
-1.03476949e-01 -1.36470902e+00 7.32847214e-01 1.76705468e+00
1.80158481e-01 -1.19994569e+00 1.07317114e+00 1.57780480e+00
2.83761561e-01 -5.53947270e-01 -9.89944518e-01 -6.15805507e-01
3.15851599e-01 -1.41722322e-01 3.02717865e-01 1.21358907e+00
7.83392251e-01 1.02433048e-01 -8.00451458e-01 -2.97259092e-01
3.14280361e-01 -7.15798438e-01 9.30594027e-01 -1.21179330e+00
1.28469691e-01 -1.30616695e-01 -4.18745816e-01 2.63791867e-02
-2.91758299e-01 -1.26629069e-01 2.12037265e-01 -1.21330798e+00
1.13107719e-01 -5.27723849e-01 6.23970628e-01 3.11426252e-01
2.13663116e-01 3.17501366e-01 8.04090917e-01 2.46244103e-01
-3.22740257e-01 2.27713093e-01 1.05229616e+00 5.01462460e-01
-2.89683640e-01 4.13965099e-02 -1.26060879e+00 5.08972108e-01
9.92037177e-01 -7.36170292e-01 -2.06327870e-01 5.57111688e-02
5.80489159e-01 4.98247482e-02 3.82556528e-01 -6.67578876e-01
6.86535761e-02 -6.00421429e-01 -6.16664886e-01 3.30690384e-01
3.55128944e-01 -9.21825469e-01 3.91733527e-01 5.68578243e-01
-2.03830525e-01 -5.74664474e-02 -2.92642087e-01 1.89903393e-01
2.92505950e-01 -5.24589419e-01 6.21597350e-01 -3.78226727e-01
4.54913646e-01 -5.56425631e-01 -1.31062198e+00 2.01565295e-01
1.07053471e+00 -3.53246123e-01 -7.50651717e-01 -1.03046644e+00
-6.71726018e-02 2.17068940e-01 1.01280391e+00 2.59872407e-01
8.79113674e-02 -8.56611311e-01 -4.31012213e-01 -3.59294593e-01
-9.47938412e-02 -5.82993031e-01 1.04708597e-01 9.23243821e-01
-9.83559012e-01 -7.25465044e-02 -4.88806486e-01 1.86768174e-01
-1.46865797e+00 -1.30615041e-01 2.13586658e-01 7.69656524e-02
-5.37388086e-01 6.55382395e-01 -2.20802948e-01 -1.25576302e-01
-5.84413111e-03 2.14644030e-01 1.37539670e-01 7.02465177e-02
3.34951699e-01 8.64929020e-01 -2.81057745e-01 -6.07397199e-01
-3.96421015e-01 -5.23982406e-01 -4.29064184e-02 -5.05408227e-01
1.20111358e+00 -1.59720868e-01 -6.66988969e-01 3.94168943e-01
7.65904784e-01 1.91146791e-01 -8.69985878e-01 5.81591308e-01
4.49312150e-01 -1.18012786e+00 -5.60748219e-01 -1.02257967e+00
-4.92646873e-01 6.82553589e-01 -3.06325614e-01 1.02792227e+00
3.31437618e-01 -3.35218936e-01 1.07869184e+00 2.92997479e-01
6.01532400e-01 -1.23953533e+00 3.62939894e-01 3.50115150e-01
8.31242263e-01 -8.85411739e-01 -2.91404933e-01 -6.22597098e-01
-1.08522022e+00 9.39377785e-01 6.20448351e-01 1.28790084e-02
2.67369330e-01 3.46904658e-02 6.18400216e-01 -2.96990931e-01
-8.47132504e-01 5.14978051e-01 -7.88970292e-01 9.51962292e-01
5.51427603e-01 1.15991116e-01 -8.27149153e-01 4.30814028e-01
-1.86828449e-01 2.01856986e-01 1.27493203e+00 1.09296155e+00
-2.61183619e-01 -1.41165459e+00 -3.92963022e-01 3.70079547e-01
-8.34332705e-01 2.37853318e-01 -1.50814199e+00 8.66819441e-01
1.94412898e-02 1.42922282e+00 -5.02870142e-01 -4.71210182e-01
-5.10582365e-02 2.47918740e-01 -5.40739745e-02 -5.51590502e-01
-1.75567782e+00 -5.31422615e-01 1.03486001e+00 -3.83818269e-01
-1.51005030e-01 -8.36275220e-01 -7.78193533e-01 -1.17077410e+00
-5.65183282e-01 5.21072567e-01 7.37954974e-01 8.89109612e-01
4.92922664e-01 -4.79134232e-01 5.92557907e-01 -6.05161190e-01
-5.87254405e-01 -9.08604503e-01 -4.41816330e-01 3.66260797e-01
-8.90351757e-02 -2.80340612e-01 -5.51004529e-01 -5.15158474e-01] | [10.424187660217285, 7.3156938552856445] |
d396014f-495a-4e95-bd20-c639cbf76f11 | fhdr-hdr-image-reconstruction-from-a-single | 1912.11463 | null | https://arxiv.org/abs/1912.11463v1 | https://arxiv.org/pdf/1912.11463v1.pdf | FHDR: HDR Image Reconstruction from a Single LDR Image using Feedback Network | High dynamic range (HDR) image generation from a single exposure low dynamic range (LDR) image has been made possible due to the recent advances in Deep Learning. Various feed-forward Convolutional Neural Networks (CNNs) have been proposed for learning LDR to HDR representations. To better utilize the power of CNNs, we exploit the idea of feedback, where the initial low level features are guided by the high level features using a hidden state of a Recurrent Neural Network. Unlike a single forward pass in a conventional feed-forward network, the reconstruction from LDR to HDR in a feedback network is learned over multiple iterations. This enables us to create a coarse-to-fine representation, leading to an improved reconstruction at every iteration. Various advantages over standard feed-forward networks include early reconstruction ability and better reconstruction quality with fewer network parameters. We design a dense feedback block and propose an end-to-end feedback network- FHDR for HDR image generation from a single exposure LDR image. Qualitative and quantitative evaluations show the superiority of our approach over the state-of-the-art methods. | ['Shanmuganathan Raman', 'Mukul Khanna', 'Zeeshan Khan'] | 2019-12-24 | null | null | null | null | ['single-image-based-hdr-reconstruction'] | ['computer-vision'] | [ 1.89344108e-01 -1.16754189e-01 -4.76094857e-02 -5.27208388e-01
-4.66338903e-01 9.73727368e-03 5.10428190e-01 -5.37930071e-01
-1.52937755e-01 5.93934178e-01 7.06506371e-01 -5.79850487e-02
4.91055474e-02 -9.49207902e-01 -8.23325574e-01 -5.57413459e-01
1.52079895e-01 -9.64593813e-02 1.37419298e-01 -3.72563839e-01
-6.46137893e-02 7.16003299e-01 -1.31222057e+00 4.73501056e-01
7.09582031e-01 1.17171979e+00 5.06292462e-01 7.68625200e-01
3.56423967e-02 1.31557047e+00 -2.75753111e-01 1.09184124e-01
4.43807125e-01 -5.83677649e-01 -6.28602862e-01 1.03323618e-02
1.07899681e-01 -8.92047465e-01 -1.10582960e+00 6.21479273e-01
9.28964972e-01 2.94488311e-01 3.35452557e-01 -6.94366753e-01
-1.15392768e+00 4.08491343e-01 -7.16322482e-01 1.58857688e-01
5.59276044e-01 3.72438282e-01 6.94126129e-01 -1.04048824e+00
7.48780370e-01 1.20215416e+00 6.85936034e-01 5.29052377e-01
-1.22231877e+00 -7.58581460e-01 -9.86485481e-02 -3.78604271e-02
-1.08310652e+00 -3.27691138e-01 8.70748758e-01 -1.19457297e-01
1.08383083e+00 -5.75333089e-03 8.40870976e-01 8.25812995e-01
3.54081273e-01 6.40808463e-01 1.21202457e+00 -4.28793132e-01
-1.60043120e-01 -2.87215412e-01 -2.89208114e-01 7.13112533e-01
-3.05498242e-01 7.13015139e-01 -5.88813484e-01 4.20330524e-01
1.42823756e+00 5.88385820e-01 -5.53319931e-01 -9.00612548e-02
-1.19306409e+00 8.30210209e-01 1.36412024e+00 2.44142801e-01
-5.32088697e-01 4.87697273e-01 1.56829625e-01 5.75665176e-01
4.27042842e-01 2.68371075e-01 -3.24856192e-02 2.96604067e-01
-1.07219791e+00 -7.57029578e-02 3.32705498e-01 5.04132867e-01
7.92398334e-01 1.93222269e-01 -3.47753435e-01 8.50401223e-01
5.01747787e-01 4.23781067e-01 5.54018378e-01 -8.42819870e-01
2.91312426e-01 4.37610537e-01 -3.63965556e-02 -7.52096057e-01
-4.37021017e-01 -5.36207080e-01 -1.37297523e+00 5.18050969e-01
-2.21786842e-01 -4.95004617e-02 -1.37154329e+00 1.31559086e+00
-3.12905051e-02 -5.25916368e-02 8.60295445e-02 1.33341038e+00
1.03495383e+00 1.10900807e+00 -2.11905181e-01 -1.76501587e-01
8.48516285e-01 -9.74459767e-01 -7.92025566e-01 -1.13453232e-01
1.76525980e-01 -6.88501120e-01 1.02442229e+00 2.63635606e-01
-1.02522123e+00 -9.60569918e-01 -1.07555640e+00 -4.61123526e-01
-1.44521207e-01 1.56223863e-01 5.17506838e-01 1.04061104e-02
-1.24404681e+00 7.03710258e-01 -3.68631124e-01 -3.99749018e-02
4.22953635e-01 3.45470041e-01 -3.20634305e-01 -4.68337744e-01
-1.43778634e+00 6.94688022e-01 1.77603424e-01 5.07642269e-01
-9.64995801e-01 -8.53755414e-01 -7.28405774e-01 -1.75662041e-01
7.42554739e-02 -8.01498771e-01 1.00379682e+00 -8.95857394e-01
-1.94951344e+00 6.80519998e-01 1.75280526e-01 -5.12811422e-01
4.79045689e-01 -4.55914378e-01 -2.54948676e-01 1.73965961e-01
-2.04380080e-01 9.27415252e-01 9.11434054e-01 -1.23411381e+00
-3.34672898e-01 -4.31795046e-03 3.81853580e-02 2.69864053e-01
4.15936448e-02 3.46376151e-02 -1.99942604e-01 -7.83917010e-01
2.96295226e-01 -8.74553502e-01 -3.17882866e-01 2.84489453e-01
-2.52159268e-01 1.28134176e-01 1.09421301e+00 -5.71061611e-01
1.10460961e+00 -2.05577755e+00 5.13365828e-02 -3.92083302e-02
4.48732227e-01 3.60917687e-01 -6.29255772e-02 3.88961315e-01
-4.20984894e-01 -2.13660806e-01 4.07856852e-02 -3.99380326e-02
-3.48263115e-01 -1.70503501e-02 -6.88767314e-01 5.22548914e-01
7.11855069e-02 1.22240758e+00 -7.41885364e-01 -1.16180845e-01
7.13390052e-01 9.04406965e-01 -2.97277212e-01 7.66347468e-01
-1.61839291e-01 7.46799052e-01 -1.03612110e-01 3.87841851e-01
4.75871772e-01 -5.55056274e-01 -2.86671817e-01 -5.64687014e-01
-3.72638136e-01 -2.67974217e-03 -7.78176129e-01 1.87142253e+00
-8.51523042e-01 7.90766954e-01 -4.51342136e-01 -5.70742786e-01
1.33152759e+00 3.18867356e-01 3.05166632e-01 -1.25260293e+00
4.12504785e-02 2.45663792e-01 -2.02391431e-01 -3.34936500e-01
4.68090355e-01 -4.18800563e-01 5.85759729e-02 5.64888835e-01
5.42820320e-02 2.20904481e-02 -2.54752517e-01 1.82061866e-01
8.54507983e-01 3.54143649e-01 1.27512112e-01 3.11684161e-01
4.09366578e-01 -6.12871230e-01 4.20584708e-01 5.77152789e-01
2.33077258e-01 1.16472137e+00 6.28260104e-03 -8.08873415e-01
-1.26786888e+00 -9.39279437e-01 7.77062848e-02 8.38914812e-01
-2.79273894e-02 -5.85707389e-02 -1.82171553e-01 -2.90357828e-01
-3.40176940e-01 1.15149058e-01 -7.00783432e-01 -3.07692260e-01
-8.70203435e-01 -5.13991237e-01 2.32300758e-01 6.40885770e-01
1.07351804e+00 -1.42177570e+00 -7.18381703e-01 3.42311263e-01
-1.49924770e-01 -1.00516009e+00 -4.28681970e-01 2.98214257e-01
-1.03240299e+00 -5.27551591e-01 -1.34397113e+00 -6.81638062e-01
4.94744271e-01 2.06525370e-01 1.11151946e+00 9.03699771e-02
-5.33207953e-01 -1.30923539e-01 -4.22211945e-01 1.68120936e-01
-3.50823164e-01 -3.10766269e-02 -2.43042320e-01 -3.19244824e-02
-3.07702273e-01 -5.60027957e-01 -1.19901538e+00 2.09839255e-01
-9.02199864e-01 5.16447663e-01 1.17132246e+00 8.94790769e-01
8.55641663e-01 -9.68792588e-02 5.62035680e-01 -7.68099904e-01
4.09068465e-01 -1.85147926e-01 -4.27255273e-01 2.14336112e-01
-5.84217787e-01 2.80598283e-01 9.33795571e-01 -4.10533309e-01
-1.27207005e+00 1.57727540e-01 -3.49964350e-01 -9.10871267e-01
1.46843970e-01 2.04787135e-01 1.94072500e-01 -2.65491694e-01
7.28312731e-01 4.30338353e-01 4.82742116e-02 -4.07623559e-01
8.07125986e-01 5.61132252e-01 6.54756010e-01 -1.62233621e-01
7.86112845e-01 4.20919746e-01 -5.15006557e-02 -5.92777908e-01
-8.31858695e-01 -2.14435846e-01 -4.56427336e-01 -5.40647686e-01
9.00642097e-01 -1.16632581e+00 -6.25292897e-01 6.97202027e-01
-8.77326131e-01 -7.62740612e-01 -6.65692151e-01 3.60028893e-01
-6.76019311e-01 -8.41238722e-02 -1.01749635e+00 -4.38260555e-01
-7.68289864e-01 -9.32707727e-01 9.96753752e-01 6.19529366e-01
1.41886786e-01 -8.85207117e-01 2.16023624e-01 -9.29679200e-02
9.55309451e-01 3.14195305e-01 7.25871086e-01 2.00950786e-01
-7.67807603e-01 -3.95226553e-02 -5.58977365e-01 2.95685560e-01
1.58334643e-01 -2.50189960e-01 -9.46363568e-01 -5.23300350e-01
-4.80506532e-02 -6.81159616e-01 1.14211452e+00 6.40521348e-01
1.20136344e+00 -3.50210518e-01 -2.73336619e-01 1.00676477e+00
1.88865399e+00 -1.77949965e-02 1.16704953e+00 2.27258757e-01
8.30258310e-01 1.87667683e-01 4.58084792e-01 3.16140443e-01
6.78591430e-02 6.50038719e-01 2.08441496e-01 -8.71427000e-01
-7.80142307e-01 -4.96053547e-01 4.06184375e-01 5.59728861e-01
-6.58788010e-02 -8.94782171e-02 -5.64511001e-01 2.39783704e-01
-1.58191597e+00 -9.51393902e-01 1.80533797e-01 1.99692583e+00
8.24302137e-01 6.12350553e-02 3.31835300e-02 2.65755178e-03
5.66387653e-01 5.90855718e-01 -6.14808083e-01 -2.25391686e-01
8.70153308e-02 1.73757017e-01 4.19950187e-01 3.18461716e-01
-5.92794240e-01 7.80515671e-01 6.54796124e+00 5.02454519e-01
-1.76055717e+00 -1.74523175e-01 9.06824529e-01 -2.30941668e-01
-3.86974216e-01 -2.44792193e-01 -6.25324547e-01 3.89258832e-01
8.72338176e-01 2.31619060e-01 4.96877044e-01 5.67796946e-01
3.41023356e-01 5.41278627e-03 -8.84577096e-01 1.12599516e+00
-5.03623672e-02 -1.50830877e+00 1.25476241e-01 7.29808658e-02
9.55333650e-01 1.68392330e-01 4.46516722e-01 2.94084996e-01
4.08078581e-01 -1.23556316e+00 4.90046114e-01 9.10599589e-01
1.45937192e+00 -9.06618953e-01 5.71327031e-01 9.65392217e-02
-1.26787436e+00 -3.50173265e-01 -5.94282389e-01 2.05257490e-01
3.50901902e-01 8.30487490e-01 -4.91795123e-01 4.08600569e-01
8.73869658e-01 8.86765599e-01 -4.42586124e-01 6.87048614e-01
-3.01925719e-01 2.42694378e-01 1.54836206e-02 3.80589128e-01
1.41618654e-01 -4.92508188e-02 2.24798620e-01 1.04923379e+00
2.62470633e-01 3.51422191e-01 -4.67850678e-02 1.09231305e+00
-4.52923387e-01 -3.47794324e-01 -8.38770330e-01 1.14406399e-01
1.41505092e-01 1.28550613e+00 -4.79993820e-01 -1.44399554e-01
-3.31830591e-01 1.28426456e+00 2.74707377e-01 3.74780744e-01
-6.94825530e-01 -6.57056212e-01 1.24261998e-01 4.97555047e-01
4.13506120e-01 -1.18718883e-02 -2.00592987e-02 -9.09450948e-01
-2.42480010e-01 -7.62822926e-01 7.21786022e-02 -1.21102154e+00
-1.19743085e+00 9.41694736e-01 -3.88509989e-01 -1.42047334e+00
-4.51780826e-01 -1.64038152e-01 -6.87028110e-01 9.86844778e-01
-1.86111462e+00 -1.23963118e+00 -6.89362764e-01 7.07969010e-01
5.27662992e-01 -5.50477728e-02 6.48119688e-01 2.35950768e-01
-2.77209193e-01 3.65186483e-01 -5.22844866e-02 2.96469957e-01
7.06635892e-01 -1.09617043e+00 4.51676220e-01 9.24706221e-01
-1.06653742e-01 5.60161650e-01 4.40868020e-01 -5.71320415e-01
-1.25543475e+00 -1.23483098e+00 5.54652393e-01 2.76901037e-01
1.73941195e-01 -2.84225732e-01 -9.05916095e-01 5.24300337e-01
2.16207474e-01 6.93920493e-01 2.41759539e-01 -1.62050441e-01
-3.23484659e-01 -5.34286201e-01 -1.09141767e+00 3.50335896e-01
1.04285562e+00 -7.30973601e-01 -2.75587708e-01 -4.11388613e-02
1.05400503e+00 -6.12497985e-01 -1.04412556e+00 3.30262452e-01
6.31998658e-01 -1.18110895e+00 1.15544569e+00 1.09347925e-01
8.71405721e-01 -4.52813089e-01 1.69360880e-02 -1.38188183e+00
-5.29355884e-01 -6.28385961e-01 -2.97336847e-01 8.88345957e-01
1.42475843e-01 -4.42189097e-01 6.24051094e-01 4.04314011e-01
-2.14358091e-01 -1.06661832e+00 -3.84247869e-01 -4.60401535e-01
-1.32194519e-01 1.46857068e-01 4.82005894e-01 4.75422263e-01
-5.15158713e-01 5.38722873e-01 -8.51990521e-01 -2.00976223e-01
6.20910227e-01 4.53603715e-01 4.57889140e-01 -9.54874516e-01
-4.34885561e-01 6.44158274e-02 -2.32255116e-01 -1.38322413e+00
-2.09400147e-01 -7.51031220e-01 2.96391815e-01 -1.94079649e+00
2.72335202e-01 -7.10786879e-01 -2.38753617e-01 4.24172729e-01
-3.00408248e-03 6.93418264e-01 3.07614535e-01 4.73566562e-01
-5.59903800e-01 7.24174559e-01 1.62432480e+00 2.18688175e-01
-3.31110477e-01 -2.08117902e-01 -5.26958346e-01 2.81777442e-01
5.64692736e-01 -3.82087171e-01 -4.58442867e-01 -4.77728128e-01
4.11615461e-01 4.75838989e-01 4.08049643e-01 -1.17696071e+00
3.20006788e-01 3.63897651e-01 1.10154784e+00 -6.81445718e-01
1.69568285e-01 -7.26990700e-01 4.92779195e-01 6.11047626e-01
-5.33550441e-01 -9.81443003e-02 -2.53988713e-01 5.95904291e-01
-3.66278768e-01 3.25050592e-01 1.22040844e+00 -2.72041291e-01
-8.42920065e-01 6.59821868e-01 -2.12333798e-02 -3.94341111e-01
7.18340278e-01 -3.34925205e-01 -2.53204554e-01 -5.45806050e-01
-7.38540173e-01 -1.14589266e-01 4.44006294e-01 4.71665442e-01
1.14894283e+00 -1.56500542e+00 -7.39916742e-01 4.62516516e-01
-8.06407481e-02 9.94869545e-02 5.04362226e-01 6.76564038e-01
-5.64207315e-01 3.27146798e-01 -5.07682443e-01 -5.96926212e-01
-9.23061132e-01 4.30062801e-01 6.31868601e-01 -5.32017231e-01
-1.31077361e+00 6.96620166e-01 2.45756447e-01 -2.34352767e-01
1.11345142e-01 -4.87432443e-02 -1.95117444e-01 -2.72323012e-01
6.81239963e-01 1.98330924e-01 -1.17865145e-01 -3.76932383e-01
2.58122515e-02 7.97601998e-01 -2.14465812e-01 -1.28762305e-01
1.68116534e+00 -2.90677637e-01 8.13912526e-02 4.60338473e-01
1.51950943e+00 -4.54828650e-01 -1.70282900e+00 -4.18384016e-01
-6.97709143e-01 -6.24530017e-01 3.84195834e-01 -9.75378275e-01
-1.50238764e+00 9.15930569e-01 1.08292472e+00 -1.42075792e-01
1.36109471e+00 -1.16111428e-01 1.19475150e+00 9.52600613e-02
3.67169499e-01 -8.39808643e-01 6.84343398e-01 2.93899119e-01
1.18987238e+00 -1.32570827e+00 1.00241497e-01 -1.68928280e-02
-5.81104875e-01 1.42118025e+00 5.37519336e-01 -4.60910976e-01
6.42649949e-01 3.66827935e-01 4.56199870e-02 -2.27943867e-01
-7.26046860e-01 -1.31402194e-01 1.97349638e-01 5.36832273e-01
6.09816551e-01 -3.56042653e-01 2.07370609e-01 -4.05553132e-02
9.44371745e-02 6.21297598e-01 4.34604347e-01 6.43164635e-01
-5.11416495e-01 -8.50094676e-01 -6.38332590e-02 4.61241633e-01
-3.45024109e-01 -2.17104591e-02 7.18356222e-02 5.93049586e-01
-1.04811251e-01 6.20928347e-01 1.17579117e-01 -5.24071574e-01
3.08374763e-01 -5.92409313e-01 5.66200614e-01 -4.29124475e-01
-5.39356470e-01 -4.81823180e-03 -4.17643607e-01 -9.12854433e-01
-4.27816242e-01 -1.00308180e-01 -1.37310576e+00 -4.51854557e-01
-2.51347780e-01 -3.38822663e-01 5.51745713e-01 6.05795383e-01
3.16373616e-01 9.21088278e-01 1.16270339e+00 -1.06329572e+00
-2.98375487e-01 -8.78632665e-01 -4.63386714e-01 1.75450161e-01
6.94940209e-01 -3.06738347e-01 -2.53461540e-01 -3.98421660e-02] | [10.88201904296875, -2.202371835708618] |
7f43a08e-279c-4003-a6ba-60e4d8637c21 | moet-interpretable-and-verifiable-1 | null | null | https://openreview.net/forum?id=BJlxdCVKDB | https://openreview.net/pdf?id=BJlxdCVKDB | MoET: Interpretable and Verifiable Reinforcement Learning via Mixture of Expert Trees | Deep Reinforcement Learning (DRL) has led to many recent breakthroughs on complex control tasks, such as defeating the best human player in the game of Go. However, decisions made by the DRL agent are not explainable, hindering its applicability in safety-critical settings. Viper, a recently proposed technique, constructs a decision tree policy by mimicking the DRL agent. Decision trees are interpretable as each action made can be traced back to the decision rule path that lead to it. However, one global decision tree approximating the DRL policy has significant limitations with respect to the geometry of decision boundaries. We propose MoET, a more expressive, yet still interpretable model based on Mixture of Experts, consisting of a gating function that partitions the state space, and multiple decision tree experts that specialize on different partitions. We propose a training procedure to support non-differentiable decision tree experts and integrate it into imitation learning procedure of Viper. We evaluate our algorithm on four OpenAI gym environments, and show that the policy constructed in such a way is more performant and better mimics the DRL agent by lowering mispredictions and increasing the reward. We also show that MoET policies are amenable for verification using off-the-shelf automated theorem provers such as Z3. | ['Sarfraz Khurshid', 'Rishabh Singh', 'Mladen Nikolic', 'Kaiyuan Wang', 'Andrija Petrovic', 'Marko Vasic'] | 2019-09-25 | null | null | null | null | ['game-of-go'] | ['playing-games'] | [-1.50718451e-01 6.87588453e-01 -4.17188674e-01 -3.11719161e-02
-2.61519700e-01 -9.70557153e-01 3.02939504e-01 1.01994105e-01
-4.86393899e-01 1.07395756e+00 -1.39953882e-01 -7.31435537e-01
-2.36983314e-01 -8.23724866e-01 -9.92278814e-01 -3.97319078e-01
-1.77765399e-01 8.16883981e-01 4.79655623e-01 -3.37162852e-01
-4.15876992e-02 3.08539450e-01 -1.44830108e+00 1.85154721e-01
1.11240613e+00 7.10216343e-01 2.79148715e-03 9.49670434e-01
4.80793655e-01 1.13727951e+00 -4.83116806e-01 -1.13329977e-01
3.65798712e-01 -5.25477350e-01 -1.10636449e+00 -1.82406589e-01
-1.28968671e-01 -4.36889887e-01 -1.34231895e-01 8.27733636e-01
1.75594077e-01 2.73118287e-01 5.63481569e-01 -1.51971698e+00
-7.75206462e-02 9.28762376e-01 -1.32766530e-01 -1.66833520e-01
2.51165032e-01 5.81442535e-01 9.48122084e-01 2.82517850e-01
7.08662927e-01 1.41224742e+00 4.57298517e-01 1.02136028e+00
-1.47111809e+00 -4.28329408e-01 4.24618632e-01 2.74271965e-01
-9.45801675e-01 -1.33224651e-01 4.51295942e-01 -3.17070782e-01
9.41530645e-01 2.34353483e-01 8.01353037e-01 1.32780468e+00
6.49882019e-01 1.00931799e+00 1.42150366e+00 -4.16988432e-01
8.56908023e-01 -1.20982170e-01 -2.66217232e-01 8.75783026e-01
1.67398483e-01 6.14076614e-01 -1.26299098e-01 -1.48502022e-01
1.03966928e+00 -3.08448434e-01 -3.66743952e-02 -8.45987320e-01
-8.76783550e-01 8.95589292e-01 2.25452721e-01 9.52146873e-02
-3.80629122e-01 5.18141150e-01 4.40190703e-01 5.23681223e-01
-1.83489919e-01 8.61329556e-01 -6.15119100e-01 -2.97625452e-01
-5.03483355e-01 7.76587844e-01 1.04398191e+00 6.02606297e-01
3.36642891e-01 1.80881768e-01 -3.84709358e-01 6.58054352e-02
2.91314363e-01 2.72020549e-01 3.08089435e-01 -1.62343609e+00
2.11608157e-01 5.59691668e-01 5.11676490e-01 -4.48693126e-01
-5.52336872e-01 -3.26882154e-01 -3.61961693e-01 9.74092364e-01
5.48918962e-01 -5.80959439e-01 -6.63368940e-01 1.99069071e+00
5.15338123e-01 2.14153901e-01 1.20827101e-01 8.87575984e-01
-9.12202373e-02 5.82282066e-01 1.37716785e-01 1.14494763e-01
1.12962103e+00 -7.20859587e-01 -4.29523379e-01 -2.07840398e-01
8.22261393e-01 1.22095048e-01 8.23558271e-01 8.34318876e-01
-1.15982747e+00 -4.78231192e-01 -1.00753009e+00 3.24214607e-01
3.80769446e-02 -7.89321810e-02 8.28427434e-01 4.91146028e-01
-9.53407288e-01 9.41999078e-01 -1.27152717e+00 -1.69449270e-01
3.09198141e-01 7.08766282e-01 -2.03495577e-01 3.39124382e-01
-1.16883576e+00 1.28467011e+00 7.38865852e-01 -6.83520958e-02
-1.44439483e+00 -2.64626265e-01 -9.73655939e-01 2.43462130e-01
8.39712381e-01 -7.77703881e-01 1.81858587e+00 -9.30380106e-01
-2.15178800e+00 4.93211597e-01 2.74008363e-01 -1.00220966e+00
6.83430731e-01 -8.83660764e-02 -6.71028718e-02 5.26649766e-02
1.19377062e-01 8.75722170e-01 7.76901364e-01 -9.25423265e-01
-8.49943042e-01 -3.12509388e-01 7.28847563e-01 2.34382421e-01
5.11522830e-01 -3.34846318e-01 2.00973660e-01 -2.14514211e-01
-4.99469221e-01 -1.17287290e+00 -6.56625032e-01 -4.59239520e-02
-3.41402799e-01 -4.76249397e-01 3.85235012e-01 -5.00827610e-01
1.18613958e+00 -1.74920237e+00 3.51011246e-01 2.11634338e-01
1.52872562e-01 3.05385292e-01 6.57044947e-02 2.20799223e-01
2.37283222e-02 5.20414263e-02 -2.16558754e-01 1.00904360e-01
4.91819948e-01 6.99025571e-01 -4.59700555e-01 3.97910595e-01
-1.74095053e-02 6.90171242e-01 -1.12832594e+00 -3.22165698e-01
3.22380364e-01 -3.04028809e-01 -9.85999227e-01 2.26976484e-01
-7.93412924e-01 7.46780157e-01 -8.33713174e-01 6.17984608e-02
7.30416849e-02 1.92161441e-01 4.90372449e-01 6.24063909e-01
-1.32147893e-01 4.64364201e-01 -1.26801682e+00 1.45319021e+00
-4.83809054e-01 2.20221862e-01 1.43000528e-01 -1.19465947e+00
6.98267162e-01 3.81799608e-01 2.18644395e-01 -3.93186212e-01
2.27927983e-01 8.64970163e-02 4.66281146e-01 -3.94625574e-01
1.47390008e-01 -3.22169363e-01 -3.51979822e-01 3.75074953e-01
1.03334561e-01 -3.16492945e-01 2.07340881e-01 -1.09192751e-01
1.22803986e+00 8.83790851e-01 6.04320168e-01 -2.65409857e-01
4.33438182e-01 3.06675494e-01 8.81880999e-01 1.02445090e+00
-3.93190414e-01 -2.10566983e-01 9.88108695e-01 -4.96792614e-01
-9.21916425e-01 -7.83010006e-01 2.79522806e-01 9.85629857e-01
6.04579002e-02 -2.99012572e-01 -1.06136131e+00 -9.16213095e-01
1.32000640e-01 1.35222471e+00 -6.17670655e-01 -4.62520599e-01
-5.99695385e-01 1.75776761e-02 7.82879412e-01 5.56848884e-01
4.35512990e-01 -1.28875411e+00 -1.27408171e+00 4.62054312e-01
-1.10791335e-02 -8.76052856e-01 -3.16345662e-01 5.03632784e-01
-9.51046407e-01 -1.24106348e+00 3.23619395e-02 -5.12040854e-01
2.78998554e-01 -5.53928792e-01 9.42784607e-01 -1.43365815e-01
5.87669536e-02 4.92931128e-01 -1.64467290e-01 -4.83931243e-01
-8.88257980e-01 -3.90777923e-02 4.49763924e-01 -3.73086154e-01
-7.32299546e-03 -5.15232742e-01 -3.70826423e-01 2.68646955e-01
-5.04411399e-01 2.94988483e-01 9.10793319e-02 7.72958219e-01
4.48613167e-01 3.00734013e-01 3.55512589e-01 -6.46190822e-01
7.00486004e-01 -1.09225288e-01 -1.06518877e+00 1.45527527e-01
-5.21875441e-01 7.48222113e-01 1.09375429e+00 -3.89506340e-01
-9.23849344e-01 5.18237837e-02 -2.10166629e-02 -2.92170435e-01
-3.84425640e-01 1.05677530e-01 -3.89135517e-02 1.37615725e-01
7.43068278e-01 -4.10963595e-03 1.27944365e-01 -2.03636438e-01
3.61915112e-01 2.44333029e-01 5.11111438e-01 -1.21088421e+00
6.10425651e-01 8.54939446e-02 1.07796416e-01 -2.34763145e-01
-6.09335959e-01 2.16420099e-01 -2.88240880e-01 -1.39033690e-01
8.37643862e-01 -5.44549942e-01 -1.66131282e+00 1.48107290e-01
-9.71739173e-01 -9.19882178e-01 -7.35924304e-01 3.57969671e-01
-1.21451199e+00 3.48667689e-02 -5.88831306e-01 -1.00973248e+00
-1.97157860e-02 -1.24407578e+00 7.66033471e-01 2.03077659e-01
-5.62618554e-01 -9.02649581e-01 3.24155658e-01 4.22060005e-02
-1.16639584e-02 3.43668610e-01 1.02781415e+00 -7.94867635e-01
-3.22501630e-01 7.93424323e-02 6.18503869e-01 2.92466938e-01
-2.30916217e-01 -3.40936109e-02 -6.18187606e-01 -3.09617251e-01
-1.21632531e-01 -4.74884808e-01 3.34301054e-01 5.82893491e-01
1.35102963e+00 -5.46645761e-01 -3.79573166e-01 1.79714873e-01
1.00937033e+00 5.54889977e-01 4.04827535e-01 5.92901409e-01
2.11615607e-01 3.85973245e-01 6.89338148e-01 2.97222376e-01
5.06007135e-01 6.11030042e-01 5.94294190e-01 2.62857854e-01
3.80204827e-01 -7.34680414e-01 8.22429836e-01 -2.53687836e-02
-3.36331517e-01 -1.26192430e-02 -7.78874218e-01 2.95274824e-01
-2.30010557e+00 -1.13092363e+00 4.42056507e-01 2.23087573e+00
9.08514440e-01 4.77489322e-01 4.20228481e-01 5.84243163e-02
3.83055508e-01 -2.99112827e-01 -9.76945281e-01 -9.87213612e-01
4.33456063e-01 5.25439739e-01 4.77618873e-01 7.07153380e-01
-9.84545588e-01 1.19254339e+00 6.52546501e+00 7.66274452e-01
-9.31391537e-01 -1.01087615e-01 4.77533996e-01 1.44994885e-01
-7.32177421e-02 1.90240383e-01 -6.28424168e-01 1.70332387e-01
9.22208130e-01 -1.56710610e-01 7.49905646e-01 1.16394234e+00
6.04379654e-01 -2.78382331e-01 -1.53852117e+00 3.68977696e-01
-7.74399400e-01 -1.20877051e+00 -3.25567573e-01 -1.74793694e-02
4.29222882e-01 -2.59219587e-01 -2.19412029e-01 8.77134681e-01
1.27775550e+00 -1.07076013e+00 1.04381585e+00 2.16904327e-01
3.95094186e-01 -9.08721685e-01 4.21906769e-01 9.22273576e-01
-8.54973137e-01 -4.41803724e-01 -5.87562323e-02 -5.06526291e-01
-3.66654158e-01 -3.31461966e-01 -1.07868934e+00 3.66286188e-01
4.60736364e-01 3.41910124e-01 -9.22656283e-02 7.64018774e-01
-8.68603706e-01 8.41817856e-01 -4.73551631e-01 -1.56077430e-01
5.57360828e-01 -3.23838025e-01 6.24415815e-01 5.99544287e-01
-6.73254132e-02 3.05314422e-01 5.45476913e-01 1.13884890e+00
4.40667480e-01 -4.77935374e-01 -6.34363294e-01 5.68836406e-02
1.68924391e-01 8.78744721e-01 -4.31007832e-01 -3.11237872e-01
2.83178449e-01 7.90975571e-01 4.91156816e-01 3.52220029e-01
-1.07188797e+00 1.71633102e-02 8.69817734e-01 1.06865978e-02
3.40366185e-01 -1.65568128e-01 -2.29153067e-01 -1.10580218e+00
-1.71667531e-01 -1.24985492e+00 3.18671197e-01 -7.13761806e-01
-7.01946855e-01 3.86577606e-01 1.05233811e-01 -1.05780280e+00
-8.08725774e-01 -6.04942143e-01 -6.31216526e-01 7.11731195e-01
-1.03959072e+00 -5.91360688e-01 4.23421115e-01 4.64724839e-01
4.86471146e-01 9.36436206e-02 9.80144322e-01 -4.47287858e-01
-6.50646031e-01 3.66344512e-01 -1.88782737e-01 -6.60363659e-02
1.12721533e-01 -1.48770380e+00 2.54767209e-01 6.95801377e-01
-2.71850377e-01 3.91913205e-01 9.35918152e-01 -5.02907813e-01
-1.23306346e+00 -8.52884591e-01 2.83224672e-01 -3.28141153e-01
7.37827122e-01 -2.41570115e-01 -6.65233791e-01 1.03770161e+00
1.09982520e-01 -4.10887361e-01 1.59129083e-01 1.57729805e-01
1.09260127e-01 -7.27538466e-02 -1.22810400e+00 9.37031567e-01
9.05659080e-01 -2.50774920e-01 -9.59963381e-01 6.86385259e-02
6.31696403e-01 -6.93384290e-01 -6.82457209e-01 1.54508799e-01
4.40705031e-01 -7.85517812e-01 5.58621228e-01 -1.18220723e+00
4.73577291e-01 -5.38424671e-01 3.11355233e-01 -1.66683781e+00
-3.17657173e-01 -9.71318364e-01 -4.38474000e-01 4.96195853e-01
1.61742806e-01 -6.78570330e-01 7.88980365e-01 7.60512412e-01
-1.65769085e-01 -8.71861398e-01 -1.12141275e+00 -8.74488294e-01
2.49567941e-01 -4.48863804e-01 4.01678801e-01 5.46033561e-01
4.86684054e-01 3.25732887e-01 -3.59455168e-01 1.74083769e-01
5.32763898e-01 2.66387939e-01 6.61192775e-01 -1.01418364e+00
-8.77759874e-01 -4.54871148e-01 -2.30320990e-02 -1.19123781e+00
4.40175384e-01 -8.04864049e-01 3.56982946e-01 -1.35608304e+00
-3.10482532e-01 -7.26515055e-01 -2.01748055e-03 1.02824318e+00
9.08235982e-02 -6.65683985e-01 3.18363696e-01 -3.47540855e-01
-6.95035279e-01 6.10181093e-01 1.61720538e+00 -1.59514114e-01
-3.52655828e-01 3.71325850e-01 -5.48594177e-01 1.09710896e+00
1.03508973e+00 -4.87804651e-01 -4.43573505e-01 -5.50240502e-02
1.02966063e-01 5.47861516e-01 3.39121699e-01 -1.13832903e+00
9.02963951e-02 -5.55375993e-01 1.50448695e-01 -6.68420941e-02
8.27268437e-02 -8.96536827e-01 1.26570016e-01 1.02054143e+00
-6.01000428e-01 5.09760855e-03 3.44418377e-01 4.22312081e-01
1.29203588e-01 -3.44583243e-01 7.12455511e-01 -2.22028553e-01
-7.14789748e-01 1.12466551e-01 -1.01613712e+00 -3.43656205e-02
1.35371721e+00 -1.67410165e-01 1.61673293e-01 -4.58064228e-01
-9.55776155e-01 8.59777033e-01 3.01327974e-01 2.41517410e-01
2.49505505e-01 -8.56828570e-01 -4.82051373e-01 -4.31851037e-02
-3.19566429e-01 2.59072989e-01 -7.67451748e-02 5.60938656e-01
-4.96781677e-01 3.55587125e-01 -4.81674910e-01 -3.27680200e-01
-1.04415143e+00 7.16221154e-01 8.24251711e-01 -8.61110032e-01
-7.67057300e-01 3.06091905e-01 1.73465222e-01 -8.22970092e-01
1.75869316e-01 -9.03270006e-01 -2.32492670e-01 -5.16016006e-01
2.83417106e-01 2.69298047e-01 -3.92030895e-01 1.03126064e-01
-3.45336437e-01 2.19521090e-01 1.78986624e-01 -2.59399891e-01
1.21929944e+00 2.78972268e-01 2.49240994e-01 3.80024731e-01
7.24273100e-02 -3.89630526e-01 -1.57205808e+00 1.96283817e-01
6.06088750e-02 1.58175588e-01 -1.23047911e-01 -1.04001999e+00
-4.70548898e-01 6.75534129e-01 3.10267568e-01 1.87713325e-01
9.41858411e-01 -3.58276814e-01 4.66425389e-01 6.85565352e-01
9.14795399e-01 -1.28604531e+00 -3.07954401e-01 6.72771037e-01
5.98710060e-01 -8.43616962e-01 -2.26193324e-01 4.45876978e-02
-8.27637792e-01 1.04463828e+00 8.93204927e-01 -5.10923624e-01
7.16681406e-02 1.86157018e-01 -4.08730716e-01 1.29167959e-01
-1.08091080e+00 -1.86220422e-01 -1.99430227e-01 6.75235093e-01
-1.28459021e-01 4.78607267e-01 -1.89090297e-01 6.48640513e-01
-1.89062864e-01 3.76603991e-01 6.06872082e-01 9.59768951e-01
-5.12723744e-01 -1.28794062e+00 -5.33386946e-01 5.10643125e-02
-9.40937176e-02 4.09765869e-01 -1.01244904e-01 1.04517329e+00
3.22963804e-01 7.94566751e-01 1.81498937e-02 -2.64348537e-01
4.11292791e-01 -2.23647226e-02 8.88702571e-01 -5.51933289e-01
-7.15311646e-01 -2.35897183e-01 3.06519121e-01 -8.89322698e-01
-2.20825803e-02 -5.49449146e-01 -1.59992158e+00 -4.51084316e-01
5.11687323e-02 3.67021233e-01 2.05509260e-01 1.16110671e+00
2.55423814e-01 8.46172988e-01 5.61230063e-01 -7.65051961e-01
-1.06901407e+00 -6.23413265e-01 -5.29092848e-01 -8.69791433e-02
3.21347892e-01 -7.85435081e-01 7.12456703e-02 -2.86177307e-01] | [4.261503219604492, 1.7179113626480103] |
cdb1a1de-18e7-4aad-a61a-e2d3693c70bd | toward-cross-theory-discourse-relation | null | null | https://aclanthology.org/W19-2702 | https://aclanthology.org/W19-2702.pdf | Toward Cross-theory Discourse Relation Annotation | In this exploratory study, we attempt to automatically induce PDTB-style relations from RST trees. We work with a German corpus of news commentary articles, annotated for RST trees and explicit PDTB-style relations and we focus on inducing the implicit relations in an automated way. Preliminary results look promising as a high-precision (but low-recall) way of finding implicit relations where there is no shallow structure annotated at all, but mapping proves more difficult in cases where EDUs and relation arguments overlap, yet do not seem to signal the same relation. | ['Olha Zolotarenko', 'Peter Bourgonje'] | 2019-06-01 | null | null | null | ws-2019-6 | ['implicit-relations'] | ['natural-language-processing'] | [ 1.43344849e-01 1.16636777e+00 -5.31276762e-01 -3.78996432e-01
-7.21055567e-01 -7.28195310e-01 1.11530399e+00 6.00067735e-01
-2.72601128e-01 1.35285437e+00 6.56415343e-01 -8.05754662e-01
-2.89791465e-01 -7.60806680e-01 -3.27835113e-01 -1.32396653e-01
-1.50495172e-01 1.10163343e+00 6.91353142e-01 -5.97579420e-01
1.26902416e-01 2.02500194e-01 -9.82284427e-01 8.38679671e-01
7.42460310e-01 3.02595735e-01 -1.41468197e-01 5.25085449e-01
-4.01888311e-01 1.38855827e+00 -7.74303794e-01 -1.08257163e+00
-2.44334757e-01 -4.86164629e-01 -1.92306542e+00 -2.64056116e-01
1.00462593e-01 2.69266158e-01 4.74876724e-02 8.72308791e-01
-1.93266179e-02 -1.61457136e-01 9.07022715e-01 -9.88752902e-01
-2.39461467e-01 1.18086588e+00 -4.33493972e-01 5.27592003e-01
7.36367047e-01 -6.14214659e-01 1.57424498e+00 -5.43009222e-01
1.11015153e+00 1.35153770e+00 8.90237093e-01 3.33819211e-01
-1.68573928e+00 -2.06909254e-01 1.30689457e-01 7.24526644e-02
-1.04183400e+00 -5.11148870e-01 4.67182130e-01 -4.70721006e-01
1.47949195e+00 6.37620509e-01 2.40154922e-01 9.19225395e-01
-1.18893601e-01 3.85248065e-01 1.31341386e+00 -9.52760518e-01
-3.70539278e-01 3.49947602e-01 5.65174341e-01 4.59032148e-01
3.85715187e-01 -1.48605421e-01 -4.24476087e-01 -3.06460828e-01
4.35109794e-01 -7.51440883e-01 -1.60287857e-01 2.44389996e-01
-1.05729795e+00 8.12939346e-01 4.14376944e-01 1.13051260e+00
-2.20782265e-01 -7.33828694e-02 5.69871604e-01 5.37241280e-01
5.64338744e-01 8.52023423e-01 -8.16045403e-01 -2.31355622e-01
-5.27372539e-01 2.18866035e-01 1.48443723e+00 1.10100806e+00
8.31249952e-01 -5.69456100e-01 -2.14061216e-02 8.48477900e-01
-6.55314997e-02 -1.38930753e-01 2.59511530e-01 -9.67985630e-01
7.34679222e-01 6.81092501e-01 3.55705440e-01 -1.21363986e+00
-3.63046020e-01 1.92501768e-01 -4.07568574e-01 -3.57518435e-01
5.89991093e-01 -1.81466192e-01 -4.45072234e-01 1.49996471e+00
2.83555776e-01 -3.29322368e-01 4.35732484e-01 4.26045746e-01
1.16905487e+00 5.16906559e-01 3.10193598e-01 -5.57028830e-01
1.49507105e+00 -6.87536478e-01 -9.44735825e-01 -3.04018170e-01
1.46986401e+00 -8.69434476e-01 9.00075316e-01 -7.28569552e-02
-9.60465848e-01 -1.25052437e-01 -6.75615311e-01 -2.72412270e-01
-5.01823306e-01 1.89481229e-02 8.13873708e-01 3.31002325e-01
-7.45923400e-01 5.60021222e-01 -4.24657106e-01 -5.93301177e-01
-8.63294452e-02 3.15646589e-01 -4.59706694e-01 6.19838953e-01
-1.48799229e+00 1.29896200e+00 7.49024272e-01 3.43422852e-02
9.74119231e-02 -1.94020733e-01 -7.96730757e-01 -2.23181963e-01
8.39159727e-01 -4.25228059e-01 1.39019001e+00 -9.34230506e-01
-1.23228633e+00 1.42622900e+00 -2.46254340e-01 -8.00192416e-01
2.07725078e-01 -4.96784568e-01 -5.36351442e-01 -2.70714015e-02
5.80047667e-01 2.92553395e-01 4.90024239e-02 -1.06173158e+00
-5.90398312e-01 3.12782847e-03 5.61381698e-01 1.92757517e-01
4.04175259e-02 7.28048086e-01 2.56151944e-01 -6.23505354e-01
1.57905325e-01 -6.18015885e-01 -2.01957256e-01 -7.04896867e-01
-7.96526849e-01 -9.35811937e-01 5.29216230e-01 -6.12583458e-01
1.45416367e+00 -1.69808459e+00 -4.62933332e-02 1.53569162e-01
1.90834090e-01 2.43981406e-01 3.64136666e-01 6.92786813e-01
-4.02910382e-01 6.19929194e-01 -1.32861108e-01 3.02245200e-01
2.53750831e-02 8.13528001e-01 -5.77654660e-01 -7.64187127e-02
3.04046273e-01 9.14002955e-01 -9.63004708e-01 -9.59027827e-01
-2.93925434e-01 -3.34532410e-01 -2.08203524e-01 5.22974506e-02
-4.58318889e-01 -9.59249958e-02 -5.12812316e-01 3.50327462e-01
-1.01293117e-01 -2.77396321e-01 8.74713480e-01 -2.06209898e-01
-3.34968656e-01 1.50177836e+00 -9.71435487e-01 9.67408597e-01
-3.44790339e-01 8.15908372e-01 -1.10201403e-01 -1.20838475e+00
8.75204146e-01 6.89810753e-01 7.11113289e-02 -3.76383394e-01
2.23450080e-01 4.94380563e-01 4.84840095e-01 -4.85820681e-01
3.26475948e-01 -5.06830692e-01 -2.53470451e-01 2.73966581e-01
-7.55349472e-02 -3.99147123e-01 5.59937954e-01 3.24191093e-01
1.05415130e+00 2.55941212e-01 1.03866196e+00 -5.85178971e-01
6.69148207e-01 4.70373690e-01 4.38418269e-01 5.39971173e-01
3.49196017e-01 6.04206808e-02 1.08160377e+00 -4.70008194e-01
-8.46202016e-01 -4.86381143e-01 -4.96712983e-01 7.78514802e-01
-6.70945644e-02 -1.17777371e+00 -4.57800388e-01 -1.10762775e+00
-4.75831866e-01 8.78828883e-01 -7.77344465e-01 6.27062321e-01
-9.48562503e-01 -4.34060037e-01 5.59055507e-01 3.84744704e-01
2.44935870e-01 -1.10828459e+00 -2.44561285e-01 5.68616986e-01
-4.51163948e-01 -1.44643319e+00 8.12986568e-02 6.13542140e-01
-6.41301334e-01 -1.25213337e+00 -8.29638541e-02 -1.01970696e+00
4.24071610e-01 -5.18005133e-01 1.44414461e+00 4.59989339e-01
1.37075081e-01 -3.81558806e-01 -5.16847789e-01 -4.31377441e-01
-5.61548412e-01 2.51059085e-01 -3.98635864e-01 -6.81087732e-01
6.76317096e-01 -4.06187415e-01 3.17792386e-01 4.07241762e-01
-2.91433156e-01 1.31132174e-02 2.18940735e-01 1.13522863e+00
2.08814606e-01 7.75925862e-03 3.40351105e-01 -1.72144389e+00
7.26551414e-01 -2.40166932e-01 -2.92859882e-01 1.93624169e-01
-5.22108018e-01 -1.60169210e-02 2.66582310e-01 -3.36784452e-01
-1.09758627e+00 -4.71168756e-01 -3.70337009e-01 7.31472313e-01
-3.04359972e-01 7.66893685e-01 -2.04329602e-02 3.10131758e-01
1.23053253e+00 -6.08782947e-01 -5.22379875e-01 -3.56079042e-01
4.15422112e-01 4.98280495e-01 1.18879110e-01 -1.08997059e+00
6.37884796e-01 -1.21679962e-01 7.95838088e-02 -8.49690020e-01
-1.51235545e+00 -4.91586715e-01 -8.98072541e-01 3.19810599e-01
9.22752440e-01 -5.48758090e-01 -3.18522006e-01 -4.24541205e-01
-1.64395285e+00 -6.04902267e-01 -5.78118384e-01 2.90776610e-01
-4.92437154e-01 3.17868829e-01 -1.03252804e+00 -6.30998075e-01
-7.77058303e-02 -6.36512935e-01 7.07363367e-01 -1.60099134e-01
-1.45149553e+00 -1.08778942e+00 2.00167343e-01 5.45087643e-02
-2.09544420e-01 2.30375826e-01 1.14352691e+00 -1.18802929e+00
-1.34303227e-01 5.57679497e-02 -4.27929640e-01 -7.65434578e-02
4.43525374e-01 -1.65351972e-01 -5.45135975e-01 5.26783824e-01
-1.81744888e-01 -4.89026815e-01 2.10708797e-01 -1.36569336e-01
4.18308020e-01 -9.39173222e-01 -6.59767568e-01 -1.19758040e-01
1.04603195e+00 1.11901276e-01 6.37070894e-01 6.85230553e-01
5.15499055e-01 1.14781821e+00 7.88561165e-01 -4.02919024e-01
3.29427153e-01 9.36391771e-01 -4.52907324e-01 -8.55950937e-02
-2.48276621e-01 -2.66237669e-02 1.54122878e-02 7.76187539e-01
-3.88815433e-01 -3.38713750e-02 -1.17897511e+00 6.32682204e-01
-1.87024701e+00 -1.15768325e+00 -9.54318702e-01 1.59311485e+00
1.64164674e+00 7.43649840e-01 7.81351402e-02 3.22568566e-01
5.78138292e-01 -3.81809883e-02 5.62251508e-01 -7.94063389e-01
-2.06398830e-01 4.54818040e-01 -8.43367260e-03 9.35038269e-01
-1.06809628e+00 1.30938745e+00 7.41232491e+00 4.87745017e-01
-5.21661043e-01 1.41590536e-01 4.55787271e-01 6.01324379e-01
-4.51774806e-01 5.92124343e-01 -8.65533769e-01 1.33049056e-01
9.51442003e-01 -2.12517351e-01 -2.64266402e-01 6.55779064e-01
-3.88656884e-01 -1.93153113e-01 -1.22123873e+00 3.99154723e-01
-3.31044197e-01 -1.50099301e+00 -2.69338220e-01 4.59590405e-02
5.61827123e-01 -2.76066124e-01 -7.21697450e-01 2.71015525e-01
6.51443064e-01 -8.18711042e-01 5.80776751e-01 -9.11101699e-02
6.74412310e-01 -3.62956166e-01 1.08259380e+00 2.57706612e-01
-1.05523348e+00 3.56136233e-01 -1.38272539e-01 -5.36999524e-01
2.93775439e-01 6.32405519e-01 -1.08777797e+00 6.32755041e-01
6.76027715e-01 6.86319411e-01 -2.96267182e-01 5.52521884e-01
-1.01584327e+00 8.60790372e-01 -4.32159364e-01 -4.33587283e-01
2.90124774e-01 -1.41976207e-01 6.80337727e-01 1.55709219e+00
-1.04615703e-01 8.25900614e-01 3.18116099e-01 5.58013499e-01
4.22574840e-02 3.91493708e-01 -9.32016075e-01 -8.84485245e-03
2.02423021e-01 1.06664002e+00 -1.03231573e+00 -6.61175728e-01
-3.37879688e-01 5.54497838e-01 6.19924188e-01 2.17689380e-01
-5.13523519e-01 -3.98832411e-01 8.46888497e-03 3.84883612e-01
6.30149990e-02 -7.58065432e-02 -1.96654528e-01 -9.89999831e-01
-1.96779832e-01 -7.19507933e-01 6.54774308e-01 -8.91314924e-01
-1.38625407e+00 9.13149834e-01 4.47105110e-01 -6.48316503e-01
-6.05926156e-01 -5.52204549e-01 -5.12686491e-01 9.99884069e-01
-1.06795788e+00 -1.03268802e+00 3.85485947e-01 1.70132056e-01
4.12021607e-01 2.92265683e-01 1.25536370e+00 6.17202669e-02
-1.03215791e-01 2.52649814e-01 -7.65258729e-01 3.97600681e-01
6.81391358e-01 -1.53689504e+00 3.30541670e-01 6.37653232e-01
4.79568541e-01 9.70789313e-01 1.12479341e+00 -9.02487099e-01
-3.28061759e-01 -4.36867148e-01 1.83200681e+00 -6.97190702e-01
1.37468314e+00 -2.72985756e-01 -9.57065880e-01 1.10541952e+00
7.43958950e-01 -2.30365649e-01 7.82239079e-01 1.08460033e+00
-4.65696037e-01 2.90293008e-01 -8.34479868e-01 5.78054070e-01
1.21364975e+00 -7.42531002e-01 -1.57929838e+00 5.55264294e-01
8.35031569e-01 -3.59380364e-01 -9.39078689e-01 6.71378314e-01
-6.44778535e-02 -5.39719760e-01 6.26946926e-01 -8.07005882e-01
5.28052688e-01 -2.82443285e-01 -1.31688371e-01 -9.74975169e-01
-2.95218527e-01 -7.18046427e-01 -5.90918474e-02 1.76045573e+00
1.19285262e+00 -7.33567476e-01 7.01167345e-01 5.91476321e-01
-1.78557009e-01 -7.45314598e-01 -7.59224534e-01 -7.90236473e-01
-1.05113424e-01 -9.11286995e-02 2.49636456e-01 1.29835880e+00
7.65223324e-01 1.25876153e+00 1.60268202e-01 -2.19139755e-01
1.36705667e-01 4.01524484e-01 7.75370955e-01 -1.49722469e+00
-4.18460757e-01 -4.84613031e-01 -3.64293754e-01 -9.31636393e-01
3.14645618e-01 -7.60875404e-01 6.75034598e-02 -1.64076436e+00
-4.84091183e-03 -7.87219226e-01 3.46379459e-01 7.98815668e-01
-3.49836238e-02 1.53190926e-01 -3.07954192e-01 2.19212458e-01
-3.93422484e-01 8.83861035e-02 8.47106218e-01 -6.38486817e-02
-2.35918257e-02 -1.81950685e-02 -7.72828877e-01 1.01881468e+00
5.88174641e-01 -6.97968602e-01 -4.07060295e-01 -1.45614132e-01
4.25680429e-01 1.17216762e-02 4.21786234e-02 -2.95400202e-01
1.07977455e-02 -3.83001864e-01 2.88249291e-02 -4.65100437e-01
9.08339098e-02 -6.96508586e-01 9.21999589e-02 3.77053231e-01
-5.01806378e-01 5.76759800e-02 2.44861051e-01 7.76941180e-02
-4.74949807e-01 -8.66172850e-01 3.22555155e-01 -3.15503359e-01
-4.72231239e-01 -2.29221106e-01 -5.23494542e-01 5.27321398e-01
8.78718317e-01 -1.88731402e-01 -3.98696005e-01 -1.40589878e-01
-1.04124701e+00 -2.18717694e-01 3.19613218e-01 -3.68504003e-02
1.07807487e-01 -1.01650441e+00 -6.50277436e-01 -5.40744901e-01
1.16615780e-01 5.02062924e-02 -7.39860892e-01 8.11140954e-01
-5.31718194e-01 3.18984747e-01 3.98747027e-01 -2.42824852e-01
-1.47649717e+00 5.23496151e-01 1.67561680e-01 -7.01183081e-01
-8.37815404e-01 9.37230229e-01 2.25909543e-03 -3.76697183e-01
-1.91868376e-02 -3.25442195e-01 -5.65637112e-01 3.01474035e-01
2.97259152e-01 -2.06046686e-01 -1.73417618e-03 -8.51703525e-01
-4.89619076e-01 2.98482955e-01 -1.18180700e-01 -2.84897774e-01
1.33376050e+00 -6.86692121e-03 -6.33388102e-01 8.14457238e-01
7.42469966e-01 4.50445294e-01 -1.10961586e-01 -2.98543543e-01
8.52595568e-01 -2.60149479e-01 -5.90128005e-01 -7.78947830e-01
-7.82918036e-02 3.96985799e-01 -5.05154371e-01 1.02339435e+00
4.39045459e-01 5.93001544e-01 4.14377511e-01 6.56099796e-01
4.05944079e-01 -9.65028763e-01 -7.28801265e-02 8.75174999e-01
1.05840373e+00 -9.91869807e-01 3.81706446e-01 -1.24071538e+00
-5.53722560e-01 1.07434142e+00 7.92497814e-01 4.58943769e-02
5.59832096e-01 5.56929052e-01 -7.75807202e-02 -7.56200850e-01
-8.59710217e-01 -3.98434162e-01 3.33882242e-01 5.32690644e-01
1.14945376e+00 3.22516635e-02 -1.01618683e+00 3.07239562e-01
-7.69062221e-01 -5.11861324e-01 4.45706695e-01 8.83608818e-01
-1.70636669e-01 -1.54951620e+00 5.82948290e-02 4.90637362e-01
-7.04072058e-01 -2.80708522e-01 -9.19956565e-01 1.27501667e+00
3.99556123e-02 9.62296784e-01 -3.14725637e-02 -2.27461606e-02
3.97335351e-01 9.99611337e-03 6.71427548e-01 -1.13871408e+00
-6.64565146e-01 1.35917172e-01 1.62921619e+00 -1.71869129e-01
-9.54968810e-01 -7.07964003e-01 -1.49298251e+00 -1.66269735e-01
-7.40883052e-01 8.90286982e-01 -4.83757729e-04 1.41518104e+00
-3.13575596e-01 2.40487963e-01 1.24311104e-01 -2.38704503e-01
2.91035417e-03 -1.00685465e+00 -1.92823395e-01 5.89716434e-01
-6.39332756e-02 -7.20580637e-01 -8.08197930e-02 1.17181465e-01] | [10.702935218811035, 9.336030960083008] |
b32dcfe6-e15f-4579-945c-8a4c121839c6 | cross-layer-attention-network-for-fine | 2210.08784 | null | https://arxiv.org/abs/2210.08784v1 | https://arxiv.org/pdf/2210.08784v1.pdf | Cross-layer Attention Network for Fine-grained Visual Categorization | Learning discriminative representations for subtle localized details plays a significant role in Fine-grained Visual Categorization (FGVC). Compared to previous attention-based works, our work does not explicitly define or localize the part regions of interest; instead, we leverage the complementary properties of different stages of the network, and build a mutual refinement mechanism between the mid-level feature maps and the top-level feature map by our proposed Cross-layer Attention Network (CLAN). Specifically, CLAN is composed of 1) the Cross-layer Context Attention (CLCA) module, which enhances the global context information in the intermediate feature maps with the help of the top-level feature map, thereby improving the expressive power of the middle layers, and 2) the Cross-layer Spatial Attention (CLSA) module, which takes advantage of the local attention in the mid-level feature maps to boost the feature extraction of local regions at the top-level feature maps. Experimental results show our approach achieves state-of-the-art on three publicly available fine-grained recognition datasets (CUB-200-2011, Stanford Cars and FGVC-Aircraft). Ablation studies and visualizations are provided to understand our approach. Experimental results show our approach achieves state-of-the-art on three publicly available fine-grained recognition datasets (CUB-200-2011, Stanford Cars and FGVC-Aircraft). | ['Huazhong Yang', 'Yu Wang', 'Ranran Huang'] | 2022-10-17 | null | null | null | null | ['fine-grained-visual-categorization'] | ['computer-vision'] | [-2.19534934e-01 -2.80249506e-01 3.50235193e-03 -4.00256634e-01
-5.54951787e-01 -4.94045287e-01 7.36853004e-01 4.81643975e-02
-3.54391992e-01 3.03601533e-01 3.54997247e-01 3.04841530e-02
-2.95590222e-01 -8.42335820e-01 -7.18829513e-01 -5.68353117e-01
-5.52958213e-02 -7.65879825e-02 4.29065406e-01 -1.05599210e-01
2.91905910e-01 7.84377813e-01 -1.71840954e+00 8.28697145e-01
6.59067929e-01 1.61286461e+00 1.68456376e-01 4.31869358e-01
-3.56960207e-01 7.70897686e-01 -2.20971823e-01 1.49449352e-02
1.25871941e-01 3.46337184e-02 -8.54053140e-01 -1.45569935e-01
7.98727274e-01 -7.27103055e-02 -3.29652518e-01 1.05092382e+00
1.66632354e-01 1.27491057e-01 6.39670491e-01 -1.15549636e+00
-1.03941834e+00 2.85165340e-01 -7.33113348e-01 4.48710829e-01
-6.04628958e-02 2.37724572e-01 1.04020166e+00 -1.29850948e+00
3.21404308e-01 1.51778936e+00 7.42619395e-01 2.23749369e-01
-1.08460009e+00 -8.65829825e-01 8.98987889e-01 2.77400225e-01
-1.74078894e+00 -2.28453428e-01 6.40082479e-01 -6.33946061e-01
1.09668660e+00 5.13093034e-03 3.53357136e-01 8.20243359e-01
2.02666864e-01 5.82693458e-01 1.23979223e+00 -1.92370281e-01
-4.45491821e-02 -2.20571771e-01 4.91379857e-01 9.74643767e-01
9.90880728e-02 4.17172313e-01 -4.67488825e-01 -6.23838278e-03
9.47553277e-01 3.35118443e-01 -5.83983548e-02 -3.19098324e-01
-1.19108641e+00 8.23302627e-01 1.16251791e+00 5.34372866e-01
-5.19765019e-01 1.52367666e-01 1.70680135e-01 2.01047495e-01
4.64604646e-01 2.29226679e-01 -3.65255654e-01 2.56097496e-01
-8.57532382e-01 -3.09626125e-02 3.40864331e-01 8.37327123e-01
1.04740739e+00 -8.00924599e-02 -6.84423566e-01 8.53375375e-01
4.07749861e-01 1.02943912e-01 5.05303741e-01 -5.15812218e-01
5.28055608e-01 1.07294774e+00 -5.72584644e-02 -1.02992868e+00
-4.17173058e-01 -7.75643528e-01 -9.09721553e-01 3.13212365e-01
1.92622498e-01 1.44389644e-01 -1.09480214e+00 1.68889105e+00
8.14108774e-02 2.55980134e-01 -2.96341151e-01 1.14381790e+00
1.13896489e+00 4.22010183e-01 4.69169468e-01 3.76163244e-01
1.41869473e+00 -1.44331372e+00 -1.39174730e-01 -3.49186391e-01
2.16206238e-01 -5.71699023e-01 1.28238952e+00 -1.13003992e-01
-6.54440582e-01 -1.29078257e+00 -1.04218650e+00 -2.10379630e-01
-9.06002283e-01 4.30659890e-01 6.65191650e-01 1.04975179e-01
-9.62730467e-01 4.16728288e-01 -5.29287696e-01 -2.02789441e-01
8.32420170e-01 2.37747565e-01 -7.02822983e-01 -2.62047023e-01
-1.05608058e+00 5.64161718e-01 1.51298523e-01 2.92711973e-01
-9.91848767e-01 -8.80428314e-01 -8.90150368e-01 4.29866344e-01
1.05845250e-01 -6.71534300e-01 6.52845800e-01 -9.27621424e-01
-1.03627491e+00 8.29142809e-01 -1.80290729e-01 -1.67151727e-02
2.46774882e-01 -2.68576503e-01 -2.08332121e-01 4.81652468e-02
2.33746082e-01 8.66634071e-01 9.11001205e-01 -1.14515281e+00
-8.59632194e-01 -3.55538726e-01 3.05855244e-01 -4.36880626e-02
-1.21494621e-01 -1.69095680e-01 -6.08476698e-01 -8.99616838e-01
-1.53921610e-02 -6.03876829e-01 -1.00283787e-01 2.64004478e-03
-2.50510633e-01 -5.45065880e-01 9.30724859e-01 -4.22881871e-01
1.21359658e+00 -2.40903497e+00 8.43856856e-02 8.37956369e-02
2.81008899e-01 2.40730062e-01 -4.86346960e-01 2.88557649e-01
-3.64627093e-01 2.40143657e-01 -3.08674388e-02 -3.12695444e-01
1.60161376e-01 -9.37663540e-02 -4.31451797e-01 3.03916484e-01
6.04995430e-01 1.31460118e+00 -7.02054143e-01 -2.37215057e-01
4.26370233e-01 5.95996737e-01 -6.03501797e-01 2.78339148e-01
5.01539111e-02 2.14778796e-01 -5.91467023e-01 7.35011220e-01
6.35932148e-01 -3.45477045e-01 -3.03416431e-01 -7.15459764e-01
-2.54119039e-01 -1.12692431e-01 -1.08801901e+00 1.66780508e+00
-6.17980540e-01 2.61783779e-01 2.86500882e-02 -6.93594456e-01
8.71640921e-01 -1.69370204e-01 -3.47246788e-02 -8.65840733e-01
7.31406361e-02 -2.20178261e-01 -5.39632216e-02 -1.91839099e-01
2.96221823e-01 -4.81268764e-03 -3.96790475e-01 2.64863253e-01
3.43951017e-01 4.03869301e-01 -7.97529966e-02 1.77204803e-01
8.66033912e-01 9.85439345e-02 3.64384145e-01 -4.26608384e-01
7.67743945e-01 -2.93606490e-01 3.82780522e-01 8.05972159e-01
-3.96813035e-01 4.96386260e-01 2.59431094e-01 -6.82289839e-01
-5.77605128e-01 -1.07946718e+00 -3.41699086e-03 1.55264485e+00
3.42054009e-01 -5.46904922e-01 -5.60616255e-01 -9.95389581e-01
3.72666806e-01 1.58912793e-01 -1.41728008e+00 -2.11830914e-01
-4.13384050e-01 -2.94571608e-01 4.37621534e-01 1.04053020e+00
8.75585914e-01 -1.30162179e+00 -3.50736946e-01 4.41650748e-02
5.24218306e-02 -1.04314089e+00 -7.25877464e-01 3.22242916e-01
-5.08722723e-01 -1.20009398e+00 -4.05215532e-01 -7.46673465e-01
7.58156717e-01 4.23279822e-01 1.06638670e+00 2.31297627e-01
-5.15038729e-01 2.55126476e-01 -3.15594912e-01 -8.51824135e-02
3.98671538e-01 1.42917603e-01 -2.23766699e-01 3.76875967e-01
3.54194134e-01 -4.33518231e-01 -5.96603453e-01 4.57176685e-01
-6.27882481e-01 -3.15862410e-02 8.92638624e-01 1.23551762e+00
8.66616845e-01 -1.14029393e-01 5.45097411e-01 -7.18378901e-01
3.79843920e-01 -3.49660426e-01 -5.71865559e-01 3.15534979e-01
-2.28535503e-01 1.11129113e-01 7.40343332e-01 -2.90738583e-01
-9.26476777e-01 7.87775666e-02 -1.54875174e-01 -7.87968338e-01
-5.19430459e-01 2.57424474e-01 -3.51597190e-01 -3.68540108e-01
3.81582558e-01 1.77069068e-01 -6.11276150e-01 -7.29409158e-01
6.82376862e-01 6.01193070e-01 6.91845834e-01 -6.59920931e-01
7.79997230e-01 3.99838746e-01 -2.82794654e-01 -5.51591992e-01
-1.19505990e+00 -4.94221300e-01 -1.02130759e+00 1.41870707e-01
1.04456711e+00 -1.04141700e+00 -7.47009695e-01 3.80411744e-01
-7.21510410e-01 -2.73798287e-01 -3.86573106e-01 9.56381634e-02
-3.08538556e-01 -1.02686681e-01 -5.62043488e-01 -3.81800443e-01
-3.78272504e-01 -1.12592697e+00 1.49620652e+00 4.57352757e-01
5.89103885e-02 -7.97544956e-01 -2.27467969e-01 1.24764629e-01
6.41377211e-01 1.17362410e-01 9.55339253e-01 -3.05101097e-01
-7.15132773e-01 -5.96366217e-03 -9.48048890e-01 1.84215873e-01
2.42217883e-01 -2.25336626e-01 -1.26277840e+00 -4.26365495e-01
-6.39842391e-01 -4.28864241e-01 1.37237144e+00 2.87853181e-01
1.61789417e+00 -1.46646857e-01 -5.11566043e-01 9.16873574e-01
1.37638819e+00 -1.14012845e-01 4.79500264e-01 1.81604758e-01
9.39002097e-01 4.99522269e-01 6.33131444e-01 2.11120561e-01
3.93806905e-01 7.18300462e-01 4.96525586e-01 -3.61837894e-01
-6.04454339e-01 -4.28336054e-01 3.17326076e-02 3.78181458e-01
1.37006715e-02 2.49724343e-01 -5.44179738e-01 6.61410511e-01
-1.82429492e+00 -8.91864657e-01 3.91572356e-01 1.73789704e+00
4.26653594e-01 1.63012501e-02 -4.10346426e-02 -1.00494452e-01
6.13455057e-01 4.65502799e-01 -5.24671614e-01 -4.58568156e-01
-1.18606016e-02 3.48911941e-01 9.08930749e-02 4.63211060e-01
-1.50611663e+00 1.26432645e+00 5.37615919e+00 1.04636168e+00
-1.21170461e+00 6.75592795e-02 6.30851865e-01 9.06299949e-02
-2.12557260e-02 -3.30693394e-01 -8.54981303e-01 4.45527554e-01
3.31662863e-01 3.79117072e-01 4.13126349e-01 9.78614986e-01
-2.11262241e-01 2.10703403e-01 -1.12057686e+00 9.71160471e-01
-5.14483824e-02 -1.50638103e+00 3.02784711e-01 -1.20360151e-01
6.31963849e-01 9.08666700e-02 6.74448088e-02 6.53779626e-01
2.21287906e-01 -1.41992342e+00 7.79883862e-01 7.58670151e-01
1.22990119e+00 -9.02426541e-01 7.12483585e-01 9.57766324e-02
-1.91774166e+00 -3.04110438e-01 -5.30790448e-01 8.94647017e-02
-3.20931345e-01 4.33111578e-01 -2.99057454e-01 6.82104409e-01
1.06845403e+00 7.83501625e-01 -9.26059365e-01 8.37532222e-01
-3.65145892e-01 4.94640589e-01 -1.96835361e-02 2.76012570e-01
6.34980559e-01 2.64200836e-01 8.26079100e-02 1.64973629e+00
-1.40966430e-01 1.14048600e-01 4.53356087e-01 1.02997780e+00
-1.50903106e-01 -1.75437823e-01 -2.52996355e-01 5.83178625e-02
2.79486895e-01 1.59651220e+00 -6.40427053e-01 -3.20082873e-01
-4.20416683e-01 8.98173451e-01 8.74270856e-01 3.03997546e-01
-6.65362298e-01 -6.50638640e-01 1.09132493e+00 -1.86695419e-02
8.95248473e-01 -3.08044720e-02 -2.46407226e-01 -1.16202307e+00
-2.31474936e-01 -7.21829355e-01 6.07913792e-01 -5.96872568e-01
-1.67643011e+00 9.67649221e-01 -3.61363292e-01 -9.39085305e-01
1.24461614e-01 -6.85024321e-01 -7.47963965e-01 1.27225637e+00
-1.67642057e+00 -1.67960596e+00 -7.99628854e-01 9.48964775e-01
5.42603195e-01 -1.26675352e-01 8.96242380e-01 2.32923388e-01
-5.20851791e-01 8.06859791e-01 -4.43863839e-01 3.66136760e-01
6.15977466e-01 -1.01855516e+00 5.43963790e-01 7.66565979e-01
1.71155751e-01 8.87416780e-01 -2.28550494e-01 -5.11054814e-01
-1.05081356e+00 -1.47230995e+00 6.85959160e-01 -3.82099658e-01
5.98753095e-01 -6.23470306e-01 -8.54156077e-01 5.78238547e-01
-9.95382816e-02 7.65752494e-01 4.79623884e-01 2.82890379e-01
-9.84459400e-01 -2.91339338e-01 -9.97268438e-01 2.42685199e-01
1.23562622e+00 -1.00170815e+00 -9.00720954e-01 -2.29626000e-01
7.73255348e-01 -8.77638310e-02 -7.36234069e-01 5.13429165e-01
8.17773104e-01 -9.30240810e-01 1.17394125e+00 -8.01077127e-01
2.99330890e-01 -6.43903673e-01 -3.75413030e-01 -1.33305907e+00
-1.18395364e+00 1.08282089e-01 4.19135652e-02 1.25562632e+00
1.12128817e-01 -5.27945518e-01 4.20654744e-01 3.04552950e-02
-3.16572130e-01 -9.47772324e-01 -8.30513954e-01 -5.74882865e-01
1.56307980e-01 -4.62889165e-01 7.45010614e-01 8.49368155e-01
-3.73767644e-01 4.10067767e-01 8.62259045e-03 3.50018829e-01
4.34060484e-01 8.01474512e-01 5.25508523e-01 -1.21367478e+00
-2.43286833e-01 -5.74425578e-01 -3.92126173e-01 -1.08215117e+00
1.90369546e-01 -9.03584421e-01 3.56175229e-02 -1.55096138e+00
3.80429834e-01 -5.82628489e-01 -8.31647813e-01 9.29465890e-01
-3.72701675e-01 6.65157676e-01 5.36698580e-01 1.41857162e-01
-8.41543019e-01 5.65320015e-01 1.32478082e+00 -3.68746102e-01
1.76288977e-01 -3.86793911e-01 -1.15194738e+00 7.65080571e-01
2.96494842e-01 -2.05594525e-01 -2.03553736e-01 -3.24150294e-01
-3.94605011e-01 -4.60958213e-01 7.24247098e-01 -9.80309665e-01
2.74870336e-01 -8.56342465e-02 1.01019549e+00 -6.09624326e-01
1.05117053e-01 -6.87514722e-01 -2.12704375e-01 3.50922376e-01
-2.70672530e-01 -7.16949031e-02 4.89417642e-01 4.44356859e-01
-5.25483847e-01 3.43414754e-01 9.42874312e-01 -6.09511063e-02
-1.26621079e+00 6.45669937e-01 -2.69276239e-02 3.03267892e-02
8.83735657e-01 -7.02720834e-03 -6.02767646e-01 5.95205128e-02
-7.53169179e-01 1.64518327e-01 2.97081143e-01 8.33640635e-01
6.02383673e-01 -1.57319498e+00 -5.88813841e-01 7.37295568e-01
5.16421437e-01 -6.67130351e-02 5.56161463e-01 6.09728515e-01
2.92144399e-02 7.59694397e-01 -3.99919927e-01 -6.82657480e-01
-9.71062601e-01 8.14530134e-01 4.17349875e-01 -5.16978800e-01
-4.62560356e-01 1.20927322e+00 1.27390301e+00 -3.63955826e-01
1.99861363e-01 -5.85622132e-01 -3.86005819e-01 -7.47314245e-02
8.18420708e-01 4.85272743e-02 7.22918138e-02 -8.58438611e-01
-8.63830447e-01 1.02463806e+00 -1.23480290e-01 4.63111609e-01
1.28733265e+00 -1.87010348e-01 5.57787903e-02 1.14758082e-01
1.26030850e+00 -9.72004980e-03 -1.62109447e+00 -2.97249824e-01
-2.33565524e-01 -3.84994358e-01 1.91874132e-01 -1.27390778e+00
-1.15673077e+00 1.08357227e+00 7.66324282e-01 -3.06824178e-01
1.30868351e+00 2.40415961e-01 4.00658906e-01 1.08289719e-01
4.81814563e-01 -5.55563509e-01 4.66806889e-02 7.98344612e-01
1.25913978e+00 -1.18543124e+00 -2.45961294e-01 -3.73044103e-01
-4.30250138e-01 9.16687310e-01 8.71738732e-01 -4.19000030e-01
9.45875764e-01 3.62527162e-01 -1.99753605e-02 -2.55974859e-01
-8.52831662e-01 -4.64280039e-01 9.05770600e-01 5.07545769e-01
3.47249687e-01 1.86496809e-01 2.46953279e-01 1.36644733e+00
6.09486699e-02 -1.54602051e-01 -4.20254618e-01 6.50687575e-01
-5.01050353e-01 -8.09780896e-01 -2.28582904e-01 3.43902141e-01
-1.55540705e-01 -3.72199595e-01 -6.40130222e-01 8.71184111e-01
7.60579169e-01 8.52643013e-01 2.66453683e-01 -6.83667541e-01
5.58428645e-01 1.50650796e-02 4.09285843e-01 -6.13364995e-01
-1.00513089e+00 -1.27699107e-01 -1.36748999e-01 -1.15849018e+00
-1.60030827e-01 -2.85484284e-01 -1.06769097e+00 -1.33316353e-01
-5.14843576e-02 9.00681391e-02 2.13631704e-01 9.66112077e-01
6.79806411e-01 8.22630882e-01 5.74558258e-01 -1.08607280e+00
-5.45857660e-02 -1.02786541e+00 -4.69614029e-01 3.82343590e-01
5.23564994e-01 -1.08490551e+00 -1.69780239e-01 -1.82679981e-01] | [9.669601440429688, 2.0472490787506104] |
c3e72cb1-5b8d-4094-b725-eae8cc20455b | enhancing-representation-learning-on-high | 2306.15661 | null | https://arxiv.org/abs/2306.15661v1 | https://arxiv.org/pdf/2306.15661v1.pdf | Enhancing Representation Learning on High-Dimensional, Small-Size Tabular Data: A Divide and Conquer Method with Ensembled VAEs | Variational Autoencoders and their many variants have displayed impressive ability to perform dimensionality reduction, often achieving state-of-the-art performance. Many current methods however, struggle to learn good representations in High Dimensional, Low Sample Size (HDLSS) tasks, which is an inherently challenging setting. We address this challenge by using an ensemble of lightweight VAEs to learn posteriors over subsets of the feature-space, which get aggregated into a joint posterior in a novel divide-and-conquer approach. Specifically, we present an alternative factorisation of the joint posterior that induces a form of implicit data augmentation that yields greater sample efficiency. Through a series of experiments on eight real-world datasets, we show that our method learns better latent representations in HDLSS settings, which leads to higher accuracy in a downstream classification task. Furthermore, we verify that our approach has a positive effect on disentanglement and achieves a lower estimated Total Correlation on learnt representations. Finally, we show that our approach is robust to partial features at inference, exhibiting little performance degradation even with most features missing. | ['Nikola Simidjievski', 'Mateja Jamnik', 'Andrei Margeloiu', 'Navindu Leelarathna'] | 2023-06-27 | null | null | null | null | ['dimensionality-reduction', 'disentanglement'] | ['methodology', 'methodology'] | [ 1.03972785e-01 2.18775511e-01 -2.10456386e-01 -3.64776671e-01
-1.03646374e+00 -4.59766120e-01 9.44591343e-01 -2.84811378e-01
-3.84319961e-01 8.95602047e-01 4.10808563e-01 2.80398056e-02
-4.36519057e-01 -5.63277900e-01 -7.31793940e-01 -8.72898459e-01
-4.88920510e-02 4.99325186e-01 -3.56814802e-01 5.62397577e-02
-3.94074246e-02 4.01149929e-01 -1.60046136e+00 -5.91644049e-02
5.50521791e-01 8.34470153e-01 -1.44345269e-01 5.60901761e-01
3.47547621e-01 3.48610282e-01 -4.15613711e-01 -3.35292786e-01
3.85153532e-01 -1.15691908e-01 -6.15523398e-01 -2.87769847e-02
4.94285941e-01 -4.71287519e-01 -2.27788761e-01 8.94968987e-01
3.01279783e-01 2.53998607e-01 1.22733510e+00 -1.26520908e+00
-5.93338311e-01 4.14310008e-01 -4.30467069e-01 -3.02368831e-02
-1.23531938e-01 5.94439469e-02 1.46967018e+00 -8.79749238e-01
4.19318795e-01 1.35021555e+00 6.55013263e-01 5.65068185e-01
-1.90883362e+00 -6.13070607e-01 1.22239523e-01 -9.78658274e-02
-1.34223068e+00 -4.48061913e-01 7.55004346e-01 -4.20558721e-01
9.56514001e-01 1.22429065e-01 4.87523198e-01 1.61467695e+00
-7.67295212e-02 9.37821209e-01 1.14372194e+00 -2.31321231e-01
3.80135149e-01 2.24883899e-01 1.53710917e-01 4.48023826e-01
3.48135024e-01 1.68322101e-01 -4.72052604e-01 -4.32409227e-01
8.65748286e-01 1.41974136e-01 -1.93946823e-01 -7.95616865e-01
-1.11348033e+00 1.27680254e+00 1.75680533e-01 1.20577723e-01
-4.89839882e-01 2.44240105e-01 1.55632094e-01 1.87890023e-01
6.18158519e-01 5.55711746e-01 -6.82966173e-01 -1.40301600e-01
-8.98373783e-01 5.63957810e-01 9.15710151e-01 5.52981436e-01
6.85312271e-01 2.68513262e-02 -1.74236938e-01 8.14384818e-01
2.58803964e-01 3.74432176e-01 4.48672444e-01 -1.09901977e+00
3.77053857e-01 3.53720963e-01 1.43850505e-01 -6.51138306e-01
-3.02503109e-01 -6.55633628e-01 -9.69162405e-01 1.62813187e-01
4.09306049e-01 -2.33745083e-01 -8.54839504e-01 2.29519868e+00
2.99311101e-01 3.00077200e-01 2.33445540e-01 6.30688190e-01
9.46484432e-02 6.34875357e-01 1.03485540e-01 -2.43569151e-01
1.37752593e+00 -5.53618312e-01 -5.67093015e-01 -2.67135888e-01
6.17706776e-01 -1.38036415e-01 1.04302990e+00 6.92533851e-01
-9.03875947e-01 -3.45984280e-01 -1.09137011e+00 -3.22818518e-01
-2.47149244e-01 2.09465116e-01 8.94580364e-01 7.15166688e-01
-8.96756887e-01 9.33563232e-01 -1.07547092e+00 -2.25024838e-02
7.34736443e-01 5.54435730e-01 -4.65315968e-01 1.10006124e-01
-9.34728801e-01 7.84501076e-01 2.13583916e-01 -2.28248939e-01
-9.97700095e-01 -1.10209024e+00 -9.03824806e-01 3.82782936e-01
3.24863911e-01 -8.46372128e-01 9.82771337e-01 -5.73010325e-01
-1.58897448e+00 4.26410556e-01 -1.31171286e-01 -5.32503903e-01
4.23884898e-01 -5.07258952e-01 -1.93028934e-02 -9.87254530e-02
-2.20713578e-02 6.48939192e-01 1.24343979e+00 -1.06870198e+00
-2.90282190e-01 -4.44892466e-01 -1.20675974e-01 1.61278574e-03
-7.64676511e-01 -4.80911404e-01 1.14693277e-01 -4.94925082e-01
-8.14340040e-02 -9.47690368e-01 -2.52791882e-01 1.37206577e-02
-3.03082287e-01 -5.96338749e-01 7.05074608e-01 -3.73533756e-01
1.06854331e+00 -2.12735939e+00 7.06609726e-01 1.52289301e-01
5.63087344e-01 3.22320044e-01 -1.88925058e-01 3.99998277e-01
-1.25684813e-01 2.33665541e-01 -3.70801777e-01 -8.98838758e-01
3.77408743e-01 4.29517299e-01 -5.80339670e-01 4.41233218e-01
6.26590967e-01 8.71994019e-01 -6.72810853e-01 -1.69056222e-01
1.15642920e-02 7.39621520e-01 -9.71558571e-01 1.49863914e-01
-2.26893649e-01 3.38243395e-01 -4.46671963e-01 1.02631003e-01
4.39576864e-01 -4.76825267e-01 2.15580374e-01 -7.35620260e-02
2.91544616e-01 1.87790111e-01 -1.19604015e+00 1.71709394e+00
-4.00542080e-01 7.41141796e-01 -6.79988489e-02 -1.37567043e+00
6.28873587e-01 2.37932131e-01 3.45519483e-01 -2.17991054e-01
1.85958728e-01 2.74941530e-02 -6.91837892e-02 -3.21375847e-01
4.39369768e-01 -3.02256882e-01 -1.07065767e-01 5.39287865e-01
3.60664785e-01 6.44634385e-03 3.13923210e-02 1.40010357e-01
9.68552351e-01 2.94440866e-01 3.62007916e-01 -3.51767659e-01
1.69012994e-01 -3.37394089e-01 4.69307214e-01 8.19865167e-01
5.46794385e-02 5.22799551e-01 9.72713113e-01 -1.27364486e-01
-1.27580833e+00 -1.21978092e+00 -4.06992435e-01 7.25724816e-01
-5.35647452e-01 -4.94603992e-01 -7.25117803e-01 -6.96653008e-01
2.02182427e-01 8.88489425e-01 -9.56163049e-01 -3.32037210e-01
-3.01516056e-01 -1.20175231e+00 5.44235170e-01 7.01759815e-01
5.20981029e-02 -4.67205375e-01 -5.83535790e-01 2.35595182e-03
1.82795748e-01 -9.80749011e-01 4.79708910e-02 1.62316591e-01
-9.90229666e-01 -8.61707032e-01 -8.48844588e-01 -2.88108647e-01
4.38251525e-01 -1.29215971e-01 8.69737804e-01 -5.61676562e-01
-1.88858628e-01 3.66157740e-01 -4.56528738e-02 -3.31025213e-01
-2.16568768e-01 2.57712126e-01 4.58411247e-01 9.85123143e-02
4.76563871e-01 -8.98713231e-01 -5.86102188e-01 -3.72186229e-02
-8.18273127e-01 -1.80824801e-01 6.84542656e-01 1.25842881e+00
4.69894260e-01 -2.28801161e-01 6.71621084e-01 -8.19347501e-01
8.48290324e-01 -6.93351984e-01 -6.00038290e-01 -1.12611890e-01
-8.86119723e-01 6.59168780e-01 5.82419157e-01 -6.54367685e-01
-9.27956283e-01 -1.94407910e-01 4.73884270e-02 -7.83806801e-01
-5.45405075e-02 3.90112221e-01 6.41771704e-02 4.32683200e-01
7.25133300e-01 1.54133081e-01 3.86702567e-01 -7.63813078e-01
5.61450243e-01 5.50104380e-01 1.22067772e-01 -6.53842986e-01
7.73257911e-01 5.42638123e-01 2.74217635e-01 -8.90604436e-01
-9.48358238e-01 3.15583311e-02 -6.29593968e-01 3.98789018e-01
7.10064650e-01 -1.06859875e+00 -8.03646147e-01 5.94579875e-02
-7.48724699e-01 -3.43977839e-01 -4.92226839e-01 7.59450257e-01
-6.98469698e-01 1.46387711e-01 -4.05094892e-01 -9.70263898e-01
-1.24273159e-01 -1.10292315e+00 1.22103631e+00 -2.75804345e-02
-3.82498652e-01 -9.21278417e-01 2.76575983e-01 1.43676654e-01
3.01690131e-01 1.79744691e-01 1.19306409e+00 -8.66845250e-01
-5.36792278e-01 -1.47707596e-01 -1.90779865e-01 6.19050026e-01
-2.70427149e-02 -1.00053683e-01 -1.16366041e+00 -4.98620689e-01
-1.04080297e-01 -4.16200340e-01 1.19473755e+00 4.36094522e-01
1.24598682e+00 -3.20530266e-01 -2.49815673e-01 7.09258258e-01
1.22633994e+00 -3.02403361e-01 4.59060520e-01 -1.15934908e-01
6.63186371e-01 6.02058470e-01 2.57286906e-01 5.44557989e-01
1.28298670e-01 5.78191161e-01 2.29682580e-01 2.75514305e-01
2.31264010e-02 -2.21860781e-01 3.97637248e-01 6.76430643e-01
-3.51492851e-03 -3.43843214e-02 -5.45715332e-01 5.19808233e-01
-1.82220149e+00 -7.45088756e-01 8.86873454e-02 2.10251784e+00
8.52613330e-01 -1.67822376e-01 1.74048051e-01 1.97196901e-01
8.22975636e-02 2.48875365e-01 -6.99735701e-01 -1.95377260e-01
9.08541232e-02 5.38192451e-01 2.21767560e-01 2.89260685e-01
-9.83327687e-01 7.82133937e-01 7.09904146e+00 8.16114247e-01
-7.70096064e-01 1.79188177e-01 4.48677331e-01 -4.69384015e-01
-3.92486095e-01 -1.24291047e-01 -7.93949962e-01 3.37464720e-01
1.05602968e+00 1.60602143e-03 5.57136238e-01 9.24208879e-01
-1.61945581e-01 1.24761134e-01 -1.50308287e+00 1.03469181e+00
-1.77337602e-02 -1.32650125e+00 7.12650344e-02 4.84557271e-01
7.25145519e-01 -4.67496142e-02 4.40914690e-01 5.64183116e-01
3.81442219e-01 -1.14847505e+00 3.97173375e-01 5.77093959e-01
7.92271912e-01 -9.00712729e-01 3.62643957e-01 4.25381035e-01
-5.38527966e-01 -2.19378233e-01 -4.61634129e-01 -1.97852597e-01
-1.36934355e-01 5.95770717e-01 -6.63871408e-01 3.88694018e-01
4.10055935e-01 6.76411331e-01 -3.09440434e-01 5.01330853e-01
-1.62016511e-01 6.15384698e-01 -4.03736562e-01 -2.23706335e-01
1.11974381e-01 -2.25406826e-01 6.13268077e-01 7.98618257e-01
2.95029163e-01 -6.54557422e-02 -2.62472302e-01 1.27243316e+00
-2.04759374e-01 -2.10992336e-01 -6.61571324e-01 -3.18367422e-01
3.44167531e-01 1.06750476e+00 -1.21075593e-01 -3.50249201e-01
-2.95205891e-01 7.84579992e-01 6.72225535e-01 5.54181159e-01
-5.83746791e-01 -2.18575463e-01 1.14889169e+00 -3.25059205e-01
5.67701399e-01 -3.07878822e-01 -1.28873631e-01 -1.45360458e+00
-1.79703385e-01 -7.93768585e-01 2.03013003e-01 -3.77904564e-01
-1.58463025e+00 1.68492362e-01 4.64626759e-01 -9.77359295e-01
-7.12603629e-01 -9.05582190e-01 -2.90672362e-01 8.87541354e-01
-1.10012043e+00 -8.85691941e-01 1.96396500e-01 4.11902040e-01
2.73529023e-01 -2.82969773e-01 1.07213223e+00 1.91776559e-01
-7.23139465e-01 7.38859475e-01 4.36867028e-01 -7.12668672e-02
4.07587349e-01 -1.28686523e+00 1.10961452e-01 6.33180976e-01
3.56846869e-01 8.09737027e-01 7.70100594e-01 -3.63763481e-01
-1.49432063e+00 -8.88840675e-01 4.17598337e-01 -7.60690033e-01
6.48454309e-01 -6.24580443e-01 -9.10501599e-01 1.03359962e+00
-1.54285505e-01 2.64196172e-02 8.74336481e-01 6.74564898e-01
-6.58234775e-01 4.91643362e-02 -1.03656697e+00 6.10956192e-01
1.08024466e+00 -6.34607017e-01 -8.21163118e-01 1.98891357e-01
5.98382354e-01 -6.91553056e-02 -1.07044661e+00 2.14434564e-01
8.34335804e-01 -8.44802499e-01 1.08406925e+00 -1.02851379e+00
6.55570447e-01 1.06131613e-01 -3.54561210e-01 -1.45811892e+00
-3.43635559e-01 -4.62923378e-01 -6.28609121e-01 1.15950656e+00
2.95796484e-01 -8.45112026e-01 8.52061391e-01 5.55135489e-01
2.49160379e-01 -9.22952235e-01 -9.75648165e-01 -8.18766892e-01
5.09470701e-01 -3.88052523e-01 5.21076679e-01 8.75470340e-01
-1.77774802e-01 5.65257013e-01 -6.07625723e-01 1.08758345e-01
9.55684543e-01 -2.31977385e-02 9.61262345e-01 -1.55158734e+00
-5.51810503e-01 -4.81543094e-01 -4.22141582e-01 -1.04716992e+00
4.48420942e-01 -9.17160571e-01 -3.93596619e-01 -1.16221714e+00
3.50159377e-01 -1.60172448e-01 -1.90727964e-01 3.27469647e-01
-1.72401324e-01 1.24650113e-01 2.85166185e-02 1.85190186e-01
-2.87115395e-01 9.85755265e-01 9.93961990e-01 2.86332984e-02
-1.21111788e-01 -2.34648332e-01 -8.52853239e-01 4.80444193e-01
5.92145801e-01 -6.36764467e-01 -4.88181084e-01 -2.25478292e-01
9.98990610e-02 -1.95663154e-01 5.13843060e-01 -7.73273647e-01
-1.82702616e-01 6.95164129e-02 7.23259091e-01 -3.54677647e-01
8.20696771e-01 -5.13784647e-01 4.63779680e-02 3.06233734e-01
-5.21135747e-01 -4.31392223e-01 9.01051611e-02 8.95543516e-01
1.56907320e-01 -1.39893875e-01 5.85495651e-01 1.34314373e-01
-1.15661554e-01 3.66678864e-01 -2.15175867e-01 5.02940007e-02
8.71848285e-01 1.44312024e-01 -1.40373155e-01 -3.09817672e-01
-7.33398855e-01 1.56660289e-01 1.32775933e-01 3.28438789e-01
3.38754922e-01 -1.34535849e+00 -7.50003159e-01 6.34562552e-01
-8.25011730e-03 1.50906080e-02 2.10117415e-01 8.41636539e-01
1.12497069e-01 4.51048464e-01 -5.62992170e-02 -7.12569237e-01
-8.90827119e-01 3.97827506e-01 1.13696471e-01 -3.99101287e-01
-6.65030360e-01 8.25473607e-01 4.13034141e-01 -3.10472399e-01
4.08502370e-02 -2.79779464e-01 -8.67058784e-02 4.85127896e-01
4.94758606e-01 4.12662774e-01 -2.08076596e-01 -3.66872430e-01
-1.11370996e-01 3.54055136e-01 -3.27474117e-01 -2.81344771e-01
1.63340998e+00 3.07806283e-02 2.18159735e-01 4.86899704e-01
1.58124924e+00 -2.27531254e-01 -1.84098029e+00 -2.67953128e-01
-1.57234728e-01 -4.27406937e-01 2.82158345e-01 -3.65935653e-01
-7.69700348e-01 1.07560277e+00 4.67231274e-01 3.90102953e-01
6.69788301e-01 1.71995685e-01 4.67433542e-01 4.63405550e-01
1.68132693e-01 -8.98149431e-01 5.75812459e-02 3.24506998e-01
7.37666845e-01 -1.20331800e+00 1.16006017e-01 2.45858580e-02
-4.89280760e-01 8.49800169e-01 7.73517564e-02 -4.98735398e-01
6.61877811e-01 8.76909588e-03 -5.05909741e-01 -1.92980692e-01
-1.02157342e+00 -3.83502916e-02 3.95543873e-01 5.88235557e-01
1.15137473e-01 6.31897748e-02 -2.34721247e-02 8.33682954e-01
-4.00581300e-01 -2.17318982e-01 2.85273165e-01 5.22112727e-01
-1.36721373e-01 -1.13705277e+00 -7.45564923e-02 7.67064691e-01
-5.73317349e-01 -3.89426835e-02 -7.41888285e-02 1.02327859e+00
-3.44653130e-01 5.96471190e-01 2.73395538e-01 -2.38982320e-01
8.93482342e-02 4.65150774e-01 6.77397788e-01 -5.23876905e-01
-5.46616688e-02 6.82590231e-02 -1.04398377e-01 -7.34893143e-01
-2.82358140e-01 -1.01712370e+00 -7.37194777e-01 -1.50420517e-01
-2.07209766e-01 1.04613192e-01 6.05664790e-01 1.08390260e+00
5.98642528e-01 6.09818041e-01 4.25765514e-01 -1.12778330e+00
-1.26227641e+00 -1.04097819e+00 -7.18562722e-01 5.01420796e-01
7.68187523e-01 -1.01590264e+00 -6.24958277e-01 -2.79854864e-01] | [7.870304107666016, 3.922283172607422] |
c3fb6f60-379e-4186-b05f-a7ba15395511 | neuroimaging-feature-extraction-using-a | 2207.10794 | null | https://arxiv.org/abs/2207.10794v1 | https://arxiv.org/pdf/2207.10794v1.pdf | Neuroimaging Feature Extraction using a Neural Network Classifier for Imaging Genetics | A major issue in the association of genes to neuroimaging phenotypes is the high dimension of both genetic data and neuroimaging data. In this article, we tackle the latter problem with an eye toward developing solutions that are relevant for disease prediction. Supported by a vast literature on the predictive power of neural networks, our proposed solution uses neural networks to extract from neuroimaging data features that are relevant for predicting Alzheimer's Disease (AD) for subsequent relation to genetics. Our neuroimaging-genetic pipeline is comprised of image processing, neuroimaging feature extraction and genetic association steps. We propose a neural network classifier for extracting neuroimaging features that are related with disease and a multivariate Bayesian group sparse regression model for genetic association. We compare the predictive power of these features to expert selected features and take a closer look at the SNPs identified with the new neuroimaging features. | ['Farouk S. Nathoo', 'Mirza Faisal Beg', 'Leno Rocha', 'Jiguo Cao', 'Michelle F. Miranda', 'Erin Gibson', 'Sidi Wu', 'Cédric Beaulac'] | 2022-07-08 | null | null | null | null | ['disease-prediction'] | ['medical'] | [ 5.4059005e-01 1.8938307e-01 -1.1918078e-01 -9.7822195e-01
-6.2547511e-01 2.6513807e-02 2.5326023e-01 2.3145869e-01
-4.6038389e-01 7.1128774e-01 4.9482775e-01 -1.7170437e-01
-7.9573244e-01 -5.2898997e-01 -3.0127102e-01 -3.0472389e-01
-5.4811686e-01 3.9508301e-01 -2.8196126e-01 4.2334509e-01
2.2862612e-01 4.8073462e-01 -1.4740931e+00 4.2971039e-01
1.0105203e+00 1.0937504e+00 -5.6125461e-03 4.2107734e-01
1.0821758e-01 3.7769285e-01 -1.6922630e-01 -3.0004075e-01
4.3384816e-02 -3.3028984e-01 -3.5755992e-01 -2.8967130e-01
4.4314778e-01 -1.9811963e-01 -3.5259802e-02 1.0273416e+00
8.6675268e-01 -2.6310650e-01 9.9441820e-01 -9.4005907e-01
-8.1904912e-01 5.8904624e-01 -6.0552835e-01 6.4315259e-01
1.0412479e-01 2.7637967e-01 8.6039931e-01 -6.8545955e-01
8.7115020e-01 1.1643171e+00 8.1309313e-01 3.5029936e-01
-1.3702358e+00 -3.7723306e-01 -5.0135471e-02 6.3037729e-01
-1.2296976e+00 -6.5361953e-01 3.5325512e-01 -1.0036446e+00
9.7605926e-01 6.7670248e-02 8.6076647e-01 1.0143312e+00
8.2499005e-02 -2.2321135e-02 1.1771886e+00 -4.0457463e-01
4.2091164e-01 -7.2699584e-02 6.2108469e-01 8.5929805e-01
2.8165174e-01 2.8553421e-02 -4.8859617e-01 -5.7306105e-01
3.8953304e-01 -1.1111955e-01 1.6425641e-01 1.3950048e-01
-9.5802259e-01 9.4355726e-01 9.3384154e-02 1.5995690e-01
-8.4246057e-01 -5.0254598e-02 2.7645770e-01 5.4732908e-02
7.5305045e-01 2.5969738e-01 -5.4612845e-01 2.5585303e-01
-1.0818733e+00 1.4330128e-01 4.1736206e-01 1.6067439e-01
5.7682192e-01 5.4636102e-02 -2.6640174e-01 1.2079314e+00
7.4974793e-01 1.8859093e-01 8.3986050e-01 -8.4466684e-01
-1.6477821e-02 6.8293840e-01 -6.6744417e-01 -9.5712709e-01
-9.8941159e-01 -2.0126362e-01 -6.8753719e-01 1.6891322e-01
1.0553473e-01 -4.7178051e-01 -5.9568512e-01 1.6873457e+00
5.5148345e-01 4.0714979e-01 -1.4823540e-01 8.1966293e-01
1.0464380e+00 8.0898963e-02 5.2848345e-01 -1.6093107e-02
1.7723783e+00 -3.6256841e-01 -3.0112508e-01 -2.3845768e-01
7.2525549e-01 -1.1925205e-01 4.4916689e-01 1.7110287e-01
-8.1476951e-01 -3.2445592e-01 -7.3671764e-01 -1.4098933e-01
-3.2159185e-01 5.5189329e-01 7.9253769e-01 8.1347865e-01
-1.0728350e+00 4.0252072e-01 -6.8222731e-01 -5.5648601e-01
8.8290292e-01 5.4699969e-01 -5.2346921e-01 -7.9637349e-02
-1.0141910e+00 1.1872176e+00 3.3085299e-01 2.3520759e-01
-4.8781994e-01 -1.0474900e+00 -5.2959239e-01 2.1815324e-02
1.0229698e-01 -1.0654286e+00 5.3483897e-01 -6.8767500e-01
-1.1525581e+00 8.4932882e-01 -2.1729238e-01 -6.5571606e-01
9.4971508e-03 -1.0771185e-01 -6.6355890e-01 4.6935987e-01
1.5669590e-01 8.0957800e-01 8.3332551e-01 -2.5904480e-01
-4.3087021e-01 -1.0884373e+00 -8.1707257e-01 -2.0830180e-01
-2.7550039e-01 3.0813059e-01 3.2387847e-01 -4.3030703e-01
-1.1282482e-01 -6.6938376e-01 -3.5941884e-01 1.8277445e-01
-4.9525699e-01 -2.2104293e-01 5.9192544e-01 -1.0516478e+00
7.5139040e-01 -1.8379247e+00 2.4757698e-02 4.8143595e-01
6.5621710e-01 1.4296153e-01 -3.3525586e-01 -6.0358819e-02
-5.4817426e-01 3.2220602e-01 -4.4250332e-02 1.5485358e-01
-1.0134494e-01 -3.0463630e-01 2.1281840e-01 4.4756114e-01
8.8179755e-01 1.0873988e+00 -3.9403912e-01 -1.7792097e-01
-1.5238939e-01 6.3198805e-01 -8.5349089e-01 1.3534683e-01
-2.0777594e-01 4.2859831e-01 -7.0494407e-01 5.9042865e-01
4.4982442e-01 -1.7154503e-01 4.4632018e-01 -3.0829218e-01
1.6736861e-02 2.4095756e-01 -5.3271908e-01 1.3231835e+00
1.4127921e-01 4.1256154e-01 -4.9651291e-02 -1.2263575e+00
1.0974184e+00 1.4788212e-01 5.1292801e-01 -4.5725390e-01
1.6359191e-01 4.1966386e-02 6.7607445e-01 -1.1309679e+00
-4.5905837e-01 -1.6926974e-02 5.7367700e-01 4.9309853e-01
2.0204368e-01 5.6341225e-01 1.1746775e-02 -1.2755910e-01
1.4348480e+00 2.5473875e-01 5.7221895e-01 -1.0064935e-01
4.3258476e-01 -7.3476240e-02 7.2506016e-01 4.6272042e-01
-1.5690739e-01 2.9472759e-01 8.4822375e-01 -4.8961818e-01
-1.0379399e+00 -5.0280207e-01 -2.4048431e-01 6.6719848e-01
-1.0500011e+00 -4.0606245e-01 -6.0675341e-01 -6.0180598e-01
1.0336142e-01 3.8532490e-01 -8.3793604e-01 -3.7820706e-01
4.5072362e-03 -1.3316386e+00 6.5976858e-01 2.9172444e-01
5.4771997e-02 -7.2403836e-01 -6.9022787e-01 -8.7566495e-02
4.6810099e-01 -9.1409659e-01 2.1147114e-01 1.4184664e-01
-1.0294770e+00 -1.2584639e+00 -4.6612215e-01 -4.1166481e-01
5.7893175e-01 -3.3196342e-01 6.5572631e-01 -1.4646642e-01
-1.0052241e+00 3.4467342e-01 -2.2993977e-01 -6.5838999e-01
-1.8639797e-01 2.0000033e-02 -1.1007287e-02 9.0435967e-02
7.7859187e-01 -8.4223366e-01 -7.0782393e-01 -3.7560070e-01
-6.6352320e-01 -1.8429074e-01 1.0222852e+00 6.2470120e-01
5.5999863e-01 -9.1241360e-02 8.7247151e-01 -7.2921687e-01
6.1231250e-01 -1.1511762e+00 -4.2711803e-01 1.1422227e-01
-6.0008502e-01 -8.2863592e-02 9.3262196e-02 -2.0295884e-01
-9.2122185e-01 3.6161065e-01 -1.5403859e-01 -4.0142611e-03
-8.6989689e-01 1.0325501e+00 -1.0152589e-01 9.3506232e-02
4.8349655e-01 -2.9438165e-01 4.7069722e-01 -7.3051143e-01
2.9237446e-01 5.1294553e-01 -7.3084041e-02 -3.5183293e-01
4.9612015e-02 2.3032174e-01 4.6247903e-01 -9.9955010e-01
-7.6039201e-01 -1.6520412e-01 -9.8448104e-01 -5.0739862e-02
1.3000550e+00 -7.6603532e-01 -6.7843485e-01 9.4545245e-02
-8.5412878e-01 6.5428406e-02 9.0283267e-02 9.7374046e-01
-2.5072020e-01 3.2280868e-01 -2.2235194e-01 -7.1569353e-01
-4.7000688e-01 -1.2297544e+00 9.4463813e-01 7.5558953e-02
-6.2836462e-01 -8.6459970e-01 5.2173346e-01 5.8450723e-01
2.9657352e-01 1.3076289e-01 1.5517292e+00 -1.1987699e+00
-6.8515100e-02 -1.6902813e-01 -4.5732546e-01 1.1265075e-01
-1.3438499e-01 -1.3842748e-01 -9.4960010e-01 5.1099187e-01
1.2979680e-02 -2.3819228e-01 9.7114009e-01 7.9207784e-01
1.0087125e+00 -1.4594954e-01 -1.9859175e-01 4.8596323e-01
1.1653575e+00 2.4132767e-01 5.6782514e-01 1.4070605e-01
6.4810538e-01 1.0202601e+00 1.6851777e-03 3.4749490e-01
3.2154417e-01 7.2104853e-01 9.1406599e-02 4.0124673e-02
-8.6793616e-02 3.5511938e-01 3.9434731e-01 3.2120588e-01
-1.9655207e-01 3.3459815e-01 -1.0591767e+00 1.3609158e-01
-1.6802686e+00 -9.1453665e-01 -5.4343402e-01 1.7472920e+00
6.5191108e-01 -4.4507647e-01 4.6076128e-01 -4.2638209e-01
8.6058128e-01 -2.2915111e-01 -6.7598796e-01 -4.2171249e-01
-1.7244554e-01 5.4299635e-01 -2.5537515e-02 -1.1412910e-01
-9.4732237e-01 4.8005104e-01 7.4435887e+00 3.9293401e-02
-8.0451751e-01 2.3473026e-01 9.8106354e-01 -4.0828311e-01
-1.5491593e-02 -1.8638724e-01 -6.6913700e-01 3.6979431e-01
1.5960244e+00 2.0283255e-01 2.6182157e-01 7.0055991e-01
7.3264605e-01 -1.9540942e-01 -1.0985039e+00 6.2335521e-01
1.3568118e-01 -1.2516177e+00 -4.7112252e-03 3.7434730e-01
3.2579917e-01 2.2081777e-01 3.1936679e-02 -3.5313699e-01
1.1468137e-01 -1.1102533e+00 7.0776545e-02 1.1425859e+00
6.0198188e-01 -5.0141394e-01 5.9982252e-01 -1.9121966e-01
-4.0648824e-01 -1.7441379e-01 -5.1817679e-01 -1.0521384e-01
-1.3101408e-03 1.3779205e+00 -1.3020434e+00 2.5154141e-01
6.1042118e-01 9.0234488e-01 -7.6003206e-01 1.2031434e+00
-8.1814537e-03 7.7852929e-01 3.4097895e-02 1.5177806e-01
-9.3103804e-02 -5.3438646e-01 3.9603448e-01 1.0010984e+00
6.8190360e-01 7.9467461e-02 -7.5376615e-02 1.5372491e+00
2.7603748e-01 4.7512811e-01 -5.8424765e-01 -4.6212599e-01
1.8038845e-01 1.4990140e+00 -7.0085287e-01 -2.8152588e-01
-5.5254114e-01 4.1242626e-01 5.3072345e-01 8.7240584e-02
-3.2962409e-01 -1.4805152e-01 6.6523051e-01 -1.4709850e-01
3.2539862e-01 6.8606138e-02 -3.8430771e-01 -8.2206231e-01
-1.6896583e-01 -8.0291915e-01 4.0443251e-01 -7.5422210e-01
-1.6727531e+00 3.2610467e-01 -2.1553446e-01 -4.1274819e-01
-2.8577927e-01 -7.2678071e-01 -7.0433766e-01 1.1665679e+00
-1.3033924e+00 -1.1408134e+00 3.8670527e-03 1.3432771e-01
-8.9532465e-02 -7.1473855e-01 1.0766145e+00 2.8984055e-01
-1.1146125e+00 -9.5413961e-02 -6.0942478e-02 -9.2725746e-02
6.6472948e-01 -9.3595463e-01 1.4209922e-01 4.6055663e-01
-7.9225183e-02 6.9879317e-01 1.0910210e-01 -1.1690425e+00
-1.2600812e+00 -1.2037045e+00 8.1561267e-01 -1.7155893e-01
8.5550481e-01 -3.2055683e-02 -7.6274401e-01 5.7097268e-01
-1.6335301e-01 -1.2910035e-01 1.4942890e+00 3.4919018e-01
-3.9088625e-01 1.0131793e-01 -1.1896597e+00 3.1172138e-01
9.3971449e-01 -4.6802801e-01 -6.8505454e-01 3.2051340e-01
6.4833120e-02 5.7865268e-01 -1.1482794e+00 1.8137962e-01
8.9596051e-01 -9.3509358e-01 1.0991087e+00 -1.1233094e+00
6.5409380e-01 2.9279443e-02 6.7066476e-02 -1.4297411e+00
-5.0652540e-01 1.4353009e-02 1.4010586e-01 1.0960143e+00
5.0404459e-01 -6.6851360e-01 5.9212917e-01 8.3547956e-01
1.4284016e-01 -6.1139899e-01 -8.9201766e-01 -2.6185831e-01
-9.8691858e-02 -4.6712878e-01 3.8823813e-01 9.1740435e-01
-1.7916262e-02 4.6622443e-01 -2.4399863e-01 1.8298538e-01
5.3172880e-01 -2.5614706e-01 1.1457630e-01 -1.8905846e+00
-2.0882846e-01 -4.9162909e-01 -7.9892051e-01 3.1021628e-01
5.6032640e-01 -1.4088316e+00 -4.2742041e-01 -1.4446317e+00
5.9812617e-01 -1.1665365e-01 -2.2045407e-01 4.5614305e-01
-1.6207303e-01 3.0609056e-01 -2.5437105e-01 -2.5274810e-01
-7.8991443e-02 2.2569948e-01 5.1102114e-01 -9.8432481e-02
-3.2183918e-01 -3.4279868e-01 -9.7485220e-01 7.9675651e-01
1.0556136e+00 -5.8354980e-01 -4.1356665e-01 -2.7890968e-01
1.8709938e-01 -2.6746744e-01 7.4290371e-01 -8.8979411e-01
-1.4731544e-01 2.1931542e-01 9.0718424e-01 -2.8795543e-01
3.6344707e-01 -4.8442939e-01 1.3284065e-01 5.4173356e-01
-6.2798584e-01 -1.5175106e-02 -1.1071931e-01 4.7490051e-01
2.7867186e-01 -5.4931450e-01 4.6428141e-01 -1.3817583e-01
-5.0715983e-01 1.8445906e-01 -7.6096302e-01 -3.0459332e-01
8.2424384e-01 -2.6986490e-03 -3.4249091e-01 -4.7037065e-02
-1.4877322e+00 -6.8172954e-02 -3.0250818e-01 3.8827278e-02
6.8282717e-01 -1.0222952e+00 -9.7347724e-01 1.8858160e-01
-1.7822629e-01 -8.5101354e-01 5.9346426e-01 1.3616697e+00
-7.6829299e-02 4.3184358e-01 -5.8017707e-01 -3.2179794e-01
-1.4887708e+00 4.9661401e-01 1.6508563e-01 1.4304018e-01
-6.5519327e-01 5.5015826e-01 -1.7553458e-01 -1.2290271e-01
-2.0591673e-03 -1.2875691e-01 -7.7395928e-01 6.0746056e-01
8.3536947e-01 6.6662806e-01 1.7441277e-01 -5.2872938e-01
-3.4977710e-01 2.5517005e-01 -1.4634936e-01 -4.4937320e-02
2.1356468e+00 9.1605656e-02 -6.7199528e-01 4.6232089e-01
1.0049362e+00 -3.0628699e-01 -5.3688461e-01 1.8864462e-02
5.0834906e-01 -8.9745082e-02 3.0263636e-01 -1.0451840e+00
-1.1013644e+00 9.4218171e-01 1.2495764e+00 7.5778291e-03
7.0036399e-01 1.6149755e-01 2.0324588e-01 4.6491709e-01
-2.6909950e-01 -9.8329544e-01 -4.2174110e-01 3.5205737e-01
7.1679527e-01 -1.0604659e+00 2.5939801e-01 -4.4393471e-01
-3.2467619e-01 1.3709846e+00 4.2078874e-01 -2.1332698e-01
8.3713478e-01 1.8343854e-01 -8.9823812e-02 -8.1019676e-01
-9.4301790e-01 -3.0223373e-01 8.1933922e-01 1.1405374e+00
6.5485275e-01 1.3069855e-02 -6.7097747e-01 1.0005740e+00
-4.4856545e-02 2.6696435e-01 4.1912038e-02 4.1189003e-01
-4.9183354e-01 -1.1959567e+00 -3.7820616e-01 1.4302131e+00
-7.1369714e-01 -3.0524212e-01 -9.1130000e-01 1.6277224e-01
4.8543113e-01 6.9119626e-01 1.8019824e-01 -3.1874247e-02
-1.6883855e-01 7.3509467e-01 2.4521519e-01 -8.4784085e-01
-4.2008603e-01 2.5909603e-01 5.6331527e-01 -6.0418671e-01
-7.3618567e-01 -1.0707657e+00 -9.3350405e-01 5.3617781e-01
1.7272484e-01 -3.1702662e-01 8.7309146e-01 1.1774495e+00
1.0540826e+00 7.2352892e-01 3.8577210e-02 -8.7350971e-01
-1.7423518e-01 -1.0006320e+00 -7.6844341e-01 -1.4160131e-01
-1.6319539e-01 -6.5970737e-01 2.9180545e-01 1.9701144e-01] | [6.681713581085205, 5.518716812133789] |
1e415517-0899-46f5-9946-5741e58a36f5 | voice-and-accompaniment-separation-in-music | 2003.08954 | null | https://arxiv.org/abs/2003.08954v1 | https://arxiv.org/pdf/2003.08954v1.pdf | Voice and accompaniment separation in music using self-attention convolutional neural network | Music source separation has been a popular topic in signal processing for decades, not only because of its technical difficulty, but also due to its importance to many commercial applications, such as automatic karoake and remixing. In this work, we propose a novel self-attention network to separate voice and accompaniment in music. First, a convolutional neural network (CNN) with densely-connected CNN blocks is built as our base network. We then insert self-attention subnets at different levels of the base CNN to make use of the long-term intra-dependency of music, i.e., repetition. Within self-attention subnets, repetitions of the same musical patterns inform reconstruction of other repetitions, for better source separation performance. Results show the proposed method leads to 19.5% relative improvement in vocals separation in terms of SDR. We compare our methods with state-of-the-art systems i.e. MMDenseNet and MMDenseLSTM. | ['Yuzhou Liu', 'Balaji Thoshkahna', 'Trausti Kristjansson', 'Ali Milani'] | 2020-03-19 | null | null | null | null | ['music-source-separation'] | ['music'] | [ 3.47811095e-02 -4.42010999e-01 -1.74431447e-02 6.13718899e-03
-5.90646684e-01 -4.32478249e-01 3.94925568e-03 -4.04504985e-01
-2.32682273e-01 4.04273421e-01 5.12033284e-01 1.04969330e-01
-1.13929711e-01 -4.34675753e-01 -6.50712907e-01 -6.84941113e-01
7.42099360e-02 -1.60078034e-01 -2.09548287e-02 -1.89805686e-01
3.44828591e-02 4.01868761e-01 -1.43308747e+00 2.88601100e-01
9.50472713e-01 1.03271472e+00 2.87719578e-01 6.15615070e-01
-1.92217037e-01 7.92979121e-01 -6.87322080e-01 -2.08059028e-01
1.95538044e-01 -8.94841313e-01 -2.89170951e-01 -1.42682850e-01
3.29026729e-01 -2.64246672e-01 -6.04045570e-01 1.18833375e+00
8.91436696e-01 2.96457350e-01 2.92074800e-01 -8.49465549e-01
-5.65181851e-01 1.29994798e+00 -7.34514236e-01 4.48063046e-01
-7.64230192e-02 4.59058546e-02 1.15111709e+00 -7.25240648e-01
3.66303585e-02 1.01901281e+00 7.46537864e-01 3.26885730e-01
-1.07207310e+00 -1.33460605e+00 -2.41832212e-02 5.45388222e-01
-1.43017030e+00 -8.25335920e-01 1.18193817e+00 -1.74419999e-01
7.66828120e-01 2.29136333e-01 6.69720829e-01 9.75538254e-01
1.16889179e-02 9.88030970e-01 6.55402005e-01 -3.03979188e-01
-6.70423955e-02 -3.77053350e-01 -1.11815467e-01 1.15660839e-01
-1.29126370e-01 -7.64598250e-02 -7.31612146e-01 3.77511650e-01
1.03222263e+00 -7.47035146e-02 -5.02767205e-01 2.76356369e-01
-1.12687063e+00 5.06157577e-01 7.80516028e-01 7.63145387e-01
-4.49788988e-01 3.59244645e-01 3.61176461e-01 2.11720258e-01
3.18030685e-01 6.19836271e-01 -2.57134408e-01 -3.03638965e-01
-1.38692307e+00 2.62270533e-02 4.25201327e-01 7.59905159e-01
1.92374527e-01 7.79607177e-01 -3.58676225e-01 1.24060714e+00
-5.86891314e-03 3.02498132e-01 9.24283445e-01 -7.85931170e-01
4.85576957e-01 2.13183761e-01 -3.04832190e-01 -8.68497431e-01
-2.91588873e-01 -1.36768866e+00 -1.35302937e+00 -1.30450174e-01
3.10727786e-02 -2.55249202e-01 -7.61353970e-01 1.86498117e+00
-1.58787325e-01 8.21955919e-01 -4.09600213e-02 1.08893001e+00
9.41490471e-01 7.76738763e-01 -4.12434876e-01 -1.98925093e-01
1.14074504e+00 -1.29651380e+00 -7.90058017e-01 -6.18490092e-02
-8.01538080e-02 -9.11348879e-01 7.50761151e-01 6.09479904e-01
-1.21954417e+00 -1.10840046e+00 -1.05695450e+00 -5.93419485e-02
2.30443329e-01 5.10427117e-01 2.88278610e-01 3.87220383e-01
-8.31934512e-01 9.80911076e-01 -6.16675019e-01 4.49826270e-02
5.12111425e-01 3.00111115e-01 3.61423045e-02 3.45985949e-01
-1.10936630e+00 1.54270083e-01 3.35898846e-01 2.91809350e-01
-9.01225328e-01 -7.14819074e-01 -6.06383145e-01 6.13217890e-01
2.67693073e-01 -4.56072718e-01 1.24939513e+00 -1.17469215e+00
-1.78076756e+00 3.88648301e-01 -2.43061468e-01 -6.81940615e-01
1.84128121e-01 -6.50305510e-01 -7.71724164e-01 -1.26664057e-01
-2.56780740e-02 4.66375083e-01 1.11003268e+00 -8.16160262e-01
-6.31906092e-01 -1.45294771e-01 -1.67703524e-01 3.01155031e-01
-3.76932532e-01 7.18028173e-02 -6.30366504e-01 -1.33639956e+00
2.17135131e-01 -9.65160429e-01 1.26094386e-01 -7.06819713e-01
-7.40694761e-01 -3.67643237e-02 5.98710001e-01 -8.53914022e-01
1.53629303e+00 -2.62679362e+00 4.42846596e-01 -8.07569847e-02
2.03760386e-01 5.54582655e-01 -3.63946319e-01 8.83298442e-02
-2.41171107e-01 -1.45477718e-02 -2.01260164e-01 -5.07882953e-01
-5.52648529e-02 -2.17577755e-01 -5.05360961e-01 2.42464572e-01
1.80420816e-01 8.05425882e-01 -6.79518044e-01 -1.19052596e-01
1.03486583e-01 5.03252089e-01 -5.53635597e-01 6.63803220e-02
-1.72811206e-02 7.58706927e-01 -8.32819100e-03 3.34802181e-01
5.84735215e-01 3.90984267e-02 -1.77974682e-02 -3.84607285e-01
-2.78761983e-01 6.80741370e-01 -1.14729559e+00 2.15769863e+00
-5.77478647e-01 8.77108991e-01 2.24329621e-01 -6.59389794e-01
9.24818158e-01 5.05500436e-01 4.84500647e-01 -6.30657494e-01
3.25760871e-01 4.49927896e-01 6.35549545e-01 -1.19614929e-01
4.71042961e-01 -2.32492298e-01 1.73574403e-01 3.54039818e-01
3.24198425e-01 4.62666415e-02 4.75832671e-02 -1.92503542e-01
8.87893498e-01 -3.89011279e-02 1.39816012e-02 -2.07266565e-02
5.19799232e-01 -6.59240246e-01 6.91953719e-01 4.40437019e-01
1.23826511e-01 8.92382264e-01 1.91080436e-01 -3.11882272e-02
-8.02953899e-01 -9.62134600e-01 8.37126821e-02 1.18032563e+00
-1.08206570e-01 -3.75937074e-01 -7.38900065e-01 -1.19938157e-01
-5.23912385e-02 6.67475104e-01 -2.80132830e-01 -2.59330928e-01
-7.98713744e-01 -3.45918119e-01 9.53808188e-01 6.87063873e-01
7.63075888e-01 -1.45922518e+00 -3.09797615e-01 5.10673761e-01
-3.25115651e-01 -7.99134552e-01 -7.46791959e-01 3.36350590e-01
-8.21066558e-01 -6.63864970e-01 -1.09430397e+00 -6.85312867e-01
-1.43437177e-01 4.10913080e-01 9.28101718e-01 -1.82922423e-01
3.46643366e-02 -2.56428719e-01 -4.17323768e-01 -3.70150238e-01
-8.93096328e-02 5.76632082e-01 1.80809706e-01 3.97928506e-01
2.30226830e-01 -1.10187840e+00 -6.41142786e-01 1.61197111e-01
-8.62844586e-01 -7.76800811e-02 8.22810650e-01 6.52983248e-01
5.81409454e-01 3.30852151e-01 8.93618047e-01 -5.61697543e-01
7.70765126e-01 -3.17029357e-01 -3.23638827e-01 -1.71013951e-01
-1.52914122e-01 -1.20187715e-01 7.76537538e-01 -5.63401818e-01
-7.94346035e-01 -1.32990912e-01 -3.69719237e-01 -1.02601027e+00
-1.07483647e-04 5.05132377e-01 -2.80374736e-01 2.55879700e-01
2.94068485e-01 2.53107548e-01 -1.75765574e-01 -9.51151192e-01
1.96201280e-01 8.63952100e-01 6.99054360e-01 -1.89694703e-01
6.22449398e-01 5.59264608e-02 -2.43142292e-01 -9.46316421e-01
-9.71410573e-01 -4.89697009e-01 -3.63832861e-01 -2.72228941e-02
8.57791305e-01 -1.09175527e+00 -7.07934678e-01 6.90188527e-01
-1.17315900e+00 -2.78004944e-01 -3.81708741e-01 7.49934316e-01
-3.36982489e-01 5.56299537e-02 -8.00539732e-01 -5.88289261e-01
-5.16842723e-01 -1.15355873e+00 7.38726735e-01 4.05003160e-01
-2.46763065e-01 -4.84060407e-01 1.18699670e-01 1.78310171e-01
5.90723157e-01 -1.49866268e-01 6.32706285e-01 -6.41525328e-01
-3.92444700e-01 8.12643692e-02 -1.66759700e-01 7.10007489e-01
4.24804717e-01 -3.65642756e-01 -1.39921129e+00 -2.12094143e-01
1.44263625e-01 3.88109833e-02 1.23451304e+00 6.38462007e-01
1.22856414e+00 -2.49825448e-01 5.35335690e-02 9.31329250e-01
1.00935566e+00 3.39797407e-01 7.95902729e-01 -1.49033433e-02
8.96627545e-01 1.86483026e-01 -4.43589408e-03 3.44038159e-01
-3.95001583e-02 7.74818063e-01 3.12723041e-01 -1.95202395e-01
-6.32645249e-01 -1.87438503e-01 3.41459274e-01 1.47631955e+00
-2.66166240e-01 -1.35354146e-01 -5.03056109e-01 5.40845394e-01
-1.68192673e+00 -1.08536696e+00 -1.57507453e-02 2.11306739e+00
9.32378948e-01 7.08768740e-02 2.09667563e-01 4.76165235e-01
9.22047377e-01 4.16009337e-01 -9.05268669e-01 -1.77449867e-01
-2.48192728e-01 5.74976325e-01 1.00015573e-01 1.37124106e-01
-1.13414240e+00 8.67050231e-01 5.20265865e+00 1.27551961e+00
-1.44917119e+00 2.99208164e-01 3.04397911e-01 -4.17432398e-01
7.23831728e-02 -5.26952267e-01 -5.53910315e-01 5.83841026e-01
7.90613353e-01 -5.28605655e-02 8.44160080e-01 4.58761781e-01
1.23167716e-01 3.49050522e-01 -9.85065222e-01 1.26086581e+00
1.71989784e-01 -1.06516039e+00 -1.35076389e-01 -5.38410842e-02
7.73018181e-01 1.05619170e-01 2.14793757e-01 4.31475759e-01
-1.67407747e-02 -1.09696174e+00 1.01556993e+00 2.84082741e-01
7.96236098e-01 -1.16283274e+00 8.65579605e-01 2.14318916e-01
-1.48778689e+00 -1.89408883e-01 -2.32955933e-01 -6.18045777e-03
1.15770452e-01 7.67507911e-01 -3.02070081e-01 7.01865017e-01
7.04064608e-01 1.03247941e+00 -1.48313597e-01 1.26192737e+00
-3.87227237e-01 1.01627839e+00 -1.78151473e-01 4.61159855e-01
1.49886772e-01 -2.33249247e-01 7.93918550e-01 1.29995763e+00
5.04217267e-01 -1.15372509e-01 -1.17561705e-01 9.45542812e-01
-4.50573027e-01 -1.46528501e-02 -1.69782251e-01 -2.05980197e-01
4.37649369e-01 1.09401953e+00 -4.98361677e-01 -1.89857706e-01
-1.58552781e-01 1.13678885e+00 5.57750240e-02 3.16657037e-01
-9.02199149e-01 -8.39485526e-01 9.11222756e-01 -1.08276762e-01
6.43429041e-01 -2.46078834e-01 -1.48561344e-01 -1.10491872e+00
-9.01181772e-02 -1.04989421e+00 3.96678261e-02 -6.85043633e-01
-1.20231473e+00 9.11873817e-01 -6.14299774e-01 -1.51019967e+00
-1.64532766e-01 -2.21976876e-01 -7.33215392e-01 1.05503786e+00
-1.36810207e+00 -7.96540856e-01 -2.22016186e-01 5.99040031e-01
7.45427132e-01 -3.19123179e-01 6.03619993e-01 7.69875526e-01
-7.62067497e-01 7.42317498e-01 1.83839768e-01 4.51770961e-01
8.63220572e-01 -1.00457370e+00 4.54313487e-01 8.75657558e-01
8.05463791e-01 6.61136031e-01 2.61106819e-01 -4.11702514e-01
-1.06537223e+00 -1.24996054e+00 5.75736523e-01 2.91781962e-01
5.58559954e-01 -3.24147671e-01 -9.53446448e-01 4.23624545e-01
3.63855600e-01 -2.89788693e-01 6.43413782e-01 3.21293443e-01
-4.18023050e-01 -5.81301451e-01 -5.11005640e-01 5.61634183e-01
9.87014830e-01 -6.67574763e-01 -5.45740545e-01 -2.46902004e-01
9.05700207e-01 -4.10367876e-01 -4.04180110e-01 3.51495057e-01
6.99268997e-01 -1.25796556e+00 1.04301715e+00 -3.17741871e-01
5.09718597e-01 -4.31414127e-01 -1.22333795e-01 -1.64858937e+00
-7.20975161e-01 -7.81312346e-01 -2.08412856e-01 1.47843862e+00
3.01229060e-01 -2.48541787e-01 4.43765521e-01 -3.14266443e-01
-6.00943565e-01 -3.35872024e-01 -9.47002947e-01 -1.05318916e+00
-1.55018061e-01 -6.14941180e-01 6.44649923e-01 9.21302497e-01
-3.43046933e-01 6.31454766e-01 -6.17637932e-01 7.93334395e-02
3.57451886e-01 3.32230657e-01 5.86124599e-01 -1.19403934e+00
-5.87974966e-01 -8.93579781e-01 -2.52339721e-01 -1.27037394e+00
2.05637261e-01 -9.73108351e-01 1.50327727e-01 -1.37874401e+00
-7.45358989e-02 -2.31474951e-01 -9.06625032e-01 3.44110906e-01
-9.29443687e-02 5.22478163e-01 5.54468036e-01 2.64916569e-01
-5.15311778e-01 7.88921177e-01 1.10861671e+00 -2.71320075e-01
-6.52644396e-01 1.71479419e-01 -6.92349076e-01 8.01458061e-01
1.06434226e+00 -4.48474646e-01 -2.39550665e-01 -6.83042765e-01
-1.62419435e-02 1.21702373e-01 2.55057335e-01 -1.59285843e+00
3.32478195e-01 3.37188840e-01 3.54050606e-01 -7.45486617e-01
5.02538979e-01 -6.78396881e-01 2.67656446e-01 3.14949334e-01
-4.00554568e-01 -2.78422713e-01 4.64339763e-01 4.61045742e-01
-6.60496771e-01 -1.94128916e-01 7.53994644e-01 6.32461011e-02
-1.77774981e-01 6.54429197e-02 -1.42569840e-01 8.21900293e-02
3.76590252e-01 1.47992268e-01 -1.70342959e-02 -5.19842267e-01
-6.98224843e-01 -1.86105818e-01 -8.61061141e-02 5.15970945e-01
4.56007391e-01 -1.57542455e+00 -8.27348113e-01 2.54039913e-01
-3.35644782e-01 -1.14075422e-01 5.93699753e-01 1.02983332e+00
-6.67946711e-02 4.07459587e-01 -3.05476397e-01 -3.79788220e-01
-1.11053514e+00 4.39268023e-01 2.25503281e-01 -8.68112966e-02
-8.02965045e-01 1.05160642e+00 3.54807138e-01 9.73523036e-02
6.13389134e-01 -6.48722768e-01 -2.76647896e-01 1.48241207e-01
6.27232552e-01 5.58087349e-01 4.23461236e-02 -6.71604097e-01
-2.92422205e-01 7.14901388e-01 1.46392271e-01 -1.03066459e-01
1.34971797e+00 -1.43174548e-02 -1.44800693e-01 7.12006867e-01
1.00980008e+00 6.04775608e-01 -1.11919892e+00 -3.45575392e-01
-2.20300600e-01 -2.03676999e-01 3.35954666e-01 -6.84883773e-01
-1.58054376e+00 1.14109468e+00 4.28743213e-01 2.35841572e-01
1.45276833e+00 -2.22708121e-01 1.13868082e+00 1.51147485e-01
8.30945224e-02 -7.75203049e-01 1.02613516e-01 6.18609071e-01
1.11162436e+00 -8.73234272e-01 -2.74529219e-01 -2.35890895e-02
-5.32145143e-01 9.70372021e-01 2.78334618e-01 -3.70974988e-01
5.66783488e-01 2.04537034e-01 4.87063117e-02 1.49620429e-01
-3.15503687e-01 -4.08117086e-01 5.78275979e-01 1.60858169e-01
6.89370990e-01 4.19199131e-02 1.46736100e-01 1.21636009e+00
-5.21583080e-01 3.18790041e-02 1.66540131e-01 2.67274201e-01
-3.37496668e-01 -8.16257954e-01 -3.89984906e-01 3.32746267e-01
-6.83525681e-01 -4.90200639e-01 -3.52955997e-01 2.92517096e-01
3.52293611e-01 1.09598804e+00 2.31471255e-01 -7.23686874e-01
3.63076240e-01 7.97303990e-02 3.67829680e-01 -5.86336613e-01
-1.00372076e+00 7.71463394e-01 -1.51749983e-01 -3.15510422e-01
-4.26425874e-01 -3.17103893e-01 -1.14181423e+00 -7.70069882e-02
-3.96303862e-01 2.06417277e-01 5.06548762e-01 5.95604002e-01
5.91437578e-01 1.32417655e+00 6.73010111e-01 -1.07430959e+00
-3.49635303e-01 -1.36481667e+00 -9.09606993e-01 1.49396688e-01
6.04649067e-01 -3.93928438e-01 -2.27255896e-01 8.65743533e-02] | [15.550687789916992, 5.500222206115723] |
f50aa5e4-62ec-4da2-8bd4-cef0cdc55536 | universal-planning-networks | 1804.00645 | null | http://arxiv.org/abs/1804.00645v2 | http://arxiv.org/pdf/1804.00645v2.pdf | Universal Planning Networks | A key challenge in complex visuomotor control is learning abstract
representations that are effective for specifying goals, planning, and
generalization. To this end, we introduce universal planning networks (UPN).
