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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
474b02c3-0a26-4279-9c2a-8864fa503b73 | microsoft-ai-challenge-india-2018-learning-to | 1906.06056 | null | https://arxiv.org/abs/1906.06056v1 | https://arxiv.org/pdf/1906.06056v1.pdf | Microsoft AI Challenge India 2018: Learning to Rank Passages for Web Question Answering with Deep Attention Networks | This paper describes our system for The Microsoft AI Challenge India 2018: Ranking Passages for Web Question Answering. The system uses the biLSTM network with co-attention mechanism between query and passage representations. Additionally, we use self attention on embeddings to increase the lexical coverage by allowing the system to take union over different embeddings. We also incorporate hand-crafted features to improve the system performance. Our system achieved a Mean Reciprocal Rank (MRR) of 0.67 on eval-1 dataset. | ['Chaitanya Sai Alaparthi'] | 2019-06-14 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [-2.62882888e-01 6.35263231e-03 -4.57297787e-02 -2.84752280e-01
-1.42053890e+00 -5.76703370e-01 6.95706010e-01 4.70833510e-01
-1.01744854e+00 5.49098492e-01 6.96657300e-01 -5.43449402e-01
1.71948686e-01 -8.01500857e-01 -8.62160087e-01 2.85733901e-02
-9.19871777e-02 3.07418823e-01 6.15225434e-01 -5.02965510e-01
3.16020548e-01 -3.59341830e-01 -1.11281157e+00 7.26074994e-01
9.42129016e-01 1.03100336e+00 1.71461791e-01 1.26065552e+00
-3.46059203e-01 1.17605591e+00 -8.03663731e-01 -7.67338276e-01
-9.22928229e-02 -2.31228739e-01 -1.38265324e+00 -7.91695952e-01
6.36612535e-01 -6.02738798e-01 -7.00060725e-01 6.83681965e-01
6.00074470e-01 3.95097554e-01 5.86072206e-01 -7.48944283e-01
-1.41362727e+00 6.83370769e-01 -2.14987665e-01 8.85515928e-01
6.15113795e-01 -1.29806399e-01 1.70803320e+00 -9.41642284e-01
3.57318789e-01 1.01951361e+00 3.25574070e-01 5.21813810e-01
-7.22316086e-01 -3.31057489e-01 -9.20839142e-03 7.87856877e-01
-1.00517452e+00 -1.66653439e-01 4.87273037e-01 -2.90673990e-02
1.58848011e+00 3.98706615e-01 2.58750647e-01 1.17636502e+00
-6.18407279e-02 9.89914536e-01 6.53032064e-01 -5.00820279e-01
-2.50147115e-02 -3.59172784e-02 5.03935456e-01 5.83340704e-01
-1.35088876e-01 -6.10525012e-01 -2.42156684e-01 -2.30216458e-01
3.53921682e-01 -1.83615953e-01 -1.84986979e-01 5.12707710e-01
-9.32394445e-01 1.08780265e+00 8.75809014e-01 2.77119994e-01
-4.66392100e-01 2.66635209e-01 6.08465791e-01 5.13217628e-01
3.41018289e-01 8.68620634e-01 -6.07913315e-01 -2.02233896e-01
-5.95130980e-01 1.51702374e-01 9.44166422e-01 5.70612013e-01
2.84178674e-01 -3.99742573e-01 -8.22751939e-01 1.18331707e+00
2.51010865e-01 3.58821064e-01 6.70362592e-01 -9.62814033e-01
5.79988658e-01 3.92098457e-01 -1.32039795e-02 -8.16012383e-01
-6.17931820e-02 -3.15664619e-01 -2.70965248e-01 -7.50425935e-01
2.10142061e-01 -7.77130425e-02 -1.06050491e+00 1.59549069e+00
-1.06885759e-02 -6.86845854e-02 1.90393507e-01 9.16152596e-01
1.33216810e+00 1.07537735e+00 3.46200496e-01 2.67711192e-01
1.48918080e+00 -1.52021551e+00 -7.55920172e-01 -3.24540026e-02
7.08372951e-01 -7.46638358e-01 1.51785219e+00 -1.69499829e-01
-9.23296034e-01 -3.14205021e-01 -9.00065720e-01 -5.16380250e-01
-6.02110326e-01 2.90587470e-02 2.01563239e-01 9.48907658e-02
-1.12128901e+00 2.07318902e-01 -5.75888276e-01 -4.33948398e-01
1.69169694e-01 7.56301582e-02 -8.34754780e-02 -1.09998889e-01
-1.96477950e+00 1.06820226e+00 1.47911131e-01 -1.04098082e-01
-6.83881938e-01 -6.06271982e-01 -5.86562514e-01 8.48498046e-02
1.12028820e-02 -7.51029432e-01 1.44882298e+00 -5.23442149e-01
-1.43819296e+00 7.58397937e-01 -2.65738994e-01 -8.23092639e-01
9.93208587e-02 -5.68832695e-01 -6.32700622e-01 4.56992567e-01
1.51096880e-01 6.44747615e-01 1.96203932e-01 -5.62688291e-01
-3.83486271e-01 2.06612185e-01 6.05416298e-01 4.10234749e-01
-5.93904018e-01 4.58042137e-02 -7.28823602e-01 -3.83838177e-01
-3.29716206e-01 -5.48418641e-01 -3.07491049e-03 -5.49643874e-01
-3.19937378e-01 -8.40575993e-01 5.71581423e-01 -1.36917615e+00
1.43767512e+00 -1.74346888e+00 -3.45104188e-01 -1.07630931e-01
1.13014475e-01 3.08867812e-01 -7.17762589e-01 7.40780652e-01
4.45624292e-01 3.20746899e-01 5.29084615e-02 4.52901199e-02
-5.70543036e-02 9.13615003e-02 -5.49292207e-01 -8.31347704e-02
3.00882161e-01 1.21477497e+00 -9.52446222e-01 -4.04813260e-01
-3.16311568e-01 4.62510705e-01 -6.56275511e-01 3.84573609e-01
-4.83473748e-01 -2.05842987e-01 -6.28632963e-01 3.20359439e-01
1.59305230e-01 -4.17398989e-01 -3.78599495e-01 5.03188483e-02
2.36030251e-01 1.20435596e+00 -3.19595307e-01 1.70358825e+00
-6.72465265e-01 6.58192277e-01 -3.44425529e-01 -5.14458776e-01
6.23794734e-01 3.26326340e-01 1.71992838e-01 -1.07129633e+00
-2.95434780e-02 7.40679950e-02 -1.02627978e-01 -6.73384964e-01
7.42429376e-01 4.04109240e-01 -1.28911650e-02 5.85375786e-01
1.23649888e-01 1.70596436e-01 2.88535625e-01 7.00839937e-01
1.43542659e+00 -1.80119112e-01 -1.95154428e-01 -1.84554040e-01
4.83373195e-01 1.24071062e-01 -1.02659594e-02 8.02219808e-01
-2.25470707e-01 6.24738812e-01 3.34664285e-01 -1.61330044e-01
-1.09756982e+00 -1.04583693e+00 4.72950004e-02 1.63548267e+00
4.49159443e-02 -3.88045430e-01 -6.52452052e-01 -9.95626450e-01
-7.00210035e-02 8.83045971e-01 -7.95927465e-01 -3.31524163e-01
-6.05446041e-01 -5.17024577e-01 7.42664754e-01 5.27961195e-01
5.12137771e-01 -1.15049648e+00 -3.21251810e-01 3.94996792e-01
-6.61093533e-01 -1.12312269e+00 -7.46774435e-01 4.45793429e-03
-5.17705023e-01 -9.24535155e-01 -8.89815092e-01 -1.04770935e+00
1.84143916e-01 6.07862845e-02 1.30215132e+00 1.28998339e-01
2.05460731e-02 4.09168124e-01 -6.44635201e-01 -1.42061323e-01
3.00395191e-02 5.89293957e-01 -2.29193956e-01 -5.06454527e-01
6.78633332e-01 -2.30133668e-01 -8.21207762e-01 -3.33286189e-02
-8.61694515e-01 -3.32845986e-01 1.75918952e-01 9.87455010e-01
3.88410956e-01 -6.51455283e-01 1.08792484e+00 -5.79191744e-01
1.27835238e+00 -7.14287937e-01 -1.84799522e-01 7.15433240e-01
-4.14908230e-01 2.10789219e-01 5.63641071e-01 -5.67572236e-01
-6.97101712e-01 -7.40883708e-01 -5.02089083e-01 -3.80817465e-02
4.72280294e-01 6.52516961e-01 2.88207620e-01 3.32753450e-01
6.87432170e-01 9.40701440e-02 -4.27755833e-01 -5.72646081e-01
7.49137282e-01 1.10232389e+00 4.29991961e-01 -4.15884405e-01
2.25359932e-01 -2.15531901e-01 -7.96116531e-01 -6.59530103e-01
-1.38755810e+00 -4.68248039e-01 -1.41567841e-01 6.08557798e-02
1.06215310e+00 -8.66067767e-01 -7.22569585e-01 -1.83547456e-02
-1.34488893e+00 -2.54541278e-01 -8.14810842e-02 3.49239379e-01
-5.26516512e-02 1.32659510e-01 -1.21607053e+00 -5.18857121e-01
-1.01632202e+00 -8.33381593e-01 6.87834620e-01 5.11816740e-01
-2.19028130e-01 -9.83140945e-01 4.36336100e-01 6.92015707e-01
8.40198755e-01 -1.76136464e-01 7.88129926e-01 -1.30683255e+00
-6.11511692e-02 -3.35870862e-01 -4.30301636e-01 4.22007293e-01
-1.46187097e-01 -1.40555575e-01 -9.16037083e-01 -1.02653950e-01
-4.70694363e-01 -6.93390489e-01 1.30383241e+00 1.48716271e-01
1.29330862e+00 -6.37172759e-01 7.21759424e-02 -6.83505908e-02
1.17098546e+00 -1.10105760e-01 3.98195177e-01 6.89799309e-01
6.16306007e-01 5.18267930e-01 1.98179826e-01 7.40398541e-02
1.02375782e+00 3.87061387e-01 2.74476141e-01 2.17223212e-01
-2.36057013e-01 -6.54125273e-01 3.27676088e-01 1.32304263e+00
3.65361094e-01 -4.24815476e-01 -1.02823055e+00 8.49200726e-01
-1.60585594e+00 -8.68169963e-01 -1.38511255e-01 1.91148996e+00
9.30470169e-01 1.39588967e-01 -7.01930672e-02 -3.88354599e-01
7.36387670e-01 2.37265110e-01 -4.57571208e-01 -7.45463729e-01
-1.02817796e-01 3.43531430e-01 4.07655120e-01 8.23005736e-01
-1.11068380e+00 1.14167440e+00 7.02979374e+00 6.74454212e-01
-9.03852582e-01 3.53802294e-01 5.54190218e-01 -8.57431144e-02
-6.32015705e-01 -2.99104452e-01 -5.62181532e-01 5.32146692e-01
1.44491088e+00 -3.93750191e-01 2.96319574e-01 5.73652983e-01
-3.61860454e-01 7.09662512e-02 -6.67704821e-01 4.84592766e-01
2.40083680e-01 -1.26078546e+00 1.96985975e-01 -3.51893812e-01
6.76043868e-01 7.80198336e-01 5.20350114e-02 8.87182117e-01
5.44902146e-01 -1.13633728e+00 3.98862511e-01 3.95064473e-01
6.00891113e-01 -6.53976262e-01 8.76727879e-01 1.93565816e-01
-7.13152289e-01 2.89443936e-02 -5.15478671e-01 -7.14995638e-02
2.10442394e-01 4.03543711e-01 -9.32945549e-01 2.49909699e-01
7.86565900e-01 3.31465006e-01 -8.41698825e-01 1.06738687e+00
-4.00023729e-01 9.95281696e-01 -4.75615233e-01 -6.29518509e-01
5.81017375e-01 1.39831737e-01 3.86065871e-01 1.35116005e+00
-1.74205918e-02 -2.26044562e-02 -1.22871295e-01 4.78346497e-01
-6.77797854e-01 3.32531691e-01 -3.43905270e-01 -1.88954577e-01
6.69248044e-01 1.05907226e+00 -4.34359089e-02 -6.63054049e-01
-3.70751113e-01 1.17097235e+00 6.92174733e-01 5.73172867e-01
-7.89799154e-01 -9.20249164e-01 4.00081307e-01 -9.27670002e-02
4.40732360e-01 -9.64271426e-02 7.58903334e-03 -1.33775377e+00
9.15125087e-02 -5.43871999e-01 7.59367108e-01 -8.47945392e-01
-1.63753891e+00 9.05638695e-01 -4.01089936e-01 -5.81229687e-01
-2.68655092e-01 -3.12064409e-01 -6.79001629e-01 1.03444040e+00
-1.72738934e+00 -1.00706184e+00 2.41901919e-01 2.71108180e-01
6.66792631e-01 -7.74942636e-02 9.75980937e-01 4.66728628e-01
-5.17011344e-01 8.07706475e-01 3.32972825e-01 5.13189614e-01
8.37101340e-01 -1.36734176e+00 6.79990709e-01 6.38622165e-01
3.96181077e-01 7.84227908e-01 4.86780465e-01 -3.95158291e-01
-1.24206161e+00 -1.07020223e+00 1.41063845e+00 -7.01183140e-01
1.00248253e+00 -1.94248706e-01 -1.10824716e+00 6.41148567e-01
6.82817519e-01 -4.94487584e-02 9.02754426e-01 4.42156732e-01
-8.59995008e-01 1.25674682e-03 -9.69962299e-01 5.65585673e-01
4.35549438e-01 -1.11651266e+00 -1.09773827e+00 3.81973505e-01
1.53006995e+00 -2.00363517e-01 -8.80540848e-01 4.17796135e-01
4.32260841e-01 -1.21785298e-01 9.45417285e-01 -9.35689807e-01
6.43627226e-01 -8.91283602e-02 -3.19458485e-01 -1.20977008e+00
-2.63053805e-01 -1.48557544e-01 -6.28089130e-01 1.18334186e+00
8.61214876e-01 -3.97403747e-01 3.93095553e-01 5.10812044e-01
5.06966487e-02 -1.05047703e+00 -9.61003721e-01 -4.44682598e-01
5.08682191e-01 -4.00399774e-01 4.17232096e-01 8.00451279e-01
1.62380487e-01 9.23041224e-01 -1.75599232e-02 7.98807964e-02
1.28788978e-01 -7.93415233e-02 4.51944359e-02 -7.12726414e-01
-1.64802700e-01 -5.41753769e-01 6.78592175e-02 -1.32069159e+00
1.86448053e-01 -1.11030853e+00 -1.62910372e-01 -2.00530481e+00
3.27057689e-01 1.51510805e-01 -8.88049603e-01 4.41051811e-01
-6.10544801e-01 4.82531846e-01 -3.81205305e-02 9.30177234e-03
-1.29919219e+00 7.54986465e-01 1.03552973e+00 -2.86146134e-01
8.37966725e-02 -3.48489225e-01 -9.49174106e-01 1.54946163e-01
1.03291488e+00 -3.30621064e-01 -1.26034588e-01 -1.22355151e+00
3.83590519e-01 -1.91416264e-01 3.47853526e-02 -6.04938328e-01
3.33256096e-01 2.80708313e-01 2.59187549e-01 -4.57198799e-01
1.23051405e-01 -2.97491997e-01 -6.91748261e-01 3.40093255e-01
-1.08826745e+00 3.74305487e-01 2.09326476e-01 4.77560937e-01
-4.03236091e-01 -1.58309877e-01 3.90391290e-01 -1.37011963e-03
-5.14075935e-01 2.42244452e-01 -2.96713442e-01 4.36646193e-01
4.22561824e-01 5.59465110e-01 -6.01581633e-01 -7.45649278e-01
-3.84565055e-01 9.31358576e-01 -7.93259516e-02 8.96820962e-01
5.38477063e-01 -1.56745553e+00 -9.49679554e-01 -3.46748441e-01
3.82465631e-01 -3.01711023e-01 1.44413352e-01 3.89837712e-01
-6.47965372e-01 6.66307330e-01 2.34971941e-01 -1.28065586e-01
-9.80440199e-01 1.81209549e-01 1.75032392e-01 -5.86438417e-01
-3.49524915e-01 1.17541826e+00 -3.64178866e-01 -7.25011408e-01
3.56380641e-01 -8.10790360e-02 -6.70898497e-01 -1.41768456e-02
8.89905632e-01 3.61010551e-01 1.27553582e-01 -2.57495880e-01
-4.63068873e-01 1.32125556e-01 -3.80497426e-01 -4.44152445e-01
1.27761090e+00 -2.89534211e-01 -1.03888623e-01 5.08722782e-01
1.71816802e+00 -1.36866882e-01 -6.80301309e-01 -5.07593215e-01
3.79146606e-01 1.35148734e-01 2.84825474e-01 -1.29888165e+00
-5.04992843e-01 8.81091654e-01 2.94407606e-01 3.78213793e-01
7.92196155e-01 1.36661693e-01 1.44531834e+00 7.29579806e-01
-2.57358134e-01 -1.05749893e+00 2.24623457e-01 1.20583129e+00
8.37504268e-01 -1.34004891e+00 -3.78410429e-01 4.92662460e-01
-7.02541828e-01 9.89629209e-01 7.69854486e-01 -3.97544026e-01
4.57932949e-01 -2.63407856e-01 2.24586323e-01 -1.29946575e-01
-1.21840405e+00 -3.28387558e-01 7.36629367e-01 1.28017753e-01
7.42204666e-01 -5.40522002e-02 -6.56950593e-01 7.48579741e-01
-3.99452060e-01 -2.48775408e-01 3.95315677e-01 6.30108058e-01
-6.98677003e-01 -7.83214808e-01 2.17500329e-01 6.66642189e-01
-1.09690523e+00 -5.31019032e-01 -2.50026464e-01 2.43668929e-01
-5.47148168e-01 1.19379723e+00 3.42783630e-01 -5.59671104e-01
3.57248962e-01 2.87517786e-01 2.54815519e-01 -5.36563277e-01
-7.16306686e-01 -4.22573745e-01 3.68499339e-01 -4.19306099e-01
7.26151243e-02 -1.21398002e-01 -1.18147266e+00 -3.00477948e-02
-3.78297925e-01 6.00224853e-01 6.90692306e-01 7.97750175e-01
5.56905329e-01 5.78485608e-01 6.07182205e-01 1.73368543e-01
-5.02507329e-01 -1.51679695e+00 -3.38173397e-02 2.36778900e-01
5.31190634e-01 -2.47118324e-02 -3.58835667e-01 -2.93874711e-01] | [11.329193115234375, 7.890735149383545] |
dc1bbf47-d87a-4535-8a57-d26ed8691446 | bowler-a-neural-approach-to-extractive-text | null | null | https://aclanthology.org/Y18-1017 | https://aclanthology.org/Y18-1017.pdf | BoWLer: A neural approach to extractive text summarization | null | ['Manish Shrivastava', 'Pranav Dhakras'] | null | null | null | null | paclic-2018-12 | ['extractive-document-summarization'] | ['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] | [-7.216468811035156, 3.598571300506592] |
c44bc9dd-ff51-4e1b-89cb-1225f52b6fc6 | musicbert-symbolic-music-understanding-with | 2106.05630 | null | https://arxiv.org/abs/2106.05630v1 | https://arxiv.org/pdf/2106.05630v1.pdf | MusicBERT: Symbolic Music Understanding with Large-Scale Pre-Training | Symbolic music understanding, which refers to the understanding of music from the symbolic data (e.g., MIDI format, but not audio), covers many music applications such as genre classification, emotion classification, and music pieces matching. While good music representations are beneficial for these applications, the lack of training data hinders representation learning. Inspired by the success of pre-training models in natural language processing, in this paper, we develop MusicBERT, a large-scale pre-trained model for music understanding. To this end, we construct a large-scale symbolic music corpus that contains more than 1 million music songs. Since symbolic music contains more structural (e.g., bar, position) and diverse information (e.g., tempo, instrument, and pitch), simply adopting the pre-training techniques from NLP to symbolic music only brings marginal gains. Therefore, we design several mechanisms, including OctupleMIDI encoding and bar-level masking strategy, to enhance pre-training with symbolic music data. Experiments demonstrate the advantages of MusicBERT on four music understanding tasks, including melody completion, accompaniment suggestion, genre classification, and style classification. Ablation studies also verify the effectiveness of our designs of OctupleMIDI encoding and bar-level masking strategy in MusicBERT. | ['Tie-Yan Liu', 'Tao Qin', 'Zeqian Ju', 'Rui Wang', 'Xu Tan', 'Mingliang Zeng'] | 2021-06-10 | null | https://aclanthology.org/2021.findings-acl.70 | https://aclanthology.org/2021.findings-acl.70.pdf | findings-acl-2021-8 | ['genre-classification'] | ['computer-vision'] | [ 3.85749102e-01 -2.23222986e-01 -2.94035494e-01 -2.73398936e-01
-6.11274898e-01 -8.72742653e-01 8.33280832e-02 -3.28375995e-02
-6.06520176e-02 2.15158716e-01 3.34142834e-01 -1.19514935e-01
-4.76142704e-01 -5.24781764e-01 -4.89173353e-01 -2.70800292e-01
-4.59379032e-02 1.82634190e-01 -2.92935967e-01 -3.49670708e-01
5.11971533e-01 1.50279656e-01 -1.83627200e+00 7.31976926e-01
6.97015882e-01 1.22088134e+00 2.67852515e-01 6.80556774e-01
-4.03018147e-01 5.81717134e-01 -7.75268376e-01 -1.70233831e-01
2.29486972e-01 -7.51511574e-01 -7.27043152e-01 -3.84195983e-01
6.39484227e-02 -2.77332515e-01 -2.13971481e-01 9.79838967e-01
5.03044307e-01 1.70179114e-01 6.15388274e-01 -1.11699796e+00
-4.13253844e-01 1.37339234e+00 -3.89258653e-01 -1.07031450e-01
6.38744891e-01 -1.95721582e-01 1.32292295e+00 -7.22741961e-01
1.94267794e-01 1.21611691e+00 8.82005274e-01 4.03430164e-01
-9.70915258e-01 -1.25527585e+00 -1.15936860e-01 4.01577652e-01
-1.63212383e+00 -4.85664427e-01 1.09603584e+00 -4.21956301e-01
6.62171066e-01 6.79616392e-01 7.89590955e-01 8.44144106e-01
-3.87206405e-01 9.36174572e-01 7.72870958e-01 -6.56601191e-01
-9.84234437e-02 -8.56654644e-02 8.95352215e-02 1.18774913e-01
-2.98735797e-01 1.58447906e-01 -1.02958763e+00 -1.18157350e-01
1.10810459e+00 -2.26212740e-01 -2.82210678e-01 9.98590365e-02
-1.39816380e+00 5.53586066e-01 2.18797967e-01 3.00124556e-01
-6.31451085e-02 2.00048178e-01 8.32507253e-01 5.73156953e-01
-1.94999069e-01 1.02390432e+00 -4.81347233e-01 -7.44715452e-01
-1.08954251e+00 2.42298618e-01 6.44478738e-01 1.04099476e+00
4.15109962e-01 4.20831174e-01 -6.40217140e-02 1.23329222e+00
1.97329260e-02 4.21555907e-01 8.65168393e-01 -1.01955783e+00
4.16169971e-01 5.94114959e-01 -2.51400381e-01 -9.37031507e-01
-3.98524970e-01 -6.08259737e-01 -9.21951771e-01 -2.47002199e-01
3.10691744e-01 1.50471047e-01 -4.28871900e-01 2.10051799e+00
-7.60805532e-02 5.93897581e-01 -1.80926062e-02 9.81582880e-01
8.58679712e-01 6.38801754e-01 -3.14990163e-01 -3.02512497e-01
1.39960039e+00 -6.54552639e-01 -7.09977448e-01 1.66938622e-02
4.82264847e-01 -9.31501806e-01 1.55012870e+00 9.51857150e-01
-8.90987277e-01 -1.00750613e+00 -1.14979422e+00 -1.95166767e-01
-1.38508797e-01 4.36742991e-01 8.14751208e-01 5.59188008e-01
-1.29403874e-01 1.03403926e+00 -4.53832537e-01 -1.81117300e-02
9.42347571e-02 5.30707181e-01 -1.13401838e-01 3.01275522e-01
-1.21822464e+00 1.09946169e-01 8.04548621e-01 -7.43924081e-02
-4.80416536e-01 -7.73951769e-01 -5.94771028e-01 2.50225633e-01
2.61572510e-01 -3.87398690e-01 1.34892046e+00 -1.12942040e+00
-1.52481413e+00 5.71710110e-01 7.32499212e-02 -2.75703371e-01
-2.03836456e-01 -3.58583361e-01 -5.47331750e-01 -6.70641810e-02
-1.58381715e-01 4.97292370e-01 9.74410534e-01 -7.76918113e-01
-5.18141329e-01 -2.11119831e-01 1.05381608e-01 2.80867130e-01
-6.58297837e-01 8.09487179e-02 -3.26470524e-01 -1.43200946e+00
4.74568188e-01 -1.04659939e+00 3.32133770e-01 -4.44697887e-01
-5.33280432e-01 -6.81503909e-03 6.07966363e-01 -6.91467106e-01
1.64684820e+00 -2.63263249e+00 2.33015925e-01 4.14127976e-01
-3.14226866e-01 2.25166939e-02 -3.08268756e-01 3.43483329e-01
-3.64233106e-01 3.99410762e-02 -7.54558817e-02 6.21228386e-03
2.94468820e-01 2.83583790e-01 -8.20154965e-01 -1.92372903e-01
-1.22774154e-01 8.28076780e-01 -6.89088762e-01 -2.57828116e-01
-1.00793518e-01 1.01718828e-01 -8.02547812e-01 1.63103659e-02
-2.56966293e-01 5.71856558e-01 -3.74971390e-01 8.00102174e-01
2.03224361e-01 1.16005532e-01 9.98763088e-03 -4.90624338e-01
1.09103784e-01 9.00114596e-01 -1.46891940e+00 2.29327774e+00
-3.44704151e-01 4.63997692e-01 -3.02136421e-01 -8.15289915e-01
1.07563984e+00 4.86323416e-01 4.94372755e-01 -6.48181856e-01
1.23761676e-01 3.43585700e-01 4.87553537e-01 -2.58868843e-01
5.93821466e-01 -3.45520496e-01 -4.08244610e-01 6.60704374e-01
4.43429872e-03 -2.65726388e-01 1.80268019e-01 -9.16954949e-02
8.36044192e-01 1.26905888e-01 3.93128037e-01 1.33033842e-01
4.70144033e-01 -2.78016739e-03 7.28160381e-01 6.54352963e-01
2.88883239e-01 6.94778442e-01 3.54479253e-01 -7.70579055e-02
-7.69525766e-01 -8.54527354e-01 -3.94332372e-02 1.62419534e+00
2.52854265e-02 -1.08856058e+00 -5.94380140e-01 -1.23407088e-01
7.05666021e-02 5.03321052e-01 -3.22500944e-01 -4.78928685e-01
-7.81042099e-01 -3.46300155e-01 1.13651538e+00 6.97494566e-01
3.46229464e-01 -1.44541717e+00 -1.82391211e-01 4.06779200e-01
-3.86086613e-01 -7.29310930e-01 -7.51800537e-01 2.18419492e-01
-1.10643995e+00 -1.01371264e+00 -2.82265812e-01 -7.69221127e-01
-1.48396671e-01 1.75391793e-01 1.08082688e+00 7.53997490e-02
-3.01281542e-01 9.24594551e-02 -7.09108770e-01 -6.32877290e-01
-2.66463637e-01 3.01753610e-01 3.67730588e-01 -1.23600140e-01
1.97359249e-01 -1.05688596e+00 -3.04157704e-01 3.40922356e-01
-7.45177984e-01 3.43030632e-01 6.54914558e-01 8.28229666e-01
8.20525825e-01 1.98227018e-01 9.41663325e-01 -6.76225841e-01
8.22297215e-01 1.16501212e-01 8.52771010e-03 4.06414643e-02
-2.59489387e-01 -1.74028218e-01 6.90571785e-01 -1.04976296e+00
-5.58516979e-01 -1.99173614e-02 -1.42263800e-01 -6.17306828e-01
-1.24544315e-02 8.87398422e-01 -2.76925504e-01 1.64848998e-01
7.11886644e-01 3.53662848e-01 -1.36716768e-01 -9.68899906e-01
3.79013419e-01 9.19062674e-01 1.07296085e+00 -8.64121199e-01
7.39847422e-01 -9.03207511e-02 -2.80492246e-01 -6.85775161e-01
-9.02963758e-01 -5.50703466e-01 -6.01339519e-01 7.73454607e-02
3.49600285e-01 -8.34880114e-01 -8.57452035e-01 1.63887233e-01
-8.73900473e-01 1.50629058e-02 -5.43418288e-01 7.07362294e-01
-9.17994022e-01 2.27699250e-01 -7.28150606e-01 -7.99938083e-01
-5.46878517e-01 -6.90137923e-01 9.00396943e-01 -5.14189936e-02
-9.66112494e-01 -3.01528871e-01 -1.51350498e-01 3.31626952e-01
1.76762953e-01 -4.12486792e-02 1.32735479e+00 -5.51936150e-01
-4.34691876e-01 -3.86626832e-02 8.01731721e-02 3.89810979e-01
2.54276812e-01 -4.00898248e-01 -1.14546096e+00 -9.41539407e-02
-2.42701010e-03 -4.96595889e-01 6.13876641e-01 7.17428550e-02
1.60385644e+00 -2.38532275e-01 1.24731168e-01 9.53794241e-01
7.06348002e-01 5.61574459e-01 6.33806884e-01 3.20642471e-01
7.30803370e-01 3.58508617e-01 9.28576708e-01 6.60209835e-01
-6.84337765e-02 9.58591998e-01 2.55489480e-02 2.42527425e-01
-2.75470585e-01 -5.92998266e-01 4.90804762e-01 1.49206126e+00
-1.44129589e-01 4.33280766e-01 -5.81936777e-01 1.95091218e-01
-1.70061719e+00 -1.00143266e+00 3.20827961e-01 2.08715725e+00
1.19627786e+00 7.49399439e-02 3.06781262e-01 9.64821517e-01
4.22590524e-01 -1.69667467e-01 -8.70341182e-01 -3.34816903e-01
-1.72456697e-01 4.72668111e-01 -1.66587308e-01 -1.16085425e-01
-9.61960495e-01 9.68611062e-01 5.86449480e+00 1.35641575e+00
-1.14061165e+00 -1.74358770e-01 4.89716418e-03 -3.02028567e-01
-2.40496054e-01 -1.77463666e-02 -4.19281155e-01 3.47806185e-01
5.62450171e-01 -1.68169945e-01 1.06345451e+00 6.54449105e-01
1.60404578e-01 4.81952459e-01 -1.35591435e+00 1.64468694e+00
-5.47699630e-02 -1.09789610e+00 3.39192063e-01 -2.22235963e-01
3.51895154e-01 -4.15837049e-01 -4.28659916e-02 7.56438971e-01
-5.68511844e-01 -1.16721869e+00 1.13466406e+00 3.96070242e-01
1.16062832e+00 -9.15056527e-01 4.09741163e-01 4.69771981e-01
-1.58651745e+00 -3.03005964e-01 -1.36032954e-01 -3.54154050e-01
-1.28497109e-02 1.29690275e-01 -5.52840889e-01 6.81049943e-01
5.71748734e-01 1.09278691e+00 -4.65381056e-01 9.90172982e-01
-1.49286285e-01 8.55657816e-01 -3.43570471e-01 2.95964092e-01
-3.33193183e-01 -2.25741729e-01 4.39984053e-01 1.05131078e+00
6.24335349e-01 3.22567642e-01 7.84831569e-02 8.97456527e-01
8.76580030e-02 3.20513040e-01 -3.11395168e-01 -6.22000277e-01
6.20570004e-01 8.70430171e-01 -4.50513810e-01 -2.31195003e-01
-4.34060916e-02 7.48150766e-01 -4.60592993e-02 1.83396667e-01
-7.41397858e-01 -6.01143241e-01 7.42873549e-01 3.81591171e-02
1.75445706e-01 -1.81049362e-01 -6.57380879e-01 -1.07975113e+00
-2.00280339e-01 -1.49244952e+00 3.80938202e-01 -8.81415069e-01
-1.12635601e+00 4.62753087e-01 -5.00744916e-02 -1.65235698e+00
-6.66557848e-01 -5.38760483e-01 -6.31071746e-01 7.49080658e-01
-8.97693217e-01 -1.05587888e+00 -9.53394026e-02 6.97322428e-01
5.56145966e-01 -4.58912224e-01 1.15825331e+00 4.32431638e-01
-2.24222049e-01 8.21297526e-01 -2.20640689e-01 3.56530160e-01
7.58034527e-01 -9.63389814e-01 1.39587909e-01 1.31326228e-01
9.72705781e-01 9.90459621e-01 4.22649980e-01 -3.21754336e-01
-1.49895382e+00 -7.22376227e-01 4.04310107e-01 -4.21393067e-02
5.21678865e-01 -4.45170999e-01 -9.82770264e-01 4.84009027e-01
-3.38695675e-01 -7.06750393e-01 1.18978608e+00 6.97150052e-01
-6.09598458e-01 -2.84031302e-01 -4.60460037e-01 6.86184108e-01
1.29044688e+00 -1.07733214e+00 -8.94496381e-01 -1.12876952e-01
7.11431563e-01 -4.42833781e-01 -9.17187691e-01 7.55416870e-01
1.07254088e+00 -6.08050346e-01 1.19405365e+00 -5.71304202e-01
2.72893608e-01 -4.45581287e-01 -4.34022546e-01 -1.13808644e+00
-3.22260499e-01 -8.07596207e-01 -2.79731840e-01 1.43689394e+00
2.88133979e-01 5.10428008e-03 6.07199728e-01 1.06965750e-02
-3.87110651e-01 -4.33673412e-01 -7.26634860e-01 -9.20375049e-01
-3.10362130e-01 -1.00286329e+00 9.11373973e-01 1.15748882e+00
4.68747377e-01 5.01609385e-01 -6.38252974e-01 -1.68921754e-01
1.16181761e-01 8.26700568e-01 1.06048596e+00 -1.35584235e+00
-9.60072756e-01 -6.15415812e-01 -3.94583136e-01 -1.21660912e+00
5.61774150e-02 -1.13492131e+00 -2.47756749e-01 -8.74958277e-01
1.46815166e-01 -5.94380856e-01 -8.49681556e-01 6.97969973e-01
1.14066489e-01 3.31175804e-01 5.28717756e-01 4.59166288e-01
-3.26706380e-01 6.03736222e-01 1.27480626e+00 -3.57635438e-01
-5.17607749e-01 8.71074647e-02 -7.79324830e-01 9.79108036e-01
9.20986354e-01 -5.27744949e-01 -6.19165659e-01 -1.84469014e-01
3.86008114e-01 2.22588047e-01 5.86831607e-02 -1.17669499e+00
1.36747465e-01 5.29694231e-03 2.30068594e-01 -7.70551026e-01
7.86348283e-01 -4.12359655e-01 3.23889345e-01 1.87225461e-01
-5.62820077e-01 -6.25662506e-02 5.66965938e-01 3.33885729e-01
-6.51890397e-01 -4.43347663e-01 2.72214115e-01 7.04540163e-02
-6.98927581e-01 -7.64518157e-02 1.26107767e-01 3.69831100e-02
1.47884876e-01 -4.07459587e-01 -1.59387626e-02 -4.60998952e-01
-8.89420629e-01 -1.81032658e-01 1.59112513e-01 6.93204999e-01
6.14775419e-01 -1.70655668e+00 -4.17364120e-01 5.42513847e-01
3.55577826e-01 -2.99104273e-01 1.82622582e-01 6.60953104e-01
-1.25085264e-02 1.92380875e-01 -2.19853282e-01 -4.43925828e-01
-1.48721147e+00 4.16634262e-01 -7.08713979e-02 1.15893081e-01
-7.05586910e-01 7.17121363e-01 3.98283094e-01 -5.34158111e-01
6.36370540e-01 -7.04095364e-01 -2.59543628e-01 -1.28129674e-02
6.69400752e-01 1.60981685e-01 -2.52677858e-01 -3.88741165e-01
-1.63039401e-01 8.65541101e-01 1.86084241e-01 -2.71802425e-01
1.04434812e+00 5.69796152e-02 -1.25395343e-01 9.87132311e-01
7.49139726e-01 2.45031372e-01 -6.14269316e-01 -8.78669620e-02
1.36766911e-01 -5.08580863e-01 -1.55713737e-01 -8.86659801e-01
-6.47449434e-01 9.36154723e-01 2.74865627e-01 9.40887555e-02
1.38414013e+00 -1.23328552e-01 9.61402953e-01 6.75555170e-01
4.49829608e-01 -1.08031595e+00 1.50755644e-01 8.17835510e-01
1.16940224e+00 -8.48768294e-01 -1.93081722e-01 -5.59689641e-01
-5.90027928e-01 1.03856409e+00 4.42903996e-01 1.22386381e-01
5.19968867e-01 1.73616603e-01 7.15288939e-03 1.20071154e-02
-6.22818232e-01 -1.96259856e-01 6.40647829e-01 2.42734879e-01
6.05717659e-01 1.40227973e-01 -6.48325384e-02 1.51814437e+00
-1.18632066e+00 7.60966912e-02 1.10425003e-01 5.19560158e-01
-4.41074461e-01 -1.23647702e+00 -6.40106022e-01 4.12411332e-01
-3.75161588e-01 -3.13106924e-01 -7.02694476e-01 7.20382571e-01
4.70272332e-01 8.90016675e-01 -3.98877598e-02 -8.76393795e-01
4.61882979e-01 3.15316111e-01 6.38377190e-01 -6.66866720e-01
-7.62678564e-01 4.85019684e-01 4.11257260e-02 -3.61694634e-01
-2.86713183e-01 -3.87435347e-01 -1.56741917e+00 -5.20100407e-02
-2.91788697e-01 2.58070827e-01 5.10693908e-01 7.81714618e-01
1.22819446e-01 7.73499131e-01 5.06259203e-01 -8.43982816e-01
-5.02032757e-01 -1.25444412e+00 -1.06997359e+00 6.36433303e-01
6.88391551e-03 -6.50525153e-01 -1.17336825e-01 8.81788060e-02] | [15.898619651794434, 5.411362648010254] |
0c951a68-60ab-4345-821a-f82e13749aba | tod-da-towards-boosting-the-robustness-of | 2112.12441 | null | https://arxiv.org/abs/2112.12441v1 | https://arxiv.org/pdf/2112.12441v1.pdf | TOD-DA: Towards Boosting the Robustness of Task-oriented Dialogue Modeling on Spoken Conversations | Task-oriented dialogue systems have been plagued by the difficulties of obtaining large-scale and high-quality annotated conversations. Furthermore, most of the publicly available datasets only include written conversations, which are insufficient to reflect actual human behaviors in practical spoken dialogue systems. In this paper, we propose Task-oriented Dialogue Data Augmentation (TOD-DA), a novel model-agnostic data augmentation paradigm to boost the robustness of task-oriented dialogue modeling on spoken conversations. The TOD-DA consists of two modules: 1) Dialogue Enrichment to expand training data on task-oriented conversations for easing data sparsity and 2) Spoken Conversation Simulator to imitate oral style expressions and speech recognition errors in diverse granularities for bridging the gap between written and spoken conversations. With such designs, our approach ranked first in both tasks of DSTC10 Track2, a benchmark for task-oriented dialogue modeling on spoken conversations, demonstrating the superiority and effectiveness of our proposed TOD-DA. | ['Haifeng Wang', 'Hua Wu', 'Fan Wang', 'Shuqi Sun', 'Jian Xie', 'Xiyuan Zhang', 'Qiang Ju', 'Liankai Huang', 'Huang He', 'Siqi Bao', 'Yingzhan Lin', 'Dongfeng He', 'Xinxian Huang', 'Xin Tian'] | 2021-12-23 | null | null | null | null | ['spoken-dialogue-systems'] | ['speech'] | [ 5.18972985e-02 3.22582155e-01 5.87973557e-02 -5.75017631e-01
-6.64280772e-01 -4.70310420e-01 9.19433355e-01 -3.14913720e-01
-2.58430332e-01 7.12594151e-01 8.71659279e-01 -3.58174741e-01
3.62531215e-01 -2.50217378e-01 6.27605692e-02 -4.24422383e-01
2.02729329e-01 9.17078018e-01 -1.85447499e-01 -7.48043418e-01
-1.46764383e-01 1.27167091e-01 -1.14538550e+00 6.05187356e-01
1.13208163e+00 8.63037765e-01 -3.33248377e-02 9.19827163e-01
-4.59447175e-01 9.59739447e-01 -8.27783465e-01 -3.60976279e-01
-2.87997480e-02 -6.95109546e-01 -1.01029563e+00 5.63860655e-01
1.22575641e-01 -5.87705612e-01 -4.73050386e-01 5.22084415e-01
7.48754263e-01 1.72016859e-01 5.58476210e-01 -1.27781177e+00
-3.21093798e-01 7.65579402e-01 -9.06333327e-02 -6.26629591e-02
4.16842788e-01 3.80893439e-01 1.00292766e+00 -1.07365036e+00
3.28798562e-01 1.65504301e+00 4.54794198e-01 1.13955426e+00
-1.19740152e+00 -3.69702369e-01 1.49232000e-01 -1.80124268e-01
-7.20751882e-01 -1.02988815e+00 8.85976911e-01 -3.78885865e-01
1.00595200e+00 4.28411037e-01 4.71462220e-01 1.61679995e+00
-6.13844395e-01 1.20634568e+00 9.79646981e-01 -3.65038931e-01
4.29702066e-02 4.10534501e-01 4.28476721e-01 3.39129269e-01
-6.73223257e-01 -2.24756837e-01 -7.56384850e-01 -3.84972483e-01
3.60448062e-01 -4.06960845e-01 -2.64239728e-01 1.00633288e-02
-1.22320521e+00 9.75230634e-01 -6.24132007e-02 1.10938668e-01
-3.67231935e-01 -6.68672383e-01 1.09590411e+00 4.55048233e-01
8.80956590e-01 6.20323002e-01 -6.92978203e-01 -7.13517010e-01
-3.48965526e-01 2.70804226e-01 1.27344477e+00 9.30824041e-01
2.63840407e-01 2.26845220e-01 -6.25619233e-01 1.66654766e+00
6.97157532e-03 2.51558036e-01 5.62369585e-01 -8.70918572e-01
8.08686018e-01 7.58758247e-01 -2.44534407e-02 -4.73390639e-01
-6.15772009e-01 6.76655471e-02 -1.16330254e+00 -4.66288537e-01
6.97900772e-01 -6.12660408e-01 -3.10619652e-01 1.80296779e+00
2.98888922e-01 -3.46076727e-01 5.70104241e-01 7.45422602e-01
1.18266106e+00 7.68491626e-01 -7.31580555e-02 -2.10467324e-01
1.34782171e+00 -1.44664681e+00 -1.09728277e+00 -2.91901469e-01
9.07608807e-01 -6.15200937e-01 1.64639878e+00 3.48980814e-01
-9.84824598e-01 -4.68042821e-01 -5.29669821e-01 -8.73721614e-02
-7.17752352e-02 1.48009598e-01 5.14963865e-01 6.40157461e-01
-7.37010837e-01 -1.25392869e-01 -3.28921020e-01 -3.02239805e-01
8.90824795e-02 1.29254702e-02 -2.74606407e-01 7.57316947e-02
-1.34039712e+00 8.28030467e-01 1.68145075e-01 1.86732337e-01
-7.46066093e-01 -6.90167010e-01 -9.81128752e-01 2.32945513e-02
4.25667584e-01 -1.69360518e-01 1.76997292e+00 -6.25649989e-01
-1.98960876e+00 7.68269956e-01 -2.65002131e-01 -5.01624405e-01
7.15367019e-01 -2.95547038e-01 -2.76289374e-01 -3.76178533e-01
-3.50281894e-01 6.59192026e-01 5.57627678e-01 -1.21422422e+00
-4.87975985e-01 -2.61309803e-01 -7.96918496e-02 4.86076266e-01
-8.00551474e-01 -1.06592298e-01 -3.41460556e-01 -4.93784785e-01
-3.02641720e-01 -8.22810352e-01 -1.66489989e-01 -3.59039098e-01
-6.63968742e-01 -6.73598289e-01 1.05370796e+00 -7.23018944e-01
1.20041049e+00 -2.07153177e+00 2.06966847e-01 -3.13756555e-01
2.79167771e-01 7.07432389e-01 -3.30100358e-01 5.57215750e-01
3.64188075e-01 -3.39654200e-02 -1.15204841e-01 -8.94827604e-01
3.29061598e-01 3.66002440e-01 -3.68107587e-01 -2.22780839e-01
3.50341707e-01 7.72007763e-01 -8.48165452e-01 -1.92438230e-01
2.18630329e-01 1.14566892e-01 -5.33316135e-01 9.18565750e-01
-5.53297222e-01 8.32716942e-01 -3.65191191e-01 2.96167821e-01
4.85255301e-01 8.97343010e-02 1.47225350e-01 -8.71621519e-02
-4.87107933e-02 9.36016202e-01 -5.97074687e-01 1.57696152e+00
-7.44572282e-01 7.24307835e-01 5.20811260e-01 -5.68021417e-01
1.25057244e+00 6.81575716e-01 2.97918886e-01 -6.94967687e-01
1.44835576e-01 6.23150282e-02 2.31091589e-01 -6.76605523e-01
7.62192249e-01 -2.22712625e-02 -3.04516494e-01 7.33100414e-01
1.17358975e-01 -4.58008766e-01 1.61009923e-01 1.69416174e-01
9.93662238e-01 -5.86202025e-01 1.04757741e-01 -4.80606928e-02
6.53869092e-01 -1.11880787e-01 5.54247022e-01 5.76251805e-01
-3.50796580e-01 4.40822154e-01 6.85760975e-01 -2.68573999e-01
-1.12985277e+00 -5.21272123e-01 -6.71133325e-02 1.55158353e+00
-5.34120440e-01 -3.07377309e-01 -9.42748785e-01 -1.03354037e+00
-9.10806432e-02 6.25172079e-01 -4.14222807e-01 -1.23417646e-01
-4.78129476e-01 -7.70827353e-01 9.61181641e-01 3.00141513e-01
6.24906957e-01 -1.28921652e+00 2.19296604e-01 3.66223395e-01
-5.96507490e-01 -1.48590374e+00 -8.20779204e-01 1.71096131e-01
-4.66537565e-01 -9.25905824e-01 -7.89721787e-01 -6.39616728e-01
3.26803535e-01 5.91610633e-02 1.23992205e+00 -1.00118704e-01
9.60699618e-02 1.94660202e-01 -5.63923955e-01 -3.76382440e-01
-1.22349238e+00 3.76367539e-01 1.47213250e-01 1.86770529e-01
3.84794325e-01 -1.77799046e-01 -8.32667425e-02 6.20667934e-01
-5.50624013e-01 2.34698743e-01 2.12101892e-01 1.34150612e+00
-3.64631176e-01 -4.40232098e-01 1.24413586e+00 -9.18568552e-01
1.33267570e+00 -2.77489394e-01 -1.31828859e-01 2.08864748e-01
-2.32611030e-01 -1.86084911e-01 7.14345813e-01 -6.52550280e-01
-1.34253848e+00 -2.93699712e-01 -5.69218993e-01 -1.32123451e-03
-3.84606272e-01 3.79334003e-01 -4.43172365e-01 4.34919685e-01
5.52530408e-01 3.43369782e-01 4.04357314e-01 -6.46702826e-01
2.90634483e-01 1.27322555e+00 4.26905751e-01 -7.86301911e-01
1.56952903e-01 -1.67801991e-01 -7.57846415e-01 -1.25955474e+00
-9.67810869e-01 -5.71160316e-01 -6.07011437e-01 -6.77574947e-02
6.69724286e-01 -8.73992145e-01 -9.10018504e-01 9.58397388e-01
-1.36534226e+00 -8.30178797e-01 -3.00121903e-01 1.92537963e-01
-5.21349072e-01 3.41164589e-01 -9.65200901e-01 -1.23883986e+00
-6.51593387e-01 -1.11393249e+00 8.35928857e-01 -3.40439752e-02
-5.62418938e-01 -9.99821305e-01 2.06207290e-01 9.12500143e-01
5.48513472e-01 -2.92372972e-01 1.00286531e+00 -1.37581897e+00
2.06097201e-01 -1.58092864e-02 -1.79174289e-01 7.60331213e-01
3.90920341e-01 -1.52677432e-01 -1.37687302e+00 -2.21106499e-01
3.32273543e-02 -1.03066516e+00 3.46986920e-01 -1.06090292e-01
8.35641742e-01 -6.39783084e-01 2.65716761e-01 1.01898171e-01
1.07255034e-01 2.39605308e-01 3.06724548e-01 -1.24270078e-02
7.41155088e-01 1.13385391e+00 6.64142787e-01 7.08615541e-01
7.41452694e-01 8.23481202e-01 8.18698332e-02 -3.22596192e-01
-1.31087095e-01 -6.85921088e-02 3.24894220e-01 1.34305847e+00
4.70145851e-01 -4.73959446e-01 -1.05159569e+00 7.55148649e-01
-1.90095305e+00 -5.20062149e-01 -1.44132972e-01 1.91047180e+00
1.40300643e+00 1.08205453e-01 6.14073694e-01 1.42054409e-02
5.59092939e-01 2.94711858e-01 -5.78140736e-01 -5.95273137e-01
-1.17201798e-01 -3.67934167e-01 -4.05271232e-01 4.82656687e-01
-1.04747415e+00 9.40642238e-01 5.63839293e+00 6.73266530e-01
-8.78996134e-01 1.23794250e-01 8.23504627e-01 6.55296892e-02
-1.07495248e-01 -5.41328847e-01 -9.01354611e-01 4.60603774e-01
9.55921471e-01 -7.75691867e-02 2.78163314e-01 7.34665573e-01
5.13070107e-01 2.74511307e-01 -1.31682646e+00 6.56036675e-01
-1.18436106e-01 -1.03471470e+00 1.30373850e-01 -4.57369722e-02
5.64828634e-01 -1.54983774e-01 -9.45234820e-02 8.39741707e-01
5.62608659e-01 -9.09442365e-01 2.35411182e-01 1.03283472e-01
4.40590113e-01 -6.01909697e-01 9.10171330e-01 7.38486826e-01
-7.32684255e-01 6.79354891e-02 3.08341924e-02 -2.00059433e-02
2.77590364e-01 3.37193549e-01 -1.48185050e+00 1.03570476e-01
3.19404095e-01 4.96838868e-01 -5.48438430e-02 4.88026172e-01
5.55414706e-02 8.70007753e-01 -2.56486565e-01 -2.06803620e-01
4.57818180e-01 -6.83052689e-02 5.16135275e-01 1.47880161e+00
-2.03799099e-01 4.17643227e-03 4.37741548e-01 6.59280837e-01
-3.91664237e-01 1.68401659e-01 -5.72562099e-01 -3.58551353e-01
7.99977899e-01 1.18469250e+00 1.51741892e-01 -2.26328805e-01
-4.66140717e-01 7.84765244e-01 3.62677932e-01 1.34435743e-01
-3.50047678e-01 1.50227686e-02 9.18645382e-01 -1.82248637e-01
-2.73565084e-01 -1.77627400e-01 -2.47686684e-01 -1.07112300e+00
-3.00725717e-02 -1.59663284e+00 1.94272071e-01 -2.87818313e-01
-1.67733860e+00 9.34927344e-01 -3.83905232e-01 -8.81285906e-01
-6.63453102e-01 -3.76138091e-01 -6.47743225e-01 9.33875024e-01
-1.19291329e+00 -1.14411843e+00 -4.20251995e-01 6.02046430e-01
1.31912291e+00 -5.15082181e-01 1.19232810e+00 2.82601923e-01
-9.09899116e-01 8.71396184e-01 4.59745042e-02 4.86192286e-01
1.01016057e+00 -1.28905678e+00 7.42820501e-01 1.34423018e-01
-2.42119625e-01 2.33975366e-01 4.23758328e-01 -3.14884722e-01
-1.17816210e+00 -8.90148997e-01 8.69826257e-01 -3.71036261e-01
7.92986751e-01 -9.57425773e-01 -1.29321873e+00 4.69384551e-01
3.37043434e-01 -4.65142131e-01 7.74084508e-01 5.21060109e-01
-2.23636582e-01 8.81110355e-02 -9.76110101e-01 7.35674202e-01
6.82761729e-01 -6.61407769e-01 -6.04465127e-01 3.57796252e-01
8.62821996e-01 -4.90174830e-01 -1.04566300e+00 4.32543010e-01
3.13423157e-01 -7.81551719e-01 6.91917181e-01 -8.20677519e-01
3.13889861e-01 5.38099229e-01 -3.34860682e-02 -1.62030339e+00
4.19483632e-01 -1.01799548e+00 -1.32579625e-01 1.77917516e+00
5.30257225e-01 -4.68641073e-01 8.27578306e-01 7.58980989e-01
-2.91883856e-01 -6.54412568e-01 -8.09314609e-01 -4.69367653e-01
1.55545563e-01 -2.32450187e-01 6.68875754e-01 1.19985580e+00
3.96367580e-01 8.46668303e-01 -7.30964243e-01 -3.45907778e-01
1.37077495e-01 -3.04523706e-01 1.30764925e+00 -1.08756196e+00
-1.90408528e-01 -4.90733087e-01 2.17197970e-01 -1.59095085e+00
4.36570406e-01 -3.94512832e-01 4.63388354e-01 -1.22769380e+00
4.79287803e-02 -7.16204643e-01 4.30106282e-01 4.65847224e-01
-4.31048095e-01 -3.25886667e-01 1.24322377e-01 1.06119312e-01
-4.84262437e-01 1.21291280e+00 1.36188376e+00 -1.89914972e-01
-5.79553962e-01 3.39359403e-01 -4.64209527e-01 7.37720311e-01
7.24868119e-01 -2.21677572e-02 -5.51715314e-01 -2.68568695e-01
-4.94496405e-01 5.35718977e-01 4.35639210e-02 -4.55001205e-01
9.97456834e-02 7.27121830e-02 -4.30199862e-01 -4.52870131e-01
7.24745154e-01 -5.54749310e-01 -7.35245407e-01 1.23548470e-01
-8.95620465e-01 -2.99557328e-01 3.39742064e-01 3.15130323e-01
-4.62634206e-01 -4.98855710e-02 6.40981555e-01 5.28070256e-02
-4.72408950e-01 1.41777396e-01 -6.77754521e-01 5.16596317e-01
4.71456051e-01 2.61499465e-01 -6.68168843e-01 -8.90468776e-01
-7.79657125e-01 6.63665056e-01 6.14359090e-03 7.43618608e-01
3.50088596e-01 -1.10102761e+00 -1.07524347e+00 3.17813247e-01
2.97372341e-01 1.89158961e-01 3.41207355e-01 8.23103130e-01
5.11086360e-02 5.25556266e-01 -1.16797872e-02 -6.82594836e-01
-1.56454408e+00 8.77811748e-04 3.80171716e-01 -6.61849260e-01
-4.41015184e-01 9.44128692e-01 3.35445166e-01 -1.33956563e+00
7.86901951e-01 -1.19390860e-01 -2.19300732e-01 1.26003966e-01
4.32351530e-01 2.71368682e-01 1.87831879e-01 -5.52511156e-01
-4.57045296e-03 -3.13535422e-01 -4.53283638e-01 -1.03646211e-01
1.19151592e+00 -3.18048626e-01 -1.57808121e-02 6.72554493e-01
1.06852865e+00 -2.15835035e-01 -1.52461386e+00 -7.38923132e-01
1.95310414e-01 -1.71497270e-01 -2.30259567e-01 -1.07786572e+00
-6.38924479e-01 1.09769988e+00 6.25170022e-02 5.70577025e-01
7.82353580e-01 -1.39265746e-01 1.10220325e+00 9.25409436e-01
6.40996248e-02 -1.06287467e+00 5.21002889e-01 1.25103176e+00
1.26435149e+00 -1.55025220e+00 -6.62485242e-01 -3.47636670e-01
-1.37028611e+00 8.89911294e-01 9.26456749e-01 5.52392721e-01
3.22206169e-01 4.38500196e-01 4.02902752e-01 1.30347980e-04
-1.10074425e+00 -1.98122505e-02 3.39993268e-01 6.04711473e-01
6.06944442e-01 -1.27978802e-01 1.98492646e-01 8.01730275e-01
-3.08300376e-01 -4.28043962e-01 3.31693202e-01 6.47892058e-01
-2.72278965e-01 -1.15012050e+00 -2.04247430e-01 3.45901400e-01
-1.23296358e-01 -2.08783999e-01 -9.41384375e-01 7.16270626e-01
-6.16773963e-01 1.43718243e+00 7.82614946e-02 -4.71380532e-01
6.75499916e-01 4.80832905e-01 3.83124501e-02 -8.12901139e-01
-9.97694671e-01 1.19784132e-01 9.72839653e-01 -1.56160101e-01
-6.02362268e-02 -4.72760975e-01 -1.04191256e+00 -4.85025316e-01
-2.88678795e-01 4.99473214e-01 5.58639765e-01 1.00636208e+00
2.97379553e-01 5.55912673e-01 9.61892486e-01 -7.84586489e-01
-9.48733211e-01 -1.62501240e+00 -4.59034085e-01 4.67628628e-01
5.62266946e-01 -3.84230554e-01 -3.53133082e-01 -1.95754021e-01] | [12.799704551696777, 7.928284645080566] |
18a6ce31-051d-413b-836d-36f8d197c670 | separating-facts-from-fiction-linguistic | null | null | https://aclanthology.org/P17-2102 | https://aclanthology.org/P17-2102.pdf | Separating Facts from Fiction: Linguistic Models to Classify Suspicious and Trusted News Posts on Twitter | Pew research polls report 62 percent of U.S. adults get news on social media (Gottfried and Shearer, 2016). In a December poll, 64 percent of U.S. adults said that {``}made-up news{''} has caused a {``}great deal of confusion{''} about the facts of current events (Barthel et al., 2016). Fabricated stories in social media, ranging from deliberate propaganda to hoaxes and satire, contributes to this confusion in addition to having serious effects on global stability. In this work we build predictive models to classify 130 thousand news posts as suspicious or verified, and predict four sub-types of suspicious news {--} satire, hoaxes, clickbait and propaganda. We show that neural network models trained on tweet content and social network interactions outperform lexical models. Unlike previous work on deception detection, we find that adding syntax and grammar features to our models does not improve performance. Incorporating linguistic features improves classification results, however, social interaction features are most informative for finer-grained separation between four types of suspicious news posts. | ['Jin Yea Jang', 'Kyle Shaffer', 'Svitlana Volkova', 'Nathan Hodas'] | 2017-07-01 | null | null | null | acl-2017-7 | ['deception-detection'] | ['miscellaneous'] | [-1.18715845e-01 2.95327276e-01 -6.32112503e-01 -4.99242634e-01
-4.36860383e-01 -5.13977706e-01 1.10445154e+00 9.45780873e-01
-6.08835578e-01 9.08246279e-01 1.08572054e+00 -5.35082400e-01
8.50428194e-02 -9.33220088e-01 -4.46484953e-01 -2.35033631e-02
3.50967079e-01 5.17981276e-02 -2.85585485e-02 -5.78527093e-01
7.60103106e-01 -2.51103759e-01 -1.28778195e+00 3.31540346e-01
8.57410967e-01 6.32656813e-01 -7.06372380e-01 5.31444132e-01
-1.45593956e-01 1.45047438e+00 -1.08516765e+00 -1.09864044e+00
-4.93842483e-01 -4.51347679e-01 -8.26461852e-01 -3.48464102e-01
8.50219786e-01 -4.39182311e-01 -4.69390541e-01 1.14841020e+00
4.05312687e-01 -1.97635666e-02 7.56529868e-01 -7.74695277e-01
-9.35951173e-01 1.29788589e+00 -6.67918921e-01 7.65798092e-01
7.40244150e-01 -2.38017336e-01 1.00513947e+00 -3.77352506e-01
7.58188069e-01 1.39136374e+00 1.04765022e+00 3.62157017e-01
-1.06513011e+00 -1.10156333e+00 2.09386587e-01 -3.06427702e-02
-8.74142468e-01 -5.13791680e-01 7.51161635e-01 -7.67495275e-01
7.84156919e-01 5.04940093e-01 9.01665866e-01 1.87391174e+00
3.80072057e-01 7.14977920e-01 1.17957115e+00 -1.33233937e-02
-4.39768583e-02 1.23736672e-02 9.20178175e-01 4.68618602e-01
7.74467766e-01 -2.55397469e-01 -9.03822780e-01 -7.13061571e-01
8.91655013e-02 -1.72190350e-02 1.15294196e-01 7.30227530e-01
-8.19470406e-01 1.50479770e+00 3.56710404e-01 4.41725731e-01
-2.13517115e-01 -5.88785633e-02 6.50474966e-01 4.15106118e-01
1.15111458e+00 4.73519176e-01 -2.77086236e-02 -5.65820873e-01
-8.62146854e-01 4.09964204e-01 1.20004976e+00 7.09521174e-02
2.15246603e-01 -2.87054479e-02 1.71105444e-01 1.00456655e+00
1.39660567e-01 4.68142241e-01 7.70231068e-01 -6.18643224e-01
7.30877757e-01 5.83088100e-01 1.10939667e-01 -1.90503955e+00
-9.15693343e-01 -7.88796186e-01 -6.76771462e-01 -3.37929696e-01
6.87472165e-01 -4.75512862e-01 -4.67200607e-01 1.56417227e+00
1.06399171e-01 -1.26356512e-01 -2.84201026e-01 5.55205286e-01
1.15923333e+00 5.24928153e-01 1.36143968e-01 -2.52829403e-01
1.34530175e+00 -2.54289031e-01 -8.64256382e-01 -6.47890925e-01
6.21335030e-01 -8.29012156e-01 5.65829694e-01 4.45823699e-01
-1.25417459e+00 2.62430571e-02 -1.05177426e+00 -1.72762141e-01
-5.36614954e-01 -3.83507758e-01 7.36544013e-01 1.07314968e+00
-6.05755866e-01 7.46915579e-01 -4.90716010e-01 -5.61783552e-01
5.37501156e-01 -1.70365244e-01 -9.31215435e-02 2.94420600e-01
-1.46669495e+00 9.98468280e-01 1.70487687e-01 -3.07297766e-01
-1.37620986e-01 -3.95346045e-01 -8.37772846e-01 -2.63385743e-01
2.67551452e-01 -1.31426066e-01 1.18269813e+00 -1.19312131e+00
-1.06206834e+00 1.04682994e+00 4.71444018e-02 -5.47201097e-01
4.08644021e-01 -3.23576301e-01 -7.65928686e-01 -1.81324959e-01
4.01040316e-01 -3.28568295e-02 8.17534506e-01 -4.62606519e-01
-4.30042535e-01 -4.46313679e-01 1.25099719e-01 -1.04949452e-01
-6.08503282e-01 7.39368379e-01 6.60539687e-01 -8.72405708e-01
2.30433926e-01 -5.64123392e-01 4.47218597e-01 -7.19359100e-01
-6.83909535e-01 -4.40319151e-01 4.04194772e-01 -9.96827304e-01
1.66185713e+00 -1.78285325e+00 -4.53088105e-01 5.17454669e-02
6.97078347e-01 1.81429863e-01 4.46290523e-01 7.10850239e-01
1.80375859e-01 7.72421181e-01 4.87321585e-01 -1.80208415e-01
5.29391170e-02 -1.19169876e-01 -2.21921787e-01 7.74785161e-01
-3.20336938e-01 6.57184780e-01 -1.08000302e+00 -5.10972850e-02
-3.02064687e-01 3.49455208e-01 -4.29749876e-01 -7.12109923e-01
-1.89952850e-02 8.46093670e-02 -4.38244849e-01 5.80380797e-01
2.72940546e-01 -4.84976619e-01 3.97347584e-02 4.56942469e-01
-3.01486373e-01 1.26845992e+00 -2.78553873e-01 6.50389194e-01
-3.77798714e-02 1.16182458e+00 1.61650792e-01 -6.79761946e-01
7.99916208e-01 -6.50728308e-03 -1.63514540e-01 -5.72513461e-01
8.67207885e-01 5.35603315e-02 2.16859668e-01 -3.78192991e-01
6.90088511e-01 -1.74217060e-01 -5.26005805e-01 9.12777364e-01
-5.03089428e-01 2.28618443e-01 4.74693766e-03 4.88741368e-01
1.10222626e+00 -6.05955243e-01 4.15503860e-01 -4.38374013e-01
1.21138170e-01 -2.94306725e-02 4.65605855e-01 1.04397881e+00
-2.95233846e-01 2.39466369e-01 7.80045152e-01 -5.76412618e-01
-6.72233403e-01 -6.87873602e-01 -1.08408645e-01 1.62723446e+00
-1.77198797e-01 -6.38594747e-01 -6.29858315e-01 -6.03535175e-01
-1.31208509e-01 1.15381336e+00 -7.35821724e-01 -2.46760964e-01
-3.15682828e-01 -1.05550182e+00 7.90925980e-01 -1.32942358e-02
7.84314573e-01 -6.70176208e-01 -5.08714139e-01 5.35628833e-02
-7.31372297e-01 -6.53686285e-01 -4.05434549e-01 -2.41290599e-01
-4.62929249e-01 -9.22457278e-01 -2.64431834e-01 -3.98084581e-01
2.24358201e-01 2.00960442e-01 9.55750167e-01 2.85538405e-01
-1.22599356e-01 6.14370070e-02 -5.81890583e-01 -5.65072060e-01
-6.61553741e-01 5.12875505e-02 1.67938799e-01 -6.99732974e-02
5.48548520e-01 -6.82060003e-01 -3.73555869e-01 2.60188859e-02
-5.80513179e-01 -1.46895796e-01 1.00181557e-01 5.21438003e-01
-6.59251273e-01 -1.00162275e-01 5.83820283e-01 -1.23313653e+00
9.99040067e-01 -1.04887879e+00 4.18376401e-02 -3.83078426e-01
-4.38209623e-01 -3.58279079e-01 3.23055506e-01 -8.11639190e-01
-7.46323824e-01 -8.74309957e-01 -2.02245697e-01 6.68500841e-01
-7.24399015e-02 8.09212983e-01 6.48969948e-01 2.85180777e-01
1.29763508e+00 -2.52044816e-02 -3.18604261e-02 -3.56916875e-01
-1.84925079e-01 9.07140553e-01 2.13081166e-01 -2.40298864e-02
5.78898072e-01 3.59257221e-01 -6.39841914e-01 -1.08373082e+00
-1.52930892e+00 -4.63937670e-01 -6.18849620e-02 -2.35686123e-01
9.17012334e-01 -8.27486455e-01 -9.94532824e-01 7.91086435e-01
-1.03452492e+00 -1.25315443e-01 5.04294872e-01 4.64922518e-01
7.17313215e-02 3.43631446e-01 -1.15765512e+00 -9.49981987e-01
-1.88476756e-01 -3.22213441e-01 2.84430325e-01 1.50872007e-01
-1.00955129e+00 -1.05403757e+00 1.84676554e-02 8.79875243e-01
4.71798003e-01 6.33520305e-01 5.88910222e-01 -1.45809090e+00
3.75712693e-01 -5.29332399e-01 -1.94781691e-01 5.86177967e-02
4.25885059e-02 -9.88027006e-02 -6.61741138e-01 7.51213878e-02
1.98475152e-01 -7.97585666e-01 7.71392107e-01 2.21781641e-01
4.73931670e-01 -1.16059279e+00 -2.90071398e-01 -1.95727050e-01
7.43919075e-01 -1.73947096e-01 4.82429266e-01 4.23413157e-01
3.60772699e-01 5.18855393e-01 5.31960614e-02 6.13348663e-01
6.35355592e-01 2.25460812e-01 2.69599438e-01 6.12238586e-01
1.53035179e-01 -6.39620841e-01 8.31366777e-01 3.98898870e-01
-5.96309565e-02 -3.53565812e-01 -1.01889610e+00 2.85143346e-01
-1.60198271e+00 -1.54011381e+00 -4.32312489e-01 1.82976305e+00
8.20986688e-01 5.64577103e-01 3.61482382e-01 3.90713394e-01
1.01406360e+00 7.49969184e-01 -1.52427003e-01 -8.13328385e-01
-7.49168396e-02 -1.35562569e-01 4.63598609e-01 9.04757977e-01
-1.08241415e+00 6.99599266e-01 6.04043865e+00 7.72851110e-01
-1.14308393e+00 2.84175962e-01 8.09664130e-01 -2.99854547e-01
-2.68791288e-01 -6.13482475e-01 -1.05682802e+00 6.65617228e-01
1.12349522e+00 -1.63027495e-01 6.08774014e-02 7.20161974e-01
2.49659479e-01 -3.44995767e-01 -5.91189146e-01 7.84507215e-01
7.77742863e-01 -1.41505992e+00 -3.10150564e-01 1.22421890e-01
7.29291499e-01 2.02956930e-01 1.27607226e-01 2.73383230e-01
8.35113168e-01 -8.85759890e-01 1.09624910e+00 1.89879924e-01
1.86197355e-01 -5.34593523e-01 6.09980345e-01 8.66635382e-01
-2.17568837e-02 -1.85354754e-01 -6.07323162e-02 -1.01173353e+00
1.42346442e-01 8.55999053e-01 -7.26908922e-01 -1.90120459e-01
6.11357331e-01 7.82018483e-01 -8.04449141e-01 4.65538323e-01
-3.72977406e-01 1.09962738e+00 -4.00354117e-01 -9.15150106e-01
3.19563836e-01 2.36939654e-01 7.43987560e-01 1.13409495e+00
8.68210662e-03 3.38481188e-01 -1.89549863e-01 5.22223949e-01
-1.23679258e-01 7.49610886e-02 -5.22950709e-01 -5.77432215e-01
2.52401590e-01 1.01551855e+00 -8.17942023e-01 -4.30648297e-01
-3.76631260e-01 7.00628519e-01 4.54594105e-01 8.75084773e-02
-7.83440411e-01 -2.14387506e-01 4.16056454e-01 7.79106200e-01
-3.33199710e-01 -1.80500999e-01 -5.08725345e-01 -1.28855598e+00
-4.30386871e-01 -7.26165473e-01 4.86551374e-01 -4.75712001e-01
-1.58981466e+00 5.84608465e-02 -1.80156603e-01 -2.64719486e-01
-2.63330966e-01 -3.33833247e-01 -8.21097970e-01 3.33442330e-01
-7.60887444e-01 -8.75887692e-01 6.82442915e-03 1.78444609e-01
4.09746349e-01 2.06581950e-01 2.93750852e-01 1.08791701e-01
-4.33549732e-01 5.30862033e-01 -2.44256496e-01 5.29261470e-01
5.41955471e-01 -8.89825821e-01 5.33813834e-01 5.37927806e-01
-1.73181847e-01 7.23094285e-01 1.03060269e+00 -1.16887414e+00
-6.42191172e-01 -7.33630419e-01 1.79865718e+00 -7.05133855e-01
1.51753199e+00 -4.20713753e-01 -5.02679169e-01 9.14730847e-01
2.86219120e-01 -8.58858168e-01 9.82635558e-01 6.49898171e-01
-7.33422637e-01 5.43066740e-01 -1.12433136e+00 6.12759411e-01
1.14785373e+00 -4.20921803e-01 -9.50587571e-01 8.28645051e-01
3.68821561e-01 -2.75155623e-03 -4.40660626e-01 -1.19907051e-01
7.39997268e-01 -1.37168491e+00 7.10368216e-01 -8.31028581e-01
9.65182602e-01 7.52305150e-01 3.29014450e-01 -1.23056328e+00
-4.55251724e-01 -7.34680176e-01 2.62347370e-01 1.13306355e+00
7.17758656e-01 -1.04499280e+00 7.57015765e-01 7.29004800e-01
-9.08299088e-02 -2.63509452e-01 -8.17868292e-01 -3.92292887e-01
3.22873980e-01 -3.42638701e-01 -1.32572558e-02 1.49940860e+00
6.81265414e-01 6.24820411e-01 -4.87118095e-01 -1.18378595e-01
4.66505319e-01 -4.29732323e-01 4.79119599e-01 -1.40554416e+00
-1.88194737e-01 -8.56444657e-01 -3.47728610e-01 -7.26217449e-01
1.20085865e-01 -7.45807171e-01 -5.11310756e-01 -1.44361973e+00
4.23196137e-01 -8.02885965e-02 3.41124058e-01 3.59869659e-01
1.50599763e-01 6.44409537e-01 -1.13715604e-01 -5.93277700e-02
-5.54935336e-01 -1.08270779e-01 8.21865201e-01 -1.92845017e-01
-2.88340151e-01 2.20211729e-01 -1.09582818e+00 1.11295426e+00
1.16395557e+00 -2.10491419e-01 3.28613166e-03 -1.97637722e-01
1.24171174e+00 -1.10268518e-01 5.58769643e-01 -6.69337630e-01
9.93196201e-03 -2.28317976e-01 4.18813735e-01 -3.54927272e-01
4.93291199e-01 -1.87425055e-02 -1.95659310e-01 7.40071356e-01
-4.77788001e-01 -3.52558047e-02 -8.90622661e-02 5.08594811e-01
-1.00941006e-02 -2.91859627e-01 6.24498487e-01 -9.18661729e-02
2.77909338e-01 -2.59669334e-01 -1.03712332e+00 3.49508673e-01
5.86040318e-01 -4.15547714e-02 -1.14162600e+00 -1.18633628e+00
-1.04978764e+00 -9.83813405e-02 1.62762403e-01 5.09301543e-01
1.63066700e-01 -9.89275932e-01 -8.95128191e-01 -1.32334754e-01
-2.17606530e-01 -7.39517808e-01 -3.07512879e-02 1.12784660e+00
-4.29503828e-01 2.30932787e-01 1.98041961e-01 2.58709520e-01
-1.36879289e+00 1.79528072e-01 -1.54318258e-01 4.17940617e-02
-3.39223534e-01 1.15773177e+00 -3.02082807e-01 6.73289318e-03
1.88534737e-01 -2.15699539e-01 -3.34509850e-01 7.80710757e-01
9.14004624e-01 8.52871954e-01 -2.54499108e-01 -9.11644876e-01
-1.39422402e-01 -1.14678323e-01 -3.99507701e-01 -5.73219359e-02
1.16677928e+00 -1.43079862e-01 -3.10235500e-01 8.68186057e-01
1.14928842e+00 4.91151989e-01 -2.75592059e-01 -6.41636923e-02
4.97495644e-02 -3.83232653e-01 -2.65444443e-02 -9.82150495e-01
-4.01924968e-01 4.30764675e-01 -5.31672180e-01 1.10751462e+00
1.81764558e-01 3.66601467e-01 8.50672543e-01 2.81111240e-01
2.44052839e-02 -1.25521791e+00 3.38043213e-01 6.12025201e-01
8.76438379e-01 -1.13964307e+00 2.99636781e-01 -5.32929838e-01
-5.81137121e-01 7.82181561e-01 2.44454473e-01 -2.48459950e-01
7.22025931e-01 -7.10287020e-02 -1.81869492e-01 -4.35924351e-01
-4.92427707e-01 2.41878793e-01 7.73744360e-02 9.69490930e-02
5.82720816e-01 1.51773900e-01 -1.16680932e+00 8.28452706e-01
-9.41065609e-01 -5.31195223e-01 8.64678502e-01 6.10317767e-01
-9.99450803e-01 -5.64654350e-01 -5.29468000e-01 1.01302302e+00
-1.23177409e+00 -1.54639378e-01 -1.17584491e+00 6.98017180e-01
1.77169487e-01 1.45953929e+00 1.72426432e-01 -5.28916180e-01
-3.40922177e-01 2.73798138e-01 8.30279961e-02 -6.30007148e-01
-1.04981303e+00 -2.23308995e-01 1.06973886e+00 -2.39821628e-01
-2.67864853e-01 -1.10947990e+00 -7.88296461e-01 -1.01864076e+00
-3.39322686e-01 1.48456648e-01 5.94295025e-01 9.67384994e-01
8.33360925e-02 -1.35047078e-01 1.82981059e-01 -3.39692742e-01
-4.70627964e-01 -1.29471982e+00 -4.92014468e-01 3.39594483e-01
4.88093495e-01 -4.73890156e-01 -8.94174278e-01 -3.86999726e-01] | [8.581302642822266, 10.273999214172363] |
e24b3788-8e53-4a47-80e3-0ab67bd12821 | time-will-tell-new-outlooks-and-a-baseline | 2210.02443 | null | https://arxiv.org/abs/2210.02443v1 | https://arxiv.org/pdf/2210.02443v1.pdf | Time Will Tell: New Outlooks and A Baseline for Temporal Multi-View 3D Object Detection | While recent camera-only 3D detection methods leverage multiple timesteps, the limited history they use significantly hampers the extent to which temporal fusion can improve object perception. Observing that existing works' fusion of multi-frame images are instances of temporal stereo matching, we find that performance is hindered by the interplay between 1) the low granularity of matching resolution and 2) the sub-optimal multi-view setup produced by limited history usage. Our theoretical and empirical analysis demonstrates that the optimal temporal difference between views varies significantly for different pixels and depths, making it necessary to fuse many timesteps over long-term history. Building on our investigation, we propose to generate a cost volume from a long history of image observations, compensating for the coarse but efficient matching resolution with a more optimal multi-view matching setup. Further, we augment the per-frame monocular depth predictions used for long-term, coarse matching with short-term, fine-grained matching and find that long and short term temporal fusion are highly complementary. While maintaining high efficiency, our framework sets new state-of-the-art on nuScenes, achieving first place on the test set and outperforming previous best art by 5.2% mAP and 3.7% NDS on the validation set. Code will be released $\href{https://github.com/Divadi/SOLOFusion}{here.}$ | ['Wei Zhan', 'Masayoshi Tomizuka', 'Kris Kitani', 'Kurt Keutzer', 'Shijia Yang', 'Chenfeng Xu', 'Jinhyung Park'] | 2022-10-05 | null | null | null | null | ['stereo-matching-1'] | ['computer-vision'] | [ 2.83342540e-01 -3.70116383e-01 -2.52987623e-01 -1.61630794e-01
-1.02718031e+00 -8.22701395e-01 8.07914019e-01 -1.25064790e-01
-4.02427256e-01 5.03170252e-01 2.91877627e-01 7.37349689e-02
-9.40066427e-02 -8.31668019e-01 -7.87883639e-01 -6.69797242e-01
4.31916793e-04 4.49403748e-02 8.17941487e-01 -1.18629508e-01
3.50273937e-01 3.71814847e-01 -2.15462208e+00 5.04081845e-01
5.82592666e-01 1.16443741e+00 5.04345238e-01 8.25025260e-01
3.86019140e-01 6.12562478e-01 9.44579020e-03 -2.62333035e-01
7.54720211e-01 -1.59815639e-01 -5.71735561e-01 2.31756523e-01
1.13199341e+00 -8.20982397e-01 -4.78418827e-01 9.12596166e-01
6.40217543e-01 2.99022555e-01 2.88260490e-01 -9.57653284e-01
-3.27348888e-01 -6.70459419e-02 -7.15549409e-01 5.44921756e-01
4.79786932e-01 6.51140451e-01 1.07952166e+00 -8.72185290e-01
9.46616709e-01 9.51092184e-01 7.76997805e-01 1.37503207e-01
-1.17473769e+00 -6.97088838e-01 1.26014650e-01 1.03147730e-01
-1.21678734e+00 -7.84448922e-01 5.17146885e-01 -6.55935585e-01
1.25403881e+00 1.98023483e-01 7.34269977e-01 8.53178263e-01
3.47614914e-01 3.97728860e-01 1.47733521e+00 -1.72522038e-01
-3.17820460e-02 -2.54096508e-01 -1.27235249e-01 5.52196085e-01
1.30655035e-01 9.17153478e-01 -7.84092784e-01 -5.25339246e-02
8.87343287e-01 4.53294702e-02 -2.63269424e-01 -3.19725811e-01
-1.26977515e+00 6.26342416e-01 5.40323377e-01 1.13210097e-01
-4.84010160e-01 1.34591416e-01 8.67703632e-02 3.62414122e-01
6.19523764e-01 3.54299486e-01 -3.79765928e-01 -9.09725577e-02
-1.03245974e+00 4.34794486e-01 3.38918418e-01 7.51930475e-01
1.03460824e+00 -2.11955473e-01 -6.11069389e-02 5.91864884e-01
-1.22810043e-01 7.16168165e-01 8.30649734e-02 -1.53745735e+00
2.86769599e-01 3.74396384e-01 2.18131453e-01 -7.15609252e-01
-2.27164686e-01 -4.30416673e-01 -4.26032603e-01 5.47227025e-01
5.54540336e-01 -6.25244458e-04 -7.79003978e-01 1.72872293e+00
5.23010552e-01 2.12691963e-01 -1.04733296e-01 9.55077291e-01
5.26377201e-01 5.00178933e-01 -1.98794186e-01 -2.14890599e-01
1.25208759e+00 -8.24374914e-01 -1.87643796e-01 -4.82000768e-01
5.05788684e-01 -9.98093903e-01 6.56525493e-01 4.02721345e-01
-1.28739953e+00 -6.71685457e-01 -9.55345273e-01 -1.33811831e-01
-2.58883804e-01 -2.39342958e-01 7.05383420e-01 3.96817833e-01
-1.19479859e+00 6.86062396e-01 -6.61897659e-01 -5.54303050e-01
2.61672705e-01 2.51618356e-01 -4.24167842e-01 -4.24554229e-01
-8.55870783e-01 8.75361443e-01 3.73919457e-01 -3.37811053e-01
-8.29381526e-01 -8.96579444e-01 -6.26632810e-01 -3.77351761e-01
6.38583004e-01 -1.04156578e+00 1.24094272e+00 -7.30468750e-01
-9.30213928e-01 1.05233812e+00 -3.16611469e-01 -5.89782000e-01
6.94229722e-01 -2.58507371e-01 -1.18423991e-01 2.94018477e-01
3.24733257e-01 1.05479038e+00 5.67743182e-01 -1.27359831e+00
-1.05403352e+00 -5.85679591e-01 3.70342255e-01 5.51892042e-01
5.47141880e-02 -1.86130226e-01 -6.83787823e-01 -3.59400392e-01
2.35088557e-01 -1.06214118e+00 -2.01831520e-01 -1.17010241e-02
1.42856121e-01 1.23300105e-01 6.72680497e-01 -6.04085982e-01
9.85023797e-01 -2.12971544e+00 1.13838464e-01 -3.32387418e-01
2.05908775e-01 -1.37381732e-01 -7.95577168e-02 3.09395581e-01
1.22183934e-01 -1.89704850e-01 8.77473429e-02 -3.90572190e-01
-3.59689921e-01 4.51431796e-02 -1.00097038e-01 6.32303774e-01
-2.18688950e-01 7.68653333e-01 -9.39833105e-01 -3.64829808e-01
8.06742251e-01 3.23732764e-01 -7.02497125e-01 5.25936410e-02
-2.21299425e-01 5.03528893e-01 -2.06043795e-01 8.15519750e-01
8.86166215e-01 -3.62405360e-01 -1.76347375e-01 -4.72204328e-01
-6.75269008e-01 1.58820182e-01 -1.37665665e+00 1.95884109e+00
-3.13654453e-01 6.67891264e-01 1.77786276e-01 -3.25401187e-01
4.47578102e-01 1.49711668e-02 7.91869283e-01 -9.49782252e-01
1.43185616e-01 3.37219596e-01 -1.74380392e-01 -3.49315107e-01
7.65849233e-01 -1.43016845e-01 2.10790098e-01 2.04264656e-01
-5.98466247e-02 -4.81130928e-01 2.58704245e-01 1.39304683e-01
1.12819815e+00 1.93215296e-01 2.02616036e-01 -1.23541981e-01
1.66424826e-01 2.20938459e-01 4.41757262e-01 7.81351805e-01
-2.95816749e-01 9.18369710e-01 -1.22329846e-01 -3.04411471e-01
-1.25809121e+00 -1.00759482e+00 -1.99223146e-01 9.91436422e-01
4.37889844e-01 -3.73907447e-01 -2.89214522e-01 -2.23055094e-01
1.61401033e-01 4.57906514e-01 -6.13169491e-01 1.36905670e-01
-3.82670283e-01 -5.65470934e-01 2.66253740e-01 4.82135653e-01
7.72998750e-01 -5.66316605e-01 -1.15129042e+00 1.92721739e-01
-3.36510241e-01 -1.44738567e+00 -3.09997827e-01 5.42962477e-02
-9.29185569e-01 -9.67738032e-01 -6.44628286e-01 -9.77772400e-02
1.48058515e-02 1.03163385e+00 1.31893396e+00 -5.41216061e-02
-1.38243705e-01 6.47679210e-01 -3.92745376e-01 -2.42771417e-01
-1.54914027e-02 -1.46502942e-01 1.67102832e-02 -4.20193493e-01
1.19741574e-01 -8.01424742e-01 -9.88007545e-01 4.43813831e-01
-9.04990733e-01 4.85980093e-01 4.97951776e-01 6.86001241e-01
6.36748374e-01 -8.47179666e-02 1.73395947e-02 -4.40412939e-01
-2.53015041e-01 -4.04311091e-01 -8.82322490e-01 -1.25824250e-02
-6.62952542e-01 -7.15823174e-02 -6.57060221e-02 -2.61168063e-01
-1.14840829e+00 3.23190540e-02 -2.66042799e-02 -6.46839201e-01
-1.48888990e-01 5.31361885e-02 2.93702930e-01 -2.83541411e-01
7.90039420e-01 6.13771416e-02 -9.59351212e-02 -2.80864239e-01
2.85185188e-01 3.00705284e-01 6.27981842e-01 -5.03763974e-01
6.34217024e-01 1.12269723e+00 -5.77343255e-02 -6.13746524e-01
-1.01321650e+00 -7.19203591e-01 -6.58867121e-01 -4.84304219e-01
1.08467054e+00 -1.32491767e+00 -4.83378768e-01 4.58695203e-01
-1.04888785e+00 -4.47259694e-01 -3.45965683e-01 5.17721415e-01
-6.02340043e-01 3.63647521e-01 -4.39215243e-01 -7.72238016e-01
-1.13403648e-01 -1.13011479e+00 1.33272469e+00 8.51893723e-02
-1.76319946e-02 -7.38931239e-01 2.26912603e-01 6.85885310e-01
3.15486401e-01 4.35901195e-01 2.43535176e-01 7.69436806e-02
-1.19780278e+00 7.78083727e-02 -4.77799952e-01 2.51932383e-01
-9.04262438e-02 -1.96490586e-01 -1.34894276e+00 -4.16956693e-01
-2.97960732e-03 -2.54285425e-01 1.03719497e+00 6.44049227e-01
5.22534251e-01 3.49913597e-01 -2.70700991e-01 8.48767936e-01
1.80768526e+00 1.07499920e-01 5.67625463e-01 6.03816748e-01
7.04887152e-01 7.12134242e-01 9.99608338e-01 5.12755394e-01
3.96357864e-01 8.68062496e-01 6.53553128e-01 -8.44760612e-03
-4.81279582e-01 -8.67861882e-03 3.47010195e-01 3.92726779e-01
-4.36801940e-01 -1.00072451e-01 -9.54787970e-01 6.86239958e-01
-1.80887306e+00 -1.15528452e+00 -1.38237894e-01 2.43141580e+00
3.69996369e-01 2.72887945e-01 1.21804133e-01 -1.32937163e-01
3.59023482e-01 4.13151801e-01 -5.74269950e-01 1.32634878e-01
-4.75177169e-01 -6.26949295e-02 1.11715865e+00 6.48213804e-01
-9.91528392e-01 8.15245509e-01 5.89043951e+00 7.54201472e-01
-1.07666719e+00 3.55417073e-01 4.77227628e-01 -5.79092860e-01
-2.66600639e-01 2.92359233e-01 -9.76631165e-01 3.69319767e-01
7.37338185e-01 -2.09028646e-02 3.82603377e-01 4.50855672e-01
2.19603464e-01 -7.11970329e-01 -1.20151556e+00 1.13158751e+00
-8.44557285e-02 -1.58928442e+00 -3.03223759e-01 4.48020726e-01
1.11849558e+00 6.71766162e-01 3.95366214e-02 -8.10791627e-02
2.39725962e-01 -7.45714486e-01 1.00924766e+00 3.07627499e-01
7.69350469e-01 -4.34054792e-01 4.86255974e-01 3.83054346e-01
-1.52185774e+00 -2.30298355e-01 -7.53142983e-02 -2.56863445e-01
2.40594447e-01 6.42562211e-01 -3.63146842e-01 8.17629933e-01
9.92204249e-01 8.86383057e-01 -7.77565956e-01 8.75800431e-01
3.09160680e-01 -2.02032533e-02 -5.74166715e-01 5.88476360e-01
2.80355603e-01 -5.94359636e-02 7.30593562e-01 1.08847427e+00
4.87370610e-01 4.16508615e-01 1.70809492e-01 5.05178690e-01
2.69287258e-01 -4.71755356e-01 -9.43607569e-01 4.48962092e-01
5.28365552e-01 8.71703327e-01 -5.91000557e-01 -3.97101671e-01
-6.76886499e-01 9.12895203e-01 1.32686198e-01 1.88670918e-01
-7.68894851e-01 3.26579273e-01 7.06140280e-01 4.15988386e-01
4.93928164e-01 -3.77539903e-01 -4.61519569e-01 -1.24215829e+00
1.72058001e-01 -6.98343515e-01 5.43998480e-01 -1.04431915e+00
-1.08709586e+00 3.65422517e-01 2.85405010e-01 -1.40685272e+00
-3.83829117e-01 -3.81220758e-01 -1.92050591e-01 8.33063245e-01
-1.61734807e+00 -1.21224189e+00 -5.61161458e-01 5.46270013e-01
7.38150299e-01 2.54195660e-01 3.75917166e-01 3.23845118e-01
-1.51297087e-02 3.14858407e-02 -6.39119670e-02 -5.04585922e-01
8.15328896e-01 -9.58828211e-01 3.38925362e-01 1.09894800e+00
-5.43373972e-02 1.14161462e-01 8.84152532e-01 -6.67176485e-01
-1.45787692e+00 -7.08857954e-01 6.90602362e-01 -8.45535159e-01
4.55027372e-01 -1.40302598e-01 -7.40985334e-01 5.48204839e-01
2.50870824e-01 1.39949853e-02 1.24206938e-01 5.96391317e-03
-3.33871275e-01 -7.16935694e-02 -1.13495588e+00 5.23562908e-01
1.40367699e+00 -6.99936748e-01 -2.89675027e-01 1.61969304e-01
6.84011996e-01 -6.35760546e-01 -1.06650984e+00 7.87851810e-01
8.28091085e-01 -1.81358278e+00 1.25011301e+00 2.03908816e-01
5.58056235e-01 -3.22320193e-01 -7.48740315e-01 -7.69410670e-01
-3.86452526e-01 -3.61188352e-01 -4.99542477e-03 1.08450890e+00
-1.23378143e-01 -5.88648677e-01 8.50485086e-01 5.97199321e-01
-1.30019024e-01 -7.75075793e-01 -9.46863830e-01 -7.82336533e-01
-1.10163637e-01 -6.16490543e-01 3.03065270e-01 8.66138697e-01
-7.74341822e-01 1.41340643e-01 -2.88713813e-01 3.61790538e-01
9.28759813e-01 3.37561756e-01 8.40284169e-01 -1.09130847e+00
-6.04446471e-01 -3.33265007e-01 -3.16449493e-01 -1.16524374e+00
-2.92002201e-01 -4.00583267e-01 -2.09040046e-01 -1.25410414e+00
3.90099883e-01 -4.39093918e-01 -3.13926600e-02 1.06031686e-01
1.76273622e-02 4.55213666e-01 4.00123715e-01 4.44934517e-01
-4.89763886e-01 2.37386599e-01 1.24869668e+00 2.04594061e-01
-1.35170400e-01 -3.84904534e-01 -3.48104358e-01 6.34763300e-01
4.74301219e-01 -2.51349807e-01 -3.82922858e-01 -7.22255051e-01
2.71746784e-01 4.03492630e-01 7.90325582e-01 -1.29697335e+00
2.89397597e-01 -2.82736838e-01 4.63556916e-01 -8.63480806e-01
8.14069867e-01 -7.01506495e-01 4.81152952e-01 3.30188334e-01
7.00879246e-02 1.63750514e-01 2.94731498e-01 5.62406540e-01
-9.64103341e-02 1.28514290e-01 9.28617060e-01 -4.58425999e-01
-1.16786444e+00 3.61924529e-01 7.58047104e-02 8.51122849e-03
1.00057685e+00 -7.27083921e-01 -3.49184334e-01 -2.04192236e-01
-5.83062470e-01 1.16233319e-01 1.11854982e+00 3.82628798e-01
3.34591389e-01 -1.20072806e+00 -5.87741494e-01 3.98864597e-02
2.56859690e-01 9.07104556e-03 6.85772538e-01 1.01130497e+00
-1.73402920e-01 3.20683211e-01 -3.48385125e-01 -1.03219843e+00
-1.40004241e+00 3.04883301e-01 3.53038937e-01 -1.79021955e-01
-6.12479091e-01 8.85919213e-01 3.88088018e-01 4.61831763e-02
-3.82853970e-02 -2.13230014e-01 3.23446274e-01 1.64088950e-01
3.23429942e-01 5.17539740e-01 1.98909983e-01 -6.63362384e-01
-2.58606434e-01 9.24179077e-01 1.76036917e-02 -3.92252773e-01
1.28047419e+00 -6.02386713e-01 2.33766213e-01 4.10725445e-01
1.22893023e+00 8.37476365e-03 -1.80748117e+00 -3.49876672e-01
-5.46015739e-01 -9.10487831e-01 3.44728589e-01 -4.99654889e-01
-8.72326314e-01 6.27678216e-01 1.01089227e+00 9.05441344e-02
1.28591585e+00 1.97035372e-01 7.70528197e-01 -5.51855331e-03
6.87564850e-01 -8.69525075e-01 2.45343111e-02 5.24476051e-01
5.69429159e-01 -1.62558889e+00 3.86081845e-01 -4.41290319e-01
-3.81553292e-01 7.20738113e-01 6.55283272e-01 -1.20739497e-01
4.40783501e-01 2.99887925e-01 -2.05996111e-01 -3.64209086e-01
-1.01338053e+00 -4.29588735e-01 2.63585210e-01 4.37740535e-01
2.56806314e-01 -1.24484688e-01 1.23244256e-01 -1.02890819e-01
-1.75933167e-01 -6.65830215e-04 1.73605844e-01 1.09055877e+00
-4.31433409e-01 -9.03191686e-01 -5.05356669e-01 4.05446529e-01
-1.35955408e-01 -1.36562675e-01 -1.81675088e-02 8.61381769e-01
5.69972873e-01 9.33778346e-01 2.55438924e-01 -4.86351758e-01
3.71382654e-01 -2.52447516e-01 8.81442070e-01 -5.26223958e-01
-5.67334354e-01 2.24973202e-01 2.77388215e-01 -9.78029013e-01
-8.16082954e-01 -9.69660997e-01 -7.38700449e-01 -6.02818906e-01
-3.14208508e-01 -4.82569575e-01 3.46490502e-01 7.73540258e-01
4.33699012e-01 2.42116079e-01 4.76117015e-01 -1.39432931e+00
-2.90211618e-01 -6.66293025e-01 -3.91065508e-01 4.60826188e-01
5.68498850e-01 -8.85640085e-01 -5.34199119e-01 9.93096828e-02] | [8.401569366455078, -2.228623628616333] |
61d0e99e-35c3-4140-8c4b-ffb1dcf2378e | proceedings-of-the-4th-international-workshop-2 | 2211.13285 | null | https://arxiv.org/abs/2211.13285v1 | https://arxiv.org/pdf/2211.13285v1.pdf | Proceedings of the 4th International Workshop on Reading Music Systems | The International Workshop on Reading Music Systems (WoRMS) is a workshop that tries to connect researchers who develop systems for reading music, such as in the field of Optical Music Recognition, with other researchers and practitioners that could benefit from such systems, like librarians or musicologists. The relevant topics of interest for the workshop include, but are not limited to: Music reading systems; Optical music recognition; Datasets and performance evaluation; Image processing on music scores; Writer identification; Authoring, editing, storing and presentation systems for music scores; Multi-modal systems; Novel input-methods for music to produce written music; Web-based Music Information Retrieval services; Applications and projects; Use-cases related to written music. These are the proceedings of the 4th International Workshop on Reading Music Systems, held online on Nov. 18th 2022. | ['Elona Shatri', 'Alexander Pacha', 'Jorge Calvo-Zaragoza'] | 2022-11-23 | null | null | null | null | ['music-information-retrieval'] | ['music'] | [ 3.87803137e-01 -4.67159003e-01 -1.32238135e-01 2.77526081e-01
-8.39178801e-01 -1.07207441e+00 1.89017847e-01 -9.23499092e-02
-5.34065552e-02 -1.91968605e-01 2.81907886e-01 -3.57777090e-03
-7.99565196e-01 -3.11198920e-01 -1.88211054e-02 -2.26325899e-01
2.27194563e-01 5.34802914e-01 1.49478197e-01 -6.20694719e-02
1.09137106e+00 6.93653226e-01 -2.06780910e+00 4.68646139e-01
2.24981755e-01 6.88056231e-01 1.96703270e-01 1.23279440e+00
-4.27113861e-01 4.89207119e-01 -1.04759097e+00 -2.96300620e-01
5.15971780e-01 -7.18956470e-01 -7.75032401e-01 -2.11381555e-01
6.71845198e-01 -8.00434947e-02 1.62936598e-02 1.29298162e+00
8.89429569e-01 1.58564091e-01 5.37267685e-01 -8.73434424e-01
-7.45458961e-01 1.20775151e+00 -7.21415579e-01 6.65092096e-02
9.37300742e-01 -2.43757337e-01 1.05954230e+00 -7.58317828e-01
5.55733860e-01 1.06180358e+00 7.32791483e-01 3.79636139e-01
-1.00402856e+00 -8.68240356e-01 -7.12258339e-01 2.96172738e-01
-1.47821498e+00 -7.31800199e-01 6.78043187e-01 -4.65490490e-01
6.98558152e-01 8.27597022e-01 8.66300225e-01 4.93573189e-01
-4.43473101e-01 9.45840657e-01 7.92266190e-01 -1.01945925e+00
-1.42772079e-01 1.77110642e-01 1.92046002e-01 3.52345884e-01
-1.35273457e-01 -1.37196124e-01 -1.18141067e+00 -3.07466816e-02
1.33384001e+00 -6.08482063e-01 -1.14341937e-01 5.57748526e-02
-1.67202544e+00 1.80679679e-01 -2.68982828e-01 6.11138940e-01
-2.56131768e-01 -3.24252881e-02 4.50736195e-01 7.51045823e-01
-2.36088485e-01 1.00773478e+00 -1.61437735e-01 -8.07447076e-01
-1.30321848e+00 4.41742957e-01 9.54432368e-01 1.02384973e+00
1.71192270e-02 1.86119471e-02 -1.01194270e-01 1.48775697e+00
4.00717795e-01 6.84522092e-01 7.79515743e-01 -1.24707925e+00
1.67979985e-01 4.21535403e-01 5.05580902e-02 -5.95761776e-01
-2.92962909e-01 -2.54523866e-02 -3.55140239e-01 4.18038636e-01
4.98351723e-01 2.63118118e-01 -4.85228390e-01 8.97602916e-01
-2.21087232e-01 1.02436200e-01 -1.55037805e-01 1.05366373e+00
1.32937181e+00 6.98455811e-01 -5.76868176e-01 -1.77004963e-01
1.43487692e+00 -8.07204187e-01 -6.88760102e-01 6.48370266e-01
-7.82668591e-03 -1.96371520e+00 1.06715298e+00 1.19281030e+00
-1.74944651e+00 -7.86531806e-01 -9.00092959e-01 -1.36191070e-01
-1.13708168e-01 4.88297611e-01 4.18438852e-01 7.90650308e-01
-1.09893560e+00 8.59155834e-01 -2.38655016e-01 -7.00422585e-01
-1.39754727e-01 5.40070951e-01 2.70786613e-01 5.57049215e-01
-4.77534056e-01 6.09266818e-01 9.59937721e-02 -4.71612394e-01
-1.72630534e-01 -7.23500550e-01 1.64982066e-01 -8.89212862e-02
4.54873918e-03 -5.14211118e-01 1.62358749e+00 -1.16370523e+00
-2.03764987e+00 1.32762969e+00 8.00814852e-02 1.40497521e-01
1.53444052e-01 -3.58466923e-01 -8.54869843e-01 -3.04027405e-02
7.79257938e-02 4.41734672e-01 6.70599759e-01 -8.17247272e-01
-6.39194965e-01 -4.18558896e-01 -2.83889860e-01 3.93989116e-01
-2.02531248e-01 6.98818386e-01 -9.69524384e-01 -9.16181684e-01
2.68352568e-01 -9.83453929e-01 6.29210353e-01 -2.17593417e-01
-6.86375856e-01 -2.17121631e-01 6.20534241e-01 -6.07617617e-01
1.34882188e+00 -2.38860822e+00 2.11097315e-01 7.28852153e-01
-5.03036022e-01 5.53457737e-01 -2.79291511e-01 6.08290732e-01
1.28403768e-01 -2.32888490e-01 6.00047588e-01 2.06244543e-01
3.05326968e-01 -4.45376992e-01 -3.30519527e-01 -8.32894295e-02
-8.68827999e-01 4.62840587e-01 -6.18801594e-01 -4.35164869e-01
3.90724331e-01 1.10388145e-01 -1.01846596e-02 -1.58878207e-01
1.57359764e-01 1.02490403e-01 -3.22191775e-01 8.63461792e-01
3.10767233e-01 5.57980537e-02 -1.10997371e-01 -6.17545284e-02
-5.47324240e-01 3.52675974e-01 -1.98713279e+00 2.11165953e+00
-7.92916119e-02 9.81941223e-01 4.16613407e-02 -1.85660273e-01
1.26745331e+00 6.23986185e-01 6.57912910e-01 -5.02034843e-01
-5.90217672e-02 6.51984155e-01 -9.07957181e-02 -5.45265317e-01
1.06297755e+00 9.21630040e-02 3.76944363e-01 9.51841474e-01
-9.65209827e-02 -3.11532527e-01 5.72023749e-01 2.18567885e-02
7.36443102e-01 3.09182525e-01 2.96609197e-02 -1.59797952e-01
5.72987497e-01 2.31414642e-02 -4.85824347e-02 8.09841514e-01
6.90168217e-02 6.68187737e-01 -1.91908225e-01 -8.28524530e-02
-1.06253743e+00 -1.21907496e+00 -3.15742344e-01 1.51749146e+00
1.72955245e-01 -7.75518894e-01 -7.76819527e-01 2.65559375e-01
-9.30218250e-02 8.55231471e-03 2.46874079e-01 3.55284303e-01
-2.64317095e-01 1.24086961e-02 1.10325861e+00 4.42202568e-01
2.17582628e-01 -1.66586101e+00 -2.93959081e-01 2.26661995e-01
1.82569206e-01 -3.33948761e-01 -7.56375253e-01 4.72043594e-03
-9.02353168e-01 -1.14269805e+00 -1.09256649e+00 -1.23048949e+00
-7.41123930e-02 2.84207612e-01 1.14304864e+00 1.50543645e-01
-5.39394200e-01 7.51248300e-01 -3.72015476e-01 -1.07087469e+00
-3.23289782e-01 3.15878689e-01 1.24734104e-01 -3.28681052e-01
8.59468102e-01 -6.49861813e-01 -4.99741793e-01 4.14732426e-01
-5.79875588e-01 1.47897735e-01 5.37597954e-01 1.78983435e-01
6.38876259e-01 -4.25207138e-01 6.65421844e-01 -6.70161843e-01
8.94050598e-01 3.24598938e-01 -6.19225442e-01 3.98795277e-01
-5.85049510e-01 -3.43012214e-01 1.39989331e-01 -5.53537011e-01
-7.06271470e-01 8.73142332e-02 1.12377495e-01 -6.31486177e-02
-2.63317347e-01 1.74512155e-02 1.15894303e-02 -1.95404172e-01
9.88740742e-01 3.18833590e-01 -1.59248233e-01 -1.01874614e+00
2.69640535e-01 1.35139215e+00 1.09715724e+00 -5.47257185e-01
7.46053398e-01 1.51392803e-01 6.07509725e-02 -1.09994638e+00
-3.96142155e-01 -1.18001890e+00 -6.52876914e-01 -5.72173357e-01
5.18921554e-01 -5.62832773e-01 -8.32424521e-01 6.42351985e-01
-8.00518215e-01 3.91410403e-02 -8.23896706e-01 7.28159487e-01
-7.54437089e-01 1.11454576e-01 -8.41913700e-01 -8.64811599e-01
-6.70428395e-01 -7.89597094e-01 1.17795646e+00 6.15228176e-01
-9.27408397e-01 -5.93868732e-01 5.77224135e-01 8.20722759e-01
3.74795824e-01 -5.26588738e-01 6.55024052e-01 -2.44874999e-01
-5.50438941e-01 -1.72068805e-01 -1.40516274e-02 1.88788667e-01
-1.28184527e-01 3.98425728e-01 -1.18459260e+00 1.30202577e-01
-8.37330401e-01 -2.56536812e-01 5.85036397e-01 5.30410469e-01
1.07664454e+00 3.79559070e-01 5.25435545e-02 5.56034207e-01
1.04072797e+00 6.02291226e-01 9.58450198e-01 5.94773471e-01
5.73460460e-01 2.41477743e-01 3.32390279e-01 6.50963545e-01
-3.35844845e-01 1.14208674e+00 -1.53635442e-01 1.57464191e-01
-4.44388032e-01 -2.01022446e-01 5.87675631e-01 1.35227668e+00
-9.21352029e-01 1.82036564e-01 -4.47902054e-01 2.69925058e-01
-1.68202865e+00 -1.28667557e+00 -7.72481740e-01 2.47771382e+00
6.19822562e-01 -4.72398043e-01 7.07565188e-01 8.90937626e-01
6.94447398e-01 -2.02542573e-01 -2.19082564e-01 -7.53027678e-01
-4.49038930e-02 8.14910352e-01 1.91745147e-01 -8.99187941e-03
-9.38878179e-01 8.24898303e-01 6.96873617e+00 9.35788810e-01
-1.01384997e+00 -1.85613796e-01 -4.79481369e-01 -4.70521212e-01
-1.03779212e-01 -1.94442660e-01 -6.73649430e-01 -6.76061446e-03
7.79443860e-01 -3.20818096e-01 1.08391762e+00 4.70424116e-01
3.49547535e-01 1.80982389e-02 -8.66454422e-01 1.48462331e+00
2.75426716e-01 -1.48028243e+00 1.82916805e-01 3.64749618e-02
1.14391088e+00 -1.50002157e-02 -1.47809070e-02 -1.68431163e-01
-2.70395428e-02 -8.69344890e-01 1.04413700e+00 1.02712333e+00
8.38098407e-01 -7.41099477e-01 3.23049635e-01 -6.81825951e-02
-1.31683826e+00 9.92604420e-02 -3.31707805e-01 -7.29014957e-03
3.41452025e-02 2.10946485e-01 -2.16581613e-01 2.28750437e-01
7.26198912e-01 8.28376055e-01 -6.92657650e-01 1.88106358e+00
6.55365884e-02 4.96703565e-01 -7.68790096e-02 -2.20352739e-01
-5.25567412e-01 -2.79462516e-01 8.58004689e-01 1.24219394e+00
8.71999085e-01 -6.89988583e-02 -1.88666463e-01 5.95182061e-01
-2.57927347e-02 7.59830952e-01 -2.70124406e-01 -3.13190609e-01
4.56987113e-01 1.23615944e+00 -7.27804005e-01 -2.83605844e-01
-1.04727253e-01 9.34083462e-01 -7.66451716e-01 1.22993127e-01
-1.57585025e-01 -7.96034038e-01 2.96165228e-01 4.20750201e-01
-1.68303087e-01 6.81535387e-03 -7.44601727e-01 -7.87689924e-01
-1.78554028e-01 -1.29758728e+00 2.41423294e-01 -1.25048137e+00
-1.15403366e+00 -4.84194942e-02 -3.58431816e-01 -1.54072666e+00
-1.60809636e-01 -7.66718626e-01 -6.51713431e-01 9.11753356e-01
-3.98453116e-01 -1.09393311e+00 -1.19523592e-01 7.84288704e-01
4.71860170e-01 -1.01219749e+00 1.36284041e+00 6.69406354e-01
1.04468279e-01 7.08791554e-01 5.13818264e-01 3.25093903e-02
1.13570464e+00 -1.28069139e+00 -1.29898846e-01 8.42962936e-02
1.28339791e+00 5.17232180e-01 1.29826754e-01 -2.01778963e-01
-1.67348516e+00 -3.46770763e-01 1.13918686e+00 -3.97423685e-01
3.85163486e-01 1.96333140e-01 -3.93462777e-01 4.24113691e-01
2.74572790e-01 -1.13668990e+00 1.06655681e+00 6.10591590e-01
-2.66487598e-01 -3.34265858e-01 -6.85910106e-01 3.25239718e-01
1.03150141e+00 -7.84868121e-01 -4.80497986e-01 3.64396185e-01
-3.84187460e-01 -4.67214674e-01 -8.69078696e-01 -4.40947294e-01
1.26932228e+00 -7.67815530e-01 9.14725006e-01 -4.11386304e-02
2.53087997e-01 -7.90097892e-01 2.34080516e-02 -8.46361041e-01
-4.74425942e-01 -1.05681205e+00 4.61615741e-01 1.43285108e+00
4.88649368e-01 1.76212534e-01 5.56878328e-01 -1.16749309e-01
-2.87942290e-01 2.84903854e-01 -5.02306104e-01 -7.11924791e-01
-4.35107946e-01 -7.95492828e-01 4.98033077e-01 7.31696784e-01
5.68259716e-01 4.70469326e-01 -4.89843905e-01 -2.24208221e-01
5.12380719e-01 2.65052795e-01 1.25459433e+00 -1.65087295e+00
-8.76236856e-01 -9.67064977e-01 -8.20401490e-01 -8.91562879e-01
-5.45752168e-01 -9.40074205e-01 -2.32029274e-01 -1.59255219e+00
3.87215108e-01 4.47482280e-02 -4.67307031e-01 3.18837792e-01
4.42386359e-01 8.50622654e-01 6.23140454e-01 7.49994397e-01
-7.89206922e-01 -6.48221016e-01 1.49012733e+00 -2.71721277e-02
-6.33565426e-01 3.70831370e-01 -7.32027709e-01 6.67146802e-01
5.48294544e-01 -2.03495234e-01 -3.72608542e-01 -2.55555689e-01
6.87382758e-01 1.16690889e-01 2.36077100e-01 -1.39397073e+00
4.32407290e-01 -8.82026553e-02 3.32340807e-01 -8.81766200e-01
1.97214171e-01 -4.11801338e-01 7.01813579e-01 -1.61517672e-02
-5.97790420e-01 1.22025840e-01 4.76365760e-02 -1.25392318e-01
-4.30953473e-01 -7.88102627e-01 5.66244841e-01 -1.96837977e-01
-6.70362175e-01 -3.65348399e-01 -6.07028067e-01 -3.29293430e-01
6.62200868e-01 -5.26614487e-01 -2.23903313e-01 -5.88522613e-01
-1.08856905e+00 -3.65245610e-01 4.62540001e-01 5.66035628e-01
3.33452672e-01 -1.39834523e+00 -6.56269789e-01 3.08373392e-01
2.23394215e-01 -8.11408103e-01 1.04663651e-02 6.40464664e-01
-8.78366828e-01 2.14229330e-01 -4.04509455e-01 -3.28947276e-01
-1.97170734e+00 3.68023254e-02 5.44197764e-03 5.82058132e-02
-5.93722045e-01 7.78539181e-01 -5.29430211e-01 -2.47804239e-01
6.15942121e-01 3.09305079e-02 -4.68364716e-01 5.46675064e-02
1.06171799e+00 8.70847821e-01 3.27322558e-02 -4.30749446e-01
1.53574944e-02 1.10003078e+00 2.90063232e-01 -7.32491970e-01
1.15134835e+00 4.03983630e-02 -1.18351825e-01 1.06879020e+00
4.10916835e-01 2.92636037e-01 -1.60250530e-01 -5.47591522e-02
1.81698337e-01 -6.24571204e-01 7.06238672e-03 -1.24661934e+00
-8.96751404e-01 7.52374589e-01 1.02949238e+00 3.87751669e-01
1.16614234e+00 -1.03580110e-01 6.43623173e-01 4.43571150e-01
3.37052017e-01 -1.72836268e+00 -1.49008036e-02 5.67545891e-01
9.16563570e-01 -5.00687480e-01 -2.69646421e-02 6.10529929e-02
-2.59198487e-01 1.75159907e+00 -2.10848171e-03 -7.59925460e-03
5.46332479e-01 2.72198439e-01 7.22052872e-01 -3.10276479e-01
-2.52573788e-01 -1.24940485e-01 7.98575282e-01 8.47603381e-01
1.20257807e+00 4.06318814e-01 -4.69077736e-01 5.61273873e-01
-8.19046378e-01 4.84122425e-01 4.64530945e-01 5.62938035e-01
-4.76450920e-01 -1.54204237e+00 -1.03367090e+00 5.36071897e-01
-5.92325687e-01 -6.31243140e-02 -1.13491130e+00 1.73985556e-01
3.24839771e-01 8.70333910e-01 -1.87142834e-03 -4.17439103e-01
6.72829211e-01 2.40711525e-01 8.51916015e-01 -6.32390797e-01
-1.22883546e+00 6.28097415e-01 2.35945389e-01 -2.99533606e-01
-6.81972802e-01 -9.02463436e-01 -1.23069406e+00 -6.36847019e-01
-4.08636630e-01 2.46495754e-01 1.19922400e+00 4.48396772e-01
1.62895277e-01 4.50491220e-01 1.86354935e-01 -1.02752578e+00
-2.36363590e-01 -1.22534657e+00 -1.52189493e+00 4.01252836e-01
-3.07094157e-01 1.04253113e-01 -2.83352472e-02 5.39320827e-01] | [16.041709899902344, 5.194366455078125] |
494e785a-1a12-4253-af1d-2e5686353f38 | zeroavatar-zero-shot-3d-avatar-generation | 2305.16411 | null | https://arxiv.org/abs/2305.16411v1 | https://arxiv.org/pdf/2305.16411v1.pdf | ZeroAvatar: Zero-shot 3D Avatar Generation from a Single Image | Recent advancements in text-to-image generation have enabled significant progress in zero-shot 3D shape generation. This is achieved by score distillation, a methodology that uses pre-trained text-to-image diffusion models to optimize the parameters of a 3D neural presentation, e.g. Neural Radiance Field (NeRF). While showing promising results, existing methods are often not able to preserve the geometry of complex shapes, such as human bodies. To address this challenge, we present ZeroAvatar, a method that introduces the explicit 3D human body prior to the optimization process. Specifically, we first estimate and refine the parameters of a parametric human body from a single image. Then during optimization, we use the posed parametric body as additional geometry constraint to regularize the diffusion model as well as the underlying density field. Lastly, we propose a UV-guided texture regularization term to further guide the completion of texture on invisible body parts. We show that ZeroAvatar significantly enhances the robustness and 3D consistency of optimization-based image-to-3D avatar generation, outperforming existing zero-shot image-to-3D methods. | ['Serena Yeung', 'Zeyu Wang', 'Zhenzhen Weng'] | 2023-05-25 | null | null | null | null | ['3d-shape-generation', 'image-to-3d'] | ['computer-vision', 'computer-vision'] | [ 3.12746555e-01 4.42005962e-01 4.26466674e-01 -3.99207287e-02
-5.23773611e-01 -3.73799741e-01 7.98489451e-01 -2.25206941e-01
3.13162394e-02 4.42032069e-01 5.59748769e-01 2.50488490e-01
1.89012051e-01 -7.96011686e-01 -8.74627411e-01 -7.85413682e-01
4.01597440e-01 6.71438158e-01 6.82478100e-02 -3.66850346e-01
6.22605756e-02 4.41048265e-01 -1.39563763e+00 -1.65609568e-01
8.61119688e-01 9.31620479e-01 2.01917347e-02 8.14724743e-01
5.22301383e-02 5.26499093e-01 -5.70653021e-01 -6.07478499e-01
4.75250572e-01 -8.32891226e-01 -4.17408168e-01 3.53704304e-01
5.64421594e-01 -4.50244755e-01 -3.35639566e-01 7.92551458e-01
8.73654008e-01 4.10768092e-01 1.17536545e+00 -8.24445307e-01
-9.48806882e-01 1.39586523e-01 -7.44814098e-01 -4.72181082e-01
3.37658703e-01 2.50730932e-01 7.53234148e-01 -8.71439755e-01
9.51600313e-01 1.41105747e+00 6.15968883e-01 8.05002272e-01
-1.35623908e+00 -1.84903443e-01 -2.95959681e-01 -3.12923014e-01
-1.34254587e+00 -4.10851359e-01 9.29796338e-01 -5.60377896e-01
6.57501757e-01 1.35693401e-01 1.03070092e+00 1.07931972e+00
2.02059165e-01 6.48900867e-01 8.72207820e-01 -6.63212180e-01
2.28796408e-01 -6.68457076e-02 -4.43179846e-01 1.06129634e+00
-8.57073218e-02 4.99042608e-02 -7.36768723e-01 -1.97741240e-01
1.20774913e+00 -4.41838712e-01 -1.15905337e-01 -7.25063801e-01
-9.30393755e-01 7.19917476e-01 3.63729805e-01 -1.88908845e-01
-5.36771417e-01 3.54957849e-01 5.45625202e-03 -2.16820434e-01
8.75666261e-01 3.85543138e-01 3.27154934e-01 6.54252321e-02
-9.55139041e-01 6.03669763e-01 7.24156022e-01 8.37634802e-01
6.20095491e-01 3.68554413e-01 -5.71082056e-01 1.01723838e+00
4.20873702e-01 8.61841679e-01 1.79062411e-02 -1.30333591e+00
1.16945095e-01 4.27903771e-01 1.60185829e-01 -8.07050586e-01
-2.01338506e-03 -1.65732995e-01 -8.26970160e-01 4.67499554e-01
3.70649427e-01 -1.51327372e-01 -1.47015762e+00 1.66322780e+00
8.79616380e-01 -1.09384544e-01 -1.16857074e-01 1.34223509e+00
7.27505863e-01 7.44937301e-01 -2.14368373e-01 2.20580801e-01
1.19635534e+00 -9.35081840e-01 -5.55325270e-01 -6.19397834e-02
1.61053091e-01 -7.19413102e-01 1.05855203e+00 8.23969916e-02
-1.58850634e+00 -2.38561064e-01 -1.01483572e+00 -1.20665208e-01
1.30672753e-01 -1.32800832e-01 3.75286341e-01 7.51395524e-01
-9.79955733e-01 4.45111483e-01 -6.60256565e-01 -3.49613935e-01
4.30387259e-01 1.55113086e-01 -2.53622800e-01 -4.56225015e-02
-9.85010266e-01 8.54160547e-01 -1.84435621e-02 -2.03693926e-01
-1.09601033e+00 -8.56836140e-01 -1.07118285e+00 -2.07249880e-01
2.62140334e-01 -1.35798860e+00 1.15220654e+00 -7.28310823e-01
-1.93262053e+00 1.06461942e+00 4.33795303e-02 -3.11595827e-01
5.46248794e-01 -2.21378207e-01 4.08894449e-01 3.58113199e-01
-6.26589507e-02 9.63679671e-01 1.20097911e+00 -1.42667365e+00
1.88579187e-01 -2.09616482e-01 -2.01300666e-01 6.95884168e-01
-8.75677094e-02 -2.46131703e-01 -8.25732768e-01 -9.46717322e-01
4.59099561e-03 -9.13349807e-01 -2.85713226e-01 4.59387511e-01
-6.04454458e-01 2.27370918e-01 5.59056997e-01 -7.82145679e-01
6.85717881e-01 -1.98412156e+00 6.22447848e-01 4.09122407e-01
3.44576001e-01 -7.17446581e-02 -1.44379377e-01 1.18906409e-01
3.52410108e-01 1.71887670e-02 -4.77889895e-01 -6.66759491e-01
6.58720210e-02 1.45993248e-01 1.15151182e-01 4.77633864e-01
2.68698335e-01 1.14587688e+00 -6.34334743e-01 -6.00508928e-01
4.41929698e-01 9.19531882e-01 -8.04266095e-01 3.30103129e-01
-4.70949143e-01 7.16867507e-01 -4.28654343e-01 5.54790378e-01
5.67709386e-01 -2.62652367e-01 -2.43995205e-01 -1.82175159e-01
9.42636505e-02 -2.58845359e-01 -8.47699106e-01 2.01551557e+00
-4.13672984e-01 2.57881135e-01 2.52438039e-01 -5.82377791e-01
9.09348369e-01 1.47335827e-01 7.79726326e-01 -7.23285317e-01
2.77616709e-01 8.99320748e-03 -3.97743076e-01 -2.90310085e-01
6.57009721e-01 -5.59805810e-01 1.65046658e-02 4.41049278e-01
1.12454146e-01 -9.17233407e-01 -1.50716022e-01 2.91865170e-01
7.45510757e-01 4.96008962e-01 -2.30884597e-01 -1.11592427e-01
1.37304783e-01 -7.63314739e-02 6.60322458e-02 6.84762955e-01
2.47296572e-01 1.39354169e+00 1.72301531e-01 -1.38416424e-01
-1.64164841e+00 -1.25678372e+00 2.18824774e-01 6.18077576e-01
2.38052711e-01 -3.10422778e-01 -1.26326299e+00 -3.80509526e-01
-2.02725362e-02 6.64697111e-01 -8.74106705e-01 -1.71367407e-01
-3.35861474e-01 -7.18624949e-01 5.61011314e-01 9.06936824e-02
4.71184403e-01 -6.01365268e-01 -6.54175580e-01 1.75407216e-01
-3.59972388e-01 -8.84529769e-01 -8.55731189e-01 -1.57654554e-01
-7.14873850e-01 -5.68482637e-01 -1.64211011e+00 -4.28718597e-01
7.87721932e-01 -5.41791953e-02 1.11223483e+00 5.92171177e-02
-6.62456214e-01 6.18940651e-01 -1.64511040e-01 -5.03915250e-01
-5.08005679e-01 -2.30317846e-01 -8.12345967e-02 2.00559542e-01
-4.03445274e-01 -3.92160028e-01 -8.35233331e-01 3.03061038e-01
-8.54712665e-01 4.09740418e-01 3.15658480e-01 7.94704854e-01
6.68927073e-01 -3.39315981e-01 -1.80545375e-02 -2.66826063e-01
6.06350541e-01 -1.18079692e-01 -2.69818872e-01 1.86074466e-01
-2.96696126e-01 3.76838595e-01 3.33514586e-02 -6.69939280e-01
-1.29566038e+00 1.64733663e-01 -2.91069329e-01 -8.33663940e-01
1.25848845e-01 -3.45354937e-02 -1.27990721e-02 -3.42265874e-01
9.07605410e-01 1.82137609e-01 3.88725370e-01 -3.04652601e-01
7.77199090e-01 2.15632498e-01 5.36884427e-01 -7.81400263e-01
9.13697064e-01 7.31064618e-01 1.13611400e-01 -1.15428543e+00
-7.73236632e-01 -1.95897132e-01 -4.66531515e-01 -6.90502644e-01
1.30210471e+00 -7.23783851e-01 -3.05191368e-01 7.04056740e-01
-1.13616323e+00 -8.49108040e-01 -6.98878765e-01 1.98079675e-01
-9.34659839e-01 4.08326209e-01 -6.66589737e-01 -1.00815296e+00
-6.04572952e-01 -8.16773057e-01 1.62945008e+00 1.05043836e-01
-3.27963054e-01 -9.89732563e-01 2.76961625e-01 5.05129755e-01
4.25784111e-01 4.99083966e-01 8.69172812e-01 4.34042841e-01
-5.16053259e-01 9.63777862e-03 -1.22284144e-02 2.54913628e-01
-1.72424272e-01 -2.14276642e-01 -8.22407722e-01 1.17025683e-02
-9.54098627e-02 -5.82479298e-01 7.17592180e-01 8.72115493e-01
6.44318044e-01 -2.62816042e-01 -1.45575598e-01 8.39199662e-01
1.01207781e+00 -3.56602788e-01 8.69472146e-01 -1.19573407e-01
8.71580720e-01 7.82780707e-01 3.87981981e-01 7.55257130e-01
2.89289981e-01 9.43962395e-01 3.21103930e-01 -3.86729002e-01
-8.29759598e-01 -7.61290491e-01 2.41046444e-01 5.92189729e-01
-4.37789470e-01 -4.11176026e-01 -4.89604324e-01 3.24117064e-01
-1.48284733e+00 -7.74180889e-01 -6.50856569e-02 2.24431324e+00
7.70117342e-01 -1.47375524e-01 1.38857156e-01 -2.46022180e-01
5.42989910e-01 2.04989925e-01 -5.15286922e-01 -1.00586936e-01
-3.30178559e-01 2.26026729e-01 4.05874401e-01 7.51170039e-01
-6.78588629e-01 1.21727097e+00 6.65949726e+00 9.60076749e-01
-9.16020930e-01 1.89806700e-01 6.67290986e-01 -2.78861612e-01
-6.89459860e-01 -1.92800120e-01 -5.87843597e-01 9.12598446e-02
3.97660881e-01 1.10726379e-01 3.78947288e-01 4.07456547e-01
3.66208762e-01 -1.36878744e-01 -5.76997399e-01 1.10387933e+00
5.66449761e-01 -1.28250539e+00 3.17235619e-01 3.17309082e-01
1.03360844e+00 -2.73732662e-01 1.43258199e-01 -1.58682108e-01
3.32537204e-01 -9.44999814e-01 1.22833645e+00 9.08969045e-01
1.12434065e+00 -5.04744768e-01 1.12775542e-01 1.87514871e-01
-9.49327946e-01 3.88309538e-01 -3.21312100e-01 3.17496449e-01
6.34792626e-01 6.68037176e-01 -6.60895407e-01 3.40518117e-01
5.43453634e-01 3.12474877e-01 -2.91438013e-01 9.97654021e-01
-2.03946531e-01 3.11487466e-01 -5.59490681e-01 -4.08511311e-02
9.40905288e-02 -3.28617036e-01 9.31931734e-01 7.61616290e-01
7.03726888e-01 2.75741071e-01 -1.18198365e-01 1.43047619e+00
-3.63633893e-02 1.83615878e-01 -7.68058419e-01 -5.65717276e-03
-1.02309272e-01 1.01956999e+00 -6.91016138e-01 -2.88973361e-01
5.54822665e-03 1.59477186e+00 1.05726525e-01 5.09802401e-01
-8.23181808e-01 -1.37853593e-01 3.86682391e-01 5.39623976e-01
1.32900462e-01 -3.90689373e-01 -2.72288084e-01 -1.11337686e+00
-1.05863675e-01 -6.41325891e-01 -1.84562460e-01 -1.28967309e+00
-1.07826674e+00 1.93694949e-01 -4.52478863e-02 -7.79086471e-01
-2.01594606e-01 -3.04558784e-01 -5.16551495e-01 9.41654801e-01
-1.03391051e+00 -1.42142260e+00 -3.03316325e-01 5.90427458e-01
5.58667958e-01 2.80707590e-02 6.09943092e-01 -1.80516392e-02
-1.12751573e-01 4.18310165e-01 -1.73194468e-01 -2.01862544e-01
7.77296126e-01 -1.00431919e+00 6.43630505e-01 6.14121795e-01
-4.64926511e-02 2.55787522e-01 6.71559572e-01 -1.09485316e+00
-1.50574458e+00 -8.36814702e-01 3.95601571e-01 -5.29345572e-01
1.58702120e-01 -5.78005016e-01 -5.40246964e-01 5.17770089e-02
1.55421898e-01 -2.50870198e-01 3.22459012e-01 -3.40053082e-01
-1.86812170e-02 2.98052192e-01 -1.03165400e+00 8.95893395e-01
1.12119210e+00 -3.37934941e-01 -3.00703019e-01 1.88566595e-01
5.39589345e-01 -7.06448317e-01 -8.48849833e-01 2.51272619e-01
6.20209336e-01 -7.61468649e-01 1.21086109e+00 -1.48731545e-01
5.06192207e-01 -2.13073254e-01 5.20817610e-03 -1.39394999e+00
-2.14677706e-01 -9.74198103e-01 -3.86731267e-01 8.70856404e-01
3.50792527e-01 1.98913775e-02 1.11912560e+00 7.53596544e-01
-2.03415692e-01 -4.17773187e-01 -8.10368419e-01 -5.02038836e-01
1.03379786e-01 -2.55985439e-01 3.26518565e-01 6.41484022e-01
-3.50423992e-01 3.31633210e-01 -9.33946192e-01 -1.38079181e-01
1.10592616e+00 -3.10036004e-01 9.85384881e-01 -1.04308355e+00
-4.21259254e-01 -3.17338198e-01 -2.52356112e-01 -1.43587351e+00
-2.52405137e-01 -7.67844200e-01 3.83850336e-01 -1.68095517e+00
1.94164097e-01 -3.52675080e-01 4.59043920e-01 1.94279969e-01
-1.95193022e-01 6.09994173e-01 2.20292121e-01 9.15165842e-02
-2.72336602e-01 1.08127415e+00 1.84053743e+00 -1.78645730e-01
-4.06315535e-01 -2.63258815e-01 -4.20598984e-01 4.50346917e-01
5.74541271e-01 -2.64194041e-01 -3.65874201e-01 -4.45605427e-01
2.21752152e-01 1.20091304e-01 7.47785747e-01 -9.73113179e-01
4.14443351e-02 -3.84906679e-02 5.17675996e-01 -4.28078413e-01
8.18434536e-01 -3.49197537e-01 1.73810720e-01 2.26335391e-01
-1.80988327e-01 -3.88653398e-01 2.18172488e-03 6.87967718e-01
2.24609077e-01 -2.10711807e-01 9.19379950e-01 -3.22683990e-01
-2.41124779e-01 5.14306426e-01 -3.32123548e-01 2.44016200e-01
6.87929988e-01 -3.14717919e-01 7.21803233e-02 -6.99237466e-01
-7.40525663e-01 -1.02874329e-02 8.07545006e-01 3.05443674e-01
9.12107110e-01 -1.46268260e+00 -8.22891712e-01 3.12170625e-01
-3.41009721e-02 1.33306786e-01 5.31524718e-01 8.15970063e-01
-7.36272812e-01 -1.28097028e-01 -1.06748827e-01 -7.62784123e-01
-1.22109294e+00 1.65472612e-01 7.32748330e-01 1.02806121e-01
-1.07697177e+00 9.13143098e-01 4.82071429e-01 -4.65814382e-01
2.89338917e-01 5.27283922e-02 2.48832732e-01 -3.36740971e-01
3.12477410e-01 1.29867688e-01 -2.88179547e-01 -7.86430478e-01
1.41679227e-01 1.07163143e+00 3.63910824e-01 -8.16107810e-01
1.16720009e+00 -2.35139146e-01 2.65920460e-01 1.67956755e-01
7.75116444e-01 3.36250067e-02 -1.80764449e+00 4.10449319e-02
-6.73309326e-01 -4.41413164e-01 1.98851824e-01 -7.09971666e-01
-1.07192314e+00 1.01665640e+00 5.27714789e-01 -2.35826299e-01
6.62799358e-01 1.09519675e-01 8.91303360e-01 -3.02390400e-02
1.94497526e-01 -1.26236212e+00 6.32688463e-01 5.45489430e-01
1.04815555e+00 -9.21396434e-01 4.70626131e-02 -4.39315021e-01
-8.22401047e-01 7.62179613e-01 3.86789978e-01 -1.11145135e-02
5.78800917e-01 1.91067964e-01 -1.42631331e-03 -4.35815871e-01
-3.51623654e-01 -2.66996384e-01 6.60473466e-01 9.01371717e-01
1.81441903e-01 -2.38855228e-01 1.07299596e-01 1.33934021e-01
-2.20418558e-01 -1.30682722e-01 2.44076520e-01 6.37800872e-01
-2.25789323e-01 -8.35089684e-01 -6.53167963e-01 1.48585796e-01
-2.87454605e-01 -2.14079961e-01 -6.01048529e-01 3.22629184e-01
-3.51270027e-02 5.92606843e-01 -6.58846721e-02 -3.17412287e-01
3.61931294e-01 1.67121440e-02 1.02555001e+00 -5.66161215e-01
-3.85901660e-01 4.75112915e-01 3.95603925e-02 -3.39429229e-01
-3.16922843e-01 -4.67225850e-01 -1.05271947e+00 -2.48733848e-01
-3.79857242e-01 -1.54153988e-01 7.98053622e-01 6.48215950e-01
3.05259228e-01 3.47717285e-01 1.89780056e-01 -1.44629109e+00
-2.05732465e-01 -7.42922366e-01 -7.09960520e-01 5.56146502e-01
8.63902718e-02 -8.38626921e-01 -2.53373653e-01 1.84797674e-01] | [9.412849426269531, -3.2164573669433594] |
dc55ca9e-0e4f-4cae-9ce0-94c701388b76 | automated-audio-captioning-by-fine-tuning | null | null | https://dcase.community/documents/workshop2021/proceedings/DCASE2021Workshop_Gontier_57.pdf | https://dcase.community/documents/workshop2021/proceedings/DCASE2021Workshop_Gontier_57.pdf | AUTOMATED AUDIO CAPTIONING BY FINE-TUNING BART WITH AUDIOSET TAGS | utomated audio captioning is the multimodal task of describing
environmental audio recordings with fluent natural language. Most
current methods utilize pre-trained analysis models to extract rele-
vant semantic content from the audio input. However, prior infor-
mation on language modeling is rarely introduced, and correspond-
ing architectures are limited in capacity due to data scarcity. In
this paper, we present a method leveraging the linguistic informa-
tion contained in BART, a large-scale conditional language model
with general purpose pre-training. The caption generation is condi-
tioned on sequences of textual AudioSet tags. This input is enriched
with temporally aligned audio embeddings that allows the model to
improve the sound event recognition. The full BART architecture
is fine-tuned with few additional parameters. Experimental results
demonstrate that, beyond the scaling properties of the architecture,
language-only pre-training improves the text quality in the multi-
modal setting of audio captioning. The best model achieves state-
of-the-art performance on AudioCaps with 46.5 SPIDEr. | ['Christophe Cerisara', 'Romain Serizel', 'F ́elix Gontier'] | 2021-11-15 | null | null | null | dcase-workshop-2021-11 | ['audio-captioning'] | ['audio'] | [ 2.79593796e-01 4.23542708e-02 -6.39001504e-02 -3.72009456e-01
-1.54907513e+00 -5.21989405e-01 2.83619702e-01 1.20853111e-01
-4.94049698e-01 5.56735158e-01 8.45076978e-01 8.47605690e-02
3.98562163e-01 -3.47884387e-01 -9.76305842e-01 -2.67637223e-01
-1.43892944e-01 4.17392194e-01 -4.79039401e-02 -1.13735072e-01
-2.42559791e-01 -1.76523194e-01 -1.65632439e+00 9.03117001e-01
3.54049116e-01 1.09072554e+00 5.40165961e-01 1.24526262e+00
-3.14001560e-01 8.43597114e-01 -4.07444954e-01 -1.69917747e-01
-3.03974271e-01 -5.11801958e-01 -8.05680096e-01 -6.01326600e-02
2.45001882e-01 -2.28784904e-01 -4.80606228e-01 5.52624285e-01
7.26886034e-01 2.71126807e-01 4.74415183e-01 -1.22081792e+00
-5.34075916e-01 1.17406738e+00 -4.80197109e-02 2.39580497e-01
5.37906289e-01 -1.12758018e-01 1.49602091e+00 -1.10095906e+00
2.61102468e-01 1.36865342e+00 5.41265190e-01 8.43368888e-01
-1.29476690e+00 -5.50445318e-01 4.54866767e-01 3.22770566e-01
-1.39742744e+00 -7.68895864e-01 8.89757872e-01 -5.22355795e-01
9.67384279e-01 3.35047632e-01 3.05762708e-01 1.63274586e+00
-2.36679152e-01 1.01991689e+00 5.74094176e-01 -5.43211639e-01
2.09357619e-01 9.70281586e-02 -1.01458728e-02 1.25065342e-01
-5.87907016e-01 -1.97405592e-01 -1.35538363e+00 -1.21871106e-01
2.26641521e-01 -4.17395204e-01 -3.04349005e-01 5.02899624e-02
-1.25837755e+00 5.72615147e-01 1.65202677e-01 3.53265584e-01
-4.12642568e-01 4.87415165e-01 7.65978813e-01 7.90982321e-02
4.50867474e-01 6.48135722e-01 -4.63030428e-01 -6.20954216e-01
-1.13078785e+00 -1.65078100e-02 5.64165115e-01 9.54972804e-01
5.03104866e-01 3.32699329e-01 -3.80809337e-01 9.91977096e-01
3.63078415e-01 7.31300712e-01 6.81572020e-01 -6.89826488e-01
8.32637787e-01 -4.30089273e-02 1.02089912e-01 -4.65562731e-01
-2.41570711e-01 -3.93200845e-01 -4.52184856e-01 -5.63059807e-01
-9.81906150e-03 -4.44630384e-01 -9.86769617e-01 1.95356882e+00
-2.13444188e-01 4.63858753e-01 2.18283549e-01 9.12548065e-01
8.27025533e-01 1.16642630e+00 4.11495149e-01 8.59809294e-02
1.50464320e+00 -1.02443457e+00 -1.06684899e+00 -5.62963486e-01
2.16405958e-01 -7.81229496e-01 1.35507596e+00 2.42793843e-01
-1.01926684e+00 -7.45808244e-01 -8.79572451e-01 -5.64835854e-02
-2.66793877e-01 1.28331691e-01 2.68573195e-01 3.60334754e-01
-8.40748191e-01 7.36452937e-02 -8.77683282e-01 -2.39721030e-01
6.30870312e-02 -2.29442306e-02 -3.06087583e-01 1.03815973e-01
-1.50953782e+00 3.83496940e-01 6.22223973e-01 7.16037825e-02
-1.26455343e+00 -9.66703773e-01 -1.03688490e+00 2.92495817e-01
2.41126910e-01 -5.31234205e-01 1.75631166e+00 -8.32553566e-01
-1.67988050e+00 3.92138749e-01 -4.70050395e-01 -6.59402609e-01
9.43083167e-02 -5.03852487e-01 -7.54431546e-01 4.94557828e-01
4.55793217e-02 9.75050986e-01 7.93027341e-01 -1.21645510e+00
-5.40122151e-01 2.04061866e-01 -1.24872066e-01 3.34672034e-01
-8.01136374e-01 3.08007240e-01 -6.08685732e-01 -7.68706858e-01
-3.47651571e-01 -7.64625430e-01 3.06162634e-03 -2.45978504e-01
-4.77573067e-01 -9.91299003e-03 7.23573625e-01 -6.73547566e-01
1.49312210e+00 -2.51440620e+00 -1.19420998e-01 -1.60981357e-01
-4.02827591e-01 3.81981730e-02 -5.57293296e-01 8.53345752e-01
-2.24143714e-01 1.75823849e-02 -1.77085161e-01 -7.92375684e-01
4.78258252e-01 9.24926996e-02 -9.31856751e-01 -1.47264287e-01
4.94280219e-01 8.24453235e-01 -9.77876604e-01 -4.78220433e-01
9.02478620e-02 6.86793149e-01 -7.06603289e-01 7.20300555e-01
-5.75679243e-01 3.51841927e-01 -3.56674463e-01 2.83380628e-01
1.37401745e-01 -1.14298880e-01 -1.91986978e-01 -1.02559932e-01
-6.81879669e-02 8.14908087e-01 -1.04429638e+00 2.23382831e+00
-8.87732804e-01 7.42178559e-01 1.04343154e-01 -6.87883496e-01
6.63630724e-01 9.57416713e-01 4.34360266e-01 -2.97640115e-01
4.54115309e-02 1.18646473e-01 -2.87592083e-01 -6.83842599e-01
5.69452047e-01 -2.38109812e-01 -3.89182240e-01 4.52679604e-01
3.83104146e-01 -1.78449914e-01 2.00413257e-01 2.17252567e-01
9.32400346e-01 -4.43051457e-02 -1.38828143e-01 2.46414244e-01
3.39552432e-01 -1.70412436e-01 2.08864376e-01 6.79235995e-01
7.53978966e-04 1.07028770e+00 4.41714860e-02 7.57743195e-02
-9.80324149e-01 -1.05770612e+00 -6.29238859e-02 1.74687648e+00
-3.13560516e-01 -9.85427380e-01 -6.85346603e-01 -2.49529511e-01
-3.30743819e-01 8.80867541e-01 -5.10797858e-01 -1.62776962e-01
-4.13754910e-01 -3.70464951e-01 1.01828921e+00 7.17748106e-01
1.25441179e-01 -1.21656108e+00 -4.81840432e-01 6.05854809e-01
-7.24225879e-01 -1.46378839e+00 -7.10844219e-01 4.45399791e-01
-3.95100951e-01 -3.95000190e-01 -7.82542646e-01 -8.23222160e-01
1.08990341e-01 -1.68300793e-01 1.05354404e+00 -6.16630495e-01
-5.97899258e-02 6.86755180e-01 -6.41460776e-01 -5.61081946e-01
-5.04612029e-01 2.45308205e-01 1.79496586e-01 3.29968929e-01
4.26048875e-01 -7.07633674e-01 -2.20565274e-01 -6.61043730e-03
-8.82497787e-01 1.91750169e-01 6.55289233e-01 7.45952010e-01
6.66814446e-01 -3.03816885e-01 8.91780138e-01 -2.33353242e-01
5.67853093e-01 -5.07431924e-01 1.67141914e-01 -5.65508008e-03
9.57224295e-02 1.92897916e-01 7.47157574e-01 -8.13950658e-01
-1.04462826e+00 2.31639832e-01 -4.10975575e-01 -5.43471038e-01
-4.30372804e-01 7.05589831e-01 -2.42848873e-01 7.27666080e-01
6.30768836e-01 2.55837619e-01 -5.14199078e-01 -8.53727996e-01
5.94156086e-01 1.04018486e+00 9.09905791e-01 -7.48295367e-01
6.08752847e-01 1.99324965e-01 -4.44656640e-01 -8.84196758e-01
-1.13159108e+00 -6.44908309e-01 -3.06711674e-01 -2.23360658e-01
9.66956079e-01 -1.29160702e+00 -3.51341486e-01 2.76704784e-02
-1.28964734e+00 -3.31675053e-01 -4.31321800e-01 8.02817464e-01
-8.11830819e-01 -5.35761230e-02 -6.35380626e-01 -1.18847549e+00
-2.78394520e-01 -7.75467753e-01 1.31264377e+00 -1.90294147e-01
-3.77041370e-01 -6.33602798e-01 2.99251974e-01 3.02273929e-01
4.25906688e-01 -1.77443504e-01 6.10509276e-01 -9.53891873e-01
-2.50107110e-01 -1.04372747e-01 1.14243641e-01 4.60507691e-01
-9.26506892e-02 -3.30263853e-01 -1.80004680e+00 -3.57896276e-02
-3.07635993e-01 -7.15198278e-01 8.24040592e-01 1.98760927e-01
1.33597243e+00 -3.67773443e-01 3.76465656e-02 2.76014179e-01
9.79920268e-01 5.98163605e-02 2.91097730e-01 9.59692895e-02
5.81679761e-01 5.84662378e-01 5.37485778e-01 6.67269111e-01
2.75582671e-01 6.74332440e-01 3.83839697e-01 6.93581328e-02
-4.21812057e-01 -8.77091110e-01 6.85382843e-01 1.26154184e+00
4.21416968e-01 -4.20776933e-01 -9.99488592e-01 9.15770411e-01
-1.76696134e+00 -1.04601502e+00 3.22121620e-01 1.76584327e+00
1.32565045e+00 9.99923423e-02 -1.78017430e-02 2.27774680e-01
7.25282252e-01 2.14958698e-01 -3.80887568e-01 -2.35065371e-01
-1.39355078e-01 6.01729415e-02 3.47641110e-02 4.48050231e-01
-1.23403800e+00 8.10356200e-01 6.17045069e+00 8.09293807e-01
-1.14830744e+00 2.43590817e-01 1.10352471e-01 -3.68407965e-01
-3.64306718e-01 -3.84288609e-01 -7.34528303e-01 4.56827372e-01
1.74655497e+00 -2.68995196e-01 4.45957720e-01 6.17364824e-01
4.50166911e-01 3.62568974e-01 -1.49466968e+00 1.19618213e+00
1.87780976e-01 -1.11531818e+00 1.40966743e-01 -3.31983298e-01
4.32446957e-01 1.94404677e-01 2.89427698e-01 5.63258827e-01
-1.07488587e-01 -9.99243736e-01 1.23125815e+00 5.31293869e-01
1.11249769e+00 -6.04120195e-01 6.19328976e-01 2.63519526e-01
-1.41611183e+00 -1.59731328e-01 -3.55478525e-02 -2.09489111e-02
6.32885039e-01 3.09523940e-01 -1.19037187e+00 5.25709867e-01
6.87965751e-01 5.57597876e-01 -3.45493853e-01 1.05894578e+00
-1.45279154e-01 1.23389995e+00 -4.73726124e-01 3.52329761e-02
3.43686074e-01 6.01977229e-01 6.57836914e-01 1.86426461e+00
4.04489011e-01 -1.29379913e-01 2.59607285e-01 6.52419984e-01
-3.85266691e-01 2.68810391e-01 -4.49540794e-01 -6.49218380e-01
6.40914083e-01 8.07743311e-01 -1.30597800e-02 -2.71644920e-01
-3.43218267e-01 8.74182522e-01 -1.02615412e-02 6.01176679e-01
-8.31685901e-01 -5.23136199e-01 6.21922910e-01 9.95506942e-02
4.47610468e-01 -1.75805047e-01 -2.14782748e-02 -1.07565451e+00
-8.75347108e-03 -7.18558908e-01 4.10748541e-01 -1.15573788e+00
-1.34739971e+00 9.44996238e-01 6.54499903e-02 -1.24669707e+00
-7.28548586e-01 -4.47649628e-01 -3.93970132e-01 9.54521894e-01
-1.61312938e+00 -1.28468585e+00 3.91085148e-02 4.61050212e-01
8.81604373e-01 -2.04963997e-01 1.23136222e+00 5.51224470e-01
-3.12460274e-01 5.81185043e-01 -8.60752687e-02 2.33426139e-01
1.07363904e+00 -1.26805604e+00 2.13451788e-01 8.06126416e-01
4.40648645e-01 4.72041398e-01 1.07462895e+00 -2.47448757e-01
-1.26806974e+00 -1.39157820e+00 1.23201346e+00 -4.55511957e-01
1.05616856e+00 -8.89491141e-01 -9.11361217e-01 7.05544353e-01
5.09558380e-01 2.55486760e-02 1.16856670e+00 1.39382765e-01
-5.78847826e-01 -1.96446955e-01 -4.22976822e-01 4.20643747e-01
6.20146275e-01 -1.32626021e+00 -8.58156979e-01 1.63680613e-01
1.24126208e+00 -2.44957164e-01 -6.61161244e-01 3.21968943e-02
3.57761353e-01 -1.76988691e-01 9.18890059e-01 -8.86770904e-01
3.75354886e-01 -3.21828961e-01 -5.49529016e-01 -1.23820531e+00
1.70270652e-02 -1.02807510e+00 -1.81405559e-01 1.51034164e+00
6.64009929e-01 7.49261603e-02 3.93643230e-01 3.82083595e-01
-5.61530828e-01 -1.63192078e-01 -1.04024315e+00 -9.34912622e-01
-1.58864960e-01 -1.08466184e+00 5.45040011e-01 6.67467594e-01
3.59707206e-01 7.44741857e-01 -6.56322539e-01 4.32312548e-01
2.24477515e-01 -8.46732184e-02 6.13435924e-01 -1.02323818e+00
-3.39214385e-01 -7.84155168e-03 -7.92174116e-02 -1.24687374e+00
5.14099836e-01 -8.38766098e-01 6.74350441e-01 -1.37365651e+00
-8.72905850e-02 -3.39503735e-02 -5.17421365e-01 6.80744529e-01
-5.01409695e-02 2.98618019e-01 3.58863890e-01 -1.19995773e-01
-8.59467983e-01 9.63150084e-01 7.50814915e-01 -3.23383391e-01
-3.23507577e-01 -1.80055737e-01 -4.74106669e-01 4.90583718e-01
7.72835314e-01 -6.12360656e-01 -5.51671863e-01 -6.35740697e-01
1.95315257e-01 1.00575902e-01 2.39820451e-01 -1.30390048e+00
3.40658575e-01 -9.91116837e-02 -5.41604050e-02 -5.63703954e-01
9.94485080e-01 -8.20102274e-01 -1.69167053e-02 -1.34399697e-01
-1.01062799e+00 -1.33702576e-01 5.76105893e-01 7.14470148e-01
-6.86875939e-01 -1.74962521e-01 4.63591546e-01 1.89694017e-01
-5.74870527e-01 2.21402913e-01 -6.82823241e-01 3.67023647e-01
3.64164650e-01 3.98815513e-01 1.79395005e-02 -6.95559919e-01
-9.48157728e-01 4.87404764e-02 -2.04333216e-01 6.70574129e-01
4.37735468e-01 -1.55974698e+00 -8.92384768e-01 1.03341211e-02
5.66334724e-01 -2.22187921e-01 3.85290891e-01 3.78635049e-01
2.11936727e-01 7.72244334e-01 6.57750890e-02 -5.06617665e-01
-9.77270901e-01 4.66023833e-01 1.05735466e-01 -3.75391394e-02
-4.64458615e-01 8.90495479e-01 2.22885728e-01 -5.56630045e-02
6.44585192e-01 -4.72426444e-01 -1.65165707e-01 1.89827874e-01
8.72861922e-01 -4.50405777e-02 6.02812730e-02 -7.54562736e-01
-3.04486722e-01 1.83109790e-01 1.92847207e-01 -8.87135148e-01
1.33583403e+00 -3.25534791e-01 3.76157850e-01 1.15369487e+00
1.21744347e+00 4.84454967e-02 -1.29537821e+00 -3.95105481e-01
-6.69770539e-02 -1.21723667e-01 2.76045889e-01 -9.84909832e-01
-4.85310078e-01 1.49129236e+00 3.35962951e-01 1.60605639e-01
9.81752336e-01 1.66930661e-01 9.26561356e-01 6.22359097e-01
8.21912736e-02 -1.16951609e+00 4.03621554e-01 8.14689100e-01
1.09607708e+00 -9.81364727e-01 -8.13837647e-01 3.41611803e-02
-9.02262747e-01 1.06289518e+00 4.96731132e-01 2.24181116e-01
5.23622513e-01 3.75160098e-01 2.22727433e-01 1.72954783e-01
-1.10540974e+00 -3.91138792e-01 5.41760325e-01 6.42544866e-01
4.54667658e-01 2.25385148e-02 4.30957526e-01 1.31511903e+00
-4.32769507e-01 -2.08108827e-01 4.46383864e-01 7.10624337e-01
-5.59871554e-01 -9.03804779e-01 -4.42399412e-01 -6.61101341e-02
-5.91302991e-01 -3.96947503e-01 -2.05098510e-01 2.69110143e-01
-2.14623481e-01 1.31379211e+00 2.58783966e-01 -3.54406774e-01
3.98010403e-01 6.92168415e-01 -1.01613231e-01 -8.23715568e-01
-4.47383165e-01 3.88165861e-01 3.25269818e-01 -4.75628138e-01
-2.92590857e-01 -3.57105672e-01 -1.33256078e+00 3.53933454e-01
-8.47583786e-02 6.09375358e-01 8.32456946e-01 8.12507570e-01
4.75706756e-01 8.90555799e-01 5.01951277e-01 -9.55941617e-01
-3.46678346e-01 -1.21861756e+00 -4.07847792e-01 1.96700841e-01
7.42078304e-01 -2.91714519e-01 -4.39667016e-01 6.27844155e-01] | [15.28309440612793, 4.953114032745361] |
70b6fa2c-11ed-4795-a922-880f60740b7f | online-k-means-clustering | 1909.06861 | null | https://arxiv.org/abs/1909.06861v1 | https://arxiv.org/pdf/1909.06861v1.pdf | Online k-means Clustering | We study the problem of online clustering where a clustering algorithm has to assign a new point that arrives to one of $k$ clusters. The specific formulation we use is the $k$-means objective: At each time step the algorithm has to maintain a set of k candidate centers and the loss incurred is the squared distance between the new point and the closest center. The goal is to minimize regret with respect to the best solution to the $k$-means objective ($\mathcal{C}$) in hindsight. We show that provided the data lies in a bounded region, an implementation of the Multiplicative Weights Update Algorithm (MWUA) using a discretized grid achieves a regret bound of $\tilde{O}(\sqrt{T})$ in expectation. We also present an online-to-offline reduction that shows that an efficient no-regret online algorithm (despite being allowed to choose a different set of candidate centres at each round) implies an offline efficient algorithm for the $k$-means problem. In light of this hardness, we consider the slightly weaker requirement of comparing regret with respect to $(1 + \epsilon) \mathcal{C}$ and present a no-regret algorithm with runtime $O\left(T(\mathrm{poly}(log(T),k,d,1/\epsilon)^{k(d+O(1))}\right)$. Our algorithm is based on maintaining an incremental coreset and an adaptive variant of the MWUA. We show that na\"{i}ve online algorithms, such as \emph{Follow The Leader}, fail to produce sublinear regret in the worst case. We also report preliminary experiments with synthetic and real-world data. | ['Vincent Cohen-Addad', 'Benjamin Guedj', 'Guy Rom', 'Varun Kanade'] | 2019-09-15 | null | null | null | null | ['online-clustering'] | ['computer-vision'] | [-1.98778361e-01 2.69129962e-01 1.79958418e-01 -4.61213201e-01
-1.16154063e+00 -8.97411108e-01 -4.87457782e-01 5.57136714e-01
-9.10611629e-01 4.62141484e-01 -5.46019495e-01 -4.51167911e-01
-8.20605934e-01 -8.56704533e-01 -1.03425241e+00 -9.15364444e-01
-5.96118212e-01 9.88148034e-01 6.31013960e-02 9.68084931e-02
4.74165916e-01 5.12340009e-01 -1.39387667e+00 -3.24755698e-01
7.59567142e-01 1.37662828e+00 1.18579529e-02 9.31681931e-01
-1.50808990e-01 1.46658897e-01 -3.22780132e-01 -6.81195319e-01
8.78060758e-01 -5.32656372e-01 -1.15956557e+00 3.62183332e-01
-1.99448280e-02 7.51803443e-02 -1.29389808e-01 1.08441043e+00
5.57684004e-01 4.39675570e-01 3.40517312e-01 -1.36091149e+00
-4.36806381e-02 9.79966640e-01 -1.17355204e+00 8.20010528e-02
1.16950378e-01 -1.94908872e-01 9.91497874e-01 -3.09318870e-01
5.41440308e-01 7.25675762e-01 6.46007299e-01 2.24456161e-01
-1.36030233e+00 -6.51175857e-01 5.09539425e-01 -1.22201942e-01
-1.70986116e+00 -1.92715630e-01 4.47653502e-01 6.97935894e-02
7.49003887e-01 6.07378125e-01 3.87561172e-01 -3.38951886e-01
-2.76484907e-01 5.47089398e-01 9.82539952e-01 -6.28665745e-01
7.89059997e-01 1.42426759e-01 1.02803960e-01 6.27150059e-01
3.59405279e-01 -1.72836885e-01 -2.97919840e-01 -3.55041891e-01
2.24005669e-01 9.33417156e-02 -6.76566809e-02 -4.80947137e-01
-6.93344712e-01 7.95465350e-01 4.89479840e-01 2.22092211e-01
1.91386305e-02 6.62254333e-01 1.95515588e-01 5.46967864e-01
3.85220826e-01 9.94166583e-02 -8.90775323e-01 -4.04199570e-01
-1.07303882e+00 1.09444000e-01 9.54597831e-01 1.22870409e+00
9.29435372e-01 -5.02952695e-01 4.24234658e-01 5.88175774e-01
-2.66890556e-01 7.63769507e-01 3.39369737e-02 -1.42872524e+00
6.68248296e-01 4.77976531e-01 5.93236327e-01 -9.70757246e-01
-4.21529084e-01 -2.72719294e-01 -7.86612451e-01 2.83055931e-01
6.40399456e-01 -4.42783117e-01 -3.65258545e-01 1.84090447e+00
5.97783744e-01 -2.24778026e-01 -4.20286179e-01 6.18206799e-01
-3.98818672e-01 6.08570158e-01 -5.24464250e-01 -7.76059687e-01
8.72101068e-01 -5.53214729e-01 -2.04497591e-01 2.62543336e-02
9.90649462e-01 -6.61956370e-01 7.16363370e-01 6.27727747e-01
-1.70180595e+00 6.30154088e-02 -6.54370725e-01 2.66639948e-01
-4.02463585e-01 -3.45542133e-01 5.04355609e-01 1.01353014e+00
-1.30614591e+00 8.03548098e-01 -8.51971090e-01 -2.42650509e-01
1.45010069e-01 7.13291287e-01 -1.22874312e-01 -1.82092488e-01
-2.72396713e-01 2.42957264e-01 6.59446940e-02 -4.06723516e-03
-4.46123630e-01 -5.19311905e-01 -4.08580989e-01 7.64445066e-02
6.49914622e-01 -3.59963745e-01 1.01652312e+00 -9.44345236e-01
-1.03179562e+00 8.50518227e-01 -2.97378927e-01 -4.30493981e-01
7.82766759e-01 2.85004228e-01 2.30544344e-01 1.52082622e-01
1.07837811e-01 2.20611736e-01 3.51797253e-01 -1.16578043e+00
-1.21722436e+00 -7.54326105e-01 1.09089054e-01 1.53070077e-01
-3.76983970e-01 1.38928741e-01 -6.96327925e-01 -1.50189728e-01
4.64741975e-01 -1.12821555e+00 -7.40598619e-01 -6.74562436e-03
-3.37183699e-02 -1.61910966e-01 1.79496661e-01 -1.54564679e-01
1.27784777e+00 -2.19320869e+00 -7.26405755e-02 9.73888814e-01
1.95267096e-01 -3.12611341e-01 2.05218673e-01 5.49197018e-01
3.07426322e-02 4.35096711e-01 -3.48416477e-01 -4.75045532e-01
3.00203025e-01 1.93770036e-01 7.14669898e-02 9.29184914e-01
-1.02815306e+00 1.21850960e-01 -7.25947320e-01 -2.97912776e-01
-4.95179929e-02 -2.19195038e-01 -7.50167966e-01 -2.38762200e-01
-1.46944216e-02 -4.30552751e-01 -2.34917089e-01 1.64826795e-01
1.13350308e+00 -4.22643840e-01 4.03558373e-01 4.01121765e-01
-2.47769639e-01 -3.81447762e-01 -2.03499460e+00 1.58830261e+00
-4.57900763e-01 7.81977251e-02 8.66746426e-01 -1.06068587e+00
3.40413183e-01 2.31538657e-02 9.94282842e-01 -4.41188961e-01
3.64699513e-01 4.09129053e-01 -4.88175631e-01 3.81729230e-02
3.14095467e-01 -1.96616113e-01 -3.55718374e-01 1.00910282e+00
-4.49582785e-01 1.09515384e-01 1.63482457e-01 4.51780230e-01
1.49877512e+00 -5.77021718e-01 -3.42227310e-01 -3.02415699e-01
1.46750652e-03 5.71690127e-02 7.64555812e-01 1.03857076e+00
-2.99045414e-01 2.60424256e-01 5.57218969e-01 -3.74602735e-01
-1.03024018e+00 -8.39689255e-01 1.74344212e-01 1.43335462e+00
3.95213723e-01 -2.37174973e-01 -9.57097411e-01 -5.27456701e-01
2.06157714e-01 5.58791816e-01 -8.88672233e-01 2.95592189e-01
-3.13009381e-01 -7.70047247e-01 6.94412813e-02 3.78429860e-01
1.92549869e-01 -4.85004067e-01 -7.48378634e-01 2.87236094e-01
1.67834312e-02 -6.02716863e-01 -9.47495937e-01 6.04916275e-01
-9.64465380e-01 -1.15314639e+00 -3.48669618e-01 -7.52830982e-01
1.32107878e+00 5.86376667e-01 7.34103560e-01 4.99762557e-02
-4.19207424e-01 6.34354591e-01 -5.76697826e-01 -3.50154400e-01
3.56113017e-01 2.07328163e-02 7.87606090e-02 -2.64837950e-01
3.77340198e-01 -5.04353106e-01 -9.74505126e-01 3.46587449e-01
-7.72541106e-01 -4.74695265e-01 2.08602652e-01 4.88072187e-01
9.72762167e-01 6.90864027e-01 1.82735726e-01 -1.07192755e+00
3.47342104e-01 -2.60690391e-01 -1.24773109e+00 2.31802434e-01
-9.25183296e-01 4.46094871e-02 9.40493047e-01 -2.11561937e-02
-5.56387067e-01 4.19770956e-01 2.21701950e-01 -3.47945869e-01
1.40177041e-01 1.87815085e-01 1.27107024e-01 -2.27820456e-01
5.25440753e-01 2.24198222e-01 -5.33864647e-02 -3.63604963e-01
6.09164238e-01 5.84548652e-01 2.89671361e-01 -8.07939231e-01
9.05144274e-01 8.63530517e-01 1.20196119e-01 -2.51038790e-01
-7.45020390e-01 -7.40601897e-01 -3.62760782e-01 4.89797667e-02
3.24653387e-01 -5.44213653e-01 -1.72698855e+00 -1.31721959e-01
-5.63715696e-01 -5.16682029e-01 -5.86072862e-01 -2.00953218e-03
-9.27251935e-01 4.64424580e-01 -4.27178919e-01 -1.41599035e+00
-3.41918766e-01 -6.98434293e-01 4.86629963e-01 6.74534738e-02
3.07033330e-01 -6.17318571e-01 -2.21729785e-01 4.00648147e-01
2.54035354e-01 2.91777104e-01 5.34557641e-01 -5.84608555e-01
-4.95334804e-01 -4.05497611e-01 -1.60375342e-01 -3.84917646e-03
-1.31320179e-01 -2.79788494e-01 -3.46775264e-01 -8.24716687e-01
-1.17179705e-02 -1.78405881e-01 4.53889519e-01 4.70128775e-01
1.70570588e+00 -8.44606459e-01 -5.12876213e-01 6.36959374e-01
1.90398121e+00 3.27338457e-01 1.75954059e-01 2.13888720e-01
3.21895233e-03 2.41374627e-01 7.74120212e-01 1.08515120e+00
3.71272326e-01 2.03084439e-01 4.34541911e-01 2.47480068e-02
7.82518625e-01 2.76050359e-01 1.01332821e-01 6.47046208e-01
7.25378171e-02 -1.19070746e-01 -9.05535698e-01 9.78460550e-01
-1.97491169e+00 -8.12424004e-01 7.22705722e-02 2.80167198e+00
8.55203092e-01 3.31941873e-01 3.54377896e-01 3.20539862e-01
7.11899757e-01 -4.50515300e-01 -6.36451244e-01 -8.73470962e-01
1.56785846e-01 6.21275961e-01 1.29718101e+00 6.60276592e-01
-7.17349410e-01 5.72126746e-01 5.55478573e+00 1.04691935e+00
-5.99471986e-01 8.93180668e-02 9.02243793e-01 -7.33799338e-01
-1.54900908e-01 1.23637021e-01 -4.41572428e-01 6.30141556e-01
1.16734040e+00 -5.21199465e-01 1.09331191e+00 1.18188846e+00
2.64447898e-01 -5.32410085e-01 -1.21136987e+00 1.15699911e+00
-1.49670869e-01 -1.23735154e+00 -6.48643911e-01 3.52384269e-01
8.92932296e-01 -8.96099135e-02 -8.19869898e-03 -8.35322663e-02
1.04331589e+00 -7.41395354e-01 5.36094308e-01 -3.32561210e-02
8.18062365e-01 -1.59765458e+00 6.20344877e-01 6.17274165e-01
-1.47725570e+00 -3.95687342e-01 -4.13153708e-01 7.04121739e-02
5.28594851e-02 7.11717904e-01 -5.73219478e-01 7.23181129e-01
1.36704481e+00 -3.25013578e-01 2.00507179e-01 1.10377431e+00
5.65643489e-01 1.03226215e-01 -1.04483628e+00 -1.50417238e-01
3.51450235e-01 -4.47733313e-01 6.34995103e-02 1.12246716e+00
3.82203519e-01 8.28965187e-01 3.67228746e-01 2.72246122e-01
-2.63933718e-01 2.64301717e-01 -9.31829810e-02 3.14848870e-01
7.87651896e-01 1.04676044e+00 -1.21967506e+00 -2.67204791e-01
7.54254833e-02 1.14951396e+00 4.58845586e-01 3.03313434e-02
-8.32923353e-01 -1.03637147e+00 6.01920366e-01 1.99661508e-01
7.44699657e-01 -3.26436818e-01 -4.76835161e-01 -4.46596026e-01
1.76868573e-01 -3.20437640e-01 9.14757013e-01 -2.97109067e-01
-1.34025717e+00 1.65091112e-01 -2.30520517e-01 -7.58783400e-01
2.04989195e-01 -2.97946036e-01 -2.48370931e-01 6.61098480e-01
-1.02013814e+00 -3.12592149e-01 -4.44447212e-02 8.56032014e-01
7.34376907e-02 4.83219236e-01 6.03966415e-01 2.54149795e-01
-2.97772318e-01 1.02384377e+00 8.17338407e-01 -1.15190804e-01
5.41204333e-01 -1.53171170e+00 -3.63595903e-01 8.05073678e-01
-2.86349475e-01 4.60903913e-01 8.82638693e-01 -1.97869778e-01
-1.55002892e+00 -1.01688313e+00 8.68654668e-01 -2.28574499e-01
4.77950007e-01 -2.69158095e-01 -1.01356082e-01 7.13166177e-01
-7.27677420e-02 1.74205139e-01 1.06212819e+00 1.86120510e-01
-1.02387168e-01 -7.84818232e-01 -1.68102896e+00 5.11528432e-01
1.13865030e+00 -1.33354351e-01 1.31099954e-01 4.96339798e-01
5.60232460e-01 -3.22131366e-01 -9.76573169e-01 2.95419972e-02
2.51631141e-01 -9.98970568e-01 5.25772989e-01 -3.38080078e-01
-3.79911751e-01 -3.51568490e-01 -1.76548705e-01 -8.87642741e-01
-2.55934715e-01 -1.11221516e+00 1.03568934e-01 8.19441736e-01
5.66056192e-01 -5.59185028e-01 1.21983361e+00 1.23315108e+00
2.23696753e-01 -8.64689112e-01 -1.42766285e+00 -7.76396215e-01
3.90734375e-01 -5.47337592e-01 5.08923709e-01 9.14829910e-01
3.94544721e-01 -4.09386992e-01 -1.26076732e-02 2.77475059e-01
8.43478262e-01 4.79212970e-01 8.22378159e-01 -9.30104434e-01
-5.87682724e-01 -4.19619352e-01 1.18092716e-01 -1.02982450e+00
-4.50608432e-01 -8.27320635e-01 1.32580608e-01 -1.46019042e+00
3.25691909e-01 -1.23662984e+00 -5.03399312e-01 6.27241611e-01
2.36891061e-01 -2.00665593e-01 2.13066369e-01 -2.07639500e-01
-1.28234410e+00 1.59107536e-01 5.10400653e-01 2.27738425e-01
-3.12126189e-01 1.19376451e-01 -7.96255231e-01 4.93937314e-01
6.53865993e-01 -7.98759699e-01 -1.92195609e-01 -5.85397243e-01
5.30234694e-01 5.56256652e-01 -4.44912791e-01 -6.98197961e-01
7.30210900e-01 -3.73553246e-01 1.66405886e-01 -6.50742233e-01
9.06580687e-02 -1.23720121e+00 2.87064373e-01 6.87492013e-01
-2.56537050e-01 3.29093695e-01 1.13899261e-02 8.57557416e-01
4.19932514e-01 -2.65331715e-01 8.77295732e-01 -3.23418975e-01
1.38820663e-01 6.17462754e-01 -1.76498115e-01 7.77867585e-02
1.50134575e+00 -5.49883306e-01 -2.90025100e-02 -3.78831536e-01
-8.65380168e-01 8.03842068e-01 7.15606451e-01 -4.92269903e-01
1.63330123e-01 -7.68747747e-01 -2.52587318e-01 -3.53581786e-01
-1.96279764e-01 3.88137013e-01 5.92682481e-01 7.66202450e-01
-6.34561002e-01 2.93302536e-01 3.72289717e-01 -2.83793330e-01
-1.15375054e+00 1.01452911e+00 2.98030168e-01 -2.75492162e-01
-2.32739776e-01 1.25907147e+00 -4.80696440e-01 -2.76081681e-01
5.41251421e-01 -4.51330002e-03 9.42929327e-01 -9.27354582e-03
3.01920950e-01 9.08962309e-01 1.10136576e-01 5.73702939e-02
-4.30653214e-01 3.58382225e-01 -2.19493940e-01 -5.12880981e-01
1.64585412e+00 -5.10293961e-01 -3.30154926e-01 5.02376705e-02
1.20692265e+00 1.63672805e-01 -1.24409235e+00 -1.57042906e-01
1.14703208e-01 -6.83987737e-01 -2.88003325e-01 -6.99409187e-01
-1.37621367e+00 3.77658099e-01 7.74503767e-01 4.94724274e-01
1.46591008e+00 -3.10612042e-02 7.32458055e-01 5.37561834e-01
1.02542055e+00 -1.95838964e+00 -2.82470942e-01 5.95078282e-02
1.65191248e-01 -8.69181454e-01 9.68786608e-03 -1.61926582e-01
-2.68626094e-01 9.86975968e-01 3.52790385e-01 -3.52175266e-01
1.04683912e+00 2.59110123e-01 -2.88354188e-01 -3.65289897e-02
-7.25829720e-01 -1.15299068e-01 -7.44893491e-01 -2.16454584e-02
2.13948563e-02 2.90396035e-01 -6.30542636e-01 6.89923763e-01
-3.71867299e-01 -8.22815523e-02 5.98079979e-01 1.34893346e+00
-8.16958785e-01 -1.18052793e+00 -3.87461573e-01 5.01139879e-01
-6.83031321e-01 2.72880048e-01 -2.21156150e-01 3.88373971e-01
2.27010727e-01 1.35285401e+00 2.32475132e-01 -1.59705624e-01
1.09335244e-01 -1.32075697e-01 3.59786481e-01 -1.93028241e-01
-7.74118364e-01 2.60784835e-01 -2.52810746e-01 -6.99181974e-01
-1.64913535e-01 -6.15657032e-01 -1.68857872e+00 -1.10547471e+00
-4.09888029e-01 7.79590428e-01 6.89618707e-01 5.68646073e-01
4.91909504e-01 -7.16191381e-02 1.34707463e+00 -3.64196897e-01
-4.90795583e-01 -3.35816711e-01 -1.05868602e+00 2.67302603e-01
-1.75952241e-01 -3.86565402e-02 -7.97601461e-01 -1.35108083e-01] | [6.628669261932373, 4.837262153625488] |
c6ae49ac-0640-41ce-924e-a48f9e19f9d3 | the-midi-degradation-toolkit-symbolic-music | 2010.00059 | null | https://arxiv.org/abs/2010.00059v1 | https://arxiv.org/pdf/2010.00059v1.pdf | The MIDI Degradation Toolkit: Symbolic Music Augmentation and Correction | In this paper, we introduce the MIDI Degradation Toolkit (MDTK), containing functions which take as input a musical excerpt (a set of notes with pitch, onset time, and duration), and return a "degraded" version of that excerpt with some error (or errors) introduced. Using the toolkit, we create the Altered and Corrupted MIDI Excerpts dataset version 1.0 (ACME v1.0), and propose four tasks of increasing difficulty to detect, classify, locate, and correct the degradations. We hypothesize that models trained for these tasks can be useful in (for example) improving automatic music transcription performance if applied as a post-processing step. To that end, MDTK includes a script that measures the distribution of different types of errors in a transcription, and creates a degraded dataset with similar properties. MDTK's degradations can also be applied dynamically to a dataset during training (with or without the above script), generating novel degraded excerpts each epoch. MDTK could also be used to test the robustness of any system designed to take MIDI (or similar) data as input (e.g. systems designed for voice separation, metrical alignment, or chord detection) to such transcription errors or otherwise noisy data. The toolkit and dataset are both publicly available online, and we encourage contribution and feedback from the community. | ['Kazuyoshi Yoshii', 'James Owers', 'Andrew McLeod'] | 2020-09-30 | null | null | null | null | ['music-transcription'] | ['music'] | [ 5.51245511e-01 -1.33823290e-01 3.41329247e-01 -1.46853015e-01
-1.18171251e+00 -1.16981006e+00 3.12233537e-01 1.87578909e-02
-1.14609212e-01 4.25289333e-01 3.96058738e-01 -1.56842172e-01
-1.61038876e-01 -2.51618892e-01 -4.89789963e-01 -5.36869586e-01
-8.54362845e-02 3.91011387e-01 2.56786913e-01 -1.05686218e-01
3.91222566e-01 3.54358584e-01 -1.63453591e+00 5.55842519e-01
7.35420167e-01 8.03803861e-01 1.55484319e-01 1.19564128e+00
3.05808157e-01 4.84885454e-01 -1.13288677e+00 -1.84388146e-01
2.44948685e-01 -8.50368619e-01 -8.40584934e-01 -1.63266063e-01
4.67182010e-01 6.76172376e-02 4.11359891e-02 9.07743931e-01
8.83388519e-01 2.70854861e-01 5.44940591e-01 -9.77103055e-01
-1.70785397e-01 9.27343726e-01 1.09782197e-01 1.02315836e-01
7.72347212e-01 2.29537040e-01 7.63779402e-01 -8.70352805e-01
5.21007180e-01 8.77934337e-01 7.86092699e-01 5.25012910e-01
-1.42044210e+00 -5.21640301e-01 -4.30762917e-01 1.28763407e-01
-1.32901943e+00 -8.81701767e-01 6.37202978e-01 -4.52679098e-01
1.02741647e+00 8.21146131e-01 4.63895500e-01 1.22438622e+00
-1.30858913e-01 5.67801952e-01 7.79965341e-01 -6.41192198e-01
2.92893201e-01 -1.80973426e-01 -1.34491697e-01 1.04579978e-01
-4.31950539e-01 1.21333450e-01 -9.21453893e-01 -2.73030341e-01
3.38223487e-01 -7.77544260e-01 -5.25686502e-01 9.37661454e-02
-1.34458470e+00 1.11328363e-01 -2.07936391e-01 2.76731879e-01
-3.11942399e-01 -7.16765150e-02 6.82074904e-01 8.54194224e-01
2.21622437e-01 1.06566513e+00 -5.28187394e-01 -9.36203301e-01
-1.22670507e+00 5.30864954e-01 8.23258102e-01 6.15671813e-01
3.20641279e-01 1.55003116e-01 -3.92512918e-01 1.17466354e+00
-2.59539574e-01 2.08359122e-01 9.19999778e-01 -1.31369901e+00
3.41029227e-01 4.33234237e-02 7.57415816e-02 -5.66138089e-01
-4.56742316e-01 -3.52179170e-01 -4.36200500e-01 2.03405365e-01
4.67319041e-01 -5.61640784e-02 -7.68575788e-01 1.74309981e+00
3.62752378e-02 1.87070251e-01 8.29371214e-02 7.12750077e-01
4.54536289e-01 5.09179652e-01 -5.93998373e-01 -4.92102772e-01
8.25866461e-01 -7.69470513e-01 -6.84257209e-01 -5.37943058e-02
6.35801256e-01 -1.41422272e+00 1.56068039e+00 1.11596763e+00
-1.29167378e+00 -8.05512965e-01 -1.02652717e+00 1.05637729e-01
7.57979527e-02 9.45311412e-02 -2.87942756e-02 7.49240398e-01
-9.66937244e-01 1.22436297e+00 -5.79880953e-01 -2.40292653e-01
-3.62444609e-01 2.97171950e-01 -1.57804992e-02 3.36778522e-01
-1.04937100e+00 6.05896235e-01 2.77554661e-01 -9.86052975e-02
-7.81498551e-01 -6.97825193e-01 -6.39054418e-01 -1.11831434e-01
8.94618779e-02 -3.06597650e-01 1.74835598e+00 -1.08496451e+00
-1.82570004e+00 6.92266345e-01 1.03008598e-01 -2.12027699e-01
4.52895343e-01 -3.05416524e-01 -6.17139697e-01 -1.67012870e-01
-1.40327692e-01 2.34067008e-01 1.12174535e+00 -9.05653954e-01
-3.76211435e-01 -2.74278343e-01 -3.94300640e-01 2.34254017e-01
-1.66737303e-01 2.51705915e-01 -3.14884335e-01 -1.08983254e+00
1.34949893e-01 -1.07991660e+00 3.53156775e-01 -6.48549199e-01
-5.93819857e-01 1.06373824e-01 6.20898306e-01 -1.12361705e+00
1.65471900e+00 -2.45755363e+00 4.02321309e-01 3.40815276e-01
-4.37716722e-01 3.80276948e-01 -4.47165132e-01 4.82492447e-01
-2.44907454e-01 5.56392409e-02 -3.52345705e-01 -3.73153389e-01
-6.45271912e-02 -6.15153685e-02 -3.78769219e-01 1.99482501e-01
-4.33671959e-02 4.41994727e-01 -9.42002416e-01 -2.71113850e-02
-8.35467428e-02 8.14107656e-02 -5.39529920e-01 4.08966810e-01
-2.35907316e-01 5.47569871e-01 5.00379086e-01 4.82613474e-01
1.18317544e-01 7.89423168e-01 1.40770644e-01 -1.41184494e-01
-2.83240795e-01 8.91082466e-01 -1.48465264e+00 1.73846245e+00
-4.61292773e-01 8.38492215e-01 5.09097055e-03 -4.91382897e-01
1.09422266e+00 5.81492305e-01 2.02171817e-01 -2.84261405e-01
-1.78955980e-02 4.33112562e-01 4.22615379e-01 -5.68716109e-01
7.22257137e-01 1.43404277e-02 -1.45740449e-01 6.02712095e-01
7.34440759e-02 -6.06099248e-01 3.49634171e-01 -1.22155510e-01
1.34297001e+00 2.31438518e-01 1.08583793e-01 1.70512408e-01
2.59663045e-01 -1.11593679e-01 5.51007092e-01 7.95688987e-01
-6.73893914e-02 1.12685978e+00 4.48860049e-01 5.11630103e-02
-1.26635838e+00 -1.07802975e+00 8.65290835e-02 1.18404949e+00
-5.40610313e-01 -8.11922252e-01 -8.92637670e-01 -1.77094370e-01
-1.01703241e-01 8.03515911e-01 -3.08854371e-01 -4.58344966e-01
-6.72047853e-01 -2.49277204e-01 1.20413947e+00 3.90035629e-01
-5.92786036e-02 -1.43953252e+00 -2.94866621e-01 3.99199694e-01
-4.41936344e-01 -4.98208284e-01 -8.45296264e-01 5.04519522e-01
-7.73861110e-01 -8.51084590e-01 -4.46274042e-01 -7.52994955e-01
-8.91428664e-02 -8.45118761e-02 9.42657232e-01 -7.91909397e-02
-1.17739417e-01 2.72144914e-01 -7.01779127e-01 -3.59782726e-01
-9.71957147e-01 2.92563796e-01 4.91108030e-01 -6.87792897e-02
-1.68985650e-01 -8.62792671e-01 -2.16939732e-01 3.48887652e-01
-1.00943899e+00 -1.21394649e-01 1.09080486e-01 6.78196490e-01
5.34829319e-01 -1.14464760e-02 6.88629389e-01 -7.07450151e-01
1.12699664e+00 -1.53741226e-01 -1.62379295e-01 6.67893216e-02
-5.66324353e-01 -5.32717630e-02 8.71086419e-01 -7.49721766e-01
-7.46594667e-01 2.27724500e-02 -3.99581403e-01 -5.71849525e-01
-1.69913545e-01 5.72671711e-01 -4.33720559e-01 3.10653597e-01
1.07750940e+00 1.82582751e-01 -1.97204769e-01 -9.75266933e-01
3.32164794e-01 1.19493985e+00 1.19011521e+00 -5.69573700e-01
6.42849684e-01 -3.00676167e-01 -4.55664933e-01 -7.80698538e-01
-5.44950187e-01 -4.61270273e-01 -8.00668657e-01 -1.83687806e-01
1.68375388e-01 -4.63554442e-01 -2.45592237e-01 7.70920992e-01
-1.03409111e+00 -7.19347000e-01 -5.44260323e-01 3.54613334e-01
-9.64256108e-01 3.96519631e-01 -5.34603179e-01 -7.52418578e-01
-3.18394721e-01 -7.57435441e-01 7.87102640e-01 1.02715239e-01
-9.63108778e-01 -4.64063048e-01 4.77053821e-01 1.85537562e-01
1.83870405e-01 7.78919086e-02 8.94690633e-01 -6.63583875e-01
9.60548297e-02 -1.06242925e-01 5.29328465e-01 7.57712960e-01
3.19961756e-01 4.42205727e-01 -1.29653370e+00 -2.17508525e-01
-2.13378039e-03 -3.19791853e-01 6.14630759e-01 1.72030970e-01
1.07265830e+00 -4.42042321e-01 2.18908921e-01 8.03116083e-01
7.17863321e-01 4.58584607e-01 8.15461695e-01 3.83444875e-01
4.12868649e-01 5.14727533e-01 8.17231596e-01 4.90101039e-01
-2.33947918e-01 1.03333056e+00 7.98512995e-02 2.40604535e-01
-3.94093037e-01 -2.33172402e-01 8.21934044e-01 1.30614400e+00
-1.28945023e-01 -1.59136951e-01 -7.33621538e-01 6.96807623e-01
-1.57452261e+00 -1.16541350e+00 -1.25227407e-01 2.62604666e+00
1.36256957e+00 1.87012270e-01 4.66298521e-01 8.70081246e-01
6.51532292e-01 4.72509861e-02 -6.09334648e-01 -1.00025296e+00
-1.23347692e-01 6.21244073e-01 1.48957567e-02 5.02528191e-01
-7.75883436e-01 8.51717591e-01 6.55714369e+00 6.77741170e-01
-1.28310978e+00 -2.04847559e-01 4.67186198e-02 -3.02622437e-01
-2.42244646e-01 3.72493416e-02 -2.52608448e-01 4.33037251e-01
1.24000871e+00 -2.84495741e-01 1.12828791e+00 5.00031829e-01
5.22102714e-01 1.36959215e-03 -1.26002860e+00 9.46113110e-01
4.38312627e-02 -9.19982910e-01 -2.40030304e-01 -2.92615652e-01
5.69875717e-01 -1.89006060e-01 2.27071419e-02 4.10262048e-01
-8.04205835e-02 -7.97270417e-01 1.10741389e+00 6.25575423e-01
1.13039184e+00 -8.06102693e-01 4.57064271e-01 2.71259129e-01
-1.01799107e+00 4.94493656e-02 -2.91911531e-02 -2.54606992e-01
-5.76825440e-02 5.01363277e-01 -1.15136635e+00 3.97153586e-01
6.56418741e-01 5.21261334e-01 -6.93804145e-01 1.30069375e+00
-3.54634255e-01 1.17891932e+00 -2.13970259e-01 3.78589541e-01
-4.08005774e-01 9.29044839e-03 8.91429901e-01 1.44370675e+00
5.65326691e-01 -3.08238953e-01 -1.15888990e-01 4.89023298e-01
-7.68354461e-02 2.12016508e-01 -3.93488973e-01 -1.39960766e-01
8.55473757e-01 1.03929031e+00 -4.30736095e-01 -3.38609457e-01
2.36591935e-01 1.47916758e+00 1.10702766e-02 3.40572774e-01
-5.87233007e-01 -8.72223973e-01 9.39232767e-01 2.36691423e-02
5.79566061e-02 -7.43784532e-02 -3.61926228e-01 -9.02869463e-01
9.02485698e-02 -1.38872397e+00 2.93974131e-01 -1.07729220e+00
-9.65292394e-01 7.01789379e-01 -2.27210701e-01 -1.30014884e+00
-8.37373853e-01 -2.87181079e-01 -8.43036950e-01 7.96352088e-01
-6.70025289e-01 -4.67912823e-01 -1.81724325e-01 3.18715781e-01
6.71759725e-01 -4.22605239e-02 1.07084882e+00 1.87925190e-01
-3.83065999e-01 6.63377881e-01 2.69344091e-01 -1.99824199e-03
1.37415195e+00 -1.54814637e+00 6.93944931e-01 7.80903518e-01
5.53929865e-01 4.60044265e-01 9.73723292e-01 -5.65684676e-01
-8.06340277e-01 -1.07004976e+00 9.00825143e-01 -4.64792252e-01
4.63086933e-01 -4.20711666e-01 -1.06473184e+00 5.54427564e-01
-3.97121757e-02 -5.44680655e-01 6.40280008e-01 2.67806709e-01
-2.88033843e-01 -8.36605057e-02 -7.88161218e-01 5.92659295e-01
1.16509402e+00 -7.45588958e-01 -9.08628404e-01 -1.79242827e-02
7.50306785e-01 -7.39226162e-01 -8.63573611e-01 2.96565145e-01
7.98168421e-01 -1.21610081e+00 7.20529497e-01 -4.43953753e-01
1.11783639e-01 -4.98932362e-01 -8.23782906e-02 -1.86154842e+00
-3.19303811e-01 -1.18020260e+00 -7.11825714e-02 1.55921459e+00
5.10344267e-01 -8.19649175e-02 2.44770184e-01 1.80100664e-01
-5.94529688e-01 -2.76841134e-01 -7.64494121e-01 -1.10122573e+00
-1.81277633e-01 -9.61233795e-01 5.44359863e-01 8.54003608e-01
2.85189539e-01 2.24615097e-01 -4.50183660e-01 -2.18094457e-02
-1.30624594e-02 8.71906336e-03 1.02954578e+00 -1.10057068e+00
-5.92212379e-01 -4.38232571e-01 -2.19414651e-01 -5.03237784e-01
-2.67956853e-01 -8.78032327e-01 3.21322918e-01 -1.04268503e+00
-3.43545556e-01 -4.03295338e-01 -2.71157950e-01 6.32520318e-01
-1.97169751e-01 4.40832704e-01 4.80265200e-01 4.43224251e-01
-1.56073853e-01 3.49885881e-01 7.05156088e-01 -2.92057008e-03
-7.95590281e-01 3.32147777e-01 -4.40224081e-01 6.63063407e-01
9.52724457e-01 -6.83372676e-01 -1.98620901e-01 -1.69244364e-01
2.21542522e-01 5.24948798e-02 1.72934309e-01 -1.47561276e+00
-1.14842243e-01 1.24664374e-01 1.36696503e-01 -3.87006074e-01
3.01794857e-01 -1.20571800e-01 4.13621753e-01 7.50728101e-02
-5.86101472e-01 2.15770528e-01 3.50140244e-01 2.90222204e-04
-3.40589225e-01 -5.71286440e-01 6.36076868e-01 7.31034204e-02
-1.87986583e-01 -3.68744671e-01 -5.98414183e-01 1.21273309e-01
3.60886186e-01 -3.59181970e-01 9.80983078e-02 -5.62398732e-01
-9.53365326e-01 -3.47675413e-01 7.56386817e-01 4.29691285e-01
3.50336432e-01 -1.22243500e+00 -7.46771395e-01 3.43214482e-01
1.35073319e-01 -2.63161331e-01 2.55943020e-03 7.12008953e-01
-3.42886209e-01 -3.59911472e-02 -1.31844565e-01 -3.37827414e-01
-1.54034483e+00 3.06182563e-01 3.77615541e-01 -1.89039055e-02
-3.32506180e-01 7.21100152e-01 -4.38526899e-01 -5.45737922e-01
4.84676927e-01 -5.63324988e-01 2.06485633e-02 1.76830947e-01
6.52263880e-01 4.82543737e-01 5.56896508e-01 -2.93539971e-01
-1.83087066e-01 3.51260573e-01 2.43541181e-01 -6.25380516e-01
1.08884180e+00 -7.65468553e-02 1.04785539e-01 1.03018212e+00
8.32782507e-01 6.15616739e-01 -9.89272833e-01 3.21213752e-02
2.81456470e-01 -4.56068546e-01 -2.60437757e-01 -1.05452120e+00
-4.78204161e-01 6.18642688e-01 5.26296198e-01 3.95519137e-01
1.39866507e+00 -2.28616625e-01 7.48898327e-01 1.30040169e-01
4.69503067e-02 -1.13188255e+00 7.34825507e-02 7.94945478e-01
1.16587400e+00 -5.07803619e-01 -3.73663992e-01 -1.82767045e-02
-6.94324017e-01 1.23769081e+00 3.54256719e-01 1.11813456e-01
1.61593914e-01 4.10860360e-01 4.53414381e-01 3.13204139e-01
-8.52488160e-01 -1.06757902e-01 4.02232558e-01 8.71798635e-01
6.03248656e-01 6.17641546e-02 -3.29984039e-01 8.03015411e-01
-8.65059733e-01 -3.86167876e-02 7.32368410e-01 6.09991193e-01
-3.65885407e-01 -1.42990220e+00 -8.33848774e-01 3.26092601e-01
-3.01973045e-01 -2.52136141e-01 -8.96921515e-01 2.39594296e-01
4.20845568e-01 1.00685441e+00 2.33240658e-03 -9.16334212e-01
5.65951645e-01 5.28628409e-01 5.07180810e-01 -7.40246713e-01
-9.87628281e-01 2.29937673e-01 3.03937048e-01 -4.82495576e-01
2.23726146e-02 -9.77505088e-01 -1.15571558e+00 1.12548336e-01
-1.62914336e-01 1.58810928e-01 6.97103500e-01 6.57456756e-01
3.19337249e-01 5.94272375e-01 6.89903200e-01 -9.43010449e-01
-5.33778548e-01 -1.44312847e+00 -6.50724530e-01 5.91521323e-01
3.19680810e-01 -1.09257221e-01 -4.77948457e-01 5.21965206e-01] | [15.727798461914062, 5.740323066711426] |
1a67469f-08e3-4370-b8f0-3b168e9170b8 | fast-hybrid-image-retargeting | 2203.13595 | null | https://arxiv.org/abs/2203.13595v1 | https://arxiv.org/pdf/2203.13595v1.pdf | Fast Hybrid Image Retargeting | Image retargeting changes the aspect ratio of images while aiming to preserve content and minimise noticeable distortion. Fast and high-quality methods are particularly relevant at present, due to the large variety of image and display aspect ratios. We propose a retargeting method that quantifies and limits warping distortions with the use of content-aware cropping. The pipeline of the proposed approach consists of the following steps. First, an importance map of a source image is generated using deep semantic segmentation and saliency detection models. Then, a preliminary warping mesh is computed using axis aligned deformations, enhanced with the use of a distortion measure to ensure low warping deformations. Finally, the retargeted image is produced using a content-aware cropping algorithm. In order to evaluate our method, we perform a user study based on the RetargetMe benchmark. Experimental analyses show that our method outperforms recent approaches, while running in a fraction of their execution time. | ['Timothy Smith', 'Oleg Muraveynyk', 'Daniel Valdez-Balderas'] | 2022-03-25 | null | null | null | null | ['image-retargeting'] | ['computer-vision'] | [ 9.03778553e-01 5.45974523e-02 1.37925074e-01 -1.59106150e-01
-6.09500527e-01 -6.90504670e-01 8.15797210e-01 4.33740586e-01
-4.79736239e-01 3.08614373e-01 1.91354588e-01 2.28978302e-02
1.37553280e-02 -7.45325446e-01 -8.02993596e-01 -3.48983914e-01
1.78955987e-01 -2.55034454e-02 9.30139244e-01 -3.73933345e-01
7.64063358e-01 5.74635208e-01 -1.61494267e+00 1.79648995e-01
1.01793242e+00 8.30780089e-01 4.80568737e-01 6.06578887e-01
2.12885007e-01 1.41061634e-01 -4.23223943e-01 -3.36156785e-01
4.24564272e-01 -5.20255148e-01 -6.99033380e-01 3.27212512e-01
4.68330264e-01 -1.91540927e-01 2.19288796e-01 1.27294040e+00
5.23861706e-01 3.32278937e-01 3.61974776e-01 -9.23316479e-01
-2.44995639e-01 5.16208708e-01 -8.12378824e-01 3.74233097e-01
2.80118942e-01 1.89398155e-01 7.42474794e-01 -9.22668040e-01
9.30918157e-01 1.17451239e+00 2.51467526e-01 1.19817093e-01
-1.58218014e+00 -4.13755387e-01 9.83422250e-02 1.43615857e-01
-1.22234130e+00 -2.19228849e-01 1.03317297e+00 -3.67701799e-01
6.48672819e-01 5.60371339e-01 7.97968149e-01 7.82238781e-01
5.08120596e-01 2.61647105e-01 1.27109087e+00 -5.48557341e-01
5.51085472e-01 6.62736408e-03 -4.37812656e-01 3.77603471e-01
1.15346350e-01 -8.22993740e-02 -5.43693602e-01 5.09778820e-02
7.71940768e-01 -3.05115134e-01 -2.22661674e-01 -7.47493327e-01
-1.47254574e+00 4.92760032e-01 5.89089870e-01 3.72644633e-01
-4.30230379e-01 -1.40634198e-02 3.68379533e-01 7.82869989e-04
7.44575799e-01 6.44541800e-01 -4.72353995e-02 -2.98205987e-02
-1.31265843e+00 4.20953840e-01 2.60953873e-01 6.01280332e-01
8.08796823e-01 -1.80718869e-01 -4.11160111e-01 6.14350736e-01
-3.34078521e-02 2.87509412e-01 5.48236489e-01 -8.58167171e-01
2.13093251e-01 7.03516901e-01 3.82977799e-02 -1.49421382e+00
-3.14290345e-01 -2.58648634e-01 -5.92212260e-01 5.53490818e-01
5.05322590e-02 1.18409611e-01 -9.83065665e-01 1.23655462e+00
5.18681586e-01 4.79550920e-02 -2.69376785e-01 1.07884133e+00
2.64435351e-01 5.19792914e-01 1.68525249e-01 -1.49806559e-01
1.48785138e+00 -7.47737527e-01 -6.04294002e-01 -2.81382650e-01
3.43096375e-01 -1.11414397e+00 1.15020669e+00 3.02282304e-01
-1.34252560e+00 -5.37866116e-01 -1.23853731e+00 -1.27693325e-01
-2.97184139e-01 -9.40223597e-03 -4.65545356e-02 6.56618118e-01
-1.17394888e+00 6.96942449e-01 -7.45976031e-01 -4.89699185e-01
2.84716249e-01 1.53523073e-01 -9.80039909e-02 2.48064548e-01
-8.75878572e-01 9.80107367e-01 7.23438323e-01 -3.85030568e-01
-6.74402356e-01 -9.01669979e-01 -1.00376856e+00 9.29202065e-02
3.42694968e-01 -5.26606977e-01 1.03304183e+00 -1.28520989e+00
-1.58372676e+00 8.21486950e-01 2.17439353e-01 -5.30121505e-01
7.25980639e-01 -2.63354391e-01 -1.45914644e-01 3.00494313e-01
-2.89074387e-02 8.89119506e-01 1.26533127e+00 -1.15793431e+00
-6.89851284e-01 -9.65844020e-02 2.26155575e-02 4.60396051e-01
-2.94348359e-01 1.47789165e-01 -7.03895092e-01 -1.04587269e+00
-8.77717435e-02 -9.27004039e-01 -2.90483624e-01 -3.17509361e-02
-3.84373099e-01 5.52126408e-01 9.85821068e-01 -8.07129979e-01
1.30223310e+00 -2.15616965e+00 3.83122861e-01 4.29662287e-01
2.40457356e-01 2.94758618e-01 -1.19399324e-01 1.98772892e-01
-1.87523425e-01 1.54218927e-03 -6.34110808e-01 -2.33707711e-01
-3.23709667e-01 -2.89205581e-01 -2.49925971e-01 3.71228039e-01
3.73896062e-01 8.65428567e-01 -8.08880508e-01 -5.06412864e-01
7.15798616e-01 5.80477238e-01 -6.68392837e-01 9.42067206e-02
-4.60608393e-01 3.93387973e-01 -4.64647897e-02 3.77971649e-01
7.71074891e-01 1.23913616e-01 -3.43964109e-03 -5.11774480e-01
-4.87562954e-01 -1.95909396e-01 -1.16929638e+00 1.96176422e+00
-3.87099057e-01 7.24180996e-01 -1.70633540e-01 -6.03944540e-01
9.45126355e-01 -2.24769413e-01 3.97482783e-01 -9.50550497e-01
3.49789113e-01 2.74014592e-01 -2.64859825e-01 -2.93715864e-01
1.00254810e+00 3.68940741e-01 6.02054000e-02 3.57967645e-01
-2.34552518e-01 -4.57109302e-01 1.92017391e-01 2.89975971e-01
7.79958963e-01 3.48830491e-01 3.27516824e-01 -5.20870209e-01
4.73484486e-01 7.10358098e-03 8.77114460e-02 1.33869976e-01
3.92516069e-02 1.02534437e+00 3.77328843e-01 -4.40259516e-01
-1.32414055e+00 -7.52836943e-01 1.07908823e-01 9.45491731e-01
5.50302982e-01 -3.58718872e-01 -1.37932646e+00 -4.28643465e-01
-4.20795798e-01 9.40426946e-01 -7.39410579e-01 -2.25797921e-01
-6.33184314e-01 -7.31069684e-01 -1.21182837e-01 1.70855224e-01
5.63822150e-01 -1.04498470e+00 -1.42173183e+00 1.55784041e-01
-1.91969618e-01 -8.68566036e-01 -8.25594783e-01 -2.67833859e-01
-8.69801641e-01 -9.03509021e-01 -1.01962054e+00 -5.38937569e-01
7.74134576e-01 4.17121232e-01 8.21999371e-01 -8.72949585e-02
-4.07619923e-01 4.59054187e-02 -4.67949569e-01 -1.63888186e-01
-4.69392806e-01 2.49303937e-01 -4.52053428e-01 2.66478091e-01
-2.13582531e-01 -5.41492045e-01 -9.18655396e-01 3.41665924e-01
-1.42353189e+00 5.04021764e-01 7.62865722e-01 2.69391209e-01
7.71645010e-01 -7.50902668e-02 1.03407696e-01 -6.32335603e-01
6.04672313e-01 -1.63577706e-01 -8.49993408e-01 1.21694990e-01
-6.45770729e-01 1.86012715e-01 3.68703097e-01 -5.71232796e-01
-1.05972135e+00 1.82624698e-01 1.25825241e-01 -1.42429128e-01
-4.22725035e-03 1.59007296e-01 -2.51080334e-01 -2.96335518e-01
8.58710051e-01 1.38393387e-01 1.00830477e-03 -3.95739555e-01
7.05862999e-01 3.36451709e-01 6.64105713e-01 -1.25760913e-01
9.11273539e-01 5.43125987e-01 -2.79083457e-02 -5.89598000e-01
-2.53962100e-01 -5.57557456e-02 -7.66389072e-01 -5.12497723e-01
7.70231724e-01 -3.89198512e-01 -1.34260252e-01 4.25929427e-01
-1.05731523e+00 -2.14645147e-01 -4.68674392e-01 2.61057436e-01
-6.00703180e-01 3.66127908e-01 -5.99995553e-02 -2.69256830e-01
-4.34963822e-01 -1.40963554e+00 1.03778231e+00 2.91404575e-01
-3.88375700e-01 -5.73216319e-01 1.03294671e-01 9.81186181e-02
6.86492443e-01 6.50662661e-01 8.06201279e-01 -4.32864875e-01
-6.56349361e-01 8.75427425e-02 -3.79070133e-01 6.63787872e-02
9.04686749e-02 1.10142738e-01 -7.56206870e-01 -2.21301630e-01
-2.36559942e-01 3.13609004e-01 6.50667608e-01 4.25385058e-01
9.62342441e-01 -2.63246983e-01 -2.55331159e-01 4.32677031e-01
1.50587320e+00 9.60638151e-02 7.23372459e-01 6.76650286e-01
5.99020958e-01 7.38294423e-01 8.45078945e-01 3.88762236e-01
2.35282302e-01 9.08287764e-01 7.78333783e-01 -2.70681411e-01
-3.90766025e-01 -8.10376108e-02 2.08554283e-01 3.99018854e-01
8.09489489e-02 -5.62635139e-02 -8.07408810e-01 7.98904836e-01
-1.66738164e+00 -4.87929940e-01 -3.36625762e-02 2.52532744e+00
8.00944209e-01 3.34441841e-01 2.44180903e-01 1.73180714e-01
8.56944501e-01 2.53393382e-01 -4.93297607e-01 -5.71384728e-01
8.45415741e-02 1.94009647e-01 6.05245471e-01 7.21626818e-01
-9.91150856e-01 1.01306903e+00 5.44331837e+00 8.81648839e-01
-1.35173142e+00 4.50931713e-02 7.28520751e-01 -1.21288612e-01
-5.17746627e-01 4.57024062e-03 -3.07417870e-01 5.99079132e-01
8.28307152e-01 -3.36508811e-01 2.93155015e-01 6.98406577e-01
3.22985649e-01 -4.70701069e-01 -5.56020021e-01 8.19911122e-01
3.50901753e-01 -1.30642271e+00 1.37729585e-01 -2.58500904e-01
7.29212403e-01 -2.58382589e-01 2.09093839e-01 -3.99737179e-01
-1.08585306e-01 -5.98609924e-01 1.15203547e+00 4.84863311e-01
7.99316466e-01 -1.15129471e+00 3.38976085e-01 -9.95844603e-02
-1.01027203e+00 1.87031537e-01 -1.35813365e-02 3.15506428e-01
2.00494260e-01 7.20035613e-01 -6.91623032e-01 4.74499136e-01
6.90100849e-01 3.96326959e-01 -8.34144533e-01 1.14156520e+00
-1.15469739e-01 2.26294354e-01 -1.61896229e-01 1.71227098e-01
-9.51161832e-02 -1.45956084e-01 8.53601933e-01 1.29595959e+00
4.50182796e-01 -1.53406337e-01 -2.03369766e-01 8.48217607e-01
1.17153890e-01 4.93557990e-01 -3.19542110e-01 3.06418687e-01
2.82178998e-01 1.39264989e+00 -1.28017759e+00 -3.06337059e-01
-1.13742143e-01 1.53929186e+00 -1.73753679e-01 2.01715022e-01
-8.02839100e-01 -4.98289824e-01 3.93426925e-01 4.98236150e-01
2.90889502e-01 -1.76522180e-01 -4.03744817e-01 -7.43041515e-01
1.09207496e-01 -8.12778354e-01 1.64883956e-03 -9.36241388e-01
-3.93038422e-01 8.47610891e-01 2.43872389e-01 -1.25430715e+00
-1.39157668e-01 -1.13675117e-01 -3.69888783e-01 8.65498185e-01
-1.44641542e+00 -1.17131495e+00 -5.50697684e-01 1.89677909e-01
8.79323244e-01 2.22742900e-01 4.48773026e-01 1.34776190e-01
-3.79148185e-01 2.45157585e-01 -2.31977299e-01 -4.40936029e-01
7.79804885e-01 -1.03610957e+00 8.14909339e-01 1.32624090e+00
-1.15083165e-01 2.03967690e-01 1.08934808e+00 -7.40606546e-01
-8.70709658e-01 -1.27598250e+00 6.50810540e-01 -2.49869332e-01
3.17370236e-01 -3.23805690e-01 -1.00228059e+00 2.57121891e-01
4.51011419e-01 -4.45401892e-02 1.27851933e-01 -7.87401199e-01
-1.23648651e-01 -3.09505453e-03 -1.29297197e+00 8.89554858e-01
7.40538180e-01 -1.24663040e-01 -3.89082968e-01 -1.02614231e-01
8.93015802e-01 -7.10613906e-01 -5.58891535e-01 2.98412412e-01
5.10411203e-01 -1.19360399e+00 9.40802395e-01 2.80809611e-01
4.63353127e-01 -5.48945844e-01 2.22210854e-01 -1.54301763e+00
-3.24547350e-01 -8.86000693e-01 1.53479829e-01 1.17156506e+00
1.57172933e-01 -2.86270648e-01 3.87090474e-01 4.79333669e-01
6.22195713e-02 -4.21317160e-01 -7.38317847e-01 -4.16195959e-01
-4.28954631e-01 -2.17160717e-01 5.91112375e-01 7.19050765e-01
-2.30836481e-01 -7.40426034e-02 -1.71379611e-01 7.79296160e-02
6.48675442e-01 9.98460650e-02 5.26665211e-01 -8.29268754e-01
9.08394232e-02 -4.80207026e-01 -5.54208875e-01 -4.08401817e-01
-3.78991246e-01 -5.71483791e-01 1.68783255e-02 -1.22282767e+00
1.46205574e-01 2.41759382e-02 -3.62287052e-02 3.78837049e-01
-2.71335959e-01 7.04099536e-01 3.61872971e-01 1.34734228e-01
-3.07489216e-01 3.24674159e-01 1.15241647e+00 -4.54306193e-02
-4.58517581e-01 -3.91871691e-01 -6.62039757e-01 5.57627857e-01
8.65074933e-01 -4.04067904e-01 -4.84788030e-01 -3.20280820e-01
2.78756201e-01 -5.01334250e-01 5.20759702e-01 -1.20376575e+00
-3.86646278e-02 -1.41723067e-01 1.36709258e-01 -4.95848238e-01
1.06880300e-01 -8.16569865e-01 2.12471053e-01 5.41099906e-01
-3.23835433e-01 3.52393389e-01 4.58287776e-01 3.58253956e-01
2.35319734e-02 -2.60778470e-04 1.16535056e+00 3.60636413e-01
-8.55812788e-01 -1.02829501e-01 -3.68949980e-01 -1.89930752e-01
1.34452343e+00 -3.54517817e-01 1.05281144e-01 -1.18905909e-01
-2.45551065e-01 -3.14588338e-01 9.14949298e-01 7.41134822e-01
5.50536931e-01 -1.09970069e+00 -6.10962152e-01 2.39961490e-01
1.49929956e-01 -6.54087290e-02 4.01255369e-01 7.70873606e-01
-8.84370327e-01 1.00081079e-01 -6.00259602e-01 -5.32932937e-01
-1.38661206e+00 7.53065646e-01 1.29868522e-01 3.39007191e-02
-6.99263215e-01 4.41636056e-01 1.80578455e-01 2.10752830e-01
-1.79965526e-01 -6.13313794e-01 -3.41969788e-01 1.45526305e-01
7.32818425e-01 5.25307417e-01 3.69635642e-01 -9.19149637e-01
-2.85181016e-01 8.14169466e-01 -8.21845233e-02 -5.04019320e-01
1.17380071e+00 -3.61596882e-01 9.28433090e-02 -1.27909720e-01
9.81995165e-01 7.54284337e-02 -1.55424356e+00 -1.52205393e-01
-8.99078399e-02 -8.14657509e-01 1.99506342e-01 -7.47177064e-01
-1.05868292e+00 5.52227914e-01 9.58686531e-01 2.21567228e-02
1.54650819e+00 -2.53421187e-01 8.05953801e-01 -5.18876910e-01
9.65863913e-02 -1.14290357e+00 1.08612150e-01 -6.37544617e-02
1.22449017e+00 -8.82693589e-01 3.24553326e-02 -4.77961302e-01
-6.60807729e-01 8.28547180e-01 4.30959672e-01 -1.72075078e-01
2.85956889e-01 3.87063533e-01 1.23426475e-01 -1.13058044e-02
-2.32068956e-01 5.08991012e-04 4.86323655e-01 5.47809899e-01
1.40827298e-01 -1.74760163e-01 -6.90905690e-01 9.84259769e-02
-2.78300703e-01 -1.25789419e-01 5.40733039e-01 8.07355881e-01
-5.58123887e-01 -1.07601702e+00 -6.08323991e-01 6.90824911e-03
-3.26994091e-01 -3.53165083e-02 -5.01608491e-01 4.41969126e-01
1.31582856e-01 7.34471798e-01 9.57200602e-02 -3.18065166e-01
5.34844816e-01 -4.06210899e-01 3.49794090e-01 -4.62706000e-01
-6.71426773e-01 1.90054700e-01 -3.56868476e-01 -9.66560900e-01
-4.05075759e-01 -5.13607621e-01 -1.12222564e+00 2.43948661e-02
-1.09939016e-01 -2.12619022e-01 9.95279789e-01 7.19496906e-01
7.23374963e-01 6.10946059e-01 5.32938123e-01 -1.32043219e+00
-2.34564003e-02 -6.84160113e-01 -1.99969202e-01 5.81258416e-01
1.89547494e-01 -4.70006853e-01 -1.19616672e-01 3.92303586e-01] | [10.94039535522461, -1.0172843933105469] |
b29e091c-b78b-4677-bfd1-3174099c4404 | learning-to-generate-equitable-text-in | 2307.04303 | null | https://arxiv.org/abs/2307.04303v1 | https://arxiv.org/pdf/2307.04303v1.pdf | Learning to Generate Equitable Text in Dialogue from Biased Training Data | The ingrained principles of fairness in a dialogue system's decision-making process and generated responses are crucial for user engagement, satisfaction, and task achievement. Absence of equitable and inclusive principles can hinder the formation of common ground, which in turn negatively impacts the overall performance of the system. For example, misusing pronouns in a user interaction may cause ambiguity about the intended subject. Yet, there is no comprehensive study of equitable text generation in dialogue. Aptly, in this work, we use theories of computational learning to study this problem. We provide formal definitions of equity in text generation, and further, prove formal connections between learning human-likeness and learning equity: algorithms for improving equity ultimately reduce to algorithms for improving human-likeness (on augmented data). With this insight, we also formulate reasonable conditions under which text generation algorithms can learn to generate equitable text without any modifications to the biased training data on which they learn. To exemplify our theory in practice, we look at a group of algorithms for the GuessWhat?! visual dialogue game and, using this example, test our theory empirically. Our theory accurately predicts relative-performance of multiple algorithms in generating equitable text as measured by both human and automated evaluation. | ['Malihe Alikhani', 'Anthony Sicilia'] | 2023-07-10 | null | null | null | null | ['fairness', 'fairness', 'text-generation', 'decision-making'] | ['computer-vision', 'miscellaneous', 'natural-language-processing', 'reasoning'] | [ 2.67042488e-01 8.93235028e-01 -3.38203669e-01 -5.22602499e-01
-5.32491148e-01 -4.69159484e-01 7.46453166e-01 9.97118875e-02
-4.95831609e-01 1.25976121e+00 5.19905925e-01 -4.53786582e-01
-3.84040759e-04 -7.19722569e-01 -9.71280560e-02 -7.69987106e-02
4.66354012e-01 6.67602599e-01 -4.98035043e-01 -6.47954881e-01
5.85705161e-01 2.24954840e-02 -1.36998379e+00 3.35942596e-01
1.41408777e+00 4.36469227e-01 -7.70036727e-02 8.87607574e-01
-2.97226727e-01 1.39327705e+00 -1.00022161e+00 -9.42745626e-01
3.24395820e-02 -1.02683866e+00 -1.49810040e+00 -9.29720104e-02
3.55310529e-01 -5.90753615e-01 1.08515076e-01 9.10243511e-01
6.02651417e-01 2.22741038e-01 9.47298825e-01 -1.49440944e+00
-6.55868053e-01 1.12403631e+00 -2.09684342e-01 -1.03185199e-01
5.42388082e-01 2.29135796e-01 1.30801380e+00 -6.81625724e-01
4.78763789e-01 1.37869692e+00 5.05704105e-01 1.07771134e+00
-1.20058727e+00 -4.87487972e-01 -1.83532819e-01 -2.77188241e-01
-7.54682302e-01 -6.05849981e-01 4.01024610e-01 -4.57645297e-01
6.79168165e-01 9.21961248e-01 6.57709479e-01 9.69450593e-01
1.80463254e-01 9.04360354e-01 1.17851090e+00 -5.84948003e-01
6.77607879e-02 6.00061238e-01 2.88947344e-01 5.40780306e-01
3.88879269e-01 1.29944146e-01 -7.77097821e-01 -2.21375883e-01
4.17105854e-01 -5.34034312e-01 -2.03856573e-01 -5.52391335e-02
-8.83966982e-01 1.21729755e+00 1.34578019e-01 7.92544410e-02
-2.54317939e-01 -1.10364415e-01 2.61690706e-01 5.38033009e-01
5.81237316e-01 1.20450592e+00 -1.20164841e-01 -3.75991464e-01
-6.55505419e-01 7.58587897e-01 1.40364623e+00 8.55229616e-01
5.24334073e-01 -5.34284525e-02 -6.43108904e-01 8.38717520e-01
1.58051252e-01 3.88153881e-01 5.95664442e-01 -1.17307818e+00
3.85070741e-01 4.47732389e-01 3.08738887e-01 -8.82535696e-01
-4.13601667e-01 -2.69313902e-01 -6.66738331e-01 1.24728978e-01
8.58479798e-01 -6.88516140e-01 6.43214211e-02 1.93592739e+00
1.53379649e-01 -6.57220066e-01 2.56403357e-01 8.48921001e-01
6.70336664e-01 4.37427133e-01 2.05586284e-01 -6.19266689e-01
1.02813601e+00 -6.87395334e-01 -9.53649104e-01 -1.25088602e-01
9.82365251e-01 -9.35710430e-01 1.41111577e+00 1.87760249e-01
-1.27629960e+00 -2.46045485e-01 -6.76472485e-01 -1.70028195e-01
2.18986049e-02 -2.62284309e-01 6.41541004e-01 1.04680026e+00
-7.69345164e-01 6.43888056e-01 1.10456161e-01 -2.03505725e-01
2.30488122e-01 2.45866776e-01 8.93707126e-02 3.35413277e-01
-1.44042623e+00 1.36098278e+00 3.00793082e-01 -2.17297494e-01
-3.93632472e-01 -6.10379517e-01 -6.25965655e-01 -7.43514001e-02
3.74407709e-01 -9.02482152e-01 1.94555223e+00 -1.37283039e+00
-1.62002611e+00 8.95980835e-01 -3.98278609e-03 -3.11675251e-01
8.98076117e-01 -2.15389848e-01 3.16307336e-01 -4.45872843e-01
2.20947996e-01 6.54502034e-01 4.46500182e-01 -1.45142961e+00
-9.01735604e-01 -1.71432108e-01 4.58941132e-01 7.46062279e-01
-4.10759896e-01 -2.16145702e-02 5.75804532e-01 -3.62262309e-01
-5.99818170e-01 -6.87522888e-01 -1.69573486e-01 -3.08281034e-01
-5.51001251e-01 -5.88598192e-01 3.00113797e-01 -4.77457553e-01
1.39001656e+00 -1.54680216e+00 -8.05813912e-03 2.06267416e-01
7.12169647e-01 3.73129249e-02 -8.61853827e-04 4.09368128e-01
1.89387396e-01 5.08864045e-01 1.46953672e-01 -1.39102638e-01
4.28036094e-01 -6.98070973e-02 -3.53470713e-01 1.42493248e-01
-5.85124753e-02 8.31271231e-01 -1.05878842e+00 -5.81450641e-01
1.13887917e-02 -1.48034260e-01 -9.01932359e-01 6.43203855e-01
-2.12038279e-01 1.31693944e-01 -3.06560010e-01 4.60292622e-02
2.53541857e-01 6.98680663e-03 2.27979898e-01 1.64476827e-01
-2.12173134e-01 4.49377924e-01 -6.86902583e-01 8.55796695e-01
-6.16888463e-01 6.53256357e-01 -2.67859936e-01 -5.88609457e-01
7.07611561e-01 9.38548744e-02 1.17853463e-01 -7.51230419e-01
2.50763863e-01 2.89806604e-01 3.85058284e-01 -5.20860195e-01
1.08903289e+00 -4.18907881e-01 -1.78173408e-01 1.13600922e+00
-2.52481401e-01 -5.47683120e-01 1.96815297e-01 5.39474726e-01
4.32151020e-01 -3.18198413e-01 5.48875690e-01 -4.48041946e-01
2.77591765e-01 1.06124252e-01 7.76287988e-02 1.16652739e+00
-2.40296990e-01 3.74558009e-02 9.06702399e-01 -3.17593813e-01
-1.26033604e+00 -6.13093615e-01 1.04791094e-02 1.65374398e+00
1.65564716e-01 -4.02381301e-01 -1.12743902e+00 -7.31950819e-01
-1.25426382e-01 1.35834098e+00 -4.93581384e-01 -3.25039893e-01
-2.85080016e-01 -7.35055685e-01 6.90985203e-01 1.14230432e-01
3.56407315e-01 -1.15039587e+00 -9.35998440e-01 -5.19912206e-02
-6.94808006e-01 -5.62960446e-01 -7.55905509e-01 -2.72505015e-01
-4.84803885e-01 -1.10805190e+00 -4.20485228e-01 -3.95102769e-01
4.35902804e-01 1.95878759e-01 1.34164691e+00 4.66652274e-01
7.17174262e-02 4.67456996e-01 -2.02437267e-01 -8.08619738e-01
-1.05885506e+00 1.53291315e-01 -5.12706675e-02 -3.15890282e-01
2.40849599e-01 -2.27718130e-02 -3.80516708e-01 2.62700886e-01
-5.78411579e-01 5.00478208e-01 1.24088816e-01 1.21202254e+00
-2.50074774e-01 -4.25007105e-01 7.75678098e-01 -1.21015680e+00
1.79740834e+00 -4.66514230e-01 -1.56480506e-01 2.55584300e-01
-1.01301920e+00 2.21308619e-01 4.89886075e-01 -2.15413749e-01
-1.28643131e+00 -4.95600432e-01 2.35232376e-02 4.99603033e-01
2.98980922e-01 3.93281013e-01 1.83536317e-02 1.50172025e-01
1.18313503e+00 -2.97034923e-02 4.29462403e-01 2.58716643e-01
6.37154758e-01 1.08599424e+00 1.87050045e-01 -9.03163612e-01
4.09740746e-01 -2.23901197e-01 -3.82214695e-01 -7.63127446e-01
-9.95282471e-01 -5.94153302e-03 -1.28625810e-01 -6.34627342e-01
5.56591868e-01 -4.95575994e-01 -1.12473118e+00 2.24200428e-01
-1.31509483e+00 -5.64333141e-01 -4.35699224e-01 1.34900421e-01
-9.55330253e-01 1.65760040e-01 -3.21688950e-01 -1.30411613e+00
-6.22213066e-01 -9.04489100e-01 6.95122719e-01 4.04815704e-01
-1.12243342e+00 -1.06036043e+00 5.96957840e-02 5.14986217e-01
5.47742188e-01 -5.16357385e-02 1.01235998e+00 -1.06037295e+00
2.47857883e-04 1.08870253e-01 -1.48910740e-02 2.55060464e-01
6.95840120e-02 -3.38085629e-02 -9.14927244e-01 7.91233778e-02
1.85382962e-01 -7.29896069e-01 2.64663249e-01 7.07409978e-02
8.63366425e-01 -1.04387021e+00 9.28871706e-02 -7.71091431e-02
8.36981654e-01 -5.80714606e-02 5.11789441e-01 1.95224181e-01
5.62023282e-01 1.28343618e+00 7.81800687e-01 8.00641418e-01
7.11268127e-01 5.79318583e-01 2.03420416e-01 9.08359420e-03
2.03590766e-01 -3.31304818e-01 2.80257523e-01 4.19248790e-01
-2.46612906e-01 -2.90657103e-01 -8.11149359e-01 2.62270987e-01
-1.81272566e+00 -1.37309659e+00 -2.55179197e-01 2.31331420e+00
1.57311368e+00 -1.03856832e-01 3.98797452e-01 -1.06388561e-01
7.60455430e-01 6.91421628e-02 -3.07135522e-01 -8.69758964e-01
9.01763737e-02 4.91394848e-02 1.61731645e-01 1.13471985e+00
-5.47706068e-01 9.17475462e-01 7.02236509e+00 7.51748085e-01
-7.45014548e-01 1.50690051e-02 1.16213000e+00 -2.51006395e-01
-1.00085616e+00 -2.37457395e-01 -4.07460719e-01 3.00347626e-01
7.96561599e-01 -9.80241060e-01 5.89466095e-01 7.79654860e-01
3.82062048e-01 -8.41584578e-02 -1.37379527e+00 8.21180701e-01
4.62277196e-02 -1.27224648e+00 1.06896713e-01 2.04240903e-02
7.96188831e-01 -8.04780483e-01 1.75952911e-02 4.92173940e-01
7.87555397e-01 -1.35930383e+00 7.93314338e-01 4.71368164e-01
6.41657114e-01 -9.82368052e-01 7.16126204e-01 7.94806719e-01
-1.05250336e-01 3.41073959e-03 -1.62861943e-01 -4.46460962e-01
-1.44096687e-01 2.18269706e-01 -1.05369258e+00 1.08224221e-01
-1.87134687e-02 -8.65895301e-02 -2.94436425e-01 3.62450510e-01
-1.12446666e-01 2.03230262e-01 1.29347816e-01 -6.63139820e-01
6.81396993e-03 5.92434593e-02 4.12892044e-01 1.08878994e+00
-1.11310132e-01 2.75870830e-01 2.07725558e-02 9.85823750e-01
-3.07135075e-01 5.18680274e-01 -6.94134772e-01 -8.53798687e-02
6.21450663e-01 1.19721079e+00 -3.02572325e-02 -3.74707103e-01
6.52681813e-02 7.33605981e-01 5.22628248e-01 1.23331450e-01
-6.26952887e-01 -1.99389651e-01 6.39490306e-01 5.12983389e-02
-8.50994170e-01 2.53466636e-01 -9.07198489e-01 -9.21021879e-01
-3.46543461e-01 -1.54405653e+00 1.58185720e-01 -4.35218364e-01
-1.38338590e+00 4.30418938e-01 -7.87473693e-02 -6.58955872e-01
-7.98024774e-01 -2.90475935e-01 -7.02774882e-01 1.24122167e+00
-1.03327584e+00 -7.63590157e-01 -3.63911361e-01 3.29730690e-01
6.50549829e-01 -3.45865995e-01 8.77310872e-01 -2.95228541e-01
-3.87010038e-01 9.39585149e-01 -2.40924835e-01 -1.32516861e-01
9.47171092e-01 -1.45187616e+00 2.70383000e-01 4.44621295e-01
-1.62038535e-01 7.03152359e-01 9.52578187e-01 -6.20231628e-01
-9.88881111e-01 -7.20315695e-01 1.35070610e+00 -7.41448879e-01
4.50401574e-01 2.97359768e-02 -4.96865690e-01 4.52969015e-01
5.38938105e-01 -1.00286448e+00 1.07089543e+00 3.23015720e-01
-1.44854382e-01 3.50830078e-01 -1.22867298e+00 1.10279655e+00
1.02856863e+00 -4.70964909e-01 -6.79551959e-01 4.49962378e-01
5.35521448e-01 -7.32978344e-01 -5.79043567e-01 -1.06368365e-03
7.10515201e-01 -1.19727123e+00 5.01037180e-01 -1.05522370e+00
9.26886857e-01 3.65884334e-01 3.45398374e-02 -1.54277897e+00
-6.03894219e-02 -1.11937296e+00 7.73182511e-02 1.00224054e+00
5.89626849e-01 -6.13392830e-01 5.89416027e-01 1.37119186e+00
3.41420583e-02 -8.14618170e-01 -4.85944003e-01 -2.31035024e-01
5.43037295e-01 -2.49579176e-01 5.78600228e-01 1.09121370e+00
6.38882637e-01 7.59317636e-01 -6.89814270e-01 -5.23614228e-01
5.88144720e-01 -1.00972690e-01 1.06223357e+00 -1.10330856e+00
-1.46189079e-01 -9.66828525e-01 3.19933444e-01 -1.02160454e+00
4.38834071e-01 -7.70332277e-01 2.30476886e-01 -1.24266541e+00
3.15227598e-01 -6.47396863e-01 6.13882244e-01 1.86944544e-01
-6.13512933e-01 -6.70207515e-02 4.90586072e-01 2.77840167e-01
-4.94937479e-01 4.53477919e-01 1.37241638e+00 1.29486830e-03
-3.20894331e-01 8.81252661e-02 -1.50380743e+00 6.04429066e-01
1.03652275e+00 -2.30538711e-01 -6.78167045e-01 -1.64829254e-01
4.31699902e-01 6.02170117e-02 1.63277432e-01 -2.70091474e-01
7.51771778e-02 -7.00439572e-01 1.47386596e-01 1.68854296e-01
-2.79606044e-01 -3.52542192e-01 -2.10750878e-01 4.54767793e-01
-1.08670926e+00 2.20862478e-02 -1.84583649e-01 7.69494176e-02
1.54956684e-01 -6.98981643e-01 7.40386009e-01 -1.65901750e-01
-1.60296634e-01 -5.05099073e-02 -6.06946468e-01 6.10446095e-01
9.12380755e-01 -1.54174477e-01 -4.80853796e-01 -9.64019120e-01
-3.07368487e-01 4.43575263e-01 3.16849589e-01 3.45925719e-01
3.99260014e-01 -1.21987736e+00 -9.81024861e-01 -1.38068929e-01
1.59452353e-02 -2.51015365e-01 -4.33781557e-02 5.03036797e-01
-4.99564558e-01 2.74502248e-01 -2.88121492e-01 -2.07496226e-01
-1.27728081e+00 -3.48357186e-02 7.10050642e-01 -2.64315397e-01
1.12292826e-01 7.07470894e-01 1.39259517e-01 -5.23425043e-01
2.60332912e-01 2.92783678e-01 -2.74182469e-01 2.01449350e-01
5.72634518e-01 4.28412884e-01 -5.12256145e-01 -4.25136387e-01
1.92826569e-01 -1.06743887e-01 -1.82818696e-01 -5.63808799e-01
6.54793739e-01 -1.91324562e-01 -2.77882338e-01 3.68182302e-01
5.63492596e-01 2.59014964e-01 -8.02299678e-01 5.17793605e-03
-2.59578172e-02 -6.15230918e-01 -3.26775432e-01 -8.39809358e-01
-3.88474584e-01 6.19318426e-01 8.06560889e-02 7.34944463e-01
6.77385926e-01 -2.34836951e-01 3.38477522e-01 5.12070060e-01
2.56571084e-01 -1.31281030e+00 3.80921572e-01 7.07341731e-01
1.10028207e+00 -1.28528571e+00 -1.41990364e-01 -2.77379602e-01
-1.33084738e+00 1.15182495e+00 1.06620526e+00 4.09914255e-01
1.21885613e-02 -1.05189914e-02 2.79422402e-01 5.52795688e-03
-9.35950696e-01 -5.57182299e-04 3.47468443e-02 6.74562871e-01
1.07816541e+00 2.92996109e-01 -9.45428967e-01 6.48484290e-01
-9.29567993e-01 -1.40210003e-01 8.69455218e-01 4.66142803e-01
-6.33835316e-01 -1.13046062e+00 -1.18514620e-01 6.39393985e-01
-3.62201452e-01 -2.96583921e-01 -8.58840346e-01 6.25044107e-01
-1.49927780e-01 1.15167212e+00 -1.38628706e-02 -3.89751792e-01
3.42505008e-01 7.63645619e-02 4.83372957e-01 -6.44839346e-01
-9.65532064e-01 -5.39144754e-01 7.46104658e-01 -2.46549860e-01
-2.52096921e-01 -6.40817821e-01 -1.10924447e+00 -1.07214594e+00
-4.21533167e-01 4.44534749e-01 2.13387057e-01 1.01428151e+00
5.08050248e-02 2.40834475e-01 8.16273510e-01 -4.50774878e-01
-1.27425611e+00 -1.03955722e+00 -3.01306188e-01 5.71692348e-01
1.01538941e-01 -1.65640771e-01 -3.71744603e-01 -1.50519967e-01] | [12.663613319396973, 8.155471801757812] |
a05c2494-623e-4f2e-bf37-a38092af856a | end-to-end-high-accuracy-license-plate | 2202.10277 | null | https://arxiv.org/abs/2202.10277v1 | https://arxiv.org/pdf/2202.10277v1.pdf | End-to-End High Accuracy License Plate Recognition Based on Depthwise Separable Convolution Networks | Automatic license plate recognition plays a crucial role in modern transportation systems such as for traffic monitoring and vehicle violation detection. In real-world scenarios, license plate recognition still faces many challenges and is impaired by unpredictable interference such as weather or lighting conditions. Many machine learning based ALPR solutions have been proposed to solve such challenges in recent years. However, most are not convincing, either because their results are evaluated on small or simple datasets that lack diverse surroundings, or because they require powerful hardware to achieve a reasonable frames-per-second in real-world applications. In this paper, we propose a novel segmentation-free framework for license plate recognition and introduce NP-ALPR, a diverse and challenging dataset which resembles real-world scenarios. The proposed network model consists of the latest deep learning methods and state-of-the-art ideas, and benefits from a novel network architecture. It achieves higher accuracy with lower computational requirements than previous works. We evaluate the effectiveness of the proposed method on three different datasets and show a recognition accuracy of over 99% and over 70 fps, demonstrating that our method is not only robust but also computationally efficient. | ['Wen-Kai Tai', 'Zheng-Yi Shen', 'Hong-Yang Shih', 'Song-Ren Wang'] | 2022-02-21 | null | null | null | null | ['license-plate-recognition'] | ['computer-vision'] | [ 5.38567044e-02 -9.78567779e-01 -1.37081593e-01 -3.42571974e-01
-7.54919767e-01 -5.48417509e-01 4.51310873e-01 -5.20965099e-01
-6.24418378e-01 7.14404941e-01 -5.57881117e-01 -2.41851032e-01
1.13816977e-01 -7.51721919e-01 -7.83334434e-01 -6.45396352e-01
2.41052702e-01 4.16269988e-01 9.57966030e-01 -2.84531713e-01
4.60685313e-01 8.16622138e-01 -1.64224660e+00 2.66934279e-04
8.05739641e-01 1.01025474e+00 -3.08720171e-02 4.76954997e-01
-1.13922663e-01 5.42292535e-01 -6.86357558e-01 -6.23925984e-01
6.61846876e-01 5.65079413e-02 -5.75161837e-02 2.43949920e-01
4.17245388e-01 -3.51393253e-01 -8.27869058e-01 8.67811680e-01
3.13825071e-01 2.28664517e-01 4.01413023e-01 -1.22810137e+00
-2.42463097e-01 -2.41347298e-01 -6.73139513e-01 2.56438583e-01
-7.40936399e-03 5.70278168e-01 4.17548537e-01 -7.78713584e-01
3.07379544e-01 5.81442058e-01 1.03708804e+00 5.35024107e-01
-6.86636746e-01 -7.69631267e-01 -8.01586881e-02 4.81330633e-01
-1.66324604e+00 -7.01786637e-01 7.15375841e-01 -2.03706011e-01
7.13234544e-01 2.99057066e-01 4.83052999e-01 1.01082599e+00
7.57399499e-02 8.79468739e-01 1.00354695e+00 -1.94701582e-01
2.07545474e-01 2.19038501e-03 9.91010815e-02 5.79244077e-01
4.05840158e-01 -2.91575760e-01 -1.87920719e-01 4.93713915e-01
9.47228789e-01 2.58612067e-01 -2.57999063e-01 1.51435032e-01
-1.10500598e+00 4.39421386e-01 8.94138515e-02 2.80315310e-01
-9.87483189e-02 8.37413445e-02 1.11886233e-01 -1.13879330e-01
2.83118844e-01 -6.38995394e-02 -9.77278650e-02 -3.80938292e-01
-1.27507603e+00 1.28141478e-01 6.37950718e-01 1.01999927e+00
6.57671928e-01 4.30963784e-01 4.56423648e-02 1.00220215e+00
3.03889006e-01 6.61743760e-01 4.48532104e-01 -6.13204479e-01
7.52305627e-01 3.52588832e-01 1.92625850e-01 -1.25635815e+00
-3.98856759e-01 -3.70390594e-01 -9.28425074e-01 4.30302143e-01
6.17862344e-01 -2.45953463e-02 -9.63491380e-01 9.23645675e-01
-1.32838547e-01 6.38681352e-01 -6.13401458e-02 1.03382730e+00
7.26520181e-01 1.07400656e+00 -2.52047598e-01 6.71566348e-04
1.11470842e+00 -1.30041111e+00 -5.92478752e-01 -4.80301380e-01
2.49264985e-01 -8.11503410e-01 8.57037306e-01 4.37513441e-01
-9.02895808e-01 -6.03582263e-01 -1.34135115e+00 1.25482813e-01
-3.65854532e-01 4.05522615e-01 3.11488092e-01 9.48599994e-01
-8.59511197e-01 2.96582609e-01 -6.08464420e-01 -3.31175357e-01
5.96272349e-01 4.78210986e-01 -3.27260852e-01 -3.30087602e-01
-9.01608109e-01 7.20587790e-01 -3.77393067e-02 5.33767164e-01
-5.17434478e-01 -3.11720192e-01 -5.76276004e-01 7.79777020e-02
4.90312159e-01 -1.53591380e-01 9.66103733e-01 -7.88109303e-01
-1.87567651e+00 6.59395635e-01 -8.60443264e-02 -4.41484749e-01
7.82703996e-01 -1.47098392e-01 -8.14946234e-01 2.47992519e-02
-4.25485760e-01 2.46015906e-01 7.80035198e-01 -9.50380325e-01
-6.77456379e-01 -2.36509383e-01 5.67703620e-02 1.21805631e-01
-2.14215830e-01 1.15195684e-01 -1.01237893e+00 -3.24468762e-01
-2.45194435e-02 -1.01154268e+00 -1.26875758e-01 3.73252071e-02
-2.27569968e-01 -7.40342811e-02 9.97469842e-01 -4.14040208e-01
7.35793710e-01 -2.25306988e+00 -9.53101456e-01 8.29223916e-02
-2.68750880e-02 9.98114884e-01 -5.15702292e-02 2.57421166e-01
3.32103968e-01 2.04052404e-01 -1.37887627e-01 -3.42832565e-01
-5.42270318e-02 8.15688819e-02 -2.63976187e-01 8.47397506e-01
1.29287019e-01 6.60448492e-01 -2.11129919e-01 -4.23644155e-01
7.10650921e-01 5.55715799e-01 -6.61205351e-02 1.85945835e-02
5.47577180e-02 1.39816582e-01 -3.57182801e-01 8.58233452e-01
1.10764587e+00 7.29851276e-02 -3.05395782e-01 -2.46448480e-02
-3.13369721e-01 -2.06990004e-01 -1.26804602e+00 1.23222911e+00
-2.26803228e-01 1.20796120e+00 1.10491095e-02 -1.07367659e+00
1.19782007e+00 8.60067159e-02 2.45602608e-01 -1.08960831e+00
1.83366612e-01 4.82501268e-01 -3.02110046e-01 -6.46699131e-01
6.14463151e-01 -3.34412493e-02 1.81811795e-01 -1.73358771e-03
-3.34713042e-01 1.00286558e-01 3.81783634e-01 -1.61304370e-01
1.04584873e+00 -4.62362655e-02 -1.47054508e-01 3.13868262e-02
9.23433661e-01 -7.14196712e-02 7.91525304e-01 7.40980387e-01
-5.85353076e-01 8.17895591e-01 1.41532898e-01 -7.59296834e-01
-1.04388726e+00 -8.10643315e-01 2.07157172e-02 5.42459428e-01
6.82491481e-01 2.42702663e-01 -6.20491385e-01 -3.80010962e-01
-2.51724809e-01 3.62427503e-01 -3.64776514e-02 1.34537697e-01
-9.28347290e-01 -8.12112808e-01 9.73725021e-01 5.30784428e-01
1.15610754e+00 -7.18265593e-01 -1.91432774e-01 1.34594306e-01
-1.33300781e-01 -1.89828849e+00 -4.04607147e-01 -5.64933181e-01
-6.97857022e-01 -9.47126210e-01 -7.26905227e-01 -8.98156822e-01
4.44325984e-01 7.12861061e-01 1.00099695e+00 2.45419875e-01
-3.86282541e-02 -4.31763493e-02 -1.08741388e-01 -2.15834156e-01
-1.66569099e-01 -6.01039082e-02 3.19559500e-02 5.42942345e-01
4.80910182e-01 -3.41063738e-01 -6.36373281e-01 8.35818172e-01
-9.36247945e-01 -8.44488591e-02 8.74353886e-01 3.35573435e-01
4.13665444e-01 2.76841253e-01 5.09330451e-01 -4.84075040e-01
3.92301470e-01 -1.38963193e-01 -1.24518681e+00 1.00895651e-01
-2.32621774e-01 -5.05109251e-01 8.98281693e-01 -1.67678490e-01
-9.69593585e-01 1.33599639e-01 -5.18341780e-01 -4.01927382e-01
-5.79696059e-01 1.44600272e-01 -9.44866091e-02 -4.80315715e-01
2.34847412e-01 7.01858044e-01 -6.84741586e-02 -1.10220246e-01
-2.13104889e-01 8.23606253e-01 7.56435990e-01 -2.98965216e-01
1.08268917e+00 6.51848257e-01 6.59949556e-02 -1.32714927e+00
-3.19515616e-01 -6.55252993e-01 -3.35856110e-01 -6.55220091e-01
6.60119832e-01 -8.34451079e-01 -1.14475632e+00 1.10255373e+00
-1.09519064e+00 -5.04769832e-02 2.15571076e-01 6.47172570e-01
-3.50387037e-01 5.58527231e-01 -5.06255448e-01 -8.23892713e-01
-7.43342489e-02 -1.28682542e+00 8.34935188e-01 6.50549829e-01
6.29360974e-01 -6.49480343e-01 -9.37753171e-02 7.55594492e-01
8.07666481e-01 1.43627882e-01 3.00683051e-01 -7.36643374e-01
-1.21138835e+00 -6.39323771e-01 -5.55197716e-01 3.47815156e-01
-2.05446079e-01 3.05333167e-01 -9.20271158e-01 -1.65198662e-03
-2.11429819e-01 -2.08906978e-01 8.44798863e-01 3.61983120e-01
1.15011203e+00 -3.78420986e-02 -2.35421270e-01 6.22054219e-01
1.70158517e+00 4.17058915e-01 1.21940899e+00 5.57599187e-01
6.04339123e-01 2.73745865e-01 2.84184605e-01 1.66087762e-01
2.44735286e-01 8.37902844e-01 3.99771661e-01 -1.78790182e-01
1.02586672e-01 1.33265167e-01 3.84375691e-01 7.83056319e-01
-3.94817203e-01 -6.34651303e-01 -1.12443054e+00 3.71439338e-01
-1.86367190e+00 -1.26724708e+00 -4.95967627e-01 2.22464395e+00
1.94739684e-01 4.42835838e-01 2.16699302e-01 1.72081128e-01
9.29934204e-01 1.88574001e-01 -3.61465901e-01 -1.95971206e-01
-3.29164743e-01 -1.47836104e-01 7.47019827e-01 1.87453449e-01
-1.16487372e+00 9.20707941e-01 6.03751373e+00 9.27941740e-01
-1.38775456e+00 -1.09294891e-01 6.88494444e-01 1.50394559e-01
2.38673151e-01 -5.28750479e-01 -8.96017611e-01 6.20337844e-01
9.59838212e-01 1.04101524e-01 1.81574643e-01 7.09260523e-01
3.81069422e-01 -5.49482927e-02 -7.44071782e-01 1.37035251e+00
3.70695621e-01 -1.68322921e+00 -1.83587596e-01 1.10751623e-03
4.78288025e-01 2.69224912e-01 1.77272141e-01 4.07810301e-01
-2.05234706e-01 -1.03967345e+00 5.65161645e-01 5.18468618e-01
6.49017632e-01 -1.00484991e+00 1.09986734e+00 5.36978483e-01
-1.13574219e+00 1.35123655e-01 -5.60735643e-01 2.10798867e-02
7.67467991e-02 5.04408896e-01 -6.45614445e-01 2.99065143e-01
3.77039701e-01 8.00216079e-01 -4.79217529e-01 1.47825074e+00
8.68394151e-02 8.01572382e-01 -4.15483356e-01 -2.77267754e-01
5.49378455e-01 -3.93269688e-01 4.68785077e-01 1.55400276e+00
4.41917896e-01 3.61131579e-02 7.83616379e-02 6.90875888e-01
-3.34934622e-01 -4.68062982e-02 -6.41092241e-01 2.07190722e-01
2.51465946e-01 1.33751714e+00 -9.21558678e-01 -4.15082239e-02
-9.10811543e-01 7.52220452e-01 -1.95694372e-01 4.17428076e-01
-1.40277708e+00 -5.04851520e-01 6.53025508e-01 2.80399740e-01
3.87737840e-01 -6.14426374e-01 -8.72098953e-02 -1.28711450e+00
4.06986803e-01 -5.48457623e-01 -1.04930945e-01 -6.32833540e-01
-1.01813889e+00 5.41065753e-01 -4.90884393e-01 -1.67643130e+00
1.73871785e-01 -1.03622949e+00 -7.28495538e-01 5.35090685e-01
-1.96324980e+00 -9.76514995e-01 -5.14337063e-01 9.06737983e-01
9.68116641e-01 -4.60256308e-01 3.45397532e-01 7.03265488e-01
-9.85327899e-01 7.23234057e-01 5.02730012e-01 5.49623549e-01
5.99611163e-01 -6.15813375e-01 5.34772575e-01 1.29223633e+00
3.79127562e-02 1.47889376e-01 3.02551121e-01 -1.50532216e-01
-1.67215359e+00 -1.21966362e+00 5.22870779e-01 -1.34895146e-01
4.13017362e-01 -4.59745467e-01 -9.08114851e-01 3.82838547e-01
4.43841629e-02 4.90933537e-01 5.73203862e-01 -4.12804812e-01
-1.67663425e-01 -6.74472868e-01 -1.28384423e+00 5.57518721e-01
7.43321538e-01 4.69279522e-03 -2.24906132e-01 2.90891409e-01
2.31100753e-01 -3.57303649e-01 -2.33003214e-01 2.07133114e-01
6.78500950e-01 -1.16662478e+00 8.60281765e-01 -1.43294513e-01
2.49231532e-01 -6.94858551e-01 -9.40256491e-02 -7.48970330e-01
-1.12411842e-01 -5.11436224e-01 2.15969101e-01 1.15369701e+00
5.20259857e-01 -7.79729068e-01 1.05575073e+00 7.41621375e-01
-3.71932417e-01 -5.77899218e-01 -1.18366766e+00 -1.11176264e+00
-2.95374870e-01 -7.60741711e-01 4.54560399e-01 6.72711730e-01
-5.72773278e-01 -7.81588703e-02 -6.62909806e-01 4.20660019e-01
6.67798758e-01 -1.54142026e-02 9.32810664e-01 -1.12902856e+00
7.43412897e-02 -4.59194571e-01 -1.16983855e+00 -1.15118515e+00
3.61125208e-02 -2.25519553e-01 3.11732352e-01 -1.51332891e+00
-2.54948568e-02 -3.19711864e-01 -1.72212407e-01 2.29784325e-01
2.61748731e-01 8.21945667e-01 3.10267180e-01 5.20770788e-01
-9.50043380e-01 2.73731798e-01 8.11660945e-01 -4.72358137e-01
-1.11765852e-02 2.42011279e-01 -1.80146232e-01 1.06480944e+00
9.10320640e-01 -2.25891143e-01 -1.29008636e-01 -5.94503641e-01
7.21558630e-02 -7.39345178e-02 2.88450867e-01 -1.55179358e+00
6.63396895e-01 -1.84887245e-01 5.12314498e-01 -5.59595168e-01
4.79818702e-01 -1.03349090e+00 -5.43636642e-02 1.46038845e-01
3.08843464e-01 -4.90866713e-02 3.70215923e-01 5.32953620e-01
-4.58899438e-01 -1.74921855e-01 1.01198542e+00 8.63151178e-02
-1.26704562e+00 3.82371008e-01 -6.12805903e-01 -1.30665407e-03
1.27323818e+00 -8.36157560e-01 -6.33841634e-01 -4.76954222e-01
-7.71433488e-02 -8.63669738e-02 4.00653780e-01 5.31417131e-01
7.84338951e-01 -1.24673533e+00 -9.27908361e-01 3.23003948e-01
9.29156542e-02 -1.29855767e-01 2.62784958e-01 7.56695151e-01
-1.18997204e+00 6.97802842e-01 -3.70277047e-01 -8.13125312e-01
-1.13382792e+00 1.40244946e-01 3.29043776e-01 -1.07356116e-01
-6.21457160e-01 4.86446023e-01 -1.60926044e-01 -2.26343468e-01
9.00356397e-02 -4.98293824e-02 -1.44174993e-01 -4.07725185e-01
6.50393069e-01 6.10450268e-01 3.70844275e-01 -8.46566081e-01
-4.58245963e-01 9.49857354e-01 -1.47535861e-01 8.87298360e-02
1.40684044e+00 6.46961853e-02 3.32463950e-01 5.71990237e-02
9.40158188e-01 4.12183907e-03 -1.35426188e+00 -4.78711873e-02
-2.84335643e-01 -8.24617445e-01 -1.58713192e-01 -4.95184124e-01
-1.27881348e+00 7.54585445e-01 4.78431135e-01 7.15055391e-02
1.11752081e+00 -5.92240930e-01 1.03926730e+00 7.40609050e-01
5.94316840e-01 -1.24622333e+00 -1.24670796e-01 6.06104136e-01
4.70145017e-01 -1.32686448e+00 -2.01031446e-01 -3.87914777e-01
-4.25137967e-01 1.43771136e+00 7.22511351e-01 -2.54215866e-01
3.64084125e-01 3.65846485e-01 2.03108907e-01 3.41633588e-01
-2.43534505e-01 -1.64654441e-02 8.05051550e-02 5.04585564e-01
1.77593678e-01 -1.88622206e-01 -1.39241159e-01 3.98259699e-01
2.64077872e-01 1.50985882e-01 8.09844434e-01 7.86463737e-01
-4.47317928e-01 -7.45779574e-01 -5.22639513e-01 3.28733772e-01
-4.44444686e-01 2.64503628e-01 -7.06627443e-02 1.21137440e+00
6.84657022e-02 9.92092192e-01 2.11671904e-01 -3.77660394e-01
3.84684175e-01 -3.04579698e-02 1.37433141e-01 2.44962741e-02
