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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
26c8f01c-f1b1-4f67-929a-c6833ab3484b | lipper-synthesizing-thy-speech-using-multi | 1907.01367 | null | https://arxiv.org/abs/1907.01367v1 | https://arxiv.org/pdf/1907.01367v1.pdf | Lipper: Synthesizing Thy Speech using Multi-View Lipreading | Lipreading has a lot of potential applications such as in the domain of surveillance and video conferencing. Despite this, most of the work in building lipreading systems has been limited to classifying silent videos into classes representing text phrases. However, there are multiple problems associated with making lipreading a text-based classification task like its dependence on a particular language and vocabulary mapping. Thus, in this paper we propose a multi-view lipreading to audio system, namely Lipper, which models it as a regression task. The model takes silent videos as input and produces speech as the output. With multi-view silent videos, we observe an improvement over single-view speech reconstruction results. We show this by presenting an exhaustive set of experiments for speaker-dependent, out-of-vocabulary and speaker-independent settings. Further, we compare the delay values of Lipper with other speechreading systems in order to show the real-time nature of audio produced. We also perform a user study for the audios produced in order to understand the level of comprehensibility of audios produced using Lipper. | ['Yifang Yin', 'Rajiv Ratn Shah', 'Khwaja Mohd. Salik', 'Roger Zimmermann', 'Yaman Kumar', 'Rohit Jain'] | 2019-06-28 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 1.95237204e-01 -6.34360081e-03 -2.48105645e-01 -3.67476434e-01
-9.56067085e-01 -5.08076310e-01 6.89875782e-01 -1.37264729e-01
-2.06504852e-01 3.60849798e-01 4.68501300e-01 -3.83073807e-01
2.24955037e-01 -8.68411139e-02 -5.57259321e-01 -4.74921972e-01
2.97205061e-01 1.07864022e-01 4.23292011e-01 -9.48893204e-02
4.43695128e-01 2.48837665e-01 -2.05341411e+00 6.42862797e-01
2.90914118e-01 6.56218767e-01 6.01174712e-01 1.19788098e+00
-5.70980646e-02 6.81485415e-01 -7.75455236e-01 -3.35642993e-01
-3.73907052e-02 -4.00091290e-01 -6.41871750e-01 2.96872139e-01
4.19097066e-01 -4.33532268e-01 -2.32746452e-01 7.34893382e-01
8.99657488e-01 -2.10907478e-02 5.86826324e-01 -1.54164386e+00
-1.40251100e-01 2.91360050e-01 -1.15500532e-01 9.40809846e-02
1.10547674e+00 1.07010923e-01 6.07155144e-01 -9.09279823e-01
4.33972478e-01 1.37665868e+00 4.60207701e-01 7.19522715e-01
-9.82206464e-01 -5.32159269e-01 -1.28232971e-01 3.47475290e-01
-1.37437487e+00 -1.45003474e+00 5.66094398e-01 -4.12459075e-01
9.30844247e-01 3.73072237e-01 3.42690527e-01 1.20461226e+00
-2.58530378e-02 8.69107306e-01 8.17190111e-01 -7.58748770e-01
6.38305545e-02 5.99405646e-01 -1.41875207e-01 2.48448238e-01
-3.52784991e-01 -1.31282613e-01 -9.46230471e-01 1.14653617e-01
3.57844383e-01 -3.66538882e-01 -6.79163337e-01 -4.32912946e-01
-1.22995269e+00 7.05460012e-01 -5.80236137e-01 6.76654279e-02
-3.71823125e-02 -6.69885352e-02 5.85242271e-01 5.16577721e-01
4.61944371e-01 -1.49806261e-01 -1.70203224e-01 -4.66584504e-01
-1.27872205e+00 -1.03455670e-01 9.62430656e-01 1.14278793e+00
2.23569229e-01 -2.33247876e-01 -1.87055618e-02 1.28063047e+00
5.25800407e-01 5.46444833e-01 7.16670513e-01 -9.60874081e-01
6.73932493e-01 -1.75703932e-02 7.32079614e-03 -5.65446734e-01
-1.73856303e-01 4.43026960e-01 -8.48572776e-02 2.25696996e-01
3.67498785e-01 -1.85056198e-02 -6.37684047e-01 1.45872676e+00
-1.21424168e-01 -4.52724919e-02 2.38390878e-01 6.80790842e-01
9.28789854e-01 7.46771395e-01 -1.66846529e-01 -5.61953843e-01
1.29543054e+00 -1.11909223e+00 -8.76178205e-01 1.55172339e-02
3.45319629e-01 -1.22420573e+00 1.17739367e+00 5.04403532e-01
-1.37855518e+00 -5.61630666e-01 -9.02564883e-01 -2.48158779e-02
-2.25053176e-01 3.15062284e-01 -1.37400836e-01 9.84862983e-01
-1.41324043e+00 9.09534991e-02 -4.46443379e-01 -8.94281507e-01
-2.74858445e-01 2.39697114e-01 -4.60461676e-01 8.10894743e-02
-1.02369618e+00 8.41773033e-01 -1.04655713e-01 -4.10361290e-01
-8.06692719e-01 -3.33080471e-01 -8.80524099e-01 1.66549683e-01
2.43034095e-01 -3.47273797e-01 1.54301608e+00 -8.20236981e-01
-1.82427287e+00 9.75137532e-01 -7.07906067e-01 -1.87570810e-01
6.96203530e-01 8.52126554e-02 -6.66112363e-01 4.99640822e-01
1.03196716e-02 7.96801448e-01 1.15046322e+00 -1.19328272e+00
-5.81654549e-01 -1.71570048e-01 1.28420919e-01 4.15276527e-01
-2.67391890e-01 3.33519012e-01 -6.68858826e-01 -3.81754339e-01
-1.14259191e-01 -7.63750613e-01 6.97898507e-01 5.98359033e-02
-1.80975541e-01 -2.69857407e-01 7.92613328e-01 -7.09277868e-01
1.12620151e+00 -2.40141106e+00 -5.73489442e-02 -1.91010743e-01
-1.12136357e-01 1.45778492e-01 2.80166585e-02 8.97021234e-01
-7.12250546e-02 2.85416543e-01 3.52570623e-01 -6.45142555e-01
-3.35563272e-02 -1.09506294e-01 -2.70136744e-01 4.20194805e-01
-2.10565552e-01 3.19028735e-01 -5.36955237e-01 -7.77499378e-01
3.81764948e-01 5.23862243e-01 -4.45246458e-01 2.95911670e-01
2.51435488e-02 1.93643086e-02 8.44634995e-02 5.90701997e-01
5.09034097e-01 1.33907825e-01 -1.44186422e-01 -3.35790939e-03
-3.27508509e-01 2.55533189e-01 -9.57847953e-01 1.58464181e+00
-7.57997990e-01 1.29728842e+00 4.53737944e-01 -7.36590505e-01
6.85777247e-01 9.05349314e-01 2.40937829e-01 -4.97584552e-01
-1.06478870e-01 6.61873892e-02 -3.03873986e-01 -1.11616611e+00
7.79516101e-01 -1.36238441e-01 4.42005247e-01 5.35863757e-01
-6.07631616e-02 -1.64939076e-01 1.96045890e-01 2.22834855e-01
5.53954005e-01 -1.14485353e-01 1.55743152e-01 -2.70524397e-02
7.67401874e-01 -4.55593437e-01 -3.23023975e-01 5.07052779e-01
-4.41791624e-01 9.76327837e-01 4.73103136e-01 2.90744454e-01
-8.84742618e-01 -1.03041029e+00 -7.51597211e-02 1.28400326e+00
2.11319968e-01 -5.00202656e-01 -1.01206565e+00 -2.06429362e-01
-3.51403028e-01 7.68678546e-01 -8.13161880e-02 1.21700324e-01
-1.54578403e-01 3.21389213e-02 8.77438366e-01 1.82877943e-01
2.31234759e-01 -1.09842014e+00 -6.94016993e-01 -2.43523028e-02
-6.89261436e-01 -1.33376086e+00 -7.44284093e-01 -2.42520615e-01
-5.28434694e-01 -9.09583271e-01 -1.13741207e+00 -8.45652819e-01
2.08717898e-01 6.15980685e-01 7.36881137e-01 -1.34120241e-01
9.41379517e-02 8.81955385e-01 -4.32838321e-01 -4.07181561e-01
-1.08432174e+00 -1.07454590e-01 2.49966741e-01 -6.30485080e-03
3.67500693e-01 -7.42398724e-02 -3.52483660e-01 7.01303542e-01
-7.97020972e-01 1.32021353e-01 4.02476579e-01 5.90052664e-01
-1.25022516e-01 -2.22563058e-01 5.58364689e-01 -6.26037642e-02
1.06937897e+00 -3.19031924e-01 -2.52526909e-01 4.69179064e-01
-3.13975751e-01 -2.91987211e-01 2.14216679e-01 -6.74875319e-01
-8.79103959e-01 -1.21486925e-01 -3.99489045e-01 -4.49748188e-01
-3.69025499e-01 8.67768228e-02 -2.16026023e-01 2.02688381e-01
3.55800897e-01 5.96757710e-01 5.81848085e-01 -3.89151573e-01
4.84118052e-02 1.59268367e+00 2.65977263e-01 -1.32928267e-01
6.25969619e-02 1.73690259e-01 -6.04018450e-01 -1.75604427e+00
1.94535013e-02 -8.54632378e-01 -2.83508927e-01 -7.21023023e-01
7.68426478e-01 -9.03189182e-01 -1.06651056e+00 6.78191841e-01
-1.21014965e+00 -3.74616295e-01 1.31509796e-01 5.27609766e-01
-1.04273796e+00 6.20862722e-01 -2.94010520e-01 -1.19032192e+00
1.70219630e-01 -1.58573020e+00 1.31257415e+00 4.76924814e-02
-2.53018469e-01 -8.24991226e-01 -7.68051073e-02 8.16436768e-01
3.31456423e-01 -6.84479952e-01 6.05111301e-01 -5.28932869e-01
-2.82442212e-01 1.33225527e-02 -1.42661929e-01 4.65483248e-01
1.25970438e-01 1.15193076e-01 -1.40385377e+00 -4.16507483e-01
-2.80169714e-02 -4.16901827e-01 5.72756529e-01 5.18142045e-01
8.13635290e-01 -1.33988827e-01 -3.13537240e-01 1.48884743e-01
1.05424178e+00 2.88615823e-01 7.43336558e-01 5.54843023e-02
5.03258407e-03 9.06357765e-01 5.23725808e-01 3.60542625e-01
2.99621850e-01 1.09641433e+00 2.05877259e-01 -4.17119376e-02
-3.74634981e-01 -4.24809635e-01 8.21346045e-01 7.25917816e-01
1.48552328e-01 -7.44233251e-01 -7.65767276e-01 6.26384497e-01
-1.36167049e+00 -1.34081018e+00 1.85269251e-01 2.35136008e+00
4.94990826e-01 4.85830717e-02 5.50411522e-01 5.45737565e-01
1.07732272e+00 8.84602517e-02 -7.55449757e-02 -6.28840744e-01
2.42407709e-01 -3.18694383e-01 3.26234192e-01 8.34023893e-01
-7.17397451e-01 7.74794638e-01 6.90337658e+00 6.96520805e-01
-1.57650948e+00 1.75487027e-02 2.52195746e-01 -1.22758783e-01
1.13032773e-01 -4.20843512e-01 -7.05680549e-01 6.51476204e-01
1.15967786e+00 -1.11291893e-01 5.10563612e-01 5.16517997e-01
8.60157013e-01 -5.29271960e-01 -1.25598967e+00 1.26564229e+00
6.50019050e-01 -9.87638295e-01 4.58453894e-02 1.21857645e-02
-6.73218444e-02 7.61454701e-02 1.18187014e-02 7.16548041e-02
-5.30707598e-01 -9.09144878e-01 9.71752405e-01 2.33128950e-01
1.19901562e+00 -3.15576077e-01 4.56742972e-01 5.84901571e-01
-1.00658512e+00 5.69778569e-02 -1.22300535e-01 1.22977577e-01
3.15693468e-01 -2.42231220e-01 -1.32137024e+00 2.14957789e-01
6.20534003e-01 3.34583730e-01 -3.23785096e-01 1.02027655e+00
2.36948818e-01 5.85588694e-01 -1.94580525e-01 -2.21760973e-01
-1.73851550e-01 2.94099540e-01 5.70573986e-01 1.41750443e+00
4.33309674e-01 -9.47154686e-02 -1.86789244e-01 3.62705261e-01
1.10895097e-01 2.82630771e-01 -7.81589031e-01 3.13291475e-02
3.72249246e-01 7.47247815e-01 -5.52418768e-01 -2.54756898e-01
-5.77768385e-01 9.13082063e-01 -4.80088800e-01 4.59433198e-01
-7.08577573e-01 -4.43124145e-01 5.48457086e-01 4.80195791e-01
1.66430399e-01 -8.88256878e-02 1.53095022e-01 -1.11180067e+00
6.25311658e-02 -9.20240402e-01 -1.84292138e-01 -1.39686179e+00
-5.55673301e-01 4.07011211e-01 2.77677804e-01 -1.29712415e+00
-5.95418096e-01 -4.14138556e-01 -4.86007124e-01 8.70931804e-01
-1.48215628e+00 -8.36869180e-01 -3.43670279e-01 6.71867490e-01
1.37103105e+00 -4.04969841e-01 7.07303524e-01 3.54954064e-01
-1.11442789e-01 6.36246920e-01 3.78857590e-02 -3.01306155e-02
1.18978393e+00 -7.04304874e-01 1.37414232e-01 4.97937351e-01
9.22902599e-02 4.24294263e-01 1.02769589e+00 -2.03480929e-01
-1.05949402e+00 -3.86584699e-01 1.23800409e+00 -2.71707118e-01
4.66298968e-01 -4.07108575e-01 -5.52712619e-01 3.33143890e-01
5.67968547e-01 -5.16606510e-01 7.03117311e-01 -2.49470204e-01
-5.76014183e-02 -5.57699939e-03 -1.13358247e+00 4.67321873e-01
8.07241499e-01 -8.34810734e-01 -5.70984006e-01 1.67482540e-01
4.95973557e-01 -1.32611185e-01 -5.07829666e-01 -8.01419616e-02
9.59178209e-01 -1.32668340e+00 9.09541190e-01 -1.86108321e-01
3.66701573e-01 8.75640735e-02 -1.89523429e-01 -1.17203927e+00
4.13026392e-01 -6.85960412e-01 4.70575899e-01 1.37945056e+00
4.36301023e-01 -5.90431273e-01 6.63075626e-01 2.12380633e-01
7.92181790e-02 -3.39001864e-01 -9.46960926e-01 -7.80685306e-01
-2.00847328e-01 -5.31937778e-01 2.03911528e-01 5.23090303e-01
6.12378418e-01 4.00822848e-01 -4.74459410e-01 3.51613015e-02
2.67022371e-01 -3.63503873e-01 9.12305176e-01 -9.90535796e-01
-1.25339568e-01 -2.90362090e-01 -5.07877827e-01 -1.42123222e+00
1.62559688e-01 -5.17718077e-01 2.74073362e-01 -1.36946499e+00
6.89079240e-02 -7.11063743e-02 3.30562860e-01 3.28611545e-02
3.45945120e-01 6.53778613e-02 4.11611378e-01 3.60955447e-01
-2.92944223e-01 1.26855910e-01 8.86787295e-01 -2.22237468e-01
-2.17966512e-01 4.96589839e-01 -2.02729180e-01 5.47998130e-01
7.72370875e-01 -1.24142624e-01 -8.00921440e-01 -3.40182513e-01
-1.28604725e-01 7.42241979e-01 3.40357572e-01 -8.93271446e-01
4.84040767e-01 6.42115623e-02 -2.67239809e-01 -7.07644165e-01
9.48437274e-01 -7.27292418e-01 -1.57567024e-01 1.77110851e-01
-6.09067321e-01 1.00958645e-01 1.24037422e-01 3.85703892e-01
-4.11803961e-01 -4.85650748e-01 7.12147593e-01 -2.41098227e-03
-6.10071480e-01 -3.10549229e-01 -1.01471663e+00 1.26616642e-01
9.07501221e-01 -7.16730475e-01 -2.73308128e-01 -1.15465546e+00
-8.76555145e-01 3.68145406e-02 6.91006184e-01 5.76858103e-01
8.15480530e-01 -6.71176493e-01 -5.91348469e-01 5.20963609e-01
2.26784170e-01 -6.33279622e-01 -1.98835470e-02 7.88025200e-01
-4.99569744e-01 7.26472735e-01 -2.22821072e-01 -9.95105624e-01
-2.01592302e+00 5.21026492e-01 2.81404525e-01 4.42378134e-01
-3.36990416e-01 4.86134946e-01 7.20567927e-02 4.22846116e-02
7.37245560e-01 -1.62691355e-01 -3.30561906e-01 1.94313541e-01
4.97446477e-01 4.59516853e-01 1.14655428e-01 -7.80546010e-01
-3.09482813e-01 5.00281572e-01 1.70998946e-01 -7.31254101e-01
7.74375439e-01 -6.17434144e-01 3.34301054e-01 6.04949117e-01
1.23990798e+00 3.18630785e-01 -9.05860007e-01 3.35394531e-01
-2.10346103e-01 -3.76541167e-01 -5.94406202e-02 -5.70299327e-01
-5.63448846e-01 1.31376803e+00 5.72086692e-01 5.83632708e-01
9.14784312e-01 1.40603796e-01 3.80060822e-01 2.80213267e-01
1.94402248e-01 -1.15198350e+00 5.51870354e-02 2.35624179e-01
9.38526094e-01 -1.28505552e+00 -3.55138183e-01 -4.88635540e-01
-6.86259329e-01 1.26950836e+00 5.36371768e-02 6.57028913e-01
4.52087879e-01 3.04260194e-01 4.31654811e-01 2.87408769e-01
-8.23700488e-01 -1.01519212e-01 -1.34538010e-01 8.82508516e-01
5.62078834e-01 -1.56817660e-01 -1.68400973e-01 -1.45927891e-02
-2.26086557e-01 2.95252800e-01 9.29241538e-01 6.23537660e-01
-5.55384517e-01 -1.00989628e+00 -5.95369637e-01 3.93324085e-02
-4.80194360e-01 -9.17573571e-02 -4.85350937e-01 8.21638763e-01
-2.90165395e-01 1.51100504e+00 -8.14388692e-02 -2.90953845e-01
2.18152508e-01 5.03230631e-01 3.40861499e-01 -5.29414713e-01
-2.68454939e-01 2.58011550e-01 3.55371624e-01 -3.13578993e-01
-5.81975877e-01 -8.87818098e-01 -7.84716189e-01 -3.52159381e-01
-4.81637061e-01 2.41334930e-01 1.18596852e+00 7.60355413e-01
1.21865124e-01 1.76150143e-01 5.12135327e-01 -8.93328965e-01
-6.76142037e-01 -1.06573188e+00 -3.61860812e-01 2.89316624e-01
6.40422344e-01 -4.55589116e-01 -7.15729594e-01 4.29562777e-01] | [14.326654434204102, 5.030131816864014] |
7a700a9d-e211-4afd-a095-3465655f5cee | sonicverse-a-multisensory-simulation-platform | 2306.00923 | null | https://arxiv.org/abs/2306.00923v1 | https://arxiv.org/pdf/2306.00923v1.pdf | Sonicverse: A Multisensory Simulation Platform for Embodied Household Agents that See and Hear | Developing embodied agents in simulation has been a key research topic in recent years. Exciting new tasks, algorithms, and benchmarks have been developed in various simulators. However, most of them assume deaf agents in silent environments, while we humans perceive the world with multiple senses. We introduce Sonicverse, a multisensory simulation platform with integrated audio-visual simulation for training household agents that can both see and hear. Sonicverse models realistic continuous audio rendering in 3D environments in real-time. Together with a new audio-visual VR interface that allows humans to interact with agents with audio, Sonicverse enables a series of embodied AI tasks that need audio-visual perception. For semantic audio-visual navigation in particular, we also propose a new multi-task learning model that achieves state-of-the-art performance. In addition, we demonstrate Sonicverse's realism via sim-to-real transfer, which has not been achieved by other simulators: an agent trained in Sonicverse can successfully perform audio-visual navigation in real-world environments. Sonicverse is available at: https://github.com/StanfordVL/Sonicverse. | ['Jiajun Wu', 'Li Fei-Fei', 'Silvio Savarese', 'Fei Xia', 'Chengshu Li', 'Zhuzhu Wang', 'Gokul Dharan', 'Hao Li', 'Ruohan Gao'] | 2023-06-01 | null | null | null | null | ['visual-navigation'] | ['robots'] | [-3.57358873e-01 -1.63489267e-01 7.33435690e-01 2.36605220e-02
-7.52432942e-01 -6.44896090e-01 5.92966855e-01 -2.02799767e-01
-6.49347067e-01 4.02681172e-01 9.69421864e-02 -9.31866989e-02
3.10259908e-01 -8.63406718e-01 -8.05238008e-01 -5.20665646e-01
-3.57147694e-01 6.19912565e-01 5.13090134e-01 -9.92098451e-01
-2.29081735e-01 1.98835284e-01 -2.42309141e+00 4.60555494e-01
7.00250804e-01 6.07540250e-01 9.72841978e-01 1.28009522e+00
1.23909444e-01 1.03823268e+00 -8.71657610e-01 1.32133022e-01
4.93738502e-02 -9.97624248e-02 -4.60884124e-01 -7.02803910e-01
1.80095822e-01 -4.12728965e-01 -3.31624091e-01 8.02883148e-01
1.03544235e+00 4.48088557e-01 4.93341029e-01 -1.94653511e+00
-5.38619518e-01 3.87503475e-01 -1.80405490e-02 -1.92331538e-01
1.24582231e+00 4.44896579e-01 7.03408837e-01 -5.90417743e-01
6.96218491e-01 1.48436940e+00 5.41548193e-01 7.86726117e-01
-7.40332127e-01 -1.04573894e+00 1.07924573e-01 4.74619210e-01
-1.13156545e+00 -3.97716254e-01 4.07486498e-01 -3.92133296e-01
1.22435355e+00 4.52790111e-01 1.06654179e+00 1.61069930e+00
1.00336842e-01 9.27311242e-01 9.88051891e-01 -2.02681333e-01
4.73957628e-01 -4.01393846e-02 -3.30855399e-01 7.92132199e-01
-5.56207657e-01 6.84175551e-01 -8.13406289e-01 6.20313212e-02
8.78256679e-01 -4.54064518e-01 -4.38377678e-01 -4.20145333e-01
-1.45390332e+00 5.54346800e-01 4.51972872e-01 1.11968685e-02
-4.04324442e-01 6.88553691e-01 5.50149739e-01 6.48429513e-01
-3.97417173e-02 2.85504103e-01 -2.20198572e-01 -4.04174268e-01
-1.60292462e-02 5.95562816e-01 7.56623566e-01 8.63103628e-01
5.42170070e-02 4.12294537e-01 3.67926121e-01 7.27272213e-01
5.96013188e-01 9.71872628e-01 2.87471861e-01 -1.52950144e+00
1.48572147e-01 -2.61326414e-02 3.61654997e-01 -5.89630008e-01
-5.84154904e-01 -1.12028264e-01 -6.25948012e-01 1.14524376e+00
2.43485779e-01 -4.17500362e-02 -7.16124296e-01 1.75914907e+00
5.70726812e-01 8.05393219e-01 4.93539721e-01 1.24558520e+00
1.51781988e+00 1.01276302e+00 2.16644630e-01 2.71695793e-01
1.52926731e+00 -1.21174026e+00 -8.18756521e-01 -2.61310309e-01
5.12324452e-01 -4.13002878e-01 1.65660512e+00 7.53268361e-01
-9.39718544e-01 -5.31476796e-01 -1.21865034e+00 1.55961029e-02
-5.81375062e-01 -4.69399601e-01 8.86193931e-01 4.84668016e-01
-1.29504800e+00 7.05831870e-02 -1.00898707e+00 -2.52572924e-01
-6.15391992e-02 1.24899201e-01 -4.85047996e-01 2.97746789e-02
-1.30373609e+00 1.00436437e+00 -3.74370441e-02 -2.69378722e-01
-1.62693572e+00 -7.25700080e-01 -1.10830939e+00 -2.36802697e-01
2.92410553e-01 -1.09554076e+00 1.98606861e+00 -5.70895553e-01
-2.06420350e+00 6.47861242e-01 1.38558254e-01 4.60387059e-02
8.25512350e-01 -3.20230395e-01 -5.53149760e-01 1.38393506e-01
8.06692392e-02 6.15175366e-01 2.95831829e-01 -1.96533418e+00
-6.37771428e-01 -1.22226596e-01 3.45890701e-01 5.38203001e-01
1.95504189e-01 2.29013160e-01 -1.83679789e-01 -3.50233018e-01
-2.76819527e-01 -8.01877379e-01 -3.52750450e-01 3.98084641e-01
1.08623222e-01 -1.13268748e-01 5.52541912e-01 -4.46069747e-01
2.58220792e-01 -2.19802427e+00 3.48510951e-01 -2.37526283e-01
2.82331526e-01 -1.17833298e-02 -4.23260599e-01 6.85143650e-01
3.66391927e-01 -3.29625607e-01 2.40204439e-01 -8.12388182e-01
4.25544322e-01 3.94060202e-02 -2.98235863e-01 2.08603948e-01
-6.19009018e-01 6.30377054e-01 -1.33366895e+00 -3.80297989e-01
5.05256534e-01 1.01279449e+00 -6.00624561e-01 3.26249421e-01
-2.97645986e-01 6.69950485e-01 -3.11720490e-01 5.08707762e-01
6.07323706e-01 2.24911999e-02 -5.45043536e-02 3.19674939e-01
-3.03894043e-01 1.20385490e-01 -1.36779320e+00 2.32074237e+00
-1.00627124e+00 5.67329943e-01 5.98790705e-01 -3.29321831e-01
3.93970370e-01 6.16912305e-01 5.42692021e-02 -1.21171403e+00
2.08558559e-01 1.90366387e-01 -3.61937642e-01 -7.43392467e-01
6.06036365e-01 -7.66761675e-02 -2.80924678e-01 2.38566831e-01
-6.71142265e-02 -7.30752289e-01 -1.33778781e-01 1.17370903e-01
1.15886283e+00 4.13933039e-01 5.71533553e-02 2.40404144e-01
3.96176651e-02 5.08501306e-02 2.24815771e-01 7.14525521e-01
-1.22622348e-01 5.27853549e-01 -3.07461888e-01 -8.23363513e-02
-6.55372918e-01 -1.86962306e+00 4.31388199e-01 1.83034623e+00
7.02659309e-01 -3.71362716e-01 -4.42153066e-01 1.20389566e-01
9.94604360e-03 9.36085880e-01 -4.70473140e-01 9.55008492e-02
-3.18617195e-01 3.27167213e-02 8.06305408e-01 5.79284787e-01
4.59680796e-01 -1.68899679e+00 -1.40826762e+00 1.00974306e-01
-3.83144200e-01 -1.15771973e+00 9.26701445e-03 2.73580253e-02
-2.10148603e-01 -8.51104259e-01 -7.06737816e-01 -1.11814678e+00
-1.47936046e-01 4.81645137e-01 1.07674336e+00 -6.31065145e-02
-2.05783024e-01 6.82440221e-01 -5.42800546e-01 -6.01823688e-01
-5.97739458e-01 -4.37887698e-01 1.96147040e-01 -8.79016459e-01
-3.98123801e-01 -7.29987860e-01 -4.38041121e-01 4.69754100e-01
-5.60325265e-01 3.68182719e-01 1.16799690e-01 3.46083403e-01
1.62082329e-01 -4.60219741e-01 6.65811121e-01 -1.16938353e-01
7.26659477e-01 -2.94343472e-01 -5.27357697e-01 3.72639000e-02
4.54146206e-01 -4.65735823e-01 5.14263988e-01 -8.33285391e-01
-1.06657386e+00 -1.98835030e-01 -5.31462967e-01 -4.10347402e-01
-4.42986161e-01 4.04769778e-01 -1.54153258e-01 -8.91121570e-03
8.13693166e-01 8.89637098e-02 8.61511230e-02 -1.08781211e-01
4.68627393e-01 6.91929221e-01 9.15951431e-01 -5.15990019e-01
5.47531188e-01 5.98879158e-01 -4.88636643e-01 -8.90466750e-01
9.07980651e-02 -3.20469558e-01 1.07292950e-01 -7.18904436e-01
9.17974591e-01 -1.26952040e+00 -1.77236652e+00 7.99872875e-01
-1.24813688e+00 -1.27305675e+00 -8.81690755e-02 7.33176649e-01
-1.04724371e+00 -1.04186177e-01 -7.49010980e-01 -9.28130984e-01
-6.89148251e-03 -1.40535319e+00 1.31783366e+00 2.41787881e-01
-4.07907158e-01 -5.69886506e-01 4.80571270e-01 2.21069202e-01
2.93539524e-01 2.43967071e-01 3.24224591e-01 -8.81608129e-02
-5.50097287e-01 2.28041798e-01 7.21779168e-02 -2.52712697e-01
-2.32882470e-01 -2.22051069e-01 -1.22281492e+00 -2.66905576e-01
-5.17721713e-01 -8.27324390e-01 6.64849222e-01 2.50230670e-01
6.45305812e-01 4.40251142e-01 -4.18391317e-01 7.80025423e-01
1.07767749e+00 7.79642463e-01 6.42073929e-01 5.80789030e-01
3.83426011e-01 3.19486141e-01 8.04788172e-01 4.37054127e-01
9.34233248e-01 7.33004212e-01 8.18313479e-01 -1.24417029e-01
-8.33378360e-02 -2.45228127e-01 6.55326188e-01 9.66375649e-01
-3.37702245e-01 -6.01383746e-01 -9.59464967e-01 2.88613856e-01
-1.76801956e+00 -1.05695367e+00 4.70767878e-02 1.73734760e+00
3.55970800e-01 -1.66700944e-01 1.88974634e-01 9.68785137e-02
4.26206082e-01 1.43544465e-01 -4.23285604e-01 -3.41524988e-01
-1.14018843e-01 3.45143914e-01 -8.45888257e-02 6.18991256e-01
-8.26922238e-01 1.14058936e+00 6.18112993e+00 4.06264454e-01
-1.05914307e+00 4.01658028e-01 -4.98076797e-01 -2.66448885e-01
-4.13494647e-01 -5.91747463e-01 2.51661032e-03 9.16739553e-02
6.28034174e-01 -1.44054264e-01 7.40480900e-01 8.86882067e-01
2.64827579e-01 -1.30733609e-01 -9.73780632e-01 1.22198331e+00
4.31360165e-03 -1.10698771e+00 -2.57121801e-01 -2.12149411e-01
2.38041982e-01 3.36416036e-01 3.06487024e-01 7.18767822e-01
9.92768645e-01 -1.28602529e+00 1.23115826e+00 3.80421698e-01
8.99451435e-01 -7.54608154e-01 4.71571088e-01 4.41396266e-01
-1.49784660e+00 -5.99022694e-02 5.18109882e-03 -2.95545578e-01
5.52148163e-01 -2.91537911e-01 -5.22830606e-01 3.58631581e-01
1.05468333e+00 2.57987350e-01 7.62193128e-02 1.33526897e+00
-2.74218678e-01 -7.49415308e-02 -3.22920263e-01 -4.54504579e-01
-5.88573441e-02 -4.50931080e-02 5.67468047e-01 1.06701434e+00
6.09405935e-01 3.50944459e-01 3.03955555e-01 4.51858550e-01
2.41461292e-01 2.16841176e-02 -7.81662345e-01 2.89802641e-01
7.02397346e-01 8.15814018e-01 -5.21945894e-01 -1.74020275e-01
-1.22080445e-01 1.13713455e+00 1.29450008e-01 4.94367480e-01
-1.15228283e+00 -5.40755570e-01 1.10896409e+00 -2.27863893e-01
-1.01044573e-01 -4.97171760e-01 1.73543066e-01 -7.78384805e-01
-1.72078088e-01 -1.06808305e+00 1.11834884e-01 -1.57223976e+00
-7.56271124e-01 9.70614731e-01 -1.33869693e-01 -1.20660877e+00
-3.76383245e-01 -4.93141234e-01 -4.68425810e-01 4.44153756e-01
-1.29087138e+00 -1.29207039e+00 -8.98617148e-01 6.77317560e-01
6.98232532e-01 -2.95438439e-01 1.34681070e+00 3.11525285e-01
1.21925719e-01 3.02624226e-01 -8.00918788e-02 -2.47151271e-01
7.98881352e-01 -1.16643417e+00 7.68224955e-01 4.70575020e-02
-5.16268238e-02 8.47534165e-02 1.21351349e+00 -3.75518680e-01
-1.44989550e+00 -5.33384979e-01 1.01149641e-01 -4.09087628e-01
5.88229418e-01 -6.70209289e-01 -6.95200086e-01 7.69723713e-01
4.29094970e-01 -1.29883558e-01 5.73850811e-01 -7.08022341e-02
-4.98769909e-01 3.33115011e-01 -1.21310306e+00 9.29637492e-01
1.75516343e+00 -6.94477022e-01 -5.44237554e-01 1.40759394e-01
1.18095529e+00 -8.08554351e-01 -3.41985643e-01 1.48633704e-01
8.97995353e-01 -1.12121427e+00 1.25076580e+00 -2.93096423e-01
3.10485899e-01 -4.73004967e-01 -4.61654872e-01 -2.07854176e+00
2.30545960e-02 -4.60774422e-01 1.57788187e-01 6.15526974e-01
3.97701263e-02 -9.57827866e-01 4.16650742e-01 -4.61501867e-01
-5.55713475e-01 -1.68338940e-02 -9.48691368e-01 -8.32781613e-01
-1.00226663e-01 -1.07785499e+00 6.91623092e-01 7.95029163e-01
2.63062328e-01 1.19804911e-01 -1.76932067e-01 7.61868417e-01
6.01868868e-01 -2.23887160e-01 9.92119968e-01 -1.33122277e+00
-4.53238755e-01 -4.10474151e-01 -4.80549812e-01 -1.03067291e+00
5.33682287e-01 -7.86678076e-01 4.07679379e-01 -1.98366904e+00
-3.31189066e-01 -4.40330863e-01 7.67844841e-02 2.03007162e-01
3.67326319e-01 1.89226821e-01 5.50556898e-01 -2.51715451e-01
-7.46040165e-01 9.20338809e-01 1.61512125e+00 -6.82134926e-02
-3.61101031e-01 -1.76874265e-01 -2.85205215e-01 8.15511942e-01
8.33600581e-01 -8.86626318e-02 -6.21531546e-01 -6.72038794e-01
4.70636368e-01 5.65944254e-01 9.66367126e-01 -1.37970448e+00
4.05368358e-01 -1.66489214e-01 -1.00484341e-01 -4.70971376e-01
1.22438979e+00 -7.13667572e-01 4.25502837e-01 4.77239102e-01
-1.39339909e-01 4.30751652e-01 4.59397763e-01 3.21197778e-01
6.54595047e-02 2.34739438e-01 3.55754524e-01 -2.08414793e-01
-1.10587239e+00 -1.52233928e-01 -9.24793601e-01 6.62300587e-02
1.10639906e+00 -1.36771798e-01 -7.50162780e-01 -7.38044381e-01
-9.96718884e-01 5.27659297e-01 4.77016687e-01 6.68178678e-01
1.02146482e+00 -1.28751922e+00 -6.67155385e-01 -1.22378161e-02
2.17352360e-01 6.63688481e-02 5.80290020e-01 1.50533572e-01
-8.82476807e-01 -9.84248221e-02 -5.35012305e-01 -6.25499368e-01
-1.70435083e+00 4.19118643e-01 3.95377874e-01 4.39842105e-01
-7.94133782e-01 1.05270410e+00 3.68390888e-01 -1.16355956e+00
6.65369689e-01 -3.65755498e-01 -1.23561010e-01 -1.36055827e-01
7.64748991e-01 3.09164822e-01 -3.01711202e-01 -5.09197891e-01
-4.40940320e-01 4.62585628e-01 6.47947609e-01 -8.25410187e-01
1.51998365e+00 1.08810246e-01 4.68825370e-01 8.30816388e-01
7.71735191e-01 -2.78479271e-02 -1.17663574e+00 2.70555794e-01
-7.13274777e-01 -4.36086088e-01 -4.75059897e-02 -9.83157992e-01
-7.74558485e-01 8.02632391e-01 8.18850815e-01 -2.11809836e-02
8.11957836e-01 1.73007488e-01 7.25729108e-01 5.38908601e-01
1.17227614e+00 -9.09881949e-01 6.47504151e-01 6.18729413e-01
1.33714712e+00 -1.05684245e+00 -7.22251177e-01 -6.77036703e-01
-9.41032648e-01 5.49202263e-01 7.58016109e-01 -6.13950565e-03
3.77888501e-01 9.94149506e-01 6.62239552e-01 -2.09483951e-01
-1.04805744e+00 -5.44455588e-01 -4.36073214e-01 1.54305172e+00
2.86448207e-02 2.64199406e-01 7.37990320e-01 5.11089921e-01
-8.11278760e-01 7.88516924e-02 6.02565110e-01 9.55954492e-01
-3.96415383e-01 -4.61080849e-01 -4.68360215e-01 -3.45086128e-01
3.00222367e-01 1.72651872e-01 -6.96224049e-02 9.46604133e-01
-2.90911160e-02 1.15549350e+00 2.53426880e-01 -5.59062064e-01
7.26820827e-01 -2.43534759e-01 8.70252728e-01 -3.59166980e-01
-8.18738163e-01 -2.31002942e-01 3.54681522e-01 -7.37116158e-01
-1.16310433e-01 -5.66203237e-01 -1.94019580e+00 -5.22741914e-01
1.13712028e-01 2.04745270e-02 9.55742240e-01 5.61929405e-01
-8.91116541e-03 1.15776849e+00 9.70720574e-02 -1.54346359e+00
-1.44401610e-01 -6.45848811e-01 -5.02262831e-01 1.91557378e-01
5.75002253e-01 -1.01190329e+00 -3.95004749e-01 -3.47481787e-01] | [4.442207336425781, 0.7960724234580994] |
60cd1740-8063-46fe-a81e-119482ab80a6 | galois-monodromy-groups-for-decomposing | 2105.04460 | null | https://arxiv.org/abs/2105.04460v1 | https://arxiv.org/pdf/2105.04460v1.pdf | Galois/monodromy groups for decomposing minimal problems in 3D reconstruction | We consider Galois/monodromy groups arising in computer vision applications, with a view towards building more efficient polynomial solvers. The Galois/monodromy group allows us to decide when a given problem decomposes into algebraic subproblems, and whether or not it has any symmetries. Tools from numerical algebraic geometry and computational group theory allow us to apply this framework to classical and novel reconstruction problems. We consider three classical cases--3-point absolute pose, 5-point relative pose, and 4-point homography estimation for calibrated cameras--where the decomposition and symmetries may be naturally understood in terms of the Galois/monodromy group. We then show how our framework can be applied to novel problems from absolute and relative pose estimation. For instance, we discover new symmetries for absolute pose problems involving mixtures of point and line features. We also describe a problem of estimating a pair of calibrated homographies between three images. For this problem of degree 64, we can reduce the degree to 16; the latter better reflecting the intrinsic difficulty of algebraically solving the problem. As a byproduct, we obtain new constraints on compatible homographies, which may be of independent interest. | ['Margaret H. Regan', 'Tomas Pajdla', 'Viktor Korotynskiy', 'Timothy Duff'] | 2021-05-10 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [ 2.08119303e-01 3.07849616e-01 9.90115106e-03 -7.60145113e-02
-5.37289083e-01 -7.87483692e-01 6.00739837e-01 -2.88035065e-01
-6.42072177e-03 4.89908069e-01 -1.33086726e-01 -1.24309361e-01
-3.55159074e-01 -4.79115844e-01 -8.00274611e-01 -8.46701086e-01
-1.78214550e-01 7.61021495e-01 -1.23737231e-01 -4.97708797e-01
4.21111494e-01 8.13499868e-01 -1.30261600e+00 -2.65310198e-01
6.73929214e-01 6.82058096e-01 -5.25410771e-02 6.86451733e-01
4.38686669e-01 3.81188095e-01 -7.90299699e-02 -3.16357374e-01
7.77345836e-01 -3.44814211e-01 -1.14877176e+00 8.07955205e-01
1.00830328e+00 -1.36317477e-01 -1.53782323e-01 1.04776442e+00
-2.35679583e-03 -6.47690594e-02 4.85315770e-01 -1.59697187e+00
-3.31545293e-01 -8.38187188e-05 -6.14499092e-01 -2.41858110e-01
7.30771363e-01 -2.23875746e-01 1.26426780e+00 -7.14882791e-01
8.78407776e-01 1.31770527e+00 6.98457897e-01 7.54260197e-02
-1.54603291e+00 -2.03132570e-01 -3.14049482e-01 1.88170940e-01
-1.43241811e+00 -3.51120621e-01 6.36376262e-01 -6.62504315e-01
5.20456851e-01 5.08433521e-01 7.00528026e-01 4.14836258e-01
1.88015550e-01 2.90036023e-01 9.32704091e-01 -5.31529725e-01
-8.03272724e-02 -7.72181675e-02 1.46866128e-01 7.70323932e-01
4.59878206e-01 -6.85891584e-02 -2.36082256e-01 -2.92769164e-01
1.16635931e+00 -5.10010496e-02 -5.52026868e-01 -9.41360354e-01
-1.60800910e+00 9.90909040e-01 2.07294375e-01 1.02313980e-01
-3.28493007e-02 -1.51953802e-01 -2.30987892e-01 4.41397816e-01
-4.12477143e-02 7.40282297e-01 -1.39804870e-01 2.46086881e-01
-4.06219751e-01 2.50720024e-01 1.31992602e+00 1.39597654e+00
1.25574088e+00 -3.35735023e-01 7.25736022e-01 3.84857446e-01
8.29950497e-02 4.41683650e-01 -3.19639564e-01 -1.41321027e+00
2.74611801e-01 3.30256850e-01 7.16763511e-02 -1.41163397e+00
-5.82137465e-01 -1.54827898e-02 -8.13635945e-01 1.41438603e-01
8.07862878e-01 2.38492861e-01 -1.89678669e-01 1.61622834e+00
2.24438310e-01 5.00238277e-02 -8.38231146e-02 7.29826331e-01
3.02876420e-02 4.76162106e-01 -8.60538125e-01 -3.43020856e-01
1.36432886e+00 -5.17778099e-01 -2.81433761e-01 -8.70489627e-02
6.09663486e-01 -1.16107357e+00 3.61047029e-01 6.10926211e-01
-1.12656093e+00 -2.41023004e-01 -1.13128710e+00 -3.40950370e-01
1.73059985e-01 2.99476862e-01 6.03036225e-01 4.48568940e-01
-1.28837240e+00 6.95390284e-01 -5.96157551e-01 -4.65236038e-01
-3.26790363e-01 7.35792458e-01 -6.82483673e-01 -4.45516817e-02
-6.02341652e-01 1.06712270e+00 7.56478757e-02 1.80965774e-02
-1.82645768e-01 -6.06425881e-01 -8.00051928e-01 -8.17861930e-02
6.44229770e-01 -8.34093332e-01 1.07773328e+00 -8.10542107e-01
-1.16335750e+00 1.20227516e+00 -1.28595501e-01 -1.17056787e-01
5.07888496e-01 2.64612198e-01 -5.42571396e-02 1.82371855e-01
1.80084229e-01 4.56554830e-01 8.14229608e-01 -1.01388001e+00
-3.09575826e-01 -7.25208640e-01 4.70341682e-01 3.10980827e-01
3.16754758e-01 -1.29367948e-01 -3.74023378e-01 -1.88074470e-01
9.04094160e-01 -1.65284598e+00 -3.80550891e-01 2.68883973e-01
-4.67338473e-01 2.65561882e-03 7.30106890e-01 -6.00941360e-01
4.26964879e-01 -2.04229426e+00 8.06657016e-01 5.20755649e-01
4.30532783e-01 -2.46742383e-01 1.35068774e-01 3.83851439e-01
-4.41001713e-01 -1.57277331e-01 -4.98089120e-02 5.42484969e-03
-8.48296732e-02 4.21808124e-01 -1.10554077e-01 1.19759071e+00
2.11642593e-01 2.54358411e-01 -5.38786411e-01 -3.41134995e-01
5.87717853e-02 3.06819171e-01 -8.77880991e-01 -1.47789851e-01
2.33569238e-02 7.40362644e-01 -4.59772348e-01 2.97920913e-01
1.12562144e+00 -1.82434231e-01 3.39108288e-01 -4.11932588e-01
-2.36752287e-01 5.09937406e-02 -1.98969972e+00 1.54169202e+00
-2.89670318e-01 5.17709732e-01 5.09846509e-01 -9.54871655e-01
7.34480977e-01 4.99823019e-02 7.86800802e-01 -8.24263319e-03
3.12192708e-01 8.87003466e-02 5.75325787e-02 -2.49465913e-01
7.52985299e-01 -9.96670220e-03 1.41574664e-03 2.88450032e-01
1.46117173e-02 -4.74841803e-01 2.40181565e-01 1.95684895e-01
7.35961914e-01 -1.74838901e-01 4.53347743e-01 -6.88083827e-01
7.32049584e-01 1.81011297e-02 5.58286607e-01 3.00196379e-01
3.91138643e-01 7.16141760e-01 1.02650046e+00 -3.51258039e-01
-1.33271706e+00 -1.04323220e+00 -2.53491670e-01 3.60703677e-01
3.27818483e-01 -6.00494325e-01 -7.63345182e-01 1.59921318e-01
2.93881819e-02 2.94938870e-02 -3.20907652e-01 7.41838738e-02
-6.60830200e-01 -4.24656659e-01 2.64284238e-02 8.01851451e-02
3.34513336e-01 -1.61689907e-01 -4.04053271e-01 -1.07481085e-01
-2.52677262e-01 -1.34083855e+00 -5.46683609e-01 -6.99730664e-02
-1.13134158e+00 -1.50567889e+00 -6.15240753e-01 -8.46209884e-01
9.20571268e-01 5.97321451e-01 8.80153537e-01 7.24222213e-02
-2.36133307e-01 8.35119426e-01 6.28397241e-02 1.62456304e-01
-4.33060884e-01 -2.06089482e-01 3.75771344e-01 1.35659799e-01
4.36523855e-02 -6.20481849e-01 -2.34898239e-01 9.34848726e-01
-8.42955589e-01 8.19889307e-02 1.78258151e-01 6.32976651e-01
5.54104984e-01 -9.62304547e-02 -3.93688798e-01 -5.46970308e-01
8.65245461e-02 -1.99288025e-01 -1.19649613e+00 3.13221037e-01
-3.77948098e-02 3.17680448e-01 4.36876774e-01 -2.28177294e-01
-5.39698303e-01 6.08873188e-01 2.95104593e-01 -2.86960632e-01
3.14406842e-01 2.73658067e-01 -5.48489511e-01 -8.06878448e-01
5.28468847e-01 -9.46450308e-02 2.79529750e-01 -3.55313003e-01
3.95244151e-01 3.34772497e-01 7.08518744e-01 -8.11527312e-01
8.39826941e-01 6.90449417e-01 8.36039543e-01 -1.34732711e+00
-3.60696465e-01 -6.81283712e-01 -1.05292261e+00 3.22320163e-02
7.72645354e-01 -9.76309299e-01 -1.04094231e+00 4.21204746e-01
-1.44864881e+00 3.25307786e-01 -2.66178027e-02 5.45995414e-01
-1.21631396e+00 9.22460020e-01 -3.70737731e-01 -4.52363282e-01
4.51435208e-01 -1.60437310e+00 1.14389038e+00 -1.87141940e-01
-1.09972939e-01 -9.23963785e-01 4.70119305e-02 3.62696201e-01
-1.71155915e-01 1.69362962e-01 7.65668631e-01 -2.10138991e-01
-1.23050821e+00 -7.07208514e-02 -5.34552522e-02 1.20549448e-01
-1.15442924e-01 7.78846145e-02 -5.56257784e-01 -3.01533997e-01
3.93582314e-01 2.78783619e-01 2.93514669e-01 1.88591093e-01
6.29856884e-01 -3.01381230e-01 -1.12139806e-01 9.14627731e-01
1.47040701e+00 -1.68819889e-01 6.84408188e-01 3.59515399e-01
8.66792679e-01 7.52294242e-01 4.57626909e-01 4.38365370e-01
3.22063297e-01 1.21305239e+00 4.14382964e-01 2.07120165e-01
2.95033872e-01 4.53849360e-02 5.03035374e-02 8.76952112e-01
-4.93911684e-01 4.07552481e-01 -8.22553158e-01 1.58386007e-01
-1.75632370e+00 -8.15218985e-01 -5.88389337e-01 2.61565161e+00
3.92827213e-01 -4.36262369e-01 2.61356175e-01 1.29897252e-01
9.10671175e-01 -1.65454745e-01 -2.41426453e-01 -1.85608432e-01
-2.14218885e-01 -1.76158249e-01 9.51973557e-01 7.59440184e-01
-8.95013213e-01 5.45642734e-01 6.37227201e+00 4.97458309e-01
-8.28009367e-01 -2.07185626e-01 1.08500674e-01 5.32710791e-01
-2.09194049e-01 6.58067763e-01 -8.82302463e-01 -1.25307634e-01
3.57749194e-01 -2.55137235e-01 7.31633306e-01 8.04526567e-01
-2.54258215e-01 -2.88805842e-01 -1.45008075e+00 1.33145094e+00
3.61636192e-01 -1.22445810e+00 -1.02642015e-01 7.29652345e-01
7.41472721e-01 -2.99868137e-01 2.56956294e-02 -4.78440553e-01
1.86370276e-02 -6.15841866e-01 5.56600511e-01 2.80069292e-01
5.35589278e-01 -5.76974511e-01 1.18900895e-01 4.07236069e-01
-1.21145594e+00 2.06113949e-01 -7.06525087e-01 -2.54328251e-01
1.86347678e-01 1.70741111e-01 -5.50617337e-01 8.01756263e-01
2.30611160e-01 8.02873552e-01 -3.73179823e-01 1.10076427e+00
1.53964236e-01 -3.21830809e-01 -5.81060529e-01 3.98945302e-01
2.66647264e-02 -1.11705720e+00 7.84217358e-01 5.18706799e-01
6.11476243e-01 5.51425099e-01 6.66997433e-02 8.15950751e-01
2.49656856e-01 1.13670915e-01 -7.76268423e-01 4.56601411e-01
-1.00143000e-01 1.25403893e+00 -7.79747188e-01 -4.77353716e-03
-4.85905975e-01 7.59928346e-01 -9.38357487e-02 1.07517876e-01
-5.12412727e-01 -1.56750545e-01 9.13270056e-01 2.09269658e-01
3.29553366e-01 -7.10440338e-01 -5.51772900e-02 -1.81753242e+00
8.43749493e-02 -1.08260155e+00 8.75274241e-02 -8.30653965e-01
-7.72593558e-01 3.17410171e-01 2.58693665e-01 -1.45050490e+00
-4.05733138e-01 -1.02410257e+00 -4.25273448e-01 6.92134440e-01
-7.87774801e-01 -8.53848100e-01 -1.21608734e-01 8.57334435e-01
9.27971452e-02 1.69573843e-01 7.00639367e-01 6.81917518e-02
-1.67589754e-01 2.57572867e-02 1.31840497e-01 -1.36577070e-01
4.92012858e-01 -1.25278234e+00 3.37150663e-01 9.78443503e-01
2.23790035e-01 7.09973574e-01 8.27220261e-01 -2.73707211e-01
-2.14545083e+00 -5.14912367e-01 8.67465019e-01 -6.32172287e-01
8.84293616e-01 -3.20361167e-01 -7.54708409e-01 1.16592240e+00
-2.46062487e-01 -1.26867071e-01 2.32181951e-01 1.03944637e-01
-3.67154092e-01 4.58634458e-02 -1.02252901e+00 5.53445160e-01
1.05077529e+00 -5.44508338e-01 -4.44067836e-01 7.21933722e-01
2.98827171e-01 -7.39993691e-01 -9.47879851e-01 9.53985304e-02
3.33602011e-01 -9.29734826e-01 1.19164670e+00 -3.79005522e-01
8.64874646e-02 -4.84521240e-01 -4.71268088e-01 -1.01682067e+00
-2.69341797e-01 -1.09306526e+00 6.20722353e-01 7.81449318e-01
-4.61489893e-03 -6.14182293e-01 7.38721371e-01 6.83010161e-01
7.54127875e-02 -1.72052607e-01 -8.94028842e-01 -1.01234889e+00
3.59912440e-02 -1.85765937e-01 5.16361713e-01 1.14533913e+00
2.07986042e-01 3.37952286e-01 -7.01738954e-01 4.19499725e-01
7.49728680e-01 4.54742871e-02 1.14105272e+00 -1.49819338e+00
-5.51309943e-01 -2.45532036e-01 -1.11538708e+00 -1.44726229e+00
1.85600102e-01 -8.74754965e-01 -2.12364763e-01 -8.12529385e-01
4.10724655e-02 -3.58834803e-01 6.44356430e-01 6.99144900e-02
5.52573621e-01 1.55281737e-01 5.35504580e-01 3.71594369e-01
-3.74617875e-01 3.80922146e-02 1.22513449e+00 1.16061747e-01
-9.74277779e-02 2.62792826e-01 -5.50202370e-01 9.06492233e-01
3.84002388e-01 -1.92616239e-01 -7.04369545e-02 -3.92495245e-01
6.52276814e-01 5.23111582e-01 5.33692658e-01 -9.45538104e-01
5.03739119e-01 -2.30137363e-01 -1.09282516e-01 -3.04739803e-01
5.40418148e-01 -1.03945422e+00 7.72397220e-01 2.95279920e-01
1.36312827e-01 3.44666004e-01 -1.67178765e-01 3.42716724e-01
-5.89091219e-02 -6.12190664e-01 7.09416091e-01 -1.67453691e-01
-5.59780180e-01 3.71899724e-01 -1.99047402e-01 -9.20838118e-02
9.16115165e-01 -4.36210096e-01 -2.56896824e-01 -6.48238242e-01
-1.04794168e+00 6.84191734e-02 1.04804790e+00 1.40880749e-01
3.47969949e-01 -1.38776088e+00 -4.78792161e-01 4.20421958e-01
1.84102178e-01 -4.30403650e-02 1.30237475e-01 1.32128930e+00
-9.61069047e-01 4.54305023e-01 -4.20319229e-01 -1.11286795e+00
-1.57596481e+00 5.43363512e-01 3.97679597e-01 1.48687497e-01
-3.44619364e-01 2.67397791e-01 3.79200518e-01 -3.88617635e-01
1.24546262e-02 -4.08369690e-01 1.79924831e-01 -1.09967098e-01
2.83151716e-01 7.14039564e-01 2.78209485e-02 -1.25638437e+00
-1.96802795e-01 1.33529508e+00 1.82473436e-01 -3.13486084e-02
1.18489254e+00 -2.65069544e-01 -7.04623401e-01 6.84535429e-02
1.55244470e+00 1.59909621e-01 -9.82963204e-01 -2.20357269e-01
-1.38837770e-01 -5.32245219e-01 -4.20299858e-01 2.59319991e-01
-9.53442931e-01 6.46087706e-01 1.31742060e-02 2.41822734e-01
1.04264212e+00 3.14707190e-01 3.49349916e-01 8.17737877e-01
5.76494157e-01 -6.46511018e-01 -3.88072610e-01 7.32864082e-01
1.10096455e+00 -1.02782881e+00 2.92345464e-01 -1.01455653e+00
-3.30632329e-01 1.65920210e+00 7.82056898e-02 -5.24475813e-01
6.47697985e-01 -1.12823404e-01 -2.41919085e-01 -8.88127983e-02
-2.36329600e-01 -2.56562717e-02 5.48539340e-01 4.33597893e-01
8.56093720e-06 -3.46297249e-02 -2.47965887e-01 -2.46399343e-01
-6.01495862e-01 -3.23963016e-01 1.03113604e+00 5.93422353e-01
-3.95061284e-01 -1.46712422e+00 -8.09964180e-01 6.25389814e-02
-1.40941054e-01 3.27503949e-01 -4.95624512e-01 1.01410770e+00
1.31151661e-01 6.96661234e-01 8.98065343e-02 -2.89559692e-01
2.78492540e-01 -4.26026940e-01 1.01715028e+00 -4.48465943e-01
-5.22364047e-04 3.77216786e-01 -5.25555834e-02 -6.91168904e-01
-6.78578794e-01 -9.60509002e-01 -6.45252109e-01 -6.46244109e-01
-4.01241958e-01 2.03326404e-01 7.09001005e-01 8.96647453e-01
-1.98871223e-03 -3.02364409e-01 6.61202610e-01 -9.83921528e-01
-8.39336336e-01 -3.68167102e-01 -9.82553601e-01 2.06761226e-01
5.89155138e-01 -6.25775695e-01 -6.30728304e-01 2.17339098e-02] | [8.00554370880127, -2.274226665496826] |
8032859e-e1cf-4e0c-a8a9-af6f5486459d | universe-points-representation-learning-for | 2212.00780 | null | https://arxiv.org/abs/2212.00780v2 | https://arxiv.org/pdf/2212.00780v2.pdf | Universe Points Representation Learning for Partial Multi-Graph Matching | Many challenges from natural world can be formulated as a graph matching problem. Previous deep learning-based methods mainly consider a full two-graph matching setting. In this work, we study the more general partial matching problem with multi-graph cycle consistency guarantees. Building on a recent progress in deep learning on graphs, we propose a novel data-driven method (URL) for partial multi-graph matching, which uses an object-to-universe formulation and learns latent representations of abstract universe points. The proposed approach advances the state of the art in semantic keypoint matching problem, evaluated on Pascal VOC, CUB, and Willow datasets. Moreover, the set of controlled experiments on a synthetic graph matching dataset demonstrates the scalability of our method to graphs with large number of nodes and its robustness to high partiality. | ['Florian Bernard', 'Frank R. Schmidt', 'Zhakshylyk Nurlanov'] | 2022-12-01 | null | null | null | null | ['graph-matching'] | ['graphs'] | [ 1.10362865e-01 1.64666578e-01 -4.47791010e-01 -3.11123610e-01
-6.98411822e-01 -4.24407572e-01 9.41105962e-01 5.02139688e-01
-1.18385434e-01 1.40355468e-01 1.12397350e-01 -8.77922177e-02
-2.98829228e-01 -1.20795524e+00 -1.06559074e+00 -2.02463597e-01
-9.81911048e-02 7.30671287e-01 5.06251216e-01 -3.58992010e-01
1.14519052e-01 4.02607381e-01 -1.57507539e+00 1.86325684e-01
4.94830281e-01 7.88844943e-01 -1.67459790e-02 5.84497511e-01
-2.06276178e-01 8.42648625e-01 -1.63640156e-01 -7.36487627e-01
5.12055218e-01 2.83904113e-02 -1.06045806e+00 -7.03398138e-02
1.42478645e+00 -2.06077367e-01 -7.64254630e-01 1.20016217e+00
3.33714038e-01 1.83648705e-01 3.27133209e-01 -1.90807271e+00
-9.94323254e-01 4.93235588e-01 -5.75145185e-01 -9.21438336e-02
7.00214624e-01 -7.59313554e-02 1.64366508e+00 -3.98750693e-01
7.87061334e-01 1.55191362e+00 9.76374030e-01 3.83494854e-01
-1.04008484e+00 -5.17420948e-01 1.32570386e-01 4.43083107e-01
-1.17812037e+00 1.44103356e-02 8.81319940e-01 -3.40630680e-01
1.24436033e+00 -2.53348183e-02 6.00270867e-01 8.39560628e-01
1.62993580e-01 6.33913696e-01 7.90484488e-01 -3.39696854e-01
-2.10960601e-02 -5.30755520e-01 4.21749264e-01 1.17195916e+00
5.21753788e-01 2.18578085e-01 -3.01866323e-01 -2.06017435e-01
5.26005030e-01 8.83030444e-02 -5.57990512e-03 -1.06433344e+00
-1.17979348e+00 8.90219688e-01 8.44619572e-01 2.15789378e-01
-5.66679649e-02 8.94816279e-01 5.18453002e-01 6.43683851e-01
2.70941079e-01 9.99368280e-02 -2.13340849e-01 3.94308925e-01
-7.08085358e-01 6.55051768e-01 1.00721800e+00 1.46167135e+00
1.07584047e+00 -2.34004214e-01 -1.10540599e-01 4.53040183e-01
4.39738035e-01 3.19461316e-01 1.51178464e-01 -6.72121048e-01
4.34721857e-01 1.14949024e+00 -3.61131877e-01 -1.43784094e+00
-4.94229585e-01 -9.89554599e-02 -7.19528615e-01 1.08495762e-03
2.39341006e-01 6.58151269e-01 -1.02053094e+00 1.69362724e+00
4.84279692e-01 5.92296541e-01 -4.34075631e-02 7.32023120e-01
1.22357702e+00 1.32566005e-01 3.86969894e-02 4.88149941e-01
1.22203851e+00 -1.27897573e+00 -3.82379174e-01 -2.25455016e-01
8.29137266e-01 -4.37526703e-01 9.93226767e-01 -9.48961973e-02
-8.28044832e-01 -6.47675812e-01 -1.26150644e+00 -4.29259777e-01
-6.67452872e-01 -4.45159078e-01 1.02450669e+00 4.06198859e-01
-1.39395726e+00 9.34274018e-01 -4.59143758e-01 -8.34286451e-01
3.90983999e-01 4.29488510e-01 -8.58926117e-01 -3.89029294e-01
-1.22708499e+00 7.77945518e-01 6.66123450e-01 -2.82867759e-01
-8.32423925e-01 -6.60377502e-01 -1.30914271e+00 1.56158075e-01
4.31199431e-01 -1.21445036e+00 9.64937389e-01 -7.04365373e-01
-9.38733816e-01 1.43598354e+00 2.94977814e-01 -6.87051117e-01
5.05416214e-01 -8.58355239e-02 -2.09830627e-01 2.80342489e-01
1.02525949e-01 6.74315870e-01 6.98955059e-01 -9.61653173e-01
-4.07110512e-01 -5.99953353e-01 5.51702976e-01 -9.65584293e-02
-1.76799536e-01 -9.06384289e-02 -4.67394054e-01 -3.94482166e-01
1.70562908e-01 -9.91994143e-01 5.60317794e-03 2.69071877e-01
-2.08676472e-01 -6.01371706e-01 7.54594505e-01 -3.17777812e-01
8.27890277e-01 -1.82986474e+00 3.52188528e-01 1.88792907e-02
6.15628123e-01 6.18589967e-02 -4.70747024e-01 7.58218169e-01
-2.71176010e-01 -2.98605263e-02 -1.72654718e-01 -4.41009939e-01
5.04955351e-01 2.64873117e-01 -3.64018790e-02 7.98396528e-01
4.59863944e-03 1.45127845e+00 -1.14854646e+00 -6.74797833e-01
2.53982097e-01 1.70523778e-01 -4.97170329e-01 3.60759258e-01
-3.73654306e-01 -2.39976719e-01 -2.03682750e-01 7.25301981e-01
9.20242488e-01 -6.96558177e-01 3.09057176e-01 -3.23170841e-01
4.28483903e-01 -1.58633038e-01 -1.29693758e+00 2.54945588e+00
-2.75639206e-01 3.44543397e-01 -1.89098697e-02 -1.31646240e+00
9.43913937e-01 7.10217804e-02 5.51160693e-01 -7.69909680e-01
-1.37046471e-01 1.16283700e-01 -3.29096735e-01 -2.71696806e-01
6.51507795e-01 2.90189862e-01 -2.54325241e-01 3.46298367e-01
4.93095636e-01 -3.27880055e-01 2.84921288e-01 6.41284168e-01
1.36339736e+00 4.08213824e-01 4.15362597e-01 -5.20572066e-01
4.05234069e-01 -4.12031729e-03 3.35295975e-01 9.05635297e-01
-2.51399547e-01 5.86145163e-01 4.93050754e-01 -8.05557430e-01
-1.09085643e+00 -9.02327955e-01 2.36754000e-01 9.55987275e-01
5.42391777e-01 -8.11036289e-01 -5.65137804e-01 -8.73643935e-01
4.75061446e-01 5.98812699e-02 -6.22796893e-01 -1.95303455e-01
-7.08166778e-01 -3.38904619e-01 5.66327810e-01 5.32473028e-01
4.29437518e-01 -8.06490183e-01 -1.80078194e-01 1.13347255e-01
1.21106245e-01 -1.67625165e+00 -4.12729979e-01 -5.20931602e-01
-5.29797852e-01 -1.64964044e+00 -3.30293298e-01 -1.05935585e+00
3.99178386e-01 4.31792825e-01 1.57135594e+00 5.99761128e-01
-6.35452151e-01 9.34134305e-01 -2.20055714e-01 -9.19639617e-02
-4.97268111e-01 1.41668871e-01 -2.32523218e-01 -6.55950904e-02
3.72725040e-01 -7.41279483e-01 -4.88259017e-01 3.66027690e-02
-9.03430521e-01 1.68379873e-01 2.21676961e-01 6.55400753e-01
6.61889732e-01 -3.22167903e-01 2.94517606e-01 -1.05625892e+00
5.40567219e-01 -4.33286607e-01 -1.11186349e+00 4.97238666e-01
-8.32725883e-01 4.10125166e-01 4.95427459e-01 -1.82371795e-01
-4.99621719e-01 9.46725085e-02 1.51061073e-01 -7.17799783e-01
-4.11216915e-03 2.99062610e-01 -2.22823396e-01 -6.03369594e-01
4.97749984e-01 5.09454831e-02 -1.74538925e-01 -3.03971857e-01
7.60572851e-01 1.20646350e-01 6.90626085e-01 -8.19472909e-01
1.04077828e+00 6.78045213e-01 6.70433044e-01 -3.79245937e-01
-4.91762996e-01 -7.53436148e-01 -8.05143595e-01 -6.93070889e-02
7.53292620e-01 -1.04721224e+00 -8.95802796e-01 5.55323124e-01
-1.31587195e+00 -2.95562178e-01 7.58258486e-03 7.75128007e-02
-8.56636167e-01 9.00899827e-01 -5.91977715e-01 -1.39733955e-01
-7.02918887e-01 -1.06751525e+00 1.61987019e+00 -2.28324812e-02
7.00043887e-02 -1.27521181e+00 4.65085059e-01 3.11015129e-01
2.97381967e-01 7.19487965e-01 1.00446558e+00 -7.32784569e-01
-1.12836969e+00 -3.05016339e-01 -5.54242194e-01 -2.77403474e-01
-7.09657371e-02 -1.61757976e-01 -6.91727817e-01 -6.35653555e-01
-6.78074419e-01 -6.48042023e-01 9.40730035e-01 -1.34921968e-01
8.45111609e-01 -1.41014263e-01 -5.27118742e-01 1.03435469e+00
1.95097673e+00 -6.35600686e-01 5.29412806e-01 3.76958132e-01
1.17076695e+00 4.77948964e-01 4.18559551e-01 9.61103290e-02
8.55057120e-01 7.48771131e-01 9.78792429e-01 -2.64831353e-02
-6.72370315e-01 -5.36121190e-01 -2.60444313e-01 6.89811349e-01
2.50179559e-01 -2.20019162e-01 -9.35640037e-01 7.55170286e-01
-2.36918426e+00 -7.76285648e-01 -3.43242705e-01 1.86921751e+00
1.48977324e-01 -9.26986709e-02 -1.30591514e-02 -2.45106637e-01
8.80119681e-01 4.46596116e-01 -3.07254851e-01 -1.99749842e-01
-1.39234498e-01 3.18247437e-01 6.05744660e-01 3.63139540e-01
-1.21122789e+00 1.18643570e+00 5.98751307e+00 6.13446295e-01
-6.81040943e-01 1.40901372e-01 1.05537297e-02 4.88909125e-01
-5.98804176e-01 4.93474573e-01 -5.80219388e-01 -4.23209518e-02
5.70938647e-01 -3.61022145e-01 5.69483936e-01 9.44782495e-01
-6.90357924e-01 2.91107476e-01 -1.49790907e+00 1.20105696e+00
2.36744776e-01 -1.69970143e+00 3.56497914e-01 -1.38726071e-01
6.51684344e-01 2.97087997e-01 -3.70122671e-01 3.22642058e-01
5.81014395e-01 -8.97456884e-01 5.27911603e-01 3.68499160e-01
7.31805503e-01 -4.03084844e-01 3.30813259e-01 7.02105984e-02
-1.82704651e+00 1.55782163e-01 -6.02283359e-01 7.45100901e-02
-1.26764998e-01 7.87160695e-02 -5.80695629e-01 1.26155579e+00
5.48321366e-01 9.88268435e-01 -9.15479183e-01 9.50383246e-01
-2.59457916e-01 -1.42410938e-02 -1.77358553e-01 1.40307397e-01
2.84002602e-01 -5.84968813e-02 5.78895450e-01 1.24512529e+00
1.09448703e-03 -2.76998907e-01 3.76163423e-01 1.09031868e+00
-4.64656591e-01 6.54487163e-02 -1.08569109e+00 4.21844758e-02
3.67837131e-01 1.48352742e+00 -6.69997633e-01 -1.94182813e-01
-7.53883183e-01 9.39660549e-01 8.78815472e-01 1.00392833e-01
-8.16848099e-01 -3.43716919e-01 4.56259966e-01 2.75212247e-02
1.10025167e-01 -2.19274148e-01 3.02444577e-01 -1.35754371e+00
9.02686641e-02 -8.97108316e-01 9.70458329e-01 -7.66618848e-01
-1.39811099e+00 4.95069772e-01 2.71150857e-01 -9.06520247e-01
-1.30477652e-01 -6.02349102e-01 -7.94744968e-01 4.39632565e-01
-1.81339180e+00 -1.88422465e+00 -7.15154350e-01 9.92517710e-01
2.15617597e-01 -1.88711077e-01 9.27265704e-01 4.64279741e-01
-1.13423206e-02 7.31847346e-01 -3.99208337e-01 3.26706737e-01
5.26359320e-01 -1.49716532e+00 1.21262264e+00 9.43607271e-01
4.35387224e-01 3.83403182e-01 4.40538675e-01 -5.79824567e-01
-2.08061624e+00 -1.21662450e+00 7.44644165e-01 -3.94703388e-01
9.42791581e-01 -5.77115715e-01 -8.79196167e-01 9.86774564e-01
2.60328472e-01 5.79855263e-01 2.99665891e-02 7.97260180e-02
-9.24743831e-01 -1.62185997e-01 -1.14385700e+00 4.14530307e-01
1.57294476e+00 -7.03654706e-01 -6.36561811e-01 6.10180974e-01
1.17240000e+00 -5.91261506e-01 -1.03220749e+00 8.04354548e-01
5.42416513e-01 -8.45141768e-01 1.08911824e+00 -8.99099648e-01
6.15890920e-02 -1.55486003e-01 -3.39742839e-01 -1.06493473e+00
-4.45199668e-01 -7.66178787e-01 -2.46729583e-01 1.05194616e+00
-1.78015932e-01 -5.26718676e-01 1.05092096e+00 3.44557881e-01
-1.21819109e-01 -3.93386126e-01 -8.79057527e-01 -8.48256052e-01
-4.15134765e-02 -1.61031634e-01 1.02325475e+00 1.16194654e+00
-1.88313529e-01 1.84218049e-01 -2.66181350e-01 2.67833441e-01
1.13728285e+00 6.69759393e-01 1.28111422e+00 -1.41636860e+00
-4.12465990e-01 -3.48115325e-01 -1.10885811e+00 -7.10774541e-01
7.76450276e-01 -1.39337146e+00 -4.67230052e-01 -1.82054913e+00
3.27954650e-01 -1.69395760e-01 -1.38418928e-01 4.36795473e-01
-6.86398819e-02 1.26938559e-02 3.44059229e-01 -1.17553525e-01
-9.97079194e-01 3.60767931e-01 1.02634871e+00 -6.04005814e-01
4.95050877e-01 -2.96412170e-01 -2.33770624e-01 3.32778007e-01
4.60485280e-01 -5.53162098e-01 -5.00790715e-01 -6.20752037e-01
4.32142377e-01 2.98235063e-02 6.48193300e-01 -8.62271547e-01
5.14870822e-01 -9.81823951e-02 -3.52982402e-01 -4.15864825e-01
2.19402730e-01 -8.12103271e-01 3.97991091e-01 5.49390435e-01
-2.36090630e-01 5.18720984e-01 2.09198400e-01 9.32444513e-01
-1.93121552e-01 9.35971662e-02 6.89037979e-01 -4.85139072e-01
-1.23020911e+00 9.72201705e-01 6.71390533e-01 3.00494254e-01
1.01823580e+00 -1.55191749e-01 -7.22364426e-01 -2.14984417e-01
-4.16779041e-01 4.15874809e-01 6.43547714e-01 8.39565337e-01
6.22490227e-01 -1.69989181e+00 -8.53049576e-01 1.32844716e-01
7.52838016e-01 -1.00610450e-01 1.37911215e-01 4.67611939e-01
-6.94737494e-01 3.29477012e-01 -2.07683891e-01 -6.94231451e-01
-1.28653896e+00 1.02225304e+00 5.81306756e-01 -4.61995810e-01
-8.71278524e-01 7.15335190e-01 1.77243754e-01 -7.66706526e-01
3.67214739e-01 -2.43216619e-01 9.21624228e-02 -2.97261089e-01
1.21591486e-01 4.03765410e-01 2.76055127e-01 -8.16869974e-01
-5.40279448e-01 7.51179814e-01 -5.16146384e-02 4.73829091e-01
1.18538976e+00 2.54982382e-01 -4.57614988e-01 2.28058081e-03
1.66029620e+00 -4.55377370e-01 -7.28936732e-01 -5.85213780e-01
2.95361072e-01 -5.39829373e-01 -2.62701720e-01 -5.90536557e-02
-1.08437860e+00 7.23971605e-01 4.68749106e-01 9.81014743e-02
6.45367205e-01 8.88791010e-02 9.89837289e-01 5.69447517e-01
5.75137973e-01 -7.71022797e-01 2.02439114e-01 2.54366070e-01
7.69866407e-01 -1.41947174e+00 5.28186448e-02 -3.61563742e-01
-6.42390773e-02 1.22836077e+00 8.39274645e-01 -5.83372653e-01
6.96552634e-01 -3.33366059e-02 -3.23366344e-01 -7.47341335e-01
-7.25215793e-01 -4.79085296e-01 4.87589002e-01 6.19290113e-01
8.29823092e-02 7.92253483e-03 -1.84961528e-01 4.68349643e-02
1.13374665e-01 -1.19175026e-02 4.08804327e-01 8.34378839e-01
-1.58762157e-01 -1.33247685e+00 1.02391176e-01 1.10822842e-01
-3.88065904e-01 -8.20436552e-02 -6.17089212e-01 1.09861851e+00
-2.55591124e-01 7.83597767e-01 -8.98451358e-02 -3.03819716e-01
4.69179541e-01 -2.69637764e-01 9.63203728e-01 -6.07014477e-01
-6.11501932e-01 -5.20136952e-01 -7.06822202e-02 -1.09197855e+00
-6.52225733e-01 -1.67731434e-01 -1.24900305e+00 -4.68091697e-01
-4.23434317e-01 -6.64104447e-02 4.66840059e-01 7.53553689e-01
4.65690047e-01 4.13003892e-01 1.59451306e-01 -4.78680938e-01
-5.68869472e-01 -5.46576500e-01 -5.59882581e-01 1.20053756e+00
3.18843096e-01 -6.07153833e-01 -2.21207321e-01 -4.31961238e-01] | [7.146459579467773, 6.306248664855957] |
0921fdcc-6798-42e8-a350-7f38851870ec | moving-object-detection-in-time-lapse-or | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Shakeri_Moving_Object_Detection_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Shakeri_Moving_Object_Detection_ICCV_2017_paper.pdf | Moving Object Detection in Time-Lapse or Motion Trigger Image Sequences Using Low-Rank and Invariant Sparse Decomposition | Low-rank and sparse representation based methods have attracted wide attention in background subtraction and moving object detection, where moving objects in the scene are modeled as pixel-wise sparse outliers. Since in real scenarios moving objects are also structurally sparse, recently researchers have attempted to extract moving objects using structured sparse outliers. Although existing methods with structured sparsity-inducing norms produce promising results, they are still vulnerable to various illumination changes that frequently occur in real environments, specifically for time-lapse image sequences where assumptions about sparsity between images such as group sparsity are not valid. In this paper, we first introduce a prior map obtained by illumination invariant representation of images. Next, we propose a low-rank and invariant sparse decomposition using the prior map to detect moving objects under significant illumination changes. Experiments on challenging benchmark datasets demonstrate the superior performance of our proposed method under complex illumination changes.
| ['Moein Shakeri', 'Hong Zhang'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['moving-object-detection'] | ['computer-vision'] | [ 6.60899103e-01 -8.61738324e-01 5.11795208e-02 -2.62390345e-01
-6.04552269e-01 -3.25244248e-01 3.55679840e-01 -2.98902541e-01
-1.68406263e-01 7.65390635e-01 2.94957429e-01 1.87277511e-01
-6.34782687e-02 -3.52064520e-01 -7.75596023e-01 -9.53182578e-01
-1.22732021e-01 -2.25241989e-01 3.38674635e-01 1.56320050e-01
3.53976518e-01 4.00655627e-01 -1.61693585e+00 2.67194003e-01
9.45055068e-01 7.93920636e-01 2.22211152e-01 5.62583089e-01
4.62801941e-02 1.10274422e+00 -6.07521117e-01 2.49096021e-01
7.09886670e-01 -4.66215879e-01 -1.46707326e-01 8.10805738e-01
1.05253458e+00 -1.86297819e-01 -5.37701607e-01 1.49507987e+00
3.50068241e-01 5.61433673e-01 2.66902536e-01 -1.11060309e+00
-4.48711842e-01 1.77009672e-01 -1.09824288e+00 6.90740466e-01
4.77596462e-01 9.29986686e-02 3.35024029e-01 -1.29559731e+00
4.91837114e-01 1.19090855e+00 4.67198312e-01 1.04487948e-01
-1.12180841e+00 -4.38022733e-01 6.76595747e-01 4.89146739e-01
-1.50739682e+00 -8.42099190e-01 9.84529257e-01 -3.76597017e-01
4.43162471e-01 6.00550592e-01 3.34647864e-01 8.66251767e-01
1.04807712e-01 8.31228197e-01 1.15337217e+00 -1.75291389e-01
1.87712774e-01 -1.94104090e-01 8.64248052e-02 4.50002491e-01
9.82200623e-01 -1.76215813e-01 -5.68981946e-01 -3.11869115e-01
6.50505841e-01 6.27178252e-01 -6.76399529e-01 -4.47529823e-01
-1.48390377e+00 4.36981916e-01 1.69179648e-01 3.90501499e-01
-6.67571783e-01 -5.07850870e-02 1.93152621e-01 3.28654051e-03
2.94424295e-01 -1.22785799e-01 -5.58306016e-02 2.41858259e-01
-1.22625840e+00 9.60036218e-02 4.12324697e-01 1.04296100e+00
5.60348809e-01 7.11490989e-01 -2.38560140e-01 8.62276375e-01
2.52022803e-01 7.39425898e-01 5.45500755e-01 -9.89327252e-01
4.63652402e-01 1.81067199e-01 4.09067243e-01 -1.76008940e+00
-1.12094313e-01 -6.80903792e-01 -1.34904981e+00 -1.31937444e-01
4.01349992e-01 1.80961251e-01 -9.38038588e-01 1.29604888e+00
4.30607557e-01 1.11160886e+00 6.70005530e-02 1.22873914e+00
7.72622228e-01 8.45302820e-01 -2.24945337e-01 -7.98963308e-01
1.02439415e+00 -7.97860146e-01 -9.63933885e-01 -3.21384788e-01
1.03965536e-01 -9.39644933e-01 6.06752157e-01 6.35606587e-01
-1.05160081e+00 -6.14866674e-01 -7.44132221e-01 3.04435283e-01
1.90920070e-01 -1.13693394e-01 5.52453160e-01 5.08842409e-01
-6.38305843e-01 1.85285598e-01 -9.05382812e-01 -3.96876842e-01
3.77472818e-01 6.85881823e-02 -3.87240112e-01 -6.29824340e-01
-6.77317262e-01 5.14023244e-01 2.66968071e-01 5.45082986e-01
-1.09273648e+00 -3.26349229e-01 -8.73751640e-01 -1.16847150e-01
5.09241700e-01 -6.17684901e-01 6.51354790e-01 -1.08161092e+00
-1.10426736e+00 4.93884504e-01 -7.02181339e-01 -4.21575755e-01
3.42969865e-01 -4.69015419e-01 -7.45548248e-01 2.43015662e-01
1.67236388e-01 3.10384315e-02 1.41233897e+00 -1.49393129e+00
-6.89831674e-01 -2.33513474e-01 -3.08919787e-01 1.69424772e-01
-2.05650389e-01 1.09810688e-01 -4.60934490e-01 -9.47965741e-01
7.07907200e-01 -8.10928762e-01 -6.31143153e-01 -2.50378370e-01
-3.22894543e-01 3.88123989e-01 1.17947209e+00 -7.02902436e-01
1.06624174e+00 -2.26765156e+00 -1.54808983e-01 1.91020459e-01
3.03214211e-02 2.50711918e-01 -1.98807955e-01 9.81335938e-02
-2.86474079e-01 -4.00538951e-01 -3.82373929e-01 -9.02289376e-02
-5.55708766e-01 1.78119600e-01 -4.14593667e-01 1.01397455e+00
8.13766965e-04 3.26612711e-01 -9.58139479e-01 -5.11656463e-01
3.13744754e-01 4.10695195e-01 -4.44033504e-01 4.21914011e-02
3.19313586e-01 7.13276803e-01 -4.43907708e-01 9.77969944e-01
1.08611119e+00 -2.16470271e-01 -2.07726628e-01 -4.41058040e-01
-2.40032017e-01 -3.77200276e-01 -1.78272796e+00 1.60030806e+00
1.22557161e-02 6.77712798e-01 3.50640506e-01 -1.17119229e+00
7.03679383e-01 1.24166995e-01 6.88788414e-01 -4.35090125e-01
-4.09652703e-02 -2.64983047e-02 -6.47027120e-02 -4.96510535e-01
7.36766577e-01 -1.40177101e-01 5.58039665e-01 -2.40497857e-01
-5.01649737e-01 2.26197183e-01 4.70414191e-01 3.66873384e-01
1.26543295e+00 3.12492624e-02 3.83860558e-01 -1.18371256e-01
9.48572278e-01 -1.81751788e-01 1.24695849e+00 8.73510838e-01
-1.86632827e-01 9.57319736e-01 -3.06515843e-01 -4.84313339e-01
-5.61499953e-01 -1.04592741e+00 -8.85639936e-02 7.86456645e-01
4.21366185e-01 -1.40249580e-01 -9.55461487e-02 -3.16993356e-01
-2.96218127e-01 4.73086596e-01 -1.48275822e-01 9.75437909e-02
-7.89590359e-01 -1.01673460e+00 -8.19059461e-02 2.81884849e-01
5.99753380e-01 -5.01049757e-01 -4.96734083e-01 3.38419676e-01
-5.25689960e-01 -1.54196310e+00 -5.84156632e-01 -2.25788474e-01
-1.18529713e+00 -9.66848135e-01 -8.85968685e-01 -7.91579068e-01
1.23760796e+00 1.21723771e+00 8.48770201e-01 1.59435086e-02
-5.33259749e-01 5.74241579e-01 -4.64825422e-01 -2.28941023e-01
2.50421494e-01 -7.11984754e-01 3.87109101e-01 6.76236868e-01
5.63439839e-02 -5.57768643e-01 -7.70058751e-01 5.26432633e-01
-1.09925926e+00 -1.35647506e-01 5.46212137e-01 7.48789668e-01
7.93240905e-01 4.81214792e-01 9.71835405e-02 -8.17502677e-01
1.11009339e-02 -4.53021109e-01 -7.41531134e-01 -1.02016358e-02
-2.06472259e-02 -4.92967993e-01 6.43074095e-01 -4.90118772e-01
-1.27403557e+00 4.12753433e-01 4.52829272e-01 -6.61345899e-01
-3.07942510e-01 3.97043020e-01 -3.48454595e-01 -3.56718987e-01
4.99517053e-01 5.99820852e-01 -4.66214865e-01 -3.58605206e-01
1.64217636e-01 1.92730635e-01 6.93987131e-01 -4.77208734e-01
1.44043112e+00 1.19440150e+00 1.64736733e-01 -1.34453511e+00
-8.30314815e-01 -9.18293059e-01 -3.52863789e-01 -1.86179399e-01
4.92845178e-01 -1.26631975e+00 -2.71485120e-01 3.86004686e-01
-8.56469810e-01 3.77889752e-01 -2.74773121e-01 8.12614441e-01
-2.79727846e-01 1.01343679e+00 -3.54606003e-01 -9.02248383e-01
3.37345861e-02 -8.44136417e-01 7.45157123e-01 3.18079650e-01
1.96868494e-01 -7.75819838e-01 -1.78529665e-01 4.30172950e-01
4.82352674e-01 5.24065614e-01 2.64354497e-01 -3.62916172e-01
-1.04793084e+00 -3.30334187e-01 -1.63243622e-01 3.45823795e-01
5.26126802e-01 -1.50819525e-01 -6.89669967e-01 -5.55895686e-01
3.71527880e-01 2.28549629e-01 9.09196913e-01 7.93483496e-01
1.10695422e+00 -3.66321802e-01 -3.70355159e-01 8.56133580e-01
1.46267080e+00 2.05933556e-01 7.05800176e-01 2.66308039e-01
9.50302124e-01 4.46439385e-01 8.78010929e-01 7.68142939e-01
-7.12536126e-02 3.62885475e-01 3.83174419e-01 -2.06147209e-01
8.95493999e-02 2.04791427e-01 6.50113463e-01 1.09935153e+00
-2.57814229e-01 -1.75710186e-01 -5.97112894e-01 7.51600564e-01
-1.96792889e+00 -1.38084114e+00 -5.73816061e-01 2.44744396e+00
2.86815822e-01 7.94726610e-02 -9.38666016e-02 9.47237089e-02
9.55667198e-01 4.70386237e-01 -5.36326826e-01 3.39871347e-01
-6.36178136e-01 -1.65341362e-01 6.95443511e-01 1.99799553e-01
-1.12550724e+00 7.12635815e-01 6.12212563e+00 4.72940475e-01
-9.67391968e-01 8.16077068e-02 3.64515990e-01 -1.48633331e-01
5.01749665e-03 2.02595294e-01 -7.21563935e-01 4.94358331e-01
4.76351172e-01 -2.10701719e-01 1.50234833e-01 6.84715211e-01
6.75191045e-01 -3.04532677e-01 -7.23411620e-01 1.34235990e+00
3.99706542e-01 -1.02884340e+00 1.39866367e-01 -1.20250858e-01
1.26584101e+00 -1.77193075e-01 1.89510942e-01 -6.43199310e-02
1.21041656e-01 -8.14566672e-01 2.74041444e-01 6.46512151e-01
1.60233945e-01 -4.18710053e-01 5.64176500e-01 3.62538040e-01
-1.22042167e+00 5.79687068e-04 -7.75147557e-01 -1.53186783e-01
3.45615894e-01 1.14661014e+00 -2.75511712e-01 6.07950509e-01
7.17480719e-01 1.26007879e+00 -4.67425495e-01 1.57874632e+00
6.64298832e-02 7.74038911e-01 -3.12716126e-01 6.62785947e-01
7.70219564e-02 -5.72293222e-01 1.09657502e+00 1.19436288e+00
5.30050337e-01 4.79056507e-01 5.48187613e-01 4.28758591e-01
1.63003907e-01 1.18739791e-01 -7.71213293e-01 1.35829732e-01
7.94079825e-02 1.24146307e+00 -9.80612099e-01 -3.91971588e-01
-8.17162097e-01 1.10321522e+00 -4.67124909e-01 8.28212380e-01
-8.38228345e-01 -2.26750039e-02 7.08375871e-01 1.54865816e-01
5.85198939e-01 -6.19333863e-01 2.08938241e-01 -1.69426715e+00
1.71007380e-01 -1.20995224e+00 3.90714645e-01 -5.69075048e-01
-1.27840686e+00 9.07741413e-02 -8.84058625e-02 -1.73722875e+00
2.73676723e-01 -3.24269086e-01 -7.10816205e-01 4.51153696e-01
-1.67594159e+00 -8.65570009e-01 -5.90954602e-01 1.07311773e+00
1.03418815e+00 -1.18094608e-01 2.47294620e-01 4.67113942e-01
-7.74316847e-01 2.82990411e-02 5.59672058e-01 4.71938886e-02
7.87843466e-01 -7.52854347e-01 -1.49579167e-01 1.64584959e+00
3.42116594e-01 9.50823128e-01 1.03334689e+00 -7.66054809e-01
-1.82512820e+00 -1.31704032e+00 2.23507062e-01 -8.28379765e-02
4.10395414e-01 1.87305715e-02 -1.16201472e+00 6.53339028e-01
1.02220915e-01 7.84966052e-01 6.81393206e-01 -2.47394785e-01
-6.49370402e-02 -2.19697237e-01 -9.18899000e-01 4.04613525e-01
1.19130516e+00 -2.79142614e-02 -5.83576143e-01 5.94077885e-01
1.88170016e-01 -4.05514598e-01 -3.85748744e-01 5.60895443e-01
2.09055707e-01 -8.82496655e-01 1.20038450e+00 -5.36562383e-01
-1.01411231e-01 -1.10968959e+00 -5.50753653e-01 -1.05078661e+00
-5.96936941e-01 -7.82296300e-01 -3.27626109e-01 1.10071421e+00
-2.66713917e-01 -5.03661692e-01 7.64893830e-01 3.24343026e-01
3.13121006e-02 6.53563291e-02 -7.90132761e-01 -1.09327173e+00
-7.58997917e-01 -3.00854206e-01 3.62509978e-03 1.30248559e+00
-5.94293892e-01 -5.59769236e-02 -9.85212624e-01 7.80986369e-01
1.13732696e+00 2.61293381e-01 9.91152585e-01 -1.17863297e+00
-1.28400654e-01 7.35128522e-02 -8.38800669e-01 -1.09331405e+00
9.04128626e-02 -4.64828402e-01 2.55461574e-01 -1.45470321e+00
3.47801954e-01 2.02199295e-02 -5.93803942e-01 -2.31546303e-03
-4.94528353e-01 4.50484782e-01 1.34246200e-01 4.64486539e-01
-1.00511205e+00 5.85753977e-01 8.57519627e-01 -3.54294002e-01
-1.36478558e-01 -6.53706491e-02 -5.11419892e-01 1.20579827e+00
4.18252856e-01 -5.84219456e-01 -2.96262771e-01 -4.76623982e-01
3.58677730e-02 -1.34859476e-02 2.46912107e-01 -1.25920725e+00
3.79277080e-01 -5.28028905e-01 6.69342399e-01 -7.44022608e-01
2.63429940e-01 -1.23685956e+00 4.27886158e-01 3.44978541e-01
1.80786848e-01 -3.98808122e-02 7.54638836e-02 1.03125048e+00
-5.85603297e-01 -7.19383359e-02 9.41683948e-01 -2.24266425e-01
-1.03968048e+00 4.01764631e-01 -5.04141510e-01 5.31139933e-02
1.05441093e+00 -6.44400477e-01 -1.12196011e-02 -5.43357730e-01
-6.53650284e-01 6.11370653e-02 3.60516936e-01 2.75670260e-01
9.98025775e-01 -1.29062021e+00 -9.71662641e-01 1.80857584e-01
5.22442535e-02 -5.16153164e-02 5.25384128e-01 1.15880215e+00
-4.26893264e-01 1.19048327e-01 -1.96183667e-01 -7.39932656e-01
-1.50389969e+00 7.19859362e-01 -8.65100920e-02 1.92475140e-01
-7.95854330e-01 6.98400199e-01 5.63063860e-01 3.98478210e-01
1.61508977e-01 -3.26077253e-01 -3.00901830e-02 -1.70498997e-01
8.01584125e-01 6.36804283e-01 -1.80888087e-01 -1.12485981e+00
-4.84991372e-01 7.95401454e-01 -1.28267199e-01 3.34453493e-01
1.30009317e+00 -3.44745904e-01 -1.56712756e-01 5.42700410e-01
9.08266664e-01 4.88374710e-01 -1.19802392e+00 -5.39147437e-01
4.31811959e-02 -1.25926566e+00 -5.49915619e-03 4.76201735e-02
-1.22498238e+00 5.96863031e-01 7.40188003e-01 -1.20567612e-01
1.37998569e+00 -6.30375504e-01 7.35085368e-01 5.55680931e-01
4.44677055e-01 -8.53182256e-01 1.46265909e-01 4.27226990e-01
7.33008921e-01 -1.44870710e+00 5.51606953e-01 -7.39021599e-01
-4.14732903e-01 8.31297576e-01 4.68363613e-01 -2.66844690e-01
3.82047653e-01 4.79060272e-03 -5.18661477e-02 1.45830736e-01
-3.82888615e-01 -4.58440900e-01 2.73351461e-01 6.65811896e-01
2.85866708e-01 -3.58007222e-01 -2.37607241e-01 1.73888095e-02
4.22324508e-01 -1.80339485e-01 6.79816008e-01 1.13317108e+00
-6.46993279e-01 -7.06806779e-01 -1.16973722e+00 3.88257980e-01
-6.52362406e-01 2.13981941e-02 -1.14442389e-02 3.06035161e-01
-1.77121367e-02 1.10531247e+00 -2.33020693e-01 1.49067283e-01
3.35794926e-01 -1.51300833e-01 3.01656991e-01 -6.36611462e-01
-1.01756319e-01 4.01011139e-01 -2.80017197e-01 -7.57817924e-01
-8.79521310e-01 -9.05309081e-01 -1.04774380e+00 2.57280264e-02
-3.82237911e-01 4.78483029e-02 4.25028592e-01 5.73541939e-01
1.92203388e-01 3.87768686e-01 6.91017210e-01 -9.44265962e-01
-1.43221363e-01 -5.89455068e-01 -9.15867448e-01 6.70058072e-01
5.06940722e-01 -5.62403917e-01 -6.60195827e-01 6.14261925e-01] | [9.02426815032959, -0.8147104978561401] |
df9cbdc5-344f-40ed-bb58-a6845adf9c93 | stochastic-learning-of-multi-instance | 1609.00817 | null | http://arxiv.org/abs/1609.00817v1 | http://arxiv.org/pdf/1609.00817v1.pdf | Stochastic Learning of Multi-Instance Dictionary for Earth Mover's Distance based Histogram Comparison | Dictionary plays an important role in multi-instance data representation. It
maps bags of instances to histograms. Earth mover's distance (EMD) is the most
effective histogram distance metric for the application of multi-instance
retrieval. However, up to now, there is no existing multi-instance dictionary
learning methods designed for EMD based histogram comparison. To fill this gap,
we develop the first EMD-optimal dictionary learning method using stochastic
optimization method. In the stochastic learning framework, we have one triplet
of bags, including one basic bag, one positive bag, and one negative bag. These
bags are mapped to histograms using a multi-instance dictionary. We argue that
the EMD between the basic histogram and the positive histogram should be
smaller than that between the basic histogram and the negative histogram. Base
on this condition, we design a hinge loss. By minimizing this hinge loss and
some regularization terms of the dictionary, we update the dictionary
instances. The experiments over multi-instance retrieval applications shows its
effectiveness when compared to other dictionary learning methods over the
problems of medical image retrieval and natural language relation
classification. | ['Ru-Ze Liang', 'Jihong Fan'] | 2016-09-03 | null | null | null | null | ['medical-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'medical'] | [-8.2056925e-02 -2.9419515e-01 -3.6674222e-01 -3.3760107e-01
-1.1146600e+00 -2.5426850e-01 1.8122363e-01 9.5216489e-01
-5.0520372e-01 6.0071445e-01 -4.1262735e-02 -7.2122574e-02
-3.7376615e-01 -1.0037881e+00 -4.9008131e-01 -1.1819712e+00
1.6929182e-01 6.9600308e-01 2.3779434e-01 -1.9574808e-01
3.6012626e-01 3.3666930e-01 -1.4933052e+00 2.4382868e-01
4.5147285e-01 1.1791937e+00 2.1422741e-01 2.3516752e-01
-7.7487051e-02 6.2422580e-01 -5.3546047e-01 -2.8976870e-01
3.2532942e-01 -3.8005581e-01 -4.9808666e-01 -8.6644098e-02
4.0591019e-01 -1.0315206e-02 -2.8893843e-01 1.2376968e+00
9.8860490e-01 4.4561619e-01 7.8840643e-01 -1.2858969e+00
-7.6476270e-01 3.2417843e-01 -7.8886110e-01 4.6858326e-01
3.4598863e-01 -6.5736783e-01 1.0623660e+00 -9.6695536e-01
3.8176939e-01 1.1784704e+00 5.5535847e-01 1.3168874e-01
-8.2383817e-01 -3.4331086e-01 -1.1172927e-02 4.1935992e-01
-1.7719476e+00 3.2409854e-02 7.7714908e-01 -3.4759009e-01
6.1697578e-01 4.9187467e-01 6.8575776e-01 2.9027888e-01
5.1420468e-01 7.3547214e-01 1.1086732e+00 -7.4243760e-01
1.2474026e-01 4.0729433e-01 1.3811889e-01 8.7427604e-01
7.5162306e-02 -1.9379556e-01 -3.0817577e-01 -4.0321454e-01
4.7478500e-01 3.1059602e-01 -1.7134686e-01 -5.7087779e-01
-8.6020601e-01 1.1027230e+00 8.8078924e-02 4.2121440e-01
-9.4548784e-02 -1.9259770e-01 6.3647324e-01 3.3858269e-01
4.2237896e-01 -4.5483444e-02 -2.0867864e-02 3.2566074e-01
-7.0141315e-01 2.4562149e-01 6.2897718e-01 9.0812063e-01
6.9666433e-01 -4.3574855e-01 -2.1093872e-01 1.1814157e+00
3.8522962e-01 6.2028819e-01 8.0779743e-01 -1.3433434e-01
3.6320749e-01 6.2730694e-01 -2.3591544e-01 -1.6364435e+00
-2.1639270e-01 -2.0262565e-01 -8.9815152e-01 -4.6232694e-01
1.6506158e-01 4.3909681e-01 -5.5376178e-01 1.4506426e+00
6.3052863e-01 2.8930536e-01 -1.0952688e-01 9.5894408e-01
9.8475718e-01 6.5366358e-01 -1.1273730e-01 -5.7987833e-01
1.3326755e+00 -8.1280053e-01 -7.9811060e-01 1.5182789e-01
6.2154037e-01 -7.7881294e-01 9.1903937e-01 2.5322470e-01
-9.8599166e-01 -4.6268919e-01 -1.1406311e+00 -4.0154427e-02
-6.4339107e-01 -2.0388375e-01 3.7853885e-01 5.2918571e-01
-7.1600729e-01 1.5412885e-01 -4.2131943e-01 -1.8788594e-01
-8.2248896e-02 2.1891676e-01 -3.4919602e-01 -2.0514323e-01
-1.3927846e+00 9.3351567e-01 6.0540098e-01 1.9538222e-02
-4.6768036e-01 -3.3587012e-01 -1.0281309e+00 -2.8578565e-01
-5.4124091e-02 -4.2589629e-01 6.6274565e-01 -4.2437053e-01
-8.6284697e-01 1.2031724e+00 6.8561032e-02 -3.0866468e-01
1.6013803e-01 3.5845885e-01 -4.5476693e-01 8.7284349e-02
2.5316209e-01 2.6189625e-01 9.3320781e-01 -1.2765465e+00
-8.1656528e-01 -5.8629656e-01 -3.2727893e-02 4.1505131e-01
-5.5159509e-01 -2.5823921e-01 -6.2744194e-01 -8.2236034e-01
3.1023303e-01 -7.8592581e-01 -1.5891206e-01 -3.0678180e-01
-1.7712499e-01 -2.8635275e-01 8.0817741e-01 -6.0094678e-01
1.6963083e+00 -2.5159550e+00 1.9807604e-01 5.6323868e-01
6.5021448e-02 -1.6891314e-01 7.6933317e-02 2.0038299e-01
-9.9880651e-02 -2.9913151e-01 -1.9750117e-01 -2.7949059e-01
4.7144528e-02 2.9573122e-01 -2.3468743e-01 6.8451709e-01
-5.2052116e-01 4.5246398e-01 -9.7656715e-01 -9.8886383e-01
3.1649008e-01 4.0127769e-01 -3.5182816e-01 1.7322749e-01
2.3732622e-01 2.1902283e-01 -2.6960537e-01 8.1599712e-01
7.4937266e-01 -1.9885872e-01 8.0710724e-02 -4.6999452e-01
1.6623843e-01 -3.3217087e-01 -1.5161531e+00 1.5737742e+00
-3.2980576e-01 3.5968125e-01 -3.5655406e-01 -1.4673581e+00
1.1335491e+00 2.3199201e-01 8.5061419e-01 -9.2943072e-01
1.5465043e-01 3.4468862e-01 -4.2625770e-01 -4.4225064e-01
4.7857186e-01 -6.4034617e-01 -2.5806287e-01 1.7288321e-01
-7.7151023e-02 -1.3752657e-01 3.0806011e-01 1.7244186e-02
5.4403967e-01 -6.3626808e-01 5.3394312e-01 -2.7863818e-01
7.6956630e-01 -2.7973264e-01 5.2965766e-01 6.9632125e-01
-1.8648103e-01 7.5927681e-01 1.4589863e-01 -5.3401911e-01
-8.5497004e-01 -9.7128278e-01 -6.5391541e-01 1.1035100e+00
7.4014914e-01 -4.1067541e-01 -3.0460915e-01 -4.6850577e-01
3.7136814e-01 3.1140533e-01 -5.0642926e-01 -3.8226166e-01
-4.1063461e-01 -1.0213348e+00 1.5918164e-01 3.4249321e-01
2.8733942e-01 -8.0851233e-01 -3.0219629e-01 4.0962514e-01
-2.0826976e-01 -7.2728556e-01 -8.8573730e-01 1.6789638e-01
-6.6172981e-01 -1.0278382e+00 -7.9227144e-01 -1.1647266e+00
6.1851722e-01 3.3497965e-01 1.0512824e+00 2.0627740e-01
-6.1152101e-01 6.6623408e-01 -4.1779912e-01 -4.4915882e-01
-2.4603772e-01 -2.2695842e-01 1.2496484e-01 6.3637249e-02
4.9942800e-01 -2.8228042e-01 -6.7022079e-01 3.6912096e-01
-1.1473308e+00 -4.0265763e-01 2.2290085e-01 9.9664235e-01
1.2839173e+00 4.3851176e-01 4.2557892e-01 -6.6370803e-01
7.1807897e-01 -6.4721620e-01 -6.0443217e-01 5.9479338e-01
-6.6803241e-01 -6.2826678e-02 3.8233817e-01 -5.0781417e-01
-3.6543515e-01 -1.5871696e-01 7.7116452e-02 -3.9831525e-01
2.8107277e-01 6.6151392e-01 -4.9648251e-02 -1.7455159e-01
2.7854607e-01 3.6296931e-01 -1.5256368e-01 -1.5613253e-01
2.8854182e-01 8.9675158e-01 2.3634206e-01 -5.2142525e-01
2.4720782e-01 4.9856696e-01 2.7124645e-02 -8.2366747e-01
-8.6978883e-01 -9.0072626e-01 -4.9887502e-01 -1.7228696e-01
8.9423412e-01 -6.9758064e-01 -6.5523618e-01 1.8246815e-01
-7.4219519e-01 1.1925697e-01 -1.9800310e-01 4.9402484e-01
-5.3459769e-01 4.8157606e-01 -4.3694243e-01 -5.8281386e-01
-4.1399279e-01 -1.1607914e+00 1.2631371e+00 5.0307699e-02
2.2247754e-01 -1.2828337e+00 3.7666726e-01 3.5454038e-01
1.2588005e-01 3.0785933e-01 1.1754920e+00 -7.2064835e-01
-3.9681893e-02 -4.9447340e-01 4.8124436e-02 3.6991745e-01
1.2836184e-01 -3.9772475e-01 -3.7985551e-01 -7.1854776e-01
3.2706526e-01 -1.7702448e-01 7.2695673e-01 4.2556027e-01
1.2843217e+00 -3.0132061e-01 -4.8321897e-01 3.6015606e-01
1.7552238e+00 5.5920643e-01 5.8295983e-01 5.2197164e-01
6.0718614e-01 1.6220231e-01 9.0390342e-01 7.8000402e-01
5.7734257e-01 7.0473254e-01 1.6387892e-01 -9.2898663e-03
2.9426664e-01 -8.0969870e-02 -1.4479896e-01 1.2984998e+00
4.3326885e-01 -2.9968238e-01 -9.4075656e-01 6.9927889e-01
-1.8363035e+00 -9.0332782e-01 1.4322527e-01 2.5013340e+00
8.5135990e-01 1.6081901e-01 8.9528583e-02 3.1550971e-01
9.3335187e-01 2.0704338e-01 -2.8437331e-01 -2.8914893e-01
-2.5673467e-01 1.3461337e-02 3.1760168e-01 5.6211948e-01
-1.3720021e+00 4.2009571e-01 5.7693853e+00 1.1687367e+00
-1.0807947e+00 7.4658655e-02 4.9254051e-01 1.2958457e-01
-2.8973821e-01 -5.4226726e-02 -8.8209635e-01 7.6702756e-01
5.2070302e-01 -1.7173688e-01 7.5720429e-02 7.8029841e-01
-2.0722139e-01 -2.2096431e-01 -1.0690100e+00 1.6476922e+00
4.6932057e-01 -1.0774059e+00 3.9401811e-01 -5.0718643e-02
5.7205880e-01 -5.5714029e-01 1.5914470e-01 3.0458206e-01
-2.9157498e-01 -9.5355499e-01 5.8852190e-01 6.3159758e-01
6.0865021e-01 -9.3379831e-01 9.3726218e-01 2.2119693e-01
-1.4689498e+00 -1.2139458e-03 -7.4494439e-01 3.8303128e-01
1.0714404e-01 7.8301549e-01 -4.1440281e-01 5.1433146e-01
5.3395361e-01 6.4054269e-01 -3.5664085e-01 1.2541295e+00
3.8912901e-01 4.6582613e-02 -1.7472614e-01 1.0305191e-01
3.6409214e-02 -4.7984830e-01 4.1432393e-01 1.2715762e+00
2.3395102e-01 3.6594453e-01 7.0027661e-01 7.8971073e-02
-9.4477586e-02 6.4527833e-01 -7.7057970e-01 3.8260493e-01
6.7479837e-01 1.0208850e+00 -7.2801799e-01 -4.4468510e-01
-5.1006681e-01 8.9710104e-01 2.2356451e-01 -1.1989622e-01
-9.8935556e-01 -6.1806679e-01 1.3566110e-01 -1.2306910e-01
5.9327960e-02 1.5143420e-01 -9.4677985e-02 -8.5371274e-01
6.9507197e-02 -9.7180629e-01 9.6775252e-01 -4.2825580e-01
-1.6308802e+00 4.3339133e-01 3.0388268e-02 -1.3661195e+00
1.9455296e-01 -4.0499300e-01 -1.6669625e-01 5.0495410e-01
-1.3087637e+00 -8.7088656e-01 -1.6989136e-01 8.2641274e-01
3.5812271e-01 -3.1267369e-01 8.1897002e-01 7.9379314e-01
-4.2846581e-01 8.3341557e-01 5.3649259e-01 9.0156198e-02
8.8134229e-01 -1.3524097e+00 -6.3924408e-01 2.1630372e-01
1.8553831e-01 5.2380830e-01 6.8220943e-01 -3.9734802e-01
-1.2837796e+00 -8.2937920e-01 7.4616605e-01 -3.7036359e-01
4.8829940e-01 -2.4992749e-02 -9.6707827e-01 4.6582302e-01
2.2520460e-02 2.1836346e-01 1.1294852e+00 -2.6725805e-01
-1.6872056e-01 -4.1353917e-01 -1.3481231e+00 2.6122275e-01
4.2384884e-01 -6.7833263e-01 -6.6806430e-01 8.0573195e-01
3.4848377e-01 -6.7626095e-01 -1.4067404e+00 5.4499924e-01
6.2573862e-01 -5.7233804e-01 1.1208764e+00 -5.5363685e-01
1.4766301e-01 -2.4178101e-01 -7.1549410e-01 -1.1671014e+00
-2.5222188e-01 8.4468843e-03 -9.8600537e-02 9.9603766e-01
2.0105924e-01 -4.5340261e-01 4.1071787e-01 4.2553774e-01
7.3215477e-02 -1.1227648e+00 -1.0095440e+00 -7.8747910e-01
1.5596803e-01 -3.4482084e-02 4.9196133e-01 1.1464263e+00
2.4593171e-01 1.4931428e-01 -4.0583089e-01 2.0546128e-01
7.7756751e-01 5.3876519e-01 3.9511332e-01 -1.0681082e+00
-2.5458685e-01 -2.9350311e-01 -8.1402707e-01 -9.5809191e-01
9.0038061e-02 -1.0933782e+00 4.5460664e-02 -1.5428852e+00
6.0090625e-01 -7.6604670e-01 -7.7949625e-01 2.3009756e-01
-2.7196291e-01 1.7516416e-01 5.3530127e-02 3.6189672e-01
-9.2272633e-01 4.3378827e-01 1.1903439e+00 -4.7656900e-01
-1.3436694e-01 -2.3533474e-01 -5.1660186e-01 4.4703022e-01
6.1172301e-01 -6.9003445e-01 -3.1171829e-01 -2.6477611e-01
8.2637005e-02 3.0696774e-01 6.8789944e-02 -8.0747807e-01
4.8307657e-01 -1.4752878e-01 1.8157154e-01 -7.7266151e-01
2.5653103e-01 -8.3320069e-01 1.2311499e-01 2.8961739e-01
-2.5572127e-01 3.4057260e-01 5.6653857e-02 7.9746830e-01
-6.2204176e-01 -4.2164889e-01 9.5509422e-01 -1.7358133e-01
-6.7515647e-01 4.8095024e-01 -4.6938853e-03 2.2680762e-01
1.1330925e+00 -3.1474501e-01 3.8141724e-02 -1.2156603e-01
-9.1404545e-01 2.3264690e-01 1.3448758e-01 3.8708025e-01
8.9203721e-01 -1.8084260e+00 -5.6492358e-01 2.2334720e-01
3.0991426e-01 -1.2136702e-01 1.4898784e-01 9.0789902e-01
-3.9657730e-01 4.3539023e-01 6.4375520e-02 -6.9324350e-01
-1.4579910e+00 7.7552837e-01 3.5189524e-01 -3.1187537e-01
-3.9755535e-01 7.9211861e-01 -5.9316857e-03 -4.1621721e-01
4.6139699e-01 1.0537065e-01 -2.5573823e-01 3.8005093e-01
4.8803622e-01 5.3375280e-01 3.7106961e-01 -7.4171078e-01
-4.3676376e-01 7.5581449e-01 -2.4711621e-01 3.6289997e-02
1.2020789e+00 -1.9844320e-01 -5.6365728e-01 6.6556036e-01
1.7177637e+00 -2.4479222e-02 -4.4330233e-01 -2.2090630e-01
-3.7464302e-03 -7.4366069e-01 6.6934608e-02 -1.7058989e-01
-1.1196531e+00 5.9325856e-01 1.1153667e+00 4.3335989e-01
1.1137110e+00 -2.6011687e-02 9.8159802e-01 2.5834987e-01
6.5632737e-01 -1.3718696e+00 2.5553566e-01 3.8832343e-01
8.5040188e-01 -1.3956915e+00 1.1330749e-01 -9.2626661e-03
-4.7066101e-01 9.7956967e-01 4.1110665e-01 -1.0140547e-01
1.0487651e+00 1.5968191e-02 -7.1662739e-02 -3.2549748e-01
-2.2593568e-01 2.6751210e-03 3.6897394e-01 2.5456330e-01
3.6177653e-01 9.7069651e-02 -7.1780205e-01 3.3856201e-01
-2.1177143e-02 -3.7353265e-01 1.6992236e-02 1.0390075e+00
-6.3458270e-01 -1.0344260e+00 -5.8126324e-01 6.1812550e-01
-4.1570598e-01 -1.5583849e-01 1.2270003e-01 6.9816959e-01
1.5741943e-01 8.9611238e-01 7.0388950e-02 -2.6871458e-01
4.7576442e-01 -7.1694806e-02 5.8618945e-01 -5.1952332e-01
-1.9505063e-01 -4.8327670e-03 -5.6282175e-01 -4.5815948e-02
-3.9771160e-01 -4.0408891e-01 -1.2505476e+00 -2.4490941e-01
-7.9964590e-01 4.4472221e-01 3.9623061e-01 8.5792691e-01
-6.2705629e-02 4.0679324e-02 8.1277502e-01 -2.8977051e-01
-5.8934283e-01 -5.2284384e-01 -1.0391979e+00 6.5287971e-01
3.2885638e-01 -8.3318830e-01 -4.4250524e-01 -1.2480474e-01] | [11.336073875427246, 0.9671360850334167] |
56a14a4c-27e7-427e-b404-97ccd3f2c99b | a-neural-approach-to-kgqa-via-sparql | null | null | https://openreview.net/forum?id=SshQlYuME_ | https://openreview.net/pdf?id=SshQlYuME_ | A Neural Approach to KGQA via SPARQL Silhouette Generation | Semantic parsing is a predominant approach to solve the Knowledge Graph Question Answering (KGQA) task where, natural language question is translated into a logic form such as SPARQL. Semantic parsing based solutions are mostly modular/pipelined where, noise introduced by the upstream modules for entity/relation linking makes it hard to solve the complex questions. Recently, Neural Machine Translation (NMT) based approaches have emerged that are capable of handling complex questions. However, NMT-based approaches struggle with handling the large number of test entities and relations that are unseen during training. In this work,we propose a modular two-stage neural approach which combines best of both the worlds - NMT and semantic parsing pipeline. Stage-I of our approach comprises an NMT-based seq2seq module that translates a question into a sketch of the desired SPARQL, called as SPARQL silhouette. This stage also contains a noise simulator which combines the masking scheme with an entity/relation linker in a novel manner so as to take care of unseen entities/relation without blowing up the vocabulary of seq2seq module.