UPNs embed differentiable planning within a goal-directed policy. This planning
computation unrolls a forward model in a latent space and infers an optimal
action plan through gradient descent trajectory optimization. The
plan-by-gradient-descent process and its underlying representations are learned
end-to-end to directly optimize a supervised imitation learning objective. We
find that the representations learned are not only effective for goal-directed
visual imitation via gradient-based trajectory optimization, but can also
provide a metric for specifying goals using images. The learned representations
can be leveraged to specify distance-based rewards to reach new target states
for model-free reinforcement learning, resulting in substantially more
effective learning when solving new tasks described via image-based goals. We
were able to achieve successful transfer of visuomotor planning strategies
across robots with significantly different morphologies and actuation
capabilities. | ['Pieter Abbeel', 'Sergey Levine', 'Allan Jabri', 'Chelsea Finn', 'Aravind Srinivas'] | 2018-04-02 | null | null | null | null | ['transfer-reinforcement-learning'] | ['methodology'] | [ 2.70481139e-01 4.64754254e-01 -3.71327460e-01 -2.24060670e-01
-7.54835546e-01 -6.56189978e-01 8.74805093e-01 -1.22263536e-01
-3.21966916e-01 7.85710871e-01 5.37891269e-01 -2.16205679e-02
-2.80472338e-01 -7.36724198e-01 -1.00887883e+00 -6.44179761e-01
-4.11765307e-01 6.59996867e-01 -1.48888618e-01 -3.84863853e-01
5.18690825e-01 7.71557808e-01 -1.35194528e+00 7.47944862e-02
5.58437586e-01 5.34115374e-01 7.32672334e-01 7.69319177e-01
1.47366136e-01 9.89757478e-01 2.51568127e-02 6.23222649e-01
3.12480032e-01 -3.47711146e-01 -9.44925308e-01 2.37754345e-01
8.50011930e-02 -5.03030121e-01 -6.19236231e-01 6.95523620e-01
5.27403280e-02 4.91130322e-01 9.60583746e-01 -1.17770267e+00
-8.25473428e-01 5.83410561e-01 1.42588824e-01 -3.57369930e-01
3.52472603e-01 8.90883803e-01 1.07627320e+00 -4.79195237e-01
9.76843894e-01 1.55959809e+00 1.56323224e-01 1.00980818e+00
-1.34872866e+00 -1.55410185e-01 2.43104160e-01 -1.09003135e-03
-5.96118331e-01 -2.45567143e-01 4.19745386e-01 -5.57394385e-01
1.41218770e+00 -3.48843068e-01 9.65803146e-01 1.18330240e+00
5.41371882e-01 7.26543009e-01 7.90176630e-01 -4.90425862e-02
4.40105587e-01 -7.74208367e-01 -4.87151206e-01 1.05116796e+00
-1.67707995e-01 7.16385603e-01 -3.85693908e-01 2.44703636e-01
1.31175113e+00 2.36505508e-01 -1.33028254e-01 -8.51676524e-01
-1.52521193e+00 8.85030031e-01 8.85273516e-01 -5.50359227e-02
-4.46212977e-01 1.04362190e+00 6.49145767e-02 3.26565087e-01
-3.22972894e-01 1.20325470e+00 -5.42110980e-01 -3.01514179e-01
-3.33034098e-01 7.30247259e-01 6.20611906e-01 1.07012534e+00
9.22034621e-01 4.72611785e-01 -2.97158360e-01 4.08974290e-01
3.97779256e-01 7.62952924e-01 3.76540810e-01 -1.83750319e+00
2.57226110e-01 3.81283462e-01 2.87707746e-01 -4.84912574e-01
-6.25886321e-01 4.33023348e-02 -6.44574687e-02 1.16473734e+00
3.07208240e-01 -2.71136224e-01 -1.30761456e+00 1.93663144e+00
-8.27144540e-04 -1.20867245e-01 2.98878968e-01 9.89019990e-01
2.11266913e-02 6.40267789e-01 2.11097986e-01 6.90172166e-02
7.04358935e-01 -1.19446290e+00 -2.52734393e-01 -3.74841273e-01
7.46588647e-01 -1.07777543e-01 1.18119609e+00 2.34232441e-01
-1.16333246e+00 -2.22873762e-01 -9.18673873e-01 -5.06153107e-02
-6.80548698e-02 -3.27832326e-02 5.34172475e-01 -3.12636256e-01
-1.27905905e+00 1.17024624e+00 -1.27144551e+00 -5.31461775e-01
6.59853995e-01 4.98091280e-01 -4.36927497e-01 9.99425421e-04
-4.30110186e-01 1.29317713e+00 4.18033093e-01 -4.46154416e-01
-2.02244854e+00 -3.31633300e-01 -1.06044126e+00 -8.74177888e-02
4.04423982e-01 -7.69321322e-01 1.53184533e+00 -6.35563314e-01
-2.12331867e+00 5.34636557e-01 1.60025194e-01 -6.28998339e-01
2.68521868e-02 -1.30145326e-01 3.60964149e-01 1.78946838e-01
3.55701655e-01 1.16728854e+00 1.03239417e+00 -1.13376880e+00
-3.91694367e-01 -1.55630186e-01 3.01238894e-01 6.17218316e-01
1.66018233e-01 -5.41223645e-01 -7.06043914e-02 -3.68554056e-01
6.19094111e-02 -1.19375265e+00 -6.12390220e-01 4.05620784e-01
-3.24611187e-01 -3.47564191e-01 5.72759867e-01 -2.86694497e-01
3.15281063e-01 -1.66018021e+00 9.81577933e-01 -1.20654747e-01
2.15388224e-01 -1.21435039e-01 -6.87142730e-01 6.65887594e-01
4.19763625e-01 -8.18066224e-02 -1.98647246e-01 -1.46717280e-01
7.39879757e-02 6.36553466e-01 -4.44160640e-01 3.81851643e-01
4.98200595e-01 1.30324376e+00 -1.34076929e+00 1.59337390e-02
3.55867714e-01 1.94307938e-01 -7.69367814e-01 5.35354078e-01
-1.05446601e+00 9.68602896e-01 -7.28572965e-01 5.42358696e-01
-2.67694652e-01 -1.45353228e-01 3.22608888e-01 2.80704498e-01
-3.44805539e-01 4.99934345e-01 -4.91600305e-01 2.28580904e+00
-6.13357902e-01 6.56742811e-01 6.55307323e-02 -1.07066393e+00
9.29898202e-01 -6.73614368e-02 4.91281211e-01 -5.02199829e-01
1.59586653e-01 2.16398612e-01 -5.05474433e-02 -5.82601070e-01
3.35812896e-01 -1.15176953e-01 -2.19999313e-01 5.26153445e-01
4.28275317e-01 -9.57046092e-01 2.09603296e-03 -4.99396957e-02
1.17302907e+00 1.12383735e+00 1.21319562e-01 -2.23288238e-01
-2.90529197e-03 6.29538000e-01 2.20186010e-01 6.89412773e-01
8.86035094e-04 4.38345730e-01 4.14997101e-01 -4.20979857e-01
-1.28880548e+00 -1.43360603e+00 3.79178196e-01 1.25795197e+00
4.58422340e-02 -1.73983887e-01 -4.51825440e-01 -4.21632111e-01
2.12719992e-01 7.72021413e-01 -6.70264125e-01 -2.76505113e-01
-8.41308713e-01 4.13053364e-01 3.96750003e-01 6.57217324e-01
1.84611250e-02 -1.60489130e+00 -1.08247566e+00 4.86351639e-01
2.69315690e-01 -7.11266637e-01 -5.02979696e-01 4.30442512e-01
-1.00209987e+00 -1.16563022e+00 -7.96949804e-01 -8.94602120e-01
8.04743350e-01 -1.57016180e-02 6.42929494e-01 -8.05087760e-02
-8.44740123e-02 8.63497078e-01 -9.78950709e-02 -1.90081224e-01
-6.93408251e-01 -4.44077235e-03 3.52164716e-01 -7.85884142e-01
-2.52764910e-01 -8.15055728e-01 -3.85194778e-01 1.06818795e-01
-3.75166357e-01 2.01094791e-01 6.19973719e-01 9.73750949e-01
6.64202750e-01 -7.38429725e-01 4.23846513e-01 -1.42494187e-01
8.26147497e-01 -3.56396079e-01 -7.76299417e-01 -9.53068677e-03
-5.56512475e-01 8.02979708e-01 8.06627274e-01 -6.07499838e-01
-6.36266410e-01 2.40739092e-01 8.74655098e-02 -5.82320929e-01
-2.38941014e-01 3.32260430e-01 2.46448547e-01 -8.30171257e-02
8.14647377e-01 3.18365753e-01 5.50630331e-01 -1.35286212e-01
9.91012573e-01 9.39806774e-02 5.80864012e-01 -9.12592947e-01
5.03109515e-01 4.26354349e-01 2.51063585e-01 -5.07498324e-01
-3.86582315e-01 8.98673609e-02 -6.18001342e-01 -1.32902592e-01
1.17698479e+00 -6.57951474e-01 -1.12144554e+00 2.59514838e-01
-1.00733018e+00 -1.20507503e+00 -7.49974787e-01 4.91472185e-01
-1.65155196e+00 -1.20254315e-01 -5.47802150e-01 -5.23548186e-01
1.75191946e-02 -1.55642581e+00 1.08622038e+00 2.18040362e-01
-3.88343513e-01 -8.66000056e-01 2.28753164e-01 -3.98264289e-01
4.44063336e-01 4.29232270e-01 9.35450554e-01 -1.79691076e-01
-9.21027124e-01 4.12964016e-01 4.89239432e-02 1.10420985e-02
1.95562765e-01 -2.13863105e-01 -4.08622384e-01 -4.10815388e-01
-5.32542646e-01 -9.68440354e-01 7.60745168e-01 4.84624803e-01
8.57473969e-01 -4.45430934e-01 -4.66023445e-01 8.00693929e-01
1.18832624e+00 1.34285927e-01 2.92340904e-01 2.79034764e-01
6.31665468e-01 5.74026644e-01 6.37652695e-01 3.12050521e-01
3.22986066e-01 6.36533737e-01 9.37066913e-01 5.46228766e-01
-2.17083007e-01 -6.73064172e-01 6.53167367e-01 2.46556252e-01
-2.48096332e-01 2.56158173e-01 -9.10890162e-01 5.46689451e-01
-2.07006526e+00 -9.44504797e-01 5.21966338e-01 1.83657753e+00
8.04651260e-01 -3.42163220e-02 7.52156526e-02 -6.09785914e-01
1.60001129e-01 9.49792862e-02 -1.24694562e+00 -5.54106116e-01
3.98859173e-01 2.38726825e-01 5.66818595e-01 6.68207049e-01
-7.16420770e-01 1.44826317e+00 7.19000244e+00 2.29080752e-01
-1.24651659e+00 -2.00570047e-01 -3.27178463e-02 -2.08064586e-01
-4.49591935e-01 1.73276126e-01 -5.81415772e-01 -5.86295500e-02
6.41862154e-01 -1.17851458e-01 1.32450187e+00 9.71997678e-01
3.85552198e-01 1.40918389e-01 -1.37756431e+00 7.65794039e-01
-3.95994246e-01 -1.81828570e+00 4.46487740e-02 2.26240218e-01
7.34771371e-01 4.41196889e-01 1.00905821e-01 5.94536543e-01
1.21645272e+00 -1.42173707e+00 8.14791441e-01 7.19824910e-01
9.70225871e-01 -4.62062329e-01 -2.96393335e-01 5.75486898e-01
-9.31008697e-01 -4.42714244e-01 -5.25601685e-01 -4.14848477e-01
3.39623600e-01 -4.06221211e-01 -1.00142276e+00 -6.57530874e-02
2.08369881e-01 1.11402869e+00 2.95668304e-01 4.98004466e-01
-7.33014703e-01 1.22146569e-01 -2.95313299e-01 -5.02105057e-01
7.29037046e-01 -2.52514422e-01 8.07326853e-01 6.94831192e-01
3.24056774e-01 -1.04434259e-01 4.14616406e-01 1.39508498e+00
3.67863402e-02 -4.00899142e-01 -1.03553069e+00 -5.53140879e-01
3.36803466e-01 9.71712828e-01 -4.04532760e-01 -1.18149452e-01
1.16345115e-01 9.22843397e-01 9.86164212e-01 6.12794936e-01
-5.53943396e-01 3.00431158e-02 1.13392258e+00 -1.41158074e-01
4.20571059e-01 -9.71085131e-01 -1.21971697e-01 -9.68329370e-01
-5.32771349e-01 -7.10791647e-01 -3.77756596e-01 -9.72361624e-01
-8.65470529e-01 2.41921008e-01 8.20531547e-02 -1.06286538e+00
-9.22375798e-01 -9.58396494e-01 -6.36872351e-01 8.49975586e-01
-1.28064775e+00 -1.20139909e+00 -9.43269432e-02 6.67345881e-01
6.98025048e-01 -4.69998956e-01 9.86573339e-01 -7.11171389e-01
9.73005369e-02 -2.16179024e-02 -1.46338958e-02 -8.66613090e-02
3.37063938e-01 -1.35842156e+00 4.07636583e-01 4.59359318e-01
6.06830567e-02 6.29415393e-01 6.94785058e-01 -6.82680428e-01
-1.72940755e+00 -1.00445902e+00 -3.07969674e-02 -4.99864370e-01
6.78405821e-01 -5.31463064e-02 -3.94751698e-01 9.90002096e-01
1.61933914e-01 -1.76122747e-02 -5.63959815e-02 -1.41666964e-01
-2.77215809e-01 3.16933692e-01 -1.05279303e+00 1.00452638e+00
1.44019580e+00 -5.14402628e-01 -7.11933851e-01 3.63192528e-01
9.61615443e-01 -6.97256505e-01 -6.40465021e-01 -4.47593955e-03
7.16416419e-01 -3.98755193e-01 9.04810667e-01 -1.18274379e+00
6.61332965e-01 -1.93873554e-01 -3.10422391e-01 -1.67810059e+00
-5.36235571e-01 -7.21013010e-01 -1.32102758e-01 3.27639133e-01
4.01401073e-01 -5.20720541e-01 6.16264224e-01 3.20376247e-01
-4.93924588e-01 -1.04786909e+00 -7.04666018e-01 -7.40082502e-01
5.52617431e-01 -3.77804458e-01 4.76190299e-01 4.64627564e-01
3.63815427e-01 9.16895345e-02 -3.42076898e-01 1.29438471e-02
5.24767280e-01 2.35129088e-01 8.24146152e-01 -7.29536176e-01
-4.44703609e-01 -6.15674257e-01 -3.02190274e-01 -1.52127218e+00
4.28803086e-01 -1.21869826e+00 5.88360071e-01 -2.07811189e+00
-2.37926930e-01 -3.58383149e-01 5.41796237e-02 9.04088914e-01
4.12559181e-01 -4.41285640e-01 5.99127412e-01 3.97383481e-01
-3.70038211e-01 8.42553258e-01 1.85492623e+00 -1.62714064e-01
-3.39803547e-01 -2.94765770e-01 -4.16836292e-01 6.72552049e-01
1.00244403e+00 -5.04624128e-01 -5.74226677e-01 -6.48980975e-01
-7.49562308e-02 3.01980644e-01 3.98667186e-01 -1.02786613e+00
1.38049155e-01 -8.21439147e-01 6.40590489e-02 -4.86408845e-02
4.74354208e-01 -4.37856466e-01 -4.35649127e-01 8.37577522e-01
-7.86716819e-01 2.92212470e-03 -4.43842672e-02 7.19589829e-01
3.09831262e-01 -8.40891078e-02 5.58876872e-01 -4.79691327e-01
-1.07729530e+00 4.42008525e-01 -7.80889452e-01 6.57078847e-02
1.05370903e+00 -4.22484614e-02 -3.05387080e-01 -5.59354186e-01
-1.05632555e+00 5.16608655e-01 7.26992369e-01 5.27782202e-01
8.35041642e-01 -1.32414818e+00 -5.41212320e-01 1.93942204e-01
1.19928329e-03 5.20377792e-02 -3.20297271e-01 4.25933510e-01
-5.98933816e-01 3.88021618e-01 -8.09733093e-01 -7.81135321e-01
-6.21868014e-01 3.76485139e-01 5.91914177e-01 -2.42396608e-01
-7.82293737e-01 6.78332746e-01 6.35953546e-02 -9.76452887e-01
4.09515463e-02 -6.44990206e-01 -6.49234354e-02 -7.42285013e-01
1.31502464e-01 1.09647989e-01 -8.28844309e-01 -2.83208996e-01
-4.46458869e-02 5.07871509e-01 3.81508172e-01 -5.43895423e-01
1.41584349e+00 1.33020547e-03 1.96970463e-01 3.33378911e-01
1.02183771e+00 -4.53397959e-01 -2.30027342e+00 1.26372114e-01
-1.56561375e-01 -5.13291806e-02 -2.51493394e-01 -7.28946149e-01
-5.06431222e-01 9.60788429e-01 1.58418417e-01 -2.58809298e-01
4.24011767e-01 9.45980549e-02 6.78229094e-01 8.83029044e-01
8.97659063e-01 -1.01178944e+00 7.25031793e-01 1.03551543e+00
1.29773891e+00 -1.16836154e+00 -1.58411771e-01 3.07916611e-01
-7.27462649e-01 1.47020280e+00 6.18497849e-01 -7.01009750e-01
2.65839785e-01 3.01753461e-01 -1.13382742e-01 -1.64015472e-01
-8.45508277e-01 -5.50667882e-01 2.02644050e-01 1.23165345e+00
6.76654861e-04 1.71280041e-01 8.73055980e-02 2.87028626e-02
-7.47177675e-02 5.08119091e-02 4.47236568e-01 9.59513426e-01
-9.01604116e-01 -1.20280683e+00 7.69292563e-02 3.93073499e-01
2.68134624e-01 2.50589669e-01 1.90203283e-02 7.37668455e-01
-2.57080019e-01 5.69557130e-01 -2.82330648e-03 -4.72312659e-01
7.31052607e-02 -7.24026859e-02 1.05622721e+00 -1.10752749e+00
-5.43106124e-02 -2.70372629e-01 4.87914830e-02 -1.14276421e+00
-9.71611217e-02 -7.78046548e-01 -2.01804471e+00 4.16405722e-02
3.28009903e-01 -2.67967224e-01 7.46656120e-01 1.05017924e+00
4.15985852e-01 6.08117938e-01 3.99863303e-01 -1.48519349e+00
-1.12777162e+00 -7.60819376e-01 -2.08559737e-01 2.84789026e-01
5.54901242e-01 -8.61002862e-01 4.09604125e-02 -5.69580961e-03] | [4.491209030151367, 0.9587392807006836] |
f51af56e-0f83-4448-a0c4-ab0275988bae | gradient-descent-type-methods-background-and | 2212.09413 | null | https://arxiv.org/abs/2212.09413v1 | https://arxiv.org/pdf/2212.09413v1.pdf | Gradient Descent-Type Methods: Background and Simple Unified Convergence Analysis | In this book chapter, we briefly describe the main components that constitute the gradient descent method and its accelerated and stochastic variants. We aim at explaining these components from a mathematical point of view, including theoretical and practical aspects, but at an elementary level. We will focus on basic variants of the gradient descent method and then extend our view to recent variants, especially variance-reduced stochastic gradient schemes (SGD). Our approach relies on revealing the structures presented inside the problem and the assumptions imposed on the objective function. Our convergence analysis unifies several known results and relies on a general, but elementary recursive expression. We have illustrated this analysis on several common schemes. | ['Marten van Dijk', 'Quoc Tran-Dinh'] | 2022-12-19 | null | null | null | null | ['type'] | ['speech'] | [-4.09685597e-02 -1.76001996e-01 9.47972089e-02 -2.26073757e-01
-5.07227302e-01 -4.66458201e-01 4.95133013e-01 -3.49569947e-01
-5.00617325e-01 1.10913062e+00 -1.49660736e-01 -5.52812517e-01
-3.65594089e-01 -3.82559180e-01 -1.59971595e-01 -1.06668532e+00
-3.49148393e-01 1.19503051e-01 9.37899388e-03 -7.61070669e-01
5.48125207e-01 5.02330482e-01 -1.31391478e+00 -3.38710904e-01
9.60285544e-01 8.28347921e-01 8.46663639e-02 1.13502133e+00
-3.66782516e-01 6.55142546e-01 -4.63060439e-01 -8.22391272e-01
2.89267600e-01 -7.69711614e-01 -8.38959754e-01 1.82261333e-01
-2.60727201e-02 -2.61354387e-01 -9.35278982e-02 1.11979640e+00
7.58177459e-01 3.63058358e-01 7.63286531e-01 -1.08390629e+00
-2.14582175e-01 5.60052454e-01 -5.11112034e-01 3.96450281e-01
2.57589132e-01 -2.82124221e-01 8.57420266e-01 -1.15338850e+00
4.78130013e-01 7.16921926e-01 1.22766256e+00 7.58819103e-01
-1.08265722e+00 1.76793292e-01 4.71384943e-01 1.23372935e-01
-1.16278625e+00 -2.98070014e-01 1.05274808e+00 -5.02463341e-01
7.26035297e-01 3.53161424e-01 5.19901752e-01 6.18375421e-01
-2.14095600e-02 1.10303378e+00 1.27391779e+00 -8.08871746e-01
2.15687111e-01 3.59534442e-01 4.60903764e-01 7.96039224e-01
2.00453341e-01 2.50791721e-02 -2.96839833e-01 -2.46521860e-01
5.63619256e-01 -3.16848814e-01 -4.90686804e-01 -6.76577866e-01
-6.64728105e-01 1.06666911e+00 -8.89601931e-03 7.38493800e-01
-3.06788296e-01 1.48371115e-01 4.66643304e-01 4.04814273e-01
8.47955465e-01 -1.10560566e-01 -4.81606275e-01 -3.68791908e-01
-1.24813688e+00 2.37946287e-01 1.22477353e+00 7.88760483e-01
5.03665149e-01 3.89481992e-01 -1.13720968e-01 8.72608185e-01
1.25063211e-01 3.60914707e-01 5.90107024e-01 -5.88469386e-01
3.17575723e-01 -2.38469169e-01 2.88364619e-01 -9.40117538e-01
-4.87332553e-01 -7.72276402e-01 -7.46749759e-01 2.48690039e-01
3.77688974e-01 -5.17759740e-01 -3.73110741e-01 1.53531098e+00
4.55711871e-01 -1.13403074e-01 -2.03812853e-01 6.45383835e-01
4.88300532e-01 5.22360921e-01 -4.54342067e-01 -6.50042593e-01
6.86540902e-01 -1.00057912e+00 -8.69963706e-01 4.23285127e-01
8.23150039e-01 -7.35523641e-01 1.06491983e+00 3.58200312e-01
-1.45868564e+00 -3.14784616e-01 -9.81837213e-01 2.03135945e-02
-3.35065633e-01 5.04128277e-01 6.27232552e-01 1.21374464e+00
-1.49143708e+00 1.10507047e+00 -9.01378274e-01 -1.62319362e-01
-1.20485790e-01 2.77343333e-01 1.56385913e-01 4.77991611e-01
-8.07841778e-01 7.46657848e-01 -4.67112623e-02 5.03847361e-01
-4.70064551e-01 -6.26601398e-01 -6.70552075e-01 -1.86966777e-01
1.99234679e-01 -7.70620823e-01 1.73815966e+00 -8.43562543e-01
-2.17888355e+00 8.62777770e-01 -4.06061858e-01 -3.13167185e-01
1.30943286e+00 -3.08399290e-01 3.45497555e-03 -7.07017770e-03
-2.74571508e-01 -3.92263472e-01 8.55686665e-01 -1.06947482e+00
-7.78952360e-01 -2.77016640e-01 -1.09888084e-01 1.80031478e-01
-2.65571296e-01 2.99754024e-01 -2.69910455e-01 -6.76126897e-01
-6.77348068e-03 -6.82180882e-01 -4.05620724e-01 -2.57062137e-01
-3.54962468e-01 -1.59790993e-01 3.91463876e-01 -5.77738464e-01
1.46736073e+00 -1.85973549e+00 3.77686024e-01 1.40556529e-01
2.44073853e-01 2.99590290e-01 3.66589010e-01 9.52856958e-01
-7.50888065e-02 -9.58468858e-03 -6.76201820e-01 -8.01376045e-01
2.91736424e-01 -1.62162215e-01 -3.32241803e-01 7.70054758e-01
-3.16665649e-01 6.49771333e-01 -1.03299606e+00 -2.73803085e-01
3.05133730e-01 6.23660564e-01 -3.12306315e-01 1.02981748e-02
4.15210247e-01 1.98062986e-01 -5.02436697e-01 3.31064999e-01
8.18160474e-01 5.59647121e-02 3.15743908e-02 9.69549790e-02
-5.08098841e-01 3.71665180e-01 -1.36390078e+00 1.45446336e+00
-7.50957727e-01 6.12250626e-01 6.83074713e-01 -1.55775189e+00
6.22961283e-01 8.73834342e-02 5.17589629e-01 -3.99799235e-02
-1.20916944e-02 6.60251379e-01 -4.56806600e-01 -2.51480669e-01
5.81990898e-01 -3.46141428e-01 3.53717387e-01 3.58410060e-01
7.33039975e-02 -1.69605345e-01 6.88856781e-01 1.89747229e-01
5.58124721e-01 2.22359851e-01 4.05854136e-01 -7.21580386e-01
1.01725006e+00 -1.86145440e-01 1.10483997e-01 1.08732629e+00
-2.85451561e-01 3.29689652e-01 5.68691254e-01 -4.64339077e-01
-9.24400866e-01 -8.64131391e-01 -3.50061417e-01 1.19598663e+00
-2.68083394e-01 -4.57350582e-01 -7.39690661e-01 -7.03074574e-01
-1.07619137e-01 6.20892882e-01 -8.97735178e-01 1.87338114e-01
-4.63975400e-01 -1.30956936e+00 2.62703538e-01 2.83196568e-01
6.18380070e-01 -3.46759409e-01 -4.18331027e-01 1.79479137e-01
3.15613866e-01 -6.63982093e-01 -2.96628386e-01 2.73556232e-01
-1.27216244e+00 -9.37632203e-01 -1.24127579e+00 -5.75992644e-01
4.29858416e-01 7.60186985e-02 1.42746317e+00 1.35179818e-01
-2.51856089e-01 8.35041821e-01 -1.73507556e-01 -2.30643123e-01
-2.48431355e-01 9.49593931e-02 1.41155869e-01 4.37213704e-02
1.32326514e-01 -6.92203820e-01 -4.81191278e-01 1.57779068e-01
-6.84669673e-01 -2.37720981e-01 2.90618539e-01 9.52617824e-01
3.07425588e-01 -1.32223696e-01 6.29548952e-02 -1.05257905e+00
9.28300083e-01 -3.62128705e-01 -1.09413207e+00 2.73692906e-02
-9.00202751e-01 2.43223891e-01 6.26314044e-01 -4.35949042e-02
-1.25745273e+00 -5.13841622e-02 -7.57787466e-01 2.02192247e-01
4.50249821e-01 3.44838619e-01 2.85808444e-01 -5.60283840e-01
4.53212142e-01 4.70813602e-01 -1.64922521e-01 -7.99936831e-01
5.51589251e-01 4.84677434e-01 2.95708209e-01 -8.31109762e-01
7.88397968e-01 7.37780690e-01 2.74329424e-01 -1.03789687e+00
-7.30819285e-01 -5.49206018e-01 -6.71948433e-01 -1.40934169e-01
3.24339449e-01 -2.64780045e-01 -7.11042583e-01 7.40091860e-01
-9.89137352e-01 -5.52770078e-01 -5.90190351e-01 5.68616450e-01
-9.53936994e-01 5.57922006e-01 -9.80190098e-01 -1.34496236e+00
-3.51794571e-01 -8.93506169e-01 7.99490988e-01 1.34129539e-01
3.85117441e-01 -1.80544078e+00 6.62300169e-01 -3.82285714e-01
6.37551904e-01 -1.82042345e-02 3.51433486e-01 -5.51260352e-01
1.26540303e-01 -8.36376846e-02 1.02868512e-01 6.26249731e-01
-5.92306182e-02 2.52893806e-01 -7.02962458e-01 -2.85263866e-01
7.30447292e-01 7.22862557e-02 9.53880548e-01 6.69020772e-01
8.32676589e-01 -6.69127330e-02 -3.55882674e-01 1.08132696e+00
1.75807333e+00 -3.57591420e-01 2.42515445e-01 3.72220904e-01
4.35743004e-01 7.32643783e-01 2.62811691e-01 6.39741182e-01
-1.00056089e-01 5.84145069e-01 1.64614648e-01 -1.11632757e-01
1.60777584e-01 -4.69208770e-02 3.47835124e-01 1.01731706e+00
-3.73448730e-01 5.98169565e-02 -5.22623956e-01 3.63727033e-01
-1.83188736e+00 -8.34045947e-01 -4.45654094e-01 2.23699641e+00
8.62478971e-01 -2.27925688e-01 4.30137128e-01 4.48940188e-01
8.73016536e-01 1.56854074e-02 -3.42012137e-01 -8.06519687e-01
-2.89596200e-01 1.26182556e-01 6.79488838e-01 9.16273475e-01
-9.74246144e-01 6.20347738e-01 8.38097668e+00 9.50345457e-01
-9.09373879e-01 2.06920773e-01 5.57700694e-01 -1.91691011e-01
-1.44770488e-01 -1.53436586e-01 -1.05679202e+00 4.15913641e-01
7.41709769e-01 -2.90031046e-01 3.93344223e-01 1.14645708e+00
2.02281296e-01 5.21632982e-03 -6.85260534e-01 1.07923484e+00
6.13579340e-02 -1.20280969e+00 -2.15056598e-01 -7.04514384e-02
1.06548917e+00 3.19146216e-02 2.34198719e-01 3.16214621e-01
2.63717622e-01 -8.01313460e-01 4.33181494e-01 3.89675587e-01
5.20779610e-01 -7.11617470e-01 7.98130095e-01 2.29653209e-01
-9.14846838e-01 3.34420539e-02 -3.98943484e-01 -5.39804220e-01
3.72105747e-01 1.06580806e+00 -4.75967266e-02 8.30101132e-01
5.12469888e-01 4.77732956e-01 -2.84407698e-02 1.10017562e+00
-3.20103079e-01 6.27650321e-01 -6.45500124e-01 -5.25301039e-01
5.63849986e-01 -8.32455158e-01 7.14750230e-01 1.56996453e+00
2.97693014e-01 2.73880325e-02 -1.91463411e-01 7.12269247e-01
2.43623152e-01 6.08273685e-01 -3.53090316e-01 3.88541937e-01
-1.74879879e-01 1.25164354e+00 -6.50547504e-01 -4.92843986e-01
-7.32153893e-01 1.12009704e+00 1.95675090e-01 4.45882648e-01
-5.89565992e-01 -9.50813115e-01 7.52907455e-01 -2.02205226e-01
5.78353822e-01 -4.90508735e-01 -3.70234758e-01 -1.43701184e+00
2.93622643e-01 -4.41735744e-01 3.47029954e-01 -1.81303978e-01
-1.03719175e+00 5.18560648e-01 2.06650078e-01 -6.51327670e-01
-4.96660441e-01 -1.08930266e+00 -7.90019393e-01 1.14778721e+00
-1.51214778e+00 -3.89344037e-01 9.21097472e-02 5.64735889e-01
5.38789868e-01 1.25988141e-01 3.58511537e-01 1.45901293e-01
-6.98091984e-01 6.39560580e-01 1.01576889e+00 -3.78138758e-02
1.88888293e-02 -1.60354567e+00 3.23487580e-01 1.09017372e+00
2.79848557e-03 7.07055688e-01 1.16354275e+00 -1.45249337e-01
-1.41091955e+00 -4.27346557e-01 1.06461442e+00 -2.56356418e-01
9.15575325e-01 -1.40434235e-01 -5.16822100e-01 5.11673510e-01
3.50767486e-02 -1.58632070e-01 4.93773490e-01 1.89149678e-01
3.60006839e-01 1.34020925e-01 -1.11211574e+00 6.21531188e-01
1.11187458e+00 -4.11645561e-01 -5.51267684e-01 4.57797736e-01
1.99508164e-02 -3.04924816e-01 -5.24267197e-01 1.46607369e-01
4.27601248e-01 -1.41876638e+00 8.50537062e-01 -7.90271699e-01
1.21681644e-02 -5.73034547e-02 2.68746596e-02 -1.32114410e+00
-6.32352307e-02 -1.40294182e+00 -1.90514237e-01 9.60347295e-01
1.97877273e-01 -9.90759611e-01 7.92971492e-01 4.45897907e-01
-1.21973939e-01 -1.06141782e+00 -9.74190414e-01 -1.11333108e+00
4.60714340e-01 -6.08118951e-01 8.21419507e-02 5.13474822e-01
2.13818744e-01 2.13246867e-01 -5.33152521e-01 -2.55529642e-01
6.62053168e-01 4.16668504e-02 5.14794469e-01 -8.70159984e-01
-3.92103940e-01 -1.01595855e+00 -3.33477676e-01 -1.65256500e+00
-1.34435579e-01 -5.32901168e-01 -1.10671520e-01 -1.16201854e+00
-1.54467791e-01 -1.41006008e-01 -3.14628780e-01 -3.25676799e-01
-3.39309573e-01 3.88359427e-01 -1.47580743e-01 1.48737371e-01
-4.04087335e-01 6.21166646e-01 6.70206666e-01 2.89907068e-01
-4.12711114e-01 6.73746347e-01 -4.14411902e-01 8.55150938e-01
9.48510587e-01 -1.36338741e-01 -3.72803867e-01 -3.68231386e-01
5.51900327e-01 -4.53212261e-01 1.16744339e-01 -7.37712741e-01
5.88282160e-02 2.88071543e-01 7.64429197e-02 -4.76516843e-01
1.62981331e-01 -2.75528520e-01 -3.30091476e-01 7.85623908e-01
-3.69749367e-01 2.88317233e-01 -4.44067270e-02 2.78558850e-01
-1.67480305e-01 -1.07762051e+00 6.93043530e-01 -1.59446523e-01
-4.22287554e-01 1.93617478e-01 -3.68524641e-01 7.86726773e-02
5.83144009e-01 -7.75862634e-02 5.24285197e-01 -5.33531368e-01
-1.03222477e+00 4.34563607e-02 4.05531853e-01 -5.21878541e-01
2.76606411e-01 -1.06219184e+00 -8.17516267e-01 -7.42031783e-02
-3.89926881e-01 -6.75894260e-01 1.94696993e-01 1.29909050e+00
-7.20074296e-01 5.90564907e-01 3.28449219e-01 -4.71351743e-01
-1.10385668e+00 7.13919759e-01 6.90257132e-01 -3.84308159e-01
-5.15528500e-01 1.05812109e+00 7.63080865e-02 -1.97063908e-01
2.17083737e-01 -2.19181597e-01 -8.77375603e-02 -2.07027838e-01
5.73631942e-01 8.26416314e-01 1.35531813e-01 -4.34455067e-01
-3.45556110e-01 9.90949929e-01 1.83196127e-01 -4.26425457e-01
1.13235211e+00 -5.62811971e-01 -2.35247403e-01 8.83701265e-01
1.48325670e+00 3.82020056e-01 -1.18740594e+00 -2.47376174e-01
1.35780973e-02 -2.87950158e-01 2.07903892e-01 -1.85679272e-01
-1.04360998e+00 9.11398828e-01 5.28589010e-01 6.18974388e-01
1.23876631e+00 -2.86980093e-01 5.18498600e-01 5.85751057e-01
1.91627130e-01 -1.39732993e+00 -6.90627277e-01 4.66446906e-01
5.87280333e-01 -9.55289423e-01 1.02665223e-01 -4.74405020e-01
-2.24044606e-01 1.44937360e+00 -2.16840237e-01 -4.35751200e-01
9.19945598e-01 8.18408579e-02 -8.48751441e-02 4.44921479e-02
-4.41612422e-01 -4.50721622e-01 4.34795059e-02 6.38599277e-01
8.06758940e-01 -3.33742321e-01 -1.16634989e+00 1.90553352e-01
-1.57190099e-01 -1.63440958e-01 3.22795272e-01 8.98236454e-01
-2.38156185e-01 -1.19993114e+00 -2.00576633e-01 8.05787966e-02
-8.13659668e-01 -3.72513264e-01 -2.65950471e-01 9.27032351e-01
-3.89057636e-01 7.51510859e-01 -3.33984613e-01 -3.91906276e-02
1.71285361e-01 1.55050918e-01 7.95882463e-01 -1.67556763e-01
-4.20770049e-01 -1.07927367e-01 8.75785053e-02 -4.92974758e-01
-4.81214941e-01 -7.95091569e-01 -6.13148034e-01 -5.62404454e-01
-4.03254181e-01 7.80036986e-01 1.19831765e+00 8.47235262e-01
-1.23234883e-01 2.22429976e-01 9.67429817e-01 -1.04892683e+00
-1.31554210e+00 -8.09184253e-01 -9.41019952e-01 -3.01986896e-02
5.14884889e-01 -4.73425090e-01 -7.09792972e-01 -1.81839213e-01] | [6.917212963104248, 4.270504474639893] |
8f9e7c5f-da5b-4a9b-a37d-fd740dc41fdb | interpretable-multimodal-sentiment-analysis | 2305.06162 | null | https://arxiv.org/abs/2305.06162v3 | https://arxiv.org/pdf/2305.06162v3.pdf | Interpretable multimodal sentiment analysis based on textual modality descriptions by using large-scale language models | Multimodal sentiment analysis is an important area for understanding the user's internal states. Deep learning methods were effective, but the problem of poor interpretability has gradually gained attention. Previous works have attempted to use attention weights or vector distributions to provide interpretability. However, their explanations were not intuitive and can be influenced by different trained models. This study proposed a novel approach to provide interpretability by converting nonverbal modalities into text descriptions and by using large-scale language models for sentiment predictions. This provides an intuitive approach to directly interpret what models depend on with respect to making decisions from input texts, thus significantly improving interpretability. Specifically, we convert descriptions based on two feature patterns for the audio modality and discrete action units for the facial modality. Experimental results on two sentiment analysis tasks demonstrated that the proposed approach maintained, or even improved effectiveness for sentiment analysis compared to baselines using conventional features, with the highest improvement of 2.49% on the F1 score. The results also showed that multimodal descriptions have similar characteristics on fusing modalities as those of conventional fusion methods. The results demonstrated that the proposed approach is interpretable and effective for multimodal sentiment analysis. | ['Shogo Okada', 'Sixia Li'] | 2023-05-07 | null | null | null | null | ['multimodal-sentiment-analysis', 'multimodal-sentiment-analysis'] | ['computer-vision', 'natural-language-processing'] | [ 1.96708009e-01 3.94647807e-01 -1.67999014e-01 -9.60197270e-01
-8.03069949e-01 -3.80308002e-01 5.34930408e-01 1.40654594e-01
-1.67601258e-01 5.94072282e-01 6.43926263e-01 1.37987584e-01
1.28556356e-01 -1.80438429e-01 -3.61206979e-01 -5.38568854e-01
4.25723463e-01 1.45585135e-01 -5.97787499e-01 -3.30201179e-01
1.85185105e-01 3.74426395e-02 -1.86479068e+00 1.05015588e+00
7.54034817e-01 1.26232719e+00 -2.56120443e-01 7.59523749e-01
-3.22400719e-01 8.25439334e-01 -6.95770264e-01 -7.41736829e-01
-3.54799181e-01 -5.29472232e-01 -8.07369530e-01 5.35922758e-02
2.78418452e-01 -4.85759228e-01 1.01121828e-01 8.36479306e-01
3.64477575e-01 1.75912589e-01 7.86049366e-01 -1.58926046e+00
-1.00659299e+00 7.68182456e-01 -5.24857402e-01 -2.61154741e-01
9.51882362e-01 -2.07481384e-01 1.20872736e+00 -9.52826440e-01
6.35020295e-03 1.64285743e+00 6.74543560e-01 7.60233581e-01
-1.06349969e+00 -8.15805256e-01 3.45188230e-01 3.13931316e-01
-1.00518727e+00 -5.83215415e-01 6.63568139e-01 -3.24271500e-01
9.71619904e-01 5.50899923e-01 5.03757179e-01 1.40746248e+00
3.07391435e-01 8.73282135e-01 1.04154420e+00 -4.63855803e-01
1.92827117e-02 6.19403780e-01 4.69992220e-01 5.85941732e-01
2.45852061e-02 -3.39190781e-01 -9.43940401e-01 -2.37659886e-01
1.21426187e-01 1.93582013e-01 -1.49828121e-01 3.03675890e-01
-1.10785162e+00 9.49899673e-01 3.72850299e-01 1.75454766e-01
-3.72247785e-01 2.02592164e-01 4.66779947e-01 2.00078711e-01
5.55602014e-01 3.83809656e-01 -3.32225204e-01 -3.20876598e-01
-4.54484433e-01 -2.13275179e-01 6.79598629e-01 4.51275766e-01
5.26478112e-01 4.95509664e-03 -1.37471750e-01 7.25342035e-01
7.85247087e-01 8.41466725e-01 7.28346586e-01 -9.95308936e-01
4.23030019e-01 6.64491177e-01 8.77684057e-02 -1.30858791e+00
-6.75997734e-01 3.25474858e-01 -7.11232305e-01 4.38234471e-02
1.34027272e-01 -4.02512163e-01 -8.99148166e-01 1.81009340e+00
2.52220239e-02 -2.50322998e-01 3.37819904e-01 1.18670213e+00
1.18894863e+00 6.89246476e-01 2.71493345e-01 3.49771678e-02
1.68976569e+00 -9.22935605e-01 -1.48310483e+00 -2.83273578e-01
4.71691698e-01 -7.88402855e-01 1.18124187e+00 4.83434051e-01
-9.38425064e-01 -6.56435251e-01 -9.51453745e-01 -9.69552323e-02
-4.20269966e-01 3.44880998e-01 9.43264246e-01 9.38153744e-01
-1.00633383e+00 2.09575370e-01 -8.93679500e-01 -4.45491761e-01
2.42652714e-01 6.36264682e-01 -5.30742288e-01 4.40641940e-01
-1.35901988e+00 9.86145854e-01 2.89779454e-01 3.48226428e-01
-2.59725213e-01 -4.23192717e-02 -1.03120995e+00 2.52449632e-01
-1.51847601e-01 -7.78160036e-01 1.45061970e+00 -1.85850799e+00
-1.82115197e+00 3.73544216e-01 -6.79522336e-01 -7.82893449e-02
-2.49884769e-01 -5.24195671e-01 -5.57350516e-01 3.03389072e-01
-1.51487499e-01 8.65524769e-01 9.99628365e-01 -1.36856782e+00
-5.87307215e-01 -4.50440109e-01 3.15575242e-01 5.34017861e-01
-7.88947880e-01 -5.83238937e-02 -1.53782070e-01 -3.60635430e-01
8.44091251e-02 -9.39154923e-01 1.79914728e-01 -1.00230008e-01
-3.06191742e-01 -3.08696091e-01 8.56818318e-01 -7.57856309e-01
1.25384235e+00 -2.09642363e+00 2.07586005e-01 2.02541053e-01
4.19824362e-01 -4.20569326e-04 -1.40100688e-01 4.35733020e-01
-2.94308543e-01 5.09390652e-01 6.82236999e-02 -8.47027123e-01
2.11012200e-01 2.67019719e-01 -4.37354058e-01 1.57585904e-01
3.37260425e-01 7.42398083e-01 -6.54995143e-01 -2.50064075e-01
2.87833095e-01 8.07669878e-01 -5.44657648e-01 3.40577513e-01
2.40715623e-01 4.21098322e-01 -4.74227816e-01 6.79721713e-01
4.26179439e-01 -2.34458983e-01 2.42699862e-01 -5.53829432e-01
3.89771968e-01 7.39167780e-02 -9.95695293e-01 1.45420218e+00
-5.57261407e-01 9.35264528e-01 -4.20974232e-02 -8.98406327e-01
6.71028078e-01 7.12966204e-01 1.97829604e-01 -4.61629003e-01
5.52836835e-01 -4.67153937e-02 -9.20428410e-02 -8.28657269e-01
5.89799762e-01 -4.45494682e-01 -3.72297972e-01 6.20489717e-01
1.37707666e-01 -1.72710687e-01 -2.85796463e-01 2.13383421e-01
4.08726364e-01 3.99078382e-03 3.54565471e-01 2.42844671e-01
5.73639750e-01 -3.43855679e-01 1.78129837e-01 5.46409130e-01
-1.53780252e-01 7.56735742e-01 5.59964180e-01 -3.71779740e-01
-3.24244827e-01 -7.09336519e-01 1.03243157e-01 1.34174848e+00
3.74619007e-01 -6.84683919e-01 -8.64486337e-01 -6.13251388e-01
-2.23083243e-01 8.85507941e-01 -8.94593894e-01 -4.31834757e-01
1.89964131e-01 -5.60664177e-01 5.50955594e-01 7.88156569e-01
2.94028521e-01 -9.37923789e-01 -3.18894714e-01 -1.90313622e-01
-8.55796635e-01 -1.14690924e+00 -2.97014058e-01 3.95974517e-02
-7.18141496e-01 -6.78591251e-01 -3.17489952e-01 -4.11526948e-01
8.50278378e-01 2.28580117e-01 7.65131533e-01 4.78650518e-02
3.32873702e-01 7.71733880e-01 -5.03535807e-01 -8.64338875e-01
-3.37285578e-01 -9.94757041e-02 3.87189120e-01 5.23587763e-01
6.25812829e-01 -2.63338350e-03 -4.35000718e-01 1.53212979e-01
-1.06396329e+00 2.04944760e-01 5.09460568e-01 9.32110548e-01
5.00696599e-02 -4.05598968e-01 8.57998312e-01 -7.07209289e-01
1.15664434e+00 -4.39931601e-01 3.14879179e-01 1.67863727e-01
-5.92582881e-01 2.85855383e-01 6.42768264e-01 -3.33854467e-01
-1.51326430e+00 6.02770224e-02 -1.84866592e-01 -1.92874283e-01
-4.09090310e-01 7.33299851e-01 3.13886479e-02 1.43759936e-01
4.40020025e-01 -2.32683361e-01 3.00004512e-01 -7.49092996e-02
3.75265002e-01 1.30131018e+00 1.47939473e-01 -4.60186779e-01
2.33639523e-01 4.53760266e-01 -4.96311069e-01 -6.66232765e-01
-8.96944225e-01 -3.34222943e-01 -5.01803815e-01 -4.06655341e-01
9.75859582e-01 -9.74334121e-01 -1.03522944e+00 2.62030482e-01
-1.34187508e+00 3.29276532e-01 1.22524343e-01 6.97243273e-01
-3.24361414e-01 5.02687275e-01 -5.92794418e-01 -1.13853621e+00
-4.53665197e-01 -1.31785452e+00 1.49385381e+00 4.33063805e-01
-1.05558968e+00 -1.24996030e+00 -1.58210710e-01 9.80100632e-01
4.64552969e-01 -1.02782091e-02 7.18273878e-01 -7.87696064e-01
2.87860394e-01 -3.50338161e-01 -2.51281947e-01 3.37093115e-01
2.89010227e-01 1.30158722e-01 -1.55992246e+00 9.26876217e-02
2.02676877e-02 -6.37580395e-01 7.52818108e-01 1.73540294e-01
1.15206444e+00 -4.07564282e-01 -1.63913757e-01 8.57363045e-02
7.92519629e-01 1.31658867e-01 5.23278475e-01 3.78088690e-02
6.04313016e-01 7.18153477e-01 6.20202303e-01 4.94934350e-01
5.65040052e-01 5.47951877e-01 5.90222180e-01 -4.00241315e-01
2.45465368e-01 -2.78926957e-02 8.99990022e-01 7.41942406e-01
-2.88648397e-01 -4.36527342e-01 -5.65074980e-01 2.32583717e-01
-2.20388365e+00 -8.59456241e-01 -1.24493659e-01 1.76909029e+00
6.57925546e-01 -1.96939483e-01 -9.81900394e-02 9.45755094e-02
6.13073170e-01 -1.21165656e-01 -3.63336325e-01 -1.13463449e+00
-5.98089881e-02 -1.76874101e-02 -1.73435494e-01 5.78067958e-01
-9.66338336e-01 6.02321863e-01 6.82116079e+00 4.70710486e-01
-1.34704387e+00 2.16196030e-02 6.48525953e-01 -1.30480677e-01
-4.80035335e-01 -3.72818589e-01 -6.06309116e-01 3.65504920e-01
1.27054036e+00 2.66712308e-01 2.36035451e-01 3.95465732e-01
5.13936579e-01 -2.39384994e-01 -1.14766133e+00 1.16951656e+00
4.95845795e-01 -7.86067665e-01 3.19246411e-01 -3.74198645e-01
5.14593124e-01 -5.29536724e-01 2.14377165e-01 3.61033291e-01
-2.83462405e-01 -1.13730872e+00 8.01956594e-01 7.91674554e-01
5.08961499e-01 -6.06490433e-01 1.27897918e+00 1.00705571e-01
-9.08704281e-01 -7.24694431e-02 9.82061103e-02 -4.76341575e-01
1.28251746e-01 9.69462171e-02 -9.24470842e-01 3.23156595e-01
7.23231614e-01 7.11326718e-01 -4.93582189e-01 1.50750145e-01
-2.82648444e-01 7.12548018e-01 -2.43525445e-01 -3.69139999e-01
6.36318773e-02 4.84794192e-02 2.19360888e-01 1.31569397e+00
3.78688395e-01 1.37019530e-01 -1.78193629e-01 6.44289613e-01
4.73274514e-02 7.77385458e-02 -6.15965486e-01 -4.01272058e-01
-3.85269080e-03 1.42762411e+00 -5.48189998e-01 -4.52651292e-01
-6.88652873e-01 9.93571401e-01 5.67719154e-02 5.29650688e-01
-1.26847279e+00 -3.31301540e-01 6.80930614e-01 -1.98581859e-01
-6.10707588e-02 9.02572274e-02 -5.42757392e-01 -1.34978795e+00
-1.37049362e-01 -9.47589040e-01 4.36488807e-01 -1.27869451e+00
-1.25858378e+00 8.40083897e-01 5.85678928e-02 -1.14878273e+00
-3.90436441e-01 -7.17438459e-01 -4.39834654e-01 7.90253341e-01
-1.22062635e+00 -1.41179585e+00 -4.08165514e-01 4.38445359e-01
5.69194257e-01 9.04273335e-03 1.34613860e+00 8.90662223e-02
-5.19100130e-01 7.79293537e-01 -1.96919277e-01 -1.82408765e-01
1.04515100e+00 -1.08291543e+00 -4.57192838e-01 5.00734270e-01
5.90227880e-02 8.83239925e-01 7.34281063e-01 -2.23778889e-01
-1.35073578e+00 -5.03346741e-01 9.03421581e-01 -5.98022044e-01
4.68448669e-01 -1.43066078e-01 -7.10026562e-01 7.29987502e-01
9.07453716e-01 -5.95203340e-01 1.48616338e+00 4.19802755e-01
-1.55715480e-01 -7.67897367e-02 -1.22003126e+00 5.59444070e-01
4.08658326e-01 -5.84967732e-01 -8.39197993e-01 1.57357939e-02
5.73953867e-01 -3.00880313e-01 -7.75613308e-01 2.83315301e-01
1.04018152e+00 -9.44913924e-01 6.01356089e-01 -7.26215780e-01
7.40507483e-01 -2.79606521e-01 -2.46402383e-01 -1.38685179e+00
-1.71136320e-01 -3.73671234e-01 -1.91013277e-01 1.23571563e+00
7.06650198e-01 -6.76628470e-01 2.84443080e-01 1.24546373e+00
-3.20903063e-02 -5.23280859e-01 -5.03750801e-01 1.31085798e-01
-5.88455737e-01 -6.64846957e-01 4.59395736e-01 8.30304861e-01
6.40031099e-01 8.54423523e-01 -4.49868768e-01 1.36867866e-01
-2.11326918e-03 3.59167010e-02 6.32052362e-01 -9.97776687e-01
1.52043290e-02 -4.35405821e-01 -4.74870890e-01 -8.36782038e-01
3.93217742e-01 -7.53493488e-01 3.38478908e-02 -1.56039977e+00
3.48563433e-01 3.27609062e-01 -3.95739943e-01 9.56591487e-01
-4.80858088e-01 5.54139376e-01 1.82512090e-01 -3.86702046e-02
-7.54224181e-01 6.71169221e-01 1.11675012e+00 -1.99401155e-01
-1.13579288e-01 -4.63929027e-02 -1.41321540e+00 9.93333995e-01
7.59846568e-01 -1.26068681e-01 -4.44485188e-01 -5.49099386e-01
5.12924314e-01 -6.17049709e-02 2.22144663e-01 -5.83682895e-01
1.05284601e-01 -1.06406406e-01 4.96064782e-01 -3.45778555e-01
8.28501225e-01 -1.09743404e+00 7.69575462e-02 8.26463625e-02
-5.89373052e-01 3.00404690e-02 6.03872657e-01 5.69331884e-01
-6.11327291e-01 -9.91310775e-02 2.51261324e-01 4.22715336e-01
-8.46369028e-01 -4.23114389e-01 -8.59858513e-01 -6.20381296e-01
8.44134748e-01 -3.74815762e-01 4.84854262e-03 -1.16070056e+00
-1.05003774e+00 1.02764942e-01 -1.01863876e-01 7.63123631e-01
9.03258026e-01 -1.46447885e+00 -4.19382870e-01 3.97205979e-01
1.26607448e-01 -4.98487115e-01 1.95209309e-01 1.02514184e+00
-1.41166985e-01 4.85511184e-01 -1.39239937e-01 -7.53415704e-01
-1.57550287e+00 -1.04664698e-01 2.00128347e-01 1.41008064e-01
1.00611635e-01 7.11479783e-01 3.13248545e-01 -4.86205190e-01
2.13488266e-01 -6.41727984e-01 -6.17733777e-01 4.61261809e-01
6.94498718e-01 4.56916392e-02 4.11901511e-02 -7.99452901e-01
-4.61958170e-01 5.56502104e-01 4.78990190e-02 -2.42368728e-01
1.16187906e+00 -5.92059851e-01 -8.29145536e-02 8.04266274e-01
1.11946940e+00 -9.04164091e-02 -6.65690124e-01 8.89153704e-02
-4.28085804e-01 -3.08912605e-01 1.06529333e-01 -1.00113475e+00
-1.06996417e+00 1.06241417e+00 6.43477261e-01 4.32505727e-01
1.30889761e+00 -8.19741264e-02 6.88410461e-01 4.84965682e-01
-7.11933225e-02 -9.88649786e-01 2.09557027e-01 5.13480544e-01
1.02952528e+00 -1.75375700e+00 -1.70752749e-01 -2.59458572e-01
-1.25231898e+00 1.66308308e+00 6.80065095e-01 4.10045356e-01
5.01164496e-01 1.01179868e-01 4.07238036e-01 -2.32719734e-01
-6.73356831e-01 9.97680426e-02 6.95144534e-01 3.45855117e-01
9.62082684e-01 1.36052892e-01 -1.29984692e-01 1.36303282e+00
-1.60823345e-01 -1.16685785e-01 3.90745223e-01 6.31529570e-01
-1.98373437e-01 -7.87238240e-01 -6.86375201e-01 4.07070249e-01
-5.98499656e-01 1.07978545e-02 -8.25572491e-01 6.09834552e-01
-1.25662163e-01 1.43934906e+00 1.98031142e-02 -7.75869608e-01
2.21722096e-01 2.71724910e-01 2.55274087e-01 -5.98525465e-01
-6.59444749e-01 1.62194658e-03 3.04339141e-01 -5.99769413e-01
-9.59076345e-01 -5.22600591e-01 -1.54292226e+00 -2.50619173e-01
-6.86937392e-01 2.31872037e-01 7.02435672e-01 1.29051828e+00
5.40435135e-01 6.05532169e-01 5.36927402e-01 -1.01703060e+00
-2.51189351e-01 -1.11423635e+00 -3.25221151e-01 6.20832503e-01
3.90843272e-01 -6.33265734e-01 -6.43635750e-01 4.60738331e-01] | [13.215063095092773, 5.227729797363281] |
59fac8e4-0784-427d-b6a4-ea080a18e6a5 | cinematic-l1-video-stabilization-with-a-log | 2011.08144 | null | https://arxiv.org/abs/2011.08144v2 | https://arxiv.org/pdf/2011.08144v2.pdf | Cinematic-L1 Video Stabilization with a Log-Homography Model | We present a method for stabilizing handheld video that simulates the camera motions cinematographers achieve with equipment like tripods, dollies, and Steadicams. We formulate a constrained convex optimization problem minimizing the $\ell_1$-norm of the first three derivatives of the stabilized motion. Our approach extends the work of Grundmann et al. [9] by solving with full homographies (rather than affinities) in order to correct perspective, preserving linearity by working in log-homography space. We also construct crop constraints that preserve field-of-view; model the problem as a quadratic (rather than linear) program to allow for an $\ell_2$ term encouraging fidelity to the original trajectory; and add constraints and objectives to reduce distortion. Furthermore, we propose new methods for handling salient objects via both inclusion constraints and centering objectives. Finally, we describe a windowing strategy to approximate the solution in linear time and bounded memory. Our method is computationally efficient, running at 300fps on an iPhone XS, and yields high-quality results, as we demonstrate with a collection of stabilized videos, quantitative and qualitative comparisons to [9] and other methods, and an ablation study. | ['Rudolph van der Merwe', 'Joseph Triscari', 'Jason Klivington', 'Arwen Bradley'] | 2020-11-16 | null | null | null | null | ['video-stabilization'] | ['computer-vision'] | [ 1.86511323e-01 2.02099085e-01 -1.89654574e-01 1.71866938e-02
-6.77851379e-01 -8.18048775e-01 1.46379516e-01 -2.74257720e-01
-3.15206587e-01 5.93380630e-01 2.03779295e-01 -1.70158654e-01
-1.41170416e-02 -2.91056514e-01 -9.54019248e-01 -4.26518023e-01
-2.45373115e-01 -1.09709300e-01 4.30961639e-01 -1.29714027e-01
4.42647606e-01 5.83048820e-01 -1.33760250e+00 3.46411573e-04
7.35132873e-01 8.78209293e-01 2.61854529e-01 9.20790732e-01
6.11360133e-01 9.29172635e-01 -2.18567237e-01 -4.04948026e-01
5.42315304e-01 -3.52178633e-01 -7.04878867e-01 6.77174926e-01
1.04577494e+00 -6.38538837e-01 -4.56669509e-01 9.47860479e-01
3.15442085e-01 3.81361544e-01 1.37051091e-01 -1.17338014e+00
-4.89060879e-01 -1.22346222e-01 -8.73972893e-01 3.65145117e-01
6.39748752e-01 2.75395848e-02 8.25870752e-01 -8.54819894e-01
1.06143987e+00 1.23226500e+00 7.43775010e-01 3.30406308e-01
-1.25474524e+00 -7.92535320e-02 2.67733842e-01 2.24826485e-02
-1.39555430e+00 -7.10332632e-01 6.61256135e-01 -3.76534760e-01
7.27296650e-01 4.80172098e-01 7.89819658e-01 5.58232844e-01
2.78419942e-01 6.50595546e-01 6.09231591e-01 -3.83389235e-01
-1.41258799e-02 1.18113019e-01 -2.18112320e-01 1.02671218e+00
-1.13280546e-02 -5.56573458e-02 -5.81378996e-01 -2.36904532e-01
1.27160501e+00 -1.13681749e-01 -7.38400877e-01 -7.69083023e-01
-1.42385709e+00 7.13255286e-01 6.53185546e-02 -1.58656668e-02
-1.88839003e-01 2.55726635e-01 7.19116256e-03 1.08372994e-01
5.72644830e-01 3.97050798e-01 -1.29383236e-01 -3.34514380e-01
-1.14830029e+00 4.66012090e-01 7.52963603e-01 1.49334693e+00
3.78351331e-01 1.00365020e-01 1.31726697e-01 4.82101500e-01
1.54767305e-01 4.90307510e-01 5.11086592e-03 -1.82654285e+00
5.01076460e-01 9.64956954e-02 5.31197608e-01 -1.45614588e+00
-3.00169468e-01 -3.66421305e-02 -3.29127938e-01 4.13517237e-01
5.15305459e-01 -2.26868764e-01 -3.41472447e-01 1.58549464e+00
3.48826796e-01 2.89767712e-01 -4.07356143e-01 1.15843177e+00
2.71236062e-01 7.78639436e-01 -5.94061434e-01 -6.33142769e-01
9.97138143e-01 -1.23918664e+00 -9.05327678e-01 -4.79152277e-02
5.47568381e-01 -9.73142564e-01 1.21044767e+00 4.56837028e-01
-1.90369201e+00 -2.35375315e-01 -9.89944160e-01 -3.33883494e-01
3.58386219e-01 1.85200632e-01 1.93925887e-01 5.05225599e-01
-1.45301092e+00 9.23339367e-01 -1.07017279e+00 -3.07475567e-01
4.93039959e-04 4.28683370e-01 -2.52503246e-01 1.78841189e-01
-5.55973589e-01 9.31454301e-01 -8.51288289e-02 -1.31584555e-01
-3.98970246e-01 -9.48965728e-01 -8.69547367e-01 9.47687402e-02
6.27634168e-01 -6.16609454e-01 1.12812960e+00 -9.06004131e-01
-1.68775213e+00 1.05710769e+00 -2.98372954e-01 -1.76298127e-01
6.92277193e-01 -3.98691535e-01 2.59028114e-02 6.55264199e-01
-4.06548083e-02 6.37563050e-01 7.75070429e-01 -1.27979541e+00
-5.24281621e-01 -9.90269035e-02 4.36125040e-01 6.21466756e-01
-4.24179524e-01 1.00527868e-01 -8.47794831e-01 -7.74272203e-01
1.34911329e-01 -1.24561501e+00 -2.07269683e-01 5.29249132e-01
-1.83055282e-01 6.08567059e-01 9.06822145e-01 -9.67085183e-01
1.45051193e+00 -2.22627473e+00 6.02656484e-01 1.15684785e-01
3.55122358e-01 7.11192861e-02 1.16727099e-01 2.78119177e-01
1.01817206e-01 -2.01404035e-01 -2.88787484e-01 -5.67219496e-01
-2.45084420e-01 2.68362835e-02 -3.22804868e-01 8.09527576e-01
-2.99730450e-01 3.85255277e-01 -8.50699008e-01 -6.10897899e-01
3.67889106e-01 4.48826879e-01 -9.32687283e-01 4.07279916e-02
1.88617840e-01 3.24755877e-01 -1.00814939e-01 5.43893695e-01
6.22450054e-01 -2.24049866e-01 1.84513509e-01 -3.33781600e-01
-4.48717743e-01 -2.48654410e-01 -1.50795877e+00 1.75187111e+00
-3.61821532e-01 8.17548633e-01 8.05651546e-01 -6.44190907e-01
5.20741463e-01 1.57607853e-01 7.90744007e-01 -8.93788114e-02
3.67245302e-02 2.87826471e-02 -5.30927479e-01 -4.72762555e-01
8.35642457e-01 -1.61252730e-02 3.76495630e-01 1.47391647e-01
-1.39754817e-01 -5.80482364e-01 3.02006960e-01 5.13483942e-01
8.50691795e-01 2.57313669e-01 7.41872191e-02 -6.27002418e-01
3.49643648e-01 5.27785383e-02 4.91282254e-01 3.04188579e-01
-2.53025204e-01 1.03801548e+00 5.39680183e-01 -2.66006976e-01
-1.23578942e+00 -9.08327639e-01 -1.67046546e-03 7.58809686e-01
5.73445559e-01 -4.28607672e-01 -9.36030567e-01 -2.31837422e-01
-7.41074607e-02 4.92331982e-01 -3.74628544e-01 2.01525688e-01
-8.72347891e-01 -1.50872663e-01 6.78794309e-02 6.09358191e-01
3.20500612e-01 -5.58025479e-01 -9.32274938e-01 7.96663240e-02
-3.12854469e-01 -1.29217184e+00 -1.28058505e+00 -3.78381729e-01
-1.11186588e+00 -1.17941988e+00 -9.08769250e-01 -8.20684135e-01
8.84698808e-01 6.29108608e-01 1.02138543e+00 1.50782377e-01
-1.88446209e-01 8.04033816e-01 -7.55355284e-02 1.80054381e-01
1.28711358e-01 -4.39797521e-01 1.90261260e-01 -8.72854963e-02
-5.78148246e-01 -5.03562450e-01 -7.08852947e-01 4.81846064e-01
-9.29874480e-01 2.46592745e-01 -1.49827316e-01 6.48834765e-01
7.49018669e-01 -2.59923995e-01 -2.95584023e-01 -5.12687147e-01
3.76569182e-01 -7.46143237e-02 -9.48340654e-01 1.02153420e-01
-3.32188904e-01 -2.99805909e-01 6.21269166e-01 -4.95946109e-01
-9.93021250e-01 2.42988437e-01 1.81312963e-01 -1.03534687e+00
5.72081745e-01 1.36427999e-01 -2.71529406e-02 -5.52079201e-01
4.29641992e-01 8.32490698e-02 1.28417835e-01 -1.80661812e-01
4.45374638e-01 6.78152144e-02 7.78504550e-01 -5.38511336e-01
7.14847088e-01 9.61805224e-01 1.03916831e-01 -1.06503594e+00
-5.80518246e-01 -4.45374131e-01 -7.78080344e-01 -4.39058632e-01
7.08390951e-01 -7.33861923e-01 -8.98028910e-01 1.90434575e-01
-1.06936407e+00 -3.19183022e-01 -3.72296542e-01 6.72422111e-01
-1.01516163e+00 7.22630084e-01 -6.84387743e-01 -7.02163756e-01
-1.31968232e-02 -1.11741126e+00 1.08726668e+00 1.57155290e-01
-2.60609239e-01 -1.16970408e+00 1.83710307e-01 2.42302924e-01
2.32180282e-01 5.04193127e-01 2.32614368e-01 2.78154910e-01
-7.83986509e-01 2.20750552e-02 -3.15811075e-02 1.89500958e-01
-1.18824713e-01 3.29698116e-01 -4.30377185e-01 -7.81391323e-01
2.45649800e-01 4.54183668e-03 3.20355237e-01 8.59849751e-01
9.92809832e-01 -5.85522890e-01 -3.21863711e-01 1.01791966e+00
1.33336818e+00 2.81248748e-01 5.50538361e-01 3.09423447e-01
7.30417252e-01 5.51680863e-01 7.11967230e-01 4.97475892e-01
3.56932789e-01 9.12090719e-01 3.70654255e-01 -1.18680589e-01
4.85014059e-02 -1.48842692e-01 4.67593908e-01 7.17350245e-01
-4.84297723e-01 -2.12186873e-01 -4.16650504e-01 6.36804223e-01
-2.03483701e+00 -1.13989806e+00 -1.92552850e-01 2.37226009e+00
3.74345601e-01 -1.59954861e-01 3.60385567e-01 -5.48722930e-02
7.79838204e-01 3.41593802e-01 -3.81518871e-01 -2.57312387e-01
-5.84004261e-02 -2.12108895e-01 7.89562166e-01 1.03998077e+00
-1.07005823e+00 6.96196914e-01 6.96512127e+00 4.32693571e-01
-1.03007805e+00 -7.75741115e-02 4.76965219e-01 -6.66924000e-01
-2.39347056e-01 8.01154077e-02 -4.76828188e-01 4.78442043e-01
5.40715635e-01 -4.18389797e-01 7.14941025e-01 8.31110299e-01
4.97932643e-01 -1.62074491e-01 -1.05672133e+00 1.18892896e+00
4.93651122e-01 -1.51393342e+00 -4.91627455e-01 1.24318294e-01
7.50789642e-01 -4.54585433e-01 1.63137943e-01 -3.69912744e-01
-6.61710417e-03 -5.09034395e-01 1.14542186e+00 4.06819403e-01
7.72068024e-01 -6.60086930e-01 1.18079267e-01 2.28059188e-01
-1.27518797e+00 9.07202214e-02 -2.50445247e-01 -8.06789845e-02
7.11124301e-01 2.14827538e-01 -1.88355908e-01 2.78504282e-01
8.04489732e-01 7.68997431e-01 -1.27147898e-01 1.04883385e+00
1.86895151e-02 2.29373381e-01 -4.07349944e-01 2.81597495e-01
2.00723693e-01 -5.99402130e-01 9.63491499e-01 1.08627558e+00
5.33484817e-01 7.83698201e-01 2.13796794e-01 4.65240747e-01
1.01571701e-01 9.65804309e-02 -5.55105209e-01 2.32323870e-01
1.81226104e-01 1.12564588e+00 -9.35987473e-01 -3.43164921e-01
-3.67423624e-01 1.30378962e+00 -1.97217334e-02 3.37180078e-01
-9.26093459e-01 -4.80354965e-01 4.99326706e-01 5.51163375e-01
3.21333081e-01 -5.53743064e-01 -2.26923674e-01 -1.63499415e+00
3.59243065e-01 -7.85988033e-01 2.39181548e-01 -1.06895840e+00
-6.03131294e-01 4.96411860e-01 2.00066298e-01 -1.46469724e+00
-1.84266120e-01 -4.32658702e-01 -5.41555941e-01 5.29528558e-01
-1.06064248e+00 -8.29678297e-01 -2.49024838e-01 8.19344997e-01
5.73611617e-01 3.81314039e-01 3.16985399e-01 4.76500064e-01
-3.08664858e-01 3.59530479e-01 1.96603656e-01 -3.73267859e-01
7.39966750e-01 -1.19600689e+00 1.01773374e-01 1.05225742e+00
-1.89020351e-01 6.99810684e-01 9.88900125e-01 -4.60086435e-01
-1.66653204e+00 -5.67312777e-01 7.12290525e-01 -3.82959038e-01
7.05953360e-01 -1.62567765e-01 -7.26411402e-01 1.10957468e+00
3.47624063e-01 1.57659635e-01 1.97544098e-01 -3.18324029e-01
2.10831895e-01 4.42171656e-02 -1.31149852e+00 7.34088838e-01
1.22236693e+00 -2.79705346e-01 -2.77904212e-01 4.72992182e-01
7.10461080e-01 -1.08129311e+00 -7.57140160e-01 4.02610339e-02
7.71512926e-01 -1.13976598e+00 1.07324934e+00 -2.24830404e-01
4.38508660e-01 -3.32261980e-01 -1.87754810e-01 -1.26491845e+00
-3.93971652e-01 -1.22855914e+00 -2.77798831e-01 7.68610656e-01
8.25056434e-02 -2.37309262e-01 9.14584637e-01 8.99676144e-01
-3.55201334e-01 -8.43565822e-01 -8.34839880e-01 -8.18931460e-01
-2.12688819e-01 -5.48316957e-03 -6.02965951e-02 9.80028033e-01
3.95813853e-01 -8.08543861e-02 -8.46078336e-01 1.83187202e-01
7.66283035e-01 -6.86668009e-02 6.60207272e-01 -6.19635761e-01
-4.23884004e-01 -2.39240199e-01 -3.29308093e-01 -1.48066247e+00
-5.54629490e-02 -3.12891304e-01 -1.02813259e-01 -1.14042008e+00
1.11251371e-02 -6.57559261e-02 3.73784065e-01 2.72687338e-02
4.60085757e-02 2.78147846e-01 7.29001403e-01 2.00561896e-01
-5.56646526e-01 3.89961034e-01 1.37900567e+00 1.63970947e-01
-4.40150589e-01 -2.99073812e-02 -4.45244223e-01 1.04788220e+00
2.83667803e-01 -1.40553311e-01 -5.27594328e-01 -7.49357045e-01
-4.30137990e-03 6.45132780e-01 3.33870530e-01 -8.83738160e-01
3.53219956e-01 -3.13426375e-01 8.54880810e-02 -6.18045032e-01
7.25119054e-01 -8.58556628e-01 2.08472192e-01 3.91730249e-01
-2.61762738e-01 5.46929419e-01 2.03496993e-01 3.41871619e-01
-1.12786576e-01 -1.21699810e-01 1.06153154e+00 5.49983047e-02
-4.41382796e-01 2.99654573e-01 -3.11000675e-01 1.08524859e-01
1.24743164e+00 -4.19462979e-01 -1.25885144e-01 -9.53364670e-01
-9.66656029e-01 2.69502223e-01 1.04625285e+00 1.98564380e-01
7.05352485e-01 -1.31392300e+00 -3.39402169e-01 -2.55650319e-02
-4.99228716e-01 -2.39846483e-01 3.76577586e-01 9.97639537e-01
-1.15581667e+00 3.22160572e-01 -6.69616833e-02 -6.01552725e-01
-1.62519205e+00 6.67270541e-01 2.39862546e-01 -8.65476355e-02
-5.55489063e-01 7.48441100e-01 1.90485582e-01 -8.74728784e-02
2.49803975e-01 -3.50446254e-01 1.80833787e-02 -1.79027066e-01
5.04870713e-01 9.15736437e-01 -1.32422850e-01 -7.32671559e-01
-4.45010543e-01 9.61550713e-01 2.41377845e-01 -5.16043246e-01
1.23141837e+00 -5.85541427e-01 7.34170899e-02 9.15285796e-02
1.28297055e+00 4.17361677e-01 -1.70267284e+00 5.91552891e-02
-4.62572157e-01 -9.63793814e-01 3.16938199e-02 -1.23727821e-01
-1.15321302e+00 5.90925276e-01 3.80978733e-01 1.24657616e-01
1.08582187e+00 -3.03696305e-01 7.50690758e-01 2.13198796e-01
2.81484365e-01 -1.09935927e+00 6.77778572e-02 3.25510919e-01
9.25754070e-01 -8.56316149e-01 3.82519037e-01 -7.38506973e-01
-7.38358140e-01 1.07184470e+00 5.05620718e-01 -4.50433254e-01
6.35613859e-01 3.99427831e-01 -1.24836430e-01 -7.38506485e-03
-5.22898257e-01 3.09198111e-01 2.10077599e-01 2.27342352e-01
3.84710759e-01 -3.24984640e-01 -5.21386147e-01 -4.91980389e-02
-7.66447857e-02 1.08540311e-01 8.53990078e-01 1.26789999e+00
-4.26425278e-01 -7.69186199e-01 -6.54968143e-01 8.73614028e-02
-4.98016566e-01 -9.24259946e-02 -7.87250623e-02 8.05404484e-01
-2.53182650e-01 9.62135077e-01 1.89821526e-01 -2.27812618e-01
6.58998907e-01 -4.95374352e-01 6.74090922e-01 -4.04824078e-01
-3.27346891e-01 4.12424147e-01 6.61432371e-02 -9.32641804e-01
-5.43103576e-01 -8.39228630e-01 -9.46536362e-01 -7.39135385e-01
-2.58447230e-01 8.98075625e-02 2.64722288e-01 5.48974454e-01
5.12010157e-01 1.47193864e-01 8.87756109e-01 -1.30250180e+00
-2.99969971e-01 -3.40564936e-01 -6.30735457e-01 3.56014311e-01
3.79957616e-01 -5.71085095e-01 -4.57719892e-01 7.10524738e-01] | [10.590492248535156, -1.376934289932251] |
84146c69-146a-4e1f-9cc4-cc65b54c6caa | historical-and-modern-features-for-buddha | 1909.12921 | null | https://arxiv.org/abs/1909.12921v2 | https://arxiv.org/pdf/1909.12921v2.pdf | Historical and Modern Features for Buddha Statue Classification | While Buddhism has spread along the Silk Roads, many pieces of art have been displaced. Only a few experts may identify these works, subjectively to their experience. The construction of Buddha statues was taught through the definition of canon rules, but the applications of those rules greatly varies across time and space. Automatic art analysis aims at supporting these challenges. We propose to automatically recover the proportions induced by the construction guidelines, in order to use them and compare between different deep learning features for several classification tasks, in a medium size but rich dataset of Buddha statues, collected with experts of Buddhism art history. | ['Matheus Oliveira Franca', 'Yutaka Fujioka', 'Jueren Wang', 'Ayaka Uesaka', 'Van Le', 'Hajime Nagahara', 'Yuta Nakashima', 'Noa Garcia', 'Jacob Chan', 'Benjamin Renoust'] | 2019-09-17 | null | null | null | null | ['art-analysis'] | ['computer-vision'] | [-2.63396114e-01 3.79600376e-02 -2.32909173e-01 -1.37540400e-01
-3.52303952e-01 -1.04516232e+00 1.24868822e+00 1.26446605e-01
-1.38789132e-01 2.87847251e-01 5.11566341e-01 -2.41063144e-02
-2.92568296e-01 -1.23518944e+00 -5.30059159e-01 -5.52214265e-01
3.83549273e-01 1.17599297e+00 -8.37194994e-02 -4.75061536e-01
4.83478963e-01 6.35302365e-01 -1.38775122e+00 4.12626475e-01
4.15959746e-01 7.14307368e-01 -3.15711387e-02 2.74061114e-01
-4.53282386e-01 8.91081035e-01 -6.01045191e-01 -9.31048334e-01
2.78290212e-01 -5.24704933e-01 -1.07054162e+00 8.15288797e-02
6.07270956e-01 -3.35140586e-01 -5.42613566e-01 6.04400635e-01
2.67034739e-01 -1.58484429e-01 9.37589467e-01 -5.86780012e-01
-1.24371135e+00 1.12012088e+00 -2.05148980e-01 -1.08609855e-01
1.58195883e-01 -1.16200879e-01 1.47924411e+00 -8.20937097e-01
1.00165617e+00 1.21489739e+00 7.46863723e-01 3.21884751e-01
-1.34727263e+00 -5.14151335e-01 -4.22076762e-01 3.25433582e-01
-1.48130202e+00 -1.95581406e-01 1.27371871e+00 -9.69892740e-01
3.84732515e-01 -6.50702268e-02 1.35260880e+00 1.25521970e+00
-3.85353379e-02 8.45639646e-01 1.04191732e+00 -4.61690068e-01
3.26364964e-01 -1.19558170e-01 -7.19378367e-02 6.81549788e-01
-6.97299987e-02 -3.30501318e-01 -5.69593251e-01 1.43295765e-01
9.35182452e-01 1.56955481e-01 1.82883292e-01 -2.43021071e-01
-8.29933703e-01 1.17609024e+00 3.69777679e-01 8.43342006e-01
-3.80774468e-01 2.77762283e-02 3.04960757e-01 3.08199227e-01
3.93332541e-01 4.72631752e-01 -2.18657628e-01 -1.40355363e-01
-1.15414417e+00 5.31370163e-01 7.09256232e-01 4.08676088e-01
8.02251518e-01 2.43823677e-02 7.37729147e-02 1.07789767e+00
6.40446320e-02 6.14544332e-01 2.46156305e-01 -9.50055778e-01
2.79711261e-02 8.05085480e-01 -2.36139894e-01 -1.08106828e+00
-2.59482980e-01 -3.84459794e-01 -6.79928243e-01 3.29212338e-01
7.14109898e-01 3.82079482e-01 -1.08718693e+00 1.35389292e+00
2.72815943e-01 -4.92242038e-01 -5.58450937e-01 7.33087540e-01
7.96373248e-01 5.41669250e-01 -2.09776625e-01 1.16692200e-01
1.43761873e+00 -5.45593679e-01 -6.55373156e-01 8.58465061e-02
5.34102082e-01 -7.55609989e-01 1.39434040e+00 6.71949625e-01
-1.16573215e+00 -3.92194688e-01 -1.00921631e+00 -4.49000806e-01
-2.67891645e-01 1.66857123e-01 7.21868813e-01 6.53325200e-01
-7.18573928e-01 8.41063619e-01 -7.28191733e-01 -4.82855022e-01
8.78674984e-01 1.49911880e-01 1.26387961e-02 2.01168463e-01
-9.28885996e-01 1.06726921e+00 3.23442847e-01 2.06963778e-01
-9.84410226e-01 -5.28160274e-01 -5.64918935e-01 -1.83342680e-01
2.93416739e-01 -3.55099469e-01 1.03264093e+00 -1.08453417e+00
-1.49943876e+00 1.61926818e+00 5.61599135e-01 -3.50451827e-01
5.61127186e-01 -3.62966299e-01 -4.60720778e-01 -1.08301148e-01
1.64416730e-02 1.35309219e-01 7.86881089e-01 -1.29166818e+00
-3.72242540e-01 -6.11409009e-01 2.75552660e-01 -2.34338388e-01
-1.64922819e-01 -3.47333625e-02 -2.72891939e-01 -8.96090209e-01
4.06205177e-01 -1.02554965e+00 2.50957191e-01 -1.52313530e-01
-2.60470062e-01 -3.56554955e-01 6.53964758e-01 -8.79941285e-01
1.11626816e+00 -1.88810718e+00 3.32882643e-01 1.90492854e-01
4.34589744e-01 -1.11651242e-01 2.61108398e-01 6.78478658e-01
5.42818248e-01 1.48395494e-01 -3.18283916e-01 5.81517890e-02
1.60687372e-01 3.48313421e-01 -5.47076702e-01 4.89580065e-01
-3.19654256e-01 8.44298840e-01 -7.18941092e-01 -5.78062415e-01
3.65682155e-01 7.40539134e-01 -2.86863923e-01 -5.99470846e-02
-2.13507146e-01 3.24302077e-01 -4.62867349e-01 9.79200542e-01
5.58141649e-01 -2.75529057e-01 6.80951178e-01 -1.09377056e-01
-6.57559931e-02 4.93599176e-01 -7.67724097e-01 1.96655715e+00
-6.46101296e-01 9.87342715e-01 -1.79136217e-01 -8.68109703e-01
8.99625659e-01 2.25859165e-01 5.25995255e-01 -7.52895892e-01
4.48589802e-01 2.44339079e-01 2.25252640e-02 -4.05934334e-01
3.44551921e-01 -3.34969252e-01 -2.95919716e-01 5.79333663e-01
1.96525246e-01 -6.08006179e-01 9.23737139e-02 6.62010759e-02
1.00984609e+00 3.24461043e-01 4.76799995e-01 -3.12623024e-01
2.30305567e-01 1.22540914e-01 3.99701089e-01 4.57864970e-01
1.19826555e-01 8.00734997e-01 2.13523924e-01 -1.41168666e+00
-1.26779437e+00 -1.61793804e+00 -3.51707667e-01 1.30782354e+00
-2.01978713e-01 -5.03238618e-01 -7.71288514e-01 -7.54994273e-01
-1.56251028e-01 7.15311348e-01 -9.34542120e-01 2.03868508e-01
-9.28221941e-01 -4.58235532e-01 5.39400220e-01 2.19692022e-01
3.85777116e-01 -1.37619722e+00 -6.65392816e-01 2.85073727e-01
3.60600464e-02 -7.29807019e-01 3.03895883e-02 -1.06991768e-01
-5.51237106e-01 -9.90534842e-01 -4.54390824e-01 -7.26507127e-01
1.53891921e-01 -2.37748489e-01 1.40159750e+00 1.55401021e-01
-2.36933500e-01 1.27898872e-01 -3.45721245e-01 -4.17733431e-01
-6.42249823e-01 3.98459435e-01 -1.55920297e-01 -4.84630577e-02
4.98492092e-01 -1.04271126e+00 -3.98873478e-01 1.04620680e-01
-6.53852463e-01 3.17128524e-02 5.48919559e-01 3.35208446e-01
4.26912338e-01 -6.18168451e-02 1.98404163e-01 -1.16164315e+00
5.26295882e-03 -4.43307847e-01 -2.45842323e-01 1.55864716e-01
-3.92602056e-01 -2.44809426e-02 4.45320100e-01 -4.28467959e-01
-8.72400403e-01 -5.98332919e-02 -2.42116854e-01 -5.90304621e-02
-6.30131140e-02 1.80488825e-01 -1.74056411e-01 2.87502527e-01
8.37218165e-01 4.65557761e-02 -5.80083847e-01 -7.01988816e-01
6.14758432e-01 5.21040916e-01 5.33865213e-01 -7.81822503e-01
9.38873887e-01 6.80925965e-01 -1.09928660e-01 -9.44937766e-01
-1.15458894e+00 1.06983989e-01 -1.13084102e+00 -4.14680600e-01
9.27202761e-01 -7.66699493e-01 -6.58850431e-01 1.58125311e-01
-1.08894813e+00 -3.05639118e-01 -5.60509086e-01 2.28390113e-01
-5.49122155e-01 -7.17682168e-02 -7.71692038e-01 -8.79002035e-01
-2.28063598e-01 -6.14459097e-01 7.88956046e-01 3.44581641e-02
-6.91875756e-01 -1.07243729e+00 5.54120064e-01 5.35204053e-01
6.67743087e-02 5.92005014e-01 1.17860091e+00 -5.75652421e-01
-4.43278462e-01 1.38532221e-01 3.26078326e-01 1.15332738e-01
4.34956044e-01 1.24586299e-01 -1.01968956e+00 1.17964901e-01
2.26582974e-01 -4.21021968e-01 8.16604376e-01 3.56906474e-01
6.93507075e-01 -5.42342961e-01 5.21841869e-02 5.00596285e-01
1.48560238e+00 2.79792160e-01 7.11539030e-01 6.66929126e-01
1.01445794e+00 5.83224297e-01 -6.57010674e-02 4.02014732e-01
2.64030904e-01 7.62977481e-01 6.12377115e-02 1.99103162e-01
-3.72027814e-01 -2.84970284e-01 2.46437833e-01 1.05282521e+00
-8.65311980e-01 9.04719830e-02 -1.22797668e+00 8.18474174e-01
-1.61149323e+00 -9.32371020e-01 -3.19153726e-01 1.83183968e+00
8.90605807e-01 4.44194853e-01 4.56052452e-01 3.14070284e-01
4.56200629e-01 3.92373890e-01 -1.78885937e-01 -3.28310281e-01
-3.67587805e-01 6.13611221e-01 2.15697721e-01 1.59973681e-01
-9.52628136e-01 1.10108352e+00 7.53358936e+00 9.89013553e-01
-9.93118167e-01 3.05468798e-01 1.34733781e-01 -2.68987745e-01
-4.20336694e-01 6.18748665e-02 -3.84404182e-01 4.39867139e-01
4.60938096e-01 2.27524340e-01 6.55043542e-01 7.39081681e-01
-1.73792884e-01 4.08438176e-01 -1.08884060e+00 1.02542067e+00
1.99088812e-01 -1.69600403e+00 3.38853180e-01 -4.71341889e-03
8.91572833e-01 -8.70726705e-02 -5.02434820e-02 1.34431109e-01
6.74858570e-01 -9.78392005e-01 1.10574126e+00 3.18894714e-01
3.59956354e-01 -7.38699615e-01 3.49502683e-01 -2.24873777e-02
-1.04864943e+00 -1.56269521e-02 -1.03409588e-01 -3.26206565e-01
1.11941926e-01 4.52658534e-01 -5.39863765e-01 3.68457139e-01
6.36066139e-01 7.00101316e-01 -6.79246008e-01 3.02034259e-01
-6.53352022e-01 9.25694287e-01 -1.78377956e-01 1.42239347e-01
2.41178572e-01 -4.63949800e-01 2.90381044e-01 9.05653358e-01
3.42289060e-01 -1.63689423e-02 -4.57134247e-02 1.10638154e+00
-2.00466156e-01 3.93160701e-01 -7.79387355e-01 -3.67370188e-01
3.49828929e-01 1.23913240e+00 -1.12153602e+00 -4.65881079e-02
-3.02293688e-01 8.56420457e-01 3.40692610e-01 1.00378752e-01
-8.35680306e-01 3.69611941e-02 4.79513615e-01 8.15715730e-01
2.33637139e-01 -4.12198871e-01 -7.21789420e-01 -8.20301116e-01
-1.24606177e-01 -7.53106236e-01 5.76793730e-01 -5.53389668e-01
-1.43100965e+00 3.66012931e-01 -2.76059270e-01 -9.46050763e-01
6.92000464e-02 -5.16125321e-01 -6.66439056e-01 1.52161866e-01
-5.35852969e-01 -1.96359253e+00 -7.70210400e-02 2.79438138e-01
8.30751836e-01 -4.32395220e-01 5.91468453e-01 3.93004477e-01
-2.21679375e-01 2.73885757e-01 2.60558784e-01 5.06901085e-01
4.57775176e-01 -1.19602323e+00 3.92027885e-01 3.52908671e-01
9.64766979e-01 4.05641258e-01 9.35968399e-01 -5.54581046e-01
-1.17732549e+00 -5.14961421e-01 7.72240996e-01 -1.00716829e+00
7.30441868e-01 -6.39643073e-01 -1.00665116e+00 8.93534720e-01
3.85686934e-01 -6.29400373e-01 6.10234857e-01 6.78731859e-01
-6.91953123e-01 -3.19820717e-02 -9.76812065e-01 4.37864363e-01
1.33521330e+00 -6.17391586e-01 -1.04445946e+00 3.94374818e-01
1.62878305e-01 6.68025017e-02 -9.91460621e-01 1.00652441e-01
1.06431150e+00 -8.63105714e-01 9.52232480e-01 -4.33079869e-01
9.41214085e-01 -1.53649092e-01 -2.63382614e-01 -1.04433262e+00
-4.91961390e-01 -4.46775466e-01 -5.71280569e-02 1.39773178e+00
1.71409205e-01 -1.01110689e-01 1.05649793e+00 2.45106354e-01
-1.62723392e-01 -4.59482670e-01 -9.00458932e-01 -7.52223492e-01
4.45444971e-01 -3.48871380e-01 7.06649303e-01 1.53358722e+00
-3.29353154e-01 6.89054906e-01 -5.15863419e-01 -3.65137607e-01
5.61918020e-01 3.10715437e-01 8.55172575e-01 -1.80420041e+00
-2.92623639e-01 -9.12383080e-01 -5.26219368e-01 -3.97079438e-01
-4.03335644e-03 -1.20185292e+00 -2.73394436e-01 -1.83536887e+00
9.62634757e-02 -2.44827241e-01 -1.29579425e-01 3.99669290e-01
3.81875634e-01 5.92227995e-01 2.43837446e-01 4.87728477e-01
-1.87803805e-01 4.90196615e-01 1.06570756e+00 -5.84654868e-01
-3.85458142e-01 -2.97103465e-01 -4.07953292e-01 1.18916082e+00
6.80639803e-01 -7.29532242e-01 -1.90179676e-01 -6.92216277e-01
7.80647278e-01 -5.18183768e-01 3.17576826e-01 -1.19546235e+00
-2.31467500e-01 -3.66913676e-01 4.51885015e-01 -7.77937412e-01
2.49505445e-01 -7.61269689e-01 7.07739592e-01 4.12249476e-01
-2.41297007e-01 -3.96398574e-01 -3.09566528e-01 1.30896345e-01
4.01907787e-03 -1.29404038e-01 9.57156003e-01 -2.48857975e-01
-5.44040740e-01 -7.29971603e-02 -3.76572818e-01 1.70955837e-01
6.84570909e-01 -3.30033302e-01 -1.44476637e-01 -4.35553417e-02
-8.21684659e-01 -3.46723348e-01 6.87630415e-01 2.38266602e-01
2.63128072e-01 -1.70183039e+00 -8.67424011e-01 -4.99173999e-02
2.62840930e-02 -5.47895171e-02 3.79017919e-01 5.03725350e-01
-1.07763290e+00 -1.98854744e-01 -8.02643657e-01 -5.28712332e-01
-7.19723582e-01 6.03090584e-01 -6.71223253e-02 -1.30590051e-01
-1.17630553e+00 5.85190833e-01 2.81818002e-01 -4.22204345e-01
-1.68633699e-01 3.86467800e-02 -4.74588335e-01 4.42417115e-01
3.73291969e-01 5.13231218e-01 -8.77325907e-02 -7.55433559e-01
-2.53873557e-01 9.93274033e-01 2.31597293e-02 -1.79462627e-01
1.55040514e+00 5.84971122e-02 -3.66022550e-02 9.94670451e-01
7.99350739e-01 4.40595597e-01 -7.61028588e-01 -1.26827031e-01
1.56451240e-01 -5.85929394e-01 -1.09299809e-01 -7.94332504e-01
-1.07401073e+00 9.57895637e-01 5.39018035e-01 4.67122346e-01
8.10042024e-01 3.80604684e-01 9.58821118e-01 5.31301856e-01
3.91283661e-01 -1.50238681e+00 5.85415401e-02 4.26903129e-01
1.01766109e+00 -8.53413999e-01 2.34383762e-01 -4.96092029e-02
-6.31806314e-01 1.14444578e+00 1.21626385e-01 -4.18531030e-01
5.67304552e-01 -2.00629961e-02 1.12661913e-01 -7.43711412e-01
-1.29337713e-01 -3.19042385e-01 3.06094140e-01 5.16992092e-01
5.38521051e-01 1.78397387e-01 -6.16451442e-01 6.20090783e-01
-8.51186752e-01 -2.66318172e-01 5.31030297e-01 7.84453034e-01
-5.02693653e-01 -1.01113379e+00 -3.91323209e-01 1.80750906e-01
-5.10740876e-01 2.40794435e-01 -9.92587626e-01 1.12929034e+00
7.56708145e-01 5.59247077e-01 2.49099731e-01 -5.32252729e-01
2.87904710e-01 2.53296513e-02 8.37859392e-01 -3.78284842e-01
-7.83441782e-01 1.05513267e-01 2.14529127e-01 2.77725477e-02
-4.53449011e-01 -4.95697141e-01 -1.13213027e+00 -7.13082314e-01
3.08239520e-01 1.90542154e-02 6.12919509e-01 1.10072017e+00
-2.89182872e-01 1.72137290e-01 5.24688363e-01 -6.47661209e-01
9.30258930e-02 -9.58816886e-01 -9.72224534e-01 7.75047541e-01
-1.49646670e-01 -7.97076404e-01 -1.80452317e-01 3.84609640e-01] | [11.33073902130127, 0.23037120699882507] |
1376a609-1d93-493e-a421-5329a2a627e4 | unsupervised-entity-alignment-for-temporal | 2302.00796 | null | https://arxiv.org/abs/2302.00796v2 | https://arxiv.org/pdf/2302.00796v2.pdf | Unsupervised Entity Alignment for Temporal Knowledge Graphs | Entity alignment (EA) is a fundamental data integration task that identifies equivalent entities between different knowledge graphs (KGs). Temporal Knowledge graphs (TKGs) extend traditional knowledge graphs by introducing timestamps, which have received increasing attention. State-of-the-art time-aware EA studies have suggested that the temporal information of TKGs facilitates the performance of EA. However, existing studies have not thoroughly exploited the advantages of temporal information in TKGs. Also, they perform EA by pre-aligning entity pairs, which can be labor-intensive and thus inefficient. In this paper, we present DualMatch which effectively fuses the relational and temporal information for EA. DualMatch transfers EA on TKGs into a weighted graph matching problem. More specifically, DualMatch is equipped with an unsupervised method, which achieves EA without necessitating seed alignment. DualMatch has two steps: (i) encoding temporal and relational information into embeddings separately using a novel label-free encoder, Dual-Encoder; and (ii) fusing both information and transforming it into alignment using a novel graph-matching-based decoder, GM-Decoder. DualMatch is able to perform EA on TKGs with or without supervision, due to its capability of effectively capturing temporal information. Extensive experiments on three real-world TKG datasets offer the insight that DualMatch outperforms the state-of-the-art methods in terms of H@1 by 2.4% - 10.7% and MRR by 1.7% - 7.6%, respectively. | ['Yunjun Gao', 'Lu Chen', 'Tianyi Li', 'Junyang Wu', 'Xiaoze Liu'] | 2023-02-01 | null | null | null | null | ['graph-matching', 'entity-alignment', 'data-integration', 'entity-alignment'] | ['graphs', 'knowledge-base', 'knowledge-base', 'natural-language-processing'] | [-5.51246963e-02 1.20223589e-01 -4.52552170e-01 -1.54670939e-01
-5.35103202e-01 -4.08033907e-01 5.09869814e-01 6.83207512e-01
-5.75151861e-01 3.63377035e-01 1.00002475e-01 -1.87098771e-01
-9.88941714e-02 -1.09963644e+00 -7.81124890e-01 -4.10911053e-01
-3.19895834e-01 2.67107189e-01 4.87401366e-01 -1.75623029e-01
-5.93567919e-03 1.15310594e-01 -1.23290813e+00 -1.33745283e-01
1.13413262e+00 9.56417918e-01 1.75707996e-01 2.46894076e-01
-2.20652610e-01 8.41317952e-01 -2.15352431e-01 -7.19521642e-01
2.19452024e-01 -2.80218840e-01 -8.00354183e-01 -3.18095922e-01
2.74948664e-02 -1.37301788e-01 -8.23489308e-01 9.65118289e-01
3.75486851e-01 1.80932030e-01 1.08878240e-01 -1.56423187e+00
-9.07556236e-01 7.69639790e-01 -5.66978514e-01 2.38416627e-01
2.46673629e-01 -8.00281614e-02 1.25636911e+00 -7.36170769e-01
6.30141556e-01 8.45030367e-01 6.45495653e-01 1.47160098e-01
-1.06733608e+00 -7.52166033e-01 2.08994001e-01 5.61177075e-01
-1.74901509e+00 -2.91092217e-01 7.35894740e-01 -3.69891554e-01
1.32029951e+00 -3.35361138e-02 6.74556136e-01 6.09388828e-01
2.16700092e-01 7.03680634e-01 7.13863909e-01 -4.01996732e-01
4.15474065e-02 -2.63696760e-01 2.33095780e-01 8.25311840e-01
3.01572472e-01 -6.01383150e-02 -6.89259946e-01 7.95101747e-03
5.45294404e-01 -2.28913948e-02 -2.87313014e-01 -3.06487560e-01
-1.39812112e+00 5.60174167e-01 7.46439219e-01 2.95764178e-01
-4.43652451e-01 3.54011148e-01 5.62931836e-01 2.46715724e-01
4.68460619e-01 1.59786865e-01 -2.89040715e-01 -2.47438066e-02
-6.84842408e-01 -5.16760312e-02 5.31285763e-01 1.31683421e+00
8.35292459e-01 -2.17168242e-01 -1.60167977e-01 5.26196241e-01
3.67601275e-01 4.28553075e-01 5.58014393e-01 -4.39044237e-01
7.14314401e-01 9.35609698e-01 -1.42912880e-01 -1.34477532e+00
-4.54645038e-01 -4.29284662e-01 -7.37700224e-01 -5.81909597e-01
2.22058609e-01 1.76302567e-01 -8.89825821e-01 2.00138998e+00
4.39871877e-01 7.03992605e-01 2.33542934e-01 7.57848263e-01
8.02913666e-01 7.01999784e-01 5.62122986e-02 -2.98653126e-01
1.43242121e+00 -1.19356287e+00 -9.01622415e-01 -2.19400495e-01
9.09210205e-01 -5.35970926e-01 7.62686908e-01 -5.75191900e-02
-8.71700525e-01 -4.32243526e-01 -1.31302834e+00 -2.33677760e-01
-5.57044983e-01 3.11993845e-02 5.85954249e-01 3.03060442e-01
-9.14335310e-01 5.89062750e-01 -1.07961833e+00 -3.68273139e-01
1.20404810e-01 4.39262331e-01 -5.27971864e-01 -1.40501841e-04
-1.65906930e+00 8.36622536e-01 8.44127297e-01 1.88201502e-01
-3.53426456e-01 -7.67478168e-01 -1.22467935e+00 1.08161397e-01
7.42769539e-01 -7.49249697e-01 1.07563508e+00 -4.98456150e-01
-1.29500997e+00 5.71656525e-01 -3.10345620e-01 -4.93850619e-01
2.25109205e-01 -1.73603788e-01 -1.15214288e+00 1.13460377e-01
2.74808854e-01 3.05696309e-01 4.10946608e-01 -8.63556385e-01
-6.63048089e-01 -3.32528234e-01 9.70056131e-02 1.47411257e-01
-4.74173427e-01 -2.64197767e-01 -1.12368095e+00 -7.55328536e-01
1.23569861e-01 -9.69834805e-01 -1.29342675e-01 -3.62498403e-01
-3.60468537e-01 -3.35717469e-01 6.18254125e-01 -8.81208777e-01
1.83510864e+00 -2.14203596e+00 2.39375636e-01 2.73715973e-01
4.38709944e-01 3.81327063e-01 -5.24896905e-02 7.61416912e-01
-3.77590396e-02 1.43873915e-02 -1.97725505e-01 -3.03520441e-01
1.54656291e-01 2.34558105e-01 -1.56134725e-01 4.07984644e-01
9.28652659e-02 1.17411506e+00 -1.27535713e+00 -5.37553430e-01
7.99405873e-02 3.12484264e-01 -2.85331398e-01 9.62285921e-02
-2.01958463e-01 -6.63709501e-03 -2.88152993e-01 5.21846235e-01
6.25564933e-01 -5.42784631e-01 6.96197331e-01 -5.59885800e-01
-9.97392461e-02 4.21105236e-01 -1.13758433e+00 1.97253573e+00
-3.52268636e-01 4.38610792e-01 -5.67122877e-01 -8.59388530e-01
8.32293987e-01 2.91106820e-01 5.09803176e-01 -1.06206405e+00
6.97222492e-03 3.94853920e-01 -2.55373180e-01 -2.52153426e-01
7.95347929e-01 7.71408454e-02 -3.71472687e-01 3.74021709e-01
1.58568963e-01 6.69562697e-01 2.98548281e-01 5.60343444e-01
1.32202005e+00 1.51778953e-02 4.26728576e-01 8.49488098e-03
5.64688504e-01 -1.15422703e-01 1.05821311e+00 2.15063959e-01
-5.90285771e-02 1.57049119e-01 2.49969140e-01 -1.25558153e-01
-7.20846713e-01 -1.04591405e+00 3.66836935e-01 6.54488087e-01
6.80767059e-01 -1.01853096e+00 -4.43137646e-01 -7.49484777e-01
2.78953582e-01 6.35080278e-01 -5.10458052e-01 -4.89644766e-01
-7.20654845e-01 -5.28489411e-01 6.22813702e-01 8.23568821e-01
6.64381325e-01 -6.20719314e-01 -2.69242108e-01 4.96975809e-01
-4.56009209e-01 -1.37044668e+00 -8.05211365e-01 7.96780288e-02
-7.71670938e-01 -1.09367883e+00 -3.66030753e-01 -6.59972250e-01
7.19286919e-01 5.32658577e-01 8.29181194e-01 8.85453299e-02
5.93582094e-02 2.08584145e-01 -5.79665959e-01 -5.53323701e-03
-4.93918844e-02 3.67283970e-01 1.17946513e-01 1.22239172e-01
5.11311948e-01 -5.84745169e-01 -5.42069554e-01 4.30811912e-01
-9.38308477e-01 2.28626534e-01 6.62868261e-01 6.52763009e-01
6.65303648e-01 1.67295933e-01 7.27346480e-01 -9.77516472e-01
3.46133888e-01 -6.14849865e-01 -7.20982134e-01 6.75022840e-01
-1.01391959e+00 3.25528830e-01 6.04549885e-01 -1.98043644e-01
-9.32999611e-01 -2.13714674e-01 2.14587778e-01 -6.13685012e-01
5.35457075e-01 9.79684770e-01 -2.39090309e-01 7.64500126e-02
2.61696607e-01 2.19556376e-01 -1.98992386e-01 -3.73959661e-01
6.27299905e-01 5.77511549e-01 8.36982310e-01 -5.69538713e-01
8.67462635e-01 2.92353004e-01 -6.94981515e-02 -2.48974428e-01
-5.59098244e-01 -6.93912089e-01 -5.19125640e-01 -1.16211183e-01
7.66056716e-01 -1.04851019e+00 -6.64209485e-01 5.10134280e-01
-9.83391345e-01 -1.99599028e-01 -6.48016259e-02 5.65526664e-01
-2.58624613e-01 6.44945383e-01 -7.21455097e-01 -3.93385798e-01
-4.70112532e-01 -9.15730774e-01 7.74161458e-01 3.32948387e-01
-3.26715000e-02 -9.82680619e-01 -1.54098263e-02 3.85548711e-01
2.33280361e-01 2.44253293e-01 8.56682539e-01 -6.66523397e-01
-7.55595148e-01 -1.67831004e-01 -3.66226554e-01 -4.54382412e-02
3.65795940e-01 -1.67380512e-01 -6.69404149e-01 -4.32544202e-01
-5.47194839e-01 9.08202305e-02 7.48653352e-01 -1.29847601e-01
7.36191511e-01 -1.65226758e-01 -6.69615507e-01 5.55044055e-01
1.57844925e+00 4.98874307e-01 5.50750852e-01 5.05559921e-01
9.46126699e-01 3.30018789e-01 9.28404152e-01 2.86255687e-01
1.03055775e+00 1.01315451e+00 3.19575578e-01 1.41123965e-01
-9.02741104e-02 -6.09195173e-01 3.88639838e-01 1.46457279e+00
-1.13603678e-02 -4.47489411e-01 -9.62747812e-01 8.68522286e-01
-2.44877243e+00 -7.44603932e-01 -1.97469190e-01 2.14619207e+00
9.32709157e-01 3.84259075e-02 -3.84805654e-03 1.49468631e-01
8.67466629e-01 2.34124705e-01 -4.86958385e-01 -5.93463099e-03
4.96723875e-02 2.97961347e-02 9.33032990e-01 2.24063367e-01
-1.04249573e+00 9.37959433e-01 4.51144838e+00 7.68078566e-01
-1.02242553e+00 1.62446499e-01 -1.43644184e-01 1.12331055e-01
-4.39453185e-01 2.56332338e-01 -6.08691573e-01 6.68722391e-01
1.03708076e+00 -6.99065804e-01 3.62672031e-01 5.63980520e-01
-2.64930785e-01 1.15966454e-01 -1.10927677e+00 1.02925658e+00
-2.83481441e-02 -1.35455966e+00 -1.72129467e-01 -6.92602172e-02
5.26431501e-01 -2.81942543e-02 -3.43706161e-01 4.27305013e-01
3.69938225e-01 -6.81908846e-01 7.56575942e-01 5.92839301e-01
7.40009069e-01 -8.66081536e-01 9.27805305e-01 1.62388347e-02
-1.87737143e+00 2.08628803e-01 -1.02090240e-01 3.25251371e-01
5.02746463e-01 8.30720365e-01 -7.13130891e-01 1.60822010e+00
6.41256094e-01 1.05261230e+00 -5.15256941e-01 9.80375111e-01
-4.08842355e-01 3.26939404e-01 -2.37238005e-01 3.15545350e-01
1.74533874e-01 -2.45191798e-01 3.55899215e-01 1.12535226e+00
4.43150699e-01 1.31293342e-01 1.61200270e-01 5.75843513e-01
-3.17648113e-01 1.94823015e-02 -4.58306223e-01 -3.92769635e-01
9.65083838e-01 1.04903710e+00 -6.20747924e-01 -3.03438067e-01
-5.32400012e-01 1.11828935e+00 4.21435714e-01 1.16495445e-01
-1.08278084e+00 -7.29161739e-01 5.30180097e-01 -1.86392531e-01
4.81135398e-01 -3.80640119e-01 4.54341285e-02 -1.33778906e+00
3.19246203e-01 -5.54871619e-01 8.35101604e-01 -5.55773973e-01
-1.09729695e+00 5.20787299e-01 1.36527903e-02 -1.23691142e+00
-2.34730300e-02 -1.04011238e-01 -2.34840140e-01 7.05081463e-01
-1.61763501e+00 -1.20314252e+00 -4.74297494e-01 5.04464447e-01
-4.74302284e-02 4.70711403e-02 6.42122805e-01 8.13130021e-01
-9.79926527e-01 8.32321525e-01 -5.49922548e-02 3.06005329e-01
8.45476985e-01 -1.28409290e+00 7.19078183e-01 1.15767884e+00
1.22795969e-01 7.06485748e-01 2.90691972e-01 -8.03202987e-01
-1.71734071e+00 -1.35961986e+00 1.33002579e+00 -1.03704065e-01
8.37713242e-01 -2.62757242e-01 -1.10090399e+00 8.95366788e-01
1.26189366e-01 2.53271908e-01 6.83104336e-01 8.90566111e-02
-6.01308763e-01 -3.85033876e-01 -9.29885924e-01 6.15693986e-01
1.39551556e+00 -7.77573645e-01 -6.21238232e-01 -8.38965252e-02
1.12261152e+00 -6.09054863e-01 -1.39626992e+00 5.69221437e-01
3.81521612e-01 -6.17022991e-01 8.87021363e-01 -3.40515733e-01
1.14322066e-01 -7.66156614e-01 -2.11416006e-01 -1.32057345e+00
-3.76723379e-01 -6.12187862e-01 -5.85897028e-01 1.64255357e+00
1.89181432e-01 -9.49038208e-01 4.64740038e-01 3.99819285e-01
-3.06042850e-01 -6.77357197e-01 -8.10018241e-01 -1.21321595e+00
-4.66014296e-01 -4.53474671e-01 9.33750510e-01 1.34946585e+00
2.89613038e-01 3.20889175e-01 -3.30355257e-01 5.04864037e-01
5.08733928e-01 3.36808234e-01 5.97595036e-01 -1.00030148e+00
-9.28389952e-02 -3.48384142e-01 -6.38381124e-01 -9.75982189e-01
2.02500850e-01 -1.24572027e+00 -1.14357240e-01 -1.73950565e+00
-5.10580800e-02 -4.73263055e-01 -5.48513532e-01 7.77686000e-01
-2.96049446e-01 -1.80408180e-01 1.33431375e-01 2.81910360e-01
-7.54480660e-01 7.94484138e-01 8.20542097e-01 -1.59852698e-01
-1.26170784e-01 -5.91608405e-01 -5.79136252e-01 3.52226079e-01
6.96122229e-01 -5.24400711e-01 -7.41942286e-01 -5.19109190e-01
3.42881233e-01 1.07376434e-01 1.61365107e-01 -8.57808411e-01
8.20682347e-01 -5.63011654e-02 -2.11151958e-01 -6.02903426e-01
5.94968535e-02 -7.91673839e-01 6.12513065e-01 3.23596209e-01
1.21431299e-01 3.30473125e-01 3.10946375e-01 9.83446062e-01
-5.36380351e-01 4.79152165e-02 3.07009459e-01 3.42919081e-01
-1.26020217e+00 4.13863897e-01 1.29821196e-01 1.38570800e-01
1.23704362e+00 -5.07364646e-02 -5.39012551e-01 -4.11795750e-02
-4.64420199e-01 6.55226529e-01 4.47552443e-01 5.37221074e-01
4.71908510e-01 -1.66390800e+00 -2.63715029e-01 -8.52865539e-03
4.70947713e-01 9.55917984e-02 3.33582431e-01 1.13532615e+00
-3.59664381e-01 4.53078866e-01 4.19232026e-02 -3.54429394e-01
-1.15907156e+00 6.98026717e-01 4.22925726e-02 -5.60930967e-01
-7.31763482e-01 5.47324002e-01 -1.04016326e-01 -3.15463543e-01
8.04389045e-02 -3.70208442e-01 1.86003987e-02 2.30037659e-01
2.13242620e-01 4.31578755e-01 3.27409953e-01 -5.67468286e-01
-7.72174060e-01 5.53643167e-01 -1.94336355e-01 1.97039083e-01
1.22238636e+00 -1.31777391e-01 -2.08603397e-01 2.54751652e-01
1.22128558e+00 -2.58851349e-02 -8.03756177e-01 -7.93686152e-01
4.71416175e-01 -4.07404959e-01 4.43941057e-02 -4.98946935e-01
-1.29879832e+00 4.72644508e-01 1.47189915e-01 9.10604522e-02
1.37559509e+00 -1.97827563e-01 1.36531925e+00 1.09583817e-01
6.46389365e-01 -9.68840778e-01 -3.95240448e-02 4.37646896e-01
4.03692305e-01 -1.03519404e+00 -9.72899329e-03 -7.08094835e-01
-5.75523615e-01 1.02162993e+00 6.34745777e-01 2.32308596e-01
5.72858989e-01 4.32512090e-02 -2.36454964e-01 -3.76708448e-01
-7.99996257e-01 -3.75241041e-01 5.66848814e-01 2.41584659e-01
1.37056217e-01 1.34591997e-01 -5.20326793e-01 9.76715982e-01
2.36816797e-02 2.36701101e-01 2.69850761e-01 1.10544789e+00
5.46595939e-02 -1.36052251e+00 2.11596906e-01 2.69959539e-01
-1.71837687e-01 -4.76945229e-02 -1.18511215e-01 9.30376589e-01
7.35266432e-02 8.96751523e-01 -4.37425561e-02 -9.25812840e-01
5.95132232e-01 -1.42757101e-02 2.94523209e-01 -4.42906946e-01
-5.10262966e-01 -1.84786350e-01 3.49778056e-01 -6.00012839e-01
-5.16541958e-01 -5.90698361e-01 -1.65888810e+00 -5.54324329e-01
-5.17287910e-01 3.40022981e-01 2.39844516e-01 8.53242755e-01
8.48977685e-01 7.29189456e-01 5.45495808e-01 -2.05509216e-01
-1.17499046e-01 -5.35137475e-01 -4.89992768e-01 3.44033748e-01
5.38729690e-02 -8.24092984e-01 -5.96917830e-02 -1.22732699e-01] | [8.664586067199707, 7.923033714294434] |
61e31f59-962a-4706-a22e-2ae14a8f5624 | blip-2-bootstrapping-language-image-pre | 2301.12597 | null | https://arxiv.org/abs/2301.12597v3 | https://arxiv.org/pdf/2301.12597v3.pdf | BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models | The cost of vision-and-language pre-training has become increasingly prohibitive due to end-to-end training of large-scale models. This paper proposes BLIP-2, a generic and efficient pre-training strategy that bootstraps vision-language pre-training from off-the-shelf frozen pre-trained image encoders and frozen large language models. BLIP-2 bridges the modality gap with a lightweight Querying Transformer, which is pre-trained in two stages. The first stage bootstraps vision-language representation learning from a frozen image encoder. The second stage bootstraps vision-to-language generative learning from a frozen language model. BLIP-2 achieves state-of-the-art performance on various vision-language tasks, despite having significantly fewer trainable parameters than existing methods. For example, our model outperforms Flamingo80B by 8.7% on zero-shot VQAv2 with 54x fewer trainable parameters. We also demonstrate the model's emerging capabilities of zero-shot image-to-text generation that can follow natural language instructions. | ['Steven Hoi', 'Silvio Savarese', 'Dongxu Li', 'Junnan Li'] | 2023-01-30 | null | null | null | null | ['generative-visual-question-answering', 'open-vocabulary-attribute-detection', 'visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'computer-vision', 'computer-vision', 'reasoning'] | [ 2.25240722e-01 2.72028059e-01 -5.05129918e-02 -4.55396265e-01
-1.45089018e+00 -4.67413425e-01 8.92097294e-01 -4.64128554e-01
-5.45865119e-01 5.54447353e-01 2.46159866e-01 -5.42322099e-01
9.09609437e-01 -5.32495975e-01 -1.12952399e+00 -3.67271423e-01
6.24310851e-01 7.00616658e-01 1.91583723e-01 -1.85218364e-01
-1.51234016e-01 -1.95988074e-01 -1.54250658e+00 6.25896335e-01
5.69754779e-01 9.98518169e-01 7.73562431e-01 1.12214386e+00
-2.22042739e-01 1.31352913e+00 -3.02035689e-01 -5.28921664e-01
2.34809458e-01 -4.59579170e-01 -7.35565484e-01 2.71382689e-01
9.55988348e-01 -8.13952267e-01 -4.36496556e-01 7.09514439e-01
4.99000043e-01 -1.97059840e-01 3.44535202e-01 -1.39218080e+00
-1.34932041e+00 3.17328423e-01 -4.08874303e-01 -1.43836513e-01
-3.97009263e-03 8.77149701e-01 8.42075944e-01 -1.27151346e+00
8.14300478e-01 1.42731190e+00 4.98263597e-01 1.08414388e+00
-1.17146289e+00 -5.53628922e-01 -1.55816019e-01 -6.02408051e-02
-1.25458908e+00 -9.87262070e-01 1.99076697e-01 -6.13970280e-01
1.36234021e+00 -4.30612028e-01 4.97693628e-01 1.35102308e+00
1.00474373e-01 7.58371174e-01 9.61421430e-01 -5.95210493e-01
8.65938663e-02 2.13775545e-01 -7.58112520e-02 1.15540111e+00
1.07551105e-01 6.68601245e-02 -7.46335506e-01 1.40990779e-01
7.75264621e-01 -2.55794585e-01 -1.23839624e-01 -3.48653048e-01
-1.03505504e+00 1.04051650e+00 3.13184112e-01 -1.83034539e-01
-3.11178863e-01 7.23836958e-01 3.61177981e-01 2.15163365e-01
3.11953068e-01 1.08237661e-01 -2.16181621e-01 -1.61835074e-01
-1.15072310e+00 8.18216875e-02 3.56688529e-01 1.39845538e+00
9.18628156e-01 3.46591145e-01 -5.66470504e-01 7.18401372e-01
5.56953728e-01 8.96012545e-01 5.23221672e-01 -1.24892938e+00
5.30050099e-01 1.52753398e-01 1.08665965e-01 3.69831286e-02
3.51401091e-01 4.04900573e-02 -6.11752629e-01 1.38902679e-01
-1.51118413e-02 -1.75760731e-01 -1.70126176e+00 1.58547699e+00
-7.63924867e-02 8.14487338e-02 4.99561578e-01 6.50989175e-01
1.13389480e+00 1.09490514e+00 1.77066058e-01 5.66602498e-02
1.43584526e+00 -1.58295953e+00 -3.31665188e-01 -6.40703678e-01
2.91799366e-01 -6.09867394e-01 1.55837297e+00 7.56904259e-02
-1.45213056e+00 -8.94203782e-01 -7.28571713e-01 -6.72999740e-01
-1.54488474e-01 2.38876998e-01 3.75480115e-01 4.70040053e-01
-1.78691220e+00 -2.30644435e-01 -7.33687162e-01 -5.13506830e-01
5.85247159e-01 5.47724357e-03 -3.05387020e-01 -5.37510574e-01
-9.24792945e-01 7.83894837e-01 3.84025544e-01 -1.87766328e-01
-1.74775302e+00 -8.71272981e-01 -1.16781628e+00 -4.57084514e-02
3.67752612e-01 -1.28265131e+00 1.77475548e+00 -9.95719433e-01
-1.69776618e+00 1.14987075e+00 -5.22189617e-01 -8.92515302e-01
5.60783803e-01 -1.25546649e-01 -5.76796196e-02 4.52156782e-01
4.11085516e-01 1.49589419e+00 1.37234163e+00 -1.24958599e+00
-4.94815797e-01 9.37519148e-02 1.08439259e-01 5.43468408e-02
-1.62661314e-01 5.65971769e-02 -1.07002234e+00 -1.33475944e-01
-6.56914592e-01 -8.39036703e-01 -1.48860499e-01 2.21346334e-01
-3.70767713e-02 -3.27199101e-01 7.16110945e-01 -6.18223369e-01
5.21654844e-01 -1.89727092e+00 -2.63872504e-01 -5.65259814e-01
2.03834742e-01 4.49659318e-01 -6.76560819e-01 4.13715959e-01
3.82932633e-01 -4.36303802e-02 -4.27467339e-02 -9.23991203e-01
-1.30885974e-01 3.49049807e-01 -6.50528252e-01 1.01855798e-02
6.25539720e-01 1.46184051e+00 -9.08537567e-01 -7.54330039e-01
4.10017073e-01 6.51835859e-01 -5.19144297e-01 5.24258912e-01
-6.79920495e-01 7.65763968e-02 1.60484703e-03 5.79152703e-01
4.70489144e-01 -7.12233782e-01 -6.04768544e-02 -1.23349346e-01
-1.14181690e-01 -3.42077315e-02 -2.87104696e-01 2.23066425e+00
-6.74395084e-01 8.92852068e-01 1.18883848e-02 -7.30264723e-01
6.06650829e-01 3.53804618e-01 1.03832789e-01 -8.72108102e-01
9.43836011e-03 -5.50097078e-02 -6.25851393e-01 -5.95124006e-01
7.88008273e-01 -1.56701133e-01 -1.14817351e-01 1.60460755e-01
7.74974763e-01 -5.35656929e-01 2.52189666e-01 7.06226110e-01
9.61732686e-01 4.04685050e-01 5.42387143e-02 1.21915750e-01
3.28255117e-01 1.71215847e-01 -2.14723852e-02 1.04665291e+00
-1.34110942e-01 8.58622789e-01 -1.96384583e-02 -1.61469325e-01
-1.27142417e+00 -1.35927498e+00 2.77621627e-01 1.23679209e+00
-2.68720776e-01 -5.15547633e-01 -7.00957537e-01 -4.48865533e-01
-1.55501530e-01 1.08264232e+00 -5.52814305e-01 -1.97930247e-01
1.34300571e-02 -2.63352413e-02 8.38635266e-01 5.10040343e-01
5.85266411e-01 -1.07845056e+00 -8.63312244e-01 2.23069619e-02
-4.22132581e-01 -1.68437445e+00 -5.77469349e-01 -9.66485441e-02
-5.76829970e-01 -6.16334379e-01 -1.06734288e+00 -8.85714531e-01
6.59157693e-01 8.12735677e-01 1.63255966e+00 -1.27399459e-01
-3.22181553e-01 9.38529909e-01 -3.68764281e-01 -5.65916896e-01
-9.18388188e-01 -1.93356901e-01 -3.90911967e-01 -1.58723354e-01
3.21261615e-01 4.44264039e-02 -5.08451879e-01 -2.54379034e-01
-8.56266558e-01 5.77146769e-01 7.14734077e-01 1.00024331e+00
6.15591586e-01 -7.36705363e-01 3.07658702e-01 -3.22578013e-01
4.85493094e-01 -2.57113606e-01 -8.08046758e-01 6.18153453e-01
-5.81211746e-01 2.63461560e-01 4.52867836e-01 -2.59167314e-01
-1.23404503e+00 2.79824346e-01 2.06946749e-02 -1.07568383e+00