-4.05347385e-02 -1.43619448e-01 -2.46546585e-02 -5.51326931e-01
-6.10199451e-01 -4.28243279e-01 -1.10515976e+00 -7.70003796e-01
-3.14424694e-01 -1.67466611e-01 8.75674486e-01 1.16387177e+00
3.30847025e-01 3.01116943e-01 7.16108441e-01 -9.39068854e-01
-2.12657616e-01 -5.46696126e-01 -5.91978610e-01 1.85682625e-01
2.45175973e-01 -4.08282369e-01 -2.18742937e-01 1.01356462e-01] | [9.843086242675781, -4.914684295654297] |
81d82453-b8fd-47ef-9fc5-36600ab092f0 | zero-shot-classification-by-logical-reasoning | 2211.03252 | null | https://arxiv.org/abs/2211.03252v2 | https://arxiv.org/pdf/2211.03252v2.pdf | Zero-Shot Classification by Logical Reasoning on Natural Language Explanations | Humans can classify data of an unseen category by reasoning on its language explanations. This ability is owing to the compositional nature of language: we can combine previously seen attributes to describe the new category. For example, we might describe a sage thrasher as "it has a slim straight relatively short bill, yellow eyes and a long tail", so that others can use their knowledge of attributes "slim straight relatively short bill", "yellow eyes" and "long tail" to recognize a sage thrasher. Inspired by this observation, in this work we tackle zero-shot classification task by logically parsing and reasoning on natural language expla-nations. To this end, we propose the framework CLORE (Classification by LOgical Reasoning on Explanations). While previous methods usually regard textual information as implicit features, CLORE parses explanations into logical structures and then explicitly reasons along thess structures on the input to produce a classification score. Experimental results on explanation-based zero-shot classification benchmarks demonstrate that CLORE is superior to baselines, which we further show mainly comes from higher scores on tasks requiring more logical reasoning. We also demonstrate that our framework can be extended to zero-shot classification on visual modality. Alongside classification decisions, CLORE can provide the logical parsing and reasoning process as a clear form of rationale. Through empirical analysis we demonstrate that CLORE is also less affected by linguistic biases than baselines. | ['Heng Ji', 'Xinya Du', 'Hengzhi Pei', 'Chi Han'] | 2022-11-07 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [ 2.71839231e-01 7.03617573e-01 -4.63195056e-01 -7.11121619e-01
-5.25572598e-01 -5.49271166e-01 8.74595106e-01 5.08626759e-01
-1.08497199e-02 4.13835913e-01 5.96632659e-01 -6.64184451e-01
-1.41131818e-01 -8.55991066e-01 -5.94436467e-01 -2.50636667e-01
3.60658973e-01 5.20143032e-01 1.37209818e-01 -2.89848536e-01
3.74312133e-01 8.65770206e-02 -1.92184460e+00 9.24682498e-01
5.70485830e-01 9.17206109e-01 -1.82317495e-01 5.33826888e-01
-4.85682666e-01 1.15704501e+00 -2.48059779e-01 -6.65263832e-01
-3.70153971e-03 -6.22550786e-01 -1.20730805e+00 3.06403071e-01
6.99534059e-01 -1.88880682e-01 3.41672212e-01 8.37051749e-01
-1.85997918e-01 1.91041812e-01 1.06225371e+00 -1.42481017e+00
-1.16682482e+00 1.07547534e+00 -3.10296476e-01 -7.73933679e-02
6.02162063e-01 2.31740206e-01 1.68279302e+00 -1.02570629e+00
7.62379229e-01 1.43748689e+00 6.56295657e-01 8.25437963e-01
-1.46112704e+00 -3.85257870e-01 5.13338149e-01 3.72931331e-01
-8.09062600e-01 -1.76251546e-01 4.68722373e-01 -5.23300052e-01
1.03920448e+00 4.90611613e-01 3.48348290e-01 1.14627230e+00
4.48629931e-02 9.65421140e-01 1.13938415e+00 -5.31610429e-01
4.44080353e-01 5.02326600e-02 8.88018668e-01 9.49706614e-01
1.69933349e-01 -9.66428369e-02 -5.35844088e-01 2.77463812e-02
-8.65381882e-02 1.56843528e-01 -9.40188244e-02 -3.96861196e-01
-1.13006222e+00 1.17329597e+00 8.83324683e-01 4.34851646e-02
-1.83569610e-01 1.89959481e-01 3.62962484e-01 2.58347154e-01
2.63059705e-01 6.19144142e-01 -3.72689784e-01 2.26608247e-01
-6.16794765e-01 2.61554837e-01 8.47094417e-01 8.86934280e-01
7.82722414e-01 -2.20410123e-01 -3.83536965e-01 5.90298295e-01
2.36203194e-01 2.46679872e-01 5.27693629e-01 -1.03186631e+00
1.00160524e-01 8.74366760e-01 -1.29545838e-01 -5.89582801e-01
-4.60231751e-01 -1.35451883e-01 -6.46925449e-01 3.67830396e-01
3.49063814e-01 1.91153094e-01 -1.04625821e+00 1.64060116e+00
-5.84582686e-02 -1.68652639e-01 3.72803122e-01 1.03085339e+00
1.17955506e+00 3.34178090e-01 2.42735580e-01 1.66408390e-01
1.77969515e+00 -1.09527922e+00 -4.67206240e-01 -6.03782475e-01
7.54589617e-01 -3.21754158e-01 1.60835052e+00 4.30723101e-01
-6.83522224e-01 -4.64658499e-01 -1.03457415e+00 -4.88446534e-01
-6.25310421e-01 -8.78923684e-02 9.49163198e-01 4.69294488e-01
-7.40041554e-01 3.79032582e-01 -4.36037421e-01 -6.53272748e-01
4.75806534e-01 -9.33572501e-02 -3.85750175e-01 -1.13199599e-01
-1.01502991e+00 8.23129654e-01 4.28448021e-01 -4.69097376e-01
-4.08904225e-01 -6.51691496e-01 -1.15622938e+00 3.40461522e-01
7.00808883e-01 -9.13536012e-01 1.47005653e+00 -9.14731443e-01
-1.01794946e+00 1.09139764e+00 -2.93214858e-01 -7.52940476e-01
4.96787935e-01 -2.53148347e-01 -8.56948569e-02 1.45790698e-02
4.70419198e-01 7.60457635e-01 6.99698150e-01 -1.28178108e+00
-4.57659692e-01 -2.82965630e-01 3.51140648e-01 -3.48611355e-01
-9.88168120e-02 -1.88275307e-01 1.87595755e-01 -5.15307188e-01
3.91909778e-01 -7.52489746e-01 9.38934237e-02 1.02688394e-01
-6.01782680e-01 -4.62774009e-01 6.41599357e-01 -1.44803658e-01
9.25858021e-01 -2.08399606e+00 -7.66225308e-02 -7.71775469e-02
5.11636555e-01 -1.85067996e-01 1.17945336e-02 3.96053553e-01
-2.54824638e-01 5.38776934e-01 -2.84347683e-01 -3.57129425e-02
4.42749500e-01 4.61187303e-01 -9.37618673e-01 7.00076967e-02
4.10600424e-01 1.14484549e+00 -9.77250814e-01 -2.83128053e-01
1.66544840e-02 1.72632057e-02 -7.58444726e-01 1.62615061e-01
-4.30164099e-01 -8.30131024e-02 -1.33670256e-01 5.77258766e-01
2.98202246e-01 -6.57266617e-01 9.73607153e-02 -1.50962770e-01
-2.35458836e-02 3.11480850e-01 -7.60218740e-01 1.32475090e+00
-4.77036536e-01 6.63425386e-01 -4.73627418e-01 -8.54692698e-01
9.34190094e-01 8.15219432e-02 -3.56787711e-01 -3.86760443e-01
2.51906663e-01 5.95429279e-02 2.54052803e-02 -6.42084897e-01
2.91032374e-01 -5.17096877e-01 -3.88490468e-01 6.12903774e-01
-2.19684780e-01 -2.21305653e-01 1.03668451e-01 6.33815110e-01
1.12777615e+00 7.15191290e-02 7.91072190e-01 -2.75831908e-01
4.14041877e-01 2.95725495e-01 2.62675941e-01 9.98704016e-01
-5.78484535e-02 5.62582970e-01 7.40389228e-01 -7.22935200e-01
-9.18818951e-01 -1.37117231e+00 -5.78483790e-02 1.42464101e+00
2.40629554e-01 -6.64541900e-01 -5.26357770e-01 -8.58112216e-01
1.92164391e-01 1.29160547e+00 -1.00807393e+00 -1.02686726e-01
-1.95846632e-02 -7.94968605e-02 3.56453300e-01 8.82324994e-01
3.27979654e-01 -9.75620329e-01 -1.06349397e+00 -1.38498589e-01
-1.06717937e-01 -1.05360222e+00 -3.05388942e-02 5.41399002e-01
-4.63617593e-01 -1.10021877e+00 -1.32731318e-01 -4.37866360e-01
6.62089884e-01 3.45989764e-01 1.19316947e+00 6.12133145e-01
-2.58140415e-01 3.78200561e-01 -7.47478247e-01 -5.00555158e-01
-4.36630756e-01 -2.73722112e-01 -1.46059334e-01 -6.83314279e-02
5.69416463e-01 -3.32205802e-01 -2.85621822e-01 1.26133179e-02
-7.91451275e-01 4.76468176e-01 4.94383067e-01 1.03322005e+00
2.29406878e-01 -3.05211246e-01 1.97617829e-01 -1.44393313e+00
5.71905017e-01 -5.76968729e-01 1.78905781e-02 4.49211031e-01
-7.62076616e-01 7.45203495e-01 7.28372157e-01 -3.11052829e-01
-1.05232000e+00 -1.12006404e-01 2.10410759e-01 -2.00954959e-01
-4.02044564e-01 4.26228374e-01 8.43642652e-02 4.79301333e-01
9.22405124e-01 -3.55433151e-02 -8.81988257e-02 -2.27478191e-01
8.26326370e-01 5.94881952e-01 7.95650244e-01 -8.09077799e-01
7.40903616e-01 6.60946190e-01 -5.13173044e-02 -7.01559901e-01
-1.53819263e+00 -4.26580012e-01 -5.83201826e-01 8.52507912e-03
1.10230565e+00 -6.44703031e-01 -9.40337837e-01 -2.57829964e-01
-1.38429379e+00 -1.18395634e-01 -2.30211034e-01 1.11119181e-01
-5.73002756e-01 2.98934340e-01 -5.00865281e-01 -9.20891047e-01
-1.60472453e-01 -7.79587269e-01 1.10029113e+00 -6.92557693e-02
-9.74039733e-01 -8.34229827e-01 -4.93404239e-01 5.51660240e-01
2.14778006e-01 2.11517543e-01 1.50902593e+00 -1.08387005e+00
-3.83884639e-01 7.02991784e-02 -4.80234951e-01 -5.53029366e-02
-3.24499995e-01 -8.08964372e-02 -1.11535752e+00 2.24759012e-01
-1.04331389e-01 -5.89545131e-01 1.15651226e+00 -1.61214381e-01
1.18246675e+00 -4.50544715e-01 -1.58843592e-01 3.91774923e-01
1.30957067e+00 -2.61609972e-01 3.16211581e-01 4.37706620e-01
4.20784533e-01 9.31683898e-01 6.51944041e-01 4.10130590e-01
5.44684708e-01 3.13797325e-01 5.69100499e-01 5.85047789e-02
-2.44558066e-01 -3.41417074e-01 2.78791636e-01 9.68100429e-02
2.63654348e-02 -2.52229750e-01 -1.15533149e+00 3.41826260e-01
-1.94079566e+00 -1.17436004e+00 -5.86536586e-01 1.76227927e+00
4.44408387e-01 2.17967287e-01 -1.27097949e-01 1.81412786e-01
4.87363487e-01 4.24223626e-03 -4.46620166e-01 -8.49511743e-01
-9.22577009e-02 7.37040639e-02 1.64301440e-01 5.96310079e-01
-8.80125701e-01 1.00024843e+00 5.94061089e+00 3.22148740e-01
-9.60502446e-01 2.32678819e-02 4.43838060e-01 2.00753167e-01
-6.49247229e-01 3.67419273e-01 -7.30607748e-01 1.06808096e-01
5.80541193e-01 -2.71125674e-01 2.33987987e-01 9.98026729e-01
-2.76288927e-01 -1.81942537e-01 -1.64930654e+00 7.06310272e-01
2.35826775e-01 -1.40813267e+00 4.57454532e-01 -1.14995830e-01
3.60145658e-01 -3.61835510e-01 4.16553617e-02 6.08479023e-01
5.58968842e-01 -1.15385652e+00 1.11900508e+00 3.26614261e-01
5.20712078e-01 -1.94648832e-01 6.25663757e-01 5.53734839e-01
-1.04715621e+00 -4.63800162e-01 -3.14142168e-01 -6.32530987e-01
-7.96597674e-02 1.43869966e-01 -9.85319674e-01 2.80752569e-01
4.67667639e-01 6.31930351e-01 -8.39484513e-01 4.31552887e-01
-9.72592652e-01 3.85162979e-01 2.50725776e-01 -3.08809012e-01
3.58900368e-01 3.19552422e-01 3.16485435e-01 1.31212068e+00
2.37176627e-01 4.22696292e-01 2.98527211e-01 1.22214222e+00
5.18351234e-02 -8.06569383e-02 -5.72932065e-01 -1.54093290e-02
2.83448070e-01 1.13051713e+00 -9.19761240e-01 -8.61147881e-01
-4.90979850e-01 9.67241883e-01 4.45720792e-01 1.82103142e-01
-9.02695596e-01 -1.08409949e-01 4.91693854e-01 2.39506617e-01
1.42917603e-01 2.85621941e-01 -5.45868158e-01 -1.32169008e+00
-2.20829383e-01 -6.23415649e-01 7.13570058e-01 -1.33200729e+00
-1.60714841e+00 6.58531964e-01 2.72369068e-02 -1.04860151e+00
-3.26848090e-01 -9.40534174e-01 -6.84755564e-01 4.91324991e-01
-1.32292986e+00 -1.37856698e+00 -5.10942042e-01 2.12375894e-01
1.00810826e+00 -1.90652069e-02 9.48499382e-01 -5.58018208e-01
-1.32880688e-01 2.21027285e-01 -6.92282081e-01 1.35212198e-01
6.45794868e-01 -1.62952065e+00 3.94064248e-01 7.60691941e-01
4.25064772e-01 9.86970186e-01 1.12087309e+00 -5.06520689e-01
-1.05369782e+00 -7.66505420e-01 1.19546664e+00 -7.60347188e-01
9.95601535e-01 -3.10729712e-01 -1.14640212e+00 8.54434788e-01
3.02474350e-01 6.57710293e-03 8.72569621e-01 5.00272512e-01
-1.27349699e+00 4.19859141e-01 -8.36423337e-01 5.88927031e-01
1.01568723e+00 -6.62041962e-01 -1.38394463e+00 2.69550622e-01
6.07799232e-01 -8.62037763e-02 -3.37455004e-01 2.51543611e-01
6.90762460e-01 -1.25457370e+00 8.80304515e-01 -1.06411028e+00
1.07326102e+00 -1.66068435e-01 -3.06495935e-01 -1.19271970e+00
-4.43571448e-01 -7.35105500e-02 1.84805661e-01 1.06560695e+00
5.65813780e-01 -4.42458451e-01 5.35288393e-01 8.51372302e-01
-1.25939012e-01 -5.21044672e-01 -4.68971819e-01 -8.32539022e-01
1.80949196e-02 -6.18091702e-01 5.67294240e-01 9.20745552e-01
4.94872630e-01 8.23341668e-01 -1.14614062e-01 3.59647237e-02
6.06911659e-01 5.88779628e-01 7.83929765e-01 -1.52625477e+00
-3.22630495e-01 -5.44564664e-01 -4.06384170e-01 -7.66975820e-01
5.28382659e-01 -1.40778863e+00 3.41874242e-01 -1.67794490e+00
6.43473744e-01 6.55613542e-02 -7.70490393e-02 9.95787263e-01
-2.24806398e-01 1.86170861e-01 5.08462310e-01 6.15024753e-02
-8.38522673e-01 6.48324639e-02 8.32458138e-01 -2.97821045e-01
2.91059434e-01 -4.88056004e-01 -1.04652059e+00 1.12582946e+00
5.24347782e-01 -3.86776507e-01 -4.65314895e-01 -3.28504562e-01
4.40022200e-01 -4.68265265e-02 8.15365195e-01 -8.63230169e-01
2.26197928e-01 -2.48201743e-01 2.38178313e-01 -1.62000492e-01
1.93347931e-01 -6.51897252e-01 -1.43476054e-01 6.31205678e-01
-7.97171831e-01 -8.45124421e-04 -1.54651865e-01 6.27222538e-01
-4.47768718e-02 -2.15165347e-01 5.23483157e-01 -2.26586431e-01
-1.12392426e+00 -2.11288810e-01 -4.17577356e-01 1.16008267e-01
8.62365782e-01 -3.67253304e-01 -7.43897438e-01 -3.69965404e-01
-8.50911319e-01 2.64821678e-01 5.59585810e-01 5.37431121e-01
6.33820891e-01 -1.12824380e+00 -5.23489416e-01 2.18339831e-01
6.90653920e-01 -4.26432580e-01 1.87828168e-02 8.10761809e-01
-2.68539459e-01 1.38732418e-01 -2.20393151e-01 -5.31608284e-01
-1.48134100e+00 8.09800625e-01 -8.62376913e-02 2.13928014e-01
-8.66350234e-01 9.82226133e-01 5.21908700e-01 -4.18168128e-01
1.67664930e-01 -6.44183576e-01 -1.73875600e-01 1.68673828e-01
6.12035334e-01 -1.09999515e-02 -3.37902516e-01 -4.39208597e-01
-4.49988753e-01 4.71467614e-01 8.51819441e-02 -8.66471883e-03
1.26385498e+00 -1.08673707e-01 -1.58923849e-01 7.76713908e-01
8.23941171e-01 1.36454046e-01 -9.20580328e-01 -5.23266830e-02
4.51348722e-01 -4.37169075e-01 -3.71874630e-01 -1.06409287e+00
-3.27618897e-01 1.27979171e+00 -3.10804006e-02 5.27136683e-01
6.37254715e-01 5.98126113e-01 4.11970139e-01 5.81401706e-01
1.98786035e-01 -6.26920104e-01 3.33690375e-01 6.70305431e-01
9.99733806e-01 -1.38688350e+00 -8.24357048e-02 -6.87401950e-01
-1.10066700e+00 1.38320446e+00 4.88661528e-01 -3.29990871e-02
1.90359190e-01 2.56843001e-01 1.76141113e-01 -5.11358917e-01
-1.19461310e+00 -5.15118957e-01 4.65006202e-01 3.89963120e-01
6.81385756e-01 2.58511513e-01 -3.15671206e-01 1.07607615e+00
-5.84886789e-01 -2.28238285e-01 7.39728868e-01 5.68268120e-01
-7.81863749e-01 -8.38987768e-01 -3.05803955e-01 3.65153700e-01
-1.74736634e-01 -2.22611591e-01 -8.19558322e-01 8.10123503e-01
2.79891282e-01 1.02178931e+00 1.67209208e-01 -3.14889759e-01
8.55417401e-02 4.78350520e-01 2.25270033e-01 -1.05760860e+00
-5.55363357e-01 -3.86584222e-01 1.27532065e-01 -4.79740620e-01
-2.27439672e-01 -4.54310745e-01 -1.85524428e+00 -2.37427294e-01
8.07230547e-02 1.22244827e-01 2.88561314e-01 1.23298430e+00
1.08983450e-01 5.24830222e-01 -4.13449332e-02 -3.07273328e-01
-7.65321314e-01 -6.27560794e-01 -2.59161979e-01 1.00161195e+00
3.84081483e-01 -7.43909776e-01 -6.87770307e-01 2.10325375e-01] | [9.575538635253906, 6.759335517883301] |
5c55f6b5-aff0-477d-9a49-6ac2b8342920 | learn-to-solve-algebra-word-problems-using | null | null | https://aclanthology.org/D15-1096 | https://aclanthology.org/D15-1096.pdf | Learn to Solve Algebra Word Problems Using Quadratic Programming | null | ['Li-Wei Chen', 'Shuaixiang Dai', 'Lipu Zhou'] | 2015-09-01 | null | null | null | emnlp-2015-9 | ['math-word-problem-solving', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'reasoning', 'time-series'] | [-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] | [-7.248019218444824, 3.773810625076294] |
9bb1bd38-5b85-4017-a20e-74d507fb530b | automated-mobile-attention-kpconv-networks | null | null | https://openreview.net/forum?id=VZC5Lzyl0le | https://openreview.net/pdf?id=VZC5Lzyl0le | Automated Mobile Attention KPConv Networks via A Wide & Deep Predictor | Kernel Point Convolution (KPConv) achieves cutting-edge performance on 3D point cloud applications. Unfortunately, the large size of KPConv network limits its usage in mobile scenarios. In addition, we observe that KPConv ignores the kernel relationship and treats each kernel point equally when formulating neighbor-kernel correlation via Euclidean distance. This leads to a weak representation power. To mitigate the above issues, we propose a module named Mobile Attention Kernel Point Convolution (MAKPConv) to improve the efficiency and quality of KPConv. MAKPConv employs a depthwise kernel to reduce resource consumption and re-calibrates the contribution of kernel points towards each neighbor point via Neighbor-Kernel attention to improve representation power. Furthermore, we capitalize Inverted Residual Bottleneck (IRB) to craft a design space and employ a predictor-based Neural Architecture Search (NAS) approach to automate the design of efficient 3D networks based on MAKPConv. To fully exploit the immense design space via an accurate predictor, we identify the importance of carrying feature engineering on searchable features to improve neural architecture representations and propose a Wide & Deep Predictor to unify dense and sparse neural architecture representations for lower error in performance prediction. Experimental evaluations show that our NAS-crafted MAKPConv network uses 96% fewer parameters on 3D point cloud classification and segmentation benchmarks with better performance. Compared with state-of-the-art NAS-crafted model SPVNAS, our NAS-crafted MAKPConv network achieves ~1% better mIOU with 83% fewer parameters and 52% fewer Multiply-Accumulates. | ['Yiran Chen', 'Hai Li', 'Feng Yan', 'Mingyuan Ma', 'Tunhou Zhang'] | 2021-09-29 | null | null | null | null | ['point-cloud-classification'] | ['computer-vision'] | [-3.84895861e-01 -2.77187258e-01 -3.30564976e-01 -2.04260588e-01
-3.95357102e-01 -4.17235672e-01 8.04043189e-02 -2.10208386e-01
-1.98482975e-01 -3.26334573e-02 1.89302508e-02 -8.89463484e-01
-2.94928551e-01 -8.89242709e-01 -9.73173559e-01 -2.65806675e-01
9.21077654e-02 2.09067613e-01 2.93223828e-01 -1.79866314e-01
3.47854257e-01 1.03471410e+00 -1.42941761e+00 4.01975602e-01
7.00360656e-01 1.28381288e+00 3.46217930e-01 4.95281905e-01
-4.36100602e-01 4.58164871e-01 -4.23042476e-01 -2.51429588e-01
5.68489254e-01 3.66700381e-01 -6.90927088e-01 -4.27740157e-01
6.56756282e-01 -3.94720227e-01 -5.76500177e-01 6.41822100e-01
5.97848117e-01 9.18572955e-03 6.85019553e-01 -1.28254807e+00
-7.83898294e-01 5.22847772e-01 -7.47636914e-01 5.13055682e-01
-4.20558274e-01 3.40969861e-01 9.04935360e-01 -1.19387543e+00
-2.88068838e-02 1.01917291e+00 1.13759935e+00 2.97185808e-01
-1.07826197e+00 -9.03191149e-01 2.62615532e-01 -3.35479081e-02
-1.67475545e+00 -2.86642879e-01 7.51110852e-01 -2.26128846e-01
1.58544815e+00 4.57760274e-01 6.57389104e-01 7.83731043e-01
1.49716198e-01 7.74088025e-01 3.08936357e-01 -1.22919165e-01
1.91196516e-01 -1.61180869e-02 3.46488297e-01 8.06309879e-01
2.44994149e-01 5.97721189e-02 -2.88843781e-01 -1.51655823e-01
1.37320435e+00 3.46902609e-01 -2.49318525e-01 -4.24757063e-01
-1.02946734e+00 7.00984001e-01 9.79964375e-01 9.08871591e-02
-4.07648832e-01 6.35501921e-01 4.42495227e-01 2.37249404e-01
2.01878011e-01 5.75433850e-01 -8.13143373e-01 -2.16036618e-01
-8.89726758e-01 1.63137719e-01 5.06139159e-01 1.15541971e+00
7.48798013e-01 3.59899133e-01 -1.45536378e-01 9.41094995e-01
4.69708890e-02 5.21527827e-01 3.10600638e-01 -8.92437935e-01
6.10270739e-01 1.05643392e+00 -3.09017867e-01 -1.01800609e+00
-3.69668007e-01 -7.33866394e-01 -7.36062884e-01 3.16930652e-01
-9.50550735e-02 1.42357424e-01 -1.00330591e+00 1.18594444e+00
2.26290926e-01 6.84805632e-01 -2.06690595e-01 1.02476847e+00
8.60944986e-01 7.05610335e-01 -1.29108205e-01 5.00548422e-01
1.14720178e+00 -1.11549413e+00 2.19572201e-01 -7.22915307e-02
8.47748697e-01 -4.72911268e-01 1.32264292e+00 4.56657149e-02
-9.73974824e-01 -7.55115747e-01 -1.04236400e+00 -1.87257499e-01
-3.38857949e-01 3.54476541e-01 1.05259037e+00 6.26657903e-01
-1.12503219e+00 7.30777383e-01 -1.06010544e+00 -3.88120674e-03
7.07197607e-01 1.02262807e+00 -7.89062083e-02 4.16438915e-02
-5.22004664e-01 5.08977890e-01 3.17967206e-01 5.66216558e-02
-5.59858322e-01 -1.36884260e+00 -6.39309824e-01 3.58495504e-01
1.05525970e-01 -7.83319294e-01 9.89391923e-01 -6.75723314e-01
-1.43549883e+00 3.26926261e-01 -5.79638928e-02 -4.61780041e-01
-6.88609034e-02 -3.80038530e-01 -2.87758052e-01 -5.27244918e-02
-2.53091663e-01 9.81246233e-01 7.64962852e-01 -8.31034839e-01
-6.53314054e-01 -1.72506839e-01 7.74942860e-02 1.16680443e-01
-5.31376004e-01 -3.22510719e-01 -9.98547316e-01 -7.19072163e-01
1.47024021e-01 -1.11692262e+00 -2.06769660e-01 5.09429770e-03
-2.04769641e-01 -3.36711228e-01 1.14195383e+00 -1.35714635e-01
1.37171364e+00 -2.41645288e+00 -2.24935830e-01 3.73628318e-01
4.22980011e-01 6.27909124e-01 -3.34040046e-01 -4.45561334e-02
-2.03172684e-01 1.41715646e-01 2.16457918e-01 -3.42659324e-01
-1.77804798e-01 2.79672444e-01 -4.84837681e-01 3.38281244e-01
3.79835755e-01 1.20393002e+00 -3.51252466e-01 -1.45412222e-01
3.83423805e-01 7.42018759e-01 -1.08675861e+00 -1.82543188e-01
-1.08593129e-01 -2.71480709e-01 -6.17989302e-01 9.90733504e-01
7.98771203e-01 -6.57828689e-01 -4.02592361e-01 -5.69739938e-01
-1.17800750e-01 1.66763008e-01 -9.06481445e-01 1.70635319e+00
-6.56778693e-01 5.43020606e-01 -1.70799389e-01 -7.28939056e-01
1.08118415e+00 -1.05883963e-01 4.07156050e-01 -5.09495020e-01
3.02354574e-01 2.46143758e-01 -1.97526008e-01 2.33972748e-03
6.22726560e-01 5.01448452e-01 5.85622452e-02 9.89132300e-02
-2.86451187e-02 2.40310296e-01 -5.79293311e-01 3.26929614e-02
1.29900515e+00 -5.93757555e-02 -2.25560799e-01 -3.41523468e-01
3.45008492e-01 1.25617176e-01 2.95603514e-01 7.15456843e-01
-7.87889659e-02 6.50579512e-01 2.72740930e-01 -8.53222013e-01
-9.65777874e-01 -9.22821224e-01 -2.83070862e-01 1.23447442e+00
1.57475874e-01 -4.90076691e-01 -5.79303145e-01 -7.05723107e-01
4.80220288e-01 7.78625786e-01 -6.19037807e-01 -2.28816345e-01
-7.85133243e-01 -6.01144373e-01 7.69666493e-01 1.07216883e+00
4.04426664e-01 -7.39918351e-01 -5.25691211e-01 5.97490296e-02
5.59585869e-01 -8.60782027e-01 -6.09550297e-01 5.13943017e-01
-1.10891664e+00 -6.25006199e-01 -4.51023430e-01 -7.77878344e-01
6.42762601e-01 5.52882075e-01 1.02789330e+00 1.69027701e-01
-6.95929006e-02 2.17067719e-01 -2.76745260e-01 -3.78256112e-01
1.15335248e-01 5.35276115e-01 1.07234091e-01 -4.96066004e-01
6.65942252e-01 -8.29206169e-01 -8.33510041e-01 4.58842546e-01
-5.49040735e-01 1.12754963e-01 9.18138981e-01 7.68215775e-01
7.93136239e-01 -1.54559866e-01 2.53584862e-01 -4.99377698e-01
4.14213896e-01 -5.53679883e-01 -7.11549759e-01 -5.84082901e-02
-8.57824504e-01 2.82122523e-01 5.68356991e-01 -6.45591617e-01
-3.56054902e-01 1.45391941e-01 -1.33643672e-01 -1.37693048e+00
8.44080597e-02 2.46255040e-01 7.96007738e-02 -5.53831995e-01
7.20857024e-01 1.83416665e-01 1.29248545e-01 -6.61889255e-01
3.47124457e-01 5.67108691e-01 5.24448156e-01 -6.54363871e-01
6.78097367e-01 2.23762080e-01 8.23919848e-02 -6.37203336e-01
-3.46308798e-01 -4.23877954e-01 -3.67779881e-01 2.70377010e-01
6.78420901e-01 -1.12281573e+00 -9.91058350e-01 9.48723480e-02
-1.08055651e+00 -3.18361193e-01 -2.33894214e-01 1.90849394e-01
-1.32296532e-01 -9.67176482e-02 -6.09946132e-01 -4.43822622e-01
-8.46120059e-01 -1.40632546e+00 1.23136234e+00 1.27377793e-01
-2.00817198e-01 -6.16115808e-01 -3.52552116e-01 1.96416423e-01
7.25643277e-01 -1.60951361e-01 1.10322881e+00 -6.80164635e-01
-9.24405396e-01 -1.96230054e-01 -7.58558691e-01 3.17057103e-01
-1.39956206e-01 -1.66100770e-01 -8.68793905e-01 -2.32874945e-01
-3.41843605e-01 2.19727069e-01 7.77672350e-01 4.76673186e-01
1.60913837e+00 -6.52857721e-02 -5.06591022e-01 1.29581630e+00
1.44764447e+00 4.91114482e-02 3.93353164e-01 3.92857760e-01
1.15525699e+00 -1.56141147e-01 1.31869853e-01 2.45736644e-01
3.20997179e-01 6.21486604e-01 7.12174416e-01 -9.15141553e-02
-2.34352291e-01 -2.12990642e-01 1.50789037e-01 9.17931736e-01
-1.41415000e-01 2.63845138e-02 -1.03978264e+00 4.75904852e-01
-1.66429448e+00 -4.07962292e-01 -2.95754001e-02 1.84298086e+00
4.51616585e-01 3.44998837e-01 -1.23772793e-01 -1.66349262e-01
3.50393236e-01 2.67293192e-02 -7.76996493e-01 -4.96640772e-01
1.12818085e-01 4.79695976e-01 9.48401690e-01 8.76251161e-02
-1.05134106e+00 1.07393801e+00 5.73879099e+00 1.17459834e+00
-1.59840906e+00 -4.27416153e-03 6.92237377e-01 -4.71135736e-01
-2.22000673e-01 -3.12158853e-01 -1.22247875e+00 4.09876257e-01
1.03700626e+00 3.41075867e-01 4.07414407e-01 1.54889858e+00
-7.12426454e-02 4.66227621e-01 -1.15238953e+00 1.38532794e+00
-2.61458814e-01 -2.01514363e+00 2.40711808e-01 2.78406382e-01
4.12800401e-01 6.41447246e-01 2.67345279e-01 5.69688857e-01
4.89720888e-02 -1.29483235e+00 6.88527286e-01 1.54149681e-01
1.00412214e+00 -1.11486101e+00 6.86962008e-01 8.73206556e-02
-1.33862710e+00 -2.17860892e-01 -6.92179620e-01 -5.05884513e-02
-2.70089537e-01 4.15627092e-01 -1.01968920e+00 1.89179823e-01
9.21029806e-01 5.34941256e-01 -6.24731839e-01 9.37861919e-01
3.86224687e-01 5.96464694e-01 -5.27118444e-01 -1.84253708e-01
5.28909266e-01 1.42129734e-01 3.40288967e-01 1.14013875e+00
4.48449850e-01 9.56757888e-02 -8.16626400e-02 9.71338212e-01
-1.37995437e-01 -1.01903684e-01 -4.51078355e-01 1.21792540e-01
8.32756877e-01 1.05421138e+00 -6.47823453e-01 -3.75242531e-02
-5.67735195e-01 1.06353486e+00 5.29525995e-01 2.51470983e-01
-1.01068294e+00 -2.98242211e-01 1.25768149e+00 2.66540378e-01
9.52884555e-01 -2.29586273e-01 -7.17582703e-01 -8.10103476e-01
-1.79176569e-01 -5.60969353e-01 -2.38201842e-02 -6.93102658e-01
-1.11070347e+00 8.54313850e-01 -1.66852072e-01 -1.11175299e+00
2.28612736e-01 -8.48899841e-01 -6.65685058e-01 1.05579209e+00
-1.39312923e+00 -1.22159374e+00 -3.24930876e-01 6.18273854e-01
6.00622952e-01 -4.58378255e-01 6.85136259e-01 4.61247444e-01
-6.81514144e-01 9.57356930e-01 -1.21162198e-01 1.04813889e-01
1.65913314e-01 -9.88946915e-01 9.66173470e-01 3.60810041e-01
3.98252383e-02 1.00598967e+00 3.44992615e-02 -4.61182505e-01
-1.99582183e+00 -1.51125836e+00 1.65906519e-01 -7.15676486e-01
4.80589867e-01 -1.75568759e-01 -9.51027215e-01 5.93000889e-01
-3.24888349e-01 5.17168641e-01 6.61691487e-01 2.80678660e-01
-4.53440577e-01 -3.26354891e-01 -8.01533520e-01 6.52361751e-01
1.26220214e+00 -4.37274516e-01 -3.48640800e-01 -1.06334560e-01
1.38101399e+00 -5.73366821e-01 -8.56833756e-01 7.36933529e-01
4.40174878e-01 -8.54042232e-01 1.37043774e+00 -4.65642959e-01
3.26291949e-01 -3.95823181e-01 -3.01851481e-01 -8.14756930e-01
-7.15117514e-01 -2.26279631e-01 -4.14139241e-01 8.79234612e-01
5.57357073e-01 -4.63370860e-01 1.29943681e+00 6.58440530e-01
-6.18388295e-01 -1.58764791e+00 -9.13855493e-01 -7.19097197e-01
4.19901386e-02 -8.66470516e-01 9.92396593e-01 9.27924633e-01
-5.92401505e-01 2.24167198e-01 -5.86578287e-02 4.44011748e-01
1.78991854e-01 1.69966191e-01 8.44529510e-01 -1.02201462e+00
-3.81697834e-01 -6.67438805e-01 -5.72791517e-01 -1.41094673e+00
-1.39094219e-02 -9.53943908e-01 -4.87752169e-01 -9.69039261e-01
-1.56042323e-01 -1.02488780e+00 -4.96366024e-01 6.37013793e-01
3.50473113e-02 2.96911091e-01 1.63088337e-01 3.66019160e-01
-4.53315735e-01 4.47140574e-01 1.05100179e+00 8.67733806e-02
-5.22903860e-01 -8.89197662e-02 -8.96885335e-01 7.42651880e-01
5.84545493e-01 -2.61482924e-01 -5.75421572e-01 -9.30276334e-01
1.42010301e-01 -2.58998692e-01 4.41785336e-01 -1.22935963e+00
4.14103419e-01 2.76152547e-02 6.93364620e-01 -9.70829129e-01
4.74244624e-01 -9.33873653e-01 1.82458237e-01 2.29642153e-01
1.46105573e-01 4.65086043e-01 6.34453475e-01 4.23679531e-01
7.00864270e-02 9.54062957e-03 4.59999084e-01 3.97396162e-02
-9.92121160e-01 5.64392984e-01 1.92712262e-01 -4.69692051e-01
9.16898549e-01 -5.86635292e-01 -2.88652033e-01 2.70060599e-01
-2.40148515e-01 1.27096206e-01 5.38149655e-01 3.57430428e-01
9.65967596e-01 -1.36179030e+00 -1.98321521e-01 5.00136077e-01
-9.68257114e-02 4.27635372e-01 2.65410513e-01 6.54265463e-01
-1.11360753e+00 6.29813015e-01 -9.31780562e-02 -7.80427337e-01
-9.80404675e-01 5.90689421e-01 3.15271944e-01 -4.70033614e-03
-9.04224873e-01 1.31558967e+00 2.02281103e-01 -4.81013298e-01
3.50107551e-01 -7.10593820e-01 1.49682820e-01 -5.18007398e-01
4.07314152e-01 4.01778728e-01 3.11081856e-01 -3.36378634e-01
-5.39804280e-01 5.90772569e-01 -2.86366940e-01 5.90192616e-01
1.44136786e+00 3.89410883e-01 7.12346584e-02 -1.18032500e-01
1.46908951e+00 2.09831055e-02 -1.33979166e+00 -7.05554634e-02
-3.26474339e-01 -3.86176318e-01 6.47685945e-01 -6.70922160e-01
-1.44347429e+00 6.86953127e-01 7.40786612e-01 -2.53859490e-01
1.11334801e+00 3.62962000e-02 1.14294457e+00 5.40422857e-01
2.49393687e-01 -6.85664177e-01 -5.90264015e-02 7.55736887e-01
7.02200651e-01 -1.00878608e+00 -4.14693449e-03 -6.08349979e-01
-4.41718221e-01 1.07995737e+00 9.84438419e-01 -4.95537907e-01
9.04244006e-01 4.48898882e-01 -2.36241162e-01 -4.05145168e-01
-7.31629908e-01 1.88907832e-01 5.29387414e-01 4.14052755e-01
1.04936011e-01 1.51166739e-02 4.76948708e-01 9.53025997e-01
-3.57538491e-01 -9.93905142e-02 -1.07571073e-01 8.19383204e-01
-2.74887860e-01 -8.38451326e-01 -2.14231491e-01 7.19324052e-01
-2.20692053e-01 -5.30749857e-01 2.04163212e-02 7.41335332e-01
1.56738713e-01 2.50617027e-01 4.63669837e-01 -9.51006413e-01
4.34316069e-01 -1.86863169e-01 1.07275926e-01 -4.33464289e-01
-9.37439322e-01 -5.22724502e-02 -2.58451939e-01 -8.33153069e-01
2.88795114e-01 -2.33662575e-01 -1.27549589e+00 -6.72197521e-01
-4.07771885e-01 -1.57359660e-01 8.83698404e-01 5.64608455e-01
1.27357829e+00 7.71396518e-01 3.76889646e-01 -9.60287094e-01
-5.30160844e-01 -5.59215665e-01 -2.73229927e-01 -1.15689516e-01
2.00318709e-01 -5.98635137e-01 -1.30862847e-01 -5.49457967e-01] | [7.883116722106934, -3.585314989089966] |
58062e39-ff7c-4965-b720-aeeef65e9a64 | unsupervised-image-semantic-segmentation | 2210.11810 | null | https://arxiv.org/abs/2210.11810v1 | https://arxiv.org/pdf/2210.11810v1.pdf | Unsupervised Image Semantic Segmentation through Superpixels and Graph Neural Networks | Unsupervised image segmentation is an important task in many real-world scenarios where labelled data is of scarce availability. In this paper we propose a novel approach that harnesses recent advances in unsupervised learning using a combination of Mutual Information Maximization (MIM), Neural Superpixel Segmentation and Graph Neural Networks (GNNs) in an end-to-end manner, an approach that has not been explored yet. We take advantage of the compact representation of superpixels and combine it with GNNs in order to learn strong and semantically meaningful representations of images. Specifically, we show that our GNN based approach allows to model interactions between distant pixels in the image and serves as a strong prior to existing CNNs for an improved accuracy. Our experiments reveal both the qualitative and quantitative advantages of our approach compared to current state-of-the-art methods over four popular datasets. | ['Eran Treister', 'Nir Ben Zikri', 'Moshe Eliasof'] | 2022-10-21 | null | null | null | null | ['unsupervised-semantic-segmentation', 'superpixels'] | ['computer-vision', 'computer-vision'] | [ 5.66612840e-01 5.66225171e-01 3.49709429e-02 -3.55892569e-01
-4.29442525e-01 -3.92133087e-01 6.32163882e-01 2.61003733e-01
-6.53485298e-01 6.14547133e-01 8.62417519e-02 -2.02148169e-01
-2.56100237e-01 -7.07216084e-01 -8.12107682e-01 -5.48912108e-01
-6.80802464e-02 5.43464422e-01 6.81769073e-01 -4.69826721e-02
1.05087668e-01 5.15061796e-01 -1.44766450e+00 1.20184653e-01
9.87405896e-01 8.67765009e-01 2.66558737e-01 5.95497906e-01
-1.56156614e-01 8.91522944e-01 -9.26376805e-02 -2.06338838e-01
3.74926090e-01 -3.82402122e-01 -1.20136333e+00 4.84703183e-01
3.88108671e-01 5.33336140e-02 -5.27239628e-02 1.02903640e+00
1.76925719e-01 2.39548445e-01 5.70438325e-01 -7.82935917e-01
-2.77160436e-01 6.93965375e-01 -7.83859074e-01 3.84864844e-02
-1.68284789e-01 1.12602048e-01 1.13325131e+00 -3.78436595e-01
8.24403107e-01 1.02265358e+00 5.88291228e-01 2.55232006e-01
-1.50850677e+00 -1.95312679e-01 1.32447392e-01 -1.09436348e-01
-1.14739978e+00 -1.61844641e-01 8.32145154e-01 -3.69982123e-01
7.79089034e-01 -1.13660209e-02 7.37253904e-01 6.53947830e-01
-3.04148972e-01 1.06695962e+00 1.03912663e+00 -6.35437965e-01
2.15023488e-01 -5.64542711e-02 1.91665798e-01 7.37031519e-01
2.01848716e-01 -1.00475177e-01 -3.91664416e-01 2.89996415e-01
9.89418566e-01 5.47549017e-02 -3.91303897e-02 -7.51091182e-01
-1.12144709e+00 7.63908565e-01 9.09989357e-01 6.38824701e-01
-6.38111711e-01 4.37822431e-01 -9.60639864e-03 -1.54801860e-01
6.63920581e-01 4.32977229e-01 -3.40790898e-01 1.31648362e-01
-1.41472983e+00 4.36598733e-02 8.82531106e-01 6.00651681e-01
1.02934980e+00 -2.27035388e-01 4.64012995e-02 6.75804853e-01
5.17904639e-01 -7.12035522e-02 1.52885571e-01 -9.90664482e-01
1.05443753e-01 9.63675022e-01 -3.30286801e-01 -8.70011270e-01
-5.15837371e-01 -6.76578522e-01 -8.25858116e-01 2.42725611e-01
5.01089334e-01 -8.11904296e-02 -1.28433239e+00 1.54849243e+00
3.35044026e-01 3.73240530e-01 -6.86313361e-02 7.21947968e-01
5.94390631e-01 2.81719148e-01 2.19783202e-01 1.73554465e-01
1.05969787e+00 -1.23619759e+00 -3.85477811e-01 -4.37107474e-01
5.47373295e-01 -5.93122125e-01 6.00170910e-01 4.72839803e-01
-1.06136787e+00 -4.08293307e-01 -8.56610119e-01 -7.57207945e-02
-6.35923862e-01 -1.39828742e-01 9.11673963e-01 4.64805484e-01
-1.44772935e+00 8.99044633e-01 -9.78047013e-01 -7.49005139e-01
9.92421627e-01 5.60354054e-01 -3.83785605e-01 7.51411030e-03
-7.03792095e-01 5.78785837e-01 9.65738535e-01 2.07890883e-01
-6.27684057e-01 -3.53400171e-01 -8.47917795e-01 -1.49656385e-01
6.94372594e-01 -7.78348684e-01 1.04683113e+00 -1.18558347e+00
-1.36502159e+00 1.02559280e+00 2.39586681e-01 -8.46239746e-01
5.77396572e-01 -2.47640252e-01 1.95077762e-01 4.24795777e-01
-1.92509875e-01 1.23651731e+00 5.34157395e-01 -1.43722892e+00
-5.96059322e-01 -3.05888295e-01 1.56932786e-01 1.09706499e-01
-2.80376554e-01 -2.21105441e-01 -7.10858822e-01 -4.31558430e-01
2.24677548e-01 -8.50439012e-01 -7.56591260e-01 -1.32130310e-01
-7.76649535e-01 -3.85558940e-02 7.47205496e-01 -5.25029063e-01
8.53997171e-01 -1.82255054e+00 3.35164815e-01 4.18529481e-01
4.10449088e-01 4.82599705e-01 -4.88545187e-03 5.35077035e-01
7.34287426e-02 1.41468644e-01 -8.62109959e-01 -6.04685605e-01
-8.28180090e-03 4.41483110e-01 4.09979969e-01 3.30257893e-01
4.11249906e-01 1.28331029e+00 -9.71576273e-01 -5.94043553e-01
4.58299518e-01 4.57402706e-01 -3.99074644e-01 1.69213563e-01
-5.69277108e-01 6.34128511e-01 -3.91488105e-01 4.16112065e-01
5.97334743e-01 -3.78783643e-01 2.46884987e-01 -7.04796389e-02
-1.57937557e-01 -5.74326999e-02 -1.01795328e+00 2.10721040e+00
-1.17839858e-01 4.66083974e-01 -4.35689650e-02 -1.33408570e+00
6.83485985e-01 4.63069566e-02 5.40479720e-01 -6.19176269e-01
2.76920497e-01 1.34323657e-01 -1.34966269e-01 -4.28152919e-01
2.39564583e-01 -4.00254913e-02 2.89290369e-01 3.63137484e-01
4.65524793e-01 -1.25475526e-01 5.30032337e-01 3.34988713e-01
1.03490019e+00 4.42791551e-01 2.28726581e-01 -5.17272770e-01
3.05156171e-01 -7.39726424e-03 1.41894072e-01 8.30776751e-01
7.19293430e-02 9.40806568e-01 4.63357180e-01 -1.69644058e-01
-9.07090485e-01 -7.32728124e-01 2.69825488e-01 8.68513405e-01
6.29270896e-02 -1.58725977e-01 -1.23287809e+00 -9.20969307e-01
-2.08115950e-01 5.26910782e-01 -6.77320242e-01 1.99441582e-01
-5.27649164e-01 -7.05876350e-01 3.09946924e-01 6.15923941e-01
5.89919984e-01 -1.31688273e+00 -7.23020434e-01 1.73842758e-01
3.95889841e-02 -1.44287300e+00 8.42704996e-02 3.95562410e-01
-1.11089897e+00 -1.11014068e+00 -7.31703460e-01 -8.12224090e-01
8.31533551e-01 2.29396820e-01 1.41193819e+00 1.23846397e-01
-3.95520508e-01 3.37198228e-01 -3.32478404e-01 -2.80795604e-01
-1.01401448e-01 3.62065881e-01 -5.94022334e-01 2.45945156e-01
1.80356175e-01 -7.75905192e-01 -7.48091757e-01 -6.67902455e-02
-1.39387655e+00 3.53339970e-01 9.02082324e-01 4.96936589e-01
6.62890553e-01 -1.38608478e-02 1.93056524e-01 -1.52354610e+00
3.36226821e-01 -5.60068548e-01 -7.21820056e-01 1.00419611e-01
-4.92296100e-01 1.77592844e-01 2.55894065e-01 3.05034723e-02
-8.59071076e-01 4.61566895e-01 -2.19449818e-01 -1.78111181e-01
-5.46711087e-01 7.12465465e-01 3.07785142e-02 -2.76947677e-01
4.65416133e-01 3.39579955e-02 1.46406949e-01 -7.34767437e-01
6.82686627e-01 3.69663954e-01 6.72974765e-01 -2.72529215e-01
6.69206560e-01 7.35799372e-01 1.60626993e-01 -8.13866913e-01
-9.34043169e-01 -9.22789812e-01 -1.13874125e+00 -2.03466415e-01
1.21676683e+00 -6.34087682e-01 -4.46190864e-01 5.64425647e-01
-1.03722560e+00 -6.48843229e-01 -3.50464255e-01 3.13769668e-01
-5.96289277e-01 5.45956135e-01 -4.67125565e-01 -7.32846797e-01
-1.49031997e-01 -9.92059648e-01 1.07421374e+00 5.22958875e-01
-2.11414266e-02 -1.24904299e+00 -4.99257706e-02 5.95946848e-01
3.48488420e-01 6.62126660e-01 5.09711206e-01 -7.84704268e-01
-8.23371649e-01 -4.84931841e-02 -6.04690254e-01 4.91707504e-01
-4.85824607e-02 5.64377755e-03 -8.89475822e-01 -2.32177693e-02
-2.92345434e-01 -2.31227249e-01 1.24288785e+00 5.01090825e-01
1.03637850e+00 -5.27996570e-02 -3.01919311e-01 5.21115303e-01
1.74162149e+00 -2.88603932e-01 8.38879704e-01 3.55434448e-01
1.00203133e+00 8.74524474e-01 3.06679755e-01 9.44607556e-02
4.26679641e-01 3.31705481e-01 8.14491093e-01 -7.66990483e-01
-2.48986572e-01 -4.96548824e-02 -2.77492970e-01 5.58862031e-01
-1.84240103e-01 -3.80442113e-01 -1.09813297e+00 9.61849630e-01
-2.33714986e+00 -4.83195275e-01 -3.41558844e-01 2.01500320e+00
6.28997624e-01 3.14825684e-01 2.06247345e-01 -6.22106977e-02
6.36556327e-01 1.71122640e-01 -2.80747205e-01 -1.48282990e-01
-4.65331413e-02 4.40157145e-01 7.89275348e-01 2.90268570e-01
-1.39004731e+00 1.26862180e+00 6.24922800e+00 7.11790919e-01
-7.39370704e-01 -1.64370015e-02 8.37796748e-01 3.56639773e-01
-9.51233432e-02 2.76232928e-01 -3.00478756e-01 1.48632184e-01
7.28725672e-01 5.36634505e-01 2.85825700e-01 6.19128644e-01
1.81935728e-01 -4.61223871e-01 -9.27529097e-01 6.42156720e-01
-1.54374698e-02 -1.35985351e+00 -9.43144560e-02 2.17902467e-01
1.09140670e+00 3.75990748e-01 -1.14569671e-01 -2.81522632e-01
7.47252524e-01 -1.07580972e+00 5.65035999e-01 6.86902821e-01
9.24387649e-02 -7.02629328e-01 8.52955937e-01 4.76584911e-01
-9.43697572e-01 2.67576754e-01 -1.70839027e-01 3.84279564e-02
1.54744089e-01 8.74854863e-01 -9.06771660e-01 8.11442912e-01
5.38342774e-01 7.40880847e-01 -7.83587813e-01 1.25888050e+00
-4.41041976e-01 8.07730138e-01 -6.31575584e-01 1.69739783e-01
9.19883907e-01 -3.54228526e-01 3.24811041e-01 1.40761685e+00
-8.42328221e-02 -1.85645968e-01 2.61226624e-01 1.14708066e+00
-2.51610637e-01 1.00003429e-01 -5.45412660e-01 -3.03131342e-01
-1.71277866e-01 1.57989717e+00 -1.49879932e+00 -2.62916327e-01
-2.80401707e-01 1.09925735e+00 5.49757481e-01 4.14823830e-01
-4.82710272e-01 -1.22646689e-01 8.83495882e-02 1.29988054e-02
7.66240478e-01 -3.75003904e-01 -1.93586439e-01 -7.47153878e-01
-1.40999705e-01 -5.76972842e-01 4.34430689e-01 -5.62146127e-01
-1.14097404e+00 4.66540217e-01 -2.07712296e-02 -7.94417560e-01
-2.87249953e-01 -3.57110620e-01 -5.89876890e-01 5.96423328e-01
-1.80085182e+00 -1.67347169e+00 -2.37262547e-01 4.02425975e-01
4.03869361e-01 3.19458395e-01 5.25347948e-01 7.00303093e-02
-5.23474693e-01 3.13553697e-04 1.49157330e-01 2.47811899e-01
1.21293925e-01 -1.62145305e+00 5.24094164e-01 1.12846136e+00
5.92895269e-01 4.80997115e-01 6.13733947e-01 -5.70673943e-01
-8.59958708e-01 -1.18397582e+00 7.35882699e-01 -1.52783498e-01
6.03881478e-01 -4.17996615e-01 -8.70270729e-01 5.92140079e-01
5.77777445e-01 3.27449918e-01 5.87370157e-01 5.33303097e-02
-9.29076597e-02 1.67130917e-01 -9.45744812e-01 4.69114661e-01
9.94310141e-01 -2.82415003e-01 -2.86270231e-01 2.59409279e-01
6.15444481e-01 -9.31853130e-02 -6.36276007e-01 3.38410258e-01
2.72552103e-01 -1.26273596e+00 9.39683795e-01 -4.13465291e-01
4.52849537e-01 -2.90005505e-01 1.65334985e-01 -1.21323943e+00
-7.19558373e-02 -6.64753497e-01 1.33508459e-01 1.26976883e+00
4.69377130e-01 -5.02268314e-01 1.04439473e+00 4.58282650e-01
-1.78332739e-02 -7.55623043e-01 -5.42157352e-01 -6.03604317e-01
-1.26787469e-01 -4.97479379e-01 2.76970267e-01 7.02229857e-01
-2.35192895e-01 9.47739109e-02 -1.92174360e-01 5.34420349e-02
9.19498861e-01 8.13548565e-02 6.52833641e-01 -1.38185990e+00
-3.20970654e-01 -5.72899401e-01 -7.77097464e-01 -9.48306978e-01
3.10484841e-02 -8.43287349e-01 2.18738616e-01 -2.00780535e+00
3.19998860e-01 -2.22605556e-01 -6.58242047e-01 4.83540297e-01
-1.40984252e-01 8.13133061e-01 2.37048283e-01 8.40730369e-02
-1.13250160e+00 2.03132227e-01 1.06147695e+00 -1.25269353e-01
-1.60465300e-01 -1.77708134e-01 -7.32711375e-01 9.35117781e-01
7.01109886e-01 -5.22243440e-01 -3.87057185e-01 -4.23387796e-01
6.15729839e-02 -4.01782423e-01 6.41894758e-01 -1.20290422e+00
4.22224164e-01 1.73517227e-01 2.23832309e-01 -5.21777332e-01
2.74526887e-02 -7.99460649e-01 -1.67905129e-02 2.98388094e-01
-3.38189542e-01 -4.92994130e-01 1.11396998e-01 7.23878622e-01
-3.74988675e-01 -2.90931493e-01 7.78202474e-01 -5.46696663e-01
-8.95268083e-01 2.53471464e-01 -9.45683196e-02 -6.13561831e-02
9.86742854e-01 -3.73228997e-01 6.24113418e-02 -2.44031042e-01
-7.09561348e-01 2.82448679e-01 4.55248892e-01 1.00794345e-01
4.48638171e-01 -8.84094596e-01 -6.00029230e-01 -1.26194865e-01
1.34223420e-02 4.24802393e-01 1.13253474e-01 1.08150887e+00
-8.19151342e-01 2.77658433e-01 -1.38211787e-01 -8.03454757e-01
-1.13233995e+00 2.99755603e-01 1.05375886e-01 -6.10134721e-01
-5.49382150e-01 1.03605068e+00 1.94442123e-01 -4.27753270e-01
1.42536923e-01 -2.43880555e-01 -4.40055490e-01 7.93658122e-02
-3.05502824e-02 5.62687591e-02 1.06931098e-01 -7.83852160e-01
-7.80393556e-02 4.10825551e-01 -1.68004453e-01 -1.08918987e-01
1.63855231e+00 -2.75072902e-01 -3.44885081e-01 2.98195273e-01
1.24786484e+00 -5.11746228e-01 -1.46651268e+00 -4.36676294e-01
4.38632697e-01 -3.37997645e-01 2.77570993e-01 -7.27870882e-01
-1.37639296e+00 8.71451795e-01 5.17436624e-01 4.05315131e-01
1.14922845e+00 2.80333996e-01 8.44929874e-01 1.61372527e-01
1.17865771e-01 -1.22489285e+00 8.21815357e-02 2.37225384e-01
3.24115962e-01 -1.42946327e+00 -9.72918123e-02 -7.27389395e-01
-4.57085013e-01 1.08125877e+00 2.32346639e-01 -2.82208949e-01
5.50250649e-01 8.37839469e-02 2.96706017e-02 -5.31744957e-01
-2.49319270e-01 -1.05974066e+00 4.13143188e-01 6.57839417e-01
2.86253452e-01 -8.75849798e-02 -2.49366656e-01 1.24171920e-01
2.80246764e-01 1.11941367e-01 8.55004713e-02 1.18979597e+00
-4.27401334e-01 -1.19084120e+00 4.05439511e-02 3.26052934e-01
-6.04058027e-01 -1.90839991e-01 -6.69635296e-01 8.05910587e-01
3.01274329e-01 8.32776606e-01 -1.39152840e-01 -1.30730316e-01
3.62393595e-02 -5.81311472e-02 4.77677286e-01 -7.13582337e-01
-6.47974014e-01 2.30293140e-01 -8.26449469e-02 -7.21608579e-01
-9.39660251e-01 -5.26267231e-01 -1.33899438e+00 3.99675250e-01
-4.54673469e-01 -1.95391506e-01 8.38724434e-01 1.28241849e+00
2.67233372e-01 6.95241451e-01 2.04723775e-01 -1.24459577e+00
8.90610144e-02 -7.99902081e-01 -5.32121897e-01 6.55169129e-01
1.15340568e-01 -4.04285729e-01 -6.75174547e-03 2.55944997e-01] | [9.57740592956543, 0.5221666693687439] |
ed0d4274-ee1c-4afd-87b2-4725b2cfd390 | sentemo-a-multilingual-adaptive-platform-for | null | null | https://aclanthology.org/2022.wassa-1.5 | https://aclanthology.org/2022.wassa-1.5.pdf | SentEMO: A Multilingual Adaptive Platform for Aspect-based Sentiment and Emotion Analysis | In this paper, we present the SentEMO platform, a tool that provides aspect-based sentiment analysis and emotion detection of unstructured text data such as reviews, emails and customer care conversations. Currently, models have been trained for five domains and one general domain and are implemented in a pipeline approach, where the output of one model serves as the input for the next. The results are presented in three dashboards, allowing companies to gain more insights into what stakeholders think of their products and services. The SentEMO platform is available at https://sentemo.ugent.be | ['Veronique Hoste', 'Olivier Parent', 'Pranaydeep Singh', 'Els Lefever', 'Cynthia Van Hee', 'Orphee De Clercq', 'Ellen De Geyndt'] | null | null | null | null | wassa-acl-2022-5 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [-1.09714255e-01 2.94872731e-01 -1.26129776e-01 -8.23939681e-01
-2.67247558e-01 -7.54212439e-01 5.95979810e-01 6.43213809e-01
1.26053095e-01 7.06567690e-02 4.55133766e-01 -3.11043710e-01
3.46913904e-01 -7.46441066e-01 -7.37044364e-02 -2.79631466e-01
5.45415819e-01 3.42718720e-01 -9.33060050e-02 -6.13560915e-01
3.96492183e-01 9.50563625e-02 -1.39996958e+00 8.24616909e-01
3.61754715e-01 1.29199290e+00 -3.22806746e-01 6.56156540e-01
-5.15025020e-01 1.09801507e+00 -5.39043248e-01 -7.79430568e-01
-3.41860838e-02 -1.12785332e-01 -6.56707883e-01 -1.46646723e-02
-3.63902688e-01 -1.02870874e-02 5.77167571e-01 8.81719828e-01
3.06142956e-01 -2.75562435e-01 3.09048176e-01 -1.41829014e+00
-4.55775082e-01 6.79466963e-01 -2.08077371e-01 -9.21476483e-02
7.67690837e-01 1.38287038e-01 1.43331647e+00 -8.71316433e-01
7.16777384e-01 1.08326757e+00 3.91779542e-01 3.08580697e-01
-9.16155040e-01 -3.58738929e-01 3.36785972e-01 -1.58535019e-01
-6.77417099e-01 -5.36935329e-01 8.88012350e-01 -7.22965539e-01
1.39970732e+00 3.18962991e-01 8.35538208e-01 9.80375826e-01
4.26722199e-01 7.76551187e-01 1.10616279e+00 -2.90331662e-01
4.81171668e-01 1.05140615e+00 6.03750229e-01 1.95388436e-01
1.26018301e-01 -5.33978641e-01 -6.28589690e-01 -4.32200789e-01
-1.35357797e-01 -2.44088378e-02 2.67944574e-01 5.32420166e-02
-5.88803709e-01 1.01952243e+00 -1.00386053e-01 3.91950667e-01
-9.11055088e-01 -3.58999819e-01 6.84194982e-01 6.26946986e-01
9.26162481e-01 4.87556994e-01 -9.26827431e-01 -6.49497271e-01
-4.23783034e-01 2.98040897e-01 1.41802788e+00 6.99778080e-01
8.17258418e-01 -3.78781885e-01 3.35200459e-01 8.14035714e-01
6.43392742e-01 9.75187793e-02 4.22114730e-01 -6.29787803e-01
1.52466476e-01 1.22449613e+00 3.11219722e-01 -1.12084854e+00
-3.68573546e-01 -5.91954589e-02 4.15101796e-02 -6.53128549e-02
-2.40116686e-01 -5.28220534e-01 -7.09099948e-01 9.75686073e-01
3.77812743e-01 -3.61904681e-01 1.02331325e-01 8.29722404e-01
1.18780804e+00 4.43228006e-01 2.13697717e-01 1.11015275e-01
1.55270433e+00 -8.98424923e-01 -9.59506392e-01 -6.16067827e-01
8.18757772e-01 -1.05630028e+00 1.02528775e+00 7.89457381e-01
-9.93245244e-01 -7.63773546e-02 -9.38119948e-01 -6.85369968e-02
-1.05084634e+00 -1.22902114e-02 6.60099685e-01 7.30240226e-01
-1.04821134e+00 3.57287318e-01 -6.58771574e-01 -5.55357814e-01
3.15663904e-01 1.56315774e-01 -1.15431458e-01 2.15281025e-01
-1.10525739e+00 8.58973920e-01 -2.26680309e-01 3.21525671e-02
-1.76208273e-01 -5.09182394e-01 -8.47127736e-01 -7.72458762e-02
9.43067223e-02 -4.02807474e-01 1.77308464e+00 -1.20446718e+00
-1.57450259e+00 1.18545091e+00 -3.12144190e-01 -1.62127584e-01
1.53729275e-01 -3.45738769e-01 -9.26549017e-01 -1.54228896e-01
1.43826352e-02 -2.05656271e-02 6.00031376e-01 -8.45474243e-01
-6.28449798e-01 -4.24644798e-01 1.77789569e-01 -6.80764690e-02
-1.17678709e-01 8.46400082e-01 -2.48107776e-01 -4.24637347e-02
-4.99783605e-01 -6.26409471e-01 -4.45313215e-01 -5.15590429e-01
-3.38526815e-01 -2.60590732e-01 6.84746981e-01 -7.31046259e-01
1.20190811e+00 -2.22463346e+00 -4.30121049e-02 2.73558617e-01
1.48064703e-01 1.40582368e-01 -5.23153180e-03 8.91519368e-01
-1.02876149e-01 4.30702537e-01 1.88234627e-01 -4.58885580e-01
3.38442773e-01 1.97838116e-02 -5.63844815e-02 6.15726933e-02
7.42284179e-01 7.31576741e-01 -7.74225175e-01 -9.41152126e-03
2.42103130e-01 6.42910719e-01 -2.93335974e-01 3.05487126e-01
-4.13031250e-01 2.18720153e-01 -6.97599888e-01 7.04553962e-01
6.12020075e-01 -2.57238805e-01 2.46644348e-01 1.96558107e-02
-2.50799924e-01 9.67974126e-01 -8.79568458e-01 1.28887475e+00
-7.67024517e-01 6.05535090e-01 5.72090626e-01 -6.02187455e-01
1.34758282e+00 3.10885042e-01 5.11057317e-01 -7.18635499e-01
5.79127550e-01 1.05974391e-01 -2.98600405e-01 -6.02386475e-01
5.99310100e-01 -2.81758159e-01 -2.76929826e-01 6.52246475e-01
-8.55616406e-02 -2.95126259e-01 2.98989773e-01 2.89684176e-01
1.03912449e+00 -1.15734532e-01 5.20821929e-01 4.76572067e-02
4.67958063e-01 1.71412364e-01 5.03009021e-01 -1.45967960e-01
-2.31354430e-01 3.12717527e-01 9.24824536e-01 -4.73085433e-01
-6.89963639e-01 -5.71662009e-01 -6.92451652e-03 1.06716919e+00
-2.73331881e-01 -1.12344480e+00 -5.29467225e-01 -6.22273505e-01
6.11686222e-02 9.09381747e-01 -7.56841302e-01 1.12280566e-02
1.01857997e-01 -4.90359366e-01 -2.76295632e-01 3.47921818e-01
-1.74567744e-01 -1.32844305e+00 -7.87776530e-01 2.49694616e-01
9.59720984e-02 -1.12945163e+00 1.24992961e-02 3.10710549e-01
-5.53146362e-01 -9.90286052e-01 8.01503733e-02 -4.64516342e-01
3.16215128e-01 -2.83510655e-01 1.56151855e+00 -1.84239849e-01
-4.13921654e-01 4.90816981e-01 -6.97930992e-01 -1.00298202e+00
-4.49359298e-01 1.04533747e-01 -2.93125838e-01 1.43869027e-01
1.32233965e+00 -3.99224043e-01 -6.52461946e-01 1.65272415e-01
-7.80994058e-01 -1.25524357e-01 2.97641754e-01 1.83596030e-01
2.02435762e-01 -2.16301650e-01 6.35518193e-01 -1.46251023e+00
1.18663979e+00 -1.03614521e+00 -3.68904173e-01 -1.26829952e-01
-8.12468350e-01 -5.50011277e-01 5.44050634e-01 1.65421173e-01
-1.03579485e+00 -9.62654725e-02 -6.21233881e-01 3.11461478e-01
-6.27596319e-01 1.02803671e+00 -1.46721706e-01 5.49889803e-01
4.29833919e-01 -5.29048979e-01 -8.18413869e-03 -6.57085001e-01
4.15335268e-01 1.24665701e+00 -1.45579040e-01 -1.38710007e-01
1.60096005e-01 4.17961568e-01 -8.52527022e-01 -5.99262416e-01
-8.43049407e-01 -8.55153978e-01 -4.58364636e-01 -4.30577904e-01
7.29888916e-01 -8.02591443e-01 -5.57225227e-01 3.08305621e-01
-9.52903569e-01 -1.99686885e-01 -4.62715328e-01 -5.81682026e-02
-7.68535212e-02 -2.06101269e-01 -5.76153815e-01 -1.02251184e+00
-6.73971593e-01 -9.41852868e-01 1.16402376e+00 4.59029019e-01
-8.96250486e-01 -1.08030081e+00 2.71586716e-01 4.21539485e-01
5.92890561e-01 3.32228839e-01 5.38533032e-01 -1.11611962e+00
2.78008282e-01 -6.90094650e-01 8.13178346e-02 6.90439582e-01
2.68358767e-01 4.09784913e-01 -1.19242895e+00 1.55966297e-01
2.44588673e-01 -4.06937152e-01 1.72871202e-01 1.30102992e-01
5.45094311e-01 -1.70175284e-01 -1.85141236e-01 -6.09018430e-02
1.27697015e+00 3.06987584e-01 5.63317657e-01 6.83807373e-01
1.54758736e-01 1.07882273e+00 1.07931304e+00 9.84641194e-01
6.01618230e-01 1.82523012e-01 4.76032674e-01 -1.86272085e-01
5.60781717e-01 1.17151238e-01 6.27614021e-01 8.78121793e-01
3.72654915e-01 -2.08570093e-01 -9.85906303e-01 7.31405795e-01
-1.91950452e+00 -4.42255616e-01 -5.01573265e-01 1.49906552e+00
4.94595110e-01 2.44547054e-01 1.44022003e-01 -8.87524560e-02
2.22653732e-01 1.78281143e-01 -5.32532632e-01 -1.62677848e+00
1.44241214e-01 3.34478199e-01 -5.86671010e-02 2.61741817e-01
-8.30149114e-01 7.48309910e-01 6.64840364e+00 -3.54878418e-02
-1.08924317e+00 1.36297792e-01 8.14735293e-01 -1.81307539e-01
-5.22251487e-01 1.54984191e-01 -5.03453016e-01 3.66240114e-01
1.43155944e+00 -3.70034486e-01 1.45382062e-01 1.22020769e+00
7.12776780e-01 -4.15048629e-01 -9.93505895e-01 5.49972713e-01
-2.01077610e-01 -1.03257287e+00 -5.46753466e-01 -8.14823061e-03
3.72489095e-01 3.92437279e-01 4.00101170e-02 1.98750854e-01
4.22622442e-01 -8.12865436e-01 4.50080216e-01 4.00257528e-01
2.41035759e-01 -7.91790724e-01 1.13417721e+00 1.03244424e-01
-7.54910707e-01 5.54404631e-02 -3.14336666e-03 -4.57706749e-01
5.20018935e-01 1.01062632e+00 -1.15661836e+00 4.74698752e-01
9.88278449e-01 1.03735423e+00 -3.10663760e-01 4.18890446e-01
-2.39682794e-01 6.95504963e-01 4.81932499e-02 -2.07832366e-01
9.54455659e-02 -5.85536301e-01 2.60445356e-01 1.58695936e+00
5.22694178e-02 2.29357276e-02 1.10804200e-01 7.39930332e-01
1.74321920e-01 4.91066247e-01 -6.40544116e-01 -6.75999701e-01
1.98736042e-02 1.93275857e+00 -5.07033408e-01 -3.53779703e-01
-7.58495271e-01 7.07942784e-01 -9.45435986e-02 1.79159954e-01
-5.47956467e-01 -5.23319542e-01 1.14592969e+00 2.67471850e-01
3.14901501e-01 1.14885300e-01 -3.76275778e-01 -1.01104701e+00
8.84446278e-02 -1.10448945e+00 3.86993378e-01 -9.10223961e-01
-1.27532685e+00 7.16290414e-01 -5.86737871e-01 -8.37804317e-01
-4.20904100e-01 -9.93685186e-01 -9.34300184e-01 9.16786253e-01
-1.54058659e+00 -9.77083743e-01 -3.05921614e-01 2.08606273e-01
3.62105876e-01 -1.19331323e-01 1.07743180e+00 3.29972923e-01
-6.94264114e-01 -1.08206451e-01 -2.84914404e-01 -5.60592525e-02
8.05529654e-01 -1.27532125e+00 6.18182242e-01 3.91575307e-01
-2.38240466e-01 7.40596950e-01 8.93764079e-01 -5.53162813e-01
-1.49756074e+00 -6.53849483e-01 1.31012571e+00 -5.96167803e-01
1.16529119e+00 -4.80599284e-01 -6.04759514e-01 8.71594846e-01
8.03629816e-01 -4.41056430e-01 1.44056022e+00 4.53920275e-01
-1.83838964e-01 -7.92786255e-02 -1.31564796e+00 1.41609520e-01
3.06336820e-01 -6.17319822e-01 -4.67425764e-01 2.92079896e-01
5.79298735e-01 -1.15091860e-01 -1.18993807e+00 -4.78054248e-02
6.92608058e-01 -1.31364715e+00 2.56784737e-01 -7.19851315e-01
7.22674727e-01 -1.73460990e-02 -1.00335822e-01 -1.54617369e+00
3.82707492e-02 -7.45745420e-01 -3.19479965e-02 1.46508229e+00
8.68657351e-01 -9.99649763e-01 4.06150430e-01 1.01747608e+00
2.50494666e-02 -7.49696612e-01 -2.33279929e-01 -1.37811489e-02
-2.45773360e-01 -9.16695356e-01 7.77251661e-01 9.15051997e-01
5.96699536e-01 5.95823646e-01 1.38235271e-01 -1.42190591e-01
-1.75430998e-01 1.62684143e-01 6.48212552e-01 -1.28830492e+00
-1.41931936e-01 -4.34929937e-01 -2.60643303e-01 -5.16315341e-01
-1.16559036e-01 -4.43452328e-01 -2.16996029e-01 -1.79024076e+00
-3.00725587e-02 -1.74773946e-01 -4.46152002e-01 6.05072737e-01
2.33299717e-01 1.56372964e-01 -3.92683633e-02 -1.92831397e-01
-4.66290891e-01 1.62217855e-01 5.91393292e-01 -1.48481177e-02
-2.01680288e-01 1.79193839e-01 -1.46781266e+00 7.98552990e-01
1.23792577e+00 -2.82156289e-01 -2.46237978e-01 -1.22960612e-01
9.03549135e-01 -5.60369194e-01 -4.27485332e-02 -2.57783324e-01
-1.30191237e-01 2.54684389e-02 -1.20848060e-01 -6.12504601e-01
2.78972387e-01 -8.54723871e-01 7.76632950e-02 1.02616623e-02
-2.40795359e-01 3.26009601e-01 3.68743628e-01 5.45294471e-02
-5.77311575e-01 -2.79242635e-01 3.98898035e-01 -9.99342650e-02
-5.20310163e-01 -1.26518428e-01 -6.51856899e-01 -3.29113036e-01
9.62352335e-01 9.57363993e-02 -3.18107486e-01 -6.57332242e-01
-9.77246881e-01 4.79692906e-01 5.95314860e-01 8.79935145e-01
5.17736495e-01 -9.30384696e-01 -4.92157072e-01 2.27404296e-01
3.31167698e-01 -3.69239956e-01 1.24302683e-02 9.75261092e-01
-2.53727555e-01 3.23272169e-01 -1.16879726e-02 -2.86505282e-01
-1.32031441e+00 2.72032470e-01 1.62079126e-01 -3.05081785e-01
-3.03340822e-01 6.47526383e-01 -2.44653583e-01 -6.92053258e-01
-1.34776190e-01 -4.56240237e-01 -5.94671071e-01 4.57204342e-01
6.85336411e-01 1.16336748e-01 5.20787835e-01 -6.57918572e-01
-7.96253741e-01 2.90075630e-01 -2.85667866e-01 -3.07458818e-01
1.83560920e+00 -1.63710490e-01 -4.75115091e-01 8.74960780e-01
1.14196742e+00 8.89316667e-03 -6.74946606e-01 1.77752227e-01
4.70127970e-01 -3.33858818e-01 2.20692575e-01 -1.29041088e+00
-1.02406156e+00 6.95467889e-01 2.91899443e-01 8.98575723e-01
1.24788117e+00 2.31010407e-01 6.86953127e-01 -3.98774147e-02
-4.70953025e-02 -1.47577238e+00 -2.98612058e-01 3.02561283e-01
8.21302116e-01 -1.11913502e+00 -8.63055736e-02 -4.79020119e-01
-1.31624568e+00 1.02863967e+00 3.96468848e-01 2.94164233e-02
9.93382215e-01 7.21098006e-01 7.57243812e-01 -7.23037779e-01
-1.35411012e+00 -7.74871260e-02 8.00914411e-03 4.74202245e-01
1.03444624e+00 1.69969708e-01 -4.47690994e-01 1.21432316e+00
-1.98403925e-01 2.24741608e-01 6.11552656e-01 1.22840106e+00
-1.74744636e-01 -1.44710696e+00 -7.02928975e-02 5.52790642e-01
-8.06396842e-01 -8.54612961e-02 -1.22804070e+00 3.13388348e-01
-9.57230385e-03 1.53430521e+00 -1.18084095e-01 -5.27396083e-01
7.38939703e-01 3.38886291e-01 -2.71340728e-01 -8.73736322e-01
-1.24218035e+00 1.15324244e-01 9.38676238e-01 -8.41819584e-01
-4.70552623e-01 -9.56516802e-01 -1.17851710e+00 -7.20257163e-02
-5.75453937e-02 2.89389044e-01 1.14642203e+00 7.10436702e-01
9.30036008e-01 5.71242571e-01 8.44154656e-01 -5.45943618e-01
1.47924665e-02 -1.02783620e+00 -5.50194800e-01 3.40431601e-01
1.64890289e-01 -1.30274221e-01 -4.29429203e-01 3.43081616e-02] | [11.284825325012207, 6.832795143127441] |
a100440e-38df-45ea-883c-ea9247ac3522 | exploiting-rich-syntax-for-better-knowledge | 2107.07940 | null | https://arxiv.org/abs/2107.07940v1 | https://arxiv.org/pdf/2107.07940v1.pdf | Exploiting Rich Syntax for Better Knowledge Base Question Answering | Recent studies on Knowledge Base Question Answering (KBQA) have shown great progress on this task via better question understanding. Previous works for encoding questions mainly focus on the word sequences, but seldom consider the information from syntactic trees.In this paper, we propose an approach to learn syntax-based representations for KBQA. First, we encode path-based syntax by considering the shortest dependency paths between keywords. Then, we propose two encoding strategies to mode the information of whole syntactic trees to obtain tree-based syntax. Finally, we combine both path-based and tree-based syntax representations for KBQA. We conduct extensive experiments on a widely used benchmark dataset and the experimental results show that our syntax-aware systems can make full use of syntax information in different settings and achieve state-of-the-art performance of KBQA. | ['Min Zhang', 'Wenliang Chen', 'Muhua Zhu', 'Yonghui Jia', 'Pengju Zhang'] | 2021-07-16 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [ 3.56084332e-02 -4.32448313e-02 -1.38567641e-01 -7.16881633e-01
-1.03223097e+00 -6.70829296e-01 1.40369877e-01 3.30417812e-01
-3.53433132e-01 6.24062300e-01 6.97084606e-01 -8.42200518e-01
-1.84139207e-01 -1.23227751e+00 -7.58056223e-01 -1.14058554e-01
1.41430333e-01 4.00731802e-01 7.99829125e-01 -7.30693758e-01
2.03848690e-01 3.21071334e-02 -1.27504909e+00 8.51234555e-01
1.13175499e+00 9.43426669e-01 1.23934411e-01 6.73823595e-01
-1.05370188e+00 1.25384080e+00 -7.39510179e-01 -6.34766936e-01
-2.09325060e-01 -4.05409098e-01 -1.55131149e+00 -3.47765565e-01
4.37378019e-01 -4.54797685e-01 -4.76717383e-01 7.19473124e-01
1.33368552e-01 1.48270592e-01 4.00476307e-01 -8.54437590e-01
-1.03274751e+00 8.95227909e-01 5.05992249e-02 5.85848510e-01
7.94971704e-01 -5.89243770e-02 1.68760741e+00 -5.46613693e-01
4.00798023e-01 1.40760362e+00 2.90297538e-01 5.00355124e-01
-5.96052706e-01 -2.88527191e-01 4.19707209e-01 8.62620234e-01
-1.04130340e+00 -2.34280266e-02 8.23875904e-01 -1.23782985e-01
1.30342638e+00 2.11758986e-01 2.97894269e-01 7.18112886e-01
5.40683754e-02 9.84283149e-01 9.56886828e-01 -4.80283111e-01
1.64650362e-02 -2.93685555e-01 1.03254068e+00 9.94979739e-01
1.66193426e-01 -4.35610443e-01 -2.27752432e-01 -1.66408926e-01
4.17804480e-01 -4.17698473e-01 -4.09068197e-01 -8.06157365e-02
-9.02279377e-01 1.13319254e+00 5.48721433e-01 3.88777584e-01
-1.96040183e-01 3.03100944e-01 4.64588583e-01 4.67310816e-01
5.40848821e-03 5.92694223e-01 -8.30367863e-01 -1.38723508e-01
-4.69188929e-01 5.80474436e-01 1.13675332e+00 8.30318749e-01
8.81589592e-01 -3.43401879e-01 -4.32195306e-01 9.19037879e-01
4.24586833e-01 4.27651107e-01 5.31843424e-01 -9.49269772e-01
8.06588113e-01 8.92610669e-01 -1.01901665e-01 -9.42194164e-01
-3.94639611e-01 3.44396457e-02 -2.99211532e-01 -7.08599448e-01
4.88936216e-01 -7.00337738e-02 -1.01225042e+00 1.62042415e+00
2.67570555e-01 2.77727172e-02 2.23624304e-01 6.88161850e-01
1.45835805e+00 1.18107224e+00 1.17783211e-01 3.37901860e-02
1.79907286e+00 -1.15548468e+00 -9.43073153e-01 -2.03292668e-01
1.04508781e+00 -2.87220061e-01 1.39093935e+00 1.40318111e-01
-9.98587012e-01 -3.46796900e-01 -9.28649008e-01 -4.82854486e-01
-6.72122359e-01 -1.65061176e-01 6.18743300e-01 7.15754330e-01
-9.58792806e-01 1.63443327e-01 -5.58459461e-01 -2.08372399e-01
2.55947083e-01 3.23792733e-02 -9.65821370e-02 -5.34205675e-01
-1.70407343e+00 8.20096850e-01 5.80316424e-01 -1.13120139e-01
-7.29271650e-01 -8.30845833e-01 -1.14583683e+00 3.08257312e-01
7.64375985e-01 -9.48657274e-01 1.69652462e+00 -3.79631937e-01
-1.44976842e+00 4.39841121e-01 -5.49625099e-01 -4.67652470e-01
-4.79169458e-01 -3.07316184e-01 -4.50888842e-01 3.44658315e-01
-1.80418283e-01 3.50857049e-01 3.29752535e-01 -1.02320528e+00
-5.36660731e-01 -2.85788834e-01 8.56736422e-01 1.44081697e-01
-4.06215400e-01 1.74860775e-01 -5.32208443e-01 -4.40692902e-01
3.13993311e-03 -3.35622430e-01 -1.72605038e-01 -8.70395303e-02
-1.23420760e-01 -7.77151108e-01 6.52211547e-01 -1.03254700e+00
1.80404460e+00 -1.66637194e+00 1.95723712e-01 -1.27370059e-01
9.05170068e-02 4.77803558e-01 -3.45457137e-01 7.02179193e-01
3.78274292e-01 3.78677040e-01 -4.71430272e-01 6.66698292e-02
1.02377258e-01 8.59844983e-01 -7.55273104e-01 -3.73420864e-01
4.29474890e-01 1.32412672e+00 -9.05182600e-01 -6.37018144e-01
-1.96811453e-01 3.04242221e-05 -7.58302331e-01 5.06027281e-01
-8.58960271e-01 -1.37477279e-01 -8.48690152e-01 5.21693587e-01
4.92607117e-01 -4.10093755e-01 1.99779928e-01 -1.26796365e-01
4.96536911e-01 1.18352091e+00 -8.04427445e-01 1.91022468e+00
-6.56900108e-01 2.76871353e-01 -1.75109535e-01 -1.04603493e+00
8.33341658e-01 1.72162354e-01 -1.78578258e-01 -1.01518941e+00
-1.10111786e-02 1.76695347e-01 1.14575895e-02 -6.56736195e-01
5.91440320e-01 -1.21563248e-01 -2.93058455e-01 5.34726441e-01
2.11673692e-01 -2.27401271e-01 3.81168336e-01 4.44090754e-01
1.22894549e+00 1.88089516e-02 3.84390563e-01 2.64828671e-02
9.27215934e-01 3.91365476e-02 2.14439064e-01 7.00310469e-01
-5.94817512e-02 1.99036583e-01 5.72139621e-01 -3.89354199e-01
-5.22964776e-01 -1.03836441e+00 1.41461849e-01 1.35512483e+00
3.33341621e-02 -9.26897645e-01 -7.24326968e-01 -1.24157894e+00
-1.08521290e-01 8.54448736e-01 -5.95547020e-01 -2.73700118e-01
-1.09084868e+00 -4.65818465e-01 8.45314324e-01 7.96410918e-01
4.88750786e-01 -1.19666553e+00 -4.58574444e-01 3.29828620e-01
-6.06127977e-01 -1.13749719e+00 -4.09698710e-02 -1.20476089e-01
-9.96835887e-01 -1.03923512e+00 -3.51238519e-01 -7.02368379e-01
1.31642669e-01 2.65205830e-01 1.41053164e+00 4.40719545e-01
2.71434262e-02 4.95262593e-01 -8.58219206e-01 -2.15561137e-01
-2.51232326e-01 3.72646183e-01 -7.46577919e-01 -3.61440271e-01
3.95541221e-01 -4.31258976e-01 -4.09042597e-01 8.42769593e-02
-1.04575932e+00 -3.92806351e-01 4.94297087e-01 6.34600997e-01
3.66984963e-01 -2.50888377e-01 7.78364956e-01 -9.21988964e-01
9.99054432e-01 -5.95024765e-01 -3.92269641e-01 6.67029202e-01
-3.81912172e-01 4.33038473e-01 7.01789021e-01 1.60615123e-03
-1.23507202e+00 -6.38028085e-01 -6.92194402e-01 1.43541917e-01
-1.56271860e-01 9.79203284e-01 -3.83358598e-01 1.17657281e-01
5.08308768e-01 3.74106020e-01 -3.51553291e-01 -5.97205997e-01
8.92536581e-01 5.47863483e-01 2.83989877e-01 -1.20626652e+00
5.20525157e-01 1.32176206e-01 -3.73423666e-01 -5.48712492e-01
-1.25286019e+00 -5.01514316e-01 -3.65882814e-01 2.04404160e-01
8.54507625e-01 -4.76411015e-01 -4.29389387e-01 1.36705920e-01
-1.27295101e+00 -9.26706269e-02 -2.78327078e-01 -5.27547859e-02
-4.14069980e-01 7.92952001e-01 -6.95825994e-01 -5.72568893e-01
-4.09877539e-01 -1.03701437e+00 1.01306283e+00 2.38958687e-01
-6.25178516e-02 -1.05992699e+00 3.53478432e-01 7.40117610e-01
4.20063883e-01 -3.50525230e-01 1.65798438e+00 -8.97315621e-01
-7.79909611e-01 1.76760957e-01 -3.61905903e-01 2.76993155e-01
2.48115629e-01 -3.69859070e-01 -6.83503509e-01 1.59180418e-01
-2.82439411e-01 -6.51087940e-01 1.04071248e+00 -3.85580286e-02
1.43104482e+00 -4.72044617e-01 -1.23599207e-03 3.99731874e-01
1.27835822e+00 2.85310507e-01 7.79301822e-01 3.71198684e-01
6.20979965e-01 7.06918120e-01 6.36323929e-01 1.65239498e-01
1.04968059e+00 5.32723010e-01 4.73295599e-01 4.91831958e-01
-2.56347895e-01 -4.93362695e-01 2.30257958e-01 1.20135081e+00
3.64058584e-01 -4.32126254e-01 -1.03576863e+00 8.73635888e-01
-1.75125146e+00 -5.46313286e-01 -1.71281084e-01 1.60760796e+00
1.06434131e+00 -1.14086792e-01 -1.41251042e-01 1.65979788e-01
3.88736166e-02 4.76655006e-01 -3.02852362e-01 -7.82173693e-01
1.07363470e-01 7.08032310e-01 1.10927867e-02 4.00382161e-01
-9.11617398e-01 1.26914680e+00 6.83233356e+00 6.90190613e-01
-5.81080675e-01 1.11726811e-02 2.32499704e-01 3.13125521e-01
-8.34590316e-01 3.76923159e-02 -9.04551208e-01 1.38815239e-01
1.22502577e+00 -3.08908850e-01 2.25134999e-01 6.71565533e-01
-3.90998989e-01 7.76637122e-02 -9.44625318e-01 6.55356050e-01
1.22898489e-01 -1.56341958e+00 7.43951738e-01 -4.23808724e-01
3.18035930e-01 -3.10796648e-01 -2.71191686e-01 7.67915726e-01
4.90266055e-01 -1.13207912e+00 3.27314854e-01 5.15308678e-01
2.14439645e-01 -7.17927933e-01 8.45652759e-01 4.02991652e-01
-1.28779101e+00 -2.19060868e-01 -6.25210524e-01 -1.64192498e-01
4.22089726e-01 4.20084655e-01 -7.10403621e-01 1.20895040e+00
8.32486928e-01 4.96185958e-01 -7.89278865e-01 8.03826749e-01
-6.71583951e-01 1.09539938e+00 -1.57415956e-01 -4.78166521e-01
5.30514538e-01 -6.53405562e-02 1.22472718e-01 1.15057755e+00
1.26358792e-01 5.77411354e-01 2.21628342e-02 6.98709488e-01
-2.46576473e-01 3.51182491e-01 -3.67601603e-01 -3.04209560e-01
3.24926823e-01 8.86747062e-01 -1.01340659e-01 -5.33341348e-01
-6.57775164e-01 6.92994297e-01 7.86347270e-01 2.43844569e-01
-7.28044689e-01 -5.26526928e-01 7.61898339e-01 -2.56671578e-01
6.51779175e-01 -4.27334607e-01 -1.08446949e-03 -1.20028698e+00
2.60969847e-01 -1.14673567e+00 1.06168270e+00 -1.00827003e+00
-1.33567858e+00 5.27074695e-01 2.56883234e-01 -3.06057632e-01
-2.02374026e-01 -8.04753780e-01 -4.90373075e-01 7.33971894e-01
-1.96504509e+00 -9.97098446e-01 -1.10870749e-01 5.87863028e-01
6.33922577e-01 2.89390922e-01 9.17408705e-01 2.03544095e-01
-3.24829549e-01 6.48508310e-01 -1.98636577e-01 4.14916426e-01
4.11919147e-01 -1.16430688e+00 4.79379207e-01 6.35246873e-01
5.47519565e-01 9.23645496e-01 2.04930454e-01 -4.85708594e-01
-1.73848367e+00 -1.01952887e+00 1.04745102e+00 -6.65784061e-01
6.75531268e-01 -9.90162492e-02 -1.50995123e+00 8.74338925e-01
4.11174029e-01 -8.05365443e-02 8.36855710e-01 3.37450683e-01
-9.48938847e-01 -1.10353194e-01 -8.40394735e-01 4.34675843e-01
1.04198527e+00 -6.51239991e-01 -1.58030939e+00 1.09730870e-01
1.53011620e+00 -2.51026541e-01 -9.42716837e-01 5.57682753e-01
3.10012549e-01 -6.10752404e-01 9.73168254e-01 -1.24998498e+00
5.38519979e-01 -3.88335526e-01 -4.02986646e-01 -1.11083281e+00
-2.20066801e-01 -8.10387284e-02 -4.53967720e-01 1.14345777e+00
6.06123328e-01 -4.62645322e-01 6.49575949e-01 1.08706303e-01
-2.77198225e-01 -1.11666489e+00 -9.40728784e-01 -5.71614563e-01
5.27433872e-01 -5.10552466e-01 9.20596361e-01 6.63171172e-01
7.11323991e-02 5.73769629e-01 2.66942140e-02 2.77058691e-01
5.04240133e-02 4.63329464e-01 5.01793921e-01 -9.83259618e-01
-3.27634573e-01 -2.49586776e-01 -1.57698914e-01 -1.69368017e+00
3.25898409e-01 -1.01644790e+00 9.20821428e-02 -2.08330870e+00
6.64674565e-02 -1.26558036e-01 -3.98133695e-01 4.63130087e-01
-6.14007056e-01 -4.36356604e-01 1.13972872e-02 -3.33410382e-01
-8.09609294e-01 6.72193050e-01 1.33570635e+00 -2.31547371e-01
2.89787084e-01 -5.04079282e-01 -9.22967553e-01 4.15838629e-01
7.57932365e-01 -2.93195128e-01 -7.32088864e-01 -8.84742558e-01
5.59297740e-01 1.23511165e-01 2.84347180e-02 -6.17770612e-01
1.71990484e-01 -3.12492520e-01 -2.31192395e-01 -6.33738041e-01
3.17182451e-01 -6.58914566e-01 -7.04406738e-01 3.12201411e-01
-5.24362981e-01 1.44392654e-01 2.37357035e-01 6.22692883e-01
-5.16688168e-01 -4.72382188e-01 3.89743626e-01 -3.36092979e-01
-9.81595337e-01 2.54644096e-01 -5.40063441e-01 6.14635646e-01
5.62592089e-01 3.71092975e-01 -7.37145483e-01 -5.49684644e-01
-4.19917643e-01 5.80550551e-01 -1.27740085e-01 5.82864404e-01
7.92417228e-01 -1.15102291e+00 -5.31810522e-01 -5.12165949e-02
3.84634078e-01 3.64676565e-02 1.50665745e-01 1.69353426e-01
-5.88745236e-01 8.13193083e-01 2.16086227e-02 -2.68100768e-01
-1.14748681e+00 6.84553146e-01 3.56450796e-01 -7.98307598e-01
-3.17456126e-01 1.03311491e+00 2.20357180e-01 -7.84835100e-01
1.63165063e-01 -9.92553592e-01 -6.67876899e-01 -9.07949209e-02
8.36534262e-01 1.04936115e-01 2.67420977e-01 -2.78247505e-01
-4.50126380e-01 6.50941670e-01 -3.42497736e-01 2.24355217e-02
9.38699484e-01 1.11989699e-01 -3.52307707e-01 3.21146876e-01
1.12055671e+00 -2.28265598e-01 -4.12956327e-01 -5.49600065e-01
4.16173071e-01 -3.10246617e-01 -1.16230264e-01 -9.65633631e-01
-8.57513309e-01 1.24399590e+00 4.33538966e-02 2.00148702e-01
1.21923971e+00 2.93465614e-01 1.36690533e+00 1.14908254e+00
4.07774866e-01 -5.86538672e-01 3.37589651e-01 1.08904219e+00
8.63697827e-01 -7.46892154e-01 -2.08228081e-01 -7.26443112e-01
-3.81826252e-01 1.07950068e+00 7.53041208e-01 -4.14089998e-03
5.62763274e-01 -1.73627093e-01 1.13953359e-01 -2.84565121e-01
-1.11596286e+00 -5.55727303e-01 4.13777858e-01 4.94129688e-01
3.96469295e-01 -1.29367411e-01 -5.40622234e-01 1.04553485e+00
-2.51535863e-01 -1.02524310e-01 4.62561995e-01 1.28184164e+00
-9.06093538e-01 -1.51721084e+00 -9.12404060e-02 4.67316002e-01
-4.82968420e-01 -4.14121836e-01 -6.12426221e-01 6.80692911e-01
-3.14677864e-01 1.11921751e+00 -1.26455262e-01 -2.59436429e-01
6.57620072e-01 6.97072268e-01 6.94365859e-01 -8.97294581e-01
-6.80606961e-01 -6.87803686e-01 5.93820572e-01 -7.85666406e-01
-2.99095601e-01 -2.72983223e-01 -1.45532370e+00 -1.14956342e-01
-2.14654163e-01 5.35021245e-01 3.93511444e-01 1.14237344e+00
4.53953803e-01 6.74666405e-01 3.62230301e-01 3.20193738e-01
-7.11226583e-01 -8.64913225e-01 -9.16397944e-02 2.93596745e-01
2.11317435e-01 -4.23062056e-01 -1.88155353e-01 -1.60494924e-01] | [10.657815933227539, 7.922436237335205] |
378e4d4e-4a25-40b3-856b-06d17119d9ca | effects-of-word-frequency-based-pre-and-post | 2009.11436 | null | https://arxiv.org/abs/2009.11436v1 | https://arxiv.org/pdf/2009.11436v1.pdf | Effects of Word-frequency based Pre- and Post- Processings for Audio Captioning | The system we used for Task 6 (Automated Audio Captioning)of the Detection and Classification of Acoustic Scenes and Events(DCASE) 2020 Challenge combines three elements, namely, dataaugmentation, multi-task learning, and post-processing, for audiocaptioning. The system received the highest evaluation scores, butwhich of the individual elements most fully contributed to its perfor-mance has not yet been clarified. Here, to asses their contributions,we first conducted an element-wise ablation study on our systemto estimate to what extent each element is effective. We then con-ducted a detailed module-wise ablation study to further clarify thekey processing modules for improving accuracy. The results showthat data augmentation and post-processing significantly improvethe score in our system. In particular, mix-up data augmentationand beam search in post-processing improve SPIDEr by 0.8 and 1.6points, respectively. | ['Kunio Kashino', 'Yasunori Ohishi', 'Yuma Koizumi', 'Daiki Takeuchi', 'Noboru Harada'] | 2020-09-24 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 3.00819188e-01 -1.03049256e-01 2.98161507e-01 -2.16706291e-01
-1.73480785e+00 -7.02198803e-01 5.60605884e-01 2.41804570e-01
-5.91622829e-01 4.55911875e-01 5.03418028e-01 -1.12641402e-01
5.15921926e-03 -9.65310559e-02 -7.54600823e-01 -3.61967266e-01
-1.50568679e-01 2.18125686e-01 2.00501248e-01 -7.84070790e-02
-5.45500889e-02 5.83738200e-02 -1.68545544e+00 8.81364286e-01
5.93778968e-01 1.11463356e+00 2.54258960e-02 9.70893323e-01
7.43246824e-02 5.91205239e-01 -7.39276052e-01 -2.17990533e-01
-2.76123315e-01 -3.79110575e-01 -8.87290418e-01 -9.64065567e-02
6.37022257e-01 -1.85530379e-01 1.25291958e-01 6.36461973e-01
8.62637401e-01 1.17956161e-01 3.50982457e-01 -1.24506807e+00
2.37662848e-02 8.65775287e-01 -5.57455063e-01 5.29491842e-01
4.54280138e-01 -5.63474372e-02 1.07274127e+00 -1.26975644e+00
9.60499570e-02 1.00908589e+00 8.27599108e-01 4.15918976e-01
-1.29506862e+00 -1.12750995e+00 -7.43503943e-02 3.46605808e-01
-1.30466843e+00 -1.13717389e+00 7.18197465e-01 -4.61257607e-01
1.05638063e+00 3.90011728e-01 2.35863224e-01 1.09716058e+00
-2.51528233e-01 9.48406637e-01 1.01420724e+00 -4.19052213e-01
1.31079957e-01 -3.20451483e-02 1.82925567e-01 3.77409309e-02
-2.24812940e-01 -1.89086124e-02 -8.82531643e-01 -1.20768614e-01
2.55519778e-01 -8.84593844e-01 -1.10100172e-01 2.43452445e-01
-9.63488877e-01 3.29376578e-01 1.59965649e-01 3.32432181e-01
-3.97690296e-01 3.06575835e-01 6.90319061e-01 3.09731007e-01
4.37060237e-01 7.37300992e-01 -5.44273555e-01 -5.97900808e-01
-9.66084003e-01 2.63741553e-01 4.10184681e-01 6.24385715e-01
2.69312501e-01 3.33950579e-01 -3.97114307e-01 1.21979809e+00
1.77401304e-01 4.51903164e-01 3.73244256e-01 -7.47191727e-01
6.67564690e-01 -1.02198094e-01 9.40688699e-02 -7.25303710e-01
-6.65695906e-01 -5.53214133e-01 -4.45086926e-01 -1.14837103e-01
3.29346716e-01 -6.39948189e-01 -9.40847099e-01 1.89276934e+00
-6.34030923e-02 4.61725414e-01 -1.74715802e-01 9.26062703e-01
9.06633317e-01 6.33811533e-01 5.22712409e-01 1.07824225e-02
1.57458889e+00 -8.81140113e-01 -8.70517969e-01 -4.11631674e-01
6.51892185e-01 -1.08705652e+00 1.15204346e+00 6.44809484e-01
-1.20572686e+00 -7.18250751e-01 -1.05781245e+00 2.32710198e-01
3.01987734e-02 6.02048300e-02 3.08788151e-01 4.57388341e-01
-9.34542298e-01 1.78595006e-01 -7.37852573e-01 -1.63960561e-01
3.52090418e-01 1.39314249e-01 -3.11191618e-01 4.20511603e-01
-1.38649964e+00 6.01209581e-01 3.74233544e-01 -2.87009358e-01
-1.22033882e+00 -1.03404999e+00 -5.48987985e-01 1.18803963e-01
5.13463140e-01 -3.52256417e-01 1.88522351e+00 -6.14928722e-01
-1.19098353e+00 6.15999758e-01 -5.82720079e-02 -6.39020920e-01
9.52549279e-02 -6.68744624e-01 -7.39500582e-01 1.25223130e-01
-1.13201797e-01 7.48271346e-01 7.74037361e-01 -1.04227519e+00
-8.95417392e-01 -1.46807298e-01 -3.55149396e-02 5.13556600e-01
-6.56648278e-01 5.89076042e-01 -7.21244395e-01 -1.03200972e+00
-1.62054136e-01 -8.23199868e-01 3.85971852e-02 -6.26291871e-01
-4.08271581e-01 -4.99077551e-02 5.79174459e-01 -8.77477229e-01
1.40324891e+00 -2.51746607e+00 -1.52830422e-01 -2.47606710e-02
1.24624498e-01 1.87632605e-01 -3.66627127e-01 2.74412155e-01
-4.26082641e-01 2.45215476e-01 -1.68222990e-02 -6.62536085e-01
-5.94286732e-02 -3.90973121e-01 -3.72271419e-01 -8.28140229e-02
4.43978190e-01 7.64409721e-01 -8.27537894e-01 -2.77436972e-01
1.07742585e-01 4.27558690e-01 -6.65814459e-01 2.06457689e-01
-1.58759683e-01 3.41290623e-01 -5.80027550e-02 5.46938121e-01
3.90191346e-01 2.00148448e-01 -2.98972219e-01 -3.44222099e-01
-1.10925622e-01 5.18995106e-01 -1.26941729e+00 1.67413008e+00
-5.98943532e-01 7.45173395e-01 4.22000051e-01 -5.83377361e-01
4.83232200e-01 7.68525481e-01 5.98556459e-01 -4.88409132e-01
-5.07101342e-02 1.51439607e-01 1.26370102e-01 -6.08654380e-01
5.13427079e-01 9.25664604e-03 -4.08005804e-01 3.68107617e-01
1.11888200e-01 -8.08204338e-02 -4.76618297e-03 2.03431979e-01
1.23948896e+00 -2.20684066e-01 -4.63219583e-02 -1.01620309e-01
2.63417929e-01 -7.68960938e-02 2.44625196e-01 8.73609424e-01
-1.89790875e-01 9.40309167e-01 2.09864199e-01 2.20061928e-01
-9.22593832e-01 -8.65554094e-01 -2.18233149e-02 1.43865335e+00
-4.33550686e-01 -8.05591941e-01 -7.41347551e-01 -3.46175581e-01
-2.38615647e-01 7.72365272e-01 -4.51167107e-01 -6.99497536e-02
-3.31224620e-01 -8.14062476e-01 1.19078457e+00 6.70211256e-01
4.33202356e-01 -1.10081708e+00 -4.69559997e-01 3.16815704e-01
-6.42094195e-01 -1.31129014e+00 -4.88314837e-01 3.14322382e-01
-4.65467572e-01 -7.89243162e-01 -4.99583334e-01 -6.32143795e-01
-1.88808702e-02 1.88568637e-01 9.12706137e-01 -3.59189153e-01
-2.91220397e-02 3.82890999e-01 -5.95455408e-01 -7.26236582e-01
-4.33922261e-01 2.68144071e-01 1.13744624e-01 -3.50520164e-02
3.70651096e-01 -6.55319393e-01 -2.49991328e-01 1.85509086e-01
-6.23858035e-01 -1.62859801e-02 7.28347540e-01 5.73790491e-01
3.94385844e-01 -6.92518055e-02 7.35869527e-01 -7.19971418e-01
9.20250475e-01 -2.54992396e-01 -2.28959993e-01 -1.46109918e-02
-4.14702266e-01 -2.24306703e-01 3.11795115e-01 -3.43817979e-01
-8.74308407e-01 1.43343344e-01 -6.47953928e-01 -2.93867022e-01
-5.08301079e-01 5.64295769e-01 -2.25054905e-01 2.38499522e-01
7.57476807e-01 -1.22390725e-01 -3.64315122e-01 -8.61644626e-01
4.13914591e-01 1.03191328e+00 7.94138134e-01 -5.29780567e-01
6.74747229e-01 4.09117341e-02 -4.35547531e-01 -8.33405197e-01
-8.61329436e-01 -5.83249211e-01 -2.62815237e-01 -1.57548219e-01
1.02552772e+00 -1.23811400e+00 -4.73409593e-01 5.83252966e-01
-1.19893837e+00 -1.94889128e-01 -2.55342782e-01 5.51035702e-01
-4.31711763e-01 5.92660345e-02 -5.19744933e-01 -1.00289381e+00
-4.69201386e-01 -8.68690908e-01 1.04091370e+00 -1.64111510e-01
-4.68381733e-01 -3.72226119e-01 1.73264802e-01 5.36616027e-01
5.20117283e-01 -6.54736236e-02 7.51568496e-01 -9.74226236e-01
-5.69238812e-02 -1.92774877e-01 -2.06145719e-02 3.87438506e-01
-1.70080706e-01 -3.27556223e-01 -1.69694304e+00 -1.33847415e-01
-1.28385350e-01 -5.05904734e-01 9.03635204e-01 3.00548077e-01
1.57069099e+00 -4.76099811e-02 -9.10437852e-02 3.17539006e-01
8.37022901e-01 2.54616410e-01 5.79446495e-01 2.38257602e-01
4.54097211e-01 6.34802520e-01 6.23619437e-01 4.09081191e-01
3.02847326e-01 1.13139200e+00 3.80447835e-01 -2.99394727e-01
-4.89405751e-01 -1.84875563e-01 4.40151155e-01 8.19749534e-01
1.83088049e-01 -1.71010494e-01 -1.12548888e+00 6.49190187e-01
-1.54933047e+00 -7.64191806e-01 -2.38814056e-01 2.22499895e+00
8.57620776e-01 3.62999082e-01 5.60496330e-01 5.35649955e-01
6.78128839e-01 -5.92303509e-03 -2.31651887e-01 -2.56136298e-01
-8.17706585e-02 1.77758873e-01 3.13057974e-02 4.80649948e-01
-1.39612508e+00 8.03180277e-01 6.89323997e+00 8.02432239e-01
-8.27387750e-01 3.60446423e-01 5.87040484e-01 -3.19971770e-01
1.92139968e-02 -3.65411580e-01 -6.83246851e-01 3.60413730e-01
1.38122952e+00 -1.16449054e-02 3.99582565e-01 5.54398417e-01
3.95072669e-01 5.54400161e-02 -9.94737983e-01 1.08289838e+00
1.65640324e-01 -9.49431837e-01 -2.50153124e-01 -2.31309682e-01
2.15294436e-01 3.46677840e-01 1.28183156e-01 5.62328994e-01
-9.21820551e-02 -6.86473370e-01 8.53743851e-01 1.56803489e-01
7.48721421e-01 -7.48319983e-01 7.71402538e-01 -7.43250875e-03
-1.10517538e+00 -1.93701282e-01 3.55685592e-01 1.20625068e-02
4.84346837e-01 5.39602697e-01 -1.10829353e+00 5.39047599e-01
1.06783056e+00 1.55016407e-01 -6.73252583e-01 1.36542940e+00
-1.78475350e-01 1.23784447e+00 -3.75658184e-01 2.84391284e-01
-7.18884617e-02 5.46657503e-01 8.66171062e-01 1.53122711e+00
2.67535686e-01 -2.70195361e-02 1.75821722e-01 3.09815168e-01
-1.74584970e-01 1.61047652e-01 -7.33413473e-02 -2.23296955e-01
8.22927237e-01 1.24919963e+00 -3.79217446e-01 -2.69321710e-01
-2.75875121e-01 6.10653758e-01 7.73528144e-02 3.08692098e-01
-9.76069629e-01 -7.42509007e-01 6.77485049e-01 2.08305031e-01
2.63612449e-01 -2.47433558e-02 -5.00605285e-01 -7.11076915e-01
-5.44190593e-02 -9.83453631e-01 5.74408710e-01 -9.20855999e-01
-9.06466603e-01 7.43469715e-01 1.15926005e-01 -1.06320667e+00
-3.24610114e-01 -1.07138760e-01 -4.69953150e-01 7.67215490e-01
-1.26025283e+00 -9.59098220e-01 -2.89785028e-01 2.76809156e-01
7.39365101e-01 -1.76807672e-01 1.06518316e+00 8.73241067e-01
-7.13917792e-01 1.04740894e+00 -2.07970425e-01 2.74900720e-03
9.38451231e-01 -1.26838040e+00 5.57264745e-01 1.02801549e+00
3.35088819e-01 2.83683300e-01 9.85694170e-01 -4.55272287e-01
-9.37077820e-01 -1.07221150e+00 8.41091096e-01 -6.41847253e-01
8.01142275e-01 -7.46684432e-01 -8.86902034e-01 3.89343947e-01
1.56391591e-01 -3.96411538e-01 7.80810058e-01 6.56584918e-01
-3.58004928e-01 -2.27932498e-01 -6.79629683e-01 3.77633065e-01
7.31256008e-01 -7.29164004e-01 -6.89500928e-01 1.90938994e-01
8.62798214e-01 -3.30302238e-01 -8.25121224e-01 6.22390509e-01
4.93485898e-01 -5.95896542e-01 9.51808572e-01 -6.05760098e-01
2.65532494e-01 -2.52313495e-01 -2.58862495e-01 -1.21306419e+00
-2.21377417e-01 -8.63174856e-01 -3.34016800e-01 1.62519908e+00
7.12518215e-01 -9.59853902e-02 6.36826932e-01 2.58871108e-01
-7.34345138e-01 -5.01702070e-01 -8.25528264e-01 -6.94129527e-01
-8.84476006e-02 -1.19960320e+00 6.10787272e-01 7.59328127e-01
5.27400114e-02 7.10659802e-01 -6.69818401e-01 1.66565910e-01
2.58049905e-01 -4.25798357e-01 8.40101957e-01 -1.02177370e+00
-4.14123386e-01 -3.70121777e-01 -2.33039811e-01 -9.11335766e-01
-3.56637180e-01 -7.72461772e-01 3.68582010e-01 -1.31643009e+00
1.07590444e-02 -2.66787976e-01 -6.25961065e-01 9.02390242e-01
-3.37607324e-01 5.27001560e-01 3.25475603e-01 1.21270582e-01
-8.01536441e-01 3.62404794e-01 6.45311236e-01 1.81874588e-01
-4.41239059e-01 2.17431158e-01 -9.85606194e-01 4.58494067e-01
9.28816199e-01 -5.11460781e-01 -3.93925697e-01 -6.55213535e-01
3.39508429e-02 -1.86379105e-01 2.69017369e-01 -1.18977010e+00
1.58042848e-01 3.35463047e-01 2.34879300e-01 -6.84367120e-01
7.95331061e-01 -6.42559052e-01 -2.06524841e-02 2.21057430e-01
-4.99635875e-01 1.02038123e-01 7.74271011e-01 3.50182414e-01
-4.34240669e-01 -3.83528098e-02 5.75429857e-01 2.97784477e-01
-6.46059394e-01 -1.61763623e-01 -6.65144205e-01 1.25522971e-01
7.74963379e-01 5.22329956e-02 -1.37933016e-01 -6.92094743e-01
-9.15697992e-01 2.74914593e-01 -2.65143096e-01 6.28986239e-01
3.53415281e-01 -1.31447875e+00 -1.06121421e+00 8.73593837e-02
2.48526379e-01 -3.55282903e-01 3.78159940e-01 9.46595371e-01
5.60422568e-03 5.68604469e-01 5.82979694e-02 -5.71932137e-01
-1.46210361e+00 1.37579232e-01 2.17716262e-01 -4.16565314e-02
-3.09702367e-01 1.15168607e+00 2.79538948e-02 -2.66850293e-01
7.42180586e-01 -3.02892625e-01 -4.36993062e-01 3.46350044e-01
7.41315246e-01 4.97570783e-01 5.57526588e-01 -5.42181015e-01
-5.12852490e-01 1.66949421e-01 -2.69737273e-01 -5.78719437e-01
1.34363306e+00 1.14010815e-02 2.94351757e-01 6.15394115e-01
7.89500892e-01 3.24557070e-04 -1.03943372e+00 -2.54203938e-02
1.95605487e-01 -2.78890699e-01 3.67723346e-01 -1.25165784e+00
-6.00924551e-01 9.55767274e-01 7.82359183e-01 4.16253805e-01
1.38898790e+00 9.94730070e-02 7.55583346e-01 1.90714687e-01
2.00411938e-02 -1.13476074e+00 1.88848630e-01 5.66742659e-01
1.06133866e+00 -1.07525754e+00 -2.94827133e-01 -4.46975470e-01
-6.35230482e-01 6.71638846e-01 8.36208045e-01 4.02585834e-01
5.27104914e-01 4.30825740e-01 8.49863663e-02 -1.01993360e-01
-8.68572235e-01 -1.75123602e-01 6.07538998e-01 4.78382021e-01
6.25176430e-01 -1.60795733e-01 2.47670636e-02 1.01747310e+00
-4.28905427e-01 -2.61922687e-01 2.46448815e-01 7.76422083e-01
-4.12888020e-01 -8.69966567e-01 -4.88247097e-01 4.79211926e-01
-6.69458747e-01 -1.47260278e-01 -4.50196117e-01 4.75094199e-01
-9.33048222e-03 1.33835030e+00 3.37834470e-03 -9.23136473e-01
6.96433842e-01 1.96347818e-01 2.39510775e-01 -7.64385641e-01
-7.89118350e-01 5.08649826e-01 5.29215813e-01 -4.29570019e-01
-1.65090486e-01 -7.96928942e-01 -1.03092659e+00 2.36467540e-01
-6.05424762e-01 4.78643984e-01 8.70793879e-01 8.92932713e-01
5.39282322e-01 9.77895200e-01 4.72484559e-01 -6.63183570e-01
-4.60715711e-01 -1.29313827e+00 -7.20844567e-02 2.36673459e-01
3.85371864e-01 -3.52386713e-01 -3.96931350e-01 2.04288270e-02] | [15.14358901977539, 5.140678405761719] |
8d22891a-6414-48e1-bc22-6f4c2cf2222d | corpus-lexicography-in-a-wider-context | null | null | https://aclanthology.org/R19-1041 | https://aclanthology.org/R19-1041.pdf | Corpus Lexicography in a Wider Context | This paper describes a set of tools that offers comprehensive solutions for corpus lexicography. The tools perform a range of tasks, including construction of corpus lexicon, integrating information from external dictionaries, internal analysis of the lexicon, and lexical analysis of the corpus. The set of tools is particularly useful for creating dictionaries for under-resourced languages. The tools are integrated in a general-purpose software that includes additional tools for various research tasks, such as linguistic development analysis. Equipped with a user-friendly interface, the described system can be easily incorporated in research in a variety of fields. | ['Chen Gafni'] | 2019-09-01 | null | null | null | ranlp-2019-9 | ['lexical-analysis'] | ['natural-language-processing'] | [-3.33929360e-01 -1.61141291e-01 -3.30888063e-01 -2.15378359e-01
-2.18565181e-01 -8.02863419e-01 5.04542887e-01 3.15730006e-01
-4.50115472e-01 6.77734613e-01 3.84039283e-01 -6.49478137e-01
1.38387114e-01 -6.46347940e-01 1.73093364e-01 -3.68829638e-01
4.00925785e-01 8.44429016e-01 1.69239536e-01 -7.38634288e-01
6.12930298e-01 4.96295094e-01 -1.80578017e+00 1.51386276e-01
7.19349802e-01 4.95890558e-01 8.75051618e-01 2.58480340e-01
-9.07882273e-01 7.48700678e-01 -7.53655136e-01 -4.05548632e-01
-1.83244973e-01 -4.94378567e-01 -9.38169837e-01 3.23330760e-02
-3.11635762e-01 1.71763465e-01 3.76709431e-01 1.26004481e+00
3.97513181e-01 -1.24411553e-01 4.62384522e-01 -3.61788660e-01
-4.11473423e-01 1.09934664e+00 1.65865898e-01 3.97473216e-01
1.09371006e+00 -3.01800758e-01 6.67729616e-01 -8.17752898e-01
1.21529639e+00 1.27703130e+00 4.14482176e-01 3.10284048e-01
-7.48161852e-01 -5.78207850e-01 -3.42985541e-01 -5.08553721e-03
-1.25142729e+00 -5.86007893e-01 7.53723860e-01 -7.34401464e-01
1.45058310e+00 3.07311535e-01 6.70996070e-01 5.21146476e-01
3.62866551e-01 2.83536196e-01 1.21716762e+00 -1.07934463e+00
-1.81711957e-01 8.61312270e-01 4.64758992e-01 5.75079381e-01
3.88936162e-01 -5.27305365e-01 -5.18875360e-01 -1.22781739e-01
8.53624880e-01 -4.59346026e-01 7.14859962e-02 2.29030505e-01
-1.05025685e+00 7.35193431e-01 -6.27688289e-01 1.43677509e+00
-2.20043778e-01 -6.27949595e-01 7.90721118e-01 6.08878255e-01
5.84311068e-01 6.53729856e-01 -6.09076440e-01 -2.95444578e-01
-6.30945861e-01 3.00119370e-01 1.23111606e+00 1.25634766e+00
3.91303658e-01 1.34248540e-01 5.05286992e-01 9.94122207e-01
5.58530986e-01 4.44286734e-01 1.06634796e+00 -2.66850501e-01
4.58752841e-01 8.16246808e-01 -3.69259268e-02 -8.78343642e-01
-5.33749580e-01 4.07032631e-02 -1.44310251e-01 -2.65732527e-01
4.23269033e-01 2.62786448e-02 -4.96702284e-01 9.59540308e-01
4.57826078e-01 -9.88444686e-01 2.44314447e-01 4.53926533e-01
1.52747905e+00 5.52300036e-01 1.53617546e-01 -7.02129245e-01
1.55293190e+00 -5.32960415e-01 -1.46748269e+00 2.07896918e-01
9.11450028e-01 -1.56576586e+00 1.32085383e+00 5.74933589e-01
-1.52696860e+00 -3.74065191e-01 -7.96316266e-01 -3.44377369e-01
-9.77032542e-01 1.98535696e-01 7.26959229e-01 8.04985464e-01
-1.11808980e+00 3.52924079e-01 -5.38355470e-01 -8.16176832e-01
-3.56468558e-01 2.48987809e-01 -3.88004512e-01 5.27646005e-01
-1.26164567e+00 1.31316340e+00 8.13017905e-01 -4.06841874e-01
-1.67141676e-01 4.37724479e-02 -1.10689425e+00 -2.93956906e-01
9.28547457e-02 -7.80294091e-03 1.08281851e+00 -1.00985050e+00
-1.75402582e+00 1.44152284e+00 9.83533263e-02 1.71434715e-01
1.14888921e-01 1.63223490e-01 -9.13332939e-01 -2.09377065e-01
5.89504004e-01 -4.55948681e-01 1.93732440e-01 -7.95288861e-01
-5.94103813e-01 -4.38208252e-01 -1.60159960e-01 4.35522705e-01
-4.25270975e-01 1.14231431e+00 -8.35067213e-01 -9.19404507e-01
-5.84424147e-03 -1.38061643e-01 -1.78555235e-01 -9.10290837e-01
-2.91636381e-02 -3.86047572e-01 3.03623706e-01 -1.13852656e+00
2.00265241e+00 -1.95158029e+00 9.50532258e-02 2.24934116e-01
-1.78204924e-01 2.52472967e-01 3.35958362e-01 9.77587998e-01
1.26606613e-01 1.64503425e-01 3.80007029e-01 -1.93798363e-01
1.55666023e-01 4.33005482e-01 1.44428894e-01 1.31318748e-01
-4.23226029e-01 3.84470493e-01 -1.04461324e+00 -8.15405488e-01
5.04561365e-01 8.31629932e-02 -8.93217176e-02 -9.84301716e-02
-2.66081095e-01 3.49784136e-01 -5.53055704e-01 8.39931309e-01
1.98053420e-01 2.52326697e-01 8.14564466e-01 3.86529982e-01
-8.53319287e-01 9.46792066e-01 -1.44292629e+00 1.81284022e+00
-8.42650175e-01 4.91235256e-01 2.73746610e-01 -7.44972110e-01
1.18999267e+00 6.65109158e-01 -1.20244578e-01 -4.52084750e-01
6.29873991e-01 5.85400462e-01 -4.92445640e-02 -1.00054777e+00
8.78910601e-01 -1.32442892e-01 -6.68852329e-02 7.82957017e-01
5.68433702e-01 -1.17790334e-01 9.57759023e-01 -1.48241028e-01
3.60308498e-01 9.73587185e-02 1.28940094e+00 -8.14093053e-01
9.56899345e-01 3.78341794e-01 3.15339684e-01 -8.38515535e-02
2.47330412e-01 -4.40311849e-01 3.37135911e-01 -6.00373030e-01
-1.18094778e+00 -4.61058587e-01 -8.48218381e-01 1.29517508e+00
-7.12575242e-02 -7.75571406e-01 -1.00549841e+00 -1.64764702e-01
-3.40798706e-01 5.18009186e-01 -2.24983960e-01 7.56578922e-01
-5.39236486e-01 -6.25944853e-01 3.22585851e-01 1.62516385e-01
1.07922651e-01 -1.64649355e+00 -4.53291178e-01 4.59559023e-01
-1.00585213e-02 -7.32913196e-01 -4.76514660e-02 1.74900115e-01
-7.14518309e-01 -1.05104017e+00 -1.92948908e-01 -1.49758601e+00
5.96267283e-01 -1.44384995e-01 1.20344567e+00 3.96599561e-01
-4.98262420e-02 1.16778083e-01 -6.49346828e-01 -5.31123102e-01
-1.07221270e+00 3.51258039e-01 7.58206323e-02 -5.86836338e-01
7.70476460e-01 -3.08950365e-01 5.07117331e-01 2.67496556e-01
-8.07460487e-01 -2.02521354e-01 7.35340640e-02 5.68439782e-01
6.16156757e-01 2.07953781e-01 7.93878317e-01 -1.17842710e+00
1.22180176e+00 -5.23374379e-01 -7.34472513e-01 2.52602667e-01
-5.16693771e-01 -4.84229922e-02 7.80132234e-01 -4.11765784e-01
-1.27570736e+00 -2.38118559e-01 -6.87227547e-01 6.46788597e-01
-3.87860954e-01 9.21035230e-01 -5.86900890e-01 -1.90484047e-01
6.03503346e-01 9.25481915e-02 -3.18262465e-02 -1.07300889e+00
9.44748670e-02 1.17333388e+00 4.36413407e-01 -4.74593252e-01
1.39099620e-02 -1.34108275e-01 -5.03921270e-01 -1.08943427e+00
-3.84154469e-01 -8.07497203e-01 -9.42869484e-01 -5.05122483e-01
6.23491585e-01 -6.85373008e-01 -2.38233566e-01 1.64927274e-01
-1.00072396e+00 -1.69801384e-01 -2.00447500e-01 5.42836070e-01
-2.83073425e-01 2.89959967e-01 -4.29734111e-01 -5.44669867e-01
-4.92677987e-01 -1.15687442e+00 5.03972411e-01 2.08418161e-01
-6.88599825e-01 -1.86399066e+00 4.40440238e-01 1.76014274e-01
1.53434435e-02 1.45681351e-01 8.92873466e-01 -9.21130717e-01
1.60134390e-01 -9.59432572e-02 5.07072866e-01 2.62463063e-01
9.78902075e-03 2.54999906e-01 -5.75207472e-01 1.63079724e-01
2.58357704e-01 -1.35179877e-01 1.15585610e-01 1.15056150e-01
4.56291914e-01 -2.43796483e-01 -4.26382482e-01 4.40593123e-01
1.51360285e+00 7.18331814e-01 4.67327118e-01 7.19268858e-01
2.60540545e-01 7.65129328e-01 7.75000632e-01 4.54893500e-01
3.53585452e-01 6.23023868e-01 -2.48530820e-01 1.58529133e-01
1.09629005e-01 1.54510662e-01 2.86514610e-01 1.63291347e+00
-2.75522828e-01 -1.62122678e-02 -1.32247591e+00 8.20804834e-01
-1.76975775e+00 -9.25901651e-01 -5.11011362e-01 1.76371872e+00
1.22315657e+00 5.04204184e-02 1.93489909e-01 4.18282092e-01
5.87964475e-01 -3.74525599e-02 1.51860863e-01 -1.08644080e+00
-1.21314906e-01 4.25578296e-01 2.20596015e-01 8.28721464e-01
-8.07451546e-01 1.37738514e+00 7.85382891e+00 8.60411108e-01
-9.92367268e-01 4.95390862e-01 -3.15306604e-01 3.00516874e-01
-2.68205434e-01 -9.66865420e-02 -1.07795835e+00 2.29059353e-01
8.75665605e-01 -7.32522726e-01 3.53121400e-01 1.08173931e+00
1.38645470e-01 -3.35832804e-01 -6.90097749e-01 6.28273249e-01
4.40697938e-01 -1.46654606e+00 3.61325704e-02 4.73420024e-02
5.52515388e-01 -4.79233777e-03 -4.26494807e-01 8.69378671e-02
1.43381700e-01 -5.94128311e-01 9.77421403e-01 2.27635220e-01
1.02088678e+00 -7.52904594e-01 9.99056160e-01 3.39564979e-01
-1.15577054e+00 4.45691377e-01 -5.44251680e-01 -6.36029899e-01
3.38620216e-01 3.19964707e-01 -7.31268167e-01 5.41828454e-01
4.09507155e-01 7.19711781e-01 -4.48755026e-01 6.72537744e-01
-3.50978911e-01 3.42288971e-01 -1.53382629e-01 -4.30372268e-01
-1.85501240e-02 -3.68329078e-01 5.96916676e-01 1.60482371e+00
-2.62453333e-02 4.82294597e-02 1.50663331e-01 4.20512080e-01
3.81077856e-01 1.42212510e+00 -8.92598689e-01 -1.57662854e-01
5.13436139e-01 1.27758920e+00 -1.13483274e+00 -5.56743801e-01
-4.25686330e-01 3.22078615e-01 1.00066945e-01 -2.67050296e-01
-1.31791949e-01 -7.57467747e-01 3.36059630e-01 2.74041444e-01
-8.70171711e-02 -2.35673368e-01 -6.17518544e-01 -1.13572514e+00
-1.67862430e-01 -1.10518157e+00 3.42287153e-01 -3.64548385e-01
-8.54427457e-01 9.79563177e-01 3.02535564e-01 -7.23780692e-01
-7.24648416e-01 -9.23719823e-01 -5.05452156e-01 1.05990982e+00
-7.59737074e-01 -9.08726811e-01 2.00330794e-01 7.26302207e-01
5.87526798e-01 -8.02895784e-01 1.26945508e+00 5.67692041e-01
-5.34359276e-01 1.81496814e-01 1.97541118e-01 -1.67655069e-02
6.44172013e-01 -1.27597666e+00 1.52036116e-01 6.88515663e-01
1.00929253e-02 9.05377328e-01 5.72840750e-01 -8.70113850e-01
-8.47007453e-01 -4.90501851e-01 1.81326115e+00 -4.27108437e-01
1.11867917e+00 -4.69955325e-01 -5.57250857e-01 6.74374461e-01
5.95286071e-01 -8.51557076e-01 1.22265911e+00 1.96310461e-01
1.97365180e-01 2.53938675e-01 -1.06960154e+00 3.48761439e-01
7.76589334e-01 -4.56928372e-01 -9.85863388e-01 6.11284018e-01
2.05092892e-01 -5.86352408e-01 -1.28673029e+00 -3.22841763e-01
5.70711553e-01 -5.07881105e-01 5.43189347e-01 -3.14040780e-01
1.99545935e-01 -1.23602077e-01 -4.06383425e-02 -9.66676772e-01
-1.96741715e-01 -9.43812251e-01 5.33299088e-01 1.48804069e+00
5.06336629e-01 -8.50162923e-01 -1.78045332e-01 3.73768322e-02
-3.81430477e-01 -1.47076070e-01 -6.21108353e-01 -4.95484680e-01
-7.62387961e-02 -4.39060092e-01 7.86142111e-01 1.22666132e+00
1.15070927e+00 6.13694966e-01 4.30591106e-02 -2.34361097e-01
1.59132928e-01 -1.16594091e-01 6.90546274e-01 -1.40006995e+00
8.29350501e-02 -6.89288616e-01 -3.46718103e-01 -5.57983637e-01
3.08591932e-01 -9.62501526e-01 -3.13977361e-01 -1.57689619e+00
-3.10054660e-01 -4.13923800e-01 3.68372023e-01 3.59710276e-01
4.02545482e-01 1.00788809e-01 -1.22696720e-01 2.55949080e-01
-2.34089002e-01 -2.03408003e-01 1.14711595e+00 4.86066192e-01
-5.68839610e-01 -3.86349827e-01 -7.30047703e-01 1.16895103e+00
8.81221533e-01 -6.03977919e-01 -2.28903338e-01 -2.68841028e-01
4.98618245e-01 -2.15581819e-01 -5.55931509e-01 -7.26877511e-01
6.93021789e-02 -3.59547287e-01 -1.22613795e-01 -7.16913402e-01
-4.99301106e-01 -6.60383880e-01 6.28929138e-01 3.64084274e-01
1.35474533e-01 5.83558023e-01 3.04378301e-01 -4.31097060e-01
-5.42928517e-01 -9.91258860e-01 6.69698238e-01 -4.94290173e-01
-7.70776033e-01 -4.66912150e-01 -7.70823419e-01 2.79025733e-02
1.25873888e+00 -4.58683610e-01 -3.98505516e-02 1.04875505e-01
-9.20184672e-01 4.30777931e-04 8.10595632e-01 2.58569777e-01
-7.49268830e-02 -1.12439489e+00 -2.38405064e-01 6.42858207e-01
7.30039692e-03 -3.90925229e-01 -3.58759969e-01 5.38455307e-01
-1.34843421e+00 7.93664038e-01 -7.34678030e-01 5.31544201e-02
-1.38540387e+00 5.06120801e-01 1.06911354e-01 -6.39581740e-01
-5.46947777e-01 3.04869741e-01 -9.95838195e-02 -4.25081223e-01
1.42428339e-01 -4.76619512e-01 -1.04108620e+00 3.75113398e-01
8.51989806e-01 3.24719131e-01 2.05075026e-01 -1.28508818e+00
-3.88370067e-01 6.43808603e-01 1.49401203e-01 -3.47817272e-01
1.23935974e+00 -3.76427114e-01 -7.11763620e-01 9.60557699e-01
5.18879056e-01 4.45384353e-01 1.80878177e-01 -2.77524926e-02
1.09957479e-01 -1.45358220e-01 1.09853178e-01 -8.97299945e-01
-5.87973773e-01 4.51729208e-01 -4.49075662e-02 6.98024809e-01
1.01685143e+00 -1.30310981e-02 2.57063568e-01 2.52676666e-01
5.79787254e-01 -1.61499107e+00 -6.14416301e-01 8.26772273e-01
1.00749040e+00 -6.86855257e-01 1.28749281e-01 -3.93826246e-01
-6.47730410e-01 1.34129000e+00 1.46286547e-01 -4.61748196e-03
1.16241407e+00 7.15197682e-01 5.52334607e-01 -4.97469455e-01
-6.30222142e-01 -5.15999258e-01 3.05593967e-01 6.50835097e-01
1.03745639e+00 1.50716558e-01 -1.69739342e+00 9.95773196e-01
-7.32592463e-01 -1.30827323e-01 5.81268549e-01 9.58364785e-01
-5.95583022e-01 -1.65176809e+00 -7.21761167e-01 9.09737051e-02
-8.69369626e-01 -2.56862015e-01 -5.57841539e-01 1.26565468e+00
5.59420168e-01 8.76459062e-01 1.48966134e-01 -9.61234942e-02
2.57792234e-01 2.99692899e-01 4.94951755e-01 -9.69023347e-01
-1.35194349e+00 6.32093549e-01 8.09958339e-01 -2.32860208e-01
-7.51543581e-01 -9.85027790e-01 -1.16767013e+00 -4.81419355e-01
-5.43728352e-01 3.78661811e-01 1.04203951e+00 1.02568376e+00
-3.32112849e-01 2.68475473e-01 6.59728646e-02 -6.87887371e-01
1.65052354e-01 -1.39120567e+00 -9.78866518e-01 8.05164352e-02
-4.13840353e-01 -6.97835565e-01 -5.56839854e-02 4.52497065e-01] | [10.31023120880127, 10.097131729125977] |
df63461c-bb99-4831-9afc-a9c0139f3cda | a-systematic-literature-review-on-the-use-of | 2009.06520 | null | https://arxiv.org/abs/2009.06520v2 | https://arxiv.org/pdf/2009.06520v2.pdf | A Systematic Literature Review on the Use of Deep Learning in Software Engineering Research | An increasingly popular set of techniques adopted by software engineering (SE) researchers to automate development tasks are those rooted in the concept of Deep Learning (DL). The popularity of such techniques largely stems from their automated feature engineering capabilities, which aid in modeling software artifacts. However, due to the rapid pace at which DL techniques have been adopted, it is difficult to distill the current successes, failures, and opportunities of the current research landscape. In an effort to bring clarity to this crosscutting area of work, from its modern inception to the present, this paper presents a systematic literature review of research at the intersection of SE & DL. The review canvases work appearing in the most prominent SE and DL conferences and journals and spans 128 papers across 23 unique SE tasks. We center our analysis around the components of learning, a set of principles that govern the application of machine learning techniques (ML) to a given problem domain, discussing several aspects of the surveyed work at a granular level. The end result of our analysis is a research roadmap that both delineates the foundations of DL techniques applied to SE research, and highlights likely areas of fertile exploration for the future. | ['Nathan Cooper', 'Kevin Moran', 'Denys Poshyvanyk', 'David Nader Palacio', 'Cody Watson'] | 2020-09-14 | null | null | null | null | ['automated-feature-engineering'] | ['methodology'] | [ 3.81179564e-02 1.31989792e-01 -4.15574193e-01 -1.84836298e-01
-2.85080582e-01 -3.34129035e-01 5.02585709e-01 8.39022547e-03
2.40273297e-01 2.13426381e-01 -4.88593020e-02 -4.94054198e-01
-4.55012560e-01 -6.26743972e-01 -5.86023390e-01 -9.95865837e-02
-4.46601957e-02 -6.47351369e-02 -9.99508649e-02 -1.28571317e-01
5.06568849e-01 4.88691360e-01 -1.76574314e+00 3.71537626e-01
5.29087543e-01 7.53717721e-01 -7.72336349e-02 2.77603596e-01
-4.65225101e-01 1.19383883e+00 -5.32047570e-01 -3.00213039e-01
4.19988036e-02 -3.03005874e-01 -8.47337484e-01 1.29057750e-01
4.60501522e-01 1.03929546e-02 -4.40530386e-03 7.31772780e-01
1.38186276e-01 -3.47306073e-01 5.49106896e-01 -1.52603757e+00
-8.07509840e-01 5.78095615e-01 -6.62290096e-01 1.77112520e-01
3.04573089e-01 -2.03024894e-01 1.29924786e+00 -9.99360859e-01
5.09153187e-01 6.93221807e-01 9.34260309e-01 4.97628033e-01
-8.62017989e-01 -5.80107391e-01 1.21015474e-01 9.33613107e-02
-1.22004497e+00 -5.19617975e-01 9.59582508e-01 -1.05720043e+00
1.47430778e+00 -2.55741149e-01 7.76350617e-01 8.88555050e-01
7.19060302e-01 1.10552990e+00 7.81208098e-01 -1.01532733e+00
2.40257412e-01 2.86278933e-01 4.98449355e-01 7.12288439e-01
4.00026649e-01 6.62160888e-02 -8.00187588e-01 -1.47378415e-01
6.33750737e-01 3.44278431e-03 1.49176568e-01 -6.69594526e-01
-6.85252428e-01 8.59862745e-01 -1.18396617e-01 8.14687133e-01
-3.00252140e-01 1.08057112e-01 3.74939501e-01 6.76317275e-01
4.72267240e-01 6.38429165e-01 -6.37840211e-01 -5.50836802e-01
-1.08815444e+00 2.91555464e-01 1.08216488e+00 8.35811973e-01
6.96498930e-01 4.13680255e-01 5.03596008e-01 5.86855888e-01
5.56602418e-01 -1.92835495e-01 2.52439201e-01 -7.63293266e-01
1.62340596e-01 7.65991628e-01 -1.24028563e-01 -8.86522353e-01
-1.45541325e-01 -6.92333281e-01 -2.08836347e-01 4.05614942e-01
-5.59101328e-02 -3.48383456e-01 -2.06070811e-01 1.35167873e+00
-2.26132080e-01 -3.27423252e-02 -2.67518640e-01 2.38365874e-01
6.47888124e-01 2.69446671e-01 3.72499973e-02 2.75871605e-02
7.22267330e-01 -9.02400136e-01 -4.58031237e-01 -4.37843740e-01
7.70314157e-01 -6.83168292e-01 8.19317281e-01 7.65322626e-01
-9.97326314e-01 -5.72172403e-01 -1.35392320e+00 1.77305311e-01
-4.43038076e-01 9.14098811e-04 8.99698913e-01 8.36073637e-01
-1.18942308e+00 5.85311472e-01 -1.10403979e+00 -4.38371867e-01
4.09205467e-01 2.47623593e-01 -2.03359481e-02 -1.29223745e-02
-7.22788453e-01 1.15289664e+00 -1.59243882e-01 -1.14818841e-01
-6.96291089e-01 -1.02455294e+00 -8.14000547e-01 -1.11442273e-02
4.14638698e-01 -4.42714542e-01 1.55965078e+00 -1.15765023e+00
-1.07355273e+00 6.12651646e-01 2.32389569e-01 -2.83469319e-01
3.80065516e-02 -5.28838098e-01 -4.07300889e-01 -5.06326377e-01
2.39011813e-02 8.94353390e-02 6.28755212e-01 -1.11530387e+00
-1.01975822e+00 -2.92226195e-01 1.85112953e-01 -2.46624365e-01
-6.76564872e-01 4.34877425e-01 -1.28808636e-02 -4.26683813e-01
-4.12262022e-01 -5.56014180e-01 -4.56254520e-02 -2.14468822e-01
1.11850850e-01 -6.20019376e-01 6.93044364e-01 -4.82291579e-01
1.72903299e+00 -2.10624981e+00 2.33145565e-01 -1.64535493e-01
7.06222475e-01 3.15556936e-02 -6.11811969e-03 1.13299537e+00
-1.81815937e-01 1.14200816e-01 -5.54952398e-02 -1.83846071e-01
3.14233512e-01 -3.48201096e-01 -2.25050002e-01 4.65704143e-01
2.37716913e-01 9.11805689e-01 -8.01857531e-01 -9.52296704e-02
3.22517037e-01 4.37372655e-01 -1.98584765e-01 -3.57395113e-02
-2.05254421e-01 -2.68424600e-01 -5.24255335e-01 6.82199121e-01
1.04140550e-01 -2.10284919e-01 9.63014364e-02 7.04071224e-02
-6.35086417e-01 5.18308818e-01 -9.92507219e-01 1.42568731e+00
-7.04438984e-01 1.05504882e+00 -2.06960607e-02 -1.05257428e+00
1.08845818e+00 3.97854477e-01 7.12979138e-01 -3.00363630e-01
6.40333518e-02 2.54939407e-01 2.19098091e-01 -5.96711099e-01
1.61541283e-01 9.84902587e-03 -8.44565481e-02 9.80409205e-01
2.09717751e-01 5.14031351e-02 9.03223902e-02 -3.00036848e-01
1.32304072e+00 5.17683387e-01 7.09914088e-01 -2.55171359e-01
4.11244422e-01 -7.11472332e-02 3.46788734e-01 4.94612366e-01
1.87446121e-02 1.81268513e-01 5.70688784e-01 -8.09588671e-01
-9.23655987e-01 -6.48797274e-01 -1.94460019e-01 1.10121751e+00
-5.90872407e-01 -7.59182811e-01 -7.12778091e-01 -9.67972934e-01
5.47515787e-02 6.90986156e-01 -6.50192261e-01 -1.99840777e-02
-4.71208006e-01 -3.97487700e-01 3.67722064e-01 8.56384397e-01
1.48583159e-01 -1.30044162e+00 -9.56817746e-01 3.09550881e-01
5.68428457e-01 -7.12437272e-01 3.22540164e-01 1.72119245e-01
-8.85544896e-01 -1.35238409e+00 -3.31762165e-01 -9.86132503e-01
2.78092861e-01 1.67400077e-01 1.41171324e+00 1.99983224e-01
-3.72290403e-01 7.12201893e-01 -6.47386372e-01 -6.49309099e-01
-6.01610601e-01 2.10721895e-01 -1.47817507e-01 -5.35860598e-01
1.29187858e+00 -6.02582872e-01 1.92458462e-03 -2.44617268e-01
-7.80166686e-01 -5.79446517e-02 7.90603578e-01 5.24321556e-01
-1.47556096e-01 2.33025789e-01 7.56139755e-01 -8.52066040e-01
8.05076182e-01 -8.42194378e-01 -3.49525064e-01 4.91089195e-01
-1.02397084e+00 -2.83904523e-01 3.44224513e-01 8.02783668e-03
-6.92806721e-01 -4.39961314e-01 -1.10049151e-01 -9.54926852e-03
-3.23749423e-01 9.45675015e-01 4.93066572e-02 -2.31071919e-01
6.36592209e-01 2.08272919e-01 -4.60272236e-03 -5.41527152e-01
4.09916363e-04 1.07841945e+00 -1.55742824e-01 -6.11779630e-01
5.91340959e-01 -1.94615856e-01 -3.14105153e-01 -8.81571472e-01
-5.31666517e-01 -3.67823653e-02 -6.05084658e-01 -1.57875225e-01
3.20977151e-01 -4.78230804e-01 -6.73362911e-02 3.72958779e-01
-9.44650054e-01 -2.33219534e-01 -2.61026740e-01 1.29549786e-01
-4.27342832e-01 2.96410471e-01 -4.54461396e-01 -6.88949883e-01
-3.48769963e-01 -1.11082923e+00 5.92598498e-01 5.65527044e-02
-7.24840760e-01 -1.35862768e+00 2.85398662e-01 2.86295742e-01
4.08121437e-01 3.31327170e-01 1.32629943e+00 -6.97403491e-01
-4.09970582e-01 -4.08429176e-01 1.78504691e-01 7.13715613e-01
4.86747563e-01 4.84530538e-01 -9.07807410e-01 -2.30019018e-01
1.38948962e-01 -1.31742418e-01 1.51247054e-01 2.67113686e-01
7.17283309e-01 2.52472848e-01 -1.49260983e-01 1.12358138e-01
1.48497057e+00 3.54893893e-01 1.22871026e-01 8.16603363e-01
4.44457531e-01 7.54834294e-01 6.70864999e-01 4.22065318e-01
5.04197776e-01 4.21087086e-01 1.91221073e-01 4.98983078e-02
-3.03410083e-01 4.55588624e-02 4.32313234e-01 9.37568605e-01
8.99560228e-02 3.55963893e-02 -1.42919135e+00 8.20449173e-01
-1.72851515e+00 -5.50667524e-01 6.24884740e-02 1.88831699e+00
5.37441373e-01 3.38652939e-01 2.56744742e-01 2.83785254e-01
3.47609997e-01 1.38372228e-01 -2.36853436e-01 -5.68621159e-01
4.65852976e-01 5.11786163e-01 -2.75592893e-01 6.96010515e-02
-8.92018080e-01 7.36526191e-01 7.11533594e+00 2.85788476e-01
-1.04949379e+00 -6.26470670e-02 1.51956767e-01 1.72214225e-01
-3.01051706e-01 -1.17957175e-01 -6.34213388e-01 2.91100204e-01
9.68351245e-01 -6.08388603e-01 2.87161291e-01 1.29484653e+00
7.65287057e-02 1.56008169e-01 -1.74017203e+00 6.06709242e-01
2.43057743e-01 -1.47229266e+00 -3.36313665e-01 -2.72971988e-02
8.13943028e-01 2.29491144e-01 3.35680395e-01 5.95411777e-01
2.51772672e-01 -1.03897107e+00 6.55561388e-01 2.09351733e-01
4.34451222e-01 -7.88863540e-01 8.02847266e-01 3.54782611e-01
-1.00338495e+00 -3.54942232e-01 -6.83862418e-02 -6.46269321e-01
-3.17569882e-01 3.59707743e-01 -7.36090064e-01 4.97760415e-01
7.54487574e-01 1.13265657e+00 -4.09841001e-01 1.02881110e+00
-3.02150190e-01 5.42579949e-01 2.43189678e-01 -1.62153676e-01
1.90806702e-01 1.25310734e-01 1.22420341e-01 1.43729186e+00
1.63606763e-01 -4.55647588e-01 9.74718183e-02 1.12437820e+00
1.55330643e-01 -8.81460011e-02 -8.17587495e-01 -4.76487488e-01
7.47842312e-01 1.18646216e+00 -3.68983299e-01 2.14622512e-01
-1.19207561e+00 1.66250437e-01 1.75814033e-01 2.00446159e-01
-4.23162997e-01 -5.40996611e-01 6.73592329e-01 1.40116170e-01
2.25024924e-01 -3.69129777e-01 -6.88479960e-01 -7.80937672e-01
1.79767698e-01 -1.37842751e+00 -5.63928206e-03 -2.33872771e-01
-1.30203915e+00 4.85033095e-01 -5.37531264e-02 -1.16429365e+00
-3.95176917e-01 -5.50338149e-01 -7.07967579e-01 6.64560378e-01
-1.44724107e+00 -1.27274871e+00 -1.87911838e-01 -2.73945093e-01
9.43091691e-01 -6.97907329e-01 8.15782547e-01 1.82396412e-01
-5.42994857e-01 4.08852190e-01 -3.94660681e-02 -4.63749319e-02
4.07009423e-01 -1.17879736e+00 9.12269652e-01 6.47097766e-01
2.60398060e-01 1.00517714e+00 5.06925642e-01 -6.40532196e-01
-1.58800960e+00 -5.99138558e-01 1.09886181e+00 -8.01769257e-01
9.54217970e-01 -1.74382553e-01 -6.66900575e-01 1.07835412e+00
2.34589979e-01 -4.06672776e-01 1.14590538e+00 4.00245756e-01
-1.79008856e-01 -9.25054476e-02 -9.33506727e-01 3.17502081e-01
7.06195295e-01 -5.79530537e-01 -9.32043314e-01 -1.71681210e-01
2.40230277e-01 -2.17255019e-02 -1.16382635e+00 3.82338315e-01
7.45767713e-01 -1.12355316e+00 5.14762044e-01 -7.01885998e-01
8.25656116e-01 1.14451453e-01 -9.06697512e-02 -9.97475863e-01
-4.63902354e-01 -6.21769726e-01 -5.24087012e-01 1.31186116e+00
3.62282157e-01 -2.29231715e-01 8.88855398e-01 5.64784408e-01
-3.01606447e-01 -1.29027760e+00 -4.59520847e-01 -4.80128080e-01
4.12930071e-01 -6.88056588e-01 6.32925570e-01 1.11352408e+00
1.49985760e-01 1.45606995e-01 3.53516154e-02 -2.03609437e-01
5.25156915e-01 6.06113672e-02 8.05617213e-01 -1.70685041e+00
-3.57936800e-01 -6.37222886e-01 -4.68848765e-01 -7.10477054e-01
2.49564394e-01 -6.71009898e-01 -2.92317718e-01 -1.62564301e+00
1.84526607e-01 -2.56280184e-01 -5.43818116e-01 5.47438502e-01
3.09316009e-01 -1.32402897e-01 3.81695032e-02 3.57259102e-02
-4.82671052e-01 9.00205821e-02 4.73975748e-01 3.10252048e-02
-3.17577034e-01 2.99611956e-01 -1.03537321e+00 9.72174644e-01
6.01432920e-01 -2.41319403e-01 -4.49476242e-01 -6.93218648e-01
7.45560646e-01 -3.16863090e-01 -6.23880662e-02 -8.88172209e-01
1.48892358e-01 -1.29701003e-01 1.53037056e-01 -2.55973279e-01
-3.51222843e-01 -8.02480936e-01 6.59310073e-02 2.38743603e-01
-3.55786771e-01 1.22194186e-01 2.04227820e-01 2.55096424e-02
-3.07814151e-01 -5.63997507e-01 4.93127853e-01 -9.62258130e-02
-1.13181913e+00 5.92192076e-02 -5.42560220e-01 -1.37009040e-01
1.17911494e+00 -4.10803020e-01 -1.40877208e-02 -1.35701954e-01
-4.65234905e-01 4.44918424e-02 5.93596637e-01 9.19981003e-01
5.02745330e-01 -9.41972315e-01 -3.81596774e-01 3.05418968e-01
8.12997594e-02 -3.61529797e-01 -7.38190934e-02 7.05771744e-01
-3.20684969e-01 6.35879099e-01 -3.85978878e-01 -1.71545133e-01
-1.02983272e+00 2.94150412e-01 3.43139440e-01 6.09896006e-03
-7.92620540e-01 8.01640570e-01 1.95543505e-02 -1.75423920e-01
5.53407609e-01 8.46675709e-02 -2.79357880e-01 -1.56067848e-01
5.10819077e-01 7.73098469e-01 3.77658457e-01 -3.29831578e-02
-4.04072136e-01 5.82620382e-01 -4.07590538e-01 1.85650319e-01
1.89185393e+00 1.18286677e-01 -2.08110571e-01 9.03697252e-01
7.51053035e-01 -1.83473915e-01 -1.16070163e+00 -1.36045665e-01
6.87242568e-01 -1.85874104e-01 3.53829473e-01 -7.94826210e-01
-7.96643317e-01 1.02842736e+00 3.18117350e-01 4.31862414e-01
8.81042182e-01 6.02307543e-02 4.85556722e-01 2.37931862e-01
6.53840542e-01 -1.03450704e+00 1.47336245e-01 4.13353324e-01
6.56871855e-01 -8.28727126e-01 1.16481572e-01 -1.37906983e-01
-3.76903147e-01 1.69506836e+00 8.37892950e-01 -2.49292120e-01
8.46982539e-01 8.58749330e-01 -1.85215235e-01 -4.39196348e-01
-8.18391800e-01 3.29141647e-01 1.30362689e-01 8.54955852e-01
9.99299765e-01 -4.16349381e-01 -3.16913933e-01 8.71143639e-01
-6.44267499e-02 4.18192804e-01 5.86642027e-01 1.48637283e+00
-6.95574880e-01 -1.50877523e+00 7.62544051e-02 4.27600533e-01
-5.73697090e-01 1.26592994e-01 -7.10816264e-01 1.22165442e+00
3.50950629e-01 8.24000478e-01 -2.74791494e-02 -5.46701193e-01
2.26062104e-01 2.86307722e-01 6.89605474e-01 -1.06794786e+00
-8.13959301e-01 -1.83935791e-01 3.81180234e-02 -3.16440970e-01
-2.65874565e-01 -7.95500576e-01 -7.48266399e-01 -2.18547419e-01
-2.97676742e-01 9.53048170e-02 7.65145481e-01 1.22621882e+00
2.72993296e-01 6.79465175e-01 5.96925020e-01 -4.73619342e-01
-6.33869648e-01 -8.40974987e-01 -6.50025964e-01 -4.30874705e-01
4.39205527e-01 -6.39560163e-01 -2.35398963e-01 -3.66809368e-02] | [7.82426643371582, 7.564787864685059] |
90ce1ec0-71c5-4872-b080-6e462724c900 | investigating-deep-learning-benchmarks-for | 2204.04420 | null | https://arxiv.org/abs/2204.04420v1 | https://arxiv.org/pdf/2204.04420v1.pdf | Investigating Deep Learning Benchmarks for Electrocardiography Signal Processing | In recent years, deep learning has witnessed its blossom in the field of Electrocardiography (ECG) processing, outperforming traditional signal processing methods in various tasks, for example, classification, QRS detection, wave delineation. Although many neural architectures have been proposed in the literature, there is a lack of systematic studies and open-source libraries for ECG deep learning. In this paper, we propose a deep learning framework, named \texttt{torch\_ecg}, which gathers a large number of neural networks, both from literature and novel, for various ECG processing tasks. It establishes a convenient and modular way for automatic building and flexible scaling of the networks, as well as a neat and uniform way of organizing the preprocessing procedures and augmentation techniques for preparing the input data for the models. Besides, \texttt{torch\_ecg} provides benchmark studies using the latest databases, illustrating the principles and pipelines for solving ECG processing tasks and reproducing results from the literature. \texttt{torch\_ecg} offers the ECG research community a powerful tool meeting the growing demand for the application of deep learning techniques. | ['Kang Jingsu', 'Wen Hao'] | 2022-04-09 | null | null | null | null | ['ecg-qrs-detection', 'ecg-classification', 'ecg-wave-delineation', 'atrial-fibrillation-detection', 'electrocardiography-ecg'] | ['medical', 'medical', 'medical', 'medical', 'methodology'] | [ 8.56479034e-02 -2.85440475e-01 1.61227584e-01 -4.17782098e-01
-6.21954203e-01 -3.96140575e-01 -1.67380199e-01 3.90425503e-01
-4.51222569e-01 5.58661342e-01 -6.50156885e-02 -4.12267655e-01
-4.41464305e-01 -5.90151191e-01 -1.00587986e-01 -8.59329998e-01
-6.19177163e-01 3.18013132e-01 -3.13117355e-01 -1.43680409e-01
-4.16453443e-02 5.72852612e-01 -1.00223100e+00 3.05337191e-01
5.80707610e-01 1.13833487e+00 -2.28876531e-01 8.39164555e-01
2.35475346e-01 4.61671799e-01 -7.95804918e-01 -3.44680786e-01
8.93904790e-02 -7.96059310e-01 -6.58391476e-01 -4.38715011e-01
-3.20820184e-03 -8.92983973e-02 -3.15646768e-01 5.30225217e-01
1.37818575e+00 -2.39323720e-01 4.21954304e-01 -8.92644465e-01
-3.02030534e-01 6.93861187e-01 -5.40197372e-01 7.70578206e-01
1.06552042e-01 1.10156812e-01 6.46683693e-01 -6.25669241e-01
3.60163361e-01 4.25389618e-01 1.21645033e+00 4.34135973e-01
-1.03439772e+00 -6.50378108e-01 -3.26799601e-01 3.94625038e-01
-1.43873048e+00 -2.51838028e-01 8.29589069e-01 -5.00685513e-01
9.51558232e-01 4.66985375e-01 9.67483640e-01 1.15832961e+00
4.53197777e-01 6.38733029e-01 8.26287687e-01 -3.50023896e-01
9.02105272e-02 -2.36363605e-01 3.16156536e-01 3.22346061e-01
1.89354464e-01 4.68598418e-02 -4.93228316e-01 -2.15329334e-01
9.75938678e-01 5.40143810e-02 -2.24041551e-01 -5.98772280e-02
-1.34187877e+00 4.71829087e-01 1.53846681e-01 5.12882411e-01
-5.76397359e-01 -4.38289270e-02 1.06566477e+00 6.00567162e-01
3.33035797e-01 5.02400815e-01 -6.89419568e-01 -2.53977507e-01
-1.13773835e+00 3.43890309e-01 7.20721781e-01 6.81492031e-01
2.06621006e-01 3.82574201e-01 -3.66555631e-01 8.21142852e-01
9.00751874e-02 1.54954553e-01 5.62709332e-01 -7.45809555e-01
2.76063144e-01 4.84229773e-01 -3.21730226e-01 -1.04582942e+00
-1.03410280e+00 -8.37263942e-01 -1.52737689e+00 -6.24415204e-02
3.29950690e-01 -5.29180765e-01 -7.36539602e-01 1.35727191e+00
-4.01284080e-03 1.43723622e-01 -5.87025397e-02 8.09564948e-01
1.23408628e+00 3.77454817e-01 1.60066456e-01 -3.94055665e-01
1.30796301e+00 -4.21257675e-01 -7.36096323e-01 9.14012194e-02
6.25813305e-01 -4.69640702e-01 7.47182369e-01 7.06662536e-01
-1.19584548e+00 -7.96015859e-01 -1.06634617e+00 2.83796005e-02
-1.62443981e-01 1.44799997e-03 7.77246892e-01 7.73857832e-01
-1.12889111e+00 8.25325370e-01 -8.97888720e-01 -3.46371979e-01
6.55838549e-01 4.14925277e-01 -2.83415794e-01 2.39528805e-01
-1.39776099e+00 8.99888456e-01 4.20870185e-01 6.08869135e-01
-8.67630363e-01 -7.04287767e-01 -7.50646889e-01 1.70939509e-02
8.71989280e-02 -9.67340529e-01 9.45458055e-01 -6.00176454e-01
-1.27506506e+00 1.20456564e+00 2.82136798e-01 -7.47431695e-01
4.45645481e-01 -2.12615147e-01 -5.66455364e-01 8.83618891e-02
-1.37715429e-01 2.87231028e-01 8.53442371e-01 -5.74273527e-01
-2.81099200e-01 -5.26548862e-01 -3.29008222e-01 3.75087336e-02
-1.65318593e-01 2.49284431e-01 -3.77309293e-01 -9.88941669e-01
2.15105824e-02 -5.38419366e-01 -4.67584282e-01 -2.87954122e-01
-3.38974983e-01 -2.75121868e-01 3.20274413e-01 -6.71719074e-01
1.66397095e+00 -2.25633669e+00 1.71571255e-01 2.60342181e-01
8.63502145e-01 4.36763525e-01 1.52621031e-01 5.70045114e-01
-4.23180729e-01 5.86108826e-02 -5.16937733e-01 -8.29749629e-02
-1.13372713e-01 2.22432554e-01 7.15651885e-02 5.19389987e-01
2.16716100e-02 1.05381882e+00 -5.88224232e-01 -4.81321216e-01
2.96785504e-01 4.45696294e-01 -3.09678495e-01 2.38128573e-01
4.41632390e-01 6.83988214e-01 -2.86410749e-01 4.85352546e-01
4.11419749e-01 -1.86999440e-01 2.37985656e-01 -1.76929221e-01
-5.09120561e-02 8.95469114e-02 -1.03727233e+00 1.95837736e+00
-2.32081059e-02 4.66536283e-01 -1.14525966e-01 -1.39312768e+00
1.06862831e+00 6.79415762e-01 9.90446389e-01 -4.77910221e-01
5.96237242e-01 1.99521869e-01 3.96044701e-01 -7.63430834e-01
-1.01350762e-01 -9.92260799e-02 -1.16358988e-01 6.39180064e-01
2.93379247e-01 7.87922814e-02 3.61513287e-01 -1.39452800e-01
1.20836246e+00 1.22002728e-01 4.10411596e-01 -3.84188235e-01
4.75781590e-01 -3.03395689e-01 6.39028549e-01 8.18091750e-01
-4.05162722e-01 8.42113912e-01 3.88380885e-01 -1.19039679e+00
-8.13870966e-01 -9.53646839e-01 -5.13158798e-01 9.00322795e-01
-3.57619673e-01 -7.75869370e-01 -8.58411014e-01 -1.97512910e-01
-2.60680258e-01 -9.96107459e-02 -6.69026792e-01 -1.55475616e-01
-7.80568600e-01 -1.30231106e+00 1.26691985e+00 8.39195728e-01
5.01295209e-01 -1.62390733e+00 -1.18456459e+00 5.49781024e-01
-2.47277662e-01 -8.34283292e-01 -4.49075401e-02 5.71756780e-01
-1.12991011e+00 -1.08901215e+00 -7.94042766e-01 -8.72467756e-01
7.73983523e-02 -4.92611825e-01 1.44519258e+00 2.54172981e-01
-8.70848656e-01 5.04022501e-02 -2.34942600e-01 -9.56844807e-01
-1.67715251e-01 2.16024712e-01 -2.09191162e-02 -1.61108188e-02
4.59030032e-01 -8.72466326e-01 -8.15411389e-01 -7.68452361e-02
-7.21681297e-01 -5.51483147e-02 6.53037369e-01 8.04805577e-01
6.55072033e-01 -1.14042677e-01 1.09431803e+00 -8.98654580e-01
9.05808806e-01 -3.36337537e-01 -3.40463035e-02 -1.70025438e-01
-7.87507474e-01 -4.29098457e-01 5.31887114e-01 5.88185191e-02
-3.60572100e-01 -1.10284641e-01 -8.17472935e-01 -1.80268154e-01
-4.48279411e-01 7.44791210e-01 1.43701613e-01 1.55925348e-01
8.20530832e-01 3.25214148e-01 -4.97176386e-02 -5.65786183e-01
6.32205978e-02 4.75370646e-01 6.95938587e-01 -3.55445564e-01
4.75991130e-01 2.17988789e-01 -3.62823084e-02 -7.89483786e-01
-5.37099183e-01 -3.16229552e-01 -1.03800535e+00 -1.21950939e-01
1.10521781e+00 -6.02726996e-01 -6.16028666e-01 6.48557723e-01
-9.70691204e-01 -2.28286892e-01 -4.60986853e-01 2.56092250e-01
-4.63896871e-01 4.66854006e-01 -8.13745916e-01 -5.20445287e-01
-1.06616843e+00 -8.83780777e-01 8.06148469e-01 7.70591274e-02
-5.29329956e-01 -1.15933251e+00 2.41070226e-01 -4.94678579e-02
5.93288958e-01 7.39075363e-01 1.09838188e+00 -6.11799300e-01
6.18431680e-02 -4.31290805e-01 1.62340458e-02 6.26576960e-01
-5.00606000e-02 -2.05342770e-01 -1.14830613e+00 -2.14187518e-01
2.36289009e-01 -1.60367280e-01 6.48851752e-01 7.41671622e-01
1.65638161e+00 1.97017282e-01 -1.69511154e-01 1.03570318e+00
1.03006363e+00 4.56590801e-01 8.51865292e-01 2.85790384e-01
5.48116148e-01 9.84775051e-02 -9.70375314e-02 5.54691911e-01
3.61568034e-01 2.54226595e-01 1.89684510e-01 -8.05708885e-01
1.97307184e-01 4.51810420e-01 -1.99195459e-01 9.49169219e-01
-6.78461611e-01 2.18099639e-01 -1.17540789e+00 2.48741314e-01
-1.57294655e+00 -8.09446633e-01 -3.32400292e-01 2.05691147e+00
8.92983317e-01 1.62525445e-01 2.24792153e-01 8.24518085e-01
3.24363917e-01 2.46268678e-02 -5.76306224e-01 -5.49365103e-01
-1.83092803e-02 8.57035220e-01 -1.93817407e-01 -2.33184978e-01
-1.37080693e+00 3.04823846e-01 7.03101873e+00 3.30256611e-01
-1.28718936e+00 6.88030943e-02 7.46049881e-01 2.19156295e-01
5.08445561e-01 -5.04334688e-01 -3.00151050e-01 2.42975712e-01
1.10414243e+00 -2.26911157e-01 2.00892203e-02 5.30462742e-01
3.81654978e-01 3.01748723e-01 -1.28469872e+00 1.44329500e+00
9.10621211e-02 -1.54185963e+00 -2.89245009e-01 -2.37464309e-01
2.73478687e-01 2.92844355e-01 -2.54055709e-01 4.02187586e-01
-4.76875782e-01 -1.18327248e+00 2.80386358e-01 5.06228387e-01
1.04300499e+00 -6.63413405e-01 1.12234926e+00 1.80430904e-01
-1.11009252e+00 -1.27987683e-01 -9.49589536e-02 -3.03997040e-01
-2.17215102e-02 6.06509626e-01 -3.51695180e-01 9.42129612e-01
1.15568686e+00 1.15319133e+00 -7.77841985e-01 1.32193339e+00
-2.02509519e-02 8.19195151e-01 1.16587378e-01 3.41476768e-01
-1.00711770e-01 -7.70981908e-02 4.16368514e-01 1.64723420e+00
5.05831093e-02 -3.91911753e-02 3.34314078e-01 6.12193346e-01
-4.00378853e-02 3.98620605e-01 -4.97189939e-01 2.07200453e-01
1.17637336e-01 1.45897233e+00 -8.71544898e-01 -3.71481299e-01
-2.35088646e-01 6.56514585e-01 -1.82548389e-01 3.18730354e-01
-8.71640623e-01 -8.80320430e-01 4.15232658e-01 1.13222897e-01
-3.50752681e-01 -1.15106635e-01 -8.51942360e-01 -1.01841187e+00
1.21423574e-02 -1.13804412e+00 7.76478231e-01 -5.07856727e-01
-1.32401562e+00 9.81934547e-01 -4.43988368e-02 -1.15368891e+00
-1.25208184e-01 -5.70700467e-01 -7.36885011e-01 1.04101491e+00
-1.26419234e+00 -6.02070034e-01 -4.63562399e-01 8.44753325e-01
5.31181991e-01 -2.61338741e-01 1.35281539e+00 8.03642929e-01
-7.55755126e-01 6.45442128e-01 -3.21926624e-01 5.99888802e-01
6.70441628e-01 -1.28381550e+00 4.47433412e-01 6.78267896e-01
8.48399252e-02 7.99650848e-01 2.96132863e-01 -6.65476844e-02
-1.33649468e+00 -9.63270485e-01 7.60039747e-01 -4.69898731e-01
3.71984392e-01 -3.25004369e-01 -9.07226741e-01 5.30985773e-01
4.06784177e-01 1.26893604e-02 1.07518744e+00 2.76984602e-01
4.57806066e-02 -6.27436459e-01 -8.41452718e-01 3.28770280e-01
8.58381271e-01 -3.55824322e-01 -5.99381745e-01 1.78151265e-01
7.23680556e-02 -6.03790641e-01 -1.23325908e+00 6.36361539e-01
7.16528296e-01 -1.04929769e+00 8.92523527e-01 -6.48245215e-01
3.17643881e-01 -1.26629367e-01 5.52037895e-01 -1.13077867e+00
-5.16305804e-01 -9.17649448e-01 -2.43096486e-01 8.08597744e-01
3.92562777e-01 -5.68834901e-01 6.09901905e-01 1.00465842e-01
-6.96040869e-01 -1.15952551e+00 -9.60853398e-01 -2.01566562e-01
9.67327058e-02 -7.87884235e-01 4.02751982e-01 9.13323820e-01
-5.87978167e-03 3.76122326e-01 -3.95878404e-01 -1.98207244e-01
5.48253119e-01 2.54908334e-02 4.93893623e-01 -1.48959947e+00
-2.56980062e-01 -6.31982684e-01 -6.60292745e-01 -5.14273345e-01
-4.64016795e-01 -1.16553891e+00 -2.03897595e-01 -1.74198341e+00
1.27616778e-01 -4.59002882e-01 -9.24405456e-01 7.40979850e-01
-1.06769852e-01 7.39725292e-01 -4.95038182e-02 1.50891870e-01
-4.55555260e-01 -1.04694530e-01 1.18952203e+00 2.11055800e-02
-4.47468579e-01 1.86771661e-01 -9.83088553e-01 8.79393220e-01
9.67903614e-01 -4.07969922e-01 -2.68918544e-01 -4.99203682e-01
3.02792877e-01 3.31411064e-02 3.15320462e-01 -1.40279758e+00
5.11879008e-03 6.04340374e-01 8.54571223e-01 -6.24103427e-01
-4.43893448e-02 -5.02159894e-01 1.94063798e-01 6.24143541e-01
-3.33282739e-01 4.70922917e-01 2.70067275e-01 1.09758608e-01
-2.83131033e-01 1.97789848e-01 7.43279219e-01 -2.62617022e-01
-4.86689299e-01 5.58920264e-01 -5.98139226e-01 1.86542705e-01
8.28092635e-01 -2.76996225e-01 2.05890745e-01 1.86247453e-02
-1.32608163e+00 2.24745095e-01 -4.63755995e-01 4.19444740e-01
7.34661758e-01 -9.96556282e-01 -9.50902581e-01 5.56631386e-01
1.18590891e-02 7.76659101e-02 4.77570564e-01 1.41987443e+00
-7.56242275e-01 1.98910341e-01 -5.84640384e-01 -8.87044251e-01
-1.09500396e+00 3.86628866e-01 6.83833539e-01 -3.85039210e-01
-1.27011764e+00 6.76238239e-01 -4.91046570e-02 -8.48838910e-02
6.15504205e-01 -6.01860464e-01 -5.09438276e-01 7.77172595e-02
6.35215282e-01 4.11502421e-01 5.64921081e-01 -1.41356945e-01
-4.57357913e-01 4.79775995e-01 1.95034146e-01 5.44385135e-01
1.57448065e+00 6.79741129e-02 -3.58033359e-01 6.70129538e-01
7.29820907e-01 -5.02589285e-01 -5.88175237e-01 1.87870786e-01
2.14548782e-01 1.37676671e-01 -9.52723026e-02 -1.12036848e+00
-1.28264475e+00 1.22504354e+00 1.00948834e+00 3.88642222e-01
1.63807762e+00 -3.47529531e-01 8.47971499e-01 3.29551756e-01
4.60627019e-01 -8.98073018e-01 -2.19813839e-01 4.54744488e-01
8.46127391e-01 -7.72318184e-01 -1.15157850e-01 -1.68457394e-03
-6.02047741e-01 1.34021819e+00 1.98531866e-01 -2.60369778e-01
9.08967733e-01 5.81283450e-01 4.93702888e-01 -4.42435414e-01
-4.47531313e-01 1.40953064e-01 1.25817150e-01 7.41882503e-01
9.25021410e-01 -2.39704903e-02 -5.52929759e-01 1.08529401e+00
-1.56587049e-01 3.84485394e-01 2.78782159e-01 8.64277661e-01
-8.89196619e-02 -1.20479715e+00 -6.14967272e-02 7.47290909e-01
-9.81710553e-01 -3.09202313e-01 -1.37781322e-01 6.84845507e-01
5.46431780e-01 6.90209448e-01 -3.69575858e-01 -3.31364840e-01
4.86528218e-01 3.57995242e-01 5.52429259e-01 -7.23923683e-01
-1.24300218e+00 3.83811682e-01 -1.04557574e-01 -3.53015274e-01
-3.19384545e-01 -4.96363372e-01 -1.22875857e+00 1.15308501e-01
1.81347162e-01 2.21356377e-01 2.73193985e-01 9.46505368e-01
4.24697876e-01 1.05660427e+00 3.13282311e-01 -1.00234640e+00
-2.42952108e-01 -1.13141763e+00 -8.85340691e-01 2.57573843e-01
2.43760645e-01 -3.25644851e-01 9.89592522e-02 2.97122389e-01] | [14.329315185546875, 3.266768455505371] |
b0e26bf0-aee6-4f29-ad88-7d8c7eaae5b9 | towards-a-broad-coverage-named-entity | 2201.12219 | null | https://arxiv.org/abs/2201.12219v2 | https://arxiv.org/pdf/2201.12219v2.pdf | Towards a Broad Coverage Named Entity Resource: A Data-Efficient Approach for Many Diverse Languages | Parallel corpora are ideal for extracting a multilingual named entity (MNE) resource, i.e., a dataset of names translated into multiple languages. Prior work on extracting MNE datasets from parallel corpora required resources such as large monolingual corpora or word aligners that are unavailable or perform poorly for underresourced languages. We present CLC-BN, a new method for creating an MNE resource, and apply it to the Parallel Bible Corpus, a corpus of more than 1000 languages. CLC-BN learns a neural transliteration model from parallel-corpus statistics, without requiring any other bilingual resources, word aligners, or seed data. Experimental results show that CLC-BN clearly outperforms prior work. We release an MNE resource for 1340 languages and demonstrate its effectiveness in two downstream tasks: knowledge graph augmentation and bilingual lexicon induction. | ['Hinrich Schütze', 'Philipp Dufter', 'Ayyoob Imani', 'Silvia Severini'] | 2022-01-28 | null | https://aclanthology.org/2022.lrec-1.417 | https://aclanthology.org/2022.lrec-1.417.pdf | lrec-2022-6 | ['transliteration'] | ['natural-language-processing'] | [-1.40112624e-01 -5.99817233e-03 -8.30215096e-01 -1.89057410e-01
-9.98021603e-01 -1.13866603e+00 6.43484354e-01 1.62236305e-04
-8.24525177e-01 1.16603255e+00 5.34434021e-01 -9.08853173e-01
4.24808323e-01 -6.62400544e-01 -7.65493631e-01 1.26365900e-01
1.38450503e-01 1.15281570e+00 -7.07542300e-02 -4.45092261e-01
-3.60084958e-02 3.69907856e-01 -4.02744740e-01 9.25504938e-02
1.09242010e+00 -2.28939012e-01 8.92780572e-02 1.78231061e-01
-4.57630336e-01 4.43062782e-01 -4.14949000e-01 -8.92752171e-01
3.88649166e-01 -6.64450467e-01 -1.28193676e+00 -4.14057761e-01
5.35074711e-01 2.22615644e-01 -1.88908890e-01 1.11085987e+00
4.00118500e-01 -4.41361010e-01 5.32266796e-01 -1.09981024e+00
-1.01238263e+00 1.50263786e+00 -5.13748944e-01 5.31734467e-01
2.34190270e-01 3.30845527e-02 1.33409333e+00 -1.05534649e+00
1.47984195e+00 1.09717798e+00 8.57830644e-01 3.94972235e-01
-1.34924531e+00 -9.86781955e-01 -2.53640972e-02 8.27715397e-02
-1.58111322e+00 -7.61204839e-01 2.93773353e-01 -3.91132534e-01
1.61454165e+00 -2.05176651e-01 4.86700207e-01 1.16342902e+00
-2.00137034e-01 6.14216745e-01 1.18793654e+00 -8.65025818e-01
-4.85062003e-01 1.69674203e-01 1.06645815e-01 8.58873487e-01
5.76824665e-01 -1.45738617e-01 -7.91489780e-01 -6.77219778e-02
7.24859893e-01 -9.10311401e-01 -1.87050804e-01 3.27894427e-02
-1.73642778e+00 6.85251057e-01 6.77620098e-02 5.14815569e-01
-2.28301272e-01 -3.33264142e-01 4.75108892e-01 6.82134807e-01
5.84709525e-01 8.69259298e-01 -9.04509783e-01 7.07912892e-02
-7.37306595e-01 -1.51935533e-01 1.17592287e+00 1.64673221e+00
8.71809840e-01 -6.20263182e-02 2.68033803e-01 1.01957393e+00
-1.21040039e-01 9.59613562e-01 3.72889429e-01 -4.49875921e-01
9.72138941e-01 4.33712870e-01 -4.13592160e-02 -6.92572236e-01
-5.69846332e-01 -2.44927555e-01 -5.20925999e-01 -6.88607812e-01
4.54187840e-01 -4.14028913e-01 -6.65674329e-01 1.96429002e+00
2.33384818e-01 -2.83706069e-01 4.59483355e-01 5.36891818e-01
1.01461399e+00 3.84321988e-01 7.54174888e-02 1.63372159e-02
1.41641700e+00 -1.06129611e+00 -6.77049458e-01 -5.48210919e-01
9.05855834e-01 -1.20275831e+00 9.67346191e-01 -2.19475269e-01
-9.13171887e-01 -1.97546110e-01 -7.64577866e-01 -1.88060194e-01
-8.01075637e-01 1.64145440e-01 1.05258918e+00 5.22773087e-01
-1.34034538e+00 1.01648919e-01 -7.89718270e-01 -8.16385925e-01
1.05162859e-01 2.66646624e-01 -7.50584900e-01 -1.66691571e-01
-1.53833163e+00 1.35358310e+00 9.20551896e-01 -1.38451025e-01
-5.42533338e-01 -5.90655208e-01 -1.14237165e+00 -3.35813940e-01
3.39510262e-01 -5.32262027e-01 9.45292592e-01 -8.81791592e-01
-1.05116653e+00 1.44108701e+00 -1.65617466e-02 -2.74296552e-01
4.96119149e-02 1.80192329e-02 -9.51248169e-01 -2.04470322e-01
7.89783180e-01 8.50526631e-01 -1.07887365e-01 -8.64336908e-01
-6.55356348e-01 4.79096919e-02 -2.19555080e-01 3.26214284e-01
-2.54840761e-01 6.18632674e-01 -8.33481550e-01 -7.02920139e-01
-2.40186490e-02 -1.12783313e+00 -2.77127594e-01 -1.08125317e+00
-6.93664312e-01 -1.60170436e-01 1.09803081e-01 -1.27053118e+00
1.12646413e+00 -1.50454795e+00 -5.39772362e-02 4.03487504e-01
-2.30685752e-02 -7.85461590e-02 -6.70829713e-01 6.57175422e-01
-1.83322474e-01 3.57643694e-01 -1.09066680e-01 4.46115844e-02
-4.77341898e-02 3.54496300e-01 1.89822086e-03 4.54113305e-01
4.71515298e-01 1.15307832e+00 -1.06093049e+00 -7.34208405e-01
-4.11890328e-01 6.08858950e-02 -5.54776669e-01 -2.60492295e-01
-1.87952742e-01 6.57837689e-01 -4.44254540e-02 8.74724805e-01
3.94125134e-01 -2.76395828e-01 8.77902210e-01 -4.39261124e-02
-2.34399110e-01 8.85890722e-01 -8.37609053e-01 1.92629552e+00
-7.98047662e-01 7.07544208e-01 -1.12964168e-01 -5.28473973e-01
7.93787181e-01 3.56636226e-01 1.88947782e-01 -7.84685671e-01
-1.58572257e-01 8.32621753e-01 2.14838043e-01 -1.20905057e-01
6.95510507e-01 -1.33284852e-01 -4.81817842e-01 6.86176777e-01
6.56674325e-01 1.29354209e-01 7.79854059e-01 4.03773516e-01
9.54955161e-01 2.20754609e-01 7.86686718e-01 -6.86965764e-01
3.33050340e-01 7.91643322e-01 8.95261645e-01 4.75832283e-01
1.13848336e-01 -3.67798135e-02 1.89461857e-01 -3.47210497e-01
-1.34584689e+00 -9.14326966e-01 -1.42006502e-01 1.24061275e+00
-1.38230577e-01 -6.01383805e-01 -5.98472178e-01 -9.77626562e-01
-2.17852145e-01 6.62322640e-01 -6.63235113e-02 3.40089619e-01
-1.15090561e+00 -1.00150883e+00 1.09756923e+00 3.22989464e-01
9.27775055e-02 -1.19935238e+00 5.92746139e-01 3.71452779e-01
-8.00394893e-01 -1.73542047e+00 -1.00657320e+00 2.00117528e-01
-5.25812149e-01 -1.22715819e+00 -2.67363012e-01 -1.38833392e+00
7.43324459e-01 -2.53416657e-01 1.94075656e+00 -1.89263836e-01
9.37973559e-02 1.05183668e-01 -6.19774684e-02 -1.44232735e-01
-9.61872876e-01 7.92103469e-01 3.36957544e-01 -6.69441164e-01
9.73027706e-01 -5.03346562e-01 9.46852341e-02 1.31818354e-01
-4.41565841e-01 2.94059098e-01 7.87081122e-01 7.89032519e-01
5.44778824e-01 -5.53440988e-01 7.79043376e-01 -1.32664263e+00
4.95179445e-01 -5.58551908e-01 -9.36037242e-01 6.17874205e-01
-5.89960217e-01 2.08225399e-01 5.49589813e-01 -3.36947948e-01
-1.01438856e+00 -1.13330267e-01 -1.37462988e-01 3.28799754e-01
2.01666951e-02 7.85124063e-01 -3.61201078e-01 -5.42102791e-02
7.06319749e-01 1.08311079e-01 -5.68954110e-01 -6.63933516e-01
8.30178440e-01 8.04841340e-01 8.76609862e-01 -9.85197246e-01
7.54574835e-01 -6.12736084e-02 -3.19297671e-01 -5.81047118e-01
-7.84296930e-01 -6.22704983e-01 -1.16953981e+00 3.67994189e-01
7.68115342e-01 -1.32520974e+00 -2.89599717e-01 -4.55674902e-03
-1.37709522e+00 -4.26777363e-01 1.80388242e-02 7.87402093e-01
-1.32384315e-01 -5.78989217e-04 -1.14223933e+00 5.42726405e-02
-5.32566607e-01 -9.11274850e-01 7.18433678e-01 -6.18849620e-02
-5.33399165e-01 -1.41070807e+00 6.46629214e-01 3.74588460e-01
8.82617608e-02 3.99866514e-03 1.22850382e+00 -1.22389567e+00
-5.17928064e-01 2.42722169e-01 -1.84365943e-01 5.90624474e-02
2.65202612e-01 -2.52684295e-01 -2.24530816e-01 -4.04593110e-01
-7.09651053e-01 -4.89819139e-01 5.03259301e-01 -2.36826539e-01
6.40562400e-02 -5.29931366e-01 -4.03887957e-01 7.80431807e-01
1.55962384e+00 9.38531533e-02 1.12692885e-01 6.64077580e-01
9.65832770e-01 4.71527815e-01 -6.50816858e-02 -3.05682749e-01
8.97110224e-01 4.11520541e-01 -4.95008171e-01 -4.68091339e-01
-4.74447399e-01 -5.05103886e-01 4.61418867e-01 2.07762575e+00
-2.62653013e-03 -3.97499986e-02 -1.68417048e+00 1.12029731e+00
-1.33874381e+00 -5.91652036e-01 -1.03168949e-01 1.93350816e+00
1.68951428e+00 -1.49566919e-01 -2.55590193e-02 -9.06790018e-01
1.01978123e+00 -1.98931590e-01 -3.16896766e-01 -1.99836642e-01
-7.75429368e-01 6.53493762e-01 1.07614982e+00 7.19820738e-01
-1.07007027e+00 1.88782382e+00 6.76229906e+00 7.18585551e-01
-7.37987816e-01 6.02554321e-01 1.38485461e-01 3.14032435e-01
-6.05441511e-01 3.87142956e-01 -1.29139996e+00 8.30974281e-02
1.07285392e+00 -3.13046038e-01 6.50419295e-01 5.38674831e-01
-4.13579553e-01 2.84270227e-01 -1.17645085e+00 7.05575943e-01
2.05823377e-01 -1.37733757e+00 1.34769052e-01 2.01002248e-02
1.30913126e+00 9.36029613e-01 -4.40868497e-01 4.86142486e-01
1.42038131e+00 -9.47446287e-01 5.47107458e-01 6.50936067e-02
1.47700834e+00 -7.23622739e-01 7.00364411e-01 1.13809533e-01
-1.12954056e+00 6.92779899e-01 -3.37879181e-01 3.24828833e-01
2.69646406e-01 3.48274946e-01 -1.10423911e+00 8.03271413e-01
2.88187921e-01 7.32961118e-01 -7.51291394e-01 5.80482423e-01
-5.74745119e-01 9.19809997e-01 -3.80180627e-01 1.24582268e-01
4.46535915e-01 -4.15371329e-01 6.00795269e-01 1.83277631e+00
7.46925920e-02 -3.67873430e-01 5.40890813e-01 6.62940681e-01
-7.35685945e-01 7.37278581e-01 -8.27478945e-01 -4.27181214e-01
7.59266317e-01 1.30411839e+00 -9.14528668e-01 -4.92633045e-01
-8.99294198e-01 1.03546703e+00 8.08334887e-01 5.49801946e-01
-5.43198347e-01 -5.19169211e-01 5.34798920e-01 -3.56801182e-01
5.22401631e-02 -3.67523581e-01 -1.08128441e-02 -1.69470441e+00
-1.39494047e-01 -1.42615199e+00 5.38294435e-01 -3.55948687e-01
-1.59356737e+00 9.74374533e-01 -2.12677479e-01 -6.92385554e-01
-3.74879032e-01 -6.87672555e-01 -4.46806289e-02 1.29348099e+00
-1.55862713e+00 -1.62205100e+00 4.84799385e-01 6.83056891e-01
3.04597229e-01 -7.44469285e-01 1.10625148e+00 6.27960980e-01
-6.16177857e-01 6.95926607e-01 8.94828141e-02 8.92154038e-01
1.17516303e+00 -1.29876971e+00 1.16551983e+00 1.13180363e+00
8.23855579e-01 9.55854595e-01 1.91190019e-01 -1.10740149e+00
-1.00693381e+00 -1.13806045e+00 1.74301636e+00 -8.55131328e-01
1.36910868e+00 -7.11597621e-01 -5.28628647e-01 1.38978064e+00
7.11636662e-01 -3.68729651e-01 9.97029126e-01 7.40035236e-01
-6.40468895e-01 2.59943783e-01 -4.72004145e-01 8.62002909e-01
1.41034162e+00 -9.84122992e-01 -8.25045705e-01 6.90807045e-01
7.50011325e-01 -4.40576047e-01 -1.17548406e+00 2.26934284e-01
1.28202721e-01 -7.49290437e-02 7.52546787e-01 -1.15860462e+00
1.61857754e-01 -1.59413278e-01 -5.97096086e-02 -1.59065259e+00
-2.04808637e-01 -7.58206606e-01 4.60134059e-01 1.48746324e+00
1.21966934e+00 -7.25881755e-01 4.14501399e-01 -1.52210370e-01
-1.87319443e-01 -1.32216260e-01 -6.78397477e-01 -9.22991335e-01
5.16676188e-01 -2.11136624e-01 6.66401684e-01 1.86925387e+00
9.14747417e-02 1.10540426e+00 -2.13641733e-01 2.31971323e-01
5.02799094e-01 1.55988365e-01 6.71788037e-01 -8.72921884e-01
-1.06184267e-01 -4.08304989e-01 1.08761363e-01 -6.78211331e-01
8.14178884e-01 -1.82337320e+00 9.45111886e-02 -1.64462996e+00
4.88554031e-01 -7.13292837e-01 -1.60808593e-01 9.77945447e-01
-3.12325746e-01 3.41658741e-01 -1.79006353e-01 4.86216187e-01
-5.44303954e-01 1.79940388e-01 9.54591870e-01 -1.96244389e-01
-1.57301828e-01 -6.63204610e-01 -6.62553132e-01 7.71390021e-01
7.41943002e-01 -8.13657582e-01 1.49479955e-02 -7.60457516e-01
6.59429073e-01 -1.73827529e-01 -2.75083810e-01 -4.09366488e-01
3.86304647e-01 -1.95305198e-01 2.45088592e-01 -4.24499601e-01
-4.78308588e-01 -3.21423322e-01 1.57813251e-01 1.57921374e-01
-8.95696804e-02 6.91793621e-01 3.55500311e-01 -2.99628377e-02
-3.04398596e-01 -2.76705362e-02 4.04400915e-01 -5.20749450e-01
-9.32075858e-01 2.30826631e-01 -2.71901727e-01 8.50125134e-01
2.99205691e-01 4.50917304e-01 -5.04453659e-01 7.51950964e-02
-4.84936774e-01 3.36700648e-01 5.52266002e-01 5.91634035e-01
-1.11045972e-01 -1.55119240e+00 -1.10362041e+00 1.92633569e-01
4.08796817e-01 -4.30667579e-01 -5.11116564e-01 7.45801568e-01
-8.09566975e-01 7.40908682e-01 -3.48102689e-01 -1.36494592e-01
-9.03650403e-01 5.09982526e-01 9.22127888e-02 -7.12311029e-01
-4.54057634e-01 5.79031348e-01 -1.06123872e-01 -1.55142248e+00
-3.36481422e-01 1.45689607e-01 -1.89645529e-01 -1.20842613e-01
1.17655173e-02 -7.64394626e-02 1.18910298e-01 -1.02579451e+00
-5.46384454e-01 3.11147332e-01 -9.86073688e-02 -4.81635332e-01
1.27154255e+00 -1.79651737e-01 -7.33001709e-01 2.47400880e-01
9.51777518e-01 8.24534476e-01 -2.48951986e-01 -7.04964936e-01
8.30042541e-01 9.33985859e-02 -3.99342299e-01 -9.84219790e-01
-8.93832922e-01 8.73048529e-02 -2.40115732e-01 -5.93628585e-01
7.52325118e-01 2.92238683e-01 9.21486676e-01 7.34847367e-01
6.45022273e-01 -1.17475951e+00 -4.15672034e-01 1.19817078e+00
4.22739148e-01 -1.38632441e+00 -2.46347055e-01 -4.07305628e-01
-5.33848882e-01 7.86895931e-01 5.94147980e-01 4.83190596e-01
4.82655495e-01 4.85181928e-01 6.38365269e-01 -3.89549166e-01
-4.12877798e-01 -4.99894708e-01 4.43062842e-01 4.30395544e-01
7.66382158e-01 2.09983394e-01 -5.54231286e-01 5.16010702e-01
-7.25158453e-01 -4.69658673e-01 3.28877211e-01 6.76853836e-01
2.77184725e-01 -1.58757949e+00 -1.53665483e-01 1.05291039e-01
-7.41978467e-01 -1.15204477e+00 -5.45382679e-01 1.25936043e+00
2.95030296e-01 7.33993113e-01 1.50754198e-01 -8.26173127e-02
9.52122137e-02 3.62437457e-01 4.40298855e-01 -8.93223405e-01
-6.48086190e-01 6.33773282e-02 7.72740662e-01 -2.74144918e-01
-6.52915776e-01 -6.30257726e-01 -1.04614866e+00 -4.32187915e-01
-3.21750194e-01 5.08426011e-01 5.30002356e-01 8.41281593e-01
2.10204378e-01 1.10018983e-01 1.91229805e-01 -1.56364143e-01
1.69740140e-01 -1.08185959e+00 -2.64116704e-01 3.45300496e-01
-3.60321492e-01 -1.34333089e-01 1.25351846e-01 2.88561583e-01] | [9.99817943572998, 9.519619941711426] |
5ff67d3d-197f-48f5-8698-49c20484798e | exploration-in-deep-reinforcement-learning-a-1 | 2205.00824 | null | https://arxiv.org/abs/2205.00824v1 | https://arxiv.org/pdf/2205.00824v1.pdf | Exploration in Deep Reinforcement Learning: A Survey | This paper reviews exploration techniques in deep reinforcement learning. Exploration techniques are of primary importance when solving sparse reward problems. In sparse reward problems, the reward is rare, which means that the agent will not find the reward often by acting randomly. In such a scenario, it is challenging for reinforcement learning to learn rewards and actions association. Thus more sophisticated exploration methods need to be devised. This review provides a comprehensive overview of existing exploration approaches, which are categorized based on the key contributions as follows reward novel states, reward diverse behaviours, goal-based methods, probabilistic methods, imitation-based methods, safe exploration and random-based methods. Then, the unsolved challenges are discussed to provide valuable future research directions. Finally, the approaches of different categories are compared in terms of complexity, computational effort and overall performance. | ['Hyondong Oh', 'Minwoo Kim', 'Lilian Weng', 'Pawel Ladosz'] | 2022-05-02 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-2.01410770e-01 1.41282991e-01 -6.25909388e-01 8.76225978e-02
-3.85422647e-01 -3.33151042e-01 5.38270950e-01 -6.45592213e-02
-6.28864646e-01 1.51059258e+00 -1.22979574e-01 -1.07696466e-02
-4.14380312e-01 -5.78378439e-01 -3.78927350e-01 -1.09240162e+00
-6.73617125e-01 4.21457976e-01 6.18837327e-02 -4.37634498e-01
6.14570081e-01 4.77578163e-01 -1.66810048e+00 -4.90858018e-01
9.22175646e-01 8.43510926e-01 4.17166024e-01 3.01627636e-01
-7.66317472e-02 9.60556507e-01 -7.91685581e-01 2.33943895e-01
3.60065222e-01 -8.45933914e-01 -7.45380998e-01 -2.61633009e-01
-8.10237527e-01 -5.59739232e-01 -3.68692607e-01 1.20317483e+00
5.10723352e-01 6.04972005e-01 6.04136407e-01 -1.36269486e+00
-2.35905260e-01 9.31368530e-01 -4.71772373e-01 2.46813342e-01
4.17432338e-01 2.55477160e-01 7.27145791e-01 -4.46383148e-01
5.24583519e-01 1.20648372e+00 2.42878720e-02 8.24168622e-01
-9.39168036e-01 -6.16371036e-01 6.07242405e-01 5.83641410e-01
-8.37978721e-01 -5.18545210e-02 8.48520815e-01 1.56446129e-01
1.08004856e+00 1.79016143e-01 1.14119375e+00 1.27908301e+00
3.63733262e-01 1.19370019e+00 1.51593864e+00 -2.34724775e-01
9.47100580e-01 -1.21592909e-01 -3.73013467e-01 2.65168697e-01
1.52743116e-01 1.00392747e+00 -6.02241814e-01 -1.27353072e-01
9.25130010e-01 -5.15449122e-02 2.57872015e-01 -8.12325001e-01
-1.21378219e+00 1.15903676e+00 5.89321017e-01 3.28305662e-01
-7.31903493e-01 4.44609851e-01 4.22613412e-01 4.05939817e-01
-3.46060812e-01 6.90854311e-01 -1.37757897e-01 -7.74165809e-01
-8.03935647e-01 7.38166034e-01 7.00368702e-01 7.42534459e-01
5.89617014e-01 7.42921591e-01 -1.41550273e-01 6.76225841e-01
3.75074118e-01 3.95918190e-01 7.07093775e-01 -1.11763942e+00
1.97611496e-01 1.03162035e-01 4.72345561e-01 -4.71423060e-01
-4.50474977e-01 -2.78292626e-01 -7.16099799e-01 9.07344103e-01
-9.72808897e-03 -4.63389933e-01 -7.09522128e-01 1.46695423e+00
3.79461080e-01 -5.96197993e-02 2.77826488e-01 1.01458478e+00
6.40682936e-01 3.46106350e-01 -2.93563027e-02 -5.47047734e-01
7.44465411e-01 -1.14279211e+00 -1.07089317e+00 -2.72089511e-01
1.74678713e-01 -3.82912755e-01 7.96113193e-01 5.21843672e-01
-1.14812934e+00 -3.18935812e-01 -1.07274187e+00 7.50340044e-01
-2.74744213e-01 -2.80825168e-01 9.33834493e-01 4.35557038e-01
-7.49474823e-01 9.67189431e-01 -9.38093781e-01 -2.32528113e-02
4.52151209e-01 4.77609009e-01 2.73751020e-01 2.79727042e-01
-1.29490626e+00 1.14238834e+00 6.87564850e-01 3.57684195e-02
-1.58275199e+00 1.71537533e-01 -7.77152896e-01 -1.31304771e-01
6.07569635e-01 1.76091436e-02 1.55868769e+00 -6.79827452e-01
-2.20122647e+00 3.76513973e-02 1.10384375e-01 -6.52277470e-01
5.80115795e-01 -2.36948267e-01 -1.56771764e-02 4.45661275e-03
5.20706270e-03 6.60348058e-01 6.65967405e-01 -1.15767825e+00
-6.97335303e-01 1.78194400e-02 3.24920177e-01 6.71961486e-01
2.34975778e-02 -3.22521701e-02 1.99693441e-01 -5.02319038e-01
-1.40802249e-01 -9.23019230e-01 -9.68153119e-01 -5.06906688e-01
-2.80676603e-01 -5.12851596e-01 5.29699743e-01 2.95114070e-02
9.77321386e-01 -1.81683707e+00 2.60041565e-01 3.30130130e-01
-6.60684556e-02 1.23573363e-01 -1.03497788e-01 8.20634067e-01
2.37762704e-01 -3.62394959e-01 -1.87512800e-01 1.86215654e-01
2.17575073e-01 5.38866520e-01 -5.15973091e-01 2.56926715e-01
-2.11328641e-02 1.12934351e+00 -1.45090187e+00 -4.55715656e-01
4.24166650e-01 1.95056517e-02 -2.21996844e-01 5.77937007e-01
-2.47950330e-01 6.95642829e-01 -9.05174911e-01 9.59432721e-01
2.61565119e-01 1.17838845e-01 2.58798391e-01 7.78789282e-01
-4.32596236e-01 4.24353898e-01 -1.51663792e+00 1.48843539e+00
-2.22990632e-01 1.46121681e-01 -3.93959470e-02 -1.19054937e+00
1.18242371e+00 2.17499107e-01 6.33837104e-01 -8.78411174e-01
2.50298381e-01 5.05902350e-01 4.16704938e-02 -2.42063776e-01
6.71014488e-01 9.32443142e-02 -1.74579531e-01 6.45297945e-01
-2.66961336e-01 -3.22768271e-01 4.25555199e-01 -1.65711626e-01
1.11190951e+00 5.94835341e-01 7.47972012e-01 -7.36694485e-02
2.46954694e-01 2.39012137e-01 5.59825361e-01 1.03053284e+00
-6.08554244e-01 -1.62427634e-01 5.69967389e-01 -4.54433620e-01
-5.76721966e-01 -1.06366169e+00 1.19463779e-01 9.28234518e-01
5.66905081e-01 -5.61694875e-02 -4.37655002e-01 -8.47818375e-01
-1.90269634e-01 5.09068191e-01 -7.63962448e-01 -2.08342284e-01
-6.80822790e-01 -6.42193198e-01 1.31533474e-01 6.26267910e-01
5.07445693e-01 -1.94098318e+00 -1.50461364e+00 4.21659857e-01
4.93225269e-02 -4.45794523e-01 2.75319070e-01 6.60598040e-01
-1.20610702e+00 -1.02848983e+00 -1.02196646e+00 -6.08257413e-01
4.07418698e-01 1.09520182e-01 8.85439634e-01 1.06188618e-01
-2.40175322e-01 2.55085558e-01 -7.49769211e-01 -4.33156788e-01
1.35486452e-02 -2.15510011e-01 2.86464930e-01 -8.65963042e-01
4.40288037e-01 -4.71836179e-01 -6.40232205e-01 2.99087644e-01
-4.83639657e-01 -5.59553802e-01 8.07215750e-01 1.33789754e+00
7.30083346e-01 8.85165781e-02 9.34246659e-01 -1.44043043e-01
9.90884542e-01 -4.91567016e-01 -8.23677599e-01 -1.08749203e-01
-7.33163059e-01 1.48870111e-01 6.49984717e-01 -7.23965645e-01
-8.90451908e-01 3.61219309e-02 -1.37755916e-01 -2.21415162e-01
-1.50830820e-01 3.29627424e-01 2.95924008e-01 -2.67197132e-01
6.59249783e-01 5.26134670e-01 1.19151376e-01 -3.01039904e-01
3.63182127e-01 4.19679701e-01 -3.21097299e-02 -5.88431656e-01
5.02491057e-01 2.79068679e-01 2.33283583e-02 -4.93582368e-01
-3.56709927e-01 -1.42645344e-01 -2.92805761e-01 -5.28131366e-01
4.21360075e-01 -4.07513380e-01 -8.58415663e-01 1.96253568e-01
-7.03087091e-01 -6.71566904e-01 -9.57296848e-01 8.36000144e-01
-1.20907104e+00 2.83444673e-01 -3.90963227e-01 -1.34558630e+00
-2.55937636e-01 -1.63908517e+00 4.87915874e-01 7.74660647e-01
-2.40676969e-01 -6.22566283e-01 3.40759009e-01 -2.36777887e-01
4.80525643e-01 2.94257402e-01 3.81256580e-01 -5.71689129e-01
-7.10948527e-01 1.92955419e-01 2.63640642e-01 -1.98694374e-02
-3.62234213e-03 -4.06465888e-01 -4.13247794e-01 -3.85154545e-01
-1.00236848e-01 -1.00321591e+00 5.74479163e-01 5.09025455e-01
9.86457705e-01 -1.96985930e-01 -3.48946124e-01 2.08272696e-01
1.19934177e+00 8.75684738e-01 6.99575901e-01 9.21334743e-01
2.29977537e-02 5.89844823e-01 1.47759104e+00 1.01936769e+00
6.49395287e-02 5.33836842e-01 1.08032787e+00 2.70020276e-01
3.52644145e-01 -9.31178629e-02 4.80010808e-01 5.99276066e-01
-3.00745517e-01 1.47798210e-01 -4.03418094e-01 6.05782390e-01
-2.09712076e+00 -1.04948711e+00 3.49080205e-01 2.06484938e+00
9.53200161e-01 6.16649799e-02 5.13646543e-01 3.22073519e-01
3.95433784e-01 1.89467400e-01 -1.19444895e+00 -8.35364759e-01
3.46927159e-02 3.46303880e-01 3.92336100e-01 3.66943151e-01
-8.64728272e-01 1.07632494e+00 7.61793184e+00 8.51095855e-01
-8.56525004e-01 -1.10508271e-01 2.90680707e-01 -2.84498066e-01
-7.45664630e-03 7.44412243e-02 -8.08044136e-01 4.16398704e-01
6.82606757e-01 -6.62191361e-02 9.17971551e-01 1.22879279e+00
1.38539582e-01 -6.80135250e-01 -7.03519821e-01 9.08913910e-01
-2.71865219e-01 -1.01079547e+00 -5.32921374e-01 4.62542735e-02
8.31570446e-01 5.84407188e-02 2.25997970e-01 7.38716424e-01
9.67863321e-01 -1.20860445e+00 5.64095795e-01 3.58034581e-01
3.38363707e-01 -1.13536465e+00 6.01794302e-01 3.93367231e-01
-1.04361689e+00 -5.03881752e-01 -5.94489455e-01 -5.02528906e-01
3.50172669e-01 2.81114697e-01 -4.75916505e-01 5.20417690e-01
9.16270614e-01 8.29122841e-01 9.56524163e-03 1.33906066e+00
-7.87697136e-01 2.96759486e-01 -2.06653908e-01 -8.95002306e-01
8.33926737e-01 -3.26376677e-01 5.61984718e-01 4.97895837e-01
3.16219747e-01 -1.42338917e-01 3.34412694e-01 8.76702845e-01
5.25719702e-01 -4.11476605e-02 -5.71382523e-01 -1.21894211e-01
6.03521585e-01 1.20367312e+00 -8.20674837e-01 -1.80675223e-01
4.38826345e-02 5.64081430e-01 4.50261533e-01 2.74862498e-01
-7.88067579e-01 -5.78956664e-01 5.78924656e-01 -5.08030355e-01
2.38126919e-01 -3.08674932e-01 -1.50129259e-01 -6.90602124e-01
-2.82887340e-01 -1.09584916e+00 2.44437516e-01 -3.17598611e-01
-8.31870675e-01 5.66657484e-01 2.18727678e-01 -1.37702107e+00
-8.33551645e-01 -3.43719929e-01 -5.45418561e-01 6.61709249e-01
-1.77730870e+00 -2.61082411e-01 -1.66536681e-02 4.47201014e-01
7.05310047e-01 -5.30767560e-01 7.83689857e-01 -2.21392855e-01
-4.05725837e-01 3.85592788e-01 5.25659621e-01 -3.89254004e-01
2.10541144e-01 -1.29289043e+00 2.64738803e-04 3.46493214e-01
-5.11891581e-02 4.64039564e-01 7.86859870e-01 -7.34859407e-01
-1.39407313e+00 -6.39564395e-01 3.78173411e-01 3.02779883e-01
5.54048538e-01 -6.46176487e-02 -6.23587132e-01 1.38955593e-01
3.73295546e-01 -1.81164801e-01 3.42901051e-01 1.43913254e-02
4.39594030e-01 2.67558396e-01 -1.19893289e+00 9.00270700e-01
8.35766912e-01 5.00336736e-02 -6.75241053e-01 2.33594418e-01
1.35941163e-01 -6.14337683e-01 -5.39019227e-01 2.61257112e-01
5.06025910e-01 -1.13780463e+00 9.30525601e-01 -5.53443432e-01
1.56409919e-01 -1.03055283e-01 9.81607884e-02 -1.59629047e+00
-2.24981338e-01 -8.62705946e-01 -7.04950631e-01 4.89648044e-01
-9.46483389e-03 -6.45479560e-01 1.02594638e+00 1.95098430e-01
-3.80935892e-02 -1.39289820e+00 -1.27537882e+00 -1.34403646e+00
1.38142452e-01 4.70070466e-02 7.39552736e-01 5.15725434e-01
5.86825550e-01 -1.45669490e-01 -6.06070340e-01 -5.98552465e-01
6.22352779e-01 3.50072086e-01 6.30122721e-01 -7.87986755e-01
-2.53854603e-01 -6.29400611e-01 -9.43226889e-02 -1.18953085e+00
8.22403729e-02 -3.61012697e-01 4.81961042e-01 -1.71854877e+00
2.65727863e-02 -7.05012143e-01 -5.83443820e-01 4.95538503e-01
5.37474593e-03 -1.62894562e-01 1.37234300e-01 3.27715993e-01
-8.91378522e-01 1.31490254e+00 1.59962785e+00 5.11343144e-02
-5.37428141e-01 1.99294955e-01 -5.02163529e-01 6.52785778e-01
1.37532330e+00 -6.07211709e-01 -5.65183461e-01 2.39757121e-01
3.25284034e-01 3.52046728e-01 1.57030270e-01 -9.77851689e-01
-1.13995284e-01 -8.26120257e-01 2.25962177e-01 -8.43750477e-01
6.48766518e-01 -5.91005921e-01 -3.31766456e-01 1.17235100e+00
-4.32688862e-01 2.62867719e-01 -7.71999285e-02 6.63216770e-01
-2.30162039e-01 -7.49154806e-01 8.24041009e-01 -4.39977944e-01
-8.48509431e-01 1.51618183e-01 -8.11601043e-01 1.12181313e-01
1.55698967e+00 -4.35843587e-01 1.40371323e-01 -4.67361063e-01
-6.44127667e-01 4.61274326e-01 7.47319311e-02 4.02089775e-01
9.91852582e-01 -1.48359203e+00 -3.01415265e-01 -1.19795658e-01
-2.64705807e-01 -9.96202156e-02 1.74143195e-01 7.98705578e-01
-7.68218189e-02 4.44497287e-01 -7.98372447e-01 -3.22444975e-01
-8.89764130e-01 6.36861920e-01 1.83284000e-01 -5.31989336e-01
-5.96250296e-01 8.03618252e-01 -2.63257205e-01 -3.71696889e-01
6.08259201e-01 -8.68263245e-02 -7.83458948e-01 -1.54301897e-01
5.95528662e-01 6.57629609e-01 -4.61738855e-01 -2.04269499e-01
-2.86970675e-01 1.97779149e-01 6.38396069e-02 -3.55421245e-01
1.19173074e+00 7.40542542e-03 1.94909677e-01 4.52246100e-01
4.24547732e-01 -6.32797301e-01 -1.54959798e+00 -1.91252287e-02
-8.75439495e-02 -5.66235304e-01 -4.49675135e-02 -8.66870344e-01
-8.10632229e-01 8.71981978e-01 7.66825438e-01 1.49014995e-01
8.95323873e-01 -2.11268693e-01 5.97860992e-01 6.97721720e-01
7.58450985e-01 -1.64662564e+00 5.69051385e-01 9.37349498e-01
9.85715330e-01 -1.19087541e+00 3.02185537e-03 3.98344994e-01
-9.65308189e-01 9.51681077e-01 8.51500690e-01 -4.77427691e-01
4.79797125e-01 1.39220044e-01 -9.23834667e-02 -8.59943256e-02
-6.42751575e-01 -3.90480280e-01 -2.38100454e-01 1.00743580e+00
1.81063280e-01 2.45161857e-02 -5.94538271e-01 2.31257543e-01
-3.08340698e-01 -1.38351962e-01 4.85586554e-01 1.32487476e+00
-6.87677145e-01 -1.40980875e+00 -4.34218138e-01 4.80328411e-01
-3.06646407e-01 1.52841255e-01 -9.63680819e-02 6.80038095e-01
-3.71120870e-01 1.03694093e+00 -2.95527339e-01 -1.75003290e-01
9.83786434e-02 -5.30794680e-01 5.95804870e-01 -4.62511659e-01
-5.72485268e-01 -7.11047510e-03 -1.40961111e-01 -9.08913612e-01
-4.71605003e-01 -8.99811268e-01 -1.58720243e+00 6.21852092e-02
-2.50712514e-01 4.89876926e-01 7.12352633e-01 8.17502081e-01
-5.04848035e-03 6.28560364e-01 7.97605634e-01 -1.08296704e+00
-1.17139971e+00 -7.29045331e-01 -7.79304266e-01 -3.60280037e-01
2.42808729e-01 -1.39210737e+00 -2.21840322e-01 -7.66638875e-01] | [3.9191527366638184, 1.8477587699890137] |
cfd9c055-370d-4654-85bf-3d564ebbd4bc | vipformer-efficient-vision-and-pointcloud | 2303.14376 | null | https://arxiv.org/abs/2303.14376v1 | https://arxiv.org/pdf/2303.14376v1.pdf | ViPFormer: Efficient Vision-and-Pointcloud Transformer for Unsupervised Pointcloud Understanding | Recently, a growing number of work design unsupervised paradigms for point cloud processing to alleviate the limitation of expensive manual annotation and poor transferability of supervised methods. Among them, CrossPoint follows the contrastive learning framework and exploits image and point cloud data for unsupervised point cloud understanding. Although the promising performance is presented, the unbalanced architecture makes it unnecessarily complex and inefficient. For example, the image branch in CrossPoint is $\sim$8.3x heavier than the point cloud branch leading to higher complexity and latency. To address this problem, in this paper, we propose a lightweight Vision-and-Pointcloud Transformer (ViPFormer) to unify image and point cloud processing in a single architecture. ViPFormer learns in an unsupervised manner by optimizing intra-modal and cross-modal contrastive objectives. Then the pretrained model is transferred to various downstream tasks, including 3D shape classification and semantic segmentation. Experiments on different datasets show ViPFormer surpasses previous state-of-the-art unsupervised methods with higher accuracy, lower model complexity and runtime latency. Finally, the effectiveness of each component in ViPFormer is validated by extensive ablation studies. The implementation of the proposed method is available at https://github.com/auniquesun/ViPFormer. | ['Deying Li', 'Xuewei Bai', 'Xudong Cai', 'Yongcai Wang', 'Hongyu Sun'] | 2023-03-25 | null | null | null | null | ['3d-shape-retrieval'] | ['computer-vision'] | [ 7.48933703e-02 -9.67421457e-02 -3.18165123e-02 -3.65747124e-01
-7.48105228e-01 -6.57976329e-01 4.05180216e-01 2.02363372e-01
-2.54469573e-01 5.78018576e-02 -5.79208076e-01 -1.03878617e-01
-2.26762630e-02 -6.93712592e-01 -7.90274262e-01 -7.50840604e-01
2.51888514e-01 6.01603746e-01 5.30385315e-01 1.20126083e-01
4.19798166e-01 7.63513267e-01 -1.69599831e+00 2.70531267e-01
1.06753647e+00 1.15583801e+00 5.95480978e-01 2.62783051e-01
-5.90790868e-01 1.78928569e-01 -2.61055440e-01 -3.51511359e-01
3.80278826e-01 1.20350450e-01 -6.82402253e-01 2.24031568e-01
6.55147314e-01 -1.09207883e-01 5.66531420e-02 1.11673129e+00
4.13400471e-01 -9.94501822e-03 5.40830255e-01 -1.57882226e+00
-3.53313029e-01 2.05697045e-01 -9.66516674e-01 5.26490547e-02
-2.89915740e-01 1.91579089e-01 8.83162081e-01 -1.26063979e+00
3.65178198e-01 8.98653388e-01 7.52883136e-01 2.55963504e-01
-1.07875395e+00 -8.50253582e-01 3.79997104e-01 5.36452942e-02
-1.44182062e+00 -3.60763669e-01 9.90870953e-01 -3.97180587e-01
1.06331432e+00 2.68063843e-01 6.35102212e-01 7.10076690e-01
-1.97353855e-01 8.34134579e-01 9.34631586e-01 -3.24570715e-01
2.86283910e-01 3.20548266e-02 1.44345202e-02 8.10692370e-01
9.44915041e-02 9.75833982e-02 -5.29664457e-01 -7.32163491e-04
8.20165098e-01 2.84384966e-01 1.12836240e-02 -6.43004775e-01
-1.09707034e+00 5.54636419e-01 7.62637854e-01 2.79845536e-01
-2.74804622e-01 2.12996155e-01 2.75141925e-01 -4.27763583e-03
6.54662848e-01 1.63064867e-01 -5.17516017e-01 3.48272212e-02
-1.20005703e+00 -9.38876644e-02 3.21799606e-01 1.23458838e+00
9.52776670e-01 2.51869801e-02 3.49062383e-01 8.45656097e-01
6.59717739e-01 6.46731973e-01 3.91972251e-02 -9.51094627e-01
5.12215436e-01 9.52466249e-01 -3.06799799e-01 -1.10528111e+00
-4.50179100e-01 -6.28086925e-01 -8.11839938e-01 3.70146722e-01
2.26116791e-01 3.16456556e-01 -1.03219318e+00 1.11266351e+00
5.07581055e-01 4.05087233e-01 -2.45906144e-01 1.13982058e+00
1.01519203e+00 7.10427642e-01 2.72998035e-01 9.16182622e-02
1.30460739e+00 -1.24679804e+00 -1.88943341e-01 -2.76341707e-01
4.29323435e-01 -9.31455255e-01 1.15999901e+00 3.85364711e-01
-1.16823900e+00 -5.68017125e-01 -8.92259300e-01 -2.28428721e-01
-3.84417504e-01 3.77465487e-01 8.28148842e-01 3.56724203e-01
-9.91766095e-01 3.04786444e-01 -1.05807602e+00 -3.55807006e-01
7.87565827e-01 4.15096700e-01 -1.54871151e-01 3.42878960e-02
-5.00709176e-01 3.60597759e-01 2.84424216e-01 9.39022899e-02
-6.58561707e-01 -1.08331013e+00 -6.92326903e-01 -3.15256901e-02
2.54900128e-01 -7.47095466e-01 1.23294127e+00 -9.38129842e-01
-1.46261239e+00 1.04234719e+00 5.11519350e-02 -1.48475289e-01
3.29335362e-01 -1.85702294e-01 -5.68231428e-03 4.18277830e-01
2.06464380e-01 9.82694030e-01 9.44594026e-01 -1.48344409e+00
-8.10516059e-01 -5.90106845e-01 -8.32435414e-02 3.52973282e-01
-1.90658212e-01 -2.06691161e-01 -1.04836071e+00 -4.87139344e-01
5.69344282e-01 -9.25262988e-01 -2.04543695e-01 3.25360417e-01
-2.63039440e-01 -2.27354258e-01 9.75443304e-01 -1.32364720e-01
7.40137339e-01 -2.27813840e+00 8.81549437e-03 2.04070359e-01
3.66544604e-01 1.91725880e-01 -7.31611103e-02 2.73187637e-01
-6.17463030e-02 -1.01967931e-01 -5.56085706e-01 -5.76443672e-01
-7.59894624e-02 2.03726083e-01 -3.50669771e-01 5.50140262e-01
2.20625371e-01 9.34771597e-01 -7.87356436e-01 -7.03371644e-01
5.59263110e-01 4.94975239e-01 -6.48436189e-01 -7.62851015e-02
-2.69908100e-01 5.48139930e-01 -4.79195744e-01 1.17393994e+00
1.01471376e+00 -5.33992589e-01 -2.04890549e-01 -5.68503082e-01
-3.64443451e-01 -1.41279116e-01 -9.04266596e-01 2.18622851e+00
-4.38114852e-01 3.47867996e-01 7.31902868e-02 -1.09502578e+00
9.10719931e-01 -1.07139740e-02 7.50019014e-01 -5.66350698e-01
1.28527239e-01 3.71205121e-01 -2.56507427e-01 -4.42834586e-01
2.45229170e-01 -2.48596929e-02 1.64905235e-01 1.45454019e-01
1.31186262e-01 -5.42102396e-01 -2.09381506e-01 1.42107487e-01
7.06049204e-01 3.90614718e-01 -1.27090901e-01 -2.62851417e-01
5.35950184e-01 3.28519285e-01 4.31825221e-01 4.07367378e-01
-1.99468717e-01 8.35854053e-01 -3.78357656e-02 -2.23883718e-01
-7.48665512e-01 -1.36245334e+00 -1.57384768e-01 9.05288398e-01
6.63642108e-01 -3.84151638e-01 -6.38768196e-01 -7.00588226e-01
4.38317060e-02 5.34961879e-01 -1.07937045e-01 1.36960402e-01
-4.72919315e-01 -4.54226375e-01 3.98212940e-01 6.97199762e-01
6.61529779e-01 -8.50677550e-01 -6.30016267e-01 -1.55329809e-01
-1.99275017e-01 -1.35302901e+00 -2.48267263e-01 -7.82099664e-02
-1.26752222e+00 -9.38408375e-01 -3.99507880e-01 -9.40075815e-01
1.01068139e+00 6.11357093e-01 1.13685107e+00 1.56256050e-01
-3.15171331e-01 6.19249165e-01 -3.44609439e-01 -6.82648957e-01
2.94079185e-01 2.48321578e-01 -6.75531924e-02 -1.23860734e-02
3.57908815e-01 -8.30833018e-01 -8.84722233e-01 2.32184589e-01
-9.06837344e-01 1.65124327e-01 7.06407964e-01 6.21821284e-01
1.13555098e+00 6.37847632e-02 1.26329750e-01 -7.32648373e-01
-1.31413778e-02 -3.03271711e-01 -8.98216426e-01 -4.19384576e-02
-5.50724387e-01 -3.25853586e-01 3.63006651e-01 -9.18590203e-02
-1.00196242e+00 3.53029996e-01 -1.20896347e-01 -8.89534533e-01
-2.60044038e-01 2.71191955e-01 -2.17084274e-01 -2.35335261e-01
3.24149817e-01 3.08206648e-01 5.95125929e-02 -5.85462749e-01
4.78881955e-01 4.62765753e-01 5.55692792e-01 -5.42433083e-01
1.13753211e+00 1.03008854e+00 -1.43850818e-01 -9.32972252e-01
-6.30056620e-01 -6.81735098e-01 -8.13038588e-01 -4.84458447e-01
9.40669298e-01 -8.81169140e-01 -6.84856892e-01 4.12777543e-01
-1.27986777e+00 -1.69067100e-01 -2.70334870e-01 3.66228789e-01
-5.82137167e-01 3.53118300e-01 -2.75068641e-01 -6.78072453e-01
-5.45248628e-01 -1.16837847e+00 1.49777520e+00 2.30685607e-01
1.06587902e-01 -8.95218790e-01 -1.92235157e-01 6.29458666e-01
1.20555721e-01 2.58750468e-02 7.47808814e-01 -3.06640238e-01
-9.96780336e-01 3.61729190e-02 -5.96358538e-01 2.19716161e-01
-1.01484936e-02 7.17025623e-02 -1.10171568e+00 -1.97790742e-01
5.81427217e-02 -1.21045791e-01 6.63803160e-01 3.60631168e-01
1.50870657e+00 1.22993656e-01 -5.13109863e-01 9.79745924e-01
1.59629178e+00 2.11659279e-02 4.71957892e-01 3.09060514e-01
1.00593877e+00 6.74449146e-01 8.01432312e-01 1.90945908e-01
4.79485750e-01 6.47988439e-01 7.26075530e-01 -3.48852426e-01
-1.11259237e-01 -1.76793292e-01 1.38803840e-01 1.19422722e+00
-1.81741163e-01 1.81324258e-01 -1.33239281e+00 5.30491590e-01
-1.82429612e+00 -7.03526556e-01 -3.17222387e-01 1.75839984e+00
4.91896421e-01 2.93748286e-02 -8.88231397e-02 1.54704303e-02
4.30110186e-01 -4.01755683e-02 -5.85087538e-01 -7.48527497e-02
7.12655708e-02 2.78078020e-01 4.40477729e-01 2.64704853e-01
-1.20911515e+00 1.22394705e+00 5.10218000e+00 9.06270444e-01
-1.42074192e+00 3.34301502e-01 5.47219217e-01 -1.56774789e-01
-1.82197675e-01 6.96433634e-02 -5.15942574e-01 4.14119780e-01
2.87254959e-01 2.73493588e-01 1.80104509e-01 9.36442614e-01
5.56611605e-02 -8.23524129e-03 -9.36984181e-01 1.57473016e+00
9.39940512e-02 -1.36229384e+00 -6.64864713e-03 -5.88544123e-02
5.94778836e-01 5.27402937e-01 8.50940347e-02 7.04755709e-02
-2.43171692e-01 -8.08455408e-01 8.12189341e-01 4.97267157e-01
7.11735308e-01 -6.06586874e-01 4.86641169e-01 4.01169866e-01
-1.31129897e+00 7.63394982e-02 -2.94773787e-01 1.67191058e-01
-5.58629818e-02 7.13049710e-01 -4.31406975e-01 7.53544331e-01
1.08378160e+00 7.36839056e-01 -5.42492807e-01 1.15907586e+00
-4.46182825e-02 3.35257322e-01 -5.24201989e-01 3.25488985e-01
4.66213495e-01 -4.15748447e-01 6.32501721e-01 1.14486992e+00
3.73479158e-01 2.54295260e-01 3.60196114e-01 9.86264288e-01
1.23776637e-01 1.21122502e-01 -5.61713219e-01 1.59926891e-01
5.77590466e-01 1.57771552e+00 -1.10779285e+00 -1.75929740e-01
-4.79816020e-01 9.31960166e-01 2.93586552e-01 2.86875933e-01
-8.70988309e-01 -1.37182742e-01 4.43210423e-01 2.07149073e-01
4.15852398e-01 -3.75450879e-01 -7.68852592e-01 -1.13699698e+00
1.59493208e-01 -5.10604203e-01 2.41583601e-01 -8.00378084e-01
-1.37617528e+00 5.94299793e-01 1.15598060e-01 -1.62836683e+00
3.49674046e-01 -5.12989700e-01 -5.72368920e-01 6.33343875e-01
-1.77572525e+00 -1.43697751e+00 -6.39704704e-01 6.03216290e-01
8.50482047e-01 -1.99252054e-01 5.09947479e-01 4.49385375e-01
-5.19612849e-01 4.31945443e-01 -2.43810803e-01 -9.70058441e-02
4.82058764e-01 -1.13394439e+00 2.53307074e-01 7.84799278e-01
2.39823967e-01 3.91728908e-01 1.56832486e-01 -5.23284376e-01
-1.57007194e+00 -1.25206149e+00 5.07264256e-01 -5.30211568e-01
5.42475462e-01 -4.23309565e-01 -8.49161327e-01 2.93039352e-01
3.74514461e-02 3.17815036e-01 5.35771251e-01 -1.96948811e-01
-3.06749880e-01 -4.12331551e-01 -1.06023932e+00 6.36799395e-01
1.27387595e+00 -4.27147299e-01 -4.04166549e-01 2.83712506e-01
5.42848945e-01 -3.84317309e-01 -7.96248496e-01 5.39719999e-01
2.13181406e-01 -9.43502426e-01 1.10864699e+00 5.81148118e-02
4.43012983e-01 -6.40095532e-01 -9.54812318e-02 -9.11706686e-01
-2.57800609e-01 -3.86837691e-01 -1.60921857e-01 1.23142731e+00
2.37994984e-01 -5.56262553e-01 1.08725727e+00 3.56189728e-01
-5.14422894e-01 -9.52039182e-01 -9.92419064e-01 -7.00398505e-01
-8.02883729e-02 -7.18201220e-01 5.43204308e-01 1.01821744e+00
-3.59875053e-01 3.59388232e-01 2.66318381e-01 4.99218106e-01
9.56012309e-01 4.79114234e-01 7.37676203e-01 -1.23467708e+00
-1.72061678e-02 -5.69507301e-01 -4.61896837e-01 -1.20798349e+00
1.31784435e-02 -1.26589096e+00 3.20594572e-02 -1.54204071e+00
8.08365177e-03 -9.32713747e-01 -1.85777485e-01 6.11446083e-01
3.16083841e-02 5.86490035e-01 4.55382228e-01 5.92589080e-01
-6.43440127e-01 6.02128685e-01 1.24383962e+00 -4.47349101e-01
-2.52276897e-01 -9.18923765e-02 -5.11035681e-01 1.02184415e+00
1.00426042e+00 -3.68568242e-01 -7.39556551e-01 -9.12398994e-01
-4.78657708e-02 -2.97527850e-01 5.16995132e-01 -1.04597402e+00
6.70682907e-01 5.02294116e-02 2.96163887e-01 -1.11262953e+00
5.04353404e-01 -1.22245884e+00 2.42382530e-02 2.06353843e-01
1.99176863e-01 1.04044586e-01 4.18710649e-01 5.30131996e-01
-2.69964278e-01 -5.89937642e-02 6.80126488e-01 -7.20024556e-02
-8.87281835e-01 5.87194383e-01 1.86949536e-01 -2.51518816e-01
1.14204097e+00 -5.76291144e-01 -8.97751302e-02 1.94723874e-01
-5.61068058e-01 3.06173533e-01 4.38350916e-01 4.81363803e-01
9.96478736e-01 -1.16623199e+00 -5.11276960e-01 3.76089394e-01
2.72160828e-01 5.56837499e-01 4.97857690e-01 1.25024629e+00
-7.92998910e-01 3.10022950e-01 -1.73097979e-02 -1.31844723e+00
-1.21814871e+00 3.75743777e-01 3.14439237e-01 9.41230953e-02
-9.02246416e-01 9.34641123e-01 6.34209275e-01 -6.99007154e-01
2.41430014e-01 -2.59121805e-01 -2.31078709e-03 -7.65390322e-02
1.28265113e-01 2.68389612e-01 3.07091743e-01 -6.65484250e-01
-6.40428364e-01 1.06404531e+00 -1.57285511e-01 7.66480267e-02
1.35405838e+00 -7.93841779e-02 -2.69064099e-01 3.22964162e-01
1.09801221e+00 -1.88028976e-01 -1.33815980e+00 -1.03529841e-01
2.74513848e-03 -5.11093855e-01 2.74237126e-01 -7.11205244e-01
-1.38978350e+00 9.05979753e-01 8.16861093e-01 -3.09122801e-02
1.27010691e+00 1.94635600e-01 8.48182380e-01 1.12586349e-01
5.75098574e-01 -1.04482782e+00 -1.93827394e-02 3.88747633e-01
7.68200457e-01 -1.43301952e+00 2.01132968e-01 -9.40120876e-01
-5.30919313e-01 9.58954930e-01 8.44748676e-01 -2.08657682e-01
8.62748802e-01 2.42685765e-01 2.07427740e-01 -6.31055534e-01
-4.91462678e-01 -1.82435632e-01 3.01367968e-01 7.00215757e-01
1.51029006e-01 -3.75985093e-02 -3.31727639e-02 4.15354282e-01
-1.78324372e-01 -1.80843011e-01 -1.34699926e-01 8.93866420e-01
-1.47025228e-01 -9.75979686e-01 -3.62707466e-01 2.81410307e-01
-5.32719120e-02 -7.70433247e-02 -3.67654830e-01 7.20858872e-01
3.57161820e-01 7.55183220e-01 4.54439342e-01 -3.85708153e-01
3.64241809e-01 -2.95761883e-01 5.17009079e-01 -6.53357029e-01
-6.62911355e-01 3.30587000e-01 -3.56079698e-01 -7.56852746e-01
-6.42721415e-01 -5.32361329e-01 -1.56775689e+00 2.29017474e-02
-3.13458949e-01 -1.42167611e-02 8.46635044e-01 7.83912659e-01
6.77085876e-01 4.41988438e-01 5.85791945e-01 -1.14132559e+00
-8.68257880e-02 -5.07978261e-01 -2.18142107e-01 2.40218952e-01
-6.14888705e-02 -7.78658926e-01 -1.58360839e-01 1.34950295e-01] | [8.026222229003906, -3.171055555343628] |
3120d4cb-1719-44fe-bc8e-27d2eaf06fe5 | flexible-dataset-distillation-learn-labels | 2006.08572 | null | https://arxiv.org/abs/2006.08572v3 | https://arxiv.org/pdf/2006.08572v3.pdf | Flexible Dataset Distillation: Learn Labels Instead of Images | We study the problem of dataset distillation - creating a small set of synthetic examples capable of training a good model. In particular, we study the problem of label distillation - creating synthetic labels for a small set of real images, and show it to be more effective than the prior image-based approach to dataset distillation. Methodologically, we introduce a more robust and flexible meta-learning algorithm for distillation, as well as an effective first-order strategy based on convex optimization layers. Distilling labels with our new algorithm leads to improved results over prior image-based distillation. More importantly, it leads to clear improvements in flexibility of the distilled dataset in terms of compatibility with off-the-shelf optimizers and diverse neural architectures. Interestingly, label distillation can also be applied across datasets, for example enabling learning Japanese character recognition by training only on synthetically labeled English letters. | ['Timothy Hospedales', 'Yongxin Yang', 'Ondrej Bohdal'] | 2020-06-15 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 6.17818117e-01 4.72019762e-01 1.25668198e-01 -5.91428936e-01
-9.39102113e-01 -7.87521124e-01 8.85685384e-01 -2.35797331e-01
-6.77452683e-01 1.08096433e+00 5.34857661e-02 -1.19005196e-01
1.82791144e-01 -5.76556802e-01 -1.02211845e+00 -6.54769838e-01
2.80603319e-01 1.06718159e+00 -4.16556865e-01 -7.89913833e-02
1.59772351e-01 4.38069701e-01 -1.45887256e+00 2.79393494e-01
8.42396200e-01 7.02323794e-01 2.49253839e-01 6.23850822e-01
3.75096081e-03 7.74573565e-01 -8.10841978e-01 -5.09587407e-01
4.63299721e-01 -3.66105288e-01 -1.02975202e+00 3.42115015e-01
8.83020461e-01 -2.75636345e-01 -1.14932723e-01 8.67058575e-01
4.77771610e-01 9.51247960e-02 6.89764261e-01 -1.04898298e+00
-9.52732563e-01 8.47760677e-01 -2.19905600e-01 -3.64710957e-01
1.28076732e-01 4.22586918e-01 8.66962850e-01 -1.01782179e+00
9.82473671e-01 1.20492148e+00 6.29722655e-01 9.19633687e-01
-1.91641164e+00 -5.49734175e-01 1.76369250e-02 -1.86183244e-01
-1.03548992e+00 -3.65832925e-01 6.55663431e-01 -6.15605354e-01
6.49299920e-01 3.36768657e-01 5.94839990e-01 1.38137984e+00
-5.73870897e-01 9.97506320e-01 1.60862076e+00 -7.25520968e-01
1.94343790e-01 5.47022223e-01 4.84571047e-03 6.08089864e-01
2.93689340e-01 1.30958989e-01 -1.70210212e-01 1.14100330e-01
6.80183113e-01 -2.16411799e-01 -1.77292615e-01 -6.69274271e-01
-1.84073102e+00 6.92406595e-01 4.69977379e-01 3.37980352e-02
-2.06034724e-02 4.14809942e-01 5.62947631e-01 4.74333078e-01
4.72800314e-01 1.38521564e+00 -5.12439966e-01 1.30204335e-01
-1.16519260e+00 3.45615417e-01 8.55190516e-01 1.26304305e+00
9.63529527e-01 3.47661942e-01 -3.07540059e-01 9.17595804e-01
8.28355737e-03 4.00530964e-01 4.42943931e-01 -1.35277545e+00
4.95627284e-01 2.51408875e-01 1.98747382e-01 -3.42769772e-01
-3.27562273e-01 -3.44464123e-01 -6.32835925e-01 3.94591331e-01
7.31181145e-01 -2.66410053e-01 -1.21758282e+00 2.05183506e+00
3.59577164e-02 9.72392038e-02 1.29877329e-01 5.15693069e-01
6.43474817e-01 3.76997650e-01 8.91910959e-03 1.07412785e-01
1.12338173e+00 -1.37404788e+00 -4.07700330e-01 -3.55592221e-01
9.89079654e-01 -5.29840112e-01 1.43663001e+00 4.80544150e-01
-1.05021346e+00 -7.24863112e-01 -1.04262888e+00 -2.39066586e-01
-6.36065841e-01 3.36564243e-01 8.05354953e-01 6.99591875e-01
-1.21751010e+00 6.96358204e-01 -4.55341071e-01 -2.78860420e-01
5.57107627e-01 5.08356929e-01 -5.12590170e-01 -2.85357118e-01
-8.94441783e-01 9.97603297e-01 8.25402498e-01 -1.08035967e-01
-1.26842368e+00 -8.36232960e-01 -9.49190617e-01 -3.04133803e-01
3.45244765e-01 -7.53793657e-01 1.40953696e+00 -1.13150012e+00
-1.63266635e+00 1.58190501e+00 2.79887974e-01 -3.59398186e-01
7.29226828e-01 -2.55378187e-02 -9.67059061e-02 -3.72594185e-02
2.49689147e-02 1.40754163e+00 9.10706222e-01 -1.52205014e+00
-2.22885072e-01 8.27877074e-02 1.51622653e-01 1.07354306e-01
-1.97086126e-01 -3.87031257e-01 -2.20148280e-01 -8.83704126e-01
-2.41725221e-01 -1.27583218e+00 -2.84953177e-01 -5.28244709e-04
-6.17345989e-01 1.52047321e-01 4.91667062e-01 -6.02048457e-01
6.50374949e-01 -2.05650854e+00 3.10525358e-01 -8.81399214e-02
3.03534836e-01 4.52941149e-01 -4.57737207e-01 1.12574153e-01
-3.60244840e-01 2.11431950e-01 -5.06255746e-01 -6.55827343e-01
1.70369178e-01 4.67093885e-01 -3.26468825e-01 2.22656041e-01
4.67958778e-01 1.21459305e+00 -1.07253695e+00 -2.44571358e-01
2.36371860e-01 1.78002924e-01 -7.25142598e-01 1.81861088e-01
-6.60930634e-01 7.47712731e-01 -4.34094705e-02 5.21595418e-01
6.28475368e-01 -3.40465367e-01 4.51090522e-02 -4.50850427e-02
1.83090836e-01 2.18334571e-01 -1.12985873e+00 2.15380669e+00
-6.07478797e-01 9.38515902e-01 -1.20290577e-01 -1.07952487e+00
9.24114823e-01 1.18537657e-01 2.13822305e-01 -4.48852301e-01
-6.48809746e-02 5.95066369e-01 -2.22715423e-01 -4.53944772e-01
7.04348505e-01 -3.38797569e-01 -1.97637528e-01 6.86962903e-01
3.49408299e-01 -7.50327945e-01 4.03855950e-01 1.41371816e-01
6.43815875e-01 5.02866805e-01 -7.34252110e-02 -3.86209756e-01
3.81572545e-01 2.35126883e-01 5.53787500e-02 8.20849359e-01
1.14318497e-01 8.65795910e-01 2.26866081e-01 -4.59909797e-01
-1.59773409e+00 -8.36353540e-01 -3.35014045e-01 1.07531989e+00
-3.00679803e-01 -4.64757346e-02 -1.01761496e+00 -7.49547839e-01
1.38813153e-01 9.64995623e-01 -8.69443297e-01 -8.81151259e-02
-7.51636147e-01 -8.88168097e-01 7.87496924e-01 5.46939850e-01
5.79376698e-01 -1.18299186e+00 -3.50713909e-01 6.25370815e-02
-6.95124343e-02 -1.02154505e+00 -3.96598727e-01 7.17933834e-01
-9.43776488e-01 -6.72523618e-01 -1.10505867e+00 -9.75808859e-01
8.64601433e-01 -4.32241224e-02 1.50296628e+00 -5.75804114e-02
-2.60149956e-01 1.93854868e-01 -4.95754741e-02 -2.43277952e-01
-8.07287097e-01 2.85762459e-01 -2.23483682e-01 -3.14657480e-01
9.13916305e-02 -3.26358557e-01 -2.98792213e-01 6.13455847e-02
-9.34961498e-01 2.74620563e-01 4.69586104e-01 1.14791369e+00
5.17059147e-01 -5.49422622e-01 4.89767671e-01 -1.46142447e+00
6.59128428e-01 -2.28504017e-01 -4.70451802e-01 3.84038657e-01
-8.13458741e-01 5.96326768e-01 6.79134786e-01 -6.57398462e-01
-9.31467891e-01 5.58274724e-02 8.64098594e-02 -1.78505391e-01
-2.59961545e-01 1.63517445e-01 -4.26028706e-02 -1.50992498e-01
9.73924458e-01 1.28130630e-01 5.30587696e-02 -4.86965120e-01
1.17812645e+00 5.13899863e-01 6.95006490e-01 -1.00966287e+00
6.05195582e-01 4.78993833e-01 -2.46885773e-02 -4.42096412e-01
-8.95987213e-01 7.22310841e-02 -8.18988264e-01 1.99455470e-01
8.43549728e-01 -9.62154746e-01 -2.63453126e-01 6.79261923e-01
-9.38404739e-01 -8.42311442e-01 -9.35793996e-01 3.91408235e-01
-9.82871890e-01 8.02924559e-02 -6.91481411e-01 -6.90647736e-02
7.31029287e-02 -1.34489942e+00 1.07056475e+00 -3.57332118e-02
-2.60370106e-01 -1.21089149e+00 2.24518806e-01 3.37963849e-01
3.40931088e-01 5.30367851e-01 7.78329909e-01 -8.65672350e-01
-6.30528033e-01 -1.36769027e-01 -3.77403557e-01 6.64572656e-01
-1.93300191e-02 -4.47019888e-03 -1.38920701e+00 -5.07280469e-01
-2.24123314e-01 -1.05964041e+00 9.31045294e-01 1.31172732e-01
1.01885915e+00 -2.24759817e-01 -3.15503329e-01 8.78000975e-01
1.37840188e+00 -3.96779738e-02 5.52699387e-01 4.90976483e-01
9.70233083e-01 5.64282656e-01 3.26007068e-01 1.37465477e-01
2.10301265e-01 5.95985591e-01 2.07082942e-01 -3.72283608e-01
-4.79383886e-01 -1.84348539e-01 -2.66322065e-02 7.33662426e-01
1.95381492e-01 -1.75590679e-01 -7.99755275e-01 4.18618858e-01
-1.59728396e+00 -7.37410426e-01 -5.55708073e-03 2.16585541e+00
1.25443077e+00 4.13426049e-02 6.87506646e-02 -9.81538072e-02
5.59678614e-01 6.82470798e-02 -6.71337247e-01 -4.94803607e-01
-3.29785615e-01 3.03864628e-01 5.05955398e-01 5.28386652e-01
-1.09292805e+00 1.09217119e+00 7.66920805e+00 7.40865111e-01
-1.08828235e+00 1.34501800e-01 8.54473233e-01 -1.59501612e-01
-6.33347988e-01 4.77077588e-02 -7.27071345e-01 4.37808543e-01
8.75855625e-01 1.26404464e-01 7.65386283e-01 9.13958192e-01
-3.22315454e-01 8.92426912e-03 -1.67367899e+00 1.08039439e+00
2.22158045e-01 -1.51846755e+00 1.17736503e-01 5.41231520e-02
1.25202417e+00 2.71466840e-02 1.46508247e-01 5.52981138e-01
7.86910772e-01 -1.21890354e+00 1.02278054e+00 6.13530219e-01
1.23890495e+00 -4.95583355e-01 3.82652193e-01 2.65880555e-01
-4.94044363e-01 6.24412633e-02 -3.59264135e-01 4.27074246e-02
-6.32976368e-02 5.36541104e-01 -8.45797658e-01 1.40140697e-01
2.23430723e-01 6.30034864e-01 -9.09774780e-01 7.69524634e-01
-3.75432104e-01 3.66277963e-01 -3.18956852e-01 1.58728868e-01
3.31191152e-01 -2.51139075e-01 1.42608061e-01 1.46958780e+00
7.41090328e-02 -3.57634515e-01 1.23252608e-01 9.82087314e-01
-4.47732121e-01 -2.54743159e-01 -8.59637201e-01 -8.96439627e-02
4.53115970e-01 1.17740464e+00 -5.94042242e-01 -7.02740490e-01
-1.77385569e-01 1.03926849e+00 7.37743497e-01 3.67376506e-01
-6.65703654e-01 -3.79532039e-01 2.49330670e-01 -1.84288442e-01
1.48571804e-01 -2.00887397e-01 -5.36274612e-01 -1.40026426e+00
-2.01702759e-01 -1.25464380e+00 5.76514937e-03 -9.72560585e-01
-1.11574852e+00 6.44306898e-01 -4.25916119e-03 -9.78734672e-01
-3.76743376e-01 -7.94740558e-01 -2.10172459e-01 1.04500496e+00
-1.36545157e+00 -1.20890570e+00 -4.63354498e-01 2.96970934e-01
5.26887059e-01 -8.96027237e-02 1.08731198e+00 2.46135965e-01
-5.13191938e-01 7.84723699e-01 4.35794741e-01 8.81225429e-03
9.32331383e-01 -1.60934865e+00 5.96441984e-01 6.41320646e-01
3.73595119e-01 5.68688571e-01 6.75597668e-01 -3.16605598e-01
-8.43427479e-01 -1.00664151e+00 5.65993190e-01 -8.17509353e-01
5.36276519e-01 -6.28884315e-01 -7.07488000e-01 9.39240515e-01
3.65133673e-01 -1.04434080e-01 6.08032048e-01 -7.02656806e-03
-5.24698436e-01 8.03067908e-02 -1.25949967e+00 7.56385922e-01
1.05209529e+00 -6.47483408e-01 -5.02286792e-01 7.29305625e-01
7.56191134e-01 -4.31978673e-01 -7.53532231e-01 3.96355420e-01
3.39913070e-01 -6.61036909e-01 1.05212617e+00 -7.01243281e-01
6.18746758e-01 -1.14237726e-01 -2.02710226e-01 -1.52703416e+00
-3.71306658e-01 -7.32557654e-01 1.23593435e-02 1.35087252e+00
7.98410714e-01 -4.50256348e-01 9.59770977e-01 8.04593801e-01
-1.59587651e-01 -6.12945080e-01 -3.50290298e-01 -9.46090043e-01
4.22991365e-01 -1.64814442e-01 6.13291800e-01 1.24794149e+00
-2.54820377e-01 2.73592353e-01 -5.94233453e-01 -5.13778925e-01
5.42250633e-01 -2.47600898e-02 1.04619384e+00 -1.02883041e+00
-6.15452349e-01 -4.23534662e-01 3.60071659e-02 -1.20630264e+00
1.07083626e-01 -1.37589622e+00 1.13670476e-01 -1.39357913e+00
1.70143038e-01 -8.74293864e-01 -1.03469662e-01 5.50889850e-01
-1.40369788e-01 5.53165853e-01 3.58997911e-01 4.51923728e-01
-3.36892068e-01 3.31731796e-01 1.68403876e+00 -3.52097809e-01
-1.03899250e-02 -3.33534658e-01 -7.15236843e-01 5.11798263e-01
6.69982433e-01 -6.55933499e-01 -5.15383065e-01 -7.08211839e-01
2.79255182e-01 -2.65627682e-01 2.30214790e-01 -8.66842330e-01
-2.85942018e-01 -2.12332923e-02 2.41348356e-01 -1.37072772e-01
4.08784330e-01 -5.97327828e-01 7.81731457e-02 3.85600567e-01
-7.95936942e-01 -1.02381207e-01 2.63196886e-01 3.95778924e-01
-1.31850373e-02 -7.53334820e-01 9.70934570e-01 -6.38873100e-01
-7.43440211e-01 -1.02015994e-01 -9.11100656e-02 5.19375265e-01
1.01423120e+00 -4.04992759e-01 -4.47231740e-01 -8.54501054e-02
-9.11019146e-01 1.26566859e-02 7.87600636e-01 1.29730806e-01
9.70003977e-02 -1.34655917e+00 -7.91604638e-01 3.94877017e-01
1.81675881e-01 1.05585150e-01 -2.58363664e-01 4.78951782e-01
-8.11441302e-01 3.88111025e-01 -3.99111867e-01 -6.01476490e-01
-8.04396987e-01 8.83755684e-01 2.00846836e-01 -4.65997487e-01
-3.34524691e-01 1.20870018e+00 1.29412785e-01 -9.85766411e-01
2.49923557e-01 -6.22728586e-01 2.27089718e-01 3.17551419e-02
1.85836539e-01 2.92896301e-01 -3.87678184e-02 -2.27783322e-01
-1.04779065e-01 5.48544824e-01 -2.57245958e-01 -3.14303786e-01
1.24899530e+00 1.08636536e-01 9.06553715e-02 5.33339560e-01
1.20477009e+00 -1.86100125e-01 -1.74386370e+00 -3.42749685e-01
1.09486803e-01 -3.25169057e-01 -3.46281528e-01 -1.04727268e+00
-8.74052048e-01 8.85011435e-01 3.49740177e-01 8.21445957e-02
8.82045269e-01 -1.25721782e-01 5.03057301e-01 8.53152871e-01
3.55313510e-01 -1.15731943e+00 4.81989175e-01 3.85897487e-01
7.03385234e-01 -1.53455460e+00 5.88428490e-02 -1.73221961e-01
-7.63492703e-01 9.38151777e-01 4.53540027e-01 -2.52889246e-01
3.35871547e-01 2.09258541e-01 1.19789101e-01 -1.40306950e-01
-4.93008107e-01 -6.60269894e-03 7.88463801e-02 8.09453011e-01
3.96733314e-01 1.47603706e-01 -3.90818641e-02 -1.98686749e-01
-3.17652076e-01 4.44364920e-02 6.17721140e-01 8.43771756e-01
-1.90823659e-01 -1.32662630e+00 -1.47659883e-01 3.31355006e-01
-2.06506699e-02 -3.13796639e-01 -3.82964343e-01 1.01150882e+00
4.12566394e-01 5.28731763e-01 -3.12819108e-02 -2.89723843e-01
9.78410468e-02 4.54263717e-01 7.97255397e-01 -9.73929882e-01
-5.98382294e-01 -2.03547046e-01 2.57179171e-01 -2.18360126e-02
-5.07639825e-01 -6.79293752e-01 -9.12243307e-01 -1.26307890e-01
-4.58361059e-01 1.65983476e-02 7.60007381e-01 8.69928122e-01
2.99783498e-01 4.91475701e-01 5.20318806e-01 -1.06057787e+00
-7.47899234e-01 -1.08531678e+00 -5.44681728e-01 7.79789746e-01
2.37680420e-01 -4.69266653e-01 -2.18463212e-01 3.98823142e-01] | [10.946633338928223, 0.5435497164726257] |
ab45ccf9-2761-44c0-b6ee-2feef24b1788 | age-and-gender-classification-with-small | null | null | https://koreascience.kr/article/JAKO202209464471634.page | https://koreascience.kr/article/JAKO202209464471634.pdf | Age and Gender Classification with Small Scale CNN | Artificial intelligence is getting a crucial part of our lives with its incredible benefits. Machines outperform humans
in recognizing objects in images, particularly in classifying people into correct age and gender groups. In this respect,
age and gender classification has been one of the hot topics among computer vision researchers in recent decades.