Stage-II of our approach comprises a Neural Graph Search (NGS) module which aims to distil the SPARQL silhouette in order to reduce the entity/relation linking noise.
Experimental results show that, the quality of generated SPARQL silhouette is impressive for an ideal scenario where entity/relation linker is noise-free. For the realistic scenario (i.e. noisy linker), the quality of the SPARQL silhouette drops but our NGS module recovers it considerably. We show that, our proposed approach improves state-of-the-art on LC-QuAD-1 dataset by an absolute margin of $3.72 \%$ $F_1$. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['graph-question-answering'] | ['graphs'] | [ 1.67946607e-01 6.96209371e-01 2.48991922e-01 -3.66937160e-01
-1.15443730e+00 -7.84221053e-01 1.27617553e-01 4.10247803e-01
-3.83210093e-01 8.12206507e-01 -5.52754402e-02 -5.44152796e-01
-1.47699893e-01 -1.47239923e+00 -1.22296858e+00 -3.26377511e-01
9.43048671e-02 7.63727248e-01 4.10176665e-01 -5.12530386e-01
-2.65035897e-01 1.94768995e-01 -1.48402750e+00 3.87173921e-01
1.03301752e+00 1.03508973e+00 2.03425050e-01 7.35366523e-01
-7.75521219e-01 7.97386527e-01 -6.36318028e-01 -8.71334195e-01
4.63452578e-01 -4.68580365e-01 -1.00472760e+00 -5.61360776e-01
6.74987376e-01 1.63839728e-01 -2.10859284e-01 1.42426288e+00
5.36425292e-01 -8.60437676e-02 2.28894223e-02 -1.03625071e+00
-3.92633349e-01 8.40729654e-01 -2.67939508e-01 1.21046573e-01
4.38525677e-01 1.42971307e-01 1.40435016e+00 -5.69678009e-01
8.53486121e-01 1.22518492e+00 6.85640216e-01 5.70489645e-01
-9.43578124e-01 -5.20391583e-01 -8.94884951e-03 1.79084420e-01
-1.38595748e+00 -4.77755703e-02 6.03052258e-01 5.82467467e-02
1.21655965e+00 2.75697380e-01 3.27690363e-01 6.15697265e-01
8.63886103e-02 3.97451252e-01 7.22796023e-01 -2.93233097e-01
2.75376022e-01 2.43840694e-01 1.74054697e-01 9.61858690e-01
3.63314360e-01 -2.36037910e-01 -4.22749490e-01 6.92246556e-02
4.56605673e-01 -5.24976492e-01 1.33637255e-02 -1.72750264e-01
-8.01297545e-01 7.97706723e-01 5.52854836e-01 2.38715604e-01
-3.90252590e-01 1.64756313e-01 5.17765939e-01 5.08907020e-01
2.98763718e-02 3.78837407e-01 -7.02003896e-01 1.83275715e-01
-7.44367361e-01 4.07565236e-01 1.20503533e+00 1.33413017e+00
9.58153784e-01 -8.14002231e-02 -2.72657186e-01 5.62613666e-01
-2.27243919e-02 2.78094053e-01 1.85675934e-01 -6.21435225e-01
1.06177592e+00 1.01610160e+00 -2.38824382e-01 -8.84259284e-01
-4.35745835e-01 -5.80439746e-01 -6.60724103e-01 -7.74296746e-02
4.07837093e-01 -3.99990559e-01 -9.29523289e-01 1.99218702e+00
6.71231091e-01 2.32111160e-02 5.75995028e-01 9.08161819e-01
1.27435863e+00 6.93923950e-01 3.47401112e-01 8.03527236e-02
1.78502119e+00 -8.89782071e-01 -5.08820057e-01 -4.60838050e-01
7.96589732e-01 -6.14686131e-01 1.10788941e+00 1.11272410e-01
-1.26106238e+00 -5.98207474e-01 -9.57814991e-01 -3.57174814e-01
-5.76562107e-01 -1.74787447e-01 5.27962625e-01 5.76435208e-01
-1.04553616e+00 5.00982225e-01 -3.36496979e-01 -5.65199196e-01
3.76881331e-01 4.39004421e-01 -6.26652658e-01 -1.60407767e-01
-1.49191523e+00 7.54060745e-01 8.32395136e-01 2.08106130e-01
-6.97012961e-01 -7.60815084e-01 -9.81863558e-01 2.41490632e-01
9.01285946e-01 -1.03333306e+00 1.05688572e+00 -8.12292576e-01
-1.21500206e+00 7.68879950e-01 -2.13760987e-01 -5.83240569e-01
3.40619951e-01 -3.44275944e-02 -4.29694027e-01 1.18769899e-01
9.60453525e-02 5.90238214e-01 4.07897621e-01 -9.97487247e-01
-7.40901828e-01 -3.83425534e-01 4.70242739e-01 2.89486140e-01
1.93496138e-01 -1.79144844e-01 -5.82652926e-01 -2.67551333e-01
1.37693971e-01 -6.46595895e-01 -2.18451172e-01 -5.22870719e-01
-5.76364756e-01 -1.98502406e-01 4.62815970e-01 -7.96114326e-01
1.06045890e+00 -1.71832907e+00 3.28531303e-02 1.85919151e-01
8.99665728e-02 4.86572772e-01 -1.57959580e-01 6.05454564e-01
-2.12216660e-01 2.98821598e-01 -3.94616783e-01 -9.58282501e-02
5.41739687e-02 3.98268640e-01 -2.75791250e-03 -1.39833093e-01
6.80029035e-01 1.31576502e+00 -8.21365476e-01 -5.55347085e-01
-2.58843154e-01 2.27711201e-01 -7.08366811e-01 3.70033741e-01
-9.04121637e-01 2.49450803e-01 -4.63945538e-01 5.65491378e-01
8.49581778e-01 -1.19270772e-01 3.35402161e-01 -3.66784424e-01
9.42652151e-02 4.37345147e-01 -1.39599597e+00 2.10974598e+00
-4.22375739e-01 2.55155973e-02 1.26276642e-01 -1.09600258e+00
1.06375778e+00 2.49025404e-01 -5.01026064e-02 -7.26961911e-01
1.32205307e-01 2.15890884e-01 -2.10037827e-02 -7.47651756e-01
4.08200920e-01 -4.46814716e-01 -3.21724385e-01 -1.19700707e-01
3.88879240e-01 -2.22758397e-01 2.56415427e-01 3.53912264e-01
1.43147779e+00 3.12097073e-01 1.65689051e-01 -1.49998441e-01
8.48143697e-01 3.50233495e-01 7.53023684e-01 6.10012710e-01
1.05933018e-01 4.21437800e-01 9.25294995e-01 -2.91941822e-01
-9.48084831e-01 -1.03206718e+00 4.38540697e-01 6.39882267e-01
-2.87915375e-02 -3.53726745e-01 -1.10925531e+00 -1.00420713e+00
-7.28636906e-02 7.65954196e-01 -3.61477077e-01 -9.64982137e-02
-8.58078718e-01 -5.95176697e-01 8.01788568e-01 3.31280559e-01
6.41234994e-01 -1.20180249e+00 -2.52772212e-01 3.03717554e-01
-9.89929438e-02 -1.45374000e+00 -8.98491815e-02 3.34311366e-01
-8.59237432e-01 -1.14935958e+00 -9.46567282e-02 -1.00480449e+00
6.01383686e-01 -3.29757601e-01 1.30482328e+00 2.55925432e-02
-2.09989160e-01 -1.14118479e-01 -2.89968044e-01 -1.02424048e-01
-4.68249232e-01 2.68552423e-01 -5.06789207e-01 -1.07932128e-01
4.64792162e-01 -6.16749406e-01 -4.71388757e-01 -1.09896496e-01
-1.12658358e+00 -4.01964039e-02 7.16707230e-01 6.29972637e-01
7.84969389e-01 3.57917637e-01 8.10498059e-01 -1.46698081e+00
5.13326049e-01 -6.45997226e-01 -8.11038136e-01 4.09678787e-01
-5.07969022e-01 3.92491758e-01 1.09129727e+00 -5.77863418e-02
-1.04240978e+00 5.51889464e-02 -5.20503163e-01 -1.10222824e-01
-2.28543967e-01 6.13752246e-01 -7.08743751e-01 4.11589779e-02
5.92205763e-01 2.04280075e-02 -3.39078575e-01 -4.79073554e-01
6.03406489e-01 3.01902413e-01 7.96440899e-01 -6.05767369e-01
1.19968700e+00 1.12498805e-01 1.67587608e-01 -3.16951513e-01
-8.38968515e-01 -1.64705873e-01 -3.57797325e-01 1.98506892e-01
1.04126489e+00 -8.42820764e-01 -1.02078104e+00 -7.82018807e-03
-1.30155528e+00 1.22866724e-02 -4.87137437e-01 2.42460929e-02
-3.17909598e-01 2.81509399e-01 -5.33777177e-01 -6.23337626e-01
-7.39807308e-01 -1.01123166e+00 9.58451271e-01 5.49413919e-01
-3.80790308e-02 -8.43564212e-01 -7.36868680e-02 7.03378022e-01
2.35835314e-01 4.19670343e-01 1.46181273e+00 -1.07587016e+00
-8.79844368e-01 -1.86767623e-01 -5.11685729e-01 5.23357153e-01
-2.51166850e-01 -6.74072623e-01 -8.71297061e-01 -2.26537678e-02
-1.02239124e-01 -2.08033696e-01 5.14692307e-01 -3.16948324e-01
6.29872859e-01 -3.98818374e-01 1.63466275e-01 5.41436434e-01
1.89812577e+00 1.54492214e-01 7.04079926e-01 2.87744880e-01
7.63901472e-01 6.83872402e-01 5.61275303e-01 -1.27354369e-01
7.60095656e-01 2.95143306e-01 5.98948359e-01 1.83448587e-02
-2.91369051e-01 -6.66041017e-01 2.17713445e-01 8.28613043e-01
4.37723190e-01 -4.47173566e-01 -1.01171792e+00 5.25397837e-01
-1.75166249e+00 -4.50987279e-01 -3.15773666e-01 2.00811219e+00
8.69807422e-01 2.97072679e-01 -2.46083006e-01 2.33658124e-02
6.10659242e-01 -4.49372269e-02 -4.39431757e-01 -5.98417699e-01
-2.88403898e-01 7.86534488e-01 6.44148290e-01 7.34688461e-01
-7.89488316e-01 1.38837802e+00 4.28778076e+00 7.75622249e-01
-7.69453764e-01 2.18189940e-01 1.17568508e-01 2.19432458e-01
-4.83084679e-01 3.13971162e-01 -9.62476075e-01 1.47675559e-01
1.06605113e+00 5.56983203e-02 4.87734973e-01 5.45494378e-01
-3.30023319e-02 -2.57844239e-01 -1.04335666e+00 6.35090888e-01
-8.38912278e-02 -1.26999283e+00 2.15198889e-01 -4.94274527e-01
3.86296242e-01 -1.11140937e-01 -3.99852306e-01 5.91463208e-01
2.97509044e-01 -1.03303826e+00 5.81997573e-01 4.90485698e-01
7.14388490e-01 -8.67933273e-01 1.12101984e+00 4.45934057e-01
-1.28844786e+00 9.10285767e-03 -3.02651823e-01 1.65224910e-01
1.98640540e-01 5.67538857e-01 -9.56669390e-01 1.27946746e+00
6.90382242e-01 8.33455380e-03 -4.67167139e-01 6.47581697e-01
-5.80774307e-01 3.06078881e-01 -2.77884334e-01 2.46076081e-02
2.74406046e-01 -1.62047267e-01 5.94351649e-01 1.12782121e+00
2.08937854e-01 2.84055829e-01 -9.79746357e-02 1.03604341e+00
-4.87324625e-01 3.15891773e-01 -4.70323831e-01 -1.51898175e-01
3.31337601e-01 1.32506251e+00 -4.19294000e-01 -3.22883993e-01
-3.97411078e-01 9.92099643e-01 6.87393546e-01 3.14085752e-01
-8.18215251e-01 -6.53169334e-01 4.12358940e-01 1.07946843e-01
4.74548310e-01 1.65088668e-01 -1.35104313e-01 -1.01565993e+00
4.30358708e-01 -1.09426904e+00 7.72598743e-01 -6.54622853e-01
-9.46375370e-01 8.42495680e-01 -2.86276162e-01 -6.12957597e-01
-7.54981861e-02 -3.82330328e-01 -5.18006682e-01 1.21464396e+00
-1.72181475e+00 -1.23182678e+00 -4.12743419e-01 4.67912227e-01
4.13226813e-01 -4.77525853e-02 8.25365543e-01 5.63672066e-01
-6.39523745e-01 7.08765328e-01 -5.68660975e-01 1.39555365e-01
4.44625735e-01 -1.31367874e+00 5.98479867e-01 9.81058896e-01
-2.96491180e-02 5.80080330e-01 6.55243576e-01 -9.07781839e-01
-1.79197145e+00 -1.38655090e+00 1.15454197e+00 -2.35865265e-01
6.61436439e-01 -4.81383681e-01 -9.98679459e-01 6.66314781e-01
3.14429030e-02 9.84981805e-02 2.95736462e-01 -1.43342689e-01
-4.45396453e-01 -3.85982156e-01 -1.47398627e+00 3.03087085e-01
1.03989077e+00 -5.54708838e-01 -6.10582054e-01 1.89401239e-01
1.04863751e+00 -5.72272956e-01 -1.06091058e+00 6.45351708e-01
6.82118386e-02 -7.76262343e-01 7.42353141e-01 -7.70573556e-01
2.39132434e-01 -6.02211297e-01 -2.64493078e-01 -9.56140459e-01
1.02399349e-01 -7.43833184e-01 -6.29347470e-03 1.36236405e+00
7.66822040e-01 -6.20718539e-01 1.16295969e+00 4.65020239e-01
-8.72990340e-02 -7.05578804e-01 -1.08618116e+00 -6.14905953e-01
2.15393621e-02 -4.18379992e-01 7.04462230e-01 7.64557958e-01
-3.03980917e-01 7.54785180e-01 1.00284472e-01 6.16333604e-01
4.61101055e-01 6.33299649e-02 7.29796708e-01 -1.07154751e+00
-4.11178559e-01 -1.62978843e-02 -3.38979155e-01 -7.81018913e-01
3.75658907e-02 -1.12573206e+00 -1.00243263e-01 -1.75504804e+00
-1.12565823e-01 -4.75561023e-01 -2.87334602e-02 4.83868957e-01
-1.92918211e-01 -1.77582249e-01 1.80178478e-01 -3.07890564e-01
-4.45054591e-01 2.07104549e-01 1.10779703e+00 -3.91896404e-02
-1.17701516e-01 -2.83305556e-01 -8.96310270e-01 4.06204104e-01
7.63010681e-01 -7.53859997e-01 -5.85872948e-01 -5.58337331e-01
7.08253562e-01 3.19059402e-01 3.78071398e-01 -8.97315264e-01
4.17054057e-01 1.78161189e-01 -1.12702057e-01 -4.54254389e-01
1.05444588e-01 -8.92726958e-01 4.02916759e-01 2.81041592e-01
-8.28036144e-02 2.34458968e-01 4.54081506e-01 5.53488016e-01
-3.65818143e-01 -4.47516620e-01 7.02131629e-01 -4.42218930e-01
-5.96854508e-01 2.49659061e-01 3.17140818e-01 5.42845666e-01
9.21182394e-01 7.16661885e-02 -4.68783557e-01 -1.84316009e-01
-7.31024086e-01 4.53625202e-01 1.40182719e-01 3.25339407e-01
3.36650640e-01 -9.68597591e-01 -6.69535458e-01 8.71004686e-02
-2.52218172e-02 6.00564718e-01 3.54622275e-01 6.78922236e-01
-8.63016129e-01 4.76892829e-01 1.71781898e-01 -6.45094067e-02
-1.07374406e+00 6.49454892e-01 4.18254226e-01 -6.86921060e-01
-3.59089047e-01 1.08138263e+00 7.82808512e-02 -7.35386014e-01
2.43825287e-01 -5.00658393e-01 -2.25798577e-01 2.85271741e-02
1.75678283e-01 7.29153231e-02 4.22276169e-01 -5.29198825e-01
-4.43527848e-01 4.53211784e-01 1.46408424e-01 7.71442279e-02
1.24753487e+00 1.77043974e-01 -3.61020446e-01 3.95653173e-02
1.16562068e+00 2.35617105e-02 -5.61116993e-01 -2.94042021e-01
3.25012833e-01 1.11602865e-01 -2.24977240e-01 -9.71298456e-01
-9.94687796e-01 8.38196695e-01 3.72534841e-01 1.15738906e-01
1.25164890e+00 7.64014646e-02 1.29523873e+00 3.56677294e-01
3.24853212e-01 -8.63439918e-01 -3.92954618e-01 5.85817337e-01
5.42515576e-01 -1.12994909e+00 -4.04734910e-01 -7.74522007e-01
-4.34940159e-01 9.41090703e-01 6.71273649e-01 -1.60187797e-03
2.62328714e-01 2.59889185e-01 -7.23584509e-03 -4.47116762e-01
-6.63299501e-01 -4.97509181e-01 2.10231140e-01 3.54709834e-01
4.08984236e-02 -1.10936716e-01 -2.98268616e-01 1.01815021e+00
-4.57549721e-01 -1.33743048e-01 1.83382407e-01 8.78249884e-01
-4.60717440e-01 -1.20691872e+00 -1.16226949e-01 1.98331565e-01
-6.75715506e-01 -5.29669344e-01 -3.22992325e-01 8.32822263e-01
4.57029581e-01 9.07756150e-01 -1.78438053e-01 -4.17988181e-01
8.36872697e-01 4.78462458e-01 3.42990905e-01 -7.64975548e-01
-1.10458732e+00 -3.67667377e-01 5.08755565e-01 -6.45677805e-01
5.96201196e-02 -2.57381946e-01 -1.74364197e+00 -1.58709690e-01
-1.33623958e-01 4.34791625e-01 6.95704341e-01 8.38044107e-01
4.23034698e-01 6.53162181e-01 1.43254817e-01 2.23990321e-01
-2.62571007e-01 -6.78841531e-01 -2.16115713e-01 3.16479445e-01
2.02122748e-01 -9.73915011e-02 -1.48567632e-01 -1.83679700e-01] | [10.011497497558594, 7.842283725738525] |
c46add53-97f0-4a01-9665-51933c6877a5 | preventing-information-leakage-with-neural | 1912.08421 | null | https://arxiv.org/abs/1912.08421v2 | https://arxiv.org/pdf/1912.08421v2.pdf | Learning to Prevent Leakage: Privacy-Preserving Inference in the Mobile Cloud | Powered by machine learning services in the cloud, numerous learning-driven mobile applications are gaining popularity in the market. As deep learning tasks are mostly computation-intensive, it has become a trend to process raw data on devices and send the deep neural network (DNN) features to the cloud, where the features are further processed to return final results. However, there is always unexpected leakage with the release of features, with which an adversary could infer a significant amount of information about the original data. We propose a privacy-preserving reinforcement learning framework on top of the mobile cloud infrastructure from the perspective of DNN structures. The framework aims to learn a policy to modify the base DNNs to prevent information leakage while maintaining high inference accuracy. The policy can also be readily transferred to large-size DNNs to speed up learning. Extensive evaluations on a variety of DNNs have shown that our framework can successfully find privacy-preserving DNN structures to defend different privacy attacks. | ['Wei Wang', 'YiXuan Wang', 'CongCong Li', 'Shuang Zhang', 'Bo Li', 'Quanshi Zhang', 'Liyao Xiang'] | 2019-12-18 | null | null | null | null | ['privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['methodology', 'natural-language-processing'] | [ 9.92803425e-02 1.70785651e-01 -2.41403952e-01 -5.31444550e-01
-6.00304723e-01 -9.88681197e-01 2.94413596e-01 -4.19508852e-03
-7.23875046e-01 8.11652541e-01 -4.07364853e-02 -5.90597034e-01
1.75381109e-01 -9.60670114e-01 -9.50980604e-01 -8.73144507e-01
7.04842433e-02 1.25423893e-01 7.64250979e-02 4.10814136e-02
-3.54239196e-02 7.89637566e-01 -1.39853740e+00 2.34016582e-01
6.73357248e-01 1.18186164e+00 4.33350205e-02 4.04071957e-01
4.62579466e-02 8.85682404e-01 -6.43167317e-01 -9.94574189e-01
8.46612334e-01 2.11292684e-01 -7.28279650e-01 -3.48254710e-01
3.75439227e-01 -8.31130683e-01 -7.40935385e-01 1.65226781e+00
4.40845221e-01 -1.86454967e-01 -3.65523934e-01 -1.67495835e+00
-5.03968894e-01 5.50027013e-01 -4.04641405e-02 3.04421782e-02
-1.76583320e-01 1.27490804e-01 8.57192755e-01 -2.49682471e-01
5.62286615e-01 8.19491684e-01 3.61916155e-01 1.03548646e+00
-7.85330474e-01 -1.15388751e+00 1.26455382e-01 1.77342981e-01
-8.37642670e-01 -4.61949944e-01 4.75942552e-01 1.76691711e-01
6.04731798e-01 5.58420300e-01 7.11867094e-01 1.14954138e+00
1.97137982e-01 9.90377486e-01 1.00013280e+00 1.61923245e-01
6.69318318e-01 3.22306663e-01 -1.18316807e-01 4.88338202e-01
5.30882180e-01 1.48298904e-01 -6.04008138e-01 -5.23338437e-01
4.25423861e-01 6.48438275e-01 -1.77081406e-01 -5.56342721e-01
-4.40490723e-01 7.71875560e-01 4.62864727e-01 -1.98786527e-01
-3.39731574e-01 3.40920627e-01 7.34452426e-01 5.32782495e-01
3.44691545e-01 6.79803789e-02 -8.11721027e-01 -2.51216263e-01
-5.85381329e-01 4.53060597e-01 8.50769222e-01 9.68010008e-01
9.32877958e-01 -1.59293771e-01 1.47868276e-01 1.61712974e-01
8.51245970e-02 4.35145795e-01 3.47073913e-01 -1.09165049e+00
5.97012281e-01 4.72380489e-01 -1.49732694e-01 -7.43324876e-01
1.88954011e-01 -3.28369111e-01 -8.58312249e-01 3.50010842e-01
2.12570742e-01 -6.34196758e-01 -4.17379349e-01 1.78701699e+00
4.67105001e-01 1.99746877e-01 4.37295467e-01 9.36861157e-01
2.85450548e-01 3.90644044e-01 6.15888573e-02 4.49179336e-02
1.24079633e+00 -6.66663110e-01 -6.46874249e-01 -1.19584732e-01
6.66788816e-01 -6.59339949e-02 7.29479194e-01 3.68113458e-01
-8.32889199e-01 1.41065821e-01 -9.76961195e-01 -2.94072598e-01
-3.82436037e-01 -2.85974085e-01 9.14801419e-01 1.06739509e+00
-1.18337142e+00 5.91983855e-01 -1.01764464e+00 1.30445389e-02
1.19130158e+00 8.37682605e-01 -5.32577991e-01 -1.86724588e-01
-1.16101503e+00 2.25377828e-01 4.23742503e-01 -2.34897695e-02
-1.08384109e+00 -8.46845686e-01 -4.91263986e-01 2.52634764e-01
4.57770199e-01 -5.54591596e-01 1.16056299e+00 -1.13358569e+00
-1.55266261e+00 6.79940581e-01 1.88040048e-01 -1.01337659e+00
8.29189479e-01 -1.78459063e-01 -3.57639849e-01 9.26293060e-02
-2.32502654e-01 4.84949350e-01 1.03876233e+00 -9.22729969e-01
-1.24059236e+00 -7.35260189e-01 2.42213234e-01 7.56032541e-02
-8.70863497e-01 -1.35902286e-01 -9.63978171e-02 -3.64944428e-01
-3.10582221e-01 -1.03563428e+00 -3.32331598e-01 3.91206890e-01
-4.29344863e-01 8.38850960e-02 1.55212629e+00 -4.82983917e-01
7.90350974e-01 -2.26625204e+00 -2.96313107e-01 3.41250867e-01
5.65227032e-01 6.58273935e-01 4.15047109e-02 -6.00736663e-02
2.62338579e-01 4.17416319e-02 -1.01932369e-01 -3.51747155e-01
-5.03850691e-02 5.69360137e-01 -8.00680995e-01 3.92535627e-01
-1.76002815e-01 8.59489977e-01 -7.81247556e-01 -3.17113586e-02
-1.34934515e-01 3.15299362e-01 -9.53771114e-01 1.75559521e-01
-2.84861028e-01 1.77422240e-01 -6.16851091e-01 5.55586874e-01
1.06788838e+00 1.80035923e-03 3.31523329e-01 3.19973499e-01
2.78408587e-01 1.85366273e-01 -6.30807340e-01 1.29268348e+00
-1.41796723e-01 5.56498766e-01 4.89657938e-01 -9.12724614e-01
7.76995540e-01 2.12043330e-01 4.97626275e-01 -4.05210614e-01
5.31481765e-02 -3.63428965e-02 -2.11825117e-01 -2.03216746e-01
4.52610791e-01 1.53163016e-01 -8.79897252e-02 3.82681727e-01
-1.07796498e-01 4.63378578e-01 -7.34658539e-01 -6.06203312e-03
1.47032785e+00 -5.08364916e-01 -1.39278471e-01 -1.43990606e-01
4.82027292e-01 -3.52045953e-01 9.55348969e-01 6.59815013e-01
-6.27840817e-01 -8.92948955e-02 5.63839793e-01 -7.96155572e-01
-8.20308626e-01 -1.11116922e+00 2.12968841e-01 1.16984212e+00
-4.93236259e-02 -4.92308378e-01 -1.17405987e+00 -1.22320247e+00
2.32945532e-01 5.42052209e-01 -3.87558192e-01 -4.10914510e-01
-1.45677239e-01 -3.17033321e-01 9.64731157e-01 3.04688632e-01
9.46640968e-01 -8.78907204e-01 -6.24632120e-01 -1.03002556e-01
1.52733907e-01 -1.09121478e+00 -4.07675326e-01 2.54278272e-01
-9.44233477e-01 -9.12899375e-01 -1.16301440e-01 -5.15521228e-01
7.37072468e-01 2.40254134e-01 4.57788527e-01 -1.02123499e-01
1.65058017e-01 2.73724169e-01 4.11908850e-02 -7.08796203e-01
-4.85986829e-01 2.87192315e-01 2.22022608e-01 1.75695583e-01
6.62575424e-01 -7.63629973e-01 -4.91651386e-01 -1.56084776e-01
-1.21693254e+00 -3.15763593e-01 2.99657881e-01 5.84974170e-01
7.02763677e-01 3.79659295e-01 4.13786262e-01 -1.17527688e+00
6.54956400e-01 -4.26709890e-01 -9.18021619e-01 1.05163813e-01
-6.16818547e-01 7.73876607e-02 1.11281383e+00 -4.02436137e-01
-8.43669355e-01 2.75936902e-01 -2.35052988e-01 -8.83918345e-01
-1.18242562e-01 -7.64947832e-02 -7.76143551e-01 -1.88481972e-01
2.18642414e-01 3.94394279e-01 1.55322671e-01 -3.82388622e-01
3.30564260e-01 7.28650033e-01 5.49228311e-01 -3.57504934e-01
6.98722243e-01 7.11671948e-01 9.25932154e-02 -5.75114012e-01
-5.41044116e-01 1.26423702e-01 2.22293921e-02 1.16750889e-01
6.99424326e-01 -1.07138562e+00 -1.17460155e+00 6.48790300e-01
-8.75500619e-01 -1.24497391e-01 -4.35534298e-01 1.96767539e-01
-3.75210702e-01 1.74402341e-01 -6.63439155e-01 -6.78475082e-01
-6.91402614e-01 -1.20489097e+00 5.37182212e-01 3.79185736e-01
3.22841644e-01 -8.09869230e-01 -2.07728177e-01 4.90026832e-01
3.39536279e-01 -1.51062921e-01 1.01993752e+00 -1.12570035e+00
-1.05515409e+00 -2.04359040e-01 9.08137020e-03 6.64641857e-01
1.25237137e-01 -2.85910368e-01 -1.19904435e+00 -5.89717507e-01
2.29932070e-01 -2.59473681e-01 5.61809838e-01 1.02060772e-02
1.93218565e+00 -1.05149984e+00 -1.64579973e-01 1.13321269e+00
1.35613239e+00 1.27000362e-01 6.99242294e-01 2.37434402e-01
8.85173917e-01 2.15492487e-01 2.97761440e-01 6.34057462e-01
2.44417146e-01 9.56408456e-02 1.06377101e+00 2.06816778e-01
7.20853329e-01 -6.82976365e-01 4.35291976e-01 4.30075377e-01
2.35180199e-01 -4.40506712e-02 -4.84219790e-01 1.46813944e-01
-1.82505107e+00 -9.35132921e-01 5.27256012e-01 2.17919374e+00
7.92501211e-01 4.81133349e-02 3.80124263e-02 -1.98981151e-01
5.32180011e-01 2.22393155e-01 -1.19004560e+00 -5.60724258e-01
1.05380073e-01 1.59073472e-01 1.17919409e+00 3.77434008e-02
-1.10318172e+00 1.06554830e+00 5.78472376e+00 7.49934554e-01
-1.38304603e+00 1.51570767e-01 7.97347307e-01 -3.73548388e-01
-5.86439431e-01 -4.09099683e-02 -8.45238566e-01 6.11236155e-01
9.62942481e-01 -4.55364048e-01 7.71514237e-01 1.49741721e+00
-2.69057155e-01 5.26469231e-01 -9.06697810e-01 1.07701099e+00
-6.31982148e-01 -1.56571400e+00 1.41398206e-01 4.10103828e-01
4.73774493e-01 3.46894145e-01 3.67023885e-01 3.08969855e-01
7.35093653e-01 -8.24430466e-01 3.79139394e-01 2.98775941e-01
7.13932335e-01 -1.50848246e+00 4.50029850e-01 6.64665997e-01
-4.89141107e-01 -5.10228217e-01 -7.86306560e-01 -2.95293089e-02
-5.80310106e-01 4.61972237e-01 -8.16912234e-01 2.36922264e-01
1.08216155e+00 5.55863798e-01 -3.61232787e-01 5.36242485e-01
-1.45072490e-01 6.02920830e-01 -2.58713156e-01 -1.82963461e-01
1.30644679e-01 -3.77683640e-01 4.88940448e-01 5.25419474e-01
2.78845519e-01 -1.18887778e-02 -9.69378427e-02 6.45014465e-01
-9.10265565e-01 -6.21043183e-02 -9.27708864e-01 -9.03495029e-02
7.36468613e-01 1.30175614e+00 -2.89097220e-01 -1.79729946e-02
-3.53479415e-01 1.28345859e+00 6.20259523e-01 2.06329226e-01
-5.78212500e-01 -1.22140460e-01 1.78586316e+00 -2.07330529e-02
4.78358775e-01 9.73197073e-02 -2.58368909e-01 -1.19751656e+00
1.87833831e-01 -1.10683024e+00 4.62388575e-01 -3.50643069e-01
-1.34494388e+00 4.51281339e-01 -6.22791350e-01 -9.86745715e-01
-5.00368588e-02 -4.37371433e-01 -4.19975758e-01 4.96242613e-01
-1.35349154e+00 -7.39827216e-01 2.48086169e-01 1.10857010e+00
-3.59825753e-02 -5.82210541e-01 9.36409295e-01 2.67118692e-01
-5.72829127e-01 1.16598129e+00 6.11381590e-01 4.91347700e-01
4.15191263e-01 -9.25253749e-01 4.27635550e-01 1.05527616e+00
2.50337362e-01 6.12292886e-01 2.66689122e-01 -5.69247723e-01
-1.83455837e+00 -1.56475794e+00 5.59320331e-01 -1.32277355e-01
3.68097246e-01 -8.41730297e-01 -8.75472546e-01 9.73716080e-01
-8.12086016e-02 5.62380672e-01 8.45968544e-01 -2.73795158e-01
-5.28682351e-01 -7.40484059e-01 -1.69524002e+00 7.49964714e-01
9.05395508e-01 -9.41709220e-01 -6.90656481e-03 2.47549579e-01
1.07896996e+00 -4.72324818e-01 -6.73235238e-01 -1.64567083e-01
2.79627949e-01 -9.59686100e-01 6.88227057e-01 -1.15791571e+00
1.72488421e-01 -6.32027760e-02 -2.37145811e-01 -1.10931671e+00
1.16812713e-01 -1.12944865e+00 -4.03113097e-01 1.20412338e+00
1.39448583e-01 -9.92558002e-01 1.51445556e+00 1.23338425e+00
4.83993053e-01 -5.42387068e-01 -1.38221514e+00 -4.83008355e-01
6.07606061e-02 -4.15585876e-01 1.08291960e+00 9.99363244e-01
-3.51457238e-01 -1.17048323e-01 -5.07295907e-01 5.83401680e-01
7.75758862e-01 -1.18544146e-01 7.73154616e-01 -1.05034971e+00
-1.04106784e-01 1.07808396e-01 -5.98553598e-01 -8.51127207e-01
5.73626280e-01 -1.13894093e+00 -3.40347975e-01 -7.69449949e-01
-3.83946137e-03 -4.80693430e-01 -6.60191894e-01 8.35943341e-01
1.15507275e-01 -3.11548293e-01 3.07226032e-01 3.45272906e-02
-5.49860358e-01 5.73855460e-01 9.64316249e-01 6.38154596e-02
-4.99849499e-04 6.55312359e-01 -1.07738519e+00 6.46204889e-01
9.82419252e-01 -8.08975756e-01 -7.90956616e-01 -3.79723519e-01
1.51578188e-01 9.00459196e-03 4.30942118e-01 -9.20474172e-01
4.26011622e-01 -7.83218145e-02 4.02338147e-01 -3.82402331e-01
5.69353066e-02 -1.66150153e+00 4.14599031e-02 6.30330801e-01
-5.07044494e-01 7.13546053e-02 -2.55789589e-02 8.61788392e-01
5.90805970e-02 1.67004317e-01 6.38643682e-01 1.21782847e-01
-5.04528880e-01 1.05956435e+00 -3.61568779e-01 3.29103880e-02
1.23182988e+00 1.13388121e-01 -3.21435958e-01 -5.59998155e-01
-2.93692201e-01 2.50794172e-01 6.87578559e-01 4.01473999e-01
6.84989989e-01 -1.23924029e+00 -2.41711572e-01 5.91865718e-01
-2.31126979e-01 2.11484998e-01 9.83809233e-02 8.07772726e-02
-3.15406471e-01 1.51598111e-01 -5.18005133e-01 -2.51850545e-01
-1.49325454e+00 9.41905141e-01 3.29598486e-01 -2.06248045e-01
-6.75890923e-01 7.38966525e-01 7.31556416e-02 -4.52882230e-01
5.33255458e-01 -4.11640793e-01 4.49257731e-01 -3.05788487e-01
8.02043080e-01 3.52176160e-01 5.85890971e-02 -6.95251375e-02
-3.82600605e-01 -4.82997984e-01 -4.74222124e-01 7.06157982e-02
1.41679704e+00 -9.61294249e-02 -1.59735739e-01 -3.97635102e-01
1.51553011e+00 7.42550567e-02 -1.75904739e+00 -3.70889217e-01
-6.50563389e-02 -5.77730298e-01 2.07057536e-01 -4.74509597e-01
-1.74710000e+00 7.34722853e-01 7.53725886e-01 -3.01929336e-04
1.39068758e+00 -4.26078886e-01 1.16213059e+00 8.37890804e-01
6.63453281e-01 -1.21528149e+00 -4.54244733e-01 3.06888223e-01
1.39728174e-01 -9.31026816e-01 -2.47080117e-01 -1.58791095e-01
-7.18007088e-01 9.75259602e-01 5.58847249e-01 -7.54094273e-02
8.93678784e-01 7.77794898e-01 1.23821296e-01 2.46789902e-01
-9.04689789e-01 4.86828119e-01 -5.71588695e-01 8.22587550e-01
-5.90318382e-01 1.74061164e-01 3.27610701e-01 9.52971697e-01
-2.77002990e-01 1.69190526e-01 3.94816279e-01 1.03192246e+00
-1.87624559e-01 -1.36397922e+00 6.89806193e-02 4.94458258e-01
-1.02243996e+00 8.47344752e-03 -2.75048882e-01 3.66348177e-01
7.53247738e-02 5.91913342e-01 1.94240436e-01 -6.62785172e-01
1.48622349e-01 2.94366648e-04 -5.89227416e-02 -9.98912603e-02
-8.98151457e-01 -5.99353611e-01 -2.31898338e-01 -9.06748891e-01
6.24325648e-02 -7.04010963e-01 -1.32750511e+00 -8.24846089e-01
2.02714339e-01 1.32231638e-01 7.74644136e-01 6.89002216e-01
7.43983507e-01 2.63482574e-02 9.70299900e-01 2.43371651e-02
-9.37545061e-01 -5.97009398e-02 -8.29741001e-01 2.53518999e-01
5.31814516e-01 1.28734395e-01 -1.78271234e-01 -3.22306633e-01] | [5.8408427238464355, 6.780174255371094] |
e2ce7e2d-4a13-4907-be72-7f8f123d4fb2 | frame-recurrent-video-super-resolution | 1801.04590 | null | http://arxiv.org/abs/1801.04590v4 | http://arxiv.org/pdf/1801.04590v4.pdf | Frame-Recurrent Video Super-Resolution | Recent advances in video super-resolution have shown that convolutional
neural networks combined with motion compensation are able to merge information
from multiple low-resolution (LR) frames to generate high-quality images.
Current state-of-the-art methods process a batch of LR frames to generate a
single high-resolution (HR) frame and run this scheme in a sliding window
fashion over the entire video, effectively treating the problem as a large
number of separate multi-frame super-resolution tasks. This approach has two
main weaknesses: 1) Each input frame is processed and warped multiple times,
increasing the computational cost, and 2) each output frame is estimated
independently conditioned on the input frames, limiting the system's ability to
produce temporally consistent results.
In this work, we propose an end-to-end trainable frame-recurrent video
super-resolution framework that uses the previously inferred HR estimate to
super-resolve the subsequent frame. This naturally encourages temporally
consistent results and reduces the computational cost by warping only one image
in each step. Furthermore, due to its recurrent nature, the proposed method has
the ability to assimilate a large number of previous frames without increased
computational demands. Extensive evaluations and comparisons with previous
methods validate the strengths of our approach and demonstrate that the
proposed framework is able to significantly outperform the current state of the
art. | ['Matthew Brown', 'Raviteja Vemulapalli', 'Mehdi S. M. Sajjadi'] | 2018-01-14 | frame-recurrent-video-super-resolution-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Sajjadi_Frame-Recurrent_Video_Super-Resolution_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Sajjadi_Frame-Recurrent_Video_Super-Resolution_CVPR_2018_paper.pdf | cvpr-2018-6 | ['multi-frame-super-resolution'] | ['computer-vision'] | [ 6.26978159e-01 -2.44234920e-01 -2.08633710e-02 -8.72076675e-02
-9.46247876e-01 -1.71148747e-01 4.27321315e-01 -4.24005747e-01
-4.68512595e-01 9.40109313e-01 2.68094927e-01 3.24118793e-01
6.06402867e-02 -7.03926444e-01 -6.71978474e-01 -6.60290360e-01
8.68743137e-02 -2.47536197e-01 7.16241241e-01 -2.15336829e-01
-3.28960158e-02 5.62876761e-01 -1.85520339e+00 5.12216747e-01
6.65914476e-01 8.50190938e-01 6.64338410e-01 8.10392618e-01
1.93264633e-01 9.97919917e-01 -3.70792389e-01 3.25301625e-02
3.50911617e-01 -7.76643574e-01 -7.26979136e-01 2.10581422e-01
8.29086781e-01 -8.06209981e-01 -4.02697206e-01 7.82133341e-01
3.97012770e-01 5.17580628e-01 -1.24474987e-01 -4.91830885e-01
-5.08570373e-01 2.56846339e-01 -8.22535574e-01 7.54272163e-01
4.59770530e-01 -9.32253301e-02 6.11228347e-01 -1.02359879e+00
9.06275690e-01 1.23763740e+00 4.67782080e-01 6.68010354e-01
-1.38558578e+00 -5.95020056e-01 1.39931515e-01 7.87702650e-02
-1.35985518e+00 -7.60757148e-01 6.50840938e-01 -9.26728621e-02
8.11691701e-01 1.61528185e-01 5.38476050e-01 9.36077893e-01
1.45607799e-01 2.85096616e-01 1.07797158e+00 -3.10848743e-01
1.26372084e-01 -3.44214350e-01 -2.72752911e-01 5.27997673e-01
-8.86116456e-03 2.89764792e-01 -7.19446778e-01 1.19501576e-01
1.40533352e+00 -1.16115339e-01 -5.72572887e-01 -1.69714596e-02
-1.25083125e+00 4.74858999e-01 3.62968236e-01 5.22169054e-01
-6.38612926e-01 8.02467689e-02 1.64540201e-01 -2.11134925e-03
6.03901565e-01 -2.13556774e-02 -9.40314606e-02 6.55918345e-02
-1.53298783e+00 2.58344501e-01 1.08371645e-01 5.29623091e-01
7.65528738e-01 3.91238958e-01 -1.35657161e-01 8.83851349e-01
-1.16695322e-01 1.02139294e-01 3.84953380e-01 -1.50621438e+00
3.77096683e-01 -1.43194541e-01 4.31600153e-01 -9.73063827e-01
-2.17683449e-01 -3.39536130e-01 -1.03194928e+00 5.52382112e-01
3.15949053e-01 6.26317132e-03 -9.10143316e-01 1.79435813e+00
3.05362076e-01 6.10167027e-01 8.13377500e-02 1.24695098e+00
7.29301989e-01 8.49541306e-01 6.49896562e-02 -7.79361486e-01
1.12450814e+00 -9.13898051e-01 -9.44256365e-01 -1.87630370e-01
-3.09963673e-01 -8.51838410e-01 5.31785131e-01 3.07780653e-01
-1.73099422e+00 -1.12607133e+00 -1.12353718e+00 -3.31217051e-01
2.43251473e-01 -7.05653206e-02 2.37424642e-01 7.73140043e-02
-1.38399959e+00 7.73604751e-01 -8.73769224e-01 -7.45721981e-02
3.87701154e-01 2.36019388e-01 -3.67904276e-01 -2.31945872e-01
-1.20461345e+00 7.39702582e-01 5.83545804e-01 1.25411272e-01
-7.76387095e-01 -6.67616606e-01 -8.68190408e-01 7.09023476e-02
2.88360894e-01 -8.91143024e-01 1.07453620e+00 -1.22049344e+00
-1.56444693e+00 4.56709683e-01 -7.63818383e-01 -5.26027322e-01
6.08500957e-01 -3.87513608e-01 -5.20556569e-01 5.42820752e-01
-3.69710810e-02 6.99064255e-01 1.12467992e+00 -1.38099015e+00
-1.04169047e+00 -1.96513124e-02 1.02666402e-02 4.29177552e-01
9.59880650e-02 2.31113255e-01 -6.86800241e-01 -8.60564888e-01
1.58105806e-01 -6.98175132e-01 -2.39761129e-01 -1.29073694e-01
2.85952955e-01 3.08670729e-01 8.71510863e-01 -7.85934687e-01
1.27443910e+00 -2.11696815e+00 5.50974309e-01 -4.34498936e-01
2.97447205e-01 4.03827339e-01 -1.29345238e-01 -9.11957696e-02
-1.33144051e-01 -8.49332511e-02 -2.39837214e-01 -5.53620338e-01
-9.20819581e-01 1.91110268e-01 -3.43273044e-01 3.85048568e-01
3.26852560e-01 6.46576583e-01 -9.25813138e-01 -5.43723345e-01
6.74468100e-01 1.17199552e+00 -2.76432276e-01 2.62605190e-01
6.15965873e-02 9.30477083e-01 -3.96002419e-02 1.46834835e-01
7.59280443e-01 -3.59785557e-01 1.63182482e-01 -4.32422191e-01
-4.02432889e-01 3.52480225e-02 -1.36849737e+00 1.88476825e+00
-5.26541412e-01 8.28202605e-01 9.27410945e-02 -5.45637608e-01
7.51122177e-01 4.82475370e-01 6.47386730e-01 -8.31627071e-01
-1.88212171e-01 7.29306862e-02 -3.65703404e-01 -2.87475497e-01
8.30638647e-01 -1.89382717e-01 4.06101286e-01 2.79923409e-01
-3.21375742e-03 2.70212084e-01 3.97588313e-01 -1.26561942e-02
7.49000072e-01 5.12673616e-01 3.38023603e-01 2.30677515e-01
8.56929481e-01 -3.77301216e-01 8.97309661e-01 4.92043078e-01
-8.38822946e-02 1.14538848e+00 -8.70886669e-02 -5.69289446e-01
-1.35071754e+00 -1.19226539e+00 -4.78139743e-02 1.00917363e+00
3.79789203e-01 -2.12533802e-01 -6.62027895e-01 -8.04512426e-02
-5.93577921e-01 4.03570354e-01 -5.40937543e-01 3.31119388e-01
-1.19190073e+00 -4.94425982e-01 8.94504562e-02 5.77653050e-01
6.90996349e-01 -9.76766050e-01 -1.05358922e+00 4.76727605e-01
-6.98126853e-01 -1.52834916e+00 -4.16297495e-01 -3.36025298e-01
-1.11310673e+00 -7.46124268e-01 -1.01687610e+00 -6.97560191e-01
4.36776727e-01 7.58091807e-01 1.11913550e+00 -2.09931973e-02
-1.64871246e-01 -1.03421234e-01 -2.66390860e-01 4.19877142e-01
-4.58886802e-01 -3.61468852e-01 2.74801888e-02 2.67226696e-01
-2.08905905e-01 -6.55512750e-01 -7.88632751e-01 2.40480796e-01
-1.20621932e+00 5.24079204e-01 4.94490653e-01 8.12385023e-01
9.47724879e-01 4.21740860e-01 4.83752072e-01 -6.01206481e-01
2.10788563e-01 -2.19084173e-01 -5.91174483e-01 7.16907606e-02
-2.89943635e-01 -1.28099471e-02 6.75316215e-01 -4.79744315e-01
-1.64750373e+00 1.26659736e-01 2.07794622e-01 -7.81888068e-01
-1.22521721e-01 6.11483678e-02 2.76094258e-01 -2.08520852e-02
4.25548553e-01 4.29938257e-01 4.31504920e-02 -4.70725179e-01
4.49495912e-01 1.59717262e-01 1.12870932e+00 -1.38021365e-01
8.24929893e-01 7.35129595e-01 1.43319547e-01 -8.13109159e-01
-9.14240360e-01 -3.00666839e-01 -9.31144476e-01 -4.10852313e-01
1.16143441e+00 -1.32744086e+00 -3.31677854e-01 3.46197307e-01
-1.12546432e+00 -1.23499177e-01 -1.42898738e-01 5.69074452e-01
-6.53906345e-01 5.36518157e-01 -8.35802257e-01 -7.14795947e-01
-3.92469376e-01 -1.12268054e+00 1.02167618e+00 5.90995550e-01
-9.86426696e-02 -6.83818102e-01 -8.13691467e-02 4.55072552e-01
6.44238830e-01 4.12103295e-01 3.40990216e-01 3.49796683e-01
-8.26412916e-01 2.93955952e-01 -3.82552713e-01 2.99197763e-01
2.06151843e-01 9.57722217e-03 -9.02988136e-01 -4.80504870e-01
1.39789939e-01 3.00263930e-02 8.67607176e-01 7.22210467e-01
7.83207476e-01 -6.41653910e-02 7.56573230e-02 6.04449213e-01
1.66252720e+00 7.14688674e-02 9.56749856e-01 3.82456183e-01
6.77056074e-01 3.53846490e-01 8.55594456e-01 2.60557890e-01
1.73168331e-01 1.02231443e+00 2.33980834e-01 -4.17626649e-01
-6.68968201e-01 2.58445665e-02 4.74626929e-01 4.62901562e-01
-7.46196747e-01 5.57226222e-03 -3.24832767e-01 4.74141389e-01
-1.96649146e+00 -1.52151251e+00 -3.03394794e-02 2.38303328e+00
7.68021703e-01 -1.07128784e-01 1.06183171e-01 9.54809710e-02
8.20833027e-01 4.33643639e-01 -3.88188481e-01 -8.64477456e-02
-3.80464196e-01 5.05001187e-01 2.77268320e-01 6.88244045e-01
-9.84189630e-01 8.49065125e-01 6.17122412e+00 5.88357806e-01
-1.24341989e+00 1.10115878e-01 6.94534957e-01 -4.40202832e-01
6.21891096e-02 -1.12904981e-01 -6.26282871e-01 3.42968702e-01
1.00739205e+00 -1.76302671e-01 6.63107991e-01 3.04020166e-01
6.16838753e-01 -4.79699761e-01 -9.16431069e-01 1.04375970e+00
1.40740007e-01 -1.54189277e+00 7.21777603e-02 -7.62360245e-02
8.95097494e-01 -4.37785573e-02 1.10949613e-01 -1.88384548e-01
-1.10320047e-01 -1.13663685e+00 8.12911093e-01 6.68138623e-01
1.12976718e+00 -1.01108205e+00 5.27074337e-01 1.51385311e-02
-1.59295106e+00 -1.00043520e-01 -3.33317101e-01 -7.09411735e-03
6.47235096e-01 4.36167061e-01 -3.52630317e-01 8.78798366e-01
1.08478379e+00 7.04956889e-01 -4.15113747e-01 6.19096279e-01
2.78671142e-02 1.02969870e-01 -6.49348497e-02 1.02804720e+00
3.40109766e-02 -1.98059127e-01 6.63629651e-01 1.11392426e+00
4.37041372e-01 4.06775385e-01 1.62703730e-02 8.56870472e-01
2.05629960e-01 -3.30204844e-01 -2.19034865e-01 4.22506213e-01
3.59218180e-01 1.21397102e+00 -6.88188672e-01 -4.82056975e-01
-6.38245881e-01 1.16686451e+00 1.60474896e-01 5.31366944e-01
-8.62941206e-01 -3.27715203e-02 5.95051527e-01 1.49524927e-01
6.44215763e-01 -3.12919319e-01 -9.16278660e-02 -1.17852354e+00
1.04582772e-01 -8.21291208e-01 4.15068179e-01 -1.01135731e+00
-7.74671793e-01 9.89497244e-01 -6.50503784e-02 -1.55215633e+00
-6.50098145e-01 5.95925376e-02 -1.53227955e-01 1.14543712e+00
-1.83556330e+00 -8.98966193e-01 -5.97988486e-01 7.27257371e-01
1.05449986e+00 1.88053653e-01 6.20285988e-01 3.58988583e-01
-4.74165320e-01 1.92561984e-01 -5.17824069e-02 -5.32414019e-02
7.60313332e-01 -7.68021107e-01 2.71482944e-01 1.37607002e+00
-1.27643600e-01 4.40483242e-01 9.07047093e-01 -5.59307992e-01
-1.04015148e+00 -1.05326307e+00 6.67237699e-01 -7.19280094e-02
1.78376630e-01 1.16038978e-01 -1.29821754e+00 4.11878794e-01
1.83852375e-01 4.29633766e-01 2.23562196e-01 -4.37461257e-01
-1.99427068e-01 -1.89013213e-01 -1.09513831e+00 4.98418182e-01
8.78333747e-01 -3.69600773e-01 -4.29887980e-01 -2.82012075e-01
7.06960320e-01 -6.55481219e-01 -1.06651735e+00 4.62821752e-01
5.86030662e-01 -1.54769289e+00 1.30335391e+00 2.85492148e-02
7.78129756e-01 -7.39498198e-01 -8.07929486e-02 -8.53730023e-01
-5.07087827e-01 -6.86771929e-01 -4.74765390e-01 1.05189753e+00
-5.76277301e-02 -3.06397229e-01 3.89571100e-01 6.20805264e-01
2.87735641e-01 -5.01273572e-01 -9.87562060e-01 -4.37161624e-01
-4.00923401e-01 -2.13620234e-02 2.72021174e-01 7.85729587e-01
-4.79742050e-01 2.59128749e-01 -7.81127930e-01 4.56256688e-01
1.04563439e+00 2.53521711e-01 4.78544652e-01 -1.04367399e+00
-3.91744554e-01 -1.74634933e-01 -1.19109102e-01 -8.50232244e-01
-6.23806864e-02 -2.23667905e-01 -2.10491940e-02 -1.31546903e+00
2.65392452e-01 -1.51758805e-01 -3.74708533e-01 1.61156043e-01
-3.65064532e-01 6.66777849e-01 4.12399113e-01 3.94684345e-01
-3.75873655e-01 1.29900455e-01 1.31485820e+00 4.34269428e-01
-4.31313574e-01 -2.54466295e-01 -4.10905302e-01 5.76570690e-01
4.53848481e-01 -1.52337149e-01 -4.63834882e-01 -4.67426091e-01
-2.26288110e-01 7.31702626e-01 4.30865049e-01 -1.07253706e+00
1.07458390e-01 -2.30136260e-01 9.90109801e-01 -4.99799341e-01
6.44593894e-01 -6.40207589e-01 6.26041830e-01 1.90155029e-01
-2.89084226e-01 -2.37423996e-03 2.82166272e-01 6.75438941e-01
-4.37341601e-01 1.26271725e-01 1.29564512e+00 -2.21073389e-01
-8.67137253e-01 2.81591475e-01 -2.06236482e-01 -3.02119792e-01
9.36914444e-01 -4.93222296e-01 -8.00585747e-02 -3.37066203e-01
-6.69899344e-01 -2.06460804e-01 7.70452738e-01 5.74238837e-01
7.46764541e-01 -1.13702404e+00 -9.25240815e-01 1.38921171e-01
-4.52219546e-01 2.31780022e-01 7.97765613e-01 6.85058594e-01
-4.50145483e-01 1.87042013e-01 -5.72165072e-01 -7.57912517e-01
-1.51775229e+00 6.08381391e-01 2.76644140e-01 -3.32434446e-01
-1.27220035e+00 4.87229139e-01 3.47114444e-01 6.15786910e-01
-7.23402202e-02 -1.32476827e-02 -5.15008807e-01 -3.12895626e-02
1.06391931e+00 4.47466552e-01 -1.70910209e-01 -1.04109371e+00
-4.88789193e-02 8.42722237e-01 -2.15329081e-01 -3.47875237e-01
1.56462812e+00 -6.66416526e-01 -2.81105284e-02 3.17555875e-01
1.03997803e+00 -1.14530131e-01 -1.69546425e+00 -3.68256986e-01
-4.35930103e-01 -8.85546982e-01 3.85543883e-01 -4.73387480e-01
-1.25581026e+00 3.62786770e-01 6.83254123e-01 -1.16922639e-01
1.69709277e+00 -2.80705899e-01 9.67356920e-01 -3.74483496e-01
3.94852042e-01 -9.08852816e-01 -2.54696198e-02 1.97495162e-01
7.26336896e-01 -1.09315956e+00 3.11428875e-01 -4.46831256e-01
-6.67753935e-01 1.40894604e+00 4.45652187e-01 -9.24570113e-02
1.90990679e-02 3.84019673e-01 -4.84524406e-02 2.65240431e-01
-8.60846341e-01 -1.61372826e-01 3.19274515e-01 4.65577304e-01
6.19261622e-01 -3.81935269e-01 -2.62782305e-01 -3.08952108e-02
1.91185534e-01 4.83582616e-01 8.57488811e-01 7.34289885e-01
-3.89140248e-01 -8.93789887e-01 -5.51504672e-01 -1.14425831e-01
-6.76125169e-01 1.96466371e-02 2.99235761e-01 5.09524167e-01
-6.09324612e-02 1.01760590e+00 1.54375479e-01 -4.66127917e-02
8.56487304e-02 -1.58261821e-01 6.27299726e-01 -2.89086431e-01
-2.29942814e-01 4.71875578e-01 -9.86419767e-02 -1.06505680e+00
-1.00470006e+00 -5.26458621e-01 -1.34895396e+00 -2.06227094e-01
9.36894715e-02 -2.84204274e-01 2.33591378e-01 7.31906950e-01
4.87308919e-01 8.38644087e-01 6.01085305e-01 -1.51367831e+00
3.27039743e-04 -6.28242314e-01 -3.09544653e-01 4.79839146e-01
8.06679010e-01 -5.82917511e-01 -2.51924902e-01 5.69356620e-01] | [11.002467155456543, -1.9012022018432617] |
cc11a34a-7db6-4203-87ba-8101c28e1423 | data-generation-for-satellite-image | 2205.14418 | null | https://arxiv.org/abs/2205.14418v1 | https://arxiv.org/pdf/2205.14418v1.pdf | Data Generation for Satellite Image Classification Using Self-Supervised Representation Learning | Supervised deep neural networks are the-state-of-the-art for many tasks in the remote sensing domain, against the fact that such techniques require the dataset consisting of pairs of input and label, which are rare and expensive to collect in term of both manpower and resources. On the other hand, there are abundance of raw satellite images available both for commercial and academic purposes. Hence, in this work, we tackle the insufficient labeled data problem in satellite image classification task by introducing the process based on the self-supervised learning technique to create the synthetic labels for satellite image patches. These synthetic labels can be used as the training dataset for the existing supervised learning techniques. In our experiments, we show that the models trained on the synthetic labels give similar performance to the models trained on the real labels. And in the process of creating the synthetic labels, we also obtain the visual representation vectors that are versatile and knowledge transferable. | ['Rachada Kongkachandra', 'Pokpong Songmuang', 'Wasit Limprasert', 'Sarun Gulyanon'] | 2022-05-28 | null | null | null | null | ['satellite-image-classification'] | ['computer-vision'] | [ 4.08549160e-01 -1.01291701e-01 -1.09977908e-01 -4.17927146e-01
-3.63687605e-01 -4.73674148e-01 5.55340767e-01 2.48137349e-03
-3.96958143e-01 8.33198488e-01 -2.09031478e-01 -3.29962432e-01
3.63327600e-02 -1.13125491e+00 -6.03070974e-01 -9.89380717e-01
2.05917194e-01 2.31915459e-01 6.40500113e-02 -1.35007352e-01
-8.66184011e-03 4.20970947e-01 -1.88777733e+00 -2.09302083e-02
9.66113865e-01 1.10267735e+00 2.40803957e-01 1.77080613e-02
-4.08065617e-01 8.02794278e-01 -1.71240330e-01 -3.63888219e-02
4.31243449e-01 -4.45261031e-01 -6.25715315e-01 4.34303790e-01
1.94182768e-01 -2.59495616e-01 -1.70356899e-01 1.28201640e+00
4.70457107e-01 1.20019309e-01 7.88795471e-01 -1.15402102e+00
-7.95581937e-01 2.64432937e-01 -6.89772546e-01 -5.69409281e-02
-2.64155716e-01 -8.30880851e-02 7.90837049e-01 -8.51554930e-01
5.61146975e-01 8.33655715e-01 4.99486297e-01 5.30509055e-01
-9.68834639e-01 -7.12984324e-01 1.26757652e-01 1.36324972e-01
-1.47297192e+00 -3.96962762e-01 9.32393551e-01 -6.07864857e-01
8.56654271e-02 1.60349056e-01 6.49963796e-01 8.04241896e-01
-2.92241633e-01 4.53878015e-01 1.30441439e+00 -6.08015418e-01
3.63551259e-01 5.15710950e-01 9.09936875e-02 5.79809606e-01
1.67809799e-01 1.68181926e-01 -3.56368087e-02 2.08495319e-01
9.68510449e-01 5.01428187e-01 -4.82850671e-01 -3.73109192e-01
-1.05401146e+00 9.07972276e-01 7.57953703e-01 5.54608643e-01
-4.02612001e-01 -2.27079511e-01 3.70901287e-01 3.10351551e-01
8.38625014e-01 1.91873565e-01 -1.99979007e-01 5.72430491e-01
-1.03946316e+00 -1.82767034e-01 3.61163139e-01 6.65912747e-01
1.09535253e+00 2.83577710e-01 1.57854378e-01 8.82240891e-01
3.66757274e-01 6.86380327e-01 6.70827389e-01 -3.33512455e-01
2.30792671e-01 9.99491930e-01 2.84097463e-01 -1.16040170e+00
-3.61879587e-01 -7.10797966e-01 -1.22680318e+00 4.60502803e-01
1.93824485e-01 -2.27430202e-02 -1.17650008e+00 1.49596775e+00
4.02461141e-01 5.80119528e-02 3.68358970e-01 9.27511632e-01
1.08914256e+00 1.07588232e+00 1.51979193e-01 -1.89854935e-01
9.66004252e-01 -1.16947830e+00 -6.17553294e-01 -3.09634835e-01
7.01292813e-01 -5.59156954e-01 1.21376777e+00 1.38788491e-01
-3.10356528e-01 -7.46672511e-01 -1.00043380e+00 2.79911339e-01
-5.99727035e-01 7.34301388e-01 6.52605057e-01 4.89817381e-01
-9.52494323e-01 4.70692188e-01 -5.36880732e-01 -4.82324183e-01
5.28520405e-01 -2.21942365e-02 -6.03619158e-01 -9.28267911e-02
-9.54731286e-01 7.78703928e-01 6.80269003e-01 4.49533373e-01
-1.07981646e+00 -9.10016224e-02 -6.75348938e-01 1.78102739e-02
2.29300469e-01 1.28455967e-01 9.33203936e-01 -1.53618824e+00
-1.12664247e+00 1.10847044e+00 2.63253808e-01 8.91088247e-02
3.84854883e-01 1.75216511e-01 -4.77672458e-01 7.73565546e-02
1.72249585e-01 6.31431520e-01 7.99689531e-01 -1.44942510e+00
-6.41006172e-01 -2.81814933e-01 2.23817620e-02 6.95373043e-02
-5.29240847e-01 -3.29301149e-01 7.15967417e-02 -5.00624597e-01
3.35091472e-01 -8.10878694e-01 -2.41610184e-01 3.51973087e-01
-2.74325043e-01 -1.52909845e-01 1.06826770e+00 -7.40198135e-01
6.75525486e-01 -2.39953136e+00 -4.68167290e-02 1.80134565e-01
-3.25452201e-02 6.37815833e-01 -1.25751704e-01 2.62627661e-01
-3.14792097e-01 1.77655458e-01 -4.04574573e-01 5.53189777e-02
-2.23895654e-01 1.49525672e-01 -3.51368785e-01 4.60464656e-01
1.75035536e-01 5.93428075e-01 -9.06276166e-01 -6.69082224e-01
2.12486446e-01 3.28209102e-01 1.87523529e-01 5.06624699e-01
-2.63849616e-01 8.76874864e-01 -6.30660176e-01 6.94239795e-01
7.52688229e-01 -2.85325557e-01 9.60552543e-02 -1.72671035e-01
-1.32013723e-01 -2.06847385e-01 -1.16220260e+00 1.46982014e+00
-4.64874595e-01 4.21838403e-01 -1.17947750e-01 -1.39303970e+00
1.34525096e+00 3.64977300e-01 1.62146002e-01 -5.30372441e-01
2.39726201e-01 4.05824244e-01 -2.72514254e-01 -6.52611136e-01
6.64781630e-02 -4.94319141e-01 2.10116595e-01 6.25906825e-01
-4.16947939e-02 4.14951220e-02 2.07534462e-01 -2.70825088e-01
3.58243108e-01 4.10702765e-01 2.00566605e-01 -5.43847606e-02
6.30032837e-01 2.92416692e-01 5.32887220e-01 2.68969357e-01
1.38677910e-01 3.31948280e-01 1.32509723e-01 -8.51915598e-01
-1.31256151e+00 -5.21503270e-01 -3.16857666e-01 9.38265979e-01
1.57741591e-01 3.16429764e-01 -6.49001181e-01 -6.87440634e-01
-3.99572849e-01 2.77679801e-01 -6.01579010e-01 1.21433280e-01
-9.43271816e-02 -7.55222142e-01 4.53052193e-01 4.76561427e-01
8.11929524e-01 -1.38972378e+00 -5.32632947e-01 1.64168477e-01
-4.09822073e-03 -7.46058822e-01 1.18981779e-01 1.67999595e-01
-9.11756873e-01 -1.03273368e+00 -9.48635936e-01 -1.13987660e+00
9.82514024e-01 5.23603022e-01 7.11643159e-01 3.28930110e-01
-9.61811543e-02 -2.89573461e-01 -7.31061041e-01 -1.69273555e-01
-3.83612841e-01 5.42222969e-02 -1.93598375e-01 4.52338427e-01
1.59649715e-01 -7.07988918e-01 -5.11958241e-01 2.65087217e-01
-1.32646394e+00 2.85818160e-01 6.97724402e-01 8.66984725e-01
5.76988041e-01 3.28686446e-01 7.63584316e-01 -8.15622211e-01
1.42462611e-01 -5.33468902e-01 -6.42832756e-01 4.60489661e-01
-3.44051391e-01 6.10565729e-02 8.14399898e-01 -5.18602788e-01
-1.15535259e+00 3.31814140e-01 -6.37921244e-02 -1.56026199e-01
-4.74536002e-01 7.13153064e-01 -1.16286121e-01 -3.04776281e-01
8.29143047e-01 4.25374269e-01 -5.52436523e-02 -6.87569320e-01
2.84089088e-01 1.22776568e+00 2.07378477e-01 -2.89704978e-01
8.73284221e-01 5.42775214e-01 -7.73440255e-03 -6.65308774e-01
-1.09613311e+00 -2.69923955e-01 -5.07438660e-01 -2.77402222e-01
8.87508929e-01 -9.97466385e-01 -1.00584552e-02 6.40986621e-01
-8.42953980e-01 -2.15584338e-01 -2.24519059e-01 5.05956888e-01
-2.03752533e-01 3.31241101e-01 -1.90249309e-01 -8.97316575e-01
-4.53628510e-01 -1.07009029e+00 9.62190390e-01 5.12985885e-01
5.35562336e-01 -8.25937569e-01 -3.29497755e-02 9.24639478e-02
5.10047019e-01 3.22634041e-01 8.97037148e-01 -6.56201363e-01
-5.19100785e-01 -1.75182432e-01 -5.01620531e-01 7.89936006e-01
2.91036874e-01 -5.85212409e-02 -9.67787027e-01 -3.23193610e-01
-4.60112989e-02 -7.29806721e-01 8.13372552e-01 -5.99623397e-02
9.74873543e-01 -2.18566909e-01 -2.60360897e-01 4.53124344e-01
1.68597710e+00 3.38039160e-01 6.56574309e-01 3.28215867e-01
7.13280559e-01 8.53661418e-01 6.15926862e-01 1.28837243e-01
8.87039825e-02 3.30196142e-01 4.51549083e-01 -5.67740142e-01
1.89084128e-01 -2.72237271e-01 -1.56577438e-01 7.75106132e-01
-2.78533965e-01 -6.99999630e-02 -1.05279148e+00 6.64301097e-01
-1.89535260e+00 -7.64322758e-01 -1.43743724e-01 2.26695728e+00
8.42724383e-01 -2.47848466e-01 -2.42811054e-01 4.19803262e-01
1.00327110e+00 2.35753193e-01 -4.47394997e-01 1.49240583e-01
-1.07400924e-01 9.08784866e-02 4.31828231e-01 1.51871890e-01
-1.40060031e+00 9.80165720e-01 5.46671629e+00 8.48924875e-01
-1.56181455e+00 1.33067250e-01 6.43981755e-01 3.77567977e-01
-1.32204652e-01 -5.61989956e-02 -5.23258209e-01 3.58311057e-01
6.12621605e-01 3.15704674e-01 2.42180526e-01 1.03525710e+00
1.30969688e-01 -1.29219502e-01 -8.03855836e-01 8.98838401e-01
4.19619791e-02 -1.09380901e+00 2.39814803e-01 -3.46397720e-02
9.70935881e-01 -1.92566693e-01 -2.06860527e-01 8.11288878e-02
1.72421053e-01 -1.07466161e+00 5.93020618e-01 5.35716534e-01
8.81230891e-01 -4.85010326e-01 1.07587206e+00 7.10960150e-01
-1.06982088e+00 -4.72644269e-02 -5.53391218e-01 -8.77886117e-02
-4.37716730e-02 5.77763498e-01 -4.57932860e-01 5.75435042e-01
7.01019764e-01 7.65343606e-01 -5.66222787e-01 1.08684707e+00
-4.73492831e-01 5.73534787e-01 -1.84798911e-01 3.04090492e-02
4.14972723e-01 -5.75587034e-01 -1.48230478e-01 7.09874570e-01
4.70421255e-01 -6.23470210e-02 4.72088188e-01 7.10283160e-01
6.88812435e-02 4.96848851e-01 -9.42851245e-01 -3.24916065e-01
3.65414828e-01 1.31167448e+00 -7.64392316e-01 -3.00981283e-01
-2.58545130e-01 6.97377920e-01 2.88569659e-01 4.29279655e-01
-5.92253327e-01 -3.97396088e-01 -1.79901823e-01 -6.56513199e-02
7.49022290e-02 -3.64168882e-02 -1.60012156e-01 -1.19916677e+00
1.49736270e-01 -7.60566473e-01 -8.55594967e-03 -8.33093405e-01
-1.30159032e+00 8.91768396e-01 -1.27548575e-01 -1.61979890e+00
4.42346977e-03 -5.66755295e-01 -4.98966873e-01 8.34245563e-01
-1.88641131e+00 -1.43718171e+00 -8.53763878e-01 4.86337721e-01
2.32280463e-01 -2.10885003e-01 1.01620400e+00 5.23240685e-01
-4.12361383e-01 9.75649431e-02 4.60580647e-01 3.41070712e-01
4.77164477e-01 -7.23132491e-01 -8.01282227e-02 6.64804697e-01
2.11367816e-01 2.25105777e-01 3.49501580e-01 -2.83884346e-01
-7.72177398e-01 -1.24612355e+00 7.92849064e-01 3.94054949e-01
4.71440434e-01 -6.46877661e-02 -9.22416747e-01 6.10370934e-01
5.06548723e-03 2.57513553e-01 8.20724666e-01 -3.35262984e-01
-2.86248237e-01 -2.81183332e-01 -1.14666760e+00 2.74767876e-01
7.47438490e-01 -3.76897097e-01 -6.44500136e-01 6.40912533e-01
4.40152228e-01 5.25847934e-02 -6.60486698e-01 3.27714056e-01
2.81640261e-01 -7.88888752e-01 6.34916544e-01 -4.56803203e-01
5.46786249e-01 -7.72074044e-01 -1.97296947e-01 -1.39712703e+00
-1.27691716e-01 2.91118234e-01 8.30197871e-01 1.35542262e+00
4.69956934e-01 -5.80138326e-01 6.32013977e-01 4.30811346e-02
1.12829842e-01 -5.30427814e-01 -7.40473032e-01 -7.38794744e-01
-1.16330609e-01 2.32985035e-01 5.67025721e-01 1.31801271e+00
-3.53045881e-01 3.54771525e-01 -5.16276956e-01 -5.66093586e-02
6.06878042e-01 3.66525710e-01 5.31504810e-01 -1.65124249e+00
-8.27775821e-02 1.04426228e-01 -3.60211819e-01 -7.14452863e-01
2.19121829e-01 -7.75102258e-01 1.16026871e-01 -1.83254480e+00
2.32777297e-01 -9.22421694e-01 -1.54516205e-01 9.09199536e-01
-1.18057886e-02 5.40838778e-01 5.50313145e-02 5.60654461e-01
-3.11108798e-01 6.76865757e-01 1.37135494e+00 -3.02361369e-01
-1.28147215e-01 -1.88062280e-01 -5.91858506e-01 7.61790276e-01
1.08882582e+00 -5.52443624e-01 -3.40036750e-01 -4.42764789e-01
1.06995068e-01 -1.88146979e-01 3.00752193e-01 -9.60519612e-01
-1.22214638e-01 -3.40235442e-01 2.27096274e-01 -4.31988448e-01
1.42892838e-01 -1.06903994e+00 2.93387532e-01 3.12055349e-01
-1.99255139e-01 -2.89663404e-01 -1.92901194e-01 5.08129418e-01
-4.80810881e-01 -4.78510827e-01 9.00468826e-01 -5.26096702e-01
-9.49169457e-01 2.56267518e-01 -8.84081349e-02 -2.88413137e-01
1.18610418e+00 -2.15569571e-01 -4.61171925e-01 -3.43961328e-01
-6.47110999e-01 2.89648604e-02 3.85757953e-01 1.58580959e-01
4.06905890e-01 -1.41760874e+00 -7.08794236e-01 3.05115819e-01
3.28773260e-01 2.40671813e-01 2.67567664e-01 5.10634720e-01
-8.97730529e-01 2.21151039e-01 -7.48411596e-01 -4.73326236e-01
-8.74914587e-01 6.44112647e-01 2.65741408e-01 -1.80593804e-01
-4.49344605e-01 4.51910049e-01 9.90349576e-02 -7.86691308e-01
2.74789751e-01 1.10598095e-01 -5.75525403e-01 2.44900897e-01
5.21543384e-01 7.23785982e-02 -2.13032253e-02 -8.60417902e-01
-1.00993074e-01 6.51138484e-01 2.44188949e-01 -4.22894321e-02
1.62823093e+00 2.10151784e-02 -3.15970153e-01 4.64961052e-01
1.08173478e+00 -4.61758107e-01 -1.05297041e+00 -4.78990287e-01
-1.69388041e-01 -4.12968218e-01 1.74036086e-01 -5.98532856e-01
-1.28616476e+00 1.15492332e+00 1.11480772e+00 2.92701095e-01
1.11382508e+00 -2.33891606e-01 4.20557916e-01 5.41107535e-01
5.60763299e-01 -8.48989427e-01 -2.06056327e-01 1.99327134e-02
7.14229107e-01 -1.40509558e+00 -1.78266540e-01 -4.68156636e-01
-5.41535258e-01 9.66983616e-01 4.79860276e-01 -1.46714911e-01
6.31078780e-01 -2.06560194e-01 6.15973949e-01 -1.45397007e-01
-1.35846615e-01 -4.16217119e-01 5.64448312e-02 6.59171999e-01
3.93765450e-01 1.79960147e-01 -7.25286365e-01 1.85292646e-01
2.33687222e-01 1.78278059e-01 3.38295519e-01 1.07616794e+00
-8.16348970e-01 -1.33252728e+00 -4.89412338e-01 4.44397300e-01
-2.31598765e-01 -1.54400155e-01 -2.04711854e-01 4.71628904e-01
4.48435396e-01 9.75537777e-01 -8.90335590e-02 -4.22063589e-01
8.16209167e-02 1.71252608e-01 4.48361076e-02 -6.44998372e-01
-2.48694569e-01 -1.41366765e-01 -3.96799743e-02 1.38684317e-01
-9.84061301e-01 -1.60038903e-01 -9.92983758e-01 4.08923738e-02
-5.79095602e-01 3.75589937e-01 8.28866303e-01 9.81149077e-01
2.57369317e-02 1.71390429e-01 8.57941687e-01 -1.06392765e+00
-6.25899553e-01 -1.24693835e+00 -9.86979008e-01 6.63704157e-01
1.30073950e-01 -7.81067073e-01 -5.14569819e-01 1.94436103e-01] | [9.795287132263184, -1.3802329301834106] |
9e2b3527-ba38-4b06-9d00-87802d7512fe | voxgraf-fast-3d-aware-image-synthesis-with | 2206.07695 | null | https://arxiv.org/abs/2206.07695v3 | https://arxiv.org/pdf/2206.07695v3.pdf | VoxGRAF: Fast 3D-Aware Image Synthesis with Sparse Voxel Grids | State-of-the-art 3D-aware generative models rely on coordinate-based MLPs to parameterize 3D radiance fields. While demonstrating impressive results, querying an MLP for every sample along each ray leads to slow rendering. Therefore, existing approaches often render low-resolution feature maps and process them with an upsampling network to obtain the final image. Albeit efficient, neural rendering often entangles viewpoint and content such that changing the camera pose results in unwanted changes of geometry or appearance. Motivated by recent results in voxel-based novel view synthesis, we investigate the utility of sparse voxel grid representations for fast and 3D-consistent generative modeling in this paper. Our results demonstrate that monolithic MLPs can indeed be replaced by 3D convolutions when combining sparse voxel grids with progressive growing, free space pruning and appropriate regularization. To obtain a compact representation of the scene and allow for scaling to higher voxel resolutions, our model disentangles the foreground object (modeled in 3D) from the background (modeled in 2D). In contrast to existing approaches, our method requires only a single forward pass to generate a full 3D scene. It hence allows for efficient rendering from arbitrary viewpoints while yielding 3D consistent results with high visual fidelity. | ['Andreas Geiger', 'Yiyi Liao', 'Michael Niemeyer', 'Axel Sauer', 'Katja Schwarz'] | 2022-06-15 | null | null | null | null | ['3d-aware-image-synthesis'] | ['computer-vision'] | [ 4.28739250e-01 -7.13730007e-02 3.94503504e-01 -2.58150280e-01
-6.68499231e-01 -6.26408517e-01 9.59426165e-01 -1.31020576e-01
3.40775065e-02 7.25854695e-01 4.24894068e-04 -2.54653066e-01
1.29394114e-01 -1.36133635e+00 -8.19738925e-01 -7.27038026e-01
1.22181572e-01 7.23514855e-01 1.62392572e-01 6.63890541e-02
-3.02159283e-02 1.21350658e+00 -1.83494043e+00 3.47565830e-01
6.72339380e-01 8.97452533e-01 1.50129959e-01 1.01638103e+00
-3.14644784e-01 6.67133749e-01 -5.50054908e-01 -8.51161405e-02
4.99615282e-01 -4.40672129e-01 -4.25048113e-01 4.01477337e-01
9.22072649e-01 -5.82823992e-01 -2.17757568e-01 6.34423971e-01
4.27373558e-01 2.13694438e-01 6.69363737e-01 -8.93817842e-01
-4.60777342e-01 5.84106445e-02 -6.31209970e-01 -2.97142684e-01
1.72559977e-01 2.75415123e-01 7.08913624e-01 -9.16506052e-01
8.94683540e-01 1.21058023e+00 7.04284310e-01 3.38624686e-01
-1.79355800e+00 -3.48634809e-01 3.53092194e-01 -4.14413482e-01
-1.26055193e+00 -3.13124150e-01 1.07864666e+00 -3.68733734e-01
1.18505919e+00 7.58466959e-01 1.22545028e+00 9.26439703e-01
2.27359071e-01 1.76153883e-01 1.11657238e+00 -3.04939806e-01
3.93567294e-01 -7.00194165e-02 -2.90740192e-01 7.60159969e-01
8.19957033e-02 1.91840276e-01 -5.36399305e-01 -3.33150715e-01
1.45745981e+00 -7.91987851e-02 -2.84938276e-01 -7.02313066e-01
-1.13685155e+00 7.63798058e-01 6.64542854e-01 -2.12218478e-01
-4.74339962e-01 7.07204580e-01 -1.11288838e-01 -4.39142287e-02
8.51328552e-01 4.82853353e-01 -2.85995096e-01 1.07685030e-01
-1.09199166e+00 6.44356370e-01 6.47934914e-01 8.27678680e-01
9.74013209e-01 4.43194866e-01 -5.57829961e-02 5.97375929e-01
2.36077905e-01 6.41954958e-01 -3.00169408e-01 -1.44925594e+00
-7.68236592e-02 4.56322193e-01 1.28965378e-01 -7.74371564e-01
-3.34395289e-01 -8.06010664e-01 -7.93820441e-01 9.82484877e-01
1.63061410e-01 2.90987324e-02 -1.16174626e+00 1.50587881e+00
7.50687718e-01 3.42121333e-01 -2.49353498e-01 8.60732794e-01
8.52873743e-01 8.01294982e-01 -2.68779874e-01 -4.01024371e-02
1.16237795e+00 -6.60994470e-01 -1.57414600e-01 -9.23692584e-02
1.53374687e-01 -8.08603644e-01 9.24936473e-01 4.88481015e-01
-1.45927763e+00 -3.73849958e-01 -1.01520813e+00 -3.49523216e-01
-7.82160312e-02 -3.44217241e-01 9.91359890e-01 5.83835423e-01
-1.18780196e+00 5.19357145e-01 -9.60992992e-01 5.12583628e-02
5.58594227e-01 1.66906178e-01 -1.69711921e-03 -2.48688504e-01
-4.80125070e-01 6.95510864e-01 -2.03645706e-01 -1.83261678e-01
-9.38274145e-01 -1.13623369e+00 -7.89784372e-01 5.46186753e-02
1.65395409e-01 -1.49282885e+00 1.21148121e+00 -6.32680297e-01
-1.58029532e+00 7.76447177e-01 -2.88550466e-01 -2.43751466e-01
6.33890986e-01 -4.62203361e-02 1.03432618e-01 1.47869721e-01
-1.62285596e-01 6.80175424e-01 8.05774868e-01 -1.77713001e+00
-2.13050663e-01 -2.29716286e-01 3.08298439e-01 4.92705375e-01
4.81336623e-01 -4.62991446e-01 -3.35272342e-01 -4.24451560e-01
5.30719340e-01 -5.53047299e-01 -4.64993954e-01 4.31224048e-01
-3.41993332e-01 5.00131845e-01 7.74784267e-01 -3.62025976e-01
4.53939140e-01 -1.82652676e+00 2.16361567e-01 2.11142480e-01
5.75421035e-01 -2.81655878e-01 1.00507602e-01 3.04114968e-01
1.56066507e-01 -1.91673897e-02 -4.02295560e-01 -7.63032854e-01
-1.33433610e-01 3.26675892e-01 -3.05308700e-01 5.21343231e-01
2.10198298e-01 8.40191782e-01 -7.19331264e-01 -2.60835499e-01
6.52626693e-01 1.26160681e+00 -1.08726740e+00 4.72287349e-02
-6.35995686e-01 8.21676493e-01 -3.60085309e-01 4.90177691e-01
9.50388134e-01 -4.87686992e-01 -4.78095487e-02 -2.60079473e-01
-3.53665173e-01 4.53816861e-01 -1.14285636e+00 1.79001141e+00
-8.46549869e-01 5.79552650e-01 3.77209872e-01 -4.51027453e-01
8.23982358e-01 1.54492185e-01 4.59088653e-01 -6.75105870e-01
4.80612777e-02 3.28068994e-02 -4.42817599e-01 2.28188246e-01
6.87613785e-01 -3.60415190e-01 1.63019508e-01 4.00946200e-01
-4.73802648e-02 -1.04292572e+00 -2.75168210e-01 1.08625323e-01
8.88223529e-01 6.10862792e-01 4.77505997e-02 -1.28752768e-01
9.87818744e-03 1.51034132e-01 1.87146023e-01 6.67646229e-01
6.99813843e-01 1.02076757e+00 3.04745257e-01 -5.26850879e-01
-1.24955046e+00 -1.33197093e+00 -2.70471543e-01 5.52406967e-01