-1.37895634e-02 1.45862684e-01 1.56536132e-01 1.57941595e-01
6.60346270e-01 7.20035613e-01 1.29479438e-01 -7.31770694e-02
1.15628266e+00 8.12530875e-01 9.93079543e-01 -4.63232607e-01
7.45956302e-01 5.12338936e-01 -4.89526361e-01 -1.06336069e+00
-9.18142915e-01 -5.58787227e-01 -2.87980318e-01 -1.12358220e-01
1.22010124e+00 -1.70433259e+00 -4.27021921e-01 5.24101913e-01
-1.50030577e+00 -7.19380140e-01 -5.25873542e-01 2.97713093e-02
-7.67584383e-01 2.65600681e-01 -5.50311923e-01 -7.94514775e-01
-1.03698933e+00 -1.11159956e+00 1.54057658e+00 2.33561143e-01
2.08018497e-01 -6.20190203e-01 2.14123249e-01 8.29743922e-01
3.63738656e-01 -3.55792016e-01 3.90881062e-01 -2.05390528e-03
-7.37067282e-01 1.04413986e-01 -5.10428786e-01 7.16885686e-01
-3.10101658e-01 -4.81714308e-02 -1.19621551e+00 -4.99910980e-01
-3.33225399e-01 -1.28708577e+00 1.10476971e+00 3.37850928e-01
7.76081383e-01 -1.09582782e-01 1.07426224e-02 6.69929564e-01
1.66746056e+00 -9.97421443e-02 7.80587912e-01 4.89475504e-02
8.28751564e-01 2.36659244e-01 3.56416404e-01 2.64024407e-01
7.78881311e-01 2.86016285e-01 4.03565198e-01 -5.25906049e-02
-4.86129612e-01 -6.54790401e-01 8.11388314e-01 6.24561131e-01
1.38609529e-01 -3.40375245e-01 -8.59767735e-01 8.51474583e-01
-1.71921217e+00 -1.07513535e+00 1.85963154e-01 1.92514694e+00
1.01227403e+00 -5.41387089e-02 -1.02994952e-03 -8.23327184e-01
4.12190408e-01 2.25614652e-01 -6.54430032e-01 -3.44036460e-01
-4.19925489e-02 2.96085149e-01 4.53540623e-01 6.58757746e-01
-7.06236839e-01 1.61393976e+00 6.28287411e+00 6.28918946e-01
-1.27978957e+00 5.53511620e-01 6.44851267e-01 -4.66457397e-01
-2.71334529e-01 1.81745321e-01 -9.67304647e-01 2.49019608e-01
1.34583712e+00 -1.78889230e-01 3.79568070e-01 1.04755688e+00
1.80394202e-01 -9.97352079e-02 -9.78039086e-01 1.31493974e+00
5.38293839e-01 -1.77655244e+00 5.38396835e-01 3.24621126e-02
7.71002710e-01 8.65829051e-01 -7.38774315e-02 6.25448346e-01
6.17712915e-01 -1.07810378e+00 1.12747848e+00 4.87684339e-01
1.49742341e+00 -3.13505024e-01 2.31073871e-01 3.61798823e-01
-9.46794689e-01 5.09326980e-02 -6.00233138e-01 2.25895852e-01
4.72035199e-01 2.38598928e-01 -1.13374650e+00 3.19539815e-01
6.92913353e-01 4.18291122e-01 -6.29082739e-01 3.95922810e-01
-2.08586931e-01 7.27945745e-01 -1.23828121e-01 1.74194932e-01
4.97583508e-01 4.10305299e-02 -5.73869981e-02 1.10805416e+00
3.62185240e-01 -9.74962562e-02 2.23528087e-01 1.06640005e+00
-2.98354477e-01 -3.03789854e-01 -8.00785244e-01 -3.68226945e-01
2.72580653e-01 1.22464776e+00 -2.55703241e-01 -9.13686812e-01
-8.33272994e-01 1.41602612e+00 2.93088347e-01 3.79254878e-01
-1.00042009e+00 -3.97560634e-02 3.60257953e-01 8.06321874e-02
6.79692447e-01 -2.81062663e-01 3.39782298e-01 -1.48086977e+00
-2.06885874e-01 -8.47736239e-01 -3.60648260e-02 -1.51289201e+00
-1.10921836e+00 7.83748448e-01 -3.95039208e-02 -8.26580405e-01
-8.12187195e-01 -5.20907164e-01 -3.72760862e-01 9.50570822e-01
-1.71476293e+00 -2.00471568e+00 -6.99027717e-01 1.00718784e+00
1.09030187e+00 -3.40257674e-01 8.86800408e-01 -1.00445911e-01
-1.16869286e-01 5.99801183e-01 -9.14043710e-02 1.63370103e-01
6.53587520e-01 -8.84468973e-01 1.02355027e+00 1.04732656e+00
4.45124656e-01 3.00725967e-01 4.64052767e-01 -5.52588284e-01
-1.55891526e+00 -1.23415542e+00 8.22505116e-01 -6.60853803e-01
7.16199577e-01 -6.39858365e-01 -4.29778308e-01 8.57131481e-01
8.54694963e-01 3.15129310e-01 3.51615876e-01 -4.95396286e-01
-7.18757272e-01 -1.24796130e-01 -7.43483722e-01 6.87338173e-01
9.35326219e-01 -9.52344239e-01 -6.63073838e-01 4.85739052e-01
1.16906595e+00 -1.64078340e-01 -3.95860612e-01 -8.94797593e-02
5.26078880e-01 -7.32851028e-01 1.01997197e+00 -6.46528482e-01
7.40788519e-01 -1.17774449e-01 -2.99551725e-01 -9.04417455e-01
-2.38590434e-01 -1.01552713e+00 -2.26953045e-01 1.13810623e+00
3.35181028e-01 -3.00944418e-01 6.38173819e-01 4.63177323e-01
-1.60139903e-01 -4.49299753e-01 -6.64895952e-01 -6.67059481e-01
1.65538594e-01 -5.34254849e-01 6.49623200e-02 4.92454290e-01
-6.30817056e-01 1.02896714e+00 -6.80182993e-01 -2.52049923e-01
7.45973587e-01 -1.10843249e-01 1.28683853e+00 -6.39198780e-01
-5.75767338e-01 1.22639567e-01 -1.38832137e-01 -1.33393681e+00
1.54758692e-01 -7.17948556e-01 5.23877561e-01 -1.85411084e+00
3.36398631e-01 2.15949640e-01 -4.86464612e-02 7.18952239e-01
-2.09583398e-02 4.11339402e-01 6.42524302e-01 2.14031786e-01
-8.82684529e-01 7.47119129e-01 1.17618704e+00 -3.85149121e-01
-2.64615496e-03 -5.92902184e-01 -7.84871578e-01 4.26767856e-01
5.18915772e-01 -1.98044643e-01 -1.08316255e+00 -8.92523944e-01
-7.22192079e-02 7.32406378e-02 7.51913786e-01 -1.07815301e+00
1.30758926e-01 -1.04937091e-01 9.58028361e-02 -6.49937689e-01
6.99188530e-01 -4.27930325e-01 -2.32984632e-01 3.32532585e-01
-2.87610501e-01 1.13122962e-01 2.22892568e-01 5.46799719e-01
5.12019657e-02 -1.09670296e-01 8.06996226e-01 -5.82351744e-01
-1.08876002e+00 2.42697537e-01 -4.44004118e-01 4.42185551e-01
9.62360978e-01 -4.77761291e-02 -6.36634707e-01 -5.98923326e-01
-2.98281014e-01 1.52914181e-01 5.99083662e-01 6.15423679e-01
9.07007575e-01 -9.73090827e-01 -8.94048214e-01 2.28280216e-01
4.82092083e-01 1.56326592e-01 3.72452468e-01 4.20673907e-01
-6.20539427e-01 6.27340257e-01 -1.02818780e-01 -9.36561465e-01
-1.29135847e+00 6.98277652e-01 2.09192976e-01 -2.35380962e-01
-7.48253882e-01 1.07938969e+00 3.59743834e-01 -4.84005222e-03
5.78414649e-02 -4.08211023e-01 4.74059373e-01 -2.71693677e-01
7.00559318e-01 -1.88107193e-01 -2.33345464e-01 -7.36660540e-01
-2.11626142e-01 3.91526550e-01 -2.93587267e-01 -7.09408879e-01
1.12747490e+00 -1.29740909e-01 1.32388666e-01 3.99892300e-01
1.10153937e+00 -3.99894089e-01 -1.68503106e+00 -2.47131586e-01
-5.68554580e-01 -1.73741966e-01 2.22293079e-01 -9.51007903e-01
-8.36071551e-01 1.17494404e+00 5.49025893e-01 -3.28430921e-01
1.00836599e+00 2.73300081e-01 1.07715297e+00 6.42513871e-01
5.55483520e-01 -1.07070279e+00 4.92096901e-01 6.85226738e-01
9.43425059e-01 -1.52227569e+00 -2.10413679e-01 1.01001002e-01
-1.04932463e+00 7.51448810e-01 7.71783650e-01 -1.94976795e-02
2.76641846e-01 1.60168722e-01 2.78085411e-01 -1.43354665e-02
-1.25772703e+00 -3.48412037e-01 1.50640130e-01 8.81347954e-01
1.30662322e-01 -1.44853458e-01 5.06729066e-01 3.14674228e-01
-6.21195287e-02 6.33637428e-01 5.00841379e-01 8.39023769e-01
-4.34823632e-01 -7.02017009e-01 -1.80460989e-01 1.84879750e-01
-6.35191053e-02 -6.42188251e-01 -2.16383278e-01 5.88720798e-01
7.24800527e-02 1.09063554e+00 2.31246099e-01 -2.13742211e-01
-7.75796324e-02 2.26482973e-01 7.22372234e-01 -8.04587841e-01
-2.91753143e-01 6.40407875e-02 8.75307024e-02 -5.85455120e-01
-4.18724477e-01 -3.20216268e-01 -1.08644307e+00 -1.29788741e-01
-2.11034976e-02 -1.98922098e-01 7.24095225e-01 8.28846812e-01
8.37999940e-01 3.54736328e-01 1.21648058e-01 -9.59273517e-01
-7.32664764e-01 -9.39600587e-01 9.93328914e-03 1.58782274e-01
4.13271576e-01 -2.36284837e-01 6.01372607e-02 7.64141381e-01] | [10.874812126159668, 1.5114136934280396] |
95a49a24-7f8f-4e44-8af2-139e39789bdd | lipo-lcd-combining-lines-and-points-for | 2009.09897 | null | https://arxiv.org/abs/2009.09897v1 | https://arxiv.org/pdf/2009.09897v1.pdf | LiPo-LCD: Combining Lines and Points for Appearance-based Loop Closure Detection | Visual SLAM approaches typically depend on loop closure detection to correct the inconsistencies that may arise during the map and camera trajectory calculations, typically making use of point features for detecting and closing the existing loops. In low-textured scenarios, however, it is difficult to find enough point features and, hence, the performance of these solutions drops drastically. An alternative for human-made scenarios, due to their structural regularity, is the use of geometrical cues such as straight segments, frequently present within these environments. Under this context, in this paper we introduce LiPo-LCD, a novel appearance-based loop closure detection method that integrates lines and points. Adopting the idea of incremental Bag-of-Binary-Words schemes, we build separate BoW models for each feature, and use them to retrieve previously seen images using a late fusion strategy. Additionally, a simple but effective mechanism, based on the concept of island, groups similar images close in time to reduce the image candidate search effort. A final step validates geometrically the loop candidates by incorporating the detected lines by means of a process comprising a line feature matching stage, followed by a robust spatial verification stage, now combining both lines and points. As it is reported in the paper, LiPo-LCD compares well with several state-of-the-art solutions for a number of datasets involving different environmental conditions. | ['Emilio Garcia-Fidalgo', 'Joan P. Company-Corcoles', 'Alberto Ortiz'] | 2020-09-03 | null | null | null | null | ['loop-closure-detection'] | ['computer-vision'] | [ 1.43841609e-01 -2.98549324e-01 2.35438108e-01 -2.07958013e-01
-5.56465149e-01 -5.97664952e-01 7.77446330e-01 9.46000457e-01
-5.42337060e-01 5.41520655e-01 -2.80185461e-01 -2.44697571e-01
-2.11749434e-01 -7.96107471e-01 -6.93823516e-01 -4.39718664e-01
-1.48664698e-01 5.57434201e-01 7.59822905e-01 -3.43600363e-01
5.64322770e-01 1.03695357e+00 -1.88847172e+00 -1.53837860e-01
7.97030151e-01 9.97180521e-01 2.73229718e-01 3.99637192e-01
-3.00398767e-01 3.87425691e-01 -1.92542329e-01 -1.41935036e-01
3.80929083e-01 -1.86504468e-01 -2.14633077e-01 2.88942724e-01
6.61910594e-01 -1.19248442e-01 -1.81108732e-02 7.44873583e-01
2.19570085e-01 2.50684083e-01 5.58441639e-01 -9.56082284e-01
4.44881350e-01 -2.80150145e-01 -9.12038922e-01 -1.43787175e-01
9.31352079e-01 -6.44249171e-02 8.02976787e-01 -1.18204403e+00
8.37946475e-01 8.71903241e-01 1.10787261e+00 -2.65517920e-01
-1.42356861e+00 -3.91109526e-01 -1.24946050e-01 1.44894555e-01
-1.62144244e+00 -5.37643135e-01 9.43945348e-01 -5.26318431e-01
8.37618113e-01 2.81304121e-01 7.63054848e-01 3.96492094e-01
1.75472781e-01 1.88898042e-01 1.00512719e+00 -1.02032340e+00
2.51418799e-01 2.04126433e-01 -7.63465390e-02 8.69593680e-01
5.18075943e-01 1.12189986e-01 -6.41049564e-01 -2.54451483e-01
5.51625073e-01 7.69403055e-02 -4.24566478e-01 -1.13848436e+00
-9.75860059e-01 6.82757795e-01 6.28309965e-01 5.01824379e-01
-4.83305752e-01 -2.00146474e-02 1.58415094e-01 6.35009632e-02
4.08206135e-01 2.45869473e-01 1.24169670e-01 1.89569697e-01
-1.46491468e+00 3.01503837e-01 4.02916878e-01 8.31579208e-01
1.46321285e+00 -5.90778828e-01 3.33142966e-01 4.42472875e-01
5.00729859e-01 3.30446899e-01 3.31365503e-02 -4.64325666e-01
4.13348764e-01 8.87453198e-01 1.97495863e-01 -1.51303911e+00
-5.51356375e-01 -2.91229874e-01 -6.89287245e-01 5.53385794e-01
3.74769896e-01 4.30210322e-01 -8.22426260e-01 1.04157364e+00
7.01086044e-01 -1.51753128e-01 -4.85627353e-02 7.27065861e-01
3.81441683e-01 4.01576847e-01 -3.87431055e-01 -6.31211251e-02
1.21304536e+00 -6.68662727e-01 -4.77544934e-01 -4.94928926e-01
6.62913144e-01 -1.14142597e+00 5.74946284e-01 2.13014349e-01
-6.35341406e-01 -5.61045468e-01 -1.27132976e+00 -2.77972594e-02
-6.02521598e-01 2.45913833e-01 4.51779425e-01 4.06319201e-01
-1.09012401e+00 4.39774603e-01 -4.50967938e-01 -8.10232103e-01
-6.77523091e-02 9.00672227e-02 -6.16305828e-01 -1.89129144e-01
-6.41823232e-01 9.47436631e-01 4.94596601e-01 4.02381539e-01
-1.67607412e-01 -3.03927302e-01 -1.15182233e+00 -2.35064223e-01
3.91139895e-01 -5.73333085e-01 6.39197409e-01 -7.98519909e-01
-9.40580368e-01 1.09132397e+00 -4.26980317e-01 -6.20795727e-01
9.67567623e-01 -6.17941841e-02 -6.88021332e-02 2.90629536e-01
1.75760493e-01 5.67430794e-01 8.72591555e-01 -1.67151165e+00
-7.77590334e-01 -4.24056262e-01 -1.66796714e-01 2.42386818e-01
4.01023626e-01 -2.31969535e-01 -7.02268600e-01 -2.16455266e-01
7.97433257e-01 -9.00321543e-01 -3.84238064e-01 3.47512335e-01
-1.27569199e-01 3.27376015e-02 8.25465858e-01 -5.91849208e-01
1.06718493e+00 -2.20221043e+00 -2.23105773e-01 9.36150074e-01
8.66493303e-03 8.15109164e-02 2.84972459e-01 7.85679877e-01
3.26313287e-01 -3.44507158e-01 -3.42972577e-01 -6.01278901e-01
-2.45156705e-01 7.23142400e-02 -1.88324541e-01 8.22145700e-01
8.62311497e-02 3.60694319e-01 -6.80244625e-01 -6.99449658e-01
8.39547276e-01 2.94159204e-01 -1.82198763e-01 -3.47831212e-02
-1.56980917e-01 3.83287996e-01 -3.12552720e-01 5.80478311e-01
1.07215917e+00 2.72571862e-01 2.78198402e-02 -1.18795902e-01
-7.89354265e-01 -1.14800185e-01 -1.42165327e+00 1.87868094e+00
-3.55006784e-01 7.47362673e-01 1.02333128e-01 -6.56773031e-01
1.47943366e+00 -1.23230316e-01 3.79752219e-01 -8.48552823e-01
6.16696104e-02 6.19372249e-01 -5.34576058e-01 -2.28165776e-01
8.57676327e-01 3.68302733e-01 7.55168349e-02 -6.22447506e-02
-4.29927438e-01 -4.86371219e-01 1.59043953e-01 9.78436321e-02
9.05525982e-01 4.88426238e-01 6.09763205e-01 3.19963656e-02
9.23613012e-01 4.11585718e-01 1.51138708e-01 7.48091102e-01
-1.38857719e-02 8.58906865e-01 2.84440815e-02 -5.10258079e-01
-1.13506114e+00 -6.57144308e-01 -2.00456813e-01 2.98042893e-01
6.26034677e-01 -5.50544381e-01 -4.80507046e-01 -2.30261296e-01
1.63344055e-01 2.90061712e-01 -5.60641766e-01 2.27960870e-01
-4.87912476e-01 -2.79306263e-01 1.00157402e-01 -6.42966926e-02
5.35532176e-01 -6.06926978e-01 -1.36094451e+00 4.19157416e-01
-1.28056303e-01 -9.47717249e-01 1.38773546e-01 3.44253778e-01
-9.22955632e-01 -1.14149892e+00 -5.72753310e-01 -4.88938898e-01
7.30865002e-01 7.81545579e-01 8.05236101e-01 4.65720266e-01
-4.38561469e-01 2.92070836e-01 -6.28374279e-01 -1.16373807e-01
-3.03915888e-01 -2.81743854e-01 -2.36514330e-01 1.69975907e-01
1.22980222e-01 -3.46845627e-01 -6.09344363e-01 3.06571722e-01
-5.96082985e-01 2.00869858e-01 7.62796044e-01 6.95760667e-01
7.00344980e-01 -2.53439873e-01 -3.28384399e-01 -2.63955683e-01
1.08004123e-01 -1.80807039e-01 -8.88871551e-01 3.28762829e-01
-5.45048058e-01 -1.19953053e-02 1.93336010e-01 7.85146132e-02
-7.55650640e-01 5.94514608e-01 -1.53162666e-02 -4.59137708e-01
-2.31942251e-01 4.34000254e-01 2.30918944e-01 -6.51871026e-01
7.11255670e-01 2.67566323e-01 7.05520390e-03 -4.33620125e-01
3.69450003e-01 6.40925467e-01 5.11021197e-01 -1.12114951e-01
1.00782955e+00 8.00397158e-01 2.46490538e-01 -1.15491784e+00
-2.95558482e-01 -1.13286805e+00 -9.03435707e-01 -5.61411560e-01
6.42276824e-01 -7.47605622e-01 -2.88211823e-01 3.86112928e-01
-1.36267650e+00 1.34586290e-01 -3.79459113e-02 4.99980211e-01
-5.57363987e-01 6.50264502e-01 -4.39015945e-04 -1.09103549e+00
-1.73945561e-01 -9.37196195e-01 1.23980808e+00 2.80087590e-01
-5.85126840e-02 -7.96251535e-01 4.05009627e-01 1.87214851e-01
1.31108984e-01 6.04244590e-01 4.37160254e-01 -1.77146405e-01
-9.08995509e-01 -5.48763752e-01 -3.12356353e-01 -1.17152452e-01
-1.84469879e-01 1.38957188e-01 -9.57872748e-01 -3.58972281e-01
-1.94107682e-01 6.95859194e-02 8.29064250e-01 9.64207649e-02
4.02785927e-01 4.11366582e-01 -6.57912076e-01 5.61017036e-01
1.83833230e+00 1.16214871e-01 7.77828097e-01 9.00167763e-01
4.19529498e-01 7.50754654e-01 1.11553979e+00 5.54479361e-01
3.49780053e-01 1.00908995e+00 5.89145064e-01 -3.96311045e-01
-1.62163675e-02 -3.28183085e-01 1.09158978e-01 2.82477915e-01
-1.62106559e-01 4.55369167e-02 -9.74981189e-01 4.58812714e-01
-2.04594779e+00 -7.71189630e-01 -6.19448006e-01 2.61976051e+00
4.57718112e-02 1.91335246e-01 -9.37047452e-02 3.85629743e-01
6.58342123e-01 1.29403412e-01 4.49076928e-02 -3.50470126e-01
-2.70283341e-01 -2.87359320e-02 9.03167188e-01 5.24822772e-01
-1.10474956e+00 6.96933389e-01 4.89161968e+00 7.14731634e-01
-1.16898978e+00 -8.11148658e-02 2.54656542e-02 5.93676865e-01
-6.47709221e-02 5.14532626e-01 -6.10967219e-01 2.09527686e-01
3.55792999e-01 3.35280538e-01 3.40529345e-02 7.29137599e-01
2.39315271e-01 -1.03937781e+00 -6.86674654e-01 1.15071797e+00
4.48780328e-01 -1.21067083e+00 -4.15947258e-01 7.17084706e-02
4.82665122e-01 7.54631907e-02 -4.79326516e-01 -3.86222273e-01
-4.54867661e-01 -4.85843033e-01 1.18319941e+00 7.12731898e-01
5.18445313e-01 -5.21404922e-01 8.00368965e-01 3.31975967e-01
-1.51637542e+00 3.37250121e-02 -2.67318606e-01 -1.58300716e-02
1.83810651e-01 6.65709317e-01 -1.01551557e+00 1.12968588e+00
7.33716488e-01 4.97596323e-01 -9.31158960e-01 1.77105224e+00
-1.01652071e-01 -9.08099934e-02 -6.59451723e-01 5.45020960e-02
4.88963544e-01 -4.00491446e-01 4.85357642e-01 1.24934828e+00
6.79978073e-01 -4.65319008e-01 1.91208959e-01 6.94789588e-01
6.46616280e-01 4.42064822e-01 -9.78692591e-01 7.49692500e-01
3.81311774e-01 1.29342711e+00 -1.17260480e+00 -2.07568649e-02
-3.89525384e-01 9.66804504e-01 2.34339461e-01 4.56933118e-02
-5.84113359e-01 -4.67205286e-01 7.08468556e-02 3.64460707e-01
4.40451592e-01 -6.18445873e-01 -3.79815139e-02 -7.95391798e-01
3.79184425e-01 -4.41258401e-01 7.51173347e-02 -8.30921769e-01
-4.73403037e-01 4.81106490e-01 -5.75910024e-02 -1.77169025e+00
-3.58873606e-01 -1.55818313e-01 -4.16746825e-01 7.75646150e-01
-1.71359003e+00 -1.40631807e+00 -7.33022094e-01 5.77875733e-01
3.23246986e-01 3.62189114e-01 6.21121943e-01 1.95482329e-01
-6.90365629e-03 2.39063930e-02 4.20523614e-01 -3.04298311e-01
8.05959404e-01 -9.38940704e-01 2.95813948e-01 1.12853336e+00
3.31562012e-01 3.92946780e-01 1.01474810e+00 -7.75146961e-01
-1.21406817e+00 -6.35606885e-01 1.02933085e+00 -2.61730235e-02
3.54378760e-01 -5.86980343e-01 -9.99097884e-01 3.02131891e-01
-1.98525656e-02 -1.84649155e-02 1.73717514e-02 -2.00535029e-01
-2.33371243e-01 -1.59130007e-01 -1.16705573e+00 2.88097054e-01
8.99148345e-01 -4.73811001e-01 -5.72501838e-01 5.36681013e-03
8.39805529e-02 -5.54995000e-01 -4.26056027e-01 4.39579338e-01
5.56953490e-01 -1.52827609e+00 9.26309228e-01 5.22606194e-01
-3.16108048e-01 -8.45458508e-01 1.12272119e-02 -1.06925726e+00
-3.09628434e-02 -4.75053221e-01 3.21164280e-01 1.20718443e+00
1.23081416e-01 -5.63720942e-01 7.30843663e-01 9.31584388e-02
-5.29503413e-02 -3.55249882e-01 -1.19535923e+00 -6.07415557e-01
-8.01983416e-01 -3.55009854e-01 3.09029490e-01 6.38888061e-01
-2.58655131e-01 -4.73550633e-02 -2.14950055e-01 1.75459534e-01
6.81728482e-01 3.60493749e-01 1.25921309e+00 -1.45853746e+00
1.50437638e-01 -3.25338632e-01 -8.52176547e-01 -8.04126740e-01
-3.72328043e-01 -5.08807778e-01 3.60639304e-01 -1.48622167e+00
-4.13958102e-01 -1.00134623e+00 2.74671435e-01 8.26349929e-02
1.89053237e-01 4.43062127e-01 2.99999148e-01 3.86777073e-01
-5.34798443e-01 3.79740059e-01 5.20716310e-01 1.64291099e-01
-3.01562130e-01 -1.80129364e-01 2.74645865e-01 8.13527644e-01
4.48075861e-01 -3.39729786e-01 -5.56812026e-02 -8.91873911e-02
3.00367296e-01 -1.28431693e-01 5.29704809e-01 -1.61318016e+00
6.61141932e-01 1.31653175e-01 4.63151366e-01 -1.18353224e+00
6.62828803e-01 -1.21465731e+00 5.21615803e-01 6.84436262e-01
2.81145871e-01 2.66224951e-01 2.13037372e-01 6.56016827e-01
-3.21433455e-01 -6.19040132e-01 5.41840315e-01 -7.74512738e-02
-1.22512650e+00 -1.72839314e-01 -3.31933945e-01 -6.31373703e-01
1.39365065e+00 -8.64537060e-01 -8.88440832e-02 -2.63672829e-01
-3.20291668e-01 1.95209563e-01 1.13922966e+00 1.84931189e-01
9.07791495e-01 -1.02016437e+00 -4.51297194e-01 4.46088970e-01
6.54929221e-01 2.25521207e-01 3.67569238e-01 1.19465292e+00
-1.21554828e+00 2.86603779e-01 -1.61477163e-01 -9.81075406e-01
-1.52545929e+00 5.88619351e-01 2.64234275e-01 -8.67088586e-02
-6.87128365e-01 3.18486214e-01 -2.55618244e-01 -1.40796065e-01
1.49799049e-01 -3.49204481e-01 -1.14493191e-01 4.78671134e-01
3.14887613e-01 2.93537170e-01 5.15756249e-01 -1.03522217e+00
-6.32727087e-01 1.36441588e+00 2.78526604e-01 -2.05624625e-01
1.09219921e+00 -4.84572679e-01 -1.88782901e-01 3.60402346e-01
9.01102901e-01 4.64258343e-01 -1.00678921e+00 -2.03636140e-01
2.64058173e-01 -7.56767213e-01 -4.15354297e-02 -2.63772279e-01
-3.85877550e-01 7.23484874e-01 8.72957706e-01 2.19459459e-01
1.02470219e+00 -1.16284572e-01 2.52786815e-01 2.05687165e-01
8.28464270e-01 -9.38357949e-01 -3.52974325e-01 3.31868947e-01
8.41057777e-01 -1.29553676e+00 2.10632339e-01 -6.23269558e-01
-1.63160548e-01 1.52212214e+00 9.12180617e-02 -3.27411704e-02
1.96866766e-01 4.79895584e-02 -1.79034695e-02 -1.63900688e-01
-8.50841254e-02 -5.93439460e-01 1.78649679e-01 5.45586228e-01
-4.85453568e-02 -1.79658160e-01 -5.70889354e-01 -4.13503379e-01
9.48828608e-02 3.42756063e-02 3.13917369e-01 1.17399955e+00
-7.81243742e-01 -9.99107957e-01 -1.03984451e+00 2.70838477e-02
1.93636149e-01 1.56267628e-01 -3.75464588e-01 9.95868266e-01
3.79712343e-01 9.14901733e-01 2.95493424e-01 -4.22715247e-01
4.58298385e-01 -1.66243687e-01 5.35118937e-01 -1.19269364e-01
-5.36356091e-01 1.23411901e-01 1.13938875e-01 -6.83264673e-01
-6.39694929e-01 -8.73383045e-01 -1.02984464e+00 -1.55111775e-01
-5.56916833e-01 8.50938931e-02 1.12102818e+00 6.44681692e-01
2.89023042e-01 -1.13715142e-01 6.42995715e-01 -1.38131166e+00
-1.21963486e-01 -5.55450678e-01 -3.40386003e-01 3.98861378e-01
4.92277145e-01 -9.39906776e-01 -3.70313257e-01 -4.17652220e-01] | [7.386581897735596, -2.1255929470062256] |
13eeabb5-9a53-48cd-968d-7377d97c97ef | broadening-the-perspective-for-sustainable-ai | 2306.13686 | null | https://arxiv.org/abs/2306.13686v1 | https://arxiv.org/pdf/2306.13686v1.pdf | Broadening the perspective for sustainable AI: Comprehensive sustainability criteria and indicators for AI systems | The increased use of AI systems is associated with multi-faceted societal, environmental, and economic consequences. These include non-transparent decision-making processes, discrimination, increasing inequalities, rising energy consumption and greenhouse gas emissions in AI model development and application, and an increasing concentration of economic power. By considering the multi-dimensionality of sustainability, this paper takes steps towards substantiating the call for an overarching perspective on "sustainable AI". It presents the SCAIS Framework (Sustainability Criteria and Indicators for Artificial Intelligence Systems) which contains a set 19 sustainability criteria for sustainable AI and 67 indicators that is based on the results of a critical review and expert workshops. This interdisciplinary approach contributes a unique holistic perspective to facilitate and structure the discourse on sustainable AI. Further, it provides a concrete framework that lays the foundation for developing standards and tools to support the conscious development and application of AI systems. | ['Ulrich Petschow', 'Marcus Voss', 'Philipp Reinhard', 'Andreas Meyer', 'Josephin Wagner', 'Friederike Rohde'] | 2023-06-22 | null | null | null | null | ['decision-making'] | ['reasoning'] | [ 4.23165739e-01 2.94380724e-01 -6.21834159e-01 1.10724472e-01
-2.22324520e-01 -5.88224649e-01 8.47875834e-01 1.78339958e-01
-1.91798478e-01 6.42649651e-01 5.32013476e-01 -3.41676503e-01
-5.72797894e-01 -6.28438711e-01 -1.99316427e-01 -7.05167115e-01
2.14468747e-01 9.71907899e-02 -3.77501905e-01 -2.92154253e-01
3.27710569e-01 5.11385620e-01 -1.56300604e+00 -1.49558350e-01
1.23450923e+00 9.81634855e-01 -2.91950941e-01 2.25248694e-01
1.81977246e-02 7.38431096e-01 -4.06290889e-01 -4.28520024e-01
-6.65078089e-02 -6.97213948e-01 -6.39807224e-01 -1.94206432e-01
-1.53454438e-01 -9.72039104e-02 3.00489902e-01 1.21516895e+00
2.20782116e-01 -1.99398130e-01 6.21906638e-01 -1.59357131e+00
-1.05352414e+00 7.00372040e-01 1.64453700e-01 -2.28381976e-01
6.92537427e-02 6.22211039e-01 7.88819730e-01 -5.01414895e-01
4.07696545e-01 1.05370343e+00 4.78068292e-01 2.17553079e-01
-7.37750411e-01 -8.77274632e-01 1.44609734e-01 3.94874573e-01
-1.18559301e+00 -6.37593508e-01 4.98002738e-01 -6.04946136e-01
8.60340714e-01 8.10474396e-01 1.54931748e+00 6.87358320e-01
3.40954244e-01 9.49307680e-02 1.09578955e+00 -8.01710486e-01
4.67244983e-01 3.64839524e-01 -2.55914964e-03 -1.74314305e-01
1.24269056e+00 1.29938930e-01 -2.04510555e-01 1.85720786e-01
2.16619328e-01 -3.37878197e-01 3.77731442e-01 -1.47279829e-01
-1.27651584e+00 5.31578839e-01 1.60049200e-01 9.14541483e-01
-7.69965887e-01 3.91048014e-01 1.44945204e-01 -1.16635703e-01
2.61761248e-01 7.19024122e-01 2.74842143e-01 -3.03354561e-01
-4.81113702e-01 2.56627768e-01 6.17513180e-01 3.34991872e-01
1.67056128e-01 6.05371535e-01 2.86974199e-02 9.96469930e-02
7.53064752e-01 1.07952428e+00 -1.94261134e-01 -1.69375455e+00
-3.06369364e-01 9.23284888e-01 4.41663206e-01 -1.09062779e+00
-3.27322423e-01 -5.52108526e-01 -3.65420371e-01 3.41895789e-01
5.09211735e-04 -2.43210882e-01 -6.54620469e-01 1.55896270e+00
1.57658592e-01 -5.73721588e-01 3.18942904e-01 8.53618801e-01
6.08239472e-01 5.70108771e-01 6.89929664e-01 -3.08969021e-01
9.81935799e-01 -4.02805150e-01 -1.27449679e+00 -2.77769595e-01
2.46327132e-01 -4.31147590e-02 8.15359414e-01 3.76598626e-01
-1.17832375e+00 1.14802584e-01 -1.15753102e+00 2.75918841e-01
-6.93719208e-01 -2.58690953e-01 6.52343094e-01 1.12573242e+00
-8.80622268e-01 1.10376887e-02 -5.45489728e-01 -6.60964370e-01
3.53509814e-01 1.57786623e-01 1.47722632e-01 1.53437302e-01
-1.13883138e+00 1.49450481e+00 5.96388221e-01 2.83101439e-01
-4.18461949e-01 -5.01684666e-01 -4.41494614e-01 3.79520643e-04
2.87230670e-01 -5.20435274e-01 6.42569959e-01 -1.30129266e+00
-1.24585295e+00 7.02713847e-01 4.08538640e-01 -4.29079592e-01
5.42785048e-01 -2.04385743e-01 -8.72649610e-01 2.05561221e-02
8.34499970e-02 3.21226239e-01 -5.46090491e-02 -1.48309636e+00
-7.38365531e-01 -6.44554734e-01 -1.71848059e-01 2.05407009e-01
-7.89100826e-01 4.25409287e-01 3.57132792e-01 -1.26453504e-01
-2.21669808e-01 -1.02619088e+00 -2.06330106e-01 -2.34621316e-01
1.15530752e-03 -2.42209375e-01 5.91738403e-01 -4.95252401e-01
1.61491525e+00 -1.84746003e+00 -6.02944680e-02 2.84541994e-01
2.30986863e-01 3.45871508e-01 2.58144915e-01 5.87598383e-01
3.10141742e-01 4.27319080e-01 -2.41423845e-01 7.60916829e-01
5.36246181e-01 1.80036426e-02 -1.29643098e-01 2.58039087e-01
2.60145515e-01 6.89431548e-01 -1.08678687e+00 -1.73153892e-01
4.47253406e-01 6.65505171e-01 -1.02698095e-01 -3.97609800e-01
-3.97314969e-03 4.08589482e-01 -4.17156190e-01 8.39003444e-01
3.46132368e-01 -2.51107272e-02 5.12421310e-01 1.47757322e-01
-8.64619553e-01 1.73162401e-01 -9.47441101e-01 1.14712405e+00
3.46699147e-03 5.56477129e-01 -3.27189756e-03 -5.40954769e-01
9.14948046e-01 4.97667998e-01 6.60682678e-01 -1.17895496e+00
3.17959607e-01 6.83298528e-01 5.73788464e-01 -6.64618313e-01
2.50740469e-01 5.43418676e-02 -2.56522030e-01 4.22598690e-01
-3.40153575e-01 -5.33999860e-01 2.34288827e-01 -1.88490972e-01
4.75563467e-01 5.20107329e-01 2.65859812e-01 -8.16103697e-01
7.05109477e-01 2.62205452e-01 5.62497795e-01 -8.26243311e-02
-6.86116755e-01 -3.59115273e-01 2.59261936e-01 -4.80540961e-01
-8.65760565e-01 -7.30849564e-01 -2.25621715e-01 7.46815681e-01
2.23964602e-01 -5.94424829e-02 -8.57538342e-01 -1.14618987e-01
-4.45178822e-02 1.50913119e+00 -5.16847014e-01 -4.14797813e-01
-1.93244383e-01 -7.12786555e-01 3.13931614e-01 1.15750410e-01
5.90241730e-01 -9.32068527e-01 -1.56851554e+00 2.40381345e-01
4.20019291e-02 -9.58179653e-01 5.55987477e-01 -1.58232555e-01
-3.63569647e-01 -1.02542067e+00 -4.50780660e-01 -1.95398837e-01
4.02906597e-01 2.13610306e-02 7.75549829e-01 -6.36626482e-02
-1.42015368e-01 3.32769752e-01 4.17683795e-02 -1.26431787e+00
-8.61060917e-01 -3.48252743e-01 1.52355701e-01 -5.95788538e-01
5.30412912e-01 -4.15838838e-01 -7.12778389e-01 -3.81316729e-02
-7.42315948e-01 6.91264048e-02 5.61725974e-01 -7.72518432e-03
2.87927240e-01 1.02524087e-01 1.40459883e+00 -3.42977405e-01
7.36229122e-01 -7.85225928e-01 -3.81257325e-01 4.06955808e-01
-1.60024178e+00 -4.60976332e-01 1.97427660e-01 1.71495572e-01
-1.16397417e+00 -4.30830002e-01 2.89211452e-01 3.36348891e-01
-1.85260266e-01 4.52436417e-01 -4.29949731e-01 -3.71164046e-02
5.87754369e-01 -2.23992050e-01 2.75750399e-01 4.69339043e-02
5.67599297e-01 6.85123801e-01 4.75150436e-01 -3.36385220e-01
6.39500201e-01 2.17704132e-01 7.33553171e-02 -8.61070156e-01
-1.20090656e-02 2.20584914e-01 -3.05609673e-01 -9.41885531e-01
8.70675921e-01 -6.50323927e-01 -5.56045890e-01 1.86184064e-01
-5.36419451e-01 -1.89511612e-01 -5.19637525e-01 6.42614067e-01
-3.45988721e-01 -1.88092381e-01 2.41084203e-01 -1.33550632e+00
-6.01026058e-01 -9.83698726e-01 7.44376928e-02 4.41702724e-01
-7.55865395e-01 -8.24216545e-01 7.75460750e-02 5.18237770e-01
8.56030166e-01 1.09662139e+00 9.77801025e-01 -4.90694255e-01
-4.44417328e-01 -1.89091980e-01 -2.01681238e-02 4.19515312e-01
1.85058802e-01 4.82163161e-01 -6.74612045e-01 4.19405878e-01
-3.98812257e-02 -4.68456857e-02 -3.67998937e-03 7.10443407e-02
2.27076516e-01 -7.67356217e-01 -3.82921547e-02 -1.21986523e-01
1.83015168e+00 1.08392823e+00 5.51117063e-01 7.60163367e-01
4.45794523e-01 1.17313099e+00 8.19443166e-01 4.05319542e-01
8.43250215e-01 9.52921584e-02 5.68942666e-01 2.00518519e-01
-4.68901806e-02 4.17834401e-01 2.26526961e-01 7.56983519e-01
-5.28909922e-01 9.43260193e-02 -1.40363026e+00 8.92055750e-01
-1.70006716e+00 -8.94329846e-01 -4.92730796e-01 2.13062835e+00
3.50526273e-01 1.39726475e-01 4.63507086e-01 1.97734743e-01
3.63500386e-01 -1.11727588e-01 -6.80037022e-01 -8.56146753e-01
1.21086195e-01 -2.53900349e-01 3.29381675e-01 7.07233846e-02
-6.03819132e-01 7.85029531e-01 7.24545479e+00 6.35685259e-03
-9.35013771e-01 -9.69396606e-02 4.03657705e-01 5.53353829e-03
-1.12790930e+00 -1.52187213e-01 -3.55861336e-02 3.48710358e-01
1.41851938e+00 -1.07688355e+00 1.18376233e-01 3.56106937e-01
5.45459688e-01 -2.45509058e-01 -5.14230549e-01 1.47319129e-02
-1.88977510e-01 -1.21656764e+00 -3.50333303e-02 3.30824882e-01
8.14220190e-01 -2.39568755e-01 3.18962425e-01 -3.67299527e-01
5.36458671e-01 -1.01640368e+00 1.22633815e+00 4.28079605e-01
6.33069634e-01 -1.00905597e+00 5.94006240e-01 -1.93070218e-01
-7.18070626e-01 -4.56423998e-01 1.93493098e-01 -3.32666785e-01
-1.32845864e-01 1.30939677e-01 -1.68150365e-01 5.44701695e-01
6.68178141e-01 1.35996699e-01 -1.52943775e-01 8.30082953e-01
-1.42397936e-02 5.08794308e-01 -2.77228892e-01 -1.34707943e-01
2.77276695e-01 -5.57153523e-01 6.87357664e-01 1.07158446e+00
3.36094081e-01 2.28712589e-01 -6.29997849e-01 1.13767028e+00
5.36977410e-01 1.29111528e-01 -7.65460014e-01 -9.60292399e-01
9.05248523e-01 1.02982759e+00 -8.06094170e-01 -1.51899248e-01
-4.42913383e-01 5.26768751e-02 -5.93653977e-01 3.88315409e-01
-6.91633761e-01 -2.65568256e-01 6.88503087e-01 1.01618715e-01
-3.12484086e-01 -1.80160478e-01 -1.10234606e+00 -4.49277848e-01
-3.94821048e-01 -1.02777410e+00 8.33994001e-02 -1.89172253e-01
-4.33149457e-01 1.94809377e-01 3.08010936e-01 -6.37963414e-01
-2.03527048e-01 -2.13231653e-01 -6.02493107e-01 4.50386584e-01
-1.49595511e+00 -1.49191082e+00 -3.07611465e-01 -6.18005767e-02
-2.99023122e-01 4.53827344e-02 1.15022922e+00 1.46067262e-01
-7.16369867e-01 6.79026842e-02 3.76691431e-01 -7.10191965e-01
2.06325069e-01 -9.30813015e-01 1.04639120e-01 9.79593635e-01
-7.55158722e-01 2.71873891e-01 8.05726528e-01 -7.95646310e-01
-1.67516923e+00 -6.44088924e-01 8.87390971e-01 -2.80122072e-01
9.40593719e-01 2.57174999e-01 -4.87891257e-01 6.32086337e-01
5.81384599e-01 -9.26046073e-01 1.23867679e+00 -3.83606493e-01
2.66263843e-01 -3.68621737e-01 -1.42756450e+00 8.07693660e-01
8.84562492e-01 1.93067770e-02 -3.78423005e-01 4.90134358e-02
6.78248286e-01 3.75218123e-01 -1.24379492e+00 3.30130279e-01
1.05918455e+00 -8.68647575e-01 9.31407154e-01 -1.06498271e-01
2.39387304e-01 -1.94864422e-01 -1.94827318e-01 -6.98030412e-01
-5.59111178e-01 -7.03632355e-01 2.89193075e-02 1.38789737e+00
2.58297235e-01 -1.13659012e+00 3.49228680e-01 1.57722688e+00
-4.61322784e-01 -5.67093849e-01 -9.88125086e-01 -6.17822409e-01
4.06535327e-01 -4.74983543e-01 1.04906702e+00 1.00898302e+00
5.32104552e-01 -3.11005086e-01 1.30611002e-01 -3.07568014e-01
9.67357695e-01 -3.57792586e-01 4.37904090e-01 -1.58296657e+00
9.63477254e-01 -9.03680563e-01 -3.77011955e-01 6.65161312e-01
-4.68353093e-01 -5.95001757e-01 -3.16672802e-01 -2.10488248e+00
-1.75501965e-02 -1.41979337e-01 -5.79666674e-01 2.98257053e-01
2.36146212e-01 1.03029206e-01 6.45835817e-01 3.87362897e-01
-3.95111352e-01 4.29309845e-01 9.95859027e-01 1.94932878e-01
-1.75699189e-01 -5.99330962e-01 -1.60341358e+00 8.62813830e-01
8.57465446e-01 -2.25540757e-01 -3.34505767e-01 -1.40589297e-01
6.54606640e-01 -4.01236206e-01 3.32321435e-01 -1.33516788e+00
3.32025327e-02 -1.09787035e+00 1.02010272e-01 -1.74623072e-01
1.05174415e-01 -1.48568833e+00 1.04603291e+00 1.19495606e+00
-2.82557368e-01 -2.61458278e-01 1.49997905e-01 -2.17423424e-01
2.89134949e-01 1.24100514e-01 5.78443885e-01 1.96374044e-01
-5.18608093e-01 -2.56554335e-01 -6.57184005e-01 -3.75492781e-01
1.67238462e+00 -6.87844694e-01 -5.20242333e-01 1.12233534e-02
-4.58217889e-01 4.35540825e-01 7.47119963e-01 3.17655087e-01
2.74764240e-01 -1.26303971e+00 -8.22228014e-01 -1.00359492e-01
-3.82538373e-03 -2.70891935e-01 1.68308362e-01 6.49833262e-01
-5.17913580e-01 7.98886538e-01 -7.27394104e-01 1.12358339e-01
-6.89433813e-01 1.65246487e-01 2.48303682e-01 3.47925723e-01
-5.17068624e-01 2.79979110e-01 -2.14029118e-01 1.94638267e-01
2.62724012e-01 -3.26563828e-02 -5.45931876e-01 3.29958290e-01
3.72172207e-01 1.15508354e+00 -5.59356928e-01 -9.67771411e-01
-7.43517637e-01 4.41129655e-01 5.93484700e-01 -4.63832080e-01
1.40764868e+00 -4.81532276e-01 -4.31904256e-01 7.91769564e-01
2.63064653e-01 -2.74872094e-01 -8.90163243e-01 4.75781262e-01
4.36403871e-01 -3.86746228e-01 6.55992627e-02 -1.46623433e+00
-5.01334906e-01 5.94034076e-01 8.31312239e-01 5.93988121e-01
1.13758671e+00 -3.72725129e-01 5.54504156e-01 3.79599184e-02
7.56734191e-03 -1.71965182e+00 -3.01885635e-01 -1.79650605e-01
1.30646586e+00 -6.72264338e-01 1.79692954e-01 -2.76397884e-01
-7.37920821e-01 1.13275445e+00 6.11148238e-01 1.99940190e-01
5.04170716e-01 2.80298412e-01 3.17952633e-02 -2.02103227e-01
-4.94298637e-01 -3.27159196e-01 4.99040559e-02 8.97126913e-01
5.07907033e-01 6.21428251e-01 -1.08506167e+00 7.66214252e-01
-2.29375996e-02 5.06879568e-01 2.99876690e-01 7.18204737e-01
-7.40353405e-01 -8.57189953e-01 -6.23444676e-01 5.00133000e-02
-6.89872622e-01 3.82824779e-01 -9.41332161e-01 9.38006997e-01
4.53894585e-01 1.14729202e+00 -6.08323738e-02 -2.71712571e-01
2.39079326e-01 1.77487701e-01 1.45478562e-01 2.54227608e-01
-7.41175830e-01 -4.12511546e-03 6.22557342e-01 -3.21999967e-01
-6.57618225e-01 -9.23467398e-01 -1.44518292e+00 -4.58277762e-01
1.33808190e-02 9.29538533e-02 1.35869479e+00 8.45131457e-01
7.10141182e-01 9.05088663e-01 2.89305568e-01 -3.07396412e-01
-2.67739058e-01 -5.70649624e-01 -2.36602396e-01 -1.50472835e-01
1.57115255e-02 -2.68668056e-01 -1.66221321e-01 -1.25897110e-01] | [9.014663696289062, 6.351232528686523] |
750b14b8-3539-420b-9ca8-eb1438a2cf7d | neural-symbolic-inference-for-robust | 2301.11459 | null | https://arxiv.org/abs/2301.11459v1 | https://arxiv.org/pdf/2301.11459v1.pdf | Neural-Symbolic Inference for Robust Autoregressive Graph Parsing via Compositional Uncertainty Quantification | Pre-trained seq2seq models excel at graph semantic parsing with rich annotated data, but generalize worse to out-of-distribution (OOD) and long-tail examples. In comparison, symbolic parsers under-perform on population-level metrics, but exhibit unique strength in OOD and tail generalization. In this work, we study compositionality-aware approach to neural-symbolic inference informed by model confidence, performing fine-grained neural-symbolic reasoning at subgraph level (i.e., nodes and edges) and precisely targeting subgraph components with high uncertainty in the neural parser. As a result, the method combines the distinct strength of the neural and symbolic approaches in capturing different aspects of the graph prediction, leading to well-rounded generalization performance both across domains and in the tail. We empirically investigate the approach in the English Resource Grammar (ERG) parsing problem on a diverse suite of standard in-domain and seven OOD corpora. Our approach leads to 35.26% and 35.60% error reduction in aggregated Smatch score over neural and symbolic approaches respectively, and 14% absolute accuracy gain in key tail linguistic categories over the neural model, outperforming prior state-of-art methods that do not account for compositionality or uncertainty. | ['Jingbo Shang', 'Jeremiah Liu', 'Zi Lin'] | 2023-01-26 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 3.95348966e-01 8.91946435e-01 -2.32672498e-01 -5.57314515e-01
-1.06103396e+00 -8.27658415e-01 5.00286460e-01 3.01703393e-01
5.89697026e-02 7.55142808e-01 3.40073615e-01 -4.87946302e-01
-2.51314163e-01 -9.70114052e-01 -1.03334975e+00 -2.18249202e-01
-3.17273468e-01 8.71282756e-01 1.45663112e-01 -1.41218886e-01
-1.31785065e-01 1.50183916e-01 -1.21496427e+00 2.53952563e-01
1.04222083e+00 1.08220232e+00 -1.20509349e-01 4.49997067e-01
-5.24974585e-01 8.25828910e-01 -6.03765070e-01 -8.18021417e-01
-7.73733333e-02 -2.80502439e-01 -7.83014655e-01 -4.20327455e-01
9.53191578e-01 3.05433404e-02 -1.35256678e-01 1.30652475e+00
5.95496953e-01 -2.59965926e-01 7.53117919e-01 -8.53623450e-01
-8.38500381e-01 1.47307849e+00 -4.20093954e-01 2.01480836e-01
4.53557611e-01 1.13190681e-01 1.64199448e+00 -3.00600976e-01
7.78312206e-01 1.71285558e+00 1.04984379e+00 7.39572465e-01
-1.55009854e+00 -5.94699562e-01 5.25152147e-01 -4.79837656e-01
-1.01136839e+00 -1.76033929e-01 6.54504120e-01 -3.59548867e-01
1.57800257e+00 -6.80112094e-02 8.39509070e-02 1.48045599e+00
1.14157774e-01 7.57107556e-01 9.77072537e-01 -3.37494075e-01
3.33770096e-01 -4.90572184e-01 5.20444572e-01 8.95117104e-01
7.41113782e-01 3.16505015e-01 -7.47855544e-01 -1.94281936e-02
3.70387226e-01 -6.62458003e-01 2.32699871e-01 -2.65093762e-02
-7.73002386e-01 7.87436485e-01 3.91635150e-01 1.47349507e-01
-3.44068438e-01 6.52231634e-01 7.59417832e-01 2.14276358e-01
7.47706115e-01 4.72073466e-01 -1.01920164e+00 -3.27673495e-01
-8.09715927e-01 4.40217435e-01 1.06191993e+00 1.35940993e+00
5.22693992e-01 3.71911854e-01 -3.92869979e-01 7.69036055e-01
1.47117376e-01 3.01295429e-01 2.77489573e-01 -8.52674186e-01
9.30044770e-01 6.98539495e-01 -6.65105224e-01 -7.33208835e-01
-6.83411777e-01 -7.65869975e-01 -6.88292861e-01 -2.49983951e-01
7.90598392e-01 -1.57914460e-01 -1.15653777e+00 2.38834810e+00
7.97446445e-02 1.38303950e-01 2.51751125e-01 3.36386532e-01
1.06107306e+00 2.74666458e-01 5.90798199e-01 2.43454799e-01
1.51405013e+00 -6.62760019e-01 -3.22862893e-01 -7.89623678e-01
9.62427914e-01 4.69193384e-02 1.30514419e+00 2.90453911e-01
-1.03602839e+00 -4.32920486e-01 -8.30962777e-01 -2.30988637e-01
-3.59091491e-01 -8.24392885e-02 8.02300215e-01 6.86875284e-01
-9.88654673e-01 8.83402824e-01 -6.75863504e-01 -3.44942063e-01
5.78167319e-01 3.74413699e-01 -1.96692899e-01 -2.02478647e-01
-1.46068358e+00 7.05695450e-01 8.09053659e-01 -1.34138986e-01
-7.39757419e-01 -1.16488123e+00 -1.14876211e+00 1.76571414e-01
5.25548518e-01 -7.45314837e-01 1.27117491e+00 -9.30122972e-01
-1.28124058e+00 1.15499234e+00 -5.82349999e-03 -9.35547233e-01
3.58685285e-01 -3.77826214e-01 -2.28441298e-01 -1.53816134e-01
3.40020299e-01 9.36716139e-01 4.33579206e-01 -8.22782993e-01
-3.63869965e-01 -5.54062605e-01 1.09680884e-01 -9.44378823e-02
1.61412120e-01 -1.38721988e-01 -2.25783870e-01 -6.97398424e-01
1.13703713e-01 -7.47796178e-01 5.23675755e-02 -7.72818029e-01
-6.96317732e-01 -4.98500317e-01 1.50089160e-01 -9.89547789e-01
1.39309728e+00 -2.01922917e+00 2.20902488e-01 3.10090221e-02
9.83069986e-02 1.99641231e-02 -3.52172256e-02 5.00790536e-01
2.41607130e-02 3.74583006e-01 -3.99611294e-01 -2.57346123e-01
5.09708941e-01 5.18440187e-01 -2.57037222e-01 -1.96157768e-02
7.71917403e-01 1.22332537e+00 -9.95099485e-01 -5.52411318e-01
-3.40237170e-01 2.19701156e-02 -8.51963282e-01 -1.18489519e-01
-9.22753870e-01 1.09081283e-01 -5.92451394e-01 9.41590309e-01
4.05021280e-01 -7.82040283e-02 3.99964958e-01 2.43952453e-01
4.32696790e-01 7.75698602e-01 -7.41831958e-01 1.98471725e+00
-4.79900151e-01 1.61392212e-01 -1.36700496e-01 -9.54427242e-01
1.04243898e+00 -7.69189149e-02 -3.50591779e-01 -7.99947441e-01
-7.62696713e-02 5.96370697e-01 4.17537332e-01 -1.92614034e-01
2.02615485e-01 -2.72025079e-01 -7.24574387e-01 -3.58461440e-02
5.25648475e-01 -1.91578805e-01 2.02700600e-01 2.30243504e-01
1.32449281e+00 6.42436266e-01 3.99266243e-01 -6.44196451e-01
2.86834300e-01 -2.47751530e-02 6.12377286e-01 9.41347241e-01
2.21987255e-02 2.71533132e-01 1.37220812e+00 -1.84905052e-01
-8.36917698e-01 -9.55015898e-01 -2.60843560e-02 1.42703092e+00
-3.17061067e-01 -3.16969633e-01 -1.10313356e+00 -9.85637665e-01
3.90974551e-01 1.30599725e+00 -8.77333403e-01 -3.66190046e-01
-8.38413239e-01 -6.13121569e-01 1.18348444e+00 9.20522094e-01
2.90649831e-01 -1.12812924e+00 -2.03472115e-02 3.19912970e-01
3.09819877e-01 -1.44178867e+00 -1.05483368e-01 5.28444171e-01
-1.02515972e+00 -7.01633275e-01 -2.28861034e-01 -8.27854693e-01
1.78874627e-01 -9.49738085e-01 1.78347123e+00 -2.36471638e-01
3.75018306e-02 3.98286730e-02 -1.50398031e-01 -4.18857485e-01
-6.20452285e-01 5.68378806e-01 -3.42108369e-01 -6.05009675e-01
5.06949782e-01 -6.97626948e-01 1.69450819e-01 -2.50131369e-01
-4.26622748e-01 -9.92233828e-02 6.36056364e-01 1.00944459e+00
5.24245560e-01 -3.77029032e-01 1.01311016e+00 -1.39005685e+00
7.46563792e-01 -7.24480152e-01 -7.22317517e-01 2.61233866e-01
-7.44132936e-01 6.27073407e-01 7.58804023e-01 -1.94686562e-01
-1.08772087e+00 -2.26986378e-01 -1.65015548e-01 -2.65522487e-02
-2.91154832e-01 6.93467975e-01 -4.02246058e-01 2.86970586e-01
6.55169904e-01 6.26156852e-03 -2.15673387e-01 -5.54416955e-01
4.15009022e-01 2.11270526e-01 5.51353872e-01 -1.23916078e+00
4.30163920e-01 -2.12800160e-01 4.08669889e-01 -4.50871348e-01
-1.16230571e+00 2.13792160e-01 -5.19319773e-01 3.68690282e-01
1.05991042e+00 -8.35731626e-01 -3.88723731e-01 3.56239021e-01
-1.05824924e+00 -6.75491750e-01 -3.41042161e-01 7.62461796e-02
-5.61793089e-01 3.19882214e-01 -7.40178764e-01 -5.92950761e-01
-4.11497414e-01 -9.22971189e-01 1.63270283e+00 -1.09561952e-02
-3.76829088e-01 -1.29014230e+00 -1.90069214e-01 1.25227179e-02
2.04999506e-01 6.57228827e-01 1.53554368e+00 -1.30231297e+00
-4.00666296e-01 6.81757629e-02 -3.54133636e-01 2.93298423e-01
-3.73293132e-01 -4.28788990e-01 -9.71626878e-01 2.06218556e-01
-6.18893087e-01 -4.22692716e-01 9.93722558e-01 5.16425669e-01
1.00584996e+00 -9.61069539e-02 -2.95934826e-01 6.62129045e-01
1.35183465e+00 -2.21284017e-01 2.08421558e-01 1.80327520e-01
7.05548346e-01 7.49591589e-01 5.16209006e-01 -2.67435871e-02
4.17579532e-01 4.68829662e-01 5.53191304e-01 4.56512928e-01
-4.45143878e-01 -8.30618620e-01 5.46656013e-01 4.04456228e-01
1.73698053e-01 -4.64845449e-01 -1.15588379e+00 4.63438869e-01
-1.55783474e+00 -5.13110280e-01 -1.69292077e-01 1.72300518e+00
9.39514041e-01 6.41777754e-01 1.45270824e-01 -1.46564916e-01
6.37665093e-01 9.60519016e-02 -6.56296253e-01 -9.09998238e-01
-2.90545970e-01 5.57455301e-01 6.28844023e-01 6.19154096e-01
-9.19650853e-01 1.61282456e+00 5.96699095e+00 1.07500565e+00
-5.96839786e-01 -1.10686636e-02 7.33155131e-01 2.38611966e-01
-5.37225723e-01 -2.97365207e-02 -1.23873258e+00 3.40457320e-01
1.38356471e+00 7.22561255e-02 4.92815912e-01 8.82518053e-01
-4.50763315e-01 1.89449340e-01 -1.47253156e+00 4.49576348e-01
-8.48115310e-02 -1.20519245e+00 1.36854053e-01 1.04305550e-01
6.54099107e-01 2.08860800e-01 -1.34385183e-01 7.45308876e-01
7.00244069e-01 -1.26956522e+00 1.00644135e+00 3.49448949e-01
9.54608381e-01 -6.22044265e-01 8.18663239e-01 4.07557070e-01
-9.61897075e-01 -1.53214529e-01 -2.39946112e-01 -8.25893730e-02
7.79342875e-02 4.93047535e-01 -7.62667298e-01 6.84322178e-01
6.11914217e-01 6.20740592e-01 -6.89862967e-01 1.22211605e-01
-5.25258958e-01 1.10435224e+00 -4.34422702e-01 -2.76619315e-01
6.12697482e-01 -9.14736744e-03 5.82997859e-01 1.48115456e+00
9.92889255e-02 -6.25541657e-02 -8.88973475e-02 1.20721555e+00
-2.76284933e-01 -2.29896396e-01 -7.35993922e-01 -3.87286454e-01
5.25076509e-01 7.29145408e-01 -4.57334250e-01 -3.75972629e-01
-2.41629586e-01 6.92975879e-01 7.44578123e-01 2.98861653e-01
-8.15130532e-01 -4.18228120e-01 5.11082113e-01 -1.26077205e-01
5.77971041e-01 -8.48386437e-02 -4.77141500e-01 -8.51838410e-01
2.05509122e-02 -8.93965840e-01 7.91161060e-01 -4.62983549e-01
-1.50734520e+00 6.26815856e-01 2.87766904e-01 -2.98140585e-01
-7.24517763e-01 -1.12163198e+00 -4.24784273e-01 9.17383015e-01
-1.46223295e+00 -1.58025098e+00 1.06574409e-01 6.19981401e-02
3.51889163e-01 -2.26766497e-01 8.90950561e-01 7.27666169e-02
-4.52226430e-01 8.87042999e-01 -2.65229166e-01 2.07371056e-01
1.09341525e-01 -1.79710782e+00 9.92325723e-01 8.77337813e-01
7.69636035e-02 4.91666943e-01 6.52536809e-01 -8.38806510e-01
-1.43283510e+00 -1.26543462e+00 1.10755837e+00 -4.35356587e-01
8.84052157e-01 -7.10373282e-01 -7.68899322e-01 1.13384330e+00
-1.78628221e-01 9.08339210e-03 1.75745219e-01 5.04535854e-01
-6.50101006e-01 1.54793590e-01 -1.24831653e+00 4.61407810e-01
1.79962766e+00 -4.51272786e-01 -7.78111637e-01 1.17681116e-01
1.36034918e+00 -5.94635725e-01 -9.82476473e-01 4.90582943e-01
2.06568345e-01 -9.07958746e-01 8.92633319e-01 -9.91506696e-01
6.15041733e-01 2.37217486e-01 -4.92224634e-01 -1.24235475e+00
-2.87148088e-01 -6.83070838e-01 -2.91956007e-01 1.67796338e+00
8.35871994e-01 -9.51542079e-01 7.76524365e-01 2.49685630e-01
-6.58437252e-01 -9.12417710e-01 -1.08228981e+00 -9.67576742e-01
6.10042989e-01 -6.97946370e-01 8.24077427e-01 5.59324741e-01
-1.89918339e-01 5.01171708e-01 1.53928310e-01 3.38281721e-01
7.41502106e-01 3.05618141e-02 4.10007775e-01 -1.23096478e+00
-5.14874578e-01 -6.83872581e-01 -5.04381120e-01 -8.27628493e-01
8.98537397e-01 -1.50216901e+00 -2.72090018e-01 -1.45511937e+00
-1.48025483e-01 -3.23190749e-01 3.79123911e-02 5.75714827e-01
-1.52229428e-01 -1.42494678e-01 -1.43841535e-01 -5.17429173e-01
-3.81224126e-01 1.70103133e-01 8.42052817e-01 1.64997935e-01
9.86409262e-02 -1.69195101e-01 -1.07201922e+00 7.77665257e-01
7.30868280e-01 -5.16692221e-01 -2.82033265e-01 -6.10965192e-01
5.40304065e-01 3.13207842e-02 4.34319496e-01 -6.74283385e-01
-2.54826784e-01 1.68588459e-01 5.31508289e-02 -3.04057479e-01
-1.43217528e-02 -3.99451584e-01 -1.33022025e-01 6.51099980e-01
-6.07338250e-01 -3.30828428e-02 5.19881368e-01 7.32146621e-01
-1.42452121e-01 -2.40661800e-01 4.24819350e-01 -1.94651395e-01
-7.32351780e-01 1.41322866e-01 2.55681694e-01 1.02159107e+00
5.63606977e-01 -2.78527260e-01 -4.87147152e-01 -7.86038414e-02
-7.53791332e-01 1.65586174e-01 6.24019466e-02 3.78765255e-01
7.31140301e-02 -9.04735208e-01 -8.97454321e-01 1.24264233e-01
1.42234430e-01 1.93476528e-01 -3.07828449e-02 6.72507882e-01
-5.14234602e-01 4.64974493e-01 1.83204785e-01 -5.69068849e-01
-7.24479437e-01 3.66486460e-01 4.77102757e-01 -5.95401108e-01
-6.01602614e-01 1.21391249e+00 2.07648247e-01 -9.63306129e-01
3.31701666e-01 -1.02329135e+00 1.89285338e-01 -1.46039858e-01
-2.03045353e-01 2.83579230e-01 3.78110409e-02 -3.85575116e-01
-5.04818141e-01 5.83715618e-01 2.62131631e-01 2.44516134e-01
1.38659847e+00 2.55082846e-01 8.04636478e-02 5.09862721e-01
1.01190472e+00 1.03490986e-01 -1.12773359e+00 -2.50270426e-01
6.15242243e-01 1.67094558e-01 -2.18411237e-01 -1.34664643e+00
-7.13045359e-01 1.13366890e+00 -2.69435085e-02 2.53062129e-01
7.74477959e-01 5.62414944e-01 7.09504247e-01 1.45577356e-01
2.68798947e-01 -8.38223934e-01 -4.01287496e-01 7.47183859e-01
7.40402460e-01 -9.52126920e-01 -3.38871956e-01 -5.96628070e-01
-8.45308900e-01 8.73622596e-01 6.33057117e-01 -4.41107661e-01
3.03856432e-01 3.23899835e-01 -5.55801570e-01 -3.83799016e-01
-9.54395354e-01 -1.89223915e-01 3.98839653e-01 7.06085265e-01
5.43756902e-01 2.71681219e-01 -9.23146456e-02 1.23291552e+00
-5.01764834e-01 -3.92053246e-01 -1.49248049e-01 4.87791777e-01
-3.45588356e-01 -8.35085034e-01 1.89821362e-01 4.87340003e-01
-8.37856948e-01 -6.01408601e-01 -6.39652550e-01 1.25526917e+00
4.30195592e-02 4.63948697e-01 8.67969356e-03 -6.08223118e-02
4.75135624e-01 4.97940958e-01 6.50284231e-01 -7.39539623e-01
-8.51334333e-01 -2.62144417e-01 9.74142313e-01 -7.07785606e-01
1.84422910e-01 -8.91695023e-01 -1.57595086e+00 3.24714594e-02
-2.38895282e-01 -1.30072758e-01 4.51205701e-01 7.59460807e-01
5.86081624e-01 7.78816342e-01 -2.76698560e-01 -5.24204075e-01
-1.13850307e+00 -1.16387081e+00 -6.87959790e-01 3.80820632e-01
-5.24708405e-02 -5.54197490e-01 -5.18583119e-01 -4.51583803e-01] | [10.489459991455078, 9.166807174682617] |
2fd7fae0-d99f-4913-a681-ea59ae6975eb | a-non-anatomical-graph-structure-for-isolated | 2207.07619 | null | https://arxiv.org/abs/2207.07619v1 | https://arxiv.org/pdf/2207.07619v1.pdf | A Non-Anatomical Graph Structure for isolated hand gesture separation in continuous gesture sequences | Continuous Hand Gesture Recognition (CHGR) has been extensively studied by researchers in the last few decades. Recently, one model has been presented to deal with the challenge of the boundary detection of isolated gestures in a continuous gesture video [17]. To enhance the model performance and also replace the handcrafted feature extractor in the presented model in [17], we propose a GCN model and combine it with the stacked Bi-LSTM and Attention modules to push the temporal information in the video stream. Considering the breakthroughs of GCN models for skeleton modality, we propose a two-layer GCN model to empower the 3D hand skeleton features. Finally, the class probabilities of each isolated gesture are fed to the post-processing module, borrowed from [17]. Furthermore, we replace the anatomical graph structure with some non-anatomical graph structures. Due to the lack of a large dataset, including both the continuous gesture sequences and the corresponding isolated gestures, three public datasets in Dynamic Hand Gesture Recognition (DHGR), RKS-PERSIANSIGN, and ASLVID, are used for evaluation. Experimental results show the superiority of the proposed model in dealing with isolated gesture boundaries detection in continuous gesture sequences | ['Sergio Escalera', 'Kourosh Kiani', 'Razieh Rastgoo'] | 2022-07-15 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition', 'boundary-detection'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 2.09622547e-01 -2.68025368e-01 -2.54415035e-01 -1.60134748e-01
-4.29036826e-01 -5.39693795e-02 4.27497298e-01 -7.85275578e-01
-7.41255283e-01 3.41323555e-01 4.33909357e-01 9.16876346e-02
-4.44306917e-02 -3.01383317e-01 -3.01854104e-01 -9.90155280e-01
-2.66918749e-01 2.59635895e-01 4.78336900e-01 3.32691334e-02
6.59238324e-02 5.68151355e-01 -1.23812997e+00 3.46293986e-01
4.07450765e-01 9.69689190e-01 2.66076237e-01 7.90986955e-01
-2.96250373e-01 5.17901063e-01 -4.35125977e-01 -9.61449593e-02
2.66159624e-01 -5.65761209e-01 -6.24981582e-01 -6.69092173e-03
3.88170898e-01 -9.26333904e-01 -9.95015144e-01 7.14078069e-01
1.05101645e+00 3.49213988e-01 6.20332956e-01 -1.15085602e+00
-2.35866189e-01 7.20014453e-01 -7.30409145e-01 3.49925280e-01
3.37859511e-01 2.62956411e-01 8.97001326e-01 -8.71873736e-01
9.76657450e-01 1.47943437e+00 4.14664954e-01 8.28712821e-01
-2.63142437e-01 -7.57021189e-01 7.05183625e-01 6.78113639e-01
-1.29664183e+00 -3.01383901e-02 7.62340188e-01 -3.65791529e-01
1.05929112e+00 1.64005011e-01 9.76294279e-01 1.76283228e+00
-1.70843095e-01 1.53300834e+00 5.94565153e-01 -3.10100198e-01
-5.58564104e-02 -1.09158528e+00 2.55696982e-01 7.65114546e-01
-2.31691524e-01 3.14389646e-01 -5.57766557e-01 2.09603414e-01
1.11877501e+00 4.77924287e-01 -4.22410220e-01 -1.07058220e-01
-1.27787364e+00 3.45326513e-01 5.83797276e-01 6.03657603e-01
-5.24037600e-01 2.53955960e-01 4.87973988e-01 3.82253528e-02
-1.06617689e-01 -5.96381962e-01 -3.13283354e-01 -2.72445649e-01
-9.76816833e-01 8.74470249e-02 5.94077766e-01 9.12730873e-01
-1.58187717e-01 6.51023462e-02 -4.49138284e-01 6.72800124e-01
6.42313480e-01 4.89104599e-01 8.82182956e-01 -2.69258589e-01
7.46423066e-01 6.45696104e-01 -2.93565780e-01 -6.83333039e-01
-7.01016188e-01 -3.89005095e-02 -8.48061562e-01 5.49547188e-02
6.42500877e-01 -8.35193992e-02 -1.45445442e+00 1.50593781e+00
2.64524132e-01 3.66202056e-01 -2.70995289e-01 1.53832459e+00
1.15441656e+00 1.98226810e-01 3.75062883e-01 1.85362231e-02
1.26424420e+00 -1.17262590e+00 -6.41302764e-01 2.73946524e-01
2.35108852e-01 -5.32935500e-01 1.03887022e+00 4.12273794e-01
-5.19613087e-01 -4.86876220e-01 -7.82234848e-01 -5.61979227e-02
-1.25983298e-01 3.92471820e-01 5.54352403e-01 3.94692123e-01
-6.80085659e-01 3.80041182e-01 -1.39086151e+00 -5.98924875e-01
5.68244159e-01 4.42293167e-01 -3.48133326e-01 7.60116195e-03
-9.12160337e-01 5.23813188e-01 5.90791464e-01 8.24407279e-01
-8.03065300e-01 1.44000184e-02 -5.50440252e-01 -1.14450425e-01
5.22158623e-01 -4.82630372e-01 8.42732072e-01 -6.50899470e-01
-1.70849872e+00 3.79545569e-01 -5.55468388e-02 8.33116993e-02
1.02299142e+00 -4.51175243e-01 -2.50324577e-01 3.45853895e-01
-5.46284556e-01 5.91851950e-01 9.88458276e-01 -8.55554342e-01
-7.04933763e-01 -8.17673147e-01 -2.21010536e-01 3.59068573e-01
-2.92346686e-01 2.21359625e-01 -8.72187495e-01 -9.57827270e-01
5.88004470e-01 -9.56297338e-01 -7.96651989e-02 -2.03132749e-01
-5.42085350e-01 -4.32061464e-01 1.00495803e+00 -1.23186159e+00
1.25166070e+00 -2.02295327e+00 5.93568802e-01 2.97308475e-01
6.02661520e-02 3.73824209e-01 -3.61899108e-01 1.24055400e-01
1.31029889e-01 -8.48178044e-02 -2.94744194e-01 -2.57271647e-01
-2.78646275e-02 4.51376706e-01 -4.91139628e-02 4.46223438e-01
9.36565772e-02 1.26815629e+00 -8.04561496e-01 -7.89232612e-01
3.53633136e-01 6.84157908e-01 -2.71283060e-01 3.22469115e-01
-1.53824568e-01 8.45039546e-01 -6.74101651e-01 9.82174218e-01
4.32606220e-01 9.43617821e-02 2.76536614e-01 -5.74415922e-02
7.40917921e-02 -2.35663906e-01 -1.21690917e+00 2.08121252e+00
8.10554475e-02 1.71977162e-01 -5.52293286e-03 -5.95175505e-01
5.02739489e-01 6.02613032e-01 5.93752980e-01 -5.75298965e-01
5.80836713e-01 2.08990559e-01 3.03755283e-01 -1.03666055e+00
5.28974012e-02 4.50380221e-02 3.48584652e-01 4.50408548e-01
2.73469269e-01 4.59078670e-01 1.69705659e-01 -1.62475273e-01
1.05597448e+00 6.03446543e-01 -1.06675461e-01 4.22958195e-01
6.23034060e-01 -2.68329233e-01 4.35335934e-01 5.73004425e-01
-4.80119705e-01 8.91323507e-01 2.44763315e-01 -5.24984777e-01
-5.86336374e-01 -8.91652882e-01 3.11499536e-01 1.23971188e+00
5.68164214e-02 -1.26928404e-01 -7.24029243e-01 -1.03545439e+00
-1.49906471e-01 -1.15544222e-01 -6.97635531e-01 2.91483879e-01
-1.33228707e+00 -7.23883510e-01 8.86698186e-01 1.12658787e+00
7.92904973e-01 -1.69296396e+00 -1.00818861e+00 1.56952500e-01
-3.57614517e-01 -1.15254629e+00 -6.78550601e-01 -1.13634109e-01
-9.45647299e-01 -1.26802838e+00 -1.38362432e+00 -1.01730478e+00
4.51662958e-01 -2.76485067e-02 1.59035534e-01 2.81311959e-01
-4.64312434e-01 4.43586558e-01 -8.21059227e-01 -5.06856712e-03
1.80253223e-01 2.73314565e-01 -5.93279041e-02 1.65416837e-01
3.23515832e-01 -5.08056045e-01 -5.77108204e-01 3.42423558e-01
-7.57953823e-01 -1.50964066e-01 9.11604881e-01 9.78397667e-01
5.06115496e-01 -6.65326655e-01 3.60808343e-01 -2.00194165e-01
4.33349371e-01 -3.96986268e-02 -1.17242955e-01 4.67447340e-01
-2.40118697e-01 -4.52752896e-02 1.99038386e-01 -8.01579416e-01
-1.14854515e+00 2.62310803e-01 -4.40776855e-01 -7.20184684e-01
-2.87510991e-01 4.49052721e-01 -3.93136382e-01 -5.72188646e-02
-2.73589045e-02 5.03289461e-01 7.32128974e-03 -8.79985332e-01
4.40467894e-01 7.27479935e-01 7.74664879e-01 -4.83680755e-01
4.15985495e-01 3.61665696e-01 -1.31778315e-01 -7.89741635e-01
-9.52848196e-02 -5.70826113e-01 -1.21772861e+00 -5.07529497e-01
1.01950145e+00 -4.84435916e-01 -9.08798635e-01 1.10498345e+00
-1.13901734e+00 -5.69803536e-01 4.86602932e-02 8.67976964e-01
-6.25931144e-01 8.21133435e-01 -6.99885666e-01 -8.14188182e-01
-6.19172037e-01 -1.11493790e+00 1.16429794e+00 4.22469229e-01
-5.24922647e-02 -5.35940111e-01 -9.69771892e-02 -6.59529492e-02
9.36779380e-02 4.82205898e-01 8.32692564e-01 -7.21040547e-01
-5.28148651e-01 -2.46163666e-01 -3.58183742e-01 1.13168493e-01
9.42356214e-02 -1.21541016e-01 -7.41860926e-01 -2.67969161e-01
-1.08517833e-01 -2.14493364e-01 1.04554999e+00 4.32173789e-01
1.02286375e+00 -7.14587942e-02 -3.34024906e-01 7.25959837e-01
9.00650442e-01 4.75559354e-01 5.29490948e-01 1.02447316e-01
1.14043474e+00 4.15426105e-01 4.28495347e-01 4.02619839e-01
2.64792919e-01 6.36702299e-01 3.99570316e-01 2.17541710e-01
-6.19597018e-01 -2.81919390e-01 3.72645408e-01 9.70294893e-01
-9.25952137e-01 -9.91472453e-02 -8.55182111e-01 2.66646951e-01
-2.04088998e+00 -6.64269388e-01 -1.29749626e-01 1.58386874e+00
4.39926654e-01 -2.29562402e-01 3.21046680e-01 1.20045297e-01
7.44464934e-01 2.52174139e-01 -7.17530787e-01 2.37387672e-01
-2.34480500e-01 3.31610411e-01 1.55028388e-01 2.49632820e-01
-1.26170158e+00 1.22057331e+00 5.58005142e+00 7.14892328e-01
-1.40696573e+00 1.08072959e-01 -7.83778280e-02 -2.74670154e-01
4.46397543e-01 -4.29647505e-01 -6.46830320e-01 3.85564238e-01
2.98888385e-01 5.67687690e-01 4.44975883e-01 5.63468814e-01
2.44479284e-01 1.11793749e-01 -1.09619784e+00 1.18660891e+00
1.86167374e-01 -6.09987557e-01 3.07837844e-01 3.66189629e-02
2.81680107e-01 1.67967811e-01 -2.06862912e-01 3.56621712e-01
-1.00357629e-01 -1.05865359e+00 6.04974866e-01 6.37335718e-01
8.23655725e-01 -3.28195721e-01 7.82976031e-01 3.49359840e-01
-1.53612685e+00 -3.35076988e-01 4.44092639e-02 8.03874582e-02
4.11350340e-01 -4.72455353e-01 -6.37858927e-01 8.36412668e-01
7.32460439e-01 7.14787602e-01 -3.89253736e-01 1.15944231e+00
-5.82055569e-01 6.05053723e-01 -6.05496407e-01 -3.47975083e-02
5.05170405e-01 1.38584292e-02 6.61828995e-01 1.31667078e+00
2.61638045e-01 5.45287609e-01 1.78084284e-01 4.60110724e-01
2.76573718e-01 1.28854960e-01 -8.55048299e-02 -1.03603266e-01
-1.09346814e-01 9.76318359e-01 -1.02395177e+00 -2.71309912e-01
-3.68527710e-01 1.20658684e+00 -1.18065223e-01 6.20537937e-01
-5.85733771e-01 -3.35073739e-01 2.43146494e-01 -4.32891965e-01
4.81220275e-01 -4.99350101e-01 -1.39207765e-01 -1.27336299e+00
2.94114679e-01 -7.31183231e-01 7.14826226e-01 -2.58119255e-01
-1.20984387e+00 6.30501747e-01 2.76137944e-02 -1.10209477e+00
-4.45845127e-01 -8.61745656e-01 -5.95965207e-01 8.17775428e-01
-1.28267050e+00 -1.49932289e+00 -6.78384602e-01 1.00387943e+00
7.35634685e-01 -7.98986033e-02 6.13253891e-01 4.12184685e-01
-8.30427527e-01 6.80889368e-01 -3.96830291e-01 7.71652758e-01
5.09849668e-01 -8.57509077e-01 5.55257857e-01 7.90868044e-01
8.92614424e-02 6.35221004e-01 -1.31446272e-01 -1.02485061e+00
-1.40164423e+00 -7.52108574e-01 5.12501359e-01 -2.15432301e-01
5.73793709e-01 -1.55482262e-01 -7.93579221e-01 7.11814165e-01
-1.59336939e-01 1.35952145e-01 2.63955742e-01 -3.50636363e-01
-2.77898043e-01 5.43567657e-01 -7.96280503e-01 5.65692663e-01
1.88152516e+00 -2.11695045e-01 -8.25221300e-01 -5.64580634e-02
3.52076441e-01 -6.99782848e-01 -7.21577764e-01 6.60555124e-01
1.20704544e+00 -4.96299237e-01 9.09732819e-01 -9.96661723e-01
1.43554851e-01 -5.08212000e-02 -2.34084859e-01 -8.29050243e-01
1.79187022e-02 -4.16765600e-01 -4.23334926e-01 8.90538454e-01
-3.10321897e-01 -2.53429890e-01 1.07758129e+00 4.17105883e-01
-1.98612168e-01 -7.35910356e-01 -1.22352076e+00 -7.75139987e-01
-1.24251597e-01 -4.83020067e-01 5.63834190e-01 5.28662205e-01
7.55756944e-02 -1.31435275e-01 -3.58145773e-01 -5.14398701e-02
5.00729918e-01 5.35304844e-02 6.72165275e-01 -1.13397706e+00
-8.10321793e-02 -7.69409955e-01 -5.71259558e-01 -1.42113888e+00
6.17241599e-02 -9.76300001e-01 1.97495893e-01 -1.67859077e+00
1.41509578e-01 -2.12144971e-01 -3.91925842e-01 7.58779407e-01
-1.15601659e-01 1.72820300e-01 6.20100200e-01 4.08863872e-01
-3.44251990e-01 6.43558443e-01 1.60349870e+00 -2.16806844e-01
-5.20423353e-01 1.43943615e-02 1.26834884e-01 8.92044485e-01
3.45011920e-01 -8.78804028e-02 -1.76449463e-01 -3.92380297e-01
-6.10518873e-01 1.90368533e-01 5.18250406e-01 -8.01981032e-01
3.19124818e-01 -5.48022538e-02 4.67049599e-01 -1.01952326e+00
1.83757886e-01 -7.40732491e-01 -2.42695615e-01 6.59858465e-01
-1.02033794e-01 -9.41278487e-02 -1.82954818e-01 5.42436242e-01
-1.72485784e-01 8.86161551e-02 3.22533816e-01 -2.10368559e-01
-1.09658277e+00 7.55733311e-01 -1.35702014e-01 -9.83527526e-02
7.25390017e-01 -3.67446870e-01 -1.30210713e-01 1.98949557e-02
-1.27468514e+00 3.35159361e-01 -2.57153004e-01 8.34124148e-01
9.68689024e-01 -1.23352098e+00 -6.33047521e-01 2.75616199e-01
-1.15827426e-01 1.97765931e-01 5.59683561e-01 1.26232684e+00
-5.41239917e-01 5.12347698e-01 -6.56700790e-01 -6.97434127e-01
-1.23413956e+00 3.02275807e-01 1.51606962e-01 -1.96044892e-01
-1.41000652e+00 9.45811868e-01 -3.13315098e-03 -2.00183973e-01
8.73632014e-01 -7.85265744e-01 -3.94235611e-01 2.82607079e-01
5.57930470e-01 4.65467989e-01 -4.84896228e-02 -8.74501824e-01
-5.73732734e-01 8.74609828e-01 2.66351947e-03 -2.47044876e-01
1.36937988e+00 1.07928291e-01 1.33195028e-01 3.20795327e-01
9.29108083e-01 -5.93948960e-01 -1.63748527e+00 -2.78552800e-01
8.51021558e-02 -3.98553342e-01 -3.18487644e-01 -1.03335452e+00
-1.45704401e+00 1.17960513e+00 8.76081407e-01 -6.02459490e-01
9.93693709e-01 -1.60315499e-01 1.17979288e+00 3.08140039e-01
5.59722245e-01 -9.22744453e-01 1.29518956e-01 6.69912636e-01
1.11609387e+00 -9.31194484e-01 -3.76172006e-01 -1.74860179e-01
-5.83401978e-01 1.39972448e+00 5.27792692e-01 -1.49116367e-01
7.21735001e-01 2.07304224e-01 2.20546454e-01 -6.65730685e-02
7.10262880e-02 -4.75983173e-01 2.76292175e-01 6.77015603e-01
1.09020062e-01 1.21675357e-01 -7.52531826e-01 1.06433105e+00
2.09544860e-02 3.94689530e-01 5.69694005e-02 9.49608028e-01
-6.33181483e-02 -1.06906402e+00 -3.51840556e-01 3.17267597e-01
-1.70536548e-01 1.53094009e-01 -5.32192290e-01 1.10465288e+00
1.20373093e-01 5.61769843e-01 -2.20578626e-01 -8.43599796e-01
3.76770079e-01 6.34589612e-01 8.59927952e-01 -2.55423397e-01
-7.01673150e-01 4.07293946e-01 -3.22538793e-01 -6.07855201e-01
-5.42190135e-01 -7.91846335e-01 -1.79324400e+00 2.54132271e-01
-2.02376068e-01 -4.29656595e-01 4.92131352e-01 1.25531435e+00
-9.96695235e-02 7.26497710e-01 -2.26517364e-01 -1.31218541e+00
-6.64847016e-01 -1.38338196e+00 -7.19247520e-01 3.24202061e-01
2.19603091e-01 -1.10999036e+00 7.55567476e-02 -8.92412961e-02] | [6.765812873840332, -0.2401949167251587] |
cb071fc9-d05b-4728-8b10-f7b087e851b5 | residual-feature-pyramid-network-for | 2306.17200 | null | https://arxiv.org/abs/2306.17200v1 | https://arxiv.org/pdf/2306.17200v1.pdf | Residual Feature Pyramid Network for Enhancement of Vascular Patterns | The accuracy of finger vein recognition systems gets degraded due to low and uneven contrast between veins and surroundings, often resulting in poor detection of vein patterns. We propose a finger-vein enhancement technique, ResFPN (Residual Feature Pyramid Network), as a generic preprocessing method agnostic to the recognition pipeline. A bottom-up pyramidal architecture using the novel Structure Detection block (SDBlock) facilitates extraction of veins of varied widths. Using a feature aggregation module (FAM), we combine these vein-structures, and train the proposed ResFPN for detection of veins across scales. With enhanced presentations, our experiments indicate a reduction upto 5% in the average recognition errors for commonly used recognition pipeline over two publicly available datasets. These improvements are persistent even in cross-dataset scenario where the dataset used to train the ResFPN is different from the one used for recognition. | ['Sebastien Marcel', 'Ketan Kotwal'] | 2023-06-29 | null | null | null | null | ['finger-vein-recognition'] | ['computer-vision'] | [ 4.75778282e-01 -1.72343060e-01 2.60604382e-01 -1.29228726e-01
-1.36698157e-01 -1.05755103e+00 6.17354453e-01 -1.12384800e-02
-1.95559472e-01 4.77108300e-01 2.02317953e-01 1.14188597e-01
5.37206940e-02 -9.56319153e-01 -1.74782917e-01 -4.47147250e-01
6.79598376e-02 1.24757975e-01 6.06293499e-01 -3.33047770e-02
6.04192555e-01 1.28532028e+00 -1.27633309e+00 5.79218268e-01
4.75605130e-01 1.08486438e+00 -3.25967491e-01 9.73830223e-01
-4.16396379e-01 4.33221668e-01 -5.46834469e-01 -7.22310185e-01
8.18011642e-01 -1.43171951e-01 -3.82188290e-01 1.79508731e-01
9.45974946e-01 -3.79447013e-01 -6.28097355e-01 5.84934652e-01
7.55725086e-01 -4.91887003e-01 6.66945100e-01 -5.31964183e-01
-1.75587863e-01 4.59032878e-02 -7.40452409e-01 4.06619459e-01
2.86788911e-01 4.91863281e-01 6.51438713e-01 -9.14604962e-01
6.30315244e-01 1.12282360e+00 8.26412976e-01 3.93816233e-01
-1.34469593e+00 -3.45040828e-01 -3.50012988e-01 -1.75050378e-01
-1.03675473e+00 -1.81274265e-01 6.06384516e-01 -4.08236444e-01
8.57472539e-01 5.09766877e-01 8.00383806e-01 9.77181971e-01
3.41955692e-01 4.27836061e-01 1.48747778e+00 -1.97156072e-01
-2.17823073e-01 2.38594577e-01 1.67515263e-01 6.97052538e-01
4.89520729e-01 1.63126022e-01 -6.02360547e-01 -1.53910428e-01
1.37850988e+00 1.86119720e-01 -5.10752872e-02 -2.06561565e-01
-7.14101493e-01 3.84609342e-01 5.52628994e-01 3.10421377e-01
-7.99412668e-01 -1.20535687e-01 3.45603824e-01 4.33118612e-01
-2.06552565e-01 5.89819014e-01 -2.65306145e-01 -8.32445323e-02
-1.13558245e+00 2.45228335e-01 9.23991263e-01 6.88038886e-01
5.23877561e-01 -1.41778469e-01 -8.07233274e-01 7.04083383e-01
1.05075672e-01 3.19996268e-01 2.61943907e-01 -5.87538719e-01
2.83124298e-01 1.11168885e+00 -4.62404545e-03 -1.15342128e+00
-4.08722699e-01 -6.89213395e-01 -8.18935871e-01 4.74509418e-01
9.61891413e-01 -8.72151256e-02 -1.10730541e+00 7.67593563e-01
2.45974362e-02 6.57456741e-02 -2.48599201e-01 7.82765448e-01
8.70894670e-01 -9.13002491e-02 -3.59261930e-02 3.77536029e-01
1.58227575e+00 -8.13774109e-01 -6.85344860e-02 -8.52528960e-02
3.94193083e-02 -1.04589856e+00 6.61219835e-01 6.92203104e-01
-9.44280446e-01 -8.21288347e-01 -9.69032526e-01 1.43325165e-01
-5.27404904e-01 2.07962677e-01 6.05213106e-01 1.13735318e+00
-9.64265168e-01 8.06041777e-01 -5.19971550e-01 -5.07535338e-01
8.24131072e-01 6.46703541e-01 -6.45704865e-01 8.81533232e-03
-5.04458487e-01 7.16237009e-01 -1.63692594e-01 4.10231918e-01
-4.22668725e-01 -7.70230114e-01 -3.59283745e-01 -1.16718113e-02
-4.86830734e-02 -4.72176135e-01 3.04994196e-01 -6.56260192e-01
-1.28814054e+00 8.02628160e-01 -2.01654285e-01 -4.72721308e-01
1.07631004e+00 -2.75397062e-01 -1.38164341e-01 4.44156945e-01
-3.06573302e-01 2.15074763e-01 1.33159661e+00 -8.22180569e-01
-4.26463962e-01 -5.75057626e-01 -2.09562019e-01 -3.57298881e-01
-2.21955284e-01 -3.90705131e-02 -3.71798992e-01 -6.93465352e-01
1.39217228e-01 -4.57197309e-01 -3.25761467e-01 1.74788862e-01
-3.13283414e-01 7.99726248e-02 8.21699798e-01 -9.70669627e-01
1.05376911e+00 -1.89365041e+00 -2.41823286e-01 8.93016756e-01
5.12591720e-01 5.34516931e-01 -2.83720046e-01 3.39262128e-01
-6.07323647e-02 -1.92194596e-01 -1.53117329e-01 1.30605742e-01
-3.86095464e-01 -3.10422685e-02 -1.10585302e-01 3.21606129e-01
7.63298392e-01 8.40395510e-01 -4.21793699e-01 -3.90569180e-01
6.05373800e-01 7.83883870e-01 -3.28355461e-01 1.54244736e-01
5.00307620e-01 4.06580836e-01 -4.45600927e-01 8.15610707e-01
1.00260985e+00 -3.41940895e-02 -7.18984811e-04 -5.93285322e-01
-1.90306216e-01 -3.15755725e-01 -1.44415164e+00 1.23546660e+00
-3.45971525e-01 4.43235904e-01 -1.89233363e-01 -3.41480255e-01
1.45360148e+00 1.23274356e-01 4.08638835e-01 -5.45036793e-01
1.48920819e-01 2.71399498e-01 1.36816159e-01 -3.17033112e-01
5.35857417e-02 1.75899327e-01 5.63689053e-01 8.71536136e-03
1.91349208e-01 3.34694028e-01 3.44326377e-01 -9.23255607e-02
1.41203439e+00 1.77509412e-01 7.86259398e-02 -1.76089540e-01
9.06878591e-01 -3.44629318e-01 1.21982247e-01 1.13290453e+00
-4.08482343e-01 9.05896306e-01 7.28563488e-01 -7.22495556e-01
-1.07413316e+00 -1.11905646e+00 -3.55696261e-01 5.43280303e-01
-1.48381352e-01 -3.90798718e-01 -5.38955510e-01 -8.68714631e-01
4.32529479e-01 -3.98980945e-01 -7.59948492e-01 3.37139338e-01
-8.82422447e-01 -7.28367031e-01 7.47125804e-01 6.42912269e-01
6.92790210e-01 -1.17102695e+00 -9.53671932e-01 3.95379037e-01
6.70795262e-01 -1.07791722e+00 -7.11958930e-02 1.50193930e-01
-7.86452293e-01 -1.41099691e+00 -1.05437648e+00 -5.62958002e-01
5.79003811e-01 -3.60251188e-01 8.63888979e-01 1.10118493e-01
-1.11040699e+00 2.50634968e-01 -1.80669799e-01 -2.51126010e-02
5.07835262e-02 -4.84259650e-02 -5.88152945e-01 4.55494285e-01
5.91364741e-01 -5.32974958e-01 -1.05969441e+00 6.95462823e-02
-3.88333201e-01 -4.79175001e-01 9.17633832e-01 8.50685179e-01
2.82543749e-01 -2.43138582e-01 1.33560658e-01 -9.98247266e-01
6.33862913e-01 1.54747635e-01 -6.37707770e-01 2.67418951e-01
-3.91534835e-01 1.14343010e-01 7.06921816e-01 -2.12126374e-01
-6.92729712e-01 2.98044741e-01 2.49487292e-02 -4.27930504e-02
-3.57559711e-01 -1.71587184e-01 1.98858649e-01 -8.13091040e-01
7.39886165e-01 2.73461610e-01 1.30531788e-01 -5.31300128e-01
-3.98294441e-02 7.64788926e-01 4.71632153e-01 -2.30515227e-01
9.12184954e-01 4.57484901e-01 5.08669376e-01 -9.54606831e-01
-1.59241751e-01 -4.51821566e-01 -1.07573271e+00 -1.71978697e-01
5.06303787e-01 -4.40688699e-01 -7.54297733e-01 6.15186393e-01
-8.91382217e-01 4.41541336e-02 -1.93317845e-01 8.59392108e-04
-4.21273857e-02 6.33199453e-01 -5.51477432e-01 -9.83249247e-01
-7.77379870e-01 -1.03686666e+00 7.69342005e-01 7.37919092e-01
-2.87259430e-01 -8.23194683e-01 -2.35365987e-01 2.20105603e-01
8.94620717e-01 6.96340382e-01 5.53145766e-01 -4.29916561e-01
-4.56085205e-01 -5.77313721e-01 -7.11275458e-01 3.96261811e-01
4.50679034e-01 2.71787494e-01 -1.12442756e+00 -1.87572062e-01
-4.77146804e-01 1.89432979e-01 9.81382549e-01 2.74699062e-01
1.12806153e+00 1.99228570e-01 -6.29518256e-02 5.04384995e-01
1.59743047e+00 -1.71210632e-01 9.37406898e-01 3.25581193e-01
6.30882561e-01 7.71452725e-01 9.89397541e-02 6.76360190e-01
-3.08112204e-01 3.40831757e-01 9.70876440e-02 -5.43366432e-01
-6.90146625e-01 2.50010639e-01 -5.05711474e-02 -2.59215117e-01
-6.23488486e-01 4.68574882e-01 -5.78267217e-01 3.12298656e-01
-1.28769457e+00 -7.07326770e-01 -3.86080742e-01 2.15307307e+00
5.34019411e-01 2.77960777e-01 5.75112462e-01 4.79349121e-02
5.56216776e-01 -1.82790346e-02 -5.98042727e-01 -4.85276222e-01
-2.84062296e-01 1.19729161e+00 8.91146064e-01 3.37324679e-01
-1.21216738e+00 9.65234697e-01 6.52369070e+00 3.95904511e-01
-1.45291591e+00 -4.00006175e-01 4.31922615e-01 1.99842691e-01
4.13151085e-01 -2.28103071e-01 -7.50557661e-01 4.03572172e-01
4.70787615e-01 4.11618739e-01 4.95962918e-01 5.07927001e-01
1.06290244e-01 1.48004638e-02 -7.58505225e-01 1.21284270e+00
-8.60776752e-02 -1.22568083e+00 6.25636652e-02 2.63751429e-02
8.33495185e-02 -2.19073087e-01 -1.46674007e-04 -1.53271899e-01
-2.77513474e-01 -1.20105588e+00 2.17772573e-02 6.85718238e-01
7.43634462e-01 -6.50208235e-01 9.19879317e-01 -4.61123765e-01
-1.41903436e+00 -1.06983274e-01 -4.80810195e-01 1.17816649e-01
-2.88137227e-01 3.58249694e-01 -9.84716117e-01 3.47138524e-01
6.10842466e-01 7.13529587e-01 -1.24296784e+00 1.25723565e+00
1.70154963e-02 4.09934282e-01 -3.68249387e-01 7.31857773e-03
8.25978722e-03 -3.59285958e-02 3.60895842e-01 1.71624935e+00
-3.22458707e-02 -4.83247787e-01 -2.95997798e-01 7.86419749e-01
1.73689738e-01 2.76684314e-01 -2.78691620e-01 -5.80477752e-02
1.93954617e-01 1.73975909e+00 -9.89726365e-01 -1.79171830e-01
-5.30499995e-01 1.38328815e+00 9.48819891e-02 2.45209843e-01
-2.24971339e-01 -7.14675367e-01 3.85010391e-01 4.24679577e-01
5.99054575e-01 9.10888314e-02 -6.46659434e-01 -8.83659959e-01
3.00870180e-01 -7.59242356e-01 4.06614453e-01 -6.50091469e-02
-1.65404117e+00 6.36049807e-01 -6.18550599e-01 -9.89715040e-01
1.54255703e-01 -1.19746208e+00 -6.10560656e-01 1.27630210e+00
-1.56551635e+00 -1.35507464e+00 -7.86124468e-01 6.47914946e-01
2.94595331e-01 -5.34263551e-01 8.77341628e-01 2.01984346e-01
-3.31796765e-01 8.77279639e-01 -4.11793321e-01 5.50489426e-01
8.44432712e-01 -1.44849813e+00 4.73520130e-01 9.32501018e-01
-9.82139856e-02 9.43578422e-01 4.70124573e-01 -7.63488948e-01
-1.57297421e+00 -9.82231498e-01 5.28727710e-01 -5.88331997e-01
3.64661455e-01 -3.28337103e-01 -7.09710598e-01 2.92145401e-01
1.79635897e-01 3.26006234e-01 7.07548380e-01 -1.28886759e-01
-6.06446147e-01 -3.96192104e-01 -1.56740451e+00 4.17055637e-01
7.37510741e-01 -1.00633927e-01 -3.41650695e-01 4.96691354e-02
-5.24748921e-01 -1.84128553e-01 -1.19308507e+00 2.30465636e-01
1.29796159e+00 -1.16070187e+00 1.02466190e+00 -4.58838314e-01
1.98981032e-01 -3.24405074e-01 4.92717363e-02 -6.21953607e-01
-5.55730402e-01 -6.22501373e-01 -1.77731767e-01 1.06655443e+00
1.86189324e-01 -7.90836036e-01 1.35207534e+00 4.73906696e-01
4.96497393e-01 -6.45569146e-01 -7.16066480e-01 -3.89496118e-01
-2.44943887e-01 1.71555862e-01 2.53406048e-01 4.19086903e-01
-4.77657765e-01 1.45660266e-02 -3.01052123e-01 7.86150545e-02
8.15002859e-01 1.25417441e-01 8.99433911e-01 -1.34210670e+00
-6.13964140e-01 -5.48639774e-01 -9.70422089e-01 -7.10653663e-01
-7.46754169e-01 -5.88021457e-01 -5.61740756e-01 -1.35991454e+00
9.01018530e-02 -2.34044373e-01 -3.91475648e-01 4.74921972e-01
-8.92508179e-02 8.72980237e-01 3.34066629e-01 3.69909555e-02
1.41390070e-01 -2.78112441e-01 1.45013142e+00 1.08907916e-01
-2.77224302e-01 2.33285781e-02 -7.67377913e-01 4.41405237e-01
6.98961198e-01 -4.04528268e-02 1.73559412e-01 4.99845818e-02
-4.19873804e-01 -4.48567897e-01 4.55203980e-01 -1.24595726e+00
5.13108782e-02 3.27604890e-01 1.27909374e+00 -6.18333399e-01
1.88844040e-01 -8.46626103e-01 -1.14918515e-01 1.09375751e+00
-6.90124854e-02 -2.72540450e-02 2.34304145e-01 1.83662400e-01
6.13656193e-02 2.21401490e-02 9.12306964e-01 -4.11942899e-01
-6.22690916e-01 1.41076565e-01 -2.36494064e-01 -5.03690600e-01
7.99333870e-01 -7.51092136e-01 -2.71161824e-01 1.34793207e-01
-8.98768961e-01 -1.42758951e-01 2.18448281e-01 1.22913018e-01
7.87415802e-01 -8.30607593e-01 -1.02558815e+00 7.52278149e-01
8.69170576e-03 -4.78962123e-01 2.64824271e-01 8.82495463e-01
-1.12220252e+00 2.73857474e-01 -7.12612092e-01 -4.58119184e-01
-1.46609092e+00 7.32884556e-02 5.78138828e-01 -3.95219743e-01
-8.76200974e-01 9.56584692e-01 -3.15501124e-01 -7.56531209e-02
2.28473932e-01 -1.40176311e-01 -4.79799598e-01 -7.82227442e-02
8.22463810e-01 5.29821038e-01 2.89298058e-01 -3.10148329e-01
-4.87221122e-01 8.67142260e-01 -3.90303344e-01 3.59607518e-01
1.37860537e+00 3.44944775e-01 -1.63441435e-01 -1.80179492e-01
7.26797938e-01 4.02554005e-01 -1.20666814e+00 4.22343705e-03
-2.28956975e-02 -6.87814891e-01 -2.72214741e-01 -1.05175579e+00
-1.09695482e+00 7.59346366e-01 1.38792467e+00 2.09730372e-01
1.03813088e+00 -4.08330113e-01 6.18193686e-01 3.41125689e-02
1.07778646e-01 -6.50283039e-01 -1.33105025e-01 1.61044151e-01
8.70765090e-01 -9.50302303e-01 1.08981291e-02 -7.26323724e-01
-4.55837041e-01 1.82110679e+00 6.90267086e-01 -9.68415082e-01
5.50324500e-01 4.40018475e-01 1.76370069e-01 -1.74218565e-01
-1.38808116e-02 -2.19520524e-01 4.85567212e-01 8.75605643e-01
6.44464612e-01 -1.06276728e-01 -4.60937589e-01 2.64843941e-01
-1.66208491e-01 -2.65270192e-02 3.14059436e-01 9.18847620e-01
-2.36347005e-01 -1.48869836e+00 -3.75851303e-01 9.10629630e-01
-4.90343481e-01 -1.84633899e-02 -7.98174620e-01 6.18364990e-01
2.12479606e-01 6.68461382e-01 1.50378998e-02 -3.25233370e-01
6.10454977e-01 -1.31972432e-01 9.55210745e-01 -4.62926567e-01
-1.35398114e+00 3.16584915e-01 -1.90427646e-01 -7.29754806e-01
-9.48500559e-02 -5.88011026e-01 -7.60171294e-01 -5.44205494e-03
1.05897345e-01 -3.88200074e-01 3.66566062e-01 6.95908487e-01
4.16563898e-01 5.26900411e-01 4.99923825e-01 -4.72919494e-01
-4.05808598e-01 -1.02899539e+00 -7.92887747e-01 4.46835309e-01
3.69806886e-01 -5.83887458e-01 5.80807813e-02 -5.20125329e-02] | [13.047359466552734, 1.0282206535339355] |
97504218-fc78-4d78-8083-0c8b155aeda6 | bridging-the-gap-providing-post-hoc-symbolic | 2002.01080 | null | https://arxiv.org/abs/2002.01080v4 | https://arxiv.org/pdf/2002.01080v4.pdf | Bridging the Gap: Providing Post-Hoc Symbolic Explanations for Sequential Decision-Making Problems with Inscrutable Representations | As increasingly complex AI systems are introduced into our daily lives, it becomes important for such systems to be capable of explaining the rationale for their decisions and allowing users to contest these decisions. A significant hurdle to allowing for such explanatory dialogue could be the vocabulary mismatch between the user and the AI system. This paper introduces methods for providing contrastive explanations in terms of user-specified concepts for sequential decision-making settings where the system's model of the task may be best represented as an inscrutable model. We do this by building partial symbolic models of a local approximation of the task that can be leveraged to answer the user queries. We test these methods on a popular Atari game (Montezuma's Revenge) and variants of Sokoban (a well-known planning benchmark) and report the results of user studies to evaluate whether people find explanations generated in this form useful. | ['Utkarsh Soni', 'Subbarao Kambhampati', 'Sarath Sreedharan', 'Mudit Verma', 'Siddharth Srivastava'] | 2020-02-04 | bridging-the-gap-providing-post-hoc-symbolic-1 | https://openreview.net/forum?id=TETmEkko7e5 | https://openreview.net/pdf?id=TETmEkko7e5 | iclr-2022-4 | ['montezumas-revenge'] | ['playing-games'] | [ 2.06826448e-01 9.54859138e-01 -1.87878788e-01 -6.40352666e-01
-4.74008322e-01 -7.39133418e-01 9.58058953e-01 2.03928441e-01
-1.56144410e-01 7.05628872e-01 4.48878795e-01 -8.66170049e-01
-4.49250489e-02 -4.09788370e-01 -8.58531147e-02 1.39473751e-01
7.03317747e-02 1.18136764e+00 2.96611458e-01 -6.86413407e-01
3.51523340e-01 5.91505840e-02 -1.43541503e+00 4.80881780e-01
9.87239778e-01 3.60147715e-01 1.08756989e-01 9.78504896e-01
-2.99688075e-02 1.01205647e+00 -7.62613654e-01 -5.77505589e-01
2.85797149e-01 -5.36059856e-01 -1.31941831e+00 -4.43140455e-02
1.37497382e-02 -5.29581726e-01 7.73717789e-03 7.76986957e-01
-6.08347170e-02 2.13564679e-01 7.19018698e-01 -1.60944235e+00
-2.44379386e-01 1.11149025e+00 1.56530827e-01 1.18813209e-01
8.79993081e-01 5.44383943e-01 1.14120984e+00 -1.51952028e-01
7.68541098e-01 1.36313140e+00 1.21282734e-01 7.55103648e-01
-1.42775822e+00 -3.21445584e-01 2.59482920e-01 2.16894060e-01
-1.15527606e+00 -5.44831753e-01 1.99106619e-01 -4.90085155e-01
1.21545267e+00 8.93830836e-01 5.52912414e-01 8.52131665e-01
2.61138797e-01 5.65899014e-01 9.66193914e-01 -5.34054875e-01
3.72333556e-01 4.08109486e-01 2.13580936e-01 5.02018034e-01
3.18043500e-01 1.37678072e-01 -7.61907220e-01 -6.39236271e-01
7.41248727e-01 -3.45812112e-01 -2.56923269e-02 -6.25356376e-01
-1.00021982e+00 9.43481445e-01 2.03642190e-01 4.56359267e-01
-4.71764952e-01 2.53096581e-01 1.14100352e-01 4.38261479e-01
-5.36533184e-02 1.33058417e+00 -4.67130721e-01 -5.63187659e-01
-4.72479612e-01 9.98832941e-01 1.35580170e+00 1.08900142e+00
2.37382770e-01 -2.48657390e-01 -1.93954840e-01 2.26260304e-01
3.90266448e-01 -9.66041931e-04 3.88099164e-01 -1.44639254e+00
2.37314835e-01 7.32421696e-01 6.88132584e-01 -7.83826470e-01
-2.83552557e-01 -1.25413850e-01 9.31952223e-02 3.88233215e-01
6.61478579e-01 -6.64317161e-02 -5.53738415e-01 1.44667411e+00
2.17981353e-01 -3.91214430e-01 4.04912204e-01 1.10434818e+00
5.69922149e-01 3.53670925e-01 2.91263133e-01 -5.00898063e-02
1.44255567e+00 -6.97625875e-01 -5.34264684e-01 -8.38490009e-01
9.46798861e-01 -3.48362803e-01 1.25735664e+00 5.10809958e-01
-1.06274426e+00 -2.61974007e-01 -1.07133281e+00 -1.98349625e-01
7.21954703e-02 -2.03176484e-01 1.04635322e+00 4.16302532e-01
-7.15256393e-01 3.40348423e-01 -7.76462317e-01 -5.73123276e-01
-1.53516889e-01 4.21268672e-01 -9.17709470e-02 1.08587870e-03
-1.13746929e+00 1.20158088e+00 4.76322204e-01 -1.28720418e-01
-7.85253525e-01 -7.45145455e-02 -8.54800463e-01 3.30428392e-01
8.52383137e-01 -7.57802606e-01 2.11055827e+00 -9.99574006e-01
-1.28669190e+00 5.16096234e-01 2.55779121e-02 -5.08352578e-01
6.86119437e-01 -1.47182688e-01 -4.84145544e-02 -2.58956671e-01
4.06796753e-01 7.05731809e-01 3.16171974e-01 -1.04292464e+00
-6.61260843e-01 -2.74965614e-01 9.47218180e-01 5.23906887e-01
5.19456863e-01 1.89678539e-02 4.67926590e-03 -1.97847769e-01
5.28579168e-02 -1.06702483e+00 -6.04424834e-01 -2.10749954e-01
-5.33762515e-01 -2.24176183e-01 1.99205622e-01 -5.13584554e-01
1.16706944e+00 -1.75428081e+00 1.90470010e-01 3.06838036e-01
5.11609495e-01 -3.51170748e-02 8.13958943e-02 5.64755201e-01
4.14192863e-02 5.08674085e-01 9.05006081e-02 -1.24294698e-01
4.65477794e-01 2.50129491e-01 -3.64292681e-01 -1.08046480e-01
-1.29407376e-01 8.87773514e-01 -7.81640708e-01 -4.34732318e-01
2.50750948e-02 -1.80684060e-01 -7.72387326e-01 5.20768821e-01
-8.28719616e-01 4.28667426e-01 -8.39253008e-01 2.61663377e-01
-1.80927396e-01 -2.32654661e-01 4.82528508e-01 3.98615479e-01
-5.88620789e-02 6.38044894e-01 -1.07035923e+00 1.45243824e+00
-8.97258222e-02 4.61960018e-01 -1.62628338e-01 -1.77842438e-01
3.86511117e-01 3.61511976e-01 -2.89642453e-01 -4.97787893e-01
1.58812121e-01 2.06179649e-01 5.73875368e-01 -6.35865450e-01
5.66536188e-01 -2.66860455e-01 -3.45976442e-01 7.05177069e-01
-4.25291449e-01 -6.14001751e-01 1.62627533e-01 5.52747428e-01
1.16388893e+00 3.62139568e-02 9.19698179e-01 -3.92633975e-01
2.11750239e-01 8.01368117e-01 1.62908435e-01 1.21159720e+00
3.10797375e-02 2.36318484e-01 8.72792363e-01 -9.33376014e-01
-8.91209543e-01 -4.36562896e-01 3.17790240e-01 1.06675661e+00
8.53196159e-02 -8.71288002e-01 -7.86109984e-01 -7.23284721e-01
-1.83905631e-01 1.75302899e+00 -5.88412464e-01 -1.35713443e-01
-2.63749629e-01 2.02481702e-01 4.45756674e-01 3.61971915e-01
3.73689085e-01 -9.39552546e-01 -1.49645257e+00 4.37702760e-02
-4.49501544e-01 -7.97142923e-01 -5.04554451e-01 1.43041000e-01
-5.93117177e-01 -1.19514501e+00 1.67125687e-01 1.51879758e-01
4.55500633e-01 1.31136298e-01 1.28417873e+00 6.09952807e-01
7.45608268e-05 4.68653679e-01 -4.17034239e-01 -6.07462525e-01
-7.95294285e-01 3.61980200e-02 -5.61695993e-02 -7.41860092e-01
5.47778368e-01 -3.26779008e-01 -2.45911434e-01 5.07762313e-01
-8.28126371e-01 7.43555248e-01 2.01068327e-01 6.11057043e-01
-4.36424725e-02 -1.82351381e-01 -7.03288615e-02 -1.35498750e+00
1.20710146e+00 -4.44622934e-01 -4.85321671e-01 4.09904718e-01
-7.94425368e-01 5.49396813e-01 4.33333427e-01 -3.30660462e-01
-9.98656154e-01 -3.90455052e-02 3.12922359e-01 1.83524460e-01
-3.02119046e-01 7.88178384e-01 -8.01396444e-02 2.37034157e-01
1.26381481e+00 -1.96181029e-01 -5.75246289e-02 4.93422449e-02
4.43127781e-01 5.26783764e-01 4.77529585e-01 -9.66104209e-01
7.54520774e-01 -1.67337894e-01 -4.67384338e-01 -4.94621873e-01
-4.37631637e-01 -9.32816416e-02 -2.68082768e-01 -2.20597386e-02
6.78606570e-01 -3.83063674e-01 -1.03262401e+00 -3.53556424e-01
-1.32260287e+00 -6.69630051e-01 -1.78932101e-01 1.59077555e-01
-8.24157655e-01 -6.74843490e-02 -6.57292232e-02 -1.05550575e+00
2.00923875e-01 -1.35851383e+00 7.25175977e-01 2.86241710e-01
-1.31172740e+00 -6.22006059e-01 4.32501212e-02 5.41344881e-01
6.42308831e-01 2.90471725e-02 1.40057206e+00 -1.46502519e+00
-6.01455212e-01 -2.05619931e-01 1.39638279e-02 -4.99585658e-01
-1.66855052e-01 -2.47679025e-01 -7.62445629e-01 2.01999366e-01
6.71075061e-02 -3.53817195e-01 -1.55501172e-01 4.20106202e-02
6.13728642e-01 -7.26814687e-01 -4.82642293e-01 -1.82283416e-01
7.52262771e-01 5.28908551e-01 6.51643813e-01 3.20038319e-01
6.30302057e-02 7.88741887e-01 9.04452980e-01 1.98395997e-01
6.99550688e-01 1.08040953e+00 2.14139923e-01 3.27853560e-01
4.67210025e-01 -3.83331478e-01 7.59533746e-03 -2.70812541e-01
-2.41463631e-01 -3.24279904e-01 -1.35513318e+00 2.30512887e-01
-2.25577521e+00 -8.32446158e-01 1.14451483e-01 2.14297748e+00
5.66115618e-01 4.61985320e-01 1.83459476e-01 -3.22017759e-01
2.61767328e-01 -1.43525422e-01 -7.75025845e-01 -6.97663724e-01
5.71284592e-01 -1.04880780e-01 -7.87401572e-04 1.07325935e+00
-5.49760342e-01 9.41030502e-01 6.76365042e+00 1.90211475e-01
-6.36649787e-01 -7.13523105e-02 7.66166866e-01 1.69354677e-02
-5.79542160e-01 4.61151719e-01 -2.64367163e-01 -8.56931731e-02
1.12555504e+00 -4.90446657e-01 5.45349479e-01 9.63733435e-01
1.85235098e-01 -5.75171769e-01 -1.79141867e+00 5.66058457e-01
-1.82242021e-01 -1.17699277e+00 -7.44272470e-02 1.11710183e-01
2.29360666e-02 -5.18688858e-01 -1.85252488e-01 6.56734109e-01
9.23520863e-01 -1.52894306e+00 9.53104615e-01 3.72778922e-01
4.67093259e-01 -5.09206057e-01 6.96024179e-01 1.03262746e+00
-5.11498034e-01 -2.03424603e-01 -2.74950336e-03 -6.83930814e-01
-7.06219077e-02 -5.21725595e-01 -1.47588444e+00 1.46857828e-01
1.84028879e-01 -1.76238850e-01 -4.94517654e-01 5.07797062e-01
-2.80073702e-01 3.48262966e-01 -3.15032274e-01 -3.97169858e-01
1.75505579e-01 1.65782915e-03 6.76277339e-01 6.02102876e-01
7.55155981e-02 9.54060555e-01 3.87228966e-01 1.10515141e+00
5.06154954e-01 -2.22833693e-01 -7.48188794e-01 -4.70164448e-01
6.07469022e-01 7.96154976e-01 -6.75848305e-01 -4.09984380e-01
-5.47387563e-02 8.01405668e-01 1.07924186e-01 3.40570122e-01
-6.44316852e-01 8.54756013e-02 7.58727610e-01 4.19624150e-01
-4.33874667e-01 -1.92253008e-01 -2.97751039e-01 -1.21977365e+00
-1.62380278e-01 -1.70335472e+00 4.19463873e-01 -1.43961811e+00
-5.95322132e-01 8.75600815e-01 5.47776461e-01 -5.62347949e-01
-1.11567032e+00 -3.04618478e-01 -6.17133796e-01 1.07619762e+00
-6.16429150e-01 -9.23012018e-01 -2.43726328e-01 1.30877912e-01
7.80812502e-01 -4.58430499e-02 1.15429962e+00 -5.44620335e-01
-1.28981203e-01 1.07446164e-01 -9.55449700e-01 -4.21236634e-01
4.12938267e-01 -1.26523817e+00 7.82360196e-01 6.98553741e-01
2.93575197e-01 1.00913012e+00 1.60966420e+00 -8.38591516e-01
-1.13076687e+00 -1.59016430e-01 1.07006478e+00 -9.70858991e-01
3.97449285e-01 -2.92884141e-01 -8.65328550e-01 1.10082364e+00
2.41572469e-01 -7.56900907e-01 5.49234271e-01 3.24229568e-01
-2.12644979e-01 5.49133897e-01 -1.12184930e+00 1.08020854e+00
1.08579803e+00 -4.42275703e-01 -1.08573472e+00 3.72298360e-01
7.28626311e-01 -8.54304373e-01 -7.88486451e-02 -1.93523496e-01
6.05962873e-01 -1.03096330e+00 6.26948655e-01 -1.26545513e+00
6.27727389e-01 -4.23595130e-01 -1.65747300e-01 -1.36831987e+00
-1.80684939e-01 -9.23923254e-01 1.97514758e-01 6.92996442e-01
6.30152524e-01 -5.07849216e-01 6.75347149e-01 2.08096051e+00
1.38980091e-01 -3.42856824e-01 -7.39134252e-01 -1.76650837e-01
-3.10236901e-01 -6.31654739e-01 7.84264982e-01 6.99033618e-01
7.19956934e-01 5.21425426e-01 -4.59698252e-02 1.96171358e-01
1.57913476e-01 -5.99166416e-02 1.01284933e+00 -1.34457350e+00
-6.56172872e-01 -2.95126408e-01 -1.25319645e-01 -1.06431973e+00
-8.57791603e-02 -6.03826463e-01 2.75581747e-01 -1.57440960e+00
7.74191171e-02 -4.74404305e-01 4.16749805e-01 6.89738691e-01
8.81883726e-02 -5.34199297e-01 4.13617909e-01 5.22176266e-01
-5.63289940e-01 9.63826329e-02 8.13086212e-01 1.48984808e-02
-3.94453287e-01 1.78981781e-01 -1.25311184e+00 8.84119868e-01
5.05826414e-01 -5.13898313e-01 -8.45541120e-01 -3.75107020e-01
5.81590116e-01 6.74974978e-01 3.55798155e-01 -8.30801725e-01
6.81194067e-01 -8.25751662e-01 -1.19255193e-01 1.72443271e-01
3.45991731e-01 -8.88350427e-01 6.65060580e-01 6.17413759e-01
-1.05910480e+00 3.27202439e-01 1.82404324e-01 3.11237305e-01
9.13953707e-02 -3.67147326e-01 2.16689229e-01 -2.23456711e-01
-6.73413217e-01 -2.78527796e-01 -8.87009025e-01 -1.62154779e-01
1.02968383e+00 -4.75581229e-01 -3.11007828e-01 -9.75442886e-01
-8.26994359e-01 6.24936521e-01 6.41780615e-01 2.44638458e-01
5.60249209e-01 -8.68188739e-01 -5.28354764e-01 4.33380939e-02
4.24136907e-01 -1.68065682e-01 -2.00130686e-01 4.54303354e-01
-6.35642588e-01 7.22092927e-01 -2.91339338e-01 -1.37902666e-02
-1.30228901e+00 1.73722655e-01 6.05560958e-01 -2.29138404e-01
-5.22731423e-01 6.84296250e-01 4.76692617e-01 -4.80776817e-01
1.31661475e-01 -4.26208496e-01 -3.62352550e-01 -4.80475128e-01
6.12980127e-01 -1.64700136e-01 -2.78515309e-01 -1.72287673e-01
-4.06249911e-01 -4.68567997e-01 -2.43539825e-01 -6.78864539e-01
1.10633898e+00 -1.25162736e-01 1.41518965e-01 4.53326225e-01
-1.12160057e-01 -1.04075044e-01 -9.29289520e-01 -1.48807257e-01
1.59493148e-01 -7.01782286e-01 -3.95545959e-01 -1.29908657e+00
-8.60173926e-02 5.86256027e-01 -9.17610750e-02 6.13795936e-01
4.93585855e-01 1.87684655e-01 5.56681380e-02 9.81464028e-01