Deployment of deep Convolutional Neural Network(: CNN) models achieved state-of-the-art performance. However, the
most of CNN based architectures are very complex with several dozens of training parameters so they require much
computation time and resources. For this reason, we propose a new CNN-based classification algorithm with
significantly fewer training parameters and training time compared to the existing methods. Despite its less complexity,
our model shows better accuracy of age and gender classification on the UTKFace dataset | ['J. H. Yoo', 'H. Yoo', 'U. Jamoliddin'] | 2022-02-28 | null | null | null | journal-of-the-kiecs-2022-2 | ['age-and-gender-classification'] | ['computer-vision'] | [-4.07548547e-01 2.14857538e-03 -2.12329209e-01 -6.78830385e-01
3.63448650e-01 -9.00982991e-02 6.01339579e-01 2.34493405e-01
-7.40855217e-01 6.59114003e-01 -1.49008140e-01 -1.74535617e-01
-6.69907108e-02 -1.07857144e+00 -9.71458703e-02 -6.43746436e-01
-3.26143252e-03 6.41456127e-01 -2.50065863e-01 -1.38663203e-01
3.13919365e-01 2.28841662e-01 -1.95311606e+00 -2.83223420e-01
1.08845246e+00 1.34284854e+00 -4.36389983e-01 5.00896037e-01
-1.76493704e-01 4.34640676e-01 -4.76509631e-01 -7.03617871e-01
1.11727305e-01 -1.43280253e-01 -6.30970299e-01 -3.71450216e-01
6.88917637e-01 -3.60662520e-01 -3.93153757e-01 9.69936132e-01
7.26029694e-01 -1.53196260e-01 8.05162966e-01 -1.44889963e+00
-7.96383381e-01 4.10591662e-01 -6.91504717e-01 2.65626311e-01
-5.10384515e-02 -3.42078298e-01 4.69165981e-01 -6.59567237e-01
-2.16103613e-01 1.33309352e+00 8.91343296e-01 8.75960946e-01
-6.16188169e-01 -1.15122759e+00 -8.33667964e-02 4.40776527e-01
-1.33890605e+00 -1.73968121e-01 4.65645701e-01 -3.08378249e-01
6.84595823e-01 2.45108202e-01 7.97922909e-01 9.65074539e-01
1.83284149e-01 4.91051972e-01 1.11841440e+00 -2.71059364e-01
-7.67736584e-02 1.77558079e-01 3.58084798e-01 8.42180014e-01
5.58858752e-01 -1.96452320e-01 -3.76628935e-01 1.67981178e-01
8.02582562e-01 4.37429100e-01 4.47791040e-01 2.01441124e-01
-9.71937954e-01 8.45837772e-01 4.55318898e-01 3.80411893e-01
-1.40405044e-01 1.53119132e-01 4.16309983e-01 1.34457961e-01
6.71012104e-01 3.51571918e-01 -5.06691754e-01 -2.77795345e-01
-8.40875924e-01 4.88598496e-01 5.14372051e-01 5.28647900e-01
4.37143505e-01 1.64892465e-01 3.93213816e-02 9.13671970e-01
-7.50220492e-02 5.52961588e-01 5.63592136e-01 -5.96470535e-01
2.11841047e-01 1.12475836e+00 -2.53938407e-01 -1.30096281e+00
-7.51868129e-01 -8.01125407e-01 -1.55778635e+00 1.06617905e-01
7.17283428e-01 -6.12062998e-02 -1.12597883e+00 1.52009964e+00
1.62917331e-01 -1.32126048e-01 -3.03445637e-01 7.18403220e-01
1.05141115e+00 4.54649627e-01 3.96789879e-01 2.15064570e-01
1.68502653e+00 -6.55383766e-01 -5.58285475e-01 -5.17401636e-01
2.67971337e-01 -4.89989370e-01 4.84854490e-01 4.99976784e-01
-1.00178099e+00 -9.26939547e-01 -9.15077686e-01 -1.20007694e-01
-7.01531827e-01 5.00662506e-01 1.42174697e+00 1.09839976e+00
-9.20629203e-01 6.72130525e-01 -5.41150808e-01 -7.06848800e-01
7.30355382e-01 8.22889388e-01 -3.79253685e-01 9.83063132e-04
-1.05595481e+00 6.41596198e-01 3.66411030e-01 1.48554131e-01
-2.48379290e-01 -6.93339407e-01 -7.90136814e-01 -1.32148322e-02
-1.71732500e-01 -7.31360793e-01 1.06663835e+00 -9.85766828e-01
-7.83079505e-01 1.15139902e+00 -1.53301060e-01 -4.71558392e-01
5.49109995e-01 -3.24280351e-01 -5.54437935e-01 -2.32393429e-01
-1.77765667e-01 9.05377924e-01 8.27083945e-01 -5.82595289e-01
-7.79662609e-01 -9.26393747e-01 -9.83981937e-02 4.19173017e-02
-8.56191099e-01 1.67018428e-01 -1.50116593e-01 -4.04263765e-01
3.50911885e-01 -6.94018841e-01 -1.73229992e-01 -2.24458706e-02
2.50775535e-02 -7.73754239e-01 9.05618012e-01 -8.17736089e-01
1.41247880e+00 -1.87485123e+00 -3.18496600e-02 -7.23867118e-02
6.62167251e-01 4.64309990e-01 5.39310873e-01 -7.66075775e-02
2.63275802e-02 1.74571857e-01 3.38920504e-01 -4.26442683e-01
-1.92264080e-01 7.34083951e-02 3.88999611e-01 2.72068948e-01
-3.88933579e-03 7.79102683e-01 -6.19881332e-01 -6.85750425e-01
3.01855892e-01 5.49789190e-01 -4.26526755e-01 9.48155224e-02
2.96974987e-01 5.03783584e-01 -4.42765623e-01 1.02002537e+00
8.66225123e-01 7.46213645e-03 -9.51729417e-02 -1.07178964e-01
-1.34146914e-01 -3.99765283e-01 -7.40298688e-01 1.20913148e+00
-2.77929932e-01 6.30491912e-01 -6.85424805e-02 -1.26448834e+00
1.34506977e+00 1.33887902e-01 2.72479773e-01 -9.25720334e-01
8.28788280e-01 3.75731528e-01 2.94731379e-01 -4.78585154e-01
5.22668660e-01 7.57454112e-02 -1.69574201e-01 -5.50024817e-03
1.09931581e-01 3.47902030e-01 2.65900910e-01 -1.10191770e-01
5.69018006e-01 -3.26640457e-01 1.81428686e-01 -2.98642457e-01
6.86672330e-01 -4.50842738e-01 7.55753815e-01 4.45904016e-01
-6.61952555e-01 5.76126993e-01 3.70549977e-01 -1.32086098e+00
-1.30926180e+00 -7.43612051e-01 -1.85821995e-01 1.11540186e+00
-3.01695198e-01 -1.53070912e-01 -8.63628924e-01 -4.75943804e-01
1.75091758e-01 -4.48115207e-02 -1.08617067e+00 -5.84179498e-02
-6.00502908e-01 -8.50070119e-01 6.75278604e-01 8.05107474e-01
1.33925498e+00 -9.52789009e-01 -5.99046230e-01 -7.10206553e-02
1.07633352e-01 -1.10928631e+00 1.37677863e-01 -1.55615553e-01
-1.16666234e+00 -1.06333053e+00 -1.07939363e+00 -1.05315661e+00
8.90302539e-01 -2.06331417e-01 1.24020457e+00 4.52355385e-01
-4.78562087e-01 -2.52136141e-01 -2.32284233e-01 -1.00787652e+00
1.95949242e-01 6.49489999e-01 3.33245248e-01 -8.88015255e-02
8.41161013e-01 -7.04894841e-01 -1.03986752e+00 1.28300652e-01
-3.51337671e-01 1.86586142e-01 7.28343725e-01 6.03511691e-01
-1.98714167e-01 2.89478600e-01 8.31076980e-01 -8.26497436e-01
3.61176759e-01 -3.35974246e-01 -3.22736681e-01 -3.71009856e-02
-7.84022093e-01 -3.64235491e-01 6.51663482e-01 -3.50121439e-01
-9.71874774e-01 -9.82341692e-02 -1.30969852e-01 8.42643669e-04
-3.13244700e-01 2.85899580e-01 1.41929895e-01 1.19491242e-01
4.72659886e-01 2.84309275e-02 9.02358815e-02 -6.27899826e-01
-2.49869481e-01 9.55652356e-01 6.50420010e-01 -5.09341836e-01
7.21910119e-01 2.54392356e-01 1.61722332e-01 -6.88407421e-01
-9.82368648e-01 -2.43751839e-01 -8.83652091e-01 -4.34103996e-01
1.04745615e+00 -8.96663249e-01 -1.10560834e+00 1.29195666e+00
-7.96693861e-01 3.47139537e-01 4.85651225e-01 5.00056326e-01
2.05034912e-01 -1.04051061e-01 -5.53684771e-01 -9.87067878e-01
-7.81391859e-01 -7.36972272e-01 7.25532174e-01 9.70705152e-01
-3.40222746e-01 -1.02133286e+00 -3.93999726e-01 5.92633665e-01
7.35661983e-01 4.76546466e-01 9.11781311e-01 -4.92205650e-01
-1.62118934e-02 -5.08668363e-01 -7.37571597e-01 2.28700668e-01
-1.92104615e-02 8.87156352e-02 -1.06054473e+00 -2.54019141e-01
-4.88128275e-01 -3.81241083e-01 9.20819342e-01 4.92445320e-01
1.74759495e+00 -9.54173878e-02 -3.29113841e-01 6.77358389e-01
1.24262977e+00 3.26390564e-01 6.66381717e-01 4.23570067e-01
8.09797764e-01 6.24247789e-01 4.75492656e-01 5.81644833e-01
5.82817793e-01 1.04614899e-01 3.43335927e-01 -3.98024052e-01
2.57318646e-01 -6.64671734e-02 -4.24582660e-01 6.85504973e-01
-8.18323255e-01 3.50116253e-01 -1.31646287e+00 6.07748449e-01
-1.71570349e+00 -8.20381463e-01 -2.83902287e-01 2.01420856e+00
6.17410898e-01 1.65366337e-01 3.08513194e-01 6.49361551e-01
7.09514081e-01 1.24892123e-01 -4.50384200e-01 -6.25497580e-01
9.01933834e-02 6.27725601e-01 5.61181545e-01 -2.92480230e-01
-1.33195531e+00 6.61531687e-01 6.24014997e+00 5.94942629e-01
-1.28063464e+00 -2.55199909e-01 1.27757168e+00 2.77363151e-01
5.67213655e-01 -5.93015015e-01 -8.75844955e-01 5.09688437e-01
8.36155593e-01 -1.08161926e-01 1.80623978e-01 1.06935740e+00
-9.10772532e-02 -2.46027887e-01 -9.68361378e-01 1.43937004e+00
1.53512791e-01 -1.02618933e+00 -1.26880422e-01 1.70211107e-01
6.79496646e-01 -5.65596104e-01 3.64603162e-01 5.70496440e-01
-1.41388729e-01 -1.65653801e+00 4.72189784e-01 2.33362854e-01
8.02418888e-01 -1.25909686e+00 1.23819590e+00 2.33136907e-01
-1.12218237e+00 -4.78357911e-01 -5.85000515e-01 -8.98858011e-01
-4.30442512e-01 8.25501740e-01 -3.62960100e-01 3.95392030e-01
1.35357475e+00 7.32116342e-01 -1.04887807e+00 9.92513061e-01
2.32013971e-01 5.13300836e-01 6.18643202e-02 -2.73855954e-01
-5.02454415e-02 -9.04622674e-02 -4.03886110e-01 8.10228646e-01
5.89529872e-01 2.35175267e-01 1.95203349e-02 4.12620306e-01
-6.52835369e-02 -5.75648760e-03 -3.77707094e-01 -1.80874854e-01
4.93211627e-01 1.52722704e+00 -8.85960639e-01 -3.25447410e-01
-2.82288224e-01 5.51036835e-01 1.49853483e-01 -1.34882405e-01
-6.63871050e-01 -4.71753180e-01 6.84241235e-01 1.30928352e-01
-2.27063090e-01 -3.30740154e-01 -7.54446089e-01 -8.31284046e-01
-3.37016523e-01 -7.97467589e-01 4.42331970e-01 -3.73282820e-01
-1.23499870e+00 3.99684161e-01 -2.11913168e-01 -9.90267038e-01
-8.33028741e-03 -9.54513371e-01 -6.41826451e-01 6.81674302e-01
-1.13308752e+00 -1.41054249e+00 -9.34366822e-01 4.66245383e-01
4.26030517e-01 -4.37446684e-01 9.16385710e-01 7.08745420e-01
-8.47083867e-01 7.40205228e-01 -2.80417442e-01 7.68118083e-01
6.59175038e-01 -1.26026738e+00 3.87258589e-01 4.49233323e-01
-6.33374035e-01 7.01435804e-01 6.61022067e-01 -3.91961247e-01
-9.38656986e-01 -8.45556438e-01 1.28858602e+00 -2.54032075e-01
2.04461306e-01 -4.74410057e-01 -4.79600459e-01 5.39757073e-01
3.18048269e-01 -2.05395728e-01 5.82619786e-01 6.61498904e-01
-3.84554744e-01 -4.12976891e-01 -1.22853768e+00 4.80618358e-01
9.78096902e-01 -1.14391021e-01 -3.16068858e-01 -9.99994278e-02
2.56675392e-01 -3.41853529e-01 -7.72453547e-01 6.38300955e-01
9.46851313e-01 -1.21189547e+00 9.81743515e-01 -3.65117639e-01
7.69905448e-01 3.80376652e-02 4.25378233e-01 -9.58075523e-01
-4.05164778e-01 9.99579951e-03 -7.55944699e-02 1.45555639e+00
1.32059708e-01 -7.11801827e-01 1.23522162e+00 7.67051816e-01
2.15946659e-01 -1.08541691e+00 -5.84499002e-01 -4.63283688e-01
2.57769674e-01 -5.07631116e-02 7.50837088e-01 1.05534768e+00
-4.07893807e-01 3.20719957e-01 -5.10799646e-01 -2.60043532e-01
8.24414432e-01 -3.25278521e-01 7.93673038e-01 -1.90630829e+00
3.72728378e-01 -4.30512041e-01 -8.61352682e-01 -1.49728045e-01
-7.93301389e-02 -3.45192522e-01 -5.94250619e-01 -1.60276961e+00
5.00437081e-01 -5.34930944e-01 -3.04763466e-01 5.05271912e-01
-2.13350996e-01 5.69440842e-01 -1.03890844e-01 -1.95272267e-01
-3.57877463e-01 2.90248007e-01 1.15764904e+00 -2.54741073e-01
1.79263681e-01 2.17910811e-01 -7.36347616e-01 9.02702987e-01
1.26845193e+00 -1.04088545e-01 -3.32593173e-01 -4.34493244e-01
4.13884521e-01 -4.84641641e-01 1.58412814e-01 -1.54875672e+00
2.52955139e-01 -1.07318394e-01 1.30450106e+00 -7.23516226e-01
3.60930771e-01 -5.59683383e-01 -1.03789419e-02 6.96865141e-01
3.86763811e-02 4.14501965e-01 1.09430971e-02 -1.59337968e-01
-2.63848305e-01 1.32868931e-01 8.04673731e-01 -2.32406899e-01
-7.70692885e-01 6.58693075e-01 -1.05921254e-01 -2.89837301e-01
9.87883627e-01 -4.91497576e-01 -3.74102205e-01 -2.38578215e-01
-4.71363217e-01 3.08613777e-01 2.93947488e-01 7.04697728e-01
4.19377595e-01 -1.26857412e+00 -6.29755557e-01 2.63027906e-01
8.91697183e-02 5.85465916e-02 4.49070454e-01 6.23302579e-01
-7.69248009e-01 5.47424316e-01 -8.15321922e-01 -5.44783890e-01
-1.75186896e+00 2.61471421e-01 2.86672026e-01 -2.63043672e-01
-1.15926802e-01 9.90878999e-01 9.78194103e-02 -5.70172429e-01
5.15871465e-01 -6.59649670e-02 -8.54392707e-01 2.61194587e-01
6.43906295e-01 6.80052519e-01 -2.06295513e-02 -4.72892225e-01
-3.15245628e-01 6.74189985e-01 -2.70900041e-01 2.95885354e-01
1.42352271e+00 1.51538849e-01 -3.18098366e-01 2.02952862e-01
8.68100524e-01 -3.64937693e-01 -7.24998116e-01 2.17692509e-01
-1.40172035e-01 -6.41795754e-01 -1.89448558e-02 -8.25841486e-01
-1.37581551e+00 1.01980650e+00 9.19825137e-01 1.39901847e-01
1.12847865e+00 -2.31561944e-01 8.07404280e-01 1.88359797e-01
4.72151846e-01 -1.46477175e+00 1.01292029e-01 5.30444741e-01
3.65822077e-01 -1.51868415e+00 2.46421963e-01 -3.93853962e-01
-1.53806031e-01 1.20228195e+00 1.20830548e+00 -6.01407774e-02
7.57840514e-01 -2.30623171e-01 2.86691576e-01 -1.27289057e-01
-2.72813767e-01 -1.35939633e-02 3.07580739e-01 7.42800653e-01
7.74348497e-01 2.86357254e-01 -6.93668544e-01 8.04804802e-01
-7.10767210e-01 7.46042430e-02 1.93011609e-03 7.52396405e-01
-5.02114713e-01 -1.12588167e+00 -3.96550775e-01 8.41100156e-01
-9.45614517e-01 1.33758634e-01 -3.19051832e-01 7.35905945e-01
6.34652853e-01 8.11233163e-01 3.55973601e-01 -7.40800500e-01
-5.06749609e-04 -1.63839124e-02 5.31766951e-01 -2.61005968e-01
-7.16907561e-01 -7.33188272e-01 9.07152519e-02 -2.52284527e-01
-4.73071724e-01 -4.55288649e-01 -1.05903840e+00 -1.07240224e+00
1.74517289e-01 3.01102269e-02 8.69934678e-01 8.67282450e-01
1.25889555e-01 4.60440874e-01 4.87339199e-01 -6.99339986e-01
-7.35858679e-02 -1.23707354e+00 -6.77595913e-01 4.12709504e-01
-1.23879962e-01 -7.21892953e-01 -2.82764435e-03 -2.25146189e-01] | [13.554365158081055, 0.9853693842887878] |
beee4495-15b8-4245-9276-87c97fcf6fbd | dynamic-multistep-reasoning-based-on-video | null | null | https://aclanthology.org/2022.naacl-main.286 | https://aclanthology.org/2022.naacl-main.286.pdf | Dynamic Multistep Reasoning based on Video Scene Graph for Video Question Answering | Existing video question answering (video QA) models lack the capacity for deep video understanding and flexible multistep reasoning. We propose for video QA a novel model which performs dynamic multistep reasoning between questions and videos. It creates video semantic representation based on the video scene graph composed of semantic elements of the video and semantic relations among these elements. Then, it performs multistep reasoning for better answer decision between the representations of the question and the video, and dynamically integrate the reasoning results. Experiments show the significant advantage of the proposed model against previous methods in accuracy and interpretability. Against the existing state-of-the-art model, the proposed model dramatically improves more than 4\%/3.1\%/2\% on the three widely used video QA datasets, MSRVTT-QA, MSRVTT multi-choice, and TGIF-QA, and displays better interpretability by backtracing along with the attention mechanisms to the video scene graphs. | ['Yong Zhu', 'Hong Liu', 'Yajuan Lyu', 'Zhifan Feng', 'Xiangdong Wang', 'Wenbin Jiang', 'Jianguo Mao'] | null | null | null | null | naacl-2022-7 | ['video-question-answering'] | ['computer-vision'] | [-6.63028508e-02 -2.59155091e-02 3.81443761e-02 -3.39699507e-01
-6.75775766e-01 -5.35241961e-01 3.94159228e-01 -1.96181193e-01
-4.15258063e-03 3.75197172e-01 5.82353234e-01 -3.95668745e-01
-3.10746670e-01 -6.44925773e-01 -8.67808223e-01 -1.22679643e-01
3.70056301e-01 4.44973588e-01 7.18243062e-01 -4.51682895e-01
2.21792951e-01 -3.11107248e-01 -1.66882515e+00 1.02888596e+00
9.73741174e-01 1.26431799e+00 4.03097570e-01 9.38242912e-01
-5.70311308e-01 1.72530210e+00 -5.54819643e-01 -8.33759964e-01
-8.51970762e-02 -6.21273279e-01 -1.31872320e+00 2.72946626e-01
7.70804346e-01 -6.97019279e-01 -6.90079510e-01 9.35899734e-01
-6.20445721e-02 2.57030040e-01 3.43170881e-01 -1.44482112e+00
-1.07714927e+00 4.84565616e-01 -2.32756779e-01 5.31265795e-01
1.15521550e+00 2.70834446e-01 1.04886055e+00 -8.91419113e-01
9.04587746e-01 1.58240175e+00 3.28819752e-01 3.14623147e-01
-3.89961690e-01 -4.47446406e-01 5.03665864e-01 1.27280140e+00
-1.10206866e+00 -3.56495351e-01 6.16670012e-01 -2.75302619e-01
1.06184638e+00 5.89648068e-01 7.98921227e-01 9.36954916e-01
-1.15401046e-02 1.06639564e+00 7.14257360e-01 -4.98296209e-02
3.33711528e-03 -3.30747396e-01 4.29386497e-01 1.13404155e+00
-2.97661781e-01 -6.16036177e-01 -6.43367231e-01 1.50889695e-01
6.78384304e-01 4.52767611e-02 -4.44640160e-01 -1.28896087e-01
-1.25947964e+00 4.54440802e-01 2.73095816e-01 2.00085938e-01
-3.54335785e-01 2.16505438e-01 4.27835733e-01 4.79054511e-01
-1.23342285e-02 1.28337666e-01 -2.14197919e-01 -3.61970514e-01
-6.12763703e-01 2.76761919e-01 7.10545182e-01 1.24553883e+00
5.28839648e-01 1.34495264e-02 -4.18543696e-01 4.79024619e-01
5.65952420e-01 5.07522345e-01 1.87659502e-01 -1.57338405e+00
8.93266797e-01 1.04386103e+00 1.15058526e-01 -1.37648022e+00
-3.55574414e-02 9.17470083e-02 -3.36321294e-01 -3.74878943e-01
3.28903794e-01 2.40015417e-01 -8.92147124e-01 1.29693031e+00
3.71744961e-01 4.28343713e-01 3.18361610e-01 9.32052612e-01
1.47315228e+00 9.04852808e-01 3.35983902e-01 -3.21110845e-01
1.69177151e+00 -1.54623318e+00 -1.19351923e+00 -2.08723947e-01
5.87991476e-01 -7.04623640e-01 1.36654663e+00 2.69600153e-01
-1.33062685e+00 -9.82133985e-01 -8.46519709e-01 -6.13990068e-01
-2.94555187e-01 -9.41668749e-02 3.28063190e-01 2.35352352e-01
-1.29202878e+00 6.62580412e-03 -3.12488705e-01 -4.14570183e-01
3.91289264e-01 7.96290189e-02 -3.64811689e-01 -7.50002742e-01
-1.38475871e+00 6.24358654e-01 3.05844218e-01 1.68033898e-01
-1.03523695e+00 -5.76890290e-01 -8.56970251e-01 2.10557133e-01
8.55388999e-01 -1.13987672e+00 1.39496768e+00 -1.26345229e+00
-1.42480028e+00 5.46078265e-01 -5.26626766e-01 -2.31273338e-01
4.51997340e-01 -6.82810307e-01 -7.66286612e-01 9.24748659e-01
7.26271644e-02 6.14188612e-01 8.73703063e-01 -1.11926663e+00
-8.33750725e-01 -2.72483557e-01 9.88796890e-01 4.70180720e-01
-1.06291503e-01 5.28312661e-02 -1.08176088e+00 -3.69827390e-01
4.91101481e-02 -4.54829812e-01 1.64306626e-01 6.17710501e-02
1.06013879e-01 -3.33089113e-01 1.34626114e+00 -1.10985422e+00
1.38195956e+00 -2.00618935e+00 4.41574365e-01 -2.78125495e-01
3.98490787e-01 2.92242408e-01 -3.57639849e-01 4.34188366e-01
4.61925827e-02 4.30481620e-02 2.76561469e-01 2.59158742e-02
-1.36675611e-01 4.23803061e-01 -1.01024836e-01 -7.81072229e-02
-2.09908932e-02 1.15059078e+00 -1.07700157e+00 -7.49622703e-01
2.78476477e-01 2.75013536e-01 -7.07867980e-01 3.90364736e-01
-6.73830032e-01 1.67132795e-01 -6.11786962e-01 6.77165568e-01
5.53116798e-01 -6.66129768e-01 3.05536836e-01 -6.08798921e-01
3.88507575e-01 -7.63467997e-02 -1.01225007e+00 1.95904899e+00
-1.29967555e-01 7.10070252e-01 -1.40092283e-01 -8.00222516e-01
5.06620228e-01 5.10588527e-01 3.08419853e-01 -1.01107645e+00
4.92902994e-02 -2.04105571e-01 -2.20270157e-01 -1.37009835e+00
5.04617572e-01 3.56923610e-01 1.47813499e-01 1.61892965e-01
2.54086405e-01 7.35668764e-02 4.11826849e-01 7.06041813e-01
9.46652353e-01 4.48254377e-01 9.83256400e-02 -5.84062971e-02
1.07014680e+00 2.92934954e-01 1.66413218e-01 6.63794398e-01
-2.85770953e-01 2.76921839e-01 6.33097827e-01 -5.22865295e-01
-5.39863765e-01 -1.09099054e+00 6.68185890e-01 1.31841803e+00
7.23168671e-01 -8.81894112e-01 -9.84128952e-01 -8.06398690e-01
-3.56301188e-01 7.66496062e-01 -5.26947677e-01 -1.97360367e-01
-6.81767046e-01 -5.75965717e-02 3.30869108e-01 7.11762726e-01
1.06961608e+00 -8.17150414e-01 -3.76569211e-01 -1.07195705e-01
-1.16091394e+00 -1.55610049e+00 -3.13623369e-01 -8.43095779e-01
-7.80170381e-01 -1.58930564e+00 -1.74645513e-01 -7.86221921e-01
5.61156213e-01 7.59969950e-01 1.39227676e+00 6.35845780e-01
1.52291819e-01 9.91590977e-01 -7.93689847e-01 1.20452821e-01
-2.42032751e-01 -4.28120613e-01 -3.68665725e-01 1.60204068e-01
2.52974063e-01 4.17966358e-02 -8.38055432e-01 6.17756069e-01
-1.04152906e+00 4.56781864e-01 1.08396493e-01 5.17749667e-01
4.21257168e-01 2.78976373e-02 4.21766341e-01 -6.37690485e-01
3.97098094e-01 -7.18990207e-01 -1.40985847e-01 7.39688337e-01
-1.87552661e-01 -9.29158647e-03 5.69987059e-01 -1.30105346e-01
-1.59901476e+00 -5.95376492e-01 -1.87008768e-01 -6.27444148e-01
2.84632184e-02 4.06297237e-01 -2.37527296e-01 4.81601879e-02
3.59700233e-01 1.11223675e-01 -1.65207520e-01 -8.81184861e-02
6.96869850e-01 4.70213264e-01 6.92280471e-01 -4.90532070e-01
5.77997565e-01 6.50739789e-01 -3.20502460e-01 -5.27180791e-01
-8.80683482e-01 -4.30526733e-01 -4.88375187e-01 -8.34554434e-01
1.43017387e+00 -8.90403807e-01 -1.04576170e+00 2.60554075e-01
-1.27235615e+00 -2.75745168e-02 -8.70026648e-03 1.91265702e-01
-6.86243296e-01 8.10798824e-01 -7.56094694e-01 -5.44150352e-01
-1.91256687e-01 -1.31722593e+00 9.18551326e-01 2.76382983e-01
-1.03809930e-01 -9.53905106e-01 -6.01915121e-01 1.37967527e+00
2.66257107e-01 -1.80080712e-01 1.15410483e+00 -1.96884945e-01
-1.01145184e+00 1.63659081e-01 -5.46582639e-01 1.21253341e-01
-1.21032884e-02 1.80456474e-01 -6.77647412e-01 3.01867425e-02
8.90357047e-02 -3.73011082e-01 6.10196352e-01 8.22842345e-02
1.44051397e+00 -2.94472933e-01 -9.29587632e-02 3.28192443e-01
1.30159914e+00 6.08189583e-01 1.01008797e+00 2.52258748e-01
8.59899163e-01 5.30110657e-01 9.06337798e-01 1.20856397e-01
1.08032358e+00 4.03483748e-01 8.27344120e-01 2.78740644e-01
-3.25335652e-01 -4.28885221e-01 5.37593782e-01 1.17959774e+00
-3.55358392e-01 -5.92378199e-01 -7.31209755e-01 4.24751937e-01
-2.33560109e+00 -1.25240040e+00 -4.64308918e-01 1.46592700e+00
9.09662917e-02 -8.84573758e-02 -9.89176184e-02 -6.05888665e-03
4.96202260e-01 3.41076165e-01 -3.87167364e-01 -3.30525398e-01
-2.16936748e-02 -2.31098056e-01 -2.48167574e-01 4.15912122e-01
-7.51596451e-01 9.89074707e-01 6.74734545e+00 7.24043787e-01
-4.69288230e-01 1.03728659e-01 5.20319462e-01 -6.52289465e-02
-4.86608326e-01 -3.66442166e-02 -3.82335305e-01 2.42857859e-01
9.48731184e-01 6.42696247e-02 3.84478807e-01 5.53184092e-01
1.53608650e-01 -3.86370480e-01 -9.76022840e-01 1.23054373e+00
6.25537217e-01 -1.70815885e+00 7.28523195e-01 -7.30912030e-01
3.75152200e-01 -4.99439985e-01 -8.88931453e-02 5.21957338e-01
4.55614552e-02 -8.49225521e-01 9.29038286e-01 7.16579974e-01
5.34874201e-01 -4.04117107e-01 7.90726364e-01 1.81787595e-01
-1.54857743e+00 -4.18018878e-01 -1.70928106e-01 -1.17509298e-01
4.30594265e-01 -2.45891437e-01 -3.15483987e-01 1.15697742e+00
1.10467517e+00 8.51209998e-01 -8.60453963e-01 5.69633126e-01
-2.10537195e-01 4.31834519e-01 2.38641784e-01 1.23671912e-01
3.05777133e-01 -2.20194355e-01 3.14048737e-01 8.50218177e-01
3.78746897e-01 4.88702685e-01 6.62633702e-02 3.34961027e-01
-5.28233014e-02 -3.91348172e-03 -2.75743812e-01 -5.23784682e-02
1.06244743e-01 8.99481297e-01 -5.39760947e-01 -7.93244600e-01
-8.47736001e-01 1.17510426e+00 3.33742708e-01 5.68889976e-01
-1.35736585e+00 3.73261943e-02 6.46840036e-01 7.95267895e-02
4.14328814e-01 -1.00713000e-01 2.83703685e-01 -1.35327685e+00
1.05494909e-01 -1.19354892e+00 1.12080991e+00 -1.81044269e+00
-8.30739617e-01 7.58023620e-01 2.89459944e-01 -1.03778934e+00
-1.68351457e-01 -7.56278574e-01 -2.63502926e-01 1.40969604e-01
-1.42873013e+00 -1.07866287e+00 -7.93081701e-01 1.10512853e+00
1.20076108e+00 5.73880933e-02 5.12594581e-01 4.77735937e-01
-3.98856014e-01 1.02206118e-01 -4.47199315e-01 2.34228070e-03
3.88120204e-01 -8.17240715e-01 1.80834204e-01 9.49938774e-01
2.44171068e-01 2.48855293e-01 4.89764690e-01 -6.12700641e-01
-1.85760188e+00 -8.80700171e-01 5.89472175e-01 -6.67714477e-01
7.07228541e-01 -5.50565263e-03 -9.23796177e-01 8.45201254e-01
7.16540635e-01 -1.66900456e-01 6.12103939e-01 -3.63329709e-01
-5.71450472e-01 -2.63527445e-02 -9.78812754e-01 8.43058765e-01
1.60499775e+00 -7.28410602e-01 -9.80760098e-01 3.76548529e-01
1.22321105e+00 -6.18205667e-01 -8.85702133e-01 2.54705518e-01
5.03349125e-01 -1.05523419e+00 1.17476916e+00 -9.13419247e-01
6.81003213e-01 -5.67055643e-01 -3.61612856e-01 -7.67796159e-01
-4.63072091e-01 -4.31781888e-01 -4.81274486e-01 9.98180330e-01
2.05385219e-02 -7.28823319e-02 4.97029871e-01 5.46962142e-01
-2.91631281e-01 -4.02163446e-01 -7.41078556e-01 -2.86788553e-01
-6.49369180e-01 -6.62242115e-01 5.54528534e-01 8.87855113e-01
1.58820080e-03 3.36797148e-01 -3.85730505e-01 2.62327909e-01
2.05569983e-01 2.45000049e-01 6.74550056e-01 -8.30102086e-01
-2.38112479e-01 -1.16613254e-01 -4.19847220e-01 -1.46741271e+00
3.39308456e-02 -5.40272534e-01 -3.34767163e-01 -2.15321636e+00
1.12350821e-01 3.95850629e-01 -1.90658867e-01 2.10922718e-01
-4.27867621e-01 1.53473184e-01 5.14414907e-01 7.29302317e-02
-1.48393941e+00 3.64656717e-01 1.69081211e+00 -2.10912168e-01
2.73002088e-01 -6.48763299e-01 -6.39425159e-01 8.04876328e-01
4.15020227e-01 -2.10622046e-03 -9.92689073e-01 -1.14400649e+00
3.87099028e-01 6.20538473e-01 4.98351842e-01 -1.07682490e+00
2.47025952e-01 -2.55139291e-01 2.86933780e-01 -6.35850668e-01
5.52697897e-01 -9.69309449e-01 3.65270942e-01 3.06978464e-01
-2.60908842e-01 5.78436732e-01 1.16044052e-01 8.76685619e-01
-5.26113629e-01 -8.94836187e-02 2.87019253e-01 -3.60493153e-01
-1.56957531e+00 2.66286522e-01 -4.62944686e-01 2.36389026e-01
1.22344005e+00 -4.83587742e-01 -9.32457387e-01 -7.98615932e-01
-7.29132414e-01 7.15739131e-01 5.89984618e-02 7.70596623e-01
1.05005777e+00 -1.33879495e+00 -3.50970328e-01 -7.37808347e-02
2.14314789e-01 -8.60419199e-02 9.29225028e-01 7.88106382e-01
-8.90805125e-01 3.83816481e-01 -3.87579143e-01 -5.86607933e-01
-1.48231304e+00 7.02736199e-01 2.73989439e-01 -4.44446281e-02
-5.07476985e-01 8.29820454e-01 3.64059538e-01 7.70707279e-02
7.84038752e-02 -5.15265167e-01 -5.66111326e-01 -9.23710540e-02
7.07632005e-01 5.58271170e-01 -2.00813681e-01 -9.00251150e-01
-2.45258659e-01 6.50584042e-01 7.20498562e-02 1.57689184e-01
9.24782336e-01 -6.31797552e-01 -1.46036804e-01 2.17278808e-01
9.11951065e-01 -4.17280257e-01 -1.15564942e+00 -1.55946612e-01
-1.60974517e-01 -7.24368989e-01 -3.63378376e-01 -7.25557446e-01
-1.12434006e+00 7.65722394e-01 3.20684373e-01 1.80673853e-01
1.36958075e+00 1.83840603e-01 9.51378763e-01 4.66381609e-01
2.60256618e-01 -1.00447178e+00 7.42810607e-01 4.07727987e-01
9.44147885e-01 -1.07079768e+00 -1.98056877e-01 -7.64031529e-01
-8.60188901e-01 1.19385564e+00 1.02425396e+00 9.79925692e-02
3.76157194e-01 -2.55498707e-01 2.53908932e-01 -4.69725132e-01
-1.12457860e+00 -1.22546874e-01 3.96179825e-01 6.73744738e-01
9.30641964e-03 -2.79271752e-01 -9.47856307e-02 5.13291478e-01
1.31085843e-01 2.24039868e-01 3.69053155e-01 6.96005464e-01
-4.67125475e-01 -6.57385707e-01 -2.65044898e-01 2.83243030e-01
-3.26042950e-01 5.57740405e-02 -2.83996075e-01 8.90349269e-01
1.22343041e-01 1.38576794e+00 1.93209618e-01 -5.53830624e-01
4.26910847e-01 2.65260905e-01 3.97905558e-01 -7.21670315e-03
-5.61547756e-01 -3.21938694e-01 4.48931694e-01 -1.11185193e+00
-8.27587008e-01 -4.49936301e-01 -1.37213612e+00 -3.66451055e-01
4.91058314e-03 1.32656604e-01 8.56642574e-02 1.32963896e+00
5.52701116e-01 7.06135869e-01 3.79938222e-02 -8.60200077e-02
1.22354977e-01 -4.80349839e-01 3.40129063e-02 7.94407189e-01
6.24041893e-02 -6.36812687e-01 -1.17577530e-01 5.40938020e-01] | [10.430496215820312, 1.079901933670044] |
1157d2a6-a97b-4ac9-8b5f-5e566116af18 | wronging-a-right-generating-better-errors-to-1 | 1810.00668 | null | http://arxiv.org/abs/1810.00668v1 | http://arxiv.org/pdf/1810.00668v1.pdf | Wronging a Right: Generating Better Errors to Improve Grammatical Error Detection | Grammatical error correction, like other machine learning tasks, greatly
benefits from large quantities of high quality training data, which is
typically expensive to produce. While writing a program to automatically
generate realistic grammatical errors would be difficult, one could learn the
distribution of naturallyoccurring errors and attempt to introduce them into
other datasets. Initial work on inducing errors in this way using statistical
machine translation has shown promise; we investigate cheaply constructing
synthetic samples, given a small corpus of human-annotated data, using an
off-the-rack attentive sequence-to-sequence model and a straight-forward
post-processing procedure. Our approach yields error-filled artificial data
that helps a vanilla bi-directional LSTM to outperform the previous state of
the art at grammatical error detection, and a previously introduced model to
gain further improvements of over 5% $F_{0.5}$ score. When attempting to
determine if a given sentence is synthetic, a human annotator at best achieves
39.39 $F_1$ score, indicating that our model generates mostly human-like
instances. | ['Riedel Sebastian', 'Stenetorp Pontus', 'Kasewa Sudhanshu'] | 2018-09-26 | wronging-a-right-generating-better-errors-to | https://aclanthology.org/D18-1541 | https://aclanthology.org/D18-1541.pdf | emnlp-2018-10 | ['grammatical-error-detection'] | ['natural-language-processing'] | [ 5.74875593e-01 4.90559846e-01 4.13904369e-01 -6.95564091e-01
-1.46833789e+00 -4.58163738e-01 3.51816684e-01 2.92168498e-01
-5.25970638e-01 1.15320349e+00 3.36383432e-02 -7.34165370e-01
7.04919279e-01 -6.48922801e-01 -1.28387558e+00 -6.17677905e-02
2.44093552e-01 7.39611804e-01 -1.41960725e-01 -3.99205089e-01
4.23352033e-01 -2.79508322e-01 -1.36800241e+00 4.74558771e-01
1.39578629e+00 2.16592044e-01 2.91141272e-01 1.07588625e+00
-2.04970464e-01 7.89295137e-01 -1.02841318e+00 -7.34197795e-01
-9.27252695e-02 -8.02670300e-01 -1.02543831e+00 -2.89649576e-01
6.24706209e-01 -7.46639147e-02 2.65618682e-01 1.06517482e+00
5.61793566e-01 -5.16928323e-02 4.04987305e-01 -7.86973774e-01
-9.39968824e-01 8.49768102e-01 2.73768464e-03 3.31137389e-01
6.51647449e-01 5.52329957e-01 8.54733169e-01 -9.41091657e-01
8.13558459e-01 1.12121904e+00 9.00563002e-01 8.69687676e-01
-1.39351785e+00 -5.50834119e-01 -2.65632361e-01 -2.71183640e-01
-8.64953041e-01 -5.99847317e-01 4.15650815e-01 -4.31959748e-01
1.49270058e+00 1.63948953e-01 3.20696443e-01 1.29669869e+00
6.03003614e-02 7.97698140e-01 1.11043525e+00 -9.79289830e-01
2.61389520e-02 2.19301939e-01 -4.72232848e-02 8.75579894e-01
1.87040269e-01 1.41016562e-02 -4.15791780e-01 -9.28057358e-02
2.95304060e-01 -6.74006581e-01 -2.58020699e-01 2.54100233e-01
-1.11874962e+00 6.81804061e-01 2.39272013e-01 2.21741438e-01
-9.86161828e-02 3.89055252e-01 6.07832551e-01 6.64044976e-01
7.08334029e-01 9.55221593e-01 -7.06362128e-01 -7.34500468e-01
-1.19285369e+00 5.08219302e-01 8.01310480e-01 1.15620399e+00
5.44285834e-01 3.31856370e-01 -1.94295958e-01 9.48249340e-01
6.12188084e-03 3.23502988e-01 5.83912730e-01 -8.83326948e-01
8.76543999e-01 3.64621073e-01 3.67987007e-01 -6.55619562e-01
-5.16787320e-02 -2.48143554e-01 -3.44995886e-01 1.65529385e-01
8.02179933e-01 -3.65714788e-01 -9.40213323e-01 1.84765756e+00
-6.86461255e-02 -2.26552218e-01 -2.44721567e-04 5.47149301e-01
4.38928425e-01 6.55118883e-01 1.96299106e-01 -3.45729440e-02
8.52944791e-01 -8.48630428e-01 -4.10576850e-01 -5.49152911e-01
1.31479228e+00 -9.71248627e-01 1.63518250e+00 3.81119162e-01
-1.32933402e+00 -4.63165551e-01 -9.20106173e-01 -2.37990484e-01
-3.22994851e-02 2.65209913e-01 4.75346118e-01 6.88445747e-01
-1.14014983e+00 9.97054517e-01 -8.15687120e-01 -3.55977148e-01
3.97144943e-01 4.37890477e-02 -3.11756104e-01 -3.24436545e-01
-1.17055583e+00 1.21715164e+00 3.48075986e-01 -4.99006957e-02
-8.21283937e-01 -7.71182299e-01 -1.10132790e+00 -2.12599367e-01
8.97790045e-02 -6.01923645e-01 1.64703786e+00 -1.22425723e+00
-1.16469336e+00 9.95645285e-01 -5.15723288e-01 -4.95272666e-01
6.52994871e-01 -3.10266525e-01 -2.62526304e-01 -4.36069548e-01
2.98654944e-01 7.41718471e-01 5.93867242e-01 -1.09847331e+00
-5.48101485e-01 -1.34680688e-01 -1.95389464e-01 -7.45296031e-02
4.72347625e-03 4.79410827e-01 1.56840071e-01 -7.72395194e-01
-2.67953366e-01 -8.06088567e-01 -1.80385500e-01 -2.74163425e-01
-3.73436093e-01 -3.70235473e-01 9.71708447e-02 -1.11112618e+00
1.25018024e+00 -1.65492749e+00 -2.45151985e-02 -1.67476773e-01
-1.89326033e-01 4.78518337e-01 -3.42908144e-01 4.13714468e-01
-1.54982671e-01 4.74919975e-01 -5.96864223e-01 -4.46314365e-01
8.72191116e-02 1.36243692e-02 -9.13939551e-02 1.19350910e-01
6.78351283e-01 9.21154797e-01 -1.32243717e+00 -2.56132424e-01
-3.47084165e-01 1.13073304e-01 -8.07139456e-01 2.65787899e-01
-5.60173750e-01 1.82344913e-01 1.00214235e-01 5.51093221e-01
3.02021921e-01 -5.78533597e-02 2.93559097e-02 5.13300240e-01
1.31148055e-01 9.88496304e-01 -7.96643794e-01 1.82546890e+00
-7.08346903e-01 8.42792869e-01 -2.11600989e-01 -9.15403247e-01
1.06243193e+00 2.78059810e-01 -3.59432936e-01 -6.36890054e-01
-1.27575127e-02 7.75668919e-01 3.80586475e-01 -6.07203484e-01
8.35990429e-01 -3.69265854e-01 -2.41026103e-01 6.29708827e-01
2.66110003e-01 -3.38379622e-01 4.77905154e-01 1.74716473e-01
1.42246115e+00 4.53966975e-01 -1.78716570e-01 -2.44708255e-01
1.44097313e-01 4.07321483e-01 6.01213872e-01 8.50264192e-01
-2.36709729e-01 7.72121191e-01 5.02761960e-01 -4.74440366e-01
-1.64094174e+00 -5.40944278e-01 8.73998404e-02 9.92529213e-01
-6.55282021e-01 -2.91872621e-01 -1.28424573e+00 -8.48080099e-01
-2.45901152e-01 1.37050784e+00 -3.76148403e-01 -2.10962325e-01
-8.39411139e-01 -7.32980609e-01 9.14666116e-01 5.32839954e-01
1.49341216e-02 -1.36600995e+00 -3.40014338e-01 5.76098323e-01
-4.25167799e-01 -7.54089355e-01 -4.76206064e-01 2.32751310e-01
-7.92955101e-01 -7.46181786e-01 -3.75569493e-01 -8.88941526e-01
8.89521182e-01 -4.85227942e-01 1.76792729e+00 4.09104526e-01
-3.61665368e-01 -3.69080514e-01 -4.38409626e-01 -5.40315926e-01
-1.21056509e+00 8.30480084e-02 -1.13103166e-01 -7.19359100e-01
7.88900495e-01 -3.34929079e-01 -5.07804751e-02 -2.53778756e-01
-6.25530005e-01 1.85083225e-02 4.01477784e-01 1.31104982e+00
1.14573695e-01 -3.78792644e-01 6.71707213e-01 -1.39599943e+00
9.59008276e-01 -2.65841722e-01 -3.06631207e-01 1.26973942e-01
-6.00357592e-01 2.64773786e-01 8.52496862e-01 -1.22686721e-01
-8.44697297e-01 -5.56035116e-02 -4.29375887e-01 3.26114632e-02
-3.04462850e-01 4.45154637e-01 1.27472848e-01 3.17412704e-01
1.15489757e+00 3.29410970e-01 -3.65292430e-02 -3.26992512e-01
3.39928061e-01 8.98380518e-01 4.17719245e-01 -8.93326998e-01
4.81886476e-01 -4.48543698e-01 -6.38689935e-01 -3.75431329e-01
-8.15040708e-01 1.37736380e-01 -5.64808965e-01 1.05934970e-01
2.89374620e-01 -8.29242885e-01 -1.89889133e-01 4.40010250e-01
-1.54357660e+00 -6.98732436e-01 -2.35294178e-01 1.44875184e-01
-6.28511250e-01 1.90787673e-01 -7.42576361e-01 -6.88625991e-01
-4.01317656e-01 -1.11497819e+00 1.11282635e+00 -1.88408315e-01
-9.29817379e-01 -8.60063195e-01 1.28972366e-01 5.40531874e-01
5.17598450e-01 2.09661886e-01 1.09255803e+00 -7.16927826e-01
-4.31806266e-01 -2.39814326e-01 -8.06459859e-02 6.86091959e-01
-1.82437152e-01 2.40388796e-01 -7.26979077e-01 -3.42089176e-01
-1.93921983e-01 -7.65967727e-01 6.26377702e-01 -1.31824926e-01
9.62185204e-01 -4.30887640e-01 -3.94166168e-03 1.08145550e-01
1.24669909e+00 8.76687765e-02 6.06778204e-01 2.63582468e-01
5.46792150e-01 5.14559925e-01 6.07086062e-01 1.27574354e-01
4.73994374e-01 4.04660702e-01 1.03235535e-01 2.19577570e-02
-1.18784964e-01 -6.24779999e-01 4.44132537e-01 1.02792919e+00
3.07363123e-01 -3.42241287e-01 -1.14408422e+00 1.00118232e+00
-1.42576075e+00 -9.86012101e-01 -3.11765671e-01 2.16151047e+00
1.51616812e+00 3.79762083e-01 -1.47996753e-01 1.56369656e-01
7.03971505e-01 -3.06702495e-01 -3.81572396e-01 -1.05822682e+00
1.89085573e-01 6.53267264e-01 2.91535944e-01 7.53599167e-01
-7.25155473e-01 1.21881270e+00 7.06227541e+00 7.30840147e-01
-1.08349872e+00 3.39821875e-02 8.30527186e-01 -1.82310775e-01
-7.06403255e-01 1.16024530e-02 -8.97791028e-01 9.95169938e-01
1.58346152e+00 -2.02627778e-02 5.47084868e-01 8.66870701e-01
3.87911111e-01 -1.08723439e-01 -1.29795706e+00 5.69356322e-01
2.54883349e-01 -1.18546283e+00 -3.62609811e-02 -2.70973742e-01
7.97346115e-01 9.67869386e-02 -3.33237231e-01 6.29746735e-01
9.34899867e-01 -1.40978897e+00 8.28585267e-01 4.61136192e-01
8.73494506e-01 -7.46472001e-01 7.60624647e-01 7.03868926e-01
-2.60411084e-01 2.50682741e-01 -5.43194592e-01 -5.27249932e-01
1.91456497e-01 8.47856522e-01 -1.07835722e+00 1.01674661e-01
5.05027652e-01 4.26221341e-01 -6.90599799e-01 7.08249211e-01
-5.07923663e-01 1.08302593e+00 -3.15290600e-01 -4.52499747e-01
2.26416558e-01 8.09363425e-02 2.84167200e-01 1.57941997e+00
5.85093975e-01 -1.25900730e-01 -8.86214674e-02 1.11593521e+00
-4.21477646e-01 1.07002951e-01 -7.43142903e-01 -2.64346391e-01
6.16142154e-01 7.70899355e-01 -1.61621451e-01 -5.74488580e-01
-1.35627881e-01 1.16078794e+00 8.82844269e-01 1.12080202e-02
-6.55309141e-01 -8.10184479e-01 4.91735846e-01 4.81746122e-02
1.50505483e-01 -1.78701758e-01 -6.46690607e-01 -1.15526223e+00
4.39079165e-01 -1.44880331e+00 -1.70319840e-01 -9.29984212e-01
-1.03765094e+00 7.74502218e-01 -6.76894486e-01 -9.08446133e-01
-6.44757450e-01 -5.86298108e-01 -6.16573513e-01 1.34501779e+00
-1.17478120e+00 -8.44045818e-01 -2.86129978e-03 -1.28686085e-01
9.42254543e-01 -5.74831478e-03 9.80840981e-01 3.60703856e-01
-5.99556863e-01 1.07620871e+00 1.19683452e-01 2.88857102e-01
8.68299186e-01 -1.52255011e+00 1.09774244e+00 1.17775846e+00
1.56449601e-01 8.18132699e-01 9.66369927e-01 -7.04894483e-01
-9.80632186e-01 -1.21025848e+00 1.71824086e+00 -9.36337829e-01
5.08584380e-01 -5.31927884e-01 -1.05083847e+00 8.82220984e-01
3.19948196e-01 -6.86812401e-02 4.40222263e-01 9.40099955e-02
-3.15818012e-01 4.05452311e-01 -1.24668074e+00 5.71219265e-01
1.20781314e+00 -6.10847771e-01 -7.75994062e-01 5.83023310e-01
7.48980403e-01 -6.72159493e-01 -7.76748717e-01 1.88445911e-01
1.05814353e-01 -7.60757804e-01 3.64242285e-01 -1.10826766e+00
1.30845857e+00 -1.84356019e-01 4.55739768e-03 -1.93196404e+00
-6.88449070e-02 -7.01090872e-01 2.85421014e-01 1.41944933e+00
9.96539414e-01 -2.27235124e-01 7.54165351e-01 8.38310003e-01
-5.67280173e-01 -7.71091580e-01 -6.96992815e-01 -7.75942922e-01
6.41367674e-01 -5.86998820e-01 5.28034985e-01 9.72049296e-01
2.79852509e-01 3.81676674e-01 -3.04623008e-01 -2.07244948e-01
3.50671649e-01 -4.49461639e-01 7.98987210e-01 -7.28042245e-01
-3.84474754e-01 -2.70023376e-01 -1.72496036e-01 -9.93516624e-01
1.61392018e-01 -1.14715719e+00 6.33310914e-01 -1.36675119e+00
-8.50253105e-02 -7.16793239e-01 1.59930393e-01 4.76917744e-01
-6.86899662e-01 2.61247277e-01 -8.58119875e-02 -2.38774031e-01
-3.56845170e-01 3.93086016e-01 9.31015670e-01 -9.66647267e-02
1.21894024e-01 -4.21121478e-01 -7.76292443e-01 5.11725247e-01
7.82279730e-01 -7.53981233e-01 -1.19317606e-01 -9.90493417e-01
3.57808828e-01 4.63587120e-02 -3.99332531e-02 -8.92093062e-01
-8.08685645e-02 8.41508061e-02 1.82139635e-01 -1.28284380e-01
-8.16011801e-02 -6.59190342e-02 -7.67194331e-02 4.23732013e-01
-7.67184198e-01 3.28206152e-01 2.50721872e-01 1.83462664e-01
-3.29411238e-01 -7.68415630e-01 8.51572514e-01 -6.50829971e-01
-3.47999185e-01 -2.41075858e-01 -4.10324484e-01 5.86979330e-01
6.61481798e-01 -4.79004532e-02 -3.84390950e-01 -1.92178518e-01
-5.82104385e-01 7.73119032e-02 7.75375426e-01 3.21439743e-01
2.65894204e-01 -9.83775139e-01 -1.07571566e+00 2.44444743e-01
2.24981532e-01 6.25443682e-02 -1.70222804e-01 3.63960594e-01
-9.03233111e-01 2.51361012e-01 -4.40638401e-02 -4.76714939e-01
-1.10759103e+00 2.06014365e-01 2.94563055e-01 -2.68494457e-01
-3.55791986e-01 1.39908743e+00 -7.12587655e-01 -1.01743424e+00
-5.93274608e-02 -3.83668333e-01 3.58327389e-01 -4.22950208e-01
4.61725235e-01 2.08347574e-01 5.43109179e-01 -2.87555635e-01
-3.62534560e-02 6.00208417e-02 -2.33821318e-01 -1.93699703e-01
1.35758388e+00 2.61705875e-01 -2.15813875e-01 3.19906801e-01
1.03581011e+00 -1.35987038e-02 -1.01634121e+00 -9.25431475e-02
2.78070122e-01 -6.24227285e-01 -4.32211608e-01 -1.24336803e+00
-4.16786641e-01 1.09887552e+00 8.39798376e-02 8.15871451e-03
5.15498519e-01 -3.39479953e-01 1.03357065e+00 4.58085448e-01
5.78282773e-01 -1.09219682e+00 -4.26622964e-02 8.43553185e-01
7.65884638e-01 -1.49302721e+00 -4.92818385e-01 -2.48111740e-01
-5.11472583e-01 9.16824996e-01 7.90831447e-01 -3.44235092e-01
-6.16336316e-02 2.80057847e-01 2.38130912e-01 1.51364729e-01
-9.94738758e-01 3.87476623e-01 -2.39082173e-01 5.79548061e-01
1.12328959e+00 5.28475940e-02 -5.74811935e-01 2.00842023e-01
-6.80467904e-01 1.51315942e-01 1.02067852e+00 8.95256639e-01
-5.23760319e-01 -1.47028828e+00 -1.83011353e-01 6.63121998e-01
-5.63193321e-01 -6.03528261e-01 -3.45264405e-01 4.70923156e-01
1.65625848e-02 9.06355083e-01 9.68263894e-02 -1.75749332e-01
3.83903444e-01 6.30241275e-01 5.35093963e-01 -1.04464936e+00
-7.84396946e-01 -5.90083539e-01 6.94515228e-01 -3.70919198e-01
7.94182941e-02 -8.66275370e-01 -1.20543528e+00 -4.11693811e-01
-2.39724383e-01 1.66450888e-01 7.71510363e-01 1.05658484e+00
4.76663798e-01 5.22172451e-01 2.62420356e-01 -5.66820323e-01
-8.35526824e-01 -1.43644643e+00 8.86383578e-02 7.69993544e-01
1.88914105e-01 -1.25396550e-01 -2.15558559e-01 2.68429101e-01] | [11.18777847290039, 10.275540351867676] |
ed03b7db-b927-4e7b-b577-b2d3d8e80dbb | scientific-evidence-extraction | 2110.00061 | null | https://arxiv.org/abs/2110.00061v3 | https://arxiv.org/pdf/2110.00061v3.pdf | PubTables-1M: Towards comprehensive table extraction from unstructured documents | Recently, significant progress has been made applying machine learning to the problem of table structure inference and extraction from unstructured documents. However, one of the greatest challenges remains the creation of datasets with complete, unambiguous ground truth at scale. To address this, we develop a new, more comprehensive dataset for table extraction, called PubTables-1M. PubTables-1M contains nearly one million tables from scientific articles, supports multiple input modalities, and contains detailed header and location information for table structures, making it useful for a wide variety of modeling approaches. It also addresses a significant source of ground truth inconsistency observed in prior datasets called oversegmentation, using a novel canonicalization procedure. We demonstrate that these improvements lead to a significant increase in training performance and a more reliable estimate of model performance at evaluation for table structure recognition. Further, we show that transformer-based object detection models trained on PubTables-1M produce excellent results for all three tasks of detection, structure recognition, and functional analysis without the need for any special customization for these tasks. Data and code will be released at https://github.com/microsoft/table-transformer. | ['Robin Abraham', 'Rohith Pesala', 'Brandon Smock'] | 2021-09-30 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Smock_PubTables-1M_Towards_Comprehensive_Table_Extraction_From_Unstructured_Documents_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Smock_PubTables-1M_Towards_Comprehensive_Table_Extraction_From_Unstructured_Documents_CVPR_2022_paper.pdf | cvpr-2022-1 | ['table-recognition', 'table-detection', 'table-extraction'] | ['computer-vision', 'miscellaneous', 'miscellaneous'] | [ 3.29321951e-01 8.82260427e-02 -4.67164069e-01 -3.10491621e-01
-1.48590374e+00 -9.53989625e-01 4.46738094e-01 7.43184626e-01
1.28168643e-01 7.22436428e-01 1.12098619e-01 -3.15461785e-01
1.12900540e-01 -8.29914451e-01 -9.41021621e-01 -2.21230149e-01
2.15904370e-01 6.42949700e-01 2.07582369e-01 1.78959534e-01
3.24803203e-01 4.23798054e-01 -1.34933639e+00 8.02335799e-01
5.60794115e-01 1.27070796e+00 4.37706895e-02 3.75870526e-01
-2.60032654e-01 6.70152485e-01 -5.10673940e-01 -6.51039302e-01
7.52957538e-02 -1.28378764e-01 -8.73969972e-01 8.57858583e-02
7.43229032e-01 -2.34425157e-01 5.86794689e-02 7.97520876e-01
1.42969489e-01 -3.88905138e-01 6.89341128e-01 -1.05013406e+00
-4.08174455e-01 8.80168021e-01 -6.58523440e-01 8.72889310e-02
4.92841303e-01 -1.83716476e-01 1.57039785e+00 -1.04939473e+00
7.95208693e-01 9.92347479e-01 6.82894051e-01 7.53421113e-02
-1.37397718e+00 -6.06806219e-01 -2.43014339e-02 4.43077199e-02
-1.45198154e+00 -6.93055332e-01 3.62871349e-01 -4.81431603e-01
9.91436660e-01 4.62709904e-01 3.02417010e-01 7.03853309e-01
1.47088274e-01 9.21795666e-01 7.44541109e-01 -4.66660976e-01
8.14307183e-02 3.18753749e-01 1.35007963e-01 7.88933635e-01
9.23240900e-01 -7.98294246e-01 -7.07524717e-01 -2.70105571e-01
5.11251926e-01 -1.89144313e-01 -8.31827372e-02 -4.26458985e-01
-1.33882260e+00 3.74364197e-01 2.22141240e-02 1.78555250e-01
-7.07934275e-02 -2.36877292e-01 4.29557234e-01 -3.57701294e-02
2.05931827e-01 7.47965276e-01 -6.61100447e-01 -1.17408656e-01
-8.03750753e-01 3.54679555e-01 9.18087780e-01 1.33662772e+00
5.58736920e-01 -2.85418749e-01 4.79919603e-03 6.85795486e-01
9.17126760e-02 3.32338542e-01 -5.04867584e-02 -8.98105264e-01
9.99331892e-01 1.02684510e+00 6.63359463e-02 -1.12311447e+00
-4.14164126e-01 -4.19593751e-01 -6.33460641e-01 -2.98370749e-01
6.95626557e-01 4.27594185e-01 -7.01139212e-01 1.36536348e+00
4.11201358e-01 -6.12741888e-01 -1.72175199e-01 2.06478491e-01
1.03000200e+00 3.58043104e-01 -1.91866338e-01 4.42954302e-02
1.65141094e+00 -4.01566327e-01 -7.75094151e-01 -1.49872124e-01
8.50072324e-01 -1.00409198e+00 1.06852484e+00 6.86285734e-01
-1.11159217e+00 -2.02846572e-01 -1.20249915e+00 -4.14900839e-01
-7.56417990e-01 3.33159000e-01 7.95645535e-01 6.60785317e-01
-5.31300604e-01 3.89351368e-01 -7.91198850e-01 -2.94619203e-01
7.46549070e-01 2.66628236e-01 -6.11016095e-01 -1.78102136e-01
-6.43153489e-01 7.71248579e-01 2.79417485e-01 -8.65676105e-02
-5.29071093e-01 -7.71481514e-01 -8.74643624e-01 2.00550780e-01
7.96486676e-01 -3.91576082e-01 1.26000631e+00 8.01821649e-02
-6.19304121e-01 1.12850726e+00 -2.68686980e-01 -2.46554852e-01
1.35393530e-01 -1.08558193e-01 -1.52869657e-01 -1.50226757e-01
4.11958426e-01 3.50815922e-01 2.77052522e-01 -1.14816725e+00
-6.64833963e-01 -4.91068035e-01 -2.86495566e-01 -1.69820502e-01
-3.43189210e-01 -1.39094726e-03 -8.36878955e-01 -7.12467551e-01
3.20680231e-01 -7.08713770e-01 2.18112797e-01 -1.12169117e-01
-8.50490987e-01 -1.78409144e-01 3.57221633e-01 -8.03111434e-01
1.37699950e+00 -1.85696065e+00 -1.45090044e-01 2.50825822e-01
5.37237644e-01 -5.96987009e-02 9.04146433e-02 6.30868912e-01
1.02405414e-01 5.50079584e-01 -3.72329772e-01 -3.14621687e-01
1.67044513e-02 -3.30831781e-02 -3.64624143e-01 2.88143694e-01
2.87637144e-01 7.89850652e-01 -5.04149616e-01 -7.80096531e-01
-1.25492424e-01 8.43773857e-02 -6.10389590e-01 -7.33213648e-02
-4.36275929e-01 3.97848599e-02 -4.77230489e-01 1.22095954e+00
6.55608594e-01 -7.42421389e-01 5.78523278e-01 -4.16788787e-01
-4.41028252e-02 6.94986999e-01 -1.33437979e+00 1.48116267e+00
9.27958917e-03 5.39541483e-01 1.30026132e-01 -7.34107435e-01
7.38050640e-01 3.17829326e-02 6.16498053e-01 -5.07827938e-01
6.52327016e-02 2.17424795e-01 -1.46604225e-01 -1.87235653e-01
5.14414847e-01 4.31272447e-01 -2.47090593e-01 3.97046596e-01
-6.63834251e-03 -1.78242609e-01 7.04459846e-01 4.67604160e-01
1.22513711e+00 -3.73055786e-02 4.24535573e-01 -2.23237723e-01
3.69604617e-01 3.96464020e-01 7.46623874e-01 7.34865367e-01
2.68956631e-01 6.83851779e-01 6.98409617e-01 -2.82502204e-01
-1.03136325e+00 -1.06115973e+00 -4.25073713e-01 7.70148337e-01
-1.95011914e-01 -1.06066084e+00 -6.81797683e-01 -5.21518767e-01
3.92287850e-01 5.79313338e-01 -5.73693395e-01 2.78988481e-01
-5.54278851e-01 -9.17773962e-01 5.02360225e-01 7.30510771e-01
3.11931968e-01 -6.08288467e-01 -1.16865270e-01 9.97894704e-02
-5.73345542e-01 -1.46630490e+00 -4.21882510e-01 2.96113759e-01
-8.38160694e-01 -1.46925104e+00 -3.82901868e-03 -6.28802657e-01
5.39308667e-01 6.51933625e-02 1.41049564e+00 1.03424124e-01
-6.51734650e-01 2.75891542e-01 -2.36431528e-02 -7.28300154e-01
-3.65676820e-01 3.50799501e-01 -2.69846231e-01 -3.54524463e-01
3.57765436e-01 -2.26203203e-01 -1.21249184e-01 4.49549556e-01
-8.12896907e-01 7.60716945e-02 5.64966083e-01 5.72588801e-01
1.01495385e+00 1.14675589e-01 3.47310007e-01 -1.17723715e+00
3.78715098e-01 -2.24667892e-01 -8.99973750e-01 5.23964047e-01
-9.31237400e-01 2.21560270e-01 4.62322325e-01 8.23897794e-02
-7.89259017e-01 1.08883128e-01 -2.29888018e-02 2.26183593e-01
1.02304302e-01 6.79881215e-01 -7.30436623e-01 2.57024795e-01
5.77173471e-01 1.20741405e-01 -9.56536829e-02 -8.22437525e-01
2.64874905e-01 6.41449630e-01 6.85958683e-01 -7.67858148e-01
6.94555581e-01 4.94079143e-01 1.94403648e-01 -7.34264076e-01
-1.02234781e+00 -4.63228106e-01 -6.35425508e-01 2.52066910e-01
5.31375349e-01 -8.27357531e-01 -7.35871673e-01 2.72791088e-01
-8.03041101e-01 -1.31909132e-01 -1.67774130e-02 1.22725368e-01
-2.79193193e-01 3.83098125e-01 -6.64070308e-01 -4.12643820e-01
-3.67157012e-01 -9.79246676e-01 1.09197748e+00 -2.53827333e-01
-5.40214181e-01 -6.79655552e-01 -2.31244490e-01 7.17441857e-01
-1.06471023e-02 3.54402214e-01 1.35884464e+00 -7.71853805e-01
-1.12247109e+00 -4.45756197e-01 -3.22525442e-01 5.12696691e-02
3.59041810e-01 2.05817744e-01 -7.36494482e-01 -3.42788845e-02
-2.99536049e-01 -4.04016048e-01 8.25868428e-01 1.21117964e-01
1.32031429e+00 -2.45423406e-01 -6.20939970e-01 5.42079210e-01
1.25174236e+00 1.21962041e-01 5.30139327e-01 5.64197123e-01
7.09534287e-01 5.51451802e-01 6.30798578e-01 4.67914373e-01
5.81626117e-01 6.20391071e-01 2.62042433e-01 -1.08362222e-02
-7.70602077e-02 -3.64697427e-01 -2.55994588e-01 5.11828840e-01
2.75021911e-01 -5.04255950e-01 -1.13665020e+00 5.54245651e-01
-1.59620118e+00 -7.61083007e-01 -2.42209703e-01 2.20535350e+00
9.82397497e-01 4.31697339e-01 1.01590224e-01 4.74324822e-01
4.65821892e-01 -1.79716200e-01 -5.54765344e-01 1.15522787e-01
-2.61440426e-01 3.77081670e-02 4.89916623e-01 1.74056247e-01
-1.22771096e+00 6.42888069e-01 6.88425636e+00 7.52409995e-01
-6.73700035e-01 -3.57574672e-01 8.22714329e-01 -1.52116232e-02
-3.23125631e-01 -8.11624378e-02 -1.30802584e+00 2.14329496e-01
7.65131891e-01 -1.92981452e-01 1.50848225e-01 7.83327043e-01
2.10322347e-02 -3.15876514e-01 -1.44887948e+00 1.03636491e+00
-7.45320395e-02 -1.78399968e+00 2.33249798e-01 2.34307870e-01
3.61246973e-01 -1.42232776e-01 1.63834523e-02 -3.92450169e-02
7.74066523e-02 -9.01008725e-01 7.38993227e-01 2.02529445e-01
9.90567207e-01 -5.13073266e-01 5.67208230e-01 9.16927457e-02
-1.28921568e+00 2.17977643e-01 -1.72613919e-01 2.43024677e-01
-2.53439844e-01 7.69638300e-01 -1.06623518e+00 3.99916738e-01
7.16927528e-01 6.67798996e-01 -9.12708521e-01 7.77553141e-01
-3.12748477e-02 6.84450865e-01 -4.98403996e-01 2.43349969e-02
-4.67057884e-01 9.69322994e-02 2.39791989e-01 1.23830509e+00
1.91699173e-02 -1.33948669e-01 7.47439489e-02 8.18616092e-01
-5.15697420e-01 1.53470978e-01 -4.90065217e-01 -4.12815332e-01
6.77946746e-01 1.34017897e+00 -1.14786339e+00 -5.09060860e-01
-5.76674044e-01 2.90047437e-01 3.94743830e-01 5.46413753e-03
-5.81798136e-01 -4.20481265e-01 5.33017099e-01 4.76764530e-01
4.00608242e-01 -1.88417688e-01 -9.05512750e-01 -1.25895727e+00
4.11633551e-01 -1.23320901e+00 7.84019291e-01 -5.99109232e-01
-1.04120386e+00 2.87646115e-01 1.11563049e-01 -9.12885964e-01
-1.63780242e-01 -9.43673015e-01 4.26062010e-02 6.04456306e-01
-9.61321056e-01 -1.05753434e+00 -1.64964139e-01 1.93360522e-01
2.92601347e-01 -1.08259216e-01 8.13191712e-01 3.47336471e-01
-8.08883071e-01 7.49785841e-01 1.60583109e-01 4.19366658e-01
8.18637967e-01 -1.39973652e+00 6.42868102e-01 7.71404624e-01
2.98685014e-01 9.47874069e-01 5.90133786e-01 -7.95877099e-01
-1.84074330e+00 -1.00009024e+00 1.03731513e+00 -1.15930569e+00
5.61996341e-01 -7.77087390e-01 -9.30671513e-01 9.50046539e-01
-1.89593166e-01 -3.57401103e-01 8.91324162e-01 4.97596473e-01
-6.65989816e-01 -3.03511590e-01 -1.20020819e+00 3.40297610e-01
9.73176599e-01 -4.89447147e-01 -3.78589183e-01 4.39571083e-01
3.84294122e-01 -6.20736599e-01 -1.20043063e+00 3.77739191e-01
7.36399889e-01 -5.99865675e-01 8.74154925e-01 -3.85421276e-01
5.86828768e-01 -3.52430403e-01 -3.53806496e-01 -4.25293267e-01
-2.54950464e-01 -3.85034919e-01 -3.65548521e-01 1.43931592e+00
9.41460848e-01 -3.74581516e-01 1.06451809e+00 9.54348564e-01
-1.57479525e-01 -8.46406758e-01 -5.46646893e-01 -6.68947697e-01
-7.59559870e-02 -3.96667808e-01 6.32322371e-01 7.45377839e-01
1.16184056e-01 4.95008618e-01 -3.25804614e-02 -1.82834303e-03
6.96229756e-01 5.42600811e-01 1.03321946e+00 -1.15959644e+00
-1.78656772e-01 -3.28404635e-01 -1.85190186e-01 -1.03083456e+00
-3.96529585e-01 -1.13239825e+00 -1.25624016e-01 -1.79763234e+00
5.73125899e-01 -1.91174537e-01 -1.03900678e-01 7.60297120e-01
-6.40434101e-02 1.91645548e-01 -5.48347160e-02 3.81631553e-01
-7.55681872e-01 -3.16525367e-03 9.87105608e-01 -1.95097700e-01
1.18350387e-01 -1.36597171e-01 -1.18657970e+00 4.79884744e-01
7.54107296e-01 -5.79354346e-01 -1.50354639e-01 -2.93490797e-01
4.10248727e-01 -1.54390320e-01 3.51778418e-03 -9.05049860e-01
4.63380516e-02 -1.08898923e-01 7.90656090e-01 -1.23527539e+00
2.17807978e-01 -5.54146051e-01 6.90501407e-02 1.42604440e-01
-3.00518632e-01 2.89925337e-01 5.48528552e-01 4.24420953e-01
-1.35205060e-01 -2.35762969e-02 4.64743197e-01 -2.59063393e-01
-3.71794134e-01 9.91827548e-02 -4.05140966e-01 2.37718567e-01
5.10182381e-01 -1.07061408e-01 -5.23905396e-01 -3.39940526e-02
-4.60218042e-01 2.35074803e-01 5.62115252e-01 2.19797820e-01
3.05223912e-01 -9.60023820e-01 -5.71715117e-01 1.58858761e-01
4.16918397e-01 1.52213261e-01 -2.42513835e-01 5.82826793e-01
-4.13989246e-01 7.67945647e-01 6.33807778e-02 -6.00100815e-01
-1.38247001e+00 5.31495452e-01 -7.84438476e-02 -5.32360137e-01
-4.30151939e-01 6.53388143e-01 1.05951615e-01 -3.04590732e-01
2.95487702e-01 -8.03314924e-01 9.31506976e-02 5.09183668e-02
5.98627508e-01 3.05402786e-01 6.15557671e-01 -2.08176941e-01
-5.61577797e-01 2.41369143e-01 -4.38255250e-01 1.48377240e-01
1.34212995e+00 1.24726566e-02 -2.48077676e-01 4.01916295e-01
8.88270676e-01 4.22941357e-01 -7.11253941e-01 -1.39934614e-01
3.84302258e-01 -3.09976280e-01 -2.83234775e-01 -1.01516771e+00
-6.77611530e-01 4.90715116e-01 7.37052709e-02 2.41758093e-01
9.33821321e-01 1.53715596e-01 6.23210967e-01 7.35309124e-01
3.43711883e-01 -8.72767210e-01 -5.26114851e-02 2.94222623e-01
8.74849796e-01 -1.30295157e+00 5.59393466e-01 -8.93091679e-01
-1.47984490e-01 9.28026259e-01 5.97958207e-01 6.22166038e-01
4.28025275e-01 6.76632226e-01 -9.61380526e-02 -3.54327887e-01
-8.68838727e-01 3.81582491e-02 3.99279714e-01 4.57090676e-01
7.57098436e-01 -9.77717414e-02 -9.26047340e-02 7.15409875e-01
-2.71659225e-01 -1.01092465e-01 4.78064090e-01 1.32784629e+00
-3.92797917e-01 -1.36265981e+00 -4.73327190e-01 9.47479904e-01
-7.35487163e-01 -1.59311011e-01 -7.00068891e-01 9.33800817e-01
-1.99412271e-01 9.16014493e-01 -1.46252960e-01 -2.46722564e-01
3.70002389e-01 1.27486184e-01 6.22429609e-01 -7.93593585e-01
-2.41493821e-01 9.89291370e-02 4.90301162e-01 -4.39551115e-01
8.02970901e-02 -9.02748346e-01 -1.24589813e+00 -4.38908786e-01
-3.53292912e-01 1.76718101e-01 5.95098317e-01 6.47822142e-01
6.73463404e-01 4.34177518e-01 6.30058488e-03 -1.90401986e-01
-4.11879361e-01 -8.29686403e-01 -4.85780507e-01 2.90265948e-01
6.72098175e-02 -6.05194509e-01 -7.00094774e-02 3.84077996e-01] | [9.634026527404785, 7.8473219871521] |
bc0d5172-986e-47a1-af7c-689e61f19ec2 | region-based-contrastive-pretraining-for | 2305.05598 | null | https://arxiv.org/abs/2305.05598v1 | https://arxiv.org/pdf/2305.05598v1.pdf | Region-based Contrastive Pretraining for Medical Image Retrieval with Anatomic Query | We introduce a novel Region-based contrastive pretraining for Medical Image Retrieval (RegionMIR) that demonstrates the feasibility of medical image retrieval with similar anatomical regions. RegionMIR addresses two major challenges for medical image retrieval i) standardization of clinically relevant searching criteria (e.g., anatomical, pathology-based), and ii) localization of anatomical area of interests that are semantically meaningful. In this work, we propose an ROI image retrieval image network that retrieves images with similar anatomy by extracting anatomical features (via bounding boxes) and evaluate similarity between pairwise anatomy-categorized features between the query and the database of images using contrastive learning. ROI queries are encoded using a contrastive-pretrained encoder that was fine-tuned for anatomy classification, which generates an anatomical-specific latent space for region-correlated image retrieval. During retrieval, we compare the anatomically encoded query to find similar features within a feature database generated from training samples, and retrieve images with similar regions from training samples. We evaluate our approach on both anatomy classification and image retrieval tasks using the Chest ImaGenome Dataset. Our proposed strategy yields an improvement over state-of-the-art pretraining and co-training strategies, from 92.24 to 94.12 (2.03%) classification accuracy in anatomies. We qualitatively evaluate the image retrieval performance demonstrating generalizability across multiple anatomies with different morphology. | ['Ivan Tarapov', 'Javier Alvarez-Valle', 'Fernando Pérez-García', 'Ozan Oktay', 'Jameson Merkow', 'Alberto Santamaria-Pang', 'Ho Hin Lee'] | 2023-05-09 | null | null | null | null | ['medical-image-retrieval', 'medical-image-retrieval', 'anatomy'] | ['computer-vision', 'medical', 'miscellaneous'] | [ 3.88362348e-01 -1.68492392e-01 -6.65030003e-01 -2.95155764e-01
-1.73952830e+00 -7.76886642e-01 2.69866288e-01 5.78451335e-01
-7.34261274e-01 4.88978207e-01 3.68241191e-01 -3.77257243e-02
-7.81296790e-01 -5.71004391e-01 -4.80609268e-01 -7.73244560e-01
-2.11070657e-01 4.88653183e-01 -1.19500071e-01 1.42327949e-01
1.86440527e-01 9.95633781e-01 -1.11682212e+00 5.29523432e-01
5.19481957e-01 1.43093610e+00 3.83509129e-01 6.15379512e-01
1.67008802e-01 5.85947394e-01 -7.45960832e-01 2.33279299e-02
3.06829542e-01 -3.60934228e-01 -1.32142019e+00 -3.52582097e-01
9.48867917e-01 -3.96782190e-01 -5.31282187e-01 7.96425402e-01
8.03722024e-01 1.74768209e-01 1.18336737e+00 -5.43518960e-01
-1.04647076e+00 1.80952385e-01 -4.46427703e-01 8.66845191e-01
7.96631649e-02 -2.33544871e-01 1.02836263e+00 -8.80589128e-01
9.79789674e-01 8.69031191e-01 3.01637173e-01 6.17084622e-01
-1.18258739e+00 -7.36539125e-01 -4.52666730e-01 1.14079557e-01
-1.84726453e+00 -1.49644375e-01 6.78870797e-01 -3.73705149e-01
7.73077905e-01 6.83861971e-01 6.13307595e-01 6.05634391e-01
7.06723630e-01 7.01968551e-01 8.77746463e-01 -3.05869818e-01
-2.58154690e-01 -6.33457452e-02 1.57599375e-02 8.58993411e-01
-6.45737275e-02 1.28150091e-01 -3.90084296e-01 -3.77339154e-01
1.01925588e+00 2.55776346e-01 -4.36228991e-01 -3.22956473e-01
-1.46604240e+00 8.46993744e-01 1.35106230e+00 6.51763618e-01
-5.67004740e-01 3.43140870e-01 4.66071814e-01 1.87462375e-01
2.72698224e-01 8.39131951e-01 -6.59919679e-02 5.86130679e-01
-1.19939661e+00 -1.44829988e-01 2.40139395e-01 6.46539032e-01
3.58261287e-01 -3.41057748e-01 -5.63424170e-01 1.09650838e+00
3.20801049e-01 5.71243346e-01 1.06125343e+00 -6.35456026e-01
2.48616636e-02 3.97417247e-01 -6.06107533e-01 -1.09257817e+00
-2.23636523e-01 -5.98097324e-01 -8.64819765e-01 -2.86884159e-01
1.35976136e-01 5.60565412e-01 -1.13664448e+00 1.54422343e+00
-3.44444476e-02 -2.85397440e-01 1.33746058e-01 1.12433469e+00
1.40724957e+00 2.46359020e-01 4.01836872e-01 -2.07097214e-02
1.48695183e+00 -1.00766039e+00 -5.23629308e-01 1.45656601e-01
6.43161476e-01 -8.10097635e-01 1.10194349e+00 -1.96728587e-01
-1.11508095e+00 -5.75169444e-01 -8.57452750e-01 -2.44224712e-01
-4.09182727e-01 5.53518713e-01 5.83377182e-01 2.05578461e-01
-1.25522053e+00 4.56591696e-01 -6.06620729e-01 -2.39757240e-01
5.95606744e-01 6.44895494e-01 -7.29098976e-01 1.12288231e-02
-1.04766977e+00 8.03074718e-01 3.94049197e-01 -3.90568636e-02
-1.19607258e+00 -1.12033284e+00 -7.93067992e-01 1.50607258e-01
-4.40984825e-03 -6.79885149e-01 8.32447827e-01 -8.33492160e-01
-8.30304921e-01 1.63756847e+00 2.40829587e-02 -2.93741316e-01
1.37888089e-01 3.20080258e-02 -4.00367498e-01 1.14299572e+00
5.09023726e-01 1.15952349e+00 7.28027582e-01 -1.16457915e+00
5.98222800e-02 -4.02167052e-01 -2.38777101e-01 4.15759742e-01
-2.97453701e-01 2.92158127e-02 -6.37876391e-01 -1.03537464e+00
1.52498513e-01 -9.02778327e-01 -2.27043480e-01 7.03133404e-01
-2.25316852e-01 3.65274772e-02 3.10356170e-01 -7.35185921e-01
1.15443492e+00 -2.18125916e+00 1.54471755e-01 6.73533559e-01
5.48108697e-01 -8.61274078e-02 -3.83105218e-01 1.12275846e-01
-4.37775671e-01 4.01403219e-01 -7.14912638e-02 6.70662969e-02
-5.92194796e-01 8.76190811e-02 -1.82151929e-01 6.24493599e-01
1.51482701e-01 1.40817130e+00 -9.30703998e-01 -1.29211497e+00
1.50342032e-01 3.41524512e-01 -6.49849951e-01 3.79498780e-01
5.37238955e-01 6.81018829e-01 -6.14993751e-01 1.01604939e+00
1.68902785e-01 -7.26755083e-01 8.19555670e-03 -7.25103319e-01
3.76433074e-01 -1.22070670e-01 -2.25530460e-01 2.16149235e+00
-6.15644395e-01 4.36784744e-01 -1.81141689e-01 -8.34349632e-01
6.95144713e-01 3.38122636e-01 1.07582140e+00 -1.01660812e+00
1.13148682e-01 4.83978480e-01 5.24844192e-02 -4.94935453e-01
2.44538754e-01 -9.01000500e-02 -2.44764417e-01 3.82946759e-01
4.85019267e-01 -7.32847378e-02 -5.05711138e-02 2.00839475e-01
1.03010762e+00 -4.12294596e-01 4.58497465e-01 -5.13957202e-01
7.55265594e-01 1.25067756e-01 -1.38249442e-01 9.62465346e-01
-3.77844125e-01 9.08355892e-01 -7.81197622e-02 -3.90987545e-01
-7.23382533e-01 -1.46868789e+00 -6.68142796e-01 9.10867274e-01
3.17850739e-01 -2.74735421e-01 -2.65161186e-01 -7.37899542e-01
4.17792313e-02 -1.16849460e-01 -1.12334085e+00 -6.30986989e-01
-6.43229127e-01 -4.18875426e-01 7.59377241e-01 5.81978321e-01
3.09660882e-01 -1.30289352e+00 -7.88958192e-01 -1.07138857e-01
-2.25674719e-01 -7.16003060e-01 -8.95750284e-01 1.68514669e-01
-9.58912134e-01 -1.15175498e+00 -1.37891543e+00 -1.16963124e+00
1.11941814e+00 1.49265632e-01 1.31503952e+00 6.00382626e-01
-1.17610228e+00 8.86653721e-01 -2.71433413e-01 1.24003597e-01
-2.70958066e-01 2.11902052e-01 -3.81611824e-01 -3.96924615e-01
-2.69233465e-01 -7.36547261e-02 -1.20325899e+00 2.24449396e-01
-1.19689000e+00 -3.43135625e-01 1.05123198e+00 1.41233420e+00
1.28377926e+00 -5.84022939e-01 2.85006136e-01 -5.99583149e-01
6.61738455e-01 -3.87249708e-01 -1.27974302e-01 8.35634291e-01
-4.78648663e-01 7.65352324e-02 2.20720798e-01 -5.33817410e-01
-4.69642311e-01 -1.32515982e-01 -1.02066360e-02 -7.76912093e-01
1.21741727e-01 7.25879908e-01 6.38712525e-01 -5.71600080e-01
8.18351865e-01 4.52854306e-01 1.81738511e-01 -8.86654630e-02
3.71875733e-01 3.98743182e-01 7.39276469e-01 -7.13958323e-01
4.98365432e-01 5.50610244e-01 1.63620457e-01 -3.92773181e-01
-8.36049736e-01 -9.01591897e-01 -8.43861639e-01 -2.12998152e-01
1.04157984e+00 -8.05468678e-01 -5.16041994e-01 -2.85599291e-01
-7.28581965e-01 -2.93211658e-02 -5.64945638e-01 6.07566476e-01
-4.85352546e-01 3.08622301e-01 -7.87609100e-01 -1.43458977e-01
-1.17155838e+00 -1.53209209e+00 1.63490391e+00 5.43940216e-02
-3.66956741e-01 -9.72068429e-01 1.42050758e-01 1.75248429e-01
6.32647634e-01 1.82071775e-01 1.20450461e+00 -8.75924289e-01
-4.72298205e-01 -3.34867686e-01 -3.96421254e-01 4.62450944e-02
3.83535981e-01 -3.63268942e-01 -6.97465360e-01 -5.81925333e-01
-4.14660722e-01 -6.11202240e-01 8.34863901e-01 5.13842583e-01
1.86390913e+00 -2.11560503e-01 -6.37351632e-01 8.70908201e-01
1.51291382e+00 1.78953096e-01 4.80976731e-01 6.27719387e-02
3.73902798e-01 4.00457174e-01 6.25676751e-01 -9.68587175e-02
-8.68298262e-02 6.44741356e-01 1.54594883e-01 -6.62759840e-01
-5.23371160e-01 -1.24706678e-01 -4.36052650e-01 8.63009095e-01
1.43285468e-01 1.02565341e-01 -9.58721817e-01 7.26144135e-01
-1.16628540e+00 -5.83257318e-01 6.22223973e-01 2.02004457e+00
1.15998149e+00 -5.51187634e-01 -3.10103178e-01 -4.50546384e-01
3.56123507e-01 1.24990754e-01 -5.07196963e-01 8.24737698e-02
1.39311418e-01 9.70312238e-01 5.18969715e-01 3.15495014e-01
-1.32867467e+00 7.77554870e-01 6.80125523e+00 1.12467313e+00
-1.41787875e+00 2.76084542e-01 9.09947217e-01 -5.38154021e-02
-2.38685191e-01 -4.52430993e-01 -3.03002626e-01 -7.58151291e-03
6.00771248e-01 -2.75918365e-01 2.21638363e-02 7.64297903e-01
-4.30563718e-01 7.59264082e-02 -1.13286150e+00 1.20446122e+00
5.37767589e-01 -1.60106897e+00 5.62381983e-01 -1.20678563e-02
5.09937286e-01 -3.43930237e-02 5.49203694e-01 2.29611427e-01
-4.46748793e-01 -1.40382028e+00 3.46286178e-01 7.18341172e-01
1.44473708e+00 -2.63717473e-01 6.88862562e-01 -4.43660289e-01
-1.28250325e+00 1.77259073e-01 -3.08804691e-01 1.09700954e+00
-4.48276490e-01 -1.79106355e-01 -9.19168711e-01 6.03533506e-01
8.78935814e-01 6.85101628e-01 -9.25462663e-01 1.17362809e+00
1.70312375e-01 2.17442214e-01 -1.04697786e-01 1.40637338e-01
2.20482007e-01 2.11595446e-01 3.40874314e-01 1.23518026e+00
4.60776761e-02 -4.39969748e-02 2.71649987e-01 1.07345271e+00
-3.21470231e-01 7.71907806e-01 -5.43096304e-01 1.87056568e-02
3.57022703e-01 1.47339070e+00 -8.69400680e-01 -5.59064507e-01
1.83866769e-01 1.01920497e+00 1.87894896e-01 3.57972562e-01
-6.70799613e-01 -2.88775176e-01 7.76038542e-02 -4.37483117e-02
-1.05045825e-01 2.62557447e-01 1.39297232e-01 -8.72029066e-01
-1.92014605e-01 -7.66908586e-01 8.11802804e-01 -8.19475114e-01
-1.48675382e+00 8.91446471e-01 2.20669925e-01 -1.38866627e+00
-1.51197881e-01 -5.29138267e-01 -1.86235771e-01 9.34569299e-01
-1.62516379e+00 -1.30631673e+00 -5.46052396e-01 8.18704784e-01
2.56802499e-01 -1.92309231e-01 1.37337863e+00 5.36509991e-01
-3.02418359e-02 9.65958178e-01 2.17274390e-02 5.19791007e-01
1.10275137e+00 -1.23850036e+00 -5.44973552e-01 6.56710118e-02
3.06352943e-01 1.20589900e+00 -2.67928928e-01 -4.59095240e-01
-1.07905507e+00 -1.00276172e+00 3.02950084e-01 -2.59784102e-01
4.75307047e-01 9.08710733e-02 -7.60818839e-01 3.64983916e-01
1.15689583e-01 6.21579766e-01 1.22210622e+00 -3.88387501e-01
-4.30388004e-01 -1.73890948e-01 -1.31495762e+00 5.78205049e-01
7.04865158e-01 -9.66039836e-01 -5.85922241e-01 5.88188469e-01
5.47095299e-01 -5.86845934e-01 -1.71995318e+00 7.42516577e-01
8.86378407e-01 -8.11588392e-02 1.49735630e+00 -8.20827603e-01
4.27701324e-01 8.89284387e-02 -2.12291032e-01 -8.50980401e-01
-2.82207191e-01 -3.85999121e-02 3.98333758e-01 4.37093019e-01
3.70469362e-01 -3.86980563e-01 5.74277520e-01 1.93306878e-01
-1.20625064e-01 -1.10037434e+00 -9.57942367e-01 -4.32742119e-01
5.11604190e-01 2.97560781e-01 7.19441995e-02 8.65154207e-01
-2.20499814e-01 -4.36102971e-02 1.69189170e-01 -1.30034640e-01
4.44352120e-01 5.54243922e-01 2.73012429e-01 -8.69362950e-01
-2.84478486e-01 -7.37266898e-01 -5.38470626e-01 -8.99084210e-01
2.91665852e-01 -1.76385081e+00 2.29014121e-02 -1.48311281e+00
6.26069903e-01 -5.84438860e-01 -9.23742056e-01 6.84470773e-01
-1.29563242e-01 7.65511572e-01 5.42445481e-03 7.22903252e-01
-6.59788251e-01 2.59736538e-01 1.63288832e+00 -5.41845322e-01
2.21503809e-01 -4.63707626e-01 -5.71047485e-01 4.17627357e-02
6.55112505e-01 -6.49822772e-01 -3.14915568e-01 -1.37739480e-01
-3.07837486e-01 2.51696020e-01 6.75606966e-01 -1.04609954e+00
1.14403449e-01 3.30225736e-01 8.72865677e-01 -6.01493001e-01
3.27859133e-01 -8.77534270e-01 -1.12426661e-01 8.35797310e-01
-9.62816656e-01 3.38531345e-01 3.75646651e-01 3.35159451e-01
-6.88456178e-01 -3.25550973e-01 6.75447643e-01 -3.76046926e-01
-6.38166904e-01 6.33742273e-01 -2.05952302e-01 1.02445893e-01
8.02979171e-01 2.83273514e-02 -1.73401967e-01 -2.32769206e-01
-1.05604148e+00 -1.45160658e-02 4.65956293e-02 2.74535358e-01
1.05953896e+00 -1.38165152e+00 -5.69553435e-01 2.62415074e-02
6.32456839e-01 -3.00848544e-01 4.00750816e-01 1.07867658e+00
-8.39395106e-01 7.47672260e-01 -3.48559111e-01 -1.01346302e+00
-1.47367322e+00 5.16223729e-01 6.70350015e-01 -7.67406106e-01
-4.79146242e-01 9.37379360e-01 4.26529944e-01 -3.53944510e-01
1.62016213e-01 -2.82062501e-01 -2.09254086e-01 -9.03715193e-02
2.25749344e-01 -4.86689717e-01 1.22480132e-01 -8.97010624e-01
-5.81622720e-01 1.09445226e+00 -3.13520640e-01 4.27730754e-02
1.03348410e+00 3.00369591e-01 -3.21640581e-01 7.52145201e-02
1.93941188e+00 1.53771918e-02 -2.38248840e-01 -3.76294702e-01
-2.83685088e-01 -3.78179938e-01 2.22923666e-01 -1.06659901e+00
-1.43323922e+00 8.70941579e-01 1.36494160e+00 -5.57840884e-01
1.23159897e+00 4.66710150e-01 6.23314261e-01 3.81672055e-01
1.83901444e-01 -5.10268748e-01 5.31884253e-01 -9.80917141e-02
1.32822239e+00 -1.28302097e+00 2.80209869e-01 -2.59046972e-01
-5.16799390e-01 1.06562877e+00 4.46428150e-01 -4.09139872e-01
9.33408201e-01 -1.12389311e-01 3.02220583e-01 -8.51896763e-01
-3.04969788e-01 -1.70230016e-01 1.21402884e+00 2.30793417e-01
6.36574388e-01 -8.51390418e-03 -4.75898743e-01 1.71017125e-01
2.13790704e-02 -3.67133260e-01 -4.50200737e-01 9.45322156e-01
1.08625457e-01 -9.54765499e-01 -1.64645702e-01 7.24771082e-01
-9.04010355e-01 -5.05738139e-01 -3.31687987e-01 9.20000732e-01
-2.97786146e-01 2.82780170e-01 8.68991315e-02 -9.90785658e-02
2.46993285e-02 -1.71399832e-01 7.45789349e-01 -7.04240620e-01
-1.01919854e+00 3.09628427e-01 -4.27693546e-01 -6.15597010e-01
-3.78209412e-01 -1.86410397e-01 -1.24490070e+00 6.72271192e-01
-4.24782842e-01 3.16300660e-01 4.63428587e-01 5.42106271e-01
3.12920004e-01 7.74422169e-01 3.65222484e-01 -5.49002469e-01
-3.09395999e-01 -8.32080126e-01 -2.27741003e-01 7.76698947e-01
2.81487137e-01 -7.19681978e-01 2.90239006e-02 -1.09613135e-01] | [14.418212890625, -1.6064180135726929] |
be2a98e5-c762-4e7c-a5b1-0d2306b4f1a3 | crossatnet-a-novel-cross-attention-based | 2104.09918 | null | https://arxiv.org/abs/2104.09918v1 | https://arxiv.org/pdf/2104.09918v1.pdf | CrossATNet - A Novel Cross-Attention Based Framework for Sketch-Based Image Retrieval | We propose a novel framework for cross-modal zero-shot learning (ZSL) in the context of sketch-based image retrieval (SBIR). Conventionally, the SBIR schema mainly considers simultaneous mappings among the two image views and the semantic side information. Therefore, it is desirable to consider fine-grained classes mainly in the sketch domain using highly discriminative and semantically rich feature space. However, the existing deep generative modeling-based SBIR approaches majorly focus on bridging the gaps between the seen and unseen classes by generating pseudo-unseen-class samples. Besides, violating the ZSL protocol by not utilizing any unseen-class information during training, such techniques do not pay explicit attention to modeling the discriminative nature of the shared space. Also, we note that learning a unified feature space for both the multi-view visual data is a tedious task considering the significant domain difference between sketches and color images. In this respect, as a remedy, we introduce a novel framework for zero-shot SBIR. While we define a cross-modal triplet loss to ensure the discriminative nature of the shared space, an innovative cross-modal attention learning strategy is also proposed to guide feature extraction from the image domain exploiting information from the respective sketch counterpart. In order to preserve the semantic consistency of the shared space, we consider a graph CNN-based module that propagates the semantic class topology to the shared space. To ensure an improved response time during inference, we further explore the possibility of representing the shared space in terms of hash codes. Experimental results obtained on the benchmark TU-Berlin and the Sketchy datasets confirm the superiority of CrossATNet in yielding state-of-the-art results. | ['Mihai Datcu', 'Avik Bhattacharya', 'Biplab Banerjee', 'Ushasi Chaudhuri'] | 2021-04-20 | null | null | null | null | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 1.67303100e-01 -2.40321115e-01 -2.48864219e-01 -2.96636134e-01
-9.10131276e-01 -5.25539279e-01 6.98215663e-01 -1.70690596e-01
-8.14621374e-02 4.70963359e-01 -6.87435865e-02 1.49843544e-01
-5.86247385e-01 -1.08134651e+00 -7.17531204e-01 -9.29783523e-01
4.79934961e-01 4.56869036e-01 1.29924908e-01 -2.37301141e-01
4.58689686e-03 4.46041703e-01 -1.71117866e+00 4.07575279e-01
6.42961621e-01 1.15565753e+00 2.64916718e-01 1.88487560e-01
-2.56255090e-01 3.11001956e-01 -3.40316743e-01 -6.34496689e-01
3.33991349e-01 -3.62976760e-01 -4.98402387e-01 2.52674609e-01
7.76872396e-01 -3.17739934e-01 -5.10231674e-01 1.13244820e+00
4.31686312e-01 2.80421913e-01 5.84469438e-01 -1.53961825e+00
-9.37220156e-01 2.25486219e-01 -3.64248365e-01 -2.77184308e-01
3.43142092e-01 -4.11980078e-02 1.33945608e+00 -1.06079936e+00
9.75036383e-01 1.16497612e+00 2.10300535e-01 4.74317491e-01
-1.28155708e+00 -6.31618500e-01 4.18097883e-01 4.41104650e-01
-1.74073136e+00 -2.51571536e-01 1.34198439e+00 -3.39248776e-01
2.88895875e-01 3.23192477e-01 5.21917701e-01 1.33319211e+00
-2.38625914e-01 8.87633562e-01 8.32580924e-01 -1.79229036e-01
3.22832733e-01 3.13167632e-01 -6.62047043e-02 7.32453287e-01
-5.88878952e-02 4.51488746e-03 -6.71793699e-01 -1.58785820e-01
8.42404306e-01 5.18021882e-01 -3.76737714e-01 -1.15664065e+00
-1.06929624e+00 8.78402948e-01 6.11047864e-01 4.70783472e-01
-4.30602878e-02 -4.42920886e-02 2.83684433e-01 2.05272391e-01
2.78050125e-01 3.83136980e-02 -6.32383302e-02 3.03880692e-01
-9.82223272e-01 8.83961245e-02 5.25360644e-01 1.09861922e+00
1.02884626e+00 -9.96696725e-02 -2.74461240e-01 8.98027539e-01
3.02392095e-01 3.57448280e-01 2.18263730e-01 -4.61811393e-01
4.77181762e-01 6.38509154e-01 -6.12077378e-02 -1.29707146e+00
2.34068826e-01 -5.42444885e-01 -9.58552301e-01 8.04903135e-02
2.30488822e-01 5.49977064e-01 -9.20553803e-01 1.95258904e+00
3.74215722e-01 1.60544977e-01 -2.08177157e-02 1.08694792e+00
8.03443074e-01 4.91706222e-01 -3.45702432e-02 1.32670864e-01
1.31717634e+00 -7.92863965e-01 -4.99734879e-01 7.85731710e-03
1.91121846e-01 -5.37479937e-01 1.28261781e+00 1.40175149e-01
-8.40155780e-01 -7.04934299e-01 -1.18184590e+00 -2.76227683e-01
-8.17772567e-01 2.23371580e-01 4.26905453e-01 4.69686687e-01
-7.35191703e-01 4.19615328e-01 -5.01986623e-01 -3.94170761e-01
3.42416435e-01 -7.20006134e-03 -5.25630772e-01 -4.94748563e-01
-1.27242827e+00 4.17764306e-01 2.97387630e-01 1.45825729e-01
-9.44320083e-01 -6.95713222e-01 -9.25459564e-01 4.32492167e-01
4.74559188e-01 -6.38385177e-01 3.76074582e-01 -8.96784365e-01
-1.09565675e+00 7.23523915e-01 -1.05643794e-01 5.47890626e-02
6.26523376e-01 1.42908201e-01 -3.65873724e-01 2.40092158e-01
6.39824420e-02 5.18144727e-01 1.09499764e+00 -1.54953480e+00
-3.51658881e-01 -6.18192732e-01 3.43017966e-01 4.35544439e-02
-5.56630254e-01 -5.64950705e-01 -6.72994733e-01 -7.73949265e-01
2.32573524e-01 -8.36323082e-01 1.81639001e-01 3.68177623e-01
-3.07727247e-01 -2.09371403e-01 1.07908893e+00 -4.15630758e-01
1.04154098e+00 -2.36366916e+00 4.60329056e-01 3.34587753e-01
1.04471140e-01 1.26841113e-01 -3.11719418e-01 5.52374125e-01
2.64794696e-02 -1.89813852e-01 -2.83058673e-01 -4.24109042e-01
1.92765728e-01 3.54593068e-01 -5.52513778e-01 4.34857190e-01
2.63834268e-01 9.61582482e-01 -1.02377832e+00 -4.59897131e-01
3.37960184e-01 8.03926170e-01 -4.92498964e-01 2.82083333e-01
-1.45399779e-01 3.04951012e-01 -5.44466078e-01 6.92894399e-01
8.22562397e-01 -3.94338697e-01 2.78678268e-01 -5.76705694e-01
2.91304111e-01 -2.50356972e-01 -1.21142983e+00 2.31325889e+00
-5.92286587e-01 1.61739632e-01 -5.24445735e-02 -1.13060153e+00
1.01181400e+00 1.88080296e-01 4.33911860e-01 -7.82501757e-01
-2.01713100e-01 1.22374453e-01 -4.59683150e-01 -2.44078547e-01
4.72922444e-01 -2.89364964e-01 -8.27226266e-02 3.17710817e-01
3.51069957e-01 2.57849008e-01 -5.34496456e-02 3.14793944e-01
5.02983153e-01 3.64420533e-01 2.70071980e-02 -3.97471711e-02
6.44916356e-01 -3.70852679e-01 4.60321397e-01 5.90998530e-01
8.58905017e-02 8.04519236e-01 3.22063744e-01 -4.65026021e-01
-9.70150232e-01 -1.37175786e+00 -1.24239646e-01 1.07074130e+00
6.16044879e-01 -4.16106194e-01 -3.47181857e-01 -8.55606973e-01
1.87048949e-02 5.40289223e-01 -7.78388500e-01 -3.54523748e-01
-2.59757340e-01 -3.19388360e-01 2.05302104e-01 3.65993381e-01
4.38675016e-01 -7.71422803e-01 -3.40038598e-01 2.73514520e-02
-2.19298765e-01 -1.14019060e+00 -6.40811026e-01 -1.45140067e-01
-4.43774760e-01 -1.09378064e+00 -8.83352518e-01 -6.58783019e-01
7.26522207e-01 5.77139139e-01 9.16802287e-01 9.56674200e-03
-4.83103484e-01 7.29314089e-01 -3.79634112e-01 3.72447729e-01
-4.72674519e-02 -4.78368998e-02 -3.68974179e-01 5.68285108e-01
3.19078654e-01 -6.89080954e-01 -6.99747980e-01 2.83563852e-01
-1.12991834e+00 5.68663850e-02 4.62700456e-01 1.13718164e+00
5.34433126e-01 -8.09355527e-02 6.16048753e-01 -5.52086294e-01
1.84309721e-01 -5.31732142e-01 -5.44498265e-01 7.92216063e-01
-4.70871866e-01 1.80189028e-01 6.44515216e-01 -5.04783213e-01
-1.08710980e+00 3.56188305e-02 1.85341656e-01 -9.67916012e-01
-1.51067257e-01 2.40478113e-01 -6.73508108e-01 1.59397498e-02
1.24131836e-01 5.64036131e-01 -1.78188439e-02 -4.45929408e-01
7.30233371e-01 4.29751068e-01 2.95754313e-01 -7.54926741e-01
9.40400958e-01 8.34901512e-01 7.12451786e-02 -7.11430967e-01
-7.12514222e-01 -5.24256647e-01 -5.43824017e-01 -1.11824214e-01
9.39932287e-01 -8.66727948e-01 -5.86461425e-01 3.12813222e-02
-1.07422149e+00 3.85966778e-01 -3.38323534e-01 1.32123142e-01
-6.20429397e-01 5.66111863e-01 -2.87096083e-01 -7.41905391e-01
-3.01286429e-01 -1.28752124e+00 1.49085915e+00 -3.31333354e-02
1.44893914e-01 -9.71516550e-01 -4.01565433e-03 2.69769818e-01
2.23789260e-01 1.36888281e-01 1.18497121e+00 -5.45355976e-01
-1.02534223e+00 -1.08900547e-01 -4.88827527e-01 1.77079648e-01
2.22611502e-01 -1.81950912e-01 -1.10054719e+00 -5.28364778e-01
-2.76014477e-01 -4.09676254e-01 9.12420630e-01 -1.79350019e-01
1.24231577e+00 -8.19597542e-02 -1.63915098e-01 6.81832373e-01
1.70619726e+00 1.90908220e-02 5.82853913e-01 6.65496886e-02
8.85464430e-01 7.06289351e-01 6.64980352e-01 5.09818673e-01
3.98106724e-01 9.12179947e-01 4.36341047e-01 -7.52836093e-02
-1.99393407e-01 -7.37388313e-01 8.97213295e-02 5.49810231e-01
2.47561067e-01 -2.09192157e-01 -3.80047441e-01 6.24609172e-01
-1.90215135e+00 -1.04789019e+00 4.69817758e-01 2.48502469e+00
5.73645592e-01 -2.43043259e-01 -1.37700930e-01 -5.75829335e-02
8.15414369e-01 4.97424692e-01 -4.08600956e-01 7.82246366e-02
-1.88055426e-01 1.33248597e-01 -1.73787437e-02 3.51301402e-01
-9.82453346e-01 7.96627462e-01 4.17974043e+00 1.29297674e+00
-1.22352278e+00 1.27874896e-01 4.07097429e-01 -3.99595723e-02
-7.49222517e-01 4.00079861e-02 -6.99485600e-01 4.32440609e-01
2.31847197e-01 -7.29086772e-02 5.39905787e-01 8.09543729e-01
-3.33148003e-01 2.88723499e-01 -1.22664225e+00 1.22465706e+00
2.51228988e-01 -1.25854242e+00 5.54120719e-01 8.17935765e-02
5.53209424e-01 -4.52432960e-01 2.04045728e-01 3.39833528e-01
-1.87239751e-01 -6.91519201e-01 7.35726893e-01 5.89977920e-01
1.06170952e+00 -7.42492080e-01 2.87368894e-01 1.51922151e-01
-1.50784814e+00 4.46147360e-02 -5.67024350e-01 4.64658409e-01
-1.56955477e-02 3.68146360e-01 -3.77863646e-01 9.97623980e-01
5.11533082e-01 9.22499657e-01 -5.69816470e-01 6.23910308e-01
3.20290811e-02 -1.39712110e-01 -5.04715405e-02 3.94039005e-01
3.08390468e-01 -4.06255841e-01 5.54578841e-01 8.00808370e-01
4.75175381e-01 -2.08579972e-01 2.10869372e-01 1.28144383e+00
-2.74368841e-02 -9.58743542e-02 -9.02658045e-01 -8.15101340e-02
3.21740925e-01 1.38162923e+00 -4.90394950e-01 -3.95188719e-01
-4.94018108e-01 1.19498253e+00 3.78177255e-01 6.87469304e-01
-8.60321403e-01 -2.40045965e-01 7.27286696e-01 -1.29717663e-01
6.36480331e-01 -1.04307495e-01 -2.41175275e-02 -1.55221844e+00
2.59404153e-01 -5.61395824e-01 4.73850846e-01 -6.46355689e-01
-1.54761255e+00 6.18957818e-01 -9.46483389e-02 -1.62725389e+00
-1.30413070e-01 -3.95228744e-01 -3.19336027e-01 8.76093090e-01
-1.71059382e+00 -1.71142864e+00 -4.82513338e-01 9.32203710e-01
5.68294704e-01 -1.31212384e-01 8.48259270e-01 5.31637430e-01
-3.44504565e-01 8.06099415e-01 7.45617151e-02 8.15868843e-03
8.21274400e-01 -9.85299349e-01 7.95725137e-02 5.16178787e-01
4.25863206e-01 8.26921046e-01 2.94842005e-01 -4.90786999e-01
-1.65474570e+00 -1.13106477e+00 5.30848384e-01 -1.38301849e-01
6.40683234e-01 -7.18191564e-01 -1.11615896e+00 3.89786750e-01
-9.79875252e-02 3.82145077e-01 5.91246605e-01 1.59032904e-02
-8.56971145e-01 -3.60169202e-01 -9.47632194e-01 4.56025720e-01
1.12617540e+00 -1.06665230e+00 -3.54716599e-01 7.94162825e-02
5.40915966e-01 5.55017479e-02 -9.29379106e-01 3.74990940e-01
8.17017853e-01 -9.30142403e-01 1.16959584e+00 -7.16148138e-01
5.10905623e-01 -3.40703428e-01 -5.33864498e-01 -1.06210482e+00
-2.69132644e-01 -6.00634068e-02 -5.79829626e-02 1.40797186e+00
-2.77944449e-02 -4.08159435e-01 8.76063466e-01 4.01583433e-01
2.12934420e-01 -6.42962635e-01 -1.05733562e+00 -8.30917597e-01
-1.28106326e-01 -1.00101583e-01 6.81624949e-01 1.05650496e+00
-3.06727320e-01 3.10680002e-01 -6.38429105e-01 1.78295314e-01
9.71322417e-01 7.75058508e-01 6.99397266e-01 -1.13746476e+00
-4.18137223e-01 -2.37346098e-01 -5.18370271e-01 -8.71518850e-01
3.63623410e-01 -9.81718540e-01 -1.83237061e-01 -1.28920317e+00
3.75329077e-01 -5.05688787e-01 -7.72469044e-01 2.87263513e-01
-8.18892568e-02 3.13887388e-01 4.46960539e-01 1.56755254e-01
-7.20645905e-01 1.06363833e+00 1.25052941e+00 -4.26219165e-01
2.50464141e-01 -3.96559477e-01 -4.10157412e-01 2.43556336e-01
2.61703700e-01 -1.80071652e-01 -7.99447596e-01 -4.64054316e-01
1.01354122e-01 2.13453725e-01 8.27685714e-01 -6.37491286e-01
3.15141052e-01 -1.75488338e-01 2.50900865e-01 -6.41020119e-01
6.95853412e-01 -1.24138677e+00 2.18323961e-01 5.53441793e-02
-2.92813540e-01 -4.22433496e-01 -1.69671535e-01 9.79723096e-01
-4.39217895e-01 7.31389150e-02 7.60996521e-01 -8.32039490e-02
-8.46116245e-01 6.20649219e-01 3.82259756e-01 -1.55322418e-01
1.00423861e+00 -3.30724925e-01 -3.38201612e-01 -2.43943781e-01
-8.06843519e-01 7.39466026e-02 6.88156605e-01 7.60528028e-01
8.80850971e-01 -1.63884664e+00 -2.85620719e-01 6.42183304e-01
7.05489337e-01 -2.51862109e-01 8.35540771e-01 5.60633302e-01
1.38382271e-01 4.89511490e-01 -3.75204802e-01 -5.74191689e-01
-9.83654201e-01 9.38265681e-01 1.87077805e-01 -1.48500860e-01
-6.19452298e-01 5.75252295e-01 8.28696072e-01 -3.76693517e-01
2.04112053e-01 2.07384527e-01 9.77285206e-02 2.26416141e-01
4.04197812e-01 2.24407956e-01 4.12371941e-02 -6.09194279e-01
-3.87729138e-01 6.91076577e-01 -2.23639701e-02 1.48695323e-03
1.16904986e+00 -2.76856571e-01 -9.91444960e-02 5.59516490e-01
1.56951797e+00 -1.47977278e-01 -1.26957262e+00 -5.02964795e-01
-1.17149219e-01 -7.80498862e-01 -1.13577060e-02 -5.25808096e-01
-1.24556100e+00 1.11701465e+00 5.50843894e-01 -5.58166113e-03
1.06669068e+00 3.50123271e-02 7.52375722e-01 1.21278219e-01
5.82657516e-01 -8.41881275e-01 2.26768002e-01 3.38295475e-02
9.51293230e-01 -1.28282821e+00 -2.43688256e-01 -3.75165492e-01
-4.80794936e-01 1.08353174e+00 5.57045460e-01 -8.20791721e-02
5.48593879e-01 -3.05236936e-01 -2.98635751e-01 -3.02713573e-01
-5.81642330e-01 -1.94773823e-01 5.73588789e-01 4.06301349e-01
1.12181760e-01 -2.26954259e-02 -1.35644153e-01 5.33334851e-01
3.49411756e-01 -1.41714185e-01 1.12738619e-02 8.43779445e-01
-3.66553776e-02 -1.24387240e+00 -1.91239029e-01 1.08258696e-02
6.74119741e-02 -4.78087738e-02 -3.31874013e-01 7.17398286e-01
9.06452313e-02 5.88973761e-01 -9.95755475e-03 -2.56559014e-01
1.81301519e-01 2.09254041e-01 5.31200588e-01 -4.01470393e-01
-9.94235799e-02 1.82094291e-01 -3.26004207e-01 -6.39425814e-01
-3.74614239e-01 -1.87692389e-01 -7.48692393e-01 -8.38075578e-02
-1.64438888e-01 1.20449789e-01 6.03515565e-01 7.48605907e-01
4.42139477e-01 3.77418309e-01 7.04751909e-01 -6.77790344e-01
-5.68529010e-01 -5.13628304e-01 -9.01190698e-01 6.87312186e-01
3.05038661e-01 -9.72383618e-01 -3.00178736e-01 -2.04002872e-01] | [11.641742706298828, 0.712063729763031] |
2b8f7a0c-817e-4a7b-b96e-244cc6b96ffb | deep-node-ranking-an-algorithm-for-structural | 1902.03964 | null | https://arxiv.org/abs/1902.03964v6 | https://arxiv.org/pdf/1902.03964v6.pdf | Deep Node Ranking for Neuro-symbolic Structural Node Embedding and Classification | Network node embedding is an active research subfield of complex network analysis. This paper contributes a novel approach to learning network node embeddings and direct node classification using a node ranking scheme coupled with an autoencoder-based neural network architecture. The main advantages of the proposed Deep Node Ranking (DNR) algorithm are competitive or better classification performance, significantly higher learning speed and lower space requirements when compared to state-of-the-art approaches on 15 real-life node classification benchmarks. Furthermore, it enables exploration of the relationship between symbolic and the derived sub-symbolic node representations, offering insights into the learned node space structure. To avoid the space complexity bottleneck in a direct node classification setting, DNR computes stationary distributions of personalized random walks from given nodes in mini-batches, scaling seamlessly to larger networks. The scaling laws associated with DNR were also investigated on 1488 synthetic Erd\H{o}s-R\'enyi networks, demonstrating its scalability to tens of millions of links. | ['Nada Lavrač', 'Marko Robnik-Šikonja', 'Blaž Škrlj', 'Jan Kralj', 'Janez Konc'] | 2019-02-11 | null | null | null | null | ['structural-node-embedding'] | ['graphs'] | [-1.26052245e-01 4.03125316e-01 -4.21145797e-01 -3.14389139e-01
6.64754063e-02 -4.01502252e-01 6.97125375e-01 4.74527448e-01
-4.64555413e-01 5.82038343e-01 -6.43092170e-02 -5.89431167e-01
-6.93179846e-01 -1.32693708e+00 -2.97029465e-01 -6.94678962e-01
-8.54747653e-01 7.98416078e-01 2.44857103e-01 -3.89480114e-01
-2.34226570e-01 7.07272947e-01 -1.14273500e+00 -4.12339866e-01
3.94041568e-01 9.35023248e-01 -2.48282865e-01 8.42280328e-01
3.99631448e-02 7.76255786e-01 -5.12499869e-01 -5.89015603e-01
2.51760662e-01 1.69811264e-01 -8.50804090e-01 -4.84723181e-01
1.25397697e-01 -4.21848744e-01 -1.15324056e+00 8.72360826e-01
4.26215321e-01 4.88602258e-02 9.04396534e-01 -1.66972673e+00
-6.73141718e-01 1.23548508e+00 -2.20404819e-01 3.38629693e-01
1.18663929e-01 -3.73710185e-01 1.54763520e+00 -5.16623139e-01
8.44585478e-01 1.23670578e+00 1.14274585e+00 2.58399129e-01
-1.57580984e+00 -6.18618190e-01 4.27080132e-03 1.72105864e-01
-1.59128523e+00 5.04911095e-02 7.81501234e-01 -3.69637579e-01
9.13885415e-01 9.99036580e-02 8.38960946e-01 1.34582818e+00
7.98309818e-02 3.54146481e-01 4.77746040e-01 -9.42930207e-02
1.64067417e-01 1.97035953e-01 4.38702583e-01 9.06586647e-01
5.16981006e-01 1.12774156e-01 -3.96547839e-02 -4.17311341e-01
7.99065828e-01 4.87788051e-01 2.04822585e-01 -4.01961297e-01
-1.11644566e+00 1.06949973e+00 1.14012218e+00 4.70089048e-01
-3.39743197e-01 6.71789348e-01 7.75135815e-01 7.74519801e-01
5.11428654e-01 2.28529215e-01 -5.94580948e-01 1.87020749e-01
-6.55685663e-01 1.38451040e-01 1.13196468e+00 7.49375284e-01
8.04683864e-01 1.78992495e-01 2.26648733e-01 5.51147282e-01
5.19095242e-01 2.36307576e-01 2.13984296e-01 -8.53519857e-01
1.75045699e-01 8.45449448e-01 -6.49844706e-01 -1.42551327e+00
-6.30035758e-01 -8.18286121e-01 -1.21995759e+00 2.24069193e-01
2.20245942e-01 -1.38537228e-01 -3.79802972e-01 1.54988980e+00
4.14161682e-01 1.74948171e-01 1.49427459e-01 3.02741736e-01
9.92558420e-01 6.64778233e-01 9.93216261e-02 2.80162960e-01
1.17612493e+00 -8.27251792e-01 -2.79341668e-01 2.97606409e-01
7.77562439e-01 1.10745534e-01 5.30044317e-01 -5.84959649e-02
-6.10220134e-01 -3.43800426e-01 -1.00395775e+00 -1.92260101e-01
-7.75730312e-01 1.82978548e-02 1.11957872e+00 5.31599998e-01
-1.54495275e+00 1.17088902e+00 -7.09416926e-01 -6.76697373e-01
5.27440190e-01 6.84756815e-01 -7.31483698e-01 1.61249965e-01
-1.48258281e+00 5.00121832e-01 5.93799353e-01 3.99754867e-02
-9.70797241e-01 -8.95960808e-01 -9.87715185e-01 5.30003309e-01
2.14283913e-01 -4.05309349e-01 7.58464575e-01 -5.90861678e-01
-1.32351851e+00 4.94226158e-01 2.36690864e-01 -6.20298982e-01
2.37285912e-01 1.64398700e-01 -6.67607367e-01 2.94263035e-01
-5.41709922e-02 6.24923050e-01 6.84710979e-01 -8.87189746e-01
-2.68730521e-01 -1.86563730e-01 2.52593696e-01 -1.45865038e-01
-1.08854854e+00 -4.38197076e-01 -1.07232086e-01 -6.38995171e-01
-1.01337649e-01 -8.62639606e-01 -3.96426409e-01 2.67957717e-01
-5.45420587e-01 -7.63089299e-01 8.24101508e-01 -3.51000994e-01
1.31915438e+00 -1.90514195e+00 2.19181553e-01 8.35439086e-01
9.68281806e-01 8.51965696e-02 -4.45513368e-01 8.77258480e-01
-2.51509070e-01 3.55694473e-01 2.61899561e-01 -1.52611598e-01
1.54457167e-01 1.76485389e-01 7.30831698e-02 4.57834482e-01
9.10239965e-02 9.98797119e-01 -1.00584209e+00 -4.59604204e-01
1.85366217e-02 7.09207416e-01 -3.87627065e-01 -1.63943529e-01
1.11956418e-01 -2.94304848e-01 -4.25703436e-01 5.99993348e-01
2.62102634e-01 -7.09474623e-01 6.03463173e-01 -2.63936102e-01
4.88993734e-01 2.05438778e-01 -8.81873548e-01 1.20693457e+00
-2.70800740e-01 1.06368613e+00 -2.55754720e-02 -1.20868504e+00
1.00751007e+00 2.18106106e-01 5.83474040e-01 -3.18364352e-01
2.59447277e-01 1.39939524e-02 1.29232749e-01 -1.33626536e-01
3.20214212e-01 3.49064052e-01 -1.56688457e-03 5.31641364e-01
4.34238553e-01 6.56555533e-01 3.37408543e-01 6.31052554e-01
1.69468117e+00 -5.84196568e-01 2.41631806e-01 -5.70330918e-01
4.11820889e-01 -2.44269818e-01 2.45396912e-01 6.98217392e-01
-8.22546706e-02 -2.27599263e-01 1.01165950e+00 -6.61035359e-01
-1.07007217e+00 -1.13709462e+00 -1.72227740e-01 1.43816900e+00
-1.10620735e-02 -7.73105860e-01 -4.95049536e-01 -7.00516045e-01
4.77548271e-01 1.44271448e-01 -1.09999478e+00 -2.99769342e-01
-5.09238660e-01 -5.33478200e-01 8.79425108e-01 6.54058576e-01
3.12065035e-02 -9.61027682e-01 -6.90556411e-03 1.96972579e-01
4.72241551e-01 -9.70343232e-01 1.36035040e-01 2.89142251e-01
-1.10897899e+00 -1.22306275e+00 -5.47271252e-01 -1.02350914e+00
7.91400850e-01 1.17544100e-01 1.40959394e+00 4.68445361e-01
-4.87994134e-01 3.29917431e-01 -2.44365051e-01 2.89902449e-01
-5.24403334e-01 7.38176525e-01 3.76184553e-01 -3.55758280e-01
2.81277448e-01 -1.05371308e+00 -5.62530816e-01 1.80692345e-01
-7.05411375e-01 -5.54604650e-01 7.63738811e-01 8.90305340e-01
1.63364992e-01 4.32801276e-01 6.37128174e-01 -1.09407687e+00
9.29398179e-01 -8.98570001e-01 -5.05233467e-01 1.26849726e-01
-1.13504267e+00 1.58374369e-01 7.83471882e-01 -4.81230259e-01
-4.00072187e-01 -4.79953289e-01 -1.90206673e-02 -4.26188260e-01
1.79176256e-01 6.74236774e-01 2.59142041e-01 -1.40368506e-01
6.61652744e-01 1.02987187e-02 3.17327291e-01 -4.32586581e-01
4.92897570e-01 4.38142598e-01 2.15967834e-01 -3.34134758e-01
1.32121658e+00 3.68257046e-01 3.79856348e-01 -9.02796686e-01
-1.96009010e-01 -2.51281261e-01 -7.05202699e-01 4.23214026e-02
5.33070624e-01 -8.67342949e-01 -1.28631830e+00 4.38543372e-02
-8.54410470e-01 -4.03648466e-01 -2.02330559e-01 1.91847011e-01
1.36380568e-01 3.11114460e-01 -1.14104629e+00 -4.45070177e-01
-5.95681727e-01 -6.73093736e-01 6.19775593e-01 7.48082250e-02
-2.84244359e-01 -1.63956106e+00 2.25843757e-01 -1.45434886e-01
5.89986801e-01 3.02414477e-01 1.32615542e+00 -1.22275102e+00
-3.96624118e-01 -6.77961946e-01 -4.53962147e-01 2.01104000e-01
-2.32863739e-01 3.19979340e-01 -6.18050516e-01 -5.49292743e-01
-9.70713258e-01 -1.57155603e-01 7.29293466e-01 8.63047466e-02
1.00925601e+00 -4.85890836e-01 -7.28068233e-01 4.52523232e-01
1.49902356e+00 -2.48634219e-01 3.31506252e-01 2.32508853e-01
8.60412002e-01 4.63271379e-01 -7.50275031e-02 2.96875596e-01
5.20633876e-01 3.93819302e-01 6.31760836e-01 -7.89650157e-02
1.89097524e-02 -3.95789802e-01 1.53512895e-01 8.69807839e-01
4.89987396e-02 -4.12756056e-01 -9.92986381e-01 4.99784708e-01
-1.61118042e+00 -8.52658689e-01 -1.73282512e-02 1.61026454e+00
4.07891691e-01 3.48505914e-01 3.53749871e-01 5.06159782e-01
6.67560220e-01 3.60951722e-01 -5.32377660e-01 -1.49383396e-01
2.44243503e-01 1.58027202e-01 4.14325684e-01 3.32151473e-01
-1.13242614e+00 7.21121252e-01 6.85477400e+00 7.97430813e-01
-8.84729207e-01 5.21426685e-02 5.03644645e-01 2.26871148e-01
-5.03300905e-01 -1.68895438e-01 -6.52591944e-01 1.64370999e-01
1.19429934e+00 -3.49288657e-02 5.36882460e-01 1.24808145e+00
-3.96817207e-01 5.40721953e-01 -1.31668770e+00 8.02040160e-01
-1.77344754e-01 -1.64924514e+00 -1.15196757e-01 3.36648852e-01
4.39542413e-01 4.48180497e-01 -8.72476250e-02 5.89746773e-01
6.14263773e-01 -1.32680881e+00 1.52675703e-01 3.11709076e-01
7.90913820e-01 -7.59635687e-01 8.85635674e-01 3.26919965e-02
-1.47585964e+00 -3.73858541e-01 -4.66735393e-01 -1.56361222e-01
-2.66590625e-01 7.29119003e-01 -1.05918169e+00 4.44910794e-01
9.18991983e-01 1.18593061e+00 -8.99197042e-01 4.66291070e-01
2.49640439e-02 6.38655186e-01 -4.81944740e-01 -5.50874233e-01
1.48315996e-01 -1.76942885e-01 4.48289335e-01 1.11773455e+00
-3.47778685e-02 -3.08087230e-01 5.80470869e-03 7.96140790e-01
-5.28140068e-01 -2.82547265e-01 -7.97734320e-01 -4.48375970e-01
9.01993692e-01 1.78715932e+00 -8.82655859e-01 -2.42563412e-01
-8.73965248e-02 6.45340264e-01 7.53381014e-01 4.51640397e-01
-5.75477481e-01 -9.25783455e-01 7.97416449e-01 2.30935752e-01
3.57463747e-01 -1.20116130e-01 1.09282449e-01 -7.73169518e-01
-3.27275842e-01 -5.20730495e-01 3.05772871e-01 -1.87311396e-01
-1.59783149e+00 8.81536186e-01 -9.73943695e-02 -1.02119982e+00
-3.27208102e-01 -8.19269896e-01 -7.67768621e-01 3.82128417e-01
-1.35509515e+00 -9.28962052e-01 -3.69042039e-01 4.16334778e-01
-1.09392770e-01 -6.01391435e-01 1.12429619e+00 5.70536673e-01
-7.73021340e-01 9.70228016e-01 5.69047630e-01 7.06836522e-01
-4.84419428e-02 -1.40680206e+00 6.17422283e-01 2.92543888e-01
3.02627295e-01 6.91680133e-01 3.93451780e-01 -4.06127036e-01
-1.46961617e+00 -1.02458596e+00 7.40791261e-01 -2.59597510e-01
1.16489899e+00 -5.62368929e-01 -7.41316259e-01 6.84626043e-01
-3.64924669e-02 4.30600911e-01 8.11668396e-01 4.75518882e-01
-6.01736188e-01 -4.97406125e-01 -1.12441528e+00 6.53300643e-01
1.25766230e+00 -8.31829309e-01 7.69848600e-02 2.69766957e-01
9.65595484e-01 2.92600185e-01 -1.64513493e+00 3.59607458e-01
6.50379539e-01 -6.70354724e-01 1.32104659e+00 -7.63569236e-01
4.58901495e-01 -7.88133144e-02 1.65440112e-01 -1.08648956e+00
-6.92526639e-01 -5.24223268e-01 -6.57228827e-01 1.17486751e+00
5.31326890e-01 -7.16138959e-01 1.01735711e+00 1.53185412e-01
3.74158353e-01 -1.06214345e+00 -8.67643774e-01 -5.13467312e-01
-1.16071915e-02 -8.07267427e-02 5.41191399e-01 1.01282930e+00
1.11530982e-02 4.88356143e-01 -7.14489892e-02 -2.61499006e-02
7.94793010e-01 -2.61952281e-01 5.96223414e-01 -2.02759171e+00
-3.18256825e-01 -5.84185839e-01 -9.52853799e-01 -8.42848003e-01
5.57605624e-01 -1.38016415e+00 -6.46103203e-01 -1.33717871e+00
-5.97836599e-02 -8.24849784e-01 -6.08020723e-01 4.64606941e-01
2.40409702e-01 3.28100532e-01 -7.16645494e-02 2.05111817e-01
-7.06554055e-01 5.89632452e-01 8.21744204e-01 -2.81431764e-01
9.43946391e-02 -5.37996069e-02 -5.72382867e-01 5.39346159e-01
7.02194512e-01 -6.51522338e-01 -5.84408104e-01 -3.88177931e-01
5.90877116e-01 1.34957926e-02 4.74601775e-01 -1.01727939e+00
3.32151800e-01 5.30871153e-01 5.09524524e-01 -5.36406577e-01
2.86532938e-01 -1.15890002e+00 1.56748503e-01 7.40726829e-01
-6.58827960e-01 3.07024777e-01 -3.23138177e-01 1.20827293e+00
6.58503622e-02 -8.35636258e-02 3.54776502e-01 2.72783846e-01
-5.63084543e-01 5.60350537e-01 -3.06883872e-01 -1.90625414e-01
1.13302827e+00 -1.61647871e-01 -3.32496077e-01 -3.37180823e-01
-7.53899693e-01 3.49142611e-01 2.04508126e-01 3.48080993e-01
5.20514369e-01 -1.53270125e+00 -5.04203975e-01 1.00732744e-01
-5.68614854e-03 -1.76903635e-01 -2.12196130e-02 5.90052843e-01
-7.83700943e-01 3.16365719e-01 -1.27142102e-01 -5.32381713e-01
-1.08046424e+00 3.55676621e-01 3.27270180e-01 -6.39612615e-01
-8.68917882e-01 7.84529984e-01 -5.43525636e-01 -9.58035886e-01
6.45951807e-01 2.82106865e-02 -4.17692214e-01 1.99096695e-01
2.46218503e-01 8.70610893e-01 -1.53339997e-01 -2.71052182e-01
-3.91063601e-01 1.85088739e-01 -2.52244145e-01 3.24528724e-01
1.59443092e+00 1.34316698e-01 -3.41106296e-01 3.19530398e-01
1.74224734e+00 -4.69895542e-01 -8.64784777e-01 -5.17250180e-01
1.99051961e-01 6.18476570e-02 1.03686906e-01 -3.45330328e-01
-1.38644016e+00 6.81975842e-01 6.25016153e-01 7.88284004e-01
5.38285315e-01 2.10842207e-01 6.43467724e-01 1.20863986e+00
1.63426831e-01 -8.41164172e-01 1.79480895e-01 4.89140272e-01
3.84609044e-01 -1.20045698e+00 1.74013853e-01 -3.44815016e-01
-1.29282728e-01 1.48476386e+00 3.89175892e-01 -5.00889540e-01
1.39474356e+00 8.10157880e-02 -2.83046156e-01 -5.13659179e-01
-8.02443683e-01 1.88200235e-01 2.71982737e-02 6.34835660e-01
3.03123862e-01 1.14707440e-01 2.73156434e-01 2.90644705e-01
-1.57370329e-01 -3.26900363e-01 2.90585965e-01 4.23293680e-01
-1.52064413e-01 -1.05413818e+00 3.74896675e-01 8.59131932e-01
-1.84023604e-01 -1.65762499e-01 -1.67644873e-01 1.09515762e+00
-4.81000602e-01 5.61878681e-01 2.81260967e-01 -7.66984701e-01
-1.04827918e-01 -1.90735623e-01 -1.54564768e-01 -3.70922267e-01
-6.89564764e-01 -5.01685202e-01 4.71048243e-02 -4.14580464e-01
-1.91492528e-01 -2.62731999e-01 -1.14640617e+00 -9.01138604e-01
-3.69650930e-01 2.82248318e-01 5.75512350e-01 5.70490122e-01
6.21688247e-01 7.73644567e-01 8.59701574e-01 -9.41886246e-01
-7.66166389e-01 -1.05833793e+00 -7.51553953e-01 4.33458120e-01
2.41891578e-01 -5.18139601e-01 -6.39306843e-01 -6.59618497e-01] | [7.0467023849487305, 6.278100490570068] |
a295451f-beb8-4201-900c-73a01fc008c4 | bmd-a-general-class-balanced-multicentric | 2204.02811 | null | https://arxiv.org/abs/2204.02811v2 | https://arxiv.org/pdf/2204.02811v2.pdf | BMD: A General Class-balanced Multicentric Dynamic Prototype Strategy for Source-free Domain Adaptation | Source-free Domain Adaptation (SFDA) aims to adapt a pre-trained source model to the unlabeled target domain without accessing the well-labeled source data, which is a much more practical setting due to the data privacy, security, and transmission issues. To make up for the absence of source data, most existing methods introduced feature prototype based pseudo-labeling strategies to realize self-training model adaptation. However, feature prototypes are obtained by instance-level predictions based feature clustering, which is category-biased and tends to result in noisy labels since the visual domain gaps between source and target are usually different between categories. In addition, we found that a monocentric feature prototype may be ineffective to represent each category and introduce negative transfer, especially for those hard-transfer data. To address these issues, we propose a general class-Balanced Multicentric Dynamic prototype (BMD) strategy for the SFDA task. Specifically, for each target category, we first introduce a global inter-class balanced sampling strategy to aggregate potential representative target samples. Then, we design an intra-class multicentric clustering strategy to achieve more robust and representative prototypes generation. In contrast to existing strategies that update the pseudo label at a fixed training period, we further introduce a dynamic pseudo labeling strategy to incorporate network update information during model adaptation. Extensive experiments show that the proposed model-agnostic BMD strategy significantly improves representative SFDA methods to yield new state-of-the-art results. The code is available at https://github.com/ispc-lab/BMD. | ['DaCheng Tao', 'wei he', 'Zhijun Li', 'Jing Zhang', 'Guang Chen', 'Sanqing Qu'] | 2022-04-06 | null | null | null | null | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 2.04117492e-01 -1.03803709e-01 -5.14443457e-01 -5.93154848e-01
-7.86005020e-01 -6.33773029e-01 3.98934841e-01 3.74761894e-02
-2.98942715e-01 8.43824327e-01 -2.14928165e-01 -5.92616433e-03
1.15380682e-01 -7.64068305e-01 -6.34821177e-01 -8.38492572e-01
3.42298359e-01 5.16506374e-01 3.61735910e-01 1.44755438e-01
-1.00937650e-01 4.47405249e-01 -1.20270777e+00 2.88506478e-01
1.03703845e+00 1.07936394e+00 3.45115691e-01 9.22266673e-03
-2.54144102e-01 1.68150872e-01 -7.14921296e-01 -3.28006357e-01
3.15689176e-01 -5.76828659e-01 -6.46372497e-01 4.01187956e-01
7.17016682e-02 -3.11050624e-01 -7.55325034e-02 1.19687629e+00
5.13634205e-01 2.30178952e-01 8.47938359e-01 -1.64420009e+00
-9.34977710e-01 4.49258745e-01 -7.28506327e-01 -4.21558246e-02
-4.75319847e-02 -2.91020107e-02 4.75932628e-01 -1.01110280e+00
6.51869416e-01 1.07374787e+00 5.24678528e-01 9.97919321e-01
-1.20424640e+00 -1.13024712e+00 4.11521137e-01 3.76285255e-01
-1.59097624e+00 -3.87097985e-01 1.01956558e+00 -1.59475237e-01
1.25305876e-01 -5.98731115e-02 4.31669384e-01 1.40079439e+00
-3.65721226e-01 8.17746162e-01 1.04646575e+00 -5.00433147e-01
5.31302691e-01 7.29487538e-01 1.30571783e-01 2.91185826e-01
2.92948842e-01 2.69593634e-02 -2.93571115e-01 -1.99484125e-01
6.13358736e-01 1.36639833e-01 -4.16541725e-01 -9.71816361e-01
-1.25092995e+00 8.43478501e-01 6.01615310e-01 1.38113871e-01
-2.02335432e-01 -3.84781629e-01 4.56002414e-01 2.75160730e-01
4.56489474e-01 -7.02931210e-02 -6.04335368e-01 3.58805597e-01
-6.62135422e-01 -1.54789761e-01 5.72295666e-01 1.28438652e+00
9.84641373e-01 6.01254851e-02 -1.85258672e-01 1.15628719e+00
2.76931822e-01 4.81093079e-01 7.72475481e-01 -7.98858821e-01
5.07590294e-01 4.80710924e-01 -2.52628475e-02 -7.45261729e-01
-1.69980019e-01 -6.45450175e-01 -1.01250279e+00 -1.21530853e-01
3.82723868e-01 -2.25955769e-01 -1.05611837e+00 1.96236181e+00
7.35050499e-01 2.63578176e-01 3.47105354e-01 8.32977414e-01
6.43638492e-01 6.23790920e-01 1.33316457e-01 -4.42639291e-01
1.10688186e+00 -9.93109882e-01 -5.12133181e-01 -1.74632728e-01
6.29953802e-01 -5.09767830e-01 1.20850742e+00 1.55407101e-01
-5.51126063e-01 -6.01301551e-01 -1.09014452e+00 3.90669227e-01
-5.30688167e-01 2.42386177e-01 2.47410432e-01 7.55180180e-01
-8.00855339e-01 8.98318663e-02 -5.41272879e-01 -4.68956202e-01
8.12559366e-01 2.06010744e-01 -3.41986001e-01 -4.44950998e-01
-1.23701394e+00 4.51407075e-01 5.76023519e-01 -1.50824279e-01
-8.72491479e-01 -5.65669894e-01 -7.74839342e-01 -7.19557330e-02
3.57761055e-01 -5.25078893e-01 1.24515712e+00 -1.33486974e+00
-1.51181459e+00 6.36027396e-01 -2.45736569e-01 -2.07056001e-01
4.22219694e-01 4.39522713e-01 -5.33100009e-01 1.90271214e-01
1.42646253e-01 9.45798934e-01 9.32139277e-01 -1.66954410e+00
-7.57176638e-01 -2.96178192e-01 -3.27997118e-01 3.80790055e-01
-8.43571961e-01 -3.73789907e-01 -5.72147667e-01 -7.27244139e-01
9.16674435e-02 -8.03541481e-01 -1.10519238e-01 2.00975150e-01
-4.01504844e-01 -1.54901162e-01 1.05660820e+00 -2.72009432e-01
9.98680532e-01 -2.31613588e+00 -1.45852655e-01 3.53804410e-01
6.31200597e-02 4.09789503e-01 -3.45266074e-01 1.09656215e-01
-9.89589840e-02 -6.81647658e-02 -4.11324084e-01 -3.12471062e-01
-1.98595509e-01 1.71501473e-01 -2.57402360e-01 3.17625910e-01
1.30075049e-02 5.06862164e-01 -8.73995662e-01 -5.16605914e-01
3.99167426e-02 5.11223912e-01 -3.96870941e-01 1.21531293e-01
-3.60106155e-02 4.96845722e-01 -5.34410894e-01 9.15992141e-01
1.02210832e+00 -4.20445561e-01 1.67155191e-01 -2.01561421e-01
4.18168813e-01 -2.72199005e-01 -1.13758004e+00 1.55845892e+00
-3.41329932e-01 2.26968914e-01 8.33778409e-04 -1.25215936e+00
9.71492887e-01 3.28795314e-01 3.28513116e-01 -6.23122573e-01
1.08869359e-01 2.90819496e-01 -1.34853259e-01 1.92106068e-02
2.02186987e-01 -1.07038721e-01 -7.03236386e-02 2.61655062e-01
5.45527749e-02 2.25766405e-01 -7.53531381e-02 1.35966748e-01
8.44679415e-01 -1.73878800e-02 2.93288827e-01 8.83069932e-02
5.14801860e-01 2.25276992e-01 9.49257076e-01 4.57019418e-01
-5.92580557e-01 8.24034452e-01 1.05256498e-01 -6.59362450e-02
-7.42431462e-01 -1.18259430e+00 -2.33489916e-01 9.61932361e-01
4.41266477e-01 1.18880421e-01 -8.79709661e-01 -1.09615409e+00
-1.41703501e-01 7.79096961e-01 -3.89940262e-01 -5.08921444e-01
-1.67009845e-01 -7.69649088e-01 3.11196536e-01 4.25226122e-01
7.55011141e-01 -8.87793720e-01 -9.68643278e-02 2.50796795e-01
-3.74334425e-01 -8.61497223e-01 -7.14941680e-01 2.33837858e-01
-8.07613373e-01 -9.45571303e-01 -1.17123663e+00 -9.12508667e-01
1.07156122e+00 5.57811141e-01 6.60261869e-01 -3.35964590e-01
1.49262026e-01 3.11811388e-01 -5.56913495e-01 -3.56859922e-01
-4.48242277e-01 2.46552587e-01 1.59409553e-01 1.73261985e-01
6.04061663e-01 -4.81931537e-01 -6.01858556e-01 5.70842147e-01
-8.80048096e-01 -6.81079105e-02 6.06581271e-01 1.03798807e+00
8.87456179e-01 7.05795586e-02 1.12895834e+00 -1.03123784e+00
5.78437805e-01 -9.00198042e-01 -4.92404938e-01 3.60818624e-01
-8.22084069e-01 -3.50995988e-01 9.12477911e-01 -8.78949463e-01
-1.16441739e+00 3.74600589e-01 1.41772762e-01 -7.34752178e-01
-4.55280393e-01 3.08026880e-01 -5.23388863e-01 -8.26643631e-02
8.39492798e-01 3.62502664e-01 4.00770940e-02 -3.60318750e-01
2.49175787e-01 1.08697987e+00 3.70832145e-01 -5.48353136e-01
8.66277277e-01 3.11710626e-01 -4.00736064e-01 -4.26862091e-01
-7.02513993e-01 -5.12449980e-01 -5.71998179e-01 1.21167459e-05
4.41298425e-01 -1.16705787e+00 -9.35420170e-02 5.27075529e-01
-8.98148179e-01 -5.69701493e-01 -3.92617822e-01 5.29251575e-01
-4.43947583e-01 3.51375192e-01 -2.75077343e-01 -4.30743039e-01
-2.62338579e-01 -9.57894862e-01 7.37001598e-01 4.21176910e-01
5.03864698e-02 -8.73680711e-01 -2.14369938e-01 1.64896339e-01
3.51165533e-01 -7.12540001e-02 8.89680028e-01 -1.04455650e+00
-3.46134245e-01 -1.87286884e-01 -3.92154098e-01 4.75714177e-01
4.85918790e-01 -3.83131474e-01 -9.39511895e-01 -5.10312021e-01
-1.32765397e-01 -3.62087578e-01 4.65794891e-01 1.81357488e-01
1.25124550e+00 -2.50498831e-01 -6.81498826e-01 5.20150065e-01
1.27054918e+00 4.37352002e-01 2.38115802e-01 2.69907832e-01
6.17872477e-01 4.79197174e-01 9.32522476e-01 5.16280353e-01
5.64264178e-01 6.39997065e-01 1.80223361e-01 -8.83743241e-02
-3.21828008e-01 -2.83098102e-01 3.03089231e-01 5.47522247e-01
5.83362520e-01 -5.15305221e-01 -8.05460632e-01 8.56989741e-01
-1.59316123e+00 -6.67888939e-01 2.57519931e-01 2.40223765e+00
1.05746806e+00 -1.19234644e-01 2.58768678e-01 5.51107489e-02
1.20526063e+00 -2.97823548e-01 -9.11571622e-01 -1.10924093e-03
2.27258410e-02 -1.62721068e-01 4.42566007e-01 4.88322191e-02
-1.00778508e+00 8.48379016e-01 5.46628428e+00 1.26463628e+00
-1.31919062e+00 4.37331259e-01 6.65694952e-01 2.72419937e-02
-1.95484206e-01 -2.22392231e-01 -9.41481352e-01 7.18582571e-01
8.24129164e-01 -2.85100073e-01 3.23737204e-01 1.10599720e+00
-8.78867954e-02 1.27988800e-01 -1.00994384e+00 8.69984925e-01
-8.70272890e-03 -9.98481750e-01 1.00343734e-01 2.18210258e-02
6.90634370e-01 6.44610077e-02 3.83730680e-02 4.38321054e-01
3.20913374e-01 -3.23562801e-01 5.51400900e-01 -2.59518567e-02
1.01375186e+00 -7.97146022e-01 5.04803658e-01 4.77918178e-01
-1.09934318e+00 -3.12919766e-01 -6.37428284e-01 5.07943690e-01
-2.24813744e-02 5.50915360e-01 -8.97556961e-01 5.78633606e-01
8.28727543e-01 6.58380449e-01 -5.48099875e-01 1.20010865e+00
5.58677921e-03 7.02992380e-01 -3.20623666e-01 2.01529846e-01
-4.86802682e-02 3.81553471e-02 4.77828354e-01 9.83827829e-01
5.99639297e-01 -1.28646955e-01 2.28824884e-01 5.94708562e-01
-2.64491111e-01 1.68667406e-01 -6.63540304e-01 2.87150532e-01
1.09054971e+00 1.14265656e+00 -7.24696815e-01 -4.27299500e-01
-4.02706832e-01 1.18181515e+00 2.50821322e-01 6.29429638e-01
-7.94476151e-01 -4.21620876e-01 4.09702748e-01 -6.33557793e-03
3.50471139e-01 1.40309826e-01 -1.51647210e-01 -1.19137287e+00
-1.49507308e-02 -7.75831103e-01 7.16302633e-01 -6.11488104e-01
-1.60014391e+00 6.40258610e-01 2.32759014e-01 -1.80028069e+00
-1.59663394e-01 -1.63774624e-01 -4.47901100e-01 6.59570813e-01
-1.61764598e+00 -1.18747711e+00 -4.20973271e-01 9.15320575e-01
6.54781044e-01 -3.76023740e-01 8.30668926e-01 5.23189485e-01
-6.30006075e-01 1.17881596e+00 5.03347754e-01 1.30013019e-01
1.09248412e+00 -8.41400564e-01 -1.16312280e-01 8.48477364e-01
-1.61792383e-01 4.80844438e-01 2.01267660e-01 -5.47435403e-01
-7.75862098e-01 -1.63746238e+00 4.70489591e-01 -1.11079976e-01
2.79267818e-01 -3.43387127e-01 -1.15241563e+00 6.53217435e-01
-2.45060287e-02 4.18464869e-01 8.00861180e-01 -2.68768132e-01
-4.64306176e-01 -5.11904955e-01 -1.63070774e+00 5.12545109e-01
1.03897071e+00 -1.25613779e-01 -3.47942889e-01 3.82409275e-01
7.22195029e-01 -2.06969678e-01 -7.01828301e-01 4.31948870e-01
1.53436631e-01 -6.96959496e-01 8.68434429e-01 -3.28522325e-02
-5.81145026e-02 -5.32986879e-01 -1.94225773e-01 -1.61699402e+00
-3.30593109e-01 -1.95516646e-01 -9.11751688e-02 1.66661429e+00
3.57645810e-01 -1.00359201e+00 9.15975928e-01 5.47370136e-01
-2.16991261e-01 -4.22314703e-01 -1.02587128e+00 -1.16639864e+00
1.06236778e-01 -4.99053076e-02 7.76591301e-01 1.21244025e+00
-8.39032754e-02 2.27332965e-01 -1.92506716e-01 2.57014036e-01
7.78741539e-01 1.28537998e-01 6.00926399e-01 -1.20006967e+00
1.52026722e-02 -1.98227614e-01 -1.60737619e-01 -9.99448836e-01
2.84267724e-01 -8.95362675e-01 1.97022408e-01 -1.25934327e+00
1.66144535e-01 -9.71997797e-01 -5.63862443e-01 7.19061852e-01
-6.07872382e-02 3.76416236e-01 1.71677489e-02 4.68964905e-01
-5.67643762e-01 7.37631381e-01 1.23524535e+00 -1.59367040e-01
-3.81521285e-01 1.99499816e-01 -9.44536150e-01 3.12103182e-01
1.02809727e+00 -6.67200863e-01 -8.88756692e-01 -3.21030170e-01
-5.35230875e-01 -2.18519047e-01 2.25157812e-01 -1.09291351e+00
2.21597508e-01 -3.93819869e-01 5.38765609e-01 -5.18496752e-01
3.17864627e-01 -1.20931494e+00 1.29002601e-01 3.66439074e-01
-1.97608724e-01 -5.36949575e-01 1.62176117e-01 8.81224036e-01
-3.19390565e-01 -1.90286115e-01 1.02803063e+00 1.42501015e-02
-9.07679439e-01 4.78098750e-01 -2.24014923e-01 -1.28719226e-01
1.35253727e+00 -4.95588362e-01 -4.63248610e-01 -2.79632479e-01
-4.09073472e-01 3.53398472e-01 6.48371696e-01 4.10714090e-01
5.81588447e-01 -1.62721789e+00 -4.29005563e-01 2.69176066e-01
4.82709557e-01 3.04575920e-01 3.61834437e-01 6.30801082e-01
2.16567218e-02 9.80954096e-02 -2.26871684e-01 -6.55884027e-01
-1.03714359e+00 8.45929980e-01 2.55157530e-01 4.59682979e-02
-3.21458846e-01 7.96290696e-01 5.21005988e-01 -8.42300832e-01
2.80357957e-01 8.56290013e-02 -2.23417968e-01 1.03740387e-01
4.51258332e-01 1.71913877e-01 -8.73519927e-02 -6.87120497e-01
-3.77745301e-01 4.80451524e-01 -2.54575521e-01 1.78757817e-01
1.05995965e+00 -6.13112330e-01 3.69259775e-01 2.58272290e-01
1.26669800e+00 -2.89123088e-01 -1.48722482e+00 -6.54803395e-01
-2.36401588e-01 -3.53766501e-01 -2.16900706e-01 -8.92915368e-01
-1.30412269e+00 8.23077857e-01 9.47070003e-01 -4.95630242e-02
1.48318481e+00 -8.08053389e-02 9.02867317e-01 3.63074750e-01
4.49759752e-01 -1.21617591e+00 1.22205198e-01 2.34843865e-01
6.35324895e-01 -1.33043551e+00 -2.56595761e-01 -6.01946235e-01
-8.73222530e-01 7.97101557e-01 9.94827688e-01 3.66403371e-01
6.67876303e-01 -2.00999334e-01 2.81699359e-01 3.13376218e-01
-4.88023311e-01 6.70181727e-03 4.71031256e-02 1.24122751e+00
-1.38568372e-01 2.35318835e-03 5.63860685e-02 9.11082923e-01
2.95894504e-01 1.94117934e-01 4.56007063e-01 8.47361267e-01
-2.41096154e-01 -1.43448997e+00 -4.79440302e-01 3.79650325e-01
-1.03420071e-01 6.45641834e-02 -2.05066890e-01 7.00365365e-01
1.62275076e-01 9.86691296e-01 -1.35696884e-02 -4.47247386e-01
3.50626171e-01 2.29001954e-01 8.74307305e-02 -7.68868387e-01
-1.90259829e-01 7.73607120e-02 -3.33478570e-01 -1.30949855e-01
-3.50055754e-01 -6.37189269e-01 -1.31193233e+00 -1.58974215e-01
-4.33845222e-01 1.83646962e-01 4.75675136e-01 6.30059779e-01
7.21568167e-01 2.17523038e-01 9.78567243e-01 -7.25986898e-01
-5.79415023e-01 -9.21974778e-01 -5.41491210e-01 3.17642123e-01
2.03530580e-01 -7.87162662e-01 -3.22577059e-01 3.12699288e-01] | [10.404645919799805, 3.0598461627960205] |
7ce95271-1d76-4315-ae68-624782065afa | exot-exit-aware-object-tracker-for-safe | 2306.05262 | null | https://arxiv.org/abs/2306.05262v1 | https://arxiv.org/pdf/2306.05262v1.pdf | EXOT: Exit-aware Object Tracker for Safe Robotic Manipulation of Moving Object | Current robotic hand manipulation narrowly operates with objects in predictable positions in limited environments. Thus, when the location of the target object deviates severely from the expected location, a robot sometimes responds in an unexpected way, especially when it operates with a human. For safe robot operation, we propose the EXit-aware Object Tracker (EXOT) on a robot hand camera that recognizes an object's absence during manipulation. The robot decides whether to proceed by examining the tracker's bounding box output containing the target object. We adopt an out-of-distribution classifier for more accurate object recognition since trackers can mistrack a background as a target object. To the best of our knowledge, our method is the first approach of applying an out-of-distribution classification technique to a tracker output. We evaluate our method on the first-person video benchmark dataset, TREK-150, and on the custom dataset, RMOT-223, that we collect from the UR5e robot. Then we test our tracker on the UR5e robot in real-time with a conveyor-belt sushi task, to examine the tracker's ability to track target dishes and to determine the exit status. Our tracker shows 38% higher exit-aware performance than a baseline method. The dataset and the code will be released at https://github.com/hskAlena/EXOT. | ['Byoung-Tak Zhang', 'Dong-Sig Han', 'Minji Kim', 'Hye Jung Yoon', 'Hyunseo Kim'] | 2023-06-08 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [-1.05092399e-01 -1.70703262e-01 -1.94110498e-01 -3.89018166e-03
-4.40940678e-01 -6.78945243e-01 2.23343968e-01 -4.35181856e-01
-3.59805763e-01 2.62185723e-01 -4.05205339e-01 -1.57926232e-01
6.59394041e-02 -1.90588266e-01 -9.29237723e-01 -7.12069750e-01
-1.00312166e-01 5.56617320e-01 7.11201072e-01 1.63165659e-01
1.63598463e-01 7.19237447e-01 -1.62204468e+00 3.65959674e-01
3.78746204e-02 1.07922363e+00 5.77068448e-01 8.58946383e-01
5.41556835e-01 9.81743574e-01 -1.01837897e+00 1.90356284e-01
6.37436628e-01 3.31711918e-02 -5.34827769e-01 1.39036760e-01
4.48811859e-01 -5.82546353e-01 -6.20462775e-01 1.04415238e+00
4.35447603e-01 8.32276302e-04 7.05181301e-01 -1.81712353e+00
-3.51136088e-01 4.57034707e-01 -6.89971626e-01 2.60441720e-01
5.28798878e-01 7.29320586e-01 4.81914252e-01 -8.55782866e-01
6.36378288e-01 1.33414769e+00 4.42297071e-01 5.42360246e-01
-8.25778067e-01 -7.73663998e-01 3.83857071e-01 3.05798292e-01
-1.47622168e+00 -3.05376261e-01 4.29513127e-01 -7.63176143e-01
6.19637191e-01 7.58593604e-02 3.64814401e-01 1.43849289e+00
5.11769235e-01 1.06065822e+00 7.30614424e-01 -1.33037657e-01
1.64136454e-01 -5.89075536e-02 1.60918474e-01 5.87632895e-01
2.68390059e-01 4.27062869e-01 -3.10762614e-01 1.06551349e-02
5.79911947e-01 1.53741494e-01 -5.08292079e-01 -6.08968437e-01
-1.63197911e+00 1.61604792e-01 3.15978438e-01 -8.74991342e-02
-4.96983081e-01 4.40656304e-01 4.89595026e-01 2.05399215e-01
-2.45535031e-01 1.54319659e-01 -5.01556575e-01 -4.61782515e-01
-1.75990000e-01 1.88447148e-01 7.33097374e-01 1.64552891e+00
1.04738884e-01 -2.53658175e-01 -4.92184579e-01 3.50678891e-01
3.03553432e-01 6.75670803e-01 2.58475482e-01 -1.06600213e+00
4.65402991e-01 4.75131899e-01 6.33533359e-01 -5.99266231e-01
-3.00829768e-01 -4.53278124e-02 -2.03070834e-01 1.01954782e+00
7.38017023e-01 -1.33626431e-01 -7.60014057e-01 1.24849701e+00
5.22250891e-01 -1.64305270e-01 1.04259618e-01 1.35628760e+00
4.98867929e-01 3.98434401e-01 -1.46728247e-01 2.02767983e-01
1.58870029e+00 -1.06042004e+00 -6.45204127e-01 -2.41992861e-01
5.39209783e-01 -8.31543267e-01 9.36012268e-01 9.93135571e-01
-3.26628923e-01 -7.06025243e-01 -9.05950010e-01 4.19459581e-01
-2.37022772e-01 8.78394544e-01 2.81476647e-01 4.77874368e-01
-4.91405845e-01 3.53209466e-01 -1.08031607e+00 -4.88664508e-01
3.30032766e-01 3.58445942e-01 -4.95181084e-01 -5.80518320e-02
-4.38996643e-01 9.55493093e-01 4.83224064e-01 5.63285835e-02
-1.18133664e+00 -2.05208987e-01 -6.46158159e-01 -2.63881505e-01
9.76702273e-01 -8.13639089e-02 1.65531540e+00 -6.06044471e-01
-1.48988366e+00 5.26804209e-01 1.62735179e-01 -1.45134076e-01
9.09017086e-01 -4.97257739e-01 -3.29346865e-01 2.00947989e-02
1.89370334e-01 4.65405405e-01 1.08243537e+00 -1.26548660e+00
-9.76941228e-01 -3.80977064e-01 6.81041852e-02 1.46410605e-02
2.39975080e-01 1.58665791e-01 -5.07496238e-01 -5.74433982e-01
-2.47487038e-01 -1.38663912e+00 2.15729281e-01 2.55767912e-01
-5.77104449e-01 -5.18140078e-01 1.33295321e+00 -2.43943885e-01
7.74320781e-01 -2.43168688e+00 -2.01870054e-01 -1.25787437e-01
3.66672762e-02 1.83924377e-01 -1.61581993e-01 2.00270385e-01
-1.26356736e-01 -6.22812033e-01 3.89906138e-01 -7.32644871e-02
6.37962073e-02 -4.32755910e-02 -3.79809827e-01 8.81326973e-01
-1.50726527e-01 4.53612477e-01 -9.86714840e-01 -3.28208774e-01
3.23767275e-01 1.58460900e-01 -1.52685314e-01 4.04729903e-01
-1.24522209e-01 3.25204790e-01 -3.95711094e-01 1.05311203e+00
4.56035167e-01 -6.43765554e-04 -3.20741646e-02 -2.87549019e-01
-1.55097932e-01 -9.64112952e-02 -1.37827373e+00 1.23362386e+00
-4.56060581e-02 8.21608424e-01 9.15667564e-02 -3.90816957e-01
7.32455909e-01 3.51831555e-01 3.96405280e-01 -1.05229259e-01
3.81648749e-01 7.40221143e-02 2.61784375e-01 -5.48497498e-01
6.84133768e-01 8.42423737e-01 -1.53551161e-01 3.16264123e-01
-1.74156398e-01 2.57783592e-01 1.80860996e-01 7.48357177e-03
1.68172324e+00 4.38853592e-01 1.30524054e-01 -4.75368164e-02
4.11305092e-02 2.55365014e-01 3.69274229e-01 1.07489753e+00
-6.66890979e-01 6.27361417e-01 3.44440907e-01 -2.40914688e-01
-7.44016767e-01 -1.06647587e+00 1.18473954e-01 1.07317948e+00
5.20578027e-01 -2.33491331e-01 -5.98366857e-01 -1.00559795e+00
3.02222967e-01 6.80549026e-01 -4.64386314e-01 -3.04220557e-01
-4.42337453e-01 -1.42118022e-01 4.81533855e-01 6.94470406e-01
2.06114292e-01 -1.32988763e+00 -1.44085026e+00 3.12712453e-02
3.74826603e-03 -1.27715445e+00 -8.25165212e-01 3.92708838e-01
-3.59014243e-01 -1.57205379e+00 -5.44612885e-01 -6.22143626e-01
7.69244850e-01 6.34068370e-01 5.38521111e-01 -7.03176335e-02
-5.30477345e-01 7.98290730e-01 -5.31765342e-01 -5.64733267e-01
-5.55955946e-01 -1.95555001e-01 5.45618176e-01 -2.44451940e-01
4.93092567e-01 1.96822524e-01 -5.44658303e-01 1.02283216e+00
-2.75680602e-01 -3.51479113e-01 6.42178297e-01 5.29714108e-01
1.57803789e-01 2.87066668e-01 2.31017143e-01 -1.83711916e-01
2.79899687e-01 -3.26779187e-01 -1.03053093e+00 1.87305599e-01
-2.18958393e-01 -1.08703077e-01 3.56724054e-01 -1.05840039e+00
-8.35497439e-01 5.02335429e-01 5.68877637e-01 -1.02783835e+00
-3.62741649e-01 -1.97186291e-01 -1.89844728e-01 1.38539091e-01
7.07048655e-01 -5.08985296e-02 -1.64903458e-02 -3.91271889e-01
1.01864465e-01 1.07540715e+00 9.21844721e-01 -5.76156616e-01
7.44042277e-01 3.22413802e-01 -2.77582109e-01 -4.38087821e-01
-3.61034960e-01 -7.54562795e-01 -5.66620171e-01 -7.45190144e-01
5.94574988e-01 -9.62342143e-01 -1.47341573e+00 8.04188192e-01
-1.41828704e+00 -7.07766712e-01 -1.21604301e-01 7.32533097e-01
-5.98145723e-01 2.11111262e-01 -5.27228951e-01 -1.05360866e+00
4.49872343e-03 -1.36611283e+00 1.32039702e+00 1.17242143e-01
-2.40431175e-01 2.29080152e-02 -3.29752028e-01 1.66033328e-01
-9.39394534e-02 5.41602112e-02 -6.25625858e-03 -6.85735285e-01
-8.10580432e-01 -4.62676525e-01 -2.56244466e-02 1.68242559e-01
3.33776534e-01 7.79703483e-02 -7.83298016e-01 -4.74890500e-01
-2.02532336e-01 -6.76546097e-02 4.39432859e-01 3.24473642e-02
1.00093627e+00 -8.83444548e-02 -7.92536378e-01 6.23533353e-02
8.97266269e-01 6.33628726e-01 5.00741839e-01 5.00734925e-01
6.30724728e-01 2.44947612e-01 1.48277819e+00 5.12335360e-01
8.46064240e-02 1.02321196e+00 6.22431159e-01 4.92708623e-01
-1.74396977e-01 6.99842200e-02 1.09852982e+00 -1.31204084e-01
-1.12764612e-01 -5.27644277e-01 -7.82333136e-01 2.99999714e-01
-2.07179880e+00 -8.98767531e-01 -2.33995035e-01 2.16003585e+00
4.92395073e-01 3.24068755e-01 3.95271212e-01 -5.22811823e-02
1.01175809e+00 -3.64466220e-01 -7.45728970e-01 2.77217984e-01
3.92607301e-01 -6.17769241e-01 8.92043173e-01 9.16790776e-03
-1.30766976e+00 9.52140570e-01 5.48171902e+00 5.86929023e-01
-9.11958277e-01 -6.98553631e-03 -8.41946155e-03 -1.71053588e-01
1.08872187e+00 -3.75890821e-01 -1.15322101e+00 7.39790857e-01
5.37054598e-01 3.19725931e-01 3.69851649e-01 1.50017142e+00
1.08797275e-01 -5.81715703e-01 -1.60885239e+00 1.09792149e+00
5.87725006e-02 -5.37759185e-01 -7.11720943e-01 1.09111689e-01
5.41017503e-02 1.48391560e-01 -1.95252604e-03 4.76947486e-01
4.27424014e-01 -6.39923334e-01 1.43815994e+00 4.42040205e-01
5.29546320e-01 -2.59954363e-01 7.52943933e-01 3.99208069e-01
-1.21590149e+00 -3.54020506e-01 -2.78609097e-01 -1.19762300e-02
5.11264950e-02 1.01111129e-01 -1.28324807e+00 1.36321142e-01
1.20180976e+00 4.65976566e-01 -4.51057255e-01 1.11218202e+00
-3.76623362e-01 2.40506276e-01 -3.47007930e-01 -1.88664235e-02
-2.45736867e-01 2.27017298e-01 1.01021755e+00 1.02425659e+00
3.59505743e-01 9.47947279e-02 4.58739042e-01 7.69292355e-01
2.75676996e-01 -4.47086096e-01 -6.94197059e-01 2.35474661e-01
6.49472952e-01 1.16666949e+00 -8.73622954e-01 -3.64037871e-01
-2.35528946e-01 1.17386830e+00 7.93875605e-02 2.80168712e-01
-1.19230103e+00 -4.48279798e-01 6.41701400e-01 -6.46084398e-02
7.33249247e-01 -2.94857979e-01 3.02986681e-01 -9.08531249e-01
4.43249375e-01 -9.64508593e-01 2.37643778e-01 -1.19722784e+00
-1.13247263e+00 5.65490603e-01 1.54439509e-01 -1.68515909e+00
-6.67408854e-02 -1.23926520e+00 -2.58520722e-01 3.11943799e-01
-7.78316081e-01 -7.80312121e-01 -7.18959033e-01 4.43995744e-01
6.80297434e-01 -4.58735555e-01 3.64132762e-01 1.43576577e-01
-5.62617540e-01 4.92674768e-01 -1.43420950e-01 4.75959361e-01
1.04881716e+00 -1.09090745e+00 1.95314229e-01 6.58663690e-01
-2.86602795e-01 3.85037541e-01 8.62482011e-01 -8.81137788e-01
-1.76793242e+00 -1.17647076e+00 -1.41795591e-01 -1.01229632e+00
6.17813170e-01 -4.58001763e-01 -6.78223014e-01 1.12876225e+00
-5.75840287e-02 3.63361955e-01 -2.66155452e-02 -1.95129856e-01
-2.46897295e-01 -5.57490923e-02 -1.03239334e+00 6.08281493e-01
1.16114783e+00 -5.86247258e-02 -8.16503108e-01 4.88417655e-01
5.35456002e-01 -7.70490229e-01 -8.47692788e-01 3.75268906e-01
8.93719435e-01 -8.00760329e-01 7.95950174e-01 -2.15366244e-01
-5.81883825e-02 -9.35366154e-01 -2.36280411e-01 -1.14942098e+00
-1.80058002e-01 -5.93920052e-01 -2.69654125e-01 1.05939925e+00
-8.40865448e-02 -5.47962666e-01 6.91403687e-01 3.73190075e-01
-1.97652265e-01 -3.97579908e-01 -8.65621865e-01 -1.44465041e+00
-6.91282153e-01 -4.49256748e-01 4.34379458e-01 3.79413098e-01
1.17260464e-01 -3.59940976e-02 -3.51618439e-01 5.20369589e-01
6.53197706e-01 -1.86569020e-01 1.33629382e+00 -8.84302080e-01
-1.16355084e-01 -2.26115301e-01 -6.69032097e-01 -1.16645646e+00
8.52073058e-02 -4.62445855e-01 8.25813532e-01 -9.70484674e-01
1.23936631e-01 -3.91100824e-01 -8.64918083e-02 5.91418028e-01
7.83250928e-02 2.83987764e-02 3.90953213e-01 3.08636278e-01
-1.01500773e+00 4.10602003e-01 8.68537784e-01 -4.05670881e-01
-1.19306423e-01 2.78590828e-01 5.58044426e-02 7.63807833e-01
7.96199441e-01 -7.55229235e-01 3.34144570e-02 -2.22389493e-02
-2.93540537e-01 1.15544751e-01 7.96838999e-01 -1.38801908e+00
4.05423999e-01 -1.26932904e-01 4.42463756e-01 -9.18527663e-01
3.42907965e-01 -1.35410666e+00 2.44872123e-01 7.56825805e-01
-7.96443820e-02 1.52599558e-01 7.58839026e-02 6.21174634e-01
3.49674106e-01 -3.16863745e-01 5.92862785e-01 4.29821946e-02
-9.57976520e-01 7.46711195e-02 -7.56111741e-01 -4.94055331e-01
1.53923488e+00 -2.83175975e-01 -6.91121161e-01 -2.32406817e-02
-5.03574491e-01 4.10688967e-01 6.49753213e-01 9.06925380e-01
5.77614963e-01 -1.19303906e+00 -3.62405270e-01 7.91657809e-03
5.22855103e-01 -4.35132533e-02 -9.06514078e-02 9.02676582e-01
-3.97251099e-01 2.62615860e-01 -1.13463409e-01 -1.05027914e+00
-1.55934346e+00 9.16818023e-01 3.45052928e-01 2.56458223e-01
-9.46940184e-01 4.10530627e-01 2.84101576e-01 -3.72406811e-01
7.06239939e-01 -6.94521725e-01 -6.59234682e-03 -3.30508649e-01
6.23402536e-01 6.65563166e-01 -1.85609996e-01 -5.59172273e-01
-7.11316109e-01 2.48542354e-01 -2.92607136e-02 1.22337326e-01
7.98866332e-01 1.16141520e-01 2.07736641e-01 4.93101478e-01
7.54317641e-01 -7.37065598e-02 -1.74437737e+00 3.40425745e-02
7.81681761e-02 -7.54925311e-01 -3.27331126e-01 -9.30003762e-01
-1.00307035e+00 2.17839241e-01 8.64522517e-01 -1.78606883e-01
6.12183928e-01 3.86352956e-01 3.56870949e-01 8.51911664e-01
9.15418208e-01 -1.02682042e+00 5.14000893e-01 5.02967954e-01
1.11950874e+00 -1.32610178e+00 1.81454383e-02 -3.78421366e-01
-7.38515258e-01 1.15807831e+00 1.19442070e+00 -2.57939193e-02
3.62144768e-01 6.02146566e-01 1.40047535e-01 -2.05929786e-01
-6.30399883e-01 -5.40319132e-03 -1.00668240e-03 7.63727069e-01
-2.73339361e-01 2.27397993e-01 5.14031053e-01 5.20767272e-01
-1.20616883e-01 1.58591509e-01 8.01116407e-01 1.24171984e+00
-4.70104575e-01 -5.54001987e-01 -9.15980041e-01 1.90745398e-01
-2.01832995e-01 4.57847893e-01 -1.71624839e-01 1.02769649e+00
-2.46424861e-02 1.11713028e+00 1.43317685e-01 -6.31984830e-01
6.54165924e-01 -2.19659671e-01 5.16181231e-01 -5.68739295e-01
-3.14812124e-01 -9.10583511e-03 -6.37283828e-03 -8.01499307e-01
-2.42062703e-01 -1.01705265e+00 -1.32037067e+00 3.75681091e-03
-5.08166313e-01 -2.43998840e-01 5.47024906e-01 6.56352937e-01
3.19792032e-01 7.01358795e-01 3.89488935e-01 -1.15536332e+00
-7.96250761e-01 -1.03937781e+00 -6.42510951e-01 1.22842096e-01
4.27186370e-01 -1.29803765e+00 -2.96434850e-01 5.17397933e-02] | [6.448986053466797, -1.1267964839935303] |
15ba72f4-a173-4df4-975b-11f8f15f01f2 | image-steganography-based-on-iteratively | 2101.05209 | null | https://arxiv.org/abs/2101.05209v1 | https://arxiv.org/pdf/2101.05209v1.pdf | Image Steganography based on Iteratively Adversarial Samples of A Synchronized-directions Sub-image | Nowadays a steganography has to face challenges of both feature based staganalysis and convolutional neural network (CNN) based steganalysis. In this paper, we present a novel steganography scheme denoted as ITE-SYN (based on ITEratively adversarial perturbations onto a SYNchronized-directions sub-image), by which security data is embedded with synchronizing modification directions to enhance security and then iteratively increased perturbations are added onto a sub-image to reduce loss with cover class label of the target CNN classifier. Firstly an exist steganographic function is employed to compute initial costs. Then the cover image is decomposed into some non-overlapped sub-images. After each sub-image is embedded, costs will be adjusted following clustering modification directions profile. And then the next sub-image will be embedded with adjusted costs until all secret data has been embedded. If the target CNN classifier does not discriminate the stego image as a cover image, based on adjusted costs, we change costs with adversarial manners according to signs of gradients back-propagated from the CNN classifier. And then a sub-image is chosen to be re-embedded with changed costs. Adversarial intensity will be iteratively increased until the adversarial stego image can fool the target CNN classifier. Experiments demonstrate that the proposed method effectively enhances security to counter both conventional feature-based classifiers and CNN classifiers, even other non-target CNN classifiers. | ['Jiwu Huang', 'Weixuan Tang', 'Bin Li', 'Shunquan Tan', 'Xinghong Qin'] | 2021-01-13 | null | null | null | null | ['steganalysis', 'image-steganography'] | ['computer-vision', 'computer-vision'] | [ 9.6573603e-01 2.3387757e-01 1.1756521e-01 1.4794651e-01
-2.9071929e-02 -7.1409172e-01 2.9870453e-01 -4.0500885e-01
-4.0117258e-01 5.1261657e-01 -3.5508761e-01 -2.4281046e-01
3.1823003e-01 -1.0786027e+00 -6.7780548e-01 -1.0613914e+00
-1.6404410e-01 -2.2923100e-01 5.6985396e-01 -5.6866413e-01
3.3243307e-01 5.6464553e-01 -1.5455540e+00 4.2028058e-01
4.2732960e-01 9.0457314e-01 -1.6714770e-01 9.6273977e-01
3.0199653e-01 5.7534909e-01 -7.4989730e-01 -1.9431776e-01
6.6148162e-01 -9.5250309e-01 -6.3690978e-01 1.9878791e-01
-1.6967152e-01 -4.3785259e-02 -4.2993993e-01 1.6025975e+00
2.0892109e-01 -2.7882969e-01 4.4552219e-01 -1.4738990e+00
-3.9937145e-01 6.1995018e-01 -3.5535750e-01 1.1724694e-01
-1.1628274e-01 4.3203369e-01 3.1255260e-01 -2.8320974e-01
5.3275400e-01 1.0509057e+00 4.6493888e-01 8.2746655e-01
-8.6689138e-01 -1.2154589e+00 -4.0252853e-02 2.6864791e-01
-1.2079538e+00 -1.5853168e-01 1.0052484e+00 1.5855424e-01
5.4890001e-01 6.0218823e-01 1.0277634e+00 5.8965647e-01
5.6575692e-01 2.0345715e-01 1.3109605e+00 -5.0340033e-01
-8.8444859e-04 2.7597442e-01 -4.4414988e-01 8.7851298e-01
4.3982410e-01 5.1759034e-01 1.5235057e-01 6.1059594e-03
4.5409939e-01 1.2131760e-01 -3.4298900e-01 1.5253289e-01
-1.0499287e+00 1.0048318e+00 6.5562427e-01 6.2453508e-01
-9.7605646e-02 3.6954471e-01 3.2915244e-01 8.3445221e-01
3.7698165e-02 2.0653421e-01 -2.7930450e-01 5.4220182e-01
-6.7822886e-01 -1.1290916e-01 7.8408694e-01 7.4580860e-01
1.0950102e+00 3.1635821e-01 4.6526051e-01 8.6092182e-02
2.3540957e-01 6.6337615e-01 6.2994945e-01 -4.0554035e-01
2.2711062e-01 7.7872145e-01 -5.4933697e-01 -1.4459436e+00
-1.3693309e-01 -2.7710816e-01 -1.1895765e+00 6.6302127e-01
1.2449002e-01 -2.2655228e-01 -1.0717533e+00 1.4048131e+00
3.3862233e-01 5.5260855e-01 5.6528187e-01 6.0712415e-01
5.8247763e-01 7.4725980e-01 -2.6575038e-01 -2.6418230e-01
1.3939041e+00 -6.8193412e-01 -4.4489655e-01 -1.2562914e-01
7.4802482e-01 -8.8446361e-01 2.9937983e-01 2.4318422e-01
-8.6942655e-01 -5.7012022e-01 -1.6251707e+00 8.0673110e-01
-7.0670944e-01 -4.1424084e-01 -1.0475998e-01 1.2008679e+00
-9.4746995e-01 5.1172960e-01 -5.0612569e-01 -3.3604894e-02
3.8173184e-01 7.4330389e-01 -3.5462216e-01 -1.1742677e-01
-1.6035727e+00 5.6137258e-01 9.7820461e-01 1.6197526e-01
-1.0333725e+00 -1.5565012e-01 -1.1118013e+00 -1.0877267e-01
3.9580844e-02 -2.1722284e-01 3.5819066e-01 -1.8674916e+00
-1.2735331e+00 7.8621489e-01 3.2738835e-01 -4.7128958e-01
4.2407227e-01 9.6799725e-01 -9.5528102e-01 2.4205363e-01
-2.6480114e-01 5.6803209e-01 1.6819088e+00 -1.2671098e+00
-9.9631429e-01 -1.2648040e-01 -5.4708999e-02 1.7240854e-01
-3.6506972e-01 -3.4300238e-02 -1.5164857e-01 -5.5304110e-01
2.6817834e-01 -1.3283131e+00 -1.8934333e-01 -3.0557987e-01
-6.0274607e-01 5.2919590e-01 1.5658407e+00 -3.9559871e-01
1.1677209e+00 -2.1512432e+00 -2.3629213e-02 8.2454318e-01
3.1355616e-01 7.7308547e-01 -1.6325636e-01 1.7388791e-01
-4.4999203e-01 5.2482456e-01 -6.2384140e-01 4.7727045e-01
-6.2714541e-01 3.9817214e-02 1.0229866e-01 8.0320662e-01
1.7255460e-01 7.5567788e-01 -7.4985766e-01 -4.6851236e-01
2.6090422e-01 5.0252211e-01 -3.8581091e-01 -6.4904615e-02
1.4135934e-01 3.6838707e-01 -2.7959517e-01 5.4849637e-01
1.2741071e+00 1.7419927e-02 4.1130403e-01 -1.5250167e-01
7.1476951e-02 -5.3745544e-01 -1.1116277e+00 5.9522885e-01
-6.7562945e-02 6.2244511e-01 6.6010199e-02 -1.3316326e+00
1.0552669e+00 3.6972842e-01 1.4063856e-01 -2.5651908e-01
5.7434541e-01 3.7512434e-01 3.9243773e-01 -5.3584898e-01
1.4706726e-01 2.6224335e-03 -1.6174544e-01 5.4042667e-01
-3.3221835e-01 -2.0715223e-01 -3.7907743e-01 -1.6842889e-02
1.0636296e+00 -3.5062626e-01 2.1694206e-01 -1.9369331e-01
1.2874796e+00 1.7418508e-01 2.8561300e-01 5.3269047e-01
-3.5522696e-01 3.0701983e-01 3.0484077e-01 -4.5788234e-01
-1.2570029e+00 -2.9877529e-01 2.2181661e-01 6.0452282e-01
4.9775094e-01 2.8430286e-01 -1.1383411e+00 -1.1883025e+00
-2.2213414e-01 -3.7554208e-02 -7.4029434e-01 -6.1613345e-01
-7.4361718e-01 -6.7052335e-01 1.0922669e+00 -2.4539044e-01
1.3784788e+00 -1.4254179e+00 -6.4647704e-01 2.9513058e-01
1.4143157e-01 -8.0551684e-01 -3.2233623e-01 6.2135994e-02
-6.5630627e-01 -1.4577143e+00 -4.9481559e-01 -1.3698765e+00
1.2054708e+00 3.0468771e-01 2.0157117e-01 9.3295026e-01
-1.7317790e-01 -2.6733324e-01 -5.3442204e-01 -2.4205180e-01
-1.0370115e+00 -8.6861588e-02 -1.7665559e-01 3.5713714e-01
1.3897614e-01 -3.4253719e-01 -6.6739011e-01 4.3443742e-01
-1.3555027e+00 -1.6397209e-01 5.6905162e-01 9.4112086e-01
2.0852575e-01 6.6002691e-01 2.9339540e-01 -1.0845376e+00
1.8485855e-01 -4.9581537e-01 -4.8758677e-01 2.3284070e-02
-7.0579422e-01 -8.2496829e-02 8.8604641e-01 -6.8799508e-01
-6.1383259e-01 -4.3856621e-02 4.3430567e-02 -4.7849318e-01
-5.7612181e-02 3.5024068e-01 -2.7595186e-01 -7.9950458e-01
5.0753641e-01 6.0704118e-01 4.6838504e-01 4.1685808e-01
-2.1780813e-03 7.5875241e-01 4.4193983e-01 4.3358979e-01
1.4796020e+00 7.0246154e-01 2.8878146e-01 -5.2805674e-01
5.7842918e-02 1.9711791e-01 -2.3450524e-01 -2.8221759e-01
1.0001113e+00 -5.2854598e-01 -5.7688558e-01 1.0175294e+00
-9.7746831e-01 2.4579410e-02 2.6671839e-01 2.3110093e-01
-1.0590902e-02 4.7645384e-01 -4.6119192e-01 -7.0411557e-01
-3.6371529e-01 -1.3309146e+00 1.9954260e-01 3.5141701e-01
3.8155431e-01 -1.1039281e+00 -4.3888339e-01 -3.6390278e-02
4.6330559e-01 9.0482348e-01 7.6279598e-01 -9.5148438e-01
-5.2594519e-01 -6.6389507e-01 -7.6321647e-02 6.4842755e-01
3.0316797e-01 1.5588186e-02 -7.2302085e-01 -6.5711790e-01
3.4461257e-01 1.4369587e-02 7.7257967e-01 2.1417266e-02
9.3468642e-01 -8.0844927e-01 -4.0363190e-01 9.6315145e-01
1.6846924e+00 7.2458166e-01 9.7455418e-01 7.6984274e-01
8.6123043e-01 6.2636435e-01 3.9133465e-01 -6.3283190e-02
-7.7112898e-02 -2.2186881e-02 8.9265752e-01 -3.7749416e-01
1.8705845e-01 4.3640286e-02 4.5509133e-01 6.3508803e-01
3.3080917e-02 -4.3624958e-01 -3.7746522e-01 2.3549376e-01
-1.3862885e+00 -1.1303964e+00 -3.2751915e-01 1.9268568e+00
5.3438091e-01 3.7041056e-01 -3.7647611e-01 7.8899062e-01
1.3802811e+00 4.1756359e-01 -4.5150271e-01 -5.9634268e-01
-2.7350223e-01 2.6928475e-01 9.8330235e-01 4.5725667e-01
-1.0530026e+00 1.0912744e+00 5.0326233e+00 1.0520368e+00
-1.5016052e+00 3.2260332e-02 6.5715212e-01 4.8567015e-01
-3.1683078e-01 3.0847055e-01 -5.4533947e-01 7.1278191e-01
6.6414690e-01 1.2147081e-01 5.2200717e-01 4.4811738e-01
-2.1625413e-01 5.5191386e-03 -4.6515876e-01 7.0337921e-01
4.2888799e-01 -1.1447772e+00 7.7911504e-02 2.9194999e-01
9.4441557e-01 -6.0477394e-01 3.4368929e-01 2.5560779e-02
1.5173841e-01 -9.0445930e-01 4.3269300e-01 8.8243736e-03
1.2447718e+00 -1.1901286e+00 1.0048490e+00 3.1303072e-01
-1.3854765e+00 -2.2412835e-01 -3.2839137e-01 1.1829398e-01
-3.9527175e-01 4.0558759e-02 -6.8898392e-01 5.3313500e-01
5.4117501e-01 3.4653997e-01 -2.4696194e-01 3.0633551e-01
-3.3921334e-01 5.4925978e-01 -1.7546037e-01 -9.3165934e-02
6.3908172e-01 2.4405394e-02 7.9153496e-01 1.0703279e+00
3.8685754e-01 4.0329537e-01 -1.6897246e-01 4.3284476e-01
-3.1686895e-02 -1.6548660e-01 -9.2003852e-01 2.1162167e-01
6.4042979e-01 1.2481387e+00 -1.1046909e+00 -8.1641567e-01
-1.3454148e-01 1.2030259e+00 -5.1749551e-01 2.3443271e-01
-7.4885589e-01 -9.6794683e-01 2.7709448e-01 -5.6747463e-02
6.8373173e-01 3.3743384e-01 -2.2744657e-02 -8.2253069e-01
-4.9839306e-01 -1.1962960e+00 3.8882765e-01 -3.5113335e-01
-6.1599702e-01 8.7225068e-01 -2.1383777e-01 -1.7015433e+00
4.8436239e-02 -2.5805479e-01 -9.4525599e-01 6.4820784e-01
-1.6270074e+00 -1.4117239e+00 -4.6416196e-01 1.1474018e+00
2.4632388e-01 -6.0016888e-01 6.7513299e-01 -2.0240127e-01
-4.8344454e-01 8.0606222e-01 -3.7476506e-02 2.6432070e-01
2.7268204e-01 -5.6585640e-01 1.6726844e-01 1.2160043e+00
-6.0034412e-01 2.2280210e-01 6.0699552e-01 -9.3757606e-01
-9.6831226e-01 -1.4546533e+00 7.4372035e-01 7.5768165e-02
3.9517185e-01 -1.9859901e-01 -7.2500902e-01 6.7617875e-01
3.0591869e-01 1.1755463e-01 3.4720522e-01 -1.2931517e+00
-3.4367999e-01 5.9980601e-02 -1.9478449e+00 4.6493617e-01
6.4420527e-01 -3.3312884e-01 -1.1692414e-01 3.4717366e-02
7.2693151e-01 -4.5353347e-01 -6.0573024e-01 4.4199228e-01
4.4585213e-01 -9.1592020e-01 7.2349972e-01 -3.2729837e-01
2.9306632e-01 -6.4357275e-01 6.3365169e-02 -1.1032999e+00
-3.2277203e-01 -9.2175853e-01 3.6495355e-01 7.8685802e-01
3.0651841e-01 -1.0602782e+00 8.4080374e-01 -1.4742850e-01
4.2839903e-02 -4.3338725e-01 -7.8891563e-01 -5.8237344e-01
-1.9955398e-01 -9.4685361e-02 9.4432664e-01 1.3579544e+00
-5.4114863e-02 -4.6126616e-01 -4.8104751e-01 5.4256880e-01
7.9942000e-01 -3.9480585e-01 5.6325960e-01 -7.6744831e-01
1.7457379e-01 -2.9696727e-01 -8.4674418e-01 -2.8670940e-01
9.2954403e-03 -9.1194284e-01 -1.6986988e-01 -6.3596600e-01
-1.5875447e-01 -4.9891835e-01 -4.0989128e-01 4.8945618e-01
-1.5964210e-01 1.0341767e+00 1.8755111e-01 3.8896498e-01
-2.8982636e-02 -2.5766036e-01 1.3289977e+00 -2.9578221e-01
-2.3732539e-01 -2.3637993e-02 -4.6025702e-01 5.9329200e-01
1.0537858e+00 -9.0980250e-01 -1.9293334e-01 3.4744599e-01
-4.9474590e-02 1.4128910e-01 3.2906297e-01 -1.2469767e+00
1.8186127e-01 -8.6271197e-02 6.6458261e-01 -2.9722315e-01
-2.1371512e-01 -1.2406656e+00 5.0282049e-01 1.5674187e+00
-7.6775834e-02 5.0633878e-02 3.4282051e-03 5.4046887e-01
-5.9080835e-02 -3.4955272e-01 1.2326199e+00 -2.5597090e-01
-6.5171731e-01 4.0323725e-01 -8.0497944e-01 -4.9562359e-01
1.5551907e+00 -9.8964179e-01 -9.8386645e-02 -2.7722096e-01
-7.3325843e-01 -2.6842363e-02 5.8688939e-01 1.1641274e-01
9.8704237e-01 -1.4242035e+00 -7.3731720e-01 8.1675899e-01
4.5317635e-02 -3.4007126e-01 2.6957703e-01 6.2756723e-01
-9.0259427e-01 -7.7046618e-02 -4.8305386e-01 -4.5509231e-01
-1.6729070e+00 8.0078262e-01 5.0649720e-01 -4.1151217e-01
-3.6006713e-01 9.7891301e-01 -4.3121368e-02 -3.6151677e-02
-1.2691426e-01 -4.0479146e-02 -5.4490668e-01 -1.4429708e-01
4.7866476e-01 2.7687523e-01 -7.4570984e-02 -9.4515383e-01
-2.0965990e-01 7.7849907e-01 -8.8605814e-02 -2.4164446e-01
9.0361607e-01 -1.9167027e-01 -3.3173570e-01 -3.5892031e-01
1.7185173e+00 -1.4138548e-01 -9.6184450e-01 -3.1288281e-02
-5.4954791e-01 -4.8902255e-01 -1.4005975e-01 -3.6646944e-01
-1.6324829e+00 4.2358389e-01 9.0589094e-01 5.4295576e-01
1.5340881e+00 -4.8102781e-01 1.0847425e+00 6.3064180e-02
1.7685995e-01 -8.5911167e-01 2.4437359e-01 4.4986135e-01
3.5209247e-01 -1.0025152e+00 -2.4795863e-01 -2.1822938e-01
-7.2040713e-01 1.2540561e+00 5.7824075e-01 -8.0084854e-01
8.6825466e-01 3.5448834e-01 8.0514558e-02 -3.3257926e-01
-3.1405577e-01 2.4812264e-02 -2.4882957e-02 9.3345881e-01
-5.5132473e-01 -1.8062705e-01 -3.0404037e-01 -5.8273770e-02
-3.9547828e-01 -3.1411874e-01 7.0415586e-01 1.1288118e+00
-7.1494108e-01 -1.0227302e+00 -9.2272991e-01 1.2166824e-01
-6.5318149e-01 -1.3517326e-01 -2.8272074e-01 8.7347639e-01
7.6664895e-01 1.0356387e+00 1.9644642e-01 -1.1142217e+00
-1.3379669e-01 -2.0210014e-01 1.6901152e-01 -3.4912044e-01
-1.0785897e+00 6.8082117e-02 -2.2338186e-01 -2.7002105e-01
-4.7538000e-01 -1.9515949e-01 -1.4245048e+00 -7.4128765e-01
-6.7999667e-01 2.2365397e-01 6.6146684e-01 8.1567568e-01
-1.4782891e-01 4.4915763e-01 1.4307884e+00 -7.3531824e-01
-1.8349575e-01 -8.3650601e-01 -3.7514415e-01 3.1402448e-01
7.3032504e-01 -1.4521401e-02 -1.0804662e+00 3.9176086e-01] | [4.293829441070557, 8.056559562683105] |
ddc7c3fa-b378-42d8-a20e-3cb1ea359aa1 | skeleton-based-zero-shot-action-recognition | 1911.11344 | null | https://arxiv.org/abs/1911.11344v1 | https://arxiv.org/pdf/1911.11344v1.pdf | Skeleton based Zero Shot Action Recognition in Joint Pose-Language Semantic Space | How does one represent an action? How does one describe an action that we have never seen before? Such questions are addressed by the Zero Shot Learning paradigm, where a model is trained on only a subset of classes and is evaluated on its ability to correctly classify an example from a class it has never seen before. In this work, we present a body pose based zero shot action recognition network and demonstrate its performance on the NTU RGB-D dataset. Our model learns to jointly encapsulate visual similarities based on pose features of the action performer as well as similarities in the natural language descriptions of the unseen action class names. We demonstrate how this pose-language semantic space encodes knowledge which allows our model to correctly predict actions not seen during training. | ['Afshaan Mazagonwalla', 'Bhavan Jasani'] | 2019-11-26 | null | null | null | null | ['zero-shot-action-recognition'] | ['computer-vision'] | [ 5.04147410e-01 2.05684960e-01 -4.75144029e-01 -5.56516469e-01
-4.06192631e-01 -3.49771023e-01 8.60289633e-01 -1.40287643e-02
-4.13938522e-01 4.37390625e-01 7.51078010e-01 4.74115223e-01
3.16707306e-02 -8.03148210e-01 -7.36565173e-01 -4.86403167e-01
-4.97771688e-02 9.32077587e-01 4.18867409e-01 -3.15229565e-01
-4.00641300e-02 2.76692837e-01 -1.79359591e+00 7.56010294e-01
-4.53334972e-02 7.10377574e-01 1.58527822e-04 6.18494511e-01
9.85669270e-02 1.51424110e+00 -5.11655688e-01 -9.55000818e-02
2.79071093e-01 -6.51223063e-01 -9.59936082e-01 4.03003603e-01
8.32684159e-01 -8.13814878e-01 -6.85049593e-01 8.30637038e-01
8.63436460e-02 5.62754869e-01 9.91483748e-01 -1.49633873e+00
-4.28897232e-01 2.10206956e-01 2.10609250e-02 2.36696199e-01
6.86758399e-01 4.81083840e-01 1.05419230e+00 -3.86068553e-01
1.04976153e+00 1.21585977e+00 2.80022085e-01 8.61141384e-01
-1.24029720e+00 -2.82178342e-01 4.91618626e-02 5.67193389e-01
-1.06696773e+00 -3.21608484e-01 6.73620284e-01 -6.80203438e-01
1.28293335e+00 -2.11249620e-01 9.73429322e-01 1.34021366e+00
2.90772498e-01 9.99883473e-01 6.02550626e-01 -2.62477219e-01
5.96388578e-01 -4.78891850e-01 1.71135128e-01 6.98546112e-01
-1.50754377e-02 4.30398554e-01 -7.87116230e-01 4.44543883e-02
6.89813793e-01 2.91923761e-01 2.75669247e-02 -1.08403790e+00
-1.21440625e+00 7.99030423e-01 5.66387534e-01 4.03869331e-01
-3.87532204e-01 7.04551458e-01 3.66265506e-01 1.02604188e-01
1.43155158e-01 3.77549529e-01 -3.30545425e-01 -3.66968423e-01
-4.00162965e-01 2.11886331e-01 6.00433171e-01 8.73370826e-01
7.97940314e-01 8.56541842e-02 -2.65032500e-01 4.61755097e-01
2.49046404e-02 4.38091516e-01 7.26469696e-01 -1.21556890e+00
1.26844436e-01 7.71042049e-01 3.44070904e-02 -6.51303709e-01
-3.05076897e-01 4.50124480e-02 -6.55423179e-02 5.23788273e-01
2.90219188e-01 4.20269132e-01 -1.35781753e+00 1.85621333e+00
2.59288132e-01 3.84629697e-01 3.26321989e-01 1.01946425e+00
8.47165525e-01 1.61932692e-01 3.85635018e-01 3.31945896e-01
1.16302669e+00 -6.46706462e-01 -5.77019155e-01 -6.98300004e-01
1.02085805e+00 -6.91953227e-02 9.08847511e-01 2.87286900e-02
-5.46206355e-01 -7.18434930e-01 -1.15557277e+00 -1.94192678e-01
-5.02783656e-01 -9.36385766e-02 4.76142645e-01 2.85740972e-01
-6.37949646e-01 9.31656897e-01 -1.05259085e+00 -7.07830489e-01
4.45998281e-01 2.88317382e-01 -8.06915879e-01 -7.54833966e-02
-1.07289743e+00 1.14348781e+00 6.76947713e-01 -4.33399618e-01
-1.57150066e+00 -2.48744965e-01 -1.40234220e+00 -1.29320681e-01
6.44785166e-01 -6.24661148e-01 1.56762707e+00 -1.14813840e+00
-1.21315277e+00 1.07829499e+00 2.87514143e-02 -6.90683007e-01
7.69341215e-02 -1.75992474e-01 -4.02468234e-01 4.53801244e-01
3.57228577e-01 6.82019055e-01 7.19352841e-01 -9.84233499e-01
-4.99686122e-01 -7.42463171e-01 4.51644957e-01 2.57847428e-01
5.24206422e-02 -3.94479930e-01 5.87910833e-03 -4.39070284e-01
2.70885199e-01 -1.04751778e+00 -1.09106861e-01 4.21342671e-01
-9.90737900e-02 6.35329913e-03 5.62671900e-01 -3.42218518e-01
4.46267933e-01 -2.23360181e+00 1.69679835e-01 -1.21533252e-01
-6.21963255e-02 1.15078501e-01 -2.35994428e-01 4.03476804e-01
-2.06256181e-01 -2.39935204e-01 -9.43033323e-02 4.72399630e-02
1.85876116e-02 8.80007625e-01 -2.73667634e-01 6.03043079e-01
7.49084800e-02 9.26744163e-01 -1.28684604e+00 -4.03229684e-01
6.57452464e-01 4.08484370e-01 -4.88355726e-01 3.59149098e-01
-3.75411689e-01 3.95528853e-01 -4.46133852e-01 6.37638092e-01
-1.49909958e-01 -1.38637647e-01 2.39026964e-01 -3.89046758e-01
3.30095619e-01 3.55141461e-02 -9.83513355e-01 1.89708197e+00
-6.68878108e-02 2.62955427e-01 -7.19953656e-01 -8.49775076e-01
6.33872986e-01 4.93689686e-01 5.38477123e-01 -8.24483752e-01
2.24957168e-01 -1.20219328e-01 8.65745395e-02 -8.20361614e-01
-4.30081300e-02 -6.36526346e-01 -3.25922281e-01 5.30975640e-01
6.41448796e-01 -1.14580560e-02 4.16772366e-02 2.54679084e-01
1.60256529e+00 6.77010953e-01 7.53889978e-01 2.86839455e-01
1.26789182e-01 3.89214993e-01 5.80492556e-01 7.59677231e-01
-6.95296407e-01 6.92411423e-01 2.45476723e-01 -7.99265265e-01
-9.94520068e-01 -1.40611970e+00 2.14144170e-01 1.29755461e+00
2.40741253e-01 -5.92225909e-01 -3.51475894e-01 -8.35808098e-01
1.54572159e-01 9.72926438e-01 -1.22550797e+00 -5.41524827e-01
-3.83543849e-01 3.01693648e-01 3.80968004e-01 9.64069486e-01
3.90381813e-01 -1.20506477e+00 -1.20451427e+00 -1.35614732e-02
6.45601526e-02 -1.05957687e+00 -6.79200813e-02 3.00741017e-01
-7.82800436e-01 -1.44112504e+00 -5.59258342e-01 -6.65225625e-01
6.18838310e-01 6.39072433e-02 1.14496958e+00 -1.60178602e-01
-6.19145095e-01 1.10037589e+00 -7.08355665e-01 -3.12157273e-01
-4.27236497e-01 -6.53490305e-01 2.14639321e-01 1.75257906e-01
8.82534206e-01 -4.43537921e-01 -5.12573421e-01 1.82469517e-01
-5.79084694e-01 -5.11961505e-02 4.95283782e-01 6.06636524e-01
7.55564094e-01 -4.44965690e-01 -8.30331966e-02 -6.71401143e-01
-6.01114072e-02 -4.33301359e-01 2.06309974e-01 4.80888009e-01
1.39254689e-01 2.95971751e-01 2.56012350e-01 -5.45443773e-01
-8.91409457e-01 7.13326752e-01 2.63092428e-01 -8.62166226e-01
-6.47480547e-01 1.01385176e-01 -2.28451386e-01 1.78741276e-01
8.34122837e-01 3.68434846e-01 1.10725053e-01 -4.25752908e-01
6.41073465e-01 2.97731727e-01 6.99204922e-01 -5.06181717e-01
6.53905094e-01 8.79621923e-01 9.37277675e-02 -7.43376911e-01
-1.08938050e+00 -7.13740289e-01 -1.24864948e+00 -4.83975142e-01
1.49917412e+00 -9.21606839e-01 -5.79942644e-01 1.58743829e-01
-9.22111034e-01 -3.96532238e-01 -7.42032230e-01 6.74630880e-01
-1.24549675e+00 1.93453893e-01 -2.04586819e-01 -7.09613204e-01
2.99786955e-01 -7.68484354e-01 1.29783404e+00 -8.57949257e-02
-7.59256780e-01 -7.51411498e-01 4.26930219e-01 2.37327874e-01
-1.51393935e-01 4.30046380e-01 8.24021578e-01 -9.73652422e-01
-3.77960563e-01 -2.38756448e-01 4.12990957e-01 2.62279838e-01
2.13678241e-01 -5.16012609e-01 -8.08404267e-01 -1.58414751e-01
3.09113320e-02 -8.14671278e-01 7.57428765e-01 1.37667403e-01
8.13646436e-01 -8.66812244e-02 -4.21499759e-01 2.96085685e-01
1.37490141e+00 3.53535980e-01 5.92816889e-01 2.00600326e-01
6.28608167e-01 5.79834580e-01 6.99024439e-01 3.37408900e-01
4.09913994e-02 7.38039136e-01 5.89883745e-01 3.12800437e-01
-2.47456610e-01 -7.16261208e-01 2.07029298e-01 -1.84999883e-01
-2.14767754e-01 8.08092952e-02 -9.76216495e-01 5.43290198e-01
-1.90948451e+00 -1.49997377e+00 2.86191463e-01 2.04944420e+00
5.60027242e-01 1.68779213e-02 1.59170717e-01 2.32927278e-02
3.65818202e-01 4.04877812e-01 -7.14689255e-01 -4.30059731e-01
1.78510576e-01 3.56659263e-01 2.86715031e-01 2.49700561e-01
-1.39016628e+00 9.92676377e-01 6.47831440e+00 2.78690875e-01
-6.26367033e-01 7.73780346e-02 2.81679649e-02 -3.34147543e-01
1.92193508e-01 2.12891832e-01 -4.01336610e-01 2.83645950e-02
7.71701932e-01 -3.23532522e-01 6.86981156e-02 1.16198635e+00
-1.75029755e-01 -2.64333725e-01 -1.79782736e+00 9.87422526e-01
6.81679964e-01 -1.14000916e+00 3.79900634e-01 -7.46598914e-02
3.59860450e-01 -4.78386395e-02 -2.51641542e-01 6.83891654e-01
6.00598633e-01 -1.14353454e+00 5.41360438e-01 7.71917284e-01
4.56947833e-01 -3.27429414e-01 2.75825322e-01 5.27098477e-01
-1.08062208e+00 -3.91687363e-01 -3.55143070e-01 -4.35690582e-01
1.91883624e-01 -5.11142731e-01 -9.51527178e-01 1.27701133e-01
4.82363969e-01 1.15826488e+00 -5.07812321e-01 9.54922080e-01
-6.71745539e-01 1.98790193e-01 1.01576909e-01 5.08354492e-02
2.87322968e-01 2.50970244e-01 5.23015678e-01 5.26968718e-01
1.19179174e-01 6.64324403e-01 4.75801229e-01 6.87839866e-01
3.01011056e-01 -1.12897143e-01 -1.05375624e+00 -8.88160169e-02
-1.11125059e-01 6.57820642e-01 -4.02126729e-01 -5.80466330e-01
-4.65180963e-01 1.29287398e+00 2.48315617e-01 2.77506590e-01
-5.74658513e-01 -4.40616481e-04 9.18554723e-01 1.80254593e-01
4.30607378e-01 -8.46801624e-02 2.30886802e-01 -1.14618552e+00
-2.65412718e-01 -4.90576744e-01 7.13168800e-01 -1.36483574e+00
-1.17186832e+00 2.95208365e-01 1.37724236e-01 -1.80590701e+00
-5.98983586e-01 -9.15989876e-01 -5.16622901e-01 2.50124961e-01
-8.21800470e-01 -1.20278513e+00 -2.92401999e-01 6.76462352e-01
7.49269545e-01 -3.62505615e-01 1.18494940e+00 -1.88447863e-01
6.24713451e-02 5.58572859e-02 -2.33335450e-01 5.81623554e-01
4.87299800e-01 -1.16074562e+00 7.35450983e-02 3.27691317e-01
7.11419642e-01 4.98201728e-01 9.05216038e-01 -9.23606336e-01
-1.12006712e+00 -7.91647196e-01 7.23930120e-01 -8.91537905e-01
6.24037504e-01 -9.60300937e-02 -7.73811877e-01 1.17163777e+00
-3.48330349e-01 3.37250322e-01 9.30324137e-01 -3.12653743e-02
-6.84303343e-01 1.13822214e-01 -9.94372845e-01 5.16288340e-01
1.33825862e+00 -7.82002091e-01 -1.41351891e+00 4.41786408e-01
3.89894664e-01 -2.86372006e-01 -5.75246811e-01 3.65959555e-01
6.42905474e-01 -8.88107419e-01 1.10749841e+00 -1.49857867e+00
5.76251328e-01 -2.63852507e-01 -5.91572523e-01 -1.28506899e+00
-2.89588690e-01 3.65141749e-01 -2.36807361e-01 5.00711381e-01
-2.98621189e-02 1.84546430e-02 9.13452327e-01 8.32400799e-01
-1.81103423e-02 -3.30096662e-01 -1.20740092e+00 -1.14739263e+00
-2.00213626e-01 -4.36922580e-01 3.74050140e-01 7.66808033e-01
1.91876829e-01 3.37416440e-01 -4.80467230e-01 -1.74728855e-01
8.23200703e-01 1.13812804e-01 9.46109414e-01 -1.23129702e+00
-4.76521313e-01 4.19280939e-02 -1.38303947e+00 -7.80842662e-01
4.07881916e-01 -1.01343310e+00 1.84918851e-01 -1.59939206e+00
5.47593057e-01 4.63508755e-01 -2.90586799e-01 9.90652442e-01
4.44248885e-01 3.24982494e-01 3.83837610e-01 3.11306417e-01
-8.66994083e-01 6.26943111e-01 1.08329093e+00 -4.30493742e-01
2.18412593e-01 -1.76283017e-01 -1.78836390e-01 9.65220869e-01
4.35239524e-01 -6.54125512e-01 -6.24301851e-01 -4.10226196e-01
-8.81523713e-02 2.33858630e-01 8.71385992e-01 -1.56216061e+00
1.36229381e-01 -3.27074468e-01 6.80780053e-01 -2.64877170e-01
8.52774620e-01 -1.00044942e+00 1.66828886e-01 6.83632851e-01
-7.48102963e-01 -4.67929214e-01 -1.46047249e-01 9.43869054e-01
-1.45016611e-01 -1.50225937e-01 5.38757026e-01 -5.90308905e-01
-1.63022017e+00 3.65855426e-01 -3.35584044e-01 1.28436089e-01
1.38944507e+00 -5.79388201e-01 -2.16800615e-01 -4.13216054e-01
-1.40569258e+00 4.31516506e-02 6.65236771e-01 5.85741460e-01
7.58675992e-01 -1.65493333e+00 -3.35700691e-01 2.87687510e-01
7.76565433e-01 -8.02206993e-01 3.42894256e-01 4.78582203e-01
-1.92986503e-01 2.69126207e-01 -6.76591992e-01 -5.63324153e-01
-1.30000138e+00 6.85229599e-01 6.80078268e-01 2.54822075e-01
-1.01677907e+00 6.68705404e-01 3.14234465e-01 -3.50812525e-01
2.14034528e-01 -1.19249940e-01 -2.40649924e-01 -2.09643483e-01
6.37792349e-01 1.25572100e-01 -4.39877838e-01 -1.34015739e+00
-7.20871866e-01 6.02011561e-01 1.58822224e-01 -3.26801866e-01
1.20904171e+00 4.04030263e-01 3.85258555e-01 9.85939085e-01
1.38865471e+00 -8.32135916e-01 -1.38053179e+00 -3.30369204e-01
-8.06531981e-02 -8.21313381e-01 -3.20868731e-01 -1.00702155e+00
-6.69100821e-01 9.69314873e-01 7.62499630e-01 -4.03554410e-01
5.86811364e-01 5.18292367e-01 5.85634649e-01 6.94019318e-01
8.27676952e-01 -1.35297465e+00 7.47043490e-01 5.84197402e-01
9.19433594e-01 -1.42299724e+00 5.47874384e-02 -8.18673335e-03
-1.11028910e+00 1.06446528e+00 7.39047408e-01 -4.30330992e-01
6.38164520e-01 -1.91779733e-01 -8.23235663e-04 -6.70807302e-01
-5.70850611e-01 -8.32836747e-01 3.35997790e-01 9.08148229e-01
1.56782582e-01 1.75774902e-01 5.00391759e-02 4.17890251e-01
-6.93647787e-02 3.04659456e-01 3.49230438e-01 1.38015640e+00
-7.74010420e-01 -8.07973027e-01 6.18877150e-02 4.24751341e-01
3.93541083e-02 5.22522688e-01 -7.63089061e-01 8.23520482e-01
2.64932245e-01 6.74394786e-01 3.68669927e-01 -6.58947349e-01
4.94392455e-01 4.39637393e-01 8.77310157e-01 -1.28186703e+00
-1.69968441e-01 -5.06255090e-01 1.70030311e-01 -1.17805576e+00
-5.28121769e-01 -7.28981256e-01 -1.68360686e+00 4.11066711e-01
3.96218479e-01 -2.54071027e-01 3.50930512e-01 1.28822529e+00
-2.03870935e-03 5.12144089e-01 4.97486964e-02 -7.06140578e-01
-6.77383184e-01 -5.64741611e-01 -7.56585121e-01 1.12642336e+00
2.74393439e-01 -1.24743545e+00 -4.71731424e-01 1.40002519e-01] | [8.465161323547363, 0.8416976928710938] |
6707f7a4-7132-4e5b-8171-aa56c54e46c6 | pareto-front-identification-with-regret | 2306.00096 | null | https://arxiv.org/abs/2306.00096v1 | https://arxiv.org/pdf/2306.00096v1.pdf | Pareto Front Identification with Regret Minimization | We consider Pareto front identification for linear bandits (PFILin) where the goal is to identify a set of arms whose reward vectors are not dominated by any of the others when the mean reward vector is a linear function of the context. PFILin includes the best arm identification problem and multi-objective active learning as special cases. The sample complexity of our proposed algorithm is $\tilde{O}(d/\Delta^2)$, where $d$ is the dimension of contexts and $\Delta$ is a measure of problem complexity. Our sample complexity is optimal up to a logarithmic factor. A novel feature of our algorithm is that it uses the contexts of all actions. In addition to efficiently identifying the Pareto front, our algorithm also guarantees $\tilde{O}(\sqrt{d/t})$ bound for instantaneous Pareto regret when the number of samples is larger than $\Omega(d\log dL)$ for $L$ dimensional vector rewards. By using the contexts of all arms, our proposed algorithm simultaneously provides efficient Pareto front identification and regret minimization. Numerical experiments demonstrate that the proposed algorithm successfully identifies the Pareto front while minimizing the regret. | ['Assaf Zeevi', 'Garud Iyengar', 'Wonyoung Kim'] | 2023-05-31 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 6.65823445e-02 -7.57191852e-02 -7.75191009e-01 -1.05590120e-01
-1.13511586e+00 -1.10452604e+00 -2.08265066e-01 2.17129290e-01
-6.05530322e-01 8.89027178e-01 -1.92747727e-01 -5.16603827e-01
-9.56209838e-01 -7.21541524e-01 -7.57850468e-01 -9.94431436e-01
-4.12545592e-01 7.80009449e-01 -2.35948265e-01 1.41196549e-01
4.55260366e-01 4.69824731e-01 -1.15708482e+00 -2.11067840e-01
1.19634390e+00 1.77084649e+00 4.02240967e-03 5.70911169e-01
-3.10274899e-01 6.41121566e-01 -5.14030874e-01 -4.38688964e-01
6.90143287e-01 -5.09912431e-01 -6.84551954e-01 1.11268088e-01
-1.09699488e-01 -3.57357860e-01 3.21348347e-02 1.21178508e+00
4.02672827e-01 3.43285263e-01 4.57268178e-01 -1.17262459e+00
-1.80752754e-01 7.42717087e-01 -9.33921933e-01 3.61923844e-01
-4.54800986e-02 -6.94638938e-02 1.30305505e+00 -3.27182740e-01
1.73934057e-01 1.08255768e+00 -7.37819122e-03 1.06032737e-01
-1.22589171e+00 -9.76886213e-01 6.65820062e-01 2.27665380e-02
-1.10125077e+00 -5.18702626e-01 7.54585624e-01 -3.12672675e-01
5.49453676e-01 6.45160437e-01 5.69245458e-01 2.56204814e-01
-1.65249154e-01 1.01294875e+00 7.96921015e-01 -5.92908859e-01
8.18001568e-01 -9.87149775e-02 3.12486827e-01 4.67698723e-01
5.16095102e-01 1.86652720e-01 -2.01514691e-01 -4.81634736e-01
5.20794153e-01 1.52774304e-01 -8.71691182e-02 -4.31897491e-01
-6.50547624e-01 1.07436216e+00 3.22475135e-02 -1.18687533e-01
-4.93662357e-01 3.85369420e-01 7.04536736e-02 4.09825325e-01
4.27041709e-01 5.68319619e-01 -3.26513439e-01 -3.13485920e-01
-7.53818274e-01 1.62622631e-01 6.36054456e-01 1.03305399e+00
5.63887298e-01 6.59155324e-02 -3.05788189e-01 7.61980176e-01
1.15644090e-01 8.47508609e-01 -1.63764790e-01 -1.32991087e+00
9.67496991e-01 7.43090808e-01 9.18506086e-01 -5.29333532e-01
-1.22325219e-01 -8.07433844e-01 -2.83161670e-01 2.65672565e-01
6.95478320e-01 -4.83808398e-01 -3.16381007e-01 1.67566288e+00
1.61281303e-01 -5.62082708e-01 -2.56716073e-01 7.33652771e-01
-1.68554112e-01 5.03894150e-01 -3.42258781e-01 -9.44196939e-01
7.92487979e-01 -6.75863206e-01 -4.68254000e-01 -4.74415720e-01
4.54375267e-01 -6.40915334e-01 8.43866587e-01 4.34870720e-01
-1.33841813e+00 2.50605941e-01 -9.57410574e-01 8.20471466e-01
2.67092377e-01 8.37859139e-02 8.45606267e-01 9.77114737e-01
-4.58328664e-01 1.99806005e-01 -4.65788662e-01 4.28467542e-01
6.67717814e-01 9.06452239e-01 2.66413987e-01 -1.69977192e-02
-4.86162633e-01 1.42528757e-01 3.96241784e-01 -3.05843130e-02
-8.01249802e-01 -3.83668453e-01 -3.81374836e-01 3.44021976e-01
8.57150376e-01 -1.03790082e-01 1.29077959e+00 -1.26265311e+00
-1.54146302e+00 1.80175364e-01 -2.17750132e-01 -4.47998583e-01
5.24413526e-01 -2.08297893e-01 -1.99410349e-01 1.39640048e-01
-5.61635150e-03 -1.52801961e-01 7.86002457e-01 -9.23526943e-01
-1.21800625e+00 -6.91324949e-01 4.95989025e-01 2.50820696e-01
-4.31396604e-01 5.48314787e-02 -2.10012376e-01 -3.17563117e-01
6.34756871e-03 -1.03798592e+00 -3.11103374e-01 -3.21137100e-01
-1.08930804e-01 -1.02582946e-01 3.12443525e-01 -2.01820612e-01
1.64294946e+00 -2.03520083e+00 2.40474977e-02 7.07261026e-01
-8.87243543e-03 4.36510134e-04 -5.97770698e-02 2.98706174e-01
1.30080909e-01 -9.25242826e-02 -5.41286059e-02 8.04337934e-02
1.76331341e-01 -1.50196543e-02 -4.02557582e-01 6.81947708e-01
-4.53064531e-01 6.40623987e-01 -7.08020449e-01 -1.79912791e-01
-5.84128276e-02 -3.80501926e-01 -5.05855501e-01 -1.00657148e-02
-4.05403078e-01 1.42778633e-02 -6.97550178e-01 7.03606784e-01
7.02137887e-01 -3.97521883e-01 3.16524178e-01 4.17312622e-01
-2.92326540e-01 4.85103838e-02 -1.64843392e+00 1.16094708e+00
-4.37107682e-01 1.04347549e-01 1.75170109e-01 -1.19154286e+00
6.59347892e-01 6.15104027e-02 7.94018090e-01 -8.76392722e-01
3.47093135e-01 3.75791490e-01 1.29494712e-01 -5.53401932e-02
1.81521773e-02 -1.07104920e-01 -4.73157726e-02 7.98030198e-01
-3.37318927e-01 7.84596682e-01 3.99509549e-01 -2.74683177e-01
1.12771547e+00 -3.40495080e-01 3.18034112e-01 -3.46194834e-01
3.16621661e-01 -1.00447247e-02 8.87707710e-01 9.34813559e-01
-1.67414650e-01 -3.42293680e-01 9.61004674e-01 -1.37186393e-01
-6.86100960e-01 -8.74437571e-01 1.98210612e-01 1.57390988e+00
3.14798623e-01 2.72980243e-01 -4.29777682e-01 -7.60964036e-01
3.53454471e-01 1.03456140e+00 -7.01745629e-01 2.52726942e-01
-5.98456323e-01 -9.13540900e-01 1.19145088e-01 3.98531646e-01
2.94344276e-01 -6.30143106e-01 -9.29942191e-01 2.61321723e-01
2.24133745e-01 -5.48706412e-01 -7.51631916e-01 5.76383591e-01
-1.02746558e+00 -1.11896086e+00 -6.61943316e-01 -5.46131372e-01
8.50028634e-01 2.85132527e-01 5.82163632e-01 -7.34867036e-01
4.45983969e-02 3.01888674e-01 -3.41758907e-01 -6.73327804e-01
1.04515933e-01 -1.81962878e-01 -3.17445546e-02 2.23638833e-01
1.53165877e-01 -3.28824013e-01 -8.94029617e-01 4.70916748e-01
-6.76281095e-01 -4.06573027e-01 6.66963577e-01 7.08860636e-01
8.16819489e-01 2.73183554e-01 6.72192335e-01 -7.23482490e-01
6.80135608e-01 -3.35037142e-01 -1.36592805e+00 5.33232033e-01
-6.60400271e-01 4.15836453e-01 8.30025554e-01 -5.38724959e-01
-9.29697096e-01 1.20148227e-01 5.31064391e-01 -2.98702270e-01
5.09043276e-01 3.57734144e-01 -1.34728968e-01 1.66756399e-02
4.31607932e-01 2.08117232e-01 -5.65198623e-02 -5.00565112e-01
9.39107612e-02 5.24062634e-01 6.86134323e-02 -6.89124107e-01
3.39461386e-01 3.18526298e-01 2.90258616e-01 -3.99221241e-01
-7.29755640e-01 -3.41953218e-01 1.10955842e-01 -1.23051569e-01
-6.76278546e-02 -3.85150850e-01 -1.74968112e+00 -2.35964075e-01
-5.36822498e-01 -1.08531043e-01 -4.52716053e-01 7.09280550e-01
-7.40598023e-01 8.67208242e-02 8.27189460e-02 -1.87524009e+00
-5.73433638e-01 -1.14480889e+00 2.90166289e-01 1.35400742e-01
4.63719741e-02 -5.91711938e-01 -1.50019318e-01 4.61609393e-01
1.31908089e-01 1.79936081e-01 1.28048897e+00 -6.34737253e-01
-5.58311701e-01 -3.81754220e-01 5.40800542e-02 1.70101672e-01
3.53011549e-01 -6.58843160e-01 -3.09528738e-01 -7.53168464e-01
6.27234429e-02 -1.59468472e-01 5.68795800e-01 7.27687240e-01
1.22927248e+00 -9.41488266e-01 -3.27772021e-01 3.47744167e-01
1.44471359e+00 1.31901693e+00 1.35629117e-01 3.04214954e-01
-1.08962871e-01 2.68598855e-01 8.73860180e-01 9.18304682e-01
-4.07074302e-01 5.37409067e-01 8.02256286e-01 2.40494251e-01
7.05838501e-01 1.86146870e-01 3.47466022e-01 4.07359982e-03
-8.00036918e-03 -3.85715932e-01 -8.24102521e-01 6.41682684e-01
-1.94707119e+00 -9.39923108e-01 4.25870776e-01 3.01710463e+00
6.16620719e-01 4.36844885e-01 6.59465015e-01 2.93741047e-01
7.86444724e-01 -2.15758100e-01 -1.16196847e+00 -7.28343070e-01
1.61451586e-02 4.33867663e-01 1.02107704e+00 6.12309039e-01
-9.87511456e-01 2.52962887e-01 5.26711607e+00 1.17202222e+00
-8.75051677e-01 -4.02015522e-02 7.67439246e-01 -1.09524739e+00
-1.49271667e-01 1.42417282e-01 -6.52389944e-01 7.70555913e-01
6.22345626e-01 -3.79483581e-01 7.87484765e-01 9.81675208e-01
2.42206231e-01 -3.63252133e-01 -1.24240029e+00 1.06233990e+00
-3.50419313e-01 -1.26558816e+00 -2.92210966e-01 4.70520228e-01
7.39661634e-01 -4.96159822e-01 4.43463713e-01 -6.90988526e-02
6.39724612e-01 -9.84497428e-01 7.66236424e-01 5.54722063e-02
9.34528232e-01 -1.53395772e+00 5.01328766e-01 6.69088721e-01
-1.12050474e+00 -1.02333057e+00 -1.18338563e-01 2.33983155e-02
-1.62420213e-01 4.84580606e-01 -7.91739702e-01 5.16701579e-01
5.61081946e-01 -7.91726112e-02 2.41666704e-01 1.29927063e+00
3.17431629e-01 4.67692196e-01 -6.09967828e-01 -4.36391324e-01
4.98646855e-01 -4.52907354e-01 7.02433228e-01 7.54108548e-01
3.58155847e-01 2.66105592e-01 2.15464130e-01 4.00919467e-01
-1.73306540e-02 3.32269520e-01 -4.10854183e-02 -3.39493483e-01
7.89978683e-01 5.55294275e-01 -7.32921660e-01 2.16635749e-01
-4.11805622e-02 6.27319694e-01 1.97162285e-01 3.13369185e-01
-7.18560040e-01 -6.85122192e-01 7.13240027e-01 -7.25784674e-02
5.85287988e-01 -2.20529232e-02 -4.07326818e-01 -5.24728358e-01
2.16554627e-01 -5.02058864e-01 6.83318734e-01 1.34977907e-01
-1.04104984e+00 1.02364190e-01 2.25809123e-03 -1.35995793e+00
-1.82671890e-01 -4.83087450e-01 -1.75450951e-01 7.44472504e-01
-1.01251614e+00 -5.00370860e-01 3.76950353e-01 6.27624273e-01
3.77582043e-01 -5.37875354e-01 7.00789452e-01 1.31650656e-01
-6.34002388e-01 1.02562749e+00 1.11659372e+00 -7.28714019e-02
8.04758072e-02 -9.48898196e-01 -4.51577693e-01 6.29632115e-01
-2.85781950e-01 4.68801558e-01 7.27126300e-01 -4.15083975e-01
-1.61226976e+00 -7.29167938e-01 3.24118882e-01 2.23647460e-01
5.51103652e-01 -1.34691924e-01 -6.47518924e-03 5.52909970e-01
-2.32117936e-01 -1.75891906e-01 9.50432122e-01 8.79813507e-02
-1.00064702e-01 -7.40795791e-01 -1.30482173e+00 5.70152342e-01
1.10761917e+00 8.46274421e-02 -1.44063700e-02 1.69688180e-01
3.65560800e-01 -2.21456751e-01 -7.11695969e-01 2.11148083e-01
6.86912775e-01 -7.81173110e-01 6.88675284e-01 -6.74588203e-01
1.08382786e-02 1.19917184e-01 -4.35104162e-01 -9.17492926e-01
-3.86836141e-01 -1.02115715e+00 -5.18111587e-01 8.25413287e-01
6.45966351e-01 -8.96710098e-01 9.82957840e-01 6.04438007e-01
2.15990573e-01 -1.14609635e+00 -1.49913275e+00 -1.01844645e+00
1.32639110e-01 -2.67178178e-01 6.47021472e-01 4.48402971e-01
4.94060367e-02 1.34952867e-03 -2.76454270e-01 -1.38745695e-01
9.56002593e-01 6.87729895e-01 2.82245070e-01 -1.16763508e+00
-8.07932258e-01 -5.80593109e-01 2.03907683e-01 -1.12547934e+00
-1.53571904e-01 -7.00060427e-01 -3.86914462e-01 -1.17707443e+00
3.27777743e-01 -8.66390347e-01 -8.34884584e-01 4.62360114e-01
2.45811999e-01 -6.14358246e-01 2.88744062e-01 1.91947520e-01
-7.47631609e-01 3.66544396e-01 1.11021352e+00 -2.65973508e-01
-7.30399787e-01 5.78871429e-01 -9.06352639e-01 5.60467660e-01
8.23459268e-01 -6.11775696e-01 -6.00995123e-01 -3.74864250e-01
4.94207203e-01 8.27605665e-01 -1.84099138e-01 -4.11876589e-01
-6.13409653e-02 -8.17342877e-01 3.66444468e-01 -7.00078905e-01
4.05818373e-01 -1.06639636e+00 1.50678292e-01 5.28884292e-01
-6.53344333e-01 -1.01420261e-01 1.94530442e-01 7.83743024e-01
3.84725094e-01 -4.18341368e-01 8.11739206e-01 3.99906561e-02
-3.53721865e-02 1.39200300e-01 -1.08474486e-01 9.20337290e-02
1.36556053e+00 -3.00450444e-01 -3.70783865e-01 -3.99951994e-01
-3.03951144e-01 5.35644531e-01 3.76503319e-02 -4.75738756e-02
3.29058766e-01 -1.15196967e+00 -4.15121168e-01 -2.05509380e-01
-4.95658442e-02 -1.51776746e-01 1.67087279e-02 7.73419738e-01
-2.91461080e-01 9.82641399e-01 -2.27796938e-02 -2.87763089e-01
-1.26540256e+00 5.26324332e-01 3.14051718e-01 -5.15523672e-01
-1.52929977e-01 9.91187513e-01 2.48359554e-02 3.26968819e-01
5.46112657e-01 8.31137970e-03 6.46256357e-02 1.67754695e-01
3.26044917e-01 9.76798773e-01 -4.17747349e-02 -3.54653560e-02
-4.10108060e-01 2.12719411e-01 -2.61997163e-01 -4.37939376e-01
1.43099093e+00 3.37119140e-02 -1.23060450e-01 2.08149761e-01
1.20495689e+00 1.63327470e-01 -1.38016021e+00 -3.92204225e-01
3.76905762e-02 -6.73324347e-01 9.08666477e-02 -1.02041471e+00
-9.74935234e-01 3.10027122e-01 8.88299942e-01 2.40836725e-01
1.44848490e+00 -1.98708698e-01 4.59657103e-01 5.17082572e-01
7.21512616e-01 -1.33816063e+00 -7.30705783e-02 2.78638572e-01
7.54868209e-01 -8.26576531e-01 8.18086714e-02 -8.07470307e-02
-5.12321293e-01 8.58370721e-01 2.90872842e-01 -1.05645403e-01
3.05098683e-01 -3.81921604e-02 -6.09377921e-01 3.14668149e-01
-5.43296576e-01 -1.67601570e-01 2.32583750e-02 -3.50339413e-02
2.46423651e-02 3.58934522e-01 -6.16066754e-01 9.04860675e-01
-1.40206978e-01 4.94911708e-02 4.20491491e-03 1.15113628e+00
-7.40476727e-01 -1.20849764e+00 -7.03536093e-01 7.87514865e-01
-6.59208298e-01 2.20937282e-01 -2.55368382e-01 4.57741469e-01
-2.65968163e-02 1.34740424e+00 1.46262959e-01 7.25900615e-03
2.68940657e-01 2.30480824e-03 6.92203462e-01 -1.30725756e-01
-3.67627144e-01 3.21029276e-01 1.03672884e-01 -4.26087976e-01
2.35901643e-02 -4.72600043e-01 -1.23883057e+00 -3.78852248e-01
-3.24855357e-01 5.89186728e-01 3.81416827e-01 8.44163060e-01
3.39749992e-01 1.61251381e-01 1.15475011e+00 -2.39826456e-01