8.44394118e-02 -2.86210090e-01 -6.66012049e-01 1.83243696e-02
-3.20649892e-02 9.52587724e-01 -5.27868509e-01 1.97935462e-01
-8.80481660e-01 -8.55055690e-01 5.86821400e-02 1.82475120e-01
3.70001674e-01 -6.89830780e-01 -1.08712244e+00 3.01065683e-01
1.53603926e-01 -9.45345759e-01 9.74063128e-02 4.10544455e-01
-1.07176638e+00 -8.79563451e-01 -6.21015429e-01 -3.35104853e-01
8.76020908e-01 3.91618103e-01 1.56211543e+00 1.01172768e-01
-3.82346988e-01 3.45032036e-01 1.50173396e-01 -7.49186724e-02
-6.02851868e-01 -2.26351246e-01 -4.07454550e-01 -3.24843019e-01
-3.93893093e-01 -1.25704956e+00 -6.13325715e-01 -1.33552402e-01
-7.32499778e-01 7.44879603e-01 8.24257508e-02 5.65948308e-01
1.09688139e+00 -6.93339184e-02 -1.55025199e-01 -1.00623727e+00
2.06542596e-01 -1.69669747e-01 -1.03528070e+00 -8.28552842e-02
-2.44705379e-01 1.49587318e-01 5.71489632e-01 -2.53558725e-01
-1.23870516e+00 3.76493298e-02 -2.76272923e-01 -6.22719765e-01
-2.87944078e-01 4.77130972e-02 9.34380293e-02 -2.87019819e-01
7.11062968e-01 2.62605667e-01 -3.38749230e-01 -5.56477010e-01
6.73145890e-01 -2.81850010e-01 5.06382465e-01 -6.87451482e-01
9.08828139e-01 8.14486027e-01 5.21608114e-01 -9.07767475e-01
-5.89035749e-01 1.78300917e-01 -6.61580980e-01 -2.70647526e-01
8.61088514e-01 -9.43843663e-01 -5.21331787e-01 3.10416132e-01
-1.35504174e+00 -5.91904998e-01 -8.13078344e-01 2.32356504e-01
-6.68441355e-01 -1.39753520e-01 -5.06953895e-01 -9.56679583e-01
-3.17070931e-02 -1.19563389e+00 1.46267998e+00 -4.51686122e-02
-2.08777696e-01 -8.57590973e-01 5.73600195e-02 3.87014858e-02
5.52661598e-01 6.55049205e-01 1.12962544e+00 3.96221638e-01
-1.36306500e+00 1.46805108e-01 -2.03184336e-01 -8.68166834e-02
-3.99582926e-03 2.36015871e-01 -1.26725376e+00 8.29691589e-02
1.64550602e-01 1.26763517e-02 6.76027596e-01 6.77798390e-01
1.12893736e+00 -1.83660969e-01 -2.05096558e-01 1.29827571e+00
1.68869185e+00 5.52313142e-02 4.39543277e-01 -9.24258027e-03
1.03159988e+00 2.64666617e-01 -7.48144761e-02 5.76504767e-01
3.22193980e-01 6.38532221e-01 8.11429560e-01 -4.09122586e-01
-5.95519900e-01 -3.20759237e-01 -3.15435141e-01 5.80666900e-01
-2.62876540e-01 -4.00366098e-01 -5.58341026e-01 1.27245501e-01
-1.28354847e+00 -7.72774100e-01 -3.10963392e-01 2.22521853e+00
6.26422882e-01 -3.81854799e-04 -1.83022365e-01 -2.35935692e-02
1.27580613e-01 5.10879993e-01 -6.78799808e-01 -4.01164562e-01
-2.44564086e-01 5.41346073e-01 5.03058970e-01 9.80054080e-01
-6.40298188e-01 7.57891238e-01 6.81730843e+00 4.95513678e-01
-1.32670152e+00 8.97979587e-02 8.15360665e-01 -7.63738334e-01
-1.12410402e+00 4.09833863e-02 -7.05063105e-01 -6.47499859e-02
5.40074706e-01 1.91851482e-01 6.68114543e-01 6.76017046e-01
2.32885242e-01 -1.88250080e-01 -9.58205700e-01 9.22588289e-01
-7.79065117e-02 -1.89284289e+00 2.22155303e-01 2.55841494e-01
8.77547562e-01 2.18022391e-01 1.09072523e-02 -2.70224690e-01
7.02350438e-01 -1.12321329e+00 9.66973722e-01 5.32693505e-01
9.12262678e-01 -7.72007167e-01 -1.49486020e-01 4.91804808e-01
-1.05371034e+00 5.44699609e-01 -2.97309816e-01 -1.48505092e-01
6.11096442e-01 9.17895973e-01 -5.86908162e-01 4.78708595e-01
5.53601980e-01 3.45479786e-01 -2.13473216e-01 7.74353623e-01
-1.11844592e-01 2.43315935e-01 -6.73041999e-01 2.64119655e-01
2.28528589e-01 -4.43789363e-01 7.02915490e-01 7.90229023e-01
5.52011371e-01 1.12076342e-01 -9.43124015e-03 1.44179738e+00
-3.07437591e-02 -2.62478590e-01 -6.70088470e-01 4.86942589e-01
2.61722744e-01 1.03706920e+00 -8.76035273e-01 -2.81194419e-01
-2.22692281e-01 9.88786757e-01 3.19711864e-01 5.62417150e-01
-8.59996796e-01 2.30134413e-01 7.55782366e-01 5.53782046e-01
5.14339983e-01 -5.94673455e-01 -6.37014210e-01 -1.07120872e+00
-3.16274986e-02 -4.87862736e-01 -3.70588094e-01 -1.11863124e+00
-9.18719590e-01 7.88243413e-01 1.38080016e-01 -9.45739985e-01
-3.33078772e-01 -3.99197400e-01 -3.66820186e-01 1.09874761e+00
-1.53124499e+00 -1.23916590e+00 -4.73759145e-01 4.02986199e-01
3.97147030e-01 4.75373358e-01 9.86343026e-01 7.01025724e-02
-3.03139375e-03 -3.04831062e-02 -4.86186817e-02 -4.61439550e-01
3.67461666e-02 -1.13045800e+00 8.40977848e-01 7.31250823e-01
3.48254114e-01 4.00691301e-01 7.00530708e-01 -5.95454872e-01
-1.46801591e+00 -9.27758992e-01 4.68679070e-01 -4.93489891e-01
1.08184136e-01 -6.61473572e-01 -8.44539702e-01 4.89971995e-01
9.67619568e-02 3.09497893e-01 3.31893891e-01 -1.61318317e-01
-3.67621958e-01 1.92146331e-01 -1.32151914e+00 7.99748242e-01
1.35149038e+00 -4.47751790e-01 1.29541948e-01 3.46594900e-01
9.10767496e-01 -8.09856117e-01 -7.33138919e-01 2.10237324e-01
4.66135532e-01 -1.44161129e+00 1.39686322e+00 -2.26476073e-01
4.71950889e-01 -4.26502287e-01 -3.87883991e-01 -1.26871371e+00
-3.99786025e-01 -6.50879502e-01 -2.82360435e-01 6.95064962e-01
9.91523489e-02 -5.39717078e-01 9.01582599e-01 5.12776196e-01
-3.94077659e-01 -7.60279357e-01 -9.54679847e-01 -3.15315902e-01
8.54123086e-02 -7.33243883e-01 8.69504094e-01 7.17586875e-01
-9.63845372e-01 1.41871080e-01 -1.58059686e-01 3.94056469e-01
8.81863058e-01 5.78136623e-01 7.49722421e-01 -1.31725025e+00
-5.99196315e-01 -5.10379732e-01 -5.85794225e-02 -1.34588063e+00
-2.08995864e-01 -7.57188439e-01 -5.62692285e-02 -1.76216078e+00
-9.24696848e-02 -8.65596473e-01 4.79716718e-01 3.40099148e-02
1.94249049e-01 6.11266434e-01 -1.44879324e-02 -1.14690445e-01
2.59786751e-02 5.38399875e-01 1.62385273e+00 1.99590862e-01
-1.96871817e-01 -1.37757495e-01 -5.19878209e-01 7.99507856e-01
5.42773187e-01 -2.73760974e-01 -6.20191097e-01 -8.08832645e-01
4.16613460e-01 2.42629036e-01 8.60279322e-01 -8.97404730e-01
-1.52945951e-01 -1.52371690e-01 7.40897417e-01 -9.32501316e-01
1.11704028e+00 -8.03234041e-01 7.63362944e-01 1.09932490e-01
-4.47273850e-02 -2.78279800e-02 2.81492800e-01 2.99923718e-01
1.42985687e-01 3.71913016e-02 7.97375143e-01 -6.13350868e-01
-1.55578226e-01 4.38030809e-01 -2.90847749e-01 -1.94424346e-01
6.37550712e-01 -4.60163027e-01 -1.52419984e-01 -3.88368905e-01
-6.62511647e-01 -3.98334354e-01 1.01293421e+00 -1.08700767e-01
4.97442037e-01 -1.31104553e+00 -6.18909776e-01 6.13097012e-01
-3.91016215e-01 7.84950316e-01 4.80948061e-01 2.31767938e-01
-1.00433791e+00 1.10599518e-01 -1.35971224e-02 -8.86358202e-01
-1.12937105e+00 2.00073391e-01 5.88328421e-01 -1.88253999e-01
-1.10271978e+00 9.74528193e-01 7.22279012e-01 -5.10602295e-01
-5.38638197e-02 -5.89058578e-01 4.62384015e-01 -2.48322874e-01
3.00918311e-01 1.30369931e-01 1.54064760e-01 -4.56329435e-01
-6.41200691e-02 7.38742292e-01 4.24567997e-01 -3.89453828e-01
1.55813670e+00 9.59950164e-02 -4.43633385e-02 4.59876716e-01
1.08972609e+00 3.79663408e-01 -1.82094753e+00 1.49486260e-02
-1.01460564e+00 -6.08247459e-01 4.16653037e-01 -7.04350650e-01
-1.25086105e+00 9.53819513e-01 1.69436380e-01 1.81147903e-01
9.63305295e-01 -2.98415534e-02 5.67994595e-01 9.40219313e-02
5.83906114e-01 -4.02410477e-01 -3.56262237e-01 3.89614344e-01
8.85362864e-01 -7.20552862e-01 4.41831797e-01 -8.53904545e-01
-1.02371678e-01 9.09195125e-01 2.58384049e-01 -4.18888509e-01
7.13541925e-01 7.49179184e-01 -1.10825203e-01 -4.10861582e-01
-6.84758365e-01 2.50088871e-01 2.79434144e-01 6.95034802e-01
3.93288374e-01 6.60179481e-02 5.61951697e-01 -2.03934520e-01
-5.76747596e-01 -3.48148584e-01 4.12335813e-01 8.61552060e-01
-2.23121971e-01 -1.03576422e+00 -4.31805253e-01 4.30437267e-01
3.19448635e-02 -2.57890195e-01 -7.82244503e-02 7.72398651e-01
9.73652750e-02 2.84743577e-01 5.08163035e-01 4.81185466e-02
1.61318675e-01 -6.82622865e-02 1.07298768e+00 -7.26233065e-01
-5.53991199e-01 5.09205639e-01 9.88542736e-02 -7.63658643e-01
-4.21653867e-01 -6.45662606e-01 -1.09247029e+00 -4.08244312e-01
-3.10217999e-02 -3.21010411e-01 8.42305958e-01 4.83254164e-01
4.43413436e-01 9.56517279e-01 3.81991655e-01 -1.54725313e+00
-1.40887061e-02 -3.99444491e-01 -5.21646976e-01 -6.66351467e-02
4.55768287e-01 -6.84345305e-01 -3.59901100e-01 -3.43256071e-02] | [9.277024269104004, -3.1850013732910156] |
ad70c211-95be-4c85-b5de-8476230a3a67 | a-read-write-memory-network-for-movie-story | 1709.09345 | null | http://arxiv.org/abs/1709.09345v4 | http://arxiv.org/pdf/1709.09345v4.pdf | A Read-Write Memory Network for Movie Story Understanding | We propose a novel memory network model named Read-Write Memory Network
(RWMN) to perform question and answering tasks for large-scale, multimodal
movie story understanding. The key focus of our RWMN model is to design the
read network and the write network that consist of multiple convolutional
layers, which enable memory read and write operations to have high capacity and
flexibility. While existing memory-augmented network models treat each memory
slot as an independent block, our use of multi-layered CNNs allows the model to
read and write sequential memory cells as chunks, which is more reasonable to
represent a sequential story because adjacent memory blocks often have strong
correlations. For evaluation, we apply our model to all the six tasks of the
MovieQA benchmark, and achieve the best accuracies on several tasks, especially
on the visual QA task. Our model shows a potential to better understand not
only the content in the story, but also more abstract information, such as
relationships between characters and the reasons for their actions. | ['Ji-Sung Kim', 'Sang-ho Lee', 'Seil Na', 'Gunhee Kim'] | 2017-09-27 | a-read-write-memory-network-for-movie-story-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Na_A_Read-Write_Memory_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Na_A_Read-Write_Memory_ICCV_2017_paper.pdf | iccv-2017-10 | ['video-story-qa'] | ['computer-vision'] | [-1.12142459e-01 1.65681526e-01 -3.11872333e-01 -2.04622343e-01
-5.08159816e-01 -6.18396223e-01 8.64293039e-01 2.11978704e-02
-3.50292146e-01 4.35514688e-01 8.76965225e-01 -2.58202195e-01
2.82262146e-01 -9.96074378e-01 -9.00661349e-01 -4.06982660e-01
4.36602443e-01 6.92902744e-01 4.53529537e-01 -3.04966867e-01
2.50054747e-01 6.04868047e-02 -1.31807542e+00 1.06619549e+00
4.24161881e-01 9.57226813e-01 3.38265598e-01 7.54440844e-01
-3.79733860e-01 1.79782951e+00 -5.11744022e-01 -4.69115168e-01
-3.71549278e-01 -5.52108049e-01 -1.34321117e+00 -1.52986929e-01
3.05258453e-01 -8.49741518e-01 -7.87947237e-01 3.71268868e-01
2.88454264e-01 5.41665375e-01 5.20849109e-01 -7.75685430e-01
-1.10044873e+00 1.05769801e+00 -3.45361620e-01 3.33138555e-01
3.52719486e-01 1.94686621e-01 1.17696416e+00 -7.55427718e-01
6.80207849e-01 1.23245251e+00 2.76502550e-01 6.70639753e-01
-8.83424759e-01 -3.13089758e-01 2.08089218e-01 6.18843377e-01
-1.04185653e+00 -5.42319655e-01 3.27965200e-01 -3.33472520e-01
1.40312672e+00 2.36235216e-01 4.56930101e-01 1.11719012e+00
6.96387962e-02 1.34962237e+00 5.44249356e-01 -3.66203859e-02
-4.63062897e-02 -2.13860899e-01 6.14901364e-01 6.95323944e-01
-2.74334162e-01 -3.76966923e-01 -9.86938655e-01 2.29932711e-01
9.32377517e-01 1.03268087e-01 -3.58988851e-01 -2.19346881e-01
-1.32968402e+00 9.87850428e-01 6.85940742e-01 6.07544146e-02
-1.02366373e-01 3.43678653e-01 5.04113138e-01 1.58432707e-01
2.23069862e-01 7.35842764e-01 -1.27917677e-01 -3.75658154e-01
-7.78317988e-01 1.33593544e-01 7.63732433e-01 9.35655475e-01
2.49299139e-01 -2.46006399e-01 -6.10861599e-01 7.87027299e-01
5.88707030e-02 2.70869732e-01 3.00168008e-01 -1.10540462e+00
6.71685994e-01 6.28650010e-01 7.85132870e-02 -8.88936937e-01
-6.13394797e-01 -2.31236219e-01 -9.68370318e-01 -3.53272676e-01
2.92989820e-01 2.93916374e-01 -9.83963728e-01 1.74839568e+00
-2.37630635e-01 -4.70454320e-02 2.08882410e-02 1.00836015e+00
1.68829501e+00 1.00465620e+00 9.30978283e-02 1.41253635e-01
1.49501443e+00 -1.68393683e+00 -7.51410961e-01 -6.08587742e-01
8.26951206e-01 -3.38732570e-01 1.22370040e+00 4.57208790e-02
-1.47446334e+00 -6.01150513e-01 -8.37158024e-01 -8.09560239e-01
-3.04213315e-01 2.34976560e-01 5.31305075e-01 -1.52310342e-01
-9.92329955e-01 2.47931719e-01 -6.72025263e-01 -4.57515925e-01
5.17261326e-01 -1.04789645e-03 -3.54022563e-01 -2.29998201e-01
-1.27382028e+00 1.03429520e+00 3.54888260e-01 9.73632634e-02
-1.02296567e+00 -5.24688423e-01 -8.74746621e-01 4.41580683e-01
4.76939172e-01 -1.02398992e+00 1.38427234e+00 -6.65602565e-01
-1.15027046e+00 9.58655834e-01 -3.86539370e-01 -5.05116343e-01
2.48752721e-02 -3.49173307e-01 -2.46019959e-01 2.23178208e-01
-1.89444095e-01 1.03145635e+00 5.10592878e-01 -1.04112899e+00
-2.79279053e-01 -1.46945432e-01 4.30415750e-01 2.26153150e-01
-2.47312397e-01 7.68258050e-02 -8.59200239e-01 -7.14072287e-01
-1.90776780e-01 -6.25472903e-01 3.16232413e-01 -1.44192412e-01
-4.33429629e-01 -8.72636959e-02 5.23528039e-01 -9.29354429e-01
1.58716142e+00 -2.32599783e+00 6.09753370e-01 -4.37964350e-01
4.77471262e-01 -4.39949520e-02 -3.99423897e-01 4.78555322e-01
1.42595932e-01 -1.52238254e-02 -2.63896137e-02 -5.72507441e-01
-6.63029253e-02 4.22715276e-01 -7.93522716e-01 -2.16112554e-01
2.45068356e-01 1.40844977e+00 -6.55664563e-01 -1.74425334e-01
-3.00672978e-01 1.99980736e-01 -4.37169343e-01 4.75248396e-01
-6.81083739e-01 2.82868564e-01 -1.49917200e-01 3.19796920e-01
3.07501227e-01 -9.59120333e-01 2.79069897e-02 -2.08597362e-01
3.11057061e-01 6.37473345e-01 -6.19513869e-01 1.66655409e+00
-3.23659897e-01 9.86813128e-01 -5.92394844e-02 -4.09384370e-01
5.13115287e-01 2.08510429e-01 -2.40062654e-01 -1.24198592e+00
1.82202291e-02 -3.77423346e-01 -2.23301306e-01 -4.97411728e-01
1.02101707e+00 2.00988308e-01 -1.18192442e-01 1.00965595e+00
-3.15579548e-02 2.48134807e-01 1.90241367e-01 7.11543500e-01
1.10978425e+00 -4.72914129e-02 3.45597044e-02 1.94254011e-01
1.75244629e-01 8.90502706e-03 1.69586137e-01 9.83343959e-01
2.04450265e-01 8.55972528e-01 7.39224672e-01 -5.61661780e-01
-1.04538953e+00 -1.15840113e+00 4.22410846e-01 1.61374378e+00
3.66261095e-01 -5.41153312e-01 -4.28199887e-01 -3.73463601e-01
-2.99942046e-01 9.23768163e-01 -8.73598456e-01 -4.17912245e-01
-6.97357476e-01 -6.07338786e-01 6.18268728e-01 1.13315535e+00
7.32949257e-01 -1.47423637e+00 -7.11769700e-01 -3.43973674e-02
-8.01827371e-01 -1.07158375e+00 -5.50312459e-01 6.75392598e-02
-6.82046652e-01 -1.03029764e+00 -5.31943023e-01 -6.21803582e-01
4.59418833e-01 3.39553356e-01 1.76149464e+00 2.52793938e-01
2.35442787e-01 3.32527518e-01 -5.36268651e-01 4.54034582e-02
-2.77593046e-01 2.40532190e-01 -5.41415453e-01 -1.82413906e-01
1.79901585e-01 -4.97929081e-02 -5.64317584e-01 3.68035793e-01
-1.04345393e+00 5.95224917e-01 4.21008050e-01 1.01920927e+00
3.54080021e-01 -2.91121930e-01 3.80669147e-01 -8.48868608e-01
7.07785845e-01 -6.14861369e-01 -2.37032305e-02 6.57349229e-01
-1.13427855e-01 8.03622156e-02 3.60247940e-01 -5.12571216e-01
-1.18466270e+00 -5.27223170e-01 -3.68584804e-02 -6.97974041e-02
3.77381518e-02 6.52222157e-01 -2.21489280e-01 5.54320514e-01
3.57455879e-01 4.68595326e-01 3.53486203e-02 -5.32369137e-01
6.59326255e-01 2.03511968e-01 6.59328878e-01 -3.32887828e-01
2.20323503e-02 3.28800440e-01 -3.49857867e-01 -3.52401495e-01
-1.10088074e+00 -2.36888632e-01 -5.08116305e-01 8.41469504e-03
1.29128742e+00 -1.04149580e+00 -7.85170197e-01 6.60826087e-01
-1.43139744e+00 -7.16990352e-01 -1.74384370e-01 -6.52206913e-02
-4.82065201e-01 9.74216685e-02 -1.27386141e+00 -4.30799842e-01
-2.44650394e-01 -1.04246092e+00 1.00624585e+00 2.88090914e-01
-6.17301166e-01 -1.01210141e+00 -1.46801114e-01 8.99016738e-01
5.71446359e-01 -3.96189123e-01 1.51940823e+00 -6.57245636e-01
-9.94385362e-01 1.26370430e-01 -5.10568559e-01 -1.47425786e-01
-3.99664283e-01 -2.91643530e-01 -8.01853478e-01 -1.56508893e-01
-3.93957853e-01 -7.95737326e-01 1.57162488e+00 2.22470433e-01
1.31672728e+00 -1.87713355e-01 -2.84568965e-01 3.85496408e-01
7.32997239e-01 5.58160730e-02 1.14895701e+00 3.89139235e-01
8.75568807e-01 4.71939772e-01 3.01633686e-01 2.91266143e-01
9.25348043e-01 7.72015393e-01 6.61947906e-01 -3.12082507e-02
-2.24705562e-01 -3.84859145e-01 3.21461022e-01 7.94871390e-01
1.46410167e-01 -8.29542935e-01 -9.24338400e-01 5.60715735e-01
-2.18657637e+00 -1.13100886e+00 -1.84587389e-01 1.82846177e+00
7.47283638e-01 -3.38754803e-02 1.07267804e-01 -5.13177216e-01
2.18298033e-01 5.84972799e-01 -4.77018237e-01 -4.13283288e-01
-5.14259696e-01 -2.67085377e-02 -6.65782019e-02 4.04636681e-01
-9.54641283e-01 9.95671332e-01 7.25756884e+00 7.38094330e-01
-8.33697021e-01 1.50805354e-01 8.72893214e-01 -5.31906605e-01
-3.54249507e-01 -1.54459417e-01 -7.10165799e-01 3.78001094e-01
7.76835322e-01 4.19454366e-01 4.08754766e-01 3.54383051e-01
-1.89411566e-01 -4.39964652e-01 -1.19605768e+00 1.00717163e+00
3.66752744e-01 -1.86226726e+00 5.64996064e-01 -2.41320789e-01
4.65494365e-01 -1.88949540e-01 1.39908940e-01 4.84992564e-01
1.55772150e-01 -1.59084010e+00 8.14311206e-01 9.03770328e-01
5.03920794e-01 -5.76524854e-01 8.42594802e-01 5.32610714e-01
-8.72818589e-01 -1.52938202e-01 -3.14365953e-01 -3.55189055e-01
4.81435776e-01 1.32876128e-01 -2.57705718e-01 1.90559626e-01
7.81268716e-01 7.05030084e-01 -7.56696582e-01 6.63279593e-01
-1.25591382e-01 5.06541371e-01 2.02638805e-01 -5.37759513e-02
1.10030815e-01 2.57398099e-01 1.03374973e-01 1.07115936e+00
5.03655262e-02 3.58536839e-01 -1.40006632e-01 9.43428516e-01
-4.10668284e-01 -2.10033953e-01 -4.11670744e-01 -4.58125979e-01
3.60701889e-01 1.06676555e+00 -6.38188958e-01 -4.21625435e-01
-5.07184863e-01 1.25732124e+00 6.44345522e-01 4.07274574e-01
-9.62373197e-01 8.83250237e-02 5.81790805e-01 1.60463527e-01
5.93247414e-01 -3.74184996e-01 -4.20775235e-01 -1.23831403e+00
-2.24937871e-01 -9.66340184e-01 6.36559129e-01 -1.37858558e+00
-1.20279241e+00 6.99077606e-01 -2.01153532e-01 -7.19866157e-01
-1.32253110e-01 -6.31000042e-01 -7.83826232e-01 6.86737835e-01
-9.51075733e-01 -1.44150662e+00 -3.91450614e-01 5.17453372e-01
6.21727645e-01 -2.16533259e-01 8.64681959e-01 1.01360589e-01
-5.67066252e-01 6.56846881e-01 -1.80016458e-01 4.18401688e-01
7.05652475e-01 -9.35911536e-01 4.51178223e-01 6.67253971e-01
2.33702198e-01 7.94067860e-01 4.33450907e-01 -5.34401238e-01
-1.20062053e+00 -9.12111878e-01 9.18052733e-01 -9.09502625e-01
5.86304009e-01 -5.30306458e-01 -1.12755454e+00 1.00137699e+00
6.45412624e-01 -5.00236034e-01 8.94606471e-01 3.47034663e-01
-6.97229862e-01 2.89457142e-01 -4.11136806e-01 5.98849297e-01
9.77465451e-01 -8.78590167e-01 -7.80886412e-01 2.01155379e-01
9.66810524e-01 -5.31814933e-01 -6.29588246e-01 1.55696362e-01
5.12668610e-01 -9.53499496e-01 9.21100259e-01 -9.34088469e-01
1.14464486e+00 -1.71814993e-01 -3.91842015e-02 -1.14808309e+00
-5.69654107e-01 -6.08094558e-02 -7.82570004e-01 1.19523597e+00
6.87855065e-01 -1.16648532e-01 6.97906315e-01 7.79355288e-01
1.54875498e-02 -8.10229123e-01 -6.74710751e-01 -3.06843787e-01
3.69399376e-02 -2.87690222e-01 5.65973163e-01 9.38923180e-01
1.27523124e-01 8.60678971e-01 -7.21683085e-01 -1.69064134e-01
-1.13174664e-02 4.99815255e-01 6.90304458e-01 -7.70524681e-01
-6.82235479e-01 -6.25061929e-01 6.39100745e-02 -1.94744885e+00
1.93508267e-01 -6.79949045e-01 -5.20183817e-02 -2.03329349e+00
9.48137939e-01 5.85956462e-02 5.43464050e-02 5.49272537e-01
-1.65008932e-01 1.75505891e-01 3.86911303e-01 2.09523812e-01
-1.27633810e+00 7.20823467e-01 1.34039617e+00 -4.06413317e-01
4.16636355e-02 -3.30255657e-01 -8.27767074e-01 3.96637738e-01
3.73890519e-01 1.45275518e-01 -5.07445931e-01 -1.29942346e+00
7.39225149e-01 2.26739526e-01 5.99350810e-01 -5.71723461e-01
5.38128734e-01 7.02788979e-02 5.21891654e-01 -8.37104678e-01
6.35392129e-01 -3.81142944e-01 1.18913621e-01 -7.10814223e-02
-7.27351904e-01 3.21067989e-01 1.95809528e-01 4.69766229e-01
-4.72709179e-01 -1.46692377e-02 3.58183742e-01 -2.58857876e-01
-1.04941356e+00 1.70029506e-01 -4.87706602e-01 5.66130802e-02
7.69293487e-01 -7.48888850e-02 -1.11944330e+00 -1.09805465e+00
-9.78757143e-01 6.30715549e-01 4.42788571e-01 7.34990060e-01
7.78975129e-01 -1.40762210e+00 -4.21052605e-01 -2.14470364e-02
3.05886030e-01 3.89892906e-02 6.55307651e-01 7.57320166e-01
-5.69553375e-01 5.58411121e-01 -2.56877303e-01 -2.77993023e-01
-1.28866363e+00 6.59360647e-01 3.25493783e-01 -3.96300882e-01
-4.73948568e-01 1.21487880e+00 5.70458949e-01 -2.60248303e-01
4.43530709e-01 -2.03772247e-01 -5.80912709e-01 2.08534837e-01
8.78962278e-01 3.57667059e-01 -3.25428426e-01 -6.01193011e-01
-1.44009754e-01 2.61526287e-01 -2.85050660e-01 1.24441348e-01
1.22314596e+00 -3.76115054e-01 -5.32131135e-01 6.07335031e-01
8.64207625e-01 -3.30669194e-01 -1.13670754e+00 -3.55279446e-01
-1.55286834e-01 -9.99835357e-02 -1.25393316e-01 -9.29092705e-01
-1.11374462e+00 1.24788666e+00 5.36659732e-02 6.29976615e-02
1.09815454e+00 4.76345867e-01 1.10999596e+00 4.91929621e-01
8.93352702e-02 -1.11339676e+00 5.79042256e-01 1.05347860e+00
1.06445527e+00 -1.26418102e+00 -2.69653857e-01 -9.50957388e-02
-1.04820323e+00 1.12866616e+00 1.07673836e+00 4.18110251e-01
2.86551654e-01 1.07646458e-01 -5.44197112e-02 -4.00702894e-01
-1.19323266e+00 -9.06053334e-02 4.91454154e-01 1.92061678e-01
4.58281696e-01 -1.13001525e-01 1.96648151e-01 1.02739668e+00
3.54960598e-02 -1.66772440e-01 5.16341448e-01 7.27300346e-01
-4.87607032e-01 -7.84095287e-01 3.50358896e-02 5.10477126e-01
-7.60552287e-02 -3.54915708e-01 -7.37553179e-01 6.19631827e-01
-1.64744109e-01 9.08528686e-01 6.03712142e-01 -5.80379486e-01
1.00895628e-01 2.01440930e-01 4.49257165e-01 -5.40256381e-01
-5.61230302e-01 -3.14194351e-01 3.46292108e-01 -7.49054432e-01
-1.72459602e-01 -2.18808249e-01 -1.39408493e+00 -7.93049514e-01
-2.70003565e-02 -1.97919115e-01 -8.13955143e-02 1.07057428e+00
5.32935023e-01 7.39193499e-01 -9.49231237e-02 -2.74730861e-01
-6.86819404e-02 -9.70490575e-01 -5.07547855e-01 5.38505197e-01
1.55524343e-01 -4.92037654e-01 1.92273736e-01 -5.04336171e-02] | [11.277138710021973, 7.816256523132324] |
c6ed5289-24b1-475c-81ca-5f61c12844e3 | fast-determinantal-point-processes-via | 1811.03717 | null | http://arxiv.org/abs/1811.03717v2 | http://arxiv.org/pdf/1811.03717v2.pdf | Fast determinantal point processes via distortion-free intermediate sampling | Given a fixed $n\times d$ matrix $\mathbf{X}$, where $n\gg d$, we study the
complexity of sampling from a distribution over all subsets of rows where the
probability of a subset is proportional to the squared volume of the
parallelepiped spanned by the rows (a.k.a. a determinantal point process). In
this task, it is important to minimize the preprocessing cost of the procedure
(performed once) as well as the sampling cost (performed repeatedly). To that
end, we propose a new determinantal point process algorithm which has the
following two properties, both of which are novel: (1) a preprocessing step
which runs in time $O(\text{number-of-non-zeros}(\mathbf{X})\cdot\log
n)+\text{poly}(d)$, and (2) a sampling step which runs in $\text{poly}(d)$
time, independent of the number of rows $n$. We achieve this by introducing a
new regularized determinantal point process (R-DPP), which serves as an
intermediate distribution in the sampling procedure by reducing the number of
rows from $n$ to $\text{poly}(d)$. Crucially, this intermediate distribution
does not distort the probabilities of the target sample. Our key novelty in
defining the R-DPP is the use of a Poisson random variable for controlling the
probabilities of different subset sizes, leading to new determinantal formulas
such as the normalization constant for this distribution. Our algorithm has
applications in many diverse areas where determinantal point processes have
been used, such as machine learning, stochastic optimization, data
summarization and low-rank matrix reconstruction. | ['Michał Dereziński'] | 2018-11-08 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 4.41354454e-01 -4.33158036e-03 2.16142207e-01 4.91971746e-02
-7.82860756e-01 -7.83707261e-01 5.30028343e-01 4.93801296e-01
-6.64953470e-01 6.69825852e-01 -2.24721566e-01 -4.07165319e-01
-4.00260210e-01 -1.21651661e+00 -8.84135127e-01 -1.10135651e+00
-1.87252387e-01 9.64949310e-01 1.53836831e-01 -1.44245148e-01
3.14872533e-01 4.25595254e-01 -1.41720331e+00 -3.64809513e-01
9.42911506e-01 9.21213210e-01 9.62658226e-03 7.24971950e-01
-2.77798027e-01 3.60005260e-01 -5.55207253e-01 -5.99162459e-01
5.37242472e-01 -6.45903528e-01 -6.39485776e-01 2.00373381e-01
-1.79786701e-02 9.82877389e-02 -4.25144844e-02 1.31785440e+00
3.78136456e-01 2.44426012e-01 9.51872766e-01 -7.50181258e-01
-1.07096151e-01 5.85884809e-01 -1.01371431e+00 -1.62836781e-03
1.18434735e-01 1.15934387e-02 8.53744745e-01 -7.82046020e-01
4.53356773e-01 9.96497929e-01 4.58881110e-01 -8.82809702e-03
-1.64297855e+00 -4.60012555e-01 -8.81799310e-02 -4.02764797e-01
-1.48402071e+00 -2.32519194e-01 5.77249408e-01 -6.49891973e-01
3.83910239e-01 5.65580606e-01 6.78507626e-01 3.40466976e-01
1.70523450e-01 4.10896599e-01 1.03886926e+00 -7.16398299e-01
5.55097342e-01 -1.64231718e-01 3.10332984e-01 6.79268539e-01
8.03520799e-01 -3.19929808e-01 -3.61530691e-01 -5.19537508e-01
8.05691361e-01 5.05943522e-02 -1.97159603e-01 -4.24549013e-01
-9.53900337e-01 8.92993927e-01 -3.38637620e-01 2.55987853e-01
-3.55479926e-01 3.94853503e-01 2.70931721e-01 1.85867801e-01
4.15766567e-01 2.91173965e-01 -2.65721291e-01 -2.76255935e-01
-1.09691513e+00 2.98264265e-01 1.00588989e+00 8.68378222e-01
1.07098424e+00 -2.43450552e-01 -1.79841056e-01 7.12828755e-01
-1.05531760e-01 1.10132682e+00 -1.28196582e-01 -9.78761971e-01
6.92481697e-01 5.82128823e-01 2.68149734e-01 -9.46631849e-01
-1.27020568e-01 -3.52997512e-01 -1.36829436e+00 5.12358248e-02
8.36031079e-01 -2.00102717e-01 -5.22488832e-01 1.89551580e+00
2.39016339e-01 -5.38474083e-01 -3.79678637e-01 3.56778175e-01
-2.50935495e-01 7.42430806e-01 -3.18968207e-01 -5.89558542e-01
1.35402167e+00 -2.81930804e-01 -3.07010263e-01 1.81882530e-02
4.87570435e-01 -8.20181429e-01 1.04822862e+00 4.56406742e-01
-1.49589646e+00 -1.45776346e-01 -6.39785647e-01 1.29042432e-01
9.46339369e-02 2.65274525e-01 5.13134539e-01 8.13211799e-01
-9.81110632e-01 5.46835005e-01 -7.73802638e-01 1.55258179e-01
2.99201965e-01 5.75989127e-01 -1.35306701e-01 -1.02720208e-01
-6.18370116e-01 2.10022613e-01 -1.45388857e-01 -1.49342455e-02
-4.29405630e-01 -5.55764914e-01 -5.81682086e-01 2.69259989e-01
6.33923948e-01 -8.00944209e-01 6.11053109e-01 -6.50319219e-01
-1.24208927e+00 8.43183756e-01 -4.95781213e-01 -4.38061148e-01
8.17541420e-01 7.33143687e-02 3.60268921e-01 1.99194223e-01
2.34068260e-01 -1.17161520e-01 1.00381160e+00 -9.54691887e-01
-4.00807202e-01 -8.03683579e-01 -2.37427756e-01 -3.63289714e-02
-9.40653384e-02 -7.37796351e-02 -5.63349009e-01 -4.79050517e-01
3.46717954e-01 -1.10221744e+00 -5.47750771e-01 -3.09480071e-01
-5.14733255e-01 -2.54129469e-01 9.86605510e-03 -3.04026097e-01
1.34400761e+00 -2.28206968e+00 4.01245743e-01 7.75698245e-01
5.36283195e-01 -3.25721949e-02 1.89209029e-01 7.39773035e-01
1.36177376e-01 1.74634114e-01 -6.51505649e-01 -4.61098135e-01
-6.03506938e-02 4.21419591e-02 -1.82967678e-01 7.49755442e-01
-2.15006620e-01 2.39390686e-01 -6.88020885e-01 -1.42583609e-01
-1.11632943e-02 1.49893016e-01 -6.78198338e-01 -2.20061108e-01
-1.37601987e-01 1.40901327e-01 -3.92049760e-01 1.25514455e-02
1.01552975e+00 -3.48704606e-01 2.78114825e-01 -1.16426358e-02
-3.36932749e-01 1.77607298e-01 -1.93232858e+00 1.20609570e+00
-2.78084874e-01 4.64941822e-02 3.64173979e-01 -1.00937450e+00
7.25403190e-01 -1.13807827e-01 7.12372601e-01 -2.32845142e-01
1.77772358e-01 4.25476283e-01 -1.76200151e-01 1.49816915e-01
4.95543242e-01 -4.35281575e-01 -3.22695881e-01 8.21413338e-01
-2.32225582e-01 -2.39471167e-01 6.17292762e-01 3.19764346e-01
1.51336396e+00 -5.60782552e-01 3.90873551e-01 -5.49639702e-01
5.75274587e-01 -9.51513052e-02 5.52442610e-01 9.14685726e-01
2.46069059e-01 4.63838786e-01 1.23271108e+00 -1.99125797e-01
-1.21969366e+00 -1.00566006e+00 -1.83615685e-02 7.08854258e-01
1.28344133e-01 -4.67641711e-01 -8.58804166e-01 -2.70498246e-01
-1.09870441e-01 5.76537669e-01 -6.84542954e-01 1.92641411e-02
-7.16740906e-01 -1.05817044e+00 3.53542745e-01 1.82881415e-01
3.89534503e-01 -5.58491409e-01 -4.81327057e-01 1.59640238e-01
1.46963835e-01 -7.10669696e-01 -8.08370471e-01 5.79191089e-01
-9.22523141e-01 -1.04306233e+00 -7.35626459e-01 -3.94542426e-01
1.09468031e+00 3.13400239e-01 8.90768349e-01 -4.29348290e-01
-8.42693523e-02 3.26076329e-01 -1.62357345e-01 -2.65356511e-01
-2.33888417e-01 6.73243776e-02 5.22401817e-02 3.28573026e-02
1.74456999e-01 -6.22346938e-01 -6.37450337e-01 2.42649972e-01
-1.14921331e+00 -1.73329502e-01 5.80867827e-01 7.99939454e-01
1.04017437e+00 3.93247813e-01 -8.50701258e-02 -1.03793550e+00
6.06698632e-01 -2.39205420e-01 -9.73391235e-01 -1.57836098e-02
-4.83228803e-01 4.55377877e-01 9.69095349e-01 -2.95943588e-01
-4.78597015e-01 1.26353949e-01 3.57608520e-03 -2.42860898e-01
3.90735805e-01 4.54045355e-01 -1.54002801e-01 6.15222342e-02
5.13448656e-01 5.63915014e-01 2.13983431e-01 -4.33543831e-01
2.83887655e-01 1.57498255e-01 2.04384953e-01 -6.78074896e-01
9.05603349e-01 7.35777617e-01 7.36197770e-01 -8.98481905e-01
-4.70604211e-01 -3.15356731e-01 -4.05872345e-01 2.06220627e-01
5.11551082e-01 -5.98079383e-01 -1.05817378e+00 4.42571938e-01
-8.79263818e-01 -2.56329805e-01 -8.34093451e-01 3.91391546e-01
-5.21110296e-01 4.72076923e-01 -3.54656339e-01 -1.08916616e+00
-3.32759708e-01 -1.13697100e+00 8.44248712e-01 -1.67491227e-01
-1.15092255e-01 -6.56938136e-01 1.64713159e-01 1.61697298e-01
2.05299929e-01 7.50249773e-02 1.28819740e+00 -4.95189965e-01
-7.61986315e-01 -3.86493504e-01 -1.70074806e-01 6.13502860e-01
-1.73999593e-02 -2.59560525e-01 -1.59467787e-01 -2.17577234e-01
4.49217945e-01 3.85846674e-01 8.24931502e-01 6.00491583e-01
1.10558033e+00 -5.92621624e-01 -2.43024692e-01 4.30855125e-01
1.51108360e+00 2.83191036e-02 7.11295426e-01 -8.60760137e-02
5.81964970e-01 3.80443245e-01 2.29603723e-01 7.59209633e-01
4.88420278e-02 4.78829265e-01 4.66967374e-02 4.03068550e-02
1.24514580e-01 -2.46421341e-02 3.58217776e-01 8.79917562e-01
-1.17903352e-01 -1.42783806e-01 -8.17206383e-01 5.45373797e-01
-1.58593273e+00 -9.27693069e-01 -4.56701815e-01 2.89703798e+00
9.21564937e-01 2.07918331e-01 3.64044309e-01 4.78221118e-01
6.94181681e-01 1.34029657e-01 -3.29903603e-01 -4.23389614e-01
-2.06953242e-01 8.94468188e-01 9.66044426e-01 6.72640681e-01
-7.17954338e-01 4.99778837e-01 4.89039516e+00 1.26682043e+00
-7.75639355e-01 -2.60583460e-02 6.56695366e-01 -3.05097193e-01
-5.51490784e-01 1.98145702e-01 -9.83943284e-01 7.47166455e-01
6.47224426e-01 -1.29654229e-01 6.49843812e-01 7.42514968e-01
2.75397778e-01 -5.78333855e-01 -1.06753874e+00 1.02990258e+00
-5.38836680e-02 -1.09068346e+00 5.32026179e-02 5.38261056e-01
6.82025790e-01 -4.31136250e-01 1.90229252e-01 -1.55292243e-01
4.34585750e-01 -7.57747829e-01 6.20932996e-01 2.79659808e-01
9.52694952e-01 -1.03218329e+00 3.89688283e-01 4.57602441e-01
-1.14202774e+00 1.35147318e-01 -2.86601305e-01 1.70663614e-02
2.75721997e-01 1.51902294e+00 -4.25946206e-01 4.35326695e-01
3.82614464e-01 5.91665842e-02 4.32458818e-02 6.11112177e-01
1.70577079e-01 5.73947489e-01 -9.22659457e-01 -1.31361261e-01
6.99405149e-02 -9.51568186e-01 9.01228011e-01 8.35361421e-01
5.66151500e-01 2.37702355e-01 1.42630907e-02 7.20319450e-01
-2.76877731e-01 3.18091899e-01 -4.76198018e-01 1.33884966e-01
3.91559094e-01 7.91138053e-01 -8.73471677e-01 -1.04046375e-01
-2.07123399e-01 9.15277779e-01 8.79437774e-02 1.04844436e-01
-5.42061090e-01 -5.20915210e-01 5.58194041e-01 6.63806379e-01
5.80809593e-01 -7.20029891e-01 -7.51830995e-01 -8.96333337e-01
3.83551061e-01 -7.00933576e-01 4.14540246e-02 1.07292226e-02
-1.18162656e+00 3.34240586e-01 -2.62776054e-02 -9.90204930e-01
-1.75400395e-02 -3.24380457e-01 -1.12542935e-01 1.06053758e+00
-9.04896975e-01 -4.10596579e-01 1.18099891e-01 4.82724428e-01
-1.94314960e-02 1.23801254e-01 6.52292192e-01 3.70463818e-01
-5.87980151e-01 5.30855238e-01 6.33472085e-01 -1.68908238e-01
4.51071769e-01 -1.19091594e+00 1.72389179e-01 8.93371642e-01
-6.45601302e-02 7.93492734e-01 7.77733028e-01 -5.34688056e-01
-1.63609684e+00 -7.78378606e-01 9.46036696e-01 -2.67586976e-01
5.14549673e-01 -4.73985076e-01 -5.51809371e-01 3.10891777e-01
-3.28441292e-01 -3.60055894e-01 7.22252846e-01 2.54719462e-02
1.14296540e-03 -4.18036640e-01 -1.13523555e+00 6.67265356e-01
9.09538627e-01 -1.70865357e-01 4.94970009e-02 3.04227293e-01
2.71614760e-01 -2.29123682e-01 -7.29676664e-01 2.04260170e-01
2.28522763e-01 -9.37536359e-01 8.76119494e-01 2.83985063e-02
3.92709196e-01 -3.76280338e-01 -2.38936067e-01 -8.96793544e-01
-8.51708204e-02 -9.93045628e-01 -6.85848519e-02 9.98409092e-01
5.25140882e-01 -7.59737074e-01 7.63076246e-01 5.21030009e-01
1.46498963e-01 -9.60545540e-01 -1.09073675e+00 -6.96156621e-01
1.72674194e-01 -3.28380018e-01 3.57619703e-01 4.75021869e-01
-4.47752833e-01 3.87757927e-01 -3.12483937e-01 -1.49538741e-01
7.05034137e-01 6.78708106e-02 9.76617455e-01 -1.28999662e+00
-7.01822937e-01 -2.85259247e-01 -9.57155898e-02 -1.39212203e+00
-4.91497636e-01 -7.26184607e-01 -2.95038044e-01 -1.27605879e+00
4.22205597e-01 -9.30661500e-01 2.52998382e-01 6.11088946e-02
-8.37500542e-02 -2.37153590e-01 8.19955096e-02 4.70583677e-01
-3.71366799e-01 5.35321116e-01 1.18878460e+00 2.46649846e-01
-5.03742754e-01 3.66731614e-01 -7.59985626e-01 7.26002097e-01
4.49208438e-01 -7.04872489e-01 -1.78508997e-01 -3.33217889e-01
8.01026165e-01 3.88999522e-01 4.21944866e-03 -9.54070449e-01
1.99382648e-01 -2.97041871e-02 -2.87414771e-02 -7.76683807e-01
3.32946122e-01 -6.36053920e-01 2.63508320e-01 4.77181792e-01
-9.34380218e-02 1.96951225e-01 -1.08143963e-01 7.49497294e-01
7.85196666e-03 -5.58052897e-01 9.08841133e-01 -1.09488867e-01
4.27644730e-01 2.62805969e-01 -4.74251568e-01 1.42469019e-01
1.10130751e+00 -2.23612845e-01 -7.76025578e-02 -3.27115923e-01
-3.62621307e-01 -8.54570791e-02 6.25111699e-01 -4.97104645e-01
1.19356580e-01 -1.00762510e+00 -5.43049634e-01 1.78804845e-01
-3.11195999e-01 5.48777997e-01 3.84165019e-01 1.08278728e+00
-7.63440609e-01 3.11707377e-01 4.43057716e-01 -6.52694821e-01
-1.02023435e+00 3.90658230e-01 -1.07537977e-01 -8.09827447e-01
-3.04624081e-01 8.33029509e-01 2.55323667e-02 -7.58037716e-02
-5.60498200e-02 -3.18138152e-01 4.43803877e-01 1.97253719e-01
3.63632172e-01 7.85359204e-01 5.62044568e-02 -2.89621025e-01
-1.10988960e-01 6.61427140e-01 -1.20516144e-01 -5.06296635e-01
1.35775447e+00 -9.40079987e-02 -6.72776461e-01 5.13273656e-01
1.20625389e+00 6.43390715e-01 -1.00816309e+00 -1.27208903e-01
-2.96047628e-01 -3.38232338e-01 -4.80870277e-01 -2.07334101e-01
-8.64326477e-01 6.69098496e-01 3.32617104e-01 5.02281249e-01
1.15286648e+00 -3.37552205e-02 7.09577501e-01 1.93246797e-01
5.30546963e-01 -1.08494174e+00 -8.25687274e-02 8.08834732e-01
4.77982402e-01 -5.72936058e-01 1.19372115e-01 -6.37610435e-01
-2.70904720e-01 8.49146545e-01 -1.01438917e-01 -4.02385503e-01
7.80578196e-01 2.10104138e-01 -6.41144872e-01 -2.06277832e-01
-2.94668883e-01 -8.44833180e-02 -1.17086202e-01 2.32631359e-02
1.73462689e-01 2.49056920e-01 -9.54091191e-01 6.24231339e-01
-4.19331014e-01 -9.05393735e-02 5.34522831e-01 8.67415011e-01
-4.67183590e-01 -1.57838273e+00 -4.98756140e-01 9.02058601e-01
-6.13399982e-01 -2.85090297e-01 -1.32065505e-01 4.38331872e-01
9.45245922e-02 6.38096333e-01 7.09312335e-02 -1.32919950e-02
2.53045499e-01 1.14638560e-01 5.28379023e-01 -6.16785288e-01
-5.24483502e-01 2.00775042e-01 -1.48161232e-01 -1.55544221e-01
-2.91344933e-02 -9.36337948e-01 -1.02462292e+00 -6.40805364e-01
-1.34954557e-01 3.45922947e-01 6.40155435e-01 7.64639795e-01
3.23192179e-01 1.15864821e-01 7.97291160e-01 -4.68406677e-01
-7.82843530e-01 -7.69678771e-01 -9.62448061e-01 1.53281808e-01
-8.58106092e-02 -3.74908537e-01 -5.98587632e-01 -1.29740909e-01] | [6.604231357574463, 4.709583759307861] |
8a1f5ede-22c8-4b6e-ba57-7ca42ee8913f | long-term-conversation-analysis-exploring | 2306.16071 | null | https://arxiv.org/abs/2306.16071v1 | https://arxiv.org/pdf/2306.16071v1.pdf | Long-term Conversation Analysis: Exploring Utility and Privacy | The analysis of conversations recorded in everyday life requires privacy protection. In this contribution, we explore a privacy-preserving feature extraction method based on input feature dimension reduction, spectral smoothing and the low-cost speaker anonymization technique based on McAdams coefficient. We assess the utility of the feature extraction methods with a voice activity detection and a speaker diarization system, while privacy protection is determined with a speech recognition and a speaker verification model. We show that the combination of McAdams coefficient and spectral smoothing maintains the utility while improving privacy. | ['Joerg Bitzer', 'Patrick A. Naylor', 'Jule Pohlhausen', 'Francesco Nespoli'] | 2023-06-28 | null | null | null | null | ['action-detection', 'activity-detection', 'dimensionality-reduction', 'speech-recognition', 'speaker-diarization', 'speaker-verification'] | ['computer-vision', 'computer-vision', 'methodology', 'speech', 'speech', 'speech'] | [ 3.74191076e-01 3.80243361e-01 1.25307962e-01 -6.25975370e-01
-8.11352313e-01 -8.92985225e-01 6.59910798e-01 9.05148685e-02
-5.41113555e-01 8.72939646e-01 4.88712788e-01 -3.45133960e-01
-2.08039522e-01 -3.30360413e-01 -1.95393220e-01 -7.66461790e-01
-2.09122986e-01 -2.19958887e-01 -7.56948441e-02 2.91315559e-03
3.34093161e-02 8.87901962e-01 -1.41778708e+00 1.60813421e-01
6.46667778e-01 9.55749094e-01 -8.03849757e-01 8.19585741e-01
1.09341867e-01 1.28712460e-01 -8.44825089e-01 -5.62165201e-01
8.03373158e-01 -2.70146996e-01 -5.82309902e-01 -2.42512062e-01
1.77480683e-01 -3.56636971e-01 -2.74167687e-01 8.71528685e-01
6.80526733e-01 1.73101231e-01 3.77601445e-01 -1.62084067e+00
-2.14069821e-02 2.98839897e-01 -8.05736333e-02 2.07138985e-01
7.87937403e-01 -1.86213374e-01 5.69781601e-01 -2.33488590e-01
4.11905825e-01 1.12165046e+00 8.22084785e-01 5.46693802e-01
-1.62432301e+00 -6.99714184e-01 -2.51250207e-01 -1.06585905e-01
-1.78989196e+00 -9.33999300e-01 8.74223769e-01 -2.72096485e-01
7.30255127e-01 1.01352489e+00 4.82209802e-01 1.00245833e+00
2.71830726e-02 3.14083844e-01 9.03119564e-01 -2.84321278e-01
3.31050247e-01 8.68983567e-01 2.71207243e-01 4.24193949e-01
2.83012748e-01 7.36460388e-02 -7.58749306e-01 -1.07088506e+00
3.34111273e-01 -1.48293942e-01 -4.52637136e-01 -4.60537225e-01
-7.86588073e-01 7.67101109e-01 -5.14720559e-01 2.36551166e-01
-1.82390720e-01 -4.00349885e-01 5.56654871e-01 6.23051584e-01
4.43163604e-01 3.25367361e-01 -3.75468582e-01 1.03263892e-02
-1.06031978e+00 3.40359479e-01 1.41952777e+00 1.22650456e+00
5.75493395e-01 -1.14852265e-01 -1.46174029e-01 3.94063532e-01
3.42760354e-01 5.32805860e-01 7.37733245e-01 -8.41241121e-01
3.07011932e-01 3.65626901e-01 1.33328572e-01 -8.56012583e-01
-2.79880941e-01 9.20485631e-02 -5.00254154e-01 4.01952453e-02
3.55058521e-01 -4.76258904e-01 -2.07334429e-01 1.86881697e+00
4.37241882e-01 -1.55244339e-02 3.49964321e-01 3.76774520e-01
2.90566921e-01 3.48975718e-01 -4.29342166e-02 -6.51674867e-01
1.39838123e+00 -2.43790090e-01 -1.21755517e+00 4.95227069e-01
4.68868703e-01 -4.74589080e-01 7.43911445e-01 5.22466719e-01
-7.68141508e-01 -1.68978851e-02 -1.13125288e+00 -6.35116771e-02
-3.89762282e-01 -3.64075780e-01 6.04493022e-01 1.92025447e+00
-9.83787596e-01 4.83453125e-01 -6.43639565e-01 -4.94101286e-01
3.42555404e-01 7.71919906e-01 -9.17647600e-01 6.25379384e-01
-1.32030547e+00 4.87478077e-01 -1.09451927e-01 -4.08460945e-01
-2.70114630e-01 -6.61077321e-01 -8.34502935e-01 1.45683005e-01
-1.98506802e-01 -4.50618982e-01 9.25727010e-01 -6.95889235e-01
-1.87657392e+00 8.54791999e-01 -2.55481362e-01 -8.66283476e-01
9.63193834e-01 7.71957487e-02 -8.54825497e-01 2.60271132e-01
-2.72301465e-01 -6.05087541e-02 1.08228564e+00 -7.85696566e-01
-5.53153753e-01 -6.81050301e-01 -6.18657649e-01 3.31633836e-02
-6.03259921e-01 2.70906895e-01 1.69271946e-01 -5.61113119e-01
1.01808555e-01 -7.52444506e-01 5.18619418e-02 -9.64929760e-02
-6.52915537e-01 3.87544692e-01 1.12251663e+00 -1.19502199e+00
1.45179987e+00 -2.65627623e+00 -3.13800067e-01 7.97399819e-01
9.36694741e-02 1.51611432e-01 3.59768957e-01 4.72514302e-01
-4.75342460e-02 4.05773908e-01 -4.20200497e-01 -5.91682374e-01
1.20871112e-01 -4.19311598e-02 -4.51022983e-01 9.71793413e-01
-1.27777740e-01 2.43413836e-01 -2.35838741e-01 -3.76601726e-01
1.71733782e-01 7.05326915e-01 -5.93137205e-01 1.32337719e-01
5.60485125e-01 2.66975075e-01 -2.86005050e-01 6.55990958e-01
9.24989641e-01 6.65591061e-01 4.03521746e-01 8.03351775e-02
-3.14076066e-01 4.22777504e-01 -1.34544802e+00 1.53394985e+00
1.04539379e-01 5.73633134e-01 6.47533298e-01 -5.00560701e-01
1.12882614e+00 7.74189591e-01 4.39181924e-01 2.52099819e-02
3.80459577e-01 2.31424086e-02 -1.85939714e-01 -5.32166123e-01
4.03590769e-01 -4.36151661e-02 -8.93774405e-02 5.32922089e-01
1.83517970e-02 2.24811450e-01 -3.99675876e-01 6.39747232e-02
1.13650012e+00 -5.48901141e-01 6.47013485e-01 -6.59429967e-01
7.26612151e-01 -6.88851595e-01 6.47639275e-01 4.95447338e-01
-9.57994223e-01 5.06832063e-01 6.07920229e-01 -4.97431606e-02
-7.32070804e-01 -9.70998406e-01 -5.40847123e-01 7.48383045e-01
-3.43242675e-01 -3.60630006e-01 -1.26280427e+00 -6.96948349e-01
4.13032472e-01 8.57352555e-01 -5.80923319e-01 -3.17494690e-01
-2.71826684e-01 -5.06796300e-01 1.10808992e+00 -7.93322846e-02
2.94786721e-01 -3.56682509e-01 -3.94817680e-01 -1.17437774e-02
7.32945800e-02 -7.98182070e-01 -7.22282767e-01 1.89112842e-01
-7.24078655e-01 -6.65447295e-01 -3.47152919e-01 -3.69141877e-01
2.44127035e-01 -9.33246091e-02 1.54985934e-01 -4.08705324e-01
-2.38634542e-01 6.92081392e-01 -1.59773454e-01 -6.20734334e-01
-6.50213599e-01 2.13774905e-01 5.22165477e-01 7.53576159e-01
7.83468425e-01 -1.11590493e+00 -3.15840989e-01 1.67742953e-01
-6.11807704e-01 -7.44311213e-01 -1.31009132e-01 3.32715482e-01
3.30396648e-03 -9.27421823e-02 6.22906327e-01 -8.85338604e-01
8.69298398e-01 -2.96180248e-01 -3.29304159e-01 1.84511721e-01
-7.98277855e-01 -1.67983528e-02 3.31882924e-01 -3.74608159e-01
-1.09080052e+00 6.53180063e-01 -1.53255403e-01 -9.80186537e-02
-2.96142012e-01 -3.50901812e-01 -9.91846740e-01 -5.10724843e-01
7.91971803e-01 4.17699337e-01 3.63443136e-01 -6.75433636e-01
3.27606261e-01 1.18799758e+00 4.52392370e-01 -1.07720055e-01
6.38576269e-01 6.57429099e-01 -2.34468654e-01 -1.34741998e+00
2.98260391e-01 -4.76830214e-01 -6.84604585e-01 6.16429299e-02
5.64592123e-01 -8.07914853e-01 -1.04008591e+00 3.55931967e-01
-1.03788340e+00 5.92169166e-01 -4.24152017e-01 6.05294943e-01
-6.02242172e-01 8.14455926e-01 -2.22690389e-01 -1.37618709e+00
-5.93623042e-01 -7.09693551e-01 8.60158503e-01 -1.07086442e-01
-6.27723575e-01 -5.27763009e-01 1.99319154e-01 3.65250051e-01
2.96288759e-01 4.22132224e-01 4.37407613e-01 -1.31863821e+00
-1.20784774e-01 -8.43764901e-01 2.93277711e-01 4.46389437e-01
3.74657899e-01 -2.05256641e-01 -1.69589698e+00 -3.66042525e-01
6.10018253e-01 3.75720382e-01 3.15305293e-01 3.13134253e-01
8.97924721e-01 -9.29205120e-01 -1.48009464e-01 1.01745224e+00
8.93118322e-01 4.87642825e-01 7.08846748e-01 -1.93540379e-02
3.78195882e-01 8.93349171e-01 -2.38961428e-01 9.01543558e-01
5.86925307e-04 4.90864247e-01 -1.63347080e-01 3.94084275e-01
3.66920948e-01 -2.50110984e-01 3.72907460e-01 4.42124784e-01
2.46276617e-01 -4.07571457e-02 -5.88369966e-01 4.85998720e-01
-1.72431314e+00 -1.00122023e+00 1.31533831e-01 2.49432206e+00
6.78590536e-01 -3.05877864e-01 5.66553056e-01 6.77428901e-01
7.88440228e-01 4.58642207e-02 -1.86043486e-01 -8.45710278e-01
-2.36073941e-01 -7.26260468e-02 9.42630172e-01 8.42747629e-01
-1.01827919e+00 4.61759239e-01 7.15743351e+00 4.07008886e-01
-7.73968637e-01 1.01486653e-01 3.48714590e-01 -3.96581799e-01
-1.77604094e-01 -2.23294526e-01 -6.58215523e-01 4.88433242e-01
1.28569806e+00 -6.81891203e-01 5.91350079e-01 8.78748775e-01
3.10083658e-01 1.95335597e-01 -1.19707823e+00 1.04099214e+00
2.66750395e-01 -8.35609019e-01 -1.71521708e-01 4.41778004e-01
-1.37138382e-01 -5.65187991e-01 -1.99996270e-02 -2.42294192e-01
-3.84546556e-02 -6.84473813e-01 3.68719935e-01 4.51120913e-01
6.32879794e-01 -9.23167050e-01 2.76543617e-01 1.22038126e-01
-9.52787817e-01 -2.27137223e-01 4.41854633e-02 5.71532957e-02
1.18241951e-01 4.33573723e-01 -1.02839291e+00 3.24320734e-01
4.26273018e-01 5.38095832e-02 -1.71436653e-01 6.44761980e-01
2.16247573e-01 6.29385293e-01 -6.74083173e-01 6.65203389e-03
-5.17351568e-01 -3.39676529e-01 1.04538953e+00 1.38057935e+00
2.93437392e-01 2.35569432e-01 -4.27824885e-01 6.31906629e-01
6.93693012e-02 4.19855088e-01 -8.52652729e-01 1.15531217e-02
8.13781738e-01 9.80225742e-01 -1.18498132e-01 1.25765935e-01
-1.73905939e-01 1.16630554e+00 -2.45105922e-01 2.00113520e-01
-1.92919880e-01 -7.65353024e-01 1.17428935e+00 3.79040211e-01
-6.97421879e-02 2.28804145e-02 -6.07955098e-01 -1.16811359e+00
2.61893839e-01 -7.53816605e-01 4.90217865e-01 6.61228448e-02
-1.00562787e+00 4.39660162e-01 -1.23279803e-01 -1.08466458e+00
-4.94387656e-01 -1.61543369e-01 -4.92721081e-01 1.15264559e+00
-9.14175272e-01 -8.80263984e-01 2.89078653e-01 1.09685791e+00
-2.09883675e-01 -5.58071733e-01 1.08516574e+00 3.17300260e-01
-7.55059481e-01 1.19270480e+00 1.06616490e-01 7.21550956e-02
4.97393072e-01 -9.97234643e-01 3.90295386e-01 8.05710673e-01
-2.58002251e-01 9.71986949e-01 8.99670780e-01 -3.59925687e-01
-1.45177829e+00 -8.32605541e-01 1.47325385e+00 -4.37209815e-01
3.26907933e-01 -9.59220231e-01 -1.07841206e+00 7.84205496e-01
6.48651272e-02 -2.01607913e-01 1.26322365e+00 1.88814811e-02
-4.77047354e-01 -3.49123627e-01 -2.12564063e+00 2.30482504e-01
6.79260194e-01 -9.96000648e-01 -7.50161946e-01 1.05185308e-01
6.65296555e-01 1.25420138e-01 -9.91863370e-01 -8.32695514e-02
9.18486416e-01 -8.76019537e-01 8.19649220e-01 -5.94745815e-01
-8.46314311e-01 -1.10818341e-01 -4.06867981e-01 -6.37965858e-01
-5.19665889e-02 -1.50156271e+00 -3.54666188e-02 1.83838260e+00
5.57347417e-01 -1.09299362e+00 9.26755309e-01 1.53139448e+00
6.21061683e-01 2.24647209e-01 -1.61796618e+00 -8.43059599e-01
-3.29862446e-01 -1.84722066e-01 7.81956673e-01 1.13606012e+00
6.65221632e-01 -8.01410377e-02 -6.20214522e-01 4.25570965e-01
7.54531085e-01 -5.95918715e-01 8.10944974e-01 -1.27905893e+00
-7.82874748e-02 5.20431176e-02 -8.12013268e-01 -4.02325839e-02
1.54062554e-01 -5.70428193e-01 -4.04310673e-01 -4.85570639e-01
-2.02068120e-01 2.99191535e-01 -3.19313377e-01 1.43160030e-01
3.95093501e-01 -4.53021258e-01 -1.15126267e-01 -8.53789598e-02
1.06406078e-01 3.99895728e-01 -2.67888382e-02 2.42983997e-01
-6.54006541e-01 3.77936959e-01 -7.65670419e-01 4.41031486e-01
9.81767118e-01 -6.72978997e-01 -5.31442881e-01 3.58126134e-01
-3.30856681e-01 1.26637325e-01 1.73979774e-01 -9.69161987e-01
8.58639181e-02 1.10579342e-01 4.46657911e-02 -6.88982084e-02
4.24750090e-01 -1.13451612e+00 2.76987940e-01 5.23366690e-01
-5.52153826e-01 -1.75957844e-01 2.82293200e-01 8.27187061e-01
2.03823447e-02 1.07316427e-01 8.34780097e-01 2.83655524e-01
3.89040285e-03 5.22558391e-02 -7.20574796e-01 -4.49588180e-01
1.07268894e+00 -5.08141220e-01 -3.25452723e-02 -4.92335409e-01
-7.50386238e-01 -1.58210024e-01 4.44323868e-01 2.25515202e-01
2.81367481e-01 -1.02565014e+00 -5.19497275e-01 8.30186188e-01
1.01498945e-03 -8.09482276e-01 1.20140664e-01 5.84864020e-01
-1.28050700e-01 3.00388336e-01 -3.39423448e-01 -1.10566005e-01
-1.94707108e+00 6.65609777e-01 3.69175971e-01 2.68059343e-01
-4.71710265e-01 7.36167371e-01 -2.20809907e-01 -3.48563313e-01
5.19452393e-01 -1.94566533e-01 4.22244854e-02 2.05083311e-01
8.96552801e-01 7.72079885e-01 3.12184006e-01 -6.48335040e-01
-7.38394499e-01 -1.58368662e-01 -4.79210401e-03 -6.12720609e-01
9.58737969e-01 -6.17795765e-01 -3.20919901e-02 2.02003136e-01
1.48581326e+00 4.74707365e-01 -9.54385459e-01 -4.05152515e-02
6.23293966e-02 -6.43885970e-01 3.13067883e-02 -5.13481677e-01
-6.34683728e-01 4.56501424e-01 1.00640607e+00 7.28973269e-01
1.08062744e+00 -4.91088390e-01 4.93936449e-01 3.76658052e-01
2.36951828e-01 -1.17360044e+00 -1.27376246e+00 -1.44498060e-02
8.05292964e-01 -9.75389481e-01 4.69911285e-02 -5.72833180e-01
-5.60703993e-01 6.06152177e-01 -2.35390902e-01 1.21573433e-01
1.16046858e+00 4.90572304e-01 1.28512099e-01 3.72896463e-01
-4.81388122e-01 3.95706534e-01 1.00987434e-01 1.07745135e+00
1.08666673e-01 2.78019637e-01 -6.87671781e-01 1.15435576e+00
-7.04069078e-01 -1.88149795e-01 4.43752110e-01 1.00135338e+00
-2.25826040e-01 -1.07102633e+00 -4.87091333e-01 3.02869469e-01
-5.03347576e-01 -7.10315444e-03 -1.13032079e+00 5.18630683e-01
-1.93257704e-01 1.22744560e+00 -1.45122245e-01 -5.50240457e-01
5.66389680e-01 6.67450726e-01 -2.99421251e-02 -1.29990447e-02
-1.11418641e+00 -1.63907439e-01 3.45548183e-01 -7.08160341e-01
-2.35099062e-01 -1.29101789e+00 -9.12198961e-01 -7.45616794e-01
-1.80213094e-01 2.68774539e-01 1.04706848e+00 6.23430490e-01
7.87007272e-01 -1.12292223e-01 9.89278555e-01 -1.97842419e-01
-7.73186564e-01 -7.85359144e-01 -1.45264208e+00 2.52152294e-01
8.44938517e-01 -1.12422280e-01 -7.75367677e-01 1.16725452e-01] | [13.937080383300781, 5.874355792999268] |
224e50e0-57ee-4008-9976-e893ce0be9c0 | the-aircraft-context-dataset-understanding | null | null | https://openaccess.thecvf.com/content/ICCV2021W/AOTW/html/Steininger_The_Aircraft_Context_Dataset_Understanding_and_Optimizing_Data_Variability_in_ICCVW_2021_paper.html | https://openaccess.thecvf.com/content/ICCV2021W/AOTW/papers/Steininger_The_Aircraft_Context_Dataset_Understanding_and_Optimizing_Data_Variability_in_ICCVW_2021_paper.pdf | The Aircraft Context Dataset: Understanding and Optimizing Data Variability in Aerial Domains | Despite their increasing demand for assistant and autonomous systems, the recent shift towards data-driven approaches has hardly reached aerial domains, partly due to a lack of specific training and test data. We introduce the Aircraft Context Dataset, a composition of two inter-compatible large-scale and versatile image datasets focusing on manned aircraft and UAVs, respectively. In addition to fine-grained annotations for multiple learning tasks, we define and apply a set of relevant meta-parameters and showcase their potential to quantify dataset variability as well as the impact of environmental conditions on model performance. Baseline experiments are conducted for detection, classification and semantic labeling on multiple dataset variants. Their evaluation clearly shows that our contribution is an essential step towards overcoming the data gap and that the proposed variability concept significantly increases the efficiency of specializing models as well as continuously and purposefully extending the dataset. | ['Christoph Sulzbachner', 'Andreas Kriegler', 'Julia Simon', 'Verena Widhalm', 'Daniel Steininger'] | 2021-10-17 | null | null | null | proceedings-of-the-ieee-cvf-international | ['robust-object-detection', 'small-object-detection', 'fine-grained-image-classification'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.40174437e-01 -1.22838765e-01 -8.89242906e-03 -5.48622727e-01
-4.40048873e-01 -1.07547951e+00 1.02713656e+00 2.24117830e-01
-4.97151494e-01 7.33002484e-01 -1.72686085e-01 -1.45769849e-01
-6.58060730e-01 -4.10775155e-01 -5.03870785e-01 -5.92509270e-01
-2.17972249e-01 7.92752266e-01 5.07224619e-01 -3.26674998e-01
5.31569542e-03 7.60286987e-01 -2.09208369e+00 3.11228067e-01
4.99701381e-01 1.05133986e+00 3.18202704e-01 5.90173125e-01
2.16279358e-01 5.21212399e-01 -7.44704485e-01 -9.13386717e-02
4.99489397e-01 1.16489017e-02 -7.27994025e-01 4.88365263e-01
5.06932259e-01 1.47280544e-01 1.08519986e-01 7.59989202e-01
3.78511041e-01 9.37646255e-03 4.15432453e-01 -1.66755760e+00
-1.51306331e-01 1.59496695e-01 -5.48443571e-02 7.79143302e-03
7.79313445e-02 3.76116276e-01 8.39484751e-01 -3.51723075e-01
7.92784691e-01 9.28376019e-01 7.86888897e-01 3.65677088e-01
-1.16266406e+00 -2.88251340e-01 4.85277534e-01 8.30144361e-02
-1.23899996e+00 -3.53278637e-01 5.83519220e-01 -6.14723861e-01
8.87022495e-01 4.08290327e-01 3.79916340e-01 1.41087806e+00
-1.23815894e-01 6.00784242e-01 1.25126934e+00 -4.03203487e-01
2.10585788e-01 4.81385171e-01 5.66497743e-02 7.95649707e-01
4.05532926e-01 2.86029071e-01 -2.75166482e-01 6.21118098e-02
4.15155858e-01 -1.81992441e-01 -2.83137500e-01 -1.04231989e+00
-1.44287324e+00 3.49192351e-01 6.86416626e-02 3.72562200e-01
-2.76828021e-01 -2.12821960e-01 5.13013363e-01 5.65342307e-01
5.93100265e-02 7.42689848e-01 -1.06485355e+00 -1.16121516e-01
-8.86417329e-01 4.54936445e-01 7.70168543e-01 1.27436149e+00
7.05419958e-01 -7.06647011e-03 -1.47835705e-02 5.64551055e-01
-7.85786845e-03 4.95969743e-01 2.67309129e-01 -8.08107316e-01
3.07144016e-01 9.82300699e-01 5.64101934e-01 -6.50585532e-01
-6.94401741e-01 -6.14327371e-01 -3.04057002e-01 2.94389337e-01
3.97086978e-01 -7.70432875e-02 -9.56085265e-01 1.75084734e+00
1.54912636e-01 -1.25659809e-01 1.64829582e-01 7.77719200e-01
7.34760523e-01 3.06673441e-02 1.57864824e-01 9.36714336e-02
1.27287555e+00 -1.02144432e+00 -3.53451729e-01 -4.74064350e-01
8.08709800e-01 -3.72586906e-01 1.17743361e+00 5.16387701e-01
-6.84586704e-01 -8.45001400e-01 -1.10101509e+00 2.70325154e-01
-9.44772422e-01 3.28925014e-01 6.19189978e-01 7.78994381e-01
-8.24000180e-01 4.05275553e-01 -6.47943139e-01 -7.52001226e-01
3.05445135e-01 2.89694637e-01 -6.73807919e-01 5.41371964e-02
-1.07159090e+00 1.09123695e+00 6.50271058e-01 -7.22686648e-02
-1.22899997e+00 -7.98020482e-01 -7.10594416e-01 -2.36589655e-01
8.03614795e-01 -5.65535367e-01 9.84061658e-01 -6.74471319e-01
-1.12540126e+00 1.08583188e+00 2.95108706e-01 -6.59224689e-01
6.65942371e-01 -1.75770596e-01 -6.65474832e-01 -1.93408057e-01
-1.37048528e-01 8.22046518e-01 9.03648019e-01 -1.48209476e+00
-9.99185801e-01 -3.40376973e-01 4.61258292e-01 -1.64761283e-02
-4.02994275e-01 -1.08349226e-01 -3.83656889e-01 -1.90993890e-01
-3.40059489e-01 -1.13010085e+00 -2.66743094e-01 -1.48907512e-01
-1.14525132e-01 3.38201135e-01 1.18403232e+00 -3.33643705e-01
1.10147655e+00 -2.11928320e+00 5.03727198e-01 1.73542202e-02
9.71129071e-03 4.19142574e-01 -3.44179064e-01 4.44419384e-01
-3.39025259e-02 -9.23995897e-02 -5.46967328e-01 -3.01487863e-01
6.78043067e-02 4.85630929e-01 -3.31179470e-01 1.31610870e-01
5.35757840e-01 7.11102307e-01 -9.00715232e-01 -1.33192271e-01
6.52157187e-01 3.66192937e-01 -2.28407279e-01 1.63999096e-01
-4.86216158e-01 7.27193356e-01 -2.72299647e-01 7.26935446e-01
4.20300514e-01 3.37683447e-02 2.69973934e-01 -2.82564282e-01
-1.93214417e-01 -8.22709575e-02 -1.23671150e+00 1.91725242e+00
-7.03023314e-01 3.71999472e-01 2.32738316e-01 -8.84112179e-01
9.81143236e-01 2.00471163e-01 3.89793664e-01 -5.28372288e-01
2.06810772e-01 1.17117926e-01 7.08020700e-04 -5.38289607e-01
7.26030111e-01 1.50979102e-01 -4.02044505e-01 -1.22120000e-01
2.54713535e-01 -3.42577726e-01 3.30875218e-01 -1.24468073e-01
9.81760800e-01 4.87788588e-01 3.59012306e-01 -4.75484967e-01
7.77605116e-01 4.43936288e-01 2.87302494e-01 5.73293269e-01
-4.75482553e-01 4.51345980e-01 3.80937785e-01 -5.91091692e-01
-9.12119150e-01 -9.30575728e-01 -1.90300480e-01 1.11899388e+00
1.11824825e-01 -4.15686399e-01 -6.27700388e-01 -9.30969596e-01
2.47950286e-01 9.19061363e-01 -7.40180910e-01 -2.03704938e-01
-1.66324839e-01 -7.64051914e-01 4.99452382e-01 3.89007300e-01
4.27076817e-01 -8.48451197e-01 -1.08113873e+00 -1.19707197e-01
1.43946290e-01 -1.48202658e+00 3.36572289e-01 4.66206223e-01
-7.30973065e-01 -1.33186042e+00 -4.17015813e-02 -3.39017779e-01
5.19277871e-01 3.85931671e-01 1.36945772e+00 -4.79491130e-02
-4.44378972e-01 7.71606505e-01 -4.65100139e-01 -7.04250276e-01
-5.91739357e-01 3.77510697e-01 1.64590433e-01 -1.72146022e-01
3.24369252e-01 -3.67346466e-01 -1.72660381e-01 5.57145596e-01
-1.04860461e+00 -1.66644871e-01 6.33503795e-01 7.02594638e-01
4.49363172e-01 9.46598724e-02 3.80748183e-01 -9.91402745e-01
4.38886911e-01 -2.67786294e-01 -9.62723136e-01 5.06757617e-01
-8.29134107e-01 9.58127975e-02 4.82119083e-01 -8.95342827e-02
-9.01834190e-01 2.76430309e-01 2.53679901e-01 -4.56084222e-01
-9.49078381e-01 2.64206946e-01 -4.63913590e-01 -1.42330751e-01
7.06550241e-01 -1.28962755e-01 -4.95199263e-02 -3.70394140e-01
4.46078688e-01 4.98222917e-01 5.70895553e-01 -6.90575063e-01
8.31951737e-01 3.43579054e-01 1.78745404e-01 -8.20200920e-01
-6.83344543e-01 -4.95085031e-01 -1.08498228e+00 -1.66766018e-01
6.86173439e-01 -8.80885661e-01 -3.00745159e-01 2.49312595e-01
-7.89821029e-01 -3.81254166e-01 -4.78879631e-01 1.35368854e-01
-6.07591569e-01 1.92286864e-01 1.55688643e-01 -7.56219745e-01
1.75279602e-01 -1.46271408e+00 1.24910629e+00 -2.98064530e-01
-1.26641303e-01 -1.20558715e+00 -1.21534832e-01 3.98811966e-01
4.85733658e-01 5.40781617e-01 8.25363755e-01 -8.60526264e-01
-4.26385492e-01 -5.65305769e-01 -8.35315734e-02 4.18479741e-01
2.94158012e-01 -6.20710701e-02 -1.24320459e+00 -5.28723240e-01
-1.33550748e-01 -4.22399521e-01 7.07746089e-01 -1.33536354e-01
1.02557874e+00 3.21804196e-01 -4.16132182e-01 2.60127455e-01
1.37764251e+00 -9.71326046e-03 2.90055156e-01 5.65558076e-01
5.31956375e-01 9.29622889e-01 9.86596227e-01 4.11201835e-01
2.61126310e-01 8.08748484e-01 9.42965925e-01 -4.83084954e-02
-2.83741448e-02 7.54060075e-02 4.54610512e-02 1.17122419e-01
-1.91823408e-01 -3.56502205e-01 -1.08958614e+00 5.97423077e-01
-1.75289905e+00 -8.01099122e-01 -9.60745811e-02 2.49149060e+00
2.21407339e-01 1.12121738e-01 2.19828054e-01 1.57276526e-01
3.30470294e-01 3.20089698e-01 -4.45938915e-01 -1.27336413e-01
-1.82323813e-01 -1.03964776e-01 6.50486648e-01 3.61327142e-01
-1.29833579e+00 1.03256989e+00 6.86079025e+00 5.43305039e-01
-9.99026358e-01 3.42021063e-02 1.02475755e-01 -1.66237503e-01
3.36361974e-02 -1.60490394e-01 -8.27463031e-01 5.94798811e-02
1.22271168e+00 3.54670137e-01 3.79443854e-01 1.00621462e+00
-1.06209464e-01 6.43679947e-02 -1.25107419e+00 6.75637782e-01
-4.70340624e-02 -9.41935420e-01 -9.92381051e-02 -1.24685308e-02
5.64445674e-01 1.79773748e-01 6.68058470e-02 3.93219352e-01
3.46366435e-01 -7.95620024e-01 9.37732697e-01 2.98671305e-01
6.29296064e-01 -5.89109361e-01 8.09002995e-01 2.83560097e-01
-1.04131401e+00 -4.35445964e-01 -6.12812005e-02 -3.84679697e-02
-7.80635551e-02 2.30979457e-01 -9.66776311e-01 1.03769839e+00
6.89337373e-01 6.01808310e-01 -9.30993736e-01 6.58897161e-01
-1.17134802e-01 2.90503621e-01 -2.02630907e-01 4.50973630e-01
3.16733032e-01 -2.93483632e-03 6.49767816e-01 1.35049355e+00
1.37192100e-01 -5.59309244e-01 4.23723757e-01 5.12666404e-01
2.67616719e-01 -2.70995796e-01 -7.25406110e-01 -2.29347423e-01
5.46301842e-01 1.55418837e+00 -6.32121146e-01 -1.57250136e-01
-6.15126610e-01 9.77625370e-01 1.72533616e-01 1.93418786e-01
-1.00472617e+00 -4.90711294e-02 9.45017636e-01 -1.25100790e-02
3.57093513e-01 -4.97321606e-01 -1.63938671e-01 -9.15978014e-01
8.34853724e-02 -1.00902629e+00 2.97879875e-01 -5.34868360e-01
-8.94209921e-01 8.07122886e-01 3.21853757e-01 -1.47853649e+00
-3.09065789e-01 -1.01867783e+00 -1.70411274e-03 6.24075353e-01
-1.68921614e+00 -1.63627982e+00 -8.68113399e-01 2.90967733e-01
4.09435481e-01 -5.54831684e-01 1.13465178e+00 2.60179907e-01
-5.20298362e-01 2.56091267e-01 -2.08571702e-02 -4.87624586e-01
7.84875393e-01 -1.20392144e+00 3.34956467e-01 7.98252344e-01
1.45586386e-01 4.03119296e-01 9.64752793e-01 -1.82269335e-01
-1.36391270e+00 -1.37584138e+00 3.09884816e-01 -1.22714448e+00
8.12425137e-01 -6.13047898e-01 -7.07042098e-01 8.32192719e-01
-1.72771607e-02 9.19060782e-02 4.48918909e-01 1.62658021e-01
-4.92578417e-01 -2.24563643e-01 -1.32506311e+00 2.08394825e-01
1.30724299e+00 -3.82414699e-01 -4.38938171e-01 4.33033615e-01
6.54626369e-01 -3.53118956e-01 -8.76439035e-01 8.04898441e-01
3.70260596e-01 -1.37323189e+00 1.00584149e+00 -8.69117916e-01
1.33005351e-01 -4.22767103e-01 -5.14836669e-01 -1.28745866e+00
-1.76346958e-01 -2.27858633e-01 1.32510602e-01 1.32538271e+00
5.61173439e-01 -5.32640934e-01 6.08953476e-01 6.21987879e-01
-5.08633912e-01 -5.59726357e-01 -5.28235614e-01 -1.04279029e+00
-2.39312157e-01 -5.91068983e-01 8.56284559e-01 9.66316462e-01
-6.54242218e-01 1.11648947e-01 -3.72360677e-01 5.28396547e-01
2.38044679e-01 1.92230999e-01 1.07764709e+00 -1.53428829e+00