8.83779526e-01 -6.67448759e-01 -1.23583011e-01 3.99769366e-01
1.24068499e+00 -1.03265011e+00 1.12170883e-01 -4.00975823e-01
-1.20973158e+00 1.21566033e+00 9.37268913e-01 2.26854056e-01
-4.26638797e-02 9.98524502e-02 2.19689384e-01 -5.73826551e-01
-1.29162133e+00 2.21065432e-01 9.32363123e-02 6.05490029e-01
5.14503300e-01 3.98212641e-01 -2.72382557e-01 1.11645710e+00
-5.89742541e-01 -3.11754830e-02 1.00095057e+00 9.84549105e-01
-4.78776276e-01 -1.22928858e+00 -2.69684523e-01 4.18606758e-01
2.01514084e-02 2.05355138e-02 -1.19022536e+00 1.08258545e+00
-4.39988285e-01 1.18682802e+00 -2.59392172e-01 -2.93925494e-01
4.41231370e-01 3.91338378e-01 2.49148145e-01 -1.14335763e+00
-8.26753080e-01 -2.77738452e-01 8.50806117e-01 -9.00516629e-01
-4.36183996e-02 -6.20699465e-01 -1.30413663e+00 -4.10320580e-01
-2.68744290e-01 6.60449326e-01 2.98635036e-01 1.32314086e+00
1.77621692e-01 1.97338402e-01 -3.19284588e-01 -4.43117321e-01
-7.81240463e-01 -8.04024816e-01 -2.08948046e-01 3.19401592e-01
-8.03584382e-02 -6.01172566e-01 -2.20073238e-02 -1.25704482e-01] | [9.292256355285645, 6.813237190246582] |
eaead7db-0f1a-423e-833c-067ae9537bb0 | privacy-preserved-neural-graph-similarity | 2210.11730 | null | https://arxiv.org/abs/2210.11730v1 | https://arxiv.org/pdf/2210.11730v1.pdf | Privacy-Preserved Neural Graph Similarity Learning | To develop effective and efficient graph similarity learning (GSL) models, a series of data-driven neural algorithms have been proposed in recent years. Although GSL models are frequently deployed in privacy-sensitive scenarios, the user privacy protection of neural GSL models has not drawn much attention. To comprehensively understand the privacy protection issues, we first introduce the concept of attackable representation to systematically characterize the privacy attacks that each model can face. Inspired by the qualitative results, we propose a novel Privacy-Preserving neural Graph Matching network model, named PPGM, for graph similarity learning. To prevent reconstruction attacks, the proposed model does not communicate node-level representations between devices. Instead, we learn multi-perspective graph representations based on learnable context vectors. To alleviate the attacks to graph properties, the obfuscated features that contain information from both graphs are communicated. In this way, the private properties of each graph can be difficult to infer. Based on the node-graph matching techniques while calculating the obfuscated features, PPGM can also be effective in similarity measuring. To quantitatively evaluate the privacy-preserving ability of neural GSL models, we further propose an evaluation protocol via training supervised black-box attack models. Extensive experiments on widely-used benchmarks show the effectiveness and strong privacy-protection ability of the proposed model PPGM. The code is available at: https://github.com/RUCAIBox/PPGM. | ['Ji-Rong Wen', 'Yaliang Li', 'Wayne Xin Zhao', 'Yupeng Hou'] | 2022-10-21 | null | null | null | null | ['graph-similarity', 'graph-matching'] | ['graphs', 'graphs'] | [ 3.43114763e-01 1.75644308e-01 -3.76901209e-01 -3.48873198e-01
-5.42922616e-01 -7.72898197e-01 3.40797752e-01 5.02238214e-01
7.81173483e-02 4.52341348e-01 2.71223076e-02 -4.99452025e-01
-2.94895947e-01 -1.07002044e+00 -8.62067580e-01 -7.60437787e-01
-2.48637125e-01 -3.14122945e-01 -1.08267114e-01 -1.00715291e-02
9.71714929e-02 5.52120149e-01 -9.95298624e-01 1.36985973e-01
6.97423100e-01 1.07087481e+00 -3.54350179e-01 2.05815479e-01
3.75662930e-02 5.91340542e-01 -3.75787050e-01 -1.00191712e+00
6.23239875e-01 -2.41503090e-01 -4.61080194e-01 -3.77191961e-01
3.40851039e-01 -3.34076524e-01 -9.36321557e-01 1.60622871e+00
4.73079443e-01 -1.34539813e-01 1.75361916e-01 -1.73760629e+00
-8.95217001e-01 7.07895160e-01 -4.35266316e-01 -6.10414855e-02
4.94956672e-01 3.64804380e-02 1.16118801e+00 -2.82760918e-01
3.52050543e-01 1.09922397e+00 7.15159953e-01 6.17970943e-01
-1.10376561e+00 -1.27048540e+00 3.40807647e-01 2.35809579e-01
-1.46944201e+00 -2.25429788e-01 1.20784569e+00 2.40157936e-02
4.19353813e-01 6.69499755e-01 4.73479658e-01 1.35490084e+00
2.44541213e-01 8.72924507e-01 9.18993950e-01 6.92302659e-02
2.17655689e-01 3.26528996e-01 2.05378011e-01 7.80043662e-01
9.09068763e-01 3.80894750e-01 -3.16716075e-01 -7.09330618e-01
4.31159109e-01 4.80183065e-01 -5.96354723e-01 -8.08229148e-01
-4.47700262e-01 7.69553661e-01 6.74779713e-01 7.26727918e-02
1.71942487e-01 2.56638855e-01 6.68399632e-01 4.86692041e-01
2.05507040e-01 2.49332726e-01 -2.09531531e-01 5.24955750e-01
-3.43769282e-01 4.54361811e-02 9.18647528e-01 1.23954916e+00
7.36285329e-01 -2.01386988e-01 -1.80239186e-01 2.72167832e-01
3.94631028e-01 2.87673593e-01 2.45446682e-01 -5.91181815e-01
6.41356707e-01 8.07291627e-01 -4.28232014e-01 -1.88995719e+00
1.08655222e-01 -2.97904611e-01 -1.30518281e+00 -3.32176656e-01
1.81561455e-01 -8.84095952e-02 -4.20204282e-01 1.85133278e+00
2.25018188e-01 5.05322099e-01 2.28147089e-01 5.47485709e-01
7.72490919e-01 4.96789306e-01 2.03401133e-01 -1.50961280e-01
1.12009692e+00 -4.99600500e-01 -5.96771836e-01 -3.18515524e-02
8.95282030e-01 -1.01139031e-01 8.82265031e-01 7.54913986e-02
-6.54581368e-01 -1.37783691e-01 -1.23652327e+00 -6.49991706e-02
-4.82047856e-01 -3.50818455e-01 9.07195926e-01 1.15068412e+00
-9.94227052e-01 8.05601239e-01 -6.97885692e-01 -2.56553352e-01
9.91238415e-01 5.61214089e-01 -6.05686903e-01 -3.05461317e-01
-1.52078450e+00 1.16761319e-01 4.68557894e-01 -3.12335081e-02
-7.32755721e-01 -5.20098150e-01 -9.87318456e-01 4.26991045e-01
3.52302849e-01 -4.38844800e-01 6.77340865e-01 -6.65041685e-01
-1.02951646e+00 8.73298109e-01 3.53795588e-01 -6.64660692e-01
4.41326469e-01 2.82110423e-01 -6.52241170e-01 1.74068883e-01
-2.00365141e-01 4.89915779e-04 8.90585124e-01 -1.26792288e+00
-3.03306252e-01 -7.21219838e-01 2.16792077e-01 -3.26942652e-02
-8.86178195e-01 -8.22334513e-02 -3.28476012e-01 -8.27179015e-01
-1.41915649e-01 -8.60363781e-01 -3.85509908e-01 4.20405358e-01
-7.48422325e-01 1.35481820e-01 1.04804051e+00 -4.94783252e-01
1.30191588e+00 -2.39926291e+00 -3.24336380e-01 6.24847174e-01
6.90083861e-01 5.59645236e-01 -1.34498820e-01 5.78179121e-01
-2.19231639e-02 4.34148639e-01 -4.06678289e-01 -2.58400917e-01
7.73050189e-02 1.40523747e-01 -5.40737927e-01 6.97478831e-01
-2.57863909e-01 1.07366991e+00 -8.82305026e-01 -2.47875988e-01
1.14639923e-01 5.42530060e-01 -4.31544781e-01 3.34564358e-01
6.52053803e-02 1.92463636e-01 -9.49591935e-01 7.03276277e-01
1.20774841e+00 -4.21800315e-01 5.46874523e-01 -1.31471127e-01
6.64582014e-01 -5.20920083e-02 -8.51606011e-01 1.60718763e+00
-1.11122549e-01 2.72665650e-01 7.76654407e-02 -1.08882260e+00
1.14161074e+00 1.43850923e-01 3.28538328e-01 -5.79047382e-01
4.91068035e-01 -8.33800882e-02 -3.17998767e-01 -1.53905153e-01
-4.70389426e-02 2.61809826e-01 -3.50806445e-01 4.86795694e-01
-8.28139633e-02 4.33745384e-01 -5.30657887e-01 3.53388399e-01
1.25960600e+00 -4.71806765e-01 5.70206165e-01 -1.19427055e-01
7.95897484e-01 -6.91001832e-01 6.98790073e-01 8.38841975e-01
-5.65323174e-01 3.95510107e-01 6.85801625e-01 -5.42439342e-01
-4.25368637e-01 -1.03808284e+00 1.89863995e-01 6.53615415e-01
6.08584583e-01 -9.01522040e-01 -8.27472746e-01 -1.18722773e+00
3.28852594e-01 4.85142112e-01 -6.55425966e-01 -9.04412508e-01
-3.38410825e-01 -5.78847587e-01 8.50590706e-01 2.56884694e-01
7.51836360e-01 -7.59732842e-01 1.05026744e-01 -2.64761329e-01
1.80977046e-01 -9.96543050e-01 -5.94254315e-01 -2.45935351e-01
-6.96736515e-01 -1.21547973e+00 -1.45941779e-01 -6.98667765e-01
9.01900887e-01 5.61611235e-01 6.34931862e-01 3.51547539e-01
-6.93456084e-02 4.80860412e-01 -2.36465320e-01 -3.72527510e-01
-3.17319185e-01 6.59810826e-02 -5.92742227e-02 3.30715984e-01
5.81418574e-01 -1.06053209e+00 -5.49920499e-01 1.06475987e-01
-8.93376172e-01 -2.28375927e-01 5.52535653e-01 5.31676888e-01
6.81671739e-01 1.78813502e-01 2.82520384e-01 -1.40494394e+00
9.53045607e-01 -6.33532107e-01 -8.09029996e-01 5.22139490e-01
-8.05730760e-01 8.58205706e-02 1.02303958e+00 -5.18133938e-01
-6.04632676e-01 1.49780184e-01 -9.96184349e-02 -7.88355052e-01
-7.85955638e-02 2.54797518e-01 -9.16808665e-01 -6.92491710e-01
4.17580158e-01 4.20723349e-01 7.48836771e-02 -3.10887575e-01
4.39666480e-01 5.29570520e-01 3.75445813e-01 -5.52879572e-01
1.20507026e+00 3.92272443e-01 3.13309014e-01 -4.04568702e-01
-5.28245509e-01 -6.38767704e-02 -8.26729054e-04 2.23735765e-01
4.89573866e-01 -7.30180204e-01 -1.18393922e+00 4.25890625e-01
-9.44962621e-01 2.30903178e-01 -1.45676777e-01 -3.39360355e-04
-3.32823157e-01 1.01700068e+00 -5.39100707e-01 -7.43445933e-01
-7.68639863e-01 -9.93209660e-01 6.78717911e-01 2.39533961e-01
1.27830893e-01 -1.12880838e+00 -2.58935869e-01 9.52448174e-02
3.02471220e-01 4.84221011e-01 9.75838602e-01 -1.21455944e+00
-8.11318696e-01 -7.07928538e-01 -4.22817528e-01 3.51012200e-01
3.29444855e-01 -5.35405576e-01 -1.08242476e+00 -7.71178603e-01
3.94918025e-01 -1.08161286e-01 6.31934404e-01 -2.10840199e-02
1.97216654e+00 -8.35010231e-01 -5.32804549e-01 1.21232879e+00
1.41289914e+00 3.61501351e-02 5.90804040e-01 2.61502545e-02
1.06277692e+00 4.44443643e-01 1.35579333e-01 4.80706841e-01
2.95009196e-01 1.79464296e-01 7.71838486e-01 1.21522754e-01
4.69606370e-01 -1.06867456e+00 1.31617978e-01 4.09588814e-01
3.38763326e-01 -4.84203666e-01 -4.05381769e-01 -2.41303388e-02
-1.91013706e+00 -7.95626044e-01 1.80664927e-01 2.34193921e+00
3.69962096e-01 -8.76688212e-03 -9.35624391e-02 9.81132612e-02
8.12047422e-01 6.98094785e-01 -6.89337432e-01 -3.36517751e-01
1.51434203e-03 -1.08763129e-01 8.74552965e-01 2.76506871e-01
-1.11375606e+00 8.16362083e-01 4.91053629e+00 9.28123474e-01
-9.22288299e-01 -5.19823954e-02 8.20085943e-01 2.66333759e-01
-6.94626331e-01 2.03033090e-01 -5.62721431e-01 6.98899329e-01
6.99915886e-01 -6.69794917e-01 6.26894474e-01 1.09297705e+00
-4.18081850e-01 8.03149104e-01 -1.13825989e+00 1.24933445e+00
6.16955273e-02 -1.42671704e+00 3.00783068e-01 3.10326129e-01
2.87280709e-01 -3.72753620e-01 2.27755442e-01 2.80787528e-01
4.55891103e-01 -9.69706178e-01 -3.57242152e-02 2.72695929e-01
9.02993917e-01 -1.01221204e+00 5.14353991e-01 2.82377630e-01
-1.33987021e+00 -3.70263636e-01 -6.67402864e-01 1.38953283e-01
-3.11732560e-01 2.74465084e-01 -5.49746871e-01 9.09840882e-01
6.63453162e-01 9.59306240e-01 -6.25799417e-01 7.12412834e-01
-4.99759465e-01 6.17755473e-01 -1.63001552e-01 -9.97484773e-02
-6.33325875e-02 -2.86780179e-01 5.20284712e-01 9.27671969e-01
2.70633310e-01 9.28868651e-02 2.24006455e-02 1.00139844e+00
-6.04702711e-01 1.13157190e-01 -1.17058468e+00 -2.89302886e-01
8.34529877e-01 1.17188871e+00 -2.16474667e-01 1.98237970e-01
-4.52062547e-01 1.09674633e+00 5.69754183e-01 3.68422627e-01
-5.44287503e-01 -5.84778845e-01 1.04918289e+00 2.16085315e-02
-4.73296866e-02 4.89577614e-02 -1.60810918e-01 -1.20387745e+00
1.62421107e-01 -8.65128577e-01 8.44485044e-01 -1.43101946e-01
-1.73222840e+00 4.96463090e-01 -1.74225613e-01 -1.28554511e+00
2.39583224e-01 -3.51932973e-01 -8.71859670e-01 7.21332073e-01
-1.34406853e+00 -1.27695489e+00 -2.71894187e-01 9.40658092e-01
-4.72883999e-01 -3.46913785e-01 1.00481308e+00 1.99514627e-01
-6.69497669e-01 1.46576953e+00 3.06540634e-03 5.47087133e-01
3.27809542e-01 -7.82145083e-01 8.34577441e-01 9.04560089e-01
1.86916932e-01 9.72953200e-01 2.64916748e-01 -6.61114752e-01
-1.70674968e+00 -1.36418521e+00 5.35495043e-01 -2.26696551e-01
6.40600085e-01 -8.49183559e-01 -1.29365683e+00 8.71573567e-01
-1.84861004e-01 5.17713904e-01 9.10423815e-01 -2.38668814e-01
-9.02473748e-01 -4.97592956e-01 -1.60950565e+00 6.07528925e-01
1.35864711e+00 -9.28740084e-01 -1.78754330e-02 3.99968535e-01
1.20434690e+00 -1.44595832e-01 -6.38843477e-01 3.56044620e-01
3.74419659e-01 -1.01138127e+00 1.14857566e+00 -7.52285719e-01
-7.83168226e-02 1.97416581e-02 -2.38421440e-01 -8.49856257e-01
-2.08811581e-01 -9.01988149e-01 -5.81760108e-01 1.37762451e+00
9.08853933e-02 -1.13899851e+00 1.28675628e+00 8.95812571e-01
5.34066916e-01 -7.21525431e-01 -8.67170215e-01 -8.18139255e-01
-6.28040954e-02 -2.11216226e-01 1.18781543e+00 1.31076920e+00
-3.98116931e-02 -1.02529258e-01 -6.37631357e-01 5.47114372e-01
9.60975647e-01 9.02180076e-02 6.72138631e-01 -1.28491604e+00
-3.17088276e-01 -2.89576769e-01 -7.91380882e-01 -6.97403789e-01
4.91059750e-01 -1.38426399e+00 -6.72922969e-01 -1.10196412e+00
4.55692261e-02 -3.89411271e-01 -7.15663254e-01 6.09960079e-01
-1.55984983e-01 -2.30009586e-01 1.14945188e-01 1.01082236e-01
-4.39883232e-01 5.70408821e-01 9.71075773e-01 -2.59925187e-01
-1.58747716e-05 3.74770820e-01 -1.34533894e+00 3.14772606e-01
1.05726707e+00 -6.06756210e-01 -9.93997097e-01 -1.72385395e-01
1.49059579e-01 1.79591142e-02 4.29286122e-01 -8.25981975e-01
3.26285571e-01 -8.50288495e-02 4.52642329e-02 -2.79574990e-02
1.03258140e-01 -1.35492969e+00 2.65440375e-01 6.07448280e-01
-5.03808975e-01 -1.25303537e-01 8.99202656e-03 1.11376441e+00
-3.11249718e-02 1.98508859e-01 6.49153352e-01 1.55254276e-02
-4.17949617e-01 1.10456240e+00 2.53376245e-01 -1.38120065e-02
1.12179077e+00 -2.77855456e-01 -4.02828425e-01 -5.59778512e-01
-3.47802788e-01 3.16662699e-01 6.81267202e-01 3.72240573e-01
7.80659258e-01 -1.34114647e+00 -3.19007188e-01 5.11987209e-01
4.55809921e-01 -2.11542875e-01 3.67833048e-01 2.15020373e-01
-2.52990663e-01 1.37897367e-02 -2.45810941e-01 -5.54592796e-02
-1.19294798e+00 1.15463138e+00 2.85544068e-01 -2.42748007e-01
-6.54923975e-01 8.56163144e-01 4.68704939e-01 -5.94377756e-01
5.34168839e-01 1.55932149e-02 2.24428326e-02 -5.29693961e-01
5.78148305e-01 2.35454410e-01 -1.75412595e-01 -4.58672643e-01
-3.63026649e-01 3.15707147e-01 -3.23813587e-01 5.69099665e-01
9.66135681e-01 -1.16858736e-01 -6.01718165e-02 -2.27692217e-01
1.73396516e+00 2.31692083e-02 -9.94719326e-01 -3.64380598e-01
-1.27823427e-01 -8.42855275e-01 -1.13491230e-01 -2.61377335e-01
-1.40907466e+00 9.63488400e-01 5.71537793e-01 3.21231455e-01
1.36775053e+00 -2.80199140e-01 9.91994381e-01 3.98974538e-01
6.11411452e-01 -4.84032422e-01 -3.68235201e-01 -1.01155683e-01
5.98171592e-01 -1.12627256e+00 -7.56608928e-03 -9.36727166e-01
-4.74253833e-01 6.65855408e-01 6.06693566e-01 -3.00575763e-01
9.42901075e-01 1.09538548e-01 -3.32948744e-01 -3.37389596e-02
-3.47408205e-01 5.25314927e-01 1.26217604e-01 9.58622336e-01
-2.09043115e-01 1.55852914e-01 -1.73373118e-01 1.11874843e+00
-4.40543294e-02 -2.65143752e-01 2.26895586e-01 8.20995152e-01
1.65802658e-01 -1.38129866e+00 2.67981380e-01 4.96435285e-01
-4.01926100e-01 4.73232903e-02 -8.50124240e-01 4.79506880e-01
-3.55719358e-01 7.12375402e-01 -4.90526050e-01 -1.01795101e+00
1.78484783e-01 -4.84566242e-01 4.31690458e-03 -3.16883951e-01
-5.58458865e-01 -7.08144367e-01 -1.28237590e-01 -8.89102459e-01
1.34116352e-01 -2.96046704e-01 -1.00587201e+00 -6.42299235e-01
-1.08245257e-02 2.07195476e-01 2.77897924e-01 4.54536676e-01
8.60382080e-01 -8.27226788e-02 1.00319409e+00 -3.09149474e-01
-6.60011292e-01 -5.66191673e-02 -8.14614236e-01 5.87471724e-01
2.83064306e-01 -2.82130182e-01 -6.53064907e-01 -7.80474544e-01] | [6.01462459564209, 7.023069858551025] |
de6b8b77-ea98-4c3f-970d-0b69d4ba4440 | differentiable-rendering-with-perturbed | 2110.09107 | null | https://arxiv.org/abs/2110.09107v1 | https://arxiv.org/pdf/2110.09107v1.pdf | Differentiable Rendering with Perturbed Optimizers | Reasoning about 3D scenes from their 2D image projections is one of the core problems in computer vision. Solutions to this inverse and ill-posed problem typically involve a search for models that best explain observed image data. Notably, images depend both on the properties of observed scenes and on the process of image formation. Hence, if optimization techniques should be used to explain images, it is crucial to design differentiable functions for the projection of 3D scenes into images, also known as differentiable rendering. Previous approaches to differentiable rendering typically replace non-differentiable operations by smooth approximations, impacting the subsequent 3D estimation. In this paper, we take a more general approach and study differentiable renderers through the prism of randomized optimization and the related notion of perturbed optimizers. In particular, our work highlights the link between some well-known differentiable renderer formulations and randomly smoothed optimizers, and introduces differentiable perturbed renderers. We also propose a variance reduction mechanism to alleviate the computational burden inherent to perturbed optimizers and introduce an adaptive scheme to automatically adjust the smoothing parameters of the rendering process. We apply our method to 3D scene reconstruction and demonstrate its advantages on the tasks of 6D pose estimation and 3D mesh reconstruction. By providing informative gradients that can be used as a strong supervisory signal, we demonstrate the benefits of perturbed renderers to obtain more accurate solutions when compared to the state-of-the-art alternatives using smooth gradient approximations. | ['Justin Carpentier', 'Cordelia Schmid', 'Ivan Laptev', 'Quentin Le Lidec'] | 2021-10-18 | null | http://proceedings.neurips.cc/paper/2021/hash/ab233b682ec355648e7891e66c54191b-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/ab233b682ec355648e7891e66c54191b-Paper.pdf | neurips-2021-12 | ['3d-scene-reconstruction'] | ['computer-vision'] | [ 5.11282325e-01 1.71051115e-01 2.03925982e-01 -4.25934762e-01
-6.93277180e-01 -4.78945464e-01 6.81884527e-01 -1.64149061e-01
-1.30398124e-01 2.56411523e-01 1.05771147e-01 -2.22075850e-01
-2.85537243e-01 -5.34240842e-01 -9.48238492e-01 -5.92224240e-01
1.22683987e-01 5.50904274e-01 -2.29344890e-02 -4.01926458e-01
3.54076952e-01 8.25282395e-01 -1.57126367e+00 -1.24213127e-02
9.23036456e-01 8.83914769e-01 2.79987752e-01 7.35740542e-01
-1.92082018e-01 3.60747874e-01 -3.25677544e-01 -2.69652098e-01
3.42002302e-01 -3.33796561e-01 -6.50677741e-01 8.26985300e-01
4.93338972e-01 -2.56677270e-01 3.38395201e-02 9.31954741e-01
2.90102661e-01 3.59723836e-01 6.96936309e-01 -9.74928498e-01
-6.16731346e-01 2.67353728e-02 -6.14531934e-01 -2.14423075e-01
3.64543408e-01 2.03874782e-02 8.41118157e-01 -9.57570255e-01
6.02052093e-01 1.41512942e+00 5.85417807e-01 5.22301018e-01
-1.66760004e+00 5.83839491e-02 3.45358700e-01 -2.05752458e-02
-1.07364333e+00 -3.85992646e-01 1.07742679e+00 -5.22238016e-01
4.58812475e-01 6.96438015e-01 5.16501129e-01 7.86004484e-01
3.83666568e-02 5.83743870e-01 1.12525797e+00 -4.48840201e-01
2.18572095e-01 3.10480148e-01 4.43200693e-02 8.27780664e-01
-3.93020548e-02 1.49805903e-01 -4.39559281e-01 -2.83587545e-01
9.91570890e-01 -1.12682126e-01 -5.47249794e-01 -6.92485452e-01
-1.03682983e+00 7.68367171e-01 5.86972177e-01 -2.03473181e-01
-6.28867209e-01 1.29830286e-01 9.89069231e-03 2.24414803e-02
8.53401244e-01 7.37070143e-01 -2.22781613e-01 1.95381194e-01
-5.03520608e-01 4.43107218e-01 6.85913622e-01 5.96906066e-01
7.07919717e-01 6.28482401e-02 -6.05535991e-02 6.88390672e-01
3.36144060e-01 3.63694191e-01 -4.37423363e-02 -1.20406520e+00
1.29420996e-01 4.77304906e-01 2.15893179e-01 -9.97013211e-01
-1.48277119e-01 -4.82443660e-01 -5.79606414e-01 5.76220214e-01
2.10440099e-01 2.28643894e-01 -8.59530747e-01 1.50264728e+00
7.63371229e-01 2.72097528e-01 -1.63350761e-01 1.12375867e+00
5.77134192e-01 4.86669868e-01 -3.64479899e-01 -8.81661475e-02
1.03286672e+00 -6.22688055e-01 -4.21895474e-01 -7.99626857e-02
1.94158450e-01 -7.42012441e-01 1.42413986e+00 2.46407107e-01
-1.17965448e+00 -3.99550796e-01 -7.75115132e-01 -3.42562139e-01
4.21090126e-02 -1.78100001e-02 4.34892297e-01 3.67059410e-01
-8.91882360e-01 8.78284693e-01 -8.52696955e-01 -2.80317478e-02
3.39408100e-01 2.79645026e-01 -7.34244958e-02 9.53909978e-02
-4.29343998e-01 8.69134843e-01 -1.69280156e-01 2.81696409e-01
-7.35964596e-01 -9.10724401e-01 -7.65019119e-01 1.38335712e-02
4.98736978e-01 -1.00297606e+00 1.02709019e+00 -8.25977504e-01
-1.69544005e+00 1.08345759e+00 -3.33944649e-01 -2.81497657e-01
7.84561276e-01 -3.54689717e-01 2.43477046e-01 1.20704502e-01
-1.69360802e-01 3.35081190e-01 1.12210608e+00 -1.58004534e+00
-1.20132975e-01 -3.62439990e-01 2.28930771e-01 3.79670531e-01
-2.12254580e-02 -3.24071527e-01 -5.47520459e-01 -4.55779165e-01
4.15719509e-01 -9.64218318e-01 -6.23398125e-01 3.62290710e-01
-5.66423357e-01 1.76022187e-01 6.35748625e-01 -6.10627472e-01
5.84634364e-01 -2.16715002e+00 6.05848372e-01 2.54418939e-01
3.83118480e-01 -1.82267018e-02 -1.40390411e-01 1.19593993e-01
-9.05431516e-04 -5.27075306e-02 -4.32468861e-01 -7.38155365e-01
1.78136617e-01 4.05203521e-01 -4.84690130e-01 6.89281881e-01
3.10791343e-01 6.86069071e-01 -6.82204604e-01 -1.48444325e-01
5.96188307e-01 8.77786875e-01 -7.48315156e-01 3.63353133e-01
-5.39550066e-01 8.96776676e-01 -6.14755690e-01 1.27549931e-01
6.24089360e-01 -3.43734622e-01 -1.27606988e-01 -3.39094639e-01
-2.76729167e-01 3.04555386e-01 -1.20027566e+00 1.46761608e+00
-6.57394588e-01 4.37905788e-01 3.90191495e-01 -1.16278064e+00
8.97243381e-01 1.36531681e-01 4.23491329e-01 -2.84297198e-01
1.62607938e-01 1.98810160e-01 -4.77469623e-01 -5.51388204e-01
5.27224183e-01 -2.25846082e-01 4.46965992e-01 2.53188550e-01
-5.78010678e-01 -7.82642305e-01 -2.28654817e-01 -8.02367646e-03
6.48641825e-01 4.54985678e-01 1.45452842e-01 -3.74740332e-01
5.66110611e-01 4.97016460e-02 1.65899128e-01 5.41477442e-01
3.91649961e-01 8.85520458e-01 4.28320199e-01 -5.42414308e-01
-1.01929128e+00 -9.08616781e-01 -3.37437660e-01 7.29874730e-01
2.94294715e-01 -7.03759268e-02 -8.17273796e-01 -3.34440500e-01
6.36841282e-02 7.74599373e-01 -6.49298966e-01 1.40109248e-02
-6.08018696e-01 -6.05153918e-01 -1.07455254e-01 1.82422921e-01
3.16203475e-01 -7.27824867e-01 -8.20560873e-01 5.64525425e-02
5.59170544e-02 -1.14972150e+00 -5.00661790e-01 9.78733823e-02
-9.93250489e-01 -1.03753150e+00 -7.10027814e-01 -3.52153867e-01
9.99792874e-01 3.99597049e-01 1.19559336e+00 2.81602263e-01
-2.24468291e-01 6.72465146e-01 -1.26464907e-02 -3.39186937e-01
-4.83375728e-01 -2.20105752e-01 -1.62029222e-01 3.58547568e-01
-5.14546990e-01 -7.13542342e-01 -3.84575874e-01 3.42732847e-01
-9.41927254e-01 4.81395751e-01 2.58754462e-01 7.43862450e-01
9.63110328e-01 -3.45173389e-01 -2.32831582e-01 -1.02892959e+00
5.77668488e-01 -1.47214001e-02 -9.01097953e-01 2.48625889e-01
-5.14799237e-01 5.69793522e-01 6.07295573e-01 -4.35053647e-01
-1.10928392e+00 1.77526966e-01 -1.33201227e-01 -7.01233029e-01
-1.52956247e-02 2.55381256e-01 -4.56589125e-02 -3.62375945e-01
6.90815032e-01 9.55797285e-02 1.85895398e-01 -5.98127604e-01
5.09329140e-01 2.04893559e-01 5.28814852e-01 -6.97193027e-01
9.59252775e-01 8.47328663e-01 5.19255936e-01 -1.17688513e+00
-9.91075456e-01 -2.23246187e-01 -3.52987587e-01 -3.14847857e-01
7.92921305e-01 -4.38131422e-01 -7.91843772e-01 1.44598782e-01
-1.20537889e+00 -5.31990111e-01 -4.67799902e-01 3.15614700e-01
-8.98174524e-01 3.35503399e-01 -3.96266490e-01 -8.30206931e-01
-1.20539598e-01 -1.50435412e+00 1.52297413e+00 -9.34187993e-02
-2.34391004e-01 -1.11391234e+00 -1.21126793e-01 2.96678782e-01
3.04702163e-01 6.60784960e-01 9.52363849e-01 1.48091346e-01
-8.91599834e-01 5.21441996e-02 -1.19579755e-01 2.18002692e-01
2.41220798e-02 5.48117831e-02 -1.01029265e+00 1.18587520e-02
3.72090518e-01 8.86452105e-03 7.23246694e-01 7.55712152e-01
1.30109406e+00 -2.57629097e-01 -1.26997843e-01 1.18926525e+00
1.31282628e+00 -2.82816857e-01 3.49789917e-01 3.65193814e-01
8.91085982e-01 9.14588451e-01 3.56318086e-01 3.91224980e-01
2.60264993e-01 9.78491366e-01 7.10480452e-01 -2.66741186e-01
-5.35782091e-02 -1.70968503e-01 -4.00372744e-02 5.06825745e-01
-2.99202144e-01 2.99661383e-02 -6.06370270e-01 1.09638058e-01
-1.73450792e+00 -5.54636896e-01 -4.81241405e-01 2.35318398e+00
4.71872360e-01 -1.19960614e-01 -1.23640522e-01 6.33132905e-02
4.83729035e-01 2.08405659e-01 -6.02308750e-01 -3.90064895e-01
1.05849579e-02 1.43040881e-01 3.67628783e-01 9.52331841e-01
-8.71165514e-01 5.84473908e-01 6.11462069e+00 4.90745306e-01
-1.29463732e+00 -1.65603086e-01 5.48482358e-01 -6.93535507e-02
-7.73241580e-01 8.23522452e-03 -4.38856065e-01 1.76759839e-01
3.90258580e-01 -1.19695552e-01 8.19229841e-01 7.86014259e-01
4.91979718e-01 5.66362850e-02 -1.22114921e+00 1.09411478e+00
-5.66252582e-02 -1.49270856e+00 4.16846462e-02 2.58802503e-01
8.28073859e-01 -2.45861068e-01 2.97083497e-01 -2.92220205e-01
1.00763232e-01 -1.00397325e+00 9.80872631e-01 5.80099046e-01
4.43855375e-01 -4.91869032e-01 8.64545330e-02 3.82192105e-01
-9.04659808e-01 2.89922476e-01 -2.40334541e-01 -5.43319322e-02
4.34951156e-01 7.57066905e-01 -6.21145844e-01 5.15364647e-01
4.00111884e-01 5.97314179e-01 -2.30780780e-01 7.55526245e-01
-2.85857350e-01 2.60847002e-01 -4.30763572e-01 7.77213126e-02
1.84917703e-01 -6.18815899e-01 9.13267970e-01 7.17666507e-01
2.64360636e-01 2.93504566e-01 2.04914227e-01 1.16281867e+00
7.64408261e-02 -3.52879390e-02 -4.83606517e-01 3.62749904e-01
-7.10929707e-02 1.04916477e+00 -6.99948072e-01 -1.75593734e-01
-1.18880138e-01 9.95423019e-01 3.22856605e-01 5.30283511e-01
-7.47681499e-01 2.62681276e-01 7.66344368e-01 2.94766098e-01
3.23582470e-01 -3.22037607e-01 -6.01228237e-01 -1.11448658e+00
2.58687824e-01 -8.17516208e-01 -3.86174098e-02 -8.39060962e-01
-1.16117084e+00 6.52567983e-01 1.00268893e-01 -1.04580271e+00
-1.23810388e-01 -6.81220889e-01 -5.72925031e-01 9.14014459e-01
-1.57021213e+00 -8.37441146e-01 -4.39290464e-01 4.31516260e-01
6.49838209e-01 5.39933622e-01 5.77714980e-01 -1.26300640e-02
-2.53731906e-01 -3.66881257e-03 8.72176811e-02 -4.69731182e-01
1.55486822e-01 -1.23920298e+00 5.15730381e-01 7.05877960e-01
3.04910034e-01 5.21006823e-01 1.14664006e+00 -3.02092403e-01
-1.66290557e+00 -8.10548484e-01 5.19639075e-01 -5.19432485e-01
3.31955224e-01 -2.89643526e-01 -1.09602404e+00 6.61129296e-01
-3.01512122e-01 1.38426453e-01 4.69370522e-02 -4.18588854e-02
-8.97319335e-03 6.69333115e-02 -1.02038622e+00 7.74866045e-01
1.13138676e+00 -4.72277820e-01 -3.04393262e-01 4.79758352e-01
5.75738490e-01 -9.17134047e-01 -4.85094190e-01 1.83664918e-01
3.97542417e-01 -1.05462635e+00 1.24397004e+00 -4.22611177e-01
5.25714695e-01 -3.07912439e-01 -1.77071154e-01 -1.41848981e+00
-3.86960283e-02 -7.99592435e-01 -5.83775081e-02 7.73659647e-01
1.62009582e-01 -7.09960163e-01 7.45994687e-01 9.65326846e-01
-3.14603269e-01 -9.20514584e-01 -6.90519273e-01 -6.29423261e-01
-3.00931603e-01 -5.20441115e-01 5.54917872e-01 6.76430285e-01
-6.45644784e-01 1.74750224e-01 -3.78786474e-01 4.77723867e-01
9.57553983e-01 3.05141449e-01 1.00162220e+00 -1.10408115e+00
-6.13228738e-01 -4.37921196e-01 -2.14594871e-01 -1.38229620e+00
1.95886865e-01 -6.61864400e-01 2.93673664e-01 -1.32583606e+00
-1.37398109e-01 -5.49286962e-01 3.84028524e-01 -1.89454779e-01
-3.18368405e-01 9.24635157e-02 2.65357584e-01 1.43728629e-01
-5.91658615e-02 7.87993908e-01 1.55123568e+00 1.13128901e-01
-3.66704047e-01 1.15375556e-01 -4.79284137e-01 9.02707756e-01
3.96381140e-01 -3.00264657e-01 -4.84336078e-01 -7.65367806e-01
1.77141801e-01 2.22060215e-02 7.51271546e-01 -5.56174099e-01
-1.56783238e-01 -2.63648808e-01 1.25313476e-01 -3.33140343e-01
6.51422143e-01 -7.97763824e-01 4.56753284e-01 2.08157286e-01
-5.98956406e-01 -9.91725456e-03 -2.28086039e-02 4.35693592e-01
-1.52371859e-03 -2.27864310e-01 7.89556265e-01 -2.13332713e-01
-5.01046240e-01 4.17314231e-01 6.84713572e-02 -9.53894705e-02
7.96720982e-01 -2.04785004e-01 1.96035162e-01 -4.71616417e-01
-7.54773319e-01 6.54530525e-02 7.36668408e-01 4.19961177e-02
7.19618797e-01 -1.10641718e+00 -6.61728203e-01 2.60715306e-01
-1.64794952e-01 2.71888793e-01 5.42807430e-02 8.99910092e-01
-6.01530373e-01 8.14155303e-03 2.31068999e-01 -8.71687591e-01
-1.30863082e+00 3.98466349e-01 4.16445792e-01 -1.36023819e-01
-9.56271410e-01 7.72150874e-01 5.83275139e-01 -4.97277349e-01
1.24108903e-01 -5.33652067e-01 1.27198517e-01 -3.68290603e-01
3.36268395e-01 3.93890172e-01 1.35165736e-01 -5.45595706e-01
-1.58994853e-01 9.03539598e-01 2.61199206e-01 -2.62517750e-01
1.56553674e+00 -1.87114567e-01 -8.13623052e-03 3.65026385e-01
1.20115340e+00 -3.11961025e-02 -1.66389167e+00 -1.25360757e-01
-1.89452425e-01 -6.75862730e-01 2.86790997e-01 -1.78079739e-01
-1.01322997e+00 8.26297164e-01 2.67569155e-01 4.05589551e-01
9.76210296e-01 2.04183683e-01 5.09757400e-01 2.06243545e-01
1.41250685e-01 -5.86707354e-01 5.41305356e-02 2.63750464e-01
1.19632554e+00 -1.14516771e+00 1.68444708e-01 -7.63887405e-01
-4.76615995e-01 1.03453469e+00 1.56140924e-01 -4.13935900e-01
4.28800076e-01 1.12826332e-01 1.93981558e-01 -4.00406480e-01
-3.96753132e-01 -1.44693062e-01 6.71988487e-01 4.04897451e-01
3.11270267e-01 -2.37147994e-02 -5.31975441e-02 -1.13267310e-01
-2.88903385e-01 -2.87250310e-01 3.26708913e-01 6.04010701e-01
-1.32556319e-01 -1.05426860e+00 -6.59050643e-01 2.15679482e-01
-1.52376324e-01 1.41909510e-01 -2.27876112e-01 6.84134185e-01
-1.68893248e-01 5.84630787e-01 1.33459177e-02 6.52232170e-02
5.52586436e-01 -2.28299126e-01 6.88872695e-01 -4.96953338e-01
-3.74188155e-01 1.39119729e-01 -1.15648732e-01 -7.08796978e-01
-5.46027303e-01 -7.17107654e-01 -1.07933533e+00 -3.11751753e-01
-3.18046808e-01 -7.79558867e-02 8.44098628e-01 9.75062549e-01
3.65057737e-01 4.75741297e-01 7.21902370e-01 -1.47703791e+00
-7.93463409e-01 -3.24349701e-01 -4.67704415e-01 7.04620600e-01
6.52028978e-01 -7.02426612e-01 -4.40071195e-01 1.42156750e-01] | [9.179031372070312, -3.1324303150177] |
2e74ee78-d450-48b5-a3de-66546b0d140c | on-anomaly-interpretation-via-shapley-values | 2004.04464 | null | https://arxiv.org/abs/2004.04464v3 | https://arxiv.org/pdf/2004.04464v3.pdf | A Characteristic Function for Shapley-Value-Based Attribution of Anomaly Scores | In anomaly detection, the degree of irregularity is often summarized as a real-valued anomaly score. We address the problem of attributing such anomaly scores to input features for interpreting the results of anomaly detection. We particularly investigate the use of the Shapley value for attributing anomaly scores of semi-supervised detection methods. We propose a characteristic function specifically designed for attributing anomaly scores. The idea is to approximate the absence of some features by locally minimizing the anomaly score with regard to the to-be-absent features. We examine the applicability of the proposed characteristic function and other general approaches for interpreting anomaly scores on multiple datasets and multiple anomaly detection methods. The results indicate the potential utility of the attribution methods including the proposed one. | ['Yoshinobu Kawahara', 'Naoya Takeishi'] | 2020-04-09 | null | null | null | null | ['supervised-anomaly-detection', 'semi-supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 3.42625141e-01 2.26127446e-01 7.42348284e-02 -7.40333974e-01
-4.69063342e-01 -3.97160530e-01 7.53659308e-01 5.49043655e-01
-2.81532764e-01 4.86720145e-01 -1.23044677e-01 -2.34794244e-01
-6.24301076e-01 -7.71561086e-01 -1.45521343e-01 -6.55957758e-01
-3.84065598e-01 4.46226299e-01 1.46287233e-01 -1.08221941e-01
7.61288345e-01 8.67373109e-01 -1.71262622e+00 1.43699143e-02
8.74985874e-01 1.21561134e+00 -5.79278529e-01 6.42501652e-01
-2.20907658e-01 6.09328508e-01 -1.05314946e+00 -2.39730105e-01
4.86725271e-01 -6.10891700e-01 -5.62018275e-01 2.61948377e-01
5.46374619e-01 -1.73770651e-01 2.19364613e-01 1.38930500e+00
-1.05612129e-01 2.27926880e-01 1.16147673e+00 -1.83567417e+00
-5.56879878e-01 2.57108390e-01 -5.11173069e-01 7.65679777e-01
5.57450235e-01 -2.78982192e-01 1.34152412e+00 -7.38353133e-01
1.62122354e-01 1.02481949e+00 4.76298332e-01 3.54110837e-01
-1.28628850e+00 -1.35276228e-01 4.85870726e-02 2.96311677e-01
-1.41839552e+00 -1.08537152e-01 8.21520925e-01 -5.57416022e-01
5.33893764e-01 5.17341852e-01 1.24464400e-01 6.69019699e-01
1.77701190e-01 5.39200127e-01 9.17963207e-01 -4.89520460e-01
3.95889163e-01 -2.55684424e-02 4.74023581e-01 7.28712022e-01
6.59121454e-01 1.83710843e-01 -2.18397528e-01 -7.61278450e-01
4.67467517e-01 1.54039800e-01 2.03400806e-01 -2.58112639e-01
-8.52549255e-01 1.02101564e+00 1.36897102e-01 1.64077759e-01
-5.79173267e-01 -9.84140262e-02 4.59826678e-01 5.61958551e-01
6.26429319e-01 5.94501913e-01 -2.72784680e-01 2.23378822e-01
-8.69330347e-01 4.15593386e-01 6.94510460e-01 5.31145334e-01
8.00271213e-01 2.72356629e-01 -1.75024822e-01 6.55363262e-01
3.92431766e-01 3.36821973e-01 4.50458050e-01 -7.95861185e-01
2.31871337e-01 9.24254119e-01 1.08660989e-01 -1.08613813e+00
-3.85545552e-01 -2.87989050e-01 -6.73796713e-01 4.51069921e-01
5.78475296e-01 -4.96149771e-02 -5.88397384e-01 1.70084500e+00
2.38478571e-01 2.13336930e-01 1.60061359e-01 7.28644550e-01
2.48506203e-01 3.25904995e-01 5.12146242e-02 -2.87541479e-01
8.72565448e-01 -4.84266728e-01 -7.82346427e-01 1.66555941e-01
7.51670301e-01 -4.60727334e-01 8.05978656e-01 5.04311323e-01
-6.83137059e-01 -1.71111867e-01 -1.00520873e+00 6.51111305e-01
-4.87050772e-01 -8.87146443e-02 3.78940165e-01 7.74529040e-01
-8.14859271e-01 8.69258583e-01 -7.06123233e-01 -4.73995686e-01
2.73323894e-01 3.47466290e-01 -3.78777921e-01 7.14083791e-01
-9.91638601e-01 1.01752818e+00 5.93755960e-01 -1.44520430e-02
-4.44538176e-01 -2.69136131e-01 -7.33839750e-01 9.31467116e-02
2.05645502e-01 -5.03835455e-03 8.45060468e-01 -1.25045693e+00
-8.65302742e-01 8.53733838e-01 -7.76169822e-02 -5.43859601e-01
6.50224030e-01 3.57911922e-03 -9.28624809e-01 -5.58859669e-02
1.54639512e-01 -7.55892992e-02 8.64258230e-01 -1.09985054e+00
-7.54192114e-01 -5.37883282e-01 -2.57133275e-01 -9.40746143e-02
-5.10168195e-01 1.21663578e-01 4.99393612e-01 -8.38780761e-01
6.46137238e-01 -5.45280039e-01 -2.53478676e-01 1.33159563e-01
-4.85840797e-01 -3.17375869e-01 1.05586946e+00 -4.41683620e-01
1.36851442e+00 -2.15500212e+00 -1.88028067e-01 1.10643756e+00
1.45254418e-01 3.04388087e-02 -5.12897819e-02 1.91714346e-01
-2.57596672e-01 1.01241939e-01 -8.38940859e-01 9.16498975e-05
3.93863916e-02 6.07175231e-01 -4.92709219e-01 6.91386044e-01
6.29675269e-01 2.68665671e-01 -7.62877107e-01 -3.77697676e-01
8.01301599e-02 -1.81910038e-01 -5.33441186e-01 4.15437251e-01
1.14569888e-01 1.98921293e-01 -6.19369388e-01 9.49320555e-01
7.14057624e-01 1.49622411e-01 -1.30038127e-01 3.86954635e-01
-7.21119111e-03 -6.47549555e-02 -1.53895700e+00 8.89480293e-01
2.23596707e-01 5.42641342e-01 -4.28665966e-01 -1.22957802e+00
1.49069297e+00 1.40054777e-01 6.20326757e-01 -2.89848119e-01
3.88486572e-02 3.30477685e-01 8.90410691e-02 -3.10925364e-01
4.90638942e-01 -1.51748538e-01 -1.85046479e-01 7.06842840e-01
2.45568193e-02 3.34006637e-01 2.89886352e-02 7.89968595e-02
1.21832144e+00 -3.02096486e-01 9.44014966e-01 -4.70177948e-01
1.06996727e+00 -1.23209238e-01 5.60286283e-01 8.74334753e-01
-4.36745107e-01 5.50383151e-01 1.07084703e+00 -6.58331037e-01
-1.04197228e+00 -1.41711950e+00 -4.96524453e-01 1.09308362e+00
-1.05058029e-01 -1.81020766e-01 -5.48826277e-01 -1.05345345e+00
4.16335076e-01 9.40922558e-01 -8.01192343e-01 -4.18434888e-01
-2.41681948e-01 -9.48179483e-01 8.15910935e-01 5.26143134e-01
1.45311847e-01 -1.10317457e+00 -5.51013291e-01 -3.56837697e-02
1.13746889e-01 -4.70894814e-01 -1.20837659e-01 4.45018917e-01
-9.14348245e-01 -1.14355659e+00 3.41136605e-02 -5.71335182e-02
1.04544270e+00 -3.17713022e-01 8.70624781e-01 4.99033928e-01
-5.84863499e-02 4.11486357e-01 -4.37294960e-01 -5.24165154e-01
-6.60903037e-01 -4.33835685e-01 4.89726543e-01 4.40351933e-01
7.21508920e-01 -3.73694956e-01 -1.08267501e-01 4.67510015e-01
-1.05969989e+00 -1.00683331e+00 1.92454457e-01 8.36722612e-01
4.59044665e-01 -1.20949432e-01 8.14643621e-01 -1.09137392e+00
8.36065054e-01 -7.75807917e-01 -6.89952075e-01 1.77877113e-01
-1.18265700e+00 2.94605732e-01 5.11989832e-01 -3.44923347e-01
-7.54347324e-01 -1.34473428e-01 1.07255436e-01 -4.45929736e-01
-3.57348233e-01 2.46362060e-01 -2.12558433e-01 -2.24589810e-01
8.20816994e-01 8.13903287e-02 9.64752808e-02 -3.32554698e-01
1.69528346e-03 4.87243772e-01 5.97560108e-01 -7.80874074e-01
8.38564754e-01 2.28298709e-01 5.26068568e-01 -6.25288546e-01
-7.40338445e-01 -4.63091284e-01 -8.33878934e-01 -2.43154287e-01
3.63363028e-01 -1.95316896e-01 -3.81886512e-01 2.69130826e-01
-9.87514555e-01 4.86447781e-01 -6.90669000e-01 9.92161632e-02
-5.11713028e-01 6.64538860e-01 -3.27068754e-02 -1.10557616e+00
-3.35647389e-02 -8.51404667e-01 7.54541397e-01 -9.13040489e-02
-4.64968443e-01 -1.29417372e+00 3.43054742e-01 -2.04974547e-01
3.88878971e-01 5.47765374e-01 8.73490989e-01 -1.72157419e+00
7.03715459e-02 -6.75651968e-01 -1.82687834e-01 6.09247208e-01
3.39431375e-01 3.46164167e-01 -1.12030745e+00 -9.38243791e-02
-8.18470307e-03 1.55867755e-01 7.34548151e-01 6.42048335e-03
1.30965710e+00 -6.38242304e-01 1.53441638e-01 4.10005063e-01
1.42444682e+00 1.78485930e-01 4.99253780e-01 6.22026026e-01
5.19751847e-01 6.89223647e-01 6.88145697e-01 8.22192669e-01
-1.64164364e-01 5.90294123e-01 6.44258082e-01 1.88291118e-01
6.03232384e-01 -4.72294241e-02 3.66736829e-01 4.61422384e-01
-1.47490516e-01 -1.20479845e-01 -1.06142139e+00 4.40258712e-01
-1.95298183e+00 -1.21489096e+00 -4.69711840e-01 2.38651848e+00
2.49557704e-01 2.82948077e-01 2.41373405e-01 4.49588418e-01
9.45763767e-01 7.82476924e-03 -4.54504967e-01 -9.44341779e-01
-3.53998691e-01 4.25356217e-02 2.11155415e-01 4.94022012e-01
-1.24802184e+00 4.24720198e-01 7.49256897e+00 5.15190899e-01
-6.63706601e-01 -1.57407284e-01 3.88646692e-01 3.28076929e-01
-1.97845072e-01 3.14033031e-01 -3.78756136e-01 4.59643006e-01
9.49875534e-01 -5.20572245e-01 -1.56763457e-02 1.00880599e+00
-5.06333960e-03 -1.10164694e-02 -1.27156973e+00 3.34291518e-01
1.60623625e-01 -6.48318589e-01 5.37703335e-01 1.75663933e-01
7.23607183e-01 -5.07214904e-01 6.86510131e-02 -2.15443894e-01
5.36835827e-02 -7.59600520e-01 4.88193840e-01 8.41641068e-01
3.20397317e-01 -8.35478663e-01 9.88366246e-01 -4.45359461e-02
-9.54363704e-01 -3.12060207e-01 -4.05409366e-01 -3.48960549e-01
-2.52497584e-01 6.18577421e-01 -9.04228449e-01 2.69623786e-01
2.03724489e-01 5.69739461e-01 -7.88416326e-01 1.09505665e+00
-8.37904140e-02 4.95032370e-01 -3.91355127e-01 1.52218908e-01
1.81515709e-01 -5.22450566e-01 8.51639450e-01 1.10432041e+00
5.56917906e-01 -2.06020638e-01 1.28530443e-01 1.07489622e+00
3.77088457e-01 2.89507449e-01 -8.74426842e-01 1.29156113e-01
4.23744351e-01 1.01141667e+00 -6.80680752e-01 -3.25732917e-01
-5.46762049e-01 7.22585797e-01 9.36895311e-02 1.27973512e-01
-4.30021673e-01 -4.04293329e-01 7.95119286e-01 5.52257895e-02
-1.40611455e-01 2.16077566e-01 -7.73750722e-01 -9.24743772e-01
2.11559534e-01 -4.91758168e-01 9.65452313e-01 -4.00316864e-01
-1.70575035e+00 5.17349899e-01 1.06400691e-01 -1.70711052e+00
-5.06462336e-01 -6.53630078e-01 -1.32207358e+00 7.45725811e-01
-1.14132214e+00 -4.83628213e-01 -2.14665607e-01 7.33025610e-01
5.39387278e-02 -7.17345238e-01 9.10100877e-01 -1.34536952e-01
-5.44417679e-01 6.92959070e-01 1.32215843e-01 1.47971272e-01
6.32894218e-01 -1.60066986e+00 2.74928749e-01 1.21940601e+00
1.19458526e-01 4.22123164e-01 1.06967974e+00 -6.38195276e-01
-4.52349305e-01 -1.04658985e+00 8.52990091e-01 -7.38996804e-01
9.81947601e-01 2.66056687e-01 -1.37686169e+00 7.16586888e-01
-4.44694519e-01 2.22572058e-01 7.57617712e-01 9.06518847e-02
-4.27414298e-01 -7.74813071e-02 -1.80632746e+00 1.76779613e-01
6.80468857e-01 -3.78075600e-01 -1.09061348e+00 5.10697514e-02
1.92685109e-02 1.76285133e-01 -8.21799517e-01 3.77897710e-01
2.09094122e-01 -1.01032329e+00 8.01501036e-01 -9.34356689e-01
3.41413885e-01 -4.12237525e-01 -4.49063122e-01 -1.19207788e+00
-3.46908182e-01 -1.14700824e-01 -1.90999299e-01 1.18388712e+00
4.15979952e-01 -9.95032132e-01 6.60826921e-01 7.05027103e-01
1.09213769e-01 -2.36481920e-01 -1.22613680e+00 -7.81203568e-01
-1.35458514e-01 -4.33677614e-01 6.22905135e-01 1.21624601e+00
1.74836323e-01 -3.88508737e-01 -2.39708886e-01 4.56326276e-01
1.09976053e+00 -1.92668393e-01 4.01414156e-01 -1.99417412e+00
1.06327511e-01 -5.19272149e-01 -1.28768408e+00 1.12016886e-01
3.42911005e-01 -9.89129782e-01 -2.36658067e-01 -6.70368016e-01
4.75478023e-02 -2.90682912e-01 -9.62558687e-01 6.40182972e-01
-1.89565659e-01 1.23701200e-01 -1.07906103e-01 3.45020682e-01
-4.89125490e-01 4.30593401e-01 6.26660824e-01 1.32267788e-01
-1.38181314e-01 3.50223333e-01 -5.97577631e-01 1.01531041e+00
9.88454342e-01 -8.52279603e-01 -6.45748824e-02 2.57003754e-01
1.80080950e-01 -7.03230128e-02 6.78688467e-01 -1.05089712e+00
2.60626227e-01 -4.49584126e-01 3.40002149e-01 -3.71131599e-01
-1.53828055e-01 -1.03083718e+00 -5.82867377e-02 3.59474510e-01
-6.31876051e-01 5.75434566e-01 -1.98820472e-01 7.22311616e-01
-4.75818068e-01 -5.98685980e-01 7.89692402e-01 9.54051763e-02
-8.01209807e-01 3.29493987e-03 -3.79164666e-01 8.69304091e-02
1.11741590e+00 -3.85042936e-01 -2.34927475e-01 -2.31910497e-01
-7.95545101e-01 1.10884786e-01 3.14378500e-01 3.49452674e-01
9.09077466e-01 -1.62280405e+00 -7.90044487e-01 6.84311569e-01
6.13329411e-01 -5.93926072e-01 -2.22698212e-01 6.18969440e-01
-4.63921964e-01 3.42911631e-02 -7.08880723e-01 -5.69283783e-01
-1.19426513e+00 3.55163783e-01 3.56393695e-01 3.22731026e-02
-6.09271765e-01 2.66178250e-01 -3.99726108e-02 -3.80233914e-01
1.64662257e-01 2.71598902e-02 -3.64341855e-01 1.12770483e-01
4.72762913e-01 8.75834703e-01 -2.47148820e-03 -7.33495951e-01
-5.37761629e-01 4.22021270e-01 -1.64479002e-01 -7.68470466e-02
1.14271295e+00 2.52841711e-02 -4.61093068e-01 5.14536321e-01
8.65109980e-01 -2.16257364e-01 -9.43242967e-01 -3.33902717e-01
6.72486186e-01 -6.78443372e-01 -9.26531330e-02 -6.73282266e-01
-1.04746711e+00 5.04808843e-01 8.07825923e-01 4.56234753e-01
1.14827275e+00 -1.22285306e-01 -1.58767119e-01 7.25988805e-01
1.48949027e-01 -1.21796012e+00 -6.75723702e-02 4.84831303e-01
6.59433365e-01 -1.29321682e+00 9.76799279e-02 -3.09247850e-03
-7.93610036e-01 1.35322523e+00 5.86787522e-01 -4.68651652e-01
7.18432069e-01 9.75404009e-02 1.43062755e-01 -2.38247603e-01
-5.26010513e-01 -3.10660936e-02 5.54920554e-01 5.57031572e-01
2.59459853e-01 3.16071689e-01 -6.17507935e-01 4.34611589e-01
-8.92888606e-02 -5.26663899e-01 8.95484209e-01 8.38435173e-01
-5.70624650e-01 -8.26714396e-01 -6.22978449e-01 8.97205412e-01
-4.32801753e-01 2.41465449e-01 -6.87111437e-01 4.96220827e-01
-5.73093593e-02 9.59606826e-01 3.85620028e-01 -3.56467932e-01
5.48096001e-01 4.43053037e-01 -2.29708686e-01 -2.88605690e-01
-4.67317820e-01 -3.55559617e-01 -1.73484176e-01 -5.86095512e-01
-3.38679731e-01 -9.52163577e-01 -1.11825645e+00 -2.84310341e-01
-3.25309575e-01 1.89474240e-01 4.92979348e-01 1.10848188e+00
-1.44798085e-01 3.36222440e-01 9.19946551e-01 -3.77215922e-01
-9.25469398e-01 -8.45229983e-01 -1.11245942e+00 8.15457046e-01
6.43714488e-01 -8.23515713e-01 -1.01241767e+00 -4.22102600e-01] | [7.623112678527832, 2.5726993083953857] |
555493b3-523b-4e9a-bc34-2d0642df5d3e | spotnet-learned-iterations-for-cell-detection | 1810.06132 | null | http://arxiv.org/abs/1810.06132v1 | http://arxiv.org/pdf/1810.06132v1.pdf | SpotNet - Learned iterations for cell detection in image-based immunoassays | Accurate cell detection and counting in the image-based ELISpot and
FluoroSpot immunoassays is a challenging task. Methodology recently proposed by
our group matches human accuracy by leveraging knowledge of the underlying
physical process of these assays and using state-of-the-art iterative
techniques to solve an inverse problem. Nonetheless, thousands of
computationally expensive iterations are often needed to reach a near-optimal
solution. In this paper, we exploit the structure of the iterations to design a
parameterized computation graph, SpotNet, that learns the characteristic
patterns embedded within several training images and their respective cell
secretion information. Further, we compare SpotNet to a customized
convolutional neural network layout for cell detection based on recent
advances. We show empirical evidence that, while both designs obtain a
detection performance far beyond that of a human expert, SpotNet is
substantially easier to train and obtains better estimates of cell secretion. | ['Joakim Jaldén', 'Vidit Saxena', 'Pol del Aguila Pla'] | 2018-10-15 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 1.45110831e-01 -1.67821273e-01 1.95696339e-01 -1.11453116e-01
-4.47752267e-01 -7.43141651e-01 3.14371824e-01 5.88852167e-01
-6.00761533e-01 7.87401557e-01 -5.46099305e-01 -3.42985928e-01
7.20481351e-02 -6.50346220e-01 -8.24701667e-01 -6.39251292e-01
-3.32076758e-01 9.10711944e-01 1.63893104e-01 1.07203983e-01
3.71085733e-01 9.08892453e-01 -9.92399395e-01 3.47689688e-01
4.01135147e-01 1.17565727e+00 -1.06627621e-01 1.27052736e+00
-1.32167384e-01 8.14256907e-01 -4.88139361e-01 -1.61593899e-01
2.33613119e-01 -3.79434884e-01 -3.64238292e-01 2.34352965e-02
2.49203920e-01 -2.29157522e-01 2.75053568e-02 6.07892454e-01
4.59142596e-01 -4.48367655e-01 6.96289599e-01 -1.03931928e+00
-3.99853468e-01 -8.54260847e-02 -5.36485255e-01 3.03604901e-01
1.47201255e-01 4.17954087e-01 4.15602446e-01 -6.83133245e-01
9.92514670e-01 7.99785376e-01 9.76427853e-01 3.17113727e-01
-1.59545314e+00 -4.86588389e-01 -3.06175649e-01 -4.05924946e-01
-1.47946060e+00 -1.03008397e-01 2.17757612e-01 -4.61177289e-01
9.47935283e-01 1.79946646e-01 1.02932656e+00 8.54407012e-01
6.99359536e-01 4.38462585e-01 1.44315135e+00 -1.65178508e-01
4.81486261e-01 2.08419710e-01 5.43427914e-02 1.09048605e+00
5.51528692e-01 1.79339275e-02 -4.15946066e-01 -2.38334551e-01
1.37024379e+00 2.88132995e-01 2.16633175e-02 -4.61507559e-01
-1.09101784e+00 7.10497856e-01 2.77027190e-01 4.42763805e-01
-2.31046945e-01 3.96366566e-01 1.99754894e-01 1.97586164e-01
4.68886137e-01 8.93042266e-01 -4.55312043e-01 2.18672022e-01
-9.26966667e-01 3.06813478e-01 1.19311142e+00 6.94665849e-01
8.94848168e-01 -2.86300600e-01 -1.24549761e-01 7.75253251e-02
9.40115750e-03 4.46829855e-01 5.60171641e-02 -9.11748528e-01
-3.37390423e-01 1.06384742e+00 3.91076386e-01 -1.04989386e+00
-8.76039207e-01 -8.32890928e-01 -7.63100863e-01 2.37884656e-01
9.71124351e-01 -3.17629963e-01 -1.00578642e+00 1.15616119e+00
4.23238844e-01 2.29741678e-01 -3.17599744e-01 9.86637115e-01
2.31211305e-01 3.69821280e-01 6.78287894e-02 -2.49720261e-01
1.35591805e+00 -5.86268425e-01 -3.16798627e-01 1.59334287e-01
9.31342602e-01 -6.83381677e-01 6.60318077e-01 5.07629693e-01
-8.79075706e-01 -9.78589654e-02 -1.16408384e+00 -1.02476485e-01
-6.63759232e-01 7.48099014e-02 9.36117411e-01 6.00894451e-01
-1.08750045e+00 5.03512681e-01 -9.34979618e-01 -6.09518111e-01
5.23092091e-01 6.30367577e-01 6.26310054e-03 1.57412305e-01
-5.64791083e-01 7.63727069e-01 -6.36602789e-02 -1.41150326e-01
-8.63633692e-01 -1.04875898e+00 -4.03754920e-01 -1.47099614e-01
1.30232960e-01 -7.61067390e-01 1.05797327e+00 -6.82164609e-01
-1.42026579e+00 1.06059563e+00 -8.73084739e-02 -6.77426279e-01
6.19408369e-01 1.24317750e-01 4.03281376e-02 5.06228089e-01
-3.66331935e-01 6.45777643e-01 3.72756302e-01 -1.17649984e+00
-5.17344892e-01 -4.85759079e-01 -1.67601526e-01 -2.16761917e-01
2.13466398e-03 -1.27327442e-01 -3.86508971e-01 -1.73710063e-01
-4.33165357e-02 -7.66600013e-01 -5.56555986e-01 3.87926012e-01
-1.89033762e-01 1.19490780e-01 6.67241096e-01 -3.29668134e-01
7.17712402e-01 -1.65026438e+00 -3.41568381e-01 5.13084054e-01
8.66113842e-01 1.67704344e-01 -1.10738762e-01 6.47427619e-01
6.11223102e-01 -1.39077425e-01 1.69298321e-01 -1.46718383e-01
-2.06579074e-01 -3.43859829e-02 1.22526558e-02 8.47256005e-01
2.00574473e-01 1.36175609e+00 -9.36475098e-01 -5.78856170e-01
1.87247515e-01 7.75953472e-01 -4.14385259e-01 1.89662606e-01
-3.08970422e-01 6.69304430e-01 -4.91250813e-01 8.62053931e-01
6.34285092e-01 -1.02543378e+00 6.39278293e-01 -1.23475939e-01
-5.08867688e-02 -1.95219561e-01 -8.15440059e-01 1.24266875e+00
-2.56904900e-01 4.94771361e-01 3.92705679e-01 -9.76002574e-01
7.35062838e-01 1.28054433e-02 6.96669817e-01 -7.40281880e-01
5.68576276e-01 3.38145316e-01 -1.55205309e-01 -2.51301408e-01
1.35258630e-01 -2.57308662e-01 2.24343121e-01 7.29526937e-01
-7.35490993e-02 -2.51304489e-02 3.07515234e-01 2.52470106e-01
1.38170230e+00 -2.70089716e-01 2.42135942e-01 -6.59508228e-01
3.58700782e-01 3.77824485e-01 1.67121202e-01 1.00864720e+00
-2.41448998e-01 5.15850902e-01 5.63320398e-01 -1.06202257e+00
-1.29799581e+00 -1.01458776e+00 -8.47881660e-02 7.43584514e-01
1.10007465e-01 -4.18723887e-03 -1.11607409e+00 -3.30997974e-01
4.19057310e-01 -3.21860686e-02 -8.61540794e-01 4.28996205e-01
-5.01946270e-01 -8.38876784e-01 6.91292882e-01 5.54587483e-01
6.23815060e-02 -6.67158067e-01 -1.00364888e+00 4.80307490e-01
3.82695168e-01 -1.31948650e+00 -3.02438438e-01 1.43959388e-01
-7.33105242e-01 -1.40002918e+00 -6.98341131e-01 -5.90813994e-01
1.16033113e+00 9.01749805e-02 1.13477170e+00 7.71094620e-01
-9.27113056e-01 4.34674889e-01 -6.77488223e-02 -5.75819314e-01
-3.26262385e-01 3.94522473e-02 -8.91967267e-02 4.47087772e-02
6.65808499e-01 -3.39440227e-01 -1.00727642e+00 2.47136652e-01
-9.08437431e-01 8.21771100e-02 7.28906989e-01 7.24392831e-01
8.01329374e-01 -3.20693314e-01 2.43197054e-01 -1.06742656e+00
5.98592579e-01 -1.63843036e-01 -8.85251343e-01 2.55831838e-01
-7.79838264e-01 4.77712527e-02 8.00811410e-01 -5.46334267e-01
-5.26490271e-01 2.46411785e-01 2.75485367e-01 -2.86094159e-01
-5.56900650e-02 1.72931388e-01 8.00281227e-01 -7.22502112e-01
6.33973658e-01 2.90121138e-01 4.49819952e-01 -5.93655482e-02
4.82662432e-02 3.12285304e-01 2.56780684e-01 -4.50818777e-01
3.61867219e-01 1.09180009e+00 4.23363477e-01 -6.00033581e-01
-6.19095683e-01 -6.23813808e-01 -3.51194859e-01 -1.80424839e-01
7.22930372e-01 -5.69784045e-01 -1.29982293e+00 6.09112263e-01
-1.19239128e+00 -4.59043741e-01 -1.33647114e-01 2.94515580e-01
-5.64696789e-01 1.94422528e-02 -1.14333367e+00 -9.58915293e-01
-4.77060944e-01 -9.37804818e-01 1.29404294e+00 1.61601245e-01
-2.00194448e-01 -1.15723348e+00 1.60639569e-01 3.25870961e-02
6.32393360e-01 5.33271909e-01 6.65715218e-01 -7.10013509e-01
-1.03368843e+00 -4.55418557e-01 -6.15378737e-01 -4.02391940e-01
-1.09078445e-01 -3.73066366e-02 -8.62259626e-01 -5.42688012e-01
7.34650642e-02 -2.05320537e-01 6.84614480e-01 5.46023726e-01
1.01174569e+00 -2.55127639e-01 -8.07754934e-01 7.21943974e-01
1.72860074e+00 1.55251309e-01 6.02645457e-01 -1.00912750e-01
4.78010029e-01 6.22691929e-01 1.00846045e-01 3.25565845e-01
1.21316381e-01 2.82545835e-01 -5.08069992e-03 -4.95448560e-01
1.64868552e-02 -1.60334677e-01 -3.84376734e-01 6.10842168e-01
8.95538460e-03 -2.58802414e-01 -9.25859690e-01 4.15935397e-01
-1.64860010e+00 -4.95125890e-01 -6.85411766e-02 1.98973918e+00
6.94961011e-01 -9.73820686e-04 1.18903793e-01 -3.48292112e-01
4.56611037e-01 -4.76220459e-01 -5.95193923e-01 -2.80288339e-01
-8.70122612e-02 4.03068662e-01 9.23530996e-01 4.15142179e-01
-6.29278660e-01 6.32248580e-01 8.01826668e+00 6.76781356e-01
-1.28139818e+00 -6.77316338e-02 1.18787098e+00 -3.06406915e-01
-9.85182226e-02 -1.48882598e-01 -7.84809947e-01 2.71101236e-01
8.82921815e-01 1.40976667e-01 6.65598035e-01 2.80602157e-01
9.99092832e-02 -3.96136880e-01 -1.28165495e+00 9.91418779e-01
-7.12944120e-02 -1.84105480e+00 -4.74567749e-02 3.52609098e-01
7.58065760e-01 -1.19756632e-01 1.61367618e-02 -1.53915554e-01
4.21074122e-01 -1.22305369e+00 1.11858845e-01 6.40694797e-01
7.47063577e-01 -5.85188746e-01 6.96749926e-01 2.59395868e-01
-1.04300547e+00 1.08952083e-01 -1.60485640e-01 -2.30057627e-01
-8.55774656e-02 8.62632632e-01 -1.30968440e+00 -3.36451568e-02
3.29926521e-01 3.71060744e-02 -5.86946964e-01 8.96268666e-01
3.92347097e-01 2.67183721e-01 -5.17560065e-01 -7.25529552e-01
2.93965876e-01 -1.78841919e-01 1.80391490e-01 1.57242525e+00
2.53710806e-01 1.30532086e-01 5.24698198e-02 1.06012571e+00
-1.68152571e-01 1.43268213e-01 -6.06778741e-01 -5.89332402e-01
1.62713602e-01 1.58229923e+00 -1.37093985e+00 -4.78853524e-01
-2.20239818e-01 6.62721097e-01 7.02785790e-01 6.40217602e-01
-5.24708629e-01 -1.61994115e-01 5.10421634e-01 5.55696189e-01
3.29765707e-01 -3.54938149e-01 -6.08446777e-01 -6.70733869e-01
-3.20078939e-01 -5.88964045e-01 1.55648291e-01 -6.21027350e-01
-1.27290547e+00 1.03924021e-01 -4.82750744e-01 -7.59337783e-01
-1.68891236e-01 -1.20183218e+00 -2.86192268e-01 6.30262315e-01
-1.31108403e+00 -1.02249837e+00 -1.95598453e-01 6.18512928e-02
7.20353350e-02 1.40043780e-01 7.13476479e-01 5.86141907e-02
-2.33541191e-01 3.55591297e-01 1.48008674e-01 1.71956882e-01
3.53964090e-01 -1.12100780e+00 3.24748546e-01 4.07393843e-01
-1.28117338e-01 8.86388600e-01 7.83560514e-01 -7.55919993e-01
-1.83948028e+00 -7.69722879e-01 7.90659010e-01 -6.17107570e-01
7.70437479e-01 -5.92234373e-01 -7.53385723e-01 4.93694574e-01
1.18850708e-01 2.39502296e-01 4.74160016e-01 -7.64904097e-02
-2.33967155e-01 -1.09085537e-01 -1.31971991e+00 5.64428806e-01
9.92749035e-01 -6.07619882e-01 1.53196886e-01 5.60242772e-01
3.20801437e-01 -6.12513542e-01 -6.37690187e-01 1.22238927e-01
9.47679698e-01 -9.92324054e-01 7.27645099e-01 -5.97534895e-01
7.85182565e-02 -3.08556229e-01 2.29022577e-01 -8.63975048e-01
-7.00572506e-02 -7.47937560e-01 -2.15321928e-01 3.62473041e-01
3.63147467e-01 -9.74611759e-01 1.18757033e+00 3.53890568e-01
3.89972925e-01 -1.28159118e+00 -7.10089624e-01 -7.76795268e-01
-1.28414258e-02 1.16841421e-01 5.04911125e-01 7.63543367e-01
9.23368409e-02 9.46417749e-02 7.16535300e-02 -2.84470171e-02
9.81274724e-01 2.05541924e-01 8.58345151e-01 -1.16429341e+00
-2.74140000e-01 -4.14177328e-01 -4.93592858e-01 -9.61503804e-01
-3.91878724e-01 -5.40299177e-01 2.46254168e-02 -1.18367708e+00
3.72407287e-01 -2.41472021e-01 -4.39189672e-01 -4.76086251e-02
2.06311807e-01 5.52873194e-01 -7.87911192e-02 8.86176229e-02
-8.32843781e-01 -3.36943150e-01 1.51467776e+00 -1.35546913e-02
-2.20445003e-02 -5.96352935e-01 -6.44709766e-01 4.45811719e-01
7.61254072e-01 -4.20340508e-01 -7.30938315e-02 -2.59621173e-01
5.40126920e-01 -8.17541555e-02 6.82520270e-01 -9.82010663e-01
6.79604471e-01 2.87454273e-03 9.05597329e-01 -6.24222219e-01
1.62551045e-01 -5.11855125e-01 3.22404981e-01 8.42791200e-01
-1.65091366e-01 3.31726633e-02 2.02972934e-01 5.25026023e-01
2.96565294e-01 1.25632226e-01 7.48595238e-01 -4.47995454e-01
-2.35102788e-01 2.25552246e-01 -5.80331922e-01 -1.26528651e-01
9.62275863e-01 -4.36536193e-01 -5.64275026e-01 -2.19898805e-01
-3.83418530e-01 2.04463482e-01 8.30595911e-01 -3.35246593e-01
5.06646514e-01 -9.15328383e-01 -4.88303155e-01 2.85123140e-01
-4.47478220e-02 -3.39906663e-01 7.00025111e-02 1.21225584e+00
-1.24702942e+00 6.58573449e-01 -2.53668100e-01 -9.10905421e-01
-7.64071286e-01 5.32702386e-01 5.59180439e-01 -6.43024802e-01
-2.70491421e-01 7.60458708e-01 4.02377635e-01 -4.27147567e-01
-1.69230804e-01 -2.85485178e-01 2.55787969e-01 -3.09791148e-01
5.36089540e-01 4.56991494e-01 4.22462262e-02 3.99516076e-02
-3.42064440e-01 5.85316777e-01 -2.35825613e-01 4.29304652e-02
1.07812738e+00 -6.06041439e-02 -1.05788238e-01 8.00535902e-02
1.19259596e+00 -2.20483720e-01 -1.51282048e+00 5.39549179e-02
-1.16781041e-01 -2.16264710e-01 -7.23475888e-02 -8.79620314e-01
-8.15059066e-01 7.51852274e-01 6.34032428e-01 3.48259747e-01
7.90539384e-01 -7.56794512e-02 8.99347842e-01 5.45749366e-01
5.80174744e-01 -1.25817776e+00 1.38734028e-01 2.00804636e-01
4.17353153e-01 -1.00907171e+00 1.68644279e-01 -6.09091878e-01
-2.54220353e-03 9.24310565e-01 5.21353662e-01 -3.86042178e-01
4.25682753e-01 8.18066418e-01 3.43328148e-01 -7.46679246e-01
-7.09496915e-01 -1.70768052e-03 -3.75804529e-02 6.96516037e-01
3.70530307e-01 -6.49809279e-03 -4.34039652e-01 1.51164904e-01
1.86151236e-01 5.10277629e-01 2.67692924e-01 1.01809144e+00
-5.30188978e-01 -6.60896182e-01 -2.86246121e-01 6.87917829e-01
-5.04322886e-01 1.38489738e-01 -6.55994117e-01 8.31359684e-01
-3.96274030e-02 6.72474504e-01 3.51601541e-01 -8.28455389e-02
2.51074582e-02 -1.64250001e-01 8.66946638e-01 -3.01345825e-01
-5.83641052e-01 1.31231397e-01 -1.54672503e-01 -7.36715257e-01
-2.72191286e-01 -2.98624575e-01 -1.55131364e+00 -5.32101572e-01
-8.69173259e-02 -1.67387575e-02 5.68827212e-01 1.05414951e+00
6.64615154e-01 2.53074735e-01 2.96691507e-01 -7.71958947e-01
-2.45434105e-01 -6.18045032e-01 -9.11598146e-01 2.98861623e-01
3.04960012e-01 -4.58936095e-01 -2.57611156e-01 5.56672961e-02] | [14.344950675964355, -3.1908881664276123] |
c4af7dbd-275a-4daa-ab48-bb28babf621e | generalized-organ-segmentation-by-imitating | 2103.16344 | null | https://arxiv.org/abs/2103.16344v1 | https://arxiv.org/pdf/2103.16344v1.pdf | Generalized Organ Segmentation by Imitating One-shot Reasoning using Anatomical Correlation | Learning by imitation is one of the most significant abilities of human beings and plays a vital role in human's computational neural system. In medical image analysis, given several exemplars (anchors), experienced radiologist has the ability to delineate unfamiliar organs by imitating the reasoning process learned from existing types of organs. Inspired by this observation, we propose OrganNet which learns a generalized organ concept from a set of annotated organ classes and then transfer this concept to unseen classes. In this paper, we show that such process can be integrated into the one-shot segmentation task which is a very challenging but meaningful topic. We propose pyramid reasoning modules (PRMs) to model the anatomical correlation between anchor and target volumes. In practice, the proposed module first computes a correlation matrix between target and anchor computerized tomography (CT) volumes. Then, this matrix is used to transform the feature representations of both anchor volume and its segmentation mask. Finally, OrganNet learns to fuse the representations from various inputs and predicts segmentation results for target volume. Extensive experiments show that OrganNet can effectively resist the wide variations in organ morphology and produce state-of-the-art results in one-shot segmentation task. Moreover, even when compared with fully-supervised segmentation models, OrganNet is still able to produce satisfying segmentation results. | ['Yefeng Zheng', 'Kai Ma', 'Yizhou Yu', 'Chixiang Lu', 'Dong Wei', 'Shilei Cao', 'Hualuo Liu', 'Hong-Yu Zhou'] | 2021-03-30 | null | null | null | null | ['one-shot-segmentation'] | ['computer-vision'] | [ 1.78681925e-01 5.96407235e-01 -6.60272315e-02 -3.63320410e-01
-6.23800755e-01 -2.95879692e-01 2.21606359e-01 3.53841007e-01
-3.65545869e-01 5.69335222e-01 -1.15888901e-01 2.94121820e-02
-1.34582162e-01 -7.75882959e-01 -6.11284375e-01 -6.58552229e-01
-6.63399771e-02 7.03913331e-01 3.81912917e-01 -2.02152923e-01
-1.42127946e-01 6.18289471e-01 -9.21229362e-01 1.07625097e-01
1.08889735e+00 7.94480562e-01 3.17773372e-01 5.64642966e-01
-3.16104501e-01 9.17569101e-01 -4.44836915e-01 -2.95770466e-01
3.15734953e-01 -6.19644225e-01 -1.05095446e+00 3.93386811e-01
2.01161698e-01 -2.32410267e-01 -3.19773525e-01 1.12438130e+00
3.08846653e-01 1.37749150e-01 1.04737771e+00 -9.90904629e-01
-6.63444400e-01 7.64614761e-01 -4.52997893e-01 3.65381122e-01
1.45471200e-01 2.42411345e-01 5.77262938e-01 -7.39176273e-01
7.41562605e-01 9.60337043e-01 6.75532222e-01 6.26780093e-01
-1.13758898e+00 -3.50758851e-01 -1.16372816e-02 -1.12833448e-01
-1.24284625e+00 1.64265968e-02 5.78814149e-01 -5.33114493e-01
5.11796296e-01 1.40761256e-01 1.02563727e+00 4.66630995e-01
4.65973496e-01 1.04273713e+00 7.75818646e-01 -8.79249945e-02
2.04087779e-01 1.39515609e-01 2.33648181e-01 9.67543721e-01
5.55824675e-02 -1.58011287e-01 2.22588569e-01 4.78877574e-02
1.14622056e+00 3.00825328e-01 -3.23581457e-01 -5.18197060e-01
-1.40401077e+00 5.80556989e-01 1.23078251e+00 7.00892270e-01
-6.68552756e-01 1.71628684e-01 2.42420614e-01 1.03096835e-01
-3.85525115e-02 4.16290522e-01 -1.54652849e-01 6.72045290e-01
-9.20136333e-01 -1.70900837e-01 8.58081996e-01 7.58575737e-01
5.90823591e-01 6.69360012e-02 -3.45467389e-01 7.12867081e-01
2.01036960e-01 2.90996820e-01 7.47014165e-01 -6.34966254e-01
-3.21225561e-02 9.68239307e-01 -3.70289713e-01 -5.79254627e-01
-7.01094866e-01 -5.67114830e-01 -1.17800379e+00 -6.77265227e-03
5.18440306e-01 -1.69145653e-03 -1.52037919e+00 1.48089671e+00
5.19459188e-01 3.01453620e-01 1.79758146e-01 1.04176521e+00
1.16738367e+00 3.78354222e-01 3.47035468e-01 -1.93378568e-01
1.42635381e+00 -1.17271709e+00 -3.48713309e-01 -1.34381086e-01
5.04929781e-01 -3.43997955e-01 7.43267417e-01 9.55517218e-03
-1.01271355e+00 -6.31061673e-01 -1.02647400e+00 8.20811912e-02
-2.25123182e-01 -4.85669775e-03 6.97562873e-01 3.28676224e-01
-9.42333817e-01 6.87814593e-01 -1.03532588e+00 -5.52458107e-01
7.67946064e-01 5.31021416e-01 -2.99931139e-01 1.44542307e-02
-9.76613104e-01 1.04739249e+00 6.33138597e-01 9.64511409e-02
-9.91840303e-01 -8.22486818e-01 -8.70840669e-01 9.20550451e-02
4.67845201e-01 -9.57621992e-01 1.21834338e+00 -1.04977810e+00
-1.26632774e+00 9.81723249e-01 3.44907343e-01 -6.36195600e-01
6.03183329e-01 1.31304637e-01 -1.72782496e-01 6.13852322e-01
1.27606809e-01 1.08203542e+00 8.51022542e-01 -1.28532350e+00
-3.25260073e-01 -2.53988266e-01 -4.42387424e-02 3.12115937e-01
3.00613372e-03 -3.99483830e-01 -4.32726145e-01 -6.02264166e-01
5.62560916e-01 -8.56094062e-01 -7.05144525e-01 4.67392325e-01
-5.33233225e-01 -7.49959797e-02 3.75109643e-01 -6.26944661e-01
6.89905822e-01 -2.07507133e+00 3.81448895e-01 3.28608036e-01
4.75429237e-01 3.03057693e-02 2.68146377e-02 -1.43070057e-01
-6.05675802e-02 -5.47964089e-02 -7.46403039e-01 1.08945608e-01
-3.44516873e-01 4.85463202e-01 8.41439515e-02 5.87913156e-01
-4.06030891e-03 1.22903526e+00 -1.00929415e+00 -1.07736790e+00
3.67312193e-01 2.79631943e-01 -3.97760183e-01 2.09585920e-01
-7.92110264e-02 9.98078346e-01 -6.11238539e-01 8.05210292e-01
3.67648214e-01 -5.24265051e-01 8.96061435e-02 -2.41679713e-01
4.03254688e-01 -5.18389642e-01 -8.27816367e-01 1.96221483e+00
-1.67642161e-01 1.12995028e-01 -1.10002011e-01 -1.22862613e+00
7.85197437e-01 5.86157203e-01 8.40114057e-01 -3.05650473e-01
5.12673616e-01 2.51457334e-01 3.51221532e-01 -5.69816649e-01
-7.88809955e-02 -5.97276211e-01 -3.47517505e-02 2.84054160e-01
3.60765427e-01 -6.33278131e-01 2.44123921e-01 1.75665706e-01
8.65053952e-01 -7.18999952e-02 7.31221735e-01 -2.58219332e-01
9.01340961e-01 2.38407463e-01 5.38755417e-01 6.79008245e-01
-4.94772434e-01 7.16444016e-01 2.84730405e-01 -7.24023104e-01
-8.39490652e-01 -1.55619359e+00 -2.90488690e-01 6.61830544e-01
4.43505555e-01 2.83466339e-01 -8.68120849e-01 -8.79491508e-01
-1.85770132e-02 4.29503292e-01 -9.26162541e-01 -1.09272845e-01
-6.55534506e-01 -6.12211823e-01 4.99965012e-01 7.02070236e-01
5.66139162e-01 -1.35118651e+00 -8.06813896e-01 2.60539323e-01
-6.01883866e-02 -9.77106810e-01 -3.87278110e-01 7.82122016e-02
-1.20781612e+00 -1.30003798e+00 -1.25698030e+00 -1.17722404e+00
1.12243426e+00 3.41721228e-03 1.04579973e+00 5.99045217e-01
-7.53239632e-01 5.45753419e-01 -1.04029089e-01 -3.25498998e-01
-7.99726248e-01 -1.05748707e-02 -1.65940374e-01 -1.39114410e-01
-6.03505895e-02 -5.90334117e-01 -7.49819040e-01 2.44249701e-01
-1.11548567e+00 2.57423013e-01 8.79622757e-01 8.73875558e-01
8.53301466e-01 -3.05325538e-01 5.24247766e-01 -8.58296275e-01
3.49633425e-01 -4.70718503e-01 -2.95834720e-01 6.24372363e-01
-2.60243714e-01 6.56838641e-02 6.97171152e-01 -5.93485236e-01
-9.99336600e-01 2.78592229e-01 -3.70099791e-03 -5.08163929e-01
-1.53948441e-01 4.92789716e-01 3.85461926e-01 -1.63641855e-01
6.69918418e-01 3.78311574e-01 1.20177791e-01 -1.77894652e-01
4.11988884e-01 2.45827481e-01 8.84918213e-01 -4.18343484e-01
7.11489558e-01 6.99215472e-01 1.04163140e-01 -7.21230805e-01
-8.44012439e-01 -4.97618884e-01 -1.13428450e+00 -3.83211941e-01
1.21688879e+00 -6.62813187e-01 -3.50240380e-01 1.88181549e-01
-8.99573386e-01 -1.88712031e-01 -7.51311898e-01 6.18837416e-01
-6.66608214e-01 3.95439804e-01 -9.60380316e-01 -3.65589231e-01
-6.11082315e-01 -1.32991302e+00 8.15407097e-01 6.16591990e-01
-6.33363873e-02 -1.10824990e+00 2.25880183e-02 1.90563619e-01
2.28287965e-01 3.95558566e-01 1.09530902e+00 -9.15553927e-01
-6.05270863e-01 -2.15768829e-01 -1.98091835e-01 2.71470666e-01
1.25074193e-01 -3.13376456e-01 -5.81921697e-01 -2.89334476e-01
1.55700237e-01 -2.33339250e-01 7.75229812e-01 6.87347472e-01
1.04009497e+00 7.05684721e-02 -5.13142347e-01 6.13062322e-01
1.36695123e+00 1.97300270e-01 5.52719414e-01 -6.16978630e-02
6.66453481e-01 4.27033395e-01 3.11836571e-01 1.34638995e-01
1.54042423e-01 1.55099452e-01 2.55341262e-01 -4.28453863e-01
-1.98935881e-01 -8.13297629e-02 -2.00116381e-01 9.87020254e-01
-1.49344534e-01 1.66727707e-01 -1.08232284e+00 4.84415859e-01
-1.69198143e+00 -5.95454812e-01 1.20560944e-01 1.76693833e+00
7.05111027e-01 1.59793913e-01 -2.92564370e-02 -1.57889917e-01
7.96876252e-01 -2.07757980e-01 -8.92378688e-01 -7.83088207e-02
2.00471565e-01 2.15523630e-01 4.27076519e-01 5.11869304e-02
-1.07803953e+00 9.30852652e-01 6.32846165e+00 5.52775979e-01
-1.04472792e+00 1.30035385e-01 7.33694196e-01 3.93201262e-01
4.69192453e-02 -1.16060808e-01 -8.13282132e-02 6.45806789e-02
4.02357459e-01 -3.93470168e-01 1.57028154e-01 8.55879068e-01
-2.81692863e-01 -1.21496424e-01 -1.41020846e+00 7.79178381e-01
1.74948260e-01 -1.20868194e+00 2.18219236e-01 -2.92516589e-01
7.51321077e-01 -1.25060081e-01 -1.85527533e-01 6.52815998e-01
1.09928191e-01 -1.15716672e+00 3.42494309e-01 6.97268009e-01
5.86045742e-01 -2.23836690e-01 7.12141633e-01 4.92383718e-01
-1.33630276e+00 -1.02815673e-01 -5.10576367e-01 3.60173434e-01
2.33190656e-01 9.28923488e-02 -1.18554902e+00 6.68590188e-01
4.69632626e-01 6.34616017e-01 -5.46884298e-01 1.36749709e+00
-3.31919700e-01 1.66221455e-01 -2.84908056e-01 1.30293295e-01
2.95158625e-01 -1.37226924e-01 5.45130968e-01 8.99436712e-01
1.90903232e-01 4.83537406e-01 4.59674269e-01 1.12393451e+00
-1.59337774e-01 3.92320633e-01 -3.88455540e-01 1.86003923e-01
6.97391387e-03 1.50377941e+00 -1.48299766e+00 -9.64635968e-01
-2.00849116e-01 1.03260374e+00 2.25113928e-01 2.27602705e-01
-8.80962908e-01 8.87642652e-02 -3.61191458e-04 -1.27401799e-01
1.22068137e-01 2.70016730e-01 -2.21041083e-01 -1.15668535e+00
-4.24837917e-01 -5.01187265e-01 6.46740496e-01 -7.89522350e-01
-1.42059112e+00 7.63647497e-01 -5.40141538e-02 -1.42662251e+00
-2.98335776e-02 -3.91108811e-01 -8.02239835e-01 4.83150512e-01
-1.18478489e+00 -1.16856658e+00 -5.30609369e-01 6.66995049e-01
8.08954597e-01 1.79956481e-02 7.67406464e-01 1.94663882e-01
-3.77165556e-01 2.65164971e-01 -2.07754955e-01 6.43696666e-01
3.42757493e-01 -1.35850096e+00 1.32494479e-01 5.66655815e-01
1.64005548e-01 4.13568974e-01 4.91361856e-01 -7.15941906e-01
-9.78841543e-01 -9.60045874e-01 1.67710066e-01 -1.09453313e-01
4.58445400e-01 1.82092920e-01 -1.11088049e+00 6.98042393e-01
2.37379465e-02 5.36926389e-01 6.03663623e-01 -5.90809405e-01
1.32311657e-01 2.23639593e-01 -1.43956363e+00 6.51159942e-01
8.41831088e-01 -1.32225662e-01 -1.27066863e+00 4.70222026e-01
6.43036306e-01 -8.00438941e-01 -1.15619290e+00 5.43364346e-01
5.58879673e-01 -7.00123370e-01 1.14917946e+00 -7.70750821e-01
4.40456390e-01 -7.40931854e-02 2.42246076e-01 -1.26613355e+00
-1.85712457e-01 -1.15165055e-01 1.05805553e-01 5.32647431e-01
3.66572112e-01 -4.89363164e-01 7.94527471e-01 6.90846860e-01
-2.95647413e-01 -9.91740763e-01 -9.29656744e-01 -6.59331381e-01
3.94388109e-01 -2.96129566e-03 4.06738579e-01 1.00584662e+00
2.99154762e-02 2.21667856e-01 5.89100681e-02 1.41001016e-01
8.75244498e-01 3.39691490e-01 5.93383908e-01 -1.36432242e+00
-2.56138682e-01 -4.38527048e-01 -6.56233370e-01 -8.76659036e-01
-5.73491156e-02 -1.46489394e+00 1.40833393e-01 -1.90015781e+00
4.52846408e-01 -3.95015091e-01 -6.61389589e-01 4.34843332e-01
-1.46568641e-01 3.50276947e-01 4.19699043e-01 3.79629910e-01
-5.39952278e-01 3.57323587e-01 1.91942203e+00 -4.13495839e-01
-4.38275933e-01 2.37981435e-02 -4.32928175e-01 1.00077188e+00
6.70133948e-01 -5.66736162e-01 -2.86504000e-01 -6.51442781e-02
-3.58887613e-01 5.84428906e-01 3.66661102e-01 -1.27905929e+00
4.74571317e-01 2.84355413e-02 7.35267341e-01 -6.05473042e-01
1.46357164e-01 -8.72249961e-01 2.43654586e-02 1.07637799e+00
-2.91251630e-01 -2.68588930e-01 5.16792648e-02 4.82485175e-01
-2.23496869e-01 -3.00865173e-01 1.06972241e+00 -7.21267045e-01
-6.62780344e-01 5.02721071e-01 -2.96654284e-01 1.44861832e-01
1.27353454e+00 -4.43946868e-01 1.15769021e-01 -8.35015252e-02
-1.21090519e+00 5.40990472e-01 1.99580640e-01 1.48616746e-01
9.17499781e-01 -1.10297787e+00 -7.15228200e-01 1.14241764e-01
-1.17319159e-01 3.54481161e-01 5.33676088e-01 1.35578322e+00
-8.73323202e-01 2.10352346e-01 -4.95718271e-01 -1.06250155e+00
-9.06153023e-01 6.77828848e-01 7.33581483e-01 -4.01033878e-01
-1.02812719e+00 7.99745977e-01 6.34598196e-01 -4.95144337e-01
1.07443079e-01 -6.58450842e-01 -3.21397841e-01 -2.22758979e-01
2.27190107e-01 4.23020348e-02 -3.10145319e-01 -6.49849772e-01
-3.02365094e-01 7.11176693e-01 -1.07899934e-01 2.11217523e-01
1.27181208e+00 8.03925470e-02 -3.82015742e-02 3.50911200e-01
1.03203464e+00 -5.10927141e-01 -1.11555970e+00 -3.65789205e-01
9.36174486e-03 5.53110847e-03 -2.51613885e-01 -7.15864122e-01
-1.38451803e+00 7.69591570e-01 7.00976014e-01 -1.41067445e-01
1.00267088e+00 3.29116523e-01 1.02488959e+00 4.48521703e-01
3.49593729e-01 -9.02229905e-01 2.18431681e-01 4.48409282e-02
7.69273996e-01 -1.34549129e+00 2.21767187e-01 -5.44078529e-01
-7.49312341e-01 1.20402944e+00 7.42474139e-01 -3.72199297e-01
7.11133838e-01 8.46238360e-02 2.25057498e-01 -4.09351408e-01
-2.55143613e-01 -2.29911014e-01 5.28466880e-01 3.06369901e-01
8.63774195e-02 9.10757184e-02 -2.43747145e-01 5.79647303e-01
2.72977538e-02 1.02575570e-01 3.28604609e-01 9.70358431e-01
-6.88319206e-01 -6.79395497e-01 -5.08310318e-01 6.42467558e-01
-3.59629065e-01 3.16059627e-02 -1.15621909e-01 1.00119686e+00
1.12648524e-01 2.46157169e-01 -2.16043857e-03 -2.15225331e-02
3.02037895e-01 2.74964236e-02 6.48382545e-01 -8.11756790e-01
-8.44907999e-01 1.10599950e-01 -6.60591245e-01 -3.10888320e-01
-4.35012132e-01 -3.22871387e-01 -1.94937050e+00 3.55653435e-01
-1.92365021e-01 2.70169750e-02 3.71263534e-01 1.12013137e+00
-2.85086483e-01 8.42968524e-01 2.69283444e-01 -8.17939758e-01
-4.84250337e-01 -9.01698053e-01 -6.10247254e-01 7.06875205e-01
1.38919726e-01 -7.63403952e-01 -2.06385657e-01 1.68434456e-01] | [14.657119750976562, -2.2725820541381836] |
2490e3e1-7e32-4bd4-8784-040042254587 | a-neural-corpus-indexer-for-document | 2206.02743 | null | https://arxiv.org/abs/2206.02743v3 | https://arxiv.org/pdf/2206.02743v3.pdf | A Neural Corpus Indexer for Document Retrieval | Current state-of-the-art document retrieval solutions mainly follow an index-retrieve paradigm, where the index is hard to be directly optimized for the final retrieval target. In this paper, we aim to show that an end-to-end deep neural network unifying training and indexing stages can significantly improve the recall performance of traditional methods. To this end, we propose Neural Corpus Indexer (NCI), a sequence-to-sequence network that generates relevant document identifiers directly for a designated query. To optimize the recall performance of NCI, we invent a prefix-aware weight-adaptive decoder architecture, and leverage tailored techniques including query generation, semantic document identifiers, and consistency-based regularization. Empirical studies demonstrated the superiority of NCI on two commonly used academic benchmarks, achieving +21.4% and +16.8% relative enhancement for Recall@1 on NQ320k dataset and R-Precision on TriviaQA dataset, respectively, compared to the best baseline method. | ['Mao Yang', 'Qi Zhang', 'Weiwei Deng', 'Hao Allen Sun', 'Xing Xie', 'Zheng Liu', 'Guoshuai Zhao', 'Chengmin Chi', 'Yuqing Xia', 'Qi Chen', 'Hao Sun', 'Shibin Wu', 'Ziming Miao', 'Haonan Wang', 'Yingyan Hou', 'Yujing Wang'] | 2022-06-06 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [ 2.85837233e-01 -4.21940356e-01 -5.18079698e-01 -2.56462932e-01
-1.47460902e+00 -6.88206851e-01 8.05568755e-01 1.03909485e-01
-7.21510351e-01 4.31395262e-01 3.51470381e-01 -1.83844209e-01
-2.21326619e-01 -5.02945542e-01 -8.15895140e-01 -3.70087892e-01
1.07354037e-01 8.36680353e-01 1.10767506e-01 -4.88394469e-01
6.44239843e-01 1.76777884e-01 -1.36900163e+00 5.21269143e-01
8.68795216e-01 1.15618265e+00 1.93877861e-01 6.90903783e-01
-2.09739879e-01 5.57976425e-01 -9.20458555e-01 -4.05450225e-01
1.33379042e-01 -7.97020793e-02 -8.14202189e-01 -7.96326160e-01
6.29945219e-01 -5.82643449e-01 -7.66322136e-01 8.91462743e-01
1.00440753e+00 -5.76740457e-03 8.23198855e-01 -7.77760386e-01
-1.17518294e+00 9.01594520e-01 -5.31857789e-01 2.88965613e-01
3.25334251e-01 -4.58400175e-02 1.40169585e+00 -7.63962269e-01
6.37046754e-01 1.08932662e+00 2.81547040e-01 5.58621407e-01
-9.31864262e-01 -8.40364397e-01 -1.57128707e-01 1.95361480e-01
-1.74823618e+00 -5.63067555e-01 4.71778572e-01 4.21604589e-02
1.36621833e+00 3.28301370e-01 -3.83957140e-02 1.10501671e+00
2.63792425e-01 1.20424080e+00 2.75343299e-01 -5.36410034e-01
-4.69300635e-02 -2.52039433e-01 3.89819860e-01 4.78588253e-01
2.50412345e-01 1.12702116e-01 -4.98837858e-01 -7.03858063e-02
2.06535146e-01 -2.66387671e-01 -3.13309759e-01 -8.25133100e-02
-1.01527047e+00 6.72903478e-01 5.79710126e-01 1.41502067e-01
-1.96661964e-01 4.54269499e-01 7.43363380e-01 2.15508312e-01
2.04708740e-01 7.70379603e-01 -3.76414955e-01 3.66501859e-03
-1.22047079e+00 4.82482165e-01 6.07236385e-01 1.24432325e+00
1.86812237e-01 -3.09567124e-01 -1.14988542e+00 9.41260159e-01
5.20489514e-01 9.86710668e-01 7.84677148e-01 -7.98326492e-01
7.88587630e-01 4.67333287e-01 3.37395370e-02 -8.08237851e-01
-3.43204349e-01 -7.28020072e-01 -7.22793043e-01 -7.12779582e-01
3.85437869e-02 1.48664564e-01 -1.10663593e+00 1.70677447e+00
-1.83714703e-01 -2.59384155e-01 2.84373343e-01 9.78806913e-01
9.78707075e-01 8.85000646e-01 -5.50324135e-02 5.81796132e-02
1.32449710e+00 -1.28838909e+00 -7.28189051e-01 -2.33567312e-01
8.81714404e-01 -8.57650459e-01 1.22822046e+00 1.18622035e-01
-1.27061236e+00 -4.77696508e-01 -1.15768468e+00 -5.56469023e-01
-3.03391993e-01 4.11961108e-01 3.47230226e-01 3.88283819e-01
-9.52236056e-01 1.97672769e-01 -5.24671018e-01 -5.78305423e-02
2.73995608e-01 4.77584422e-01 7.16119707e-02 -1.98039725e-01
-1.39542413e+00 4.88914251e-01 5.98660767e-01 -3.27128693e-02
-9.78483737e-01 -8.98372829e-01 -4.48114038e-01 4.86487538e-01
2.76767910e-01 -6.76745713e-01 1.53461921e+00 -2.99981594e-01
-1.28638875e+00 7.73043871e-01 -2.67249167e-01 -6.00172997e-01
2.29882710e-02 -5.44653594e-01 -5.13366818e-01 2.66513348e-01
5.47103509e-02 8.23798358e-01 5.28968632e-01 -1.09186804e+00
-5.96034527e-01 -2.93150872e-01 -2.61017811e-02 2.59275198e-01
-6.80745304e-01 1.13920890e-01 -1.39244437e+00 -6.97761297e-01
-2.43050426e-01 -8.48862827e-01 1.07650988e-01 -3.98139149e-01
-5.87982535e-01 -5.71377277e-01 4.88185614e-01 -8.03045034e-01
1.86605406e+00 -2.08122945e+00 1.49184912e-01 3.04030567e-01
-2.82498915e-02 7.57507026e-01 -6.63669884e-01 5.47428310e-01
4.71619308e-01 9.74686742e-02 1.34995962e-02 -2.97465444e-01
4.05156404e-01 -1.72910437e-01 -6.07235193e-01 4.01927345e-02
1.88933328e-01 1.31045794e+00 -7.91430831e-01 -5.33758044e-01
-4.31287348e-01 5.98964155e-01 -6.46078765e-01 3.10694546e-01
-6.03143096e-01 -2.90952832e-01 -6.50365174e-01 6.57385647e-01
3.99537593e-01 -4.09397632e-01 9.80585814e-02 -2.18607813e-01
3.28634232e-01 6.27277195e-01 -6.81394100e-01 2.18651414e+00
-4.86087501e-01 6.23406649e-01 -2.22705781e-01 -5.15958130e-01
8.58571231e-01 2.10264534e-01 1.25136465e-01 -1.64221203e+00
-5.51009327e-02 5.66054940e-01 -2.63308674e-01 -3.10363859e-01
1.13701224e+00 7.37394869e-01 -2.43806049e-01 6.33672178e-01
-1.52324751e-01 1.09133035e-01 4.13049638e-01 4.48707581e-01
1.22901440e+00 -1.01456054e-01 -3.82187545e-01 -2.31550753e-01
7.63522089e-01 -2.10357085e-03 3.62681337e-02 1.13919950e+00
2.85166830e-01 6.74139857e-01 2.34767467e-01 -8.62350836e-02
-1.05908883e+00 -7.65700698e-01 -1.34860337e-01 1.45141721e+00
3.38194042e-01 -3.43344092e-01 -8.08517158e-01 -6.15282476e-01
5.17891571e-02 7.10486054e-01 -2.72860259e-01 -6.77417457e-01
-9.43261683e-01 -6.77975893e-01 1.01658618e+00 4.50139433e-01
3.94457489e-01 -1.11776793e+00 -2.96818793e-01 2.48755082e-01
-2.53273278e-01 -9.99960780e-01 -1.10108185e+00 5.08817434e-02
-5.93014240e-01 -6.05115354e-01 -1.05222261e+00 -8.47293258e-01
4.46100026e-01 2.15903535e-01 1.46151507e+00 3.84225875e-01
-5.22828586e-02 1.87002048e-01 -3.42252702e-01 -2.57599026e-01
-2.20215812e-01 1.11000288e+00 -2.21354559e-01 -4.62677538e-01
6.06518388e-01 1.02079399e-02 -9.01862264e-01 1.50910169e-01
-1.28677058e+00 -4.37503457e-01 7.80596793e-01 9.37469006e-01
5.95779896e-01 -4.65644538e-01 6.95452452e-01 -5.37894189e-01
1.12645400e+00 -2.57616401e-01 -7.58895278e-01 9.05778229e-01
-1.17851472e+00 5.56002676e-01 5.02385259e-01 -3.12144428e-01
-7.94382393e-01 -3.23597044e-01 -9.26665515e-02 -2.69636244e-01
5.11710823e-01 7.10437357e-01 9.08107981e-02 1.23931423e-01
7.38954306e-01 4.95210558e-01 -2.25951299e-01 -5.48769712e-01
4.59699392e-01 1.14648902e+00 6.30788326e-01 -7.60592043e-01
4.14406240e-01 1.79471318e-02 -2.33219326e-01 -1.23138256e-01
-9.75293100e-01 -6.17503047e-01 -2.96816498e-01 2.36241847e-01
4.69729483e-01 -1.02545261e+00 -7.22631991e-01 2.45009378e-01
-1.25538170e+00 -2.37103492e-01 1.28870094e-02 1.38959005e-01
-1.98251769e-01 2.80108482e-01 -6.47608280e-01 -3.59898567e-01
-1.42248261e+00 -1.41964877e+00 1.54469597e+00 2.85627216e-01
-2.52080392e-02 -6.20176435e-01 1.26713231e-01 4.05927867e-01
7.49950588e-01 -6.34521008e-01 1.10251760e+00 -9.01613235e-01
-6.30326867e-01 -4.60887104e-01 -6.76244080e-01 1.40184209e-01
-3.43494415e-01 -2.32030600e-01 -8.56456280e-01 -5.48394263e-01
-6.85336113e-01 -5.42982459e-01 1.08296013e+00 1.19699650e-01
1.40444016e+00 -2.05242828e-01 -6.11650467e-01 7.57199049e-01
1.44041562e+00 4.82605517e-01 7.27848649e-01 5.67488194e-01
6.92413807e-01 1.52899817e-01 4.34624195e-01 2.61206239e-01
4.18356091e-01 1.12100530e+00 1.62126020e-01 3.71832967e-01
-4.71148372e-01 -3.43851328e-01 2.37383440e-01 8.69289696e-01
5.97291350e-01 -7.67617524e-01 -1.03437579e+00 6.33810341e-01
-1.64760649e+00 -5.37841320e-01 2.73259878e-01 2.09009910e+00
1.27582300e+00 2.02046826e-01 -3.45229864e-01 -1.88511863e-01
4.12074864e-01 6.91442564e-02 -6.45391464e-01 -2.46258110e-01
-8.12382251e-02 4.95358020e-01 6.70763910e-01 4.08557534e-01
-9.34302151e-01 1.16879439e+00 6.38095808e+00 1.34181690e+00
-1.15901923e+00 -2.24278614e-01 5.09030998e-01 -4.67572540e-01
-5.25074899e-01 -3.36833835e-01 -1.45589471e+00 4.76257712e-01
1.31040013e+00 -2.83990026e-01 5.36362827e-01 5.38199425e-01
-3.20579231e-01 3.85076880e-01 -1.22215581e+00 7.68981636e-01
2.79544413e-01 -1.41822612e+00 5.12825727e-01 -6.82274178e-02
6.71074986e-01 1.74354166e-01 3.41651827e-01 8.05894792e-01
9.42754075e-02 -1.11439824e+00 6.21781349e-01 7.95005262e-01
8.56105089e-01 -7.71679282e-01 7.39083946e-01 -1.39485355e-02
-8.65146697e-01 2.06983034e-02 -3.37511957e-01 6.37502015e-01
1.40311673e-01 3.37341785e-01 -6.72040284e-01 4.50511962e-01
5.65765500e-01 2.73589820e-01 -6.58221304e-01 1.13726914e+00
-4.38982658e-02 4.23913509e-01 -3.90218258e-01 -2.67981023e-01
5.07918537e-01 2.03543156e-01 2.78261989e-01 1.38603306e+00
3.67867738e-01 -1.88883796e-01 -1.71009138e-01 6.69924796e-01
-7.19383478e-01 6.22914918e-02 -1.71915039e-01 -2.80612916e-01
7.91878045e-01 9.67639267e-01 -2.09016085e-01 -3.47948045e-01
-1.36724755e-01 1.03662360e+00 3.89652491e-01 5.27534008e-01
-8.08535337e-01 -9.41489339e-01 4.29427952e-01 -3.30502510e-01
5.46405911e-01 -2.59211585e-02 -2.74177101e-02 -9.17616546e-01
3.54170680e-01 -1.28510821e+00 4.74300504e-01 -6.83885217e-01
-9.90359366e-01 5.45948327e-01 -7.37335980e-02 -8.78071010e-01
-4.63905960e-01 -3.79762888e-01 -1.76989194e-03 1.06217420e+00
-1.77926362e+00 -9.21697319e-01 -3.43718342e-02 2.30612516e-01
4.44466710e-01 -5.02699316e-01 8.41551781e-01 8.06019008e-01
-6.28976941e-01 1.43427765e+00 8.44001234e-01 3.07777673e-01
7.92391717e-01 -9.20103192e-01 6.92224443e-01 6.12236202e-01
2.71334529e-01 8.85867417e-01 1.43948227e-01 -4.27259296e-01
-1.93945360e+00 -1.06940866e+00 1.03599203e+00 -4.35014427e-01
4.22581911e-01 -2.58074254e-01 -8.37389410e-01 2.48382419e-01
3.50234151e-01 -1.34518445e-01 4.30222780e-01 3.42429765e-02
-6.07741535e-01 -3.80709022e-01 -7.57912278e-01 7.69647002e-01
8.91440213e-01 -7.37163424e-01 -3.37310404e-01 4.73550647e-01
1.27729952e+00 -6.63412094e-01 -9.91818488e-01 5.71310699e-01
8.06084514e-01 -2.83065051e-01 1.38208497e+00 -6.55551195e-01
4.73401010e-01 -1.34314552e-01 -3.24825853e-01 -8.63565564e-01
-7.59120509e-02 -3.69510740e-01 -7.30608702e-01 1.25287163e+00
6.49363577e-01 -2.98900250e-02 7.55327344e-01 4.13716942e-01
-3.30999851e-01 -9.11529243e-01 -7.80055940e-01 -8.15837026e-01
3.71804208e-01 -2.82844692e-01 9.59238768e-01 3.14449668e-01
-3.22583854e-01 6.29376411e-01 -1.28199592e-01 4.60845083e-02
4.56471026e-01 1.66516975e-01 5.36607385e-01 -8.54889512e-01
-2.02434868e-01 -8.80757630e-01 1.99961290e-01 -1.66277385e+00
3.42701256e-01 -1.11068618e+00 1.14253208e-01 -1.49941182e+00
2.64114499e-01 -5.89434862e-01 -5.58562636e-01 3.16570967e-01
-3.23175579e-01 2.21925870e-01 2.63360962e-02 5.12545645e-01
-1.06198800e+00 8.58056247e-01 1.20048892e+00 -7.19362497e-01
-1.75608322e-01 -3.40182453e-01 -8.54507625e-01 -3.64434749e-01
5.62795103e-01 -5.99956036e-01 -5.29162645e-01 -1.30837119e+00
4.10126358e-01 1.29718602e-01 -1.54294953e-01 -8.56062055e-01
5.19695342e-01 3.10177743e-01 9.91156995e-02 -8.58532071e-01
2.58892417e-01 -5.57777226e-01 -4.04464692e-01 3.98369133e-01
-1.06186819e+00 1.39348477e-01 3.45536053e-01 3.90578687e-01
-3.43780756e-01 -5.51906645e-01 1.75165698e-01 2.42890328e-01
-4.83956158e-01 3.18476140e-01 6.66173324e-02 4.13786739e-01
3.99311572e-01 4.19042021e-01 -7.23659337e-01 -2.71959543e-01
1.67386085e-01 6.09476447e-01 -1.26038063e-02 6.60914242e-01
5.31372130e-01 -1.38657784e+00 -8.01158190e-01 9.57845375e-02
4.39237952e-01 -6.65810928e-02 -4.99889627e-02 3.46075237e-01
-6.45578921e-01 1.39193535e+00 1.62245914e-01 -4.38366681e-01
-9.46094215e-01 5.77789187e-01 1.52635023e-01 -7.05940485e-01
-3.60302776e-01 9.87015247e-01 -2.25542277e-01 -4.66688335e-01
8.07345450e-01 -2.12859452e-01 -2.86394089e-01 -1.76038265e-01
8.28560889e-01 1.42042369e-01 4.70799714e-01 -3.88129503e-01
-2.52565175e-01 4.71671015e-01 -7.62479544e-01 -1.39465660e-01
9.22724485e-01 2.53795274e-02 -4.29064110e-02 -2.70894706e-01
1.72342598e+00 -3.88059914e-01 -5.98265767e-01 -6.46653771e-01
3.80126029e-01 -1.38482168e-01 4.26679313e-01 -1.10113823e+00
-1.26624572e+00 8.23356092e-01 8.14205587e-01 -1.02961026e-01
1.23700929e+00 -2.75713563e-01 1.40235317e+00 1.03274858e+00
2.01369762e-01 -9.95075583e-01 5.48343360e-02 8.81704152e-01
8.66245151e-01 -1.17411792e+00 -3.48138027e-02 2.66187966e-01
-2.59257585e-01 9.23354924e-01 5.55523515e-01 1.17467910e-01
3.50013584e-01 -5.26097901e-02 -6.27921894e-02 -4.28509027e-01
-8.54542196e-01 1.14431819e-02 9.84989107e-01 1.72685608e-01
6.55414462e-01 -3.45069468e-01 -4.17724848e-01 4.75569576e-01
-1.65728033e-01 6.13434762e-02 -7.77469054e-02 8.44621897e-01
-2.59801447e-01 -1.20122564e+00 -1.13494471e-01 5.41894257e-01
-8.30391467e-01 -6.45270288e-01 -1.48130730e-01 4.37324762e-01
-6.05233908e-01 7.94923365e-01 2.35589921e-01 -3.69530410e-01
3.89601856e-01 8.45549703e-02 2.99086213e-01 -3.38948190e-01
-7.74894655e-01 -8.44354630e-02 5.77565618e-02 -5.27971148e-01
-9.68771875e-02 -2.30910599e-01 -1.13777280e+00 -1.39672443e-01
-4.97482687e-01 2.82510549e-01 7.60079920e-01 7.03903019e-01
8.66685331e-01 6.22466326e-01 4.44451600e-01 -3.65503073e-01
-9.18412507e-01 -1.06357872e+00 -7.54792616e-02 1.92178175e-01
2.85377800e-01 -3.00447315e-01 -6.59848303e-02 -3.11067015e-01] | [11.446245193481445, 7.63766622543335] |
b28561d8-306f-49be-97c1-9f7a47e025e5 | isia-food-500-a-dataset-for-large-scale-food | 2008.05655 | null | https://arxiv.org/abs/2008.05655v1 | https://arxiv.org/pdf/2008.05655v1.pdf | ISIA Food-500: A Dataset for Large-Scale Food Recognition via Stacked Global-Local Attention Network | Food recognition has received more and more attention in the multimedia community for its various real-world applications, such as diet management and self-service restaurants. A large-scale ontology of food images is urgently needed for developing advanced large-scale food recognition algorithms, as well as for providing the benchmark dataset for such algorithms. To encourage further progress in food recognition, we introduce the dataset ISIA Food- 500 with 500 categories from the list in the Wikipedia and 399,726 images, a more comprehensive food dataset that surpasses existing popular benchmark datasets by category coverage and data volume. Furthermore, we propose a stacked global-local attention network, which consists of two sub-networks for food recognition. One subnetwork first utilizes hybrid spatial-channel attention to extract more discriminative features, and then aggregates these multi-scale discriminative features from multiple layers into global-level representation (e.g., texture and shape information about food). The other one generates attentional regions (e.g., ingredient relevant regions) from different regions via cascaded spatial transformers, and further aggregates these multi-scale regional features from different layers into local-level representation. These two types of features are finally fused as comprehensive representation for food recognition. Extensive experiments on ISIA Food-500 and other two popular benchmark datasets demonstrate the effectiveness of our proposed method, and thus can be considered as one strong baseline. The dataset, code and models can be found at http://123.57.42.89/FoodComputing-Dataset/ISIA-Food500.html. | ['Zhiling Wang', 'Xiaolin Wei', 'Weiqing Min', 'Zhengdong Luo', 'Xiaoming Wei', 'Linhu Liu', 'Shuqiang Jiang'] | 2020-08-13 | null | null | null | null | ['food-recognition'] | ['computer-vision'] | [ 2.50720561e-01 -4.83788252e-01 -4.41367269e-01 -4.82686192e-01
-7.11793244e-01 -5.25811851e-01 1.45621434e-01 8.14255118e-01
-2.84568459e-01 1.08884990e-01 5.32411814e-01 5.20767346e-02
4.24723588e-02 -1.31273985e+00 -1.07998383e+00 -9.88268614e-01
-1.72921091e-01 -1.85144365e-01 1.18104391e-01 -1.62760586e-01
-4.42450047e-02 3.13861072e-02 -1.77907801e+00 5.83866656e-01
8.32094252e-01 1.31851351e+00 5.61204791e-01 3.42658103e-01
-1.91240534e-01 5.68964660e-01 -3.68286856e-02 -1.22245923e-01
1.72718763e-01 -4.07632768e-01 -4.38642174e-01 2.27012709e-01
3.54223609e-01 -3.74231458e-01 -2.43959874e-01 1.58749938e+00
3.53545547e-01 1.19237550e-01 4.23278570e-01 -8.49968791e-01
-1.48330295e+00 9.08881426e-01 -6.33703172e-01 2.85620093e-01
4.84415084e-01 3.02197158e-01 6.91710949e-01 -7.50447392e-01
1.55195460e-01 1.35294974e+00 3.71685207e-01 5.91926761e-02
-9.20205653e-01 -8.42930913e-01 5.27659595e-01 2.52288967e-01
-1.27337229e+00 -3.00891936e-01 6.53475881e-01 -3.95443320e-01
8.26811373e-01 2.10041150e-01 9.22573626e-01 9.04056609e-01
-8.88754334e-03 1.10064352e+00 8.37076843e-01 -3.03401887e-01
-1.81234717e-01 -9.00553726e-03 5.33331692e-01 8.77635479e-01
6.93195581e-01 1.42489389e-01 -2.52276033e-01 2.83805460e-01
8.71555805e-01 4.68451053e-01 -2.57049650e-01 -3.00240248e-01
-1.54359627e+00 1.07416034e+00 1.08041728e+00 3.33508790e-01
-6.24167264e-01 -1.57700628e-01 4.14489061e-01 2.27384433e-01
4.85325217e-01 -7.94061124e-02 -3.58170390e-01 5.79210639e-01
-2.25550175e-01 7.78823197e-02 3.78760576e-01 9.77311850e-01
7.01040983e-01 -3.00810561e-02 -1.99324086e-01 1.00380528e+00
6.40462875e-01 1.02893627e+00 5.83780289e-01 -4.94271725e-01
5.48815906e-01 8.89320493e-01 9.17009041e-02 -1.44859028e+00
-6.59244776e-01 -3.50573003e-01 -1.14330029e+00 -3.14915687e-01
3.32872778e-01 3.41898352e-01 -8.95697057e-01 1.50604784e+00
5.14009893e-01 -1.03973232e-01 1.62618250e-01 1.31690192e+00
1.60020649e+00 9.69648004e-01 5.70269823e-01 6.12205938e-02
1.87716699e+00 -1.25574708e+00 -4.25602168e-01 -9.41011012e-02