-9.74877536e-01 -5.63756883e-01 -8.69458616e-01 -2.66346596e-02
3.65169823e-01 -8.81286800e-01 -2.40242109e-01 -5.98798215e-01] | [4.568022727966309, 3.337733507156372] |
984ac012-58ea-4a1b-8136-cd6315d4999c | exploiting-all-samples-in-low-resource | 2111.06971 | null | https://arxiv.org/abs/2111.06971v1 | https://arxiv.org/pdf/2111.06971v1.pdf | Exploiting all samples in low-resource sentence classification: early stopping and initialization parameters | In low resource settings, deep neural models have often shown lower performance due to overfitting. The primary method to solve the overfitting problem is to generalize model parameters. To this end, many researchers have depended on large external resources with various manipulation techniques. In this study, we discuss how to exploit all available samples in low resource settings, without external datasets and model manipulation. This study focuses on natural language processing task. We propose a simple algorithm to find out good initialization parameters that improve robustness to a small sample set. We apply early stopping techniques that enable the use of all samples for training. Finally, the proposed learning strategy is to train all samples with the good initialization parameters and stop the model with the early stopping techniques. Extensive experiments are conducted on seven public sentence classification datasets, and the results demonstrate that the proposed learning strategy achieves better performance than several state-of-the-art works across the seven datasets. | ['Hyunju Lee', 'HongSeok Choi'] | 2021-11-12 | null | null | null | null | ['sentence-classification'] | ['natural-language-processing'] | [ 5.77740967e-02 -3.08752775e-01 -3.45771343e-01 -6.24117255e-01
-5.01085341e-01 -2.27738395e-01 1.90178618e-01 -3.18824947e-02
-9.26709354e-01 8.37916970e-01 -2.63769150e-01 -2.67779976e-01
7.00048753e-04 -7.03057766e-01 -4.64534551e-01 -7.11030066e-01
3.06855857e-01 1.79777518e-01 1.97537333e-01 5.57247363e-02
5.13314068e-01 3.57949346e-01 -1.14775503e+00 7.81767815e-02
1.03451872e+00 8.59886527e-01 3.70774060e-01 4.13913518e-01
-2.80010283e-01 6.78129852e-01 -7.56355703e-01 -3.77052695e-01
2.71073043e-01 -1.63231432e-01 -5.94718039e-01 3.04871853e-02
-5.49270585e-02 -4.70167160e-01 -1.36557415e-01 1.16234982e+00
7.52844453e-01 4.08719778e-01 3.82640660e-01 -1.08835149e+00
-5.88914394e-01 7.38007247e-01 -5.71569026e-01 3.80120695e-01
-5.74856997e-02 2.04385385e-01 5.93854427e-01 -9.05320525e-01
1.54555529e-01 1.06347990e+00 6.59118950e-01 6.53126240e-01
-8.68939281e-01 -8.58091772e-01 3.52271467e-01 1.54226124e-01
-1.53362906e+00 -5.28369963e-01 8.61736953e-01 -1.26442179e-01
1.06176090e+00 4.04976122e-02 1.03978135e-01 1.22720027e+00
1.05368728e-02 1.01735139e+00 9.29539382e-01 -6.40795410e-01
3.47838432e-01 3.18807036e-01 7.17112184e-01 6.10862970e-01
3.74706298e-01 -3.86520565e-01 -4.40409631e-01 -1.86784998e-01
5.38981259e-01 1.58551678e-01 -2.53204018e-01 1.54344857e-01
-7.86979020e-01 8.21086705e-01 2.41940662e-01 3.45545590e-01
-2.88785309e-01 -1.26860037e-01 6.23053014e-01 3.50211263e-01
5.53631306e-01 3.93620104e-01 -9.50529158e-01 -3.11796576e-01
-7.38366008e-01 -6.33178875e-02 7.05475092e-01 9.66746390e-01
5.22538126e-01 1.48504246e-02 -1.00778960e-01 1.33392155e+00
1.91207051e-01 8.91448036e-02 1.00769520e+00 -3.49574834e-01
9.80905890e-01 7.49135137e-01 -1.07655875e-01 -1.04375148e+00
-4.92688894e-01 -5.79768777e-01 -1.24588895e+00 -4.84482884e-01
3.19725454e-01 -4.55850929e-01 -7.91862488e-01 1.67482162e+00
2.81443089e-01 3.32438767e-01 2.25424558e-01 5.65658391e-01
8.88956726e-01 7.46906459e-01 1.96576238e-01 -3.72670710e-01
1.16869724e+00 -1.09485912e+00 -6.78112209e-01 -4.86183703e-01
9.31367517e-01 -5.83175838e-01 1.49076641e+00 6.14897251e-01
-7.88379252e-01 -7.09967077e-01 -9.91358578e-01 9.47983786e-02
-3.27745289e-01 7.75267422e-01 6.82679236e-01 6.93602204e-01
-5.25554001e-01 6.30709350e-01 -8.86377990e-01 -3.44264627e-01
5.79825759e-01 3.51891547e-01 -2.14625150e-01 -8.30691084e-02
-1.22617757e+00 5.64072490e-01 9.35330927e-01 4.74875778e-01
-5.25531769e-01 -5.23392260e-01 -5.92777133e-01 6.18621781e-02
6.44415736e-01 -5.15482426e-01 1.28918064e+00 -9.49404418e-01
-1.68723774e+00 6.88639641e-01 -1.69304952e-01 -4.64125216e-01
5.21524727e-01 -5.05920410e-01 -2.14328304e-01 -1.47440806e-01
-3.11159670e-01 3.38116139e-01 6.76772118e-01 -1.02740109e+00
-2.83836186e-01 -2.46549562e-01 2.07951516e-02 1.22372180e-01
-1.07433069e+00 2.39744395e-01 -6.19415879e-01 -6.03397191e-01
9.44920033e-02 -7.03699529e-01 -3.92253220e-01 -5.00037968e-01
-5.16146302e-01 -4.18294191e-01 3.98973882e-01 -4.66611743e-01
1.80477476e+00 -1.90624416e+00 -2.00749874e-01 6.00082353e-02
-1.98984686e-02 6.06137633e-01 -1.27670929e-01 2.31496587e-01
9.04715583e-02 3.50464851e-01 -1.77198157e-01 -4.73232508e-01
-1.29810214e-01 1.11851715e-01 -2.35593677e-01 1.18618384e-01
1.25017434e-01 5.42027652e-01 -6.54344738e-01 -6.64195001e-01
-3.27809229e-02 1.85325205e-01 -4.81485188e-01 3.26124936e-01
-8.72934163e-02 -3.05058472e-02 -8.69872272e-01 4.88507658e-01
8.41095984e-01 -3.49030286e-01 1.46015868e-01 -1.81496635e-01
1.13090269e-01 2.50557333e-01 -1.36944866e+00 1.48468685e+00
-4.00249481e-01 2.35498011e-01 -9.15215239e-02 -1.09680915e+00
1.24527013e+00 1.19471587e-02 1.34763360e-01 -3.66307050e-01
5.88047862e-01 1.54934466e-01 1.36817828e-01 -9.28661823e-01
3.32641155e-01 1.07359312e-01 5.25768735e-02 4.70665067e-01
-1.95199251e-02 2.61710346e-01 1.95780590e-01 -4.80551869e-02
8.69216263e-01 -3.11919721e-03 3.98148447e-01 -9.24724489e-02
7.49918222e-01 -2.38594383e-01 8.64012003e-01 9.98250365e-01
-3.39811295e-01 3.87719184e-01 4.34657514e-01 -5.64866364e-01
-8.62094402e-01 -4.34837908e-01 -1.06784999e-01 1.15586817e+00
-3.88857961e-01 -3.68086934e-01 -9.48198736e-01 -6.76242352e-01
-3.70378524e-01 6.67715847e-01 -5.42249084e-01 -4.48175102e-01
-6.57575548e-01 -1.37575769e+00 6.26079381e-01 5.76584101e-01
1.03806055e+00 -1.13809741e+00 -4.94754672e-01 1.78025246e-01
-1.45776039e-02 -9.61696982e-01 -1.91793561e-01 1.73370451e-01
-1.12769186e+00 -7.29977131e-01 -5.77155173e-01 -8.95413756e-01
9.40701962e-01 2.30449632e-01 9.92029786e-01 5.22753894e-01
-2.44596656e-02 -4.26223248e-01 -5.14518559e-01 -3.93203855e-01
-2.74126798e-01 6.88694060e-01 1.93361118e-01 -4.47291136e-02
8.16658676e-01 -1.78685233e-01 -3.46410751e-01 2.08978444e-01
-9.17861164e-01 2.10130587e-01 6.44325793e-01 1.06182516e+00
5.43359399e-01 1.97627231e-01 8.93929541e-01 -1.02195609e+00
1.12265432e+00 -5.46668649e-01 -5.35681844e-01 5.40574253e-01
-5.89810193e-01 1.29100919e-01 1.24126244e+00 -8.26729000e-01
-9.67712224e-01 -1.25845999e-01 -7.93968514e-02 -3.58303457e-01
-2.63106763e-01 8.24839652e-01 -4.11243677e-01 2.08276704e-01
6.17041767e-01 2.79898793e-01 -2.14354098e-01 -7.82740295e-01
-1.49383321e-01 1.08799076e+00 1.05289720e-01 -7.69483149e-01
4.21873003e-01 -1.75157696e-01 -2.58530051e-01 -7.69613683e-01
-1.08188403e+00 -2.54842758e-01 -6.94142163e-01 1.20863564e-01
4.22916621e-01 -6.86321199e-01 -6.52907550e-01 8.39465082e-01
-9.88041818e-01 -3.97051364e-01 2.59707659e-01 6.05477035e-01
-1.86926752e-01 4.72782642e-01 -6.11150146e-01 -9.97681022e-01
-7.55875885e-01 -1.10891092e+00 5.77326953e-01 6.83313310e-01
1.10886514e-01 -9.66426075e-01 -2.15606943e-01 2.32617646e-01
4.14661646e-01 -2.19446406e-01 6.47020519e-01 -1.12153161e+00
-2.42083430e-01 -2.36611113e-01 -9.70130488e-02 5.38176179e-01
2.25738034e-01 1.52041033e-01 -1.03126156e+00 -3.83454233e-01
3.46935332e-01 -4.38251406e-01 8.03930938e-01 2.75869489e-01
1.80517960e+00 -3.55055600e-01 -2.49987856e-01 7.73095727e-01
1.35025179e+00 2.75478125e-01 4.74161327e-01 7.52440393e-01
5.50582230e-01 2.84432381e-01 7.87795007e-01 4.96585518e-01
1.27638459e-01 2.69023061e-01 -3.76410852e-03 1.20246619e-01
4.88902956e-01 -2.52497792e-01 1.67685643e-01 1.23005676e+00
2.09428459e-01 -4.57639694e-01 -1.06228864e+00 3.05579573e-01
-1.91176653e+00 -7.59865463e-01 7.42262006e-02 2.23194504e+00
1.11968029e+00 5.49212039e-01 -1.21752501e-01 2.12701738e-01
6.02475345e-01 -7.65794367e-02 -6.95881367e-01 -3.17970484e-01
-1.21885963e-01 1.23973517e-02 3.93662989e-01 3.35378796e-01
-1.16617382e+00 1.20807576e+00 6.17272758e+00 9.88544464e-01
-1.07891154e+00 -8.15044865e-02 1.01217866e+00 -2.02004477e-01
2.06754193e-01 -1.32281825e-01 -1.21268165e+00 5.86376131e-01
1.01230836e+00 -1.22221373e-01 3.06883484e-01 9.59297001e-01
4.92903531e-01 -3.14699225e-02 -8.81515741e-01 9.35184062e-01
1.06408350e-01 -1.04531395e+00 4.01227362e-02 -4.58793312e-01
5.49934030e-01 -7.66518712e-02 -1.91396326e-01 5.80594599e-01
5.50339855e-02 -9.08052683e-01 2.12280765e-01 7.10774958e-01
3.77636880e-01 -8.95250380e-01 9.86197948e-01 9.53266919e-01
-8.31527114e-01 -3.31851572e-01 -8.84626627e-01 -1.82691380e-01
-1.86660811e-01 6.17699027e-01 -8.52863729e-01 4.83112842e-01
6.66838288e-01 6.61393642e-01 -8.41139197e-01 1.10760772e+00
-1.19423412e-01 9.10275996e-01 -4.81128305e-01 -5.97694337e-01
1.37179106e-01 -3.33155930e-01 5.21476753e-02 1.18832004e+00
1.96742535e-01 1.66092962e-01 1.32540941e-01 6.16422713e-01
-1.99874461e-01 5.36684394e-01 -4.51987296e-01 1.25947043e-01
6.80137455e-01 1.17129719e+00 -6.31841183e-01 -3.89146090e-01
-3.98093104e-01 5.93433857e-01 7.83335507e-01 3.04769754e-01
-7.72913218e-01 -5.76121390e-01 8.90947208e-02 -2.74294645e-01
4.80749346e-02 -1.71565175e-01 -3.94202948e-01 -1.28786814e+00
2.46416420e-01 -1.13939810e+00 4.58102614e-01 -7.19499826e-01
-1.42067552e+00 6.35899723e-01 9.22544077e-02 -9.88412380e-01
-1.99361471e-03 -7.16816306e-01 -7.99055755e-01 8.54136765e-01
-1.46056664e+00 -6.87458873e-01 -3.26243490e-01 3.42226982e-01
7.82959759e-01 -4.83022481e-01 6.98498785e-01 3.42470080e-01
-1.33366847e+00 1.08256555e+00 2.78806925e-01 4.20281589e-01
6.45108521e-01 -7.32596338e-01 1.11798808e-01 1.05617356e+00
-2.46090189e-01 1.24960184e+00 4.78491545e-01 -5.05552232e-01
-1.22169971e+00 -9.05544102e-01 7.21633673e-01 2.25333516e-02
4.83633697e-01 -4.26622063e-01 -1.16606557e+00 6.24518991e-01
9.85671133e-02 -3.29276696e-02 8.30215812e-01 1.72375455e-01
4.22853753e-02 -3.35990816e-01 -1.12699163e+00 6.12458825e-01
7.47419417e-01 -5.20965643e-02 -5.87798655e-01 3.40701967e-01
6.19390249e-01 -3.36571157e-01 -6.83961689e-01 4.59440231e-01
3.15203547e-01 -4.69477683e-01 7.12220728e-01 -1.01049674e+00
3.32272440e-01 -1.09913476e-01 -6.32399693e-02 -1.24770415e+00
-1.63008630e-01 -6.83147371e-01 -1.76498875e-01 1.63657594e+00
4.16901141e-01 -6.32433832e-01 7.67327368e-01 9.78953063e-01
7.44263306e-02 -1.12701404e+00 -7.30378211e-01 -5.84750712e-01
2.37179488e-01 -3.23855430e-01 6.88989997e-01 9.72588241e-01
-3.21486533e-01 4.84860957e-01 -3.41266125e-01 -9.13996175e-02
3.35975140e-01 -1.18230522e-01 9.49389517e-01 -1.07157600e+00
-2.71585673e-01 -3.33080530e-01 6.49744086e-03 -1.18344784e+00
2.22658351e-01 -6.50059104e-01 6.74229581e-03 -1.24722219e+00
4.54284757e-01 -6.28158391e-01 -5.09997129e-01 6.44845068e-01
-6.69683576e-01 -3.31736058e-01 7.44459033e-02 1.58495486e-01
-6.01432025e-01 6.44919872e-01 1.13673186e+00 2.69411821e-02
-4.28316623e-01 9.63023454e-02 -8.26635718e-01 7.71955967e-01
1.28325534e+00 -6.53087795e-01 -5.12031496e-01 -7.85412312e-01
2.56863385e-01 -4.25767392e-01 -7.68185928e-02 -1.00916553e+00
3.02134573e-01 -3.35656375e-01 4.43880945e-01 -6.79106355e-01
-3.66656557e-02 -7.71021783e-01 -2.54133493e-01 3.43991488e-01
-5.85938394e-01 1.89940572e-01 3.57471228e-01 3.07014495e-01
5.47424518e-02 -8.99396658e-01 8.03097188e-01 -1.32719249e-01
-3.61485928e-01 2.62069643e-01 -1.55935556e-01 1.47123486e-01
9.32036102e-01 -1.00074403e-01 -2.65784591e-01 1.86443627e-02
-3.79417837e-01 4.63534147e-01 8.03828835e-02 5.16425848e-01
5.64614177e-01 -1.22384846e+00 -6.29668117e-01 3.57581109e-01
-4.68817241e-02 2.13203549e-01 1.48513883e-01 6.20969117e-01
-4.70613241e-01 3.64044428e-01 4.11275513e-02 -4.59145844e-01
-1.32563663e+00 6.52475297e-01 4.63986099e-01 -4.18947548e-01
-4.22409564e-01 8.80083025e-01 -2.54608333e-01 -6.36351943e-01
5.25684118e-01 -5.35844862e-01 -4.03138399e-01 -1.52281910e-01
6.70933485e-01 2.09582850e-01 6.08261153e-02 -2.44501337e-01
-2.54493207e-01 3.69775265e-01 -5.97152650e-01 3.59966040e-01
1.41932714e+00 -7.64124840e-02 2.80178264e-02 5.69229066e-01
1.29874551e+00 -2.89341271e-01 -1.16686809e+00 -5.00953257e-01
-1.39941648e-01 -5.05827546e-01 5.61532676e-02 -5.39049745e-01
-1.16359985e+00 9.72294986e-01 5.29720724e-01 2.65869051e-01
1.24281371e+00 -6.84297502e-01 7.26703167e-01 9.98270452e-01
2.14820281e-01 -1.54112911e+00 -4.02705073e-02 7.53080130e-01
4.82821047e-01 -1.36120069e+00 1.71470419e-01 -1.01045564e-01
-6.09665990e-01 1.26384401e+00 1.06238854e+00 -1.25551209e-01
7.37612009e-01 1.98755011e-01 -8.69531929e-03 1.86115593e-01
-8.92811179e-01 3.36678982e-01 -1.19950727e-01 1.45432696e-01
4.98835653e-01 -2.23262146e-01 -5.65273345e-01 1.12891018e+00
-2.04578280e-01 3.53066146e-01 3.89408439e-01 9.42691147e-01
-5.41565359e-01 -1.24882603e+00 -2.70453572e-01 6.82984054e-01
-7.20018327e-01 -2.11773038e-01 -2.41423383e-01 6.60976648e-01
-1.47389948e-01 1.09904289e+00 -2.07358748e-01 -3.02199274e-01
3.65934551e-01 1.98582023e-01 1.38002038e-01 -4.31719542e-01
-6.95415854e-01 -1.91464633e-01 -2.35577747e-02 -1.77291662e-01
-3.02818954e-01 -5.09287059e-01 -1.23134637e+00 -2.05821425e-01
-6.85283720e-01 1.51627705e-01 4.90941674e-01 9.31905150e-01
4.17801440e-01 4.73888189e-01 7.81340420e-01 -4.42737192e-01
-1.14368856e+00 -1.40789878e+00 -1.53029591e-01 3.10503572e-01
3.33941765e-02 -5.44977546e-01 -4.12664831e-01 -9.29030105e-02] | [9.103506088256836, 3.951772451400757] |
c7af6f68-a6e3-484c-b90b-ad8dd83c1f44 | learning-to-combine-instructions-in-llvm | 2202.12379 | null | https://arxiv.org/abs/2202.12379v1 | https://arxiv.org/pdf/2202.12379v1.pdf | Learning to Combine Instructions in LLVM Compiler | Instruction combiner (IC) is a critical compiler optimization pass, which replaces a sequence of instructions with an equivalent and optimized instruction sequence at basic block level. There can be thousands of instruction-combining patterns which need to be frequently updated as new coding idioms/applications and novel hardware evolve over time. This results in frequent updates to the IC optimization pass thereby incurring considerable human effort and high software maintenance costs. To mitigate these challenges associated with the traditional IC, we design and implement a Neural Instruction Combiner (NIC) and demonstrate its feasibility by integrating it into the standard LLVM compiler optimization pipeline. NIC leverages neural sequence-to-sequence (Seq2Seq) models for generating optimized encoded IR sequence from the unoptimized encoded IR sequence. To the best of our knowledge, ours is the first work demonstrating the feasibility of a neural instruction combiner built into a full-fledged compiler pipeline. Given the novelty of this task, we built a new dataset for training our NIC neural model. We show that NIC achieves exact match results percentage of 72% for optimized sequences as compared to traditional IC and neural machine translation metric Bleu precision score of 0.94, demonstrating its feasibility in a production compiler pipeline. | ['Dibyendu Das', 'Sandya Mannarswamy'] | 2022-02-22 | null | null | null | null | ['compiler-optimization'] | ['computer-code'] | [ 3.91974300e-01 -1.89828441e-01 -7.49276698e-01 -4.43609029e-01
-8.24584961e-01 -5.98831117e-01 1.47281334e-01 1.13690831e-01
-3.70036751e-01 4.11840230e-01 3.41288924e-01 -1.19598556e+00
6.97163165e-01 -6.91863716e-01 -1.18216443e+00 1.17772758e-01
-3.37313041e-02 9.35216025e-02 -1.02751575e-01 -4.39232558e-01
5.20415545e-01 2.56575942e-01 -1.64957809e+00 5.36611259e-01
9.15391505e-01 6.10707819e-01 2.98607051e-01 1.11149800e+00
-3.84594291e-01 8.64533961e-01 -8.32855880e-01 -3.60672623e-01
4.66425389e-01 -2.32636526e-01 -7.46046245e-01 -7.03203797e-01
6.50844991e-01 -5.18925190e-01 -2.48228312e-01 8.57727766e-01
4.33192879e-01 -2.02938057e-02 2.45209143e-01 -8.86066318e-01
-8.44417036e-01 1.12186885e+00 -5.75934649e-01 1.83303162e-01
2.32885823e-01 4.07642990e-01 1.03299308e+00 -3.35715473e-01
4.81089890e-01 9.12081063e-01 6.69148862e-01 2.95878261e-01
-1.12010860e+00 -6.59657419e-01 -3.20334136e-01 1.17860418e-02
-1.07883632e+00 -5.55923164e-01 3.61624241e-01 -1.97191104e-01
1.92576969e+00 2.83587515e-01 4.05125678e-01 8.20292890e-01
8.69744778e-01 6.88988030e-01 5.86416304e-01 -5.83317399e-01
1.86679661e-01 -3.04905534e-01 4.80058193e-01 9.71045375e-01
1.70250714e-01 3.94940823e-01 -4.29656476e-01 6.93420768e-02
1.85621336e-01 -1.82963327e-01 -9.98896956e-02 2.15608552e-01
-1.19166994e+00 5.87392032e-01 5.82393587e-01 2.64149576e-01
-5.18358834e-02 7.01767504e-01 9.39791679e-01 5.23381412e-01
-2.05776945e-01 9.05362606e-01 -5.70233107e-01 -5.93430042e-01
-1.02155101e+00 7.52905384e-02 9.15929556e-01 9.86390173e-01
8.18118930e-01 5.04960716e-01 -1.93029448e-01 5.72804928e-01
-1.22668901e-02 3.44391495e-01 9.16972280e-01 -7.53373742e-01
4.36921209e-01 7.61237383e-01 -5.69320142e-01 -7.55413651e-01
-2.64878690e-01 -6.81797504e-01 -4.56340194e-01 1.79924026e-01
-5.29291257e-02 -8.95158295e-03 -7.43461192e-01 1.64811563e+00
-1.18002787e-01 1.09129682e-01 6.22556731e-02 5.25056541e-01
6.43166423e-01 8.27508330e-01 -1.68243662e-01 1.88031301e-01
1.31354868e+00 -1.17084682e+00 -4.11807150e-01 -4.70099509e-01
1.27990329e+00 -1.02518177e+00 1.31916857e+00 2.43366838e-01
-1.05012429e+00 -6.34297431e-01 -1.75676465e+00 -3.38896871e-01
-2.28733584e-01 1.28901288e-01 1.02141500e+00 7.56215513e-01
-1.32105362e+00 6.19529903e-01 -9.17800665e-01 2.91354489e-02
1.50787741e-01 6.60494506e-01 -2.83156782e-01 1.75564542e-01
-6.64854705e-01 7.69188583e-01 6.97666109e-01 -2.53258795e-01
-5.22099435e-01 -1.16544092e+00 -9.08453465e-01 1.44580096e-01
1.38891831e-01 -6.12178922e-01 1.62747812e+00 -1.16952765e+00
-1.84793711e+00 4.97146755e-01 -3.65641117e-01 -6.83003664e-01
-4.99105871e-01 -3.51112127e-01 -4.14122224e-01 -4.92022663e-01
-1.56333745e-01 5.16983390e-01 4.77371395e-01 -6.83385193e-01
-6.02551699e-01 -1.67112686e-02 -3.55416648e-02 -1.63470596e-01
4.91599133e-03 1.60493523e-01 -2.15330869e-01 -7.26932406e-01
-4.47250128e-01 -9.91296589e-01 -1.05255976e-01 -6.94263816e-01
-2.19534189e-01 3.52882862e-01 7.24624038e-01 -8.19232464e-01
1.56834543e+00 -2.08532429e+00 1.00780495e-01 -1.18123017e-01
1.19435839e-01 5.65566123e-01 -3.35374713e-01 1.27287298e-01
-1.71972662e-01 1.99564800e-01 -2.42079228e-01 2.30431631e-02
1.24787971e-01 1.25292093e-01 -4.89566594e-01 8.96081999e-02
8.67679343e-02 1.46919513e+00 -8.87959898e-01 -1.50633961e-01
-1.72060832e-01 1.70705557e-01 -1.03732979e+00 3.39780778e-01
-4.91546690e-01 -4.01224121e-02 -5.63809313e-02 8.80334258e-01
3.61204952e-01 -4.34751436e-02 1.73060581e-01 -4.54646647e-01
-3.29410344e-01 8.00062001e-01 -3.93189788e-01 1.91574693e+00
-1.03934681e+00 9.21559393e-01 -2.47907668e-01 -9.31383848e-01
8.70326638e-01 -1.34648029e-02 -6.08137101e-02 -9.76233423e-01
2.11900100e-01 5.20325363e-01 3.29644859e-01 -1.12697974e-01
1.12471366e+00 2.88561195e-01 -5.87597847e-01 7.88128734e-01
2.80104131e-02 -2.44819388e-01 1.45906657e-01 2.70699169e-02
1.46052885e+00 3.23376000e-01 5.97940087e-01 -1.69227690e-01
3.93765092e-01 1.91380724e-01 3.35525095e-01 7.53818750e-01
-6.67389631e-02 9.35028717e-02 4.05604780e-01 -7.06834912e-01
-1.20964539e+00 -8.66650939e-01 1.92687601e-01 1.47660089e+00
-5.95545113e-01 -6.91079199e-01 -8.86640489e-01 -5.93122184e-01
-2.12400287e-01 1.10563946e+00 -2.33383924e-01 -4.37581748e-01
-1.31604648e+00 -7.54093051e-01 1.11736810e+00 4.26140755e-01
4.94654775e-01 -7.52741277e-01 -9.74438190e-01 5.13662040e-01
2.46804550e-01 -7.86388755e-01 -9.79326189e-01 6.33378923e-01
-8.81240845e-01 -5.52288175e-01 1.58098727e-01 -7.26055205e-01
3.39382887e-01 -7.24674612e-02 1.74064851e+00 2.40310773e-01
-4.28432971e-01 -4.18045282e-01 -1.60524964e-01 -1.09161720e-01
-1.18464494e+00 4.70593214e-01 -2.34113187e-01 -8.19173157e-01
4.72352862e-01 -5.52420437e-01 -3.76324713e-01 -6.14753366e-02
-9.37607825e-01 2.42722780e-01 9.46739137e-01 1.05310774e+00
4.07852739e-01 -4.25540268e-01 2.37506196e-01 -7.93650448e-01
5.60282528e-01 -2.95688957e-01 -1.05761719e+00 8.23608339e-02
-6.92168355e-01 7.61023462e-01 1.01855016e+00 -1.91806778e-01
-8.58989298e-01 2.22898275e-01 -4.61210787e-01 1.26163661e-01
5.71377110e-03 5.78242958e-01 5.68054318e-02 -2.08086088e-01
9.21522141e-01 1.37652144e-01 2.14879543e-01 -1.76889464e-01
6.27484143e-01 7.91135490e-01 9.76383746e-01 -8.15796316e-01
5.05304575e-01 -2.31040671e-01 1.14204235e-01 -2.39976361e-01
-5.44088721e-01 -1.31045720e-02 -4.61383551e-01 2.56714821e-01
5.93547106e-01 -6.53295398e-01 -7.98078597e-01 2.08142594e-01
-1.25218403e+00 -6.06123745e-01 -5.40976860e-02 8.08263943e-02
-4.35975522e-01 4.80591446e-01 -7.97594249e-01 -3.50839645e-02
-7.98764229e-01 -2.00015283e+00 1.07401013e+00 1.42633155e-01
-6.66365206e-01 -7.16360450e-01 5.13824448e-02 7.26203620e-02
8.27129781e-01 1.44070596e-01 1.25251484e+00 -4.28227246e-01
-9.05169487e-01 -1.24647491e-01 -2.05522746e-01 3.47896039e-01
5.01858480e-02 9.46152508e-02 -5.38640738e-01 -7.76503831e-02
-1.38732880e-01 -1.63958713e-01 6.63347006e-01 6.60251677e-02
1.16412616e+00 -4.19624567e-01 -1.35105088e-01 1.22438025e+00
1.57281733e+00 3.56797725e-01 4.93785769e-01 5.31146049e-01
7.55205035e-01 -9.04735550e-02 2.69553393e-01 1.70550868e-01
2.76134759e-01 8.05448949e-01 3.21015298e-01 1.60179883e-01
-1.97720036e-01 -2.20408767e-01 9.35393810e-01 1.30829155e+00
4.34419721e-01 -5.98679762e-03 -1.16171861e+00 1.54908359e-01
-1.26089823e+00 -6.05865836e-01 2.34733857e-02 2.32664394e+00
1.26478362e+00 6.63108677e-02 -3.59131545e-01 -1.83665887e-01
7.68996626e-02 2.38064211e-02 -5.17218649e-01 -1.54477954e+00
1.91661745e-01 1.05887008e+00 1.14113808e+00 6.83506548e-01
-8.01406622e-01 1.21094060e+00 6.71499348e+00 8.19970489e-01
-1.39251637e+00 7.15706721e-02 7.67875612e-01 -1.81901664e-01
-3.75474572e-01 1.78531900e-01 -1.08755958e+00 3.54753077e-01
1.82797635e+00 -2.03373834e-01 8.56421590e-01 9.35703397e-01
-1.82840332e-01 -8.78865793e-02 -1.35949349e+00 6.46119654e-01
1.47462979e-01 -1.66932404e+00 -6.83187395e-02 -1.22027390e-01
6.63928390e-01 4.01544303e-01 2.37290457e-01 7.13779271e-01
6.74793661e-01 -1.38968313e+00 5.40860176e-01 2.51764745e-01
8.83246541e-01 -9.91372585e-01 7.32779622e-01 -1.66466702e-02
-1.13420749e+00 -1.07455209e-01 -8.35964233e-02 -8.82585272e-02
1.53835014e-01 3.76928478e-01 -1.08993745e+00 1.11113653e-01
3.30396950e-01 4.36956018e-01 -7.75889874e-01 5.13779283e-01
-2.68940739e-02 6.97028875e-01 -9.22940448e-02 -1.18855849e-01
2.71083027e-01 1.59678444e-01 3.41497958e-01 1.52485895e+00
5.11836469e-01 -4.66612339e-01 -1.63341835e-01 9.86017346e-01
-2.93254197e-01 8.52950513e-02 -5.92431664e-01 -3.33927602e-01
3.82854372e-01 1.09222353e+00 -3.23374599e-01 -4.21240091e-01
-3.97591740e-01 1.17923224e+00 5.58563828e-01 -1.12722646e-02
-1.11740124e+00 -5.47615528e-01 8.99027288e-01 -4.12162483e-01
1.55740842e-01 -4.59363788e-01 -7.28425980e-01 -8.81604671e-01
-1.66187987e-01 -1.45371604e+00 9.66542633e-04 -3.14168006e-01
-5.36983311e-01 5.67544103e-01 -2.57783979e-01 -6.84270442e-01
-5.82384288e-01 -7.54004121e-01 -5.58318436e-01 8.82090092e-01
-1.29096961e+00 -7.97494590e-01 8.97905007e-02 -1.80248246e-01
5.60730457e-01 -5.27692080e-01 8.02217424e-01 2.77926594e-01
-6.76543295e-01 1.17284906e+00 2.21237123e-01 -2.29440644e-01
6.37443244e-01 -1.21783519e+00 1.24187410e+00 1.00460041e+00
-2.22633302e-01 1.22097218e+00 5.76596737e-01 -5.75334668e-01
-2.35385251e+00 -1.13727927e+00 7.50808001e-01 -3.47054809e-01
7.50334680e-01 -4.02880013e-01 -8.69898856e-01 8.34577322e-01
4.92736757e-01 -4.02549833e-01 9.25238550e-01 -4.83653694e-02
-5.02994776e-01 -6.59644790e-03 -6.78318799e-01 1.02863932e+00
8.27910721e-01 -4.81382221e-01 -4.39396083e-01 1.48937345e-01
1.30156875e+00 -7.99627185e-01 -1.00922489e+00 3.90074432e-01
6.73391998e-01 -9.07417715e-01 1.08012056e+00 -3.09102714e-01
7.76440740e-01 -3.93327177e-01 -5.51470459e-01 -1.23534560e+00
-1.93216905e-01 -7.09426939e-01 -4.13951546e-01 8.44793320e-01
6.14898503e-01 -5.75976491e-01 5.13720453e-01 2.19668090e-01
-7.65659511e-01 -7.84512579e-01 -6.28112674e-01 -7.06811845e-01
2.09830120e-01 -5.59151769e-01 1.13963151e+00 5.26698768e-01
7.23584741e-02 4.24103230e-01 -1.42983571e-01 -6.59009963e-02
2.36675397e-01 8.53548124e-02 9.33352292e-01 -3.05183917e-01
-1.08952737e+00 -8.46034348e-01 -1.86224729e-01 -1.15474200e+00
4.99110758e-01 -1.41444051e+00 1.67952955e-01 -7.70532727e-01
4.72710431e-02 -3.61740857e-01 -1.59226939e-01 5.63253641e-01
-4.29033220e-01 1.10463731e-01 4.48551439e-02 -1.14237443e-02
-2.05148697e-01 2.27986380e-01 6.72057569e-01 -1.24718823e-01
-2.49756142e-01 -6.81358516e-01 -9.09531176e-01 3.46480489e-01
1.07976472e+00 -3.25223297e-01 -1.65040895e-01 -7.46078789e-01
4.54296678e-01 -1.04634978e-01 -2.10218772e-01 -1.04750347e+00
8.73963833e-02 1.87616274e-02 -1.39513269e-01 -4.69812125e-01
-2.28779331e-01 -4.12026554e-01 3.72499108e-01 8.09173584e-01
-3.73158962e-01 8.33834767e-01 7.67032266e-01 -2.44834628e-02
-2.07806733e-02 -4.41263050e-01 7.51188993e-01 -7.45363310e-02
-1.03821123e+00 -1.00877412e-01 -4.20989096e-01 -8.68250504e-02
8.09252918e-01 -1.92166746e-01 -4.76910353e-01 2.68996924e-01
1.78300917e-01 -2.94729531e-01 7.82992721e-01 4.78946328e-01
3.45291078e-01 -9.73445237e-01 -4.76685375e-01 4.30413038e-01
9.14401710e-02 -4.07016426e-01 -1.98522642e-01 6.02806926e-01
-1.22317231e+00 7.46852219e-01 -2.92948782e-01 -4.31296676e-01
-1.20593941e+00 6.81405187e-01 2.84864277e-01 -5.30323327e-01
-5.01502633e-01 6.29572451e-01 -7.02529177e-02 -7.49538660e-01
-2.66090155e-01 -6.51316106e-01 4.49552894e-01 -6.15363121e-01
7.08759964e-01 4.78634946e-02 3.40402365e-01 -5.37935197e-01
-9.92801785e-02 1.01222239e-01 -3.00431728e-01 9.39857811e-02
1.02505445e+00 4.34449613e-01 -6.38781905e-01 2.30548397e-01
1.67871714e+00 4.99276146e-02 -5.61391830e-01 7.05400780e-02
8.94750953e-02 -1.92754269e-01 2.98100799e-01 -7.92400718e-01
-8.85803580e-01 6.44542813e-01 5.44685423e-01 -7.38066316e-01
1.19279277e+00 -4.97489989e-01 1.34307742e+00 6.50554776e-01
3.38689864e-01 -7.60660708e-01 -1.14302687e-01 1.01659584e+00
3.56518388e-01 -9.10003960e-01 -5.92361167e-02 7.32780993e-02
-1.76224411e-01 1.29320312e+00 5.49054384e-01 -1.71783984e-01
1.36874288e-01 1.09129620e+00 -1.61258459e-01 2.79194653e-01
-8.69510412e-01 2.46938542e-01 3.58469963e-01 4.34242219e-01
1.05319774e+00 3.12170982e-01 -3.56513113e-01 2.16519773e-01
-6.90506697e-01 2.55689174e-01 5.79307199e-01 1.16344738e+00
-3.39918792e-01 -1.56238532e+00 -2.36540988e-01 7.09775925e-01
-6.87585652e-01 -7.30950594e-01 4.88959514e-02 5.56863964e-01
-2.09029004e-01 5.28997004e-01 9.16265249e-02 -6.76727891e-01
7.53761753e-02 1.82052791e-01 4.35768127e-01 -5.66906154e-01
-1.33720601e+00 -4.94236797e-01 1.50325119e-01 -9.08457696e-01
4.26044792e-01 -2.92849272e-01 -1.32519901e+00 -5.82836688e-01
-5.74430525e-02 -3.40175897e-01 1.00841987e+00 5.79408169e-01
9.72269475e-01 1.04009032e+00 2.97763139e-01 -8.57803643e-01
-5.20869315e-01 -4.32665020e-01 2.58797199e-01 -7.80954026e-03
3.15925539e-01 -3.84812765e-02 1.04813807e-01 1.25537544e-01] | [7.805032253265381, 7.5995659828186035] |
23e8b2fb-4c81-4821-90d4-8729fd42cf3a | read-before-generate-faithful-long-form | 2203.00343 | null | https://arxiv.org/abs/2203.00343v1 | https://arxiv.org/pdf/2203.00343v1.pdf | Read before Generate! Faithful Long Form Question Answering with Machine Reading | Long-form question answering (LFQA) aims to generate a paragraph-length answer for a given question. While current work on LFQA using large pre-trained model for generation are effective at producing fluent and somewhat relevant content, one primary challenge lies in how to generate a faithful answer that has less hallucinated content. We propose a new end-to-end framework that jointly models answer generation and machine reading. The key idea is to augment the generation model with fine-grained, answer-related salient information which can be viewed as an emphasis on faithful facts. State-of-the-art results on two LFQA datasets, ELI5 and MS MARCO, demonstrate the effectiveness of our method, in comparison with strong baselines on automatic and human evaluation metrics. A detailed analysis further proves the competency of our methods in generating fluent, relevant, and more faithful answers. | ['Pascale Fung', 'Qun Liu', 'Xin Jiang', 'Lifeng Shang', 'Jindi Zhang', 'Xiaoguang Li', 'Dan Su'] | 2022-03-01 | null | https://aclanthology.org/2022.findings-acl.61 | https://aclanthology.org/2022.findings-acl.61.pdf | findings-acl-2022-5 | ['long-form-question-answering'] | ['natural-language-processing'] | [ 1.08449504e-01 6.06375277e-01 2.46828943e-01 -4.61879700e-01
-1.55398047e+00 -6.31272435e-01 9.33300495e-01 3.67139690e-02
-1.47227570e-01 1.09020066e+00 1.06982076e+00 -2.22846150e-01
3.37498009e-01 -9.32374358e-01 -6.74456775e-01 8.95831883e-02
6.71767294e-01 9.45583284e-01 1.00932933e-01 -8.61180365e-01
3.74781728e-01 -1.68639094e-01 -1.27088749e+00 9.93690014e-01
1.37712193e+00 7.04056919e-01 8.13643634e-02 1.09980726e+00
-4.68558192e-01 1.66169524e+00 -1.07610202e+00 -1.10742843e+00
-1.28562272e-01 -1.02398586e+00 -1.71112359e+00 -2.54674345e-01
9.10351396e-01 -5.96525848e-01 -1.71815902e-02 5.27123094e-01
5.85247397e-01 2.57979542e-01 9.07316267e-01 -6.99043453e-01
-1.30469358e+00 8.24681997e-01 5.03616966e-02 2.57910907e-01
9.94255245e-01 4.47082818e-01 1.29655409e+00 -1.04976237e+00
7.67310917e-01 1.65432966e+00 3.61420512e-01 1.00395787e+00
-1.09106946e+00 -1.69524446e-01 -1.11560233e-01 2.21941993e-01
-7.64239132e-01 -5.10707915e-01 5.72102368e-01 -1.21042296e-01
1.05428946e+00 3.65865201e-01 1.96124941e-01 1.18646359e+00
2.09288135e-01 1.04328418e+00 1.07341516e+00 -5.86814821e-01
-6.20148294e-02 -5.26046790e-02 1.74140841e-01 5.68822682e-01
-1.23835430e-01 -2.16309488e-01 -6.63257241e-01 -4.00803797e-02
2.81010389e-01 -5.33213854e-01 -4.22229946e-01 4.31790560e-01
-1.24597907e+00 1.20009613e+00 4.97222692e-01 2.08805539e-02
-6.46840334e-01 1.64788812e-01 1.21354245e-01 4.28442180e-01
5.29927492e-01 1.28342104e+00 -2.15081334e-01 -1.78914294e-01
-8.95448923e-01 9.42861974e-01 1.05826175e+00 8.18555057e-01
6.21790707e-01 -3.29964422e-02 -1.29097962e+00 7.58992910e-01
1.57329485e-01 6.74012005e-01 3.33474606e-01 -1.39757574e+00
7.05954850e-01 6.44152701e-01 4.69778448e-01 -6.87192380e-01
3.68012418e-03 -3.64968926e-01 -4.60157156e-01 -2.00033605e-01
4.55845386e-01 -3.62489134e-01 -7.58374095e-01 1.70975626e+00
2.17031896e-01 -4.09063518e-01 3.90465438e-01 7.20103323e-01
1.51547313e+00 1.05335617e+00 2.53665656e-01 1.11792728e-01
1.42882538e+00 -1.44377601e+00 -9.44306970e-01 -5.56439042e-01
3.42479318e-01 -8.60027492e-01 1.67944705e+00 -6.33653626e-02
-1.70692194e+00 -7.91207016e-01 -6.53070271e-01 -8.50332737e-01
-5.90408631e-02 -5.05512170e-02 2.70888388e-01 2.08688200e-01
-1.23663723e+00 1.58908695e-01 7.92009830e-02 -1.07414193e-01
2.54379153e-01 -3.11045617e-01 -6.32415861e-02 -3.44338506e-01
-1.58503127e+00 1.29444432e+00 2.68278778e-01 -2.26640701e-01
-9.34914649e-01 -8.94034564e-01 -9.79513347e-01 1.91239510e-02
1.78539291e-01 -1.49745405e+00 2.04467034e+00 -6.20337605e-01
-1.56705952e+00 8.85754883e-01 -5.67816138e-01 -6.59784198e-01
3.11876029e-01 -4.04802978e-01 -2.74418861e-01 6.09688044e-01
5.19079864e-01 1.33277380e+00 8.20854545e-01 -1.24340093e+00
-5.10756791e-01 9.83670056e-02 6.79942846e-01 4.00029212e-01
3.26361954e-02 3.15958373e-02 5.13727814e-02 -6.47442997e-01
-3.56980979e-01 -4.26350564e-01 -4.90059629e-02 -5.83811164e-01
-2.90040791e-01 -8.59590530e-01 3.08240712e-01 -1.04460227e+00
1.25514758e+00 -1.32221639e+00 1.54201046e-01 -5.63427329e-01
1.45451069e-01 2.18328148e-01 -5.95773757e-01 8.60628963e-01
3.29940408e-01 1.20641947e-01 -1.80831790e-01 -2.34540686e-01
2.88291425e-01 -4.28845994e-02 -1.02596509e+00 -5.92795968e-01
7.20828354e-01 1.66561520e+00 -1.29609036e+00 -6.06432438e-01
-2.70257503e-01 1.31871447e-01 -7.41981149e-01 7.25260675e-01
-9.12778795e-01 3.88981521e-01 -3.59723210e-01 3.66698682e-01
2.07511321e-01 -5.08505762e-01 -5.74968934e-01 -3.04430015e-02
3.53411436e-01 9.99586582e-01 -3.87186587e-01 1.77631068e+00
-7.45509923e-01 3.26665789e-01 -3.94825816e-01 3.42574753e-02
8.21496367e-01 4.83603925e-01 -5.48351109e-01 -9.40509856e-01
-5.67010157e-02 2.30942816e-01 -1.60449475e-01 -6.08555436e-01
1.13769174e+00 -3.95293444e-01 -2.79608548e-01 8.10331583e-01
2.65814275e-01 -1.03391957e+00 6.71882272e-01 7.95142710e-01
9.47865188e-01 2.62795061e-01 1.69841543e-01 -2.59142548e-01
5.78758240e-01 2.58407623e-01 -1.88041478e-01 9.83662963e-01
-6.94020931e-03 8.16116571e-01 3.41048956e-01 -1.74018949e-01
-6.95655167e-01 -1.21704733e+00 3.66781861e-01 1.08648777e+00
-2.12103784e-01 -4.42425221e-01 -9.99637246e-01 -9.47839677e-01
-3.21965843e-01 1.53650928e+00 -7.24496841e-01 -2.05622509e-01
-7.34685123e-01 -5.84149361e-02 6.43807709e-01 4.33501035e-01
6.13154590e-01 -1.65282309e+00 -6.06553257e-01 3.73652250e-01
-1.12204671e+00 -9.61964667e-01 -5.82047939e-01 -5.18596709e-01
-7.76961625e-01 -7.31744945e-01 -9.01747346e-01 -7.90972352e-01
4.35259789e-01 1.31945699e-01 2.31854868e+00 2.04476684e-01
3.10990304e-01 4.53440547e-01 -5.77194273e-01 -5.52796721e-01
-8.15276265e-01 2.01051861e-01 -7.62833476e-01 -3.78111064e-01
3.14509183e-01 -9.15679932e-02 -7.92688429e-01 -1.00195698e-01
-1.03837514e+00 2.77250528e-01 4.29652005e-01 7.96080232e-01
3.21475774e-01 -7.69009531e-01 1.29842377e+00 -1.11278307e+00
1.43043256e+00 -5.11769474e-01 2.18831211e-01 5.36692798e-01
-2.29923293e-01 3.52204829e-01 7.25559056e-01 1.83074530e-02
-1.61061490e+00 -6.45435154e-01 -6.10151947e-01 4.44270968e-01
3.95654999e-02 3.69332850e-01 4.61795628e-02 5.74009776e-01
1.18504333e+00 1.14610411e-01 -2.30638608e-01 -8.20812434e-02
1.12582684e+00 5.19822180e-01 8.22839081e-01 -9.05547917e-01
6.39044404e-01 1.93248078e-01 -4.77803618e-01 -3.53509992e-01
-1.78666353e+00 -2.70190746e-01 -7.03213289e-02 -1.75093561e-01
9.05249715e-01 -8.72899115e-01 -2.82942057e-01 2.09195301e-01
-1.64852405e+00 -4.84277397e-01 -7.76195526e-01 -2.41626531e-01
-6.81031108e-01 1.71681419e-01 -9.10326064e-01 -4.70316023e-01
-1.01007783e+00 -6.59957409e-01 1.21056211e+00 3.81574869e-01
-9.15983260e-01 -1.04834962e+00 2.97973186e-01 1.15176475e+00
6.66164815e-01 2.94614971e-01 1.13534832e+00 -4.56263155e-01
-5.82959771e-01 -1.11790612e-01 -1.29435554e-01 3.14688653e-01
-1.93017676e-01 -2.18638256e-01 -9.38343823e-01 1.47191852e-01
2.00267673e-01 -1.23688269e+00 9.48236525e-01 -1.99005436e-02
6.79384172e-01 -6.93707764e-01 4.20268625e-01 -1.96410269e-01
9.26950216e-01 -2.48259634e-01 9.22311842e-01 -1.42711058e-01
4.03683901e-01 9.18161154e-01 6.32762134e-01 5.29613718e-02
9.77042675e-01 2.71717489e-01 1.60508901e-01 7.45948181e-02
-8.63354862e-01 -7.80021846e-01 3.71406168e-01 7.91814566e-01
2.70167053e-01 -4.61590201e-01 -6.01691127e-01 8.85283411e-01
-1.61506736e+00 -1.33911169e+00 -4.13205922e-01 1.60876429e+00
1.51347911e+00 -8.24373960e-02 2.36832188e-03 -2.45410189e-01
2.79519379e-01 3.28748316e-01 -2.50088155e-01 -6.71463192e-01
-2.96136081e-01 8.04637730e-01 -6.96393549e-01 9.02133584e-01
-5.12422919e-01 1.16102111e+00 6.57535076e+00 8.34346950e-01
-5.20204842e-01 1.44690305e-01 7.86644220e-01 -3.74939926e-02
-9.84044433e-01 -2.96808966e-02 -7.60736704e-01 1.07577190e-01
1.07160378e+00 -2.72853643e-01 1.96778893e-01 5.17507613e-01
7.28918165e-02 -9.82282087e-02 -1.21440327e+00 3.98168325e-01
5.33817947e-01 -1.60416412e+00 7.85324931e-01 -6.62441909e-01
1.17515004e+00 -4.31513339e-01 1.09650776e-01 8.43375146e-01
6.04996383e-01 -1.25957382e+00 1.01986539e+00 6.09489143e-01
6.73021436e-01 -7.33908296e-01 7.35280097e-01 3.73897254e-01
-7.00614274e-01 3.13718379e-01 -3.10708553e-01 -5.60622156e-01
6.67854130e-01 4.39380556e-01 -9.81283367e-01 4.01049912e-01
3.42419475e-01 1.14094928e-01 -8.78108799e-01 5.67096174e-01
-1.12201893e+00 6.95769787e-01 3.01063210e-01 -4.48814392e-01
4.30928856e-01 2.22147897e-01 1.80899948e-01 1.11567998e+00
7.01153874e-02 4.36285973e-01 -1.14499629e-01 1.25116193e+00
-5.70563257e-01 1.10596865e-01 -3.82610321e-01 -5.94073907e-02
4.80085582e-01 1.30500257e+00 -3.03823706e-02 -6.21182680e-01
-2.00131670e-01 1.14070141e+00 7.81181335e-01 4.43341345e-01
-4.97605085e-01 -4.40331787e-01 1.10186879e-02 5.54360524e-02
1.59349516e-02 1.78854108e-01 -4.41000938e-01 -1.06573272e+00
3.15830149e-02 -1.57699585e+00 6.37681603e-01 -1.28680980e+00
-1.51632595e+00 7.43578851e-01 -1.84572503e-01 -5.85394740e-01
-9.90243137e-01 -2.86915243e-01 -8.95683944e-01 1.28027463e+00
-1.63971150e+00 -1.41214073e+00 -2.24112049e-01 4.83990461e-01
1.00609255e+00 1.86271563e-01 9.56095397e-01 -2.14665785e-01
1.85862422e-01 4.46193308e-01 -6.60477161e-01 -8.55880827e-02
9.04446721e-01 -1.57608998e+00 7.92171955e-01 1.10739779e+00
3.90103459e-01 7.79728949e-01 8.53677988e-01 -6.62533164e-01
-8.92180264e-01 -1.05280912e+00 1.75310290e+00 -1.22059500e+00
3.77027601e-01 -8.03905539e-03 -9.52927768e-01 4.73064929e-01
1.11675215e+00 -5.99574149e-01 6.06806576e-01 7.54521117e-02
-5.49153149e-01 2.64657348e-01 -1.23062706e+00 9.26762879e-01
6.98275030e-01 -6.81395650e-01 -1.42946923e+00 5.41622758e-01
1.26971138e+00 -4.41667140e-01 -4.74620402e-01 2.95193911e-01
-2.42585465e-02 -9.90907907e-01 8.34767461e-01 -8.17956090e-01
1.28912914e+00 -3.71284246e-01 1.09534577e-01 -1.57246232e+00
-3.67564946e-01 -8.14057648e-01 -3.72234017e-01 1.17530453e+00
7.21978545e-01 -1.18441030e-01 4.08191025e-01 6.50040567e-01
-4.29197550e-01 -7.75399268e-01 -5.23231149e-01 -3.15730929e-01
5.66858470e-01 6.92913588e-03 6.76463544e-01 4.19139743e-01
1.75757252e-03 1.29905879e+00 -2.30172575e-01 -3.75808150e-01
1.80761218e-01 4.11186397e-01 8.62899601e-01 -8.65018070e-01
-2.91770011e-01 -4.22361791e-01 5.88614345e-01 -1.36268687e+00
2.83757627e-01 -8.43733072e-01 2.79070407e-01 -2.32569718e+00
2.38895133e-01 2.16929153e-01 3.80127281e-01 1.15478933e-01
-9.86515939e-01 3.79438937e-01 2.57438630e-01 -1.23787597e-01
-9.63587701e-01 7.25584805e-01 1.90119076e+00 -1.68773904e-02
2.11851284e-01 -1.58455893e-01 -1.27683413e+00 4.85637993e-01
6.25244081e-01 -8.20849314e-02 -7.10873723e-01 -9.40528214e-01
5.45029044e-01 2.64414221e-01 3.58167529e-01 -7.26344228e-01
4.23040278e-02 -1.02838561e-01 2.45593160e-01 -6.75315082e-01
2.62276709e-01 1.15111642e-01 -4.90558296e-01 2.80575722e-01
-8.31432998e-01 3.38348061e-01 1.06964275e-01 1.96711332e-01
-4.79428411e-01 -5.39957464e-01 6.11504197e-01 -4.48718250e-01
-4.58889425e-01 5.69786392e-02 -1.82554945e-01 1.12469661e+00
4.18259174e-01 2.26999283e-01 -8.87384653e-01 -1.17962980e+00
-1.14350535e-01 4.50812429e-01 1.97694868e-01 5.54934859e-01
8.04877877e-01 -1.48467624e+00 -1.57127178e+00 -4.88621891e-01
4.10554558e-01 3.23512149e-03 3.79662067e-01 4.96278703e-02
-6.32898271e-01 7.07433105e-01 1.17889091e-01 -6.17336407e-02
-7.01930165e-01 2.94737935e-01 3.12907726e-01 -7.96223700e-01
-2.03026384e-01 1.23868406e+00 1.31427616e-01 -5.47414184e-01
-1.04876682e-01 -3.48182350e-01 -3.69441479e-01 6.07304946e-02
9.81373489e-01 3.09153378e-01 5.27143218e-02 -3.64745587e-01
2.51204789e-01 1.47814125e-01 -1.42085582e-01 -5.41169286e-01
7.68777192e-01 -1.85024485e-01 -3.04542571e-01 1.23816840e-01
8.39619756e-01 2.36606508e-01 -1.10211897e+00 -2.27656901e-01
-6.75421162e-03 -1.66063428e-01 -5.29795349e-01 -1.51020932e+00
-2.57040441e-01 1.08000457e+00 -2.45179772e-01 1.67750776e-01
8.81013036e-01 2.62806535e-01 1.34340107e+00 6.30206883e-01
1.69365555e-01 -8.70111406e-01 7.76958466e-01 1.05551600e+00
1.74052429e+00 -1.15348780e+00 -5.47822356e-01 -1.42359212e-01
-9.38033223e-01 8.36498201e-01 9.68134165e-01 -1.21066570e-01
-3.14441115e-01 -3.62511337e-01 4.55018908e-01 -3.11473876e-01
-1.08642626e+00 -2.46599510e-01 7.16826558e-01 4.84866530e-01
8.74362469e-01 -5.47118820e-02 -3.84105444e-01 6.34305775e-01
-8.73108983e-01 -1.70948133e-01 6.13174200e-01 6.03973746e-01
-6.52985632e-01 -9.34082866e-01 -1.83142573e-01 3.65512192e-01
-6.15420461e-01 -5.70825756e-01 -8.54341030e-01 3.88064593e-01
-2.83208817e-01 1.67335975e+00 -1.80309713e-01 1.92161143e-01
3.94945383e-01 3.65369231e-01 7.02384949e-01 -1.07406867e+00
-1.05964768e+00 -6.69182897e-01 6.06129587e-01 -3.40995908e-01
-1.36942923e-01 -1.32799432e-01 -1.34485662e+00 -2.14643806e-01
7.14766607e-02 5.01312137e-01 -1.34599907e-02 1.08932674e+00
4.96511698e-01 6.03310227e-01 2.27963731e-01 -1.85825542e-01
-9.23311710e-01 -1.30197561e+00 1.86954334e-01 9.27186966e-01
3.79594892e-01 1.84215635e-01 -1.82918653e-01 1.08059056e-01] | [11.485321998596191, 8.158233642578125] |
e03e8c94-40c9-4130-bc34-a737a6f3fe4d | multi-channel-transformers-for-multi | 2009.00299 | null | https://arxiv.org/abs/2009.00299v1 | https://arxiv.org/pdf/2009.00299v1.pdf | Multi-channel Transformers for Multi-articulatory Sign Language Translation | Sign languages use multiple asynchronous information channels (articulators), not just the hands but also the face and body, which computational approaches often ignore. In this paper we tackle the multi-articulatory sign language translation task and propose a novel multi-channel transformer architecture. The proposed architecture allows both the inter and intra contextual relationships between different sign articulators to be modelled within the transformer network itself, while also maintaining channel specific information. We evaluate our approach on the RWTH-PHOENIX-Weather-2014T dataset and report competitive translation performance. Importantly, we overcome the reliance on gloss annotations which underpin other state-of-the-art approaches, thereby removing future need for expensive curated datasets. | ['Simon Hadfield', 'Oscar Koller', 'Richard Bowden', 'Necati Cihan Camgoz'] | 2020-09-01 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 3.38595331e-01 1.50156943e-02 -1.74731508e-01 -1.50037974e-01
-9.04066503e-01 -1.05689299e+00 1.08771813e+00 -5.00068367e-01
-6.59249187e-01 7.49849677e-01 7.66191185e-01 -1.18895501e-01
4.64019887e-02 -4.23921376e-01 -6.29230440e-01 -3.59636754e-01
2.28013590e-01 6.60458326e-01 1.26746103e-01 -2.45474905e-01
-2.50964224e-01 5.77858128e-02 -1.18938196e+00 5.79394519e-01
4.62358445e-01 7.98001766e-01 -1.13020398e-01 6.62650824e-01
-2.73589551e-01 8.06166232e-01 -4.52052474e-01 -7.96158671e-01
3.38072509e-01 -4.09009248e-01 -6.96245313e-01 -4.90482271e-01
7.77635157e-01 -5.37305295e-01 -1.79608867e-01 5.19382119e-01
8.83406162e-01 -3.79235983e-01 3.62524927e-01 -1.05158281e+00
-5.64453006e-01 8.86138022e-01 -2.09648326e-01 -3.03878516e-01
2.99874604e-01 1.04045764e-01 1.42271495e+00 -7.15616405e-01
1.23012555e+00 1.07158566e+00 8.57160628e-01 6.97378993e-01
-9.89786148e-01 -8.45251620e-01 3.07609946e-01 1.73465461e-01
-1.15560746e+00 -7.35866904e-01 6.18961692e-01 -1.03923701e-01
9.80860949e-01 1.15145616e-01 8.52795243e-01 1.66818643e+00
-1.96532726e-01 9.77582037e-01 1.42535698e+00 -4.07060206e-01
-2.84731090e-01 -4.72605914e-01 -2.94902146e-01 6.17867589e-01
-1.15579307e-01 1.32135957e-01 -1.20703626e+00 -1.62833571e-01
7.28989780e-01 -3.86659384e-01 1.39640085e-03 -1.34647712e-01
-1.56272089e+00 2.99513161e-01 1.79944783e-01 3.09831977e-01
-4.87023473e-01 7.94649541e-01 7.04245508e-01 6.37336969e-01
1.87330209e-02 -1.98854804e-01 -7.42964923e-01 -6.90193713e-01
-8.17198873e-01 1.26719967e-01 8.79633963e-01 1.23241007e+00
1.86154187e-01 -1.64835617e-01 -1.26986921e-01 7.16971099e-01
5.61197579e-01 8.93677592e-01 8.51035491e-02 -6.67083144e-01
9.61048722e-01 1.50563031e-01 -6.46107271e-02 -2.84500182e-01
-2.34859943e-01 -2.76008546e-01 -5.64711690e-01 -1.97194293e-01
7.24204123e-01 -2.02387199e-01 -1.17726886e+00 1.93131185e+00
1.38352469e-01 1.15632430e-01 -7.91505799e-02 9.99999344e-01
8.27282429e-01 -5.96903078e-02 2.52848297e-01 1.37800381e-01
1.51423156e+00 -9.35975194e-01 -8.28628242e-01 -5.71040176e-02
4.10689801e-01 -1.25094712e+00 7.81401753e-01 2.63308108e-01
-1.09398925e+00 8.40575546e-02 -6.22832716e-01 -3.41827720e-01
-3.32917601e-01 3.97099286e-01 8.52247536e-01 4.03010786e-01
-1.20562637e+00 1.67659760e-01 -8.55580330e-01 -4.97043401e-01
3.40697438e-01 4.15596217e-01 -5.53071439e-01 3.79848629e-02
-1.17007434e+00 1.07149374e+00 -2.77997553e-01 4.94456977e-01
-6.31108642e-01 -5.64756393e-01 -5.24870515e-01 -4.89602149e-01
1.55162752e-01 -9.44297910e-01 1.35033190e+00 -8.65051568e-01
-2.13494349e+00 6.93299174e-01 -3.13131481e-01 -2.69976884e-01
1.07065213e+00 -4.06785160e-01 -4.63964701e-01 5.86900450e-02
-1.37632161e-01 5.84151983e-01 8.59777749e-01 -1.06260979e+00
-5.71484625e-01 -2.24775538e-01 8.39785486e-02 2.57732123e-01
1.10752899e-02 4.13333476e-01 -6.42287731e-01 -7.41779864e-01
7.30838254e-02 -1.34055221e+00 1.87232509e-01 4.22508746e-01
-3.24104160e-01 1.81795850e-01 8.31789196e-01 -1.05353606e+00
1.01199615e+00 -1.84988999e+00 4.64453727e-01 5.60898662e-01
4.27170619e-02 5.24773113e-02 -3.44813228e-01 6.61102474e-01
2.96280414e-01 -3.52783203e-01 -3.22804570e-01 -6.05808139e-01
3.34066033e-01 6.99233055e-01 -3.64684463e-01 4.66376454e-01
-2.49115210e-02 1.34864652e+00 -7.47434139e-01 -4.53966707e-01
9.01433975e-02 9.75295782e-01 -3.98175746e-01 -3.19385648e-01
-5.35821356e-02 6.45846725e-01 -2.70478845e-01 1.08830988e+00
3.38106006e-01 1.24502152e-01 5.67581952e-01 -1.77447855e-01
-3.15799356e-01 7.27263451e-01 -1.07135534e+00 2.35453176e+00
-8.82175267e-01 7.44861364e-01 3.93358052e-01 -4.48129803e-01
5.68446934e-01 7.23744154e-01 4.76667702e-01 -9.42934275e-01
1.05551347e-01 6.06032193e-01 1.26982197e-01 -1.81756496e-01
4.59476203e-01 -1.75542876e-01 -3.11019182e-01 7.22656548e-01
7.24405348e-02 -5.16362377e-02 -6.93029985e-02 5.19293807e-02
1.02165329e+00 6.52727783e-01 -5.10415323e-02 -2.11495906e-03
2.08004490e-01 -2.60514796e-01 2.90175200e-01 4.98078138e-01
-2.00669900e-01 3.39679688e-01 1.34073526e-01 -1.12077199e-01
-8.67503643e-01 -1.26179588e+00 6.68854453e-03 1.17913926e+00
-2.30945289e-01 -5.99018633e-01 -4.19905066e-01 -6.07618928e-01
9.70062315e-02 1.87654018e-01 -4.99932736e-01 4.87944543e-01
-1.02055335e+00 -2.78685927e-01 1.50749028e+00 6.85027421e-01
3.59497011e-01 -7.93451428e-01 -7.89509058e-01 3.18766594e-01
-7.11603343e-01 -1.36381948e+00 -5.67339599e-01 1.13420524e-01
-6.88826561e-01 -8.61224115e-01 -8.97406816e-01 -5.08933783e-01
2.62421787e-01 -4.45186436e-01 8.52481961e-01 -1.59363270e-01
-1.20084949e-01 5.22370040e-01 -3.54123145e-01 -3.41999590e-01
-1.74177900e-01 2.74531841e-01 -1.40564442e-01 -9.54930857e-02
1.13068461e-01 -6.70615375e-01 -3.48705709e-01 3.45307857e-01
-4.62710053e-01 1.83976367e-01 7.45761156e-01 9.34161305e-01
4.23712939e-01 -1.03079164e+00 4.02079254e-01 -5.55457771e-01
5.47739387e-01 -3.22138444e-02 -3.32174569e-01 3.91822994e-01
-3.44491780e-01 2.40154877e-01 3.65166664e-01 -3.51738125e-01
-1.03883851e+00 6.79140687e-02 -2.40987428e-02 -1.40134115e-02
2.23518640e-01 3.80097985e-01 -3.92567329e-02 -2.95927227e-01
1.52348027e-01 3.10480118e-01 -5.14654294e-02 -5.19719839e-01
9.09051716e-01 8.18565607e-01 6.85960770e-01 -8.85049284e-01
8.78638983e-01 6.84040010e-01 -6.00030869e-02 -5.83503544e-01
-1.97992429e-01 -2.16456398e-01 -8.89404058e-01 -4.73713189e-01
5.46778798e-01 -9.80320275e-01 -7.64606118e-01 6.71526670e-01
-1.46233189e+00 -4.33135718e-01 -2.95772761e-01 4.43416476e-01
-6.42754555e-01 2.74869561e-01 -8.80883574e-01 -5.11823535e-01
-5.53368449e-01 -9.53032613e-01 1.42323899e+00 -2.52802938e-01
-5.74122727e-01 -8.45852077e-01 -2.96962801e-02 4.76331145e-01
9.00775075e-01 1.67963073e-01 6.39219344e-01 -1.04710877e-01
-8.35949481e-01 -1.06891729e-01 -4.21999156e-01 8.39992091e-02
1.50183201e-01 -1.35659412e-01 -1.07862663e+00 -1.66497931e-01
-9.16481555e-01 -2.51649201e-01 8.12115729e-01 1.72444861e-02
1.78666934e-01 -1.31237939e-01 -1.06219798e-01 5.55993378e-01
1.09535909e+00 -3.51568073e-01 5.40497363e-01 1.84785172e-01
7.86562502e-01 4.61611897e-01 1.90594882e-01 2.99696863e-01
9.67545867e-01 9.51350689e-01 3.45872641e-02 -5.87235671e-03
-7.97977149e-01 -5.13782918e-01 6.50740027e-01 1.38719153e+00
-7.52186835e-01 -7.45877251e-02 -7.19565332e-01 7.95974433e-01
-1.89398646e+00 -9.47367072e-01 -2.27124095e-01 1.80251765e+00
1.14033079e+00 -2.94442415e-01 -8.33334923e-02 -1.34614054e-02
1.34691447e-01 2.24023372e-01 -2.14417160e-01 -5.87491810e-01
-4.38455880e-01 7.65397310e-01 7.97163427e-01 5.90746701e-01
-9.59846199e-01 1.27914023e+00 6.75318766e+00 3.15895349e-01
-1.32157624e+00 3.96210790e-01 -6.28648162e-01 -4.54233319e-01
-3.54987085e-01 8.78241658e-02 -5.93745351e-01 2.41908282e-01
7.66606390e-01 2.83134609e-01 8.72428298e-01 -7.18474714e-03
1.23325065e-01 7.17092007e-02 -1.07979250e+00 8.72484565e-01
1.48525536e-02 -1.11855221e+00 -5.92777506e-02 1.76681519e-01
5.54470718e-01 6.29292011e-01 -8.41222033e-02 -1.14176847e-01
5.29507935e-01 -7.78411090e-01 1.28586459e+00 6.46533370e-01
1.12089384e+00 -1.25975862e-01 4.20693398e-01 -1.58579543e-01
-1.55972457e+00 3.41113210e-02 4.44218725e-01 -5.27154095e-02
5.51442444e-01 1.75215036e-01 -4.96907145e-01 8.12741041e-01
4.03380990e-01 8.24575722e-01 -1.50474355e-01 8.11647475e-01
-7.67026842e-01 5.22966921e-01 -8.48172605e-01 1.18381150e-01
2.79091775e-01 1.97374206e-02 5.99960566e-01 1.36787820e+00
3.35799038e-01 -4.18950677e-01 -2.16737911e-01 4.82220173e-01
-2.00621128e-01 2.07146659e-01 -5.82156420e-01 2.77736522e-02
3.56063277e-01 7.86418617e-01 -3.30929875e-01 -3.31464529e-01
-6.05712593e-01 1.42832899e+00 6.38495013e-02 5.51116168e-01
-8.10765624e-01 9.01535600e-02 9.50884938e-01 -1.90751955e-01
4.77065086e-01 -5.39530456e-01 -2.36415699e-01 -1.26682234e+00
5.10633469e-01 -8.83490384e-01 3.79053175e-01 -6.03752017e-01
-1.08336914e+00 2.40493208e-01 -1.07335694e-01 -1.33065641e+00
-3.69043529e-01 -5.12622893e-01 -5.99356107e-02 9.11053717e-01
-2.03662801e+00 -2.15192580e+00 1.28595725e-01 9.25838828e-01
-1.39775258e-02 1.02409668e-01 1.18718624e+00 9.94908929e-01
4.21962738e-02 9.60761368e-01 -8.91342387e-02 4.18945998e-01
1.15960479e+00 -8.84669363e-01 5.22409379e-01 6.39825583e-01
3.61589015e-01 8.24048698e-01 3.68906915e-01 -5.17823935e-01
-1.60469544e+00 -7.33287871e-01 1.54351890e+00 -6.18448555e-01
8.87609601e-01 -6.21540964e-01 -1.10687152e-01 8.54553461e-01
2.98085093e-01 5.39606139e-02 5.98940849e-01 2.46911064e-01
-9.64594960e-01 1.82058737e-02 -8.56108546e-01 6.02711558e-01
1.71547627e+00 -1.20796466e+00 -4.18119550e-01 -9.24054682e-02
1.58730641e-01 -6.42058611e-01 -9.32589769e-01 9.94676501e-02
1.44002473e+00 -3.79630834e-01 8.48885536e-01 -5.55332422e-01
3.12530190e-01 -4.09113795e-01 -2.76700020e-01 -1.33324087e+00
3.52683693e-01 -9.37029362e-01 -2.24574700e-01 1.13363612e+00
5.64783335e-01 -7.11889386e-01 4.22006637e-01 4.46939290e-01
-1.03274815e-01 -1.79088935e-01 -1.56374478e+00 -6.28705323e-01
-8.56067389e-02 -8.07936788e-01 5.41020751e-01 1.03010416e+00
1.14691574e-02 1.69033021e-01 -9.16586637e-01 -6.12520240e-02
6.54746473e-01 9.82799158e-02 8.80043983e-01 -1.11826289e+00
-3.60791206e-01 -4.38888222e-01 -3.52710277e-01 -9.80215609e-01
1.31888613e-01 -1.12696183e+00 -3.77199007e-03 -1.66408610e+00
-2.49934688e-01 -4.25862491e-01 -1.22733183e-01 1.06581140e+00
5.47140419e-01 5.26721001e-01 4.73682612e-01 2.32554018e-01
-4.06537235e-01 5.21993697e-01 1.48484206e+00 -1.60249427e-01
1.00697823e-01 -3.57174337e-01 -3.00845921e-01 4.92946625e-01
4.23692942e-01 -3.46214801e-01 -6.91888556e-02 -1.26485288e+00
2.78227419e-01 -1.57567903e-01 6.13076210e-01 -4.87177193e-01
4.45365310e-01 1.57327279e-01 -2.67737117e-02 -3.37342262e-01
4.89246577e-01 -1.08565247e+00 3.52643937e-01 3.23646933e-01
-2.41570696e-01 1.48461953e-01 2.19078377e-01 3.42138380e-01
-1.18504152e-01 5.61246574e-01 2.60709852e-01 6.48206025e-02
-5.13759315e-01 1.84454322e-02 -4.16054189e-01 -3.74954566e-02
5.65727115e-01 1.25297327e-02 -6.24703944e-01 -3.28364998e-01
-7.34008253e-01 6.29463717e-02 3.40563595e-01 7.05769062e-01
2.94779122e-01 -1.36389709e+00 -6.61046982e-01 2.71506220e-01
3.61617327e-01 -3.85461032e-01 -1.07828207e-01 1.20748580e+00
-4.59915906e-01 4.25125241e-01 -4.00994927e-01 -2.99033970e-01
-1.32602406e+00 -5.82942486e-01 3.52469802e-01 -2.82394290e-01
-8.14785600e-01 7.82519937e-01 -6.32765770e-01 -8.81810248e-01
1.93590000e-01 -6.92593098e-01 3.73174161e-01 3.40656489e-01
9.46360677e-02 7.13175312e-02 5.53071238e-02 -9.28951204e-01
-5.77012599e-01 8.11349273e-01 1.69720605e-01 -7.04878688e-01
1.20343184e+00 -5.45115657e-02 -2.22662583e-01 3.23998541e-01
8.85141671e-01 2.48235673e-01 -7.49801219e-01 -4.78309423e-01
1.83618087e-02 -3.25706780e-01 -1.73682962e-02 -1.44109166e+00
-9.36609209e-01 8.05356443e-01 4.80355263e-01 -6.91559553e-01
9.04379487e-01 -6.64933845e-02 1.12103152e+00 4.37181324e-01
7.71143258e-01 -1.52287340e+00 -5.89660645e-01 8.32693756e-01
8.43802035e-01 -8.00445437e-01 -1.72368556e-01 -2.56991655e-01
-6.14363313e-01 9.34163690e-01 -6.26628771e-02 2.32917726e-01
5.48135459e-01 7.63164818e-01 7.05159068e-01 -1.26744047e-01
-8.17189395e-01 -4.79731292e-01 7.61703104e-02 6.37336850e-01
6.95967078e-01 3.88013363e-01 -6.34867370e-01 5.14326811e-01
-2.24415168e-01 3.91376168e-01 5.82147241e-02 1.21433103e+00
4.98478740e-01 -1.78434229e+00 -8.53489190e-02 2.79880762e-01
-3.50532323e-01 -5.52717745e-01 -8.19501281e-01 7.67106831e-01
2.47308269e-01 4.95311677e-01 -1.60768509e-01 -2.15324327e-01
5.70134044e-01 4.48088646e-01 9.23538208e-01 -3.05278040e-03
-9.56410050e-01 1.97327808e-01 7.58267999e-01 -6.88261807e-01
-9.27967548e-01 -1.16088474e+00 -1.32381773e+00 -2.34946996e-01
-1.64504677e-01 -6.02255344e-01 9.98499691e-01 1.18620217e+00
4.44627553e-01 4.06384677e-01 -1.81253165e-01 -5.93252599e-01
-4.18080777e-01 -1.01845086e+00 -1.93475425e-01 1.64695650e-01
3.71477485e-01 -6.03297293e-01 1.41053721e-01 1.51696965e-01] | [9.216373443603516, -6.543091297149658] |
82a7f2d9-94d6-47f3-b040-05c2da6289b6 | inherently-interpretable-multi-label | 2303.00500 | null | https://arxiv.org/abs/2303.00500v1 | https://arxiv.org/pdf/2303.00500v1.pdf | Inherently Interpretable Multi-Label Classification Using Class-Specific Counterfactuals | Interpretability is essential for machine learning algorithms in high-stakes application fields such as medical image analysis. However, high-performing black-box neural networks do not provide explanations for their predictions, which can lead to mistrust and suboptimal human-ML collaboration. Post-hoc explanation techniques, which are widely used in practice, have been shown to suffer from severe conceptual problems. Furthermore, as we show in this paper, current explanation techniques do not perform adequately in the multi-label scenario, in which multiple medical findings may co-occur in a single image. We propose Attri-Net, an inherently interpretable model for multi-label classification. Attri-Net is a powerful classifier that provides transparent, trustworthy, and human-understandable explanations. The model first generates class-specific attribution maps based on counterfactuals to identify which image regions correspond to certain medical findings. Then a simple logistic regression classifier is used to make predictions based solely on these attribution maps. We compare Attri-Net to five post-hoc explanation techniques and one inherently interpretable classifier on three chest X-ray datasets. We find that Attri-Net produces high-quality multi-label explanations consistent with clinical knowledge and has comparable classification performance to state-of-the-art classification models. | ['Christian F. Baumgartner', 'Lisa M. Koch', 'Andreas Maier', 'Stefano Woerner', 'Susu Sun'] | 2023-03-01 | null | null | null | null | ['clinical-knowledge'] | ['miscellaneous'] | [ 5.77488184e-01 1.02951252e+00 -7.32586026e-01 -8.03976059e-01
-6.10812783e-01 -3.42453569e-01 3.12931687e-01 4.88602847e-01
-1.90771855e-02 9.94891584e-01 3.57107848e-01 -9.03998733e-01
-3.43693882e-01 -2.86871821e-01 -6.67788804e-01 -3.18297386e-01
3.27601492e-01 6.36197984e-01 -4.59966749e-01 3.89966995e-01
1.02323078e-01 7.67531022e-02 -1.20429611e+00 9.54089761e-01
9.16204691e-01 8.07716608e-01 -4.85615700e-01 5.96652925e-01
1.15640201e-01 1.11587584e+00 -5.10433257e-01 -8.19144845e-01
-1.51235461e-01 -5.53744972e-01 -9.97915387e-01 -5.45725301e-02
4.62836236e-01 -3.94541591e-01 2.37677798e-01 8.76796544e-01
4.45076264e-02 -4.81505454e-01 1.07180083e+00 -1.75729370e+00
-1.06630981e+00 8.68007779e-01 -3.29978555e-01 1.02764396e-02
2.82170296e-01 5.49712330e-02 1.20569634e+00 -5.78766584e-01
6.93855226e-01 1.29942107e+00 8.75162125e-01 9.02822316e-01
-1.24564540e+00 -9.17700768e-01 2.48565711e-02 2.60084063e-01
-6.53574944e-01 -1.49485534e-02 4.11734402e-01 -5.21013677e-01
6.44172132e-01 7.50515699e-01 6.10050857e-01 1.36950099e+00
9.54003334e-01 6.39960527e-01 1.58443189e+00 -4.16742980e-01
1.62824363e-01 3.62189770e-01 2.34055042e-01 9.00269389e-01
5.71668565e-01 3.27920586e-01 -5.20220041e-01 -4.74750519e-01
5.46095490e-01 2.75469214e-01 -2.63309628e-01 1.41525595e-02
-1.45122373e+00 1.23110950e+00 7.89878249e-01 2.42946185e-02
-5.57887912e-01 2.47370318e-01 1.68256983e-01 1.84595600e-01
6.40830278e-01 9.28872705e-01 -6.01389468e-01 2.92727888e-01
-8.44823360e-01 1.63754091e-01 5.41487038e-01 4.19357777e-01
1.87250957e-01 -2.66246617e-01 -1.92209482e-01 4.56127584e-01
4.17081326e-01 2.25640193e-01 6.33268297e-01 -1.20835149e+00
1.03645220e-01 5.98070025e-01 -2.35265508e-01 -1.11846530e+00
-7.45153964e-01 -6.12983584e-01 -1.01379240e+00 5.23380816e-01
2.89002001e-01 4.13548015e-03 -9.10935998e-01 1.58214009e+00
1.39473602e-01 -1.08879633e-01 1.70814246e-01 9.85116124e-01
1.09954906e+00 -8.79646316e-02 4.44668829e-01 -1.01309448e-01
1.55678594e+00 -1.13559377e+00 -9.02851820e-01 -3.31387848e-01
1.00543821e+00 -4.18965548e-01 9.52465475e-01 4.21889096e-01
-8.13782275e-01 -6.82983249e-02 -9.62548077e-01 9.63121876e-02
-2.23018169e-01 -1.59811005e-02 9.73138690e-01 6.57999694e-01
-6.92556202e-01 6.12187326e-01 -3.25219750e-01 -4.43137959e-02
1.05083144e+00 3.29060763e-01 -6.68559611e-01 -2.97168959e-02
-1.01369905e+00 1.28087461e+00 2.63862848e-01 -6.34499416e-02
-6.25938535e-01 -9.47940767e-01 -8.27434182e-01 1.21283308e-01
3.22817802e-01 -1.07252300e+00 1.44319320e+00 -1.47535312e+00
-7.36516953e-01 1.07886755e+00 -1.37393132e-01 -4.96324748e-01
7.05544114e-01 1.05588369e-01 -4.42874283e-01 2.60791838e-01
5.06137371e-01 9.90968585e-01 7.40843117e-01 -1.47546828e+00
-6.10581517e-01 -2.41554186e-01 -3.24132293e-02 6.61810338e-02
1.72765180e-02 -8.01750347e-02 5.32472372e-01 -8.12079668e-01
2.50730246e-01 -1.09026873e+00 -4.16260481e-01 4.01671469e-01
-8.66166592e-01 -3.90591770e-02 6.33201003e-01 -5.19229472e-01
8.48346114e-01 -1.75164521e+00 -3.61757547e-01 6.81280047e-02
1.03164744e+00 -5.62899709e-02 3.24293345e-01 -5.10969162e-01
-6.19234383e-01 7.83803284e-01 -1.69128180e-01 -2.32406676e-01
-1.00930035e-02 3.85340750e-01 -2.72236973e-01 3.86041075e-01
1.68861479e-01 1.10841560e+00 -1.00884032e+00 -8.45091701e-01
2.16892734e-01 2.60368679e-02 -6.06061459e-01 -3.79218571e-02
-1.31900191e-01 3.83234650e-01 -2.69609571e-01 6.44612432e-01
3.09764922e-01 -8.83118093e-01 2.31334642e-01 -1.05124200e-02
4.64475423e-01 1.21271290e-01 -3.95438254e-01 1.17789173e+00
-3.57864767e-01 7.42610812e-01 -4.67804551e-01 -6.86862767e-01
3.62829566e-01 6.58927202e-01 1.98462814e-01 -3.03308785e-01
1.92129374e-01 3.60458732e-01 2.65185267e-01 -5.88010013e-01
5.59684522e-02 -8.85834455e-01 2.98018358e-03 8.46591353e-01
-2.64390856e-01 -2.69894451e-02 -6.12337530e-01 2.89676130e-01
1.07762611e+00 -2.83957332e-01 9.49931324e-01 -8.38801265e-02
9.39383060e-02 2.43028760e-01 6.06244564e-01 9.71603453e-01
-2.55143911e-01 8.83483469e-01 6.89663231e-01 -8.96605134e-01
-9.01250780e-01 -8.60338211e-01 -3.27519000e-01 5.75291574e-01
-3.44585404e-02 -9.78243351e-02 -6.06842399e-01 -1.29179549e+00
1.05617568e-01 1.33221817e+00 -1.22119427e+00 -1.99190617e-01
1.26681522e-01 -6.63559020e-01 5.00631392e-01 6.11592472e-01
1.25804931e-01 -1.25759661e+00 -9.96966660e-01 -5.47728548e-03
-3.30824107e-01 -8.59609962e-01 -1.67432562e-01 3.23899269e-01
-7.87903309e-01 -1.35984230e+00 -3.65831047e-01 -2.41929486e-01
1.08269513e+00 7.65415952e-02 1.37747896e+00 8.03407192e-01
-2.91235596e-01 1.21604554e-01 -2.58558124e-01 -8.65102708e-01
-9.78777885e-01 -2.29176968e-01 -1.96512584e-02 -2.66846508e-01
3.52417856e-01 -1.44824222e-01 -6.63186908e-01 4.69343066e-01
-9.00529921e-01 6.29218042e-01 6.84642136e-01 1.20857632e+00
5.52567422e-01 -2.54187703e-01 5.86694777e-01 -1.58887243e+00
6.86257482e-01 -6.60114944e-01 2.39151239e-01 4.75340366e-01
-1.25762522e+00 5.72360121e-02 5.54560304e-01 -4.53798205e-01
-9.91580665e-01 -1.67885795e-01 1.79186076e-01 -3.80071193e-01
-4.39672589e-01 5.41696072e-01 3.81333590e-01 2.71298289e-02
1.02856743e+00 -4.71145242e-01 1.65291905e-01 7.29703158e-02
3.50331873e-01 8.35493207e-01 4.27361012e-01 -1.27472103e-01
5.95038295e-01 4.60758924e-01 7.32140169e-02 -6.13808408e-02
-1.58019555e+00 4.82995249e-02 -4.11173642e-01 -3.54531229e-01
1.06745756e+00 -5.74315548e-01 -7.71435499e-01 -1.96861863e-01
-1.31409085e+00 -1.93826824e-01 -3.17638993e-01 5.80793023e-01
-5.76738656e-01 -1.13903256e-02 -3.04375112e-01 -4.74528432e-01
-3.21773857e-01 -1.30027425e+00 8.73127043e-01 1.38312235e-01
-1.14628863e+00 -1.22699082e+00 -2.90577859e-01 8.78037155e-01
1.75551683e-01 4.54667062e-01 1.31621444e+00 -1.01859879e+00
-3.67528647e-01 -2.66446114e-01 -3.94728005e-01 -3.29485014e-02
5.01251519e-02 -1.64888054e-01 -1.08518410e+00 2.96534359e-01
3.16856727e-02 -6.26951098e-01 5.84533095e-01 6.77023590e-01
1.69303215e+00 -6.25307083e-01 -6.10796809e-01 3.59002173e-01
1.06765652e+00 -1.05159834e-03 3.91658425e-01 3.10623288e-01
5.63125968e-01 8.10615778e-01 6.72906816e-01 2.38493811e-02
5.08448064e-01 4.97200012e-01 6.97386086e-01 -6.55138314e-01
-1.82958245e-02 -2.23995581e-01 -2.11728647e-01 3.46831605e-02
1.12444133e-01 -2.82939076e-01 -9.59458172e-01 3.50394487e-01
-2.15491796e+00 -8.49339187e-01 -5.17725289e-01 1.71607232e+00
5.42508364e-01 2.31573001e-01 -3.63892823e-01 -8.27514232e-05
5.16298056e-01 -2.92986006e-01 -6.85803235e-01 -8.32571685e-01
-1.60827160e-01 -1.05624862e-01 4.25566047e-01 3.56524378e-01
-7.82234013e-01 3.95084232e-01 7.15717411e+00 5.13817549e-01
-8.62373233e-01 5.99572241e-01 1.28823960e+00 -7.87617713e-02
-8.52993250e-01 -2.95597557e-02 -8.71936232e-02 3.33193749e-01
9.40122008e-01 2.12453865e-02 -4.57635708e-02 8.48286211e-01
2.27967292e-01 -1.67269066e-01 -1.36001194e+00 8.35933745e-01
2.38001794e-01 -1.83572316e+00 1.24321029e-01 2.07346275e-01
6.85004830e-01 -4.40944135e-01 2.41660297e-01 -5.78297302e-02
3.92038196e-01 -1.67032921e+00 8.83671999e-01 4.11658287e-01
1.02119482e+00 -2.98179328e-01 1.14176381e+00 2.25490153e-01
-2.53455341e-01 -1.66284084e-01 -8.11240375e-02 -5.36134578e-02
-2.34532114e-02 5.85161507e-01 -1.21720481e+00 3.45096320e-01
5.35995543e-01 4.89513725e-01 -6.05420649e-01 6.93997145e-01
-7.47652292e-01 6.86144292e-01 1.43038288e-01 1.52785601e-02
1.73554793e-01 5.23837924e-01 1.67996213e-01 8.54587734e-01
4.29455601e-02 4.51503545e-01 -2.66198099e-01 1.15311682e+00
-1.98439032e-01 -5.07562868e-02 -7.17898726e-01 2.63599694e-01
5.61714768e-02 1.22248268e+00 -8.33469093e-01 -5.77411711e-01
-2.11419702e-01 8.86319578e-01 1.40880525e-01 -6.00920618e-02
-9.82966602e-01 3.54010195e-01 3.06146413e-01 7.43400455e-02
-3.84542912e-01 7.88671672e-01 -1.04857814e+00 -1.15272677e+00
-2.40026414e-01 -1.08468056e+00 7.22051203e-01 -1.41546392e+00
-1.29365134e+00 5.74619234e-01 -9.19000506e-02 -1.01638460e+00
-4.21017945e-01 -6.97128236e-01 -5.32983959e-01 6.00796878e-01
-1.38781440e+00 -1.54314601e+00 -2.32244745e-01 1.07189015e-01
4.78389621e-01 -1.46410897e-01 1.24809992e+00 -2.61944801e-01
-3.87657374e-01 6.76853955e-01 -2.88111091e-01 -8.52212682e-02
7.76829064e-01 -1.46686494e+00 -9.11566615e-03 2.79360086e-01
1.84464723e-01 4.44051296e-01 8.19795132e-01 -8.24377656e-01
-3.79076153e-01 -9.42120492e-01 1.25065506e+00 -8.35204959e-01
4.06897575e-01 1.83944643e-01 -9.71620381e-01 9.00089800e-01
9.09725800e-02 1.22722633e-01 1.38012660e+00 3.59054774e-01
-4.23499554e-01 4.65445191e-01 -1.50817394e+00 6.06885552e-01
7.42659569e-01 -3.23001623e-01 -8.65737200e-01 8.16091776e-01
7.23475993e-01 -4.34011668e-01 -7.46664882e-01 1.30932063e-01
7.97259748e-01 -1.09536743e+00 8.49178612e-01 -1.28367376e+00
1.31823254e+00 1.08887121e-01 9.72525328e-02 -1.45018137e+00
-3.64945501e-01 -2.26663888e-01 2.51636863e-01 5.04035950e-01
1.11932755e+00 -7.65280008e-01 8.05988848e-01 1.25369465e+00
1.39519759e-02 -1.07918227e+00 -9.21667933e-01 -2.53745884e-01
-8.01440794e-03 -4.77679729e-01 6.67485058e-01 1.40717840e+00
2.86665499e-01 2.16792822e-01 -3.23618174e-01 2.10981846e-01
8.27390015e-01 2.28537649e-01 3.65116090e-01 -1.34126270e+00
-1.40405342e-01 -3.30358416e-01 -2.95977831e-01 -7.96756223e-02
5.34470916e-01 -9.68240559e-01 -7.26806670e-02 -1.59141421e+00
7.72870779e-01 -5.32635629e-01 -2.61410564e-01 1.09653723e+00
-4.09374565e-01 4.45606768e-01 1.06763065e-01 4.28617507e-01
-4.28479582e-01 6.06789254e-02 1.25728822e+00 -2.02201456e-01
3.50743592e-01 -6.57159612e-02 -1.31282246e+00 1.15184665e+00
7.40326881e-01 -1.08980811e+00 -2.64430821e-01 -1.43508956e-01
3.11212271e-01 3.51611137e-01 8.50966513e-01 -4.81573671e-01
-5.04759699e-02 -3.61253619e-01 6.24473095e-01 -1.91095620e-01
1.17885076e-01 -9.47875261e-01 4.27153766e-01 8.42152536e-01
-9.78018761e-01 6.86689690e-02 2.19201110e-02 6.08151615e-01
1.55626796e-02 -4.31585580e-01 5.81985474e-01 -3.33002329e-01
-1.51530698e-01 -3.77369933e-02 -2.90030479e-01 -1.64319113e-01
1.08093357e+00 -3.72348696e-01 -5.62155008e-01 -5.60527682e-01
-8.89966905e-01 1.23830810e-01 3.53701264e-01 3.03635240e-01
7.91432798e-01 -1.19174612e+00 -7.30964065e-01 -1.79660961e-01
2.70354271e-01 -2.37204865e-01 3.46922219e-01 8.23471785e-01
-4.39220339e-01 5.71101487e-01 -2.52230704e-01 -5.04619420e-01
-1.46884108e+00 4.32489783e-01 4.82996851e-01 -4.33677256e-01
-6.87359333e-01 6.49492264e-01 5.63970268e-01 -5.47567368e-01
-2.08446637e-01 -4.01676357e-01 2.36412808e-02 -2.49963701e-01
4.16836262e-01 -1.60914660e-02 -1.76441550e-01 -4.38669264e-01
-3.75344902e-01 6.82706237e-02 -1.07067816e-01 -6.14387728e-02
1.32526302e+00 6.44298941e-02 -6.98866695e-02 4.17840779e-01
7.62594461e-01 -5.12839198e-01 -6.97294056e-01 1.56873941e-01
-3.33964080e-01 -5.13491213e-01 2.05464840e-01 -1.45676064e+00
-8.81597757e-01 8.38168442e-01 3.70187283e-01 9.65191647e-02
8.27014744e-01 6.70308471e-02 3.43465894e-01 3.08198780e-02
-2.18345840e-02 -6.89317465e-01 9.22719687e-02 -5.14054716e-01
9.13435817e-01 -1.72623420e+00 3.33393186e-01 -4.54660177e-01
-9.96994495e-01 1.15769064e+00 6.42839909e-01 5.33605576e-01
3.81335497e-01 -1.30420744e-01 3.68203729e-01 -7.07115650e-01
-9.44510579e-01 5.27303636e-01 5.37821949e-01 4.31137621e-01
4.31114614e-01 6.53251827e-01 -1.77885696e-01 7.54686892e-01
-3.70335102e-01 7.61634707e-02 8.21237683e-01 4.69071418e-01
-3.77820805e-02 -8.01140547e-01 -6.94840074e-01 1.14559305e+00
-7.45564699e-01 -6.78873360e-02 -5.61014950e-01 8.30397367e-01
2.80042022e-01 9.42006588e-01 -1.90636262e-01 -2.85938978e-01
-4.08299603e-02 9.46540311e-02 1.11552484e-01 -7.00425565e-01
-8.81318688e-01 -4.02544856e-01 3.12332273e-01 -5.62228024e-01
-4.86172020e-01 -6.47053003e-01 -1.33388245e+00 -3.13542664e-01
-5.38057745e-01 5.20245023e-02 6.08872354e-01 1.34629774e+00
2.07715690e-01 7.61001706e-01 1.49868727e-01 -2.70850033e-01
-4.65412557e-01 -5.90647697e-01 -3.21081102e-01 6.91428602e-01
4.93787557e-01 -6.99058771e-01 -5.23110986e-01 -3.21061909e-02] | [8.596667289733887, 5.715461730957031] |
1f41588a-4d60-46a1-af31-bf7528177b35 | one-class-svm-on-siamese-neural-network | 2304.08058 | null | https://arxiv.org/abs/2304.08058v1 | https://arxiv.org/pdf/2304.08058v1.pdf | One-Class SVM on siamese neural network latent space for Unsupervised Anomaly Detection on brain MRI White Matter Hyperintensities | Anomaly detection remains a challenging task in neuroimaging when little to no supervision is available and when lesions can be very small or with subtle contrast. Patch-based representation learning has shown powerful representation capacities when applied to industrial or medical imaging and outlier detection methods have been applied successfully to these images. In this work, we propose an unsupervised anomaly detection (UAD) method based on a latent space constructed by a siamese patch-based auto-encoder and perform the outlier detection with a One-Class SVM training paradigm tailored to the lesion detection task in multi-modality neuroimaging. We evaluate performances of this model on a public database, the White Matter Hyperintensities (WMH) challenge and show in par performance with the two best performing state-of-the-art methods reported so far. | ['Carole Lartizien', 'Robin Trombetta', 'Nicolas Pinon'] | 2023-04-17 | null | null | null | null | ['outlier-detection'] | ['methodology'] | [ 3.22658122e-01 1.76161062e-02 1.92787454e-01 -2.61198014e-01
-7.74800122e-01 1.31423429e-01 6.41786039e-01 3.98993462e-01
-4.83951688e-01 4.62760687e-01 3.71166505e-02 -1.04395546e-01
-1.59947142e-01 -1.79722890e-01 -6.86128497e-01 -8.12340021e-01
-6.66939497e-01 7.45480835e-01 3.26444030e-01 2.20721796e-01
2.68037558e-01 6.99982166e-01 -1.60033476e+00 3.87257934e-01
1.05191112e+00 1.00862610e+00 -1.61020473e-01 3.76262575e-01
1.37737319e-01 5.82767367e-01 -3.61993760e-01 2.23410805e-03
3.97264421e-01 -2.87556738e-01 -5.94910145e-01 9.99079645e-02
8.55226338e-01 -2.06615761e-01 -4.53496188e-01 1.28681970e+00
5.60539603e-01 7.92573690e-02 1.16790962e+00 -1.17716348e+00
-5.25613487e-01 2.94243902e-01 -4.57654715e-01 1.08024907e+00
2.73970872e-01 2.31865361e-01 7.87803233e-01 -1.00606334e+00
6.75183952e-01 8.14888716e-01 7.33129323e-01 2.63001263e-01
-1.80851507e+00 -1.96298063e-01 -7.44560286e-02 7.40995705e-01
-1.12942874e+00 -2.91876197e-01 5.99308431e-01 -1.03986752e+00
1.32922029e+00 -1.23941809e-01 4.02809083e-01 1.57512236e+00
6.36158705e-01 8.42176616e-01 1.18856299e+00 -9.53486487e-02
5.71692228e-01 -3.36320698e-01 6.00174963e-02 6.32335901e-01
2.97834843e-01 -7.99373835e-02 -4.46250200e-01 -5.57136536e-01
4.21521634e-01 2.62139291e-01 -1.16934404e-02 -6.27931118e-01
-1.30500519e+00 1.01678741e+00 4.86883193e-01 5.53764343e-01
-9.54605103e-01 -5.30875213e-02 8.41566563e-01 5.78013897e-01
8.19970191e-01 4.78596330e-01 -1.37399212e-01 1.43804203e-03
-1.45412576e+00 1.99766621e-01 4.47143793e-01 3.82030994e-01
9.74290147e-02 4.56528902e-01 -2.52385855e-01 7.63491452e-01
6.00045808e-02 9.81319919e-02 1.19947994e+00 -4.19686168e-01
2.21297935e-01 6.20913982e-01 -3.73136759e-01 -9.26479518e-01
-5.35880208e-01 -3.70755076e-01 -1.03567457e+00 6.32872999e-01
4.53237861e-01 3.56740147e-01 -1.16943038e+00 1.17862856e+00
2.32219353e-01 6.48760855e-01 -1.38163567e-01 8.12708080e-01
5.11662543e-01 5.88431619e-02 -4.95052934e-02 -8.47336128e-02
1.20403683e+00 -7.57722378e-01 -5.83945096e-01 -5.17761856e-02
6.72596931e-01 -4.85227197e-01 9.15952504e-01 7.94966042e-01
-8.32629204e-01 -1.10643446e-01 -1.07963026e+00 1.37071714e-01
-6.24571800e-01 6.26505092e-02 4.06877875e-01 5.19416869e-01
-8.50212455e-01 8.47731292e-01 -1.36950743e+00 -3.00931185e-01
1.04969716e+00 2.13619441e-01 -8.48277450e-01 -1.73002958e-01
-6.14554226e-01 1.16678047e+00 3.18107605e-01 8.44556186e-03
-1.10212433e+00 -8.75065625e-01 -8.57783496e-01 -1.38714001e-01
2.34309673e-01 -4.56188172e-01 7.22372472e-01 -7.88396120e-01
-1.29576182e+00 1.10301733e+00 1.84455693e-01 -1.13070560e+00
9.49179232e-01 -3.19414258e-01 -7.06450224e-01 5.75516105e-01
3.45351964e-01 2.66730100e-01 1.35235751e+00 -5.19825935e-01
5.32897748e-03 -4.40859824e-01 -6.89175665e-01 -3.33787471e-01
-9.96058434e-03 1.72109902e-01 6.41394198e-01 -8.43647242e-01
5.75661838e-01 -6.50317132e-01 -3.36775690e-01 9.48407128e-02
-4.52231079e-01 -3.59476537e-01 8.84296536e-01 -8.70587409e-01
7.67015278e-01 -2.21332669e+00 2.53326595e-01 3.95224482e-01
3.11025620e-01 1.64190263e-01 1.70175862e-02 6.25626892e-02
-7.43261099e-01 -3.29797000e-01 -7.57615566e-01 -4.34471756e-01
1.27742290e-01 2.46033236e-01 -1.61539316e-01 1.06693852e+00
5.23899019e-01 7.00004697e-01 -8.74571919e-01 -3.01806569e-01
1.79237902e-01 1.84664756e-01 -5.98627985e-01 2.99599737e-01
1.30900496e-03 5.38505733e-01 -8.01494122e-02 7.29154885e-01
4.65045273e-01 -7.35950246e-02 -3.27247918e-01 -9.79550108e-02
8.11231509e-02 -4.63133343e-02 -1.12219679e+00 2.00216150e+00
4.46887733e-03 5.16851068e-01 -3.45680147e-01 -1.45181906e+00
6.60328984e-01 3.52977395e-01 8.35619926e-01 -6.67073309e-01
2.84498986e-02 6.06132567e-01 5.12756050e-01 -7.35076845e-01
-1.30848542e-01 -8.86990800e-02 3.05071950e-01 3.55453104e-01
6.44297063e-01 1.60337418e-01 2.74550706e-01 3.23153019e-01
1.72815096e+00 -1.69793442e-02 5.01844883e-01 -4.08863962e-01
4.67372447e-01 -2.70136893e-01 2.99643129e-01 8.49889338e-01
-5.15536487e-01 9.30070043e-01 8.08629453e-01 -5.99949777e-01
-1.32977951e+00 -1.24886322e+00 -6.76632166e-01 6.18756473e-01
-8.67848039e-01 8.04950297e-03 -5.94040692e-01 -9.43664253e-01
2.90655315e-01 5.89662492e-01 -8.51315677e-01 -2.46719763e-01
-3.56082529e-01 -8.31905127e-01 4.55168247e-01 5.39875746e-01
1.04500011e-01 -1.21401143e+00 -6.38556063e-01 3.27541173e-01
4.28420722e-01 -1.17261624e+00 -1.02522209e-01 3.57272536e-01
-9.61087406e-01 -1.14587176e+00 -7.03369975e-01 -4.54438359e-01
6.93975866e-01 -4.94993776e-01 9.15449023e-01 -2.86750972e-01
-1.02596951e+00 3.44677478e-01 -3.77389044e-01 -3.31039846e-01
-1.93317071e-01 -6.98958989e-03 5.08994341e-01 2.75623679e-01
3.58672112e-01 -8.86027277e-01 -6.25491083e-01 -5.29552251e-02
-1.22747266e+00 -8.01173031e-01 6.99728847e-01 1.15690660e+00
6.68229759e-01 -2.11942777e-01 6.74030423e-01 -9.93072152e-01
3.71292800e-01 -1.04550445e+00 -4.74940598e-01 -3.35309237e-01
-5.89623511e-01 8.60165656e-02 6.22524321e-01 -4.21586037e-01
-4.50337917e-01 1.35629103e-01 -1.57070130e-01 -9.81259525e-01
-6.08244479e-01 3.84046674e-01 1.63776264e-01 -2.00935215e-01
9.79023099e-01 5.36181629e-02 2.68489242e-01 -6.69702291e-01
1.93255320e-01 2.71057695e-01 8.13483059e-01 -1.69817239e-01
6.42617524e-01 5.79878867e-01 2.13445842e-01 -9.32444274e-01
-7.00870275e-01 -7.36850262e-01 -9.99488115e-01 1.18447721e-01
8.01156998e-01 -6.34555161e-01 -8.55026990e-02 4.17052120e-01
-5.45271993e-01 -1.29505441e-01 -6.52698338e-01 7.34671652e-01
-7.55086243e-01 6.00615263e-01 -5.14202952e-01 -5.12143314e-01
-2.66312659e-01 -1.25379050e+00 9.72531021e-01 -3.28399956e-01
-1.77536413e-01 -8.04531157e-01 3.42074335e-01 -8.58182013e-02
6.33305192e-01 6.40102386e-01 9.05651331e-01 -1.46161318e+00
-3.07765186e-01 -4.38611835e-01 -1.10756885e-02 6.27177715e-01
-6.78321123e-02 -4.07421708e-01 -9.75667298e-01 -4.62513387e-01
1.16874985e-01 -3.02079141e-01 1.11152887e+00 4.09157097e-01
1.42846441e+00 -1.99959233e-01 -5.78123741e-02 6.35602176e-01
1.28927743e+00 -3.98953408e-01 6.10278368e-01 5.89372516e-01
5.17955244e-01 3.27009946e-01 1.94158629e-02 3.68391842e-01
-1.94451243e-01 5.74286938e-01 5.85492015e-01 2.39211515e-01
-3.02530676e-02 1.98452696e-01 3.44882399e-01 6.75958574e-01
1.47529081e-01 4.43092883e-01 -1.14395857e+00 6.53610826e-01
-1.88944042e+00 -1.11077869e+00 -1.77576944e-01 2.07791877e+00
3.74566674e-01 1.63800761e-01 2.11289987e-01 3.17637235e-01
3.03592086e-01 -1.03260335e-02 -7.64630973e-01 -2.90174365e-01
-3.16111863e-01 3.73492986e-01 1.96984261e-01 -8.07585567e-02
-1.36932337e+00 4.80587900e-01 6.59998846e+00 6.18816137e-01
-9.41618919e-01 6.46562278e-01 3.81688774e-01 -1.98389873e-01
2.11895108e-01 -3.49389344e-01 8.97696149e-03 7.56394684e-01
1.15368140e+00 1.68742314e-01 1.58510476e-01 9.69212174e-01
-3.64884846e-02 -2.92490005e-01 -1.30685365e+00 1.15324879e+00
4.70913976e-01 -9.70402181e-01 -1.86723396e-01 -4.46855947e-02
7.03066468e-01 6.01554573e-01 2.06035018e-01 1.19925544e-01
-2.64663607e-01 -1.28669620e+00 3.19135845e-01 9.18277204e-01
4.39089358e-01 -4.69362795e-01 9.97404456e-01 9.11372676e-02
-5.13194084e-01 -1.26260743e-01 -5.82129061e-01 3.50216717e-01
-5.50499670e-02 9.25164640e-01 -7.66848505e-01 5.38378000e-01
9.72242713e-01 8.89865339e-01 -7.19955444e-01 1.71059585e+00
-2.17086583e-01 6.76999390e-01 -4.77861613e-01 7.18156457e-01
3.85723054e-01 -2.76708186e-01 1.11406696e+00 1.18744457e+00
3.99819106e-01 -3.08870137e-01 2.14128166e-01 1.09335399e+00
1.11952081e-01 3.43534201e-01 -9.86539900e-01 8.78620297e-02
-1.78334802e-01 1.27057552e+00 -7.35761166e-01 -3.22941989e-01
-4.52445269e-01 1.27094352e+00 3.84264857e-01 1.51201695e-01
-4.28451061e-01 -1.57258585e-01 5.30937791e-01 2.66607031e-02
4.55038309e-01 -9.36866254e-02 -8.79821181e-02 -1.53534496e+00
2.01088309e-01 -8.19027662e-01 7.43657827e-01 -5.20508111e-01
-2.03500271e+00 7.28230596e-01 -3.65018025e-02 -1.31454742e+00
-3.53466213e-01 -9.77909565e-01 -8.35445583e-01 6.48737490e-01
-1.49632692e+00 -8.34990382e-01 8.71982127e-02 7.17437565e-01
2.58940458e-01 -8.44520986e-01 9.70296502e-01 3.54055494e-01
-5.92441022e-01 4.82325524e-01 4.43431556e-01 8.61815587e-02
7.02747941e-01 -1.56950021e+00 -2.16575917e-02 1.11420619e+00
2.78635561e-01 3.53902578e-01 6.69947088e-01 -6.43727839e-01
-1.00294232e+00 -1.03462195e+00 4.25943494e-01 -4.86481518e-01
1.07995284e+00 -2.95748293e-01 -1.49830830e+00 8.24610889e-01
1.51359931e-01 7.55528569e-01 7.76504219e-01 -1.00553811e-01
-5.37717879e-01 2.32613429e-01 -1.56070328e+00 1.94881067e-01
7.21713364e-01 -5.77917457e-01 -1.14913571e+00 7.04520941e-01
1.38397306e-01 -4.23091829e-01 -9.37970638e-01 2.03467429e-01
1.09706841e-01 -8.44456494e-01 9.98384655e-01 -9.73758936e-01
3.88930947e-01 -1.97360918e-01 -4.96191643e-02 -1.72934365e+00
-2.63891339e-01 -4.08140272e-01 -5.26479840e-01 6.11185133e-01
5.60077615e-02 -8.63885343e-01 6.57745123e-01 2.46885672e-01
-5.07910132e-01 -7.42639601e-01 -1.65132535e+00 -9.59505260e-01
1.82845682e-01 -4.33359027e-01 3.17351758e-01 9.63163257e-01
4.62568458e-03 -1.86555326e-01 -1.00306071e-01 4.17766929e-01
9.17596042e-01 -2.31719464e-01 1.79294899e-01 -1.41159034e+00
-2.80082934e-02 -6.62145674e-01 -1.49390662e+00 5.83941303e-02
3.88927549e-01 -1.43450010e+00 -2.41172701e-01 -9.97182012e-01
-4.51820670e-03 8.36740285e-02 -7.83105075e-01 5.77404141e-01
1.57142114e-02 3.56784075e-01 -3.29754472e-01 2.90359735e-01
-6.01793706e-01 5.49923897e-01 3.93856734e-01 -2.51556545e-01
1.50477648e-01 -2.06420824e-01 8.14967752e-02 8.37239087e-01
7.46192336e-01 -7.00851858e-01 7.02568665e-02 -4.74722832e-02
-4.93005291e-02 -6.28333688e-01 6.38994038e-01 -1.43935788e+00
2.44194001e-01 5.40219903e-01 5.46025693e-01 -3.84717315e-01
1.03853652e-02 -8.48978102e-01 -3.77944350e-01 5.07632852e-01
-1.86189860e-01 3.76242310e-01 1.58370603e-02 7.26059794e-01
-2.51417488e-01 -3.82450551e-01 8.71843994e-01 -7.54998182e-04
-6.07117832e-01 5.96330822e-01 -6.60001636e-01 8.96701366e-02
1.15541327e+00 -1.09273680e-01 -2.57814359e-02 1.65841684e-01
-1.30567539e+00 -1.00019924e-01 1.33892611e-01 2.82812208e-01
8.44272852e-01 -1.48564839e+00 -8.94856751e-01 6.33897781e-01
5.23588240e-01 -4.11333710e-01 4.90983069e-01 1.48444867e+00
-4.30050343e-01 1.34690240e-01 -8.23028743e-01 -9.31050599e-01
-9.40334082e-01 5.70912778e-01 4.76456255e-01 -2.60040134e-01
-1.26064098e+00 5.84201574e-01 -3.91716033e-01 -3.21734220e-01
1.74871355e-01 -4.59911168e-01 -1.88533992e-01 2.61087924e-01
5.66744030e-01 4.54201370e-01 6.87686682e-01 -6.33281946e-01
-3.96941453e-01 1.11546339e-02 -3.62427682e-01 2.32024878e-01
1.74949062e+00 4.06590253e-01 -2.61686951e-01 7.47513771e-01
1.22463250e+00 -5.27642846e-01 -9.67824221e-01 -5.01728058e-01
4.87032712e-01 -5.94768286e-01 1.76130831e-01 -4.74469692e-01
-9.94159639e-01 1.04943907e+00 1.12263358e+00 3.28607470e-01
7.17461288e-01 2.08868049e-02 4.61410850e-01 3.45797598e-01
2.16777250e-01 -1.20324242e+00 4.49828953e-01 4.09585714e-01
9.73454535e-01 -1.47520304e+00 9.32508260e-02 1.97925597e-01
-5.61849535e-01 1.13051653e+00 3.93182129e-01 -7.43276834e-01
8.39516222e-01 2.96042394e-02 -1.58661634e-01 -5.23052633e-01
-5.14302313e-01 -2.69217759e-01 6.70919836e-01 7.50347674e-01
1.90685719e-01 3.36574651e-02 -1.08706921e-01 2.78113157e-01
5.51655255e-02 -2.52246410e-01 5.21748543e-01 7.58414149e-01
-1.79762691e-01 -9.24721897e-01 -2.10637003e-01 1.31344116e+00
-7.91653097e-01 -1.28213435e-01 1.48783237e-01 4.19897705e-01
1.30434066e-01 3.16946834e-01 1.35691255e-01 3.52517605e-01
3.70641917e-01 6.77994609e-01 3.78858268e-01 -7.78018296e-01
-5.03778279e-01 -1.35548100e-01 -3.13596368e-01 -1.17226446e+00
-2.86292136e-01 -1.15234244e+00 -9.57453430e-01 3.02876025e-01
3.42383310e-02 -1.80008560e-01 5.25271654e-01 9.71247077e-01
4.33172077e-01 7.03845739e-01 2.53199846e-01 -9.00100231e-01
-7.96057701e-01 -1.15404248e+00 -9.73019958e-01 9.46904957e-01
5.55648863e-01 -1.04960227e+00 -5.08875310e-01 -3.02502334e-01] | [7.682313442230225, 2.2269413471221924] |
70571ae3-fad1-45a0-8676-c74202819704 | a-scalable-method-for-quantifying-the-role-of | null | null | https://aclanthology.org/W19-5934 | https://aclanthology.org/W19-5934.pdf | A Scalable Method for Quantifying the Role of Pitch in Conversational Turn-Taking | Pitch has long been held as an important signalling channel when planning and deploying speech in conversation, and myriad studies have been undertaken to determine the extent to which it actually plays this role. Unfortunately, these studies have required considerable human investment in data preparation and analysis, and have therefore often been limited to a handful of specific conversational contexts. The current article proposes a framework which addresses these limitations, by enabling a scalable, quantitative characterization of the role of pitch throughout an entire conversation, requiring only the raw signal and speech activity references. The framework is evaluated on the Switchboard dialogue corpus. Experiments indicate that pitch trajectories of both parties are predictive of their incipient speech activity; that pitch should be expressed on a logarithmic scale and Z-normalized, as well as accompanied by a binary voicing variable; and that only the most recent 400 ms of the pitch trajectory are useful in incipient speech activity prediction. | ['Kornel Laskowski', 'Mattias Heldner', 'Marcin Wlodarczak'] | 2019-09-01 | null | null | null | ws-2019-9 | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 3.37065935e-01 1.94432050e-01 -2.82217771e-01 -3.50693911e-01
-6.89626217e-01 -8.98843765e-01 9.31218803e-01 3.83840501e-01
-3.78854692e-01 5.77409685e-01 7.19829917e-01 -5.13948262e-01
-1.43648982e-01 -4.08760309e-01 2.16672882e-01 -6.11387849e-01
-1.41078785e-01 4.03862417e-01 1.72097787e-01 -3.44420880e-01
5.18251002e-01 3.78671348e-01 -1.54284275e+00 8.13450888e-02
3.14463258e-01 8.68711054e-01 7.40450770e-02 9.95842636e-01
-3.53799134e-01 7.37972200e-01 -1.32463956e+00 -1.30664989e-01
-9.98709053e-02 -8.32899332e-01 -9.49057043e-01 1.83568850e-01
-3.07409436e-01 -1.89005941e-01 1.68036986e-02 5.02556920e-01
7.68194079e-01 2.93427855e-01 3.10282499e-01 -9.84588265e-01
4.04115289e-01 6.38345897e-01 1.21851020e-01 7.78329909e-01
8.33729386e-01 1.20978855e-01 1.20825529e+00 -3.05211931e-01
4.27236736e-01 9.87855554e-01 5.85399151e-01 2.41115287e-01
-1.30383444e+00 -4.12904263e-01 1.26801357e-01 -2.18330890e-01
-1.08704734e+00 -9.31993723e-01 8.40690792e-01 -5.29324651e-01
1.40043259e+00 5.16526937e-01 1.02486253e+00 9.00660336e-01
1.23081282e-02 4.65009272e-01 8.04400861e-01 -6.02196813e-01
8.40706974e-02 2.01667041e-01 -7.99824372e-02 3.62161733e-02
-5.52965939e-01 -6.81389347e-02 -9.01159883e-01 -4.77941811e-01
5.18443525e-01 -5.92346013e-01 -3.85089457e-01 3.66441488e-01
-1.09356725e+00 7.75382102e-01 -3.33835572e-01 6.94563746e-01
-4.74655390e-01 -3.12070310e-01 7.32383490e-01 7.02578962e-01
6.38710499e-01 5.70121706e-01 -3.69934589e-01 -1.33197892e+00
-9.23696756e-01 4.04031664e-01 1.27239227e+00 4.82661098e-01
2.76982248e-01 -9.48615931e-03 2.19704255e-01 1.04269528e+00
1.60077140e-02 7.46823177e-02 4.64692652e-01 -1.03689969e+00
5.64235866e-01 3.45436782e-01 2.53333956e-01 -7.82461822e-01
-5.16817331e-01 -5.17724305e-02 -4.27134670e-02 -2.45826945e-01
7.69770443e-01 -3.32105726e-01 -1.30934209e-01 1.71085513e+00
2.56876647e-01 -2.92704195e-01 1.67954236e-01 6.07292533e-01
3.75791788e-01 7.89922893e-01 -4.70983759e-02 -8.27101350e-01
1.11686277e+00 -3.64722699e-01 -9.39547956e-01 -1.14917621e-01
3.93837869e-01 -1.16183543e+00 1.23095858e+00 6.65836453e-01
-1.26563871e+00 -3.21313143e-01 -9.47515190e-01 1.30716100e-01
-5.32310754e-02 -5.22840381e-01 6.34357333e-01 1.09069991e+00
-8.72947633e-01 6.61432564e-01 -6.73536777e-01 -3.04434776e-01
-5.00129879e-01 3.70325893e-01 -9.78979990e-02 6.69019282e-01
-1.18145502e+00 8.71860981e-01 3.78556289e-02 1.13785118e-01
-1.63868725e-01 -3.03094864e-01 -5.42706370e-01 3.94348195e-03
5.98772094e-02 1.30856603e-01 1.85402298e+00 -8.16379488e-01
-1.90606689e+00 5.15718699e-01 -1.77031144e-01 -3.05018127e-01
3.52300137e-01 9.92194116e-02 -6.99130833e-01 2.10966229e-01
-1.67825550e-01 1.75617203e-01 7.02463448e-01 -7.05253184e-01
-9.95796263e-01 -1.12231098e-01 1.52535602e-01 6.32895052e-01
-2.11619139e-01 5.45004964e-01 -2.02410221e-01 -4.91251051e-01
2.18162119e-01 -9.35303390e-01 2.88301438e-01 -6.59205496e-01
-4.19837311e-02 -3.97023678e-01 6.59954429e-01 -7.45382965e-01
1.71822894e+00 -2.20972872e+00 -1.12902887e-01 1.82062149e-01
-2.02134863e-01 2.23293528e-02 4.03213829e-01 9.23069358e-01
1.54209077e-01 1.71225429e-01 2.65490599e-02 -1.85688511e-01
3.60414051e-02 2.24646613e-01 -5.14589787e-01 5.86289823e-01
-1.29633501e-01 4.74363267e-01 -8.13290238e-01 -3.00996870e-01
2.21021548e-01 4.77012545e-01 -4.05439228e-01 3.76806110e-01
2.70331558e-03 5.69516599e-01 -1.51114389e-01 5.23132086e-01
-2.02513114e-01 2.61044532e-01 3.65571290e-01 2.82875150e-01
-6.39586806e-01 1.13171875e+00 -9.34045374e-01 1.18031085e+00
-5.21177828e-01 1.00737548e+00 4.45444673e-01 -5.75010896e-01
1.05489600e+00 8.19674909e-01 6.54864013e-01 -5.49514353e-01
1.16140530e-01 1.31889179e-01 5.98347843e-01 -4.34131980e-01
6.37137234e-01 -4.02732760e-01 -2.66182572e-01 7.36791134e-01
-2.23828927e-01 -4.29112375e-01 2.65783072e-01 -1.42416820e-01
9.78106260e-01 -3.95058662e-01 4.51063782e-01 -1.13664322e-01
3.61340791e-01 -7.84669891e-02 4.55364645e-01 3.71553987e-01
-4.93560463e-01 3.38140756e-01 8.75841379e-01 -4.09711674e-02
-8.83568704e-01 -7.77843356e-01 -2.35021874e-01 1.30482197e+00
-1.16237864e-01 -4.72839445e-01 -7.92129993e-01 -6.23383261e-02
-3.26584518e-01 8.05307150e-01 -3.15523714e-01 1.91856667e-01
-7.37978101e-01 -4.61269796e-01 7.64340460e-01 2.91334838e-01
6.74213693e-02 -1.37660301e+00 -7.08607197e-01 6.85784757e-01
-4.41612542e-01 -8.65939796e-01 -6.02316976e-01 3.93112957e-01
-6.86104178e-01 -9.68792140e-01 -2.22073838e-01 -2.65977383e-01
2.62600426e-02 1.88604072e-01 7.21756995e-01 2.76173837e-03
1.35588469e-02 5.12330592e-01 -5.26521862e-01 -4.39509392e-01
-7.22896636e-01 1.72019303e-01 1.90735294e-03 -7.65119195e-02
2.98108965e-01 -7.68864810e-01 -4.44298446e-01 4.72788692e-01
-3.75612676e-01 -4.52473640e-01 9.70828682e-02 6.12488151e-01
-1.69795044e-02 6.28603697e-02 9.76154447e-01 -8.38124096e-01
1.28280687e+00 -5.13657689e-01 -2.35836864e-01 -2.99149573e-01
-5.58826923e-01 -5.38942754e-01 3.92071515e-01 -4.26839590e-01
-1.04951000e+00 -2.74692237e-01 -5.04412353e-01 2.04372317e-01
-1.53732508e-01 5.89835405e-01 -1.37844503e-01 2.62107849e-01
5.41750371e-01 1.32915705e-01 2.48331010e-01 -3.92885685e-01
2.99612787e-02 1.19719076e+00 5.85949659e-01 -4.84719753e-01
2.22102895e-01 1.15627035e-01 -5.65225482e-01 -1.53135931e+00
-1.93044931e-01 -8.86170983e-01 -6.73645854e-01 -5.12297750e-01
5.07532179e-01 -5.03453612e-01 -7.84156740e-01 3.26646596e-01
-8.73161912e-01 -3.70848507e-01 -1.16920345e-01 5.47372162e-01
-7.72259355e-01 3.96763057e-01 -5.81777990e-01 -1.28714287e+00
-3.16554010e-01 -9.28145885e-01 6.45663440e-01 5.27437069e-02
-1.09918678e+00 -8.52927089e-01 1.57388404e-01 4.81435478e-01
1.73325285e-01 -5.62808141e-02 7.42542982e-01 -7.40333915e-01
1.20089553e-01 -4.20899600e-01 4.27834302e-01 2.34194800e-01
5.03631532e-01 2.45325759e-01 -1.17913544e+00 -1.61891058e-01
2.91413367e-01 -1.75205320e-01 1.00150466e-01 2.46030197e-01
3.95558417e-01 -3.77880216e-01 5.39357178e-02 6.68791905e-02
6.16162956e-01 7.11570144e-01 4.16901946e-01 1.83302641e-01
-2.84319315e-02 9.38379169e-01 6.48159564e-01 6.28868461e-01
1.97983623e-01 7.97988534e-01 3.88210192e-02 4.76711631e-01
1.82682544e-01 -2.11979896e-01 4.37922746e-01 1.24407744e+00
-1.15143061e-01 -3.22442651e-01 -7.79303432e-01 7.37522602e-01
-1.35331774e+00 -1.21596801e+00 4.29151356e-02 2.19424248e+00
8.44204247e-01 5.28120875e-01 6.27438962e-01 8.25731695e-01
5.04239917e-01 5.40300310e-01 -3.28797072e-01 -9.09446120e-01
1.42724171e-01 1.91327073e-02 2.25184970e-02 7.74058282e-01
-7.64513135e-01 6.18433416e-01 7.64698458e+00 3.72109532e-01
-1.28890002e+00 -2.09445462e-01 2.20903233e-01 -1.60118580e-01
-1.35480165e-01 -3.62348557e-02 -5.14831543e-01 5.31731904e-01
1.41994536e+00 -3.86131704e-01 5.56939840e-01 6.17671013e-01
7.92089999e-01 -4.58671451e-01 -9.39584732e-01 7.19658732e-01
-2.49840230e-01 -9.05933678e-01 -6.76199794e-01 2.22886622e-01
5.35582379e-02 -2.05656260e-01 -1.19241737e-01 2.18433708e-01
-1.54631823e-01 -8.11425805e-01 7.53261626e-01 5.61067611e-02
6.14045322e-01 -8.37240100e-01 4.84638989e-01 5.77912390e-01
-1.15772653e+00 -1.27074778e-01 1.71798557e-01 -4.57679003e-01
5.91549873e-01 1.57342792e-01 -1.39015579e+00 1.07846811e-01
4.69231814e-01 3.94059978e-02 2.31242269e-01 5.89680076e-01
-9.20824260e-02 1.12311542e+00 -3.61287445e-01 -4.53237966e-02
2.44449332e-01 -4.02898639e-02 6.41253531e-01 1.22739971e+00
1.42225370e-01 3.58305991e-01 1.15608618e-01 2.42170975e-01
4.69259948e-01 3.17198396e-01 -4.06598777e-01 -5.68822861e-01
9.60826159e-01 9.41120088e-01 -8.62331748e-01 2.77615469e-02
-5.63831925e-01 5.48986673e-01 -1.20917045e-01 7.00447336e-02
-3.61026853e-01 -5.14934480e-01 9.65498567e-01 2.55375147e-01
2.14482769e-02 -4.46234494e-01 -2.47757882e-01 -3.45469564e-01
-4.84432653e-02 -8.91024768e-01 1.08669668e-01 -2.41188496e-01
-7.92295456e-01 6.36471808e-01 9.87514760e-03 -1.00655723e+00
-1.08758581e+00 -1.45831078e-01 -7.70468116e-01 1.13013220e+00
-9.37875569e-01 -4.72453415e-01 2.08516791e-02 3.94673824e-01
7.38372803e-01 -1.87453404e-01 1.03619695e+00 2.60315150e-01
-4.23945516e-01 4.51613814e-01 -2.99229026e-01 -4.67293710e-02
4.89153057e-01 -1.20125699e+00 4.93156940e-01 2.21817881e-01
7.22228140e-02 7.54908144e-01 1.14243412e+00 -4.90976751e-01
-1.05646443e+00 -3.27339768e-01 1.30880475e+00 -4.47829425e-01
7.95931280e-01 -2.73874789e-01 -7.59298503e-01 4.20366764e-01
1.28926978e-01 -6.83078587e-01 9.24027801e-01 4.18366045e-01
1.40763953e-01 7.03726411e-02 -9.84765232e-01 4.49936450e-01
6.78142726e-01 -9.97065246e-01 -8.38386118e-01 -4.27781604e-02
5.57864785e-01 -3.46219540e-01 -1.12149143e+00 -8.15887749e-02
6.67596579e-01 -1.16214955e+00 5.25449216e-01 -1.44356757e-01
-2.32367858e-01 -4.03765030e-02 -8.98911431e-02 -1.13441443e+00
1.59751430e-01 -1.26231027e+00 8.72320756e-02 1.33644032e+00
4.89234030e-01 -5.88988841e-01 7.61013985e-01 5.77956021e-01
-4.06852543e-01 -6.19443297e-01 -1.06366062e+00 -5.34513891e-01
-2.92775542e-01 -6.69197023e-01 4.88675833e-01 7.92806983e-01
7.43027210e-01 3.70069534e-01 -2.97657251e-01 -1.45136550e-01
-1.10091671e-01 -1.59446537e-01 7.26039290e-01 -1.21954000e+00
-2.03175262e-01 -7.16082096e-01 -3.52239877e-01 -1.06646597e+00
-7.52503276e-02 -2.96538562e-01 1.82434246e-01 -1.09867477e+00
-4.38096404e-01 -3.66729617e-01 7.03299418e-03 7.04015195e-02
1.98190436e-01 -1.80945903e-01 -1.98649205e-02 3.51563364e-01
2.35738903e-01 2.09765121e-01 8.28431010e-01 1.69166520e-01
-6.37297928e-01 6.59037888e-01 -4.48284358e-01 7.01025903e-01
9.46701407e-01 -5.41099943e-02 -7.71243691e-01 2.36145556e-01
-4.51322012e-02 6.29059017e-01 -2.10028142e-01 -7.12335467e-01
2.27079615e-01 -3.41198981e-01 -1.04354404e-01 -5.65163136e-01
8.65717709e-01 -4.79181767e-01 2.16313273e-01 8.12923387e-02
-5.15003264e-01 1.83061421e-01 1.84847981e-01 3.57706368e-01
-4.64645624e-01 -2.82272071e-01 5.07725954e-01 3.96970995e-02
-4.67291385e-01 -3.31467211e-01 -7.64969051e-01 2.64866762e-02
7.12676585e-01 -5.82206130e-01 9.43083093e-02 -7.97959030e-01
-7.91473031e-01 -2.90688872e-01 2.21869886e-01 4.42859977e-01
6.61289394e-02 -7.12890983e-01 -2.34074935e-01 1.97823673e-01
-3.14779058e-02 -4.63051289e-01 9.23548862e-02 7.41684616e-01
-4.19639289e-01 7.31116235e-01 3.90722565e-02 -4.39959407e-01
-1.56909740e+00 1.34694874e-01 2.26053342e-01 -2.99182087e-02
-8.00682008e-01 7.77354896e-01 -4.08771396e-01 -1.17394723e-01
5.39164662e-01 -2.93996394e-01 -2.72539884e-01 6.70938969e-01
5.84108353e-01 4.55861866e-01 1.82677448e-01 -9.15863872e-01
-1.72984332e-01 -3.48844007e-02 7.53795132e-02 -7.64795303e-01
9.77535963e-01 -5.67799449e-01 2.92569455e-02 1.07143867e+00
9.11320508e-01 3.04513633e-01 -1.30265319e+00 7.17620924e-02
2.75245368e-01 -4.29839849e-01 -1.04896454e-02 -5.68258941e-01
-3.14462841e-01 7.20477879e-01 3.69457528e-02 9.99391615e-01
9.31207120e-01 -1.34313121e-01 6.69564605e-01 2.24965021e-01
3.22054386e-01 -1.33255231e+00 -2.37478688e-02 6.65453196e-01
8.29883516e-01 -7.33585536e-01 -1.16972588e-01 -3.59137595e-01
-7.41776526e-01 1.01995647e+00 9.09452513e-02 4.83204305e-01
5.92582345e-01 2.79394180e-01 3.69657874e-01 1.22098010e-02
-9.58896279e-01 4.44621369e-02 -9.26064551e-02 4.18400228e-01
9.54845607e-01 2.38566682e-01 -5.41032672e-01 4.33614165e-01
-8.42525184e-01 -3.89120400e-01 4.70627487e-01 1.06498933e+00
-7.21769750e-01 -1.30596960e+00 -3.81812871e-01 4.25656646e-01
-6.61563039e-01 4.37069088e-02 -7.02401757e-01 8.65671635e-01
-2.40358695e-01 1.44197249e+00 1.26801014e-01 -4.61666018e-01
5.33217430e-01 4.75101590e-01 1.96630746e-01 -6.70717239e-01
-9.69557345e-01 4.69251692e-01 7.29237378e-01 -2.86445439e-01
-3.33125710e-01 -1.09213245e+00 -1.22305548e+00 -4.96615678e-01
-3.36186618e-01 6.18792176e-01 8.61793518e-01 9.57037330e-01
-9.43085775e-02 3.56883347e-01 8.30872476e-01 -8.38341594e-01
-6.55067682e-01 -1.33816218e+00 -8.39232564e-01 3.55206057e-02
4.94474262e-01 -4.99649137e-01 -6.11396611e-01 -1.15341000e-01] | [14.249629020690918, 6.710218906402588] |
64e3dbfb-5d6f-4ef0-abcd-733bfac51c89 | spectral-embedding-for-dynamic-networks-with | 2106.01282 | null | https://arxiv.org/abs/2106.01282v2 | https://arxiv.org/pdf/2106.01282v2.pdf | Spectral embedding for dynamic networks with stability guarantees | We consider the problem of embedding a dynamic network, to obtain time-evolving vector representations of each node, which can then be used to describe changes in behaviour of individual nodes, communities, or the entire graph. Given this open-ended remit, we argue that two types of stability in the spatio-temporal positioning of nodes are desirable: to assign the same position, up to noise, to nodes behaving similarly at a given time (cross-sectional stability) and a constant position, up to noise, to a single node behaving similarly across different times (longitudinal stability). Similarity in behaviour is defined formally using notions of exchangeability under a dynamic latent position network model. By showing how this model can be recast as a multilayer random dot product graph, we demonstrate that unfolded adjacency spectral embedding satisfies both stability conditions. We also show how two alternative methods, omnibus and independent spectral embedding, alternately lack one or the other form of stability. | ['Patrick Rubin-Delanchy', 'Andrew Jones', 'Ian Gallagher'] | 2021-06-02 | null | http://proceedings.neurips.cc/paper/2021/hash/5446f217e9504bc593ad9dcf2ec88dda-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/5446f217e9504bc593ad9dcf2ec88dda-Paper.pdf | neurips-2021-12 | ['stochastic-block-model'] | ['graphs'] | [ 2.48865440e-01 2.60342032e-01 -2.18718704e-02 8.54533389e-02
2.11088225e-01 -9.34766233e-01 8.91276777e-01 3.95016998e-01
-1.03559978e-01 4.95425314e-01 4.50825423e-01 -3.76196384e-01
-7.59406209e-01 -1.00566483e+00 -3.35336387e-01 -8.84533823e-01
-7.35408902e-01 3.65751505e-01 1.78330094e-01 -2.20319048e-01
-1.27042264e-01 6.06043458e-01 -1.01279247e+00 -1.64032385e-01
2.21915036e-01 2.72548974e-01 -3.42747241e-01 7.74547517e-01
2.31037557e-01 7.73638010e-01 -2.66911179e-01 -2.44376943e-01
1.41764283e-01 -4.08590555e-01 -7.09521353e-01 1.29349411e-01
1.76369213e-02 1.48689389e-01 -7.37296641e-01 1.01414442e+00
9.69824493e-02 1.46342918e-01 1.01794171e+00 -1.46337128e+00
-8.88985336e-01 6.90747201e-01 -5.68832636e-01 2.83426315e-01
4.53132957e-01 -2.81911910e-01 1.26110828e+00 -3.13593864e-01
7.30512559e-01 9.89147782e-01 9.17161047e-01 2.69761860e-01
-2.13749552e+00 -1.07644491e-01 2.51718879e-01 -3.10725361e-01
-1.35998535e+00 -3.10132325e-01 1.00753057e+00 -8.12163353e-01
3.29114497e-01 6.60970211e-01 8.98985684e-01 1.07969368e+00
3.38860273e-01 -7.38238357e-03 9.84117448e-01 -1.74812362e-01
1.17888018e-01 -4.62366119e-02 4.15904939e-01 7.23707616e-01
4.20068324e-01 -3.19102444e-02 -2.33209088e-01 -5.58519959e-01
5.19816875e-01 2.58022219e-01 -3.47121239e-01 -8.25101018e-01
-1.32092345e+00 8.38265717e-01 4.97779131e-01 8.75604630e-01
-2.14840785e-01 2.85958022e-01 2.35740736e-01 6.40688062e-01
5.09255409e-01 2.50370085e-01 -5.34000285e-02 1.21724851e-01
-7.22725332e-01 2.75663808e-02 8.24259222e-01 4.28570300e-01
9.05244589e-01 -1.16415687e-01 2.23195821e-01 3.40808809e-01
2.99678832e-01 4.31701183e-01 3.20823461e-01 -9.56955194e-01
6.27185926e-02 7.22871542e-01 -2.88213175e-02 -1.78214324e+00
-6.42863870e-01 -5.07292688e-01 -1.28430045e+00 -9.26257577e-03
3.60169679e-01 -2.48171613e-02 -3.42380345e-01 2.49323583e+00
7.07621872e-02 3.79245907e-01 -1.11723907e-01 2.45774418e-01
1.46784544e-01 7.52232671e-01 -1.51831686e-01 -5.40612519e-01
1.02317190e+00 -2.23123819e-01 -6.81128621e-01 1.68183684e-01
6.14159644e-01 -1.96980998e-01 6.23051286e-01 -3.97966146e-01
-1.06056571e+00 -1.47111773e-01 -1.10204983e+00 4.07326877e-01
-3.79804462e-01 -5.56587875e-01 2.94472843e-01 5.66295147e-01
-1.71728480e+00 8.30631554e-01 -9.74460423e-01 -6.05849564e-01
-1.51648119e-01 2.22655654e-01 -6.69760048e-01 3.68833870e-01
-1.27660835e+00 5.59941769e-01 -2.18227640e-01 1.15594596e-01
-5.51152050e-01 -5.62448084e-01 -7.29756415e-01 2.29328692e-01
-1.22182868e-01 -7.18228817e-01 4.95948792e-01 -9.19902742e-01
-8.12506139e-01 7.51931787e-01 9.48126055e-03 -1.95994899e-01
4.82679009e-01 6.64916992e-01 -6.05233312e-01 2.12858006e-01
3.22447985e-01 -5.03277071e-02 6.36231661e-01 -1.32767367e+00
-1.18276915e-02 -6.24752402e-01 7.03242719e-02 1.15936846e-01
-6.46468878e-01 -1.36612222e-01 4.21248823e-02 -5.14763534e-01
5.18895030e-01 -1.13391304e+00 -2.05496579e-01 1.51588097e-01
-3.58502239e-01 -2.78221630e-03 8.31564128e-01 -4.55366492e-01
1.49992442e+00 -2.19816589e+00 6.80043936e-01 7.22904801e-01
7.43965209e-01 -4.04802680e-01 -2.02180266e-01 9.62073743e-01
-5.15147924e-01 5.85776925e-01 -5.87642312e-01 -3.04849267e-01
1.03362352e-01 2.03135833e-01 -1.04510263e-02 1.08700669e+00
-2.86180396e-02 8.60700071e-01 -1.13660991e+00 -2.52701640e-01
-3.03926263e-02 4.68311459e-01 -3.91139925e-01 -3.24623674e-01
5.81543922e-01 1.45539343e-01 -3.60466123e-01 1.01773031e-01
2.87608564e-01 -4.36895460e-01 6.48253500e-01 -3.99674103e-02
-3.08344029e-02 -1.50671199e-01 -1.19981074e+00 1.07115889e+00
-9.23310593e-02 8.52361441e-01 2.77933061e-01 -1.17873597e+00
5.59193492e-01 4.41829711e-01 7.93140471e-01 -2.95891434e-01
-2.44924068e-01 -3.95229831e-02 1.30851403e-01 -2.20337078e-01
3.14309716e-01 -1.37674138e-01 -2.71367013e-01 9.50589299e-01
-2.97705293e-01 5.56850851e-01 -2.44736429e-02 6.25114262e-01
1.60890627e+00 -4.88814235e-01 1.02306873e-01 -5.41486025e-01
2.12605670e-01 -4.19014692e-01 3.79178286e-01 6.40830398e-01
-4.61391479e-01 1.51453212e-01 1.02576196e+00 -3.42172571e-02
-1.31502199e+00 -1.47095513e+00 -2.08195522e-01 6.81444407e-01
2.02539600e-02 -4.28639233e-01 -4.28860456e-01 -2.68898964e-01
2.65616506e-01 2.31501669e-01 -1.28622842e+00 -4.34369892e-01
-3.60315591e-01 -8.41213167e-01 5.13512492e-01 9.52939242e-02
-2.07255222e-02 -6.12056255e-01 -4.52758700e-01 1.29199505e-01
1.31975040e-02 -3.27443540e-01 -5.34426212e-01 4.93632853e-02
-6.56430840e-01 -9.71361399e-01 -7.19037712e-01 -6.80461287e-01
7.98492372e-01 2.64548331e-01 8.55789483e-01 2.44193614e-01
-5.07867262e-02 7.50656366e-01 -8.80947486e-02 6.22938871e-01
-5.78652501e-01 -8.68862793e-02 5.47978044e-01 3.91687989e-01
-2.87843853e-01 -1.11577821e+00 -5.56426227e-01 2.26112470e-01
-1.19711268e+00 -1.84261397e-01 -4.96427417e-02 6.19301856e-01
1.34373546e-01 2.06820980e-01 3.30295801e-01 -6.40697718e-01
8.90722454e-01 -8.88599992e-01 -9.89137366e-02 4.00959969e-01
-7.67450690e-01 1.80093609e-02 3.52360308e-01 -5.31218588e-01
-3.28946590e-01 -4.10099328e-01 4.67293292e-01 -2.68804371e-01
4.28040564e-01 9.03388321e-01 7.24941343e-02 -2.86268815e-02
4.99115944e-01 2.60463178e-01 4.19881523e-01 -3.23679566e-01
6.67069852e-01 3.99724126e-01 3.48668337e-01 -2.58631825e-01
9.86244321e-01 7.65060842e-01 3.54513377e-01 -8.59949350e-01
-1.48922801e-01 -2.41104484e-01 -8.65477085e-01 -2.52723336e-01
6.44187629e-01 -5.18335402e-01 -7.07246065e-01 1.96825370e-01
-6.88433409e-01 -1.49600118e-01 -4.37555015e-01 6.91588819e-02
-5.07040322e-01 5.37584901e-01 -6.29033744e-01 -8.99004936e-01
3.24868411e-01 -5.74738264e-01 5.00097930e-01 -3.69635671e-01
-4.81462866e-01 -1.76213348e+00 7.20202982e-01 -2.46135995e-01
3.79252970e-01 6.15045309e-01 1.11327553e+00 -5.99640071e-01
-1.26270235e-01 -4.55438673e-01 -2.95638870e-02 -8.55907947e-02
6.35079563e-01 4.38276857e-01 -3.25443953e-01 -6.58040047e-01
1.90200116e-02 3.51216644e-01 6.97853982e-01 3.18626076e-01
3.82272393e-01 -7.83609033e-01 -6.89713240e-01 4.22848165e-01
1.33474684e+00 6.09434471e-02 3.57659370e-01 7.35291690e-02
6.67830050e-01 8.86907697e-01 -5.22859693e-01 1.21122763e-01
3.27858776e-01 7.94339061e-01 1.91567376e-01 -1.33624360e-01
2.03586474e-01 -3.67679507e-01 3.70018512e-01 1.04230022e+00
2.95736194e-02 -2.15414882e-01 -9.87242699e-01 8.81751359e-01
-2.06732368e+00 -1.49156511e+00 -2.79218763e-01 2.14079571e+00
4.76345956e-01 -2.60558818e-02 4.52684045e-01 2.11681068e-01
8.18592191e-01 5.14249682e-01 -3.19464296e-01 -2.50809908e-01
-4.33617085e-01 -4.44493473e-01 5.73717594e-01 7.60338008e-01
-7.48970091e-01 3.38456541e-01 7.38504696e+00 3.12019020e-01
-9.13644016e-01 2.29138613e-01 4.66227204e-01 1.37022743e-03
-7.99094737e-01 3.29728395e-01 1.62398741e-01 5.04811406e-01
1.26769543e+00 -6.02432668e-01 4.90064740e-01 1.45050928e-01
3.05923939e-01 3.24613959e-01 -1.02170420e+00 4.90933985e-01
-7.05509633e-02 -1.37619948e+00 -1.53276339e-01 4.03383791e-01
7.35447168e-01 -2.08049044e-01 2.13867843e-01 -3.45989972e-01
7.22546995e-01 -8.67316902e-01 5.75780869e-01 6.62691116e-01
6.94831014e-01 -4.95430082e-01 3.04470181e-01 1.92794517e-01
-1.36328852e+00 3.42893675e-02 1.85303483e-02 -1.63343295e-01
3.02434295e-01 4.51476902e-01 -1.87854707e-01 7.07356930e-01
3.95407081e-01 9.47949588e-01 -5.02511203e-01 5.19217670e-01
2.43122041e-01 4.31148738e-01 -2.71321297e-01 1.77593961e-01
1.78415924e-01 -6.26330912e-01 9.23750222e-01 7.70010650e-01
2.01030746e-01 -1.24418348e-01 -4.58608046e-02 9.72660065e-01
1.80482581e-01 -2.25636929e-01 -1.03134727e+00 -3.83990586e-01
4.78346586e-01 1.04715264e+00 -8.56983066e-01 -1.33661136e-01
-3.55460376e-01 1.18590450e+00 3.68415236e-01 7.33700991e-01
-3.96386743e-01 -2.69924372e-01 9.57144141e-01 4.86219168e-01
-8.87346193e-02 -5.10580361e-01 3.30610201e-02 -1.13658988e+00
-6.34745285e-02 -3.43025476e-01 2.95249969e-01 -5.54205716e-01
-1.44583714e+00 3.59837472e-01 2.46204872e-04 -1.01981652e+00
-1.45396873e-01 -3.05004776e-01 -6.89910710e-01 8.60162616e-01
-6.82573855e-01 -1.02602422e+00 3.40419441e-01 6.02932990e-01
-4.66258228e-01 6.22503497e-02 8.22891295e-01 1.15205772e-01
-8.17596912e-01 3.11941385e-01 6.18828952e-01 1.79234460e-01
1.12048760e-01 -1.33758223e+00 2.36937046e-01 7.53706157e-01
3.54454845e-01 1.04430771e+00 8.22383881e-01 -7.66573727e-01
-1.18523502e+00 -9.96378362e-01 1.12747931e+00 -6.05484545e-01
1.41293478e+00 -5.36460280e-01 -8.47310603e-01 1.05947590e+00
1.01926379e-01 8.48427939e-04 7.34162152e-01 3.33009064e-01
-3.12753022e-01 7.65526965e-02 -9.88849640e-01 9.64602113e-01
1.20039690e+00 -9.27274346e-01 -3.88279110e-01 2.49279037e-01
7.94984877e-01 5.21956265e-01 -1.06482077e+00 1.95352092e-01
6.91118300e-01 -1.06184697e+00 9.39889073e-01 -9.00234163e-01
8.34708810e-02 -3.48251969e-01 -3.11427325e-01 -1.36775172e+00
-9.21831489e-01 -7.85329103e-01 1.65813919e-02 1.13270795e+00
3.99529964e-01 -1.02318811e+00 5.38982093e-01 5.14319360e-01
3.02800417e-01 -5.75429559e-01 -1.29871488e+00 -8.48251104e-01
2.67838031e-01 6.25439733e-02 4.04219002e-01 1.33354700e+00
4.41960871e-01 3.58977318e-01 -3.10852259e-01 8.05664361e-02
5.30756354e-01 -3.50994058e-02 2.73683012e-01 -1.53078830e+00
-1.96871385e-01 -7.54626632e-01 -9.74617779e-01 -5.32251060e-01
3.07732135e-01 -1.03513598e+00 -5.48948824e-01 -1.45466924e+00
4.42472279e-01 -4.40390617e-01 -4.31470245e-01 1.64612874e-01
8.58347416e-02 3.08862895e-01 3.65670323e-02 6.73932314e-01
-2.23649770e-01 4.54788566e-01 6.74545705e-01 -1.81359842e-01
-2.43446231e-01 -1.05071567e-01 -7.62595117e-01 4.79947031e-01
3.44275683e-01 -3.70130062e-01 -6.27689242e-01 5.87530085e-04
7.69575834e-01 3.59218866e-01 5.74751675e-01 -7.08662093e-01
1.76945746e-01 -1.03086449e-01 7.57784173e-02 -3.41449119e-02
3.46598536e-01 -8.95076990e-01 1.00239623e+00 7.73601174e-01
-5.64974844e-01 5.69215775e-01 -9.36272964e-02 1.03266978e+00
1.77109793e-01 1.14057921e-02 5.54190993e-01 8.68086740e-02
6.88472390e-02 3.29727203e-01 -5.92276156e-01 -1.83507040e-01
1.17565215e+00 -4.23246384e-01 -3.31547171e-01 -7.27120996e-01
-1.21762431e+00 1.49197042e-01 8.61729920e-01 1.21169098e-01
2.62137353e-01 -1.76705539e+00 -7.15131998e-01 1.67965740e-01
-6.06567413e-02 -9.59489107e-01 3.37303132e-01 1.10947669e+00
-2.28693634e-01 1.88236654e-01 -7.88594782e-02 -4.47351336e-01
-1.18929982e+00 6.80259585e-01 5.51767230e-01 -1.88197121e-01
-6.52048469e-01 5.56829989e-01 1.39792189e-01 -4.11954492e-01
-2.20196024e-01 1.73221111e-01 -8.80780518e-02 6.08527899e-01
1.99656427e-01 3.86916548e-01 -2.94325262e-01 -1.14269495e+00
-4.31542307e-01 7.42342591e-01 2.99291790e-01 -4.94483858e-01
1.13650453e+00 -6.20951056e-01 -6.58270121e-01 1.14497435e+00
1.42515314e+00 2.43127927e-01 -9.96275127e-01 -1.73547789e-01
-1.11387698e-02 -2.00151652e-01 -1.90338179e-01 -1.29864484e-01
-8.65565658e-01 4.21348244e-01 2.96829462e-01 1.30769598e+00
7.52755344e-01 8.30956027e-02 -7.06211179e-02 6.07561283e-02
7.76720494e-02 -4.43271220e-01 -9.76965353e-02 1.00185551e-01
8.15378666e-01 -5.61985731e-01 -3.21791768e-02 -1.36222497e-01
-1.93887904e-01 8.30936313e-01 -2.33440384e-01 -3.04293215e-01
1.15417469e+00 -3.83076631e-02 -2.77095646e-01 -3.87427598e-01
-8.59576464e-01 6.34993687e-02 1.63701370e-01 6.29565716e-01
4.35464680e-01 3.17126572e-01 -2.67643899e-01 -1.10412516e-01
-1.79280698e-01 -7.58218110e-01 7.93725610e-01 6.59164488e-01
-1.45574063e-01 -9.31735575e-01 -8.77860636e-02 5.18226802e-01
-1.34886861e-01 6.41086176e-02 -7.51372159e-01 9.28140044e-01
-2.92990118e-01 8.83785129e-01 5.61106026e-01 -5.36529243e-01
8.06064382e-02 2.19580382e-01 2.32817218e-01 -5.15220940e-01
-1.44057825e-01 -1.76723287e-01 -3.01184971e-02 -2.69799888e-01
-6.77348256e-01 -1.00229669e+00 -7.27423131e-01 -8.60718310e-01
-1.47804916e-01 2.47146651e-01 2.27103323e-01 5.01735270e-01
3.09274733e-01 5.47114074e-01 9.00585055e-01 -6.43695235e-01
-4.64918613e-01 -7.55889356e-01 -1.10279155e+00 5.93591511e-01
7.83201993e-01 -5.28949976e-01 -1.02347338e+00 -6.19644299e-02] | [7.061751365661621, 5.410226821899414] |
7e7f13df-3609-4a74-965b-8346d5632713 | deep-probabilistic-kernels-for-sample | 1910.05858 | null | https://arxiv.org/abs/1910.05858v2 | https://arxiv.org/pdf/1910.05858v2.pdf | Deep Kernels with Probabilistic Embeddings for Small-Data Learning | Gaussian Processes (GPs) are known to provide accurate predictions and uncertainty estimates even with small amounts of labeled data by capturing similarity between data points through their kernel function. However traditional GP kernels are not very effective at capturing similarity between high dimensional data points. Neural networks can be used to learn good representations that encode intricate structures in high dimensional data, and can be used as inputs to the GP kernel. However the huge data requirement of neural networks makes this approach ineffective in small data settings. To solves the conflicting problems of representation learning and data efficiency, we propose to learn deep kernels on probabilistic embeddings by using a probabilistic neural network. Our approach maps high-dimensional data to a probability distribution in a low dimensional subspace and then computes a kernel between these distributions to capture similarity. To enable end-to-end learning, we derive a functional gradient descent procedure for training the model. Experiments on a variety of datasets show that our approach outperforms the state-of-the-art in GP kernel learning in both supervised and semi-supervised settings. We also extend our approach to other small-data paradigms such as few-shot classification where it outperforms previous approaches on mini-Imagenet and CUB datasets. | ['T. Yong-Jin Han', 'Gauri Joshi', 'Chaitanya Dwivedi', 'Ankur Mallick', 'Bhavya Kailkhura'] | 2019-10-13 | null | null | null | null | ['small-data'] | ['computer-vision'] | [-2.84555584e-01 2.33599152e-02 -2.50232369e-01 -6.03296101e-01
-8.82759869e-01 -3.55533302e-01 7.15670884e-01 3.82463843e-01
-3.73286724e-01 4.16354567e-01 9.95039493e-02 -2.16219411e-03
-3.24403673e-01 -8.95626664e-01 -8.61553729e-01 -7.11993039e-01
-2.48677507e-01 8.26856196e-01 4.54356611e-01 4.70924318e-01
2.80294478e-01 4.41199720e-01 -1.42180037e+00 1.16302975e-01
7.14441359e-01 1.27284813e+00 2.85212956e-02 5.89924395e-01
-2.74353564e-01 8.39033723e-01 -1.53191388e-01 -3.94374877e-01
1.48950279e-01 1.59064442e-01 -4.61371094e-01 -2.57322103e-01
4.34292197e-01 -9.61153880e-02 -4.46350038e-01 9.24754143e-01
5.79702914e-01 4.42396849e-01 1.24974346e+00 -1.36671567e+00
-1.22130477e+00 3.70135725e-01 -4.16898251e-01 2.47351095e-01
6.71425462e-02 -4.83147427e-02 8.92239213e-01 -1.14561331e+00
-3.54343606e-03 1.41013682e+00 1.12759674e+00 3.37949425e-01
-1.67388022e+00 -3.08036536e-01 -1.63895741e-01 9.30935219e-02
-1.36521792e+00 -2.91830033e-01 4.59565848e-01 -5.09413302e-01
1.20320487e+00 -5.34660459e-01 3.34707499e-01 1.20173800e+00
2.85142988e-01 9.79878724e-01 9.04075384e-01 -1.45258337e-01
8.28675687e-01 8.40202123e-02 3.72665852e-01 7.43676841e-01
2.17660621e-01 1.19841471e-01 -5.76044679e-01 -7.24395752e-01
7.97993600e-01 3.92524689e-01 -2.02071667e-01 -8.83897424e-01
-1.05563807e+00 1.20887947e+00 3.96310747e-01 -1.36102960e-01
-4.38880414e-01 5.98860025e-01 3.17288697e-01 2.01067984e-01
7.53601372e-01 2.76538849e-01 -6.18730366e-01 -4.41151470e-01
-7.98941731e-01 3.06046933e-01 1.26420712e+00 8.16036165e-01
8.78091335e-01 -2.26898864e-01 -2.30760008e-01 8.74873638e-01
3.82580966e-01 3.40187371e-01 7.10744023e-01 -6.43203557e-01
3.76374066e-01 3.08323294e-01 5.30964732e-02 -7.58268237e-01
-2.70724505e-01 1.85481742e-01 -7.22858846e-01 9.48892757e-02
3.39349627e-01 -1.00575969e-01 -1.12304688e+00 1.50441766e+00
2.89911836e-01 8.66843462e-01 3.36280912e-01 5.24173439e-01
4.16322351e-01 7.38403082e-01 2.36595303e-01 2.37161979e-01
1.17430151e+00 -8.73758256e-01 -2.93057084e-01 -2.96548605e-01
7.51202643e-01 -3.61192733e-01 1.20678604e+00 1.27611980e-01
-7.12177038e-01 -4.22895283e-01 -7.69081831e-01 2.61217579e-02
-6.44653440e-01 1.87682495e-01 8.28946173e-01 7.65888929e-01
-9.94021475e-01 1.02739835e+00 -1.21842384e+00 -4.08017933e-01
1.06924021e+00 4.05869782e-01 -3.42662752e-01 -3.76947641e-01
-1.05625057e+00 9.83103275e-01 7.99776375e-01 -2.16969535e-01
-8.30308795e-01 -1.15941060e+00 -1.08168137e+00 2.66885847e-01
2.12086365e-01 -7.44503915e-01 1.13700593e+00 -2.29278341e-01
-1.62853181e+00 4.83925819e-01 2.14766622e-01 -7.79233158e-01
1.13794021e-01 -4.25277442e-01 -1.68231532e-01 1.19603135e-01
-3.29703875e-02 7.99953103e-01 1.28423381e+00 -6.98684454e-01
-6.64179027e-01 -2.71273643e-01 -2.63101518e-01 1.04560815e-01
-3.42488796e-01 -7.85626546e-02 -2.69111156e-01 -3.25763106e-01
1.54097844e-02 -1.02668822e+00 -1.70352712e-01 1.12842910e-01
-2.89402083e-02 -5.75730383e-01 9.14526820e-01 -1.98062703e-01
6.77260339e-01 -1.95605671e+00 -1.26927011e-02 1.59630105e-01
1.86080083e-01 2.37447858e-01 2.78052501e-02 2.80965149e-01
-5.88053055e-02 -1.35244429e-01 -2.84601688e-01 -6.04711831e-01
4.31857228e-01 5.18763125e-01 -6.23408496e-01 6.09921396e-01
5.85013986e-01 1.14811945e+00 -1.09566879e+00 -1.24476098e-01
3.09311420e-01 7.12996423e-01 -2.88134515e-01 3.27511668e-01
-9.80572551e-02 -3.26412678e-01 -2.66750991e-01 6.40601456e-01
5.43761075e-01 -3.97414505e-01 -5.81689000e-01 -7.84518868e-02
5.94616175e-01 3.04708928e-02 -1.27164137e+00 1.74373007e+00
-3.98580998e-01 4.67526466e-01 -6.44636571e-01 -1.20736516e+00
9.49529588e-01 1.44791409e-01 2.22859636e-01 -1.71862226e-02
-1.87685918e-02 2.56436821e-02 -2.93560892e-01 -1.42468363e-01
4.03400451e-01 -2.28683114e-01 -1.41549245e-01 5.67743242e-01
5.38773358e-01 -3.96141052e-01 7.87041932e-02 1.53095759e-02
1.04611599e+00 1.57269448e-01 1.02157697e-01 -3.28097731e-01
1.28128365e-01 -2.00649381e-01 1.75140202e-01 9.33587611e-01
-9.48637873e-02 6.92339778e-01 7.15142667e-01 -6.49518549e-01
-1.05002177e+00 -1.39253199e+00 -2.57175952e-01 1.13267994e+00
-1.66492656e-01 -4.10952330e-01 -4.76453453e-01 -8.20304573e-01
5.66388130e-01 8.32778394e-01 -8.98423910e-01 -5.72099864e-01
1.05411202e-01 -1.06124663e+00 6.33881748e-01 9.84287322e-01
1.68556079e-01 -7.88904011e-01 -3.12785566e-01 2.42934883e-01
5.23341119e-01 -1.07232499e+00 -3.79172295e-01 6.37915015e-01
-1.03395867e+00 -1.00996327e+00 -6.34538233e-01 -6.52796984e-01
5.47844768e-01 1.32787958e-01 9.55447078e-01 -8.88361931e-01
-2.69597530e-01 6.53327644e-01 -8.15782845e-02 -6.74264193e-01
-1.02183901e-01 -1.03043311e-03 5.70532262e-01 -2.25686759e-01
9.75750506e-01 -5.84290445e-01 -3.20164382e-01 1.64673880e-01
-8.72244418e-01 -5.63506067e-01 5.29148400e-01 1.02935898e+00
7.23269701e-01 2.64140695e-01 3.49243224e-01 -6.33539021e-01
1.22605956e+00 -7.73237705e-01 -5.36862314e-01 1.82711512e-01
-6.01273358e-01 6.52223647e-01 5.55165231e-01 -1.01248968e+00
-7.71363735e-01 -5.30290604e-02 3.18544865e-01 -1.04865539e+00
-2.60754675e-01 5.06308019e-01 2.26789549e-01 -4.15136851e-02
8.99881542e-01 1.44714087e-01 3.88567001e-02 -3.88928115e-01
7.73637533e-01 5.14161885e-01 4.27720666e-01 -8.91783655e-01
4.89447147e-01 5.17198026e-01 1.94892678e-02 -5.54715157e-01
-9.67299700e-01 -6.04026616e-01 -7.07212925e-01 5.41376114e-01
7.34551966e-01 -1.09371173e+00 -5.63933969e-01 3.07751924e-01
-7.43192852e-01 -3.67616713e-01 -7.07881391e-01 8.61295044e-01
-1.03268933e+00 2.38801718e-01 -6.93310142e-01 -6.36335731e-01
-3.47372413e-01 -8.52551401e-01 1.27423370e+00 2.58492202e-01
-1.57641023e-01 -1.32661963e+00 4.29673135e-01 -3.09455931e-01
3.48069429e-01 -1.97196871e-01 8.75169277e-01 -1.03490138e+00
-1.15953334e-01 -5.48554182e-01 -3.70969951e-01 5.59363842e-01
1.97683796e-01 -1.41302034e-01 -1.02704525e+00 -9.11465511e-02
-1.85665451e-02 -8.20504606e-01 1.12111652e+00 5.47156990e-01
1.24785125e+00 -5.26575651e-03 -4.13465351e-01 7.52464592e-01
1.36062610e+00 -3.82086277e-01 5.39124727e-01 3.82017270e-02
8.28243315e-01 4.07930762e-01 3.70109171e-01 3.66409600e-01
3.77782106e-01 2.34377399e-01 8.69509056e-02 3.81222963e-01
3.98601949e-01 -5.71453750e-01 3.08819294e-01 3.85626704e-01
2.02302977e-01 7.09301829e-02 -1.16420174e+00 5.29075205e-01
-2.25717711e+00 -9.72904146e-01 3.00174385e-01 2.22515726e+00
8.11300159e-01 2.11669132e-01 -1.03354171e-01 -4.03684855e-01
5.42100310e-01 3.49119008e-02 -6.18935525e-01 -3.32256883e-01
1.90373644e-01 3.66665065e-01 6.12166047e-01 4.59566992e-03
-1.53148234e+00 1.06466115e+00 6.48341322e+00 1.13795531e+00
-9.29309249e-01 1.68465078e-02 5.67251503e-01 -1.67264551e-01
2.28355631e-01 -1.58038124e-01 -9.82595980e-01 6.17582321e-01
1.49433196e+00 -2.20837116e-01 2.28775948e-01 1.48802638e+00
-2.39082709e-01 -1.95066378e-01 -1.48907578e+00 1.47091115e+00
4.84971330e-02 -1.43287516e+00 -9.76179242e-02 -6.81316108e-02
8.07850122e-01 4.07460630e-01 1.68216422e-01 6.72051549e-01
1.01545155e+00 -1.18540823e+00 2.96389401e-01 7.51318991e-01
4.05649424e-01 -9.79085028e-01 7.79801726e-01 4.74581867e-01
-9.38071728e-01 -3.24894302e-02 -1.29133439e+00 5.86507469e-02
1.26850046e-02 7.33768225e-01 -9.56744432e-01 1.92031823e-02
7.87871301e-01 8.03292334e-01 -4.06901002e-01 1.13490689e+00
-2.70081758e-01 5.03370762e-01 -6.75482273e-01 -4.30674516e-02
4.09135133e-01 -2.57040888e-01 -9.16857459e-03 1.10586369e+00
4.80210304e-01 -2.05590516e-01 1.61406800e-01 1.15514028e+00
-1.51930466e-01 -1.52042449e-01 -8.42742682e-01 -4.60097104e-01
3.59005719e-01 1.12979078e+00 -6.43688858e-01 -4.79337603e-01
-5.65695643e-01 1.17469835e+00 9.74627435e-01 3.67420346e-01
-6.04988158e-01 -3.59313190e-01 1.15603757e+00 -3.37971717e-01
8.58886242e-01 -3.41443449e-01 2.50494704e-02 -1.26326275e+00
-1.91107705e-01 -2.33737007e-01 3.67651045e-01 -7.43575513e-01
-2.03992414e+00 2.45930091e-01 1.74592316e-01 -1.07701898e+00
-4.86455768e-01 -9.01753068e-01 -7.01175749e-01 1.13640845e+00
-1.50080884e+00 -1.18092394e+00 7.03744516e-02 6.59944952e-01
2.48964533e-01 -2.10150704e-01 1.12941432e+00 -2.66931713e-01
-2.74776995e-01 2.45025277e-01 6.64074302e-01 -1.12847224e-01
8.77824605e-01 -1.58669627e+00 7.49480128e-01 4.02039587e-01
4.52770263e-01 6.81837142e-01 3.85966122e-01 -6.23299718e-01
-1.50632823e+00 -1.16703570e+00 5.21477103e-01 -9.14830029e-01
1.10614836e+00 -4.62411135e-01 -1.17860484e+00 5.89652717e-01
-3.76371890e-01 6.89791322e-01 1.13332975e+00 2.62990922e-01
-7.01331198e-01 2.12326497e-01 -1.22812510e+00 3.58303815e-01
6.42917931e-01 -9.34148610e-01 -1.02193248e+00 3.95869315e-01
8.37258756e-01 -3.81263733e-01 -1.16831875e+00 6.24163076e-02
3.58143896e-01 -5.29651344e-01 1.10508764e+00 -9.75590110e-01
3.31135035e-01 -1.29576772e-01 -2.01074332e-01 -1.74230158e+00
-4.01720971e-01 -3.46549332e-01 -8.58820677e-01 7.99120724e-01
2.52203703e-01 -5.92309773e-01 9.40914452e-01 1.21505630e+00
1.28750578e-01 -9.87341642e-01 -7.22244382e-01 -9.04674709e-01
1.78493962e-01 -8.41240406e-01 4.94936794e-01 8.28441501e-01
1.22167341e-01 1.85296386e-01 -2.52070338e-01 4.20830101e-01
8.41035485e-01 3.52780558e-02 6.99872196e-01 -1.44740772e+00
-2.41624624e-01 -3.08973551e-01 -9.15106893e-01 -8.57359290e-01
4.53805685e-01 -7.06314743e-01 5.64032719e-02 -1.47548425e+00
-5.18128043e-03 -3.98030251e-01 -3.50737423e-01 3.73203367e-01
-3.89320433e-01 -3.74335758e-02 -1.22983165e-01 3.85787070e-01
-6.61171973e-01 8.54098916e-01 4.02454734e-01 -2.36225314e-02
-3.91629785e-01 1.07331052e-01 -6.04509711e-01 8.34453046e-01
7.68729866e-01 -6.48766339e-01 -7.08056986e-01 -1.35081410e-01
-1.90274268e-02 -5.26556909e-01 3.28208804e-01 -1.15973270e+00
4.74883974e-01 -5.10235354e-02 7.60300040e-01 -5.58004677e-01
5.40647447e-01 -7.89110839e-01 -2.54981220e-01 5.32176308e-02
-3.33576679e-01 -2.61541426e-01 2.28268757e-01 1.21110570e+00
-1.52577370e-01 -3.46758664e-01 6.26258135e-01 -5.95239326e-02
-8.82733285e-01 8.33306074e-01 -9.67094973e-02 1.86680406e-01
1.13504183e+00 -2.16409802e-01 -1.91872433e-01 -1.52199686e-01
-6.61722660e-01 3.28345060e-01 5.40857911e-01 4.82697427e-01
7.44567096e-01 -1.52943420e+00 -3.16475570e-01 3.25581789e-01
4.80453163e-01 1.52562454e-01 8.74153748e-02 6.32359862e-01
-3.17489207e-01 4.10570264e-01 -4.64293733e-02 -7.54669905e-01
-7.41886377e-01 8.05869401e-01 2.04172403e-01 -1.95928827e-01
-6.38119340e-01 1.06575775e+00 9.95062664e-02 -6.55160725e-01
3.52482647e-01 -8.26549828e-01 6.34185672e-02 9.66239572e-02
8.45148981e-01 2.28505969e-01 -8.86626020e-02 -2.76423097e-01
-6.83680177e-02 2.71555483e-01 -2.08668455e-01 1.02061287e-01
1.59346175e+00 1.16085030e-01 1.83671266e-01 7.86728442e-01
1.52527833e+00 -7.60827005e-01 -1.78088069e+00 -6.78832293e-01
2.27073744e-01 -5.57591379e-01 1.46154493e-01 -1.45942926e-01
-4.85770583e-01 1.13950717e+00 5.81685841e-01 -2.52592228e-02
5.43570280e-01 3.78876716e-01 4.91186380e-01 9.69821811e-01
3.35125238e-01 -1.25514698e+00 1.97370157e-01 6.32036746e-01
3.91591996e-01 -1.45601606e+00 -1.47013277e-01 -8.85928646e-02
-9.69278932e-01 1.30837858e+00 2.84446716e-01 -5.45450091e-01
1.08289039e+00 1.69391379e-01 -4.93303388e-01 -2.64628768e-01
-9.48615193e-01 -2.46418729e-01 4.21748757e-01 1.14292419e+00
7.24885671e-04 3.63997146e-02 4.17177051e-01 5.89438200e-01
1.01256199e-01 1.95796534e-01 3.05752933e-01 9.86406326e-01
-5.55978835e-01 -9.10539329e-01 -2.03778818e-01 9.88687694e-01
-2.49598593e-01 -3.33654344e-01 1.38686970e-01 4.82701629e-01
-3.20029110e-01 3.27840000e-01 3.50338131e-01 -1.04085408e-01
1.61439016e-01 5.21254182e-01 4.02531445e-01 -8.47203791e-01
7.52742290e-02 -3.44893694e-01 -2.12377533e-01 -6.97887003e-01
-2.19805360e-01 -5.53586245e-01 -1.07126844e+00 -2.71305650e-01
-1.95999652e-01 6.84935674e-02 7.24799871e-01 8.04866254e-01
4.92364347e-01 7.77308047e-02 1.08949162e-01 -1.06231380e+00
-1.24199188e+00 -1.00556421e+00 -1.02279615e+00 4.49106157e-01
-1.78019740e-02 -9.51567709e-01 -4.36722934e-01 2.99345963e-02] | [7.403618812561035, 3.7359812259674072] |
342d75d8-c1bb-47e1-9e01-9492a98cd07c | fragment-graphical-variational-autoencoding | 1910.13325 | null | https://arxiv.org/abs/1910.13325v2 | https://arxiv.org/pdf/1910.13325v2.pdf | Fragment Graphical Variational AutoEncoding for Screening Molecules with Small Data | In the majority of molecular optimization tasks, predictive machine learning (ML) models are limited due to the unavailability and cost of generating big experimental datasets on the specific task. To circumvent this limitation, ML models are trained on big theoretical datasets or experimental indicators of molecular suitability that are either publicly available or inexpensive to acquire. These approaches produce a set of candidate molecules which have to be ranked using limited experimental data or expert knowledge. Under the assumption that structure is related to functionality, here we use a molecular fragment-based graphical autoencoder to generate unique structural fingerprints to efficiently search through the candidate set. We demonstrate that fragment-based graphical autoencoding reduces the error in predicting physical characteristics such as the solubility and partition coefficient in the small data regime compared to other extended circular fingerprints and string based approaches. We further demonstrate that this approach is capable of providing insight into real world molecular optimization problems, such as searching for stabilization additives in organic semiconductors by accurately predicting 92% of test molecules given 69 training examples. This task is a model example of black box molecular optimization as there is minimal theoretical and experimental knowledge to accurately predict the suitability of the additives. | ['Henning Sirringhaus', 'Gareth Conduit', 'Ivan Dimov', 'Lorena Marañón', 'Ian E. Jacobs', 'John Armitage', 'Iyad Nasrallah', 'Dimitrios Simatos', 'Mark Nikolka', 'Malgorzata Nguyen', 'Leszek J. Spalek', 'Christian B. Nielsen', 'Iain McCulloch', 'Guillaume Schweicher'] | 2019-10-21 | null | null | null | null | ['small-data'] | ['computer-vision'] | [ 4.93473619e-01 -5.69437444e-02 -3.84173512e-01 -3.30381602e-01
-7.61641741e-01 -6.02337301e-01 4.09623027e-01 6.52449489e-01
-2.76445508e-01 1.28544497e+00 -2.33546048e-01 -6.59585238e-01
-2.86264569e-01 -9.68190014e-01 -1.00909579e+00 -8.01568925e-01
1.00765740e-02 6.27765954e-01 -1.64262980e-01 -1.94154456e-01
5.63462973e-01 6.54361010e-01 -1.74092066e+00 2.54326522e-01
1.23912776e+00 1.07624471e+00 3.39250058e-01 6.12087727e-01
-8.09279084e-02 4.46048468e-01 -5.61666489e-01 -3.26856643e-01
8.92585889e-02 -4.45127994e-01 -5.62108338e-01 -3.71601880e-01
4.43487406e-01 -2.49618795e-02 4.84941304e-02 8.61129165e-01
7.59434462e-01 2.51768082e-01 7.86915839e-01 -5.90307891e-01
-5.86689949e-01 5.49797893e-01 1.91465557e-01 -2.64354140e-01
6.59263015e-01 2.29446799e-01 1.08561182e+00 -7.02333272e-01
7.77523994e-01 9.39129412e-01 6.76776171e-01 3.49726170e-01
-1.41245008e+00 -4.26054448e-01 -2.90484548e-01 2.58984447e-01
-1.10777879e+00 -4.98765558e-01 5.47373772e-01 -5.74647546e-01
1.19870210e+00 3.02282900e-01 6.74473047e-01 8.15262198e-01
2.83999890e-01 1.96107015e-01 8.66660357e-01 -6.47917926e-01
6.29868925e-01 3.53536129e-01 -3.47443596e-02 7.20141053e-01
4.12615567e-01 4.57047433e-01 -4.59299713e-01 -4.25719351e-01
3.60177606e-01 -7.65324682e-02 -2.50935584e-01 -4.41527843e-01
-8.20249081e-01 1.08898664e+00 5.58560252e-01 1.35437816e-01
-3.75115812e-01 -1.23466123e-02 1.37060657e-01 4.77714151e-01
1.10783398e-01 1.36579406e+00 -6.16199672e-01 -1.12489089e-01
-7.76846588e-01 2.07392722e-01 1.04846179e+00 5.01203299e-01
9.94675636e-01 1.94691062e-01 2.26000220e-01 4.94289577e-01
5.31434119e-02 1.33120149e-01 4.62999284e-01 -6.57164931e-01
3.25444698e-01 6.29183888e-01 2.31503695e-01 -6.99818611e-01
-4.56610233e-01 -3.12677622e-01 -3.04562300e-01 3.16555142e-01
3.65982831e-01 -8.00354779e-02 -9.34427142e-01 1.35143805e+00
3.06520671e-01 -4.35660243e-01 2.45867535e-01 4.99296755e-01
8.86911452e-01 9.62928414e-01 1.46251097e-01 -3.46998483e-01
7.68683255e-01 -5.49062848e-01 -2.55619824e-01 7.58744627e-02
9.51479852e-01 -6.58072829e-01 9.92698252e-01 5.17853498e-01
-7.90466726e-01 -4.99598503e-01 -1.40583539e+00 2.01205790e-01
-8.04816008e-01 1.76543221e-01 1.15195608e+00 1.03324974e+00
-6.81144357e-01 1.44123423e+00 -6.64060771e-01 -4.94670793e-02
2.81540871e-01 1.08264494e+00 -6.11025751e-01 -1.43074632e-01
-1.01647663e+00 9.66946781e-01 8.59507561e-01 -3.54895517e-02
-5.72817743e-01 -8.22390378e-01 -8.32076371e-01 1.29935160e-01
1.80798426e-01 -5.10636628e-01 7.09799707e-01 -1.08094823e+00
-1.81506956e+00 2.53020704e-01 2.07231697e-02 -5.36332130e-01
8.89959559e-02 1.89971939e-01 -3.53872746e-01 1.69904362e-02
-2.85921276e-01 7.67028391e-01 6.31224930e-01 -8.80320430e-01
-1.55001372e-01 -1.49618715e-01 -5.02027944e-02 -1.23805068e-01
-3.66162121e-01 -1.69520199e-01 1.20471366e-01 -2.79356837e-01
-1.67609379e-01 -9.19279635e-01 -5.38710177e-01 -3.65237474e-01
-4.36710715e-01 2.92980280e-02 5.28091371e-01 -4.72194076e-01
1.03375280e+00 -1.52323747e+00 3.05003256e-01 5.37731588e-01
7.71289244e-02 3.70719075e-01 -2.02508241e-01 8.32765520e-01
-3.56666505e-01 2.53051519e-01 -1.63940176e-01 4.39115882e-01
-2.36561537e-01 -1.19739980e-01 3.41377570e-03 3.66123110e-01
2.64784783e-01 8.65516186e-01 -7.42474258e-01 -2.02621460e-01
3.42858940e-01 4.88270372e-01 -8.26427221e-01 1.97061509e-01
-6.52711749e-01 3.18916023e-01 -4.92804825e-01 9.04615879e-01
2.57760346e-01 -3.63649249e-01 2.14752629e-01 -1.84847593e-01
-2.52266616e-01 3.85883629e-01 -9.49264765e-01 1.31998956e+00
-2.72129297e-01 5.15976131e-01 -5.35103917e-01 -1.09240258e+00
8.87201965e-01 1.01176284e-01 4.99805808e-01 -6.86693013e-01
2.24963650e-02 4.63623583e-01 4.28084761e-01 -3.31494004e-01
2.21889004e-01 -4.17994350e-01 2.26558283e-01 3.30641568e-01
3.07043910e-01 -3.69954497e-01 3.17467153e-01 -2.29813218e-01
1.03764963e+00 1.27824917e-01 3.77145916e-01 -3.13405544e-02
4.18744981e-01 1.43501163e-01 2.67150491e-01 7.78430820e-01
3.93906027e-01 2.90904105e-01 3.55655581e-01 -7.84683764e-01
-1.33629346e+00 -4.31192160e-01 -3.41512322e-01 8.36642563e-01
-1.78671733e-01 -7.00568318e-01 -5.40641546e-01 -2.44402140e-01
9.28014144e-02 4.78268892e-01 -5.34531295e-01 -4.13603038e-01
-2.47738764e-01 -1.06372857e+00 8.74151811e-02 2.54059017e-01
-3.13144684e-01 -1.12041283e+00 -2.98905015e-01 5.75385571e-01
6.49807870e-01 -6.99106574e-01 2.54284471e-01 7.33969092e-01
-8.26129019e-01 -9.65357363e-01 -4.10956949e-01 -3.84223819e-01
6.81893408e-01 -2.92296559e-01 9.57924247e-01 8.65599737e-02
-7.06981778e-01 -9.04850438e-02 -2.42056474e-01 -4.28754330e-01
-6.40116930e-01 2.32174154e-02 1.55608162e-01 -2.70580709e-01
9.36067849e-02 -6.29170597e-01 -6.45811558e-01 1.09648734e-01
-4.91451561e-01 -3.59970212e-01 8.49732637e-01 1.10529578e+00
7.51298130e-01 7.42082596e-02 5.42391896e-01 -7.61729419e-01
7.40124285e-01 -2.32370034e-01 -8.16269875e-01 3.80855024e-01
-8.22898149e-01 6.33300900e-01 8.45219195e-01 -5.58498025e-01
-5.39090514e-01 3.63540441e-01 -3.40159953e-01 -8.87503400e-02
-1.71232268e-01 9.30215299e-01 -3.46818566e-01 -8.73964906e-01
7.91585028e-01 9.06458795e-02 -1.27119258e-01 -3.81618589e-01
1.11271575e-01 4.61397797e-01 1.94770359e-02 -8.76628399e-01
3.29913467e-01 -1.46601409e-01 3.70748758e-01 -1.06035388e+00
-4.03340638e-01 -1.44780397e-01 -6.57884955e-01 1.01023898e-01
4.84160274e-01 -5.32364607e-01 -1.16583097e+00 -1.57813460e-01
-8.56278419e-01 -2.87841976e-01 -2.93913931e-01 6.23580456e-01
-5.59842587e-01 4.06694442e-01 -2.74324358e-01 -6.95827901e-01
-4.36674863e-01 -1.56023848e+00 9.92985606e-01 2.07598582e-01
-3.33184540e-01 -8.03930819e-01 3.80572528e-02 2.48979419e-01
3.14453244e-01 4.06324893e-01 1.36884391e+00 -8.42225730e-01
-7.70432353e-01 -7.74956405e-01 1.52057454e-01 1.76200196e-01
1.60006553e-01 1.71380967e-01 -7.32225657e-01 -2.08694890e-01
-4.09101397e-01 -5.67470372e-01 8.26384842e-01 5.47367036e-01
1.24725461e+00 -3.52090865e-01 -4.26958054e-01 6.13354087e-01
1.39148772e+00 6.35618925e-01 6.47791982e-01 2.91184545e-01
7.51916647e-01 6.55022502e-01 4.04550105e-01 3.46991420e-01
-3.02310854e-01 7.09952176e-01 2.07188144e-01 7.34575931e-03
3.90043586e-01 -3.03709805e-01 2.94825047e-01 2.54377455e-01
-3.13214839e-01 -4.08103913e-01 -9.14874792e-01 -4.22460474e-02
-1.42957568e+00 -1.03425443e+00 1.03789352e-01 2.64014816e+00
1.05073714e+00 2.72941023e-01 2.43752286e-01 2.09605739e-01
4.97157991e-01 -3.42552811e-01 -7.27226079e-01 -6.58669770e-01
-1.39383614e-01 6.25176311e-01 5.52664101e-01 5.68892360e-01
-8.92808437e-01 9.45047557e-01 7.18749142e+00 8.79456937e-01
-1.43372858e+00 -4.30616170e-01 7.59688914e-01 6.08623552e-04
-2.53656507e-01 1.63494572e-01 -8.67412031e-01 4.41210002e-01
1.48602891e+00 2.62038894e-02 5.25909841e-01 8.26420367e-01
-2.39353385e-02 -2.05320995e-02 -1.37003386e+00 8.26779127e-01
-2.94290870e-01 -1.82739604e+00 2.56279290e-01 3.06690782e-01
6.57094955e-01 -1.86839774e-01 1.82117492e-01 7.80602098e-02
5.34878150e-02 -1.58559203e+00 3.40540588e-01 5.10891914e-01
8.70129466e-01 -8.41684103e-01 5.16098619e-01 1.43129647e-01
-6.56294703e-01 -2.30001986e-01 -6.33829415e-01 -8.65868479e-02
-1.39015004e-01 4.33582217e-01 -1.12968993e+00 3.27262819e-01
1.20641276e-01 4.18595463e-01 -4.86121029e-01 1.18656826e+00
2.60759890e-01 4.91955131e-01 -6.26253963e-01 -4.81905460e-01
3.10835809e-01 -4.99224395e-01 2.02204570e-01 7.87144661e-01
5.37857413e-01 -3.87466997e-02 1.46509960e-01 9.67312694e-01
-6.56339079e-02 3.01414847e-01 -6.85991347e-01 -7.19055712e-01
3.40577245e-01 8.42041433e-01 -5.14160752e-01 -1.11331306e-01
-2.12242827e-01 5.79229712e-01 1.56891048e-01 2.11991444e-01
-3.63326967e-01 -5.83598077e-01 3.04931313e-01 8.62104595e-02
4.45408255e-01 -1.47294402e-01 7.11244252e-03 -7.61028409e-01
-3.06317151e-01 -1.08229494e+00 1.80332243e-01 -5.81763208e-01
-7.86535382e-01 4.66317415e-01 -4.53923494e-01 -8.56026113e-01
-2.25880936e-01 -1.18868732e+00 -5.37198246e-01 7.56286860e-01
-1.09034562e+00 -9.06726718e-01 2.04454899e-01 1.49207473e-01
1.90684110e-01 -4.23236728e-01 1.08369184e+00 1.11059852e-01
-5.50686181e-01 5.68847477e-01 7.13581324e-01 -6.30688071e-01
5.98003507e-01 -1.37240732e+00 5.22159077e-02 1.75692797e-01
1.73574731e-01 7.81340003e-01 9.60942447e-01 -6.77758813e-01
-1.81761503e+00 -6.50613546e-01 4.68492955e-01 -4.87258434e-01
5.55100918e-01 -3.82042229e-01 -6.77864432e-01 6.75126314e-02
-3.55671823e-01 4.33996879e-02 1.12981033e+00 2.90949136e-01
-9.55761820e-02 -2.31351424e-02 -9.78959680e-01 3.69380265e-01
6.73273385e-01 -5.97994983e-01 -7.61158625e-03 8.07421088e-01
6.65372252e-01 -3.26427490e-01 -1.26530576e+00 3.62184644e-01
6.94886029e-01 -7.28801489e-01 1.23425972e+00 -1.13600862e+00
4.74997669e-01 -5.35828359e-02 -2.02564344e-01 -1.15297973e+00
3.49041484e-02 -7.40847588e-01 -1.24434620e-01 7.61120737e-01
9.45063472e-01 -4.48351383e-01 1.04425538e+00 8.85024428e-01
-5.43521941e-02 -9.75064039e-01 -8.95375252e-01 -7.56060421e-01
-2.98742950e-02 -1.73592404e-01 7.59947002e-01 6.47816241e-01
1.41038343e-01 3.44194293e-01 -3.57397258e-01 -6.10315837e-02
3.06613892e-01 3.85647416e-01 5.16934633e-01 -1.36778140e+00
-5.83416760e-01 -3.37746143e-01 -7.12614000e-01 -5.99899054e-01
2.60653645e-01 -8.05755496e-01 -2.56735951e-01 -1.02416956e+00
-6.15192913e-02 -6.15422547e-01 -3.65053535e-01 1.86641589e-01
1.54721275e-01 -1.32495776e-01 -2.64286816e-01 -1.19349044e-02
-4.33829129e-01 6.10733628e-01 9.84673798e-01 -3.98704648e-01
-6.56801283e-01 1.10184610e-01 -7.46076047e-01 2.09905833e-01
5.78287542e-01 -3.89510393e-01 -2.12989941e-01 1.83819070e-01
6.78325236e-01 2.90793896e-01 8.86163786e-02 -1.14540935e+00
2.14272290e-02 -2.85318404e-01 7.09623694e-01 -3.74509007e-01
4.59746987e-01 -6.10601246e-01 2.07629800e-01 5.17095864e-01
-4.25810307e-01 -2.83714533e-01 2.09532306e-01 7.17197120e-01
-1.40820310e-01 -4.80182350e-01 4.86337215e-01 -1.19532540e-01
-4.74090308e-01 5.08921504e-01 -2.35337630e-01 -5.60492814e-01
7.69782424e-01 -5.10218501e-01 -1.15646422e-01 -2.79471010e-01
-6.61708593e-01 -2.58613050e-01 5.93545020e-01 -6.53769672e-02
6.28328502e-01 -9.58312154e-01 -1.94831237e-01 1.08099878e-01
1.14962645e-01 -4.26224947e-01 -9.49998051e-02 4.30904686e-01
-7.40895271e-01 8.21512520e-01 -3.19158584e-01 -3.53675425e-01
-1.21075809e+00 8.97371650e-01 4.94519204e-01 -1.32377684e-01
-1.90671355e-01 9.35199976e-01 -1.27336770e-01 -3.55992079e-01
-2.68354192e-02 -1.91541925e-01 -5.06564304e-02 -4.35741507e-02
3.38289231e-01 3.43644284e-02 4.67390031e-01 -2.41614312e-01
-2.47521132e-01 5.11486471e-01 -2.13422656e-01 3.32369685e-01
1.76184034e+00 5.52287996e-01 -5.46477661e-02 -3.45607311e-03
1.29878986e+00 -1.29486501e-01 -1.02206075e+00 2.44611129e-01
1.55386314e-01 -2.44245842e-01 2.25301102e-01 -7.98862755e-01
-3.96368533e-01 9.87136841e-01 7.77730048e-01 -5.02335690e-02
7.01931894e-01 -8.38791877e-02 4.93390471e-01 1.17397833e+00
8.20988268e-02 -1.12572169e+00 -3.42624150e-02 2.04826996e-01
8.05768609e-01 -1.35560191e+00 3.48701954e-01 -2.16530189e-01
-3.17514032e-01 1.46382284e+00 3.65399361e-01 3.13489497e-01
3.21197838e-01 2.55045686e-02 -5.50685465e-01 -4.23769951e-01
-6.21303022e-01 -7.97752943e-03 3.82254839e-01 6.85353696e-01
6.20514750e-01 -1.89157519e-02 -1.47127405e-01 3.56548280e-01
-3.47264677e-01 -2.66019642e-01 2.74911761e-01 1.04637527e+00
-6.90931678e-01 -1.64747536e+00 -1.93027556e-01 6.99704885e-01
-4.50885355e-01 -4.32431132e-01 -6.86824381e-01 4.00850505e-01
-8.02240334e-03 6.43631756e-01 -4.70747322e-01 -3.29163194e-01
1.28594693e-03 1.71875820e-01 7.27655232e-01 -6.15835428e-01
-3.10744047e-01 -2.05078945e-01 3.52088630e-01 -2.69056737e-01
-1.40909538e-01 -3.47822458e-01 -8.21564496e-01 -2.16881990e-01
-9.20025885e-01 5.05115092e-01 9.14257407e-01 9.20193791e-01
5.43606937e-01 1.48306817e-01 4.87412035e-01 -1.13964486e+00
-4.55341071e-01 -4.65199709e-01 -4.56100076e-01 1.83832049e-01
3.49771380e-01 -7.20595777e-01 -3.89644988e-02 -7.31699541e-02] | [5.153909206390381, 5.498390197753906] |
cc0dd151-ccbc-41b0-9fff-7a42bc7bb1a2 | heat-a-highly-efficient-and-affordable | 2304.07334 | null | https://arxiv.org/abs/2304.07334v2 | https://arxiv.org/pdf/2304.07334v2.pdf | HEAT: A Highly Efficient and Affordable Training System for Collaborative Filtering Based Recommendation on CPUs | Collaborative filtering (CF) has been proven to be one of the most effective techniques for recommendation. Among all CF approaches, SimpleX is the state-of-the-art method that adopts a novel loss function and a proper number of negative samples. However, there is no work that optimizes SimpleX on multi-core CPUs, leading to limited performance. To this end, we perform an in-depth profiling and analysis of existing SimpleX implementations and identify their performance bottlenecks including (1) irregular memory accesses, (2) unnecessary memory copies, and (3) redundant computations. To address these issues, we propose an efficient CF training system (called HEAT) that fully enables the multi-level caching and multi-threading capabilities of modern CPUs. Specifically, the optimization of HEAT is threefold: (1) It tiles the embedding matrix to increase data locality and reduce cache misses (thus reduces read latency); (2) It optimizes stochastic gradient descent (SGD) with sampling by parallelizing vector products instead of matrix-matrix multiplications, in particular the similarity computation therein, to avoid memory copies for matrix data preparation; and (3) It aggressively reuses intermediate results from the forward phase in the backward phase to alleviate redundant computation. Evaluation on five widely used datasets with both x86- and ARM-architecture processors shows that HEAT achieves up to 45.2X speedup over existing CPU solution and 4.5X speedup and 7.9X cost reduction in Cloud over existing GPU solution with NVIDIA V100 GPU. | ['Dingwen Tao', 'Yuxiong He', 'Shuaiwen Leon Song', 'Xiaodong Yu', 'Jonathan Soifer', 'Jiannan Tian', 'Baixi Sun', 'Shaden Smith', 'Chengming Zhang'] | 2023-04-14 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [-2.78268337e-01 -8.39574993e-01 -1.28869817e-01 -2.29393423e-01
-6.37850285e-01 -5.61638534e-01 4.22080934e-01 2.34500274e-01
-4.73024964e-01 4.45620328e-01 3.31801087e-01 -8.68614733e-01
2.02068631e-02 -1.06932151e+00 -4.66356248e-01 -7.30565488e-01
-1.13376901e-01 2.44247362e-01 3.68836999e-01 -2.00887889e-01
4.90435600e-01 2.82388121e-01 -1.73022091e+00 3.85429442e-01
9.38972712e-01 1.22918677e+00 2.01328099e-02 5.39836586e-01
-3.99104446e-01 4.64994609e-01 -1.28245071e-01 -3.96180153e-01
4.61684942e-01 -5.44250421e-02 -4.41340983e-01 -2.94805229e-01
1.97180882e-01 -4.07113791e-01 -5.68108037e-02 9.06005681e-01
6.04825675e-01 2.29356483e-01 2.54928291e-01 -9.88586545e-01
-3.13077807e-01 3.10311168e-01 -1.01994610e+00 1.73458427e-01
8.73272642e-02 2.04870194e-01 9.09947157e-01 -1.13333166e+00
2.33788297e-01 9.05321658e-01 6.64744198e-01 1.99454534e-03
-1.02687335e+00 -7.20254004e-01 9.82709602e-02 2.35309321e-02
-1.43006670e+00 -1.54513568e-01 3.66532594e-01 -3.00647855e-01
1.21145427e+00 5.78519046e-01 8.89065444e-01 6.54291332e-01
2.94053465e-01 6.79374218e-01 1.19597936e+00 -3.79044920e-01
4.94855225e-01 2.07371622e-01 5.24321318e-01 6.81058466e-01
4.51409519e-01 2.22828746e-01 -5.79618156e-01 -6.85678422e-01
5.24860382e-01 2.35723808e-01 -4.32870835e-02 -1.26600593e-01
-1.03777766e+00 1.02905130e+00 2.12899208e-01 3.31559628e-02
-3.87766689e-01 4.13675047e-02 8.81403983e-01 2.52521485e-01
5.25794566e-01 7.15074688e-02 -7.29411423e-01 -2.86382824e-01
-1.13848531e+00 2.93990105e-01 1.13030005e+00 6.10742450e-01
7.41014600e-01 1.77389495e-02 -3.23682427e-01 7.50839353e-01
4.32378352e-01 4.60112333e-01 4.59001958e-01 -5.95369577e-01
5.24384201e-01 6.99242473e-01 -1.49390455e-02 -1.06102490e+00
-3.10269326e-01 -8.22970033e-01 -1.00943482e+00 2.20664799e-01
1.73033267e-01 -2.13872582e-01 -3.07076395e-01 1.25693285e+00
7.29318857e-01 5.03222942e-01 -2.59976149e-01 9.89212155e-01
5.72933018e-01 6.29489899e-01 -9.15802121e-02 -1.27701670e-01
1.45067513e+00 -1.40083051e+00 -2.79962450e-01 -6.72842488e-02
9.04612601e-01 -1.14546156e+00 1.24319696e+00 4.83778924e-01
-8.00942898e-01 -5.04586995e-01 -1.18125749e+00 -1.20623469e-01
-3.31486940e-01 4.42227423e-01 1.14291036e+00 1.02294493e+00
-9.12296593e-01 5.53972125e-01 -1.13189960e+00 -8.81183334e-03
2.59307802e-01 5.56927979e-01 8.92215129e-03 5.27257062e-02
-8.29077244e-01 2.28401139e-01 9.25227404e-02 -3.02265417e-02
-2.39893362e-01 -1.29107606e+00 -3.73784333e-01 2.19232723e-01
3.51401567e-01 -9.26751316e-01 7.70684242e-01 -9.54728842e-01
-1.76102841e+00 4.17465717e-01 -2.35273525e-01 -3.34401935e-01
3.73587430e-01 -6.06008649e-01 -2.40604281e-01 -3.60680342e-01
-3.01726222e-01 5.19461222e-02 7.31015861e-01 -7.31464922e-01
-7.48408437e-01 -5.90343177e-01 -1.83973163e-01 2.87097007e-01
-7.63131440e-01 8.13261196e-02 -7.01514542e-01 -7.65913665e-01
1.12520427e-01 -1.09687769e+00 -3.51568222e-01 -1.32937044e-01
-2.16960683e-01 -2.91108005e-02 5.51784754e-01 -5.10202289e-01
1.72488260e+00 -2.16051769e+00 -1.79822445e-01 5.35042465e-01
1.93869427e-01 4.91521627e-01 3.00149798e-01 4.01820511e-01
2.99214363e-01 -3.21689814e-01 1.19413793e-01 -3.53288740e-01
-8.03406630e-03 1.05251469e-01 -4.14694399e-01 5.51980317e-01
-4.16077852e-01 5.67346811e-01 -6.64339542e-01 -1.98406249e-01
2.34521240e-01 6.10653102e-01 -1.04784775e+00 6.49907291e-02
6.08318076e-02 1.06995083e-01 -5.34995615e-01 4.79374200e-01
1.03269327e+00 -3.53986621e-01 2.92938441e-01 -4.72150952e-01
-4.85917896e-01 5.16212046e-01 -1.73146188e+00 1.38476586e+00
-6.44246280e-01 5.26359789e-02 2.15089813e-01 -7.63677418e-01
6.62283719e-01 -5.98867349e-02 3.92789155e-01 -5.50282955e-01
1.52589217e-01 3.29077780e-01 -2.89605558e-01 -8.86663273e-02
6.16057873e-01 3.84243518e-01 2.49137431e-01 8.62844229e-01
-4.62633967e-01 3.61386895e-01 1.48569211e-01 1.77051947e-01
1.12209082e+00 -1.89738329e-02 1.22225679e-01 -5.75405955e-01
7.69602239e-01 1.05780557e-01 5.03421605e-01 6.21063650e-01
1.09110750e-01 8.73965546e-02 1.58327430e-01 -6.80102766e-01
-8.35887492e-01 -9.79503632e-01 3.32671814e-02 1.56715882e+00
2.82796305e-02 -9.51363444e-01 -5.22341073e-01 -3.55620116e-01
2.37890527e-01 5.97318769e-01 -2.11850837e-01 1.16611555e-01
-5.28888106e-01 -1.20784688e+00 2.00967461e-01 5.47932684e-01
6.71868205e-01 -4.37036186e-01 -6.87816501e-01 3.33232462e-01
4.17600214e-01 -5.38895071e-01 -6.53125346e-01 5.68112405e-03
-1.11933494e+00 -8.45341444e-01 -6.20806292e-02 -4.11025733e-01
5.17655671e-01 6.98019505e-01 1.15556753e+00 3.32202971e-01
-2.13228479e-01 -1.12058580e-01 -3.16159159e-01 3.19069475e-02
1.79631054e-01 1.39218807e-01 1.66646570e-01 -8.43708813e-02
5.50098598e-01 -5.97817183e-01 -1.01345611e+00 1.93646267e-01
-6.64840221e-01 2.20191166e-01 4.34610635e-01 1.11912107e+00
7.98400700e-01 9.10596736e-03 -1.59334093e-02 -1.14415598e+00
7.13657200e-01 -4.91385788e-01 -9.30889308e-01 1.76820993e-01
-1.08742368e+00 -3.95667255e-02 7.95049787e-01 -2.61068970e-01
-1.18721199e+00 5.43392003e-02 -3.62895310e-01 -9.11447108e-02
3.74649584e-01 4.58436549e-01 3.86886328e-01 -2.01369002e-01
5.22851884e-01 2.65759766e-01 -6.28593639e-02 -6.70767486e-01
3.77015054e-01 8.27994287e-01 1.67674258e-01 -9.16826427e-01
4.98187542e-01 5.75776935e-01 -3.66592146e-02 -6.62960112e-01
-7.36866236e-01 -5.95948875e-01 -2.10487857e-01 1.85496256e-01
5.14214993e-01 -1.09787560e+00 -1.07799494e+00 2.35308841e-01
-6.07187867e-01 -4.25846502e-02 1.36656404e-01 6.89319253e-01
-3.59830633e-02 4.56863761e-01 -9.04058218e-01 -7.12796211e-01
-1.21011162e+00 -1.24580753e+00 8.35700035e-01 2.04107821e-01
-1.80433840e-01 -7.22616792e-01 8.88563246e-02 3.70403737e-01
8.31470490e-01 -1.66587904e-01 8.43887925e-01 -4.27454978e-01
-5.71832836e-01 -2.62227058e-02 -4.72801268e-01 3.57460916e-01
-2.50847667e-01 2.09559977e-01 -5.66350341e-01 -7.02069163e-01
5.89246824e-02 3.37260626e-02 5.14006555e-01 1.43883005e-01
1.12633920e+00 -1.97497860e-01 -4.01394576e-01 9.90626872e-01
1.66003728e+00 -8.57179984e-02 3.75487894e-01 4.20860440e-01
6.92391694e-01 2.31002063e-01 7.02669203e-01 7.85958648e-01
3.30101997e-01 6.26251638e-01 1.30370781e-01 -8.60271305e-02
1.08207084e-01 8.93094460e-04 4.61638808e-01 1.28433442e+00
4.11679149e-02 2.09487230e-01 -7.88583696e-01 1.15323246e-01
-1.97176182e+00 -6.02987170e-01 -6.22899473e-01 2.52467632e+00
6.80310965e-01 1.23003632e-01 7.77069330e-02 1.66234002e-01
2.76069313e-01 -8.18999112e-02 -4.49847162e-01 -6.80995226e-01
2.42444754e-01 6.58334792e-01 6.94534123e-01 4.14201498e-01
-1.09037793e+00 7.91523159e-01 5.44772816e+00 1.05134594e+00
-1.23000276e+00 4.66742188e-01 5.41672528e-01 -5.47289729e-01
-1.07017100e-01 1.48137957e-01 -1.14854872e+00 6.22379541e-01
8.83294880e-01 -3.36847752e-02 4.69548881e-01 1.23258054e+00
2.49602217e-02 -1.76562995e-01 -8.08900297e-01 1.10546386e+00
-1.49733767e-01 -1.47748780e+00 -2.90313158e-02 1.43868759e-01
8.30040932e-01 2.93105006e-01 -2.88921595e-02 3.75404149e-01
3.36465895e-01 -6.35278404e-01 4.25008774e-01 1.28914997e-01
3.34470302e-01 -9.83242393e-01 8.43585014e-01 2.81129271e-01
-1.30600500e+00 -5.14079519e-02 -3.71218354e-01 -2.76552111e-01
1.30577991e-02 1.17418289e+00 -1.72829688e-01 4.57328260e-01
9.66920137e-01 4.24938768e-01 -4.54212993e-01 8.65107477e-01
6.97895065e-02 9.21537519e-01 -7.04959214e-01 -2.21340925e-01
3.10595959e-01 -8.01952899e-01 2.30872318e-01 1.35398519e+00
3.97234201e-01 -4.91038561e-02 3.20011646e-01 3.71941894e-01
1.71059325e-01 6.27793849e-01 5.19621894e-02 3.52101088e-01
5.74532509e-01 1.34673643e+00 -6.71039581e-01 -4.56294894e-01
-9.12795544e-01 9.80614364e-01 4.18623954e-01 4.28135544e-02
-8.35143566e-01 -4.14650202e-01 1.10602462e+00 -2.06082407e-03
4.07584786e-01 -4.76652235e-01 -7.23412395e-01 -1.10678411e+00
3.71224880e-02 -8.52107286e-01 4.13182110e-01 -1.26900196e-01
-1.25776756e+00 4.59896743e-01 -5.00115275e-01 -9.35407877e-01
9.27298367e-02 -4.70933378e-01 -4.04087037e-01 1.19353533e+00
-1.35581505e+00 -7.99581051e-01 -3.58846635e-01 6.18640125e-01
2.88778156e-01 -8.59520361e-02 9.23271716e-01 6.06956422e-01
-7.54642308e-01 8.11308444e-01 6.05220914e-01 -3.79985422e-01
6.67078495e-01 -9.76290226e-01 4.93020386e-01 7.17236161e-01
6.85406290e-03 1.25253177e+00 4.94565815e-01 -6.29738092e-01
-2.17044520e+00 -1.01583946e+00 7.44018674e-01 -1.35856509e-01
7.75509715e-01 -4.34704483e-01 -8.84616375e-01 3.15879434e-01
-1.12906545e-01 1.78154737e-01 1.05134487e+00 3.24397564e-01
-3.89423400e-01 -3.39591682e-01 -9.69107330e-01 7.14911103e-01
8.26351225e-01 -3.33786309e-01 -4.23734076e-02 5.90140939e-01
2.85394132e-01 -6.53541148e-01 -1.00370431e+00 1.49634540e-01
8.03120255e-01 -1.24615169e+00 1.11738217e+00 -2.86301553e-01
2.02897564e-01 -4.46487039e-01 -2.80441195e-01 -7.86457598e-01
-4.79988724e-01 -5.64686596e-01 -3.42138410e-01 1.09108222e+00
1.37617007e-01 -7.76628673e-01 8.54886770e-01 5.83804309e-01
6.89900815e-02 -1.34725475e+00 -6.46415770e-01 -6.10950470e-01
-2.47184724e-01 -6.21291697e-01 6.65312409e-01 7.99812734e-01
-2.22824916e-01 3.78759891e-01 -2.86485821e-01 1.11085460e-01
5.42853296e-01 4.77820128e-01 1.01640272e+00 -8.73028755e-01
-7.45080888e-01 -3.20348173e-01 1.15241833e-01 -1.25604677e+00
-2.80565411e-01 -8.80807400e-01 -5.80016613e-01 -1.15583062e+00
1.61716640e-01 -6.99552596e-01 -8.01026449e-02 2.71579862e-01
-3.21637064e-01 3.24013919e-01 -3.13985869e-02 2.75702685e-01
-3.70047450e-01 3.58086556e-01 7.79418588e-01 3.74122262e-01
-3.63958240e-01 -5.93564427e-03 -5.95978856e-01 5.94907522e-01
5.95958889e-01 -3.45353395e-01 -4.79632378e-01 -5.44898152e-01
5.10170341e-01 -2.16841817e-01 2.43463293e-02 -7.88666844e-01
3.60097468e-01 9.42574516e-02 3.86496097e-01 -6.69827759e-01
1.12059869e-01 -6.63311005e-01 3.38497937e-01 6.76045954e-01
1.02634050e-01 3.64473164e-01 1.31330624e-01 3.98767084e-01
-6.32261857e-02 -9.44391191e-02 6.49880469e-01 1.47677153e-01
-3.93559009e-01 3.19886982e-01 -2.10125506e-01 -2.21863240e-01
9.74270940e-01 6.07109927e-02 -1.17680974e-01 2.18877956e-01
-1.69989154e-01 1.36216909e-01 4.17771667e-01 1.55551955e-01
2.25749791e-01 -1.28178787e+00 -5.85655570e-01 4.49007779e-01
-2.76764393e-01 -3.91869694e-01 3.35775107e-01 9.70369577e-01
-9.11337972e-01 3.69907737e-01 1.66284695e-01 -5.03592491e-01
-1.54103291e+00 6.06934667e-01 -1.81585476e-01 -5.36490679e-01
-7.93314934e-01 1.11279881e+00 -8.43910575e-02 -2.41730496e-01
9.16310251e-02 -7.61180967e-02 1.22381754e-01 -6.65423721e-02
8.66334975e-01 8.41363549e-01 5.05402863e-01 -2.36005381e-01
-4.28794861e-01 5.38323998e-01 -3.38828236e-01 1.91141561e-01
1.27471018e+00 -4.30069976e-02 -4.89689589e-01 2.91859377e-02
1.35875678e+00 3.10641527e-01 -8.75302315e-01 -2.57156432e-01
-1.55804396e-01 -7.11441100e-01 5.80396056e-01 -5.28824747e-01
-1.20416069e+00 6.33618534e-01 8.35996568e-01 -1.49156183e-01
1.20386171e+00 -7.01764584e-01 1.33424354e+00 1.99327305e-01
5.12738466e-01 -1.25560045e+00 -4.22417223e-01 5.57131052e-01
3.48921150e-01 -1.07271910e+00 5.39018989e-01 -6.18348956e-01
-3.82425517e-01 1.06521988e+00 3.74326736e-01 -3.94273967e-01
1.07953835e+00 5.57387531e-01 -1.58505231e-01 -1.22919463e-01
-8.61262381e-01 4.44801450e-02 2.66766697e-01 -4.63500991e-02
7.17926264e-01 2.81890929e-01 -9.68165636e-01 8.01309645e-01
-2.77487904e-01 1.99286252e-01 -1.59226596e-01 1.12559569e+00
-6.64898381e-02 -1.28598893e+00 -3.41362745e-01 8.78934026e-01
-4.46598858e-01 -4.20337349e-01 1.59142822e-01 4.82065260e-01
5.90167120e-02 1.00001252e+00 -6.49886355e-02 -5.69262087e-01
2.39721090e-01 -1.99737489e-01 3.68946999e-01 -5.36785722e-01
-1.35386992e+00 2.60777469e-03 1.23947533e-02 -8.56583357e-01
9.67382565e-02 -6.55047834e-01 -1.14869714e+00 -7.13156223e-01
-3.06028903e-01 2.71846980e-01 1.02567005e+00 7.35035241e-01
8.25128496e-01 4.04850930e-01 6.52394831e-01 -8.25625420e-01
-9.56919909e-01 -6.05035603e-01 -4.90408659e-01 1.91136435e-01
-3.55384201e-01 -6.29701793e-01 -3.73410940e-01 -4.19191718e-01] | [8.432901382446289, 3.4005234241485596] |
5cf20bf9-519e-460f-8e29-ce2c55725a23 | semantic-aware-contrastive-learning-for-more | 2301.07919 | null | https://arxiv.org/abs/2301.07919v1 | https://arxiv.org/pdf/2301.07919v1.pdf | Semantic-aware Contrastive Learning for More Accurate Semantic Parsing | Since the meaning representations are detailed and accurate annotations which express fine-grained sequence-level semtantics, it is usually hard to train discriminative semantic parsers via Maximum Likelihood Estimation (MLE) in an autoregressive fashion. In this paper, we propose a semantic-aware contrastive learning algorithm, which can learn to distinguish fine-grained meaning representations and take the overall sequence-level semantic into consideration. Specifically, a multi-level online sampling algorithm is proposed to sample confusing and diverse instances. Three semantic-aware similarity functions are designed to accurately measure the distance between meaning representations as a whole. And a ranked contrastive loss is proposed to pull the representations of the semantic-identical instances together and push negative instances away. Experiments on two standard datasets show that our approach achieves significant improvements over MLE baselines and gets state-of-the-art performances by simply applying semantic-aware contrastive learning on a vanilla Seq2Seq model. | ['Le Sun', 'Xianpei Han', 'Bo Chen', 'Chunlei Xin', 'Shan Wu'] | 2023-01-19 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 6.70813799e-01 1.10642791e-01 -2.01614574e-01 -9.51043189e-01
-1.46613967e+00 -6.80550933e-01 4.39679176e-01 1.10212699e-01
-5.55762291e-01 5.86910963e-01 4.29334342e-01 -9.85159501e-02
3.79689671e-02 -5.27033985e-01 -8.85805428e-01 -7.30899632e-01
1.43001169e-01 6.13652349e-01 -1.46089755e-02 -7.75263309e-02
3.53873432e-01 2.22721193e-02 -1.44372904e+00 6.61039054e-01
1.08743286e+00 8.69865835e-01 6.76647842e-01 6.15811884e-01
-7.03214526e-01 7.43869483e-01 -6.46118343e-01 -3.29597920e-01
-5.06972671e-02 -7.39522696e-01 -9.31770265e-01 -1.39592156e-01
8.21471512e-01 1.79517478e-01 2.31684316e-02 1.32986188e+00
3.14122677e-01 2.61069536e-01 5.74515522e-01 -8.14031839e-01
-7.74565458e-01 5.84922612e-01 -4.09026831e-01 2.36181632e-01
2.05695570e-01 2.24384531e-01 1.49613023e+00 -1.08106983e+00
4.90456372e-01 1.69752395e+00 3.11591744e-01 7.44896710e-01
-1.38114870e+00 -6.87401116e-01 7.60930657e-01 2.15973094e-01
-1.07340324e+00 -2.25639731e-01 6.66352212e-01 -1.37890100e-01
1.19487250e+00 2.76374042e-01 2.51481056e-01 1.40230644e+00
5.56690618e-02 1.18068945e+00 9.28512096e-01 -1.50365397e-01
5.99940956e-01 -3.82949054e-01 3.21353436e-01 5.95186949e-01
9.56975967e-02 -1.52906492e-01 -6.25180185e-01 -2.39853218e-01
4.34771508e-01 5.81432581e-02 -3.68319117e-02 -4.43424970e-01
-9.67752695e-01 9.92022157e-01 4.24480200e-01 2.15118662e-01
-4.49335009e-01 3.27581525e-01 5.75195074e-01 3.54304075e-01
6.75961852e-01 5.28010964e-01 -9.36077714e-01 -3.11301708e-01
-7.31445849e-01 3.55469286e-01 5.33326685e-01 9.90934551e-01
8.90506864e-01 -1.16890468e-01 -5.07973909e-01 1.09273803e+00
8.15724358e-02 3.47738057e-01 6.40841782e-01 -8.94533932e-01
5.94456017e-01 4.43835348e-01 -2.53762484e-01 -5.00519276e-01
-1.92131549e-02 -3.54308099e-01 -5.70338786e-01 -2.01712936e-01
2.80614018e-01 -4.14610170e-02 -1.03293884e+00 2.10750771e+00
9.56686139e-02 5.44762969e-01 1.38460815e-01 9.98934269e-01
5.87962508e-01 6.03335738e-01 7.52655208e-01 8.57784823e-02
1.46616423e+00 -9.83423769e-01 -5.79913676e-01 -7.75036275e-01
7.42048442e-01 -3.70829314e-01 1.55363321e+00 1.35593593e-01
-9.50341165e-01 -4.25883979e-01 -9.13809717e-01 -3.35519999e-01
-2.27730855e-01 -1.19002901e-01 6.74990416e-01 2.88572162e-01
-6.60818756e-01 7.47101605e-01 -8.54256630e-01 -1.46931455e-01
6.89008951e-01 -1.40707955e-01 -2.08949164e-01 -4.74643737e-01
-1.26761043e+00 6.21482670e-01 6.41552866e-01 -7.12908581e-02
-9.57411408e-01 -9.42285895e-01 -1.17015934e+00 3.26803476e-01
3.85675400e-01 -8.22388351e-01 1.28005385e+00 -1.30114591e+00
-1.39129031e+00 9.91522253e-01 -5.23843765e-01 -3.72043461e-01
2.08254978e-01 -5.26248276e-01 -1.34916008e-01 9.13116634e-02
3.12406301e-01 8.15564871e-01 5.52382350e-01 -9.54834878e-01
-5.12672246e-01 -4.35766250e-01 3.83427255e-02 3.59745741e-01
3.81975621e-02 -2.90334076e-02 -2.72913516e-01 -7.96541989e-01
3.10002696e-02 -6.82731867e-01 -4.33167845e-01 -4.89224494e-02
-2.76712447e-01 -5.10904133e-01 2.94856548e-01 -6.51534915e-01
9.38401222e-01 -2.32787919e+00 2.90601403e-01 -2.15609044e-01
-5.63044809e-02 2.04524249e-01 -7.23828197e-01 1.98911861e-01
-6.80058263e-04 -1.40134254e-02 -6.87688768e-01 -1.85804591e-01
6.87201694e-02 5.30070603e-01 -2.35527635e-01 1.20465443e-01
8.09757113e-01 1.01933968e+00 -1.50729430e+00 -2.35495970e-01
-9.07155573e-02 3.44207257e-01 -7.44829774e-01 4.00669813e-01
-6.32297158e-01 2.95175165e-01 -7.31354296e-01 3.78683567e-01
6.13712192e-01 -3.80677998e-01 4.70894814e-01 1.42616436e-01
4.32955861e-01 8.60211611e-01 -6.62591517e-01 2.38702059e+00
-6.77911639e-01 2.53445387e-01 -1.59127459e-01 -1.34987712e+00
9.44218099e-01 -6.60280138e-02 -1.59420520e-01 -7.68512189e-01
-1.74623191e-01 2.09891960e-01 -1.98826596e-01 -1.59205586e-01
5.29641330e-01 -4.45200562e-01 -5.67708015e-01 -7.05315406e-03
2.01797605e-01 -6.23049997e-02 -1.99507430e-01 1.78795546e-01
1.26913464e+00 4.67202097e-01 3.90448660e-01 -5.07586241e-01
3.77321094e-01 7.79876532e-03 9.68342662e-01 8.38755667e-01
-5.46732545e-02 6.62308514e-01 6.29489720e-01 -4.59580183e-01
-7.74877906e-01 -1.32821476e+00 2.33972684e-01 1.42935872e+00
2.86085337e-01 -3.63586307e-01 -7.60732889e-01 -1.30364037e+00
-4.22794791e-03 1.18528092e+00 -5.68226993e-01 -4.78682697e-01
-6.12606525e-01 -5.37126303e-01 3.50279897e-01 7.29219913e-01
2.15831459e-01 -1.08972645e+00 -3.77880216e-01 3.50666016e-01
-2.02556372e-01 -9.74034131e-01 -5.20697236e-01 4.53186631e-01
-9.60683823e-01 -8.60777557e-01 -6.02640569e-01 -7.45682538e-01
5.13188958e-01 2.47327700e-01 1.67587912e+00 -2.80820757e-01
-1.32714435e-01 -3.25072467e-01 -5.08071721e-01 -1.48571447e-01
-2.47526288e-01 8.87841880e-02 -4.89854544e-01 -2.22026512e-01
7.63207376e-01 -4.66702491e-01 -5.69813967e-01 -9.95052978e-02
-7.25845814e-01 -9.03424844e-02 8.87498200e-01 1.15587723e+00
1.05434930e+00 -4.12160695e-01 7.38826334e-01 -1.33406568e+00
4.46642995e-01 -6.60062253e-01 -4.18915331e-01 4.24239486e-01
-3.71095479e-01 7.01562524e-01 9.16688263e-01 -2.71901846e-01
-1.23347235e+00 7.13496730e-02 -2.21715480e-01 -4.92430627e-01
-3.32253039e-01 2.47166783e-01 -5.25204182e-01 6.59712553e-01
3.84151578e-01 4.28836733e-01 -2.87968636e-01 -6.92123115e-01
7.94863999e-01 5.30195415e-01 4.84306335e-01 -8.30376089e-01
1.71649933e-01 1.59228981e-01 -3.70888889e-01 -5.34536481e-01
-1.60329103e+00 -4.66392040e-01 -3.38724554e-01 5.07378876e-01
9.88376915e-01 -1.03157103e+00 -1.66375816e-01 2.02995911e-01
-1.16218352e+00 -3.53585064e-01 -3.82813364e-01 3.43406230e-01
-6.54030502e-01 2.23798513e-01 -5.13271809e-01 -5.48505068e-01
-3.15789759e-01 -1.02666092e+00 1.50603271e+00 8.09320658e-02
-3.52278471e-01 -7.66017139e-01 9.57799107e-02 2.56145686e-01
2.05524504e-01 -5.86486906e-02 1.10937178e+00 -9.63978112e-01
-5.63793242e-01 1.92166463e-01 -3.21715713e-01 4.24183547e-01
9.15770605e-02 -6.57356322e-01 -1.06881142e+00 -1.17925838e-01
-3.00332963e-01 -5.69365263e-01 1.30714476e+00 3.12022269e-01
1.66159809e+00 -2.83655375e-01 -2.69702196e-01 5.88103712e-01
1.41112161e+00 -1.36646643e-01 6.12564087e-01 1.00247171e-02
7.54136682e-01 5.46079099e-01 9.81785893e-01 1.45482242e-01
2.99991637e-01 4.64964867e-01 3.25643450e-01 2.19305471e-01
-6.30556867e-02 -6.45647764e-01 3.74814421e-01 8.23856175e-01
5.76026261e-01 -2.25850955e-01 -4.27266777e-01 6.15685463e-01
-1.76422763e+00 -1.04914391e+00 1.71695203e-01 2.10006881e+00
1.01159263e+00 8.26450586e-02 -2.02559993e-01 -5.29545009e-01
7.56569326e-01 4.68442708e-01 -9.10592616e-01 -6.24292016e-01
-1.41843900e-01 5.32547891e-01 3.66363138e-01 3.32004666e-01
-1.08290601e+00 1.43098092e+00 6.04510546e+00 1.19508004e+00
-8.42376053e-01 2.94873476e-01 7.15864658e-01 -1.81989774e-01
-8.42168510e-01 2.98483372e-02 -6.45889938e-01 5.17084420e-01
1.04306722e+00 1.56792939e-01 2.44722933e-01 9.95333850e-01
-4.39199470e-02 1.44018188e-01 -1.09517574e+00 8.51089954e-01
-3.91944796e-02 -1.13597989e+00 2.87507921e-01 -2.18723372e-01
5.91681004e-01 2.34999627e-01 -2.27637202e-01 6.23148024e-01
7.23539233e-01 -9.87443268e-01 6.01486862e-01 5.52352130e-01
6.98985457e-01 -8.55261266e-01 6.24131978e-01 2.36081779e-01
-1.12283194e+00 5.86824156e-02 -6.94958687e-01 3.19117755e-02
1.33702636e-01 6.62331343e-01 -5.68797231e-01 5.47095597e-01
5.23670018e-01 8.40565503e-01 -1.96458802e-01 5.11223495e-01
-6.85774326e-01 8.17670345e-01 5.06098345e-02 -3.71784419e-01
5.27392745e-01 -3.00427172e-02 3.28156352e-01 1.59785199e+00
1.35978132e-01 -2.75038388e-02 2.81653166e-01 1.11746430e+00
-1.83366671e-01 4.13350947e-02 -2.21018597e-01 -2.32937247e-01
6.79697752e-01 1.17156339e+00 -3.79828334e-01 -6.27597451e-01
-5.06575644e-01 1.47177029e+00 8.41019690e-01 1.83722794e-01
-5.41449070e-01 -4.29695487e-01 1.35513961e+00 -4.39822882e-01
5.58965921e-01 1.01856291e-01 -2.89979041e-01 -1.30504274e+00
2.09794641e-02 -7.72343993e-01 6.57527149e-01 -5.65806568e-01
-1.76780438e+00 4.29715186e-01 -2.73654312e-01 -8.87928128e-01
-3.92767906e-01 -5.96810997e-01 -5.53032875e-01 1.02677488e+00
-1.71409154e+00 -8.97355378e-01 3.52987982e-02 3.35806608e-02
1.13022614e+00 1.61760058e-02 9.86057580e-01 -7.11256415e-02
-2.91952968e-01 8.09136748e-01 7.18755648e-02 -3.62896430e-03
4.58065689e-01 -1.51771891e+00 7.55636334e-01 7.46835470e-01
2.01354250e-01 6.22182786e-01 5.55647492e-01 -5.30255437e-01
-1.11695123e+00 -1.25080287e+00 1.12205005e+00 -3.00929993e-01
4.34527844e-01 -6.76736832e-01 -1.29738593e+00 6.35268688e-01
-1.23840630e-01 1.79251745e-01 6.64141059e-01 4.22176689e-01
-9.69565332e-01 1.69160619e-01 -1.03520060e+00 4.41837579e-01
1.47425199e+00 -8.09971690e-01 -1.05737936e+00 1.68323398e-01
1.06664407e+00 -6.14132546e-02 -5.06183207e-01 2.01265782e-01
4.66143936e-01 -7.43807435e-01 8.93046439e-01 -1.06956959e+00
5.70648551e-01 -3.97941098e-02 -3.90102923e-01 -1.52462316e+00
-3.61095279e-01 -2.74063766e-01 7.26373345e-02 1.15093064e+00
4.47989464e-01 -4.45704281e-01 7.55521297e-01 2.13233158e-01
-2.62318373e-01 -7.63224363e-01 -7.70953774e-01 -9.83741999e-01
3.80075753e-01 -3.70645493e-01 6.54482841e-01 9.60718870e-01
-5.93345873e-02 7.87042022e-01 -8.21136385e-02 1.36299580e-01
6.47634625e-01 2.58702278e-01 1.83969885e-01 -1.12012768e+00
-5.43579042e-01 -4.18507755e-01 -4.84381288e-01 -1.51594186e+00
8.37162137e-01 -1.16025805e+00 4.27355647e-01 -1.38663673e+00
5.93037426e-01 -2.32151344e-01 -7.69660056e-01 4.12137598e-01
-9.16448593e-01 -2.75370300e-01 -3.09216436e-02 -8.01830217e-02
-9.75755930e-01 8.31922770e-01 9.90812302e-01 -1.64736360e-01
2.01304451e-01 -3.47948045e-01 -9.43550527e-01 6.52276278e-01
7.28989720e-01 -7.36180782e-01 -6.40696049e-01 -6.60010278e-01
1.53334722e-01 -1.15563296e-01 2.28312403e-01 -4.98059571e-01
-2.38898858e-01 -2.00267434e-01 2.12147549e-01 -5.04846811e-01
2.93541431e-01 -3.09081942e-01 -3.41114670e-01 2.63109982e-01
-9.30149496e-01 -2.51486540e-01 -6.06322102e-02 7.88298011e-01
-3.36267203e-01 -2.05501676e-01 7.61124849e-01 -4.45893943e-01
-1.09569788e+00 2.08563790e-01 -1.36378363e-01 7.51196921e-01
5.29992700e-01 1.52044967e-01 -4.48209569e-02 -1.13029696e-01
-5.15039623e-01 2.17562273e-01 3.46120507e-01 6.09789491e-01
6.06121898e-01 -1.24750292e+00 -8.23313594e-01 1.29310474e-01
5.77702761e-01 1.42751798e-01 4.76160020e-01 1.33004338e-01
7.49610290e-02 1.76011786e-01 4.92781661e-02 -4.50616866e-01
-1.02308416e+00 4.68707770e-01 1.76676735e-01 -3.53741646e-01
-5.72796643e-01 1.30538392e+00 8.23620021e-01 -8.95803392e-01
-3.47960480e-02 -3.61005634e-01 9.45047140e-02 -2.46053636e-01
7.59069562e-01 8.70148279e-03 -5.51501289e-02 -3.31448913e-01
-4.74800706e-01 3.54143083e-01 -3.15199912e-01 2.30048895e-01
1.34039271e+00 -1.24029003e-01 1.62030950e-01 4.98364270e-01
1.58553159e+00 -3.54399145e-01 -1.70216167e+00 -3.47099692e-01
5.91508210e-01 -4.53357339e-01 -5.97867481e-02 -9.82743859e-01
-8.76467764e-01 1.08758974e+00 3.01374733e-01 -3.60326558e-01
1.04719067e+00 4.94242013e-01 9.63292956e-01 2.91555047e-01
1.25908121e-01 -1.02158535e+00 1.23217270e-01 8.26063335e-01
6.32160246e-01 -1.03091764e+00 -6.47344947e-01 -5.16838312e-01
-7.70616174e-01 6.79220676e-01 5.27267456e-01 -4.30878282e-01
7.08178505e-02 1.68283403e-01 2.23939624e-02 -1.50812104e-01
-1.12479818e+00 -3.23223650e-01 1.30852744e-01 5.54547012e-01
6.77552760e-01 2.74618089e-01 -4.10649359e-01 8.79570484e-01
2.40348410e-02 -2.58442461e-01 -3.24683040e-02 7.33300149e-01
-7.06557274e-01 -1.17737317e+00 2.63047010e-01 4.78283018e-01
-6.61933601e-01 -5.46234012e-01 -3.09072167e-01 2.89319247e-01
-1.66250721e-01 7.74297833e-01 3.73054177e-01 -9.97005254e-02
3.63820642e-01 4.49518800e-01 5.07843792e-01 -7.89070368e-01
-4.28370982e-01 -1.65470630e-01 3.46877843e-01 -9.77727354e-01
-1.11438006e-01 -7.61007488e-01 -1.53354132e+00 2.30704963e-01
1.06870398e-01 1.48385376e-01 5.48952401e-01 1.09544706e+00
6.77433431e-01 6.33263767e-01 4.88043875e-01 -5.19429207e-01