-1.91872358e-01 -3.65180373e-01 -7.31034219e-01 -4.45596308e-01
1.11779764e-01 -6.15208924e-01 2.33298242e-02 -1.24398613e+00
-2.38475755e-01 -6.46349370e-01 -3.91708761e-01 4.50690508e-01
9.42723453e-02 9.67421234e-02 4.06533554e-02 1.30232722e-01
-7.70230770e-01 1.98799580e-01 7.82111168e-01 3.10267825e-02
-7.16474131e-02 -7.16404170e-02 -5.18854618e-01 6.49797142e-01
7.52015948e-01 -1.67313427e-01 -7.80572236e-01 -5.81316113e-01
-1.24615513e-01 -4.93404269e-01 6.67396128e-01 -1.39817595e+00
-2.38613814e-01 -1.39243022e-01 1.53747469e-01 -2.19491467e-01
5.05554974e-01 -1.39891040e+00 3.00989330e-01 4.33613896e-01
-2.71741837e-01 1.00157924e-01 5.50362945e-01 6.30613089e-01
-2.68920481e-01 -1.43750921e-01 7.46484101e-01 -2.34854519e-01
-1.37542343e+00 1.96143508e-01 -4.26409021e-02 -1.05924353e-01
1.40462959e+00 -2.75238752e-01 -1.73143551e-01 1.71842113e-01
-6.36931658e-01 2.37143382e-01 7.87626922e-01 1.05562985e+00
-6.77085444e-02 -7.45528042e-01 -5.00351727e-01 4.58592713e-01
7.46098638e-01 1.85255725e-02 2.14553505e-01 5.96220374e-01
-1.96204290e-01 8.85623395e-01 -4.56958503e-01 -7.28360116e-01
-1.46127856e+00 7.12940276e-01 3.36658269e-01 -2.64973968e-01
-3.24992687e-01 7.31474400e-01 -1.65544339e-02 -8.54658544e-01
3.25759053e-01 -4.78760928e-01 -7.55831450e-02 1.04688331e-01
3.04157674e-01 2.55092889e-01 4.30319220e-01 -5.95214069e-01
-5.15771568e-01 2.65036404e-01 3.91720794e-02 1.60003781e-01
1.23033643e+00 -8.92201439e-02 1.36893690e-01 5.06882131e-01
6.20728850e-01 -1.64889887e-01 -1.30972302e+00 -4.79533896e-02
4.59277809e-01 -4.08964634e-01 7.33968467e-02 -1.29161716e+00
-8.63374829e-01 8.40260684e-01 8.86906266e-01 2.48433173e-01
1.12513411e+00 -3.57417837e-02 1.97536781e-01 6.49946988e-01
8.36573303e-01 -8.73866081e-01 -2.79092580e-01 3.92044067e-01
9.70426321e-01 -1.47360671e+00 3.02615613e-02 -4.56447393e-01
-8.36792052e-01 8.70293260e-01 6.34206176e-01 3.05152148e-01
1.94653451e-01 4.13103402e-01 2.22404972e-01 -2.95033336e-01
-7.40451634e-01 -5.08248746e-01 1.91982046e-01 1.06083679e+00
4.06236529e-01 2.67470535e-03 1.65920898e-01 4.99692082e-01
-8.63410160e-02 1.03566699e-01 2.14268044e-01 1.10045695e+00
-4.81250137e-01 -1.39736581e+00 -1.21346749e-01 2.04297125e-01
3.30870748e-02 3.46489847e-01 -6.13202095e-01 1.25875103e+00
4.32969242e-01 8.74647081e-01 -9.56665054e-02 -5.94557643e-01
7.87702858e-01 2.32814118e-01 6.39418244e-01 -5.55362999e-01
-6.75436318e-01 -6.39853179e-01 3.87445629e-01 -5.53308725e-01
-5.61857700e-01 -7.68328309e-01 -8.94074023e-01 1.68345407e-01
-2.38194957e-01 5.26639745e-02 9.43623364e-01 8.51072967e-01
6.85773253e-01 8.39063585e-01 2.05380902e-01 -8.32968354e-01
-6.48987889e-01 -9.21904027e-01 -3.39643359e-01 5.21762550e-01
1.59659281e-01 -1.07293653e+00 -2.60541618e-01 7.73949623e-02] | [8.042759895324707, -1.1248924732208252] |
374af84b-e2eb-49e0-bedc-7cec21a1aa9c | exploring-domain-invariant-parameters-for | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wang_Exploring_Domain-Invariant_Parameters_for_Source_Free_Domain_Adaptation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wang_Exploring_Domain-Invariant_Parameters_for_Source_Free_Domain_Adaptation_CVPR_2022_paper.pdf | Exploring Domain-Invariant Parameters for Source Free Domain Adaptation | Source-free domain adaptation (SFDA) newly emerges to transfer the relevant knowledge of a well-trained source model to an unlabeled target domain, which is critical in various privacy-preserving scenarios. Most existing methods focus on learning the domain-invariant representations depending solely on the target data, leading to the obtained representations are target-specific. In this way, they cannot fully address the distribution shift problem across domains. In contrast, we provide a fascinating insight: rather than attempting to learn domain-invariant representations, it is better to explore the domain-invariant parameters of the source model. The motivation behind this insight is clear: the domain-invariant representations are dominated by only partial parameters of an available deep source model. We devise the Domain-Invariant Parameter Exploring (DIPE) approach to capture such domain-invariant parameters in the source model to generate domain-invariant representations. A distinguishing method is developed correspondingly for two types of parameters, i.e., domain-invariant and domain-specific parameters, as well as an effective update strategy based on the clustering correction technique and a target hypothesis is proposed. Extensive experiments verify that DIPE successfully exceeds the current state-of-the-art models on many domain adaptation datasets. | ['Yilong Yin', 'Yongshun Gong', 'Zhongyi Han', 'Fan Wang'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 4.64965582e-01 -1.23759523e-01 -6.00494862e-01 -4.51974720e-01
-9.67036545e-01 -8.25107574e-01 8.87868524e-01 -6.61926493e-02
-2.55698711e-01 1.07407546e+00 3.08392555e-01 -5.65777682e-02
-2.00013056e-01 -5.31494737e-01 -6.15575016e-01 -1.01441717e+00
3.17523509e-01 4.34741497e-01 2.41473526e-01 -3.43669713e-01
9.70620811e-02 3.85183334e-01 -1.15970802e+00 4.87278961e-02
1.18306386e+00 9.37814713e-01 -1.04420155e-01 4.50026095e-02
-2.04689860e-01 2.40951315e-01 -6.09816790e-01 -4.51389641e-01
5.50043106e-01 -6.94617391e-01 -6.36431754e-01 1.07655928e-01
1.21164300e-01 -2.36588821e-01 -2.39392832e-01 1.15357757e+00
3.97724926e-01 1.88986093e-01 1.11599243e+00 -1.33559728e+00
-7.89072514e-01 1.43846199e-01 -4.43780541e-01 3.73938717e-02
-3.83134885e-03 -1.75678536e-01 8.00873518e-01 -6.53294742e-01
8.27082276e-01 9.68990564e-01 2.89211154e-01 7.97077000e-01
-1.29435956e+00 -8.90538990e-01 3.38602036e-01 1.55942857e-01
-1.49912179e+00 -4.96651828e-01 1.13976562e+00 -4.48712677e-01
2.18193993e-01 -7.26696402e-02 1.56020358e-01 1.56718254e+00
-1.43135041e-02 5.52958488e-01 1.11410928e+00 -4.33610469e-01
6.30432487e-01 8.11415195e-01 2.70234235e-02 1.03874691e-01
4.51792359e-01 2.00273320e-01 -5.36804080e-01 -5.44429243e-01
7.67485082e-01 1.68754950e-01 -3.87598842e-01 -1.34453380e+00
-9.23767328e-01 9.66313064e-01 2.11810440e-01 3.70947003e-01
-2.86816746e-01 -5.64452767e-01 4.89127278e-01 4.03000653e-01
5.47704518e-01 2.73458511e-01 -7.49954641e-01 1.53164983e-01
-7.84975469e-01 2.01364145e-01 8.10762942e-01 1.25450277e+00
9.60811257e-01 -4.98325936e-02 -2.27647826e-01 9.16147172e-01
2.07912952e-01 2.81838983e-01 7.51422584e-01 -5.67274451e-01
4.84221578e-01 6.45325422e-01 2.97316998e-01 -6.85944557e-01
1.58868179e-01 -5.51491141e-01 -9.55077767e-01 -3.08181290e-02
4.73151773e-01 -1.61689728e-01 -8.31277549e-01 2.16079521e+00
4.70587730e-01 2.41169915e-01 3.41457486e-01 6.67502642e-01
4.34392124e-01 2.97686815e-01 1.18884362e-01 -1.69502541e-01
1.20069468e+00 -6.99442983e-01 -6.34436965e-01 -3.97343606e-01
4.79896694e-01 -3.78159672e-01 9.20763075e-01 1.07458808e-01
-6.28539324e-01 -3.95894051e-01 -1.16313314e+00 1.58993304e-01
-7.19207704e-01 -1.30240126e-02 3.54170620e-01 7.90638924e-01
-8.14214289e-01 3.73359889e-01 -3.00449967e-01 -5.71601391e-01
6.36255205e-01 2.90190339e-01 -6.48504853e-01 -3.64424407e-01
-1.48079073e+00 7.83099651e-01 5.22422850e-01 -5.09163380e-01
-8.06957781e-01 -8.95501614e-01 -7.41904557e-01 -4.56590988e-02
3.41015458e-01 -7.55915463e-01 9.09364700e-01 -1.14633560e+00
-1.51619017e+00 9.75883961e-01 -3.89319718e-01 -5.36309898e-01
6.76564574e-01 -2.72128843e-02 -5.66740334e-01 1.65748283e-01
5.69120198e-02 3.14131260e-01 1.25855434e+00 -1.55455065e+00
-4.97703165e-01 -5.13294458e-01 -2.44078562e-01 1.86773628e-01
-7.90880978e-01 -1.89234674e-01 -5.06543517e-01 -7.09168971e-01
-1.76232561e-01 -8.33151460e-01 -1.57195523e-01 2.94958353e-01
-1.80850789e-01 9.96900573e-02 9.19638515e-01 -6.76950932e-01
1.09988832e+00 -2.38488817e+00 3.86812836e-02 3.74899119e-01
9.27993804e-02 6.03588641e-01 -1.85002565e-01 3.77562106e-01
-2.70146489e-01 -8.99346024e-02 -6.82387650e-01 -2.43552312e-01
3.69413644e-02 3.10829818e-01 -5.52536190e-01 5.80397308e-01
3.72027636e-01 5.39054215e-01 -8.00831556e-01 -3.77906710e-01
1.60393506e-01 6.26442373e-01 -6.30718768e-01 2.03772977e-01
-9.98908095e-03 8.66884470e-01 -7.26173818e-01 4.90417808e-01
1.22861969e+00 -2.25706384e-01 2.87966579e-01 1.17501887e-02
2.93843269e-01 2.10464790e-01 -1.17208111e+00 1.84271955e+00
-1.90650985e-01 1.01751707e-01 1.60952389e-01 -1.39604795e+00
1.28075767e+00 3.10171157e-01 5.93839943e-01 -5.23812771e-01
5.60147613e-02 3.54943782e-01 -2.74264187e-01 1.01537950e-01
2.62834787e-01 -3.59063357e-01 -1.36132583e-01 2.71539807e-01
3.41679126e-01 2.69275606e-01 -3.38691115e-01 4.46702205e-02
9.51799214e-01 1.56747848e-01 8.35458100e-01 -4.11841124e-01
7.80672789e-01 -1.53093383e-01 7.98530102e-01 5.67324996e-01
-5.56054115e-01 8.34848821e-01 3.01411867e-01 -1.90603018e-01
-9.10625398e-01 -1.14035368e+00 -3.30297321e-01 1.04633844e+00
2.14724898e-01 4.98142699e-03 -7.31215835e-01 -1.03471971e+00
-1.98254362e-02 6.69476330e-01 -7.32242942e-01 -5.71957290e-01
-3.50099742e-01 -6.92513406e-01 3.33611786e-01 2.56992638e-01
7.18102694e-01 -4.97637302e-01 -2.90221535e-02 1.19028017e-01
-1.20629758e-01 -1.04018223e+00 -5.28784096e-01 3.22083086e-01
-1.02895069e+00 -8.82352650e-01 -1.21201479e+00 -5.87464750e-01
7.64300168e-01 3.72419924e-01 9.32877243e-01 -5.05312443e-01
2.88859308e-01 3.67726147e-01 -4.24190223e-01 -3.30095053e-01
-4.43946481e-01 4.08419937e-01 2.66464874e-02 4.70929801e-01
7.85429299e-01 -8.15802276e-01 -5.15420675e-01 3.15152436e-01
-1.07049882e+00 -5.16924441e-01 8.05624008e-01 9.39856529e-01
5.13383269e-01 -4.44653332e-02 9.05234396e-01 -1.21038616e+00
5.83987892e-01 -8.18567693e-01 -3.31896544e-01 3.47874284e-01
-7.17314065e-01 2.02976137e-01 9.38283682e-01 -7.23499238e-01
-1.29357147e+00 2.98943162e-01 3.14588659e-02 -7.30232954e-01
-4.13960934e-01 4.74435352e-02 -6.56594515e-01 -2.69442666e-02
8.99827480e-01 6.07253134e-01 -2.44783200e-02 -5.87815344e-01
3.28599244e-01 6.50642812e-01 4.58397716e-01 -7.58970201e-01
1.19585240e+00 7.00608194e-01 -2.78041333e-01 -6.38366699e-01
-5.91287315e-01 -7.52888978e-01 -9.42573607e-01 4.59832579e-01
4.54256624e-01 -1.10152531e+00 2.37072617e-01 2.31674001e-01
-9.38274562e-01 1.34140164e-01 -5.85000336e-01 3.69728923e-01
-6.09723926e-01 4.12350655e-01 1.93888232e-01 -5.76658547e-01
-9.63897109e-02 -8.87545705e-01 8.77241850e-01 1.45093441e-01
-7.37232566e-02 -1.27159953e+00 2.81421065e-01 9.81140062e-02
4.11310107e-01 2.15232328e-01 9.91976261e-01 -1.31348264e+00
-3.78272355e-01 -2.00409427e-01 -2.82821655e-01 5.51246226e-01
3.75094116e-01 -5.84378719e-01 -1.10519588e+00 -5.11081219e-01
1.51097655e-01 4.78400029e-02 7.97714949e-01 2.94352472e-01
1.12496746e+00 -4.10347551e-01 -5.56877732e-01 8.07386220e-01
1.44725025e+00 1.77836940e-01 5.93785703e-01 3.16731513e-01
4.10532266e-01 4.29116488e-01 5.46704352e-01 6.50771677e-01
3.42896074e-01 7.77243137e-01 1.91358998e-01 -3.10586262e-02
3.53013910e-02 -6.60167217e-01 3.79848659e-01 2.92414427e-01
3.23203027e-01 -7.04056174e-02 -5.93056381e-01 8.16514671e-01
-1.75584579e+00 -6.83776498e-01 5.72195053e-01 2.56691718e+00
1.02679706e+00 -2.05544814e-01 2.43413240e-01 -1.79175466e-01
9.19471502e-01 2.37787336e-01 -1.11508167e+00 -2.32055321e-01
-1.95855513e-01 2.56491512e-01 6.24634504e-01 1.19303882e-01
-1.20369589e+00 8.53565156e-01 6.00417614e+00 1.03588998e+00
-9.99051094e-01 1.09407492e-01 2.78336644e-01 2.36054942e-01
-5.62547445e-01 2.47253850e-02 -8.54890168e-01 5.19914985e-01
8.04814875e-01 -6.23080730e-01 2.83499897e-01 1.22353673e+00
-1.81960747e-01 3.84571791e-01 -1.18936276e+00 8.78211856e-01
8.84733722e-02 -1.17587662e+00 3.28152001e-01 4.42836404e-01
7.10343361e-01 -3.03953588e-01 2.31995642e-01 4.96861875e-01
3.96738350e-01 -5.83728194e-01 2.74806470e-01 3.16907912e-01
9.42902505e-01 -7.94539154e-01 4.88700479e-01 4.58801568e-01
-8.81323278e-01 -1.62333876e-01 -6.35031641e-01 2.41540641e-01
-2.96067953e-01 5.46199143e-01 -8.39650571e-01 9.23699021e-01
5.31117260e-01 9.09613490e-01 -4.70509648e-01 8.55101824e-01
3.00365090e-02 5.36816180e-01 -1.71687260e-01 4.57640797e-01
-9.40715596e-02 -2.77078986e-01 5.59347689e-01 9.70213175e-01
4.44407046e-01 -1.19051471e-01 -1.94541410e-01 9.00883734e-01
-3.42193812e-01 1.10122897e-01 -1.12438965e+00 4.87672277e-02
8.33887577e-01 8.82764101e-01 -1.40491873e-01 -1.23787157e-01
-4.52144593e-01 1.15161598e+00 2.42744908e-01 6.25001132e-01
-7.06120551e-01 -1.44927233e-01 1.14995074e+00 1.59463778e-01
5.80266654e-01 -5.99198751e-02 -3.23461950e-01 -1.46805775e+00
1.06900208e-01 -1.05599010e+00 6.23001397e-01 -3.52401324e-02
-1.69024634e+00 4.41471696e-01 6.40478134e-02 -1.71180546e+00
-3.49500179e-01 -3.24468642e-01 -4.81920630e-01 1.02994692e+00
-2.02645302e+00 -1.17923725e+00 5.65632060e-02 1.15504527e+00
2.72589266e-01 -4.72531110e-01 1.05071807e+00 3.78835917e-01
-4.39947218e-01 1.20252824e+00 8.28564227e-01 9.73472148e-02
1.23551333e+00 -8.43169987e-01 1.23786345e-01 8.87530982e-01
-9.12050381e-02 1.05840373e+00 5.59198320e-01 -3.51711512e-01
-1.15645802e+00 -1.26678360e+00 7.21130311e-01 -5.18914104e-01
4.32887405e-01 -5.22842526e-01 -1.25424850e+00 6.81309998e-01
-2.32748035e-02 2.32071713e-01 9.22350407e-01 4.27525043e-02
-7.25659311e-01 -2.95790642e-01 -1.67672956e+00 3.18664372e-01
1.04057062e+00 -4.80739206e-01 -6.42129540e-01 -4.56958031e-03
6.73153400e-01 -1.49919167e-01 -6.90884590e-01 2.07211912e-01
3.10330331e-01 -8.94685745e-01 1.25133252e+00 -7.25072205e-01
1.18289500e-01 -3.42419833e-01 -3.01592737e-01 -1.40358055e+00
-3.17825168e-01 -5.86477816e-01 -1.86474472e-01 1.62161732e+00
1.44766420e-01 -9.99214828e-01 8.63720775e-01 7.29558051e-01
2.36465737e-01 -3.07235271e-01 -1.23197782e+00 -9.45123553e-01
5.65106153e-01 1.54431671e-01 9.16342139e-01 1.36643624e+00
-3.13765496e-01 2.93296337e-01 -4.55030441e-01 3.03230464e-01
7.45186448e-01 7.07609653e-02 8.96147192e-01 -1.51027322e+00
2.36557014e-02 -1.80660009e-01 -2.19763100e-01 -1.17792797e+00
4.29988235e-01 -8.10071468e-01 -2.44066209e-01 -1.05912602e+00
1.25506714e-01 -5.61397433e-01 -6.76764011e-01 2.68334895e-01
-1.57147855e-01 -3.37257117e-01 7.39750117e-02 3.65434945e-01
-4.32155848e-01 8.87202442e-01 1.02458513e+00 -1.02314137e-01
-2.92587608e-01 2.21277431e-01 -1.39416122e+00 4.92585599e-01
8.94844472e-01 -5.82306445e-01 -7.28353500e-01 -1.01465300e-01
-4.21304792e-01 -2.80749381e-01 3.56076807e-01 -8.92450035e-01
1.14603542e-01 -3.93579513e-01 3.90311807e-01 -2.90578425e-01
2.60781944e-01 -1.12202132e+00 -9.51881930e-02 1.63507253e-01
-4.09203768e-01 -6.58450484e-01 1.51021615e-01 9.20778453e-01
-3.56204033e-01 -1.49967089e-01 1.19356716e+00 -9.13554281e-02
-1.01039767e+00 4.42968100e-01 5.41901886e-02 2.63862222e-01
9.81737018e-01 -3.69073868e-01 -3.28437537e-01 -4.53093946e-01
-3.82462770e-01 -9.27150771e-02 6.10487223e-01 4.91197586e-01
3.83072227e-01 -1.56690764e+00 -7.29946971e-01 6.17256403e-01
5.90551972e-01 -2.85743997e-02 3.11681747e-01 3.13999891e-01
2.84831256e-01 3.96310627e-01 -3.44528943e-01 -4.53582764e-01
-7.48036623e-01 7.26684749e-01 2.65887707e-01 -4.18486536e-01
-4.10686851e-01 7.81262279e-01 8.10030699e-01 -7.51365721e-01
1.41913623e-01 1.62554681e-02 -1.56809211e-01 9.61877704e-02
5.72354436e-01 1.16964392e-02 -2.63145175e-02 -6.87506199e-01
-5.13732612e-01 5.03219485e-01 -3.14973295e-01 1.34778336e-01
1.24626935e+00 -3.53451937e-01 3.22028577e-01 7.90519044e-02
1.36898327e+00 -1.11261427e-01 -1.59715366e+00 -8.14016342e-01
-3.69890593e-02 -6.36980176e-01 -1.93335071e-01 -8.06610227e-01
-7.67664254e-01 9.34619009e-01 6.87694371e-01 -1.90149039e-01
1.43195641e+00 -9.45693254e-02 6.64017081e-01 1.40777737e-01
4.77101088e-01 -1.24590456e+00 -1.18722185e-01 3.95594209e-01
7.94624507e-01 -1.34666574e+00 -1.67503431e-02 -2.24866167e-01
-8.67936492e-01 7.31679618e-01 4.99705702e-01 -6.94718137e-02
7.04901755e-01 -2.33214244e-01 -8.38443562e-02 3.42402786e-01
-4.66960281e-01 -1.54589996e-01 3.23870033e-01 1.39538193e+00
9.64398012e-02 -1.30355924e-01 -2.92949881e-02 9.28269088e-01
1.37266800e-01 1.58461213e-01 1.92373112e-01 7.47314930e-01
-2.18167618e-01 -1.60086751e+00 -4.18471783e-01 1.61097810e-01
-2.37890273e-01 5.44651374e-02 -6.88082695e-01 7.91532040e-01
-2.07947269e-02 6.97757602e-01 -2.85295784e-01 -2.04287887e-01
3.44948888e-01 3.95643920e-01 1.19241588e-01 -6.29009008e-01
-2.57397622e-01 -1.52119666e-01 -3.36169511e-01 -3.26107144e-01
-3.46517146e-01 -8.13081980e-01 -8.00596654e-01 -2.08447725e-01
-6.34840131e-02 1.45129085e-01 4.63157922e-01 8.00951004e-01
7.89526463e-01 1.67579159e-01 8.70754004e-01 -5.14068484e-01
-8.14059019e-01 -6.53932214e-01 -7.81228840e-01 6.65856421e-01
5.31516612e-01 -8.65093589e-01 -4.63955581e-01 1.27944080e-02] | [10.37838363647461, 3.1559929847717285] |
26f4f866-be64-4451-8466-c9319f35903d | a-comparative-study-of-conversion-aided | null | null | https://aclanthology.org/W14-4507 | https://aclanthology.org/W14-4507.pdf | A Comparative Study of Conversion Aided Methods for WordNet Sentence Textual Similarity | null | ['Muhidin Mohamed', 'Mourad Oussalah'] | 2014-08-01 | null | null | null | ws-2014-8 | ['text-clustering'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.138737201690674, 3.687035083770752] |
ba370f2b-2c40-4d6f-9a17-e08754950ede | investigating-the-importance-of-shape | 2002.03779 | null | https://arxiv.org/abs/2002.03779v2 | https://arxiv.org/pdf/2002.03779v2.pdf | Investigating the Importance of Shape Features, Color Constancy, Color Spaces and Similarity Measures in Open-Ended 3D Object Recognition | Despite the recent success of state-of-the-art 3D object recognition approaches, service robots are frequently failed to recognize many objects in real human-centric environments. For these robots, object recognition is a challenging task due to the high demand for accurate and real-time response under changing and unpredictable environmental conditions. Most of the recent approaches use either the shape information only and ignore the role of color information or vice versa. Furthermore, they mainly utilize the $L_n$ Minkowski family functions to measure the similarity of two object views, while there are various distance measures that are applicable to compare two object views. In this paper, we explore the importance of shape information, color constancy, color spaces, and various similarity measures in open-ended 3D object recognition. Towards this goal, we extensively evaluate the performance of object recognition approaches in three different configurations, including \textit{color-only}, \textit{shape-only}, and \textit{ combinations of color and shape}, in both offline and online settings. Experimental results concerning scalability, memory usage, and object recognition performance show that all of the \textit{combinations of color and shape} yields significant improvements over the \textit{shape-only} and \textit{color-only} approaches. The underlying reason is that color information is an important feature to distinguish objects that have very similar geometric properties with different colors and vice versa. Moreover, by combining color and shape information, we demonstrate that the robot can learn new object categories from very few training examples in a real-world setting. | ['Wessel van der Rest', 'Jits Schilperoort', 'S. Hamidreza Kasaei', 'Maryam Ghorbani'] | 2020-02-10 | null | null | null | null | ['3d-object-recognition', 'color-constancy'] | ['computer-vision', 'computer-vision'] | [-2.36858204e-02 -6.19469523e-01 1.75100923e-01 -3.25257957e-01
-1.02672741e-01 -8.33365440e-01 4.41537172e-01 8.20923895e-02
-5.02294481e-01 3.58334154e-01 -6.50515795e-01 -9.91302803e-02
-2.97481298e-01 -5.72977304e-01 -5.06646693e-01 -8.20124507e-01
-6.41491711e-02 6.95257425e-01 3.30003798e-01 -9.67109650e-02
4.02627170e-01 1.23339152e+00 -1.98259628e+00 -2.61624128e-01
5.25815010e-01 1.40494061e+00 3.76128614e-01 4.48011667e-01
-2.65583038e-01 1.37542486e-01 -6.01930261e-01 -1.11670226e-01
7.97200322e-01 -1.67290956e-01 -4.52438176e-01 3.89887303e-01
2.53063917e-01 -2.26750314e-01 -3.27235669e-01 1.20314229e+00
5.10047793e-01 3.17233473e-01 6.45442724e-01 -1.57284391e+00
-7.02441394e-01 -2.48329818e-01 -4.72992778e-01 -1.34621136e-04
5.00776291e-01 3.02683800e-01 5.31421125e-01 -9.54877317e-01
3.88629347e-01 1.15385473e+00 2.70499140e-01 4.35288548e-01
-1.16111028e+00 -5.15604854e-01 4.02797699e-01 5.19871414e-01
-1.25348711e+00 -2.76130617e-01 9.27513778e-01 -3.07112157e-01
8.05305302e-01 2.56366014e-01 6.11057758e-01 6.74562216e-01
-1.08408280e-01 8.05860519e-01 1.29480731e+00 -4.24267322e-01
3.86226267e-01 2.30765730e-01 -2.83904430e-02 7.23280609e-01
6.04095280e-01 1.61213800e-01 -3.06629241e-01 8.17019045e-02
8.59425128e-01 3.88896406e-01 -2.56615907e-01 -9.94329572e-01
-1.37648797e+00 4.40514922e-01 3.63281995e-01 1.26151651e-01
-1.93855211e-01 -1.42347246e-01 1.55203477e-01 3.46544892e-01
5.22121787e-02 5.96375108e-01 -2.16645554e-01 -2.71459937e-01
-1.16484582e-01 -1.45632988e-02 8.91480625e-01 1.39498651e+00
8.98984015e-01 1.00753032e-01 3.55673224e-01 9.80839133e-01
8.94752666e-02 8.76205921e-01 2.75523305e-01 -8.54828238e-01
3.45572352e-01 7.77304292e-01 3.94892901e-01 -1.20483613e+00
-6.40952110e-01 -1.44979224e-01 -6.12840116e-01 6.76726043e-01
5.46679735e-01 2.63989627e-01 -1.01145375e+00 1.37358201e+00
4.71597642e-01 -3.87648016e-01 9.67216343e-02 1.11953688e+00
5.84208548e-01 2.24152505e-01 -4.57448781e-01 -9.00892355e-03
1.12130928e+00 -8.96800935e-01 -2.43336931e-01 -3.34129810e-01
4.09980118e-01 -7.09075332e-01 1.02176774e+00 3.91951382e-01
-7.75446773e-01 -4.79857087e-01 -1.13377094e+00 3.06708992e-01
-6.95489585e-01 2.68022716e-01 5.68892717e-01 7.16876447e-01
-7.00779200e-01 3.05462360e-01 -7.57197917e-01 -5.83514273e-01
5.02413698e-02 4.09171492e-01 -6.95170999e-01 -4.54185754e-01
-3.85738641e-01 1.11358345e+00 4.00394499e-01 2.01752529e-01
-5.01272500e-01 -4.60729077e-02 -7.22406745e-01 -1.78242922e-01
5.40748417e-01 -1.44485444e-01 1.01114798e+00 -7.34220743e-01
-1.66425240e+00 8.66453946e-01 3.85754369e-02 1.72893226e-01
6.03661835e-01 -3.56984586e-02 -2.37172633e-01 3.58406216e-01
-1.53461277e-01 3.46945167e-01 7.10927725e-01 -1.61163437e+00
-6.36414111e-01 -7.22393095e-01 1.62643462e-01 3.38328421e-01
-1.32736608e-01 -3.06684315e-01 -6.34089172e-01 -2.58391529e-01
6.97490811e-01 -1.10107589e+00 6.51325509e-02 4.22700018e-01
-1.60966501e-01 -1.30161420e-01 1.04718041e+00 -8.04258510e-02
3.12998354e-01 -2.27454233e+00 5.04676625e-02 3.15223038e-01
-1.98138431e-01 1.85541064e-01 -2.60974646e-01 2.57746905e-01
1.88546777e-01 -2.35114768e-01 2.80854944e-03 3.86857912e-02
1.72447637e-01 4.13811207e-01 1.34588212e-01 6.20312154e-01
1.46468341e-01 5.65677047e-01 -8.58385801e-01 -2.54930466e-01
4.63527352e-01 2.02616140e-01 -2.22899377e-01 2.20243588e-01
3.83412256e-03 3.53163719e-01 -5.41702628e-01 8.68614376e-01
8.26105952e-01 1.44933620e-02 1.44557625e-01 -1.74637780e-01
-1.50908381e-01 -2.21283823e-01 -1.40971768e+00 1.37708545e+00
-2.62091696e-01 3.65376234e-01 1.61383644e-01 -1.27565432e+00
1.45808673e+00 6.50966540e-03 6.65686429e-01 -1.01367331e+00
2.86682636e-01 3.98638844e-01 -9.15579274e-02 -4.98026192e-01
3.74883711e-01 5.88239096e-02 8.11917707e-02 5.01809061e-01
-2.70975024e-01 -4.63850617e-01 3.45046133e-01 -3.87715608e-01
9.16254878e-01 2.67801583e-01 1.04944982e-01 -6.66283146e-02
3.85305375e-01 -1.01680212e-01 4.04397339e-01 7.97408283e-01
-4.80202109e-01 7.09229708e-01 2.27127492e-01 -4.75797921e-01
-1.02048349e+00 -1.13365543e+00 9.06981975e-02 8.80265594e-01
1.01058710e+00 3.63983840e-01 -1.62939325e-01 -6.65404975e-01
2.76757121e-01 4.88047481e-01 -5.07501125e-01 -3.41532469e-01
-5.27455270e-01 -4.14820552e-01 2.28728652e-01 5.37621558e-01
6.27484441e-01 -8.92629564e-01 -1.21366000e+00 1.00254258e-02
1.38031870e-01 -1.17914426e+00 -1.61555544e-01 3.79941314e-01
-8.29788744e-01 -1.16816533e+00 -7.03272402e-01 -7.53486931e-01
9.61585224e-01 8.88071120e-01 6.03533864e-01 -1.48886191e-02
-4.35453594e-01 8.88437152e-01 -7.13474214e-01 -4.84756142e-01
-3.97462733e-02 -2.23958030e-01 1.95380241e-01 2.13948619e-02
3.97812605e-01 -5.28565884e-01 -4.46847409e-01 8.87535095e-01
-8.84709477e-01 2.98031643e-02 8.72705519e-01 6.47603452e-01
5.13071060e-01 -1.03569806e-01 1.87245935e-01 -2.36634448e-01
1.86037824e-01 -9.89190117e-02 -6.55777812e-01 4.66169566e-01
-5.85776508e-01 9.77384895e-02 4.98083800e-01 -8.17689300e-01
-7.94100165e-01 2.70859241e-01 4.23114210e-01 -6.82858229e-01
-4.45598990e-01 1.97820529e-01 -2.94785857e-01 -2.61755824e-01
3.03302824e-01 5.03531277e-01 2.90297747e-01 -4.97615963e-01
8.38347897e-02 6.27825379e-01 3.69741052e-01 -6.76664650e-01
8.63187134e-01 6.65736616e-01 9.87916663e-02 -7.96970606e-01
-4.28719938e-01 -5.55104494e-01 -8.81776035e-01 -4.32168841e-01
5.03198147e-01 -3.85718554e-01 -1.05088031e+00 6.16219640e-01
-1.03050041e+00 -1.01053618e-01 -2.18129247e-01 6.82309330e-01
-7.06079245e-01 4.25756663e-01 -7.48109818e-02 -1.12766445e+00
1.00840412e-01 -1.27989221e+00 1.02042377e+00 4.03861284e-01
4.08438683e-01 -4.53993529e-01 -4.27010864e-01 1.69775784e-01
3.25574428e-01 2.39262477e-01 9.91689205e-01 -6.27581775e-01
-6.09991968e-01 -4.80609179e-01 -5.98093092e-01 9.37001482e-02
5.48681080e-01 -2.26546302e-01 -6.30021453e-01 -4.26833689e-01
-1.48254126e-01 -1.40098870e-01 3.97775620e-01 -2.35973835e-01
7.96299338e-01 1.26356855e-02 -2.15298072e-01 3.44990999e-01
1.29538298e+00 8.21384788e-01 2.91753531e-01 5.87203681e-01
4.75451797e-01 5.39798677e-01 7.07790256e-01 5.53129971e-01
3.29515219e-01 7.18549728e-01 7.92542279e-01 1.20490178e-01
1.82442129e-01 2.56004274e-01 2.62656540e-01 5.45929551e-01
-2.48982608e-01 -5.88920377e-02 -8.99296582e-01 3.42463732e-01
-1.75359130e+00 -8.00146401e-01 2.98553586e-01 2.41422820e+00
3.97783637e-01 1.86569635e-02 5.76996021e-02 1.75985128e-01
8.20895135e-01 -2.97156721e-01 -8.84410739e-01 -2.01879412e-01
-8.52231830e-02 -2.39107788e-01 5.05483985e-01 -1.45292267e-01
-1.00540137e+00 4.94093895e-01 5.55622625e+00 3.12063932e-01
-1.42885244e+00 -3.99135530e-01 2.21886337e-01 1.14117665e-02
1.34456918e-01 -2.53659934e-01 -5.56678414e-01 1.67001069e-01
5.39695024e-02 9.58924741e-02 5.11221230e-01 1.19834602e+00
-9.11732540e-02 -5.16799331e-01 -1.33127844e+00 1.32124233e+00
4.13021326e-01 -6.36549652e-01 -3.57320122e-02 -1.15530798e-02
3.99826437e-01 -2.05930188e-01 6.47238716e-02 1.65957227e-01
1.04294933e-01 -6.52922392e-01 1.17443287e+00 5.51393986e-01
5.59595525e-01 -3.70626837e-01 7.13020980e-01 2.96941221e-01
-1.21514511e+00 -1.95527628e-01 -3.95200104e-01 6.23179823e-02
-1.79863140e-01 2.54058748e-01 -7.12022781e-01 6.88154876e-01
9.81883109e-01 5.02081931e-01 -5.24382532e-01 1.40285265e+00
5.12271151e-02 -1.83175340e-01 -5.94691813e-01 -4.12875742e-01
8.14539567e-02 -3.60770643e-01 6.23528302e-01 8.58215213e-01
3.79405648e-01 2.64474005e-01 4.58174706e-01 7.44635999e-01
3.93857449e-01 -7.86251128e-02 -5.04577458e-01 -2.04128604e-02
4.11208928e-01 9.48336780e-01 -1.13028705e+00 -2.70133074e-02
-4.78852361e-01 9.37928915e-01 1.26360372e-01 4.80417073e-01
-5.77989638e-01 -4.96211946e-01 5.55875540e-01 -2.67544061e-01
5.76160669e-01 -9.32657301e-01 -3.01496357e-01 -1.00533569e+00
2.74410009e-01 -6.18497610e-01 1.91945910e-01 -8.16831827e-01
-1.31310201e+00 5.43469310e-01 1.69084474e-01 -1.57815039e+00
1.04783833e-01 -1.13166201e+00 -1.37852803e-01 4.40836579e-01
-1.60088193e+00 -8.71728241e-01 -6.14804924e-01 6.59201503e-01
4.10727322e-01 -6.32863194e-02 6.67545795e-01 1.31644055e-01
-3.92915636e-01 4.63616312e-01 4.98632401e-01 9.03018191e-02
5.29672444e-01 -9.96661782e-01 -1.51515841e-01 4.75381017e-01
2.53739711e-02 3.60505879e-01 5.89730740e-01 -3.48844886e-01
-2.09475756e+00 -6.92813337e-01 3.46934833e-02 -1.87518954e-01
2.14893430e-01 -3.17733407e-01 -6.30703866e-01 2.47769445e-01
-4.69171315e-01 1.31641507e-01 2.74273723e-01 -2.19899878e-01
-5.73176980e-01 -3.24588120e-01 -1.26638675e+00 5.22185385e-01
1.16797030e+00 -2.74182320e-01 -3.93746257e-01 -3.37334983e-02
3.00457180e-01 -3.31754476e-01 -6.66283309e-01 5.02198875e-01
8.30612183e-01 -1.13387847e+00 8.84135664e-01 -2.57883340e-01
-1.98227018e-01 -5.46431422e-01 -4.90474373e-01 -1.08105099e+00
-1.34641200e-01 -1.34692997e-01 2.01328933e-01 8.28962326e-01
2.63055027e-01 -9.31028366e-01 7.15924561e-01 7.70508528e-01
-1.50881067e-01 -4.70952481e-01 -9.15112257e-01 -1.29855824e+00
-3.33537847e-01 -2.79224217e-01 5.24224281e-01 6.07882857e-01
-1.27898812e-01 -1.90749526e-01 -4.64005321e-02 2.21808866e-01
5.30206978e-01 7.42976248e-01 1.02991962e+00 -1.38783002e+00
-4.84733246e-02 -5.55499017e-01 -8.58162582e-01 -1.02529836e+00
-2.03809947e-01 -4.70527083e-01 3.49078625e-01 -1.51654685e+00
1.64704785e-01 -7.91107953e-01 -2.88659513e-01 6.10232830e-01
1.05628848e-01 3.08558851e-01 6.36060715e-01 3.51239204e-01
-5.91002107e-01 7.05424368e-01 1.17107606e+00 -3.93430918e-01
-1.96998000e-01 4.74546570e-03 -1.27774075e-01 5.37234426e-01
6.14136577e-01 -1.40997708e-01 -1.69074968e-01 -6.49942338e-01
-1.32304341e-01 -1.82788134e-01 4.31487381e-01 -1.16795421e+00
2.84731746e-01 -5.24682224e-01 3.80620033e-01 -5.54742813e-01
7.13876009e-01 -1.25445879e+00 1.53558119e-03 2.95456648e-01
1.22136690e-01 1.42650083e-01 3.08373302e-01 5.60315549e-01
-9.32247266e-02 -2.26611108e-01 7.93967187e-01 -3.83779377e-01
-9.66648221e-01 2.25337222e-01 -2.69076556e-01 -2.71749526e-01
1.21900260e+00 -9.98465657e-01 -1.94230795e-01 -3.66850853e-01
-5.66884518e-01 6.51052669e-02 7.09504962e-01 6.62151277e-01
7.20791578e-01 -1.22034049e+00 -2.77754605e-01 4.31299895e-01
4.17212486e-01 -2.25472748e-02 -7.30057135e-02 8.93839896e-01
-4.90727097e-01 4.03075665e-01 -5.33545911e-01 -8.25065911e-01
-1.15970767e+00 5.82205296e-01 2.89653152e-01 2.76852608e-01
-3.58411670e-01 5.13013959e-01 2.61435509e-01 -5.96722543e-01
4.72494423e-01 -2.35970065e-01 1.14035085e-01 -1.88593850e-01
1.97415486e-01 5.25342524e-01 8.23826715e-02 -6.04687154e-01
-4.27585274e-01 8.92212868e-01 1.13734178e-01 -4.06305306e-02
1.34772372e+00 -2.86465585e-01 1.31831393e-01 5.27315319e-01
9.61052060e-01 -4.88968939e-01 -1.28077376e+00 -2.86908239e-01
1.30840272e-01 -7.40676045e-01 -3.41246635e-01 -8.27261269e-01
-1.10431921e+00 8.37240458e-01 8.98607492e-01 3.17050219e-01
1.14088738e+00 -4.09976281e-02 2.42949262e-01 8.68142962e-01
9.69216704e-01 -1.15529561e+00 4.36509162e-01 7.15068042e-01
8.94095719e-01 -1.23927104e+00 2.66869720e-02 -3.44827473e-01
-4.40843195e-01 1.31753159e+00 7.99273849e-01 1.07265308e-01
3.53996545e-01 5.47358543e-02 1.04524069e-01 -4.76392871e-03
-2.07083672e-01 -4.41296011e-01 1.40052512e-01 7.76536345e-01
-1.72651619e-01 -9.16280970e-02 1.17970653e-01 1.25305593e-01
-2.08969526e-02 -4.28938597e-01 3.59784901e-01 1.37377763e+00
-6.51491523e-01 -8.82607341e-01 -5.75533807e-01 2.70667672e-01
2.22055122e-01 5.16657174e-01 -5.01652777e-01 8.77376199e-01
-3.56527790e-02 9.70771909e-01 3.21949162e-02 -3.31706613e-01
6.74167037e-01 4.36204374e-02 7.98183322e-01 -3.53549629e-01
-6.00193180e-02 -8.70143175e-02 -2.68473923e-01 -5.68919361e-01
-5.16784787e-01 -7.70823419e-01 -1.32976770e+00 -6.62179962e-02
-5.88842213e-01 6.64493814e-02 1.09823120e+00 9.23027158e-01
3.98776054e-01 -4.82746214e-02 7.66275167e-01 -1.25633824e+00
-6.36224806e-01 -7.06311584e-01 -6.89137936e-01 5.62737823e-01
1.70972258e-01 -1.23449028e+00 -3.23995143e-01 -6.60118535e-02] | [7.613922119140625, -1.3060818910598755] |
91cf0252-50ee-42b4-b68f-cbcd7c9d1a43 | book2movie-aligning-video-scenes-with-book | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Tapaswi_Book2Movie_Aligning_Video_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Tapaswi_Book2Movie_Aligning_Video_2015_CVPR_paper.pdf | Book2Movie: Aligning Video Scenes With Book Chapters | Film adaptations of novels often visually display in a few shots what is described in many pages of the source novel. In this paper we present a new problem: to align book chapters with video scenes. Such an alignment facilitates finding differences between the adaptation and the original source, and also acts as a basis for deriving rich descriptions from the novel for the video clips. We propose an efficient method to compute an alignment between book chapters and video scenes using matching dialogs and character identities as cues. A major consideration is to allow the alignment to be non-sequential. Our suggested shortest path based approach deals with the non-sequential alignments and can be used to determine whether a video scene was part of the original book. We create a new data set involving two popular novel-to-film adaptations with widely varying properties and compare our method against other text-to-video alignment baselines. Using the alignment, we present a qualitative analysis of describing the video through rich narratives obtained from the novel. | ['Rainer Stiefelhagen', 'Martin Bauml', 'Makarand Tapaswi'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['video-alignment'] | ['computer-vision'] | [ 4.12318438e-01 -1.94632098e-01 -2.32850656e-01 -3.53239596e-01
-1.06577241e+00 -9.86299276e-01 9.05255139e-01 2.15131044e-01
-1.72995254e-01 5.19409478e-01 9.95589435e-01 3.76711667e-01
1.94043443e-02 -3.44248980e-01 -6.60058498e-01 -2.44507357e-01
3.08333457e-01 3.52525771e-01 4.90987659e-01 -2.44539842e-01
7.79787779e-01 3.20832670e-01 -1.21693885e+00 9.92875040e-01
1.45690173e-01 4.32439446e-01 3.86876553e-01 1.07894456e+00
-4.75709528e-01 9.26650405e-01 -8.66242349e-01 -9.81005490e-01
3.10436219e-01 -1.12914467e+00 -9.16676283e-01 5.65979958e-01
1.10294735e+00 -3.74736309e-01 -4.68913674e-01 1.02595854e+00
2.91946322e-01 4.60381806e-01 6.87169194e-01 -1.11754525e+00
-4.63595599e-01 9.28963900e-01 -5.21114647e-01 5.76857150e-01
1.20732856e+00 -1.54387474e-01 1.03731883e+00 -8.86224210e-01
1.57063174e+00 1.08735609e+00 8.51612926e-01 2.88170546e-01
-1.15889370e+00 -3.89329970e-01 4.67932373e-02 6.68834150e-01
-1.43980825e+00 -1.06111526e+00 9.03858125e-01 -6.98487580e-01
7.35674202e-01 5.09185016e-01 1.09538472e+00 1.51730573e+00
2.05486745e-01 6.16130829e-01 6.21421933e-01 -4.88155425e-01
4.75343270e-03 2.23397031e-01 -3.01899940e-01 5.30274272e-01
-4.10432547e-01 -5.77430725e-01 -1.09775388e+00 1.96612161e-02
7.50177979e-01 -2.54117072e-01 -2.69426405e-01 -3.85393761e-02
-1.36072266e+00 6.29960656e-01 -2.10273266e-01 4.23983306e-01
-2.27465123e-01 -4.40120399e-02 5.49232006e-01 2.87967145e-01
2.62720019e-01 7.46431768e-01 3.57994258e-01 -7.29382455e-01
-1.41978753e+00 6.90293074e-01 1.04764509e+00 1.43922758e+00
4.99144793e-01 -2.31466323e-01 -2.18953028e-01 5.83994865e-01
-2.27446929e-01 2.07221527e-02 3.95155668e-01 -1.39499593e+00
6.81683540e-01 3.40718091e-01 -1.46610243e-02 -1.46524429e+00
-5.18381707e-02 2.35788867e-01 -1.81644887e-01 -2.94305921e-01
3.60034734e-01 3.70380104e-01 -1.83943480e-01 1.15532899e+00
1.63639218e-01 1.48258060e-01 -9.55076739e-02 6.79409027e-01
1.01391387e+00 7.07028568e-01 -5.13876259e-01 -6.96894944e-01
1.22821593e+00 -1.23107696e+00 -1.15476155e+00 -2.29624081e-02
4.66820985e-01 -1.26568246e+00 1.26639795e+00 1.54010996e-01
-1.49016631e+00 -5.48261106e-01 -1.10578501e+00 -3.12330097e-01
7.34972879e-02 -2.83608019e-01 -2.54352093e-01 3.52532938e-02
-9.04600203e-01 7.19503760e-01 -3.12951684e-01 -9.67693388e-01
1.03821874e-01 -2.68100381e-01 -6.49362564e-01 5.73766418e-02
-5.07003605e-01 7.33480096e-01 2.97819704e-01 -4.69826728e-01
-5.90358555e-01 -5.72360575e-01 -7.82208383e-01 -9.27684829e-02
7.06021130e-01 -4.81659740e-01 1.57641017e+00 -1.48466408e+00
-1.52803218e+00 9.23834383e-01 -5.02030551e-01 -1.84595689e-01
6.02625489e-01 -2.97747970e-01 -4.14357513e-01 6.67510211e-01
3.44712675e-01 3.86197239e-01 8.94840240e-01 -1.12108457e+00
-8.06811929e-01 4.56408225e-02 2.68016189e-01 5.25831163e-01
-3.28871697e-01 4.86720204e-01 -8.91798794e-01 -1.03395951e+00
2.43889596e-02 -7.59894490e-01 1.51165500e-01 -3.36931832e-02
-5.62105656e-01 3.60314399e-01 6.12150252e-01 -1.12866783e+00
1.62146783e+00 -2.09581971e+00 5.69962680e-01 7.14483783e-02
1.93485737e-01 -4.66906816e-01 -1.67795569e-01 1.00019217e+00
2.61417627e-01 2.00484201e-01 -1.06120020e-01 -4.12653983e-01
-1.74814388e-01 -8.35222006e-03 -2.36295074e-01 5.01260579e-01
-2.15301469e-01 6.15591824e-01 -9.49215055e-01 -9.40896511e-01
-2.60137737e-01 1.35685697e-01 -4.78122860e-01 4.62477863e-01
-2.04128608e-01 3.93141776e-01 -2.12832168e-02 5.20967424e-01
1.62738681e-01 1.77735031e-01 1.06532149e-01 -3.74395937e-01
-4.43969071e-01 1.45122662e-01 -1.03873277e+00 2.05091429e+00
2.28954539e-01 1.44691193e+00 -2.89699852e-01 -5.39379239e-01
7.97533333e-01 3.64731073e-01 5.14055252e-01 -6.53231919e-01
-1.16595015e-01 -1.90150291e-01 -1.90223694e-01 -1.11123109e+00
1.01179838e+00 5.48160868e-04 -1.93785325e-01 4.83785033e-01
5.11560403e-02 -3.78574938e-01 7.33643174e-01 5.02302647e-01
1.00339317e+00 4.38157976e-01 8.15179050e-01 -6.66881576e-02
4.45949227e-01 5.96675098e-01 5.28781056e-01 8.36549461e-01
-1.21356972e-01 1.04197919e+00 7.40201533e-01 -4.81691211e-01
-1.56655133e+00 -9.78842199e-01 2.88776964e-01 9.88273501e-01
1.92052022e-01 -1.26347315e+00 -9.92290378e-01 -4.39155132e-01
-5.82128704e-01 6.27662063e-01 -6.46132112e-01 2.83306658e-01
-8.67687583e-01 1.85051143e-01 5.15847743e-01 3.90376270e-01
3.53403330e-01 -7.65176952e-01 -5.29222131e-01 2.11061910e-01
-5.26934206e-01 -1.51130438e+00 -9.62912083e-01 -4.19800729e-01
-4.75646287e-01 -1.16545236e+00 -4.77911144e-01 -7.57278800e-01
5.75998724e-01 3.54604900e-01 1.09544384e+00 -2.14322396e-02
2.18751300e-02 6.76849723e-01 -8.13316166e-01 -1.36397528e-02
-7.07562566e-01 6.98507801e-02 1.77439488e-02 -1.32168353e-01
2.65848547e-01 -5.10629296e-01 -6.21633194e-02 3.35097492e-01
-8.44836056e-01 6.75868809e-01 -6.05731085e-02 5.20996034e-01
6.24300599e-01 -1.92316905e-01 -6.71325028e-02 -9.76513445e-01
5.11120498e-01 -4.48856980e-01 -1.48315609e-01 4.03104007e-01
5.30291758e-02 -2.79441714e-01 7.27213442e-01 -5.68532705e-01
-1.15652835e+00 1.76625818e-01 2.59380452e-02 -4.62283045e-01
-1.38027491e-02 6.40043736e-01 4.13888618e-02 2.09448650e-01
6.78398788e-01 -6.85155019e-02 -3.60135764e-01 -4.78291422e-01
3.67924362e-01 5.36633432e-01 1.07318008e+00 -5.49058795e-01
1.07965195e+00 5.32540441e-01 -2.62316436e-01 -1.14185989e+00
-7.59875059e-01 -4.78556216e-01 -1.30605161e+00 -7.33797550e-01
8.89289260e-01 -6.64594889e-01 -1.57847524e-01 -9.62648168e-02
-1.44600749e+00 -2.23102927e-01 -5.01283407e-01 3.35945249e-01
-8.29277337e-01 4.33325946e-01 -5.93438029e-01 -3.65383834e-01
1.91632986e-01 -9.99442577e-01 5.45518517e-01 3.59359592e-01
-9.31413829e-01 -9.65034366e-01 3.78859490e-01 3.04431796e-01
-6.41202554e-02 6.52435005e-01 6.54646039e-01 -8.31493139e-01
-3.85495186e-01 -8.54786932e-02 9.76171419e-02 -3.89435053e-01
2.33297244e-01 7.61324883e-01 -6.71280444e-01 -8.55869725e-02
4.69307154e-02 1.32127613e-01 2.23102704e-01 1.64892405e-01
5.51348090e-01 -8.61116052e-01 -1.40697747e-01 9.08595383e-01
1.22346890e+00 5.41823268e-01 6.09346092e-01 8.30215216e-01
8.93467963e-01 9.03446078e-01 7.41130650e-01 7.25003660e-01
3.58259946e-01 1.03438199e+00 -6.76303729e-02 5.03585458e-01
-2.46337801e-01 -5.58937669e-01 7.32094288e-01 1.36079741e+00
-1.36643037e-01 -3.21853101e-01 -9.74822402e-01 6.02960229e-01
-2.07407403e+00 -1.66578901e+00 -2.59705752e-01 1.82461023e+00
6.98772669e-01 1.98837351e-02 4.85944957e-01 -1.90679654e-01
8.28559577e-01 2.98421651e-01 -1.02125138e-01 -5.23391008e-01
-4.43154752e-01 -3.57209086e-01 1.29846884e-02 4.08986896e-01
-8.81392300e-01 8.53729010e-01 7.06812954e+00 6.73901737e-01
-6.50420129e-01 1.12309195e-01 2.59748101e-01 -6.30235970e-01
-2.68326849e-01 2.50561267e-01 -6.58205569e-01 3.61808807e-01
9.82566357e-01 -7.75132060e-01 5.03644526e-01 5.66078246e-01
4.06633198e-01 -2.57692993e-01 -1.63498938e+00 1.10620666e+00
8.00433517e-01 -1.88866556e+00 1.24283835e-01 -3.51397872e-01
9.79820728e-01 -7.31019378e-01 -2.82335639e-01 -1.89247176e-01
-2.44322702e-01 -8.28597426e-01 1.40048540e+00 5.83142102e-01
7.65039086e-01 -6.53237641e-01 4.85055119e-01 1.66497547e-02
-1.43042791e+00 2.15804785e-01 -2.81714976e-01 -1.60913989e-01
4.48197752e-01 -2.62463242e-01 -8.78600180e-01 3.80800873e-01
8.87831509e-01 1.29045081e+00 -8.64871621e-01 1.02527606e+00
-9.40564051e-02 2.03004077e-01 2.16824129e-01 5.05696647e-02
8.06968138e-02 -3.90250862e-01 1.16195428e+00 1.69738090e+00
5.44306874e-01 1.82964146e-01 1.87905580e-01 7.18506992e-01
-1.74751505e-01 4.43039864e-01 -8.69161606e-01 -4.12813812e-01
6.92557633e-01 1.01515913e+00 -1.05599058e+00 -3.89108181e-01
-6.21998668e-01 1.39066875e+00 1.85279921e-01 3.42809618e-01
-7.69014418e-01 -2.97003835e-01 3.92858922e-01 3.74854773e-01
2.70205975e-01 -1.06503598e-01 -1.53479472e-01 -1.12187123e+00
1.09471172e-01 -1.14225972e+00 3.74956280e-01 -1.05070627e+00
-1.21622849e+00 8.01605225e-01 4.94977325e-01 -1.44165778e+00
-3.64220709e-01 -2.84903739e-02 -7.69746244e-01 3.22303742e-01
-7.47764289e-01 -9.63991463e-01 -5.39563894e-01 6.62552416e-01
1.31964660e+00 -3.67142111e-01 5.39755642e-01 4.23531644e-02
-5.92630386e-01 4.88401026e-01 3.87794137e-01 1.53315350e-01
1.25798988e+00 -1.04351532e+00 4.38324422e-01 1.29550982e+00
5.11345863e-01 4.21955734e-01 1.19821072e+00 -8.92073870e-01
-1.15033841e+00 -5.91520905e-01 8.06617796e-01 -5.97057462e-01
8.97846937e-01 -2.49699309e-01 -9.32675481e-01 9.57114220e-01
8.31057012e-01 -7.00960159e-01 9.53740239e-01 -3.30397457e-01
-2.63851672e-01 1.41442433e-01 -7.64804602e-01 1.01113951e+00
1.27358985e+00 -8.25567365e-01 -8.84837091e-01 3.61854434e-01
5.58917642e-01 -8.59534919e-01 -8.02722156e-01 -4.69338566e-01
8.22228849e-01 -1.11547148e+00 7.03299820e-01 -5.30197859e-01
1.19472957e+00 -1.59786493e-01 -4.73228633e-01 -1.19917917e+00
-2.62266666e-01 -1.24356270e+00 -5.71781248e-02 1.74259973e+00
5.89131676e-02 3.35184842e-01 5.84100425e-01 6.23583794e-01
-3.34939420e-01 -2.15084359e-01 -8.11977088e-01 -6.93977296e-01
-4.07216668e-01 -2.51967311e-01 3.61714691e-01 1.06043959e+00
3.79772514e-01 4.29124296e-01 -5.02188146e-01 -2.88296998e-01
2.78784335e-01 -8.72859061e-02 1.14271092e+00 -8.55100572e-01
-1.39240488e-01 -5.59742510e-01 -5.15350282e-01 -8.19220781e-01
2.05951214e-01 -6.07633173e-01 1.16982773e-01 -1.38751376e+00
6.77954912e-01 3.69096011e-01 3.20881516e-01 -3.10069900e-02
6.03523366e-02 3.30162227e-01 6.23896718e-01 6.38600171e-01
-7.82917559e-01 7.31377751e-02 9.36351240e-01 -1.46211803e-01
-4.41961467e-01 -5.40508270e-01 -5.42509973e-01 9.63408291e-01
4.78530884e-01 -7.08678901e-01 -2.38440782e-01 -4.11158144e-01
4.22997326e-01 2.66543239e-01 5.08446060e-03 -1.15330780e+00
5.29205441e-01 -5.38310230e-01 2.70685762e-01 -6.36686146e-01
4.22014087e-01 -6.68309987e-01 8.49329710e-01 -3.62630859e-02
-8.09456170e-01 7.90741682e-01 -1.24705303e-02 5.15162170e-01
-2.35922650e-01 -7.30575860e-01 4.77162778e-01 -3.77914339e-01
-1.04088676e+00 -1.54988825e-01 -6.02279782e-01 2.64048249e-01
1.18882811e+00 -9.54249561e-01 -4.02937233e-01 -7.70592034e-01
-4.73999113e-01 -1.22947499e-01 1.15854573e+00 4.09859598e-01
7.65197217e-01 -1.56346357e+00 -8.25800478e-01 -3.56074810e-01
2.36101270e-01 -4.22501326e-01 9.76906940e-02 7.84037709e-01
-9.81898546e-01 -3.23416665e-02 -6.79889441e-01 -4.82425392e-01
-1.73373580e+00 4.44944918e-01 -1.56273887e-01 2.27560893e-01
-1.07764816e+00 7.39033163e-01 6.83715940e-02 5.57659984e-01
2.82778293e-01 1.50624096e-01 -4.49835300e-01 4.76810783e-01
9.48875546e-01 5.32244205e-01 -3.82142812e-01 -1.25631285e+00
-1.94133282e-01 7.02768147e-01 -2.10101947e-01 -5.06248116e-01
1.41044283e+00 -7.08139300e-01 4.99278009e-02 1.11913490e+00
1.22563362e+00 4.59573358e-01 -1.33883440e+00 -3.00108582e-01
2.58449525e-01 -1.04910779e+00 -5.18478692e-01 -2.49198720e-01
-6.09799623e-01 4.04188037e-01 -1.03310645e-01 2.13874862e-01
8.80177021e-01 4.59369197e-02 8.08469772e-01 4.98359263e-01
-7.80313015e-02 -1.34311354e+00 4.36984330e-01 3.43830675e-01
1.20442879e+00 -1.08510327e+00 4.76228416e-01 -2.82470286e-01
-1.07254493e+00 1.73338091e+00 5.85806131e-01 -1.12728231e-01
-5.64215779e-02 3.93889219e-01 2.92362422e-02 -2.22370073e-01
-6.22992039e-01 1.35403812e-01 4.94502962e-01 6.48805082e-01
4.38198328e-01 -4.75666285e-01 -1.30866513e-01 6.99212611e-01
-5.83488107e-01 -3.93301874e-01 1.31081402e+00 9.99239028e-01
-2.32993588e-01 -8.50072503e-01 -5.69097996e-01 2.22085521e-01
-4.54793602e-01 3.01370434e-02 -9.57215726e-01 9.12506104e-01
-1.46223083e-01 7.64331162e-01 4.09732789e-01 -5.69144130e-01
4.09025639e-01 1.21084054e-03 5.87451637e-01 -7.91911542e-01
-9.50495481e-01 3.39081585e-01 3.79099250e-01 -6.47910774e-01
-9.28165913e-01 -9.69264030e-01 -7.44651318e-01 -7.59088397e-01
8.32301602e-02 1.54735163e-01 2.82706320e-01 1.02686632e+00
-2.47158110e-03 3.96286547e-01 3.92353475e-01 -1.16865921e+00
1.55733749e-01 -5.84852278e-01 -2.99113125e-01 8.41551185e-01
2.69180655e-01 -3.47753406e-01 -3.61724272e-02 9.40354526e-01] | [10.704846382141113, 0.6417894959449768] |
8adec524-4058-4e8d-ab56-b059ca901730 | linguistically-informed-self-attention-for-1 | null | null | https://openreview.net/forum?id=Bk7GGldiz | https://openreview.net/pdf?id=Bk7GGldiz | Linguistically-Informed Self-Attention for Semantic Role Labeling | The current state-of-the-art end-to-end semantic role labeling (SRL) model is a deep neural network architecture with no explicit linguistic features.
However, prior work has shown that gold syntax trees can dramatically improve SRL, suggesting that neural network models could see great improvements from explicit modeling of syntax.
In this work, we present linguistically-informed self-attention (LISA): a new neural network model that combines
multi-head self-attention with multi-task learning across dependency parsing, part-of-speech, predicate detection and SRL. For example, syntax is incorporated by training one of the attention heads to attend to syntactic parents for each token. Our model can predict all of the above tasks, but it is also trained such that if a high-quality syntactic parse is already available, it can be beneficially injected at test time without re-training our SRL model.
In experiments on the CoNLL-2005 SRL dataset LISA achieves an increase of 2.5 F1 absolute over the previous state-of-the-art on newswire with predicted predicates and more than 2.0 F1 on out-of-domain data. On ConLL-2012 English SRL we also show an improvement of more than 3.0 F1, a 13% reduction in error. | ['Anonymous'] | 2018-04-08 | null | null | null | null | ['predicate-detection', 'semantic-role-labeling'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.43462080e-01 8.44218969e-01 -3.44492674e-01 -9.72815692e-01
-1.34939659e+00 -6.95214868e-01 2.66038090e-01 2.78284580e-01
-6.43117964e-01 7.53730297e-01 7.15303063e-01 -4.55326498e-01
1.92275137e-01 -5.90088904e-01 -1.10623085e+00 -2.29061127e-01
4.61351611e-02 8.55253637e-01 3.22014779e-01 -4.05215204e-01
-2.80521899e-01 4.75398004e-02 -1.32811558e+00 8.66957784e-01
5.96027255e-01 8.98704946e-01 2.15034947e-01 7.24818885e-01
-3.38203222e-01 1.26328349e+00 -7.36631155e-01 -2.63490111e-01
-9.98341218e-02 -1.40414879e-01 -1.25881851e+00 -4.56582874e-01
5.58557093e-01 -1.03501447e-01 3.47864628e-02 7.66794622e-01
5.81464708e-01 5.64656146e-02 -2.48292834e-02 -7.98980355e-01
-9.12161291e-01 1.31306660e+00 -3.02211016e-01 3.35449308e-01
2.85246164e-01 -1.00149913e-02 1.77583838e+00 -7.77818620e-01
6.48115337e-01 1.74825025e+00 6.42998576e-01 8.29019785e-01
-1.22188067e+00 -6.17150903e-01 3.70130271e-01 2.86447890e-02
-6.32784843e-01 -3.95315140e-01 6.27208412e-01 -8.83436725e-02
1.57703352e+00 -1.12083122e-01 8.14037398e-02 9.79929209e-01
-5.33889532e-02 1.26064825e+00 9.82112527e-01 -5.61610520e-01
-8.85276347e-02 -3.81491572e-01 6.49045289e-01 7.59954751e-01
9.96179432e-02 -9.84868109e-02 -5.49155354e-01 6.46419898e-02
1.41129360e-01 -6.12012923e-01 1.53385535e-01 2.65078545e-01
-9.97304082e-01 9.97060657e-01 5.17954350e-01 3.23878229e-01
-1.78034350e-01 3.24796557e-01 6.49853289e-01 1.71396375e-01
6.66741073e-01 6.29108608e-01 -1.16362739e+00 -8.50317031e-02
-6.47155523e-01 1.88995421e-01 7.01853275e-01 5.85453153e-01
3.87946546e-01 4.67087179e-02 -2.61876464e-01 1.13225937e+00
3.35317165e-01 2.94042826e-01 6.17122889e-01 -1.08684790e+00
5.66798210e-01 6.25751436e-01 -2.36947313e-01 -6.17560931e-02
-7.53169537e-01 -3.39341074e-01 -2.09289148e-01 6.90018386e-02
6.13281488e-01 -3.45309764e-01 -1.12300396e+00 2.14793921e+00
1.01841792e-01 -1.03658132e-01 4.06889588e-01 6.07274711e-01
1.16934097e+00 7.19834685e-01 7.30584145e-01 2.43025631e-01
1.51004922e+00 -1.19040847e+00 -7.24438190e-01 -8.01806986e-01
1.05892527e+00 -4.92615700e-01 1.23270667e+00 1.75076678e-01
-1.26314855e+00 -7.10350871e-01 -8.61774862e-01 -5.54234028e-01
-3.24378997e-01 6.21274821e-02 6.99178576e-01 4.68122959e-01
-1.15272546e+00 6.58580601e-01 -8.25967729e-01 -1.42213151e-01
5.92737556e-01 5.57148755e-01 -3.96246880e-01 -2.40958974e-01
-1.45167863e+00 8.21815729e-01 6.69936180e-01 -2.04685286e-01
-9.30704832e-01 -9.00667429e-01 -1.26829326e+00 1.80720791e-01
5.07241249e-01 -3.38201791e-01 1.74250054e+00 -1.04066443e+00
-1.30944824e+00 1.20704198e+00 -2.30740398e-01 -7.41461694e-01
-3.73824574e-02 -6.09164417e-01 -2.77050525e-01 -1.70070753e-01
5.83363235e-01 9.37161565e-01 2.01348975e-01 -1.06584573e+00
-8.92132044e-01 -3.22806060e-01 3.15174907e-01 2.15291366e-01
1.03261404e-01 5.44875383e-01 -1.39575049e-01 -4.75282609e-01
-2.55540669e-01 -6.82107627e-01 -1.89853936e-01 -4.54639494e-01
-2.85458148e-01 -8.30886424e-01 5.50407588e-01 -8.70864272e-01
8.89999330e-01 -2.18216681e+00 -1.29590318e-01 -4.74675149e-01
-9.22541320e-02 4.90580499e-01 -4.75985944e-01 -8.46528932e-02
-3.84056360e-01 4.70269531e-01 -3.81856322e-01 -6.66042447e-01
-1.73639700e-01 5.07461369e-01 -4.35075909e-03 2.55157024e-01
5.98908961e-01 8.41973364e-01 -1.12830269e+00 -3.30931485e-01
-4.36551347e-02 2.35180423e-01 -7.10980415e-01 2.40995288e-01
-5.50675213e-01 4.37832713e-01 -9.89859030e-02 6.56876802e-01
2.61425525e-01 -3.42525840e-01 2.85150439e-01 1.10983536e-01
-4.50550728e-02 1.24016488e+00 -6.35301828e-01 1.77622819e+00
-7.88665533e-01 4.73903984e-01 1.72689542e-01 -9.75126684e-01
6.74374759e-01 3.97077024e-01 2.55016476e-01 -8.78786743e-01
1.62052974e-01 3.39139134e-01 5.48210084e-01 -9.81619582e-02
2.08126023e-01 -1.36944696e-01 -4.16714430e-01 2.63648063e-01
3.98626298e-01 3.36591095e-01 3.83425117e-01 -1.59464125e-02
1.36909616e+00 2.59395629e-01 3.15609694e-01 -4.44325238e-01
5.62049031e-01 2.08623931e-01 9.27924991e-01 7.77765393e-01
-1.65747672e-01 4.11443800e-01 5.25966644e-01 -4.16653931e-01
-8.08731198e-01 -7.08149374e-01 -9.21568796e-02 1.87828696e+00
-3.52363616e-01 -2.45861799e-01 -6.50005043e-01 -1.07630491e+00
7.30284676e-02 1.09378707e+00 -6.35652065e-01 1.04416370e-01
-1.16886318e+00 -7.60648966e-01 8.80056739e-01 9.65814888e-01
5.21704078e-01 -1.68619084e+00 -6.30099058e-01 6.18004143e-01
-2.38881279e-02 -1.40056634e+00 -9.72518399e-02 6.68354213e-01
-5.34243584e-01 -1.07761252e+00 -3.59680444e-01 -1.02457488e+00
2.36007690e-01 -3.75285983e-01 1.55298960e+00 1.47493482e-01
7.07838610e-02 -1.72710881e-01 -5.55999994e-01 -4.60691273e-01
-6.82827234e-01 2.12961450e-01 -3.14593762e-01 -2.25742057e-01
3.51374865e-01 -2.43071154e-01 -1.91712305e-01 -2.30300218e-01
-2.87675649e-01 4.04702872e-02 4.50473219e-01 1.08015561e+00
7.09041893e-01 -1.88767523e-01 7.20703125e-01 -1.53869951e+00
2.99650669e-01 -4.52950388e-01 -4.24327165e-01 1.65596768e-01
-3.19361001e-01 3.53717625e-01 7.98060238e-01 -4.51274030e-02
-1.38825691e+00 1.64707839e-01 -7.35430598e-01 7.26344483e-03
-4.29970562e-01 5.24758339e-01 -5.05742729e-01 4.82437968e-01
5.32534897e-01 -2.59894520e-01 -3.09007198e-01 -7.19876587e-01
6.00971758e-01 4.88320500e-01 6.03946328e-01 -8.14750850e-01
1.83911964e-01 1.00743018e-01 -2.07427889e-01 -5.08181393e-01
-1.72138870e+00 -4.26897079e-01 -5.92600346e-01 4.14463967e-01
1.20807183e+00 -1.09299362e+00 -6.52819991e-01 3.55805308e-01
-1.30696023e+00 -1.04441226e+00 -3.73598605e-01 1.59491412e-02
-4.30611759e-01 -1.26161557e-02 -9.55551207e-01 -5.64351916e-01
-4.84771550e-01 -1.00083268e+00 1.18597209e+00 -3.65285687e-02
-2.94455141e-01 -1.03930736e+00 -1.06488921e-01 6.23216271e-01
2.92555362e-01 6.29091859e-02 1.14024365e+00 -1.40741956e+00
-1.50487915e-01 2.26533085e-01 -2.44661897e-01 4.80182648e-01
-3.16941589e-02 -5.18168926e-01 -1.14903498e+00 -1.65048003e-01
-3.07032853e-01 -6.58781826e-01 1.05713558e+00 3.66710126e-01
1.20935619e+00 -1.69085607e-01 -1.89367056e-01 4.34918493e-01
1.23810434e+00 3.43803138e-01 2.54802823e-01 4.38049942e-01
8.32319200e-01 6.70214415e-01 7.67942548e-01 -1.53731674e-01
6.31209016e-01 3.94971669e-01 5.02034187e-01 -2.59613663e-01
-5.68538189e-01 -2.86350608e-01 4.20696706e-01 3.00604254e-01
4.00587350e-01 -4.67047662e-01 -1.15055978e+00 8.51352811e-01
-1.75516939e+00 -6.77789569e-01 -2.22762048e-01 1.75562549e+00
1.07323408e+00 6.08892977e-01 -3.22563909e-02 1.04276396e-01
5.73449492e-01 3.99513394e-01 -5.65720439e-01 -6.80793107e-01
-1.49705470e-01 7.78278649e-01 7.20465660e-01 7.67147005e-01
-1.41737735e+00 1.70169687e+00 6.01356840e+00 5.91520309e-01
-8.65435481e-01 5.58074236e-01 8.36121976e-01 1.62366077e-01
-2.47293815e-01 7.65382797e-02 -1.42440772e+00 3.14120263e-01
1.37897050e+00 1.63441643e-01 6.57121912e-02 1.08254576e+00
-1.21664226e-01 -2.44420045e-03 -1.04683435e+00 3.76047015e-01
8.78461823e-02 -1.18711245e+00 -1.10399820e-01 -2.11867705e-01
3.50183815e-01 5.74638247e-01 -4.41327095e-01 8.57563972e-01
1.00880015e+00 -1.08405626e+00 8.68950784e-01 -3.25765371e-01
8.48390818e-01 -6.91560805e-01 9.04901505e-01 4.49500829e-01
-1.12601364e+00 -3.17958385e-01 -2.47775450e-01 -2.87221342e-01
4.10771668e-01 4.75799114e-01 -8.75544846e-01 2.60435879e-01
7.39620745e-01 7.81712890e-01 -5.54552972e-01 5.44272423e-01
-1.00553393e+00 1.17548048e+00 -3.29501778e-01 -7.15866759e-02
5.07559955e-01 6.32396638e-01 3.59619141e-01 1.42706454e+00
-2.65145123e-01 1.35734111e-01 4.68665361e-01 6.66708708e-01
-4.16379899e-01 1.60865560e-01 -4.08768028e-01 -2.10733209e-02
4.14299875e-01 1.04321837e+00 -4.87014264e-01 -4.59652781e-01
-5.24575293e-01 8.13590765e-01 7.92618811e-01 -1.26367375e-01
-6.74217045e-01 -1.21091723e-01 7.24844754e-01 5.61081618e-02
2.98699141e-01 3.22009996e-02 -4.33337063e-01 -7.25350380e-01
-3.94634128e-01 -7.19827116e-01 7.59667337e-01 -7.31018901e-01
-1.25387859e+00 6.89714551e-01 -7.82940090e-02 -2.93624371e-01
-2.67058194e-01 -1.03887558e+00 -5.01564622e-01 9.38613474e-01
-1.91313910e+00 -1.51912761e+00 3.07010651e-01 2.59449005e-01
1.03052807e+00 -1.17093213e-01 9.40412104e-01 3.20718348e-01
-4.56538737e-01 7.47351646e-01 -3.97961527e-01 6.61922157e-01
6.13912344e-01 -1.76099300e+00 9.33327138e-01 1.00701571e+00
8.83520842e-02 3.51708800e-01 3.67650092e-01 -7.12100387e-01
-7.61447012e-01 -1.33480036e+00 1.52058637e+00 -6.39340758e-01
6.97764874e-01 -4.04729724e-01 -1.00623441e+00 1.16848075e+00
2.53729254e-01 2.28162855e-01 7.19267905e-01 5.73194504e-01
-3.73483747e-01 1.59861684e-01 -9.89439428e-01 1.22146577e-01
1.16827095e+00 -3.64913195e-01 -9.99147773e-01 4.12331283e-01
1.28176570e+00 -5.89438260e-01 -7.55544186e-01 6.54966831e-01
1.44320861e-01 -6.35238469e-01 7.78429449e-01 -8.46784711e-01
5.91514111e-01 1.12776319e-02 -3.42856109e-01 -1.22414351e+00
-4.36992466e-01 -2.62186289e-01 2.08281368e-01 1.37834728e+00
8.88221979e-01 -3.94572556e-01 6.52142406e-01 5.42944431e-01
-7.57216215e-01 -6.53046429e-01 -9.14969683e-01 -6.34719253e-01
3.57708305e-01 -6.21037602e-01 3.08126032e-01 8.22373450e-01
-2.95610815e-01 1.00415695e+00 -4.41578105e-02 2.18071327e-01
4.83156323e-01 -1.09889992e-01 3.01215619e-01 -1.33412111e+00
-4.74761069e-01 -2.64334321e-01 1.50485158e-01 -9.72750068e-01
9.98303652e-01 -1.31746864e+00 3.23222697e-01 -1.75584996e+00
9.50169004e-03 -6.03516340e-01 -4.30967301e-01 1.32841575e+00
-3.66065502e-01 -1.11609429e-01 2.47939110e-01 -2.53905386e-01
-7.41132677e-01 2.65226573e-01 9.24120843e-01 9.50209349e-02
-9.25229937e-02 -1.98802292e-01 -1.00172162e+00 9.45379496e-01
8.79126251e-01 -7.67140746e-01 -1.11419417e-01 -7.84380257e-01
3.46677080e-02 9.90053266e-02 3.56996804e-02 -8.53307307e-01
-1.68218911e-01 1.97394282e-01 1.75703555e-01 -4.73549873e-01
3.49515885e-01 -4.45545614e-01 -6.26840949e-01 4.32836235e-01
-7.26394355e-01 -2.33139899e-02 4.86059099e-01 3.28159422e-01
-1.80869713e-01 -4.15482312e-01 8.54328990e-01 -5.04763961e-01
-9.81890440e-01 -2.72543728e-02 -2.74955839e-01 5.13634920e-01
4.91803110e-01 3.22092384e-01 -5.10989845e-01 -3.55983004e-02
-9.21138704e-01 5.17243207e-01 -1.10849991e-01 5.33985674e-01
1.52555794e-01 -8.91293108e-01 -9.36997652e-01 2.56451190e-01
-1.56909004e-02 4.04698521e-01 -6.10576309e-02 1.67482883e-01
-3.24289292e-01 4.80936646e-01 1.28457248e-01 -3.84636223e-01
-1.34239376e+00 2.79468775e-01 2.58218676e-01 -8.65365386e-01
-4.74454939e-01 1.28708315e+00 1.97970197e-01 -7.94873476e-01
1.38904363e-01 -4.60500449e-01 -4.14811939e-01 -1.26043096e-01
5.11052549e-01 -1.32917538e-01 4.16843332e-02 -7.11159647e-01
-6.13657355e-01 2.36393377e-01 -2.38563225e-01 -1.15603685e-01
1.54921031e+00 3.62703890e-01 -6.65376484e-02 3.55029553e-01
1.01369822e+00 7.39705265e-02 -1.22651267e+00 -2.64655799e-01
5.09354174e-01 1.12319037e-01 6.05401807e-02 -1.39451098e+00
-9.35098588e-01 1.01541805e+00 2.13250443e-01 -9.61978510e-02
8.98010552e-01 4.64140981e-01 9.63965595e-01 3.59024286e-01
2.03758299e-01 -8.58032048e-01 -7.10805820e-04 1.14822936e+00
6.90876305e-01 -1.42111146e+00 -4.57933426e-01 -5.30039787e-01
-6.39114201e-01 6.99363649e-01 8.04259479e-01 -2.70690084e-01
5.91439545e-01 4.65450227e-01 4.84185554e-02 -2.28244081e-01
-8.85549307e-01 -5.51572561e-01 7.36076981e-02 3.96077722e-01
9.09980595e-01 1.48221985e-01 -2.58720547e-01 1.05805993e+00
-3.39506984e-01 -3.45774174e-01 2.51333922e-01 9.41059887e-01
-6.42724276e-01 -1.39718759e+00 1.04509801e-01 2.59428680e-01
-1.13681400e+00 -5.54779172e-01 -2.77155310e-01 9.05672371e-01
1.71868801e-01 1.02328491e+00 2.03071222e-01 6.25276715e-02
5.46428919e-01 4.90918934e-01 1.85822114e-01 -1.32775497e+00
-9.36815917e-01 -3.36403847e-02 8.34030628e-01 -7.08097577e-01
-4.78956729e-01 -6.38170898e-01 -1.93825901e+00 2.79770136e-01
-3.35258320e-02 -6.57894015e-02 5.57726383e-01 1.10021257e+00
2.65896350e-01 7.74713099e-01 1.59265041e-01 -5.84265649e-01
-3.55113804e-01 -1.03539729e+00 -1.21197015e-01 4.87502873e-01
2.84977317e-01 -5.51400840e-01 -1.93014145e-01 -8.38585347e-02] | [10.451192855834961, 9.442742347717285] |
19bdc023-e5aa-4f44-b0f3-c2ba39cd3f4c | lstm-cf-unifying-context-modeling-and-fusion | 1604.05000 | null | http://arxiv.org/abs/1604.05000v3 | http://arxiv.org/pdf/1604.05000v3.pdf | LSTM-CF: Unifying Context Modeling and Fusion with LSTMs for RGB-D Scene Labeling | Semantic labeling of RGB-D scenes is crucial to many intelligent applications
including perceptual robotics. It generates pixelwise and fine-grained label
maps from simultaneously sensed photometric (RGB) and depth channels. This
paper addresses this problem by i) developing a novel Long Short-Term Memorized
Context Fusion (LSTM-CF) Model that captures and fuses contextual information
from multiple channels of photometric and depth data, and ii) incorporating
this model into deep convolutional neural networks (CNNs) for end-to-end
training. Specifically, contexts in photometric and depth channels are,
respectively, captured by stacking several convolutional layers and a long
short-term memory layer; the memory layer encodes both short-range and
long-range spatial dependencies in an image along the vertical direction.