3.78668725e-01 -8.86728942e-01 9.46436405e-01 2.12006569e-01
-9.21996772e-01 -8.59178066e-01 -1.14887953e+00 -2.31958881e-01
-8.17823470e-01 2.36022562e-01 8.27472150e-01 2.48510957e-01
-6.84800088e-01 1.65482432e-01 -5.60864449e-01 -6.87886417e-01
4.98366982e-01 3.03451687e-01 -2.04455480e-01 -5.92533886e-01
-1.18456435e+00 2.83408999e-01 6.66882634e-01 2.06023470e-01
-7.22510695e-01 -5.19602656e-01 -1.40751088e+00 8.72951560e-03
1.72991499e-01 -7.02975690e-01 9.32847321e-01 -1.03384459e+00
-1.14636099e+00 9.06735957e-01 2.83937901e-01 -8.99024978e-02
-1.35918885e-01 -1.62787326e-02 -8.01362038e-01 1.05804153e-01
4.08407986e-01 1.05237794e+00 4.31604624e-01 -9.96345103e-01
-8.92187893e-01 -5.57516575e-01 3.28193545e-01 1.42186537e-01
-3.75259399e-01 8.41437057e-02 -4.69857872e-01 -7.59116232e-01
1.61867872e-01 -5.86489260e-01 -1.67135879e-01 9.59089696e-02
-4.25399691e-01 -3.45703304e-01 2.00798377e-01 -7.12079644e-01
1.13157296e+00 -2.29416299e+00 -2.42175981e-01 8.23592246e-02
-2.37643108e-01 1.75543115e-01 -6.73528850e-01 -1.60308033e-02
-1.99800476e-01 -1.98848471e-01 -1.69675909e-02 4.51190352e-01
-1.27327498e-02 -1.29522353e-01 1.54514283e-01 4.67276305e-01
2.85166860e-01 1.16479683e+00 -1.25004959e+00 -3.58939022e-01
4.84765500e-01 6.04051113e-01 -3.17275077e-01 -1.58980221e-01
-4.05023903e-01 1.91080227e-01 -9.29133117e-01 9.61995721e-01
6.99127674e-01 -4.25704271e-01 -1.13085002e-01 -7.77707279e-01
-3.10386956e-01 -1.16292745e-01 -1.03605199e+00 1.92007351e+00
-2.18689501e-01 -2.48356134e-01 7.09634423e-02 -1.06320190e+00
9.14333642e-01 -6.17579184e-02 5.77886343e-01 -1.15619552e+00
4.31230873e-01 1.87879309e-01 -1.06848434e-01 -6.98036134e-01
4.34931427e-01 3.10622752e-01 -3.49832803e-01 8.39422736e-03
-1.40646249e-02 6.61037743e-01 6.90515101e-01 -2.04725817e-01
5.59505403e-01 2.91930944e-01 3.92176390e-01 -6.62179768e-01
7.28723824e-01 2.46911317e-01 4.50789034e-01 3.77435803e-01
-3.79426003e-01 2.66729236e-01 -3.42759401e-01 -7.85537302e-01
-8.48233163e-01 -7.95319021e-01 -8.03992972e-02 1.20437336e+00
4.49674964e-01 -6.75869212e-02 -5.27457058e-01 -6.14025533e-01
3.30719143e-01 2.85186261e-01 -6.86945796e-01 -3.58390719e-01
-4.86701429e-01 -8.67975295e-01 3.56514305e-02 7.59460926e-01
8.94384325e-01 -1.48819184e+00 -4.49631214e-01 5.39368629e-01
-4.97596681e-01 -6.24836624e-01 -9.39830601e-01 1.46568343e-01
-5.42537272e-01 -1.43276918e+00 -8.93180907e-01 -1.31575775e+00
5.11750460e-01 1.02856231e+00 1.34129417e+00 1.01441711e-01
-3.80717963e-01 2.39052281e-01 -6.18135571e-01 -3.79245549e-01
-2.49141470e-01 -4.82790358e-02 -2.67312676e-01 1.02789633e-01
9.41043139e-01 -1.51242027e-02 -1.04832697e+00 3.10146332e-01
-1.12194848e+00 5.95118515e-02 5.58377206e-01 6.17112100e-01
1.07283378e+00 8.53079185e-03 7.23798931e-01 -4.75271106e-01
1.21354416e-01 -8.28388214e-01 -6.52697921e-01 2.71214128e-01
-4.82583605e-02 -1.95463508e-01 6.68488920e-01 -5.27681887e-01
-8.01524222e-01 2.53493994e-01 -3.33720148e-01 -5.03178723e-02
-4.34668094e-01 6.99039638e-01 -4.34803754e-01 7.40111768e-02
2.59619027e-01 4.04622912e-01 -1.82329744e-01 -6.36495531e-01
3.47114235e-01 5.71867943e-01 5.27739525e-01 -4.16778505e-01
3.68895113e-01 1.33560672e-01 -3.64249706e-01 -7.08533049e-01
-1.13251495e+00 -7.33487844e-01 -3.63031238e-01 -1.55363798e-01
1.12114239e+00 -1.30156076e+00 -7.66378343e-01 6.69646621e-01
-6.98383808e-01 -1.74615189e-01 -1.50780931e-01 5.57115376e-01
-2.36018866e-01 1.18912168e-01 -8.74266982e-01 -2.34848842e-01
-4.94488746e-01 -1.08339274e+00 1.03941500e+00 6.89687371e-01
2.56765425e-01 -7.31798232e-01 -2.82256454e-01 3.76496941e-01
2.61230081e-01 4.27624702e-01 9.40989673e-01 -4.20604348e-01
-5.43449819e-01 1.32509872e-01 -5.96450269e-01 -3.89293395e-02
4.21215564e-01 -2.14780524e-01 -5.59560597e-01 -3.01012993e-01
-4.88250881e-01 -4.46726799e-01 1.26251936e+00 7.19758034e-01
1.14212120e+00 -1.63362220e-01 -3.09800744e-01 4.78309959e-01
1.62566483e+00 5.49062133e-01 2.03131020e-01 4.77320373e-01
8.58845890e-01 4.13626224e-01 7.05594718e-01 2.88712412e-01
7.93375969e-01 5.28583467e-01 5.53134501e-01 -4.38884765e-01
-2.92442679e-01 -3.56525868e-01 5.90804875e-01 1.25145137e+00
2.61007190e-01 -6.73908042e-03 -2.53187180e-01 6.67951643e-01
-1.76794553e+00 -8.61678660e-01 -2.16459140e-01 1.97417068e+00
5.73004782e-01 -3.18620801e-01 5.31976700e-01 -2.32124571e-02
7.82622218e-01 9.04295594e-02 -9.69446182e-01 -6.20668754e-02
-2.76497215e-01 -1.64013937e-01 5.05333483e-01 -2.97429301e-02
-1.83466446e+00 5.49387157e-01 4.84317017e+00 7.15270102e-01
-8.88939917e-01 7.72557259e-02 8.13548744e-01 1.01979665e-01
-3.12236045e-02 -8.95391881e-01 -6.60981417e-01 6.34500444e-01
7.48138309e-01 1.63943380e-01 5.15654087e-01 8.23954403e-01
7.74396509e-02 1.26030698e-01 -7.93448269e-01 8.56421590e-01
2.72192836e-01 -1.05706286e+00 2.09075779e-01 -1.00635044e-01
6.95013046e-01 6.33021072e-02 9.49724987e-02 4.02349442e-01
6.45821750e-01 -6.55363083e-01 9.75220799e-01 3.16391438e-01
8.47542822e-01 -7.55266547e-01 7.43786216e-01 2.24826615e-02
-2.04249930e+00 -3.03149402e-01 -7.60838628e-01 2.06528530e-01
-2.63041079e-01 6.04151547e-01 3.05231780e-01 8.01578224e-01
1.08122420e+00 1.18831825e+00 -5.15615165e-01 1.21406353e+00
9.54215005e-02 3.79481167e-01 -5.64108610e-01 -1.56225950e-01
5.20331681e-01 -3.87862325e-01 -1.26803339e-01 1.14851356e+00
4.90282148e-01 3.81823033e-01 8.10945570e-01 6.81143880e-01
-9.22130421e-02 6.60229146e-01 -4.43390667e-01 -1.41295493e-01
7.55588040e-02 1.42590785e+00 -1.02778077e+00 -4.94817883e-01
-9.05535758e-01 6.86531305e-01 9.23927501e-02 1.63847134e-01
-7.50166893e-01 -1.46082297e-01 6.26908004e-01 -2.57871896e-01
6.12265408e-01 2.40250602e-01 3.81273538e-01 -1.28823841e+00
-1.41372845e-01 -8.76128972e-01 7.55779803e-01 -6.86757386e-01
-1.62058127e+00 4.36732590e-01 -1.67175725e-01 -1.02012873e+00
4.24354166e-01 -5.64575195e-01 -2.13946939e-01 6.27160311e-01
-1.84806836e+00 -1.50990283e+00 -5.49742818e-01 7.59041071e-01
7.53819764e-01 1.61551192e-01 9.50022340e-01 7.70046473e-01
-7.85211444e-01 5.06132901e-01 1.78832546e-01 4.06061709e-01
4.31327134e-01 -1.01395571e+00 3.26984912e-01 4.74869102e-01
1.64316714e-01 4.22132462e-01 1.95314378e-01 -6.52830541e-01
-1.62659407e+00 -1.73179758e+00 6.49131238e-01 9.39088911e-02
5.15178680e-01 -1.47374213e-01 -8.45635116e-01 3.44868302e-01
1.13241903e-01 5.11094294e-02 6.09682977e-01 -2.40540564e-01
-4.84748155e-01 -4.77062792e-01 -1.15629292e+00 8.26337039e-02
1.06482947e+00 -7.65611753e-02 -2.58522570e-01 5.68004847e-01
8.89250755e-01 -1.62656114e-01 -1.04695821e+00 3.79530519e-01
5.54952025e-01 -7.58916259e-01 1.18364131e+00 -2.22440049e-01
4.61490124e-01 -4.37556267e-01 -4.24744338e-01 -1.28854322e+00
-1.00027812e+00 2.59314895e-01 1.40368268e-01 1.34070563e+00
1.76562652e-01 -6.29781902e-01 4.32060301e-01 -1.82056263e-01
-4.04126078e-01 -7.26191163e-01 -2.90735662e-01 -4.42517906e-01
1.61565579e-02 7.79944286e-02 1.03162539e+00 9.64643955e-01
-2.75663674e-01 1.69814810e-01 -4.50059399e-02 2.98825443e-01
5.94012678e-01 7.87202358e-01 1.84779808e-01 -1.02226341e+00
2.35490203e-02 -4.43053901e-01 -2.67819703e-01 -1.30946112e+00
-1.74081430e-01 -1.17165577e+00 2.66339332e-01 -1.67169154e+00
6.02746665e-01 -2.97478586e-01 -9.30484474e-01 7.32391179e-01
-2.44141519e-01 5.96868038e-01 1.91313684e-01 7.90850371e-02
-7.44791687e-01 3.90507340e-01 1.47807682e+00 -7.58509576e-01
-1.52851213e-02 -3.64729822e-01 -1.03375959e+00 6.69034302e-01
7.88373172e-01 -1.63976043e-01 -3.26247633e-01 -4.50837344e-01
-2.54972447e-02 -2.67644763e-01 4.29531157e-01 -1.03240049e+00
-1.14507444e-01 -2.20420018e-01 7.81834185e-01 -6.36176288e-01
-1.52501211e-01 -9.21117902e-01 1.82625547e-01 6.89887464e-01
-3.03256273e-01 -6.42264932e-02 1.36903450e-01 5.44828773e-01
-3.77664298e-01 1.24036409e-01 5.99339068e-01 -4.62911755e-01
-1.09066856e+00 5.82627118e-01 9.01099946e-03 -2.96686292e-01
9.31199491e-01 -4.45236918e-03 -4.51362669e-01 2.20511898e-01
-6.76743269e-01 4.18908089e-01 2.90584236e-01 8.63397121e-01
4.53888923e-01 -1.85415399e+00 -8.37977529e-01 5.81001401e-01
5.29330492e-01 -1.21863253e-01 6.55509830e-01 4.75140482e-01
-3.64106327e-01 7.47268677e-01 -3.09449613e-01 -5.76575696e-01
-1.04815042e+00 1.06774426e+00 1.39905084e-02 -1.14748314e-01
-7.96915829e-01 6.20080590e-01 6.15935266e-01 -2.20706329e-01
1.99157178e-01 -8.89966130e-01 -6.23538315e-01 3.95537354e-02
1.03279078e+00 -2.72257011e-02 -1.03602558e-01 -1.02123797e+00
-3.42076153e-01 9.35895383e-01 9.80930179e-02 9.87835169e-01
1.46198893e+00 -4.23798919e-01 1.07328549e-01 3.11023355e-01
1.11828709e+00 -5.83511114e-01 -1.14997041e+00 -3.99598151e-01
-4.25946951e-01 -3.24445724e-01 -7.59897232e-02 -1.00123727e+00
-1.65114093e+00 6.91492379e-01 9.86644208e-01 3.53070408e-01
1.45095575e+00 1.06276557e-01 1.18642652e+00 2.72730514e-02
5.65879703e-01 -6.47556543e-01 -5.23898564e-02 1.56472936e-01
8.22221279e-01 -1.38784766e+00 -2.16176271e-01 -6.27317667e-01
-1.86852634e-01 9.23354506e-01 5.13729870e-01 -2.51709670e-01
7.71102786e-01 -1.34060001e-02 2.48708203e-01 -1.72979623e-01
-2.11640328e-01 -3.49180371e-01 4.28228915e-01 6.99822605e-01
4.10111248e-01 4.25342619e-01 -1.18495170e-02 1.06728375e+00
1.69100493e-01 1.48064747e-01 -2.00655922e-01 7.22583532e-01
-6.66258514e-01 -8.00551713e-01 -4.89473671e-01 5.25129557e-01
-1.42623901e-01 -1.91870376e-01 3.55616271e-01 5.15491605e-01
6.11139953e-01 1.14229083e+00 3.37073281e-02 -2.71534145e-01
5.21065891e-01 -1.93187624e-01 4.38521475e-01 -3.40224892e-01
-8.03244352e-01 4.16438937e-01 -1.41829122e-02 -6.83493853e-01
-9.15420711e-01 -6.76755011e-01 -1.14578748e+00 -2.69954801e-01
-1.51189536e-01 -9.38889906e-02 4.84134078e-01 5.94827592e-01
2.41804525e-01 6.14712715e-01 5.03053665e-01 -9.20691788e-01
-2.86400408e-01 -8.34787726e-01 -6.62807882e-01 8.49821448e-01
1.16609424e-01 -7.87105024e-01 5.05806506e-02 3.04868937e-01] | [11.540565490722656, 4.374851703643799] |
2d258325-0ece-4c3e-a2eb-5b7bb0d14d03 | unsupervised-foreground-background | 2104.00483 | null | https://arxiv.org/abs/2104.00483v2 | https://arxiv.org/pdf/2104.00483v2.pdf | Learning Foreground-Background Segmentation from Improved Layered GANs | Deep learning approaches heavily rely on high-quality human supervision which is nonetheless expensive, time-consuming, and error-prone, especially for image segmentation task. In this paper, we propose a method to automatically synthesize paired photo-realistic images and segmentation masks for the use of training a foreground-background segmentation network. In particular, we learn a generative adversarial network that decomposes an image into foreground and background layers, and avoid trivial decompositions by maximizing mutual information between generated images and latent variables. The improved layered GANs can synthesize higher quality datasets from which segmentation networks of higher performance can be learned. Moreover, the segmentation networks are employed to stabilize the training of layered GANs in return, which are further alternately trained with Layered GANs. Experiments on a variety of single-object datasets show that our method achieves competitive generation quality and segmentation performance compared to related methods. | ['Xiangyang Ji', 'Wing Yin Cheung', 'Qiran Zou', 'Hakan Bilen', 'Yu Yang'] | 2021-04-01 | null | null | null | null | ['unsupervised-object-segmentation'] | ['computer-vision'] | [ 7.37775624e-01 4.77584958e-01 6.56556115e-02 -2.72097468e-01
-1.14331102e+00 -5.66477776e-01 4.52279866e-01 -7.07721293e-01
-1.14705518e-01 8.83140206e-01 -2.91511208e-01 -1.84210911e-01
5.10659039e-01 -1.05824709e+00 -1.00279570e+00 -1.02283204e+00
5.08697331e-01 5.48425078e-01 4.36683036e-02 2.15688005e-01
-2.21955448e-01 3.20477366e-01 -1.33158481e+00 1.91775575e-01
1.26628578e+00 9.38175917e-01 2.65395880e-01 7.31854856e-01
-9.58174020e-02 8.13004434e-01 -7.93088734e-01 -6.46413624e-01
5.31698167e-01 -8.90618086e-01 -7.62098134e-01 7.13422239e-01
3.53866279e-01 -4.88121390e-01 -5.02011441e-02 1.02265227e+00
2.31204078e-01 -1.63512817e-03 6.88993871e-01 -1.30096114e+00
-6.80104017e-01 5.47858298e-01 -7.01959550e-01 -2.99227685e-01
-6.61693215e-02 5.64262152e-01 7.79000521e-01 -5.94467163e-01
4.21797663e-01 1.10576558e+00 3.96919608e-01 8.69935095e-01
-1.45805597e+00 -6.62398279e-01 1.27578557e-01 -3.51498961e-01
-1.10999095e+00 -3.71904880e-01 9.16040838e-01 -4.65181440e-01
2.31079906e-01 1.61066309e-01 6.50098085e-01 1.22483599e+00
-1.56003326e-01 9.27094400e-01 1.12107265e+00 -2.78344601e-01
1.86082259e-01 9.54038650e-02 -4.97490197e-01 9.01006937e-01
2.42379084e-01 -1.77159637e-01 -1.06315635e-01 2.00554371e-01
1.18179226e+00 -6.64418861e-02 -2.96068937e-01 -2.35459149e-01
-1.11361933e+00 8.29911053e-01 5.92375398e-01 1.74660794e-02
-5.79968512e-01 1.52581066e-01 -1.60165429e-01 -1.13336109e-01
5.81251681e-01 4.10611510e-01 -1.26826048e-01 2.18265533e-01
-1.23765612e+00 1.01451308e-01 3.82763773e-01 1.09178066e+00
1.04269671e+00 4.68161106e-01 -2.66266793e-01 8.57896984e-01
3.02682638e-01 5.47850847e-01 2.98767716e-01 -1.09287536e+00
5.05272031e-01 6.54850245e-01 1.58808082e-01 -7.65407801e-01
1.47059247e-01 -3.56095821e-01 -1.15827382e+00 3.21795762e-01
3.69497985e-01 -2.30177835e-01 -1.44368708e+00 1.64364636e+00
3.90744388e-01 3.22897464e-01 1.52695611e-01 8.25581789e-01
6.53406680e-01 7.75944293e-01 9.84779522e-02 -1.18765876e-01
8.86979997e-01 -1.40005910e+00 -6.73309565e-01 -4.49694127e-01
1.63410828e-01 -6.85863733e-01 1.20185149e+00 2.25209460e-01
-1.50475001e+00 -7.80933201e-01 -1.03890407e+00 -1.42315984e-01
-2.69792322e-02 4.19704616e-01 6.21197045e-01 5.99540532e-01
-9.69730437e-01 4.78466779e-01 -1.10512197e+00 2.18991384e-01
8.62027526e-01 1.62553698e-01 2.02295594e-02 -2.00885218e-02
-8.22684228e-01 4.68809217e-01 4.72999215e-01 3.91245842e-01
-1.41948628e+00 -5.01790762e-01 -9.46469069e-01 -9.07380655e-02
2.07810447e-01 -9.78332102e-01 1.09539258e+00 -1.44200492e+00
-1.91773570e+00 9.68274057e-01 -8.19040369e-03 -3.92601192e-01
8.51153195e-01 2.43304539e-02 -4.81902361e-02 2.39654198e-01
2.98372269e-01 1.02887464e+00 1.28459287e+00 -1.68564296e+00
-5.31253815e-01 -5.55877239e-02 1.49137378e-02 9.09986421e-02
1.46614732e-02 -3.92098784e-01 -6.97285056e-01 -7.21481264e-01
1.78938612e-01 -9.31694567e-01 -5.01234114e-01 -1.68159589e-01
-7.85794199e-01 8.03684294e-02 7.67104745e-01 -8.92237425e-01
6.83282852e-01 -2.05122328e+00 3.91769677e-01 1.03061602e-01
1.40058622e-01 2.36821264e-01 -2.23869711e-01 -1.81279853e-01
8.68439302e-02 3.17802399e-01 -6.91605747e-01 -6.88081682e-01
-2.00776055e-01 4.66142565e-01 -4.08541471e-01 2.23616466e-01
5.66048801e-01 1.23977172e+00 -7.98881292e-01 -6.73648417e-01
1.99506372e-01 4.25855190e-01 -4.15319532e-01 7.74521232e-01
-6.22654557e-01 1.05702245e+00 -3.80016714e-01 6.21104240e-01
6.74250245e-01 -5.18335581e-01 5.97104803e-02 -9.74701867e-02
3.57185811e-01 1.37132585e-01 -8.30548346e-01 1.76648986e+00
-4.93642926e-01 3.87509316e-01 5.02614230e-02 -9.71349537e-01
8.29870582e-01 1.92227408e-01 2.26259455e-01 -3.91642272e-01
2.17202321e-01 7.06372410e-03 -2.31410846e-01 -2.02177003e-01
2.01196149e-01 -1.25283733e-01 -7.35340565e-02 3.92488867e-01
6.71417639e-02 -5.01326501e-01 2.28638187e-01 1.73695967e-01
7.03014433e-01 5.82032979e-01 -2.24054515e-01 1.28681406e-01
5.66313565e-01 -1.37727454e-01 7.19477117e-01 4.07308608e-01
2.53716230e-01 1.04667318e+00 4.97448474e-01 -2.27723256e-01
-1.30144286e+00 -1.20676184e+00 1.88035920e-01 6.25609100e-01
1.79117680e-01 -9.52887759e-02 -1.33954537e+00 -9.15189743e-01
-4.37388420e-01 7.42407262e-01 -5.42351305e-01 -1.47565994e-02
-6.28070354e-01 -7.34021902e-01 5.74076891e-01 6.30857527e-01
7.77497709e-01 -1.22866714e+00 -3.80170554e-01 1.54292718e-01
-4.70656484e-01 -1.33309710e+00 -4.05195028e-01 -7.78438943e-03
-9.23879564e-01 -9.32317972e-01 -1.02882195e+00 -8.18531573e-01
1.10155487e+00 1.60364732e-02 1.23589718e+00 2.13555112e-01
-1.92163467e-01 -8.93686786e-02 -1.36027440e-01 -3.06770056e-01
-6.65628910e-01 1.69204697e-01 -4.96425122e-01 2.87691742e-01
-3.12825888e-01 -5.86856961e-01 -7.07230210e-01 3.02722663e-01
-1.11226141e+00 6.54483438e-01 8.61410141e-01 9.26060498e-01
9.17739928e-01 1.03940770e-01 4.04974431e-01 -1.16844559e+00
1.57671466e-01 -1.91285029e-01 -9.83832777e-01 2.72240192e-01
-3.91823113e-01 4.73945849e-02 6.59302413e-01 -2.85961837e-01
-1.44535136e+00 2.72218525e-01 -1.81334153e-01 -5.77075183e-01
-2.94797152e-01 5.03890067e-02 -5.89506865e-01 3.12523581e-02
4.66337949e-01 3.33274782e-01 -2.45571598e-01 -2.97011077e-01
5.91016412e-01 3.50217521e-01 7.03984201e-01 -5.62073529e-01
1.13294053e+00 4.90011990e-01 -2.20888272e-01 -4.53229994e-01
-1.03950143e+00 1.51697353e-01 -7.16669023e-01 -1.34987816e-01
1.30260861e+00 -9.45262492e-01 -1.12521693e-01 8.29027534e-01
-1.13735998e+00 -8.19011629e-01 -2.93021709e-01 4.30527814e-02
-7.13639319e-01 1.92017660e-01 -8.17564249e-01 -6.89360380e-01
-2.52713978e-01 -1.27567029e+00 1.29987061e+00 5.15824080e-01
2.62851506e-01 -9.07710016e-01 -2.64435500e-01 8.94374669e-01
1.09110333e-01 6.89507067e-01 7.31424510e-01 -6.20057918e-02
-1.23123133e+00 2.21782643e-02 -2.93206543e-01 6.59923732e-01
3.02831352e-01 1.34739891e-01 -1.04809296e+00 -2.69069463e-01
-1.32767305e-01 -4.97561723e-01 8.02546740e-01 3.48763198e-01
1.49144173e+00 -4.81606275e-01 -2.88569063e-01 9.81520236e-01
1.27494919e+00 1.90677151e-01 9.05796409e-01 -1.25317611e-02
1.18484998e+00 4.52507198e-01 4.70053792e-01 2.12606997e-03
2.73475051e-01 3.03568214e-01 3.41787606e-01 -7.29796052e-01
-2.72670716e-01 -4.28534687e-01 1.58479318e-01 7.25613058e-01
-1.13581531e-01 -4.30393398e-01 -6.73055112e-01 5.36776960e-01
-1.69883847e+00 -7.56564319e-01 3.12985815e-02 1.89907181e+00
1.09471405e+00 1.61772668e-01 6.13062382e-02 -4.19045165e-02
7.46751249e-01 1.43414915e-01 -6.19251251e-01 -4.59947390e-03
-9.80251729e-02 3.40891808e-01 4.72772002e-01 3.81552011e-01
-1.06836843e+00 1.20093536e+00 6.14993858e+00 6.82586014e-01
-1.13391531e+00 1.44647658e-01 1.19316268e+00 8.15407336e-02
-4.38980877e-01 -7.48746619e-02 -5.17346382e-01 6.34462118e-01
6.20877802e-01 2.76578516e-01 4.85776305e-01 8.41571629e-01
-4.45308723e-02 -2.92730276e-02 -1.01905751e+00 8.35119665e-01
-2.78593674e-02 -1.30131805e+00 2.17792034e-01 1.49725854e-01
1.37683308e+00 -3.00813913e-01 1.26004502e-01 7.51632378e-02
7.11822748e-01 -1.14257622e+00 6.21007979e-01 5.10876119e-01
7.56965280e-01 -7.55503416e-01 4.98148322e-01 4.24699575e-01
-8.49760234e-01 3.08502346e-01 -2.39143893e-01 3.09941173e-01
3.96268368e-01 6.42115831e-01 -7.57645249e-01 6.23877943e-01
5.99509776e-01 4.39842343e-01 -4.83327270e-01 6.24647975e-01
-7.96453655e-01 7.54367292e-01 -5.90640418e-02 5.29268980e-01
2.90054619e-01 -5.92977524e-01 1.47511095e-01 8.73378813e-01
1.50531620e-01 5.33191785e-02 3.16725224e-01 1.35788727e+00
-3.74299407e-01 -2.73672104e-01 -5.49863100e-01 -1.74934283e-01
7.32577443e-02 1.55718720e+00 -1.08796382e+00 -5.77737033e-01
-2.03814834e-01 1.40186071e+00 3.11248720e-01 6.11010194e-01
-1.16926372e+00 -9.66072381e-02 3.65922242e-01 -2.15451941e-02
4.13462996e-01 -6.49566725e-02 -4.79704082e-01 -1.36014712e+00
7.08226487e-02 -1.02385056e+00 6.12193458e-02 -9.09394681e-01
-1.29611897e+00 7.85868883e-01 -2.68137842e-01 -1.11125517e+00
-4.44318742e-01 -2.63784111e-01 -7.38407135e-01 9.85881090e-01
-1.57941508e+00 -1.44410241e+00 -5.69270134e-01 5.59974551e-01
5.88638246e-01 1.70773570e-03 6.12379253e-01 2.28561595e-01
-7.83355772e-01 5.21465898e-01 -2.91292340e-01 4.02219325e-01
3.80868554e-01 -1.38043058e+00 6.02116048e-01 1.21136403e+00
2.47069359e-01 3.36299628e-01 3.81606698e-01 -6.85910165e-01
-8.21184397e-01 -1.51966131e+00 2.41000518e-01 -3.35957974e-01
1.72197640e-01 -5.46389580e-01 -8.51162553e-01 8.15052986e-01
4.05268937e-01 -2.31068023e-02 5.55764616e-01 -4.66500551e-01
-1.25113754e-02 -2.00499911e-02 -1.13444698e+00 6.55828118e-01
8.94661486e-01 -4.29735541e-01 -2.99895197e-01 3.55463952e-01
7.85333097e-01 -6.22908294e-01 -6.50191128e-01 4.57506359e-01
1.12919658e-01 -9.57041740e-01 9.61438239e-01 -3.35658848e-01
7.60964811e-01 -5.13410687e-01 2.25332126e-01 -1.29903805e+00
-3.79063338e-02 -6.85341895e-01 1.68537229e-01 1.46187222e+00
5.51243424e-01 -3.56661677e-01 1.10804224e+00 6.70126081e-01
-1.48447707e-01 -5.88372827e-01 -4.36426848e-01 -5.31559348e-01
-2.79198010e-02 -6.43250644e-02 6.54527605e-01 8.34740222e-01
-9.39993441e-01 4.35540497e-01 -4.10612702e-01 1.96760952e-01
9.25165176e-01 4.16875154e-01 1.10557735e+00 -8.10057223e-01
-4.40003961e-01 -2.66925693e-01 -5.94191663e-02 -1.16228855e+00
2.74572194e-01 -6.31209373e-01 4.40791249e-01 -1.55407643e+00
-1.97332855e-02 -5.21824539e-01 1.13123199e-02 5.08460581e-01
-4.36981887e-01 6.37498796e-01 1.47848397e-01 2.32929081e-01
-4.92041320e-01 7.75014639e-01 1.79411709e+00 -3.53737444e-01
-1.37241304e-01 1.64463505e-01 -6.07005715e-01 8.35396171e-01
7.33438492e-01 -4.50572282e-01 -6.13894880e-01 -4.69695956e-01
-1.22986443e-01 1.78515226e-01 5.48436165e-01 -8.93296242e-01
-1.18767977e-01 -2.30270207e-01 6.84515595e-01 -5.88513732e-01
3.55338097e-01 -6.80799901e-01 2.49479547e-01 2.17182636e-01
-2.87769675e-01 -2.93322682e-01 -6.44233748e-02 4.50377017e-01
-3.37106586e-01 -8.03369805e-02 1.02661121e+00 -3.87857527e-01
-2.70451099e-01 5.79128206e-01 -7.36304373e-02 1.66575059e-01
1.08570373e+00 -9.10489783e-02 6.97615892e-02 -4.09156203e-01
-6.27192557e-01 1.42410830e-01 7.15859175e-01 2.10658640e-01
5.39846003e-01 -1.37866485e+00 -6.18142366e-01 3.50793332e-01
-2.36175671e-01 8.18229914e-01 1.74119532e-01 5.17203867e-01
-5.87709904e-01 -8.81361030e-03 -2.56840974e-01 -7.59348929e-01
-8.82356286e-01 4.97493565e-01 2.75689930e-01 -3.64834338e-01
-4.59973931e-01 9.46361899e-01 5.79122126e-01 -3.76578778e-01
6.53009117e-02 -2.65695095e-01 2.36536220e-01 -1.57110348e-01
2.78052777e-01 8.87936726e-02 -2.00951353e-01 -4.63551790e-01
2.91770678e-02 3.69230717e-01 1.59404173e-01 -1.73687175e-01
1.19713414e+00 -1.27143532e-01 -1.80428028e-01 1.77628696e-01
1.04423869e+00 -1.84244379e-01 -1.88362992e+00 -8.30973163e-02
-2.65014172e-01 -5.58591306e-01 -1.05069920e-01 -6.78471148e-01
-1.63686645e+00 9.32206750e-01 3.00369650e-01 2.44186863e-01
1.35209429e+00 -4.82677333e-02 1.11781454e+00 -8.45994800e-02
1.86005011e-01 -8.96209717e-01 3.07242095e-01 -1.39272837e-02
7.37030029e-01 -1.14751542e+00 -1.61625534e-01 -6.26031280e-01
-6.81627214e-01 7.73684442e-01 9.67128932e-01 -3.28018665e-01
6.28983229e-02 3.40859890e-01 3.11879277e-01 9.80249643e-02
-3.98265481e-01 -1.93154082e-01 4.17371929e-01 6.56711817e-01
2.33413786e-01 4.00306918e-02 8.98713320e-02 4.47513133e-01
-1.89867005e-01 -1.65284306e-01 2.54910171e-01 6.41317606e-01
8.48196261e-03 -1.31814337e+00 -1.68487310e-01 2.49236763e-01
-4.04845834e-01 -1.50508583e-01 -3.89770061e-01 6.26654625e-01
3.02663833e-01 7.18236029e-01 9.56545845e-02 -1.38136253e-01
-2.42191181e-02 -3.97557877e-02 4.94225591e-01 -6.67359889e-01
-4.28614616e-01 3.47374648e-01 -1.53770581e-01 -5.81232250e-01
-6.44087732e-01 -5.24801910e-01 -1.12029362e+00 1.21501740e-02
-3.46232384e-01 -1.02687934e-02 5.02828062e-01 9.17488813e-01
2.31778234e-01 8.47490549e-01 6.95836604e-01 -1.02856827e+00
-2.12462619e-01 -8.26136827e-01 -2.77129680e-01 5.87217271e-01
2.01638713e-01 -4.03705865e-01 -2.43298903e-01 6.26220167e-01] | [11.31605052947998, -0.40196460485458374] |
d7ccdf0f-23a7-4b03-a879-0660e268d54f | exploring-open-vocabulary-semantic | 2306.00450 | null | https://arxiv.org/abs/2306.00450v1 | https://arxiv.org/pdf/2306.00450v1.pdf | Exploring Open-Vocabulary Semantic Segmentation without Human Labels | Semantic segmentation is a crucial task in computer vision that involves segmenting images into semantically meaningful regions at the pixel level. However, existing approaches often rely on expensive human annotations as supervision for model training, limiting their scalability to large, unlabeled datasets. To address this challenge, we present ZeroSeg, a novel method that leverages the existing pretrained vision-language (VL) model (e.g. CLIP) to train open-vocabulary zero-shot semantic segmentation models. Although acquired extensive knowledge of visual concepts, it is non-trivial to exploit knowledge from these VL models to the task of semantic segmentation, as they are usually trained at an image level. ZeroSeg overcomes this by distilling the visual concepts learned by VL models into a set of segment tokens, each summarizing a localized region of the target image. We evaluate ZeroSeg on multiple popular segmentation benchmarks, including PASCAL VOC 2012, PASCAL Context, and COCO, in a zero-shot manner (i.e., no training or adaption on target segmentation datasets). Our approach achieves state-of-the-art performance when compared to other zero-shot segmentation methods under the same training data, while also performing competitively compared to strongly supervised methods. Finally, we also demonstrated the effectiveness of ZeroSeg on open-vocabulary segmentation, through both human studies and qualitative visualizations. | ['Sean Chang Culatana', 'Mohamed Elhoseiny', 'Fanyi Xiao', 'Chenchen Zhu', 'Zhicheng Yan', 'Bernard Ghanem', 'Guocheng Qian', 'Deyao Zhu', 'Jun Chen'] | 2023-06-01 | null | null | null | null | ['zero-shot-segmentation'] | ['computer-vision'] | [ 4.47108150e-01 2.96169281e-01 -3.41523767e-01 -3.69907618e-01
-7.64223456e-01 -7.02731252e-01 4.29159433e-01 2.68507004e-01
-5.81989884e-01 2.17784122e-01 -1.20042801e-01 -8.86180177e-02
5.22253692e-01 -7.97731638e-01 -8.69278729e-01 -5.64003706e-01
4.60724205e-01 3.97670001e-01 8.22858930e-01 -9.71649066e-02
1.95312798e-01 1.00350797e-01 -1.63216710e+00 -4.08284692e-03
9.49107349e-01 1.00172198e+00 3.13366354e-01 4.49566215e-01
-4.95443135e-01 4.89005089e-01 -4.46600199e-01 -2.96365738e-01
2.02640489e-01 -3.37702274e-01 -1.02125156e+00 4.55591530e-01
6.49992764e-01 4.77880910e-02 1.10740602e-01 1.25116920e+00
9.96488854e-02 3.17619860e-01 4.93320227e-01 -1.11054325e+00
-8.05550635e-01 4.54816461e-01 -5.36019742e-01 -3.78950424e-02
1.21020600e-01 2.85770208e-01 1.06614316e+00 -9.43219900e-01
1.02574575e+00 1.01280820e+00 6.71896160e-01 6.22071922e-01
-1.59063995e+00 -3.41310263e-01 3.23900312e-01 -6.11150153e-02
-1.39274848e+00 -1.61634877e-01 7.51551688e-01 -7.48878062e-01
8.08285117e-01 4.97378223e-02 7.33223915e-01 9.97731924e-01
-2.64512569e-01 9.89104688e-01 9.21461880e-01 -3.82511020e-01
5.67253411e-01 5.95321506e-02 5.87984145e-01 6.07243121e-01
1.04466066e-01 -1.68654814e-01 -3.24950963e-01 2.19672576e-01
6.00158155e-01 6.60469979e-02 -2.16442645e-01 -8.10995758e-01
-1.20744157e+00 8.31692278e-01 6.89199626e-01 2.19554901e-01
-1.69644371e-01 1.52576819e-01 5.61586976e-01 -1.42138913e-01
6.11974597e-01 4.66289997e-01 -3.81016016e-01 1.73683062e-01
-1.28979182e+00 8.33251849e-02 7.22214818e-01 1.09986687e+00
1.23457813e+00 2.09699087e-02 -4.93012577e-01 8.64192426e-01
1.97632328e-01 3.86432171e-01 4.43529457e-01 -9.84332800e-01
1.26334697e-01 7.01666296e-01 -8.77659395e-02 -5.80862701e-01
-1.48708954e-01 -2.60388851e-01 -5.13392985e-01 1.68613866e-01
3.60489070e-01 3.85446660e-02 -1.70324504e+00 1.64732838e+00
4.53557581e-01 4.72672850e-01 1.08249664e-01 9.52781081e-01
1.04723179e+00 5.89972496e-01 5.11930168e-01 1.79298475e-01
1.25816953e+00 -1.49367154e+00 -5.92363596e-01 -4.62280840e-01
6.65778637e-01 -5.14329255e-01 1.58779824e+00 1.74513161e-01
-7.49708593e-01 -5.26149690e-01 -9.43010747e-01 -3.27887177e-01
-7.28730142e-01 -2.51475811e-01 4.80784237e-01 4.71511632e-01
-8.91620815e-01 4.53913063e-01 -8.52129996e-01 -5.32506227e-01
8.29729557e-01 -6.42912164e-02 -1.60228595e-01 -1.70335233e-01
-9.55747485e-01 3.48439425e-01 7.25322247e-01 -4.13407773e-01
-1.03658950e+00 -9.15301979e-01 -1.40016282e+00 8.97222832e-02
7.95245767e-01 -4.19654965e-01 1.23466969e+00 -1.23888135e+00
-1.21677506e+00 1.08966351e+00 -3.41093205e-02 -5.40571690e-01
3.88699770e-01 -1.84891000e-01 1.56459704e-01 3.28615516e-01
5.27716339e-01 1.21938550e+00 7.58563995e-01 -1.45533514e+00
-3.52538943e-01 -2.93129474e-01 2.12047771e-01 -8.32281113e-02
-1.93022341e-01 -3.05122674e-01 -1.03160524e+00 -7.52513051e-01
6.69441819e-02 -8.45728993e-01 -5.90001106e-01 1.65102974e-01
-5.65803885e-01 -2.19720438e-01 1.11695588e+00 -5.16116321e-01
9.95706618e-01 -2.13217592e+00 1.55740961e-01 1.54922912e-02
9.75721553e-02 5.62213778e-01 -1.56806722e-01 1.22820206e-01
1.71175912e-01 3.07557791e-01 -9.41697419e-01 -6.06749713e-01
-4.62354049e-02 4.25777137e-01 -6.43687844e-02 2.11885765e-01
1.75330058e-01 1.40828168e+00 -1.05518150e+00 -8.45554411e-01
5.69993615e-01 3.76246154e-01 -5.86337507e-01 2.13041738e-01
-7.04909146e-01 5.27686119e-01 -3.39828998e-01 8.88734400e-01
4.40174252e-01 -4.93260503e-01 -1.92753375e-01 -1.09485053e-01
-5.08614853e-02 -4.39161777e-01 -8.84586394e-01 2.19992638e+00
-2.50474423e-01 5.72930455e-01 -1.73349559e-01 -1.27247310e+00
6.75390422e-01 1.21303074e-01 4.50865418e-01 -7.14472890e-01
2.33839184e-01 -2.57774312e-02 -6.17982388e-01 -5.00387847e-01
3.44989598e-01 -2.93097906e-02 -3.58580798e-01 2.18771383e-01
4.46966350e-01 -3.87630254e-01 5.19216835e-01 4.29039031e-01
6.64537072e-01 3.74136388e-01 2.69809127e-01 -3.45406920e-01
2.96645850e-01 4.51195300e-01 6.48352742e-01 8.10432911e-01
-3.92581969e-01 1.00595617e+00 4.73377526e-01 -1.71858758e-01
-1.05682147e+00 -1.16668272e+00 -7.67625570e-02 1.15842450e+00
5.97153962e-01 -4.09470916e-01 -1.33706617e+00 -7.80255616e-01
-1.75505087e-01 7.91991591e-01 -7.97898829e-01 6.20732419e-02
-1.76870644e-01 -1.58315569e-01 4.33833003e-01 7.34371781e-01
5.60880601e-01 -1.13886726e+00 -9.17966962e-01 4.17983048e-02
-5.49454950e-02 -1.52077472e+00 -5.86888015e-01 5.00672199e-02
-5.90817451e-01 -1.18063164e+00 -9.68343794e-01 -9.18491006e-01
7.24078238e-01 4.61563021e-01 1.20623159e+00 2.85393018e-02
-7.03348398e-01 7.02526093e-01 -5.85029423e-01 -2.94493675e-01
-3.69270258e-02 4.66841683e-02 -6.28668487e-01 6.27112314e-02
3.90994161e-01 -2.21325591e-01 -5.94556391e-01 2.00179949e-01
-1.25895369e+00 2.98878968e-01 3.25813144e-01 7.73793638e-01
1.11332512e+00 -5.82942009e-01 3.99578065e-01 -1.23107290e+00
3.15320849e-01 -4.21285748e-01 -6.42954707e-01 4.34577167e-01
-3.87656659e-01 -1.18265443e-01 4.95503098e-01 -3.42198342e-01
-9.55903649e-01 4.07921016e-01 -4.56023328e-02 -8.42741370e-01
-4.18474823e-01 5.08546531e-01 -1.61459714e-01 -1.55569643e-01
6.89362526e-01 1.76349014e-01 -1.72841147e-01 -3.91192734e-01
9.27086949e-01 5.54932833e-01 8.31953883e-01 -4.09568995e-01
6.63658798e-01 6.54104650e-01 -3.57503027e-01 -1.04919624e+00
-1.23281395e+00 -9.93872344e-01 -9.67371702e-01 -6.00529499e-02
1.51262963e+00 -8.36136222e-01 8.09133705e-03 4.42769080e-01
-9.46102381e-01 -6.83583498e-01 -7.20110714e-01 -3.32886837e-02
-7.17739463e-01 5.10807395e-01 -3.83188277e-01 -3.03628296e-01
-3.19636971e-01 -1.17614233e+00 1.34009004e+00 4.34736878e-01
-1.34122595e-01 -1.29087043e+00 -7.33282715e-02 5.32637954e-01
1.25200674e-01 5.46830654e-01 6.39020324e-01 -5.44395387e-01
-5.51222503e-01 -9.09645185e-02 -4.76984054e-01 5.06058097e-01
-6.13039657e-02 -1.18835866e-01 -1.10608780e+00 -6.61357120e-02
-3.57930571e-01 -7.19939291e-01 1.16682243e+00 4.41032410e-01
1.41834235e+00 1.76850289e-01 -3.77480149e-01 7.81167865e-01
1.60857964e+00 -9.10650566e-02 5.37243664e-01 1.58311531e-01
1.05015469e+00 5.82792699e-01 9.07430828e-01 1.24248236e-01
3.38264912e-01 5.67157745e-01 4.25389469e-01 -5.45624971e-01
-3.01162452e-01 -3.93934101e-01 -1.10840641e-01 4.19190288e-01
2.80835062e-01 -1.43071532e-01 -1.13082993e+00 8.64635587e-01
-2.01992440e+00 -6.07354820e-01 -5.49163744e-02 2.01637030e+00
7.66559124e-01 7.69906864e-02 -3.63252461e-02 -1.48295090e-01
6.39481843e-01 4.28976297e-01 -8.07197392e-01 -3.01831275e-01
-7.70368129e-02 3.21341842e-01 6.75594151e-01 4.41933990e-01
-1.39935744e+00 1.68878603e+00 5.83096552e+00 8.98307621e-01
-1.18956590e+00 4.89951581e-01 6.66372061e-01 1.05696037e-01
-2.47156575e-01 1.51456684e-01 -5.71598709e-01 2.91914940e-01
5.54413617e-01 -1.22052491e-01 6.43481016e-02 1.01388574e+00
6.02731369e-02 -3.10058534e-01 -9.23303545e-01 9.23011124e-01
3.40926945e-01 -1.45313179e+00 1.42664596e-01 -3.07173848e-01
1.16098177e+00 2.15943873e-01 -2.01074973e-01 4.66766864e-01
3.25494528e-01 -1.07059431e+00 7.55694330e-01 3.06349576e-01
8.45140100e-01 -3.65084916e-01 5.51514924e-01 3.41270119e-01
-1.34263849e+00 2.46428490e-01 -2.47933775e-01 2.95524955e-01
3.94946843e-01 4.35615361e-01 -5.42643666e-01 4.57556158e-01
7.08331168e-01 1.03354132e+00 -7.70839453e-01 1.02023363e+00
-4.22066361e-01 7.54422963e-01 -7.91822821e-02 3.63324970e-01
6.46226823e-01 -2.26194561e-01 3.87217820e-01 1.36238289e+00
-1.26161456e-01 8.61240551e-02 7.24053442e-01 1.31357944e+00
-1.38214692e-01 2.14806750e-01 -5.64850092e-01 -1.85309529e-01
1.04183875e-01 1.24955225e+00 -1.29616618e+00 -6.88010335e-01
-6.46364510e-01 1.12956071e+00 2.00024500e-01 6.80433750e-01
-7.31327653e-01 -5.59188843e-01 4.79532391e-01 1.79300159e-02
5.05536795e-01 -2.25684166e-01 -3.65463972e-01 -1.20498502e+00
-1.36723489e-01 -4.60402101e-01 3.10883641e-01 -7.48562455e-01
-1.01978934e+00 3.72037917e-01 3.70039158e-02 -9.92283285e-01
9.60890669e-03 -3.85551006e-01 -6.74577832e-01 5.55909872e-01
-1.60101914e+00 -1.38021994e+00 -6.34869337e-01 6.64358437e-01
1.05025423e+00 3.23940039e-01 6.66664600e-01 5.23058213e-02
-5.56544721e-01 2.87953556e-01 -1.29701197e-01 3.11863035e-01
4.83358562e-01 -1.38760567e+00 6.45129144e-01 9.76770997e-01
3.37343633e-01 3.57362449e-01 5.13824999e-01 -6.39637530e-01
-8.76575649e-01 -1.45105898e+00 4.17749286e-01 -3.39489400e-01
6.08139575e-01 -6.01994157e-01 -1.20687997e+00 7.22779095e-01
1.16423748e-01 6.03515506e-01 6.22952342e-01 -1.60240307e-01
-3.71731550e-01 4.45838302e-01 -9.65788245e-01 5.38394451e-01
1.10091460e+00 -5.26102364e-01 -7.87317038e-01 2.01737747e-01
1.29863262e+00 -4.07683849e-01 -7.39384651e-01 3.47617984e-01
2.05037773e-01 -7.69045949e-01 1.08459389e+00 -5.88930905e-01
4.76051271e-01 -2.95701146e-01 -1.30237028e-01 -1.23712170e+00
1.55689925e-01 -2.23201081e-01 1.56004533e-01 1.20391881e+00
2.66071230e-01 -3.39681029e-01 8.42042327e-01 5.66199064e-01
-3.09652835e-01 -7.57434607e-01 -6.95183098e-01 -8.40887129e-01
5.38503267e-02 -5.37345588e-01 2.13339016e-01 1.00303566e+00
-2.87465394e-01 3.62518042e-01 -4.76961769e-02 -5.12125269e-02
8.50008607e-01 2.25863174e-01 8.94791365e-01 -1.34526026e+00
-4.07245383e-02 -4.25502360e-01 -5.93199015e-01 -1.02616966e+00
4.74797815e-01 -1.07838142e+00 3.33639979e-01 -1.87855434e+00
1.67993143e-01 -3.07011932e-01 -3.29609573e-01 6.82225525e-01
-3.12697589e-01 5.99258661e-01 3.30401331e-01 1.49864748e-01
-1.08552563e+00 6.12765074e-01 1.35032344e+00 -4.60001141e-01
-3.51597130e-01 -2.69820213e-01 -6.02445602e-01 8.44738126e-01
6.76005125e-01 -4.21475202e-01 -6.57383800e-01 -3.01552713e-01
-3.07685226e-01 -2.75710613e-01 5.85412264e-01 -1.01231503e+00
2.05431357e-01 -2.59693772e-01 -8.81338120e-02 -4.49327052e-01
1.96667358e-01 -4.68738645e-01 -3.54280174e-01 2.56791741e-01
-3.27670246e-01 -6.73307538e-01 3.03822935e-01 7.62647390e-01
-4.30414021e-01 -3.01729679e-01 9.44286048e-01 -2.61735618e-01
-1.40379369e+00 4.92637008e-01 -5.61562963e-02 5.83956301e-01
1.36610830e+00 -4.43178922e-01 -1.34234786e-01 3.27869765e-02
-9.36971426e-01 5.32750130e-01 8.12564611e-01 5.70222139e-01
5.36397994e-01 -9.16619241e-01 -2.55147338e-01 2.39315867e-01
6.67927086e-01 5.12921154e-01 3.11533570e-01 7.76279330e-01
-5.70673287e-01 3.11345637e-01 -1.34205014e-01 -1.10480022e+00
-1.21440578e+00 7.55722344e-01 2.62207240e-01 -5.49351564e-03
-9.33942854e-01 8.33159983e-01 6.83516264e-01 -4.89351034e-01
2.96146750e-01 -3.84632319e-01 -1.55419484e-01 5.39522395e-02
3.43652815e-01 -2.27389894e-02 -1.67050108e-01 -6.71100259e-01
-6.75256848e-02 9.90674555e-01 2.94126738e-02 -9.32874084e-02
1.03880417e+00 -1.80245340e-01 1.26739785e-01 6.60439253e-01
1.17026675e+00 -4.15794820e-01 -1.56379128e+00 -4.61247116e-01
2.09682867e-01 -4.07453030e-01 5.57836294e-02 -6.10400796e-01
-1.16930234e+00 1.14422143e+00 4.83910710e-01 4.50606504e-03
8.30246031e-01 2.90502876e-01 9.68472004e-01 1.77963808e-01
3.52612913e-01 -1.27847397e+00 2.77875930e-01 4.61156964e-01
3.88454944e-01 -1.65647960e+00 -3.47948134e-01 -7.29116201e-01
-9.09345388e-01 7.49160171e-01 5.71740568e-01 -1.93837494e-01
6.28047168e-01 -6.49419576e-02 3.20728213e-01 -3.23316842e-01
-2.42883891e-01 -6.98434353e-01 3.03375065e-01 7.09161282e-01
1.48669466e-01 9.81299952e-02 -1.94363326e-01 5.63247144e-01
1.13545045e-01 1.12342782e-01 3.13939840e-01 1.01744378e+00
-5.94948769e-01 -8.31475198e-01 -9.14092809e-02 2.72346795e-01
-2.44244039e-01 -9.51903835e-02 -4.40960556e-01 7.87521720e-01
1.61855847e-01 8.33856523e-01 1.40928000e-01 -3.99915427e-02
2.69260108e-01 5.79275675e-02 5.89185692e-02 -1.14128363e+00
-4.45179403e-01 3.07030845e-02 -4.44460571e-01 -8.79181802e-01
-5.29454589e-01 -4.44474608e-01 -1.56835163e+00 3.90828580e-01
-1.76461190e-01 1.59184001e-02 6.12090588e-01 1.03595579e+00
1.96587190e-01 7.07444787e-01 3.04275639e-02 -1.00196981e+00
-3.88038941e-02 -6.27996325e-01 -4.74557877e-01 9.04794931e-01
1.29275531e-01 -7.26293862e-01 -2.48785645e-01 4.04038131e-01] | [9.617168426513672, 0.8470037579536438] |
b7dd1141-decd-4550-804f-3fc865944c9f | safeml-safety-monitoring-of-machine-learning | 2005.13166 | null | https://arxiv.org/abs/2005.13166v1 | https://arxiv.org/pdf/2005.13166v1.pdf | SafeML: Safety Monitoring of Machine Learning Classifiers through Statistical Difference Measure | Ensuring safety and explainability of machine learning (ML) is a topic of increasing relevance as data-driven applications venture into safety-critical application domains, traditionally committed to high safety standards that are not satisfied with an exclusive testing approach of otherwise inaccessible black-box systems. Especially the interaction between safety and security is a central challenge, as security violations can lead to compromised safety. The contribution of this paper to addressing both safety and security within a single concept of protection applicable during the operation of ML systems is active monitoring of the behaviour and the operational context of the data-driven system based on distance measures of the Empirical Cumulative Distribution Function (ECDF). We investigate abstract datasets (XOR, Spiral, Circle) and current security-specific datasets for intrusion detection (CICIDS2017) of simulated network traffic, using distributional shift detection measures including the Kolmogorov-Smirnov, Kuiper, Anderson-Darling, Wasserstein and mixed Wasserstein-Anderson-Darling measures. Our preliminary findings indicate that the approach can provide a basis for detecting whether the application context of an ML component is valid in the safety-security. Our preliminary code and results are available at https://github.com/ISorokos/SafeML. | ['Declan Whiting', 'Koorosh Aslansefat', 'Yiannis Papadopoulos', 'Ramin Tavakoli Kolagari', 'Ioannis Sorokos'] | 2020-05-27 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 2.73437127e-02 -6.85131177e-02 -4.04929668e-02 -3.31701010e-01
-3.39943975e-01 -9.26651001e-01 6.30815983e-01 6.31722510e-01
-2.69304752e-01 6.26083791e-01 -4.36866283e-01 -1.19799840e+00
-7.47280598e-01 -6.89679980e-01 -5.77010393e-01 -4.96527135e-01
-3.38166356e-01 6.91592917e-02 3.52465123e-01 -1.20007493e-01
4.68278259e-01 8.99902880e-01 -1.81547153e+00 3.33799958e-01
4.21395898e-01 1.15005898e+00 -6.77464306e-01 1.00833273e+00
2.95830667e-01 7.05850601e-01 -7.41430819e-01 -2.00085685e-01
3.93136263e-01 -1.30155161e-01 -7.48066485e-01 -3.94625992e-01
1.21074729e-02 -9.99108329e-02 8.12896490e-02 1.19588602e+00
-1.04631074e-02 -2.16504484e-01 4.97654587e-01 -2.26119089e+00
-8.52148328e-03 3.12884361e-01 -2.36636609e-01 2.45452613e-01
1.22220911e-01 3.43140304e-01 6.69737875e-01 -3.25240791e-01
4.66938876e-02 9.93642569e-01 3.75202447e-01 2.91672558e-01
-1.23629999e+00 -6.84182525e-01 -9.72015560e-02 2.52775609e-01
-1.29553318e+00 -2.12151766e-01 4.57509577e-01 -6.70397580e-01
8.95081997e-01 6.39397204e-01 8.10019597e-02 1.04419243e+00
6.63751662e-01 -8.26476701e-03 9.42351043e-01 -6.12284601e-01
4.99663353e-01 7.83612072e-01 4.36348945e-01 2.45260805e-01
8.56842518e-01 6.14459097e-01 -6.80666268e-02 -5.14144003e-01
7.92969838e-02 5.96419647e-02 1.48705900e-01 -4.09276873e-01
-6.76522195e-01 7.91119695e-01 -4.18021470e-01 4.33890224e-01
-1.14858113e-01 -7.69849261e-03 6.93391800e-01 5.87608099e-01
1.65986046e-01 1.21864468e-01 -8.84766579e-01 -4.53534812e-01
-4.86667722e-01 1.68668956e-01 1.04412687e+00 9.04642582e-01
4.02106404e-01 -3.77964079e-02 1.30528092e-01 -4.41885963e-02
5.73359430e-01 3.26379091e-01 -6.92572817e-03 -5.77020466e-01
1.40824482e-01 6.82389975e-01 -6.24915212e-02 -7.47259796e-01
-2.71189690e-01 -1.05916649e-01 -4.91189182e-01 6.67196095e-01
3.06775182e-01 -1.68367773e-01 -2.09600762e-01 1.40228772e+00
3.46154809e-01 9.64356363e-02 -3.41413170e-02 1.26067638e-01
-7.96323344e-02 4.13929790e-01 3.11994225e-01 -3.61975491e-01
1.00225699e+00 -4.54431167e-03 -7.49088943e-01 2.32940897e-01
9.73406971e-01 -4.76445228e-01 8.99891913e-01 7.91982234e-01
-8.16752195e-01 -5.52575529e-01 -1.24484897e+00 7.95952260e-01
-7.59039044e-01 -6.40247881e-01 1.03959352e-01 1.26172221e+00
-7.95219898e-01 5.78262985e-01 -5.52793086e-01 -2.61431217e-01
2.70744622e-01 3.33392709e-01 -4.29499120e-01 2.61603534e-01
-1.30088151e+00 1.07020080e+00 4.92223710e-01 -1.97309688e-01
-9.97516870e-01 -7.72016823e-01 -7.09723175e-01 -1.29478738e-01
3.80541712e-01 1.71511248e-01 1.05137885e+00 -4.26620275e-01
-8.27646852e-01 4.95693713e-01 5.80392540e-01 -6.16156876e-01
5.88593900e-01 -1.00682825e-01 -1.05277467e+00 -2.04309940e-01
-3.48177075e-01 -8.55801478e-02 6.05446517e-01 -1.12860250e+00
-9.02858317e-01 -4.45723653e-01 -9.34736878e-02 -6.56830549e-01
-3.49184901e-01 4.02411044e-01 6.07370198e-01 -2.19352156e-01
-6.83872402e-01 -7.39256382e-01 2.36642342e-02 -2.64720827e-01
-3.87262166e-01 5.63878566e-02 1.16740012e+00 -4.61051017e-01
1.72237015e+00 -2.31329417e+00 -6.37343884e-01 6.74101293e-01
-2.29332969e-01 5.68855584e-01 2.46036455e-01 7.43815064e-01
-5.98586261e-01 4.66673553e-01 -4.21905577e-01 5.01038015e-01
1.46785319e-01 2.15101615e-02 -4.95926857e-01 9.19726789e-01
2.01203480e-01 3.21027994e-01 -4.37523961e-01 -1.60651401e-01
5.54413080e-01 1.07201479e-01 -2.97674477e-01 2.44594574e-01
-8.59152675e-02 2.69029260e-01 -1.89899474e-01 5.52980900e-01
6.57010555e-01 4.10697043e-01 1.50310611e-02 1.75829187e-01
-3.06654871e-01 1.53055312e-02 -1.29760540e+00 6.74852133e-01
-3.18026513e-01 4.24175829e-01 -1.18033394e-01 -1.15011692e+00
9.69771385e-01 4.88440931e-01 4.05564040e-01 -6.34811938e-01
3.52336913e-01 9.71018597e-02 2.28139296e-01 -6.33715391e-01
7.25057274e-02 -8.24246630e-02 -3.29696238e-01 5.46803951e-01
-2.21879110e-01 1.02796229e-02 1.84357334e-02 2.11871602e-03
1.27228773e+00 -1.77890286e-01 5.32716393e-01 -6.57463610e-01
9.15322959e-01 -1.41145006e-01 3.35073382e-01 3.40311080e-01
-4.93707567e-01 -6.67975172e-02 1.15394843e+00 -3.57558668e-01
-9.66549993e-01 -1.14750445e+00 -3.92116845e-01 5.78603685e-01
-2.00930551e-01 -2.76471227e-01 -8.93546402e-01 -1.10455346e+00
3.51779729e-01 1.48331714e+00 -5.72373569e-01 -8.04611564e-01
-2.84947306e-01 -2.78737754e-01 8.13274980e-01 3.04924190e-01
3.30417491e-02 -9.47369277e-01 -1.02590823e+00 -1.56805530e-01
5.26898265e-01 -9.82055545e-01 -4.95767370e-02 3.71994138e-01
-7.70904541e-01 -1.48534632e+00 4.84791696e-01 -1.44175217e-01
5.04025102e-01 -3.66606981e-01 7.30318308e-01 2.89753973e-01
-7.53203511e-01 5.03284931e-01 -3.01419318e-01 -7.25535154e-01
-9.17960346e-01 -4.67757821e-01 3.83869320e-01 -1.89991277e-02
7.06367314e-01 -3.04633349e-01 -2.34684587e-01 6.60826147e-01
-1.16465044e+00 -9.17673588e-01 2.06821859e-01 2.22650498e-01
4.01724018e-02 5.99298656e-01 7.90883541e-01 -7.29524970e-01
8.64260674e-01 -6.38934314e-01 -1.00651908e+00 2.73093402e-01
-1.16416228e+00 -1.07298575e-01 6.60915732e-01 -3.76279563e-01
-5.44521272e-01 -3.21079552e-01 -6.56787977e-02 -2.75339961e-01
-9.14883375e-01 2.36507297e-01 -5.69409132e-01 6.32958934e-02
7.18109965e-01 -7.39383996e-02 2.66567469e-01 -1.93773016e-01
4.84344456e-03 9.60893989e-01 1.70637578e-01 -6.08797967e-01
1.01129735e+00 2.32591793e-01 2.75948405e-01 -8.28194082e-01
-1.58391580e-01 -4.27062631e-01 -5.59312999e-01 -3.24094683e-01
5.27857363e-01 -1.38652042e-01 -1.18438756e+00 3.22165042e-01
-9.14683521e-01 -1.20243289e-01 -4.01910603e-01 3.22733015e-01
-4.05831128e-01 5.01693726e-01 -1.44311652e-01 -1.45335388e+00
-1.17123447e-01 -9.61249053e-01 5.84279120e-01 -2.48392537e-01
-5.27439773e-01 -1.00493097e+00 2.48384491e-01 7.64614344e-02
2.68693507e-01 5.68560004e-01 1.20640433e+00 -1.34851742e+00
6.78187832e-02 -8.86481225e-01 2.06427664e-01 1.05257070e+00
1.91251710e-01 5.51121891e-01 -1.00302565e+00 -4.04328585e-01
4.23003018e-01 7.90203735e-02 -1.19777821e-01 1.18517708e-02
1.29389763e+00 -5.60700893e-01 8.76233876e-02 -8.56693834e-03
1.28574085e+00 7.92297482e-01 6.62447810e-01 3.26726347e-01
8.93677119e-03 1.03931153e+00 1.13691223e+00 7.04060733e-01
-2.35501140e-01 4.15759236e-01 7.88202107e-01 3.25324357e-01
7.51098692e-01 -4.33285050e-02 6.50821984e-01 -3.63052860e-02
4.78069931e-01 -1.13826476e-01 -1.29107738e+00 3.38824481e-01
-1.58616364e+00 -1.08579314e+00 -3.97607952e-01 2.68944430e+00
4.22012746e-01 8.13277721e-01 3.41329873e-01 8.67829084e-01
6.97056890e-01 -4.18983102e-01 -2.97333121e-01 -9.80788350e-01
5.52500010e-01 1.25055745e-01 6.29643381e-01 4.65048164e-01
-1.00455022e+00 2.91526556e-01 5.49994135e+00 6.79987311e-01
-7.71364689e-01 -1.13914602e-01 5.72586119e-01 1.42145827e-01
9.03945118e-02 1.47291362e-01 -6.34160876e-01 7.01627374e-01
1.78467989e+00 -3.15296412e-01 -3.70817855e-02 1.02816045e+00
4.62819576e-01 -1.68806821e-01 -1.40661323e+00 3.80141765e-01
-2.89406389e-01 -7.74486661e-01 -3.22940439e-01 3.69253695e-01
-2.28376221e-02 -4.79544044e-01 3.15197468e-01 1.03161938e-01
-3.43479891e-03 -1.05055571e+00 7.40093350e-01 5.36420524e-01
6.76863909e-01 -1.25505245e+00 9.15901959e-01 5.43500900e-01
-8.36318314e-01 -2.83147246e-01 8.91789719e-02 -8.23874399e-02
-4.38691862e-02 4.87449944e-01 -7.46253490e-01 4.77846473e-01
7.44925082e-01 1.19275264e-01 -5.48302948e-01 7.27234840e-01
9.36306044e-02 6.77127004e-01 4.23548520e-02 3.01929377e-02
-8.23069736e-02 -6.52635619e-02 4.69440311e-01 1.07766902e+00
9.46779773e-02 -3.62123698e-01 -2.12160096e-01 7.43965805e-01
7.71603227e-01 -3.22937310e-01 -8.70215714e-01 -3.24807703e-01
5.16508281e-01 1.21420527e+00 -4.99457359e-01 1.34887651e-01
-3.89631510e-01 -1.14346351e-02 -5.30200362e-01 7.35955611e-02
-1.13897872e+00 -6.40482008e-01 9.86517429e-01 4.72884983e-01
-1.06007181e-01 -2.60474950e-01 -4.74515468e-01 -3.50384980e-01
7.54569769e-02 -1.01731312e+00 6.21693134e-01 -1.50556013e-01
-1.11692762e+00 4.99045700e-01 4.81846064e-01 -1.54226851e+00
-1.30004227e-01 -7.70751357e-01 -7.12490678e-01 8.55256200e-01
-1.00675619e+00 -8.20084333e-01 5.40956520e-02 6.96313083e-01
1.20770358e-01 -3.94938231e-01 7.53318906e-01 3.93382013e-01
-6.34015918e-01 5.45834601e-01 -2.78724074e-01 -1.33162484e-01
5.93243957e-01 -9.57551837e-01 3.77278477e-01 1.03477514e+00
-1.51463091e-01 6.91576779e-01 1.00952959e+00 -8.08780909e-01
-1.23395014e+00 -1.23470008e+00 6.59227431e-01 -6.81681812e-01
9.92832005e-01 -5.05830646e-01 -1.02886569e+00 5.32205284e-01
5.73879145e-02 8.01363885e-02 9.06174898e-01 -1.89179525e-01
-2.65581071e-01 -2.44203478e-01 -1.64981818e+00 3.39418173e-01
4.91158187e-01 -5.29724479e-01 -4.53937173e-01 1.75982788e-01
6.54034555e-01 3.49846810e-01 -9.49293017e-01 5.73284566e-01
3.11237723e-01 -1.16052628e+00 6.42565608e-01 -1.04338872e+00
-1.78430423e-01 -3.72668535e-01 -1.80024013e-01 -7.45438457e-01
2.82252058e-02 -7.08946705e-01 -3.28356534e-01 1.31251729e+00
5.47326922e-01 -9.67320442e-01 4.86102521e-01 6.94659233e-01
1.20611705e-01 -5.96158147e-01 -1.09901464e+00 -1.16245985e+00
1.41404316e-01 -1.02474868e+00 5.67704856e-01 9.01381254e-01
4.23020273e-01 -2.11413279e-01 1.99359953e-02 4.98516768e-01
7.61896431e-01 -5.80292761e-01 6.26681924e-01 -1.33683181e+00
-6.26766533e-02 -5.00199318e-01 -7.89427996e-01 3.61140847e-01
1.07791357e-01 -5.59507012e-01 -1.18293941e-01 -8.71634066e-01
-3.05774659e-01 -1.82798982e-01 -5.80028355e-01 4.08644289e-01
4.24366623e-01 -4.43552911e-01 -1.07238695e-01 -2.74260163e-01
-2.68400609e-01 9.13168788e-02 2.00714141e-01 1.25137165e-01
8.39703158e-02 4.64919925e-01 -4.99990135e-01 6.00954294e-01
1.00865304e+00 -6.85395002e-01 -5.55217326e-01 4.98510450e-01
2.73616403e-01 5.41537069e-02 5.55402815e-01 -1.07710898e+00
3.07143867e-01 -4.58430409e-01 -1.97260350e-01 -4.92691427e-01
-3.78646910e-01 -1.59634066e+00 4.30256933e-01 1.03801405e+00
-4.27532524e-01 4.03950125e-01 5.49272656e-01 4.95220333e-01
-8.57578963e-02 -3.25271726e-01 9.16083992e-01 4.11921680e-01
-3.53185773e-01 2.52208114e-01 -6.36876523e-01 -1.53736949e-01
1.94321060e+00 -3.09259057e-01 -4.11492974e-01 -8.95084590e-02
-3.15300226e-01 1.23673096e-01 3.14530551e-01 6.36650026e-01
7.06996381e-01 -9.54546571e-01 -4.35618728e-01 7.06714809e-01
2.52801597e-01 -5.37669003e-01 2.01803818e-01 9.09390450e-01
-5.17769396e-01 5.80508113e-01 -3.79182935e-01 -3.75730634e-01
-1.56535041e+00 1.14176977e+00 3.24684262e-01 1.09248728e-01
-2.32359469e-01 3.81433785e-01 -1.17793828e-01 -3.77432406e-01
5.87172389e-01 -1.33210927e-01 -1.77934282e-02 -2.91001704e-03
5.60828507e-01 8.77492368e-01 4.46581542e-01 -4.42947239e-01
-7.32781112e-01 2.10596863e-02 3.76430457e-03 -1.21779941e-01
1.01324463e+00 2.78179914e-01 -3.03041250e-01 5.02546966e-01
1.08276260e+00 -2.27312669e-01 -8.42855632e-01 4.28215057e-01
7.18439162e-01 -4.30193096e-01 -1.82940334e-01 -8.69207323e-01
-6.46171510e-01 1.01635003e+00 7.62502015e-01 1.03253794e+00
9.09064889e-01 -3.12945664e-01 2.30825976e-01 4.67660651e-03
3.84637535e-01 -1.13210797e+00 -2.21374243e-01 2.48766780e-01
6.43960238e-01 -8.02749157e-01 -1.06877923e-01 -1.53927147e-01
-7.72553086e-01 1.09949386e+00 6.17509604e-01 1.67724043e-01
1.30860305e+00 7.88556218e-01 -1.32759959e-01 -7.87839573e-03
-1.01999247e+00 3.46359670e-01 1.26231804e-01 7.20859408e-01
1.96142003e-01 1.86483562e-02 -2.31432930e-01 6.85847342e-01
1.68641314e-01 -2.10414261e-01 6.81814909e-01 1.33668029e+00
-4.12488788e-01 -1.26174808e+00 -5.05864441e-01 3.85852724e-01
-4.54871684e-01 3.00460041e-01 -4.89341736e-01 1.08290219e+00
4.14892919e-02 1.39909303e+00 -9.71311554e-02 -7.80076921e-01
7.20257401e-01 2.35010237e-01 -1.31069750e-01 -4.83636439e-01
-6.56264782e-01 -4.29371238e-01 1.20755546e-01 -6.61769509e-01
1.85631886e-01 -8.25151622e-01 -1.03885508e+00 -5.29131770e-01
-2.33088791e-01 5.02361655e-01 1.01480293e+00 8.43227327e-01
2.49196500e-01 5.60015678e-01 9.55136657e-01 4.47556861e-02
-1.11655724e+00 -6.47384286e-01 -8.22587609e-01 2.21168041e-01
4.44322169e-01 -6.28601670e-01 -8.93595219e-01 -2.09117532e-01] | [5.358952045440674, 7.227952480316162] |
88d173ee-6be9-4901-8032-b8f789ec1c7c | free-as-in-free-word-order-an-energy-based | 1809.01446 | null | http://arxiv.org/abs/1809.01446v2 | http://arxiv.org/pdf/1809.01446v2.pdf | Free as in Free Word Order: An Energy Based Model for Word Segmentation and Morphological Tagging in Sanskrit | The configurational information in sentences of a free word order language
such as Sanskrit is of limited use. Thus, the context of the entire sentence
will be desirable even for basic processing tasks such as word segmentation. We
propose a structured prediction framework that jointly solves the word
segmentation and morphological tagging tasks in Sanskrit. We build an energy
based model where we adopt approaches generally employed in graph based parsing
techniques (McDonald et al., 2005a; Carreras, 2007). Our model outperforms the
state of the art with an F-Score of 96.92 (percentage improvement of 7.06%)
while using less than one-tenth of the task-specific training data. We find
that the use of a graph based ap- proach instead of a traditional lattice-based
sequential labelling approach leads to a percentage gain of 12.6% in F-Score
for the segmentation task. | ['Sasi Prasanth Bandaru', 'Pavankumar Satuluri', 'Gaurav Sahu', 'Vishnu Dutt Sharma', 'Amrith Krishna', 'Pawan Goyal', 'Bishal Santra'] | 2018-09-05 | free-as-in-free-word-order-an-energy-based-1 | https://aclanthology.org/D18-1276 | https://aclanthology.org/D18-1276.pdf | emnlp-2018-10 | ['morphological-tagging'] | ['natural-language-processing'] | [ 2.28529006e-01 3.56272578e-01 -1.28154337e-01 -4.88171101e-01
-8.66468906e-01 -8.37246776e-01 3.10925394e-01 5.48847258e-01
-7.88863361e-01 7.15023518e-01 1.27834484e-01 -8.59122753e-01
2.61199921e-01 -7.86742091e-01 -4.40047383e-01 -5.64563453e-01
1.79852515e-01 6.27665520e-01 5.39474368e-01 -1.23634420e-01
4.72665966e-01 3.44226629e-01 -8.22689533e-01 2.07136348e-01
1.08263624e+00 5.73458910e-01 6.45349920e-01 7.51630127e-01
-4.79742169e-01 3.83547425e-01 -4.25905645e-01 -6.28765941e-01
1.70785397e-01 -5.27979136e-01 -1.03273571e+00 1.06569216e-01
3.79336029e-01 3.24499011e-02 -9.84199494e-02 1.00693667e+00
2.72228032e-01 4.09702986e-01 4.10949945e-01 -1.62704036e-01
-9.34476703e-02 6.65958345e-01 -6.92497253e-01 4.47563589e-01
9.40515921e-02 -1.85446873e-01 1.40836740e+00 -5.08240521e-01
8.10220182e-01 8.46362948e-01 5.08444488e-01 5.16570926e-01
-1.13701069e+00 -1.17895938e-01 3.41314107e-01 -1.37802079e-01
-1.15582275e+00 -2.11091429e-01 7.12944567e-01 -2.45483100e-01
1.58376479e+00 3.03047836e-01 6.43930078e-01 3.08354646e-01
4.33302373e-01 8.28571022e-01 1.19011581e+00 -7.76619732e-01
3.16385388e-01 -3.90070349e-01 5.65338612e-01 8.84913683e-01
2.49544173e-01 -3.29750270e-01 -3.57252538e-01 -6.60483018e-02
5.03227711e-01 -4.61658865e-01 3.03253680e-01 -5.94500676e-02
-7.81506717e-01 1.01628888e+00 1.76971182e-01 6.04384959e-01
-2.95124978e-01 1.43571094e-01 2.84940094e-01 1.48944721e-01
7.82093346e-01 5.59781015e-01 -8.37587595e-01 -3.04516941e-01
-1.26742220e+00 1.90430880e-01 7.06549466e-01 6.69943094e-01
5.04785776e-01 2.08985191e-02 1.93878442e-01 8.42879057e-01
4.43725705e-01 1.74448520e-01 2.20782906e-01 -8.23243797e-01
6.23699188e-01 4.42955762e-01 -1.61329985e-01 -5.95858276e-01
-4.90049988e-01 -4.55810010e-01 -3.22920054e-01 -4.53298092e-01
7.53078580e-01 -2.77247190e-01 -1.33056140e+00 1.45585823e+00
4.80772167e-01 -1.16421402e-01 -5.51148914e-02 5.13074577e-01
4.04131144e-01 8.92299950e-01 2.96517432e-01 -4.92083162e-01
1.45928049e+00 -9.98434246e-01 -5.31862617e-01 -5.84261179e-01
7.35589385e-01 -8.61373007e-01 9.42642450e-01 5.72066605e-01
-1.11365402e+00 -2.10730344e-01 -8.46068919e-01 -9.32179019e-02
4.66063991e-02 -2.38328293e-01 8.82343590e-01 9.66058254e-01
-9.80003178e-01 6.51587427e-01 -1.24843514e+00 -4.79751825e-01
2.02217013e-01 5.57818890e-01 -1.17053986e-01 -4.83098477e-02
-6.74105167e-01 7.46786177e-01 6.55935824e-01 -1.47072703e-01
-2.50819832e-01 -3.47124457e-01 -9.03598547e-01 -1.23865969e-01
7.34144211e-01 -2.01833799e-01 1.39472401e+00 -7.05179036e-01
-1.47173285e+00 8.11126709e-01 -3.94663960e-01 -3.92297745e-01
1.80223376e-01 -2.13270992e-01 -1.02429226e-01 2.08443120e-01
-1.00073390e-01 6.01834297e-01 6.40095294e-01 -8.81871045e-01
-5.74778497e-01 -4.22539979e-01 1.06052503e-01 3.16252232e-01
-6.28372654e-02 4.95302975e-01 -6.15858912e-01 -6.15716398e-01
4.96518075e-01 -1.04724622e+00 -7.13852108e-01 -8.21867049e-01
-3.90154064e-01 -1.85328618e-01 2.19251409e-01 -1.31381810e+00
1.43362975e+00 -1.68763494e+00 1.34882122e-01 1.88122317e-01
3.59785855e-02 2.98658013e-01 -7.73118669e-03 4.81314182e-01
1.25639558e-01 4.51431096e-01 -5.54993868e-01 -1.11550130e-01
-2.16010779e-01 4.20614153e-01 1.97216153e-01 2.34663308e-01
1.67820230e-01 1.01103175e+00 -9.09920752e-01 -5.36983728e-01
1.37142599e-01 9.74950120e-02 -5.49287379e-01 -1.04518734e-01
-4.34132814e-01 4.60175633e-01 -5.70602894e-01 6.08900666e-01
4.39847589e-01 1.02251414e-02 8.74173939e-01 3.05035830e-01
-1.82779431e-01 9.85482812e-01 -9.47661877e-01 1.79623568e+00
-4.03127193e-01 1.79321513e-01 3.23157422e-02 -8.75011563e-01
7.68785655e-01 1.29556879e-01 1.84072152e-01 -6.82376146e-01
1.23803839e-02 3.39688033e-01 5.56518912e-01 -1.39210984e-01
7.28154480e-01 -3.13760102e-01 -3.83080661e-01 4.18994516e-01
6.28941357e-02 -1.65740177e-01 5.17871737e-01 2.00119242e-01
1.38634777e+00 1.57933444e-01 5.62195599e-01 -5.20698905e-01
2.49904022e-01 2.17560604e-01 7.42004156e-01 6.75388992e-01
-4.18500789e-02 4.22945738e-01 5.62183261e-01 -3.02500814e-01
-1.21724761e+00 -5.40013492e-01 -2.13644989e-02 1.28099167e+00
-3.24590206e-01 -6.49649858e-01 -1.05545413e+00 -1.13954186e+00
-4.02825534e-01 9.27288711e-01 -1.85108408e-01 3.66109967e-01
-1.16393757e+00 -1.13462532e+00 2.90988684e-01 5.69167435e-01
2.84396350e-01 -1.05527318e+00 -5.60562015e-01 5.12879491e-01
-2.65599545e-02 -1.09642732e+00 -4.23737586e-01 5.81633508e-01
-1.22861004e+00 -6.58793271e-01 -3.91201586e-01 -8.99049878e-01
6.07615650e-01 -1.35803118e-01 1.15137684e+00 3.95130098e-01
2.02483341e-01 -5.73408715e-02 -6.77383900e-01 -3.96148562e-02
-4.77862328e-01 4.51973081e-01 -4.54193324e-01 -4.32761401e-01
3.42273235e-01 -2.04528540e-01 -3.93386841e-01 -1.77891299e-01
-7.08288133e-01 5.15268110e-02 5.19905031e-01 7.88477004e-01
7.27388322e-01 8.01843628e-02 3.58843118e-01 -1.27036619e+00
4.02544200e-01 -1.14031285e-01 -6.59548879e-01 1.30546227e-01
-8.26119065e-01 1.86117992e-01 4.90817219e-01 -1.22208796e-01
-1.12466943e+00 3.59598219e-01 -5.97305119e-01 6.11842573e-01
-2.71017969e-01 7.10876107e-01 -2.63382405e-01 -8.13455507e-02
3.32173146e-02 1.19964154e-02 -5.58626771e-01 -8.46327960e-01
1.60644725e-01 6.47724271e-01 1.98246270e-01 -5.93199611e-01
5.37592590e-01 -1.16221748e-01 -3.47549133e-02 -8.34034801e-01
-7.94283509e-01 -6.84901893e-01 -9.56138015e-01 5.87055907e-02
9.98553157e-01 -4.85224873e-01 -5.06284647e-02 3.92972350e-01
-1.14286160e+00 -5.98219097e-01 -1.98690668e-02 2.88553327e-01
-2.57789493e-01 8.38077843e-01 -1.01509106e+00 -9.17350709e-01
-3.04933548e-01 -8.77185583e-01 1.06512749e+00 1.20209113e-01
-2.35490814e-01 -9.84800398e-01 1.04756735e-01 6.92828655e-01
-8.91232193e-02 6.94600726e-03 1.18948865e+00 -1.08170760e+00
-3.19711536e-01 1.25757670e-02 5.36159202e-02 1.68844372e-01
-1.87976044e-02 -2.74282694e-01 -5.97379088e-01 -1.08103663e-01
1.12314954e-01 -9.40506235e-02 1.02505779e+00 3.56058329e-01
8.00921917e-01 -2.17912927e-01 -1.39598861e-01 1.37761643e-04
1.42008495e+00 5.37242651e-01 5.68588674e-01 2.12136760e-01
7.57526517e-01 8.35694551e-01 8.80546212e-01 -9.94339809e-02
5.21751404e-01 6.41875029e-01 8.84323642e-02 3.08061808e-01
-1.76298469e-02 -2.63031125e-01 2.93288201e-01 1.34517312e+00
-1.97652146e-01 -5.72018921e-01 -1.36054480e+00 7.37973511e-01
-1.87201166e+00 -5.05148947e-01 -4.35960025e-01 2.01508045e+00
8.74893129e-01 4.10104215e-01 1.23046879e-02 1.56256676e-01
7.05864429e-01 3.52884859e-01 -1.32250801e-01 -9.72172141e-01
2.31408536e-01 6.76477790e-01 6.94692075e-01 8.68355691e-01
-1.11754453e+00 1.54294205e+00 6.10488892e+00 1.02290106e+00
-7.71315098e-01 1.76411614e-01 6.80769980e-01 2.48943478e-01
-3.57508630e-01 2.88333207e-01 -8.29255641e-01 4.79368031e-01
1.32713330e+00 3.51300359e-01 4.11557585e-01 4.03322995e-01
2.20519766e-01 -5.02048612e-01 -6.10644042e-01 5.49612224e-01
-8.54343176e-02 -9.27916169e-01 -3.18365872e-01 4.22107667e-01
6.43174171e-01 1.84102997e-01 -3.91185492e-01 1.44405216e-01
5.07178903e-01 -9.83998120e-01 6.32538497e-01 3.04234196e-02
4.15374875e-01 -8.24528158e-01 8.33950043e-01 7.08074272e-01
-1.11863244e+00 3.42697650e-01 -2.84003645e-01 -2.90276021e-01
5.33577800e-01 7.80878961e-01 -8.48585010e-01 6.70853972e-01
3.44636887e-01 2.96599716e-01 -4.59220558e-01 9.05074835e-01
-6.19502306e-01 1.30418050e+00 -4.33080614e-01 -1.70944422e-01
5.70781708e-01 -3.24847877e-01 4.02174324e-01 1.39203930e+00
9.98685434e-02 2.88876086e-01 4.87101734e-01 2.44693495e-02
-6.68472424e-02 5.04813194e-01 -1.27566546e-01 -3.40365708e-01
1.88318059e-01 1.30669463e+00 -1.58536422e+00 -2.14329332e-01
-4.25555944e-01 1.09042072e+00 4.69239444e-01 -5.29827476e-02
-4.67610747e-01 -1.38988525e-01 -5.42094484e-02 5.20244129e-02
8.46327841e-01 -7.87465394e-01 -5.54731905e-01 -9.86687005e-01
6.61183670e-02 -8.57682347e-01 2.47894540e-01 -2.51789182e-01
-9.97324944e-01 5.42939663e-01 -4.33310382e-02 -3.62202585e-01
-4.80857044e-01 -7.18993127e-01 -5.00419796e-01 7.79751301e-01
-1.36788023e+00 -1.12046456e+00 4.89120871e-01 -1.81421995e-01
6.72659516e-01 1.38583928e-01 9.47587907e-01 9.59736258e-02
-5.55572927e-01 3.39502007e-01 5.19604310e-02 3.12335372e-01
9.72013324e-02 -1.66902280e+00 9.44476426e-01 1.14824343e+00
5.63562155e-01 5.26946545e-01 6.75263703e-01 -1.04473853e+00
-1.27924657e+00 -8.71387422e-01 1.54451549e+00 -3.73691827e-01
7.58520603e-01 -5.55216491e-01 -8.96342278e-01 5.94140112e-01
2.67696679e-01 -2.77672768e-01 8.20026755e-01 3.55138838e-01
-7.02451766e-02 3.00858617e-01 -1.14478505e+00 3.99757087e-01
1.21472371e+00 -2.57358432e-01 -7.74037421e-01 1.68584958e-01
6.34658694e-01 -4.45839763e-01 -9.12315845e-01 3.04554701e-01
3.22092950e-01 -6.51312351e-01 4.53746080e-01 -6.07904971e-01
5.42596802e-02 -1.01216324e-01 -3.10462117e-01 -1.15883493e+00
-3.36607963e-01 -6.93122566e-01 -2.96673756e-02 1.22857344e+00
7.31579900e-01 -4.32679981e-01 1.03860295e+00 5.34926295e-01
-3.10764909e-01 -8.23071063e-01 -1.07535124e+00 -6.45540833e-01
3.15929055e-01 -7.18181670e-01 1.79628447e-01 7.39079237e-01
-6.03405461e-02 6.57482266e-01 -4.16068137e-02 -1.02848828e-01
3.80593598e-01 1.43439576e-01 1.50750577e-01 -1.23368096e+00
-5.73158920e-01 -1.75884385e-02 -2.52805442e-01 -1.11940658e+00
3.12702358e-01 -9.64864433e-01 2.45694861e-01 -1.77724934e+00
2.53911614e-01 -5.60621321e-01 -3.66237342e-01 7.01139033e-01
-1.15006514e-01 3.21797192e-01 3.50586027e-01 -1.93892449e-01
-7.48279989e-01 -8.36957339e-03 1.00350797e+00 -1.25936558e-02
-2.42978588e-01 -2.07391039e-01 -8.17632854e-01 8.89482200e-01
1.10405171e+00 -6.95759773e-01 -1.54704899e-01 -5.91220796e-01
3.51686239e-01 1.12355329e-01 -2.59275764e-01 -4.57662642e-01
-1.07461020e-01 -4.00760621e-01 9.21747163e-02 -5.58854997e-01
6.64437637e-02 -4.55257326e-01 -3.57427709e-02 5.53599000e-01
6.25199545e-03 8.85295793e-02 2.00652242e-01 4.28531259e-01
-4.09681089e-02 -8.48880351e-01 5.39134204e-01 -3.28705043e-01
-7.98283637e-01 -8.62978771e-02 -5.12168169e-01 8.24818090e-02
8.07212651e-01 -2.14132115e-01 4.79282103e-02 -4.01398577e-02
-6.57066524e-01 -7.51657933e-02 5.32013178e-01 -1.08564366e-02
2.44903520e-01 -6.04833007e-01 -4.51977342e-01 -1.41226739e-01
-3.48916918e-01 4.41562161e-02 -1.25950739e-01 5.68792641e-01
-8.36854041e-01 6.66802227e-01 1.41770303e-01 -1.64986238e-01
-1.38596451e+00 2.54142880e-01 -4.80044149e-02 -9.55305278e-01
-4.76571769e-01 8.30806851e-01 -1.58638656e-01 -4.82368857e-01
-3.65814120e-01 -4.08811718e-01 -1.21748693e-01 6.67565316e-02
1.20141700e-01 2.81258404e-01 4.88070130e-01 -8.23046386e-01
-5.49854279e-01 4.84740466e-01 -3.31000149e-01 -3.87081623e-01
1.49982560e+00 3.54839787e-02 -2.19516948e-01 2.97346979e-01
6.39465690e-01 5.21641970e-01 -9.98880744e-01 6.58457261e-03
6.33135974e-01 -2.98281103e-01 3.09768707e-01 -1.07492769e+00
-8.28194857e-01 8.07964325e-01 3.95685770e-02 3.16194624e-01
1.01661968e+00 2.49485806e-01 1.08155262e+00 4.27676857e-01
6.29774392e-01 -1.18597901e+00 -3.87343198e-01 9.56692815e-01
2.44962230e-01 -9.95393038e-01 1.62185039e-02 -7.70070076e-01
-6.24325812e-01 8.00670326e-01 2.24019617e-01 -1.58385456e-01
4.19460654e-01 3.36690485e-01 -2.06502900e-01 -1.06683098e-01
-5.99581540e-01 -4.09651637e-01 4.65260237e-01 2.64492720e-01
7.10015297e-01 3.02049071e-01 -1.19139981e+00 7.02095807e-01
-2.48792470e-01 -6.05880260e-01 3.20654303e-01 1.20446146e+00
-7.39784598e-01 -1.85793495e+00 -4.33771983e-02 6.70819700e-01
-1.04285669e+00 -5.32115400e-01 -6.52022958e-01 5.51407814e-01
5.37342913e-02 9.96481001e-01 1.14154056e-01 -2.32955039e-01
-2.15866286e-02 5.50125837e-01 7.49287426e-01 -1.22370136e+00
-8.36461961e-01 6.51430547e-01 6.80722415e-01 -1.63515538e-01
-5.72368681e-01 -9.68025923e-01 -1.68535614e+00 -1.24862514e-01
-5.54771125e-01 3.68001580e-01 7.88897753e-01 1.13591456e+00
-3.60400975e-02 3.94781768e-01 1.74104467e-01 -5.12929261e-01
-1.80364519e-01 -9.16093647e-01 -5.98864019e-01 -1.74006119e-01
-1.63409412e-01 -2.82466710e-01 1.68256029e-01 -3.93179990e-02] | [10.342808723449707, 9.77509880065918] |
b27901a6-28f0-4674-ae1e-b0a2bdd6d62d | hle-upc-at-semeval-2021-task-5-multi-depth | 2104.00639 | null | https://arxiv.org/abs/2104.00639v3 | https://arxiv.org/pdf/2104.00639v3.pdf | HLE-UPC at SemEval-2021 Task 5: Multi-Depth DistilBERT for Toxic Spans Detection | This paper presents our submission to SemEval-2021 Task 5: Toxic Spans Detection. The purpose of this task is to detect the spans that make a text toxic, which is a complex labour for several reasons. Firstly, because of the intrinsic subjectivity of toxicity, and secondly, due to toxicity not always coming from single words like insults or offends, but sometimes from whole expressions formed by words that may not be toxic individually. Following this idea of focusing on both single words and multi-word expressions, we study the impact of using a multi-depth DistilBERT model, which uses embeddings from different layers to estimate the final per-token toxicity. Our quantitative results show that using information from multiple depths boosts the performance of the model. Finally, we also analyze our best model qualitatively. | ['Albert Rial-Farràs', 'Rafel Palliser-Sans'] | 2021-04-01 | null | https://aclanthology.org/2021.semeval-1.131 | https://aclanthology.org/2021.semeval-1.131.pdf | semeval-2021 | ['toxic-spans-detection'] | ['natural-language-processing'] | [-1.38894007e-01 -2.80760556e-01 1.85510274e-02 1.64988369e-01
-6.90543592e-01 -9.45999861e-01 7.93891907e-01 8.06905866e-01
-6.56797171e-01 7.91223288e-01 8.87539923e-01 -7.38620162e-02
-3.89029458e-02 -6.51443720e-01 -6.03742361e-01 -6.76069260e-01
-2.54894495e-01 2.27270201e-01 9.87558588e-02 -2.92917162e-01
1.72094986e-01 6.45737529e-01 -1.10273540e+00 5.08551896e-01
5.34572244e-01 6.01766407e-01 -3.10072839e-01 5.13019025e-01
-6.36369884e-02 8.34385216e-01 -8.99605930e-01 -7.18720436e-01
-8.75614211e-02 -2.33668089e-01 -7.50336051e-01 -3.11077833e-01
3.31503838e-01 -1.03046745e-01 -2.53077805e-01 1.06027865e+00
7.54758418e-01 2.86089871e-02 1.22784543e+00 -9.56941366e-01
-5.29835939e-01 1.03942144e+00 -5.22879422e-01 3.15177411e-01
6.52595162e-01 3.59037042e-01 1.28042877e+00 -8.07457149e-01
5.58649421e-01 1.50663519e+00 8.00679386e-01 6.51166618e-01
-1.25352418e+00 -6.26341820e-01 2.45897457e-01 6.57374114e-02
-1.34380329e+00 -2.21763536e-01 3.51469189e-01 -8.18484902e-01
1.19576919e+00 2.82152623e-01 3.43874335e-01 1.73855424e+00
5.14876783e-01 5.90157688e-01 1.01941025e+00 3.36503424e-02
2.40126953e-01 -1.70401290e-01 2.24423751e-01 4.25816476e-01
3.65961343e-01 -1.11448802e-01 -5.59715509e-01 -3.38289738e-01
-3.17969799e-01 -2.04034969e-01 -3.47837716e-01 6.22086287e-01
-5.58819532e-01 8.61176670e-01 2.65106291e-01 4.69221383e-01
-4.31521773e-01 4.53174323e-01 7.21722960e-01 1.59371402e-02
8.40008497e-01 6.10939503e-01 -6.46118283e-01 -1.93818659e-01
-8.43072474e-01 4.76783514e-01 7.08271921e-01 3.78420442e-01
3.26205045e-01 -3.46485317e-01 -5.48165143e-01 8.44460726e-01
2.66216367e-01 1.60571244e-02 2.55689591e-01 -1.16301313e-01
5.77855706e-01 6.36497974e-01 -6.14401465e-03 -6.07698917e-01
-8.67430925e-01 -3.64032090e-01 -5.35978019e-01 8.53303000e-02
4.70580220e-01 -5.51237047e-01 -8.66778672e-01 1.88203990e+00
4.94882464e-02 -1.44954771e-01 -3.65438730e-01 6.08322263e-01
7.45837629e-01 5.38071334e-01 8.70265424e-01 -2.83003077e-02
1.48445487e+00 -4.12876308e-01 -5.41845083e-01 -2.03886107e-01
1.20560765e+00 -5.77514946e-01 1.04510629e+00 6.42417908e-01
-7.40284085e-01 1.29214779e-01 -1.10419130e+00 -1.95856959e-01
-1.04986596e+00 -3.91648769e-01 2.78477311e-01 7.54274070e-01
-7.88015246e-01 8.54854703e-01 -1.60001516e-01 -2.49762326e-01
6.05870485e-01 1.91196203e-01 -3.65750343e-01 -6.68596663e-03
-1.94367957e+00 1.50988197e+00 2.31343433e-01 -4.36268449e-01
-1.00008130e+00 -1.04321599e+00 -6.73834085e-01 7.04179704e-02
-3.09815109e-02 -4.51489836e-01 8.09977472e-01 -3.32194835e-01
-7.56484866e-01 6.32506251e-01 8.18114430e-02 -3.49314719e-01
6.86022937e-01 -4.48031634e-01 -1.82737440e-01 -1.62753925e-01
-8.59624967e-02 3.48531336e-01 1.66858539e-01 -8.63123238e-01
-2.06641853e-01 -5.65236986e-01 1.79854184e-01 1.31114051e-01
-7.91244507e-01 4.70906615e-01 -1.41279027e-01 -4.99047816e-01
-9.85984981e-01 -6.57872021e-01 -3.32772195e-01 -2.84785986e-01
-7.21886158e-01 -6.57480776e-01 3.95253330e-01 -8.04398119e-01
1.61663556e+00 -2.01871014e+00 2.46318519e-01 -1.21483430e-01
3.77794832e-01 6.97411038e-03 -3.09078485e-01 8.71882856e-01
-4.11656111e-01 6.79911852e-01 -3.24928671e-01 -5.59577942e-01
2.34284371e-01 -1.39857799e-01 -3.87628615e-01 6.39320552e-01
4.32581127e-01 9.42323625e-01 -9.34039950e-01 -3.25335890e-01
-1.62638962e-01 5.38408399e-01 -7.12029561e-02 -2.66718268e-01
-4.62818235e-01 -1.23297691e-01 -3.18545848e-01 5.13403654e-01
6.13630056e-01 4.36362535e-01 -2.75920421e-01 2.28281226e-02
-2.56356686e-01 5.46607077e-01 -6.69223964e-01 1.35234797e+00
-5.72638810e-01 4.56542969e-01 -3.97669256e-01 -4.74274427e-01
3.76357704e-01 3.37257445e-01 4.44118232e-01 -6.95688188e-01
2.50042975e-01 2.79289042e-03 -1.15224048e-01 -4.14153367e-01
3.18797380e-01 -6.70939922e-01 -4.53815192e-01 5.90603352e-01
-2.53550112e-02 -1.13493383e-01 5.74362934e-01 3.05454791e-01
1.73570621e+00 -2.49208540e-01 8.29681158e-02 -3.78278226e-01
2.94013530e-01 -3.69664371e-01 4.66566980e-01 3.05101335e-01
-1.34369656e-02 5.59342265e-01 1.16465843e+00 -1.35239050e-01
-7.93431699e-01 -1.04053128e+00 -2.36749977e-01 1.11898780e+00
-3.70655954e-01 -7.94445992e-01 -7.61103213e-01 -1.02048063e+00
1.72397107e-01 9.64923143e-01 -1.07432449e+00 -3.48821998e-01
-9.49661061e-02 -1.30849349e+00 1.15209496e+00 5.28142035e-01
-2.22648293e-01 -9.46774364e-01 -5.73279619e-01 2.30143994e-01
-7.32768252e-02 -9.26353335e-01 -5.45395553e-01 7.69496799e-01
-5.21229923e-01 -9.80707228e-01 -5.69678068e-01 -6.26086816e-02
1.93444967e-01 -3.12084407e-01 1.14031911e+00 1.63737878e-01
-5.49693048e-01 -1.20944381e-01 -3.59009475e-01 -7.27032423e-01
-3.45842421e-01 -3.10116801e-02 1.12564832e-01 -2.77405024e-01
5.71860552e-01 -3.16339552e-01 -2.87187159e-01 -1.45450264e-01
-1.34036791e+00 -5.74056506e-01 4.28119302e-01 2.92837888e-01
8.78863335e-02 2.04697326e-01 6.91693664e-01 -1.05975413e+00
1.14705026e+00 -9.53650415e-01 -1.38147131e-01 9.70756263e-02
-3.38899523e-01 5.89816123e-02 8.42738569e-01 -5.47501147e-01
-6.64821923e-01 3.30080539e-02 -4.61868554e-01 -1.35261542e-03
-6.57687858e-02 6.66279376e-01 -5.01960218e-01 3.52407634e-01
1.02211535e+00 -1.29530013e-01 -4.76213604e-01 -4.28807527e-01
7.02199519e-01 7.28335559e-01 -1.15194216e-01 -5.13377309e-01
6.30174696e-01 1.69737101e-01 7.27741569e-02 -9.12653863e-01
-9.66553450e-01 -4.55041379e-01 -4.95988697e-01 -2.78292775e-01
9.53535140e-01 -8.97510409e-01 -5.76942325e-01 6.28553867e-01
-1.65007162e+00 -5.01756966e-01 -2.98659243e-02 -3.53467092e-02
-1.55043051e-01 4.63001072e-01 -9.34379399e-01 -7.15565383e-01
-3.46776605e-01 -7.86481142e-01 9.00558293e-01 -1.31546691e-01
-8.28836381e-01 -1.23783135e+00 3.76823485e-01 -5.67723028e-02
1.31569415e-01 7.20952213e-01 1.21706450e+00 -7.94999540e-01
2.47950584e-01 -3.76150131e-01 -1.23044923e-01 3.19565594e-01
1.53979212e-01 8.48609135e-02 -1.12543380e+00 6.97518662e-02
-1.84746474e-01 -5.93279123e-01 1.26410246e+00 -2.62446832e-02
9.44607913e-01 -4.75403756e-01 -2.10521564e-01 1.31623387e-01
1.56868577e+00 -2.44066790e-01 9.22518492e-01 3.07602257e-01
6.36561275e-01 9.72508669e-01 2.48205796e-01 7.89857447e-01
1.74225837e-01 4.20719057e-01 6.56726420e-01 9.90186185e-02
-2.94127941e-01 -2.08132625e-01 8.85047376e-01 1.39284223e-01
2.93632656e-01 -6.88542724e-01 -8.70952070e-01 8.40018690e-01
-1.35311377e+00 -9.95304406e-01 -6.70772672e-01 2.01519060e+00
1.21495140e+00 2.48812720e-01 3.16987306e-01 2.67598212e-01
2.84673035e-01 3.40760916e-01 -3.92630607e-01 -1.02930808e+00
-3.76290172e-01 4.08729136e-01 5.89839578e-01 4.38241869e-01
-1.07781947e+00 8.15014184e-01 6.37498808e+00 1.18023825e+00
-8.60797048e-01 2.50030130e-01 6.31445467e-01 -3.41126233e-01
-6.09900475e-01 -4.03718680e-01 -7.13838458e-01 7.23707736e-01
1.15783393e+00 -1.93453163e-01 2.60088712e-01 2.34562427e-01
3.88730586e-01 -1.78069621e-01 -1.32883322e+00 3.64769340e-01
1.35243863e-01 -6.82708859e-01 2.05462217e-01 2.14574441e-01
5.37112355e-01 2.28008404e-01 1.94622278e-01 8.29834715e-02
5.87591231e-01 -1.42850435e+00 1.02626336e+00 2.49198467e-01
5.89137316e-01 -9.81198847e-01 6.98006034e-01 3.59831661e-01
-7.71604657e-01 -1.89483151e-01 -3.36513668e-01 -2.66742021e-01
5.08175492e-02 1.25122237e+00 -4.89418954e-01 1.89914376e-01
3.90915394e-01 6.11535490e-01 -6.64791286e-01 1.11161220e+00
-7.47395277e-01 6.24836802e-01 -2.26223037e-01 -4.29090649e-01
3.24241251e-01 2.70082355e-01 3.98413956e-01 1.92590892e+00
2.09380358e-01 -8.13449621e-02 -4.64042395e-01 8.71997952e-01
-4.09593642e-01 1.14358574e-01 -8.76608193e-01 -4.05609280e-01
2.52968103e-01 1.25235081e+00 -4.77415323e-01 -2.35387962e-02
-2.98529267e-01 1.02211559e+00 3.23182255e-01 1.90006629e-01
-8.34392250e-01 -5.32564104e-01 9.13174033e-01 2.73030877e-01
3.61425988e-02 -1.90086424e-01 -6.48999929e-01 -6.57592535e-01
-2.33879641e-01 -6.05293989e-01 4.33964998e-01 -5.11791527e-01
-1.51644015e+00 5.21406770e-01 -7.50374421e-02 -8.80017698e-01
1.93999931e-01 -8.43823135e-01 -7.22979426e-01 1.03774655e+00
-1.23196840e+00 -9.63176608e-01 3.43499541e-01 2.18698457e-01
3.14430922e-01 5.93795955e-01 8.98326039e-01 5.11909187e-01
-5.65147400e-01 6.67583346e-01 -2.02921540e-01 -1.09656185e-01
7.47694135e-01 -1.35832834e+00 5.04776120e-01 7.50955880e-01
9.16673522e-03 6.08816028e-01 9.74605024e-01 -7.60859072e-01
-9.16261971e-01 -1.26356745e+00 1.53033352e+00 -1.17690146e+00
1.08945239e+00 -6.87746167e-01 -6.78640485e-01 1.41539961e-01
2.45310009e-01 -5.52393854e-01 8.39513361e-01 2.96193659e-01
-8.02768350e-01 3.69768143e-01 -1.13501441e+00 8.35652232e-01
9.12778556e-01 -5.78470349e-01 -4.70852643e-01 4.88497615e-01
7.39210546e-01 7.77235925e-02 -6.86088085e-01 8.19490254e-02
2.50107020e-01 -6.86086714e-01 7.21330345e-01 -9.11326408e-01
9.98394728e-01 -1.01843476e-01 -6.56124949e-02 -1.76509130e+00
-3.65725785e-01 -2.06474483e-01 1.48132026e-01 1.25613725e+00
8.16532671e-01 -3.27636540e-01 3.90566558e-01 5.81032336e-01
-1.18236482e-01 -8.35154355e-01 -1.07463968e+00 -9.15692985e-01
7.89337873e-01 -6.58191383e-01 4.99659687e-01 8.50501895e-01
3.33422184e-01 7.10871696e-01 -1.40869230e-01 -5.95570058e-02
5.10063648e-01 -6.81555152e-01 -6.73358142e-02 -1.27064538e+00
-2.72891000e-02 -7.31558323e-01 -3.25021774e-01 -4.00557727e-01
3.13722789e-01 -8.43810081e-01 2.65189528e-01 -1.53324497e+00
5.22871852e-01 6.41945228e-02 -5.88434994e-01 6.95595205e-01
-3.17206591e-01 4.64998603e-01 1.50387153e-01 -4.41199511e-01
-4.48362708e-01 4.18459356e-01 7.43771017e-01 -3.69948626e-01
1.77728146e-01 -3.70837271e-01 -1.15712404e+00 6.77095473e-01
7.40829349e-01 -8.13646913e-01 -1.63786471e-01 -4.54089373e-01
8.28954279e-01 -5.52403629e-01 1.62522614e-01 -6.76614523e-01
-5.21914363e-02 -1.04230359e-01 3.29700887e-01 -3.29862237e-01
2.36930713e-01 -3.70995730e-01 -8.21576715e-02 6.71149850e-01
-3.61054540e-01 -2.09200848e-02 4.33802903e-01 4.91286725e-01
2.17923075e-01 -4.69341934e-01 7.30502963e-01 -1.32437825e-01
-1.41977519e-01 2.24762008e-01 -7.54224122e-01 1.68083534e-01
7.90974379e-01 1.34525046e-01 -6.24381423e-01 -1.05454430e-01
-4.03226703e-01 2.87275881e-01 5.06205440e-01 5.77815533e-01
5.50341249e-01 -1.13730764e+00 -9.74165559e-01 -7.07597435e-01
3.10950786e-01 -4.68918860e-01 1.72811374e-01 8.19188654e-01
-2.82915682e-01 2.75631964e-01 1.54241726e-01 2.50466555e-01
-1.28914177e+00 7.37582207e-01 2.13761076e-01 -5.00228465e-01
-1.51247099e-01 1.11992538e+00 2.98494577e-01 -1.65350720e-01
3.41367096e-01 -1.99796587e-01 -5.11701167e-01 8.20386171e-01
7.16493905e-01 5.17055333e-01 3.09495538e-01 -6.28420353e-01
-7.95033991e-01 4.64246809e-01 -7.30001852e-02 -1.04263611e-01
1.30917847e+00 3.40288669e-01 -3.40426058e-01 3.95762384e-01
1.46604717e+00 2.06985340e-01 -7.43027568e-01 2.44557545e-01
3.74334931e-01 -3.01531460e-02 -1.11259555e-03 -1.34329915e+00
-6.31878853e-01 1.21493042e+00 1.56241134e-01 3.69665176e-02
7.99827218e-01 3.87778729e-02 8.09614897e-01 -6.04107045e-02
9.95387882e-02 -1.19570851e+00 1.26842618e-01 7.17525303e-01
1.02461720e+00 -3.94859880e-01 1.33823887e-01 -1.19880259e-01
-5.33655703e-01 8.77744436e-01 3.09051156e-01 1.64187387e-01
3.47854763e-01 3.83990288e-01 -2.36633494e-02 -3.33181858e-01
-1.18244576e+00 -1.92858726e-01 3.83424647e-02 4.43005830e-01
6.44934475e-01 2.23055050e-01 -9.39161539e-01 8.87774706e-01
1.50022190e-02 -3.10176671e-01 9.20491099e-01 4.06747162e-01
-4.15846467e-01 -1.17797482e+00 -9.65716317e-02 4.95560795e-01
-9.93290484e-01 -4.09040302e-01 -1.22354043e+00 3.13760608e-01
5.42909741e-01 1.13084555e+00 -3.66198689e-01 -6.92288101e-01
5.27002871e-01 1.26005888e-01 4.05231386e-01 -7.79148281e-01
-1.07408869e+00 -4.10397083e-01 4.21584338e-01 -3.62999171e-01
1.79671794e-01 -6.66812062e-01 -1.35033798e+00 -3.18750083e-01
-1.72032595e-01 7.56102800e-02 8.60347092e-01 9.60452080e-01
-4.59987932e-04 6.34263098e-01 5.47456622e-01 -2.67924100e-01
-7.48872340e-01 -1.14711893e+00 -7.04788923e-01 4.88576293e-01
2.01217562e-01 -5.18259168e-01 -6.84353888e-01 -2.98474789e-01] | [8.914163589477539, 10.635584831237793] |
2e3484af-1c0e-4e49-8b37-6ea4ef860d4c | layernas-neural-architecture-search-in | 2304.11517 | null | https://arxiv.org/abs/2304.11517v1 | https://arxiv.org/pdf/2304.11517v1.pdf | LayerNAS: Neural Architecture Search in Polynomial Complexity | Neural Architecture Search (NAS) has become a popular method for discovering effective model architectures, especially for target hardware. As such, NAS methods that find optimal architectures under constraints are essential. In our paper, we propose LayerNAS to address the challenge of multi-objective NAS by transforming it into a combinatorial optimization problem, which effectively constrains the search complexity to be polynomial. For a model architecture with $L$ layers, we perform layerwise-search for each layer, selecting from a set of search options $\mathbb{S}$. LayerNAS groups model candidates based on one objective, such as model size or latency, and searches for the optimal model based on another objective, thereby splitting the cost and reward elements of the search. This approach limits the search complexity to $ O(H \cdot |\mathbb{S}| \cdot L) $, where $H$ is a constant set in LayerNAS. Our experiments show that LayerNAS is able to consistently discover superior models across a variety of search spaces in comparison to strong baselines, including search spaces derived from NATS-Bench, MobileNetV2 and MobileNetV3. | ['Erik Vee', 'Da-Cheng Juan', 'Fotis Iliopoulos', 'Xin Wang', 'Yun Long', 'Keshav Kumar', 'Daiyi Peng', 'Jingyue Shen', 'Dana Alon', 'Yicheng Fan'] | 2023-04-23 | null | null | null | null | ['combinatorial-optimization', 'architecture-search'] | ['methodology', 'methodology'] | [-8.25872645e-02 -3.52196664e-01 -4.58640367e-01 -2.52975464e-01
-9.44898784e-01 -6.89159393e-01 -8.58063623e-02 -2.19713360e-01
-7.40090489e-01 6.06228232e-01 -3.53019476e-01 -6.42502010e-01
-3.50759059e-01 -5.49164593e-01 -7.06869602e-01 -4.06941593e-01
-2.76835352e-01 5.73266089e-01 3.32419366e-01 1.65896281e-03
3.77078295e-01 5.12098908e-01 -1.44707918e+00 -1.81935206e-01
7.11124837e-01 1.52347875e+00 3.71675760e-01 5.82701504e-01
-2.66948342e-01 4.84214216e-01 -7.68553793e-01 -3.61478925e-01
4.88690853e-01 -2.42435992e-01 -7.97688663e-01 -4.27475601e-01
5.44584692e-01 -1.01151853e-03 -2.99207479e-01 1.12799668e+00
6.89107537e-01 3.87995876e-02 3.08731526e-01 -1.29058528e+00
1.20777339e-01 7.87604868e-01 -4.47206497e-01 6.25752985e-01
-5.52548170e-01 2.69504964e-01 1.41369164e+00 -6.87088490e-01
3.20141613e-01 1.33753300e+00 3.50458235e-01 4.98539925e-01
-1.37890172e+00 -1.16216135e+00 6.57639027e-01 3.00198644e-01
-1.67977405e+00 -6.91800296e-01 5.51633775e-01 -1.82067901e-01
1.29203987e+00 4.42291349e-01 3.95968050e-01 6.39975369e-01
-5.31739406e-02 6.61577165e-01 6.54035032e-01 -3.11414391e-01
6.80034578e-01 -2.08414361e-01 3.63699496e-01 7.52490938e-01
3.75875264e-01 7.36979321e-02 -5.65064430e-01 -3.40245634e-01
6.08923912e-01 -4.50572491e-01 -1.78530172e-01 -1.40790597e-01
-9.40197885e-01 7.57009327e-01 4.85089660e-01 1.39876693e-01
-3.18774849e-01 9.32908833e-01 8.75752419e-02 3.40718895e-01
1.58665664e-02 9.54949796e-01 -6.98316693e-01 -1.96556553e-01
-1.08826506e+00 2.23937526e-01 5.59865594e-01 9.92700160e-01
8.99860859e-01 4.21759814e-01 -6.03767857e-02 7.74023890e-01
3.85263175e-01 5.74293196e-01 1.70662194e-01 -1.10876846e+00
7.00993001e-01 6.73600554e-01 -7.12523460e-02 -5.43748021e-01
-4.27506924e-01 -7.85256028e-01 -6.63541257e-01 1.29428461e-01
6.75438568e-02 -3.92949015e-01 -9.46471334e-01 2.11679101e+00
-1.84318908e-02 3.28382075e-01 -2.98449814e-01 7.69146323e-01
4.82732922e-01 6.70613289e-01 -1.90789253e-01 -2.54417956e-01
1.25258815e+00 -1.08210313e+00 -1.27225310e-01 -6.81389749e-01
5.78913808e-01 -5.30498147e-01 1.08062947e+00 3.47555935e-01
-1.43163061e+00 -2.19850063e-01 -1.13780355e+00 4.05324250e-01
4.62453812e-02 1.36194453e-01 6.55861616e-01 8.12807739e-01
-1.37943041e+00 2.40153834e-01 -1.01128221e+00 1.30563825e-01
3.10156643e-01 9.71965671e-01 3.87695312e-01 1.13845341e-01
-8.46009433e-01 6.55104756e-01 3.79037112e-01 8.92157406e-02
-1.07874668e+00 -6.16459310e-01 -4.48272586e-01 5.10275543e-01
6.64364278e-01 -5.50568938e-01 1.10855734e+00 -5.26295066e-01
-1.41791439e+00 3.95969063e-01 -3.33562732e-01 -8.33930016e-01
2.59930617e-03 3.12825829e-01 -3.76093507e-01 -1.32184982e-01
-2.66449958e-01 7.89076269e-01 7.91662037e-01 -1.11251569e+00
-6.42745018e-01 -1.91595435e-01 4.14749682e-01 6.49863528e-03
-6.49206519e-01 2.91034728e-01 -1.02129114e+00 -3.90658766e-01
3.29748601e-01 -1.27855682e+00 -6.40818477e-01 -3.12997609e-01
-3.44624847e-01 -8.99859369e-02 2.19015136e-01 -2.04882011e-01
1.87876379e+00 -2.00345302e+00 2.19636530e-01 8.23485315e-01
5.42996168e-01 1.44127771e-01 -3.68706435e-01 -1.76200926e-01
4.14055288e-01 4.58348036e-01 2.24990323e-02 -4.54438299e-01
1.78964600e-01 1.06985688e-01 2.49227863e-02 -2.71266530e-04
-2.18991533e-01 8.76695275e-01 -3.07515860e-01 -4.70094711e-01
-3.14508229e-01 7.18862712e-02 -9.43627834e-01 -2.65483819e-02
-5.86968362e-01 -1.29108056e-01 -5.66424429e-01 6.26371145e-01
5.37754893e-01 -7.48730123e-01 3.04950505e-01 -1.42852440e-01
-2.07247764e-01 6.13522530e-01 -1.35820937e+00 1.54388905e+00
-6.35299981e-01 4.99795556e-01 3.45369816e-01 -6.88924670e-01
8.56560111e-01 3.66739230e-03 4.03762668e-01 -7.57260442e-01
2.40677863e-01 5.03549576e-01 3.12307745e-01 2.40566105e-01
1.24111868e-01 3.06141227e-01 -7.21825510e-02 4.97944176e-01
-1.39628738e-01 1.93908200e-01 2.64627337e-01 -1.86476991e-01
1.53703237e+00 -5.77773452e-01 -1.41641095e-01 -4.88125801e-01
4.34085518e-01 -1.13383889e-01 8.93793046e-01 9.41752493e-01
-1.12068422e-01 1.87106073e-01 5.51873088e-01 -4.93370354e-01
-7.68246353e-01 -9.38402653e-01 2.74567217e-01 1.14019024e+00
3.11347544e-01 -5.67190588e-01 -8.24816763e-01 -5.85178316e-01
-4.08971310e-01 4.87555206e-01 -2.87093580e-01 -6.44225404e-02
-1.06518841e+00 -9.00450349e-01 6.64826453e-01 4.34529662e-01
5.10041237e-01 -9.92142618e-01 -9.05392885e-01 3.10636848e-01
2.17247438e-02 -1.03045642e+00 -7.35779762e-01 6.78644776e-01
-9.03178632e-01 -7.15411186e-01 -4.25764769e-01 -8.48647118e-01
6.70620203e-01 1.54151276e-01 1.30208671e+00 4.78397727e-01
-2.67568350e-01 -1.64477572e-01 3.92797366e-02 -2.35524654e-01
-1.33976355e-01 6.30023956e-01 1.27578482e-01 -3.01884145e-01
-7.03498870e-02 -5.70416391e-01 -7.53760636e-01 5.08264959e-01
-5.81564963e-01 -2.63031311e-02 8.23189914e-01 6.97894394e-01
8.70707333e-01 2.17039764e-01 3.18834513e-01 -4.47857946e-01
4.88119006e-01 -1.50726408e-01 -1.27684700e+00 5.04574478e-01
-1.08855760e+00 4.57894742e-01 5.41632652e-01 -5.06327212e-01
-3.04949284e-01 -1.27341412e-02 -3.55479377e-03 -8.05203915e-01
3.73661518e-01 6.57906055e-01 -2.58652985e-01 -3.31896573e-01
6.72899246e-01 1.72357887e-01 -2.08614588e-01 -6.01380050e-01
-2.40249075e-02 1.48048162e-01 3.06540489e-01 -6.64818048e-01
7.98523486e-01 -8.72854367e-02 1.65215030e-01 -4.40038115e-01
-3.27401966e-01 -2.15169266e-01 -1.75857823e-02 1.29239559e-01
5.68994761e-01 -7.96723425e-01 -1.05845737e+00 -3.14737274e-03
-1.00613225e+00 -5.56838751e-01 1.33783162e-01 4.80573028e-01
-1.66150734e-01 -2.37296596e-02 -4.01171982e-01 -7.08915353e-01
-6.05750382e-01 -1.73380041e+00 6.25603914e-01 9.53375623e-02
-6.71802387e-02 -5.99936843e-01 -3.59381646e-01 2.40936890e-01
6.95860922e-01 -2.91358382e-01 1.48682511e+00 -5.58166504e-01
-1.25415504e+00 -8.35592858e-03 -1.83830127e-01 1.94058254e-01
-4.48039502e-01 -3.95407021e-01 -5.43070078e-01 -5.46262860e-01
-2.58173436e-01 -1.83454484e-01 8.56140137e-01 5.80361664e-01
1.47757566e+00 -6.15436792e-01 -4.19134706e-01 9.11070466e-01
1.51167047e+00 5.01289725e-01 4.84286785e-01 5.19199848e-01
4.68882501e-01 -3.32243592e-02 2.22050920e-01 4.14561480e-01
2.79121339e-01 9.76673365e-01 8.46682429e-01 -1.01427794e-01
3.03018820e-02 1.87610164e-01 2.67601252e-01 6.53011143e-01
2.11130559e-01 -5.94001710e-01 -1.16696262e+00 3.21300149e-01
-1.65380144e+00 -4.58941877e-01 4.64733392e-01 2.36691236e+00
6.36343122e-01 6.43348634e-01 6.99864104e-02 -1.67283460e-01
5.76390207e-01 2.14408040e-01 -8.95807087e-01 -3.53254288e-01
-2.92418282e-02 5.94990253e-01 8.67601752e-01 4.44962084e-01
-6.87026441e-01 1.07545137e+00 6.29491472e+00 1.16590953e+00
-1.27400899e+00 -5.89960329e-02 6.93935275e-01 -7.06733108e-01
-3.49209249e-01 8.78211483e-02 -1.11660242e+00 5.30250430e-01
1.12980127e+00 -2.26880327e-01 9.30573821e-01 8.37095380e-01
3.78330536e-02 2.84626931e-01 -1.17173588e+00 1.19890380e+00
-2.03408957e-01 -1.69735599e+00 -8.95059779e-02 4.01825726e-01
6.58501506e-01 3.85596305e-01 2.20738694e-01 4.08705115e-01
2.85261065e-01 -1.16483796e+00 7.99672306e-01 -2.30917215e-01
7.18336403e-01 -8.72552097e-01 4.72529948e-01 3.45165670e-01
-1.52574944e+00 -4.10925746e-01 -2.00074524e-01 3.23891938e-01
1.13726787e-01 5.94605841e-02 -8.64479423e-01 2.67652143e-02
8.64444911e-01 -1.10883124e-01 -5.32404542e-01 1.26526892e+00
2.73330122e-01 6.32534862e-01 -6.03107512e-01 -3.04768562e-01
4.15620863e-01 1.16178319e-01 4.52341199e-01 8.81440341e-01
4.38435018e-01 5.14900647e-02 2.85972565e-01 9.42785740e-01
-4.23536301e-01 -8.45229849e-02 1.07881106e-01 -7.73931071e-02
1.06724429e+00 8.19942534e-01 -8.31893086e-01 -5.11267819e-02
-7.52067566e-02 4.73724693e-01 2.15761825e-01 3.02768767e-01
-1.12107027e+00 -2.96826333e-01 9.20413077e-01 4.88719121e-02
3.96760315e-01 -3.47948164e-01 -4.89557147e-01 -7.69722819e-01
1.19124644e-01 -1.11755574e+00 3.92206043e-01 -2.38781586e-01
-5.57536304e-01 1.08159018e+00 -2.48243473e-02 -1.06136060e+00
-9.34110358e-02 -3.86471510e-01 -5.08214772e-01 8.04466903e-01
-1.42809570e+00 -6.66072309e-01 -1.61126420e-01 3.91094536e-01