-1.07869041e+00 -9.24667656e-01 -5.77683806e-01 8.38799059e-01
1.53731480e-01 -4.30176616e-01 -3.18937153e-01 -2.16380596e-01] | [10.759471893310547, 8.790302276611328] |
607f45ca-8d99-4ccc-98df-d7e1a76797e1 | exarn-self-attending-rnn-for-target-speaker | 2212.01106 | null | https://arxiv.org/abs/2212.01106v2 | https://arxiv.org/pdf/2212.01106v2.pdf | ExARN: self-attending RNN for target speaker extraction | Target speaker extraction is to extract the target speaker, specified by enrollment utterance, in an environment with other competing speakers. Therefore, the task needs to solve two problems, speaker identification and separation, at the same time. In this paper, we combine self-attention and Recurrent Neural Networks (RNN). Further, we exploit various ways to combining different auxiliary information with mixed representations. Experimental results show that our proposed model achieves excellent performance on the task of target speaker extraction. | ['Xueliang Zhang', 'Shulin He', 'Pengjie Shen'] | 2022-12-02 | null | null | null | null | ['target-speaker-extraction', 'speaker-identification'] | ['audio', 'speech'] | [ 3.48832458e-01 7.79006854e-02 -1.39231116e-01 -5.29237747e-01
-1.10508955e+00 -1.82409957e-01 2.44452760e-01 -5.90105712e-01
-9.84296501e-02 5.79713523e-01 5.51923275e-01 -2.30998680e-01
3.01780552e-01 -7.67839700e-02 8.73718858e-02 -8.22579205e-01
1.47623032e-01 1.58064604e-01 -3.48340660e-01 -3.01031113e-01
8.49525556e-02 5.68559587e-01 -1.47667873e+00 2.22899780e-01
6.38030767e-01 8.30530345e-01 1.95919752e-01 8.35130692e-01
-2.51911342e-01 1.06489420e+00 -9.52094257e-01 7.88888708e-02
-1.67230114e-01 -6.67089462e-01 -7.30372250e-01 1.28517970e-01
1.20716326e-01 -2.65254736e-01 -2.95862585e-01 1.12678504e+00
8.73532236e-01 2.27000237e-01 5.39398491e-01 -1.01051807e+00
-6.22191668e-01 1.29517472e+00 -6.68630779e-01 6.53909504e-01
2.89016575e-01 -2.58846313e-01 8.11513364e-01 -1.07624495e+00
-1.99910119e-01 1.40328538e+00 3.62832516e-01 1.01529086e+00
-9.86748695e-01 -1.24987221e+00 4.81869608e-01 1.51325241e-02
-1.52777481e+00 -1.28835213e+00 1.12923849e+00 -1.22995652e-01
7.89312363e-01 4.85087603e-01 -1.98792666e-04 1.26032770e+00
-5.75626612e-01 1.08778298e+00 7.74007499e-01 -4.31456476e-01
-2.46129379e-01 5.73495090e-01 7.16586769e-01 1.21623911e-01
-2.36722648e-01 -5.95903909e-03 -5.99845767e-01 -6.36850595e-02
3.96781355e-01 -9.09650698e-02 -6.89804733e-01 4.92643356e-01
-8.78454506e-01 8.54623258e-01 -5.15189655e-02 6.34496748e-01
-2.38401130e-01 -3.69600624e-01 3.10877711e-01 3.17199439e-01
3.71950299e-01 -8.78051892e-02 -3.58389288e-01 1.26482606e-01
-9.27974284e-01 -2.82305807e-01 8.55500400e-01 1.08857238e+00
2.83894181e-01 7.52661228e-01 -2.19092607e-01 1.01343405e+00
6.46809638e-01 5.66783547e-01 1.21826589e+00 -1.22248828e-01
5.25683105e-01 2.70213395e-01 -3.16468365e-02 -5.83846152e-01
-2.29072466e-01 -7.48421192e-01 -1.02969790e+00 -5.66415973e-02
-1.16318427e-01 -5.74063241e-01 -1.03206015e+00 1.82255495e+00
2.59747267e-01 4.51415986e-01 8.15145314e-01 7.41652489e-01
1.59322929e+00 9.78520572e-01 1.93307661e-02 -6.28747702e-01
1.29580796e+00 -1.22130382e+00 -1.32078886e+00 -5.20975173e-01
5.21424562e-02 -8.42485487e-01 4.35245603e-01 7.67446458e-02
-1.04521096e+00 -7.54894435e-01 -1.06751966e+00 1.61424026e-01
-1.90217301e-01 4.22118425e-01 2.02762067e-01 8.65189552e-01
-8.46574128e-01 2.86848899e-02 -4.34033066e-01 4.65583317e-02
-1.19056128e-01 8.29281509e-01 -1.34042859e-01 6.20001435e-01
-1.45761144e+00 7.02089012e-01 2.44641736e-01 6.08437836e-01
-6.90034091e-01 -1.97686896e-01 -1.12249613e+00 2.81718761e-01
7.67146572e-02 -6.19577579e-02 1.45907259e+00 -1.19871008e+00
-1.96056747e+00 5.50503552e-01 -6.11427248e-01 -4.80242729e-01
6.88004419e-02 -1.89027712e-01 -1.02188575e+00 -3.15058976e-01
-5.90708060e-03 1.78156734e-01 1.13140225e+00 -1.02412164e+00
-7.46075153e-01 -6.86433315e-01 -6.58899844e-01 4.22279000e-01
-4.12498742e-01 6.56826079e-01 -3.52619529e-01 -5.72640598e-01
2.72390097e-01 -6.18003428e-01 -3.26465629e-02 -1.24840033e+00
-7.61343300e-01 -4.98371035e-01 1.20657480e+00 -1.01063454e+00
1.32560587e+00 -2.54984355e+00 2.68635541e-01 2.97115892e-01
1.85120106e-01 4.64172035e-01 -2.23632261e-01 -1.22772150e-01
-5.10152638e-01 1.87115073e-02 -4.93081696e-02 -6.86742485e-01
-6.49476051e-02 -3.34163547e-01 -4.93087530e-01 2.42269889e-01
2.12506220e-01 7.47400045e-01 -2.60834903e-01 -6.18249178e-01
-1.78336203e-02 7.27789700e-01 1.41047001e-01 5.29776871e-01
5.50881088e-01 4.46706623e-01 -4.35274035e-01 6.71286583e-01
6.99809015e-01 1.44860089e-01 1.52828917e-01 8.06617513e-02
-9.31041837e-02 5.90686619e-01 -1.43237388e+00 1.15138769e+00
-3.08065534e-01 8.11100066e-01 6.41639411e-01 -8.99744809e-01
1.08661139e+00 7.62030303e-01 1.32549837e-01 -2.67605275e-01
4.44094628e-01 1.80458412e-01 3.01623940e-01 -3.73512030e-01
3.70332032e-01 -1.54978886e-01 -3.75424102e-02 4.92249340e-01
1.43249169e-01 4.79245037e-01 -1.74378276e-01 -1.04285561e-01
5.42283833e-01 -4.26752657e-01 3.98725897e-01 -2.35449057e-02
9.78267193e-01 -7.22123086e-01 6.98119164e-01 6.32949352e-01
-4.69888836e-01 4.05634135e-01 2.19841301e-03 8.71798396e-03
-2.77221292e-01 -7.48013675e-01 2.62708757e-02 1.47275305e+00
-8.14342275e-02 -6.41468540e-02 -7.05794871e-01 -5.80690920e-01
-3.61141294e-01 6.33690119e-01 -4.97291893e-01 -6.19346760e-02
-1.01191318e+00 -6.38821065e-01 7.03917265e-01 5.49422204e-01
2.68725902e-01 -1.26697278e+00 -1.49896905e-01 4.53442663e-01
-3.27643245e-01 -9.25883114e-01 -9.47722912e-01 5.11940181e-01
-6.16787732e-01 -4.94989157e-01 -9.95342791e-01 -1.37654877e+00
4.35381442e-01 4.96404886e-01 8.68498325e-01 -6.90740645e-02
1.95186034e-01 -2.26073619e-02 -1.96436681e-02 -5.58422327e-01
-4.64310139e-01 4.14495260e-01 2.85897434e-01 3.38704199e-01
7.13014662e-01 -5.79523146e-01 8.67469981e-02 3.14451665e-01
-4.56511825e-01 -2.12068692e-01 7.58109868e-01 7.86854088e-01
8.96197855e-02 2.11399630e-01 1.07145882e+00 -5.18711388e-01
8.20144176e-01 -2.83009559e-01 -3.55487496e-01 3.53953272e-01
-6.70580789e-02 3.70508581e-02 3.91773492e-01 -8.12300205e-01
-1.24798965e+00 3.69868398e-01 -3.35725665e-01 -3.98862928e-01
-3.76885563e-01 4.81894344e-01 -7.71428764e-01 1.98581144e-01
2.28521779e-01 6.76227152e-01 9.17515233e-02 -6.07164025e-01
2.99742762e-02 1.51659763e+00 3.82576942e-01 7.56273791e-03
6.38944268e-01 -2.02639759e-01 -9.03510511e-01 -1.03354800e+00
-8.39894831e-01 -7.40589797e-01 -5.72512925e-01 7.54739121e-02
5.29118717e-01 -1.16881502e+00 -9.20847178e-01 4.97080714e-01
-1.31250572e+00 2.67199814e-01 6.77655041e-02 8.13180268e-01
-1.85411777e-02 1.57910720e-01 -4.96248007e-01 -1.57537186e+00
-7.44918168e-01 -1.19259906e+00 9.34508502e-01 5.89381635e-01
-1.89110115e-01 -5.97596347e-01 -1.55582204e-02 2.64463872e-01
4.21431303e-01 -3.76635939e-01 1.69455096e-01 -1.39119577e+00
-2.33192042e-01 -8.51358175e-02 -7.98286647e-02 4.32382584e-01
7.56830037e-01 -1.52985826e-01 -1.39111936e+00 -2.13205025e-01
6.88258946e-01 2.20352039e-02 9.24028635e-01 4.84253317e-01
7.20645905e-01 -5.41050375e-01 -3.16155612e-01 3.64632130e-01
8.06101024e-01 5.87887526e-01 4.25514191e-01 -1.94886968e-01
8.39123547e-01 8.92959177e-01 1.51846064e-02 -2.44408697e-02
1.65969864e-01 4.26717252e-01 -2.31736898e-01 -1.09365426e-01
1.77546591e-03 2.19564363e-01 7.58748233e-01 1.29630017e+00
2.17802241e-01 -1.85916141e-01 -8.20389450e-01 4.81583923e-01
-1.77551615e+00 -1.08870208e+00 1.51887015e-01 1.99705541e+00
8.76368225e-01 1.73094630e-01 2.77103394e-01 2.47968033e-01
1.42021525e+00 1.83626488e-01 -6.86711848e-01 -2.30520412e-01
-1.56572163e-01 -1.37729263e-02 1.38723865e-01 6.20891035e-01
-1.24887216e+00 8.56127024e-01 7.01005697e+00 7.52943814e-01
-1.43001032e+00 -1.63746122e-02 5.94693959e-01 -1.43374428e-01
5.26803434e-02 -5.53409517e-01 -1.46704757e+00 3.85700792e-01
1.39008200e+00 -3.95551682e-01 2.26359323e-01 7.85286367e-01
3.12234852e-02 6.05796933e-01 -1.20157290e+00 1.29695880e+00
6.80483162e-01 -6.75447404e-01 1.70459151e-02 -8.88017714e-02
1.56572327e-01 -1.89704970e-01 1.26528382e-01 8.36870253e-01
2.97693759e-01 -1.22035646e+00 4.73671287e-01 2.58698553e-01
2.47594520e-01 -8.84635687e-01 7.61033535e-01 5.09248614e-01
-1.48606348e+00 -1.25609413e-01 1.26868114e-01 1.66396454e-01
1.48253873e-01 1.00755662e-01 -8.73647988e-01 4.12386030e-01
3.58005196e-01 6.49666429e-01 -3.24172020e-01 6.21556044e-01
-9.80864540e-02 6.01397216e-01 -2.21242547e-01 -1.64070025e-01
5.99441379e-02 8.37621763e-02 7.42286503e-01 1.32604218e+00
1.91427469e-01 1.15825571e-01 1.54103234e-01 6.75510228e-01
-1.97372690e-01 3.76818538e-01 -5.34595132e-01 -4.86698411e-02
5.69267273e-01 1.17271769e+00 -3.00388575e-01 -4.58854169e-01
-2.19695330e-01 8.14018250e-01 3.98846358e-01 5.95650256e-01
-8.27610195e-01 -6.52004480e-01 5.17511427e-01 -3.86386871e-01
3.10695261e-01 1.46612659e-01 -5.15948720e-02 -1.15046155e+00
1.80560872e-02 -1.08949363e+00 3.55761170e-01 -3.93253565e-01
-1.19376278e+00 1.05243647e+00 -4.91566926e-01 -1.08229601e+00
-5.09619594e-01 -2.48709574e-01 -8.81994009e-01 1.36800134e+00
-1.56586087e+00 -1.09310615e+00 6.38225749e-02 6.96204185e-01
1.04313397e+00 -5.78773320e-01 8.68682683e-01 3.79565865e-01
-1.17055964e+00 8.20804358e-01 -1.60832722e-02 7.20648587e-01
5.04719734e-01 -1.10589576e+00 3.19378912e-01 8.70522320e-01
3.12013656e-01 8.11772048e-01 5.15395880e-01 -4.51295227e-01
-9.79039729e-01 -9.36786473e-01 9.21056271e-01 -1.46613970e-01
3.31754714e-01 -4.80427235e-01 -1.00960171e+00 8.91166031e-01
5.23509026e-01 -3.94668192e-01 1.07785451e+00 3.70182365e-01
-1.93750307e-01 -3.46764587e-02 -8.39135349e-01 4.83972579e-01
5.32672822e-01 -8.06994140e-01 -9.26917374e-01 -5.59876747e-02
1.03506076e+00 -3.49442393e-01 -4.30747956e-01 4.73887801e-01
4.50433522e-01 -5.26694775e-01 9.89288032e-01 -7.07696259e-01
-2.23908201e-01 -2.35433504e-01 -2.63123125e-01 -1.14440930e+00
-3.92984241e-01 -7.61771202e-01 -3.29764605e-01 1.82538188e+00
7.54786313e-01 -6.36368096e-01 6.55374706e-01 7.41886854e-01
3.93217020e-02 -2.58805990e-01 -9.38476086e-01 -6.89784288e-01
-3.26295465e-01 -5.58890663e-02 9.11569118e-01 1.03538382e+00
-6.93625286e-02 1.16497874e+00 -8.30187619e-01 4.46009785e-01
4.87284571e-01 1.25342473e-01 5.41013300e-01 -1.34174120e+00
-1.18933000e-01 -6.72334433e-01 -5.80212660e-02 -1.14081919e+00
7.79432595e-01 -4.38169122e-01 3.62354755e-01 -1.07876241e+00
-7.87581224e-03 8.90401974e-02 -8.81480396e-01 2.76773363e-01
-5.22033036e-01 -9.33804139e-02 -3.61275263e-02 2.09881231e-01
-4.30607796e-01 7.37695396e-01 5.49014747e-01 -6.09893382e-01
-7.58595765e-01 4.86500353e-01 -1.14069951e+00 7.39336669e-01
1.07221603e+00 -5.68313360e-01 -2.37441495e-01 -2.37382531e-01
-7.30969250e-01 2.41385117e-01 -2.31271908e-01 -7.66124964e-01
4.55247283e-01 4.24459875e-02 3.95580262e-01 -9.86419439e-01
5.11088490e-01 -7.46042430e-01 -9.03917775e-02 3.42369646e-01
-6.86157405e-01 -2.18893170e-01 2.48994291e-01 3.23671371e-01
-5.14616609e-01 -4.82228279e-01 6.64548695e-01 9.99876633e-02
-4.35921520e-01 4.07150201e-02 -5.83868086e-01 -2.92574286e-01
6.56826735e-01 -1.19674593e-01 1.13598131e-01 -4.97110754e-01
-1.03419733e+00 1.81722760e-01 -5.08727729e-01 6.38287842e-01
7.69102275e-01 -1.38190877e+00 -1.01845264e+00 4.55680132e-01
-1.81826890e-01 -2.68994123e-01 2.31341407e-01 5.61303079e-01
4.02074248e-01 3.30106735e-01 1.27444804e-01 -3.95710945e-01
-1.89615452e+00 5.21642983e-01 4.99814183e-01 -1.50463209e-01
-2.96302050e-01 9.89609718e-01 3.36597830e-01 -5.82117617e-01
7.94021070e-01 -8.46656784e-03 -1.03095591e+00 3.93035382e-01
1.11655164e+00 7.15594590e-02 -9.31547657e-02 -1.19583237e+00
-5.60459435e-01 2.65572727e-01 -5.63744962e-01 -9.93482247e-02
1.12351167e+00 -4.09879833e-01 -1.88054249e-01 6.93020344e-01
1.19592476e+00 4.42327172e-01 -5.24913847e-01 -5.55692911e-01
1.76440716e-01 -1.19372882e-01 2.13832498e-01 -3.86300564e-01
-1.12519157e+00 6.62634254e-01 7.41185665e-01 4.91991013e-01
1.01443601e+00 -1.07053339e-01 6.61149204e-01 4.28653955e-01
-4.47608918e-01 -1.03588593e+00 -2.77698725e-01 5.94511449e-01
8.98870289e-01 -1.55219662e+00 -3.43196630e-01 -3.27906847e-01
-5.58901072e-01 8.95545185e-01 9.13987637e-01 3.15858871e-01
8.94025147e-01 2.46518329e-01 5.40988624e-01 8.96970630e-02
-6.90436423e-01 -3.88700634e-01 4.63024884e-01 5.69556653e-01
7.81102955e-01 3.42987627e-02 4.16474968e-01 1.04769576e+00
-3.56378943e-01 -6.34570181e-01 3.62972409e-01 7.21027315e-01
-5.38449407e-01 -1.03337371e+00 -8.41093361e-01 2.49123871e-01
-7.62581766e-01 -3.45994622e-01 -6.26324832e-01 2.80979007e-01
-4.25341457e-01 1.33312976e+00 -8.88579562e-02 -2.47674227e-01
2.23712623e-01 2.38252565e-01 -1.31265804e-01 -7.67899513e-01
-7.90286720e-01 6.21592343e-01 3.79160382e-02 1.68812141e-01
-5.83343685e-01 -5.22401094e-01 -1.06833947e+00 1.59031704e-01
-8.36803973e-01 7.54412353e-01 7.93179810e-01 9.94160473e-01
1.95026055e-01 9.14530098e-01 1.04108846e+00 -8.05434525e-01
-5.43635905e-01 -1.45194340e+00 -7.59640872e-01 -8.57113004e-02
9.79705632e-01 -2.85563976e-01 -5.80716372e-01 6.30135322e-03] | [14.515252113342285, 6.055954456329346] |
190f9c39-c6a8-4515-b816-51e229cfedc9 | cross-attentive-pooling-for-speaker | 2008.05983 | null | https://arxiv.org/abs/2008.05983v2 | https://arxiv.org/pdf/2008.05983v2.pdf | Cross attentive pooling for speaker verification | The goal of this paper is text-independent speaker verification where utterances come from 'in the wild' videos and may contain irrelevant signal. While speaker verification is naturally a pair-wise problem, existing methods to produce the speaker embeddings are instance-wise. In this paper, we propose Cross Attentive Pooling (CAP) that utilizes the context information across the reference-query pair to generate utterance-level embeddings that contain the most discriminative information for the pair-wise matching problem. Experiments are performed on the VoxCeleb dataset in which our method outperforms comparable pooling strategies. | [] | 2020-08-13 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [ 2.29009479e-01 -1.20551199e-01 -9.42999497e-02 -1.02491295e+00
-1.50429356e+00 -5.21067262e-01 6.13934517e-01 -2.65385568e-01
-3.70317489e-01 3.58327419e-01 8.15645576e-01 1.34980917e-01
1.62747741e-01 1.25491107e-02 -5.54500759e-01 -8.37361813e-01
-1.40178502e-01 -6.78968057e-02 -3.69292237e-02 -2.23089918e-01
2.59196281e-01 2.08671585e-01 -1.30219829e+00 4.59905148e-01
3.20348978e-01 1.14885569e+00 -1.60533890e-01 8.89449894e-01
7.26045854e-03 6.58053100e-01 -8.35664511e-01 -4.76804793e-01
2.37244725e-01 -4.09955114e-01 -7.00483739e-01 -6.07829681e-03
9.98566866e-01 -2.47662827e-01 -4.87458855e-01 9.10175562e-01
9.96506393e-01 3.52719724e-01 4.23210502e-01 -1.65295732e+00
-6.98943436e-01 6.56614065e-01 -4.05912042e-01 4.91095364e-01
6.15239203e-01 -2.69297976e-02 1.23416805e+00 -1.37174094e+00
5.40907085e-01 1.44559479e+00 6.03953183e-01 9.44403410e-01
-9.68290269e-01 -8.38855684e-01 3.38433176e-01 5.52479267e-01
-1.70678926e+00 -1.31262124e+00 1.06458139e+00 -1.78783923e-01
8.53487492e-01 3.75860214e-01 4.65111509e-02 1.34193718e+00
-1.20810434e-01 1.07752275e+00 6.85661435e-01 -3.92107576e-01
5.27377287e-03 4.48258191e-01 4.24467862e-01 2.37701282e-01
-4.83270884e-01 1.85239255e-01 -1.23943830e+00 -8.70975256e-02
1.99481145e-01 -2.00837001e-01 -6.84089541e-01 -1.55791387e-01
-1.37584162e+00 7.12649882e-01 3.03403199e-01 4.05484259e-01
-2.01567098e-01 -1.42466258e-02 5.57014465e-01 6.17088199e-01
4.41797554e-01 1.44222289e-01 -3.41249943e-01 -6.16277568e-02
-1.13976061e+00 2.67079711e-01 7.82537639e-01 1.20165181e+00
4.16936368e-01 5.03306054e-02 -6.35582983e-01 9.00126278e-01
6.01270914e-01 2.56828338e-01 7.71636784e-01 -5.31560838e-01
6.69911861e-01 1.11321099e-01 1.35942206e-01 -6.92247510e-01
2.10161611e-01 -1.66479215e-01 -3.40940177e-01 -4.20001268e-01
1.50690362e-01 -1.86745390e-01 -8.94169033e-01 1.86241949e+00
2.88973570e-01 5.41815996e-01 4.37759697e-01 1.07047796e+00
1.38790238e+00 6.78192377e-01 1.42087176e-01 -2.26927139e-02
1.35458612e+00 -1.26393449e+00 -1.07526600e+00 -1.79917514e-01
1.76354289e-01 -8.70869040e-01 6.67176604e-01 -1.22828968e-01
-7.70780861e-01 -6.89346850e-01 -1.19207430e+00 -1.77379742e-01
-3.74019414e-01 -1.01889223e-01 -1.73541248e-01 8.12585652e-01
-1.24040461e+00 8.17804411e-02 -1.89618647e-01 -2.25093395e-01
2.79196054e-01 1.39593571e-01 -7.66487002e-01 -1.71748012e-01
-1.38263476e+00 9.77150083e-01 1.57853276e-01 4.50352639e-01
-1.18278205e+00 -7.95599699e-01 -1.15575767e+00 -8.81727319e-03
8.94782916e-02 -1.03697784e-01 1.49370003e+00 -1.03323519e+00
-1.45982313e+00 8.80015552e-01 -8.44629526e-01 -5.12720823e-01
3.87666464e-01 2.21772324e-02 -6.49573088e-01 1.32780716e-01
1.25769880e-02 9.99012709e-01 1.24755454e+00 -9.53311265e-01
-3.85156035e-01 -3.63087296e-01 -2.12632135e-01 2.44003057e-01
-2.08613768e-01 5.35269380e-01 -3.86917561e-01 -5.54350257e-01
2.11556852e-01 -4.41614598e-01 2.00949222e-01 -8.52127373e-02
-3.84263694e-01 -6.19115591e-01 1.22840428e+00 -1.05758178e+00
8.77442420e-01 -2.84979463e+00 7.19318539e-02 -1.12024531e-01
-2.37860620e-01 -3.42151038e-02 -4.83102918e-01 4.00560737e-01
-1.84369460e-01 1.09936997e-01 -1.27218291e-01 -9.36913311e-01
3.68207127e-01 -1.42331675e-01 -5.44242144e-01 6.61118865e-01
5.42451262e-01 8.73623550e-01 -5.56233466e-01 -1.00978482e+00
-3.47271785e-02 7.22016811e-01 -2.83724010e-01 6.39218926e-01
2.22830638e-01 3.21979344e-01 -1.15380108e-01 9.09176886e-01
6.07042134e-01 4.47223186e-01 -8.76690894e-02 -3.43363047e-01
7.17893392e-02 4.45257157e-01 -9.04850364e-01 1.90908790e+00
-2.84532368e-01 1.01238918e+00 3.67800683e-01 -8.38104904e-01
9.40205336e-01 8.54510248e-01 -6.65224493e-02 -4.79744762e-01
2.00343966e-01 2.22905166e-02 -8.94907564e-02 -7.30301440e-01
5.96664310e-01 -4.62275952e-01 -2.36716822e-01 1.42888531e-01
5.53511977e-01 -1.30904332e-01 -1.66117787e-01 1.65470973e-01
7.99692988e-01 7.26623274e-03 1.04301445e-01 -1.02715045e-01
7.78480947e-01 -5.25688887e-01 8.08312237e-01 6.07607126e-01
-1.00315273e+00 8.47026229e-01 1.81993946e-01 -1.18594199e-01
-6.22559071e-01 -1.06713533e+00 -2.45414838e-01 1.27728271e+00
5.02814241e-02 -2.27146879e-01 -5.98123372e-01 -7.50022709e-01
-5.79961464e-02 6.79148734e-01 -8.09635758e-01 1.03986710e-01
-4.83330756e-01 -1.08801268e-01 6.21489227e-01 5.76540351e-01
4.48315531e-01 -8.97518337e-01 -1.18740335e-01 2.43172362e-01
-4.01327431e-01 -1.26848161e+00 -1.17077923e+00 2.07823403e-02
-4.19222414e-01 -8.13694060e-01 -8.24482977e-01 -1.03627110e+00
4.98923421e-01 3.44298869e-01 8.44477355e-01 -3.02650720e-01
-2.89994746e-01 5.43949544e-01 -2.83930719e-01 -6.36189401e-01
-1.78358227e-01 -2.59419471e-01 6.73274323e-02 6.02318466e-01
8.45526576e-01 -1.30599886e-01 -3.04252684e-01 3.95936787e-01
-5.81221640e-01 -6.79180086e-01 2.88145781e-01 1.08909535e+00
4.73658860e-01 -3.41370493e-01 8.96184802e-01 -2.16211677e-01
7.53745258e-01 -3.20532084e-01 -2.00469866e-01 5.33452570e-01
1.42097414e-01 -4.62436676e-02 2.06712782e-01 -5.95668018e-01
-1.16705537e+00 1.13868155e-01 -3.25453170e-02 -8.09582233e-01
-3.06702256e-01 1.18726328e-01 -6.18493736e-01 -9.72962379e-02
3.87601823e-01 4.99245524e-01 -3.26693282e-02 -4.07202393e-01
2.05879912e-01 1.12920392e+00 4.17562306e-01 -2.50715703e-01
5.21688223e-01 1.41933993e-01 -7.35830486e-01 -1.18040574e+00
-6.38078630e-01 -9.42324162e-01 -4.98525918e-01 -4.01820898e-01
7.86028028e-01 -1.12668574e+00 -4.87505734e-01 1.68973386e-01
-1.33851933e+00 2.41856426e-01 -1.38970330e-01 5.24339795e-01
-2.55763084e-01 2.47870848e-01 -3.64150703e-01 -1.14789808e+00
-3.87860805e-01 -1.37278903e+00 1.26471353e+00 2.77832508e-01
-2.44207516e-01 -6.70463979e-01 2.04246640e-01 4.77393091e-01
5.22331595e-01 -3.39167893e-01 1.86285987e-01 -1.24486744e+00
-6.38704121e-01 -5.86649835e-01 -1.43508688e-01 5.08427322e-01
1.91566706e-01 -6.06751777e-02 -1.75497687e+00 -3.89215440e-01
2.44795486e-01 -2.82562047e-01 8.34126413e-01 2.47333467e-01
9.27894413e-01 -4.07010823e-01 -3.67402762e-01 2.75038540e-01
1.06633055e+00 1.10957868e-01 5.12012899e-01 -2.84315884e-01
3.43850791e-01 1.02517498e+00 4.61654067e-01 -7.71498531e-02
3.54840517e-01 9.20585096e-01 -5.41190105e-03 2.38962054e-01
-1.46405593e-01 -3.24985445e-01 7.19253659e-01 8.96105707e-01
5.95754981e-01 -1.88706502e-01 -5.25692999e-01 1.06928051e+00
-1.49233854e+00 -1.35726631e+00 3.55603158e-01 1.92323411e+00
7.42845476e-01 -3.01419646e-01 6.13989048e-02 1.45189688e-01
1.00635028e+00 3.43648285e-01 -2.83140361e-01 -4.28299487e-01
-1.14730231e-01 -6.19665831e-02 -1.12787858e-02 6.13076210e-01
-1.33869755e+00 6.85478866e-01 6.88008833e+00 5.72802842e-01
-1.32309973e+00 4.93709862e-01 2.64982402e-01 -3.73674959e-01
6.71415997e-04 -2.64887571e-01 -1.13050342e+00 4.48341370e-01
1.15678358e+00 -2.84965634e-01 -1.75038591e-01 8.92837465e-01
6.93113431e-02 2.41903886e-01 -1.51025283e+00 1.29466152e+00
8.34820807e-01 -9.46635187e-01 -2.19097570e-01 -3.32621008e-01
4.19226587e-01 -3.73527259e-02 1.79238588e-01 4.16142434e-01
-1.89376414e-01 -9.17694211e-01 8.40640306e-01 3.28773171e-01
5.53342581e-01 -5.27200878e-01 9.22143519e-01 -2.37488542e-02
-1.25186050e+00 -8.12178478e-03 -6.85241669e-02 4.98084009e-01
3.67310345e-01 8.45962241e-02 -1.14603710e+00 4.52286631e-01
8.49302709e-01 6.06050193e-01 -3.84039283e-01 1.20050240e+00
-2.71972418e-01 6.35230601e-01 -8.23372230e-02 -1.78401843e-01
3.61978374e-02 2.87562549e-01 7.49327660e-01 1.63895297e+00
2.89560854e-02 -4.59062159e-02 -1.90051392e-01 6.00981176e-01
-3.15330565e-01 2.76311159e-01 -8.21931005e-01 5.57964407e-02
5.92422605e-01 9.25821602e-01 -1.56337451e-02 -2.52613276e-01
-4.79744494e-01 9.50048506e-01 6.01944961e-02 5.17580211e-01
-7.06234574e-01 -7.18546331e-01 9.34602559e-01 -4.73815620e-01
6.16878033e-01 1.28609464e-01 1.47851974e-01 -1.09968424e+00
3.92200142e-01 -7.27578461e-01 3.00983429e-01 -5.23935914e-01
-1.63981342e+00 7.88621187e-01 -8.61642659e-02 -1.30345941e+00
-1.96361348e-01 -4.10198659e-01 -8.56740355e-01 1.00606143e+00
-1.75921941e+00 -1.20909202e+00 -1.09523490e-01 6.08235419e-01
9.30749655e-01 -2.10768789e-01 7.80281842e-01 3.97681832e-01
-6.02507889e-01 1.20439065e+00 -2.98455566e-01 5.80955923e-01
1.02534580e+00 -8.43638122e-01 2.85048693e-01 9.24971879e-01
4.13423538e-01 6.84790492e-01 8.17604303e-01 -1.58149540e-01
-1.36636496e+00 -9.28068340e-01 1.35458732e+00 -4.33752269e-01
5.24096310e-01 -7.17885077e-01 -1.06214893e+00 7.39591897e-01
5.88840842e-01 8.51068422e-02 1.05961394e+00 1.86638590e-02
-8.41986954e-01 -3.65030318e-01 -1.40027976e+00 2.77472317e-01
6.78690314e-01 -1.18934333e+00 -1.09390926e+00 7.98786432e-02
7.74762869e-01 -2.56876469e-01 -6.95695698e-01 1.89196199e-01
6.59850121e-01 -5.44556677e-01 9.15344834e-01 -7.09093332e-01
-1.40067056e-01 -3.03969413e-01 -6.23602211e-01 -1.18433642e+00
2.00915873e-01 -5.73413968e-01 2.00921297e-01 1.66330242e+00
6.57560110e-01 -5.20781994e-01 4.50700432e-01 6.24441028e-01
-2.17547163e-01 -1.57941446e-01 -1.67333555e+00 -8.81208837e-01
-1.71073273e-01 -4.05273259e-01 9.03096378e-01 9.84106779e-01
3.85430127e-01 2.85603583e-01 -3.51642370e-01 4.41483885e-01
6.78763568e-01 1.30173909e-02 6.14457667e-01 -7.59346783e-01
-3.45823504e-02 -3.19750965e-01 -7.75371373e-01 -1.05016077e+00
5.99394023e-01 -7.15701699e-01 7.94289172e-01 -9.43409026e-01
1.29349515e-01 4.47131954e-02 -4.94607121e-01 3.34182918e-01
-2.03467354e-01 1.15472794e-01 1.58140764e-01 -1.76453575e-01
-5.29145420e-01 7.52121389e-01 6.71509445e-01 -7.25677073e-01
-6.67976066e-02 -1.93827853e-01 -4.54453230e-01 6.44094050e-02
6.96488440e-01 -3.45496148e-01 -1.43845305e-01 -4.72549170e-01
-6.95227504e-01 1.93038970e-01 3.77483159e-01 -7.25246370e-01
4.40896869e-01 1.77242100e-01 7.82488585e-02 -7.15039790e-01
8.70265484e-01 -7.12278128e-01 -2.79462039e-01 -8.75550136e-02
-7.06763983e-01 -1.59145176e-01 3.61443311e-01 6.13109171e-01
-8.69953871e-01 -2.90272415e-01 7.21249104e-01 2.70692974e-01
-7.02332854e-01 2.65397310e-01 -8.90053585e-02 -2.36366484e-02
9.41934168e-01 -4.71012555e-02 -1.17996544e-01 -5.72954834e-01
-6.29874408e-01 3.27161074e-01 1.97655186e-02 8.30293179e-01
9.47916150e-01 -1.42626345e+00 -1.09178841e+00 2.53948629e-01
5.03045619e-01 -4.46953267e-01 5.44060111e-01 7.20773995e-01
2.65279621e-01 5.95713317e-01 8.13923627e-02 -6.79751813e-01
-1.69159341e+00 5.13483644e-01 7.01664686e-01 4.53818470e-01
-1.97209075e-01 1.52456677e+00 2.02680722e-01 -4.49890614e-01
6.87704802e-01 -6.05647415e-02 -1.45356238e-01 3.28592002e-01
1.01303899e+00 -9.55942273e-02 5.71347699e-02 -1.17549086e+00
-7.98000872e-01 2.40459859e-01 -3.70780885e-01 -4.63005871e-01
1.15476787e+00 -1.56239599e-01 1.39328673e-01 6.75239384e-01
1.56309938e+00 -8.19830447e-02 -1.08523655e+00 -3.77208680e-01
-2.46809563e-03 -6.78269386e-01 1.12043932e-01 -5.32157362e-01
-1.05370760e+00 1.20450878e+00 7.93664455e-01 -5.70802838e-02
8.79023850e-01 1.51559591e-01 4.39463258e-01 1.04863457e-01
1.23003557e-01 -1.06113029e+00 -1.08960882e-01 6.06690943e-02
1.36452687e+00 -1.51231968e+00 -3.18109542e-01 -2.43068755e-01
-6.59519017e-01 7.57416725e-01 6.59253597e-01 1.98175028e-01
7.35641956e-01 -2.52413284e-02 3.82357270e-01 5.36244139e-02
-8.43677044e-01 -3.84055860e-02 3.83124053e-01 7.47354269e-01
6.16120875e-01 2.72780675e-02 1.26088813e-01 6.35126412e-01
-3.12663205e-02 -2.55101293e-01 2.00007454e-01 1.00888324e+00
-7.26980269e-02 -1.19343889e+00 -5.80499291e-01 5.26945889e-02
-3.49011064e-01 -8.16768482e-02 -6.32496238e-01 4.51137930e-01
-2.91073203e-01 1.29919636e+00 9.17134956e-02 -3.32410514e-01
4.49097723e-01 7.30963111e-01 3.58899981e-01 -4.33652341e-01
-6.91138029e-01 -7.55358189e-02 1.08328916e-01 -4.83864427e-01
-5.25609553e-01 -8.06777179e-01 -8.01586747e-01 2.00754199e-02
-4.96597171e-01 2.62571871e-01 9.58376169e-01 9.00862992e-01
4.33880955e-01 2.82200336e-01 8.59901309e-01 -8.74136567e-01
-6.77257955e-01 -1.32382119e+00 -4.66956824e-01 4.94067043e-01
8.99668634e-01 -6.69586182e-01 -7.31335044e-01 8.64353180e-02] | [14.355745315551758, 6.065332889556885] |
2233c474-5c53-41bf-a82c-da8005921fba | building-a-corpus-for-biomedical-relation | 2306.08403 | null | https://arxiv.org/abs/2306.08403v1 | https://arxiv.org/pdf/2306.08403v1.pdf | Building a Corpus for Biomedical Relation Extraction of Species Mentions | We present a manually annotated corpus, Species-Species Interaction, for extracting meaningful binary relations between species, in biomedical texts, at sentence level, with a focus on the gut microbiota. The corpus leverages PubTator to annotate species in full-text articles after evaluating different Named Entity Recognition species taggers. Our first results are promising for extracting relations between species using BERT and its biomedical variants. | ['Samuel Chaffron', 'Solen Quiniou', 'Oumaima El Khettari'] | 2023-06-14 | null | null | null | null | ['relation-extraction'] | ['natural-language-processing'] | [ 5.38472772e-01 -8.43823254e-02 -7.18472123e-01 1.87167525e-01
-3.49439412e-01 -9.66869414e-01 5.68221986e-01 1.41041112e+00
-5.56263506e-01 1.30091763e+00 5.53804636e-01 -3.81797135e-01
-1.88282251e-01 -5.12691498e-01 -6.38872683e-01 -6.98742926e-01
-7.37783790e-01 2.85007566e-01 -2.17490673e-01 -2.50795752e-01
4.66672629e-01 2.16363877e-01 -9.28281784e-01 7.64500439e-01
9.56385136e-01 5.13292968e-01 4.16320860e-02 8.35498929e-01
-7.53490686e-01 8.07538629e-01 -7.41820931e-01 -9.00974154e-01
-1.19177453e-01 -2.68864483e-01 -9.56291854e-01 -6.88374162e-01
2.93525964e-01 7.06703961e-01 9.14474502e-02 1.18084610e+00
2.87206739e-01 -5.49018979e-01 1.15564978e+00 -9.34550703e-01
-3.80361348e-01 1.12615597e+00 -4.56691474e-01 7.79330134e-01
5.70513248e-01 1.36896446e-01 1.47852349e+00 -1.23500258e-01
1.50450873e+00 1.22864163e+00 1.16809249e+00 2.96626657e-01
-1.40442526e+00 -5.96841693e-01 -3.16632569e-01 -1.02282830e-01
-1.56965852e+00 -1.64979156e-02 8.28948170e-02 -1.25186050e+00
1.98889649e+00 7.31703877e-01 4.83231544e-01 1.75243104e+00
8.46816838e-01 1.32938817e-01 7.98228502e-01 -2.14886189e-01
-4.34095003e-02 1.05385482e-01 3.25657040e-01 3.57889116e-01
9.79140401e-01 -1.99501738e-01 -7.13553190e-01 -1.00789881e+00
1.94496408e-01 -2.73968965e-01 -2.66619235e-01 3.41792792e-01
-1.62853229e+00 4.88267541e-01 4.58558738e-01 8.44270706e-01
-7.28071570e-01 -2.52121150e-01 1.12174559e+00 7.80706555e-02
9.66749191e-01 1.41007018e+00 -1.08814192e+00 3.66234004e-01
-5.81575513e-01 -1.77574471e-01 1.37065709e+00 1.04590225e+00
-1.29741251e-01 -9.58597064e-01 -2.20782876e-01 9.32926536e-01
-3.58529650e-02 2.86805719e-01 6.98993742e-01 -1.84408963e-01
2.00856194e-01 5.41655004e-01 -6.12534061e-02 -1.15803897e+00
-7.21295297e-01 -2.70574749e-01 -8.03725719e-01 -1.18292630e+00
-1.90698668e-01 -3.04983646e-01 -4.69178647e-01 1.54317451e+00
6.87305987e-01 5.75359225e-01 2.42038026e-01 1.10642783e-01
1.42525232e+00 2.87116945e-01 1.10283911e+00 -4.98722136e-01
2.00908113e+00 -5.40244937e-01 -9.07109559e-01 -5.64051494e-02
7.80581176e-01 -9.14339066e-01 -1.99857261e-02 4.23173398e-01
-2.76359111e-01 1.06929258e-01 -9.11778748e-01 -1.97328534e-03
-1.19403636e+00 -5.46619654e-01 7.76201963e-01 2.11978868e-01
-7.77547300e-01 5.64085543e-01 -7.22225070e-01 -7.29254425e-01
4.68622088e-01 2.15184838e-01 -6.21889591e-01 4.08008724e-01
-1.57745337e+00 1.26149642e+00 5.64954937e-01 -3.14943284e-01
-2.85365105e-01 -1.38285446e+00 -7.50576615e-01 -3.84086430e-01
-8.72011781e-02 -1.54210472e+00 8.95560563e-01 2.07704287e-02
-5.41016400e-01 1.49760163e+00 -1.34426087e-01 -8.00379455e-01
-2.29181454e-01 -7.14808851e-02 -8.80066574e-01 3.74051213e-01
5.36942244e-01 4.56865430e-01 -2.75976479e-01 -5.71897745e-01
-4.76927191e-01 -2.37692609e-01 -2.57778555e-01 -3.94747347e-01
-6.17904127e-01 2.89101452e-01 3.83589715e-01 -8.44184577e-01
-6.28840327e-01 -4.80483741e-01 -9.53402460e-01 -3.70487869e-01
-9.62050259e-01 -3.94322723e-01 -3.99392694e-01 -8.68360758e-01
1.32464600e+00 -1.75663817e+00 2.34315395e-01 -2.81718254e-01
5.38012147e-01 -3.37556511e-01 -2.52614349e-01 7.93643534e-01
-2.54839718e-01 8.57252240e-01 -2.21297130e-01 2.60430664e-01
-3.69414151e-01 7.35986531e-02 -3.72129798e-01 5.14006674e-01
6.39122128e-01 7.90625274e-01 -1.63803303e+00 -1.26558566e+00
-8.93034860e-02 4.45436448e-01 -6.52180493e-01 -5.49563169e-02
-3.83414090e-01 1.76378608e-01 -6.50568068e-01 7.30728984e-01
5.27785838e-01 -2.48901814e-01 5.01251400e-01 -5.78406632e-01
-1.19026572e-01 7.59366453e-01 -1.16510861e-01 1.83668125e+00
-4.18783069e-01 3.27072084e-01 1.25868581e-02 -7.51634598e-01
5.00450611e-01 3.40920866e-01 1.14452469e+00 -7.26944627e-03
2.38963559e-01 4.26500440e-01 1.95120305e-01 -9.94741738e-01
2.31786475e-01 -4.33689356e-01 -3.96177739e-01 -6.23610556e-01
4.13448662e-01 -2.92652220e-01 6.79975629e-01 2.42399618e-01
1.54309893e+00 -1.43107225e-03 1.37921214e+00 -1.00591564e+00
6.32367194e-01 7.37248421e-01 4.38203931e-01 3.80196363e-01
-4.86493818e-02 -1.02780141e-01 4.99889791e-01 -2.53524959e-01
-6.97376013e-01 -7.91971743e-01 -1.15204346e+00 5.96462071e-01
-2.24466965e-01 -8.62818599e-01 -4.56961960e-01 -5.60804963e-01
4.09880549e-01 5.13502717e-01 -1.01221383e+00 9.80206951e-02
-2.33238786e-01 -1.36778820e+00 1.17715704e+00 -2.08022594e-01
-2.22787648e-01 -8.13565850e-01 -3.88971180e-01 4.46605623e-01
-4.39486742e-01 -1.40238464e+00 -4.91247118e-01 6.76706791e-01
-5.91203153e-01 -1.58741832e+00 -4.98314142e-01 -8.56569648e-01
1.62052229e-01 -4.43698257e-01 1.64405990e+00 -2.36423910e-01
-1.10146952e+00 -1.73210353e-01 -2.51647800e-01 -7.23071337e-01
-7.01624513e-01 2.36652985e-01 2.22267345e-01 -7.66780436e-01
7.68325865e-01 -4.69462454e-01 -6.36834383e-01 -2.04743966e-01
-6.46262586e-01 -4.56527770e-02 4.50765997e-01 7.65398502e-01
6.60668015e-01 -5.07869363e-01 5.14605522e-01 -1.13817155e+00
5.02301693e-01 -1.24472070e+00 1.84212372e-01 2.71876454e-01
-1.55990288e-01 1.05505623e-01 5.19762456e-01 -5.83084106e-01
-3.35033983e-01 -1.93766236e-01 -3.48058730e-01 3.41954172e-01
-2.14434028e-01 9.58314300e-01 4.08430308e-01 3.55220079e-01
1.26047850e+00 -1.59033507e-01 -3.64652693e-01 -4.59084868e-01
4.36558276e-01 1.22941744e+00 6.26348257e-01 -4.62033778e-01
-1.39444634e-01 1.63051978e-01 2.97664821e-01 -8.87278140e-01
-9.85291302e-01 -1.14638960e+00 -7.93774247e-01 5.35106957e-01
1.27701926e+00 -1.12228560e+00 -8.53312194e-01 -1.53289706e-01
-1.39365280e+00 8.42852890e-02 -7.18619227e-02 5.22648752e-01
1.12269260e-02 2.78817385e-01 -1.20445454e+00 -4.69000548e-01
-9.59699929e-01 -4.62439656e-01 1.40313435e+00 -4.21033323e-01
-1.00361466e+00 -1.36315119e+00 7.62777150e-01 6.52032122e-02
-2.58768320e-01 9.20788586e-01 1.09815240e+00 -1.20776391e+00
5.83696961e-01 5.85767031e-02 1.12294041e-01 -2.75540769e-01
8.24536383e-01 3.20970595e-01 -4.89204466e-01 3.46998215e-01
-1.66825935e-01 2.23625705e-01 8.85646760e-01 3.24746400e-01
9.24225211e-01 -8.15867186e-01 -1.21844053e+00 4.09665883e-01
1.25338185e+00 -1.77574530e-01 -3.08621884e-03 4.35369551e-01
5.33380508e-01 8.72300506e-01 3.80375087e-01 2.01339066e-01
-3.68944034e-02 1.11109078e-01 -1.03633031e-01 -4.63402607e-02
-2.69489121e-02 -5.53578995e-02 -4.70326841e-02 1.14429939e+00
2.63356954e-01 -3.70140344e-01 -1.18338299e+00 8.90745103e-01
-1.14907837e+00 -8.07018876e-01 -7.04235256e-01 1.57145178e+00
1.77506208e+00 -2.51703531e-01 -4.65811081e-02 -6.22233331e-01
8.48240972e-01 -4.08820927e-01 -5.30747235e-01 -4.93335336e-01
-3.95386130e-01 4.75416392e-01 8.35649550e-01 1.78857669e-01
-1.25654173e+00 8.51665556e-01 7.72264481e+00 7.27041364e-01
-8.09258759e-01 -1.04266461e-02 5.84610283e-01 -5.60018159e-02
-1.27876043e-01 -3.40789527e-01 -6.29810989e-01 5.16095996e-01
1.41677666e+00 -5.90752125e-01 -1.10506967e-01 5.93788326e-01
-9.52758789e-02 2.38624156e-01 -1.28494108e+00 6.88023150e-01
-1.72161013e-02 -1.65203106e+00 3.68310213e-02 1.20049149e-01
5.60196936e-01 5.76443315e-01 -5.59294581e-01 -1.39148429e-01
4.25311565e-01 -8.83520067e-01 2.78695554e-01 2.65428483e-01
7.00389981e-01 -2.06259415e-01 7.86414623e-01 -1.64878438e-03
-9.25801396e-01 3.85059625e-01 -3.22836310e-01 2.68331110e-01
-5.04471920e-03 1.21121252e+00 -1.43074632e+00 9.30989981e-01
7.28675187e-01 1.35311592e+00 -6.26278639e-01 1.16667306e+00
1.07984804e-01 4.91744161e-01 -3.82406175e-01 -4.72512126e-01
-1.93220153e-01 2.22054899e-01 8.15656960e-01 2.34895086e+00
2.09893808e-02 1.85791895e-01 7.71103352e-02 7.24426091e-01
1.34677052e-01 1.12082243e+00 -6.73633933e-01 -7.46321440e-01
6.23601973e-01 1.31363297e+00 -9.64715362e-01 -7.14436293e-01
1.06983230e-01 7.53793240e-01 1.29207537e-01 -1.49040431e-01
-4.28237259e-01 -2.97290713e-01 1.07641768e+00 -5.67723811e-02
4.52176034e-02 3.20338577e-01 -5.79449296e-01 -1.03107297e+00
-7.39355326e-01 -8.22567999e-01 1.16388547e+00 -5.82107663e-01
-2.60533094e+00 6.96307540e-01 2.96775382e-02 -1.22757876e+00
-2.21455395e-02 -1.02641511e+00 2.59112567e-01 8.77963960e-01
-9.16207016e-01 -9.47284818e-01 4.78668272e-01 5.69876432e-02
2.03729227e-01 2.82593459e-01 1.22087741e+00 4.15717572e-01
-6.99920893e-01 3.96238148e-01 -1.78702995e-01 -1.87151745e-01
1.09107959e+00 -1.24315822e+00 6.68312609e-01 -1.33465856e-01
1.34637028e-01 1.46865320e+00 1.17684245e+00 -1.34484565e+00
-1.09581470e+00 -1.20387447e+00 1.35634398e+00 -8.25991035e-01
1.56748450e+00 -5.33428311e-01 -5.88755965e-01 1.92175537e-01
5.00430942e-01 -5.85737705e-01 1.69288337e+00 3.38941723e-01
-8.18072081e-01 7.74896145e-01 -1.42029786e+00 5.43413341e-01
1.41734469e+00 -5.40530801e-01 -9.08219695e-01 1.17093933e+00
1.27706003e+00 -3.00291866e-01 -1.75205088e+00 5.06455958e-01
2.90338516e-01 1.93749949e-01 1.54487777e+00 -1.35506511e+00
9.16150868e-01 1.43265016e-02 -4.44492772e-02 -1.26588547e+00
-1.95743099e-01 -2.89217174e-01 1.72921360e-01 8.92802298e-01
8.50053132e-01 -8.50246072e-01 -1.11762024e-01 -6.81514069e-02
-1.25044882e-01 -5.45601368e-01 -9.87758279e-01 -6.73496366e-01
5.22119999e-01 6.58953264e-02 6.03912115e-01 1.76074517e+00
9.33022201e-01 8.92008007e-01 2.27109134e-01 7.91985989e-02
3.14547509e-01 -1.08527571e-01 7.97847360e-02 -1.37505388e+00
-3.84976208e-01 -8.19992602e-01 -7.03039110e-01 -6.43040314e-02
2.22044557e-01 -1.29512417e+00 3.03188473e-01 -1.61098230e+00
7.93252826e-01 9.17584226e-02 -5.84455609e-01 4.24427539e-01
-7.17844844e-01 2.66448766e-01 -6.27719283e-01 1.59456402e-01
-2.85825342e-01 -4.39052641e-01 8.79597247e-01 -4.78166848e-01
8.28226507e-02 -8.50558341e-01 -7.90445864e-01 5.60303688e-01
2.28863046e-01 -8.43636036e-01 3.30665350e-01 1.83869630e-01
5.52899837e-01 -2.72260725e-01 1.06804445e-01 -3.28757167e-01
-6.30977601e-02 -4.79299992e-01 -6.35671243e-02 -3.70430738e-01
-3.08310330e-01 -2.80659229e-01 6.76814914e-01 1.22915089e+00
-7.41460323e-01 -1.44375935e-01 6.15628719e-01 4.53364044e-01
9.24141407e-02 -3.04064274e-01 2.55872905e-01 -2.93818593e-01
-2.20697328e-01 8.74619260e-02 -3.03798348e-01 1.67607620e-01
6.37318611e-01 3.96329701e-01 -6.49612784e-01 7.69281626e-01
-5.40196717e-01 1.00249268e-01 1.97977424e-01 5.68393826e-01
-3.53501946e-01 -8.16570044e-01 -1.11308420e+00 -9.23546791e-01
4.81536686e-01 -5.78114212e-01 1.57545090e-01 1.00489330e+00
-6.66629374e-01 8.05853784e-01 1.24269158e-01 -7.05794990e-01
-1.31636214e+00 1.29597569e+00 1.57857522e-01 -5.88897288e-01
-3.69190216e-01 1.19950974e+00 4.01889473e-01 -6.05981827e-01
-4.10228878e-01 -6.12921178e-01 -4.28308100e-01 5.11772573e-01
5.19038796e-01 -2.16211453e-01 1.78957850e-01 -5.64910352e-01
-1.14126396e+00 4.36083376e-02 2.69692749e-01 3.04006130e-01
1.37630796e+00 2.84825742e-01 -9.79818165e-01 8.30081999e-01
1.60624444e+00 2.16345698e-01 2.85252512e-01 1.93792313e-01
4.53966677e-01 9.93345529e-02 -3.71613801e-01 -1.12896931e+00
-3.62735569e-01 2.77512491e-01 1.22391291e-01 2.51950830e-01
4.75943953e-01 2.45676294e-01 5.76054633e-01 3.17181677e-01
2.93069392e-01 -3.70238423e-01 -5.16676664e-01 1.38524085e-01
9.58447278e-01 -8.43298197e-01 3.73225391e-01 -1.18174410e+00
-2.00919405e-01 7.13214636e-01 2.57265806e-01 1.73717380e-01
6.24782979e-01 6.81071460e-01 -1.22546747e-01 -7.31990099e-01
-6.88296676e-01 2.49478202e-02 5.74037313e-01 4.30452138e-01
1.12341106e+00 3.65477681e-01 -1.08572161e+00 7.13260770e-01
-5.89312553e-01 -2.93928534e-01 7.46471286e-02 6.48871601e-01
1.62311599e-01 -1.15291703e+00 -1.29359603e-01 8.42592597e-01
-1.05125034e+00 -1.11716890e+00 -1.26500070e+00 4.58990604e-01
5.14222682e-01 1.07255960e+00 1.19790867e-01 -2.41338447e-01
2.50800967e-01 7.68144615e-03 5.83333969e-01 -9.25270736e-01
-1.30754542e+00 -8.87812395e-03 8.48642588e-01 -9.67939496e-02
-7.72521794e-01 -9.13624525e-01 -1.11899412e+00 3.33105326e-02
-5.43670118e-01 4.96859461e-01 5.52393198e-01 7.25905240e-01
5.37364841e-01 9.13740754e-01 2.27738321e-01 -1.81398064e-01
1.89030573e-01 -1.37757349e+00 -2.69417435e-01 6.57877982e-01
7.09735528e-02 -3.72783929e-01 -3.64695519e-01 3.68676424e-01] | [8.443771362304688, 8.756741523742676] |
6417c529-fc67-4021-b611-e503213e746a | deep-reinforcement-learning-for-surgical | 1806.08089 | null | http://arxiv.org/abs/1806.08089v1 | http://arxiv.org/pdf/1806.08089v1.pdf | Deep Reinforcement Learning for Surgical Gesture Segmentation and Classification | Recognition of surgical gesture is crucial for surgical skill assessment and
efficient surgery training. Prior works on this task are based on either
variant graphical models such as HMMs and CRFs, or deep learning models such as
Recurrent Neural Networks and Temporal Convolutional Networks. Most of the
current approaches usually suffer from over-segmentation and therefore low
segment-level edit scores. In contrast, we present an essentially different
methodology by modeling the task as a sequential decision-making process. An
intelligent agent is trained using reinforcement learning with hierarchical
features from a deep model. Temporal consistency is integrated into our action
design and reward mechanism to reduce over-segmentation errors. Experiments on
JIGSAWS dataset demonstrate that the proposed method performs better than
state-of-the-art methods in terms of the edit score and on par in frame-wise
accuracy. Our code will be released later. | ['Tingting Jiang', 'Daochang Liu'] | 2018-06-21 | null | null | null | null | ['surgical-gesture-recognition'] | ['medical'] | [ 4.08952355e-01 5.05070627e-01 -4.85203117e-01 -4.53110725e-01
-7.35145152e-01 -2.51371026e-01 4.14741904e-01 2.27112636e-01
-7.93191314e-01 6.63224518e-01 2.98318386e-01 -3.63374293e-01
-8.41124207e-02 -3.05400103e-01 -5.57790637e-01 -7.64367521e-01
2.53693815e-02 4.77111161e-01 3.54177386e-01 7.18045160e-02
4.97476041e-01 3.48514199e-01 -1.14388311e+00 2.00734183e-01
7.20415771e-01 7.98530459e-01 2.16814727e-01 8.99277151e-01
-7.30822682e-02 1.27057576e+00 -4.01224554e-01 -1.08118266e-01
1.55437320e-01 -7.43276119e-01 -1.10764492e+00 1.17675327e-02
-1.56032249e-01 -3.90289932e-01 -2.48112082e-01 7.59238005e-01
5.35631657e-01 1.12649553e-01 3.39467824e-01 -4.28684711e-01
-2.59873047e-02 5.79712629e-01 -1.88199431e-01 3.03738024e-02
2.58496642e-01 1.89521521e-01 4.20624882e-01 -2.98839539e-01
8.81931484e-01 7.88336873e-01 6.08369529e-01 7.94627845e-01
-9.26141739e-01 -3.28425437e-01 1.64167643e-01 5.46488166e-02
-8.20463300e-01 -8.31862539e-02 4.18398321e-01 -4.89503503e-01
1.03485572e+00 8.27922523e-02 9.68056738e-01 1.05049992e+00
8.58326674e-01 1.00757301e+00 1.27493393e+00 -4.69650924e-01
2.89564699e-01 -3.98126036e-01 -6.48169518e-02 1.11526918e+00
-1.27796188e-01 5.27798951e-01 -5.59033036e-01 3.18061262e-01
1.07161105e+00 2.55919695e-01 1.48106039e-01 -3.82463723e-01
-1.36532855e+00 7.96388686e-01 3.81377548e-01 6.44555151e-01
-6.58408403e-01 4.65161592e-01 5.06667316e-01 8.45533013e-02
-6.58100173e-02 3.77005190e-01 -2.26423547e-01 -6.50559604e-01
-1.21325934e+00 4.60152104e-02 8.03743124e-01 8.73481512e-01
1.92827418e-01 -1.76022619e-01 -4.42239285e-01 5.37577689e-01
3.57861400e-01 -3.18473786e-01 9.38073754e-01 -1.21758592e+00
-9.51453969e-02 3.88189673e-01 -1.44878194e-01 -3.67038280e-01
-7.97110915e-01 -2.75612235e-01 -6.15125775e-01 5.23646295e-01
2.94465929e-01 -1.87680826e-01 -1.57843196e+00 1.42937160e+00
1.56125233e-01 2.74270326e-01 -1.04881867e-04 7.64077783e-01
6.53241813e-01 1.00809284e-01 3.78898650e-01 -3.36727083e-01
1.09301078e+00 -1.42609239e+00 -1.06534398e+00 -2.28772074e-01
9.00089383e-01 -7.70486474e-01 4.22671258e-01 5.10811329e-01
-1.28620541e+00 -3.71325612e-01 -8.11072290e-01 7.62899071e-02
7.10028782e-03 2.66279310e-01 6.99426472e-01 6.19505465e-01
-1.09235620e+00 9.13955867e-01 -1.65194511e+00 -3.90082389e-01
3.85646135e-01 6.91370904e-01 -1.94543064e-01 1.92141756e-01
-6.87800229e-01 1.25446427e+00 3.54884267e-01 3.27581078e-01
-1.15985024e+00 -2.97243834e-01 -1.01473486e+00 -3.65012348e-01
4.91603553e-01 -7.10183680e-01 1.97911990e+00 -9.82769907e-01
-2.35577321e+00 9.44076777e-01 -1.14030093e-01 -8.10628355e-01
7.71790445e-01 -5.69112837e-01 2.72192925e-01 3.81434895e-02
-3.95995557e-01 6.70015037e-01 6.64566815e-01 -9.82371986e-01
-6.63938046e-01 -1.98745266e-01 5.38730510e-02 4.05705720e-01
3.52428138e-01 1.10839635e-01 -5.10085702e-01 -6.01687312e-01
4.92693074e-02 -1.44345772e+00 -1.02976644e+00 -1.33768320e-01
-4.18366641e-01 -1.03318959e-01 2.77355969e-01 -8.86938512e-01
1.32062542e+00 -1.72132361e+00 2.67331094e-01 2.94944253e-02
2.69294195e-02 2.39757165e-01 3.91922832e-01 3.51565838e-01
5.93043379e-02 -1.07415006e-01 -3.17461163e-01 -3.95069093e-01
-3.28200281e-01 6.23747110e-01 3.06473970e-01 4.91906464e-01
-4.83982526e-02 7.82804728e-01 -1.00534678e+00 -8.13617706e-01
4.83910888e-01 1.76449001e-01 -4.28108603e-01 4.84400243e-01
-1.18720070e-01 9.60316956e-01 -4.56837982e-01 6.35513306e-01
-1.07638553e-01 -4.33572009e-02 5.75358212e-01 2.73479044e-01
-2.93179333e-01 3.67092967e-01 -6.84309781e-01 2.54272914e+00
-5.71926534e-01 2.40697786e-01 -3.02243024e-01 -1.11541760e+00
6.67682528e-01 5.29505610e-01 6.44461453e-01 -5.94082534e-01
4.06498909e-01 1.45209044e-01 2.81084508e-01 -6.75558567e-01
5.52384734e-01 -3.09526682e-01 -5.80817275e-02 4.95684594e-02
1.27766058e-01 -1.45455778e-01 6.75621927e-02 -7.57008046e-02
1.27556789e+00 9.71367359e-01 6.58210993e-01 5.74985966e-02
1.75503537e-01 2.19383553e-01 8.06253016e-01 1.00818610e+00
-4.56582218e-01 6.96952939e-01 4.96133447e-01 -4.30512935e-01
-7.81908512e-01 -6.03726923e-01 6.56190515e-02 6.38305962e-01
1.85198888e-01 -2.54528165e-01 -9.30420637e-01 -9.16059196e-01
-5.46609461e-01 7.65660048e-01 -8.46866012e-01 -1.30795389e-01
-8.85974646e-01 -4.68991965e-01 2.63584435e-01 8.08308661e-01
-3.84601089e-03 -1.58730602e+00 -1.30318570e+00 7.02048481e-01
1.09698415e-01 -1.26297712e+00 -1.97324827e-01 4.16120976e-01
-1.30887318e+00 -9.23976064e-01 -8.90665531e-01 -8.90799701e-01
8.08140516e-01 -5.05862117e-01 9.88965690e-01 1.44954085e-01
-5.98631561e-01 1.46342188e-01 -5.38668513e-01 -4.56806481e-01
-5.38740635e-01 1.39437066e-02 -2.78876692e-01 -4.22540992e-01
-6.93201600e-03 -2.01141298e-01 -9.25217271e-01 1.15024075e-01
-8.53938520e-01 5.14434516e-01 1.11584425e+00 1.14558804e+00
6.71667457e-01 -5.49706697e-01 -2.51297392e-02 -1.23279703e+00
2.59622157e-01 -1.97311472e-02 -4.52620476e-01 2.04764783e-01
-7.41271973e-01 3.12794268e-01 2.14279547e-01 -3.65488052e-01
-1.15004933e+00 6.92248225e-01 -1.83242679e-01 -4.13343191e-01
-4.17660594e-01 5.73204458e-01 7.45630920e-01 -5.39241023e-02
3.63361746e-01 1.76572293e-01 2.48797327e-01 -1.85579956e-01
5.17398529e-02 1.49787039e-01 4.80888277e-01 -3.06840628e-01
2.39189461e-01 1.75274402e-01 -2.21467335e-02 -3.02164942e-01
-7.04813898e-01 -4.32394385e-01 -8.58959556e-01 -4.80232447e-01
1.23097754e+00 -6.11802638e-01 -7.62156010e-01 5.79772949e-01
-1.07620609e+00 -7.44237244e-01 -2.87638217e-01 9.25180018e-01
-1.08322155e+00 1.40243039e-01 -1.00428462e+00 -7.55236328e-01
-2.78911531e-01 -1.57503581e+00 1.11067080e+00 5.57305574e-01
-2.96031624e-01 -8.88734698e-01 4.26340103e-01 2.91381001e-01
2.69928873e-01 6.05509520e-01 6.12801731e-01 -5.76482654e-01
-5.80231965e-01 -4.63815153e-01 2.77939081e-01 1.52913228e-01
2.09757853e-02 7.65813366e-02 -4.98654634e-01 1.57819707e-02
-2.45071724e-01 -3.00610781e-01 7.09850490e-01 9.38425064e-01
1.29875982e+00 4.39096205e-02 -4.23439413e-01 3.30475152e-01
1.40461481e+00 5.30953228e-01 7.94840157e-01 4.34871614e-01
3.91496629e-01 5.91607988e-01 9.78185475e-01 2.72078693e-01
2.89262563e-01 7.62019098e-01 5.08189321e-01 -2.11326092e-01
-1.23350434e-01 -4.72328216e-02 2.32125193e-01 8.33407938e-01
-4.94157970e-01 1.80804238e-01 -1.02392578e+00 5.06286442e-01
-2.38360143e+00 -7.94846177e-01 1.79143220e-01 2.05839562e+00
9.55218673e-01 2.72407174e-01 2.13216757e-03 -2.13868126e-01
3.91116232e-01 3.31835798e-03 -3.34081084e-01 -7.09830880e-01
5.28446496e-01 4.49678749e-01 7.21560717e-01 4.72067088e-01
-1.06763768e+00 1.21497762e+00 6.48097849e+00 5.01121104e-01
-1.08836985e+00 6.60683066e-02 5.16248405e-01 -1.93368390e-01
3.67853403e-01 1.70393407e-01 -5.99693954e-01 1.81584433e-01
1.01078880e+00 3.28575313e-01 2.95312912e-03 7.49780834e-01
3.47393960e-01 -5.68429172e-01 -1.02275670e+00 7.12056100e-01
2.05076020e-02 -1.47035515e+00 -5.83090961e-01 5.89709692e-02
6.11735225e-01 -5.17633446e-02 -1.76108003e-01 3.71966213e-01
7.62296975e-01 -1.16649973e+00 6.66199267e-01 1.00482929e+00
4.63951319e-01 -5.30481517e-01 1.07597697e+00 2.08458617e-01
-8.17695141e-01 1.17674105e-01 1.30344227e-01 1.91653386e-01
3.77247512e-01 7.61994570e-02 -1.07909286e+00 8.24562669e-01
5.13500929e-01 7.78872669e-01 -1.19353727e-01 1.29112995e+00
-4.96985376e-01 5.46213984e-01 -8.59063268e-02 -1.64909467e-01
6.10946476e-01 -1.00391924e-01 2.69189268e-01 1.39054620e+00
1.19409204e-01 2.41542175e-01 3.36143374e-01 2.79138118e-01
2.20747322e-01 5.13058109e-03 -4.36064929e-01 -1.35267735e-01
-3.27619433e-01 1.18872428e+00 -1.02168393e+00 -3.63375515e-01
-3.80840987e-01 1.13169956e+00 2.29372799e-01 1.08579524e-01
-9.25153613e-01 -1.74812272e-01 2.88532913e-01 -2.51592904e-01
3.26517165e-01 -4.46856678e-01 -5.20617187e-01 -7.50212193e-01
-1.18046843e-01 -7.44249463e-01 3.39292884e-01 -5.23679674e-01
-4.63904768e-01 7.75637269e-01 -2.44276524e-01 -1.46452022e+00
-7.20263362e-01 -4.33270186e-01 -4.77311820e-01 4.89632815e-01
-1.59500408e+00 -1.06788123e+00 -2.86070079e-01 3.73406142e-01
8.86352718e-01 2.22911373e-01 1.07871306e+00 -1.95137873e-01
-5.93800128e-01 3.20297122e-01 -1.15552790e-01 2.39925981e-01
5.73987424e-01 -1.29120493e+00 6.23128973e-02 8.41640353e-01
-9.28109139e-02 3.99075419e-01 7.77720153e-01 -7.36046970e-01
-1.13909900e+00 -6.58844352e-01 7.63815522e-01 -1.34975776e-01
4.10479367e-01 1.25840589e-01 -5.41366994e-01 7.35922396e-01
4.45498466e-01 -1.54105742e-02 6.31678998e-01 -2.71636005e-02
4.57310021e-01 3.39027554e-01 -9.05622005e-01 5.47732532e-01
1.03188491e+00 -1.16812944e-01 -6.41698122e-01 3.82771194e-01
5.44178188e-01 -1.18267488e+00 -9.83933210e-01 6.64899170e-01
8.46592665e-01 -9.87036169e-01 5.03246546e-01 -7.84444511e-01
7.41811991e-01 -4.70850207e-02 4.44784284e-01 -1.08558536e+00
-1.31998911e-01 -6.86296225e-01 3.90585549e-02 4.63793129e-01
4.90961075e-01 -1.94387324e-02 1.06165874e+00 7.55331993e-01
-5.11413336e-01 -1.15690494e+00 -8.77922535e-01 -5.76665401e-01
-2.88714886e-01 -3.32519799e-01 -1.96209595e-01 6.15802765e-01
3.89252529e-02 -1.96445838e-01 -4.76593375e-01 1.50137926e-02
3.33666712e-01 1.45974636e-01 4.22527552e-01 -8.29425454e-01
-3.05013627e-01 -5.44050872e-01 -5.55327237e-01 -7.10666299e-01
1.06618702e-01 -6.05413556e-01 7.40035295e-01 -1.82950842e+00
2.47500185e-02 -2.74841011e-01 -3.00605029e-01 7.55991459e-01
1.00585945e-01 -4.33876887e-02 8.42576548e-02 6.35308251e-02
-6.85343146e-01 3.58803779e-01 1.33551753e+00 3.26890588e-01
-5.34259498e-01 1.84776902e-01 -2.00767234e-01 8.94223750e-01
6.21513784e-01 -7.29061902e-01 8.67795795e-02 -3.38739127e-01
-1.88654646e-01 6.20411932e-01 6.16164058e-02 -9.76303101e-01
6.19195342e-01 -2.44242772e-01 1.60957471e-01 -3.86767745e-01
1.76972747e-01 -6.90356791e-01 1.20100498e-01 1.13341594e+00
-5.73147655e-01 4.00199629e-02 2.08454862e-01 5.63908279e-01
-4.42809969e-01 -3.45541805e-01 8.63550007e-01 -3.56395841e-01
-6.54407799e-01 3.57764125e-01 -7.87658870e-01 -4.67571020e-01
1.22618592e+00 -3.49470377e-01 2.30119914e-01 -1.95634946e-01
-1.26001751e+00 -5.88803971e-03 2.30510563e-01 4.11173552e-01
6.37001097e-01 -8.81851435e-01 -5.41851342e-01 -1.27444953e-01
-1.60775080e-01 1.87189937e-01 2.51912624e-01 1.39833295e+00
-9.06587303e-01 5.08126557e-01 -2.71724671e-01 -6.54090822e-01
-1.35613227e+00 1.65570781e-01 5.88931262e-01 -9.61761057e-01
-8.09183061e-01 8.72306585e-01 5.21184411e-03 -3.03760231e-01
5.95526516e-01 -4.28395778e-01 -3.36216629e-01 -3.35200250e-01
2.34081268e-01 -1.17475607e-01 6.42179549e-02 -1.88458890e-01
-4.25509572e-01 5.54847121e-01 -3.18757534e-01 -2.66869366e-01
1.31206608e+00 1.19124964e-01 1.89952388e-01 3.62086058e-01
6.80686951e-01 -5.06839931e-01 -1.34586370e+00 -2.33222514e-01
3.50511074e-01 -1.78681314e-01 2.58133590e-01 -1.15427351e+00
-8.67553651e-01 6.85130894e-01 8.77182961e-01 -2.45533586e-01
1.04499483e+00 -1.15393914e-01 7.88570285e-01 5.89911863e-02
3.75526726e-01 -1.23061943e+00 4.22960036e-02 4.45535839e-01
6.13699853e-01 -1.18278635e+00 -1.67787537e-01 -1.33357406e-01
-1.09541988e+00 1.26941228e+00 5.56404829e-01 -3.66355360e-01
5.77048659e-01 4.83932614e-01 3.33572835e-01 -1.11829393e-01
-7.11943567e-01 -2.66787559e-01 2.82481939e-01 1.82055935e-01
7.37540603e-01 3.76659818e-02 -7.82778919e-01 3.73357326e-01
-5.42708002e-02 6.33178711e-01 4.24049675e-01 1.53197515e+00
-2.50778377e-01 -1.34909201e+00 1.46017238e-01 3.99823785e-01
-8.38282526e-01 -5.32514341e-02 2.94926073e-02 6.95697546e-01
-1.99184015e-01 4.81396824e-01 -9.89339054e-02 -2.32569918e-01
3.57296228e-01 1.04659162e-01 7.19138801e-01 -7.33736157e-01
-1.00849986e+00 3.15525949e-01 2.27140591e-01 -1.16549420e+00
-8.56776476e-01 -8.73911738e-01 -1.62212026e+00 2.08569169e-01
-2.06072509e-01 -1.31840244e-01 9.86478925e-01 1.23294103e+00
-3.41529362e-02 1.06299114e+00 3.17577779e-01 -8.76462638e-01
-3.84366035e-01 -9.02111590e-01 -2.47419953e-01 1.32414547e-03
3.06458265e-01 -4.10451442e-01 2.70884216e-01 1.71440989e-01] | [14.060250282287598, -3.341071605682373] |
eec73d30-b835-4e9a-9f10-c2459b7219c5 | generative-adversarial-networks-based-on | null | null | https://aclanthology.org/2021.rocling-1.36 | https://aclanthology.org/2021.rocling-1.36.pdf | Generative Adversarial Networks based on Mixed-Attentions for Citation Intent Classification in Scientific Publications | We propose the mixed-attention-based Generative Adversarial Network (named maGAN), and apply it for citation intent classification in scientific publication. We select domain-specific training data, propose a mixed-attention mechanism, and employ generative adversarial network architecture for pre-training language model and fine-tuning to the downstream multi-class classification task. Experiments were conducted on the SciCite datasets to compare model performance. Our proposed maGAN model achieved the best Macro-F1 of 0.8532. | ['Lung-Hao Lee', 'Chao-Yi Chen', 'Yuh-Shyang Wang'] | null | null | null | null | rocling-2021-10 | ['citation-intent-classification'] | ['natural-language-processing'] | [-1.56671539e-01 2.82526553e-01 -2.00238094e-01 -3.72326165e-01
-1.04338598e+00 -4.49953079e-01 1.00940013e+00 -4.59709823e-01
-4.39689040e-01 1.03517449e+00 3.96790832e-01 -7.73633063e-01
1.50005594e-01 -8.88737798e-01 -9.06555235e-01 -3.97202641e-01
3.87575120e-01 4.59310353e-01 -2.98131436e-01 9.74458456e-02
3.87463212e-01 5.10433435e-01 -6.29378974e-01 2.38290176e-01
1.15980256e+00 4.96701270e-01 3.04522645e-02 9.05507505e-01
-1.96560964e-01 9.76093292e-01 -1.19208002e+00 -8.38195682e-01
-1.49408817e-01 -5.28158128e-01 -7.16475248e-01 -6.01222634e-01
6.65554255e-02 -1.03504151e-01 -6.42941535e-01 9.15776193e-01
7.21775651e-01 1.64489836e-01 1.11159718e+00 -1.10119689e+00
-1.70514989e+00 9.48732793e-01 -1.82045475e-01 4.51511383e-01
-1.38520943e-02 3.08408082e-01 6.45003080e-01 -9.06010509e-01
7.54062414e-01 1.51172113e+00 4.19874161e-01 8.99703562e-01
-8.96825910e-01 -1.01567054e+00 1.32036984e-01 2.29079425e-01
-1.24797428e+00 -1.87266558e-01 8.75154376e-01 -4.49566185e-01
1.14984822e+00 8.39989036e-02 4.00349572e-02 1.89530742e+00
8.42258155e-01 5.66827059e-01 9.85073388e-01 -1.99926451e-01
1.74762830e-01 -1.87531933e-01 1.92190677e-01 3.85279864e-01
2.30710909e-01 8.12943131e-02 2.30196044e-02 -1.97673693e-01
8.12186360e-01 1.27698168e-01 -9.80100185e-02 9.55436349e-01
-9.50831950e-01 1.10562289e+00 5.91854453e-01 4.49208409e-01
-6.36090040e-01 2.01531917e-01 2.12485656e-01 3.49194497e-01
7.82104492e-01 7.96595812e-01 -2.39063472e-01 3.59570608e-02
-9.12705541e-01 2.68931329e-01 7.76001036e-01 1.14016640e+00
8.66577178e-02 4.34959859e-01 -9.79054213e-01 8.73452842e-01
2.84095585e-01 5.34659028e-01 8.53273988e-01 -7.87446022e-01
2.66081125e-01 3.95462930e-01 -4.43394661e-01 -6.20287240e-01
-5.82128298e-03 -8.35226476e-01 -1.28236556e+00 4.30742949e-02
-1.93113983e-01 -5.04999459e-01 -1.59248686e+00 1.56918132e+00
-4.44523007e-01 2.51859307e-01 3.79395604e-01 6.97549939e-01
1.53540969e+00 9.63362157e-01 8.46718907e-01 8.55709761e-02
1.23657238e+00 -1.26223195e+00 -7.85883367e-01 -3.23217690e-01
2.44435713e-01 -7.16689825e-01 8.87391627e-01 5.07039465e-02
-1.16159892e+00 -7.55588114e-01 -8.17784011e-01 -4.10281807e-01
-6.78824842e-01 5.77837378e-02 4.92302269e-01 2.15676799e-01
-9.76608396e-01 3.51623029e-01 -4.82062697e-01 9.23284963e-02
7.36121535e-01 2.26753876e-01 -5.46608381e-02 -2.39389371e-02
-1.54392123e+00 6.86637521e-01 3.62130821e-01 -9.76247489e-02
-1.11711264e+00 -7.40105867e-01 -5.33211827e-01 3.71271908e-01
-1.25938952e-01 -1.29541206e+00 1.02035666e+00 -8.09934318e-01
-1.80041766e+00 8.36813331e-01 7.20361248e-02 -4.56916451e-01
4.45156366e-01 -2.29090929e-01 -6.83481812e-01 -2.70135850e-01
2.04897702e-01 6.58595741e-01 7.37572432e-01 -8.97762299e-01
7.50935310e-03 2.04780381e-02 -2.97329035e-02 -9.16171819e-02
-1.11145228e-01 1.87731087e-01 -4.77762878e-01 -1.18030858e+00
-7.44578123e-01 -6.98775649e-01 -4.33572978e-01 -7.89864242e-01
-7.69761920e-01 -6.69057131e-01 6.05684936e-01 -8.46376956e-01
1.34287953e+00 -1.77794242e+00 2.07495764e-01 -1.37504235e-01
2.64289916e-01 3.86147469e-01 -3.93173367e-01 -7.94593059e-03
-1.25155061e-01 5.69201052e-01 -9.89522040e-02 -3.83555710e-01
-1.40325770e-01 9.79612861e-03 -3.27793866e-01 -1.22847125e-01
6.09815121e-01 1.58945251e+00 -8.18027079e-01 -4.05733049e-01
-2.27725953e-01 6.03365123e-01 -5.86521745e-01 7.87281632e-01
-3.43330532e-01 3.49790573e-01 -8.47083449e-01 7.08255172e-01
6.86947584e-01 -4.46203381e-01 -2.70272195e-01 1.14980787e-01
1.29450813e-01 1.22818813e-01 -1.53228387e-01 1.47464108e+00
-6.54479563e-01 6.13954425e-01 -2.51901239e-01 -9.15246546e-01
1.04381132e+00 1.30676523e-01 -1.40943453e-01 -6.24372423e-01
3.71092498e-01 -6.94489479e-03 7.02045634e-02 -3.47098827e-01
2.59176284e-01 -9.80076790e-02 -3.94427836e-01 1.10122807e-01
4.88316178e-01 2.73446381e-01 -1.84522241e-01 2.76883155e-01
1.23601651e+00 6.75540268e-02 7.39220530e-02 -4.20748651e-01
4.66632485e-01 -2.35736161e-01 2.72900999e-01 1.02000499e+00
1.00242689e-01 6.10721052e-01 5.10782540e-01 -2.53427953e-01
-9.29841757e-01 -7.84453630e-01 -2.00041197e-02 1.22092390e+00
-4.93911088e-01 -3.73558737e-02 -8.92212987e-01 -1.02937961e+00
-5.73266670e-02 1.19563222e+00 -8.95660281e-01 -6.28819227e-01
-4.14914072e-01 -8.94149899e-01 7.45993912e-01 6.15546286e-01
4.67019528e-01 -1.57416737e+00 2.77305424e-01 2.68211961e-01
3.41938913e-01 -8.22657585e-01 -5.56944251e-01 1.36634499e-01
-5.79779387e-01 -7.00028181e-01 -1.39303851e+00 -8.66896033e-01
3.99318814e-01 -5.93917608e-01 1.38197386e+00 -1.78259000e-01
1.28580675e-01 1.45259798e-02 4.14739251e-02 -6.55898750e-01
-6.67009294e-01 5.89456260e-01 -5.01540780e-01 -3.92711282e-01
5.13996840e-01 -3.67453277e-01 -5.75954080e-01 -3.83166254e-01
-6.26177788e-01 -6.27690181e-02 9.87100720e-01 1.24487507e+00
5.06100595e-01 -5.79735518e-01 1.21499944e+00 -1.29674792e+00
1.24874103e+00 -1.02431750e+00 -3.59211564e-01 2.24092007e-01
-9.71129179e-01 2.70397961e-02 9.44109201e-01 -6.14162087e-01
-1.25411546e+00 -7.62153625e-01 -6.00813806e-01 -7.69382536e-01
-2.51422524e-01 8.23022902e-01 -2.95343190e-01 1.59138232e-01
4.51418221e-01 1.41774565e-01 -3.55338544e-01 -6.88103676e-01
4.78079110e-01 9.50995028e-01 8.28964829e-01 -3.81416261e-01
4.33044791e-01 -2.91921318e-01 -7.44283125e-02 -4.02694494e-01
-1.00520492e+00 1.68243915e-01 -1.66536108e-01 2.77062297e-01
1.08872235e+00 -1.08719683e+00 -5.13672113e-01 4.21699375e-01
-1.53536844e+00 -2.49604449e-01 6.27279133e-02 2.83525169e-01
-2.73194879e-01 -2.89099395e-01 -8.27091217e-01 -3.22451323e-01
-1.17592478e+00 -1.06929195e+00 8.46039355e-01 4.79301006e-01
-1.25499770e-01 -1.23003924e+00 1.65774882e-01 2.35439092e-01
7.02789366e-01 2.99541771e-01 9.95551109e-01 -1.26951993e+00
-3.69323879e-01 -3.08964670e-01 -4.71672833e-01 2.96801031e-01
-6.65919110e-02 2.93675661e-02 -1.00321877e+00 1.53943263e-02
-2.49230862e-01 -5.04564457e-02 1.15681481e+00 6.40273631e-01
2.02291322e+00 -5.32195091e-01 -4.45499420e-01 9.30660963e-01
1.16945016e+00 4.00347739e-01 8.89891982e-01 3.75333935e-01
1.00528514e+00 -1.71286792e-01 -8.21579173e-02 7.33930320e-02
1.80356771e-01 3.05036604e-01 2.09468231e-02 -2.50215083e-01
-4.34466511e-01 -2.77888656e-01 6.08828664e-02 9.16107476e-01
-3.08343619e-01 -8.67051959e-01 -8.27766716e-01 4.58228707e-01
-1.50770330e+00 -9.53759432e-01 6.07785806e-02 1.41467285e+00
1.05785632e+00 3.91038567e-01 -2.22676456e-01 -7.41218686e-01
6.19372308e-01 -3.84511687e-02 -5.95799685e-01 -8.59623075e-01
-2.89751410e-01 7.44307876e-01 5.16405880e-01 4.28850681e-01
-1.11401391e+00 1.22347665e+00 7.20361614e+00 1.09028292e+00
-1.28562868e+00 4.92514163e-01 1.05682933e+00 3.97660658e-02
-5.63520789e-01 -3.76679629e-01 -7.47009575e-01 9.87699926e-01
1.71254468e+00 -6.88336790e-01 1.67136014e-01 9.50647652e-01
-1.43819660e-01 8.20427060e-01 -7.98972964e-01 6.07020915e-01
1.48894601e-02 -1.52502763e+00 5.96686900e-01 -4.63166162e-02
8.99720490e-01 2.67827153e-01 1.51849970e-01 1.07570052e+00
7.67864347e-01 -1.56184185e+00 1.83573619e-01 9.41726923e-01
1.07249117e+00 -9.76449966e-01 8.70318294e-01 1.61140069e-01
-2.70731866e-01 1.18388779e-01 -4.26881999e-01 2.67904162e-01
-1.44767836e-01 6.09587491e-01 -6.90544844e-01 5.86578131e-01
3.65554363e-01 8.16346228e-01 -5.88464618e-01 6.56805456e-01
-3.63442630e-01 1.12439096e+00 1.84300497e-01 -3.98308128e-01
4.15072650e-01 1.92867480e-02 5.28971553e-01 1.73669839e+00
3.62797678e-01 1.28992766e-01 -2.31861264e-01 1.34167457e+00
-1.01070046e+00 -5.32636121e-02 -5.22987425e-01 -4.29623604e-01
3.63076150e-01 1.10854268e+00 -1.94006234e-01 -7.68668234e-01
-4.07489717e-01 1.24348712e+00 3.98998320e-01 6.52921081e-01
-1.02822292e+00 -8.77050459e-01 4.81258333e-01 -2.49962583e-01
2.22421244e-01 3.31312180e-01 -4.73477900e-01 -1.13542950e+00
-5.72738528e-01 -7.36120820e-01 3.02496076e-01 -9.52457428e-01
-1.70784819e+00 8.99495721e-01 -2.96527565e-01 -6.66742086e-01
-4.82611358e-01 -5.36909640e-01 -1.16374576e+00 1.47977388e+00
-1.38852489e+00 -1.42870593e+00 -9.73525047e-02 2.71196395e-01
7.85225511e-01 -7.22167730e-01 1.09672618e+00 4.32703078e-01
-8.16890836e-01 1.07491994e+00 4.05460566e-01 2.90307313e-01
7.77058959e-01 -1.33647239e+00 6.77994967e-01 6.21698976e-01
-3.06806415e-01 9.06284630e-01 3.17593127e-01 -8.99214149e-01
-1.20822489e+00 -1.67146945e+00 1.07047892e+00 -4.37606305e-01
6.37263238e-01 -2.26785079e-01 -1.15680587e+00 8.87527168e-01
7.27562606e-01 -2.33252898e-01 6.98345900e-01 9.08525512e-02
-3.00488621e-01 3.68975103e-01 -1.04675031e+00 5.52045345e-01
9.68584657e-01 -4.65500325e-01 -8.39754045e-01 3.09495807e-01
1.21082604e+00 -3.04364860e-01 -1.25512969e+00 3.29830974e-01
2.91190982e-01 1.48691803e-01 1.12158680e+00 -1.25417089e+00
1.30446160e+00 3.25568050e-01 3.71221483e-01 -1.65580392e+00
-1.16343820e+00 -6.32148862e-01 -3.04323167e-01 1.43287718e+00
5.87899566e-01 -5.31944215e-01 3.31224740e-01 3.74342620e-01
-5.01827121e-01 -8.88130307e-01 -7.14612961e-01 -5.69271028e-01
8.92821074e-01 -2.89676599e-02 7.29403079e-01 1.27622664e+00
-4.67856407e-01 7.61846542e-01 -2.43067339e-01 -1.48811162e-01
5.35795689e-01 5.65871559e-02 3.74590218e-01 -1.07545352e+00
-2.09313557e-01 -9.25456285e-01 -1.00600965e-01 -7.33229101e-01
7.48998344e-01 -1.17412210e+00 -3.08764488e-01 -1.58956516e+00
3.08479935e-01 -8.94794315e-02 -8.65009964e-01 5.04479825e-01
-8.25905919e-01 1.64119095e-01 -2.99512386e-01 8.56498405e-02
-4.33560431e-01 7.28551149e-01 1.33779061e+00 -4.82567728e-01
1.29496604e-01 -1.81546230e-02 -1.16407502e+00 2.95713216e-01
7.53809392e-01 -4.95410025e-01 -7.00136647e-02 -4.00987625e-01
-2.41274893e-01 -7.36476183e-02 2.82621443e-01 -6.56355500e-01
4.42470573e-02 -4.07748669e-01 8.35479617e-01 -3.85097802e-01
-1.67957678e-01 -1.89438269e-01 6.62203133e-02 2.82894462e-01
-7.97861159e-01 -7.52181485e-02 4.67854053e-01 5.45459151e-01
-4.07986045e-02 -2.07682312e-01 5.84340930e-01 -1.19146153e-01
-4.14170265e-01 3.90584886e-01 -3.72772485e-01 2.23888367e-01
7.61708498e-01 5.63002646e-01 -6.11388683e-01 -2.69972980e-01
-6.75850332e-01 1.82612792e-01 1.04416423e-01 7.20488608e-01
5.37959695e-01 -1.42014778e+00 -1.11592722e+00 1.93879381e-01
-2.05915384e-02 -8.33292529e-02 1.12702332e-01 3.40506174e-02
-2.41805330e-01 6.23912930e-01 -2.56619006e-01 1.33585874e-02
-7.54125476e-01 6.70267701e-01 2.62520522e-01 -4.71485943e-01
-2.10593626e-01 1.04084909e+00 3.47681642e-01 -1.55061170e-01
-3.94484811e-02 -8.54542255e-02 -4.23034161e-01 -4.40094620e-01
3.49055737e-01 2.17743874e-01 -2.27043405e-02 -2.54322529e-01
-2.96393842e-01 9.97608975e-02 -3.10832083e-01 1.32506400e-01
1.45795143e+00 4.89949673e-01 1.64839346e-02 1.38594761e-01
1.49723291e+00 -1.05697505e-01 -6.00122392e-01 -6.22459278e-02
-3.96209389e-01 1.82465807e-01 6.71909928e-01 -1.32359958e+00
-1.23154974e+00 8.77531767e-01 3.42312157e-01 -1.99461449e-02
1.01905715e+00 7.07125142e-02 6.24830842e-01 1.10357240e-01
-3.90196681e-01 -6.24717116e-01 -1.75098389e-01 6.64580643e-01
1.24011314e+00 -1.06236434e+00 -9.17862654e-02 1.27901584e-01
-5.03864408e-01 8.87664676e-01 8.11420858e-01 -4.34499472e-01
7.45906830e-01 2.81774133e-01 3.08869015e-02 -2.51215369e-01
-1.01037610e+00 3.06368649e-01 7.25992918e-01 4.17342782e-01
8.95151854e-01 1.72676980e-01 -6.02423012e-01 1.33377314e+00
-2.18097627e-01 4.57172185e-01 2.07707182e-01 5.36113441e-01
9.41250250e-02 -9.67087507e-01 -4.92575066e-03 8.41396451e-01
-1.11558938e+00 -4.96334761e-01 -5.89776337e-01 4.88121033e-01
-2.03109995e-01 4.07852769e-01 2.20518589e-01 -3.98449004e-01
2.75449485e-01 3.15674722e-01 1.13836378e-01 -4.70636994e-01
-9.25928831e-01 7.85371363e-02 -9.71993059e-02 -7.84031078e-02
4.85330820e-02 -1.25554651e-01 -8.62551451e-01 -4.10772741e-01
5.21636335e-03 3.16310763e-01 5.99461854e-01 7.20056534e-01
1.05861139e+00 1.39635456e+00 4.97985512e-01 -4.85599220e-01
-4.39716548e-01 -1.66387510e+00 9.83319655e-02 4.65157509e-01
1.78643122e-01 -3.42294186e-01 -3.72531891e-01 1.44395858e-01] | [10.262409210205078, 8.15455436706543] |
02c815a8-95f5-4edb-9eb8-94373c4c79cc | incorporating-orientations-into-end-to-end | 2103.05846 | null | https://arxiv.org/abs/2103.05846v1 | https://arxiv.org/pdf/2103.05846v1.pdf | Incorporating Orientations into End-to-end Driving Model for Steering Control | In this paper, we present a novel end-to-end deep neural network model for autonomous driving that takes monocular image sequence as input, and directly generates the steering control angle. Firstly, we model the end-to-end driving problem as a local path planning process. Inspired by the environmental representation in the classical planning algorithms(i.e. the beam curvature method), pixel-wise orientations are fed into the network to learn direction-aware features. Next, to handle the imbalanced distribution of steering values in training datasets, we propose an improvement on a cost-sensitive loss function named SteeringLoss2. Besides, we also present a new end-to-end driving dataset, which provides corresponding LiDAR and image sequences, as well as standard driving behaviors. Our dataset includes multiple driving scenarios, such as urban, country, and off-road. Numerous experiments are conducted on both public available LiVi-Set and our own dataset, and the results show that the model using our proposed methods can predict steering angle accurately. | ['Jianfeng Lu', 'Zhenbo Song', 'Peng Wan'] | 2021-03-10 | null | null | null | null | ['steering-control'] | ['computer-vision'] | [ 5.35104014e-02 -1.82156786e-02 -2.33323470e-01 -1.25526285e+00
-3.69262755e-01 -3.66324782e-01 5.27206421e-01 -6.36137962e-01
-5.86656570e-01 7.27651894e-01 6.11858629e-02 -5.59988856e-01
-7.19713373e-03 -1.09665537e+00 -8.92614067e-01 -5.56694627e-01
1.57179937e-01 3.20187062e-01 3.83628219e-01 -5.57204366e-01
3.92629772e-01 3.54442388e-01 -1.61528420e+00 -2.13255182e-01
1.07262528e+00 1.05817437e+00 5.40395081e-01 4.82423663e-01
2.74370879e-01 5.10211349e-01 1.68300658e-01 -1.38899028e-01
4.49421823e-01 1.04352348e-01 -3.36371571e-01 -1.46498412e-01
5.18154800e-01 -7.19929934e-01 -8.30707848e-01 8.36178422e-01
6.78941309e-01 1.12565987e-01 5.85728526e-01 -1.66376925e+00
-2.17635378e-01 1.62085906e-01 -3.93242031e-01 -2.03887790e-01
-1.43801132e-02 7.40727842e-01 6.59819961e-01 -9.01079476e-01
8.03910911e-01 1.31304228e+00 3.74188542e-01 1.67977780e-01
-6.04729235e-01 -8.49041700e-01 3.12285215e-01 6.05020106e-01
-1.17506564e+00 -4.58189011e-01 9.76784170e-01 -4.13635314e-01
3.79764766e-01 -3.40749949e-01 3.86636108e-01 1.04651356e+00
6.01470053e-01 1.08772063e+00 9.24265742e-01 2.15188161e-01
4.49720360e-02 -9.61168706e-02 2.43946277e-02 6.86750412e-01
3.21110412e-02 7.80204177e-01 -2.82626510e-01 5.05289137e-01
3.24364066e-01 -9.16360915e-02 -7.58024231e-02 -1.00133502e+00
-1.28328466e+00 8.82841527e-01 8.92854452e-01 -6.91812217e-01
-2.69081652e-01 3.63975585e-01 3.80359918e-01 -2.16736240e-04
1.47180095e-01 -2.72974700e-01 -5.06682158e-01 -1.26007453e-01
-3.56715739e-01 8.68688166e-01 1.94079623e-01 1.26298082e+00
1.11119163e+00 -8.63419026e-02 -1.43446505e-01 7.60016680e-01
6.23816729e-01 7.90880084e-01 1.46539226e-01 -1.18418145e+00
9.46524918e-01 3.19617748e-01 3.90607089e-01 -1.14729238e+00
-8.70814860e-01 -4.17908460e-01 -7.56440043e-01 4.00401115e-01
2.13901773e-01 -4.07778114e-01 -1.03438425e+00 2.05190802e+00
4.06052738e-01 -3.13166045e-02 3.93103957e-02 1.21211231e+00
9.54548597e-01 4.90980357e-01 -1.99041054e-01 2.66921341e-01
8.21426988e-01 -1.29116464e+00 -6.23603463e-01 -5.32656133e-01
5.84460616e-01 -3.40458363e-01 1.04986906e+00 3.07249963e-01
-7.50002742e-01 -8.43820572e-01 -1.17002380e+00 -3.54395539e-01
-5.39178967e-01 2.82390743e-01 4.56459701e-01 1.92176446e-01
-9.03081536e-01 3.22638035e-01 -5.11152148e-01 -1.68451637e-01
5.93902528e-01 2.05113981e-02 -2.76432991e-01 -5.30475855e-01
-1.36979377e+00 8.21888983e-01 2.85095185e-01 2.45956719e-01
-1.05890441e+00 -5.24931669e-01 -1.17708409e+00 -3.60263973e-01
4.12502408e-01 -6.87639356e-01 1.24733329e+00 -1.39688879e-01
-1.64059711e+00 5.04178584e-01 -4.45467174e-01 -3.93110782e-01
8.04788709e-01 -3.78238827e-01 -2.31220648e-01 -4.66515541e-01
3.90784949e-01 1.27375495e+00 5.21927238e-01 -1.35061443e+00
-1.02275610e+00 -5.10211051e-01 1.90080568e-01 4.37659115e-01
3.13458651e-01 -6.40124261e-01 -5.35705209e-01 -1.07420841e-02
2.93560714e-01 -1.21689606e+00 -4.87811983e-01 2.55948365e-01
-7.04238176e-01 -2.23284721e-01 9.12700236e-01 -3.37337673e-01
7.87064850e-01 -2.08141303e+00 -3.50678414e-02 8.04709569e-02
2.42230035e-02 -1.31632105e-01 -3.27528030e-01 1.78523913e-01
1.82425424e-01 -3.82963598e-01 -4.41572696e-01 -1.75344616e-01
2.03310013e-01 3.73808891e-01 -4.49085087e-01 4.60985243e-01
1.49591669e-01 1.04727721e+00 -1.02548599e+00 -2.96495229e-01
3.78782332e-01 2.35567331e-01 -4.46294665e-01 1.31084740e-01
-8.97246674e-02 4.63295162e-01 -4.63369548e-01 4.71873343e-01
1.20681369e+00 4.22455728e-01 -4.19491827e-01 -2.49091890e-02
-4.84317452e-01 4.76080537e-01 -1.09267688e+00 1.91127479e+00
-5.26256323e-01 7.82899618e-01 -2.09530383e-01 -6.81028128e-01
1.21810532e+00 -3.41621041e-01 1.85178757e-01 -1.06071138e+00
4.09752280e-02 2.88050056e-01 -6.01063892e-02 -4.02900487e-01
7.60034323e-01 3.42277795e-01 -3.29883397e-01 1.31146014e-02
-4.18201774e-01 -4.68320191e-01 1.53932229e-01 -1.63341418e-01
7.33481705e-01 6.93872571e-01 -1.53707027e-01 -1.20663896e-01
7.11468041e-01 2.58264303e-01 8.25075746e-01 4.59111512e-01
-5.90690076e-01 6.81626916e-01 3.11525792e-01 -7.67303824e-01
-9.61505473e-01 -1.08782816e+00 -2.17454940e-01 7.78881788e-01
7.80166328e-01 -1.39144743e-02 -4.35928404e-01 -6.14326894e-01
2.81215131e-01 8.87295365e-01 -4.07424122e-01 -2.87706256e-01
-7.12164760e-01 -5.36359191e-01 2.56214947e-01 6.99337363e-01
9.71944094e-01 -1.00399470e+00 -7.33061194e-01 2.39504188e-01
-3.96645963e-01 -1.24361145e+00 -4.45528060e-01 4.43718098e-02
-4.65236515e-01 -9.65211451e-01 -2.66278327e-01 -6.71791196e-01
2.94142663e-01 6.01208806e-01 9.25267458e-01 -5.13563037e-01
6.90927804e-02 -3.30141872e-01 5.71322702e-02 -7.67525673e-01
7.48889372e-02 3.40833366e-01 1.65891871e-02 -1.66567892e-01
5.26835859e-01 -4.37442839e-01 -9.72065091e-01 7.89055765e-01
-2.00371295e-01 3.18787783e-01 6.79696023e-01 5.86197674e-01
6.19855344e-01 -1.28096327e-01 6.72200441e-01 -5.76322436e-01
3.80017340e-01 -5.69594920e-01 -8.15595210e-01 -5.15002668e-01
-7.23951221e-01 -6.33026287e-02 5.92015743e-01 2.24242106e-01
-1.08715355e+00 4.11702961e-01 -2.84952670e-01 -1.77638799e-01
-4.67865050e-01 4.10071969e-01 -4.36669469e-01 4.58416492e-02
4.35142517e-01 2.72978753e-01 -1.45596415e-02 -6.45720512e-02
6.29453063e-01 6.48780942e-01 8.60449672e-01 -8.31561461e-02
1.01249123e+00 6.22129440e-01 2.96552986e-01 -4.50977027e-01
-5.98311663e-01 -3.16067815e-01 -7.59888887e-01 -2.91453779e-01
7.37303734e-01 -1.24617648e+00 -6.24949813e-01 8.12645912e-01
-1.06453121e+00 -5.33543646e-01 1.05209604e-01 5.37503660e-01
-1.02338910e+00 8.00241306e-02 3.30977961e-02 -6.18251383e-01
-4.17052507e-02 -1.43451762e+00 1.21666658e+00 2.79837549e-01
3.05913031e-01 -5.49867928e-01 1.81326225e-01 2.29670927e-01
3.57262671e-01 4.01889473e-01 6.53916895e-01 5.74835464e-02
-8.64353299e-01 -1.13550074e-01 -5.40514231e-01 6.15245290e-02
-1.79270491e-01 -1.57116815e-01 -8.68415236e-01 -8.35927203e-02
-5.21044493e-01 -4.02491510e-01 1.11687720e+00 5.31428993e-01
1.35488379e+00 1.17574386e-01 -5.22948980e-01 9.83783007e-01
1.29109597e+00 3.89429390e-01 7.60389388e-01 5.41410685e-01
8.36406469e-01 8.31011534e-01 1.24440360e+00 1.89699173e-01
1.13783050e+00 8.56258988e-01 9.56042230e-01 1.79837644e-02
1.04390018e-01 -5.40047050e-01 3.76809776e-01 3.31783801e-01
3.19162875e-01 -2.92464763e-01 -9.11955297e-01 7.06422746e-01
-2.05359316e+00 -7.77736068e-01 -4.03794348e-01 1.94109368e+00
2.97063649e-01 4.49996293e-01 8.82306844e-02 -4.15733419e-02
3.60730529e-01 4.17497933e-01 -1.03126383e+00 -4.04483348e-01
3.05674952e-02 -3.96483213e-01 9.60729301e-01 6.54109061e-01
-1.39457178e+00 1.07286453e+00 5.68495512e+00 5.26695490e-01
-1.35221887e+00 -2.37905055e-01 4.73721176e-01 -9.99264792e-02
-3.13266397e-01 -1.44925043e-01 -9.50059891e-01 6.24849439e-01
7.24352837e-01 3.96438688e-02 2.93652087e-01 1.12185824e+00
7.06495404e-01 -1.32496431e-01 -8.45329702e-01 7.89273202e-01
-2.95238137e-01 -1.01377034e+00 -3.77479196e-01 1.09138370e-01
6.18956923e-01 6.32395029e-01 1.75274670e-01 5.03456652e-01
4.55080241e-01 -1.01941764e+00 8.63530278e-01 3.99201095e-01
8.54028583e-01 -1.02379084e+00 8.86979163e-01 5.85937083e-01
-1.28828955e+00 -3.05588096e-01 -4.88556892e-01 -1.47404686e-01
5.54509223e-01 6.15521073e-01 -7.67125666e-01 7.35500216e-01
6.29066050e-01 1.15513229e+00 -3.55555534e-01 1.09950876e+00
-4.79899049e-01 2.13650480e-01 -2.71223009e-01 -9.82105359e-02
6.47417784e-01 -4.59757298e-01 4.75479007e-01 8.05685043e-01
3.19385588e-01 -2.59805977e-01 1.61123857e-01 8.66773725e-01
1.33152112e-01 -2.26026684e-01 -1.08864713e+00 6.84952497e-01
4.31127429e-01 1.29708946e+00 1.30725242e-02 2.79549826e-02
-3.11620027e-01 5.28356552e-01 2.65488088e-01 4.40462947e-01
-9.48332727e-01 -7.63285875e-01 1.02008200e+00 5.26013635e-02
3.69034022e-01 -4.19915140e-01 -5.06995976e-01 -7.03424811e-01
2.29533672e-01 -2.53233641e-01 -3.83043617e-01 -1.06790090e+00
-7.97133207e-01 3.71556938e-01 1.04536720e-01 -1.37983763e+00
-3.74966323e-01 -6.77017629e-01 -8.43914986e-01 9.24668074e-01
-2.34013128e+00 -1.04798472e+00 -9.09996152e-01 3.44359517e-01
5.98552644e-01 -1.97604835e-01 3.51673625e-02 4.37919497e-01
-4.60122734e-01 4.05647486e-01 4.82689552e-02 -7.71986395e-02
8.54990602e-01 -1.09423590e+00 8.49803925e-01 7.15215921e-01
-5.68686843e-01 1.25485525e-01 6.99815452e-01 -3.41203511e-01
-1.25005519e+00 -1.69799817e+00 9.05289590e-01 -3.59693229e-01
2.13786379e-01 -2.59738028e-01 -3.88696313e-01 6.32296324e-01
4.98632062e-03 4.25120741e-02 -1.76319495e-01 -3.27148080e-01
1.48927674e-01 -4.61441636e-01 -1.06589818e+00 8.15006673e-01
1.64827991e+00 -1.48402685e-02 -2.61899382e-01 2.29059130e-01
7.67481506e-01 -7.40008414e-01 -8.90954062e-02 7.90609658e-01
6.38118863e-01 -1.06027532e+00 9.65383768e-01 -4.57592994e-01
8.12893152e-01 -4.74278033e-01 -3.43005329e-01 -1.57949531e+00
-4.74952370e-01 -3.96979392e-01 3.19739223e-01 6.92617059e-01
4.23541993e-01 -6.68065429e-01 9.17331338e-01 1.28530651e-01
-7.11011589e-01 -1.15864801e+00 -9.34610903e-01 -5.77141345e-01
2.12513000e-01 -7.22163439e-01 9.41984117e-01 3.15040171e-01
-4.84084427e-01 3.50448489e-01 -3.80962729e-01 1.83995038e-01
6.32540345e-01 2.52973050e-01 1.35589981e+00 -1.10039687e+00
4.33286607e-01 -3.62952650e-01 -5.11024177e-01 -1.70232368e+00
3.26195776e-01 -8.26095700e-01 7.18964458e-01 -1.66578734e+00
-2.85849512e-01 -5.71691930e-01 -3.01501751e-02 2.34200880e-01
-1.22175373e-01 1.69194993e-02 -1.99838988e-02 -2.24830300e-01
-2.77365625e-01 9.95281339e-01 1.60198510e+00 -2.59848535e-01
-9.50580686e-02 1.72703430e-01 -5.60763776e-01 7.54112840e-01
8.85996103e-01 -1.73746079e-01 -8.03368926e-01 -5.96659899e-01
2.19525665e-01 -2.71855369e-02 4.35321957e-01 -9.80017424e-01
1.25182763e-01 -5.95054269e-01 2.34520882e-01 -1.57270312e+00
3.61478865e-01 -6.99488997e-01 -4.25007492e-01 3.99614662e-01
-2.41549909e-01 1.13821968e-01 -8.86026174e-02 6.34585738e-01
-8.54635313e-02 1.92153960e-01 7.91900933e-01 2.39141062e-01
-1.19641483e+00 8.14649463e-01 -1.63198650e-01 1.19138271e-01
1.07044852e+00 -2.46618927e-01 -4.85235900e-01 -5.17597318e-01
-1.77616701e-01 1.06551337e+00 3.80403638e-01 8.31707835e-01
8.35789025e-01 -1.75790775e+00 -8.34046900e-01 4.72654283e-01
5.82172573e-01 6.01004660e-01 1.73540920e-01 6.53909266e-01
-4.61327016e-01 7.60267854e-01 -4.43375707e-01 -9.31086004e-01
-6.60908639e-01 3.13953429e-01 2.86700249e-01 1.78947777e-01
-5.13222277e-01 6.37405753e-01 3.84380698e-01 -1.27289951e+00
2.32893024e-02 -4.53774095e-01 -2.28722647e-01 -3.14918309e-01
1.10363975e-01 3.53330582e-01 -1.17107600e-01 -6.67502165e-01
-5.28025270e-01 7.78318167e-01 1.60074711e-01 -2.19795495e-01
1.15781713e+00 -5.35314798e-01 4.67815131e-01 3.66760373e-01
1.32774627e+00 -3.57899010e-01 -1.73164928e+00 -1.56182081e-01
-4.20701832e-01 -6.53528810e-01 1.73778683e-01 -7.07473814e-01
-1.13413382e+00 1.19316947e+00 7.24737287e-01 -4.32410121e-01
5.99213243e-01 -4.35233772e-01 1.16776037e+00 5.88286638e-01
5.34079731e-01 -1.27219379e+00 -2.86163449e-01 1.09622300e+00
9.72578347e-01 -1.73402584e+00 -2.60333627e-01 -3.45317781e-01
-6.96528435e-01 8.07670951e-01 1.03931630e+00 -7.80985355e-02
6.79752707e-01 -1.04414627e-01 3.37566495e-01 5.84122948e-02
-7.32205153e-01 -2.45341808e-01 7.39577562e-02 8.14378500e-01
-7.28032738e-02 2.24295199e-01 -2.93677121e-01 3.65067482e-01
-6.50139272e-01 3.32008630e-01 5.82418144e-01 7.76939929e-01
-6.31140828e-01 -7.13667214e-01 7.92347491e-02 4.18100715e-01
5.06471515e-01 3.61732692e-01 -7.92769268e-02 7.39056826e-01
3.79429936e-01 9.85356212e-01 2.08910167e-01 -7.26774454e-01
6.38924479e-01 -1.62042931e-01 -8.85510966e-02 -2.00232506e-01
3.40995580e-01 -6.42542779e-01 2.00061098e-01 -9.50179160e-01
-5.14759198e-02 -6.82167172e-01 -1.37292230e+00 -3.67456436e-01
1.44528896e-02 -5.03256142e-01 9.54013407e-01 9.02260423e-01
5.31790972e-01 5.35960197e-01 1.13519180e+00 -1.16718304e+00
-5.80579281e-01 -6.75391555e-01 -2.11306781e-01 2.91018546e-01
5.93942225e-01 -1.06759429e+00 -1.72665492e-01 -4.03101385e-01] | [8.042851448059082, -1.593914270401001] |
7fd4e16e-1e34-4629-b1c0-0c63268e569d | challenges-in-designing-games-with-a-purpose | null | null | https://aclanthology.org/2021.hcinlp-1.10 | https://aclanthology.org/2021.hcinlp-1.10.pdf | Challenges in Designing Games with a Purpose for Abusive Language Annotation | In this paper we discuss several challenges related to the development of a 3D game, whose goal is to raise awareness on cyberbullying while collecting linguistic annotation on offensive language. The game is meant to be used by teenagers, thus raising a number of issues that need to be tackled during development. For example, the game aesthetics should be appealing for players belonging to this age group, but at the same time all possible solutions should be implemented to meet privacy requirements. Also, the task of linguistic annotation should be possibly hidden, adopting so-called orthogonal game mechanics, without affecting the quality of collected data. While some of these challenges are being tackled in the game development, some others are discussed in this paper but still lack an ultimate solution. | ['Sara Tonelli', 'Federico Bonetti'] | null | null | null | null | eacl-hcinlp-2021-4 | ['abusive-language'] | ['natural-language-processing'] | [-1.55483603e-01 6.30759776e-01 1.06036499e-01 -1.73033684e-01
-2.33312264e-01 -4.95302260e-01 6.77905902e-02 5.58454633e-01
-5.17994046e-01 4.71464932e-01 2.26461813e-01 -4.53324849e-03
-2.23498657e-01 -7.83179045e-01 -1.59303367e-01 -2.27369457e-01
1.85677130e-02 1.81536272e-01 3.44739407e-01 -4.49843675e-01
3.43720615e-01 2.32528463e-01 -1.73908341e+00 2.43474599e-02
6.94245696e-01 5.27302980e-01 -1.14445277e-01 2.91860461e-01
-3.24200600e-01 6.20974243e-01 -6.76438451e-01 -1.12466133e+00
1.95777521e-01 -1.77153543e-01 -1.05308294e+00 5.85914329e-02
1.58802703e-01 -5.23332894e-01 6.16993487e-01 1.29697073e+00
5.06837010e-01 7.32255280e-02 1.90171257e-01 -9.46311057e-01
-1.56092495e-01 5.65482497e-01 -3.95766199e-01 8.01239279e-04
7.85520673e-01 -2.41483271e-01 8.60828638e-01 -4.20020163e-01
6.14664972e-01 9.98226881e-01 4.74157214e-01 6.65465593e-01
-1.04237819e+00 -7.56280005e-01 1.28681913e-01 -1.92606002e-01
-1.55187595e+00 -3.68889749e-01 9.17630672e-01 -6.62837327e-01
3.91406924e-01 4.61092800e-01 9.25766587e-01 9.48324740e-01
-3.61400455e-01 2.78962672e-01 9.33130026e-01 -5.79638302e-01
5.12416422e-01 5.70062816e-01 2.06164658e-01 3.47967952e-01
6.65194333e-01 -4.42784637e-01 -7.66263604e-01 -1.73081458e-01
3.25746864e-01 -7.66592324e-01 1.29798278e-01 -3.51101607e-01
-1.61431387e-01 9.72691298e-01 -6.65042341e-01 8.70330393e-01
-3.68375443e-02 -3.75579536e-01 9.53047693e-01 9.28872451e-02
9.01568353e-01 4.19816464e-01 -1.70177862e-01 -9.71513808e-01
-6.85031176e-01 3.83322269e-01 9.02160943e-01 6.38197303e-01
6.27176821e-01 -3.09284866e-01 3.85331690e-01 9.23431814e-01
5.55267394e-01 -1.93850487e-01 -6.57768399e-02 -6.81728244e-01
5.11394382e-01 8.80198479e-01 -3.34902674e-01 -1.32735598e+00
-4.31056172e-01 -1.22394353e-01 -3.04136038e-01 4.71355498e-01
6.97281599e-01 -3.16128075e-01 -1.47130698e-01 1.64500403e+00
6.98175490e-01 -1.05636127e-01 -3.83143693e-01 7.65333414e-01
9.61398363e-01 -4.84193023e-03 5.50690889e-01 -1.67244613e-01
1.55712020e+00 4.14301008e-02 -9.37954783e-01 -3.04159969e-01
7.96884835e-01 -8.70099127e-01 9.39343333e-01 5.81431329e-01
-1.26959980e+00 -2.26347908e-01 -7.18161643e-01 -2.10935265e-01
-4.61621165e-01 -3.46375555e-01 5.11483133e-01 1.71587515e+00
-9.18716550e-01 2.37181768e-01 -6.70428753e-01 -7.62926280e-01
3.39335836e-02 5.53748071e-01 -5.36932826e-01 3.43049049e-01
-1.38983083e+00 9.72838461e-01 5.69911838e-01 1.53459430e-01
-3.38313468e-02 -3.87533575e-01 -1.07369328e+00 -3.07896525e-01
5.50868273e-01 -1.84000321e-02 9.20573354e-01 -7.26968348e-01
-1.56341279e+00 1.51162291e+00 2.80061841e-01 8.13867897e-02
6.23457789e-01 -4.79124784e-01 -3.60367388e-01 -1.28397778e-01
3.53763074e-01 2.20504865e-01 2.39215136e-01 -9.39003885e-01
-7.37920582e-01 -4.06614244e-01 4.41058159e-01 3.76837254e-01
-6.06500089e-01 6.95502043e-01 -5.82942307e-01 -4.58890855e-01
-3.59221660e-02 -7.11781085e-01 -6.02274239e-02 -1.48071408e-01
-3.19058150e-01 -8.39998350e-02 5.40729403e-01 -9.26177800e-01
1.57528424e+00 -2.20544934e+00 -5.74359149e-02 4.87241268e-01
2.63216197e-01 5.58126450e-01 3.30112249e-01 6.06164396e-01
6.40145317e-02 6.03243470e-01 4.42022145e-01 -3.89682174e-01
6.17910624e-02 1.17585426e-02 2.82089144e-01 5.27687669e-01
-3.38943638e-02 4.23156291e-01 -6.94843590e-01 -5.30201674e-01
2.40738764e-01 1.14309475e-01 -6.57380044e-01 -7.05391392e-02
1.02817185e-01 2.23519862e-01 -5.68754315e-01 3.91720414e-01
8.14566374e-01 7.97567487e-01 2.32924491e-01 4.59049314e-01
-4.42972988e-01 3.96185637e-01 -1.41462409e+00 1.45206499e+00
-2.25567386e-01 1.96576446e-01 5.14239550e-01 -8.47658455e-01
1.00429893e+00 3.51070642e-01 7.38079667e-01 -6.14633560e-01
3.11284810e-01 1.85631067e-01 1.54450014e-01 -8.84564459e-01
7.84658730e-01 -2.36861363e-01 -3.18145603e-01 7.24049032e-01
-2.45569600e-03 4.55133952e-02 3.45307022e-01 -1.97770521e-01
5.66129088e-01 1.59622595e-01 2.32394829e-01 -3.01225781e-01
5.40552437e-01 -1.14329748e-01 6.69184089e-01 3.73754084e-01
-2.46616036e-01 3.42640221e-01 1.00500906e+00 -1.70333073e-01
-6.23200059e-01 -4.52225536e-01 1.02814615e-01 1.03528059e+00
-1.26001537e-01 -8.94799769e-01 -1.24502337e+00 -5.79218924e-01
-7.28111267e-01 6.38135314e-01 -4.54064220e-01 -1.74486428e-01
-2.55970448e-01 -3.92270654e-01 6.50777519e-01 -4.28482652e-01
2.15575039e-01 -7.10535109e-01 -1.06230938e+00 4.85559940e-01
-1.15780994e-01 -1.10772908e+00 -1.60848326e-03 -2.04120070e-01
-2.54499227e-01 -1.13747382e+00 -2.61575490e-01 -5.89716733e-01
4.13889736e-01 4.82257782e-03 6.82573795e-01 4.04345006e-01
1.42804310e-02 5.30862510e-01 -8.91687751e-01 -8.26873600e-01
-5.42037368e-01 4.03839439e-01 -8.58082622e-02 -3.62168998e-02
7.76324272e-01 -7.07656562e-01 3.34258750e-02 -3.34248953e-02
-8.71075034e-01 -4.03818972e-02 -1.31092414e-01 1.75348714e-01
7.15426654e-02 3.76721114e-01 1.64865687e-01 -1.00500178e+00
9.05780137e-01 -4.75394100e-01 -3.07589233e-01 -6.60956698e-03
-3.55257243e-02 -5.53819001e-01 3.31856996e-01 -4.42365021e-01
-7.04812825e-01 -1.13846987e-01 -8.48532557e-01 2.66460329e-01
-5.51357985e-01 4.89410490e-01 -5.87849081e-01 -2.21061260e-01
4.38969642e-01 -3.52378011e-01 2.98273444e-01 -5.39119065e-01
-4.62315083e-02 6.13272429e-01 -2.32840136e-01 -7.45481431e-01
6.88519180e-01 2.31812656e-01 -1.89540178e-01 -1.17089820e+00
-6.26230478e-01 -3.70496869e-01 -7.62081087e-01 -7.97020972e-01
9.02850270e-01 -5.53570151e-01 -8.17062795e-01 4.82140243e-01
-9.64354575e-01 2.71640811e-03 -1.91521615e-01 4.29965764e-01
-3.52330059e-01 4.06825483e-01 -2.01799601e-01 -1.28601861e+00
-2.40548342e-01 -1.03448701e+00 6.12208307e-01 4.87032160e-02
-8.71907294e-01 -1.09836197e+00 1.61186203e-01 7.23696709e-01
5.74854203e-02 4.86382514e-01 8.09968531e-01 -8.96417499e-01
2.05772504e-01 -1.34824932e-01 2.51870334e-01 3.57607007e-01
3.42935771e-01 9.00659040e-02 -8.11649442e-01 -5.62471524e-03
2.83631176e-01 -2.79817164e-01 -3.34213912e-01 -7.47931972e-02
6.20485008e-01 -2.33359739e-01 1.91752419e-01 -7.49279335e-02
1.14285278e+00 1.96124762e-01 5.73203683e-01 4.78888571e-01
4.49296921e-01 1.36365998e+00 9.88235891e-01 3.79942834e-01
5.58839738e-01 8.01331043e-01 3.95724148e-01 1.69407755e-01
2.04375714e-01 -2.49268174e-01 2.76144773e-01 7.67579198e-01
-1.67732701e-01 3.12234163e-01 -1.12517536e+00 3.92947793e-01
-1.70404887e+00 -9.34272051e-01 -6.45578682e-01 2.16622519e+00
5.42975605e-01 4.30384159e-01 6.30380809e-01 5.05307257e-01
5.89878559e-01 5.12367412e-02 1.76907063e-01 -1.06447756e+00
2.64204353e-01 5.72654605e-01 -6.73612282e-02 5.94989419e-01
-9.84781027e-01 8.90869319e-01 5.97597218e+00 7.31565177e-01
-1.22511792e+00 3.60789448e-01 4.97497499e-01 -2.34182607e-02
-3.62747908e-01 1.55740961e-01 -7.53206313e-01 4.50904757e-01
6.22051358e-01 -2.48812884e-01 1.75383031e-01 8.19975436e-01
5.50848424e-01 -5.00941336e-01 -4.93416816e-01 5.75488269e-01
3.01965736e-02 -6.82002962e-01 -6.24163628e-01 4.37643677e-01
-1.83331836e-02 -4.86776084e-01 -1.85538650e-01 2.16028765e-01
-7.21637830e-02 -7.73695469e-01 1.04335237e+00 1.11304671e-01
5.90687633e-01 -9.43495691e-01 6.62181973e-01 6.03797197e-01
-9.98969257e-01 3.33452880e-01 -1.81369439e-01 -7.05465078e-01
1.33576050e-01 4.98396933e-01 -6.12455368e-01 4.72309649e-01
8.52676392e-01 -4.79559824e-02 -6.19767308e-01 1.02019453e+00
-3.43052208e-01 5.40276885e-01 -3.55863810e-01 -4.57258284e-01
1.71450928e-01 -4.21745211e-01 4.35310066e-01 9.01110590e-01
3.32028478e-01 2.32359916e-01 -5.97759895e-02 6.76672876e-01
5.81681013e-01 8.37765157e-01 -8.77513230e-01 3.66392061e-02
3.78739178e-01 1.13762534e+00 -8.29543293e-01 6.51290774e-01
-6.86851680e-01 6.49145842e-01 4.04022574e-01 -3.23633879e-01
-6.23825192e-01 -1.10351935e-01 1.05568779e+00 8.60953212e-01
-1.58135936e-01 -2.04899728e-01 -4.62000549e-01 -6.86652958e-01
4.51994240e-02 -9.36104596e-01 4.00730520e-01 -3.02801013e-01
-8.44575763e-01 5.96495606e-02 1.07641347e-01 -1.15403831e+00
-3.32567602e-01 -4.54892844e-01 -5.29707551e-01 6.52521729e-01
-7.12607563e-01 -1.13119078e+00 2.94174820e-01 3.41373891e-01
1.13854613e-02 2.28693679e-01 9.53734875e-01 5.32052279e-01
-4.98902172e-01 4.95888889e-01 -5.35371244e-01 -2.21500639e-02
4.18806791e-01 -8.69024098e-01 1.13489397e-01 9.30009305e-01
9.00797769e-02 7.48439729e-01 7.96354353e-01 -7.74185777e-01
-9.31584120e-01 -6.81049883e-01 1.24855947e+00 -1.97603047e-01
7.36841917e-01 -8.91118050e-01 -8.19364369e-01 2.80633360e-01
8.36836174e-03 -6.27442122e-01 1.29828060e+00 7.47149110e-01
1.04939215e-01 2.45386824e-01 -1.37372100e+00 7.89718628e-01
9.69766200e-01 -5.66452801e-01 -5.65797687e-01 2.35520862e-02
2.99193680e-01 -5.43439507e-01 -9.21729028e-01 -1.29193231e-01
3.82540882e-01 -1.13927138e+00 5.86616397e-01 -2.77315527e-01
1.06109805e-01 -2.77299080e-02 3.75199407e-01 -8.64208877e-01
5.63519076e-03 -6.61908746e-01 5.06846726e-01 2.07046151e+00
3.26162279e-01 -2.84031719e-01 1.17797649e+00 1.46683526e+00
3.67351621e-01 -4.18259948e-01 -1.33702767e+00 -4.49515551e-01
3.37143987e-01 -1.29958355e+00 3.45805287e-01 1.09599364e+00
6.07669711e-01 1.32181168e-01 -6.62168741e-01 1.92626774e-01
3.14873248e-01 -8.10388565e-01 8.84631455e-01 -1.31074667e+00
1.84741184e-01 -2.97663093e-01 -7.13959992e-01 -1.13585062e-01
9.87096280e-02 -3.60949963e-01 -3.74203801e-01 -1.26999283e+00
-1.94592446e-01 -2.25748181e-01 3.89943361e-01 3.04196745e-01
6.29325435e-02 2.92466998e-01 2.02187508e-01 -5.65545380e-01
-4.11521107e-01 -3.31305410e-03 7.94224143e-01 4.53377962e-01
-2.28153974e-01 2.89291799e-01 -1.24078989e+00 9.38932240e-01
9.41480041e-01 -5.09489298e-01 -5.49199581e-01 -2.34141365e-01
6.91981196e-01 -3.03488582e-01 1.27542168e-01 -8.91727030e-01
2.70694077e-01 -4.42046434e-01 -3.60748351e-01 -3.56561333e-01
3.86496663e-01 -9.46029246e-01 5.05828142e-01 3.39806408e-01
7.63803348e-02 -3.14798564e-01 2.78698772e-01 -2.79248238e-01
-2.03346804e-01 -9.33252990e-01 5.04140854e-01 -2.01073602e-01
-5.46343803e-01 6.14205524e-02 -8.18098783e-01 1.74161449e-01
1.48567116e+00 -7.69383729e-01 2.38723129e-01 -5.55122495e-01
-7.92492926e-01 1.59643888e-01 6.81819499e-01 5.94236791e-01
2.07913384e-01 -9.94652867e-01 -4.12542313e-01 6.66052550e-02
1.49144948e-01 -1.37579367e-01 4.54044104e-01 5.19121051e-01
-7.07530797e-01 -7.91796371e-02 -3.08142453e-01 4.77644950e-02
-1.63344622e+00 2.35114306e-01 -7.71658644e-02 -3.39152515e-01
-3.57304573e-01 7.87208617e-01 -2.77937323e-01 -5.13585865e-01
3.04627657e-01 1.31936371e-01 -7.74486780e-01 5.41569889e-01
5.89941263e-01 3.71462256e-01 2.26014763e-01 -1.00968611e+00
-2.76005447e-01 5.76289952e-01 1.08965673e-01 -2.05367178e-01
1.51952744e+00 -4.58442181e-01 -5.46355307e-01 3.72373581e-01
7.78823853e-01 6.62562191e-01 -4.72994328e-01 1.40926063e-01
4.24501687e-01 -6.00917995e-01 -3.25328410e-01 -3.74731570e-01
-9.91686344e-01 8.44267309e-01 2.80492961e-01 8.41388166e-01
1.04426694e+00 -1.20029496e-02 2.97587156e-01 -3.45744580e-01
7.10905910e-01 -1.52366579e+00 -3.06067973e-01 5.87608337e-01
5.20866215e-01 -7.68175542e-01 -3.37270871e-02 -8.76714110e-01
-6.43559515e-01 1.09524059e+00 8.00634503e-01 2.59988725e-01
7.20432997e-01 4.63870645e-01 7.91665390e-02 -4.57491130e-01
-1.22226253e-01 -4.63932842e-01 9.00490880e-02 1.01047862e+00
6.06900990e-01 -2.18971476e-01 -1.29462409e+00 8.90410960e-01
-6.24644458e-01 -2.01866776e-01 6.55078292e-01 8.02705348e-01
-2.22557560e-01 -1.57512510e+00 -6.67313993e-01 1.29233047e-01
-1.10027635e+00 2.09465057e-01 -7.92717874e-01 1.02102101e+00
8.96626472e-01 1.25164056e+00 -3.26010257e-01 -5.02352357e-01
6.11872554e-01 -2.22316943e-02 8.82091895e-02 -1.05370963e+00
-1.04812598e+00 -2.47535352e-02 8.03282320e-01 -3.68793964e-01
-3.31651330e-01 -9.25473511e-01 -7.34381557e-01 -5.19703269e-01
-2.71275073e-01 3.54978442e-01 8.09010983e-01 1.02921093e+00
-1.24182060e-01 -4.45184894e-02 2.04146326e-01 -3.40573996e-01
6.97852522e-02 -4.88756508e-01 -8.43629539e-01 1.01450264e-01
-2.67022371e-01 -5.93685925e-01 3.73749249e-02 -6.27929032e-01] | [8.737125396728516, 10.46005630493164] |
85bd7830-94ba-4e4d-8220-f77df62d6f57 | polytrack-tracking-with-bounding-polygons | 2111.01606 | null | https://arxiv.org/abs/2111.01606v1 | https://arxiv.org/pdf/2111.01606v1.pdf | PolyTrack: Tracking with Bounding Polygons | In this paper, we present a novel method called PolyTrack for fast multi-object tracking and segmentation using bounding polygons. Polytrack detects objects by producing heatmaps of their center keypoint. For each of them, a rough segmentation is done by computing a bounding polygon over each instance instead of the traditional bounding box. Tracking is done by taking two consecutive frames as input and computing a center offset for each object detected in the first frame to predict its location in the second frame. A Kalman filter is also applied to reduce the number of ID switches. Since our target application is automated driving systems, we apply our method on urban environment videos. We trained and evaluated PolyTrack on the MOTS and KITTIMOTS datasets. Results show that tracking polygons can be a good alternative to bounding box and mask tracking. The code of PolyTrack is available at https://github.com/gafaua/PolyTrack. | ['Nicolas Saunier', 'Guillaume-Alexandre Bilodeau', 'Hughes Perreault', 'Gaspar Faure'] | 2021-11-02 | null | null | null | null | ['multi-object-tracking-and-segmentation'] | ['computer-vision'] | [-1.21121161e-01 -1.33966774e-01 -4.53632027e-02 -2.86818087e-01
-5.16418040e-01 -7.89805770e-01 4.07763541e-01 1.86466873e-01
-5.07303774e-01 3.09638619e-01 -4.90621209e-01 -8.55867341e-02
3.94794554e-01 -8.62109303e-01 -8.11240673e-01 -5.07660925e-01
2.94002779e-02 6.48841262e-01 1.25333273e+00 1.95816278e-01
1.23810060e-01 6.97810650e-01 -1.69911325e+00 2.22113371e-01
6.20388150e-01 9.95124400e-01 4.15818185e-01 9.51290190e-01
-1.33757129e-01 4.36159343e-01 -5.05613446e-01 -1.95879787e-01
5.68991303e-01 -1.30759954e-01 -4.76918250e-01 -1.54387318e-02
6.72530532e-01 -3.24207842e-01 -2.72159964e-01 9.36359167e-01
8.15741271e-02 2.81405896e-01 3.31204236e-01 -1.49223650e+00
2.91838288e-01 3.43836933e-01 -6.52359188e-01 3.92131954e-01
1.38552159e-01 2.02772349e-01 5.68064690e-01 -7.90746510e-01
7.70638645e-01 1.11251366e+00 6.92490280e-01 3.01084310e-01
-1.04139340e+00 -8.99033666e-01 4.48052846e-02 1.80818453e-01
-1.56049979e+00 -4.04277146e-01 3.30365688e-01 -6.14179909e-01
6.31067038e-01 4.02245462e-01 8.44465852e-01 2.63559192e-01
3.00447315e-01 9.20493782e-01 1.02096868e+00 -1.85488313e-01
1.81193054e-01 5.01084365e-02 3.32953870e-01 8.46278310e-01
2.67108828e-01 1.21196747e-01 -1.33185893e-01 -1.48795143e-01
6.01982296e-01 1.72833115e-01 1.71360806e-01 -4.78351116e-01
-1.35718071e+00 6.02845073e-01 4.29729789e-01 -1.55527284e-02
-2.66733855e-01 5.39529383e-01 2.12393537e-01 -9.02399272e-02
2.78364152e-01 -1.46449208e-02 -1.73559472e-01 -2.37177134e-01
-1.24012208e+00 4.15890306e-01 5.58117330e-01 1.04240835e+00
1.08825111e+00 -3.44567925e-01 -2.09887967e-01 3.34650576e-01
3.09042126e-01 6.16871834e-01 1.16313122e-01 -1.13366401e+00
1.49448663e-01 6.20770812e-01 3.81883919e-01 -9.19458449e-01
-6.20671690e-01 1.85785860e-01 -9.84131843e-02 6.67448461e-01
7.39092886e-01 -2.76501894e-01 -1.03202462e+00 9.22854066e-01
9.06690538e-01 4.40039933e-01 -3.42248887e-01 9.06174183e-01
7.31700420e-01 7.84908116e-01 -1.59840435e-02 1.17082432e-01
1.41446805e+00 -1.05336535e+00 -5.21581113e-01 -1.21082857e-01
6.57409668e-01 -7.74766505e-01 4.73600715e-01 2.94805318e-01
-8.76808822e-01 -7.64651179e-01 -7.48881221e-01 1.15578964e-01
-5.44800460e-01 2.66268224e-01 4.00991857e-01 5.41548550e-01
-7.44085908e-01 5.05566597e-01 -1.35390365e+00 -2.89426029e-01
4.20307398e-01 3.21159363e-01 2.42225593e-03 4.07014741e-03
-7.30527043e-01 7.39657044e-01 5.97844243e-01 -8.56638551e-02
-7.31462538e-01 -3.91653359e-01 -7.25289345e-01 -2.19578117e-01
3.87858599e-01 -2.03393698e-01 1.17850947e+00 -6.40948117e-01
-1.32853854e+00 7.14362442e-01 -3.34272295e-01 -7.35354006e-01
7.78594673e-01 -4.44921345e-01 -2.87049741e-01 2.60101072e-02
2.59956717e-01 9.35934842e-01 1.04622614e+00 -1.07871485e+00
-1.14762843e+00 -1.35536209e-01 -7.27130175e-02 -3.99060622e-02
2.99396217e-01 2.11111277e-01 -9.76597726e-01 -1.51925102e-01
1.43446013e-01 -1.19632423e+00 -3.24683219e-01 1.66507941e-02
-4.34888870e-01 -2.33641192e-01 1.27291179e+00 -4.57673252e-01
1.23697841e+00 -2.33235192e+00 -2.23811939e-01 1.87468424e-01
2.18999013e-01 3.32733572e-01 2.42285490e-01 1.13140509e-01
4.58689600e-01 -2.47320578e-01 -1.29635364e-01 -3.10692608e-01
-1.18681028e-01 7.54218549e-02 -2.24629477e-01 7.19380319e-01
-1.26247674e-01 7.11307764e-01 -8.84869039e-01 -6.85234487e-01
6.86740637e-01 4.32212085e-01 -4.34047818e-01 -1.29826084e-01
-3.22958589e-01 4.17046517e-01 -3.27699065e-01 6.32329047e-01
9.91705239e-01 -9.88844689e-03 -3.04783195e-01 -6.63442537e-02
-6.43467605e-01 6.25569001e-02 -1.65798783e+00 1.20870709e+00
-8.65036398e-02 9.94978249e-01 -1.61068037e-01 -3.37700546e-01
8.05601001e-01 -1.02646425e-01 6.15011752e-01 -4.88803506e-01
2.88521856e-01 4.25206544e-03 -1.72519624e-01 -2.49615967e-01
8.94800544e-01 4.51252133e-01 -1.62447646e-01 2.69211859e-01
-3.02444160e-01 -2.17866883e-01 6.47974312e-01 1.00483999e-01
1.09767640e+00 3.59599262e-01 1.61165342e-01 -7.08571030e-03
3.90748620e-01 5.83312929e-01 6.75585508e-01 8.11295748e-01
-4.23430413e-01 5.78135788e-01 4.97965142e-02 -6.68304503e-01
-8.37731302e-01 -9.92222428e-01 -3.39749128e-01 9.82739687e-01
6.69063151e-01 -6.13161385e-01 -8.80697370e-01 -6.16079628e-01
2.56278962e-01 4.09973800e-01 -4.11879480e-01 4.66356516e-01
-7.99019039e-01 -3.91712874e-01 3.38726252e-01 5.24353564e-01
5.35642922e-01 -9.45441604e-01 -1.44889510e+00 3.51821661e-01
-8.27519372e-02 -1.21167493e+00 -5.04401922e-01 1.72070995e-01
-7.59682119e-01 -1.20603573e+00 -3.41659963e-01 -3.79506081e-01
7.18226492e-01 4.24301714e-01 7.75992215e-01 1.67290807e-01
-4.26821202e-01 1.58669904e-01 -3.62819642e-01 -5.96908391e-01
-3.49801600e-01 3.54414550e-03 -5.02230376e-02 -3.97225730e-02
4.49022859e-01 2.59646736e-02 -5.59216678e-01 7.64862657e-01
-4.06084955e-01 3.41316134e-01 2.88985282e-01 1.61123067e-01
1.00492644e+00 4.27327305e-02 -7.95163363e-02 -7.50199318e-01
-1.12398818e-01 -2.97752738e-01 -1.54767978e+00 -1.96232736e-01
-7.03789890e-02 -3.66111279e-01 1.84140503e-01 -3.65010858e-01
-6.19532704e-01 8.60814452e-01 7.33144805e-02 -5.69466472e-01
-4.04485822e-01 -1.25707462e-01 3.26909930e-01 -1.03322029e-01
4.49793071e-01 -1.09753460e-01 -2.28509337e-01 -3.40135723e-01
4.91087824e-01 5.12601912e-01 6.17373765e-01 -1.11425176e-01
9.15261447e-01 8.46161485e-01 -2.49836996e-01 -7.78448880e-01
-5.50776482e-01 -8.89917254e-01 -9.59459186e-01 -6.85181260e-01
1.13953638e+00 -8.61730516e-01 -7.75696754e-01 3.03349406e-01
-1.09873283e+00 -6.15833223e-01 -1.97844058e-01 5.30233681e-01
-3.78675163e-01 -9.42354184e-03 -3.21594477e-01 -8.45201135e-01
-5.19240052e-02 -1.19840252e+00 1.15809214e+00 5.82821310e-01
-1.75084099e-01 -5.63079894e-01 2.17401221e-01 1.65059939e-01
5.90119958e-02 3.62201899e-01 -2.06136238e-02 -4.19465482e-01
-1.04569912e+00 -3.67991984e-01 -1.99469909e-01 -2.58477151e-01
-6.36693761e-02 4.29635227e-01 -1.03980076e+00 -1.24835327e-01
-5.21183312e-01 3.05282861e-01 8.74670804e-01 6.14438772e-01
1.03157127e+00 6.53827339e-02 -1.00897658e+00 5.73757827e-01
1.36890531e+00 5.06459832e-01 4.57320720e-01 5.33719003e-01
6.74955130e-01 1.34700000e-01 1.07756078e+00 3.35591167e-01
5.42104125e-01 9.36759949e-01 4.40483063e-01 -6.29124269e-02
-1.76623285e-01 9.94189009e-02 4.83310342e-01 2.01060519e-01
-4.47413251e-02 -9.54210311e-02 -1.20375001e+00 7.39709258e-01
-1.86458147e+00 -1.07194626e+00 -7.25378215e-01 2.16649246e+00
2.98465729e-01 3.99895966e-01 6.61774814e-01 3.70433144e-02
9.64064181e-01 -8.34508315e-02 -3.97969663e-01 -1.92812651e-01
2.91425765e-01 -2.10644484e-01 1.04235995e+00 6.03315413e-01
-1.48264432e+00 1.26351571e+00 5.90801430e+00 5.50533533e-01
-1.09420931e+00 2.73929119e-01 2.57650673e-01 -1.17626704e-01
5.24067760e-01 5.83222285e-02 -1.42918003e+00 6.54918969e-01
8.57897878e-01 2.11692527e-01 2.22714484e-01 1.05903411e+00
2.00780302e-01 -7.97991633e-01 -8.05988312e-01 8.24156225e-01
-3.48622680e-01 -1.35878348e+00 -6.65658832e-01 -5.12258001e-02
6.63733959e-01 5.72422028e-01 -2.43858799e-01 1.25927016e-01
5.50221801e-01 -4.74254817e-01 1.01643085e+00 2.99485624e-01
4.27821308e-01 -5.88276744e-01 3.90328020e-01 2.81456560e-01
-1.72087979e+00 4.60555926e-02 -3.66739303e-01 9.78029240e-03
2.07565010e-01 5.22386014e-01 -1.08826542e+00 2.43274763e-01
9.52125847e-01 4.54034358e-01 -7.53610194e-01 1.71490777e+00
-8.59921426e-02 5.22426903e-01 -7.76727855e-01 1.45111546e-01
1.40499786e-01 -2.05217123e-01 6.61401510e-01 1.57947421e+00
2.48445809e-01 -4.90495889e-03 6.02314949e-01 7.99444139e-01
2.55785853e-01 -1.86405107e-01 -5.62608004e-01 3.96919549e-01
7.49995530e-01 1.60266423e+00 -1.49877965e+00 -6.14160299e-01
-4.67924118e-01 7.55861819e-01 -7.58549124e-02 3.80158387e-02
-1.23077750e+00 -5.79496443e-01 6.29480898e-01 2.75983125e-01
6.89867079e-01 -5.25414705e-01 -2.66961604e-01 -7.27111220e-01
-2.62941241e-01 -1.58611178e-01 4.45719868e-01 -7.18932867e-01
-2.61083573e-01 5.82260072e-01 1.75092936e-01 -1.47363961e+00
3.68182920e-02 -3.44350845e-01 -6.67899489e-01 4.28719044e-01
-1.41836965e+00 -9.27146614e-01 -6.20159626e-01 5.01403570e-01
5.22248387e-01 4.50834960e-01 3.27925593e-01 4.27691549e-01
-6.31626844e-01 1.08795568e-01 2.34588847e-01 3.54943663e-01
4.78020370e-01 -1.19299424e+00 7.61165798e-01 1.02849281e+00
2.50803232e-01 6.54566437e-02 8.16581428e-01 -9.59004045e-01
-1.19646370e+00 -1.34606659e+00 4.56998467e-01 -7.71148384e-01
4.59146082e-01 -5.00030816e-01 -8.52867782e-01 1.02284467e+00
-1.35475984e-02 4.85864103e-01 1.65658370e-01 -4.09621060e-01
2.47869730e-01 2.69983280e-02 -1.01304281e+00 3.17061275e-01
7.81127930e-01 -1.62980780e-02 -2.00156674e-01 3.37235987e-01
5.47099888e-01 -1.08586657e+00 -6.94114864e-01 9.33653042e-02
6.68511033e-01 -8.37838709e-01 7.80372798e-01 2.22244427e-01
-2.54677206e-01 -1.03748572e+00 1.19670421e-01 -1.00787759e+00
-7.82412961e-02 -5.99524140e-01 -1.59713596e-01 8.75801921e-01
2.58090734e-01 -4.35414255e-01 7.96703041e-01 4.47297454e-01
-2.48256475e-01 -4.41515267e-01 -9.08031046e-01 -8.72822523e-01
-5.31738698e-01 -6.35153115e-01 5.86529434e-01 5.55576384e-01
-2.42021799e-01 -2.97102451e-01 2.05352809e-02 6.08332038e-01
6.52472436e-01 3.87684852e-01 1.26148307e+00 -1.28264749e+00
5.96889853e-02 -1.70723245e-01 -6.20047331e-01 -9.09407020e-01
-1.66311666e-01 -7.16370463e-01 3.99019808e-01 -1.49260247e+00
-1.77713603e-01 -7.03453720e-01 4.48625535e-02 6.08078301e-01
2.57040951e-02 6.88274324e-01 5.34404933e-01 2.03507155e-01
-1.00861931e+00 -9.82024893e-02 9.63851750e-01 -1.19198917e-03
-5.17634034e-01 1.39082178e-01 1.24753736e-01 8.63579750e-01
1.00705993e+00 -8.43598247e-01 1.71329260e-01 -2.86234528e-01
-1.18645892e-01 3.33476588e-02 4.55688149e-01 -1.65386903e+00
4.58417088e-01 -1.93723619e-01 4.05241311e-01 -1.25880361e+00
4.64634299e-01 -8.59333634e-01 4.17594135e-01 6.56700552e-01
3.35233599e-01 2.27674425e-01 6.36193573e-01 3.37772399e-01
8.20816755e-02 -7.23139346e-02 9.31115091e-01 -2.21763775e-02
-1.02777946e+00 1.85320392e-01 -6.17164195e-01 -2.55202651e-01
1.47505283e+00 -5.72361946e-01 -1.92153633e-01 1.04009770e-01
-8.85152459e-01 5.47193825e-01 6.09723628e-01 4.15855289e-01
4.99074280e-01 -1.24888372e+00 -4.04196918e-01 2.60991126e-01
3.21486266e-03 1.89259261e-01 1.65334009e-02 1.00869477e+00
-7.58627713e-01 3.36385965e-01 -1.69311807e-01 -1.02735436e+00
-1.70078230e+00 4.48592603e-01 4.10819978e-01 7.10034668e-02
-9.65946138e-01 7.19214737e-01 2.11282074e-01 -1.52302921e-01
1.25106290e-01 -7.17272520e-01 -1.36869788e-01 1.14337072e-01
7.22550392e-01 6.06088161e-01 1.04448553e-02 -9.06837225e-01
-5.99893272e-01 7.50855386e-01 -8.58828947e-02 -1.12381674e-01
1.00507879e+00 -2.54465222e-01 1.89659163e-01 5.13758898e-01
6.71643376e-01 -1.63471829e-02 -1.61305273e+00 -1.45799154e-02
2.04221010e-01 -7.54712999e-01 3.02877165e-02 -4.09810036e-01
-9.97338414e-01 4.64432985e-01 1.01169217e+00 1.80453956e-01
6.29102886e-01 1.56775475e-01 7.52830207e-01 2.18057513e-01
4.78789538e-01 -1.17272317e+00 -2.70309925e-01 5.86063206e-01
4.08364236e-01 -1.21711731e+00 -8.40929225e-02 -4.14858729e-01
-4.57564980e-01 1.08708084e+00 7.40570009e-01 -1.84636295e-01
5.43191314e-01 6.16634846e-01 1.49422169e-01 -3.38568777e-01
-1.77999601e-01 -5.37503600e-01 5.62481061e-02 4.41886544e-01
7.27042332e-02 2.46622413e-01 -4.50869203e-02 -1.12572953e-01
-2.40833685e-01 7.02770278e-02 6.27494574e-01 1.07931173e+00
-8.93664181e-01 -7.87344694e-01 -9.69610333e-01 3.95949185e-01
-3.36812049e-01 3.44507337e-01 -1.34988680e-01 8.69383931e-01
3.66946757e-01 8.48700643e-01 5.42865157e-01 -4.28475171e-01
2.95377046e-01 5.37780002e-02 3.27392817e-01 -6.20580971e-01
-4.82192427e-01 9.57048088e-02 -7.47645157e-04 -8.57489407e-01
-3.33766729e-01 -1.05870402e+00 -1.80285454e+00 -2.16128245e-01
-5.12505412e-01 3.05682961e-02 8.25571775e-01 6.37898147e-01
3.16932708e-01 5.21612704e-01 3.67817402e-01 -1.34465909e+00
4.14142728e-01 -5.99159956e-01 -2.42085442e-01 9.64446142e-02
4.60224003e-01 -8.62739086e-01 5.91598079e-02 1.97474152e-01] | [7.052350997924805, -1.940996766090393] |
06bc52aa-6cb8-4776-b5a8-916ad0ee47f7 | causal-reinforcement-learning-a-survey | 2307.01452 | null | https://arxiv.org/abs/2307.01452v1 | https://arxiv.org/pdf/2307.01452v1.pdf | Causal Reinforcement Learning: A Survey | Reinforcement learning is an essential paradigm for solving sequential decision problems under uncertainty. Despite many remarkable achievements in recent decades, applying reinforcement learning methods in the real world remains challenging. One of the main obstacles is that reinforcement learning agents lack a fundamental understanding of the world and must therefore learn from scratch through numerous trial-and-error interactions. They may also face challenges in providing explanations for their decisions and generalizing the acquired knowledge. Causality, however, offers a notable advantage as it can formalize knowledge in a systematic manner and leverage invariance for effective knowledge transfer. This has led to the emergence of causal reinforcement learning, a subfield of reinforcement learning that seeks to enhance existing algorithms by incorporating causal relationships into the learning process. In this survey, we comprehensively review the literature on causal reinforcement learning. We first introduce the basic concepts of causality and reinforcement learning, and then explain how causality can address core challenges in non-causal reinforcement learning. We categorize and systematically review existing causal reinforcement learning approaches based on their target problems and methodologies. Finally, we outline open issues and future directions in this emerging field. | ['Chengqi Zhang', 'Guodong Long', 'Jing Jiang', 'Zhihong Deng'] | 2023-07-04 | null | null | null | null | ['transfer-learning'] | ['miscellaneous'] | [ 2.27056116e-01 1.82428584e-01 -7.51891136e-01 -2.63416439e-01
-4.76141453e-01 -5.53957820e-01 7.27114141e-01 3.23034912e-01
-3.55959296e-01 1.37845254e+00 5.19131087e-02 -2.99155325e-01
-8.73456120e-01 -9.25868332e-01 -6.86113119e-01 -6.94151700e-01
-5.46634853e-01 4.06966895e-01 1.30580977e-01 -4.18220401e-01
5.64191103e-01 3.34797293e-01 -1.54015291e+00 -1.04790360e-01
1.02822518e+00 6.45366490e-01 1.33006036e-01 6.60652518e-01
7.02216625e-02 1.37386703e+00 -5.82322359e-01 -1.34531140e-01
3.42166498e-02 -7.57949710e-01 -9.81380880e-01 -5.48819415e-02
-6.74420819e-02 -4.82889265e-01 -5.92994750e-01 9.83809352e-01
1.64101586e-01 2.73166239e-01 6.94115579e-01 -1.64963567e+00
-6.98142946e-01 8.71420264e-01 -3.49405646e-01 2.51844525e-01
5.50126851e-01 1.79965451e-01 1.21898019e+00 -3.16855580e-01
4.37801003e-01 1.30925822e+00 4.18115348e-01 7.31855810e-01
-1.18940711e+00 -7.16490805e-01 7.70241380e-01 9.79923904e-01
-7.72973061e-01 5.16291410e-02 8.08686256e-01 -3.39425415e-01
1.04234922e+00 -1.62497118e-01 8.37562203e-01 1.13723195e+00
4.04477298e-01 9.84410048e-01 1.49159837e+00 -6.89822793e-01
4.84657258e-01 -4.51767385e-01 -4.37382385e-02 5.57137609e-01
9.20900255e-02 1.09992456e+00 -9.50612128e-01 -2.96276864e-02
9.04946327e-01 -1.86585382e-01 6.46593794e-02 -8.07121217e-01
-1.04528964e+00 1.13602054e+00 4.79821891e-01 2.16658711e-02
-5.61558247e-01 6.54869616e-01 1.36074573e-01 5.48641562e-01
-3.25345308e-01 5.78524768e-01 -6.23026431e-01 -2.71701753e-01
-3.78918082e-01 8.20286989e-01 6.98710740e-01 5.40612638e-01
7.60905206e-01 3.47713172e-01 1.11380756e-01 5.34989536e-01
5.43072820e-01 5.98581672e-01 2.36763030e-01 -1.09716129e+00
1.83213681e-01 2.04575628e-01 2.28649691e-01 -4.43692356e-01
-3.32534105e-01 -7.92317316e-02 -3.84290993e-01 7.51100719e-01
3.36143047e-01 -4.32941079e-01 -7.80310273e-01 1.79226255e+00
3.37752134e-01 2.04187438e-01 3.11051279e-01 7.99384356e-01
3.88639718e-01 3.43305498e-01 2.96895623e-01 -5.02041996e-01
8.28285396e-01 -7.88762033e-01 -9.43975449e-01 -3.81168306e-01
-8.33308250e-02 -5.37391961e-01 6.94729328e-01 4.71122116e-01
-8.38634908e-01 -3.03193659e-01 -1.20176589e+00 4.20347869e-01
-2.99923033e-01 -7.53164887e-01 1.11440957e+00 4.19888020e-01
-6.97240412e-01 9.35655713e-01 -8.60564709e-01 -1.77826941e-01
3.43410403e-01 5.44430912e-01 5.58304153e-02 1.42039254e-01
-1.71609545e+00 1.25158107e+00 4.18613881e-01 -7.24933520e-02
-1.43612027e+00 -7.18339682e-01 -8.70072663e-01 -2.75475830e-01
8.58684480e-01 -6.44386351e-01 1.88819993e+00 -7.92919099e-01
-1.86471641e+00 4.66108546e-02 1.10512003e-01 -6.87771618e-01
5.53663552e-01 -3.63594174e-01 -3.56866181e-01 5.55288829e-02
1.11295953e-01 6.71503782e-01 8.89054179e-01 -1.30335033e+00
-1.15312266e+00 -2.34195977e-01 4.22879547e-01 4.12339836e-01
1.73696041e-01 -2.92616189e-01 2.66062677e-01 -5.46035230e-01
-2.99839109e-01 -8.31014991e-01 -6.29046798e-01 -5.05321980e-01
2.24382449e-02 -5.40811181e-01 7.29637325e-01 1.53600752e-01
1.10704398e+00 -1.61508584e+00 2.93165833e-01 1.96201831e-01
2.04374656e-01 -1.42910510e-01 -1.40568927e-01 8.58244896e-01
-2.27662280e-01 -1.30955979e-01 -2.42960304e-01 4.56426948e-01
4.64318581e-02 6.41248465e-01 -7.99936950e-01 2.16730684e-01
3.93421620e-01 1.05554211e+00 -1.57627606e+00 -2.46034160e-01
6.02176964e-01 1.19846255e-01 -6.59713387e-01 5.04881442e-01
-5.77108681e-01 5.81062913e-01 -5.76083720e-01 4.02238250e-01
1.04501143e-01 -2.67021582e-02 5.94211221e-01 4.52872843e-01
-2.23115474e-01 4.58842874e-01 -1.46475983e+00 1.40776467e+00
-3.07688445e-01 4.65154946e-01 -3.96474957e-01 -1.30149555e+00
6.48930669e-01 4.89704788e-01 7.82677770e-01 -9.07248318e-01
-1.60503104e-01 2.45169476e-01 2.11418852e-01 -5.25414050e-01
1.46214500e-01 -5.98542392e-01 -1.93464130e-01 5.98049343e-01
1.16400398e-01 -4.94580120e-01 4.10129845e-01 -4.56944965e-02
1.06253779e+00 5.02960265e-01 8.45787764e-01 -4.43977378e-02
3.20289016e-01 1.94923893e-01 7.33996093e-01 1.03910816e+00
-4.60870713e-01 -1.20341346e-01 5.73926210e-01 -7.18092561e-01
-4.70520377e-01 -1.38946342e+00 1.30252736e-02 1.13455331e+00
1.55729353e-01 -2.16712132e-01 -3.21736932e-01 -9.95249867e-01
3.43966961e-01 8.54505479e-01 -1.04562283e+00 -2.61311769e-01
-4.44903374e-01 -7.68394589e-01 1.81214884e-01 7.38093615e-01
3.61022174e-01 -1.52435744e+00 -8.47138226e-01 4.48471099e-01
-4.30069724e-03 -6.53648019e-01 2.41807953e-01 6.16554439e-01
-9.56498265e-01 -1.54106116e+00 -1.71159133e-01 -3.91739726e-01
3.16487014e-01 1.94409832e-01 1.18110299e+00 -4.76763695e-02
-2.40092590e-01 8.23861599e-01 -3.52472872e-01 -7.37204909e-01
-4.84913439e-01 -3.11451823e-01 4.45005178e-01 -6.90799892e-01
4.81488138e-01 -6.09968960e-01 -4.43764836e-01 1.43641397e-01
-8.54193449e-01 -4.20087665e-01 4.78150517e-01 1.15407097e+00
4.01815474e-01 4.01781708e-01 1.15582967e+00 -8.37412655e-01
9.63369966e-01 -4.71971959e-01 -8.76752198e-01 2.40862876e-01
-1.15809751e+00 3.33651751e-01 4.72100377e-01 -1.71062693e-01
-1.20034957e+00 -1.24413565e-01 1.09286860e-01 1.11825150e-02
-2.82639027e-01 6.73137665e-01 2.54365027e-01 8.83671865e-02
8.02499712e-01 -7.57153286e-03 3.68923172e-02 6.22638948e-02
7.06502557e-01 1.09795824e-01 1.73220560e-01 -7.84886062e-01
7.61182368e-01 1.76982343e-01 1.75697431e-01 -4.67746824e-01
-9.60321963e-01 -5.65808341e-02 -7.99722075e-01 -4.65774357e-01
7.04014182e-01 -5.33125460e-01 -1.12511873e+00 2.72077888e-01
-7.83372045e-01 -8.76160920e-01 -5.74303687e-01 5.18379450e-01
-1.16684234e+00 5.12088537e-02 -4.92674261e-01 -8.12211394e-01
3.02853733e-01 -1.15984285e+00 2.60636717e-01 4.84908074e-01
-3.92958075e-01 -1.28699875e+00 3.38527650e-01 1.55879334e-01
1.82862192e-01 8.40737820e-02 1.04043710e+00 -1.84702978e-01
-6.77101076e-01 3.19831818e-01 1.14432812e-01 -9.64582781e-04
4.46718991e-01 4.32563992e-03 -7.33467340e-01 1.00878710e-02
-1.75160930e-01 -8.15158427e-01 6.84020460e-01 5.28997302e-01
8.20136905e-01 -5.02365977e-02 -9.90192145e-02 -1.35131449e-01
1.29860020e+00 6.09617829e-01 1.54091209e-01 5.70393622e-01
3.21580231e-01 8.07829499e-01 1.05883360e+00 5.90267718e-01
6.25453055e-01 3.90808374e-01 7.46054113e-01 2.05093548e-01
-6.22731186e-02 -5.14805853e-01 3.39694083e-01 5.46169281e-01
-2.29378715e-01 1.31475091e-01 -9.50573504e-01 2.96721160e-01
-2.28266501e+00 -1.21918786e+00 1.10986963e-01 1.86544120e+00
1.08617198e+00 7.86214992e-02 1.92932263e-01 2.64070600e-01
3.76552910e-01 -4.08642925e-02 -9.54674482e-01 -3.59493792e-01
2.78230309e-02 2.14023367e-01 3.75813454e-01 8.59674692e-01
-1.08277488e+00 1.13680804e+00 8.19652462e+00 3.02017570e-01
-8.72004092e-01 -3.76276106e-01 1.70088291e-01 1.72399774e-01
-2.26613149e-01 1.35132566e-01 -5.70233941e-01 -1.10179417e-01
7.54684508e-01 -2.67648101e-01 7.90854394e-01 7.25518346e-01
3.40974092e-01 -3.85157764e-01 -1.38970780e+00 5.61657548e-01
-3.38973731e-01 -1.33190870e+00 1.35809377e-01 -1.99793324e-01
1.05447936e+00 -1.32826701e-01 2.29267150e-01 5.46312988e-01
1.23494422e+00 -1.18640637e+00 6.31919384e-01 2.70163089e-01
3.03425580e-01 -9.51333046e-01 3.89238328e-01 2.00286373e-01
-9.40931976e-01 -5.75345457e-01 -2.37905204e-01 -7.12870300e-01
-1.04351774e-01 2.62031764e-01 -8.30870748e-01 6.75369620e-01
6.64577425e-01 8.88542354e-01 -8.54327679e-02 8.15162301e-01
-1.03675377e+00 5.81956685e-01 1.61953583e-01 -2.97910184e-01
4.51132834e-01 -6.33158758e-02 2.41020516e-01 7.95436621e-01
-1.40555799e-01 2.02938870e-01 4.12608743e-01 6.43111825e-01
2.64405757e-01 -1.85418949e-01 -6.13819182e-01 1.74182821e-02
5.55922985e-01 6.66146755e-01 -5.15422821e-01 -2.10453689e-01
-4.97162849e-01 4.08329308e-01 5.95622182e-01 5.13801932e-01
-8.46661687e-01 -1.71031296e-01 7.82046914e-01 -3.45150888e-01
2.23884821e-01 -2.95568466e-01 -2.37774819e-01 -9.36929226e-01
-4.81375575e-01 -1.33865762e+00 6.78383112e-01 -3.74273360e-01
-1.39260638e+00 1.26163929e-03 3.08571041e-01 -9.69264150e-01
-6.63936257e-01 -6.51714146e-01 -3.19329321e-01 4.50408310e-01
-1.92653298e+00 -6.38658822e-01 -2.88271494e-02 6.94319427e-01
7.31486261e-01 -1.41167015e-01 9.91632760e-01 -2.13133827e-01
-3.61123025e-01 2.74154469e-02 3.34240757e-02 -1.34918869e-01
8.29401374e-01 -1.72010088e+00 -3.94407054e-03 4.86936122e-01
1.75014094e-01 6.57726109e-01 8.85615945e-01 -7.41973162e-01
-1.43516111e+00 -7.24668324e-01 4.74434763e-01 -3.46734673e-01
1.14154708e+00 1.33399293e-02 -6.75544143e-01 6.86792374e-01
4.69932199e-01 -3.24980825e-01 7.65986502e-01 5.58225095e-01
-3.82344216e-01 -1.41777754e-01 -7.81728864e-01 9.50466514e-01
8.73261452e-01 -3.12631965e-01 -1.08006883e+00 7.25044757e-02
4.03671741e-01 -2.80428290e-01 -7.73501575e-01 4.73650873e-01
5.40349901e-01 -9.68613625e-01 8.99411678e-01 -8.99529397e-01
6.07362926e-01 -4.48401660e-01 9.83863100e-02 -1.82986498e+00
-4.52697188e-01 -8.67425919e-01 -3.00206780e-01 7.03716338e-01
2.49331400e-01 -5.87059796e-01 7.92804062e-01 4.54678267e-01
1.54856279e-01 -6.81806982e-01 -7.27697790e-01 -8.55627418e-01
5.53650200e-01 -7.03765571e-01 4.52051848e-01 9.66849267e-01
4.69597101e-01 3.81053627e-01 -3.91186506e-01 1.82909276e-02
1.01124394e+00 2.88534552e-01 6.24377251e-01 -1.03580439e+00
-2.71276087e-01 -6.00874484e-01 -7.22521544e-02 -9.08922672e-01
3.88188273e-01 -5.36667585e-01 3.10337931e-01 -1.63272440e+00
6.22481517e-02 -3.64733905e-01 -6.10242009e-01 6.49221182e-01
-4.62443948e-01 -2.80613363e-01 1.17339619e-01 -4.34226245e-02
-6.31764352e-01 7.88858235e-01 1.48642552e+00 -2.66337860e-02
-2.80306280e-01 7.96395540e-02 -6.67057097e-01 8.81421506e-01
1.11917961e+00 -5.20204663e-01 -8.85590553e-01 -1.87593266e-01
5.10665953e-01 2.44856715e-01 4.03170735e-01 -7.05741107e-01
3.68858784e-01 -1.02481031e+00 3.86933118e-01 -3.67241025e-01
-7.19148815e-02 -8.05965245e-01 -4.21591073e-01 7.77042031e-01
-6.34950817e-01 1.72569096e-01 2.54741818e-01 9.36372459e-01
-3.74731392e-01 -3.07989538e-01 7.19415605e-01 -2.06762776e-01
-1.21336412e+00 1.08449481e-01 -9.67894971e-01 1.24079540e-01
1.35866570e+00 7.88004696e-02 -2.41935328e-01 -4.28932011e-01
-6.98966682e-01 5.74790716e-01 -6.16145618e-02 6.62396848e-01
7.12476730e-01 -1.27861619e+00 -5.12880385e-01 -8.40083957e-02
7.06684366e-02 -2.59112597e-01 1.98023006e-01 4.41300333e-01
2.27187425e-02 6.50252461e-01 -5.42675078e-01 -2.84464896e-01
-8.74720871e-01 7.41504133e-01 4.39558685e-01 -3.84486020e-01
-2.84292847e-01 5.79835773e-01 1.71876252e-01 -4.81037021e-01
3.60930622e-01 -1.02865435e-01 -3.92610580e-01 2.27249227e-02
4.45842296e-01 4.27154332e-01 -2.62800246e-01 1.92713380e-01
-3.57564598e-01 3.94693464e-01 -9.30176452e-02 -4.13027257e-01
1.39369273e+00 -7.55749941e-02 8.15114081e-02 6.24001265e-01
3.32251340e-01 -3.89198780e-01 -1.73653531e+00 -1.65137380e-01
3.38863939e-01 -3.04539442e-01 -3.36735807e-02 -1.12702882e+00
-6.40420973e-01 8.93194318e-01 3.95128429e-01 2.65517294e-01
9.08809066e-01 -9.28380638e-02 9.89143327e-02 6.13299072e-01
6.41455114e-01 -1.43327177e+00 6.43285632e-01 9.12350476e-01
1.03932643e+00 -1.31691539e+00 1.18039109e-01 -1.84553340e-01
-6.99857950e-01 1.36464489e+00 7.81812131e-01 -2.84204930e-01
7.65199959e-01 2.78563946e-01 2.72825778e-01 -1.08503386e-01
-1.05698121e+00 -4.20790672e-01 -3.97366807e-02 1.02313459e+00
4.54044402e-01 1.36376172e-01 -1.00912489e-01 1.83669209e-01
-2.08722115e-01 1.24019377e-01 6.50885582e-01 1.13613236e+00
-5.03277719e-01 -1.74586844e+00 -4.32124197e-01 1.49359792e-01
-2.98669338e-01 1.72031835e-01 -2.44797379e-01 8.59391987e-01
-7.90632591e-02 1.21860838e+00 -2.42721617e-01 -2.15393752e-01
4.16369647e-01 -1.30957186e-01 9.41573739e-01 -6.14460230e-01
-3.11899394e-01 -1.80740163e-01 -2.29519755e-01 -6.61769986e-01
-6.55850768e-01 -9.71781313e-01 -1.61884487e+00 -1.81190312e-01
1.35288805e-01 2.47367471e-01 3.79724234e-01 1.09224832e+00
-1.17510714e-01 9.65338469e-01 5.82119346e-01 -5.55585265e-01
-9.63092983e-01 -5.25195420e-01 -3.57756585e-01 1.23435408e-01
4.75579590e-01 -1.27369213e+00 -9.59485173e-02 6.43799081e-02] | [3.9748973846435547, 1.745336890220642] |
037216d5-94a1-44b7-b3d6-d135a92d78bb | using-spatio-temporal-dual-stream-network | 2305.02719 | null | https://arxiv.org/abs/2305.02719v2 | https://arxiv.org/pdf/2305.02719v2.pdf | Using Spatio-Temporal Dual-Stream Network with Self-Supervised Learning for Lung Tumor Classification on Radial Probe Endobronchial Ultrasound Video | The purpose of this study is to develop a computer-aided diagnosis system for classifying benign and malignant lung lesions, and to assist physicians in real-time analysis of radial probe endobronchial ultrasound (EBUS) videos. During the biopsy process of lung cancer, physicians use real-time ultrasound images to find suitable lesion locations for sampling. However, most of these images are difficult to classify and contain a lot of noise. Previous studies have employed 2D convolutional neural networks to effectively differentiate between benign and malignant lung lesions, but doctors still need to manually select good-quality images, which can result in additional labor costs. In addition, the 2D neural network has no ability to capture the temporal information of the ultrasound video, so it is difficult to obtain the relationship between the features of the continuous images. This study designs an automatic diagnosis system based on a 3D neural network, uses the SlowFast architecture as the backbone to fuse temporal and spatial features, and uses the SwAV method of contrastive learning to enhance the noise robustness of the model. The method we propose includes the following advantages, such as (1) using clinical ultrasound films as model input, thereby reducing the need for high-quality image selection by physicians, (2) high-accuracy classification of benign and malignant lung lesions can assist doctors in clinical diagnosis and reduce the time and risk of surgery, and (3) the capability to classify well even in the presence of significant image noise. The AUC, accuracy, precision, recall and specificity of our proposed method on the validation set reached 0.87, 83.87%, 86.96%, 90.91% and 66.67%, respectively. The results have verified the importance of incorporating temporal information and the effectiveness of using the method of contrastive learning on feature extraction. | ['Yun-Chien Cheng', 'Chin-Wen Chen', 'Ching-Kai Lin'] | 2023-05-04 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [ 3.56283523e-02 -3.47188503e-01 -1.04081258e-01 8.36918354e-02
-5.12801766e-01 -3.85979623e-01 9.49981362e-02 -1.54644996e-01
-4.76292104e-01 2.46346533e-01 -1.91949680e-01 -6.64882600e-01
-4.83097702e-01 -7.48921335e-01 -9.39399078e-02 -1.03356385e+00
-2.90805310e-01 4.24878627e-01 6.50078177e-01 3.09486896e-01
-1.99300364e-01 8.24957907e-01 -1.29461420e+00 3.31671029e-01
5.86597323e-01 1.26936638e+00 5.69505394e-01 9.60448384e-01
-1.00999705e-01 8.37654531e-01 -3.16382349e-01 6.09081276e-02
2.70322621e-01 -5.57859540e-01 -5.72905421e-01 8.58485997e-02
3.53552178e-02 -7.00434804e-01 -6.88312888e-01 7.02015281e-01
9.67595041e-01 -1.25299335e-01 7.33466446e-01 -5.50219119e-01
-2.31823266e-01 1.17164738e-01 -3.76953214e-01 6.53048337e-01
-1.07580729e-01 3.61378849e-01 3.30439508e-01 -6.87449813e-01
3.15032780e-01 5.73321700e-01 7.44892180e-01 6.38062179e-01
-5.09302735e-01 -6.29757166e-01 -5.04694104e-01 3.76313835e-01
-1.16896892e+00 -3.27283330e-02 4.38725561e-01 -6.71373963e-01
6.00722015e-01 5.71490765e-01 1.08758759e+00 6.47700727e-01
6.35154188e-01 5.35011947e-01 9.27201807e-01 -2.79527187e-01
-1.27774045e-01 1.59565747e-01 -1.56111717e-01 1.05518472e+00
3.10305238e-01 4.16373968e-01 2.16726452e-01 -1.65004566e-01
1.03014529e+00 4.41352218e-01 -6.04881704e-01 -2.76653320e-01
-1.22155559e+00 6.51363015e-01 4.22434121e-01 9.15947556e-01
-5.59528232e-01 -1.09734930e-01 4.44659650e-01 -9.19303447e-02
-2.87749153e-03 3.42533767e-01 -5.53598292e-02 -1.42786086e-01
-7.46488273e-01 -3.68980080e-01 6.44351602e-01 4.12212074e-01
-5.03730634e-03 -1.98490266e-02 -3.51511180e-01 7.72388399e-01
2.44081348e-01 6.83547854e-01 1.07303178e+00 -7.83288300e-01
5.25571816e-02 3.65158200e-01 -8.97619724e-02 -1.01140094e+00
-6.91312313e-01 -6.45371199e-01 -1.28805482e+00 -1.96483567e-01
2.27542996e-01 -4.24353741e-02 -1.18285751e+00 1.07067287e+00
1.67987064e-01 3.43787342e-01 -9.48724002e-02 1.05150139e+00
1.13239884e+00 3.96482497e-01 -1.21728040e-01 -5.40171981e-01
1.38648248e+00 -7.76784897e-01 -8.10654104e-01 1.95592836e-01
7.07454741e-01 -5.17413199e-01 7.06167042e-01 2.09071144e-01
-1.07782423e+00 -6.10070765e-01 -9.86914694e-01 5.03481209e-01
-3.05538289e-02 4.91592824e-01 6.12673044e-01 6.73650742e-01
-8.75216305e-01 4.96976167e-01 -1.19997025e+00 -1.14232212e-01
3.96630168e-01 6.37859702e-01 -2.63778299e-01 -1.71141893e-01
-1.25911188e+00 9.40816104e-01 2.29283139e-01 3.19061607e-01
-7.11435020e-01 -6.91721320e-01 -5.31946123e-01 1.32857963e-01
2.69886047e-01 -6.46230757e-01 1.40045941e+00 -6.83357298e-01
-1.17940450e+00 6.94724441e-01 5.98343648e-02 -1.10518217e-01
5.05814433e-01 3.77040744e-01 -3.83203983e-01 6.72449529e-01
-1.08285718e-01 3.36405754e-01 4.66590375e-01 -7.36717045e-01
-9.18950677e-01 -2.10854247e-01 -3.05823594e-01 2.14251745e-02
-1.98728859e-01 -2.88838238e-01 -7.75863171e-01 -4.19794828e-01
3.09454829e-01 -1.02227080e+00 -2.27801055e-01 3.68757248e-02
1.95704833e-01 -1.24089913e-02 1.07122111e+00 -9.29414272e-01
1.35737658e+00 -2.21527076e+00 -3.31754088e-01 3.51796418e-01
4.45730090e-01 8.84206593e-01 2.32737571e-01 -2.19420940e-01
-1.08795270e-01 1.85033113e-01 8.76507238e-02 5.98793685e-01
-5.58686972e-01 4.85975593e-02 5.01288593e-01 3.94642830e-01
5.76458983e-02 9.73970354e-01 -8.15824628e-01 -7.28013098e-01
5.74566305e-01 5.39750218e-01 -1.09143600e-01 3.52752388e-01
3.78437072e-01 4.00844067e-01 -5.29131472e-01 5.47884107e-01
4.04703587e-01 -4.55309182e-01 1.20139509e-01 -5.03458679e-01
8.59001055e-02 1.15102559e-01 -1.07275343e+00 9.54764307e-01
-5.15317023e-01 5.67903161e-01 6.43416420e-02 -9.40742493e-01
6.34416878e-01 9.53890622e-01 9.96708155e-01 -6.71845734e-01
4.59804565e-01 4.98739630e-01 5.38639486e-01 -1.36887074e+00
-3.82005453e-01 -2.85078526e-01 5.18493831e-01 3.55014890e-01
-2.48339981e-01 -1.72471836e-01 8.24933499e-02 -2.00263366e-01
1.25988162e+00 -4.45540994e-01 3.04017991e-01 8.66058543e-02
6.68040752e-01 -1.29671142e-01 2.87862092e-01 7.17465222e-01
-3.69609445e-01 7.06652641e-01 2.30936602e-01 -5.59769690e-01
-6.01343036e-01 -6.71122432e-01 -3.01628858e-01 2.57283211e-01
1.13553107e-01 2.81971663e-01 -4.24766660e-01 -8.95410240e-01
-6.81890622e-02 1.64274439e-01 -4.61028963e-01 -2.01400638e-01
-7.71558166e-01 -6.47176504e-01 4.32063669e-01 6.91417396e-01
6.32115602e-01 -8.59448612e-01 -9.04144704e-01 3.01118910e-01
-3.59743893e-01 -7.26443946e-01 -4.67877448e-01 9.06987637e-02
-1.02931964e+00 -1.34124565e+00 -9.18798506e-01 -9.84642863e-01
6.19783163e-01 6.61106646e-01 7.01588035e-01 4.34640944e-01
-7.30141938e-01 4.44453180e-01 -1.40195295e-01 -2.65079826e-01
-5.02521098e-01 -1.64136708e-01 -1.87296540e-01 -2.60062069e-01
2.49039322e-01 -2.20453158e-01 -7.28849828e-01 5.17018735e-01
-8.94830048e-01 -2.28399993e-03 1.02412248e+00 1.14626372e+00
6.24181747e-01 3.73561025e-01 5.65761805e-01 -7.93314397e-01
5.44977307e-01 -4.32783782e-01 -3.85717750e-01 1.20301403e-01
-5.18947721e-01 -4.16696578e-01 4.92718786e-01 -5.84237754e-01
-8.58308494e-01 2.01644213e-03 -1.43432960e-01 -6.42272830e-01
9.69394110e-03 5.40551007e-01 4.84907001e-01 -2.06635505e-01
6.96209013e-01 2.61159956e-01 3.41051430e-01 -2.40634773e-02
-3.95403773e-01 1.00057089e+00 3.94654691e-01 1.97940826e-01
5.65908372e-01 3.10977131e-01 2.72328734e-01 -7.58077443e-01
-4.94664192e-01 -8.57308507e-01 -5.96839488e-01 -3.59477550e-01
8.59957576e-01 -7.73659229e-01 -5.86763084e-01 3.28279197e-01
-7.15451002e-01 -9.44758486e-03 -2.76423603e-01 1.16204464e+00
-3.42825025e-01 6.96061134e-01 -7.48380542e-01 -8.10237527e-01
-5.87137580e-01 -1.32018840e+00 4.88399953e-01 3.13230634e-01
-3.63843553e-02 -8.76655817e-01 -4.25304621e-01 3.55731159e-01
8.29934001e-01 -7.08148116e-03 9.33700740e-01 -6.85566306e-01
-6.57020509e-01 -6.93970203e-01 -3.56678516e-01 3.94974291e-01
6.09252870e-01 8.90765786e-02 -6.83163464e-01 -2.60894805e-01
4.43747580e-01 -9.04018357e-02 6.10324621e-01 8.31886888e-01
1.44859445e+00 -7.20072016e-02 -4.63674605e-01 6.74352467e-01
1.13618624e+00 9.66573000e-01 4.97870594e-01 1.53313694e-03
4.48044986e-01 2.47247130e-01 7.91383207e-01 1.09577201e-01
-1.33076325e-01 2.96764046e-01 3.89034986e-01 -3.94671261e-01
-3.15540910e-01 2.85550058e-01 -2.00036481e-01 9.58392620e-01
-2.60073751e-01 -1.28664389e-01 -1.06566715e+00 6.68409228e-01
-1.50678396e+00 -9.49875653e-01 -1.51872545e-01 2.01332402e+00
7.04208732e-01 -2.01502480e-02 -1.32112280e-01 2.34183222e-01
6.83654726e-01 -2.71304697e-01 -2.39023194e-01 3.54894139e-02
4.09963489e-01 2.46086240e-01 5.47589123e-01 1.86199099e-01
-1.23585570e+00 4.02828120e-02 6.01361132e+00 7.83276975e-01
-1.81239986e+00 1.07400874e-02 6.24493241e-01 -2.12461520e-02
4.56229523e-02 -6.04806781e-01 -4.05429721e-01 5.26295066e-01
6.54566526e-01 1.77074075e-01 8.06621686e-02 6.06062412e-01
4.36950594e-01 -2.81079650e-01 -8.01169693e-01 1.00484204e+00
-7.34929135e-03 -1.29671717e+00 -4.19863433e-01 -1.46779686e-01
4.66941148e-01 -2.76821423e-02 1.24785323e-02 1.26035720e-01
-3.53139132e-01 -1.10705364e+00 -1.23149142e-01 4.95999694e-01
1.03028452e+00 -3.82872492e-01 1.41145992e+00 5.32550633e-01
-1.01179242e+00 -1.01264052e-01 -1.01340868e-01 3.84462476e-01
-2.79436082e-01 4.55946863e-01 -1.52927554e+00 3.96430224e-01
6.67261839e-01 3.30892295e-01 -4.01301235e-01 1.37839913e+00
7.17715994e-02 7.26421654e-01 -3.63887757e-01 -3.47595125e-01
1.62466764e-01 8.89911205e-02 2.44957745e-01 1.33242083e+00
7.15036988e-01 4.14509416e-01 3.80527414e-02 3.62894326e-01
2.35677451e-01 2.23323733e-01 -6.16302609e-01 -7.80498907e-02
2.83765078e-01 1.20960188e+00 -8.81778777e-01 -3.07366878e-01
-5.69860399e-01 4.35632318e-01 -4.69627291e-01 1.16901606e-01
-8.57771158e-01 -3.71743977e-01 -4.01194602e-01 3.32017303e-01
3.69237721e-01 8.70435759e-02 -1.70602262e-01 -7.26220787e-01
-8.19285884e-02 -9.25178587e-01 4.22517806e-01 -7.60349333e-01
-7.82684743e-01 5.72411656e-01 -2.10595474e-01 -1.47150183e+00
-3.47269684e-01 -6.50179565e-01 -5.79508185e-01 8.38197470e-01
-1.24298608e+00 -7.49366522e-01 -6.47398531e-01 2.68432200e-01
3.91131222e-01 -2.25531086e-01 7.66436458e-01 4.48241681e-01
-3.60268742e-01 2.92865783e-01 1.86410457e-01 2.42631495e-01
4.98844862e-01 -1.08504426e+00 -4.10133868e-01 4.76632088e-01
-4.88305777e-01 2.10361227e-01 8.86117145e-02 -5.50915599e-01
-1.29685462e+00 -1.01037550e+00 6.29561603e-01 -6.64627180e-02
1.88797578e-01 4.60287541e-01 -8.39278698e-01 2.22548723e-01
-2.75465131e-01 3.90245885e-01 7.12347865e-01 -6.47160947e-01
2.72346675e-01 -1.93140760e-01 -1.28252625e+00 4.48947281e-01
5.46148300e-01 -1.17669389e-01 -2.50262737e-01 3.27285022e-01
2.51530319e-01 -7.02572644e-01 -1.03428566e+00 7.74742723e-01
7.99444437e-01 -8.93339574e-01 8.89580250e-01 -1.74719483e-01
2.65762031e-01 -1.95853472e-01 2.81791449e-01 -8.68523538e-01
-7.69893110e-01 5.81554994e-02 -6.82299510e-02 4.87006247e-01
3.85749221e-01 -5.52677095e-01 7.99354970e-01 4.22012866e-01
-4.00076061e-02 -1.25608253e+00 -9.52594459e-01 -3.59454185e-01
-3.79132628e-01 -2.17490077e-01 1.33596227e-01 6.95802808e-01
-2.78552413e-01 -4.39079925e-02 3.59113440e-02 1.77513018e-01
1.11675002e-02 -1.08525246e-01 1.73448801e-01 -1.06182742e+00
-4.14539367e-01 -5.61066031e-01 -4.79118586e-01 -7.13906586e-01
-5.96788347e-01 -7.28077471e-01 -5.24526648e-02 -1.84751678e+00
3.11508715e-01 -7.06214309e-01 -3.95426065e-01 3.14270616e-01
-3.02398503e-01 2.08985239e-01 -1.17136784e-01 4.84409243e-01
-1.90972909e-01 -6.98465900e-03 1.89501739e+00 -2.12171078e-01
-1.11843400e-01 6.50303245e-01 -3.90827268e-01 6.68396354e-01
7.07299352e-01 -3.65366966e-01 -5.21050513e-01 -1.70367077e-01
-2.55740374e-01 6.28833115e-01 2.34469235e-01 -1.01955485e+00
1.70195490e-01 -7.97517225e-02 7.55449891e-01 -7.88986981e-01
2.00379282e-01 -1.22790349e+00 4.00897145e-01 1.19366360e+00
6.79888278e-02 -3.29479545e-01 1.31622255e-01 3.97245139e-01
-3.66828501e-01 -7.34003425e-01 1.04225838e+00 -3.72916579e-01
-4.44168925e-01 4.39175963e-01 -6.69974148e-01 -3.34872842e-01
1.23547602e+00 -4.36475486e-01 7.66545087e-02 -4.93172854e-01
-7.18406260e-01 4.30979095e-02 2.33728122e-02 8.51293211e-04
6.82654977e-01 -1.20443273e+00 -4.24773008e-01 4.85201538e-01
-2.78010726e-01 3.04565042e-01 5.69036961e-01 1.52108407e+00
-1.01237953e+00 6.34542048e-01 9.25307348e-02 -9.73238349e-01
-1.45476317e+00 4.38853681e-01 9.58892763e-01 -4.97916043e-01
-5.38145483e-01 8.26353133e-01 8.31135362e-02 -7.27984607e-02
4.74312365e-01 -4.04797912e-01 -4.23151791e-01 -7.62655959e-02
3.02169055e-01 2.71836281e-01 4.18728590e-01 -2.32846230e-01
-2.70421654e-01 6.39051437e-01 -1.18131176e-01 2.96156317e-01
9.91621017e-01 1.76025897e-01 6.70126826e-02 1.27244860e-01
1.09371734e+00 -2.81803403e-02 -7.35284984e-01 -1.88474044e-01
-3.82562399e-01 -4.89618242e-01 3.82081658e-01 -9.33513999e-01
-1.26706183e+00 1.02682674e+00 1.14566994e+00 5.38361967e-01
1.42061675e+00 -1.62018299e-01 7.71235228e-01 2.91882932e-01
-1.51599213e-01 -7.39698291e-01 1.05979189e-01 2.75875360e-01
5.70416212e-01 -1.20248234e+00 -4.55765836e-02 -4.31007355e-01
-5.51798403e-01 1.53766143e+00 5.99403381e-01 2.83426851e-01
8.59539151e-01 2.83941716e-01 3.78801614e-01 -1.90657228e-01
-6.66464806e-01 -6.23670518e-02 3.97336006e-01 5.83094895e-01
5.43536961e-01 3.02945934e-02 -4.33028430e-01 7.38342941e-01
3.78871530e-01 3.55860353e-01 2.67124802e-01 1.01478910e+00
-7.01980472e-01 -5.80465555e-01 -4.48299944e-01 9.34601247e-01
-9.53480899e-01 2.54579365e-01 6.71960935e-02 9.34261262e-01
2.63975680e-01 6.64420307e-01 -1.29559740e-01 -3.66460353e-01
3.72789949e-01 -1.45733720e-02 4.54123706e-01 -4.76926267e-01
-4.80571538e-01 3.84481609e-01 -2.31030677e-02 -6.69151545e-02
-3.77613455e-01 -3.60651135e-01 -1.19340384e+00 5.92233315e-02
-6.81079745e-01 2.32916594e-01 6.10826731e-01 7.89930463e-01
1.77621365e-01 9.00652051e-01 8.25982630e-01 -5.60071230e-01
-7.49906123e-01 -9.94378030e-01 -4.90565717e-01 2.50580430e-01
4.28741127e-01 -5.00556648e-01 -4.43866134e-01 -5.95020577e-02] | [15.23033618927002, -2.116334915161133] |
d1ccbb38-ffde-4379-a427-bce6c44fb668 | instruct2act-mapping-multi-modality | 2305.11176 | null | https://arxiv.org/abs/2305.11176v3 | https://arxiv.org/pdf/2305.11176v3.pdf | Instruct2Act: Mapping Multi-modality Instructions to Robotic Actions with Large Language Model | Foundation models have made significant strides in various applications, including text-to-image generation, panoptic segmentation, and natural language processing. This paper presents Instruct2Act, a framework that utilizes Large Language Models to map multi-modal instructions to sequential actions for robotic manipulation tasks. Specifically, Instruct2Act employs the LLM model to generate Python programs that constitute a comprehensive perception, planning, and action loop for robotic tasks. In the perception section, pre-defined APIs are used to access multiple foundation models where the Segment Anything Model (SAM) accurately locates candidate objects, and CLIP classifies them. In this way, the framework leverages the expertise of foundation models and robotic abilities to convert complex high-level instructions into precise policy codes. Our approach is adjustable and flexible in accommodating various instruction modalities and input types and catering to specific task demands. We validated the practicality and efficiency of our approach by assessing it on robotic tasks in different scenarios within tabletop manipulation domains. Furthermore, our zero-shot method outperformed many state-of-the-art learning-based policies in several tasks. The code for our proposed approach is available at https://github.com/OpenGVLab/Instruct2Act, serving as a robust benchmark for high-level robotic instruction tasks with assorted modality inputs. | ['Hongsheng Li', 'Peng Gao', 'Yu Qiao', 'Hao Dong', 'Zhengkai Jiang', 'Siyuan Huang'] | 2023-05-18 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 2.37813070e-01 7.64782727e-02 -5.47347307e-01 -2.54435003e-01
-6.82885945e-01 -7.56781578e-01 6.55139804e-01 1.62200462e-02
-2.89551169e-01 2.81053275e-01 4.16602306e-02 -4.28657591e-01
-2.69597709e-01 -6.97797775e-01 -9.36646521e-01 -3.20661217e-01
1.73280701e-01 4.74740028e-01 4.77355689e-01 -3.86210352e-01
5.75774193e-01 2.60858148e-01 -2.13561964e+00 6.56847954e-01
1.02857494e+00 8.73911083e-01 1.02934015e+00 7.92559326e-01
-3.20009649e-01 7.72449732e-01 -4.83495057e-01 2.85790283e-02
2.97340125e-01 9.25618038e-02 -8.63073409e-01 -1.59693122e-01
3.51147830e-01 -4.27962631e-01 -4.21217531e-02 9.27856922e-01
3.57823581e-01 4.19726610e-01 7.06213832e-01 -1.36115265e+00
-5.55837572e-01 9.77121413e-01 -8.42092559e-02 -3.42041641e-01
7.73192286e-01 7.25345969e-01 7.71492958e-01 -7.41229832e-01
8.09657633e-01 1.50470376e+00 2.99325995e-02 7.21607804e-01
-8.26985478e-01 -4.29467529e-01 2.53143311e-01 7.31014088e-02
-8.10156524e-01 -2.52668887e-01 3.09855819e-01 -7.00507164e-01
1.14580274e+00 3.47791642e-01 5.07257760e-01 1.26249170e+00
4.12067294e-01 1.23203039e+00 1.05845928e+00 -3.43763143e-01
2.81867087e-01 -3.62512581e-02 3.22723120e-01 6.95494056e-01
-2.61027247e-01 1.18557967e-01 -6.44259393e-01 1.25409096e-01
8.42623949e-01 -9.55650490e-03 -1.25159666e-01 -6.41874552e-01
-1.66860735e+00 3.96067142e-01 2.38502339e-01 1.05941348e-01
-4.64331925e-01 2.93108135e-01 4.39452142e-01 9.94294360e-02
-1.39988169e-01 7.11147904e-01 -4.67826664e-01 -6.15491152e-01
-3.89573991e-01 4.10304219e-01 9.15323496e-01 1.65387702e+00
6.39263153e-01 -2.26287261e-01 -7.08171189e-01 9.00272965e-01
1.45743549e-01 3.31898749e-01 4.77374673e-01 -1.61714888e+00
6.70527995e-01 5.80051482e-01 2.82825708e-01 -4.50340480e-01
-2.50366658e-01 9.53650028e-02 -7.67384171e-02 4.28264976e-01
3.38443577e-01 -1.90591943e-02 -1.13216913e+00 1.33687687e+00
4.91131514e-01 -3.05706352e-01 1.70045778e-01 1.04354930e+00
8.81484449e-01 6.64324343e-01 3.28779727e-01 2.94980586e-01
1.51785588e+00 -1.24944019e+00 -6.60499930e-01 -3.50487024e-01
4.82075155e-01 -8.43001068e-01 1.82660747e+00 6.72200024e-01
-1.10262132e+00 -8.01129222e-01 -9.62845743e-01 -3.12397987e-01
-6.42733693e-01 3.87607276e-01 8.65941584e-01 3.84905860e-02
-8.15759957e-01 5.54933846e-01 -8.98392737e-01 -5.14381647e-01
3.14420253e-01 3.58169347e-01 -5.02685010e-02 -2.15727150e-01
-6.23480439e-01 7.84068882e-01 7.75856793e-01 -2.26749092e-01
-1.30884385e+00 -5.83361328e-01 -9.27724063e-01 1.44207841e-02
8.17107916e-01 -5.15829325e-01 1.77531683e+00 -4.50859576e-01
-2.03653836e+00 6.09872818e-01 1.10162079e-01 -1.59959167e-01
3.75260532e-01 -4.18668151e-01 1.24550283e-01 8.77814889e-02
8.71302485e-02 1.18123150e+00 7.73572147e-01 -1.32397163e+00
-7.46426165e-01 -1.18238494e-01 6.71022296e-01 3.96053791e-01
1.27668343e-02 -2.23438218e-02 -6.80899560e-01 -3.14235926e-01
-1.64336711e-01 -9.94449258e-01 -6.65450320e-02 -1.99662820e-02
-3.76430482e-01 -5.31477988e-01 7.39231884e-01 -2.76677877e-01
9.54566479e-01 -2.22623825e+00 4.68507409e-01 -1.87143624e-01
-1.58627748e-01 2.14462057e-01 -3.79442841e-01 5.95470011e-01
4.78966922e-01 -2.80648261e-01 -1.14164516e-01 -1.83270350e-01
5.31341434e-01 4.93175983e-01 -4.86386269e-01 -1.34577245e-01
-4.41580489e-02 9.26455498e-01 -1.12753963e+00 -6.81801915e-01
6.75967276e-01 9.31961983e-02 -7.00780451e-01 4.63182926e-01
-9.16343391e-01 4.68293905e-01 -7.29428947e-01 8.59656811e-01
2.47632265e-01 1.16344482e-01 1.88392513e-02 4.78976499e-03
-3.35158199e-01 2.01467216e-01 -1.02490497e+00 2.49246025e+00
-7.60796189e-01 2.07049236e-01 2.61829793e-01 -6.50786042e-01
7.46434391e-01 5.60660623e-02 3.53642493e-01 -3.75720829e-01
2.30905950e-01 2.57622749e-01 -1.09886698e-01 -1.10026169e+00
7.80603290e-01 7.69826174e-01 -5.26785731e-01 3.33949268e-01
1.93507567e-01 -9.40250456e-01 5.01350343e-01 2.01306671e-01
9.19791698e-01 9.95956063e-01 1.95963562e-01 -2.00777158e-01
2.46427134e-01 4.09630001e-01 1.54335275e-01 8.95533681e-01
-3.55386078e-01 1.87987462e-01 3.74226958e-01 -1.75927162e-01
-5.62198400e-01 -1.07485545e+00 9.39890221e-02 1.77636969e+00
3.41403335e-01 -2.98318654e-01 -9.26116586e-01 -5.12787402e-01
3.94329242e-02 1.16459513e+00 -3.07116538e-01 4.33921590e-02
-4.39686805e-01 -7.42436871e-02 2.05435336e-01 4.07753497e-01
3.97223026e-01 -1.66957951e+00 -1.28016341e+00 2.91011836e-02
-1.14613831e-01 -1.22408843e+00 -5.27507544e-01 2.38378212e-01
-5.36009073e-01 -1.34243786e+00 -3.45815331e-01 -1.03348351e+00
6.26961231e-01 3.52642685e-01 9.53543544e-01 -2.05133289e-01
-3.06074709e-01 9.99238610e-01 -5.53397417e-01 -5.72712779e-01
-7.66146481e-01 1.25425607e-01 -1.81781635e-01 -4.86682296e-01
2.13123366e-01 -2.29332045e-01 -4.51124430e-01 3.01218361e-01
-1.08674788e+00 3.30600649e-01 6.27966762e-01 4.63904828e-01
6.05225623e-01 -5.15519679e-01 1.97694764e-01 -5.58784127e-01
1.09309757e+00 -3.72157514e-01 -7.42631257e-01 3.50505352e-01
-2.25107536e-01 9.28305686e-02 4.60324407e-01 -6.43522322e-01
-1.13424301e+00 4.58202600e-01 1.37250558e-01 -3.97890002e-01
-5.54184854e-01 3.65993917e-01 -7.51153082e-02 4.26367745e-02
6.61558867e-01 2.66044110e-01 3.32099125e-02 -3.17050040e-01
7.89959550e-01 1.13654208e+00 7.17462301e-01 -1.31751335e+00
2.36383200e-01 7.60739073e-02 -3.68706405e-01 -5.12960017e-01
-5.82506955e-01 -3.37944537e-01 -6.80261731e-01 -4.35004979e-01
9.78096426e-01 -6.47456288e-01 -1.07945561e+00 4.41915214e-01
-1.23305690e+00 -1.03209484e+00 -2.45800659e-01 4.32577431e-01
-1.07129049e+00 -6.38331810e-04 -7.30921745e-01 -6.36436939e-01
-2.57843584e-01 -1.88526309e+00 1.31241310e+00 3.78169090e-01
-3.42827201e-01 -3.68337810e-01 -3.11843604e-01 3.94775718e-01
3.24355781e-01 8.14606026e-02 9.52207029e-01 -4.92275596e-01
-8.45758915e-01 -8.06336626e-02 -1.76419001e-02 1.02463335e-01
2.95985453e-02 1.37492180e-01 -7.29436517e-01 2.62920652e-02
-4.75100607e-01 -8.79331470e-01 4.68236208e-01 1.84184819e-01
1.61373007e+00 -1.40621349e-01 -3.03642124e-01 4.12614107e-01
1.05142951e+00 4.35552895e-01 4.67297673e-01 3.37936074e-01
5.71881354e-01 8.00989151e-01 1.28882384e+00 3.60019207e-01
6.00695550e-01 8.15785944e-01 7.83095181e-01 4.65543598e-01
-7.61745572e-02 -3.07038873e-01 5.78984499e-01 6.95998132e-01
4.73525748e-02 -8.19479525e-02 -9.45016205e-01 5.94703853e-01
-2.09824467e+00 -7.28525341e-01 1.29423350e-01 2.01793981e+00
9.38871503e-01 5.28767183e-02 -9.85123813e-02 -4.88843381e-01
3.52419883e-01 2.28977371e-02 -8.15688312e-01 -7.19258010e-01
5.40229619e-01 1.04952805e-01 2.65944809e-01 3.78258437e-01
-1.10520971e+00 1.42871118e+00 5.76812220e+00 9.89525139e-01
-9.90240157e-01 4.54413034e-02 7.34930560e-02 -8.17564130e-02
-5.31818578e-03 -2.08835825e-01 -7.94668257e-01 3.94542545e-01
5.61723769e-01 3.66967767e-02 8.53125274e-01 1.19672561e+00
2.61659622e-01 -4.76950377e-01 -1.30732489e+00 6.84101939e-01
-8.09859484e-02 -1.23845696e+00 4.85543869e-02 -2.33017132e-01
6.21257901e-01 2.61031091e-01 8.65793824e-02 8.91492605e-01
7.46944785e-01 -7.98113525e-01 1.02741301e+00 4.30306077e-01
6.99091017e-01 -3.64752799e-01 1.40213922e-01 6.16461873e-01
-9.97639775e-01 -4.06906784e-01 -1.98870972e-01 -9.06568617e-02
1.93898063e-02 -1.72798872e-01 -8.30423176e-01 3.72498959e-01
7.40276039e-01 4.97181922e-01 -9.92240980e-02 8.51636648e-01
-5.77907801e-01 6.06572106e-02 -1.41393557e-01 -1.49218649e-01
2.47666150e-01 -9.74267945e-02 4.82035071e-01 1.04272187e+00
3.11681777e-01 3.75919752e-02 6.65851235e-01 9.95299876e-01
1.59737885e-01 1.39987633e-01 -6.11820042e-01 -1.26330942e-01
7.41259277e-01 1.28757811e+00 -6.25928760e-01 -5.33483684e-01
-3.00237864e-01 8.55177045e-01 2.90268898e-01 5.33642530e-01
-9.34903562e-01 -6.21178925e-01 8.20410311e-01 -1.43035114e-01
2.80732661e-01 -5.49332023e-01 -1.25589043e-01 -8.77595127e-01
-8.82484242e-02 -1.13605666e+00 1.06890462e-01 -1.13190031e+00
-6.90587997e-01 4.75932330e-01 6.09229624e-01 -1.23761272e+00
-3.03577721e-01 -1.00004435e+00 -5.50656378e-01 4.21062440e-01
-1.33309019e+00 -1.23294199e+00 -5.61682642e-01 5.72256029e-01
1.06498218e+00 -9.92483422e-02 8.53454828e-01 -1.37756750e-01
-6.26360953e-01 1.28550738e-01 -3.42448950e-01 -1.97803155e-01
6.62301540e-01 -1.18603826e+00 1.23335846e-01 5.78738213e-01
-2.26562634e-01 7.88755178e-01 6.42940521e-01 -6.49803221e-01
-1.87316179e+00 -8.60062182e-01 3.35310586e-03 -4.31767762e-01
7.40080416e-01 -4.46672380e-01 -6.12777472e-01 8.54983926e-01
3.88082951e-01 -1.35400802e-01 2.00185269e-01 -3.38676751e-01
-3.95474061e-02 2.11045057e-01 -1.06543636e+00 7.54806757e-01
1.21237540e+00 -2.77753174e-01 -7.94372678e-01 4.75978434e-01
1.01742482e+00 -1.11721098e+00 -9.25987363e-01 2.62815446e-01
5.98716319e-01 -7.03745067e-01 9.11440015e-01 -4.25367832e-01
8.48486006e-01 -3.33474964e-01 -1.76850751e-01 -1.24778354e+00
5.15654534e-02 -6.54084265e-01 -3.16015542e-01 8.47346067e-01
2.71957338e-01 -5.60612142e-01 3.71280044e-01 7.20868707e-01
-8.05118322e-01 -8.76022756e-01 -6.18772566e-01 -6.55517220e-01
-1.56082511e-01 -6.97599351e-01 7.90722191e-01 5.10783851e-01
3.24155331e-01 -1.43615022e-01 6.33697137e-02 7.48987347e-02
2.75548667e-01 3.92694116e-01 1.20296681e+00 -8.15615356e-01
-4.09505427e-01 -4.92890686e-01 9.12397131e-02 -1.38018775e+00
3.08700949e-01 -1.01435018e+00 6.35123014e-01 -1.73784637e+00
-9.98692065e-02 -4.77430850e-01 -6.80058450e-02 9.26760256e-01
3.79518270e-02 -3.00556421e-01 5.82145095e-01 2.79728472e-01
-6.96512163e-01 4.49561238e-01 1.69088662e+00 -3.86900306e-01
-6.16863966e-01 -1.78630743e-02 -4.20412928e-01 6.72393978e-01
1.04199445e+00 -1.09891267e-02 -5.77957451e-01 -7.84991384e-01
-9.87232924e-02 -2.29864046e-02 3.25108171e-01 -1.17938745e+00
2.66765177e-01 -6.63059592e-01 -2.47140869e-01 -4.24336165e-01
5.09898841e-01 -7.95553863e-01 -3.26440066e-01 4.43348080e-01
-5.33006847e-01 -7.69804716e-02 3.47696573e-01 2.92093128e-01
1.26680672e-01 -3.41286361e-01 3.09111834e-01 -4.27620620e-01
-1.29421866e+00 1.57375224e-02 -4.92622584e-01 -2.11184710e-01
1.44115746e+00 -1.08765200e-01 -5.81964731e-01 1.89669412e-02
-5.77966750e-01 5.75393677e-01 5.21021485e-01 1.01501799e+00
6.98306143e-01 -8.60534310e-01 -1.98651806e-01 2.39848733e-01
4.50396031e-01 2.52171874e-01 6.60650879e-02 7.60032058e-01
-6.35450125e-01 5.84591925e-01 -4.19303507e-01 -7.37509668e-01
-1.05587709e+00 5.96708536e-01 1.45854026e-01 -1.55388806e-02
-4.38417524e-01 7.06141233e-01 1.05974630e-01 -8.92740250e-01
5.23604095e-01 -9.95719433e-01 -1.34005964e-01 -2.86976099e-01
3.67296696e-01 2.04246938e-01 -2.93374777e-01 -2.52309412e-01
-6.52097687e-02 4.96035486e-01 9.22264084e-02 -3.79320122e-02
1.10584021e+00 1.44234762e-01 -6.61119521e-02 4.81597275e-01
5.64239204e-01 -2.39207998e-01 -1.64143050e+00 1.23959340e-01
-1.56391948e-01 -2.39098847e-01 -1.35319263e-01 -1.06170809e+00
-6.98196411e-01 8.33946407e-01 3.16059202e-01 -1.39875397e-01
7.59093821e-01 9.14427340e-02 6.96663618e-01 6.56475544e-01
1.03557503e+00 -1.23139405e+00 1.39707029e-01 7.50334442e-01
1.07718945e+00 -1.28230989e+00 -2.29243457e-01 -5.07851362e-01
-7.84147501e-01 1.15423560e+00 1.14385009e+00 1.28821924e-01
1.21119007e-01 1.41065910e-01 1.26442714e-02 -2.53857672e-01
-7.00426400e-01 -2.14263439e-01 1.62534669e-01 6.75303459e-01
2.64954001e-01 3.60185891e-01 -3.14659953e-01 5.38307250e-01
-2.53041387e-01 3.91932845e-01 5.33737302e-01 1.43818629e+00
-5.64083457e-01 -1.05627036e+00 -3.94732803e-01 2.91178167e-01
7.80034736e-02 8.61553252e-02 -1.30378753e-01 8.05492520e-01
1.26054436e-01 9.11921084e-01 -7.88769871e-02 -4.34906125e-01
5.17267227e-01 3.37562896e-02 6.18220091e-01 -1.06343353e+00
-6.76884294e-01 -3.04514080e-01 2.86848247e-02 -1.11964917e+00
-2.83814311e-01 -5.11403441e-01 -1.55306840e+00 1.81087360e-01
2.28251234e-01 -1.85592443e-01 9.36465979e-01 8.20415974e-01
6.94492102e-01 8.15651000e-01 6.43924531e-03 -1.55648780e+00
-7.27517009e-01 -1.06076646e+00 1.12647966e-01 2.77478427e-01
-1.02439553e-01 -8.68855000e-01 7.36929923e-02 1.20972291e-01] | [4.543042182922363, 0.767650842666626] |
13e83562-5c49-429c-8513-d92f2f84f18f | how-well-can-text-to-image-generative-models | 2210.15230 | null | https://arxiv.org/abs/2210.15230v1 | https://arxiv.org/pdf/2210.15230v1.pdf | How well can Text-to-Image Generative Models understand Ethical Natural Language Interventions? | Text-to-image generative models have achieved unprecedented success in generating high-quality images based on natural language descriptions. However, it is shown that these models tend to favor specific social groups when prompted with neutral text descriptions (e.g., 'a photo of a lawyer'). Following Zhao et al. (2021), we study the effect on the diversity of the generated images when adding ethical intervention that supports equitable judgment (e.g., 'if all individuals can be a lawyer irrespective of their gender') in the input prompts. To this end, we introduce an Ethical NaTural Language Interventions in Text-to-Image GENeration (ENTIGEN) benchmark dataset to evaluate the change in image generations conditional on ethical interventions across three social axes -- gender, skin color, and culture. Through ENTIGEN framework, we find that the generations from minDALL.E, DALL.E-mini and Stable Diffusion cover diverse social groups while preserving the image quality. Preliminary studies indicate that a large change in the model predictions is triggered by certain phrases such as 'irrespective of gender' in the context of gender bias in the ethical interventions. We release code and annotated data at https://github.com/Hritikbansal/entigen_emnlp. | ['Kai-Wei Chang', 'Masoud Monajatipoor', 'Da Yin', 'Hritik Bansal'] | 2022-10-27 | null | null | null | null | ['culture'] | ['speech'] | [ 3.85827094e-01 5.89360178e-01 -9.29298252e-02 -5.19852519e-01
-2.25464150e-01 -6.71382487e-01 1.04928637e+00 -1.70016184e-01
-3.05332899e-01 8.54474902e-01 6.90468609e-01 -1.40719235e-01
1.79298714e-01 -8.95804167e-01 -8.21506917e-01 -6.55154228e-01
5.34229279e-01 2.37794831e-01 -4.89571750e-01 -3.11011851e-01
5.00056922e-01 -1.12556329e-03 -1.41701818e+00 3.60448480e-01
9.51751471e-01 3.63224894e-01 -6.29602820e-02 6.93725586e-01
1.47596374e-01 8.55056643e-01 -7.07591712e-01 -1.04985321e+00
3.26159775e-01 -9.27268982e-01 -6.86245441e-01 4.32877289e-03
6.89583480e-01 -5.40674090e-01 -1.90989375e-01 1.18787491e+00
8.17116559e-01 -2.08298698e-01 9.55621481e-01 -1.71426606e+00
-1.48422897e+00 8.24097335e-01 -7.68726885e-01 -2.52052844e-01
5.77786922e-01 4.63626325e-01 8.74372721e-01 -4.83093560e-01
1.30625880e+00 1.76443195e+00 4.33026344e-01 9.17662323e-01
-1.33428466e+00 -9.18792129e-01 -4.00097370e-02 -3.04047354e-02
-1.07179439e+00 -4.55358148e-01 7.16105342e-01 -6.99658155e-01
5.16435146e-01 5.54374814e-01 7.09809899e-01 1.83067787e+00
4.95762080e-01 3.14928263e-01 1.57140005e+00 -4.18180406e-01
-7.67375305e-02 4.22550231e-01 -2.05516726e-01 4.93777186e-01
3.68347406e-01 1.16566680e-01 -6.87502563e-01 -1.74789369e-01
5.62679589e-01 -4.92736250e-01 4.44064140e-02 2.65057206e-01
-1.21209943e+00 9.63629305e-01 3.64532948e-01 3.31610143e-02
-3.47287059e-01 3.32475662e-01 2.14297265e-01 1.12820469e-01
7.53490865e-01 4.08945829e-01 2.17494398e-01 -3.60199320e-03
-6.36338353e-01 4.91535813e-01 5.24377823e-01 8.77947032e-01
6.26103401e-01 -2.05504186e-02 -5.95498860e-01 6.71992838e-01
2.80303448e-01 8.59617889e-01 4.04974967e-02 -1.34556174e+00
1.56645834e-01 3.55513602e-01 2.05408543e-01 -1.42401588e+00
-3.83779407e-02 -8.10560286e-02 -9.90458012e-01 9.99879390e-02
2.11035639e-01 -5.05108476e-01 -7.90013790e-01 2.03471780e+00
1.70919493e-01 -3.36182117e-01 -1.25471771e-01 7.95105815e-01
9.03723240e-01 7.99193740e-01 6.06820226e-01 -2.39737242e-01
1.32064342e+00 -4.00547534e-01 -7.19710767e-01 -2.77277499e-01
3.71899039e-01 -1.07863867e+00 1.23439872e+00 2.85705030e-02
-1.04730844e+00 -4.76513058e-01 -4.66640383e-01 -4.79292609e-02
-3.10051054e-01 8.63916799e-02 4.71290886e-01 9.02325928e-01
-1.30069673e+00 3.30104083e-01 9.88415703e-02 -9.33341980e-01
6.43713951e-01 -1.80210441e-01 -1.58415496e-01 -5.75927012e-02
-1.14071178e+00 6.74934030e-01 8.61127004e-02 -1.81658402e-01
-8.86990011e-01 -5.85221231e-01 -6.45126522e-01 -3.52764398e-01
-4.79871267e-03 -1.02255678e+00 9.40778077e-01 -1.57652962e+00
-1.09474468e+00 1.42601085e+00 -4.26079035e-02 -2.78967172e-01
1.02001488e+00 -1.50111511e-01 -2.46743977e-01 1.47903815e-01
5.37189901e-01 1.60280144e+00 9.27034020e-01 -1.65750408e+00
-3.72995406e-01 -1.59446374e-01 2.50172734e-01 3.52631480e-01
-3.46358955e-01 5.52686036e-01 2.84773201e-01 -6.97160661e-01
-5.75795650e-01 -1.20347643e+00 -1.17405638e-01 -1.39420733e-01
-8.66342485e-01 -2.02038571e-01 3.67319405e-01 -7.29709268e-01
1.09912336e+00 -2.08412194e+00 -4.97524701e-02 3.62794138e-02
2.14275181e-01 -2.24377021e-01 -8.06493908e-02 6.34174287e-01
-1.16203621e-01 8.90694201e-01 -1.35637280e-02 -1.80436790e-01
3.80206198e-01 8.25058445e-02 -2.26100713e-01 3.70327979e-01
7.24843293e-02 9.10437703e-01 -7.54051328e-01 -8.76454055e-01
-7.89108649e-02 4.32105988e-01 -6.72536075e-01 -8.17091204e-03
-3.28193791e-02 5.29815137e-01 -1.05787069e-02 3.92476678e-01
7.27348566e-01 -1.51904985e-01 -1.75765585e-02 -6.46621138e-02
-2.39082828e-01 -3.16910893e-01 -5.73074520e-01 1.04167867e+00
2.48339381e-02 8.21421802e-01 -1.03631794e-01 -2.84876108e-01
7.34879494e-01 1.96552902e-01 1.22664884e-01 -6.41387880e-01
2.41924122e-01 2.84815021e-02 2.62723714e-01 -4.40718323e-01
5.93351305e-01 -3.87400687e-01 -2.83229381e-01 5.98293304e-01
-4.08408880e-01 -3.44579726e-01 3.88984263e-01 6.18032694e-01
4.94150639e-01 1.31308153e-01 1.15168512e-01 -4.64795381e-01
1.18794246e-02 2.01117564e-02 5.06138206e-01 9.31355536e-01
-3.77645254e-01 6.59899294e-01 8.39453578e-01 -2.15117022e-01
-1.45881510e+00 -1.00175118e+00 2.63282415e-02 1.30813777e+00
8.54018554e-02 -2.08560213e-01 -1.02769542e+00 -3.84621382e-01
-1.45329893e-01 1.26042867e+00 -9.35804546e-01 -1.92586035e-01
-1.61336526e-01 -8.59691978e-01 5.67928016e-01 -1.54488988e-03
7.34865725e-01 -1.29184496e+00 -8.00785303e-01 -2.78443664e-01
-3.24917555e-01 -9.94382143e-01 -5.15889764e-01 -7.93740630e-01
-2.64662027e-01 -7.43957043e-01 -7.58364260e-01 -4.62655127e-01
1.04616189e+00 -1.97480723e-01 1.15145719e+00 1.16862476e-01
-3.73705298e-01 6.33604765e-01 -3.97595346e-01 -6.84178174e-01
-7.99177170e-01 -1.38960421e-01 -1.79502025e-01 -6.72744308e-03
-7.25644529e-02 -3.47196192e-01 -9.22737777e-01 2.53321975e-01
-9.38585341e-01 7.28902161e-01 3.42575401e-01 4.50934231e-01
1.24650629e-04 -2.53400505e-01 6.17770791e-01 -1.05252779e+00
9.97468531e-01 -5.48428178e-01 1.68866850e-02 2.28208363e-01
-7.33114362e-01 -3.96308899e-01 3.29764605e-01 -3.90487790e-01
-1.42773688e+00 -3.64511341e-01 1.65074989e-01 1.03324145e-01
-3.19797546e-01 1.33091182e-01 1.56183183e-01 2.60426462e-01
8.54803026e-01 6.58265203e-02 -1.81912363e-01 2.53861070e-01
3.24176997e-01 6.22723162e-01 2.78428018e-01 -7.10004330e-01
8.85776281e-01 5.07180274e-01 -1.12412624e-01 -6.48382366e-01
-5.43410838e-01 2.98754990e-01 -1.44475624e-01 -9.56142068e-01
1.15710330e+00 -7.91808009e-01 -5.93012154e-01 7.05893934e-01
-1.20995247e+00 -4.73560750e-01 -1.69826120e-01 2.92090595e-01
-4.91895169e-01 1.97052613e-01 -5.15692949e-01 -8.49232137e-01
-4.19994801e-01 -8.49793196e-01 9.94507730e-01 5.06605089e-01
-5.81101894e-01 -8.25341582e-01 -1.00888416e-01 6.36527896e-01
3.34476918e-01 7.44848609e-01 1.02859879e+00 -1.70917779e-01
-3.51582378e-01 1.64943799e-01 -4.66008514e-01 2.69921720e-01
5.63776568e-02 7.20723867e-01 -7.68654287e-01 -6.65363297e-02
-4.14203912e-01 -3.98038536e-01 4.07514364e-01 6.24596536e-01
6.81956232e-01 -8.17139506e-01 -1.37213737e-01 2.94743419e-01
1.37323999e+00 1.46050274e-01 9.93691862e-01 1.97197795e-01
6.40559256e-01 9.02529180e-01 6.35250509e-01 6.65692687e-01
5.65981030e-01 9.51674804e-02 2.18806639e-01 -1.76305473e-01
-7.18403384e-02 -6.43192530e-01 5.20369172e-01 2.42436960e-01
-3.33695620e-01 -5.85237801e-01 -9.86035526e-01 5.35009861e-01
-1.63793874e+00 -1.18499041e+00 -2.90571839e-01 1.73083627e+00
8.63797188e-01 -1.65961847e-01 2.53649410e-02 -4.88266975e-01
1.03309631e+00 3.74616742e-01 -2.81239718e-01 -8.22516382e-01
-5.01753926e-01 -2.71874577e-01 4.33349431e-01 5.82286775e-01
-6.59353793e-01 1.00116146e+00 6.18966579e+00 8.75090301e-01
-9.08733428e-01 6.99305013e-02 1.59061944e+00 -1.11584917e-01
-8.72705281e-01 1.88029841e-01 -5.91165245e-01 7.05677211e-01
7.24445641e-01 -6.58926487e-01 3.99416029e-01 4.03056502e-01
6.27282262e-01 -4.13893491e-01 -8.18774521e-01 8.68283153e-01
3.71898323e-01 -1.04010904e+00 1.95951819e-01 1.48244694e-01
1.06063855e+00 -5.82682133e-01 5.48136234e-01 1.04714455e-02
5.69880664e-01 -1.22968543e+00 1.07386518e+00 5.19020379e-01
1.09883118e+00 -7.12551892e-01 3.28729838e-01 1.98153883e-01
-2.89347231e-01 -1.86034311e-02 -2.99385101e-01 -1.14153966e-01
2.06900030e-01 4.20119524e-01 -9.50845897e-01 1.84253469e-01
7.97895193e-01 4.06894475e-01 -7.97338367e-01 2.16175407e-01
-4.06518131e-01 5.58439910e-01 8.18650275e-02 -8.73551145e-03
7.38664269e-02 -2.90041149e-01 6.21439934e-01 1.11462331e+00
5.97913861e-01 2.39561230e-01 -3.80616218e-01 1.08352423e+00
-1.31063268e-01 3.97528529e-01 -9.23659444e-01 -2.22079024e-01
2.54682362e-01 1.23465812e+00 -7.84860611e-01 -3.07555199e-01
4.12370265e-02 9.44900036e-01 -1.17927924e-01 5.10131121e-01
-1.05638647e+00 8.14765319e-02 3.09291869e-01 5.57742298e-01
-4.19590265e-01 2.20095441e-01 -3.75915647e-01 -6.84409201e-01
-4.67151374e-01 -1.10680532e+00 4.49968651e-02 -1.53078687e+00
-1.23941004e+00 3.58620644e-01 3.55193853e-01 -6.98737860e-01
-2.39457816e-01 -1.96271479e-01 -5.34194529e-01 7.27237403e-01
-8.69460821e-01 -1.57946134e+00 -4.01341289e-01 3.13194185e-01
2.33857587e-01 -7.10415244e-02 3.96221638e-01 8.55588466e-02
-5.27107239e-01 4.46710616e-01 -3.56328756e-01 -1.62346885e-02
1.23870730e+00 -9.97514486e-01 1.67063892e-01 8.11675489e-01
-3.38924140e-01 5.01966119e-01 1.07890677e+00 -1.08362854e+00
-8.72008145e-01 -9.20809805e-01 1.11577749e+00 -3.55971128e-01
3.02425116e-01 -3.25603068e-01 -1.36991590e-01 8.25647891e-01
1.13205099e+00 -6.47931576e-01 8.18883300e-01 -3.17377329e-01
-1.62394613e-01 1.39177755e-01 -1.44117022e+00 1.26293671e+00
1.40023637e+00 -2.68508226e-01 -6.45302236e-02 4.70573753e-01
5.72604775e-01 -5.77386189e-03 -6.16085887e-01 2.07245559e-01
5.74366510e-01 -1.35743129e+00 7.47281373e-01 -4.84211802e-01
1.18757772e+00 6.93852752e-02 -8.11677873e-02 -1.36271489e+00
-4.60003346e-01 -8.45112681e-01 7.53817022e-01 1.67208993e+00
3.48766178e-01 -7.43724346e-01 4.38284010e-01 9.13059652e-01
2.50832021e-01 -3.52575749e-01 -6.69912457e-01 -2.34231859e-01
1.97651073e-01 1.06345555e-02 5.82924545e-01 1.06909251e+00
-2.99098969e-01 2.99693674e-01 -7.68157959e-01 -1.93559498e-01
6.21340275e-01 -9.16117430e-02 9.24624741e-01 -8.56914043e-01
8.55783150e-02 -4.42862511e-01 -1.22964650e-01 -1.15547501e-01
2.74181604e-01 -7.25114524e-01 -1.01607718e-01 -1.67070198e+00
7.75970221e-01 -1.16064623e-01 2.32140467e-01 3.28871965e-01
-4.05228823e-01 4.27998722e-01 5.36590397e-01 7.94176459e-02
-4.28801507e-01 2.67433882e-01 1.60786152e+00 1.03182057e-02
2.48947263e-01 -5.50983846e-01 -1.27432454e+00 6.63176239e-01
1.13440168e+00 -5.14472604e-01 -5.48712015e-01 -3.69404912e-01
8.37957799e-01 -2.00252369e-01 7.61477470e-01 -4.85484928e-01
-2.23704621e-01 -6.01728320e-01 3.78491849e-01 -6.29428476e-02
4.76870611e-02 -4.23311383e-01 7.10278809e-01 5.40324390e-01
-5.64070702e-01 1.63796857e-01 -4.72526270e-04 1.53605253e-01
1.40096664e-01 -1.92962497e-01 8.10737610e-01 -2.82054335e-01
-2.33915478e-01 -2.16067340e-02 -4.15305793e-01 2.75551975e-01
1.16321158e+00 -3.47609848e-01 -9.50807631e-01 -9.79142308e-01
-3.47666085e-01 -7.79260788e-03 8.92566741e-01 3.91973913e-01
5.00781238e-01 -1.39989650e+00 -1.31046438e+00 -2.87545413e-01
-1.49776787e-02 -4.47356075e-01 5.62773764e-01 7.72875249e-01
-6.27280414e-01 -1.96817257e-02 -6.07831538e-01 -3.53206396e-01
-1.45276511e+00 1.98078990e-01 8.86930525e-02 3.30026299e-02
6.37195855e-02 9.37530637e-01 7.19720244e-01 -2.12038204e-01
-4.30554897e-01 8.45367610e-02 -1.06429696e-01 4.13283288e-01
6.35893568e-02 4.98166054e-01 -8.21746886e-01 -1.04407847e+00
-1.62643045e-01 4.75772142e-01 1.80790305e-01 -4.41561818e-01
1.22560894e+00 -2.95439184e-01 -3.78320396e-01 2.07436547e-01
9.12213087e-01 1.31556526e-01 -1.06149626e+00 5.46498060e-01
-5.58870971e-01 -9.09061849e-01 -5.10065258e-01 -9.52219069e-01
-1.01690054e+00 5.36089301e-01 6.26845121e-01 2.98416644e-01
1.09310472e+00 -5.93723357e-03 3.10602337e-01 -2.73589849e-01
9.77363586e-02 -1.27714431e+00 2.50591278e-01 -5.40936440e-02
1.43755925e+00 -1.25631201e+00 1.50863409e-01 -3.97215873e-01
-1.20208478e+00 4.60794240e-01 7.77885914e-01 8.00141096e-02
2.69679755e-01 -1.41463280e-01 1.66430324e-01 -9.98933613e-02
-8.47694576e-01 2.77580738e-01 -1.13290586e-02 8.26319456e-01
5.59278727e-01 4.32620138e-01 -7.64109612e-01 2.98597068e-01
-6.70977056e-01 -1.34742841e-01 8.30883443e-01 4.77900416e-01
-9.39514562e-02 -8.86426747e-01 -5.77580154e-01 4.56927717e-01
-5.99107146e-01 -2.31359482e-01 -8.99807453e-01 7.22860992e-01
6.51047826e-01 9.35773194e-01 -7.12743867e-03 -1.45382896e-01
-6.16282411e-03 -1.13393180e-02 4.18783754e-01 -4.53000396e-01
-7.84918785e-01 1.39263093e-01 4.45213228e-01 -1.39113292e-01
-6.12805426e-01 -9.51040328e-01 -7.59207904e-01 -7.63325036e-01
3.17588538e-01 -2.80702889e-01 4.28850293e-01 4.12044078e-01
4.47794110e-01 2.26263613e-01 5.94468355e-01 -5.05791426e-01
-4.61031795e-02 -9.45749283e-01 -4.27971721e-01 8.80707383e-01
-2.06560373e-01 -2.13497579e-01 -4.95738238e-01 5.06414115e-01] | [12.087080955505371, 0.9470548033714294] |
7a20dab2-f9c9-40b7-9784-787965850883 | in-situ-3d-spatiotemporal-measurement-of | 2209.01164 | null | https://arxiv.org/abs/2209.01164v3 | https://arxiv.org/pdf/2209.01164v3.pdf | In Situ 3D Spatiotemporal Measurement of Soluble Biomarkers in Organoid Culture | Advanced cell culture techniques such as 3D bio-printing and hydrogel-based cell embedding techniques harbor many new and exciting opportunities to study cells in environments that closely recapitulate in-vivo conditions. Researchers often study these environments using fluorescence microscopy to visualize the protein association with objects such as cells within the 3D environment, yet quantification of concentration profiles in the microenvironment has remained elusive. Here, we present a method to continuously measure the time-dependent concentration gradient of various biomarkers within a 3D cell culture assay using bead-based immunoassays to sequester and concentrate the fluorescence intensity of these tagged proteins. This assay allows for near real-time in situ biomarker detection and enables spatiotemporal quantification of biomarker concentration. Snapshots of concentration profiles can be taken, or time series analysis can be performed enabling time-varying biomarker production estimation. Example assays utilize an osteosacroma tumoroid as a case study for a quantitative single-plexed gel encapsulated assay, and a qualitative multi-plexed 3D bioprinted assay. In both cases, a time-varying cytokine concentration gradient is measured. An estimation for the production rate of the IL-8 cytokine per second per osteosarcoma cell results from fitting an analytical function for continuous point source diffusion to the measured concentration gradient and reveals that each cell produces approximately two IL-8 cytokines per second. Proper calibration and use of this assay is exhaustively explored for the case of diffusion-limited Langmuir kinetics of a spherical adsorber. | ['W. Gregory Sawyer', 'Jack E Famiglietti', 'Eric O McGhee', 'Alexander J McGhee'] | 2022-09-02 | null | null | null | null | ['culture'] | ['speech'] | [ 1.89403445e-02 -5.79885066e-01 8.11098740e-02 4.36898887e-01
-7.58236587e-01 -6.97391987e-01 2.84765065e-01 5.49249232e-01
-4.76725787e-01 9.44403052e-01 -4.85426746e-02 -1.01607665e-01
6.51058853e-02 -6.37228966e-01 -5.29734313e-01 -1.20837784e+00
-2.76869118e-01 6.76537275e-01 3.13184381e-01 1.77912906e-01
6.24251552e-02 1.01533616e+00 -1.11644077e+00 1.54458908e-02
2.46646613e-01 8.60666454e-01 2.57638812e-01 9.35993135e-01
-3.58175874e-01 5.52759588e-01 -6.67879939e-01 3.54325086e-01