Another long short-term memorized fusion layer is set up to integrate the
contexts along the vertical direction from different channels, and perform
bi-directional propagation of the fused vertical contexts along the horizontal
direction to obtain true 2D global contexts. At last, the fused contextual
representation is concatenated with the convolutional features extracted from
the photometric channels in order to improve the accuracy of fine-scale
semantic labeling. Our proposed model has set a new state of the art, i.e.,
48.1% and 49.4% average class accuracy over 37 categories (2.2% and 5.4%
improvement) on the large-scale SUNRGBD dataset and the NYUDv2dataset,
respectively. | ['Liang Lin', 'Hui Cheng', 'Yukang Gan', 'Xiaodan Liang', 'Yizhou Yu', 'Zhen Li'] | 2016-04-18 | null | null | null | null | ['scene-labeling'] | ['computer-vision'] | [ 4.43350703e-01 -9.10431445e-02 8.47286582e-02 -8.93692791e-01
-8.41506660e-01 -4.12525535e-01 5.97871602e-01 2.27023941e-02
-5.54033577e-01 6.21597230e-01 1.52267674e-02 7.21717775e-02
1.31491736e-01 -1.17210448e+00 -9.51442003e-01 -9.09627914e-01
3.64809871e-01 1.43154217e-02 4.73541677e-01 2.56072562e-02
1.73421532e-01 6.32560670e-01 -1.92175210e+00 6.62435293e-01
5.32049060e-01 1.84377789e+00 6.49602950e-01 5.66270411e-01
-6.40225112e-01 9.06776071e-01 -3.70136350e-01 1.04568131e-01
5.64408600e-02 -3.33835976e-03 -8.16967666e-01 1.79440454e-01
7.25917935e-01 -3.21117163e-01 -1.29132509e-01 1.05311418e+00
4.04343784e-01 2.67750740e-01 2.83810705e-01 -7.84309506e-01
-8.80120814e-01 5.61333224e-02 -5.35382092e-01 -1.22798942e-01
5.07984422e-02 2.01498687e-01 6.01358712e-01 -8.58461142e-01
4.96771365e-01 1.30423188e+00 3.45474690e-01 3.37563425e-01
-7.41099358e-01 -6.05348229e-01 4.23464060e-01 2.36815557e-01
-1.25120616e+00 3.68645974e-02 9.20887351e-01 -3.80525738e-01
1.16514277e+00 -3.07284333e-02 6.66075408e-01 9.21941996e-01
1.94591880e-01 5.51157653e-01 1.60488331e+00 -3.21535349e-01
2.74949253e-01 -2.95391709e-01 1.44590050e-01 7.38615811e-01
-2.14456558e-01 2.56445885e-01 -5.59877634e-01 4.08177406e-01
8.67915213e-01 2.28757307e-01 1.61417499e-01 3.03051248e-02
-1.15736055e+00 6.15674853e-01 1.20188105e+00 2.82793850e-01
-4.51516509e-01 2.90762484e-01 -5.16575016e-02 -4.34995405e-02
5.41573882e-01 2.45238729e-02 -6.34469271e-01 2.43861139e-01
-6.31939650e-01 -5.22378981e-02 2.17529446e-01 9.18861449e-01
1.35983145e+00 -2.61286408e-01 -6.81749210e-02 9.97637093e-01
5.82863152e-01 7.68167377e-01 4.57546860e-01 -1.07886088e+00
3.65235120e-01 7.98475802e-01 3.10662333e-02 -8.58193398e-01
-6.70342207e-01 -3.99109155e-01 -6.64918840e-01 3.59221905e-01
1.88903987e-01 1.39106244e-01 -1.37691951e+00 1.80727601e+00
4.65119541e-01 3.63856614e-01 1.00313641e-01 1.10155714e+00
1.12287879e+00 6.89698577e-01 3.27419847e-01 1.72466338e-01
1.25932634e+00 -1.02532673e+00 -4.68364269e-01 -6.14628255e-01
3.17455113e-01 -8.10175359e-01 1.09108412e+00 1.53420746e-01
-8.49178195e-01 -9.80921268e-01 -1.15263295e+00 -7.07466125e-01
-8.42129767e-01 2.70059705e-01 4.26629633e-01 1.01592436e-01
-1.22834110e+00 2.34795690e-01 -7.27744699e-01 -2.51901686e-01
3.05190653e-01 2.70052850e-01 -2.93308407e-01 -3.79974902e-01
-1.19903398e+00 6.28973067e-01 5.00193715e-01 4.09137100e-01
-8.30067933e-01 -2.82472253e-01 -1.04884267e+00 -2.92922050e-01
1.32767886e-01 -6.58781946e-01 1.09051371e+00 -9.06821847e-01
-1.36705887e+00 1.10467982e+00 -4.21594292e-01 -9.37853083e-02
7.49562159e-02 -2.20145687e-01 -2.77345657e-01 8.42659324e-02
2.33252347e-01 1.11081123e+00 5.91379821e-01 -1.43593395e+00
-9.72158730e-01 -8.26755822e-01 1.99598566e-01 4.89963174e-01
-1.20981656e-01 -5.75966239e-01 -6.43462181e-01 -5.61706185e-01
6.56828701e-01 -8.76730859e-01 -6.29470497e-02 1.23391105e-02
-3.43864769e-01 -2.23458514e-01 9.41776156e-01 -6.71644807e-01
5.50323367e-01 -2.08066654e+00 9.83551815e-02 -7.99592435e-02
-7.94737637e-02 7.69572556e-02 -1.05848707e-01 -1.41187772e-01
1.68915480e-01 -3.29687923e-01 -3.37469429e-01 -6.56675696e-01
-5.89041412e-02 4.56787169e-01 -2.35046566e-01 2.15168491e-01
1.89033270e-01 9.49916005e-01 -9.54273582e-01 -2.20178962e-01
8.61045480e-01 6.59758449e-01 -1.32903785e-01 2.88158238e-01
-4.84862268e-01 7.17618585e-01 -3.00214648e-01 7.90626705e-01
8.99054408e-01 -5.31516857e-02 -1.03403047e-01 -4.20526981e-01
-2.40644440e-01 3.46043646e-01 -9.46799517e-01 2.37439036e+00
-8.18024218e-01 5.06216884e-01 -1.30468979e-01 -6.24550462e-01
1.17260575e+00 -1.61974952e-01 2.15839401e-01 -1.23349643e+00
3.11629742e-01 2.76073068e-01 -7.43415177e-01 -4.92297232e-01
5.73116779e-01 -1.80202425e-01 -2.72938967e-01 4.62083854e-02
1.43454403e-01 -3.37959737e-01 -3.85428458e-01 -3.75898659e-01
5.45081139e-01 4.13743258e-01 -1.76986650e-01 6.30471185e-02
6.20573819e-01 -6.49514273e-02 4.53318179e-01 5.24830878e-01
-1.52267709e-01 7.11707413e-01 -1.36964455e-01 -6.20802045e-01
-6.40136898e-01 -1.08828294e+00 -5.00807073e-04 1.04282761e+00
6.28082633e-01 1.38767272e-01 -5.39884448e-01 -5.08301914e-01
1.23138785e-01 5.87290823e-01 -7.88535058e-01 -2.08125278e-01
-4.38712060e-01 -5.11154234e-01 2.71761924e-01 7.79722512e-01
1.34302127e+00 -1.12866378e+00 -6.34194672e-01 1.95426241e-01
-1.91085458e-01 -1.34100473e+00 1.15471128e-02 4.24139947e-01
-8.23984444e-01 -8.92924309e-01 -3.94331396e-01 -9.44180191e-01
2.81989157e-01 4.45059896e-01 8.87128234e-01 -4.21038479e-01
-9.25828665e-02 -1.88286323e-02 -2.55728871e-01 -1.63067847e-01
2.51567960e-01 -7.69444974e-03 -4.35204506e-01 7.15157092e-02
5.78704000e-01 -4.82792497e-01 -6.67297363e-01 4.34592932e-01
-9.66826379e-01 2.78577924e-01 6.17091358e-01 7.57478416e-01
1.24362683e+00 -2.22786054e-01 3.30363929e-01 -5.36499202e-01
-9.20081884e-02 -2.74953574e-01 -7.07814753e-01 2.71718591e-01
-2.94190854e-01 5.23654334e-02 2.00729683e-01 4.37321514e-02
-1.52604389e+00 3.07464033e-01 -5.00129998e-01 -4.62697715e-01
-6.65268481e-01 2.83225834e-01 -3.90687197e-01 -1.77464485e-01
3.79650980e-01 1.42668366e-01 -4.73295927e-01 -5.78008294e-01
7.88839221e-01 8.60844135e-01 9.36777413e-01 -5.66251397e-01
2.17398152e-01 7.86517441e-01 -3.04376874e-02 -5.37819982e-01
-1.11162746e+00 -4.76669252e-01 -8.46495748e-01 -2.26519912e-01
1.43973482e+00 -1.19044852e+00 -4.62469280e-01 1.10281336e+00
-1.06099904e+00 -6.69340670e-01 -2.65747935e-01 3.68368536e-01
-5.06162643e-01 -9.28319916e-02 -5.96023917e-01 -5.19547880e-01
-1.56372607e-01 -1.30316353e+00 1.67481434e+00 6.19731784e-01
4.11390424e-01 -7.81676531e-01 -2.09208012e-01 3.88362318e-01
1.70344323e-01 4.58147138e-01 7.12205291e-01 1.59936503e-01
-7.78385580e-01 4.20858972e-02 -7.94964969e-01 5.10529280e-01
1.75388858e-01 -3.16512972e-01 -1.63047194e+00 1.95435554e-01
3.52397896e-02 -4.31677312e-01 1.26833129e+00 5.03490508e-01
1.36056399e+00 3.37421656e-01 -1.65497079e-01 8.90425563e-01
1.59855580e+00 2.82070369e-01 4.87908095e-01 4.44369793e-01
1.09839785e+00 6.52940631e-01 7.17008948e-01 9.33291018e-02
8.12669337e-01 5.32985568e-01 7.92958915e-01 -1.13365017e-01
-6.60669923e-01 -1.83357939e-01 1.23819793e-02 4.97512668e-01
1.94115549e-01 4.15917858e-02 -8.13840628e-01 5.87103605e-01
-1.82389534e+00 -4.94532734e-01 -4.08554226e-02 2.04912210e+00
6.22096479e-01 2.29147404e-01 -4.23389584e-01 -1.04843900e-02
7.31148124e-01 3.03113222e-01 -7.58288383e-01 -3.65184188e-01
-5.53921461e-01 1.69552714e-01 6.87620580e-01 6.63810372e-01
-1.35343671e+00 1.16748297e+00 4.68627787e+00 7.25058734e-01
-1.55960047e+00 2.59002805e-01 7.99195290e-01 -5.83430082e-02
-1.98344305e-01 -1.85977712e-01 -7.17805386e-01 2.84365296e-01
6.51454329e-01 8.50429773e-01 2.62162715e-01 8.68855715e-01
-1.15030542e-01 -4.26704019e-01 -7.25776494e-01 1.16233623e+00
9.22501236e-02 -1.14766526e+00 -1.22243017e-01 -4.88759838e-02
1.12913799e+00 4.97748643e-01 8.24411958e-02 9.52835083e-02
2.91138887e-01 -8.03225756e-01 1.05306280e+00 7.65069485e-01
1.00680780e+00 -7.67762899e-01 7.37697482e-01 2.37386242e-01
-1.43891144e+00 -3.32659721e-01 -4.28456247e-01 -2.53668368e-01
1.54447556e-01 8.37618649e-01 -3.57799768e-01 6.17854595e-01
9.62759793e-01 9.20682430e-01 -6.47920728e-01 6.08237267e-01
-6.42618299e-01 6.39419556e-02 -3.82935137e-01 3.02461326e-01
6.33812606e-01 -2.56595016e-02 -4.59112115e-02 1.00715756e+00
2.85811216e-01 8.65645483e-02 8.53732601e-02 6.62582397e-01
-1.66740641e-01 -4.89963919e-01 -4.60306793e-01 4.21139121e-01
4.29691583e-01 1.34698069e+00 -6.55952811e-01 -5.36930978e-01
-3.78252685e-01 1.10931242e+00 4.49135721e-01 5.43399692e-01
-7.17025399e-01 -3.14934045e-01 6.84032083e-01 -4.80803221e-01
3.39678615e-01 -4.28244233e-01 -9.09329295e-01 -9.75273371e-01
-3.59905623e-02 1.51018389e-02 3.90933990e-01 -1.27903795e+00
-1.16931629e+00 7.73653507e-01 -2.57888675e-01 -8.81852567e-01
-4.18138318e-02 -7.96378195e-01 -2.82038450e-01 1.20601916e+00
-2.07639241e+00 -1.65827012e+00 -9.84297335e-01 7.62336433e-01
4.44444448e-01 4.19020832e-01 7.55477667e-01 3.12945366e-01
-3.09844553e-01 2.44296551e-01 -2.08501760e-02 -1.35660738e-01
6.95880294e-01 -1.30895829e+00 2.59039670e-01 7.00651884e-01
-5.44216670e-02 2.79329956e-01 6.45209998e-02 -5.46422720e-01
-9.67620969e-01 -1.65541232e+00 9.04821098e-01 -2.76576996e-01
2.30035648e-01 -2.76128322e-01 -6.64054871e-01 5.13512373e-01
-1.29511118e-01 3.25023264e-01 2.87531793e-01 -2.57360429e-01
-5.49997151e-01 -4.04193521e-01 -1.32906032e+00 9.22901407e-02
1.13733470e+00 -1.11948991e+00 -4.72135305e-01 1.20801456e-01
1.10758877e+00 -6.67229354e-01 -8.49179327e-01 6.79106355e-01
5.96675396e-01 -1.10108829e+00 1.10016024e+00 8.94569084e-02
5.10581017e-01 -5.69970548e-01 -8.99735451e-01 -9.69252884e-01
-1.14616781e-01 2.76949316e-01 1.20197169e-01 1.13475645e+00
9.31196436e-02 -5.50506294e-01 5.92047513e-01 4.42191601e-01
-7.10333109e-01 -8.41880620e-01 -9.88169372e-01 -3.01226825e-01
7.87357092e-02 -7.75442541e-01 7.78102398e-01 7.88638413e-01
-6.91444278e-01 2.06828058e-01 1.03452867e-02 3.03354859e-01
5.21647632e-01 6.73828304e-01 3.51732731e-01 -9.68227923e-01
3.40724811e-02 -2.97652543e-01 -4.60900187e-01 -1.28854322e+00
8.79434496e-02 -8.09400797e-01 2.33508974e-01 -1.81890893e+00
-1.93635866e-01 -6.85095310e-01 -5.99402845e-01 5.46905339e-01
-2.00824052e-01 7.26431131e-01 6.22242168e-02 7.70924985e-02
-5.63934505e-01 6.10185266e-01 1.24536753e+00 -3.43420506e-01
-3.95601802e-02 -4.07475740e-01 -4.41286385e-01 7.87449777e-01
7.24921048e-01 -1.99299846e-02 -4.46530938e-01 -1.03636527e+00
6.00937009e-03 -1.69008166e-01 7.18446553e-01 -1.21489775e+00
1.92029402e-01 -1.76001340e-01 8.87971520e-01 -9.27583754e-01
6.99414074e-01 -7.83521771e-01 -1.05313271e-01 7.83450454e-02
-2.75728464e-01 -2.92478561e-01 2.24448845e-01 5.43334961e-01
-4.09671396e-01 2.95664251e-01 7.89492190e-01 -2.46320263e-01
-1.37652898e+00 3.15338731e-01 1.52004108e-01 -3.87689710e-01
9.53630745e-01 -1.45342201e-01 -5.72092533e-01 9.28648114e-02
-8.95388722e-01 1.48307011e-01 5.96702397e-01 4.52337801e-01
6.67203605e-01 -1.48354411e+00 -4.53072861e-02 4.77540195e-01
3.28715205e-01 6.71575963e-01 7.73635685e-01 4.81373787e-01
-3.94744307e-01 4.35762137e-01 -3.15571189e-01 -9.93339956e-01
-8.34197700e-01 4.45153117e-01 4.08038974e-01 8.70023668e-02
-2.73198932e-01 1.51118374e+00 3.20459396e-01 -6.80173755e-01
2.37015441e-01 -7.74246573e-01 -1.26079693e-01 8.28693435e-02
4.69534934e-01 7.93759748e-02 3.94603997e-01 -9.72367227e-01
-3.52121770e-01 1.08353472e+00 3.78632873e-01 -1.32092521e-01
1.21054935e+00 -6.05250895e-01 -2.42993549e-01 6.52283669e-01
1.50268018e+00 -5.42329609e-01 -1.66682112e+00 -5.84257185e-01
-2.65603006e-01 -4.51886714e-01 3.21213812e-01 -1.11717212e+00
-1.31516671e+00 1.28085709e+00 1.02109098e+00 -1.75235435e-01
1.60744739e+00 1.63607866e-01 1.03476501e+00 2.04281196e-01
5.00285268e-01 -9.72922683e-01 -1.74196865e-02 7.08892643e-01
5.70100784e-01 -1.31674922e+00 -3.63218725e-01 -3.00744593e-01
-3.88145536e-01 1.04194736e+00 7.59537578e-01 -1.19851954e-01
6.83304012e-01 6.45515323e-02 3.24756265e-01 -2.58533776e-01
-3.69926006e-01 -4.59910125e-01 3.06625575e-01 6.03820980e-01
1.35440633e-01 2.78219104e-01 2.01206133e-01 5.52652776e-01
8.78992118e-03 3.70616913e-02 -5.09765483e-02 9.54558015e-01
-6.65629804e-01 -9.30385411e-01 -3.28194827e-01 1.58930406e-01
6.33506775e-02 7.59318843e-03 -3.96655917e-01 3.01026046e-01
9.41040874e-01 1.15078330e+00 2.88256258e-01 -8.18904757e-01
2.32390076e-01 1.84083030e-01 4.37608510e-01 -4.80072230e-01
-2.97767043e-01 1.95788085e-01 -4.28923294e-02 -9.15621877e-01
-8.15685928e-01 -3.54481459e-01 -1.57042849e+00 -5.28618544e-02
-1.72537770e-02 -3.90309483e-01 1.17207944e+00 1.16786206e+00
3.55039299e-01 7.94066668e-01 5.78108132e-01 -1.19985044e+00
1.77483425e-01 -9.85451519e-01 -4.40757006e-01 4.12237316e-01
5.08446097e-01 -9.57862854e-01 1.60616255e-04 6.19980395e-02] | [9.380117416381836, -1.1819326877593994] |
24f62278-5b57-4810-90ab-e7ca38628949 | context-empowered-visual-attention-prediction | 2210.16933 | null | https://arxiv.org/abs/2210.16933v1 | https://arxiv.org/pdf/2210.16933v1.pdf | Context-empowered Visual Attention Prediction in Pedestrian Scenarios | Effective and flexible allocation of visual attention is key for pedestrians who have to navigate to a desired goal under different conditions of urgency and safety preferences. While automatic modelling of pedestrian attention holds great promise to improve simulations of pedestrian behavior, current saliency prediction approaches mostly focus on generic free-viewing scenarios and do not reflect the specific challenges present in pedestrian attention prediction. In this paper, we present Context-SalNET, a novel encoder-decoder architecture that explicitly addresses three key challenges of visual attention prediction in pedestrians: First, Context-SalNET explicitly models the context factors urgency and safety preference in the latent space of the encoder-decoder model. Second, we propose the exponentially weighted mean squared error loss (ew-MSE) that is able to better cope with the fact that only a small part of the ground truth saliency maps consist of non-zero entries. Third, we explicitly model epistemic uncertainty to account for the fact that training data for pedestrian attention prediction is limited. To evaluate Context-SalNET, we recorded the first dataset of pedestrian visual attention in VR that includes explicit variation of the context factors urgency and safety preference. Context-SalNET achieves clear improvements over state-of-the-art saliency prediction approaches as well as over ablations. Our novel dataset will be made fully available and can serve as a valuable resource for further research on pedestrian attention prediction. | ['Christian Mueller', 'Ahmed Abouelazm', 'Nils Lipp', 'Lorena Hell', 'Philipp Mueller', 'Igor Vozniak'] | 2022-10-30 | null | null | null | null | ['saliency-prediction'] | ['computer-vision'] | [ 6.59960210e-02 9.79506746e-02 -2.55750954e-01 -3.66744041e-01
-5.30590773e-01 -7.90361315e-02 4.49269325e-01 3.13844323e-01
-5.19765496e-01 6.86579645e-01 6.42899275e-01 -2.32234314e-01
1.90623522e-01 -4.30040509e-01 -6.80018961e-01 -2.94755846e-01
-1.35063171e-01 1.73210561e-01 5.85687518e-01 -5.33752978e-01
1.47154927e-01 1.57637447e-01 -1.89720547e+00 1.36743769e-01
8.93232346e-01 8.58809650e-01 6.74508333e-01 7.03292489e-01
5.11084437e-01 7.20307052e-01 -1.37609541e-01 -4.73616570e-01
7.88842514e-02 -1.94824666e-01 -6.89755619e-01 -2.87088424e-01
7.17478812e-01 -2.25640923e-01 -6.40613079e-01 7.33374238e-01
7.95383036e-01 3.96065891e-01 4.41165596e-01 -1.48836207e+00
-8.95689130e-01 2.21459299e-01 -1.43843740e-01 7.92724371e-01
3.22009355e-01 5.29684961e-01 1.16098976e+00 -8.47391129e-01
4.17041034e-01 1.23090136e+00 4.98220921e-01 7.47638106e-01
-1.23019946e+00 -2.64605194e-01 6.37102902e-01 9.98482525e-01
-1.32597530e+00 -4.10822660e-01 5.49799740e-01 -6.09721005e-01
1.13842571e+00 5.72792053e-01 1.01962972e+00 1.37838757e+00
3.24172050e-01 1.16659629e+00 6.39554918e-01 1.97837744e-02
3.18721950e-01 2.33933106e-01 1.99364081e-01 2.62529910e-01
7.71546438e-02 4.68407243e-01 -6.49895012e-01 1.64637342e-01
5.84090054e-01 -1.25901893e-01 -9.00469646e-02 -7.61644542e-01
-1.13667166e+00 8.52077127e-01 8.86616707e-01 -2.05614954e-01
-4.88667637e-01 2.29024112e-01 3.43947768e-01 -1.72937915e-01
2.54442245e-01 2.94155180e-01 -3.98445785e-01 -1.51565194e-01
-8.76587927e-01 6.09308898e-01 1.73483014e-01 1.27562606e+00
5.24674356e-01 2.02855542e-01 -8.86919975e-01 7.15961099e-01
3.34158897e-01 6.21768773e-01 1.61292255e-01 -8.06685209e-01
3.73200059e-01 2.32785210e-01 3.59090209e-01 -1.05716419e+00
-5.76704741e-01 -6.21083498e-01 -5.08613288e-01 1.64564550e-01
1.95171267e-01 -8.39882717e-02 -8.11496079e-01 2.03720856e+00
-6.79329783e-02 3.12754363e-01 -2.01667175e-01 1.48268700e+00
1.08906412e+00 2.00278848e-01 7.05121279e-01 3.35688561e-01
1.44174278e+00 -1.05227673e+00 -4.11106974e-01 -9.29896832e-01
4.13573347e-02 -4.35538143e-01 1.15755427e+00 -1.99856564e-01
-1.27873051e+00 -6.60165012e-01 -8.81663799e-01 -2.76827276e-01
-4.10680443e-01 5.05597815e-02 7.59759009e-01 6.24028504e-01
-1.37957680e+00 1.93174034e-01 -5.83361745e-01 -5.35933435e-01
5.86258829e-01 3.52866173e-01 -6.53072447e-02 8.82300362e-02
-1.54058611e+00 1.41821468e+00 -6.48742616e-02 2.00159457e-02
-1.03275168e+00 -8.59251440e-01 -1.21500826e+00 4.47800905e-01
2.78674126e-01 -8.94978762e-01 1.26217747e+00 -8.15116465e-01
-9.47859168e-01 6.49571061e-01 -5.39786160e-01 -8.17544341e-01
4.35614049e-01 -3.91938090e-01 -3.68918538e-01 -1.83153987e-01
4.46176350e-01 1.29144704e+00 6.62364900e-01 -1.06436777e+00
-7.29948163e-01 1.19136564e-01 1.53719470e-01 2.59339780e-01
1.25134230e-01 -5.55640310e-02 -4.03588116e-01 -6.54194355e-01
-3.62480283e-01 -9.14280891e-01 -6.05803132e-01 -3.88925299e-02
-2.69748628e-01 1.44626200e-01 4.63029295e-01 -7.24230349e-01
1.21784556e+00 -1.90888619e+00 2.57954687e-01 -1.15491346e-01
1.59121558e-01 4.52733696e-01 -2.63007253e-01 1.00301854e-01
-1.66790783e-01 -5.08592017e-02 -4.82653044e-02 -5.51222622e-01
1.74802154e-01 2.56995466e-02 -3.76724571e-01 1.82812050e-01
4.77157801e-01 1.13386762e+00 -1.18427491e+00 -4.85493273e-01
6.29626393e-01 7.39288270e-01 -1.01776659e+00 -3.26311886e-02
-1.11358710e-01 4.91558403e-01 -2.35119030e-01 4.96313453e-01
5.57745218e-01 -1.03545047e-01 -4.61557329e-01 -8.71997252e-02
-3.62639248e-01 4.21228707e-01 -9.91073191e-01 1.49941981e+00
-3.29911560e-01 1.04361689e+00 -3.51248652e-01 -5.57803869e-01
4.88514215e-01 5.25263436e-02 2.21599936e-01 -1.08826470e+00
1.23856477e-01 -1.84644371e-01 5.64681813e-02 -2.70296901e-01
1.05550706e+00 3.21849078e-01 -1.28372744e-01 -1.92775115e-01
5.75414263e-02 3.12313855e-01 1.39982998e-01 3.06402922e-01
7.34596789e-01 -4.98372130e-02 2.07061425e-01 -5.31727314e-01
5.54753840e-01 -1.56412676e-01 5.18163025e-01 1.05097699e+00
-7.55641580e-01 9.54805434e-01 3.79830122e-01 -4.99833822e-01
-1.05400372e+00 -1.14068472e+00 -9.74928145e-04 1.37556219e+00
4.40642983e-01 -4.13723648e-01 -6.54380739e-01 -4.71950710e-01
-1.34046465e-01 1.28907931e+00 -9.19427097e-01 -4.88824278e-01
-4.66634661e-01 -4.30002242e-01 3.25668752e-02 8.27094674e-01
4.80183810e-01 -1.22789776e+00 -1.10416234e+00 1.40616912e-02
-4.73026335e-01 -9.54239786e-01 -7.52741933e-01 -3.89233828e-02
-8.83596018e-02 -9.18680727e-01 -8.97919118e-01 -6.14821017e-01
2.32427955e-01 4.64179873e-01 1.36782706e+00 1.48169268e-02
-2.74132699e-01 6.32144630e-01 -6.62702322e-02 -5.21186709e-01
2.50047326e-01 3.06105971e-01 -1.16341621e-01 -6.74647167e-02
5.73496997e-01 -1.80007219e-01 -1.03157020e+00 4.36089903e-01
-2.31254563e-01 3.10564607e-01 2.91389853e-01 6.95281267e-01
4.82761323e-01 -6.38578594e-01 5.98506927e-01 -7.11274743e-02
4.82615381e-01 -6.77120507e-01 -4.16432977e-01 4.98918742e-02
-2.34092489e-01 3.83800752e-02 7.74497166e-02 -3.12592715e-01
-1.03241801e+00 5.88738592e-03 -2.43138358e-01 -2.26094097e-01
-2.63363123e-01 2.02088833e-01 -1.76956862e-01 1.59636021e-01
5.72349608e-01 1.02772653e-01 -5.00367105e-01 -1.88931748e-01
3.16262454e-01 2.48320594e-01 4.50333476e-01 -2.49129027e-01
2.50690132e-01 1.20985277e-01 7.40242302e-02 -8.24090958e-01
-7.47913718e-01 -5.18395960e-01 -4.77398366e-01 -3.55385005e-01
1.09092784e+00 -1.06659842e+00 -8.74266148e-01 1.38301119e-01
-1.02649057e+00 -5.98317921e-01 -3.03053319e-01 3.33575547e-01
-8.38229597e-01 1.55652687e-01 -2.18638688e-01 -8.38643610e-01
-7.07158595e-02 -1.40610361e+00 1.06807423e+00 3.14772785e-01
-4.80881393e-01 -9.30952847e-01 -2.94577897e-01 1.12517178e-01
8.21139574e-01 -6.23372151e-03 5.26627123e-01 -2.28869811e-01
-7.35368371e-01 3.47769372e-02 -4.41306591e-01 -1.23054504e-01
-2.27096632e-01 -4.61264759e-01 -8.39357078e-01 -2.26875290e-01
-4.54842597e-01 -6.07419088e-02 1.17981422e+00 8.73808861e-01
9.26260650e-01 -9.65969265e-02 -3.88838977e-01 2.45245889e-01
9.88988101e-01 -1.15113877e-01 8.69001389e-01 4.78992075e-01
6.51395142e-01 5.85431993e-01 7.46223927e-01 4.44646537e-01
1.13542330e+00 1.16072464e+00 6.69710457e-01 -1.12730928e-01
-3.85835886e-01 -3.66921633e-01 1.33647844e-01 -2.02343181e-01
-1.36714518e-01 -2.94140697e-01 -9.13375616e-01 1.01963210e+00
-2.10500360e+00 -1.27328205e+00 -2.23877534e-01 2.31617641e+00
3.98635089e-01 2.54424721e-01 4.23695773e-01 -1.21509679e-01
6.69246376e-01 1.69498995e-01 -5.80416501e-01 -4.98989046e-01
-1.12189628e-01 -1.31840259e-01 7.28109717e-01 8.20678115e-01
-1.40440881e+00 1.16644120e+00 6.75348473e+00 4.72181082e-01
-9.78644013e-01 1.52440801e-01 5.48297346e-01 -3.03705126e-01
-2.39246368e-01 -8.71354789e-02 -1.02686083e+00 6.74170315e-01
8.70085001e-01 -4.39094380e-02 3.08395207e-01 1.01124430e+00
2.55623519e-01 -2.51680911e-01 -1.24303734e+00 8.94721270e-01
1.21863820e-01 -1.03197706e+00 -6.01079091e-02 -1.28481120e-01
4.23777103e-01 2.09923699e-01 6.64783478e-01 5.33424020e-01
2.50344872e-01 -1.10257697e+00 1.23565447e+00 8.18926871e-01
4.49194640e-01 -6.84845448e-01 6.69466078e-01 1.51122242e-01
-1.14952207e+00 -2.84082055e-01 -4.67969805e-01 -2.69691855e-01
5.69611251e-01 2.74346769e-02 -6.56294584e-01 -3.99613641e-02
7.85656452e-01 7.65865028e-01 -1.04412615e+00 1.48291385e+00
-1.08301908e-01 3.41987997e-01 -6.69366494e-02 -1.40482724e-01
3.39835733e-01 4.27290052e-01 8.18558276e-01 1.39936721e+00
3.04317296e-01 1.09347235e-02 4.14759628e-02 9.27708507e-01
6.05119228e-01 -9.69363078e-02 -3.49368453e-01 5.20764768e-01
2.05471307e-01 7.37632334e-01 -3.89819920e-01 -2.20801443e-01
-5.39077401e-01 8.77247155e-01 3.91112000e-01 5.77254593e-01
-1.10776460e+00 1.50259212e-01 1.16841781e+00 3.98038417e-01
7.52766252e-01 -1.29771322e-01 -5.66244245e-01 -1.04154849e+00
-1.78699747e-01 -4.39782530e-01 2.70709783e-01 -1.05392742e+00
-9.52626526e-01 5.63776195e-01 2.02689260e-01 -1.06224608e+00
-1.80630207e-01 -4.18443441e-01 -6.28584921e-01 1.21096563e+00
-1.67346120e+00 -1.28743553e+00 -4.08457726e-01 5.41610181e-01
8.85627985e-01 -1.03412002e-01 6.65586174e-01 2.80672073e-01
-4.32803988e-01 7.06073940e-01 -6.30758464e-01 -3.09221089e-01
5.80921113e-01 -1.22730780e+00 8.11931252e-01 9.19921815e-01
-3.40095341e-01 4.72766221e-01 1.25260532e+00 -6.50822401e-01
-7.75863111e-01 -1.10446012e+00 1.26510537e+00 -8.19716752e-01
2.28701621e-01 -4.15560335e-01 -6.98514998e-01 8.58751476e-01
2.01839224e-01 5.77635057e-02 5.77939987e-01 2.34083071e-01
-2.33292375e-02 1.87100694e-01 -9.49660063e-01 1.15736926e+00
1.14460266e+00 -3.36224169e-01 -6.15323842e-01 2.12645549e-02
6.29521847e-01 -5.50780833e-01 -1.82699338e-01 2.48662114e-01
4.20762241e-01 -1.03422463e+00 1.48920667e+00 -7.59255230e-01
3.88915300e-01 -1.63151756e-01 -4.02627349e-01 -1.37874579e+00
-9.56090927e-01 -1.95046276e-01 -2.10070565e-01 8.22249234e-01
4.05212343e-01 -2.57927865e-01 7.34763682e-01 1.15265048e+00
-3.31548542e-01 -5.68619668e-01 -1.08624136e+00 -3.57892901e-01
-5.31525612e-02 -4.95060623e-01 6.26192868e-01 3.98669302e-01
6.76306561e-02 3.36486280e-01 -7.48566687e-01 1.62508965e-01
6.21272206e-01 -2.73693204e-01 5.04376829e-01 -1.09002733e+00
1.82374148e-03 -6.60391271e-01 -7.11427331e-01 -1.21445847e+00
6.28893748e-02 -6.47103906e-01 8.33020434e-02 -1.68055415e+00
2.13002801e-01 -3.71388137e-01 -4.92329240e-01 3.56614500e-01
-4.59677160e-01 1.45532280e-01 4.29370493e-01 -4.15378213e-01
-9.82196987e-01 8.41304481e-01 1.00143647e+00 -2.46673554e-01
-2.45325699e-01 2.09907517e-01 -1.01380479e+00 6.32800341e-01
8.03778470e-01 -2.97155734e-02 -5.65695345e-01 -4.61837202e-01
9.57374349e-02 -5.94397634e-02 9.98318017e-01 -1.23870158e+00
2.93811232e-01 -4.09962237e-01 4.37923998e-01 -7.11189032e-01
6.47631466e-01 -7.21858442e-01 -1.53407767e-01 1.75167575e-01
-5.82929969e-01 2.70948168e-02 5.06064475e-01 5.60110092e-01
1.67958364e-01 6.33184910e-02 6.64860308e-01 1.92338862e-02
-1.20207727e+00 3.24525356e-01 -7.12874055e-01 2.06403181e-01
9.39321995e-01 -1.66575372e-01 -1.65673926e-01 -6.36061490e-01
-1.04198885e+00 5.64849317e-01 3.38326782e-01 9.43625391e-01
8.17995608e-01 -1.56982839e+00 -7.19498694e-01 2.46519506e-01
1.51210710e-01 -4.30381685e-01 6.54761672e-01 7.17587352e-01
6.58962801e-02 8.91369224e-01 -4.84733760e-01 -7.01562524e-01
-1.16479957e+00 9.61053371e-01 3.08881283e-01 6.81190193e-02
-4.70652968e-01 8.85680795e-01 5.31440198e-01 -2.31548864e-02
2.68825114e-01 -2.37915620e-01 -6.30300760e-01 -1.83737293e-01
6.08335435e-01 4.03008014e-01 -1.97530791e-01 -1.27873218e+00
-5.85919678e-01 2.28735015e-01 1.14147253e-01 -2.91241314e-02
9.55547035e-01 -4.88939643e-01 5.52049637e-01 1.96154356e-01
6.68435156e-01 -2.95716792e-01 -1.82803893e+00 -4.81793731e-02
-2.75611758e-01 -7.15463698e-01 1.66494340e-01 -9.37016606e-01
-8.82156312e-01 1.11832500e+00 9.25359070e-01 -1.88646212e-01
8.54443669e-01 8.29686970e-02 5.86079001e-01 1.44514004e-02
2.53615737e-01 -9.70095992e-01 -7.90325105e-02 6.92566574e-01
1.14584732e+00 -1.49257135e+00 -2.67811954e-01 -3.77849996e-01
-1.20034254e+00 3.38571787e-01 1.01083934e+00 4.13612789e-03
6.32502019e-01 -2.68633962e-01 -1.80244863e-01 5.95504045e-02
-8.06203187e-01 -7.19389856e-01 6.49518847e-01 1.01611125e+00
2.67170519e-01 1.20778635e-01 -1.35986961e-03 9.90987241e-01
-3.42678279e-01 -1.01185158e-01 2.64040858e-01 5.21090984e-01
-5.29046237e-01 -7.16501296e-01 -7.78722316e-02 2.58062690e-01
-7.00630769e-02 -4.36621070e-01 1.14566740e-02 3.85986149e-01
2.32162371e-01 9.70315456e-01 1.06623620e-01 -3.53442162e-01
6.58932865e-01 -1.41702399e-01 2.51590341e-01 -3.79170895e-01
-4.28092271e-01 -3.77291322e-01 8.07977021e-02 -6.72939062e-01
-1.14448868e-01 -9.06265378e-01 -8.24171305e-01 -2.35801026e-01
9.27555338e-02 -1.63613006e-01 2.69852579e-01 7.88272500e-01
5.94303846e-01 8.17549169e-01 -4.35126610e-02 -1.38903332e+00
-7.07049901e-03 -6.41341507e-01 -1.30583167e-01 4.12874043e-01
6.80646598e-01 -1.21687913e+00 -6.73679560e-02 -1.06601574e-01] | [6.292829513549805, 0.6111217141151428] |
33dad568-1ede-4e00-b2d0-c2d084d95e73 | adaptive-message-quantization-and | 2306.01381 | null | https://arxiv.org/abs/2306.01381v1 | https://arxiv.org/pdf/2306.01381v1.pdf | Adaptive Message Quantization and Parallelization for Distributed Full-graph GNN Training | Distributed full-graph training of Graph Neural Networks (GNNs) over large graphs is bandwidth-demanding and time-consuming. Frequent exchanges of node features, embeddings and embedding gradients (all referred to as messages) across devices bring significant communication overhead for nodes with remote neighbors on other devices (marginal nodes) and unnecessary waiting time for nodes without remote neighbors (central nodes) in the training graph. This paper proposes an efficient GNN training system, AdaQP, to expedite distributed full-graph GNN training. We stochastically quantize messages transferred across devices to lower-precision integers for communication traffic reduction and advocate communication-computation parallelization between marginal nodes and central nodes. We provide theoretical analysis to prove fast training convergence (at the rate of O(T^{-1}) with T being the total number of training epochs) and design an adaptive quantization bit-width assignment scheme for each message based on the analysis, targeting a good trade-off between training convergence and efficiency. Extensive experiments on mainstream graph datasets show that AdaQP substantially improves distributed full-graph training's throughput (up to 3.01 X) with negligible accuracy drop (at most 0.30%) or even accuracy improvement (up to 0.19%) in most cases, showing significant advantages over the state-of-the-art works. | ['Chuan Wu', 'Juntao Zhao', 'Borui Wan'] | 2023-06-02 | null | null | null | null | ['quantization'] | ['methodology'] | [-2.73501929e-02 2.31435239e-01 -3.71115029e-01 -3.55669379e-01
-5.07415175e-01 -4.09802079e-01 1.73560940e-02 4.65635985e-01
-6.38519645e-01 5.22236645e-01 -6.52014971e-01 -8.79699707e-01
-2.39780471e-01 -1.01173913e+00 -8.98385167e-01 -7.50662684e-01
-6.22974098e-01 3.99212003e-01 4.33570385e-01 1.60243288e-01
-9.93245915e-02 6.17845178e-01 -1.11335170e+00 -3.65202904e-01
6.44186735e-01 1.18028212e+00 1.49167180e-02 9.52730536e-01
-1.53589442e-01 8.17454636e-01 -7.78508961e-01 -7.64606595e-01
3.17114383e-01 -5.82223274e-02 -5.73363781e-01 -2.11779207e-01
5.71943343e-01 -6.50585890e-01 -7.66202450e-01 1.31026435e+00
7.08020449e-01 -7.71498904e-02 7.90480450e-02 -1.43146384e+00
-7.04252660e-01 9.68905628e-01 -6.40723109e-01 2.78562397e-01
-4.98126984e-01 -1.31736979e-01 9.05651331e-01 -3.15933943e-01
3.90615076e-01 8.13208342e-01 9.51950133e-01 6.07077599e-01
-1.06586742e+00 -9.91919219e-01 1.00359984e-01 7.55562037e-02
-1.58123744e+00 -3.94495696e-01 7.47539520e-01 1.55060574e-01
1.14530659e+00 3.82584691e-01 7.57579267e-01 5.19895315e-01
2.15261176e-01 2.98548460e-01 2.64153659e-01 -3.07057321e-01
5.15031636e-01 -6.77948073e-03 1.89531818e-01 1.03896511e+00
8.32575142e-01 -3.94050151e-01 -5.37557065e-01 -1.70494586e-01
9.55540299e-01 9.80725661e-02 -2.16591984e-01 -2.44887903e-01
-7.26630092e-01 8.52391660e-01 9.41479802e-01 1.12067945e-01
-5.38240552e-01 1.03242409e+00 8.09101641e-01 5.42667150e-01
5.81167638e-01 -3.29313725e-01 -4.79703039e-01 -2.05808029e-01
-6.65092170e-01 -2.52083898e-01 8.96033883e-01 1.38021111e+00
9.65970755e-01 3.50716382e-01 3.34733427e-01 6.55009210e-01
1.92795530e-01 7.49892116e-01 3.18581820e-01 -5.93729615e-01
8.29974413e-01 6.34323955e-01 -4.86925364e-01 -1.32960737e+00
-4.62584138e-01 -5.99248171e-01 -1.41203773e+00 -4.26430911e-01
1.32865265e-01 -6.61536574e-01 -6.28910542e-01 1.67446935e+00
6.18537188e-01 4.18852091e-01 -1.09298110e-01 6.84561193e-01
6.94822848e-01 7.73475111e-01 8.58176500e-02 -1.22993752e-01
1.15559900e+00 -8.78731251e-01 -4.66084480e-01 -1.36549965e-01
1.23947835e+00 -3.20083559e-01 7.15748549e-01 -5.04174232e-02
-8.79178405e-01 -2.38166720e-01 -1.04437459e+00 -9.99822915e-02
-3.15919340e-01 2.18636274e-01 1.13843799e+00 1.08039510e+00
-1.67572188e+00 7.96668828e-01 -1.10590291e+00 -2.17029423e-01
6.00871801e-01 1.11684668e+00 -1.37276530e-01 9.53983050e-03
-7.53672957e-01 -9.89002809e-02 2.25544527e-01 2.78268635e-01
-5.41380107e-01 -8.60725820e-01 -6.95955575e-01 1.28439859e-01
1.26213267e-01 -6.25146449e-01 9.71548557e-01 -6.79570735e-01
-1.40813625e+00 3.53608727e-01 9.98624489e-02 -7.21772254e-01
1.07682139e-01 1.30326256e-01 -4.99462485e-01 2.50725955e-01
-2.55653083e-01 4.21203494e-01 5.53360224e-01 -4.90259618e-01
-5.31690121e-01 -5.80971897e-01 -2.05685608e-02 1.24680335e-02
-1.16024363e+00 -2.59206265e-01 -8.56183767e-01 -2.91527808e-01
6.57416135e-02 -1.01697528e+00 -3.62572074e-01 2.20998496e-01
-3.44190896e-01 -3.14468563e-01 9.92804289e-01 -2.41741076e-01
1.29038286e+00 -2.11815190e+00 -6.16952777e-01 4.96096820e-01
8.73938799e-01 4.60911393e-01 -5.11741564e-02 2.37337679e-01
5.46141565e-01 3.68588232e-02 3.74763548e-01 -4.66299295e-01
5.29302983e-03 4.12313730e-01 -1.39358670e-01 7.65615225e-01
-2.60169595e-01 9.34188306e-01 -7.82789052e-01 -4.14092124e-01
1.34376615e-01 8.71319115e-01 -6.65511608e-01 -3.04911882e-01
1.06319852e-01 -4.04451430e-01 -6.59390330e-01 3.36658329e-01
8.61195982e-01 -9.67392147e-01 6.46867394e-01 -2.81376034e-01
3.91799241e-01 4.26556677e-01 -1.13628423e+00 1.33381343e+00
-5.53647637e-01 7.32158005e-01 3.46194923e-01 -8.90974998e-01
8.73778522e-01 3.98753695e-02 4.66150910e-01 -6.69566691e-01
4.09592897e-01 2.62827754e-01 -4.77291286e-01 6.60557598e-02
4.99319196e-01 3.28182280e-01 1.36332974e-01 3.36314887e-01
7.03831241e-02 5.80389559e-01 -6.60096779e-02 3.43600422e-01
1.49473107e+00 -9.99218345e-01 -2.11682513e-01 -2.73330748e-01
-9.47088376e-03 -1.32975802e-01 2.79915124e-01 7.42395699e-01
-2.79797018e-01 -3.23527306e-01 5.49131036e-01 -5.51342368e-01
-8.70882630e-01 -6.89850032e-01 2.89776802e-01 1.25989521e+00
2.41321236e-01 -8.56657565e-01 -8.55115712e-01 -5.44030070e-01
8.06361735e-02 3.49272378e-02 -3.39866340e-01 -1.54306918e-01
-5.94868541e-01 -6.52946830e-01 7.35944808e-01 6.56745136e-01
6.48566782e-01 -5.79826951e-01 -5.63787639e-01 2.98145562e-01
6.46563351e-01 -1.29252577e+00 -5.93713105e-01 4.50377375e-01
-1.23585725e+00 -6.83784902e-01 -4.90353942e-01 -1.09537458e+00
1.11076140e+00 4.36770082e-01 7.73161471e-01 2.42891565e-01
-1.81753531e-01 2.38528714e-01 -1.57410383e-01 -9.45805833e-02
-4.56495509e-02 2.73506492e-01 -2.21139956e-02 -1.58241719e-01
3.33187789e-01 -7.73031831e-01 -7.26829946e-01 3.88515927e-02
-6.39669776e-01 -5.15389889e-02 5.82496047e-01 6.43832266e-01
5.92916012e-01 3.66618067e-01 2.20227078e-01 -1.14437389e+00
5.82073510e-01 -2.95225561e-01 -1.15376842e+00 4.66888808e-02
-8.26100886e-01 4.53860015e-02 1.07730794e+00 -6.50292218e-01
-1.12415709e-01 -4.55264095e-03 -9.24133789e-03 -7.09195077e-01
4.92526650e-01 2.75140911e-01 6.60942793e-02 -7.95551956e-01
5.27283669e-01 3.68903838e-02 -1.11269034e-01 -1.45632550e-01
4.43263978e-01 6.34369314e-01 3.85859847e-01 -3.21844995e-01
6.15945041e-01 2.70674556e-01 2.56852835e-01 -9.48812962e-01
-8.82392526e-02 -1.90858543e-01 -7.27228895e-02 -8.30576643e-02
2.43688986e-01 -1.16785920e+00 -1.28478730e+00 2.22782075e-01
-9.39125955e-01 -7.86785424e-01 -2.06880108e-01 4.70444977e-01
2.07872584e-01 1.79802194e-01 -1.04203796e+00 -7.58662224e-01
-9.04816270e-01 -9.12442505e-01 9.50599670e-01 2.86458522e-01
4.33803499e-01 -1.27294052e+00 -5.65941453e-01 -9.74408165e-02
8.02529693e-01 6.67466372e-02 7.86396563e-01 -5.89328587e-01
-9.11869466e-01 -3.93597573e-01 -7.23527849e-01 2.36014977e-01
-1.08426116e-01 -1.94891959e-01 -5.82864285e-01 -6.56475842e-01
-3.32001001e-01 -1.83952674e-01 4.35284913e-01 3.45862567e-01
1.30294752e+00 -7.55436659e-01 -5.71755350e-01 8.39736104e-01
1.76636553e+00 -9.63880643e-02 1.77570835e-01 -3.26039135e-01
1.15015638e+00 -2.20803484e-01 -1.06753156e-01 6.44405067e-01
4.12857801e-01 2.52406508e-01 5.30259490e-01 -2.68136337e-02
-2.65326083e-01 -3.19361061e-01 2.70625591e-01 1.30140054e+00
3.58944237e-01 -5.81793010e-01 -7.35644400e-01 5.61224043e-01
-1.48568892e+00 -3.83908659e-01 -1.34397805e-01 2.20959306e+00
6.26631677e-01 2.76443988e-01 -2.15967163e-03 1.60924807e-01
8.91007066e-01 2.58392274e-01 -7.35337317e-01 -4.50504452e-01
3.91213953e-01 3.65219444e-01 1.36775517e+00 4.34346020e-01
-6.71797693e-01 7.94771314e-01 5.61523724e+00 1.16056943e+00
-1.46725321e+00 2.67241180e-01 7.18880773e-01 -2.01514870e-01
-2.06462801e-01 -2.02667981e-01 -1.10281706e+00 4.50125247e-01
1.60968268e+00 -2.15331540e-01 7.22014904e-01 1.31378448e+00
-3.36879641e-01 4.00741696e-01 -9.14777100e-01 1.35504901e+00
-2.01879248e-01 -1.64027441e+00 9.62338895e-02 4.70193356e-01
7.28157043e-01 6.67655289e-01 -1.55269772e-01 2.27722317e-01
4.98563617e-01 -5.56268334e-01 3.17183971e-01 -4.76171911e-01
1.18101776e+00 -1.04320133e+00 6.64601147e-01 2.07767904e-01
-1.60146880e+00 -7.10205585e-02 -8.76283228e-01 3.54165323e-02
-1.40439346e-01 1.15103710e+00 -8.93094599e-01 3.97933304e-01
5.27990520e-01 4.13874984e-01 -4.03746307e-01 6.00835979e-01
2.86490858e-01 7.48775363e-01 -9.52887654e-01 -7.49492109e-01
3.23987275e-01 -9.25131738e-02 -1.07378501e-03 1.01398993e+00
3.35057527e-01 5.60391136e-02 -3.37451026e-02 4.29126829e-01
-8.99581313e-01 2.52043754e-01 -4.72534657e-01 -8.55376497e-02
1.09127915e+00 1.22036457e+00 -9.26280379e-01 -4.31571752e-01
-3.36338699e-01 1.06288350e+00 6.47461176e-01 1.53994232e-01
-8.83777678e-01 -1.01378882e+00 4.46925521e-01 4.10841517e-02
5.35273492e-01 -3.47845316e-01 -8.17104951e-02 -5.05173922e-01
3.85072201e-01 -2.84829050e-01 2.16970488e-01 -1.73724711e-01
-1.08310699e+00 6.50910914e-01 -4.61691141e-01 -7.44633973e-01
6.28941506e-02 -5.20359218e-01 -3.61246347e-01 5.09636164e-01
-1.37495148e+00 -8.07366550e-01 -4.30951089e-01 8.17685783e-01
-1.28417939e-01 9.20615792e-02 6.99952602e-01 8.91361058e-01
-6.76100314e-01 1.29868245e+00 3.56579334e-01 3.79695296e-01
-2.79468596e-02 -9.21718776e-01 8.73316169e-01 6.70774221e-01
2.64511853e-01 6.70643508e-01 2.66344696e-01 -5.22773147e-01
-2.17737818e+00 -1.43574190e+00 9.55609620e-01 3.61814648e-01
8.69662166e-01 -6.53562307e-01 -6.03129804e-01 5.98805249e-01
-2.21901163e-01 7.89615989e-01 5.91287076e-01 2.12204859e-01
-3.12859774e-01 -7.54271328e-01 -1.09104097e+00 6.88660383e-01
1.20972776e+00 -6.56278491e-01 7.61309326e-01 7.28541613e-01
9.61949885e-01 -6.40320599e-01 -1.01670122e+00 2.92046368e-02
2.67066538e-01 -6.14411056e-01 6.38143301e-01 -1.61117017e-01
-4.36338007e-01 -1.11268032e-02 -1.16961397e-01 -8.69422674e-01
-3.89363132e-02 -1.05869126e+00 -4.10175353e-01 1.02819097e+00
4.11393940e-01 -9.54449415e-01 1.48977351e+00 5.16077340e-01
5.82191534e-02 -9.21516180e-01 -1.18230987e+00 -7.75532782e-01
-4.01121408e-01 -6.45606995e-01 6.94856763e-01 8.09153855e-01
-1.08936012e-01 5.94967902e-01 -1.04244426e-01 6.00184083e-01
5.49150169e-01 -1.91601321e-01 8.91488671e-01 -1.10887241e+00
-2.36484289e-01 -2.18898609e-01 -8.63583922e-01 -1.64465368e+00
-2.09674854e-02 -9.82514203e-01 -3.98225069e-01 -1.39788437e+00
-3.04326266e-01 -9.38018143e-01 -2.03996688e-01 8.07700157e-01
2.73256004e-01 2.42234185e-01 -7.10314810e-02 3.21689844e-02
-1.00286448e+00 3.21694523e-01 9.68845844e-01 1.11153528e-01
-2.55807906e-01 -1.71987370e-01 -6.12306833e-01 2.74002194e-01
8.38075757e-01 -5.82774997e-01 -8.73478234e-01 -8.94959509e-01
3.29608619e-01 2.30702251e-01 3.24303955e-01 -1.34591234e+00
6.56280518e-01 4.07486111e-01 1.28445417e-01 -5.03812194e-01
2.51625985e-01 -1.05086613e+00 1.79390863e-01 6.12561524e-01
-1.77617893e-01 4.55004245e-01 2.43984461e-01 8.86585534e-01
2.25754589e-01 3.39122623e-01 4.99212712e-01 5.28241575e-01
-3.48751038e-01 7.24078119e-01 9.18502957e-02 -2.33417787e-02
9.86647248e-01 -1.71243045e-02 -6.09109640e-01 -3.24674249e-01
-3.19548547e-01 1.75151587e-01 1.31948769e-01 -8.87361094e-02
5.59176683e-01 -1.25348723e+00 -2.09215745e-01 4.46999639e-01
-2.69823402e-01 1.62045240e-01 4.16075945e-01 7.71430969e-01
-8.76328051e-01 5.69657207e-01 3.12672734e-01 -5.73645711e-01
-1.19688606e+00 3.25848520e-01 1.68480873e-01 -2.99683422e-01
-9.00025249e-01 1.33133459e+00 -3.55011851e-01 -1.39866129e-01
7.09254861e-01 -5.57185411e-01 6.28130138e-01 -5.51391244e-01
4.63791907e-01 5.32223523e-01 4.00277793e-01 -6.76981658e-02
-4.18780118e-01 3.12795997e-01 -2.29049891e-01 4.35183734e-01
1.16415966e+00 -1.04675956e-01 -1.37279272e-01 2.08654720e-02
1.87652624e+00 -3.00787389e-01 -1.08961821e+00 -3.78136992e-01
-4.29694772e-01 -2.16887355e-01 5.45546591e-01 -1.11951649e-01
-1.65650892e+00 6.49136961e-01 6.74824178e-01 4.64697689e-01
1.14384484e+00 -5.89781962e-02 1.25540721e+00 7.42776573e-01
7.00728595e-01 -9.74474132e-01 -1.84697881e-01 2.58517921e-01
-9.54325050e-02 -6.59578145e-01 3.70512493e-02 -5.61079979e-01
-1.57520220e-01 9.49796200e-01 5.50151527e-01 -1.29474878e-01
1.01212168e+00 4.62225974e-01 -2.60326385e-01 -4.69302654e-01
-8.64596367e-01 4.15615082e-01 -2.12975070e-01 5.59932768e-01
5.12769483e-02 2.49466985e-01 4.29206565e-02 3.08315396e-01
-2.93867022e-01 -4.89576869e-02 1.05707511e-01 8.42751443e-01
-5.50083458e-01 -8.33598554e-01 3.12998891e-01 7.80034065e-01
-3.40159714e-01 -2.35196814e-01 1.09768108e-01 7.84839332e-01
-1.76300630e-01 7.89346278e-01 4.57608104e-01 -6.90769672e-01
-4.07090783e-02 -4.13946599e-01 3.18991065e-01 -3.55972797e-01
-4.71206456e-01 -5.31144766e-03 1.07723475e-01 -7.44671106e-01
1.06675131e-02 1.09597087e-01 -1.54952836e+00 -1.14307463e+00
-6.18755043e-01 2.86927253e-01 1.10742140e+00 2.86849976e-01
8.07640672e-01 5.61684012e-01 6.78742409e-01 -6.56512141e-01
-6.04115546e-01 -4.40784216e-01 -9.42394912e-01 -1.78084955e-01
3.39265853e-01 -3.55395824e-02 -4.91144240e-01 -3.09393495e-01] | [6.981847763061523, 5.699334621429443] |
23e8da33-6aa2-4dae-aeaf-9733b89ca7df | towards-better-validity-dispersion-based | 1906.01308 | null | https://arxiv.org/abs/1906.01308v1 | https://arxiv.org/pdf/1906.01308v1.pdf | Towards better Validity: Dispersion based Clustering for Unsupervised Person Re-identification | Person re-identification aims to establish the correct identity correspondences of a person moving through a non-overlapping multi-camera installation. Recent advances based on deep learning models for this task mainly focus on supervised learning scenarios where accurate annotations are assumed to be available for each setup. Annotating large scale datasets for person re-identification is demanding and burdensome, which renders the deployment of such supervised approaches to real-world applications infeasible. Therefore, it is necessary to train models without explicit supervision in an autonomous manner. In this paper, we propose an elegant and practical clustering approach for unsupervised person re-identification based on the cluster validity consideration. Concretely, we explore a fundamental concept in statistics, namely \emph{dispersion}, to achieve a robust clustering criterion. Dispersion reflects the compactness of a cluster when employed at the intra-cluster level and reveals the separation when measured at the inter-cluster level. With this insight, we design a novel Dispersion-based Clustering (DBC) approach which can discover the underlying patterns in data. This approach considers a wider context of sample-level pairwise relationships to achieve a robust cluster affinity assessment which handles the complications may arise due to prevalent imbalanced data distributions. Additionally, our solution can automatically prioritize standalone data points and prevents inferior clustering. Our extensive experimental analysis on image and video re-identification benchmarks demonstrate that our method outperforms the state-of-the-art unsupervised methods by a significant margin. Code is available at https://github.com/gddingcs/Dispersion-based-Clustering.git. | ['Guodong Ding', 'Salman Khan', 'Fatih Porikli', 'Zhenmin Tang', 'Jian Zhang'] | 2019-06-04 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [-1.69910371e-01 -4.42804933e-01 5.48008643e-02 -5.27604640e-01
-4.68566686e-01 -5.26917696e-01 4.30470586e-01 3.61379802e-01
-5.19739330e-01 3.81719112e-01 7.76961669e-02 4.25733067e-02
-4.62609857e-01 -6.09418333e-01 -4.33090538e-01 -9.39444721e-01
-2.88615674e-02 7.13946819e-01 -3.03327709e-01 2.13583007e-01
1.36936218e-01 3.05815846e-01 -1.66567230e+00 4.72139195e-03
9.71268535e-01 5.13487101e-01 -6.17901944e-02 3.73024762e-01
2.34168261e-01 1.69762924e-01 -6.54751897e-01 -7.09193110e-01
3.95734698e-01 -2.72790700e-01 -7.45509386e-01 3.44466925e-01
4.38708007e-01 -2.87127942e-01 -1.52890787e-01 1.24523485e+00
6.92156315e-01 1.71709836e-01 5.83710372e-01 -1.44647920e+00
-5.03865361e-01 4.67362285e-01 -9.30670440e-01 2.36521766e-01
2.30675444e-01 7.67385587e-02 8.08636069e-01 -6.95927203e-01
4.62191291e-02 9.38252270e-01 9.66302872e-01 4.77180213e-01
-1.32011271e+00 -7.07052052e-01 1.93744376e-01 3.96860093e-01
-1.91253722e+00 -4.59497690e-01 7.32023358e-01 -5.58025658e-01
3.47653776e-01 4.47576672e-01 5.52022278e-01 1.04215837e+00
-5.71398973e-01 5.10304868e-01 8.98816109e-01 -3.36132556e-01
1.77038684e-01 8.70873556e-02 3.35682690e-01 3.19456905e-01
4.91699129e-01 -1.96733460e-01 -3.22269052e-01 -1.07963100e-01
5.96514404e-01 2.84162939e-01 -1.22448988e-01 -3.97373974e-01
-1.20887160e+00 5.20183384e-01 3.05650413e-01 3.16055089e-01
-1.69392928e-01 -8.15024972e-02 4.73923802e-01 7.40952641e-02
4.38267112e-01 8.31237808e-02 -7.62036890e-02 -1.66005030e-01
-1.19267476e+00 2.15348750e-01 4.45339710e-01 9.05475318e-01
6.61586583e-01 -4.75270867e-01 3.76909343e-03 9.02281344e-01
1.38482690e-01 1.64654806e-01 4.92673784e-01 -9.26679850e-01
4.57340598e-01 7.27247894e-01 1.15320936e-01 -1.27536619e+00
-4.78175461e-01 -6.29771650e-01 -1.34739912e+00 -1.02008604e-01
6.72386706e-01 -6.00884892e-02 -4.00747001e-01 1.72677100e+00
4.55688536e-01 5.27073383e-01 -1.09076113e-01 9.93416548e-01
5.89375794e-01 1.19613439e-01 -4.52191159e-02 -2.47499496e-01
1.24664080e+00 -6.44965947e-01 -5.35839856e-01 1.20657369e-01
5.76817274e-01 -4.54210639e-01 9.35037196e-01 2.55440891e-01
-6.97853386e-01 -7.60397673e-01 -9.48420465e-01 2.79980481e-01
-2.83700019e-01 4.63615388e-01 3.25270116e-01 8.90608549e-01
-1.02806246e+00 4.76874262e-01 -7.16736495e-01 -6.35970712e-01
3.73771489e-01 6.79530144e-01 -5.41978419e-01 -3.21872942e-02
-8.55982780e-01 2.77168274e-01 4.19043839e-01 4.35764194e-01
-3.39546263e-01 -6.31779432e-01 -5.75707793e-01 -6.06714562e-03
2.44982153e-01 -6.47196531e-01 6.74962819e-01 -9.04562414e-01
-1.15865421e+00 1.09891677e+00 -2.01261833e-01 -2.91370571e-01
7.46659338e-01 -1.02901474e-01 -3.77935171e-01 2.29988638e-02
3.11300278e-01 3.23351145e-01 5.41617036e-01 -1.40562475e+00
-6.29214287e-01 -6.32970870e-01 -1.27558395e-01 2.54298240e-01
-7.37525403e-01 1.26399964e-01 -7.32175469e-01 -6.27227962e-01
-7.33126923e-02 -1.00357652e+00 -9.78983268e-02 -3.49129081e-01
-6.24456704e-01 -4.04508322e-01 4.81840581e-01 -6.84350371e-01
1.41277421e+00 -2.23030472e+00 2.00744569e-01 4.81408358e-01
5.28920412e-01 2.84950465e-01 1.90247431e-01 3.30338687e-01
-2.45137870e-01 1.44846544e-01 -2.88585395e-01 -8.52470756e-01
1.07702106e-01 -2.19726801e-01 1.80939272e-01 8.64692032e-01
-2.07699582e-01 6.02436960e-01 -7.18707144e-01 -5.76576650e-01
3.89484346e-01 2.96893537e-01 -5.00856340e-01 3.23751926e-01
5.64075828e-01 6.98090017e-01 -7.99219981e-02 5.79014301e-01
9.14257407e-01 -3.83860081e-01 4.69945610e-01 -2.52131045e-01
-8.40677619e-02 -3.44448447e-01 -1.48082888e+00 1.51000392e+00
6.13620207e-02 3.94587010e-01 2.89852023e-02 -1.30548871e+00
8.18674326e-01 1.01352789e-01 7.56169498e-01 -3.39529186e-01
1.16985045e-01 7.41517693e-02 -7.69359916e-02 -3.26740414e-01
4.37410474e-01 1.79391086e-01 -1.87908590e-01 5.11938751e-01
-2.41512239e-01 8.63817871e-01 1.39493927e-01 1.44745424e-01
8.28402281e-01 -2.15229288e-01 1.49196997e-01 -3.52089375e-01
5.89145482e-01 -2.52365649e-01 7.92718172e-01 7.46221900e-01
-5.72630167e-01 7.95376658e-01 3.26947272e-02 -4.53435928e-01
-1.03641033e+00 -9.77318048e-01 -1.60488650e-01 9.22007918e-01
4.34148580e-01 -5.06273627e-01 -1.02639222e+00 -5.05392134e-01
1.16038686e-02 1.90443948e-01 -5.77120900e-01 -1.12533741e-01
-4.16503221e-01 -1.34758842e+00 7.12094545e-01 5.25389731e-01
7.38273501e-01 -4.60582614e-01 -1.37213781e-01 -8.10434297e-02
-6.18635595e-01 -1.08652520e+00 -5.34094214e-01 -3.01769763e-01
-5.41090548e-01 -1.24320149e+00 -8.28206480e-01 -7.41570532e-01
8.97712529e-01 4.88091737e-01 7.86759913e-01 3.74311954e-01
-2.03834429e-01 4.74169135e-01 -3.71985137e-01 6.71945233e-03
-3.34887654e-02 1.55764326e-01 6.99415088e-01 5.60238242e-01
1.05815327e+00 -7.11333573e-01 -7.94615149e-01 5.63528180e-01
-5.53369820e-01 -7.46714547e-02 3.13959271e-01 6.12653434e-01
4.58417326e-01 6.42577171e-01 6.24122441e-01 -6.26888335e-01
6.55214369e-01 -5.28438389e-01 -3.29854786e-01 4.21136707e-01
-5.61300695e-01 -4.24345076e-01 5.46440363e-01 -3.79358113e-01
-8.17443430e-01 1.04412563e-01 1.49993390e-01 -4.92256492e-01
-6.05823874e-01 3.38857204e-01 -4.84851956e-01 1.36614814e-01
4.21735972e-01 2.55919635e-01 -1.04426257e-01 -5.80245554e-01
2.13979274e-01 1.04446793e+00 8.42919946e-01 -6.38105214e-01
9.51658964e-01 7.59464800e-01 -3.38829547e-01 -7.35789299e-01
-4.91272300e-01 -1.03539705e+00 -1.32098114e+00 -3.93168926e-01
7.63043225e-01 -1.11060846e+00 -1.12448072e+00 7.29176641e-01
-8.44842613e-01 -1.73054159e-01 7.80198500e-02 4.51375812e-01
-2.68958002e-01 8.84867668e-01 -3.70355576e-01 -8.62689137e-01
-2.67281562e-01 -8.98711085e-01 9.14267480e-01 2.96312690e-01
-4.14988875e-01 -1.03089774e+00 5.59509769e-02 7.67113090e-01
-8.12992379e-02 2.86590010e-01 3.98152560e-01 -8.70067596e-01
-2.88763791e-01 -2.65248001e-01 -3.00547540e-01 6.98226541e-02
3.61039698e-01 -5.92744723e-02 -9.43334401e-01 -6.45557880e-01
-3.39761943e-01 6.16562329e-02 6.60469830e-01 3.00993025e-01
1.41991639e+00 -1.81581169e-01 -4.45431262e-01 7.90885150e-01
1.24233150e+00 -1.91504359e-01 5.10549963e-01 5.56585133e-01
1.14673877e+00 8.41580451e-01 4.45133954e-01 6.72525048e-01
8.80903065e-01 8.55826259e-01 9.60937068e-02 -1.49296165e-01
1.32078335e-01 -4.36925478e-02 2.53770705e-02 8.19976509e-01
-5.21583617e-01 -2.11588636e-01 -1.06103206e+00 6.50524437e-01
-2.21264482e+00 -1.17815685e+00 -2.78975934e-01 2.60640979e+00
6.60484076e-01 -2.42049292e-01 5.22897542e-01 4.42943841e-01
1.31511128e+00 -2.58049726e-01 -4.87753063e-01 2.38837898e-01
-1.54718095e-02 -3.88977975e-01 4.22406673e-01 2.25296333e-01
-1.38929355e+00 6.70334518e-01 5.33482122e+00 8.34630311e-01
-6.05519593e-01 3.27503569e-02 7.96027362e-01 -1.30842268e-01
1.54932350e-01 -3.00818563e-01 -8.12818646e-01 9.22795475e-01
7.94731259e-01 -4.92665693e-02 3.88945878e-01 5.70825994e-01
3.50446671e-01 7.31227323e-02 -1.24518156e+00 1.63194680e+00
9.06769335e-02 -9.47912335e-01 -1.22196235e-01 4.54036355e-01
5.03698468e-01 -3.51289243e-01 6.13239892e-02 -1.17652845e-02
2.62263149e-01 -9.83092308e-01 4.48634267e-01 5.53020179e-01
7.43736804e-01 -1.01568663e+00 9.30275202e-01 3.50987703e-01
-1.34828484e+00 -2.59519875e-01 -3.87026340e-01 -6.75213635e-02
3.60671207e-02 7.44352460e-01 -5.74778795e-01 8.66919160e-01
1.23128331e+00 9.42247152e-01 -7.79938161e-01 1.11130142e+00
3.53948504e-01 4.42564934e-01 -1.99165314e-01 4.66463655e-01
-2.71845877e-01 -4.45058852e-01 3.61586958e-01 1.29872370e+00
3.19077790e-01 1.12910971e-01 2.98691809e-01 8.37867498e-01
-2.40666140e-02 -6.34650374e-03 -2.50143826e-01 2.75443733e-01
7.39417076e-01 1.30205238e+00 -8.25127006e-01 -1.46475539e-01
-3.11449170e-01 1.18604136e+00 5.27356029e-01 3.46290529e-01
-8.69430304e-01 -1.23713538e-01 7.45831490e-01 1.79091498e-01
6.08237684e-02 -2.89924145e-01 -2.43014231e-01 -1.24100935e+00
2.17723250e-01 -8.24320793e-01 7.55163252e-01 -2.25558281e-01
-1.80284131e+00 4.01627004e-01 7.43087754e-02 -1.46314228e+00
-9.75902230e-02 -2.56368577e-01 -6.54213190e-01 6.83402240e-01
-1.24802959e+00 -1.23900473e+00 -7.92244732e-01 8.68028641e-01
1.72519475e-01 -3.76410037e-01 6.62532151e-01 6.76360786e-01
-1.13255930e+00 1.11647809e+00 3.39095563e-01 6.11389935e-01
9.19616878e-01 -1.25969374e+00 1.87826395e-01 1.17243922e+00
-2.17925329e-02 9.03323591e-01 5.33319712e-01 -6.04705393e-01
-9.92668390e-01 -1.13575983e+00 7.06773400e-01 -6.50678039e-01
3.80877703e-01 -4.50047910e-01 -8.85841131e-01 5.74848592e-01
-1.34661704e-01 -1.80795416e-01 1.32926476e+00 3.99896443e-01
-2.64411151e-01 -3.25862885e-01 -1.02586067e+00 4.07881439e-01
1.27906871e+00 -5.43684959e-01 -2.37561733e-01 2.39387512e-01
2.12527901e-01 -4.08962136e-03 -1.12201357e+00 3.33649665e-01
4.57847625e-01 -1.19953108e+00 1.17093909e+00 -2.69715250e-01
-3.83993015e-02 -6.19339347e-01 -3.07462290e-02 -1.07604265e+00
-5.83866477e-01 -5.40021718e-01 -7.37420917e-02 1.73470557e+00
-1.18490659e-01 -5.25605083e-01 8.65560770e-01 9.21235561e-01
1.96720198e-01 -2.56638169e-01 -9.89588976e-01 -7.28703737e-01
-9.18785334e-02 -1.26739860e-01 8.59101534e-01 1.38531399e+00
3.80166546e-02 -5.64153455e-02 -5.32400072e-01 6.36717737e-01
1.13759220e+00 5.48902228e-02 1.02607679e+00 -1.55196166e+00
-1.58814088e-01 -5.48104763e-01 -7.01489627e-01 -7.83938646e-01
2.22999960e-01 -8.11161518e-01 -2.34175622e-01 -1.26719177e+00
6.41114712e-01 -7.41927445e-01 -3.56275946e-01 2.29203597e-01
-4.55252886e-01 4.15322334e-01 1.48493335e-01 8.59692574e-01
-8.91293466e-01 4.93957072e-01 5.04190445e-01 -1.62346542e-01
-2.01761156e-01 2.15416119e-01 -8.87282610e-01 6.41000390e-01
9.91479576e-01 -2.00145617e-01 -2.69977927e-01 -3.87792259e-01
8.62444285e-03 -6.53712749e-01 7.08908916e-01 -1.23922265e+00
6.22625828e-01 1.25277057e-01 5.94545960e-01 -5.02412379e-01
1.20686524e-01 -8.61586034e-01 3.56940895e-01 8.50332379e-02
-1.85092866e-01 1.40243381e-01 -1.09002881e-01 7.14728475e-01
-1.76498875e-01 7.65006393e-02 6.23138487e-01 -4.48879711e-02
-6.41406715e-01 4.79225606e-01 -8.94093290e-02 -1.69125974e-01
1.16526115e+00 -5.08859038e-01 -2.13918805e-01 -3.07500094e-01
-7.37583816e-01 3.60656023e-01 7.60836124e-01 3.28974605e-01
3.44282448e-01 -1.52400172e+00 -8.59516263e-01 8.17994997e-02
3.25386167e-01 1.02068891e-03 5.25156260e-01 9.72311139e-01
-2.67704219e-01 1.94849357e-01 -1.46008566e-01 -8.08904767e-01
-1.57027245e+00 6.82983220e-01 4.82516974e-01 -8.29965174e-02
-5.11818409e-01 6.21034265e-01 1.85751498e-01 -6.69998527e-01
3.75030220e-01 2.27737620e-01 -3.80214959e-01 1.32031634e-01
7.09354579e-01 6.11308813e-01 -5.03800698e-02 -1.07293487e+00
-5.05712986e-01 7.43804693e-01 -1.33577764e-01 2.97018141e-01
1.11146009e+00 -6.00357413e-01 -2.78678447e-01 3.02357793e-01
1.07071877e+00 -1.68947473e-01 -1.26304233e+00 -3.00937980e-01
1.37233585e-01 -5.66555619e-01 -3.21347833e-01 -3.35766554e-01
-9.66390550e-01 6.77044392e-01 7.67876983e-01 2.44548976e-01
1.03716493e+00 -8.13761055e-02 6.37225926e-01 1.40588567e-01
2.95270234e-01 -1.41172624e+00 7.34200701e-02 -6.96897209e-02
3.46095949e-01 -1.48863089e+00 9.71214324e-02 -4.00253564e-01
-5.22154093e-01 8.70315969e-01 6.64331079e-01 1.87093854e-01
6.19555831e-01 -1.61055326e-01 -5.36858812e-02 -1.70548171e-01
4.84517477e-02 -2.62791723e-01 2.30590120e-01 8.06898117e-01
2.53121585e-01 3.40342969e-01 1.03475966e-01 8.37961555e-01
-2.83305377e-01 -2.80271947e-01 3.21757376e-01 4.66267407e-01
-2.72046831e-02 -1.09052265e+00 -6.23055696e-01 2.55222678e-01
-3.12845469e-01 5.95946908e-02 -4.74880159e-01 6.24191344e-01
4.00273502e-01 1.21436226e+00 3.05976570e-01 -6.38605535e-01
5.96122928e-02 -2.21095353e-01 2.49320403e-01 -3.46046329e-01
-4.41621900e-01 -5.94574101e-02 -1.41624555e-01 -3.47846389e-01
-9.55414236e-01 -8.61652374e-01 -1.01126230e+00 -8.57392013e-01
-1.99282572e-01 2.04695538e-01 2.59077549e-01 9.14492965e-01
5.71329474e-01 1.78139701e-01 7.29159236e-01 -8.55440259e-01
-2.65720129e-01 -8.52458835e-01 -5.69802880e-01 7.83838212e-01
1.50049374e-01 -6.90506816e-01 -3.55028003e-01 2.89070278e-01] | [14.769647598266602, 1.0612694025039673] |
543deea0-066c-43e6-820f-c3e79a254938 | english-language-spelling-correction-as-an | null | null | https://aclanthology.org/2022.lrec-1.48 | https://aclanthology.org/2022.lrec-1.48.pdf | English Language Spelling Correction as an Information Retrieval Task Using Wikipedia Search Statistics | Spelling correction utilities have become commonplace during the writing process, however, many spelling correction utilities suffer due to the size and quality of dictionaries available to aid correction. Many terms, acronyms, and morphological variations of terms are often missing, leaving potential spelling errors unidentified and potentially uncorrected. This research describes the implementation of WikiSpell, a dynamic spelling correction tool that relies on the Wikipedia dataset search API functionality as the sole source of knowledge to aid misspelled term identification and automatic replacement. Instead of a traditional matching process to select candidate replacement terms, the replacement process is treated as a natural language information retrieval process harnessing wildcard string matching and search result statistics. The aims of this research include: 1) the implementation of a spelling correction algorithm that utilizes the wildcard operators in the Wikipedia dataset search API, 2) a review of the current spell correction tools and approaches being utilized, and 3) testing and validation of the developed algorithm against the benchmark spelling correction tool, Hunspell. The key contribution of this research is a robust, dynamic information retrieval-based spelling correction algorithm that does not require prior training. Results of this research show that the proposed spelling correction algorithm, WikiSpell, achieved comparable results to an industry-standard spelling correction algorithm, Hunspell. | ['Markus Hofmann', 'Kyle Goslin'] | null | null | null | null | lrec-2022-6 | ['spelling-correction'] | ['natural-language-processing'] | [ 4.25888926e-01 -3.92887414e-01 5.23998998e-02 1.85924117e-02
-5.56934536e-01 -9.75799441e-01 6.71044827e-01 9.20432031e-01
-6.28512979e-01 7.79180586e-01 2.00213253e-01 -3.42294186e-01
-5.47064364e-01 -5.32151461e-01 -5.25681019e-01 -8.48030373e-02
5.99858105e-01 6.43461525e-01 2.92987853e-01 -6.02443457e-01
1.15944958e+00 5.31862020e-01 -1.83282185e+00 3.44006300e-01
1.44514418e+00 2.68866390e-01 5.98474264e-01 8.65604520e-01
-6.45697117e-01 3.75528008e-01 -9.16999280e-01 -3.41405571e-01
3.75482589e-01 -6.67019933e-02 -8.41754019e-01 -6.81282699e-01
7.95477211e-01 1.01222523e-01 -2.49880925e-03 1.16096008e+00
5.21174192e-01 2.57952303e-01 5.11109591e-01 -9.08416986e-01
-8.00983727e-01 5.94984591e-01 1.55138671e-01 3.68345886e-01
1.01972914e+00 -2.81152219e-01 7.50694692e-01 -9.37307000e-01
7.55370498e-01 7.69268751e-01 1.00398779e+00 2.95728147e-01
-6.59945011e-01 -6.47384465e-01 -5.74486613e-01 5.83935194e-02
-1.64572096e+00 3.28113399e-02 1.42628908e-01 -5.09531379e-01
1.48618197e+00 7.63080895e-01 6.04272842e-01 4.66887772e-01
4.14543957e-01 4.26611990e-01 1.09835529e+00 -1.42002678e+00
-1.46667451e-01 6.18006051e-01 7.84846663e-01 3.94157171e-01
7.90072858e-01 -9.02376845e-02 -7.37912953e-01 -1.85881391e-01
1.18822671e-01 -6.74094856e-02 -2.44755775e-01 -7.07997307e-02
-9.85252857e-01 2.31278434e-01 -3.17310572e-01 6.77174270e-01
-3.45024854e-01 -3.16146165e-01 6.10041559e-01 3.78907651e-01
9.42801386e-02 1.07797372e+00 -4.25505757e-01 -4.18336719e-01
-1.55043316e+00 7.63477504e-01 9.00009871e-01 1.33013523e+00
5.80980122e-01 -4.02797014e-01 -6.68303430e-01 8.47917795e-01
4.60447401e-01 8.79273534e-01 1.14542842e+00 -1.50721401e-01
4.56370950e-01 1.09007549e+00 4.85913277e-01 -1.20960033e+00
-9.73520279e-02 -4.06484932e-01 1.09765001e-01 1.67029016e-02
1.78377405e-01 2.24193811e-01 -1.09715712e+00 1.02610052e+00
-8.89167413e-02 -6.04655027e-01 -1.57840461e-01 4.19810057e-01
6.60899699e-01 5.05190492e-01 1.35926716e-02 -7.13167340e-02
1.17408943e+00 -1.10266948e+00 -9.99390960e-01 -2.09204853e-02
8.21009874e-01 -1.45496285e+00 1.28697860e+00 4.93017763e-01
-1.09467649e+00 -3.61763686e-01 -1.18783402e+00 -8.53705034e-02
-1.09304762e+00 1.75368264e-01 4.50649381e-01 1.00817657e+00
-8.76937628e-01 8.39313269e-01 -3.36115569e-01 -7.70641446e-01
-4.32768375e-01 6.59649149e-02 -2.51367480e-01 -5.38274236e-02
-1.30046666e+00 1.37665844e+00 7.18767762e-01 -4.53349918e-01
-1.41990110e-01 -9.94352400e-01 -5.76819897e-01 7.54647255e-02
4.57823388e-02 -4.32375163e-01 1.36753368e+00 -1.05666947e+00
-9.15181816e-01 8.89724374e-01 -2.46616960e-01 -3.94020706e-01
3.52467239e-01 -5.96045196e-01 -9.05880928e-01 -2.57088184e-01
4.06697124e-01 -2.50430852e-01 4.55544800e-01 -9.21725512e-01
-9.54871416e-01 -5.89742437e-02 -4.55169857e-01 5.59740722e-01
-2.76920110e-01 4.26093787e-01 -4.37617868e-01 -1.20602310e+00
-2.00713366e-01 -8.28738332e-01 3.56425613e-01 -5.42085290e-01
-6.82438314e-02 -3.19610201e-02 3.92583072e-01 -1.19363475e+00
2.36525846e+00 -1.75804579e+00 -2.56809711e-01 6.88863575e-01
-4.71759677e-01 1.03451073e+00 1.23680733e-01 1.11189198e+00
-1.31413534e-01 2.91995436e-01 -1.37855217e-01 5.09839281e-02
3.79031859e-02 -1.11137696e-01 -3.05231899e-01 -1.16857532e-02
-7.11647749e-01 4.84209746e-01 -1.16662800e+00 -4.22262073e-01
2.15814054e-01 5.43783717e-02 -1.22636482e-01 -6.92964345e-02
7.04981619e-03 -4.25305814e-01 1.05568811e-01 7.78070748e-01
6.55066967e-01 3.89693767e-01 -2.87890047e-01 -7.86700938e-03
-6.69819653e-01 5.50666749e-01 -1.60815525e+00 1.61919200e+00
-5.47444463e-01 5.68865120e-01 -4.91098225e-01 -1.52194872e-01
8.03149521e-01 2.17547253e-01 3.57024409e-02 -7.38766372e-01
-9.65737700e-02 9.77147937e-01 -4.49919730e-01 -6.24096692e-01
1.19317245e+00 4.15883750e-01 -2.80451272e-02 4.61078584e-01
6.95345877e-03 -2.68210500e-01 7.91158676e-01 3.30617428e-01
1.02459323e+00 2.90331304e-01 6.14879966e-01 -3.24554980e-01
9.82271492e-01 5.21647811e-01 1.43910065e-01 1.12086082e+00
8.47063139e-02 4.55305010e-01 -5.48991024e-01 -2.01574221e-01
-1.16055882e+00 -8.27265441e-01 -2.40885824e-01 1.01532710e+00
-2.18904205e-02 -1.01283884e+00 -1.08023894e+00 -4.28381324e-01
2.23861277e-01 1.22690761e+00 -2.32925743e-01 -1.32366925e-01
-2.62347192e-01 -2.14349523e-01 9.00864899e-01 -2.96276733e-02
2.53859967e-01 -1.15220881e+00 -3.83720875e-01 2.98247546e-01
2.74115168e-02 -2.21572027e-01 -7.87398219e-01 1.47713676e-01
-5.47589362e-01 -9.02596772e-01 -6.68138027e-01 -8.50711763e-01
6.24271214e-01 4.14148331e-01 9.62847769e-01 4.84027416e-01
-3.94530982e-01 5.04210293e-01 -9.07415271e-01 -8.59841466e-01
-7.46574163e-01 2.35928088e-01 1.01733126e-01 -7.49889255e-01
9.11807477e-01 -2.09607869e-01 -2.98925191e-01 -8.68791491e-02
-1.12788475e+00 -3.83826405e-01 5.96853495e-01 6.30861878e-01
4.28151518e-01 2.51748301e-02 1.36145756e-01 -1.12288415e+00
1.20976257e+00 -3.08824837e-01 -5.09777606e-01 8.33545327e-01
-1.45259929e+00 2.07628042e-01 4.73879874e-01 9.01548658e-03
-9.31095600e-01 -1.14467770e-01 5.03104739e-02 2.69812226e-01
-1.11608662e-01 6.15780056e-01 1.80976108e-01 -5.36729276e-01
9.48353708e-01 6.45004034e-01 -2.44170889e-01 -8.42710912e-01
1.01377331e-01 1.18808651e+00 5.18840790e-01 -1.31676137e-01
7.62901783e-01 -2.75698930e-01 -4.88560200e-01 -7.87744522e-01
-4.46997136e-01 -1.05363452e+00 -7.38490403e-01 -1.81195885e-01
3.13098937e-01 -5.65659106e-01 4.85281199e-02 8.67801487e-01
-1.07400489e+00 1.73852772e-01 -1.74173981e-01 2.94613063e-01
-1.10679707e-02 8.04621041e-01 -1.31081641e-01 -6.12205505e-01
-9.61650968e-01 -5.92704475e-01 6.62311316e-01 5.35464525e-01
-7.14682817e-01 -7.86671281e-01 5.89362323e-01 4.88945246e-01
4.77790922e-01 -2.65737802e-01 9.12876844e-01 -1.03762066e+00
-1.12606570e-01 -6.41046107e-01 2.05564916e-01 4.91235673e-01
3.56281430e-01 2.59425312e-01 -3.67081732e-01 -1.31312147e-01
-4.29674745e-01 3.65161180e-01 3.55426490e-01 -2.26068243e-01
5.93196213e-01 -3.92351449e-01 -3.62375408e-01 5.44868231e-01
1.57513046e+00 3.54655355e-01 6.40771627e-01 1.24730480e+00
5.53700328e-01 1.03532374e-01 1.09792268e+00 4.40344572e-01
5.39818034e-03 6.15440607e-01 -5.99314272e-02 5.97664952e-01
-3.89623940e-01 -4.86321509e-01 1.77768573e-01 9.95891929e-01
8.72603133e-02 -3.53203356e-01 -1.19193935e+00 6.65839016e-01
-1.51416993e+00 -1.02759242e+00 -4.21022594e-01 2.71828651e+00
1.11529338e+00 9.75446701e-02 -2.36266375e-01 5.14025390e-01
7.11830139e-01 -4.12190169e-01 1.25386864e-01 -1.19726253e+00
-1.43491521e-01 5.50142229e-01 1.27780175e+00 6.66291416e-01
-9.08493698e-01 1.17883122e+00 6.04273272e+00 9.14889038e-01
-9.42645907e-01 -1.96981475e-01 -7.29619205e-01 5.99658154e-02
-4.95473385e-01 6.80576339e-02 -1.08699322e+00 8.56284380e-01
7.35310256e-01 -6.08046055e-01 7.17730582e-01 8.62374127e-01
2.29922712e-01 -2.47680113e-01 -6.61837101e-01 9.89371538e-01
6.71299040e-01 -1.30716980e+00 6.05750561e-01 -6.39032006e-01
8.69658887e-01 -2.61505127e-01 -3.55912179e-01 1.94891393e-01
1.92848340e-01 -8.90987694e-01 1.15914226e+00 7.51080096e-01
3.94799680e-01 -7.06350267e-01 8.08651984e-01 3.10325086e-01
-6.87989712e-01 1.15785979e-01 -1.53486341e-01 -1.31521732e-01
-1.18632086e-01 3.31980467e-01 -9.89693999e-01 5.36480606e-01
6.60559356e-01 2.63115168e-01 -1.06389964e+00 1.77209210e+00
-3.26723546e-01 3.38348538e-01 -1.80895269e-01 -3.43959630e-01
3.81006487e-03 -2.29992606e-02 8.93869460e-01 1.82209241e+00
6.96218610e-01 -3.37328136e-01 -3.92049700e-01 3.30071062e-01
3.20306480e-01 7.09728360e-01 -3.86970311e-01 -1.38528660e-01
1.16913593e+00 8.89968395e-01 -4.48585421e-01 -2.53079474e-01
-1.83853701e-01 1.35657632e+00 -1.11744776e-01 -1.63261686e-02
-4.40021783e-01 -1.14274144e+00 4.34773654e-01 2.64787704e-01
1.34649500e-01 -8.19247141e-02 -5.98962128e-01 -7.13111162e-01
2.18127653e-01 -1.33408129e+00 3.52156162e-01 -8.05038452e-01
-7.44172156e-01 2.07524344e-01 1.53616771e-01 -1.25800359e+00
5.13583496e-02 -5.19688487e-01 -4.94703591e-01 1.36207473e+00
-1.15429056e+00 -1.04533434e+00 -3.61711979e-01 2.15518773e-01
6.79106236e-01 -3.73264521e-01 8.53740036e-01 7.03874290e-01
-2.99929857e-01 9.90633428e-01 7.23170638e-01 -2.30886102e-01
1.24249768e+00 -1.55287278e+00 1.84082627e-01 1.32001293e+00
1.12054177e-01 1.46638596e+00 1.05575550e+00 -1.34208953e+00
-1.18545926e+00 -8.48477662e-01 1.74477386e+00 -3.95172775e-01
4.97003555e-01 -1.03371544e-02 -8.88938069e-01 4.82458353e-01
3.09446514e-01 -8.81846905e-01 6.73633218e-01 -2.98515797e-01
-1.44137576e-01 -1.28583893e-01 -1.26854515e+00 5.59768200e-01
5.86870432e-01 -5.78883767e-01 -1.26472998e+00 3.57650369e-01
2.57146031e-01 -6.12821579e-01 -4.97365892e-01 -1.45780174e-02
6.93844438e-01 -4.62904721e-01 7.44670928e-01 -2.19795629e-01
-5.35925925e-02 -6.66854560e-01 1.34517118e-01 -1.44940078e+00
-4.49995309e-01 -7.61963665e-01 1.25452369e-01 1.75426519e+00
8.49866033e-01 -2.86926985e-01 2.70004034e-01 8.73347759e-01
-2.96295911e-01 1.30664045e-02 -5.36673963e-01 -8.33398819e-01
-9.22825113e-02 -1.39207140e-01 5.80814302e-01 8.36360812e-01
2.45451480e-01 -1.64646953e-01 -1.31504387e-01 1.36510357e-01
2.22055018e-01 -3.43272060e-01 5.49828708e-01 -1.05191934e+00
-9.90714505e-02 -4.81789052e-01 -2.26140186e-01 -4.33094144e-01
-3.98250759e-01 -1.23944449e+00 1.05631642e-01 -1.69377863e+00
1.68092661e-02 -4.62333441e-01 -1.89904243e-01 3.24758857e-01
-3.12052101e-01 -4.23564576e-02 2.33224332e-01 4.71716493e-01
-2.87048310e-01 -1.93226054e-01 4.89744157e-01 2.91532446e-02
-3.15531611e-01 5.61329946e-02 -5.76272786e-01 6.39857292e-01
6.05206370e-01 -8.42584252e-01 -3.61903340e-01 -5.39487123e-01
8.07535112e-01 -7.81668186e-01 -8.11768919e-02 -1.25991058e+00
6.00440562e-01 -1.72119036e-01 -9.96962469e-03 -6.79556251e-01
-5.50604224e-01 -8.87123287e-01 3.46066922e-01 4.76702631e-01
-5.43994725e-01 6.79631412e-01 3.39859635e-01 1.83227986e-01
3.79359126e-02 -1.28180707e+00 4.00532603e-01 -2.42524743e-01
-1.14385009e+00 -3.82182181e-01 -7.42266357e-01 6.30823523e-02
1.13918793e+00 -7.34807909e-01 -1.37915000e-01 2.95909524e-01
-1.09234810e-01 -1.65835962e-01 9.93396759e-01 6.11233413e-01
1.66467309e-01 -1.14657056e+00 -3.69487911e-01 6.40466362e-02
3.51195633e-01 -4.49712753e-01 -1.12351209e-01 2.91469157e-01
-1.53802204e+00 8.86966050e-01 -5.48272967e-01 2.98204988e-01
-1.80190074e+00 5.75003922e-01 1.28437474e-01 -2.29956448e-01
-3.13398391e-01 8.72988522e-01 -9.09681499e-01 -4.30607408e-01
5.23869038e-01 -3.17125022e-01 -6.14578366e-01 8.58558640e-02
9.65220630e-01 7.69319713e-01 7.96715379e-01 -5.56548715e-01
-5.91473401e-01 5.45451462e-01 -3.89629990e-01 -1.39995486e-01
7.89460123e-01 -5.49040973e-01 -3.12877476e-01 3.61476302e-01
6.48044586e-01 7.48993635e-01 -3.68262008e-02 -6.00281656e-02
5.65143764e-01 -7.15241790e-01 2.79412307e-02 -1.57854640e+00
-1.29325077e-01 2.15708196e-01 7.93392777e-01 -2.45451741e-02
8.95928144e-01 -8.18665981e-01 8.08680952e-01 6.57443404e-01
3.98246914e-01 -1.76123798e+00 -7.68345237e-01 8.98667157e-01
8.41188848e-01 -9.09751534e-01 1.47365123e-01 -1.67382240e-01
-4.52123106e-01 1.18385947e+00 3.34914923e-01 6.91228062e-02
4.05357599e-01 1.74879283e-01 5.28600812e-02 -2.43238397e-02
-1.01046897e-01 3.40775996e-02 5.73318124e-01 6.84581757e-01
8.19951713e-01 -1.05246782e-01 -1.34615529e+00 6.66283429e-01
-4.47612405e-01 3.82196516e-01 5.29768407e-01 1.28704941e+00
-6.44463301e-01 -1.56499279e+00 -6.25316083e-01 5.45681834e-01
-6.05572164e-01 -9.10179734e-01 -8.94687116e-01 6.14976346e-01
4.54719841e-01 8.40944111e-01 -4.22743499e-01 -2.04492897e-01
4.75845456e-01 3.26624632e-01 3.29490334e-01 -9.19331253e-01
-1.42181766e+00 -8.40540707e-01 2.04096183e-01 -2.76459843e-01
-2.08677091e-02 -8.69445264e-01 -1.00906217e+00 -4.13695753e-01
-6.18448138e-01 5.21959424e-01 9.84878480e-01 9.47435737e-01
4.20470476e-01 2.20350131e-01 6.91412613e-02 -3.43550414e-01
-5.74499488e-01 -1.02530754e+00 -3.56197894e-01 6.50060236e-01
8.64634104e-03 -5.71066737e-01 -1.70529902e-01 1.16178654e-01] | [10.892062187194824, 10.607683181762695] |
24a9bb42-d7cd-4d9b-8141-9ff74a9e3f19 | epic-fields-marrying-3d-geometry-and-video | 2306.08731 | null | https://arxiv.org/abs/2306.08731v1 | https://arxiv.org/pdf/2306.08731v1.pdf | EPIC Fields: Marrying 3D Geometry and Video Understanding | Neural rendering is fuelling a unification of learning, 3D geometry and video understanding that has been waiting for more than two decades. Progress, however, is still hampered by a lack of suitable datasets and benchmarks. To address this gap, we introduce EPIC Fields, an augmentation of EPIC-KITCHENS with 3D camera information. Like other datasets for neural rendering, EPIC Fields removes the complex and expensive step of reconstructing cameras using photogrammetry, and allows researchers to focus on modelling problems. We illustrate the challenge of photogrammetry in egocentric videos of dynamic actions and propose innovations to address them. Compared to other neural rendering datasets, EPIC Fields is better tailored to video understanding because it is paired with labelled action segments and the recent VISOR segment annotations. To further motivate the community, we also evaluate two benchmark tasks in neural rendering and segmenting dynamic objects, with strong baselines that showcase what is not possible today. We also highlight the advantage of geometry in semi-supervised video object segmentations on the VISOR annotations. EPIC Fields reconstructs 96% of videos in EPICKITCHENS, registering 19M frames in 99 hours recorded in 45 kitchens. | ['Andrea Vedaldi', 'Dima Damen', 'Diane Larlus', 'Iro Laina', 'David Fouhey', 'Zhifan Zhu', 'Ahmad Darkhalil', 'Vadim Tschernezki'] | 2023-06-14 | null | null | null | null | ['neural-rendering', 'video-understanding'] | ['computer-vision', 'computer-vision'] | [ 4.21978325e-01 1.77644715e-01 -1.01293273e-01 -3.19724143e-01
-6.85896695e-01 -7.71635890e-01 7.68741906e-01 -5.67619562e-01
-4.31987315e-01 3.83526236e-01 4.20921862e-01 -2.32271254e-01
1.13286473e-01 -5.85112154e-01 -1.23276198e+00 -4.52970415e-01
1.32158205e-01 5.89911878e-01 1.44169018e-01 -7.06471279e-02
-8.25911984e-02 5.39464176e-01 -1.76086950e+00 7.10964084e-01
1.78363070e-01 9.59010422e-01 2.18763649e-02 9.78499591e-01
1.24192238e-02 9.73386168e-01 -3.88123453e-01 -7.51339018e-01
4.79713559e-01 -3.55119020e-01 -1.01670241e+00 3.98349941e-01
1.23944509e+00 -7.87981629e-01 -6.25684619e-01 5.19782901e-01
3.64782482e-01 2.78426677e-01 5.80774307e-01 -1.30115688e+00
-3.36685300e-01 3.81526977e-01 -5.02299249e-01 5.72624616e-02
4.52539325e-01 3.68052095e-01 8.71490836e-01 -5.85065246e-01
9.75213885e-01 1.27680516e+00 1.01611018e+00 5.68505049e-01
-1.13227952e+00 -5.10437012e-01 2.02100754e-01 1.77539080e-01
-9.23617482e-01 -6.27428770e-01 6.56141043e-01 -6.94704473e-01
1.18364954e+00 3.32054108e-01 1.17663622e+00 1.64282477e+00
-2.46493161e-01 1.02129066e+00 9.24412906e-01 9.90891084e-03
1.86327808e-02 -3.53698194e-01 -2.19716698e-01 7.55880058e-01
-1.95310012e-01 3.77160460e-01 -5.77026069e-01 4.58411276e-01
1.06386602e+00 -4.71591316e-02 -4.33817923e-01 -5.00518203e-01
-1.39496219e+00 6.10274732e-01 1.42404869e-01 -1.56845585e-01
-1.95031673e-01 6.76686704e-01 6.41401529e-01 -7.03128939e-03
7.04949677e-01 2.63654053e-01 -4.77978587e-01 -6.26396060e-01
-1.21435559e+00 3.31005216e-01 7.61124372e-01 1.02677560e+00
4.99191374e-01 2.51521796e-01 2.61706650e-01 5.24341226e-01
-6.65487256e-03 4.08962190e-01 -4.34737056e-02 -1.74816489e+00
6.74008667e-01 4.46366102e-01 -6.46342244e-03 -7.01799512e-01
-4.70774472e-01 -4.98730615e-02 -5.33678412e-01 3.67020547e-01
8.36814642e-01 -1.45606354e-01 -8.23230743e-01 1.33634257e+00
3.61686200e-01 3.93288285e-01 -1.54211834e-01 1.02565753e+00
9.45282221e-01 6.21485710e-01 -2.62135826e-02 3.10071111e-01
1.18524146e+00 -1.03486121e+00 -3.07794839e-01 -1.52686685e-01
5.55390954e-01 -6.77712023e-01 1.07881391e+00 7.50086069e-01
-1.35304487e+00 -6.09148085e-01 -8.23893964e-01 -4.74310368e-01
-2.83684403e-01 1.28112838e-01 1.01202166e+00 6.79564714e-01
-1.20567906e+00 8.73203635e-01 -1.01501310e+00 -4.71625984e-01
8.09872866e-01 1.27816722e-01 -7.76793659e-01 -2.02202454e-01
-7.17005193e-01 6.59763336e-01 3.93588871e-01 9.94717106e-02
-1.13537061e+00 -1.10920167e+00 -9.61505532e-01 -2.63959497e-01
3.97374600e-01 -9.11615610e-01 1.20965874e+00 -1.14399719e+00
-1.48293257e+00 1.13978779e+00 3.89457077e-01 -4.55765992e-01
1.16263628e+00 -4.53270346e-01 -2.16566809e-02 4.39231008e-01
-1.94821700e-01 1.16698277e+00 7.60572553e-01 -1.15444207e+00
-6.61960483e-01 -2.06954226e-01 4.83356982e-01 2.36156151e-01
9.68739986e-02 8.52649473e-03 -9.26456690e-01 -5.38305938e-01
2.73386040e-03 -9.99246716e-01 -7.16022030e-02 1.05776042e-01
-2.95177400e-01 7.77896419e-02 8.20505917e-01 -9.51047361e-01
3.45478982e-01 -1.92113054e+00 3.90061796e-01 -3.16339284e-01
1.46658286e-01 2.24820226e-01 -1.82547346e-01 2.88825125e-01
-2.81914562e-01 1.16915330e-01 -2.10344553e-01 -7.39450455e-01
5.30868284e-02 4.23537374e-01 -4.88373071e-01 6.39763534e-01
2.92763144e-01 1.03118813e+00 -8.02459180e-01 -4.04251784e-01
6.22091174e-01 8.56227994e-01 -8.59517097e-01 3.14178392e-02
-5.31258523e-01 6.86985672e-01 5.50943278e-02 8.26072335e-01
5.62862635e-01 -1.23797141e-01 9.95386485e-03 -3.51959318e-01
-1.99200530e-02 -1.07295178e-02 -1.09595132e+00 2.27928472e+00
-2.76998699e-01 9.97689784e-01 -2.24802922e-02 -1.05314553e+00
3.82436067e-01 2.25403398e-01 8.06066632e-01 -6.64029360e-01
3.11255723e-01 -2.30337963e-01 -7.25635231e-01 -8.62073302e-01
5.84388971e-01 3.66251692e-02 1.02089979e-01 3.40795159e-01
2.39230260e-01 -6.21883333e-01 1.44838646e-01 2.12462127e-01
1.15732491e+00 1.25616753e+00 -2.17050344e-01 1.58418491e-01
4.44781929e-02 3.33108217e-01 3.86346638e-01 4.62608546e-01
-6.35980219e-02 1.07687426e+00 5.80278754e-01 -7.60470688e-01
-1.33153737e+00 -1.00804698e+00 -9.36947856e-03 9.60868716e-01
-4.31017689e-02 -5.45310199e-01 -1.13480949e+00 -7.97834575e-01
-9.05374363e-02 5.66015780e-01 -7.38675416e-01 2.40967527e-01
-8.99073422e-01 -5.47827661e-01 7.50358224e-01 7.65146673e-01
4.83068287e-01 -9.63484704e-01 -8.58030021e-01 -2.38390919e-02
-3.97044569e-01 -1.68480706e+00 -1.60154909e-01 -9.80804786e-02
-9.86190856e-01 -1.41936529e+00 -7.62727380e-01 -3.32811177e-01
3.40597481e-01 3.70370984e-01 1.73042214e+00 -8.61289501e-02
-3.77926230e-01 1.05576158e+00 -2.93381304e-01 -2.30004877e-01
-4.50453430e-01 -2.67203748e-02 -2.31845096e-01 -1.71906665e-01
8.08679760e-02 -7.09614515e-01 -6.32952988e-01 2.98285365e-01