6.78839743e-01 -4.95018065e-01 6.52040958e-01 4.97482479e-01
-8.70923042e-01 1.24711049e+00 -1.43572837e-01 -3.91384989e-01
2.66415536e-01 -8.17886174e-01 7.59445369e-01 8.90787244e-01
1.48031026e-01 7.57470727e-01 6.36704564e-01 -2.36285150e-01
-1.47548580e+00 -8.49000037e-01 6.80070937e-01 4.52007093e-02
4.56918120e-01 -4.43342000e-01 -6.54908180e-01 3.80876005e-01
-9.59450379e-02 -1.10329457e-01 5.81989765e-01 1.80276543e-01
-4.39957172e-01 -2.82960534e-01 -7.99098253e-01 9.59206760e-01
1.28559458e+00 -1.52341500e-01 3.92459005e-01 1.05728582e-01
1.03620994e+00 -3.99251997e-01 -6.47892475e-01 6.33833349e-01
4.48175550e-01 -6.59726441e-01 1.10165131e+00 -4.29509878e-01
-2.48445153e-01 -4.40820545e-01 -4.13224101e-01 -9.41206455e-01
-3.12506467e-01 -9.35863137e-01 -4.17981297e-01 6.96594894e-01
1.23160708e+00 -5.21967947e-01 1.15158188e+00 6.67491496e-01
-3.00933808e-01 -1.10880280e+00 -1.23545253e+00 -9.99163091e-01
-6.93185851e-02 -4.23162907e-01 1.02347147e+00 3.98759454e-01
-6.06355309e-01 3.56878668e-01 -1.90539747e-01 3.14675063e-01
5.48425496e-01 7.62814581e-02 5.99359810e-01 -1.31601334e+00
-7.63183117e-01 -1.04264343e+00 -8.91799107e-02 -1.36924732e+00
1.76197112e-01 -7.55013585e-01 -2.77578421e-02 -1.20793283e+00
-1.15811322e-02 -1.11225510e+00 -5.64630687e-01 5.61961114e-01
8.43082592e-02 -1.00674048e-01 4.88058031e-01 4.28855211e-01
-8.87826800e-01 3.73208553e-01 7.39129424e-01 -2.61256903e-01
-4.40397292e-01 1.69015646e-01 -8.33479345e-01 6.19450450e-01
8.29857707e-01 -5.04271626e-01 -5.84367335e-01 -8.87589931e-01
4.34133708e-01 1.90510571e-01 -4.31640409e-02 -1.15242410e+00
4.21425402e-01 -4.36875165e-01 -1.95415810e-01 -5.76580703e-01
5.62571168e-01 -6.77758455e-01 2.31703594e-01 5.91637313e-01
-4.51476008e-01 6.26511872e-01 3.57343704e-01 3.42934191e-01
7.69992918e-02 -1.99385822e-01 7.33802915e-01 6.05579875e-02
-7.97405005e-01 5.22092402e-01 -2.58182213e-02 2.40269862e-02
6.69021845e-01 -2.88792759e-01 -4.55905318e-01 -7.43243843e-02
-4.07327205e-01 7.38978922e-01 3.26786876e-01 2.58887380e-01
5.90148807e-01 -1.09544611e+00 -3.89046907e-01 8.46214443e-02
7.51300976e-02 4.00749147e-02 1.83571503e-01 6.98468566e-01
-6.05518520e-01 6.95635378e-01 7.50612020e-02 -5.37270367e-01
-1.32230234e+00 3.38301837e-01 4.56886262e-01 -6.74784601e-01
-1.16856784e-01 1.32018340e+00 1.60097688e-01 -1.73753947e-01
6.54584944e-01 -3.08518440e-01 1.23228997e-01 -3.89208615e-01
4.12834644e-01 5.07697821e-01 1.81373179e-01 -2.39462659e-01
-6.08085036e-01 3.49119127e-01 -3.69411260e-02 -2.76517361e-01
1.13184953e+00 -5.46068884e-02 -9.76844281e-02 -1.43936068e-01
1.15643239e+00 -2.18589365e-01 -1.20998478e+00 -2.97323346e-01
3.02696347e-01 -1.53444275e-01 3.88628900e-01 -8.94618332e-01
-1.24767852e+00 5.86659610e-01 8.45062733e-01 -6.64530993e-02
1.41878915e+00 -1.89826973e-02 8.73657703e-01 6.69182122e-01
8.08045268e-01 -9.95768368e-01 1.85406506e-01 6.96503758e-01
4.81084943e-01 -8.76428545e-01 -1.60955682e-01 -2.85919756e-01
-2.38475859e-01 6.46954536e-01 1.06847978e+00 1.86883956e-01
5.78799486e-01 2.73062229e-01 -1.56314656e-01 -1.93492740e-01
-1.11747909e+00 -1.03882238e-01 2.38442764e-01 5.75838201e-02
4.77230288e-02 1.23821557e-01 -2.37989753e-01 6.50976360e-01
-3.00831378e-01 -4.17460829e-01 9.94135998e-03 8.13061357e-01
-6.54513180e-01 -1.36572289e+00 -1.47776589e-01 4.84317183e-01
-4.25371170e-01 -4.56878155e-01 -1.72526583e-01 4.80790645e-01
3.02543193e-02 8.77142668e-01 1.18511654e-02 -6.25175893e-01
2.64221519e-01 -1.01302683e-01 2.79548556e-01 -4.90656227e-01
-7.46562064e-01 3.26477945e-01 1.84049055e-01 -6.08812094e-01
3.24036688e-01 -1.98026240e-01 -1.35711443e+00 -3.12330544e-01
-5.04558682e-01 2.66762346e-01 8.49885821e-01 7.21255541e-01
6.57600701e-01 5.40666521e-01 7.18211353e-01 -6.11660063e-01
-7.45556772e-01 -4.78623599e-01 -2.17306569e-01 -4.85993475e-01
1.61513284e-01 -6.91267312e-01 -2.39863709e-01 -5.03706694e-01] | [8.516645431518555, 3.1861915588378906] |
a03bfbe7-5ff4-47f8-97ed-f1819647dbd4 | video-reconstruction-by-spatio-temporal | 2010.10052 | null | https://arxiv.org/abs/2010.10052v2 | https://arxiv.org/pdf/2010.10052v2.pdf | Video Reconstruction by Spatio-Temporal Fusion of Blurred-Coded Image Pair | Learning-based methods have enabled the recovery of a video sequence from a single motion-blurred image or a single coded exposure image. Recovering video from a single motion-blurred image is a very ill-posed problem and the recovered video usually has many artifacts. In addition to this, the direction of motion is lost and it results in motion ambiguity. However, it has the advantage of fully preserving the information in the static parts of the scene. The traditional coded exposure framework is better-posed but it only samples a fraction of the space-time volume, which is at best 50% of the space-time volume. Here, we propose to use the complementary information present in the fully-exposed (blurred) image along with the coded exposure image to recover a high fidelity video without any motion ambiguity. Our framework consists of a shared encoder followed by an attention module to selectively combine the spatial information from the fully-exposed image with the temporal information from the coded image, which is then super-resolved to recover a non-ambiguous high-quality video. The input to our algorithm is a fully-exposed and coded image pair. Such an acquisition system already exists in the form of a Coded-two-bucket (C2B) camera. We demonstrate that our proposed deep learning approach using blurred-coded image pair produces much better results than those from just a blurred image or just a coded image. | ['S Anupama', 'Kaushik Mitra', 'Abhishek Pal', 'Prasan Shedligeri'] | 2020-10-20 | null | null | null | null | ['video-reconstruction'] | ['computer-vision'] | [ 6.18685305e-01 -3.27609718e-01 1.71569467e-01 -1.13551132e-01
-7.01140285e-01 -3.48485887e-01 1.96573645e-01 -7.55547643e-01
-4.67441708e-01 8.47329259e-01 2.39804834e-01 1.41953245e-01
-2.08072752e-01 -4.37800586e-01 -7.91119277e-01 -9.14560616e-01
1.05521463e-01 -1.85630381e-01 2.27407485e-01 1.11531511e-01
1.94778249e-01 1.06814466e-01 -1.41560018e+00 3.57962668e-01
8.81820619e-01 1.01118994e+00 1.08937693e+00 7.35853970e-01
1.22155830e-01 1.08010697e+00 -2.92335749e-01 4.30509858e-02
3.59399736e-01 -6.01799190e-01 -7.87983954e-01 3.45304430e-01
5.76332092e-01 -1.07829511e+00 -9.40487623e-01 1.27651370e+00
1.11612774e-01 2.40819588e-01 1.86339661e-01 -8.28267753e-01
-8.09347689e-01 1.62759975e-01 -7.44877994e-01 4.42652345e-01
4.73486543e-01 2.67201513e-01 1.80793762e-01 -9.36774015e-01
6.75850749e-01 9.90612686e-01 2.58771688e-01 3.40657860e-01
-9.88485038e-01 -3.32411319e-01 -2.43084878e-01 3.94225448e-01
-1.21412015e+00 -6.73474133e-01 8.77078235e-01 -4.43398327e-01
3.78840953e-01 1.64313257e-01 6.04182541e-01 6.75032675e-01
4.53766406e-01 4.42454249e-01 1.10573423e+00 -2.78965324e-01
8.29360560e-02 -8.58120024e-02 -8.69407551e-04 4.76480484e-01
2.67357796e-01 3.83087814e-01 -4.54660207e-01 1.78856477e-01
1.00185299e+00 5.58552206e-01 -1.30807340e+00 -2.39579991e-01
-1.16481018e+00 1.40149966e-01 4.62576568e-01 4.06367421e-01
-4.31314021e-01 1.52513444e-01 -1.03452645e-01 2.68938154e-01
3.67249757e-01 2.73841798e-01 -1.71816856e-01 -1.54034600e-01
-1.35103393e+00 1.22013800e-01 4.33015466e-01 7.70031750e-01
1.11064398e+00 2.20464811e-01 2.25113127e-02 6.23789728e-01
2.20360368e-01 5.79125762e-01 5.51055372e-01 -1.24577188e+00
5.28001547e-01 -5.27617410e-02 6.35963023e-01 -8.98599088e-01
2.07391784e-01 -2.13932186e-01 -1.03615737e+00 4.59566236e-01
4.48855877e-01 -1.75554797e-01 -8.48606765e-01 1.80759144e+00
-3.58058177e-02 5.38972616e-01 9.98849198e-02 1.58720517e+00
4.25513834e-01 8.72724891e-01 -4.32867855e-01 -7.21295834e-01
1.18567586e+00 -1.02192795e+00 -1.12242758e+00 -4.89898682e-01
-3.49033982e-01 -8.52709532e-01 7.15277731e-01 3.37433517e-01
-1.47783840e+00 -8.42746258e-01 -1.29174936e+00 -3.27088058e-01
1.70702502e-01 -2.37534165e-01 -1.47915706e-01 1.74121246e-01
-1.06682110e+00 6.08329177e-01 -5.74452758e-01 1.51705444e-01
2.32920647e-01 3.86401899e-02 -6.27122283e-01 -7.68940508e-01
-1.23775828e+00 9.52664852e-01 1.93898574e-01 1.03237063e-01
-9.45115030e-01 -7.13112950e-01 -9.36452985e-01 2.54433930e-01
4.43748981e-01 -7.21654236e-01 1.00431085e+00 -1.24275565e+00
-1.24648714e+00 4.45340425e-01 -3.14257979e-01 -1.26812592e-01
4.44966584e-01 -3.56137186e-01 -3.09506238e-01 6.24580026e-01
2.23303717e-02 2.80175865e-01 1.54766572e+00 -1.29997253e+00
-4.02019203e-01 -2.16597065e-01 2.78172102e-02 2.80674994e-01
1.98613390e-01 -1.62968040e-01 -4.40413117e-01 -6.33856118e-01
1.54885367e-01 -6.30340993e-01 7.29487538e-02 1.68635666e-01
9.89740714e-02 7.98553050e-01 1.21824753e+00 -1.15035701e+00
1.13168347e+00 -2.35875225e+00 3.67586225e-01 -7.07912922e-01
3.43423426e-01 3.49193633e-01 -1.40912324e-01 3.66070479e-01
-2.84936756e-01 -3.31471980e-01 -7.52295792e-01 -4.46286112e-01
-6.10749662e-01 5.46637084e-03 -4.82399344e-01 7.34635174e-01
4.46335524e-02 7.62338400e-01 -1.18865776e+00 -2.65901208e-01
6.38734758e-01 6.87424064e-01 -3.86937231e-01 6.04097724e-01
8.77035931e-02 7.02068865e-01 -7.50876323e-04 3.54788631e-01
1.31826365e+00 -2.43904486e-01 1.93721367e-04 -3.48223239e-01
-3.88392627e-01 -1.82227731e-01 -9.83291149e-01 1.99943316e+00
-3.85429472e-01 9.82927620e-01 5.27177989e-01 -6.84064031e-01
4.14509654e-01 3.82283360e-01 4.39301550e-01 -6.18106484e-01
-1.25568345e-01 1.92991003e-01 -2.70589560e-01 -1.07990324e+00
5.99770427e-01 -3.56100559e-01 2.96730608e-01 3.41858745e-01
-6.30891100e-02 -1.74908981e-01 -1.08526893e-01 1.80571124e-01
1.07834435e+00 1.95759282e-01 2.38057345e-01 3.96403447e-02
5.91873944e-01 -4.93071109e-01 5.40959835e-01 6.04261458e-01
-3.22648913e-01 1.33757949e+00 -8.57229382e-02 -3.49823862e-01
-1.24137449e+00 -9.54996943e-01 -2.44161561e-02 3.28661829e-01
7.23700762e-01 -2.19846498e-02 -7.66154587e-01 -1.93389282e-01
-2.99133897e-01 4.29722250e-01 -2.75758475e-01 -2.13967219e-01
-7.55265355e-01 -3.41687053e-01 1.38135105e-02 9.54302177e-02
9.35306132e-01 -8.59085858e-01 -8.22293639e-01 6.83540180e-02
-7.27200985e-01 -1.27269268e+00 -1.12500322e+00 -1.05044596e-01
-6.97192609e-01 -1.19804645e+00 -1.07413054e+00 -8.09880078e-01
5.38008511e-01 9.87543046e-01 6.42833173e-01 -7.75219128e-02
-1.30942911e-01 -6.70821231e-04 -1.29651338e-01 4.23303485e-01
-3.06652665e-01 -8.16468120e-01 2.77671423e-02 5.64788342e-01
-2.71767586e-01 -5.89401722e-01 -9.25751150e-01 2.55175810e-02
-1.27656686e+00 4.35898751e-01 5.69942117e-01 9.93502200e-01
6.77698925e-02 2.92921931e-01 2.55001247e-01 -3.67657304e-01
1.55845568e-01 -4.47297841e-01 -4.98017550e-01 -7.31894700e-03
-3.92561048e-01 8.06125030e-02 7.17933416e-01 -5.27813315e-01
-1.51123762e+00 5.39838485e-02 2.30333835e-01 -1.00293076e+00
9.81033072e-02 3.19572568e-01 -2.29054347e-01 -1.85969640e-02
2.70332694e-01 5.81009626e-01 2.05183983e-01 -6.78262055e-01
2.41818339e-01 8.95468056e-01 9.97402251e-01 -4.00810093e-02
7.52251804e-01 6.80501044e-01 -1.58494502e-01 -8.45214248e-01
-3.83832991e-01 -4.43714559e-01 -5.11154771e-01 -2.84945905e-01
8.77395391e-01 -1.07463527e+00 -4.61433709e-01 8.85777652e-01
-1.36545038e+00 -2.47504249e-01 -1.61986813e-01 6.88823640e-01
-5.08719444e-01 9.32339847e-01 -9.97780263e-01 -6.33496583e-01
-5.64223155e-02 -1.35904491e+00 9.86847699e-01 3.68595690e-01
3.41868818e-01 -7.10769296e-01 -2.27959156e-02 4.87606049e-01
6.53069317e-01 -1.10669784e-01 5.88131249e-01 4.56322908e-01
-1.17279482e+00 1.05716012e-01 -4.90244716e-01 5.29905379e-01
4.81146783e-01 -4.95977998e-01 -1.13935781e+00 -5.20561039e-01
8.87759447e-01 -1.69292428e-02 1.06832421e+00 6.91576600e-01
9.59622860e-01 -3.98783773e-01 -5.36233112e-02 9.60822701e-01
1.75264382e+00 2.90344238e-01 1.01780581e+00 -1.52450195e-02
7.55260587e-01 3.85362864e-01 5.25925994e-01 1.95345148e-01
2.02744622e-02 8.31026912e-01 4.60995525e-01 4.76198364e-03
-6.19888365e-01 -1.10844217e-01 4.76522833e-01 6.30560577e-01
9.30039212e-03 -3.79371315e-01 -3.80447656e-01 6.02493286e-01
-1.68274188e+00 -1.41789627e+00 -7.46159032e-02 2.38213015e+00
9.31994617e-01 -3.75494212e-01 -5.82873583e-01 9.17668492e-02
8.94894540e-01 5.66410661e-01 -5.13196468e-01 6.22015968e-02
-2.48345971e-01 -1.00967273e-01 3.88386041e-01 9.98632193e-01
-7.78383076e-01 3.95157516e-01 6.11223459e+00 6.80915296e-01
-1.39510179e+00 2.30624601e-01 4.07369673e-01 -1.98505461e-01
-1.69836909e-01 1.70393825e-01 -7.21488148e-02 1.15729916e+00
6.86120152e-01 -1.28290519e-01 8.73530269e-01 4.09379065e-01
4.55062270e-01 -5.93888402e-01 -1.00006056e+00 1.52828240e+00
2.82180041e-01 -1.23642409e+00 -1.53345168e-01 1.32501766e-01
7.24176466e-01 -3.16310793e-01 1.45055756e-01 -1.65889084e-01
-3.95678073e-01 -9.32330549e-01 8.40849996e-01 9.00312781e-01
1.31850350e+00 -3.96463752e-01 6.72702491e-01 5.89760005e-01
-9.72064614e-01 -1.71849400e-01 -2.86213070e-01 -2.90177852e-01
5.50132751e-01 7.52086818e-01 -1.40814349e-01 6.92841828e-01
6.77110493e-01 8.57455134e-01 -1.55591920e-01 1.06393337e+00
1.37287587e-01 1.61978826e-01 1.80772081e-01 8.76540780e-01
-1.71025970e-03 -5.22702634e-01 7.88590610e-01 9.16031361e-01
5.55139482e-01 5.51660597e-01 -1.96158648e-01 9.31609571e-01
4.52806056e-03 -7.72775531e-01 -6.77234054e-01 4.23493266e-01
2.03493193e-01 1.02513885e+00 -1.00126527e-01 -5.30515969e-01
-6.35872304e-01 1.48513150e+00 1.32269273e-03 7.18110085e-01
-8.04231107e-01 -3.92939955e-01 5.97461581e-01 1.65102527e-01
4.57482696e-01 -1.08717628e-01 1.56328127e-01 -1.62172270e+00
7.07019866e-02 -6.55318499e-01 -1.59263372e-01 -1.52301669e+00
-1.01370525e+00 6.46230817e-01 -3.83380502e-02 -1.52845490e+00
-1.69091210e-01 -2.59366482e-01 -3.11168730e-01 1.38452876e+00
-1.64681435e+00 -6.61393166e-01 -8.11081707e-01 5.55504858e-01
9.07504976e-01 1.07305169e-01 2.39741027e-01 5.69175303e-01
-3.06245863e-01 -1.24427728e-01 3.27339053e-01 -8.93671289e-02
1.00289202e+00 -1.03750503e+00 -6.93868026e-02 1.20949006e+00
-4.60346103e-01 4.63813335e-01 7.50903189e-01 -6.81996167e-01
-1.52836609e+00 -9.62429762e-01 5.90876818e-01 -1.99678093e-01
9.46281999e-02 8.35007802e-02 -1.24723816e+00 5.42733908e-01
5.88173449e-01 2.85104007e-01 -5.47695048e-02 -9.58914161e-01
1.89801138e-02 -2.69564956e-01 -1.16518366e+00 2.25056008e-01
7.19447255e-01 -8.45160961e-01 -8.07155609e-01 -2.50983268e-01
9.67675269e-01 -5.71560025e-01 -3.17235470e-01 2.37750351e-01
4.71696138e-01 -1.25552893e+00 1.06783175e+00 2.23469511e-01
7.53244102e-01 -7.14890182e-01 -2.17917770e-01 -1.28409576e+00
-4.93799567e-01 -7.96362936e-01 -3.71770531e-01 7.73684919e-01
-3.72752696e-01 -4.32247430e-01 4.32695597e-01 6.03647828e-01
-3.17979664e-01 -3.48903179e-01 -7.67742693e-01 -4.57426369e-01
-4.50468093e-01 -5.49172349e-02 2.64206588e-01 9.07749951e-01
-5.81467897e-02 2.13858217e-01 -1.01642084e+00 2.99116433e-01
1.03945994e+00 2.80159712e-01 2.39174217e-01 -6.12121940e-01
-4.37728465e-01 1.62936881e-01 -1.19396813e-01 -1.60827267e+00
-1.82869941e-01 -3.87776911e-01 2.99765468e-01 -1.26191175e+00
5.55146456e-01 -1.88696175e-03 -1.05515197e-01 -2.33556360e-01
-3.19210708e-01 2.25348577e-01 2.48261243e-01 6.04977369e-01
-2.27633566e-01 4.28687006e-01 1.57996190e+00 -9.60899219e-02
-9.00672451e-02 -2.80681610e-01 -4.62930411e-01 4.82715160e-01
2.36689121e-01 -2.92423874e-01 -5.07382154e-01 -7.76935697e-01
-2.98643380e-01 9.04366136e-01 6.32636786e-01 -9.27854717e-01
4.83897626e-01 -2.71535993e-01 6.23729587e-01 -5.04468501e-01
5.93182921e-01 -9.12295043e-01 5.42548895e-01 3.65069479e-01
-1.29882589e-01 -1.86113328e-01 -1.52736038e-01 8.51752043e-01
-4.89013791e-01 -4.04064357e-01 1.05906260e+00 -4.02389735e-01
-7.88324416e-01 2.22210392e-01 -3.06214064e-01 -1.49327338e-01
8.43699217e-01 -3.72283578e-01 -4.01108652e-01 -5.00161111e-01
-5.35291076e-01 -2.04072952e-01 9.59359586e-01 1.57594949e-01
9.52906966e-01 -1.05467451e+00 -5.04642606e-01 4.92043376e-01
-3.74323756e-01 8.46194252e-02 8.44996214e-01 7.46044934e-01
-8.27820182e-01 4.25991386e-01 -3.76548648e-01 -5.34455717e-01
-1.08337486e+00 1.07432711e+00 4.60438073e-01 1.73634276e-01
-8.04612219e-01 4.64381069e-01 5.93313873e-01 4.73302424e-01
-2.22434998e-02 -3.42725039e-01 1.56589180e-01 -2.79614300e-01
1.04012275e+00 4.03275490e-01 -2.41568953e-01 -7.77732193e-01
-1.93897948e-01 7.76890397e-01 2.60231495e-01 -4.02850509e-01
1.12410092e+00 -8.11229825e-01 -1.72931075e-01 2.98981994e-01
1.42755735e+00 1.40478060e-01 -1.90428901e+00 -2.62681186e-01
-5.82293510e-01 -1.10254204e+00 2.88769573e-01 -5.39758503e-01
-1.11525774e+00 9.39165413e-01 6.06962323e-01 9.10952222e-03
1.48928320e+00 -2.63698637e-01 1.01445687e+00 -3.62863928e-01
3.99782836e-01 -6.23528540e-01 3.18861723e-01 2.14577958e-01
9.38950896e-01 -1.39116299e+00 6.82445168e-02 -2.25859866e-01
-4.75596339e-01 1.13766062e+00 3.70815486e-01 -1.91106483e-01
4.42338377e-01 1.55409828e-01 -1.99556872e-01 1.65749103e-01
-7.26733148e-01 -2.27594655e-02 1.55328050e-01 5.75669944e-01
4.21855450e-02 -3.84058893e-01 2.78799217e-02 2.82228529e-01
3.72433633e-01 3.22517574e-01 9.28108215e-01 7.22282767e-01
-4.87871677e-01 -5.88060439e-01 -7.15914965e-01 -9.38196406e-02
-3.44109714e-01 -8.71932283e-02 2.51251906e-01 4.18281049e-01
3.23881865e-01 1.04827726e+00 1.32360294e-01 -1.52262479e-01
-1.88466292e-02 -4.38700505e-02 6.82874203e-01 -3.78722638e-01
1.21451996e-01 1.97430283e-01 -4.65409338e-01 -7.05451429e-01
-5.01799703e-01 -4.68627930e-01 -9.90512967e-01 -4.25280333e-01
-1.05643906e-01 -5.56667857e-02 3.73344332e-01 7.89430320e-01
2.20624477e-01 4.63185787e-01 6.96655512e-01 -1.23341095e+00
-1.29901528e-01 -9.17077959e-01 -7.39896178e-01 5.39203823e-01
1.37316775e+00 -2.97264129e-01 -7.45561481e-01 4.83046293e-01] | [11.302306175231934, -2.4624574184417725] |
fceab5a2-7627-4626-a049-37e719cdddfd | an-efficient-algorithm-for-mining-frequent | 1604.01166 | null | http://arxiv.org/abs/1604.01166v1 | http://arxiv.org/pdf/1604.01166v1.pdf | An Efficient Algorithm for Mining Frequent Sequence with Constraint Programming | The main advantage of Constraint Programming (CP) approaches for sequential
pattern mining (SPM) is their modularity, which includes the ability to add new
constraints (regular expressions, length restrictions, etc). The current best
CP approach for SPM uses a global constraint (module) that computes the
projected database and enforces the minimum frequency; it does this with a
filtering algorithm similar to the PrefixSpan method. However, the resulting
system is not as scalable as some of the most advanced mining systems like
Zaki's cSPADE. We show how, using techniques from both data mining and CP, one
can use a generic constraint solver and yet outperform existing specialized
systems. This is mainly due to two improvements in the module that computes the
projected frequencies: first, computing the projected database can be sped up
by pre-computing the positions at which an symbol can become unsupported by a
sequence, thereby avoiding to scan the full sequence each time; and second by
taking inspiration from the trailing used in CP solvers to devise a
backtracking-aware data structure that allows fast incremental storing and
restoring of the projected database. Detailed experiments show how this
approach outperforms existing CP as well as specialized systems for SPM, and
that the gain in efficiency translates directly into increased efficiency for
other settings such as mining with regular expressions. | ['John O. R. Aoga', 'Tias Guns', 'Pierre Schaus'] | 2016-04-05 | null | null | null | null | ['sequential-pattern-mining'] | ['natural-language-processing'] | [ 4.44427013e-01 1.16084307e-01 -5.18389583e-01 1.14464760e-03
-1.67091250e-01 -5.43534160e-01 1.13031462e-01 4.82135385e-01
-4.67554212e-01 8.05432022e-01 -1.29901990e-01 -5.54804444e-01
-4.77824539e-01 -1.12970781e+00 -4.33901906e-01 -4.85018432e-01
-4.89146292e-01 9.15639162e-01 8.38591397e-01 -2.01798752e-01
4.35799211e-01 6.69170856e-01 -1.88480532e+00 4.63384330e-01
5.73374391e-01 9.79067981e-01 1.74167991e-01 2.81711042e-01
-4.89981145e-01 2.53654033e-01 -3.87509525e-01 -1.27443582e-01
4.63811815e-01 -1.97229296e-01 -1.05449045e+00 2.17481047e-01
-7.58297741e-02 1.67253748e-01 3.07818860e-01 6.94105327e-01
-1.13735795e-01 -5.13936952e-02 -7.24458694e-02 -1.18558812e+00
1.44039705e-01 7.58107066e-01 -7.18404949e-01 1.57311007e-01
8.57267916e-01 -2.16890767e-01 1.18835354e+00 -7.36620486e-01
9.80727792e-01 9.37586486e-01 6.32485390e-01 2.17857361e-01
-1.40358388e+00 -4.14023161e-01 4.41620976e-01 3.54059011e-01
-1.46354890e+00 -1.62520781e-01 1.94210142e-01 -2.23457795e-02
1.60372329e+00 7.26966441e-01 8.46844256e-01 2.41016313e-01
-1.32433474e-01 7.02198625e-01 9.30555880e-01 -8.53093624e-01
4.02562827e-01 2.22975001e-01 3.33578497e-01 5.93823731e-01
4.19886172e-01 1.38907507e-01 -7.75198638e-01 -5.37354887e-01
1.42348975e-01 -1.68051630e-01 -1.76524848e-01 -5.35905123e-01
-9.26407814e-01 7.69893944e-01 -3.65411848e-01 4.64335591e-01
-5.81673048e-02 -2.88719803e-01 5.91975689e-01 6.14627779e-01
-1.72993138e-01 5.71232378e-01 -8.67860079e-01 -2.37436712e-01
-1.19813299e+00 6.50916874e-01 1.47566497e+00 1.13955927e+00
8.76005352e-01 -3.69297862e-01 8.46313611e-02 4.05582517e-01
-4.72155921e-02 9.89345089e-02 5.74568398e-02 -7.52563596e-01
4.26556021e-01 9.05422211e-01 6.64945319e-02 -8.19896936e-01
-5.55472672e-01 -2.29642674e-01 -2.82394648e-01 1.47997111e-01
4.17087525e-01 -1.48411430e-02 -5.45298398e-01 1.62670851e+00
3.34145695e-01 -1.27312928e-01 -1.73692361e-01 4.43524897e-01
-1.67113662e-01 6.51771843e-01 -1.21073619e-01 -9.35596824e-01
1.44480944e+00 -5.83092988e-01 -3.41370553e-01 -2.19976678e-01
9.36005235e-01 -5.12705028e-01 4.53636348e-01 1.05570948e+00
-1.29336572e+00 -1.68415591e-01 -1.32554054e+00 2.57869244e-01
-4.58342105e-01 -4.68857616e-01 9.82650876e-01 7.07028329e-01
-8.10824096e-01 5.91528833e-01 -8.76718342e-01 -2.88774759e-01
1.61273956e-01 7.03527451e-01 -3.45067680e-01 -6.70917481e-02
-9.93470430e-01 1.09585452e+00 8.68251085e-01 -2.85252661e-01
-2.25612279e-02 -9.77368176e-01 -7.22676396e-01 2.11806595e-01
1.13231099e+00 -3.53253961e-01 1.07643068e+00 -8.37922096e-01
-1.14482105e+00 7.47492611e-01 -4.54715639e-01 -5.66624284e-01
2.92918861e-01 1.59087732e-01 -3.71679395e-01 -1.06740162e-01
-9.73297507e-02 6.50528818e-02 2.43513495e-01 -8.26992631e-01
-1.04934871e+00 -3.84344131e-01 2.62083253e-03 1.91310362e-03
-1.38045520e-01 3.51768374e-01 -7.92790771e-01 -1.77754313e-01
2.32011005e-01 -9.54141736e-01 -6.53140247e-01 -2.16596499e-01
-3.61093670e-01 -2.16400996e-01 6.23739481e-01 -3.83602500e-01
1.87084913e+00 -1.98926461e+00 3.29888642e-01 1.12553465e+00
-2.26061583e-01 3.44411403e-01 2.09663704e-01 9.65931535e-01
-8.66430476e-02 1.27760708e-01 -5.45322299e-01 4.42817248e-02
-6.17208034e-02 7.53382564e-01 -2.29065150e-01 2.38720089e-01
1.31426856e-01 4.19162989e-01 -7.31204033e-01 -5.00825167e-01
-9.10097063e-02 -2.92943984e-01 -1.04585648e+00 -3.95041943e-01
-6.24602914e-01 -1.89825624e-01 -2.14808732e-01 5.11363089e-01
8.09127331e-01 7.57165179e-02 1.05419016e+00 4.98264462e-01
-6.94239616e-01 5.13257921e-01 -2.03385782e+00 1.49578071e+00
-2.41584435e-01 6.35505393e-02 2.35384271e-01 -1.22427559e+00
7.91498482e-01 6.20694943e-02 6.81071877e-01 -6.89241529e-01
-2.97477663e-01 3.46383601e-01 -9.98496935e-02 -3.13088626e-01
3.98874938e-01 6.42624348e-02 -1.46997735e-01 6.30837977e-01
-3.90448451e-01 1.19889438e-01 8.35084260e-01 1.69464111e-01
1.31765497e+00 -4.40818332e-02 5.32785177e-01 -5.60233653e-01
9.82664406e-01 5.11131406e-01 1.07957172e+00 7.04402566e-01
6.02028608e-01 -3.99326570e-02 9.65456367e-01 -5.92616498e-01
-7.92786598e-01 -6.40025318e-01 -3.78514260e-01 9.57564473e-01
-1.23842254e-01 -1.11013424e+00 -3.60713333e-01 -4.20539320e-01
3.34775090e-01 4.50597584e-01 -3.30194890e-01 1.84156418e-01
-9.35538292e-01 -7.78386950e-01 1.25679418e-01 4.60921258e-01
8.96320492e-02 -1.06431293e+00 -9.81421530e-01 6.09477162e-01
1.65096804e-01 -8.51879358e-01 -1.20001249e-01 8.04455519e-01
-9.04872835e-01 -1.19180441e+00 3.46305996e-01 -7.38753140e-01
5.73165178e-01 1.29544281e-03 9.90581393e-01 5.71251333e-01
-4.96438414e-01 -2.27584109e-01 -5.12021422e-01 -5.38021266e-01
-2.23599747e-01 1.75142944e-01 -1.13832414e-01 -4.08132792e-01
7.53651321e-01 -7.71503210e-01 1.86233126e-04 4.20453072e-01
-9.51233983e-01 -1.69568703e-01 4.35386091e-01 8.24932933e-01
8.42132449e-01 4.90972459e-01 1.58138841e-01 -1.45498681e+00
4.31574494e-01 -4.79897887e-01 -1.01572192e+00 2.48858765e-01
-1.18507767e+00 2.01394454e-01 5.00574470e-01 -1.84842706e-01
-7.23615885e-01 3.78867179e-01 -1.25086457e-01 7.52068684e-02
-1.45862103e-02 9.28708494e-01 -1.66258872e-01 8.26011226e-02
4.49798495e-01 1.74458638e-01 6.76336884e-02 -4.67641592e-01
1.04914702e-01 3.21421504e-01 4.51162547e-01 -8.59590352e-01
8.16891253e-01 4.08893019e-01 3.21787506e-01 -6.32404149e-01
-3.65963876e-01 -7.53643692e-01 -5.49188137e-01 4.16285634e-01
2.47760803e-01 -3.43436778e-01 -1.02281833e+00 -7.34971687e-02
-8.62032354e-01 -1.43461272e-01 -5.35185933e-01 -5.79883270e-02
-5.81719577e-01 7.06672311e-01 -4.07301277e-01 -8.45855951e-01
-8.18519741e-02 -7.02223480e-01 5.11769474e-01 -1.59305111e-01
-5.07164180e-01 -3.89193118e-01 1.47577986e-01 -1.05657130e-01
2.69919246e-01 3.25714558e-01 1.21444607e+00 -7.45439708e-01
-5.66842258e-01 -1.72381461e-01 4.12083678e-02 -4.71995063e-02
-3.42691332e-01 2.03701109e-02 -1.29992425e-01 -2.27256328e-01
-6.48154840e-02 1.55658811e-01 5.08960187e-01 3.13092172e-02
1.18751240e+00 -2.76525259e-01 -8.01220059e-01 5.13023078e-01
1.70930231e+00 3.03908676e-01 7.14590728e-01 6.93329096e-01
5.18289581e-02 7.02825665e-01 9.16729152e-01 7.35869944e-01
1.23553500e-01 1.02865744e+00 1.97675109e-01 1.50792897e-01
3.66842687e-01 1.00141011e-01 1.28367662e-01 3.71056646e-01
-2.01442122e-01 1.41798377e-01 -9.61644053e-01 5.99649668e-01
-2.13217592e+00 -1.13577163e+00 -3.82203996e-01 2.41595960e+00
1.02685797e+00 5.12239397e-01 5.14634013e-01 9.66922641e-01
1.93089470e-01 -3.49639803e-01 -1.92683876e-01 -1.29744852e+00
-2.90710106e-02 9.76663113e-01 8.59363616e-01 5.86123407e-01
-6.78939939e-01 7.59291887e-01 6.84430885e+00 6.01284683e-01
-5.62461555e-01 -8.24424624e-02 -1.35773242e-01 -2.82908201e-01
-2.05730900e-01 3.64326239e-01 -1.09228241e+00 2.64613748e-01
9.85968113e-01 -2.98033983e-01 6.03901088e-01 9.14850116e-01
9.54995956e-03 -4.56443191e-01 -1.35796332e+00 4.99202460e-01
-1.76238030e-01 -1.46156538e+00 -2.03911111e-01 4.25419509e-01
4.16500658e-01 -2.64285952e-01 -5.42709410e-01 1.03911027e-01
1.59418181e-01 -6.97752655e-01 7.32256055e-01 1.28046125e-01
4.88411933e-01 -1.04443777e+00 6.20604634e-01 5.56724131e-01
-1.17492867e+00 -4.30257469e-01 -3.72031420e-01 -3.98846328e-01
1.74506664e-01 7.38761008e-01 -5.89027345e-01 8.71618867e-01
6.81908131e-01 2.74424493e-01 -7.39309564e-02 1.05192113e+00
-8.32611695e-02 2.12959170e-01 -8.29764903e-01 4.35581757e-03
9.79076028e-02 -3.45567614e-01 6.06662989e-01 1.61586654e+00
9.69694853e-02 4.47334982e-02 4.77949411e-01 5.76397002e-01
5.28310478e-01 1.07840121e-01 -3.05250317e-01 2.78777182e-01
7.10549772e-01 8.84702861e-01 -5.76886475e-01 -1.27391472e-01
-7.76733994e-01 3.55115116e-01 2.47120112e-01 -2.20368765e-02
-4.56761062e-01 -4.31976140e-01 6.12688482e-01 3.76969188e-01
7.93475211e-01 -3.69662821e-01 -4.29077327e-01 -6.94063246e-01
4.25524801e-01 -1.24845362e+00 9.35735643e-01 4.33104113e-02
-8.37637603e-01 2.97759891e-01 2.44078323e-01 -7.55764544e-01
-3.09216350e-01 -6.09510839e-01 -4.40149724e-01 8.95329833e-01
-1.43292141e+00 -6.08365655e-01 2.54234821e-01 8.09112072e-01
1.54127643e-01 2.91368395e-01 1.03025758e+00 3.06302577e-01
-5.06071389e-01 6.90971434e-01 -3.98643166e-01 -3.85483474e-01
1.59329861e-01 -1.25205529e+00 2.06974506e-01 1.06692874e+00
8.30924213e-02 8.79697263e-01 7.43038595e-01 -5.00448346e-01
-1.85141480e+00 -5.13489068e-01 1.18242049e+00 -2.20963806e-01
6.57543182e-01 -3.94959956e-01 -9.00562525e-01 6.99159443e-01
-8.95803720e-02 -3.15260559e-01 7.11945236e-01 6.43540800e-01
-4.67086434e-01 -1.99918091e-01 -1.03018439e+00 3.50800782e-01
1.35983145e+00 -4.46490496e-02 -5.67534029e-01 2.22517088e-01
3.29738349e-01 -4.61554617e-01 -8.39033902e-01 3.50761950e-01
5.28244913e-01 -1.00786257e+00 7.87454844e-01 -6.12386107e-01
1.87460016e-02 -5.29497206e-01 -6.42749965e-02 -6.13634169e-01
-4.15416151e-01 -8.89498413e-01 -4.75708604e-01 8.45775723e-01
8.38561177e-01 -7.21942246e-01 9.84300077e-01 6.78382814e-01
-2.12238446e-01 -1.06516325e+00 -9.49184000e-01 -1.06960356e+00
-3.13952625e-01 -6.35594964e-01 8.15250158e-01 8.33988428e-01
8.68934333e-01 -7.83652738e-02 -2.39622578e-01 1.08413145e-01
2.50245005e-01 5.69812715e-01 6.26497626e-01 -1.45062232e+00
-8.74296308e-01 -4.42496568e-01 -2.64783174e-01 -8.49025607e-01
-2.33796462e-01 -1.13878655e+00 -2.11597368e-01 -1.13474512e+00
1.03960156e-01 -8.53958547e-01 -2.01510474e-01 8.36891353e-01
3.99558932e-01 -1.51728570e-01 1.54173791e-01 2.19896398e-02
-4.63120908e-01 -4.81897771e-01 7.05203474e-01 2.90005654e-01
-6.28947437e-01 -6.40341491e-02 -8.41030419e-01 4.89257932e-01
5.45819223e-01 -6.89899325e-01 -1.41860902e-01 -7.57360160e-02
9.23448384e-01 1.86351836e-01 -3.09018195e-01 -8.01882505e-01
5.00481069e-01 -5.19796848e-01 -2.72295717e-02 -7.01192021e-01
-9.10176113e-02 -9.89884377e-01 6.12610221e-01 8.95374179e-01
5.51965199e-02 1.20347142e-01 4.67246205e-01 4.23750699e-01
-2.10420921e-01 -5.82478702e-01 5.43496966e-01 -2.19761670e-01
-7.27787316e-01 1.23906478e-01 -6.93542361e-01 -3.53084058e-01
1.24538922e+00 -6.68873191e-01 -4.54319790e-02 2.71738648e-01
-8.63636851e-01 5.56374252e-01 5.50367892e-01 1.05953053e-01
1.67109773e-01 -7.65448987e-01 -3.29428017e-01 3.94289315e-01
1.30894050e-01 1.25754789e-01 -3.09903443e-01 1.01584959e+00
-5.22740304e-01 6.27076149e-01 -2.13028044e-01 -3.86531204e-01
-1.37136829e+00 1.01505053e+00 -7.08547980e-02 -8.02015960e-01
-7.58739471e-01 6.29958153e-01 -6.61477387e-01 -6.19199052e-02
2.28351995e-01 -3.41479927e-01 1.86624408e-01 5.14436476e-02
8.41183126e-01 3.01207453e-01 6.09768987e-01 1.15879633e-01
-8.06042254e-01 5.03745675e-01 -2.65993237e-01 3.54877077e-02
1.56215775e+00 1.90270215e-01 -5.65854609e-01 -1.14000581e-01
5.19953251e-01 4.29522902e-01 -5.08335233e-01 -2.93280333e-01
8.10426652e-01 -4.31275457e-01 -2.32039139e-01 -7.05105960e-01
-8.61479461e-01 1.08023830e-01 -1.76893458e-01 4.81771469e-01
1.43400395e+00 -1.28832057e-01 6.97870135e-01 6.23073757e-01
7.57219195e-01 -1.28611493e+00 -6.48462057e-01 5.82451582e-01
5.77233851e-01 -3.17179799e-01 4.01094168e-01 -1.01980901e+00
-4.14670736e-01 1.20370901e+00 3.33004206e-01 -9.87222269e-02
5.94880044e-01 9.10345733e-01 -6.49479747e-01 -2.61068732e-01
-9.97550070e-01 -3.57496768e-01 -2.99397141e-01 3.69108498e-01
-3.45021822e-02 1.54009671e-03 -1.05613089e+00 6.15740299e-01
-3.46010923e-01 2.02362150e-01 5.37451148e-01 1.53686070e+00
-5.93262672e-01 -2.00908399e+00 -3.21133584e-01 5.72091520e-01
-4.98782456e-01 1.86713915e-02 -3.40785950e-01 9.85533476e-01
5.45786500e-01 9.15414393e-01 8.80226418e-02 -1.97482347e-01
6.08693242e-01 3.13004643e-01 6.74439847e-01 -9.88349140e-01
-7.80007422e-01 -4.75711264e-02 7.02015996e-01 -8.97371352e-01
-3.15325201e-01 -1.04205203e+00 -1.43427682e+00 -5.06806374e-01
-5.14575899e-01 3.69807571e-01 5.09098768e-01 9.47988331e-01
3.15843254e-01 1.33731321e-01 3.89814317e-01 -2.67211527e-01
-3.19560945e-01 -2.89296478e-01 -6.19959712e-01 1.39838010e-01
-1.73377424e-01 -6.69035256e-01 -1.47992596e-01 -2.61344284e-01] | [8.317964553833008, 6.31557559967041] |
3cbe5855-e755-4d68-8553-9115caf35906 | optimized-projection-for-sparse | 1502.00115 | null | http://arxiv.org/abs/1502.00115v1 | http://arxiv.org/pdf/1502.00115v1.pdf | Optimized Projection for Sparse Representation Based Classification | Dimensionality reduction (DR) methods have been commonly used as a principled
way to understand the high-dimensional data such as facial images. In this
paper, we propose a new supervised DR method called Optimized Projection for
Sparse Representation based Classification (OP-SRC), which is based on the
recent face recognition method, Sparse Representation based Classification
(SRC). SRC seeks a sparse linear combination on all the training data for a
given query image, and make the decision by the minimal reconstruction
residual. OP-SRC is designed on the decision rule of SRC, it aims to reduce the
within-class reconstruction residual and simultaneously increase the
between-class reconstruction residual on the training data. The projections are
optimized and match well with the mechanism of SRC. Therefore, SRC performs
well in the OP-SRC transformed space. The feasibility and effectiveness of the
proposed method is verified on the Yale, ORL and UMIST databases with promising
results. | ['De-Shuang Huang', 'Can-Yi Lu'] | 2015-01-31 | null | null | null | null | ['sparse-representation-based-classification'] | ['computer-vision'] | [ 2.28040203e-01 -1.62664443e-01 -3.02633286e-01 -5.33373296e-01
-3.50104243e-01 2.18772724e-01 2.96728939e-01 -7.81744897e-01
1.20946109e-01 3.78535509e-01 6.40935659e-01 3.22199672e-01
-4.70549643e-01 -6.04630589e-01 4.31193039e-03 -1.03845453e+00
3.78193706e-01 9.30180326e-02 -2.03558281e-01 -1.35747895e-01
3.50702494e-01 8.92281711e-01 -1.90288734e+00 4.59818006e-01
5.13499320e-01 1.09522772e+00 9.98486504e-02 1.46149807e-02
-2.17981115e-01 5.94210625e-01 -2.00334951e-01 1.40734345e-01
2.70595670e-01 -5.26933491e-01 -3.64437282e-01 3.41778040e-01
2.67809719e-01 1.15684588e-02 -7.00555921e-01 1.06370533e+00
7.04951942e-01 2.47143164e-01 8.14672172e-01 -1.02527225e+00
-6.38894498e-01 1.70173094e-01 -8.41752827e-01 3.12554181e-01
3.81224960e-01 -5.03560960e-01 5.07303357e-01 -1.38235307e+00
6.71664000e-01 1.62742221e+00 3.49463224e-01 7.24631727e-01
-1.22752357e+00 -6.93337917e-01 -1.03197612e-01 3.95444632e-01
-1.74008524e+00 -7.26683736e-01 1.22347391e+00 -3.61038774e-01
6.02265120e-01 2.96110302e-01 3.87039512e-01 6.00904822e-01
3.16815600e-02 5.94523966e-01 1.18725526e+00 -5.73497355e-01
2.53824472e-01 2.27996737e-01 3.32431138e-01 6.42435431e-01
1.76263139e-01 1.95170462e-01 -6.66217029e-01 -3.28602314e-01
6.46370828e-01 3.39825064e-01 -4.30302650e-01 -5.63461542e-01
-7.19013929e-01 1.02798498e+00 1.98463649e-01 4.42783087e-01
-3.81423295e-01 -4.54175264e-01 1.03267401e-01 3.57182696e-02
2.68651694e-01 -2.06617236e-01 -7.74950087e-02 2.93082446e-01
-8.48746896e-01 -9.14625600e-02 5.31496346e-01 6.47094309e-01
7.43920505e-01 3.75116140e-01 -3.70093547e-02 1.19937229e+00
7.57069707e-01 7.62327850e-01 1.03219128e+00 -8.34051669e-01
2.46166870e-01 7.17796206e-01 -4.24961239e-01 -1.55380392e+00
-2.31175065e-01 -1.32407784e-01 -9.99659002e-01 2.20613971e-01
-2.86829263e-01 1.15192704e-01 -7.76723683e-01 1.42747104e+00
5.20808995e-01 4.17867869e-01 4.22530383e-01 1.08417141e+00
1.14389479e+00 8.10316741e-01 -2.48958275e-01 -6.51492000e-01
1.21246517e+00 -4.05579925e-01 -8.71632397e-01 9.98867378e-02
4.33865190e-01 -7.38199413e-01 7.52766550e-01 5.01859248e-01
-4.57427114e-01 -6.81924641e-01 -8.62864792e-01 2.10680366e-01
3.42762470e-02 6.13872468e-01 5.14069498e-01 7.47610629e-01
-7.91560471e-01 3.43454748e-01 -5.88819206e-01 -4.49983656e-01
4.39073563e-01 5.04490972e-01 -7.05528796e-01 -5.05902052e-01
-8.09588432e-01 6.58939540e-01 2.14396179e-01 3.55526417e-01
-6.02614462e-01 -3.85933280e-01 -1.00590336e+00 4.46588062e-02
1.00927672e-03 8.55688080e-02 5.92180312e-01 -7.33270943e-01
-1.71701527e+00 8.44840407e-01 -6.05642200e-01 7.47964531e-03
-3.43657881e-01 9.57863852e-02 -7.77132034e-01 3.84912103e-01
-6.91297650e-02 3.96206528e-01 1.17495799e+00 -9.96561825e-01
-9.07191932e-02 -8.87474120e-01 -6.81874394e-01 2.85514057e-01
-5.06638050e-01 1.50056928e-01 -1.29494026e-01 -6.53808951e-01
8.07554007e-01 -8.57610285e-01 -4.28068526e-02 -5.44678457e-02
-3.31711918e-02 -1.77104592e-01 1.58970952e+00 -6.55722022e-01
1.12622368e+00 -2.64995408e+00 3.17324281e-01 5.21282494e-01
-9.30873752e-02 4.40704972e-01 -2.11995557e-01 1.64399028e-01
-5.73956311e-01 -3.44954669e-01 -2.39782795e-01 3.15540098e-02
-6.30564332e-01 2.01824561e-01 -3.71043265e-01 7.53671527e-01
8.52191225e-02 4.29687709e-01 -3.56373668e-01 -6.54746890e-01
2.39898264e-01 7.71617889e-01 -4.13354009e-01 2.58614033e-01
3.37348789e-01 3.98173690e-01 -6.18143499e-01 8.35831523e-01
8.62675786e-01 -1.83719099e-02 2.66566485e-01 -4.97452110e-01
7.70567060e-02 -2.92058170e-01 -1.66781271e+00 1.48390913e+00
3.63029279e-02 4.11754370e-01 3.04348506e-02 -1.26547003e+00
1.57903159e+00 3.50009471e-01 6.96535408e-01 -7.87186086e-01
7.26349875e-02 1.64529413e-01 -2.03265607e-01 -6.23419523e-01
-3.41893695e-02 -3.71173829e-01 4.12520081e-01 3.51957589e-01
8.98604542e-02 2.47355297e-01 -1.88782051e-01 9.31950882e-02
5.61061978e-01 4.09741178e-02 6.04194701e-01 -4.41472977e-01
1.10011542e+00 -3.04906070e-01 8.56887639e-01 -1.58584177e-01
1.57843493e-02 6.41897202e-01 3.79871398e-01 -4.27394718e-01
-7.09600091e-01 -7.01135993e-01 -4.98429835e-01 4.34440583e-01
2.13666081e-01 -1.80165946e-01 -4.30972785e-01 -7.01509297e-01
1.22841941e-02 4.15211856e-01 -5.41428089e-01 -3.71969730e-01
-5.14117360e-01 -6.31901205e-01 8.54308084e-02 1.68177679e-01
6.45111442e-01 -9.29688632e-01 1.94799490e-02 -1.48924097e-01
1.13759920e-01 -8.29403520e-01 -3.60433549e-01 -3.30614716e-01
-1.02426541e+00 -1.05464077e+00 -6.75951660e-01 -8.76317143e-01
1.10752642e+00 7.68637657e-01 2.91621655e-01 -1.55894145e-01
-4.70517308e-01 3.06002200e-01 -4.09296095e-01 -8.05852562e-03
6.42559603e-02 -4.89112407e-01 4.42002177e-01 6.92040622e-01
6.27952218e-01 -5.84074378e-01 -3.93657237e-01 6.71570301e-01
-6.91620588e-01 -3.28559190e-01 6.70918465e-01 8.62954617e-01
9.17635918e-01 3.54005426e-01 4.86902058e-01 -7.07952082e-01
4.09931540e-01 -4.18701500e-01 -5.63956082e-01 1.46860078e-01
-7.29260504e-01 1.32510498e-01 5.18024981e-01 -4.48088527e-01
-1.18710375e+00 4.35300589e-01 7.29389638e-02 -8.98775935e-01
2.22209077e-02 3.88510227e-01 -5.84490478e-01 -4.57706064e-01
7.28883803e-01 6.62104309e-01 4.10623282e-01 -8.07680368e-01
4.87947613e-02 1.03899539e+00 3.09550285e-01 -4.03618515e-01
6.20870173e-01 6.62797749e-01 3.18557441e-01 -1.18498433e+00
-5.92110336e-01 -6.16540790e-01 -5.57913542e-01 -7.61626512e-02
5.85522890e-01 -9.63429391e-01 -4.99430239e-01 1.07770756e-01
-7.34966516e-01 5.02206743e-01 -2.92808443e-01 9.98566747e-01
-5.00435352e-01 5.80780923e-01 -1.75570384e-01 -9.54958379e-01
-4.02399987e-01 -1.06587124e+00 9.05185282e-01 4.18712020e-01
1.51495054e-01 -5.70511103e-01 1.58725247e-01 2.93743700e-01
1.15701482e-01 1.07987739e-01 8.11747134e-01 -5.91001630e-01
-2.09956273e-01 -2.03581437e-01 -2.46493116e-01 5.26823461e-01
3.26740831e-01 -6.80403486e-02 -8.55733335e-01 -3.51154625e-01
4.69391614e-01 -4.87247966e-02 6.34547532e-01 2.20837265e-01
1.22464561e+00 -3.69421542e-01 -2.30235562e-01 6.85874164e-01
1.59607399e+00 4.11621749e-01 9.64365065e-01 -1.99358299e-01
5.97054124e-01 6.39992476e-01 7.40792692e-01 4.47224677e-01
-9.39805508e-02 8.02312016e-01 -2.11639088e-02 8.32781196e-02
-3.29045653e-01 -3.38427782e-01 3.65139037e-01 8.96434903e-01
1.20387286e-01 3.44820499e-01 -5.60007989e-01 1.05479904e-01
-1.80456531e+00 -1.24878907e+00 6.62752092e-02 2.13767290e+00
5.77530682e-01 -4.96383578e-01 -1.12085149e-01 3.79185110e-01
8.25051367e-01 1.56995326e-01 -5.06467938e-01 -3.10322106e-01
-1.60437107e-01 1.58043295e-01 -1.62510537e-02 2.68405914e-01
-7.26511121e-01 8.78680110e-01 6.64642382e+00 9.07424986e-01
-1.42438209e+00 -7.48808607e-02 4.49018419e-01 6.57446682e-02
-1.78810969e-01 -7.56860971e-02 -1.26310694e+00 5.37889361e-01
6.34724081e-01 -1.55589983e-01 5.27492702e-01 1.18024278e+00
4.07099783e-01 9.33130309e-02 -8.98850918e-01 1.63318002e+00
6.02727532e-01 -1.14157677e+00 4.21145350e-01 2.03554228e-01
7.31319785e-01 -4.94235069e-01 1.02890983e-01 1.44326717e-01
-3.05685043e-01 -1.01740205e+00 -9.73506831e-03 6.59780383e-01
9.24013376e-01 -6.64523423e-01 6.53814912e-01 3.58148187e-01
-1.17323530e+00 -2.37075359e-01 -7.56942034e-01 2.97476858e-01
-2.81033993e-01 7.37176955e-01 -4.41279799e-01 3.60552639e-01
5.58223248e-01 1.08743429e+00 -3.49797308e-01 7.51743138e-01
-1.39105126e-01 4.12868023e-01 -3.00595909e-01 3.01959157e-01
-2.75711387e-01 -5.77701688e-01 5.13872862e-01 7.53229916e-01
4.87739533e-01 6.50259912e-01 1.61221735e-02 5.90299726e-01
1.19457141e-01 4.62791204e-01 -8.67487550e-01 -7.62521774e-02
4.80101913e-01 1.16449177e+00 -3.53437930e-01 -1.14400752e-01
-3.56817961e-01 6.51535034e-01 1.26472205e-01 3.58778328e-01
-2.97950059e-01 -2.02080354e-01 6.03164196e-01 1.23677896e-02
6.16607845e-01 -1.59044236e-01 -2.73872077e-01 -1.23789740e+00
8.58855024e-02 -1.06102860e+00 5.47575891e-01 -6.14532232e-01
-1.12474120e+00 4.66769338e-01 1.42071079e-02 -1.51918399e+00
-3.11995827e-04 -4.54820335e-01 -4.76051301e-01 8.52191687e-01
-1.51875532e+00 -8.21552932e-01 -3.51961285e-01 9.68852520e-01
4.67184454e-01 -8.42053533e-01 8.39902163e-01 3.71388525e-01
-9.00315881e-01 4.96760309e-01 1.91259190e-01 -3.05053405e-02
4.44092035e-01 -3.60801905e-01 -6.35380447e-01 7.01374590e-01
1.87219918e-01 6.41258597e-01 3.32286358e-01 -5.62691033e-01
-1.68247175e+00 -8.32369506e-01 7.24671423e-01 1.53305337e-01
6.87978268e-02 1.09949291e-01 -8.63795936e-01 3.03091645e-01
-5.61028242e-01 3.50207686e-01 8.89689207e-01 -1.34865075e-01
-6.34375095e-01 -6.66572869e-01 -1.60649955e+00 2.74233609e-01
7.81475663e-01 -5.30968189e-01 -7.01059103e-01 3.87008309e-01
4.97639477e-01 -9.58142877e-02 -1.03414452e+00 5.08336723e-01
5.17679393e-01 -6.73699021e-01 1.00990093e+00 -3.88391882e-01
2.56718457e-01 -6.45882249e-01 -5.81975162e-01 -9.86559331e-01
-6.19145215e-01 -1.88151732e-01 -3.70749116e-01 1.15735543e+00
-1.46580935e-01 -6.54634774e-01 1.07210362e+00 4.30711746e-01
2.14861870e-01 -8.43341887e-01 -1.13286781e+00 -7.80961752e-01
-6.13138318e-01 6.60019815e-02 6.22806847e-01 1.00327730e+00
-1.48879802e-02 3.08828294e-01 -4.10283655e-01 2.94929922e-01
8.87225509e-01 2.09060580e-01 5.28393209e-01 -1.32521451e+00
-8.70580301e-02 -6.22485429e-02 -9.18008447e-01 -7.20232010e-01
3.02068174e-01 -7.84989953e-01 -3.42188776e-01 -1.27816916e+00
2.60154009e-01 -5.07783115e-01 -2.38845617e-01 4.77940917e-01
1.55314907e-01 2.29612276e-01 1.14275254e-01 8.29005241e-01
-8.67909268e-02 8.76680672e-01 1.14030409e+00 -5.35482727e-02
-3.95438582e-01 -8.74522626e-02 -7.96154380e-01 3.56784880e-01
6.48717284e-01 -4.35454249e-01 -5.99127769e-01 -1.93190351e-02
-4.61459339e-01 1.82683453e-01 -1.63634434e-01 -9.77579653e-01
1.07504345e-01 -2.75795221e-01 5.45375228e-01 -5.40056765e-01
5.64346313e-01 -1.18384123e+00 3.73585016e-01 4.62973833e-01
-2.92703509e-01 -4.58705515e-01 -1.94400728e-01 6.28376961e-01
-4.94933099e-01 -1.89553499e-01 1.22329807e+00 1.55524999e-01
-9.23654735e-01 4.30431128e-01 1.63247705e-01 -4.09168154e-01
1.29530609e+00 -5.48747659e-01 -7.31962174e-02 -1.45428628e-01
-7.54442036e-01 -1.80588022e-01 6.31540418e-02 3.26783776e-01
1.22433603e+00 -1.80428100e+00 -7.15350032e-01 8.79526496e-01
1.84459582e-01 -2.99182534e-01 3.52562726e-01 6.82265759e-01
-3.36235553e-01 6.67160571e-01 -2.86851168e-01 -5.37472427e-01
-1.66797137e+00 7.04810381e-01 2.84463227e-01 9.40135643e-02
-8.18578362e-01 5.44036388e-01 1.30512968e-01 -3.76148313e-01
7.26032853e-02 3.62987906e-01 -8.95501077e-01 -4.76095416e-02
9.78524208e-01 4.50878173e-01 4.30231057e-02 -1.16430926e+00
-4.26204592e-01 1.04587364e+00 -1.15023404e-01 -5.34241125e-02
1.42277503e+00 5.90947531e-02 -3.96738857e-01 2.32214518e-02
1.68271875e+00 -8.54253694e-02 -8.37137401e-01 -5.15947104e-01
-2.69450516e-01 -9.94938791e-01 3.13163042e-01 -2.93415666e-01
-1.40162218e+00 7.00611234e-01 9.52262998e-01 -3.31739455e-01
1.20322299e+00 -1.19608536e-01 4.48669225e-01 4.24140364e-01
4.40837622e-01 -8.44388485e-01 1.93951145e-01 1.20751373e-01
1.11964059e+00 -1.07064831e+00 4.15724516e-01 -7.01766133e-01
-7.40311146e-01 1.17594278e+00 6.83441997e-01 -2.59304225e-01
1.05943751e+00 -1.43553555e-01 -2.02020675e-01 -3.29823315e-01
-5.60938179e-01 2.68442351e-02 2.76149511e-01 8.01890254e-01
1.53704613e-01 -9.01691541e-02 -6.65827394e-01 3.13817441e-01
2.82424316e-02 1.90293655e-01 1.10681713e-01 7.84722507e-01
-6.34336948e-01 -1.02541947e+00 -8.07471216e-01 5.05405545e-01
-7.92642310e-02 2.57787973e-01 -2.67084092e-01 1.65001705e-01
-4.02789488e-02 8.83442640e-01 -7.02857450e-02 -5.54985642e-01
2.73223877e-01 -4.37944941e-02 1.89510390e-01 -7.68226683e-01
1.73804685e-01 9.93393734e-02 -2.32443899e-01 -7.36712813e-01
-4.00890529e-01 -7.82783806e-01 -1.33436179e+00 -7.19484016e-02
-2.47537106e-01 3.48680645e-01 9.45001364e-01 8.12906384e-01
6.46301508e-01 -8.01512003e-02 1.42402065e+00 -5.53431273e-01
-6.66280866e-01 -6.84090555e-01 -1.01967049e+00 2.67856956e-01
-2.93234847e-02 -7.35964715e-01 -5.59113264e-01 4.98260483e-02] | [12.478577613830566, 0.41213634610176086] |
bef91b72-7820-459e-9ffc-6fbe67e275d9 | linkbert-pretraining-language-models-with | 2203.15827 | null | https://arxiv.org/abs/2203.15827v1 | https://arxiv.org/pdf/2203.15827v1.pdf | LinkBERT: Pretraining Language Models with Document Links | Language model (LM) pretraining can learn various knowledge from text corpora, helping downstream tasks. However, existing methods such as BERT model a single document, and do not capture dependencies or knowledge that span across documents. In this work, we propose LinkBERT, an LM pretraining method that leverages links between documents, e.g., hyperlinks. Given a text corpus, we view it as a graph of documents and create LM inputs by placing linked documents in the same context. We then pretrain the LM with two joint self-supervised objectives: masked language modeling and our new proposal, document relation prediction. We show that LinkBERT outperforms BERT on various downstream tasks across two domains: the general domain (pretrained on Wikipedia with hyperlinks) and biomedical domain (pretrained on PubMed with citation links). LinkBERT is especially effective for multi-hop reasoning and few-shot QA (+5% absolute improvement on HotpotQA and TriviaQA), and our biomedical LinkBERT sets new states of the art on various BioNLP tasks (+7% on BioASQ and USMLE). We release our pretrained models, LinkBERT and BioLinkBERT, as well as code and data at https://github.com/michiyasunaga/LinkBERT. | ['Percy Liang', 'Jure Leskovec', 'Michihiro Yasunaga'] | 2022-03-29 | null | https://aclanthology.org/2022.acl-long.551 | https://aclanthology.org/2022.acl-long.551.pdf | acl-2022-5 | ['medical-relation-extraction', 'triviaqa', 'pico'] | ['medical', 'miscellaneous', 'natural-language-processing'] | [-1.57010972e-01 5.72884619e-01 -7.40471363e-01 -2.20216274e-01
-1.17364371e+00 -7.06573009e-01 6.52817905e-01 5.45340717e-01
-6.05761409e-01 1.34807241e+00 4.53664631e-01 -4.44337308e-01
-1.44423708e-01 -6.12775862e-01 -1.17734039e+00 -4.01185393e-01
-1.65703401e-01 1.09175384e+00 3.18033546e-01 -1.46314576e-01
-1.28090560e-01 1.05512500e-01 -3.94092321e-01 8.47442567e-01
1.02473044e+00 3.47827584e-01 1.25275403e-01 7.89177537e-01
-4.94529158e-01 9.99155164e-01 -4.44632769e-01 -9.68590140e-01
-3.61683428e-01 -4.96624500e-01 -1.31424463e+00 -7.79334307e-01
4.50426668e-01 1.09230645e-01 -5.60105562e-01 7.63591349e-01
5.91104865e-01 -2.65047878e-01 8.07775676e-01 -1.14723694e+00
-8.35667074e-01 1.15922177e+00 -5.06836712e-01 5.18898368e-01
3.17466073e-02 1.07402369e-01 1.20309973e+00 -7.93089032e-01
1.18560934e+00 1.25131118e+00 7.00305521e-01 6.63824975e-01
-1.22811115e+00 -5.92078328e-01 -3.04916762e-02 3.13044697e-01
-1.20480597e+00 -3.44517052e-01 1.60500601e-01 -4.15835261e-01
1.52165592e+00 3.51904482e-02 8.81863534e-02 1.31100166e+00
6.32733047e-01 8.12393188e-01 7.67809868e-01 -3.98073763e-01
-2.00419441e-01 -7.38348290e-02 4.59760517e-01 1.22755539e+00
1.64737836e-01 -3.65403086e-01 -6.79484487e-01 -3.14140379e-01
8.58280957e-02 -2.36489102e-01 -3.95368546e-01 2.32286021e-01
-1.22350717e+00 7.03965425e-01 5.38392246e-01 2.33870357e-01
-4.61987108e-02 2.40763709e-01 5.94433546e-01 3.14519405e-01
6.35490298e-01 4.10250247e-01 -7.38325179e-01 1.80778921e-01
-7.96390355e-01 -2.33895466e-01 1.08469176e+00 1.11155999e+00
5.66202164e-01 -8.16029906e-01 -4.93445784e-01 9.26234484e-01
3.01109731e-01 4.04846966e-01 5.36698997e-01 -7.03782022e-01
6.85708106e-01 4.38038826e-01 -3.70394737e-01 -6.10507727e-01
-6.06417656e-01 -5.85526884e-01 -8.78826857e-01 -5.96658945e-01
2.09249079e-01 -2.85951972e-01 -1.01587725e+00 1.80404437e+00
1.11783683e-01 3.49178970e-01 2.70731181e-01 4.12754089e-01
1.55279863e+00 8.07371974e-01 5.28053343e-01 -8.37742388e-02
1.32794416e+00 -1.43699169e+00 -8.50152314e-01 -2.35686719e-01
1.23044920e+00 -7.00546622e-01 6.77874565e-01 3.62351090e-01
-1.20989966e+00 -2.68030465e-01 -8.38074684e-01 -6.95584059e-01
-7.82644451e-01 2.03419328e-01 5.82507908e-01 -1.84665486e-01
-1.14734447e+00 4.40876365e-01 -9.38855529e-01 -5.00430703e-01
4.62865770e-01 2.77745783e-01 -4.05372977e-01 -3.46598834e-01
-1.61637390e+00 1.21574199e+00 5.47689795e-01 -7.16656819e-02
-1.01863670e+00 -9.16903496e-01 -5.84798455e-01 1.54869333e-01
3.71194243e-01 -1.04103279e+00 1.07968938e+00 -1.69259235e-01
-1.05908310e+00 1.20179880e+00 -2.48895898e-01 -5.99384725e-01
4.19524223e-01 -3.40519130e-01 -4.88576531e-01 2.50132233e-01
1.06667519e-01 1.04204023e+00 -1.29522160e-01 -9.22299325e-01
-2.30407596e-01 -1.42702669e-01 -3.84991355e-02 6.53409809e-02
-3.13826382e-01 -2.10054994e-01 -9.42447245e-01 -2.32881233e-01
-3.71869445e-01 -7.26317346e-01 -7.11161047e-02 -7.28575289e-02
-9.99119878e-01 -4.68568742e-01 3.75445575e-01 -8.69020581e-01
9.83240485e-01 -1.58824420e+00 3.18781286e-01 -8.57294425e-02
4.28039759e-01 2.85336494e-01 -5.92245281e-01 7.73029387e-01
-7.82116577e-02 3.77602756e-01 -2.43882969e-01 -3.62364709e-01
-1.95430815e-01 4.76627409e-01 -1.62779763e-01 1.91122323e-01
2.80013531e-01 1.44530499e+00 -1.18002665e+00 -9.11467254e-01
-5.76668441e-01 5.78158021e-01 -4.30949748e-01 3.59640643e-02
-7.11329699e-01 3.24240178e-01 -4.74427670e-01 5.11534274e-01
2.35902831e-01 -7.29418993e-01 3.17280769e-01 -4.32693154e-01
2.66057879e-01 6.46791220e-01 -1.64064199e-01 2.09741521e+00
-3.27083141e-01 5.72080791e-01 -1.16176665e-01 -8.55703413e-01
7.22359419e-01 3.31051111e-01 3.92354190e-01 -5.21812439e-01
-1.21074878e-01 1.02914222e-01 6.23860061e-02 -7.26327837e-01
-7.10345134e-02 7.46249855e-02 3.42424601e-01 4.16789055e-01
5.97407997e-01 4.94095236e-02 5.82135916e-01 9.50644672e-01
1.50134146e+00 1.56522810e-01 1.41860917e-01 -1.73835918e-01
4.61959124e-01 1.91365406e-01 2.69230723e-01 6.59552872e-01
3.31162035e-01 2.16574728e-01 9.94517624e-01 3.70293111e-02
-6.06774032e-01 -1.05642748e+00 -2.24027440e-01 1.07974124e+00
-3.23285520e-01 -7.97183096e-01 -5.32258987e-01 -1.11627424e+00
4.72078733e-02 7.08493948e-01 -5.16255558e-01 -4.04219270e-01
-4.97586727e-01 -1.14470947e+00 1.02418447e+00 3.93541664e-01
2.81861216e-01 -8.88705730e-01 3.10340613e-01 1.91871867e-01
-4.45122570e-01 -1.08760071e+00 -5.53384781e-01 4.68138844e-01
-9.25304770e-01 -1.19201493e+00 -5.76160967e-01 -8.05369318e-01
6.92203462e-01 -3.34447503e-01 1.59016836e+00 8.37829635e-02
-5.13664126e-01 3.43748748e-01 -1.89834580e-01 -3.86709422e-01
-4.95844245e-01 3.21445346e-01 -3.30228746e-01 -7.57809341e-01
3.42637360e-01 -2.73554564e-01 -5.37196457e-01 -1.04296587e-01
-8.02332103e-01 2.60438472e-01 8.19524169e-01 1.15197182e+00
8.21036935e-01 -3.18008035e-01 7.79109001e-01 -1.51275504e+00
6.47991002e-01 -9.95327055e-01 -2.96561897e-01 8.02297115e-01
-6.08505070e-01 1.86794400e-01 3.96629661e-01 -1.48188725e-01
-9.26721692e-01 -2.87621349e-01 -2.96182185e-01 1.70531478e-02
2.47218117e-01 1.24417245e+00 -6.33696690e-02 3.58797729e-01
7.43536055e-01 -6.91049770e-02 -1.70751378e-01 -5.96287072e-01
7.92068839e-01 5.30321360e-01 5.77802539e-01 -7.11759984e-01
1.80328220e-01 2.22270951e-01 1.42888024e-01 -3.47802073e-01
-1.25794935e+00 -4.02963966e-01 -9.44163203e-01 2.90260881e-01
1.00179100e+00 -9.71856654e-01 -5.45486331e-01 -1.27555445e-01
-1.41355479e+00 -6.96711183e-01 1.11790098e-01 3.36343557e-01
-1.97528094e-01 1.32077232e-01 -1.22616148e+00 -1.02583490e-01
-7.16234684e-01 -8.21602285e-01 9.62735891e-01 -2.09963962e-01
-2.64588773e-01 -1.39739513e+00 5.04074991e-01 4.55586672e-01
-1.74175352e-01 1.14097141e-01 1.51207769e+00 -9.82316196e-01
-4.19792920e-01 -3.86250243e-02 -3.49096239e-01 2.11374871e-02
1.41599089e-01 -6.73036501e-02 -8.10876667e-01 -1.56103328e-01
-7.58399665e-01 -7.48855174e-01 1.36250782e+00 1.68870687e-01
1.24553680e+00 -3.87937129e-01 -1.02448666e+00 5.53322494e-01
1.18514049e+00 -6.79210424e-02 3.66824716e-01 1.25907823e-01
7.48517513e-01 5.69630980e-01 3.24535877e-01 -1.87139317e-01
4.36130375e-01 2.86249369e-01 9.62946489e-02 -3.28025907e-01
-5.28917015e-01 -3.67862582e-01 2.83620238e-01 1.18171060e+00
1.91273674e-01 -8.48042071e-01 -1.20626283e+00 3.86461258e-01
-1.89354897e+00 -5.45287430e-01 -2.49164417e-01 1.58223736e+00
1.49154532e+00 2.83984154e-01 -2.32093260e-01 -7.41541684e-01
3.72085005e-01 -3.63712668e-01 -7.06132174e-01 -1.07485168e-01
-3.63891155e-01 3.03936005e-01 3.75170738e-01 7.18551040e-01
-7.71880209e-01 1.13947129e+00 5.71879482e+00 7.97573090e-01
-8.33699584e-01 5.13597190e-01 9.27720964e-01 -2.38731608e-01
-4.05607760e-01 -1.47222325e-01 -1.04286122e+00 3.82925987e-01
1.36942434e+00 -2.56368160e-01 -7.06162304e-02 3.19061458e-01
-4.30949032e-03 1.44248843e-01 -1.43355703e+00 5.88366628e-01
1.06189691e-01 -1.82178986e+00 2.41869450e-01 -1.37070477e-01
6.64648473e-01 6.84880197e-01 -3.84654589e-02 5.85056186e-01
8.00456882e-01 -1.28598654e+00 -7.16250092e-02 8.74915421e-01
7.88972676e-01 -4.03758854e-01 9.85692739e-01 3.19309741e-01
-6.14824533e-01 5.45172155e-01 -4.68273431e-01 5.79771578e-01
8.42667744e-02 8.69631827e-01 -1.09666848e+00 7.80649841e-01
4.22644079e-01 9.99072194e-01 -5.85097134e-01 1.04374397e+00
-6.57415926e-01 9.45993304e-01 -6.64762557e-02 3.76370512e-02
2.89374322e-01 1.78040996e-01 1.77709863e-01 1.81239450e+00
3.67579702e-03 2.63697803e-01 2.68334746e-01 7.74553061e-01
-7.52286673e-01 2.12703243e-01 -3.56420636e-01 -3.93475533e-01
5.15178502e-01 1.30430710e+00 -5.40995181e-01 -5.50828874e-01
-5.41800022e-01 9.40432191e-01 9.22840953e-01 5.79957187e-01
-7.70752847e-01 -2.95263112e-01 1.97224081e-01 -2.67826945e-01
-2.05439895e-01 1.32021972e-03 4.65957820e-02 -1.02591527e+00
-4.11375940e-01 -7.21377432e-01 9.56008017e-01 -8.46018076e-01
-1.69280136e+00 6.72995389e-01 -5.91192618e-02 -6.36555195e-01
-8.75279382e-02 -8.44076455e-01 -4.25671577e-01 9.21876132e-01
-1.65695381e+00 -1.23650563e+00 1.59610435e-01 2.59934872e-01
3.73574793e-01 -7.84103796e-02 9.07198429e-01 4.42641824e-01
-6.88968778e-01 6.63677573e-01 3.17465186e-01 2.48005882e-01
1.27635145e+00 -1.30096281e+00 3.09000164e-01 1.68736443e-01
1.98256671e-01 1.02631855e+00 2.64662266e-01 -1.06622016e+00
-1.23515964e+00 -1.44775677e+00 1.28662610e+00 -7.89491951e-01
9.24761117e-01 -2.72337914e-01 -1.17761040e+00 1.06650949e+00
5.14621437e-01 -3.56971398e-02 9.98128533e-01 2.96855986e-01
-3.80745828e-01 2.00917557e-01 -9.24990892e-01 4.07446444e-01
1.20342088e+00 -5.31159997e-01 -7.04607666e-01 1.09188211e+00
9.56351340e-01 -4.99553174e-01 -1.12136221e+00 2.45538563e-01
1.26967773e-01 -2.53529400e-01 1.04592192e+00 -1.22521889e+00
8.73815417e-01 1.15094937e-01 2.71974146e-01 -1.49506915e+00
-3.81908655e-01 -2.46677309e-01 -4.92391050e-01 1.27994728e+00
1.10119152e+00 -5.92841506e-01 6.35659158e-01 1.84864938e-01
-5.02632201e-01 -1.00849235e+00 -6.78861618e-01 -7.50169575e-01
5.47282994e-01 8.32858458e-02 6.26654774e-02 1.22961998e+00
3.86302143e-01 8.70734692e-01 -3.37816551e-02 -1.41500048e-02
4.15729195e-01 -2.55809456e-01 2.26011232e-01 -1.05423248e+00
-5.53214729e-01 -3.31239372e-01 2.17221588e-01 -8.99423361e-01
4.96631384e-01 -1.74180233e+00 -3.73890251e-02 -2.08681393e+00
7.16288865e-01 -2.74886787e-01 -4.52414781e-01 9.50254202e-01
-4.31182295e-01 7.02292770e-02 -3.45250577e-01 2.85656720e-01
-8.69612753e-01 4.16082144e-01 1.27778530e+00 -5.52095711e-01
1.28875583e-01 -5.40335476e-01 -4.24457133e-01 5.39525747e-01
6.45551205e-01 -5.36154091e-01 -5.54009855e-01 -9.21401978e-01
4.94416982e-01 1.96414888e-02 1.27293408e-01 -4.98390526e-01
5.95574021e-01 2.06590712e-01 4.90057260e-01 -7.28888273e-01
2.34563828e-01 -1.74071670e-01 -5.69853373e-02 7.30778515e-01
-9.13495243e-01 -1.87444501e-02 5.44957340e-01 5.47912776e-01
-3.96829657e-02 -3.97805333e-01 3.74552697e-01 -4.05386001e-01
-3.68861407e-01 3.80662322e-01 -1.14856139e-01 3.83760035e-01
3.56057405e-01 7.56089926e-01 -1.18447280e+00 -1.47004455e-01
-1.05528665e+00 8.75859320e-01 -2.10844204e-01 1.12521060e-01
5.06129265e-01 -9.01401043e-01 -9.00379837e-01 -5.57909608e-01
6.20756298e-02 -1.32673867e-02 5.55537418e-02 1.10491383e+00
-4.10889238e-01 1.03011835e+00 5.41773140e-02 -3.90952229e-01
-1.22837377e+00 5.74317992e-01 2.07047537e-01 -7.05362797e-01
-3.72843891e-01 1.27911401e+00 2.72896796e-01 -6.71333492e-01
3.81866455e-01 -1.87111333e-01 -2.43016571e-01 1.12511031e-01
4.67991292e-01 1.77954584e-01 3.31547290e-01 -2.19770651e-02
-3.17271173e-01 2.18526348e-01 -4.49557304e-01 1.27836823e-01
1.35154474e+00 6.67567626e-02 -6.81718349e-01 4.84743655e-01
1.36945736e+00 -8.69252905e-02 -6.82784557e-01 -4.16748822e-01
4.15467739e-01 3.02758485e-01 4.08221669e-02 -1.53153944e+00
-7.50783622e-01 9.32530940e-01 2.16539308e-01 -3.65170658e-01
5.84777534e-01 5.58929145e-01 8.72215688e-01 8.67790878e-01
1.49411902e-01 -5.86056471e-01 2.51921862e-01 5.99389255e-01
8.91777217e-01 -1.22523427e+00 1.51131496e-01 -5.64471364e-01
-4.23455745e-01 9.47971165e-01 7.95556068e-01 4.65179592e-01
5.85543752e-01 3.99317533e-01 9.61664878e-03 -5.97701192e-01
-1.31033885e+00 -1.21764824e-01 7.66345680e-01 3.09170663e-01
1.13311672e+00 -8.11279044e-02 -6.42879844e-01 7.07164586e-01
-2.87387669e-02 1.42503902e-01 2.20274940e-01 7.51185298e-01
-2.05393553e-01 -1.33737302e+00 1.75067902e-01 7.09943295e-01
-5.49605012e-01 -7.39772499e-01 -7.76434839e-01 7.47377038e-01
1.85982376e-01 7.14933634e-01 1.82223134e-02 -1.03203557e-01
-1.70123260e-02 3.67232382e-01 5.48085392e-01 -1.02613723e+00
-3.82033736e-01 1.26029432e-01 6.05870306e-01 -3.06431860e-01
-2.43629292e-01 -2.93811291e-01 -1.74992001e+00 -5.96589409e-02
-1.31248131e-01 5.70892453e-01 3.37747753e-01 9.45677996e-01
5.63166857e-01 8.99229348e-01 -2.98543841e-01 3.40829790e-02
-1.84211545e-02 -1.15417325e+00 -3.39413397e-02 5.62058799e-02
7.03961030e-02 -4.11505908e-01 2.94736791e-02 2.55593151e-01] | [8.802821159362793, 8.567546844482422] |
a6572bb9-40f1-4de7-a840-a939e81e9fb0 | h2-golden-retriever-methodology-and-tool-for | 2211.08614 | null | https://arxiv.org/abs/2211.08614v1 | https://arxiv.org/pdf/2211.08614v1.pdf | H2-Golden-Retriever: Methodology and Tool for an Evidence-Based Hydrogen Research Grantsmanship | Hydrogen is poised to play a major role in decarbonizing the economy. The need to discover, develop, and understand low-cost, high-performance, durable materials that can help maximize the cost of electrolysis as well as the need for an intelligent tool to make evidence-based Hydrogen research funding decisions relatively easier warranted this study.In this work, we developed H2 Golden Retriever (H2GR) system for Hydrogen knowledge discovery and representation using Natural Language Processing (NLP), Knowledge Graph and Decision Intelligence. This system represents a novel methodology encapsulating state-of-the-art technique for evidence-based research grantmanship. Relevant Hydrogen papers were scraped and indexed from the web and preprocessing was done using noise and stop-words removal, language and spell check, stemming and lemmatization. The NLP tasks included Named Entity Recognition using Stanford and Spacy NER, topic modeling using Latent Dirichlet Allocation and TF-IDF. The Knowledge Graph module was used for the generation of meaningful entities and their relationships, trends and patterns in relevant H2 papers, thanks to an ontology of the hydrogen production domain. The Decision Intelligence component provides stakeholders with a simulation environment for cost and quantity dependencies. PageRank algorithm was used to rank papers of interest. Random searches were made on the proposed H2GR and the results included a list of papers ranked by relevancy score, entities, graphs of relationships between the entities, ontology of H2 production and Causal Decision Diagrams showing component interactivity. Qualitative assessment was done by the experts and H2GR is deemed to function to a satisfactory level. | ['Gregory Renard', 'Rozhin Yasaei', 'Loveneesh Rana', 'Lorien Pratt', 'Joseph Wiggins', 'Olusola Olabanjo', 'Paul Seurin'] | 2022-11-16 | null | null | null | null | ['lemmatization'] | ['natural-language-processing'] | [-5.26254952e-01 4.33336079e-01 -2.88945526e-01 1.85960189e-01
-3.95261586e-01 -5.43580234e-01 8.57679784e-01 6.90828562e-01
-3.03449154e-01 7.93086410e-01 4.90719259e-01 -5.95121980e-01
-7.60712802e-01 -1.29276133e+00 -2.38949761e-01 -5.62881112e-01
3.49112577e-03 8.94399464e-01 1.55854687e-01 -1.48540974e-01
8.49208415e-01 8.02335978e-01 -1.50225627e+00 -3.07250163e-03
6.91326499e-01 5.35191953e-01 7.66012669e-01 5.67616642e-01
-4.23976481e-01 6.88538074e-01 -6.68219388e-01 -2.09792750e-03
1.68830782e-01 -2.26466373e-01 -8.17143202e-01 -5.47518432e-01
-7.15931356e-01 2.10493729e-01 -1.67444587e-01 8.96048069e-01
6.74436390e-01 2.11365566e-01 9.47351515e-01 -1.12239408e+00
-6.05495811e-01 6.04939461e-01 -1.24247834e-01 1.70709968e-01
5.09095907e-01 -1.33392960e-01 7.18715429e-01 -1.02376878e+00
1.27323103e+00 1.12284410e+00 1.87599853e-01 -1.42638668e-01
-4.19498622e-01 -7.37960041e-01 -4.31560993e-01 7.13673353e-01
-1.50178957e+00 3.77466120e-02 4.62525368e-01 -7.13210046e-01
1.68793631e+00 2.32443020e-01 9.29628432e-01 3.43699247e-01
3.12260062e-01 3.51147503e-02 1.16784382e+00 -8.69552135e-01
5.35659671e-01 4.37482536e-01 2.78308928e-01 3.59084517e-01
6.92248225e-01 -1.32910505e-01 -7.05924332e-01 -9.09073055e-02
3.19145203e-01 -4.96301651e-02 -1.48241399e-02 1.11715458e-01
-8.49093139e-01 6.22945189e-01 1.50867507e-01 8.90971780e-01
-1.00632787e+00 -4.00466442e-01 2.41474852e-01 3.29158679e-02
3.14543582e-02 8.89360011e-01 -2.01038405e-01 -5.51903099e-02
-7.52192080e-01 2.17098311e-01 1.16137528e+00 1.01689625e+00
5.89603662e-01 -3.57418239e-01 -3.49085122e-01 3.63720894e-01
5.50959170e-01 4.79032636e-01 4.92562294e-01 -7.38659978e-01
1.43414676e-01 1.15104568e+00 2.20106333e-01 -1.10346937e+00
-3.67404997e-01 -1.34822642e-02 -4.33901191e-01 3.52910557e-03
-2.79443532e-01 -2.43720755e-01 -8.32819164e-01 1.00584078e+00
2.80851007e-01 -3.93613786e-01 5.00590205e-01 5.55751026e-01
1.12855506e+00 1.14100897e+00 5.14512479e-01 -4.29867059e-01
1.60756600e+00 -3.65728587e-01 -1.07871354e+00 7.12542236e-01
4.48007286e-01 -7.93825328e-01 3.63955170e-01 3.70460749e-01
-9.77046072e-01 -3.79838757e-02 -1.19150722e+00 -1.14044093e-01
-1.39960325e+00 -2.78854489e-01 5.80254138e-01 3.56568843e-01
-7.59525180e-01 6.82613254e-01 -3.57606262e-01 -8.72959852e-01
1.32927954e-01 3.30593407e-01 -1.60754353e-01 -1.30015716e-01
-1.65027511e+00 1.46144617e+00 7.07124710e-01 -2.10226372e-01
-8.41865540e-01 -8.18674743e-01 -3.52720827e-01 3.89418826e-02
2.19734311e-01 -6.25436544e-01 7.42311716e-01 1.12662584e-01
-1.15410018e+00 6.78314984e-01 -4.08855900e-02 -5.13314128e-01
1.94501519e-01 2.04084396e-01 -7.62863100e-01 2.99202114e-01
3.47969472e-01 3.07314634e-01 -6.75355792e-02 -8.93007994e-01
-8.47677886e-01 -5.14375031e-01 -4.28844959e-01 4.19914246e-01
-2.98048288e-01 2.77643234e-01 -3.18903804e-01 -2.08268806e-01
-8.51699859e-02 -6.51208878e-01 4.00147550e-02 -8.37172508e-01
-3.24926704e-01 -9.07260478e-01 8.69929552e-01 -1.12624276e+00
1.34353030e+00 -1.56159019e+00 1.36749521e-01 5.69224000e-01
1.12688206e-01 -3.95353138e-02 6.44086599e-01 9.46328461e-01
8.98255333e-02 6.23514652e-01 2.16121197e-01 7.42646635e-01
3.34367692e-01 -4.54496145e-02 2.52020448e-01 2.21785322e-01
-1.94407940e-01 5.84709644e-01 -8.38943362e-01 -2.78114706e-01
2.17615575e-01 5.04882872e-01 1.73961863e-01 1.58488035e-01
-1.80956647e-01 -2.61230230e-01 -7.06657052e-01 8.86102557e-01
4.99801971e-02 7.23506585e-02 4.51661855e-01 -3.12775403e-01
-9.28510487e-01 2.58085608e-01 -1.18543530e+00 1.30375457e+00
-3.47015113e-01 5.30190706e-01 -3.00172746e-01 -9.35193300e-01
1.13485396e+00 5.89554727e-01 5.70528269e-01 -1.02974606e+00
-8.93093050e-02 4.47913945e-01 -3.22404832e-01 -8.96873295e-01
4.02095169e-01 -1.42732948e-01 -9.13389325e-02 4.74349447e-02
-4.82770056e-02 -2.15748936e-01 5.25259435e-01 4.37279135e-01
1.19583738e+00 -6.88081831e-02 6.83950186e-01 -6.72962785e-01
1.87646255e-01 5.64529121e-01 3.44306558e-01 3.92904162e-01
1.92032143e-01 -2.42522620e-02 4.18489337e-01 -1.38774276e-01
-1.28978217e+00 -5.82294226e-01 -1.57471746e-01 4.08212274e-01
-7.94764981e-03 -3.89090210e-01 -4.63967443e-01 -1.49123624e-01
6.13373406e-02 1.16716611e+00 -2.49919847e-01 7.71586522e-02
1.16433993e-01 -4.88400400e-01 1.30953461e-01 -2.70841997e-02
5.23958623e-01 -1.07495236e+00 -7.61483550e-01 3.47223639e-01
3.08158159e-01 -5.69472909e-01 4.94848073e-01 3.57044727e-01
-5.17397106e-01 -1.27757108e+00 -7.67464221e-01 -6.17762089e-01
2.75261253e-01 -6.09055459e-02 1.01419067e+00 -3.28634530e-01
-6.58836901e-01 1.49016768e-01 -2.77177721e-01 -1.06746471e+00
-2.66670287e-01 1.12832308e-01 -6.28886893e-02 -1.09443414e+00
8.30522716e-01 -4.64317232e-01 -6.76169634e-01 -4.83123921e-02
-6.45276427e-01 -5.90667389e-02 7.49873281e-01 1.42123237e-01
6.79806590e-01 9.11780059e-01 9.14734542e-01 -7.46211886e-01
8.94598007e-01 -1.14866650e+00 -6.63172901e-01 4.01003331e-01
-1.39190710e+00 1.22881427e-01 8.33764002e-02 2.33029872e-01
-1.08716011e+00 -3.31589580e-01 4.37353522e-01 2.65938919e-02
-2.35197961e-01 8.79117608e-01 -5.36566019e-01 4.13068235e-01
5.45127988e-01 -2.14543983e-01 -3.66354257e-01 -5.08347511e-01
4.61815029e-01 8.65611911e-01 2.73790896e-01 -3.80222827e-01
5.84050357e-01 2.12140121e-02 2.69201815e-01 -9.40995097e-01
-7.10345283e-02 -9.36408699e-01 -3.76663476e-01 -4.66984421e-01
1.14878821e+00 -8.77549708e-01 -5.70010364e-01 -1.45054720e-02
-1.14448166e+00 1.99866280e-01 -2.09523186e-01 5.54163516e-01
1.03664748e-01 -1.45419016e-01 -9.03323516e-02 -1.03619969e+00
-6.94842517e-01 -5.55230319e-01 1.66279823e-01 5.70667863e-01
-3.44068468e-01 -6.01742148e-01 1.22360796e-01 4.18552160e-01
4.46095020e-01 4.10284400e-01 1.40223479e+00 -9.59027052e-01
-5.78007162e-01 -2.74950951e-01 -1.09800883e-01 -6.80847615e-02
-1.49884308e-02 1.86556410e-02 -3.97469312e-01 2.98879534e-01
-1.96388289e-01 2.86682606e-01 3.96367759e-01 2.19397575e-01
5.72501004e-01 -5.48133016e-01 -5.83091080e-01 -1.98353991e-01
1.72369087e+00 9.11242723e-01 1.01536047e+00 7.83935547e-01
5.82004189e-01 8.96142364e-01 5.51633835e-01 5.20249128e-01
4.21103209e-01 3.12485367e-01 -5.31318337e-02 2.80810744e-01
-1.95414171e-01 -9.69349295e-02 1.57660127e-01 7.96405375e-01
-3.94450992e-01 -1.08593732e-01 -1.28884912e+00 8.13507080e-01
-1.62857032e+00 -9.39507544e-01 -4.71037507e-01 2.14551497e+00
5.66706121e-01 1.56561926e-01 6.23571724e-02 9.75853056e-02
7.45354533e-01 -5.41633248e-01 -3.92132521e-01 -4.16415840e-01
-2.42993057e-01 3.18336993e-01 9.56100345e-01 2.81643897e-01
-2.98529208e-01 7.52971470e-01 5.43554258e+00 6.92615867e-01
-5.58623314e-01 1.11469246e-01 3.74475777e-01 -5.59261702e-02
-6.32136762e-01 4.60309356e-01 -9.15749133e-01 1.71205029e-01
1.29540205e+00 -8.65860522e-01 3.36296052e-01 6.67103171e-01
7.25760400e-01 -3.57770443e-01 -4.92176861e-01 6.02103174e-01
-1.30285576e-01 -1.75533450e+00 2.64381189e-02 1.76960796e-01
8.21322978e-01 3.51264179e-02 -7.07264960e-01 1.10819854e-01
5.24020433e-01 -8.59828591e-01 6.35179758e-01 1.05672574e+00
2.65154809e-01 -9.25200284e-01 8.38773727e-01 -3.19218673e-02
-1.14616239e+00 -8.40628147e-02 -3.20121258e-01 1.10060588e-01
2.50630081e-02 8.24783146e-01 -9.89593506e-01 1.05070662e+00
7.73465514e-01 3.01750571e-01 -2.33266696e-01 8.94993424e-01
-1.76399961e-01 2.94744164e-01 -5.20070136e-01 -4.63789582e-01
1.87159027e-03 -4.02528077e-01 4.81451213e-01 1.18057930e+00
7.56198287e-01 5.59156716e-01 -2.10205600e-01 8.27193558e-01
-2.31816635e-01 6.35188699e-01 -8.91175449e-01 -6.95148051e-01
1.05070364e+00 1.13428879e+00 -1.02772856e+00 -4.78418708e-01
-1.99766606e-01 3.67257953e-01 -2.91468054e-01 3.08261931e-01
-2.07635343e-01 -9.22018349e-01 -1.09561067e-02 6.42103076e-01
1.64661407e-01 -1.84595510e-01 -3.74571025e-01 -3.29303563e-01
-3.82705897e-01 -5.54476678e-01 5.32328904e-01 -8.46093237e-01
-8.88107002e-01 1.36221856e-01 3.75306904e-01 -4.52519089e-01
-7.46401623e-02 -2.89476126e-01 -4.20961112e-01 1.15652573e+00
-1.33713579e+00 -6.42329335e-01 -2.58811504e-01 1.81386366e-01
4.43642139e-01 -4.28474039e-01 8.18630636e-01 3.01904112e-01
-5.39605975e-01 -4.79206592e-01 2.65621752e-01 -5.94151080e-01
3.76641691e-01 -1.38432801e+00 -3.41766298e-01 4.84756947e-01
-3.76852542e-01 7.15603769e-01 7.41909087e-01 -1.31631577e+00
-1.71920025e+00 -7.36594200e-01 1.33516252e+00 -7.82788470e-02
7.91156352e-01 7.57923722e-02 -6.61245823e-01 -8.85505006e-02
5.93771875e-01 -8.64133835e-01 7.75095224e-01 -4.44998033e-02
3.18115562e-01 5.72696142e-02 -1.15684032e+00 3.96456718e-01
5.91335893e-01 -3.71514440e-01 -7.24503994e-01 6.90285087e-01
4.49112594e-01 2.42944151e-01 -1.44205618e+00 5.20010525e-03
3.31795722e-01 -3.88401330e-01 6.68631673e-01 -4.63956475e-01
2.38753423e-01 -4.24126327e-01 -1.09388702e-01 -8.97665918e-01
-2.75067389e-01 -7.68278599e-01 -3.19577530e-02 1.58123100e+00
8.19990158e-01 -2.27922037e-01 4.02605027e-01 1.00910532e+00
-2.56586462e-01 -4.18767333e-01 -7.90147483e-01 -4.77266848e-01
2.09057776e-04 -1.32694216e-02 5.57904720e-01 1.19821167e+00
4.47171777e-01 4.47862893e-01 1.99199766e-01 1.44840643e-01
5.98411024e-01 -1.71073273e-01 3.65956724e-01 -1.51757300e+00
3.55113804e-01 -4.28267777e-01 -3.20438653e-01 1.73676848e-01
-1.36120245e-01 -1.02499282e+00 -2.51447141e-01 -2.54480124e+00
2.26510510e-01 -3.05878669e-01 -3.08143198e-01 4.36682612e-01
4.12211955e-01 -5.81467211e-01 1.70952026e-02 5.37020087e-01
-2.05127105e-01 1.69349059e-01 6.89052403e-01 6.83717942e-03
-3.29783559e-01 -6.09243155e-01 -7.77309656e-01 4.61568296e-01
8.51466894e-01 -7.10439205e-01 -3.08435082e-01 2.90728807e-01
8.08141530e-01 -5.21797128e-02 6.81361277e-03 -8.39533329e-01
6.20293915e-01 -3.23485017e-01 2.92780995e-01 -7.34204471e-01
-3.14415306e-01 -7.20876873e-01 9.80500519e-01 3.95637780e-01
-1.09267756e-01 2.05151021e-01 -9.98181254e-02 4.67378229e-01
6.27219006e-02 -5.38310945e-01 2.05974162e-01 -4.34852332e-01
-8.72919142e-01 -1.40493482e-01 -7.28697300e-01 -4.68381435e-01
1.28232467e+00 -1.49911746e-01 -4.54789340e-01 8.37476458e-03
-5.76668620e-01 3.25691730e-01 3.21582705e-01 3.15040410e-01
2.94418186e-01 -1.09871686e+00 -5.91288447e-01 -5.90183496e-01
-1.23087682e-01 -1.30533785e-01 1.33669063e-01 4.32161033e-01
-7.26981997e-01 8.04171145e-01 -1.15319721e-01 2.81886280e-01
-9.54822481e-01 6.29354835e-01 -5.16181923e-02 -3.81989926e-01
-6.34496629e-01 3.41294140e-01 -4.67685550e-01 1.05851278e-01
2.21353054e-01 -2.48371780e-01 -8.31731796e-01 5.63900173e-01
2.21821114e-01 8.89974058e-01 1.97725371e-01 -4.02485609e-01
-5.38859427e-01 3.40917945e-01 2.96605229e-01 -3.10343295e-01
1.60597026e+00 3.87587138e-02 -3.83134693e-01 5.24004161e-01
8.24290872e-01 -1.34801582e-01 -2.81318575e-01 4.76498276e-01
6.76181555e-01 1.43765360e-01 5.05775809e-01 -1.16973186e+00
-6.08990312e-01 2.47895837e-01 6.14671707e-01 4.79630470e-01
7.67435968e-01 1.21112823e-01 3.65045607e-01 5.47807038e-01
2.63252556e-02 -1.65934825e+00 -4.98117179e-01 1.15806110e-01
9.90809202e-01 -7.66904473e-01 4.19790953e-01 -2.69941270e-01
-4.79424745e-01 9.76960897e-01 6.70718541e-03 3.19459200e-01
9.29564178e-01 2.46618629e-01 -5.04554451e-01 -9.89197135e-01
-4.87766653e-01 -1.57744363e-01 1.06644645e-01 5.08494735e-01
4.27757531e-01 2.74497986e-01 -1.06902981e+00 9.22326624e-01
-9.17396247e-02 2.27356404e-01 2.78471828e-01 1.11339474e+00
-7.08601534e-01 -1.04132533e+00 -4.20652330e-01 3.99904698e-01
-5.87043643e-01 -1.42319664e-01 -6.55119538e-01 8.58897269e-01
1.96274042e-01 1.05595505e+00 -2.75360823e-01 -7.26483241e-02
7.47880042e-01 2.51128614e-01 1.14841759e-01 -4.06546652e-01
-5.16079247e-01 -1.09627537e-01 6.07081413e-01 -9.04819965e-02
-4.17342395e-01 -6.99138224e-01 -1.77017605e+00 -1.80744708e-01
-3.02865684e-01 9.19578075e-01 1.58324063e+00 9.51742411e-01
7.16898143e-01 6.90059304e-01 2.54143625e-01 -3.20221275e-01
1.67476729e-01 -1.12514019e+00 -8.73427391e-01 2.21694529e-01
-6.82560980e-01 -6.74984694e-01 -3.16698104e-01 7.39321634e-02] | [9.497227668762207, 8.207371711730957] |
fe598a82-1bc9-4517-9007-3dad3e7687e5 | tonet-tone-octave-network-for-singing-melody | 2202.00951 | null | https://arxiv.org/abs/2202.00951v1 | https://arxiv.org/pdf/2202.00951v1.pdf | TONet: Tone-Octave Network for Singing Melody Extraction from Polyphonic Music | Singing melody extraction is an important problem in the field of music information retrieval. Existing methods typically rely on frequency-domain representations to estimate the sung frequencies. However, this design does not lead to human-level performance in the perception of melody information for both tone (pitch-class) and octave. In this paper, we propose TONet, a plug-and-play model that improves both tone and octave perceptions by leveraging a novel input representation and a novel network architecture. First, we present an improved input representation, the Tone-CFP, that explicitly groups harmonics via a rearrangement of frequency-bins. Second, we introduce an encoder-decoder architecture that is designed to obtain a salience feature map, a tone feature map, and an octave feature map. Third, we propose a tone-octave fusion mechanism to improve the final salience feature map. Experiments are done to verify the capability of TONet with various baseline backbone models. Our results show that tone-octave fusion with Tone-CFP can significantly improve the singing voice extraction performance across various datasets -- with substantial gains in octave and tone accuracy. | ['Shlomo Dubnov', 'Taylor Berg-Kirkpatrick', 'Wei Li', 'Cheng-i Wang', 'Shuai Yu', 'Ke Chen'] | 2022-02-02 | null | null | null | null | ['melody-extraction', 'music-information-retrieval'] | ['music', 'music'] | [ 2.31139049e-01 -5.44739902e-01 -2.01404884e-01 -2.49041408e-01
-1.08154213e+00 -6.23513162e-01 2.10250378e-01 2.28412207e-02
-9.65305120e-02 3.35760415e-01 5.88299930e-01 2.37809986e-01
-3.75343621e-01 -6.31625116e-01 -3.13498497e-01 -3.79828990e-01
2.84499768e-03 -2.84659058e-01 1.99350387e-01 -5.57495356e-01
3.23944509e-01 -9.32647586e-02 -1.77881253e+00 6.42761469e-01
5.92701256e-01 1.25489354e+00 2.10307509e-01 8.90871108e-01
1.07028536e-01 5.84082603e-01 -8.55118513e-01 -3.40301096e-02
4.73848879e-01 -9.34916735e-01 -4.52801824e-01 -3.53881627e-01
3.94124448e-01 -1.76231116e-01 -1.79952666e-01 9.58293319e-01
8.43442917e-01 2.89390326e-01 5.15396297e-01 -8.14531982e-01
-4.55363452e-01 1.17964876e+00 -2.95320868e-01 1.75275594e-01
4.93979067e-01 6.27737567e-02 1.49730194e+00 -9.84670341e-01
2.64506251e-01 1.01041067e+00 8.79085422e-01 2.30898052e-01
-1.15637612e+00 -7.23199010e-01 -2.42554873e-01 3.77182454e-01
-1.52711999e+00 -3.28762084e-01 1.32448435e+00 -3.23526800e-01
9.55948830e-01 6.69690907e-01 9.32268500e-01 5.94943583e-01
9.28586572e-02 5.81550062e-01 1.04597676e+00 -6.57325983e-01
8.35799333e-03 1.31230867e-02 -8.10998529e-02 1.52987182e-01
-2.32217640e-01 4.15496111e-01 -1.26784217e+00 1.92952633e-01
8.49679708e-01 -4.16435152e-01 -3.87567222e-01 -7.26906881e-02
-1.07430351e+00 5.84586740e-01 5.91271043e-01 4.84429985e-01
-4.09521341e-01 2.42056653e-01 5.00744879e-01 4.46963489e-01
2.23271713e-01 8.92171264e-01 -9.52491611e-02 -3.93174350e-01
-1.44280720e+00 5.83153009e-01 6.94164753e-01 3.62798154e-01
3.17665547e-01 6.01171792e-01 -5.24395406e-01 1.00072694e+00
-3.18042282e-03 3.58971655e-01 5.51064491e-01 -8.62139881e-01
3.17437232e-01 1.16588742e-01 -1.55347213e-01 -1.09345281e+00
-2.82518029e-01 -8.10777128e-01 -5.55858433e-01 1.84472829e-01
1.40444979e-01 3.23412120e-02 -6.08632743e-01 2.00238252e+00
2.74364986e-02 1.13561295e-01 -2.58347690e-01 1.46741068e+00
6.93968415e-01 6.83682024e-01 -2.17514142e-01 -1.63959697e-01
1.51957130e+00 -8.56115520e-01 -9.72111285e-01 5.67011237e-02
-2.10000038e-01 -9.91245925e-01 1.38897455e+00 6.76991045e-01
-1.45572090e+00 -1.09383249e+00 -1.54038763e+00 -2.81968564e-01
-2.11969957e-01 1.71483219e-01 4.97324884e-01 6.21961772e-01
-6.29528105e-01 9.11284626e-01 -2.62732774e-01 1.83094040e-01
-8.89005419e-03 3.39339644e-01 8.75062868e-02 7.44852066e-01
-1.44579709e+00 5.53543270e-01 4.63683158e-01 -6.44999295e-02
-7.96409786e-01 -1.00371826e+00 -7.15058863e-01 4.05253857e-01
9.31079984e-02 -5.69035351e-01 1.54352582e+00 -7.11282372e-01
-1.62918413e+00 3.14484239e-01 1.21548943e-01 -5.43773413e-01
-7.80787989e-02 -4.12980497e-01 -6.17202282e-01 1.94727853e-01
-6.03855029e-02 6.08891308e-01 9.50383306e-01 -1.00354147e+00
-6.38439059e-01 -8.11367929e-02 -1.03587054e-01 5.14790058e-01
-6.26979947e-01 -1.77827254e-01 -1.33162260e-01 -1.17395210e+00
2.22642437e-01 -6.29208267e-01 1.75234526e-01 -3.96347493e-01
-3.82199556e-01 5.63106947e-02 3.98569882e-01 -7.34433532e-01
2.07394719e+00 -2.41366935e+00 2.06101641e-01 1.81151658e-01
1.39318600e-01 1.64533556e-01 -9.03066546e-02 5.13733685e-01
-2.16788143e-01 -5.98863959e-02 -1.15224831e-01 -4.44314480e-02
3.18615437e-01 -3.22360307e-01 -5.83198845e-01 -7.11533353e-02
3.09522569e-01 8.53658020e-01 -6.39627278e-01 -2.90847540e-01
9.14939940e-02 5.39296150e-01 -1.08419824e+00 1.18066527e-01
-1.48633271e-01 2.28163823e-01 1.14238877e-02 4.74952459e-01
4.05979306e-01 3.04974467e-01 5.92203885e-02 -6.07693911e-01
-3.13900054e-01 1.00790226e+00 -1.27563989e+00 1.87248528e+00
-4.03033018e-01 6.21375203e-01 -1.53652951e-01 -2.93957531e-01
1.14548981e+00 5.36226273e-01 4.51200932e-01 -7.79048324e-01
2.26579919e-01 4.30664062e-01 4.77794468e-01 -1.91630706e-01
9.03449118e-01 -5.73081851e-01 -4.60456163e-01 3.33262444e-01
1.91933841e-01 -4.89762366e-01 1.95280671e-01 -4.12400812e-01
6.16702318e-01 9.40804183e-02 2.72017241e-01 -7.32527226e-02
3.31278980e-01 -2.07998186e-01 5.53955853e-01 5.65740347e-01
-9.45349187e-02 1.02846956e+00 3.68590117e-01 -1.68771088e-01
-1.04248405e+00 -1.13220763e+00 -1.33348897e-01 1.41912270e+00
-7.62043074e-02 -8.66853118e-01 -8.63983393e-01 -3.02889821e-04
-5.76242507e-02 6.44873619e-01 -3.33043694e-01 -4.11239356e-01
-5.91987550e-01 -3.27754915e-01 9.01637793e-01 5.47641814e-01
4.48216259e-01 -1.21070671e+00 -5.86166203e-01 3.96019459e-01
-4.12891120e-01 -6.10900998e-01 -9.88248467e-01 5.53289592e-01
-3.93214941e-01 -6.02088690e-01 -5.10368586e-01 -9.42806184e-01
-2.57742107e-01 4.26825918e-02 1.00574577e+00 -2.85933584e-01
-2.68469453e-01 -4.23455909e-02 -5.20910382e-01 -4.94314462e-01
-1.84908479e-01 2.74860978e-01 2.28671834e-01 -8.38671848e-02
4.29174215e-01 -1.01178074e+00 -7.42502391e-01 8.25926811e-02
-8.90057027e-01 3.51094678e-02 4.65092808e-01 7.67635643e-01
6.90285623e-01 2.85383165e-01 8.97409976e-01 -4.82802629e-01
1.20812845e+00 -2.40801368e-02 -4.66387868e-02 -2.18975112e-01
-4.87725973e-01 -2.67926723e-01 5.91324806e-01 -5.03887594e-01
-6.25033557e-01 -1.12940989e-01 -3.52788448e-01 -5.46145082e-01
1.33964285e-01 7.05360889e-01 -5.89275211e-02 2.06326500e-01
9.22442794e-01 3.05147260e-01 -2.11508110e-01 -7.97173738e-01
4.60945219e-01 9.86788750e-01 8.89429688e-01 -5.64964056e-01
7.57077217e-01 -8.05734396e-02 -2.32281163e-01 -6.19841874e-01
-9.40507472e-01 -5.01462340e-01 -5.31962574e-01 -2.48541892e-01
7.21033812e-01 -1.01434481e+00 -6.54769599e-01 2.16669202e-01
-8.14122915e-01 1.18732505e-01 -6.23045802e-01 6.65692925e-01
-7.13782728e-01 9.94374529e-02 -7.65228987e-01 -8.97363603e-01
-6.32025957e-01 -8.12703609e-01 9.07498837e-01 2.73455530e-01
-6.72241986e-01 -3.18237454e-01 1.07998148e-01 8.16819295e-02
6.74358130e-01 1.10922679e-01 9.56043839e-01 -4.19300914e-01
-2.44336829e-01 -5.30136265e-02 1.70653507e-01 4.23990786e-01
3.33711684e-01 -4.33598399e-01 -1.20836592e+00 -1.71067178e-01
2.13492334e-01 -4.02931601e-01 9.82764661e-01 4.16644305e-01
1.21533430e+00 -2.42114574e-01 4.15068924e-01 6.39864683e-01
1.06617010e+00 3.72933358e-01 5.62176526e-01 -3.85057461e-03
5.09201229e-01 3.46919477e-01 6.06346846e-01 6.02770865e-01
1.86661482e-01 1.02929544e+00 -1.84899326e-02 -5.70211709e-02
-7.36115158e-01 -7.63877034e-01 3.89598429e-01 1.48915935e+00
-7.17303306e-02 1.39809176e-01 -4.69082922e-01 6.79769516e-01
-1.58778548e+00 -1.15554750e+00 3.31757218e-01 2.26725769e+00
1.43486392e+00 3.34470749e-01 4.68549013e-01 7.91615903e-01
4.22908872e-01 2.72383660e-01 -4.49432045e-01 -6.38676226e-01
-1.35806993e-01 6.96379125e-01 -2.42561940e-02 4.35288996e-01
-1.04304850e+00 9.00254667e-01 6.72431040e+00 1.13319862e+00
-1.32790256e+00 -1.47960797e-01 1.45797431e-01 -4.35385913e-01
-4.71214980e-01 -1.18898399e-01 -6.02411985e-01 2.69673437e-01
8.18636656e-01 -2.98353136e-01 8.45442116e-01 6.02420390e-01
2.54149258e-01 3.40996534e-01 -1.08292568e+00 1.13146973e+00
2.72104800e-01 -1.13070583e+00 2.14539200e-01 -2.25277945e-01
4.80661064e-01 -5.63297272e-01 3.57951641e-01 4.95865494e-01
-4.29672509e-01 -1.04076111e+00 1.13589621e+00 5.24004877e-01
1.06908238e+00 -9.72765326e-01 3.00381452e-01 3.39579061e-02
-1.59773016e+00 -9.37023386e-02 -1.11688130e-01 -3.00969183e-01
1.09979138e-01 2.07863301e-01 -9.30297613e-01 5.05960166e-01
6.75667644e-01 5.26767135e-01 -3.07457209e-01 1.15625787e+00
-1.13916975e-02 7.34431684e-01 -2.06820935e-01 -1.11912657e-02
-1.68023556e-01 1.34563804e-01 6.84601605e-01 1.40665078e+00
3.64966005e-01 -5.12768738e-02 2.08529919e-01 1.06256604e+00
1.11691847e-01 1.70778200e-01 -2.75170118e-01 -2.56209731e-01
6.92411602e-01 1.08550978e+00 -3.08666497e-01 -1.66859686e-01
1.07749216e-02 6.46666884e-01 -6.16496429e-02 9.82918739e-02
-8.16221654e-01 -8.64327490e-01 8.10120225e-01 1.96469367e-01
5.74407220e-01 -1.82348952e-01 -4.44991201e-01 -7.13656902e-01
-1.71281025e-01 -1.26497257e+00 9.73034129e-02 -7.90631890e-01
-1.22164214e+00 8.09197009e-01 -1.35764211e-01 -1.67976832e+00
-6.69741929e-01 -4.49369222e-01 -4.23729956e-01 1.04327559e+00
-1.46024120e+00 -8.96585941e-01 6.45573810e-02 4.63154674e-01
7.27600932e-01 -2.11463720e-01 9.78564978e-01 5.44582427e-01
-1.20320618e-01 8.01440239e-01 -4.78976905e-01 5.38784117e-02
7.33053029e-01 -1.37510192e+00 1.82527304e-01 5.53405225e-01
7.46389687e-01 6.83271945e-01 7.38576353e-01 -3.39709312e-01
-1.16465747e+00 -8.24958205e-01 8.49995017e-01 -2.24982843e-01
4.62645888e-01 -4.65289384e-01 -7.26612151e-01 2.98516154e-02
3.51344824e-01 -4.92297590e-01 1.12498140e+00 2.41493911e-01
-4.29674745e-01 -3.90749097e-01 -6.29088521e-01 5.94103038e-01
6.99302316e-01 -1.11447668e+00 -8.79208684e-01 -3.91150892e-01
9.09445643e-01 -4.79436666e-01 -1.04889989e+00 5.74122488e-01
8.97638559e-01 -1.03401542e+00 1.08954942e+00 -3.59651953e-01
4.91970986e-01 -7.63406634e-01 -4.20951545e-01 -1.41978872e+00
-6.80450320e-01 -7.21374571e-01 -2.00571030e-01 1.14662671e+00
4.12413180e-01 1.12335615e-01 3.12915862e-01 -1.60443678e-01
-2.78056324e-01 -5.92569947e-01 -8.88466656e-01 -6.75336957e-01
-3.27980309e-03 -7.50533462e-01 8.49151075e-01 7.87443697e-01
4.89165217e-01 6.43721759e-01 -7.57948875e-01 -2.55567312e-01
2.94052988e-01 4.59892899e-01 5.33743978e-01 -9.97526765e-01
-8.06216538e-01 -7.91838527e-01 -2.32331678e-01 -1.01199532e+00
-5.36090970e-01 -9.33177888e-01 2.51600057e-01 -1.13331079e+00
2.71804053e-02 1.01825155e-01 -9.18131530e-01 4.31954831e-01
-8.44004750e-02 7.13670850e-01 6.66576445e-01 1.61339000e-01
-2.78781444e-01 6.80727661e-01 1.40469992e+00 3.49418190e-03
-5.00317812e-01 -1.50234280e-02 -8.62759054e-01 7.11331844e-01
8.83760691e-01 -3.12077612e-01 -4.22592044e-01 -2.36676618e-01
2.32762352e-01 2.40804389e-01 2.95650512e-01 -1.46414804e+00
2.71737486e-01 1.23423979e-01 5.03658593e-01 -8.01912248e-01
8.98976684e-01 -4.91073459e-01 9.31082666e-02 2.20815226e-01
-6.78417444e-01 -6.04469553e-02 3.61053348e-01 3.70480984e-01
-6.38559878e-01 -2.28071585e-02 4.65360284e-01 1.12645604e-01
-5.45295894e-01 -2.32866660e-01 -8.20342898e-02 -1.38268396e-01
3.46935630e-01 -2.35673100e-01 -1.21825144e-01 -4.66237992e-01
-7.46072471e-01 -3.05776507e-01 -1.50584534e-01 7.44550228e-01
7.20324337e-01 -1.77686214e+00 -7.60813415e-01 5.39981186e-01
1.33846581e-01 -6.95863664e-01 1.56128958e-01 5.77632785e-01
-2.48961642e-01 4.24032092e-01 -3.49571377e-01 -4.25016254e-01
-1.15616012e+00 2.85650522e-01 3.71654809e-01 -7.05052316e-02
-3.38206321e-01 7.97508121e-01 1.58674315e-01 -3.03899467e-01
3.51318240e-01 -6.23305380e-01 -3.33674699e-01 1.82293281e-01
5.75427115e-01 -4.61375341e-03 -1.02934293e-01 -5.55350006e-01
-1.04876161e-01 7.74671435e-01 1.58798248e-01 -4.72696722e-01
1.11195707e+00 -7.33576268e-02 1.26612410e-01 9.19972837e-01
8.50084305e-01 4.26410764e-01 -9.85048413e-01 -1.98965222e-01
-8.14812705e-02 -5.86233675e-01 2.51758814e-01 -1.11618316e+00
-7.32373834e-01 8.89293611e-01 5.66437006e-01 4.58771020e-01
1.58003688e+00 -3.99230540e-01 1.03905129e+00 -3.53296697e-02
1.39252841e-01 -1.22430921e+00 1.69499204e-01 4.94277418e-01
1.05988419e+00 -6.64742947e-01 -2.08723202e-01 -4.09111887e-01
-7.15441406e-01 1.03502727e+00 5.12631238e-01 -2.35906184e-01
7.46095598e-01 2.09848672e-01 2.37900630e-01 -5.24441190e-02
-7.54369318e-01 -5.57248890e-01 9.75572288e-01 2.09246695e-01
8.62782359e-01 2.12179095e-01 -4.57942009e-01 1.32666254e+00
-1.20344853e+00 -1.27400815e-01 1.10162735e-01 4.54717547e-01
-7.08338857e-01 -1.23140812e+00 -3.69093210e-01 1.69353411e-01
-4.91111070e-01 -4.57198977e-01 -5.37079513e-01 4.42838132e-01
4.30858225e-01 9.20705140e-01 7.63837546e-02 -1.09814417e+00
5.32903254e-01 2.28578284e-01 5.05876839e-01 -7.14296877e-01
-1.14976168e+00 7.71951973e-01 1.29936263e-02 -2.10293785e-01
-1.63220212e-01 -1.21532097e-01 -1.14546931e+00 -4.82622236e-02
-3.61002177e-01 3.51681888e-01 5.49280107e-01 5.13166606e-01
2.92692184e-01 1.21042216e+00 8.99103701e-01 -7.89008558e-01
-5.67777991e-01 -1.20833015e+00 -8.77334177e-01 3.53933454e-01
4.50790703e-01 -4.15198505e-01 -2.32871205e-01 7.25516751e-02] | [15.742176055908203, 5.442770957946777] |
3d35234c-5c2b-4bc1-8366-81b867bafa03 | channel-spatial-based-few-shot-bird-sounds | 2306.10499 | null | https://arxiv.org/abs/2306.10499v2 | https://arxiv.org/pdf/2306.10499v2.pdf | Channel-Spatial-Based Few-Shot Bird Sound Event Detection | In this paper, we propose a model for bird sound event detection that focuses on a small number of training samples within the everyday long-tail distribution. As a result, we investigate bird sound detection using the few-shot learning paradigm. By integrating channel and spatial attention mechanisms, improved feature representations can be learned from few-shot training datasets. We develop a Metric Channel-Spatial Network model by incorporating a Channel Spatial Squeeze-Excitation block into the prototype network, combining it with these attention mechanisms. We evaluate the Metric Channel Spatial Network model on the DCASE 2022 Take5 dataset benchmark, achieving an F-measure of 66.84% and a PSDS of 58.98%. Our experiment demonstrates that the combination of channel and spatial attention mechanisms effectively enhances the performance of bird sound classification and detection. | ['Chenlei Jin', 'Xin Pan', 'Yajie Yang', 'Haitao Fu', 'Yuxuan Feng', 'Lingwen Liu'] | 2023-06-18 | null | null | null | null | ['sound-event-detection', 'sound-classification', 'few-shot-learning'] | ['audio', 'audio', 'methodology'] | [-1.60474643e-01 -6.24008894e-01 -3.59048918e-02 -2.34130114e-01