-2.91618612e-02 -1.30856410e-01 8.80789459e-02 2.13311240e-01
-2.48093247e-01 -1.72511369e-01 1.47794336e-01 3.59227568e-01
3.79020572e-01 -4.51285571e-01 9.27896559e-01 -8.56111944e-01
-1.32184386e-01 -4.04162079e-01 -3.89447987e-01 -8.18559974e-02
5.73918998e-01 5.06898649e-02 -9.70811695e-02 -9.62056220e-01
9.85583365e-01 1.00735807e+00 5.82314432e-01 4.58220392e-01
-1.51422179e+00 -2.50498772e-01 -5.12063980e-01 -7.67362118e-01
-1.22756290e+00 -1.66605338e-01 3.67039204e-01 -9.37610805e-01
5.41065574e-01 5.69653988e-01 9.87339795e-01 7.37617254e-01
1.25240278e+00 2.02294409e-01 1.66852689e+00 -1.66362911e-01
5.37804961e-01 1.80712223e-01 -9.12641883e-02 7.74801910e-01
3.65543008e-01 2.96756983e-01 -2.58613139e-01 -3.86964440e-01
1.14205289e+00 6.53908074e-01 -1.70072868e-01 -3.47387433e-01
-1.16408241e+00 2.62129277e-01 -2.33659014e-01 6.86083615e-01
-2.71169662e-01 3.64174932e-01 8.82562771e-02 1.42353728e-01
5.45366704e-01 3.13734561e-01 -2.40462005e-01 -2.68069059e-01
-8.02959263e-01 3.76112938e-01 9.24047709e-01 3.38700384e-01
4.47679907e-01 -5.06076574e-01 -1.14478856e-01 5.16530633e-01
4.10565913e-01 9.28058982e-01 3.37058902e-01 -9.77360547e-01
-3.72948796e-01 5.15721381e-01 9.25345242e-01 -6.18198931e-01
-4.64513451e-01 -2.15338901e-01 -1.52963698e-01 4.37407821e-01
1.09941185e+00 -3.28589439e-01 -6.92163885e-01 1.13696265e+00
6.28432930e-01 -7.18403235e-02 3.17911245e-02 9.81149316e-01
1.34989589e-01 3.60176146e-01 -8.51790770e-04 -7.73046017e-01
1.36359584e+00 -1.61863163e-01 -6.75183356e-01 3.21232677e-01
6.78208768e-01 -5.06079555e-01 1.01998854e+00 5.18218055e-02
-8.61196756e-01 1.81535989e-01 -9.00763512e-01 3.21704924e-01
-5.52810550e-01 -2.95722187e-01 4.40780699e-01 5.66979051e-01
-6.84040308e-01 6.02164805e-01 -1.18283534e+00 -6.71377838e-01
2.57933885e-01 3.05639625e-01 -2.64006406e-01 -2.98427463e-01
-5.98669052e-01 6.28711522e-01 -5.42106092e-01 -8.00668970e-02
-7.30658770e-01 -9.82748389e-01 -4.31305051e-01 -2.24629283e-01
-6.41140223e-01 -6.51784062e-01 7.49270380e-01 -1.30832508e-01
-2.02160907e+00 9.91393745e-01 -3.63265157e-01 -6.70784712e-02
4.07064497e-01 1.06609590e-01 -8.91683921e-02 2.33294651e-01
2.29145780e-01 3.97932231e-02 2.00824440e-02 -1.31812823e+00
-8.72527212e-02 -6.98153377e-01 -5.06219149e-01 -1.03730358e-01
4.08241972e-02 2.18696728e-01 1.88603565e-01 -1.82626039e-01
2.72824645e-01 -8.83014858e-01 -3.27359974e-01 2.40037113e-01
4.21859398e-02 3.78859758e-01 7.88843453e-01 -6.06238879e-02
6.06050968e-01 -1.86958420e+00 -1.83727607e-01 4.84377705e-02
2.67172098e-01 -3.70414168e-01 1.46766245e-01 1.11906779e+00
5.40254593e-01 1.00163318e-01 -1.07822262e-01 -1.41481785e-02
-9.57118347e-02 -1.86774924e-01 1.68663993e-01 9.63438451e-01
2.71250103e-02 7.89929509e-01 -1.14568591e+00 -4.50761765e-01
-9.91055369e-02 7.36043751e-01 1.50382295e-01 1.33177161e-01
-3.05663109e-01 7.45597839e-01 -5.18656731e-01 1.04916239e+00
5.94527721e-01 -3.57602894e-01 3.12548757e-01 1.05619989e-01
-4.23193723e-01 -4.34946477e-01 -5.33121943e-01 1.21116257e+00
-1.01421997e-01 3.67596388e-01 4.87693846e-01 -2.14143828e-01
1.17381811e+00 2.38048062e-01 9.96636569e-01 -7.32497096e-01
1.83674917e-01 7.83084869e-01 -4.57565784e-01 -6.27904594e-01
3.56174482e-04 -1.10968971e+00 4.01410274e-02 4.68266219e-01
-6.24225795e-01 2.06398237e-02 9.40962657e-02 -4.17218544e-02
1.45915544e+00 -1.88262254e-01 -3.64857197e-01 -5.89317560e-01
3.10363621e-01 2.25136280e-01 -5.35973050e-02 3.69468361e-01
-2.61763752e-01 3.20722699e-01 6.07648432e-01 -2.60806918e-01
-7.76860476e-01 -1.10533643e+00 -3.81706804e-01 3.56015891e-01
4.03499275e-01 1.20033212e-01 -3.50300401e-01 2.84657907e-02
8.38579476e-01 -1.77343220e-01 -7.31659591e-01 4.01118606e-01
-2.13195726e-01 -6.58628047e-01 4.39262450e-01 1.11153327e-01
-1.08357266e-01 -1.60779044e-01 -7.67128825e-01 6.44545019e-01
3.48936349e-01 -8.44873190e-01 7.19349906e-02 1.20757446e-01
-1.11765075e+00 -1.21599281e+00 -1.15970635e+00 -4.58628297e-01
6.51084661e-01 -1.75282359e-02 5.34177482e-01 2.71880142e-02
-4.50142860e-01 8.95860255e-01 -4.93069068e-02 -5.75070202e-01
-3.41320872e-01 -6.34181321e-01 8.94368440e-02 -7.90278539e-02
4.54781950e-01 -4.69918400e-01 -9.91319656e-01 4.45077986e-01
-4.87300187e-01 -3.15900296e-01 3.85078788e-01 6.40514433e-01
1.11057317e+00 -6.14061713e-01 6.28720343e-01 -7.73333013e-01
7.99488842e-01 -5.92186093e-01 -5.34194291e-01 -9.08120424e-02
-3.95794868e-01 -2.64675498e-01 2.77742028e-01 -8.20568025e-01
-6.56547308e-01 -6.28641322e-02 4.08450603e-01 -3.35421950e-01
-2.00790018e-01 9.37204361e-01 2.19154119e-01 -1.71618998e-01
6.58285439e-01 6.33450896e-02 9.48447645e-01 -7.56009743e-02
-2.94133186e-01 7.14684248e-01 1.74359828e-01 -7.57990301e-01
2.68033594e-01 1.11199594e+00 5.28387010e-01 -8.06405962e-01
-2.62094945e-01 -5.84309101e-01 -1.20153293e-01 -6.14527643e-01
5.18626511e-01 -6.95929050e-01 -1.10811186e+00 6.26279056e-01
-7.72318721e-01 -1.08560669e+00 -1.56880721e-01 7.05984116e-01
-6.35871112e-01 -9.11104381e-02 -1.17282081e+00 -1.18247211e+00
1.74691412e-03 -1.04555249e+00 1.15915942e+00 1.44078091e-01
-1.49839774e-01 -1.16321635e+00 5.05999923e-01 6.86143562e-02
7.85928309e-01 1.23197043e+00 6.90978706e-01 5.71161434e-02
-4.96193051e-01 -6.71938300e-01 1.14791937e-01 -6.29797339e-01
3.60209048e-01 2.78090574e-02 -7.18896747e-01 -2.76778907e-01
1.73713863e-01 1.67119399e-01 2.80969650e-01 7.19460666e-01
4.65210646e-01 -1.39010891e-01 -1.21169519e+00 2.21968919e-01
1.69633722e+00 3.54007035e-01 7.61294186e-01 2.80308336e-01
2.29951799e-01 5.82095504e-01 7.31428862e-01 3.34234327e-01
-1.71141684e-01 3.58231694e-01 1.61114112e-01 -1.66579604e-01
8.62629116e-02 7.60569870e-02 3.60338956e-01 4.84111726e-01
1.32041471e-02 -3.48591477e-01 -9.25063431e-01 3.89250457e-01
-1.17226970e+00 -7.31350720e-01 -6.41646683e-01 2.27229500e+00
9.52444732e-01 -1.30718678e-01 3.38379979e-01 -2.51388222e-01
4.09326792e-01 -5.09465158e-01 -5.05868256e-01 -8.92179161e-02
-1.22542262e-01 1.56618923e-01 6.74177527e-01 7.66557992e-01
-4.04576808e-01 2.08520338e-01 7.35640526e+00 2.03483850e-01
-1.66738892e+00 1.05184600e-01 5.47317386e-01 -4.49902415e-01
-5.56630492e-01 1.85684338e-01 -4.81527656e-01 6.83914781e-01
1.08058858e+00 -6.43063113e-02 -2.96279788e-01 1.26656935e-01
6.69840276e-01 -9.04493868e-01 -1.00561523e+00 7.04192877e-01
-6.24281704e-01 -1.69520593e+00 -4.23256874e-01 8.97785664e-01
4.78030205e-01 -2.17566237e-01 -1.01791978e-01 -4.27731335e-01
-1.11392021e-01 -7.71395147e-01 3.35621357e-01 1.19996059e+00
1.15345037e+00 -2.87234396e-01 8.06122363e-01 3.95449638e-01
-7.95454979e-01 1.23913437e-02 -9.16626081e-02 -3.48294675e-01
2.59102553e-01 1.22532213e+00 -1.13654900e+00 -2.30377406e-01
1.64932087e-01 3.95375669e-01 -2.54816320e-02 9.43174601e-01
5.81293523e-01 4.42706525e-01 -6.15717351e-01 -6.94218040e-01
-2.01378614e-01 -4.99318361e-01 4.66651499e-01 1.21861041e+00
4.65103149e-01 5.16446412e-01 -2.92539507e-01 9.04160261e-01
3.43800336e-01 1.63627505e-01 -6.23015642e-01 -4.84060168e-01
5.84438026e-01 1.10307670e+00 -1.07099926e+00 -2.08897199e-02
-1.57042488e-01 4.44256812e-01 1.29359832e-03 7.23091960e-01
-4.23512697e-01 -1.34544134e-01 7.71967947e-01 8.94204974e-01
-4.09121066e-02 -7.77413487e-01 -1.19870983e-01 -5.49014986e-01
-3.33601236e-01 -1.48600951e-01 -4.50863689e-02 -6.18829727e-01
-1.15713418e+00 -3.58436078e-01 -3.78630221e-01 -9.44174886e-01
4.68457043e-01 -7.68845320e-01 -4.76962924e-01 8.61940026e-01
-9.97906089e-01 -7.59965479e-01 -3.23940814e-01 1.40876442e-01
-1.55480638e-01 4.12701309e-01 1.00284672e+00 5.28782122e-02
-4.28418368e-01 4.94519509e-02 5.64520299e-01 -2.57091403e-01
7.49609947e-01 -9.43920910e-01 -7.53309190e-01 2.66357094e-01
-8.02728832e-01 7.16880500e-01 8.22907865e-01 -1.14658701e+00
-2.06124544e+00 -7.18633056e-01 3.04812640e-01 -7.02057540e-01
6.83029175e-01 -3.89583975e-01 -6.67901397e-01 1.16652727e-01
-3.06635559e-01 1.70942619e-01 1.18446875e+00 -4.19275075e-01
1.75221026e-01 -1.61050141e-01 -1.39701653e+00 3.90822619e-01
5.46566904e-01 -4.62463260e-01 3.97098958e-01 6.34071589e-01
1.90699697e-01 -8.14981461e-01 -1.61481631e+00 1.98259920e-01
1.33595741e+00 -7.17569351e-01 3.72377902e-01 -5.19235611e-01
1.27474636e-01 -5.72132707e-01 -2.93603867e-01 -5.69885731e-01
-1.68209374e-01 -5.52801251e-01 -1.39820591e-01 6.75431430e-01
2.72210926e-01 -8.78096402e-01 9.70220089e-01 9.53776002e-01
6.11348562e-02 -1.31789351e+00 -9.11020041e-01 -8.44306111e-01
3.35382432e-01 2.70613760e-01 3.44109118e-01 6.17946684e-01
5.16157210e-01 -3.72868508e-01 6.39899671e-01 1.07830856e-02
5.95383942e-01 -3.54911126e-02 7.87587821e-01 -1.14874983e+00
-4.50837463e-02 -1.33535624e-01 -5.46845555e-01 -5.60238123e-01
-2.98243165e-01 -5.12386560e-01 9.12303291e-03 -1.55698431e+00
1.33723736e-01 -7.10155666e-01 -1.26611724e-01 -3.25133026e-01
3.46859813e-01 -4.94160466e-02 -2.81402320e-01 5.12058914e-01
-2.04607584e-02 4.24176902e-02 1.53076732e+00 -1.50783375e-01
-3.53372991e-01 -4.66755062e-01 -8.64214301e-02 -3.71495783e-02
5.17121673e-01 -5.46831131e-01 -2.71507382e-01 2.68703606e-03
2.21444637e-01 5.08412302e-01 5.42284012e-01 -1.01378679e+00
3.51192802e-01 -4.22568858e-01 5.15598178e-01 -5.17202377e-01
3.43003601e-01 -7.09181368e-01 8.90142381e-01 6.16629362e-01
5.19020483e-02 -4.44863945e-01 1.05737343e-01 9.35866773e-01
9.13326442e-02 2.70193636e-01 6.61646366e-01 -3.54630053e-01
1.67914242e-01 2.96093281e-02 -1.08079863e+00 -3.13102663e-01
1.27005160e+00 -8.72746170e-01 -8.67486954e-01 9.87353623e-02
-8.74908268e-01 -5.60120754e-02 1.22234285e+00 -6.60795808e-01
6.02521479e-01 -1.30795956e+00 -3.99327576e-01 -1.35278702e-01
-1.10458083e-01 -1.55456319e-01 3.82129282e-01 1.45253181e+00
-8.97081792e-01 1.79671541e-01 -1.84328362e-01 -1.04531074e+00
-7.55688071e-01 4.53541845e-01 8.79644215e-01 -2.83114463e-01
-1.93447381e-01 7.05858707e-01 1.17826499e-02 1.91068143e-01
-4.77711886e-01 -1.47793293e-01 2.44325817e-01 1.86488323e-03
4.99064147e-01 5.40417910e-01 -1.61668211e-02 -1.57279342e-01
-3.91997755e-01 6.92195475e-01 2.13438958e-01 -2.53189504e-01
1.26516092e+00 -2.48353586e-01 -3.60986799e-01 9.80949283e-01
1.23287392e+00 1.76304281e-01 -1.57024586e+00 3.03720623e-01
-1.80298686e-01 -2.22357899e-01 -2.87560821e-01 -8.12626302e-01
-6.47745013e-01 3.29776883e-01 7.04705477e-01 2.40614325e-01
6.84905469e-01 3.02907407e-01 7.50808418e-01 -9.20625031e-02
4.95588481e-01 -1.00711203e+00 1.45747691e-01 1.40148237e-01
8.66741538e-01 -6.12109423e-01 1.80733845e-01 -4.86715227e-01
2.90401816e-01 1.20926929e+00 1.54904455e-01 -4.05538350e-01
1.00925910e+00 9.86381650e-01 4.04524714e-01 -7.71130502e-01
-7.19199061e-01 3.03321958e-01 -4.98521984e-01 5.94269156e-01
6.48198068e-01 3.44076335e-01 -6.84393764e-01 4.97336596e-01
3.19767892e-01 2.64913052e-01 8.47304761e-01 1.28108168e+00
-6.27823770e-01 -8.41189146e-01 -6.04735613e-01 5.63462138e-01
-4.54649210e-01 7.53716230e-01 -4.06334490e-01 9.02715683e-01
-2.77196348e-01 7.85830438e-01 1.49745569e-01 -1.26264840e-01
3.10462326e-01 3.34646314e-01 7.73815811e-01 -4.93155479e-01
-1.25284836e-01 6.18925929e-01 -9.08053741e-02 -5.47453642e-01
-7.29930162e-01 -8.84332299e-01 -1.65484238e+00 -1.52079891e-02
-3.69557321e-01 1.64024144e-01 1.00108421e+00 8.44317734e-01
6.10260189e-01 1.52663225e-02 3.84560198e-01 -1.00222456e+00
1.13007464e-01 -6.02024436e-01 -1.37775290e+00 9.15233269e-02
4.57396507e-01 -9.82635736e-01 -5.94506443e-01 1.36358783e-01] | [13.727827072143555, -3.0993454456329346] |
28a7ebeb-f865-4558-a33c-3a064db29061 | improving-automatic-source-code-summarization | 1811.07234 | null | http://arxiv.org/abs/1811.07234v1 | http://arxiv.org/pdf/1811.07234v1.pdf | Improving Automatic Source Code Summarization via Deep Reinforcement Learning | Code summarization provides a high level natural language description of the
function performed by code, as it can benefit the software maintenance, code
categorization and retrieval. To the best of our knowledge, most
state-of-the-art approaches follow an encoder-decoder framework which encodes
the code into a hidden space and then decode it into natural language space,
suffering from two major drawbacks: a) Their encoders only consider the
sequential content of code, ignoring the tree structure which is also critical
for the task of code summarization, b) Their decoders are typically trained to
predict the next word by maximizing the likelihood of next ground-truth word
with previous ground-truth word given. However, it is expected to generate the
entire sequence from scratch at test time. This discrepancy can cause an
\textit{exposure bias} issue, making the learnt decoder suboptimal. In this
paper, we incorporate an abstract syntax tree structure as well as sequential
content of code snippets into a deep reinforcement learning framework (i.e.,
actor-critic network). The actor network provides the confidence of predicting
the next word according to current state. On the other hand, the critic network
evaluates the reward value of all possible extensions of the current state and
can provide global guidance for explorations. We employ an advantage reward
composed of BLEU metric to train both networks. Comprehensive experiments on a
real-world dataset show the effectiveness of our proposed model when compared
with some state-of-the-art methods. | ['Yao Wan', 'Philip S. Yu', 'Jian Wu', 'Guandong Xu', 'Zhou Zhao', 'Haochao Ying', 'Min Yang'] | 2018-11-17 | null | null | null | null | ['code-summarization'] | ['computer-code'] | [ 1.94113314e-01 2.92286783e-01 -3.99026692e-01 -1.81600571e-01
-8.26221347e-01 -3.35642308e-01 2.27829888e-01 2.45904431e-01
-5.99760823e-02 6.19847596e-01 5.03836632e-01 -4.18414891e-01
3.66235912e-01 -6.86304271e-01 -1.02754474e+00 -3.77604753e-01
-2.66978405e-02 -2.98651517e-03 2.68677801e-01 -1.38637275e-01
6.05089784e-01 -2.92815238e-01 -1.52617157e+00 2.85351306e-01
1.34366262e+00 7.46824622e-01 7.23755836e-01 5.52576363e-01
-3.89748663e-01 1.06040692e+00 -7.98543990e-01 -5.45144141e-01
-1.15780473e-01 -7.65528202e-01 -7.79769361e-01 2.28599068e-02
-1.54402643e-01 -4.35887486e-01 -3.03533316e-01 1.38904178e+00
2.28221446e-01 -2.70983458e-01 4.01982933e-01 -9.66107249e-01
-7.19582677e-01 9.68774736e-01 -5.26682734e-01 1.97739806e-02
5.27853847e-01 4.18100178e-01 1.06735575e+00 -5.73281527e-01
4.52389538e-01 9.06329274e-01 3.12172860e-01 5.74191988e-01
-8.03883791e-01 -4.52299595e-01 3.52474868e-01 1.13921106e-01
-1.18921447e+00 -3.09950173e-01 7.70782650e-01 -5.45705378e-01
1.22171462e+00 -1.42442852e-01 7.07166731e-01 9.11916256e-01
6.67275667e-01 9.78862762e-01 4.94277030e-01 -4.19413835e-01
5.80879867e-01 -2.31306981e-02 -2.97785597e-03 1.05998778e+00
1.90680876e-01 -1.56955838e-01 -5.10186911e-01 -1.17255621e-01
2.45541185e-01 1.09467283e-02 -2.39244401e-01 -3.19661796e-01
-9.87910271e-01 7.14705706e-01 3.49571556e-01 1.00434780e-01
-4.88706261e-01 4.75088775e-01 5.92680812e-01 2.58558001e-02
7.47404024e-02 3.40140998e-01 -3.17955345e-01 -4.98260409e-01
-1.02086687e+00 1.85217142e-01 6.94506049e-01 1.19052982e+00
9.06625450e-01 3.72871310e-01 -4.00083542e-01 5.94927788e-01
6.33079648e-01 2.73039639e-01 9.14394855e-01 -6.69505358e-01
8.29873323e-01 9.36600268e-01 -1.14501975e-01 -6.54741168e-01
6.34291098e-02 -6.94269001e-01 -5.15547037e-01 1.92196563e-01
-2.35419422e-01 -2.61282682e-01 -7.18220770e-01 1.70158207e+00
-1.95132107e-01 2.09373936e-01 2.53914505e-01 5.48779070e-01
4.60634410e-01 9.38682377e-01 -1.71887472e-01 -4.04801846e-01
1.13981485e+00 -1.25356042e+00 -8.22770000e-01 -7.63697028e-01
6.38831556e-01 -5.02861738e-01 9.99804139e-01 4.28565204e-01
-1.09059715e+00 -4.72077191e-01 -1.41622889e+00 3.69985044e-01
1.20477214e-01 3.25462520e-01 3.92801821e-01 5.64431846e-01
-9.94766414e-01 5.01253903e-01 -8.82829428e-01 -9.72742215e-02
3.06606203e-01 8.14669058e-02 1.31802991e-01 -8.70119408e-02
-1.03493643e+00 6.79725170e-01 6.73323035e-01 4.46403679e-03
-1.31596923e+00 -1.88639820e-01 -9.34645772e-01 3.68108928e-01
6.13104045e-01 -4.57284719e-01 1.60019374e+00 -1.09029126e+00
-1.68791115e+00 2.32692480e-01 -2.77621716e-01 -5.95746517e-01
2.42411792e-01 -2.94458419e-01 -2.56913006e-01 -2.58014768e-01
1.64658278e-01 3.76966357e-01 7.32331097e-01 -1.13253188e+00
-5.77426016e-01 1.99084152e-02 1.05587080e-01 1.43281996e-01
-4.44688857e-01 -2.95321077e-01 -7.13110805e-01 -6.16061747e-01
-2.07531169e-01 -7.45025814e-01 -2.69405127e-01 -3.75122398e-01
-5.68932950e-01 -2.27356091e-01 4.21773195e-01 -8.03904951e-01
1.98566711e+00 -2.23661613e+00 2.20541760e-01 -2.06826657e-01
-2.65005697e-02 2.72870243e-01 -1.53822258e-01 6.50117815e-01
4.05816957e-02 2.82974094e-01 -4.70582992e-01 -1.44703299e-01
-6.12701215e-02 5.99054880e-02 -4.54482585e-01 2.92239010e-01
4.07109261e-01 8.95729363e-01 -1.07163370e+00 -4.37551618e-01
-1.83956340e-01 -5.36601506e-02 -9.85667109e-01 5.38311064e-01
-6.95802212e-01 1.12830624e-01 -6.03062451e-01 4.45254922e-01
3.62346619e-01 -3.09276760e-01 1.79666609e-01 2.44084150e-01
-1.40205279e-01 4.54751104e-01 -8.94677758e-01 2.04613519e+00
-6.44887269e-01 4.72661734e-01 -4.78621423e-01 -8.96532416e-01
1.10895741e+00 2.11524472e-01 -3.93600129e-02 -6.78672016e-01
-1.79768540e-02 2.54921228e-01 1.01136118e-01 -7.64244080e-01
6.03358448e-01 1.58199847e-01 -3.07906181e-01 5.08308887e-01
-1.39695257e-01 8.78051668e-02 1.62575275e-01 1.55793205e-01
1.33633029e+00 4.40331936e-01 5.35808146e-01 3.09874639e-02
5.39683282e-01 -3.83979548e-03 7.49348879e-01 6.74305201e-01
1.29593357e-01 4.45542157e-01 9.82518792e-01 -1.93420410e-01
-9.05683875e-01 -7.61526346e-01 3.56481880e-01 8.71208727e-01
2.46272951e-01 -6.69797063e-01 -1.02065039e+00 -9.11582887e-01
-4.34726030e-01 1.27212596e+00 -4.28934485e-01 -5.80314159e-01
-5.39053500e-01 -4.29782182e-01 5.30121803e-01 4.22168523e-01
4.57989365e-01 -1.21245766e+00 -1.09295535e+00 4.00386930e-01
-2.61048883e-01 -7.02745497e-01 -6.74120128e-01 2.55296350e-01
-8.65753710e-01 -8.61037016e-01 -4.92767692e-01 -7.82615781e-01
6.98721111e-01 -1.49696410e-01 1.02612066e+00 3.87716442e-01
5.96291618e-03 -1.52377695e-01 -6.16082907e-01 -1.90423861e-01
-1.01003528e+00 9.19359773e-02 -4.34717536e-01 -1.98099092e-01
1.14780933e-01 -3.08007896e-01 -5.94048917e-01 -7.39395767e-02
-9.61279273e-01 3.59333247e-01 9.88672197e-01 8.67377102e-01
2.94719249e-01 1.79357439e-01 6.49525881e-01 -7.54047871e-01
8.29204261e-01 -6.96211934e-01 -6.69387281e-01 4.89811629e-01
-7.83749998e-01 6.82625234e-01 8.87054980e-01 -1.85555518e-01
-1.06914151e+00 1.00632638e-01 -2.68358499e-01 -5.58628626e-02
1.73488244e-01 8.84869576e-01 -1.61845461e-01 4.99450594e-01
5.99740744e-01 8.13241661e-01 -7.90814236e-02 -2.45166615e-01
1.28880963e-01 8.34479094e-01 3.14268798e-01 -4.72628057e-01
4.49892461e-01 -1.83197483e-01 -4.67464209e-01 -2.33806551e-01
-5.36479712e-01 -1.34733366e-02 -1.67998686e-01 -2.44310126e-01
8.56296778e-01 -8.36316168e-01 -4.87253189e-01 2.24445790e-01
-1.55438471e+00 -2.02755183e-01 -2.74802484e-02 1.84990570e-01
-6.93075895e-01 5.79502642e-01 -3.65075648e-01 -9.51585114e-01
-4.36317027e-01 -1.72446823e+00 9.68250453e-01 3.16823870e-01
-2.11432189e-01 -5.35342693e-01 6.24500923e-02 -9.00430456e-02
3.01123530e-01 4.72529829e-02 1.35205364e+00 -5.54948688e-01
-7.71681249e-01 -2.15900078e-01 1.23854555e-01 3.95323217e-01
-1.19286954e-01 4.49355468e-02 -6.65313423e-01 -2.47875988e-01
5.68536595e-02 -4.17246461e-01 7.21505046e-01 1.64268762e-01
1.19966495e+00 -5.96203625e-01 -2.51366854e-01 2.61025488e-01
1.67751276e+00 5.13882935e-01 8.91408503e-01 2.74390191e-01
4.09589797e-01 3.08571130e-01 7.69621670e-01 7.85506845e-01
3.80013973e-01 5.85044324e-01 9.33533013e-01 4.08414960e-01
-6.43692305e-03 -5.18226683e-01 9.26141202e-01 1.10701990e+00
5.80090046e-01 -5.42496681e-01 -1.09153533e+00 6.08843863e-01
-1.90364861e+00 -6.08807027e-01 1.43000662e-01 2.18913102e+00
1.00240242e+00 4.43006337e-01 -3.23151350e-01 -9.20695215e-02
7.54330456e-01 1.61081761e-01 -8.29465210e-01 -5.47192454e-01
3.61559182e-01 -1.32860348e-01 3.14871192e-01 3.54948521e-01
-4.79940325e-01 8.49674881e-01 5.59610033e+00 8.57767582e-01
-1.00627041e+00 -2.01120577e-03 4.26500946e-01 1.15069434e-01
-5.23743808e-01 3.17502081e-01 -6.57004833e-01 6.85834229e-01
9.95114744e-01 -4.24242705e-01 6.51822925e-01 1.03379214e+00
5.65518662e-02 -2.41948485e-01 -1.22974467e+00 7.18103528e-01
2.04843417e-01 -1.21403980e+00 1.34094030e-01 -1.77950054e-01
7.57599354e-01 -6.36947006e-02 -1.15067482e-01 5.93202651e-01
3.23572189e-01 -8.40024769e-01 1.09774840e+00 5.10566413e-01
7.34341741e-01 -8.40114415e-01 7.39806414e-01 7.62249351e-01
-1.08593309e+00 -3.87370735e-01 -2.98292547e-01 -4.52410057e-02
2.38767769e-02 4.34708953e-01 -9.97173727e-01 5.57581663e-01
1.90181524e-01 6.86099291e-01 -7.04396367e-01 1.07084692e+00
-4.82081413e-01 6.48092508e-01 3.15725893e-01 -4.41900253e-01
3.95255059e-01 1.56771913e-01 4.50652927e-01 1.18919921e+00
7.10113645e-01 -3.13469827e-01 1.95691720e-01 1.15353608e+00
-1.43562019e-01 8.03461671e-02 -4.99056548e-01 -2.69828677e-01
5.78603983e-01 7.84836054e-01 -6.13421202e-01 -3.26208174e-01
-4.84851837e-01 9.76715386e-01 3.86294842e-01 3.19902092e-01
-9.19147491e-01 -5.47091961e-01 2.74873942e-01 -1.22886546e-01
5.01662612e-01 1.20561793e-01 -3.26908141e-01 -1.08157921e+00
2.71474272e-01 -1.04632974e+00 1.15839541e-02 -8.13986659e-01
-6.59998178e-01 7.88611591e-01 -1.64904445e-01 -1.52795553e+00
-4.74974781e-01 -1.80746138e-01 -8.12467456e-01 7.65546441e-01
-1.44852400e+00 -4.87544566e-01 -1.78750172e-01 -1.41545698e-01
1.10438800e+00 -4.79408652e-01 6.26592159e-01 -2.65314002e-02
-7.20379591e-01 5.61933875e-01 1.11161254e-01 -9.93528217e-03
3.45007420e-01 -1.13235891e+00 7.13387907e-01 1.15734088e+00
-1.79907098e-01 7.94164300e-01 8.61017048e-01 -9.28438425e-01
-1.56413221e+00 -1.10558927e+00 7.25320816e-01 2.78648455e-02
4.98975068e-01 -2.44603157e-01 -9.12678182e-01 4.92640704e-01
4.71146703e-01 -3.69513124e-01 4.46760923e-01 -2.65645474e-01
-1.40623093e-01 3.12117934e-02 -6.99402750e-01 6.85939014e-01
7.25782096e-01 -3.85073334e-01 -5.78061700e-01 1.29552588e-01
9.95382428e-01 -4.70859110e-01 -4.07931298e-01 1.63170114e-01
3.52261275e-01 -1.13670623e+00 3.08104604e-01 -2.62989610e-01
1.15458965e+00 -4.36618119e-01 -5.18308803e-02 -1.44961381e+00
-1.53172418e-01 -5.10167599e-01 -3.77763271e-01 1.16991878e+00
5.74033082e-01 -1.50364697e-01 6.64440870e-01 2.76300609e-01
-5.27845562e-01 -1.08364141e+00 -7.10013807e-01 -5.99680185e-01
-1.69676989e-01 -5.04031777e-01 8.09360683e-01 3.30155700e-01
3.02350909e-01 4.07078177e-01 -4.55118567e-01 2.48972420e-02
2.36443356e-01 1.26263693e-01 6.08185947e-01 -7.41098404e-01
-6.04317427e-01 -5.11534095e-01 -1.99968398e-01 -1.42918146e+00
3.95849645e-01 -9.98995543e-01 5.64119577e-01 -1.70867491e+00
4.92945969e-01 -4.77315709e-02 -1.13705516e-01 3.45905393e-01
-3.42396736e-01 -5.41902840e-01 7.19883591e-02 1.38682678e-01
-8.48388493e-01 8.33467126e-01 1.06209135e+00 -2.79779881e-01
-1.69223428e-01 -2.47150268e-02 -8.16425204e-01 5.60681164e-01
6.87374473e-01 -6.75826609e-01 -6.37787580e-01 -7.48306751e-01
5.87506652e-01 7.17319548e-01 -1.57165136e-02 -1.13047147e+00
4.83579308e-01 -1.96708769e-01 -1.41678184e-01 -5.75636983e-01
-1.46512538e-01 -7.27921903e-01 -3.02088186e-02 8.81922007e-01
-6.32418811e-01 3.14327806e-01 1.56940952e-01 8.91043305e-01
-2.32864931e-01 -9.73508716e-01 5.49598515e-01 -1.69892445e-01
-6.85726523e-01 -4.13043648e-02 -6.90890789e-01 1.50702566e-01
1.00777292e+00 -1.25413865e-01 -2.67211139e-01 -3.42323810e-01
5.13369441e-02 2.52219617e-01 6.09413326e-01 5.88791609e-01
7.53189564e-01 -1.22821987e+00 -6.21602237e-01 1.65332392e-01
4.01302546e-01 1.89276766e-02 7.00299963e-02 3.82109195e-01
-5.57309985e-01 2.49867946e-01 -8.64851549e-02 -4.48189884e-01
-7.98505723e-01 6.22296035e-01 1.48689777e-01 -5.14190853e-01
-5.09915292e-01 4.94524479e-01 2.09826097e-01 5.36379181e-02
3.71336281e-01 -3.14730078e-01 -2.00585887e-01 -3.03034812e-01
4.93448049e-01 9.34018120e-02 2.58257426e-02 -1.61016509e-01
-3.29278588e-01 2.95852900e-01 -2.39848703e-01 -5.33911288e-02
1.22779977e+00 6.54109716e-02 -1.98607787e-01 4.94752586e-01
1.22240329e+00 -4.63663526e-02 -1.28817892e+00 -1.78668305e-01
2.40690768e-01 -2.42734492e-01 2.21235175e-02 -8.19204211e-01
-8.70525956e-01 9.43761051e-01 3.76541078e-01 6.38327152e-02
1.02697992e+00 -2.43962213e-01 8.43139529e-01 4.42976683e-01
4.08020675e-01 -1.03404510e+00 4.60889041e-01 5.38160264e-01
7.16924787e-01 -8.89738083e-01 -6.92333281e-02 -1.09706149e-02
-9.19797897e-01 1.23857367e+00 8.81570518e-01 -6.70689121e-02
1.31780639e-01 2.83668458e-01 -3.78657043e-01 -5.25441729e-02
-1.09732366e+00 1.43954992e-01 2.73033157e-02 2.80086428e-01
5.07525921e-01 -2.47506931e-01 -3.60058576e-01 7.53219664e-01
-2.82069929e-02 8.54666252e-03 9.16865468e-01 1.08398044e+00
-7.83625126e-01 -1.05186725e+00 -1.01513796e-01 5.49249768e-01
-3.42656076e-01 -2.39861876e-01 -4.74451222e-02 1.65506735e-01
2.84454152e-02 7.70248890e-01 -1.34183690e-01 -3.81134301e-01
2.65232116e-01 1.88050076e-01 5.79468086e-02 -9.81761873e-01
-6.21115565e-01 -9.28204227e-03 -1.32389814e-01 -3.63039941e-01
1.62287697e-01 -6.22894108e-01 -1.50669587e+00 8.29271302e-02
-4.93111074e-01 2.57739693e-01 7.06004202e-01 9.36034560e-01
4.33471590e-01 1.10538626e+00 7.20666349e-01 -4.50731546e-01
-8.64111304e-01 -9.49655712e-01 -1.39746308e-01 6.03626296e-02
4.12446558e-01 -3.34954023e-01 4.88842763e-02 2.19941452e-01] | [7.671348571777344, 7.89597225189209] |
54d8bdd0-9096-4cf8-b7d5-14d9e3871407 | saliency-detection-with-moving-camera-via | 2111.01681 | null | https://arxiv.org/abs/2111.01681v1 | https://arxiv.org/pdf/2111.01681v1.pdf | Saliency detection with moving camera via background model completion | To detect saliency in video is a fundamental step in many computer vision systems. Saliency is the significant target(s) in the video. The object of interest is further analyzed for high-level applications. The segregation of saliency and the background can be made if they exhibit different visual cues. Therefore, saliency detection is often formulated as background subtraction. However, saliency detection is challenging. For instance, dynamic background can result in false positive errors. In another scenario, camouflage will lead to false negative errors. With moving camera, the captured scenes are even more complicated to handle. We propose a new framework, called saliency detection via background model completion (SD-BMC), that comprises of a background modeler and the deep learning background/foreground segmentation network. The background modeler generates an initial clean background image from a short image sequence. Based on the idea of video completion, a good background frame can be synthesized with the co-existence of changing background and moving objects. We adopt the background/foreground segmenter, although pre-trained with a specific video dataset, can also detect saliency in unseen videos. The background modeler can adjust the background image dynamically when the background/foreground segmenter output deteriorates during processing of a long video. To the best of our knowledge, our framework is the first one to adopt video completion for background modeling and saliency detection in videos captured by moving camera. The results, obtained from the PTZ videos, show that our proposed framework outperforms some deep learning-based background subtraction models by 11% or more. With more challenging videos, our framework also outperforms many high ranking background subtraction methods by more than 3%. | ['Kwok-Leung Chan', 'Yupei Zhang'] | 2021-10-30 | null | null | null | null | ['foreground-segmentation'] | ['computer-vision'] | [ 7.50134170e-01 -3.52056146e-01 -9.91457477e-02 7.98819065e-02
-3.35063934e-01 -2.24650681e-01 2.23188266e-01 -1.02773190e-01
-2.65918523e-01 6.65182292e-01 -1.43217191e-01 -5.83210066e-02
5.76443791e-01 -6.03338182e-01 -8.83844137e-01 -9.37030673e-01
2.82258123e-01 -1.00879088e-01 1.12206471e+00 -1.05659693e-01
3.33495408e-01 1.56188935e-01 -1.83299136e+00 3.13938558e-01
9.44874823e-01 1.03519917e+00 9.70996201e-01 6.08908594e-01
-3.50646883e-01 1.10387433e+00 -6.37639105e-01 4.14196961e-02
2.89935470e-01 -5.86537659e-01 -3.63301307e-01 3.86050195e-01
5.49926877e-01 -5.54468572e-01 -2.45759249e-01 1.52053809e+00
3.07628125e-01 1.27344728e-01 6.31739050e-02 -1.39911044e+00
-1.04681373e-01 4.46279168e-01 -8.38609099e-01 7.98054993e-01
2.32081458e-01 1.93923861e-01 3.39925498e-01 -8.83494496e-01
3.93817306e-01 1.24548781e+00 1.44892901e-01 7.26278722e-01
-9.31730270e-01 -6.79957747e-01 5.45580268e-01 4.16273683e-01
-1.10923886e+00 -4.92425829e-01 1.10395825e+00 -3.91103476e-01
7.65156820e-02 4.06332195e-01 6.70699954e-01 9.52489197e-01
1.15688473e-01 1.34710240e+00 8.59368563e-01 -2.31336892e-01
3.07804912e-01 2.48637609e-02 1.13532290e-01 2.19479263e-01
4.68167037e-01 -1.55735850e-01 -3.26677740e-01 3.08148354e-01
7.64667153e-01 3.62540156e-01 -7.10348070e-01 -4.03016418e-01
-1.30765748e+00 3.13367486e-01 3.36959302e-01 2.00973466e-01
-3.73028636e-01 6.03473410e-02 2.12840125e-01 -8.55634063e-02
2.19859853e-01 -1.03607729e-01 -2.18035400e-01 2.19402492e-01
-1.51105189e+00 1.27853379e-01 4.29826260e-01 9.31093514e-01
6.89504981e-01 5.44087529e-01 -3.27771276e-01 4.76024538e-01
1.92848384e-01 7.62208104e-01 5.80028892e-01 -7.08443582e-01
1.74594030e-01 4.52798575e-01 4.76589710e-01 -1.38708794e+00
-1.86633110e-01 -4.59822267e-01 -7.56805778e-01 2.71231562e-01
3.50496054e-01 -1.31246880e-01 -9.86690283e-01 1.53280902e+00
5.82692206e-01 8.18242967e-01 9.19131637e-02 1.26074600e+00
1.09112847e+00 7.33924448e-01 1.64367277e-02 -5.63362718e-01
1.14692724e+00 -1.26377678e+00 -9.42538619e-01 -6.34688437e-01
-3.35145220e-02 -8.75883818e-01 7.50585675e-01 3.21576327e-01
-1.09126604e+00 -8.47158909e-01 -1.12712467e+00 2.27396190e-01
-3.93440723e-02 1.41673207e-01 3.07552636e-01 4.34577554e-01
-9.69651401e-01 1.15880363e-01 -8.32160175e-01 -2.96382993e-01
4.13159430e-01 2.07187504e-01 1.07382387e-01 -1.08347610e-01
-1.10116863e+00 7.88912177e-01 6.03033721e-01 2.36022145e-01
-1.42946815e+00 -3.65183413e-01 -7.40151107e-01 -1.39911780e-02
8.94287646e-01 -4.28281486e-01 1.17899084e+00 -1.91363323e+00
-1.06547666e+00 5.97840369e-01 -4.13435459e-01 -4.56821650e-01
6.14772677e-01 -3.39475363e-01 -5.06935298e-01 3.14004213e-01
3.00123692e-01 6.04928851e-01 1.37986612e+00 -1.36630785e+00
-1.02291799e+00 2.20123753e-02 -7.72161037e-02 3.07718903e-01
1.13275431e-01 2.72392869e-01 -7.13068664e-01 -6.62503481e-01
1.54357687e-01 -6.93149149e-01 -2.71872640e-01 -2.24332765e-01
-1.35979563e-01 3.43535542e-01 1.22208202e+00 -9.13827360e-01
1.24818265e+00 -2.15898132e+00 1.41319269e-02 -3.35873097e-01
2.90656000e-01 5.77637196e-01 -9.09704715e-02 -4.67832536e-01
-8.71055499e-02 -1.88243940e-01 -3.06495816e-01 1.95145845e-01
-5.77149510e-01 -1.45418018e-01 -4.21948493e-01 3.21755290e-01
4.10628229e-01 7.11333036e-01 -1.21595657e+00 -7.42113531e-01
4.17928368e-01 2.72183657e-01 -3.42784911e-01 1.36789680e-01
-3.11854929e-01 5.45063853e-01 -3.22690338e-01 8.63308907e-01
9.79311764e-01 -1.67377830e-01 -3.86158191e-02 -1.59682855e-01
-1.91754386e-01 -6.09460995e-02 -1.49943078e+00 1.33430386e+00
2.32222587e-01 8.99066448e-01 4.31746572e-01 -1.05816197e+00
6.49981260e-01 5.53357974e-02 3.51533324e-01 -4.79689777e-01
1.93804905e-01 2.66147316e-01 1.15790717e-01 -4.88718390e-01
7.00206757e-01 1.81831062e-01 3.29742700e-01 -2.36236304e-01
-2.67230242e-01 1.05423704e-01 3.60800982e-01 1.14093915e-01
9.05239999e-01 2.02655092e-01 1.39100015e-01 -1.47768468e-01
8.16459596e-01 3.92798707e-02 1.03361785e+00 6.67930722e-01
-5.37345588e-01 7.78483272e-01 2.22927585e-01 -2.92409569e-01
-6.66500688e-01 -1.02462590e+00 2.17770636e-01 1.02537870e+00
9.07273293e-01 6.22713938e-02 -9.57753062e-01 -3.76127541e-01
-4.23718065e-01 5.66073120e-01 -2.32857764e-01 -2.01108664e-01
-7.49888122e-01 -7.36648440e-01 -3.51148769e-02 2.36220837e-01
8.70851099e-01 -1.24513197e+00 -9.63697672e-01 3.60950500e-01
-5.84959626e-01 -1.27721620e+00 -5.55701017e-01 -1.15692772e-01
-9.81116474e-01 -1.16600549e+00 -9.78230894e-01 -1.08714700e+00
6.62787139e-01 1.10045886e+00 8.37646484e-01 3.52658898e-01
-1.78447925e-02 1.55903727e-01 -2.44510919e-01 -3.79266679e-01
-4.70101297e-01 -3.92894745e-01 2.48700622e-02 4.06143039e-01
3.25403988e-01 -2.69952357e-01 -7.17972696e-01 4.19943213e-01
-1.26343548e+00 4.84476864e-01 5.40325880e-01 6.17956579e-01
4.94615883e-01 4.27531928e-01 3.09574991e-01 -3.48148286e-01
3.59564796e-02 -3.34249228e-01 -8.74253273e-01 2.26843297e-01
1.20676607e-01 -4.40875173e-01 3.15890759e-01 -6.14253640e-01
-1.08053803e+00 1.45183742e-01 3.69643927e-01 -8.13218057e-01
-1.81673571e-01 5.48922792e-02 -6.35726631e-01 -4.07528220e-04
3.46893907e-01 6.07710421e-01 -1.71379551e-01 -1.91575333e-01
1.49218189e-02 5.03540754e-01 7.43392825e-01 -6.96300715e-02
7.89435744e-01 5.23304224e-01 -2.21589059e-01 -9.61890578e-01
-6.75790310e-01 -4.86670703e-01 -5.61070502e-01 -5.52821040e-01
9.93891537e-01 -1.21697414e+00 -1.50804192e-01 8.11714947e-01
-1.28289104e+00 -3.03732067e-01 1.32110000e-01 3.37887406e-01
-2.76493430e-01 5.70533037e-01 -3.41802955e-01 -9.48615849e-01
-2.02321604e-01 -1.35224521e+00 9.47871923e-01 6.69200242e-01
2.42467180e-01 -6.68655097e-01 -5.03568232e-01 2.73671091e-01
3.52789789e-01 2.71461874e-01 2.97650665e-01 -4.25541520e-01
-1.12178707e+00 5.98975793e-02 -2.70845026e-01 4.58265692e-01
3.80448520e-01 2.42960766e-01 -9.86493111e-01 -1.77088901e-01
3.69889408e-01 3.47093016e-01 1.13173699e+00 7.00763404e-01
8.93269300e-01 -1.78574920e-01 -5.00837743e-01 3.52403283e-01
1.33064663e+00 6.63973808e-01 7.53577054e-01 4.90820885e-01
9.67175364e-01 4.57663059e-01 1.09112620e+00 8.18861276e-02
1.44021884e-01 5.97190261e-01 6.08159781e-01 -3.78244847e-01
-2.88987309e-01 4.19955775e-02 8.52998137e-01 6.87946200e-01
9.41776633e-02 -3.64016928e-02 -8.09153497e-01 6.42567039e-01
-1.84430540e+00 -1.29675376e+00 -2.87356675e-01 2.31530786e+00
7.00851738e-01 4.42356557e-01 9.87777933e-02 5.49204536e-02
1.38446283e+00 1.79744765e-01 -7.68995881e-01 3.04876894e-01
-5.23236334e-01 -1.76115513e-01 4.19974297e-01 4.02571291e-01
-1.38257897e+00 1.29111147e+00 5.06994677e+00 7.85542965e-01
-1.57547116e+00 1.12668872e-01 7.57216334e-01 -1.87341452e-01
1.02917720e-02 7.32407644e-02 -9.06821489e-01 1.06510925e+00
3.77432883e-01 -2.23027319e-01 3.99932116e-01 1.05228508e+00
4.58350778e-01 -5.91933489e-01 -7.34334886e-01 1.10021758e+00
3.64090532e-01 -1.18456089e+00 4.47382554e-02 -4.20084625e-01
8.77001047e-01 -7.60185793e-02 -2.47933045e-02 2.15211123e-01
-1.43490985e-01 -5.04548311e-01 1.00353360e+00 3.90150011e-01
2.32492149e-01 -4.05070364e-01 6.21895254e-01 4.41954374e-01
-1.16681361e+00 -1.72468022e-01 -4.81726944e-01 -1.58874661e-01
2.57727921e-01 7.48033047e-01 -5.75344503e-01 3.20649326e-01
7.81875908e-01 8.27959418e-01 -7.27257967e-01 1.35600495e+00
-1.69535130e-01 5.46915352e-01 -1.17168486e-01 1.89795107e-01
7.21074315e-03 -2.29610130e-01 9.30226624e-01 1.08174694e+00
2.67608017e-01 1.24426894e-02 4.92115229e-01 8.69088948e-01
2.37938106e-01 6.08788580e-02 -4.09199566e-01 1.05526499e-01
2.51762033e-01 1.33210135e+00 -1.22451508e+00 -7.59840786e-01
-4.31373656e-01 1.07094133e+00 -3.57480496e-01 5.49514651e-01
-1.14639533e+00 -5.52153029e-02 4.46375400e-01 1.15557991e-01
5.30597866e-01 4.08104174e-02 -8.80073309e-02 -1.56768465e+00
9.22346264e-02 -9.58703578e-01 -3.17465374e-03 -1.19123089e+00
-6.06403291e-01 5.05825579e-01 -8.82118642e-02 -1.36683619e+00
3.26746881e-01 -3.67215931e-01 -8.18486571e-01 6.08852148e-01
-1.69984877e+00 -8.04489017e-01 -7.58214235e-01 4.74699646e-01
1.06244338e+00 -3.03148460e-02 -4.33472879e-02 3.67973864e-01
-8.78312290e-01 -7.44064711e-03 -5.68585433e-02 1.86891451e-01
6.11829519e-01 -9.21524525e-01 2.11906269e-01 1.62149143e+00
-1.66668221e-01 3.23849648e-01 9.91755962e-01 -1.06907499e+00
-1.08647585e+00 -1.26792657e+00 2.74567157e-01 -2.91554272e-01
4.58615601e-01 -2.20775664e-01 -1.11128819e+00 3.71787399e-01
1.47503212e-01 -6.23273998e-02 1.62286237e-01 -8.18093061e-01
3.47690582e-01 -2.42350757e-01 -8.54738712e-01 9.13698494e-01
7.77772844e-01 -3.93218994e-02 -7.33097374e-01 9.77464989e-02
1.03950691e+00 -5.18217683e-01 1.01962082e-01 6.21854007e-01
2.27206171e-01 -1.00764406e+00 9.36361670e-01 -2.79351115e-01
4.85078424e-01 -1.03249621e+00 -1.47318915e-01 -8.71132493e-01
-2.02555746e-01 -5.98951757e-01 -1.71820030e-01 1.30129981e+00
-1.71860501e-01 -1.52947187e-01 6.82446480e-01 4.63516384e-01
-9.27790254e-02 -1.80813953e-01 -5.95646322e-01 -5.72018266e-01
-5.31824529e-01 -3.98937225e-01 3.84712756e-01 8.33994210e-01
-5.38579881e-01 1.84788972e-01 -5.03935099e-01 4.47861403e-01
6.89817727e-01 2.57952690e-01 9.33772981e-01 -1.08685112e+00
4.21542488e-02 -4.73463923e-01 -5.00971735e-01 -1.16532218e+00
-5.80640044e-03 -4.46950912e-01 3.68657351e-01 -1.44507349e+00
1.97982073e-01 7.30821863e-02 -5.14770865e-01 1.43786162e-01
-7.73764789e-01 1.56437635e-01 3.30717027e-01 1.73519239e-01
-8.99080455e-01 4.37267870e-01 1.35472524e+00 -4.12398785e-01
-3.52714479e-01 1.25786811e-01 -5.62445581e-01 9.92355824e-01
9.29072380e-01 -3.37816805e-01 -3.02459568e-01 -2.18968302e-01
-2.41441473e-01 5.40219322e-02 5.27912021e-01 -1.22688293e+00
2.72004187e-01 -4.71628189e-01 5.50298035e-01 -8.51407647e-01
7.49461204e-02 -8.53679955e-01 1.88059211e-01 5.41625857e-01
7.26660937e-02 -1.31198257e-01 4.31311488e-01 6.22264385e-01
-2.49465257e-01 -2.00995237e-01 8.71744037e-01 -2.77321666e-01
-1.29438722e+00 3.80246714e-03 -6.03218913e-01 4.20525717e-03
1.16156924e+00 -6.18806601e-01 -2.39027962e-01 -3.21731120e-01
-5.49891055e-01 2.70219535e-01 6.52673364e-01 5.44939339e-01
7.67410755e-01 -1.22110081e+00 -4.90310162e-01 2.47399613e-01
-2.11061448e-01 2.02316195e-01 3.70529592e-01 1.15472138e+00
-5.90713263e-01 1.31400283e-02 -3.17587793e-01 -8.42649460e-01
-1.44472003e+00 1.02314699e+00 2.71379203e-01 3.07910711e-01
-4.40673143e-01 6.59545183e-01 8.49928677e-01 4.63066548e-01
2.77713805e-01 -6.20262504e-01 -3.29824179e-01 7.81269670e-02
1.02232933e+00 4.85304534e-01 -2.84039229e-01 -9.57558036e-01
-3.55635345e-01 4.22322869e-01 -2.15226188e-02 1.98971912e-01
8.95279408e-01 -3.34096581e-01 -2.81298667e-01 5.26934266e-01
6.52363360e-01 6.60256520e-02 -1.54585707e+00 -1.29404113e-01
3.34427990e-02 -6.18508637e-01 1.05974652e-01 -4.55773234e-01
-1.30666733e+00 8.25073123e-01 7.62719393e-01 7.66039342e-02
1.45660853e+00 -3.70315611e-01 7.58563876e-01 5.16228415e-02
3.89615774e-01 -1.08902740e+00 2.66072810e-01 3.10691863e-01
6.77253067e-01 -1.53050923e+00 -1.49513334e-01 -4.87165213e-01
-7.11271882e-01 9.38080370e-01 9.71965909e-01 -1.02658428e-01
3.37978005e-01 1.44117042e-01 2.23731786e-01 2.25539491e-01
-3.96586478e-01 -5.34240246e-01 2.72154421e-01 5.83169699e-01
-7.91198313e-02 -1.31531879e-01 -4.65774648e-02 5.80575764e-01
2.17816874e-01 6.72333911e-02 8.47330451e-01 8.18160295e-01
-9.70885396e-01 -6.07696712e-01 -8.08481693e-01 2.10173696e-01
-5.48557043e-01 -2.70950258e-01 -2.16844574e-01 4.28551525e-01
2.01936275e-01 1.21536577e+00 2.71928813e-02 -1.71498880e-01
-1.21551812e-01 -1.15281574e-01 1.80720225e-01 -5.41239023e-01
-1.95972621e-01 5.49748003e-01 -4.04982567e-01 -4.74692494e-01
-8.86770070e-01 -6.65156305e-01 -1.29582036e+00 9.45094749e-02
-5.40026486e-01 -3.53800580e-02 3.74513924e-01 7.52638757e-01
7.89194331e-02 7.90258944e-01 3.57527018e-01 -1.12789214e+00
1.63873553e-01 -7.26635277e-01 -5.59366465e-01 3.91111642e-01
6.98788106e-01 -8.52569938e-01 -4.92387235e-01 4.65768069e-01] | [9.160443305969238, -0.4844220280647278] |
237e4eae-43de-4387-b032-7e7a480c3b36 | frigate-frugal-spatio-temporal-forecasting-on | 2306.08277 | null | https://arxiv.org/abs/2306.08277v1 | https://arxiv.org/pdf/2306.08277v1.pdf | FRIGATE: Frugal Spatio-temporal Forecasting on Road Networks | Modelling spatio-temporal processes on road networks is a task of growing importance. While significant progress has been made on developing spatio-temporal graph neural networks (Gnns), existing works are built upon three assumptions that are not practical on real-world road networks. First, they assume sensing on every node of a road network. In reality, due to budget-constraints or sensor failures, all locations (nodes) may not be equipped with sensors. Second, they assume that sensing history is available at all installed sensors. This is unrealistic as well due to sensor failures, loss of packets during communication, etc. Finally, there is an assumption of static road networks. Connectivity within networks change due to road closures, constructions of new roads, etc. In this work, we develop FRIGATE to address all these shortcomings. FRIGATE is powered by a spatio-temporal Gnn that integrates positional, topological, and temporal information into rich inductive node representations. The joint fusion of this diverse information is made feasible through a novel combination of gated Lipschitz embeddings with Lstms. We prove that the proposed Gnn architecture is provably more expressive than message-passing Gnns used in state-of-the-art algorithms. The higher expressivity of FRIGATE naturally translates to superior empirical performance conducted on real-world network-constrained traffic data. In addition, FRIGATE is robust to frugal sensor deployment, changes in road network connectivity, and temporal irregularity in sensing. | ['Sayan Ranu', 'Hariprasad Kodamana', 'Mridul Gupta'] | 2023-06-14 | null | null | null | null | ['spatio-temporal-forecasting'] | ['time-series'] | [ 2.73739785e-01 3.47619176e-01 -1.99491516e-01 -7.04451501e-02
-1.04592763e-01 -3.91208619e-01 6.71114862e-01 3.43219608e-01
-5.21588981e-01 8.63773048e-01 1.57304123e-01 -5.94033957e-01
-5.73338091e-01 -1.55422926e+00 -1.00013757e+00 -5.34412444e-01
-6.84141874e-01 5.04014254e-01 5.70395231e-01 -3.30239654e-01
-2.26894721e-01 8.48132372e-01 -1.41998851e+00 -4.02561575e-01
4.24206257e-01 1.00887346e+00 1.09783292e-01 7.71534681e-01
-1.46184787e-01 6.44656539e-01 -2.56644338e-01 -8.22141990e-02
1.88141719e-01 2.40897432e-01 -6.30496800e-01 9.37549304e-03
-1.10706314e-01 -2.76873976e-01 -9.76983547e-01 7.51548767e-01
2.32071415e-01 2.07407370e-01 4.43026155e-01 -1.55514491e+00
-7.48622596e-01 6.72132134e-01 -4.67852384e-01 2.80470043e-01
-7.48526081e-02 -9.40942243e-02 7.94519842e-01 -5.00274301e-01
4.40820307e-01 1.19447136e+00 8.17528069e-01 2.33356073e-01
-8.70216370e-01 -3.58479202e-01 5.86816490e-01 5.09536490e-02
-1.39161420e+00 -3.30127239e-01 7.85042048e-01 -9.11056101e-02
7.88158000e-01 8.84700865e-02 6.29065335e-01 1.07566047e+00
2.61806726e-01 4.10673320e-01 4.94750679e-01 -1.48908094e-01
3.29113871e-01 -2.18504354e-01 -2.93453902e-01 4.60742086e-01
2.67662525e-01 1.70086905e-01 -3.56880873e-01 8.23051482e-02
8.72591913e-01 3.71315688e-01 -6.65253475e-02 -2.07772642e-01
-1.01258540e+00 5.63889742e-01 9.69130516e-01 3.73452693e-01
-7.42294252e-01 6.98716223e-01 2.78827548e-01 4.40421164e-01
6.41429722e-01 -2.04473659e-01 -2.14602500e-01 -2.48051614e-01
-7.13792384e-01 -1.55576557e-01 6.71035171e-01 1.04913092e+00
7.91358650e-01 8.57500508e-02 2.73498058e-01 4.35882837e-01
2.17586458e-01 6.77236855e-01 -2.10833624e-02 -8.38419795e-01
6.03513956e-01 5.70443571e-01 1.94260888e-02 -1.40283430e+00
-7.29454517e-01 -4.12429512e-01 -1.42196643e+00 -1.01632029e-01
1.73792124e-01 -3.53121221e-01 -8.70491803e-01 1.81096351e+00
1.74545482e-01 7.05476105e-01 -5.02959304e-02 6.16411030e-01
3.79331529e-01 7.00878263e-01 1.25853643e-01 -8.61463137e-03
7.94258893e-01 -4.67837363e-01 -7.51738966e-01 -1.74362838e-01
5.76365352e-01 -1.02407910e-01 5.14800787e-01 -5.98701201e-02
-9.26800549e-01 -4.20188129e-01 -7.93797255e-01 2.09029421e-01
-9.50739264e-01 -3.22329879e-01 6.91932499e-01 6.20519102e-01
-1.49916744e+00 4.34722215e-01 -8.78962517e-01 -6.22944474e-01
4.74058777e-01 3.88406724e-01 -3.13129753e-01 -2.81530052e-01
-1.56651878e+00 7.21358240e-01 1.06055938e-01 7.89185584e-01
-8.58897746e-01 -4.91994351e-01 -8.16195965e-01 -7.42823854e-02
5.23063362e-01 -4.04354542e-01 7.97083676e-01 -4.02035266e-01
-1.15869367e+00 2.56568283e-01 -3.02690744e-01 -7.08553910e-01
6.07624114e-01 1.78759307e-01 -8.66335690e-01 -1.39311820e-01
1.82751636e-03 5.61571777e-01 4.37801152e-01 -1.10401154e+00
-7.28364289e-01 -1.92889884e-01 4.02443618e-01 -4.17511649e-02
-4.21934575e-01 -4.17659074e-01 -3.72653753e-01 -1.68977931e-01
8.16503391e-02 -9.24534380e-01 -5.24415076e-01 1.96120366e-01
-3.65709096e-01 -3.68226588e-01 1.13972855e+00 -1.43145591e-01
1.31370699e+00 -1.99803710e+00 -1.53044194e-01 5.15400529e-01
3.41630250e-01 3.51062894e-01 -3.32945198e-01 1.01011920e+00
3.20560157e-01 5.24857402e-01 -2.53908902e-01 -4.05863881e-01
-4.92295623e-02 7.67990708e-01 -2.88454175e-01 6.26721323e-01
5.71272969e-02 9.81945813e-01 -1.13022935e+00 -3.18101943e-01
5.72395384e-01 6.99725330e-01 -2.00118218e-02 -3.58502150e-01
-3.36363092e-02 3.08906019e-01 -5.25887072e-01 5.94529927e-01
6.61480963e-01 -1.38963014e-01 7.79385120e-02 3.35731179e-01
-2.54435509e-01 2.33772978e-01 -1.16399801e+00 1.33896637e+00
-6.64364994e-01 8.60694587e-01 -1.30774938e-02 -1.04926813e+00
8.07308614e-01 3.92372072e-01 5.53332627e-01 -8.25552166e-01
-1.39110116e-02 4.14767265e-02 -3.31098646e-01 -3.41361463e-01
4.38474298e-01 1.38097003e-01 6.63133245e-03 3.38299304e-01
-4.94306803e-01 1.95752531e-01 1.31242961e-01 1.35734200e-01
1.58748806e+00 -4.19198155e-01 -2.84275591e-01 1.35888875e-01
3.57285529e-01 -2.93276966e-01 5.40322006e-01 8.03522944e-01
-2.12312117e-01 1.98248580e-01 4.48486835e-01 -5.32999635e-01
-8.73886704e-01 -1.22784019e+00 9.50948372e-02 8.60320628e-01
1.86845928e-01 -6.67650923e-02 -2.58266568e-01 -3.06124151e-01
7.06887320e-02 3.16683382e-01 -8.61866176e-01 -3.46156098e-02
-6.42470360e-01 -5.76450527e-01 8.90361428e-01 7.17332006e-01
6.01369441e-01 -8.98239315e-01 -4.49205518e-01 5.89615703e-01
-5.27446344e-02 -1.32989383e+00 -1.38350821e-03 -3.61219049e-02
-7.81362116e-01 -1.09201717e+00 -3.05950165e-01 -5.86986065e-01
6.32395744e-01 5.12634635e-01 8.25105727e-01 2.07128134e-02
1.20603636e-01 4.75107163e-01 -2.63428867e-01 -4.75014687e-01
-1.44476429e-01 3.84508938e-01 3.67167778e-02 1.85346439e-01
2.76835263e-01 -8.72316957e-01 -4.48536694e-01 1.78749412e-01
-1.05346096e+00 -2.28698581e-01 5.76273978e-01 2.61747122e-01
6.13136947e-01 8.43274772e-01 6.42647088e-01 -5.65429509e-01
7.08837509e-01 -9.11480486e-01 -5.12638628e-01 4.91040312e-02
-3.70503008e-01 -2.14894056e-01 5.68009317e-01 -3.25026423e-01
-6.41730964e-01 -2.33270347e-01 -6.09210506e-02 -3.13319027e-01
-2.09768429e-01 8.30214381e-01 -8.09169039e-02 -4.17960063e-03
3.15477163e-01 -2.80124731e-02 -1.66716650e-01 -1.19327895e-01
3.93327743e-01 5.72749794e-01 3.09462696e-01 -2.68223137e-01
1.03159487e+00 9.09585834e-01 4.82505977e-01 -1.27997506e+00
-4.11013126e-01 -2.92364448e-01 -5.99196374e-01 -4.60532397e-01
4.77680266e-01 -8.39471519e-01 -8.60614657e-01 7.30271757e-01
-1.11768997e+00 -7.12581396e-01 -2.99692124e-01 3.64581108e-01
-1.48316652e-01 1.03012644e-01 -5.94144166e-01 -1.14453685e+00
1.34646356e-01 -5.93169868e-01 7.71209002e-01 -4.09313478e-02
7.39593506e-02 -1.63644171e+00 -1.36348620e-01 -2.79186189e-01
7.84709096e-01 6.26336157e-01 5.42981982e-01 -7.74046108e-02
-7.67702878e-01 -3.81625593e-01 -4.78727221e-01 1.54387400e-01
4.58847106e-01 8.05018321e-02 -6.76281333e-01 -9.45332274e-02
-5.21835029e-01 1.04217708e-01 8.28294396e-01 6.00514770e-01
1.13765895e+00 -3.16677511e-01 -6.06010914e-01 3.21529806e-01
1.37232041e+00 -3.45331728e-02 7.20942676e-01 1.31635219e-01
8.32912147e-01 7.79405177e-01 3.21019351e-01 2.66329288e-01
9.15623367e-01 2.61299670e-01 1.01296031e+00 -3.10390562e-01
-8.88120383e-02 -4.64358687e-01 2.66252428e-01 5.76758742e-01
-1.18873261e-01 -8.49653184e-01 -1.17239392e+00 1.10936368e+00
-1.92897820e+00 -9.46530282e-01 -5.74034333e-01 2.04091763e+00
9.66810733e-02 4.01494831e-01 1.34415496e-02 4.29094791e-01
7.20277250e-01 4.84734893e-01 -5.35459280e-01 -3.61847639e-01
-2.50475675e-01 -1.25225559e-01 1.27675581e+00 6.12558901e-01
-1.01228023e+00 7.66251981e-01 5.90364408e+00 4.26769197e-01
-1.02756011e+00 1.41373128e-01 2.30594903e-01 -1.05065610e-02
-4.16285008e-01 -2.09875658e-01 -5.58939040e-01 4.98010248e-01
1.36102653e+00 1.29353151e-01 4.92808193e-01 3.69980365e-01
6.57869220e-01 4.92276251e-02 -6.91657722e-01 5.36575198e-01
-4.00487304e-01 -1.38466179e+00 -1.81403473e-01 4.15375948e-01
6.09004498e-01 6.66504323e-01 8.07924122e-02 1.13841638e-01
7.99226642e-01 -1.10416341e+00 5.26718736e-01 6.33841038e-01
8.22196484e-01 -7.63327479e-01 8.07628453e-01 4.09670711e-01
-1.64268160e+00 -2.60531813e-01 -2.46026009e-01 -4.31250215e-01
6.02580845e-01 9.70140755e-01 -5.86177111e-01 7.46305525e-01
6.57524407e-01 1.00532138e+00 -2.11008310e-01 7.73953140e-01
-2.78855056e-01 6.83228493e-01 -8.26526523e-01 -7.57839734e-05
7.07360864e-01 1.80816557e-02 4.60167229e-01 1.02261317e+00
2.38736168e-01 -2.49688148e-01 1.92805946e-01 5.84174395e-01
-1.71405345e-01 -5.97893476e-01 -1.15432465e+00 1.18127957e-01
9.67542350e-01 8.52756023e-01 -5.81926703e-01 9.41190217e-03
-3.97483379e-01 4.94218171e-01 2.72414476e-01 8.74879777e-01
-8.56811643e-01 -2.21466586e-01 8.15106273e-01 5.03564954e-01
2.66712815e-01 -6.58519387e-01 -1.40755847e-01 -5.22549212e-01
1.39765143e-01 -7.70305023e-02 5.98901585e-02 -5.48303723e-01
-1.21818388e+00 3.93702298e-01 -1.22632898e-01 -9.24090683e-01
4.78264131e-02 -2.50013292e-01 -5.36301017e-01 5.76502681e-01
-2.02266455e+00 -1.37868631e+00 -2.97853053e-01 6.63492680e-01
1.89013720e-01 3.11337799e-01 5.25501549e-01 7.12887645e-01
-5.88286102e-01 1.38472632e-01 -2.77450494e-02 3.72968435e-01
7.23429024e-02 -9.47695374e-01 7.66663194e-01 1.02862072e+00
-3.54298167e-02 2.82216907e-01 4.69003081e-01 -6.93888545e-01
-1.33391607e+00 -1.71948123e+00 1.25337005e+00 -2.22032651e-01
1.02625132e+00 -4.33650255e-01 -8.63923311e-01 9.56130505e-01
-2.60489970e-01 4.28412229e-01 2.70023197e-01 1.30468160e-01
-1.08362213e-02 -3.85697961e-01 -1.02331364e+00 6.28167510e-01
1.52698994e+00 -6.39891446e-01 6.53693313e-03 2.19300926e-01
6.85154259e-01 -1.29830852e-01 -7.80004621e-01 4.70149577e-01
2.93647677e-01 -5.88538766e-01 8.18616927e-01 -3.33904088e-01
-7.31956065e-02 -3.71608764e-01 -1.34824365e-01 -1.31170118e+00
-2.07431585e-01 -5.57870328e-01 -1.03701137e-01 1.18208778e+00
3.78362536e-01 -1.13575351e+00 7.87891448e-01 5.02685010e-01
-1.20096639e-01 -7.14896381e-01 -1.35692370e+00 -9.21153724e-01
4.25180681e-02 -9.33790505e-01 8.93480420e-01 6.69377863e-01
-3.59647393e-01 2.15096660e-02 -3.98458928e-01 5.65284133e-01
5.28892159e-01 -6.34011745e-01 6.70037627e-01 -1.60417938e+00
4.40330058e-01 -4.29664969e-01 -6.14011407e-01 -1.18842316e+00
2.19675764e-01 -5.04603684e-01 -5.11437058e-02 -2.11093831e+00
-6.25961304e-01 -1.12868285e+00 -4.40633386e-01 5.20918787e-01
5.74856639e-01 2.54926775e-02 -2.76534349e-01 1.03903189e-02
-5.52879393e-01 6.37224853e-01 9.85717595e-01 -1.77031681e-01
-2.81866550e-01 1.38799354e-01 -3.39173406e-01 6.13890171e-01
1.15861630e+00 -4.55515087e-01 -7.59090602e-01 -7.06993878e-01
7.19774127e-01 1.98733836e-01 6.75343275e-01 -1.20861232e+00
5.92859626e-01 -1.30621493e-01 -1.67711690e-01 -7.62738347e-01
4.94171977e-01 -1.23286462e+00 3.12408805e-01 2.84405917e-01
-9.17471647e-02 2.36879334e-01 1.16361201e-01 1.25452340e+00
-1.50625303e-01 5.17693758e-01 2.33478591e-01 8.98698941e-02
-7.45746672e-01 7.20477819e-01 -7.94532895e-01 -1.75902784e-01
1.02591324e+00 -5.13861835e-01 -3.19308162e-01 -6.43703878e-01
-4.25880820e-01 5.75075090e-01 2.99209803e-01 6.04550838e-01
5.82304776e-01 -1.20670402e+00 -5.67027569e-01 1.12020008e-01
3.42066325e-02 2.46301249e-01 2.55887061e-01 9.18182075e-01
-2.66741335e-01 4.73919362e-01 9.72541273e-02 -3.40888679e-01
-8.24111342e-01 5.13675809e-01 2.39835888e-01 -4.07077014e-01
-6.12055004e-01 5.98406613e-01 -4.16143268e-01 -5.48636615e-01
3.38203251e-01 -6.59912348e-01 -4.01217304e-02 1.08336192e-02
1.47918954e-01 6.56867743e-01 4.15961118e-03 -6.70900524e-01
-3.75794441e-01 4.56313610e-01 5.55869162e-01 -1.52285218e-01
1.31126118e+00 -7.02120423e-01 -1.82228908e-02 7.12057233e-01
9.24462676e-01 -2.62093514e-01 -1.11013258e+00 -5.12662888e-01
-8.39749258e-03 -4.02400978e-02 2.05498204e-01 -3.09651643e-01
-1.26631165e+00 9.28487301e-01 2.84921199e-01 6.82522714e-01
9.94136930e-01 -1.30050018e-01 1.03216910e+00 4.46248293e-01
6.81364894e-01 -1.15268958e+00 -3.99228483e-01 6.65831506e-01
4.72301215e-01 -1.08132541e+00 -3.19558263e-01 -3.53861034e-01
1.68358367e-02 7.89856434e-01 2.36395940e-01 3.37503143e-02
1.17752695e+00 2.72540182e-01 -2.18504310e-01 -3.72247458e-01
-8.27413857e-01 -5.84536850e-01 -3.29139590e-01 8.85700405e-01
-1.25690266e-01 2.76870787e-01 1.93667784e-01 -1.85239896e-01
7.06314072e-02 1.20156355e-01 7.52001822e-01 8.50021601e-01
-4.89702135e-01 -9.60736990e-01 -9.09699723e-02 3.66971046e-01
-1.30949944e-01 1.27400473e-01 -5.54729328e-02 9.96982992e-01
1.83959335e-01 1.41683650e+00 3.86687845e-01 -3.80394608e-01
3.49355906e-01 -4.39462036e-01 1.09168841e-02 -3.56735229e-01
4.54213694e-02 -5.51908493e-01 1.94136679e-01 -4.33437645e-01
-5.61953306e-01 -5.50788581e-01 -1.30368555e+00 -9.27392185e-01
-2.41621155e-02 -1.12510398e-01 7.68316627e-01 9.59441543e-01
5.45774400e-01 8.52371037e-01 5.38719952e-01 -6.68018639e-01
1.73721716e-01 -7.60632634e-01 -6.58011019e-01 -1.16936870e-01
7.04364300e-01 -8.32205951e-01 -3.43903631e-01 -3.86314869e-01] | [6.459964752197266, 2.058907985687256] |
0ae85a04-3b52-41b8-8b1b-6d7da3f95642 | keyphrase-generation-with-gans-in-low | null | null | https://aclanthology.org/2020.sustainlp-1.12 | https://aclanthology.org/2020.sustainlp-1.12.pdf | Keyphrase Generation with GANs in Low-Resources Scenarios | Keyphrase Generation is the task of predicting Keyphrases (KPs), short phrases that summarize the semantic meaning of a given document. Several past studies provided diverse approaches to generate Keyphrases for an input document. However, all of these approaches still need to be trained on very large datasets. In this paper, we introduce BeGanKP, a new conditional GAN model to address the problem of Keyphrase Generation in a low-resource scenario. Our main contribution relies in the Discriminator’s architecture: a new BERT-based module which is able to distinguish between the generated and humancurated KPs reliably. Its characteristics allow us to use it in a low-resource scenario, where only a small amount of training data are available, obtaining an efficient Generator. The resulting architecture achieves, on five public datasets, competitive results with respect to the state-of-the-art approaches, using less than 1% of the training data. | ['Carlo Tasso', 'Giuseppe Serra', 'Beatrice Portelli', 'Saida S.Mohamed', 'Giuseppe Lancioni'] | null | null | null | null | emnlp-sustainlp-2020-11 | ['keyphrase-generation'] | ['natural-language-processing'] | [ 3.77767026e-01 3.34963143e-01 -2.51522452e-01 2.23664150e-01
-1.30881667e+00 -9.49554145e-01 1.26477468e+00 3.78687650e-01
-3.78817499e-01 1.09241056e+00 3.44457835e-01 -1.81301743e-01
2.74738550e-01 -9.23624337e-01 -8.24374318e-01 -5.75879753e-01
2.89179951e-01 6.53609157e-01 2.40476847e-01 -4.61889476e-01
4.63333398e-01 1.91496149e-01 -1.59432566e+00 5.65636158e-01
6.86427891e-01 9.69343066e-01 1.51206121e-01 1.04967535e+00
-3.01869750e-01 1.03373098e+00 -9.91858423e-01 -5.62528729e-01
1.51157260e-01 -6.95730090e-01 -9.79588747e-01 -3.83526951e-01
5.53837240e-01 -3.74801487e-01 -1.84674084e-01 7.05870986e-01
5.05616307e-01 -2.06279650e-01 8.37848425e-01 -1.26902580e+00
-5.41541874e-01 1.19451237e+00 5.78908287e-02 1.27737463e-01
5.49033105e-01 -1.73045963e-01 1.47578478e+00 -8.92825603e-01
8.82120848e-01 8.90292346e-01 3.39195013e-01 5.44025242e-01
-1.23502231e+00 -6.29675210e-01 -2.02632055e-01 6.94290102e-02
-1.24982417e+00 -2.20083371e-01 6.67726636e-01 -1.30039468e-01
8.19388330e-01 4.16820765e-01 6.10683084e-01 1.60966730e+00
2.18891785e-01 1.16475880e+00 1.22650051e+00 -5.52510560e-01
2.77305961e-01 3.28197062e-01 -2.35545095e-02 3.45385760e-01
2.75570422e-01 -2.28928998e-01 -8.34810317e-01 -4.87069428e-01
3.19004208e-01 -3.58927548e-01 -3.01363736e-01 1.36532532e-02
-1.59225750e+00 8.69644165e-01 6.67271838e-02 6.99892342e-01
-4.71588224e-01 3.15657437e-01 1.43820226e-01 4.03026193e-01
6.48173571e-01 1.04767442e+00 -5.86548328e-01 -4.63888079e-01
-1.29724848e+00 8.50080371e-01 1.23270488e+00 9.95564342e-01
6.04730129e-01 -3.30257505e-01 -7.65358984e-01 2.67386138e-01
-3.33516717e-01 5.09126008e-01 5.62445223e-01 -2.98082173e-01
6.41030312e-01 5.27576029e-01 3.46637964e-01 -9.74482954e-01
-1.00003906e-01 -2.95951247e-01 -7.84046352e-01 -3.58458608e-01
1.18683405e-01 -3.09298456e-01 -9.67673421e-01 1.50023329e+00
-1.87151961e-03 4.18712705e-01 2.94381708e-01 2.54111022e-01
8.89313161e-01 9.70594049e-01 -1.02254443e-01 -7.37816561e-03
1.42050076e+00 -9.10730302e-01 -5.47495306e-01 -1.44045770e-01
3.71399850e-01 -8.13400745e-01 1.19988501e+00 4.93103653e-01
-9.43807721e-01 -3.10521722e-01 -9.93956447e-01 -1.03648484e-01
-7.82930315e-01 4.41934198e-01 6.02153182e-01 4.88980472e-01
-1.20652092e+00 5.70212245e-01 -3.02243888e-01 -2.47481748e-01
2.34019414e-01 4.30845730e-02 -2.82317728e-01 2.06719801e-01
-1.49566615e+00 6.88053608e-01 8.84454608e-01 -3.64401102e-01
-9.61642802e-01 -1.07830644e+00 -4.77159590e-01 2.08022043e-01
7.31473207e-01 -9.86192048e-01 1.51354265e+00 -5.38241804e-01
-1.44601285e+00 7.11347520e-01 -9.17518139e-03 -6.51499152e-01
5.81528008e-01 -4.44870859e-01 -1.57590657e-01 2.74538070e-01
2.31451228e-01 7.02852666e-01 1.20611131e+00 -1.20336652e+00
-8.45017195e-01 2.40559638e-01 1.24518692e-01 3.05746701e-02
-4.55862582e-01 -1.18027674e-02 -4.48449731e-01 -1.21925747e+00
-5.27681708e-01 -9.06600416e-01 -1.73700899e-01 -6.94563568e-01
-7.95105398e-01 -3.96088958e-01 4.73946095e-01 -5.95190525e-01
1.29440701e+00 -1.75195837e+00 2.06444606e-01 1.75762355e-01
1.36486679e-01 2.76711375e-01 -1.45736381e-01 1.11142004e+00
1.55661196e-01 3.99049163e-01 -2.41566986e-01 -3.94945621e-01
1.46811426e-01 -1.54935475e-02 -1.21400928e+00 -3.04744154e-01
3.77081275e-01 1.14087379e+00 -1.07068729e+00 -4.07395810e-01
-2.54138142e-01 1.90305904e-01 -2.45526299e-01 6.07458591e-01
-7.80004025e-01 2.15085328e-01 -5.98642170e-01 2.07560673e-01
3.87068808e-01 -2.35277936e-01 -5.76060973e-02 -1.76859368e-02
1.51399061e-01 5.36418974e-01 -1.20069540e+00 1.77855921e+00
-5.49357295e-01 4.59383756e-01 -4.74359900e-01 -7.07801342e-01
8.40475380e-01 5.90501189e-01 4.01243836e-01 -2.89688200e-01
5.08684963e-02 4.63966906e-01 -6.41200185e-01 9.32194442e-02
1.08561039e+00 1.18889496e-01 -4.58587259e-01 9.66240346e-01
3.08065265e-01 -5.79290211e-01 6.64995849e-01 6.71267271e-01
1.35897875e+00 -2.71332651e-01 4.42669809e-01 -1.62446201e-01
5.82554936e-01 5.50254025e-02 -1.78317368e-01 1.23450208e+00
7.64895201e-01 7.22009301e-01 7.14255571e-01 -2.85402179e-01
-1.03472912e+00 -7.09375441e-01 4.20086384e-01 9.22755599e-01
-1.01926841e-01 -1.11017108e+00 -9.39129174e-01 -1.02941775e+00
-1.03342988e-01 9.69036937e-01 -6.67271912e-01 -1.42473221e-01
-5.61809480e-01 -5.99158406e-01 8.63024473e-01 3.06349903e-01
2.76688397e-01 -1.29213798e+00 -5.89776635e-01 2.60699749e-01
-3.48973334e-01 -1.37906229e+00 -2.41424054e-01 1.42759860e-01
-3.68508190e-01 -9.67238963e-01 -9.92732584e-01 -3.72230947e-01
5.92753291e-01 2.21513212e-02 1.65536916e+00 -2.91677807e-02
-8.37347135e-02 5.11611879e-01 -8.36757421e-01 -5.53877354e-01
-8.89075398e-01 7.62418807e-01 -2.31412619e-01 5.76242954e-02
6.42417520e-02 -4.68237460e-01 -3.85694772e-01 -3.25688064e-01
-1.14523387e+00 3.75399441e-01 8.83467376e-01 9.02288139e-01
4.08336252e-01 2.12253496e-01 3.76941800e-01 -1.23157752e+00
1.18597531e+00 -4.18401986e-01 -3.93411249e-01 4.02754128e-01
-8.38037670e-01 4.57199484e-01 1.02090085e+00 -3.82874578e-01
-7.74415791e-01 -4.24523532e-01 5.77149447e-03 3.75814103e-02
-4.73380610e-02 5.16986787e-01 1.40819728e-01 2.55434811e-01
7.08158016e-01 7.12867558e-01 -5.60856819e-01 -5.86610019e-01
9.02427077e-01 6.20178699e-01 6.89345777e-01 -7.44223952e-01
1.17502952e+00 2.68586397e-01 -5.00171771e-03 -5.34216881e-01
-1.14308977e+00 -4.90873456e-01 -4.48089212e-01 2.02976987e-01
3.78204733e-01 -9.67348397e-01 -1.86601311e-01 4.39658850e-01
-1.24403083e+00 -2.01597348e-01 -6.80061340e-01 1.12414956e-02
-6.41838253e-01 2.33866900e-01 -3.67733240e-01 -4.85892326e-01
-1.09172833e+00 -6.67979658e-01 1.41639590e+00 1.05744809e-01
-3.99735749e-01 -7.13986456e-01 2.61812627e-01 5.96685037e-02
5.09303927e-01 2.52634138e-01 8.65480900e-01 -1.11808586e+00
-5.74008942e-01 -4.06233490e-01 -1.13691054e-01 4.65095162e-01
2.48829290e-01 -6.93190321e-02 -9.06893432e-01 -2.12992027e-01
-1.31649211e-01 -6.60691082e-01 1.11027384e+00 -2.00869858e-01
1.18423676e+00 -7.55659461e-01 -1.52847320e-01 3.86662245e-01
1.18392324e+00 -2.30723128e-01 4.57567722e-01 2.50664353e-01
6.51103914e-01 3.83869886e-01 5.55037200e-01 4.62601066e-01
5.22649705e-01 5.27963281e-01 1.15394562e-01 -2.44186111e-02
-1.77727371e-01 -8.65354955e-01 4.03412580e-01 7.30058551e-01
6.30843565e-02 -5.78364670e-01 -8.04627061e-01 6.33457780e-01
-1.87663615e+00 -9.57801342e-01 1.19349457e-01 1.82418215e+00
1.37837780e+00 2.59279370e-01 -5.25407195e-02 3.37111235e-01
1.51791662e-01 4.30128217e-01 -8.99498835e-02 -1.96341753e-01
-2.08041564e-01 6.95694983e-01 4.89391029e-01 1.91874266e-01
-1.13258886e+00 1.17706263e+00 6.58432341e+00 1.15231073e+00
-1.11310923e+00 -1.75242752e-01 4.45014685e-01 2.93699116e-01
-7.16977000e-01 -1.72256934e-03 -1.13586223e+00 5.96238196e-01
1.16963661e+00 -5.81875801e-01 3.39624256e-01 9.55509484e-01
-2.26585746e-01 -1.25494525e-01 -1.18683255e+00 9.80179906e-01
3.17079574e-01 -1.63230586e+00 4.79839921e-01 -3.59011233e-01
7.81575561e-01 -2.92103022e-01 6.67985082e-02 3.03031385e-01
5.04685283e-01 -1.08157825e+00 8.18371534e-01 4.98380154e-01
8.51234257e-01 -6.19614959e-01 7.60720372e-01 5.13715565e-01
-9.98870373e-01 2.10500211e-01 -2.16092363e-01 2.45495349e-01
-6.23371638e-02 8.79912913e-01 -1.34850097e+00 8.13236833e-01
4.06101108e-01 5.23565769e-01 -9.80258167e-01 5.31724751e-01
-7.48219311e-01 7.41342604e-01 -2.96040565e-01 -3.47004950e-01
4.00020868e-01 1.34934783e-01 4.41696405e-01 1.44946134e+00
4.29934591e-01 -2.32524365e-01 1.28252134e-01 9.25802290e-01
-3.37629527e-01 2.79293597e-01 -5.30760944e-01 -5.90101540e-01
3.76675904e-01 1.52756441e+00 -7.93851674e-01 -7.25533783e-01
-2.36220807e-02 1.26186037e+00 1.35143861e-01 1.27088726e-01
-4.80076224e-01 -3.80776256e-01 2.67515600e-01 -4.80760122e-04
4.54340488e-01 -7.98188224e-02 -2.60684583e-02 -1.42638004e+00
3.09136301e-01 -1.30143297e+00 4.12825137e-01 -7.15940058e-01
-1.17951906e+00 7.02916384e-01 2.70702899e-01 -9.67427909e-01
-9.52014208e-01 -4.80245233e-01 -5.30592859e-01 9.16982830e-01
-1.61654663e+00 -1.59777665e+00 -3.19609940e-01 4.46653515e-01
3.91247511e-01 -3.55986357e-01 1.25958240e+00 -2.00376749e-01
9.48065370e-02 5.69621146e-01 4.50925063e-03 7.42487162e-02
8.02368701e-01 -1.76889753e+00 1.09685874e+00 9.55951273e-01
6.81451082e-01 5.18546224e-01 8.33146214e-01 -4.79710340e-01
-1.38203120e+00 -9.19091046e-01 1.31992590e+00 -6.76114440e-01
8.58374357e-01 -7.57876933e-01 -5.29038310e-01 3.80029082e-01
2.86803454e-01 -1.87129736e-01 6.58048570e-01 -1.91685155e-01
-4.78687584e-01 1.45044634e-02 -7.66074717e-01 7.91605353e-01
7.82181203e-01 -7.05928266e-01 -7.48260081e-01 3.71423066e-01
9.33883727e-01 -4.29828733e-01 -5.76113105e-01 2.15032592e-01
2.90951818e-01 -7.86050856e-01 7.98089325e-01 -6.27606273e-01
7.54378974e-01 -2.23767266e-01 1.28209189e-01 -1.59708142e+00
8.08514729e-02 -1.16492903e+00 -6.23910069e-01 1.37495005e+00
5.13243556e-01 -2.81063437e-01 7.43683815e-01 2.00302780e-01
2.55506724e-01 -6.92463994e-01 -6.08491302e-01 -5.66423714e-01
-9.12663899e-03 -4.07972455e-01 8.16970766e-01 6.13808334e-01
-2.36167535e-01 7.58219361e-01 -6.06477320e-01 -3.46824497e-01
2.98773646e-01 2.88556635e-01 1.22430456e+00 -1.14604664e+00
-3.79831284e-01 -2.22403228e-01 -1.24179594e-01 -9.34556782e-01
2.39056066e-01 -9.16138649e-01 -1.11421019e-01 -1.53971779e+00
5.01716509e-02 -2.41491944e-01 -1.96061388e-01 6.02696776e-01
-5.50509095e-01 1.59785062e-01 3.27832431e-01 1.26302272e-01
-3.50404859e-01 5.11721730e-01 9.56455290e-01 -3.69623154e-01
-8.42330009e-02 1.68034568e-01 -1.11167967e+00 3.57997894e-01
7.61639595e-01 -6.14477456e-01 -5.47223628e-01 4.04579900e-02
6.72645390e-01 -2.32096225e-01 3.67854536e-01 -9.33405519e-01
3.08453053e-01 -6.77578971e-02 1.82328910e-01 -6.95539057e-01
2.00178102e-01 -6.25904441e-01 7.90597871e-02 3.09744298e-01
-5.63060343e-01 1.44324258e-01 7.78460726e-02 3.85987252e-01
-3.36240172e-01 -3.37732613e-01 2.65579104e-01 -3.75216514e-01
-7.22833157e-01 3.01104903e-01 -2.67337501e-01 3.28306556e-01
8.10105920e-01 3.68360668e-01 -3.39166850e-01 -6.34048760e-01
-7.57910684e-02 -2.52295792e-01 2.98373789e-01 6.11759126e-01
5.45624733e-01 -1.16067970e+00 -1.02614427e+00 1.45313755e-01
3.65377039e-01 -7.10688066e-04 -2.23772049e-01 2.01500088e-01
-4.79595542e-01 7.59218037e-01 2.71685347e-02 -2.39505079e-02
-8.67143929e-01 3.95627081e-01 -1.66523173e-01 -1.02754581e+00
-6.87324047e-01 8.41933191e-01 -3.00065070e-01 -1.36979878e-01
-9.13646072e-02 -5.36092818e-01 -2.77513236e-01 4.21950758e-01
8.57906580e-01 -2.24715099e-03 3.06807220e-01 -2.67735869e-01
7.72179291e-02 2.49416560e-01 -3.81124705e-01 -3.64513874e-01
1.14650524e+00 4.10967112e-01 -3.38424712e-01 4.81431693e-01
1.01985228e+00 2.77177453e-01 -7.25584090e-01 -3.31009209e-01
1.83358386e-01 -3.19503397e-01 -1.00508250e-01 -8.78092110e-01
-5.33382833e-01 4.46574032e-01 9.60016027e-02 5.45027852e-01
1.01970971e+00 1.33289769e-01 1.06687558e+00 6.72172546e-01
5.22801161e-01 -1.07138360e+00 2.83711076e-01 5.88074982e-01
1.02126980e+00 -1.10237861e+00 1.52794629e-01 -3.04711550e-01
-6.81880951e-01 1.12714684e+00 8.45929757e-02 -6.93271542e-03
4.54046667e-01 1.86784238e-01 -3.48734222e-02 -1.21179067e-01
-1.08332765e+00 -1.77376300e-01 6.04175389e-01 5.35196006e-01
2.83812076e-01 7.80001059e-02 -3.59361917e-01 7.64654994e-01
-8.99550796e-01 -4.29452583e-03 6.52365506e-01 8.82166803e-01
-1.07805915e-01 -1.36552119e+00 1.79673284e-02 5.24764895e-01
-7.47455657e-01 -5.24010301e-01 -8.91296685e-01 4.74364728e-01
-1.18495174e-01 8.20414782e-01 -2.30453804e-01 -4.58326966e-01
1.98058009e-01 2.50831634e-01 2.08061531e-01 -9.23063159e-01
-9.40014660e-01 -3.75280082e-01 2.13450164e-01 -4.71134275e-01
-3.52748394e-01 -2.19207644e-01 -7.20190525e-01 -3.05656582e-01
-8.23861361e-02 5.71196735e-01 5.61366498e-01 8.16888332e-01
4.86656100e-01 3.08807552e-01 7.18680084e-01 -7.18186915e-01
-8.35500121e-01 -1.03515446e+00 -5.48727810e-01 5.25266588e-01
1.96067169e-01 -6.03475086e-02 -1.61435217e-01 2.84441620e-01] | [12.272200584411621, 8.910719871520996] |
3cc59e96-a330-4153-ad91-289b7b07e50c | table-retrieval-does-not-necessitate-table | null | null | https://openreview.net/forum?id=Z1Vgh4I_5ur | https://openreview.net/pdf?id=Z1Vgh4I_5ur | Table Retrieval Does Not Necessitate Table-specific Model Design | Tables are an important form of structured data for both human and machine readers alike, providing answers to questions that cannot, or cannot easily, be found in texts. Recent work designs special models and trains for table-related tasks such as table-based question answering and table retrieval. Though effective, they add model-data dual complexity to generic text solutions and obscure which elements are truly beneficial. In this work, we focus on the task of table retrieval, and ask: ``are table-specific model designs necessary for table retrieval, or can a text-generic model be effectively used to achieve a similar result?’’ We start by analyzing NQ-table, a set of table-answerable questions in the Natural Questions (NQ) dataset, and find 90\% of the questions can match tables in content with little concern for table structure. Motivated by this, we experiment with a general-purpose Dense Passage Retriever (DPR) for text and a special-purpose Dense Table Retriever (DTR) for tables. We show that DPR, without any design for or training on tables, can perform comparably well to the state-of-the-art DTR model, and neither adding DTR-like table-specific embeddings nor perturbing cell orders lead to significant changes. Both results strongly indicate that table retrieval does not necessitate table-specific model design, as well as the potential of directly applying powerful text-generic retrievers to structured tables. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['table-retrieval'] | ['natural-language-processing'] | [-7.22019151e-02 9.36399847e-02 -3.66076007e-02 -1.64335132e-01
-1.54523039e+00 -9.52487826e-01 4.56316501e-01 7.88325787e-01
-4.87914830e-01 5.57624459e-01 7.22277403e-01 -5.75824440e-01
-2.62557477e-01 -1.06243563e+00 -9.35066998e-01 -9.70050246e-02
3.18631560e-01 1.14002442e+00 4.86602843e-01 -9.06754434e-01
1.19716361e-01 2.70408541e-01 -1.35417032e+00 8.15449238e-01
8.50131333e-01 9.29304719e-01 2.83516645e-02 7.83290386e-01
-6.04241133e-01 1.18243515e+00 -7.85304308e-01 -7.58557916e-01
1.21474832e-01 -2.55946457e-01 -1.05245662e+00 -4.25945669e-01
8.07959020e-01 -4.05406386e-01 -7.54534304e-01 4.07900631e-01
5.86389124e-01 2.37726688e-01 6.88284755e-01 -6.48262739e-01
-1.17234993e+00 9.01120186e-01 -2.10135081e-03 4.67323422e-01
6.97357178e-01 -9.22170281e-02 1.64756739e+00 -8.31951499e-01
8.43449295e-01 1.38555884e+00 6.12296879e-01 4.84318584e-01
-1.26640165e+00 -1.61876723e-01 -1.87209919e-01 1.26811266e-01
-1.22377372e+00 -5.53084016e-01 3.88287216e-01 5.43565210e-03
1.29745102e+00 7.88078427e-01 2.46478990e-01 9.68436182e-01
2.06698582e-01 1.00969505e+00 4.95083898e-01 -2.79412627e-01
5.05723478e-03 4.26548302e-01 1.03616089e-01 3.89910817e-01
6.13340199e-01 -6.01473153e-01 -5.21403491e-01 -3.98696102e-02
3.35100025e-01 -1.96733817e-01 -4.47223097e-01 -2.73019671e-01
-1.22910011e+00 8.38248193e-01 4.88733262e-01 3.41734648e-01
-1.51268244e-01 4.33859490e-02 6.23062968e-01 6.94139421e-01
1.96852908e-01 1.13074625e+00 -7.28873491e-01 -1.66009277e-01
-7.33392477e-01 7.84449399e-01 1.24063253e+00 1.24795294e+00
6.29917264e-01 -3.03598583e-01 -6.33049965e-01 7.43808508e-01
-3.47218961e-02 6.70410872e-01 3.45449835e-01 -7.82810211e-01
9.06345785e-01 8.63646746e-01 1.71034470e-01 -1.11019623e+00
-1.94799796e-01 -2.12483078e-01 -4.86031115e-01 -7.11673319e-01
6.65129900e-01 2.46575892e-01 -6.81999922e-01 1.40787733e+00
1.17244400e-01 -7.89812207e-01 2.46603012e-01 5.53087115e-01
1.37748289e+00 9.40204859e-01 -2.95000494e-01 2.11548805e-01
1.59330964e+00 -6.95357084e-01 -9.24006701e-01 -2.85434097e-01
1.12502277e+00 -8.67656350e-01 1.64805400e+00 5.56154810e-02
-1.15800869e+00 -4.02404130e-01 -9.43415880e-01 -1.02096999e+00
-9.67937648e-01 -1.93450034e-01 4.65975791e-01 6.58963978e-01
-1.26608634e+00 2.85623312e-01 -4.21142519e-01 -5.97866893e-01
6.11464456e-02 9.68867242e-02 -3.36536735e-01 -5.25727272e-01
-1.37038541e+00 1.18139923e+00 3.22152004e-02 2.78117303e-02
-4.97132897e-01 -9.35538948e-01 -9.55598652e-01 4.61261183e-01
7.32368946e-01 -7.69998133e-01 1.18730116e+00 6.73016608e-02
-8.33810210e-01 8.16100538e-01 -3.93534064e-01 -4.98861045e-01
5.33091240e-02 -2.04367667e-01 -3.11825484e-01 2.25311235e-01
7.97165409e-02 4.63771820e-01 2.19810486e-01 -1.13237727e+00
-1.49231032e-01 -2.51471072e-01 5.47223210e-01 3.90513331e-01
-4.49607462e-01 -6.46564886e-02 -6.75199032e-01 -5.10746300e-01
-6.47189394e-02 -4.90204155e-01 9.64502469e-02 -1.34766065e-02
-3.86685371e-01 -4.42451388e-01 4.64800298e-01 -1.02974236e+00
1.71268785e+00 -1.89415598e+00 -7.76355639e-02 -4.24994109e-03
3.39189619e-01 1.40277356e-01 -2.77044386e-01 1.02851284e+00
4.68818635e-01 6.16158009e-01 -1.82471216e-01 -1.17856316e-01
3.98610115e-01 3.29612106e-01 -6.85339749e-01 2.16525104e-02
2.39275739e-01 1.32723391e+00 -6.50222659e-01 -6.24023020e-01
-7.66860247e-02 2.71858573e-01 -8.11859429e-01 1.26960233e-01
-6.06563866e-01 -4.34157550e-01 -4.91442442e-01 7.28651583e-01
2.47374237e-01 -4.79234010e-01 1.69591494e-02 -2.30321974e-01
3.37834537e-01 1.02287817e+00 -9.36616659e-01 1.43102944e+00
-6.07700944e-01 7.05820382e-01 -8.95895809e-02 -4.79456365e-01
7.55714059e-01 1.90447778e-01 8.33122954e-02 -1.28515303e+00
-1.46906078e-01 3.75538111e-01 -8.42197835e-02 -5.22434115e-01
1.01364577e+00 1.08300619e-01 -2.65410930e-01 4.88459826e-01
-8.03963095e-02 -4.56482798e-01 5.50423384e-01 6.71173930e-01
1.61788189e+00 -2.49531403e-01 -7.06522837e-02 -3.94594848e-01
4.69867647e-01 3.37189645e-01 -1.41447708e-01 1.03129125e+00
2.15039864e-01 6.94804311e-01 4.55170542e-01 -2.63227880e-01
-1.04204059e+00 -1.12230968e+00 -2.20178977e-01 1.25324082e+00
1.52033925e-01 -7.63674557e-01 -4.85915363e-01 -6.35155141e-01
3.78249258e-01 8.42779934e-01 -7.30221629e-01 -1.34619027e-01
-8.57428670e-01 -4.65029329e-01 3.73025954e-01 6.50968909e-01
1.77164510e-01 -1.02498305e+00 -2.75358081e-01 3.01510274e-01
-5.35557210e-01 -9.97007608e-01 -7.38644719e-01 3.90349597e-01
-7.19993472e-01 -9.30892289e-01 -5.17874837e-01 -7.03945100e-01
2.12944061e-01 3.42009693e-01 1.93344700e+00 2.84005940e-01
-5.79094552e-02 7.04456568e-01 -4.71892178e-01 -2.47285008e-01
-4.57751244e-01 5.58578670e-01 -4.41673040e-01 -5.55134237e-01
5.72221100e-01 -1.97053745e-01 -4.39383149e-01 2.90462613e-01
-1.20268989e+00 -4.18708891e-01 3.97707999e-01 6.95937157e-01
4.12055939e-01 -6.39194176e-02 6.46930873e-01 -1.18130302e+00
1.06003809e+00 -3.80592674e-01 -3.92309695e-01 6.37552798e-01
-5.11363208e-01 5.53751171e-01 8.93542588e-01 -2.21940786e-01
-5.54565907e-01 -7.10227072e-01 -4.51353401e-01 -1.67556912e-01
4.48548943e-01 6.91205382e-01 -3.55034798e-01 3.02157044e-01
1.01331246e+00 2.16180563e-01 -2.85375923e-01 -6.06297553e-01
5.41064501e-01 5.73063910e-01 1.63841754e-01 -8.09801042e-01
9.09736633e-01 3.76756757e-01 -3.71639937e-01 -6.10198855e-01
-1.04466295e+00 -6.89940751e-01 -2.24706426e-01 2.44922861e-01
9.27115798e-01 -8.61744463e-01 -7.51006246e-01 -2.51991302e-01
-9.95700002e-01 -4.21131164e-01 -6.38454735e-01 -1.05829991e-01
-3.90345395e-01 -1.20282797e-02 -8.47269416e-01 -4.92361665e-01
-3.81616712e-01 -8.56580079e-01 1.20839608e+00 -7.39028007e-02
-3.66648525e-01 -1.20246148e+00 3.62523012e-02 5.44717431e-01
8.79544735e-01 -8.04778636e-02 1.46434355e+00 -1.09716403e+00
-9.75098491e-01 -2.67597497e-01 -2.55485266e-01 5.33449836e-02
1.21471159e-01 -2.63805568e-01 -9.12719131e-01 -1.68997273e-01
-1.24659956e-01 -6.97650254e-01 9.25619185e-01 -3.37408222e-02
1.05844676e+00 -6.20976210e-01 -1.04739331e-01 2.65604824e-01
1.41607332e+00 8.93690288e-02 8.11342359e-01 3.91455799e-01
7.98245311e-01 7.36672461e-01 2.75174171e-01 3.71165723e-02
8.79787266e-01 5.39891660e-01 2.25708932e-02 1.33249789e-01
-2.14124262e-01 -7.62682557e-01 9.11256522e-02 1.10058856e+00
7.90766120e-01 -9.05780435e-01 -9.71252024e-01 7.89948642e-01
-1.43815386e+00 -7.69901991e-01 -1.36416912e-01 2.18020129e+00
1.13990510e+00 2.69222379e-01 -1.02410071e-01 2.27791727e-01
2.07347393e-01 2.98072159e-01 -4.33465451e-01 -6.03999734e-01
-2.58005202e-01 5.49157321e-01 4.74994451e-01 4.35349911e-01
-7.49580979e-01 8.84671509e-01 6.44270802e+00 1.00105762e+00
-5.81939101e-01 -2.26938993e-01 5.22268534e-01 -6.31584674e-02
-1.09222722e+00 -8.97140577e-02 -9.79465485e-01 1.41675835e-02
1.17318714e+00 -4.27791238e-01 6.37647510e-01 4.59210992e-01
-2.99105763e-01 8.80816430e-02 -1.42970395e+00 7.92863905e-01
2.32353568e-01 -1.57076824e+00 5.92179775e-01 -2.37477466e-01
1.68712288e-01 -4.37880605e-01 1.48967609e-01 9.81955588e-01
2.58317798e-01 -1.29735076e+00 6.20059431e-01 3.07827353e-01
6.62683129e-01 -5.11441469e-01 6.54426455e-01 1.66618854e-01
-1.15776086e+00 1.48982435e-01 -6.34325445e-01 1.75315320e-01
-2.59322494e-01 2.50937998e-01 -7.18910456e-01 4.64887321e-01
8.81306469e-01 4.15363729e-01 -1.11096680e+00 6.27929270e-01
7.04523697e-02 5.87787569e-01 -5.14752448e-01 -5.60635448e-01
2.87309885e-01 1.92586526e-01 2.50085086e-01 1.08376813e+00
3.93719189e-02 -6.19190931e-02 -2.97840536e-01 8.82377028e-01
-5.72157919e-01 3.17827195e-01 -7.53032327e-01 -3.23157609e-01
6.13973796e-01 1.03192484e+00 -6.66458368e-01 -4.52827007e-01
-3.77906710e-01 7.43310153e-01 4.52507198e-01 3.43264878e-01
-4.12214309e-01 -6.40966833e-01 3.84153008e-01 4.24684733e-01
4.77997333e-01 9.89209116e-03 -4.84931260e-01 -1.35807133e+00
5.41858554e-01 -1.41688800e+00 6.77487314e-01 -9.02627230e-01
-1.37000275e+00 5.76326728e-01 -9.01428536e-02 -8.45214903e-01
-2.86980510e-01 -5.23995399e-01 -9.67186093e-02 8.78386736e-01
-1.46397018e+00 -7.10376084e-01 7.46250600e-02 5.78458607e-01
4.35930669e-01 2.29421183e-01 7.93249726e-01 4.83090341e-01
-1.67709514e-01 9.63486612e-01 3.20275366e-01 3.41227412e-01
8.36521745e-01 -1.63146460e+00 7.65567362e-01 6.99843705e-01
4.74663109e-01 1.16094637e+00 8.02356541e-01 -3.73626828e-01
-2.15485311e+00 -9.96769071e-01 1.38958859e+00 -1.50972605e+00
6.77465498e-01 -9.90503192e-01 -1.26547444e+00 6.57525599e-01
2.40629748e-01 -1.97391927e-01 5.34359813e-01 5.23213089e-01
-7.33718753e-01 -4.09855485e-01 -1.17570996e+00 9.57638085e-01
8.84809375e-01 -9.24811780e-01 -8.83657932e-01 4.81281668e-01
1.39982355e+00 -5.81654072e-01 -9.89072978e-01 2.48894230e-01
3.87114316e-01 -5.99043727e-01 1.09210098e+00 -6.62229776e-01
5.83870113e-01 -1.11152418e-01 -4.54137951e-01 -9.47771192e-01
-9.00761485e-02 -4.03427392e-01 -3.78781408e-01 1.25739515e+00
8.73744667e-01 -5.05682707e-01 6.72782719e-01 8.36118162e-01
-1.54451713e-01 -9.49754894e-01 -8.07718694e-01 -7.19457984e-01
6.25138879e-01 -1.95723400e-01 7.72737980e-01 7.12302864e-01
-6.96476698e-02 6.57121778e-01 1.61499485e-01 -1.95479766e-02
2.89756339e-02 9.31321383e-02 8.87760282e-01 -9.21073318e-01
-2.75144726e-01 -3.67274225e-01 3.43906507e-02 -1.35819936e+00
-1.38763070e-01 -1.05395710e+00 -5.55829145e-04 -2.22646642e+00
4.81255352e-02 -4.42704529e-01 -5.02070738e-03 3.33031356e-01
-3.02925020e-01 -8.42555389e-02 2.92241335e-01 6.76771626e-02
-8.48656178e-01 5.00041485e-01 1.46496105e+00 -3.78855348e-01
8.64808932e-02 -2.91612118e-01 -1.29865098e+00 -2.21679583e-01
4.01952207e-01 -4.53944147e-01 -6.93863869e-01 -7.52088130e-01
8.74616325e-01 4.77217615e-01 -8.17503110e-02 -9.25143778e-01
3.11660558e-01 1.90999433e-01 2.47450113e-01 -6.38498306e-01
2.73745000e-01 -6.87232554e-01 -4.16241378e-01 -2.12635156e-02
-7.97874451e-01 5.72607100e-01 6.69445634e-01 3.95407051e-01
-3.70861262e-01 -2.41388828e-01 2.56906897e-01 -2.45068282e-01
-9.67960134e-02 2.46284917e-01 -5.41254222e-01 9.65222418e-01
9.22582597e-02 -1.31177828e-02 -6.59404695e-01 -7.04517782e-01
-9.23598409e-02 5.48068464e-01 4.79459316e-01 5.37377119e-01
5.45168161e-01 -1.08138108e+00 -7.27684736e-01 -5.17659113e-02
4.86063242e-01 1.03015408e-01 8.27866718e-02 2.37482041e-01
-6.75241888e-01 8.93024266e-01 4.43039209e-01 -3.12136710e-01
-6.25693083e-01 8.27619493e-01 2.44289234e-01 -8.12717557e-01
-6.21824861e-01 9.01905417e-01 8.29137117e-02 -9.04424071e-01
3.06838274e-01 -9.31038618e-01 -1.07246198e-01 1.65249214e-01
5.05371392e-01 -2.82582492e-02 6.34186447e-01 2.79027857e-02
-3.00301611e-01 2.29434699e-01 -4.27828819e-01 -8.57260078e-02
1.07263815e+00 -1.99977458e-01 6.22532703e-02 6.37937784e-01
1.40295005e+00 3.65891993e-01 -3.73285621e-01 -2.31069848e-01
1.60825223e-01 -2.74402052e-01 -2.45774209e-01 -1.04020321e+00
-6.73599780e-01 9.94950116e-01 6.73518106e-02 4.92421567e-01
9.59464252e-01 1.27887288e-02 1.23046684e+00 1.08260417e+00
3.04703504e-01 -6.74653530e-01 1.01965018e-01 8.13392401e-01
9.22242641e-01 -1.13184810e+00 -1.39519528e-01 -2.20104423e-03
-4.12906587e-01 7.90695727e-01 5.56087375e-01 6.35494962e-02
3.76518011e-01 2.00943127e-01 1.53429255e-01 -3.87972772e-01
-1.26707530e+00 -4.22005296e-01 4.74422812e-01 4.61557657e-01
6.18921936e-01 -3.83002400e-01 -1.98691990e-02 3.54420960e-01
-6.02285802e-01 -3.96272928e-01 5.77066004e-01 1.07017612e+00
-4.43496227e-01 -9.28171694e-01 -1.69811174e-01 9.86737192e-01
-4.52054769e-01 -6.10665083e-01 -4.35121953e-01 9.77824509e-01
-5.36490619e-01 1.07481992e+00 1.54807836e-01 -3.70669514e-01
5.57601750e-01 1.99697450e-01 3.87576431e-01 -7.76158571e-01
-1.04085040e+00 -5.55991948e-01 4.13783640e-01 -4.15460855e-01
3.23486686e-01 -3.87916595e-01 -9.57995057e-01 -6.75789595e-01
-2.48603120e-01 4.65214372e-01 2.34253034e-01 6.70884669e-01
4.38945621e-01 4.07479137e-01 2.04760596e-01 1.28763735e-01
-6.73427105e-01 -9.38768566e-01 -3.25505406e-01 3.90275270e-01
4.24475998e-01 -2.07469374e-01 -5.60353279e-01 -3.14301044e-01] | [10.157649993896484, 7.945359230041504] |
9ef0d0fa-ccac-4542-a049-eb46f25460e8 | raytran-3d-pose-estimation-and-shape | 2203.13296 | null | https://arxiv.org/abs/2203.13296v2 | https://arxiv.org/pdf/2203.13296v2.pdf | RayTran: 3D pose estimation and shape reconstruction of multiple objects from videos with ray-traced transformers | We propose a transformer-based neural network architecture for multi-object 3D reconstruction from RGB videos. It relies on two alternative ways to represent its knowledge: as a global 3D grid of features and an array of view-specific 2D grids. We progressively exchange information between the two with a dedicated bidirectional attention mechanism. We exploit knowledge about the image formation process to significantly sparsify the attention weight matrix, making our architecture feasible on current hardware, both in terms of memory and computation. We attach a DETR-style head on top of the 3D feature grid in order to detect the objects in the scene and to predict their 3D pose and 3D shape. Compared to previous methods, our architecture is single stage, end-to-end trainable, and it can reason holistically about a scene from multiple video frames without needing a brittle tracking step. We evaluate our method on the challenging Scan2CAD dataset, where we outperform (1) recent state-of-the-art methods for 3D object pose estimation from RGB videos; and (2) a strong alternative method combining Multi-view Stereo with RGB-D CAD alignment. We plan to release our source code. | ['Vittorio Ferrari', 'Stefan Popov', 'Kevis-Kokitsi Maninis', 'Michał J. Tyszkiewicz'] | 2022-03-24 | null | null | null | null | ['3d-pose-estimation'] | ['computer-vision'] | [ 7.34555721e-02 6.70632273e-02 2.05088988e-01 -4.16592032e-01
-9.13821816e-01 -5.50855696e-01 3.93303394e-01 -2.82666415e-01
-2.70410269e-01 4.18868586e-02 9.57576782e-02 -2.14967623e-01
1.08470425e-01 -5.57969630e-01 -1.31447864e+00 -3.28888059e-01
2.40061179e-01 1.08983612e+00 5.56399584e-01 -3.56292538e-02
2.83560544e-01 7.83582687e-01 -1.60193157e+00 2.40395695e-01
2.35494912e-01 1.53383744e+00 4.23651934e-01 6.92270160e-01
1.89534768e-01 9.06935036e-01 2.60685291e-02 -1.79112062e-01
4.84049559e-01 -1.12841971e-01 -9.23213542e-01 5.48283517e-01
9.99357283e-01 -6.67656004e-01 -4.35985833e-01 7.54220009e-01
4.49478656e-01 -5.44708297e-02 2.87237376e-01 -9.20229912e-01
-4.10735399e-01 2.18926802e-01 -6.88791990e-01 -3.92505899e-02
4.52308685e-01 2.40151912e-01 8.83358419e-01 -1.27633786e+00
8.07709098e-01 1.17783606e+00 6.78839087e-01 4.48065728e-01
-1.15380299e+00 -3.75693351e-01 3.14312309e-01 3.55587274e-01
-1.13247848e+00 -5.66590369e-01 1.09194756e+00 -5.27975082e-01
1.32472420e+00 2.61276457e-02 1.14913404e+00 8.42186630e-01
1.71769522e-02 8.60297859e-01 1.05769575e+00 -2.72473663e-01
1.27503678e-01 -3.41525644e-01 -2.09425818e-02 9.95108426e-01
-1.79787606e-01 -2.40706187e-02 -7.77501762e-01 9.46739763e-02
1.20893073e+00 1.67784348e-01 -3.35995436e-01 -1.06010520e+00
-1.31050158e+00 5.06242037e-01 7.28602111e-01 -1.23585202e-01
-5.73637545e-01 4.96610910e-01 9.64072645e-02 -1.99479125e-02
4.18138564e-01 9.57555249e-02 -7.40938962e-01 2.35097832e-03
-5.78233778e-01 2.53553569e-01 5.61482072e-01 1.05423105e+00
9.46823180e-01 -8.84911790e-02 4.15118873e-01 3.86571854e-01
4.38065469e-01 4.69455242e-01 2.40766242e-01 -1.08963180e+00
4.24792618e-01 6.88137352e-01 1.15147635e-01 -7.18847990e-01
-4.85275328e-01 -1.72012255e-01 -4.03343409e-01 3.90534401e-01
2.46083692e-01 4.08072263e-01 -1.12957513e+00 1.38312376e+00
6.19499326e-01 1.05327383e-01 -4.38377470e-01 1.25654280e+00
7.49097884e-01 3.80403042e-01 -6.35442317e-01 3.40981901e-01
1.32679868e+00 -1.05690694e+00 -1.51064381e-01 -6.03933275e-01
2.44713351e-01 -6.58413947e-01 8.85125101e-01 4.81110066e-01
-1.58478129e+00 -5.90330780e-01 -9.16973829e-01 -6.17182434e-01
-2.03831583e-01 6.73610196e-02 6.10770822e-01 2.39530727e-01
-1.07415199e+00 6.25626922e-01 -1.22157669e+00 -1.34796336e-01
5.99100471e-01 5.84324360e-01 -6.62657797e-01 -2.81746626e-01
-3.94407600e-01 9.55894351e-01 2.83353299e-01 3.81942332e-01
-1.01606202e+00 -7.12381244e-01 -1.06194520e+00 -4.33763750e-02
6.73552811e-01 -1.04878557e+00 1.34963572e+00 -7.06337690e-01
-1.58264661e+00 1.10805726e+00 -9.53068584e-02 -8.51722509e-02
4.79512334e-01 -6.24393821e-01 3.08536440e-01 1.85709611e-01
-7.27022663e-02 7.08564341e-01 8.16504657e-01 -1.43856716e+00
-4.41934049e-01 -7.82132924e-01 1.24263488e-01 3.08487326e-01
1.10601649e-01 -2.15670258e-01 -1.07817841e+00 -2.41048440e-01
6.59024417e-01 -1.07204592e+00 -2.63361365e-01 4.55295026e-01
-3.63401473e-01 -5.74733540e-02 1.02762783e+00 -7.07131684e-01
3.91118437e-01 -1.82267511e+00 5.78981221e-01 6.68493658e-02
3.28670204e-01 -3.67847160e-02 4.23559733e-02 3.92014766e-03
-1.17175169e-02 -3.43268543e-01 -8.74074996e-02 -5.87287366e-01
-2.88477093e-02 2.38245293e-01 -1.67640463e-01 7.24676490e-01
3.85175645e-01 1.00782669e+00 -6.72918916e-01 -2.14814037e-01
5.74673414e-01 7.01913118e-01 -9.16610420e-01 4.26702917e-01
-3.80642027e-01 3.97619903e-01 -3.93992901e-01 8.09314907e-01
6.30448222e-01 -5.69154859e-01 1.07793443e-01 -7.22869873e-01
-7.19530061e-02 4.69845772e-01 -1.26023233e+00 2.34968328e+00
-5.90511799e-01 4.22714114e-01 1.49200797e-01 -8.86396050e-01
6.99547470e-01 9.50992480e-02 5.01779497e-01 -5.09201169e-01
3.01052332e-01 1.10337488e-01 -3.67821097e-01 -3.81413341e-01
3.77898604e-01 1.43445909e-01 -7.91626349e-02 4.72505867e-01
4.08649325e-01 -5.14816642e-01 -2.03001559e-01 2.98870653e-02
1.02675045e+00 7.52437294e-01 1.77485734e-01 -4.55264300e-02
3.50899756e-01 -1.83823675e-01 3.51647526e-01 4.01840836e-01
2.45194107e-01 9.38098133e-01 2.49328196e-01 -9.87599075e-01
-1.21870768e+00 -1.07974005e+00 2.49022871e-01 6.94431782e-01
1.83081150e-01 -3.58833462e-01 -3.22181225e-01 -6.90806687e-01
8.74186456e-02 3.22739959e-01 -6.98939800e-01 1.24806844e-01
-9.57020223e-01 -1.28048912e-01 -1.74851641e-01 9.06299889e-01
3.52647454e-01 -7.25915372e-01 -1.33315551e+00 2.40786448e-01
-2.42469124e-02 -1.34366643e+00 -4.94968086e-01 6.01719022e-01
-9.05819356e-01 -1.33001482e+00 -4.00829375e-01 -5.81382394e-01
6.60296917e-01 2.50833720e-01 1.46551299e+00 4.22584377e-02
-2.12716892e-01 6.35102689e-01 -1.22732744e-01 -2.53687501e-01
-1.39386415e-01 -1.71431765e-01 -8.06802660e-02 -6.86153919e-02
6.52986988e-02 -6.72641039e-01 -7.63745844e-01 4.11671489e-01
-6.53406322e-01 3.45584840e-01 6.65005088e-01 7.60947883e-01
8.49457681e-01 -4.34259832e-01 -2.77359366e-01 -5.60175478e-01
-4.05015767e-01 -2.56213993e-02 -9.50025737e-01 4.89752106e-02
-3.62433270e-02 9.06350240e-02 2.91183233e-01 -2.50053644e-01
-6.59054101e-01 8.14199209e-01 -2.12149084e-01 -1.00997484e+00
-1.23230249e-01 2.14771301e-01 -1.62581414e-01 -3.25549006e-01
7.96084255e-02 1.95644364e-01 -3.30046937e-02 -6.06559157e-01
4.59524572e-01 3.01028162e-01 7.38427639e-01 -4.92305189e-01
7.44793892e-01 6.14419520e-01 1.19433373e-01 -4.95515108e-01
-1.02992201e+00 -3.27397585e-01 -1.11172473e+00 -3.79120052e-01
9.83888745e-01 -1.12348151e+00 -9.89588678e-01 5.94467521e-01
-1.48274732e+00 -5.20641327e-01 -1.41963422e-01 4.14316386e-01
-1.02423751e+00 1.26184180e-01 -6.01958871e-01 -4.49564010e-01
-2.43548036e-01 -1.40662599e+00 1.63882971e+00 -2.06084281e-01
-3.66075374e-02 -5.89948475e-01 -2.15605929e-01 4.96176302e-01
1.87296256e-01 4.05524999e-01 7.15539217e-01 -8.45759362e-02
-1.15394771e+00 -1.63104057e-01 -1.56535327e-01 1.77768439e-01
-2.14066237e-01 -2.52709925e-01 -1.00133073e+00 -3.32006723e-01
2.01892421e-01 -7.66023099e-01 7.42219329e-01 2.87592769e-01
1.44535530e+00 -9.01407003e-02 -2.60022163e-01 8.25749338e-01
1.52481902e+00 -2.17337772e-01 3.22828501e-01 3.66643697e-01
1.03687835e+00 3.23342264e-01 2.80556530e-01 3.79629791e-01
6.96061134e-01 1.00997150e+00 1.05528283e+00 -1.31871670e-01
-1.94517508e-01 -2.38018259e-01 2.07099795e-01 8.32718730e-01
-2.24956036e-01 -5.47670089e-02 -1.11448729e+00 3.61642241e-01
-1.78196442e+00 -6.23728395e-01 8.67768824e-02 2.10234857e+00
4.48608905e-01 2.11396381e-01 5.68313897e-03 5.34286499e-02
2.51213014e-01 1.13921784e-01 -8.89559388e-01 9.94282141e-02
1.50138125e-01 2.45584726e-01 4.89123434e-01 4.25652474e-01
-1.02096438e+00 8.77223492e-01 5.78706074e+00 2.69789249e-01
-1.25428677e+00 -4.76335287e-02 5.45255899e-01 -4.49083745e-01
-1.38888851e-01 -9.32153314e-04 -7.25953817e-01 2.10628901e-02
4.25930113e-01 4.09237087e-01 3.43833357e-01 9.51301157e-01
-3.57056499e-01 -5.94396964e-02 -1.64013958e+00 1.16535735e+00
3.02152276e-01 -1.56141424e+00 2.85491552e-02 9.63059068e-02
5.11857927e-01 5.86858392e-01 -2.46686459e-01 -1.05038777e-01
3.14059943e-01 -8.07512164e-01 1.42376947e+00 5.05892336e-01
7.02541113e-01 -4.90309656e-01 4.66825366e-01 2.96358466e-01
-1.24808085e+00 -1.42671973e-01 -2.62944758e-01 2.57819630e-02
4.64481115e-01 3.58851969e-01 -5.52319527e-01 6.27216160e-01
1.10585356e+00 8.76675725e-01 -6.52554929e-01 7.95946658e-01
-9.89062861e-02 5.65699115e-02 -6.44010365e-01 3.42263252e-01
2.67796814e-01 9.83711779e-02 3.46692294e-01 6.30033195e-01
5.72446346e-01 1.28759414e-01 2.53526539e-01 8.89412522e-01
-7.13146105e-02 -4.41226304e-01 -4.64430511e-01 3.62445533e-01
2.88477898e-01 1.16838467e+00 -7.86481142e-01 -2.85673022e-01
-5.74192464e-01 1.10917139e+00 5.66680908e-01 -1.97910164e-02
-7.39765465e-01 1.07126594e-01 4.73899156e-01 1.83722466e-01
1.06488597e+00 -4.17364717e-01 -1.57857686e-01 -1.21321130e+00
3.66538405e-01 -7.08542168e-01 1.78684890e-01 -1.26392639e+00
-1.06735897e+00 7.67737448e-01 -2.01725557e-01 -1.29301810e+00
-2.76164055e-01 -9.41713810e-01 -1.67302474e-01 6.97369337e-01
-1.50201237e+00 -1.47150147e+00 -5.09060740e-01 6.97573543e-01
5.60001433e-01 2.22554535e-01 7.22981930e-01 1.66140020e-01
-2.21802711e-01 1.84006616e-01 -5.45187652e-01 2.12742202e-02
3.08433115e-01 -1.14449823e+00 7.14029253e-01 6.27474189e-01
4.21728492e-01 2.00075999e-01 4.18863028e-01 -2.99979120e-01
-2.16292334e+00 -8.17263484e-01 5.54962635e-01 -9.30395842e-01
4.67616141e-01 -8.08167756e-01 -6.82993591e-01 9.71486568e-01
1.24539718e-01 4.69277412e-01 2.74719983e-01 9.77801904e-02
-4.80248541e-01 -1.49846613e-01 -8.22972476e-01 2.02112317e-01
1.28697801e+00 -4.73222882e-01 -5.37926137e-01 3.16953152e-01
7.11556792e-01 -1.14325202e+00 -9.53965962e-01 4.23302233e-01
7.21278429e-01 -1.23480833e+00 1.14935839e+00 -3.70835602e-01
5.68770409e-01 -5.02833843e-01 -3.56060058e-01 -1.05736721e+00
-3.24473351e-01 -4.45404500e-01 -4.02060658e-01 6.24829888e-01
6.84927031e-02 -2.60022283e-01 1.02998114e+00 5.64954519e-01
-5.11239350e-01 -1.03151834e+00 -1.08037663e+00 -4.66831625e-01
-2.89858490e-01 -6.85725927e-01 5.53723335e-01 5.36208510e-01
-5.37714779e-01 4.09917295e-01 -3.38032544e-01 4.29889977e-01
6.07716918e-01 5.75946450e-01 1.11240852e+00 -1.20267451e+00
-4.87894565e-01 -3.61708254e-01 -6.36214137e-01 -1.53916740e+00
-8.90882162e-04 -6.68748021e-01 1.20722249e-01 -1.33436430e+00
1.39442563e-01 -2.05790177e-01 -2.05529351e-02 6.50170445e-01
2.12106034e-01 5.34776092e-01 3.66420388e-01 9.58808325e-03
-7.30900407e-01 6.58081949e-01 1.10272098e+00 3.61808240e-02
6.23423159e-02 -3.75410795e-01 -3.47656846e-01 9.14457202e-01
2.69663125e-01 -3.71062875e-01 -2.11121868e-02 -1.02892590e+00
3.11863869e-01 2.85305262e-01 9.09963608e-01 -1.08477175e+00
2.78582484e-01 2.12908313e-01 7.24850237e-01 -1.19578457e+00
7.81129777e-01 -1.14929295e+00 1.89247608e-01 2.93401092e-01
-1.63933057e-02 3.16971004e-01 1.78172320e-01 5.20199656e-01
2.56545171e-02 2.74979472e-01 6.76715314e-01 -4.01656419e-01
-8.12943339e-01 6.97293222e-01 5.06853797e-02 -2.97376841e-01
8.57848346e-01 -3.71942997e-01 4.55965735e-02 -2.38148481e-01
-7.19338715e-01 1.81109071e-01 8.55330825e-01 4.52178031e-01
8.26468527e-01 -1.33811927e+00 -4.04414654e-01 6.13772035e-01
1.32043064e-02 6.67898715e-01 3.29684615e-01 7.60778368e-01
-5.71257055e-01 3.10625345e-01 -3.79882127e-01 -1.13409066e+00
-1.18429482e+00 5.55733085e-01 4.59447145e-01 -9.13290456e-02
-8.60844791e-01 1.00283527e+00 2.28890181e-01 -5.39599299e-01
2.49507219e-01 -4.47831720e-01 3.16126972e-01 -2.58350223e-01
3.92156363e-01 5.91863021e-02 3.21076691e-01 -6.71831608e-01
-5.07484138e-01 1.22952354e+00 -8.31329301e-02 5.40462183e-03
1.83066380e+00 -1.82755902e-01 -1.49074122e-01 4.75907832e-01
1.24444628e+00 -3.08833241e-01 -1.79469430e+00 -2.84817964e-01
-2.95007199e-01 -6.76724613e-01 3.62519324e-01 -6.45311654e-01
-1.33170104e+00 8.53421986e-01 5.10197163e-01 -9.62144360e-02
1.17348719e+00 4.22502577e-01 7.20981836e-01 4.96445537e-01
5.46538711e-01 -7.28218675e-01 4.09432888e-01 5.98827243e-01
1.08861804e+00 -1.27549613e+00 2.05535173e-01 -3.44982028e-01
-3.07549745e-01 1.36225533e+00 7.48698235e-01 -3.77438843e-01
5.32252312e-01 3.52176398e-01 -2.19841544e-02 -6.13870680e-01
-8.93301308e-01 -5.37025556e-02 4.05223101e-01 4.73000586e-01
1.99984983e-01 -4.60832953e-01 6.66876435e-01 2.25872159e-01
-1.64830595e-01 1.04272952e-02 3.45081419e-01 1.02750969e+00
-1.31707340e-01 -8.40766430e-01 -2.47454569e-01 2.33007908e-01
-4.02906239e-02 1.58670843e-01 -3.38057786e-01 8.59705627e-01
7.51740262e-02 2.46392936e-01 4.33122486e-01 -3.56868833e-01
5.88430941e-01 1.04439706e-02 9.86231506e-01 -5.00581026e-01
-3.79735172e-01 2.17031702e-01 -1.48924917e-01 -1.28063881e+00
-6.67733669e-01 -7.94819891e-01 -9.17287409e-01 -4.18567248e-02
-4.20232564e-01 -5.24431169e-01 7.15921819e-01 9.23254430e-01
4.65636164e-01 3.89792144e-01 5.08334100e-01 -1.82343721e+00
-3.48915815e-01 -6.28901899e-01 -2.08276674e-01 2.57334977e-01
4.23348516e-01 -8.17926943e-01 8.50659311e-02 7.81488568e-02] | [7.62324857711792, -2.569611072540283] |
3d3183d0-649c-4112-b40e-9cca933d3fac | priberam-compressive-summarization-corpus-a | null | null | https://aclanthology.org/L14-1193 | https://aclanthology.org/L14-1193.pdf | Priberam Compressive Summarization Corpus: A New Multi-Document Summarization Corpus for European Portuguese | In this paper, we introduce the Priberam Compressive Summarization Corpus, a new multi-document summarization corpus for European Portuguese. The corpus follows the format of the summarization corpora for English in recent DUC and TAC conferences. It contains 80 manually chosen topics referring to events occurred between 2010 and 2013. Each topic contains 10 news stories from major Portuguese newspapers, radio and TV stations, along with two human generated summaries up to 100 words. Apart from the language, one important difference from the DUC/TAC setup is that the human summaries in our corpus are {\textbackslash}emph{compressive}: the annotators performed only sentence and word deletion operations, as opposed to generating summaries from scratch. We use this corpus to train and evaluate learning-based extractive and compressive summarization systems, providing an empirical comparison between these two approaches. The corpus is made freely available in order to facilitate research on automatic summarization. | ["Cl{\\'a}udia Pinto", "Andr{\\'e} F. T. Martins", 'Pedro Mendes', 'Miguel B. Almeida', 'Mariana S. C. Almeida', 'Helena Figueira'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['sentence-compression'] | ['natural-language-processing'] | [ 6.34973824e-01 5.43844938e-01 -2.06039548e-01 -4.76290882e-02
-1.46140528e+00 -8.07797730e-01 8.27207625e-01 7.14842796e-01
-5.31946182e-01 1.28995430e+00 1.29213655e+00 6.75044805e-02
-5.62630296e-02 -2.40910873e-01 -4.46793377e-01 -4.44411516e-01
2.11612694e-02 5.70277572e-01 6.76793158e-02 -2.49201059e-01
7.91334391e-01 2.79663992e-03 -1.14322841e+00 7.54828990e-01
1.16045535e+00 2.96422005e-01 4.05379683e-01 1.20501494e+00
8.91332850e-02 9.50915456e-01 -1.39494574e+00 -6.46589339e-01
-9.64510888e-02 -9.28656220e-01 -1.01896560e+00 2.59841412e-01
5.05663097e-01 -2.71607190e-01 -2.34040231e-01 9.26634848e-01
8.60626698e-01 2.37895906e-01 7.78113723e-01 -6.32472277e-01
-4.13695276e-01 1.24459302e+00 -6.09362423e-01 5.03134429e-01
8.29820573e-01 -3.98653477e-01 1.28277469e+00 -6.08420491e-01
9.97176945e-01 9.91596222e-01 4.80205029e-01 4.96331960e-01
-8.62328589e-01 -1.24758007e-02 -2.42294833e-01 1.06110811e-01
-9.10200953e-01 -8.03504825e-01 7.29781032e-01 -8.91162753e-02
1.27618814e+00 8.05015624e-01 6.40771270e-01 1.24717736e+00
3.00259262e-01 1.22877800e+00 4.81821626e-01 -8.04056883e-01
2.85766810e-01 1.46850580e-02 3.72026145e-01 8.20491090e-02
4.88566756e-01 -7.61843145e-01 -5.06342411e-01 -3.93089503e-01
2.92897243e-02 -7.01384366e-01 -5.24051487e-01 3.97050321e-01
-1.52770627e+00 8.05875719e-01 -4.36110467e-01 7.69906521e-01
-7.97904849e-01 -1.53150484e-01 9.81343210e-01 2.35103101e-01
5.93439758e-01 6.62483096e-01 -2.01501891e-01 -4.38430429e-01
-1.54018664e+00 4.99755621e-01 1.23510981e+00 1.18372798e+00
2.77984470e-01 2.44708836e-01 -5.34765244e-01 1.02234900e+00
-2.81000644e-01 4.72910404e-01 7.66316891e-01 -1.22194719e+00
1.07539475e+00 -3.43952887e-02 1.76455036e-01 -1.02602458e+00
-3.10909748e-01 -3.14009041e-01 -9.01055694e-01 -7.42081404e-01
-1.11993790e-01 -6.08544886e-01 -4.74905699e-01 1.09492552e+00
-2.34097913e-01 -3.36495429e-01 4.14528906e-01 3.61745536e-01
1.20361817e+00 1.10885298e+00 -4.00984794e-01 -1.00245655e+00
1.42120993e+00 -1.21608138e+00 -1.11719358e+00 -2.41958544e-01
4.63191390e-01 -1.07167041e+00 6.88684702e-01 5.96793771e-01
-1.59804058e+00 -2.94319063e-01 -1.20400858e+00 -2.48572662e-01
4.06317617e-04 4.47451234e-01 1.13659106e-01 4.60235924e-01
-1.01864254e+00 4.23381269e-01 -6.03490651e-01 -6.18040562e-01
2.34972835e-01 -1.42008707e-01 -1.45530775e-01 1.63804233e-01
-1.02672327e+00 9.69864786e-01 7.98931003e-01 -2.53890842e-01
-4.52141374e-01 -3.82385820e-01 -8.44777167e-01 1.87120959e-01
4.47784513e-01 -6.05352938e-01 1.77409804e+00 -8.26505244e-01
-1.48617089e+00 8.90592158e-01 -2.91259766e-01 -9.38476384e-01
5.08360863e-01 -5.56443334e-01 -3.26005608e-01 6.75604820e-01
4.09084380e-01 3.40198159e-01 6.35997236e-01 -1.06617570e+00
-7.66326249e-01 9.37976763e-02 -1.71636313e-01 3.87671709e-01
-1.52753577e-01 4.88150746e-01 -3.92506421e-01 -1.17152059e+00
-3.35634112e-01 -6.28912508e-01 -2.08923444e-01 -1.16784847e+00
-1.04064512e+00 -1.44547045e-01 6.33601904e-01 -1.12954986e+00
1.68524861e+00 -1.78428805e+00 3.57322842e-01 -2.48291418e-01
3.54940916e-04 2.77243108e-01 -2.09000617e-01 1.09660232e+00
1.58826292e-01 2.81201005e-01 -5.26966989e-01 -4.44730639e-01
2.40102317e-02 -1.06864132e-03 -6.01065040e-01 2.31611982e-01
4.55416925e-02 7.75564849e-01 -9.87628102e-01 -7.42901504e-01
-8.15730914e-02 5.89140542e-02 -4.02996540e-01 -8.80617797e-02
-3.25251341e-01 3.17016393e-01 -4.95084643e-01 3.04773152e-01
4.06383723e-01 3.06621462e-01 1.32374853e-01 -4.93693491e-03
-4.08601940e-01 6.14198983e-01 -8.36118162e-01 1.86602545e+00
-1.34307787e-01 9.43783045e-01 -3.92030627e-02 -1.14758956e+00
7.25650370e-01 6.12417459e-01 3.26174289e-01 -3.70380312e-01
2.55508780e-01 2.79502451e-01 -3.07435483e-01 -5.74426830e-01
1.57535160e+00 1.89076718e-02 -5.28853178e-01 5.74537337e-01
1.33482158e-01 -7.86046028e-01 1.24579823e+00 6.34447396e-01
1.03917348e+00 -2.79464871e-01 9.03174639e-01 -3.18832397e-01
6.92134500e-01 3.53960782e-01 5.52243888e-01 8.18697572e-01
2.05864638e-01 1.02871180e+00 7.93599784e-01 2.04703078e-01
-1.19275403e+00 -7.89231181e-01 1.04393855e-01 9.06092346e-01
-3.66629183e-01 -9.52537358e-01 -9.92918611e-01 -3.90485287e-01
-5.51728964e-01 1.40447211e+00 -4.03329521e-01 1.18515432e-01
-9.58337545e-01 -5.94353914e-01 8.04922581e-01 2.54922479e-01
3.53134066e-01 -1.28319800e+00 -6.71603620e-01 3.47359508e-01
-8.24493647e-01 -1.01767504e+00 -7.71110594e-01 4.18678559e-02
-7.52509892e-01 -8.56884181e-01 -9.85787749e-01 -8.34376156e-01
2.32409507e-01 9.49233249e-02 1.22324729e+00 -2.77546018e-01
7.61389881e-02 4.85533446e-01 -8.25252473e-01 -8.46182823e-01
-1.11647296e+00 5.43130636e-01 -1.50150508e-01 -4.48108166e-01
2.37369258e-02 -3.76154661e-01 1.26940021e-02 -4.61510122e-01
-1.15598834e+00 1.66236952e-01 5.93255341e-01 7.22873688e-01
4.94784653e-01 -1.44421935e-01 8.41133714e-01 -1.12065232e+00
1.33443689e+00 -3.25273395e-01 -1.13369785e-01 2.87519276e-01
6.35841042e-02 -1.80586845e-01 6.87952578e-01 -4.15599421e-02
-1.20759737e+00 -3.75452787e-01 -3.06397587e-01 4.06187683e-01
-8.70416611e-02 9.24345553e-01 2.10175924e-02 7.88186550e-01
7.62225270e-01 4.64655787e-01 -4.32488650e-01 -3.84549171e-01
3.62998843e-01 1.01343155e+00 1.04819822e+00 -3.84861380e-01
4.03311521e-01 2.02146679e-01 -6.25568926e-01 -1.54621696e+00
-1.03970742e+00 -5.17948747e-01 -6.80574536e-01 -1.25624195e-01
8.78885090e-01 -7.98594892e-01 -2.36558989e-02 1.97378337e-01
-1.54398823e+00 6.35121539e-02 -7.97919929e-01 5.18197715e-01
-8.25253308e-01 7.05039144e-01 -8.34612191e-01 -5.40065706e-01
-9.09297228e-01 -6.00998759e-01 1.11465001e+00 3.30705404e-01
-7.67824709e-01 -9.49046552e-01 3.34782422e-01 3.15859854e-01
2.22350180e-01 4.29450870e-01 5.78348339e-01 -1.07005751e+00
2.54668683e-01 -3.50640923e-01 3.21717888e-01 4.51549381e-01
4.73852567e-02 1.09524287e-01 -7.32519388e-01 -3.41073334e-01
4.07068193e-01 -5.13337433e-01 1.08919513e+00 5.80396771e-01
5.62925816e-01 -7.60078251e-01 -1.06129497e-01 -1.72520071e-01
1.02132273e+00 2.19246335e-02 7.67821133e-01 2.65179515e-01
3.47667634e-01 6.32613301e-01 6.32463992e-01 8.08205724e-01
3.92362982e-01 2.66968697e-01 -1.83436543e-01 3.81598622e-01
-1.15659885e-01 -5.77477366e-02 5.33163667e-01 1.73642385e+00
-3.40627506e-02 -7.77498126e-01 -6.98201001e-01 8.69261742e-01
-1.70502651e+00 -1.44516098e+00 -3.03508401e-01 1.87539566e+00
1.01473808e+00 5.08157238e-02 1.00773565e-01 2.77998030e-01
7.42768288e-01 6.91872239e-01 -1.34085666e-03 -7.50891566e-01
-6.59514844e-01 9.98781174e-02 1.29621401e-01 5.72937310e-01
-1.05352640e+00 6.11602962e-01 6.10174704e+00 8.84861052e-01
-7.19157040e-01 3.64437625e-02 3.56413454e-01 -3.30349624e-01
-3.63448292e-01 -1.29193082e-01 -6.08993471e-01 3.21462065e-01
1.19699931e+00 -8.61257493e-01 -7.49531984e-02 3.76231641e-01
4.32175398e-01 -5.43427706e-01 -9.08670306e-01 5.64964235e-01
6.61896825e-01 -1.60846639e+00 1.75221860e-01 -3.14501166e-01
1.17541921e+00 1.09436557e-01 -4.31432784e-01 2.26294339e-01
1.44503981e-01 -5.53235888e-01 1.03585255e+00 4.63734269e-01
6.16723418e-01 -8.14920723e-01 8.47781062e-01 6.32025719e-01
-8.02917302e-01 1.91402897e-01 -4.06115621e-01 8.54289979e-02
5.62254906e-01 7.05177009e-01 -6.14905775e-01 1.12039638e+00
2.42780164e-01 8.23935688e-01 -5.25718749e-01 9.73730206e-01
-4.30023104e-01 9.18551505e-01 -1.27712354e-01 -2.63999820e-01
2.81215250e-01 -1.80719331e-01 1.21626675e+00 1.85401773e+00
3.67737472e-01 1.44768909e-01 9.07055289e-02 3.02846611e-01
-2.75758415e-01 3.66834611e-01 -4.60438728e-01 -2.76373684e-01
4.10408050e-01 1.13041079e+00 -8.88822854e-01 -8.09139252e-01
-2.11445689e-01 1.15442991e+00 -6.76577911e-02 4.03132856e-01
-6.10008061e-01 -8.90389264e-01 -3.10702503e-01 -2.05449462e-01
5.14355600e-01 -1.43421158e-01 -2.41182342e-01 -1.39838386e+00
1.63023829e-01 -1.08587241e+00 3.63094956e-01 -6.28474176e-01
-8.94154847e-01 6.32546067e-01 3.52853119e-01 -1.17249274e+00
-5.49905181e-01 2.45414242e-01 -1.02133238e+00 3.74477804e-01
-1.09047699e+00 -8.28688860e-01 1.62647769e-01 -1.53941274e-01
1.17428982e+00 -3.10275286e-01 8.02861631e-01 1.37844101e-01
-5.31722963e-01 2.54255980e-01 5.35181820e-01 -1.72197044e-01
7.18556225e-01 -1.27662718e+00 4.85709459e-01 9.51346099e-01
1.57595888e-01 3.07868123e-01 1.30901778e+00 -7.07919896e-01
-1.00572789e+00 -1.00671721e+00 1.62918329e+00 -1.76546693e-01
6.41388655e-01 -9.65413526e-02 -7.28343368e-01 7.94962883e-01
1.21216607e+00 -9.82793391e-01 5.93292296e-01 -4.25907671e-01
3.54658186e-01 3.49566340e-01 -8.66901457e-01 5.69071651e-01
6.70555711e-01 -1.37969404e-01 -1.40297401e+00 6.81887329e-01
6.69183135e-01 -5.92445910e-01 -6.15094364e-01 6.09943643e-02
1.45567685e-01 -7.91297853e-01 4.45501149e-01 -2.39930585e-01
8.97121370e-01 -7.04976544e-02 -6.77436963e-02 -1.71707463e+00
-2.81737968e-02 -1.05208361e+00 -1.40901461e-01 1.63125789e+00
5.31406045e-01 -2.42294192e-01 4.64072078e-01 -2.19696969e-01
-6.73506737e-01 -3.10549915e-01 -7.36078143e-01 -4.57405984e-01
1.90145984e-01 -1.17577404e-01 1.77139789e-01 7.22630262e-01
4.24482048e-01 8.12319636e-01 -4.18411583e-01 -3.93403679e-01
3.50348890e-01 1.42555133e-01 7.50117362e-01 -9.29276049e-01
-2.04275608e-01 -6.61502242e-01 1.69719994e-01 -1.04197931e+00
2.03437224e-01 -8.59632134e-01 1.10585824e-01 -2.02605844e+00
3.95037860e-01 5.11876404e-01 5.35359919e-01 1.26959831e-01
-1.62830561e-01 2.90776361e-02 3.06396246e-01 2.42438868e-01
-9.17274654e-01 7.06477106e-01 1.03993583e+00 -4.37504888e-01
-2.71776319e-01 1.23582661e-01 -9.47478592e-01 8.50010276e-01
1.02348685e+00 -5.14474154e-01 -2.39307269e-01 -4.21212852e-01
7.61491060e-02 3.30064803e-01 -2.80128956e-01 -8.28307807e-01
3.99400592e-01 1.55031517e-01 -1.36020944e-01 -1.19873500e+00
2.41198912e-02 -8.96404386e-02 -4.69645076e-02 4.98678952e-01
-6.15593076e-01 2.49776408e-01 3.47212732e-01 3.29571903e-01
-5.70606351e-01 -7.86587000e-01 5.06424308e-01 -2.33932167e-01
-1.64065987e-01 -4.82206672e-01 -1.01500404e+00 6.00244582e-01
9.05028164e-01 -1.90170780e-01 -4.77908254e-01 -7.57629812e-01
-3.91685843e-01 2.61978418e-01 2.49594614e-01 2.06582144e-01
4.76573169e-01 -9.32582557e-01 -1.54609752e+00 -3.37245077e-01
-1.69627234e-01 3.29491794e-02 2.22334325e-01 9.27963793e-01
-8.73925447e-01 7.23291159e-01 -9.37862918e-02 -1.62992433e-01
-1.29883909e+00 2.49650463e-01 -2.94837594e-01 -6.20819151e-01
-7.20323861e-01 3.74581069e-01 -3.59785467e-01 -1.28094271e-01
1.46975860e-01 -3.56990099e-01 -5.98778009e-01 5.90100944e-01
9.40052927e-01 8.00575435e-01 1.12183936e-01 -7.22846925e-01
5.34388125e-02 8.73297453e-02 -2.44744882e-01 -4.68667090e-01
1.46386552e+00 -2.94359952e-01 -4.31106567e-01 7.20582604e-01
1.09247863e+00 6.79738462e-01 -3.69199157e-01 -1.10994810e-02
4.62223202e-01 1.22108594e-01 -3.30433697e-01 -5.76190054e-01
-4.41119969e-01 3.75144541e-01 -5.44379354e-01 6.02169812e-01
9.96429026e-01 -4.65821773e-02 9.65379357e-01 6.96860850e-01
-9.14783403e-02 -1.46414423e+00 -1.28840189e-02 8.26045930e-01
1.49022377e+00 -6.71792150e-01 6.24282360e-01 -1.90609112e-01
-1.01308382e+00 1.12147462e+00 -6.24326766e-02 -1.96740210e-01
2.75456235e-02 1.84367180e-01 -2.22985774e-01 -1.13675095e-01
-6.78915501e-01 2.15547472e-01 1.74848706e-01 4.09776807e-01
6.01061881e-01 -6.10596091e-02 -1.15075278e+00 8.47639441e-01
-8.73520970e-01 -2.86548823e-01 1.31430995e+00 1.08452606e+00
-6.78049862e-01 -9.68997180e-01 -4.62269574e-01 7.22477555e-01
-8.70672643e-01 -1.37853011e-01 -7.95504153e-01 6.78062797e-01
-5.15284359e-01 1.24154973e+00 2.08643004e-02 1.75786600e-01
5.36658764e-01 -1.73184752e-01 4.19019818e-01 -9.11428392e-01
-8.88747513e-01 3.53212386e-01 7.98023045e-01 1.63516421e-02
-5.80909669e-01 -1.11335111e+00 -1.25073826e+00 -2.02260301e-01
-3.54253650e-01 6.93922877e-01 4.03883845e-01 8.69093776e-01
4.58833687e-02 8.03610802e-01 4.65116620e-01 -1.21823680e+00
-5.83065927e-01 -1.53631508e+00 -6.43313646e-01 1.71656430e-01
3.61892462e-01 2.28859290e-01 -3.67829949e-01 3.39060128e-01] | [12.46518611907959, 9.526551246643066] |
7e57d063-e072-4169-9f54-ea854b1e0014 | cause-effect-inference-in-location-scale | 2301.12930 | null | https://arxiv.org/abs/2301.12930v2 | https://arxiv.org/pdf/2301.12930v2.pdf | Cause-Effect Inference in Location-Scale Noise Models: Maximum Likelihood vs. Independence Testing | A fundamental problem of causal discovery is cause-effect inference, learning the correct causal direction between two random variables. Significant progress has been made through modelling the effect as a function of its cause and a noise term, which allows us to leverage assumptions about the generating function class. The recently introduced heteroscedastic location-scale noise functional models (LSNMs) combine expressive power with identifiability guarantees. LSNM model selection based on maximizing likelihood achieves state-of-the-art accuracy, when the noise distributions are correctly specified. However, through an extensive empirical evaluation, we demonstrate that the accuracy deteriorates sharply when the form of the noise distribution is misspecified by the user. Our analysis shows that the failure occurs mainly when the conditional variance in the anti-causal direction is smaller than that in the causal direction. As an alternative, we find that causal model selection through residual independence testing is much more robust to noise misspecification and misleading conditional variance. | ['Oliver Schulte', 'Xiangyu Sun'] | 2023-01-26 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 4.15348589e-01 -2.07827300e-01 -4.58421707e-01 -3.52451861e-01
-8.75681520e-01 -6.43244028e-01 6.90433800e-01 1.22610658e-01
-1.98654249e-01 9.41857219e-01 4.57949102e-01 -6.24825358e-01
-5.46828330e-01 -7.88770556e-01 -9.87848938e-01 -6.54426277e-01
-2.09479064e-01 3.35010141e-01 -1.04776137e-01 4.15734321e-01
1.51976228e-01 1.34656563e-01 -1.37134111e+00 -8.60128030e-02
7.86684096e-01 4.05593246e-01 1.17753856e-01 5.58538795e-01
1.60928980e-01 6.44692779e-01 -5.70439339e-01 -2.21728876e-01
1.25339240e-01 -6.47700667e-01 -4.71928716e-01 -4.08011645e-01
2.09828109e-01 -2.93109994e-02 -1.12865046e-01 1.04706812e+00
4.65948105e-01 -8.77180323e-02 8.86951983e-01 -1.13820970e+00
-5.34721017e-01 9.49996471e-01 -5.73993206e-01 8.42927545e-02
-3.84312100e-03 2.67109782e-01 1.12522876e+00 -5.98863006e-01
4.83786732e-01 1.47165799e+00 6.72449768e-01 2.04446360e-01
-1.76643455e+00 -9.59987044e-01 1.96351200e-01 -1.68828577e-01
-1.61774063e+00 -4.04449850e-01 2.82270879e-01 -6.20035589e-01
6.28567696e-01 3.59873652e-01 1.45102084e-01 1.21347511e+00
4.17001277e-01 2.41973132e-01 1.03017187e+00 -4.37032521e-01
5.10145426e-01 -1.38033256e-01 4.70735021e-02 3.05062592e-01
5.45831144e-01 5.38575411e-01 -5.18901348e-01 -7.29215086e-01
9.78268325e-01 -3.08162391e-01 -3.27525258e-01 -1.75893322e-01
-8.59406769e-01 9.78606880e-01 -6.44571856e-02 1.56922281e-01
-2.72709280e-01 7.59007931e-01 4.96823005e-02 1.83735684e-01
4.49328810e-01 4.94119465e-01 -7.10202336e-01 -1.75816253e-01
-8.75455797e-01 3.93169671e-01 5.87380767e-01 5.55598736e-01
5.04345775e-01 -5.28242886e-02 -4.27513570e-01 5.02076566e-01
4.08558965e-01 6.25109792e-01 8.03672150e-02 -8.89804006e-01
2.00752452e-01 2.31735095e-01 4.31640029e-01 -7.41065085e-01
-4.59936261e-01 -6.64479971e-01 -7.22069800e-01 -8.18164088e-03
7.15637863e-01 -5.12832105e-01 -1.04609573e+00 2.24619174e+00
2.65739650e-01 4.65239853e-01 -3.85744095e-01 6.97802842e-01
2.99935013e-01 3.35568637e-01 5.58361530e-01 -4.46169019e-01
1.18332744e+00 1.74334586e-01 -8.00417244e-01 -2.02295050e-01
6.22897863e-01 -6.40456140e-01 9.60905254e-01 2.27777913e-01
-7.87931681e-01 -7.54780397e-02 -6.81653619e-01 2.61237174e-01
2.28253618e-01 -2.12408394e-01 8.21657896e-01 7.88707018e-01
-6.20040953e-01 4.45095003e-01 -8.14037919e-01 4.53437269e-02
2.85010159e-01 3.31862092e-01 -2.34524921e-01 -9.93405655e-02
-1.34390402e+00 6.01364136e-01 2.33842004e-02 -7.97627196e-02
-1.16294324e+00 -1.33138394e+00 -6.11074090e-01 3.33055794e-01
4.76331174e-01 -8.33855450e-01 1.04544163e+00 -7.65645087e-01
-8.58623922e-01 3.24922800e-01 -3.08859169e-01 -2.92688191e-01
4.63904709e-01 5.10429516e-02 -4.12935495e-01 -3.99519622e-01
3.30306321e-01 -1.40740955e-03 7.52655983e-01 -1.18224049e+00
-5.93442798e-01 -4.38227087e-01 -3.21181178e-01 -2.32429102e-01
1.67204976e-01 9.97026339e-02 -1.71048522e-01 -7.39077866e-01
7.23101795e-02 -8.39189827e-01 -2.42064700e-01 -2.77256697e-01
-7.24434972e-01 -1.21252187e-01 3.80153865e-01 -3.71036023e-01
1.55445564e+00 -2.15936017e+00 -1.86285898e-01 5.92041910e-01
1.20673656e-01 -3.72162938e-01 9.15371031e-02 3.95223349e-01
-4.87400800e-01 5.01422644e-01 -2.65315920e-01 1.35260001e-01
-1.53066702e-02 -1.44562323e-03 -1.57247037e-01 6.82175696e-01
2.83818960e-01 6.17362976e-01 -6.76657379e-01 9.47235990e-03
-1.84642807e-01 4.47918594e-01 -4.44400609e-01 -1.81190018e-02
-7.45075718e-02 3.90268981e-01 -3.07939768e-01 1.92330092e-01
6.28739893e-01 -4.37250972e-01 4.75867003e-01 1.31067872e-01
-2.49530777e-01 4.29063350e-01 -1.55902922e+00 1.00867271e+00
-2.62780041e-01 3.90796334e-01 1.13224901e-01 -7.83994079e-01
5.17941117e-01 3.55527043e-01 2.14698255e-01 -3.21621597e-01
-1.02680348e-01 1.90559030e-01 4.62644100e-01 -4.83634204e-01
-2.40226239e-01 -5.22088885e-01 -2.73128003e-02 2.98660040e-01
-2.17077509e-01 2.72363335e-01 5.39109632e-02 -5.26619069e-02
1.48898494e+00 -2.16009736e-01 3.42293471e-01 -5.98218441e-01
-2.56516248e-01 -1.71972230e-01 9.25880730e-01 1.32112861e+00
4.14165199e-01 5.82283556e-01 9.64151323e-01 1.03001609e-01
-7.74001181e-01 -1.34197903e+00 -5.97599268e-01 6.67187572e-01
-3.94226581e-01 -3.86539757e-01 -4.70940620e-01 -5.15305936e-01
3.35858166e-01 1.08068466e+00 -8.74282897e-01 -2.52274126e-01
-2.67259181e-01 -1.27276719e+00 5.70444107e-01 6.62002742e-01
-7.56537467e-02 -3.98536354e-01 -2.94164389e-01 1.22240253e-01
-2.22626086e-02 -5.56964636e-01 -2.60928661e-01 4.39847052e-01
-6.32937729e-01 -1.24248338e+00 -2.80849516e-01 -8.81049782e-02
6.00323617e-01 -1.57640904e-01 1.00798345e+00 -7.45975748e-02
-2.38587350e-01 2.86163539e-01 1.04161337e-01 -5.46015620e-01
-3.16511035e-01 -4.30917442e-01 2.03886256e-02 -6.35818914e-02
3.17215353e-01 -6.45125151e-01 -4.56737339e-01 2.15020880e-01
-8.59437287e-01 -4.62370634e-01 4.29928571e-01 8.70780766e-01
4.42193061e-01 6.33814037e-01 7.21047997e-01 -1.11079037e+00
4.93310869e-01 -7.46565878e-01 -7.64860630e-01 1.43371642e-01
-7.34920681e-01 2.91676223e-01 1.66271433e-01 -5.66238523e-01
-1.36666048e+00 -1.79394975e-01 1.56986371e-01 -9.70204771e-02
-2.62267232e-01 8.36534858e-01 -5.29138148e-01 5.18802881e-01
6.85304463e-01 -4.01972711e-01 -3.05939674e-01 -5.44643998e-01
7.89513439e-02 2.01046333e-01 3.02153677e-01 -6.07232809e-01
6.10305548e-01 4.29486930e-01 3.83463174e-01 -5.56930721e-01
-6.57738984e-01 -6.74372837e-02 -2.10278988e-01 2.61090070e-01
7.36215055e-01 -9.71985996e-01 -7.49513328e-01 1.36110350e-01
-9.31356192e-01 -4.66972351e-01 -3.12154561e-01 8.08483243e-01
-3.03814083e-01 -1.79906413e-01 -2.76417971e-01 -1.13039756e+00
2.77510643e-01 -1.10384262e+00 8.05401802e-01 -1.52196556e-01
-5.21715641e-01 -1.12581146e+00 2.99426913e-02 -2.64414903e-02
1.70682758e-01 8.15866664e-02 1.49378991e+00 -7.34106183e-01
-3.52386981e-01 -3.18087965e-01 -2.03149989e-01 -2.07464918e-01
2.62420446e-01 2.03343093e-01 -8.32478583e-01 1.32191023e-02
-4.00497280e-02 2.41195723e-01 9.04560924e-01 1.16240454e+00
1.01612008e+00 -3.82466763e-01 -4.97648329e-01 3.25791210e-01
1.37806404e+00 2.14431018e-01 5.48540056e-01 -8.10316429e-02
6.21043622e-01 5.87651908e-01 1.60114795e-01 5.21721482e-01
1.94726571e-01 5.25093257e-01 1.78454816e-01 -1.18036270e-01
4.09373902e-02 -5.58200955e-01 3.23555209e-02 -2.35981718e-01
2.94362277e-01 -4.79882419e-01 -9.19870496e-01 4.78666186e-01
-1.81774843e+00 -1.08120644e+00 -6.51613176e-01 2.67594123e+00
1.07513976e+00 2.98861563e-01 1.29581764e-01 -7.07792491e-02
7.46390224e-01 -3.74780595e-01 -6.01966441e-01 1.87616795e-01
-3.72872740e-01 -1.89514353e-03 7.60959744e-01 7.90897191e-01
-7.90381253e-01 3.78718555e-01 7.70825005e+00 9.23520088e-01
-7.81322122e-01 4.06270102e-02 8.17016602e-01 -1.19329214e-01
-8.01983476e-01 3.59246373e-01 -7.63031721e-01 5.56213796e-01
1.02334917e+00 -2.64262371e-02 -1.03433393e-02 2.29350433e-01
9.28694308e-01 -4.37884003e-01 -1.14897406e+00 5.41451752e-01
-5.52069724e-01 -1.12815094e+00 -1.82390988e-01 3.43010336e-01
7.61843503e-01 -4.01205957e-01 1.62037164e-01 -1.83800027e-01
8.55998218e-01 -1.31675124e+00 5.40042639e-01 8.31193388e-01
9.09042597e-01 -9.12989974e-01 7.64086008e-01 3.01368147e-01
-7.37495840e-01 -2.34691486e-01 -2.24474892e-01 -4.05511320e-01
6.82212487e-02 1.32010949e+00 -7.39079297e-01 9.33082178e-02
6.13195598e-01 2.00482741e-01 -2.08727419e-01 1.04549396e+00
-4.17174280e-01 1.38432491e+00 -5.27697682e-01 2.92466760e-01
-3.18985999e-01 9.59538445e-02 6.90792680e-01 1.15241456e+00
3.20538610e-01 1.19736359e-01 -8.56156051e-02 1.29215395e+00
4.29395624e-02 -1.69986159e-01 -5.60039461e-01 -2.19969284e-02
7.46831477e-01 7.19745994e-01 -7.48299778e-01 -6.78537712e-02
-3.55903119e-01 3.35079640e-01 -1.90589484e-02 7.90691257e-01
-5.54045379e-01 -2.28931587e-02 7.16751039e-01 2.86238939e-01
1.87819421e-01 1.31623462e-01 -7.47543037e-01 -8.32791686e-01
-1.52134404e-01 -6.97249353e-01 5.35067141e-01 -5.45913398e-01
-1.46967876e+00 -4.14663076e-01 4.15734708e-01 -5.44647038e-01
-3.10071170e-01 -3.31279486e-01 -5.49837947e-01 1.19827008e+00
-7.64374733e-01 -7.51246691e-01 3.84489864e-01 1.99906543e-01
1.33747324e-01 2.81697482e-01 6.35580063e-01 1.72951877e-01
-5.69885552e-01 5.07038832e-01 1.12035237e-01 -1.12558648e-01
6.80691481e-01 -1.26277339e+00 1.00798100e-01 8.97185266e-01
-1.20288514e-01 9.14650798e-01 1.21066046e+00 -1.28309178e+00
-1.15404654e+00 -7.85913408e-01 9.94300425e-01 -5.09743392e-01
9.23303604e-01 -4.92161602e-01 -9.15738583e-01 8.00781488e-01
-3.43540639e-01 -1.94218770e-01 6.56917870e-01 5.79346836e-01
-4.14054036e-01 5.53230718e-02 -1.06480706e+00 6.12690330e-01
9.98458624e-01 -2.77189851e-01 -3.58385406e-02 1.17532030e-01
6.90208554e-01 9.16193649e-02 -7.09731877e-01 5.97262740e-01
4.80470955e-01 -7.71582305e-01 6.85808718e-01 -7.51276612e-01
3.16104382e-01 -4.13962930e-01 -3.62430036e-01 -1.20312357e+00
-5.59062421e-01 -5.36031306e-01 1.92753330e-01 1.44291556e+00
7.36211300e-01 -3.94731045e-01 4.38860953e-01 1.07391512e+00
3.61751467e-01 -2.76132524e-01 -1.00585341e+00 -6.54499292e-01
1.22128800e-01 -8.43725264e-01 4.72178787e-01 9.14688051e-01
-2.02378288e-01 5.46759725e-01 -4.50674355e-01 4.28401411e-01
7.07600653e-01 -1.41358882e-01 3.79561990e-01 -1.51578474e+00
-5.30517519e-01 -3.20300549e-01 -2.09654793e-01 -6.93400145e-01
1.24735579e-01 -3.74468476e-01 1.32545546e-01 -1.19127643e+00
5.46863377e-01 -4.06766117e-01 -1.72440857e-02 5.90388775e-01
-4.87800151e-01 -2.20550582e-01 -2.69746125e-01 -1.85852051e-01
-5.78438118e-03 2.52525479e-01 8.45548928e-01 5.54046780e-02
-3.66557807e-01 3.34568650e-01 -8.90845418e-01 8.33710492e-01
5.80597341e-01 -8.84281278e-01 -6.51077926e-01 -1.23504423e-01
5.38824737e-01 3.89832854e-01 6.90041006e-01 -3.34825426e-01
2.82398215e-03 -4.95674878e-01 5.09663939e-01 -3.31361443e-01
1.94835793e-02 -7.11706161e-01 7.99816966e-01 2.25398093e-01
-6.06566608e-01 8.28143880e-02 3.13841641e-01 8.19154441e-01
2.36275852e-01 -3.00251484e-01 4.76256877e-01 4.82838787e-02
-8.57420452e-03 -2.06171125e-01 -6.05442524e-01 2.01447502e-01
3.53229821e-01 3.54889929e-01 -3.10320526e-01 -6.42674923e-01
-6.95003986e-01 -4.18025553e-02 1.31165147e-01 1.02100439e-01
3.80316317e-01 -1.14120853e+00 -9.45816219e-01 -2.13229991e-02
-1.79745376e-01 -4.18408632e-01 2.33653724e-01 8.85225415e-01
3.66982430e-01 2.82950759e-01 6.14391685e-01 -3.81581485e-01
-1.10249853e+00 3.84067386e-01 3.21712285e-01 -1.14301130e-01
-2.81624824e-01 8.74254763e-01 6.78772151e-01 -1.72052786e-01
1.24721296e-01 -1.22548476e-01 3.02546024e-01 -4.13221046e-02
6.11392200e-01 6.65692210e-01 -3.67920250e-02 -1.55563474e-01
-4.39128339e-01 7.51967877e-02 3.13205987e-01 -3.53820205e-01
1.12321305e+00 -1.57985136e-01 -1.92415163e-01 7.95553744e-01
1.00146484e+00 2.88820177e-01 -1.22876525e+00 4.19640802e-02
9.70427468e-02 -3.93130809e-01 2.73565680e-01 -1.22454798e+00
-7.36683130e-01 4.57031816e-01 6.04253054e-01 2.24562392e-01
8.96369338e-01 2.70839743e-02 -1.68136060e-01 -1.97706595e-01
9.80650932e-02 -7.61801302e-01 -4.34001744e-01 1.78621337e-01
6.47803247e-01 -1.14945245e+00 1.71341166e-01 -4.77650851e-01
-1.62583649e-01 6.31781936e-01 3.50220531e-01 8.86810571e-02
9.51456964e-01 6.36165917e-01 -1.31311223e-01 -2.86422163e-01
-9.22081888e-01 -6.36924654e-02 4.77367491e-01 6.67598486e-01
6.34203851e-01 3.02911580e-01 -4.25440311e-01 8.61716151e-01
-1.22812815e-01 -2.70706981e-01 4.33474928e-01 4.89582628e-01
-1.84936538e-01 -8.52884412e-01 -6.21671796e-01 7.65920699e-01
-8.45105052e-01 -4.93170857e-01 -2.87075907e-01 9.73400056e-01
3.43617201e-02 1.38490856e+00 2.74846494e-01 2.36825999e-02
3.10113341e-01 5.80372252e-02 1.58428475e-01 -6.37606680e-01
-2.31985182e-01 6.60715103e-01 2.34422788e-01 -3.80742848e-01
-1.46176890e-01 -8.84029984e-01 -1.24324632e+00 -2.97104686e-01
-5.63336015e-01 1.56755418e-01 3.77694190e-01 1.22470140e+00
3.03703398e-01 5.75391531e-01 2.63720155e-01 -9.58716348e-02
-6.01962268e-01 -7.40333676e-01 -8.20033550e-01 1.46240100e-01
4.66171533e-01 -8.11731935e-01 -7.32447147e-01 -7.61653110e-02] | [7.842119216918945, 5.24251651763916] |
08194ab3-17af-496b-a153-0ca2a64e908d | improved-neural-text-attribute-transfer-with | 1711.09395 | null | http://arxiv.org/abs/1711.09395v2 | http://arxiv.org/pdf/1711.09395v2.pdf | Improved Neural Text Attribute Transfer with Non-parallel Data | Text attribute transfer using non-parallel data requires methods that can
perform disentanglement of content and linguistic attributes. In this work, we
propose multiple improvements over the existing approaches that enable the
encoder-decoder framework to cope with the text attribute transfer from
non-parallel data. We perform experiments on the sentiment transfer task using
two datasets. For both datasets, our proposed method outperforms a strong
baseline in two of the three employed evaluation metrics. | ['Cicero Nogueira dos santos', 'Igor Melnyk', 'Inkit Padhi', 'Kahini Wadhawan', 'Abhishek Kumar'] | 2017-11-26 | null | null | null | null | ['text-attribute-transfer'] | ['natural-language-processing'] | [ 4.74642217e-01 1.80194169e-01 -3.09539080e-01 -9.55280066e-01
-1.13258564e+00 -5.13272166e-01 1.15147161e+00 5.13985038e-01
-8.82961750e-01 1.00792766e+00 4.47432578e-01 -6.47916496e-02
3.43674600e-01 -6.07260823e-01 -7.28996515e-01 -2.74023890e-01
1.90965891e-01 1.00981927e+00 -5.91943823e-02 -3.38889360e-01
1.98702097e-01 -6.43630922e-02 -1.52979732e+00 8.46637845e-01
6.64436579e-01 1.29810929e+00 -2.15463549e-01 5.51907599e-01
-2.96772420e-01 7.61007786e-01 -2.62200892e-01 -9.40337598e-01
2.24668205e-01 -2.68157870e-01 -1.03356659e+00 -3.38687539e-01
6.65496886e-01 -1.59198344e-01 7.72094801e-02 7.94342577e-01
6.92402840e-01 -1.08652949e-01 1.10252690e+00 -1.26518631e+00
-7.96058893e-01 6.46801531e-01 -4.73127633e-01 1.30774513e-01
3.50301683e-01 -2.81402558e-01 1.21007800e+00 -1.18313742e+00
4.22607511e-01 1.32936478e+00 6.33058190e-01 5.10794222e-01
-1.35775435e+00 -1.02300680e+00 -2.10405737e-02 2.82892346e-01
-1.23263204e+00 -6.91318929e-01 5.78557074e-01 -3.66757780e-01
1.14204121e+00 -1.12645619e-01 1.98700681e-01 1.41157103e+00
1.81091517e-01 1.01564813e+00 1.59460449e+00 -6.26336873e-01
-1.90499470e-01 6.16526902e-01 2.05886140e-02 1.26201198e-01
-1.13434620e-01 1.34206876e-01 -1.01351237e+00 -5.88378273e-02
-1.14012323e-01 -4.67161536e-01 2.74454415e-01 -3.16548318e-01
-1.42026341e+00 9.78882551e-01 9.27837640e-02 -1.52763864e-02
-3.37226629e-01 3.31591591e-02 1.12327719e+00 8.34646046e-01
1.07522321e+00 3.86709183e-01 -9.74374413e-01 -2.36909181e-01
-7.49045908e-01 4.21097316e-02 8.87602746e-01 1.41559720e+00
7.93010652e-01 -2.02991039e-01 -5.06614387e-01 6.44976676e-01
2.80462146e-01 4.99842793e-01 5.41810095e-01 -2.72704273e-01
7.72369385e-01 5.08358538e-01 -1.58440262e-01 -3.69468480e-01
-2.22639486e-01 -5.14894873e-02 -7.94401765e-01 1.03592880e-01
1.26659542e-01 -3.87507111e-01 -6.08775616e-01 1.68467522e+00
3.17077726e-01 -1.64530158e-01 4.00628686e-01 4.38534886e-01
7.44380236e-01 4.51811761e-01 7.39224792e-01 2.09290639e-01
1.59133565e+00 -1.04849720e+00 -1.06220412e+00 -2.71493256e-01
7.74130166e-01 -9.60248351e-01 1.34278798e+00 2.19841555e-01
-1.04883003e+00 -5.62292933e-01 -1.17106557e+00 -4.07242328e-01
-7.75339365e-01 4.67776179e-01 4.71076280e-01 4.54332858e-01
-9.24323976e-01 3.33891720e-01 -5.19637048e-01 -3.09277534e-01
4.16455895e-01 5.60400903e-01 -6.74701035e-01 2.38477007e-01
-1.57772982e+00 1.15089333e+00 7.80448139e-01 -1.76589221e-01
-4.77343798e-01 -7.09753335e-01 -9.05485511e-01 -2.21831556e-02
-1.33412918e-02 -7.97242165e-01 1.25541091e+00 -1.31974173e+00
-1.75486588e+00 1.32307208e+00 -1.87499151e-02 -3.97249639e-01
6.31009340e-01 -7.54907131e-01 -2.77516305e-01 -3.07337642e-01
6.79199770e-02 8.41227472e-01 6.89547956e-01 -8.14802885e-01
-6.72897518e-01 -5.81188500e-01 -1.48331732e-01 5.42067111e-01
-9.61094022e-01 2.65087247e-01 1.22564264e-01 -8.41167808e-01
-7.31863141e-01 -9.28432822e-01 3.45094264e-01 -1.30688518e-01
-5.11755288e-01 -3.49315464e-01 6.75902188e-01 -4.54186857e-01
7.83187509e-01 -2.23022056e+00 3.59044701e-01 -2.88146883e-01
-2.34070346e-01 2.13111237e-01 -6.05836511e-02 7.26482153e-01
-7.97076449e-02 -2.68899240e-02 -2.86265671e-01 -8.62237811e-01
2.99926728e-01 1.10225916e-01 -2.26247266e-01 3.20227981e-01
3.77999306e-01 8.56817365e-01 -4.29488778e-01 -6.71255052e-01
-1.56771392e-01 6.08601630e-01 -3.99944335e-01 4.98851389e-01
-2.87655871e-02 4.43954200e-01 -4.77277786e-01 2.34535649e-01
5.82659423e-01 1.78743854e-01 -1.28162518e-01 -1.93770334e-01
5.60540408e-02 6.05273724e-01 -7.25593746e-01 2.06003761e+00
-8.65841985e-01 8.05496931e-01 -1.22694895e-01 -8.15737784e-01
9.78070796e-01 7.01621830e-01 2.71431834e-01 -7.67532885e-01
1.52955487e-01 2.53647476e-01 -2.12728173e-01 -4.14526016e-01
5.29826283e-01 -5.55043399e-01 -4.02789265e-01 7.26772010e-01
4.88992423e-01 1.42022520e-01 -1.42511457e-01 4.10603248e-02
6.87919378e-01 3.21440428e-01 3.62201661e-01 -6.37284756e-01
9.79773521e-01 -2.11935714e-01 1.92396954e-01 3.56218040e-01
-1.35184765e-01 3.46555978e-01 3.86607051e-01 -3.52020621e-01
-1.15671837e+00 -8.32470596e-01 -2.95335144e-01 1.81137669e+00
-2.62037635e-01 -3.82970989e-01 -5.83393991e-01 -1.16541851e+00
2.05226839e-01 9.40761507e-01 -9.49761033e-01 -3.13361645e-01
-3.23639721e-01 -7.68770039e-01 7.66374230e-01 8.17461014e-01
4.49175537e-01 -1.05564046e+00 -1.96911260e-01 1.41813338e-01
-3.37338060e-01 -1.47751606e+00 -4.73158062e-01 3.54567170e-01
-5.79461277e-01 -5.79713225e-01 -3.73788118e-01 -7.96542764e-01
2.95766085e-01 -4.29116368e-01 1.31242573e+00 -3.05201083e-01
1.82558149e-01 1.06534354e-01 -4.08315927e-01 -9.58094060e-01
-6.10501766e-01 9.48614836e-01 -1.02824755e-01 2.76571244e-01
9.68009710e-01 -5.26919067e-01 -3.83818388e-01 2.70615276e-02
-7.04403758e-01 2.24291638e-01 9.36252713e-01 6.69839740e-01
3.94218266e-01 -7.46953785e-01 1.00772357e+00 -1.37374532e+00
7.10757554e-01 -6.89397931e-01 -2.48465121e-01 2.62765825e-01
-8.63786340e-01 5.43329537e-01 6.92867100e-01 -3.58279765e-01
-1.23269057e+00 1.25203073e-01 -1.75441787e-01 1.11283183e-01
-3.62224102e-01 4.42008793e-01 -3.70214134e-01 3.15729320e-01
4.09972936e-01 3.67083177e-02 -1.84411302e-01 -5.66932619e-01
5.10969758e-01 1.10306811e+00 1.08804233e-01 -7.12716758e-01
4.62755352e-01 3.44123423e-01 -1.13284290e-01 -1.22046635e-01
-1.02123427e+00 -3.57714206e-01 -1.17261446e+00 6.60983622e-02
1.05609047e+00 -1.04329908e+00 -5.95144629e-01 3.58117014e-01
-1.30273676e+00 -1.79277673e-01 -2.60077924e-01 5.71486413e-01
-7.71800518e-01 -2.64816493e-01 -5.87107956e-01 -3.53223592e-01
-6.92521989e-01 -9.86652911e-01 1.28842366e+00 -3.90762925e-01
-3.90886128e-01 -1.24129629e+00 3.87788057e-01 1.92903712e-01
7.65769422e-01 1.18175782e-01 1.11829925e+00 -1.45496154e+00
1.78842157e-01 -3.98450106e-01 -2.77197927e-01 4.61327255e-01
8.16668794e-02 -3.67984056e-01 -1.37794757e+00 -3.27796847e-01
-5.49991280e-02 -9.67665911e-01 7.95292199e-01 -2.15869382e-01
1.00704420e+00 -4.36505854e-01 -1.42499611e-01 5.94299138e-01
1.24280095e+00 -2.95228511e-01 3.47641230e-01 6.61696613e-01
7.90488064e-01 1.00351119e+00 6.94270849e-01 2.51702756e-01
8.08938205e-01 7.32672274e-01 7.97749460e-02 -1.64177567e-01
-2.06473365e-01 -2.91708052e-01 5.26018023e-01 1.10563076e+00
1.23084158e-01 -2.44481444e-01 -8.30245018e-01 5.34129620e-01
-1.72948349e+00 -4.40770984e-01 -1.66812047e-01 1.84375942e+00
1.25317276e+00 4.47649956e-02 -5.34057505e-02 -5.37645891e-02
4.65867639e-01 -2.75362283e-02 -4.52017069e-01 -8.01205814e-01
-9.90959853e-02 3.09649080e-01 6.34442508e-01 4.67599332e-01
-1.44006145e+00 1.02827883e+00 6.64272881e+00 4.91348088e-01
-8.73183370e-01 3.85074914e-01 4.50950533e-01 -1.62087977e-02
-2.06990153e-01 -1.16906852e-01 -9.43461299e-01 2.74156421e-01
1.33269966e+00 -5.17518222e-01 -1.06542155e-01 6.59887135e-01
-4.29676712e-01 3.77451688e-01 -1.82002342e+00 5.47802031e-01
3.24322939e-01 -6.41460061e-01 3.37841839e-01 -7.44990557e-02
4.24819857e-01 1.70923695e-01 3.65991086e-01 4.88251090e-01
4.21830565e-01 -1.08772814e+00 7.31905162e-01 2.08042488e-01
1.01946414e+00 -7.85696745e-01 9.90779042e-01 1.22301489e-01
-8.96355987e-01 1.58978984e-01 -1.18606031e-01 4.81138267e-02
2.64370302e-03 2.34031662e-01 -8.34015310e-01 6.64798319e-01
5.16358197e-01 9.73416030e-01 -6.02546871e-01 1.80666104e-01
-4.33752090e-01 6.24199450e-01 1.46772251e-01 5.00208698e-02
1.55256703e-01 2.33279616e-02 2.21592426e-01 1.71534884e+00
1.22933038e-01 -2.34100640e-01 -1.85332328e-01 4.52529639e-01
-4.68979150e-01 5.89998543e-01 -6.81094050e-01 2.69283533e-01
3.19273680e-01 1.15901530e+00 -1.64812788e-01 -5.84690630e-01
-7.29755342e-01 1.05850828e+00 6.11999512e-01 7.00848624e-02
-6.10069275e-01 -4.97202456e-01 5.93870819e-01 -1.48785308e-01
2.64307469e-01 -6.69238251e-03 -3.54201168e-01 -1.18941975e+00
-5.12052700e-02 -9.71469641e-01 6.23957098e-01 -6.21231139e-01
-1.59245932e+00 9.08587694e-01 1.16405956e-01 -1.08837748e+00
-5.75862944e-01 -6.31540120e-01 -8.34186673e-02 8.88753116e-01
-2.01318908e+00 -1.72305465e+00 2.24691965e-02 8.39898288e-01
5.85385382e-01 -3.66892219e-01 1.29384565e+00 6.53629780e-01
-2.05556124e-01 1.02852798e+00 8.66123885e-02 2.02806681e-01
1.42534471e+00 -1.52217054e+00 5.22752821e-01 3.49618644e-01
1.54293254e-01 3.75886150e-02 6.93791091e-01 -3.23344231e-01
-1.12621212e+00 -1.08658826e+00 1.30133641e+00 -8.17755699e-01
6.98720157e-01 -8.68817449e-01 -9.34366107e-01 1.10618114e+00
8.65056217e-01 -1.07593440e-01 1.25240600e+00 3.30919385e-01
-5.65482676e-01 -1.59689844e-01 -9.59959567e-01 6.56024739e-02
5.91148436e-01 -7.99449444e-01 -7.15492249e-01 2.09085464e-01
6.57030284e-01 -5.26409149e-02 -1.29728317e+00 3.53778720e-01
6.39547944e-01 -4.35166955e-01 7.75443435e-01 -1.01071751e+00
7.37754583e-01 3.29699129e-01 -3.32823575e-01 -1.58673716e+00
-4.02495544e-03 -4.07339364e-01 -4.55583067e-04 1.66153383e+00
8.12512696e-01 -6.09520376e-01 3.87683272e-01 6.08206570e-01
-3.96459810e-02 -3.30901444e-01 -9.43962038e-01 -6.42035961e-01
6.95264697e-01 -2.21732467e-01 7.51096666e-01 1.04471052e+00
4.16560397e-02 1.11873257e+00 -4.65504467e-01 -3.71567369e-01
5.55401325e-01 -1.48533374e-01 6.77904904e-01 -1.32994080e+00
2.57806573e-03 -2.32038349e-01 -4.55994010e-01 -6.19716167e-01
8.21933687e-01 -1.32155120e+00 -1.08042002e-01 -1.26485956e+00
3.32205921e-01 -2.81408310e-01 -4.69213724e-01 5.44819891e-01
-2.36467600e-01 2.18795389e-01 1.56106278e-02 2.39841402e-01
-5.69984496e-01 1.11323333e+00 9.65699255e-01 -7.11964816e-02
3.63502383e-01 -1.85118943e-01 -6.65625215e-01 4.97594595e-01
9.75628912e-01 -9.44734573e-01 -3.18606913e-01 -9.06803429e-01
5.34582973e-01 -2.85108149e-01 -8.49883333e-02 -6.67768180e-01
2.03813791e-01 1.72221959e-01 1.60439268e-01 -3.05821240e-01
3.16348076e-01 -9.11748827e-01 -5.66640377e-01 1.28502890e-01
-1.02739573e+00 5.30560732e-01 5.74136436e-01 4.12654161e-01
-5.54665625e-01 7.16768652e-02 5.78214467e-01 3.84393364e-01
-2.51409799e-01 1.37466729e-01 -3.98947120e-01 4.50498343e-01
1.00562692e+00 3.33123922e-01 -2.13165566e-01 -2.77037084e-01
-6.68972552e-01 1.81762874e-01 2.32463539e-01 6.72184765e-01
3.19248229e-01 -1.63211417e+00 -1.36464834e+00 2.21512869e-01
5.53521872e-01 -4.28866625e-01 -2.41254613e-01 8.37925971e-01
-1.21845976e-01 5.08738577e-01 -5.90604603e-01 -2.97137618e-01
-1.16569018e+00 6.98816478e-01 7.90543333e-02 -3.64286542e-01
-3.11476231e-01 5.35441339e-01 1.50476307e-01 -8.15463662e-01
1.34807289e-01 8.84189159e-02 -4.17742193e-01 3.21942031e-01
3.92047316e-01 2.44556427e-01 2.91278750e-01 -8.76412392e-01
-3.08328092e-01 4.30440754e-01 -5.05426645e-01 -4.85864609e-01
1.34381294e+00 -3.69611800e-01 -8.21503177e-02 8.61685932e-01
1.67544508e+00 -3.02165866e-01 -9.96393800e-01 -5.12825191e-01
3.10374469e-01 -1.13110617e-01 1.91374533e-02 -8.95533025e-01
-8.59436750e-01 1.33677673e+00 5.97697973e-01 1.10679530e-02
1.15880513e+00 -3.94293070e-02 7.08186865e-01 3.78324747e-01
-6.67424202e-02 -1.14679193e+00 -8.98005590e-02 6.15072429e-01
8.25417697e-01 -1.48080957e+00 -1.75664216e-01 -1.99027836e-01
-1.07729220e+00 9.36325788e-01 8.15593183e-01 -1.96456313e-02
8.27877402e-01 4.83815640e-01 2.98296243e-01 -2.01807648e-01
-1.27630353e+00 -1.32348448e-01 5.26690722e-01 3.96353394e-01
1.13427317e+00 -1.04451589e-01 -4.55665618e-01 4.13512766e-01
-3.23188871e-01 -2.50532515e-02 2.23290175e-01 7.64244318e-01
1.40536964e-01 -1.57141590e+00 1.39788613e-01 2.20579401e-01
-8.70061100e-01 -3.58929038e-01 -6.79276228e-01 9.23553526e-01
-2.38997832e-01 7.01268673e-01 2.49362797e-01 -2.83674091e-01
5.71309984e-01 4.59067106e-01 3.45553786e-01 -6.03355348e-01
-8.96788359e-01 -3.84768546e-01 3.58843237e-01 -3.50957602e-01
-7.64218450e-01 -9.32694316e-01 -9.33286607e-01 -8.22745413e-02
-1.97545037e-01 1.29745394e-01 9.00918186e-01 9.76803899e-01
5.37795484e-01 4.35188651e-01 5.81013978e-01 -4.11444515e-01
-6.42223120e-01 -1.43024802e+00 -1.48004398e-01 7.16338575e-01
3.81959498e-01 -5.67882717e-01 -1.15177229e-01 2.77176082e-01] | [11.294747352600098, 9.546908378601074] |
72e442e4-16ac-4e83-9005-584b5a819dfb | uncertainty-driven-6d-pose-estimation-of | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Brachmann_Uncertainty-Driven_6D_Pose_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Brachmann_Uncertainty-Driven_6D_Pose_CVPR_2016_paper.pdf | Uncertainty-Driven 6D Pose Estimation of Objects and Scenes From a Single RGB Image | In recent years, the task of estimating the 6D pose of object instances and complete scenes, i.e. camera localization, from a single input image has received considerable attention. Consumer RGB-D cameras have made this feasible, even for difficult, texture-less objects and scenes. In this work, we show that a single RGB image is sufficient to achieve visually convincing results. Our key concept is to model and exploit the uncertainty of the system at all stages of the processing pipeline. The uncertainty comes in the form of continuous distributions over 3D object coordinates and discrete distributions over object labels. We give three technical contributions. Firstly, we develop a regularized, auto-context regression framework which iteratively reduces uncertainty in object coordinate and object label predictions. Secondly, we introduce an efficient way to marginalize object coordinate distributions over depth. This is necessary to deal with missing depth information. Thirdly, we utilize the distributions over object labels to detect multiple objects simultaneously with a fixed budget of RANSAC hypotheses. We tested our system for object pose estimation and camera localization on commonly used data sets. We see a major improvement over competing systems. | ['Carsten Rother', 'Michael Ying Yang', 'Frank Michel', 'Eric Brachmann', 'Stefan Gumhold', 'Alexander Krull'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['camera-localization'] | ['computer-vision'] | [ 1.40947074e-01 -2.08853245e-01 1.06249996e-01 -5.61733663e-01
-1.11141896e+00 -1.00524318e+00 3.57278258e-01 -2.56823525e-02
-4.77856904e-01 3.40180755e-01 -1.96379364e-01 9.84084699e-03
2.24329621e-01 -2.00693011e-01 -9.32966650e-01 -6.88098609e-01
3.56877565e-01 7.99421787e-01 4.24119592e-01 5.02825677e-01
4.21303034e-01 9.44960892e-01 -1.59957469e+00 6.91499338e-02
1.48702219e-01 1.18536055e+00 4.02020872e-01 1.00710177e+00
2.50349697e-02 6.58519149e-01 -2.91400552e-01 -1.69698253e-01
3.91517907e-01 -5.63446954e-02 -6.28595173e-01 7.74442136e-01
7.40545809e-01 -5.90115011e-01 -3.85475866e-02 9.16290283e-01
3.26165050e-01 2.79408116e-02 7.78831005e-01 -1.33433807e+00
1.65469872e-04 -1.57371655e-01 -8.34342420e-01 -1.29383251e-01
3.63440305e-01 1.01725116e-01 7.99216688e-01 -1.05404866e+00
5.41661978e-01 1.15966022e+00 7.26778328e-01 2.83269703e-01
-1.48188746e+00 -2.85609037e-01 3.06565702e-01 -1.13294750e-01
-1.65635240e+00 -3.78651857e-01 7.18874097e-01 -6.79491043e-01
8.42188060e-01 2.10348815e-01 4.78874058e-01 7.53439546e-01
1.17539585e-01 7.22460985e-01 1.14090073e+00 -4.87971842e-01
2.92647958e-01 3.85344684e-01 -1.14040012e-02 6.07885301e-01
1.28884032e-01 -1.82135284e-01 -5.50372481e-01 -7.57247955e-02
8.49261820e-01 2.13709608e-01 8.29033088e-04 -9.45466995e-01
-1.26521766e+00 5.27122080e-01 3.64433885e-01 -1.91319823e-01
-1.52966797e-01 3.81326795e-01 -1.30611554e-01 -3.83173436e-01
4.27634388e-01 3.39194238e-01 -5.30188084e-01 -1.63265765e-01
-7.03752100e-01 -6.47305511e-03 6.11130059e-01 1.52384317e+00
8.63085866e-01 -4.51551318e-01 2.66134351e-01 5.71255147e-01
5.45567155e-01 8.87973845e-01 -1.38391480e-01 -1.34903145e+00
2.88239568e-01 3.16809952e-01 4.07550991e-01 -8.93524826e-01
-4.06009078e-01 -9.88077149e-02 -3.37797314e-01 3.89373928e-01
6.28812551e-01 4.85585965e-02 -9.79352534e-01 1.29551625e+00
5.48856556e-01 5.38601307e-04 -3.26353818e-01 1.08072877e+00
2.52542168e-01 4.95262831e-01 -3.55063319e-01 8.88566077e-02
1.23424029e+00 -8.26856494e-01 -2.92632639e-01 -4.97733027e-01
2.70072341e-01 -1.13433266e+00 7.77511477e-01 6.27814353e-01
-1.06406617e+00 -3.77782643e-01 -7.60125041e-01 -3.88898611e-01
-1.72886178e-01 4.73877013e-01 6.22272611e-01 5.12504876e-01
-9.32448626e-01 4.15066570e-01 -1.24041724e+00 -3.69466931e-01
2.26020649e-01 6.75068378e-01 -5.75017214e-01 -1.98055223e-01
-1.15847424e-01 1.02990699e+00 2.60198653e-01 2.61634827e-01
-8.55628133e-01 -3.75552475e-01 -8.94215524e-01 -3.23101491e-01
6.31082058e-01 -4.12550062e-01 1.49672782e+00 -5.85751951e-01
-1.39242566e+00 9.63020921e-01 -4.11896408e-01 -3.72532615e-03
4.30927247e-01 -4.14206296e-01 3.66023242e-01 1.21949159e-01
-1.53674735e-02 8.58315945e-01 9.90357816e-01 -1.46791387e+00
-7.40521073e-01 -6.07718468e-01 -1.53784633e-01 3.71079057e-01
2.75216609e-01 -8.90514925e-02 -1.00786805e+00 -1.51090577e-01
6.75905406e-01 -1.22527742e+00 -3.29345524e-01 4.26363587e-01
-4.64567393e-01 8.08944255e-02 5.71410656e-01 -5.05883694e-01
3.60709131e-01 -2.08092809e+00 3.49957526e-01 2.10443169e-01
7.83695132e-02 -4.41955715e-01 1.94553077e-01 3.72267589e-02
1.30087152e-01 -2.71292478e-01 -7.03459978e-02 -9.61151719e-01
3.63866389e-02 3.35466176e-01 -3.09100658e-01 9.72452343e-01
3.22458386e-01 6.15321398e-01 -5.85781038e-01 -5.22202909e-01
6.45066679e-01 6.96220219e-01 -5.47867596e-01 2.57158309e-01
-3.02213490e-01 6.12034678e-01 -3.25498372e-01 8.93719137e-01
9.47522640e-01 -3.17880839e-01 -1.15521885e-01 -4.41748977e-01
-2.31646314e-01 6.64417595e-02 -1.51465929e+00 1.84896946e+00
-4.51233864e-01 6.53875947e-01 1.91663742e-01 -3.89638931e-01
6.58565402e-01 -1.01657748e-01 6.11564159e-01 -1.22836024e-01
3.16752642e-01 1.27460703e-01 -4.65218812e-01 -1.86935559e-01
5.92034280e-01 -6.47012666e-02 -5.86300865e-02 4.38690126e-01
1.10699460e-01 -7.49295235e-01 -2.40467116e-01 3.07415798e-02
9.01958823e-01 4.51615393e-01 2.62425750e-01 2.18782332e-02
9.34926271e-02 1.80370122e-01 2.48755410e-01 6.36673272e-01
-1.44900391e-02 1.20245886e+00 3.77126634e-01 -3.13037336e-01
-1.22239113e+00 -1.15760708e+00 -1.62886769e-01 6.56686544e-01
4.98313427e-01 -1.27010584e-01 -4.80113566e-01 -6.22632861e-01
2.34982222e-01 5.24836302e-01 -4.49896604e-01 1.82276279e-01
-4.13517088e-01 -3.88927460e-01 -1.88560009e-01 7.63732493e-01
3.46292704e-02 -4.24376488e-01 -8.61488283e-01 -1.32059872e-01
1.03302911e-01 -1.29381216e+00 -4.38274294e-01 6.45920873e-01
-9.19795156e-01 -1.00396121e+00 -5.25237083e-01 -3.70371193e-01
1.13107765e+00 5.49002290e-01 9.69688475e-01 -3.42507899e-01
-6.05172515e-01 8.42749536e-01 -1.38431966e-01 -1.77223697e-01
2.95283813e-02 -1.88985899e-01 1.62970483e-01 -1.35911226e-01
2.75730222e-01 -2.92305779e-02 -5.13732553e-01 6.55376732e-01
-6.38056278e-01 8.71786252e-02 6.58993363e-01 4.80658531e-01
1.06480789e+00 -1.46708230e-03 -4.83020693e-01 -5.38244069e-01
-2.60112911e-01 -9.45364013e-02 -1.13558948e+00 1.98450103e-01
-2.72617638e-01 2.57035881e-01 3.39337736e-02 -4.01627392e-01
-9.82385933e-01 1.05810785e+00 7.24728703e-02 -7.27490723e-01
-3.77965808e-01 -2.26678643e-02 -2.30209649e-01 -3.14374268e-01
4.02883053e-01 -1.61733150e-01 2.10631546e-02 -6.54891074e-01
4.72950965e-01 7.13699758e-01 5.65398514e-01 -6.62349403e-01
6.60068631e-01 7.74548471e-01 2.05398783e-01 -6.68689430e-01
-8.89708638e-01 -9.85903323e-01 -1.07549274e+00 -2.15576887e-01
9.95313168e-01 -1.07035530e+00 -1.03473461e+00 3.84741962e-01
-1.41263437e+00 -2.62877673e-01 -2.86509424e-01 7.31648386e-01
-7.41761208e-01 2.15950087e-01 -3.52150381e-01 -1.01093459e+00
3.72470260e-01 -1.36472261e+00 1.75954711e+00 1.17078073e-01
-1.72169972e-02 -7.38097191e-01 -1.79473221e-01 2.62703866e-01
5.29507734e-02 1.01383716e-01 1.47415593e-01 -1.62792608e-01
-1.04458284e+00 -4.58315641e-01 -4.25880969e-01 3.12966138e-01
1.54278144e-01 1.64619386e-02 -1.17285025e+00 -2.78640002e-01
1.97069973e-01 -2.98594952e-01 6.40361309e-01 4.32598621e-01
1.16402054e+00 1.99034765e-01 -4.59508836e-01 6.38874412e-01
1.44327354e+00 -4.08335254e-02 1.70039460e-01 1.62076786e-01
9.30611968e-01 6.31743193e-01 8.43080938e-01 4.26666081e-01
3.10098112e-01 9.41498756e-01 6.13340378e-01 1.05889469e-01
4.61683460e-02 -1.87965930e-01 1.91322550e-01 4.32468951e-01
8.64784345e-02 -9.91435274e-02 -1.14910924e+00 2.71842152e-01
-1.70529509e+00 -3.24032754e-01 -1.63966089e-01 2.35114813e+00
4.37444985e-01 8.14874545e-02 -2.49268897e-02 -2.39704609e-01
6.88949883e-01 -3.93277913e-01 -7.26414204e-01 2.68924743e-01
8.66962522e-02 -2.30033532e-01 1.10868371e+00 7.56047487e-01
-1.17135537e+00 7.90587842e-01 6.55079412e+00 3.16662401e-01
-8.61346602e-01 -6.31885510e-03 4.20998156e-01 -2.19337568e-01
1.27383582e-02 1.91051647e-01 -1.31395423e+00 2.79088438e-01
4.34067369e-01 3.41084719e-01 3.72007400e-01 1.19409776e+00
-1.86840177e-01 -6.62225902e-01 -1.51278365e+00 1.32325780e+00
4.03895110e-01 -8.18638742e-01 -4.46406811e-01 2.29681537e-01
6.09287500e-01 1.44958496e-01 1.30855083e-01 -2.00405598e-01
3.33873242e-01 -7.40197122e-01 1.07379329e+00 6.03219390e-01
7.08222508e-01 -6.20762944e-01 8.12927783e-01 4.72840160e-01
-9.90761280e-01 6.23574145e-02 -5.65562844e-01 1.17595131e-02
2.92998910e-01 6.09956443e-01 -1.12025785e+00 1.88461170e-01
7.13775516e-01 6.50930464e-01 -6.23886108e-01 1.13639390e+00
-2.30239600e-01 -1.01150544e-02 -8.72719109e-01 1.39014453e-01
-8.06490481e-02 -1.35820881e-01 2.21224248e-01 8.95132005e-01
4.01991367e-01 2.40520269e-01 2.26286292e-01 7.90241897e-01
2.13872522e-01 -4.19887513e-01 -4.43243980e-01 2.47019827e-01
2.61762738e-01 1.31422460e+00 -1.12678146e+00 1.25275552e-01
-3.65001857e-01 1.16594982e+00 2.20311642e-01 1.14467502e-01
-8.13940048e-01 4.24364395e-03 3.95712733e-01 9.61264446e-02
5.51217258e-01 -7.10216463e-01 -2.10696682e-01 -1.21731377e+00
1.57989204e-01 -4.36162800e-01 2.11268701e-02 -1.09088182e+00
-1.20372272e+00 3.66066933e-01 1.74562663e-01 -1.27956617e+00
-3.12082201e-01 -1.13325834e+00 1.40581250e-01 8.85128319e-01
-1.21613193e+00 -1.16546714e+00 -4.94433790e-01 4.09491420e-01
5.74006200e-01 3.33962411e-01 6.53089762e-01 1.11723058e-01
-2.18900278e-01 8.53188261e-02 1.69685155e-01 -9.76494104e-02
7.83294976e-01 -1.43659377e+00 2.24996746e-01 6.20963454e-01
3.92985016e-01 4.65350240e-01 8.11788976e-01 -5.14721811e-01
-1.81030750e+00 -7.55965114e-01 4.60033208e-01 -1.23265493e+00
4.62069899e-01 -7.55928040e-01 -3.78894120e-01 9.42351580e-01
-2.74062902e-01 4.45810795e-01 3.24180901e-01 2.85727158e-03
-2.00845435e-01 -1.66185766e-01 -1.05685294e+00 1.90998912e-01
6.61786616e-01 -6.11842215e-01 -2.28797391e-01 4.33082789e-01
5.54407179e-01 -9.79920626e-01 -6.03614330e-01 2.79228717e-01
6.08504832e-01 -9.82017994e-01 1.01640892e+00 -1.01564132e-01
7.04394802e-02 -8.26756716e-01 -5.51540375e-01 -9.53283250e-01
8.12551454e-02 -2.41030931e-01 1.37175890e-02 1.08706355e+00
2.73881674e-01 -1.85390130e-01 1.02421451e+00 1.17169094e+00
-1.07234553e-03 -3.48385841e-01 -9.90506530e-01 -6.28976345e-01
-4.28433478e-01 -7.68992066e-01 1.34553403e-01 3.38598669e-01
-4.74017233e-01 2.01612994e-01 -3.07223558e-01 5.09023607e-01
9.26387310e-01 3.07953835e-01 1.02019143e+00 -1.00588214e+00
-5.05378842e-01 -7.41676241e-02 -6.68343306e-01 -1.42649138e+00
5.21583445e-02 -3.86363000e-01 5.34016490e-01 -1.30001593e+00
3.62094223e-01 -6.09806418e-01 1.93014756e-01 2.60523707e-01
1.10575236e-01 4.52209979e-01 2.01361611e-01 2.68137962e-01
-8.85296464e-01 1.89639002e-01 8.14573169e-01 3.44751738e-02
1.30195664e-02 2.57009119e-01 -2.36384466e-01 9.21653926e-01
5.57832956e-01 -5.28027058e-01 -1.81340620e-01 -7.67773747e-01
3.22916627e-01 -4.18731607e-02 6.68503225e-01 -8.83302391e-01
5.56355178e-01 4.07685805e-03 8.04254591e-01 -1.17823482e+00
7.51435399e-01 -1.44672632e+00 2.40340278e-01 1.85843930e-02
-1.78882048e-01 -5.52668311e-02 2.40791708e-01 5.66422701e-01
3.16617712e-02 -2.86975592e-01 8.15446138e-01 -1.51574999e-01
-7.03235447e-01 1.33923307e-01 -6.35654852e-02 -3.58266711e-01
1.22620356e+00 -2.53443658e-01 1.20069399e-01 -2.41065741e-01
-9.14490402e-01 1.58708468e-01 8.65509570e-01 3.18193018e-01
6.02102518e-01 -1.16987729e+00 -2.28774771e-01 3.87541622e-01
2.19885796e-01 6.25460982e-01 5.79804815e-02 7.67545402e-01
-7.42769241e-01 4.08642173e-01 8.69693235e-02 -1.31141782e+00
-1.40345085e+00 6.35720730e-01 2.12821200e-01 3.31854761e-01
-3.15973043e-01 1.05939174e+00 1.23359740e-01 -4.05081630e-01
5.22473335e-01 -5.96946895e-01 4.67163891e-01 -1.39886320e-01
3.64027768e-01 3.73183608e-01 8.07615668e-02 -7.67992496e-01
-5.43289781e-01 1.04814732e+00 -1.29184023e-01 -3.51607949e-01
1.24534929e+00 -5.36402524e-01 -1.03672430e-01 8.27710330e-01
1.22039175e+00 2.90540922e-02 -1.91950870e+00 -2.12354839e-01
-7.17809871e-02 -8.77024531e-01 2.20154732e-01 -5.52955925e-01
-8.15066040e-01 1.11370766e+00 6.75563574e-01 -1.27701163e-01
7.44448483e-01 5.12018561e-01 1.04727894e-01 3.65711778e-01
6.90205395e-01 -9.39096749e-01 7.16795623e-02 3.92777532e-01
6.52868748e-01 -1.60211873e+00 3.08168590e-01 -5.28214216e-01
-6.40999854e-01 1.19933963e+00 5.77877700e-01 -1.37902070e-02
5.06605208e-01 5.43680966e-01 -6.16641678e-02 -5.31896576e-02
-3.91022533e-01 -1.54293537e-01 3.59325707e-01 4.15833473e-01
1.76350057e-01 -1.72067687e-01 7.84578621e-01 1.94646001e-01
1.88192800e-02 -2.94000685e-01 3.08514893e-01 1.00596797e+00
-4.93096322e-01 -9.56154943e-01 -7.64766753e-01 6.98259845e-02
-1.97219431e-01 1.55604735e-01 -4.32981521e-01 6.84074104e-01
1.16668396e-01 7.57062674e-01 2.87070692e-01 -1.72129199e-01
3.19075853e-01 -1.57012597e-01 8.05544972e-01 -8.12116146e-01
5.09811826e-02 3.91865134e-01 -3.48957717e-01 -8.61279368e-01
-4.79974508e-01 -9.78089094e-01 -1.06310666e+00 -4.52764274e-04
-7.38010347e-01 -1.27690166e-01 1.22435308e+00 8.19940090e-01
1.88468426e-01 1.01739377e-01 5.94599068e-01 -1.45311344e+00
-5.13573706e-01 -6.31242573e-01 -7.41971791e-01 -8.47800672e-02
5.98909795e-01 -7.59173751e-01 -6.38760984e-01 3.22491974e-01] | [7.601783752441406, -2.554365634918213] |
f79eba9e-1805-40f7-b1a7-6db5ac07b67d | eeg-cortical-source-feature-based-hand | 2304.06321 | null | https://arxiv.org/abs/2304.06321v1 | https://arxiv.org/pdf/2304.06321v1.pdf | EEG Cortical Source Feature based Hand Kinematics Decoding using Residual CNN-LSTM Neural Network | Motor kinematics decoding (MKD) using brain signal is essential to develop Brain-computer interface (BCI) system for rehabilitation or prosthesis devices. Surface electroencephalogram (EEG) signal has been widely utilized for MKD. However, kinematic decoding from cortical sources is sparsely explored. In this work, the feasibility of hand kinematics decoding using EEG cortical source signals has been explored for grasp and lift task. In particular, pre-movement EEG segment is utilized. A residual convolutional neural network (CNN) - long short-term memory (LSTM) based kinematics decoding model is proposed that utilizes motor neural information present in pre-movement brain activity. Various EEG windows at 50 ms prior to movement onset, are utilized for hand kinematics decoding. Correlation value (CV) between actual and predicted hand kinematics is utilized as performance metric for source and sensor domain. The performance of the proposed deep learning model is compared in sensor and source domain. The results demonstrate the viability of hand kinematics decoding using pre-movement EEG cortical source data. | ['Lalan Kumar', 'Anant Jain'] | 2023-04-13 | null | null | null | null | ['eeg', 'eeg'] | ['methodology', 'time-series'] | [ 9.36344191e-02 -3.09895664e-01 -1.02668904e-01 -2.21598614e-02
-6.01834178e-01 -3.47528867e-02 3.39159936e-01 -5.13234735e-01
-5.50531447e-01 1.01593471e+00 4.76672679e-01 -3.14849168e-02
-3.01107198e-01 -3.46761525e-01 -7.83953190e-01 -7.47011960e-01
-4.06459033e-01 -2.39736274e-01 -1.83576196e-01 -7.01727644e-02
4.90532666e-01 4.32651818e-01 -1.21184826e+00 5.00969708e-01
5.58682442e-01 1.09613276e+00 1.02474320e+00 6.59761965e-01
2.90906169e-02 5.51035881e-01 -6.79849982e-01 3.78315538e-01
-8.33512023e-02 -4.55461919e-01 -8.22383940e-01 -6.08301699e-01
-3.95450622e-01 -3.94262642e-01 -6.18520319e-01 9.38992202e-01
9.16050732e-01 -9.67373773e-02 7.72791266e-01 -1.06883788e+00
-5.85020304e-01 4.91250813e-01 -3.77247304e-01 6.62838519e-01
3.69238317e-01 2.62984544e-01 2.32884049e-01 -8.02277982e-01
4.30110306e-01 5.40929854e-01 5.82551777e-01 5.11307240e-01
-6.58914208e-01 -9.13962066e-01 -1.05610006e-01 8.53533864e-01
-1.31374514e+00 -1.80229675e-02 9.19262171e-01 -4.96640831e-01
1.67799246e+00 -1.59165442e-01 1.00383759e+00 1.60084045e+00
1.37492490e+00 9.51886296e-01 1.04238129e+00 -1.87611952e-01
1.47279382e-01 -3.52827400e-01 3.77402723e-01 1.06133871e-01
-6.36313036e-02 1.88838840e-01 -1.14602375e+00 2.49379590e-01
9.91236567e-01 -1.93620473e-01 -6.75216913e-01 5.29815316e-01
-1.48484290e+00 1.17084578e-01 5.83615601e-01 4.48661715e-01
-1.18645442e+00 6.15463555e-01 4.87746954e-01 5.40163875e-01
2.36384079e-01 2.27047935e-01 -5.63720167e-01 -9.24593508e-01
-1.08745944e+00 1.05496712e-01 5.15363336e-01 9.78841424e-01
1.11843469e-02 3.95613045e-01 -2.29487851e-01 6.43430233e-01
2.98427165e-01 5.61985552e-01 1.04336965e+00 -3.72183710e-01
6.66324258e-01 3.53816867e-01 -1.78892255e-01 -7.28170335e-01
-6.97442472e-01 -2.20888585e-01 -6.14618659e-01 2.87309051e-01
1.49541110e-01 -2.77070969e-01 -8.76750350e-01 1.52540421e+00
-2.70189941e-01 5.01088917e-01 6.55719787e-02 1.45662212e+00
7.96506047e-01 4.64155078e-01 1.85939118e-01 -6.75390363e-02
1.32935870e+00 -2.57901579e-01 -7.96945632e-01 -2.37945274e-01
2.93949604e-01 -4.18104142e-01 8.60747814e-01 5.66548824e-01
-9.41832721e-01 -4.37264055e-01 -1.21815014e+00 1.33638307e-01
-1.46581665e-01 4.67684507e-01 6.56126440e-01 2.55677283e-01
-9.35279369e-01 8.10771585e-01 -1.48547840e+00 -2.28403673e-01
3.37786734e-01 9.46939528e-01 -6.72312438e-01 5.09430885e-01
-1.19403553e+00 1.22060168e+00 3.46524477e-01 4.58721071e-01
-9.28635240e-01 -5.33530235e-01 -3.44985336e-01 -2.34941542e-01
-4.14745867e-01 -4.25255686e-01 9.05430019e-01 -6.28716588e-01
-1.97594261e+00 3.89736176e-01 -1.89618200e-01 -2.56141394e-01
-6.90161949e-03 -3.98247778e-01 -2.34244034e-01 5.03956862e-02
-1.46413624e-01 5.53103149e-01 1.02069533e+00 -4.58069146e-01
-3.31032097e-01 -7.36292839e-01 -5.81245363e-01 3.28230470e-01
-3.44206423e-01 1.38966888e-01 6.65704235e-02 -8.41743469e-01
2.26794034e-01 -1.00589490e+00 6.58606112e-01 -4.88161206e-01
-3.08211654e-01 -3.11780930e-01 4.77630079e-01 -1.33077669e+00
1.03427911e+00 -1.76113498e+00 6.44403219e-01 3.29493284e-01
-4.39696126e-02 -8.78857970e-02 -1.57818064e-01 2.23619878e-01
-3.06287318e-01 -4.60739821e-01 1.78105846e-01 3.44213217e-01
-2.42131278e-01 -1.01010464e-01 -1.96711004e-01 6.13543570e-01
1.64978161e-01 1.15614212e+00 -5.10475397e-01 2.09725961e-01
2.51587391e-01 6.96404457e-01 -3.28822173e-02 1.77272603e-01
2.35506862e-01 7.80158997e-01 -1.96120173e-01 6.75807595e-01
3.10858101e-01 2.23721489e-01 -1.99357107e-01 -5.64476490e-01
-1.44378722e-01 5.05606174e-01 -7.52150655e-01 2.50731564e+00
-5.80921352e-01 1.06356943e+00 -1.31348997e-01 -8.32022667e-01
6.29098058e-01 7.81374454e-01 7.18567967e-01 -8.89011919e-01
1.00173032e+00 3.88895661e-01 4.59925473e-01 -8.70691180e-01
-1.94403604e-01 1.45143285e-01 4.09035265e-01 5.51641583e-01
3.91480178e-01 2.60258794e-01 -4.43408132e-01 -4.82023358e-01
1.24252856e+00 6.75404191e-01 -6.44810125e-02 -3.50999445e-01
1.87223136e-01 1.00269951e-01 -6.15188554e-02 2.31194034e-01
-1.18156284e-01 4.56201464e-01 7.44453892e-02 1.60685834e-02
-7.06805706e-01 -1.01359117e+00 9.23894439e-03 7.72299647e-01
9.41369310e-02 -1.00917779e-01 -8.67043078e-01 2.69572526e-01
5.38286790e-02 4.66855735e-01 -4.54871297e-01 -5.04845917e-01
-8.50390017e-01 -5.35335362e-01 7.29529440e-01 1.11407292e+00
3.97282094e-01 -1.62437475e+00 -1.12866056e+00 7.37143695e-01
-1.31690994e-01 -8.88191819e-01 2.36067306e-02 4.74328965e-01
-1.14604616e+00 -1.02991450e+00 -1.17502427e+00 -9.53911424e-01
1.58132032e-01 -2.01444179e-01 2.40153610e-03 -6.62973881e-01
-4.53635037e-01 4.26766038e-01 -3.35545152e-01 -5.28033555e-01
2.51464933e-01 8.35608989e-02 4.30033952e-01 -5.58709860e-01
8.45885754e-01 -1.02206981e+00 -6.02072060e-01 -4.59929220e-02
-1.65906906e-01 1.79665774e-01 1.01893997e+00 7.34862506e-01
4.00902092e-01 -4.12222952e-01 9.86388683e-01 4.07962173e-01
1.31680810e+00 -5.17180085e-01 -1.11405089e-01 2.20275238e-01
-3.30269307e-01 6.87033311e-02 3.56236994e-01 -6.64874971e-01
-9.62390423e-01 -1.46956429e-01 -1.54543683e-01 -5.23531675e-01
-6.85022399e-02 8.40408623e-01 2.19030045e-02 -8.85890797e-02
6.04654014e-01 6.23394310e-01 1.01860501e-02 -4.94719297e-01
-1.31518096e-01 1.16245747e+00 1.04181802e+00 -6.85254395e-01
-2.52784699e-01 -2.32868586e-02 -1.93298861e-01 -1.03740120e+00
3.35081190e-01 -2.35045284e-01 -7.85757780e-01 -5.99102020e-01
1.03115737e+00 -8.95285487e-01 -1.02697742e+00 8.50864053e-01
-1.60799634e+00 -6.36214912e-01 5.71160316e-01 1.47937489e+00
-7.30303586e-01 -1.23620853e-01 -6.99525774e-01 -7.91484237e-01
-8.91037226e-01 -1.40182078e+00 7.91910827e-01 -7.57241175e-02
-5.54886401e-01 -3.51597726e-01 -2.72797942e-01 -1.99619517e-01
2.45364279e-01 2.38614842e-01 6.42464221e-01 -1.09787084e-01
-4.32545356e-02 -5.39265454e-01 -1.41080618e-01 2.77414620e-01
2.73876280e-01 -8.18570614e-01 -8.60034406e-01 -2.41969720e-01
2.85316050e-01 -1.94616094e-01 3.88070911e-01 8.05058956e-01
1.05045211e+00 2.05639713e-02 -3.07123870e-01 5.12330711e-01
1.23559272e+00 6.23866618e-01 8.92577052e-01 3.62979501e-01
7.87441254e-01 4.17518318e-01 -3.68874916e-03 1.94429711e-01
1.26462698e-01 3.43011379e-01 5.54087050e-02 6.63827240e-01
-2.72693723e-01 1.11398883e-01 6.08606279e-01 1.10065854e+00
-8.56194019e-01 8.92227590e-02 -1.02985752e+00 4.50262040e-01
-1.64561617e+00 -6.32142842e-01 -3.13757122e-01 1.96142125e+00
7.25118995e-01 -5.45799546e-02 -3.20109040e-01 4.99213696e-01
4.01234150e-01 -5.83050311e-01 -8.68935823e-01 -3.10405195e-01
2.42633760e-01 1.02588475e+00 6.58350348e-01 6.27601594e-02
-4.83019888e-01 8.22657347e-01 5.20001936e+00 6.22567594e-01
-1.63134348e+00 5.48241615e-01 -4.37999815e-01 -4.79023963e-01
4.45122331e-01 -5.80092847e-01 -4.09055114e-01 7.40310192e-01
1.14722311e+00 5.31648733e-02 8.42379630e-01 7.19672382e-01
5.78346610e-01 -3.85694027e-01 -1.30016673e+00 1.58302426e+00
-6.24012277e-02 -1.07694364e+00 -4.59811687e-01 1.35378186e-02
1.75991625e-01 5.84180593e-01 -1.27651751e-01 4.97875400e-02
-6.31073773e-01 -1.25942361e+00 8.72427642e-01 6.87770069e-01
1.31120205e+00 -7.45772839e-01 6.58290029e-01 3.62164885e-01
-1.47229290e+00 -2.06697613e-01 -2.71623224e-01 -4.30996567e-01
2.33688489e-01 -1.50672376e-01 -5.40018737e-01 3.02079469e-01
7.00106442e-01 8.30679476e-01 -7.75455460e-02 8.37237656e-01
-2.86294073e-01 7.58887649e-01 -3.81298691e-01 -3.17845732e-01
2.21568391e-01 1.28604621e-01 6.37078404e-01 1.08493650e+00
7.12969601e-01 2.89859295e-01 -6.20619595e-01 9.13997471e-01
2.13125080e-01 -1.32611230e-01 -5.67730963e-01 -2.00111374e-01
3.48523349e-01 5.26786149e-01 -4.62016284e-01 1.28813520e-01
-2.32766420e-01 1.42764056e+00 2.24116389e-04 4.87909317e-01
-7.75112510e-01 -7.77043819e-01 5.75586081e-01 -2.57720798e-01
-4.90713835e-01 -6.05598092e-01 -6.86657012e-01 -9.13718283e-01
4.90641266e-01 -4.01714355e-01 -3.03408533e-01 -1.12547231e+00
-8.54977190e-01 5.64837933e-01 3.06313246e-01 -1.10967565e+00
-2.67853349e-01 -1.08370364e+00 -7.13527739e-01 1.45428741e+00
-1.12461865e+00 -9.35120404e-01 -5.61980367e-01 1.02775896e+00
6.94672227e-01 -3.53132337e-01 7.91146159e-01 3.47283900e-01
-4.27628100e-01 4.08203900e-01 9.68236998e-02 1.33026838e-01
4.11266655e-01 -8.01162422e-01 3.20553407e-02 5.22942543e-01
-1.82907447e-01 9.90441740e-01 3.94499660e-01 -1.03248656e+00
-1.96570337e+00 -7.24275470e-01 4.12142813e-01 1.71859842e-02
5.25542140e-01 -1.29757747e-01 -7.60190725e-01 5.99809825e-01
2.90953189e-01 -4.57672268e-01 3.26868087e-01 -4.81822938e-01
4.36175317e-01 5.21021374e-02 -8.85921538e-01 4.66214716e-01
1.13525987e+00 -6.41064048e-01 -1.00650692e+00 5.69036789e-02
-9.71710905e-02 -3.49002987e-01 -8.87742460e-01 2.82846212e-01
1.19216013e+00 -2.88084090e-01 8.51191223e-01 -5.55364072e-01
2.40956366e-01 1.21640377e-01 -1.29462123e-01 -1.68090749e+00
-1.80487126e-01 -3.81721348e-01 -3.24044645e-01 5.21758676e-01
6.19434007e-02 -5.86174488e-01 6.99689388e-01 5.03099978e-01
-6.44717216e-01 -8.06378365e-01 -1.16367126e+00 -9.85290587e-01
1.03438683e-01 -1.05517304e+00 3.51632833e-01 1.51029766e-01
8.15225124e-01 9.15565491e-02 -3.44668776e-01 1.87532101e-02
2.74817079e-01 -5.16489983e-01 2.88940847e-01 -8.91121089e-01
1.98743179e-01 -5.42173862e-01 -7.60352314e-01 -8.40249002e-01
4.01721776e-01 -1.33409083e+00 2.62972116e-01 -1.99111247e+00
-7.13569149e-02 1.14221439e-01 -6.42634392e-01 5.19529283e-01
2.46742517e-01 4.40328084e-02 -1.28042787e-01 4.18664664e-01
6.08490944e-01 4.93735701e-01 1.06536031e+00 -9.19619203e-02
-5.62979460e-01 -3.23352367e-01 1.38943240e-01 4.33690310e-01
1.19576573e+00 -4.16252047e-01 -5.83064795e-01 -6.21633291e-01
-2.85959959e-01 4.07872826e-01 5.23144722e-01 -1.45031130e+00
5.27006209e-01 2.45047227e-01 8.54130626e-01 -7.88105190e-01
5.21431923e-01 -5.33963203e-01 2.22731069e-01 5.41201591e-01
-1.54185146e-01 1.02334604e-01 3.71502608e-01 5.43312550e-01
-8.10293853e-02 1.48205742e-01 3.15230876e-01 -3.04911137e-02
-8.91509950e-01 1.22098081e-01 -8.80370319e-01 -4.99219298e-01
9.60543931e-01 -8.50499749e-01 5.51106408e-02 1.85856611e-01
-9.03791845e-01 -1.28730267e-01 -3.93431842e-01 5.66883326e-01
1.06834257e+00 -1.58760583e+00 -5.29667199e-01 4.44909036e-01
3.54017504e-02 -5.68166375e-01 3.12599152e-01 1.44384074e+00
-5.86362302e-01 8.60905051e-01 -1.06994092e+00 -5.87294996e-01
-9.37850773e-01 2.52427664e-02 1.11393444e-01 6.26290321e-01
-1.19850802e+00 1.21416688e+00 -5.78249216e-01 3.40245724e-01
5.05007744e-01 -7.40388155e-01 -2.10612282e-01 -1.32829472e-01
4.31663632e-01 4.30842370e-01 3.64979804e-01 -6.05807006e-01
-7.18428969e-01 3.90535027e-01 4.11850214e-01 -4.10881877e-01
1.48324418e+00 1.28006458e-01 1.97635721e-02 5.15890896e-01
1.34435439e+00 -7.68222451e-01 -1.28032207e+00 3.45297188e-01
1.26987427e-01 -1.06768303e-01 5.57862282e-01 -1.08234417e+00
-1.03356695e+00 1.25606775e+00 1.08664751e+00 -9.65964079e-01
1.02620769e+00 -4.81625736e-01 1.34378123e+00 6.75853789e-01
9.26071823e-01 -1.40113318e+00 5.89401871e-02 2.40080371e-01
1.38592076e+00 -9.35336590e-01 -2.01839656e-01 4.13464934e-01
-5.68462014e-01 1.66281891e+00 6.38006330e-01 -5.86152017e-01
9.27935004e-01 4.53782946e-01 -2.48811722e-01 -3.21040779e-01
-2.43472412e-01 1.11752875e-01 4.91795003e-01 5.61619759e-01
5.32900512e-01 2.76776969e-01 -5.47798991e-01 1.19660711e+00
-3.28388363e-01 7.48478711e-01 1.64080501e-01 1.46024024e+00
-2.46549904e-01 -4.99015808e-01 -4.59870666e-01 9.13002193e-01
-3.64582956e-01 -1.35763884e-01 -1.27608806e-01 6.75400555e-01
-6.92158416e-02 8.06683719e-01 -1.59797713e-01 -9.81283009e-01
1.05823494e-01 4.55493182e-01 1.18434870e+00 -4.60868806e-01
-6.22859657e-01 8.65894929e-02 -3.18364352e-01 -8.23652804e-01
-2.05475271e-01 -2.72649497e-01 -1.74214971e+00 -9.17957872e-02
-4.36444521e-01 -3.62817854e-01 1.34554160e+00 1.19837141e+00
3.55493695e-01 8.50494862e-01 -8.26697424e-02 -1.36176550e+00
-2.13326693e-01 -1.52220690e+00 -6.66685820e-01 -1.82896376e-01
7.41718635e-02 -1.25734806e+00 1.05808109e-01 1.66092634e-01] | [12.989822387695312, 3.4111857414245605] |
5a870447-de76-484c-9b9f-6b6f3f4ea86d | systems-challenges-for-trustworthy-embodied | 2201.03413 | null | https://arxiv.org/abs/2201.03413v2 | https://arxiv.org/pdf/2201.03413v2.pdf | Systems Challenges for Trustworthy Embodied Systems | A new generation of increasingly autonomous and self-learning embodied systems is about to be developed. When deploying embodied systems into a real-life context we face various engineering challenges, as it is crucial to coordinate the behavior of embodied systems in a beneficial manner, ensure their compatibility with our human-centered social values, and design verifiably safe and reliable human-machine interaction. We are arguing that traditional systems engineering is coming to a climacteric from embedded to embodied systems, and with assuring the trustworthiness of dynamic federations of situationally aware, intent-driven, explorative, ever-evolving, largely non-predictable, and increasingly autonomous embodied systems in uncertain, complex, and unpredictable real-world contexts. We are therefore identifying a number of urgent systems challenges for trustworthy embodied systems, including robust and human-centric AI, cognitive architectures, uncertainty quantification, trustworthy self-integration, and continual analysis and assurance. | ['Harald Rueß'] | 2022-01-10 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [-2.39693537e-01 2.05350190e-01 5.20423234e-01 -3.03256541e-01
-3.70676927e-02 -7.02705085e-01 8.61543059e-01 4.55965847e-02
-1.37788653e-01 7.69531250e-01 1.89907074e-01 -1.08089678e-01
-4.77078736e-01 -4.91645813e-01 -3.59676093e-01 -2.67892867e-01
-3.04067791e-01 4.19533908e-01 -4.79015559e-02 -7.52577305e-01