-8.72651875e-01 6.38804853e-01 4.09398139e-01 3.58807534e-01
6.01234555e-01 -4.05681700e-01 -3.82562242e-02 -1.10502267e+00
-2.78273106e-01 -4.77717847e-01 -6.67055249e-01 -9.80832130e-02
-5.13853952e-02 -3.07052463e-01 3.54426891e-01 -3.30316663e-01
-1.25358927e+00 5.24917662e-01 -1.58518285e-01 -6.94753110e-01
-5.00429451e-01 1.45913064e-02 -1.24851547e-01 1.21097947e-02
6.71276987e-01 -1.82853639e-01 -2.76323315e-02 -5.75537384e-01
8.37799430e-01 5.84105030e-02 1.00960267e+00 -6.78278625e-01
6.03398442e-01 9.06379759e-01 1.26426499e-02 -7.42041290e-01
-7.29429483e-01 -1.52039796e-01 -9.26824987e-01 -6.18316412e-01
1.28361619e+00 -1.18007672e+00 -8.32824707e-01 3.40456367e-01
-1.25500572e+00 -8.13002467e-01 -5.93930185e-01 3.74638885e-01
-1.07454002e+00 4.45375025e-01 -5.78758776e-01 -5.03591180e-01
3.62766050e-02 -1.17568576e+00 1.36573029e+00 -2.70874381e-01
-2.06739962e-01 -9.16806042e-01 -1.38225496e-01 9.87093925e-01
-8.96391422e-02 9.01794255e-01 4.98966813e-01 3.51759866e-02
-1.04076624e+00 -1.61777034e-01 -3.74099225e-01 5.38139462e-01
-4.10567850e-01 1.41112402e-01 -1.35119045e+00 9.58531722e-02
-8.65381584e-02 -4.67994303e-01 9.78408396e-01 4.68287081e-01
1.30934441e+00 2.48663388e-02 -2.70911660e-02 1.28380811e+00
1.17162383e+00 -6.06392231e-03 8.44227612e-01 2.18830094e-01
1.28708196e+00 8.82390559e-01 4.55943197e-01 2.08385468e-01
7.08288968e-01 6.23750031e-01 8.35864663e-01 -2.99439251e-01
-5.81138611e-01 -3.44801813e-01 3.49857599e-01 6.18506134e-01
-7.39679456e-01 -1.56131133e-01 -8.56577218e-01 4.36038792e-01
-1.68899167e+00 -1.00206769e+00 -3.55673909e-01 1.84753489e+00
6.10753775e-01 -1.15453959e-01 2.33491853e-01 2.10620612e-01
2.61124283e-01 3.75326425e-01 -4.75750506e-01 -1.12858042e-01
-3.09230357e-01 6.45640045e-02 5.98217309e-01 3.21044654e-01
-1.25813603e+00 1.09714460e+00 6.09015703e+00 5.79156756e-01
-8.05572569e-01 1.90541938e-01 6.94322467e-01 -4.33579326e-01
-1.27946571e-01 5.23571633e-02 -6.31728351e-01 2.39199325e-01
9.18031931e-01 6.93903744e-01 8.10954154e-01 8.63699317e-01
3.24976027e-01 -2.15731487e-01 -1.51862228e+00 1.26054657e+00
4.32906687e-01 -1.74726665e+00 -3.50137539e-02 1.15763471e-01
8.76637638e-01 3.71679515e-01 4.88109738e-02 2.60048091e-01
4.20672297e-01 -1.16270709e+00 1.18057013e+00 6.54774606e-01
9.23847854e-01 -6.01753354e-01 3.23881924e-01 1.73239708e-01
-9.64706957e-01 2.30822079e-02 -1.87162176e-01 -6.19270280e-02
4.53891456e-01 9.67695862e-02 -6.70822501e-01 5.17531991e-01
1.04659545e+00 1.19803309e+00 -5.65594614e-01 7.22072780e-01
-2.03616291e-01 5.25983989e-01 -2.23413810e-01 6.35175169e-01
3.71601731e-01 -4.77540046e-01 3.50643843e-01 1.24475551e+00
1.10128008e-01 -3.82917449e-02 -4.28607725e-02 8.22631955e-01
-2.02749982e-01 -2.26850092e-01 -7.33063996e-01 -3.60155515e-02
-1.95687532e-01 1.23921227e+00 -9.58029330e-01 -3.82252783e-01
-4.22287315e-01 1.18277347e+00 1.71518788e-01 3.54218215e-01
-1.11530769e+00 2.89985508e-01 6.67565763e-01 2.37981290e-01
1.63376838e-01 -4.65863049e-01 -2.23788291e-01 -1.21790946e+00
-1.14714541e-01 -1.01701558e+00 2.66393483e-01 -1.14541256e+00
-6.87339008e-01 3.13709140e-01 2.20602050e-01 -1.02877450e+00
-3.41304600e-01 -6.45441473e-01 -1.35154411e-01 2.71215767e-01
-1.45690489e+00 -1.54901230e+00 -7.63990939e-01 5.76850474e-01
1.02160132e+00 7.29182735e-02 4.93292838e-01 4.97852772e-01
-4.63867962e-01 1.34830922e-01 -1.98813275e-01 1.83310643e-01
5.96615374e-01 -1.31492913e+00 9.71679926e-01 7.55510509e-01
5.51608503e-01 -2.49777865e-02 4.46976602e-01 -4.34538305e-01
-1.63992548e+00 -1.15155923e+00 2.86506385e-01 -1.23793304e+00
4.05098349e-01 -6.12168074e-01 -6.82738662e-01 1.02802491e+00
-2.62519927e-03 6.52247593e-02 3.42314154e-01 -2.69614875e-01
-4.57795024e-01 1.55231893e-01 -7.97223270e-01 5.92999876e-01
1.74203718e+00 -5.59384406e-01 -1.60642281e-01 6.41251385e-01
7.56817043e-01 -8.88283491e-01 -9.63247240e-01 2.80742258e-01
7.86254227e-01 -1.37773073e+00 1.45024121e+00 -6.23862088e-01
7.96965420e-01 -1.26588300e-01 -2.49393091e-01 -9.14064527e-01
3.00402641e-01 -7.70078123e-01 -7.02173710e-02 1.23807085e+00
-1.54368147e-01 -1.09899014e-01 1.26408958e+00 7.02673197e-01
-4.47994649e-01 -6.01161361e-01 -7.12609231e-01 -4.59650218e-01
5.02817333e-04 -1.10716653e+00 5.90619385e-01 9.02979255e-01
-7.24577606e-01 -1.57008730e-02 -5.13875127e-01 -2.06279889e-01
5.00635743e-01 -2.33262807e-01 1.29317105e+00 -1.13228023e+00
-5.33367455e-01 -4.07538742e-01 -4.70967650e-01 -1.13398159e+00
3.38037878e-01 -8.63020480e-01 -2.60277331e-01 -1.59141994e+00
1.98241547e-02 -2.83138603e-01 3.98028433e-01 3.80805969e-01
2.92102307e-01 9.25790548e-01 2.96596348e-01 7.76895061e-02
-6.43562496e-01 2.21904501e-01 1.27716136e+00 4.67825718e-02
-1.00587821e-02 -2.81312734e-01 -3.51408035e-01 1.16202497e+00
5.09866953e-01 -1.43577486e-01 -5.50407767e-01 -8.80009949e-01
4.38805163e-01 1.57785207e-01 8.27957571e-01 -1.16035414e+00
-1.12712398e-01 1.31438076e-02 6.22550726e-01 -8.33205044e-01
8.84747684e-01 -8.78214419e-01 5.98424137e-01 1.09330513e-01
-9.21480805e-02 1.52939036e-01 2.60701329e-01 6.64017141e-01
2.14830920e-01 1.03940569e-01 5.22107303e-01 -3.83247375e-01
-1.05336511e+00 3.75384271e-01 -1.01720095e-01 3.71017963e-01
9.47840869e-01 -5.43729484e-01 -3.36050600e-01 -3.83548230e-01
-5.89151144e-01 -5.50184958e-02 6.86713398e-01 5.91933966e-01
5.68308115e-01 -9.28416848e-01 -5.07262051e-01 1.62254021e-01
-2.12678373e-01 4.20886666e-01 5.74445724e-01 7.89646447e-01
-1.17175269e+00 1.67788848e-01 -3.07971805e-01 -8.75309765e-01
-1.28269386e+00 4.79166985e-01 3.64389569e-01 2.13769302e-01
-1.16628683e+00 9.04544473e-01 3.30060333e-01 -3.24167937e-01
3.55937451e-01 -6.71170175e-01 -1.43930882e-01 2.29414046e-01
3.30192685e-01 5.94674468e-01 4.17214483e-02 -8.76569510e-01
4.82385419e-02 8.46813142e-01 3.59261841e-01 -1.86379120e-01
1.52629733e+00 2.59444993e-02 1.49917051e-01 2.96383649e-01
1.13586295e+00 -2.82234192e-01 -2.13972974e+00 1.38535425e-01
-2.06983536e-01 -5.73193491e-01 1.28686577e-01 -7.46224463e-01
-1.42694676e+00 1.06588745e+00 4.18374538e-01 -1.88592404e-01
9.07108545e-01 -1.37483496e-02 9.40798044e-01 1.52576774e-01
2.59536654e-01 -1.13347948e+00 1.63081959e-01 3.39118958e-01
9.35215831e-01 -1.23258209e+00 1.14821829e-01 -3.90950233e-01
-7.26318598e-01 1.15969872e+00 3.95162016e-01 -2.77481437e-01
1.94075763e-01 2.15947062e-01 1.93158332e-02 -2.49633402e-01
-3.32413971e-01 -7.12031648e-02 2.12379843e-01 8.99022281e-01
2.12143987e-01 -9.56182331e-02 4.79731977e-01 2.26122662e-01
-5.10447085e-01 -7.84123614e-02 4.94370103e-01 6.19771719e-01
1.92551628e-01 -8.26582968e-01 -2.64940262e-01 8.91927704e-02
-5.08474529e-01 2.42559779e-02 -3.34122598e-01 1.21106422e+00
3.14110905e-01 7.62056172e-01 2.90384889e-01 -3.84942204e-01
2.88276613e-01 -8.40555876e-02 7.44560778e-01 -3.14242661e-01
-5.79733133e-01 5.89647070e-02 4.01480615e-01 -1.08957791e+00
-6.51994467e-01 -8.43340158e-01 -9.44577396e-01 -4.76612031e-01
8.23972523e-02 -4.64554518e-01 1.04157603e+00 8.37739706e-01
1.81147054e-01 8.07347596e-01 4.91811670e-02 -1.72807646e+00
-7.19180107e-02 -5.01309514e-01 -2.38422111e-01 5.97205222e-01
2.56844372e-01 -6.36285424e-01 -1.94883525e-01 6.52689219e-01] | [8.547784805297852, -1.8715797662734985] |
1ecb441a-3f96-47d7-bdbf-e2de5f8b551b | complex-logical-reasoning-over-knowledge | 2305.01157 | null | https://arxiv.org/abs/2305.01157v2 | https://arxiv.org/pdf/2305.01157v2.pdf | Complex Logical Reasoning over Knowledge Graphs using Large Language Models | Reasoning over knowledge graphs (KGs) is a challenging task that requires a deep understanding of the complex relationships between entities and the underlying logic of their relations. Current approaches rely on learning geometries to embed entities in vector space for logical query operations, but they suffer from subpar performance on complex queries and dataset-specific representations. In this paper, we propose a novel decoupled approach, Language-guided Abstract Reasoning over Knowledge graphs (LARK), that formulates complex KG reasoning as a combination of contextual KG search and logical query reasoning, to leverage the strengths of graph extraction algorithms and large language models (LLM), respectively. Our experiments demonstrate that the proposed approach outperforms state-of-the-art KG reasoning methods on standard benchmark datasets across several logical query constructs, with significant performance gain for queries of higher complexity. Furthermore, we show that the performance of our approach improves proportionally to the increase in size of the underlying LLM, enabling the integration of the latest advancements in LLMs for logical reasoning over KGs. Our work presents a new direction for addressing the challenges of complex KG reasoning and paves the way for future research in this area. | ['Chandan K. Reddy', 'Nurendra Choudhary'] | 2023-05-02 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [-1.27250567e-01 5.29574335e-01 -4.92021233e-01 -3.35607857e-01
-6.54809654e-01 -6.40226543e-01 6.71617568e-01 8.81095290e-01
-7.36330748e-02 3.80820423e-01 2.04271764e-01 -7.41923988e-01
-4.90759343e-01 -1.55053055e+00 -9.57794726e-01 1.43628135e-01
-3.12694967e-01 6.67936146e-01 6.01989567e-01 -4.12627727e-01
1.09488713e-02 5.50769329e-01 -1.35288250e+00 3.51233959e-01
7.39556551e-01 1.21100760e+00 -6.05246305e-01 3.98637921e-01
-5.89528799e-01 1.53086329e+00 -2.88781971e-01 -9.38497007e-01
5.22902161e-02 1.69268861e-01 -1.18679714e+00 -5.38664639e-01
6.68724358e-01 -3.34620625e-02 -5.32356501e-01 9.08764303e-01
2.81163119e-02 -5.08694798e-02 4.83428419e-01 -1.45594811e+00
-5.77064753e-01 1.06766570e+00 -1.36536360e-01 -1.46991596e-01
6.23365283e-01 -2.94579029e-01 1.61049449e+00 -8.78935874e-01
5.86541891e-01 1.48238087e+00 5.71358740e-01 -8.00993592e-02
-1.11318517e+00 -3.98322821e-01 3.24230731e-01 5.18761933e-01
-1.68619967e+00 -2.04093337e-01 4.59058166e-01 -2.43651465e-01
1.57290971e+00 3.34129542e-01 5.15739024e-01 2.78224707e-01
5.66258170e-02 8.26930106e-01 6.46911204e-01 -6.03407681e-01
3.35875094e-01 1.92589328e-01 5.12129128e-01 1.21666408e+00
9.10827160e-01 -3.44968289e-01 -7.34469056e-01 -4.35643852e-01
4.67146516e-01 -2.24410951e-01 1.38734914e-02 -8.63571048e-01
-1.04260874e+00 9.78157461e-01 5.56980073e-01 1.23320609e-01
-3.31670076e-01 5.03870428e-01 5.45202494e-01 3.45043689e-01
5.17894365e-02 6.65299654e-01 -6.40987694e-01 1.50863916e-01
-7.62377739e-01 4.95439976e-01 1.49066436e+00 1.20868063e+00
7.75641382e-01 -1.96868703e-01 -1.75228462e-01 2.92318851e-01
4.93604153e-01 5.70040703e-01 -1.86010331e-01 -7.13622630e-01
8.06828856e-01 1.38944221e+00 -1.72165141e-01 -1.24319351e+00
-3.45110148e-01 -2.68254369e-01 -3.60493988e-01 -1.61204264e-01
1.51808634e-01 2.33943656e-01 -5.00614882e-01 1.52154434e+00
3.88991714e-01 -9.36458632e-02 6.03841126e-01 2.85765976e-01
9.52824652e-01 3.85390788e-01 2.19753936e-01 1.50540859e-01
1.58548152e+00 -8.24125171e-01 -5.54163098e-01 -1.19732119e-01
1.02260900e+00 -1.24180935e-01 9.39305365e-01 2.58253336e-01
-1.14506221e+00 -4.65316102e-02 -1.18758178e+00 -2.52248913e-01
-1.02937281e+00 -9.96721759e-02 1.37315345e+00 5.54772139e-01
-1.00494540e+00 2.31382623e-01 -7.20057905e-01 -4.21949267e-01
4.89451885e-01 5.27493179e-01 -3.70888114e-01 -3.83759290e-01
-1.50942445e+00 8.83609354e-01 9.85110581e-01 2.21597403e-02
-3.50742191e-01 -9.17132139e-01 -1.31507707e+00 2.98566878e-01
1.22072053e+00 -8.86528909e-01 1.04359460e+00 3.16931576e-01
-9.53571081e-01 5.80366135e-01 1.47182971e-01 -7.87086725e-01
2.39045814e-01 -4.51003611e-01 -5.18071413e-01 8.89299065e-02
-1.44534707e-02 1.83033377e-01 3.33742976e-01 -1.03521442e+00
-4.79182363e-01 -4.36427355e-01 8.37152421e-01 1.75570548e-02
-2.28339389e-01 -2.94711649e-01 -7.71470606e-01 -2.10854828e-01
6.30652234e-02 -5.78939557e-01 -2.00889394e-01 -2.88510174e-01
-6.41871750e-01 -5.13647258e-01 5.27844608e-01 -2.28073925e-01
1.63322115e+00 -1.84391427e+00 2.84189731e-01 5.93602896e-01
4.96444702e-01 2.12362334e-01 2.53415316e-01 1.04928136e+00
2.40292698e-01 2.29374930e-01 1.23726716e-02 1.24195017e-01
5.44292390e-01 4.27350849e-01 -7.98780739e-01 -6.43798783e-02
2.58394629e-01 1.51429975e+00 -9.48794961e-01 -7.85613358e-01
3.02976519e-02 -3.27005200e-02 -6.15914762e-01 6.90946132e-02
-8.25269699e-01 -6.21581256e-01 -5.66232800e-01 9.26053524e-01
5.32967031e-01 -5.54533064e-01 6.23206377e-01 -5.64897239e-01
2.90700614e-01 2.67016828e-01 -1.46040940e+00 1.69834566e+00
-5.02625465e-01 8.31705332e-02 -2.38660827e-01 -9.27977383e-01
7.01894343e-01 6.03007525e-02 2.48946652e-01 -5.12026012e-01
-2.61914432e-01 2.23979935e-01 -4.22631115e-01 -4.38966870e-01
5.75171590e-01 8.51105750e-02 -4.69827503e-01 3.31304461e-01
6.27820119e-02 -4.36033487e-01 5.23390174e-01 7.05960691e-01
1.26792538e+00 1.32630402e-02 6.07151926e-01 -2.26063794e-03
6.64078057e-01 2.10269630e-01 1.93188116e-01 1.05534887e+00
5.56665778e-01 -4.85662162e-01 9.14054513e-01 -5.13584971e-01
-3.30432951e-01 -1.19416070e+00 1.43657967e-01 8.90284836e-01
1.48320422e-01 -1.31645811e+00 -3.04987371e-01 -9.57694948e-01
5.54207027e-01 8.07453990e-01 -3.21440130e-01 -3.12175721e-01
-5.90137661e-01 -4.79271084e-01 1.02936220e+00 8.94271255e-01
4.12562579e-01 -7.73411334e-01 -5.64272523e-01 1.18471608e-01
1.60992116e-01 -1.63929093e+00 4.43061799e-01 1.71148423e-02
-8.08894932e-01 -1.49304414e+00 2.31450006e-01 -4.97568816e-01
5.38404346e-01 -5.04869446e-02 1.42421210e+00 2.69700825e-01
-2.32171580e-01 7.74902582e-01 -2.81100482e-01 -4.97292697e-01
-2.44703889e-01 2.32324585e-01 -1.48402229e-01 -3.01912189e-01
4.40827549e-01 -4.42144454e-01 -1.01945952e-01 -1.23045973e-01
-1.18057120e+00 3.17272358e-02 6.68901920e-01 5.20461380e-01
6.30245745e-01 7.49794245e-01 1.92975640e-01 -1.43060327e+00
7.24933922e-01 -4.85339195e-01 -9.74528313e-01 8.53762150e-01
-9.87418950e-01 7.25759327e-01 5.22349834e-01 7.72724673e-02
-8.08070838e-01 -3.72101933e-01 1.99006170e-01 -1.66071713e-01
3.19669902e-01 1.26440561e+00 -2.45438769e-01 -2.68565238e-01
4.75244045e-01 -4.50437143e-03 -5.69176495e-01 -2.30223984e-01
9.40121055e-01 1.13937864e-02 5.55089235e-01 -1.20321131e+00
1.11923099e+00 4.22875077e-01 6.19267285e-01 -3.72824252e-01
-7.92679548e-01 -3.56658667e-01 -4.46972638e-01 2.46476084e-01
5.49611568e-01 -9.08042371e-01 -1.05214942e+00 6.46575689e-02
-1.00061476e+00 -2.02079222e-01 -4.22543377e-01 2.68799394e-01
-5.12816131e-01 4.38375592e-01 -4.52344060e-01 -6.83908105e-01
-3.37548614e-01 -9.52462196e-01 1.14610398e+00 -1.00657083e-01
-2.16629177e-01 -1.15406191e+00 -4.98799570e-02 4.71161664e-01
3.32740068e-01 2.74634033e-01 1.76433420e+00 -7.99598932e-01
-1.20989597e+00 -5.40964961e-01 -5.02084315e-01 -9.50550567e-03
-3.75930071e-02 -2.24298537e-01 -5.18329740e-01 -9.70031023e-02
-5.55535913e-01 -5.26364982e-01 6.56649768e-01 -3.44011068e-01
9.55582976e-01 -4.87466693e-01 -4.93147224e-01 4.44846213e-01
1.82090557e+00 -2.24737570e-01 5.71455657e-01 2.08764166e-01
9.11728621e-01 4.61382359e-01 7.32778251e-01 3.42627674e-01
8.31252158e-01 6.08574450e-01 4.73869354e-01 1.45316064e-01
2.04249900e-02 -6.93533182e-01 -1.48194637e-02 5.06786108e-01
-5.06592169e-02 -1.51745126e-01 -1.32209873e+00 4.28962916e-01
-2.13327932e+00 -7.00914502e-01 -9.69523340e-02 1.96194887e+00
8.04149032e-01 3.26518327e-01 -2.22499385e-01 2.55524963e-01
-1.32521719e-01 1.88386470e-01 -4.93137300e-01 -4.82885152e-01
-6.03485852e-02 5.57482123e-01 6.81032062e-01 6.40666783e-01
-9.28459227e-01 1.37247550e+00 6.34840441e+00 5.83909452e-01
-6.89464986e-01 -2.70762622e-01 -4.15440053e-01 1.91126004e-01
-5.99174798e-01 3.82847100e-01 -9.92380261e-01 -3.46217602e-01
9.17480886e-01 -3.90224993e-01 5.70054591e-01 9.81364250e-01
-7.10051119e-01 -1.03784405e-01 -1.41376078e+00 8.91851187e-01
-2.65789335e-03 -1.65744150e+00 4.85510230e-01 -2.71207206e-02
3.04786056e-01 -2.76999176e-01 -2.61070043e-01 7.66292453e-01
7.11878419e-01 -1.06919181e+00 4.82294500e-01 6.63202524e-01
4.81433332e-01 -5.68803251e-01 7.71593213e-01 9.15077701e-02
-1.65500140e+00 -1.10850126e-01 -1.36348158e-01 -7.38059822e-03
-1.33315906e-01 4.12016690e-01 -1.11875534e+00 1.37691045e+00
4.59704041e-01 5.78667164e-01 -8.89775038e-01 4.31639314e-01
-4.67372656e-01 3.11171472e-01 -3.62838864e-01 -3.96770500e-02
9.39914212e-02 1.04294330e-01 1.55515820e-01 1.33350062e+00
-1.61593750e-01 -8.91337916e-03 2.46400148e-01 8.38836193e-01
-2.88552761e-01 2.59869337e-01 -7.78592885e-01 -5.18212140e-01
5.46336234e-01 9.87171590e-01 -4.22076613e-01 -6.33311927e-01
-7.27097511e-01 4.73313153e-01 7.48205483e-01 5.05154014e-01
-7.60032713e-01 -5.02808034e-01 4.68207330e-01 2.04090886e-02
4.51718122e-01 -3.08020890e-01 -8.89793932e-02 -1.31681049e+00
5.06072104e-01 -8.81786823e-01 1.07142293e+00 -6.45591915e-01
-1.08050978e+00 3.08504134e-01 5.89403808e-01 -5.74095130e-01
-4.11840796e-01 -6.91519320e-01 -1.71693832e-01 6.62331462e-01
-1.74827647e+00 -1.48254478e+00 -2.54951209e-01 6.82819605e-01
-9.33296159e-02 5.23711555e-02 1.11152756e+00 1.45884290e-01
-1.91065967e-01 5.98067284e-01 -4.44244176e-01 2.35499322e-01
2.45473936e-01 -1.40603101e+00 4.18134630e-01 6.47515297e-01
4.54220653e-01 9.63889718e-01 3.17616314e-01 -6.93940282e-01
-2.22941780e+00 -1.08507025e+00 1.08838189e+00 -5.75603306e-01
1.02733171e+00 -5.08010268e-01 -8.70543659e-01 1.16825581e+00
-1.18451014e-01 2.15608746e-01 7.58495569e-01 6.20985091e-01
-9.86831844e-01 -3.29929858e-01 -9.13151562e-01 7.12750733e-01
1.07665169e+00 -7.89333344e-01 -7.66632140e-01 2.97751367e-01
1.12490785e+00 -4.64097649e-01 -1.47323704e+00 7.82633781e-01
5.60466409e-01 -5.92409134e-01 1.25505316e+00 -1.12048471e+00
7.33354613e-02 -5.40794253e-01 -6.42309546e-01 -7.39071786e-01
-2.57927645e-02 -3.04831058e-01 -9.39074814e-01 1.01607502e+00
4.39315408e-01 -6.90478504e-01 7.33579516e-01 8.20562601e-01
2.98225105e-01 -1.21380067e+00 -5.20649672e-01 -7.08613753e-01
-2.18466848e-01 -8.25503349e-01 1.02114880e+00 7.80303180e-01
3.40149611e-01 4.36599672e-01 1.89151451e-01 6.80959046e-01
5.90085208e-01 5.55854917e-01 9.83849585e-01 -1.33290613e+00
-3.28005284e-01 -2.72704959e-01 -9.42013144e-01 -7.58803070e-01
3.83183926e-01 -1.19870400e+00 -7.09681809e-01 -2.05832577e+00
-1.67056635e-01 -4.63558316e-01 -3.38891834e-01 8.76528919e-01
8.00301731e-02 -3.10278833e-01 2.07981974e-01 -1.13530338e-01
-9.47586656e-01 3.20199043e-01 7.35574067e-01 -5.32851815e-01
-7.55592138e-02 -2.52865613e-01 -9.15251374e-01 5.96549749e-01
3.16497803e-01 -2.23661721e-01 -9.25344467e-01 -3.37763101e-01
9.32817042e-01 1.95511803e-01 4.97831732e-01 -1.01431453e+00
6.70561552e-01 -2.88558960e-01 -1.89340681e-01 -7.06422806e-01
3.86198252e-01 -8.71739447e-01 1.42753899e-01 3.04003209e-01
-3.66830796e-01 8.92432109e-02 4.18365866e-01 6.18146658e-01
-4.98395860e-01 8.69227424e-02 1.94360968e-02 -8.70869160e-02
-1.02810609e+00 2.90035427e-01 2.03094244e-01 4.08161461e-01
9.03375208e-01 2.22418919e-01 -4.72106159e-01 -1.41548678e-01
-5.80198348e-01 5.35848916e-01 2.24368334e-01 2.95442253e-01
6.85606956e-01 -1.29061174e+00 -3.65648448e-01 -2.96302810e-02
5.60226381e-01 3.25975865e-01 -1.49142712e-01 7.41052866e-01
-6.29053235e-01 9.48836982e-01 3.77951682e-01 -1.85257643e-01
-1.04760528e+00 8.61591458e-01 1.65282935e-01 -8.48766029e-01
-5.39939344e-01 6.78890586e-01 -3.48475054e-02 -4.32076633e-01
4.86050904e-01 -7.96248555e-01 -4.94088121e-02 -1.09693520e-01
3.12298626e-01 2.76686013e-01 3.65520418e-01 -1.44059733e-01
-6.65644705e-01 4.77887779e-01 -2.35691667e-01 2.57998914e-01
1.22429919e+00 3.92754257e-01 -5.45544088e-01 3.84108305e-01
8.56046319e-01 2.72399992e-01 -2.01131850e-01 -6.95378542e-01
5.01442134e-01 -3.22687864e-01 -2.08036333e-01 -8.65219593e-01
-7.78166652e-01 4.93723661e-01 -1.85079828e-01 2.87133992e-01
9.67850685e-01 5.82589388e-01 5.92668831e-01 1.08731222e+00
7.55506396e-01 -6.92284703e-01 -2.20409676e-01 5.29163122e-01
7.04890549e-01 -9.20058310e-01 3.89793098e-01 -8.82212758e-01
-3.46766561e-01 1.10578179e+00 5.14813244e-01 1.31042466e-01
6.17923737e-01 2.43886739e-01 -1.97844908e-01 -9.36119020e-01
-9.17427540e-01 -2.92767316e-01 3.31686080e-01 3.86748284e-01
7.76525363e-02 1.92588612e-01 1.40425125e-02 5.61221540e-01
-2.69461989e-01 2.33059928e-01 3.32085192e-02 1.30035031e+00
-8.82581770e-02 -1.39778471e+00 -9.25785452e-02 4.40387189e-01
-1.99816093e-01 -4.07139659e-01 -4.77242082e-01 1.23966074e+00
-1.28755897e-01 8.00212860e-01 -4.68512863e-01 -3.62599254e-01
7.77694583e-01 3.28109443e-01 7.41477847e-01 -7.01153815e-01
-1.30027205e-01 -8.72467458e-01 4.68287319e-01 -7.65879810e-01
-1.33254081e-01 -2.30618343e-01 -1.59705210e+00 -2.40644053e-01
-1.90115392e-01 3.25470567e-01 3.87998730e-01 8.06532323e-01
4.93642002e-01 4.72405553e-01 -2.44843159e-02 2.09186628e-01
-6.48882151e-01 -3.35585475e-01 -6.12412870e-01 3.19322914e-01
9.94863827e-03 -8.12861502e-01 1.63903460e-01 -3.10765266e-01] | [9.1358003616333, 7.695498466491699] |
5598e8cb-fe55-401b-b1f4-48d56d02c991 | inducing-point-allocation-for-sparse-gaussian | 2301.10123 | null | https://arxiv.org/abs/2301.10123v2 | https://arxiv.org/pdf/2301.10123v2.pdf | Inducing Point Allocation for Sparse Gaussian Processes in High-Throughput Bayesian Optimisation | Sparse Gaussian Processes are a key component of high-throughput Bayesian Optimisation (BO) loops; however, we show that existing methods for allocating their inducing points severely hamper optimisation performance. By exploiting the quality-diversity decomposition of Determinantal Point Processes, we propose the first inducing point allocation strategy designed specifically for use in BO. Unlike existing methods which seek only to reduce global uncertainty in the objective function, our approach provides the local high-fidelity modelling of promising regions required for precise optimisation. More generally, we demonstrate that our proposed framework provides a flexible way to allocate modelling capacity in sparse models and so is suitable broad range of downstream sequential decision making tasks. | ['Victor Picheny', 'Sebastian W. Ober', 'Henry B. Moss'] | 2023-01-24 | null | null | null | null | ['point-processes', 'bayesian-optimisation'] | ['methodology', 'methodology'] | [ 3.33327740e-01 2.59894967e-01 4.83261757e-02 -1.26338884e-01
-1.21753860e+00 -5.01008570e-01 8.66429567e-01 5.59328385e-02
-2.67850637e-01 6.70404494e-01 1.85027674e-01 -4.73211378e-01
-7.75031090e-01 -5.19727886e-01 -5.83700120e-01 -1.19131088e+00
-1.12446018e-01 8.83682072e-01 1.55538157e-01 1.21274628e-01
4.83188212e-01 6.18295312e-01 -1.12520862e+00 6.55544698e-02
6.68455958e-01 9.44295168e-01 3.57434660e-01 9.80664253e-01
1.24823198e-01 4.83318329e-01 -4.19253469e-01 -3.18187743e-01
5.51224232e-01 -2.57046580e-01 -4.58948255e-01 -1.15331873e-01
-1.84193850e-01 1.50753841e-01 -9.92882773e-02 6.92591786e-01
9.11450446e-01 5.38690865e-01 1.11096382e+00 -7.57226825e-01
2.56614506e-01 6.15138113e-01 -6.19440913e-01 4.89070088e-01
-2.11427927e-01 4.64529961e-01 1.01086915e+00 -7.98353612e-01
1.79210246e-01 1.52168298e+00 6.61163151e-01 2.33066410e-01
-1.50883412e+00 -1.71308607e-01 2.68078685e-01 -2.74837315e-01
-1.30898714e+00 -1.00405800e+00 4.37087595e-01 -4.67431545e-01
9.78058219e-01 4.65742290e-01 4.18868661e-01 8.51730108e-01
6.90629125e-01 7.09062338e-01 1.07339489e+00 -3.04764956e-01
6.37254238e-01 -4.23879802e-01 -1.61590979e-01 3.14509302e-01
-6.50641397e-02 4.44996238e-01 -8.28668416e-01 -6.22599423e-01
9.21601832e-01 -3.20659548e-01 -1.20552100e-01 -6.67468131e-01
-1.31659532e+00 8.22007298e-01 1.06758639e-01 -2.84016460e-01
-5.83418906e-01 6.00452721e-01 2.59323239e-01 -1.10653706e-01
6.24548495e-01 5.91968596e-01 -5.50696015e-01 -5.69170177e-01
-1.17891359e+00 4.04954284e-01 9.30745184e-01 1.06526518e+00
4.22546893e-01 -2.06676617e-01 -7.68705845e-01 6.59352481e-01
9.43989992e-01 3.69364619e-01 -8.74008015e-02 -1.12857974e+00
6.57863140e-01 -1.58274963e-01 1.21401235e-01 -7.81510115e-01
-4.15668964e-01 -9.68654990e-01 -1.10005724e+00 -1.62178978e-01
3.93587828e-01 -3.19570512e-01 -8.61393929e-01 1.29948807e+00
6.70857489e-01 1.24302192e-03 2.37350888e-03 6.94144726e-01
1.41661078e-01 7.08077133e-01 8.43807682e-02 -3.80248785e-01
1.21827888e+00 -8.89973700e-01 -4.66599286e-01 -2.95933515e-01
3.70412707e-01 -8.43950570e-01 4.76136595e-01 6.44584656e-01
-1.26446295e+00 -1.86207741e-01 -4.49514717e-01 3.20916712e-01
2.69552916e-01 -1.90076023e-01 9.76898909e-01 8.65349770e-01
-1.14279222e+00 6.83129847e-01 -1.03243732e+00 1.08402312e-01
6.10667229e-01 5.01448810e-01 2.86140621e-01 9.72646400e-02
-7.00613976e-01 8.27906191e-01 3.11383039e-01 6.16143346e-01
-1.28563142e+00 -8.85883689e-01 -5.04679501e-01 2.57089227e-01
4.22617763e-01 -9.90534365e-01 1.34477735e+00 -1.06145188e-01
-1.98192298e+00 1.75418362e-01 -3.72115672e-01 -5.37657320e-01
6.51116669e-01 -8.74153301e-02 2.77523160e-01 -2.01487076e-02
-2.04836294e-01 6.31236017e-01 1.32247424e+00 -1.05240142e+00
-6.69540405e-01 -2.68976599e-01 -2.58084506e-01 4.91249442e-01
2.65282333e-01 2.04647034e-01 -5.72957397e-01 -5.83824873e-01
4.21193510e-01 -7.99650431e-01 -1.06851590e+00 -3.53931844e-01
-4.09262478e-01 5.52502014e-02 8.47580731e-02 -4.77230638e-01
1.20819616e+00 -1.87694299e+00 5.20438194e-01 7.64303267e-01
3.20473284e-01 -2.03352094e-01 2.47559309e-01 4.08912778e-01
2.98301518e-01 7.19623789e-02 -4.99184579e-01 -4.86938477e-01
3.68057787e-01 3.22822690e-01 -1.75077841e-01 6.52996540e-01
3.90759915e-01 8.59969854e-01 -1.04268920e+00 -4.09568042e-01
2.35414952e-01 2.96003342e-01 -6.51161671e-01 2.98492834e-02
-2.45248616e-01 7.33946919e-01 -7.13370144e-01 6.56409383e-01
5.75849593e-01 -1.23726971e-01 -5.87417297e-02 2.55483627e-01
-1.86277423e-02 2.62399346e-01 -1.41246557e+00 1.53651786e+00
-4.13706869e-01 3.08952957e-01 5.94052374e-01 -9.54210460e-01
8.17545414e-01 1.80267394e-01 5.79773426e-01 -2.74071619e-02
2.69079749e-02 -1.08382944e-02 1.80478022e-01 2.70191312e-01
4.16207850e-01 -4.24242735e-01 -1.32179022e-01 1.20705165e-01
-8.50065425e-02 -7.65557945e-01 2.41917232e-03 -2.72487011e-02
1.32990777e+00 2.94459015e-01 3.64165843e-01 -7.07457542e-01
2.30555236e-01 -1.71135783e-01 5.62819302e-01 1.36686456e+00
-2.84806192e-01 6.45595074e-01 7.31026232e-01 -5.75793674e-03
-1.03155708e+00 -9.76121008e-01 -5.20609200e-01 1.22564995e+00
-3.08347464e-01 -3.50148946e-01 -4.03096408e-01 -3.88619602e-01
2.44626012e-02 6.12792552e-01 -2.90776581e-01 5.83282560e-02
-4.55408216e-01 -1.37526166e+00 4.09635186e-01 2.97747582e-01
-1.13549652e-02 -6.39812231e-01 -4.26142722e-01 5.40017903e-01
3.03202927e-01 -5.84962189e-01 -4.50429052e-01 8.31823051e-01
-1.15086234e+00 -5.86177111e-01 -8.74325633e-01 -1.20542929e-01
4.19683397e-01 9.23804659e-03 1.21510088e+00 -6.88559115e-01
1.02184854e-01 4.03034031e-01 5.69989160e-02 -5.88855863e-01
-3.52607310e-01 1.51720718e-01 -1.61005437e-01 -1.59441710e-01
-5.72183877e-02 -4.14531648e-01 -5.78914642e-01 3.93570960e-01
-6.13184154e-01 -1.49843479e-02 8.02466631e-01 9.74685550e-01
6.51019692e-01 1.62674263e-01 3.48541945e-01 -6.31890237e-01
8.86798918e-01 -5.98363698e-01 -9.01661396e-01 -5.90935610e-02
-5.63156486e-01 3.54942173e-01 1.54673710e-01 -1.92860469e-01
-1.25663495e+00 8.87288600e-02 -1.84481680e-01 -2.88856924e-01
-1.19229905e-01 5.00781894e-01 -5.88342845e-02 -1.17707700e-01
5.52397668e-01 -9.44795758e-02 -1.07863382e-01 -4.67307925e-01
5.14669120e-01 3.45514566e-01 1.22509435e-01 -1.00205112e+00
4.16203856e-01 3.53784680e-01 6.19898140e-01 -5.37273049e-01
-5.02815902e-01 -8.74581099e-01 -6.77690566e-01 -1.84471712e-01
4.51539993e-01 -9.85018194e-01 -7.91541934e-01 2.21027136e-01
-1.12401021e+00 -4.16490644e-01 -3.81521165e-01 3.50048482e-01
-1.02845836e+00 2.87086666e-01 -3.80186617e-01 -1.33169174e+00
-3.01303744e-01 -1.44168091e+00 1.60135520e+00 -2.28836954e-01
-1.56822950e-01 -1.09855354e+00 7.11178929e-02 1.20700657e-01
4.07532305e-01 4.89426032e-02 5.42025149e-01 -3.75661522e-01
-7.45337188e-01 1.15000680e-01 -2.26729922e-02 -5.66560915e-03
-2.45500043e-01 -4.57265079e-02 -8.45638573e-01 -1.74053982e-01
2.15238854e-01 1.92682371e-01 9.97819364e-01 9.69600499e-01
1.08735001e+00 -1.63093165e-01 -5.23664355e-01 7.81460822e-01
1.04395008e+00 1.34773285e-03 5.13229549e-01 3.02945167e-01
6.68117046e-01 6.56226635e-01 6.83605015e-01 7.59259224e-01
8.16825032e-02 7.54627347e-01 2.17344150e-01 5.68557531e-02
3.49480182e-01 4.12079971e-03 3.47039014e-01 6.61049008e-01
-1.45693228e-01 -2.19069570e-01 -1.17273247e+00 4.81647879e-01
-2.20487189e+00 -7.16395319e-01 -2.46822104e-01 2.25335407e+00
8.95677149e-01 1.85808793e-01 -1.54197002e-02 -1.45114079e-01
6.16300106e-01 1.23619281e-01 -4.71160978e-01 -6.85631990e-01
2.04040587e-01 2.85674751e-01 1.08783817e+00 6.26397550e-01
-1.11814904e+00 7.62375116e-01 8.02787399e+00 1.36761415e+00
-2.44431928e-01 2.75246143e-01 7.03951299e-01 -5.25094807e-01
-1.76640779e-01 1.38310745e-01 -1.21925402e+00 4.93059725e-01
1.35483384e+00 1.42772391e-01 3.62519950e-01 6.54019296e-01
8.01735401e-01 -2.93742359e-01 -1.06817985e+00 6.96272552e-01
-5.88774025e-01 -1.17446256e+00 -2.48192117e-01 4.54486936e-01
8.68280709e-01 2.50259880e-02 9.83775184e-02 2.40225881e-01
9.33797061e-01 -1.23596632e+00 6.93025768e-01 8.32308531e-01
2.47590154e-01 -9.44476545e-01 6.73910677e-01 4.91725445e-01
-7.50464439e-01 -8.75419229e-02 -5.68373382e-01 7.39376098e-02
6.03120506e-01 1.11051762e+00 -1.01416075e+00 5.88311493e-01
5.29427946e-01 3.01082164e-01 -1.33215621e-01 1.26518118e+00
-1.03969619e-01 7.51292825e-01 -7.69177854e-01 1.01284981e-01
5.48611641e-01 -4.24435854e-01 9.88342047e-01 1.29512846e+00
4.23972458e-01 -2.39687294e-01 -4.41864468e-02 6.99717462e-01
4.26767170e-01 -6.91054761e-02 -5.41522443e-01 1.88618973e-01
1.69285491e-01 9.31913912e-01 -6.97312832e-01 -3.67106833e-02
5.94459027e-02 4.61347014e-01 2.98825622e-01 4.05772001e-01
-6.37320220e-01 2.58345246e-01 7.57204890e-01 -1.92986831e-01
6.38982594e-01 -5.38199902e-01 -6.71459675e-01 -8.96418869e-01
-2.56292492e-01 -9.60822284e-01 3.45683306e-01 -4.46179152e-01
-1.31991696e+00 -1.15513943e-01 1.15448095e-01 -7.54780293e-01
-4.62176532e-01 -5.12343287e-01 -4.85816479e-01 1.33956289e+00
-1.33245647e+00 -9.42105174e-01 3.66050214e-01 1.64196312e-01
6.54224753e-01 -1.31378487e-01 4.64388579e-01 -1.96857721e-01
-4.92269814e-01 1.25069007e-01 7.86942661e-01 -7.65672743e-01
4.56531107e-01 -1.37860966e+00 5.31776249e-01 7.58482933e-01
1.06451276e-03 8.43946338e-01 1.01634395e+00 -6.50281966e-01
-1.54653430e+00 -8.77851188e-01 3.18093300e-01 -6.96966350e-01
7.93276846e-01 -4.00047183e-01 -5.34491003e-01 3.61593336e-01
-2.96677738e-01 -4.64876801e-01 6.17561698e-01 4.66548860e-01
5.19167125e-01 6.79472759e-02 -1.03714037e+00 5.59003532e-01
9.02438402e-01 -6.71946034e-02 -3.33950937e-01 4.78518575e-01
4.15026993e-01 -5.58116674e-01 -1.07156777e+00 5.14420509e-01
2.56331712e-01 -5.84291995e-01 1.15510857e+00 -5.68335772e-01
2.46936157e-01 -3.39741886e-01 -2.01246440e-01 -1.38516057e+00
-5.53670943e-01 -1.31637943e+00 -5.36886632e-01 1.11396503e+00
4.69118983e-01 -5.76131582e-01 8.51047933e-01 7.83792555e-01
-2.88345397e-01 -8.06781292e-01 -1.28987515e+00 -4.96723533e-01
2.20192969e-01 -7.93044567e-01 4.20529038e-01 2.13825941e-01
-3.51073056e-01 2.07290068e-01 -2.68641382e-01 4.22976017e-01
1.02111089e+00 -3.50093156e-01 7.24205136e-01 -1.02211142e+00
-8.15613925e-01 -7.02710450e-01 -2.19885126e-01 -1.54832709e+00
-1.73690066e-01 -5.57448864e-01 5.55752754e-01 -1.30635464e+00
5.31895868e-02 -9.83235657e-01 -2.54479736e-01 -1.71784803e-01
-3.52362573e-01 -1.35470092e-01 3.43191363e-02 3.89082521e-01
-6.03022158e-01 8.08713853e-01 1.30546880e+00 1.12779446e-01
-3.06971490e-01 5.15251517e-01 -8.85373533e-01 4.93617266e-01
5.90474665e-01 -5.83989680e-01 -2.32529163e-01 -2.75321275e-01
3.59230548e-01 3.11937872e-02 3.89496237e-01 -6.59110069e-01
3.63813818e-01 -5.26121736e-01 3.92568707e-01 -7.96950340e-01
4.09762114e-01 -6.42943203e-01 3.12183291e-01 1.83353070e-02
-2.52241254e-01 -3.29642296e-01 2.45566830e-01 1.16824532e+00
9.66757312e-02 -5.52428544e-01 6.61450446e-01 -2.57772088e-01
-2.88123339e-01 1.74537063e-01 -6.06421173e-01 -1.83305502e-01
8.61365318e-01 7.58930445e-02 1.77325398e-01 -3.05521280e-01
-9.39857125e-01 5.00288486e-01 2.83919871e-01 -1.72220260e-01
2.61867762e-01 -9.39324558e-01 -7.27613151e-01 -9.43008885e-02
-3.79985869e-01 7.25383162e-01 -4.37937044e-02 1.12508249e+00
-4.47308928e-01 6.64235234e-01 4.66459990e-01 -9.78958368e-01
-7.14173615e-01 3.59582335e-01 3.88707519e-01 -6.41866207e-01
-5.18029511e-01 1.08543348e+00 3.68238717e-01 -4.02863353e-01
8.55708569e-02 -2.81529933e-01 1.45609573e-01 -4.72679660e-02
2.75764972e-01 5.84623754e-01 2.64825076e-01 -3.71363401e-01
-1.43996611e-01 1.13949463e-01 2.38763288e-01 -6.48866475e-01
1.35803866e+00 -4.47453797e-01 -1.13538265e-01 4.19445693e-01
5.10830522e-01 -1.88688755e-01 -1.76257408e+00 -4.48978506e-02
1.81591496e-01 -5.30719399e-01 6.62845850e-01 -3.51842701e-01
-4.96893913e-01 9.20405686e-01 2.40091845e-01 6.85404241e-02
7.29677022e-01 -1.94590446e-02 6.38390109e-02 4.11190897e-01
4.68805820e-01 -1.03478897e+00 -5.49340367e-01 7.67557681e-01
8.65375340e-01 -9.13546026e-01 2.08820105e-01 -5.60621142e-01
-5.98978341e-01 7.92706490e-01 -3.76591906e-02 2.20737144e-01
6.00893259e-01 3.56874943e-01 -3.73063326e-01 -3.64851594e-01
-1.03532541e+00 -2.12268591e-01 3.52564007e-01 8.61773193e-01
1.91406161e-01 -1.41526177e-03 -1.54312104e-01 3.90812427e-01
-7.87366778e-02 -4.29734379e-01 1.04433596e-01 9.19620097e-01
-4.48535293e-01 -1.00137329e+00 -8.13350856e-01 8.08119476e-01
-5.14529169e-01 -5.00362515e-01 1.45700082e-01 2.06254125e-01
-3.88125807e-01 1.00970411e+00 -3.37691000e-03 3.20708245e-01
1.28296033e-01 6.50340989e-02 4.12533164e-01 -7.29923368e-01
-6.96509302e-01 5.73901534e-01 3.21424901e-01 -5.10202527e-01
-1.26248404e-01 -9.52646911e-01 -5.20021260e-01 -2.97573507e-01
-4.36163455e-01 1.43668637e-01 7.31593251e-01 1.05715942e+00
4.62126464e-01 7.62928903e-01 4.68907267e-01 -1.32326949e+00
-1.02997398e+00 -1.10233402e+00 -5.96757650e-01 -3.73955339e-01
1.91562846e-01 -8.85246754e-01 -4.44913238e-01 -9.31995735e-02] | [6.389586448669434, 3.8425145149230957] |
9dbe6d3d-c591-4f82-958a-563a7d6d064d | what-when-and-where-self-supervised-spatio | 2303.16990 | null | https://arxiv.org/abs/2303.16990v1 | https://arxiv.org/pdf/2303.16990v1.pdf | What, when, and where? -- Self-Supervised Spatio-Temporal Grounding in Untrimmed Multi-Action Videos from Narrated Instructions | Spatio-temporal grounding describes the task of localizing events in space and time, e.g., in video data, based on verbal descriptions only. Models for this task are usually trained with human-annotated sentences and bounding box supervision. This work addresses this task from a multimodal supervision perspective, proposing a framework for spatio-temporal action grounding trained on loose video and subtitle supervision only, without human annotation. To this end, we combine local representation learning, which focuses on leveraging fine-grained spatial information, with a global representation encoding that captures higher-level representations and incorporates both in a joint approach. To evaluate this challenging task in a real-life setting, a new benchmark dataset is proposed providing dense spatio-temporal grounding annotations in long, untrimmed, multi-action instructional videos for over 5K events. We evaluate the proposed approach and other methods on the proposed and standard downstream tasks showing that our method improves over current baselines in various settings, including spatial, temporal, and untrimmed multi-action spatio-temporal grounding. | ['Hilde Kuehne', 'James Glass', 'Rogerio Feris', 'Shih-Fu Chang', 'Samuel Thomas', 'Daniel Kondermann', 'Andrew Rouditchenko', 'Nina Shvetsova', 'Brian Chen'] | 2023-03-29 | null | null | null | null | ['spatio-temporal-video-grounding'] | ['computer-vision'] | [ 3.15182090e-01 -1.22716971e-01 -4.56172079e-01 -4.69100833e-01
-1.26638854e+00 -7.57366776e-01 7.62356937e-01 5.68965793e-01
-4.10249770e-01 6.07666254e-01 8.85367692e-01 -1.52521715e-01
-1.78983882e-01 -2.90663928e-01 -1.14844263e+00 -5.45664907e-01
-1.45600855e-01 9.63089690e-02 4.91912723e-01 -2.15901151e-01
1.97791487e-01 1.52707085e-01 -1.66629219e+00 9.53771770e-01
5.83477855e-01 7.86753118e-01 1.95036575e-01 8.64485741e-01
3.88951212e-01 1.43260348e+00 -4.96738076e-01 -8.55431706e-02
-1.17003381e-01 -4.37008440e-01 -1.09181201e+00 2.82040089e-01
1.26954234e+00 -4.72246081e-01 -6.07409716e-01 5.51372111e-01
2.32791558e-01 8.41478825e-01 3.36656243e-01 -1.37121105e+00
-4.62113410e-01 5.78127861e-01 -3.78290296e-01 3.86036277e-01
9.20948565e-01 1.81646928e-01 1.16471708e+00 -5.12234569e-01
8.48279357e-01 1.16136420e+00 5.44717610e-01 2.77685314e-01
-1.11159027e+00 -3.41966391e-01 6.31254852e-01 4.12793547e-01
-1.27658665e+00 -2.67312676e-01 4.52250063e-01 -6.30932570e-01
1.21083438e+00 1.02037132e-01 2.94619441e-01 1.62398231e+00
1.49613753e-01 1.01359749e+00 8.98110569e-01 -2.94913173e-01
2.88667213e-02 -2.61336058e-01 2.13382542e-01 6.81496382e-01
-2.43223652e-01 -2.90352732e-01 -1.21589208e+00 2.56293684e-01
8.55208933e-01 3.29957724e-01 -3.66609991e-01 -4.79903907e-01
-1.66661525e+00 5.79950333e-01 4.98764455e-01 4.09087360e-01
-2.14203209e-01 3.76963794e-01 7.02731609e-01 9.45430547e-02
5.77403426e-01 3.65179628e-01 -3.13533545e-01 -6.13798499e-01
-1.27517676e+00 3.26406360e-01 2.69648373e-01 1.13735640e+00
4.50818449e-01 -3.18915516e-01 -7.14359581e-01 4.21663553e-01
6.71915859e-02 6.69173449e-02 5.58226824e-01 -8.92698169e-01
1.06822777e+00 5.35810351e-01 1.77338883e-01 -8.28084290e-01
-4.06625718e-01 1.50635811e-02 -3.81483406e-01 -1.84905693e-01
5.35051882e-01 1.71731576e-01 -9.95000720e-01 1.79909301e+00
4.39502358e-01 1.05532062e+00 -9.19794813e-02 1.00136924e+00
1.12384510e+00 8.34181488e-01 3.85374933e-01 8.25194418e-02
1.33583999e+00 -1.56436861e+00 -9.40465689e-01 3.51425260e-02
1.08960617e+00 -4.36111659e-01 1.16942358e+00 3.07908386e-01
-1.05654633e+00 -7.18276620e-01 -9.90142524e-01 -4.65732604e-01
-7.68199384e-01 3.83426905e-01 1.84726238e-01 -7.95417577e-02
-1.08560753e+00 6.84298396e-01 -1.08857703e+00 -6.95730150e-01
1.70052052e-01 5.83088361e-02 -9.48902965e-01 -3.18617791e-01
-1.08634961e+00 6.96554601e-01 4.93926942e-01 -4.83513735e-02
-1.24459815e+00 -6.79590642e-01 -1.51048934e+00 -1.21902056e-01
6.70084357e-01 -3.61253679e-01 1.23240960e+00 -3.64211470e-01
-1.34669626e+00 7.14349568e-01 -1.43075138e-01 -5.19468844e-01
3.33018929e-01 -6.64300740e-01 -1.09139256e-01 4.54313368e-01
3.76565099e-01 5.98270655e-01 5.41620672e-01 -8.90863001e-01
-4.91272569e-01 -2.64917731e-01 5.77916086e-01 2.92886198e-01
-2.88930625e-01 1.01256870e-01 -3.98470193e-01 -7.30493605e-01
-1.49302319e-01 -8.23264718e-01 -1.47755682e-01 -3.79053473e-01
-4.79377687e-01 -3.39793205e-01 8.25962663e-01 -8.47137451e-01
1.31794894e+00 -2.24073076e+00 6.09589458e-01 -3.70005488e-01
-1.17900316e-02 -3.04496121e-02 -2.95752317e-01 6.83022201e-01
-1.96944565e-01 -1.41811997e-01 8.75876397e-02 -7.97387123e-01
2.25282937e-01 2.63356835e-01 -3.59760225e-01 5.23641944e-01
1.98337942e-01 8.96300733e-01 -1.30258894e+00 -7.04157233e-01
5.05339921e-01 5.45784354e-01 -5.06553113e-01 5.13976514e-01
-1.66013837e-01 7.32329190e-01 -2.24604338e-01 5.87414086e-01
-1.13302283e-02 -3.09623033e-01 1.61343738e-02 -1.87215820e-01
-2.43457660e-01 3.84736568e-01 -1.12857485e+00 2.61823297e+00
-5.45260131e-01 7.89389312e-01 -1.38744041e-01 -9.75600660e-01
3.68306637e-01 6.65018499e-01 6.25896931e-01 -6.29173636e-01
-2.56405652e-01 -2.07703054e-01 -7.70543516e-01 -8.00957918e-01
8.02131295e-01 2.57534236e-01 -4.97014463e-01 3.96112353e-01
7.71303236e-01 2.17704877e-01 2.94816822e-01 6.31965995e-01
1.38878846e+00 9.41785336e-01 3.18576992e-01 1.17439635e-01
3.12887371e-01 6.01873733e-02 1.56552449e-01 7.63006628e-01
-3.51028085e-01 8.43948305e-01 4.02723432e-01 -3.49398643e-01
-7.59688914e-01 -7.92120993e-01 9.88684371e-02 1.90332115e+00
2.07436472e-01 -9.98806596e-01 -7.05138624e-01 -1.00593626e+00
-2.84264266e-01 3.86264920e-01 -7.81232476e-01 -2.21001394e-02
-7.27609873e-01 4.58048806e-02 4.18473065e-01 9.04630840e-01
2.66234964e-01 -8.95158827e-01 -6.45113111e-01 1.40223429e-01
-4.87794518e-01 -1.75314462e+00 -5.92115939e-01 1.79268062e-01
-6.61611855e-01 -9.62721109e-01 -7.44926751e-01 -5.22315085e-01
3.58450800e-01 3.94857854e-01 1.19353461e+00 -1.45279288e-01
-1.92447379e-01 8.80185962e-01 -7.31828094e-01 1.24171123e-01
2.12237835e-01 9.10101458e-03 -2.25629538e-01 9.65121668e-04
1.24890320e-01 -2.83280790e-01 -4.45947975e-01 2.67221898e-01
-9.80746090e-01 2.25448146e-01 2.98743874e-01 7.38784373e-01
6.21153653e-01 -2.23189235e-01 2.41461977e-01 -5.66597223e-01
1.11477137e-01 -6.99941516e-01 -2.26066485e-01 3.74686718e-01
3.20699543e-01 -1.70368508e-01 3.67221445e-01 -4.32089090e-01
-1.04748642e+00 1.20052069e-01 2.56485134e-01 -7.37325490e-01
-6.70160055e-01 5.46782494e-01 -1.84140146e-01 1.21487632e-01
4.88190889e-01 5.88429123e-02 -5.58785975e-01 -5.40030956e-01
5.93617618e-01 3.99399072e-01 7.67605722e-01 -8.48766148e-01
4.50940907e-01 4.44510072e-01 1.03013068e-01 -7.04776168e-01
-1.21569514e+00 -1.03612936e+00 -1.37544072e+00 -3.76743644e-01
1.50256598e+00 -1.34059608e+00 -5.61229587e-01 3.53775211e-02
-1.14885175e+00 -7.12948442e-01 -2.50902504e-01 7.74521589e-01
-9.69368339e-01 2.64621556e-01 -7.50447154e-01 -4.63737845e-01
5.32481670e-01 -1.09244096e+00 1.95815206e+00 -7.99026340e-02
-3.17869633e-01 -1.15841949e+00 2.74166107e-01 7.79896319e-01
-1.48687035e-01 7.51445413e-01 3.47138792e-01 -7.98350930e-01
-7.80744612e-01 -1.78232834e-01 -6.06873184e-02 4.39043380e-02
-2.04587192e-03 -2.43359596e-01 -9.08049583e-01 -2.42370099e-01
-5.44099212e-01 -8.25141430e-01 1.02345765e+00 1.95309356e-01
1.22358382e+00 -3.04017454e-01 -1.46984488e-01 4.85948652e-01
1.23738253e+00 -1.98934972e-01 4.25145358e-01 4.22797471e-01
1.20573449e+00 7.52008080e-01 1.06700420e+00 4.92344916e-01
5.33521593e-01 1.12219346e+00 5.64433157e-01 -5.91527671e-03
-2.06090868e-01 -4.61107433e-01 5.36805093e-01 5.86283267e-01
-2.74316192e-01 -4.94971395e-01 -8.98515046e-01 9.31085169e-01
-2.50288010e+00 -1.36246395e+00 5.38764223e-02 1.99192953e+00
8.36072862e-01 -2.64203727e-01 3.56321573e-01 6.02418743e-02
5.49273372e-01 6.45874441e-01 -1.25989109e-01 -2.75479853e-01
1.92756817e-01 -3.00276745e-02 2.66127467e-01 4.11626190e-01
-1.73126698e+00 1.02116215e+00 5.59402227e+00 6.22194469e-01
-8.93474042e-01 3.63513827e-01 2.00912610e-01 -3.92581135e-01
-2.08560396e-02 -3.54539126e-01 -7.97415793e-01 2.70795047e-01
1.20720375e+00 1.87797859e-01 5.10327928e-02 5.92605174e-01
3.93484354e-01 1.79559253e-02 -1.61960423e+00 8.31591964e-01
4.47642058e-01 -1.34010160e+00 7.10299890e-03 -1.68149337e-01
9.87043142e-01 -2.41017878e-01 1.75636820e-02 6.10810697e-01
3.26900721e-01 -1.27774978e+00 9.30497110e-01 4.97538120e-01
5.95728457e-01 -3.74057591e-01 7.30278671e-01 2.01819941e-01
-1.66815853e+00 1.13904305e-01 3.21552038e-01 -1.61866337e-01
3.11181933e-01 -2.06736743e-01 -4.69058424e-01 8.86688232e-01
7.47170985e-01 1.49097693e+00 -6.03215098e-01 8.60187352e-01
-3.88983011e-01 5.08548319e-01 -9.29043293e-02 4.88650352e-01
9.43477511e-01 1.24908656e-01 2.84066588e-01 1.59389329e+00
2.72933185e-01 1.26408517e-01 6.48744643e-01 4.25266087e-01
9.85288024e-02 -3.17484662e-02 -8.70863020e-01 -1.08845040e-01
2.34101236e-01 1.07313335e+00 -6.42963648e-01 -4.45031077e-01
-6.57510757e-01 1.00314391e+00 3.27515304e-01 6.50370181e-01
-1.21058726e+00 -1.12947866e-01 5.95885396e-01 2.27776200e-01
3.33889455e-01 -5.56927204e-01 2.97772437e-01 -1.28310442e+00
-2.25587515e-03 -6.85734332e-01 8.28131974e-01 -1.08318472e+00
-9.28348362e-01 3.14623863e-01 5.69234192e-01 -1.38088048e+00
-4.39493656e-01 -7.14959025e-01 -5.30313134e-01 4.60190207e-01
-1.41740990e+00 -1.58326089e+00 -5.11566281e-01 9.53743339e-01
9.11258101e-01 1.71985418e-01 4.86589551e-01 4.06414658e-01
-8.07480931e-01 4.62145358e-01 -2.42136791e-01 1.45927370e-01
1.03371680e+00 -1.47511220e+00 -1.33901432e-01 8.86939049e-01
5.54017186e-01 5.87117374e-01 6.21048212e-01 -4.21769232e-01
-1.12821162e+00 -1.47742653e+00 9.68252838e-01 -8.74121428e-01
8.86471272e-01 -4.77145880e-01 -8.83804321e-01 1.33026350e+00
4.72436547e-01 3.89105976e-01 7.99852848e-01 1.60814553e-01
-4.69107568e-01 3.31771582e-01 -7.98532724e-01 4.59949106e-01
1.26402068e+00 -9.80654776e-01 -7.15733230e-01 6.48499787e-01
1.04096973e+00 -8.46348882e-01 -1.13883781e+00 2.51841843e-01
1.99990615e-01 -7.36891627e-01 1.27545023e+00 -1.06539559e+00
8.95626664e-01 -3.18440229e-01 -4.41395462e-01 -1.09209633e+00
-1.48765177e-01 -7.21150398e-01 -2.17733994e-01 1.25442398e+00
3.57126221e-02 2.39450261e-01 5.72090387e-01 2.87453353e-01
-6.85400844e-01 -6.47338092e-01 -1.09981835e+00 -8.59825075e-01
-1.97309956e-01 -4.35912520e-01 4.16643888e-01 1.09945571e+00
3.96120965e-01 2.75646567e-01 -7.16253221e-01 3.98589671e-01
3.48572195e-01 1.02672592e-01 8.14508438e-01 -7.35040247e-01
-6.90808073e-02 -9.36502963e-02 -8.72147799e-01 -1.20615160e+00
6.69656038e-01 -6.74238741e-01 1.99971914e-01 -1.60047424e+00
2.61423796e-01 3.27044845e-01 -4.76465166e-01 7.29542077e-01
-1.11199312e-01 3.67301583e-01 1.82437196e-01 -2.21151058e-02
-1.62862182e+00 4.97397125e-01 1.16140628e+00 -1.45584419e-01
2.42517013e-02 -5.66288948e-01 7.32042193e-02 6.88317239e-01
3.39024544e-01 -1.40415490e-01 -4.94766563e-01 -5.12062728e-01
-1.06499664e-01 1.59931973e-01 8.55033934e-01 -1.19030595e+00
3.79005641e-01 -4.39331740e-01 4.90658507e-02 -5.41211843e-01
5.77031493e-01 -8.31796467e-01 -4.27953869e-01 -3.50532383e-01
-8.74225318e-01 1.23534743e-02 3.56018543e-01 8.24360430e-01
-7.20163465e-01 1.25873491e-01 2.53149450e-01 -8.06993097e-02
-1.25445223e+00 2.34227091e-01 -1.47056460e-01 9.91691798e-02
1.45102096e+00 -2.86465675e-01 -4.90404040e-01 -4.27330643e-01
-1.28300822e+00 2.02687472e-01 2.22491235e-01 5.52782714e-01
5.45478284e-01 -1.61968756e+00 -4.74836916e-01 -9.79709700e-02
5.64116538e-01 -4.55826782e-02 6.03991032e-01 9.95748520e-01
-3.01279604e-01 8.31793845e-01 -2.52845466e-01 -1.03387952e+00
-1.46246064e+00 5.62878311e-01 8.11490193e-02 -4.05855864e-01
-7.49479115e-01 9.38893616e-01 2.59966254e-01 -2.43283197e-01
5.93670905e-01 -7.06807494e-01 -5.29097795e-01 2.16980994e-01
4.54317778e-01 2.80329794e-01 -6.24810439e-03 -1.05827665e+00
-3.48026961e-01 6.97976351e-01 2.10419327e-01 -9.82065275e-02
1.26966226e+00 -1.95178568e-01 3.20074379e-01 9.59147573e-01
1.32985020e+00 -2.43898094e-01 -1.55277348e+00 -3.60639721e-01
4.20924462e-02 -5.94314516e-01 7.37038478e-02 -5.09347737e-01
-5.59116125e-01 1.12615597e+00 1.97162598e-01 7.74641111e-02
8.95197928e-01 2.66590655e-01 7.39606321e-01 2.50331372e-01
4.09542620e-01 -9.44801629e-01 8.56257439e-01 5.19868433e-01
9.39855099e-01 -1.54819202e+00 -4.41962443e-02 -2.32108966e-01
-8.28321099e-01 9.67457473e-01 8.51792634e-01 1.68767460e-02
2.96684623e-01 -8.71792361e-02 -3.40714246e-01 -2.73978949e-01
-9.97312307e-01 -4.74654406e-01 6.95554733e-01 3.80433172e-01
7.08571494e-01 -6.56548589e-02 3.79902750e-01 5.51617205e-01
2.51433343e-01 -5.80225489e-04 4.35845882e-01 1.31803513e+00
-2.00588390e-01 -7.73135662e-01 -3.03152829e-01 -1.81579724e-01
-3.78499568e-01 6.21805340e-02 -2.16604158e-01 8.93827081e-01
3.21735919e-01 1.06410205e+00 2.99768716e-01 -2.58073241e-01
4.08216596e-01 9.74744707e-02 4.48822707e-01 -9.78350580e-01
-6.96601987e-01 -4.02327292e-02 1.75357506e-01 -1.34621739e+00
-7.17409492e-01 -7.32707739e-01 -1.09526646e+00 -2.72864904e-02
1.74730688e-01 6.60254359e-02 4.01346117e-01 1.02730167e+00
2.49360040e-01 9.58770633e-01 6.41276613e-02 -1.21780002e+00
-1.94260940e-01 -9.10826266e-01 -4.60549772e-01 8.79318118e-01
6.01401806e-01 -8.33657384e-01 -1.74106672e-01 4.01201218e-01] | [10.057876586914062, 0.7600241303443909] |
8defa087-4682-4264-a302-239e57072a8e | mathematical-capabilities-of-chatgpt | 2301.13867 | null | https://arxiv.org/abs/2301.13867v1 | https://arxiv.org/pdf/2301.13867v1.pdf | Mathematical Capabilities of ChatGPT | We investigate the mathematical capabilities of ChatGPT by testing it on publicly available datasets, as well as hand-crafted ones, and measuring its performance against other models trained on a mathematical corpus, such as Minerva. We also test whether ChatGPT can be a useful assistant to professional mathematicians by emulating various use cases that come up in the daily professional activities of mathematicians (question answering, theorem searching). In contrast to formal mathematics, where large databases of formal proofs are available (e.g., the Lean Mathematical Library), current datasets of natural-language mathematics, used to benchmark language models, only cover elementary mathematics. We address this issue by introducing a new dataset: GHOSTS. It is the first natural-language dataset made and curated by working researchers in mathematics that (1) aims to cover graduate-level mathematics and (2) provides a holistic overview of the mathematical capabilities of language models. We benchmark ChatGPT on GHOSTS and evaluate performance against fine-grained criteria. We make this new dataset publicly available to assist a community-driven comparison of ChatGPT with (future) large language models in terms of advanced mathematical comprehension. We conclude that contrary to many positive reports in the media (a potential case of selection bias), ChatGPT's mathematical abilities are significantly below those of an average mathematics graduate student. Our results show that ChatGPT often understands the question but fails to provide correct solutions. Hence, if your goal is to use it to pass a university exam, you would be better off copying from your average peer! | ['Julius Berner', 'Alexis Chevalier', 'Philipp Christian Petersen', 'Thomas Lukasiewicz', 'Tommaso Salvatori', 'Ryan-Rhys Griffiths', 'Luca Pinchetti', 'Simon Frieder'] | 2023-01-31 | null | null | null | null | ['selection-bias', 'elementary-mathematics'] | ['natural-language-processing', 'reasoning'] | [-2.48873636e-01 2.86447525e-01 -1.80018023e-02 -4.44148034e-02
-8.32660675e-01 -8.61932993e-01 7.78266251e-01 6.48185372e-01
-4.08924967e-01 6.96435392e-01 -1.83450997e-01 -1.30967176e+00
-5.25042951e-01 -1.18756294e+00 -9.09233570e-01 -1.05440527e-01
-2.99269576e-02 5.79186618e-01 8.87323543e-02 -2.33920112e-01
6.18149400e-01 2.89998651e-01 -1.42076647e+00 1.22871779e-01
1.60385025e+00 2.14686066e-01 4.79352921e-01 7.98036098e-01
-3.46499592e-01 1.05151415e+00 -7.57890642e-01 -8.51836801e-01
-1.45562157e-01 -4.05599385e-01 -1.31911850e+00 -3.90464753e-01
9.52147484e-01 -1.25090837e-01 -2.05136448e-01 1.15231156e+00
1.12404801e-01 -1.03798741e-03 3.47244054e-01 -1.36368144e+00
-6.69650257e-01 1.10375857e+00 8.07764232e-02 2.28061780e-01
6.41575396e-01 1.84407413e-01 1.09861636e+00 -3.74285996e-01
8.49014163e-01 1.35288823e+00 5.66551447e-01 2.74832994e-01
-1.27715611e+00 -7.49286830e-01 -5.67488782e-02 3.13033491e-01
-1.13775218e+00 -1.89585641e-01 5.94138503e-01 -4.08960670e-01
6.25057936e-01 4.09213126e-01 8.10453534e-01 9.34876978e-01
1.17898665e-01 6.01725280e-01 1.36757076e+00 -5.34471333e-01
-3.53282727e-02 5.05513608e-01 3.10015231e-01 8.60869706e-01
3.13304543e-01 -4.74471748e-01 -4.90747213e-01 -2.67801434e-01
7.05036342e-01 -5.07886171e-01 -4.00743425e-01 -1.22452654e-01
-1.67741442e+00 6.45785153e-01 -4.65735607e-02 7.37931550e-01
-4.54411320e-02 2.03026980e-01 1.22710109e-01 9.97818410e-01
5.86131364e-02 8.95026624e-01 -5.85863113e-01 -8.34798455e-01
-9.81038451e-01 7.70619988e-01 1.39300931e+00 1.04438090e+00
3.97369564e-01 -3.39588404e-01 1.11817680e-01 5.81606865e-01
1.50491491e-01 2.77462035e-01 3.30129057e-01 -1.33040428e+00
7.12182403e-01 9.04293120e-01 -1.96218058e-01 -1.02122951e+00
8.94829631e-02 -5.63729227e-01 -5.34460008e-01 8.09411705e-02
1.14165878e+00 2.18786940e-01 -1.07910998e-01 1.65730739e+00
5.53972833e-02 -3.20779309e-02 -1.02795258e-01 6.25024378e-01
1.15274036e+00 5.57576537e-01 1.60073474e-01 -2.71303691e-02
1.19278944e+00 -5.36182940e-01 -3.53182703e-01 2.26871192e-01
1.16088772e+00 -9.36336935e-01 1.24859226e+00 8.59539390e-01
-1.48367071e+00 -3.99697453e-01 -8.15875471e-01 -2.19847128e-01
-5.54010451e-01 -1.17665648e-01 8.95174682e-01 8.74970078e-01
-1.07663310e+00 9.05800104e-01 -5.41902304e-01 -2.36749306e-01
4.04550076e-01 9.13354233e-02 -3.74376953e-01 -2.57932216e-01
-1.12616551e+00 9.87537503e-01 1.38240919e-01 -4.30647522e-01
-7.97084391e-01 -1.35238850e+00 -6.48611426e-01 1.46463349e-01
5.39188504e-01 -7.02766001e-01 1.22883701e+00 -5.00867844e-01
-1.18408132e+00 1.12535834e+00 7.41472393e-02 -4.04311895e-01
8.31157565e-01 8.94162655e-02 4.83010374e-02 1.09943800e-01
1.64753050e-01 2.90600657e-01 2.24654391e-01 -9.14669275e-01
-3.25730413e-01 -8.04418847e-02 6.80796504e-01 -2.85284281e-01
-2.00224265e-01 1.91648409e-01 -8.69107693e-02 -5.35579145e-01
5.29255457e-02 -4.72609490e-01 2.48234160e-02 -4.93914559e-02
-2.19634488e-01 -8.87288272e-01 6.07595325e-01 -5.93422830e-01
1.36095405e+00 -1.49915242e+00 4.98575382e-02 2.25684404e-01
6.97177947e-01 5.30024916e-02 -7.52031524e-03 5.43928146e-01
-4.68012132e-02 7.72912860e-01 1.76204562e-01 -1.66592017e-01
5.92241883e-01 -9.20803621e-02 -2.69762635e-01 3.28623295e-01
-1.26398116e-01 1.03424001e+00 -1.14838779e+00 -7.88870037e-01
2.30692059e-01 -2.95403041e-02 -5.87924302e-01 -1.75772868e-02
-4.22820300e-01 1.51267633e-01 -5.24707317e-01 7.59355009e-01
4.52643991e-01 -3.73791873e-01 5.63005684e-03 6.25374436e-01
-3.49414021e-01 8.94920230e-01 -1.01689744e+00 1.68472588e+00
-5.60523987e-01 9.48247850e-01 2.50994265e-01 -1.06706834e+00
7.26585209e-01 3.39997232e-01 -4.15917039e-02 -4.36607867e-01
9.59886163e-02 3.29091012e-01 5.34376502e-01 -3.68160844e-01
4.33632672e-01 9.04900357e-02 1.50138721e-01 8.12179208e-01
1.58085495e-01 -8.23476434e-01 6.98051035e-01 6.91241205e-01
1.24030542e+00 1.40487179e-01 -6.05596341e-02 -5.74575484e-01
6.77713811e-01 9.29954201e-02 4.04285640e-02 1.15791643e+00
-1.20860346e-01 3.06007452e-02 7.93814480e-01 -2.21951470e-01
-1.06071675e+00 -9.75319386e-01 -2.45209590e-01 1.04987955e+00
-2.79231697e-01 -9.83985722e-01 -7.81947017e-01 -4.11215663e-01
-1.80455759e-01 9.18001652e-01 -7.04041496e-02 2.80066222e-01
-4.39023525e-01 -2.21042648e-01 6.37805223e-01 -2.14984417e-01
7.27186918e-01 -8.36164713e-01 -3.47274423e-01 1.80953607e-01
-2.40104511e-01 -1.17295837e+00 1.18377924e-01 -2.16773629e-01
-7.26596534e-01 -1.20780993e+00 -5.18325388e-01 -7.77043700e-01
4.50882345e-01 -5.74288331e-03 1.61208582e+00 8.36207330e-01
-2.08393112e-01 7.91077912e-01 -2.89684147e-01 -6.57878041e-01
-9.32119906e-01 3.03518951e-01 -1.76685467e-01 -9.55606282e-01
3.20101708e-01 -8.37759316e-01 -9.13315639e-02 5.11564836e-02
-7.27405012e-01 3.03331792e-01 5.33108056e-01 4.27684873e-01
-2.59847701e-01 3.94409329e-01 4.26206976e-01 -8.21884215e-01
8.06977570e-01 -3.04312348e-01 -7.97097147e-01 3.15151334e-01
-5.47767222e-01 -1.35904998e-01 5.81068158e-01 -5.62000453e-01
-6.13501251e-01 -9.40900326e-01 3.91640365e-02 -1.47317380e-01
-2.70284712e-01 7.74002790e-01 5.46332076e-02 -4.09535021e-01
5.27398050e-01 3.26285064e-01 -1.50345042e-01 -4.29578245e-01
-6.36915341e-02 3.80698860e-01 3.52507323e-01 -1.37729287e+00
1.34371829e+00 -3.79495114e-01 1.47781059e-01 -1.11570287e+00
-5.60113847e-01 -1.93818256e-01 -4.00515705e-01 -3.38319361e-01
1.93990514e-01 -6.63084745e-01 -1.37932956e+00 2.92049646e-01
-1.21371210e+00 -6.65131211e-01 1.52657241e-01 6.21274769e-01
-4.25471634e-01 4.70616490e-01 -6.42334998e-01 -9.34596181e-01
6.64114133e-02 -9.69232619e-01 6.57047868e-01 1.18580408e-01
-6.31337285e-01 -1.37946987e+00 -3.02226692e-01 9.52554166e-01
3.17086935e-01 1.35106027e-01 1.49114609e+00 -7.34348774e-01
-9.82251346e-01 6.13232441e-02 -1.61379829e-01 1.70887247e-01
-6.17839873e-01 2.47773424e-01 -7.23334253e-01 -1.82597693e-02
-8.80086944e-02 -5.24139225e-01 4.86397088e-01 -1.97076425e-03
1.26426208e+00 -3.93847257e-01 -3.26132476e-02 1.19034704e-02
1.15284526e+00 -4.26203191e-01 4.86650079e-01 6.74178064e-01
2.94117272e-01 6.76473796e-01 2.26645932e-01 -3.89618315e-02
7.21736908e-01 2.05380708e-01 1.72912747e-01 4.09607321e-01
1.48126587e-01 -4.25889909e-01 3.56734365e-01 1.06514001e+00
-3.18932116e-01 -8.00229385e-02 -1.36256075e+00 7.72020459e-01
-1.52340746e+00 -8.85010600e-01 -6.22174025e-01 2.17326403e+00
1.19642186e+00 4.28912610e-01 1.51965197e-03 3.11345845e-01
2.27936104e-01 -2.35834718e-01 2.71462888e-01 -6.10118985e-01
1.11504495e-01 7.53509283e-01 7.90993273e-02 5.74257970e-01
-6.18472993e-01 6.92537844e-01 6.22466326e+00 1.11784124e+00
-7.13783264e-01 -1.30497590e-02 4.49955285e-01 1.84754178e-01
-4.65126425e-01 1.94106385e-01 -4.52161580e-01 2.62061208e-01
1.25457633e+00 -5.97262383e-01 5.77116549e-01 5.86099088e-01
1.81265101e-01 -2.05073208e-01 -1.46980107e+00 7.24066973e-01
-1.71333626e-01 -1.49563718e+00 4.78648432e-02 2.71806270e-01
5.83455443e-01 -2.73646176e-01 -1.26818748e-04 7.26177931e-01
4.56482977e-01 -1.51118708e+00 7.00529039e-01 4.61018652e-01
2.62089282e-01 -5.65891266e-01 6.90350235e-01 9.83245552e-01
-6.64452314e-01 1.55049771e-01 -4.90602329e-02 -7.84535885e-01
-4.65338558e-01 4.18285787e-01 -9.16567564e-01 7.17155814e-01
5.17711639e-01 4.27174747e-01 -8.67087245e-01 1.05305278e+00
-2.24476859e-01 9.19596851e-01 -2.13386789e-01 -5.21909297e-01
2.82046676e-01 -4.16377336e-01 6.51805222e-01 1.00990415e+00
2.46315435e-01 3.26802224e-01 1.66439116e-01 1.70216811e+00
-2.09225819e-01 2.61029512e-01 -5.32014847e-01 -6.75795913e-01
5.10372698e-01 1.27834475e+00 -4.64877278e-01 -4.79065001e-01
-4.24918860e-01 3.35086107e-01 2.42216453e-01 1.59028560e-01
-5.53490341e-01 -5.67275763e-01 2.97718257e-01 4.43785906e-01
-3.38510334e-01 -5.10895550e-01 -3.05495769e-01 -1.14522254e+00
6.59487545e-02 -1.20354927e+00 -7.27045834e-02 -9.40842748e-01
-1.24049926e+00 -1.64076872e-02 6.82627410e-02 -7.03115165e-01
-4.17116769e-02 -8.46319318e-01 -8.36596131e-01 1.09665132e+00
-1.28705549e+00 -1.02600861e+00 -2.84100711e-01 3.68486881e-01
1.73505500e-01 -5.81913590e-02 8.73412251e-01 1.58559635e-01
-1.47274733e-01 5.57469070e-01 -3.46534848e-01 4.05067235e-01
5.42401791e-01 -1.49217391e+00 1.10711835e-01 5.48271000e-01
2.67503649e-01 1.09556532e+00 9.74383473e-01 -3.87257010e-01
-1.77577937e+00 -6.38933599e-01 1.44636083e+00 -9.56882417e-01
1.16884804e+00 -3.92077416e-01 -1.01002359e+00 5.34382582e-01
1.36960194e-01 -4.53996539e-01 4.16429847e-01 3.34586561e-01
-2.28987440e-01 3.05218309e-01 -9.22617555e-01 6.39949739e-01
1.01454604e+00 -6.06759965e-01 -1.15193403e+00 5.03755927e-01
6.08790457e-01 -5.16795695e-01 -1.27557707e+00 -7.21850572e-03
4.60762054e-01 -9.65042710e-01 9.32653964e-01 -5.95497787e-01
8.72534037e-01 1.37326151e-01 1.43064097e-01 -1.01910818e+00
1.86488330e-01 -8.44585240e-01 1.72354653e-01 1.31457293e+00
3.27554375e-01 -6.10482752e-01 7.68214941e-01 7.71449685e-01
-5.98010607e-02 -7.53116786e-01 -8.97945225e-01 -7.43576169e-01
8.05446923e-01 -8.49013567e-01 4.29092914e-01 1.41789389e+00
2.92404413e-01 1.29602656e-01 5.38373530e-01 -1.11005127e-01
6.27063334e-01 3.91344041e-01 1.20420849e+00 -1.62576449e+00
-2.17154488e-01 -1.11164749e+00 -5.52807212e-01 -8.11703920e-01
5.16402662e-01 -1.26500249e+00 -5.23710012e-01 -1.47461319e+00
2.64625728e-01 -6.55040085e-01 3.30408186e-01 3.77279878e-01
1.82876021e-01 -1.81276891e-02 1.61197647e-01 -2.75037050e-01
-5.11734366e-01 6.89230487e-02 1.62567854e+00 -6.77286834e-02
1.01474255e-01 -1.14165053e-01 -7.82704413e-01 9.68613923e-01
5.42805672e-01 -1.02694817e-01 -1.47717685e-01 -4.13555242e-02
5.51201344e-01 7.99325854e-02 9.65231717e-01 -9.60413873e-01
3.39799613e-01 -5.62159657e-01 1.22818418e-01 -3.06384563e-01
-1.60376340e-01 -5.50583065e-01 -2.42185205e-01 3.86666268e-01
-3.98993343e-01 -1.53186554e-02 2.21598268e-01 2.78387684e-02
-1.36467770e-01 -5.16754866e-01 3.17405403e-01 -5.21565318e-01
-2.45997474e-01 6.01645671e-02 -5.64564288e-01 3.33612293e-01
6.28265440e-01 -1.21109992e-01 -5.48259079e-01 -6.17345333e-01
-3.57445240e-01 3.60087514e-01 4.68067139e-01 2.00875342e-01
3.12889010e-01 -9.67012286e-01 -9.85470712e-01 -1.85639486e-01
-1.51340976e-01 6.88918829e-02 -2.34677806e-01 1.07363772e+00
-6.79789543e-01 8.35209250e-01 1.16027430e-01 -4.08619314e-01
-1.44238591e+00 4.64518107e-02 1.27096534e-01 -4.51051801e-01
-5.68989635e-01 8.67984712e-01 -1.67276219e-01 -8.01556349e-01
3.20486784e-01 -8.43533278e-01 1.58359766e-01 -2.68838048e-01
4.61669743e-01 3.10868084e-01 3.24337557e-02 9.33959894e-03
-9.09561384e-03 1.42101884e-01 1.87659338e-01 -2.90552527e-02
1.37249732e+00 1.28252074e-01 -5.63235998e-01 7.17349172e-01
7.49040663e-01 4.62286741e-01 -1.04586810e-01 -1.44969881e-01
9.98104587e-02 -4.15100664e-01 -9.08908322e-02 -9.97621715e-01
-3.26824307e-01 9.07007337e-01 -2.46922255e-01 5.58476806e-01
4.51694071e-01 5.56406798e-03 3.75574589e-01 8.92405987e-01
5.56927443e-01 -8.18964839e-01 -4.78586778e-02 7.19639242e-01
7.63459086e-01 -1.22875273e+00 1.94003195e-01 -4.74212646e-01
1.18992645e-02 1.25379074e+00 5.40102243e-01 8.31818208e-02
2.69371480e-01 -1.15346879e-01 -3.30062300e-01 -2.68900573e-01
-1.21489906e+00 1.69826567e-01 2.68988699e-01 3.06955367e-01
9.43979919e-01 4.14735153e-02 -3.19193870e-01 7.51841187e-01
-1.12804115e+00 2.21235007e-01 9.28183556e-01 7.19853342e-01
-4.20371741e-01 -1.24935961e+00 -5.51486075e-01 3.81254464e-01
-5.90738595e-01 -2.91120201e-01 -6.35120571e-01 1.18655705e+00
1.60950363e-01 9.91410077e-01 -1.29800215e-01 9.99324098e-02
-1.61107212e-01 4.09201652e-01 1.01466036e+00 -6.50300086e-01
-8.00642908e-01 -5.86916506e-01 3.08816463e-01 -4.09694836e-02
-2.80299902e-01 -4.51444447e-01 -9.86653507e-01 -1.14379537e+00
-2.85397470e-01 6.52650535e-01 5.69468915e-01 1.18567348e+00
-1.62967637e-01 5.31813681e-01 -8.99145752e-02 -3.35598856e-01
-6.26786053e-01 -1.17634475e+00 -4.47565794e-01 1.46057770e-01
1.65668756e-01 -3.54049891e-01 -5.29165626e-01 -1.09823132e-02] | [9.622729301452637, 7.317354679107666] |
411a1c35-797a-4773-89ac-6414a6410565 | vonda-a-framework-for-ontology-based-dialogue | 1910.00340 | null | https://arxiv.org/abs/1910.00340v1 | https://arxiv.org/pdf/1910.00340v1.pdf | VOnDA: A Framework for Ontology-Based Dialogue Management | We present VOnDA, a framework to implement the dialogue management functionality in dialogue systems. Although domain-independent, VOnDA is tailored towards dialogue systems with a focus on social communication, which implies the need of long-term memory and high user adaptivity. For these systems, which are used in health environments or elderly care, margin of error is very low and control over the dialogue process is of topmost importance. The same holds for commercial applications, where customer trust is at risk. VOnDA's specification and memory layer relies upon (extended) RDF/OWL, which provides a universal and uniform representation, and facilitates interoperability with external data sources, e.g., from physical sensors. | ['Christophe Biwer', 'Anna Welker', 'Bernd Kiefer'] | 2019-10-01 | null | null | null | null | ['dialogue-management'] | ['natural-language-processing'] | [-6.32321954e-01 1.05267560e+00 -1.74632519e-01 -3.76322687e-01
2.29854994e-02 -4.76080716e-01 7.17303097e-01 7.33743489e-01
-6.36849523e-01 9.74252820e-01 5.73761821e-01 -7.39931166e-02
-2.80673325e-01 -1.05431306e+00 2.17204511e-01 -9.19280276e-02
1.54818937e-01 4.40301746e-01 6.01995528e-01 -9.00161743e-01
-9.90379378e-02 2.89616585e-01 -1.38691199e+00 2.73839355e-01
6.09022677e-01 1.03256607e+00 9.97113883e-02 6.40282094e-01
-4.39832091e-01 1.15227532e+00 -6.29430234e-01 -3.69392037e-01
-4.77292597e-01 -1.34411186e-01 -1.41569340e+00 -2.63769627e-01
-5.45171916e-01 -6.19830787e-01 2.30195224e-01 5.65211535e-01
6.99032605e-01 -1.27974436e-01 1.03302084e-01 -1.15049744e+00
-1.87404037e-01 4.16828781e-01 6.02392197e-01 -4.63056058e-01
1.17386091e+00 -1.21779321e-02 7.09373593e-01 -4.83680397e-01
8.09423506e-01 1.08484983e+00 6.66212380e-01 9.22664165e-01
-9.58246589e-01 1.36728898e-01 -7.47987628e-02 -1.10915788e-01
-8.60487700e-01 -7.90637434e-01 3.08880866e-01 -2.94732094e-01
1.29246628e+00 5.81516802e-01 6.34747982e-01 1.07079911e+00
1.91977516e-01 3.80124182e-01 8.45475554e-01 -6.17343962e-01
6.61484361e-01 9.30414438e-01 4.72788960e-01 2.67076463e-01
1.62376568e-01 -4.75436777e-01 -7.51318634e-01 -6.02099776e-01
3.46882939e-01 -2.56144553e-01 -2.91926682e-01 -4.44598079e-01
-8.30411017e-01 7.07939744e-01 -2.23104402e-01 6.81062222e-01
-5.14700115e-01 -5.33769190e-01 8.15576553e-01 4.91325051e-01
3.56903374e-02 2.12761298e-01 -5.49883246e-01 -8.05511951e-01
-1.47982806e-01 5.70334978e-02 1.77978683e+00 9.75401521e-01
3.45352292e-01 -4.66370285e-01 7.92532787e-02 1.04003763e+00
8.88310552e-01 1.67742252e-01 4.75503415e-01 -1.08396530e+00
-7.26119503e-02 1.05120909e+00 5.90177536e-01 -5.44386983e-01
-7.64354289e-01 2.36937910e-01 -5.51243305e-01 2.19826967e-01
3.32912385e-01 -2.89948314e-01 -7.23434687e-02 1.50705659e+00
6.51480138e-01 -8.96274209e-01 8.84094775e-01 6.05738997e-01
1.37390411e+00 1.04185037e-01 2.80744672e-01 -5.52354634e-01
1.58536720e+00 -4.37777191e-01 -1.20517457e+00 -2.51100361e-01
7.29585528e-01 -4.59807605e-01 1.09923697e+00 1.92766517e-01
-1.22967684e+00 1.12590037e-01 -9.14497077e-01 -1.77343711e-01
-6.73069596e-01 -5.80226541e-01 4.17118162e-01 9.71871376e-01
-1.06421828e+00 2.82134086e-01 -4.37418759e-01 -1.05003595e+00
-3.25758874e-01 3.96054566e-01 -5.79941511e-01 6.63188100e-01
-1.79491889e+00 1.48385882e+00 3.62431258e-01 5.48356101e-02
5.44382667e-04 -9.98661667e-02 -8.11743855e-01 -2.22499609e-01
3.32028680e-02 -5.63702643e-01 1.34484279e+00 -5.53592741e-01
-2.20979571e+00 8.64896536e-01 4.42225814e-01 -5.20345569e-01
8.02900732e-01 -2.19730213e-01 -7.59322882e-01 9.05284137e-02
-1.53971121e-01 1.16046473e-01 1.61678761e-01 -9.64064777e-01
-5.07853746e-01 -2.56736547e-01 7.12332368e-01 2.22484976e-01
-6.59452081e-01 3.46955299e-01 -2.56743103e-01 2.97688186e-01
-4.21228349e-01 -4.65493977e-01 -1.15774475e-01 1.53533602e-02
1.20718122e-01 -1.36960074e-01 7.47711003e-01 -6.15087032e-01
1.65318942e+00 -1.87014854e+00 -2.00478896e-01 2.67248005e-01
2.02339277e-01 3.91838700e-01 6.16706252e-01 1.09548247e+00
7.41733015e-01 -5.01447497e-03 1.61712959e-01 2.42269170e-02
4.81321961e-01 4.07382220e-01 1.87817484e-01 -1.91181283e-02
-2.80266762e-01 3.11862499e-01 -8.25630605e-01 -5.23463845e-01
2.84352541e-01 3.89800698e-01 -2.15833619e-01 5.13937294e-01
-4.25354421e-01 4.32678759e-01 -5.31447053e-01 2.26552263e-01
3.32083166e-01 -2.99754888e-01 7.80296206e-01 1.25114784e-01
-3.28026295e-01 4.71378863e-01 -1.12434173e+00 1.72000277e+00
-8.20526123e-01 -2.17429817e-01 7.02855170e-01 -8.59466717e-02
9.67070580e-01 9.83819425e-01 4.18767631e-01 -9.05483961e-01
1.20040826e-01 3.66641432e-01 -2.82319129e-01 -1.07466793e+00
7.68705249e-01 1.38502032e-01 -2.44444370e-01 6.49125576e-01
-1.18134886e-01 1.92284986e-01 -8.59234631e-02 2.52160132e-01
9.71061409e-01 2.28863895e-01 8.00397635e-01 -3.83434147e-01
9.07970488e-01 -1.99942872e-01 4.83669668e-01 2.28086337e-01
-2.41080165e-01 -6.03270605e-02 5.53885400e-01 -2.64876544e-01
-8.51705492e-01 -6.72795475e-01 -3.47144485e-01 9.84734118e-01
-1.61532667e-02 -7.85223901e-01 -9.27800238e-01 -5.11562645e-01
-8.93864259e-02 6.73380673e-01 -2.78473049e-01 -2.89137121e-02
-2.69293040e-01 -8.75193402e-02 5.47685027e-01 2.06150666e-01
6.18180811e-01 -1.16547275e+00 -1.28722811e+00 6.67410076e-01
-1.08453274e-01 -1.14237309e+00 1.84025839e-01 -1.57384366e-01
-5.22881091e-01 -1.08383679e+00 -2.35326901e-01 -3.57225448e-01
7.53611252e-02 -4.91395265e-01 1.21923029e+00 1.30626470e-01
2.75532842e-01 1.05223608e+00 -3.96651417e-01 -3.62763375e-01
-8.24614942e-01 5.45000173e-02 -7.04324991e-02 -2.84396797e-01
4.05563921e-01 -2.12778091e-01 -6.44930959e-01 4.66838568e-01
-8.06159675e-01 -6.06172122e-02 -1.60004631e-01 7.52091825e-01
-1.82871684e-01 -4.81467128e-01 1.18373668e+00 -1.07306886e+00
1.16478646e+00 -5.62744021e-01 -2.30659992e-01 5.05016744e-01
-7.78220534e-01 -1.67056203e-01 5.36083400e-01 4.48415317e-02
-1.30175400e+00 -3.84049922e-01 -5.81606209e-01 6.21404827e-01
-3.15795183e-01 6.00092888e-01 -5.84527135e-01 1.75618097e-01
6.00311399e-01 -3.66656244e-01 6.48369670e-01 -4.01609391e-01
1.37173131e-01 1.28396606e+00 5.01009040e-02 -4.50580209e-01
-2.51207232e-01 2.96987444e-01 -4.31415230e-01 -1.08147037e+00
-1.16168536e-01 -2.08327368e-01 -4.86706913e-01 -5.71432829e-01
9.14917648e-01 -6.07281029e-01 -1.25458491e+00 3.78967434e-01
-1.06498516e+00 -5.37838519e-01 -4.90247041e-01 3.80115598e-01
-6.79514408e-01 3.85718167e-01 -6.05191946e-01 -1.13552272e+00
-6.96593940e-01 -6.42852664e-01 3.70900929e-01 3.23118210e-01
-9.68374789e-01 -1.24417067e+00 9.31161270e-02 3.57511818e-01
6.87724352e-01 1.95742339e-01 8.64987314e-01 -8.90797794e-01
8.54124650e-02 -3.27144474e-01 1.82390958e-01 3.32590580e-01
3.36674273e-01 -1.48961127e-01 -1.05045497e+00 -1.30394414e-01
6.77445978e-02 -4.85017598e-01 -4.60777491e-01 -3.64024967e-01
2.92139649e-02 -3.97292882e-01 -7.56692886e-02 -5.65787494e-01
9.92498398e-01 3.59697491e-01 7.44187057e-01 6.00356877e-01
-9.41488077e-04 1.16134918e+00 1.04596984e+00 9.04661775e-01
9.32138562e-01 8.39347661e-01 3.09939265e-01 8.28829110e-02
2.26533487e-01 1.74741507e-01 3.31725866e-01 7.90613949e-01
1.11076288e-01 -8.03808272e-02 -8.91296566e-01 3.53810072e-01
-2.40429091e+00 -7.12028027e-01 -3.52371901e-01 2.42765498e+00
1.06997228e+00 -2.00212672e-02 3.99128079e-01 1.55222937e-01
4.95002747e-01 3.31263728e-02 -3.14025074e-01 -9.47983623e-01
1.61628500e-01 -6.29709288e-02 7.66836107e-02 6.83070600e-01
-4.66693729e-01 4.77530479e-01 6.59216976e+00 6.90712854e-02
-7.12099433e-01 4.40380782e-01 -9.63452533e-02 1.58081487e-01
-2.46297166e-01 -9.96729955e-02 -4.61487144e-01 4.28445131e-01
1.51161313e+00 -2.98061013e-01 8.04272369e-02 5.75522006e-01
4.58426565e-01 -3.98075491e-01 -9.35925245e-01 4.63478148e-01
-4.18744385e-01 -1.14518154e+00 -6.47756875e-01 5.25058880e-02
-1.87423840e-01 -3.00415248e-01 -5.64751804e-01 -7.79260769e-02
2.91610271e-01 -5.57279587e-01 6.68777049e-01 7.70338356e-01
5.11786520e-01 -6.89158261e-01 1.01169574e+00 4.60613519e-01
-7.84295559e-01 -7.88366720e-02 -6.39637113e-02 4.90885340e-02
4.87946302e-01 5.19606888e-01 -6.92431211e-01 3.69820386e-01
7.92173564e-01 8.31929594e-02 -2.30688117e-02 5.97398758e-01
1.02415621e-01 -3.68849635e-02 -3.09598297e-01 -2.38427505e-01
-1.96957245e-01 -2.25258604e-01 2.68872350e-01 1.22828913e+00
-5.67511432e-02 1.07859068e-01 -4.77628857e-02 3.33125181e-02
2.87025243e-01 6.12756968e-01 -6.64439380e-01 1.56728014e-01
6.90300167e-01 1.17923510e+00 -2.29428142e-01 -1.83649078e-01
-7.73641765e-01 7.22969174e-01 -7.41181988e-03 -1.08712912e-01
-4.23970222e-01 -2.90647775e-01 6.68418705e-01 3.60423177e-01
-1.00910164e-01 -5.61295524e-02 1.23837955e-01 -7.87043750e-01
1.19868897e-01 -9.11600769e-01 5.47734857e-01 -3.34567904e-01
-9.02237177e-01 6.52339816e-01 -2.57247210e-01 -1.14912069e+00
-7.65378475e-01 -2.60876834e-01 -3.07911009e-01 6.37181461e-01
-1.11620450e+00 -1.33144820e+00 -2.67137945e-01 7.35100210e-01
-1.25520363e-01 -9.46333073e-03 1.76513672e+00 5.04198492e-01
-2.45246187e-01 2.51741767e-01 -1.48431897e-01 -1.82076275e-01
8.63427818e-01 -1.15699828e+00 -8.46045241e-02 -1.16863623e-01
-1.05871153e+00 7.01259434e-01 7.74442732e-01 -5.81563175e-01
-1.66247046e+00 -5.34061134e-01 1.25605834e+00 -9.98164490e-02
7.59706914e-01 -3.66614550e-01 -9.47342455e-01 2.59186953e-01
6.85981631e-01 -2.72037655e-01 1.08205354e+00 2.45045036e-01
-2.87481192e-02 -1.46619424e-01 -1.79796541e+00 3.46720189e-01
8.03986430e-01 -9.59810495e-01 -5.78577399e-01 2.12905854e-01
3.73396963e-01 -3.19229394e-01 -1.80133641e+00 8.82217139e-02
8.61133337e-01 -1.43462098e+00 5.41356027e-01 -3.42434675e-01
-2.85664290e-01 -1.74180672e-01 6.79787202e-03 -7.73194849e-01
2.45870247e-01 -7.33567536e-01 -4.42329824e-01 1.66583240e+00
4.28539008e-01 -1.24189174e+00 5.04740298e-01 1.71344590e+00
2.02541754e-01 -3.58306468e-01 -8.07275891e-01 -6.14811659e-01
-3.50076109e-01 -1.78732514e-01 8.91848087e-01 8.63902986e-01
9.94388878e-01 5.26644647e-01 -4.33760732e-01 -7.53195733e-02
3.14810798e-02 -4.36809301e-01 7.94942141e-01 -1.44482279e+00
-2.51719356e-01 1.32505894e-01 -3.55068028e-01 -5.43281436e-01
-9.86196473e-02 -1.33518830e-01 -1.39198944e-01 -2.09935546e+00
-6.04126930e-01 -4.22885031e-01 1.87735349e-01 3.02009493e-01
5.28147280e-01 -3.80709261e-01 -7.67457038e-02 1.36175051e-01
-6.50348902e-01 4.31471586e-01 6.93200350e-01 2.83192128e-01
-4.52446669e-01 1.67969584e-01 -6.27529919e-01 6.54986441e-01
8.65060270e-01 -7.82048702e-03 -4.84296113e-01 -1.23476245e-01
6.62784874e-01 3.56709987e-01 -1.41677652e-02 -7.05454946e-01
6.42585397e-01 -8.80965516e-02 -3.88268709e-01 -1.97167564e-02
5.62639236e-01 -1.23583019e+00 5.75551331e-01 5.19657612e-01
-2.71412343e-01 -1.75185561e-01 -1.80545911e-01 1.27036750e-01
-2.88988084e-01 -6.90231562e-01 6.58771217e-01 -1.42071009e-01
-5.30843735e-01 -3.18404049e-01 -8.11533868e-01 -2.55684942e-01
1.05754089e+00 -3.59142065e-01 -7.28414178e-01 -5.73891699e-01
-1.12959766e+00 4.03856397e-01 7.93499351e-01 3.09117913e-01
4.29268092e-01 -9.83628809e-01 -1.93966657e-01 -1.83100428e-03
3.71133655e-01 -1.00925617e-01 1.83219209e-01 7.76252985e-01
-3.96128327e-01 2.61930048e-01 -2.66500890e-01 -2.01144844e-01
-1.36727750e+00 2.13295862e-01 4.10368145e-01 -9.44832489e-02
-6.74491465e-01 2.64296439e-02 -6.54479980e-01 -6.72568023e-01
3.17550033e-01 5.47462702e-02 -7.17153490e-01 5.16727507e-01
7.42229104e-01 5.32470584e-01 2.85775512e-01 -5.77753067e-01
-6.18520141e-01 5.71089052e-03 5.75852655e-02 -3.45843196e-01
1.33532453e+00 -6.52038455e-01 -5.09214699e-01 8.48555267e-01
5.82327008e-01 2.45955676e-01 -8.30649912e-01 -1.60298735e-01
4.73416328e-01 -6.01582192e-02 -2.08297759e-01 -8.42327714e-01
-3.50890249e-01 2.07689285e-01 6.76284850e-01 9.72931743e-01
6.72562599e-01 -8.76363590e-02 6.75574422e-01 4.87636834e-01
6.94968462e-01 -1.81075919e+00 -3.94809335e-01 4.91788715e-01
1.03633845e+00 -9.52407718e-01 -3.28421406e-02 -2.91255236e-01
-9.69977260e-01 1.19056833e+00 3.90163392e-01 6.45833850e-01
1.01974964e+00 5.29810607e-01 4.17698711e-01 -3.61199826e-01
-9.98464584e-01 -1.39750391e-01 -5.38136423e-01 6.37601554e-01
6.02048635e-01 3.46077941e-02 -7.95793593e-01 7.03281999e-01
7.15673715e-02 3.00725818e-01 5.82368255e-01 1.58972001e+00
-4.60386366e-01 -1.70480394e+00 -2.92760432e-01 2.48973042e-01
-5.72935820e-01 3.31739575e-01 -5.87709367e-01 6.23507500e-01
-6.21265285e-02 1.41703987e+00 -1.88145727e-01 -1.19229509e-02
7.54764259e-01 3.64154398e-01 1.41686335e-01 -5.33644438e-01
-1.28490305e+00 -1.92037165e-01 1.14287364e+00 -4.53234494e-01
-7.26484716e-01 -5.45322239e-01 -1.74223006e+00 -4.68888402e-01
-2.44039193e-01 5.34797370e-01 8.14089298e-01 8.19955051e-01
5.71804345e-01 9.13060009e-02 4.60549682e-01 6.28556237e-02
-3.11074257e-01 -8.67182732e-01 -6.43128276e-01 2.34581634e-01
1.82167247e-01 -3.18360537e-01 1.79061010e-01 -3.16272050e-01] | [12.86268424987793, 7.975837230682373] |