-8.01841319e-01 -3.06509018e-01 3.68602455e-01 3.37503999e-02
-4.90373194e-01 2.87436724e-01 2.46383533e-01 -5.37035242e-02
-3.04676834e-02 -8.58339369e-01 -5.72677195e-01 -5.47694087e-01
-4.19576317e-01 -5.94591856e-01 6.84919178e-01 1.81715470e-02
6.22186139e-02 2.22706705e-01 -1.74459743e+00 1.31800532e-01
5.60589373e-01 1.35841393e+00 3.18779498e-01 1.17785251e+00
-6.64056605e-03 7.74521589e-01 -9.48794007e-01 6.90168738e-02
2.53333420e-01 -5.79338193e-01 -1.96462855e-01 -7.55265713e-01
2.46864691e-01 -5.85918784e-01 -4.35086489e-01 9.48423028e-01
1.13591278e+00 6.13044500e-01 5.88108957e-01 -1.09190810e+00
-6.46346927e-01 8.69080067e-01 -5.77050984e-01 9.41312432e-01
1.55136570e-01 3.80078495e-01 1.38048136e+00 -1.06288755e+00
-2.14610979e-01 8.07501554e-01 8.09715390e-01 5.33052802e-01
-8.67936850e-01 -1.13972926e+00 -2.44331341e-02 4.23181176e-01
-1.47763610e+00 -3.82563919e-01 7.93214321e-01 -3.01704019e-01
1.26857960e+00 1.72194555e-01 5.85256696e-01 7.83010542e-01
8.22781306e-03 6.20001018e-01 5.15217483e-01 -4.82534677e-01
3.31517607e-01 -8.78080949e-02 3.54234517e-01 4.71385032e-01
-3.24694999e-02 7.92756081e-01 -8.62093508e-01 -7.57884011e-02
5.14679730e-01 1.79083779e-01 -2.29643092e-01 5.55772245e-01
-6.11246288e-01 8.27973723e-01 9.37762678e-01 3.54948640e-01
-2.66533941e-01 5.28704166e-01 4.29794163e-01 9.42799821e-02
5.41414440e-01 6.06499791e-01 -2.47743949e-01 -2.67282963e-01
-1.01890540e+00 3.86537239e-02 6.53900325e-01 9.01510417e-01
5.52233458e-01 3.89593124e-01 -7.76177049e-01 9.60822105e-01
2.27123797e-01 7.67551422e-01 3.29169542e-01 -4.19147223e-01
1.68365508e-01 1.06361158e-01 -1.54624447e-01 -8.28042626e-01
-4.17513520e-01 -8.27022910e-01 -6.37134016e-01 -2.16211081e-01
-7.09862635e-02 -3.94547820e-01 -8.37952018e-01 1.79257703e+00
5.93455397e-02 9.24205542e-01 -2.59212106e-01 8.99965167e-01
1.21073115e+00 1.08389080e+00 2.11997852e-01 -1.34158656e-01
1.25126362e+00 -1.17609668e+00 -6.22355282e-01 -7.72669762e-02
2.16358542e-01 -4.79522973e-01 1.28824651e+00 1.35301888e-01
-7.77371109e-01 -8.37719142e-01 -1.23388731e+00 1.24300621e-01
-4.37150091e-01 1.44101093e-02 4.62196112e-01 7.55166650e-01
-7.03501046e-01 4.91557389e-01 -6.06739402e-01 -6.56379387e-02
6.50372446e-01 -3.36415470e-02 4.01623398e-01 5.55614978e-02
-1.30848408e+00 1.78595316e-02 7.07286596e-02 -1.40551746e-01
-1.28049421e+00 -1.02841520e+00 -6.06282830e-01 6.75610840e-01
4.24694046e-02 -3.57768536e-01 1.45972359e+00 -1.72639430e-01
-1.32624435e+00 1.17016234e-01 2.02317342e-01 -7.58106709e-01
-7.52317011e-02 -3.56788963e-01 -7.34833062e-01 2.58401752e-01
4.66242880e-02 5.88354409e-01 8.80905271e-01 -6.92626655e-01
-9.56262589e-01 1.83896810e-01 -5.44515625e-02 -9.47757214e-02
-6.60457194e-01 3.52302313e-01 -1.54121757e-01 -8.76616955e-01
-4.89544362e-01 -4.06954944e-01 -1.38639286e-01 6.49430836e-03
-1.84159920e-01 -1.99896157e-01 6.17784739e-01 -3.54697138e-01
1.79231381e+00 -2.58322668e+00 -6.11439705e-01 1.93827786e-02
1.97427616e-01 4.30861115e-01 -3.08032632e-01 2.91171402e-01
2.04765230e-01 7.50288658e-04 -2.97246650e-02 -2.81505734e-01
1.64283305e-01 -3.14952403e-01 -6.50783420e-01 1.28178433e-01
2.05073684e-01 8.02546382e-01 -1.02409852e+00 -2.07362384e-01
1.22631360e-02 4.09940928e-01 -9.09680486e-01 5.06051242e-01
2.61462685e-02 -1.39389411e-01 -1.84795812e-01 6.40489459e-01
6.89327598e-01 -2.27585901e-02 -7.47262657e-01 1.17343433e-01
-2.09279060e-01 3.24376196e-01 -8.82466137e-01 1.62167859e+00
-5.06218374e-01 6.79564595e-01 -3.48801851e-01 -4.12716120e-01
1.11777711e+00 2.52523750e-01 1.63125053e-01 -9.61148143e-01
2.09421262e-01 6.92671537e-02 1.18255429e-01 -3.03033948e-01
4.00573343e-01 -3.95214528e-01 -1.97860628e-01 5.46892226e-01
4.77883548e-01 1.10596400e-02 -4.77001444e-02 2.05691516e-01
1.26083601e+00 -5.06673694e-01 1.95911989e-01 -4.49311584e-02
1.12323821e-01 -6.69980168e-01 4.29162472e-01 1.30654895e+00
-4.30638224e-01 5.61366260e-01 1.60223022e-01 -2.75179952e-01
-5.72826028e-01 -1.07806206e+00 -3.16039592e-01 1.77925718e+00
8.87848586e-02 -7.97457099e-01 -6.80322349e-01 -5.90897083e-01
-3.75601165e-02 8.83043170e-01 -5.96846879e-01 -8.15017998e-01
-1.81001201e-01 -8.90802860e-01 1.00316465e+00 9.43716049e-01
6.14075005e-01 -1.16660452e+00 -9.98020113e-01 3.52747500e-01
2.67087847e-01 -7.07529247e-01 -6.36552632e-01 6.69116795e-01
-2.19091490e-01 -5.60740292e-01 -5.83382726e-01 -7.50397921e-01
-6.83643818e-02 4.98270422e-01 9.15485799e-01 -5.88172756e-04
-5.51455081e-01 3.54083955e-01 -6.30350173e-01 -8.38318408e-01
1.55152917e-01 -3.35187800e-02 5.85796162e-02 -1.52443662e-01
4.03799236e-01 -6.11131847e-01 -8.00639033e-01 1.99390575e-01
-6.41135156e-01 -4.28542435e-01 3.35725754e-01 1.02019274e+00
2.56498218e-01 5.81463128e-02 1.05931151e+00 -2.66801387e-01
7.57134914e-01 -5.76922715e-01 -3.74305576e-01 -5.30710109e-02
-3.43118370e-01 -4.32832301e-01 7.43095517e-01 -5.80061376e-01
-8.42379332e-01 -1.78680867e-01 -4.70744610e-01 -4.54920381e-01
-1.37591273e-01 2.69831926e-01 8.46261457e-02 -6.42330796e-02
8.17341268e-01 2.88626343e-01 -3.80805135e-01 -5.25709748e-01
3.58948976e-01 8.84140849e-01 4.51151103e-01 -2.75495291e-01
6.37927353e-01 5.11769615e-02 -4.42530811e-01 -9.91145194e-01
-9.10294414e-01 -7.62483597e-01 -1.66823491e-01 -2.71276265e-01
7.17095017e-01 -9.58382607e-01 -4.71462220e-01 4.60494906e-01
-9.58038509e-01 -3.27927232e-01 -5.00163436e-01 7.68262446e-01
-3.30112070e-01 -2.54166067e-01 -7.54171669e-01 -9.29724216e-01
-6.41112328e-01 -6.15927041e-01 8.89565051e-01 5.19909561e-01
1.18808210e-01 -4.81548846e-01 3.57947230e-01 -5.21029055e-01
8.12402070e-01 -2.88014203e-01 6.02270782e-01 -7.68222749e-01
-4.58829433e-01 -3.19462419e-01 -3.42347980e-01 4.37813312e-01
-2.41445228e-02 -1.36053920e-01 -1.54105902e+00 -1.19531348e-01
-3.23165692e-02 -9.37178358e-02 1.38266754e+00 5.68757534e-01
1.36882997e+00 -1.32532462e-01 -1.29117996e-01 8.16241324e-01
1.26728690e+00 5.10538697e-01 4.19829488e-01 -2.55450189e-01
4.79462713e-01 1.43989921e-02 5.28548479e-01 7.56973267e-01
-3.45199518e-02 3.98450404e-01 4.00522530e-01 3.74870636e-02
-3.95451635e-01 -5.73456645e-01 1.49203762e-01 7.98065960e-01
1.61082521e-01 -4.09905046e-01 -7.99035013e-01 5.57575583e-01
-1.48397219e+00 -1.28259885e+00 2.69581139e-01 1.98798192e+00
7.08933175e-01 1.94732219e-01 2.87751287e-01 2.36682534e-01
7.04835653e-01 3.70071739e-01 -3.55220824e-01 -2.71261007e-01
1.81315631e-01 4.06110227e-01 2.20256343e-01 2.32005537e-01
-1.34864378e+00 8.81638408e-01 6.92835283e+00 1.17022145e+00
-1.15222216e+00 2.07665920e-01 3.25146675e-01 -4.59371269e-01
1.11246027e-01 -5.37442923e-01 -1.09957325e+00 7.39624977e-01
1.24015033e+00 -3.45814258e-01 5.26036203e-01 8.27393413e-01
3.51412073e-02 1.97041944e-01 -8.76662552e-01 1.18236065e+00
1.43054217e-01 -1.24997997e+00 -2.14484185e-01 -1.53796107e-01
5.78956068e-01 1.67568296e-01 3.43802452e-01 7.94418514e-01
-4.21038941e-02 -1.01809454e+00 6.46541595e-01 5.90839624e-01
9.56856549e-01 -8.58633339e-01 6.00718856e-01 3.86247963e-01
-1.64411914e+00 -5.94508052e-01 -6.74769998e-01 -2.43988618e-01
1.29966974e-01 4.93334800e-01 -8.22136998e-01 1.01965375e-01
9.36266184e-01 6.74764454e-01 -6.05903685e-01 1.82854211e+00
-8.12033564e-02 1.32947767e+00 -5.04823565e-01 -5.94747543e-01
2.61418879e-01 4.78434563e-01 5.96776783e-01 1.52406204e+00
4.24963087e-01 1.67849645e-01 -1.93872731e-02 1.07413125e+00
-2.30690196e-01 1.29122570e-01 -6.06908679e-01 5.06340563e-02
7.64134765e-01 1.22721183e+00 -5.70443094e-01 -2.41316319e-01
-5.37990749e-01 4.69957083e-01 6.38549179e-02 2.97155261e-01
-1.24496031e+00 -1.03021348e+00 5.30541897e-01 -3.50046277e-01
7.74649143e-01 2.47215077e-01 -4.95535433e-02 -6.43102169e-01
-4.40190494e-01 -3.75863373e-01 2.77213782e-01 -5.53300858e-01
-1.40714657e+00 5.95607400e-01 -1.39514714e-01 -1.38498569e+00
1.50655404e-01 -3.15543860e-01 -1.21599960e+00 6.67075038e-01
-1.56063271e+00 -8.51305187e-01 -2.97080338e-01 3.94388199e-01
3.98533314e-01 -3.29806358e-01 8.15266252e-01 5.84040701e-01
-6.37317121e-01 1.06723046e+00 6.33607805e-02 1.69208914e-01
5.20529449e-01 -1.30200005e+00 5.61061740e-01 8.08132052e-01
5.91317832e-01 4.88168299e-01 3.93414915e-01 -2.65902132e-01
-8.87273669e-01 -1.48131251e+00 4.08172041e-01 -3.44269983e-02
5.96265554e-01 -5.52782416e-01 -9.81137156e-01 1.76471248e-01
8.47969204e-02 5.59441566e-01 9.20012474e-01 1.37367174e-01
-4.71853048e-01 -4.21876699e-01 -8.99965644e-01 1.01132706e-01
9.85563338e-01 -7.85998404e-01 -5.37175536e-01 5.26151732e-02
1.04061997e+00 4.46484089e-02 -5.21342337e-01 1.93371221e-01
5.17942548e-01 -7.25672543e-01 1.11066031e+00 -5.35956681e-01
9.53471661e-02 -2.57753491e-01 -3.52223158e-01 -1.48503828e+00
-5.82318068e-01 -4.82329518e-01 -3.42901498e-01 1.39761376e+00
4.17561620e-01 -3.22640628e-01 1.64039746e-01 -1.81329802e-01
-2.90835232e-01 -6.25137031e-01 -9.75214660e-01 -9.41934109e-01
1.22462492e-02 -8.13109875e-01 7.69365251e-01 3.74084353e-01
-6.46760538e-02 5.31268895e-01 -6.02990627e-01 1.68429866e-01
3.85153085e-01 9.41167995e-02 5.05921543e-01 -1.17364037e+00
-5.17166972e-01 -4.81209785e-01 -2.59864241e-01 -1.36800373e+00
-1.09066278e-01 -8.00842941e-01 5.93662262e-01 -1.01595402e+00
2.20252737e-01 -1.24205627e-01 -1.11599207e+00 3.67351413e-01
-2.67944276e-01 3.61525983e-01 9.80276391e-02 -2.01849669e-01
-8.54808211e-01 8.12863410e-01 7.99864292e-01 -2.26161368e-02
-2.32158005e-01 2.54385412e-01 -7.11713433e-01 5.88653147e-01
8.81371796e-01 -6.91761434e-01 -3.18483233e-01 -1.25521228e-01
-6.40090257e-02 -7.73314834e-02 4.95970547e-01 -1.42825413e+00
6.13781512e-01 1.86660528e-01 3.11887592e-01 -6.87524498e-01
2.69381016e-01 -4.92042810e-01 -5.65949798e-01 3.87925744e-01
-5.59833109e-01 -4.80198830e-01 4.71944511e-01 1.02064621e+00
-1.24952279e-01 -6.63959980e-02 8.25016439e-01 1.44386292e-01
-8.34948123e-01 3.66192579e-01 -2.29278758e-01 2.59524822e-01
9.38532531e-01 1.85951039e-01 -5.14587283e-01 -2.14417905e-01
-3.10064316e-01 6.81601763e-02 -3.53615016e-01 4.38742042e-01
8.03436100e-01 -1.55658650e+00 -4.62426692e-01 6.43513441e-01
3.16396236e-01 -4.01533872e-01 5.46326458e-01 5.63661933e-01
-2.79429466e-01 3.27404737e-01 -1.11335441e-01 -5.41832507e-01
-1.00180686e+00 4.91423637e-01 4.65092331e-01 3.92280638e-01
-5.67967057e-01 1.38176501e+00 2.68155068e-01 -1.77278295e-01
6.03068888e-01 -6.21683598e-01 -3.08274239e-01 -6.81939572e-02
1.14433336e+00 6.43927813e-01 -2.99153179e-01 -1.92475155e-01
-4.06462789e-01 4.05424982e-01 1.98724195e-01 -1.73692144e-02
1.39146125e+00 7.05094635e-02 3.78563195e-01 6.04503810e-01
1.15481174e+00 8.32324997e-02 -1.21502364e+00 -2.89475620e-01
-3.82584453e-01 -5.31767488e-01 4.29424316e-01 -1.04515052e+00
-9.82133389e-01 1.36155844e+00 9.31048453e-01 5.56839466e-01
1.24563539e+00 2.05672696e-01 9.34641480e-01 3.52893084e-01
4.68124785e-02 -9.83669639e-01 3.43745679e-01 6.73323572e-01
6.82155490e-01 -1.09537053e+00 -4.52884555e-01 8.55688527e-02
-3.57389092e-01 9.13739920e-01 7.18258619e-01 -2.97025472e-01
1.06182575e+00 5.98853290e-01 -2.91286260e-01 -1.57149732e-01
-1.02060819e+00 -4.72404987e-01 4.42428112e-01 4.45922673e-01
1.57441735e-01 2.31780231e-01 9.57200453e-02 1.34690750e+00
-1.85098842e-01 -2.30234817e-01 1.30308330e-01 7.42750525e-01
-1.00521564e+00 -3.38104337e-01 -4.88279425e-02 6.19522631e-01
-4.26730961e-01 -5.57567060e-01 -8.39861184e-02 3.02478075e-01
1.54899150e-01 1.08438408e+00 3.28012586e-01 -8.34013700e-01
4.33640599e-01 2.31197681e-02 5.01031801e-02 -8.37555289e-01
-7.30740249e-01 4.65578318e-01 -2.88504094e-01 -2.72347301e-01
3.46340686e-02 -2.28803977e-01 -1.04052413e+00 1.39796343e-02
-8.28708231e-01 2.38998622e-01 2.84529835e-01 4.89613175e-01
4.42409158e-01 1.05668879e+00 9.37335968e-01 -6.81678176e-01
-7.47693002e-01 -1.07993364e+00 -1.04363811e+00 -1.27237171e-01
6.43823147e-01 -6.29373431e-01 -7.60461211e-01 -3.77536058e-01] | [15.159820556640625, 5.2250165939331055] |
093bf689-a80f-4fe8-b18d-48b6a26d5de1 | rule-of-thumb-deep-derotation-for-improved | 1507.05726 | null | http://arxiv.org/abs/1507.05726v1 | http://arxiv.org/pdf/1507.05726v1.pdf | Rule Of Thumb: Deep derotation for improved fingertip detection | We investigate a novel global orientation regression approach for articulated
objects using a deep convolutional neural network. This is integrated with an
in-plane image derotation scheme, DeROT, to tackle the problem of per-frame
fingertip detection in depth images. The method reduces the complexity of
learning in the space of articulated poses which is demonstrated by using two
distinct state-of-the-art learning based hand pose estimation methods applied
to fingertip detection. Significant classification improvements are shown over
the baseline implementation. Our framework involves no tracking, kinematic
constraints or explicit prior model of the articulated object in hand. To
support our approach we also describe a new pipeline for high accuracy magnetic
annotation and labeling of objects imaged by a depth camera. | ['Aaron Wetzler', 'Ron Slossberg', 'Ron Kimmel'] | 2015-07-21 | null | null | null | null | ['fingertip-detection'] | ['computer-vision'] | [ 2.79645231e-02 1.43423185e-01 1.23277634e-01 -1.82128206e-01
-7.02123702e-01 -8.08626294e-01 4.33562130e-01 -5.54957509e-01
-7.89728999e-01 3.78176779e-01 -1.59719978e-02 2.56079137e-01
-7.16262236e-02 -1.26730904e-01 -6.43219292e-01 -6.33243978e-01
2.21207246e-01 1.18223703e+00 5.40784359e-01 1.38940178e-02
4.04031396e-01 1.21323180e+00 -1.19437659e+00 1.35674834e-01
-3.62030178e-01 8.88252795e-01 2.79789507e-01 1.02720404e+00
3.69029194e-01 3.69879782e-01 -6.01678908e-01 -1.90997675e-01
4.42234427e-01 3.14455628e-01 -1.03172314e+00 2.36527234e-01
1.12814629e+00 -9.59260643e-01 -5.29669464e-01 4.76460040e-01
1.22129989e+00 -3.26528400e-01 5.51945269e-01 -7.79428661e-01
1.70390368e-01 2.05879420e-01 -4.21753734e-01 -4.27671298e-02
5.81522703e-01 1.85272157e-01 5.52261293e-01 -1.09730089e+00
1.05421209e+00 1.32887268e+00 9.93472815e-01 5.62041581e-01
-9.99437988e-01 -2.22399920e-01 5.57947010e-02 -1.51590303e-01
-1.20086074e+00 -2.30482891e-01 7.39117622e-01 -6.70497775e-01
1.17182922e+00 2.93238834e-02 7.26811111e-01 1.38638568e+00
2.76585463e-02 8.04618418e-01 1.02205527e+00 -7.40165472e-01
-1.25649691e-01 -2.95849770e-01 1.45097924e-02 1.10532820e+00
8.16037133e-02 1.02590598e-01 -5.52807570e-01 1.82620455e-02
1.50589776e+00 -2.43180692e-01 1.72595680e-01 -7.04256177e-01
-1.34436536e+00 2.37364501e-01 2.21324310e-01 5.37960883e-03
-5.51002264e-01 6.15096450e-01 4.55854535e-01 -3.51980329e-01
2.10191742e-01 1.74809933e-01 -6.74651563e-01 -1.80147305e-01
-8.33442688e-01 6.07290685e-01 8.12389553e-01 9.39326584e-01
7.03058243e-02 -2.31345832e-01 -3.15878928e-01 3.15098703e-01
6.17882907e-01 4.54685003e-01 -3.00368935e-01 -1.42793810e+00
3.86532456e-01 2.05165207e-01 3.21781039e-01 -4.18761611e-01
-8.96282732e-01 -2.62980193e-01 -8.20104182e-02 7.72360921e-01
9.14752543e-01 -1.88923061e-01 -8.78207326e-01 1.23909926e+00
5.66187143e-01 -5.00360727e-01 -5.75809717e-01 1.23561370e+00
4.56652641e-01 -3.82629931e-01 -3.46088767e-01 3.83327752e-01
1.48680854e+00 -8.90651941e-01 -5.57486773e-01 4.44211476e-02
-2.69800914e-03 -1.00306046e+00 7.39848673e-01 8.80423605e-01
-1.18326807e+00 -4.48543280e-01 -7.73095965e-01 -2.82889396e-01
-3.24130118e-01 5.89536905e-01 8.15498829e-01 8.10571492e-01
-7.86358297e-01 5.61030149e-01 -1.03983259e+00 -3.93559307e-01
5.67893267e-01 8.81123245e-01 -3.82032156e-01 4.15098250e-01
-4.14908171e-01 1.05682516e+00 1.97245538e-01 2.07023844e-01
-8.34182560e-01 -3.79560411e-01 -4.09019589e-01 -4.42340970e-01
3.97980362e-01 -7.51940370e-01 1.60682809e+00 -3.14518988e-01
-1.79410040e+00 1.26842773e+00 -7.58982897e-02 2.92361584e-02
1.19766569e+00 -7.76367903e-01 4.22389001e-01 5.91595411e-01
-2.50790775e-01 8.00909281e-01 1.19613099e+00 -1.18641961e+00
-3.39549512e-01 -7.91839540e-01 1.35599419e-01 7.20862299e-02
6.21429831e-02 2.45458916e-01 -5.79321563e-01 -7.34209597e-01
2.91999221e-01 -1.17506456e+00 1.17171325e-01 4.96243685e-01
-4.65543687e-01 -4.21270281e-01 9.73024368e-01 -9.95803177e-01
4.69040871e-01 -1.51403952e+00 4.40703452e-01 3.01961958e-01
2.80426830e-01 5.13175726e-02 1.25019953e-01 8.46510828e-02
1.78945124e-01 -5.61780691e-01 2.13182375e-01 -3.92003179e-01
1.58637032e-01 9.64402780e-02 2.11688690e-02 6.62163615e-01
8.21864605e-02 7.99914479e-01 -6.72686338e-01 -4.88851964e-01
3.71502906e-01 8.51966202e-01 -4.90465909e-01 2.29714870e-01
-2.32814372e-01 7.77040660e-01 -1.07815430e-01 1.09336615e+00
5.80750167e-01 1.04482494e-01 2.41584137e-01 -6.46891177e-01
-2.94185191e-01 3.11072409e-01 -1.44739652e+00 2.18475008e+00
-1.73193082e-01 4.25523639e-01 3.39029551e-01 -6.40609860e-01
4.89119112e-01 6.46374106e-01 6.34038627e-01 2.74834186e-02
4.25064415e-01 2.81582415e-01 -5.78614995e-02 -6.46906614e-01
1.34899765e-02 3.71504351e-02 2.98718095e-01 4.09736365e-01
3.52097958e-01 -4.18934564e-04 -2.66422689e-01 -2.41177082e-01
7.56361842e-01 1.09063196e+00 -1.91907644e-01 -2.51619697e-01
3.63264829e-01 -2.07994543e-02 -1.02257751e-01 7.27996707e-01
-2.04488888e-01 1.05001533e+00 3.73815089e-01 -6.38678789e-01
-1.35179794e+00 -1.04354560e+00 -2.65974760e-01 1.05834687e+00
-2.91211665e-01 1.25037469e-02 -1.08449471e+00 -7.13885427e-01
3.92533720e-01 -3.26124191e-01 -5.79434335e-01 6.55682147e-01
-1.03189647e+00 -3.00338387e-01 7.69180894e-01 9.91412401e-01
3.17812234e-01 -1.04738212e+00 -1.27223229e+00 2.10563675e-01
1.16527319e-01 -1.38056052e+00 -2.72506237e-01 2.07713336e-01
-7.82455444e-01 -1.39162385e+00 -1.13228655e+00 -8.37074816e-01
3.94320011e-01 -4.43806112e-01 9.02787745e-01 -5.13829052e-01
-9.27768826e-01 1.03045642e+00 4.01227316e-03 -6.47121429e-01
1.64323330e-01 3.69362295e-01 3.78083855e-01 -4.66773629e-01
1.75647393e-01 -3.66023004e-01 -7.55019248e-01 2.10979998e-01
-9.41566452e-02 -2.79759169e-01 4.94471878e-01 5.36331058e-01
4.14646477e-01 -8.05692017e-01 6.69671893e-02 -9.44976807e-02
3.15269917e-01 5.30466080e-01 -5.69591761e-01 1.65002108e-01
-1.76232919e-01 2.20296636e-01 -2.42725134e-01 -5.07456779e-01
-9.47505534e-01 8.37481260e-01 -4.34521943e-01 -5.55876553e-01
-2.97190607e-01 -2.09746316e-01 -2.56982110e-02 -5.97133338e-01
4.66492027e-01 -2.54362583e-01 3.87350351e-01 -8.85696411e-01
2.55365044e-01 6.43551171e-01 8.92775238e-01 -8.01771164e-01
3.98496896e-01 7.87733853e-01 5.75163960e-01 -8.03068399e-01
-7.03804314e-01 -3.79957199e-01 -1.75716746e+00 -5.14640272e-01
9.74065363e-01 -5.35408974e-01 -1.37619853e+00 8.68456721e-01
-1.69426239e+00 -2.60885596e-01 -2.60771289e-02 6.02484226e-01
-8.50477636e-01 5.97899735e-01 -5.17387271e-01 -1.07014251e+00
-5.32912552e-01 -1.14249337e+00 1.71752465e+00 -1.39537364e-01
-4.14943099e-01 -7.41482139e-01 -6.12480640e-02 2.12626517e-01
-8.38055555e-03 4.70523000e-01 4.30056214e-01 -2.72434592e-01
-6.52840316e-01 -5.19020975e-01 -1.34281188e-01 1.14054523e-01
-8.83102417e-02 -4.73856628e-02 -1.22142863e+00 -3.41278076e-01
-3.90163481e-01 -4.41278040e-01 4.14690495e-01 5.48163295e-01
1.27730119e+00 2.87246276e-02 -2.85189778e-01 5.28241992e-01
1.07026052e+00 -2.67585248e-01 4.38957363e-01 5.89214087e-01
1.03656638e+00 7.94957399e-01 5.00519753e-01 5.83943367e-01
7.49731883e-02 1.09698701e+00 6.05504334e-01 9.99485254e-02
-4.88680840e-01 3.10683995e-01 3.15052979e-02 9.38409790e-02
-1.06240427e+00 4.60139900e-01 -1.08689702e+00 3.15427035e-01
-1.57186508e+00 -6.17743254e-01 -1.62947863e-01 2.01835752e+00
6.71591520e-01 1.03225648e-01 7.48677969e-01 3.42388541e-01
4.15501535e-01 -4.07280713e-01 -5.52953839e-01 3.93246859e-02
1.72204897e-01 7.30065107e-01 6.28158689e-01 6.19256020e-01
-1.29795635e+00 9.88067627e-01 7.33037329e+00 1.20516364e-02
-1.02224541e+00 2.67238259e-01 -3.58099312e-01 -1.89892486e-01
7.07694292e-01 -3.52458835e-01 -1.21652758e+00 -4.63839620e-02
1.68003216e-01 1.04142904e+00 4.27810729e-01 8.00406516e-01
3.22596431e-02 6.34848094e-03 -1.30065405e+00 1.09567320e+00
2.90491939e-01 -1.03833008e+00 -2.65854418e-01 1.15007669e-01
1.67610440e-02 1.97455183e-01 -1.54839039e-01 -3.36131394e-01
-3.20975929e-01 -7.49472201e-01 9.54914927e-01 8.74216199e-01
9.42173004e-01 -5.44669926e-01 3.33986700e-01 1.02239348e-01
-1.17997026e+00 1.08946607e-01 -1.24818429e-01 -6.27591908e-02
1.06473662e-01 1.81330312e-02 -1.02884603e+00 2.03915328e-01
9.28396881e-01 1.04521245e-01 -5.15629709e-01 7.23281384e-01
-1.88484743e-01 5.46989255e-02 -5.11706233e-01 2.46940926e-01
1.78458374e-02 2.81250209e-01 7.61333764e-01 1.43827963e+00
-1.32095501e-01 -1.77352130e-01 1.40130043e-01 5.99883735e-01
2.36903429e-01 -5.82950473e-01 -1.93610132e-01 1.41815811e-01
3.05823758e-02 1.44724381e+00 -6.93942368e-01 -2.46636122e-01
-3.61709669e-02 1.32347810e+00 8.77278075e-02 1.82016566e-02
-4.22503263e-01 -4.51376379e-01 4.13498461e-01 3.93544316e-01
4.23701823e-01 -7.26269841e-01 -3.97705346e-01 -1.01744699e+00
3.85772914e-01 -5.40437460e-01 -3.99891734e-02 -8.53734016e-01
-9.50570822e-01 1.32244661e-01 1.02455877e-01 -7.25115001e-01
-4.44603026e-01 -1.52083600e+00 -1.50065839e-01 1.06707251e+00
-1.12361813e+00 -1.56125617e+00 -7.31289089e-01 6.80210650e-01
5.40216684e-01 -2.90698469e-01 1.00366163e+00 1.03039473e-01
-9.55298841e-02 4.89775389e-01 -3.13848823e-01 4.32893157e-01
9.21098590e-01 -1.60146856e+00 3.47081751e-01 3.38203847e-01
-1.60465300e-01 6.99169040e-01 4.73426759e-01 -5.62369883e-01
-1.69358194e+00 -5.33767641e-01 5.78213573e-01 -1.20611525e+00
2.27029055e-01 -5.45660257e-01 -3.30412447e-01 9.02942419e-01
5.61603494e-02 1.81761369e-01 1.95542529e-01 2.24101841e-02
-3.46686006e-01 2.46372938e-01 -1.26932383e+00 1.25820771e-01
1.29043555e+00 -6.90019906e-01 -7.54625320e-01 7.87166834e-01
7.37730414e-02 -9.60303605e-01 -1.13302827e+00 4.83485967e-01
1.60155463e+00 -6.78530812e-01 1.44115579e+00 -7.46565878e-01
7.87368491e-02 5.29406127e-03 1.89011469e-02 -4.43402201e-01
-2.58341610e-01 -5.50420165e-01 -7.22033918e-01 8.00688982e-01
-3.72886568e-01 -1.86154678e-01 1.21514452e+00 2.71220356e-01
3.07329856e-02 -7.02417135e-01 -1.09239256e+00 -7.92775691e-01
-6.21259287e-02 -3.01687360e-01 -9.25200880e-02 3.37360531e-01
-2.05228195e-01 4.25412580e-02 -1.62345201e-01 2.94990569e-01
1.15073872e+00 -1.94618851e-01 8.33759248e-01 -1.48954642e+00
-3.75960022e-01 -3.04685682e-01 -6.53945804e-01 -9.64473069e-01
6.00696616e-02 -6.48457229e-01 -1.90870136e-01 -1.52094054e+00
3.20246033e-02 -5.93681782e-02 1.77682579e-01 5.84900975e-01
3.25256616e-01 5.46991765e-01 5.30538894e-02 3.52067113e-01
-2.38819703e-01 -2.16244891e-01 1.06805658e+00 1.22899532e-01
1.89410001e-01 -1.62210688e-02 2.98792273e-01 9.91248965e-01
7.20983267e-01 -2.40714386e-01 4.50913429e-01 -4.12932336e-01
7.78634008e-03 -2.26124287e-01 1.04318118e+00 -9.86865342e-01
6.95865825e-02 3.71876448e-01 8.40492845e-01 -1.15515804e+00
5.35208046e-01 -7.71372437e-01 -3.14893901e-01 8.36608469e-01
-1.55346468e-01 2.50480678e-02 1.27703711e-01 5.83595224e-02
5.38933992e-01 -3.20373148e-01 8.77891660e-01 -4.10542130e-01
-4.69555974e-01 1.84064850e-01 -2.74691850e-01 -4.04240727e-01
6.98086321e-01 -3.37070942e-01 7.15987980e-02 6.50265515e-02
-1.33117676e+00 -2.52228379e-01 2.42670909e-01 3.82609606e-01
4.27833945e-01 -1.12883270e+00 -5.55640757e-01 4.27470133e-02
-9.35932398e-02 1.49093583e-01 -3.93968225e-01 1.03860176e+00
-9.29427862e-01 7.62970984e-01 -5.25209665e-01 -1.04257023e+00
-1.57150018e+00 2.17258126e-01 8.13954353e-01 1.57602027e-01
-7.80694485e-01 9.01461840e-01 -4.77185786e-01 -8.44176292e-01
7.24081039e-01 -3.79955709e-01 6.50879145e-02 9.79510322e-02
2.66064286e-01 7.83665359e-01 3.80677760e-01 -4.53650266e-01
-5.54836810e-01 9.88093495e-01 4.36559394e-02 -4.56541210e-01
1.32957804e+00 1.48102447e-01 -2.03982189e-01 3.22426051e-01
9.22222793e-01 -5.07668853e-02 -1.71099806e+00 5.56009933e-02
4.19914983e-02 -3.98523897e-01 -5.00885164e-03 -1.20299470e+00
-6.71969533e-01 1.13966477e+00 1.26792860e+00 -4.25028950e-01
3.49660635e-01 2.41748393e-01 3.15569729e-01 8.90234828e-01
3.92257810e-01 -1.41091812e+00 3.54358226e-01 7.20826685e-01
1.30445218e+00 -1.10049176e+00 1.66257679e-01 -5.05139291e-01
-2.61346921e-02 1.67990685e+00 5.40178061e-01 -3.46392184e-01
5.13208926e-01 6.57428384e-01 6.68570697e-02 -4.44069713e-01
2.32095987e-01 -2.03925684e-01 4.99516904e-01 9.14965928e-01
5.91599345e-01 -2.23498419e-01 -1.52567327e-01 1.25547439e-01
-2.18352377e-02 1.79251403e-01 -6.85581118e-02 1.45157826e+00
-1.82169229e-01 -1.28178978e+00 -7.42751598e-01 7.67622292e-02
-7.28160799e-01 2.81461567e-01 -7.88134694e-01 9.47710454e-01
3.18870813e-01 9.57290083e-02 9.59144011e-02 -1.71012878e-01
4.59318042e-01 4.09649789e-01 1.72907436e+00 -5.83885908e-01
-7.25296557e-01 5.72952271e-01 -9.44993645e-02 -5.98808050e-01
-6.84842288e-01 -9.06388998e-01 -1.23060775e+00 2.65139967e-01
-3.47987503e-01 -7.28843391e-01 1.21678412e+00 1.06342649e+00
1.39941484e-01 7.19343960e-01 4.54369839e-03 -1.93690157e+00
-8.93677056e-01 -1.05918944e+00 -5.49621284e-01 -9.24835131e-02
2.45500907e-01 -1.10866678e+00 2.94049978e-01 1.58796068e-02] | [6.5501179695129395, -0.8015610575675964] |
3a235c80-bcdd-4f41-9d4e-068ce65555d6 | a-survey-on-knowledge-graph-based-methods-for | 2210.08119 | null | https://arxiv.org/abs/2210.08119v1 | https://arxiv.org/pdf/2210.08119v1.pdf | A Survey on Knowledge Graph-based Methods for Automated Driving | Automated driving is one of the most active research areas in computer science. Deep learning methods have made remarkable breakthroughs in machine learning in general and in automated driving (AD)in particular. However, there are still unsolved problems to guarantee reliability and safety of automated systems, especially to effectively incorporate all available information and knowledge in the driving task. Knowledge graphs (KG) have recently gained significant attention from both industry and academia for applications that benefit by exploiting structured, dynamic, and relational data. The complexity of graph-structured data with complex relationships and inter-dependencies between objects has posed significant challenges to existing machine learning algorithms. However, recent progress in knowledge graph embeddings and graph neural networks allows to applying machine learning to graph-structured data. Therefore, we motivate and discuss the potential benefit of KGs applied to the main tasks of AD including 1) ontologies 2) perception, 3) scene understanding, 4) motion planning, and 5) validation. Then, we survey, analyze and categorize ontologies and KG-based approaches for AD. We discuss current research challenges and propose promising future research directions for KG-based solutions for AD. | ['Lavdim Halilaj', 'Cory Henson', 'Sebastian Monka', 'Juergen Luettin'] | 2022-09-30 | null | null | null | null | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [-7.91987926e-02 3.75699759e-01 -4.82905656e-01 -4.83597100e-01
7.58307502e-02 -2.52057612e-01 4.90762949e-01 5.95534682e-01
-3.22275847e-01 4.39537883e-01 8.92019719e-02 -4.93806690e-01
-5.11161327e-01 -1.25599277e+00 -6.17478609e-01 -3.43067914e-01
-2.90787876e-01 6.33394957e-01 5.32354414e-01 -5.91343641e-01
3.28027606e-02 7.01587617e-01 -2.09532046e+00 -7.26703703e-02
8.91604125e-01 1.09774387e+00 2.00142175e-01 3.78378302e-01
-2.73436934e-01 1.13144696e+00 -1.05176970e-01 -3.70815247e-01
-2.41252519e-02 -8.07848945e-03 -7.08112121e-01 -4.11147485e-03
3.35255146e-01 1.57326013e-01 -8.34845185e-01 9.88114834e-01
7.76193514e-02 4.21385944e-01 3.70389283e-01 -1.86668098e+00
-7.51910806e-01 3.12984258e-01 4.43004072e-02 1.09046347e-01
-4.43220139e-03 9.80942994e-02 8.90278399e-01 -3.73155564e-01
6.05658650e-01 1.17372084e+00 3.58150184e-01 4.18469697e-01
-4.56719726e-01 -4.76040453e-01 2.63654202e-01 1.36503756e+00
-1.20647383e+00 -2.10983485e-01 8.73284817e-01 -6.37511730e-01
1.15356731e+00 -6.81082904e-02 9.76807177e-01 5.23428261e-01
4.26302046e-01 8.49271476e-01 6.59143507e-01 -2.32048869e-01
3.63277823e-01 7.70789906e-02 5.33994973e-01 8.80992711e-01
4.55851495e-01 2.60857642e-01 -6.87006474e-01 2.74012893e-01
1.03440329e-01 -2.73502409e-01 1.66388452e-01 -8.44074607e-01
-9.61715877e-01 1.04576647e+00 7.08726645e-01 1.16921596e-01
-2.79076397e-01 2.62845993e-01 5.42001188e-01 8.57355893e-02
1.26080513e-01 3.40825379e-01 -2.13270724e-01 -1.33762375e-01
-2.33402908e-01 4.10351455e-01 6.66706622e-01 1.18524039e+00
1.08307886e+00 2.96267509e-01 3.21642905e-01 5.90762794e-01
1.58672184e-01 3.81125987e-01 2.42220029e-01 -7.17061877e-01
5.10536075e-01 1.08309829e+00 -2.19183251e-01 -1.42060518e+00
-6.85692489e-01 -2.93730199e-01 -6.30191505e-01 1.44472033e-01
-3.90735865e-02 2.58684456e-01 -7.35626519e-01 1.32398188e+00
3.58677059e-01 3.08871150e-01 2.85983384e-01 8.33249927e-01
1.13360786e+00 3.46293241e-01 1.75036862e-01 2.82971442e-01
1.18849051e+00 -1.06661522e+00 -9.89180148e-01 -6.56327188e-01
7.97397316e-01 -6.69667050e-02 7.89878070e-01 9.64411646e-02
-5.68777680e-01 -7.56736577e-01 -1.26001644e+00 -3.28322411e-01
-1.04224396e+00 -3.16382647e-01 9.19637263e-01 4.06691104e-01
-9.88538444e-01 3.04477096e-01 -7.17099011e-01 -5.85610271e-01
7.35155702e-01 5.36674440e-01 -4.90865231e-01 -4.70368266e-01
-1.53957391e+00 1.22083485e+00 7.83053219e-01 2.12440059e-01
-8.02082479e-01 -2.64584154e-01 -1.47879446e+00 -1.15329556e-01
7.42738724e-01 -6.52946055e-01 8.99678767e-01 -2.87126184e-01
-1.09703195e+00 7.05070138e-01 -1.14023082e-01 -9.38181460e-01
1.22380145e-01 -3.63079309e-01 -8.28658998e-01 -1.42151624e-01
1.00377034e-02 6.18773401e-01 4.32863683e-01 -1.01036763e+00
-9.01370287e-01 -5.65754950e-01 2.66624957e-01 1.72268465e-01
-4.43752617e-01 -3.91920745e-01 -3.62966388e-01 1.27490059e-01
-2.16154847e-02 -8.98830533e-01 -3.22974294e-01 -1.08421415e-01
-1.76621720e-01 -6.50048018e-01 1.13653159e+00 -4.20706123e-01
1.15449655e+00 -2.20806170e+00 2.43472651e-01 2.47057411e-03
4.13003147e-01 5.20008922e-01 -3.09337750e-02 5.11577427e-01
2.72540540e-01 -2.58847326e-01 -1.42694443e-01 1.54469803e-01
1.38716683e-01 8.52455199e-01 -1.37104869e-01 2.84755319e-01
2.69602120e-01 1.29652274e+00 -1.14593422e+00 -4.30631995e-01
7.79983342e-01 2.86763757e-01 -3.61148238e-01 1.83156326e-01
-4.40354615e-01 2.36210406e-01 -7.44098902e-01 4.55747247e-01
4.09225553e-01 -1.14714410e-02 1.68318108e-01 -3.25759828e-01
-6.66000769e-02 1.79666713e-01 -9.01465178e-01 1.49932480e+00
-3.45972538e-01 9.67885852e-01 -3.63605827e-01 -1.56711113e+00
1.07893550e+00 -1.55781314e-01 5.02252996e-01 -1.09981549e+00
2.21677497e-01 2.88928282e-02 1.17110856e-01 -1.02365577e+00
8.39551210e-01 6.33473555e-03 -1.89342141e-01 -2.03906059e-01
-1.77670956e-01 -3.27507108e-01 2.93217719e-01 9.77256969e-02
9.56609964e-01 7.67013505e-02 3.41292381e-01 -8.15402437e-03
6.64311826e-01 5.91408908e-01 3.89688939e-01 2.06296310e-01
-5.49922049e-01 -9.28391814e-02 2.50736982e-01 -8.65678847e-01
-6.29894257e-01 -7.40126133e-01 1.88845366e-01 7.36094356e-01
7.14004517e-01 -5.00373542e-01 -5.37092328e-01 -6.36956573e-01
3.66731226e-01 9.45125520e-01 -5.54162562e-01 -6.80547297e-01
-6.22467995e-01 -3.11664671e-01 3.62452894e-01 8.00394714e-01
6.31697178e-01 -1.21071780e+00 -6.73763037e-01 2.96337187e-01
-1.82555646e-01 -1.77627432e+00 1.61144629e-01 6.67118952e-02
-6.67878807e-01 -1.46780539e+00 2.83933997e-01 -8.34637761e-01
3.52154851e-01 4.87076759e-01 1.09673607e+00 5.28247878e-02
-3.47653270e-01 5.83218634e-01 -4.13579017e-01 -8.08831036e-01
-3.65765750e-01 3.20254117e-02 1.29093811e-01 2.50890218e-02
7.78307617e-01 -2.71366537e-01 -1.27000272e-01 4.78460103e-01
-7.50075579e-01 -3.82194109e-02 4.35576797e-01 4.15207803e-01
5.91300249e-01 6.52635992e-01 6.60361409e-01 -5.48594832e-01
4.78098959e-01 -4.17013168e-01 -8.18936765e-01 1.25369683e-01
-5.81535280e-01 9.04862508e-02 4.30658340e-01 1.16095327e-01
-6.59782529e-01 3.64127243e-03 -5.50066754e-02 -3.03354949e-01
-2.83869773e-01 7.94020414e-01 -4.41012770e-01 -1.77919865e-01
5.50244033e-01 3.92802916e-02 6.97978437e-02 4.27372120e-02
6.85766280e-01 5.95009029e-01 5.33035815e-01 -3.88070047e-01
6.86203837e-01 6.27587259e-01 4.08775508e-01 -1.07927227e+00
-9.36583936e-01 -5.94056010e-01 -8.22839439e-01 -5.20224690e-01
1.15061355e+00 -7.92538822e-01 -7.49735177e-01 2.51463026e-01
-9.06929910e-01 -3.26947242e-01 -4.73478466e-01 5.72117329e-01
-8.24891567e-01 3.71751606e-01 4.85182405e-02 -6.72278881e-01
1.79098353e-01 -1.14020956e+00 9.10093486e-01 1.73236966e-01
1.85147431e-02 -1.11928499e+00 -1.73773274e-01 7.86184430e-01
3.75175387e-01 3.28611076e-01 1.12705588e+00 -4.05352592e-01
-8.11758280e-01 -3.25902134e-01 -3.13017219e-01 3.96531880e-01
2.29513869e-01 -3.58302355e-01 -8.17838132e-01 4.42574359e-02
-4.10406888e-01 -2.67931283e-01 7.45384276e-01 2.37841189e-01
1.17406499e+00 8.10027346e-02 -6.45610094e-01 4.01912868e-01
1.19181144e+00 2.32087076e-01 6.45727754e-01 5.28748095e-01
1.07269609e+00 1.10990250e+00 9.13792074e-01 8.63597915e-03
1.05106831e+00 6.62194908e-01 1.00592232e+00 1.82214484e-01
-3.19447935e-01 -3.54587972e-01 1.78273916e-01 1.05232382e+00
-3.52554917e-02 -3.02024662e-01 -1.09081316e+00 8.91907454e-01
-2.25785875e+00 -8.00448000e-01 -4.36001450e-01 1.90325546e+00
1.16528414e-01 2.64499336e-01 -6.26733974e-02 3.09062600e-01
6.02694690e-01 1.78418830e-01 -7.17737079e-01 -4.06069636e-01
-1.18471622e-01 7.57743837e-03 5.71876585e-01 3.16923499e-01
-1.18784630e+00 1.37652647e+00 5.88722324e+00 5.25867164e-01
-9.65094864e-01 -8.26148391e-02 6.56798258e-02 4.15184587e-01
-1.68100253e-01 -7.09984452e-02 -9.29557860e-01 4.29348201e-02
1.08792830e+00 -3.13916385e-01 3.90947461e-01 1.20438743e+00
8.03330913e-02 -1.01667024e-01 -1.19283831e+00 1.03092468e+00
1.64643496e-01 -1.54363132e+00 1.17245264e-01 2.01484282e-02
6.59967542e-01 2.92888522e-01 -9.02609453e-02 5.19796014e-01
3.15216601e-01 -1.08177364e+00 6.59353673e-01 3.31780881e-01
4.12734807e-01 -8.34453583e-01 8.90353143e-01 4.17396367e-01
-1.56911814e+00 -2.52219349e-01 -5.95914483e-01 -2.37392470e-01
3.14782292e-01 5.83752990e-01 -7.58715272e-01 1.00423539e+00
7.25725591e-01 1.20995045e+00 -5.15034199e-01 9.21913564e-01
-3.59970808e-01 2.79379427e-01 -3.63607630e-02 -1.78138867e-01
3.17342341e-01 -2.59780556e-01 3.84311259e-01 8.60355973e-01
8.04656744e-03 -8.24865028e-02 2.96620369e-01 6.04614079e-01
2.19167560e-01 -1.03683263e-01 -1.08893764e+00 -4.44067657e-01
2.17533857e-01 1.05290580e+00 -5.19316435e-01 -2.36512586e-01
-6.30417585e-01 2.70819217e-01 5.55616856e-01 2.81103402e-01
-8.27849686e-01 -2.81091869e-01 1.01839769e+00 2.65975565e-01
2.55447686e-01 -9.39844370e-01 -7.88589343e-02 -7.20474780e-01
4.97041568e-02 -4.58328098e-01 6.06435955e-01 -7.68836617e-01
-1.03447556e+00 6.10471070e-01 2.30857193e-01 -1.10871005e+00
-1.39598906e-01 -9.28320408e-01 -9.98238921e-02 4.19223189e-01
-2.24455762e+00 -1.23306942e+00 -7.90620506e-01 7.02400267e-01
4.34103191e-01 -2.60546952e-01 6.72686160e-01 3.47113669e-01
-2.97068596e-01 6.74255714e-02 -3.06621403e-01 1.38771776e-02
1.77717373e-01 -9.67018008e-01 5.65680146e-01 6.07016683e-01
1.82146773e-01 3.04896049e-02 4.27251607e-01 -6.62096620e-01
-1.97441578e+00 -1.52609909e+00 8.68947923e-01 -5.69172859e-01
8.23972046e-01 -2.72497416e-01 -8.23934078e-01 7.61223674e-01
-1.08666532e-02 2.58294165e-01 4.59862530e-01 1.04904562e-01
-1.43950745e-01 -2.78228194e-01 -8.96478057e-01 5.58372259e-01
1.27953887e+00 -4.77615714e-01 -5.19600809e-01 4.18269753e-01
8.14300060e-01 -3.97393525e-01 -6.34409249e-01 6.04416966e-01
3.04856479e-01 -8.91204178e-01 7.90707469e-01 -9.13881719e-01
6.90224580e-03 -5.97203195e-01 -3.10631603e-01 -1.21455121e+00
-3.02766055e-01 -2.14612648e-01 -3.54050159e-01 5.35498202e-01
4.00511771e-02 -8.03379595e-01 8.46992433e-01 4.23604548e-01
-6.42115831e-01 -7.11886644e-01 -7.90773749e-01 -9.77935374e-01
-1.97561488e-01 -1.11208415e+00 7.68715560e-01 7.57878125e-01
5.18694483e-02 3.58886987e-01 -1.71923786e-01 4.44128633e-01
6.11962020e-01 8.07638466e-02 1.00840509e+00 -1.46385562e+00
2.07767040e-01 -2.08531111e-01 -1.30755126e+00 -8.27485025e-01
4.94824171e-01 -9.76706505e-01 -5.30705564e-02 -2.31597757e+00
-2.94565022e-01 -4.68144834e-01 -1.52877361e-01 6.21976495e-01
5.57463132e-02 -5.32838777e-02 -1.58043094e-02 -2.71032512e-01
-7.31029332e-01 8.34180832e-01 1.28522456e+00 -5.38032353e-01
6.95181312e-03 -2.93905616e-01 -4.94540751e-01 5.56066513e-01
8.79245281e-01 -1.35882184e-01 -7.72415221e-01 -5.47498345e-01
4.92822886e-01 -2.80896127e-01 4.21290606e-01 -1.24016345e+00
4.73883152e-01 -4.59040523e-01 -3.46011728e-01 -6.55404806e-01
3.44564974e-01 -1.11360025e+00 2.69078873e-02 3.56533498e-01
1.26540795e-01 -8.99398699e-03 6.05793178e-01 7.73348749e-01
-5.89141786e-01 2.75044758e-02 5.06302059e-01 1.41184032e-01
-1.73521018e+00 6.43003404e-01 -3.63932371e-01 2.29594149e-02
1.44868910e+00 -3.97963852e-01 -4.31564927e-01 -3.04684609e-01
-8.41542661e-01 7.19845831e-01 3.48072238e-02 1.07773018e+00
7.53217638e-01 -1.32103717e+00 -4.48134571e-01 2.99124241e-01
6.54396653e-01 2.12415293e-01 3.81273240e-01 4.75497246e-01
-4.62940782e-01 6.63055301e-01 -3.58393550e-01 -7.36532748e-01
-1.06061804e+00 8.25334370e-01 1.71494350e-01 1.65889803e-02
-7.34393597e-01 5.24956226e-01 1.44113213e-01 -4.63625729e-01
2.24419296e-01 -4.16096330e-01 -6.16724610e-01 -2.04539061e-01
2.23083928e-01 4.88591373e-01 3.54981989e-01 -9.35779214e-01
-6.00452960e-01 6.66178167e-01 1.73877865e-01 5.47323823e-01
1.18410027e+00 -6.50144788e-03 -1.64736318e-03 3.48767251e-01
9.18608606e-01 -5.45127034e-01 -7.56676912e-01 -1.77365616e-01
3.26781601e-01 -2.32440963e-01 1.30021885e-01 -4.03169602e-01
-1.12758696e+00 1.14889073e+00 3.80794227e-01 3.38289768e-01
7.69818902e-01 1.56626657e-01 8.18809927e-01 6.54421687e-01
8.25019062e-01 -1.16282594e+00 1.31324917e-01 6.38854802e-01
8.13030422e-01 -1.27416921e+00 2.72370595e-02 -5.99053621e-01
-6.51705861e-01 1.10625660e+00 7.47799218e-01 7.32178390e-02
9.59466338e-01 5.05551649e-03 -1.80314947e-02 -6.22799516e-01
-4.49134618e-01 -7.19988585e-01 4.85977113e-01 1.05165792e+00
-7.14455247e-02 2.54061103e-01 8.48533362e-02 3.22125763e-01
-4.05253470e-01 1.59583509e-01 2.95401245e-01 9.79103923e-01
-6.66346014e-01 -1.17550969e+00 4.78917919e-02 4.39286172e-01
4.06412005e-01 3.01273167e-01 -3.00458223e-01 9.67476845e-01
3.57514501e-01 1.21225607e+00 2.06770021e-02 -6.95001662e-01
6.72228456e-01 -1.17694050e-01 4.69015002e-01 -6.18213892e-01
1.71938285e-01 -8.16267252e-01 4.21272486e-01 -6.99633598e-01
-4.87602323e-01 -3.89407665e-01 -1.59494519e+00 -1.08055629e-01
-2.99010158e-01 1.66864827e-01 9.86102700e-01 1.16725731e+00
5.93694746e-01 9.45499659e-01 1.75234959e-01 -4.75343019e-01
-9.96016804e-03 -4.47088063e-01 -4.36863095e-01 1.71715215e-01
8.75298083e-02 -1.18500268e+00 -4.22771387e-02 -9.18957219e-03] | [7.9755401611328125, -2.1430611610412598] |
74544db0-8a74-436e-8162-015ded4a92e5 | acoustic-scene-classification-using-audio | 2003.09164 | null | http://arxiv.org/abs/2003.09164v2 | http://arxiv.org/pdf/2003.09164v2.pdf | Acoustic Scene Classification using Audio Tagging | Acoustic scene classification systems using deep neural networks classify
given recordings into pre-defined classes. In this study, we propose a novel
scheme for acoustic scene classification which adopts an audio tagging system
inspired by the human perception mechanism. When humans identify an acoustic
scene, the existence of different sound events provides discriminative
information which affects the judgement. The proposed framework mimics this
mechanism using various approaches. Firstly, we employ three methods to
concatenate tag vectors extracted using an audio tagging system with an
intermediate hidden layer of an acoustic scene classification system. We also
explore the multi-head attention on the feature map of an acoustic scene
classification system using tag vectors. Experiments conducted on the detection
and classification of acoustic scenes and events 2019 task 1-a dataset
demonstrate the effectiveness of the proposed scheme. Concatenation and
multi-head attention show a classification accuracy of 75.66 % and 75.58 %,
respectively, compared to 73.63 % accuracy of the baseline. The system with the
proposed two approaches combined demonstrates an accuracy of 76.75 %. | [] | 2020-04-19 | null | null | null | null | ['audio-tagging'] | ['audio'] | [ 5.66377461e-01 -2.38321811e-01 7.95655191e-01 -5.52092671e-01
-8.90508592e-01 -4.74925280e-01 5.04775882e-01 2.56109685e-01
-8.01170230e-01 -1.46364584e-03 2.45261103e-01 1.57858580e-01
1.12677336e-01 -6.00413918e-01 -3.60243410e-01 -9.37062562e-01
1.14944823e-01 -1.18736714e-01 3.37184727e-01 1.60724178e-01
5.35923123e-01 1.82941869e-01 -2.11728787e+00 6.53266132e-01
6.38358593e-02 1.51708722e+00 6.76928937e-01 9.91859972e-01
-9.30333361e-02 7.52334714e-01 -9.56134021e-01 -8.78152549e-02
-7.56795481e-02 -2.55469412e-01 -6.69570625e-01 -1.35605216e-01
5.73046684e-01 -4.01322776e-03 5.42677119e-02 1.12921524e+00
8.73094857e-01 5.37378430e-01 5.65389037e-01 -1.27146327e+00
-3.04837823e-01 6.83065295e-01 -1.47869065e-01 3.29352617e-01
4.58244890e-01 -3.16178381e-01 9.57316160e-01 -9.25076425e-01
-4.64829654e-02 1.07205749e+00 7.82393575e-01 3.09408754e-01
-6.97901130e-01 -8.44744444e-01 -1.07167475e-02 5.86903811e-01
-1.54486501e+00 -6.42640889e-01 1.02714610e+00 -6.83681488e-01
9.07094359e-01 4.04835850e-01 2.62442917e-01 5.58805883e-01
2.65665591e-01 6.26355529e-01 1.03493738e+00 -6.40114069e-01
2.88264543e-01 2.50310957e-01 4.35376346e-01 4.45931643e-01
-3.06009769e-01 -1.65537879e-01 -6.94980383e-01 -8.43877196e-02
6.22316636e-02 -5.82728125e-02 5.13660349e-02 1.72027737e-01
-9.81786370e-01 6.01563811e-01 3.00590336e-01 6.26364112e-01
-3.92312884e-01 8.93528461e-02 7.68897057e-01 -6.58903942e-02
3.48324001e-01 4.14985776e-01 -1.75472051e-01 -3.35455947e-02
-8.74615848e-01 -2.17756063e-01 5.13173580e-01 6.70088768e-01
6.19347751e-01 3.80058259e-01 -1.38177499e-01 1.24899220e+00
4.71987695e-01 4.38751578e-01 8.31801414e-01 -6.44676745e-01
1.63748279e-01 2.09905222e-01 -1.87422037e-01 -1.17988718e+00
-6.32623434e-01 -5.51630616e-01 -6.69100404e-01 -9.73818675e-02
-8.69963318e-02 7.86583871e-04 -9.76235509e-01 1.67542779e+00
2.79826522e-01 5.76635122e-01 3.35084856e-01 8.08276951e-01
1.20357871e+00 9.72154796e-01 6.24134243e-01 4.55150530e-02
1.47546625e+00 -8.30461144e-01 -9.31112468e-01 1.71237126e-01
5.25114775e-01 -1.11155427e+00 1.04719353e+00 4.17911142e-01
-7.65457511e-01 -1.25666666e+00 -1.16172695e+00 1.87586159e-01
-6.94589317e-01 5.44683576e-01 2.04401523e-01 6.98412478e-01
-1.17713189e+00 1.76975012e-01 -5.00738502e-01 -6.59503043e-01
7.05454126e-02 3.94440949e-01 -1.17376372e-01 3.11337084e-01
-1.11390448e+00 4.61228669e-01 5.16171932e-01 3.44294220e-01
-1.08147812e+00 -3.23829621e-01 -9.26458001e-01 2.04818830e-01
-8.49931464e-02 -1.15324922e-01 1.45907903e+00 -7.66583860e-01
-1.48263204e+00 6.89134598e-01 -9.98725742e-02 -4.04040664e-01
-2.66194552e-01 -3.11502993e-01 -6.71132445e-01 1.45344481e-01
2.82599241e-01 4.80848998e-01 4.89510655e-01 -1.21212280e+00
-1.05292225e+00 -1.05161205e-01 -6.12192117e-02 3.99978936e-01
-3.59616578e-01 3.21813434e-01 -4.17870395e-02 -4.90810454e-01
2.87156433e-01 -8.80130947e-01 -2.67453454e-02 -6.69639349e-01
-6.07259214e-01 -4.20892328e-01 9.20343161e-01 -4.54762518e-01
1.13847816e+00 -2.64505816e+00 -4.85362977e-01 6.23001829e-02
-1.17265284e-01 2.69997001e-01 -1.27226636e-01 3.85774910e-01
-6.65694252e-02 -1.35230273e-01 -2.22470164e-01 -5.96651316e-01
1.98742598e-01 -1.08124897e-01 -4.47664171e-01 2.01683745e-01
-6.38616905e-02 2.12488949e-01 -8.26595187e-01 -5.04256010e-01
4.70558226e-01 4.06967968e-01 -4.57861751e-01 5.72611868e-01
4.74522144e-01 1.04268685e-01 -3.79862428e-01 5.69119871e-01
6.04836166e-01 3.55735272e-01 6.87256381e-02 -4.45595264e-01
-3.09043705e-01 3.67568523e-01 -1.27781904e+00 1.57718980e+00
-6.58889830e-01 7.92558968e-01 -1.84546575e-01 -1.13683915e+00
1.10172307e+00 5.75705051e-01 3.14132631e-01 -5.46331704e-01
2.96201348e-01 9.45679247e-02 1.17326051e-01 -6.93948984e-01
7.90695250e-01 -2.29151756e-01 -4.64632332e-01 3.21208417e-01
3.50146830e-01 1.04839932e-02 -3.16124827e-01 9.03042182e-02
1.03233731e+00 -2.70562470e-02 2.53322482e-01 -2.10474893e-01
7.90676534e-01 -3.82165551e-01 4.16154921e-01 1.09120023e+00
-4.81424332e-01 4.92429495e-01 -1.84007451e-01 -3.97500038e-01
-5.57378113e-01 -8.32185686e-01 -2.94232577e-01 1.60730290e+00
1.19365171e-01 -4.50906694e-01 -7.66590297e-01 -3.59932572e-01
-3.12033623e-01 7.92070210e-01 -6.52217269e-01 -3.82121682e-01
-2.72260845e-01 -7.16171026e-01 9.06226635e-01 6.18763268e-01
7.28112102e-01 -1.47654223e+00 -8.00698340e-01 1.89074114e-01
-2.24600792e-01 -1.24980676e+00 7.40291253e-02 4.15172011e-01
-1.42829120e-01 -6.11309886e-01 -2.15352833e-01 -1.18460798e+00
1.69653848e-01 2.25295931e-01 7.54992068e-01 -1.00243375e-01
-1.28517389e-01 6.53481185e-01 -6.33567393e-01 -8.43815506e-01
-3.52106631e-01 -1.03249468e-01 3.58050227e-01 6.32190466e-01
5.44373095e-01 -3.92731458e-01 -4.68244076e-01 1.21607997e-01
-7.84701884e-01 -2.06746474e-01 3.62015784e-01 4.50523317e-01
4.78739679e-01 1.90906584e-01 7.14728653e-01 -5.72556198e-01
4.11029994e-01 -4.03116435e-01 -1.13691606e-01 1.23925723e-01
-1.02338068e-01 -4.77609426e-01 5.35945892e-01 -2.80152619e-01
-1.06827044e+00 5.03312409e-01 -4.19217765e-01 -1.80247381e-01
-8.82784188e-01 2.40653470e-01 -2.91581362e-01 -2.59255171e-02
2.65824348e-01 3.33620608e-01 -6.44052029e-01 -5.85741758e-01
1.11785896e-01 1.26152503e+00 6.79654598e-01 -2.17481926e-01
3.62309217e-01 1.70510247e-01 -2.95430630e-01 -1.06514478e+00
-8.99521708e-01 -9.42249954e-01 -7.76923478e-01 -5.80895603e-01
1.36687160e+00 -8.50325286e-01 -7.72722542e-01 8.75992179e-01
-1.11772287e+00 -6.09545000e-02 -5.90212503e-03 6.95519447e-01
-3.61240864e-01 2.22660542e-01 -4.03369039e-01 -1.22392190e+00
-1.90211520e-01 -1.14932597e+00 1.43827140e+00 8.13113451e-02
-1.81874827e-01 -6.70049965e-01 1.29408583e-01 2.00629294e-01
3.05219352e-01 -6.64608777e-02 5.75560987e-01 -1.15406942e+00
-1.14458598e-01 -1.91148251e-01 -4.62173745e-02 3.55702162e-01
2.44818911e-01 -2.13212103e-01 -1.98591936e+00 8.05655122e-03
5.54791428e-02 -2.01811865e-01 9.20513630e-01 3.98235261e-01
1.38253808e+00 1.01032536e-02 -9.27254111e-02 3.68303180e-01
1.36502099e+00 1.02687037e+00 4.99265432e-01 4.07364696e-01
8.26692462e-01 6.21311963e-01 6.34805083e-01 4.41192180e-01
3.32989007e-01 8.62865031e-01 3.22199792e-01 -1.40980154e-01
-1.40296385e-01 -6.57235011e-02 3.39090824e-01 1.39313483e+00
1.95871890e-01 -3.40363204e-01 -1.01426077e+00 4.91076171e-01
-1.53511870e+00 -9.19037759e-01 2.18171217e-02 2.00999212e+00
4.86654222e-01 5.48565462e-02 -1.05201811e-01 7.05094814e-01
8.88483405e-01 2.16503575e-01 1.25777975e-01 -7.42954910e-01
4.74161394e-02 2.97812730e-01 3.15762833e-02 4.05221552e-01
-1.68015897e+00 9.71413910e-01 5.74393463e+00 8.41527402e-01
-1.28254855e+00 2.68795937e-01 2.64576584e-01 3.03907990e-01
3.66885304e-01 -2.77264118e-01 -7.57302761e-01 4.97036844e-01
1.27801156e+00 2.46773213e-01 -1.33694395e-01 6.95902824e-01
1.02690533e-02 -1.22123912e-01 -9.21983063e-01 1.12725663e+00
3.96317482e-01 -8.17375124e-01 1.06812976e-01 -2.86778897e-01
3.07098955e-01 -2.31559858e-01 1.58756793e-01 4.96974230e-01
-8.26608241e-02 -5.99133730e-01 9.98567760e-01 5.41043639e-01
4.53489095e-01 -7.66621888e-01 1.03991151e+00 -7.97798485e-03
-1.60306513e+00 -1.56366408e-01 -5.00569880e-01 -2.24605575e-01
1.52522057e-01 1.96261734e-01 -1.20367324e+00 3.93599361e-01
1.06613421e+00 4.94977117e-01 -6.42254472e-01 1.30983078e+00
1.55158669e-01 1.13199759e+00 -2.64217705e-01 -2.33476847e-01
3.68833959e-01 1.57979935e-01 5.13870597e-01 1.82921767e+00
3.39406937e-01 6.78383261e-02 2.45659247e-01 3.34332734e-01
2.53834844e-01 4.29080933e-01 -8.17053497e-01 2.68288523e-01
4.88446087e-01 1.29169476e+00 -8.38300824e-01 -6.34075284e-01
-1.76884085e-01 6.72658682e-01 -1.16925895e-01 2.67551631e-01
-8.99429679e-01 -8.56064141e-01 3.20759833e-01 -4.95191395e-01
2.72773862e-01 5.80091262e-03 -2.50847727e-01 -4.06023115e-01
-3.09222400e-01 -3.49092245e-01 4.28768098e-01 -1.20540297e+00
-9.11824465e-01 1.12145913e+00 -1.38111249e-01 -1.44898403e+00
-1.70291457e-02 -4.87952173e-01 -6.38621032e-01 5.31493902e-01
-1.20919633e+00 -1.10510552e+00 -5.10344028e-01 4.93681520e-01
8.28383148e-01 -3.73956293e-01 1.02533007e+00 5.67365229e-01
-5.68567097e-01 7.59251833e-01 -5.39380983e-02 2.49210030e-01
6.19391918e-01 -1.19344175e+00 1.53829949e-02 8.21530104e-01
4.50854093e-01 3.73345762e-01 4.54370409e-01 -3.03529292e-01
-7.96379566e-01 -1.29449022e+00 1.07044792e+00 -2.50812560e-01
5.03945529e-01 -5.46651244e-01 -9.45160449e-01 5.00177920e-01
5.12581110e-01 -4.80365783e-01 1.18728197e+00 -9.62513126e-03
-2.11313650e-01 -3.70732069e-01 -8.59923720e-01 1.55243114e-01
5.82257807e-01 -9.53447282e-01 -7.45611548e-01 2.06836537e-01
7.22567677e-01 -5.24576828e-02 -7.11094081e-01 4.50264543e-01
6.25264406e-01 -6.80028379e-01 8.19254577e-01 -3.24645936e-01
9.37420279e-02 -5.73320389e-01 -9.33541477e-01 -1.10517657e+00
-4.75360841e-01 -2.48910964e-01 5.70791483e-01 1.53470564e+00
2.83095539e-01 -5.29625058e-01 1.07779190e-01 1.95055574e-01
-5.82498014e-01 -7.55648240e-02 -8.98379147e-01 -5.01563013e-01
-5.23704112e-01 -9.79921877e-01 5.28145313e-01 1.09885013e+00
-2.88212180e-01 5.29160559e-01 -2.50453472e-01 6.01059616e-01
4.45938557e-01 -2.83535980e-02 5.55368125e-01 -1.16330385e+00
-4.87441830e-02 -2.07656607e-01 -9.95475054e-01 -6.82676375e-01
1.63367242e-01 -9.20635164e-01 5.47308862e-01 -1.28249216e+00
2.58911908e-01 -4.59413528e-01 -1.12418354e+00 6.00905180e-01
-1.05941659e-02 7.37673819e-01 4.18426841e-01 1.70339003e-01
-9.26705480e-01 4.77885097e-01 4.84645724e-01 -9.70618427e-02
-1.16867214e-01 8.40401426e-02 -4.11299855e-01 7.98384488e-01
1.01707745e+00 -5.15079677e-01 -2.39008889e-01 -3.07127744e-01
-2.41936252e-01 -2.43963361e-01 5.25887132e-01 -1.56169176e+00
4.53221083e-01 3.25377762e-01 3.09239745e-01 -8.17673922e-01
6.21101022e-01 -7.97872543e-01 7.17462599e-02 2.61271626e-01
-6.37161851e-01 -4.06837203e-02 5.38955808e-01 5.29340148e-01
-5.96127331e-01 -3.32371116e-01 6.95286453e-01 1.35418311e-01
-1.33010995e+00 -4.30115551e-01 -8.73984158e-01 -5.26931047e-01
6.80471480e-01 -1.69375390e-01 -1.02338277e-01 -3.77735496e-01
-1.00087047e+00 -3.49115670e-01 -4.21450883e-01 5.66156983e-01
6.22235060e-01 -1.51905417e+00 -4.04417157e-01 2.27114707e-01
2.83613294e-01 -5.60370147e-01 4.53884959e-01 5.65744042e-01
-2.64990777e-01 5.06308794e-01 -1.89472064e-01 -8.54236960e-01
-1.47653425e+00 2.36763999e-01 4.32024777e-01 4.35706466e-01
-1.68404520e-01 1.08298409e+00 7.46726036e-01 -4.06568497e-01
4.53420162e-01 -4.17382926e-01 -8.41709495e-01 1.78912461e-01
4.56631660e-01 2.63548136e-01 2.35399112e-01 -1.13964140e+00
-6.48187160e-01 8.91846240e-01 2.38984257e-01 -1.78770453e-01
1.11902499e+00 -1.65928870e-01 1.01898082e-01 9.91591454e-01
1.42616081e+00 -3.40897255e-02 -5.84746122e-01 -2.49392658e-01
-8.34773332e-02 -1.43427357e-01 2.84992635e-01 -9.86484885e-01
-8.97678792e-01 1.28040862e+00 1.27847731e+00 6.52298033e-01
1.35540271e+00 -6.07255250e-02 5.73602974e-01 3.49756747e-01
1.34591356e-01 -1.03139293e+00 -1.22465730e-01 4.66956764e-01
8.30446422e-01 -1.14875674e+00 -4.67665523e-01 -2.52030969e-01
-7.42093027e-01 9.36492503e-01 5.90768814e-01 -3.77656333e-02
9.07049954e-01 1.40045002e-01 4.88700122e-01 -2.37803683e-01
-5.98824263e-01 -4.50727284e-01 5.03368676e-01 5.29201090e-01
3.88345808e-01 2.62309462e-01 -1.83220487e-02 8.89472246e-01
-4.64980215e-01 -5.27297854e-01 4.07004148e-01 1.06283867e+00
-7.64660895e-01 -5.15630245e-01 -3.92749846e-01 6.17390797e-02
-6.61622822e-01 -1.81769446e-01 -4.36774641e-01 4.00050193e-01
4.10050690e-01 1.46122718e+00 3.41570526e-01 -1.08854759e+00
5.27795494e-01 4.01804417e-01 -1.03180222e-01 -6.75511360e-01
-8.30213726e-01 3.89141321e-01 1.94817409e-02 -4.09172684e-01
-8.19126308e-01 -7.08627522e-01 -1.15636981e+00 4.89630342e-01
-3.36834341e-01 3.83725315e-01 1.06175780e+00 8.63604367e-01
3.03114086e-01 8.88616323e-01 8.47766340e-01 -9.27185178e-01
-1.00450933e-01 -1.08328331e+00 -6.41821980e-01 3.81882191e-01
4.58745211e-01 -8.28324258e-01 -4.77622509e-01 4.04457211e-01] | [15.12771224975586, 5.112823009490967] |
42b85084-cd23-4f3d-99d7-4540f4a35c3e | bayesian-networks-for-named-entity-prediction | 2302.13253 | null | https://arxiv.org/abs/2302.13253v1 | https://arxiv.org/pdf/2302.13253v1.pdf | Bayesian Networks for Named Entity Prediction in Programming Community Question Answering | Within this study, we propose a new approach for natural language processing using Bayesian networks to predict and analyze the context and how this approach can be applied to the Community Question Answering domain. We discuss how Bayesian networks can detect semantic relationships and dependencies between entities, and this is connected to different score-based approaches of structure-learning. We compared the Bayesian networks with different score metrics, such as the BIC, BDeu, K2 and Chow-Liu trees. Our proposed approach out-performs the baseline model at the precision metric. We also discuss the influence of penalty terms on the structure of Bayesian networks and how they can be used to analyze the relationships between entities. In addition, we examine the visualization of directed acyclic graphs to analyze semantic relationships. The article further identifies issues with detecting certain semantic classes that are separated in the structure of directed acyclic graphs. Finally, we evaluate potential improvements for the Bayesian network approach. | ['Sergey Kovalchuk', 'Alexey Gorbatovski'] | 2023-02-26 | null | null | null | null | ['community-question-answering', 'community-question-answering'] | ['miscellaneous', 'natural-language-processing'] | [ 2.31003109e-03 5.56873024e-01 4.90834750e-02 -7.33401597e-01
-2.24308163e-01 -5.72062850e-01 5.96615911e-01 8.26239109e-01
-1.95915118e-01 5.58243275e-01 5.92961729e-01 -5.30989110e-01
-9.73268747e-01 -1.05502236e+00 -3.06656599e-01 -1.73231643e-02
-4.61135864e-01 5.93766451e-01 8.91052306e-01 3.69207077e-02
6.17251873e-01 3.88716161e-01 -1.35855484e+00 4.02298629e-01
7.98620820e-01 5.27520359e-01 -9.86598283e-02 6.31743908e-01
-5.28087199e-01 9.94379520e-01 -7.16307580e-01 -5.33630073e-01
-3.37664366e-01 -8.55361372e-02 -1.13124526e+00 -4.38181818e-01
5.22372782e-01 -1.18792511e-03 -1.41119212e-01 9.54887509e-01
1.75492257e-01 2.57792532e-01 9.86450553e-01 -1.24999869e+00
-2.41111353e-01 8.99681568e-01 -2.97200590e-01 5.76795280e-01
9.52333868e-01 -6.31124854e-01 1.35283732e+00 -7.20698357e-01
6.57667100e-01 1.96191657e+00 6.39886081e-01 -1.42457113e-01
-1.11370838e+00 -5.53904474e-01 3.47161353e-01 7.00316906e-01
-1.31697083e+00 -9.91963670e-02 5.73058724e-01 -7.78290451e-01
9.39164102e-01 1.41568854e-01 2.23691642e-01 5.44083357e-01
-2.18087494e-01 5.05433559e-01 1.16153002e+00 -7.67890811e-01
2.31542408e-01 6.32773638e-02 1.02269661e+00 1.00435448e+00
5.50211906e-01 -1.30104572e-01 -6.71710968e-01 -5.65732777e-01
3.75487685e-01 -3.99594218e-01 3.98488902e-03 -2.80218184e-01
-8.61106932e-01 9.52544808e-01 2.16295689e-01 5.16204596e-01
-1.83230508e-02 -6.73160404e-02 4.75350246e-02 -4.27573696e-02
3.86883169e-01 4.44944173e-01 -4.74534243e-01 2.54743814e-01
-7.74969041e-01 2.25219250e-01 1.19172943e+00 8.24015439e-01
6.48136318e-01 -5.56319773e-01 -3.97293597e-01 7.17709124e-01
9.41860259e-01 7.88488016e-02 -5.28308213e-01 -1.11776221e+00
3.40406984e-01 6.71161234e-01 4.12494317e-03 -1.64783180e+00
-5.25620520e-01 -5.08531556e-02 -2.31351420e-01 -1.94498494e-01
8.29267681e-01 -4.86072004e-02 -4.84177381e-01 1.42305934e+00
4.29536343e-01 6.31882697e-02 -1.90213770e-01 3.97141784e-01
1.08024216e+00 4.78546739e-01 3.79831612e-01 7.06533194e-02
1.57559788e+00 -6.28384590e-01 -1.00226927e+00 -1.03615448e-02
6.19872451e-01 -7.47664511e-01 5.61052740e-01 2.03871831e-01
-7.44256258e-01 -1.67945266e-01 -8.65871251e-01 1.61732391e-01
-6.05602980e-01 -2.62470424e-01 5.92567384e-01 8.53302598e-01
-8.78310919e-01 7.75312006e-01 -7.62933373e-01 -7.31459498e-01
1.91296369e-01 2.03085765e-01 6.08362071e-02 1.52978739e-02
-1.49495041e+00 1.15448868e+00 7.38864601e-01 -2.01158240e-01
-2.69757658e-01 -5.45672297e-01 -7.87152052e-01 4.79525208e-01
5.06632566e-01 -5.52810431e-01 9.62630630e-01 -2.64822692e-01
-1.00960970e+00 5.19042909e-01 -4.13717955e-01 -2.07333446e-01
4.12473194e-02 6.10061064e-02 -5.35238683e-01 5.14688551e-01
1.46656126e-01 4.01500314e-01 1.97727084e-01 -1.32433093e+00
-7.22923219e-01 -4.64784086e-01 3.43560100e-01 -1.82076953e-02
-1.41660050e-01 5.33919513e-01 -3.66933048e-01 -5.16069829e-01
2.74149299e-01 -5.33964276e-01 -1.83602646e-01 -2.28087664e-01
-4.50911134e-01 -7.66906679e-01 7.32057333e-01 -7.78874278e-01
1.70721006e+00 -1.73181081e+00 -3.02180022e-01 9.88588929e-01
3.44413668e-01 -1.89638600e-01 1.08904414e-01 4.54512298e-01
-2.05898389e-01 5.77941120e-01 -1.51167125e-01 2.39385590e-01
-8.38762056e-03 3.24652195e-01 6.00821599e-02 3.40169900e-05
5.63968085e-02 4.11552399e-01 -9.63989258e-01 -9.38386202e-01
-1.17453381e-01 3.90274972e-01 -5.56700766e-01 -1.37824137e-02
-2.57504642e-01 7.18968036e-03 -5.70997715e-01 4.99328613e-01
5.09189844e-01 -3.59416813e-01 8.38342726e-01 -2.23239422e-01
9.11388397e-02 5.99881291e-01 -1.47894824e+00 1.01900816e+00
2.00024337e-01 7.65938342e-01 1.94192193e-02 -1.14009917e+00
1.06358027e+00 1.41061470e-01 9.80862826e-02 -2.09201649e-01
-2.96243340e-01 -1.01091832e-01 6.01242296e-02 -6.81882918e-01
3.03221405e-01 1.50049180e-01 1.76487610e-01 4.16401207e-01
6.45408928e-02 9.61336195e-02 5.80946267e-01 7.00313151e-01
1.13488007e+00 -1.47563577e-01 4.38544929e-01 -5.15948832e-01
5.46193004e-01 -4.77784947e-02 4.50823933e-01 9.63086426e-01
-1.75585881e-01 1.88433286e-02 1.08603501e+00 -1.97154850e-01
-6.31581426e-01 -1.09543586e+00 -2.35611260e-01 1.30308294e+00
1.57203469e-02 -9.40821230e-01 -4.89233226e-01 -1.03353298e+00
-1.34391621e-01 1.00703859e+00 -5.06247580e-01 1.73055321e-01
-4.34257954e-01 -7.16802239e-01 6.31715298e-01 4.99486774e-01
2.25610748e-01 -6.25259638e-01 -2.31671646e-01 2.11903825e-02
-4.63189095e-01 -1.02405965e+00 7.95493424e-02 4.45578322e-02
-1.00112092e+00 -1.51764882e+00 -8.55028536e-03 -7.30473101e-01
5.47530293e-01 -2.67150486e-03 1.35615528e+00 1.44912243e-01
1.55581534e-01 6.88262284e-01 -5.42460322e-01 -4.42862004e-01
-4.96609449e-01 -5.11092842e-02 -2.64071345e-01 -4.03549582e-01
8.61446619e-01 -5.20796120e-01 -4.44758087e-01 4.82323945e-01
-7.42889225e-01 -4.88126487e-01 1.24069145e-02 3.69506121e-01
9.49502662e-02 4.17748213e-01 5.08088291e-01 -1.10113358e+00
9.12670016e-01 -7.04283476e-01 -5.27886629e-01 7.86797941e-01
-8.92831683e-01 4.02363420e-01 -1.20066993e-01 4.07905057e-02
-1.34885049e+00 -2.64791936e-01 1.43091336e-01 3.75553578e-01
-2.67575115e-01 9.25336301e-01 -9.99225974e-02 2.03458861e-01
8.06514382e-01 -8.34593296e-01 -2.22786590e-01 -4.40962732e-01
3.53761137e-01 5.53462386e-01 -1.14061581e-02 -7.71566451e-01
5.07097423e-01 2.80458540e-01 1.91442132e-01 -6.50032222e-01
-1.13786221e+00 -6.82066023e-01 -7.73061097e-01 -3.16970378e-01
8.82593215e-01 -5.26272058e-01 -7.13867784e-01 -1.28140390e-01
-1.43079913e+00 2.64182955e-01 1.49348363e-01 6.00029171e-01
-5.23999445e-02 8.47098291e-01 -5.18395782e-01 -8.69821429e-01
2.17677355e-01 -5.99031627e-01 5.18450379e-01 2.86450446e-01
-6.81305110e-01 -1.38128698e+00 2.16868922e-01 7.70335734e-01
2.48432070e-01 7.24490955e-02 1.55769587e+00 -1.21594322e+00
-4.27286446e-01 2.49040648e-01 -4.45287496e-01 -1.25258267e-01
-2.30798975e-01 3.48168045e-01 -6.59571588e-01 1.92642733e-01
-3.43573868e-01 2.85320997e-01 8.27255487e-01 4.14539933e-01
8.32954884e-01 -3.43465447e-01 -6.81448698e-01 -2.19771653e-01
1.05368376e+00 3.21758568e-01 5.03889084e-01 3.00251305e-01
4.73627657e-01 1.49017799e+00 5.72298706e-01 1.80680826e-01
7.76946664e-01 4.33339357e-01 1.98107854e-01 3.97203803e-01
2.26033535e-02 -3.61855805e-01 1.68678239e-01 8.24682772e-01
-3.67653482e-02 -6.36829948e-03 -1.43705821e+00 5.14453650e-01
-2.04650235e+00 -1.06192982e+00 -9.03649151e-01 1.87861907e+00
6.63905799e-01 2.40098238e-01 1.34540468e-01 7.07357749e-02
1.21568871e+00 -4.55026589e-02 1.26166984e-01 -4.16202784e-01
3.23407687e-02 2.66274005e-01 2.18932956e-01 8.50723207e-01
-1.01632237e+00 7.67331004e-01 8.05747032e+00 6.53353512e-01
-5.51562905e-02 3.00998893e-02 3.73479694e-01 4.92761910e-01
-4.01797265e-01 5.88431835e-01 -8.51548076e-01 5.18763252e-02
1.11708999e+00 -1.79241188e-02 -1.72849506e-01 6.15418077e-01
3.52691650e-01 -5.62691271e-01 -1.02391982e+00 4.00368959e-01
-1.93581898e-02 -1.27296269e+00 -1.04972079e-01 -1.04750015e-01
5.65668344e-01 -2.50868112e-01 -6.90761566e-01 8.24812725e-02
7.75369763e-01 -9.87113178e-01 1.71682447e-01 8.70024383e-01
-6.49812520e-02 -6.36983275e-01 7.35660434e-01 1.96309030e-01
-1.19011915e+00 -5.70046902e-02 -1.94117472e-01 -1.39928430e-01
2.61740871e-02 9.67171907e-01 -1.18726957e+00 6.69255555e-01
1.01668692e+00 5.94901443e-01 -9.75880861e-01 1.23283899e+00
-7.11558759e-01 8.73003781e-01 -3.77761662e-01 -3.57315511e-01
-1.04346573e-01 -1.90468624e-01 6.67920947e-01 1.47335601e+00
1.11433938e-01 7.53049776e-02 1.46587953e-01 9.34790790e-01
4.60899651e-01 1.74186349e-01 -5.47810137e-01 1.03914127e-01
8.58771324e-01 9.40942168e-01 -1.00506938e+00 -5.29595077e-01
-3.84735614e-01 4.41608019e-02 1.28197566e-01 3.67232949e-01
-2.85290778e-01 -5.08908212e-01 -4.56784014e-03 2.47263879e-01
2.74950951e-01 -1.76846847e-01 -3.16797704e-01 -9.29902315e-01
-1.63551286e-01 -7.36631989e-01 1.05998409e+00 -8.24866354e-01
-1.53521597e+00 6.18099943e-02 7.27929473e-01 -4.91091371e-01
-1.64651781e-01 -5.73139787e-01 -5.65431297e-01 6.41446054e-01
-1.22872102e+00 -6.49082899e-01 -1.13397017e-01 4.86253738e-01
-4.04278748e-02 6.34864792e-02 8.69247198e-01 1.18574321e-01
-4.96513754e-01 1.00871980e-01 -1.89218074e-02 3.06534469e-01
7.14557588e-01 -1.44695950e+00 2.24551037e-02 8.16106260e-01
1.13266252e-01 7.38059700e-01 7.51106620e-01 -8.70363474e-01
-3.10162008e-01 -6.25315785e-01 1.37809062e+00 -6.52613461e-01
8.80078912e-01 -1.36999011e-01 -1.12520266e+00 5.73576987e-01
2.62730151e-01 -5.64249635e-01 1.01343846e+00 9.03082252e-01
-7.57105231e-01 2.15973675e-01 -1.26071274e+00 4.07393575e-01
8.17322135e-01 -5.13604403e-01 -1.21945369e+00 3.11849803e-01
5.92390776e-01 4.29917395e-01 -1.04295206e+00 4.97678161e-01
5.53137958e-01 -1.16575015e+00 9.07529712e-01 -7.53727496e-01
2.13036120e-01 -4.71056163e-01 -1.50559217e-01 -1.07429874e+00
-3.94883633e-01 -1.89530671e-01 -2.76863128e-01 1.32083297e+00
6.68163538e-01 -4.39015567e-01 8.79925251e-01 7.91161060e-01
4.60180819e-01 -2.77448237e-01 -7.21054018e-01 -4.98398900e-01
5.88666983e-02 -4.00861442e-01 2.88660020e-01 1.19733202e+00
3.58557194e-01 6.75781190e-01 2.10619956e-01 3.14480245e-01
7.33828545e-01 -1.70706570e-01 2.01018572e-01 -1.85663307e+00
-2.46304810e-01 -4.41653728e-01 -4.24296975e-01 -6.75727963e-01
2.28983268e-01 -8.88749838e-01 -2.75864512e-01 -1.92428875e+00
2.34851003e-01 -2.76696891e-01 -1.23961486e-01 2.44560450e-01
-3.92071247e-01 -5.33327341e-01 2.36583501e-01 8.93004984e-02
-7.05954790e-01 2.91671734e-02 7.86050498e-01 -7.35437870e-02
-6.60122791e-03 -1.55463247e-02 -7.02448964e-01 1.22631538e+00
8.32445800e-01 -9.27121758e-01 -3.34630400e-01 -7.96065629e-02
7.52391875e-01 -6.10096157e-02 2.59944975e-01 -4.83242512e-01
6.32999301e-01 -2.74302159e-02 1.61870629e-01 -9.14819598e-01
-1.26329333e-01 -7.90708721e-01 -8.41738954e-02 3.35349500e-01
-5.38679421e-01 4.54824641e-02 -2.52960443e-01 8.54448974e-01
-3.71880800e-01 -8.22398424e-01 4.67000037e-01 -2.10988969e-01
-5.28859019e-01 -4.80371147e-01 -7.32652783e-01 1.16920240e-01
7.73662150e-01 -1.80337712e-01 -3.59460980e-01 -5.19167006e-01
-1.17644167e+00 6.28576398e-01 -3.46444339e-01 1.94769248e-01
4.70298469e-01 -1.02262282e+00 -5.46050251e-01 -6.08212233e-01
-6.59262482e-03 -1.87177271e-01 -1.02631636e-01 7.19986439e-01
-4.83667940e-01 5.35355330e-01 1.85338423e-01 -4.51137841e-01
-1.59338093e+00 2.17806995e-01 2.06315145e-01 -8.03668976e-01
8.53365436e-02 5.82059860e-01 -1.64101049e-01 -6.67766988e-01
5.89208245e-01 -1.23416007e-01 -1.02690721e+00 4.91125554e-01
3.99216622e-01 8.49514663e-01 -6.45093322e-02 -2.78337002e-01
-6.76157653e-01 3.85189563e-01 -1.13961041e-01 -2.30510652e-01
1.27117527e+00 -3.29826981e-01 -6.75778031e-01 3.84413719e-01
7.73740113e-01 7.17119277e-02 -4.38256025e-01 -3.74679744e-01
1.01318026e+00 -2.70065099e-01 -5.73403575e-02 -9.61990535e-01
-5.24881363e-01 7.02701092e-01 4.17651832e-01 7.28905618e-01
6.75729275e-01 3.79969269e-01 -4.54538204e-02 7.34529674e-01
4.18483689e-02 -1.18403471e+00 4.10167575e-02 8.30987751e-01
5.35121560e-01 -1.07389724e+00 3.11852753e-01 -9.96908903e-01
-3.00665408e-01 1.41881943e+00 4.04176086e-01 1.15141772e-01
1.15505111e+00 1.46640942e-01 -3.79009902e-01 -6.84690475e-01
-7.79364228e-01 -3.44139159e-01 4.45103139e-01 6.99054837e-01
6.74802661e-01 1.57078117e-01 -6.67304039e-01 4.47763085e-01
-1.95729569e-01 -2.80385911e-01 5.96881032e-01 8.18326771e-01
-6.15887046e-01 -1.32097805e+00 -7.41015673e-01 4.86030132e-01
-6.27461791e-01 -2.53986984e-01 -8.43188763e-01 6.98185325e-01
-7.67872483e-02 1.60093868e+00 2.18315665e-02 -2.74784356e-01
4.18216228e-01 3.84598911e-01 3.26492906e-01 -7.71189392e-01
-6.10117316e-01 1.05401722e-03 8.31406415e-01 -2.27487117e-01
-1.07149816e+00 -6.95440590e-01 -1.08970094e+00 -1.72256768e-01
-5.91293693e-01 5.47609389e-01 5.75618029e-01 1.06961977e+00
3.82108927e-01 3.35324645e-01 1.88240469e-01 -6.12121727e-03
-1.40830129e-01 -1.07521319e+00 -3.22318405e-01 1.07812040e-01
-2.09338918e-01 -6.37951553e-01 -5.37658334e-01 1.01902932e-02] | [9.312521934509277, 8.537528038024902] |
1f683495-da9c-4fc8-ab74-d54c6bce9ec7 | a-scalable-second-order-method-for-ill | 2106.02119 | null | https://arxiv.org/abs/2106.02119v1 | https://arxiv.org/pdf/2106.02119v1.pdf | A Scalable Second Order Method for Ill-Conditioned Matrix Completion from Few Samples | We propose an iterative algorithm for low-rank matrix completion that can be interpreted as an iteratively reweighted least squares (IRLS) algorithm, a saddle-escaping smoothing Newton method or a variable metric proximal gradient method applied to a non-convex rank surrogate. It combines the favorable data-efficiency of previous IRLS approaches with an improved scalability by several orders of magnitude. We establish the first local convergence guarantee from a minimal number of samples for that class of algorithms, showing that the method attains a local quadratic convergence rate. Furthermore, we show that the linear systems to be solved are well-conditioned even for very ill-conditioned ground truth matrices. We provide extensive experiments, indicating that unlike many state-of-the-art approaches, our method is able to complete very ill-conditioned matrices with a condition number of up to $10^{10}$ from few samples, while being competitive in its scalability. | ['Claudio Mayrink Verdun', 'Christian Kümmerle'] | 2021-06-03 | null | null | null | null | ['low-rank-matrix-completion'] | ['methodology'] | [ 2.10270748e-01 2.49566764e-01 -2.02393383e-02 -1.29203618e-01
-1.45709789e+00 -4.41243649e-01 4.14975554e-01 -1.59827083e-01
-4.53873247e-01 7.94920325e-01 2.07189143e-01 -2.79445410e-01
-4.81769472e-01 -2.10119560e-01 -8.17934752e-01 -6.81137025e-01
-3.47560912e-01 7.86172867e-01 -3.00708022e-02 -4.06065613e-01
3.20088953e-01 3.87174457e-01 -1.09708738e+00 -5.90610392e-02
9.53257263e-01 9.87577677e-01 7.39689916e-02 6.17584586e-01
4.77946371e-01 6.23571932e-01 3.05633783e-01 -5.38868196e-02
6.90750301e-01 -2.12313667e-01 -9.05596197e-01 1.12617008e-01
8.05836022e-01 -2.13042930e-01 -2.73587823e-01 1.14613211e+00
4.06603366e-01 3.60426515e-01 5.22384465e-01 -7.23338068e-01
-3.07170570e-01 2.73529589e-01 -7.38319159e-01 -1.23442084e-01
6.30004585e-01 -2.35376507e-01 1.25030971e+00 -1.64343119e+00
6.68619871e-01 1.26218903e+00 9.81223583e-01 1.70294270e-01
-1.74398506e+00 -2.52923459e-01 1.15488216e-01 -1.95285603e-02
-1.61384702e+00 -8.10542226e-01 4.15197223e-01 -5.70839167e-01
5.69227397e-01 5.42813659e-01 3.17573667e-01 5.59796989e-01
-2.42043123e-01 7.81729758e-01 1.30645716e+00 -4.12517875e-01
3.37769270e-01 -5.97575419e-02 2.54301220e-01 9.22471166e-01
3.59973341e-01 1.60967335e-01 -6.67111695e-01 -6.21766329e-01
7.84713387e-01 -4.83380556e-01 -3.37572366e-01 -6.97076440e-01
-1.68615377e+00 9.14776206e-01 3.80405664e-01 3.97074111e-02
-4.51601982e-01 3.96045774e-01 1.57285228e-01 2.76495010e-01
5.90827346e-01 3.40913504e-01 -3.94661069e-01 1.03540644e-01
-1.26568198e+00 3.35686594e-01 9.25891161e-01 8.79321039e-01
8.80554140e-01 2.45810196e-01 -4.05092910e-02 6.97268665e-01
4.92175400e-01 9.83349204e-01 -7.80120865e-02 -1.14401245e+00
4.52381164e-01 1.22750029e-01 3.90391529e-01 -1.30049694e+00
-5.96368253e-01 -4.52262461e-01 -9.52013910e-01 2.48871222e-01
5.50913811e-01 -2.98246201e-02 -4.99525100e-01 1.61852992e+00
3.73259932e-01 3.94020706e-01 -1.90488175e-01 1.19303799e+00
3.96435410e-01 5.09108782e-01 -5.48567712e-01 -5.87866724e-01
1.18499005e+00 -7.77733207e-01 -6.50843203e-01 -5.38785636e-01
5.02267182e-01 -8.52748930e-01 9.76739228e-01 8.01737726e-01
-1.32239521e+00 -1.79573268e-01 -9.80763495e-01 -2.31373589e-02
2.80570775e-01 3.42880338e-01 8.04341495e-01 4.77678597e-01
-1.10961294e+00 9.41520751e-01 -8.27926636e-01 -2.58969590e-02
1.80647299e-02 5.57490706e-01 -6.49664164e-01 -2.38566816e-01
-6.25883460e-01 6.89092398e-01 5.10864928e-02 4.76104379e-01
-9.79409158e-01 -8.45022559e-01 -6.51240706e-01 -2.38344461e-01
5.25805175e-01 -6.72556758e-01 9.07636583e-01 -6.75514162e-01
-1.39466321e+00 8.35953891e-01 -5.94217300e-01 -3.07418883e-01
7.42022276e-01 -5.29474139e-01 -1.56366006e-01 1.14978865e-01
8.08786750e-02 1.82076059e-02 9.99875963e-01 -1.36464274e+00
-3.61515790e-01 -3.51370871e-01 -3.44542950e-01 5.42352647e-02
-1.88107625e-01 3.61798666e-02 -4.85933483e-01 -6.37844503e-01
8.14659536e-01 -1.29839122e+00 -9.11266208e-01 6.82023466e-02
-4.38928306e-01 2.22046271e-01 2.80055672e-01 -8.34932745e-01
1.28856671e+00 -1.99564588e+00 5.56235254e-01 6.48278952e-01
3.34090769e-01 -8.74767751e-02 -2.83919305e-01 5.84871352e-01
-2.13071972e-01 -1.58881247e-01 -5.29724479e-01 -4.13684100e-01
-1.36779204e-01 -3.07838898e-03 -2.83690900e-01 1.07152629e+00
-1.63478315e-01 5.29882908e-01 -1.21600330e+00 -2.01428935e-01
-6.70367759e-03 3.89546543e-01 -6.29021466e-01 -8.43806788e-02
1.49983093e-01 4.06145394e-01 -3.09253395e-01 3.72080445e-01
8.03142369e-01 -4.28278774e-01 3.11270922e-01 -5.78560889e-01
-6.46057725e-02 -9.94397625e-02 -2.09366846e+00 1.84582484e+00
-2.41013318e-01 4.59601671e-01 7.21222639e-01 -1.07589591e+00
6.81239724e-01 2.69615799e-01 6.45247519e-01 -2.08448887e-01
-2.45504268e-02 7.57432818e-01 -4.43095714e-01 -1.36786282e-01
5.05684555e-01 -2.57504374e-01 1.24490485e-01 2.49556527e-01
-1.88071325e-01 -2.65595093e-02 4.73447412e-01 4.06039685e-01
1.12889612e+00 -5.63364401e-02 2.40999565e-01 -8.64749551e-01
7.81163514e-01 -2.19699908e-02 6.08251095e-01 8.63835633e-01
2.59671539e-01 7.45095968e-01 2.06307650e-01 -3.78344327e-01
-1.08159661e+00 -1.13798845e+00 -2.14147523e-01 9.30940509e-01
6.57103583e-03 -5.47327995e-01 -4.61720526e-01 -1.74874112e-01
-5.10597937e-02 2.00196669e-01 -3.39938492e-01 2.73680001e-01
-5.68741202e-01 -8.94470215e-01 1.35175496e-01 2.33953953e-01
1.62102640e-01 -2.59938627e-01 2.50400305e-01 3.14698100e-01
-1.24450110e-01 -1.12781215e+00 -4.77577329e-01 -2.15049200e-02
-1.23338354e+00 -1.09694004e+00 -7.83267438e-01 -6.88372791e-01
1.13297081e+00 3.65051895e-01 1.10636365e+00 -6.92266449e-02
-8.35326314e-02 2.68473476e-01 -4.11218666e-02 3.20160568e-01
-1.84440240e-01 -2.82640040e-01 6.40536189e-01 2.41365746e-01
-1.61286011e-01 -6.03154361e-01 -6.16155982e-01 3.91442031e-01
-7.02807546e-01 1.34012569e-02 5.15143812e-01 1.09898937e+00
9.79340911e-01 -1.72329202e-01 1.57172948e-01 -1.18570125e+00
6.72240376e-01 -2.71787614e-01 -9.44560707e-01 1.48843482e-01
-8.82279456e-01 3.49472135e-01 6.82495832e-01 -2.51021683e-01
-7.14573205e-01 6.58403933e-01 2.34118223e-01 -3.35817069e-01
6.19169891e-01 8.55555058e-01 2.24434674e-01 -5.76136947e-01
1.17969060e+00 1.33525044e-01 1.19655281e-01 -6.08559728e-01
8.79110694e-01 3.11210990e-01 7.75730610e-01 -9.12494302e-01
1.34505451e+00 1.00031865e+00 4.16353911e-01 -8.41765881e-01
-9.99577403e-01 -8.37083280e-01 -6.55946612e-01 -1.23329930e-01
2.48897225e-01 -1.17106068e+00 -7.33462095e-01 1.11568481e-01
-6.39063120e-01 -2.84521133e-01 -2.24215135e-01 6.38538063e-01
-9.02951419e-01 7.72530198e-01 -6.88484132e-01 -9.01745379e-01
-2.07097620e-01 -8.87989879e-01 1.06909227e+00 -4.40293312e-01
-1.21748790e-01 -9.64550972e-01 3.56535643e-01 3.65543246e-01
2.13740304e-01 3.34995240e-01 3.22107285e-01 -2.29942545e-01
-6.01175308e-01 -5.18054903e-01 -3.81409228e-01 3.07626575e-01
-2.24551454e-01 -1.88977867e-01 -5.54775357e-01 -9.65100825e-01
-1.58680141e-01 -3.03706706e-01 8.66403699e-01 3.07280600e-01
5.73490202e-01 -5.39655924e-01 -2.94406563e-01 6.98612809e-01
1.70747256e+00 -6.64489746e-01 5.77026844e-01 4.10046019e-02
7.20990360e-01 4.10196543e-01 7.90792644e-01 6.58307374e-01
7.67915547e-02 5.81209838e-01 2.52652049e-01 -3.04495543e-01
1.55878082e-01 2.43819244e-02 2.79801399e-01 1.11289430e+00
-3.55190396e-01 5.68458974e-01 -1.01066065e+00 6.33816838e-01
-2.39426279e+00 -8.73917282e-01 -8.42322588e-01 2.72442555e+00
8.64120841e-01 -2.01914936e-01 1.02004707e-01 2.42746353e-01
4.46922392e-01 1.08160816e-01 -3.95907402e-01 -4.06863354e-02
-2.21382201e-01 1.55943111e-01 9.57291722e-01 9.22283888e-01
-1.03250289e+00 9.34763014e-01 7.84240913e+00 9.59512711e-01
-5.41818142e-01 8.03907737e-02 1.58032700e-01 -1.68211415e-01
-3.78645957e-01 3.27206343e-01 -5.45935810e-01 1.23648206e-02
7.58679271e-01 -1.78655297e-01 1.01334703e+00 9.53477621e-01
4.91366655e-01 2.11794116e-02 -1.05681288e+00 1.13909638e+00
4.35829833e-02 -1.33398056e+00 -3.62924606e-01 2.12955460e-01
1.27316749e+00 2.30561048e-01 -2.72393692e-02 -1.08754151e-01
5.04984677e-01 -1.03925455e+00 5.41604459e-01 7.44567871e-01
8.73292089e-01 -5.52881122e-01 3.76453191e-01 3.10713202e-01
-1.24613678e+00 -7.28180930e-02 -5.25694847e-01 -8.25015828e-02
3.84153396e-01 1.23923481e+00 -3.48592132e-01 6.06809258e-01
5.44457138e-01 5.78127444e-01 -2.71512300e-01 1.08795464e+00
-1.66259751e-01 5.67267239e-01 -8.30717921e-01 3.74117851e-01
2.09982738e-01 -6.15976751e-01 7.51802266e-01 1.02267599e+00
3.34914476e-01 3.14471871e-01 4.56900388e-01 6.02090538e-01
1.37999952e-01 3.60088646e-01 -3.73268455e-01 1.32066071e-01
2.74420381e-01 1.31485820e+00 -4.81612086e-01 -2.82385111e-01
-3.41676593e-01 9.88104224e-01 3.92135620e-01 4.87173438e-01
-4.46043849e-01 -1.47249639e-01 3.72471064e-01 1.18812315e-01
2.62932450e-01 -6.17372811e-01 -4.05202061e-01 -1.43124652e+00
2.22277194e-01 -1.17500293e+00 3.28221858e-01 -6.12275422e-01
-1.42588139e+00 3.09809536e-01 -2.69724935e-01 -1.36603379e+00
-2.26997748e-01 -7.54688382e-01 -1.98985800e-01 7.56277263e-01
-1.21102667e+00 -9.17169213e-01 -1.77415788e-01 6.18998647e-01
-7.41681755e-02 -8.70244503e-02 6.98460400e-01 4.25288856e-01
-4.82197046e-01 4.57984090e-01 6.98690951e-01 -1.86076954e-01
6.53986096e-01 -1.28604901e+00 4.01688144e-02 1.04664135e+00
6.12919815e-02 7.66646624e-01 1.03276789e+00 -5.32784224e-01
-1.90051007e+00 -6.32282734e-01 8.57167244e-01 -2.29387492e-01
1.00430346e+00 -2.01281965e-01 -7.04034388e-01 6.71090841e-01
-3.31188023e-01 3.33182931e-01 5.29130995e-01 5.05361199e-01
-3.72262269e-01 -1.77406847e-01 -9.02225137e-01 6.82541847e-01
1.17631412e+00 -5.60706496e-01 -3.68324935e-01 9.24952209e-01
1.09909765e-01 -5.87816238e-01 -1.03997469e+00 5.21141052e-01
3.98327738e-01 -6.34864509e-01 1.36141539e+00 -5.84785700e-01
1.92731038e-01 -6.27455413e-01 -3.99542689e-01 -9.04002666e-01
-6.87644720e-01 -1.15776026e+00 -4.36437964e-01 7.41936445e-01
5.20328224e-01 -3.91191781e-01 9.66158211e-01 5.91262460e-01
3.92488651e-02 -6.47151589e-01 -1.07588041e+00 -1.08375132e+00
-1.00760996e-01 -5.79779565e-01 -1.48858026e-01 9.55984473e-01
1.04090549e-01 4.20511603e-01 -8.52547109e-01 3.90419424e-01
1.12627339e+00 3.46783921e-02 8.81011069e-01 -1.25243890e+00
-5.18273652e-01 -1.76606864e-01 -1.92910388e-01 -1.35402250e+00
-4.74617060e-04 -1.04190481e+00 1.81485057e-01 -1.46942258e+00
2.19255432e-01 -6.05424881e-01 -2.49535590e-01 2.83491969e-01
-1.70184970e-01 5.16909897e-01 8.65444243e-02 3.40396583e-01
-7.00031579e-01 4.03054595e-01 1.02555275e+00 -2.10895110e-02
-1.54641241e-01 -2.09368065e-01 -6.39275074e-01 7.28603065e-01
2.24289387e-01 -4.78204876e-01 -1.83073133e-01 -2.63923824e-01
8.31251085e-01 3.15563828e-01 2.64449894e-01 -1.01054025e+00
4.01428968e-01 -7.97558725e-02 -4.06144373e-02 -5.38147449e-01
3.68679702e-01 -6.40624762e-01 3.50104213e-01 3.84333044e-01
-1.90392137e-01 -2.26040468e-01 -1.12405427e-01 7.92530894e-01
-7.80290514e-02 -2.57049024e-01 7.88316131e-01 -3.93183082e-02
-4.96033728e-01 3.92478287e-01 -2.97057927e-01 1.47101209e-01
5.45098424e-01 -1.45866707e-01 9.55485031e-02 -5.24415195e-01
-8.77278507e-01 3.91488105e-01 4.59109962e-01 -2.30065361e-01
4.84640867e-01 -1.57002330e+00 -1.10491836e+00 -4.03862670e-02
4.52494361e-02 -4.55845967e-02 1.67227328e-01 1.15899003e+00
-6.19160831e-01 2.65706182e-01 4.48433042e-01 -6.49443746e-01
-1.17305684e+00 5.28156638e-01 6.30084500e-02 -5.44679642e-01
-5.47642529e-01 8.89024734e-01 -1.67864636e-01 -6.44738257e-01
1.53506678e-02 9.20099318e-02 3.60711157e-01 -5.67095205e-02
6.71942830e-01 8.76052618e-01 9.32982937e-02 -6.61872625e-01
-4.40238118e-01 8.20447505e-01 2.03129947e-01 -4.09061849e-01
1.39969528e+00 -1.38176650e-01 -3.80738646e-01 2.71907449e-01
1.27886891e+00 4.54419047e-01 -1.33651960e+00 -5.45331836e-01
1.11803554e-01 -6.66454852e-01 3.84163648e-01 -3.81068200e-01
-1.05562353e+00 3.10871214e-01 4.03751791e-01 -5.31834215e-02
9.22387540e-01 -2.10371330e-01 6.00037932e-01 9.34806108e-01
5.83236635e-01 -1.23502958e+00 -2.92652790e-02 5.65724432e-01
1.13527930e+00 -1.30086374e+00 7.71333158e-01 -7.08157778e-01
-1.96389601e-01 1.01079762e+00 -1.35582998e-01 -6.80481791e-01
6.51182652e-01 2.73240972e-02 -1.12642534e-01 -2.21597329e-01
-4.21393186e-01 -1.80493757e-01 6.52941942e-01 3.87170762e-01
2.94252574e-01 -1.42281055e-02 -6.20341480e-01 1.77224040e-01
-2.39591286e-01 -9.19157788e-02 4.22251880e-01 5.79921782e-01
-4.18871999e-01 -1.04996026e+00 -6.08577430e-01 4.53145176e-01
-2.00499579e-01 -2.73139924e-01 -1.43506154e-01 5.62486649e-01
-6.53448522e-01 1.04263926e+00 -4.86109495e-01 -1.93728626e-01
3.89542818e-01 -2.95267522e-01 4.93991554e-01 -4.32642817e-01
-3.18546981e-01 2.86363423e-01 3.89455736e-01 -9.96117234e-01
-3.94486964e-01 -7.80206800e-01 -1.10918093e+00 -5.85555732e-01
-5.15188038e-01 1.36223286e-01 5.66661000e-01 8.69536638e-01
2.22202852e-01 -5.82296960e-02 7.94823289e-01 -1.00787437e+00
-7.30019867e-01 -5.89885175e-01 -8.11502099e-01 4.37567443e-01
2.54553437e-01 -3.83197814e-01 -6.31485999e-01 -6.91502094e-02] | [6.994635581970215, 4.597374439239502] |
2dfa78e6-3e46-48c8-8fca-4ff9721b456b | align-smatch-a-novel-evaluation-method-for | null | null | https://aclanthology.org/2022.lrec-1.638 | https://aclanthology.org/2022.lrec-1.638.pdf | Align-smatch: A Novel Evaluation Method for Chinese Abstract Meaning Representation Parsing based on Alignment of Concept and Relation | Abstract Meaning Representation is a sentence-level meaning representation, which abstracts the meaning of sentences into a rooted acyclic directed graph. With the continuous expansion of Chinese AMR corpus, more and more scholars have developed parsing systems to automatically parse sentences into Chinese AMR. However, the current parsers can’t deal with concept alignment and relation alignment, let alone the evaluation methods for AMR parsing. Therefore, to make up for the vacancy of Chinese AMR parsing evaluation methods, based on AMR evaluation metric smatch, we have improved the algorithm of generating triples so that to make it compatible with concept alignment and relation alignment. Finally, we obtain a new integrity metric align-smatch for paring evaluation. A comparative research then was conducted on 20 manually annotated AMR and gold AMR, with the result that align-smatch works well in alignments and more robust in evaluating arcs. We also put forward some fine-grained metric for evaluating concept alignment, relation alignment and implicit concepts, in order to further measure parsers’ performance in subtasks. | ['Weiguang Qu', 'Junsheng Zhou', 'Minxuan Feng', 'Kairui Huo', 'Zhixing Xu', 'Bin Li', 'Liming Xiao'] | null | null | null | null | lrec-2022-6 | ['concept-alignment', 'amr-parsing'] | ['computer-vision', 'natural-language-processing'] | [ 4.18567151e-01 3.58851969e-01 -4.36671339e-02 -6.86095893e-01
-6.01365328e-01 -5.56836307e-01 3.73814732e-01 5.23251414e-01
-2.11577788e-01 5.72096765e-01 6.23741329e-01 -6.23323619e-01
8.57715160e-02 -1.11058295e+00 -1.09613568e-01 -1.59416795e-01
3.37942809e-01 3.83151323e-01 1.83559045e-01 -4.53333259e-01
5.21043837e-01 1.02324396e-01 -9.81124520e-01 4.97018039e-01
9.13490117e-01 3.66778761e-01 2.19676390e-01 5.26061356e-01
-9.98453140e-01 9.33294773e-01 -8.23325515e-01 -8.45360875e-01
-2.60021299e-01 -7.13616014e-01 -1.47398996e+00 -1.46731779e-01
-8.18192884e-02 3.70406896e-01 7.10986601e-03 1.06787097e+00
9.58308578e-02 -1.45984873e-01 3.41627628e-01 -9.36653614e-01
-1.03026962e+00 1.60020638e+00 -5.14892519e-01 1.98030844e-01
9.77204204e-01 -2.54504979e-01 1.39900982e+00 -4.98517215e-01
6.17271602e-01 1.49364138e+00 2.61295080e-01 6.25444174e-01
-6.88723683e-01 -4.28039193e-01 5.21589041e-01 9.35548693e-02
-1.12942505e+00 7.07863420e-02 4.78183240e-01 -4.40845490e-02
1.32477605e+00 4.99770552e-01 6.18484855e-01 7.97601223e-01
2.18493462e-01 5.79179287e-01 1.00548422e+00 -8.37511182e-01
7.79203000e-03 -1.62237629e-01 7.83657372e-01 3.60773534e-01
4.52283591e-01 -6.99628234e-01 -3.37867767e-01 2.07584888e-01
5.21870792e-01 -4.97043759e-01 1.15923427e-01 4.78649020e-01
-1.33937013e+00 7.51766860e-01 1.96921065e-01 9.04812574e-01
3.42888646e-02 -5.31295091e-02 3.15275550e-01 2.97843784e-01
2.20369682e-01 6.57459080e-01 -3.42039645e-01 -3.35807562e-01
-1.89306021e-01 -6.24627359e-02 8.45774829e-01 1.36380804e+00
6.54894412e-01 -4.05465588e-02 -2.48080805e-01 7.02680111e-01
3.46497923e-01 4.09649700e-01 6.62479103e-01 -6.90758288e-01
8.92789245e-01 1.14903295e+00 -4.30968910e-01 -9.71862793e-01
-3.12416583e-01 -1.33194104e-01 -7.44111359e-01 -6.44443572e-01
6.23319559e-02 -7.83752725e-02 -7.61814058e-01 1.54861200e+00
1.08696528e-01 -3.43693465e-01 5.45824349e-01 7.16018200e-01
1.11111081e+00 7.41560161e-01 3.19629371e-01 -3.07574272e-01
1.79311347e+00 -8.34093094e-01 -1.03681993e+00 -4.37359899e-01
1.26117516e+00 -8.25363576e-01 1.31361258e+00 1.07506245e-01
-9.83607650e-01 -5.11464119e-01 -1.21837580e+00 -2.07783803e-01
-4.48857099e-01 -1.18939362e-01 9.21593785e-01 8.00998092e-01
-8.01950395e-01 3.30467254e-01 -5.08579791e-01 -2.76687860e-01
-2.75147166e-02 5.57000078e-02 -5.35027623e-01 -1.42655507e-01
-1.38232505e+00 1.09456670e+00 8.44314575e-01 3.86399835e-01
-4.37689498e-02 3.08234002e-02 -1.00696599e+00 1.92397181e-02
3.06462735e-01 -6.45296216e-01 1.17900467e+00 -9.06207085e-01
-1.34036469e+00 9.12587404e-01 -1.73910946e-01 -3.12628925e-01
-2.46365249e-01 -3.14818919e-01 -8.70378256e-01 -4.85070199e-02
2.64095247e-01 4.35944438e-01 -6.19624294e-02 -9.88675177e-01
-7.35149682e-01 -4.50580180e-01 2.69229949e-01 4.80633497e-01
-1.53168887e-01 3.66278797e-01 -3.28623325e-01 -5.89945436e-01
7.04627395e-01 -6.06485903e-01 -3.75880539e-01 -1.29291403e+00
-3.99164885e-01 -5.64567029e-01 4.44494635e-01 -6.78840041e-01
1.64501560e+00 -1.97664917e+00 1.64163336e-01 2.42257804e-01
1.38598859e-01 6.24712706e-02 -3.56403291e-01 6.71232045e-01
-2.87731588e-01 6.30906343e-01 -5.54118335e-01 -2.60051303e-02
-7.52237663e-02 6.99765801e-01 -2.14703351e-01 -3.26814651e-01
5.39012849e-01 1.00500083e+00 -9.27875578e-01 -8.28527510e-01
-2.21093595e-01 -1.25383645e-01 -1.75776929e-01 1.78417861e-01
-3.01988900e-01 3.83659303e-01 -8.81166041e-01 6.35397434e-01
4.99710709e-01 6.52061254e-02 5.78638196e-01 2.81565517e-01
-7.65243871e-03 5.89636445e-01 -1.11707532e+00 1.93650746e+00
-3.98964971e-01 4.03038561e-01 -4.26251322e-01 -1.02003610e+00
1.45965338e+00 3.65610957e-01 6.99100494e-02 -8.72871220e-01
-6.88170257e-04 3.05470139e-01 3.12652081e-01 -5.04917383e-01
9.50706661e-01 -1.12937056e-02 -5.18186808e-01 2.10897699e-01
-2.38483518e-01 -2.81518042e-01 3.99439096e-01 4.55197632e-01
1.17622674e+00 2.62482017e-01 4.87853140e-01 -2.71289438e-01
9.56721306e-01 4.08416241e-01 6.37900770e-01 3.67327005e-01
3.44354451e-01 5.60003817e-01 6.89980984e-01 -5.11325598e-01
-7.36856341e-01 -8.89084637e-01 -7.18364045e-02 7.27687061e-01
1.92513421e-01 -8.27216387e-01 -9.72826183e-01 -7.68366992e-01
-6.57884181e-01 1.19258451e+00 -1.91346169e-01 9.69696045e-02
-1.19450605e+00 -7.59068072e-01 5.91902494e-01 6.84096277e-01
5.09331584e-01 -1.38178647e+00 -2.06696615e-01 5.38926601e-01
-6.20941043e-01 -1.49104559e+00 -2.83887959e-03 -1.02124713e-01
-8.94018233e-01 -1.09130526e+00 -4.14906349e-03 -9.95102406e-01
6.70777977e-01 1.03885807e-01 1.46221483e+00 4.51858729e-01
4.29916419e-02 1.47599772e-01 -1.19625700e+00 -3.96834493e-01
-7.24282265e-01 2.20252186e-01 -3.81920248e-01 -5.99275231e-01
8.01486015e-01 -3.91284794e-01 -1.61342442e-01 2.43534878e-01
-8.35858166e-01 2.18049839e-01 8.00745964e-01 3.18776220e-01
5.29145181e-01 -1.81244630e-02 6.27882719e-01 -1.15586340e+00
8.63647521e-01 -1.54717445e-01 -2.21398070e-01 6.30364180e-01
-5.88200510e-01 2.66217887e-01 4.90589678e-01 7.32037872e-02
-1.28494358e+00 -4.00150388e-01 -6.98696673e-01 5.71044087e-01
-1.41098291e-01 6.98064923e-01 -5.71823478e-01 5.42593598e-01
4.59874809e-01 9.24979076e-02 -3.24209243e-01 -2.68466413e-01
5.47851861e-01 9.30459738e-01 5.50221324e-01 -9.48652685e-01
6.22584581e-01 -1.85972571e-01 5.60504012e-02 -6.20249271e-01
-9.44241881e-01 -4.24257964e-01 -6.84960246e-01 1.94208235e-01
1.39896464e+00 -7.43946433e-01 -3.02316070e-01 5.00287116e-02
-1.34908974e+00 8.85758102e-02 -2.24834919e-01 5.91645420e-01
-3.33146334e-01 7.72420287e-01 -7.07568884e-01 -6.70333743e-01
-5.83786905e-01 -9.35846031e-01 8.40714157e-01 1.65695399e-01
-3.12799096e-01 -7.35000968e-01 -1.42605314e-02 4.17612076e-01
1.14027105e-01 3.08903366e-01 1.30310619e+00 -8.47949445e-01
-3.93661559e-01 -2.42619872e-01 -2.52064079e-01 3.14678997e-01
1.11167066e-01 9.14919004e-03 -5.62324405e-01 2.44364873e-01
-4.18155901e-02 2.97445301e-02 4.25049812e-01 -6.31144270e-02
1.02097785e+00 -4.00208950e-01 -1.73105985e-01 -1.05161667e-02
1.14779294e+00 4.80268270e-01 1.17551637e+00 6.24890804e-01
7.29380250e-01 7.11283088e-01 1.12096095e+00 -2.75691628e-01
5.90005338e-01 4.12956804e-01 2.36629650e-01 1.14829473e-01
-4.00102660e-02 -3.95952791e-01 2.73864776e-01 1.46613705e+00
-3.10452640e-01 -2.25509390e-01 -9.44496393e-01 1.86476499e-01
-1.84422672e+00 -6.39309704e-01 -7.09098577e-01 1.96073377e+00