-1.05564460e-01 3.63536835e-01 -1.86323631e+00 -4.16570336e-01
8.97770047e-01 1.30461037e+00 -3.45317274e-01 7.95902431e-01
6.35125399e-01 1.32985485e+00 -3.33028197e-01 -2.06893012e-01
-1.89395517e-03 -3.47588301e-01 -7.87273705e-01 -3.56215715e-01
-4.78109777e-01 -1.53827608e-01 9.47620720e-02 9.64175403e-01
3.45467746e-01 2.11615592e-01 3.12985659e-01 -1.83584952e+00
-1.05480301e+00 6.64651752e-01 1.47653535e-01 -3.44557434e-01
4.26934749e-01 5.63524067e-01 2.88861513e-01 -5.17865539e-01
5.89626491e-01 1.05470085e+00 6.02951348e-01 5.95248818e-01
-8.05047631e-01 -4.57509547e-01 1.83556691e-01 -7.58481026e-02
-1.58325052e+00 -1.31363618e+00 2.40676150e-01 -5.96196234e-01
8.55886519e-01 4.85157281e-01 8.51846695e-01 1.34003842e+00
8.69337380e-01 6.92679882e-02 9.44193721e-01 -3.88607562e-01
8.53133440e-01 4.50244784e-01 -1.86869085e-01 5.89128375e-01
4.69083607e-01 5.63995600e-01 -7.83637464e-01 -2.67346621e-01
1.82090759e-01 -3.49186629e-01 2.79958844e-02 -3.32095176e-01
-1.66550183e+00 1.29698599e-02 6.37163594e-02 7.52236366e-01
-7.06791818e-01 7.55012453e-01 3.26740175e-01 2.47170746e-01
2.20318198e-01 7.25476980e-01 -1.86422080e-01 -6.97084308e-01
-4.99402016e-01 3.03045034e-01 1.01750541e+00 1.00955284e+00
1.56478226e-01 3.62508237e-01 4.37488586e-01 -2.10847318e-01
6.64645791e-01 6.93669200e-01 2.36089051e-01 -1.33422554e+00
-6.78514957e-01 5.47045231e-01 4.90997225e-01 -1.09968197e+00
-4.39657748e-01 -1.97194517e-01 -4.31213230e-01 7.07954168e-01
-1.75187126e-01 -4.56328392e-01 -4.76904809e-01 1.80510318e+00
6.65684283e-01 -2.84374326e-01 5.79074383e-01 1.00437522e+00
2.06824049e-01 4.34957415e-01 2.66713947e-01 -1.30457446e-01
1.04659605e+00 -2.26020217e-01 -9.26031172e-01 -2.11532757e-01
2.42157793e-03 -3.76084089e-01 6.36134267e-01 1.51973531e-01
-1.19875729e+00 3.90906706e-02 -1.16332960e+00 1.89987242e-01
-4.09507424e-01 -6.31008327e-01 5.37888706e-01 5.61278880e-01
-1.23853719e+00 3.53930980e-01 -1.09425044e+00 -5.42612553e-01
-4.26333211e-03 3.41162026e-01 -5.72608531e-01 6.14433765e-01
-1.25806487e+00 1.49247897e+00 2.99986064e-01 2.78181940e-01
-1.10456693e+00 -2.93703407e-01 -7.26351678e-01 -1.48976490e-01
4.42300051e-01 -7.42836535e-01 1.28908777e+00 -1.01814425e+00
-1.39274824e+00 3.81649703e-01 3.31149459e-01 -3.41214538e-01
4.66687053e-01 8.02014023e-03 -9.31476951e-01 -2.80196637e-01
-1.73963726e-01 6.24281690e-02 5.88615477e-01 -1.61210024e+00
-5.19315958e-01 -4.79587287e-01 -1.78001404e-01 3.25860709e-01
-2.63124347e-01 7.42305219e-02 8.52551699e-01 1.07044742e-01
-4.29725042e-03 -1.14341712e+00 -2.88949817e-01 1.17271721e-01
6.86125234e-02 1.96307406e-01 9.43087935e-01 -4.17643487e-01
7.50035346e-01 -2.07504368e+00 2.20171943e-01 2.75488913e-01
7.37767756e-01 1.43537372e-02 3.17174137e-01 7.03012168e-01
5.59585214e-01 2.31146649e-01 2.21601248e-01 -7.79007468e-03
7.14062274e-01 -1.52186990e-01 -2.82178104e-01 3.58139843e-01
-7.37571046e-02 8.75278831e-01 -1.16453218e+00 -3.76740456e-01
1.21554002e-01 5.33336222e-01 1.06545761e-01 3.52110058e-01
-6.63703084e-02 5.91476500e-01 -7.27516413e-01 8.10239136e-01
-1.40662506e-01 -4.68478501e-02 2.92322874e-01 2.89115578e-01
-5.19423544e-01 -3.68259966e-01 -9.66087401e-01 1.22287822e+00
-6.89746261e-01 6.40396893e-01 5.86292744e-01 5.14172912e-02
9.55938816e-01 6.64116800e-01 4.47444886e-01 -1.05911839e+00
7.06675887e-01 4.95978743e-01 9.61251110e-02 -5.70442021e-01
9.25015748e-01 -2.89552480e-01 -4.46269274e-01 8.99677932e-01
-1.16488889e-01 -3.65661591e-01 -3.30308378e-01 2.18263760e-01
1.30342174e+00 2.82110125e-01 2.56951511e-01 -6.68103933e-01
9.10484344e-02 -1.25875711e-01 4.23654824e-01 2.16048673e-01
-9.13219392e-01 5.57012856e-02 -6.30561309e-03 -6.39696598e-01
-7.61674523e-01 -8.78068149e-01 1.69271663e-01 9.77214158e-01
4.12302852e-01 8.20606798e-02 -8.72816384e-01 3.70328948e-02
-3.14060040e-02 1.27612031e+00 -5.22149980e-01 -8.73580635e-01
8.25777352e-02 1.17216811e-01 5.25674462e-01 2.85489887e-01
4.40442502e-01 -9.34733331e-01 -1.45484614e+00 5.39857447e-01
2.25684434e-01 -9.97439027e-01 -1.69403434e-01 2.00701907e-01
7.37223476e-02 -6.53847516e-01 1.62455559e-01 -1.75374359e-01
3.52672011e-01 -8.76852870e-02 1.01799703e+00 2.72491109e-02
8.55331216e-03 9.04986978e-01 -2.31374204e-01 -6.72155440e-01
-9.88769233e-01 -4.05250400e-01 7.19838858e-01 -1.36868834e-01
9.28364173e-02 -6.77186966e-01 -6.24341607e-01 2.88404852e-01
-7.20333219e-01 1.43604338e-01 8.64819661e-02 4.00482535e-01
-5.86167872e-02 2.34654650e-01 1.06500947e+00 -7.52480030e-02
8.65726650e-01 -7.56853580e-01 -3.03662896e-01 9.56834793e-01
-1.15874672e+00 -2.48025298e-01 5.15130520e-01 -3.35278779e-01
-1.28579283e+00 -3.25849950e-01 3.71188164e-01 9.81989577e-02
-3.41983706e-01 2.08002225e-01 -3.04664880e-01 -2.53081888e-01
9.56355393e-01 -1.55558273e-01 4.57175195e-01 6.19806290e-01
8.94346297e-01 1.14979243e+00 6.96100354e-01 -1.02024567e+00
6.71141505e-01 4.29459989e-01 -3.77498537e-01 -4.46747720e-01
2.02628046e-01 4.86435831e-01 -2.82615781e-01 -9.40618575e-01
6.45146191e-01 -7.75904298e-01 -1.00005972e+00 5.25151968e-01
-8.94556642e-01 -5.38043559e-01 -6.50361419e-01 1.92658469e-01
-6.36411548e-01 -2.58722216e-01 -2.08614543e-01 -1.45192802e+00
-6.43226445e-01 -1.00089347e+00 7.87875593e-01 4.65516984e-01
-1.09857118e+00 -6.59387231e-01 3.70801657e-01 2.40560696e-01
1.09160483e+00 8.35861683e-01 4.06018615e-01 -4.95468050e-01
-5.57322502e-01 -4.27159548e-01 3.79772753e-01 1.25207871e-01
3.33350122e-01 4.93692696e-01 -9.03259754e-01 -6.91353455e-02
1.95367083e-01 -5.00325680e-01 -5.91208696e-01 -4.19830084e-01
-3.73002626e-02 -5.74090719e-01 -1.01227589e-01 -6.01141676e-02
9.95956004e-01 8.05824339e-01 2.11076438e-01 1.41481698e-01
2.14158222e-01 9.32188630e-01 6.45013332e-01 4.80495334e-01
1.06860542e+00 -5.39427549e-02 3.86329561e-01 3.74974310e-01
4.85203922e-01 -4.62991446e-02 1.95669517e-01 6.47129059e-01
-3.00249141e-02 -1.92577317e-01 -1.21690011e+00 4.69222546e-01
-2.23118925e+00 -9.32514310e-01 5.73764622e-01 1.93073189e+00
6.27270103e-01 -1.04480445e-01 1.32293291e-02 -1.98153015e-02
5.03419161e-01 -2.57039070e-01 -8.61624718e-01 -6.34068847e-01
2.59464502e-01 -4.31664556e-01 -4.36216667e-02 8.57347474e-02
-6.31893575e-01 7.88170099e-01 6.53117418e+00 -1.15367651e-01
-1.13884628e+00 3.10536325e-01 6.38491809e-01 -1.28659710e-01
-7.26000309e-01 6.31261617e-03 2.74303466e-01 4.44069564e-01
1.59911633e+00 -8.27535212e-01 8.09758425e-01 8.43251586e-01
1.90176025e-01 -4.16171670e-01 -1.05957532e+00 4.50039029e-01
-2.41657853e-01 -1.17882109e+00 -9.58719134e-01 -1.50440156e-01
6.45110130e-01 -1.94401875e-01 3.32976282e-02 -8.53239670e-02
8.15470457e-01 -1.20193124e+00 1.82438147e+00 1.10257065e+00
3.83319974e-01 -8.75615180e-01 3.70658517e-01 4.03983921e-01
-9.19842362e-01 -2.66437143e-01 5.48203111e-01 -2.65823334e-01
2.75656819e-01 2.97731221e-01 -1.97035838e-02 1.26673117e-01
7.59556234e-01 1.24153337e-02 -3.68112661e-02 3.55587929e-01
5.77817142e-01 -2.49072671e-01 -1.42298654e-01 -4.43020552e-01
-1.08717881e-01 3.22681665e-02 4.70860213e-01 2.86013156e-01
2.15772808e-01 4.40404236e-01 -3.71291786e-01 9.95385289e-01
4.39850092e-01 -4.44032043e-01 -1.02032149e+00 -8.75977755e-01
8.28371525e-01 1.33147955e+00 -9.93271172e-01 6.66746730e-03
1.44803017e-01 4.64036882e-01 -6.59218952e-02 1.00855075e-01
-9.31239963e-01 -1.29377395e-01 8.29401433e-01 -1.74170971e-01
-7.23901749e-01 -7.15750754e-01 -3.16816241e-01 -7.85357356e-01
-9.64168161e-02 -1.33892167e+00 -2.77685016e-01 -7.91818082e-01
-1.04205501e+00 1.09520221e+00 -2.93866277e-01 -7.60281503e-01
-5.42445540e-01 1.01566121e-01 -3.44504237e-01 1.96869016e-01
-7.07182288e-01 -1.55267274e+00 -4.50293779e-01 2.70939708e-01
-4.43826050e-01 -2.78356224e-01 1.12073946e+00 -5.60166016e-02
-3.49675149e-01 5.84741309e-03 7.56273270e-02 -7.46128619e-01
3.34753245e-01 -8.23357165e-01 4.16883677e-01 8.50378811e-01
-4.08672363e-01 8.06260109e-01 8.90042186e-01 -5.65799773e-01
-2.18170142e+00 -7.41631091e-01 3.75265986e-01 -8.03160906e-01
9.65451717e-01 -4.74174589e-01 -5.93780279e-01 7.45424509e-01
3.67010325e-01 -1.87474057e-01 5.77427566e-01 -1.34402052e-01
-1.36056244e-01 -2.05548376e-01 -1.57055104e+00 8.66455555e-01
1.03881454e+00 -8.72370660e-01 -3.98218066e-01 -1.51144918e-02
1.12704480e+00 -3.26952398e-01 -1.01813197e+00 2.70857185e-01
9.68383551e-01 -8.95046055e-01 5.25538087e-01 -2.76540011e-01
-1.53742105e-01 -5.66218793e-01 -3.11310828e-01 -1.02606726e+00
-1.53750017e-01 -1.28198826e+00 -2.00943097e-01 1.42582369e+00
6.01956025e-02 -1.16632414e+00 1.05213486e-01 1.97430933e+00
-7.26658553e-02 -4.40038890e-01 -8.29011798e-01 -7.36333370e-01
-1.56446412e-01 -3.05795908e-01 1.34078109e+00 1.03468943e+00
7.76442409e-01 -2.12477237e-01 -1.75609708e-01 1.70524195e-01
5.78170896e-01 -1.50817230e-01 4.59524453e-01 -8.93079102e-01
3.06368470e-01 -2.76582867e-01 -6.25437498e-01 5.47223151e-01
-2.39244569e-02 -2.49175668e-01 7.96121478e-01 -1.46799290e+00
-2.75978446e-01 -6.29964232e-01 -7.16003254e-02 3.41663361e-01
3.01391304e-01 -2.72324532e-01 2.21747234e-01 7.65864477e-02
-9.35112417e-01 9.75161076e-01 7.47230232e-01 1.76401466e-01
-5.98427616e-02 -5.43887973e-01 -1.14711213e+00 6.17394924e-01
4.71385419e-01 -8.39907750e-02 -6.15225792e-01 -1.78489640e-01
5.26984334e-01 1.27234951e-01 5.15357435e-01 -1.39140534e+00
8.89598727e-01 -8.13035369e-01 1.62494313e-02 2.32045725e-01
2.46283323e-01 -1.34123969e+00 1.18010783e+00 4.71792519e-01
-2.11019173e-01 -5.34234196e-02 3.08130216e-02 2.23160252e-01
2.44922638e-01 4.13344383e-01 4.66805398e-01 1.05191479e-02
-8.78239691e-01 4.06381451e-02 -6.11104310e-01 -1.94441631e-01
1.83544874e+00 -1.61839783e-01 -7.17134655e-01 -4.88764554e-01
-5.09852409e-01 6.16601825e-01 9.76596057e-01 6.62658453e-01
4.78972375e-01 -9.10467744e-01 -3.32599103e-01 1.21834986e-01
7.44277164e-02 -5.56067675e-02 4.04688269e-01 2.30646715e-01
-3.85118574e-01 4.66417111e-02 -5.29843807e-01 -1.26723707e-01
-6.29295647e-01 2.87058473e-01 6.24734342e-01 8.43259633e-01
-3.85086268e-01 5.27711809e-01 -4.09050226e-01 -4.00827259e-01
1.45183578e-01 9.76787955e-02 4.19785678e-01 -1.86772883e-01
5.16488075e-01 5.50000966e-01 -1.68545127e-01 -8.13591063e-01
-8.24495018e-01 8.09671730e-03 6.26054645e-01 -5.95866323e-01
1.20750368e+00 -3.21881920e-01 -5.21619856e-01 7.01408088e-01
4.09311086e-01 -3.40493888e-01 -1.23772275e+00 4.48948443e-01
2.05286205e-01 -2.57668197e-01 2.36131549e-01 -1.01259661e+00
-5.35923660e-01 4.38609888e-04 6.70830846e-01 7.29051590e-01
9.99808848e-01 -1.18528545e-01 7.42626488e-01 3.47469538e-01
1.17144227e+00 -1.42116976e+00 5.58363721e-02 9.70888063e-02
1.13711572e+00 -8.81254017e-01 -3.18228640e-02 3.13523084e-01
-9.71865773e-01 8.43615294e-01 7.78710127e-01 3.36632878e-01
7.82387495e-01 9.44105506e-01 2.09995314e-01 -3.29914451e-01
-1.08210659e+00 2.62088683e-02 -2.11272091e-01 7.81037509e-01
1.15515761e-01 3.85720253e-01 2.20128089e-01 5.91060758e-01
-1.99902266e-01 2.19852045e-01 5.49962103e-01 1.39513433e+00
-5.02063155e-01 -6.96874201e-01 -3.26382816e-01 -2.42542863e-01
1.79080665e-01 5.71701288e-01 -6.32623851e-01 8.03008080e-01
1.84165150e-01 1.27161932e+00 -9.61340219e-02 -7.41097450e-01
2.80814916e-01 1.37096211e-01 3.00191548e-02 -2.33610615e-01
-7.25685835e-01 -5.77284276e-01 3.23958725e-01 -7.27851152e-01
-6.21046983e-02 -6.84042811e-01 -1.76892102e+00 -1.05427825e+00
-2.92789668e-01 1.44057140e-01 1.34636259e+00 8.96382868e-01
1.10044861e+00 4.69093710e-01 6.66797698e-01 -6.07871830e-01
-5.31527579e-01 -2.63271034e-01 -3.63878518e-01 -2.73839593e-01
3.42993796e-01 -4.69612837e-01 -3.42394471e-01 -1.87105238e-01] | [8.930144309997559, 6.2303009033203125] |
94e30143-e07c-47e1-b96e-4f3a1f71a6fa | analysing-mixed-initiatives-and-search | 2109.05955 | null | https://arxiv.org/abs/2109.05955v1 | https://arxiv.org/pdf/2109.05955v1.pdf | Analysing Mixed Initiatives and Search Strategies during Conversational Search | Information seeking conversations between users and Conversational Search Agents (CSAs) consist of multiple turns of interaction. While users initiate a search session, ideally a CSA should sometimes take the lead in the conversation by obtaining feedback from the user by offering query suggestions or asking for query clarifications i.e. mixed initiative. This creates the potential for more engaging conversational searches, but substantially increases the complexity of modelling and evaluating such scenarios due to the large interaction space coupled with the trade-offs between the costs and benefits of the different interactions. In this paper, we present a model for conversational search -- from which we instantiate different observed conversational search strategies, where the agent elicits: (i) Feedback-First, or (ii) Feedback-After. Using 49 TREC WebTrack Topics, we performed an analysis comparing how well these different strategies combine with different mixed initiative approaches: (i) Query Suggestions vs. (ii) Query Clarifications. Our analysis reveals that there is no superior or dominant combination, instead it shows that query clarifications are better when asked first, while query suggestions are better when asked after presenting results. We also show that the best strategy and approach depends on the trade-offs between the relative costs between querying and giving feedback, the performance of the initial query, the number of assessments per query, and the total amount of gain required. While this work highlights the complexities and challenges involved in analyzing CSAs, it provides the foundations for evaluating conversational strategies and conversational search agents in batch/offline settings. | ['Nick Craswel', 'Paul Thomas', 'Evangelos Kanoulas', 'Hamed Zamani', 'Leif Azzopardi', 'Mohammad Aliannejadi'] | 2021-09-13 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [ 2.29542747e-01 2.48544544e-01 -1.09411843e-01 -2.48976827e-01
-1.02195740e+00 -1.01940000e+00 1.14142501e+00 3.71014297e-01
-8.06844175e-01 4.47487354e-01 5.74078321e-01 -5.70294499e-01
-5.95847130e-01 -4.05042499e-01 -1.08631186e-01 -5.53784549e-01
1.51100397e-01 9.39193308e-01 3.91025186e-01 -4.77142662e-01
4.07290578e-01 2.97942519e-01 -1.57541585e+00 1.88145936e-01
7.23347247e-01 6.24891877e-01 4.77955580e-01 9.84416604e-01
-6.79010272e-01 7.73381531e-01 -7.24334180e-01 -3.53314847e-01
9.13145319e-02 -5.17481565e-01 -1.35298693e+00 3.12009212e-02
-9.53767151e-02 -3.72955233e-01 9.68971625e-02 5.81358433e-01
4.87146437e-01 2.94854611e-01 4.89965171e-01 -1.23832738e+00
1.29588068e-01 3.88109416e-01 8.36454779e-02 4.01493192e-01
9.75624323e-01 3.51148933e-01 1.18729365e+00 -3.36112231e-01
7.30181217e-01 1.18085480e+00 2.48977631e-01 4.52814460e-01
-1.14842093e+00 -1.02839462e-01 8.99531916e-02 -1.73087101e-02
-8.82164896e-01 -7.15322912e-01 4.55562890e-01 -4.29494053e-01
9.60843205e-01 9.06349540e-01 6.01307750e-01 9.15179253e-01
-4.16272640e-01 7.45831072e-01 1.07350910e+00 -5.05301893e-01
3.61714065e-01 7.09164977e-01 3.43671918e-01 6.65062014e-03
-2.65923530e-01 3.24696861e-02 -4.98236060e-01 -8.11290145e-01
2.90787727e-01 -7.04932734e-02 -3.43714893e-01 -1.89571366e-01
-8.49640787e-01 7.30706036e-01 1.20318085e-01 6.38266683e-01
-6.59792960e-01 -2.72994161e-01 1.90314814e-01 6.82918727e-01
1.34271890e-01 8.28973234e-01 -2.55946040e-01 -7.56232738e-01
-7.28324890e-01 6.86721742e-01 1.40948272e+00 5.81591785e-01
8.45366180e-01 -9.49907899e-01 -4.23781276e-01 1.07990050e+00
2.63640165e-01 1.14364646e-01 3.07622075e-01 -1.07345164e+00
4.72415000e-01 6.70080245e-01 7.91435361e-01 -8.70608389e-01
-2.64625877e-01 -1.58483505e-01 3.11695114e-02 -9.48743522e-02
7.67626047e-01 -2.07787365e-01 -2.73137242e-01 1.63176095e+00
3.04577380e-01 -8.14901352e-01 1.19253527e-02 8.61590981e-01
5.39207041e-01 4.51953441e-01 6.61194250e-02 -5.22708237e-01
1.57429528e+00 -8.05920780e-01 -7.81871438e-01 -2.70190209e-01
8.36194992e-01 -9.37799513e-01 1.27337706e+00 -2.07270142e-02
-1.34711742e+00 -1.73979215e-02 -4.46307093e-01 4.35906872e-02
-3.33609939e-01 -5.39758027e-01 2.97398001e-01 7.55044520e-01
-1.23439407e+00 1.38205647e-01 -5.64211786e-01 -6.66415274e-01
-5.70452809e-01 2.96234459e-01 1.72076732e-01 1.30505309e-01
-1.16935444e+00 9.19808984e-01 -2.46482819e-01 -1.91211738e-02
-2.71772057e-01 -4.07839745e-01 -2.23856777e-01 4.45007116e-01
6.44238651e-01 -4.89932716e-01 1.87837231e+00 -8.16480815e-01
-1.49914312e+00 5.30540943e-01 -4.54612166e-01 -1.94618419e-01
7.23312020e-01 6.44612238e-02 1.23714522e-01 -2.99368426e-03
5.37439547e-02 5.39538622e-01 3.65928411e-02 -1.13978720e+00
-7.71315932e-01 -4.98376220e-01 6.48204625e-01 9.11129475e-01
-2.45477930e-02 3.48448157e-01 -5.28490365e-01 1.25068292e-01
-1.89340208e-02 -1.07036173e+00 -1.15415700e-01 -3.83129120e-01
-1.25980094e-01 -7.07271874e-01 7.35496879e-01 -4.81937110e-01
1.43490338e+00 -1.96430206e+00 -3.76860872e-02 2.94361800e-01
7.62561942e-03 7.90276825e-02 -4.10245433e-02 1.23564792e+00
3.77426982e-01 3.22074175e-01 2.15022400e-01 -3.34345460e-01
2.35128343e-01 2.08944827e-02 -3.51118222e-02 -2.14970201e-01
-4.02586311e-01 7.63580143e-01 -8.95538807e-01 -4.37081516e-01
-8.48289579e-02 3.67286682e-01 -5.05307257e-01 2.66257048e-01
-3.58898431e-01 3.73766452e-01 -5.29975951e-01 3.38832259e-01
1.62063450e-01 -4.54236656e-01 2.82835394e-01 2.88416088e-01
-2.80282646e-01 9.80458438e-01 -1.02474058e+00 1.13219655e+00
-7.00296402e-01 7.00088263e-01 6.25860512e-01 -5.69215834e-01
5.05183935e-01 5.97745717e-01 3.11192036e-01 -8.46807420e-01
-6.59152344e-02 3.21324110e-01 1.18186757e-01 -5.87414145e-01
6.23066247e-01 1.70107663e-01 1.24299169e-01 1.13458776e+00
-5.08429468e-01 -2.52661765e-01 3.59013379e-01 5.13829827e-01
1.24814415e+00 -4.65283036e-01 4.30280790e-02 -2.04678521e-01
4.90990788e-01 1.15605459e-01 -4.67337444e-02 1.11817741e+00
-3.08853835e-01 1.28729781e-02 6.72881782e-01 -1.96951896e-01
-6.03364527e-01 -4.76081371e-01 1.51642010e-01 1.39542103e+00
2.96052247e-01 -5.10746419e-01 -8.22642148e-01 -5.36824763e-01
-1.83894768e-01 8.05022240e-01 -2.05864266e-01 1.13543987e-01
-3.59121680e-01 -3.04630280e-01 1.80861011e-01 -1.56647503e-01
4.54953760e-01 -1.21609116e+00 -1.00271583e+00 1.60282016e-01
-8.08327079e-01 -7.23522544e-01 -7.77334332e-01 -5.72331585e-02
-6.58207297e-01 -1.02643275e+00 -6.67727232e-01 -3.27055067e-01
2.42477074e-01 5.90875328e-01 1.17202461e+00 4.55610514e-01
-1.15795257e-02 1.03234661e+00 -3.61934334e-01 -1.48018196e-01
-5.00284553e-01 2.30417714e-01 -3.61605853e-01 -2.07512021e-01
4.06599790e-01 -4.72569585e-01 -9.00416195e-01 6.44781947e-01
-8.96279812e-01 -9.29244459e-02 4.45403755e-01 6.17797732e-01
-2.69607306e-01 -1.14646181e-01 3.81083220e-01 -8.45104277e-01
1.49933088e+00 -4.32000995e-01 -5.22254288e-01 4.51286733e-01
-8.63502264e-01 1.76654294e-01 -5.86908348e-02 -2.60035813e-01
-1.23140478e+00 -3.21327716e-01 -1.96893793e-03 3.73480797e-01
-2.85217632e-02 5.52689135e-01 2.99316078e-01 5.76267168e-02
7.85859227e-01 8.04698616e-02 1.41520619e-01 -4.90820825e-01
1.86277837e-01 1.02379608e+00 -1.00184605e-01 -5.70181727e-01
2.39324749e-01 1.44974992e-01 -8.15807879e-01 -9.99565065e-01
-2.13530883e-01 -8.91794443e-01 -2.57665869e-02 -5.01557767e-01
6.27421200e-01 -2.76435465e-01 -1.27033544e+00 1.65215373e-01
-1.09307826e+00 -3.96315783e-01 8.87678377e-03 1.83737293e-01
-4.49520618e-01 3.66530329e-01 -4.73780632e-01 -1.44598508e+00
-2.49382868e-01 -1.50765216e+00 8.04084003e-01 3.27836484e-01
-7.64075994e-01 -8.72715592e-01 1.61190331e-01 8.54028463e-01
8.63090634e-01 -4.90878552e-01 8.93404365e-01 -1.03492916e+00
-8.93909097e-01 -2.63644338e-01 -7.99293742e-02 -3.52772027e-01
2.27662206e-01 -3.38955849e-01 -8.07420611e-01 -2.17339367e-01
-1.12031261e-02 -3.35460037e-01 1.34943902e-01 1.21701397e-01
3.57077807e-01 -7.43619084e-01 -4.14985925e-01 -3.15719277e-01
1.06973767e+00 5.62318921e-01 4.35485095e-01 4.66707408e-01
-2.89361000e-01 1.20167375e+00 4.16577220e-01 1.83615595e-01
5.42034507e-01 1.10605896e+00 1.44625872e-01 2.17556015e-01
3.07486504e-01 -1.06131554e-01 1.62715241e-01 2.42119595e-01
7.17708319e-02 -3.88389647e-01 -9.33023274e-01 5.61785519e-01
-1.83883965e+00 -1.00754809e+00 2.89048225e-01 2.39188099e+00
9.52516973e-01 1.25128865e-01 5.18132687e-01 2.84496043e-02
4.11125839e-01 8.32115859e-02 -3.89100909e-01 -5.77029467e-01
3.76211226e-01 -1.31239727e-01 5.99138737e-02 1.16895258e+00
-3.48513812e-01 4.68942940e-01 6.27984285e+00 4.10989285e-01
-8.27289104e-01 7.71158448e-05 5.36257744e-01 -1.28901765e-01
-6.15965664e-01 3.15320760e-01 -6.46519065e-01 4.22500759e-01
1.01857126e+00 -3.05348635e-01 9.20099020e-01 5.35282135e-01
2.31439829e-01 -6.56008601e-01 -1.23400700e+00 6.10269248e-01
-3.82739514e-01 -1.10496271e+00 -3.12119752e-01 3.07449728e-01
7.75397047e-02 -1.12937791e-02 -3.49262118e-01 3.82177383e-01
5.51964104e-01 -6.07157826e-01 3.12042981e-01 3.79601240e-01
2.75490675e-02 -1.39905572e-01 6.54224634e-01 9.50273037e-01
-8.25775921e-01 -2.35586628e-01 3.66224527e-01 -2.17246309e-01
2.02760488e-01 1.17763020e-01 -1.07351089e+00 2.86120772e-01
6.72562003e-01 -3.46224397e-01 -3.04633886e-01 9.11372781e-01
3.64979178e-01 2.54344821e-01 -6.10645354e-01 -5.34995258e-01
3.07454109e-01 -4.52940762e-01 6.44244969e-01 1.06959796e+00
2.08625328e-02 2.79426783e-01 1.91649958e-01 7.75385976e-01
3.75738233e-01 2.81772703e-01 -3.99261326e-01 -2.49578357e-01
9.47775543e-01 1.07423890e+00 -6.58676505e-01 -3.29148263e-01
-2.19857112e-01 6.80810153e-01 1.17680550e-01 5.96937180e-01
-4.87161651e-02 -2.31694087e-01 4.05994892e-01 3.59355986e-01
-2.18704864e-01 2.62623783e-02 4.02592532e-02 -6.05067909e-01
3.01568151e-01 -1.16044557e+00 4.12858874e-01 -6.78927898e-01
-8.01859915e-01 5.10803223e-01 3.40134978e-01 -4.93767381e-01
-8.31038237e-01 3.32501680e-02 -6.16654098e-01 1.24244642e+00
-1.15894330e+00 -3.18603694e-01 -3.19607615e-01 1.08434841e-01
7.52148151e-01 3.93442482e-01 7.22740591e-01 1.47470608e-01
-4.93985303e-02 2.78870702e-01 -2.11794823e-01 -2.94046462e-01
4.43512827e-01 -1.08356869e+00 -5.82954176e-02 6.36471063e-02
-4.30269241e-02 9.51893985e-01 9.26073968e-01 -3.16323966e-01
-1.28954220e+00 1.85638621e-01 1.51567292e+00 -4.76688802e-01
3.13483596e-01 -2.68042237e-01 -8.99217725e-01 3.71072501e-01
5.72116137e-01 -1.01565480e+00 5.88390529e-01 5.37758410e-01
4.86654639e-02 9.81823578e-02 -1.05644858e+00 8.40811253e-01
7.08207667e-01 -8.28056574e-01 -4.72978294e-01 5.27201712e-01
6.99132681e-01 -9.01807770e-02 -4.34775800e-01 1.48114234e-01
7.53547430e-01 -1.31089890e+00 8.22916150e-01 -2.26794794e-01
-6.34440854e-02 1.45037621e-01 1.45145487e-02 -9.98020828e-01
1.22958370e-01 -9.67385173e-01 5.21921515e-01 1.14149606e+00
7.58175075e-01 -8.22536469e-01 7.62450397e-01 1.50354826e+00
4.04439092e-01 -8.12166035e-01 -7.54308164e-01 -3.99840027e-01
-2.64724523e-01 -1.43462345e-01 5.36576509e-01 6.00831330e-01
5.28991878e-01 5.78159332e-01 1.06009319e-01 -1.84989169e-01
5.33338301e-02 -5.19677065e-02 8.48581910e-01 -1.14526165e+00
-3.91850621e-01 -6.99887514e-01 3.90964121e-01 -1.45892715e+00
-4.58578199e-01 -3.46544236e-01 9.74515453e-02 -1.55008519e+00
1.79457322e-01 -6.80953741e-01 2.36780256e-01 1.89281344e-01
-1.51442394e-01 -7.27525234e-01 2.68585414e-01 5.79000771e-01
-5.20696163e-01 1.00730330e-01 8.67945254e-01 1.23919621e-01
-8.43119562e-01 6.13620520e-01 -7.01818168e-01 2.20037028e-01
5.08603156e-01 -1.41115934e-01 -7.31073260e-01 -1.46768436e-01
6.69647992e-01 7.81455636e-01 2.04341426e-01 -2.98889846e-01
8.01893413e-01 -2.19721735e-01 -5.15496492e-01 -5.02469003e-01
6.08870804e-01 -8.19648743e-01 1.92391783e-01 4.36127782e-01
-9.68204975e-01 3.38071793e-01 -1.72218308e-01 5.85476339e-01
-2.28394970e-01 -4.88065571e-01 2.42444173e-01 -4.23333704e-01
-6.15470344e-05 -2.59320587e-01 -8.34253669e-01 1.63371573e-04
5.09842336e-01 -3.82866502e-01 -2.39501655e-01 -1.26389265e+00
-8.52177918e-01 5.80671728e-01 3.84506077e-01 3.57136428e-01
-3.97016220e-02 -7.53325939e-01 -3.19834471e-01 -2.25079078e-02
3.18959653e-02 -2.73817569e-01 2.81574409e-02 9.13327336e-01
-1.33843720e-01 8.70239258e-01 3.85608286e-01 -6.42002165e-01
-1.48024046e+00 1.46075130e-01 3.44750583e-01 -6.12838089e-01
-1.69905741e-02 6.40077949e-01 1.52803674e-01 -6.24517739e-01
7.36905277e-01 -4.26595174e-02 -2.32736588e-01 2.52835929e-01
3.74098718e-01 3.85508567e-01 1.96695492e-01 -1.24787293e-01
-1.77559897e-01 1.08529791e-01 -3.42558593e-01 -7.43594408e-01
7.40070283e-01 -4.42895800e-01 -1.38862714e-01 4.14026320e-01
9.02205288e-01 1.64702293e-02 -6.90681696e-01 -4.45167601e-01
4.45892632e-01 -5.35799742e-01 -1.97703645e-01 -1.05901027e+00
-3.62568408e-01 4.50580835e-01 3.32506508e-01 1.06551504e+00
9.30606425e-01 2.03171924e-01 5.66188037e-01 7.04388559e-01
3.23931307e-01 -1.10971391e+00 1.63100690e-01 2.69415677e-01
8.99826884e-01 -9.78802562e-01 -2.75464028e-01 -2.95801669e-01
-6.05710745e-01 6.49138331e-01 3.13410640e-01 6.33793175e-01
5.03815591e-01 -2.21432179e-01 3.18482518e-01 -5.38815618e-01
-1.12602484e+00 -1.88491657e-01 -4.82853837e-02 4.95196804e-02
5.47697544e-01 -1.70733348e-01 -7.76196599e-01 1.03673086e-01
-4.11637574e-01 -2.16977611e-01 1.36941105e-01 1.04257035e+00
-4.06867176e-01 -1.30682349e+00 -3.05801004e-01 3.65365088e-01
-2.34756708e-01 -3.88263203e-02 -9.97479796e-01 7.21832395e-01
-5.89590728e-01 1.59157360e+00 -3.22696716e-02 -9.87351909e-02
4.02364761e-01 5.01742065e-01 1.00116730e-02 -4.34686035e-01
-1.14002573e+00 1.70816720e-01 6.08209312e-01 -4.82861489e-01
-4.76320297e-01 -7.65173674e-01 -6.90451860e-01 -1.93038329e-01
-7.10052669e-01 9.69097972e-01 8.94209027e-01 9.53192890e-01
6.76875234e-01 -9.56377015e-02 6.25829339e-01 -5.21360576e-01
-9.14700568e-01 -1.13801825e+00 -7.67092928e-02 5.10083675e-01
4.81972396e-01 -4.56911802e-01 -8.27480793e-01 -4.12264973e-01] | [12.223397254943848, 7.78875732421875] |
a745dcf5-5300-4c03-bbd3-0f5b33fa2cdd | hunting-for-spammers-detecting-evolved | 1512.02573 | null | http://arxiv.org/abs/1512.02573v2 | http://arxiv.org/pdf/1512.02573v2.pdf | Hunting for Spammers: Detecting Evolved Spammers on Twitter | Once an email problem, spam has nowadays branched into new territories with
disruptive effects. In particular, spam has established itself over the recent
years as a ubiquitous, annoying, and sometimes threatening aspect of online
social networks. Due to its prevalent existence, many works have tackled spam
on Twitter from different angles. Spam is, however, a moving target. The new
generation of spammers on Twitter has evolved into online creatures that are
not easily recognizable by old detection systems. With the strong tangled
spamming community, automatic tweeting scripts, and the ability to massively
create Twitter accounts with a negligible cost, spam on Twitter is becoming
smarter, fuzzier and harder to detect. Our own analysis of spam content on
Arabic trending hashtags in Saudi Arabia results in an estimate of about three
quarters of the total generated content. This alarming rate makes the
development of adaptive spam detection techniques a very real and pressing
need. In this paper, we analyze the spam content of trending hashtags on Saudi
Twitter, and assess the performance of previous spam detection systems on our
recently gathered dataset. Due to the escalating manipulation that
characterizes newer spamming accounts, simple manual labeling currently leads
to inaccurate results. In order to get reliable ground-truth data, we propose
an updated manual classification algorithm that avoids the deficiencies of
older manual approaches. We also adapt the previously proposed features to
respond to spammers evading techniques, and use these features to build a new
data-driven detection system. | ['Alaboodi Saad', 'El-Mawass Nour'] | 2015-12-15 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [-1.08648904e-01 -1.73137471e-01 1.26585111e-01 -2.14522555e-01
-2.10550413e-01 -7.41473794e-01 1.08735287e+00 3.23290914e-01
-3.24131221e-01 7.52698123e-01 1.68969750e-01 -3.61972481e-01
8.22097138e-02 -1.02002537e+00 5.40064313e-02 -5.99194467e-01
-2.91436221e-02 6.02357149e-01 8.21195662e-01 -8.86323452e-01
6.17347538e-01 5.60334027e-01 -1.81144655e+00 2.08736151e-01
1.06328237e+00 7.04370499e-01 -1.09429888e-01 3.41247767e-01
-4.78997529e-01 6.41551495e-01 -7.76616454e-01 -6.09142303e-01
1.39056280e-01 -4.03060913e-01 -6.17843926e-01 9.15208012e-02
3.66772652e-01 -1.40619352e-01 -3.44549626e-01 1.11858904e+00
2.05521956e-01 -5.90620674e-02 3.31953466e-01 -1.10749018e+00
-1.24509558e-01 5.98578155e-01 -6.00150049e-01 1.42770320e-01
3.91569197e-01 1.89761132e-01 6.34505928e-01 -4.48811144e-01
6.12230480e-01 1.25260627e+00 7.33043075e-01 4.49352205e-01
-1.01530778e+00 -9.33353424e-01 -7.62470663e-02 2.01777905e-01
-9.88045335e-01 -2.18148932e-01 4.76846486e-01 -3.34618270e-01
2.14436233e-01 5.67778766e-01 7.72743285e-01 1.12480092e+00
-3.20924968e-02 6.21451557e-01 1.39769256e+00 -1.15601778e-01
3.41616943e-02 5.90739965e-01 1.75371706e-01 5.81031322e-01
7.09341526e-01 -2.23082870e-01 -3.36018354e-01 -5.47053576e-01
1.72496156e-03 4.00789753e-02 -8.36247504e-02 5.73586784e-02
-7.11746573e-01 1.00745642e+00 2.97833622e-01 7.59075344e-01
1.69751551e-02 -2.32241824e-01 5.97185433e-01 5.22898674e-01
6.96828187e-01 4.72370863e-01 -2.18757585e-01 -5.28210521e-01
-1.10974050e+00 3.02056819e-01 1.20826578e+00 4.25330400e-01
7.48885751e-01 -3.52771580e-02 5.36266506e-01 7.39101946e-01
1.68682680e-01 4.96006757e-01 9.25164163e-01 -3.23612511e-01
1.76162869e-01 1.25823355e+00 2.06382915e-01 -1.67625654e+00
-3.92375916e-01 -5.17408907e-01 -6.09913290e-01 6.24557547e-02
8.47116947e-01 1.29597962e-01 -4.81441736e-01 1.15757525e+00
4.48943645e-01 -1.76249489e-01 -7.29177594e-01 7.55885422e-01
2.51846939e-01 4.86978590e-01 3.71134398e-03 -1.66281119e-01
1.21642458e+00 -3.67148012e-01 -5.98221898e-01 -1.13679267e-01
5.99955380e-01 -9.26488578e-01 9.42499280e-01 6.21817827e-01
-4.87110049e-01 1.24901459e-02 -7.96847582e-01 4.87002820e-01
-9.19715106e-01 -4.40352499e-01 5.48572123e-01 1.30531561e+00
-7.10322857e-01 7.85019219e-01 -4.50065374e-01 -7.05627561e-01
3.98202181e-01 3.62915725e-01 -2.51151145e-01 3.41737002e-01
-1.29468024e+00 1.12727332e+00 1.17552698e-01 -2.00888515e-01
-2.03175664e-01 -1.00888982e-01 -6.97183013e-02 -2.80619949e-01
4.67215031e-01 -1.19163260e-01 1.06455457e+00 -1.37369525e+00
-1.21614921e+00 1.02302873e+00 5.62362969e-02 -5.42047560e-01
9.90758777e-01 1.04507431e-01 -7.74674535e-01 5.81688397e-02
1.74764201e-01 -1.17083840e-01 1.32220721e+00 -1.22925663e+00
-7.94679940e-01 -4.32930648e-01 -4.30733234e-01 -3.24098021e-01
-7.76357055e-01 1.92956641e-01 1.91273287e-01 -6.49208486e-01
3.42034727e-01 -8.22604299e-01 -2.33666245e-02 -4.56008106e-01
-2.73416758e-01 -1.97981432e-01 1.33430994e+00 -5.62381566e-01
1.66471446e+00 -1.98441410e+00 -4.21864122e-01 4.86309141e-01
4.68946606e-01 6.63973451e-01 3.74632537e-01 7.49586582e-01
2.24822566e-01 2.23204657e-01 -2.93744653e-01 6.52081296e-02
-1.10361002e-01 2.74795610e-02 -4.92458135e-01 6.32594883e-01
-1.70060202e-01 5.21028101e-01 -1.17580891e+00 -4.55760449e-01
1.07854970e-01 2.30605174e-02 -1.13463365e-01 -1.39448345e-01
-6.07970320e-02 1.26874045e-01 -4.75053936e-01 8.94415796e-01
7.52216220e-01 -1.45682141e-01 1.83595434e-01 2.89673924e-01
-2.85296500e-01 1.91213906e-01 -1.00500262e+00 5.58588445e-01
-1.02955922e-01 7.06813335e-01 3.14480543e-01 -9.31165993e-01
9.82475579e-01 2.25563496e-02 3.45367312e-01 -3.16934973e-01
6.04786396e-01 8.31214428e-01 2.86218207e-02 -4.18620199e-01
7.29553521e-01 -2.10734338e-01 -1.17811479e-01 7.50778913e-01
-3.27909470e-01 -1.49702355e-01 4.75425005e-01 4.25677150e-01
1.24161863e+00 -2.73882866e-01 3.46790254e-01 -2.39637077e-01
8.94987404e-01 7.77766183e-02 -4.84321564e-02 7.51736701e-01
-5.14874816e-01 3.89806718e-01 4.74132329e-01 -5.38699985e-01
-6.97406232e-01 -8.54388475e-01 -2.12326854e-01 1.14359546e+00
9.03332978e-02 -2.81949252e-01 -8.66731942e-01 -9.50636685e-01
3.60150039e-01 3.52566272e-01 -3.35987449e-01 -9.07499194e-02
-7.64926434e-01 -1.14250469e+00 6.23578429e-01 -4.11153883e-01
6.03268385e-01 -8.76626253e-01 -2.16969445e-01 5.96080184e-01
-1.85912892e-01 -7.51235008e-01 -1.16546884e-01 -1.45442560e-01
-1.01154387e+00 -1.23700166e+00 -4.51971799e-01 -5.31416357e-01
5.93565583e-01 6.31624401e-01 8.88516784e-01 5.67918301e-01
-2.98125803e-01 3.22016180e-02 -6.13623559e-01 -3.49969387e-01
-9.56881881e-01 3.57107222e-01 1.42799169e-01 3.59757006e-01
6.27401531e-01 -5.81683099e-01 -3.62840652e-01 6.67958319e-01
-1.07068050e+00 -5.28340936e-01 3.92094821e-01 6.77862644e-01
-7.46272624e-01 1.63891658e-01 1.16681039e+00 -1.28715789e+00
8.55416656e-01 -8.40324104e-01 -4.77708012e-01 -3.12306076e-01
-8.16787660e-01 -3.05214643e-01 7.52483904e-01 -4.55769479e-01
-1.11087298e+00 -1.11338615e-01 -5.48336133e-02 6.22296035e-01
-1.63768753e-01 1.79828212e-01 1.49432614e-01 -2.35839292e-01
1.13240206e+00 2.16831565e-01 5.01078784e-01 -5.26680350e-01
3.77876520e-01 1.40168667e+00 1.01147816e-01 -1.15280338e-01
1.25078011e+00 7.86885619e-01 -3.45780730e-01 -1.23794329e+00
-5.40939331e-01 -8.52352917e-01 -2.66434699e-01 -4.58241612e-01
4.65008728e-02 -3.46859485e-01 -9.52132761e-01 9.25604165e-01
-8.57089579e-01 3.38688016e-01 9.55393836e-02 -2.37325385e-01
4.51860651e-02 1.02030027e+00 -5.93266368e-01 -1.00375569e+00
-2.82612503e-01 -7.22215354e-01 5.74018657e-01 -2.23806649e-02
-7.11940169e-01 -9.14512038e-01 -5.54690473e-02 5.31285465e-01
9.19580162e-01 2.19248440e-02 5.91287136e-01 -1.03942263e+00
-5.22018254e-01 -9.68191445e-01 -5.13860285e-01 3.81802350e-01
4.13624048e-01 -1.09080665e-01 -8.65643144e-01 -1.46387577e-01
3.77068430e-01 2.92161331e-02 7.30310202e-01 -4.66711640e-01
6.94858372e-01 -4.61406767e-01 -3.99399459e-01 -1.38872057e-01
1.12333739e+00 -1.68123394e-02 4.26027894e-01 8.16683590e-01
2.15172574e-01 8.36417019e-01 4.80320781e-01 5.01198709e-01
3.37091871e-02 3.85714382e-01 6.73542142e-01 4.93531793e-01
1.76837280e-01 -1.95459247e-01 3.51890922e-01 9.39161181e-01
1.27903685e-01 -1.00476556e-01 -8.59412432e-01 3.81778181e-01
-1.76255715e+00 -1.19151032e+00 -6.35611832e-01 2.16486883e+00
1.02997708e+00 3.87582093e-01 4.53417063e-01 4.37180281e-01
9.66395557e-01 1.28266364e-01 7.23716840e-02 -2.78115273e-01
-1.45885289e-01 9.10679325e-02 6.18543148e-01 4.58287299e-01
-1.16422379e+00 1.02571690e+00 5.75126410e+00 1.00421286e+00
-1.27213800e+00 1.14055492e-01 3.17451268e-01 2.55241334e-01
1.46122323e-02 4.61543277e-02 -7.93640316e-01 1.08905160e+00
8.98214519e-01 -2.76161760e-01 1.98396832e-01 8.45162034e-01
4.87768263e-01 -5.88939250e-01 -3.24402958e-01 7.77648807e-01
2.52144903e-01 -1.18930531e+00 8.41635466e-02 1.08869903e-01
5.09444177e-01 -6.81628585e-02 -1.23043537e-01 1.52128041e-01
1.64490446e-01 -6.48902118e-01 4.60425317e-01 2.78183013e-01
1.06813155e-01 -5.41286945e-01 9.70743895e-01 5.57797790e-01
-5.50604463e-01 -3.82070392e-01 7.89250508e-02 -3.77789885e-01
3.08249384e-01 9.56879139e-01 -1.22030783e+00 1.20602079e-01
4.16108936e-01 4.91024405e-01 -9.37281251e-01 1.28701818e+00
-2.47983895e-02 8.49603951e-01 -5.82647800e-01 -7.86533415e-01
3.24045897e-01 -4.34785903e-01 9.02621627e-01 1.30347764e+00
1.25480279e-01 -5.26602805e-01 -1.76074699e-01 4.15461689e-01
4.76220772e-02 3.39312106e-01 -7.44248450e-01 -2.79859543e-01
2.89488405e-01 1.58246577e+00 -1.16534638e+00 -4.47960943e-01
-1.32167965e-01 7.62828410e-01 -9.65045393e-02 -1.03860036e-01
-6.37813270e-01 -4.76472139e-01 5.49784005e-01 8.38490129e-01
-4.32360202e-01 -2.69285589e-01 -3.85543108e-01 -1.04730594e+00
-2.35297486e-01 -1.22962165e+00 1.45487010e-01 -5.81637360e-02
-1.64727235e+00 5.44824839e-01 -3.00118536e-01 -1.35062706e+00
-9.00935531e-02 -5.91823101e-01 -4.92495835e-01 5.77909946e-01
-8.94194126e-01 -8.80112290e-01 -3.10954958e-01 1.71020582e-01
3.86142880e-01 -3.86834204e-01 5.08467734e-01 4.85336512e-01
-1.86130285e-01 2.16913626e-01 2.51956254e-01 -1.12441704e-01
8.08132350e-01 -1.11076128e+00 5.61599493e-01 5.60755551e-01
-2.07925364e-01 7.19132721e-01 1.09685600e+00 -9.68872249e-01
-1.05172491e+00 -7.19708920e-01 1.13318110e+00 -8.19210231e-01
1.54857886e+00 -4.60092396e-01 -1.00790763e+00 2.00153351e-01
-1.49459302e-01 -7.44771779e-01 1.91600949e-01 -1.96620762e-01
-3.09424967e-01 -1.60989359e-01 -1.15867937e+00 7.21387386e-01
7.33645499e-01 -3.53354573e-01 -6.27831638e-01 5.09511828e-01
4.90023801e-03 1.97807789e-01 3.90316136e-02 -1.23767756e-01
5.45444667e-01 -1.27614105e+00 4.74049360e-01 -3.40026140e-01
-1.06090799e-01 -2.13502184e-01 3.82567644e-01 -1.04949725e+00
6.50290474e-02 -1.02279556e+00 2.33720973e-01 1.13216090e+00
2.95328736e-01 -1.15770519e+00 1.25807273e+00 1.90546557e-01
1.52972415e-01 -2.91895270e-01 -9.12561119e-01 -7.08026648e-01
-1.80871353e-01 -2.43426070e-01 2.32982129e-01 1.09524679e+00
4.77491736e-01 1.12391621e-01 -3.86691153e-01 -4.92676497e-01
6.22596562e-01 -1.36325404e-01 9.86490250e-01 -1.69233990e+00
1.53806150e-01 -7.76801229e-01 -6.33517921e-01 -7.59172082e-01
-1.69992670e-01 -8.99813831e-01 -2.25040004e-01 -9.13986981e-01
-9.28840339e-02 -3.63456637e-01 4.93502975e-01 9.06768441e-02
1.82406344e-02 5.83877385e-01 -4.04050946e-03 3.78606081e-01
-3.80144686e-01 1.79495104e-02 9.09855008e-01 -9.45985094e-02
-6.94894269e-02 4.76403832e-01 -5.68965733e-01 1.00811100e+00
9.46906447e-01 -4.98863429e-01 1.34558454e-01 4.69830722e-01
7.46181607e-01 -7.88234174e-01 7.21519813e-02 -9.00584996e-01
1.54049322e-01 1.25371277e-01 -1.44016623e-01 -4.20577019e-01
1.95134982e-01 -9.69943821e-01 -8.99731517e-02 8.37993622e-01
3.26585025e-01 -1.70148090e-01 -2.23151609e-01 5.59231341e-01
-2.35260203e-01 -4.56669360e-01 9.98720646e-01 -2.53500342e-01
-3.35591286e-01 -1.14938356e-01 -1.12302685e+00 -4.08008136e-02
9.19780731e-01 -5.65623224e-01 -5.72177947e-01 -6.67110443e-01
-6.17899716e-01 6.06395006e-02 6.33458376e-01 6.80450201e-01
2.08078861e-01 -7.68624961e-01 -4.89797890e-01 2.58948117e-01
-9.60691497e-02 -7.80344665e-01 -1.04418963e-01 9.65317547e-01
-8.76892507e-01 5.49815968e-02 -6.54453486e-02 -3.25662136e-01
-1.23747838e+00 1.86193049e-01 1.00542754e-01 -3.68151069e-02
-4.27944273e-01 4.13949817e-01 -5.44485211e-01 -3.46255273e-01
1.37039289e-01 1.95149273e-01 -2.40694419e-01 7.51880109e-01
5.71792960e-01 9.44166899e-01 2.52159715e-01 -8.09757590e-01
-9.24443454e-02 4.60869558e-02 -2.50523508e-01 1.18866332e-01
1.05215442e+00 -1.31956100e-01 -7.68141925e-01 3.66244495e-01
9.06087875e-01 4.23518419e-01 -2.39869669e-01 1.29417777e-01
6.50785565e-01 -8.48049879e-01 -4.22848552e-01 -5.92868626e-01
-4.96513277e-01 5.13856888e-01 4.07099664e-01 1.41251230e+00
5.36306381e-01 -1.60575092e-01 1.06287324e+00 2.74612367e-01
5.35888851e-01 -1.34925807e+00 1.11626387e-01 7.99724519e-01
4.81038243e-01 -1.41128278e+00 5.41340932e-02 -6.03958011e-01
-1.66881606e-01 1.15427315e+00 4.14758474e-01 -1.01583898e-01
7.00745046e-01 -1.53547134e-02 9.99726728e-02 -1.73534974e-01
-2.28547752e-01 -9.75898430e-02 -2.83125490e-01 3.67073298e-01
3.63441855e-01 -1.40791414e-02 -9.63717997e-01 2.24180102e-01
-3.64633501e-01 -2.70607531e-01 8.84445012e-01 8.81751239e-01
-1.28344727e+00 -1.41075408e+00 -7.06911743e-01 9.86466229e-01
-5.56690931e-01 1.06739923e-01 -9.10090208e-01 8.42258036e-01
-3.49248908e-02 1.08542335e+00 -2.59176582e-01 -3.49692523e-01
2.55570173e-01 2.61487722e-01 -6.65020943e-02 -4.53533471e-01
-1.02357709e+00 -3.98634464e-01 2.36240640e-01 -2.50234306e-01
-2.02407226e-01 -6.61824048e-01 -9.55171883e-01 -9.30501044e-01
-7.55579472e-01 4.25965518e-01 9.12879944e-01 8.46745431e-01
5.88049404e-02 -2.89064497e-01 8.90840709e-01 -8.24012876e-01
-1.04110980e+00 -1.19675744e+00 -6.44409239e-01 7.89428234e-01
6.14043362e-02 -5.56142569e-01 -1.05037534e+00 -3.44167113e-01] | [7.909165859222412, 10.057793617248535] |
7ff75174-e469-4d3c-9c9b-d0ea6974ece8 | grounding-language-representation-with-visual | null | null | https://openreview.net/forum?id=Mdn3eM7VHFn | https://openreview.net/pdf?id=Mdn3eM7VHFn | Grounding Language Representation with Visual Object Information via Cross Modal Pretraining | Previous studies of visual grounded language learning use a convolutional neural network (CNN) to extract features from the whole image for grounding with the sentence description. However, this approach has two main drawbacks: (i) the whole image usually contains more objects and backgrounds than the sentence itself; thus, matching them together will confuse the grounded model; (ii) CNN only extracts the features of the image but not the relationship between objects inside that, limiting the grounded model to learn complicated contexts. To overcome such shortcomings, we propose a novel object-level grounded language learning framework that empowers the language representation with visual object-grounded information. The framework is comprised of three main components: (i) ObjectGroundedBERT captures the visual-object relations and literary portrayals by cross-modal pretraining via a Text-grounding mechanism, (ii) Visual encoder represents a visual relation between objects and (iii) Cross-modal Transformer helps the Visual encoder and ObjectGroundedBERT learn the alignment and representation of image-text context. Experimental results show that our proposed framework consistently outperforms the baseline language models on various language tasks of GLUE and SQuAD datasets. | ['Tho Quan', 'Anh Tuan Luu', 'Cong-Duy T Nguyen'] | 2021-09-29 | null | null | null | null | ['grounded-language-learning'] | ['natural-language-processing'] | [ 4.30092700e-02 -4.01205830e-02 -1.84278682e-01 -5.14937162e-01
-6.08155429e-01 -3.84636283e-01 8.98853302e-01 1.22197539e-01
-4.11469311e-01 3.63435358e-01 4.67348456e-01 -1.18836932e-01
2.77042508e-01 -1.03949821e+00 -1.12437630e+00 -6.26298666e-01
3.42347473e-01 3.57652307e-01 4.17483449e-01 -6.21152699e-01
2.27240130e-01 1.25961542e-01 -1.29040551e+00 1.12386227e+00
5.70149720e-01 1.05869699e+00 6.36765242e-01 2.83985198e-01
-8.57318819e-01 1.57396007e+00 -6.11402631e-01 -6.90224469e-01
-1.35767311e-01 -6.02431118e-01 -8.13106239e-01 3.43089432e-01
5.53222835e-01 -5.13226032e-01 -4.92611110e-01 1.16033149e+00
2.59105384e-01 -4.31086272e-02 6.74460530e-01 -1.47694683e+00
-1.41019404e+00 9.93842185e-01 -6.00521326e-01 2.02749580e-01
2.75371373e-01 1.71226859e-01 1.01593161e+00 -1.35030901e+00
7.23397315e-01 1.62891197e+00 4.44116175e-01 3.38624895e-01
-8.38721931e-01 -5.51034749e-01 6.04322672e-01 1.38321593e-01
-1.42256057e+00 -2.17052549e-01 1.00838220e+00 -5.25319219e-01
9.11216795e-01 -1.39846906e-01 1.03827238e+00 1.29626250e+00
1.37521386e-01 9.78882253e-01 1.13738275e+00 -4.92533296e-01
-4.53559272e-02 3.40405703e-01 4.72155996e-02 1.03351450e+00
-3.15028615e-02 -1.19742453e-01 -8.27037096e-01 3.13421875e-01
7.76192963e-01 2.56917298e-01 -1.32100567e-01 -5.56341410e-01
-1.48139596e+00 8.53005409e-01 9.62536991e-01 4.45513219e-01
-2.08556890e-01 4.19408500e-01 4.39688146e-01 -5.42361289e-02
3.91053706e-01 3.81789589e-03 1.36401638e-01 4.31247652e-01
-8.67838383e-01 2.11423710e-01 3.66444170e-01 1.33896065e+00
1.09599698e+00 1.36761427e-01 -4.42670494e-01 5.38562000e-01
6.06580436e-01 6.80571258e-01 1.60246313e-01 -3.06538165e-01
1.15587127e+00 1.16568065e+00 -2.65148878e-01 -1.47462869e+00
-1.63850665e-01 -1.60058618e-01 -9.37450051e-01 3.55083286e-03
1.50332689e-01 2.16747716e-01 -6.63541198e-01 1.53971672e+00
-6.07978292e-02 -1.22363986e-02 2.44751632e-01 1.08341110e+00
1.76112080e+00 1.00549710e+00 3.48030776e-01 1.61441132e-01
1.64246058e+00 -1.06236303e+00 -8.18482757e-01 -4.61574882e-01
4.20656502e-01 -6.93849862e-01 1.63429797e+00 -1.48343503e-01
-1.24994898e+00 -7.37459362e-01 -1.37028599e+00 -7.54447520e-01
-6.84708655e-01 1.19679369e-01 4.87418383e-01 5.55497892e-02
-9.42559063e-01 -6.68317527e-02 -3.33623260e-01 -3.32767308e-01
5.49417853e-01 -6.65603951e-02 -3.18712234e-01 1.16993422e-02
-1.08838820e+00 7.97684729e-01 7.59093463e-01 2.55169630e-01
-9.36816216e-01 -3.51968199e-01 -1.32539177e+00 1.24381989e-01
3.59857768e-01 -7.95590758e-01 7.91368544e-01 -1.31864357e+00
-9.33657050e-01 1.37647247e+00 1.17202885e-02 -2.53120750e-01
3.32157493e-01 -1.77559167e-01 -2.37945735e-01 3.76487821e-01
1.75708428e-01 8.97147417e-01 6.02355003e-01 -1.57766736e+00
-5.75026751e-01 -2.41754085e-01 5.03175378e-01 3.92499864e-01
-5.38219333e-01 1.26205876e-01 -6.32787824e-01 -7.39137650e-01
1.15559645e-01 -5.15616715e-01 2.61883676e-01 2.71667182e-01
-5.64050972e-01 -9.89424288e-02 1.10738778e+00 -5.05438566e-01
8.82027388e-01 -2.33739924e+00 1.64140671e-01 -3.39952201e-01
4.28117245e-01 -3.95109830e-03 -2.71203905e-01 6.88400388e-01
4.60157096e-02 1.30417511e-01 4.08737808e-02 -3.45442533e-01
-6.90299645e-02 3.13096642e-01 -7.39880323e-01 2.58015007e-01
5.15898526e-01 1.53891110e+00 -1.06062472e+00 -1.07327962e+00
3.86676818e-01 5.49797058e-01 -4.46091115e-01 4.47183162e-01
-2.73537070e-01 2.22533941e-01 -3.24481517e-01 5.26798785e-01
4.96440798e-01 -3.23433608e-01 -8.71102810e-02 -6.33764744e-01
-8.44916999e-02 2.57060379e-01 -9.70434844e-01 1.93414915e+00
-4.55260098e-01 7.64402568e-01 -3.51392090e-01 -1.09648323e+00
1.04453051e+00 2.00777322e-01 6.26498908e-02 -1.06826842e+00
1.35341302e-01 -2.14122802e-01 -3.56539309e-01 -9.33517396e-01
4.77119654e-01 -3.23471725e-01 -2.39848465e-01 3.11627328e-01
3.31664622e-01 -2.41739944e-01 -4.51209210e-02 6.74993157e-01
2.42832810e-01 5.50104976e-01 1.44442558e-01 -2.10591793e-01
5.84081173e-01 1.16029987e-02 1.76563293e-01 4.36222017e-01
-5.73163964e-02 5.52985907e-01 5.81074357e-01 -6.30977809e-01
-9.97389078e-01 -1.18949807e+00 2.92452067e-01 1.27435958e+00
7.34331310e-01 -4.79832172e-01 -4.98744816e-01 -5.67612052e-01
-2.85451055e-01 7.78063893e-01 -7.89111853e-01 -1.41124144e-01
-6.62962914e-01 -1.86777562e-01 3.59935015e-01 8.49737525e-01
9.59180951e-01 -1.40195036e+00 -4.74319786e-01 -2.43640527e-01
-6.96156621e-01 -1.48340189e+00 -6.43841684e-01 1.63370013e-01
-5.20747185e-01 -9.78848398e-01 -4.51213926e-01 -1.39279306e+00
8.68813634e-01 5.32531738e-01 1.47165251e+00 3.44929934e-01
-8.61904100e-02 3.71699005e-01 -3.80784363e-01 -5.54926276e-01
-1.49831221e-01 -1.96562931e-01 -6.05595231e-01 1.40467882e-01
3.89777035e-01 -7.60039389e-02 -5.31732678e-01 -5.01905307e-02
-7.39178360e-01 5.90370953e-01 5.51703930e-01 9.58035827e-01
5.74356496e-01 -1.05392084e-01 1.38512105e-01 -7.90635824e-01
2.74640799e-01 -3.49882632e-01 -2.91550845e-01 5.39431989e-01
1.21870562e-01 -1.41863242e-01 3.52952480e-01 -5.71749449e-01
-1.00418901e+00 -2.47191284e-02 1.77148893e-01 -4.47490871e-01
8.96797851e-02 4.74778742e-01 -4.66120690e-01 2.95360386e-01
4.50472087e-01 4.89546865e-01 -1.93851784e-01 -2.42516249e-02
6.94586635e-01 5.82679868e-01 6.57029986e-01 -8.61796677e-01
8.33553195e-01 5.62450349e-01 -4.81322780e-02 -7.44867921e-01
-1.10741115e+00 -2.23272100e-01 -9.80118454e-01 -4.27914947e-01
1.28331053e+00 -1.11142194e+00 -6.65806472e-01 3.71099934e-02
-1.50114942e+00 -2.16483265e-01 -3.41194838e-01 2.34295502e-01
-7.22321391e-01 1.29555672e-01 -5.69448709e-01 -7.23178625e-01
-3.99187267e-01 -1.17725587e+00 1.34153974e+00 8.82808492e-02
1.21018797e-01 -1.00202870e+00 -5.94635427e-01 2.66462117e-01
-5.55923618e-02 5.62087178e-01 1.30639863e+00 -1.79603606e-01
-7.59351611e-01 1.57998055e-01 -5.42035401e-01 1.01214007e-01
-7.96072185e-02 9.27376598e-02 -8.68322968e-01 -1.41301915e-01
-9.31114033e-02 -8.11376154e-01 6.99741900e-01 2.83957296e-03
1.10451734e+00 -5.35888255e-01 -1.78589121e-01 6.76202238e-01
1.58260012e+00 1.75496116e-02 7.21494019e-01 4.84830528e-01
1.02030087e+00 7.34844148e-01 5.14338434e-01 4.62741870e-03
7.56022394e-01 6.05517030e-01 6.80436492e-01 -7.91621625e-01
-5.43446422e-01 -7.82094955e-01 4.44201797e-01 7.22665906e-01
3.29995416e-02 -1.34281561e-01 -1.01009059e+00 5.58890641e-01
-1.92647779e+00 -1.24537337e+00 -1.89258561e-01 1.63054609e+00
7.89322138e-01 2.01422691e-01 1.97870761e-01 -6.31614849e-02
6.43061101e-01 1.54243931e-01 -2.46494308e-01 -2.69459814e-01
-5.37009597e-01 -2.40540266e-01 -1.07532226e-01 1.70665175e-01
-9.41457450e-01 1.23804796e+00 5.21083689e+00 3.57557565e-01
-1.34327245e+00 1.94728702e-01 6.84496880e-01 5.90267479e-02
-2.89579183e-01 -5.58706932e-02 -7.34877169e-01 1.93127200e-01
1.74805880e-01 -1.57864243e-02 1.17145769e-01 8.52401912e-01
-1.03044026e-01 3.05388808e-01 -1.22252178e+00 1.23637462e+00
4.42436099e-01 -1.48485005e+00 7.35641003e-01 -6.38061836e-02
5.50503910e-01 -3.59664381e-01 1.98314533e-01 4.27938163e-01
1.12779386e-01 -1.09616554e+00 1.59870517e+00 5.83160400e-01
7.86773443e-01 -8.71564686e-01 5.71526229e-01 3.08174014e-01
-1.45813429e+00 -1.28826842e-01 -4.70843762e-01 -5.43128178e-02
3.90503332e-02 3.61728184e-02 -5.78254223e-01 4.82293159e-01
8.29347014e-01 9.82610643e-01 -7.55197942e-01 3.72216016e-01
-2.86701113e-01 2.71728277e-01 4.55385238e-01 -1.96864411e-01
5.92583358e-01 -1.25778452e-01 2.45115265e-01 1.32745969e+00
-2.59192381e-02 5.98454811e-02 3.68476212e-01 1.31900859e+00
9.67244059e-02 3.56719345e-01 -8.42669070e-01 -2.89869159e-01
2.98433006e-01 1.26315725e+00 -8.59936059e-01 -4.44604665e-01
-9.07395124e-01 8.42851102e-01 6.62552357e-01 5.58067024e-01
-8.49530101e-01 -2.66278565e-01 1.71319917e-02 3.31873506e-01
3.37524146e-01 -3.66966754e-01 -2.97647446e-01 -1.29233968e+00
-1.98244583e-02 -8.47621262e-01 4.24657524e-01 -1.45644140e+00
-1.25205517e+00 7.64513791e-01 1.69237539e-01 -1.35339499e+00
1.72181413e-01 -8.21022093e-01 -4.02667642e-01 7.99530029e-01
-1.38425064e+00 -1.98767340e+00 -5.77469468e-01 9.33467269e-01
7.46206820e-01 -2.57318079e-01 6.39058053e-01 1.31616250e-01
-3.61925781e-01 3.96450162e-01 -6.84791625e-01 7.52135217e-01
4.41145718e-01 -1.13359380e+00 1.51661947e-01 8.85432541e-01
5.68449199e-01 9.54197764e-01 2.84883708e-01 -5.14802217e-01
-1.49878490e+00 -1.10195327e+00 9.85347092e-01 -3.72980326e-01
6.80582941e-01 -9.54299271e-01 -7.90713370e-01 8.07401299e-01
6.71467721e-01 8.44214782e-02 5.51960707e-01 8.87050480e-03
-6.99895978e-01 -1.37314245e-01 -7.05669820e-01 9.59596157e-01
1.09004366e+00 -8.78880203e-01 -8.82962525e-01 4.70901757e-01
7.41203785e-01 -5.06554365e-01 -4.80374753e-01 1.55151248e-01
3.49739552e-01 -9.40908790e-01 1.11761463e+00 -6.65466547e-01
9.79775727e-01 -4.29138601e-01 -5.85980058e-01 -7.99658835e-01
-3.45220625e-01 -4.95099612e-02 -1.40737206e-01 1.67163479e+00
9.45121720e-02 -1.42323390e-01 2.49625921e-01 8.85574892e-02
-8.44253600e-02 -8.40977907e-01 -5.75800717e-01 -4.02778089e-01
1.32355720e-01 -2.78241068e-01 6.14408493e-01 1.07787454e+00
-6.07470311e-02 6.83311403e-01 -2.80006051e-01 8.78154412e-02
3.35177362e-01 1.45749778e-01 7.61644363e-01 -7.22976625e-01
2.01420728e-02 -3.50204080e-01 -6.03252530e-01 -1.08744097e+00
3.51061940e-01 -1.08037341e+00 9.08310264e-02 -1.86814952e+00
6.99174464e-01 -2.30498821e-01 -1.93976134e-01 7.61532426e-01
-2.63780206e-01 4.73099023e-01 5.71336508e-01 1.53632686e-01
-8.72209668e-01 5.06052434e-01 1.49810684e+00 -6.38205707e-01
2.01251805e-01 -7.28502393e-01 -7.42049396e-01 7.70011127e-01
2.74705082e-01 -2.30434433e-01 -6.81549311e-01 -8.06276143e-01
5.34649014e-01 -5.31370640e-02 1.00532210e+00 -7.78004766e-01
1.37948498e-01 -2.83856809e-01 8.41080248e-01 -1.11091697e+00
3.17783743e-01 -1.00410473e+00 -3.70582670e-01 3.72070253e-01
-5.24532437e-01 2.97585785e-01 2.20443964e-01 3.98169518e-01
-4.32554752e-01 -3.54682319e-02 7.60390460e-01 -3.81784260e-01
-9.76209164e-01 1.60628974e-01 -1.36479158e-02 2.62157559e-01
8.90549421e-01 -3.23371202e-01 -6.40799582e-01 -3.24585438e-01
-3.34951758e-01 9.58526880e-02 4.06286269e-01 8.17936778e-01
8.64376962e-01 -1.62150192e+00 -5.65730333e-01 5.53032607e-02
5.73060691e-01 1.79353297e-01 1.94419920e-01 5.37286460e-01
-5.52067995e-01 2.75237292e-01 -2.54861951e-01 -1.01008093e+00
-1.22751057e+00 1.04501975e+00 4.72492605e-01 2.78730482e-01
-1.01563561e+00 9.70468342e-01 8.10549855e-01 -1.96150139e-01
2.35401511e-01 -4.39159721e-01 -2.92480201e-01 1.69674382e-01
5.85380077e-01 -3.08710694e-01 -2.73060948e-01 -1.37679958e+00
-3.44656795e-01 8.82797837e-01 -4.34749648e-02 -5.36113009e-02
1.20703995e+00 -2.40677044e-01 -2.82607913e-01 8.89443815e-01
1.28063667e+00 -2.64313430e-01 -1.12490118e+00 -5.41533947e-01
-1.23690955e-01 -3.21287274e-01 -9.26242734e-04 -5.99715114e-01
-9.57177818e-01 1.39845991e+00 3.41839164e-01 -1.45624667e-01
1.09226358e+00 2.54594803e-01 5.95585525e-01 3.02576721e-01
2.83934504e-01 -7.75022984e-01 8.55974376e-01 3.84628862e-01
1.33521450e+00 -1.12875032e+00 1.38300195e-01 -4.79715288e-01
-1.05816329e+00 1.23834813e+00 7.98690200e-01 -8.55197832e-02
3.41800153e-01 1.99417874e-01 2.94096261e-01 -4.84120280e-01
-5.80209196e-01 -3.08951616e-01 7.30294108e-01 6.42362595e-01
5.55347502e-01 -3.57388586e-01 8.20365474e-02 6.64280474e-01
-2.01716110e-01 -3.28870773e-01 9.09221843e-02 8.50970924e-01
-3.26235861e-01 -6.03797019e-01 -2.67875195e-01 -1.82290584e-01
-7.34467357e-02 -3.41342032e-01 -6.91324353e-01 1.09351146e+00
6.15297377e-01 9.15732622e-01 3.97056401e-01 -9.45026651e-02
4.28085119e-01 -8.50583315e-02 5.66510439e-01 -7.32694268e-01
-6.37251198e-01 6.93882629e-02 -2.07676455e-01 -4.44659948e-01
-6.50392890e-01 -1.42672613e-01 -1.55655110e+00 1.64907631e-02
-4.17097099e-02 -1.54152870e-01 5.76591671e-01 1.12259197e+00
-2.33847853e-02 6.74723029e-01 2.59643555e-01 -6.64373815e-01
-3.30574848e-02 -7.30323672e-01 -4.98817235e-01 8.11280727e-01
3.35822701e-01 -7.69870579e-01 6.86969683e-02 4.56544876e-01] | [10.705766677856445, 1.589190125465393] |
2ee30179-1286-459b-89e7-f296313fcf4a | webcaricature-a-benchmark-for-caricature | 1703.03230 | null | http://arxiv.org/abs/1703.03230v4 | http://arxiv.org/pdf/1703.03230v4.pdf | WebCaricature: a benchmark for caricature recognition | Studying caricature recognition is fundamentally important to understanding
of face perception. However, little research has been conducted in the computer
vision community, largely due to the shortage of suitable datasets. In this
paper, a new caricature dataset is built, with the objective to facilitate
research in caricature recognition. All the caricatures and face images were
collected from the Web. Compared with two existing datasets, this dataset is
much more challenging, with a much greater number of available images, artistic
styles and larger intra-personal variations. Evaluation protocols are also
offered together with their baseline performances on the dataset to allow fair
comparisons. Besides, a framework for caricature face recognition is presented
to make a thorough analyze of the challenges of caricature recognition. By
analyzing the challenges, the goal is to show problems that worth to be further
investigated. Additionally, based on the evaluation protocols and the
framework, baseline performances of various state-of-the-art algorithms are
provided. A conclusion is that there is still a large space for performance
improvement and the analyzed problems still need further investigation. | ['Yang Gao', 'Wenbin Li', 'Jing Huo', 'Yinghuan Shi', 'Hujun Yin'] | 2017-03-09 | null | null | null | null | ['caricature'] | ['computer-vision'] | [ 2.53344089e-01 -3.57550383e-01 -2.40397394e-01 -6.24826133e-01
-4.53765213e-01 -5.05723774e-01 7.01141298e-01 -5.70222020e-01
2.05999210e-01 4.19894904e-01 -1.52060285e-01 1.10325269e-01
-1.85910061e-01 -4.50579852e-01 -5.70994258e-01 -8.94950926e-01
-7.26882294e-02 1.77048668e-01 -3.72575462e-01 -1.12462848e-01
4.21236932e-01 1.05595529e+00 -2.12195683e+00 5.92531621e-01
4.03824270e-01 1.30176842e+00 2.21571952e-01 3.81733209e-01
-1.40254036e-01 5.97350121e-01 -7.86213517e-01 -7.82171667e-01
2.27793589e-01 -2.70038903e-01 -6.56873703e-01 6.60347641e-01
1.24112737e+00 -2.30740830e-01 -2.87004918e-01 9.57543492e-01
4.79537100e-01 -4.19425778e-02 6.66527569e-01 -1.45773697e+00
-1.09674561e+00 2.41146952e-01 -5.28557539e-01 1.94964614e-02
2.68901765e-01 -1.28948689e-01 6.93338275e-01 -1.26057661e+00
4.12692279e-01 1.62743926e+00 3.88143182e-01 8.51859331e-01
-8.49125385e-01 -8.43947709e-01 2.92948902e-01 5.47502935e-01
-1.67128468e+00 -7.85937607e-01 1.11650455e+00 -2.40682334e-01
6.20533764e-01 5.51303804e-01 4.38929647e-01 1.30393207e+00
-2.87651092e-01 8.56649339e-01 1.24293482e+00 -7.08658278e-01
1.06911317e-01 5.87251723e-01 1.96402848e-01 6.41777813e-01
1.26586799e-02 8.58912915e-02 -3.96890342e-01 7.20821097e-02
7.38697469e-01 7.18614534e-02 -2.78269887e-01 -5.42946815e-01
-6.71319425e-01 6.84448540e-01 2.94466525e-01 5.92698514e-01
9.34534222e-02 -6.98249266e-02 7.23551512e-02 2.70508140e-01
2.13481680e-01 6.94944337e-02 -1.48864537e-01 1.20863304e-01
-1.06150925e+00 2.18362257e-01 7.03600347e-01 1.02841544e+00
4.85223591e-01 2.53725380e-01 2.88810078e-02 1.21529722e+00
2.89395064e-01 6.15833938e-01 2.28457555e-01 -7.28762686e-01
2.53098100e-01 4.34073150e-01 -2.03984261e-01 -1.33305764e+00
8.21316689e-02 -1.96668610e-01 -7.76680708e-01 3.56047809e-01
5.26546061e-01 4.11817640e-01 -7.43750989e-01 1.21627510e+00
-1.96279567e-02 5.02099454e-01 1.25866067e-02 8.91681671e-01
1.05918932e+00 4.94053453e-01 -1.36845887e-01 -1.50521532e-01
1.73605192e+00 -9.45458531e-01 -1.00751805e+00 -1.22311316e-01
4.17945646e-02 -1.14803946e+00 1.11680734e+00 5.53196669e-01
-9.21062469e-01 -7.49410927e-01 -1.06157267e+00 5.34422025e-02
-5.68177104e-01 7.58599818e-01 8.26477528e-01 1.28199053e+00
-1.03090584e+00 2.77609617e-01 -3.05719793e-01 -4.99582887e-01
8.68472159e-01 2.63530552e-01 -4.90303367e-01 -3.39042217e-01
-6.37809873e-01 9.61095273e-01 2.17034109e-02 3.54134589e-01
-9.36601460e-01 -6.25039339e-01 -7.66433418e-01 -1.76960260e-01
5.15423656e-01 1.82488877e-02 9.82428491e-01 -1.28533459e+00
-1.44779372e+00 1.22141409e+00 -1.82556495e-01 -1.78479150e-01
3.54864389e-01 8.52897093e-02 -1.00675154e+00 2.47358590e-01
-3.87486398e-01 7.24860787e-01 1.29257286e+00 -1.54974222e+00
-5.24790287e-01 -6.63419843e-01 3.33061367e-02 -2.13166580e-01
-4.85948384e-01 3.48221987e-01 -1.04639328e+00 -8.68722975e-01
-3.90223384e-01 -1.12265253e+00 4.65893477e-01 2.14061752e-01
-1.13837518e-01 -3.79923642e-01 1.22088921e+00 -4.31255847e-01
1.20270169e+00 -2.48265386e+00 -2.61405464e-02 1.32612810e-01
-1.91223696e-01 4.01144445e-01 -3.60066324e-01 3.12472522e-01
-2.78403252e-01 1.48473144e-01 -1.60838753e-01 -5.08593082e-01
-9.50901061e-02 2.52757400e-01 -5.83084345e-01 4.09441382e-01
4.66808170e-01 6.88539386e-01 -3.83239716e-01 -4.74103183e-01
3.12832117e-01 7.50780582e-01 -3.05026233e-01 9.77678895e-02
5.11504523e-02 1.15769431e-01 -3.21363539e-01 1.41874576e+00
1.08782935e+00 1.35257676e-01 2.29257680e-02 -3.99075717e-01
2.44642980e-02 -4.12502885e-01 -1.19481254e+00 1.48649168e+00
-3.89863521e-01 7.91149676e-01 3.79986346e-01 -1.10677493e+00
1.06833816e+00 3.62762332e-01 1.65206879e-01 -3.36132854e-01
2.17070594e-01 2.62314916e-01 -9.36597809e-02 -6.04855061e-01
3.04621130e-01 4.32376079e-02 2.34904721e-01 1.13862365e-01
-4.61195735e-03 -2.64402241e-01 2.89645880e-01 -1.18581519e-01
3.27762127e-01 1.01515450e-01 -4.13033441e-02 -3.63512397e-01
1.06524825e+00 -4.08872403e-02 6.49175942e-02 3.57804984e-01
-4.23596740e-01 7.24548221e-01 1.57705650e-01 -6.96344376e-01
-6.35819256e-01 -8.98831010e-01 -5.53247094e-01 7.56228626e-01
2.73856167e-02 -3.86902332e-01 -1.02683902e+00 -6.74500465e-01
1.36779800e-01 4.70991999e-01 -9.13161576e-01 2.11010531e-01
-5.15556812e-01 -7.00520635e-01 5.30764461e-01 4.34829652e-01
7.33069777e-01 -1.28680861e+00 -1.79912075e-01 -4.45656270e-01
-4.92557734e-02 -1.15776777e+00 -5.24320424e-01 -5.84796369e-01
-8.10434043e-01 -1.23616004e+00 -6.36551559e-01 -1.05184197e+00
7.98178136e-01 8.12035143e-01 1.16137373e+00 4.36681598e-01
-6.37719750e-01 7.77466238e-01 -2.07845926e-01 -6.00828290e-01
-2.21982092e-01 -2.74114043e-01 7.09389802e-03 4.39174682e-01
7.89643109e-01 -2.61760235e-01 -3.69108915e-01 5.62401474e-01
-9.37819958e-01 -4.48564112e-01 6.62967265e-01 7.26757944e-01
4.54819560e-01 1.30857557e-01 5.08497775e-01 -7.51914382e-01
5.19674301e-01 -3.15195382e-01 -6.02253616e-01 5.78690588e-01
-3.43676776e-01 -3.32540751e-01 3.98292989e-01 -6.04281902e-01
-1.14201725e+00 7.28184581e-02 -1.20590609e-02 -6.51615202e-01
-4.50655699e-01 -3.78693826e-02 -4.73830402e-01 -3.33865374e-01
4.43685651e-01 3.03165168e-01 4.70179141e-01 -7.18542397e-01
3.35958332e-01 8.23966324e-01 6.85718060e-01 -5.98336160e-01
6.57564044e-01 6.66142583e-01 -2.22639218e-01 -1.32959902e+00
-5.23504734e-01 -1.65868118e-01 -6.01342022e-01 -3.78497481e-01
5.89137971e-01 -8.27169597e-01 -9.37728643e-01 6.03143930e-01
-8.78158092e-01 2.49978945e-01 9.46941450e-02 1.82291061e-01
-3.14247310e-01 4.65222150e-01 -2.69772977e-01 -1.04148173e+00
-1.26421571e-01 -1.30061877e+00 9.63797629e-01 1.78906709e-01
1.75518826e-01 -9.05032933e-01 -4.35862571e-01 8.55486751e-01
4.08449382e-01 3.78124192e-02 6.97266996e-01 -2.66921133e-01
-7.67239392e-01 -3.11391205e-01 -1.44614965e-01 5.81496835e-01
4.12267387e-01 5.10808170e-01 -1.35806632e+00 -4.28877532e-01
1.31327406e-01 -3.90523612e-01 7.54496276e-01 1.78158835e-01
1.55245137e+00 -9.32571664e-02 -3.01057518e-01 5.22211075e-01
1.26005161e+00 3.67730767e-01 1.02768385e+00 1.07764013e-01
4.76286203e-01 1.11972785e+00 5.03020763e-01 2.10195899e-01
5.49428500e-02 8.86525333e-01 5.11476040e-01 1.78363677e-02
-4.27190214e-01 -3.38646658e-02 3.39446127e-01 4.58645880e-01
-2.93126762e-01 -1.41470805e-02 -5.00415683e-01 3.61422390e-01
-1.21284187e+00 -1.21039295e+00 -4.36078310e-02 2.06504679e+00
4.59233612e-01 -3.66166621e-01 3.05000752e-01 4.12008047e-01
7.76607871e-01 -4.21512732e-03 -2.46900171e-01 -3.45943511e-01
-3.81801307e-01 2.43824437e-01 4.09464585e-03 1.87139332e-01
-1.16501737e+00 8.57221186e-01 6.78901434e+00 1.19261479e+00
-1.32572627e+00 -2.55544245e-01 7.19076037e-01 7.96713084e-02
1.29606113e-01 -2.56755501e-01 -1.09342885e+00 5.12446105e-01
4.78537977e-01 3.31150085e-01 7.08745956e-01 1.03866625e+00
-1.27569571e-01 8.49499404e-02 -1.19265389e+00 1.54770637e+00
8.14469934e-01 -1.32636654e+00 2.02882916e-01 5.07207550e-02
5.57787657e-01 -4.96694744e-01 4.73541737e-01 1.87122568e-01
-3.64673227e-01 -1.33714068e+00 6.61144853e-01 3.82463872e-01
8.37806225e-01 -7.66855538e-01 5.34574330e-01 3.27869914e-02
-1.08698916e+00 -1.62790388e-01 -5.64647675e-01 7.86055326e-02
-4.35228318e-01 1.09697923e-01 -4.07716453e-01 5.38080812e-01
9.79209185e-01 7.30870962e-01 -8.88142228e-01 7.80311108e-01
3.63812894e-02 4.38934952e-01 -1.11811183e-01 -1.58674091e-01
1.60106681e-02 -4.27904248e-01 4.09263037e-02 1.13478267e+00
3.39337468e-01 3.57429078e-03 -2.30068658e-02 7.82243311e-01
-1.67733595e-01 3.78332466e-01 -6.71328962e-01 -1.24741189e-01
6.29482329e-01 1.49197066e+00 -4.78132635e-01 -2.46855363e-01
-7.93404758e-01 7.78576255e-01 2.19616801e-01 1.81147888e-01
-6.04698896e-01 -4.09445390e-02 9.14657831e-01 1.23106670e-02
2.44816989e-01 4.02092636e-02 -4.36880440e-02 -1.13844645e+00
4.38041501e-02 -1.32883859e+00 4.66441602e-01 -6.11471951e-01
-1.49481595e+00 6.94252908e-01 1.57848045e-01 -1.15857852e+00
-2.00318117e-02 -9.91849780e-01 -4.86960828e-01 7.10816979e-01
-1.62533975e+00 -1.46900594e+00 -6.04669988e-01 8.57518315e-01
8.96676958e-01 -8.27179074e-01 9.72202361e-01 4.68981802e-01
-8.91727924e-01 9.53361630e-01 -5.96037321e-03 8.96694064e-02
7.00425982e-01 -7.77654529e-01 1.03807583e-01 6.92656338e-01
5.07743299e-01 6.65405571e-01 5.15766740e-01 -2.43544415e-01
-1.99701023e+00 -9.11096454e-01 5.86893559e-01 -7.88003564e-01
3.99970621e-01 -4.64723080e-01 -9.59855437e-01 4.81138676e-01
2.64636368e-01 1.82630286e-01 7.53859818e-01 -1.51817262e-01
-4.25629497e-01 -3.58355671e-01 -1.38967514e+00 5.88254929e-01
1.01181078e+00 -3.53829950e-01 -4.61733282e-01 1.84442222e-01
-1.84443891e-01 -6.81777671e-02 -6.48204207e-01 2.53065437e-01
7.58281112e-01 -8.67222071e-01 1.31824124e+00 -2.80515939e-01
2.91061461e-01 -2.80839503e-01 -3.74211460e-01 -8.78247261e-01
-2.45595858e-01 -3.48190159e-01 -2.50220925e-01 1.59704292e+00
6.43642694e-02 -4.58302259e-01 9.95228112e-01 4.13884431e-01
1.38768375e-01 -6.09270632e-01 -7.47825027e-01 -1.05486333e+00
9.58930403e-02 -4.84944522e-01 7.11197257e-01 1.00918496e+00
-3.38764459e-01 1.88556388e-01 -7.06781685e-01 8.71423259e-03
7.00322092e-01 3.37144345e-01 1.01793504e+00 -1.17553627e+00
7.35800937e-02 -5.66193759e-01 -4.19771284e-01 -9.06220317e-01
3.16435158e-01 -9.06380177e-01 -4.14371848e-01 -8.47345769e-01
1.92881718e-01 -3.26969594e-01 -1.25143126e-01 3.75008285e-01
-2.32850807e-03 7.54878759e-01 4.77177352e-01 3.38150024e-01
-5.36707602e-02 3.87315363e-01 1.51009440e+00 -4.76440787e-01
1.53927326e-01 -5.40925190e-02 -7.80951917e-01 6.28843546e-01
7.48046577e-01 3.93046513e-02 -5.21318555e-01 -3.70430380e-01
-7.06859708e-01 -2.89660543e-01 3.87312531e-01 -8.38519633e-01
-3.00068650e-02 -1.18781559e-01 6.70086086e-01 -8.15367877e-01
8.37291121e-01 -1.17645180e+00 1.84199154e-01 1.03296272e-01
-8.60958695e-02 4.93039284e-03 3.90861899e-01 3.92474324e-01
-3.82543415e-01 -3.47759634e-01 9.53845561e-01 -5.93540072e-02
-9.71921027e-01 2.52214998e-01 -1.63559452e-01 -2.50398576e-01
1.26424706e+00 -4.89180744e-01 -1.04102358e-01 -1.09924994e-01
-5.31104684e-01 -1.14523381e-01 4.17094111e-01 9.43981409e-01
7.96897352e-01 -1.60679674e+00 -8.03018987e-01 5.74651837e-01
4.17040676e-01 -7.08464086e-01 5.30481756e-01 4.79496866e-01
-5.66201925e-01 6.97785139e-01 -5.57469845e-01 -5.87639570e-01
-1.80446291e+00 9.43546772e-01 2.51124531e-01 4.60848898e-01
-3.96343142e-01 7.33016014e-01 2.43705302e-01 -1.62468895e-01
5.15753627e-01 -5.54761067e-02 -6.30528986e-01 1.13004930e-01
1.06056023e+00 3.13613325e-01 1.44160166e-01 -1.06243098e+00
-5.07115901e-01 8.77590358e-01 -2.24947259e-01 4.08422321e-01
1.20606840e+00 3.07717118e-02 -2.30171010e-01 4.13494557e-02
1.20796525e+00 1.41402960e-01 -1.09643161e+00 -1.65723160e-01
-6.27489686e-02 -1.10211992e+00 -6.41414374e-02 -7.63639629e-01
-1.44181514e+00 9.99528706e-01 8.60500693e-01 6.85271621e-02
1.34474492e+00 -4.98881489e-02 2.55752295e-01 3.54224473e-01
4.02867556e-01 -1.05605114e+00 2.89164543e-01 1.06785677e-01
1.55182469e+00 -1.38367307e+00 -5.02477773e-02 -9.99091744e-01
-6.40854180e-01 1.36616910e+00 5.37895858e-01 -1.53119974e-02
7.32139766e-01 2.16187119e-01 1.82556525e-01 -1.51895881e-01
-5.08221388e-01 -1.21056773e-02 4.08322304e-01 8.42348218e-01
5.20066500e-01 -1.39530785e-02 -1.25272915e-01 4.87494111e-01
-2.59676754e-01 -3.95599902e-02 3.08565468e-01 6.53276563e-01
-7.40665793e-02 -1.45760810e+00 -9.39692855e-01 2.07195982e-01
-7.12714911e-01 1.46378204e-01 -7.32690513e-01 1.07974160e+00
1.08674608e-01 1.29896748e+00 2.26681046e-02 -1.45747438e-01
4.10606384e-01 -7.34118447e-02 9.00085926e-01 -3.31326842e-01
-3.52859646e-01 -1.10908523e-01 -9.75721180e-02 -3.27345848e-01
-5.31823456e-01 -6.88734233e-01 -6.11936510e-01 -6.66246831e-01
-2.76424378e-01 4.22633399e-04 9.20405865e-01 5.33972442e-01
1.98569283e-01 2.68409729e-01 6.25313401e-01 -8.92782629e-01
-4.10399497e-01 -9.20819402e-01 -6.54423118e-01 6.47877336e-01
1.42523190e-02 -9.15749967e-01 -3.36138874e-01 3.41100484e-01] | [13.212163925170898, 0.7032554745674133] |
5bdc8eec-5158-4551-a7ff-46817ee0f64e | augmentation-techniques-analysis-with-removal | null | null | https://www.researchgate.net/profile/Allena-Venkata-Sai-Abhishek-2/publication/364345327_Augmentation_Techniques_Analysis_with_Removal_of_Class_Imbalance_Using_PyTorch_for_Intel_Scene_Dataset/links/634d3f1612cbac6a3ed4a645/Augmentation-Techniques-Analysis-with-Removal-of-Class-Imbalance-Using-PyTorch-for-Intel-Scene-Dataset.pdf | https://www.researchgate.net/profile/Allena-Venkata-Sai-Abhishek-2/publication/364345327_Augmentation_Techniques_Analysis_with_Removal_of_Class_Imbalance_Using_PyTorch_for_Intel_Scene_Dataset/links/634d3f1612cbac6a3ed4a645/Augmentation-Techniques-Analysis-with-Removal-of-Class-Imbalance-Using-PyTorch-for-Intel-Scene-Dataset.pdf | Augmentation Techniques Analysis with Removal of Class Imbalance Using PyTorch for Intel Scene Dataset | although best-in-class AI can deliver extraordinary outcomes in experimentation, data scientists struggle to duplicate these outcomes on actual-world data. It's nothing unexpected-actual data mirrors the messy world that made it, containing many biases and gaps. A painful element of real data is that it tends to be imbalanced. An imbalanced dataset is a dataset with a lot more examples in one class than others. This exploration features the broad study about taking care of class Imbalance issues utilizing Random Sampling and Data Augmentation Techniques. The critical angle featured is to grasp how Under-Sampling, Over-Sampling, and Data Augmentation use images and custom datasets. The model performance is improved with the expulsion of class imbalance issue utilizing different Augmentation approaches utilizing an augmentation library. The accuracy contrasts with taking care of the Class imbalance issue to boost accuracy, lessen error, and track down an ideal technique to tackle it. | ['Anusha Kanukolanu', 'Dr. S Phani Kumar', 'Allena Venkata Sai Abhishek'] | 2022-05-01 | null | null | null | international-journal-of-innovative-research | ['image-manipulation-detection', 'image-smoothing', 'image-matting', 'image-variation', 'image-stitching', 'image-augmentation', 'image-manipulation', 'image-morphing', 'roi-based-image-generation', 'image-cropping', 'detecting-image-manipulation', 'data-visualization', 'data-visualization'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'methodology', 'miscellaneous'] | [ 3.08126658e-01 2.74351060e-01 -4.30359632e-01 -5.60810685e-01
-4.18189466e-01 -3.14288735e-01 3.08941990e-01 2.92902261e-01
-2.77057678e-01 7.92617023e-01 3.38112146e-01 -2.29166433e-01
-6.73996508e-02 -9.31193888e-01 -6.51805878e-01 -5.26411653e-01
8.04257840e-02 7.14482069e-01 -2.04935834e-01 -6.04123175e-01
6.28909528e-01 2.66730189e-01 -1.80009139e+00 4.44723070e-01
6.77641511e-01 7.28915036e-01 -5.47621429e-01 3.37217152e-01
-4.31849569e-01 8.44932854e-01 -1.04076552e+00 -3.57312053e-01
6.26852334e-01 -4.12963420e-01 -7.50392199e-01 -4.62495576e-04
6.63936198e-01 -4.65373360e-02 9.29033384e-02 9.40784395e-01
5.02467096e-01 -3.87122571e-01 3.26187760e-01 -1.84920585e+00
-5.59582591e-01 3.89742315e-01 -9.76670921e-01 4.43796098e-01
3.04610223e-01 4.52834368e-01 6.89225554e-01 -5.19040823e-01
6.19053721e-01 1.15547931e+00 1.11893666e+00 3.84768873e-01
-1.32100654e+00 -9.36090171e-01 1.35108158e-01 -2.22064659e-01
-1.06397748e+00 -2.89937228e-01 5.70188999e-01 -6.60861969e-01
7.95557737e-01 5.73093176e-01 1.01436317e+00 9.46618557e-01
9.89917070e-02 4.95246023e-01 1.13773036e+00 -3.35513979e-01
1.69132054e-01 4.02160972e-01 2.49185801e-01 1.18829735e-01
8.12172830e-01 6.26366809e-02 -6.71090066e-01 -1.44664105e-02
2.74190634e-01 2.82705188e-01 -7.16521889e-02 -3.28841656e-01
-1.13117909e+00 7.28395402e-01 3.83474976e-01 2.55877674e-01
-1.41950861e-01 -2.16492593e-01 6.79286778e-01 5.07673502e-01
6.37277603e-01 1.13742971e+00 -7.04383790e-01 -3.13456953e-01
-9.67551649e-01 5.29678047e-01 5.34100890e-01 6.92577600e-01
8.53173733e-01 1.05367444e-01 1.69500142e-01 7.60265589e-01
-2.46908903e-01 2.58100927e-01 8.21439147e-01 -6.94828510e-01
4.93293345e-01 1.40357018e+00 -8.05698261e-02 -1.29371357e+00
-6.45691276e-01 -6.18532538e-01 -9.11859810e-01 5.91761589e-01
7.05935121e-01 -2.29525438e-04 -1.11137581e+00 1.80193055e+00
4.00891244e-01 -3.16993743e-01 -2.16165349e-01 8.16471279e-01
3.95545334e-01 3.58252466e-01 1.07865922e-01 3.74850221e-02
1.15671217e+00 -4.15878534e-01 -8.98638189e-01 -6.37463093e-01
9.50084746e-01 -6.53754115e-01 1.39696109e+00 6.14399374e-01
-8.04426253e-01 -3.96394670e-01 -1.41503274e+00 -1.30111426e-01
-6.95425451e-01 -2.88851798e-01 1.04144239e+00 9.71695840e-01
-5.69141805e-01 5.00824034e-01 -3.85610849e-01 -2.51243830e-01
8.83938670e-01 4.03148472e-01 -6.73497140e-01 -3.64995487e-02
-8.29613626e-01 1.04469955e+00 3.48683059e-01 -2.93681741e-01
-3.18436056e-01 -1.35527515e+00 -6.77517772e-01 -2.05361471e-01
3.60329479e-01 -5.13309121e-01 8.53936851e-01 -1.36984503e+00
-4.97659892e-01 1.07673609e+00 8.10523257e-02 -3.28487873e-01
5.49896181e-01 -2.08476454e-01 -4.35101539e-01 -6.51790917e-01
1.75562665e-01 5.45053244e-01 3.94873321e-01 -1.26406538e+00
-4.28933442e-01 -9.00461972e-01 -3.12008202e-01 4.36446443e-02
-2.60869563e-01 -2.28134975e-01 4.33316082e-01 -8.07087243e-01
4.18374062e-01 -5.84095716e-01 -2.52404720e-01 -2.88904719e-02
-2.86723226e-01 2.03147352e-01 9.63248014e-01 -2.88468003e-01
1.19650638e+00 -2.09105515e+00 -4.75831091e-01 -2.77737528e-02
5.14458299e-01 3.28174114e-01 -2.80400757e-02 4.90664721e-01
-4.81116176e-01 6.41805410e-01 -1.41113505e-01 1.65519491e-01
-2.15064913e-01 1.54267728e-01 -3.43339235e-01 3.83964241e-01
6.27679110e-01 5.64493835e-01 -7.27549791e-01 -1.88844025e-01
-1.04619823e-01 2.75440812e-01 -9.07550752e-01 -1.12350412e-01
-3.68919745e-02 2.54236758e-01 -4.40175086e-02 8.83040130e-01
8.50523710e-01 -2.14280531e-01 7.82417506e-02 -1.74654603e-01
-8.80602449e-02 3.09941351e-01 -1.31165338e+00 1.35926247e+00
7.17295259e-02 7.68893719e-01 -3.00713956e-01 -1.14236498e+00
1.18414509e+00 5.68014048e-02 3.78747731e-01 -9.69086587e-01
2.54294485e-01 3.22917551e-01 4.34290558e-01 -7.22694933e-01
5.88788748e-01 -3.81961524e-01 3.30172405e-02 5.06659269e-01
-1.49470538e-01 -4.29660499e-01 7.42347091e-02 8.48697051e-02
1.09508264e+00 -2.30929907e-02 3.21291953e-01 -3.17258388e-01
-2.54226714e-01 4.86794561e-01 9.17321622e-01 5.55066884e-01
-1.57604933e-01 8.29272270e-01 7.23763883e-01 -8.93606544e-01
-1.41009903e+00 -6.11060739e-01 -1.66675463e-01 9.24121976e-01
-2.82845974e-01 -3.15961748e-01 -4.83709872e-01 -3.93397719e-01
3.32431972e-01 3.69892716e-01 -9.90320146e-01 -4.29886878e-01
-2.09935695e-01 -1.37966681e+00 4.04757470e-01 4.07117039e-01
3.28537077e-01 -8.45893383e-01 -7.03042805e-01 6.20865524e-02
1.27694933e-02 -6.52456164e-01 8.16990808e-02 3.38397682e-01
-1.09620655e+00 -1.13325930e+00 -1.79682031e-01 -3.99063438e-01
7.79916883e-01 2.66695656e-02 1.54934692e+00 2.63991743e-01
-7.84739494e-01 -3.23386550e-01 -1.74303204e-01 -1.29834998e+00
-4.26029652e-01 3.60096693e-02 1.18519872e-01 -3.28042507e-01
1.02232099e+00 -7.87936211e-01 -5.46936393e-01 2.20494926e-01
-9.69606102e-01 2.23511867e-02 4.78355706e-01 8.41716349e-01
2.81888962e-01 7.84646869e-02 9.58389521e-01 -1.27951968e+00
4.74902511e-01 -6.51715636e-01 -1.17998987e-01 -3.61700319e-02
-9.49261606e-01 -2.75869697e-01 3.50057304e-01 -5.03887534e-01
-4.05871034e-01 -2.94234633e-01 1.82293132e-01 7.11690933e-02
-8.88208672e-02 8.56745541e-02 -2.62980521e-01 3.76858950e-01
1.11396849e+00 -4.46626097e-01 4.43361372e-01 -5.16458333e-01
-9.98238847e-02 8.38887870e-01 2.36049086e-01 -4.68966216e-01
5.68190694e-01 5.59166670e-01 -2.41240293e-01 -6.28500998e-01
-1.07448745e+00 -6.75875619e-02 -6.45561099e-01 1.78017318e-02
3.38923961e-01 -8.43797266e-01 -6.61508143e-01 5.14805138e-01
-6.68646753e-01 -1.43919185e-01 -9.07982409e-01 2.15387464e-01
-2.56927946e-04 -3.10161322e-01 4.51520905e-02 -8.47940922e-01
-9.05620977e-02 -9.36760426e-01 6.44263029e-01 3.61129582e-01
-6.13410175e-01 -4.87559408e-01 1.43024281e-01 6.03333294e-01
6.53965712e-01 5.48074901e-01 8.94884109e-01 -1.16537964e+00
-3.64480495e-01 -6.47025704e-01 -3.07319939e-01 4.63322043e-01
4.44837302e-01 6.56036660e-02 -1.24074745e+00 -1.90638781e-01
1.44846775e-02 -4.74257171e-01 4.29414809e-01 2.75523383e-02
1.36252904e+00 -4.25234675e-01 -1.33548439e-01 3.21650773e-01
1.59463739e+00 3.52505535e-01 7.75472403e-01 6.22403502e-01
4.92816538e-01 7.64451742e-01 6.53545499e-01 4.71797913e-01
3.16536546e-01 2.87335485e-01 4.03281301e-01 -4.55396235e-01
-1.10141672e-01 -2.50445724e-01 -3.07365656e-01 4.01829213e-01
2.34618947e-01 3.25722881e-02 -1.36805570e+00 7.71319509e-01
-1.41236258e+00 -8.99873674e-01 -1.60716072e-01 2.27021575e+00
1.14987433e+00 5.14152467e-01 2.97011077e-01 8.87001932e-01
3.74971449e-01 -9.03424993e-02 -6.59819305e-01 -6.94529116e-01
-2.43647218e-01 1.60341002e-02 3.13308656e-01 2.28745133e-01
-7.93057024e-01 3.99192363e-01 7.16113615e+00 3.34273636e-01
-1.21774256e+00 -2.95034677e-01 1.21845698e+00 -2.71167845e-01
-3.61693412e-01 -4.41590622e-02 -6.63545430e-01 5.85335374e-01
6.63266480e-01 -1.62044302e-01 1.43011570e-01 7.84360945e-01
-4.14777808e-02 -4.40442830e-01 -1.17976856e+00 6.76147938e-01
1.94171637e-01 -1.30106616e+00 7.31292069e-02 7.63767362e-02
6.90129220e-01 -4.14454676e-02 9.13544521e-02 1.63898438e-01
3.36234003e-01 -1.48401678e+00 4.16604787e-01 3.02547067e-01
5.37597418e-01 -7.20250010e-01 9.20102537e-01 2.57717431e-01
-4.44602147e-02 -2.31344983e-01 -2.12754041e-01 -8.95649076e-01
-4.54366535e-01 1.05142975e+00 -1.04612255e+00 7.11381622e-03
1.02173984e+00 4.00770336e-01 -8.37008953e-01 1.07327986e+00
4.98806149e-01 3.84473443e-01 -3.77964795e-01 1.21529676e-01
-1.38772428e-01 1.62218027e-02 3.20732862e-01 9.89627182e-01
2.74607912e-02 -1.19961157e-01 -1.73944794e-02 4.56338018e-01
8.48527476e-02 1.11725852e-01 -1.01355839e+00 -2.47812286e-01
5.79883039e-01 9.92199183e-01 -5.27056277e-01 -3.38655531e-01
-1.41182199e-01 2.41501912e-01 1.65405553e-02 -2.04753265e-01
-3.30076247e-01 -3.75592053e-01 8.40176046e-01 4.89296585e-01
-2.96689600e-01 2.98223734e-01 -1.19758499e+00 -7.57934511e-01
1.16324410e-01 -1.55574346e+00 3.30337137e-01 -7.36507833e-01
-1.44481492e+00 2.77346820e-01 -3.82816494e-01 -1.31449771e+00
1.30466089e-01 -4.07884359e-01 -2.76452601e-01 6.43620968e-01
-1.11182451e+00 -8.61812651e-01 -5.75701952e-01 -1.46138817e-01
2.73597479e-01 -1.49833456e-01 8.05917978e-01 6.02009594e-01
-5.24312615e-01 5.91757655e-01 -2.55679816e-01 2.09826857e-01
8.07333112e-01 -1.17436373e+00 2.98991501e-01 3.63511056e-01
-1.49529546e-01 5.50518095e-01 9.42778528e-01 -6.11992776e-01
-1.00078058e+00 -6.20253921e-01 6.70232534e-01 -9.02374864e-01
4.00973886e-01 -2.96215534e-01 -1.18925333e+00 7.68856168e-01
-7.20314607e-02 -1.98274497e-02 9.54051852e-01 3.52872282e-01
-6.97543025e-01 -4.22240824e-01 -1.59049475e+00 4.32782143e-01
8.87159348e-01 -1.11189403e-01 -5.90212226e-01 2.25576580e-01
6.29647553e-01 -4.26725805e-01 -7.26557672e-01 6.87374830e-01
8.02090824e-01 -1.08993733e+00 8.60867023e-01 -1.15357184e+00
8.22967887e-01 -1.73071519e-01 -1.22424513e-01 -1.38596833e+00
1.43987462e-01 -2.05597758e-01 5.94136238e-01 1.42020905e+00
5.26007533e-01 -6.54887617e-01 1.07934976e+00 8.15379143e-01
1.50431320e-01 -9.30662155e-01 -5.60369670e-01 -5.87204635e-01
2.90298283e-01 -3.02048326e-01 1.00560272e+00 1.57171702e+00
-1.85730755e-02 1.89635068e-01 -2.51601636e-01 -1.93765938e-01
5.25811672e-01 -1.25218257e-01 1.17941296e+00 -1.46926594e+00
1.85026720e-01 -2.43303150e-01 -7.13804841e-01 -3.11808307e-02
-6.96366668e-01 -6.29775524e-01 -2.95278043e-01 -1.15265024e+00
2.73526192e-01 -9.60988522e-01 -7.53910169e-02 6.16647601e-01
-2.93637067e-01 6.48149073e-01 1.67884618e-01 1.78277325e-02
-9.42348242e-02 4.71423008e-02 1.19015694e+00 -4.24055681e-02
-2.93080360e-01 -1.86015919e-01 -1.35917306e+00 8.77964199e-01
1.07209790e+00 -7.20828354e-01 -4.54411566e-01 -3.91779929e-01
7.75369108e-01 -3.44194502e-01 2.95936972e-01 -1.14591563e+00
-3.82486023e-02 -2.02060163e-01 8.25442612e-01 -5.27971864e-01
2.41362646e-01 -8.25998008e-01 3.18765223e-01 4.42510545e-01
-3.83227587e-01 4.71570134e-01 5.21169424e-01 1.43469110e-01
-1.16949759e-01 2.35959347e-02 9.51918960e-01 -3.64876390e-01
-3.97436500e-01 -1.64119720e-01 -1.09239221e-01 6.51697397e-01
1.01126075e+00 -3.88731658e-01 -6.96931779e-01 -1.50700495e-01
-4.47174728e-01 2.36941814e-01 7.05696166e-01 6.04678035e-01
2.52485424e-01 -1.15503931e+00 -6.41230285e-01 6.45799458e-01
1.82739004e-01 3.22441697e-01 1.33011267e-01 6.29556179e-01
-4.40605134e-01 -1.33968830e-01 -5.86344898e-01 -6.49803579e-01
-1.24242985e+00 4.20823216e-01 3.24703574e-01 -1.43861219e-01
-4.26735222e-01 9.08144772e-01 -1.57661527e-01 -7.18875349e-01
2.88619578e-01 -9.00650471e-02 -3.22395898e-02 5.87518096e-01
8.23997438e-01 3.84374887e-01 3.74181688e-01 6.79978654e-02
-3.53954166e-01 2.95524180e-01 -2.95345157e-01 1.67077184e-01
1.66996121e+00 1.26630321e-01 -3.11691791e-01 8.34705472e-01
9.92855549e-01 -2.49389112e-02 -8.44058156e-01 1.21271916e-01
1.48776501e-01 -9.25922930e-01 -1.72991768e-01 -1.35282731e+00
-9.78568733e-01 8.63446236e-01 8.38130176e-01 4.86555576e-01
1.11972046e+00 -5.26329517e-01 3.01197290e-01 8.12442005e-02
1.14253521e-01 -1.23374701e+00 1.11397564e-01 1.36665910e-01
8.12846720e-01 -1.64093673e+00 5.71588755e-01 -9.07154307e-02
-8.10392141e-01 6.03663623e-01 9.63867188e-01 -1.69328228e-01
6.26093686e-01 6.35796368e-01 3.25276911e-01 -3.73099327e-01
-8.08648169e-01 2.11911857e-01 -2.13977590e-01 7.43567884e-01
6.23390257e-01 -1.18304521e-01 -2.82426059e-01 3.52016598e-01
-6.06569886e-01 3.22832733e-01 7.50805736e-01 1.00235891e+00
-3.62294674e-01 -9.86795723e-01 -5.91528952e-01 9.20731902e-01
-6.78765476e-01 5.28942496e-02 -5.17752051e-01 1.07696521e+00
5.97325087e-01 7.32230783e-01 5.23219109e-01 -5.02384245e-01
4.17775393e-01 2.42111564e-01 2.34361976e-01 -5.69277704e-01
-8.60669255e-01 -6.44624591e-01 4.68052402e-02 -5.22534490e-01
-1.99034229e-01 -4.49579865e-01 -9.92900431e-01 -6.16705656e-01
-3.12586933e-01 -3.79573777e-02 1.02863729e+00 6.85085297e-01
4.61456805e-01 6.07610047e-01 5.78089535e-01 -4.79859859e-01
-4.09063369e-01 -9.61587191e-01 -3.69660914e-01 7.59739101e-01
6.95243061e-01 -5.65823257e-01 -6.55840099e-01 -3.09981167e-01] | [8.807780265808105, 4.3252272605896] |
04b97981-bf55-4889-a961-1119340b5996 | deep-ordinal-regression-with-label-diversity | 2006.15864 | null | https://arxiv.org/abs/2006.15864v1 | https://arxiv.org/pdf/2006.15864v1.pdf | Deep Ordinal Regression with Label Diversity | Regression via classification (RvC) is a common method used for regression problems in deep learning, where the target variable belongs to a set of continuous values. By discretizing the target into a set of non-overlapping classes, it has been shown that training a classifier can improve neural network accuracy compared to using a standard regression approach. However, it is not clear how the set of discrete classes should be chosen and how it affects the overall solution. In this work, we propose that using several discrete data representations simultaneously can improve neural network learning compared to a single representation. Our approach is end-to-end differentiable and can be added as a simple extension to conventional learning methods, such as deep neural networks. We test our method on three challenging tasks and show that our method reduces the prediction error compared to a baseline RvC approach while maintaining a similar model complexity. | ["Mark O'Connor", 'Magnus Oskarsson', 'Axel Berg'] | 2020-06-29 | null | null | null | null | ['head-pose-estimation', 'historical-color-image-dating'] | ['computer-vision', 'computer-vision'] | [ 5.19980669e-01 1.03150770e-01 -4.38875228e-01 -6.30730748e-01
-9.52035069e-01 -4.68653023e-01 5.40488958e-01 1.76704124e-01
-4.45905834e-01 9.47824240e-01 -2.71141231e-01 -4.43538219e-01
-2.27132425e-01 -7.23560929e-01 -8.18918049e-01 -6.43987298e-01
1.05660483e-01 6.25194490e-01 -3.78718041e-03 -1.26291737e-01
2.12046310e-01 7.03537583e-01 -1.39944708e+00 3.58331382e-01
6.46567643e-01 1.34390545e+00 -2.94375300e-01 5.49212635e-01
6.66205063e-02 1.02142918e+00 -8.31925511e-01 -8.20700154e-02
3.84947151e-01 -4.02114004e-01 -7.80911684e-01 -2.63625830e-01
7.02593863e-01 -7.10541978e-02 -7.54058585e-02 6.23571754e-01
1.30707383e-01 3.40342313e-01 1.01561677e+00 -1.21337819e+00
-4.91097420e-01 3.04773778e-01 -5.70969522e-01 2.39308528e-03
-2.48544365e-01 -3.24664980e-01 1.09631467e+00 -5.68581581e-01
3.01255286e-01 1.14312756e+00 9.94313478e-01 5.82490146e-01
-1.84682894e+00 -8.78633142e-01 3.91876400e-01 -9.67237428e-02
-1.10598636e+00 -2.17698678e-01 7.35025167e-01 -5.74827075e-01
1.00869787e+00 1.85352415e-01 3.36251765e-01 9.77551162e-01
-9.33566038e-03 7.05602646e-01 1.17076719e+00 -6.19823396e-01
3.95456165e-01 2.08806470e-01 5.77815950e-01 5.32789052e-01
9.30843428e-02 1.70788035e-01 -1.25436066e-02 -1.83951318e-01
6.85115159e-01 1.97027400e-01 -1.60927132e-01 -4.84840661e-01
-6.15235507e-01 1.46156919e+00 6.13804281e-01 4.34511900e-02
-2.52920479e-01 7.28091896e-01 5.03095686e-01 5.98217607e-01
8.38691294e-01 5.76291561e-01 -8.41225207e-01 5.10182604e-03
-1.00815403e+00 3.75177681e-01 9.45066452e-01 6.14615858e-01
7.08363950e-01 2.06438571e-01 -1.35640547e-01 1.22309685e+00
-2.79782563e-02 4.21926565e-02 4.40569729e-01 -9.98237014e-01
4.20169622e-01 4.28986967e-01 2.62000039e-02 -9.41810369e-01
-4.09970254e-01 -5.23113191e-01 -8.85226369e-01 6.63282573e-01
6.60213351e-01 -3.93937111e-01 -1.05064082e+00 1.72129703e+00
-6.42627105e-02 9.23812613e-02 -7.54306046e-03 6.40179873e-01
7.03314245e-01 5.45354545e-01 -5.92184514e-02 -1.71530157e-01
6.07831955e-01 -1.07053137e+00 -4.30416971e-01 -1.67753428e-01
6.50052726e-01 -3.14726561e-01 6.39786661e-01 8.28568399e-01
-7.16430247e-01 -4.82097924e-01 -1.17128515e+00 -1.77455291e-01
-6.44126236e-01 1.05942450e-01 7.40476072e-01 8.09491575e-01
-8.45550895e-01 1.06330693e+00 -7.15741098e-01 9.79345664e-02
6.75016403e-01 8.26184571e-01 -3.80326539e-01 -6.86845705e-02
-9.34235692e-01 1.04457211e+00 2.66129732e-01 3.44956061e-04
-5.67280889e-01 -7.83821464e-01 -8.18659544e-01 -3.10925264e-02
2.34767482e-01 -3.41344237e-01 1.36031699e+00 -1.39029455e+00
-1.49344695e+00 5.95930755e-01 -1.91459328e-01 -8.61116767e-01
5.82558095e-01 -2.63261199e-01 7.72560909e-02 -2.27595821e-01
-3.26005369e-01 7.02383280e-01 9.07363713e-01 -1.14297175e+00
-7.22548544e-01 -3.10879916e-01 6.26464635e-02 -6.67473748e-02
-1.38333797e-01 -5.74366096e-03 -2.55988687e-01 -6.96876585e-01
-1.04429364e-01 -9.28418100e-01 -5.07714033e-01 3.59370448e-02
-2.51754731e-01 -4.38151568e-01 7.27090776e-01 -6.03598297e-01
1.02118397e+00 -1.81968081e+00 3.08841527e-01 4.14689064e-01
3.98732930e-01 1.75789639e-01 -2.83045880e-03 3.56320670e-04
-4.55448896e-01 3.25643092e-01 -4.68977839e-01 -3.86038899e-01
-1.87169105e-01 3.14319313e-01 -3.69088441e-01 6.98988438e-01
5.10309696e-01 6.81100130e-01 -6.30534887e-01 -1.15601070e-01
9.65051427e-02 7.03327298e-01 -4.30278748e-01 -4.83109169e-02
-2.36372083e-01 3.85145783e-01 -3.30834270e-01 3.35294008e-01
5.55561662e-01 -2.78322473e-02 5.69777749e-02 3.33486885e-01
1.94477782e-01 3.02165031e-01 -9.63967085e-01 1.07903910e+00
-7.62939453e-01 1.04761827e+00 -3.92255396e-01 -1.83415675e+00
1.40462923e+00 2.19952196e-01 5.07584691e-01 -4.62680936e-01
5.06701097e-02 2.60712445e-01 7.52323791e-02 3.29944417e-02
1.57119110e-02 -1.71014667e-01 -8.86088312e-02 1.54210016e-01
4.38955752e-03 3.25018279e-02 2.27666311e-02 -5.28284192e-01
1.16015851e+00 3.23400617e-01 4.92526263e-01 -1.20740958e-01
4.09978062e-01 2.09029838e-01 6.09694839e-01 8.58314931e-01
1.39587685e-01 8.12330425e-01 9.20674145e-01 -5.77432632e-01
-7.97002792e-01 -6.98407948e-01 -4.39336747e-01 1.12062669e+00
-3.26583892e-01 -1.55819312e-01 -5.86066604e-01 -9.90143180e-01
4.40846503e-01 7.47637510e-01 -1.07412243e+00 -2.19537422e-01
-7.70877957e-01 -7.38970876e-01 5.78259885e-01 8.04423094e-01
-4.74335402e-02 -9.93986547e-01 -4.71218675e-01 2.24683344e-01
3.35650474e-01 -9.25330579e-01 -5.78992665e-02 9.43900168e-01
-1.20035732e+00 -1.06252968e+00 -5.79350293e-01 -7.86700189e-01
6.64255917e-01 4.68043387e-02 1.33387160e+00 1.76133305e-01
-1.71211436e-01 1.58207595e-01 -2.66753525e-01 -5.21735907e-01
-4.24050242e-01 2.33244210e-01 -2.17750281e-01 -1.17101625e-01
4.58726168e-01 -3.88767600e-01 -2.97890395e-01 7.51573034e-03
-4.88524824e-01 -3.56817432e-02 2.89657325e-01 1.07492077e+00
6.53939426e-01 -1.31989315e-01 9.20405567e-01 -1.53314793e+00
6.74804747e-01 -6.62141204e-01 -7.32188761e-01 1.36943802e-01
-7.94070363e-01 2.34281152e-01 1.06656039e+00 -7.09864140e-01
-5.94826162e-01 3.28870744e-01 3.45703103e-02 -5.98037124e-01
-1.10243797e-01 6.86545372e-01 4.16682631e-01 -1.97429135e-01
6.60993397e-01 -1.34715557e-01 3.44844431e-01 -5.75588346e-01
2.70265460e-01 5.15742004e-01 1.89689636e-01 -6.01013064e-01
5.63868999e-01 -1.06104970e-01 3.82736325e-01 -7.24344075e-01
-7.58280098e-01 -3.16390961e-01 -9.50634480e-01 5.73176555e-02
5.06223202e-01 -7.24968374e-01 -6.68219388e-01 1.96023844e-03
-1.11404872e+00 -6.83533490e-01 -1.95047230e-01 3.54018152e-01
-6.14225149e-01 -1.46441534e-01 -3.64961028e-01 -7.61652887e-01
-1.83291644e-01 -1.22139156e+00 7.35983133e-01 -1.88112911e-02
-3.53958786e-01 -1.31486750e+00 -5.40656932e-02 1.32000536e-01
3.40102226e-01 6.70951843e-01 1.01426625e+00 -1.02179432e+00
-1.19482882e-01 -5.14146209e-01 -1.24221593e-01 6.61954343e-01
8.71328339e-02 9.43562314e-02 -9.59237278e-01 -2.82256812e-01
-2.64188826e-01 -6.34854138e-01 1.42380440e+00 5.59465349e-01
1.66074908e+00 -2.33756587e-01 -1.84624657e-01 7.60194123e-01
1.52332246e+00 2.87119657e-01 6.15072906e-01 4.44500893e-01
6.02722764e-01 6.32129550e-01 6.03862703e-01 9.79852974e-02
-8.73152241e-02 7.64590740e-01 3.80493611e-01 -3.98649186e-01
1.27777830e-01 9.01866481e-02 2.29446277e-01 9.26498771e-02
-2.04764172e-01 9.01025757e-02 -8.72637689e-01 2.34284744e-01
-1.97294021e+00 -7.87915111e-01 -1.88935727e-01 2.42686725e+00
9.34878826e-01 1.31261885e-01 3.05966049e-01 3.98834318e-01
3.98257464e-01 -2.37714455e-01 -7.16858625e-01 -8.02893996e-01
2.35686645e-01 7.82747567e-01 8.66520703e-01 5.66692948e-01
-1.31093514e+00 8.24541569e-01 7.01698589e+00 6.77802444e-01
-1.47663331e+00 -2.47257762e-02 9.98170376e-01 -1.95536688e-01
1.18191481e-01 -2.96464801e-01 -8.73768628e-01 1.52939349e-01
1.02621210e+00 8.73510391e-02 4.72542942e-01 1.02737772e+00
5.46537265e-02 2.30817512e-01 -1.59806752e+00 8.31205964e-01
6.83550164e-03 -1.25251853e+00 -4.22626942e-01 1.71076506e-02
5.92192292e-01 5.49588315e-02 5.55593334e-02 8.47962499e-01
5.98084271e-01 -1.78596723e+00 4.69400615e-01 4.02253807e-01
7.05565512e-01 -9.50190842e-01 5.89528799e-01 4.15895700e-01
-8.39715362e-01 -2.87899405e-01 -3.95303309e-01 -2.49148294e-01
-7.23749340e-01 2.61041522e-01 -9.70117927e-01 2.94577628e-01
5.84585547e-01 9.41243649e-01 -7.89265752e-01 8.31176519e-01
-4.14435333e-03 1.02716804e+00 -2.47695908e-01 -1.17209882e-01
2.70344257e-01 -2.82745153e-01 1.01944236e-02 1.32560956e+00
-6.70181811e-02 -1.86772287e-01 3.02329838e-01 8.56160104e-01
-1.50721654e-01 1.27931284e-02 -7.05975056e-01 2.03490570e-01
2.97840476e-01 9.67920959e-01 -6.73107386e-01 -1.70083866e-01
-5.74527979e-01 5.99331796e-01 8.40041757e-01 5.53603172e-01
-8.07875156e-01 -4.47497666e-01 5.39673626e-01 -7.51731619e-02
5.57480752e-01 3.11231203e-02 -7.72840559e-01 -7.05286026e-01
-7.30865356e-03 -9.72269714e-01 3.60828221e-01 -2.10824683e-01
-1.22900093e+00 4.02897239e-01 4.26323526e-02 -1.15114832e+00
-4.37930048e-01 -8.95497978e-01 -3.74831378e-01 1.13744831e+00
-1.60896599e+00 -8.42473447e-01 -1.16460688e-01 3.19639862e-01
5.58012366e-01 -3.95331800e-01 9.99995768e-01 4.18084897e-02
-6.79529905e-01 7.99258292e-01 5.92525601e-01 2.23213837e-01
5.44453979e-01 -1.59464908e+00 2.35039443e-01 4.25771654e-01
5.33482842e-02 5.54287970e-01 6.13120079e-01 -2.61159986e-01
-1.01363945e+00 -1.34450150e+00 5.43646991e-01 -3.29919428e-01
3.47730994e-01 -3.98764163e-01 -1.13468981e+00 8.63603115e-01
-1.03067778e-01 4.02295202e-01 7.16352820e-01 5.26505411e-01
-5.31450629e-01 -3.77389759e-01 -1.15359986e+00 2.28318095e-01
4.39204872e-01 -2.19030619e-01 -4.37385082e-01 2.22322613e-01
6.40271842e-01 -2.48078004e-01 -9.57298636e-01 4.88642126e-01
7.34869480e-01 -5.20808339e-01 1.05498457e+00 -1.08613837e+00
7.81954825e-01 6.73335865e-02 -2.62474060e-01 -1.52997673e+00
-2.48167768e-01 -2.54298270e-01 -3.55878443e-01 8.50940406e-01
6.34574592e-01 -8.27712953e-01 8.45902741e-01 6.90665543e-01
1.85843632e-01 -1.23508072e+00 -7.58213460e-01 -6.87762320e-01
8.02626848e-01 -4.43104357e-01 1.44223869e-01 9.05179262e-01
-4.03697312e-01 5.18755019e-01 -3.99577171e-01 -6.97106495e-02
3.80399913e-01 1.55928433e-01 8.20442975e-01 -1.87035310e+00
-4.19608057e-01 -7.53755569e-01 -4.02531236e-01 -9.15444911e-01
5.13362408e-01 -1.07652736e+00 -8.56864383e-04 -1.51744890e+00
-1.72756627e-01 -8.16715539e-01 -5.85276663e-01 8.36208642e-01
5.22873960e-02 3.53938133e-01 1.82917714e-01 5.78339361e-02
-1.80629730e-01 2.54156321e-01 9.23770308e-01 -3.26476485e-01
-3.91238064e-01 4.20021832e-01 -9.00971770e-01 5.67747116e-01
9.79642153e-01 -7.81666994e-01 -3.80020440e-01 -5.95712177e-02
2.50997059e-02 2.82261431e-01 1.85561612e-01 -6.74009144e-01
-6.49215654e-02 -3.30370069e-01 7.32100070e-01 -3.19562793e-01
4.46039349e-01 -8.76393735e-01 -2.51419991e-01 5.23919046e-01
-8.99293542e-01 -3.10270607e-01 2.40978673e-01 6.18056417e-01
-1.59803063e-01 -5.06651759e-01 9.58581328e-01 8.15554932e-02
-2.85547554e-01 4.30229604e-02 -3.01374882e-01 -2.84862537e-02
9.76180315e-01 -1.53558001e-01 4.32643257e-02 -3.27129155e-01
-6.81171060e-01 2.26966679e-01 6.27674609e-02 3.92654777e-01
6.21735513e-01 -1.21605480e+00 -7.66897082e-01 5.42817600e-02
-6.66453391e-02 1.23135015e-01 -5.41507244e-01 7.34304488e-01
-4.74301785e-01 4.19352561e-01 3.86803970e-02 -5.54188132e-01
-1.53624475e+00 4.68965769e-01 7.00319767e-01 -4.43631500e-01
-6.61777437e-01 6.42127931e-01 -1.19188102e-02 -7.29464769e-01
6.48986757e-01 -4.91988927e-01 -5.38100064e-01 1.30331039e-01
2.52422273e-01 2.49930114e-01 1.59508109e-01 -2.67216593e-01
-7.90099204e-02 5.19680798e-01 -3.00316513e-01 1.74877569e-01
1.67370641e+00 5.34423292e-01 1.05480455e-01 8.41042161e-01
1.59709847e+00 -4.37014163e-01 -1.24111128e+00 -2.78166443e-01
1.59883678e-01 -2.76064873e-01 2.02974379e-01 -8.69404614e-01
-1.19960272e+00 9.19731557e-01 5.45436740e-01 2.24689960e-01
1.10034227e+00 -3.70852977e-01 2.51124889e-01 8.09319496e-01
4.32807952e-02 -9.68035817e-01 -2.01104388e-01 5.63862503e-01
1.05461931e+00 -1.51057100e+00 1.63844138e-01 -3.35560888e-01
-5.72496891e-01 1.60179472e+00 5.20694673e-01 -7.43510604e-01
7.27896452e-01 2.35611930e-01 -8.06598961e-02 2.19182909e-01
-9.84693885e-01 -9.41614136e-02 4.57338274e-01 6.68887019e-01
7.98594177e-01 1.15210367e-02 -1.47964329e-01 5.33739924e-01
-2.10927635e-01 1.55484840e-01 4.09874350e-01 7.92712510e-01
-1.63969025e-01 -1.16234970e+00 -3.90752763e-01 9.24475074e-01
-6.76194310e-01 3.04608457e-02 -3.96050125e-01 1.12751877e+00
-3.40752341e-02 9.36726034e-01 2.39528790e-01 -2.84338504e-01
2.04741731e-01 7.29103684e-02 5.42822897e-01 -7.50850439e-01
-7.47116327e-01 -1.62952617e-01 1.89467207e-01 -2.43055403e-01
-3.14327568e-01 -5.00293314e-01 -1.27899837e+00 2.51903608e-02
-3.05559605e-01 -4.19308655e-02 5.82102180e-01 1.02894521e+00
-5.47401160e-02 7.73597717e-01 8.06768596e-01 -7.52836227e-01
-9.20730650e-01 -9.51282024e-01 -2.35435218e-01 2.31343418e-01
9.10825968e-01 -9.24559295e-01 -3.36497903e-01 -1.83759946e-02] | [8.875709533691406, 3.4338619709014893] |
93974c99-e64e-4ad4-be4f-9a782d13a2cf | counterfactual-inference-for-consumer-choice | 1906.02635 | null | https://arxiv.org/abs/1906.02635v1 | https://arxiv.org/pdf/1906.02635v1.pdf | Counterfactual Inference for Consumer Choice Across Many Product Categories | This paper proposes a method for estimating consumer preferences among discrete choices, where the consumer chooses at most one product in a category, but selects from multiple categories in parallel. The consumer's utility is additive in the different categories. Her preferences about product attributes as well as her price sensitivity vary across products and are in general correlated across products. We build on techniques from the machine learning literature on probabilistic models of matrix factorization, extending the methods to account for time-varying product attributes and products going out of stock. We evaluate the performance of the model using held-out data from weeks with price changes or out of stock products. We show that our model improves over traditional modeling approaches that consider each category in isolation. One source of the improvement is the ability of the model to accurately estimate heterogeneity in preferences (by pooling information across categories); another source of improvement is its ability to estimate the preferences of consumers who have rarely or never made a purchase in a given category in the training data. Using held-out data, we show that our model can accurately distinguish which consumers are most price sensitive to a given product. We consider counterfactuals such as personally targeted price discounts, showing that using a richer model such as the one we propose substantially increases the benefits of personalization in discounts. | ['David Blei', 'Susan Athey', 'Francisco R. Ruiz', 'Rob Donnelly'] | 2019-06-06 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [-3.82479250e-01 -1.50902376e-01 -9.31041121e-01 -6.67246401e-01
-7.94182062e-01 -8.36062610e-01 4.58849341e-01 4.93490428e-01
-5.97512245e-01 5.01766682e-01 4.98233557e-01 -3.24627817e-01
-2.54696935e-01 -9.45310652e-01 -6.52034521e-01 -5.26277602e-01
-3.23540628e-01 7.66986132e-01 -1.05444923e-01 -1.81638867e-01
1.06105566e-01 2.01638743e-01 -1.39194286e+00 3.54973197e-01
5.80857575e-01 1.09648919e+00 -3.21182162e-02 3.36865425e-01
-1.29414394e-01 4.90777910e-01 -1.29004821e-01 -8.58554065e-01
7.00954735e-01 1.04058765e-01 -4.11547214e-01 4.12828237e-01
-5.65773584e-02 -5.20006955e-01 6.13817535e-02 9.98603463e-01
7.27980863e-03 9.11210105e-02 9.09976184e-01 -1.30462158e+00
-8.14508736e-01 1.07365847e+00 -6.00737035e-01 6.25129091e-03
1.06777661e-02 -1.83761969e-01 1.50202966e+00 -2.93358833e-01
1.61424145e-01 1.41017735e+00 5.02464235e-01 3.16013455e-01
-1.94127119e+00 -4.72446233e-01 8.03068519e-01 -2.31108248e-01
-1.07627392e+00 -3.07667673e-01 6.07763112e-01 -6.73445106e-01
4.93971020e-01 2.85348624e-01 5.77793360e-01 7.65700340e-01
2.83278555e-01 1.01014078e+00 1.08436656e+00 -3.35866094e-01
6.52621925e-01 5.34894943e-01 3.14743370e-01 -1.15395583e-01
5.60297072e-01 3.88208389e-01 -1.38088837e-01 -6.51628256e-01
6.91852391e-01 3.88874799e-01 1.25439182e-01 -7.07179308e-01
-8.83803725e-01 1.45312285e+00 1.44990325e-01 -9.18357521e-02
-6.42550886e-01 -1.26253247e-01 2.92640984e-01 6.15036249e-01
7.16752112e-01 1.94413260e-01 -9.71436739e-01 2.37601176e-01
-8.73488069e-01 6.82164609e-01 1.13692105e+00 8.83434236e-01
6.53682649e-01 -3.00563097e-01 -9.05182809e-02 8.03061247e-01
3.54270458e-01 3.36785704e-01 6.64064288e-01 -1.06853163e+00
4.14745122e-01 1.75423220e-01 9.54980373e-01 -6.17766023e-01
-3.34034801e-01 -2.64445782e-01 -3.17362726e-01 2.75182813e-01
5.30411780e-01 -4.52639282e-01 -6.66756213e-01 2.04370952e+00
3.05829942e-02 -4.74854797e-01 1.53300408e-02 5.13791203e-01
-2.08179072e-01 6.29983902e-01 4.68449295e-01 -5.35142899e-01
1.45438480e+00 -5.20983756e-01 -3.90146196e-01 -9.72690433e-02
6.56379282e-01 -5.91557503e-01 6.68141127e-01 6.85592890e-01
-1.01191545e+00 -2.73545265e-01 -6.77574754e-01 2.28816032e-01
-2.33006150e-01 -2.76632994e-01 1.25949252e+00 9.81345654e-01
-8.36194992e-01 8.39677215e-01 -7.36170590e-01 -4.21448126e-02
2.16806963e-01 4.64964360e-01 1.78871185e-01 3.93996537e-02
-1.24478400e+00 6.99213207e-01 3.13506238e-02 -5.40236294e-01
-5.95438182e-01 -7.99014688e-01 -7.91561604e-01 3.52149427e-01
2.84461975e-01 -6.38932884e-01 1.84318054e+00 -1.40827775e+00
-1.28617811e+00 1.87895909e-01 -4.96397354e-02 -6.26039803e-01
4.78821456e-01 1.41260833e-01 -6.86867893e-01 -5.37367404e-01
4.56233919e-01 4.82010424e-01 5.68627596e-01 -1.06513786e+00
-1.35968411e+00 -5.49297631e-01 3.49998355e-01 2.61282682e-01
-5.79476058e-02 1.73786975e-04 1.84968054e-01 -6.73928022e-01
-7.16592371e-02 -1.07771158e+00 -5.08622468e-01 -3.07668597e-01
-1.55740231e-01 -4.11532193e-01 -5.19728065e-02 -2.11887494e-01
9.06902254e-01 -1.98893273e+00 -2.71145165e-01 2.42224097e-01
-2.93050736e-01 -5.58903635e-01 6.39177114e-02 4.85004574e-01
-7.73457140e-02 3.86618555e-01 1.64514020e-01 -4.94945467e-01
6.43901885e-01 4.14579868e-01 -3.65003318e-01 4.72451031e-01
-1.85846120e-01 4.90466863e-01 -6.32452130e-01 4.83243428e-02
-5.99047504e-02 5.41073792e-02 -6.77418530e-01 -3.44200045e-01
-2.36587167e-01 -2.04438061e-01 -4.90448266e-01 4.04898614e-01
8.08705032e-01 5.54882698e-02 3.48267823e-01 1.01133160e-01
-3.64866287e-01 6.02658808e-01 -1.56108356e+00 1.18226504e+00
-5.76290786e-01 -1.65512770e-01 1.08756855e-01 -8.36961389e-01
1.98568702e-01 5.13780951e-01 5.36099970e-01 -4.00510103e-01
8.25065300e-02 2.05401450e-01 3.95869240e-02 -9.39559117e-02
4.03029501e-01 -7.48030961e-01 -3.98267418e-01 7.10475862e-01
-4.25107367e-02 1.02920063e-01 3.65133643e-01 1.56299278e-01
2.93239534e-01 -1.90754414e-01 3.51380408e-01 -7.29459107e-01
-2.57319510e-01 -1.18807696e-01 9.20149863e-01 8.58549237e-01
3.35101448e-02 2.36621629e-02 5.14252484e-01 -3.52048576e-01
-1.08204305e+00 -1.22505844e+00 -3.67706418e-01 1.53865695e+00
-6.30025268e-02 7.68919960e-02 -2.21802503e-01 -4.96118456e-01
7.88512468e-01 1.00931466e+00 -9.04081464e-01 2.69273669e-01
3.29357862e-01 -1.17600572e+00 -4.55518335e-01 7.08652198e-01
5.80085963e-02 -3.95412236e-01 -2.62251914e-01 2.95805454e-01
6.07727356e-02 -2.71550059e-01 -7.44884074e-01 3.21779817e-01
-8.86635959e-01 -6.99997067e-01 -6.86244726e-01 -3.69889528e-01
2.61114150e-01 1.97695106e-01 1.20768929e+00 -5.60697556e-01
3.66966128e-01 4.66967821e-01 -1.65300667e-01 -8.10388684e-01
-1.52222380e-01 -1.20198950e-01 3.93729925e-01 3.47006589e-01
7.60707676e-01 -4.13554549e-01 -6.41787589e-01 3.42556536e-01
-7.82565117e-01 -4.51130122e-01 3.72173250e-01 6.55344546e-01
2.67390728e-01 5.87001026e-01 5.41443288e-01 -1.30059183e+00
6.40742123e-01 -9.83277321e-01 -9.22128141e-01 1.12547159e-01
-9.19421792e-01 2.84844756e-01 3.12519073e-01 -7.96513617e-01
-1.36065996e+00 1.60063103e-01 1.00904785e-01 3.93710107e-01
-7.11679831e-02 5.34933686e-01 -3.59160632e-01 4.58731890e-01
3.74049097e-01 -3.05921197e-01 -1.06352508e-01 -7.75336862e-01
5.94971180e-01 4.85293299e-01 1.31021947e-01 -7.29140401e-01
5.55233300e-01 4.41568315e-01 -2.79351324e-01 -2.70188451e-01
-5.62509298e-01 -4.09730583e-01 -4.23441917e-01 5.77632785e-01
3.83633941e-01 -1.36708307e+00 -1.05531847e+00 4.43498008e-02
-5.22495389e-01 -2.71435499e-01 -5.26430190e-01 9.53643262e-01
-6.87605917e-01 3.44609027e-03 -8.67827535e-01 -1.27627420e+00
3.19795430e-01 -1.05518711e+00 5.27480364e-01 1.19776465e-01
-3.21143776e-01 -1.34154236e+00 3.80909033e-02 -7.86758065e-02
2.98718661e-01 -2.22918659e-01 9.57692087e-01 -9.09095824e-01
-3.55222464e-01 -3.71082127e-01 3.41385067e-01 1.61413789e-01
4.18620646e-01 -7.83648789e-02 -6.24127209e-01 -3.60187590e-01
2.43270546e-01 -1.54764667e-01 9.33906138e-01 1.02614021e+00
6.43532157e-01 -6.92442656e-01 -4.73582387e-01 9.46957543e-02
1.51064360e+00 6.10161126e-01 1.61419958e-01 2.36371681e-01
9.72527787e-02 9.43483591e-01 4.85793859e-01 6.93135321e-01
7.67631054e-01 7.03603446e-01 9.35152248e-02 6.39836788e-02
8.94563973e-01 -3.45995873e-01 3.11159939e-01 -8.13620463e-02
7.98002034e-02 -2.19611794e-01 -6.70524910e-02 8.79138887e-01
-2.02274561e+00 -1.08950341e+00 2.44103089e-01 2.52295566e+00
6.42057896e-01 2.43987128e-01 1.02195811e+00 1.60728414e-02
6.48122370e-01 -2.36526981e-01 -7.19226241e-01 -6.98219895e-01
8.55675638e-02 -1.11103833e-01 1.18208158e+00 8.69130194e-01
-1.07957852e+00 2.50187904e-01 7.68676329e+00 6.07856631e-01
-6.80622876e-01 2.49668062e-02 1.14617372e+00 -3.61291558e-01
-9.00700629e-01 1.65959835e-01 -1.00055707e+00 6.66885555e-01
1.02433944e+00 -5.57305217e-01 4.68987972e-01 1.07487822e+00
2.04055086e-01 -1.58797294e-01 -1.64877164e+00 4.88678724e-01
-4.36466366e-01 -9.03122008e-01 -1.93564013e-01 6.18950367e-01
8.72283518e-01 -3.07336539e-01 6.07329369e-01 3.56528014e-01
1.38024962e+00 -6.32372081e-01 9.89566267e-01 2.13031545e-01
1.13703817e-01 -1.15371823e+00 6.01542115e-01 5.57309926e-01
-8.73998106e-01 -6.73388898e-01 -5.08297145e-01 -5.56713820e-01
2.83882976e-01 7.11256981e-01 -4.16996479e-01 2.31374413e-01
4.85053718e-01 2.25090310e-01 -4.67000678e-02 7.90359020e-01
1.92356899e-01 5.28594613e-01 -4.66567099e-01 1.49652272e-01
1.51039258e-01 -3.36797029e-01 -8.87545571e-02 8.93175900e-01
4.92802888e-01 2.68583953e-01 4.35946375e-01 7.31163144e-01
-1.94871314e-02 1.51605874e-01 -3.95547748e-01 5.91245517e-02
5.11131525e-01 9.47720885e-01 -5.36824286e-01 -4.73378599e-01
-1.02412093e+00 6.56399429e-01 -1.59930229e-01 3.88578683e-01
-4.25141841e-01 6.43738359e-02 9.99637544e-01 3.98081154e-01
4.57735211e-01 5.20190001e-02 -2.15518981e-01 -1.25134790e+00
-2.64546156e-01 -5.92527032e-01 7.79622972e-01 -2.01893568e-01
-1.80479217e+00 -1.23935640e-01 3.65282387e-01 -9.19473648e-01
-6.01573050e-01 -6.45286977e-01 -3.31152350e-01 1.17005539e+00
-1.31824338e+00 -8.08315396e-01 9.38585639e-01 5.43346345e-01
6.14983499e-01 3.89342383e-02 7.58942127e-01 1.55230891e-02
-3.97557989e-02 4.08173770e-01 5.23357451e-01 -1.75930753e-01
3.97860825e-01 -1.57383835e+00 5.68251312e-01 4.42246705e-01
1.51177362e-01 9.38301146e-01 8.45194876e-01 -6.49797499e-01
-8.44336510e-01 -6.71255291e-01 1.05808318e+00 -4.61158335e-01
8.13727200e-01 -3.73948097e-01 -5.57200968e-01 1.25129604e+00
5.32447658e-02 -4.76448417e-01 1.21040297e+00 7.89438665e-01
-4.65067953e-01 -2.89445311e-01 -1.54895806e+00 4.35560644e-01
4.77712929e-01 -2.38405839e-01 -4.86895680e-01 3.87391657e-01
7.10303307e-01 2.17892051e-01 -1.01206195e+00 -1.81868672e-01
9.19984221e-01 -1.05985606e+00 9.73282039e-01 -6.17656529e-01
-5.96823841e-02 9.87197310e-02 -5.58718979e-01 -1.69040298e+00
-1.04519498e+00 -5.25878787e-01 5.69769204e-01 1.33921778e+00
7.97861397e-01 -9.16298211e-01 7.04357505e-01 1.70455337e+00
3.84194463e-01 -2.72067070e-01 -7.48789251e-01 -8.28128636e-01
6.93058610e-01 -3.05287480e-01 1.13490856e+00 8.14640343e-01
4.09379810e-01 1.79948658e-01 -4.32027072e-01 1.18693247e-01
7.79573619e-01 3.60556781e-01 2.46396571e-01 -1.68760121e+00
-6.97278142e-01 -3.63594174e-01 4.72648665e-02 -1.17792761e+00
-3.78239378e-02 -4.45328414e-01 -3.23389351e-01 -1.09044969e+00
4.34455305e-01 -6.68154180e-01 -5.17270029e-01 2.24669024e-01
1.58959895e-01 -3.66402119e-02 2.60890990e-01 1.83763593e-01
-8.51332098e-02 -1.52060408e-02 7.48119891e-01 -2.23401397e-01
-5.07465780e-01 7.59288132e-01 -1.59781885e+00 7.66182959e-01
7.02568114e-01 -3.86881530e-01 -7.07057655e-01 -1.33289665e-01
3.37783217e-01 2.36076832e-01 -1.93109922e-02 -2.71034122e-01
-5.04401661e-02 -6.70360625e-01 7.02209949e-01 -3.39642346e-01
3.21716964e-01 -8.93324137e-01 5.69063008e-01 2.44640425e-01
-7.41152346e-01 2.92044997e-01 4.23310734e-02 7.70264208e-01
2.54974306e-01 -2.77379543e-01 5.52041650e-01 -5.42326808e-01
-4.61399376e-01 4.91952598e-01 -6.49061263e-01 -5.20499110e-01
7.18761384e-01 1.41351342e-01 1.07946441e-01 -7.61287153e-01
-1.17743361e+00 2.76142657e-01 7.75678039e-01 1.88704446e-01
-1.54719546e-01 -1.47994661e+00 -4.92356181e-01 7.06736520e-02
8.77305269e-02 -5.39935887e-01 1.90443128e-01 2.79492915e-01
3.23721021e-01 4.31400567e-01 -6.18034378e-02 -8.63431618e-02
-7.24856675e-01 1.24838459e+00 1.37218013e-01 -2.57181793e-01
1.55143319e-02 7.83881545e-01 4.82167929e-01 -8.14005584e-02
-6.13548271e-02 -5.35661697e-01 -5.34698032e-02 4.37309802e-01
5.93657017e-01 3.13435525e-01 -1.78324744e-01 -7.10749149e-01
-1.28244311e-01 2.54350156e-01 -4.59108293e-01 -5.81217885e-01
1.21768641e+00 -5.98820210e-01 2.49222383e-01 8.64413798e-01
9.05680597e-01 1.60730526e-01 -1.64532137e+00 -1.99258924e-01
1.53492391e-02 -7.73506165e-01 -2.19389657e-03 -8.59937012e-01
-1.04368126e+00 3.91880155e-01 4.71277297e-01 7.96253383e-01
8.67136240e-01 -8.25212002e-02 4.33517218e-01 -8.73119086e-02
7.72447467e-01 -1.21460259e+00 -5.51811695e-01 -3.28222573e-01
3.15166652e-01 -1.27526546e+00 2.18732767e-02 -2.43118003e-01
-1.03255558e+00 6.11949980e-01 -4.02544945e-01 -3.24184567e-01
1.08760715e+00 6.65944591e-02 5.81308790e-02 3.13585281e-01
-9.33301449e-01 -1.35208458e-01 1.84680130e-02 8.00108314e-01
5.83958268e-01 7.78754711e-01 -4.66605872e-01 1.10102117e+00
-3.32902312e-01 -5.05303107e-02 7.80147076e-01 6.82395756e-01
-4.21659499e-01 -1.45650065e+00 -2.84621716e-01 7.35686243e-01
-7.95716286e-01 -9.73553658e-02 4.49609123e-02 7.87711322e-01
2.12171748e-01 1.26654494e+00 4.50608641e-01 -7.94382393e-02
4.12569344e-01 2.68721521e-01 4.36696082e-01 -7.38928676e-01
-2.12426066e-01 4.96606857e-01 5.07022031e-02 -2.42002919e-01
-4.16592360e-01 -1.38309348e+00 -6.69249833e-01 -6.54114604e-01
-2.82383919e-01 4.78532255e-01 5.76601207e-01 7.13016689e-01
1.25885680e-01 -1.51379645e-01 1.04768181e+00 -8.15852225e-01
-1.12939978e+00 -6.77225888e-01 -1.58877897e+00 6.49089098e-01
5.08496702e-01 -9.36385512e-01 -6.68996513e-01 9.93474200e-02] | [9.143209457397461, 5.503706455230713] |
2623fee8-e905-4c96-89ec-5cdbbeb0334c | fine-tuned-neural-models-for-propaganda | 1910.09702 | null | https://arxiv.org/abs/1910.09702v1 | https://arxiv.org/pdf/1910.09702v1.pdf | Fine-Tuned Neural Models for Propaganda Detection at the Sentence and Fragment levels | This paper presents the CUNLP submission for the NLP4IF 2019 shared-task on FineGrained Propaganda Detection. Our system finished 5th out of 26 teams on the sentence-level classification task and 5th out of 11 teams on the fragment-level classification task based on our scores on the blind test set. We present our models, a discussion of our ablation studies and experiments, and an analysis of our performance on all eighteen propaganda techniques present in the corpus of the shared task. | ['Tariq Alhindi', 'Smaranda Muresan', 'Jonas Pfeiffer'] | 2019-10-22 | fine-tuned-neural-models-for-propaganda-1 | https://aclanthology.org/D19-5013 | https://aclanthology.org/D19-5013.pdf | ws-2019-11 | ['propaganda-detection'] | ['natural-language-processing'] | [ 1.99403018e-02 1.89528629e-01 -1.49412453e-01 -3.17576900e-02
-1.32686615e+00 -6.96718097e-01 1.12558401e+00 4.04054016e-01
-9.23198402e-01 6.82244122e-01 9.57426488e-01 -6.30153894e-01
-2.07378119e-01 -2.01332137e-01 -2.88155198e-01 -3.14904720e-01
-1.12217635e-01 2.59013563e-01 7.58552849e-02 -4.40706074e-01
8.04198861e-01 8.71651024e-02 -6.10657632e-01 9.59967494e-01
9.01310205e-01 4.00220543e-01 -4.12855923e-01 1.00933039e+00
2.47497827e-01 1.32586884e+00 -1.29430151e+00 -6.05450511e-01
-3.54581103e-02 -4.24668938e-01 -1.24797690e+00 -6.92913949e-01
1.07767797e+00 -6.17489517e-02 -3.09080660e-01 9.61086094e-01
6.38403654e-01 -1.49993852e-01 8.47387254e-01 -5.18778324e-01
-7.94989765e-01 1.06679988e+00 -4.88700628e-01 8.29843700e-01
5.12075365e-01 3.37274708e-02 9.96286750e-01 -9.21093822e-01
8.76063943e-01 1.22398245e+00 6.71157062e-01 6.94921076e-01
-1.12625468e+00 -7.27923095e-01 8.64141062e-02 2.73077279e-01
-9.45896983e-01 -7.44899094e-01 4.65871036e-01 -8.37282658e-01
1.34906471e+00 1.10674813e-01 4.48321164e-01 1.85908580e+00
3.33839744e-01 7.01352596e-01 1.42546332e+00 -4.20877278e-01
-5.84667586e-02 -1.95178151e-01 6.18280292e-01 8.73608887e-01
2.56831288e-01 1.08211726e-01 -8.43514919e-01 -5.92486978e-01
4.17657554e-01 -1.17452228e+00 -1.38912544e-01 7.43035138e-01
-1.20328712e+00 1.14321315e+00 1.92038968e-01 6.17260277e-01
-1.35830328e-01 1.54139683e-01 5.86729348e-01 4.46162939e-01
1.03936088e+00 7.75902092e-01 -1.20423064e-01 -3.88476789e-01
-1.15079224e+00 6.95081592e-01 6.94173157e-01 1.99815109e-01
-1.09735988e-01 1.09874517e-01 -8.30555499e-01 7.02186108e-01
-1.55190423e-01 7.84645319e-01 3.46749842e-01 -6.60909474e-01
9.00147259e-01 1.81579068e-02 2.47022867e-01 -8.54639411e-01
-8.49676967e-01 -6.09959781e-01 -4.39170837e-01 2.94046253e-01
4.84813005e-01 -4.69052792e-01 -1.12111938e+00 1.50381482e+00
-1.20935299e-01 -4.21622485e-01 -4.03934866e-01 7.74402559e-01
7.77715147e-01 5.06573319e-01 3.99664044e-01 -5.00497401e-01
1.28087986e+00 -1.03769195e+00 -7.98398018e-01 -3.53614122e-01
1.18828976e+00 -1.02920711e+00 8.36898983e-01 4.55348372e-01
-8.73596966e-01 -2.43472263e-01 -9.78892028e-01 3.15538794e-02
5.49258515e-02 -2.34473776e-02 3.95492315e-01 7.31470764e-01
-7.22761571e-01 3.87170523e-01 -3.52604449e-01 -2.61350781e-01
5.04900455e-01 -3.84909123e-01 -2.06807226e-01 3.35052580e-01
-1.25443089e+00 1.38578165e+00 1.49687052e-01 -1.05905265e-01
-9.07879531e-01 -5.33775091e-01 -4.86953735e-01 -3.65847081e-01
4.87325713e-03 -3.36832017e-01 1.17671168e+00 -4.21734601e-01
-9.73526299e-01 1.33045900e+00 1.29019424e-01 -6.34419322e-01
6.32753849e-01 -7.14102626e-01 -6.77092731e-01 8.06543753e-02
5.63062251e-01 -5.22490144e-02 1.07376873e+00 -6.56646550e-01
-5.57860374e-01 -2.56773513e-02 -2.49799535e-01 -9.63476370e-04
-1.14930399e-01 8.67868602e-01 6.84056580e-01 -9.92653251e-01
-3.01855862e-01 -5.61141491e-01 8.19737613e-02 -1.11211503e+00
-8.45721066e-01 -5.46144664e-01 4.40626025e-01 -1.16641307e+00
1.30349636e+00 -2.28360319e+00 3.43848020e-01 -5.61014749e-02
4.23206031e-01 3.44368696e-01 -4.80196476e-01 5.96894085e-01
-3.50500077e-01 7.05951333e-01 9.72038880e-02 -4.05214727e-01
-9.88750532e-02 -6.04936838e-01 -8.74315023e-01 7.39641726e-01
4.71662730e-02 6.55261219e-01 -8.85118246e-01 -3.38771045e-01
-2.76237726e-01 -4.09172736e-02 -3.79712224e-01 -1.89951912e-01
-4.54216860e-02 5.20690680e-01 -3.44510168e-01 2.68958449e-01
2.35859141e-01 -9.62096092e-04 -3.18543643e-01 2.24223495e-01
-3.40712577e-01 1.07139540e+00 -2.32341409e-01 1.56067193e+00
-2.50262320e-01 7.95822918e-01 2.64117807e-01 -4.71755207e-01
3.24714243e-01 4.99025911e-01 -1.12892181e-01 -7.69622624e-01
1.87601760e-01 1.68717727e-01 4.52655643e-01 -4.23041403e-01
2.43916944e-01 -2.31020272e-01 -4.94794399e-01 1.08314753e+00
-1.68419093e-01 -1.97603963e-02 4.68108684e-01 7.00382769e-01
1.75891709e+00 -1.93126798e-01 4.37314421e-01 -5.16011119e-01
3.64614516e-01 3.78537565e-01 2.29289621e-01 1.47079170e+00
-4.65689600e-01 4.15123463e-01 6.57032788e-01 -7.50339568e-01
-8.10520232e-01 -1.07597435e+00 -3.12555194e-01 1.42735720e+00
-7.01311588e-01 -9.28419352e-01 -6.55521393e-01 -1.16520166e+00
-4.09730017e-01 9.50393975e-01 -9.36583102e-01 1.71581879e-01
-8.76642346e-01 -7.54643023e-01 1.31958175e+00 1.27637774e-01
3.44604611e-01 -1.29872477e+00 -5.52949369e-01 1.02399044e-01
-4.56609130e-01 -9.79927957e-01 -1.67265952e-01 9.25431550e-02
-2.49533772e-01 -1.30894196e+00 -3.42070043e-01 -8.65739584e-01
-2.75288709e-02 -1.23393916e-01 7.53888905e-01 8.61166418e-02
-2.59583712e-01 -1.20424941e-01 -6.67168319e-01 -2.85109550e-01
-7.49159098e-01 2.72558331e-01 1.73249364e-01 -5.92370570e-01
2.31663242e-01 -2.53199697e-01 8.53166804e-02 -2.89480448e-01
-7.90677741e-02 7.65157267e-02 2.52703160e-01 8.49658966e-01
-3.40270072e-01 -2.89045513e-01 7.56889820e-01 -8.02090168e-01
1.34350824e+00 -1.28425673e-01 -3.36237997e-01 6.64032325e-02
-1.25759289e-01 -1.47453308e-01 4.67867374e-01 -1.13851793e-01
-9.46412921e-01 -5.31369627e-01 -1.93082735e-01 2.62780577e-01
-6.14441782e-02 4.88694429e-01 6.79909647e-01 2.04273418e-01
1.66630185e+00 -1.46418720e-01 -4.91967440e-01 -6.92785800e-01
4.27681446e-01 7.01777935e-01 4.88391578e-01 -4.76176918e-01
9.86375749e-01 5.24296761e-02 -5.32290280e-01 -9.02864873e-01
-1.54960537e+00 -3.18244040e-01 -3.78962874e-01 1.32543355e-01
1.13196695e+00 -8.42711926e-01 -2.40599185e-01 6.98000371e-01
-1.76611233e+00 -5.47030568e-01 9.69287455e-02 4.61298525e-01
-3.29121083e-01 2.88121641e-01 -9.67208087e-01 -9.14163947e-01
-6.89928591e-01 -7.31859326e-01 7.58832455e-01 -4.90182847e-01
-7.77320921e-01 -7.78740823e-01 6.80579960e-01 7.68037915e-01
2.36068785e-01 1.11409076e-01 1.30971920e+00 -9.71227765e-01
3.52895468e-01 -2.87275732e-01 -2.66440302e-01 2.66675204e-01
-4.87360246e-02 -2.85558373e-01 -9.10823822e-01 -1.97895855e-01
2.07436681e-01 -7.81841338e-01 1.48632348e+00 2.62588739e-01
6.69381857e-01 -6.55361712e-01 -2.86744326e-01 4.52089868e-02
6.85507357e-01 -3.06264549e-01 5.66562414e-01 4.30661380e-01
6.10525906e-01 5.62364042e-01 3.92297834e-01 2.43157968e-01
-3.87830019e-01 6.31302357e-01 2.02127337e-01 4.01243448e-01
-5.22502542e-01 -1.27928972e-01 7.77880669e-01 5.63997626e-01
-2.23492905e-01 -4.05322194e-01 -1.17039824e+00 5.49892366e-01
-1.69557405e+00 -1.60205555e+00 -7.81750858e-01 1.80141616e+00
7.48607278e-01 2.50949264e-01 4.13574368e-01 1.38232410e-01
8.16551685e-01 7.50278831e-01 4.39899445e-01 -9.43218768e-01
-1.16755262e-01 5.56550801e-01 2.98087038e-02 9.93625700e-01
-1.59318125e+00 1.25196493e+00 7.97577095e+00 1.07924509e+00
-5.72426558e-01 7.17602551e-01 1.21776111e-01 -6.79459393e-01
4.10464592e-03 -1.83776125e-01 -7.29767323e-01 5.24425864e-01
9.49138880e-01 -5.04008085e-02 3.89197737e-01 2.60449678e-01
1.31668955e-01 -1.88023373e-01 -9.21707749e-01 5.78793108e-01
3.68767828e-01 -1.44471014e+00 -3.98734286e-02 1.29224181e-01
8.13107431e-01 6.00683093e-01 6.50344864e-02 3.99209291e-01
4.14338797e-01 -1.40886104e+00 1.12079346e+00 -1.18884079e-01
6.97974622e-01 -4.53956991e-01 6.97276533e-01 6.30224705e-01
-2.64674723e-01 -5.59143350e-02 1.08312286e-01 -6.58233941e-01
4.45149332e-01 7.28340805e-01 -8.55810642e-01 3.35438669e-01
3.64197314e-01 4.52786982e-01 -7.78983176e-01 7.33278751e-01
-1.28057003e+00 1.25205159e+00 -1.46843418e-01 -1.12723053e-01
4.89466906e-01 6.22109115e-01 1.21988797e+00 1.61938024e+00
-2.76098579e-01 -6.26605004e-02 5.86927347e-02 5.25390685e-01
-1.28282025e-01 1.48866534e-01 -4.29163605e-01 -2.56712556e-01
2.83611625e-01 1.11054516e+00 -2.55338043e-01 -1.45223305e-01
2.50056356e-01 6.39884651e-01 8.51704895e-01 1.72494292e-01
-9.20418978e-01 -5.78300714e-01 1.58387125e-01 1.80049002e-01
-5.59427328e-02 -3.35860133e-01 -5.71174860e-01 -1.05774522e+00
-1.07691668e-01 -8.42223525e-01 5.10905445e-01 -4.46747363e-01
-1.57617998e+00 7.11135507e-01 -7.33270496e-02 -4.96709943e-01
-2.53848195e-01 -6.04244530e-01 -8.77901912e-01 8.26684237e-01
-6.14782631e-01 -1.00377250e+00 3.25090379e-01 3.14166307e-01
4.54757005e-01 -4.46279824e-01 9.97659028e-01 -2.76326966e-02
-3.11089039e-01 3.46265763e-01 -3.57447684e-01 4.39112276e-01
1.14216852e+00 -9.34599161e-01 5.85246801e-01 1.11750102e+00
2.13749826e-01 7.25413978e-01 9.44723547e-01 -9.20540631e-01
-3.06295663e-01 -8.33643019e-01 1.51542902e+00 -9.41373765e-01
1.29205358e+00 -4.76333469e-01 -5.13692737e-01 7.12630093e-01
5.73359728e-01 -7.10388184e-01 5.85877836e-01 9.33480203e-01
-9.75709677e-01 7.74573088e-01 -9.17682171e-01 4.38636541e-01
1.33250368e+00 -7.83413887e-01 -1.21404576e+00 1.16662300e+00
3.77860606e-01 -3.40808243e-01 -3.55585426e-01 7.57611319e-02
5.78956366e-01 -6.57769084e-01 3.91114354e-01 -9.66458321e-01
1.01134789e+00 1.22302502e-01 7.28922933e-02 -1.39685810e+00
-8.71694505e-01 -5.70426285e-01 1.79228142e-01 7.72261083e-01
6.62911057e-01 -4.77176189e-01 2.35515103e-01 -4.44045544e-01
-2.01656684e-01 -2.33841076e-01 -1.47607601e+00 -8.42358887e-01
6.12418175e-01 -4.67345864e-01 -6.45841122e-01 1.00685573e+00
4.72091258e-01 9.64676797e-01 -5.27216792e-01 -1.42762318e-01
7.90131390e-01 -1.10078700e-01 4.59740549e-01 -9.65588152e-01
-4.61218357e-01 -6.25289977e-01 -2.72730291e-01 -4.21195537e-01
3.17531765e-01 -1.12809503e+00 3.49931940e-02 -1.72753155e+00
5.00407338e-01 4.38454933e-02 -2.66750436e-02 8.17235589e-01
-5.36678825e-03 4.10843641e-01 1.89481497e-01 4.65928853e-01
-6.17387354e-01 1.01854958e-01 7.34113634e-01 -4.95701402e-01
-7.17455670e-02 -4.53170091e-01 -8.67512226e-01 7.42702842e-01
9.52758789e-01 -1.06022596e+00 1.36722431e-01 -9.08244193e-01
4.55617934e-01 -4.20880556e-01 6.56895280e-01 -9.00501192e-01
-7.47515634e-02 -1.93426013e-01 2.11948991e-01 -5.50149798e-01
2.07952738e-01 4.84212041e-01 -5.22708178e-01 9.60741520e-01
-3.24908912e-01 -3.33492517e-01 1.49419039e-01 2.90945023e-01
2.38079011e-01 -2.66814590e-01 5.98495543e-01 5.39183319e-02
-7.09497109e-02 -4.70635653e-01 -9.76317704e-01 5.77533126e-01
6.19154990e-01 4.93541628e-01 -1.52244580e+00 -2.08074495e-01
-9.63451624e-01 5.83082736e-02 1.68128967e-01 3.77433330e-01
2.12416589e-01 -1.09117556e+00 -1.42827320e+00 -2.61701316e-01
-8.03527907e-02 -9.02313054e-01 5.32824248e-02 1.20573473e+00
-5.02887189e-01 6.05020344e-01 -3.49194445e-02 -9.79045779e-02
-1.28463626e+00 3.13154131e-01 2.41479222e-02 -7.40206003e-01
-5.95812917e-01 1.06019080e+00 1.39714018e-01 1.83650672e-01
-2.25599453e-01 4.67240423e-01 -2.10621774e-01 5.37973940e-02
1.02654231e+00 6.33256793e-01 1.52132779e-01 -5.34076154e-01
-8.14197481e-01 1.78753272e-01 -6.05025768e-01 -6.50269449e-01
1.08721817e+00 5.52743554e-01 -3.14066648e-01 2.51083851e-01
8.54439020e-01 7.85561085e-01 -2.84674704e-01 2.45607600e-01
2.10559182e-02 2.35501714e-02 1.62456438e-01 -1.57935488e+00
-1.70377102e-02 7.46349514e-01 6.54802397e-02 4.12112951e-01
2.52232909e-01 1.12140805e-01 5.13772249e-01 5.00626504e-01
3.55993092e-01 -1.33535194e+00 1.66167527e-01 1.22476733e+00
1.45703220e+00 -6.68017447e-01 3.69597226e-01 -4.36464161e-01
-3.25706929e-01 6.58246875e-01 2.57460505e-01 -5.40370107e-01
1.53466851e-01 3.34337130e-02 7.12271482e-02 -6.44768298e-01
-8.37680161e-01 1.70265764e-01 1.79500818e-01 6.09111965e-01
6.51651800e-01 1.15837209e-01 -1.18735313e+00 6.27329051e-01
-4.94952261e-01 -5.51481485e-01 4.77507353e-01 9.08009410e-01
-7.97657669e-01 -6.51875198e-01 -3.50596398e-01 3.78407568e-01
-8.09098482e-01 -5.83308399e-01 -1.23105156e+00 6.43889487e-01
7.86588714e-02 1.39604032e+00 -1.50562063e-01 -5.79782724e-01
3.96053456e-02 2.51313448e-01 9.07652318e-01 -1.05974400e+00
-1.14218342e+00 -4.60384674e-02 1.10302329e+00 -5.72887182e-01
-1.96139246e-01 -5.43892026e-01 -8.12088609e-01 -5.19068301e-01
-1.62025020e-01 2.82638788e-01 4.57668990e-01 1.19907653e+00
-6.98063821e-02 3.81123364e-01 2.46072993e-01 -7.58843064e-01
-9.42553282e-01 -1.57133996e+00 -2.87089139e-01 1.19259216e-01
3.79891932e-01 -4.44629580e-01 -7.98499167e-01 -1.89152941e-01] | [8.462806701660156, 10.679157257080078] |
31d6c939-a65b-4312-b98d-78d28d57f543 | oberta-improving-sparse-transfer-learning-via | 2303.17612 | null | https://arxiv.org/abs/2303.17612v3 | https://arxiv.org/pdf/2303.17612v3.pdf | oBERTa: Improving Sparse Transfer Learning via improved initialization, distillation, and pruning regimes | In this paper, we introduce the range of oBERTa language models, an easy-to-use set of language models which allows Natural Language Processing (NLP) practitioners to obtain between 3.8 and 24.3 times faster models without expertise in model compression. Specifically, oBERTa extends existing work on pruning, knowledge distillation, and quantization and leverages frozen embeddings improves distillation and model initialization to deliver higher accuracy on a broad range of transfer tasks. In generating oBERTa, we explore how the highly optimized RoBERTa differs from the BERT for pruning during pre-training and finetuning. We find it less amenable to compression during fine-tuning. We explore the use of oBERTa on seven representative NLP tasks and find that the improved compression techniques allow a pruned oBERTa model to match the performance of BERTbase and exceed the performance of Prune OFA Large on the SQUAD V1.1 Question Answering dataset, despite being 8x and 2x, respectively faster in inference. We release our code, training regimes, and associated model for broad usage to encourage usage and experimentation | ['ChengXiang Zhai', 'Mark Kurtz', 'Alexandre Marques', 'Daniel Campos'] | 2023-03-30 | null | null | null | null | ['model-compression'] | ['methodology'] | [-2.68466901e-02 3.09455395e-01 -4.34906870e-01 -5.01099944e-01
-1.02139354e+00 -7.68564641e-01 6.43932641e-01 3.50247651e-01
-7.90222824e-01 6.41438067e-01 2.23899364e-01 -8.22997928e-01
-4.06907588e-01 -7.33258605e-01 -7.99714029e-01 -1.04659423e-01
-2.91063059e-02 9.47545350e-01 1.41245827e-01 -1.39034614e-01
5.01534678e-02 2.67559201e-01 -1.27858722e+00 4.97515827e-01
6.70832813e-01 9.11062777e-01 1.76620916e-01 1.01079309e+00
-2.11824939e-01 8.22398961e-01 -5.18240035e-01 -6.86894417e-01
2.27060899e-01 1.39416486e-01 -1.27062714e+00 -5.97661436e-01
8.13412368e-01 -4.74334180e-01 -3.35751444e-01 3.91521901e-01
4.43829238e-01 2.22849339e-01 5.85165858e-01 -1.01209795e+00
-9.10355568e-01 1.19844949e+00 -1.87076330e-01 5.36946714e-01
-1.19755557e-02 2.31940776e-01 1.38055301e+00 -7.29537547e-01
5.06222010e-01 1.39229226e+00 7.69085467e-01 4.15161312e-01
-1.42436576e+00 -6.39147460e-01 -1.23462744e-01 1.74364328e-01
-1.39274800e+00 -7.69323826e-01 7.21259266e-02 -1.43419027e-01
1.76046479e+00 1.67221680e-01 4.47530210e-01 9.91824448e-01
-8.01260546e-02 9.55810249e-01 6.26453459e-01 -6.58637047e-01
2.22103730e-01 6.83770627e-02 5.88941216e-01 8.65679443e-01
2.43060634e-01 -1.02148667e-01 -4.23146397e-01 -5.96579015e-01
3.59667391e-01 -3.89163703e-01 -1.75130919e-01 -1.97274178e-01
-1.16781104e+00 9.29133892e-01 4.37304974e-01 1.53101042e-01
-1.95553407e-01 6.30808651e-01 4.42254215e-01 4.17580366e-01
1.57173604e-01 1.03041661e+00 -9.75625694e-01 -5.42613328e-01
-1.12472165e+00 6.63330138e-01 1.17940414e+00 9.83814538e-01
6.76035345e-01 -1.71993941e-01 -3.41996610e-01 9.25477087e-01
8.13758895e-02 1.99943736e-01 6.85924053e-01 -1.48053932e+00
6.60638809e-01 4.33190972e-01 3.51727307e-02 -5.92216313e-01
-2.68779695e-01 -4.62697089e-01 -4.35869694e-01 -3.55542183e-01
5.80285370e-01 -1.72820643e-01 -1.10531938e+00 1.66829133e+00
-4.27012555e-02 2.54745572e-03 5.79834990e-02 3.40784132e-01
6.50162518e-01 8.04610133e-01 3.16842020e-01 2.23487616e-01
1.48182201e+00 -1.16744912e+00 -4.03629303e-01 -5.65994799e-01
1.12989390e+00 -6.34699285e-01 1.53051603e+00 5.27678490e-01
-1.21685982e+00 -5.50679505e-01 -1.07018971e+00 -6.90700233e-01
-5.17232299e-01 5.19179367e-02 9.48438764e-01 4.94965285e-01
-9.29857552e-01 7.13648379e-01 -1.05229294e+00 -3.86536270e-01
4.91287291e-01 1.85368061e-01 -1.79095194e-01 -2.84831703e-01
-1.32738340e+00 1.37919784e+00 7.27989495e-01 -2.34559566e-01
-7.82742560e-01 -1.15447438e+00 -7.16152728e-01 2.21409231e-01
2.93114007e-01 -9.86748576e-01 1.65597427e+00 -2.02013895e-01
-1.27255726e+00 6.59263551e-01 -2.40952402e-01 -1.10419071e+00
6.52895644e-02 -5.37198365e-01 -5.79602793e-02 1.55368000e-01
-3.84498268e-01 1.18574202e+00 4.08519804e-01 -6.04235530e-01
-3.74662340e-01 4.88698483e-02 2.97999233e-01 1.22321062e-01
-4.05682623e-01 -1.07382998e-01 -6.36139333e-01 -3.65428716e-01
-3.44579279e-01 -6.00720823e-01 -2.42365412e-02 -4.97926809e-02
-2.44564697e-01 -5.72173178e-01 3.37837219e-01 -5.29733777e-01
1.50309539e+00 -1.79554093e+00 -8.29514116e-02 1.84940733e-02
3.48380417e-01 5.36492407e-01 -4.54096258e-01 4.63193148e-01
2.18304083e-01 5.30820668e-01 -2.76857525e-01 -4.92960155e-01
2.32972339e-01 6.87611461e-01 -5.09203792e-01 3.26377563e-02
2.49298677e-01 1.23106027e+00 -7.45890677e-01 -5.85437834e-01
-5.39831929e-02 3.75119627e-01 -1.01071084e+00 9.81262922e-02
-6.69646680e-01 -4.29344445e-01 -3.16549510e-01 5.92647254e-01
3.39621603e-01 -4.57860202e-01 -1.18536003e-01 -1.59592837e-01
2.20640868e-01 1.02763164e+00 -7.17729926e-01 1.83301711e+00
-6.12770438e-01 6.16505682e-01 -5.52834608e-02 -6.13473952e-01
5.91760516e-01 6.27078488e-02 -8.25927109e-02 -3.55139881e-01
2.11082567e-02 1.25780851e-01 2.10120648e-01 -5.05349338e-01
6.29082561e-01 6.47328421e-02 1.72790796e-01 6.86552584e-01
3.88628960e-01 -4.18194085e-01 5.63388705e-01 3.93149525e-01
1.36611712e+00 -7.88571239e-02 2.18448371e-01 -3.88229102e-01
8.66567269e-02 2.01805457e-01 2.19916314e-01 1.02327669e+00
-1.26887694e-01 3.27100962e-01 4.51616615e-01 -3.41361552e-01
-1.07655060e+00 -1.03392029e+00 -4.88664955e-01 1.59015906e+00
-4.48982388e-01 -8.91670644e-01 -6.82595253e-01 -6.32909000e-01
3.53147715e-01 1.14051044e+00 -6.29374504e-01 -2.59675354e-01
-6.83550000e-01 -7.11288333e-01 1.15610778e+00 6.00699961e-01
3.95738304e-01 -9.60289121e-01 -4.99416322e-01 2.05196410e-01
-2.00936347e-01 -8.97814929e-01 -2.73619890e-01 4.41537589e-01
-9.50833321e-01 -8.89802516e-01 -2.72030681e-01 -6.19471550e-01
1.17735967e-01 -4.70687635e-02 1.66445827e+00 1.78557053e-01
-1.50814667e-01 2.32145071e-01 -3.24844509e-01 -3.97238910e-01
-5.58212399e-01 7.87689388e-01 -1.02085270e-01 -9.38388288e-01
8.33139598e-01 -5.34809232e-01 -3.38180035e-01 -1.56691805e-01
-9.31267262e-01 1.00375846e-01 6.64746583e-01 8.15114439e-01
5.69920838e-01 -1.60179064e-01 4.09841537e-01 -1.01216900e+00
9.91454959e-01 -6.37659609e-01 -4.39638466e-01 5.07708013e-01
-8.92129600e-01 5.38929105e-01 5.39038181e-01 -4.56594676e-01
-6.32941902e-01 -4.25544262e-01 -3.72652054e-01 -3.69257867e-01
-1.39727350e-02 6.20274544e-01 2.55148590e-01 1.25311986e-01
1.15636277e+00 -6.49084896e-02 -1.01868846e-01 -7.61199832e-01
9.09404278e-01 6.62719429e-01 4.05952662e-01 -8.93736720e-01
5.24147570e-01 8.94941855e-03 -5.87689400e-01 -5.27678013e-01
-1.13095105e+00 -3.15378696e-01 -3.40961933e-01 7.40743935e-01
4.91893053e-01 -9.29245949e-01 -4.89353478e-01 -1.29636645e-01
-1.14796913e+00 -6.70048535e-01 -5.71580768e-01 3.01084667e-01
-5.41983485e-01 1.85032830e-01 -1.02784133e+00 -3.96929711e-01
-7.76618481e-01 -9.52449262e-01 9.57091749e-01 -8.46634284e-02
-5.84511757e-01 -1.03759873e+00 2.79450119e-01 5.46505213e-01
6.86183274e-01 -3.70446742e-01 1.43325126e+00 -8.45110118e-01
-3.58977169e-01 2.45871185e-03 -2.63880044e-01 4.90351886e-01
-3.45069945e-01 -6.04584161e-03 -9.76816118e-01 -3.31970960e-01
-3.13668102e-01 -8.64272475e-01 1.10928178e+00 1.51652336e-01
1.36116898e+00 -4.16335970e-01 -2.05554649e-01 8.25059414e-01
1.22913957e+00 -2.18683317e-01 4.75967646e-01 5.17971456e-01
3.50751340e-01 2.37072021e-01 2.15080634e-01 1.48744151e-01
5.46495318e-01 4.10256654e-01 1.81871146e-01 3.97291750e-01
-6.92617074e-02 -6.21234596e-01 2.80193567e-01 1.01852393e+00
3.64795715e-01 -2.82242507e-01 -1.18604267e+00 8.26530993e-01
-1.50404000e+00 -7.08098114e-01 4.14389104e-01 1.83405709e+00
1.52070725e+00 3.57449591e-01 -1.23632714e-01 -1.78025886e-01
8.96567777e-02 -5.44126560e-05 -4.69974726e-01 -7.21217036e-01
1.38012365e-01 7.00267315e-01 5.59073627e-01 8.90310466e-01
-8.69355381e-01 1.24532521e+00 7.12764072e+00 1.06792414e+00
-5.90662956e-01 1.98296070e-01 7.29526401e-01 -4.48883384e-01
-3.53089571e-01 9.25680026e-02 -1.29133201e+00 1.78715706e-01
1.60396266e+00 -2.94746190e-01 6.99035466e-01 9.66979444e-01
-2.87298530e-01 4.48647663e-02 -1.41837668e+00 7.83647001e-01
-1.32485524e-01 -1.53616655e+00 1.93829075e-01 -2.74348229e-01
4.64117706e-01 6.01614833e-01 -1.56822532e-01 1.07106233e+00
7.15345919e-01 -1.35836315e+00 4.68907773e-01 3.07820976e-01
5.05682290e-01 -5.46506226e-01 5.68164289e-01 6.29251182e-01
-6.58096433e-01 -2.09893331e-01 -7.43037581e-01 -1.15557961e-01
1.50504559e-01 4.60167557e-01 -1.12229705e+00 7.79774366e-03
6.25248432e-01 2.01601550e-01 -9.47200000e-01 8.43432009e-01
-1.20890245e-01 1.13642871e+00 -8.02896798e-01 -1.85614660e-01
3.87689799e-01 2.38010570e-01 1.06794290e-01 1.53586161e+00
1.26454066e-02 -7.70247653e-02 -8.58126506e-02 9.09213424e-01
-4.29336160e-01 3.08574568e-02 -3.02535862e-01 -3.53061676e-01
9.65731084e-01 1.05482173e+00 3.81095707e-02 -6.96901858e-01
-8.32751319e-02 4.92466241e-01 8.63827705e-01 2.55155325e-01
-8.12032342e-01 -5.61740279e-01 6.83161080e-01 1.63097173e-01
5.07011056e-01 -3.48267764e-01 -3.01298618e-01 -1.20840323e+00
-1.06295787e-01 -1.09823847e+00 5.05281508e-01 -8.88360322e-01
-1.22133505e+00 6.34090245e-01 5.04929364e-01 -2.85628021e-01
-5.18448353e-01 -7.02940524e-01 -2.68964231e-01 9.73133445e-01
-1.56491542e+00 -8.96681666e-01 5.53755760e-02 1.66714102e-01
4.03372914e-01 -3.08897495e-02 9.97283518e-01 1.80717498e-01
-4.51122820e-01 9.40611005e-01 2.18357101e-01 6.91456050e-02
5.97646773e-01 -1.22280335e+00 7.54156232e-01 5.73399425e-01
4.11917001e-01 1.19635522e+00 6.70053363e-01 -2.94592172e-01
-1.36759949e+00 -1.05234599e+00 1.28647912e+00 -8.19007277e-01
9.33061540e-01 -4.81913567e-01 -9.72661376e-01 1.05193293e+00
2.98128873e-01 -2.35217780e-01 6.61296308e-01 5.34470081e-01
-6.95872843e-01 -1.00435853e-01 -1.12115765e+00 5.99569738e-01
9.07594502e-01 -6.05477512e-01 -1.05596614e+00 4.47610378e-01
1.03230000e+00 -3.83566409e-01 -1.15159714e+00 2.99738556e-01
4.55364913e-01 -5.37766755e-01 1.03728187e+00 -9.20300841e-01
4.39233720e-01 6.57033175e-02 -2.26037994e-01 -1.12591815e+00
-4.56059664e-01 -5.57523608e-01 -7.25748360e-01 1.03405559e+00
7.32866824e-01 -6.13411784e-01 8.73762548e-01 7.88912117e-01
1.07912049e-01 -1.18554509e+00 -8.20021808e-01 -5.98740280e-01
6.98708236e-01 -7.51440942e-01 7.26741672e-01 7.59496033e-01
-5.79818431e-03 5.65740466e-01 1.17706418e-01 -2.05107555e-01
3.64690930e-01 3.18026803e-02 7.78403640e-01 -8.69056284e-01
-9.21342432e-01 -5.15173376e-01 1.23591289e-01 -1.70946002e+00
-1.32948821e-02 -1.13509738e+00 -1.38968617e-01 -1.44882309e+00
1.15502231e-01 -7.84213543e-01 -1.80483073e-01 9.34459269e-01
-1.38427928e-01 4.45189476e-02 2.37758026e-01 1.59124613e-01
-5.30563653e-01 4.87674385e-01 9.37666535e-01 -7.11480975e-02
-7.33110309e-02 -3.56432378e-01 -1.18266070e+00 5.66344917e-01
8.80866885e-01 -4.74934578e-01 -6.97875440e-01 -1.16265130e+00
4.03295040e-01 -2.56729722e-01 8.61778930e-02 -8.95348549e-01
2.07535028e-01 9.50980633e-02 2.66024917e-01 -4.56201404e-01
5.20684361e-01 -4.32689905e-01 -2.76312470e-01 2.82119542e-01
-8.77539515e-01 3.02715451e-01 5.16038716e-01 2.47211084e-01
-1.11621469e-02 -6.24260068e-01 6.71439528e-01 -2.21189663e-01
-3.57507318e-01 2.73765206e-01 -2.21164227e-01 5.30639231e-01
3.92765135e-01 1.08607344e-01 -6.41373575e-01 -3.61931473e-01
-4.56773520e-01 5.29535890e-01 2.33386144e-01 3.77560347e-01
3.30072254e-01 -1.06213200e+00 -6.93479180e-01 1.16180457e-01
4.56690192e-02 2.85367727e-01 -1.12518020e-01 5.87252975e-01
-8.06246340e-01 8.94851387e-01 2.76517004e-01 -2.76441753e-01
-9.19558406e-01 3.06225866e-01 3.82548869e-01 -7.76266932e-01
-5.17979920e-01 1.23081040e+00 -2.28845939e-01 -7.47643113e-01
4.67878193e-01 -9.86113250e-01 9.83659402e-02 -1.93327904e-01
7.44305193e-01 3.12166482e-01 2.28183880e-01 9.75095555e-02
-7.26619139e-02 1.34140598e-02 -5.80016077e-01 -2.18712073e-02
1.29391646e+00 1.50154904e-01 2.01254059e-02 3.06812435e-01
1.36022866e+00 -1.45387992e-01 -1.08470142e+00 -3.79922062e-01
9.86672491e-02 -1.50749773e-01 2.42095202e-01 -1.10666931e+00
-5.94591975e-01 9.45349813e-01 2.03053191e-01 8.45102742e-02
7.95065045e-01 -7.59546980e-02 1.15358102e+00 1.15418184e+00
2.72645384e-01 -8.83613527e-01 -4.12196428e-01 8.61420155e-01
8.01101089e-01 -8.11659157e-01 3.06977123e-01 8.77581686e-02
-3.34427416e-01 9.58992124e-01 5.46996117e-01 -1.02735177e-01
7.25917876e-01 3.57701689e-01 -1.44374922e-01 -4.35105488e-02
-1.55656064e+00 2.54331492e-02 6.63044825e-02 4.66876715e-01
5.34261763e-01 -1.52348941e-02 1.84186287e-02 6.35132492e-01
-8.42302620e-01 1.58830807e-01 1.88476235e-01 8.78077924e-01
-6.01888657e-01 -1.09918463e+00 2.61897836e-02 7.67048061e-01
-5.51677704e-01 -6.71450675e-01 -1.23656802e-01 1.06828332e+00
1.20692320e-01 7.28707194e-01 2.52698481e-01 -2.92754143e-01
2.42447227e-01 5.05152404e-01 7.03895092e-01 -7.38647759e-01
-7.26066172e-01 -5.70129514e-01 4.45189685e-01 -4.49251443e-01
1.23096816e-01 -3.70943338e-01 -1.10113251e+00 -7.90937781e-01
-4.59130853e-01 2.94099987e-01 5.95267713e-01 8.80687118e-01
7.79507756e-01 9.82245058e-02 -3.01819086e-01 -4.49366987e-01
-1.06869924e+00 -1.35244024e+00 -9.63668674e-02 2.69035906e-01
2.30851904e-01 -4.39156711e-01 -4.11080539e-01 -3.43942568e-02] | [10.808128356933594, 8.08434009552002] |
9e18b825-a309-4411-8496-1e7d018226fd | unimelb-at-semeval-2019-task-12-multi-model | null | null | https://aclanthology.org/S19-2231 | https://aclanthology.org/S19-2231.pdf | UniMelb at SemEval-2019 Task 12: Multi-model combination for toponym resolution | This paper describes our submission to SemEval-2019 Task 12 on toponym resolution over scientific articles. We train separate NER models for toponym detection over text extracted from tables vs. text from the body of the paper, and train another auxiliary model to eliminate misdetected toponyms. For toponym disambiguation, we use an SVM classifier with hand-engineered features. The best setting achieved a strict micro-F1 score of 80.92{\%} and overlap micro-F1 score of 86.88{\%} in the toponym detection subtask, ranking 2nd out of 8 teams on F1 score. For toponym disambiguation and end-to-end resolution, we officially ranked 2nd and 3rd, respectively. | ['Martin Tomko', 'Timothy Baldwin', 'Minghan Wang', 'Haonan Li', 'Maria Vasardani'] | 2019-06-01 | null | null | null | semeval-2019-6 | ['toponym-resolution'] | ['natural-language-processing'] | [ 4.61048586e-03 3.23152930e-01 -2.33593345e-01 -5.75398803e-02
-1.15877581e+00 -9.01838779e-01 6.57786250e-01 1.06900275e+00
-8.97825181e-01 1.06151366e+00 -3.89762111e-02 2.58941855e-03
-4.39338237e-01 -4.82225060e-01 -6.24390960e-01 -2.00365275e-01
2.79900640e-01 9.53982413e-01 -5.14217429e-02 -1.11191608e-01
4.92233753e-01 2.64801234e-01 -1.14671564e+00 5.54402530e-01
1.09565306e+00 7.07436740e-01 1.13758989e-01 4.89996254e-01
-2.53021389e-01 -3.62026542e-02 -1.15009761e+00 -9.44612741e-01
7.97965080e-02 9.07244682e-02 -7.40252674e-01 -9.99700367e-01
1.10084248e+00 5.18644273e-01 -1.07247785e-01 1.24648869e+00
4.76630837e-01 -9.13661942e-02 9.16634917e-01 -9.40017223e-01
-3.56082916e-01 1.06106722e+00 -5.63155353e-01 6.86018229e-01
7.61708558e-01 -4.96526241e-01 1.48379171e+00 -1.28395116e+00
1.02632678e+00 9.50509250e-01 5.69791973e-01 4.12848890e-01
-1.29052258e+00 -1.11109781e+00 -5.07792011e-02 -1.41898543e-01
-1.58413577e+00 -1.31151572e-01 3.90122086e-02 -5.68430722e-01
1.27505791e+00 3.01014841e-01 2.32913747e-01 1.16259682e+00
4.03029770e-01 1.92881301e-01 9.25255060e-01 -5.68303883e-01
1.05845675e-01 2.68188715e-01 5.89089572e-01 4.09621060e-01
9.62060511e-01 -4.63239551e-01 -8.74807358e-01 -5.19928515e-01
4.55848902e-01 -3.74702543e-01 -1.17184624e-01 1.91633120e-01
-1.45349431e+00 4.90844399e-01 2.14225188e-01 3.11660618e-01
-2.93882251e-01 -4.99741971e-01 4.27603662e-01 1.63792267e-01
4.70243441e-03 1.47954726e+00 -7.94152021e-01 1.82219408e-02
-1.19252002e+00 4.16435927e-01 7.95932949e-01 1.24967647e+00
4.30242747e-01 -6.08142138e-01 -5.44398785e-01 1.08520699e+00
-2.74460822e-01 4.86630678e-01 5.38337767e-01 -7.66553342e-01
6.90849662e-01 6.11955285e-01 2.15075284e-01 -7.19502449e-01
-8.90524387e-01 -9.20944095e-01 -6.70743048e-01 -3.83674502e-01
4.74080801e-01 -1.21981278e-01 -9.24722195e-01 1.64129174e+00
-2.37099811e-01 -3.00965726e-01 1.90414526e-02 5.04794657e-01
1.46088302e+00 2.83111334e-01 3.70974600e-01 -2.54839122e-01
1.82049251e+00 -6.64683938e-01 -7.49516904e-01 -1.89005151e-01
8.02637696e-01 -9.60476637e-01 8.90155792e-01 6.39628589e-01
-9.47554946e-01 -3.97855490e-01 -1.20724845e+00 -7.63808414e-02
-8.48169982e-01 5.13547301e-01 3.10820699e-01 4.51291114e-01
-7.26764083e-01 9.25555587e-01 -3.21062207e-01 -4.77630824e-01
1.09473310e-01 5.29028833e-01 -6.06393099e-01 1.39651164e-01
-1.67069399e+00 1.15161693e+00 7.56833851e-01 -5.00931561e-01
-2.48247907e-01 -1.37449527e+00 -4.88622934e-01 5.31677082e-02
9.36148167e-02 -8.63219321e-01 7.68032074e-01 -1.15555368e-01
-8.31926286e-01 1.45285952e+00 8.64419341e-03 -4.18700457e-01
3.05042475e-01 -5.61780214e-01 -9.74685550e-01 -1.04145862e-01
8.15525115e-01 5.14628947e-01 8.76487717e-02 -7.29579985e-01
-7.75788546e-01 -3.17353666e-01 -3.18157524e-01 2.77506113e-01
-3.21856081e-01 2.39173263e-01 -5.28580785e-01 -7.63673544e-01
1.60717174e-01 -8.06597471e-01 1.07236050e-01 -7.39517391e-01
-7.32536137e-01 -5.53282261e-01 -1.80165395e-02 -8.13181102e-01
1.55955327e+00 -1.75518584e+00 3.11927497e-02 3.97364050e-01
6.47869706e-01 3.00667077e-01 -5.46049364e-02 3.44145507e-01
-6.15943670e-01 3.68270725e-01 1.08671635e-02 -1.62690550e-01
-5.02122082e-02 -4.16091919e-01 -3.83541048e-01 4.12029028e-02
2.44642690e-01 5.82020640e-01 -1.03626752e+00 -3.70669693e-01
-1.26638114e-01 2.16316804e-01 -1.41371131e-01 -3.11043620e-01
9.56919044e-02 -1.09541811e-01 -4.51590657e-01 5.57074964e-01
5.66912532e-01 -3.15429986e-01 1.89986184e-01 -1.36008725e-01
1.62091196e-01 5.41517794e-01 -1.41966867e+00 1.71870971e+00
-5.22516906e-01 5.03398120e-01 -2.18975231e-01 -5.81862986e-01
1.03936267e+00 1.61025912e-01 3.05328459e-01 -5.83978236e-01
-2.46020764e-01 6.73814714e-01 -9.73755270e-02 4.32810821e-02
5.71678042e-01 2.09639430e-01 -5.24449229e-01 -3.29662442e-01
3.92304093e-01 7.71803632e-02 6.28045559e-01 5.93485951e-01
1.10329139e+00 -2.07027629e-01 5.29493511e-01 -6.94230020e-01
6.12432122e-01 7.63513818e-02 5.97264111e-01 1.11957967e+00
6.23047759e-04 7.24256575e-01 6.60237074e-01 -5.37021697e-01
-7.32185125e-01 -1.06271803e+00 -6.89611435e-01 1.09111595e+00
-1.26404420e-01 -1.23088789e+00 -6.49117649e-01 -8.41193497e-01
2.96754181e-01 1.10074878e+00 -5.34832776e-01 -1.64167970e-01
-6.16855383e-01 -8.83957028e-01 8.49198639e-01 2.76166856e-01
1.57601386e-02 -7.09274292e-01 -3.30327392e-01 -5.37091643e-02
-3.43800545e-01 -1.43102503e+00 -3.93356860e-01 7.50589252e-01
-4.26936924e-01 -1.06706226e+00 -6.84825540e-01 -6.59874856e-01
2.40893245e-01 -2.44098693e-01 1.50744343e+00 -3.33356082e-01
-3.88942361e-01 -3.29049796e-01 -4.86243106e-02 -6.08816862e-01
-3.03757209e-02 5.33230126e-01 4.64481711e-01 -8.32275987e-01
8.77378404e-01 -2.40305007e-01 -2.62215674e-01 2.78934129e-02
-2.50143230e-01 -4.17325884e-01 5.89222789e-01 9.50147510e-01
7.65370667e-01 -4.80461180e-01 6.55315101e-01 -1.11264944e+00
5.80654800e-01 -2.47500032e-01 -6.05772197e-01 4.08285260e-01
-1.00542712e+00 1.60935506e-01 9.43135917e-01 -3.65962721e-02
-5.24841011e-01 -3.90738621e-02 6.39003590e-02 -1.86475724e-01
-7.60604069e-02 4.38529819e-01 -3.13004911e-01 3.97643358e-01
1.05434465e+00 -9.07735974e-02 -8.95525694e-01 -6.87631011e-01
2.96169460e-01 6.73652053e-01 7.81662881e-01 -5.82843244e-01
5.45500159e-01 -2.98372597e-01 1.94116488e-01 -5.96486628e-01
-1.21607208e+00 -6.34723425e-01 -6.05461478e-01 6.07693493e-01
8.56890261e-01 -1.29103076e+00 -6.01877987e-01 -1.76905766e-01
-1.30371988e+00 4.47999448e-01 2.21046917e-02 5.87087154e-01
5.87777793e-03 2.16303930e-01 -1.85175970e-01 -2.61634775e-02
-7.92650580e-01 -8.93643975e-01 1.20879793e+00 5.12689531e-01
-1.04801393e+00 -6.33158565e-01 2.26609692e-01 3.85581821e-01
-5.68965562e-02 1.54810354e-01 1.00134099e+00 -1.40407217e+00
2.21679181e-01 -4.94387329e-01 -3.60083550e-01 -3.26298326e-01
-3.91048864e-02 1.53509974e-02 -1.00435317e+00 -4.31853831e-02
-8.21318805e-01 -5.76236136e-02 1.26277089e+00 2.90071756e-01
1.18124950e+00 6.62578791e-02 -9.21504736e-01 6.42523766e-01
1.08491755e+00 -4.01237756e-02 1.31967321e-01 6.35855973e-01
6.01609349e-01 3.52136225e-01 5.73324680e-01 4.03704554e-01
-1.14755407e-02 1.09846532e+00 -8.31190124e-02 2.31110618e-01
-4.56519909e-02 -3.81364733e-01 -1.30401909e-01 2.25737169e-01
3.06096226e-01 -1.29514143e-01 -1.24674916e+00 6.02119386e-01
-1.42239797e+00 -7.83985674e-01 -4.69110489e-01 2.41996241e+00
1.07167554e+00 4.81949687e-01 1.57422274e-01 -2.02474549e-01
9.05722141e-01 -1.37207612e-01 -9.98608023e-02 -6.12326264e-01
-5.89801967e-01 5.73662221e-01 7.07272708e-01 4.51578140e-01
-1.52032232e+00 1.21038425e+00 6.10103369e+00 1.08364499e+00
-6.41146421e-01 -2.87461191e-01 3.86647224e-01 -2.87916750e-01
-2.33449802e-01 3.89076024e-02 -1.43594754e+00 5.79329014e-01
9.63352740e-01 -3.66452426e-01 2.51433570e-02 6.89752221e-01
-3.91890764e-01 1.93595886e-01 -1.27303362e+00 1.31059253e+00
-1.00601591e-01 -1.42864752e+00 2.70420521e-01 -2.34457374e-01
7.85088301e-01 4.06336375e-02 -1.42788097e-01 3.91437024e-01
4.18840647e-01 -1.35158157e+00 4.25289184e-01 4.34256792e-01
1.07107115e+00 -8.21148217e-01 9.09420550e-01 1.06145285e-01
-1.02581131e+00 2.02807561e-01 -5.19866765e-01 3.21389556e-01
-3.63841802e-01 1.23935115e+00 -7.30487823e-01 9.28481758e-01
9.97393370e-01 7.86363542e-01 -8.17879140e-01 1.05789471e+00
-4.32287335e-01 1.47937939e-01 -5.49358904e-01 -5.91561384e-02
-2.64466465e-01 2.89621621e-01 8.78655791e-01 1.82509577e+00
1.19849041e-01 -2.19839185e-01 -1.59857348e-02 7.55526960e-01
-5.77515662e-01 3.45430911e-01 -2.99387753e-01 2.07464918e-02
1.07039118e+00 1.51165891e+00 -6.42076015e-01 -4.22828913e-01
1.17708873e-02 8.76356006e-01 3.85186017e-01 1.17110893e-01
-7.38761485e-01 -1.28880227e+00 4.67020214e-01 -3.95673215e-01
3.74071226e-02 4.67501074e-01 -8.31097424e-01 -1.19923425e+00
5.31621985e-02 -8.22039187e-01 1.06414914e+00 -3.54706794e-01
-1.35463822e+00 6.29392385e-01 -1.51212782e-01 -9.73114669e-01
8.15171748e-02 -8.64599705e-01 -4.78096545e-01 1.16164243e+00
-9.30570483e-01 -7.43308902e-01 -1.93593409e-02 -6.16237298e-02
1.82302743e-01 -4.47671980e-01 1.14636779e+00 3.80598396e-01
-6.93613529e-01 1.32874048e+00 2.73189396e-01 2.49670967e-01
1.51323164e+00 -1.72739732e+00 4.12294120e-01 6.30565882e-01
2.57719934e-01 1.26582634e+00 9.03559208e-01 -9.81998146e-01
-9.41620946e-01 -1.04329717e+00 1.78129601e+00 -1.13081849e+00
5.91287971e-01 -3.25597048e-01 -6.38851106e-01 4.40410048e-01
1.42170966e-01 -3.34880829e-01 8.26669991e-01 6.19634330e-01
-7.94804692e-01 -1.48756793e-02 -9.55015361e-01 4.70867872e-01
8.58448207e-01 -4.61550564e-01 -9.96025741e-01 5.28624058e-01
6.38725936e-01 -6.39550686e-01 -1.24854159e+00 3.61705661e-01
6.71845019e-01 -9.58998278e-02 1.28491282e+00 -8.77900064e-01
7.39458919e-01 -2.79233217e-01 -1.59015581e-01 -1.22367656e+00
-5.85626006e-01 -5.54853499e-01 9.55004171e-02 1.10716164e+00
1.11392057e+00 -2.10844532e-01 4.57247704e-01 2.52644092e-01
-2.91927785e-01 -5.06060123e-01 -8.99153590e-01 -8.91346037e-01
6.38251126e-01 2.31563494e-01 2.79407978e-01 1.17546427e+00
4.19066519e-01 7.14759171e-01 1.95603315e-02 1.27140164e-01
5.31133294e-01 1.01829931e-01 3.51025432e-01 -1.69506598e+00
1.66676147e-03 -9.01865780e-01 -7.12230727e-02 -3.26176673e-01
1.41815603e-01 -1.33410394e+00 -3.91067296e-01 -1.59285605e+00
4.69252557e-01 2.05509886e-01 -9.02355731e-01 6.02561474e-01
-5.93030691e-01 1.25150457e-01 -6.23906497e-04 1.61300436e-01
-7.31350183e-01 -3.44132818e-02 5.43489277e-01 -4.29581106e-02
-1.32500097e-01 -4.72866923e-01 -1.02864408e+00 4.79482770e-01
4.93060499e-01 -8.54084790e-01 1.22489564e-01 -2.55897492e-01
6.01480305e-01 -2.76551902e-01 9.40434933e-02 -8.20870578e-01
1.18677318e-01 -1.27507076e-01 8.62020910e-01 -5.27952671e-01
-3.07948124e-02 -9.57650617e-02 -1.34046748e-01 4.84360039e-01
-7.18390286e-01 4.06714082e-01 4.84553725e-01 3.25036496e-01
1.08976904e-02 -1.09346911e-01 6.14362180e-01 -1.14158913e-02
-3.43091428e-01 -2.32565716e-01 -2.11892396e-01 6.04212642e-01
4.92264479e-01 -2.99322442e-03 -8.33260715e-01 3.47852677e-01
-8.20373714e-01 4.10533875e-01 2.75435388e-01 5.94343960e-01
1.96700886e-01 -9.40044224e-01 -9.73029673e-01 -2.27885246e-01
7.52488256e-01 -4.98820990e-01 -9.00483206e-02 7.54614532e-01
-3.85504007e-01 8.42347920e-01 -2.44196370e-01 -1.64926633e-01
-1.30286455e+00 4.26457793e-01 1.38120025e-01 -7.23825216e-01
-4.68476027e-01 1.33594131e+00 1.43586874e-01 -5.73570192e-01
2.82707274e-01 -7.77583793e-02 -5.10222316e-01 4.43347067e-01
7.02669263e-01 4.62862730e-01 6.64671957e-01 -2.43457451e-01
-1.16947854e+00 7.33725607e-01 -5.52311897e-01 6.80038123e-04
1.01279402e+00 4.04476672e-01 -1.06893323e-01 1.18355915e-01
1.00814641e+00 4.41655725e-01 5.92368562e-03 -8.47253278e-02
5.34850657e-01 5.05651580e-03 2.04277094e-02 -1.49181974e+00
-4.33661699e-01 5.68021715e-01 4.62217867e-01 -2.07278848e-01
5.79374075e-01 -4.34496440e-02 3.40763211e-01 8.10824811e-01
-2.22390704e-02 -1.33155775e+00 -3.61605257e-01 9.36946869e-01
8.58791649e-01 -1.07405150e+00 3.50131094e-01 -4.73305374e-01
-4.44185287e-01 1.13133836e+00 6.59038067e-01 -1.64219048e-02
3.25804293e-01 1.95336029e-01 -1.40169784e-01 -2.09020987e-01
-6.62592113e-01 2.85356939e-01 8.33662510e-01 4.05770183e-01
7.26678431e-01 1.07763767e-01 -9.11297381e-01 1.47012496e+00
-6.00863636e-01 -3.26661706e-01 4.27693069e-01 2.89353669e-01
-3.72930646e-01 -1.05978751e+00 -2.01717570e-01 7.98105836e-01
-8.23093295e-01 -5.71110606e-01 -9.76179063e-01 6.22204900e-01
1.48825318e-01 7.16648638e-01 -1.41995758e-01 -4.90509748e-01
4.87996042e-01 3.21729541e-01 3.08780342e-01 -7.21785128e-01
-8.84229124e-01 -1.17551656e-02 4.78136867e-01 -5.04651964e-01
3.10481519e-01 -5.73704422e-01 -1.36882329e+00 6.09271117e-02
-1.29057214e-01 5.58433950e-01 3.28186661e-01 9.24464345e-01
6.12644613e-01 4.96417880e-01 1.35661647e-01 8.07220563e-02
-4.88500714e-01 -9.45185304e-01 -4.67611223e-01 3.37481588e-01
2.54462403e-03 -7.24989414e-01 -1.76894173e-01 -2.77221709e-01] | [9.208452224731445, 9.114164352416992] |
39da45a8-6d77-403a-b749-89d654667db4 | arguably-at-comma-icon-detection-of | null | null | https://aclanthology.org/2021.icon-multigen.7 | https://aclanthology.org/2021.icon-multigen.7.pdf | ARGUABLY at ComMA@ICON: Detection of Multilingual Aggressive, Gender Biased, and Communally Charged Tweets Using Ensemble and Fine-Tuned IndicBERT | The proliferation in Social Networking has increased offensive language, aggression, and hate-speech detection, which has drawn the focus of the NLP community. However, people’s difference in perception makes it difficult to distinguish between acceptable content and aggressive/hateful content, thus making it harder to create an automated system. In this paper, we propose multi-class classification techniques to identify aggressive and offensive language used online. Two main approaches have been developed for the classification of data into aggressive, gender-biased, and communally charged. The first approach is an ensemble-based model comprising of XG-Boost, LightGBM, and Naive Bayes applied on vectorized English data. The data used was obtained using an Indic Transliteration on the original data comprising of Meitei, Bangla, Hindi, and English language. The second approach is a BERT-based architecture used to detect misogyny and aggression. The proposed model employs IndicBERT Embeddings to define contextual understanding. The results of the models are validated on the ComMA v 0.2 dataset. | ['Jatin Bedi', 'Prabsimran Kaur', 'Guneet Kohli'] | null | null | null | null | icon-2021-12 | ['transliteration'] | ['natural-language-processing'] | [-6.04330003e-01 -3.26396455e-03 1.49245560e-01 -4.63086069e-01
1.16265461e-01 -3.92020941e-01 8.15462708e-01 4.09330457e-01
-6.47579432e-01 6.51347399e-01 5.92399716e-01 5.38669433e-03
-2.38917246e-01 -7.21611202e-01 2.29722574e-01 -4.22217339e-01
2.01046497e-01 4.16136295e-01 -1.44680053e-01 -7.72774220e-01
7.34773397e-01 3.00828964e-01 -1.32008898e+00 2.11981863e-01
9.05572474e-01 4.67418790e-01 -2.92675465e-01 7.75753438e-01
-5.32953680e-01 1.15489483e+00 -8.90989900e-01 -8.45047355e-01
8.23327973e-02 -5.25900245e-01 -8.56713533e-01 -2.87091702e-01
2.30933487e-01 -2.88427830e-01 -2.27560624e-01 1.18447530e+00
7.54235387e-01 7.88558573e-02 7.79991686e-01 -1.32674539e+00
-9.33486462e-01 8.50488782e-01 -6.24419928e-01 5.49510479e-01
4.44292098e-01 -2.18825459e-01 7.02916265e-01 -5.77615857e-01
6.00308895e-01 1.72343922e+00 4.62974519e-01 6.53499305e-01
-8.03061843e-01 -8.68778288e-01 -3.95949125e-01 4.26549941e-01
-1.40716052e+00 -4.26629037e-02 9.82663095e-01 -8.69493723e-01
6.52005255e-01 4.50904608e-01 9.28608358e-01 1.32502556e+00
3.81497629e-02 5.82150817e-01 1.59204412e+00 -5.10246754e-01
5.50715141e-02 8.43604088e-01 5.72031498e-01 4.25917625e-01
-2.42944844e-02 -1.46315902e-01 -5.54059148e-01 -3.06326091e-01
-1.72649667e-01 -1.66902140e-01 1.97668850e-01 3.41992944e-01
-4.14272785e-01 1.48587501e+00 1.63424149e-01 7.01881707e-01
-5.67688644e-02 -3.96148413e-01 8.61566067e-01 1.36523888e-01
7.79147863e-01 6.61324263e-01 1.64221391e-01 -7.35663831e-01
-7.40678370e-01 2.71709293e-01 7.08257020e-01 2.17735574e-01
1.82493716e-01 2.15945225e-02 2.43970543e-01 1.24285865e+00
5.25505960e-01 3.75850230e-01 8.65393221e-01 -2.68167853e-01
2.78326958e-01 9.38771129e-01 -3.99605036e-01 -1.68838763e+00
-5.09988904e-01 -2.44545922e-01 -4.08342242e-01 2.63163626e-01
4.47452143e-02 -3.02774727e-01 -6.08590364e-01 1.36471295e+00
4.08096254e-01 -3.80237430e-01 -4.19290811e-02 7.14352310e-01
9.55435991e-01 9.13024485e-01 1.95473760e-01 7.98621178e-02
1.05522299e+00 -8.32208037e-01 -8.87058020e-01 1.80450410e-01
7.44124711e-01 -7.72699177e-01 7.81339049e-01 5.59463680e-01
-7.51360178e-01 -3.64014536e-01 -9.94778335e-01 -1.39031366e-01
-9.92453873e-01 -2.92997360e-01 3.12078536e-01 1.25138652e+00
-6.42088532e-01 2.52827764e-01 -1.01302184e-01 -7.35128820e-01
3.16592216e-01 5.15038520e-02 -4.05232251e-01 4.31884825e-01
-1.35071659e+00 1.25246394e+00 5.03060997e-01 -1.18830815e-01
-3.10492188e-01 -2.89815247e-01 -4.63123888e-01 -4.23352808e-01
-2.69121855e-01 2.03823984e-01 4.53568339e-01 -1.29411125e+00
-1.33386314e+00 1.13468039e+00 5.43374896e-01 -3.16730320e-01
5.56134284e-01 -1.49212390e-01 -8.66009116e-01 -1.79046139e-01
6.30430132e-02 3.47609162e-01 7.12022603e-01 -9.55291033e-01
-6.93813562e-01 -7.56319165e-01 1.02695562e-01 2.61207461e-01
-8.36954415e-01 8.13137829e-01 3.31085414e-01 -5.66787481e-01
-3.57955724e-01 -7.87069917e-01 3.03144991e-01 -3.94917876e-01
-6.47524476e-01 -3.67725611e-01 1.00743997e+00 -1.18542814e+00
1.72186708e+00 -2.45562840e+00 3.12282324e-01 3.44397247e-01
5.26700258e-01 4.87122267e-01 4.66815501e-01 7.03547299e-01
2.75235195e-02 5.38311422e-01 1.55316249e-01 -1.17318921e-01
8.83676782e-02 2.99007267e-01 5.82368746e-02 4.75689054e-01
-2.13675305e-01 7.65440986e-02 -6.74238563e-01 -7.33111501e-01
1.67225495e-01 3.77488911e-01 -6.27314448e-01 2.75040090e-01
5.20210445e-01 3.15347433e-01 -3.03290129e-01 4.82599854e-01
6.65106654e-01 7.55010188e-01 -7.23406896e-02 3.98159586e-02
-5.78972340e-01 -2.39687953e-02 -1.01214790e+00 8.43069553e-01
-3.83934647e-01 8.79998446e-01 4.46137320e-03 -7.53546536e-01
1.25715601e+00 -9.23593789e-02 1.09695971e-01 -6.94966912e-01
6.19714439e-01 3.14171642e-01 4.28102821e-01 -8.44347119e-01
7.14441240e-01 -1.17377959e-01 -3.83940637e-01 2.52012581e-01
2.75168493e-02 6.11797459e-02 3.55712771e-01 2.73384571e-01
6.06351256e-01 -2.15073749e-01 3.61977518e-01 -3.33318144e-01
6.77663565e-01 -9.45921764e-02 3.60052556e-01 3.86290729e-01
-6.40642405e-01 1.25766769e-01 7.59994984e-01 -5.37555516e-01
-1.01809788e+00 -7.83399582e-01 -3.38650346e-01 1.48248541e+00
-6.93799779e-02 -4.63066965e-01 -7.52217114e-01 -6.38604701e-01
3.30007970e-02 1.00793588e+00 -7.83103287e-01 -2.82566816e-01
-2.97754079e-01 -7.53621578e-01 8.70967627e-01 -8.03069398e-02
6.67229652e-01 -7.33418524e-01 -5.26780963e-01 7.08476081e-02
1.19326569e-01 -5.46459019e-01 1.14325672e-01 -3.12640257e-02
-4.45075408e-02 -7.05663800e-01 -2.39738569e-01 -4.44773346e-01
2.44551212e-01 -3.28520626e-01 6.12078786e-01 2.41448671e-01
-4.28901792e-01 -5.61298691e-02 -8.41474950e-01 -7.06696451e-01
-7.61704266e-01 3.25922929e-02 2.40256980e-01 2.42321968e-01
8.32962692e-01 -2.86766589e-01 -1.02388777e-01 3.21375094e-02
-6.70090079e-01 -2.73391288e-02 8.54003951e-02 6.46457732e-01
-4.76114452e-01 5.70160635e-02 3.56986046e-01 -8.38666260e-01
1.04763496e+00 -8.86937737e-01 1.14308730e-01 -1.48733586e-01
-4.24963474e-01 -4.02851462e-01 6.71956778e-01 -3.26396137e-01
-9.39605594e-01 -6.40054703e-01 -4.95654523e-01 1.12156101e-01
-2.18687147e-01 3.57563168e-01 9.68328640e-02 2.36171996e-03
7.78169274e-01 -4.72645536e-02 1.56312779e-01 -5.00821173e-01
1.07980020e-01 1.55982530e+00 -6.42495453e-02 -1.38602078e-01
7.02961683e-01 1.10611981e-02 -4.94441181e-01 -1.43839240e+00
-4.90521073e-01 -5.15785396e-01 -4.92765665e-01 -6.59557402e-01
1.34680820e+00 -3.52108598e-01 -6.19894922e-01 7.43551850e-01
-9.86312330e-01 3.04221898e-01 1.92881092e-01 4.73333627e-01
1.57704398e-01 3.93710315e-01 -5.10185897e-01 -1.25902140e+00
-5.12727141e-01 -8.35809529e-01 2.90438294e-01 2.88479239e-01
-8.11910629e-01 -8.88680637e-01 4.59908098e-01 8.16648424e-01
5.34593284e-01 3.13010961e-01 1.03045702e+00 -1.35669875e+00
5.94014645e-01 -2.06772059e-01 -1.20247923e-01 7.27488458e-01
-6.74252808e-02 6.61353827e-01 -7.04072595e-01 1.01767674e-01
-7.36859739e-02 -6.66985989e-01 2.50735670e-01 -2.84689724e-01
6.86494589e-01 -5.02839625e-01 1.42158613e-01 2.01123208e-01
1.13381159e+00 4.90387648e-01 5.86367726e-01 6.28195405e-01
7.49510586e-01 7.50648439e-01 4.49515134e-01 7.50275195e-01
3.09416413e-01 6.29065573e-01 3.00264925e-01 1.54804870e-01
7.39425570e-02 -2.61860877e-01 3.61200154e-01 1.12572765e+00
-1.13213286e-01 -7.19546974e-02 -1.26003981e+00 3.16336364e-01
-1.38400578e+00 -1.29124260e+00 -2.26791829e-01 1.60188746e+00
7.30526447e-01 4.90628928e-02 6.38365746e-01 4.19418067e-01
6.67231023e-01 2.64167517e-01 -4.74547297e-02 -1.58174026e+00
2.55679395e-02 3.19585651e-02 3.01141620e-01 5.78856111e-01
-1.07643926e+00 8.40617061e-01 5.34582329e+00 8.22043359e-01
-1.34170485e+00 2.53399134e-01 6.86647654e-01 -1.43966660e-01
5.45546748e-02 -5.52010179e-01 -7.74067104e-01 9.80629623e-01
1.03537369e+00 -4.24189240e-01 2.88269877e-01 9.72018003e-01
1.78558007e-01 -1.49792120e-01 -3.93615395e-01 1.02403498e+00
7.33913481e-01 -7.90936530e-01 -2.25545034e-01 1.20828778e-01
2.85278201e-01 1.53428875e-02 1.82553202e-01 6.79516017e-01
2.26720378e-01 -1.04468393e+00 6.96681678e-01 2.89315999e-01
6.49042204e-02 -8.33136380e-01 9.38041627e-01 4.84913170e-01
-3.00277352e-01 -4.02808696e-01 -1.70843452e-01 -3.58293444e-01
-1.25466466e-01 -8.99368823e-02 -7.90929139e-01 8.22458714e-02
8.53938639e-01 4.14036691e-01 -8.79091918e-01 6.95799530e-01
1.60990521e-01 7.66515017e-01 -3.21358263e-01 -8.74785662e-01
3.39223504e-01 -5.36966324e-01 5.91807663e-01 1.35787106e+00
5.36436997e-02 -2.35929340e-01 -1.77450404e-02 4.47135121e-01
3.61116320e-01 7.56604612e-01 -7.08227932e-01 -4.45855439e-01
2.63630390e-01 1.22823179e+00 -3.48498732e-01 -1.01992495e-01
-2.87885994e-01 7.41647601e-01 5.19098938e-01 -3.62819582e-01
-1.10245228e+00 -5.94158947e-01 4.49793756e-01 1.89496949e-01
-4.55303371e-01 -1.26679465e-01 -1.70870289e-01 -7.14288592e-01
-5.08969545e-01 -1.12758625e+00 4.94058728e-01 -4.52855289e-01
-1.38549030e+00 7.21152365e-01 2.39718229e-01 -7.87676096e-01
-1.74777448e-01 -6.91896200e-01 -4.48579341e-01 7.99652934e-01
-6.57226205e-01 -1.14981461e+00 -2.14308083e-01 2.38310769e-01
4.92117614e-01 -6.04938507e-01 6.56217098e-01 6.58831716e-01
-9.73018229e-01 4.04118687e-01 5.42328581e-02 4.89670783e-01
4.78945017e-01 -1.17159188e+00 -3.12404335e-01 6.44794345e-01
-2.28216052e-01 4.25984353e-01 1.08202457e+00 -6.42208219e-01
-8.18122387e-01 -7.58023560e-01 9.54445362e-01 -4.16337878e-01
1.10940492e+00 -3.71377766e-01 -6.93531334e-01 2.10023999e-01
4.97540683e-01 -6.54598057e-01 1.24556255e+00 1.71283320e-01
-3.72532040e-01 -1.30914701e-02 -1.47160470e+00 6.61207795e-01
6.59335971e-01 -3.07146460e-01 -9.45608854e-01 3.86257738e-01
-6.11241385e-02 2.26252712e-02 -6.53836191e-01 -4.77376610e-01
5.71218252e-01 -1.11066258e+00 4.26041961e-01 -8.92551363e-01
8.52136195e-01 1.33308470e-01 -3.50539804e-01 -1.40579903e+00
-4.83564675e-01 -2.37308308e-01 3.82590383e-01 1.47161710e+00
4.43298966e-01 -6.28092706e-01 4.97460634e-01 6.72674894e-01
5.78470416e-02 -5.17424345e-01 -1.06800401e+00 -4.10447568e-01
3.84525388e-01 -1.28638849e-01 1.96432769e-01 1.31256354e+00
3.28837365e-01 5.37990987e-01 -6.16517425e-01 -2.59770483e-01
4.01713759e-01 -4.89644289e-01 9.96225178e-01 -1.21900868e+00
-6.97737327e-03 -6.57435477e-01 -1.13239098e+00 -9.67985913e-02
2.34290943e-01 -9.74946439e-01 -6.21999264e-01 -1.27572989e+00
3.06783676e-01 -1.93123698e-01 2.43493751e-01 2.05880776e-01
1.07604869e-01 4.39684153e-01 4.33670878e-01 -1.33894801e-01
-2.58615196e-01 5.20808816e-01 5.83025396e-01 -5.55622503e-02
-6.26617372e-02 -4.41501915e-01 -6.21368527e-01 9.41448987e-01
1.10224843e+00 -4.11866903e-01 -3.24054867e-01 -1.34250879e-01
4.57789809e-01 -5.18332362e-01 9.34161171e-02 -1.10733521e+00
-1.18501693e-01 -2.99937874e-01 2.35754088e-01 -4.22171175e-01
5.96696019e-01 -6.33528411e-01 -9.63565409e-02 5.63056827e-01
-4.53576237e-01 3.41522336e-01 -1.30099371e-01 1.69167265e-01
-2.32226998e-01 -5.25632262e-01 9.97495770e-01 3.17510180e-02
-6.64733887e-01 -2.63272915e-02 -5.59827149e-01 2.22104162e-01
1.38336337e+00 -3.74246716e-01 -5.34672379e-01 -4.68720764e-01
-5.78896105e-01 -8.73404890e-02 1.97973415e-01 6.94646060e-01
4.20980215e-01 -1.10189831e+00 -1.06563818e+00 1.62669972e-01
1.33223116e-01 -1.09890270e+00 3.43874991e-01 7.04443097e-01
-1.04215813e+00 -6.44227564e-02 -7.99392164e-01 -8.14486369e-02
-1.72107422e+00 3.27214479e-01 2.00653285e-01 1.73958346e-01
-6.21000230e-02 7.68741846e-01 -4.34903711e-01 -6.00569785e-01
-9.47298408e-02 5.36612809e-01 -9.40472484e-01 5.44620156e-01
4.54750031e-01 8.25583518e-01 -3.84782135e-01 -1.53031123e+00
-3.71503472e-01 2.17891678e-01 -3.64488214e-01 -2.27695525e-01
1.05643737e+00 3.29899825e-02 -5.74178994e-01 7.28319287e-01
1.54074430e+00 4.32047099e-01 7.31068701e-02 4.83981699e-01
-2.02566385e-02 -9.23133254e-01 1.66176274e-01 -8.86016905e-01
-5.49959004e-01 8.58209312e-01 7.58038044e-01 9.11433280e-01
5.81406832e-01 -2.93635726e-01 4.95843083e-01 -4.37158644e-02
3.44470888e-02 -1.79316854e+00 -1.87542692e-01 5.88648140e-01
1.06525791e+00 -1.17014539e+00 -4.57948484e-02 -1.91592589e-01
-9.17266846e-01 1.15047777e+00 8.74370635e-01 6.30873218e-02
7.27722168e-01 -6.31538928e-02 3.03881407e-01 -1.99791476e-01
-3.87173861e-01 1.22585103e-01 -1.35420151e-02 6.66176140e-01
7.47362018e-01 3.30088288e-01 -1.07257521e+00 6.13007903e-01
-8.09502959e-01 -4.34648633e-01 7.94456601e-01 6.77497745e-01
-3.87113273e-01 -7.94582844e-01 -4.36497390e-01 5.88016510e-01
-7.53243923e-01 1.44533008e-01 -1.02043605e+00 9.83206332e-01
4.74068761e-01 1.08981085e+00 2.26918817e-01 -8.85930181e-01
2.03714237e-01 1.02156721e-01 4.30822521e-02 -4.00328070e-01
-9.97497678e-01 -5.52345037e-01 6.66288078e-01 -4.88459542e-02
-1.92024574e-01 -5.23579717e-01 -8.33326697e-01 -8.27749312e-01
-6.04900457e-02 3.77132565e-01 1.08188236e+00 7.37178862e-01
1.83089972e-01 4.11770791e-02 8.95381689e-01 -3.59801829e-01
-6.04775786e-01 -1.15993536e+00 -6.99768603e-01 7.80841708e-01
-1.14584781e-01 -7.32341349e-01 -6.52050853e-01 -5.74326634e-01] | [8.816947937011719, 10.633091926574707] |
344e79dc-86bf-4671-988c-0d8eb0790295 | proceedings-of-the-13th-international-9 | 2112.14770 | null | https://arxiv.org/abs/2112.14770v1 | https://arxiv.org/pdf/2112.14770v1.pdf | Proceedings of the 13th International Conference on Automated Deduction in Geometry | Automated Deduction in Geometry (ADG) is a forum to exchange ideas and views, to present research results and progress, and to demonstrate software tools at the intersection between geometry and automated deduction. Relevant topics include (but are not limited to): polynomial algebra, invariant and coordinate-free methods; probabilistic, synthetic, and logic approaches, techniques for automated geometric reasoning from discrete mathematics, combinatorics, and numerics; interactive theorem proving in geometry; symbolic and numeric methods for geometric computation, geometric constraint solving, automated generation/reasoning and manipulation with diagrams; design and implementation of geometry software, automated theorem provers, special-purpose tools, experimental studies; applications of ADG in mechanics, geometric modelling, CAGD/CAD, computer vision, robotics and education. Traditionally, the ADG conference is held every two years. The previous editions of ADG were held in Nanning in 2018, Strasbourg in 2016, Coimbra in 2014, Edinburgh in 2012, Munich in 2010, Shanghai in 2008, Pontevedra in 2006, Gainesville in 2004, Hagenberg in 2002, Zurich in 2000, Beijing in 1998, and Toulouse in 1996. The 13th edition of ADG was supposed to be held in 2020 in Hagenberg, Austria, but due to the COVID-19 pandemic, it was postponed for 2021, and held online (still hosted by RISC Institute, Hagenberg, Austria), September 15-17, 2021 (https://www.risc.jku.at/conferences/adg2021). | ['Zoltán Kovács', 'Predrag Janičić'] | 2021-12-28 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [-4.26028520e-01 3.55471343e-01 3.60188931e-01 2.35233635e-01