014187f8-2200-4d14-adea-cd1e4d1db602 | c-cnn-contourlet-convolutional-neural | null | null | https://ieeexplore.ieee.org/document/9145825 | https://sci-hub.mksa.top/10.1109/TNNLS.2020.3007412 | C-CNN: Contourlet Convolutional Neural Networks | Extracting effective features is always a challenging problem for texture classification because of the uncertainty of scales and the clutter of textural patterns. For texture classification, spectral analysis is traditionally employed in the frequency domain. Recent studies have shown the potential of convolutional neural networks (CNNs) when dealing with the texture classification task in the spatial domain. In this article, we try combining both approaches in different domains for more abundant information and proposed a novel network architecture named contourlet CNN (C-CNN). The network aims to learn sparse and effective feature representations for images. First, the contourlet transform is applied to get the spectral features from an image. Second, the spatial-spectral feature fusion strategy is designed to incorporate the spectral features into CNN architecture. Third, the statistical features are integrated into the network by the statistical feature fusion. Finally, the results are obtained by classifying the fusion features. We also investigated the behavior of the parameters in contourlet decomposition. Experiments on the widely used three texture data sets (kth-tips2-b, DTD, and CUReT) and five remote sensing data sets (UCM, WHU-RS, AID, RSSCN7, and NWPU-RESISC45) demonstrate that the proposed approach outperforms several well-known classification methods in terms of classification accuracy with fewer trainable parameters. | ['Shuyuan Yang', 'Fang Liu', 'Lingling Li', 'Xu Liu', 'Licheng Jiao', 'Mengkun Liu'] | 2020-07-21 | null | null | null | ieee-transactions-on-neural-networks-and-5 | ['texture-classification'] | ['computer-vision'] | [ 2.35655144e-01 -5.85956872e-01 1.08667284e-01 -3.23073626e-01
-5.18900156e-01 4.34488468e-02 4.55180943e-01 -1.16679169e-01
-3.87226641e-01 7.11247504e-01 -4.75424267e-02 5.39309718e-02
-6.52272642e-01 -1.04339528e+00 -2.14117184e-01 -1.11248589e+00
-3.91482234e-01 -2.74343401e-01 3.15844938e-02 -4.76403236e-01
1.57662094e-01 7.19023168e-01 -1.88496983e+00 2.97980577e-01
9.21550751e-01 1.84889245e+00 6.37172088e-02 2.63091475e-01
-1.29913121e-01 4.67455477e-01 -1.96534410e-01 5.84719144e-03
2.19405442e-01 -1.48569122e-01 -4.75295991e-01 1.81522414e-01
1.38319507e-01 1.96656957e-02 -1.70773357e-01 9.87857223e-01
6.54264748e-01 3.19009304e-01 7.59613037e-01 -8.16923082e-01
-6.56879067e-01 1.75018638e-01 -8.53055716e-01 2.23251030e-01
-1.85266018e-01 -3.47487032e-01 5.78419745e-01 -1.04922593e+00
3.50857079e-01 1.21513557e+00 9.67790067e-01 -1.20488733e-01
-8.17927897e-01 -5.41920722e-01 -3.96966934e-01 5.88049173e-01
-1.62691712e+00 -2.76758730e-01 1.06366169e+00 -4.16591614e-01
5.99772811e-01 3.14843029e-01 5.26625574e-01 8.36820781e-01
4.52253819e-01 4.48340803e-01 1.50166726e+00 -4.46070850e-01
1.02595337e-01 -3.34330499e-01 1.20026536e-01 7.31921256e-01
1.30364284e-01 1.84394345e-01 -6.68335974e-01 -3.08326352e-02
7.50688016e-01 2.03704536e-02 -3.10155183e-01 1.04927301e-01
-8.68087173e-01 9.98221576e-01 5.32610297e-01 6.95483446e-01
-5.03980279e-01 -1.19134948e-01 4.33197528e-01 4.48066324e-01
1.12166393e+00 3.32671963e-02 -4.09405023e-01 1.10821560e-01
-8.62066090e-01 5.81975281e-02 5.21484137e-01 3.77722055e-01
8.14444721e-01 3.71433645e-01 -1.18624354e-02 1.15206182e+00
1.27387568e-01 7.66164422e-01 5.51785469e-01 -6.26979828e-01
1.04673915e-01 3.75599176e-01 -9.26959664e-02 -1.68825197e+00
-6.12432659e-01 -6.13143682e-01 -1.49199843e+00 8.48827213e-02
7.42712244e-02 -1.40178159e-01 -8.21874976e-01 1.11199605e+00
1.83000758e-01 8.56560320e-02 1.50898695e-01 1.09785032e+00
1.06983519e+00 6.67543769e-01 2.25825496e-02 -1.22021355e-01
1.32822168e+00 -4.84519601e-01 -7.63881028e-01 1.78222805e-01
2.20358849e-01 -9.36112344e-01 6.05284512e-01 4.38799560e-01
-4.18206871e-01 -9.23064828e-01 -1.05374956e+00 2.44873539e-01
-5.02237678e-01 6.53999925e-01 7.52785385e-01 3.94372970e-01
-8.95092905e-01 8.18591416e-01 -6.53546095e-01 -3.80912483e-01
5.82123578e-01 1.39289364e-01 -4.14706856e-01 -1.98933277e-02
-1.34857631e+00 6.53220356e-01 5.01746714e-01 6.16430879e-01
-4.18239534e-02 -3.77656162e-01 -8.43820930e-01 2.94436961e-01
-8.74187332e-03 -5.67465201e-02 4.81528938e-01 -1.08090580e+00
-1.45121455e+00 4.54949558e-01 2.13397518e-01 -1.79218695e-01
8.19294900e-02 3.43633406e-02 -7.30944276e-01 3.90516937e-01
-4.57787998e-02 3.17312747e-01 1.03830099e+00 -8.30390215e-01
-5.38505733e-01 -3.29247981e-01 -5.19456446e-01 1.20066836e-01
-4.53705966e-01 -1.73052654e-01 2.22511768e-01 -8.49786639e-01
4.75959778e-01 -5.15756607e-01 7.83748105e-02 4.46517132e-02
-1.03771843e-01 -1.60833925e-01 1.11776316e+00 -8.26769292e-01
8.99053872e-01 -2.58721423e+00 -4.00422178e-02 4.46313918e-01
-1.70608610e-03 2.48083755e-01 -1.80631161e-01 5.02617657e-01
-1.50137246e-01 -2.78725401e-02 -2.46535853e-01 6.45342842e-02
-1.94505379e-01 1.07142717e-01 -7.60676041e-02 6.37155592e-01
4.45080757e-01 5.67380369e-01 -4.34342802e-01 -5.18270850e-01
3.71060133e-01 6.02156639e-01 -2.06150547e-01 -5.08909933e-02
1.08954147e-01 3.55408967e-01 -5.95099270e-01 8.33642602e-01
1.01570940e+00 -4.84287292e-02 -8.98274407e-02 -7.36637235e-01
-4.60877091e-01 -6.25847638e-01 -1.19105327e+00 1.21125281e+00
-4.70699370e-01 6.62768841e-01 8.46813321e-02 -1.49861515e+00
1.34107924e+00 2.85052270e-01 7.30600715e-01 -9.80238140e-01
2.46453598e-01 4.47826773e-01 -1.88330393e-02 -8.54893804e-01
2.35971794e-01 -1.35250524e-01 1.46954373e-01 1.25415474e-01
3.20541203e-01 1.17620490e-01 -5.92280850e-02 -7.96672285e-01
4.85438526e-01 3.56465504e-02 2.52491504e-01 -5.92149198e-01
7.15181410e-01 -2.90072076e-02 4.78090525e-01 4.30811942e-01
1.74691575e-03 4.53081012e-01 2.98534751e-01 -9.95271504e-01
-9.34730411e-01 -4.53971684e-01 -5.37974894e-01 1.03675640e+00
3.26912999e-02 1.34289846e-01 -2.57815063e-01 -1.38995320e-01
3.47162671e-02 -7.89798945e-02 -8.53115320e-01 -8.63280222e-02
-3.70508373e-01 -1.00926614e+00 5.03060162e-01 3.04013699e-01
1.32418668e+00 -1.06156147e+00 -6.28677607e-01 1.89451844e-01
-2.68420935e-01 -8.72335315e-01 -1.28405215e-02 2.88721979e-01
-7.49864638e-01 -9.45642412e-01 -7.76646316e-01 -7.06502676e-01
2.45149717e-01 3.34852189e-01 4.57250446e-01 2.98574492e-02
-4.46265399e-01 5.99348545e-02 -7.05104530e-01 -3.84756714e-01
1.42718732e-01 1.74357042e-01 -2.75510907e-01 5.90053201e-01
3.39170992e-01 -6.26663506e-01 -5.04747510e-01 1.98574930e-01
-9.46833789e-01 2.30699833e-02 7.04214513e-01 1.23524463e+00
4.87937987e-01 5.77698350e-01 6.39378965e-01 -4.10771072e-01
5.76227546e-01 -4.51332957e-01 -3.05068672e-01 1.49003267e-01
-1.48840621e-01 -9.82377157e-02 6.16648793e-01 -3.28414887e-01
-1.30054951e+00 -1.31537706e-01 -2.71443307e-01 -2.48533234e-01
-2.79190242e-01 1.15012074e+00 2.97642529e-01 -5.95133901e-01
7.86308587e-01 5.28503060e-01 2.85316974e-01 -5.20177543e-01
-1.72612265e-01 8.58065069e-01 2.12075338e-01 -6.38282239e-01
5.93386114e-01 6.12284064e-01 1.73851386e-01 -1.36274040e+00
-8.59023571e-01 -3.83934438e-01 -5.72325826e-01 -3.13355476e-01
8.45070243e-01 -7.69672632e-01 -6.63331330e-01 9.53434646e-01
-8.64433289e-01 -6.27203956e-02 -8.37357417e-02 6.58716202e-01
-3.32019776e-01 4.22977507e-01 -4.88579541e-01 -8.06324780e-01
-5.36290348e-01 -8.62470806e-01 1.14016044e+00 3.28908920e-01
3.72466683e-01 -1.03020179e+00 -2.08860472e-01 -4.86876816e-02
1.07007480e+00 6.02524340e-01 9.49075103e-01 -4.22651708e-01
1.22027174e-01 -6.76031485e-02 -5.19308627e-01 3.82507324e-01
3.88757050e-01 4.03535329e-02 -1.06356359e+00 -1.98579520e-01
1.98262036e-01 -3.93313110e-01 1.09078372e+00 6.16966367e-01
1.40169632e+00 -2.97044933e-01 1.37026027e-01 8.37747574e-01
1.67975152e+00 2.97887981e-01 5.72413206e-01 3.22884411e-01
3.90978754e-01 5.64363182e-01 4.34533417e-01 5.29178917e-01
-3.18394378e-02 3.91166240e-01 2.00111702e-01 -5.02842307e-01
1.07746702e-02 4.54813361e-01 -1.23065859e-01 9.69205678e-01
-6.19402051e-01 1.39188215e-01 -7.81694889e-01 1.55310139e-01
-1.82859135e+00 -1.09588170e+00 -2.71452386e-02 1.70120537e+00
6.18665755e-01 -1.92145824e-01 -1.87616915e-01 5.00113785e-01
6.26846433e-01 3.14670801e-01 -1.74258068e-01 -1.96637347e-01
-4.93865848e-01 7.90888190e-01 4.83375549e-01 2.05355272e-01
-1.59926414e+00 6.98764622e-01 5.70209599e+00 1.12010622e+00
-1.66049838e+00 -6.11754647e-03 7.44559288e-01 5.69874167e-01
2.53289014e-01 -5.59420705e-01 -1.76723480e-01 2.81775057e-01
5.61302781e-01 1.48664668e-01 3.11885238e-01 5.28278530e-01
1.45682231e-01 -2.86798984e-01 -1.73170045e-01 1.08412969e+00
-1.37167107e-02 -1.23983634e+00 -8.23342055e-02 -1.51404649e-01
5.31138003e-01 -1.08653186e-02 1.56958684e-01 1.53749287e-01
-1.80731520e-01 -1.11273563e+00 4.67735916e-01 9.50256109e-01
1.05947769e+00 -8.81394625e-01 1.31219077e+00 1.25902995e-01
-1.46761012e+00 -2.00957298e-01 -7.47002661e-01 -1.07260771e-01
-3.54055375e-01 8.61595809e-01 -1.81669310e-01 1.13373077e+00
9.75506425e-01 1.04286742e+00 -5.86600900e-01 9.47962463e-01
2.74746716e-01 7.03760982e-01 -3.33183706e-01 -4.43204725e-03
4.33749110e-01 -4.46785480e-01 9.49059278e-02 1.15366316e+00
7.47462451e-01 3.92768741e-01 1.22532524e-01 7.47164071e-01
3.80444974e-01 2.35927984e-01 -4.55396533e-01 7.35281035e-02
1.88559115e-01 1.54774606e+00 -8.71836066e-01 -5.11921346e-02
-2.44351953e-01 3.78733367e-01 -1.10187322e-01 4.83556539e-01
-5.69521666e-01 -4.70436752e-01 3.58961672e-01 -3.31449419e-01
4.78179961e-01 -3.41363162e-01 -1.45313144e-01 -9.69795883e-01
-1.28211215e-01 -7.64239609e-01 1.97396874e-01 -7.35352278e-01
-1.42309928e+00 7.52216756e-01 -4.56440635e-02 -1.35064280e+00
3.87161463e-01 -8.59266222e-01 -4.46033865e-01 1.12181640e+00
-1.74773371e+00 -1.19496465e+00 -7.36980438e-01 7.81549692e-01
3.65158826e-01 -3.83149147e-01 9.28920567e-01 3.84249151e-01
-3.78653973e-01 9.89152715e-02 2.99182683e-01 3.03182930e-01
4.77345109e-01 -8.14407945e-01 -1.99608177e-01 4.36764479e-01
-3.85739386e-01 1.35738492e-01 3.44347298e-01 -4.43207592e-01
-1.23233449e+00 -1.00747752e+00 4.97901917e-01 7.17704952e-01
5.44415593e-01 -9.65304449e-02 -9.43652093e-01 4.96618031e-03
9.44186747e-02 3.37599784e-01 7.25120366e-01 -1.02358498e-01
-1.55864164e-01 -4.30981874e-01 -1.09713948e+00 5.80263659e-02
5.91097772e-01 -4.09306526e-01 -2.72069424e-01 1.37672365e-01
2.44638920e-01 -2.42630810e-01 -1.20590830e+00 6.25128031e-01
7.39951432e-01 -9.05645013e-01 7.96992719e-01 -2.91799545e-01
6.36859298e-01 -8.00785646e-02 -5.48813224e-01 -1.48079443e+00
-5.99615395e-01 7.64201656e-02 2.88960129e-01 7.38898814e-01
-8.01849142e-02 -7.90199339e-01 4.00978476e-01 -3.45184207e-01
-8.96070823e-02 -8.61336946e-01 -1.09434247e+00 -5.87155223e-01
-1.17025815e-01 -2.31206939e-01 5.24839461e-01 1.14282596e+00
-3.78153682e-01 -3.32584940e-02 -4.46319222e-01 1.05484940e-01
6.36876583e-01 2.97993869e-01 1.56171471e-01 -1.75034201e+00
1.44434050e-01 -5.20404458e-01 -5.76482534e-01 -3.30834121e-01
2.81153861e-02 -6.66432977e-01 -2.32970983e-01 -1.25030947e+00
-8.90071243e-02 -5.78007340e-01 -2.83936679e-01 6.18189573e-01
4.35273610e-02 3.37694794e-01 -2.84916069e-02 3.46153647e-01
-9.64513868e-02 7.50224054e-01 1.50990081e+00 -2.39477456e-01
1.40221208e-01 -9.16377604e-02 -3.28481585e-01 4.94945616e-01
8.68170261e-01 -9.75062400e-02 -1.59116849e-01 -1.93349138e-01
1.18430130e-01 2.13625655e-01 4.97217625e-01 -1.31093144e+00
2.16037869e-01 -2.35165343e-01 8.19210291e-01 -6.49990082e-01
1.66165218e-01 -9.17211771e-01 1.31065145e-01 4.17495847e-01
4.02753241e-03 -3.00247788e-01 3.14746052e-01 2.91770279e-01
-6.92683697e-01 4.14450280e-02 7.97969878e-01 3.62141170e-02
-9.65798974e-01 3.62764508e-01 -4.38422114e-01 -4.82598454e-01
6.72522426e-01 -3.42875719e-01 -3.81081879e-01 -1.73895404e-01
-7.63226867e-01 -1.02981947e-01 -2.31668770e-01 2.18879282e-01
7.31061757e-01 -1.56455314e+00 -8.53622675e-01 5.38627326e-01
1.04084924e-01 -1.64444968e-01 6.33519948e-01 1.12967956e+00
-7.96510458e-01 1.98615193e-01 -8.23275924e-01 -7.00365424e-01
-8.80364358e-01 1.39063418e-01 7.10405052e-01 -8.96690339e-02
-5.27274549e-01 4.98523235e-01 -2.29138523e-01 -3.83335650e-01
-1.00652277e-01 -2.76668757e-01 -6.89647436e-01 3.93006235e-01
1.64524049e-01 3.21024060e-01 4.06940490e-01 -7.29404747e-01
-1.97731316e-01 8.96755219e-01 4.06812847e-01 2.38853276e-01
1.84296024e+00 1.57919466e-01 -4.42126065e-01 2.39402249e-01
1.39246309e+00 -3.63819331e-01 -9.82419014e-01 -5.44314206e-01
-3.42845507e-02 -3.07905763e-01 2.57053673e-01 -6.00022376e-01
-1.37197673e+00 8.97424698e-01 1.00447166e+00 4.76150900e-01
1.45916307e+00 -5.22541463e-01 4.67347592e-01 4.48062360e-01
2.79854149e-01 -1.17114854e+00 -7.69039840e-02 6.66091323e-01
1.04221106e+00 -1.21851957e+00 7.58008957e-02 -3.70065749e-01
-2.74599105e-01 1.72120965e+00 3.68187994e-01 -1.67085603e-01
1.07422340e+00 8.02796260e-02 1.94701401e-03 -4.52220589e-01
-3.15357000e-01 -2.94022292e-01 5.92078984e-01 4.60701913e-01
5.58574796e-01 1.47453278e-01 -4.18995261e-01 5.23935974e-01
-1.45957455e-01 -7.70779932e-03 1.21747896e-01 9.05599475e-01
-6.56717420e-01 -6.08212292e-01 -5.95957160e-01 6.68423891e-01
-5.06598651e-01 2.21015550e-02 -8.97305235e-02 8.50791395e-01
4.44501430e-01 8.00333917e-01 1.12607181e-02 -5.00890851e-01
1.31600261e-01 9.95706990e-02 2.36952573e-01 -1.70740977e-01
-3.27954233e-01 3.94061476e-01 -6.95360824e-02 -3.84965986e-01
-9.46366668e-01 -4.01117235e-01 -6.35818303e-01 -3.15062493e-01
-3.77391756e-01 2.14710161e-01 5.76645792e-01 9.29987848e-01
1.44080922e-01 6.61735833e-01 8.99048030e-01 -1.09135699e+00
-4.69062120e-01 -1.29631972e+00 -1.06949770e+00 2.02970311e-01
3.62390667e-01 -1.08645809e+00 -4.50151861e-01 -5.49934246e-02] | [10.138111114501953, -0.704339325428009] |
34fcc7c1-560f-42d8-8d8a-3d0e857611f7 | magicpony-learning-articulated-3d-animals-in | 2211.12497 | null | https://arxiv.org/abs/2211.12497v3 | https://arxiv.org/pdf/2211.12497v3.pdf | MagicPony: Learning Articulated 3D Animals in the Wild | We consider the problem of predicting the 3D shape, articulation, viewpoint, texture, and lighting of an articulated animal like a horse given a single test image as input. We present a new method, dubbed MagicPony, that learns this predictor purely from in-the-wild single-view images of the object category, with minimal assumptions about the topology of deformation. At its core is an implicit-explicit representation of articulated shape and appearance, combining the strengths of neural fields and meshes. In order to help the model understand an object's shape and pose, we distil the knowledge captured by an off-the-shelf self-supervised vision transformer and fuse it into the 3D model. To overcome local optima in viewpoint estimation, we further introduce a new viewpoint sampling scheme that comes at no additional training cost. MagicPony outperforms prior work on this challenging task and demonstrates excellent generalisation in reconstructing art, despite the fact that it is only trained on real images. | ['Andrea Vedaldi', 'Christian Rupprecht', 'Tomas Jakab', 'Ruining Li', 'Shangzhe Wu'] | 2022-11-22 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_MagicPony_Learning_Articulated_3D_Animals_in_the_Wild_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_MagicPony_Learning_Articulated_3D_Animals_in_the_Wild_CVPR_2023_paper.pdf | cvpr-2023-1 | ['viewpoint-estimation'] | ['computer-vision'] | [ 2.29714662e-01 2.90720254e-01 2.33274177e-01 -3.86286706e-01
-6.01805925e-01 -8.47314596e-01 7.21254289e-01 -2.62917072e-01
-3.99212949e-02 2.42333665e-01 3.89504805e-02 2.21408844e-01
1.20937169e-01 -4.36484069e-01 -1.17099798e+00 -3.71744692e-01
3.75105888e-01 1.07630861e+00 3.97639841e-01 -6.87858164e-02
7.33634755e-02 7.14500904e-01 -1.55806208e+00 3.69109273e-01
2.52252847e-01 1.12609184e+00 3.62874866e-01 7.32347846e-01
2.99534351e-01 6.75173640e-01 -8.31269845e-02 -5.98596334e-01
4.34444189e-01 -1.86516881e-01 -9.34208095e-01 5.94283164e-01
8.89283955e-01 -4.95422125e-01 -1.45421416e-01 5.15045047e-01
1.92526907e-01 -1.96759403e-01 8.36450636e-01 -6.94475889e-01
-2.55349547e-01 -1.39990091e-01 -3.45514864e-01 -2.23805696e-01
3.30023646e-01 2.94150949e-01 1.03397095e+00 -1.01297617e+00
1.01272047e+00 1.16154242e+00 9.98025119e-01 3.57376277e-01
-1.60943878e+00 8.83718133e-02 1.33937821e-01 -2.99765617e-02
-1.02800465e+00 -5.75099707e-01 1.03728640e+00 -8.21894467e-01
9.10660803e-01 1.52552426e-01 8.95290315e-01 9.85996783e-01
1.68997973e-01 8.19760442e-01 1.11822605e+00 -4.38988805e-01
2.14813977e-01 2.13384368e-02 -5.52126765e-01 9.88236368e-01
-1.88365862e-01 2.49933288e-01 -4.18644547e-01 -9.98377055e-02
1.11401272e+00 -2.55333930e-01 -1.52319506e-01 -1.32661533e+00
-1.20585918e+00 4.41259384e-01 5.21428585e-01 -1.73472255e-01
-4.15351033e-01 2.13137805e-01 -4.95796017e-02 1.19242735e-01
6.68497324e-01 5.41637480e-01 -7.82893538e-01 8.00410062e-02
-8.04044008e-01 3.58559042e-01 7.89301991e-01 5.59771299e-01
8.00349176e-01 2.74881292e-02 3.40261072e-01 6.70173347e-01
4.10311133e-01 6.06252432e-01 -6.28440827e-02 -1.44267976e+00
1.42972529e-01 6.43413603e-01 1.61388144e-01 -7.98316717e-01
-2.87176728e-01 -5.07357001e-01 -2.38934904e-01 4.94346499e-01
5.35843074e-01 1.47478223e-01 -1.01332021e+00 1.35819519e+00
6.41148746e-01 2.75986921e-03 -4.30813819e-01 9.50965524e-01
4.30804491e-01 2.40882620e-01 -4.20311213e-01 2.02141568e-01
1.25470197e+00 -9.59598601e-01 3.98428328e-02 -7.26900458e-01
-4.81772562e-03 -8.35628986e-01 7.60754704e-01 6.07656956e-01
-1.40016329e+00 -5.61586738e-01 -9.01305556e-01 -3.11303556e-01
-1.35353416e-01 1.86370656e-01 4.66124028e-01 2.12077066e-01
-9.53394830e-01 9.05749977e-01 -9.85749185e-01 -1.99275583e-01
4.67133790e-01 4.69876558e-01 -5.93455076e-01 -1.54611051e-01
-3.26273650e-01 1.28195715e+00 -1.67309567e-02 1.08955652e-01
-1.00219476e+00 -7.65368223e-01 -8.96471024e-01 -3.83479565e-01
4.62352633e-01 -1.08073688e+00 1.29718387e+00 -1.29660416e+00
-1.66512775e+00 1.18949306e+00 3.02280728e-02 2.56283041e-02
7.34148324e-01 -2.31308937e-01 2.93870091e-01 2.51272291e-01
-1.31553233e-01 8.70239317e-01 1.17231774e+00 -1.62527072e+00
-9.92662311e-02 -6.61310017e-01 1.93513379e-01 3.13993573e-01
3.49061430e-01 -4.72168207e-01 -4.66979206e-01 -6.04641080e-01
2.16842353e-01 -9.61206794e-01 -2.13314787e-01 6.12516463e-01
-1.67604014e-01 1.48765892e-01 7.60746241e-01 -7.70838141e-01
1.22160472e-01 -1.91465724e+00 6.54641926e-01 1.53945446e-01
-1.50460461e-02 1.20028265e-01 -8.99580866e-02 2.37652302e-01
1.37295499e-01 -4.96470034e-01 -2.12131649e-01 -5.80577731e-01
-2.33763009e-01 6.03019893e-01 -3.99713144e-02 6.83050454e-01
4.88081664e-01 9.07079160e-01 -7.60086238e-01 -2.70603955e-01
4.56438333e-01 7.72703528e-01 -9.42870855e-01 4.94892508e-01
-6.61411345e-01 8.73819888e-01 -3.40616614e-01 5.13385057e-01
4.60033417e-01 -3.70664567e-01 2.38805994e-01 -4.06991035e-01
2.69885119e-02 2.49555171e-01 -1.13741338e+00 2.06727004e+00
-6.14600241e-01 4.59048152e-01 4.78546530e-01 -1.09262228e+00
7.57214308e-01 2.19598636e-01 3.88022244e-01 -2.58767813e-01
2.27296129e-01 2.79934436e-01 -2.44217038e-01 -6.51664197e-01
6.83538336e-03 -3.74463588e-01 4.27786797e-01 2.93436825e-01
4.05158699e-01 -6.36407554e-01 -4.40749735e-01 -2.11899102e-01
8.82661521e-01 1.00311816e+00 1.89070761e-01 -1.74006507e-01
2.26609081e-01 9.55896676e-02 5.53564914e-02 1.93164080e-01
4.76040870e-01 1.08496141e+00 2.23503813e-01 -7.49799669e-01
-1.32585537e+00 -1.03774333e+00 -8.04543197e-02 8.84584367e-01
1.72360968e-02 -1.04818404e-01 -6.60667419e-01 -5.92496812e-01
1.90514833e-01 3.43753546e-01 -7.46742547e-01 1.91194192e-01
-6.26452386e-01 -1.57093018e-01 1.01925088e-02 6.10079944e-01
7.25485981e-02 -9.72095728e-01 -1.13498580e+00 9.11552012e-02
-1.00947052e-01 -1.19231546e+00 -4.21597332e-01 3.02117079e-01
-9.67137277e-01 -1.04564726e+00 -5.37633002e-01 -7.47046947e-01
8.96496534e-01 -5.31133637e-02 1.38546598e+00 6.58070073e-02
-4.50165778e-01 6.29321873e-01 -8.60639364e-02 -2.52640784e-01
-3.10453802e-01 -2.03705370e-01 -2.97385335e-01 1.12685136e-01
-2.62179524e-01 -8.57856154e-01 -7.30526030e-01 3.39334786e-01
-4.73547220e-01 2.67594516e-01 6.27296209e-01 8.57404351e-01
7.55276024e-01 -3.81402582e-01 -1.54509425e-01 -7.35355198e-01
-2.44747847e-01 -2.27943212e-01 -4.82715219e-01 1.36941895e-01
-2.46680826e-01 2.66739964e-01 5.67418039e-01 -4.26042259e-01
-9.10464942e-01 7.48390675e-01 -1.00937583e-01 -7.27587223e-01
-3.19323033e-01 1.18118495e-01 -9.57785249e-02 -2.35198587e-01
5.82755804e-01 1.70255825e-02 4.55483466e-01 -9.14152861e-01
4.52840298e-01 2.57625043e-01 7.66838372e-01 -5.25691628e-01
7.16078937e-01 8.30916762e-01 1.57014996e-01 -6.68238997e-01
-1.01416600e+00 -3.40121508e-01 -1.19356871e+00 -1.49739221e-01
6.95252359e-01 -1.00049174e+00 -6.23023152e-01 4.12204891e-01
-1.29161489e+00 -4.97833759e-01 -4.25577253e-01 3.20546001e-01
-1.08545411e+00 1.97863162e-01 -3.17229837e-01 -6.07552946e-01
-3.84232700e-02 -9.57946062e-01 1.57509851e+00 -4.07733142e-01
-1.79376513e-01 -9.85089958e-01 6.62910938e-02 7.29581773e-01
1.37100041e-01 4.54730898e-01 9.05353069e-01 -2.82973737e-01
-7.28365839e-01 -8.14006403e-02 9.02808458e-02 5.48005998e-01
-1.31095573e-01 -4.98456955e-02 -1.17400312e+00 -3.11207980e-01
2.09761202e-01 -5.76445341e-01 5.45655191e-01 3.44018251e-01
9.74672079e-01 -3.13870877e-01 -8.37843865e-02 8.06250811e-01
1.41508436e+00 -3.26760471e-01 2.38593936e-01 1.53250128e-01
9.09647763e-01 8.32215428e-01 2.95462966e-01 2.70736843e-01
4.11177993e-01 9.80481803e-01 8.18374038e-01 -1.22058541e-01
-4.03317690e-01 -4.53079343e-01 3.08766067e-01 5.69664776e-01
-2.53221869e-01 8.24294835e-02 -8.68198216e-01 3.80336076e-01
-1.49090850e+00 -6.61792159e-01 1.67348847e-01 2.00383735e+00
6.92133665e-01 1.64635733e-01 1.61231846e-01 1.85972210e-02
1.30189955e-01 8.46210215e-03 -7.16147184e-01 -3.89350057e-01
1.83191448e-01 2.20737532e-01 3.45381230e-01 6.36664450e-01
-8.99590731e-01 5.82271338e-01 6.63591051e+00 3.02690476e-01
-1.24479175e+00 -8.16611126e-02 5.19594908e-01 1.28040910e-01
-4.41857219e-01 1.98748529e-01 -3.06259871e-01 6.80254474e-02
5.81959903e-01 5.36054730e-01 7.55060434e-01 8.01291943e-01
-1.22830257e-01 7.39945890e-03 -1.57958281e+00 6.82457209e-01
5.01221895e-01 -1.34162152e+00 -4.99082617e-02 9.23789963e-02
5.02472997e-01 2.60055035e-01 -1.47653772e-02 -2.20460817e-01
2.13691458e-01 -1.00575805e+00 1.11551583e+00 5.43711424e-01
9.07975078e-01 -2.87861586e-01 5.21110594e-01 6.22053206e-01
-8.22111070e-01 -6.73014894e-02 -3.70241627e-02 -1.05903491e-01
1.90384109e-02 1.37082651e-01 -9.74752784e-01 5.07543921e-01
5.26517034e-01 7.66547441e-01 -5.97259521e-01 7.06044436e-01
-4.59907092e-02 2.45208591e-01 -5.11457086e-01 5.49291790e-01
6.13035671e-02 -8.50304961e-02 5.12809515e-01 7.14218974e-01
3.96638773e-02 5.33622801e-02 3.02686334e-01 8.27433228e-01
9.83640924e-02 -1.83566660e-01 -6.00635171e-01 3.14617217e-01
-6.61573112e-02 1.10197604e+00 -5.56319475e-01 1.85536873e-02
-2.73299485e-01 8.76146674e-01 4.60411102e-01 6.66458309e-02
-3.15553010e-01 5.33908129e-01 1.85783580e-01 7.28523314e-01
1.04468715e+00 -2.38192216e-01 -3.65389943e-01 -1.15975177e+00
1.30857170e-01 -8.32256019e-01 -6.63108155e-02 -1.20976400e+00
-1.22382426e+00 4.80343848e-01 4.25238907e-02 -1.22072589e+00
-4.00708765e-01 -9.34981525e-01 -3.35310370e-01 6.73623323e-01
-1.22691667e+00 -1.55688357e+00 -2.00909913e-01 3.56064439e-01
7.56129324e-01 1.07835777e-01 8.98034453e-01 -2.12553844e-01
1.08074300e-01 5.99223338e-02 2.84102280e-02 -1.23476915e-01
4.09257799e-01 -1.29332840e+00 4.34681237e-01 3.35339546e-01
4.27290529e-01 2.72427022e-01 7.68931985e-01 -3.65659118e-01
-1.67943180e+00 -7.90509403e-01 5.50469041e-01 -1.07669926e+00
4.54474092e-01 -4.74363446e-01 -6.50431693e-01 7.22158790e-01
-8.81514177e-02 4.29233640e-01 2.18097642e-01 3.51920351e-02
-5.56927204e-01 -5.47942221e-02 -1.18680608e+00 2.53949553e-01
1.02835810e+00 -6.23692691e-01 -7.42162645e-01 2.58047193e-01
1.70175955e-01 -7.48145282e-01 -8.42410803e-01 4.36360240e-01
9.15628910e-01 -1.04542649e+00 1.24300826e+00 -4.78573352e-01
6.08666956e-01 -2.47809976e-01 -2.44671762e-01 -1.34159231e+00
-1.92594960e-01 -5.18566310e-01 -1.39896423e-01 6.58229470e-01
1.22468926e-01 -1.42327622e-01 9.24580455e-01 3.75526994e-01
-3.44126485e-02 -1.02508390e+00 -8.79740775e-01 -4.53926921e-01
5.87746920e-03 -1.21301316e-01 2.43588939e-01 6.01612389e-01
-4.83854204e-01 6.14203930e-01 -4.86608624e-01 2.94137672e-02
7.56643653e-01 3.79246145e-01 9.85044241e-01 -1.50393438e+00
-6.15048170e-01 1.30205929e-01 -5.76673746e-01 -1.20550549e+00
2.12895754e-03 -7.84526348e-01 2.14287460e-01 -1.20258510e+00
9.57842991e-02 -1.97958335e-01 2.42436975e-01 4.79559630e-01
2.39125952e-01 4.87679660e-01 1.40252918e-01 3.26625258e-01
-3.74710947e-01 5.00068307e-01 1.43734360e+00 1.60069540e-01
1.95392072e-01 1.89805645e-02 -2.75010586e-01 1.10922384e+00
3.03672284e-01 -3.91092598e-01 -4.77011412e-01 -9.56378102e-01
1.81083336e-01 2.75722653e-01 8.10781240e-01 -7.64682591e-01
-9.53736529e-03 -9.65610817e-02 6.36534452e-01 -4.95414287e-01
8.57536495e-01 -1.13149953e+00 5.06458640e-01 3.35037529e-01
-3.75765771e-01 1.43681079e-01 -4.76775877e-02 7.41005182e-01
1.87809646e-01 -7.38268485e-03 8.35674584e-01 -4.41356212e-01
-3.91243130e-01 1.90189093e-01 2.01100737e-01 1.33439973e-01
6.80926204e-01 -4.56760913e-01 3.20466310e-02 -1.63471445e-01
-9.32703376e-01 -2.96097230e-02 1.09536052e+00 2.55600452e-01
5.39444327e-01 -1.00063789e+00 -7.02484906e-01 6.30595446e-01
-5.72148599e-02 5.79142630e-01 4.31258790e-02 6.64993465e-01
-7.61574149e-01 6.11345023e-02 -4.17784274e-01 -9.48295891e-01
-1.16831088e+00 4.45407331e-01 5.23102880e-01 -1.16374996e-02
-8.88901472e-01 7.90170133e-01 2.23017395e-01 -6.62030101e-01
6.84998333e-02 -4.81591895e-02 1.24157473e-01 -2.83548355e-01
4.32824977e-02 2.04522368e-02 2.56757975e-01 -9.13129270e-01
-2.87981808e-01 9.81728196e-01 1.40187070e-01 -2.14719966e-01
1.65594792e+00 2.90599577e-02 -2.99134366e-02 5.92415750e-01
1.06303906e+00 -4.30718698e-02 -1.92929149e+00 -2.70252615e-01
-2.37933025e-01 -7.87172258e-01 4.46688943e-02 -9.25229311e-01
-1.07791770e+00 9.59564984e-01 2.99302250e-01 -1.73132405e-01
6.47276223e-01 3.14053148e-01 5.36013842e-01 3.27440798e-01
4.16720688e-01 -8.89651358e-01 3.61820757e-01 3.33055496e-01
1.14738536e+00 -1.03777313e+00 4.00886446e-01 -3.78878683e-01
-6.49615943e-01 1.13415980e+00 4.27869886e-01 -5.61940730e-01
6.73466504e-01 2.20841706e-01 5.01775518e-02 -4.92871314e-01
-8.09292018e-01 1.33827820e-01 6.38823330e-01 7.72490680e-01
3.93651687e-02 -1.03787020e-01 4.96712506e-01 -8.97451770e-03
-2.90768474e-01 -2.64835432e-02 1.79199651e-01 8.19336116e-01
-2.68645346e-01 -1.02016592e+00 -2.28701800e-01 3.41399312e-01
-3.96584094e-01 2.39258349e-01 -5.72649896e-01 5.90695798e-01
4.65449065e-01 3.69007707e-01 1.98640093e-01 -2.62198806e-01
5.05133748e-01 2.70658899e-02 9.98425007e-01 -7.62415648e-01
-4.80592638e-01 2.42082298e-01 7.00417459e-02 -6.42454445e-01
-7.33970463e-01 -8.02884877e-01 -6.82637632e-01 1.38127329e-02
-2.69114375e-01 -2.80123144e-01 7.03633428e-01 1.13669384e+00
3.09108615e-01 2.02407494e-01 6.06916070e-01 -1.55967057e+00
-7.31929064e-01 -5.81368029e-01 -2.71402657e-01 4.94312972e-01
6.66906536e-01 -9.15654242e-01 -2.92394549e-01 3.75177383e-01] | [8.263776779174805, -2.8350346088409424] |
40fdf84a-fcd9-41a6-8ad9-e35ae4fa1752 | an-empirical-study-of-finding-similar | 2111.08322 | null | https://arxiv.org/abs/2111.08322v1 | https://arxiv.org/pdf/2111.08322v1.pdf | An Empirical Study of Finding Similar Exercises | Education artificial intelligence aims to profit tasks in the education domain such as intelligent test paper generation and consolidation exercises where the main technique behind is how to match the exercises, known as the finding similar exercises(FSE) problem. Most of these approaches emphasized their model abilities to represent the exercise, unfortunately there are still many challenges such as the scarcity of data, insufficient understanding of exercises and high label noises. We release a Chinese education pre-trained language model BERT$_{Edu}$ for the label-scarce dataset and introduce the exercise normalization to overcome the diversity of mathematical formulas and terms in exercise. We discover new auxiliary tasks in an innovative way depends on problem-solving ideas and propose a very effective MoE enhanced multi-task model for FSE task to attain better understanding of exercises. In addition, confidence learning was utilized to prune train-set and overcome high noises in labeling data. Experiments show that these methods proposed in this paper are very effective. | ['Xihua Li', 'Tongwen Huang'] | 2021-11-16 | null | null | null | null | ['paper-generation'] | ['natural-language-processing'] | [ 4.16417062e-01 -3.64052951e-01 1.32444099e-01 -3.62793207e-01
-6.07653975e-01 -6.59048200e-01 8.79440829e-02 -3.92715409e-02
-5.16686857e-01 1.06784689e+00 1.05974443e-01 -2.74799854e-01
-6.77317262e-01 -8.90013039e-01 -7.05917656e-01 -2.27043033e-01
5.67535400e-01 2.29510546e-01 1.24015965e-01 -5.00508547e-01
5.48441052e-01 9.12369341e-02 -1.76140034e+00 2.64301956e-01
1.48010933e+00 5.44335544e-01 6.18348539e-01 5.18526673e-01
-8.52659404e-01 1.08047664e+00 -9.11711514e-01 -7.01643944e-01
-5.29403798e-02 -1.04908180e+00 -1.02460301e+00 6.37079477e-02
3.94529104e-01 2.00926229e-01 -1.12405546e-01 1.17300093e+00
5.72042525e-01 5.78370094e-01 7.71755338e-01 -1.30066037e+00
-1.09180140e+00 8.38959217e-01 -7.20760465e-01 3.12089473e-01
2.36968309e-01 -3.34655374e-01 7.40527928e-01 -6.08239710e-01
4.55613405e-01 9.53137636e-01 8.04377556e-01 7.23490715e-01
-5.58281720e-01 -7.25738704e-01 -1.29856229e-01 6.88345015e-01
-1.24029255e+00 1.16118833e-01 1.03537905e+00 -3.51775885e-01
3.72607052e-01 3.92832309e-01 3.96204770e-01 7.70641744e-01
-9.78919938e-02 8.38044822e-01 1.59193873e+00 -8.30622554e-01
-3.30359131e-01 4.88472790e-01 3.43034416e-01 8.90525222e-01
5.63284606e-02 -3.91797781e-01 -4.05488729e-01 6.01361156e-01
5.00684559e-01 2.04492435e-01 -2.31003717e-01 1.55457065e-01
-8.89237940e-01 7.31860220e-01 1.19564727e-01 5.22532940e-01
1.42580524e-01 -3.24214369e-01 3.84974808e-01 5.35913885e-01
1.24121029e-02 5.65030396e-01 -8.88753772e-01 -1.51687562e-01
-8.44671309e-01 1.88510552e-01 5.45908213e-01 1.11487162e+00
6.49590850e-01 2.64197111e-01 -3.84073436e-01 1.00267053e+00
1.22110695e-01 3.51997375e-01 8.48927617e-01 -6.28272355e-01
7.07239389e-01 8.74536216e-01 -4.52811629e-01 -9.78922725e-01
-2.84478068e-01 -8.72904778e-01 -7.53944814e-01 1.88959703e-01
4.45542961e-01 -4.72156912e-01 -7.59562492e-01 1.78010499e+00
1.64274108e-02 5.48749924e-01 1.85292885e-01 3.99446011e-01
1.31928730e+00 5.16161919e-01 2.70153612e-01 -1.98915020e-01
1.39845896e+00 -1.33555186e+00 -1.10358870e+00 -3.41260135e-02
7.62133718e-01 -1.13781643e+00 1.34214568e+00 5.47246993e-01
-1.01026154e+00 -1.14843798e+00 -9.68235016e-01 -1.36322364e-01
-8.66378486e-01 4.73611295e-01 6.10196948e-01 9.21593368e-01
-5.08455753e-01 6.73868120e-01 -4.61910665e-03 2.19064001e-02
5.22255778e-01 1.94848210e-01 9.12837163e-02 -1.30509466e-01
-1.07634330e+00 9.43279743e-01 7.04514921e-01 -6.20475747e-02
-5.02483666e-01 -8.31282079e-01 -9.15217638e-01 2.10723266e-01
8.07632685e-01 -2.96253294e-01 1.10234594e+00 -9.09954965e-01
-1.54742026e+00 7.20234275e-01 1.70998901e-01 -4.61847521e-02
3.95731330e-01 -6.53313249e-02 -6.95939720e-01 -3.32939833e-01
1.01293504e-01 1.60751238e-01 1.63816422e-01 -8.74657333e-01
-7.85778761e-01 -2.26779833e-01 -4.79993373e-02 3.30430478e-01
-5.51772714e-01 8.16010591e-03 -2.24330127e-01 -6.35242701e-01
1.10637464e-01 -6.67838454e-01 -7.38059357e-02 -4.52204496e-01
-6.56389669e-02 -7.09524035e-01 7.27580070e-01 -8.12938631e-01
1.56016648e+00 -1.60528719e+00 -1.90508962e-02 -5.71978763e-02
4.15055119e-02 4.24267143e-01 -2.35458836e-01 -6.89622015e-02
-1.08040646e-01 3.36590946e-01 -6.59193993e-02 -1.15115931e-02
1.36332735e-01 9.24558863e-02 9.70160812e-02 -2.27583557e-01
1.19187087e-01 8.08849454e-01 -8.85654807e-01 -9.72245753e-01
3.73044938e-01 -9.27856937e-02 -5.18857956e-01 3.35324228e-01
-3.00252438e-01 4.48518276e-01 -4.86953169e-01 6.34911716e-01
6.22568250e-01 -3.51774283e-02 -8.96740034e-02 -3.90958749e-02
-2.11267918e-02 8.16163570e-02 -1.79586649e+00 2.00710630e+00
-6.77135408e-01 3.64584059e-01 -3.20700109e-01 -1.39208937e+00
1.07211220e+00 1.98137715e-01 2.47527614e-01 -7.61078060e-01
2.46640176e-01 2.03886598e-01 2.78321236e-01 -1.12276793e+00
4.94598776e-01 -2.01719254e-01 -6.35087565e-02 2.14932159e-01
3.62332553e-01 -2.76416034e-01 6.36516094e-01 2.75184605e-02
6.67981088e-01 7.93522000e-01 6.83127865e-02 -2.12977111e-01
1.05846536e+00 7.73416162e-02 6.65658593e-01 5.91385305e-01
-8.75022933e-02 2.97809005e-01 2.99604447e-03 1.49183115e-02
-6.46695495e-01 -5.94577610e-01 1.14438180e-02 1.30932379e+00
-1.54349089e-01 -3.93823981e-01 -6.56538308e-01 -7.08111048e-01
-4.38877374e-01 7.42408156e-01 -3.09238106e-01 -2.85052042e-02
-5.42047501e-01 -6.69973671e-01 4.31702107e-01 3.91036510e-01
1.00385344e+00 -1.28111124e+00 -2.29363561e-01 3.19516152e-01
-2.08217263e-01 -8.73164654e-01 -3.77029568e-01 2.73312777e-01
-6.89977229e-01 -1.08657610e+00 -5.57350039e-01 -1.19956350e+00
5.90269685e-01 7.88064748e-02 1.22959328e+00 3.24979186e-01
-3.44682932e-01 -1.75939370e-02 -3.76600444e-01 -7.63041258e-01
-1.74957037e-01 1.56413376e-01 -3.42902690e-01 -5.18476665e-01
7.21814573e-01 -4.14576292e-01 6.19261563e-02 -1.96451411e-01
-8.62106264e-01 -2.34180111e-02 5.35657942e-01 8.53121996e-01
4.83379751e-01 8.18235457e-01 7.94995904e-01 -1.29137850e+00
8.29560757e-01 -3.90699804e-01 -3.87782991e-01 7.37186670e-01
-7.16667831e-01 2.54217684e-01 6.79785371e-01 -5.67768097e-01
-1.37051976e+00 -2.80349534e-02 -3.03464711e-01 -1.75589174e-01
-3.02745283e-01 8.51237297e-01 -4.62856233e-01 -1.46871105e-01
5.11228561e-01 3.45620692e-01 -3.11106384e-01 -6.74474716e-01
1.07458927e-01 5.10545731e-01 4.67758566e-01 -1.16363931e+00
7.94262052e-01 -4.94581312e-01 2.56927609e-01 -5.32390594e-01
-1.25489020e+00 -2.64967382e-01 -6.06401026e-01 -3.45758319e-01
6.74235940e-01 -7.71592855e-01 -1.01899838e+00 3.73137563e-01
-9.83203650e-01 -1.34879053e-01 -2.09499061e-01 5.74558020e-01
-2.06502117e-02 3.27268898e-01 -4.92844880e-01 -8.96649837e-01
-2.24712007e-02 -1.20106435e+00 1.47167072e-01 1.03496015e+00
1.84966743e-01 -1.09892499e+00 1.50427353e-02 8.49449337e-01
3.92114431e-01 1.02051228e-01 1.18618095e+00 -8.01029980e-01
-3.35545272e-01 4.49214250e-01 -1.97117194e-03 4.55571234e-01
2.48652741e-01 -6.90215975e-02 -7.60007381e-01 2.84619004e-01
8.40148553e-02 -6.65341020e-01 7.51193583e-01 2.06758276e-01
1.68081069e+00 -3.79450396e-02 4.68631014e-02 5.61751187e-01
1.68787384e+00 3.03951651e-01 6.92025959e-01 3.97777408e-01
8.16998601e-01 6.93900228e-01 6.45329595e-01 -1.51732042e-02
3.01750153e-01 3.43238622e-01 -1.28469795e-01 -1.32733220e-02
-3.71522099e-01 -2.66718417e-01 5.61231151e-02 1.51040077e+00
-1.48800150e-01 -2.42407098e-01 -6.42650723e-01 5.80093086e-01
-1.45560133e+00 -1.02737749e+00 -6.24957681e-01 1.79950249e+00
1.29500246e+00 2.59035021e-01 -2.38217011e-01 2.56200671e-01
6.45034373e-01 -2.32292891e-01 -6.10883161e-02 -5.04670560e-01
-2.72767901e-01 1.05190122e+00 1.79825574e-01 4.29640621e-01
-1.04286885e+00 1.03893471e+00 5.10240555e+00 1.32953250e+00
-5.94562054e-01 3.21576029e-01 3.74231756e-01 3.39680970e-01
-2.26950154e-01 -2.45937049e-01 -1.08992434e+00 5.48447132e-01
9.21401083e-01 -1.93213195e-01 3.14489245e-01 7.57100940e-01
-3.27612348e-02 1.06581405e-01 -5.36028743e-01 9.46831584e-01
4.91117477e-01 -1.20932770e+00 2.10985560e-02 -4.33091938e-01
1.13416958e+00 -8.06847274e-01 -5.37206866e-02 1.20032644e+00
3.76541287e-01 -1.11468315e+00 2.26705268e-01 6.55512989e-01
4.75328445e-01 -1.00809574e+00 9.06485140e-01 4.35716301e-01
-1.13950741e+00 4.85333540e-02 -5.02883971e-01 -4.97154206e-01
-3.08660626e-01 9.41256583e-02 -4.72733557e-01 9.53223348e-01
5.10576785e-01 4.89701658e-01 -1.06756210e+00 1.18157101e+00
-4.77667063e-01 5.42560577e-01 3.76408175e-02 -4.27607536e-01
7.57445693e-02 -6.62128448e-01 -1.62010863e-01 1.09983933e+00
6.35639668e-01 4.16596264e-01 4.67557877e-01 9.32546020e-01
-3.15926462e-01 4.04579669e-01 -3.62083286e-01 -1.36376232e-01
3.23286533e-01 1.35003364e+00 -7.58015156e-01 -2.73047954e-01
-5.75882137e-01 8.76618207e-01 2.59114712e-01 3.43853235e-02
-9.62616265e-01 -8.06690812e-01 4.08468656e-02 -2.95545220e-01
7.00702742e-02 4.96845953e-02 -5.11974752e-01 -9.58965003e-01
-1.78138599e-01 -1.07637715e+00 5.37584484e-01 -8.14033151e-01
-1.23963416e+00 4.01979744e-01 -2.22516924e-01 -1.19146764e+00
8.70115086e-02 -8.01414907e-01 -7.84968734e-01 8.73113692e-01
-1.60164058e+00 -1.00132406e+00 -6.05271339e-01 8.06402504e-01
1.04770589e+00 -4.54080135e-01 7.00228274e-01 6.90668583e-01
-8.34741414e-01 7.72625804e-01 4.84142862e-02 3.25754911e-01
8.24088931e-01 -1.54664528e+00 -6.85775042e-01 1.07834625e+00
4.23191935e-01 4.39146131e-01 4.76824284e-01 -6.37775719e-01
-1.05923223e+00 -9.19594824e-01 9.48960483e-01 -2.67430425e-01
4.79354739e-01 1.80704489e-01 -9.03484106e-01 5.80281734e-01
4.31187063e-01 -4.29532707e-01 1.11023879e+00 2.37799987e-01
9.70660672e-02 -2.16608688e-01 -9.77121949e-01 7.00794458e-01
8.67692232e-01 -2.40477249e-01 -1.12654006e+00 4.73424554e-01
5.84846795e-01 -6.43964767e-01 -7.80324042e-01 4.31182534e-01
-3.44630587e-03 -5.48006356e-01 8.87340307e-01 -9.90428090e-01
9.10799026e-01 -3.44668925e-01 1.64328620e-01 -1.29510212e+00
-3.60817969e-01 -3.04545969e-01 7.56124258e-02 1.92190099e+00
4.39614415e-01 1.72412440e-01 9.11304355e-01 4.16601747e-01
-4.27182198e-01 -5.72748065e-01 -4.00138289e-01 -4.93953913e-01
1.03384167e-01 -3.11458975e-01 6.18267536e-01 1.45404303e+00
-2.45025739e-01 8.09627473e-01 -3.38649690e-01 -1.15335412e-01
4.34518993e-01 1.80616170e-01 5.77378273e-01 -1.47840130e+00
-2.67657280e-01 -5.18189728e-01 1.27655491e-01 -7.15419054e-01
3.58332336e-01 -1.08891499e+00 -2.11590137e-02 -1.45740974e+00
3.37880939e-01 -5.22168815e-01 -6.66071296e-01 4.37996179e-01
-5.99442184e-01 1.50442019e-01 1.93034440e-01 -5.17811120e-01
-8.35094869e-01 3.74617100e-01 1.63196087e+00 -1.79268956e-01
-1.96375936e-01 1.99522778e-01 -1.01850367e+00 7.48737574e-01
7.21280754e-01 -5.72370768e-01 -5.70286870e-01 -5.12989283e-01
8.17962661e-02 -8.78084265e-03 2.78498698e-02 -1.33866572e+00
4.88575906e-01 -4.86514986e-01 7.65022576e-01 -6.95902586e-01
-9.41093341e-02 -9.58315432e-01 -1.09771393e-01 3.90895814e-01
-4.93812770e-01 1.04435422e-01 3.60897213e-01 2.67825216e-01
-1.99371800e-01 -1.11779070e+00 5.56669652e-01 -5.28694808e-01
-9.64426875e-01 3.91366147e-03 -1.07826274e-02 4.57794547e-01
8.87737215e-01 -1.18678078e-01 -3.02173167e-01 2.96463985e-02
-5.87635517e-01 4.62273717e-01 -3.07848454e-01 4.18174237e-01
4.27465796e-01 -1.30190802e+00 -9.36841130e-01 -9.98455212e-02
-1.97037578e-01 7.34760687e-02 6.53942227e-01 5.60170770e-01
-5.54763019e-01 4.24627751e-01 -5.40671647e-01 -2.03353465e-01
-1.51826906e+00 6.01318955e-01 1.88093171e-01 -7.91391969e-01
-1.41421035e-02 1.10010195e+00 -4.83993918e-01 -9.53040302e-01
4.60451722e-01 -2.56862968e-01 -8.26382041e-01 1.62348613e-01
4.79132384e-01 4.85028923e-01 9.07603558e-03 -2.47803867e-01
9.27249938e-02 7.85911202e-01 -6.85612559e-02 2.37385869e-01
1.43997324e+00 -5.81229664e-03 -9.60504636e-02 2.68942863e-01
8.73377025e-01 2.31687978e-01 -7.44936943e-01 -3.39278102e-01
3.16856712e-01 -2.60524213e-01 -3.68267223e-02 -1.34525812e+00
-1.02784991e+00 1.06782281e+00 6.62118018e-01 -2.49511540e-01
1.15993989e+00 -6.13467872e-01 6.59578502e-01 5.42630672e-01
8.08620453e-02 -1.40034556e+00 2.83532590e-01 7.00697124e-01
4.12904412e-01 -1.39968586e+00 1.20962963e-01 -5.33250391e-01
-5.24484992e-01 1.37971139e+00 1.19702876e+00 1.83140367e-01
3.89943898e-01 2.60457397e-01 -1.29285648e-01 -4.87004891e-02
-1.90879032e-01 -4.38772082e-01 4.81808603e-01 5.66594660e-01
9.37541604e-01 -1.84252724e-01 -9.12676096e-01 1.16795671e+00
-4.15935367e-01 2.63988048e-01 5.48499763e-01 9.03912783e-01
-4.49327737e-01 -1.46639431e+00 -2.98395842e-01 5.03629148e-01
-6.70283258e-01 -2.97794253e-01 -8.18736702e-02 8.53950381e-01
9.22463715e-01 1.03773391e+00 -3.31276923e-01 -3.62381607e-01
3.70869011e-01 5.42450011e-01 7.70470321e-01 -8.49549949e-01
-9.76964474e-01 -1.82575330e-01 -1.20420724e-01 3.15707773e-01
-6.06292605e-01 -1.91878587e-01 -1.28858304e+00 -1.26377270e-01
-5.29512584e-01 6.03041351e-01 4.13641185e-01 1.40072250e+00
-2.02115819e-01 1.45949829e+00 3.41567427e-01 1.87999621e-01
-7.53665090e-01 -1.18162477e+00 -6.91317379e-01 7.31998444e-01
-3.60573083e-01 -5.74193835e-01 -2.43971884e-01 2.66359568e-01] | [10.043954849243164, 7.278232574462891] |
0ddbb0f7-ae8c-4232-a811-653c40f9a679 | efficient-hybrid-modeling-and-sorption-model | 2303.13555 | null | https://arxiv.org/abs/2303.13555v3 | https://arxiv.org/pdf/2303.13555v3.pdf | Efficient hybrid modeling and sorption model discovery for non-linear advection-diffusion-sorption systems: A systematic scientific machine learning approach | This study presents a systematic machine learning approach for creating efficient hybrid models and discovering sorption uptake models in non-linear advection-diffusion-sorption systems. It demonstrates an effective method to train these complex systems using gradient based optimizers, adjoint sensitivity analysis, and JIT-compiled vector Jacobian products, combined with spatial discretization and adaptive integrators. Sparse and symbolic regression were employed to identify missing functions in the artificial neural network. The robustness of the proposed method was tested on an in-silico data set of noisy breakthrough curve observations of fixed-bed adsorption, resulting in a well-fitted hybrid model. The study successfully reconstructed sorption uptake kinetics using sparse and symbolic regression, and accurately predicted breakthrough curves using identified polynomials, highlighting the potential of the proposed framework for discovering sorption kinetic law structures. | ['Chris Rackauckas', 'Idelfonso B. R. Nogueira', 'Ana Mafalda Ribeiro', 'Carine M. Rebello', 'Erbet Costa', 'Vinicius V. Santana'] | 2023-03-22 | null | null | null | null | ['model-discovery'] | ['miscellaneous'] | [ 3.37613449e-02 -5.49348295e-01 -1.38981476e-01 1.58546083e-02
-4.56894606e-01 -2.77184814e-01 1.58579364e-01 6.14442110e-01
-5.24760067e-01 1.23769152e+00 -4.86988634e-01 -7.16055751e-01
-9.65904593e-01 -6.06445253e-01 -8.30213785e-01 -1.01576388e+00
-7.14718103e-01 8.23695064e-01 -3.49550128e-01 -1.07930338e+00
2.86799669e-01 9.17404294e-01 -1.22489941e+00 6.11118460e-03
1.27058589e+00 8.43538880e-01 3.90347213e-01 6.03249609e-01
-7.14008063e-02 8.47051263e-01 -8.38577189e-03 2.94993490e-01
1.19264446e-01 3.11980825e-02 -5.08147895e-01 -4.60721195e-01
-3.36467922e-01 -1.57306910e-01 -1.68550059e-01 5.41055739e-01
5.62318385e-01 2.35880643e-01 1.01139545e+00 -5.93761206e-01
-3.37864965e-01 5.61647236e-01 -2.86718935e-01 -4.79635000e-02
2.98029751e-01 1.49780244e-01 5.74394703e-01 -1.42160833e+00
2.03064144e-01 8.00908685e-01 1.00834215e+00 -1.48069948e-01
-1.66423392e+00 -3.57419640e-01 -2.65077621e-01 5.30931875e-02
-1.87424362e+00 -2.69705951e-01 3.70567739e-01 -6.88196778e-01
1.25280285e+00 5.60581565e-01 1.19249260e+00 6.18634403e-01
4.35173661e-01 5.20119548e-01 1.33306253e+00 -3.11660558e-01
6.99967623e-01 2.60008126e-01 2.18227893e-01 4.64647621e-01
3.40711564e-01 5.83780348e-01 -6.08237386e-02 -4.15461928e-01
6.51807964e-01 1.94997877e-01 -2.24317074e-01 -2.52026260e-01
-5.83187640e-01 1.02635431e+00 5.50622046e-01 5.76043464e-02
-1.07707286e+00 -2.97945261e-01 2.77752250e-01 4.57366467e-01
4.36595172e-01 7.02734768e-01 -9.16331708e-01 1.44685104e-01
-1.20731819e+00 2.95677423e-01 1.48926651e+00 9.09932613e-01
8.92583251e-01 7.49319732e-01 1.59853816e-01 7.01878726e-01
4.76639777e-01 1.19609571e+00 2.38793999e-01 -3.26706856e-01
-6.91950917e-02 3.65550339e-01 4.95722592e-01 -6.34227395e-01
-6.05188727e-01 -4.56617177e-01 -9.60517764e-01 3.58953103e-02
5.69655374e-02 -7.83927217e-02 -7.06773877e-01 7.59501338e-01
3.01646918e-01 -3.66484225e-01 4.56241429e-01 9.65741575e-01
4.85429615e-01 7.60634303e-01 2.28491738e-01 -5.45772314e-01
4.88210797e-01 -5.09172320e-01 -5.94079614e-01 4.43989784e-01
5.96342742e-01 -4.46934015e-01 9.44348931e-01 2.68062443e-01
-1.30601215e+00 -1.35667115e-01 -9.81885374e-01 3.90301794e-01
-5.28877199e-01 2.93817937e-01 8.84172261e-01 2.83068508e-01
-5.97003162e-01 1.31461334e+00 -1.05913270e+00 6.37699291e-02
1.92570180e-01 9.12971735e-01 -1.08746186e-01 4.02697660e-02
-1.28689814e+00 1.14158928e+00 3.97973418e-01 7.93922186e-01
-1.08821130e+00 -9.42469299e-01 -8.07132602e-01 1.24947764e-01
-1.78158749e-02 -3.71229410e-01 8.98228467e-01 -5.49400449e-01
-1.90618002e+00 -4.43469621e-02 -3.05391043e-01 -4.41624939e-01
5.62283039e-01 -1.52040571e-01 -3.22763830e-01 -1.08899109e-01
-3.09462577e-01 -1.96436897e-01 7.72659421e-01 -1.21934426e+00
-1.23610027e-01 -4.13817376e-01 -2.17066750e-01 3.50731239e-02
-1.08933210e-01 -5.29863238e-01 5.62394917e-01 1.15215279e-01
3.55464220e-01 -5.09039402e-01 -5.89908838e-01 -6.31988406e-01
-3.60006154e-01 2.64102042e-01 3.59320611e-01 -7.76329219e-01
9.31974769e-01 -1.66405761e+00 4.14962083e-01 9.90224242e-01
-1.09296359e-01 1.10721335e-01 8.01829025e-02 1.01007414e+00
-1.66228741e-01 -2.78602131e-02 -6.31364644e-01 2.66846180e-01
5.76892905e-02 4.23235208e-01 -2.46285990e-01 6.76363707e-01
2.16595396e-01 7.87467182e-01 -8.54479671e-01 9.33139250e-02
5.03418386e-01 7.93979824e-01 -2.65033960e-01 2.16645643e-01
-2.54093200e-01 5.16210914e-01 -3.68210077e-01 1.06735098e+00
9.04073536e-01 7.36020803e-02 2.61265785e-01 -2.78522521e-01
-6.66439652e-01 -1.25506163e-01 -1.37773311e+00 1.29737675e+00
-7.49180198e-01 4.39492539e-02 4.57644999e-01 -1.08876610e+00
1.13477850e+00 1.94773585e-01 8.94442201e-01 -8.25233042e-01
6.47103116e-02 9.21478868e-01 -1.01803824e-01 -8.82037401e-01
3.41169357e-01 -3.32090855e-01 6.64959848e-01 -2.06868902e-01
-1.93276495e-01 -3.90789598e-01 3.32087874e-02 -2.87864208e-01
4.79443699e-01 2.72971839e-01 9.76313353e-02 -8.86833251e-01
9.35160935e-01 7.32975245e-01 -2.94127643e-01 7.53089130e-01
6.43018901e-01 -2.01553121e-01 -5.18914834e-02 -5.92396140e-01
-1.43998539e+00 -2.86703020e-01 -6.23885632e-01 9.22452986e-01
1.71437323e-01 1.12955250e-01 -4.86549228e-01 5.44278264e-01
5.00497758e-01 6.13132834e-01 -3.56983244e-01 1.38398400e-02
-2.43589923e-01 -9.32434678e-01 6.93376899e-01 3.44575197e-01
-4.18923609e-03 -7.09140003e-01 -1.86245054e-01 8.50285351e-01
6.24473333e-01 -4.01354402e-01 5.67536116e-01 1.05305815e+00
-1.29049766e+00 -1.18590522e+00 -7.48956501e-01 -5.14811158e-01
5.38525224e-01 -2.73907650e-02 7.35432148e-01 1.06239930e-01
-4.71362889e-01 3.91183086e-02 7.25013390e-02 -3.62238169e-01
-6.19102657e-01 3.43855955e-02 2.12661311e-01 -3.83142352e-01
1.05189040e-01 -3.39423984e-01 -4.97815996e-01 4.17621732e-01
-7.98272014e-01 -3.77309918e-01 6.37116134e-01 1.35356152e+00
8.55089784e-01 2.45871600e-02 3.02057803e-01 -7.68117666e-01
1.07859731e+00 -8.61672163e-01 -8.05678904e-01 3.21892619e-01
-1.16030395e+00 1.58052281e-01 7.95113087e-01 -3.90318513e-01
-7.76094437e-01 3.10595810e-01 -4.14038092e-01 -3.09246659e-01
-6.65278956e-02 1.52261555e+00 4.58494842e-01 -5.83198309e-01
1.00513089e+00 3.55450898e-01 3.77806991e-01 -2.96355188e-01
-1.80196632e-02 5.60110867e-01 7.66778737e-02 -1.01831853e+00
3.43153983e-01 2.24026531e-01 5.30695021e-01 -1.53707075e+00
-9.60730761e-03 -5.65776110e-01 -7.24402010e-01 1.01080768e-01
6.24682419e-02 -1.04159760e+00 -1.17004895e+00 4.74303722e-01
-8.29050720e-01 -7.02508330e-01 -5.11130869e-01 6.27313435e-01
-2.83117056e-01 4.96448241e-02 -6.14935696e-01 -1.47136295e+00
-5.73840559e-01 -1.25156403e+00 7.84753680e-01 3.04476102e-03
1.46728948e-01 -1.18845594e+00 -4.19921875e-02 -2.48760059e-01
9.88726020e-01 1.03158720e-01 8.45092773e-01 -6.16633117e-01
-1.98024586e-01 -4.25805002e-01 2.60195196e-01 2.09101006e-01
-1.30289555e-01 4.20329988e-01 -8.17141473e-01 -6.58832550e-01
9.79497358e-02 -2.60145158e-01 6.58522666e-01 7.94381380e-01
7.04928815e-01 -3.97259504e-01 -2.82890528e-01 8.31898510e-01
1.68135238e+00 4.66277093e-01 3.98286521e-01 2.34097332e-01
6.08451903e-01 3.71318251e-01 3.10205787e-01 7.45235205e-01
5.52302673e-02 2.59321332e-01 5.49632490e-01 -4.52640712e-01
8.87867510e-01 -1.01988554e-01 1.15298145e-01 5.28059721e-01
-8.17529738e-01 2.33465791e-01 -1.02498436e+00 1.23640828e-01
-1.63753510e+00 -8.82361531e-01 -8.02775800e-01 2.13081646e+00
6.14384234e-01 -2.32399285e-01 1.67908017e-02 4.35054362e-01
9.61755440e-02 -5.72815418e-01 -6.75592244e-01 -5.40556610e-01
-3.57603669e-01 7.89300382e-01 1.11092234e+00 1.01730955e+00
-8.01203132e-01 7.31531799e-01 6.83306980e+00 4.62258309e-01
-1.65719295e+00 -3.24667105e-03 1.08268544e-01 2.57204711e-01
-4.40264076e-01 2.85008997e-02 -6.31611705e-01 -6.62289038e-02
1.03908813e+00 1.58053748e-02 9.42239165e-01 7.91596472e-01
7.72685766e-01 -9.01966691e-02 -7.33612537e-01 5.06759524e-01
-4.85752374e-01 -1.26710689e+00 -3.19753706e-01 -3.00014101e-04
6.88362300e-01 8.18870068e-02 -1.85248107e-01 3.96285534e-01
1.47803634e-01 -1.20626223e+00 6.37887239e-01 9.81281161e-01
7.21863568e-01 -3.22801203e-01 8.90506327e-01 4.80062097e-01
-9.03837442e-01 -4.16598529e-01 -6.55112207e-01 -5.07083952e-01
-8.62780139e-02 4.00245607e-01 -1.10776365e+00 9.14523304e-01
2.06320554e-01 7.60177851e-01 6.81831734e-03 1.17596245e+00
7.11180791e-02 6.58647597e-01 -7.99108803e-01 -4.77697581e-01
2.06954271e-01 -7.90490687e-01 2.42626876e-01 1.29785025e+00
3.87708843e-01 3.82065356e-01 -1.25652745e-01 9.58573580e-01
9.17336404e-01 6.79676592e-01 -6.32243335e-01 -2.28818670e-01
3.36067349e-01 7.53842056e-01 -2.09251180e-01 9.78994668e-02
6.20874539e-02 4.22662914e-01 -3.34748060e-01 9.97720301e-01
-6.73996806e-01 -1.06080726e-01 3.43807399e-01 1.65313184e-01
7.46821612e-02 -3.49577725e-01 -3.99381757e-01 -9.81393516e-01
-4.06688601e-01 -5.36364436e-01 1.54912338e-01 -7.76037872e-01
-1.13836694e+00 3.62358481e-01 1.43409431e-01 -6.34142458e-01
-9.02752876e-02 -1.01013947e+00 -3.48486364e-01 1.47515047e+00
-1.70365858e+00 -8.03646922e-01 -4.90262032e-01 7.42780566e-01
3.17257553e-01 -2.25426689e-01 1.25760257e+00 3.61152321e-01
-6.28979743e-01 -4.65647280e-02 1.10703170e+00 -8.37343812e-01
1.12329563e-02 -1.17267764e+00 -4.39623892e-01 6.21239357e-02
-6.19205534e-01 8.28578770e-01 8.87724876e-01 -6.98403239e-01
-2.34580564e+00 -5.77245891e-01 3.69516611e-01 2.80702412e-01
7.49536753e-01 1.59432456e-01 -1.20843077e+00 1.16943032e-01
-3.31380278e-01 -2.00388849e-01 4.24167752e-01 4.37008627e-02
4.69681948e-01 -4.04971361e-01 -9.99433458e-01 -3.19894850e-02
2.73202091e-01 -3.52251619e-01 -1.12629624e-03 6.22737408e-01
1.06395796e-01 -8.69318306e-01 -1.38583410e+00 4.07517582e-01
8.87806296e-01 -3.00888240e-01 1.33478236e+00 -9.32942212e-01
4.95680213e-01 -1.47277832e-01 -2.52624810e-01 -1.28097463e+00
-1.59812286e-01 -4.25949186e-01 -2.72204041e-01 2.65847743e-01
7.73389995e-01 -5.96375942e-01 4.73148137e-01 9.88489747e-01
-3.41709584e-01 -1.09510887e+00 -6.88120425e-01 -6.35317743e-01
1.37077838e-01 -8.41097385e-02 6.08007371e-01 8.55468452e-01
2.38881651e-02 2.09849942e-02 -3.16928685e-01 4.52771723e-01
2.17914462e-01 -3.78954172e-01 6.88446462e-01 -1.53687346e+00
-1.58837169e-01 -3.62032890e-01 7.00365752e-02 -8.31721187e-01
4.50583100e-02 -8.98692727e-01 -4.51713800e-02 -1.49615788e+00
-5.43679535e-01 -7.81361520e-01 -3.77862155e-01 5.13518155e-01
3.20805579e-01 -4.11525220e-01 -4.32226181e-01 7.70089388e-01
3.90504032e-01 7.86918759e-01 1.24054039e+00 -7.58020699e-01
-7.51618505e-01 2.62169391e-01 -8.54639709e-02 1.79655641e-01
5.48919261e-01 -4.09002304e-01 -4.23795074e-01 5.98842949e-02
4.32388514e-01 5.57468653e-01 5.03051937e-01 -7.64839470e-01
2.87241012e-01 -4.85821366e-01 4.98676568e-01 -4.18844432e-01
3.36038202e-01 -1.35151649e+00 7.86513090e-01 7.78163254e-01
-7.81429484e-02 -9.25900042e-02 7.33883917e-01 3.69355857e-01
-3.47889990e-01 -3.89992625e-01 2.35539526e-01 -1.72379725e-02
-6.34041011e-01 3.05434406e-01 -3.41519147e-01 -8.78545105e-01
8.66990983e-01 -3.92138332e-01 -1.11835599e-01 6.67608455e-02
-1.09363127e+00 1.75131783e-01 1.60663247e-01 -5.46006918e-01
6.83125675e-01 -7.03096628e-01 -6.12137258e-01 5.09407759e-01
-4.49050456e-01 1.12793200e-01 2.49965161e-01 1.29714000e+00
-1.20109987e+00 5.12170970e-01 -1.15328185e-01 -8.97515774e-01
-6.16657376e-01 3.18378419e-01 8.37163627e-01 -6.96272403e-02
-3.14184695e-01 5.40430903e-01 -8.16237211e-01 -4.88505453e-01
-9.02604982e-02 -3.33992213e-01 -8.96653980e-02 1.59105621e-02
-2.77492166e-01 5.68878829e-01 5.25643289e-01 -3.35663170e-01
-2.11934209e-01 2.38957837e-01 3.17779630e-01 1.15545236e-01
1.82632267e+00 2.05789924e-01 -4.36010033e-01 4.42606241e-01
8.97905111e-01 -3.87798965e-01 -1.18220150e+00 1.23088665e-01
-1.15480579e-01 -1.87252849e-01 6.76744580e-02 -9.29179192e-01
-4.59162414e-01 6.15040362e-01 6.14703238e-01 4.54211205e-01
8.83525610e-01 -8.14999640e-01 5.03973551e-02 1.22083879e+00
2.76717216e-01 -9.80861962e-01 -5.34776807e-01 9.07921493e-01
1.01477396e+00 -1.18217921e+00 3.74545127e-01 -2.09017560e-01
-5.14869690e-01 1.60296655e+00 4.01060104e-01 -1.47147849e-03
7.96184003e-01 1.99477598e-01 -1.06297888e-01 -6.69842839e-01
4.80754748e-02 2.60855943e-01 -1.43595368e-01 2.22303614e-01
4.40112561e-01 2.11853415e-01 -5.01558363e-01 4.29946184e-01
9.52430889e-02 9.42878202e-02 2.71472812e-01 1.10046721e+00
-2.54466027e-01 -8.14566374e-01 -4.55227703e-01 3.77152890e-01
3.07618529e-01 -4.51651812e-01 1.16067573e-01 1.12303126e+00
-4.47775871e-01 5.70482135e-01 -5.35137653e-01 -1.50313959e-01
8.80294323e-01 2.30840549e-01 1.27326697e-01 -1.93217263e-01
-5.26913106e-01 2.19624028e-01 -1.96303651e-01 -1.08140178e-01
-4.74891216e-01 -4.33022380e-01 -1.27742445e+00 -4.24348772e-01
-8.19262266e-01 7.44392574e-01 1.41932845e+00 1.03384483e+00
1.69748530e-01 4.17390853e-01 7.24635541e-01 -1.11535907e+00
-1.18937719e+00 -1.14983666e+00 -1.07259083e+00 1.25566065e-01
4.59025204e-01 -8.24361801e-01 -3.15056056e-01 -2.76593238e-01] | [6.383156776428223, 3.381610870361328] |
aa6cf3ed-7785-4b5a-8cbc-f61aad191df9 | multimodal-web-navigation-with-instruction | 2305.11854 | null | https://arxiv.org/abs/2305.11854v1 | https://arxiv.org/pdf/2305.11854v1.pdf | Multimodal Web Navigation with Instruction-Finetuned Foundation Models | The progress of autonomous web navigation has been hindered by the dependence on billions of exploratory interactions via online reinforcement learning, and domain-specific model designs that make it difficult to leverage generalization from rich out-of-domain data. In this work, we study data-driven offline training for web agents with vision-language foundation models. We propose an instruction-following multimodal agent, WebGUM, that observes both webpage screenshots and HTML pages and outputs web navigation actions, such as click and type. WebGUM is trained by jointly finetuning an instruction-finetuned language model and a vision transformer on a large corpus of demonstrations. We empirically demonstrate this recipe improves the agent's ability of grounded visual perception, HTML comprehension and multi-step reasoning, outperforming prior works by a significant margin. On the MiniWoB benchmark, we improve over the previous best offline methods by more than 31.9%, being close to reaching online-finetuned SoTA. On the WebShop benchmark, our 3-billion-parameter model achieves superior performance to the existing SoTA, PaLM-540B. We also collect 347K high-quality demonstrations using our trained models, 38 times larger than prior work, and make them available to promote future research in this direction. | ['Izzeddin Gur', 'Shixiang Shane Gu', 'Yutaka Matsuo', 'Kuang-Huei Lee', 'Ofir Nachum', 'Hiroki Furuta'] | 2023-05-19 | null | null | null | null | ['instruction-following'] | ['natural-language-processing'] | [-5.20012043e-02 9.34242979e-02 -3.19941759e-01 -3.08379322e-01
-9.82650578e-01 -1.03061426e+00 6.97163582e-01 -1.97329193e-01
-6.48069978e-01 5.20641565e-01 2.27013841e-01 -8.49154949e-01
1.45568892e-01 -5.08168221e-01 -1.39054227e+00 -1.88090712e-01
-1.46647066e-01 5.44975638e-01 4.62075055e-01 -3.12243581e-01
1.94431573e-01 -1.73530057e-01 -1.65436375e+00 3.91859293e-01
1.00891066e+00 7.91092753e-01 5.69453061e-01 1.30468380e+00
3.09976824e-02 9.99022305e-01 -2.32192963e-01 -1.80666104e-01
2.93491513e-01 -2.25931376e-01 -6.48947001e-01 -6.31248206e-02
9.49941456e-01 -1.03814340e+00 -5.08000612e-01 8.06996465e-01
1.74777776e-01 9.83503386e-02 4.85920221e-01 -1.42827821e+00
-7.77103186e-01 6.74497187e-01 -4.27400053e-01 -4.47730394e-03
3.80173087e-01 9.21563029e-01 1.18935680e+00 -5.37676692e-01
6.79683447e-01 1.36840451e+00 4.54793632e-01 7.37016976e-01
-1.18753254e+00 -6.11078680e-01 4.47266430e-01 2.71985739e-01
-3.93065631e-01 -3.47339869e-01 2.22636744e-01 -3.73489767e-01
1.19292402e+00 -1.61222547e-01 8.10765088e-01 1.86219954e+00
-2.26800278e-01 1.08438957e+00 1.29297101e+00 -3.89934778e-01
2.38570556e-01 -1.98919073e-01 2.72420466e-01 1.52133429e+00
2.22151473e-01 4.09632146e-01 -8.27485502e-01 1.23688407e-01
8.33484471e-01 -6.62745014e-02 -1.33823738e-01 -9.53806281e-01
-1.27943456e+00 7.98105240e-01 4.75994766e-01 -5.24914324e-01
-2.98221737e-01 4.36448097e-01 2.00199917e-01 5.58193564e-01
-4.28172529e-01 5.01819968e-01 -6.81237578e-01 -8.03903341e-01
-2.13928625e-01 4.02217954e-01 1.21525347e+00 1.29535413e+00
5.64510345e-01 -8.58523697e-02 2.85852522e-01 6.42086148e-01
5.01368642e-01 9.44036961e-01 7.26530790e-01 -1.59877920e+00
8.50029528e-01 5.99581659e-01 2.92506397e-01 -1.99056640e-01
-2.58187741e-01 1.71795100e-01 1.38395607e-01 7.86916971e-01
9.50004697e-01 -2.75937408e-01 -1.18999398e+00 1.80778134e+00
2.44936123e-01 -3.83480906e-01 2.24254161e-01 8.15146863e-01
6.71958804e-01 5.78920782e-01 2.31110230e-01 4.28469837e-01
1.34900153e+00 -1.75610745e+00 -2.24821880e-01 -3.79615426e-01
6.32473648e-01 -3.90614480e-01 2.02935433e+00 6.44279659e-01
-9.97953832e-01 -3.66010070e-01 -1.11875284e+00 -2.72163868e-01
-6.10215485e-01 -1.49124056e-01 8.65089476e-01 3.49107653e-01
-9.67193842e-01 2.18157321e-01 -1.16711485e+00 -6.45296216e-01
3.46067011e-01 6.46713376e-02 -1.84761852e-01 -5.32468110e-02
-6.18211865e-01 7.43459642e-01 7.94867426e-02 -4.73448426e-01
-1.78769612e+00 -8.03618014e-01 -8.52408707e-01 1.34373382e-01
8.73907804e-01 -8.08844864e-01 2.00812149e+00 -4.38701928e-01
-1.87069809e+00 5.82409561e-01 1.38961837e-01 -6.82082236e-01
7.16415584e-01 -5.27240336e-01 -6.94344863e-02 1.38550848e-01
5.58194565e-03 9.79280770e-01 6.59706831e-01 -1.18266439e+00
-1.04950702e+00 -3.30693275e-01 6.46276355e-01 3.28522235e-01
-4.33267832e-01 -6.88966155e-01 -6.74357533e-01 -1.50041915e-02
-4.62203890e-01 -1.14470494e+00 -7.94194452e-03 1.71942398e-01
-4.02684718e-01 -3.55825394e-01 7.62426436e-01 -6.56015098e-01
5.48555553e-01 -1.82347667e+00 7.30964243e-02 -2.43537247e-01
3.21911842e-01 1.12435974e-01 -5.56912065e-01 6.01382732e-01
4.49234992e-01 -5.72627559e-02 1.30800873e-01 -1.47702500e-01
6.47858679e-01 2.85286844e-01 -3.36947978e-01 -1.04026265e-01
-2.81791985e-01 1.12586117e+00 -8.70123267e-01 -4.78856504e-01
2.50888020e-01 7.52100274e-02 -8.68306279e-01 5.32553971e-01
-9.88741338e-01 2.89232042e-02 -5.18655896e-01 6.88188791e-01
-3.08078690e-03 -5.88872254e-01 2.66041011e-01 -4.69123535e-02
-1.56792209e-01 4.36033130e-01 -5.11783659e-01 2.21060777e+00
-8.38447869e-01 7.96960354e-01 2.20126078e-01 -4.43313777e-01
2.78664470e-01 -1.31181583e-01 6.23131506e-02 -1.12839735e+00
-3.68436962e-01 5.90746514e-02 1.30794257e-01 -9.80594993e-01
2.29376197e-01 6.35736227e-01 3.98903489e-02 4.33381259e-01
2.17141956e-01 -1.56289756e-01 5.21727145e-01 5.54224849e-01
1.52860987e+00 6.97963059e-01 -7.16318563e-02 1.71771750e-01
-1.88230738e-01 5.44880331e-01 -1.34114563e-01 1.13970971e+00
-3.70719254e-01 2.61079036e-02 4.77737546e-01 -2.78361470e-01
-1.34479082e+00 -1.45043504e+00 4.87153471e-01 1.73879778e+00
1.29417807e-01 -5.51320612e-01 -1.00522113e+00 -1.09289479e+00
3.00633907e-01 1.00388622e+00 -4.98340875e-01 1.36537254e-01
-5.88544607e-01 -1.88608408e-01 4.38889056e-01 6.86560750e-01
6.94702983e-01 -1.11742496e+00 -8.22848797e-01 -7.62929320e-02
-4.37547565e-02 -1.28685308e+00 -5.60427368e-01 3.48003089e-01
-7.20134139e-01 -1.44531786e+00 -5.02807617e-01 -9.58146572e-01
4.24856186e-01 4.80359346e-01 1.28906035e+00 3.34285535e-02
-2.57289112e-01 1.03981411e+00 -3.02005649e-01 -3.05145293e-01
-5.57977855e-01 2.65300125e-01 -1.48422971e-01 -8.78475368e-01
2.44707793e-01 -6.40829742e-01 -7.18094587e-01 2.16299891e-01
-5.65233052e-01 3.81280631e-01 1.05564868e+00 1.04985404e+00
7.16276020e-02 -3.77998710e-01 3.32463622e-01 -7.61960030e-01
7.05381036e-01 -3.05929810e-01 -1.07932949e+00 2.26096973e-01
-9.82115328e-01 3.30007344e-01 8.12660098e-01 -5.55737674e-01
-1.09080482e+00 -2.67645177e-02 1.16086908e-01 -1.04780652e-01
-3.42708945e-01 2.17021525e-01 -1.15123671e-02 -6.47757798e-02
7.89032876e-01 3.63078982e-01 1.81421012e-01 -4.08734143e-01
8.69926810e-01 4.35120553e-01 6.44610524e-01 -8.23494613e-01
8.35501373e-01 2.07397297e-01 -3.86770636e-01 -7.42716372e-01
-7.13495612e-01 -1.98777333e-01 -4.07141633e-02 -3.42590809e-02
8.06389451e-01 -9.30067897e-01 -1.56402040e+00 2.26739675e-01
-7.42818356e-01 -1.46042621e+00 -4.37275060e-02 4.55968201e-01
-1.00197709e+00 2.34731853e-01 -9.12193775e-01 -5.08778036e-01
-1.01322075e-02 -1.28097701e+00 8.60647440e-01 4.60633188e-01
1.39292285e-01 -7.96342850e-01 1.97732463e-01 7.54308939e-01
3.19345862e-01 -3.93578291e-01 1.10326934e+00 -6.98477685e-01
-1.04134154e+00 1.76182121e-01 -3.76252472e-01 1.22599907e-01
-1.77758813e-01 -2.12086484e-01 -6.64197087e-01 -2.93211877e-01
-4.91979450e-01 -1.12245607e+00 9.55759704e-01 -9.61883110e-04
1.06489873e+00 -4.68420863e-01 -2.77520388e-01 6.79243624e-01
1.20462811e+00 1.51699141e-01 1.33562043e-01 8.71861517e-01
6.63781464e-01 2.64944345e-01 5.93063295e-01 2.53673017e-01
8.55967879e-01 4.09732699e-01 9.04865980e-01 2.78130800e-01
-4.65840489e-01 -9.10719037e-01 6.26011670e-01 4.97515857e-01
7.95312971e-02 -4.11743671e-01 -6.90056205e-01 5.13408422e-01
-1.95473123e+00 -8.21112692e-01 3.06101948e-01 1.87851012e+00
9.68347490e-01 4.19151992e-01 3.56163442e-01 -7.00127006e-01
2.91072596e-02 7.02047944e-02 -1.15370500e+00 -1.54183373e-01
3.65409732e-01 -1.73225403e-01 6.22067094e-01 7.95371830e-01
-9.49084103e-01 1.10077620e+00 6.08369255e+00 5.43254137e-01
-5.66311538e-01 -1.26268387e-01 9.68661234e-02 -2.51572698e-01
-1.65662482e-01 -1.62615821e-01 -9.48814154e-01 4.42714840e-01
1.05940378e+00 3.01072299e-01 1.13543367e+00 1.17372632e+00
1.68077424e-01 -2.82685965e-01 -1.40473235e+00 7.65483797e-01
-1.02028754e-02 -1.18230009e+00 -9.10745487e-02 4.03571904e-01
6.25031412e-01 5.55694640e-01 3.27678353e-01 9.57445741e-01
1.47078013e+00 -9.38570440e-01 5.04871666e-01 2.29485601e-01
4.74153340e-01 -2.30697051e-01 -4.54862379e-02 5.40063381e-01
-5.59993744e-01 -2.56990224e-01 -5.38458675e-02 5.17155863e-02
-1.83439538e-01 -3.74902815e-01 -1.24413073e+00 -2.81914920e-01
8.96177173e-01 3.65109026e-01 -6.50207996e-01 8.52949500e-01
-4.69690651e-01 9.32803869e-01 -4.13655579e-01 -6.94330990e-01
5.56020498e-01 -6.75205439e-02 4.82958794e-01 9.16521966e-01
5.05291782e-02 -1.38793662e-01 2.98674554e-01 8.30526650e-01
-2.17116594e-01 -2.28909299e-01 -6.04437709e-01 -4.53554004e-01
4.24808621e-01 1.28573322e+00 -1.35445669e-01 -5.24938941e-01
-6.44129992e-01 1.09579194e+00 6.78834260e-01 7.17207432e-01
-1.09084952e+00 -2.88918316e-01 8.16466928e-01 -1.96689442e-01
7.51237690e-01 -7.92774916e-01 1.27962679e-01 -1.19895494e+00
1.75897535e-02 -1.33548486e+00 3.14721018e-01 -1.14241660e+00
-1.10086679e+00 2.78526336e-01 -2.95043051e-01 -6.96732700e-01
-5.39202034e-01 -1.22899926e+00 -2.81537682e-01 3.87758523e-01
-1.63499069e+00 -1.33890676e+00 -7.00236499e-01 5.72466075e-01
1.08988225e+00 -3.34320515e-01 9.05926406e-01 -1.09216340e-01
-3.64145428e-01 5.00412285e-01 1.35377899e-01 1.53938800e-01
8.70509446e-01 -1.73336053e+00 6.06208563e-01 4.16343749e-01
3.52396041e-01 6.16934061e-01 8.53972018e-01 -6.38605416e-01
-2.09328365e+00 -5.44578731e-01 3.35997939e-02 -8.33176672e-01
1.13720250e+00 -4.73135561e-01 -4.35963809e-01 1.13741601e+00
6.63028181e-01 -3.93016607e-01 3.83864075e-01 4.05029863e-01
-9.48137105e-01 -3.63757797e-02 -7.59783924e-01 1.32231069e+00
1.40552270e+00 -3.42165530e-01 -7.26853728e-01 5.28085470e-01
1.05644882e+00 -4.65874761e-01 -6.45475924e-01 -1.70370623e-01
9.84903097e-01 -9.65054750e-01 1.06829333e+00 -1.02579486e+00
6.31285369e-01 -1.52559299e-02 -3.26487601e-01 -1.53958881e+00
-6.57031834e-02 -6.02370799e-01 -3.98326546e-01 6.19582295e-01
6.28977299e-01 -5.56781232e-01 1.03551710e+00 4.39765215e-01
-2.66827434e-01 -5.05946100e-01 -2.40313753e-01 -7.84369767e-01
-6.73424080e-02 -2.70579964e-01 3.28378916e-01 3.23205203e-01
2.68658072e-01 6.12236738e-01 -1.56741470e-01 1.76456660e-01
9.10769105e-01 2.59305030e-01 1.35367918e+00 -7.31172860e-01
-6.71364784e-01 -6.13730788e-01 3.53783011e-01 -1.88236284e+00
6.19320422e-02 -6.24530792e-01 2.77896941e-01 -1.44441891e+00
3.13741595e-01 -5.36502823e-02 7.19725490e-02 6.41465664e-01
1.35561824e-01 -1.66554227e-01 2.25583911e-01 5.81546267e-03
-1.17155135e+00 5.05872369e-01 1.53895652e+00 -2.79150635e-01
-8.30823779e-02 -2.64064282e-01 -5.42904317e-01 8.23505640e-01
6.62238359e-01 1.09019861e-01 -6.29151285e-01 -6.59678161e-01
1.25014499e-01 1.01789087e-01 6.59783006e-01 -1.01184583e+00
4.09858376e-01 -3.43101621e-01 3.91698241e-01 -4.63614851e-01
5.69920957e-01 -7.69863248e-01 -7.90767908e-01 4.90883529e-01
-8.72188449e-01 1.97370678e-01 1.90846026e-01 9.28854287e-01
3.54312778e-01 -1.47654861e-01 1.86939344e-01 -4.50997233e-01
-1.24683440e+00 9.39973071e-02 -5.72635829e-01 2.80251890e-01
7.53318131e-01 -8.63486994e-03 -9.43085730e-01 -6.93969548e-01
-4.65973765e-01 6.14912152e-01 6.70570135e-01 5.11360466e-01
5.08143604e-01 -9.77783680e-01 -2.15125307e-01 1.05024174e-01
2.56201118e-01 -2.05922738e-01 4.46105003e-02 5.31253934e-01
-5.18201947e-01 6.24041617e-01 -2.75449991e-01 -5.25545359e-01
-1.12061179e+00 6.89804852e-01 6.26795813e-02 -7.47720823e-02
-8.43155742e-01 6.61696196e-01 3.20208102e-01 -6.60192847e-01
7.08534777e-01 -6.63685918e-01 1.80381462e-01 -6.81337893e-01
3.57444018e-01 3.33894312e-01 -3.90745848e-01 3.46502721e-01
1.03704236e-01 2.97905505e-01 -2.64524162e-01 -4.80323374e-01
1.26366484e+00 -1.99817121e-01 5.76701701e-01 4.10309613e-01
9.41740334e-01 5.84187992e-02 -2.10233545e+00 2.31466349e-02
-1.91432029e-01 -1.74871832e-01 -2.92237878e-01 -1.62577772e+00
-4.08374220e-01 8.08755994e-01 4.72127765e-01 2.69993216e-01
4.82664227e-01 4.49629873e-02 1.01281023e+00 1.22342312e+00
3.92178535e-01 -1.24450803e+00 7.96877623e-01 5.38566351e-01
7.12032676e-01 -1.89892161e+00 -4.19342726e-01 -5.06504765e-03
-6.10475361e-01 1.08942437e+00 1.17594481e+00 -1.51751190e-01
5.07592037e-02 2.44881973e-01 1.17841423e-01 -1.17625266e-01
-1.20287311e+00 -1.82154059e-01 1.40530029e-02 8.84155095e-01
6.20370470e-02 7.05554411e-02 3.18515360e-01 4.18001711e-01
-3.88625562e-01 -5.69916843e-03 3.57990712e-01 8.37162673e-01
-7.44494617e-01 -8.43144238e-01 -6.85738921e-02 2.49992982e-01
1.89666331e-01 -1.77388310e-01 -1.37378365e-01 1.23145235e+00
-6.11669540e-01 9.69447970e-01 -1.62631422e-01 -1.88334227e-01
3.00625741e-01 1.03202738e-01 7.87887633e-01 -3.17242444e-01
-9.10414010e-02 -1.86448634e-01 4.67645079e-01 -1.06988454e+00
2.44265109e-01 -6.02214396e-01 -1.45404053e+00 -2.58844346e-01
1.76606104e-01 -9.82807204e-02 9.87604797e-01 6.01291418e-01
4.35521305e-01 6.02064133e-01 1.78080630e-02 -6.28019869e-01
-1.29821420e+00 -9.47593689e-01 -1.03165228e-02 4.10388976e-01
4.68976945e-01 -6.61534131e-01 -3.52127522e-01 1.16177998e-01] | [4.411139011383057, 0.8343736529350281] |
6f8949a2-afca-4dca-8f03-f32891ab766a | going-denser-with-open-vocabulary-part | 2305.11173 | null | https://arxiv.org/abs/2305.11173v1 | https://arxiv.org/pdf/2305.11173v1.pdf | Going Denser with Open-Vocabulary Part Segmentation | Object detection has been expanded from a limited number of categories to open vocabulary. Moving forward, a complete intelligent vision system requires understanding more fine-grained object descriptions, object parts. In this paper, we propose a detector with the ability to predict both open-vocabulary objects and their part segmentation. This ability comes from two designs. First, we train the detector on the joint of part-level, object-level and image-level data to build the multi-granularity alignment between language and image. Second, we parse the novel object into its parts by its dense semantic correspondence with the base object. These two designs enable the detector to largely benefit from various data sources and foundation models. In open-vocabulary part segmentation experiments, our method outperforms the baseline by 3.3$\sim$7.3 mAP in cross-dataset generalization on PartImageNet, and improves the baseline by 7.3 novel AP$_{50}$ in cross-category generalization on Pascal Part. Finally, we train a detector that generalizes to a wide range of part segmentation datasets while achieving better performance than dataset-specific training. | ['Zhicheng Yan', 'Saining Xie', 'Ping Luo', 'Fanyi Xiao', 'Chenchen Zhu', 'Shoufa Chen', 'Peize Sun'] | 2023-05-18 | null | null | null | null | ['semantic-correspondence'] | ['computer-vision'] | [ 1.76880807e-01 2.06708044e-01 -1.71865240e-01 -8.28812242e-01
-1.00997734e+00 -7.20912814e-01 4.78617340e-01 -1.42454952e-01
-2.54538417e-01 2.46282071e-01 -1.13329589e-01 5.37504405e-02
4.24684346e-01 -6.72232151e-01 -1.11648107e+00 -2.66361237e-01
2.33872458e-01 7.28635132e-01 1.01278174e+00 -2.27410585e-01
-2.02834718e-02 3.65404248e-01 -1.67357457e+00 3.98151726e-01
8.01950216e-01 1.22460485e+00 4.85885799e-01 4.23069656e-01
-7.06698895e-01 2.75455058e-01 -4.19997424e-01 -4.74769920e-01
4.83512551e-01 6.89888522e-02 -9.38457489e-01 4.70457613e-01
1.07774866e+00 -3.19229931e-01 5.39061204e-02 1.24641228e+00
5.98710813e-02 -2.68158764e-01 6.51342213e-01 -1.14867759e+00
-7.05851078e-01 6.60787463e-01 -6.63724959e-01 2.99668312e-03
1.50445793e-02 1.83335021e-01 1.03622699e+00 -7.74696231e-01
7.07878888e-01 1.40312481e+00 6.90205693e-01 6.89676642e-01
-1.11133242e+00 -7.57059932e-01 5.70429862e-01 -4.06385101e-02
-1.46086705e+00 -1.53954968e-01 4.66559649e-01 -6.85241163e-01
1.05761254e+00 -7.52523318e-02 6.03086770e-01 8.00715804e-01
-1.12887666e-01 1.10913479e+00 9.89070714e-01 -1.03842579e-01
3.52808880e-03 2.95550019e-01 6.16488397e-01 7.36876130e-01
5.67811847e-01 -2.81995028e-01 -9.76419300e-02 2.77547568e-01
6.57480359e-01 7.77137559e-03 -6.04546517e-02 -6.27502561e-01
-9.62809801e-01 6.51752234e-01 7.22156346e-01 9.40824766e-03
-6.26516119e-02 6.55520707e-02 2.91408002e-01 -5.15176952e-02
2.89949149e-01 4.16986972e-01 -7.97456026e-01 3.01256716e-01
-8.56894672e-01 1.55698970e-01 8.96590531e-01 1.62601125e+00
1.05597699e+00 -1.58127770e-01 -1.24446452e-01 9.61961269e-01
5.33103466e-01 7.51461446e-01 3.62314045e-01 -8.22251439e-01
3.52965385e-01 8.83664191e-01 -1.89483747e-01 -1.68429628e-01
-4.22611058e-01 -4.59751070e-01 -3.73647392e-01 2.49623265e-02
4.23520416e-01 1.91981941e-01 -1.46084952e+00 1.52580452e+00
4.27472204e-01 -2.42797270e-01 -1.21854253e-01 8.01368654e-01
9.60508406e-01 4.13678378e-01 3.52628231e-01 3.62421811e-01
1.81843877e+00 -1.22479272e+00 -2.20492989e-01 -5.95708013e-01
4.47373420e-01 -7.21340239e-01 9.70595062e-01 1.46412134e-01
-7.77337849e-01 -9.92367685e-01 -1.11228859e+00 -3.72297794e-01
-5.49438775e-01 -7.59627596e-02 6.41798198e-01 5.10972261e-01
-8.96041095e-01 2.18495756e-01 -7.85604596e-01 -5.86537063e-01
9.00763214e-01 2.85351396e-01 -3.44177157e-01 -2.41302907e-01
-6.56571090e-01 6.95330858e-01 7.90518820e-01 -6.58244669e-01
-6.70333982e-01 -7.23421574e-01 -8.73673737e-01 2.04534288e-02
5.23671746e-01 -9.21864688e-01 1.32135272e+00 -1.02613258e+00
-9.67519462e-01 1.12587881e+00 -4.89145070e-02 -6.02530718e-01
3.07876468e-01 -2.97679693e-01 -2.94633836e-01 2.17731714e-01
5.50837040e-01 1.54386902e+00 8.13340306e-01 -1.30692971e+00
-1.10079598e+00 -6.38293266e-01 1.13083370e-01 6.23063445e-02
5.74736334e-02 -1.68328747e-01 -1.01038051e+00 -4.39769417e-01
3.93592834e-01 -1.02266455e+00 -9.90177765e-02 2.73068756e-01
-3.39960217e-01 -5.36385715e-01 8.69030774e-01 -3.37730289e-01
5.16114593e-01 -2.21895838e+00 -1.83081821e-01 -2.91863799e-01
1.84149206e-01 2.61199027e-01 -1.99031889e-01 -1.09852165e-01
2.13013425e-01 6.30479306e-02 -2.18201235e-01 -2.92456776e-01
8.05242956e-02 3.26286644e-01 -3.43706578e-01 1.90736040e-01
4.73308235e-01 1.20777225e+00 -5.05666852e-01 -7.95476556e-01
-3.50811817e-02 7.91738778e-02 -6.51137888e-01 5.83201600e-03
-6.51684880e-01 7.30857700e-02 -6.70171499e-01 8.62143517e-01
8.19721162e-01 -4.91820812e-01 -1.94408312e-01 -2.56415874e-01
3.69619280e-02 2.59873942e-02 -1.03255057e+00 1.94643259e+00
5.44407265e-03 4.43074107e-01 1.76500350e-01 -7.67678916e-01
9.02212322e-01 -1.66796654e-01 3.17877054e-01 -5.67519605e-01
1.94757760e-01 2.15433121e-01 1.60666816e-02 -2.50907630e-01
5.38400948e-01 -3.26914340e-02 -2.86969543e-01 1.81247577e-01
4.32953805e-01 -5.02064228e-01 3.31320941e-01 2.22330093e-01
8.39159071e-01 3.09576124e-01 2.49778762e-01 -4.91888940e-01
3.58042508e-01 4.50925589e-01 5.35778940e-01 8.94144297e-01
-5.42494178e-01 6.95101023e-01 1.00219361e-01 -3.31248850e-01
-1.05425417e+00 -1.17585909e+00 -5.64511597e-01 1.27217793e+00
5.40221214e-01 -1.71543971e-01 -1.08611250e+00 -8.08144569e-01
3.00344825e-01 5.18985748e-01 -3.84668916e-01 2.94132717e-02
-4.25941855e-01 -3.39244455e-01 5.18808067e-01 9.63765979e-01
8.79299641e-01 -8.93850744e-01 -4.93808478e-01 -6.86657503e-02
-1.91926628e-01 -1.54404891e+00 -5.94988525e-01 3.01158845e-01
-8.98063064e-01 -1.06980968e+00 -5.14263809e-01 -1.18536294e+00
5.37014365e-01 5.40482759e-01 1.31183314e+00 -1.67895183e-01
-5.50150514e-01 4.31751847e-01 -4.51315165e-01 -7.04809904e-01
-3.42411906e-01 5.76286912e-02 -2.27038451e-02 -1.99502975e-01
8.29197407e-01 -1.12703308e-01 -4.71335351e-01 5.31558394e-01
-7.76022613e-01 7.45150074e-02 9.63677406e-01 5.73826373e-01
1.02187645e+00 -2.45237380e-01 3.17922950e-01 -7.12847650e-01
-1.16914779e-01 -2.71962225e-01 -8.16729844e-01 3.23688298e-01
-3.46505344e-01 3.15050095e-01 2.32490271e-01 -4.35000628e-01
-9.76935387e-01 4.43093807e-01 -1.86981469e-01 -4.25367296e-01
-4.76127803e-01 -2.59428442e-01 -5.49551666e-01 -2.22231783e-02
6.42513156e-01 4.10474658e-01 -1.45693421e-01 -6.63775861e-01
7.78103352e-01 8.30503047e-01 7.16313660e-01 -5.20387232e-01
7.97852814e-01 6.86868906e-01 -4.05203760e-01 -7.44744778e-01
-1.12531400e+00 -9.58836854e-01 -8.92244458e-01 2.38351181e-01
1.28249884e+00 -1.33150446e+00 -2.19419420e-01 4.43873465e-01
-1.08848333e+00 -3.47730011e-01 -5.38007140e-01 2.04559326e-01
-5.33793449e-01 3.83225769e-01 -4.27946448e-01 -2.74441898e-01