8.52953851e-01 4.07159030e-01 -3.19967389e-01 1.46420389e-01
7.63329983e-01 -6.72368333e-02 2.65040040e-01 -9.68624830e-01
-3.33266377e-01 2.95005590e-01 3.19308378e-02 5.05174041e-01
-5.99324584e-01 1.33527708e+00 5.86583424e+00 5.43080688e-01
-4.66047853e-01 -1.16948344e-01 4.96985435e-01 8.83494377e-01
-8.62782896e-01 5.30612469e-01 -9.86699522e-01 1.15822330e-01
9.25084531e-01 2.37504058e-02 1.18453510e-01 7.19337106e-01
-1.44789547e-01 3.54187051e-03 -1.03526735e+00 6.56332076e-01
-1.41213477e-01 -1.14668524e+00 3.20264965e-01 -2.14251995e-01
2.65076339e-01 -3.97451669e-01 -6.01166487e-01 4.11988497e-01
2.67782509e-01 -1.08751905e+00 5.04315078e-01 1.54310763e-01
5.33438623e-01 -7.38649309e-01 1.24062109e+00 1.61299974e-01
-1.61260700e+00 1.60003915e-01 -6.78359270e-01 -2.67464101e-01
2.41887420e-01 3.01346421e-01 -6.52877092e-01 1.28202462e+00
4.35869038e-01 5.78352153e-01 -8.08846593e-01 3.26645732e-01
-7.86490858e-01 4.63126719e-01 1.79160550e-01 -4.47737306e-01
3.06446910e-01 -5.44496536e-01 4.26999211e-01 1.52536142e+00
2.31982723e-01 4.76365298e-01 3.26160491e-02 6.50674343e-01
2.16449335e-01 5.23987055e-01 -4.76538688e-01 -3.71861696e-01
7.36806989e-01 1.39915669e+00 -9.09137666e-01 -4.05472308e-01
-5.05570948e-01 5.96896827e-01 3.78947020e-01 8.08835700e-02
-6.32323086e-01 -4.66134995e-01 3.10703993e-01 -8.58100131e-02
-2.57604599e-01 -5.56955993e-01 -4.93900657e-01 -1.14587212e+00
2.95204043e-01 -9.14733469e-01 5.51290929e-01 -8.63977134e-01
-1.12491167e+00 9.43003774e-01 3.50589216e-01 -9.74093080e-01
-2.03064248e-01 -6.42669678e-01 -9.07722831e-01 9.03713524e-01
-1.15427780e+00 -1.22367275e+00 -2.23256901e-01 1.19552121e-01
6.12338305e-01 -1.14606991e-01 1.19643533e+00 -6.66065887e-02
-7.19350219e-01 3.68769288e-01 -8.96869123e-01 5.05355597e-01
1.80359498e-01 -1.53337574e+00 9.56339180e-01 1.33629870e+00
4.48191583e-01 9.89414930e-01 6.00976527e-01 -5.95083475e-01
-1.30063069e+00 -7.56635368e-01 1.23888302e+00 -4.99680072e-01
6.78613544e-01 -2.92469233e-01 -1.03291821e+00 8.79481673e-01
4.09959406e-01 -5.91420770e-01 7.43626595e-01 5.01584232e-01
-3.92720908e-01 -7.82605335e-02 -8.06694150e-01 6.57253265e-01
1.34019816e+00 -1.01437353e-01 -1.36686325e+00 2.29374334e-01
1.37720406e+00 -3.27210993e-01 -1.12977636e+00 4.94965643e-01
2.95607578e-02 -6.03608191e-01 5.63143253e-01 -6.92229092e-01
6.01691604e-01 -4.75870401e-01 -2.46843681e-01 -9.07505572e-01
-2.77069598e-01 -4.13226783e-01 3.08982879e-01 1.58642161e+00
8.19731474e-01 -7.55506933e-01 5.49679041e-01 5.10864019e-01
-5.83409369e-01 -4.79940623e-01 -6.91835165e-01 -5.54402769e-01
7.64591917e-02 -5.23113668e-01 8.21157873e-01 9.17017519e-01
4.61189240e-01 1.02138162e+00 2.86029875e-01 7.10018426e-02
3.11856866e-01 2.58139908e-01 5.09512961e-01 -9.86342132e-01
-1.28331810e-01 -2.37366185e-01 -4.39440966e-01 -7.79537022e-01
1.84177682e-01 -9.31235790e-01 -1.60236627e-01 -2.20919108e+00
1.71867996e-01 -5.23212135e-01 1.57562494e-02 7.72754133e-01
-4.11562383e-01 -2.18102276e-01 9.79037359e-02 1.47664577e-01
-4.40085709e-01 2.08641440e-01 1.36469543e+00 6.10230453e-02
-8.97720829e-02 -1.26853585e-01 -9.39050376e-01 4.86900300e-01
9.51907098e-01 -4.75169986e-01 -4.91286069e-01 -6.73974395e-01
6.55977428e-01 1.41436353e-01 -3.40609163e-01 -8.55945528e-01
-4.98610176e-03 -2.82025576e-01 1.15725838e-01 -6.82182789e-01
-1.52654231e-01 -5.94678164e-01 1.08203486e-01 3.35171551e-01
-2.92051345e-01 7.67498970e-01 -1.02520041e-01 1.22577652e-01
-4.84968990e-01 -6.92638695e-01 2.33312443e-01 -3.16620082e-01
-1.02545857e+00 -1.02977894e-01 -3.08446109e-01 2.53781229e-01
7.98104703e-01 -4.70409811e-01 -5.29740095e-01 5.63703887e-02
-4.91126835e-01 4.30898994e-01 1.21981144e-01 6.84112310e-01
7.96232998e-01 -1.03962016e+00 -8.39039445e-01 -4.72285189e-02
2.12060660e-01 3.75203609e-01 2.05368595e-03 3.90879840e-01
-7.27122784e-01 4.05116856e-01 -1.59441128e-01 -3.58541191e-01
-1.33937883e+00 4.92134750e-01 -5.83901554e-02 -4.22905535e-01
-8.22799623e-01 4.42837000e-01 3.02102929e-03 -5.04447758e-01
-1.40627965e-01 -6.20605886e-01 -7.65832722e-01 -3.89571786e-01
4.92758751e-01 -1.05827581e-02 1.13073260e-01 -5.60369670e-01
-2.97602654e-01 5.77526748e-01 7.40549937e-02 -9.17029753e-02
9.74968672e-01 -2.47337699e-01 -6.92192554e-01 1.75831527e-01
6.25222683e-01 2.73799211e-01 -1.58348650e-01 1.28689483e-01
6.36256754e-01 -2.18912631e-01 -5.54211199e-01 -5.76015055e-01
-4.83210415e-01 5.93616426e-01 -2.12019458e-02 6.69286191e-01
1.16003668e+00 3.91051993e-02 8.01179647e-01 5.91024101e-01
4.38774377e-01 -1.02012944e+00 1.00405730e-01 8.08411777e-01
6.84709370e-01 -7.58593380e-01 -4.11745682e-02 -9.43332434e-01
-9.64854479e-01 1.36048353e+00 7.72991359e-01 2.31715113e-01
3.53372097e-02 2.38710225e-01 1.05274990e-01 -1.41051278e-01
-4.18715924e-01 -2.83127069e-01 1.47266369e-02 6.50685668e-01
8.54405403e-01 3.15435737e-01 -1.00974953e+00 7.59541810e-01
-7.58015692e-01 -4.66412485e-01 6.13015473e-01 9.22164977e-01
-8.04587483e-01 -1.70587635e+00 2.58172047e-03 7.60880485e-02
-6.11603796e-01 -2.79222280e-01 -6.00800812e-01 1.07443631e+00
-1.54467240e-01 1.28943551e+00 1.82006538e-01 -4.65399384e-01
5.49356341e-01 -2.86219511e-02 4.11342651e-01 -1.10400856e+00
-6.03295445e-01 -2.44602621e-01 8.29294145e-01 -2.80103117e-01
-6.13526642e-01 -2.96677977e-01 -1.90369296e+00 -1.76651821e-01
-3.73175621e-01 7.10967720e-01 6.36220932e-01 1.15098917e+00
-5.14025651e-02 6.38954222e-01 5.36521018e-01 1.41832680e-01
-3.45249712e-01 -1.23143423e+00 -2.58876503e-01 4.02496845e-01
-7.31574357e-01 8.30200762e-02 2.75110248e-02 -2.04534739e-01] | [10.408651351928711, 9.391386985778809] |
4ce4839d-cdb7-488a-8f52-0ce2af33afcd | a-deep-learning-search-for-technosignatures | 2301.12670 | null | https://arxiv.org/abs/2301.12670v1 | https://arxiv.org/pdf/2301.12670v1.pdf | A deep-learning search for technosignatures of 820 nearby stars | The goal of the Search for Extraterrestrial Intelligence (SETI) is to quantify the prevalence of technological life beyond Earth via their "technosignatures". One theorized technosignature is narrowband Doppler drifting radio signals. The principal challenge in conducting SETI in the radio domain is developing a generalized technique to reject human radio frequency interference (RFI). Here, we present the most comprehensive deep-learning based technosignature search to date, returning 8 promising ETI signals of interest for re-observation as part of the Breakthrough Listen initiative. The search comprises 820 unique targets observed with the Robert C. Byrd Green Bank Telescope, totaling over 480, hr of on-sky data. We implement a novel beta-Convolutional Variational Autoencoder to identify technosignature candidates in a semi-unsupervised manner while keeping the false positive rate manageably low. This new approach presents itself as a leading solution in accelerating SETI and other transient research into the age of data-driven astronomy. | ['S. Pete Worden', 'Sofia Z. Sheikh', 'Danny C. Price', 'Imke de Pater', 'David MacMahon', 'Matt Lebofsky', 'Howard Isaacson', 'John Hoang', 'Vishal Gajjar', 'Jamie Drew', 'Daniel Czech', 'Bryan Brzycki', 'Andrew P. V. Siemion', 'Steve Croft', 'Leandro Rizk', 'Cherry Ng', 'Peter Xiangyuan Ma'] | 2023-01-30 | null | null | null | null | ['astronomy'] | ['miscellaneous'] | [-1.78152062e-02 -9.43867937e-02 5.43555617e-02 -1.28291070e-01
-7.91188419e-01 -5.95596194e-01 1.04360259e+00 -7.49773562e-01
-4.83150899e-01 6.92710400e-01 1.98395297e-01 -3.86662424e-01
-5.87407291e-01 -4.53613907e-01 -4.50793803e-01 -8.92325580e-01
4.49652448e-02 9.36072469e-01 -2.49512017e-01 -3.75636667e-01
2.08086342e-01 8.25367093e-01 -1.50671029e+00 -1.10175595e-01
3.90391737e-01 1.20977926e+00 -4.39018793e-02 8.12510967e-01
2.99672782e-01 7.85100222e-01 -9.51255441e-01 1.20178059e-01
2.12145641e-01 -4.78361785e-01 -4.96270150e-01 -5.94038248e-01
5.39978817e-02 7.10891336e-02 -5.67071676e-01 8.72946680e-01
8.53957474e-01 3.54755163e-01 1.04459929e+00 -9.77607787e-01
-8.75798166e-02 2.80418038e-01 -1.19493097e-01 8.32219064e-01
1.67048112e-01 -1.13280311e-01 8.83825302e-01 -7.95758963e-01
4.88399416e-01 6.43768013e-01 8.42822969e-01 3.68338197e-01
-8.79627049e-01 -5.61066687e-01 -6.93033576e-01 -6.63448796e-02
-1.31928372e+00 -4.81521755e-01 7.70076156e-01 -6.10781968e-01
1.20492315e+00 4.83312070e-01 6.61351740e-01 1.60651052e+00
2.84205526e-01 4.07526493e-01 5.50405800e-01 -6.46369278e-01
4.62200373e-01 2.36533493e-01 -7.00313896e-02 6.24286294e-01
4.73339587e-01 7.61137962e-01 -8.52102518e-01 -4.74778414e-01
8.04172099e-01 -5.03615201e-01 -5.32190979e-01 -1.72667235e-01
-1.38097203e+00 1.05451882e+00 1.11263230e-01 9.26475286e-01
-3.71790409e-01 2.75042593e-01 1.75285950e-01 5.07659197e-01
5.48070133e-01 1.33565116e+00 -7.63144612e-01 -4.32052203e-02
-9.42152560e-01 1.18535124e-01 7.48141646e-01 5.68460703e-01
2.37941712e-01 7.61961222e-01 3.09383273e-01 6.03023350e-01
4.28307176e-01 6.62428021e-01 7.85379589e-01 -7.88393915e-01
-3.31432402e-01 -5.86081073e-02 3.24490398e-01 -7.04535544e-01
-9.18832541e-01 -1.29564571e+00 -5.30289590e-01 1.74356803e-01
-7.74073675e-02 -3.58861357e-01 -8.76894772e-01 1.26590168e+00
3.23981166e-01 1.06193729e-01 1.13421008e-01 1.10571396e+00
9.79833186e-01 6.17661834e-01 -3.32415938e-01 9.03705321e-03
1.34132493e+00 -3.73652428e-01 -2.00666711e-01 -3.91536921e-01
5.25215089e-01 -5.87493360e-01 1.66043237e-01 9.05996501e-01
-2.70603299e-01 -4.55225915e-01 -1.46270001e+00 2.51228988e-01
-3.38098139e-01 1.14664309e-01 1.27450287e+00 1.06288910e+00
-5.64248443e-01 5.75287521e-01 -5.47271788e-01 -1.70844555e-01
-4.84748818e-02 1.97323516e-01 -8.37908238e-02 5.74142814e-01
-1.13763416e+00 9.95521903e-01 5.44507444e-01 -1.49553344e-01
-1.20955873e+00 -7.98675656e-01 -4.11289722e-01 1.03318825e-01
5.89303002e-02 -7.15105116e-01 1.30274796e+00 -8.12224329e-01
-1.40281403e+00 1.03020632e+00 4.76080120e-01 -9.40627098e-01
2.80392736e-01 -1.68748483e-01 -1.14422655e+00 7.53810182e-02
-1.17062055e-01 2.72665292e-01 1.16862321e+00 -8.39366257e-01
-5.30204535e-01 -9.13428366e-02 -7.45804429e-01 -2.05487773e-01
-2.40278214e-01 2.21300811e-01 2.98004802e-02 -6.23534560e-01
4.59450573e-01 -7.31794894e-01 1.97384004e-02 -5.43509960e-01
-1.16674625e-03 -4.16647971e-01 4.41901565e-01 -5.83525658e-01
5.70565343e-01 -1.99706089e+00 3.34647715e-01 1.28497362e-01
4.41066623e-01 2.35299002e-02 2.10477784e-01 4.86133285e-02
-3.24901402e-01 -2.51961440e-01 2.62368247e-02 1.09367184e-01
-2.64148675e-02 -3.74862254e-01 -5.58261395e-01 8.27225685e-01
-1.08962364e-01 5.53564906e-01 -8.73359561e-01 3.34695429e-01
4.00792956e-01 3.90334576e-01 -2.79520273e-01 4.64941077e-02
-3.95153701e-01 5.84308863e-01 -5.51296234e-01 9.16228235e-01
5.77324867e-01 -1.23504452e-01 -9.81145948e-02 -1.84863374e-01
-5.02493501e-01 3.56187969e-01 -5.63273013e-01 1.54408801e+00
-1.35899737e-01 1.03789771e+00 3.29228900e-02 -1.20892596e+00
1.20033967e+00 2.13329926e-01 8.55118692e-01 -8.66100192e-01
4.81505990e-01 4.15903062e-01 -1.14441954e-01 -2.53614336e-01
6.15268230e-01 -6.46575034e-01 -1.98816523e-01 2.05194592e-01
7.22988427e-01 -2.69006431e-01 -5.70179522e-01 -2.76219216e-03
1.17155790e+00 2.25815490e-01 1.16532512e-01 -4.52995121e-01
1.56698003e-01 1.98254421e-01 5.07741094e-01 1.15631199e+00
-1.84247941e-01 4.03381646e-01 -8.03148970e-02 -6.62833273e-01
-1.03374553e+00 -9.10752356e-01 -4.03504878e-01 8.24602723e-01
-5.22786319e-01 3.71024720e-02 -1.52249644e-02 -1.80375993e-01
-7.33453706e-02 9.30896401e-01 -3.10988277e-01 -3.92084390e-01
-1.46410167e-01 -1.13921428e+00 8.51346672e-01 5.79753444e-02
2.84938812e-01 -8.88796508e-01 -8.00416946e-01 2.11952746e-01
-1.97574139e-01 -6.85020983e-01 4.57339436e-01 8.22037101e-01
-4.93871599e-01 -6.76213086e-01 -9.77996945e-01 -5.02670586e-01
-2.93478996e-01 3.86714906e-01 1.25520861e+00 -4.46004301e-01
-9.83015716e-01 4.92163539e-01 -3.24975878e-01 -8.00818563e-01
-2.43683547e-01 -2.25555841e-02 7.27813601e-01 -2.41356000e-01
5.14172077e-01 -3.99297088e-01 -3.70710075e-01 2.21841052e-01
-1.72585160e-01 -6.76627338e-01 4.75019664e-01 9.15360868e-01
4.12375063e-01 3.83813143e-01 6.56449318e-01 -7.99913481e-02
1.27657413e-01 -7.06757128e-01 -1.01898777e+00 -3.69561128e-02
-2.98109204e-01 1.73663184e-01 3.05179387e-01 -4.18226957e-01
-1.18393600e+00 -3.36039603e-01 -1.42155573e-01 -5.66588521e-01
6.57164827e-02 5.51943362e-01 2.59181321e-01 -6.05052054e-01
1.27332246e+00 1.98672786e-01 -3.31513643e-01 -4.21058178e-01
1.24517374e-01 6.96193218e-01 8.65131915e-01 -2.64548302e-01
7.65813708e-01 3.41020852e-01 -3.81724089e-02 -1.48077273e+00
-1.06376266e+00 -9.42525685e-01 1.05905071e-01 -6.16159201e-01
6.71070397e-01 -1.13340306e+00 -8.74406755e-01 1.60048142e-01
-7.48238802e-01 1.05921350e-01 -1.23170629e-01 1.00909317e+00
-4.75465775e-01 3.18565398e-01 7.22934976e-02 -1.21875679e+00
-4.88734186e-01 -4.91760999e-01 1.09603548e+00 3.48635495e-01
-1.17452510e-01 -6.11493170e-01 6.09944463e-01 4.03441280e-01
4.37369704e-01 2.23982215e-01 4.39509153e-01 -7.26502657e-01
-4.96235341e-01 -4.97152716e-01 5.95326647e-02 1.15039751e-01
-3.96799833e-01 -4.45663750e-01 -1.31463373e+00 -3.51551861e-01
1.04414666e+00 -4.84022319e-01 1.25900912e+00 5.13359666e-01
8.22632730e-01 3.44643712e-01 -3.48665535e-01 9.71394122e-01
1.28641891e+00 3.24580967e-01 4.93227005e-01 5.92879891e-01
2.14760318e-01 2.53029883e-01 5.91091812e-01 7.96411276e-01
-7.11750865e-01 2.57948071e-01 2.81954139e-01 2.08212018e-01
3.28657806e-01 1.49202883e-01 2.44582459e-01 4.07484889e-01
-2.75015086e-01 -5.40554464e-01 -8.89025629e-01 3.60105902e-01
-1.22202587e+00 -1.43326735e+00 -1.63148761e-01 2.33833623e+00
-1.56897068e-01 1.85389236e-01 -1.04036525e-01 4.71452773e-02
2.53675729e-01 1.68918893e-01 -5.05029082e-01 1.01091430e-01
-3.06371272e-01 4.49218065e-01 9.15126503e-01 1.60871491e-01
-1.20812309e+00 5.53666651e-01 7.10235262e+00 5.51569223e-01
-1.14416194e+00 9.30291489e-02 1.03437982e-01 -1.66072637e-01
-5.13575375e-01 -3.73086005e-01 -6.21893406e-01 1.89422056e-01
1.39743125e+00 1.34226596e-02 5.58367193e-01 1.14754629e+00
-3.23742121e-01 -1.05822735e-01 -6.62510693e-01 1.10643864e+00
2.46559605e-01 -1.62426114e+00 -7.46995032e-01 1.07122799e-02
1.59426630e-01 7.96827614e-01 9.44957659e-02 7.37084508e-01
4.35971171e-02 -8.38937521e-01 8.53298485e-01 8.31539750e-01
8.14692676e-01 -9.44467366e-01 6.36267304e-01 3.72615695e-01
-5.00823915e-01 -1.46851510e-01 -6.83298767e-01 1.30559236e-01
-1.29648700e-01 8.18093181e-01 -1.09324312e+00 6.85391128e-01
9.98327255e-01 9.18668211e-02 -7.71450922e-02 1.08714521e+00
1.66934937e-01 8.01276028e-01 -8.19525898e-01 -5.10512471e-01
1.41388774e-01 -1.31311283e-01 1.04973459e+00 8.84261966e-01
7.86730647e-01 1.23525541e-02 -3.17950219e-01 7.48711288e-01
-9.78504792e-02 -6.60688519e-01 -7.96358645e-01 -6.58568323e-01
3.44268233e-01 1.10604668e+00 -6.84053779e-01 -8.49223733e-02
-2.24165529e-01 8.33642781e-01 -2.05029398e-01 7.05282912e-02
-1.00850928e+00 -5.76347649e-01 5.65042257e-01 -4.07180488e-01
3.80936712e-01 -5.14823556e-01 -2.96042860e-01 -1.29040885e+00
-3.66828263e-01 -7.83859849e-01 4.44593996e-01 -1.12165451e+00
-1.06847155e+00 5.74881494e-01 -4.55344081e-01 -1.34460676e+00
-1.81347728e-01 -9.09229100e-01 -4.04316634e-01 7.66727626e-01
-8.00731301e-01 -5.73775589e-01 -2.91179836e-01 1.31820858e-01
2.87916899e-01 -1.05449772e+00 1.01783943e+00 1.46658778e-01
-1.63238779e-01 2.75075376e-01 7.18228400e-01 -4.80167329e-01
5.01684487e-01 -8.81527126e-01 3.40564519e-01 6.16399407e-01
4.52095509e-01 4.09957916e-01 1.41457999e+00 -7.27841735e-01
-1.74854100e+00 -7.28965640e-01 5.75788796e-01 -8.16788852e-01
8.66214335e-01 -4.81165081e-01 -6.19581580e-01 6.75776839e-01
-1.66620594e-02 -4.09289747e-01 6.01578951e-01 5.69584370e-01
-2.51199275e-01 2.58132160e-01 -1.16866732e+00 8.68581533e-02
7.15969503e-01 -6.75172269e-01 -8.04840147e-01 8.63570035e-01
7.32483029e-01 -3.51842083e-02 -6.94357157e-01 4.83884066e-01
4.70025629e-01 -8.20337713e-01 1.26161993e+00 -4.47618246e-01
-1.53805122e-01 -1.84939742e-01 -1.21175304e-01 -1.08809912e+00
-7.85676062e-01 -9.92947638e-01 -1.71644121e-01 5.76524079e-01
2.37118840e-01 -6.51287556e-01 9.98608410e-01 -1.81994125e-01
-5.15002310e-01 1.69300944e-01 -1.29342198e+00 -1.05373073e+00
-2.55218476e-01 -6.99560344e-01 6.69798106e-02 9.49018180e-01
-3.68420213e-01 5.34255683e-01 -8.38641644e-01 3.93283784e-01
9.11996305e-01 -8.72837901e-02 4.14211839e-01 -1.66890216e+00
-7.63754725e-01 -3.55939597e-01 -4.47510213e-01 -5.95919669e-01
-1.50722474e-01 -9.47444439e-01 1.80633143e-01 -8.45003188e-01
-5.05495965e-01 -1.47750452e-01 -4.76347208e-01 -1.75880224e-01
6.63994253e-01 -6.36407807e-02 -5.72962344e-01 4.39405106e-02
-2.35274464e-01 7.97753155e-01 2.20509440e-01 -3.80593270e-01
-4.91818190e-02 1.26419663e-01 -5.81480980e-01 7.58542120e-01
8.62865627e-01 -5.35944760e-01 -1.07929923e-01 -2.65725583e-01
4.51732188e-01 2.68727690e-01 3.80469173e-01 -1.78853798e+00
1.05546527e-01 -1.03816744e-02 8.64565909e-01 -9.30830359e-01
5.25801003e-01 -2.04528242e-01 3.35777342e-01 4.12919432e-01
1.75011456e-01 -6.77659750e-01 1.85958877e-01 7.00329483e-01
6.13687895e-02 -5.39838493e-01 9.35051084e-01 -1.88640937e-01
-7.64826596e-01 -2.63292436e-02 -9.31671441e-01 -1.61014795e-01
5.20288229e-01 1.69653267e-01 -7.85379335e-02 -1.57328442e-01
-4.29315746e-01 -1.79363564e-01 -3.96655276e-02 5.04967272e-01
5.12050927e-01 -9.23291922e-01 -4.88580793e-01 2.41430894e-01
2.69193411e-01 -6.26813829e-01 4.62953120e-01 3.93997699e-01
-5.98241448e-01 9.50233817e-01 -2.77251959e-01 -3.87809277e-01
-7.40290165e-01 6.55074120e-01 7.79090464e-01 1.33937612e-01
-7.98103690e-01 1.63780534e+00 -8.61257166e-02 -4.14982289e-01
2.07987368e-01 2.75373787e-01 -1.59054294e-01 1.96589157e-01
7.02321947e-01 1.89162835e-01 4.45378661e-01 -1.84442267e-01
-5.77780724e-01 -4.49545607e-02 5.71090505e-02 -5.02659082e-01
1.57636321e+00 5.15459001e-01 1.89270884e-01 4.16831940e-01
9.72760201e-01 -1.25656381e-01 -5.67123115e-01 5.47559336e-02
1.60574079e-01 -1.15982592e-01 8.46358240e-01 -8.89897287e-01
-3.74194384e-01 4.98578250e-01 1.07753503e+00 3.34360689e-01
7.34766126e-01 2.30323687e-01 2.82151163e-01 1.03156841e+00
8.60769570e-01 -1.26385415e+00 4.80142310e-02 6.32023454e-01
1.19060481e+00 -1.07622910e+00 2.15301365e-01 3.78715575e-01
-2.68097639e-01 1.10520136e+00 2.22876027e-01 -1.48826027e-02
4.57108170e-01 5.92108108e-02 -2.35551327e-01 -6.25340402e-01
-5.87834358e-01 -3.12418062e-02 2.16317967e-01 7.04875827e-01
3.77634376e-01 1.72654927e-01 -1.27059370e-01 6.14177585e-01
-2.52026141e-01 -2.73883253e-01 4.67809409e-01 8.19425285e-01
-1.08272648e+00 -3.98620546e-01 -7.92007029e-01 6.31411493e-01
-1.80718511e-01 1.23908982e-01 -3.85548145e-01 4.60216552e-01
-1.45919174e-01 8.28719556e-01 -1.33935004e-01 -4.55655813e-01
-5.24487235e-02 1.92007333e-01 5.56235254e-01 -1.78652927e-01
-3.29231739e-01 2.02296749e-01 5.55427969e-01 -2.38403410e-01
-1.80617362e-01 -7.06826150e-01 -7.87595272e-01 6.55353665e-02
-5.10066092e-01 3.78521323e-01 1.23754013e+00 8.54663849e-01
1.88831180e-01 8.95509124e-01 6.55850053e-01 -1.06571829e+00
-6.74975574e-01 -9.12844598e-01 -6.63536012e-01 -5.24595261e-01
4.23525304e-01 -1.12550259e+00 -9.56525445e-01 -6.86024904e-01] | [7.568843841552734, 3.119579553604126] |
b686dc2c-081b-457e-955e-e69571352732 | a-theory-of-unsupervised-speech-recognition | 2306.07926 | null | https://arxiv.org/abs/2306.07926v1 | https://arxiv.org/pdf/2306.07926v1.pdf | A Theory of Unsupervised Speech Recognition | Unsupervised speech recognition (ASR-U) is the problem of learning automatic speech recognition (ASR) systems from unpaired speech-only and text-only corpora. While various algorithms exist to solve this problem, a theoretical framework is missing from studying their properties and addressing such issues as sensitivity to hyperparameters and training instability. In this paper, we proposed a general theoretical framework to study the properties of ASR-U systems based on random matrix theory and the theory of neural tangent kernels. Such a framework allows us to prove various learnability conditions and sample complexity bounds of ASR-U. Extensive ASR-U experiments on synthetic languages with three classes of transition graphs provide strong empirical evidence for our theory (code available at cactuswiththoughts/UnsupASRTheory.git). | ['Chang D. Yoo', 'Mark Hasegawa-Johnson', 'Liming Wang'] | 2023-06-09 | null | null | null | null | ['automatic-speech-recognition', 'unsupervised-speech-recognition'] | ['speech', 'speech'] | [ 3.02361608e-01 2.68620312e-01 -2.34046564e-01 -3.33879948e-01
-1.05130076e+00 -4.88193870e-01 4.73346323e-01 -1.34750441e-01
-1.21182710e-01 5.56858301e-01 -6.85165375e-02 -1.03238189e+00
-1.77320868e-01 -3.67074609e-01 -7.69486845e-01 -7.68534184e-01
-3.15626860e-01 3.33287656e-01 2.51570672e-01 -3.05728137e-01
-3.04567013e-02 5.36442161e-01 -1.16086841e+00 -7.62376785e-02
7.75154293e-01 6.18106246e-01 1.88728660e-01 1.25308442e+00
6.65437803e-02 8.49359512e-01 -2.33198971e-01 -1.51266068e-01
4.34806138e-01 -4.36977118e-01 -8.42797279e-01 2.91208655e-01
1.98828131e-01 1.44514116e-02 -1.00377142e+00 1.28005147e+00
2.58750051e-01 3.71159673e-01 7.09486961e-01 -1.11375332e+00
-8.17856669e-01 9.07983780e-01 -2.22395048e-01 3.94035399e-01
3.34704787e-01 -9.63917896e-02 1.11160541e+00 -9.65008914e-01
1.98302478e-01 1.09291804e+00 3.92450929e-01 8.28994453e-01
-1.12191308e+00 -4.72794682e-01 -3.53542566e-02 1.45203881e-02
-1.51779521e+00 -1.11210513e+00 6.37395203e-01 -3.22539270e-01
1.09546578e+00 4.95755941e-01 6.10014498e-02 1.06783414e+00
-2.70860642e-01 1.22504187e+00 1.09747386e+00 -6.90903902e-01
2.99563289e-01 3.05776268e-01 7.25560486e-01 9.13510382e-01
1.27865672e-01 2.09910542e-01 -5.76380491e-01 -1.99158356e-01
7.65166223e-01 -4.33385253e-01 -3.14030051e-01 -2.16325149e-01
-9.88020718e-01 6.66195095e-01 -7.74917826e-02 3.79188597e-01
-3.92149724e-02 -6.16213530e-02 1.59318119e-01 7.19505489e-01
3.83491993e-01 -4.63239336e-03 -4.23015743e-01 -2.09035933e-01
-7.46237636e-01 -2.68755734e-01 1.01730847e+00 1.35957181e+00
6.17541850e-01 6.56009197e-01 4.54483807e-01 9.79123950e-01
3.68011951e-01 7.18513072e-01 5.33209205e-01 -6.07024491e-01
7.08520293e-01 -1.09083846e-01 -2.17056274e-01 -5.88171065e-01
-1.70399204e-01 -1.24391638e-01 -7.61242747e-01 -2.60777056e-01
5.40111721e-01 -3.19569260e-01 -9.71098006e-01 1.42436266e+00
2.08720863e-02 3.55264693e-01 5.21933377e-01 6.80032969e-01
3.74525815e-01 9.93066609e-01 -4.68731135e-01 -4.04681206e-01
8.84621084e-01 -7.20719218e-01 -6.21756554e-01 -8.19502547e-02
9.55935776e-01 -5.06129384e-01 1.07956839e+00 2.89262742e-01
-9.20275688e-01 -7.39489198e-02 -9.75803435e-01 2.05643862e-01
-1.98616534e-01 1.47546634e-01 4.38778698e-01 1.00273442e+00
-1.27685070e+00 3.94344985e-01 -1.09976220e+00 -5.04837394e-01
-2.02664509e-01 3.29550713e-01 -4.72938955e-01 -3.18574831e-02
-1.13065648e+00 6.39873624e-01 3.34499478e-01 2.89338142e-01
-7.88619280e-01 -1.44695729e-01 -8.43460739e-01 -1.08283311e-01
5.28130889e-01 -4.76195030e-02 1.56996846e+00 -7.27080762e-01
-1.75076699e+00 4.39619541e-01 -2.48157218e-01 -6.64761603e-01
2.97136545e-01 3.24676260e-02 -4.89510417e-01 2.75523186e-01
-5.62593520e-01 -1.68808565e-01 7.74463236e-01 -9.38904464e-01
-2.94595569e-01 -4.77250159e-01 -2.92858511e-01 2.23153550e-02
-3.24191183e-01 1.92856833e-01 -3.34502831e-02 -7.18565464e-01
4.23733234e-01 -1.07447338e+00 -1.94522724e-01 -7.27739990e-01
-5.39221823e-01 -2.68397003e-01 6.65807068e-01 -6.86534524e-01
1.07714736e+00 -2.16929722e+00 5.48153743e-02 4.09834236e-01
-1.40020177e-01 4.98084247e-01 -1.92970857e-01 6.19123697e-01
-3.76430750e-01 1.57912970e-01 -3.00268829e-01 -5.08330613e-02
1.99664533e-01 3.79176289e-01 -6.81980908e-01 6.65072978e-01
4.90642563e-02 4.85912442e-01 -8.21146846e-01 -1.89917296e-01
1.06347583e-01 2.51433998e-01 -3.69917393e-01 1.93219945e-01
-7.32274503e-02 -3.57695967e-02 -5.56741118e-01 6.18232489e-01
2.05254599e-01 -8.06651339e-02 2.09829658e-01 3.72601062e-01
7.29283020e-02 5.58431983e-01 -1.18790996e+00 9.61034417e-01
-3.73854518e-01 8.32162261e-01 3.63602340e-01 -1.15373170e+00
9.04135764e-01 5.95898926e-01 8.58278051e-02 -1.60470560e-01
1.15622275e-01 3.57339859e-01 1.74050331e-01 -3.42336565e-01
5.23110569e-01 -4.69683483e-02 2.18897566e-01 5.06016016e-01
2.24886298e-01 1.46592230e-01 3.35191153e-02 3.76776636e-01
1.21015739e+00 -4.50656503e-01 1.46516725e-01 -3.29729915e-01
3.11772376e-01 -3.72267306e-01 2.39894286e-01 9.92056906e-01
-3.24089050e-01 4.19288129e-01 2.85572022e-01 3.52222435e-02
-1.10392845e+00 -1.17548025e+00 -2.60735333e-01 8.80634010e-01
-3.63974899e-01 -4.86928225e-01 -9.01235521e-01 -4.12912816e-01
-3.56863350e-01 8.31447363e-01 -2.56493628e-01 -1.13096043e-01
-3.91181767e-01 -6.97154760e-01 8.52470994e-01 4.46582824e-01
7.93529749e-02 -7.74614096e-01 2.50946969e-01 -6.67022243e-02
1.58798456e-01 -1.40571547e+00 -7.46418536e-01 4.94114190e-01
-8.62294912e-01 -8.36520314e-01 -7.22020030e-01 -8.77682567e-01
7.14129746e-01 3.86810720e-01 4.33495671e-01 -1.22143164e-01
5.96610568e-02 7.66286790e-01 -3.60435575e-01 1.93767831e-01
-1.12751508e+00 2.31558040e-01 5.83012402e-01 2.47368485e-01
3.20941955e-01 -4.96276259e-01 1.14483953e-01 4.37201530e-01
-9.59933281e-01 -1.53674141e-01 5.31282902e-01 8.38686466e-01
2.28965163e-01 1.14022680e-01 5.10988712e-01 -7.69767642e-01
8.11888158e-01 -3.43681484e-01 -8.91730309e-01 5.11077583e-01
-6.10887170e-01 2.57287443e-01 6.62656963e-01 -4.65341419e-01
-8.44324291e-01 1.57211632e-01 -2.91872323e-01 -5.19731641e-01
-2.73331881e-01 3.97442341e-01 -3.31185848e-01 2.20749825e-02
5.15934825e-01 8.35422754e-01 1.16644613e-02 -3.67230535e-01
4.49631065e-01 1.31138086e+00 4.76858050e-01 -7.68293977e-01
9.81230557e-01 -1.13222435e-01 -3.99967551e-01 -1.94206429e+00
-5.93595445e-01 -6.08331263e-01 -4.75678593e-01 -9.13495105e-03
4.41929936e-01 -8.01103890e-01 -5.87917447e-01 4.53575552e-01
-8.56004119e-01 -4.53433335e-01 4.42573763e-02 7.56535769e-01
-7.00693011e-01 6.79177463e-01 -1.03978252e+00 -1.55201006e+00
-1.33991078e-01 -9.66395080e-01 6.19516194e-01 -7.39933131e-03
9.39770862e-02 -9.86687481e-01 1.21813677e-01 3.76173496e-01
1.57380164e-01 -7.14421391e-01 8.66996527e-01 -1.04819441e+00
-6.56073332e-01 -2.54825026e-01 7.42727220e-02 7.41675258e-01
3.20952795e-02 1.46869689e-01 -1.01360357e+00 -3.74043733e-01
1.05598129e-01 -1.23166278e-01 6.82071686e-01 1.95199117e-01
8.83317709e-01 -7.23842204e-01 -1.56543553e-01 4.17402416e-01
1.02985179e+00 3.67151320e-01 5.55610120e-01 -1.71484649e-01
5.57289958e-01 3.97864312e-01 3.24176028e-02 1.60204157e-01
4.00676019e-02 4.62073207e-01 -3.78249913e-01 3.49100322e-01
1.10541463e-01 -4.17332590e-01 8.90940189e-01 1.79198110e+00
6.06960058e-02 -3.84589493e-01 -9.93186891e-01 4.89522755e-01
-1.74775302e+00 -9.89452541e-01 -3.34067866e-02 2.44563961e+00
7.45835125e-01 3.35896492e-01 2.15575531e-01 2.90588796e-01
1.02618110e+00 4.11334932e-02 -2.46604413e-01 -3.03457677e-01
-2.73088068e-01 -4.98415381e-02 6.27380431e-01 9.39912200e-01
-7.18933105e-01 1.04303861e+00 7.10961771e+00 8.74060512e-01
-8.59650195e-01 1.76529456e-02 4.97763813e-01 2.49003232e-01
-3.49000633e-01 6.04664311e-02 -8.86055112e-01 1.15530260e-01
1.41840041e+00 -2.13456661e-01 7.13841915e-01 7.17625976e-01
1.66583598e-01 3.25122446e-01 -9.93763864e-01 1.01969981e+00
-4.85332608e-02 -8.69894326e-01 -1.64083779e-01 8.74102190e-02
3.26667845e-01 3.74485999e-01 1.61695197e-01 2.70565838e-01
6.34806871e-01 -6.47392809e-01 5.02789259e-01 3.06080319e-02
7.26248980e-01 -5.32697201e-01 2.12465778e-01 4.72083271e-01
-1.00935853e+00 -8.78712535e-02 -4.42484409e-01 5.73820993e-02
-1.18394762e-01 4.01401818e-01 -1.19263601e+00 4.03832138e-01
1.31524354e-01 4.94896501e-01 -3.47581595e-01 7.00820386e-01
-1.49342224e-01 1.32669091e+00 -4.34044808e-01 -3.35050911e-01
8.19612294e-02 -4.52786148e-01 8.97682965e-01 1.28313231e+00
2.80368745e-01 2.65555054e-01 -4.21537124e-02 3.60802472e-01
4.57253009e-02 1.98425293e-01 -8.45934331e-01 -6.97654843e-01
4.03332889e-01 8.94649327e-01 -7.55744398e-01 -2.07152769e-01
-4.39955592e-01 8.38659406e-01 3.88414234e-01 6.70702577e-01
-3.47676396e-01 -4.93876934e-01 5.76048136e-01 2.41658986e-01
2.38282725e-01 -8.96127403e-01 7.45876655e-02 -1.60845280e+00
8.48691389e-02 -8.81920815e-01 2.10707247e-01 -5.78738749e-01
-1.24518359e+00 8.03807199e-01 -9.57001746e-02 -8.86116326e-01
-3.77214015e-01 -8.07171047e-01 -2.32353389e-01 4.00342435e-01
-1.07845664e+00 -7.35916495e-01 5.94658852e-01 5.65107346e-01
5.53637147e-01 -6.13701999e-01 7.33153164e-01 5.60398810e-02
-7.97685266e-01 8.07243645e-01 4.75242108e-01 4.55636322e-01
8.17043856e-02 -1.39084172e+00 5.89391232e-01 1.28755951e+00
6.96308196e-01 8.37310076e-01 7.68558681e-01 -4.30375069e-01
-1.76882970e+00 -6.74147367e-01 8.35046768e-01 -4.49622124e-01
1.25483274e+00 -7.64167666e-01 -9.98784184e-01 9.20694709e-01
3.38649340e-02 -1.54050767e-01 5.65799296e-01 1.37240529e-01
-4.85140115e-01 -7.97433183e-02 -4.45578009e-01 7.92908192e-01
9.57363129e-01 -1.15837300e+00 -5.91790259e-01 4.36853677e-01
8.44742298e-01 -2.66759396e-01 -7.60262489e-01 2.29636105e-04
2.41630644e-01 -5.88063180e-01 5.87086618e-01 -7.32146680e-01
-9.38603282e-02 -1.47136360e-01 -6.14006162e-01 -1.17578709e+00
1.52366653e-01 -1.29915214e+00 -2.78147161e-01 1.05264688e+00
7.39469111e-01 -9.48506296e-01 5.97760677e-01 6.72422051e-01
-3.15292507e-01 -4.15402442e-01 -9.90765572e-01 -1.17259991e+00
4.30666544e-02 -7.63260782e-01 1.19550824e-01 8.81777227e-01
2.31231347e-01 5.08837223e-01 -4.32812303e-01 6.48426294e-01
6.00475788e-01 -2.70382494e-01 5.13418913e-01 -7.20671296e-01
-5.77225983e-01 -2.11647913e-01 -4.34995234e-01 -1.23794532e+00
4.74266469e-01 -1.12773919e+00 1.74232006e-01 -9.81134057e-01
-1.18066847e-01 -4.22227979e-01 -3.78101289e-01 4.26767468e-01
1.74678832e-01 -4.72103328e-01 1.17968783e-01 1.85261369e-01
-4.39806521e-01 4.58488405e-01 6.39798224e-01 -6.60772771e-02
-1.02591500e-01 4.67190653e-01 -2.88365483e-01 4.71593231e-01
8.34464729e-01 -4.21481460e-01 -6.44032896e-01 -2.15629444e-01
-5.20900674e-02 5.77960312e-01 1.77126631e-01 -9.41349089e-01
3.40278715e-01 -5.22073209e-02 -2.30946347e-01 -4.47811335e-01
3.10967654e-01 -6.13933921e-01 -1.23112768e-01 3.21277678e-01
-6.55753374e-01 -6.65246770e-02 -1.49945349e-01 7.01431036e-01
-1.35119319e-01 -3.61018568e-01 7.28715122e-01 1.69170827e-01
-5.27030826e-01 3.33799273e-01 -9.06564116e-01 9.61119011e-02
6.50673985e-01 6.22946173e-02 -1.73730060e-01 -6.98617041e-01
-8.23022366e-01 -1.32751301e-01 1.75897792e-01 3.49914491e-01
7.64261127e-01 -1.06585991e+00 -6.22236967e-01 4.44634348e-01
1.64631996e-02 -4.85240966e-01 7.27385804e-02 6.57417417e-01
-2.32176781e-01 7.56399989e-01 4.55581725e-01 -4.24512833e-01
-1.30497980e+00 5.50961018e-01 4.58140373e-01 1.80528253e-01
-5.26353061e-01 5.37187696e-01 3.13088261e-02 -5.55798411e-01
4.06024903e-01 -2.84843266e-01 2.81829327e-01 -4.56968963e-01
5.34139931e-01 2.41600960e-01 2.08497599e-01 -6.75426304e-01
-2.57589847e-01 1.17277935e-01 -1.33865729e-01 -5.08422315e-01
1.11592042e+00 -2.26354495e-01 1.56773821e-01 7.67331183e-01
1.43231106e+00 1.52121261e-01 -8.39498281e-01 -4.22990799e-01
2.80318260e-01 -1.26933977e-01 -7.01019019e-02 -3.27915773e-02
-6.96245551e-01 9.92865920e-01 4.31643575e-01 7.18753219e-01
7.71681368e-01 4.88174856e-02 7.31661737e-01 1.02234626e+00
4.35739964e-01 -1.19599843e+00 -1.25522852e-01 7.93978214e-01
5.62011421e-01 -1.07395172e+00 -6.01603389e-01 -3.78205746e-01
-4.84472364e-01 1.16732395e+00 1.82027087e-01 2.97723170e-02
9.51827347e-01 4.03523207e-01 9.00708735e-02 2.99030185e-01
-9.84505594e-01 -4.23939824e-01 1.45020410e-02 6.99206829e-01
4.63228673e-01 3.29934806e-01 1.12978086e-01 4.21775401e-01
-3.37337196e-01 -3.77760231e-01 7.43098021e-01 7.80094326e-01
-5.44440269e-01 -1.01252353e+00 -3.70234281e-01 2.56547153e-01
-2.95804590e-01 -1.91908985e-01 -4.05888468e-01 5.30555308e-01
-9.72680449e-01 1.02540874e+00 -3.45182717e-01 -7.18534768e-01
1.48392981e-02 2.97690511e-01 4.70902324e-01 -5.89661956e-01
4.48448211e-02 2.67744035e-01 2.40254253e-01 -4.81715389e-02
3.77317774e-03 -7.44636834e-01 -1.09961736e+00 -9.36450958e-02
-6.32613420e-01 5.10823131e-01 5.81289649e-01 8.89290452e-01
1.43626064e-01 4.15289290e-02 9.18420613e-01 -1.29509509e-01
-1.10572636e+00 -8.28980803e-01 -1.11868775e+00 -8.42619464e-02
4.88047659e-01 -8.94495100e-02 -5.97114503e-01 7.18633160e-02] | [14.418644905090332, 6.684770584106445] |
81eaf081-ac86-44b7-9e86-eccbe4e210ea | accounting-ngrams-and-multi-word-terms-can | null | null | https://aclanthology.org/W16-1806 | https://aclanthology.org/W16-1806.pdf | Accounting ngrams and multi-word terms can improve topic models | null | ['Natalia Loukachevitch', 'Michael Nokel'] | 2016-08-01 | null | null | null | ws-2016-8 | ['text-clustering'] | ['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] | [-7.325442314147949, 3.6668591499328613] |
3248dca3-c8c7-4c12-9ac1-a3b197d0fb1b | adversarial-occlusion-aware-face-detection | 1709.05188 | null | http://arxiv.org/abs/1709.05188v6 | http://arxiv.org/pdf/1709.05188v6.pdf | Adversarial Occlusion-aware Face Detection | Occluded face detection is a challenging detection task due to the large
appearance variations incurred by various real-world occlusions. This paper
introduces an Adversarial Occlusion-aware Face Detector (AOFD) by
simultaneously detecting occluded faces and segmenting occluded areas.
Specifically, we employ an adversarial training strategy to generate
occlusion-like face features that are difficult for a face detector to
recognize. Occlusion mask is predicted simultaneously while detecting occluded
faces and the occluded area is utilized as an auxiliary instead of being
regarded as a hindrance. Moreover, the supervisory signals from the
segmentation branch will reversely affect the features, aiding in detecting
heavily-occluded faces accordingly. Consequently, AOFD is able to find the
faces with few exposed facial landmarks with very high confidences and keeps
high detection accuracy even for masked faces. Extensive experiments
demonstrate that AOFD not only significantly outperforms state-of-the-art
methods on the MAFA occluded face detection dataset, but also achieves
competitive detection accuracy on benchmark dataset for general face detection
such as FDDB. | ['Yujia Chen', 'Ran He', 'Lingxiao Song'] | 2017-09-15 | null | null | null | null | ['occluded-face-detection'] | ['computer-vision'] | [ 1.19860232e-01 2.52913624e-01 2.84068137e-02 -2.44960323e-01
-4.82161820e-01 -5.42185783e-01 2.96283096e-01 -5.59593499e-01
-1.96284316e-02 5.09325802e-01 -3.11558574e-01 -7.72814592e-03
5.35914540e-01 -7.67217100e-01 -6.88086808e-01 -9.12515879e-01
-1.20170906e-01 3.50229055e-01 8.72682333e-02 1.18634522e-01
-2.12890729e-01 9.79317248e-01 -1.60903037e+00 1.78966194e-01
7.17579424e-01 1.13784444e+00 -8.84501338e-02 1.08712956e-01
7.40170013e-03 4.85635400e-01 -7.94600189e-01 -5.61312556e-01
6.76719606e-01 -8.27192292e-02 -2.31358498e-01 4.86860693e-01
7.27369308e-01 -6.08059824e-01 -4.35320795e-01 1.25252581e+00
3.68209273e-01 -3.21546972e-01 7.74089396e-01 -1.23805285e+00
-4.44082737e-01 2.17749014e-01 -1.26673877e+00 2.61223555e-01
4.65540260e-01 1.85580283e-01 3.68007660e-01 -1.31398475e+00
5.16394258e-01 1.67728627e+00 3.06681871e-01 7.74896502e-01
-1.23301697e+00 -1.07228041e+00 5.19180536e-01 -6.15088269e-02
-1.80702043e+00 -1.00069022e+00 9.82989728e-01 -3.40231210e-01
1.37813821e-01 4.11969781e-01 3.09359610e-01 1.09070110e+00
-1.08589400e-02 7.93608308e-01 8.97418797e-01 -3.63909453e-01
-8.72073099e-02 3.11384112e-01 -3.07014674e-01 1.12911665e+00
2.81411648e-01 3.13722283e-01 -1.53696492e-01 -8.19790959e-02
7.56474376e-01 -3.85706164e-02 -3.72390389e-01 -2.66764849e-01
-4.83340263e-01 6.43398404e-01 4.83722508e-01 -2.97705494e-02
-3.62617135e-01 -3.42616826e-01 2.37906948e-01 2.46665731e-01
5.48211157e-01 -2.73737162e-01 -1.14808239e-01 8.48722577e-01
-8.33424747e-01 7.06732795e-02 3.32139134e-01 1.01920307e+00
6.67482853e-01 4.89443243e-01 -6.17215633e-01 7.43521154e-01
5.17101169e-01 6.66476548e-01 -9.56133679e-02 -6.87782168e-01
2.92815506e-01 6.75951183e-01 2.17448071e-01 -1.17791128e+00
-1.55567989e-01 -7.20456600e-01 -7.47578442e-01 5.43182075e-01
4.91910070e-01 -2.26515874e-01 -1.03732598e+00 1.79139018e+00
9.25395906e-01 7.94510394e-02 -1.76953375e-01 9.68418181e-01
9.50611770e-01 4.13665980e-01 4.53233197e-02 -6.37832224e-01
1.57783437e+00 -8.12908709e-01 -7.63685226e-01 -5.22739232e-01
-2.89298762e-02 -9.86851990e-01 6.31676972e-01 3.03645521e-01
-8.83113027e-01 -7.65691519e-01 -1.00648916e+00 2.32166544e-01
3.72755677e-02 7.30189204e-01 5.55719376e-01 1.03432059e+00
-8.36295128e-01 -1.09972410e-01 -6.15089774e-01 2.45397892e-02
1.07187176e+00 4.22944278e-01 -5.79100013e-01 -3.16774100e-01
-6.70593500e-01 5.03811896e-01 -2.49149720e-03 4.29725438e-01
-1.38767374e+00 -4.04673725e-01 -9.53236997e-01 9.72418208e-03
4.94011700e-01 -2.85636336e-01 6.50581598e-01 -1.13641465e+00
-1.18427444e+00 1.12871408e+00 -4.11921203e-01 -1.87699825e-01
9.41651285e-01 -4.93488498e-02 -6.14608645e-01 3.08938354e-01
2.75903940e-01 8.14873874e-01 1.65535164e+00 -1.38911903e+00
-4.50364262e-01 -7.60419548e-01 -1.46458253e-01 -3.01852147e-03
-4.13218558e-01 3.18249345e-01 -5.64282715e-01 -8.22931409e-01
2.61709213e-01 -6.97230160e-01 6.56524003e-02 6.09127104e-01
-6.42990112e-01 -3.04319173e-01 1.44716692e+00 -6.03189170e-01
8.51990283e-01 -2.29909396e+00 -1.75970063e-01 1.88348979e-01
2.08544984e-01 4.83826846e-01 -3.45398158e-01 -2.39744425e-01
-6.55838847e-02 -1.35647416e-01 -8.94075409e-02 -5.34277380e-01
-2.01387942e-01 5.40399086e-03 -3.79464597e-01 9.72763777e-01
6.24456882e-01 6.34657204e-01 -5.72984397e-01 -7.33224094e-01
2.31262520e-01 5.31919420e-01 -4.15301561e-01 3.24809760e-01
2.64726169e-02 6.55547082e-01 -6.36204422e-01 1.27982950e+00
1.36181903e+00 3.64937425e-01 5.15284762e-02 -2.92362154e-01
1.79447263e-01 -4.43380684e-01 -1.26000059e+00 1.04642439e+00
-2.28912532e-01 4.33843404e-01 7.33517945e-01 -7.88638651e-01
9.95990574e-01 2.20332503e-01 1.18508086e-01 -3.43077689e-01
4.80736375e-01 -1.20296059e-02 2.02109087e-02 -4.31221068e-01
-2.14610279e-01 2.26037815e-01 3.04967195e-01 1.40002504e-01
-1.40872106e-01 3.57734114e-01 8.60849395e-03 1.45570943e-02
8.64803553e-01 -2.99444757e-02 1.90500408e-01 -3.19115639e-01
1.04250252e+00 -5.78216910e-01 8.80949616e-01 4.57964838e-01
-6.61196053e-01 4.05913651e-01 2.59971321e-01 -5.04259765e-01
-2.84505785e-01 -1.20673740e+00 -5.10017097e-01 1.06259692e+00
2.84618169e-01 2.17334419e-01 -9.74550545e-01 -1.06316888e+00
1.38309374e-01 1.93799600e-01 -7.75334060e-01 -1.07854955e-01
-5.44640899e-01 -5.92352509e-01 5.08060813e-01 3.28293443e-01
6.08834028e-01 -1.08667982e+00 -2.40365684e-01 -8.55277777e-02
5.89816039e-03 -1.28972375e+00 -6.06214345e-01 -1.35393620e-01
-3.67022187e-01 -1.27363276e+00 -5.59059739e-01 -1.19782794e+00
1.22051322e+00 5.40301442e-01 8.60665381e-01 3.48966420e-01
-6.62334204e-01 -1.37863800e-01 -2.65086684e-02 -5.39295733e-01
-4.24357563e-01 -5.77599823e-01 2.04781875e-01 6.09774470e-01
3.82203788e-01 -2.94325501e-01 -7.52632856e-01 6.15130961e-01
-6.12917483e-01 -5.54365575e-01 5.77450812e-01 9.31384802e-01
6.05333507e-01 2.77399153e-01 8.34394455e-01 -9.18729901e-01
2.62282975e-02 -2.85229325e-01 -8.32460582e-01 2.13868231e-01
-5.48593178e-02 -5.32680035e-01 6.16055489e-01 -5.04527211e-01
-1.26267600e+00 4.87163901e-01 -1.22418292e-01 -8.26357961e-01
-4.36025590e-01 -5.62219739e-01 -8.18616390e-01 -4.70170349e-01
5.84762573e-01 1.12332225e-01 -8.90355278e-03 -3.47845048e-01
2.01254904e-01 7.18924522e-01 5.98132133e-01 -4.14988697e-01
1.26593971e+00 8.86616409e-01 -4.56587635e-02 -7.84836829e-01
-7.15882063e-01 -2.35519812e-01 -5.78726590e-01 -3.89922649e-01
5.88508666e-01 -1.14971650e+00 -5.77836990e-01 4.95086938e-01
-1.23564768e+00 3.84879619e-01 8.03929865e-02 -2.43911818e-02
8.92063752e-02 3.46907228e-01 -5.80425858e-01 -1.07049048e+00
-3.59145403e-01 -1.32290828e+00 1.40081966e+00 2.77829379e-01
2.59880543e-01 -5.09189904e-01 -7.30185330e-01 3.89716506e-01
7.96250105e-02 5.03490806e-01 5.22056699e-01 -3.04975301e-01
-7.26494789e-01 -5.24138689e-01 -2.48498246e-01 3.07407945e-01
3.32527578e-01 2.49757677e-01 -1.46055734e+00 -6.49811208e-01
1.44268438e-01 -1.75267667e-01 7.91972101e-01 2.96778172e-01
1.28947043e+00 -3.11562538e-01 -6.22940779e-01 6.19427860e-01
1.00642049e+00 1.63082972e-01 5.41192889e-01 -2.69013196e-01
5.51999509e-01 8.22939217e-01 8.31170022e-01 2.96166688e-01
-2.64566839e-01 7.13858604e-01 6.97312951e-01 -4.11103189e-01
-2.52268881e-01 -1.79306060e-01 4.27429050e-01 -1.28754854e-01
4.01697725e-01 -2.57864565e-01 -3.40643436e-01 2.68493503e-01
-1.30667675e+00 -7.24671602e-01 -2.30053887e-02 2.04007769e+00
5.86307585e-01 6.01499490e-02 8.93043578e-02 2.15198547e-01
1.26986337e+00 2.63582796e-01 -6.78844273e-01 -1.72567721e-02
-2.23628595e-01 2.21211448e-01 1.63399965e-01 2.12299719e-01
-1.55170536e+00 1.15818942e+00 5.68943548e+00 1.13105524e+00
-1.04518676e+00 3.92934859e-01 9.78987634e-01 -2.70384569e-02
2.43876442e-01 -3.57391328e-01 -9.68826532e-01 5.33924460e-01
-4.20968495e-02 2.37095103e-01 3.08211923e-01 1.07053578e+00
3.54652070e-02 9.89601016e-02 -9.61269617e-01 9.52948570e-01
1.53028235e-01 -6.30684733e-01 1.04878917e-01 1.49867535e-02
7.14220941e-01 -4.77814525e-01 4.27520961e-01 1.25286177e-01
-1.03863567e-01 -1.13809478e+00 7.69427001e-01 2.51340326e-02
9.90100384e-01 -8.14384222e-01 5.97671449e-01 1.94184572e-01
-1.37504447e+00 -3.01067531e-01 -5.99598348e-01 2.07448706e-01
-5.79731017e-02 5.77429473e-01 -6.40868485e-01 2.75381833e-01
4.32545513e-01 3.33936751e-01 -5.25848806e-01 8.27253520e-01
-5.28862000e-01 4.58829045e-01 -2.46869534e-01 7.20876455e-01
-2.25361839e-01 -5.00311926e-02 7.59427726e-01 8.04151237e-01
3.76745872e-02 -2.59987283e-02 4.40194458e-01 9.57982242e-01
-4.73712713e-01 2.04527140e-01 -7.75877118e-01 4.45819288e-01
8.06279719e-01 1.51864564e+00 -7.25206375e-01 -6.29887655e-02
-4.09535319e-01 9.91350353e-01 2.23788023e-01 3.41541916e-01
-1.05640030e+00 3.42386477e-02 7.43795455e-01 1.82758272e-01
4.29267883e-01 3.84510100e-01 1.03005253e-01 -9.45122600e-01
3.15213114e-01 -8.63280714e-01 3.73050332e-01 -2.75369912e-01
-1.16520715e+00 9.28580582e-01 -5.76583505e-01 -1.28729856e+00
1.58054724e-01 -5.01581848e-01 -7.78359771e-01 7.62265384e-01
-1.61368227e+00 -1.27337706e+00 -5.77428043e-01 6.81977510e-01
7.56969273e-01 -3.70697379e-01 6.87874794e-01 6.14556551e-01
-1.21219897e+00 1.19047487e+00 -2.70241857e-01 4.96532410e-01
7.07760692e-01 -6.70094907e-01 3.10284737e-02 1.16605365e+00
1.61626652e-01 3.98867548e-01 5.28979242e-01 -5.86375654e-01
-1.38737059e+00 -1.66461980e+00 1.57832101e-01 -3.10342312e-01
1.29686385e-01 -5.75518668e-01 -8.80175591e-01 5.55428147e-01
-2.07255468e-01 6.47704959e-01 1.97882980e-01 -4.01995331e-01
-4.92692977e-01 -4.71456438e-01 -1.68241775e+00 4.27949101e-01
1.22592199e+00 -4.46195424e-01 -2.21451551e-01 5.34848392e-01
2.86775023e-01 -3.71918321e-01 -3.59517395e-01 6.76047325e-01
5.49407899e-01 -1.00910449e+00 1.16352916e+00 -2.57704437e-01
-2.71692537e-02 -4.86356795e-01 -4.03283676e-03 -7.07325041e-01
-1.45247474e-01 -6.20693326e-01 -2.47187868e-01 1.72380602e+00
-8.72468129e-02 -6.65649831e-01 8.12226057e-01 1.17711134e-01
1.51515037e-01 -5.96131742e-01 -1.12040091e+00 -8.93694341e-01
-3.37485373e-01 1.11643270e-01 5.82542002e-01 6.97198391e-01
-6.82827830e-01 -3.39407176e-02 -2.85048813e-01 7.58095324e-01
8.72561574e-01 2.40956903e-01 7.87287176e-01 -1.19641316e+00
1.05873808e-01 -2.48989165e-01 -5.80953598e-01 -9.83832002e-01
6.11429393e-01 -6.19718611e-01 1.10753300e-03 -5.43076992e-01
4.91391197e-02 -3.97191256e-01 -1.40009910e-01 4.89776969e-01
-3.12219352e-01 9.60998654e-01 -9.35944095e-02 1.60778284e-01
-3.45130444e-01 5.31618357e-01 1.25902915e+00 -4.87039506e-01
-3.76353264e-02 1.77069023e-01 -6.80849969e-01 9.56601322e-01
3.66489589e-01 -4.42321181e-01 -2.84016609e-01 -2.32241154e-01
-7.87233353e-01 1.33439219e-02 4.49460894e-01 -9.87660408e-01
-8.35333914e-02 2.01324642e-01 1.04869151e+00 -5.54939091e-01
4.51053292e-01 -8.10542285e-01 1.20291494e-01 5.60099840e-01
2.45548263e-01 -3.74396533e-01 3.23494911e-01 7.37930298e-01
-2.02314466e-01 4.40794714e-02 1.32338524e+00 6.25886247e-02
-5.84762692e-01 5.22090733e-01 -1.14238091e-01 -2.06581339e-01
1.48687387e+00 -2.06354067e-01 -3.05805951e-01 -3.72528993e-02
-6.77035868e-01 2.48678565e-01 4.66319084e-01 5.55777311e-01
6.29938602e-01 -1.10625744e+00 -8.45012248e-01 8.81350994e-01
-6.82769567e-02 -9.23582986e-02 3.42831433e-01 5.75425684e-01
-3.42799366e-01 1.67653129e-01 -1.75548956e-01 -7.32113421e-01
-1.60799813e+00 8.51665437e-01 4.01432484e-01 2.10406259e-01
-4.24212247e-01 1.12413609e+00 8.15588057e-01 5.73316924e-02
4.82252210e-01 3.32566410e-01 -4.01320934e-01 1.42553076e-01
8.00168455e-01 1.48614928e-01 1.42621528e-03 -1.06843126e+00
-5.60612857e-01 5.79733014e-01 -3.93912941e-01 6.80495799e-01
7.90004671e-01 -1.89466715e-01 -1.17886923e-02 -5.20030200e-01
9.62204278e-01 3.82372081e-01 -1.49084914e+00 -2.47846961e-01
-5.24736106e-01 -9.90272284e-01 3.55471447e-02 -5.18374324e-01
-1.67205453e+00 8.12112629e-01 1.04587567e+00 2.05738097e-03
1.42624867e+00 1.57816242e-02 4.28507984e-01 -5.38421683e-02
6.05044842e-01 -6.07533336e-01 1.26023635e-01 -1.88696042e-01
1.10311568e+00 -1.35102415e+00 -4.07811888e-02 -1.18476653e+00
-2.42192611e-01 9.23194647e-01 1.05921924e+00 1.18636973e-02
4.69695985e-01 3.01604569e-01 2.23072574e-01 -9.53030437e-02
-2.19902709e-01 -2.94202149e-01 2.41147831e-01 6.94904983e-01
-1.60259157e-02 -5.11269160e-02 -9.46524926e-03 4.45806473e-01
1.31508231e-01 -5.60838044e-01 9.30943713e-02 4.90617543e-01
-3.49068195e-01 -9.35123026e-01 -7.66395032e-01 3.90385419e-01
-6.48637652e-01 2.05238864e-01 -4.41469580e-01 8.04183543e-01
4.87038612e-01 1.15500343e+00 1.08444490e-01 -6.45899922e-02
2.51177745e-03 -1.55263498e-01 4.68930691e-01 -6.77923381e-01
-2.41606295e-01 3.08177918e-01 -2.54542649e-01 -5.83038747e-01
-1.98847562e-01 -6.35753751e-01 -9.43946183e-01 -1.16653480e-01
-7.02124119e-01 5.98507002e-04 2.31656551e-01 7.07565725e-01
3.62228811e-01 4.68079180e-01 1.25770760e+00 -9.36956763e-01
-5.55520535e-01 -9.68767703e-01 -6.74683630e-01 2.81930655e-01
4.60474193e-01 -1.18228996e+00 -4.63197798e-01 -1.44432783e-01] | [13.352252006530762, 0.5669070482254028] |
bb863397-0485-41f8-98e7-9b4061268715 | examining-performance-of-sketch-to-image | 1811.00249 | null | http://arxiv.org/abs/1811.00249v1 | http://arxiv.org/pdf/1811.00249v1.pdf | Examining Performance of Sketch-to-Image Translation Models with Multiclass Automatically Generated Paired Training Data | Image translation is a computer vision task that involves translating one
representation of the scene into another. Various approaches have been proposed
and achieved highly desirable results. Nevertheless, its accomplishment
requires abundant paired training data which are expensive to acquire.
Therefore, models for translation are usually trained on a set of paired
training data which are carefully and laboriously designed. Our work is focused
on learning through automatically generated paired data. We propose a method to
generate fake sketches from images using an adversarial network and then pair
the images with corresponding fake sketches to form large-scale multi-class
paired training data for training a sketch-to-image translation model. Our
model is an encoder-decoder architecture where the encoder generates fake
sketches from images and the decoder performs sketch-to-image translation.
Qualitative results show that the encoder can be used for generating
large-scale multi-class paired data under low supervision. Our current dataset
now contains 61255 image and (fake) sketch pairs from 256 different categories.
These figures can be greatly increased in the future thanks to our weak
reliance on manually labeled data. | ['Dichao Hu'] | 2018-11-01 | null | null | null | null | ['sketch-to-image-translation'] | ['computer-vision'] | [ 5.24432242e-01 3.24958146e-01 -2.23945886e-01 -5.82522154e-01
-1.02368951e+00 -6.98131144e-01 8.25626910e-01 -7.30973840e-01
-5.58155142e-02 7.19314277e-01 -5.02032638e-02 -1.64250508e-01
6.95674837e-01 -8.19598854e-01 -1.20784497e+00 -3.33094925e-01
6.33942246e-01 5.53204298e-01 -7.08965287e-02 -2.52181023e-01
3.51731360e-01 3.56720597e-01 -1.18982506e+00 6.21978879e-01
5.81115365e-01 6.18969142e-01 5.10400832e-02 7.54420042e-01
-1.42143264e-01 6.71519279e-01 -7.35633314e-01 -1.00749063e+00
6.72635198e-01 -8.04055154e-01 -3.67614090e-01 2.33713090e-01
8.71021986e-01 -8.22758853e-01 -4.25053716e-01 1.09728086e+00
2.89233148e-01 -6.10516489e-01 7.51497269e-01 -1.87123311e+00
-1.27366006e+00 3.36258858e-01 -3.19249719e-01 -5.84342778e-01
2.02180699e-01 3.16800028e-01 6.09264135e-01 -1.15619361e+00
8.80682170e-01 1.27947867e+00 4.40093398e-01 9.03185785e-01
-1.27621818e+00 -1.11601567e+00 -3.88343900e-01 -1.63540855e-01
-1.34102297e+00 -5.81539154e-01 1.15570092e+00 -3.21709007e-01
4.27869886e-01 8.65725726e-02 5.34517884e-01 1.54471779e+00
-6.32892502e-03 8.83249879e-01 1.10310984e+00 -4.87267405e-01
-6.83154613e-02 5.15306890e-01 -8.42891991e-01 8.82169425e-01
1.72928974e-01 2.99482256e-01 -5.12963414e-01 -7.10991770e-02
1.30834901e+00 8.35266709e-02 -9.85817686e-02 -4.30440456e-01
-1.35678411e+00 8.97466004e-01 6.40212834e-01 8.52876753e-02
-1.78538427e-01 4.54305232e-01 2.49810115e-01 6.27738059e-01
3.05512398e-01 6.59569442e-01 -2.17635240e-02 1.91318125e-01
-1.06729579e+00 4.82504338e-01 5.14130056e-01 1.27617264e+00
7.17115462e-01 1.68915153e-01 5.90196364e-02 7.41958559e-01
5.26041090e-02 9.32781398e-01 3.67332697e-01 -9.05386627e-01
8.00684273e-01 4.11237985e-01 3.13943088e-01 -1.40822721e+00
5.63087463e-01 3.42262030e-01 -8.87844682e-01 3.60076010e-01
4.69630688e-01 1.80707335e-01 -8.57887030e-01 1.54247153e+00
-1.72358528e-01 1.38598964e-01 1.88489601e-01 9.57981825e-01
6.83688164e-01 6.86191142e-01 -2.20220760e-01 2.36869916e-01
9.07711267e-01 -1.07797146e+00 -6.61880672e-01 -2.83406675e-01
2.58850634e-01 -1.11187041e+00 1.35233319e+00 2.53842652e-01
-1.18827152e+00 -6.16508067e-01 -1.11676633e+00 -2.76523530e-01
-4.11393374e-01 4.06436414e-01 2.65762717e-01 6.09758019e-01
-8.24230492e-01 5.46087742e-01 -3.31032664e-01 -4.83686253e-02
8.98347199e-01 2.36075222e-01 -5.88114619e-01 -3.43053162e-01
-9.34988856e-01 9.89929855e-01 1.17837913e-01 -2.50022225e-02
-1.01219225e+00 -3.70356798e-01 -7.30853558e-01 -2.48977050e-01
1.39728084e-01 -6.45140767e-01 1.09041953e+00 -1.47928703e+00
-1.52344239e+00 1.14741361e+00 6.34968057e-02 -1.99542701e-01
1.02952874e+00 2.09835172e-01 -2.78752476e-01 3.18260789e-01
2.84910440e-01 1.29363251e+00 1.49683249e+00 -1.78248656e+00
-1.90087155e-01 -1.31826922e-01 -2.32545003e-01 -6.81759417e-02
-2.47674569e-01 -1.55325472e-01 -3.25795650e-01 -9.89577651e-01
2.28002248e-03 -1.06525362e+00 6.87888637e-02 5.33932745e-01
-3.47409427e-01 3.36507350e-01 1.14134765e+00 -6.73205435e-01
5.40872633e-01 -1.96057677e+00 9.56240073e-02 -1.30941719e-01
1.34294614e-01 4.60455894e-01 -5.56503236e-01 5.38818419e-01
-5.46897762e-02 2.63987660e-01 -2.24189654e-01 -5.37328541e-01
3.33132371e-02 1.88783318e-01 -8.69513035e-01 2.41234183e-01
5.48317671e-01 1.43672323e+00 -1.10745811e+00 -6.30549848e-01
2.69099742e-01 4.13759708e-01 -5.13636589e-01 6.71944916e-01
-2.38148093e-01 5.18030286e-01 -2.19905287e-01 7.97438025e-01
7.49513030e-01 1.48650780e-02 3.57952230e-02 -2.66932845e-01
2.90794700e-01 -2.34158412e-01 -5.07794142e-01 1.65838051e+00
-5.21726727e-01 9.15540159e-01 -1.85490966e-01 -8.37156951e-01
1.23819685e+00 3.31183791e-01 9.83103961e-02 -6.35444880e-01
1.77846596e-01 5.45208454e-01 -4.00608599e-01 -5.30511796e-01
6.14285469e-01 -6.14156604e-01 -3.76772404e-01 6.37986243e-01
-3.32607254e-02 -1.01288033e+00 -2.70304501e-01 2.96987683e-01
8.07823718e-01 3.60716820e-01 -1.09918602e-01 2.73359746e-01
3.04071426e-01 1.21210322e-01 6.08109795e-02 4.99818325e-01
-2.19048008e-01 9.20249999e-01 3.59123260e-01 -6.86639369e-01
-1.94390130e+00 -1.18645227e+00 2.99522012e-01 4.02384877e-01
6.60641268e-02 -6.81556910e-02 -7.23545909e-01 -8.66832018e-01
3.06298696e-02 5.24658263e-01 -5.09737968e-01 -3.63799363e-01
-6.70138836e-01 6.01254962e-02 9.90652084e-01 3.37565362e-01
6.72093689e-01 -1.30446410e+00 -1.88418388e-01 1.40230088e-02
-2.97308952e-01 -1.16309333e+00 -7.74897814e-01 -4.23826218e-01
-6.58155024e-01 -9.87223387e-01 -1.00306368e+00 -1.22611582e+00
1.20452499e+00 5.49190402e-01 9.60960925e-01 1.72545582e-01
-2.25515664e-01 8.27968717e-02 -2.41662294e-01 -3.56965750e-01
-1.01135182e+00 -3.74051958e-01 -2.98320651e-02 9.04038996e-02
3.07785086e-02 -5.03583789e-01 -3.42680663e-01 5.00793159e-01
-1.15603483e+00 5.53019285e-01 9.20694053e-01 1.22805178e+00
2.34057754e-01 -3.73735100e-01 6.49598122e-01 -7.68239260e-01
5.27088284e-01 -5.05948029e-02 -6.07163489e-01 3.60863417e-01
-3.37016761e-01 1.30302221e-01 1.00670910e+00 -6.25440598e-01
-1.03208959e+00 2.68112481e-01 3.57590467e-01 -1.05119658e+00
-3.78282033e-02 -1.60483822e-01 -1.98456436e-01 -2.76501149e-01
6.26549542e-01 4.07023758e-01 3.51670861e-01 -8.82955566e-02
8.60000074e-01 9.00825262e-01 8.11660588e-01 -5.91537178e-01
1.27125120e+00 5.66376030e-01 -2.15485469e-01 -5.13432324e-01
-5.01091778e-01 3.15253943e-01 -6.62272632e-01 -3.19617569e-01
6.42086804e-01 -8.73772562e-01 -4.52582538e-01 5.21042764e-01
-1.50274444e+00 -3.32387269e-01 -8.53053927e-02 2.79104970e-02
-7.15709925e-01 2.59457499e-01 -5.40530860e-01 -4.63355690e-01
-2.55116940e-01 -1.16224408e+00 1.46186721e+00 -2.05619648e-01
-1.27493501e-01 -7.07865119e-01 -1.17052235e-01 6.77789390e-01
3.85019273e-01 2.69339085e-01 8.14233303e-01 -7.45395720e-02
-1.00689018e+00 -5.47168195e-01 -6.18865669e-01 5.13189018e-01
2.51593739e-01 1.28427539e-02 -7.39488661e-01 -1.97675675e-01
-2.48760611e-01 -9.04361308e-01 5.67422450e-01 -1.69652924e-01
1.18433189e+00 -6.00058913e-01 -2.02923328e-01 4.27864462e-01
1.49061501e+00 1.89189371e-02 8.05559933e-01 -1.40206963e-01
9.82549012e-01 5.83033800e-01 5.75364172e-01 2.44570207e-02
3.02504659e-01 7.07887471e-01 4.33652014e-01 -2.22083315e-01
-2.94407010e-01 -9.70143676e-01 3.60645592e-01 7.00533748e-01
1.93290979e-01 -2.54937559e-01 -6.14802778e-01 6.56030715e-01
-1.66908658e+00 -1.20285881e+00 2.40890868e-03 2.09127378e+00
9.57762182e-01 -7.16673967e-04 -1.13714613e-01 2.11183894e-02
8.44554901e-01 6.96993023e-02 -3.97882432e-01 -3.78900558e-01
3.67331952e-02 1.65442392e-01 3.86649996e-01 4.81034189e-01
-8.84426117e-01 1.37689698e+00 6.42994499e+00 7.22130060e-01
-1.28821409e+00 -7.75826797e-02 7.12791562e-01 -1.88124143e-02
-4.67558801e-01 3.36310477e-03 -1.36147380e-01 6.35712326e-01
5.17478168e-01 1.78591888e-02 7.07608342e-01 7.18456566e-01
8.08331594e-02 1.58199698e-01 -1.18545532e+00 1.31631279e+00
2.84613848e-01 -1.63429809e+00 4.67326134e-01 3.04128453e-02
1.04044986e+00 -4.74573046e-01 1.79935649e-01 1.61899239e-01
2.71347791e-01 -1.09335017e+00 9.47966695e-01 5.41071296e-01
1.29121435e+00 -5.93204796e-01 4.68616039e-01 2.72750795e-01
-6.69083118e-01 2.62591213e-01 -4.81866270e-01 4.04740237e-02
1.41192824e-01 2.71618932e-01 -1.12115097e+00 2.27911472e-01
1.60053000e-01 7.32918978e-01 -5.58310032e-01 3.13922793e-01
-3.02035779e-01 2.18837962e-01 9.87297576e-03 -3.44101782e-03
1.50583521e-01 -3.33628297e-01 1.49897680e-01 7.92849839e-01
4.32020247e-01 -1.73587129e-01 -5.07753268e-02 1.33227968e+00
-4.54665273e-01 -2.03003287e-01 -1.31525517e+00 -3.65181714e-01
4.68730509e-01 1.03834069e+00 -5.34129083e-01 -7.95341074e-01
-2.12456122e-01 1.55569017e+00 2.48068392e-01 9.90210250e-02
-1.07440889e+00 -3.46751720e-01 3.95489454e-01 3.38407531e-02
1.27299234e-01 -9.19022635e-02 -5.19871831e-01 -1.41653013e+00
2.11468577e-01 -9.84849870e-01 -3.74404460e-01 -1.23264444e+00
-1.32141519e+00 6.41232312e-01 -2.81153321e-01 -1.54066813e+00
-2.74538100e-01 -5.75811982e-01 -5.07462204e-01 6.87510252e-01
-1.19388580e+00 -1.73265886e+00 -4.57400382e-01 5.86002350e-01
4.99516815e-01 -4.25579518e-01 9.46556449e-01 3.65412205e-01
-2.32656315e-01 7.89215088e-01 -9.86799896e-02 4.88248944e-01
1.06941450e+00 -9.06738162e-01 6.79333210e-01 6.46248877e-01
3.86760801e-01 3.16302925e-01 5.58100581e-01 -7.13631392e-01
-1.39043605e+00 -1.21642947e+00 1.09527779e+00 -7.11317122e-01
4.74995434e-01 -5.72200596e-01 -6.16932273e-01 6.19537294e-01
2.70900488e-01 3.20883989e-01 3.33117217e-01 -8.77637029e-01
-7.57025599e-01 2.63380799e-02 -1.34930718e+00 7.81520665e-01
9.54305291e-01 -9.16783094e-01 -5.21336317e-01 3.11601967e-01
5.08107603e-01 -3.71015608e-01 -5.72520375e-01 -1.37094289e-01
9.17236924e-01 -9.13309872e-01 1.00674319e+00 -6.04017258e-01
1.19880939e+00 -3.17982823e-01 -1.29654959e-01 -1.30006921e+00
1.67565152e-01 -5.34580231e-01 4.17135835e-01 1.02805662e+00
2.60930896e-01 -2.95282274e-01 1.09166157e+00 7.30659127e-01
1.22914113e-01 -3.89222622e-01 -5.97063839e-01 -9.51950192e-01
1.45026073e-01 -8.65264684e-02 7.55539000e-01 1.11799097e+00
-2.08436012e-01 2.95512885e-01 -9.91713464e-01 -1.63508028e-01
7.15218186e-01 3.22676510e-01 1.28862393e+00 -5.63217103e-01
-1.10081449e-01 -3.16553324e-01 -3.93457800e-01 -8.82949769e-01
2.71508425e-01 -9.56275403e-01 1.57192931e-01 -9.79940772e-01
5.30794784e-02 -4.69837755e-01 5.45746505e-01 6.02858663e-01
-3.14937681e-02 9.49344218e-01 4.49943602e-01 4.81138766e-01
-1.76253363e-01 7.56191313e-01 1.74527407e+00 -3.56002122e-01
2.91332126e-01 -2.88584769e-01 -4.58306462e-01 5.02166450e-01
7.33934045e-01 -6.81349814e-01 -2.84187764e-01 -5.01252532e-01
-3.77102122e-02 3.26520503e-01 8.90643775e-01 -6.69576108e-01
-1.06089965e-01 -2.69829184e-01 5.10303080e-01 -2.67808527e-01
5.30059993e-01 -9.19913232e-01 2.21248358e-01 5.09339154e-01
-4.42047268e-01 1.12819243e-02 -1.64511099e-01 4.81955260e-01
-3.51715863e-01 -4.10506576e-02 9.18807387e-01 -3.59240025e-01
-3.07556689e-01 2.69323230e-01 6.98919445e-02 -2.15691090e-01
1.07125962e+00 -2.09874138e-01 -2.93176442e-01 -7.47559428e-01
-4.45991278e-01 -3.71856093e-01 9.63112891e-01 6.70063674e-01
9.51549530e-01 -1.92576098e+00 -6.74660504e-01 4.60027099e-01
3.15113336e-01 -1.31843150e-01 -2.02431202e-01 8.04134309e-02
-7.60195494e-01 2.98401594e-01 -6.70215547e-01 -3.41479361e-01
-1.22431910e+00 8.19179177e-01 5.15581667e-02 9.20842886e-02
-3.95109713e-01 6.41314566e-01 -4.37515229e-02 -5.45604944e-01
-1.47954956e-01 -1.67086899e-01 6.71160877e-01 -3.58942509e-01
3.18654746e-01 -6.27746508e-02 -3.17758799e-01 -7.36014068e-01
-4.46558110e-02 5.19846976e-01 1.44364178e-01 -3.32933754e-01
1.21092284e+00 2.54845083e-01 -1.10136054e-01 -2.85201520e-02
1.63475084e+00 -2.53797919e-01 -1.42939341e+00 -3.39916311e-02
-4.49738771e-01 -1.05321109e+00 -3.27649236e-01 -7.69455552e-01
-1.20545554e+00 9.84726548e-01 3.12512308e-01 -1.96952045e-01
8.61646831e-01 -1.96169510e-01 9.78179097e-01 3.66165161e-01
6.98948860e-01 -8.32556188e-01 7.37300694e-01 -1.42223924e-01
1.31627488e+00 -1.63228047e+00 -1.51379675e-01 -3.39988619e-01
-8.36256564e-01 1.16162300e+00 5.97493112e-01 -6.35403454e-01
2.62832880e-01 5.21783307e-02 3.85682285e-01 -3.42442244e-02
-5.06535113e-01 3.06013584e-01 1.75676182e-01 6.41263723e-01
7.03130737e-02 1.90915123e-01 -2.51606721e-02 1.64442375e-01
-3.34902585e-01 2.30686173e-01 6.29826128e-01 7.23312199e-01
5.97860664e-03 -1.68225992e+00 -5.47539175e-01 3.37291360e-01
4.32921201e-02 2.97257043e-02 -8.47742438e-01 7.08245754e-01
2.05960363e-01 7.12747753e-01 -9.09209326e-02 -5.60376346e-01
3.95474248e-02 -8.55717137e-02 9.46535945e-01 -4.03813988e-01
-3.89924049e-01 -2.94084519e-01 -1.32556394e-01 -5.28011143e-01
-2.62488246e-01 -4.16677058e-01 -6.85031712e-01 -5.46533763e-01
-1.31484449e-01 -2.13723660e-01 7.39216506e-01 6.74261391e-01
4.00639623e-01 1.45687740e-02 8.16625357e-01 -1.22812700e+00
-5.14162540e-01 -7.06836522e-01 -3.11362803e-01 8.38340342e-01
2.92885423e-01 -4.61259633e-01 -2.40339369e-01 4.68258232e-01] | [11.900199890136719, -0.09539579600095749] |
786e4524-00d4-48c3-b80e-a2aff3b12caa | theme-transformer-symbolic-music-generation | 2111.04093 | null | https://arxiv.org/abs/2111.04093v2 | https://arxiv.org/pdf/2111.04093v2.pdf | Theme Transformer: Symbolic Music Generation with Theme-Conditioned Transformer | Attention-based Transformer models have been increasingly employed for automatic music generation. To condition the generation process of such a model with a user-specified sequence, a popular approach is to take that conditioning sequence as a priming sequence and ask a Transformer decoder to generate a continuation. However, this prompt-based conditioning cannot guarantee that the conditioning sequence would develop or even simply repeat itself in the generated continuation. In this paper, we propose an alternative conditioning approach, called theme-based conditioning, that explicitly trains the Transformer to treat the conditioning sequence as a thematic material that has to manifest itself multiple times in its generation result. This is achieved with two main technical contributions. First, we propose a deep learning-based approach that uses contrastive representation learning and clustering to automatically retrieve thematic materials from music pieces in the training data. Second, we propose a novel gated parallel attention module to be used in a sequence-to-sequence (seq2seq) encoder/decoder architecture to more effectively account for a given conditioning thematic material in the generation process of the Transformer decoder. We report on objective and subjective evaluations of variants of the proposed Theme Transformer and the conventional prompt-based baseline, showing that our best model can generate, to some extent, polyphonic pop piano music with repetition and plausible variations of a given condition. | ['Yi-Hsuan Yang', 'Meinard Müller', 'Frank Zalkow', 'Shih-Lun Wu', 'Yi-Jen Shih'] | 2021-11-07 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 7.85852969e-01 -4.61672479e-03 3.11339617e-01 7.10958689e-02
-1.01943696e+00 -8.02958786e-01 8.10609877e-01 -4.20114063e-02
-3.02447490e-02 5.71812749e-01 4.78643566e-01 -4.23860438e-02
7.42210075e-02 -7.61952400e-01 -8.12061667e-01 -7.57334888e-01
3.04202110e-01 7.44039297e-01 -2.21937373e-01 -4.76258606e-01
4.19297963e-01 -5.51625751e-02 -1.71327627e+00 9.65106845e-01
8.97977650e-01 8.13582182e-01 8.18205297e-01 8.15605223e-01
-2.10635681e-02 7.71810174e-01 -1.07193410e+00 -5.25266826e-01
1.16481438e-01 -1.16909468e+00 -9.05707538e-01 -1.06286379e-02
2.01734066e-01 -4.18328866e-03 2.71608353e-01 7.54185140e-01
7.82312691e-01 1.48913026e-01 8.72580826e-01 -7.92666554e-01
-6.00070238e-01 1.33373153e+00 -1.91146463e-01 -1.97817728e-01
6.65084660e-01 1.00842841e-01 1.40152681e+00 -8.81869197e-01
5.60190320e-01 9.92884278e-01 5.04739523e-01 6.97597325e-01
-1.42765629e+00 -5.46252310e-01 -1.21161349e-01 2.40902945e-01
-1.25941145e+00 -3.60598713e-01 9.68842983e-01 -4.46040720e-01
8.84992540e-01 3.89112294e-01 9.48408723e-01 1.50669694e+00
4.33469526e-02 8.65016937e-01 1.04971468e+00 -6.77546740e-01
1.87794551e-01 -2.17760026e-01 -4.94002253e-01 1.22721881e-01
-4.94928598e-01 3.42713207e-01 -8.44689608e-01 1.35069057e-01
6.81301177e-01 -6.75500095e-01 -4.26584691e-01 2.91889459e-01
-1.26910102e+00 6.65888667e-01 2.20737860e-01 5.23154795e-01
-7.17923224e-01 3.02319348e-01 5.24233282e-01 1.02671340e-01
1.72895536e-01 8.04848135e-01 -3.73910517e-01 -2.87937284e-01
-1.30273485e+00 5.53983808e-01 7.46874869e-01 8.92416358e-01
4.21065152e-01 2.31445953e-01 -8.09993744e-01 7.67078638e-01
1.83875859e-01 2.47968480e-01 8.87372613e-01 -5.99391222e-01
5.08531094e-01 1.41029105e-01 1.00562930e-01 -5.50707459e-01
3.72797213e-02 -5.27010143e-01 -6.01219237e-01 1.42875001e-01
1.87776178e-01 -3.57358083e-02 -7.99358845e-01 2.04413342e+00
-8.80071372e-02 4.84571278e-01 1.12592563e-01 9.65988040e-01
6.45250022e-01 8.26901853e-01 1.38817966e-01 -3.01576346e-01
1.26894855e+00 -9.07961309e-01 -7.36990988e-01 7.17095379e-03
2.31628850e-01 -1.02077925e+00 1.51767588e+00 6.64958596e-01
-1.31440866e+00 -8.93623888e-01 -1.18675196e+00 6.77214637e-02
3.84774692e-02 2.34091908e-01 1.93099037e-01 4.45602626e-01
-9.40265656e-01 9.80914533e-01 -2.74959624e-01 8.25283176e-04
2.99028549e-02 1.35096908e-01 1.79084986e-01 4.21873063e-01
-1.44351482e+00 6.29482269e-01 5.25268614e-01 2.05071762e-01
-1.18747461e+00 -7.59395361e-01 -5.50385773e-01 2.19379365e-01
-4.93309498e-02 -9.08773839e-01 1.50918853e+00 -1.48226762e+00
-1.83132601e+00 8.65085006e-01 -9.97169018e-02 -4.54879165e-01
1.68121085e-01 -3.13079655e-01 -1.78167090e-01 7.04954341e-02
3.57539281e-02 6.45206332e-01 1.12120140e+00 -1.40411639e+00
-3.37924212e-01 1.22456878e-01 -1.13129184e-01 4.34725940e-01
1.84234664e-01 1.04167163e-01 -3.13337177e-01 -1.13885725e+00
-1.72100738e-01 -9.48572636e-01 -5.18920906e-02 -6.92783177e-01
-7.34658778e-01 -1.10651888e-01 1.63334206e-01 -5.81310213e-01
1.31043029e+00 -2.11771107e+00 5.04321158e-01 4.44177203e-02
-3.98357034e-01 2.57692873e-01 -4.28562045e-01 6.17137909e-01
-4.01386976e-01 1.48723781e-01 -4.38318938e-01 -5.92317760e-01
1.07985765e-01 -6.26886487e-02 -8.71238887e-01 -1.38947904e-01
4.14757192e-01 1.01636183e+00 -8.59566450e-01 -1.11767322e-01
-9.37219039e-02 6.58893704e-01 -7.38422513e-01 5.18534899e-01
-8.13332498e-01 7.52319813e-01 -3.15354317e-02 1.73546568e-01
4.08280760e-01 2.20737442e-01 2.33963251e-01 -2.17296362e-01
-2.84557968e-01 9.20342267e-01 -1.03583109e+00 2.07095695e+00
-5.56966722e-01 4.40311939e-01 -1.74613610e-01 -9.30958331e-01
8.70720029e-01 7.65672266e-01 2.40066573e-01 -6.82112873e-01
1.89545989e-01 3.19984794e-01 1.35821894e-01 -3.52562726e-01
7.96134531e-01 -6.31194592e-01 -2.30342701e-01 7.02490389e-01
7.21023306e-02 -3.31892580e-01 2.80063748e-01 -1.80137679e-02
8.46089721e-01 6.75508797e-01 9.48831439e-02 -7.74934217e-02
6.11997902e-01 -1.29239812e-01 4.03927565e-01 6.89613640e-01
3.40338141e-01 1.18381846e+00 4.26643461e-01 7.33410567e-02
-1.09074998e+00 -1.04008865e+00 3.82959902e-01 1.14338553e+00
-1.96405619e-01 -7.27614224e-01 -8.55328858e-01 -4.15801316e-01
-5.65954983e-01 1.01539552e+00 -5.69425881e-01 -1.71249002e-01
-8.77872586e-01 -4.76679891e-01 6.96808577e-01 4.32736546e-01
2.24255338e-01 -1.86289310e+00 -5.54443598e-01 5.10981679e-01
-7.96022952e-01 -4.58753228e-01 -6.11324847e-01 2.87657887e-01
-6.67584479e-01 -8.01831067e-01 -8.57490122e-01 -8.74758065e-01
2.09125519e-01 -3.76875848e-01 1.32878757e+00 -2.15390436e-02
-6.61431998e-02 7.52059445e-02 -6.37194097e-01 -3.33195359e-01
-8.02166641e-01 1.27861962e-01 -2.60160953e-01 1.33675322e-01
-2.19506204e-01 -7.92042732e-01 -5.41195452e-01 -8.68639424e-02
-1.02140546e+00 4.26616639e-01 6.96538687e-01 9.05888617e-01
7.38132656e-01 -1.01395033e-01 7.84124017e-01 -7.43404686e-01
7.79772699e-01 -2.80829370e-01 -2.11947650e-01 5.89457415e-02
-1.26496390e-01 2.03953221e-01 9.20369446e-01 -7.09871292e-01
-1.06885672e+00 6.47252575e-02 -5.13582289e-01 -5.10939181e-01
-1.40138352e-02 8.14536870e-01 -3.40147913e-01 7.46200562e-01
8.67419899e-01 5.47820091e-01 -2.28472710e-01 -4.88621265e-01
4.73673940e-01 4.36151892e-01 6.70887887e-01 -7.72328019e-01
7.06559837e-01 -1.82674509e-02 -2.19360501e-01 -3.18098038e-01
-7.43922114e-01 -3.98046970e-02 -4.89802361e-01 -3.73080939e-01
8.26978743e-01 -8.70891094e-01 -6.00798726e-01 4.62913275e-01
-1.44196975e+00 -7.79470265e-01 -6.09573126e-01 2.34671548e-01
-1.06563032e+00 4.64619845e-02 -5.96979856e-01 -7.45626926e-01
-6.19414091e-01 -9.26921606e-01 1.21116602e+00 -7.43208006e-02
-5.05663395e-01 -6.30557358e-01 3.67589384e-01 -7.53825577e-03
3.94639045e-01 1.42104030e-01 1.04664993e+00 -4.83677745e-01
-6.10462785e-01 2.21691594e-01 2.58234113e-01 3.15150023e-01
1.53051943e-01 -2.52975058e-02 -1.23726583e+00 -6.79979399e-02
1.19122155e-01 -2.69658715e-01 8.42726469e-01 1.71854064e-01
9.85432863e-01 -3.82945061e-01 7.06785172e-02 4.39895868e-01
1.30548036e+00 2.84446806e-01 1.06026518e+00 1.52982175e-02
5.07831931e-01 5.90722203e-01 5.42215705e-01 5.12603819e-01
1.48140248e-02 8.32780302e-01 1.20440997e-01 1.63389996e-01
-5.56676209e-01 -5.77623904e-01 8.18493664e-01 1.01027608e+00
-1.43999040e-01 -5.14054477e-01 -5.58429539e-01 6.12543821e-01
-1.57708955e+00 -1.35334718e+00 -2.47608155e-01 2.37825251e+00
1.20415151e+00 -1.03445640e-02 1.50873542e-01 6.42107725e-01
8.04311633e-01 -2.88971122e-02 2.01663133e-02 -5.79792440e-01
-2.43627056e-01 8.70410860e-01 -3.70345026e-01 3.89893949e-01
-8.20646346e-01 1.04810846e+00 6.06857300e+00 1.01104915e+00
-1.39160001e+00 3.21780704e-02 2.60882825e-01 -5.49107380e-02
-6.16023600e-01 8.53604004e-02 -4.40331101e-01 4.99119937e-01
1.08070338e+00 -2.74576813e-01 6.44996524e-01 4.23182249e-01
5.22574127e-01 9.58963260e-02 -1.40016735e+00 8.67435813e-01
1.51739255e-01 -1.22606444e+00 4.95941967e-01 -1.55732244e-01
7.37552226e-01 -5.54966807e-01 3.12397838e-01 4.15706694e-01
-1.83917910e-01 -1.13553178e+00 1.27685010e+00 5.55076659e-01
9.71560180e-01 -6.82715893e-01 2.99601793e-01 1.11531995e-01
-1.35971701e+00 6.09637015e-02 6.13184758e-02 -1.95506349e-01
4.42917854e-01 3.52944106e-01 -1.05577326e+00 7.04805791e-01
1.96706459e-01 4.49706614e-01 -2.69598871e-01 1.06626999e+00
-5.77485681e-01 1.09184599e+00 -1.27044204e-03 8.05081576e-02
-5.29954359e-02 -7.23531321e-02 7.66715348e-01 1.32111204e+00
7.36089945e-01 -1.03870384e-01 -1.74259588e-01 1.26115191e+00
6.71022981e-02 3.85557264e-01 -4.27060366e-01 -1.32685274e-01
3.04418087e-01 1.09994316e+00 -5.75833976e-01 -3.31065744e-01
1.69832945e-01 1.47386742e+00 1.23392120e-01 2.64298856e-01
-1.00097227e+00 -2.06059203e-01 3.12201202e-01 9.79245678e-02
7.54985809e-01 -1.09151490e-02 -3.56588870e-01 -8.42518449e-01
-2.70113405e-02 -1.07212281e+00 1.51667893e-01 -1.16424835e+00
-1.16674888e+00 8.94946158e-01 -3.58543575e-01 -1.44035029e+00
-6.09236121e-01 -1.84014589e-01 -9.56428230e-01 1.14051366e+00
-1.18638992e+00 -1.18026733e+00 3.52931805e-02 5.59658229e-01
6.55373514e-01 -8.45260918e-02 1.11187160e+00 3.79599124e-01
-1.01091675e-01 4.44157779e-01 -4.48144108e-01 -7.99844041e-02
8.29020441e-01 -1.44833076e+00 2.79035211e-01 7.33762920e-01
5.29165387e-01 5.93055069e-01 7.77981341e-01 -7.47537971e-01
-9.55493867e-01 -1.06724143e+00 1.26218402e+00 -2.01702088e-01
2.87638724e-01 -5.18260956e-01 -7.77325451e-01 3.85893673e-01
6.71906352e-01 -7.96735168e-01 6.60103202e-01 -1.69462394e-02
-3.09429348e-01 -7.91511163e-02 -5.73424280e-01 7.45589912e-01
8.62361908e-01 -6.44192815e-01 -8.13209474e-01 -6.78739091e-03
7.36218870e-01 -3.20065856e-01 -3.76348704e-01 3.19562882e-01
4.13345963e-01 -9.88508165e-01 5.24429917e-01 -2.38296703e-01
8.45970571e-01 -6.37528956e-01 -8.45647044e-03 -1.54210472e+00
-3.10516924e-01 -1.06258476e+00 1.32591918e-01 1.51627481e+00
6.79771960e-01 1.02733687e-01 4.46830124e-01 -6.52616695e-02
-5.56095600e-01 -3.97935629e-01 -8.29540014e-01 -6.11596882e-01
6.08519092e-02 -6.97688460e-01 5.69851577e-01 5.66567004e-01
6.27376586e-02 7.85656691e-01 -5.46496212e-01 -1.59004867e-01
9.33732912e-02 2.76615560e-01 5.15500665e-01 -9.57797527e-01
-6.33966625e-01 -5.19713342e-01 4.35938798e-02 -1.09317935e+00
1.07964255e-01 -1.25893509e+00 5.61329961e-01 -1.44232678e+00
-5.40593937e-02 -5.10344505e-01 -3.14905882e-01 3.06143910e-01
-2.55896121e-01 3.07476372e-01 4.20842409e-01 1.36019424e-01
-1.49252519e-01 9.21963811e-01 1.17635715e+00 -1.10721432e-01
-4.19823319e-01 1.36882082e-01 -7.27504790e-01 1.15824908e-01
6.87010765e-01 -3.89847726e-01 -4.56459731e-01 -2.17338264e-01
4.83007550e-01 2.19051555e-01 3.58107477e-01 -1.11577988e+00
4.96902540e-02 2.13422358e-01 1.71121582e-01 -5.99431813e-01
4.73255366e-01 -3.52353811e-01 4.42664653e-01 3.05588394e-01
-6.46089494e-01 8.44360068e-02 3.12674254e-01 2.03241974e-01
-3.09088230e-01 -3.27084541e-01 4.82373625e-01 -1.76524103e-01
-3.09638292e-01 -1.16770744e-01 -6.94932520e-01 5.70624769e-02
5.52746117e-01 -5.66503452e-03 9.41778868e-02 -4.90043879e-01
-8.16349387e-01 -3.72103810e-01 8.45867768e-02 3.53441685e-01
7.06189215e-01 -1.50272727e+00 -1.05925775e+00 1.43794581e-01
-5.25259711e-02 -2.17250749e-01 2.46236667e-01 6.63044155e-01
-1.92304075e-01 1.52933121e-01 -1.42963678e-01 -4.19896096e-01
-1.08183324e+00 6.97230101e-01 2.45285302e-01 -3.65097523e-01
-5.22708356e-01 9.01378930e-01 2.35923499e-01 -1.64443329e-01
6.19203858e-02 -2.52809435e-01 -2.66914517e-01 3.49593580e-01
5.33743203e-01 -1.51048318e-01 3.72386463e-02 -6.56915903e-01
1.70806162e-02 3.95344138e-01 2.44019955e-01 -9.18237090e-01
9.83141243e-01 1.89650476e-01 -1.90446898e-01 7.49704421e-01
7.99118876e-01 3.72569174e-01 -1.01293099e+00 2.02341855e-01
-6.55051023e-02 -1.13941498e-01 -4.04566318e-01 -1.15775216e+00
-8.56472313e-01 8.43627691e-01 2.49759585e-01 1.27794653e-01
1.29645848e+00 -5.44229522e-02 7.18619704e-01 -1.16830012e-02
5.65435328e-02 -1.00019312e+00 3.95590127e-01 6.91457033e-01
1.48372400e+00 -4.83587205e-01 -5.68842888e-01 -2.42067575e-01
-8.10848296e-01 1.05298615e+00 4.60694343e-01 -1.01140745e-01
2.10761011e-01 1.77106544e-01 1.13631494e-01 5.52609339e-02
-9.66049254e-01 -4.64184582e-01 3.68140668e-01 6.06075823e-01
8.65977108e-01 8.87618866e-03 -4.32047725e-01 9.91314352e-01
-7.51734316e-01 -6.54953392e-03 3.89355153e-01 3.58640343e-01
-1.29970610e-01 -1.48332345e+00 -3.83851141e-01 -3.58793810e-02
-4.26626891e-01 -6.75943732e-01 -4.48161513e-01 2.68036366e-01
4.48194772e-01 9.57286835e-01 1.49422884e-01 -5.15277207e-01
3.07216763e-01 4.55434471e-01 6.75915539e-01 -7.81339943e-01
-1.19021952e+00 5.69726288e-01 1.63471580e-01 -2.98534870e-01
-4.80934143e-01 -7.09069550e-01 -1.23175943e+00 2.62192547e-01
-2.05753729e-01 1.61063433e-01 5.58694601e-01 9.05004680e-01
1.96540341e-01 9.57708359e-01 6.38599515e-01 -1.08720899e+00
-2.80655771e-01 -1.24307609e+00 -3.74183863e-01 6.01493418e-01
9.81996357e-02 -3.65383178e-01 -1.65451050e-01 3.93483281e-01] | [15.92122745513916, 5.617412567138672] |
68e20659-7d92-465b-9689-feec464bf1a4 | group-attention-single-shot-detector-ga-ssd | 1812.07166 | null | https://arxiv.org/abs/1812.07166v2 | https://arxiv.org/pdf/1812.07166v2.pdf | Group-Attention Single-Shot Detector (GA-SSD): Finding Pulmonary Nodules in Large-Scale CT Images | Early diagnosis of pulmonary nodules (PNs) can improve the survival rate of patients and yet is a challenging task for radiologists due to the image noise and artifacts in computed tomography (CT) images. In this paper, we propose a novel and effective abnormality detector implementing the attention mechanism and group convolution on 3D single-shot detector (SSD) called group-attention SSD (GA-SSD). We find that group convolution is effective in extracting rich context information between continuous slices, and attention network can learn the target features automatically. We collected a large-scale dataset that contained 4146 CT scans with annotations of varying types and sizes of PNs (even PNs smaller than 3mm were annotated). To the best of our knowledge, this dataset is the largest cohort with relatively complete annotations for PNs detection. Our experimental results show that the proposed group-attention SSD outperforms the classic SSD framework as well as the state-of-the-art 3DCNN, especially on some challenging lesion types. | ['Bjoern H. Menze', 'Wei-Shi Zheng', 'Jiechao Ma', 'Sen Liang', 'Rongguo Zhang', 'Xiang Li', 'Hongwei Li'] | 2018-12-18 | null | null | null | null | ['finding-pulmonary-nodules-in-large-scale-ct'] | ['medical'] | [ 2.17654392e-01 2.92728513e-01 -2.11317122e-01 -1.46758005e-01
-8.79208326e-01 -9.23167318e-02 2.43923664e-01 -1.25694156e-01
-4.88816112e-01 1.84177950e-01 3.75892967e-01 -5.01827657e-01
-2.36097917e-01 -3.99823546e-01 -4.82389331e-01 -7.85090625e-01
-8.14585611e-02 6.02142274e-01 9.53275323e-01 4.63573903e-01
-2.34133288e-01 6.17040455e-01 -9.78256285e-01 3.74423712e-01
7.46298611e-01 1.04119456e+00 7.76817322e-01 6.91947103e-01
-1.85704622e-02 1.01779234e+00 -9.07883048e-02 -1.47668514e-02
1.93750292e-01 -4.31339532e-01 -8.93137872e-01 9.24232230e-02
3.10710669e-01 -7.38288820e-01 -7.12956071e-01 8.70344579e-01
8.21163893e-01 -1.12100570e-02 7.34033585e-01 -6.40066564e-01
-1.00401342e+00 5.41188240e-01 -5.37573338e-01 9.88323152e-01
-3.65764886e-01 3.38866800e-01 8.63484323e-01 -1.19383693e+00
4.52739865e-01 7.28650093e-01 8.95166278e-01 7.24918544e-01
-4.05543864e-01 -4.57806230e-01 -1.91388309e-01 2.95582473e-01
-1.22358716e+00 1.49298862e-01 1.39073178e-01 -4.60771769e-01
9.48820531e-01 3.12460124e-01 7.78342843e-01 1.10002398e+00
2.20655680e-01 8.38517606e-01 5.77921271e-01 -3.14461768e-01
-1.03092574e-01 -1.29332796e-01 2.86946576e-02 1.01534390e+00
5.18601656e-01 1.68318301e-01 9.08970386e-02 -1.11289389e-01
1.12726915e+00 5.65618157e-01 -5.48050821e-01 -2.19426826e-01
-1.54099178e+00 6.07428372e-01 1.03927875e+00 6.68223679e-01
-6.63990796e-01 2.66727239e-01 2.94801146e-01 -3.59645605e-01
1.32734001e-01 2.12328658e-01 -2.16241062e-01 2.54140854e-01
-3.90995443e-01 -3.73590022e-01 5.35979271e-01 8.52737427e-01
-2.83608109e-01 -1.36526868e-01 -8.90464425e-01 6.78336143e-01
2.16124579e-01 3.70094150e-01 9.71124411e-01 -4.60849762e-01
9.53339413e-02 4.37188774e-01 -2.28512436e-01 -3.12015295e-01
-6.62363946e-01 -8.23031008e-01 -1.03436804e+00 -9.16082263e-02
3.85260254e-01 -1.26243398e-01 -1.36985457e+00 1.25704038e+00
7.99933299e-02 4.81005013e-01 -3.08983415e-01 1.14399850e+00
1.50394070e+00 1.95967376e-01 2.70026565e-01 -1.63262740e-01
1.53914118e+00 -1.48952436e+00 -7.02433527e-01 -1.11743174e-01
6.64220750e-01 -3.62330675e-01 1.01611233e+00 -1.84422836e-01
-1.06891501e+00 -5.24310887e-01 -9.07424450e-01 -6.36627451e-02
1.77001804e-01 3.43999982e-01 4.77382541e-01 4.67699915e-01
-9.15212333e-01 4.97426629e-01 -1.34517097e+00 -5.98883629e-01
9.37384367e-01 4.54919487e-01 -3.93208265e-02 -1.06995843e-01
-7.72241712e-01 8.80165756e-01 1.38010398e-01 -5.48778586e-02
-9.70413983e-01 -1.04848230e+00 -3.91186476e-01 3.29956174e-01
8.54323983e-01 -1.10330558e+00 1.92969537e+00 -5.68217456e-01
-1.12593520e+00 8.58677387e-01 7.58536905e-02 -4.78010625e-01
5.29918313e-01 -4.61624637e-02 -1.19795367e-01 3.05160820e-01
5.21670319e-02 5.62064469e-01 6.32599056e-01 -6.91346288e-01
-6.05862498e-01 -2.80916899e-01 -4.64857280e-01 2.22991377e-01
-1.71573937e-01 1.53169811e-01 -5.77329338e-01 -6.16248012e-01
4.62449104e-01 -9.24161792e-01 -6.10804260e-01 6.26300335e-01
-4.19475555e-01 -2.76302725e-01 9.46729481e-01 -6.06845617e-01
1.04223537e+00 -1.99661469e+00 -1.67871341e-01 -1.16519988e-01
7.17927694e-01 4.97529864e-01 3.35412860e-01 -3.90563875e-01
-1.90518111e-01 3.10225904e-01 -2.66370326e-01 -5.10885455e-02
-4.30722654e-01 3.25230062e-01 3.52330953e-01 3.85342896e-01
2.65657812e-01 1.45438194e+00 -8.65793049e-01 -8.84653211e-01
1.55924678e-01 3.59612167e-01 -4.81274039e-01 2.87346154e-01
9.66058224e-02 6.55477405e-01 -7.74033248e-01 7.57275581e-01
3.80031407e-01 -1.04230940e+00 -8.70964900e-02 -8.77553746e-02
5.89402653e-02 1.20828480e-01 -6.30395591e-01 1.57639396e+00
-2.00127289e-01 5.86202800e-01 -3.08038592e-01 -3.96412730e-01
3.62198621e-01 6.42777979e-01 8.70279849e-01 -2.67445564e-01
3.22569519e-01 4.18552488e-01 7.92714715e-01 -8.26569498e-01
-1.28490373e-01 -4.75681908e-02 4.58544403e-01 5.23783922e-01
2.57576518e-02 2.61209086e-02 -1.81748539e-01 -2.28056014e-02
1.62731481e+00 -6.08957112e-01 8.46378803e-01 -1.74384922e-01
3.39946091e-01 -1.60948604e-01 4.11577135e-01 1.16225863e+00
-6.00647449e-01 9.23133016e-01 2.85515100e-01 -4.66362119e-01
-1.04149246e+00 -1.19163859e+00 -3.51654649e-01 7.20157743e-01
-1.51809335e-01 8.16945732e-02 -4.42917913e-01 -1.20838940e+00
-4.28898595e-02 2.13077590e-01 -8.20888281e-01 1.87712591e-02
-6.67031407e-01 -7.21734822e-01 4.72689569e-01 1.23059690e+00
6.65675104e-01 -1.33182836e+00 -7.78086305e-01 1.04285814e-01
4.68708724e-02 -1.09360099e+00 -7.42678225e-01 3.06191742e-01
-9.92816269e-01 -1.38564396e+00 -1.16307259e+00 -1.14790833e+00
9.42269206e-01 5.04665911e-01 1.02703917e+00 2.61755943e-01
-7.64965713e-01 1.81005836e-01 -2.64797807e-01 -6.33614659e-01
-2.62670428e-01 4.35383767e-02 -5.05969822e-01 -4.16495055e-01
2.22859904e-01 -1.91740379e-01 -8.95852864e-01 2.85365909e-01
-7.72196829e-01 8.95033702e-02 1.43344212e+00 1.05275154e+00
8.32637072e-01 -4.17249762e-02 3.60613257e-01 -1.17940772e+00
2.61563659e-01 -5.46324074e-01 -2.43224204e-01 1.00205980e-01
-2.06500381e-01 -1.85718179e-01 3.06255192e-01 -3.19313139e-01
-1.09796524e+00 4.68750298e-03 -1.40308738e-01 -6.68339193e-01
-1.86293051e-01 3.10074538e-01 4.23795938e-01 -3.97910058e-01
7.94336677e-01 9.88751575e-02 -1.58629656e-01 -4.36020195e-01
-2.57744372e-01 6.90230191e-01 5.70533633e-01 2.13826403e-01
4.86018389e-01 4.90101308e-01 2.31389925e-01 -6.46245122e-01
-1.14322221e+00 -8.50336254e-01 -7.35059977e-01 -3.90484370e-02
9.81142938e-01 -6.22942448e-01 -2.17414409e-01 3.50331455e-01
-9.69770849e-01 -2.67009825e-01 -6.03198588e-01 8.19790125e-01
-3.38220209e-01 3.09106052e-01 -8.72958362e-01 -3.40057909e-01
-7.87583828e-01 -1.24877071e+00 9.41305935e-01 1.80806696e-01
1.33032426e-01 -6.79285169e-01 -2.08123058e-01 -5.80316037e-03
6.05072856e-01 -6.94517568e-02 8.35526884e-01 -9.94173706e-01
-8.19019139e-01 -1.76625282e-01 -7.99449801e-01 1.87593505e-01
2.08189666e-01 -1.41855925e-01 -6.86143398e-01 -1.50986582e-01
3.48245919e-01 -1.17941294e-02 1.12717974e+00 9.76412416e-01
2.03507447e+00 1.19355127e-01 -6.93989158e-01 7.33283460e-01
1.23830187e+00 3.70619416e-01 4.54956055e-01 -4.21479568e-02
8.26668084e-01 -1.84532091e-01 4.22335267e-01 3.37630391e-01
-1.74831957e-01 3.12995613e-01 6.07071161e-01 -3.33827138e-01
-6.06331825e-01 1.59380600e-01 -3.52759272e-01 7.53856719e-01
-3.58265370e-01 -4.56796765e-01 -1.32588184e+00 5.38551509e-01
-1.59976304e+00 -7.18401790e-01 -4.49434787e-01 1.76602054e+00
2.89333522e-01 6.89739212e-02 -2.84030944e-01 -2.34317213e-01
7.90680408e-01 -1.55103644e-02 -6.52367115e-01 2.45098948e-01
3.10944706e-01 3.51814866e-01 7.10606694e-01 -8.91804844e-02
-1.42126644e+00 2.34956339e-01 6.36139345e+00 6.39079809e-01
-8.93305361e-01 5.83162189e-01 4.23865944e-01 -1.03235207e-01
1.84499010e-01 -5.60650229e-01 -5.98834336e-01 3.74578923e-01
5.60288787e-01 -1.03888623e-01 -2.44284794e-01 9.94604409e-01
-8.57998282e-02 1.11124190e-02 -1.09875584e+00 7.89645135e-01
5.21694683e-02 -1.23050129e+00 -3.28923523e-01 4.63928655e-02
8.98012578e-01 5.85633695e-01 1.58022866e-01 2.85829306e-01
4.05291282e-03 -1.11984515e+00 2.09043380e-02 4.27158386e-01
9.06983495e-01 -2.61926323e-01 1.31048882e+00 3.75889391e-01
-1.19374382e+00 -2.22824335e-01 -3.44378263e-01 6.87606156e-01
7.77903944e-02 3.54652584e-01 -1.42757487e+00 3.04965079e-01
8.28007638e-01 8.11142921e-01 -7.17391789e-01 1.69957471e+00
-2.42175415e-01 6.50462210e-01 -3.36239278e-01 -1.23565502e-01
4.31369334e-01 5.02756715e-01 6.72174752e-01 1.13011944e+00
7.76036263e-01 6.73132956e-01 1.58299714e-01 9.00692761e-01
-2.39226043e-01 3.24582495e-02 -4.03604954e-01 -1.07181288e-01
1.81033939e-01 1.32365012e+00 -8.34795177e-01 -3.90491158e-01
-8.18638563e-01 7.08236694e-01 4.66969162e-02 -1.87852827e-03
-1.15639162e+00 5.18133231e-02 -6.52049705e-02 3.24830920e-01
5.50388098e-01 2.16555610e-01 -2.55401641e-01 -7.71675348e-01
-2.14260742e-02 -3.11806619e-01 7.82753229e-01 -9.38130319e-01
-1.39459109e+00 6.96461976e-01 -3.84382129e-01 -1.35405099e+00
9.04757157e-02 -7.41249919e-01 -8.46514761e-01 5.01393259e-01
-1.39172602e+00 -1.09924376e+00 -1.00480521e+00 4.37577128e-01
6.79845393e-01 -8.72739106e-02 6.66759551e-01 2.25361198e-01
-4.87534821e-01 3.96282762e-01 9.62465852e-02 2.93736309e-01
5.80414474e-01 -1.36690259e+00 2.53207475e-01 5.27637839e-01
-2.26795614e-01 9.54561532e-02 -3.14982049e-02 -7.27777779e-01
-9.01955187e-01 -1.35415351e+00 5.13778925e-01 -3.08842301e-01
4.72055078e-01 3.44380736e-01 -1.12730324e+00 9.88710582e-01
6.26378506e-02 8.20840657e-01 5.86883724e-01 -5.10490596e-01
3.13909441e-01 4.66917902e-01 -1.18422759e+00 4.67401624e-01
1.39361382e+00 -9.57423374e-02 -5.49950957e-01 6.69415295e-01
9.14308310e-01 -8.94641221e-01 -8.41779530e-01 8.67282510e-01
3.13303709e-01 -8.64257574e-01 1.02872467e+00 -5.22370815e-01
4.01575267e-01 2.65115313e-02 1.73534483e-01 -9.92464721e-01
-8.84641349e-01 1.41018212e-01 -2.52043366e-01 3.09793919e-01
2.90992081e-01 -5.30436575e-01 1.06342864e+00 3.29200506e-01
-9.53969777e-01 -1.08666432e+00 -7.18894899e-01 -5.39112031e-01
-1.03216268e-01 -1.45409495e-01 2.90606767e-01 6.15938365e-01
-5.58652461e-01 -2.00438201e-01 3.09870571e-01 3.23251188e-01
5.01478136e-01 -1.11638166e-01 3.14961225e-02 -1.08790183e+00
-5.54238677e-01 -6.04205012e-01 -4.18799132e-01 -8.19673359e-01
-4.04055536e-01 -1.14671767e+00 1.80232413e-02 -1.73878431e+00
7.29662776e-01 -2.12746873e-01 -4.90708828e-01 4.39059824e-01
-3.48543376e-01 3.50511037e-02 -1.92279935e-01 2.78725088e-01
-7.40727544e-01 3.51749003e-01 2.02716327e+00 -9.71618388e-03
8.84102955e-02 3.52070898e-01 -5.49651265e-01 9.70899999e-01
6.01752937e-01 -6.65052652e-01 -2.02388242e-01 -2.74339080e-01
-5.35292804e-01 1.44844726e-01 5.53797364e-01 -1.18297625e+00
5.87349534e-01 2.55418289e-02 6.88275933e-01 -1.02597356e+00
5.86287826e-02 -8.94312918e-01 -1.50034592e-01 1.00244284e+00
-2.64137924e-01 -4.51691508e-01 1.21984519e-01 7.24824667e-01
-4.17001918e-02 -3.38541538e-01 1.05529630e+00 -5.77411830e-01
-8.62410903e-01 8.52492392e-01 -4.20582920e-01 1.82179332e-01
1.30955839e+00 -1.51239410e-01 -2.56578952e-01 -1.72236208e-02
-9.72925901e-01 1.63366765e-01 -3.07218824e-02 8.78210068e-02
7.71728992e-01 -1.42544341e+00 -7.45786011e-01 3.08612168e-01
1.32471249e-01 7.09715664e-01 5.69762290e-01 1.48820567e+00
-7.45131552e-01 8.65323842e-01 -2.74232738e-02 -1.05007863e+00
-1.30337882e+00 4.23083663e-01 7.23565280e-01 -5.56588054e-01
-1.07240617e+00 1.28421748e+00 6.37785494e-01 -1.44439161e-01
4.15885329e-01 -9.29922700e-01 3.27343941e-02 -5.45370281e-01
4.93047804e-01 1.88409507e-01 3.18465590e-01 -1.01967365e-01
-3.87417644e-01 3.59411687e-01 -3.11536700e-01 5.13087928e-01
1.22145629e+00 1.76660031e-01 3.25622886e-01 6.53643301e-03
8.86222064e-01 -4.14363265e-01 -1.14679170e+00 -6.54664278e-01
-9.23919231e-02 -5.90348780e-01 3.96254063e-01 -8.31353128e-01
-1.35642159e+00 7.59452164e-01 8.30730498e-01 -2.42173523e-02
1.00807488e+00 5.52045107e-01 9.41859543e-01 4.48565632e-01
6.26388416e-02 -4.50622022e-01 6.38228476e-01 3.63519490e-01
9.94406223e-01 -1.59951735e+00 -3.41144055e-02 -6.23272836e-01
-7.33265042e-01 9.66461480e-01 9.95805204e-01 -8.45886990e-02
7.69851983e-01 2.77048320e-01 -2.28683472e-01 -4.87515599e-01
-7.70396590e-01 -4.87182349e-01 6.61968648e-01 4.92002398e-01
4.87535685e-01 6.90391064e-02 -5.86437574e-03 1.04119039e+00
3.62181664e-01 6.64102286e-02 4.22603846e-01 1.04011405e+00
-7.19357014e-01 -5.92614174e-01 -2.65054047e-01 1.13656259e+00
-6.90393567e-01 -4.24948484e-02 -3.07608813e-01 1.14379919e+00
9.39737111e-02 1.74664646e-01 1.82918593e-01 1.06713250e-01
3.85228068e-01 -2.06216887e-01 5.17329633e-01 -1.00366306e+00
-6.51400447e-01 1.99883118e-01 -1.02970123e-01 -3.52506191e-01
-6.21603131e-02 -7.87401736e-01 -1.57249522e+00 5.42372353e-02
-5.72774470e-01 -3.68862331e-01 3.20260257e-01 7.07504749e-01
1.03912786e-01 1.28795636e+00 3.54509115e-01 -6.05827272e-01
-8.31356525e-01 -1.07780671e+00 -6.45837903e-01 2.37694934e-01
2.45631561e-01 -7.99451649e-01 -2.67487258e-01 -2.99921095e-01] | [15.387787818908691, -2.1294288635253906] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.