-3.48894268e-01 -8.10237229e-01 6.22185051e-01 4.26362514e-01
8.42320248e-02 7.31384695e-01 -2.59808511e-01 -8.03411067e-01
-6.87726736e-01 -1.09684551e+00 -5.90525985e-01 -3.28226209e-01
-7.10617959e-01 6.40676618e-01 1.07269406e-01 -7.57701322e-02
2.42482588e-01 1.09771955e+00 -1.54562163e+00 -3.08874816e-01
7.22693324e-01 5.85299253e-01 2.16577519e-02 8.11509311e-01
1.70606077e-01 5.77049017e-01 -1.11258879e-01 -3.12480360e-01
2.77519882e-01 -2.87996978e-01 -9.89814043e-01 -4.65891033e-01
3.96507699e-03 -3.11200619e-01 -2.45154917e-01 1.15083265e+00
1.78960726e-01 -5.90921305e-02 6.58029020e-01 -1.51819634e+00
-4.86828685e-02 6.29903913e-01 -3.96604478e-01 -5.02250731e-01
8.00152302e-01 -1.19510870e-02 6.46983385e-01 -9.19657290e-01
7.75671601e-01 1.25385988e+00 4.92648751e-01 4.90193814e-02
-8.80639613e-01 -7.55786598e-01 -1.67900011e-01 3.05505037e-01
-1.69083273e+00 2.68077374e-01 6.60083890e-01 -7.40947306e-01
4.04354751e-01 6.22373819e-01 9.18255568e-01 2.71557450e-01
1.35209590e-01 6.55076921e-01 9.31738198e-01 -9.07498777e-01
4.10281032e-01 5.15202954e-02 -1.90286666e-01 9.79398012e-01
5.06591499e-01 2.89446890e-01 3.68463546e-02 -2.10139602e-01
1.09577525e+00 -3.93553287e-01 1.33494660e-01 -4.39308435e-01
-1.21299601e+00 9.09591973e-01 1.33924738e-01 2.73820370e-01
-2.63311982e-01 3.01280290e-01 2.01359123e-01 3.25331569e-01
-9.44744572e-02 4.53703612e-01 -4.23313230e-01 -1.84401438e-01
-6.78757071e-01 1.13302934e+00 1.01783538e+00 1.10110915e+00
7.53735304e-01 -2.10396364e-01 3.62136900e-01 2.01631263e-01
5.18697977e-01 5.09737551e-01 -3.67873907e-01 -1.25503886e+00
3.14594090e-01 2.40062043e-01 2.64230907e-01 -1.24442995e+00
-6.14888310e-01 2.96916306e-01 -7.46217132e-01 3.18025589e-01
4.77258354e-01 -3.23673695e-01 -4.33817238e-01 1.20825875e+00
6.28970802e-01 -1.58285901e-01 -1.50974751e-01 8.99756253e-01
5.65984666e-01 7.01653898e-01 -3.11982125e-01 -3.18948656e-01
9.03001904e-01 -3.02899837e-01 -3.59056085e-01 6.56346023e-01
7.05254495e-01 -9.80750024e-01 3.79437655e-01 9.37727094e-01
-1.24314201e+00 -1.98492706e-01 -7.86970675e-01 3.98455888e-01
-4.94054288e-01 -1.03789926e-01 6.94311857e-01 6.54763281e-01
-1.00640941e+00 6.27124965e-01 -9.64094579e-01 -4.55176562e-01
2.64772564e-01 4.14851487e-01 -4.51303907e-02 -1.82989910e-01
-1.29950953e+00 8.88286471e-01 3.26273561e-01 -1.77722722e-02
-5.94343960e-01 -9.16623175e-01 -8.22085321e-01 -2.54173011e-01
2.95099318e-01 -4.52696025e-01 9.25771773e-01 -5.14110923e-02
-1.46457970e+00 4.59768534e-01 1.59542426e-01 -2.58070022e-01
7.94626117e-01 1.20815106e-01 -2.53564805e-01 -1.55502865e-02
8.74777362e-02 5.01244903e-01 -5.97655550e-02 -1.03893554e+00
-3.78417492e-01 -4.88394290e-01 4.11488175e-01 -2.22698823e-01
5.41856527e-01 5.77133149e-02 -7.38514634e-03 -3.05704683e-01
1.91527948e-01 -8.07927668e-01 -3.12828660e-01 -1.72951162e-01
-6.01410270e-01 -2.97233194e-01 9.29765463e-01 -5.67267239e-01
7.69051790e-01 -1.81294811e+00 3.33814286e-02 6.70547783e-01
-1.53814331e-01 1.48504395e-02 4.88719195e-01 1.21148837e+00
5.68363443e-02 1.35912329e-01 -1.71395659e-01 4.54827279e-01
3.33369106e-01 -7.17413500e-02 -1.50855094e-01 6.74468875e-01
-1.71004623e-01 6.54204607e-01 -1.08522391e+00 -7.08555520e-01
8.39811921e-01 -2.73017376e-03 -3.35067511e-01 -3.28116179e-01
-3.18904698e-01 3.83446127e-01 -7.24872470e-01 8.11572373e-01
9.17795241e-01 2.58548528e-01 2.90730536e-01 7.10120350e-02
-7.91153133e-01 -2.75542792e-02 -1.91192865e+00 1.37779295e+00
-3.37640196e-01 4.57097083e-01 4.00176644e-01 -1.09731328e+00
9.97120380e-01 4.51264650e-01 6.50814176e-01 -3.20401043e-01
2.16563180e-01 4.08961177e-01 -8.30806196e-02 -3.88159037e-01
5.44700325e-01 3.55098814e-01 -7.92304873e-02 4.63413328e-01
-2.18912631e-01 -1.21245110e+00 4.78228211e-01 3.54390234e-01
8.38984430e-01 4.76664007e-01 4.11462098e-01 -3.19349229e-01
6.14817441e-01 2.22861543e-01 7.06216171e-02 4.00080323e-01
-2.33646296e-02 2.71493733e-01 4.52532619e-01 -4.14437830e-01
-1.05342889e+00 -1.19137383e+00 -2.84410655e-01 2.00133353e-01
2.92019576e-01 -6.24622047e-01 -7.92381465e-01 -1.46443859e-01
-3.33252028e-02 9.02437508e-01 -2.14819506e-01 1.34512737e-01
-5.23300350e-01 2.50630438e-01 4.48099464e-01 4.57031950e-02
8.62294257e-01 -9.04548347e-01 -4.91780639e-01 1.66919660e-02
2.97208965e-01 -7.96578109e-01 9.44723412e-02 -3.26399133e-02
-6.39566481e-01 -1.34538877e+00 -5.87276161e-01 -5.54040611e-01
4.51798320e-01 -1.55892879e-01 4.86368358e-01 -6.32698908e-02
-7.25210607e-01 3.58485281e-01 -1.97116137e-01 -9.13707197e-01
-3.03721279e-01 -2.27440804e-01 3.55476514e-02 -6.90492630e-01
-1.59528121e-01 -6.55515254e-01 -4.94527459e-01 3.74663144e-01
-7.02247083e-01 4.52871472e-02 3.96504611e-01 4.17190552e-01
3.69015068e-01 3.23661953e-01 4.17346597e-01 -5.19093573e-01
1.01031393e-01 -1.04698986e-01 -1.24022412e+00 1.31672025e-01
-5.47092974e-01 -3.99738252e-01 7.52651751e-01 1.60019413e-01
-6.93219721e-01 5.68891726e-02 3.56318876e-02 -2.96082616e-01
-8.62416551e-02 4.72374111e-01 -3.30141425e-01 -1.15895130e-01
4.47443098e-01 8.65743533e-02 5.01154270e-03 -1.57315508e-01
4.71282780e-01 5.02660036e-01 3.51497710e-01 -1.05284965e+00
1.20544028e+00 3.73071611e-01 6.97136939e-01 -7.83164263e-01
-1.78564548e-01 -1.10985480e-01 -6.73554480e-01 -5.37280142e-01
5.69498181e-01 -4.01501060e-01 -8.52492630e-01 5.10563731e-01
-9.93648648e-01 -2.41457596e-01 -4.37774748e-01 8.04646909e-01
-5.50320208e-01 1.53462276e-01 -2.79311717e-01 -1.28639340e+00
1.44506842e-01 -1.09473753e+00 7.64325798e-01 2.80949414e-01
-5.24974823e-01 -1.05591655e+00 8.76252279e-02 1.43292680e-01
5.32038957e-02 7.09538341e-01 1.05652618e+00 -5.27738154e-01
-8.30198884e-01 -2.81719029e-01 -2.10725918e-01 2.30204374e-01
-1.04612812e-01 6.27298653e-01 -3.31266582e-01 2.06433926e-02
-6.23031497e-01 -1.06029816e-01 -5.06486185e-02 4.93135780e-01
1.21511102e+00 -3.21171403e-01 -5.82237244e-01 -4.52353247e-02
1.30190015e+00 2.90343642e-01 7.96466470e-01 3.67826790e-01
2.24767655e-01 5.25973737e-01 8.90420318e-01 3.61980110e-01
6.30956590e-01 7.42102444e-01 5.53150356e-01 3.01269352e-01
1.33411974e-01 -2.73211092e-01 2.03737080e-01 7.21190035e-01
-5.38384795e-01 4.83103871e-01 -1.08466744e+00 5.39070129e-01
-1.72745419e+00 -7.84985602e-01 -8.16116512e-01 2.73606396e+00
5.92068076e-01 2.39150766e-02 3.25205386e-01 4.72218394e-01
7.33824074e-01 -3.97850841e-01 1.63542286e-01 -7.27330029e-01
4.20407921e-01 5.27636826e-01 5.77345729e-01 5.89308918e-01
-7.75793374e-01 7.64429927e-01 4.70337296e+00 1.10104930e+00
-8.66696775e-01 -2.63776839e-01 1.87791005e-01 3.90642673e-01
-4.76050586e-01 5.82315683e-01 -3.79637718e-01 2.96071172e-01
7.30138600e-01 -6.01446271e-01 5.83986878e-01 1.01798880e+00
3.94100666e-01 -3.40254307e-01 -7.71209478e-01 6.57799482e-01
-4.11175609e-01 -1.49793005e+00 -5.42926490e-01 3.89192432e-01
9.96975124e-01 -2.39750937e-01 -3.59847993e-01 1.39109448e-01
4.99272734e-01 -1.07819939e+00 1.01157546e+00 5.04238665e-01
5.44544876e-01 -1.26313162e+00 6.28158867e-01 2.43868664e-01
-9.39846873e-01 3.27682346e-01 7.80122727e-03 -4.35323983e-01
1.28934324e-01 9.49588299e-01 -9.66064095e-01 1.40520871e+00
6.14034235e-01 1.15701839e-01 -1.23615861e-01 1.09402800e+00
-3.71633261e-01 5.43102384e-01 -4.56212759e-01 -6.29928052e-01
5.47023751e-02 -4.84633923e-01 7.44313896e-01 8.14264596e-01
3.15225363e-01 1.98252186e-01 2.14031056e-01 1.10386467e+00
5.03547311e-01 -1.44959673e-01 -6.17246687e-01 2.18880773e-01
6.51536703e-01 1.25983071e+00 -9.22809780e-01 7.38394111e-02
1.78748332e-02 5.58685139e-02 -6.46627605e-01 4.71418887e-01
-1.07107460e+00 -9.84944999e-01 1.21631563e-01 5.65248549e-01
2.63544731e-02 -7.00266957e-01 -5.38578749e-01 -5.55282235e-01
-1.21551424e-01 -5.91714203e-01 6.75009340e-02 -8.14107776e-01
-6.42574489e-01 1.18619055e-01 8.24966133e-01 -9.39455807e-01
-6.29578650e-01 -7.79199302e-01 -5.57926893e-01 9.37694371e-01
-5.14652312e-01 -1.09707499e+00 1.26752481e-01 2.68256247e-01
-5.20829558e-02 -2.99335718e-02 8.77199471e-01 2.17907459e-01
6.01928756e-02 2.44253159e-01 3.67888451e-01 1.32186517e-01
1.74735561e-02 -1.20867991e+00 1.64144009e-01 4.73025709e-01
-4.04448695e-02 4.48149472e-01 8.92885447e-01 -6.07732236e-01
-1.68464589e+00 -1.24286532e+00 8.97543550e-01 1.36390999e-01
8.65464270e-01 -4.35990393e-01 -3.16085398e-01 7.04680741e-01
-3.48970115e-01 -1.84497669e-01 -1.65660307e-01 -1.66473389e-01
1.29079685e-01 1.96269364e-03 -1.52452421e+00 6.40650332e-01
6.75346255e-01 -1.34505540e-01 -1.53247550e-01 6.94796324e-01
4.98966277e-01 -9.06655550e-01 -9.12143230e-01 3.34044397e-01
5.49190342e-01 -6.47124112e-01 6.64536059e-01 6.54437765e-02
2.77702808e-01 -5.63329041e-01 -2.24047631e-01 -8.14341128e-01
1.38900623e-01 -1.06284273e+00 2.76387990e-01 1.27646935e+00
3.66636217e-01 -1.14693511e+00 6.11191392e-01 1.96713045e-01
-4.01592106e-01 -8.03798556e-01 -1.18726611e+00 -1.07958078e+00
4.63177055e-01 -7.60738730e-01 6.27304733e-01 7.50611186e-01
1.63790211e-01 -4.15043712e-01 4.21075672e-01 8.16609189e-02
5.61867476e-01 2.11943224e-01 9.42572236e-01 -1.25464857e+00
-2.19160989e-02 -4.71886039e-01 -8.22326124e-01 -3.55262607e-01
-5.82194217e-02 -6.78744555e-01 -1.75284415e-01 -1.95543778e+00
-4.74781275e-01 -6.02822065e-01 7.28807330e-01 2.09073305e-01
4.96943027e-01 -1.94885969e-01 2.00532332e-01 -8.06669518e-02
-1.87292948e-01 8.74357224e-02 1.24286377e+00 -2.10438911e-02
-9.18145329e-02 3.49056721e-01 -1.92548260e-01 6.72661483e-01
6.34564817e-01 -2.06527784e-02 -1.72864065e-01 2.35769778e-01
3.31222236e-01 1.29107028e-01 9.23956692e-01 -9.90719199e-01
2.85930902e-01 -5.60698628e-01 2.19769225e-01 -8.92721295e-01
1.99379295e-01 -8.09309125e-01 6.62595332e-01 4.91494328e-01
1.63059771e-01 -1.54458612e-01 2.89980054e-01 3.87036540e-02
-1.18567785e-02 -5.19399166e-01 3.74926835e-01 -1.99851811e-01
-2.46085137e-01 1.38067201e-01 -6.24673128e-01 -1.02031402e-01
1.52829027e+00 -3.17083418e-01 -1.76107422e-01 -4.45457458e-01
-8.94596219e-01 3.84719998e-01 5.04524589e-01 -1.54763237e-01
4.18962657e-01 -1.20570898e+00 -6.89582527e-01 -3.30394238e-01
-3.62705022e-01 5.43310285e-01 3.66146922e-01 9.48528171e-01
-1.15824318e+00 4.89390910e-01 2.12281600e-01 -4.83085066e-01
-8.44482899e-01 3.32658917e-01 -8.09219033e-02 -4.74615932e-01
-5.26009083e-01 4.39048409e-01 -2.21274137e-01 -7.10709572e-01
1.39081836e-01 -3.41702342e-01 2.93680847e-01 -3.96548957e-01
2.42928967e-01 8.26861978e-01 2.41632581e-01 -2.28305310e-01
-5.57754397e-01 5.71697772e-01 4.13985699e-01 -5.27617455e-01
1.21754420e+00 6.45632595e-02 -4.52826530e-01 5.31125784e-01
8.34779322e-01 4.91508804e-02 -4.72465426e-01 6.16374850e-01
-6.01160824e-02 -1.03976771e-01 -2.55312055e-01 -8.16567183e-01
-5.41791439e-01 5.64534545e-01 1.49280921e-01 5.05977869e-01
8.68861318e-01 -9.62354336e-03 4.40863162e-01 1.06877036e-01
9.69385684e-01 -1.16516769e+00 -6.03258431e-01 7.48565793e-01
1.11587358e+00 -3.23576659e-01 3.37467909e-01 -7.00616360e-01
-2.29969680e-01 1.59691894e+00 1.71375811e-01 -2.29500353e-01
1.01585352e+00 2.53714949e-01 -4.03363198e-01 -2.84964949e-01
-2.99437582e-01 -1.00689471e-01 -3.57717788e-03 7.52725124e-01
2.69415438e-01 1.78973064e-01 -3.35265756e-01 4.30623263e-01
-7.74486303e-01 4.46665406e-01 8.31800997e-01 1.01168787e+00
1.71267502e-02 -1.08687198e+00 -7.98352599e-01 3.31296831e-01
-1.40399501e-01 1.51669875e-01 -4.95822430e-01 1.58974922e+00
2.12816149e-01 7.21361876e-01 2.70553045e-02 -3.09585661e-01
3.38532090e-01 -1.62749901e-01 9.01474714e-01 -3.35340440e-01
-3.25668305e-01 -9.90521237e-02 3.07293087e-01 -2.08467767e-01
-8.41909051e-02 -1.00516737e+00 -1.79854083e+00 -6.17803812e-01
-4.06921655e-01 8.91416013e-01 1.04224062e+00 8.57598066e-01
1.35684937e-01 3.30176830e-01 6.28898740e-01 -7.46396840e-01
-2.83960730e-01 -8.66122365e-01 -8.37468565e-01 -8.60277891e-01
-2.15976626e-01 -6.96257532e-01 -4.26983535e-01 -5.04964963e-02] | [8.77348804473877, 6.787961006164551] |
1fea07c9-e53a-485a-b391-0f3293059b3b | deepspectrumlite-a-power-efficient-transfer | 2104.11629 | null | https://arxiv.org/abs/2104.11629v1 | https://arxiv.org/pdf/2104.11629v1.pdf | DeepSpectrumLite: A Power-Efficient Transfer Learning Framework for Embedded Speech and Audio Processing from Decentralised Data | Deep neural speech and audio processing systems have a large number of trainable parameters, a relatively complex architecture, and require a vast amount of training data and computational power. These constraints make it more challenging to integrate such systems into embedded devices and utilise them for real-time, real-world applications. We tackle these limitations by introducing DeepSpectrumLite, an open-source, lightweight transfer learning framework for on-device speech and audio recognition using pre-trained image convolutional neural networks (CNNs). The framework creates and augments Mel-spectrogram plots on-the-fly from raw audio signals which are then used to finetune specific pre-trained CNNs for the target classification task. Subsequently, the whole pipeline can be run in real-time with a mean inference lag of 242.0 ms when a DenseNet121 model is used on a consumer-grade Motorola moto e7 plus smartphone. DeepSpectrumLite operates decentralised, eliminating the need for data upload for further processing. By obtaining state-of-the-art results on a set of paralinguistics tasks, we demonstrate the suitability of the proposed transfer learning approach for embedded audio signal processing, even when data is scarce. We provide an extensive command-line interface for users and developers which is comprehensively documented and publicly available at https://github.com/DeepSpectrum/DeepSpectrumLite. | ['Björn W. Schuller', 'Sandra Ottl', 'Maurice Gerczuk', 'Tobias Hübner', 'Shahin Amiriparian'] | 2021-04-23 | null | null | null | null | ['audio-signal-processing'] | ['audio'] | [ 1.61941588e-01 -1.27881125e-01 1.17947310e-01 -3.20120573e-01
-9.44420099e-01 -5.30616879e-01 1.47596881e-01 -2.15885848e-01
-5.56068301e-01 2.53403544e-01 -3.16763461e-01 -6.88696444e-01
8.99344161e-02 -4.80613232e-01 -8.66859555e-01 -5.33842802e-01
-1.75411075e-01 1.77479357e-01 1.23193897e-01 -4.83975112e-02
-2.77670532e-01 3.24002117e-01 -1.84560525e+00 6.29335165e-01
2.93951601e-01 1.40150702e+00 3.64181519e-01 1.12919188e+00
6.29222333e-01 5.57237685e-01 -8.71410608e-01 -7.10912794e-02
7.26009905e-02 6.58454448e-02 -3.94752383e-01 -3.29844713e-01
2.92003572e-01 -5.65625310e-01 -3.87102932e-01 8.78040850e-01
1.08896005e+00 9.67373401e-02 1.88946620e-01 -1.26372159e+00
-6.02868535e-02 5.94612122e-01 1.78861871e-01 4.60621357e-01
1.35930374e-01 5.49443483e-01 7.20117569e-01 -8.50567818e-01
-2.06467301e-01 9.79800105e-01 8.22633982e-01 4.03533041e-01
-1.03312624e+00 -1.09240723e+00 -3.24263901e-01 3.65603805e-01
-1.35736907e+00 -1.12612605e+00 7.04333603e-01 -3.91343266e-01
1.51019084e+00 2.57530808e-01 5.19088089e-01 1.44421756e+00
1.67057648e-01 5.19449770e-01 6.71892405e-01 -3.12340617e-01
4.01867867e-01 7.83122703e-02 -2.79852420e-01 5.26270449e-01
-5.04205406e-01 1.56156003e-01 -7.23681629e-01 -8.41669366e-02
7.01343834e-01 -2.40720004e-01 -2.66071200e-01 3.01991105e-01
-1.11830962e+00 5.62328339e-01 2.57326186e-01 2.54538745e-01
-5.21816492e-01 3.43955666e-01 8.62537980e-01 6.06230497e-01
5.60914695e-01 -4.38916311e-02 -8.89561534e-01 -7.73950994e-01
-1.03953600e+00 -1.35548040e-01 6.67197526e-01 7.76427805e-01
5.20543814e-01 4.31165844e-01 4.20415252e-01 1.06811154e+00
2.00291499e-01 4.37857747e-01 8.13583910e-01 -8.07546377e-01
4.71230924e-01 -2.95151211e-02 -4.05967861e-01 -5.98039508e-01
-4.65668738e-01 -4.25240934e-01 -9.40936387e-01 2.32021019e-01
2.00193092e-01 -4.76712078e-01 -6.67984247e-01 1.42062294e+00
1.87295064e-01 5.17033041e-01 -2.86236797e-02 8.89945567e-01
7.66262114e-01 8.74268770e-01 -1.38461515e-01 2.06431419e-01
1.38663197e+00 -1.01948309e+00 -5.46295404e-01 -1.89960375e-01
4.62307811e-01 -8.79090428e-01 1.36971390e+00 7.04890251e-01
-1.08441865e+00 -1.00005853e+00 -1.20178175e+00 -1.86191559e-01
-4.04459506e-01 5.86938858e-01 4.13364142e-01 6.67248189e-01
-1.33038986e+00 7.18120337e-01 -1.26272857e+00 -9.70839560e-02
3.71023238e-01 9.38341796e-01 -1.56782329e-01 5.58710158e-01
-1.24441576e+00 3.53260010e-01 3.87374610e-01 2.57947177e-01
-1.16732121e+00 -9.49976802e-01 -7.25908697e-01 2.19209701e-01
2.77004927e-01 -5.13383031e-01 1.77496934e+00 -1.05745614e+00
-2.16866350e+00 5.62848210e-01 2.46861935e-01 -6.75535202e-01
2.31496811e-01 -4.04113650e-01 -7.75244951e-01 1.23650342e-01
-2.24799529e-01 5.99156499e-01 1.33191597e+00 -3.42364520e-01
-7.77625501e-01 -3.79289128e-02 2.03689933e-03 -1.46732360e-01
-6.49501503e-01 2.55322754e-01 -3.72675598e-01 -6.86519384e-01
-5.84559679e-01 -9.78743792e-01 2.06351221e-01 -2.47478351e-01
-2.23654747e-01 -1.87420070e-01 8.88907552e-01 -6.85378730e-01
1.10492456e+00 -2.57037640e+00 -2.11027160e-01 8.63866974e-03
-2.75629852e-02 7.87461996e-01 -1.75968423e-01 2.41865695e-01
-3.53472590e-01 -2.21074209e-01 1.67319164e-01 -3.84593517e-01
1.46782979e-01 -1.09662279e-01 -2.25627840e-01 3.25043529e-01
2.56448030e-01 5.76911807e-01 -6.09522164e-01 4.82654199e-02
5.47234237e-01 7.88278282e-01 -6.56365633e-01 3.74092460e-01
7.92398583e-03 4.50646818e-01 8.96616746e-03 3.85104060e-01
3.13802660e-01 -1.00857824e-01 -6.34811223e-02 -2.75961041e-01
-1.20292269e-01 1.02060139e+00 -9.87984419e-01 1.83985960e+00
-1.11767507e+00 1.08261967e+00 4.74757046e-01 -1.02446854e+00
6.85727835e-01 8.03717136e-01 3.57070476e-01 -5.31876802e-01
5.23147583e-01 4.26107019e-01 1.38797104e-01 -5.35376251e-01
3.30837607e-01 2.50214338e-01 1.08493276e-01 4.67245579e-01
4.60521698e-01 -3.15154232e-02 -2.01129809e-01 -3.29070717e-01
1.21857941e+00 -2.32133999e-01 1.20489029e-02 -6.26515970e-02
4.79844987e-01 -5.12773514e-01 1.32582188e-01 2.90330946e-01
-5.35006309e-03 4.21968102e-01 4.88681979e-02 -2.65101939e-01
-1.13075614e+00 -9.45838690e-01 -1.07743606e-01 1.59585786e+00
-8.32517922e-01 -5.66126227e-01 -1.11035979e+00 -1.40177533e-01
-3.02311778e-01 3.07453305e-01 -2.19478965e-01 -1.93810672e-01
-4.48968053e-01 -3.60251874e-01 1.07955742e+00 5.23206949e-01
4.20540959e-01 -1.40595329e+00 -7.73251832e-01 4.88078326e-01
4.33836244e-02 -1.17419219e+00 -3.86793971e-01 5.35176039e-01
-6.27242625e-01 -7.18991339e-01 -6.14706695e-01 -8.29875112e-01
1.43046468e-01 -4.45150323e-02 9.14791644e-01 -1.20114498e-01
-2.57831842e-01 3.21458876e-01 -7.59256780e-02 -5.63482404e-01
-4.22696650e-01 5.56984246e-01 3.96504849e-01 1.22260913e-01
3.42703640e-01 -1.02054024e+00 -6.47872508e-01 2.15321720e-01
-7.01173484e-01 -3.76934446e-02 4.44522232e-01 6.99335635e-01
3.73041540e-01 1.96806252e-01 9.61935043e-01 -3.96531582e-01
7.73107290e-01 -4.88077283e-01 -8.74438107e-01 -3.70339036e-01
-1.85799465e-01 -4.49428141e-01 8.85200262e-01 -7.26478457e-01
-6.68113589e-01 1.58186167e-01 -6.32815778e-01 -6.76291227e-01
-3.57652605e-01 5.51613152e-01 -6.12386726e-02 2.56622612e-01
6.56694174e-01 -4.80221957e-02 5.10877781e-02 -5.06398857e-01
1.96012214e-01 1.47418165e+00 7.92194128e-01 -2.50038147e-01
3.34057122e-01 1.54423267e-01 -5.95320702e-01 -1.20451415e+00
-4.10913110e-01 -3.58432174e-01 -4.28221762e-01 -2.10139275e-01
6.33575201e-01 -1.41088116e+00 -1.04041660e+00 8.35703611e-01
-1.08321607e+00 -1.09853935e+00 -3.35361436e-02 5.87440431e-01
-7.11705625e-01 -2.21470058e-01 -6.83445215e-01 -6.83110833e-01
-7.50219047e-01 -1.24158430e+00 1.08140159e+00 -1.27673158e-02
-3.76376748e-01 -8.84723783e-01 -1.88572466e-01 3.35783243e-01
5.41548967e-01 -2.83016950e-01 5.29156208e-01 -6.11890495e-01
-3.23372841e-01 -1.91409290e-01 5.10703474e-02 9.09769893e-01
1.56161010e-01 6.95695207e-02 -1.62193739e+00 -4.26863790e-01
1.08463041e-01 -4.10405457e-01 5.17854273e-01 4.62104797e-01
1.52414382e+00 -3.05620551e-01 1.32380933e-01 6.57141566e-01
7.64662504e-01 2.67672896e-01 4.56128925e-01 2.10871056e-01
5.79182506e-01 3.03919435e-01 2.89936453e-01 4.98578817e-01
2.73192674e-01 8.33870888e-01 2.51466811e-01 -1.13973342e-01
-1.34106025e-01 6.89017773e-02 6.37169600e-01 1.34204543e+00
5.83624430e-02 -5.06424308e-02 -8.50820839e-01 5.14859915e-01
-1.60045385e+00 -6.76582634e-01 6.74573854e-02 2.15335464e+00
1.03745651e+00 2.49444112e-01 2.98272967e-01 6.07712984e-01
6.01323426e-01 -1.76441789e-01 -4.36658293e-01 -4.65596825e-01
4.78170246e-01 7.86945581e-01 1.76136404e-01 3.78949046e-01
-1.20573127e+00 9.07928169e-01 5.43414974e+00 1.12614226e+00
-1.84456909e+00 4.32956964e-01 5.45737267e-01 -4.88400787e-01
4.74534780e-01 -5.42878568e-01 -4.81114239e-01 6.28093481e-01
1.99380219e+00 1.51829928e-01 8.89229476e-01 1.05181372e+00
5.52302361e-01 2.16339722e-01 -1.35199594e+00 1.35872340e+00
-4.68023777e-01 -1.18109083e+00 -6.93500459e-01 5.62822036e-02
1.94299534e-01 6.17653549e-01 2.89718628e-01 4.56090599e-01
-1.48184553e-01 -9.49602842e-01 8.25270355e-01 1.20191611e-02
1.17313266e+00 -9.60320234e-01 6.03179574e-01 2.69070834e-01
-1.12033594e+00 -1.62581176e-01 -2.01742291e-01 -3.15575123e-01
-7.23684505e-02 6.15078807e-01 -1.12996292e+00 1.54650792e-01
1.08224189e+00 6.26901865e-01 -3.22362483e-01 6.36026621e-01
-8.63778964e-02 1.07387054e+00 -6.04982138e-01 5.31550385e-02
1.01977654e-01 3.31746995e-01 1.57402590e-01 1.36021233e+00
5.15997648e-01 -1.95639312e-01 -3.21498185e-01 5.09393871e-01
-1.53824344e-01 -1.26654357e-01 -3.09428126e-01 -1.77535996e-01
5.07649899e-01 1.40962934e+00 -4.73000646e-01 -4.33132142e-01
-3.20875257e-01 8.13846827e-01 3.23674046e-02 2.62888908e-01
-1.04889631e+00 -6.26531541e-01 9.10615385e-01 4.37539369e-02
4.76848274e-01 -2.38674909e-01 -5.15845343e-02 -8.90842438e-01
1.87510833e-01 -1.30003762e+00 3.57139530e-03 -8.99604797e-01
-1.03956091e+00 1.05321646e+00 -1.53066024e-01 -1.26590025e+00
-7.33097911e-01 -9.15636063e-01 -6.97042465e-01 7.27122784e-01
-1.22992659e+00 -8.44781160e-01 -2.56054431e-01 9.04338181e-01
6.55326307e-01 -5.13955951e-01 1.05069053e+00 9.40159678e-01
-5.70825040e-01 7.29767382e-01 -8.58613774e-02 1.45344555e-01
6.06969297e-01 -9.79638338e-01 7.83159673e-01 4.95557725e-01
2.05697015e-01 4.43636924e-01 4.16717172e-01 -8.17061290e-02
-1.25658035e+00 -1.28832138e+00 5.16742110e-01 -1.19991645e-01
1.04629993e+00 -7.77693450e-01 -8.96792710e-01 6.29328489e-01
3.59933972e-01 1.85321510e-01 8.66656065e-01 1.38793200e-01
-2.90238976e-01 -4.42301601e-01 -5.95325470e-01 4.01065290e-01
6.31828964e-01 -1.10552812e+00 -1.93833828e-01 3.12800974e-01
6.22065008e-01 -5.42124629e-01 -8.54647636e-01 1.22940913e-01
5.71540415e-01 -9.82704163e-01 8.58383358e-01 -3.72901112e-02
1.57121539e-01 -1.15155354e-01 -4.88012582e-02 -1.33383358e+00
2.00922072e-01 -1.14582849e+00 -1.65003687e-01 1.12181234e+00
5.47207654e-01 -7.91620255e-01 5.52965939e-01 1.53178513e-01
-5.63779175e-01 -6.50358737e-01 -1.22047210e+00 -6.69367552e-01
-3.03752959e-01 -1.10031426e+00 5.66969037e-01 5.48268020e-01
4.74699698e-02 5.19042969e-01 -2.00351447e-01 3.64720404e-01
2.26710215e-02 -4.96547014e-01 9.10210311e-01 -9.97518778e-01
-5.18373191e-01 -2.07571328e-01 -4.13501740e-01 -1.01142716e+00
2.47506171e-01 -6.72555387e-01 1.28112376e-01 -7.99535811e-01
-6.45784259e-01 -4.50950027e-01 -3.52815270e-01 8.36815655e-01
4.25240844e-01 4.86177593e-01 5.24818450e-02 2.40833554e-02
-3.74503464e-01 4.42381591e-01 5.15405774e-01 -1.02899909e-01
-3.67128789e-01 3.38175416e-01 -1.35768995e-01 7.59676397e-01
1.15766335e+00 -4.22898442e-01 -5.80975235e-01 -4.26911771e-01
6.94663003e-02 3.00075803e-02 6.95007622e-01 -1.53977787e+00
3.13602895e-01 6.65466726e-01 1.08145341e-01 -3.44432801e-01
8.42503011e-01 -9.41778183e-01 1.60131052e-01 2.76326656e-01
-1.84669763e-01 -1.74697256e-03 6.82702363e-01 1.35390088e-01
-3.16282868e-01 -9.90281031e-02 6.05898678e-01 2.70007968e-01
-4.16245997e-01 2.36459151e-01 -7.67833173e-01 -2.52643228e-01
6.03644669e-01 7.34044984e-02 -9.59056616e-02 -5.22500217e-01
-8.92246544e-01 -4.18170989e-01 -1.02882914e-01 5.72214127e-01
4.06704664e-01 -1.30212474e+00 -4.49558944e-01 6.44472659e-01
-1.86418772e-01 1.06445000e-01 5.02307475e-01 7.67206073e-01
-4.31546688e-01 5.75432539e-01 -2.67607421e-01 -7.48005033e-01
-1.41642892e+00 2.31967762e-01 4.60269779e-01 1.84688509e-01
-6.28926218e-01 7.85068750e-01 1.81864249e-03 -4.58142012e-01
4.80782628e-01 -7.69187689e-01 1.66197866e-01 -3.65130268e-02
6.70460701e-01 3.01806450e-01 8.42812300e-01 -3.64909917e-01
-3.95813555e-01 3.46865021e-02 1.23979062e-01 -3.15470129e-01
1.42308974e+00 1.83797091e-01 5.98746128e-02 5.99775016e-01
1.46479881e+00 -3.05109113e-01 -1.50248766e+00 -3.72727066e-02
-3.33851486e-01 1.30279303e-01 5.28035045e-01 -5.34987032e-01
-1.30735457e+00 1.21432090e+00 9.18089569e-01 3.12667429e-01
1.45740271e+00 -2.33617857e-01 9.64400172e-01 4.91540492e-01
1.93590820e-01 -1.10307837e+00 1.68567017e-01 5.37918746e-01
9.67978716e-01 -9.46399033e-01 -4.84029293e-01 -1.30679131e-01
-3.16206664e-01 1.26701558e+00 3.27200562e-01 -1.54767990e-01
9.91580248e-01 7.63795793e-01 3.83638501e-01 2.04878335e-04
-1.00337625e+00 1.66656554e-01 2.46958837e-01 6.84599638e-01
4.70650315e-01 5.01536131e-02 5.31403720e-01 8.27798307e-01
-8.09627652e-01 1.54229775e-01 2.52754360e-01 6.54565156e-01
7.20732957e-02 -8.00462186e-01 -3.15338463e-01 1.70590580e-01
-7.03148127e-01 -3.55747551e-01 1.09097093e-01 4.05797362e-01
9.83869210e-02 1.15785599e+00 4.16059613e-01 -7.31576085e-01
2.17672557e-01 8.57758820e-02 1.57172367e-01 -6.25712216e-01
-9.24182892e-01 4.52922225e-01 9.43044052e-02 -5.90280414e-01
-1.64239824e-01 -4.01244909e-01 -1.12358904e+00 -3.67342114e-01
-2.26295441e-01 -1.16759881e-01 1.05308485e+00 7.37247646e-01
8.48382950e-01 1.03969753e+00 5.32058299e-01 -1.41702521e+00
-5.61129153e-01 -1.26172304e+00 -3.04428577e-01 -2.43930444e-01
5.53819358e-01 -3.66895139e-01 -3.03875983e-01 3.93431008e-01] | [14.916571617126465, 5.566709041595459] |
43f6a82d-6079-432a-b6cf-1c9df360ef52 | caption-feature-space-regularization-for | 2204.08409 | null | https://arxiv.org/abs/2204.08409v1 | https://arxiv.org/pdf/2204.08409v1.pdf | Caption Feature Space Regularization for Audio Captioning | Audio captioning aims at describing the content of audio clips with human language. Due to the ambiguity of audio, different people may perceive the same audio differently, resulting in caption disparities (i.e., one audio may correlate to several captions with diverse semantics). For that, general audio captioning models achieve the one-to-many training by randomly selecting a correlated caption as the ground truth for each audio. However, it leads to a significant variation in the optimization directions and weakens the model stability. To eliminate this negative effect, in this paper, we propose a two-stage framework for audio captioning: (i) in the first stage, via the contrastive learning, we construct a proxy feature space to reduce the distances between captions correlated to the same audio, and (ii) in the second stage, the proxy feature space is utilized as additional supervision to encourage the model to be optimized in the direction that benefits all the correlated captions. We conducted extensive experiments on two datasets using four commonly used encoder and decoder architectures. Experimental results demonstrate the effectiveness of the proposed method. The code is available at https://github.com/PRIS-CV/Caption-Feature-Space-Regularization. | ['Yuan Dong', 'Zhanyu Ma', 'Ruoyi Du', 'Hong Yu', 'Yiming Zhang'] | 2022-04-18 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 3.63759637e-01 2.14658044e-02 -2.58326419e-02 -4.57546532e-01
-1.15635443e+00 -2.87757635e-01 1.87299415e-01 -9.85093694e-03
-5.65412156e-02 6.62534773e-01 7.96949923e-01 3.48214120e-01
1.69214785e-01 -2.73122430e-01 -9.35589135e-01 -7.70344377e-01
1.99880660e-01 2.46110260e-01 -3.05888951e-02 -4.52220216e-02
1.32519305e-02 -3.29671681e-01 -1.34582174e+00 5.02363682e-01
6.48496926e-01 1.10008550e+00 4.87484038e-01 3.06737363e-01
-1.37447804e-01 7.53867984e-01 -4.57310617e-01 -3.20440471e-01
8.91743004e-02 -7.15835810e-01 -4.94541794e-01 1.75041616e-01
3.55085671e-01 -9.38847438e-02 -3.94852072e-01 1.08347809e+00
5.72771907e-01 1.18354395e-01 2.86384225e-01 -1.33774221e+00
-7.65678704e-01 8.77543569e-01 -6.42016888e-01 3.45908552e-02
5.23096085e-01 2.42256615e-02 1.33635139e+00 -9.70671475e-01
1.65662333e-01 1.26102579e+00 3.19907725e-01 6.90218210e-01
-1.10380793e+00 -8.40525270e-01 3.61485749e-01 2.15289280e-01
-1.61841297e+00 -5.44368386e-01 1.12136829e+00 -4.42716837e-01
1.31953672e-01 3.68678957e-01 5.83814144e-01 1.14106572e+00
-2.02211082e-01 8.16788673e-01 6.60605907e-01 -1.88746914e-01
6.44686222e-02 3.52995157e-01 -1.71357710e-02 2.04493105e-01
-9.55356583e-02 -1.80770814e-01 -7.69521356e-01 -1.29792556e-01
3.94580990e-01 -1.42611563e-01 -5.62272847e-01 -1.42945677e-01
-1.25103569e+00 7.50860095e-01 4.11233723e-01 2.38840535e-01
-3.94663185e-01 2.18902141e-01 4.70422924e-01 8.18903968e-02
3.34279358e-01 4.19170260e-01 -2.44932145e-01 -1.45195723e-01
-6.71056986e-01 1.55673102e-01 4.13823724e-01 8.54502320e-01
8.03383648e-01 -6.09305725e-02 -4.61320639e-01 1.09378719e+00
6.40033901e-01 3.65999401e-01 5.29914141e-01 -7.73544312e-01
9.75779057e-01 1.91938877e-01 1.76034182e-01 -1.15862870e+00
-1.10837407e-02 -5.72500825e-01 -8.47445250e-01 -5.49833477e-01
2.89639458e-02 -3.89799058e-01 -6.38377666e-01 2.05299783e+00
1.20950624e-01 7.12692320e-01 2.52675004e-02 1.45845997e+00
8.38390887e-01 1.18603635e+00 1.92762852e-01 -2.27020994e-01
1.35294569e+00 -1.07063150e+00 -7.81925917e-01 -5.95902562e-01
2.74938971e-01 -1.00229609e+00 1.37231934e+00 1.03445873e-01
-8.85160327e-01 -5.75079441e-01 -9.60659504e-01 1.98978812e-01
2.30239004e-01 9.73136276e-02 2.64491975e-01 1.79369777e-01
-7.15249240e-01 6.32005036e-02 -4.86632347e-01 3.26864384e-02
2.73135781e-01 5.05950674e-03 -9.25802365e-02 -3.40650640e-02
-1.65808177e+00 3.67214054e-01 5.44118166e-01 2.22618908e-01
-8.79987121e-01 -6.06154561e-01 -8.12445998e-01 2.83181310e-01
2.58912355e-01 -5.19863248e-01 1.27406752e+00 -1.40263975e+00
-1.31719398e+00 4.46460962e-01 -2.85466194e-01 -3.24292988e-01
3.27673703e-01 -5.16396940e-01 -5.38332820e-01 1.23221070e-01
2.37251624e-01 1.00140738e+00 9.00266290e-01 -1.55473828e+00
-4.53491122e-01 5.77746928e-02 -5.25439382e-02 5.86995482e-01
-6.36905611e-01 5.25763370e-02 -7.61598825e-01 -7.91044235e-01
3.30714248e-02 -9.82891500e-01 6.88028634e-02 -1.50927454e-01
-4.45868731e-01 1.53558284e-01 5.56155026e-01 -6.75187707e-01
1.43703735e+00 -2.60315919e+00 2.13942289e-01 1.42725348e-01
-1.19424269e-01 5.32809868e-02 -3.49774778e-01 3.08674246e-01
-2.40972713e-01 1.18494198e-01 -2.66443610e-01 -4.88523334e-01
-1.14389762e-01 1.22399412e-01 -5.14883637e-01 1.53005973e-01
3.64892215e-01 5.18248081e-01 -1.02864814e+00 -7.01596916e-01
-5.06571271e-02 5.95595717e-01 -6.41219854e-01 4.52521265e-01
-2.19841406e-01 6.18661880e-01 -6.67246640e-01 3.55410159e-01
6.07609987e-01 -3.63583654e-01 -6.13902286e-02 -2.51013070e-01
1.22744657e-01 4.88027751e-01 -1.34107733e+00 1.69522643e+00
-5.00608206e-01 5.72615147e-01 5.46897724e-02 -8.85240078e-01
9.47800040e-01 7.36688375e-01 5.05352974e-01 -4.65660453e-01
9.95431244e-02 1.94972068e-01 -8.64472538e-02 -4.83794391e-01
3.08634847e-01 -1.27956420e-01 -1.64034560e-01 3.06298912e-01
-1.20898902e-01 9.62256193e-02 -6.18400685e-02 6.14734367e-02
6.01591349e-01 -2.07065225e-01 -8.23683888e-02 2.08029449e-01
6.87530041e-01 -4.00736123e-01 7.85389066e-01 5.08839607e-01
-1.74879596e-01 1.06678545e+00 4.48060513e-01 -5.10868952e-02
-8.88896704e-01 -9.88382518e-01 3.51496898e-02 1.13892734e+00
4.00278509e-01 -4.71544027e-01 -6.72482312e-01 -2.84125745e-01
-2.96596140e-01 7.25545168e-01 -5.13828814e-01 -3.27148050e-01
-4.90407109e-01 -4.65062618e-01 4.25907284e-01 3.43195677e-01
4.19662774e-01 -1.18273175e+00 -2.71492656e-02 2.28842765e-01
-6.55709505e-01 -1.00991774e+00 -1.00221181e+00 -7.90374428e-02
-5.22230387e-01 -5.26679397e-01 -8.33030224e-01 -1.06976032e+00
5.04020691e-01 3.52233768e-01 8.32903028e-01 -1.48852365e-02
3.69377822e-01 1.55574709e-01 -5.67820966e-01 -2.97272086e-01
-2.71230787e-01 3.23991776e-02 -9.83628072e-03 5.17147064e-01
2.90733576e-01 -5.64786136e-01 -4.61837292e-01 2.43152693e-01
-8.72014880e-01 2.18069196e-01 5.65907478e-01 8.72593641e-01
6.78963840e-01 -2.66549349e-01 7.09256887e-01 -4.61291462e-01
8.09999704e-01 -7.64007986e-01 -2.93995082e-01 6.99470937e-02
-2.14030191e-01 -1.17115572e-01 8.79807830e-01 -7.13904500e-01
-7.97664404e-01 3.35950136e-01 3.50265065e-03 -7.09452868e-01
-2.96689272e-02 7.58746028e-01 -4.39365059e-01 4.60668027e-01
3.13256204e-01 3.96873027e-01 -7.52206296e-02 -4.31725174e-01
2.17670590e-01 8.94034386e-01 4.59430665e-01 -4.99837279e-01
9.22154963e-01 9.91195738e-02 -5.20601511e-01 -6.62820935e-01
-1.14715385e+00 -4.88046139e-01 -3.58127505e-02 -3.61878216e-01
9.01900232e-01 -1.28777647e+00 -1.64497778e-01 1.40202001e-01
-1.31838679e+00 1.93977699e-01 -6.70591146e-02 8.49773824e-01
-4.75861460e-01 1.99427277e-01 -4.30251420e-01 -7.40583599e-01
-2.64586240e-01 -1.16137409e+00 1.12438428e+00 4.77198929e-01
-3.46922576e-01 -5.12397766e-01 1.25005662e-01 3.97173822e-01
8.23297873e-02 -5.79691157e-02 6.37200475e-01 -6.80824816e-01
-4.30577129e-01 -5.82768470e-02 -2.11882859e-01 4.86382335e-01
2.67425299e-01 -2.21023574e-01 -1.11661458e+00 -1.21550165e-01
1.66193508e-02 -4.25053418e-01 5.96543193e-01 2.23447710e-01
1.34936905e+00 -5.28584540e-01 2.60810181e-03 4.39007759e-01
1.13300800e+00 2.19826043e-01 5.92821598e-01 2.90340036e-01
7.60456920e-01 5.12144208e-01 8.02163780e-01 4.76482451e-01
1.76339060e-01 8.35067749e-01 3.68606448e-01 -6.48539811e-02
1.07188700e-02 -7.13748813e-01 5.01163840e-01 1.17420578e+00
3.53986979e-01 -2.75639266e-01 -7.11716712e-01 6.36558354e-01
-1.91496348e+00 -9.23473477e-01 1.30465612e-01 2.22089124e+00
1.04781556e+00 1.51605666e-01 3.34293842e-02 1.77383423e-02
1.28604937e+00 3.24298143e-01 -4.10261393e-01 -9.86450464e-02
-6.49121329e-02 -2.96440512e-01 5.40682748e-02 5.29238343e-01
-9.53456044e-01 7.18441486e-01 5.01414633e+00 7.68151760e-01
-1.29233265e+00 7.71877766e-02 6.12737298e-01 -2.03435346e-01
-4.60418224e-01 4.01106067e-02 -5.55449963e-01 9.51782823e-01
7.98843145e-01 -4.00274336e-01 4.85498071e-01 6.76515818e-01
5.18115759e-01 4.58088517e-01 -9.29026127e-01 1.18998098e+00
7.32506812e-02 -9.88767505e-01 2.92056561e-01 -3.62457097e-01
5.35295844e-01 -1.36905193e-01 1.00121230e-01 2.18357131e-01
-3.86376292e-01 -7.63930500e-01 1.06244564e+00 4.24885392e-01
5.23305655e-01 -6.89572811e-01 7.21419156e-01 1.43495470e-01
-1.28067064e+00 3.18442211e-02 -4.03568894e-01 1.32736713e-01
4.62901115e-01 5.09429812e-01 -7.83559024e-01 4.14143115e-01
6.82141423e-01 6.86895728e-01 -3.22500318e-01 1.24189925e+00
-2.17184439e-01 8.99046123e-01 -1.91466987e-01 -7.57998005e-02
2.82498479e-01 -1.99968174e-01 8.09744120e-01 1.22019911e+00
3.60894442e-01 -1.06575370e-01 2.36569062e-01 7.89202392e-01
-2.64072001e-01 3.66962194e-01 -3.34312797e-01 -1.20444164e-01
8.11940312e-01 9.29668188e-01 -1.71843797e-01 -1.09190047e-01
-5.37487864e-01 8.20688009e-01 6.16703928e-02 3.98809195e-01
-1.23894465e+00 -3.86469394e-01 5.39134026e-01 1.58039212e-01
2.48116016e-01 1.71383888e-01 -1.46191329e-01 -1.10320759e+00
4.10382718e-01 -8.64299059e-01 2.26252496e-01 -9.03549969e-01
-1.43696427e+00 7.38453031e-01 -9.96072590e-02 -1.65883899e+00
-5.98890856e-02 9.00556147e-02 -6.39149547e-01 8.58979106e-01
-1.47113252e+00 -7.99319983e-01 -3.46832752e-01 6.10748172e-01
5.99118948e-01 -1.22068115e-01 6.40497088e-01 7.03380108e-01
-5.70793986e-01 7.11324513e-01 6.93905726e-03 2.12110579e-01
1.07347703e+00 -8.94230366e-01 6.61340430e-02 5.70526600e-01
1.95660800e-01 4.21718836e-01 9.53935981e-01 -4.09986198e-01
-9.85202312e-01 -1.29837871e+00 8.75684679e-01 1.02927074e-01
7.33494163e-01 -4.44148868e-01 -1.20523417e+00 5.87188244e-01
9.05294418e-02 -2.06920460e-01 9.21553969e-01 2.33425070e-02
-3.57839167e-01 -4.46218610e-01 -6.07282400e-01 6.24019980e-01
6.21691823e-01 -6.50368094e-01 -7.16516495e-01 3.94625187e-01
1.11584151e+00 -2.90472746e-01 -6.05195284e-01 6.49310201e-02
4.26907986e-01 -5.85173249e-01 7.82884836e-01 -3.83234411e-01
6.57580495e-01 -5.95613122e-01 -2.83674300e-01 -1.36565137e+00
-2.64343530e-01 -7.72763729e-01 3.16096097e-02 1.66858518e+00
6.74355745e-01 -3.67707193e-01 5.08287668e-01 3.37781370e-01
-3.91190648e-01 -6.24056458e-01 -9.62603450e-01 -6.99821234e-01
-2.23992497e-01 -2.87233323e-01 6.97151721e-01 9.14135218e-01
1.93466600e-02 6.67372227e-01 -8.72945666e-01 3.43065977e-01
2.73798019e-01 6.40048683e-02 6.25749826e-01 -9.56084847e-01
-4.34900343e-01 -1.45009756e-01 -2.70238489e-01 -1.36486769e+00
2.32495099e-01 -7.27371454e-01 4.60593581e-01 -1.27034152e+00
2.22462028e-01 -3.92014802e-01 -5.07862329e-01 4.03568029e-01
-4.93748784e-01 1.09835580e-01 3.27254295e-01 4.47623700e-01
-7.60700762e-01 8.43963802e-01 1.16432118e+00 -2.23773956e-01
-3.59768778e-01 2.93913931e-02 -1.02684331e+00 5.23785949e-01
9.88774836e-01 -7.44347572e-01 -5.67921758e-01 -6.80657268e-01
1.78567737e-01 1.19867548e-01 3.11683983e-01 -1.10713923e+00
1.07287578e-01 -2.43041515e-01 1.44197345e-01 -2.34838352e-01
6.63521647e-01 -8.64930987e-01 1.28713444e-01 1.87780887e-01
-6.59995258e-01 -1.38596714e-01 1.30618930e-01 6.89803779e-01
-7.37753868e-01 -2.80796736e-01 7.47814655e-01 1.65592879e-01
-3.59200150e-01 2.32998013e-01 -7.91375190e-02 2.83038020e-01
7.25705028e-01 1.04204021e-01 -1.11132205e-01 -8.77139688e-01
-5.53634703e-01 4.76046532e-01 2.13660091e-01 6.91970348e-01
5.58145046e-01 -1.70301354e+00 -1.03691304e+00 -1.58026442e-01
3.22415113e-01 -6.83697127e-03 2.83994496e-01 5.74441552e-01
-6.58234060e-02 3.27221662e-01 -6.83488846e-02 -6.71477795e-01
-1.17478061e+00 3.70528400e-01 2.04045385e-01 -4.15319391e-03
-3.33970845e-01 8.50232184e-01 4.58257526e-01 1.06432401e-02
3.86043489e-01 7.75935352e-02 -3.16297203e-01 3.95922847e-02
6.29272580e-01 -1.21267706e-01 -1.81984633e-01 -8.42375576e-01
-4.04882163e-01 4.86104161e-01 -2.87817776e-01 -2.31405243e-01
1.18084526e+00 -3.13196391e-01 9.62081626e-02 6.46537185e-01
1.37076259e+00 1.04992941e-01 -1.35218096e+00 -1.18412741e-01
-2.66987860e-01 -5.20010471e-01 1.32356100e-02 -4.74399209e-01
-1.09007096e+00 7.60551810e-01 5.03541231e-01 2.47000918e-01
1.27700663e+00 9.03152078e-02 9.00440454e-01 8.81528556e-02
-4.49418798e-02 -9.73426819e-01 2.22963840e-01 4.39919651e-01
1.04460025e+00 -1.11269474e+00 -3.45559001e-01 -4.58477825e-01
-1.18742180e+00 7.64176309e-01 7.97211587e-01 -4.46695164e-02
3.88161898e-01 -9.32071880e-02 2.42781043e-01 8.91574025e-02
-7.96401620e-01 5.46208173e-02 3.27806562e-01 4.30457622e-01
6.17118895e-01 2.32361164e-02 -2.92914182e-01 9.50770974e-01
-2.53778458e-01 -1.09319352e-01 4.88530695e-01 6.80695295e-01
-3.81760955e-01 -1.05475152e+00 -5.46839356e-01 2.22027719e-01
-4.00863528e-01 -2.21440911e-01 -4.10930961e-01 2.77827919e-01
-2.94361338e-02 1.16901445e+00 2.65031829e-02 -5.02123237e-01
3.15699965e-01 1.05007716e-01 -9.27006602e-02 -6.93367064e-01
-3.99984479e-01 3.58234078e-01 -7.93520212e-02 -2.77208507e-01
-2.29614466e-01 -6.47431850e-01 -1.28760421e+00 2.47130781e-01
-4.17947412e-01 6.78594172e-01 3.91882151e-01 6.69032156e-01
4.79624927e-01 5.84142625e-01 1.02560985e+00 -6.57398403e-01
-5.24767399e-01 -1.04290986e+00 -4.61989999e-01 5.89237809e-01
4.08569336e-01 -5.92667162e-01 -6.50795877e-01 2.36246943e-01] | [15.259134292602539, 4.888879299163818] |
0582eae6-77a1-4716-ba20-669d705fd3c9 | efficient-generative-modeling-of-protein | 2103.03292 | null | https://arxiv.org/abs/2103.03292v3 | https://arxiv.org/pdf/2103.03292v3.pdf | Efficient generative modeling of protein sequences using simple autoregressive models | Generative models emerge as promising candidates for novel sequence-data driven approaches to protein design, and for the extraction of structural and functional information about proteins deeply hidden in rapidly growing sequence databases. Here we propose simple autoregressive models as highly accurate but computationally efficient generative sequence models. We show that they perform similarly to existing approaches based on Boltzmann machines or deep generative models, but at a substantially lower computational cost (by a factor between $10^2$ and $10^3$). Furthermore, the simple structure of our models has distinctive mathematical advantages, which translate into an improved applicability in sequence generation and evaluation. Within these models, we can easily estimate both the probability of a given sequence, and, using the model's entropy, the size of the functional sequence space related to a specific protein family. In the example of response regulators, we find a huge number of ca. $10^{68}$ possible sequences, which nevertheless constitute only the astronomically small fraction $10^{-80}$ of all amino-acid sequences of the same length. These findings illustrate the potential and the difficulty in exploring sequence space via generative sequence models. | ['Martin Weigt', 'Francesco Zamponi', 'Andrea Pagnani', 'Guido Uguzzoni', 'Jeanne Trinquier'] | 2021-03-04 | null | null | null | null | ['protein-design'] | ['medical'] | [ 4.07047898e-01 2.31794834e-01 2.71311760e-01 -3.28373045e-01
-9.09301460e-01 -6.56407714e-01 3.51925343e-01 -6.14890121e-02
-5.84090054e-01 1.30169415e+00 -6.17963932e-02 -6.00307524e-01
-2.43794397e-02 -6.50034964e-01 -8.51339161e-01 -1.29787886e+00
-8.98325518e-02 7.21482396e-01 -4.59731705e-02 -3.78185034e-01
2.43226245e-01 3.98104370e-01 -1.46398747e+00 -6.17882088e-02
9.03214216e-01 7.46406674e-01 6.41122699e-01 7.57999122e-01
-2.48741284e-01 3.83125097e-01 -6.17085695e-01 -5.23559809e-01
-1.86466500e-01 -7.02329099e-01 -6.42042339e-01 -1.43133029e-01
-1.76831067e-01 1.17226675e-01 -2.58470714e-01 9.55447733e-01
4.19322222e-01 2.00660810e-01 8.33650529e-01 -6.34979188e-01
-7.13210464e-01 3.07461888e-01 -2.52945632e-01 1.08147778e-01
2.94352710e-01 3.92855018e-01 1.08651185e+00 -8.16159904e-01
6.83647215e-01 8.20523977e-01 4.97429907e-01 6.84549630e-01
-1.40192020e+00 -3.14031661e-01 5.88140413e-02 1.30856812e-01
-1.30408800e+00 -2.96485990e-01 4.71782565e-01 -5.25848866e-01
1.52766562e+00 3.72820109e-01 7.26497948e-01 9.85293210e-01
4.18565363e-01 2.69464284e-01 8.99397850e-01 -4.47184443e-01
4.99869883e-01 -1.72517076e-01 1.61994454e-02 6.29952729e-01
1.52305111e-01 -1.02857865e-01 -4.30782914e-01 -7.27940261e-01
5.58188856e-01 7.63927400e-02 -2.17431679e-01 -4.27335948e-01
-9.30493593e-01 1.00418329e+00 -4.79866005e-02 1.27027363e-01
-5.14476717e-01 2.40242824e-01 1.15199603e-01 -1.79635465e-01
3.16541702e-01 3.56043220e-01 -6.56407118e-01 -5.10538280e-01
-8.39786530e-01 4.63592261e-01 8.08073878e-01 8.74695897e-01
7.84492731e-01 1.50759712e-01 4.51536566e-01 5.88322043e-01
2.66776145e-01 4.12253857e-01 5.05570531e-01 -5.51841140e-01
4.05601077e-02 3.41605663e-01 4.12610173e-01 -5.21053553e-01
-1.65976867e-01 -2.20070705e-01 -6.91236138e-01 -5.64853065e-02
4.89964634e-01 -3.41387764e-02 -9.31284368e-01 1.88697088e+00
2.72046655e-01 -2.17894554e-01 -1.67511404e-02 6.07420743e-01
2.06230223e-01 9.09558237e-01 4.39290375e-01 -4.93081868e-01
1.48315513e+00 -3.52530330e-01 -2.01027110e-01 -3.18336695e-01
3.43070745e-01 -6.15850747e-01 7.77765989e-01 2.68364578e-01
-1.15993333e+00 -2.72433281e-01 -8.96701276e-01 4.60875668e-02
-2.65563071e-01 -1.76189706e-01 9.96564746e-01 7.89931774e-01
-1.00130200e+00 7.06508934e-01 -8.21536362e-01 -8.76052603e-02
2.95629859e-01 4.49871927e-01 -2.31283695e-01 1.06238388e-01
-1.06133425e+00 8.58376324e-01 4.83373016e-01 3.61547023e-02
-7.61184454e-01 -4.88853693e-01 -6.58888698e-01 3.66024405e-01
1.46826759e-01 -6.65115058e-01 1.06905472e+00 -5.91941059e-01
-1.34725952e+00 7.67990530e-01 -6.21546865e-01 -5.77423871e-01
1.87702626e-02 7.37187639e-02 -4.74471785e-02 -9.31414664e-02
-4.09702450e-01 5.77276289e-01 4.44488525e-01 -7.05747128e-01
-1.92978069e-01 -4.25376475e-01 -2.80597538e-01 -1.51256518e-02
2.52975255e-01 8.63295123e-02 1.08564742e-01 -4.52308983e-01
5.11291400e-02 -1.08102596e+00 -8.51153374e-01 -3.65448654e-01
-1.80330858e-01 -1.23194292e-01 -1.89006589e-02 -8.06476355e-01
1.03970611e+00 -1.79907405e+00 4.66495037e-01 3.54324877e-01
2.20811740e-01 3.89518112e-01 1.00460619e-01 7.50784934e-01
-2.65782356e-01 1.09676093e-01 -5.47337472e-01 4.58576709e-01
4.83503379e-02 2.42565468e-01 -2.52731085e-01 3.10487509e-01
3.58170748e-01 1.16227567e+00 -7.43151248e-01 6.82245120e-02
5.34294918e-02 6.27728820e-01 -4.14544761e-01 7.67408535e-02
-5.74730873e-01 3.18738341e-01 -5.11197090e-01 3.17981094e-01
4.45095092e-01 -5.67225993e-01 7.87152469e-01 2.31112331e-01
7.33686984e-02 3.49815518e-01 -6.15422070e-01 1.42258012e+00
2.80749705e-02 3.51148158e-01 -2.65084177e-01 -1.19268262e+00
1.07928765e+00 1.55205891e-01 4.88199621e-01 -3.52014840e-01
-1.01048835e-02 1.03862338e-01 4.30423111e-01 -2.19086051e-01
4.34408665e-01 -7.30268240e-01 1.97023321e-02 5.06636977e-01
1.80903211e-01 4.78687175e-02 2.39735022e-01 1.09971957e-02
1.17355359e+00 2.56251931e-01 5.83826959e-01 -3.22509199e-01
2.47153595e-01 -1.09289840e-01 6.69126213e-01 6.79460168e-01
-1.34609893e-01 3.46925735e-01 5.77290654e-01 -5.70194840e-01
-1.63149023e+00 -8.97871971e-01 8.42714384e-02 1.13296092e+00
-2.22208142e-01 -4.95548368e-01 -1.00741148e+00 -1.75628662e-01
-2.15731442e-01 6.30903721e-01 -4.69878942e-01 -3.96382749e-01
-6.89331770e-01 -1.42556179e+00 3.52063745e-01 4.82535690e-01
-2.00600207e-01 -1.26846397e+00 -7.48050570e-01 5.00934780e-01
-2.61764258e-01 -5.55180609e-01 -2.77789980e-01 5.53203642e-01
-8.59617591e-01 -7.71114945e-01 -9.40871656e-01 -5.95592558e-01
6.24479830e-01 -8.44830051e-02 1.17127204e+00 -2.28962407e-01
-7.71793425e-01 -2.04663515e-01 -5.86198829e-03 -4.35697228e-01
-5.71236134e-01 -1.56644702e-01 -2.76661180e-02 -4.25739914e-01
8.00939322e-01 -8.27503800e-01 -7.71163046e-01 -3.21023501e-02
-7.59531975e-01 1.00632086e-02 4.79075640e-01 1.05051744e+00
7.57730007e-01 -1.92520007e-01 5.59663117e-01 -9.01497126e-01
5.69753468e-01 -5.26748359e-01 -6.63762331e-01 2.66226083e-01
-6.02111042e-01 3.80642682e-01 5.91500998e-01 -4.68937337e-01
-9.44843709e-01 2.70494670e-01 -6.25917614e-01 8.69885013e-02
-1.82441115e-01 6.05223179e-01 -1.22390911e-01 1.11435622e-01
6.22368753e-01 9.81639445e-01 1.39072344e-01 -5.03717124e-01
4.80287284e-01 2.48102471e-01 1.78061560e-01 -5.89626133e-01
3.69613975e-01 2.65305161e-01 -1.02680521e-02 -9.30854559e-01
-2.82734334e-01 -2.21006274e-01 -4.75861549e-01 2.24636778e-01
6.90580070e-01 -6.61875129e-01 -1.13843739e+00 1.30966872e-01
-8.87463212e-01 -1.21258376e-02 -1.67604834e-01 3.34090292e-01
-1.06682849e+00 6.81599498e-01 -7.30381370e-01 -1.13264275e+00
-4.46641147e-01 -1.24393463e+00 1.03255749e+00 1.47130281e-01
-4.78493214e-01 -6.70958519e-01 1.89802095e-01 2.55342752e-01
3.76211196e-01 2.19376534e-01 1.58612502e+00 -5.72607219e-01
-6.17261529e-01 -1.85855165e-01 3.41543593e-02 1.12513728e-01
-1.00035198e-01 -1.25609199e-02 -5.77643514e-01 -2.11069912e-01
3.97354104e-02 -1.82944328e-01 9.39280748e-01 5.01890063e-01
9.93083417e-01 -3.58054310e-01 -3.20495814e-01 3.12365800e-01
1.30936706e+00 7.87729800e-01 1.01263154e+00 3.14322226e-02
3.02477986e-01 4.18191016e-01 3.81980270e-01 5.24211526e-01
-1.50505185e-01 5.84698319e-01 1.77231267e-01 8.35800245e-02
4.49264646e-01 -2.82691985e-01 1.76268697e-01 7.67365158e-01
-3.35151374e-01 -3.18348140e-01 -8.25031698e-01 3.79793495e-01
-1.74327195e+00 -1.13491178e+00 1.96189448e-01 2.27544928e+00
1.12874830e+00 3.32769156e-02 3.41056108e-01 -2.17198431e-01
5.94453871e-01 -1.23324245e-01 -7.51219392e-01 -5.25434852e-01
-2.38218218e-01 8.65318716e-01 4.67487484e-01 4.02791321e-01
-6.63810611e-01 8.05371881e-01 7.75361395e+00 1.04035962e+00
-8.63221407e-01 -3.21167856e-01 8.41707706e-01 -1.67989448e-01
-5.23110509e-01 1.86360523e-01 -8.37499619e-01 7.45693028e-01
1.49775016e+00 -3.99287283e-01 3.06414127e-01 9.42273557e-01
1.26764942e-02 -9.36787799e-02 -8.28532875e-01 8.85756612e-01
-1.32944256e-01 -1.58359289e+00 8.25356543e-02 5.41406631e-01
3.84057164e-01 4.02287729e-02 6.46875128e-02 3.02211083e-02
4.68821317e-01 -1.37585735e+00 4.27480340e-01 5.44367194e-01
6.61239505e-01 -1.04245925e+00 4.56431299e-01 5.95227718e-01
-6.12468302e-01 1.15535535e-01 -8.60996604e-01 2.11359244e-02
3.81362945e-01 7.71396518e-01 -8.64143968e-01 8.62717554e-02
3.49061370e-01 4.31860378e-03 5.68795076e-04 6.97325826e-01
7.15672374e-02 6.51743770e-01 -3.19155723e-01 -4.86949682e-01
2.85576642e-01 -5.53015471e-01 3.08827192e-01 1.15322471e+00
4.92922217e-01 4.45019960e-01 -2.00479224e-01 9.40020561e-01
5.31974323e-02 7.27159157e-02 -4.00289774e-01 -5.59360862e-01
2.12729618e-01 9.00060058e-01 -6.60564661e-01 -3.88670474e-01
-1.20553583e-01 9.08924937e-01 4.31689769e-01 4.05095309e-01
-8.79537404e-01 -4.09685940e-01 8.40489864e-01 -5.91211617e-02
6.39114439e-01 -4.33200657e-01 1.99131697e-01 -1.03390086e+00
-1.93949565e-02 -1.03266740e+00 2.76217679e-03 -6.22032046e-01
-1.06642199e+00 4.92912441e-01 -3.20852906e-01 -5.79655290e-01
-6.75659597e-01 -8.24645579e-01 -3.40104401e-02 1.17836463e+00
-8.83320212e-01 -6.03755832e-01 4.63351518e-01 -5.27777821e-02
4.95104879e-01 -8.44946802e-02 1.16673028e+00 -2.77246386e-02
-3.97297084e-01 3.06861639e-01 7.92755544e-01 -3.41064572e-01
5.14965598e-03 -1.18563294e+00 1.04870868e+00 3.97374690e-01
2.04707474e-01 1.07771885e+00 9.70486581e-01 -5.46541214e-01
-1.43919802e+00 -5.73051095e-01 1.01513731e+00 -5.54954052e-01
4.53779161e-01 -6.45034194e-01 -1.14671147e+00 5.31726837e-01
-2.30169699e-01 -5.52274764e-01 1.18981147e+00 8.37790817e-02
-2.34044492e-01 5.40753305e-01 -1.08648586e+00 5.69188058e-01
1.01247525e+00 -4.85841483e-01 -3.82115930e-01 3.25418770e-01
5.73327184e-01 -1.05377361e-01 -8.30120802e-01 2.23155499e-01
8.74204755e-01 -1.09028447e+00 1.20429111e+00 -9.94012833e-01
3.09719980e-01 1.63643174e-02 -1.81223333e-01 -1.05044651e+00
-7.09002674e-01 -7.76232243e-01 -1.39320478e-01 7.06571758e-01
6.05907500e-01 -5.14457703e-01 1.15927219e+00 8.22569728e-01
2.17070714e-01 -1.07324946e+00 -1.09632123e+00 -6.32364631e-01
3.91531289e-01 -2.61342861e-02 5.80489516e-01 4.76348251e-01
2.39563912e-01 1.19096324e-01 -4.53566402e-01 -5.34859955e-01
3.94393593e-01 1.34605169e-01 2.73295403e-01 -9.63984907e-01
-7.90744364e-01 -2.80143261e-01 -5.46576321e-01 -1.20189822e+00
-1.89105030e-02 -5.25839865e-01 2.03592092e-01 -1.10549867e+00
6.05915308e-01 -2.17071742e-01 -4.07727778e-01 1.41610950e-01
-3.04460794e-01 -9.30379257e-02 -2.18262762e-01 9.75840446e-03
-2.37840742e-01 5.38495421e-01 7.27127254e-01 1.93867177e-01
2.22150683e-01 -5.62135205e-02 -9.07476485e-01 5.90689361e-01
7.88132012e-01 -2.86036879e-01 -3.50294799e-01 6.21241964e-02
5.20693898e-01 1.43117845e-01 1.02682382e-01 -3.30158144e-01
-3.39362115e-01 -5.55847585e-01 5.10410726e-01 -4.94301796e-01
5.91865480e-01 -3.01280469e-01 5.79071462e-01 6.26784027e-01
-3.12695742e-01 1.40694171e-01 9.74892899e-02 7.47921646e-01
1.79978848e-01 -3.09801817e-01 7.40810275e-01 -6.38882935e-01
-4.57035094e-01 1.17823943e-01 -8.99449706e-01 -2.87147105e-01
1.04785109e+00 -3.18946242e-01 -8.01845267e-02 -3.92651618e-01
-7.77506590e-01 -5.10408700e-01 5.27695537e-01 1.03583150e-01
5.90722144e-01 -7.89310575e-01 -4.62799013e-01 1.87319979e-01
2.63243299e-02 -4.26010132e-01 4.33130413e-01 4.22394753e-01
-6.82607055e-01 8.17861021e-01 -1.87132940e-01 -3.10853690e-01
-1.20088089e+00 7.35264182e-01 1.46567807e-01 -2.76454985e-01
-3.22669655e-01 1.08422959e+00 4.15412009e-01 -1.55758873e-01
-3.51785153e-01 -1.94699666e-03 2.68933535e-01 -4.07079577e-01
5.28842211e-01 1.74913540e-01 -9.44409370e-02 -6.91856921e-01
-3.51576447e-01 2.85536200e-01 -3.00450593e-01 -1.38982590e-02
1.36512744e+00 8.66010934e-02 -1.92184210e-01 2.44687587e-01
9.20418203e-01 -2.88304001e-01 -1.16018927e+00 1.42545998e-02
1.08303219e-01 -1.59125119e-01 -6.01780295e-01 -7.20867515e-01
-3.20913553e-01 9.16454375e-01 4.33168501e-01 4.32233028e-02
6.97730541e-01 2.22965285e-01 8.33420932e-01 2.96636283e-01
6.08604491e-01 -7.79155850e-01 -3.70287001e-01 3.34756672e-01
3.65903974e-01 -8.11376870e-01 -8.49205777e-02 -3.14496636e-01
-3.72463465e-01 8.76333892e-01 1.43052936e-01 6.87049842e-03
2.26524055e-01 2.18921974e-01 -3.10968548e-01 -1.56095028e-01
-1.12905467e+00 2.69709737e-04 -2.84418724e-02 8.65301430e-01
8.75526369e-01 3.34247798e-01 -5.51456392e-01 3.75910282e-01
-3.66181612e-01 -8.14273134e-02 2.83503652e-01 9.28808510e-01
-9.20661747e-01 -1.37410009e+00 -2.61283964e-02 4.60251063e-01
-6.37067437e-01 -4.39906687e-01 -4.04711306e-01 2.70184100e-01
-2.32485369e-01 6.77348912e-01 -1.86357915e-01 -4.77730632e-02
-1.42635956e-01 6.53886139e-01 7.85314381e-01 -4.41040635e-01
-1.99476033e-01 2.51997173e-01 -5.11361547e-02 -3.13409239e-01
-4.91457409e-04 -6.10748410e-01 -1.22855902e+00 -3.28193367e-01
-3.13942373e-01 5.33817172e-01 8.51169944e-01 8.91233265e-01
6.22584283e-01 1.34216383e-01 2.25065783e-01 -6.46477938e-01
-7.83330798e-01 -8.13547075e-01 -5.30378640e-01 2.51402766e-01
-7.85872564e-02 -3.75791848e-01 -9.80355777e-03 2.03864455e-01] | [4.714311599731445, 5.5777268409729] |
50bc9f24-c84f-4947-b137-b13dc287647d | joint-acoustic-echo-cancellation-and-blind | 2205.06473 | null | https://arxiv.org/abs/2205.06473v2 | https://arxiv.org/pdf/2205.06473v2.pdf | Joint Acoustic Echo Cancellation and Blind Source Extraction based on Independent Vector Extraction | We describe a joint acoustic echo cancellation (AEC) and blind source extraction (BSE) approach for multi-microphone acoustic frontends. The proposed algorithm blindly estimates AEC and beamforming filters by maximizing the statistical independence of a non-Gaussian source of interest and a stationary Gaussian background modeling interfering signals and residual echo. Double talk-robust and fast-converging parameter updates are derived from a global maximum-likelihood objective function resulting in a computationally efficient Newton-type update rule. Evaluation with simulated acoustic data confirms the benefit of the proposed joint AEC and beamforming filter estimation in comparison to updating both filters individually. | ['Walter Kellermann', 'Zbyněk Koldovský', 'Thomas Haubner'] | 2022-05-13 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [ 1.16654217e-01 -4.48765129e-01 1.02085471e+00 -2.83063531e-01
-1.50493121e+00 -5.56846678e-01 3.99940014e-01 -1.74176350e-01
-7.19053030e-01 4.09450263e-01 5.53651989e-01 -1.45006344e-01
-2.40814209e-01 2.49047596e-02 -5.31496346e-01 -9.39399004e-01
-3.22802395e-01 -3.44297439e-02 1.99239224e-01 3.46920311e-01
-5.97668923e-02 3.90427679e-01 -1.44432974e+00 -3.33658904e-01
5.91945767e-01 9.75210428e-01 4.61806148e-01 1.51306093e+00
1.17179878e-01 2.45484769e-01 -9.64086652e-01 -3.09315950e-01
3.07875872e-01 -3.13712001e-01 2.90214688e-01 -1.59271240e-01
3.39126766e-01 -2.08653718e-01 -8.09916556e-02 1.23505902e+00
1.10932481e+00 4.12814081e-01 6.75477803e-01 -8.43073428e-01
1.55238211e-01 4.16575730e-01 -2.45881304e-01 2.52681196e-01
2.98962831e-01 -8.46834630e-02 6.51374519e-01 -1.46172047e+00
-2.73672432e-01 1.15106893e+00 1.03095007e+00 2.74757624e-01
-7.85967350e-01 -9.60539281e-01 -1.91528872e-01 -9.39608142e-02
-1.56595027e+00 -1.14997852e+00 7.47908652e-01 -1.68800324e-01
7.64520645e-01 5.71432233e-01 3.13754886e-01 7.32975900e-01
-8.36759359e-02 6.12505257e-01 1.08608484e+00 -6.19398594e-01
5.39238691e-01 3.14767808e-02 -1.04223393e-01 4.49475378e-01
3.59415233e-01 3.30559641e-01 -8.88928950e-01 -5.25041878e-01
2.91533619e-01 -6.47209823e-01 -5.78664899e-01 9.03838575e-02
-1.00797951e+00 3.32559764e-01 -3.20203096e-01 3.04755241e-01
-6.34501874e-01 4.65824604e-01 -1.01822965e-01 3.84520902e-03
3.77459645e-01 5.72716370e-02 -4.85200062e-02 -1.66959152e-01
-1.19326878e+00 -8.37110728e-03 1.26773643e+00 9.31791246e-01
3.14573020e-01 9.72416103e-01 5.01097701e-02 1.07611334e+00
1.13269162e+00 1.70997930e+00 1.76850066e-01 -9.22671258e-01
3.41678351e-01 -6.10725701e-01 4.73754078e-01 -1.01681805e+00
-3.67966443e-01 -1.04233158e+00 -7.21766889e-01 1.69157565e-01
3.42629910e-01 -7.06718862e-01 -6.45134807e-01 1.57474387e+00
4.98516351e-01 7.10829794e-01 2.26391584e-01 9.32122111e-01
6.44998968e-01 8.26452792e-01 -1.45873249e-01 -5.61910391e-01
9.79156137e-01 -4.20475841e-01 -1.27486479e+00 -5.01939178e-01
-2.29923710e-01 -1.23437107e+00 -2.47290656e-02 6.32414460e-01
-1.31410539e+00 -3.80898088e-01 -1.22410369e+00 5.61349094e-01
-2.65988186e-02 -8.54690839e-03 3.09859794e-02 1.44006443e+00
-1.18921709e+00 -1.71354443e-01 -6.31792247e-01 4.87282455e-01
-1.98832870e-01 4.15730804e-01 -7.13237971e-02 6.23134598e-02
-7.82887876e-01 5.45061409e-01 -2.22269967e-01 6.48137033e-01
-1.09826660e+00 -9.00543451e-01 -7.63087988e-01 1.56587400e-02
8.99775848e-02 -2.76446134e-01 1.46534967e+00 -7.15629339e-01
-1.79123557e+00 -4.68735285e-02 -7.42154002e-01 -4.12251711e-01
2.11392224e-01 -6.59860551e-01 -1.06103444e+00 1.52233958e-01
-2.11167186e-01 -2.28709862e-01 1.48742056e+00 -1.74946880e+00
-7.53933609e-01 -2.11784579e-02 -8.69517922e-01 4.21922922e-01
-2.23799691e-01 2.53454328e-01 -4.30562466e-01 -8.90600979e-01
4.51141149e-01 -4.55374867e-01 -2.85117984e-01 -3.62952679e-01
-2.87050277e-01 5.79516053e-01 4.80821788e-01 -1.18534636e+00
1.39421403e+00 -2.24405956e+00 -1.75008267e-01 7.56390929e-01
2.36225259e-02 3.31763506e-01 -1.85826972e-01 3.12013984e-01
7.41714910e-02 -4.97846007e-01 -3.19414854e-01 -7.52826095e-01
2.65862681e-02 -2.36698896e-01 -6.14751764e-02 8.64049792e-01
-2.40319565e-01 9.42698643e-02 -9.29241121e-01 -3.75330627e-01
3.00768107e-01 7.92042434e-01 -3.72428179e-01 5.40931761e-01
5.69077134e-01 3.52466494e-01 -2.09051967e-02 6.42315626e-01
1.04960895e+00 4.70713198e-01 -4.46348712e-02 -2.39275321e-01
-3.06076407e-01 -9.43179205e-02 -2.19416451e+00 1.11189616e+00
-7.19889104e-01 7.65271306e-01 1.51132703e+00 -5.87852478e-01
8.33733916e-01 7.23362446e-01 2.41482496e-01 -2.91944414e-01
1.59251451e-01 8.03276420e-01 2.18596961e-02 -2.55388647e-01
3.92252952e-01 -4.71618772e-01 2.82754213e-01 3.52743953e-01
3.94175261e-01 -1.39657721e-01 -3.01124066e-01 1.33240253e-01
1.08706129e+00 -6.13438189e-01 3.81139666e-01 -5.03983736e-01
8.98467481e-01 -9.08778727e-01 3.96078557e-01 1.13895619e+00
-1.84585422e-01 5.01087308e-01 -8.03080976e-01 5.64871550e-01
-3.49880546e-01 -1.18428826e+00 9.99900922e-02 9.42259610e-01
-6.28522784e-02 -1.70921445e-01 -7.55160391e-01 1.05342176e-02
-3.10276169e-02 9.75542188e-01 -7.14866910e-03 2.29661062e-01
-7.59110630e-01 -5.27459919e-01 5.38253784e-01 9.77903754e-02
6.96936250e-02 -2.86022604e-01 -2.30104178e-01 6.72915339e-01
-1.37997821e-01 -1.02109206e+00 -8.63388777e-01 4.03770953e-01
-4.96547729e-01 -5.50636649e-01 -9.19915199e-01 -5.88963687e-01
5.51130831e-01 5.68880379e-01 7.41872370e-01 -2.81146467e-01
-5.23254611e-02 1.30324614e+00 -3.61003131e-02 -7.50124633e-01
-7.10807741e-01 -1.01802325e+00 5.92496276e-01 6.14246249e-01
-1.25556961e-01 -6.64445102e-01 -5.97989321e-01 3.50570560e-01
-5.38246155e-01 -5.38288653e-01 3.97622377e-01 4.70419168e-01
2.09920987e-01 1.98229387e-01 4.59683061e-01 -7.01813400e-02
8.36185753e-01 -4.04620856e-01 -9.90978360e-01 1.12527862e-01
-4.48385566e-01 -2.31730789e-01 3.86773080e-01 -4.79470968e-01
-1.56380999e+00 2.71272302e-01 -3.93439144e-01 -3.74136209e-01
-1.83853000e-01 1.54447764e-01 -4.10829931e-01 -3.85918587e-01
2.83279419e-01 4.64085311e-01 -2.83344179e-01 -6.06337547e-01
4.25386518e-01 8.95044565e-01 1.08768559e+00 -2.63350874e-01
1.25825787e+00 3.23412031e-01 2.86472961e-02 -1.34093499e+00
-1.79006130e-01 -1.04792547e+00 -2.14253277e-01 -6.01204753e-01
5.14176369e-01 -1.08082056e+00 -8.43118072e-01 8.57811809e-01
-1.28970480e+00 2.07838848e-01 2.70729028e-02 1.20155454e+00
-8.29526260e-02 5.16854584e-01 -2.11181611e-01 -1.88921165e+00
-3.66453648e-01 -8.48653913e-01 8.30669105e-01 4.34835851e-02
-1.03584863e-01 -1.13850987e+00 2.97266573e-01 1.77805796e-02
7.94225156e-01 -2.94238448e-01 -3.11200358e-02 -6.16753817e-01
-3.90944034e-01 -5.74006557e-01 1.96803659e-01 6.31923616e-01
5.68072423e-02 -3.12288821e-01 -1.39818215e+00 -3.17414373e-01
4.20349568e-01 4.37234938e-01 5.29991865e-01 7.75235653e-01
2.62768298e-01 -3.59348029e-01 -1.90745592e-01 5.74034035e-01
1.57150471e+00 4.24201965e-01 9.38908458e-02 -3.14034551e-01
3.63974601e-01 2.84065306e-01 1.87280521e-01 7.39772260e-01
-1.39921695e-01 3.39235872e-01 2.44990215e-01 -1.16209984e-02
-2.97961980e-01 1.32197350e-01 4.79973793e-01 1.52056694e+00
2.97684640e-01 -6.94109917e-01 -6.17199838e-01 6.62814498e-01
-1.17504549e+00 -7.95145273e-01 -6.18198931e-01 2.48902869e+00
6.38914168e-01 -2.76464313e-01 -2.93434143e-01 3.76558393e-01
7.93542981e-01 -6.32490814e-02 -2.61466175e-01 -1.53886348e-01
-3.01374614e-01 4.95803058e-01 1.08347833e+00 1.34861267e+00
-9.55724061e-01 1.65431574e-01 7.26507616e+00 9.28412974e-01
-4.99534100e-01 5.55968583e-01 -1.77880153e-01 -1.16447747e-01
-3.12077224e-01 -4.06820685e-01 -8.67436588e-01 2.46573940e-01
1.19716287e+00 -5.56737520e-02 5.07159829e-01 3.08288395e-01
3.60365838e-01 -2.99231052e-01 -7.32562304e-01 1.30591691e+00
2.67951250e-01 -8.52202117e-01 -5.44595897e-01 -9.93571207e-02
3.77830535e-01 -7.13866111e-03 1.82109196e-02 -1.14791654e-01
4.58434135e-01 -5.04540324e-01 1.12703240e+00 7.44781315e-01
3.80792290e-01 -4.34068084e-01 4.04834539e-01 1.60469994e-01
-1.26748383e+00 -2.68732965e-01 3.70703526e-02 1.88102722e-01
6.38825059e-01 9.88848031e-01 -7.57354021e-01 4.06987250e-01
6.06135607e-01 -1.49389610e-01 4.25022542e-02 1.72194445e+00
-8.71826634e-02 1.26390934e+00 -8.54136288e-01 -2.02240586e-01
7.41214072e-03 -1.87803015e-01 1.43555176e+00 1.95215642e+00
7.08162367e-01 4.73895431e-01 -3.71332109e-01 4.86705124e-01
9.99548808e-02 2.32161433e-01 3.01723331e-02 2.87361324e-01
8.01927269e-01 1.21863711e+00 -5.10888398e-01 -7.81102329e-02
-4.02527660e-01 7.35591590e-01 -5.44156492e-01 7.98548639e-01
-5.70510149e-01 -4.97681558e-01 6.09552324e-01 -3.62704188e-01
5.40768981e-01 -4.21952367e-01 -4.40521389e-02 -5.48288107e-01
-1.39338076e-01 -7.61978149e-01 6.32824004e-02 -6.12030029e-01
-1.05156577e+00 2.81129807e-01 -1.42888919e-01 -1.35530043e+00
-2.37286806e-01 -5.48249841e-01 -8.85054290e-01 1.30344677e+00
-1.47070611e+00 -4.45654929e-01 1.41756415e-01 7.02796757e-01
3.37896705e-01 -3.42067152e-01 8.08204353e-01 5.36855698e-01
-3.00017715e-01 6.94760680e-01 8.32092226e-01 -1.55309737e-01
4.42645252e-01 -1.32518816e+00 2.72475444e-02 1.40868461e+00
-1.56812277e-02 7.65685141e-01 1.32377303e+00 -5.10722280e-01
-1.72529924e+00 -6.94484532e-01 7.08090067e-01 -1.26089752e-01
6.76351011e-01 -6.01874053e-01 -7.35253513e-01 -7.13551864e-02
4.12452996e-01 -1.28217489e-01 1.01792097e+00 -2.51748174e-01
-1.34174466e-01 -2.49373868e-01 -9.75597978e-01 1.87742859e-01
5.18343449e-01 -3.69038045e-01 -3.45448613e-01 2.25738674e-01
2.72990555e-01 -3.24976534e-01 -4.19098407e-01 6.15074001e-02
6.75983965e-01 -7.41533101e-01 1.24580944e+00 1.99401632e-01
-5.80107749e-01 -5.70764959e-01 -6.43837452e-01 -1.30566108e+00
-2.40038365e-01 -1.42247725e+00 -2.41451800e-01 1.53619778e+00
5.60918748e-01 -6.91950083e-01 2.70033360e-01 3.87255222e-01
-3.42911690e-01 2.07843661e-01 -1.33186877e+00 -9.64487612e-01
-4.82525200e-01 -1.23740101e+00 8.01905021e-02 2.16319486e-01
-3.47703695e-01 3.91683839e-02 -7.77025938e-01 1.12815654e+00
1.35515344e+00 -5.83177507e-01 5.85680187e-01 -9.68691349e-01
-8.12093675e-01 -2.01228321e-01 -1.50651606e-02 -1.19892251e+00
-6.91801533e-02 -4.34317142e-01 9.25364733e-01 -1.41076481e+00
-5.41665733e-01 -6.39828965e-02 -3.78526092e-01 -2.48770371e-01
-3.87205601e-01 1.11127935e-01 1.53618082e-01 -3.04441929e-01
-4.60321099e-01 5.21822274e-01 6.19084477e-01 -5.75038679e-02
-1.47631302e-01 6.74098432e-01 -3.44943345e-01 7.54752815e-01
2.58940935e-01 -7.78119206e-01 -2.12227955e-01 -4.88542229e-01
1.38650566e-01 2.63186067e-01 2.84959376e-01 -1.24144018e+00
8.47309589e-01 4.63525593e-01 2.25902945e-01 -5.44662356e-01
8.48767519e-01 -1.10770321e+00 3.76569122e-01 2.98951715e-01
-7.17249438e-02 -4.26598787e-01 3.79679412e-01 9.98283386e-01
-1.68012679e-01 -5.55977345e-01 7.30220139e-01 2.03613907e-01
-3.62547606e-01 -3.89985055e-01 -1.00883627e+00 -8.46683010e-02
2.68976450e-01 -1.20593823e-01 2.31096402e-01 -1.09139895e+00
-7.50616491e-01 -1.91202555e-02 -6.59483969e-01 5.15497813e-04
9.00285661e-01 -1.02736366e+00 -1.19562662e+00 3.34529787e-01
-4.02344823e-01 -5.45563400e-01 2.99417675e-01 7.98894048e-01
-8.10069889e-02 3.66334230e-01 7.03023851e-01 -4.99071509e-01
-1.48577571e+00 -3.54556181e-02 5.74634731e-01 3.35875601e-01
-1.17496401e-01 1.48199069e+00 -8.94822925e-03 -2.26329803e-01
6.76451981e-01 -1.56327654e-02 1.34795634e-02 -2.26878196e-01
9.36271667e-01 8.75100136e-01 1.97097033e-01 -9.42787349e-01
-4.47414547e-01 5.83714902e-01 6.76375687e-01 -9.16621387e-01
9.29814279e-01 -5.84606826e-01 -1.20312050e-01 3.93360436e-01
1.29389989e+00 9.92013097e-01 -1.04910672e+00 -3.77520978e-01
-2.08409607e-01 -5.57774603e-01 5.66580176e-01 -1.01603675e+00
-6.87670350e-01 6.57431126e-01 1.14168775e+00 1.22236144e-02
1.24459481e+00 -3.45810443e-01 4.48867083e-01 3.26641440e-01
1.07732229e-01 -9.01510715e-01 -3.33149612e-01 2.80050695e-01
8.74633968e-01 -7.69383907e-01 -4.82959524e-02 -2.59466916e-01
-2.61549670e-02 9.78040397e-01 -9.38587040e-02 1.49698183e-01
1.25974500e+00 9.25029874e-01 5.31587720e-01 1.92437589e-01
-2.76989698e-01 -1.41145229e-01 5.71029544e-01 9.14381623e-01
1.95381969e-01 8.43634456e-03 -7.32993893e-03 7.50200152e-01
-1.66732669e-01 -5.36097527e-01 2.68669486e-01 7.84461081e-01
-7.53902733e-01 -7.11744547e-01 -1.21095610e+00 1.64410934e-01
-6.21604264e-01 -4.33573306e-01 4.71667871e-02 2.19359979e-01
-2.67030299e-01 1.85779583e+00 -2.14324877e-01 -1.10344164e-01
4.04267013e-01 6.50694221e-02 1.19202271e-01 -1.43032908e-01
-6.40766919e-01 1.12769175e+00 1.29786670e-01 -4.35094416e-01
-4.73657191e-01 -9.28917229e-01 -1.05697453e+00 2.18026266e-01
-7.86084533e-01 5.72091520e-01 1.31195593e+00 7.78415859e-01
9.66144446e-03 6.59117341e-01 7.72747338e-01 -1.00895035e+00
-7.36871481e-01 -8.84421051e-01 -7.90852070e-01 -1.92859933e-01
7.68555403e-01 -2.36521810e-01 -1.06601810e+00 8.80800560e-02] | [15.135476112365723, 5.764640808105469] |
1295625b-d370-4c69-802f-ca1b5d36ae24 | tsk-fuzzy-system-towards-few-labeled | 2110.05610 | null | https://arxiv.org/abs/2110.05610v3 | https://arxiv.org/pdf/2110.05610v3.pdf | TSK Fuzzy System Towards Few Labeled Incomplete Multi-View Data Classification | Data collected by multiple methods or from multiple sources is called multi-view data. To make full use of the multi-view data, multi-view learning plays an increasingly important role. Traditional multi-view learning methods rely on a large number of labeled and completed multi-view data. However, it is expensive and time-consuming to obtain a large number of labeled multi-view data in real-world applications. Moreover, multi-view data is often incomplete because of data collection failures, self-deficiency, or other reasons. Therefore, we may have to face the problem of fewer labeled and incomplete multi-view data in real application scenarios. In this paper, a transductive semi-supervised incomplete multi-view TSK fuzzy system modeling method (SSIMV_TSK) is proposed to address these challenges. First, in order to alleviate the dependency on labeled data and keep the model interpretable, the proposed method integrates missing view imputation, pseudo label learning of unlabeled data, and fuzzy system modeling into a single process to yield a model with interpretable fuzzy rules. Then, two new mechanisms, i.e. the bidirectional structural preservation of instance and label, as well as the adaptive multiple alignment collaborative learning, are proposed to improve the robustness of the model. The proposed method has the following distinctive characteristics: 1) it can deal with the incomplete and few labeled multi-view data simultaneously; 2) it integrates the missing view imputation and model learning as a single process, which is more efficient than the traditional two-step strategy; 3) attributed to the interpretable fuzzy inference rules, this method is more interpretable. Experimental results on real datasets show that the proposed method significantly outperforms the state-of-the-art methods. | ['Shitong Wang', 'Kup-Sze Choi', 'Te Zhang', 'Qiongdan Lou', 'Zhaohong Deng', 'Wei zhang'] | 2021-10-08 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [-9.51270610e-02 -3.21049422e-01 -4.88734394e-01 -4.62126017e-01
-5.87976694e-01 -5.53188980e-01 1.61202878e-01 -9.61634740e-02
8.46879855e-02 6.63776755e-01 8.02151263e-02 1.51791573e-01
-3.17236096e-01 -7.43949175e-01 -4.11792994e-01 -8.13854814e-01
8.14598203e-01 7.21145868e-01 5.93413822e-02 -1.21166401e-01
1.99376389e-01 1.59179401e-02 -1.47134364e+00 6.09670937e-01
1.13967025e+00 9.00265217e-01 2.54133552e-01 -1.06755570e-01
-3.08365494e-01 8.19351435e-01 -9.98839885e-02 -2.78570592e-01
1.76829770e-01 -2.92103827e-01 -5.36980093e-01 4.20037419e-01
-6.86619133e-02 -2.81049311e-01 3.60334694e-01 1.20330989e+00
3.61665040e-01 -3.51988673e-02 4.78652686e-01 -1.35085225e+00
-6.74141347e-01 3.85952324e-01 -9.02596116e-01 -2.32727483e-01
1.47857398e-01 -3.46737981e-01 6.46803617e-01 -1.21399784e+00
5.86252630e-01 1.33560562e+00 5.31499565e-01 2.88440883e-01
-1.17421365e+00 -7.99801469e-01 3.12323511e-01 5.70327997e-01
-1.43417740e+00 -3.87322336e-01 9.97089148e-01 -4.61094767e-01
1.40582517e-01 6.91992044e-02 5.68496466e-01 7.40325928e-01
-1.16046062e-02 5.45277536e-01 1.64427543e+00 -4.29537535e-01
1.48939312e-01 4.24677819e-01 2.39817679e-01 5.82716346e-01
4.11837339e-01 -9.45783705e-02 -2.01376215e-01 -6.90678358e-02
5.50277293e-01 6.77611768e-01 -2.88052708e-01 -4.48535025e-01
-1.38430679e+00 5.93465507e-01 4.53617312e-02 2.83329606e-01
-3.35273445e-01 -5.90922475e-01 5.56345940e-01 1.46118641e-01
3.75744104e-01 -2.20860288e-01 -5.02136171e-01 2.69443214e-01
-7.24094331e-01 -1.81634575e-01 4.53801662e-01 9.71186459e-01
1.03334749e+00 -1.49878254e-02 5.23895681e-01 8.39148104e-01
5.98785758e-01 7.42939174e-01 4.24826622e-01 -1.09306586e+00
7.22022891e-01 1.06140661e+00 1.79333314e-01 -1.10639071e+00
-2.63489932e-01 -6.78798780e-02 -1.27221501e+00 9.80750248e-02
1.10776976e-01 5.48610240e-02 -7.38358676e-01 1.52809000e+00
5.35581410e-01 1.43339336e-01 4.11306977e-01 8.05910289e-01
8.65406036e-01 6.77348435e-01 -2.57638127e-01 -8.77768397e-01
1.23668182e+00 -7.79573619e-01 -1.12754059e+00 9.70286801e-02
2.80716479e-01 -8.07173550e-01 7.75221109e-01 6.49203837e-01
-8.69037390e-01 -7.82733321e-01 -1.13275254e+00 8.39670822e-02
-3.78664732e-01 2.66947925e-01 2.92870432e-01 2.92876899e-01
-3.89890224e-01 1.09127708e-01 -6.34445786e-01 -6.38520867e-02
1.09768130e-01 4.82168406e-01 -6.22994065e-01 -4.73472863e-01
-1.16626501e+00 8.37685823e-01 7.32404530e-01 3.21779341e-01
-4.61619794e-01 -4.31770861e-01 -6.50674105e-01 -1.43983573e-01
8.24532151e-01 -5.43046355e-01 6.79547846e-01 -9.33694482e-01
-9.87848699e-01 5.15815914e-01 -4.56200987e-01 3.89124155e-01
3.11790913e-01 1.15721330e-01 -7.93866932e-01 9.19303596e-02
1.01647489e-01 7.13432357e-02 6.43798292e-01 -1.98545241e+00
-7.08321333e-01 -9.15582299e-01 -6.73809722e-02 3.99210811e-01
-4.25464928e-01 -3.81848961e-01 -4.48300272e-01 -4.31982249e-01
5.95692813e-01 -9.74215806e-01 1.47885621e-01 -4.00219113e-01
-1.07154496e-01 -9.80577767e-02 1.05132723e+00 -8.10063362e-01
1.21087146e+00 -1.87602055e+00 4.90328461e-01 3.32743257e-01
1.60816431e-01 4.04615730e-01 3.17791939e-01 2.51622349e-01
-1.46033600e-01 2.37468585e-01 -2.60169625e-01 -9.82354879e-02
-4.62089151e-01 5.30175209e-01 5.05109839e-02 1.36047781e-01
-3.42977077e-01 3.91050488e-01 -7.59755492e-01 -9.22381699e-01
3.73593301e-01 2.34536171e-01 -1.86795965e-01 1.78945333e-01
1.66339576e-01 6.59728825e-01 -5.27000189e-01 7.81181812e-01
1.03730786e+00 -2.89292693e-01 4.86200809e-01 -7.17172563e-01
-6.52897879e-02 -5.99929094e-01 -1.84680128e+00 1.59693503e+00
-4.92755771e-01 -3.99480492e-01 1.66294530e-01 -1.05562294e+00
8.85491014e-01 6.22127593e-01 5.91662586e-01 -4.41115290e-01
3.65794487e-02 4.37300682e-01 -2.67135084e-01 -8.62287581e-01
9.91813242e-02 -4.67884213e-01 1.08384900e-01 5.55935025e-01
-7.30234608e-02 2.74559319e-01 2.28219070e-02 1.07656643e-02
1.20169178e-01 2.07873106e-01 4.63004380e-01 -4.77929674e-02
1.02016544e+00 -3.39025706e-02 1.36135137e+00 -2.51409393e-02
4.93095070e-02 3.48483741e-01 1.53408751e-01 -5.61803877e-01
-9.42019045e-01 -7.95011699e-01 7.57821649e-02 5.54626048e-01
5.56242168e-01 -2.08536446e-01 -5.37162900e-01 -7.73170352e-01
-2.32124701e-01 4.96008515e-01 -4.47654247e-01 -8.63158610e-03
-4.81577277e-01 -7.27845967e-01 4.14106809e-02 6.08485520e-01
7.93989062e-01 -8.04154575e-01 4.07417528e-02 1.02942646e-01
-7.99703479e-01 -9.69535947e-01 -4.04028237e-01 -2.68853396e-01
-1.13753736e+00 -1.32636774e+00 -2.66383111e-01 -7.00307906e-01
9.31383073e-01 7.02381611e-01 7.34757423e-01 2.19212815e-01
5.00882149e-01 1.53378159e-01 -4.81797487e-01 -2.29835555e-01
-5.15151739e-01 -2.58197814e-01 4.35317457e-01 3.87597054e-01
4.82955337e-01 -5.61230958e-01 -2.11447105e-01 5.79572320e-01
-1.17338884e+00 3.75640690e-01 4.94048983e-01 1.17770958e+00
1.09856391e+00 4.88119811e-01 9.79578853e-01 -1.10516477e+00
1.80889353e-01 -4.58095789e-01 -6.18003607e-01 9.53598797e-01
-1.16716611e+00 -1.30761191e-01 9.91704285e-01 -2.45437279e-01
-1.56116712e+00 6.38642535e-02 2.89817601e-01 -7.46742964e-01
-6.90128356e-02 7.62533844e-01 -8.92882943e-01 7.68744573e-02
9.97320339e-02 3.48315418e-01 2.68944949e-01 -4.46485519e-01
2.67906189e-01 9.57632601e-01 3.42129052e-01 -4.06886607e-01
6.53062761e-01 6.12785935e-01 1.72440693e-01 -3.13460141e-01
-7.79773116e-01 -3.67788702e-01 -9.35409606e-01 -4.13086474e-01
8.91504586e-01 -1.20973742e+00 -7.79462397e-01 5.52751780e-01
-9.23860133e-01 6.50595725e-01 2.03391239e-01 6.54045045e-01
-3.17463905e-01 6.29293919e-01 -4.39073741e-01 -7.52306759e-01
-3.52861822e-01 -1.16833675e+00 5.30416012e-01 3.90661180e-01
3.37236732e-01 -9.78360415e-01 -2.48608813e-01 9.46358442e-01
-1.00838907e-01 1.96954519e-01 1.09826028e+00 -5.92335045e-01
-5.30108392e-01 7.01121837e-02 -8.93795341e-02 5.81148386e-01
5.15230775e-01 2.65843794e-02 -6.88800573e-01 -2.34688237e-01
5.98103225e-01 -4.00439531e-01 4.33753312e-01 1.64997160e-01
9.66047168e-01 -2.44488031e-01 -1.84997901e-01 3.29759657e-01
1.71038473e+00 5.60402453e-01 3.79813224e-01 2.56857008e-01
1.04083109e+00 7.31229722e-01 1.07304227e+00 4.14510936e-01
8.41989577e-01 3.62248331e-01 4.33974624e-01 -4.50779535e-02
4.56312686e-01 -2.08023950e-01 3.24899866e-03 1.61510170e+00
-3.84790212e-01 -4.92812283e-02 -7.15391636e-01 3.02216589e-01
-2.35366488e+00 -1.34974802e+00 -3.45328838e-01 2.24111772e+00
5.24189711e-01 -2.86783297e-02 -8.70024636e-02 3.65293831e-01
1.12506771e+00 -1.57226652e-01 -7.41927683e-01 -4.82990593e-02
-1.70236290e-01 -5.99234164e-01 4.37983833e-02 2.82945216e-01
-7.93081224e-01 3.58609259e-01 4.75759220e+00 6.90415621e-01
-8.63518417e-01 3.30385447e-01 5.75560331e-01 1.05457112e-01
-7.03478158e-01 1.67314216e-01 -7.23328292e-01 6.55915499e-01
3.43091220e-01 6.39990047e-02 6.44112468e-01 6.22410774e-01
4.04380590e-01 -1.47463143e-01 -8.62640202e-01 1.21414852e+00
3.02976161e-01 -1.14032245e+00 1.64010331e-01 9.93908942e-02
9.06034112e-01 -3.93073380e-01 -3.27418178e-01 6.96635395e-02
-4.45500873e-02 -4.54508483e-01 5.02000809e-01 8.36989164e-01
6.80324852e-01 -1.06294203e+00 9.93883729e-01 9.00840044e-01
-1.28893220e+00 -2.87923723e-01 -3.12017143e-01 1.35252506e-01
2.08853364e-01 7.12576032e-01 2.11955351e-03 1.38785696e+00
5.94083846e-01 8.98394048e-01 -4.22083318e-01 4.13700521e-01
5.74400648e-02 2.21675113e-01 -3.95310819e-02 6.42015517e-01
-2.77916074e-01 -7.39581227e-01 1.52042121e-01 3.06848377e-01
3.00785750e-01 5.45662999e-01 6.00385249e-01 4.96330827e-01
1.38811573e-01 1.01842791e-01 -6.48383021e-01 3.59810323e-01
7.58766770e-01 1.40779889e+00 -5.34887850e-01 -5.67073584e-01
-7.62985051e-01 5.38160205e-01 1.59860313e-01 2.53467917e-01
-7.92508900e-01 3.65498662e-02 -5.69955222e-02 -1.21543415e-01
8.23718011e-02 5.85611723e-02 -5.53484440e-01 -1.57822418e+00
2.91436255e-01 -1.10040140e+00 5.65445304e-01 -8.73356700e-01
-1.47213519e+00 3.97090793e-01 9.07464251e-02 -1.69299507e+00
-2.04319507e-01 -6.42270967e-02 -2.25486323e-01 8.79022241e-01
-1.37632501e+00 -1.59169030e+00 -4.73096400e-01 8.80230904e-01
3.87819141e-01 -1.97849512e-01 8.43828559e-01 5.28502643e-01
-6.23101711e-01 2.21645936e-01 3.34084958e-01 -1.03757098e-01
9.17713106e-01 -9.16323960e-01 -5.65413058e-01 7.91490197e-01
-2.53673404e-01 7.31869698e-01 1.34070069e-01 -8.37522566e-01
-1.38831174e+00 -9.72418725e-01 7.05824077e-01 -1.88190356e-01
1.44166946e-01 1.08476430e-02 -1.07457507e+00 8.09617221e-01
1.18371606e-01 2.21842099e-02 1.07929313e+00 9.73804444e-02
-2.66842097e-01 -6.93956792e-01 -1.30855107e+00 2.48652786e-01
5.81292093e-01 -3.16470474e-01 -8.34940434e-01 1.19060807e-01
3.68452758e-01 -2.79650658e-01 -1.27428019e+00 7.22173095e-01
5.08920312e-01 -1.16228163e+00 8.29562664e-01 -2.42952749e-01
3.64217222e-01 -9.36216235e-01 -3.78495216e-01 -1.27910292e+00
-5.88100851e-01 1.16774313e-01 -1.08305044e-01 1.82054484e+00
1.53587326e-01 -7.13273466e-01 4.62363034e-01 6.67399704e-01
-1.02393828e-01 -6.85253501e-01 -7.67954469e-01 -5.55217266e-01
-3.54332775e-01 1.97866425e-01 7.33619392e-01 1.35245275e+00
-1.50830492e-01 3.86456639e-01 -6.91508293e-01 2.71904022e-01
8.34530532e-01 7.27072716e-01 4.88348305e-01 -1.53751671e+00
-9.74002294e-03 2.79573917e-01 4.02014777e-02 -3.22911561e-01
1.33660197e-01 -6.81683004e-01 -2.42291868e-01 -1.70821381e+00
5.10581374e-01 -5.81533909e-01 -6.00119829e-01 5.04354835e-01
-3.10431212e-01 6.09521708e-03 1.70443743e-01 7.87599742e-01
-6.22822642e-01 4.36997592e-01 1.43933380e+00 -5.37816398e-02
-2.17565089e-01 -8.05582479e-02 -8.28453720e-01 9.57812667e-01
5.55536151e-01 -3.73847038e-01 -7.61913538e-01 -5.02008319e-01
1.54002011e-01 5.73149085e-01 3.49297486e-02 -7.37431288e-01
2.84890562e-01 -5.72313309e-01 5.51654637e-01 -9.83982921e-01
2.43896395e-01 -1.32596552e+00 6.40329182e-01 3.29806298e-01
1.90221205e-01 3.04819584e-01 -2.44377613e-01 7.06021011e-01
-4.74645048e-01 -1.23453394e-01 7.28156149e-01 -5.18199623e-01
-8.55642259e-01 1.78395554e-01 1.12943053e-01 -7.76080834e-03
1.14997172e+00 -2.98105687e-01 -3.32576275e-01 -7.90043622e-02
-8.58736575e-01 5.87660015e-01 6.65427923e-01 4.71031576e-01
5.67323148e-01 -1.66531599e+00 -5.51835954e-01 1.96518883e-01
2.25918934e-01 2.71759808e-01 8.99371088e-01 8.96516740e-01
-1.24844439e-01 1.63714528e-01 -3.67416501e-01 -6.35071337e-01
-1.32641613e+00 9.53539014e-01 -1.33206561e-01 -3.06252211e-01
-2.20944330e-01 1.31754149e-02 -8.31802338e-02 -6.50665045e-01
-3.07924598e-01 2.08034709e-01 -5.34331083e-01 4.95564371e-01
3.86451036e-01 6.65727913e-01 1.18568420e-01 -8.77257109e-01
-3.55366856e-01 9.95952129e-01 -7.12992996e-02 1.03243060e-01
1.27250266e+00 -6.35342598e-01 -6.35491252e-01 8.43040586e-01
9.05614078e-01 -9.76263434e-02 -8.86951745e-01 -2.91756362e-01
-3.73368442e-01 -4.38941300e-01 -1.15239486e-01 -8.21459353e-01
-1.15252340e+00 6.81412399e-01 5.28385520e-01 5.77039681e-02
1.45801580e+00 -4.51791108e-01 8.06012332e-01 2.52154976e-01
5.80910325e-01 -1.30665708e+00 -1.29381344e-01 1.36022106e-01
5.18720269e-01 -1.50977993e+00 2.76318699e-01 -6.36338949e-01
-8.86289418e-01 1.05146956e+00 9.07395542e-01 3.38607430e-01
7.19267130e-01 -5.40138893e-02 3.13897938e-01 -8.46650451e-03
-6.38825178e-01 2.69341916e-01 7.51520321e-02 4.80239481e-01
-8.41471776e-02 5.57929017e-02 -3.78200859e-01 9.17663217e-01
4.76077497e-01 3.86157960e-01 6.43860042e-01 8.15045714e-01
-2.13460758e-01 -1.28365743e+00 -7.27093875e-01 4.69730824e-01
-3.28773528e-01 2.02448770e-01 -3.55561078e-02 5.66292405e-01
6.92980409e-01 1.37356293e+00 -4.14199650e-01 -6.94937408e-01
3.28504086e-01 1.80084035e-01 2.75355011e-01 -4.65538234e-01
-2.85396367e-01 3.40744048e-01 -3.35803449e-01 -1.56391263e-01
-1.05909979e+00 -6.67887032e-01 -1.42335415e+00 -3.08830261e-01
-7.94103980e-01 1.84980124e-01 4.89118040e-01 1.24485517e+00
4.44635808e-01 3.38086754e-01 8.84002447e-01 -4.42253679e-01
-4.19026673e-01 -6.40027881e-01 -6.73809886e-01 6.40671730e-01
-1.17748260e-01 -7.83760667e-01 -2.08423510e-01 4.00877416e-01] | [8.294673919677734, 4.626969814300537] |
1015d850-f0bc-4b52-acb8-469fd82ca056 | explaining-knowledge-graph-embedding-via | null | null | https://openreview.net/forum?id=RCyHECZIUFb | https://openreview.net/pdf?id=RCyHECZIUFb | Explaining Knowledge Graph Embedding via Latent Rule Learning | Knowledge Graph Embeddings (KGEs) embed entities and relations into continuous vector space following certain assumption, and are a powerful tools for representation learning of knowledge graphs. However, following vector space assumptions makes KGE a one step reasoner that directly predict final results without reasonable multi-hop reasoning steps. Thus KGEs are black-box models and explaining predictions made by KGEs remains unsolved. In this paper, we propose KGExplainer, the first general approach of providing explanations for predictions from KGE models. KGExplainer is a multi-hop reasoner learning latent rules for link prediction and is encouraged to behave similarly to KGEs during prediction through knowledge distillation. For explanation, KGExplainer outputs a ranked list of rules for each relation. Experiments on benchmark datasets with two target KGEs show that our approach is faithfulness to replicate KGEs behaviors for link prediction and is good at outputting quality rules for effective explanations. | ['Huajun Chen', 'Yushan Zhu', 'Zezhong Xu', 'Mingyang Chen', 'Wen Zhang'] | 2021-09-29 | null | null | null | null | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [-2.37040028e-01 1.37518156e+00 -8.86188745e-01 -3.76176715e-01
7.91819990e-02 -2.27793753e-01 5.74059188e-01 4.06388074e-01
4.39159602e-01 7.12151647e-01 5.18569469e-01 -8.51733923e-01
-6.13987744e-01 -1.29379880e+00 -9.20948625e-01 -3.69399823e-02
-2.28093550e-01 8.74586523e-01 1.18183233e-01 -3.25849205e-01
-1.09677045e-02 3.97996157e-01 -1.16186249e+00 4.95097041e-01
7.89098620e-01 3.84619176e-01 -5.10798335e-01 9.51990366e-01
-6.88096344e-01 1.53259313e+00 3.86495888e-02 -1.05237460e+00
-9.75670367e-02 -2.06129149e-01 -1.14465475e+00 -3.18518400e-01
2.74700642e-01 -7.47375265e-02 -8.02522421e-01 6.74154818e-01
-3.11950058e-01 1.18060283e-01 1.19812953e+00 -1.90234053e+00
-1.56469667e+00 1.56953132e+00 2.49932796e-01 -1.93042576e-01
4.70045388e-01 -1.59312218e-01 1.56243861e+00 -1.21312439e+00
7.61951208e-01 1.31905007e+00 7.82948136e-01 8.04652035e-01
-1.32422674e+00 -4.65720683e-01 3.00275177e-01 5.84313929e-01
-1.26121318e+00 -1.06826439e-01 6.73146009e-01 -4.75558221e-01
1.50008810e+00 3.21422666e-01 7.34398544e-01 1.01778090e+00
2.60280758e-01 8.13521087e-01 4.34752703e-01 -3.00064743e-01
3.18947405e-01 6.22372627e-01 8.98308158e-01 1.01506567e+00
8.08813453e-01 2.34033614e-01 -9.49773848e-01 -4.03038830e-01
5.65504849e-01 2.67794549e-01 -4.13079917e-01 -9.20021236e-01
-1.00224626e+00 1.15694380e+00 7.97520041e-01 -1.00076169e-01
-4.44577754e-01 4.24187154e-01 1.00064449e-01 7.06007481e-01
1.53856561e-01 6.78709686e-01 -8.03725898e-01 1.56687051e-01
-2.39352435e-01 2.77402818e-01 1.13183177e+00 1.11479354e+00
8.28101218e-01 3.25712323e-01 -1.74962670e-01 3.41523618e-01
6.72828317e-01 1.92522898e-01 3.16206694e-01 -5.08366048e-01
3.26069027e-01 1.17857409e+00 2.48410292e-02 -1.36220908e+00
-2.36568645e-01 -3.07238698e-01 -7.20984161e-01 7.87605941e-02
-6.07447214e-02 1.35669246e-01 -8.02212358e-01 1.24275577e+00
2.20632523e-01 1.87529057e-01 7.82250345e-01 5.92902243e-01
1.06350327e+00 8.35929811e-01 1.71086803e-01 -1.39526263e-01
8.02370727e-01 -1.31743777e+00 -6.88958526e-01 -2.24576667e-01
1.02584851e+00 4.48754691e-02 9.43511486e-01 3.92683566e-01
-6.85232997e-01 -5.37107050e-01 -9.38104093e-01 2.16381941e-02
-8.22607398e-01 -2.38074616e-01 1.35314250e+00 2.90979296e-01
-9.53915417e-01 9.29877162e-01 -7.21728921e-01 -5.22165895e-01
2.70927876e-01 4.27063674e-01 -7.16607273e-01 -8.93903598e-02
-1.34774911e+00 9.66253042e-01 9.76935446e-01 -2.09926903e-01
-5.99366546e-01 -9.63478386e-01 -1.22852445e+00 4.13921505e-01
5.54608822e-01 -7.93432176e-01 7.72860408e-01 -1.05864383e-01
-1.17826676e+00 2.73062974e-01 4.00964767e-02 -7.34171927e-01
1.44170389e-01 -2.54831195e-01 -8.72424960e-01 -2.94107109e-01
-1.74263999e-01 5.45571208e-01 5.39090276e-01 -1.36109519e+00
-2.56100714e-01 -3.24537382e-02 4.49943244e-02 -8.14498812e-02
-4.99412745e-01 -7.50655830e-01 -1.95003048e-01 -2.01789454e-01
1.12974115e-01 -7.30715513e-01 -3.49122763e-01 -4.86152098e-02
-8.65675509e-01 -6.68893158e-01 6.35955155e-01 -4.09589916e-01
1.44583654e+00 -1.73362756e+00 9.38754976e-02 5.49244881e-01
8.03641379e-01 1.96142301e-01 -7.24350587e-02 7.24210382e-01
-4.70685065e-01 4.47206914e-01 2.87340522e-01 1.66767344e-01
4.13310260e-01 6.58668697e-01 -7.71319509e-01 -8.65166783e-02
1.21457530e-02 1.56092167e+00 -1.13419700e+00 -2.54058748e-01
1.92958891e-01 3.11443955e-01 -8.25266123e-01 3.33427072e-01
-5.46594858e-01 -3.92244995e-01 -4.71611619e-01 4.72187191e-01
2.44442612e-01 -7.42942750e-01 5.50511181e-01 -2.80374736e-01
5.04896581e-01 1.33246809e-01 -1.12153149e+00 1.09112322e+00
-1.09103300e-01 4.60916936e-01 -9.10636246e-01 -8.64899814e-01
1.26735747e+00 2.70077407e-01 -3.17964047e-01 1.16922345e-03
-1.54622436e-01 7.58719742e-02 -1.73734769e-01 -3.73273283e-01
5.76251268e-01 2.45736558e-02 6.49700239e-02 2.83802450e-01
1.24568120e-01 -2.09604483e-02 -1.10689171e-01 9.23234463e-01
1.21729159e+00 -2.64347941e-02 7.64225900e-01 1.60947233e-01
1.81611344e-01 3.78685951e-01 5.58772206e-01 9.75251675e-01
2.15062231e-01 -7.59332404e-02 6.22640610e-01 -9.89167213e-01
-8.25873315e-01 -1.24763739e+00 4.76499081e-01 7.15643346e-01
1.11956954e-01 -1.16924429e+00 1.24524318e-01 -1.22881567e+00
4.71882135e-01 1.47381079e+00 -8.94639790e-01 -5.90990067e-01
-1.20701203e-02 2.59047262e-02 3.98942053e-01 8.97722185e-01
5.98852057e-03 -9.62154865e-01 8.85060877e-02 2.68254697e-01
3.15434009e-01 -9.60652113e-01 -9.56512168e-02 3.73807907e-01
-8.07611704e-01 -1.42538595e+00 1.72421157e-01 -4.88434643e-01
8.17390800e-01 1.26743183e-01 1.30862153e+00 4.24035668e-01
-1.23838134e-01 6.94592059e-01 -5.67018926e-01 -1.73206642e-01
-6.73066795e-01 -2.56763905e-01 3.84440929e-01 -3.55855972e-01
6.62174404e-01 -6.59467340e-01 -2.62504309e-01 3.94338667e-02
-5.72573304e-01 3.99970204e-01 5.08124888e-01 7.57695854e-01
5.66467226e-01 2.70678580e-01 3.22001785e-01 -1.56978154e+00
9.47955072e-01 -6.84817791e-01 -3.07840586e-01 7.69411981e-01
-1.55767822e+00 5.92268884e-01 8.76640618e-01 -2.94642538e-01
-6.14988327e-01 -2.70246565e-01 1.79513156e-01 -7.44177401e-01
3.30702402e-02 7.72765160e-01 1.26706272e-01 2.71022823e-02
9.54305351e-01 3.08518201e-01 -3.19728702e-01 -2.53079772e-01
1.07621217e+00 2.58525759e-01 5.54360628e-01 -4.74280387e-01
1.23118985e+00 -1.48064762e-01 -3.54788043e-02 -2.33052477e-01
-9.01397645e-01 -6.24091215e-02 -2.71320850e-01 3.94855887e-02
7.30469644e-01 -5.92106879e-01 -9.61456120e-01 -5.88437855e-01
-1.12755203e+00 -2.70694703e-01 -6.98864520e-01 4.87245500e-01
-4.35770154e-01 4.36053239e-02 -4.95964378e-01 -6.04993999e-01
-3.53407800e-01 -4.08581197e-01 2.74143904e-01 1.66946411e-01
-6.15940809e-01 -1.41777611e+00 1.71746656e-01 3.22194010e-01
2.36763239e-01 7.73020834e-02 1.54336214e+00 -1.21451259e+00
-6.96167171e-01 -1.87063724e-01 -1.50455281e-01 7.48406425e-02
2.57299207e-02 1.06572784e-01 -6.42497361e-01 1.69378191e-01
-9.18826938e-01 -3.99322480e-01 7.69951642e-01 -1.55570060e-01
9.80047405e-01 -8.23744714e-01 -7.75273800e-01 4.24315125e-01
1.37861943e+00 -6.89852685e-02 4.78130639e-01 9.08534899e-02
1.03728914e+00 2.38732025e-01 5.57989478e-01 2.89840490e-01
8.12300444e-01 4.13364828e-01 5.19243360e-01 2.91638851e-01
-2.39954010e-01 -1.14796710e+00 3.69001210e-01 8.55927169e-01
-8.00649822e-02 -3.49176049e-01 -1.21937215e+00 6.24957740e-01
-2.26923108e+00 -9.83681560e-01 -4.10918802e-01 1.65834475e+00
6.15734339e-01 2.06693605e-01 -2.84307063e-01 1.54529527e-01
4.15742368e-01 -1.08156681e-01 -5.87875009e-01 -8.47634137e-01
1.56542957e-01 -6.87481463e-02 3.58269483e-01 8.56527388e-01
-5.05203366e-01 1.45831478e+00 6.42784166e+00 4.51344281e-01
-4.58757490e-01 -2.24701166e-01 -4.33344729e-02 5.20043552e-01
-9.86899912e-01 3.01571906e-01 -7.44495809e-01 -6.15667067e-02
9.33167756e-01 -6.08805180e-01 5.52844763e-01 1.28869426e+00
-4.09805447e-01 6.51925981e-01 -1.54331899e+00 9.46107626e-01
-2.79315978e-01 -2.02444029e+00 6.95356786e-01 -2.08531588e-01
7.91102707e-01 -3.45466733e-01 -8.33577216e-02 9.36791241e-01
1.23279870e+00 -1.26870453e+00 2.46172279e-01 7.45546281e-01
3.90293270e-01 -6.98425770e-01 6.84386492e-01 3.02153498e-01
-1.12706959e+00 -1.40645266e-01 -7.73858845e-01 -2.55511492e-01
-5.52457087e-02 5.72707415e-01 -1.59818304e+00 8.97822380e-01
2.57776320e-01 8.84560943e-01 -6.40762627e-01 5.16130030e-01
-8.96788359e-01 8.70728195e-01 2.10092038e-01 -1.40437707e-01
1.44504875e-01 1.00329414e-01 2.99706340e-01 1.09254205e+00
3.41124415e-01 2.17568368e-01 2.26539478e-01 9.42401886e-01
-1.27501324e-01 6.91784248e-02 -1.10129392e+00 -5.94904423e-01
5.31714022e-01 9.73507941e-01 -4.70556617e-01 -6.33492172e-01
-4.10863161e-01 9.76205885e-01 9.47106481e-01 5.23423970e-01
-7.02038109e-01 -2.07640976e-01 7.17611074e-01 3.65029015e-02
3.72151405e-01 -1.56307574e-02 -1.40370890e-01 -1.31852031e+00
-4.37940180e-01 -6.67271972e-01 7.92692840e-01 -1.22816956e+00
-1.30585563e+00 6.87851369e-01 -1.27972648e-01 -8.30648839e-01
-3.59095395e-01 -7.80653894e-01 -7.83260167e-01 3.17440599e-01
-1.61804199e+00 -1.11435962e+00 -2.68933266e-01 8.52742136e-01
3.90418798e-01 -4.38871056e-01 1.28830647e+00 -3.67076546e-01
-3.03980201e-01 5.71037591e-01 -3.05134296e-01 9.17037874e-02
1.17067344e-01 -1.54900289e+00 5.56854546e-01 4.87484485e-01
7.90093064e-01 1.04609203e+00 1.07337213e+00 -1.00798309e+00
-1.59590495e+00 -1.44221556e+00 1.27387786e+00 -5.99686801e-01
9.68550146e-01 -2.98972533e-04 -1.16211605e+00 1.33996379e+00
1.04546711e-01 2.58152455e-01 1.00781941e+00 8.15620005e-01
-8.07458222e-01 1.85934022e-01 -7.73865163e-01 8.40053082e-01
1.26558697e+00 -6.10500574e-01 -1.11149752e+00 5.98693013e-01
1.11043453e+00 -5.80760650e-02 -1.05539834e+00 3.48427296e-01
1.59375921e-01 -8.24512243e-01 9.42476749e-01 -1.41290128e+00
6.44552588e-01 -4.12938625e-01 -1.33659959e-01 -1.59666467e+00
-6.32120788e-01 -4.59598631e-01 -1.12756300e+00 1.05414617e+00
8.47126126e-01 -6.39580250e-01 1.18595147e+00 8.69178951e-01
3.70630696e-02 -8.86335790e-01 -1.86685711e-01 -9.43794310e-01
-2.71384120e-01 -7.05434978e-01 8.14086258e-01 1.32751298e+00
5.49768507e-01 5.27291954e-01 -5.82229197e-01 5.75765908e-01
7.60260224e-01 3.80228043e-01 1.03907156e+00 -1.45625126e+00
-4.79765087e-01 -1.08972877e-01 -7.79296219e-01 -8.18543434e-01
5.55298388e-01 -1.57874835e+00 -4.73651439e-01 -2.21542907e+00
1.21571667e-01 -2.24769607e-01 -3.18570822e-01 1.06814396e+00
-2.97569275e-01 -2.49620542e-01 7.31876343e-02 1.21495470e-01
-5.93551755e-01 7.09982693e-01 9.74999607e-01 -4.28722769e-01
-1.34525448e-01 -2.26264313e-01 -1.02772999e+00 7.27979064e-01
5.22166073e-01 -6.48352087e-01 -1.15593481e+00 -1.97824072e-02
8.70048165e-01 2.34986231e-01 4.85491127e-01 -6.77018106e-01
6.35065198e-01 -3.74744028e-01 2.54669636e-01 -3.81886154e-01
6.40842319e-02 -8.70610714e-01 5.35059988e-01 3.74769449e-01
-7.26340115e-01 -2.21817359e-01 -5.04048914e-02 9.28223550e-01
-2.11150810e-01 -6.82013631e-02 -3.34856138e-02 2.21265823e-01
-1.31902325e+00 4.82758373e-01 -7.76118040e-02 -1.94406703e-01
9.59139705e-01 -3.14031065e-01 -6.00902259e-01 -5.68994105e-01
-1.33977938e+00 4.32427138e-01 4.70039696e-02 5.81144214e-01
1.17506731e+00 -1.75997281e+00 -4.18601900e-01 1.02726042e-01
6.30836070e-01 -4.07541394e-01 -1.14502423e-02 2.66292751e-01
-2.26664245e-01 6.25155568e-01 1.10032819e-01 9.60206613e-03
-1.15358138e+00 1.01157928e+00 7.12337568e-02 -4.68547076e-01
-1.07629406e+00 1.14388299e+00 1.46330735e-02 -1.02533245e+00
1.78113550e-01 -4.05903846e-01 -4.41931546e-01 -3.72321486e-01
3.64676952e-01 2.23320484e-01 -3.20043862e-01 8.76285285e-02
-3.19733679e-01 2.39046425e-01 -2.90031701e-01 5.09271085e-01
1.47211266e+00 2.34173238e-01 1.26792207e-01 4.67133611e-01
8.77520800e-01 -1.02713659e-01 -7.72494018e-01 -3.47526520e-01
2.47417644e-01 -4.44944024e-01 -7.52462149e-02 -8.91068041e-01
-7.63318539e-01 6.32604003e-01 -2.37186298e-01 3.83135408e-01
2.84254670e-01 4.08910036e-01 2.77292311e-01 9.81332660e-01
3.14493656e-01 -5.88505387e-01 3.31229270e-02 4.55594927e-01
9.93232906e-01 -1.08899772e+00 2.42508128e-02 -7.34418929e-01
-9.54589546e-01 1.27956998e+00 9.54814196e-01 9.36787501e-02
6.88154519e-01 -5.61051108e-02 -1.31997289e-02 -7.28381753e-01
-1.39336419e+00 5.56087792e-02 5.09540617e-01 7.70761847e-01
1.92741558e-01 4.45673972e-01 1.43656328e-01 1.04371178e+00
-4.29793686e-01 1.27838016e-01 4.97981936e-01 4.57138300e-01
-5.07928193e-01 -1.08572638e+00 4.17034179e-02 6.18478537e-01
4.41240489e-01 -2.57279128e-01 -5.64735115e-01 9.58805203e-01
-2.62750566e-01 8.75981331e-01 -5.01014411e-01 -9.82327580e-01
3.08619708e-01 4.64349240e-01 3.12982164e-02 -7.84048617e-01
-7.47561604e-02 -1.03657305e+00 2.81953752e-01 -8.33098233e-01
2.79924333e-01 5.79772070e-02 -1.82719243e+00 -5.50772965e-01
-3.78743917e-01 8.72438610e-01 1.51441470e-01 8.81214738e-01
5.34805119e-01 7.46755123e-01 3.36542010e-01 6.47820607e-02
-3.86201471e-01 -3.40695977e-01 -9.02017295e-01 4.67129081e-01
-1.53802456e-02 -7.22432256e-01 -3.42397481e-01 -7.33837113e-02] | [8.870599746704102, 7.76581335067749] |
2100ba09-4a18-4d13-8da6-b9f0b0ac84e9 | how-to-prevent-the-poor-performance-clients | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Qu_How_To_Prevent_the_Poor_Performance_Clients_for_Personalized_Federated_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Qu_How_To_Prevent_the_Poor_Performance_Clients_for_Personalized_Federated_CVPR_2023_paper.pdf | How To Prevent the Poor Performance Clients for Personalized Federated Learning? | Personalized federated learning (pFL) collaboratively trains personalized models, which provides a customized model solution for individual clients in the presence of heterogeneous distributed local data. Although many recent studies have applied various algorithms to enhance personalization in pFL, they mainly focus on improving the performance from averaging or top perspective. However, part of the clients may fall into poor performance and are not clearly discussed. Therefore, how to prevent these poor clients should be considered critically. Intuitively, these poor clients may come from biased universal information shared with others. To address this issue, we propose a novel pFL strategy, called Personalize Locally, Generalize Universally (PLGU). PLGU generalizes the fine-grained universal information and moderates its biased performance by designing a Layer-Wised Sharpness Aware Minimization (LWSAM) algorithm while keeping the personalization local. Specifically, we embed our proposed PLGU strategy into two pFL schemes concluded in this paper: with/without a global model, and present the training procedures in detail. Through in-depth study, we show that the proposed PLGU strategy achieves competitive generalization bounds on both considered pFL schemes. Our extensive experimental results show that all the proposed PLGU based-algorithms achieve state-of-the-art performance. | ['Lixing Chen', 'Chengchao Shen', 'Rui Duan', 'Xiao Han', 'Xingyu Li', 'Zhe Qu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['personalized-federated-learning', 'generalization-bounds'] | ['methodology', 'methodology'] | [-2.03076243e-01 -1.15600802e-01 -4.14445966e-01 -4.94743347e-01
-1.16360414e+00 -3.17439944e-01 2.04916075e-01 -1.46052778e-01
-6.32439852e-02 7.94019103e-01 3.47046256e-01 4.29326808e-03
-5.68186820e-01 -6.71640933e-01 -8.57667863e-01 -1.22178388e+00
4.09775376e-02 6.60892069e-01 1.02620780e-01 1.33736217e-02
1.75971359e-01 6.07345581e-01 -1.33706820e+00 6.07519209e-01
1.12717605e+00 1.05580020e+00 2.74392128e-01 4.35104430e-01
-3.61171693e-01 8.02074492e-01 -2.38118231e-01 -6.78268850e-01
4.21624064e-01 -1.33246496e-01 -8.11622918e-01 2.47777954e-01
4.13518608e-01 -3.81377131e-01 -2.63264533e-02 8.91141951e-01
5.98354161e-01 3.73033762e-01 4.08767164e-01 -1.18700421e+00
-8.08215857e-01 1.05076253e+00 -5.49193025e-01 2.25902542e-01
-1.86784238e-01 7.90253580e-02 1.14035094e+00 -8.25904012e-01
2.07650468e-01 1.23324287e+00 6.37982011e-01 5.52013159e-01
-1.08439827e+00 -5.48990190e-01 6.81032002e-01 4.98735338e-01
-1.38792336e+00 -3.95737171e-01 9.34894562e-01 1.06175601e-01
5.82415223e-01 5.55726111e-01 2.95174886e-02 8.47637296e-01
-7.16084912e-02 1.07976735e+00 1.01266527e+00 -8.69560465e-02
2.85847634e-01 4.95899320e-01 1.49839282e-01 5.77974081e-01
1.01651214e-01 -2.59009689e-01 -5.12266636e-01 -5.36404967e-01
4.65462804e-01 3.62976879e-01 -3.79295021e-01 -5.60804248e-01
-5.62740684e-01 4.69357133e-01 3.75905216e-01 2.45913174e-02
-6.13094807e-01 -5.78899011e-02 2.76673287e-01 2.55480230e-01
4.50440466e-01 -1.40027627e-01 -7.12066293e-01 1.41762540e-01
-7.82667696e-01 3.05348635e-01 8.09807301e-01 1.27268457e+00
1.28242743e+00 -2.43415356e-01 -3.53765190e-01 9.48627591e-01
1.17269658e-01 7.59889409e-02 4.59383011e-01 -1.08092391e+00
5.96478403e-01 6.60538137e-01 1.25126913e-01 -7.28752017e-01
-1.27980128e-01 -8.97690654e-01 -6.26122594e-01 -2.05328420e-01
-9.12545919e-02 -3.24370176e-01 -3.19924355e-01 1.76591313e+00
7.26122856e-01 4.47666377e-01 2.81690639e-02 8.97122920e-01
3.53438973e-01 4.86782998e-01 2.42172867e-01 -5.38213015e-01
7.42723048e-01 -1.43069029e+00 -2.93617517e-01 1.68289661e-01
4.79186207e-01 -7.57157385e-01 9.49711263e-01 4.01115209e-01
-1.16600394e+00 -1.16294377e-01 -5.11538327e-01 2.76844084e-01
-7.78780580e-02 -2.74855942e-01 6.09435797e-01 7.98340201e-01
-1.10745227e+00 7.48714685e-01 -6.40321672e-01 -6.02802098e-01
6.92049623e-01 3.00808370e-01 -1.62914753e-01 -2.11794242e-01
-7.90938675e-01 5.55017054e-01 5.77942692e-02 -2.19802156e-01
-9.48632360e-01 -1.04137349e+00 -1.20833129e-01 3.70152563e-01
5.16097724e-01 -9.38245296e-01 1.41947997e+00 -8.70917320e-01
-1.40313208e+00 3.94712806e-01 -2.65686482e-01 -3.08980107e-01
9.21932995e-01 -1.05025016e-01 -2.50134885e-01 -1.78776518e-01
-2.31267482e-01 -9.12285447e-02 5.79447389e-01 -1.72828364e+00
-1.08098412e+00 -6.48042858e-01 1.70919776e-01 4.20987964e-01
-8.93151879e-01 2.75085792e-02 -8.28364253e-01 -4.65596706e-01
2.97797727e-03 -5.75597107e-01 -3.64990532e-01 -1.56651348e-01
-2.00161144e-01 -4.44475740e-01 1.01104450e+00 -3.16722542e-01
1.39906895e+00 -1.89412379e+00 -1.13647878e-01 2.32674807e-01
1.12640336e-01 9.65128839e-02 -1.38235271e-01 7.37601638e-01
4.02345866e-01 3.75455208e-02 -5.56331798e-02 -9.12397385e-01
2.89186150e-01 6.02969170e-01 -1.16385490e-01 5.01491427e-01
-4.53333110e-01 4.86299723e-01 -7.60867178e-01 -7.50703990e-01
-8.58748481e-02 4.91412848e-01 -9.94548321e-01 4.34923440e-01
-3.47215444e-01 4.58980739e-01 -8.41938615e-01 8.04517388e-01
1.24049258e+00 -2.08140522e-01 2.44113237e-01 -4.50576901e-01
-1.46055639e-01 -4.71440963e-02 -1.21610165e+00 1.44000995e+00
-6.87592685e-01 -3.93349499e-01 5.14150143e-01 -9.72152174e-01
5.69413006e-01 1.96266577e-01 7.33743429e-01 -3.17159444e-01
3.02891191e-02 1.64476857e-01 -7.50465870e-01 -5.31621695e-01
4.70690876e-01 -7.38056377e-02 2.93502450e-01 6.08327687e-01
-8.47816467e-02 8.35820198e-01 -2.68941790e-01 2.72392362e-01
9.22198057e-01 1.09918620e-02 -1.98004946e-01 -3.91891479e-01
5.95339537e-01 -4.16531891e-01 8.96678090e-01 9.97789741e-01
-3.82093102e-01 4.92830485e-01 1.75034627e-01 -2.14936227e-01
-6.32480621e-01 -1.00633490e+00 6.27094880e-02 1.83062840e+00
4.69352931e-01 -2.11172655e-01 -7.65292704e-01 -9.11075592e-01
4.87810448e-02 6.65554941e-01 -5.19042253e-01 7.56076053e-02
-4.96597886e-01 -9.52380061e-01 2.59855956e-01 3.46426815e-01
6.12160504e-01 -7.14388549e-01 -7.66450018e-02 2.98943639e-01
-9.19132233e-02 -8.46878946e-01 -8.94153714e-01 5.86432889e-02
-8.38781953e-01 -8.10713708e-01 -6.24674022e-01 -6.27153873e-01
4.23014730e-01 6.45667195e-01 7.61020541e-01 -2.04713661e-02
1.70114949e-01 4.71693933e-01 -5.01435637e-01 2.84009092e-02
3.40303965e-02 3.25977713e-01 3.59905921e-02 6.88518524e-01
1.31402522e-01 -1.05816174e+00 -9.75587785e-01 5.46069324e-01
-8.05688024e-01 -1.02849878e-01 6.64978981e-01 6.23456836e-01
6.88768744e-01 8.38899687e-02 8.60903561e-01 -1.27484417e+00
5.42484462e-01 -9.02730882e-01 -4.48718406e-02 6.00026429e-01
-1.08470094e+00 4.73144790e-03 9.39827502e-01 -2.74061948e-01
-1.69405580e+00 -3.08059096e-01 -9.57245082e-02 -5.69586456e-01
-5.64062269e-04 1.51511490e-01 -6.37960255e-01 -3.21718633e-01
5.03525496e-01 2.49771416e-01 -3.45326275e-01 -1.06764424e+00
5.23882329e-01 9.58620965e-01 4.26537871e-01 -1.23043132e+00
4.72887069e-01 5.58783710e-01 -2.92850822e-01 -3.19409609e-01
-8.01970124e-01 -5.01584828e-01 -2.85997428e-03 -1.34109259e-01
1.51353687e-01 -6.65207028e-01 -8.54584932e-01 3.13202620e-01
-6.95336223e-01 -2.09158912e-01 -3.93769220e-02 8.15605372e-02
-5.89736998e-01 7.11060762e-01 -5.64357936e-01 -9.57642078e-01
-8.19889903e-01 -1.05739248e+00 6.96194589e-01 5.57546020e-01
4.38374281e-01 -1.00717747e+00 5.63412309e-02 5.31605422e-01
8.33858669e-01 -1.37360960e-01 8.12414289e-01 -8.43966663e-01
-7.10098326e-01 -1.54383257e-02 -3.73564601e-01 4.34322029e-01
2.36164540e-01 -3.51887733e-01 -1.10662866e+00 -4.18249696e-01
-4.69400026e-02 -3.44478711e-02 6.57742143e-01 -4.69255149e-02
1.48132932e+00 -7.84710824e-01 -3.95884722e-01 9.04598415e-01
1.72062314e+00 -8.86391923e-02 2.64207929e-01 2.53289282e-01
7.71330655e-01 4.59620148e-01 3.08972359e-01 1.12100375e+00
7.56136239e-01 5.54426074e-01 7.10516989e-01 2.91434884e-01
9.16724801e-02 -1.66137844e-01 2.83038706e-01 5.86007297e-01
-1.44561395e-01 -4.46438372e-01 -4.11224157e-01 6.41380548e-01
-2.29929852e+00 -7.44341612e-01 7.73739591e-02 2.10397482e+00
6.70012116e-01 -3.76501799e-01 1.31262302e-01 -3.01479191e-01
9.21639681e-01 -3.28635215e-03 -7.44003713e-01 -5.00726223e-01
-1.87045708e-01 -1.23520553e-01 7.71961451e-01 3.43684286e-01
-7.65343964e-01 8.30042422e-01 5.36020327e+00 1.09646785e+00
-9.23402727e-01 5.54962397e-01 6.51189566e-01 -3.20711344e-01
-6.62375629e-01 -4.94398437e-02 -8.29241872e-01 5.15003860e-01
5.95495760e-01 -5.75750113e-01 6.86326861e-01 1.15663564e+00
4.01575357e-01 3.63432854e-01 -1.07423627e+00 8.82237375e-01
-1.16774626e-02 -1.16083813e+00 3.19995463e-01 -1.58670217e-01
1.12820661e+00 1.15241103e-01 3.02454293e-01 4.36706841e-01
3.59362721e-01 -4.29799199e-01 4.72377032e-01 7.81890690e-01
2.43638307e-01 -1.09096014e+00 5.33750236e-01 5.48156977e-01
-1.04007947e+00 -4.15173948e-01 -4.49026376e-01 4.11204785e-01
1.38297854e-02 3.94294679e-01 -4.12703663e-01 1.04678595e+00
9.61416245e-01 2.56149113e-01 -3.66399229e-01 1.25216889e+00
4.96213615e-01 6.15023792e-01 -3.60135883e-01 2.39279911e-01
3.25940341e-01 -5.23667820e-02 5.34362555e-01 1.26937461e+00
3.26709270e-01 1.22785367e-01 2.84962982e-01 5.86192071e-01
-3.14835638e-01 5.86105168e-01 1.70284376e-01 4.40191537e-01
6.13417268e-01 1.35540402e+00 -1.02091851e-02 -3.72562855e-01
-4.60401982e-01 1.00390792e+00 6.60539925e-01 4.36794937e-01
-8.33832741e-01 1.53503502e-02 9.16004598e-01 1.34789079e-01
3.61988544e-01 3.69033456e-01 -3.45398098e-01 -1.15795541e+00
1.83258802e-02 -7.37484097e-01 8.54648232e-01 -3.17075133e-01
-1.80592501e+00 5.78782618e-01 -1.92138374e-01 -8.68346572e-01
2.92507410e-01 9.38141020e-04 -9.36776817e-01 8.99466455e-01
-1.57963216e+00 -1.37120044e+00 -4.66784596e-01 1.03626001e+00
6.13397121e-01 -4.34591016e-03 4.57760394e-01 8.03283870e-01
-8.56332958e-01 1.33870983e+00 5.20915985e-01 -7.59311974e-01
9.51916575e-01 -9.28132713e-01 -2.97205329e-01 5.91946542e-01
-5.82642138e-01 7.64969826e-01 8.28363538e-01 -2.70153522e-01
-1.52674997e+00 -1.48300135e+00 6.61132812e-01 -9.63133425e-02
4.75387752e-01 3.04693758e-01 -9.86256719e-01 8.30161929e-01
1.61361694e-01 2.30762288e-02 6.42852664e-01 8.65110680e-02
-3.20954114e-01 -5.98597407e-01 -1.69227362e+00 4.65008527e-01
1.36248636e+00 -1.06436007e-01 -1.42234191e-01 4.83442754e-01
6.85296297e-01 -1.15634128e-01 -8.65212679e-01 3.68417650e-01
5.15653074e-01 -1.29698443e+00 7.01039433e-01 -7.34009385e-01
-1.33797780e-01 -1.18259871e-02 -5.01509249e-01 -1.03616977e+00
-5.70528150e-01 -8.38488340e-01 -3.29364300e-01 1.72641516e+00
1.04627125e-01 -9.64812577e-01 1.14583480e+00 8.94034088e-01
-4.61302429e-01 -1.11935055e+00 -6.91263139e-01 -7.57311404e-01
1.18061215e-01 -3.57734591e-01 1.16609931e+00 8.17003727e-01
-2.56371573e-02 -1.66394159e-01 -6.24162912e-01 4.89912510e-01
1.03411174e+00 3.96095455e-01 7.85972953e-01 -8.15884888e-01
-5.72341204e-01 -5.75416505e-01 1.07935257e-01 -1.31190252e+00
-7.83251002e-02 -7.41097748e-01 -2.71466315e-01 -1.52821803e+00
6.67272806e-01 -8.57394516e-01 -7.58572519e-01 4.42034274e-01
-3.08996469e-01 -9.98676643e-02 9.36929435e-02 4.52836782e-01
-1.00659585e+00 6.00995541e-01 1.04650557e+00 1.42963424e-01
-2.14154765e-01 3.15424860e-01 -1.13670123e+00 3.62930894e-01
8.07467520e-01 -2.33468041e-01 -4.84631121e-01 -5.36879838e-01
-1.85306773e-01 -1.70641199e-01 4.04042676e-02 -7.25727558e-01
5.07286370e-01 -6.25920177e-01 -1.33387774e-01 -4.45460320e-01
1.11006228e-02 -8.76109779e-01 3.58228594e-01 -8.97238702e-02
-1.99162394e-01 -2.51239687e-01 -2.06860363e-01 6.67015374e-01
1.79978222e-01 5.39334677e-02 9.58529890e-01 -1.54299334e-01
-5.65141261e-01 8.84222984e-01 6.76338822e-02 -2.52422512e-01
9.44826424e-01 1.61640346e-01 -4.29061472e-01 -4.75838006e-01
-7.50201523e-01 6.78102672e-01 5.76767445e-01 9.99484435e-02
1.16344184e-01 -1.19188833e+00 -6.65694475e-01 -1.54973343e-01
2.23683659e-02 -2.24779949e-01 1.03452361e+00 9.85012650e-01
-1.80515036e-01 2.88721681e-01 1.96902752e-01 -3.10046911e-01
-1.02750742e+00 8.82895052e-01 4.33695585e-01 -3.09247017e-01
-4.80488688e-01 1.03157806e+00 4.26733643e-01 -4.79362041e-01
4.76629794e-01 1.86089963e-01 4.98203114e-02 -1.95330203e-01
4.18429524e-01 8.58345330e-01 1.68189913e-01 -5.59781313e-01
-3.57756525e-01 3.72613072e-01 -3.92898142e-01 4.08235162e-01
1.39543223e+00 -5.70839942e-01 -8.78890157e-02 -4.60216030e-02
1.33874023e+00 1.37675509e-01 -1.38413846e+00 -7.29375839e-01
-3.51969987e-01 -5.44282734e-01 -3.63643914e-02 -7.49823570e-01
-1.38091075e+00 3.06496173e-01 6.49263680e-01 1.79153413e-01
1.45825613e+00 1.43516377e-01 1.09786296e+00 8.29119533e-02
7.07111299e-01 -1.00147724e+00 -3.36174965e-01 1.38021395e-01
6.11179292e-01 -9.64424014e-01 -1.90516070e-01 -3.17147464e-01
-5.92401505e-01 7.18806386e-01 7.98249304e-01 -3.17048654e-02
5.69966912e-01 -1.24409936e-01 -1.48678482e-01 1.49127439e-01
-1.04563630e+00 -7.31345713e-02 -1.86390042e-01 2.93330282e-01
-3.14970165e-02 1.07451707e-01 -5.63717663e-01 1.18590522e+00
1.06641471e-01 -1.38581544e-01 8.79521295e-02 8.65234196e-01
-5.60472250e-01 -1.21252501e+00 -5.00280023e-01 3.27978432e-01
-5.94614089e-01 1.74750358e-01 -2.14460604e-02 1.66564196e-01
3.11613858e-01 9.15083885e-01 -2.59533226e-01 -5.04555762e-01
3.25629622e-01 6.50689527e-02 3.86497408e-01 -2.79865772e-01
-8.06272209e-01 1.01649515e-01 -6.83315918e-02 -5.92892468e-01
-1.21347286e-01 -5.61719358e-01 -1.05380189e+00 -8.19756329e-01
-3.28446358e-01 4.07375306e-01 4.45836157e-01 9.16530788e-01
6.16498709e-01 2.76425421e-01 1.16380179e+00 -7.93383420e-01
-8.71693194e-01 -8.09262931e-01 -6.50290668e-01 2.62099355e-01
1.54533446e-01 -4.24418151e-01 -5.68823457e-01 -2.57185638e-01] | [5.824272632598877, 6.27525520324707] |
697fb346-2d14-4610-8b7b-8c0fe84cf2dd | a-survey-of-embodied-ai-from-simulator-to | 2103.04918 | null | https://arxiv.org/abs/2103.04918v8 | https://arxiv.org/pdf/2103.04918v8.pdf | A Survey of Embodied AI: From Simulators to Research Tasks | There has been an emerging paradigm shift from the era of "internet AI" to "embodied AI", where AI algorithms and agents no longer learn from datasets of images, videos or text curated primarily from the internet. Instead, they learn through interactions with their environments from an egocentric perception similar to humans. Consequently, there has been substantial growth in the demand for embodied AI simulators to support various embodied AI research tasks. This growing interest in embodied AI is beneficial to the greater pursuit of Artificial General Intelligence (AGI), but there has not been a contemporary and comprehensive survey of this field. This paper aims to provide an encyclopedic survey for the field of embodied AI, from its simulators to its research. By evaluating nine current embodied AI simulators with our proposed seven features, this paper aims to understand the simulators in their provision for use in embodied AI research and their limitations. Lastly, this paper surveys the three main research tasks in embodied AI -- visual exploration, visual navigation and embodied question answering (QA), covering the state-of-the-art approaches, evaluation metrics and datasets. Finally, with the new insights revealed through surveying the field, the paper will provide suggestions for simulator-for-task selections and recommendations for the future directions of the field. | ['Hui Li Tan', 'Jiafei Duan', 'Cheston Tan', 'Hongyuan Zhu', 'Samson Yu'] | 2021-03-08 | null | null | null | null | ['embodied-question-answering'] | ['computer-vision'] | [ 1.58905089e-01 3.61357033e-01 2.60071427e-01 -6.73554465e-02
-1.61077991e-01 -6.82561994e-01 8.91471684e-01 -1.61967084e-01
-6.90865636e-01 5.94045460e-01 2.72991359e-01 -4.21798617e-01
-3.02096099e-01 -5.88993251e-01 -6.50919318e-01 -4.79909837e-01
-4.25607204e-01 2.71439731e-01 -2.61473745e-01 -7.33610332e-01
4.66857821e-01 2.90166825e-01 -2.10318756e+00 1.96331933e-01
9.82277930e-01 8.75150323e-01 5.35154641e-01 8.57192874e-01
-1.86917618e-01 1.27263808e+00 -6.16766751e-01 -2.64793307e-01
1.03879003e-02 -4.71817493e-01 -9.92250204e-01 -3.20440292e-01
1.78198457e-01 -4.38432455e-01 -6.75869584e-01 8.72714520e-01
6.93631768e-01 6.06759846e-01 4.39549714e-01 -1.82164514e+00
-1.12569249e+00 2.74378508e-01 -7.44729722e-03 2.29394600e-01
9.26814258e-01 4.44817185e-01 6.32347882e-01 -9.05500770e-01
1.07450175e+00 1.24295127e+00 2.15405464e-01 7.51052439e-01
-4.39957500e-01 -3.31639022e-01 5.43577790e-01 5.83916247e-01
-1.15236390e+00 -6.15016937e-01 6.48433924e-01 -3.73393863e-01
1.18669450e+00 3.76606345e-01 1.01181865e+00 1.37050939e+00
2.19782650e-01 1.09722459e+00 9.83230650e-01 -5.86370707e-01
4.93090153e-01 3.67850482e-01 -7.64986128e-02 1.02047729e+00
1.31065369e-01 4.87761050e-01 -9.28685963e-01 1.79955721e-01
5.60045242e-01 -5.25623858e-01 -2.47858852e-01 -1.00834274e+00
-1.67311525e+00 5.71175277e-01 7.04302311e-01 4.54556674e-01
-5.68038344e-01 5.04586637e-01 4.39280480e-01 5.14642715e-01
3.10897231e-01 5.57585716e-01 1.12250097e-01 -4.83400732e-01
-1.96757555e-01 7.06375659e-01 9.88702416e-01 1.21098602e+00
4.30949122e-01 3.01447242e-01 9.09053758e-02 2.79697210e-01
5.79049289e-01 5.49820781e-01 2.81106710e-01 -1.07435954e+00
1.19093060e-01 6.68213785e-01 3.08985651e-01 -1.09828055e+00
-4.15615290e-01 -2.37514134e-02 -2.38309547e-01 7.96941817e-01
1.23529650e-01 -2.89896369e-01 -9.44810569e-01 1.60811603e+00
4.02503997e-01 -3.42644572e-01 4.42663550e-01 1.29037833e+00
1.36911952e+00 7.14222729e-01 3.69414181e-01 4.56572503e-01
1.41253650e+00 -1.39161122e+00 -9.96069849e-01 -5.57816267e-01
6.04502439e-01 -1.21554688e-01 1.00835228e+00 1.67208821e-01
-9.92005229e-01 -4.06239003e-01 -1.00283861e+00 -1.77096829e-01
-1.13909173e+00 -1.86537310e-01 8.68134320e-01 4.85241860e-01
-1.29936969e+00 7.12556415e-04 -6.01075292e-01 -9.69875515e-01
2.83620834e-01 2.42007151e-01 -3.65738451e-01 -1.94361922e-03
-1.11344838e+00 1.38314271e+00 3.78186375e-01 2.76553839e-01
-1.10915625e+00 -2.91737199e-01 -1.01672602e+00 -4.24520224e-01
5.40540099e-01 -1.05701888e+00 1.28459823e+00 -1.15500677e+00
-1.56154239e+00 1.08184981e+00 -4.85085882e-02 -3.34399551e-01
4.66667652e-01 -3.69406790e-01 -6.56243801e-01 2.15961307e-01
-9.82891694e-02 1.01578522e+00 3.91596675e-01 -1.68741083e+00
-6.59235418e-01 -3.14626098e-01 6.76701784e-01 8.42966259e-01
1.17271923e-01 -3.93959172e-02 -2.12857068e-01 -3.56512994e-01
-3.62146139e-01 -1.10606897e+00 -1.31192133e-01 3.61327469e-01
1.79008007e-01 -1.97394297e-01 9.59700704e-01 -4.22300577e-01
7.34472096e-01 -2.04072595e+00 6.27285242e-01 -1.82014942e-01
6.14408493e-01 9.32281241e-02 -2.08486393e-01 9.51734900e-01
3.86861652e-01 -1.19198740e-01 5.31413518e-02 -1.30530849e-01
4.10799116e-01 1.39218330e-01 -1.99409544e-01 2.45288551e-01
-5.52622259e-01 1.37076688e+00 -1.36874223e+00 -4.34207022e-01
6.05279148e-01 6.18077695e-01 -1.68686405e-01 2.68938988e-01
-1.81758344e-01 6.15792513e-01 -5.28343022e-01 8.06000710e-01
2.26427242e-01 8.15783534e-03 -3.10411960e-01 4.08613980e-02
-5.56342125e-01 -2.13664502e-01 -6.92583919e-01 2.41984725e+00
-4.78612095e-01 1.18837166e+00 3.44261944e-01 -7.68289268e-01
6.59495234e-01 4.67567027e-01 4.12092209e-01 -1.10813677e+00
3.47602487e-01 1.96688160e-01 4.13330160e-02 -9.58144605e-01
6.86481297e-01 1.76928788e-01 -2.53731627e-02 3.30319017e-01
-1.44491389e-01 -1.72127798e-01 1.71742320e-01 2.37583265e-01
1.01045716e+00 6.10026062e-01 2.25312665e-01 -3.22610140e-01
3.75669569e-01 7.20938563e-01 -4.11982089e-01 9.24965620e-01
-5.38368225e-01 7.19680563e-02 -3.82311016e-01 -6.13267660e-01
-6.48360431e-01 -1.07258546e+00 2.49544218e-01 1.30447638e+00
8.56508076e-01 -2.52408177e-01 -9.98313963e-01 -3.58926207e-01
-2.17792913e-01 1.09780407e+00 -8.50770712e-01 -2.38888726e-01
-2.22108230e-01 -6.25266135e-02 5.30704319e-01 4.64690089e-01
1.09942937e+00 -1.89849639e+00 -1.81903207e+00 -6.78198189e-02
-3.44035476e-01 -8.48853290e-01 2.96396822e-01 -5.04443869e-02
-5.59172869e-01 -8.82344365e-01 -8.34380865e-01 -9.10980463e-01
4.88798141e-01 5.48504829e-01 1.18880129e+00 4.16796319e-02
-2.82675743e-01 1.41081512e+00 -8.17909539e-01 -9.22944963e-01
-1.31212279e-01 -2.02533275e-01 -8.02654922e-02 -4.87484097e-01
5.37813544e-01 -1.51350468e-01 -7.86094010e-01 2.14732096e-01
-8.85870814e-01 4.22034085e-01 6.16770864e-01 4.16183174e-01
1.94336642e-02 -4.55541313e-01 3.97101372e-01 -3.31554502e-01
9.91420507e-01 -7.31153548e-01 9.77680758e-02 4.23355043e-01
-3.45376343e-01 -1.41969115e-01 1.26200542e-01 -4.91968185e-01
-1.34090149e+00 -5.97755075e-01 4.20863181e-02 -2.05154717e-01
-4.70149398e-01 6.96869731e-01 -6.64134249e-02 -4.62764919e-01
8.77301693e-01 1.16970226e-01 3.91918048e-02 -2.58583203e-02
6.69799566e-01 6.67419314e-01 6.74253166e-01 -5.24803579e-01
4.18109030e-01 5.55202961e-01 -4.45932806e-01 -1.09086692e+00
-1.84647739e-01 -6.12106659e-02 -2.31972709e-01 -8.98375154e-01
1.04036570e+00 -6.27505422e-01 -9.22566593e-01 4.87315416e-01
-1.02243114e+00 -5.95112622e-01 -5.07749557e-01 5.15882432e-01
-9.26549375e-01 -8.19314346e-02 -2.14713186e-01 -1.09680283e+00
-2.26916596e-01 -1.16983032e+00 9.53374982e-01 5.48477769e-01
-5.14754415e-01 -1.13109756e+00 3.64517450e-01 4.69850242e-01
8.26201677e-01 4.09526587e-01 6.39441192e-01 -3.73263091e-01
-7.81896293e-01 4.03253781e-03 -1.33902803e-02 -3.39055687e-01
-1.81270123e-01 -3.65483612e-01 -9.85242486e-01 -2.11573467e-01
-2.97395498e-01 -5.16717613e-01 2.92233080e-01 1.82492018e-01
7.02516794e-01 2.70045903e-02 -5.87917566e-01 5.57972670e-01
1.22074854e+00 1.03249395e+00 7.14182138e-01 9.16810751e-01
4.59208965e-01 8.96845043e-01 8.79146039e-01 2.90972382e-01
7.01985896e-01 4.39098537e-01 7.23795772e-01 3.57697047e-02
-8.32454190e-02 -2.54767299e-01 2.91622609e-01 5.59847057e-01
-5.72729826e-01 -8.81189167e-01 -1.16166389e+00 5.05978465e-01
-1.90624249e+00 -8.82517397e-01 3.26424271e-01 1.73405385e+00
-1.93066046e-01 -5.29659450e-01 -4.84308787e-02 -2.48941585e-01
4.91456270e-01 1.42918289e-01 -9.18786943e-01 -6.80987895e-01
-1.06804878e-01 -2.23012194e-01 -3.81175429e-02 9.50019881e-02
-8.58929634e-01 1.11828160e+00 6.80030727e+00 8.13834444e-02
-7.95340478e-01 -7.53884315e-02 -2.42215898e-02 2.41904184e-01
-2.45518714e-01 -5.74789420e-02 1.10798813e-01 1.66315049e-01
7.93107152e-01 -4.96968955e-01 8.34737241e-01 8.14535320e-01
-1.54232308e-01 -6.72060668e-01 -1.20074821e+00 1.32990992e+00
2.18411759e-01 -1.13836455e+00 4.09571007e-02 5.42030111e-02
4.45039749e-01 2.88068745e-02 5.02077222e-01 5.04143059e-01
3.18183780e-01 -1.27132094e+00 1.19148564e+00 8.31982613e-01
4.48407233e-01 -5.78491569e-01 4.29878056e-01 3.31812739e-01
-1.24003220e+00 -2.45797604e-01 -7.78211355e-02 -2.57886022e-01
2.93409348e-01 -5.92301250e-01 -2.04517320e-01 6.10972941e-01
9.80482161e-01 3.36042404e-01 -1.86540380e-01 1.07364655e+00
3.23418468e-01 -1.55911028e-01 1.19684100e-01 -5.43752372e-01
4.89615232e-01 -3.58451068e-01 8.60766888e-01 8.78515542e-01
4.00673449e-01 4.85818148e-01 -2.00997233e-01 7.96168983e-01
2.81514734e-01 8.88120309e-02 -1.18287361e+00 -5.63483179e-01
3.57361555e-01 9.13952231e-01 -7.56364584e-01 -1.75949484e-01
-4.75058496e-01 1.28977847e+00 2.82331496e-01 7.11498380e-01
-1.00567532e+00 -6.28897309e-01 7.06699669e-01 -3.69033277e-01
-3.12928438e-01 -4.96195227e-01 1.69601440e-01 -8.98935616e-01
-2.27924109e-01 -1.29285038e+00 8.53509232e-02 -1.52718127e+00
-7.74700284e-01 9.21789944e-01 4.06043589e-01 -9.04558241e-01
-4.54696536e-01 -7.37146437e-01 -3.01910758e-01 4.69682038e-01
-1.12094629e+00 -1.33196449e+00 -1.02834129e+00 3.09629619e-01
6.67476475e-01 -3.51439416e-01 9.84814644e-01 -2.03768667e-02
-1.83494434e-01 1.09010860e-01 5.68290837e-02 -1.65280536e-01
4.03442264e-01 -9.47616756e-01 6.28620446e-01 2.73405641e-01
1.04460336e-01 4.72736388e-01 9.81874585e-01 -4.89396691e-01
-2.00266552e+00 -2.41578490e-01 2.47336596e-01 -7.30690897e-01
4.69359040e-01 -2.17783868e-01 -4.23859596e-01 1.02257657e+00
8.32344234e-01 -4.35027868e-01 5.04283726e-01 -4.01760131e-01
1.94032341e-01 4.10011351e-01 -1.22927332e+00 1.18354785e+00
1.66175497e+00 -5.46772778e-01 -6.95456266e-01 -6.52543232e-02
5.40708899e-01 -5.89873850e-01 -4.91887242e-01 2.13257328e-01
9.42988694e-01 -1.10547256e+00 1.03267562e+00 -4.52253252e-01
1.60854295e-01 -2.50884473e-01 -1.49485633e-01 -1.72127616e+00
-2.76465476e-01 -5.01093686e-01 -2.96897650e-01 6.02346182e-01
9.17209983e-02 -8.55186403e-01 5.72396100e-01 8.24140549e-01
-3.95047873e-01 -5.80193579e-01 -5.59434831e-01 -4.91565824e-01
-1.74952626e-01 -3.01398218e-01 5.42154551e-01 6.81296170e-01
3.08453411e-01 8.34472924e-02 -1.16943181e-01 -3.44578512e-02
4.74385858e-01 -1.08773418e-01 8.98389935e-01 -1.04226279e+00
2.98781693e-01 -5.45487046e-01 -6.35539591e-01 -1.22507668e+00
1.72447026e-01 -6.54050231e-01 1.09663911e-01 -2.13230348e+00
-8.06620121e-02 -1.59799263e-01 4.00374755e-02 1.91417914e-02
3.23594153e-01 8.35830495e-02 3.77174556e-01 5.22341859e-03
-9.01157379e-01 7.79086113e-01 1.54254627e+00 -3.08022872e-02
-1.87730923e-01 -6.85017765e-01 -7.04709172e-01 6.83748960e-01
6.43330574e-01 1.39256701e-01 -8.83446336e-01 -7.77529955e-01
6.13675535e-01 1.02320112e-01 7.82169223e-01 -1.41449475e+00
6.15123034e-01 -6.12305291e-02 1.46615610e-01 -5.83314240e-01
7.33491361e-01 -1.18753242e+00 3.83471698e-01 2.76735991e-01
-4.01155204e-01 5.16129971e-01 4.95030016e-01 5.96025169e-01
-1.42101914e-01 -1.63314119e-01 9.35763121e-02 -3.58842731e-01
-1.70626736e+00 -8.22486579e-02 -9.93640542e-01 1.25586882e-01
1.53737235e+00 -7.83965468e-01 -4.29727018e-01 -7.25248337e-01
-6.39795840e-01 6.51368082e-01 6.04687691e-01 7.74416089e-01
8.70192409e-01 -1.09980416e+00 -3.34595501e-01 -7.71382526e-02
4.03505623e-01 -2.11957529e-01 5.59153736e-01 5.12457371e-01
-7.62913942e-01 7.18672156e-01 -5.99092782e-01 -2.34049127e-01
-1.12222433e+00 8.08146298e-01 5.17949164e-01 5.52758455e-01
-6.58019185e-01 7.99031615e-01 4.32184249e-01 -5.94943345e-01
5.55572271e-01 -5.00832759e-02 -3.40063393e-01 -2.56285250e-01
3.65673989e-01 7.11762786e-01 -4.03024197e-01 -8.11151445e-01
-5.49632192e-01 3.99399459e-01 4.41150457e-01 -5.08430600e-01
9.24104095e-01 -4.52514410e-01 1.26025528e-01 7.27064133e-01
8.14150751e-01 -3.86381119e-01 -1.04434752e+00 2.33777031e-01
-1.61949992e-01 -3.38773876e-01 6.15332015e-02 -1.21264279e+00
-6.00764632e-01 8.18662763e-01 8.28321993e-01 1.53372392e-01
8.99247229e-01 1.06886216e-01 4.53716964e-01 8.59046221e-01
8.07150245e-01 -1.07467985e+00 3.13282996e-01 4.14651692e-01
1.27380800e+00 -1.04405212e+00 -4.16064829e-01 8.36569890e-02
-1.11998475e+00 7.84896612e-01 7.18203247e-01 6.02041259e-02
3.58501226e-01 1.98620707e-01 2.28505716e-01 -7.56964803e-01
-5.90910554e-01 -3.40364188e-01 3.25472057e-01 1.16112292e+00
3.09886307e-01 -2.23936364e-01 1.82439223e-01 9.03504714e-02
-3.32323104e-01 7.41419420e-02 1.00823067e-01 1.44143057e+00
-4.52546716e-01 -4.97600108e-01 -2.46240392e-01 3.83717529e-02
2.70030916e-01 2.55082369e-01 -7.66964316e-01 1.38193667e+00
-8.80393479e-03 1.12003326e+00 1.70076117e-01 -4.20704544e-01
4.73718256e-01 1.92605987e-01 7.18518198e-01 -1.25755705e-02
-5.30992687e-01 -9.45159972e-01 1.98308870e-01 -7.03734756e-01
-7.37320423e-01 -4.02649403e-01 -1.42691398e+00 -5.99849939e-01
-1.21468030e-01 1.98259324e-01 1.13725424e+00 7.72442937e-01
4.79741812e-01 5.92689335e-01 -2.63227820e-01 -1.18919742e+00
4.38055843e-02 -5.49537539e-01 -1.89203709e-01 2.02024523e-02
3.33026737e-01 -1.08243334e+00 -3.71584892e-01 -1.89838380e-01] | [4.48035192489624, 0.726983904838562] |
fd689272-ff6b-4817-8fda-2df6dc0fda62 | rankmix-data-augmentation-for-weakly | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_RankMix_Data_Augmentation_for_Weakly_Supervised_Learning_of_Classifying_Whole_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_RankMix_Data_Augmentation_for_Weakly_Supervised_Learning_of_Classifying_Whole_CVPR_2023_paper.pdf | RankMix: Data Augmentation for Weakly Supervised Learning of Classifying Whole Slide Images With Diverse Sizes and Imbalanced Categories | Whole Slide Images (WSIs) are usually gigapixel in size and lack pixel-level annotations. The WSI datasets are also imbalanced in categories. These unique characteristics, significantly different from the ones in natural images, pose the challenge of classifying WSI images as a kind of weakly supervise learning problems. In this study, we propose, RankMix, a data augmentation method of mixing ranked features in a pair of WSIs. RankMix introduces the concepts of pseudo labeling and ranking in order to extract key WSI regions in contributing to the WSI classification task. A two-stage training is further proposed to boost stable training and model performance. To our knowledge, the study of weakly supervised learning from the perspective of data augmentation to deal with the WSI classification problem that suffers from lack of training data and imbalance of categories is relatively unexplored. | ['Chun-Shien Lu', 'Yuan-Chih Chen'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['whole-slide-images'] | ['computer-vision'] | [ 8.83241296e-01 2.54438460e-01 -5.32804489e-01 -6.04188740e-01
-1.00747728e+00 -5.21134734e-01 6.82580769e-01 3.59623164e-01
-3.16936225e-01 7.79276192e-01 2.80132174e-01 -5.66238388e-02
-1.61498442e-01 -3.66342723e-01 -6.53996766e-01 -1.09347963e+00
1.27521828e-01 2.98068047e-01 2.74021775e-01 -1.18409224e-01
3.05775076e-01 3.34870428e-01 -1.75269532e+00 7.06214845e-01
8.77411664e-01 1.13108706e+00 6.51795301e-04 6.00894749e-01
-3.82764667e-01 7.34381557e-01 -6.18537664e-01 5.92614375e-02
1.98561519e-01 -3.43098819e-01 -8.46938014e-01 2.72687465e-01
8.86511087e-01 -1.37576889e-02 -1.84179723e-01 9.81871128e-01
5.27742207e-01 -2.05275238e-01 8.06666911e-01 -1.52101052e+00
-1.96469754e-01 4.49093610e-01 -9.39583957e-01 4.63879645e-01
3.49846925e-03 -1.01841139e-02 1.02473342e+00 -7.91955352e-01
7.62101769e-01 7.77531266e-01 4.76726919e-01 6.49067700e-01
-1.07454395e+00 -6.67829216e-01 2.02729017e-01 2.06728444e-01
-1.41096044e+00 -4.13526982e-01 8.54722500e-01 -3.82582515e-01
3.49778861e-01 7.14924335e-01 4.81544524e-01 9.15461242e-01
-2.95528024e-01 1.17008257e+00 1.45644510e+00 -4.92923260e-01
1.57632604e-01 3.19517165e-01 4.99189168e-01 2.63657004e-01
3.02141815e-01 -1.70581490e-01 -9.90848064e-01 -6.90107234e-03
2.37849876e-01 1.31414652e-01 -1.41966537e-01 -3.21565300e-01
-1.32699561e+00 3.36780161e-01 4.25099611e-01 5.59913039e-01
1.30671978e-01 -2.71328986e-01 3.81566793e-01 3.61286610e-01
7.26893961e-01 3.84297162e-01 -5.31700909e-01 2.61949182e-01
-1.18895388e+00 1.02889158e-01 3.03089619e-01 6.72529697e-01
6.41703963e-01 -3.30972791e-01 -2.77871907e-01 8.39213729e-01
1.11233152e-01 1.20624840e-01 6.32414520e-01 -4.11273390e-01
4.82011557e-01 1.21922338e+00 -1.46544978e-01 -6.64825618e-01
-5.56147873e-01 -9.68136489e-01 -1.07121921e+00 2.73187190e-01
6.01355195e-01 4.45329428e-01 -1.07884288e+00 1.59522414e+00
4.92987543e-01 5.25585592e-01 5.10934740e-02 9.54437792e-01
1.16844690e+00 5.24403393e-01 -9.14362669e-02 -1.81895614e-01
1.30410683e+00 -8.18204045e-01 -6.80630863e-01 -2.03878731e-01
6.20336711e-01 -6.43573463e-01 1.28730011e+00 4.20354724e-01
-9.97604907e-01 -5.39793015e-01 -1.25474620e+00 -1.02463558e-01
-5.61424971e-01 5.17480791e-01 4.38092291e-01 2.08964244e-01
-8.72458339e-01 1.11965917e-01 -6.47817254e-01 -2.64303803e-01
8.60179782e-01 3.69856209e-01 -6.21507466e-01 -1.12570211e-01
-9.26576138e-01 6.94865406e-01 2.82024026e-01 1.48717403e-01
-8.11214685e-01 -9.38945889e-01 -5.92179477e-01 -5.76936081e-02
5.68553865e-01 -2.07934573e-01 8.40714276e-01 -1.01302707e+00
-1.07313585e+00 1.36881733e+00 -7.92752355e-02 -1.66983575e-01
6.50888860e-01 8.63246471e-02 -1.73833489e-01 -9.98293012e-02
7.21216574e-02 4.33491051e-01 7.91807532e-01 -1.48303783e+00
-8.29056501e-01 -9.75222111e-01 -6.88373223e-02 3.68649393e-01
-7.52742648e-01 -4.07544971e-01 -5.36821723e-01 -6.95829034e-01
4.14189100e-01 -8.82510722e-01 -2.08382502e-01 3.71980928e-02
-8.23439658e-01 -1.99541822e-01 1.00194240e+00 -3.19746882e-01
1.05429566e+00 -2.21730471e+00 -3.17094326e-02 2.33924866e-01
4.78503317e-01 3.98011833e-01 -2.86097020e-01 -1.24801733e-01
-2.93211132e-01 1.06492236e-01 -3.36143374e-01 -3.94589841e-01
-2.36694112e-01 3.08363028e-02 1.69876609e-02 5.50153196e-01
3.49181592e-01 7.50269771e-01 -7.30794072e-01 -6.72440290e-01
1.24423830e-02 5.76284453e-02 -2.70041615e-01 4.30523485e-01
1.03681959e-01 6.60999060e-01 -2.45346993e-01 7.43649721e-01
8.79529119e-01 -3.90715331e-01 -7.32327253e-02 -5.40252447e-01
-1.14479199e-01 -7.99405873e-02 -1.26309812e+00 1.80566967e+00
4.07409109e-02 5.24998546e-01 1.73475534e-01 -1.55438304e+00
6.89872921e-01 9.90484357e-02 6.00470781e-01 -4.62489307e-01
1.30149364e-01 2.53453821e-01 -2.97376662e-02 -5.35207927e-01
2.04076856e-01 -2.14396283e-01 -1.50291294e-01 2.04750806e-01
1.38143629e-01 -1.64474875e-01 5.32931685e-01 2.05534786e-01
8.80939901e-01 -2.53624804e-02 2.91629076e-01 -3.08767021e-01
6.16556466e-01 6.09606728e-02 7.51812100e-01 7.26295292e-01
-2.16494858e-01 8.42061520e-01 5.69655597e-01 -2.55589902e-01
-1.08066869e+00 -1.02651453e+00 -4.96869445e-01 1.11932123e+00
3.27086121e-01 -7.47106075e-02 -6.19043112e-01 -8.44589055e-01
-5.57673536e-02 -9.51949954e-02 -8.02546978e-01 -2.77479798e-01
-2.91010588e-01 -1.37259376e+00 4.00898844e-01 5.46359956e-01
2.84532666e-01 -5.38634002e-01 -1.65392563e-01 -1.64753839e-01
-2.29217619e-01 -1.07267809e+00 -2.20707506e-01 5.06677091e-01
-6.99341774e-01 -1.32245958e+00 -7.77494609e-01 -1.06579423e+00
1.13898861e+00 3.65229726e-01 1.09148026e+00 2.14706376e-01
-8.25101674e-01 -1.94291040e-01 -2.58662075e-01 -5.95221579e-01
-2.32638925e-01 2.84417748e-01 -1.94131836e-01 2.69864500e-01
3.23091000e-01 -3.24430853e-01 -7.95309544e-01 2.84511209e-01
-1.41606283e+00 3.79783720e-01 7.76540577e-01 1.08863819e+00
8.34637702e-01 8.85147676e-02 5.80036819e-01 -1.23852324e+00
1.44316554e-01 -4.02596921e-01 -3.29935312e-01 3.10150117e-01
-4.01525050e-01 -1.20072134e-01 3.36818814e-01 -3.76985043e-01
-1.19056356e+00 1.12449110e-01 -5.99473417e-02 1.71635538e-01
-4.08141345e-01 3.64855886e-01 -3.34027618e-01 -9.48469639e-02
6.51008546e-01 1.55465022e-01 2.09239230e-01 -1.99801371e-01
4.81962003e-02 8.78041327e-01 5.30478001e-01 -1.43507645e-01
8.82222652e-01 7.81415939e-01 -1.41513515e-02 -8.20426524e-01
-1.29077995e+00 -9.22170103e-01 -7.61012197e-01 -1.39290884e-01
5.20191669e-01 -8.55613291e-01 -3.02100867e-01 8.38926375e-01
-4.78358686e-01 -3.61825734e-01 -5.19393802e-01 2.10302025e-01
-3.57369661e-01 2.96409994e-01 -6.47949457e-01 -5.38268864e-01
-2.71126270e-01 -1.05512440e+00 1.21898127e+00 3.94565105e-01
3.51592265e-02 -5.60433209e-01 1.00202583e-01 7.86457300e-01
4.09975737e-01 3.82860303e-01 7.98999488e-01 -9.38207865e-01
-4.56670046e-01 -4.93478894e-01 -3.78357112e-01 4.92181927e-01
1.73360080e-01 -1.64417867e-02 -1.13976896e+00 -4.23573643e-01
-1.55603081e-01 -7.59100199e-01 1.08027256e+00 4.35184181e-01
1.31836188e+00 -2.27483995e-02 -4.02068377e-01 5.65478086e-01
1.50794578e+00 -1.15527950e-01 2.90796727e-01 2.91366875e-01
5.70502520e-01 8.42314124e-01 9.83052373e-01 3.78166825e-01
2.14757025e-01 5.36106765e-01 5.50903976e-01 -7.98394024e-01
-4.16982204e-01 1.26367792e-01 -2.13440463e-01 6.37691736e-01
2.05202743e-01 -2.80877829e-01 -8.15849721e-01 5.89942396e-01
-1.58847749e+00 -4.78351682e-01 -1.77214280e-01 2.01290631e+00
9.55722153e-01 1.75785214e-01 -1.38852790e-01 6.76879346e-01
5.89268684e-01 2.79287368e-01 -6.82541609e-01 4.27191287e-01
-4.11632419e-01 -1.42085284e-01 4.63158697e-01 1.15233123e-01
-1.28829014e+00 5.45502841e-01 5.50080204e+00 1.00948119e+00
-1.14670408e+00 1.06458180e-01 1.24091804e+00 -2.96984673e-01
-1.25327379e-01 -3.06221664e-01 -9.61257219e-01 5.79563677e-01
4.66711372e-01 4.89992768e-01 -9.54928100e-02 4.76730317e-01
1.24871535e-02 -9.59027186e-02 -8.56850624e-01 7.61040270e-01
3.32510352e-01 -1.28152537e+00 -2.79485583e-02 3.45742516e-02
9.94331181e-01 -1.61387362e-02 3.72357130e-01 2.17019185e-01
-4.70025055e-02 -9.79548991e-01 4.50432003e-01 3.46850783e-01
8.72223258e-01 -5.62500954e-01 1.16061175e+00 3.54876816e-01
-9.44198072e-01 -1.68863937e-01 -2.88158447e-01 1.37208909e-01
-2.09453493e-01 8.76047611e-01 -8.14351141e-01 3.29782158e-01
8.11454833e-01 6.68540359e-01 -8.92338216e-01 1.03856969e+00
1.18312001e-01 6.60556555e-01 -3.33028257e-01 2.84735531e-01
9.68400389e-02 -2.71110088e-02 2.35837474e-01 1.08680201e+00
1.48408473e-01 -1.60187352e-02 2.77024031e-01 3.97298366e-01
-1.00434534e-01 2.82732606e-01 -3.65259230e-01 5.93437254e-02
1.77798256e-01 1.46182704e+00 -9.79282856e-01 -4.07092214e-01
-3.17769974e-01 7.87297487e-01 2.39458412e-01 2.81050324e-01
-4.04479980e-01 -3.99075821e-02 4.39848721e-01 3.51988763e-01
-2.12486446e-01 1.95954859e-01 -5.51981688e-01 -1.29030156e+00
1.52154863e-01 -7.16339171e-01 6.12160444e-01 -2.28845030e-01
-1.35214758e+00 5.41708291e-01 -1.81627631e-01 -1.59109807e+00
2.97960103e-01 -6.03894711e-01 -6.67441428e-01 2.60112762e-01
-2.27219009e+00 -1.51041651e+00 -1.02475023e+00 4.67138916e-01
4.79779065e-01 -1.51372582e-01 4.76419479e-01 4.30218577e-01
-7.06352949e-01 6.08249724e-01 3.54102224e-01 4.86157415e-03
9.70942199e-01 -1.56546581e+00 -1.58232391e-01 5.55675745e-01
-7.99232945e-02 2.56710410e-01 6.20030820e-01 -2.71640390e-01
-1.09802234e+00 -1.08621562e+00 5.33381522e-01 -2.12647364e-01
3.46493661e-01 -2.79599279e-01 -1.01011944e+00 2.56732255e-01
-9.53797996e-02 6.25147879e-01 1.32589912e+00 -6.57702908e-02
-6.43405989e-02 -4.89334315e-01 -1.18445122e+00 3.33707601e-01
1.04661226e+00 -3.61219645e-01 -3.25002968e-01 5.35746455e-01
1.90999895e-01 -4.92295712e-01 -8.39612424e-01 6.05757892e-01
2.70399839e-01 -7.63411582e-01 1.17780554e+00 -6.70690894e-01
3.46011370e-01 -5.92888474e-01 1.08178675e-01 -1.24445927e+00
-1.28480699e-02 2.98223887e-02 1.46730736e-01 1.62702465e+00
4.00966018e-01 -2.16020122e-01 1.34568667e+00 2.36500755e-01
-1.65506437e-01 -6.98460698e-01 -9.69783366e-01 -3.96418154e-01
-1.28308490e-01 -1.04697153e-01 4.36607748e-01 9.84842777e-01
-2.15275884e-02 1.40190601e-01 -2.33654857e-01 1.13225192e-01
8.66842508e-01 1.68599159e-01 7.98619568e-01 -1.50353122e+00
5.08750044e-02 -4.68111932e-01 -6.09462321e-01 -4.93291140e-01
7.62457624e-02 -1.02255106e+00 2.14195982e-01 -1.38573992e+00
6.30936921e-01 -7.05196559e-01 -7.43623257e-01 3.19621801e-01
-4.53697979e-01 9.23429966e-01 -9.70964953e-02 2.92615265e-01
-7.13164568e-01 4.27295476e-01 1.24451530e+00 -5.71821690e-01
6.19181059e-02 8.24052691e-02 -8.36426556e-01 7.72087455e-01
8.09680998e-01 -1.65104821e-01 -5.52873373e-01 -1.81286167e-02
1.03786089e-01 -1.52637124e-01 6.29745275e-02 -1.00708616e+00
3.22126687e-01 -1.08453147e-01 5.21175146e-01 -8.18532109e-01
1.64136693e-01 -7.51832128e-01 -2.81110168e-01 3.36809486e-01
-8.01710486e-01 -5.66017151e-01 -1.10607408e-01 5.02747715e-01
-6.61596417e-01 -3.09956521e-01 1.01491070e+00 -1.50127351e-01
-6.58079922e-01 5.46777308e-01 -1.80762023e-01 8.13061371e-02
1.19251835e+00 -3.81179184e-01 -3.68907899e-01 4.53723669e-02
-7.63044178e-01 2.79074222e-01 3.30801874e-01 1.81562483e-01
6.15563929e-01 -1.10175705e+00 -8.76512349e-01 4.03104335e-01
6.54843569e-01 5.42674124e-01 5.63253224e-01 9.37478423e-01
-2.86754906e-01 1.56817317e-01 -3.16624254e-01 -8.17401290e-01
-1.47710240e+00 3.47665548e-01 -1.24853356e-02 -7.63694406e-01
-2.24095568e-01 1.09139228e+00 6.10868931e-01 -6.90879881e-01
4.90818232e-01 -2.65995294e-01 -5.72780490e-01 4.59416538e-01
6.72079086e-01 7.34356865e-02 3.80011231e-01 -4.60348427e-01
-2.32083708e-01 5.73822260e-01 -3.70409727e-01 3.52335393e-01
1.58218718e+00 -3.07118654e-01 -2.62461573e-01 3.95657897e-01
1.37862945e+00 -2.69654483e-01 -1.34527993e+00 -6.02222741e-01
2.80995637e-01 -4.67077225e-01 6.18735142e-02 -8.57701123e-01
-1.13439393e+00 9.09589231e-01 9.98867691e-01 -1.20813027e-01
1.23914790e+00 1.35736451e-01 6.29025936e-01 -8.54687542e-02
-3.60541195e-02 -1.28141570e+00 4.28805649e-01 -1.08172987e-02
6.14111722e-01 -1.70179188e+00 -1.81230962e-01 -5.50443649e-01
-5.19456804e-01 9.13334787e-01 9.28248942e-01 1.08508885e-01
4.35057014e-01 6.06903076e-01 1.81094781e-01 -1.50372297e-01
-5.76158822e-01 -3.50192010e-01 4.48639095e-01 6.35198295e-01
3.56957614e-01 -1.48130849e-01 -3.72804612e-01 5.67873180e-01
1.76502079e-01 1.20825700e-01 3.15858901e-01 9.78568435e-01
-4.39224064e-01 -9.38442647e-01 -2.96545357e-01 9.16728675e-01
-5.06591856e-01 2.18492314e-01 -2.17136487e-01 4.82461065e-01
3.48588437e-01 7.64695466e-01 1.63451165e-01 -2.05830723e-01
3.04710060e-01 -9.48017389e-02 1.17340833e-01 -4.94566202e-01
-3.90592158e-01 -5.77393547e-02 -1.99880689e-01 -3.20704281e-01
-6.02984905e-01 -4.90620583e-01 -1.12044775e+00 2.71086276e-01
-4.81535465e-01 1.06348544e-01 6.72556877e-01 1.09128439e+00
2.13840026e-02 7.49793470e-01 7.60756373e-01 -7.42935240e-01
-4.13882464e-01 -1.00135887e+00 -7.97353566e-01 9.53699410e-01
6.54396236e-01 -5.64390421e-01 -4.88620460e-01 1.29334569e-01] | [15.09763240814209, -2.778594970703125] |
5b904b9c-d8fc-427b-bb7e-a8f3467d3f87 | sharp-sparsity-and-hidden-activation-replay | 2305.18563 | null | https://arxiv.org/abs/2305.18563v1 | https://arxiv.org/pdf/2305.18563v1.pdf | SHARP: Sparsity and Hidden Activation RePlay for Neuro-Inspired Continual Learning | Deep neural networks (DNNs) struggle to learn in dynamic environments since they rely on fixed datasets or stationary environments. Continual learning (CL) aims to address this limitation and enable DNNs to accumulate knowledge incrementally, similar to human learning. Inspired by how our brain consolidates memories, a powerful strategy in CL is replay, which involves training the DNN on a mixture of new and all seen classes. However, existing replay methods overlook two crucial aspects of biological replay: 1) the brain replays processed neural patterns instead of raw input, and 2) it prioritizes the replay of recently learned information rather than revisiting all past experiences. To address these differences, we propose SHARP, an efficient neuro-inspired CL method that leverages sparse dynamic connectivity and activation replay. Unlike other activation replay methods, which assume layers not subjected to replay have been pretrained and fixed, SHARP can continually update all layers. Also, SHARP is unique in that it only needs to replay few recently seen classes instead of all past classes. Our experiments on five datasets demonstrate that SHARP outperforms state-of-the-art replay methods in class incremental learning. Furthermore, we showcase SHARP's flexibility in a novel CL scenario where the boundaries between learning episodes are blurry. The SHARP code is available at \url{https://github.com/BurakGurbuz97/SHARP-Continual-Learning}. | ['Constantine Dovrolis', 'Jean Michael Moorman', 'Mustafa Burak Gurbuz'] | 2023-05-29 | null | null | null | null | ['class-incremental-learning', 'incremental-learning'] | ['computer-vision', 'methodology'] | [ 1.30678862e-01 -1.27941251e-01 -1.38920128e-01 -1.83316782e-01
-5.96313588e-02 -5.67235827e-01 7.14866579e-01 -9.49671045e-02
-6.32240117e-01 9.61967289e-01 8.38693008e-02 -6.60271272e-02
-1.96696043e-01 -7.98150182e-01 -1.11095941e+00 -7.80592322e-01
-6.03908161e-03 2.66985238e-01 4.57955986e-01 -8.77739191e-02
1.32549867e-01 4.76888567e-01 -1.50195289e+00 4.70618546e-01
6.11712515e-01 5.60289323e-01 3.08406264e-01 3.95987600e-01
-1.49237812e-01 9.23763394e-01 -4.27570939e-01 -1.85825348e-01
1.69095472e-01 -5.13769448e-01 -8.20795536e-01 -1.64564058e-01
5.95984936e-01 -4.03269172e-01 -6.85117126e-01 9.66790497e-01
3.56970698e-01 4.21196401e-01 2.26563007e-01 -1.17398489e+00
-9.53235626e-01 8.91455412e-01 -5.99436998e-01 8.64624143e-01
2.48515774e-02 3.07139337e-01 6.07025921e-01 -1.18453658e+00
6.06924295e-01 9.42691445e-01 7.70858347e-01 1.03326178e+00
-1.31245971e+00 -8.81478131e-01 6.96240485e-01 1.79446921e-01
-1.08665073e+00 -6.84769511e-01 7.74174213e-01 -7.76430219e-02
1.08221841e+00 -1.21295854e-01 1.10456574e+00 1.51233196e+00
1.89386234e-01 1.17440116e+00 8.93036306e-01 -2.40157470e-01
4.37556744e-01 -3.13819259e-01 4.50081646e-01 6.33590281e-01
4.11009401e-01 1.30934894e-01 -1.00738347e+00 1.42790839e-01
9.21871901e-01 6.60995364e-01 -5.63258529e-01 -1.32898211e-01
-1.38706076e+00 2.67910451e-01 6.67823732e-01 3.58443111e-01
-5.47971070e-01 4.03805226e-01 3.04067165e-01 5.99349439e-01
2.33625218e-01 4.56671804e-01 -5.09353995e-01 -2.94849053e-02
-1.00586224e+00 -2.25866735e-02 4.66830820e-01 6.33406639e-01
1.05303586e+00 3.50000530e-01 1.26544118e-01 9.30798292e-01
-9.47061405e-02 2.13745922e-01 9.43199396e-01 -8.62893820e-01
1.44160688e-01 5.74972510e-01 -3.46485853e-01 -4.95837003e-01
-1.44007951e-01 -6.54258966e-01 -1.03183377e+00 -5.14086224e-02
2.51177251e-01 -1.68715835e-01 -1.20548320e+00 2.06445241e+00
9.64966137e-03 8.89730692e-01 1.86806768e-01 6.44580841e-01
7.94713557e-01 6.33650422e-01 -8.47192705e-02 -1.73196137e-01
7.68924892e-01 -1.05806839e+00 -4.29680586e-01 -4.11776692e-01
8.54717195e-02 -1.43876811e-02 1.15483880e+00 5.63457370e-01
-1.24718034e+00 -5.78845263e-01 -1.16444492e+00 2.44741142e-01
-3.42866302e-01 -3.68682772e-01 8.39103937e-01 2.11088419e-01
-1.50041234e+00 7.35008776e-01 -1.10776377e+00 -2.28494167e-01
7.53550529e-01 4.12875235e-01 -3.86629790e-01 -2.29135111e-01
-1.01824772e+00 4.69949961e-01 5.06588519e-01 2.48111397e-01
-1.17738211e+00 -1.00856125e+00 -6.01646423e-01 1.23753980e-01
3.16647053e-01 -8.57201159e-01 1.33307004e+00 -1.19045937e+00
-1.43836570e+00 5.57420135e-01 -2.05107719e-01 -7.74171352e-01
3.21840405e-01 -5.91840804e-01 -2.56789327e-01 2.58137941e-01
-2.69106567e-01 1.00124466e+00 8.10361087e-01 -1.22312212e+00
-2.80478925e-01 -3.07638168e-01 -2.20056579e-01 1.52575240e-01
-5.25664747e-01 -6.99219346e-01 -4.85344708e-01 -6.65950060e-01
2.94612914e-01 -7.53439844e-01 -2.55302293e-04 -4.38201204e-02
-7.07742665e-03 -5.44959716e-02 1.08414674e+00 7.95060322e-02
1.00324416e+00 -2.26646638e+00 1.65584996e-01 -3.79266515e-02
4.17453945e-01 5.19535661e-01 -3.75143528e-01 2.84774333e-01
-2.57820994e-01 -1.09963819e-01 -4.12230104e-01 -3.85581195e-01
-4.18621600e-01 4.89973426e-01 -8.03252876e-01 1.49069101e-01
1.43303007e-01 1.19159067e+00 -1.31835020e+00 7.47861937e-02
-7.51455352e-02 5.66603065e-01 -5.36399603e-01 -1.65827245e-01
-1.75834373e-01 3.30958784e-01 3.91778611e-02 6.90377057e-01
6.42948866e-01 -3.66754740e-01 2.41019335e-02 1.06071226e-01
-4.52573970e-02 1.10416643e-01 -8.99856269e-01 2.01222134e+00
-1.40137866e-01 6.59308016e-01 -2.82905132e-01 -1.03536844e+00
8.36607277e-01 3.18433106e-01 4.85480070e-01 -8.37560952e-01
-1.66148052e-01 1.45542353e-01 -8.55176672e-02 5.46696223e-03
2.44903460e-01 8.67184624e-02 3.72473747e-01 8.16296160e-01
4.23784465e-01 1.48405731e-01 6.36281148e-02 4.53353375e-01
1.55402946e+00 2.46991124e-02 -1.89245027e-02 -7.11012334e-02
1.66381940e-01 -2.72687733e-01 8.71481478e-01 9.95111465e-01
-3.46890032e-01 5.74965417e-01 2.31753200e-01 -7.79269040e-01
-7.06115127e-01 -1.58494496e+00 4.29419056e-02 1.06010830e+00
1.13488264e-01 -3.19892727e-02 -4.60392803e-01 -6.60261810e-01
-3.18869948e-02 6.92983449e-01 -8.42143834e-01 -6.00116372e-01
-6.39197826e-01 -9.12302613e-01 4.52902734e-01 5.87726653e-01
9.86300290e-01 -1.46007073e+00 -7.55855083e-01 3.75967294e-01
1.34298086e-01 -6.82045698e-01 -4.16656494e-01 3.56049299e-01
-1.35192621e+00 -9.30097520e-01 -8.62735093e-01 -8.26359987e-01
9.01499987e-01 6.76281333e-01 1.15155947e+00 3.07578743e-01
-1.88866735e-01 7.02978849e-01 -1.88403517e-01 -2.46983171e-01
3.39285508e-02 2.32564524e-01 2.63374627e-01 4.26085666e-02
3.93039584e-01 -1.26102245e+00 -9.24468696e-01 9.19332877e-02
-1.14936018e+00 2.60941625e-01 8.55957210e-01 1.00500309e+00
7.41133749e-01 -9.80408341e-02 1.18419659e+00 -1.22361207e+00
6.04783356e-01 -7.52554357e-01 -1.83768332e-01 2.82791346e-01
-5.22899568e-01 2.73758844e-02 4.41750854e-01 -9.51101422e-01
-1.08214903e+00 5.71532398e-02 -4.00327109e-02 -6.06700301e-01
-4.43474762e-02 5.09753346e-01 2.92812437e-01 5.60453199e-02
8.62292588e-01 7.28490114e-01 -1.01023972e-01 -2.42204547e-01
1.63121909e-01 -5.57121933e-02 8.66604924e-01 -6.87964499e-01
4.98275757e-01 5.60725331e-01 -5.39521515e-01 -6.69387043e-01
-8.94970179e-01 -3.01377147e-01 -6.50800645e-01 -4.25880730e-01
3.78803343e-01 -9.75513458e-01 -5.59759259e-01 9.28687990e-01
-9.47833061e-01 -9.72302020e-01 -6.32939577e-01 3.52313131e-01
-3.99314523e-01 5.61764911e-02 -9.46666241e-01 -4.83675033e-01
-4.69743252e-01 -6.79802299e-01 4.52425957e-01 5.05401373e-01
-3.59178782e-02 -8.96776676e-01 2.71012723e-01 -2.65497744e-01
6.24051809e-01 1.66793093e-01 6.41688645e-01 -4.38938737e-01
-5.90112984e-01 1.12823151e-01 6.82494463e-03 1.43109575e-01
3.60486805e-01 -1.56275630e-01 -1.06172419e+00 -6.13330126e-01
2.76389532e-02 -6.48886442e-01 1.56853533e+00 2.58552223e-01
1.29915917e+00 -4.17601109e-01 -4.17255312e-01 5.82676828e-01
1.27046871e+00 1.17934108e-01 6.91493332e-01 2.28870422e-01
4.47278798e-01 1.22024804e-01 -1.17559820e-01 2.92124510e-01
2.36110508e-01 -5.71205728e-02 4.81557459e-01 8.92241076e-02
-4.14805353e-01 -2.69657344e-01 4.52746302e-01 1.05524170e+00
-5.00559583e-02 -6.78364607e-03 -9.52508628e-01 7.68368542e-01
-2.04460692e+00 -1.35921574e+00 5.11779904e-01 2.12378144e+00
1.22945845e+00 4.13610846e-01 -1.61129519e-01 1.70593485e-01
5.37074029e-01 1.27417102e-01 -1.23718226e+00 -9.05569047e-02
-3.44334602e-01 3.05254281e-01 6.81896433e-02 3.83257270e-01
-6.77933395e-01 1.03028226e+00 6.14562750e+00 2.88864285e-01
-1.32596588e+00 1.98351786e-01 6.49736106e-01 -5.68118513e-01
-4.27198946e-01 -2.16894016e-01 -8.01775455e-01 4.18527275e-01
8.76112878e-01 -4.36909385e-02 7.18510151e-01 6.46057844e-01
-3.67502391e-01 -2.12354377e-01 -1.15005672e+00 8.68803501e-01
2.93823723e-02 -1.76362026e+00 4.43719745e-01 -2.82499343e-01
1.01696742e+00 5.77650905e-01 4.34769034e-01 6.12250805e-01
5.78846693e-01 -8.10820222e-01 6.43479288e-01 8.47956181e-01
4.90488231e-01 -6.28018916e-01 3.15288752e-01 5.31365931e-01
-1.01171505e+00 -3.57826114e-01 -3.65919024e-01 -2.71905642e-02
-1.18387200e-01 7.38684356e-01 -6.39669001e-01 4.93191704e-02
9.02678907e-01 1.01413476e+00 -6.14819169e-01 1.13907003e+00
-3.48051637e-01 9.08449292e-01 -2.37465620e-01 2.64777392e-01
2.10998684e-01 1.77050948e-01 3.81187916e-01 1.25396872e+00
2.77809262e-01 2.33496219e-01 -1.92089483e-01 8.76743376e-01
-3.60531121e-01 -5.04887938e-01 -7.02981293e-01 1.78479195e-01
8.85239065e-01 9.79160666e-01 -8.46101463e-01 -4.02854145e-01
-2.85224825e-01 1.09609044e+00 6.61612511e-01 7.95992255e-01
-5.39832592e-01 -1.98013172e-01 3.85231584e-01 9.76121277e-02
2.77733415e-01 -4.56879079e-01 -1.56693652e-01 -1.24739838e+00
-1.75132975e-01 -7.02344656e-01 5.19226670e-01 -8.39889944e-01
-1.23183966e+00 7.59364784e-01 -1.25915125e-01 -7.71188080e-01
6.49814978e-02 -3.90827745e-01 -8.99081230e-01 3.30598176e-01
-1.72602177e+00 -9.64625239e-01 -4.28641737e-01 1.08255982e+00
9.24124777e-01 -1.86136544e-01 6.54687226e-01 2.01821670e-01
-8.06065917e-01 5.99593282e-01 -1.63764626e-01 6.61653355e-02
6.57921195e-01 -1.08274138e+00 5.18894136e-01 8.72190356e-01
5.45620501e-01 1.10980070e+00 2.94878364e-01 -6.32408500e-01
-1.43443489e+00 -1.08400893e+00 3.83632571e-01 -2.88312703e-01
5.50238609e-01 -3.60057086e-01 -1.42177737e+00 9.32871580e-01
3.29782993e-01 3.78258944e-01 6.54947400e-01 8.29785094e-02
-6.32871151e-01 -3.70880455e-01 -8.41871202e-01 7.22404003e-01
1.31091058e+00 -3.63559395e-01 -7.41433680e-01 1.30150123e-02
7.41045058e-01 -2.39246234e-01 -2.59912580e-01 4.01071697e-01
4.30799961e-01 -1.02083898e+00 8.80364716e-01 -3.91256154e-01
9.52085033e-02 -2.83496261e-01 1.07240304e-01 -1.43887246e+00
-3.67671847e-01 -7.17870057e-01 -6.57470584e-01 9.02515471e-01
4.64983016e-01 -9.82880533e-01 9.04925048e-01 2.66714275e-01
-4.44135368e-01 -8.75992954e-01 -8.59742463e-01 -7.74247169e-01
2.78891940e-02 -2.44856656e-01 5.69706321e-01 1.07203472e+00
-1.06236696e-01 1.35159835e-01 -1.25091553e-01 1.91892032e-03
4.79534060e-01 -3.51069458e-02 4.92572457e-01 -1.28162134e+00
-3.90993267e-01 -4.25208777e-01 9.85014662e-02 -1.04230881e+00
9.85005274e-02 -9.98192906e-01 3.58377099e-02 -1.38286221e+00
3.25409710e-01 -4.05624658e-01 -9.64492083e-01 1.12366128e+00
-9.68503281e-02 2.30551168e-01 2.74082199e-02 5.73394835e-01
-8.20981205e-01 6.94120169e-01 1.22458816e+00 -1.64474115e-01
-5.56683362e-01 -6.34675473e-02 -7.60411859e-01 5.33364594e-01
9.41154957e-01 -4.06779706e-01 -8.23394120e-01 -8.02278996e-01
4.25700575e-01 -3.93451780e-01 5.31506419e-01 -1.30142236e+00
8.41872871e-01 -8.18658993e-02 7.12411582e-01 -5.55305779e-01
3.24666709e-01 -3.78282070e-01 1.18295297e-01 6.93563581e-01
-2.54062086e-01 3.60036761e-01 5.52377999e-01 7.90974259e-01
-1.63768336e-01 -4.85901944e-02 6.49315596e-01 -5.78143716e-01
-9.92383897e-01 6.41249895e-01 -5.11684775e-01 7.74297267e-02
8.16219568e-01 -4.33484674e-01 -6.90945745e-01 -5.59003875e-02
-8.05104911e-01 2.63442010e-01 2.82590061e-01 3.12981069e-01
1.12927043e+00 -1.24554324e+00 -4.77488220e-01 3.18716198e-01
-2.36641645e-01 3.77731532e-01 3.64869177e-01 7.57624865e-01
-2.74973750e-01 -3.75194810e-02 -5.33460557e-01 -5.85132360e-01
-6.01271927e-01 5.35664499e-01 4.50028956e-01 6.96733594e-02
-8.59976053e-01 1.22159100e+00 1.75647631e-01 -2.50513792e-01
3.76103044e-01 -1.37002349e-01 3.73727418e-02 1.97347298e-01
7.02621579e-01 4.32192162e-02 1.45309135e-01 1.30062744e-01
-2.38532200e-01 1.78136468e-01 -7.25259542e-01 -2.80093133e-01
1.79143858e+00 1.46736363e-02 -6.81207553e-02 7.56990433e-01
7.18693733e-01 -3.61096591e-01 -1.90928102e+00 -6.93593085e-01
-1.76321156e-02 -1.59194231e-01 -3.84904188e-03 -8.01374316e-01
-1.47778594e+00 8.33000481e-01 3.03686500e-01 -1.95590913e-01
1.34314418e+00 -1.96860105e-01 9.64612126e-01 8.84389997e-01
3.45799029e-01 -1.01485860e+00 9.02148247e-01 8.35838735e-01
9.60443914e-01 -9.19233263e-01 -1.08855240e-01 3.25370520e-01
-5.19752324e-01 1.05408800e+00 9.93049562e-01 -3.09183866e-01
8.94937396e-01 2.96251833e-01 -7.62233138e-02 -2.77513862e-01
-1.29401100e+00 1.56979471e-01 -1.45538673e-01 4.84854162e-01
1.38907492e-01 -4.27983284e-01 2.68030673e-01 4.06314433e-01
-1.02776289e-01 2.03751892e-01 5.20649672e-01 1.19418371e+00
-5.44315338e-01 -7.07884014e-01 8.37122500e-02 3.42052490e-01
-1.77531689e-01 -2.10079417e-01 -3.62445384e-01 4.90130216e-01
4.88857962e-02 4.52523708e-01 1.53713271e-01 -3.09327513e-01
1.01407103e-01 1.84823707e-01 5.12806237e-01 -8.14091682e-01
-4.60726440e-01 -2.99516320e-01 -6.81080997e-01 -4.64175731e-01
-3.49547714e-01 -8.01234603e-01 -1.49182785e+00 -3.45103413e-01
-5.44028133e-02 5.42623997e-02 2.90502608e-01 7.30318427e-01
4.87445563e-01 6.47531807e-01 3.78751606e-01 -8.96921575e-01
-2.49413028e-01 -6.88336372e-01 -3.48007202e-01 1.50769249e-01
6.51878417e-01 -5.88892996e-01 -3.30598146e-01 1.15945712e-01] | [9.850516319274902, 3.4058480262756348] |
53880652-b1b2-4a1a-9552-f381910b661b | exploring-sparse-expert-models-and-beyond | 2105.15082 | null | https://arxiv.org/abs/2105.15082v5 | https://arxiv.org/pdf/2105.15082v5.pdf | M6-T: Exploring Sparse Expert Models and Beyond | Mixture-of-Experts (MoE) models can achieve promising results with outrageous large amount of parameters but constant computation cost, and thus it has become a trend in model scaling. Still it is a mystery how MoE layers bring quality gains by leveraging the parameters with sparse activation. In this work, we investigate several key factors in sparse expert models. We observe that load imbalance may not be a significant problem affecting model quality, contrary to the perspectives of recent studies, while the number of sparsely activated experts $k$ and expert capacity $C$ in top-$k$ routing can significantly make a difference in this context. Furthermore, we take a step forward to propose a simple method called expert prototyping that splits experts into different prototypes and applies $k$ top-$1$ routing. This strategy improves the model quality but maintains constant computational costs, and our further exploration on extremely large-scale models reflects that it is more effective in training larger models. We push the model scale to over $1$ trillion parameters and implement it on solely $480$ NVIDIA V100-32GB GPUs, in comparison with the recent SOTAs on $2048$ TPU cores. The proposed giant model achieves substantial speedup in convergence over the same-size baseline. | ['Hongxia Yang', 'Jingren Zhou', 'Lin Qu', 'Wei Lin', 'Di Zhang', 'Yong Li', 'Jiamang Wang', 'Jie Zhang', 'Ang Wang', 'Xianyan Jia', 'Le Jiang', 'Chang Zhou', 'Rui Men', 'Junyang Lin', 'An Yang'] | 2021-05-31 | null | null | null | null | ['2048'] | ['playing-games'] | [-4.49818820e-01 -5.20078652e-02 -3.02209795e-01 -3.27074707e-01
-6.99313283e-01 -4.07646894e-01 1.63859352e-01 -3.16121787e-01
-6.85606360e-01 5.60270011e-01 -2.04551760e-02 -4.01146591e-01
-5.03373668e-02 -5.42257309e-01 -8.52612376e-01 -5.56571364e-01
2.98615415e-02 7.61346936e-01 3.56031537e-01 -8.69586766e-02
7.87395462e-02 3.28158289e-01 -1.43965948e+00 3.72495890e-01
1.11723912e+00 1.04753947e+00 -4.50437441e-02 6.57955527e-01
-5.36210053e-02 8.39552760e-01 -4.86500651e-01 -6.05203152e-01
5.81816554e-01 7.02661872e-02 -6.61997557e-01 -1.70082062e-01
5.59799492e-01 -4.62566942e-01 -1.99029461e-01 6.56793773e-01
6.38310134e-01 1.36929825e-01 4.70040023e-01 -1.01868904e+00
-2.50655532e-01 7.99360871e-01 -6.90600634e-01 3.32629710e-01
-3.14935297e-01 3.57869089e-01 1.11722338e+00 -1.04769063e+00
2.30384156e-01 1.37111509e+00 7.90382981e-01 4.44458306e-01
-9.95884895e-01 -8.05670321e-01 8.06455493e-01 4.74423878e-02
-1.41934502e+00 -4.70209450e-01 2.32558936e-01 -3.60003322e-01
1.27726316e+00 1.66749209e-01 6.24835789e-01 1.01728737e+00
-7.95998946e-02 9.16317523e-01 1.14653039e+00 -2.50687748e-01
4.01923627e-01 4.35222626e-01 3.63921374e-01 8.44037414e-01
4.96728361e-01 -9.84794721e-02 -7.08322942e-01 -4.95912045e-01
8.51714313e-01 -1.22613654e-01 -7.92096704e-02 2.22433314e-01
-7.17360795e-01 9.30110872e-01 3.29577744e-01 6.67311922e-02
-4.71739352e-01 4.08636183e-01 3.45127612e-01 2.07904115e-01
5.16210258e-01 7.19149172e-01 -6.27769232e-01 -3.01167518e-01
-1.17966592e+00 1.85634568e-01 7.82148123e-01 7.35471904e-01
7.02395856e-01 4.58797216e-01 3.28151286e-02 7.84572363e-01
1.44409388e-01 4.94726449e-01 3.47089976e-01 -1.38548791e+00
3.43262255e-01 7.05488205e-01 3.55704993e-01 -8.01632464e-01
-1.27924427e-01 -7.69504368e-01 -8.70313287e-01 3.00830334e-01
1.97622761e-01 -4.52838957e-01 -8.27616274e-01 1.59625864e+00
5.06134331e-01 2.76718676e-01 -4.39065814e-01 9.49541211e-01
3.27327788e-01 6.79339051e-01 1.12207912e-01 1.77694321e-01
1.31251597e+00 -1.38029218e+00 -6.84577972e-02 -6.90011740e-01
4.85952586e-01 -6.43219471e-01 1.23414314e+00 7.97575951e-01
-1.28312588e+00 -6.92059994e-01 -8.40061426e-01 2.24430516e-01
-1.09165817e-01 4.54740614e-01 1.22936904e+00 8.07973444e-01
-1.25192046e+00 6.46753609e-01 -1.08043349e+00 1.55500891e-02
4.96828258e-01 6.48366630e-01 -2.72849761e-02 -2.32861474e-01
-1.00801182e+00 9.13463831e-01 2.49563679e-01 -5.78698069e-02
-7.77909458e-01 -8.86807442e-01 -3.72878015e-01 3.10905069e-01
5.35535634e-01 -9.94773030e-01 1.13102925e+00 -1.04661751e+00
-1.68191290e+00 4.51258719e-01 -3.28471601e-01 -5.38657725e-01
4.81858164e-01 -4.73911196e-01 -2.91885376e-01 5.24848327e-02
-4.46747005e-01 8.85491788e-01 1.06297922e+00 -1.11055410e+00
-7.98970163e-01 -3.29837680e-01 2.45891929e-01 2.33625948e-01
-7.45690286e-01 1.71699405e-01 -7.13105381e-01 -3.75902981e-01
-6.79834187e-02 -1.09171200e+00 -6.61811292e-01 -7.59279132e-02
-1.47347627e-02 -2.49035269e-01 2.24459589e-01 -4.67005342e-01
1.66711605e+00 -1.79509759e+00 -8.35444853e-02 2.81250566e-01
4.01951075e-01 6.37494504e-01 -7.36900792e-02 5.32448515e-02
2.41463542e-01 2.36516997e-01 -1.35529399e-01 -5.42109847e-01
1.84037179e-01 5.64583391e-02 -2.63025612e-01 5.73232248e-02
1.21612974e-01 9.00903821e-01 -4.96997952e-01 -4.46432531e-01
-1.19247705e-01 6.34877622e-01 -9.84276831e-01 -1.76389311e-02
-2.26477385e-01 -7.29383007e-02 -5.79616010e-01 6.36006355e-01
7.93549836e-01 -8.97678137e-01 2.39677459e-01 1.72386859e-02
6.07689135e-02 2.99570084e-01 -1.44614661e+00 1.44390893e+00
-5.92451453e-01 4.54352230e-01 5.06992340e-01 -9.59211349e-01
5.57474673e-01 2.41002619e-01 1.48208365e-01 -4.74861473e-01
1.34384081e-01 4.65584546e-01 1.81043580e-01 -2.98292488e-02
6.49757206e-01 1.08377993e-01 1.53345615e-01 4.90282536e-01
-1.83869451e-01 5.02695367e-02 2.30538279e-01 3.50740373e-01
1.09761405e+00 -1.76677018e-01 -9.46284682e-02 -4.78282571e-01
5.36988443e-03 1.38510793e-01 5.63400209e-01 9.18283165e-01
6.83427528e-02 1.83044493e-01 3.16203117e-01 -5.72841287e-01
-1.00518823e+00 -6.39340937e-01 5.41231893e-02 1.43193686e+00
-3.28544766e-01 -3.88026118e-01 -8.00537348e-01 -5.29402316e-01
6.59132749e-02 4.79929626e-01 -5.74178159e-01 8.36392567e-02
-8.00562561e-01 -1.21982014e+00 4.99183446e-01 8.62271726e-01
5.45992494e-01 -6.97121561e-01 -8.70098174e-01 2.48398289e-01
2.16153204e-01 -9.56969142e-01 -3.54233801e-01 1.45053580e-01
-1.10647690e+00 -7.61593282e-01 -7.37949908e-01 -3.26532155e-01
7.72148788e-01 4.13496941e-01 1.65037739e+00 1.72027990e-01
-3.09534669e-01 1.58463150e-01 -2.43304089e-01 -3.93219322e-01
2.82770232e-03 3.29916745e-01 1.30460188e-01 -2.27109149e-01
2.92154640e-01 -6.39726758e-01 -1.21273673e+00 2.83970773e-01
-7.23935604e-01 -4.07818630e-02 7.82314420e-01 7.16979384e-01
3.83541554e-01 7.27291927e-02 4.52474922e-01 -1.31842482e+00
5.88339925e-01 -5.67191482e-01 -5.77231884e-01 2.77461410e-01
-1.00392270e+00 7.34496862e-02 9.66237128e-01 -6.19584978e-01
-1.11218035e+00 -8.00093561e-02 -3.00435126e-01 -6.61254168e-01
3.17441434e-01 4.14568067e-01 2.19766274e-01 -7.53483921e-02
9.33417737e-01 -8.59627351e-02 -4.40398365e-01 -4.26703095e-01
1.73501074e-01 5.51644146e-01 1.46369189e-02 -9.74126458e-01
4.68881935e-01 6.67817414e-01 -4.81805325e-01 -5.05973935e-01
-5.98967850e-01 -3.29592735e-01 2.92365104e-02 2.51302361e-01
3.62313956e-01 -1.43621218e+00 -6.99012280e-01 3.67260039e-01
-9.40235913e-01 -6.67239368e-01 -1.95707157e-01 3.62870783e-01
1.98039673e-02 1.64103091e-01 -8.28345716e-01 -9.20258403e-01
-5.85382223e-01 -1.19577706e+00 8.16897094e-01 3.52034092e-01
-1.71057314e-01 -8.64064872e-01 1.50653347e-02 4.26687896e-01
7.84667015e-01 -3.71237129e-01 7.52296150e-01 -3.98565561e-01
-5.92269182e-01 -2.23616660e-01 -2.55428016e-01 5.32844663e-01
-5.33820987e-01 1.29137486e-01 -1.00808263e+00 -6.43728137e-01
8.80225096e-04 -4.84652728e-01 9.92250860e-01 3.96996260e-01
1.24501812e+00 -2.37121254e-01 -4.42023069e-01 5.98221064e-01
1.57738507e+00 1.63359828e-02 5.17836988e-01 -1.37154236e-01
6.48449898e-01 2.96315610e-01 3.22737128e-01 6.71038508e-01
3.68524253e-01 5.61666489e-01 1.37806028e-01 -2.92520493e-01
1.18269905e-01 -5.18050455e-02 3.13177198e-01 1.00576329e+00
-2.49052763e-01 -3.35793465e-01 -9.48294461e-01 5.17577589e-01
-1.85671198e+00 -4.46627140e-01 1.38192862e-01 2.20409679e+00
7.32847154e-01 3.47482085e-01 8.02241564e-02 -2.70980537e-01
4.48664218e-01 1.31488264e-01 -8.38502467e-01 -3.94672453e-01
1.29076485e-02 6.58660889e-01 4.80567068e-01 6.20780289e-01
-6.70512915e-01 9.82328713e-01 6.56869936e+00 1.24861312e+00
-1.29343081e+00 3.77193660e-01 1.09473288e+00 -6.85428381e-01
-2.51009047e-01 7.43717477e-02 -1.25596523e+00 7.59254456e-01
1.30811799e+00 1.64475873e-01 5.51494896e-01 1.31574345e+00
8.30839053e-02 -3.00340265e-01 -9.64020193e-01 9.97381628e-01
1.66480348e-03 -1.30036807e+00 4.64086942e-02 7.56371915e-02
9.14658487e-01 2.43576318e-01 1.78437561e-01 5.85999906e-01
8.09766412e-01 -1.21859241e+00 5.01063108e-01 1.42007366e-01
5.48887551e-01 -6.98768079e-01 4.73700792e-01 5.04619002e-01
-1.11391711e+00 -2.48713389e-01 -7.23395526e-01 -1.38103232e-01
2.04526678e-01 7.38042772e-01 -5.41341007e-01 7.43851438e-02
1.09136641e+00 4.07241546e-02 -5.64054251e-01 7.18064964e-01
4.99825031e-02 8.64710748e-01 -7.18308806e-01 4.28111069e-02
4.26876187e-01 -2.61262413e-02 5.20304814e-02 1.20695710e+00
4.11019057e-01 3.37020516e-01 -4.90642600e-02 5.89245558e-01
-1.83996066e-01 -4.34260108e-02 -7.79862180e-02 -6.22402802e-02
7.22112060e-01 1.24772704e+00 -7.13745117e-01 -7.86168575e-01
-3.96259218e-01 9.83651280e-01 6.29034221e-01 3.10514599e-01
-1.07962930e+00 1.92607846e-02 5.59680343e-01 2.94127971e-01
3.82099003e-01 -1.23628944e-01 -3.07419509e-01 -1.24290264e+00
3.49989273e-02 -1.25841284e+00 3.36502105e-01 -4.72722322e-01
-1.18790841e+00 7.29166687e-01 -2.28794083e-01 -8.33826959e-01
-4.15202565e-02 -7.77316630e-01 -5.49991548e-01 8.26717615e-01
-1.38795042e+00 -6.86894357e-01 -2.48287529e-01 5.02152324e-01
6.05680466e-01 4.79520261e-02 7.10592747e-01 4.71291989e-01
-7.86808491e-01 9.82945681e-01 3.35083067e-01 -3.86822134e-01
6.76784456e-01 -1.03514636e+00 5.50342143e-01 7.10982621e-01
3.29050422e-02 1.00837016e+00 5.83289027e-01 -4.28647935e-01
-1.36761689e+00 -7.61231840e-01 5.58430731e-01 -5.23139298e-01
3.72644097e-01 -2.74745166e-01 -9.08621669e-01 6.43414736e-01
2.17203617e-01 3.80926654e-02 6.74526989e-01 3.21508497e-01
-4.04217720e-01 -9.82275829e-02 -9.23793137e-01 6.38167441e-01
1.07599616e+00 -2.20763788e-01 9.84911174e-02 3.43581229e-01
6.63841009e-01 -3.51317316e-01 -7.93118000e-01 4.20636415e-01
6.42561615e-01 -1.22515237e+00 1.10688996e+00 -5.33291399e-01
4.16988432e-01 4.65275496e-02 -8.90681297e-02 -1.26932383e+00
-4.10556197e-01 -5.20981908e-01 -4.03022230e-01 7.75949717e-01
6.88634932e-01 -8.49009037e-01 1.21396255e+00 1.12253261e+00
-9.70348194e-02 -1.26582515e+00 -7.09009647e-01 -5.52154362e-01
4.76688385e-01 -3.47513467e-01 5.99726677e-01 7.37605095e-01
-2.75693953e-01 2.77265549e-01 -5.04584312e-01 -5.39901154e-03
3.64100218e-01 -8.12673420e-02 7.71466196e-01 -1.13436866e+00
-9.69206214e-01 -5.05221903e-01 4.85314071e-01 -1.49101806e+00
2.21526641e-02 -4.03016537e-01 -2.70272046e-01 -1.20684111e+00
2.98084140e-01 -9.96538401e-01 -2.66096681e-01 2.12619618e-01
-6.48380935e-01 1.65671200e-01 1.44262567e-01 2.69091010e-01
-8.61677945e-01 3.98118138e-01 9.71042633e-01 1.13607429e-01
-1.91075832e-01 -1.78267494e-01 -9.73737001e-01 8.34547937e-01
6.89918220e-01 -3.87355089e-01 -5.37705779e-01 -1.07269871e+00
6.28425360e-01 -4.86194752e-02 -9.15712491e-02 -1.04830134e+00
4.34818864e-01 -1.62189230e-01 3.50047648e-01 -2.48830449e-02
6.00572169e-01 -7.16939688e-01 2.39816099e-01 4.78916675e-01
4.57790941e-02 3.20938468e-01 3.37703705e-01 4.21959728e-01
-1.15863711e-01 -2.56450117e-01 9.06312704e-01 -6.32440507e-01
-6.17960155e-01 3.69763106e-01 -2.16128722e-01 9.57898945e-02
6.81823492e-01 -2.09165365e-01 -4.12846357e-01 -2.00845525e-01
-3.41213495e-01 3.82995367e-01 6.14737391e-01 8.68050288e-03
2.35601395e-01 -8.01967263e-01 -7.22880363e-01 -2.69574188e-02
-3.90148163e-01 2.12155744e-01 7.47757852e-01 7.37416387e-01
-6.64320886e-01 5.34788132e-01 1.71517923e-01 -4.63348538e-01
-9.96050477e-01 2.41050988e-01 3.10536027e-01 -7.93256283e-01
-2.23593727e-01 1.48503447e+00 1.69732675e-01 -3.68286014e-01
2.09685296e-01 -5.45685478e-02 2.22677886e-01 -1.61660224e-01
5.04826903e-01 7.80379474e-01 7.22394660e-02 -1.06314972e-01
-3.55077475e-01 5.23939729e-01 -4.32797402e-01 -1.49731636e-01
1.25813293e+00 2.28365347e-01 1.29686907e-01 7.03560561e-02
7.74638236e-01 1.25657409e-01 -1.66338456e+00 -1.18921503e-01
-3.75199944e-01 -4.15464669e-01 8.12017769e-02 -8.57631207e-01
-1.22170687e+00 1.11621654e+00 5.34034312e-01 -1.06146909e-01
1.00251579e+00 -6.58469573e-02 9.09136176e-01 3.21748167e-01
5.96310616e-01 -1.43843341e+00 1.61501229e-01 4.50988650e-01
4.20599967e-01 -1.43549490e+00 2.02051923e-01 -2.72397995e-01
-7.26128519e-01 4.48880047e-01 1.03358388e+00 -4.16527063e-01
5.99990785e-01 3.83931875e-01 -8.55626315e-02 -2.35297933e-01
-1.21928287e+00 1.23351268e-01 9.22390167e-03 2.89103817e-02
3.06741863e-01 2.42941841e-01 -2.58530289e-01 9.18495595e-01
-7.70998821e-02 3.74422558e-02 3.67144682e-02 8.92762721e-01
-5.62178910e-01 -1.20915806e+00 -2.41925925e-01 7.15037167e-01
-5.93187332e-01 -4.27155018e-01 4.71367352e-02 8.14378142e-01
9.69210267e-02 8.96061122e-01 1.34685278e-01 -3.98082763e-01
2.40446284e-01 9.58496034e-02 3.86796683e-01 -5.15278041e-01
-1.09096670e+00 -6.01958819e-02 -2.49112025e-02 -6.65187478e-01
-2.25993190e-02 -1.31578580e-01 -9.79751647e-01 -8.41397345e-01
-3.35622221e-01 3.58531535e-01 6.67271674e-01 6.47553682e-01
9.54041481e-01 3.39963228e-01 4.93543386e-01 -8.16728830e-01
-7.43972242e-01 -9.75878358e-01 -4.52963859e-01 1.24217771e-01
-3.25656503e-01 -5.70153892e-01 -7.51186788e-01 -3.07779521e-01] | [8.751225471496582, 3.4969873428344727] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.