-3.50262314e-01 -1.16742933e+00 1.49880290e+00 2.99312651e-01
1.43068299e-01 -4.27518517e-01 -4.21129942e-01 7.13372588e-01
-8.87830462e-03 -1.85572773e-01 5.63991725e-01 -1.05693460e+00
-1.05574405e+00 -1.92273125e-01 -6.95466876e-01 6.69817030e-01
-2.38427036e-02 -2.39561990e-01 -1.04416132e+00 -1.96639612e-01
-9.03245211e-02 -4.73163605e-01 1.12615657e+00 2.88041830e-01
1.19232905e+00 1.38096586e-01 -7.19749391e-01 7.67168283e-01
1.28147197e+00 6.62622601e-02 3.58910501e-01 2.14513615e-01
8.37823689e-01 5.28093815e-01 6.52196586e-01 4.77363728e-02
4.74576771e-01 5.22849262e-01 3.70404363e-01 8.98372158e-02
-5.92406452e-01 -3.07878733e-01 1.35993853e-01 4.31124508e-01
4.10335720e-01 -5.93541451e-02 -8.98407519e-01 6.40062630e-01
-1.47597480e+00 -7.77017415e-01 -2.10484624e-01 1.71691430e+00
7.21726656e-01 6.46279871e-01 1.88411340e-01 -4.66096789e-01
8.21782947e-01 -1.62488297e-01 -8.88315558e-01 -1.01696432e-01
-1.48608625e-01 -2.74307169e-02 8.49807382e-01 1.81871444e-01
-1.37884295e+00 1.41160667e+00 6.16859961e+00 9.57759142e-01
-8.72130871e-01 9.89810899e-02 4.22692448e-01 1.30589515e-01
4.25697081e-02 -8.83642733e-02 -1.51886141e+00 1.65283501e-01
5.59408426e-01 2.32818097e-01 -1.88148972e-02 1.30697894e+00
-5.43271542e-01 -4.97813635e-02 -1.35493231e+00 9.77001011e-01
2.75788367e-01 -1.20333302e+00 3.10719728e-01 1.36699095e-01
8.61327529e-01 5.12203097e-01 -2.87327111e-01 7.00502872e-01
3.83550078e-01 -7.44431198e-01 1.05201745e+00 2.01938197e-01
7.98981667e-01 -2.16890082e-01 5.27214289e-01 5.33735752e-01
-1.45472693e+00 -4.76973653e-02 -4.64358807e-01 1.20382532e-01
6.61616474e-02 2.11318135e-01 -9.52796876e-01 3.05297375e-01
8.71603370e-01 5.72431326e-01 -8.22489142e-01 9.63784337e-01
-1.92344949e-01 4.12383258e-01 -4.45218951e-01 1.06538244e-01
2.33242318e-01 9.29944068e-02 3.79056901e-01 1.27989519e+00
-1.51517168e-01 3.04239780e-01 6.98570669e-01 8.79275024e-01
-2.42253006e-01 -1.83393419e-01 -3.69819641e-01 6.84525073e-02
4.44003314e-01 1.28681707e+00 -9.26164985e-01 -7.51183033e-01
-6.95742130e-01 1.01770246e+00 3.24094862e-01 5.06489426e-02
-9.00501609e-01 -3.15761238e-01 9.15575862e-01 1.74883202e-01
9.12525773e-01 -7.77328461e-02 -3.95343512e-01 -1.11988163e+00
1.43066391e-01 -5.84743142e-01 3.23064208e-01 -7.93696225e-01
-1.45675468e+00 6.78193450e-01 8.98611024e-02 -1.03147984e+00
2.76886076e-02 -1.05221343e+00 -9.34299082e-03 6.85973048e-01
-1.49085987e+00 -1.58426261e+00 -3.63064647e-01 5.55458724e-01
9.93807018e-01 -8.15081596e-02 5.10632753e-01 2.77142912e-01
-2.90603906e-01 5.20499527e-01 -1.58398345e-01 6.63750112e-01
5.88309467e-01 -1.38985777e+00 8.08330536e-01 8.31122220e-01
3.96897703e-01 4.78530765e-01 4.19150919e-01 -6.96740925e-01
-1.14260757e+00 -1.41443241e+00 4.85692084e-01 -8.95911396e-01
6.85568690e-01 -5.45472741e-01 -9.14189041e-01 8.93765748e-01
-1.08257681e-01 3.81087542e-01 3.76596659e-01 1.13535345e-01
-9.92206216e-01 -1.42226234e-01 -1.16737330e+00 3.88318121e-01
1.37123394e+00 -5.15610933e-01 -1.02658534e+00 3.49461257e-01
1.11735308e+00 -4.07425672e-01 -9.33303833e-01 5.24155676e-01
5.28539896e-01 -7.06763566e-01 1.02086377e+00 -6.15078926e-01
9.26128849e-02 -4.47269708e-01 -5.40026963e-01 -7.56844044e-01
-4.94584739e-01 1.84000880e-02 1.65174723e-01 1.17795336e+00
3.33036035e-01 -3.75646949e-01 6.20781898e-01 4.85091001e-01
-6.08272076e-01 -5.57224870e-01 -6.81585789e-01 -1.06892216e+00
2.38201573e-01 -6.91811740e-01 5.52860141e-01 3.81924748e-01
-4.99421597e-01 5.85605025e-01 3.04770947e-01 3.29282582e-01
6.63994014e-01 5.53443968e-01 8.98764431e-01 -1.33907127e+00
-3.22014451e-01 -6.52240992e-01 -7.81683803e-01 -1.76033390e+00
1.38004079e-01 -1.05316830e+00 3.13635647e-01 -1.60998440e+00
5.62002420e-01 -5.40902674e-01 -1.47452176e-01 5.75829089e-01
-1.31838575e-01 5.17970800e-01 3.80874604e-01 3.21896046e-01
-1.11202204e+00 2.75013089e-01 1.34100425e+00 -4.55303729e-01
6.80091511e-03 -9.91804525e-02 -8.88233364e-01 9.50151265e-01
6.05723679e-01 -3.26062977e-01 -1.77835822e-01 -5.55094838e-01
-4.05657709e-01 -4.14971948e-01 4.31343824e-01 -1.28747547e+00
1.59521643e-02 4.96808589e-02 3.07554841e-01 -9.50664580e-01
4.40357149e-01 -7.33265460e-01 -1.37246624e-01 3.90147150e-01
-5.28857410e-02 -5.46036959e-01 4.50450867e-01 7.86676526e-01
-7.06705973e-02 -7.24579841e-02 1.01975548e+00 -2.80214489e-01
-1.38057518e+00 3.89854640e-01 1.42142147e-01 4.27429199e-01
1.04683125e+00 -6.15783572e-01 -4.08102542e-01 2.81199127e-01
-7.55029857e-01 5.48759520e-01 7.23466933e-01 7.50550926e-01
2.51898676e-01 -9.16079223e-01 -5.50868273e-01 3.71315598e-01
6.62072957e-01 3.34636986e-01 1.57603309e-01 4.72167701e-01
-4.26002711e-01 6.86460555e-01 -2.21433893e-01 -1.12459898e+00
-1.20622492e+00 7.85699129e-01 3.87883216e-01 1.55064523e-01
-6.90153420e-01 1.28812385e+00 7.23841310e-01 -5.01398623e-01
2.22693667e-01 -6.91410780e-01 1.09010160e-01 -7.30499476e-02
5.09351194e-01 1.92940459e-02 -9.70759317e-02 -8.54650915e-01
-4.86996680e-01 9.13845956e-01 -2.96871483e-01 2.22971529e-01
9.19239938e-01 -1.20006360e-01 1.41051427e-01 3.25778365e-01
1.19112921e+00 -2.01174721e-01 -1.73032331e+00 -4.23512369e-01
2.47954074e-02 -8.80291611e-02 -2.35197574e-01 -9.30855930e-01
-8.58040512e-01 7.25995421e-01 6.89451277e-01 -4.87106759e-03
7.70309687e-01 7.97807515e-01 8.19886029e-01 5.43964267e-01
6.32763386e-01 -1.01122761e+00 6.22155704e-02 6.74062371e-01
4.95237142e-01 -1.62062275e+00 1.89833771e-02 -7.47491539e-01
-5.85268319e-01 8.22990894e-01 9.07245159e-01 -1.47151619e-01
6.98198438e-01 1.34283215e-01 1.54252440e-01 -3.13021064e-01
-4.06559914e-01 -7.71341205e-01 5.70547462e-01 6.75733984e-01
1.23088300e-01 3.42391223e-01 4.47870865e-02 6.78871214e-01
-2.47598022e-01 -2.31567070e-01 1.82568535e-01 7.78785288e-01
-1.04134369e+00 -1.02331364e+00 -3.45910907e-01 4.92407620e-01
-1.96410030e-01 2.52856165e-02 -3.72054815e-01 1.05884027e+00
5.90124428e-01 7.69057274e-01 3.50375503e-01 -8.61999840e-02
3.80636722e-01 2.22436473e-01 4.50187683e-01 -1.00007427e+00
-2.04594254e-01 3.50575820e-02 -3.99215519e-02 -6.32819235e-01
-3.87495816e-01 -7.71368206e-01 -1.53390455e+00 1.99584916e-01
-4.26251501e-01 -1.61121458e-01 7.69437909e-01 1.03337348e+00
3.99816602e-01 5.17425478e-01 9.01271310e-03 -7.36964226e-01
-5.90426385e-01 -9.55792487e-01 -6.74982667e-01 6.08189642e-01
1.31312817e-01 -6.92395449e-01 3.03264223e-02 2.11520150e-01] | [9.532453536987305, 1.3039400577545166] |
35fdf62f-6f8c-4417-b4a1-028616762fc4 | can-chatgpt-assess-human-personalities-a | 2303.01248 | null | https://arxiv.org/abs/2303.01248v2 | https://arxiv.org/pdf/2303.01248v2.pdf | Can ChatGPT Assess Human Personalities? A General Evaluation Framework | Large Language Models (LLMs) especially ChatGPT have produced impressive results in various areas, but their potential human-like psychology is still largely unexplored. Existing works study the virtual personalities of LLMs but rarely explore the possibility of analyzing human personalities via LLMs. This paper presents a generic evaluation framework for LLMs to assess human personalities based on Myers Briggs Type Indicator (MBTI) tests. Specifically, we first devise unbiased prompts by randomly permuting options in MBTI questions and adopt the average testing result to encourage more impartial answer generation. Then, we propose to replace the subject in question statements to enable flexible queries and assessments on different subjects from LLMs. Finally, we re-formulate the question instructions in a manner of correctness evaluation to facilitate LLMs to generate clearer responses. The proposed framework enables LLMs to flexibly assess personalities of different groups of people. We further propose three evaluation metrics to measure the consistency, robustness, and fairness of assessment results from state-of-the-art LLMs including ChatGPT and InstructGPT. Our experiments reveal ChatGPT's ability to assess human personalities, and the average results demonstrate that it can achieve more consistent and fairer assessments in spite of lower robustness against prompt biases compared with InstructGPT. | ['Chunyan Miao', 'Cyril Leung', 'Haocong Rao'] | 2023-03-01 | null | null | null | null | ['answer-generation'] | ['natural-language-processing'] | [-2.66142815e-01 1.82500228e-01 -1.87697922e-04 -5.99388897e-01
-5.78244627e-01 -4.98743236e-01 4.99078304e-01 1.10473973e-03
-5.98951817e-01 7.22046673e-01 1.72034264e-01 -4.60780948e-01
-1.70574948e-01 -8.01285446e-01 -1.79281577e-01 -4.47752088e-01
2.74983257e-01 3.80727470e-01 2.03002304e-01 -1.30724996e-01
5.89028537e-01 -3.57781649e-02 -1.22415328e+00 4.07966673e-01
1.50235665e+00 6.48840427e-01 -4.06836085e-02 7.59791791e-01
-3.27039808e-01 1.05219924e+00 -9.81149971e-01 -1.06864953e+00
1.72627017e-01 -3.47285837e-01 -1.00150442e+00 -2.51774698e-01
6.47724390e-01 -5.06645083e-01 -7.42077455e-02 1.10258758e+00
7.09928751e-01 2.11105198e-01 5.49612522e-01 -1.53473318e+00
-1.19350505e+00 6.64386094e-01 -3.48269343e-01 2.03901064e-02
7.80626118e-01 5.34194827e-01 1.12819588e+00 -3.50747257e-01
1.40783772e-01 1.69802821e+00 5.86712480e-01 5.75047672e-01
-1.05197620e+00 -7.23751128e-01 2.60915905e-01 2.53960252e-01
-1.07778907e+00 -6.08489737e-02 3.95085424e-01 -3.77959371e-01
3.24900836e-01 7.73296595e-01 9.85292792e-02 1.26153672e+00
1.63220838e-01 6.97271347e-01 1.63720429e+00 -2.24805132e-01
3.38255703e-01 7.26979256e-01 6.28678083e-01 4.54161644e-01
1.01751491e-01 -3.57036382e-01 -4.78306711e-01 -7.24413991e-01
7.13885486e-01 -2.33865604e-01 -8.59479979e-02 9.92309153e-02
-1.22172034e+00 8.52579176e-01 2.84643680e-01 1.54647395e-01
-4.70705718e-01 -3.19338381e-01 1.52764142e-01 4.16621476e-01
2.51307458e-01 6.06635153e-01 -2.30466351e-01 -1.24109417e-01
-5.69960296e-01 5.73325574e-01 1.13901222e+00 8.89017999e-01
6.27866387e-01 -4.39339846e-01 -9.59958673e-01 1.04698884e+00
1.76371589e-01 5.73622584e-01 6.52181327e-01 -1.21326792e+00
5.23955524e-01 7.84104049e-01 4.06983227e-01 -1.20759964e+00
-2.32002139e-01 -3.67725432e-01 -5.73496163e-01 -6.94709420e-02
4.43542778e-01 -2.72156715e-01 -2.59690974e-02 1.78353918e+00
3.02586615e-01 -4.24117267e-01 -2.69470870e-01 8.43754888e-01
8.41536403e-01 4.50268477e-01 4.76951689e-01 5.52268960e-02
1.38391542e+00 -7.98550069e-01 -5.67119420e-01 -1.58232182e-01
6.85718536e-01 -6.04503751e-01 1.93406832e+00 3.38329613e-01
-9.20278013e-01 -6.47795737e-01 -5.16891897e-01 -3.95539068e-02
7.19336700e-03 -4.04514708e-02 5.21020591e-01 1.29536712e+00
-1.01067495e+00 3.72664541e-01 -5.49334735e-02 -3.97948325e-01
2.36955091e-01 2.58012682e-01 -2.18591299e-02 1.68929666e-01
-1.52916634e+00 8.22256505e-01 9.75894742e-03 -3.30523670e-01
-3.90201092e-01 -6.46306157e-01 -3.31125945e-01 8.77744555e-02
4.51342762e-01 -7.75612533e-01 1.46067333e+00 -5.88956296e-01
-1.59724951e+00 8.04054916e-01 -1.66226849e-01 -1.71947956e-01
8.71599257e-01 -1.96181253e-01 -3.71102422e-01 4.51391041e-02
2.56074816e-01 4.92956489e-01 3.68126929e-01 -9.67212260e-01
-5.55166721e-01 -1.74695015e-01 4.75653440e-01 1.88024521e-01
-7.39029765e-01 3.23341697e-01 -1.25826057e-02 -3.51906747e-01
-3.49379003e-01 -6.61640048e-01 -2.20700681e-01 -1.94586501e-01
-4.11959976e-01 -5.59938610e-01 2.05539301e-01 -6.91693723e-01
1.57096124e+00 -1.68484211e+00 -4.86816049e-01 1.39812425e-01
6.31490469e-01 3.59631985e-01 -3.78682882e-01 4.75581408e-01
3.99020135e-01 3.68303001e-01 3.67165893e-01 -1.89701602e-01
5.07955313e-01 -6.98798895e-02 -1.12299547e-01 -2.75660632e-03
-2.00062871e-01 1.12518632e+00 -7.89848328e-01 -7.28632867e-01
-1.53602853e-01 -1.47707999e-01 -7.99805522e-01 5.26453137e-01
-1.24477476e-01 2.01969415e-01 -5.54550231e-01 4.30761606e-01
7.25967586e-01 -4.08038318e-01 -3.53563875e-02 1.77949712e-01
-5.95674617e-03 4.93679196e-01 -9.54217374e-01 7.53954411e-01
-3.55996430e-01 2.23510429e-01 3.30882473e-03 -3.29328507e-01
1.25310361e+00 4.57763160e-03 -1.55754864e-01 -7.68292844e-01
1.32155746e-01 3.14601474e-02 1.46469653e-01 -8.66874099e-01
6.14363313e-01 -1.03529401e-01 -2.27534249e-01 8.07868540e-01
-2.37612918e-01 2.71847576e-01 6.83981255e-02 3.81119370e-01
8.98856223e-01 -3.23712707e-01 2.78797448e-01 -4.38555956e-01
7.71037817e-01 -5.37686825e-01 3.66346717e-01 1.26224637e+00
-5.50915599e-01 2.24351257e-01 8.13928545e-01 -1.47056952e-01
-6.45045578e-01 -1.17734623e+00 1.25159785e-01 1.72141540e+00
2.80860666e-04 -3.11382681e-01 -1.01261330e+00 -1.02178609e+00
1.37476139e-02 1.14227593e+00 -5.07541955e-01 4.49149385e-02
-2.35240385e-01 -8.09222758e-01 7.29749560e-01 1.15321547e-01
8.51343870e-01 -1.00881243e+00 -3.51142526e-01 -1.90942794e-01
-7.55834281e-01 -8.68643343e-01 -5.75716496e-01 -6.69951677e-01
-4.71653312e-01 -8.19206774e-01 -7.18548119e-01 -5.00051320e-01
4.37805057e-01 1.78757131e-01 1.14567542e+00 1.87372133e-01
3.13001037e-01 4.33562011e-01 -3.38698149e-01 -4.10458237e-01
-5.56500614e-01 1.47296071e-01 1.29978508e-01 3.68478633e-02
7.15548038e-01 -5.24424732e-01 -7.52370000e-01 8.16494107e-01
-6.69692039e-01 -1.88260451e-01 6.40984356e-01 4.48392719e-01
-2.72357047e-01 -4.09907013e-01 9.46587503e-01 -1.14470732e+00
1.46948624e+00 -6.23839557e-01 -8.22127536e-02 6.31691933e-01
-6.84676468e-01 -6.86152726e-02 5.98397732e-01 -7.09524930e-01
-1.17377615e+00 -1.15918171e+00 -1.84147701e-01 2.90029913e-01
-2.07718387e-01 1.52913049e-01 -2.28616923e-01 -2.87264407e-01
8.27854514e-01 -7.74679184e-02 3.12175173e-02 -4.30738240e-01
1.34367451e-01 1.20275629e+00 4.45613116e-01 -9.73826349e-01
6.95264161e-01 -1.62977576e-01 -5.77534854e-01 -5.84152222e-01
-7.49206126e-01 -2.68670171e-01 -6.57573193e-02 -4.43528205e-01
5.30277789e-01 -5.17400920e-01 -1.53640103e+00 4.88300323e-01
-1.07316816e+00 -2.33804032e-01 1.96040109e-01 2.56015956e-01
-2.92951375e-01 7.76868641e-01 -8.06454301e-01 -9.85053480e-01
-5.28160512e-01 -8.16290855e-01 6.46780312e-01 4.23200637e-01
-9.44375932e-01 -1.03800070e+00 -5.46390787e-02 9.77171719e-01
5.19452870e-01 -1.76591679e-01 1.13536274e+00 -1.07966876e+00
-5.00442922e-01 -3.34271938e-01 -1.21068306e-01 3.58300716e-01
-7.97292739e-02 -2.78970659e-01 -9.02913749e-01 2.22065635e-02
3.56527805e-01 -3.67199033e-01 2.25944698e-01 -8.29688907e-02
1.02687240e+00 -9.85564828e-01 6.83561387e-03 9.86887813e-02
7.49459565e-01 -1.44706607e-01 7.71569431e-01 4.90291178e-01
4.00557697e-01 1.17544091e+00 6.86223865e-01 4.92956609e-01
8.21617126e-01 3.13274562e-01 8.30086991e-02 3.17732155e-01
3.40543509e-01 -5.17772257e-01 6.00323081e-01 8.02049875e-01
4.96936552e-02 -3.31276089e-01 -6.82764113e-01 2.06492394e-01
-1.73676538e+00 -8.70588124e-01 -3.56506675e-01 2.42887902e+00
9.82546568e-01 1.65163968e-02 2.60282636e-01 -1.84654921e-01
8.33396435e-01 4.40414026e-02 -3.02741557e-01 -6.25590444e-01
5.80090955e-02 -1.69799551e-01 7.24076778e-02 5.32624245e-01
-4.29671526e-01 5.80146134e-01 6.52742052e+00 9.41157222e-01
-5.94181776e-01 1.86531991e-01 9.32472169e-01 -9.96884704e-02
-7.06011295e-01 3.03794276e-02 -8.62254262e-01 7.73184419e-01
9.31476593e-01 -6.18275702e-01 1.99623555e-01 7.46451497e-01
2.36207277e-01 -9.30636153e-02 -1.01280284e+00 8.47357929e-01
-3.04115675e-02 -6.32009625e-01 2.15398088e-01 8.13908118e-04
7.10307837e-01 -6.65261924e-01 2.22027332e-01 7.81510115e-01
4.31468844e-01 -8.72688234e-01 5.92024565e-01 7.50339746e-01
3.27491671e-01 -3.89482409e-01 8.37909937e-01 9.98962641e-01
-4.08291757e-01 -2.30498374e-01 -5.62778354e-01 -5.17253399e-01
-5.82695380e-02 3.79022837e-01 -9.89133596e-01 3.49377990e-01
5.45050681e-01 -2.14707121e-01 -9.78730261e-01 8.25193226e-01
-2.39947826e-01 6.30483568e-01 9.19130743e-02 -8.51775467e-01
3.42770927e-02 -2.08969161e-01 2.38274336e-01 9.78420317e-01
2.95634121e-01 9.11227092e-02 1.81965247e-01 1.14398897e+00
-5.74388504e-02 5.47092617e-01 -1.60194021e-02 3.18691768e-02
8.25438261e-01 1.40481353e+00 -8.51218104e-02 -4.02886897e-01
-2.68373787e-01 6.75862789e-01 4.47482318e-01 3.83273244e-01
-7.74390340e-01 -2.91670322e-01 4.54534203e-01 4.01937306e-01
-5.58752596e-01 1.03842281e-01 -5.83124280e-01 -1.02737713e+00
5.27033359e-02 -1.24517775e+00 3.88977230e-01 -8.36888552e-01
-1.81841612e+00 5.00411928e-01 9.88423750e-02 -8.64528954e-01
-2.09383130e-01 -5.06780505e-01 -9.58820343e-01 1.31396735e+00
-9.62377787e-01 -8.92956734e-01 -5.81780255e-01 3.94956380e-01
1.95466146e-01 -2.01375820e-02 4.88325477e-01 8.67307559e-02
-4.99996901e-01 1.14622426e+00 -3.40707600e-01 -1.86859332e-02
1.14316356e+00 -1.16271639e+00 4.22410548e-01 5.22175789e-01
-4.25270736e-01 1.28701603e+00 7.87928343e-01 -6.96476638e-01
-7.64379084e-01 -7.42021322e-01 1.23843634e+00 -9.49678838e-01
5.58737516e-01 -3.21493179e-01 -1.17157745e+00 5.59333026e-01
2.30636120e-01 -7.65459061e-01 8.38314354e-01 4.29336935e-01
-2.97562629e-01 2.75962725e-02 -1.19078755e+00 1.06902218e+00
8.87188375e-01 -5.68689585e-01 -6.32659256e-01 2.67307281e-01
7.15639830e-01 4.76582609e-02 -7.90639997e-01 1.08791173e-01
5.54720461e-01 -1.43886840e+00 7.89173126e-01 -5.53666890e-01
4.06972378e-01 2.46120095e-02 9.30149481e-02 -9.97093618e-01
-6.28214419e-01 -8.90421093e-01 2.84048796e-01 1.69594133e+00
1.73210174e-01 -1.14426541e+00 7.48425841e-01 1.36935115e+00
3.30359131e-01 -7.40947783e-01 -4.59456533e-01 -6.53124452e-01
2.09312841e-01 -2.00755045e-01 8.86121631e-01 8.80317390e-01
2.69213110e-01 3.16384315e-01 -4.78572011e-01 6.49713576e-02
6.17083132e-01 -3.15573141e-02 1.07941604e+00 -1.19712222e+00
-4.58431751e-01 -5.04409254e-01 -1.40909469e-02 -1.09218097e+00
1.00297414e-01 -6.69156730e-01 -2.63244629e-01 -1.36105835e+00
5.94185114e-01 -4.47809756e-01 -1.07123889e-01 2.52944440e-01
-8.88265073e-01 -1.79114372e-01 2.06604406e-01 7.39858225e-02
-8.50902557e-01 5.32322824e-01 1.28095365e+00 2.02226624e-01
-1.08508550e-01 2.34784022e-01 -1.21670699e+00 6.88423634e-01
9.84884977e-01 -2.85138786e-01 -4.62095827e-01 -7.83654526e-02
3.01901877e-01 -6.44995123e-02 6.78287566e-01 -9.57578897e-01
3.31417859e-01 -4.65485513e-01 1.85648397e-01 -8.48376229e-02
-1.30500659e-01 -1.47171646e-01 -1.29063264e-01 2.48312384e-01
-8.03768337e-01 1.69456303e-01 -2.80162960e-01 2.61707544e-01
9.71469656e-02 -5.14863253e-01 4.14284050e-01 -3.18567216e-01
-3.06193113e-01 2.48980954e-01 -3.33157480e-01 2.06813514e-01
7.13629246e-01 -1.61391854e-01 -6.53378308e-01 -8.42667699e-01
-9.76841822e-02 5.62704742e-01 4.30851340e-01 5.80870688e-01
2.99046606e-01 -1.31736469e+00 -8.38075221e-01 -4.84721288e-02
1.57513306e-01 -7.95431674e-01 6.46644294e-01 8.32527518e-01
-2.58640826e-01 4.14830118e-01 -2.35899210e-01 -2.58156836e-01
-1.40279830e+00 4.48860645e-01 9.64628235e-02 -3.97289485e-01
1.42716542e-01 9.00059819e-01 6.33709013e-01 -7.88968801e-01
1.86183050e-01 -8.99696797e-02 -3.36600780e-01 -1.19208016e-01
6.94201827e-01 7.33021736e-01 -2.79767066e-01 -1.63702071e-01
-4.91820313e-02 5.37153818e-02 -2.27418736e-01 -1.64254919e-01
6.42170072e-01 -4.54094708e-01 -3.42279971e-01 3.62094164e-01
9.15542305e-01 3.86615187e-01 -8.79654646e-01 -3.49023312e-01
-1.20931761e-02 -7.22672284e-01 -7.18107879e-01 -1.03310740e+00
-2.70844132e-01 8.32445085e-01 1.52591184e-01 5.00710726e-01
7.34691322e-01 -1.64860889e-01 7.03025341e-01 4.36019063e-01
5.43615103e-01 -1.12343347e+00 3.90388310e-01 3.72613400e-01
8.90197575e-01 -1.11504507e+00 -2.73089141e-01 -3.56767118e-01
-1.04463446e+00 8.54460418e-01 1.04571056e+00 2.95507491e-01
3.18878032e-02 -2.49742657e-01 3.52948815e-01 1.42651394e-01
-9.18098211e-01 3.36802661e-01 3.62950176e-01 6.22232080e-01
6.63169622e-01 1.86833233e-01 -6.63373649e-01 1.25965452e+00
-7.41609156e-01 -9.47913378e-02 5.45173049e-01 4.58995640e-01
-6.28778934e-01 -1.14816225e+00 -6.40294433e-01 6.07618809e-01
-4.32028145e-01 -1.28953248e-01 -5.87511122e-01 4.83913451e-01
-1.17284551e-01 1.31554878e+00 -4.45773840e-01 -6.60938144e-01
3.47824574e-01 2.21220076e-01 1.26824632e-01 -4.94602799e-01
-1.00724483e+00 -4.78220910e-01 2.87668347e-01 -3.73512983e-01
-3.67382844e-03 -5.06670892e-01 -7.60714948e-01 -8.96485090e-01
-1.20456822e-01 4.53300923e-01 -7.39341080e-02 8.37820470e-01
2.82784611e-01 1.41905949e-01 6.39900088e-01 -2.16409594e-01
-1.18879604e+00 -1.19430625e+00 -5.20441532e-01 6.28391445e-01
-6.56701028e-02 -3.42981011e-01 -4.21464086e-01 -4.78875250e-01] | [11.765214920043945, 8.17653751373291] |
0b265114-dd4b-4994-beac-ab5d30134bdd | falcon-fast-visual-concept-learning-by-1 | 2203.16639 | null | https://arxiv.org/abs/2203.16639v1 | https://arxiv.org/pdf/2203.16639v1.pdf | FALCON: Fast Visual Concept Learning by Integrating Images, Linguistic descriptions, and Conceptual Relations | We present a meta-learning framework for learning new visual concepts quickly, from just one or a few examples, guided by multiple naturally occurring data streams: simultaneously looking at images, reading sentences that describe the objects in the scene, and interpreting supplemental sentences that relate the novel concept with other concepts. The learned concepts support downstream applications, such as answering questions by reasoning about unseen images. Our model, namely FALCON, represents individual visual concepts, such as colors and shapes, as axis-aligned boxes in a high-dimensional space (the "box embedding space"). Given an input image and its paired sentence, our model first resolves the referential expression in the sentence and associates the novel concept with particular objects in the scene. Next, our model interprets supplemental sentences to relate the novel concept with other known concepts, such as "X has property Y" or "X is a kind of Y". Finally, it infers an optimal box embedding for the novel concept that jointly 1) maximizes the likelihood of the observed instances in the image, and 2) satisfies the relationships between the novel concepts and the known ones. We demonstrate the effectiveness of our model on both synthetic and real-world datasets. | ['Joshua B. Tenenbaum', 'Chuang Gan', 'Ziqi Wang', 'Jiayuan Mao', 'Lingjie Mei'] | 2022-03-30 | falcon-fast-visual-concept-learning-by | https://openreview.net/forum?id=htWIlvDcY8 | https://openreview.net/pdf?id=htWIlvDcY8 | iclr-2022-4 | ['novel-concepts'] | ['reasoning'] | [ 3.54723960e-01 2.31321201e-01 4.00225334e-02 -6.70887828e-01
-6.26327336e-01 -6.96686566e-01 6.31424665e-01 5.94331563e-01
-1.61892906e-01 4.59441066e-01 2.44453296e-01 -5.92153333e-02
-1.34340059e-02 -7.85510480e-01 -1.12782252e+00 -5.01898944e-01
-4.49545905e-02 4.54305559e-01 7.79377595e-02 1.20473407e-01
2.62539595e-01 2.15582460e-01 -1.88345194e+00 6.69323683e-01
3.72643769e-01 8.95147920e-01 6.17946684e-01 6.72254741e-01
-5.58828175e-01 7.44081914e-01 -7.85666287e-01 -3.60392213e-01
4.86944467e-02 -4.73704070e-01 -8.82268488e-01 5.76365292e-01
7.41230607e-01 -3.12973708e-01 -1.59981236e-01 1.02755046e+00
-2.51157165e-01 3.19140345e-01 6.48766339e-01 -1.44458854e+00
-1.11827695e+00 3.60608697e-01 -6.31088376e-01 2.13092372e-01
4.34454381e-01 2.33647674e-01 1.12706387e+00 -1.17777443e+00
8.95093977e-01 1.42624605e+00 -2.12131278e-03 7.99194753e-01
-1.56738329e+00 -4.36017781e-01 6.74253583e-01 3.22418660e-01
-1.27664363e+00 -2.79003710e-01 8.71744752e-01 -4.29702550e-01
7.59123743e-01 2.19896629e-01 7.73289680e-01 1.16187119e+00
-6.99036941e-02 8.74046028e-01 9.76939559e-01 -3.53160799e-01
7.29626298e-01 6.41594708e-01 8.65205228e-02 5.75320601e-01
1.54282853e-01 -1.48519203e-01 -7.36814380e-01 1.30894510e-02
3.42539459e-01 2.80452430e-01 7.07228258e-02 -5.97367227e-01
-1.42629123e+00 6.39537990e-01 6.15735769e-01 2.29187414e-01
-2.59190351e-01 2.96879113e-01 1.35060295e-01 -7.36440271e-02
4.18264657e-01 6.12724185e-01 -4.94798779e-01 1.43135205e-01
-5.13677418e-01 2.06320345e-01 3.88024211e-01 1.27475309e+00
1.12647188e+00 -4.23881263e-01 9.69113111e-02 4.43822891e-01
3.39665890e-01 6.97994173e-01 2.50644833e-01 -1.02623069e+00
4.93718058e-01 8.01285803e-01 8.32620710e-02 -1.07003641e+00
-1.32295815e-02 -1.00362353e-01 -5.79762578e-01 8.61526951e-02
1.58851400e-01 1.12541609e-01 -7.27888405e-01 1.92734730e+00
3.72148246e-01 2.57136077e-01 5.39115787e-01 8.74794066e-01
1.05831754e+00 1.04949272e+00 2.22081780e-01 -1.44251451e-01
1.57635581e+00 -8.53950322e-01 -3.58761132e-01 -6.96045697e-01
2.41847098e-01 -2.40588143e-01 1.27596521e+00 5.58485426e-02
-9.03328359e-01 -7.81207919e-01 -9.06417966e-01 -4.40787643e-01
-5.61056435e-01 4.87766782e-04 2.93379605e-01 1.94350258e-01
-9.40460026e-01 1.84413850e-01 -3.75467837e-01 -7.37877965e-01
5.80929756e-01 -8.07965100e-02 -3.84325624e-01 -1.46049798e-01
-7.47569919e-01 5.09467661e-01 4.74886924e-01 -1.69226706e-01
-1.15577400e+00 -7.79957771e-01 -1.12835515e+00 3.85171801e-01
6.29640818e-01 -1.05442524e+00 1.15757298e+00 -1.40951037e+00
-7.94215858e-01 1.22419286e+00 -7.81326354e-01 -3.97221118e-01
6.83127567e-02 -2.25852713e-01 -5.10639429e-01 4.21108544e-01
3.42341304e-01 1.19519317e+00 1.20272923e+00 -1.95646834e+00
-7.64328897e-01 -5.91010332e-01 6.17224276e-01 2.81921327e-01
-2.13601664e-01 -1.63013786e-01 -2.34200612e-01 -3.93904001e-01
3.89704227e-01 -4.06045258e-01 9.58246272e-03 7.44679451e-01
-4.14957225e-01 -1.12073250e-01 9.72196817e-01 -1.89805418e-01
6.20413125e-01 -2.53153276e+00 2.11091712e-02 2.07881145e-02
5.28287351e-01 -4.98041809e-01 -7.91373178e-02 3.04172426e-01
-3.42123210e-01 1.10112533e-01 -2.19298095e-01 -3.56690824e-01
-8.30683112e-02 4.52321917e-01 -8.27705264e-01 8.99437666e-02
4.18722808e-01 9.86305296e-01 -1.33602130e+00 -4.64297563e-01
2.61879206e-01 7.42178932e-02 -1.99357346e-01 4.50568348e-01
-7.98044622e-01 2.90023625e-01 -2.10950315e-01 3.53458017e-01
5.44236720e-01 -5.61769068e-01 9.63701010e-02 -2.63624281e-01
5.23007363e-02 -1.70002043e-01 -9.98913348e-01 1.62597060e+00
-3.23367327e-01 8.41035426e-01 -4.37357843e-01 -8.73449981e-01
1.14538348e+00 -1.10394448e-01 1.56093594e-02 -4.75221038e-01
-2.09208950e-01 -3.29825222e-01 -4.47493315e-01 -1.17523563e+00
4.18908417e-01 -3.86962295e-01 -6.22936003e-02 5.87664366e-01
5.88877301e-04 -1.58542871e-01 1.72089785e-01 6.25241280e-01
5.24913311e-01 1.38630912e-01 4.58141714e-01 -1.73488274e-01
4.86210436e-01 -4.67261150e-02 1.96322933e-01 9.55175281e-01
-6.96762577e-02 5.15453279e-01 5.91317773e-01 -6.78431273e-01
-1.07506013e+00 -1.67366576e+00 2.66692907e-01 1.27395415e+00
6.69458866e-01 -5.35458505e-01 -3.07566106e-01 -5.44128418e-01
9.32779536e-02 8.93757403e-01 -1.02941811e+00 -5.57365641e-02
-1.69781789e-01 2.46596858e-02 -7.19218552e-02 8.09472084e-01
3.68533552e-01 -1.10836279e+00 -1.07055390e+00 -2.12271050e-01
-1.47131532e-01 -1.02304101e+00 -2.75521010e-01 1.66543737e-01
-9.17382419e-01 -1.10857928e+00 -2.84550130e-01 -1.09780681e+00
1.15941477e+00 4.24881548e-01 1.15042686e+00 1.30388230e-01
-4.37952161e-01 6.16500258e-01 -3.45894158e-01 -3.39957565e-01
-2.36063763e-01 -7.30058908e-01 -8.76529142e-02 3.29218656e-01
4.33949202e-01 -2.22586542e-01 -4.67493415e-01 -1.49217933e-01
-1.02032101e+00 5.22594631e-01 3.51165533e-01 6.03205204e-01
5.78728318e-01 -1.15802716e-02 3.96848500e-01 -9.54571247e-01
2.95766115e-01 -6.30877912e-01 -4.50948119e-01 6.06739879e-01
-2.04216942e-01 3.36879909e-01 5.86413503e-01 -4.78118837e-01
-1.15208530e+00 8.55012611e-02 5.09500325e-01 -3.92193496e-01
-7.11772025e-01 4.75054950e-01 -3.66148084e-01 8.96189928e-01
7.19089985e-01 4.66666043e-01 -1.87824175e-01 -1.62664905e-01
9.31347847e-01 2.85639316e-01 8.53454709e-01 -9.08910036e-01
8.58345509e-01 1.04702222e+00 -9.32319462e-02 -1.03812540e+00
-1.15492177e+00 -4.41688091e-01 -9.48028862e-01 -2.47153237e-01
1.11695027e+00 -8.01438808e-01 -7.20290899e-01 -3.03849671e-02
-1.48865581e+00 1.03595391e-01 -6.40697241e-01 1.60164282e-01
-6.55628920e-01 2.82110363e-01 4.75432165e-02 -1.03846288e+00
2.36655876e-01 -8.37349832e-01 1.28865349e+00 4.90492851e-01
-2.64465511e-01 -1.02302957e+00 -1.62874281e-01 1.93001673e-01
1.04967477e-02 2.53974885e-01 1.48111212e+00 -4.74748343e-01
-7.60936558e-01 2.57480443e-01 -5.07263005e-01 -1.33629870e-02
1.94044679e-01 1.28446490e-01 -1.26815617e+00 -2.09262580e-01
-2.68539246e-02 -2.99356729e-01 7.16257036e-01 1.48794904e-01
1.47350276e+00 -4.35959876e-01 -4.64576304e-01 3.77154708e-01
1.48997271e+00 3.62139583e-01 3.17661554e-01 1.88946456e-01
2.60279924e-01 8.91366303e-01 3.25255543e-01 5.07550240e-01
3.57453614e-01 1.82436451e-01 5.08145750e-01 -1.18139647e-01
2.14322627e-01 -5.41667163e-01 1.92373708e-01 2.19508976e-01
5.83013058e-01 -2.51804084e-01 -7.03007638e-01 6.68423116e-01
-1.94786143e+00 -1.08515131e+00 3.49557489e-01 1.91410351e+00
6.54730737e-01 1.32004350e-01 -4.07724455e-02 -4.42865402e-01
8.43551397e-01 1.69232816e-01 -1.15385079e+00 -2.97934294e-01
-3.37611020e-01 1.65943634e-02 -2.85070240e-01 3.83298606e-01
-9.92094934e-01 8.62989008e-01 5.93372059e+00 3.48477922e-02
-8.56392324e-01 -1.37073964e-01 8.34298134e-01 -1.00514978e-01
-6.36235178e-01 4.49338704e-01 -4.60728914e-01 7.12434798e-02
3.61419469e-01 -6.63762629e-01 4.27775532e-01 9.78307903e-01
9.39551182e-03 -2.95918018e-01 -1.78531981e+00 1.00457978e+00
7.70474017e-01 -1.50098777e+00 5.99464178e-01 -3.67201835e-01
6.20757461e-01 -5.98518670e-01 3.49097580e-01 2.23638460e-01
2.98292637e-02 -9.80061054e-01 8.73261511e-01 7.70179212e-01
8.29938412e-01 -3.87541980e-01 2.70395517e-01 4.46274251e-01
-1.07241893e+00 -1.82401791e-01 -4.68674779e-01 -2.90656596e-01
-1.11132637e-01 1.84904456e-01 -9.49187994e-01 1.42989889e-01
8.77172112e-01 1.00681615e+00 -8.91010702e-01 7.15067506e-01
-5.22830486e-01 5.73200025e-02 4.25199196e-02 -2.19541267e-01
5.60899153e-02 -1.39018461e-01 4.88712251e-01 8.89256358e-01
4.34668884e-02 2.93859333e-01 -2.58616433e-02 1.62124908e+00
-2.30570018e-01 -8.42371583e-02 -7.94435859e-01 3.95015851e-02
5.58286190e-01 1.12906337e+00 -7.45570719e-01 -8.37244749e-01
-5.27555764e-01 9.27092731e-01 3.72464240e-01 7.83700049e-01
-5.03366113e-01 -1.43211365e-01 6.89078689e-01 3.32090333e-02
3.17489505e-01 1.15641800e-03 -2.54139543e-01 -1.16052198e+00
1.55667990e-01 -4.60662752e-01 6.33588016e-01 -1.78403020e+00
-1.50598323e+00 5.01990438e-01 2.59865701e-01 -1.30067182e+00
-7.06156567e-02 -7.14666188e-01 -6.29824162e-01 6.17614985e-01
-1.26217902e+00 -9.98640358e-01 -5.35801411e-01 5.71739256e-01
9.14016008e-01 -1.66183654e-02 9.71173048e-01 -5.04141331e-01
-3.92230079e-02 1.32588595e-01 3.23781418e-03 1.30416781e-01
5.79663634e-01 -1.47788095e+00 5.35065174e-01 7.49540567e-01
5.77503920e-01 9.18029070e-01 7.93577313e-01 -5.25529802e-01
-1.21566844e+00 -1.11698794e+00 8.08990479e-01 -6.93100214e-01
6.85303748e-01 -8.68967891e-01 -1.03351390e+00 7.38609731e-01
2.80585676e-01 6.00395538e-03 7.96481669e-01 2.53231991e-02
-8.92557740e-01 -2.61907727e-01 -1.10146117e+00 8.01060259e-01
8.61101806e-01 -8.29299152e-01 -1.22374666e+00 3.69304508e-01
1.12437344e+00 -2.49130234e-01 -1.91149712e-01 -8.16843733e-02
4.37432081e-01 -9.11855757e-01 1.02234530e+00 -1.40424204e+00
8.90733898e-01 -4.82879132e-01 -4.01227921e-01 -1.30136418e+00
-2.80828118e-01 6.86000362e-02 -1.88629776e-01 8.94681334e-01
5.38151085e-01 -1.95037231e-01 7.84264088e-01 7.30917871e-01
1.66658059e-01 -6.26203060e-01 -8.46450627e-01 -6.38432264e-01
-1.75301611e-01 -4.65935588e-01 7.95410573e-01 8.08132708e-01
1.66951388e-01 6.30701423e-01 -1.64030075e-01 3.03290874e-01
5.64184904e-01 7.38493741e-01 8.92392457e-01 -1.10154498e+00
-2.60161072e-01 -7.75136501e-02 -5.88634491e-01 -1.25360954e+00
3.21641356e-01 -7.50769317e-01 1.81278765e-01 -1.53026998e+00
8.49752605e-01 2.76219361e-02 -8.39165151e-02 4.21526611e-01
-2.44393423e-01 -1.77785620e-01 4.22788143e-01 7.23483134e-03
-8.95637870e-01 5.44589818e-01 1.02372837e+00 -6.92530990e-01
-1.14905750e-02 -4.76766169e-01 -9.77288961e-01 7.42072582e-01
4.17727113e-01 -2.59138286e-01 -5.41972697e-01 -7.47989058e-01
4.96359646e-01 5.68642691e-02 7.21649289e-01 -7.22517788e-01
2.69457459e-01 -4.13052857e-01 7.52056420e-01 -6.51426494e-01
4.71222550e-01 -1.20907390e+00 -1.86436668e-01 2.64284700e-01
-8.66677701e-01 2.55383223e-01 1.90797880e-01 8.39687049e-01
-6.94032609e-02 -2.18577459e-01 5.61022878e-01 -2.46464431e-01
-1.18804610e+00 1.06705457e-01 -1.27044380e-01 2.30137855e-01
1.27075720e+00 -4.52279180e-01 -7.70008624e-01 -5.42350292e-01
-1.05451715e+00 3.36371988e-01 5.62929273e-01 8.39508414e-01
1.24607873e+00 -1.15881598e+00 -5.44801056e-01 4.51482624e-01
9.52881634e-01 2.76745856e-01 4.19675052e-01 -3.07265297e-02
-2.52283603e-01 2.43461840e-02 -1.42550677e-01 -9.28481996e-01
-1.23030841e+00 1.27500319e+00 2.38976777e-01 6.26899600e-01
-8.43671024e-01 9.18715537e-01 9.33408082e-01 -8.19538459e-02
2.87484348e-01 -5.57298839e-01 -6.98102415e-02 -1.89296693e-01
7.49170244e-01 -1.83325753e-01 -6.30157292e-01 -6.56822979e-01
-4.25916195e-01 5.74626625e-01 -1.84255451e-01 -2.55242139e-01
1.24457586e+00 -3.91534150e-01 -2.95463830e-01 1.20449424e+00
1.38531744e+00 -3.78975958e-01 -1.40647709e+00 -5.60330033e-01
4.81853709e-02 -7.89534271e-01 -6.02549613e-01 -7.97222972e-01
-5.61808407e-01 1.12165844e+00 4.59017634e-01 -1.43091485e-01
1.05790222e+00 8.76253486e-01 3.32621008e-01 6.35578871e-01
1.94072381e-01 -8.63554180e-01 8.21121991e-01 2.29844943e-01
9.54259574e-01 -1.26912320e+00 7.96710048e-03 -2.63882756e-01
-8.43577445e-01 1.28428173e+00 9.61967468e-01 9.70382690e-02
4.35817361e-01 -8.93842205e-02 1.85047716e-01 -5.43359697e-01
-1.19645238e+00 -1.33293532e-02 2.40497410e-01 7.13333547e-01
9.85235255e-03 1.19989522e-01 5.06646931e-01 5.49163520e-01
-8.99327919e-02 -4.15965915e-01 5.37766516e-01 7.99498618e-01
-6.39257133e-01 -3.57713044e-01 -3.60185653e-01 3.99243861e-01
4.74450231e-01 -5.22867553e-02 -6.86652422e-01 6.21092379e-01
3.98949921e-01 7.83145070e-01 7.56156147e-01 -1.24419935e-01
2.88167685e-01 3.85937840e-02 4.32618558e-01 -1.05561233e+00
5.54324016e-02 -2.35835999e-01 -5.20845532e-01 -2.92699605e-01
-5.05638421e-01 -6.41735077e-01 -1.46338820e+00 2.25431249e-01
1.15607046e-01 1.35018483e-01 5.54425597e-01 1.25095201e+00
3.43772829e-01 2.75312036e-01 6.21293485e-01 -1.69898033e-01
-1.29906535e-01 -5.21600783e-01 -5.15389085e-01 9.74017262e-01
7.30463803e-01 -5.02267420e-01 -3.92007619e-01 6.75489187e-01] | [10.223636627197266, 1.9073407649993896] |
4134c2ef-1a51-4ea3-8727-db9f624c8d0c | small-text-active-learning-for-text | 2107.10314 | null | https://arxiv.org/abs/2107.10314v6 | https://arxiv.org/pdf/2107.10314v6.pdf | Small-Text: Active Learning for Text Classification in Python | We introduce small-text, an easy-to-use active learning library, which offers pool-based active learning for single- and multi-label text classification in Python. It features numerous pre-implemented state-of-the-art query strategies, including some that leverage the GPU. Standardized interfaces allow the combination of a variety of classifiers, query strategies, and stopping criteria, facilitating a quick mix and match, and enabling a rapid and convenient development of both active learning experiments and applications. With the objective of making various classifiers and query strategies accessible for active learning, small-text integrates several well-known machine learning libraries, namely scikit-learn, PyTorch, and Hugging Face transformers. The latter integrations are optionally installable extensions, so GPUs can be used but are not required. Using this new library, we investigate the performance of the recently published SetFit training paradigm, which we compare to vanilla transformer fine-tuning, finding that it matches the latter in classification accuracy while outperforming it in area under the curve. The library is available under the MIT License at https://github.com/webis-de/small-text, in version 1.3.0 at the time of writing. | ['Martin Potthast', 'Andreas Niekler', 'Lydia Müller', 'Christopher Schröder'] | 2021-07-21 | null | null | null | null | ['classification'] | ['methodology'] | [ 2.02802964e-03 -3.48439097e-01 -4.85726476e-01 -3.83454889e-01
-1.23279178e+00 -9.76483107e-01 6.86581254e-01 6.66461587e-01
-7.39224672e-01 6.35167241e-01 -2.39746869e-01 -4.10801619e-01
1.69409830e-02 -6.20390356e-01 -3.38660628e-01 -7.51105726e-01
3.02012920e-01 8.17426383e-01 4.94038850e-01 -2.87011433e-02
2.88220495e-01 2.69303471e-01 -1.67649698e+00 4.62517738e-01
7.01683283e-01 7.31809318e-01 8.02224055e-02 8.50878119e-01
-3.97183865e-01 8.09447348e-01 -3.15123826e-01 -4.67085928e-01
3.19763720e-02 9.49139968e-02 -8.87225270e-01 -3.83550882e-01
4.99886692e-01 2.89924303e-03 5.32282703e-02 2.42988482e-01
9.12553847e-01 1.59717023e-01 2.77756184e-01 -1.21591175e+00
-9.48226377e-02 1.65224433e-01 -4.63620201e-02 1.20106861e-01
5.68423271e-01 4.04334128e-01 1.12868738e+00 -1.05598927e+00
4.80765224e-01 7.06100345e-01 7.94409931e-01 4.26854670e-01
-1.28503823e+00 -7.66093075e-01 -1.28520370e-01 1.25205517e-01
-1.28071296e+00 -1.01836860e+00 3.87595773e-01 -5.00633061e-01
1.27233577e+00 6.10132337e-01 6.68470979e-01 1.31345975e+00
1.81736350e-02 1.20337999e+00 1.03524911e+00 -7.95659006e-01
4.55807418e-01 4.46358532e-01 5.35791934e-01 8.43419373e-01
-9.24748406e-02 -1.45392627e-01 -8.44942570e-01 -8.97700429e-01
6.84482828e-02 6.95013776e-02 -1.28048167e-01 -5.04504204e-01
-1.17231357e+00 8.22624385e-01 5.10221766e-03 1.99073225e-01
1.65337846e-02 -6.50436208e-02 6.00893021e-01 9.87584293e-02
8.90702903e-01 5.14085770e-01 -7.82412648e-01 -5.84579647e-01
-8.77194881e-01 2.18975931e-01 1.05621529e+00 8.13887537e-01
7.70530045e-01 -5.93131125e-01 -1.98909417e-01 9.05889809e-01
4.43465829e-01 -7.85621181e-02 3.81576896e-01 -7.68753707e-01
2.48134494e-01 9.13991630e-01 1.36865780e-03 -4.77469862e-02
-4.40974683e-01 -4.49495763e-01 -1.61779061e-01 3.09664428e-01
5.44815302e-01 8.29257630e-03 -4.59214211e-01 1.10088587e+00
5.66873968e-01 4.19491716e-02 -3.70646745e-01 2.58216619e-01
1.06499386e+00 2.21245930e-01 4.43890363e-01 -3.28374773e-01
1.13949788e+00 -1.03590012e+00 -3.80454123e-01 -2.43530124e-01
1.13005471e+00 -9.86827314e-01 1.34093595e+00 4.10040140e-01
-9.24185574e-01 -1.35895014e-01 -9.82527018e-01 -2.74832129e-01
-8.28792751e-01 -3.71249579e-02 8.22957695e-01 8.37363183e-01
-1.02021325e+00 5.70604563e-01 -1.14439976e+00 -5.23390412e-01
4.88018006e-01 3.76840919e-01 -2.51527727e-01 1.04272909e-01
-8.97533298e-01 6.82563066e-01 4.02235419e-01 -4.17861968e-01
-5.73726594e-01 -9.34681594e-01 -5.56044638e-01 -8.67478251e-02
5.35722136e-01 -6.22680485e-01 1.73940051e+00 -7.72711813e-01
-1.52630770e+00 1.27947354e+00 -7.96994045e-02 8.38547349e-02
7.24068165e-01 -2.16413274e-01 -2.14742012e-02 -1.26014709e-01
-4.47318703e-02 5.49208879e-01 3.03248078e-01 -5.57052612e-01
-3.65789354e-01 -4.46323752e-01 -5.13262637e-02 4.80510324e-01
-4.06881064e-01 3.01570028e-01 -5.94370544e-01 -2.77274102e-01
-3.34924638e-01 -8.47297370e-01 7.48890117e-02 4.58116978e-01
-1.14784434e-01 -5.71811438e-01 8.84733975e-01 -1.63903788e-01
1.06290901e+00 -2.15387607e+00 -3.78220558e-01 -1.65873483e-01
3.38474303e-01 2.79187351e-01 2.07298070e-01 6.89400256e-01
-6.00746507e-03 3.53383273e-02 -2.74162650e-01 -8.59272122e-01
1.05353005e-01 -3.51852365e-02 9.08837765e-02 5.53056061e-01
-2.11487710e-02 9.20010090e-01 -9.89866376e-01 -7.67593801e-01
3.75423044e-01 3.38658959e-01 -2.79664159e-01 3.63625288e-01
-4.49162275e-01 3.61496508e-01 -2.68868327e-01 6.13792956e-01
3.77648711e-01 -5.05669236e-01 2.12354600e-01 2.40768030e-01
-4.75128710e-01 6.61466002e-01 -1.04577851e+00 1.83182633e+00
-3.51914614e-01 5.26587069e-01 -4.93618846e-02 -6.79240644e-01
7.19585359e-01 4.62743253e-01 5.08244455e-01 -5.35769284e-01
6.63255379e-02 3.62020612e-01 -4.86075044e-01 -3.34053308e-01
3.95555824e-01 4.48018163e-01 1.12906642e-01 8.64350855e-01
3.30194116e-01 2.38337703e-02 3.96920353e-01 4.46486861e-01
1.15029168e+00 3.76279444e-01 5.90360165e-01 -2.50279129e-01
1.60551056e-01 1.49341047e-01 1.84745789e-01 9.27411675e-01
-1.61987916e-01 3.54403824e-01 1.75901398e-01 -8.70563313e-02
-7.14813709e-01 -9.06300962e-01 -4.37459141e-01 1.99099064e+00
-4.64730322e-01 -7.65216053e-01 -3.87228191e-01 -7.06759334e-01
-1.50499567e-01 7.82938540e-01 -3.98446918e-01 2.60130733e-01
-1.61293209e-01 -1.01326025e+00 5.44193506e-01 1.86999157e-01
3.24133992e-01 -9.64491844e-01 -3.98866355e-01 8.56584087e-02
-1.97207648e-02 -5.11284292e-01 -2.74593532e-01 7.03240693e-01
-8.21714640e-01 -1.21262825e+00 -5.08110046e-01 -5.22387087e-01
2.34992579e-01 8.74067843e-02 1.49200583e+00 2.93139428e-01
-3.86570126e-01 7.02201903e-01 -4.70831603e-01 -6.15792692e-01
-3.35494429e-01 6.60580337e-01 -3.83002937e-01 -4.30117249e-01
4.24208462e-01 -2.48740256e-01 -5.84703565e-01 2.20768243e-01
-6.42222404e-01 1.00699835e-01 6.82571679e-02 7.50155210e-01
5.23122728e-01 -5.17973483e-01 1.18086874e-01 -1.10391736e+00
4.67931151e-01 -5.74268997e-01 -6.93541646e-01 4.55283701e-01
-1.02077734e+00 -1.39617696e-01 2.96754122e-01 -2.17379496e-01
-8.54514658e-01 3.60193342e-01 -6.06420457e-01 7.35256076e-02
-1.79203302e-01 2.64388800e-01 -1.74206957e-01 -2.30195358e-01
9.98544633e-01 1.31011084e-01 -5.33508696e-02 -4.97994423e-01
1.65268242e-01 9.78441954e-01 2.50123590e-02 -4.81184483e-01
1.92568719e-01 1.86610997e-01 -5.32806516e-01 -5.08374214e-01
-9.13031638e-01 -1.04333258e+00 -6.43499196e-01 -1.91493481e-01
3.94995451e-01 -8.01398695e-01 -8.19496751e-01 8.76633525e-01
-7.06188917e-01 -8.40650558e-01 -1.69972196e-01 1.84836820e-01
-4.70905632e-01 3.01313192e-01 -5.01183867e-01 -6.78798437e-01
-8.55024934e-01 -1.12474406e+00 1.11737728e+00 1.54502109e-01
-5.14147282e-01 -1.25429893e+00 5.50456583e-01 5.54019868e-01
5.06793022e-01 -1.83611900e-01 5.39785802e-01 -1.27550423e+00
-4.11042631e-01 -3.87235403e-01 2.34818950e-01 -3.79267335e-02
-2.20511913e-01 3.63955170e-01 -1.42661560e+00 -4.50090587e-01
-3.67572159e-01 -5.77668607e-01 6.94404125e-01 5.69928996e-02
1.09579253e+00 -1.10946208e-01 -5.76482773e-01 4.97674972e-01
1.18305993e+00 7.96475187e-02 6.45699799e-01 4.68944669e-01
3.72261822e-01 1.62265375e-01 4.58070457e-01 4.63235915e-01
2.47765914e-01 9.10199881e-01 1.75596505e-01 -2.73512751e-01
3.68229061e-01 1.54549897e-01 1.62453324e-01 7.20678508e-01
3.37336987e-01 -1.78817615e-01 -1.24476445e+00 1.08538363e-02
-1.94225800e+00 -8.66264939e-01 -3.90940577e-01 2.57814956e+00
1.25283670e+00 2.77483165e-01 3.11922938e-01 3.46621215e-01
2.60548264e-01 -5.88031225e-02 -6.10269487e-01 -1.79559007e-01
6.60041720e-02 6.07863963e-01 2.78365940e-01 5.64995825e-01
-1.18420577e+00 6.08726144e-01 6.06456852e+00 1.11875033e+00
-1.34371495e+00 5.74873269e-01 6.09351516e-01 -2.93853492e-01
9.05375406e-02 1.10159956e-01 -1.00994682e+00 4.24025655e-01
9.80682611e-01 -1.01348139e-01 1.75265059e-01 1.00226736e+00
-1.49657973e-03 -3.52561712e-01 -1.07358491e+00 7.51484454e-01
-2.25002974e-01 -1.46805239e+00 -6.41647756e-01 -2.06654027e-01
1.20422728e-01 4.63316202e-01 -2.50993252e-01 6.15798235e-01
1.99936181e-01 -5.77716470e-01 7.04251945e-01 3.93951714e-01
8.92642498e-01 -3.47155541e-01 5.01031756e-01 5.68537176e-01
-9.31773603e-01 6.56195208e-02 1.19105048e-01 -8.30772966e-02
-3.33990306e-01 6.31305337e-01 -9.55671787e-01 3.12250763e-01
5.99347472e-01 5.24801731e-01 -9.38055813e-01 1.40697908e+00
1.23112835e-01 9.33246493e-01 -5.12865782e-01 -2.92311966e-01
-3.01425815e-01 1.94573373e-01 3.71972203e-01 1.47164786e+00
-1.89624250e-01 -1.45659998e-01 3.03257972e-01 5.57316482e-01
6.33634701e-02 5.38304865e-01 -2.13204771e-01 1.14326730e-01
6.98897064e-01 1.55474937e+00 -8.13258469e-01 -4.02857155e-01
-3.62578154e-01 7.53252089e-01 5.15266478e-01 2.13116072e-02
-6.00075662e-01 -4.17774260e-01 1.73322603e-01 2.26424918e-01
-1.79515705e-01 -1.75921127e-01 -3.90555680e-01 -1.07308090e+00
-3.04245234e-01 -9.77144718e-01 7.39104748e-01 -8.46644700e-01
-1.27964151e+00 1.90190837e-01 1.15931302e-01 -9.81229663e-01
-1.56424910e-01 -6.56131864e-01 -5.78889012e-01 7.46886671e-01
-9.30681765e-01 -1.00725996e+00 -5.00471175e-01 6.32195115e-01
4.32591468e-01 -1.98918983e-01 1.35712945e+00 4.30199265e-01
-8.46133828e-01 8.94282341e-01 5.16047955e-01 7.40053058e-02
1.05973446e+00 -1.20826566e+00 4.68744040e-01 2.46179879e-01
1.51609793e-01 6.90918922e-01 3.73988420e-01 -3.78071934e-01
-1.26644850e+00 -8.54274929e-01 9.57791924e-01 -6.81549311e-01
6.24149144e-01 -9.54762995e-01 -9.08642650e-01 7.96485126e-01
1.23443559e-01 -4.09964621e-02 1.15899873e+00 5.02916217e-01
-3.58393937e-01 2.55958848e-02 -1.06778240e+00 3.96124363e-01
7.61195123e-01 -6.61896169e-01 -1.33628011e-01 7.28455663e-01
3.09343219e-01 -5.46647787e-01 -7.17295051e-01 1.80818960e-01
6.20450974e-01 -1.03880894e+00 7.49037445e-01 -2.12364793e-01
-1.77373216e-02 1.54789716e-01 1.08543158e-01 -8.35612059e-01
-3.96803051e-01 -8.13997746e-01 -1.23264328e-01 1.32743895e+00
7.62779236e-01 -1.10660076e+00 8.28069806e-01 5.27132154e-01
-2.70825803e-01 -9.76380587e-01 -9.64353204e-01 -2.58165479e-01
-2.07154965e-03 -6.03994548e-01 2.33533606e-01 1.18917751e+00
2.64870703e-01 5.39633989e-01 2.99635619e-01 -4.81459290e-01
4.61914718e-01 -2.69235790e-01 7.46596754e-01 -1.37337983e+00
-5.79905808e-01 -4.28206861e-01 -6.66862875e-02 -7.84733951e-01
-8.83644894e-02 -1.14153457e+00 -2.33379290e-01 -1.27201462e+00
4.22298640e-01 -8.55395079e-01 -1.22384347e-01 1.22506571e+00
-3.93893450e-01 3.94210696e-01 5.45502501e-03 6.04587317e-01
-8.87346148e-01 3.17834407e-01 6.56752110e-01 -1.34469315e-01
-3.54189485e-01 2.23825738e-01 -4.98033077e-01 3.63363385e-01
9.93149519e-01 -6.41962588e-01 -1.78385600e-01 -5.09811454e-02
4.51757997e-01 -3.59377533e-01 2.74934232e-01 -9.48625326e-01
3.09802383e-01 6.54820427e-02 2.89031386e-01 -4.44382668e-01
4.10408080e-01 -4.38234717e-01 2.56439567e-01 2.09649399e-01
-6.64987743e-01 -7.05578923e-02 4.29693758e-01 1.60342261e-01
3.27252686e-01 -4.59871620e-01 8.35621059e-01 -2.25929856e-01
-4.95363742e-01 3.67647201e-01 -4.85537797e-01 2.03091204e-01
9.19429064e-01 -2.66901404e-01 -7.19804943e-01 -2.29358468e-02
-5.77769578e-01 2.14337669e-02 5.56074858e-01 1.66641623e-01
-9.78733897e-02 -8.16123903e-01 -6.90787375e-01 6.47349507e-02
6.20384634e-01 -2.05192089e-01 2.92371884e-02 1.09509838e+00
-5.02674580e-01 2.79413193e-01 1.97434559e-01 -7.86710203e-01
-1.66461039e+00 1.78252548e-01 6.32265925e-01 -4.46179420e-01
-2.69185722e-01 8.67315829e-01 -1.53693825e-01 -9.45348799e-01
3.48608375e-01 2.59078145e-01 -4.16646563e-02 1.56189010e-01
4.22338814e-01 4.43449140e-01 7.18741953e-01 3.51331122e-02
-4.07651097e-01 -9.04142391e-03 -2.35003129e-01 5.63874058e-02
1.13802361e+00 1.67732358e-01 -7.38490373e-02 9.18436468e-01
1.17585897e+00 7.26982355e-02 -6.78746343e-01 -2.79434443e-01
6.37756512e-02 -2.43575186e-01 3.09570372e-01 -1.10807121e+00
-5.84227085e-01 6.39530241e-01 1.01084483e+00 1.69917062e-01
8.74151886e-01 2.38677841e-02 2.49663875e-01 4.26791519e-01
2.81907916e-01 -9.41661775e-01 1.66368410e-02 6.34384334e-01
4.60445017e-01 -1.19507015e+00 4.30739254e-01 -3.73697221e-01
-3.28796625e-01 9.82988834e-01 4.14615154e-01 4.36046898e-01
5.61471164e-01 6.13340020e-01 2.66138971e-01 -2.73144990e-01
-1.21519864e+00 -1.36320293e-01 6.04425110e-02 3.15672994e-01
1.02575421e+00 -5.16158119e-02 -2.51678765e-01 9.51496884e-02
4.75910679e-02 2.54211605e-01 9.28517580e-02 1.40636599e+00
-2.36554131e-01 -1.35930145e+00 -3.61836642e-01 6.37778282e-01
-4.31546718e-01 -2.94047207e-01 -5.87408960e-01 6.52766168e-01
-7.69189373e-02 8.54751229e-01 1.78619415e-01 -3.02784946e-02
1.72665671e-01 5.18226922e-01 4.37102318e-01 -8.68978560e-01
-1.15201545e+00 -9.46503729e-02 4.87389117e-01 -1.99710205e-01
-1.51681617e-01 -6.31732702e-01 -8.97523224e-01 -4.56195384e-01
-7.49644876e-01 1.33401886e-01 7.55851448e-01 8.13657045e-01
5.65462470e-01 -7.00339004e-02 2.89438546e-01 -7.46538758e-01
-3.88691872e-01 -1.12634158e+00 -3.17906469e-01 1.54452160e-01
-1.88708812e-01 -5.88325441e-01 -5.42254686e-01 -1.86035901e-01] | [9.635421752929688, 4.255773544311523] |
6855060b-5ae4-470f-9018-936d7a18cc6c | vatex-a-large-scale-high-quality-multilingual | 1904.03493 | null | https://arxiv.org/abs/1904.03493v3 | https://arxiv.org/pdf/1904.03493v3.pdf | VATEX: A Large-Scale, High-Quality Multilingual Dataset for Video-and-Language Research | We present a new large-scale multilingual video description dataset, VATEX, which contains over 41,250 videos and 825,000 captions in both English and Chinese. Among the captions, there are over 206,000 English-Chinese parallel translation pairs. Compared to the widely-used MSR-VTT dataset, VATEX is multilingual, larger, linguistically complex, and more diverse in terms of both video and natural language descriptions. We also introduce two tasks for video-and-language research based on VATEX: (1) Multilingual Video Captioning, aimed at describing a video in various languages with a compact unified captioning model, and (2) Video-guided Machine Translation, to translate a source language description into the target language using the video information as additional spatiotemporal context. Extensive experiments on the VATEX dataset show that, first, the unified multilingual model can not only produce both English and Chinese descriptions for a video more efficiently, but also offer improved performance over the monolingual models. Furthermore, we demonstrate that the spatiotemporal video context can be effectively utilized to align source and target languages and thus assist machine translation. In the end, we discuss the potentials of using VATEX for other video-and-language research. | ['Yuan-Fang Wang', 'Lei LI', 'Xin Wang', 'William Yang Wang', 'Junkun Chen', 'Jiawei Wu'] | 2019-04-06 | vatex-a-large-scale-high-quality-multilingual-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Wang_VaTeX_A_Large-Scale_High-Quality_Multilingual_Dataset_for_Video-and-Language_Research_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Wang_VaTeX_A_Large-Scale_High-Quality_Multilingual_Dataset_for_Video-and-Language_Research_ICCV_2019_paper.pdf | iccv-2019-10 | ['video-description'] | ['computer-vision'] | [-1.01159662e-01 -6.33348823e-01 -6.01067424e-01 -2.78129429e-01
-1.32492614e+00 -8.70341063e-01 6.07729793e-01 -4.38804388e-01
-3.09766650e-01 9.54084635e-01 5.16703367e-01 -1.81927949e-01
7.62315512e-01 -2.02430397e-01 -9.82777119e-01 -3.16232383e-01
1.98759213e-01 4.65251058e-01 -3.94708067e-02 -1.35333478e-01
-2.50028938e-01 1.06716245e-01 -1.26226091e+00 7.46207178e-01
6.48107767e-01 8.77900839e-01 7.30445027e-01 6.57102883e-01
-6.67979717e-02 1.07190096e+00 -2.67365336e-01 -5.01160741e-01
6.31503537e-02 -7.05520034e-01 -6.96223617e-01 1.75098568e-01
6.82693720e-01 -4.96167302e-01 -7.77711570e-01 9.46341813e-01
4.28076267e-01 -3.57628949e-02 3.43588561e-01 -1.40462303e+00
-1.25772643e+00 6.95264757e-01 -2.99298078e-01 8.41779411e-02
1.00517821e+00 1.33497536e-01 9.26853776e-01 -1.09810579e+00
1.35278332e+00 1.20819151e+00 2.99497545e-01 5.55250883e-01
-6.97852731e-01 -9.21223998e-01 6.97806850e-02 5.74397564e-01
-1.78463960e+00 -5.29012918e-01 2.62784660e-01 -3.89980137e-01
8.64377797e-01 1.79320186e-01 7.77318299e-01 1.61427295e+00
2.09591985e-01 8.71086955e-01 7.59572387e-01 -1.46831736e-01
-4.22527075e-01 1.62150875e-01 -7.05482185e-01 6.27354383e-01
-1.41674042e-01 -1.21509925e-01 -6.39344335e-01 2.16730207e-01
8.61345768e-01 -2.56592810e-01 -4.43427861e-01 -6.58272952e-02
-2.11163235e+00 5.81944287e-01 -7.11061992e-03 4.98527735e-01
-2.49524251e-01 1.93696946e-01 7.55401015e-01 3.36615145e-01
3.25919509e-01 3.37379277e-01 -2.78522730e-01 -5.65756798e-01
-8.21672499e-01 6.63381219e-02 5.09381592e-01 1.88739133e+00
7.43403256e-01 1.51030809e-01 -3.78589749e-01 7.11510062e-01
8.51002410e-02 1.18052185e+00 5.51950932e-01 -1.16488647e+00
9.85971212e-01 -1.91235617e-02 1.16320372e-01 -8.54149520e-01
6.49904832e-02 2.73996443e-01 -5.27976036e-01 -9.82871294e-01
-1.38312265e-01 -1.61295786e-01 -5.83861709e-01 1.79635870e+00
-1.37285352e-01 4.48980451e-01 4.12038147e-01 1.13558531e+00
1.12408662e+00 1.28673100e+00 1.17847636e-01 -4.69862312e-01
1.32077193e+00 -1.28242207e+00 -1.09780240e+00 -1.34032667e-01
6.68136358e-01 -1.16336381e+00 1.02678037e+00 -2.43253142e-01
-1.14669275e+00 -6.28961086e-01 -7.41737604e-01 -3.65750253e-01
-1.78249449e-01 3.70014906e-01 2.97647029e-01 5.27991354e-02
-1.30091250e+00 -2.43977562e-01 -5.87055743e-01 -5.59021950e-01
7.29627348e-03 -1.56542361e-01 -7.40406334e-01 -6.96621060e-01
-1.48458290e+00 8.20486307e-01 4.68317002e-01 -2.16187850e-01
-1.05982828e+00 -4.29430872e-01 -1.37226057e+00 -2.36744821e-01
2.64581174e-01 -5.47505438e-01 1.29956698e+00 -1.27539837e+00
-1.25262511e+00 8.77338767e-01 -5.97487688e-01 -2.33649939e-01
3.65604788e-01 5.89216650e-02 -7.45509744e-01 6.95933878e-01
6.23934686e-01 1.22144341e+00 4.81514037e-01 -9.63234961e-01
-8.81017447e-01 2.63227284e-01 1.39319986e-01 5.75657964e-01
-3.29856932e-01 5.82277596e-01 -1.38891757e+00 -9.40524340e-01
-4.40959990e-01 -1.20853710e+00 2.19824135e-01 2.08444744e-02
-1.98694855e-01 1.20393835e-01 8.83154154e-01 -1.15316033e+00
1.23996019e+00 -2.10241938e+00 5.16263008e-01 -4.38201755e-01
-6.51282370e-02 2.47364305e-02 -5.68227351e-01 6.19213521e-01
7.68668354e-02 2.02956006e-01 4.96425293e-02 -2.28017986e-01
-1.79809049e-01 3.58307481e-01 -3.58426452e-01 3.55506003e-01
3.89764979e-02 1.28346884e+00 -1.21161902e+00 -8.67374063e-01
6.00032359e-02 5.61240673e-01 -4.13431585e-01 3.11213434e-01
-1.36944801e-01 6.22017503e-01 -4.71249104e-01 8.69105220e-01
2.95209050e-01 -1.66553825e-01 3.58484983e-01 -2.98136413e-01
-2.30052233e-01 6.20756559e-02 -4.24894094e-01 2.13790798e+00
-7.48625398e-01 1.33042598e+00 -4.16456535e-02 -5.87404609e-01
5.69420218e-01 9.40788805e-01 6.84079409e-01 -8.22510481e-01
-4.94515412e-02 4.90721166e-01 -5.05616128e-01 -9.34688210e-01
6.75713539e-01 2.04132020e-01 -4.82997030e-01 4.40458089e-01
1.63900852e-01 -6.13602325e-02 6.50303721e-01 4.15006518e-01
6.78231001e-01 5.58031261e-01 1.67137444e-01 6.22153692e-02
5.26955128e-01 4.34282929e-01 6.26662791e-01 3.88429761e-01
-3.22705567e-01 6.97806954e-01 1.67379245e-01 -3.18022013e-01
-1.41058803e+00 -1.07166672e+00 1.58542454e-01 9.12537932e-01
4.33415651e-01 -6.40796423e-01 -6.95652127e-01 -3.61520112e-01
-3.84972453e-01 5.72284937e-01 -1.16980463e-01 1.97654832e-02
-8.44933748e-01 -4.75814044e-02 7.93318987e-01 5.82597494e-01
7.66500831e-01 -1.04070973e+00 1.19480211e-02 1.47680655e-01
-1.34394062e+00 -2.21174479e+00 -1.36075616e+00 -6.92650199e-01
-4.25249755e-01 -8.38842630e-01 -1.07616961e+00 -1.42086613e+00
5.22463799e-01 7.03849435e-01 1.19447374e+00 -1.99833542e-01
1.33784309e-01 7.67937481e-01 -8.21650147e-01 2.30488300e-01
-7.22941875e-01 -1.75736889e-01 3.71061981e-01 -1.03480518e-01
4.24302608e-01 -2.93247532e-02 4.23636548e-02 4.93992567e-01
-8.53871763e-01 6.50159955e-01 5.85680068e-01 6.68032527e-01
7.39063382e-01 -5.60513198e-01 3.97820830e-01 -2.28043780e-01
3.47762406e-01 -7.22481847e-01 -3.37240636e-01 6.60551071e-01
-5.72548807e-03 -3.49294782e-01 8.38462353e-01 -7.65069544e-01
-9.18480515e-01 -1.73103064e-03 5.27951196e-02 -9.80277061e-01
9.39829499e-02 7.76271462e-01 -1.04239203e-01 2.91183386e-02
8.74863267e-02 7.64111638e-01 -1.84398159e-01 -9.29634720e-02
3.76810998e-01 8.72639894e-01 8.45783651e-01 -6.88309252e-01
8.03602457e-01 2.89123595e-01 -4.71349359e-01 -9.83023167e-01
-4.24296230e-01 -5.00628352e-01 -7.51593947e-01 -4.75297391e-01
1.28411126e+00 -1.64473832e+00 -1.74263045e-01 3.65800321e-01
-1.49736619e+00 -2.83384621e-01 2.24486560e-01 8.12460184e-01
-9.50537562e-01 3.86763543e-01 -9.05265570e-01 -3.06728464e-02
-1.24523684e-01 -1.62778687e+00 1.28360617e+00 -1.61313772e-01
1.42081797e-01 -1.01087034e+00 -6.87171817e-02 5.87640047e-01
4.06012908e-02 -4.26667556e-02 5.57876587e-01 -3.18280339e-01
-6.94088399e-01 -9.98658463e-02 -4.99390006e-01 1.88112810e-01
1.79863557e-01 2.95523517e-02 -2.37477005e-01 -3.50970685e-01
-3.63275498e-01 -5.33679485e-01 2.86493182e-01 1.35591179e-01
7.52198458e-01 -3.36980641e-01 -1.51562572e-01 8.27548742e-01
1.29693651e+00 4.48901951e-01 7.05868602e-01 3.72914195e-01
1.04447103e+00 3.69579852e-01 1.03318489e+00 1.09465003e-01
8.65116239e-01 1.05590618e+00 -1.27298340e-01 -4.60327230e-03
-3.66711468e-01 -6.14271700e-01 9.91710544e-01 1.65216625e+00
-1.20796196e-01 -3.93146813e-01 -8.58769476e-01 6.48102880e-01
-1.98679733e+00 -1.07860041e+00 1.86262339e-01 1.68187952e+00
7.79094577e-01 -6.60360456e-01 -2.44541410e-02 -8.41170311e-01
1.07403421e+00 8.47648978e-02 -3.92298698e-01 -1.90916479e-01
-6.02986693e-01 -4.51486021e-01 4.60768104e-01 2.57118225e-01
-9.79729533e-01 1.38369632e+00 6.56234312e+00 8.71419191e-01
-1.28670144e+00 4.20509219e-01 3.62417489e-01 -1.21136181e-01
-3.71629894e-01 -1.00034393e-01 -6.18778586e-01 4.37343299e-01
1.21078515e+00 -5.41186690e-01 6.97071791e-01 6.28855705e-01
5.44621468e-01 4.22365487e-01 -1.08651757e+00 1.42112064e+00
5.98226786e-01 -1.73542452e+00 6.54403806e-01 -1.49166346e-01
8.78330588e-01 2.97439516e-01 -1.33256257e-01 5.31653643e-01
-2.36187726e-01 -7.18522191e-01 1.30922174e+00 5.30258715e-02
1.58976340e+00 -4.59843457e-01 6.41255021e-01 1.24504186e-01
-1.85947096e+00 1.54803529e-01 -4.14859504e-01 2.84320056e-01
6.03224933e-01 -2.73935854e-01 -3.64707232e-01 7.32492208e-01
7.67977834e-01 1.50933874e+00 -3.12464982e-01 6.57319427e-01
-2.79370338e-01 2.71131694e-01 -4.19231132e-03 -1.31506911e-02
5.99347830e-01 -4.02398348e-01 5.68420887e-01 1.45645547e+00
8.54921699e-01 -6.32202160e-03 4.78615165e-01 5.16890705e-01
-3.05605382e-01 4.02691215e-01 -9.13698673e-01 -5.40779114e-01
6.24886632e-01 8.74547601e-01 -3.64186704e-01 -6.61483228e-01
-9.11274374e-01 1.32721353e+00 4.07049134e-02 7.07630038e-01
-1.37914109e+00 3.77628487e-03 7.66756058e-01 -1.00758933e-01
2.59239107e-01 -4.06136990e-01 4.92674917e-01 -2.01215982e+00
1.74946323e-01 -1.13792491e+00 1.22139767e-01 -1.48965478e+00
-9.09561634e-01 1.02699506e+00 2.85642505e-01 -1.79807448e+00
-5.44750214e-01 -5.01868188e-01 -1.87656023e-02 5.93340635e-01
-1.48952317e+00 -1.61980748e+00 -1.85409933e-01 9.44262862e-01
1.17835176e+00 -4.31329846e-01 6.45475090e-01 8.70508194e-01
-5.91314852e-01 4.84296501e-01 1.83248296e-01 3.58519703e-01
1.13063860e+00 -5.35383284e-01 3.38007271e-01 1.01710916e+00
3.01843107e-01 3.07058513e-01 5.28251648e-01 -8.49955499e-01
-1.79946959e+00 -1.46584749e+00 1.20306396e+00 -3.65993768e-01
9.87367988e-01 -4.07815009e-01 -4.33660984e-01 1.18457282e+00
6.12881780e-01 -1.52185991e-01 5.00186205e-01 -7.59249747e-01
-4.59644377e-01 1.73214331e-01 -7.01664448e-01 8.84257555e-01
1.14967847e+00 -9.60823953e-01 -3.12202036e-01 6.01765156e-01
1.21230614e+00 -6.18178725e-01 -1.00611281e+00 4.16977406e-02
6.64116025e-01 -4.12414789e-01 9.50610876e-01 -4.85958934e-01
8.44067872e-01 -1.79313034e-01 -6.51669860e-01 -1.13135028e+00
-1.65979743e-01 -7.50210643e-01 9.96269733e-02 1.20039058e+00
4.39374387e-01 -1.95662320e-01 1.04493126e-01 1.22271411e-01
-4.04530048e-01 -4.11500841e-01 -9.39972758e-01 -9.88962650e-01
5.84540237e-03 -5.55201054e-01 4.22902644e-01 1.11725116e+00
-6.08626846e-03 3.36533070e-01 -1.00317788e+00 -1.11092599e-02
2.34365702e-01 1.71945006e-01 5.44305086e-01 -3.18883896e-01
1.26568496e-01 -1.78590596e-01 -3.94950062e-01 -1.46648741e+00
6.17608070e-01 -1.04709148e+00 1.83909357e-01 -1.47464299e+00
4.82798398e-01 -2.63599101e-02 1.86308354e-01 4.42495883e-01
9.00188908e-02 6.88291550e-01 3.85629296e-01 7.01517999e-01
-9.41799879e-01 5.20532489e-01 1.55177331e+00 -3.58485401e-01
1.03849746e-01 -6.70638442e-01 -4.38150525e-01 2.73551673e-01
2.96518713e-01 -2.24659115e-01 -2.75272012e-01 -9.86198187e-01
5.23141138e-02 5.11179924e-01 -2.11662874e-02 -6.08747900e-01
9.48889181e-02 -4.77688819e-01 -4.08750996e-02 -4.55439836e-01
4.68932569e-01 -8.07454050e-01 3.54253173e-01 1.95326000e-01
-3.12175900e-01 7.62983561e-01 3.35498631e-01 4.17255640e-01
-6.72553778e-01 1.71622723e-01 4.14213538e-01 -1.37707070e-01
-1.30530643e+00 6.26615703e-01 -6.86665177e-01 1.44706398e-01
1.37782073e+00 -1.46894202e-01 -4.92376328e-01 -8.41552854e-01
-4.78090912e-01 4.45868909e-01 7.50522435e-01 1.12654746e+00
8.09987962e-01 -1.96473312e+00 -1.05707359e+00 -7.34089911e-02
5.98925948e-01 -6.76586866e-01 2.12217480e-01 8.19382668e-01
-9.19084251e-01 9.57063794e-01 -4.64945972e-01 -7.38530874e-01
-1.17116141e+00 7.10884809e-01 1.98629685e-04 9.13644210e-02
-5.91490507e-01 2.55394578e-01 3.39106798e-01 -2.55857885e-01
-8.47473890e-02 -7.91189726e-03 -1.01862408e-01 -1.05570018e-01
7.19128132e-01 -2.60889735e-02 -4.91028666e-01 -1.66527414e+00
-3.05047870e-01 7.56505132e-01 1.63163394e-01 -2.73533821e-01
9.37881112e-01 -7.56038725e-01 -3.11621666e-01 4.23069537e-01
1.51323700e+00 -1.48808807e-01 -9.97803211e-01 -2.40829289e-01
-3.39884311e-01 -4.91400480e-01 -2.52030253e-01 -3.69940877e-01
-1.06606817e+00 5.89083433e-01 1.52408659e-01 -4.77290094e-01
1.18083024e+00 1.18213572e-01 1.36424720e+00 4.03348088e-01
7.53056645e-01 -9.24154460e-01 4.47464809e-02 6.05952322e-01
9.84456539e-01 -1.25676680e+00 -3.32846761e-01 -3.63083363e-01
-1.10380757e+00 1.19481885e+00 6.00256026e-01 3.90840292e-01
1.76061645e-01 7.69505203e-02 3.84174943e-01 3.45941275e-01
-1.01186275e+00 -5.04857078e-02 3.25184196e-01 6.75335288e-01
3.17582756e-01 6.47064298e-02 -1.67714134e-01 4.78430361e-01
6.57564700e-02 5.63123450e-02 6.87209129e-01 7.96399891e-01
5.83653338e-04 -1.11605954e+00 -3.61404836e-01 1.89577676e-02
-2.83276916e-01 -3.61427069e-01 -1.53085247e-01 8.12424362e-01
-2.38543358e-02 9.84383643e-01 1.47803783e-01 -5.35503626e-01
3.76100093e-02 -2.12622672e-01 3.51982474e-01 -6.67492151e-01
-2.70224065e-01 2.80821562e-01 1.60340637e-01 -7.45921552e-01
-7.71396637e-01 -7.58654773e-01 -9.43152905e-01 -4.65493768e-01
2.03527540e-01 2.72039086e-01 6.16513252e-01 1.03749049e+00
3.71005982e-01 2.18356684e-01 4.97400999e-01 -1.04730070e+00
1.40318766e-01 -7.43168235e-01 -2.45055467e-01 4.64109927e-01
1.00626208e-01 -3.84059280e-01 -1.66169658e-01 6.89635515e-01] | [10.811234474182129, 1.0890910625457764] |
dc6ef5a2-b6e2-4ea0-9ff8-9877f8638815 | model-based-trajectory-stitching-for-improved-1 | 2212.04280 | null | https://arxiv.org/abs/2212.04280v1 | https://arxiv.org/pdf/2212.04280v1.pdf | Model-based trajectory stitching for improved behavioural cloning and its applications | Behavioural cloning (BC) is a commonly used imitation learning method to infer a sequential decision-making policy from expert demonstrations. However, when the quality of the data is not optimal, the resulting behavioural policy also performs sub-optimally once deployed. Recently, there has been a surge in offline reinforcement learning methods that hold the promise to extract high-quality policies from sub-optimal historical data. A common approach is to perform regularisation during training, encouraging updates during policy evaluation and/or policy improvement to stay close to the underlying data. In this work, we investigate whether an offline approach to improving the quality of the existing data can lead to improved behavioural policies without any changes in the BC algorithm. The proposed data improvement approach - Trajectory Stitching (TS) - generates new trajectories (sequences of states and actions) by `stitching' pairs of states that were disconnected in the original data and generating their connecting new action. By construction, these new transitions are guaranteed to be highly plausible according to probabilistic models of the environment, and to improve a state-value function. We demonstrate that the iterative process of replacing old trajectories with new ones incrementally improves the underlying behavioural policy. Extensive experimental results show that significant performance gains can be achieved using TS over BC policies extracted from the original data. Furthermore, using the D4RL benchmarking suite, we demonstrate that state-of-the-art results are obtained by combining TS with two existing offline learning methodologies reliant on BC, model-based offline planning (MBOP) and policy constraint (TD3+BC). | ['Giovanni Montana', 'Charles A. Hepburn'] | 2022-12-08 | null | null | null | null | ['d4rl'] | ['robots'] | [ 3.59126091e-01 2.49516934e-01 -4.70696449e-01 -4.10995744e-02
-8.38356316e-01 -7.60511696e-01 9.53350425e-01 2.38895655e-01
-7.24310398e-01 1.31603110e+00 1.27740771e-01 -5.06965995e-01
-3.88469636e-01 -5.42212367e-01 -1.15290916e+00 -7.95313418e-01
-4.85682875e-01 7.02544630e-01 5.35056770e-01 -1.24643967e-01
3.27784687e-01 6.93815589e-01 -1.87403893e+00 8.89634565e-02
9.78481650e-01 6.61730707e-01 2.46168062e-01 8.82131457e-01
1.49716258e-01 7.01386929e-01 -6.40527368e-01 1.90133929e-01
5.30285001e-01 -5.11759102e-01 -7.53502846e-01 -2.26651225e-02
-1.70145094e-01 -6.35554135e-01 -2.68718570e-01 7.88537979e-01
2.65919507e-01 4.45238054e-01 1.68976560e-01 -1.39945710e+00
1.31378636e-01 6.25761688e-01 7.21816719e-03 -6.24785051e-02
5.40427744e-01 7.51208246e-01 5.03579855e-01 4.53453287e-02
9.12969053e-01 1.17821360e+00 4.77306306e-01 6.34679735e-01
-1.31339192e+00 -4.36449647e-01 4.08127010e-01 3.90236408e-01
-8.90051901e-01 -3.65454555e-01 5.64124942e-01 -1.97323829e-01
1.14120293e+00 2.05806270e-01 1.09585690e+00 1.56419873e+00
3.85132372e-01 8.68994772e-01 1.53329241e+00 -4.18314248e-01
7.38339126e-01 7.00790882e-02 -6.60172403e-01 4.68311638e-01
4.93069962e-02 1.06472898e+00 -3.68247598e-01 -2.57781357e-01
5.85503101e-01 -2.07512960e-01 5.07775396e-02 -6.20969474e-01
-1.26886427e+00 4.37497914e-01 1.29695728e-01 1.09185554e-01
-6.29937887e-01 4.17923421e-01 4.65090483e-01 5.18398404e-01
5.02501912e-02 7.18917668e-01 -5.69661081e-01 -8.68248940e-01
-6.88818634e-01 8.62741411e-01 8.05641413e-01 9.85400498e-01
6.52234733e-01 1.28778204e-01 -2.85346329e-01 8.98525864e-02
-1.62661832e-04 3.70113134e-01 4.39098537e-01 -1.33208954e+00
4.09986854e-01 3.99776012e-01 6.07857525e-01 -3.65483373e-01
-1.63727164e-01 -5.69560975e-02 -2.07774252e-01 6.77918792e-01
3.52342039e-01 -4.16838914e-01 -1.00991237e+00 1.79071081e+00
6.24277413e-01 5.29642820e-01 3.28382850e-01 5.24499118e-01
-3.02511543e-01 7.03310013e-01 -6.34816736e-02 -3.19018692e-01
2.91008443e-01 -7.33569443e-01 -4.59530592e-01 -1.22129619e-01
8.05592716e-01 -2.27835551e-01 8.96680415e-01 5.62368751e-01
-8.39368820e-01 -5.85294187e-01 -8.84411573e-01 7.71842659e-01
-1.94685176e-01 -3.59514743e-01 3.76918942e-01 3.83408308e-01
-9.33167756e-01 1.28031743e+00 -1.26555872e+00 -2.05228657e-01
3.35189193e-01 4.87925231e-01 -2.75691628e-01 2.47041360e-02
-9.25044954e-01 1.21638262e+00 8.70839536e-01 -1.51797161e-01
-1.80088961e+00 -7.49825299e-01 -5.63978672e-01 -3.83891970e-01
1.06858301e+00 -3.38406622e-01 1.66497934e+00 -9.10082757e-01
-2.21542311e+00 -1.19027898e-01 8.05906132e-02 -7.74998307e-01
8.00235868e-01 -2.86740035e-01 -3.27468306e-01 -1.28999293e-01
-9.25032124e-02 7.61128485e-01 9.36186969e-01 -1.42781341e+00
-1.29916668e+00 -7.67773315e-02 2.32368425e-01 2.30533689e-01
2.16084078e-01 -3.55797589e-01 -1.31669015e-01 -1.52809784e-01
-6.30973577e-01 -1.38077271e+00 -6.03173196e-01 -4.37012881e-01
-2.28751272e-01 -2.79987872e-01 8.15860033e-01 -7.11949110e-01
1.07996845e+00 -1.72088397e+00 6.21684909e-01 4.09548134e-01
-5.44257224e-01 5.01680970e-01 -1.01601839e-01 7.06813872e-01
3.01380605e-01 -2.98566706e-02 -3.95086348e-01 -2.17960507e-01
1.69972345e-01 8.57331872e-01 -4.34108049e-01 4.57506746e-01
4.57562208e-02 8.25645626e-01 -1.38253510e+00 -2.19551116e-01
5.02495587e-01 -5.84134385e-02 -5.68641186e-01 4.72524583e-01
-8.18633676e-01 1.07796991e+00 -4.52580214e-01 3.21158797e-01
1.03251919e-01 5.27886808e-01 4.52955693e-01 5.61422586e-01
-4.63381171e-01 3.46042395e-01 -1.10801911e+00 1.90271926e+00
-4.25807893e-01 3.07970107e-01 -1.65114775e-01 -9.77283359e-01
6.60630524e-01 3.03714216e-01 6.78366721e-01 -7.72668302e-01
-8.88388678e-02 1.82781696e-01 1.51078328e-01 -6.04786932e-01
3.60715270e-01 6.03670850e-02 -1.76610008e-01 3.52020353e-01
2.23164018e-02 -3.99844944e-01 2.31594637e-01 -1.93186149e-01
1.22664285e+00 1.11600518e+00 3.33081543e-01 7.21843913e-02
4.56246257e-01 4.37492043e-01 7.07973123e-01 9.27418292e-01
-2.16814786e-01 -1.58085421e-01 6.59465969e-01 -1.88109741e-01
-1.34013367e+00 -7.75301993e-01 4.76074696e-01 8.06276560e-01
1.24309650e-02 -1.97609723e-01 -8.14621091e-01 -1.06336486e+00
1.17413245e-01 1.24804807e+00 -6.03083253e-01 -4.15514886e-01
-8.34215879e-01 -4.43088152e-02 5.60861528e-01 4.61295366e-01
3.91545564e-01 -1.34149945e+00 -1.24444139e+00 7.06681967e-01
2.90541440e-01 -8.47534537e-01 -5.55272698e-02 2.23547995e-01
-9.54289377e-01 -9.83538270e-01 -3.55192214e-01 -2.56547958e-01
5.33176661e-01 -2.28179663e-01 5.86215615e-01 -8.16359222e-02
2.26139605e-01 4.72860605e-01 -5.59841275e-01 -3.64413649e-01
-1.12591398e+00 1.34330569e-02 2.74481326e-01 -3.51589590e-01
-3.43435228e-01 -6.02932513e-01 -1.88548297e-01 2.18313247e-01
-9.47046995e-01 3.85299809e-02 6.64643347e-01 7.70358920e-01
6.75107181e-01 3.18768770e-01 4.87152338e-01 -5.88042438e-01
6.58080578e-01 -3.54742259e-01 -9.34801877e-01 2.84564167e-01
-8.89642954e-01 6.77696526e-01 9.81875002e-01 -7.25849092e-01
-1.22453392e+00 1.53681457e-01 -9.68064293e-02 -6.64550185e-01
-5.90592802e-01 4.01966512e-01 1.21479951e-01 9.34384912e-02
5.47697127e-01 4.39880937e-01 2.67945558e-01 -4.20203000e-01
6.09201133e-01 4.28142488e-01 5.49937785e-01 -1.02174473e+00
6.12339914e-01 3.03795338e-01 1.41435847e-01 -3.23916942e-01
-3.61691475e-01 -4.07452017e-01 -6.82012320e-01 -5.11131883e-01
4.61662292e-01 -3.46953064e-01 -7.98593462e-01 4.70136672e-01
-8.38656664e-01 -1.03100836e+00 -6.66075170e-01 4.75751698e-01
-1.07736290e+00 2.61383086e-01 -3.15349996e-02 -1.12215137e+00
2.00166672e-01 -1.39653921e+00 8.13330054e-01 1.55804038e-01
-1.89277679e-01 -7.84748197e-01 6.11808062e-01 -2.33874306e-01
1.10767506e-01 6.48503482e-01 8.13765883e-01 -6.40812039e-01
-3.88775319e-01 -5.49171306e-02 5.50763786e-01 4.15300578e-01
1.61589593e-01 -5.25096208e-02 -6.59141898e-01 -6.89983368e-01
-3.99332017e-01 -2.81739712e-01 3.01919609e-01 9.99048725e-02
9.88928318e-01 -6.89199686e-01 -4.86231178e-01 6.14598207e-02
1.22245693e+00 7.61147141e-01 5.06257772e-01 6.13155007e-01
4.07002717e-01 6.69096172e-01 1.15923643e+00 5.59103787e-01
1.22013777e-01 8.25631618e-01 6.35653615e-01 4.12375301e-01
7.35688061e-02 -7.37312436e-01 8.81651878e-01 3.50199789e-01
-2.96316028e-01 4.57880162e-02 -7.99850643e-01 7.01966584e-01
-2.20807862e+00 -1.13355637e+00 3.39241862e-01 2.62328553e+00
8.95919979e-01 4.63778824e-01 6.35817349e-01 6.31956607e-02
3.08984220e-01 -2.20172644e-01 -9.69713748e-01 -6.84715867e-01
4.64777023e-01 3.02790344e-01 7.65412331e-01 6.39533758e-01
-7.29822993e-01 1.08073092e+00 5.74795818e+00 7.05158889e-01
-1.04111552e+00 -8.42094570e-02 -2.02753507e-02 -1.24491297e-01
-1.26004741e-01 3.16006422e-01 -6.77054405e-01 6.10685706e-01
1.40840268e+00 -1.72856987e-01 1.01456320e+00 8.44790280e-01
5.48397779e-01 -3.89003932e-01 -1.29683256e+00 2.95402050e-01
-5.22723138e-01 -1.35297978e+00 -2.96753615e-01 2.44430274e-01
9.32954431e-01 -5.73183149e-02 -1.03783667e-01 7.31213093e-01
8.64510834e-01 -7.13550627e-01 1.07411468e+00 6.40482008e-01
5.36488593e-01 -1.19137800e+00 6.08567417e-01 8.96061301e-01
-9.46622252e-01 -4.93168592e-01 -1.24171553e-02 -1.99872237e-02
1.73078060e-01 -1.70842201e-01 -1.33226049e+00 9.31239367e-01
5.79949558e-01 6.90230608e-01 -3.16057533e-01 1.09265423e+00
-5.29254675e-01 6.67948008e-01 -3.61469507e-01 -2.90997118e-01
6.92176163e-01 -2.35507816e-01 9.41876948e-01 7.62366951e-01
2.27797598e-01 -3.38075966e-01 3.60040158e-01 7.65822470e-01
5.56036770e-01 -5.52134931e-01 -8.95620227e-01 -1.02650993e-01
4.73929256e-01 6.83278561e-01 -3.57884943e-01 -4.62709635e-01
1.38872877e-01 8.34371626e-01 3.42495024e-01 2.70192295e-01
-1.00054550e+00 -5.00614084e-02 7.66781747e-01 -2.93995738e-01
4.75319594e-01 -4.13667619e-01 2.61994183e-01 -5.25399804e-01
-4.12838124e-02 -1.25044215e+00 1.02371782e-01 -2.75860012e-01
-6.82339072e-01 2.54717767e-01 5.77447057e-01 -1.37610853e+00
-9.44207370e-01 -1.47504419e-01 -3.94951820e-01 4.75583851e-01
-1.33818638e+00 -6.97225332e-01 2.64890164e-01 2.47498482e-01
7.40155935e-01 -4.37538512e-03 6.92795157e-01 -2.88267821e-01
-4.20445591e-01 1.40485346e-01 4.65096742e-01 -6.31074786e-01
2.34837353e-01 -1.44193494e+00 4.07649279e-01 8.63368928e-01
-3.76641229e-02 2.38157943e-01 1.03848422e+00 -1.01980281e+00
-1.51990771e+00 -1.13258648e+00 1.18516207e-01 -3.27163011e-01
6.42836869e-01 -1.50294647e-01 -8.58991981e-01 7.53611267e-01
1.74839020e-01 -4.43625957e-01 -8.18769112e-02 -3.38938653e-01
2.86658376e-01 -1.16630299e-02 -1.15957201e+00 7.75854588e-01
1.13622749e+00 -1.65064901e-01 -7.53720403e-01 1.48971587e-01
7.27783859e-01 -5.85686326e-01 -9.17619944e-01 5.21175444e-01
5.36727250e-01 -8.75048101e-01 7.95526147e-01 -1.07328868e+00
-1.41593805e-02 -2.89159417e-01 1.88382864e-01 -1.85303020e+00
1.11523844e-01 -1.19684362e+00 -5.22512138e-01 8.99644673e-01
2.44942099e-01 -5.84737241e-01 7.55785286e-01 3.97508174e-01
-3.44183922e-01 -9.70272005e-01 -1.16864610e+00 -1.36467612e+00
8.38617384e-02 -5.45079648e-01 9.79978740e-01 4.32870299e-01
3.35770212e-02 -2.50243366e-01 -4.51801360e-01 1.52212366e-01
5.27900100e-01 1.66926917e-03 1.19866657e+00 -7.09441006e-01
-5.05158901e-01 -4.27379280e-01 -4.61560711e-02 -9.96459067e-01
4.33678240e-01 -5.11942446e-01 5.69357753e-01 -1.50172174e+00
-3.68334025e-01 -7.14613676e-01 -2.73345262e-01 4.99580085e-01
1.14949286e-01 -8.34254384e-01 2.32395515e-01 7.38499761e-02
-6.11919165e-01 7.77661741e-01 1.34951031e+00 1.65161565e-01
-7.29646504e-01 3.24089795e-01 1.28480103e-02 4.88269567e-01
1.00350833e+00 -7.02727973e-01 -6.17549598e-01 -2.11927164e-02
-1.71203483e-02 2.77269691e-01 2.99426645e-01 -1.16396022e+00
1.11965999e-01 -6.21797383e-01 -1.19546846e-01 -5.25219500e-01
2.69984543e-01 -9.76342976e-01 4.56251085e-01 9.87933397e-01
-5.07403493e-01 1.16664767e-01 4.81731564e-01 1.13329852e+00
2.02181712e-01 -2.78755367e-01 4.10739034e-01 -1.60680294e-01
-1.07223797e+00 2.12129317e-02 -7.98697531e-01 -2.07856387e-01
1.48738682e+00 -3.18441212e-01 -1.04065547e-02 -2.12458074e-01
-5.48107028e-01 4.16442692e-01 5.02991199e-01 4.62595403e-01
4.58380938e-01 -1.12431049e+00 -3.54660183e-01 -5.03930748e-02
-7.55977333e-02 1.52392253e-01 -1.70924410e-01 7.89741397e-01
-1.89007476e-01 5.31954587e-01 -4.51208383e-01 -4.52103496e-01
-9.34834182e-01 6.56397402e-01 2.84123421e-01 -5.10906219e-01
-8.78026068e-01 2.90570021e-01 -7.47081697e-01 -6.90687716e-01
3.86476755e-01 -8.34577441e-01 -2.26458739e-02 -3.13335717e-01
1.57779992e-01 7.06520081e-01 -4.22321334e-02 -2.32480973e-01
-1.55065835e-01 6.33853450e-02 2.10769102e-01 -6.63905501e-01
1.39858580e+00 1.62268475e-01 4.69405979e-01 5.40113389e-01
6.51538491e-01 -3.24929297e-01 -2.09610033e+00 -6.76610693e-02
4.40007269e-01 -5.33466518e-01 -8.51355046e-02 -9.91149783e-01
-4.14455563e-01 3.39857668e-01 4.59692836e-01 2.36770049e-01
8.01835716e-01 -3.01212460e-01 6.87022507e-01 5.60701787e-01
9.91691232e-01 -1.50037110e+00 2.53777988e-02 3.59712929e-01
8.45629930e-01 -9.16246235e-01 -8.24458059e-03 3.08165014e-01
-8.49197984e-01 7.86008418e-01 6.37762547e-01 -1.08858772e-01
2.00809650e-02 8.39451328e-02 -4.27735150e-01 2.96434909e-01
-9.16950643e-01 -3.34587008e-01 1.07627891e-01 9.65425730e-01
-6.37542903e-01 1.68717101e-01 1.11310780e-02 -5.69168106e-02
3.99224982e-02 1.61044106e-01 4.93150651e-01 1.51438427e+00
-5.06790638e-01 -1.51593041e+00 -4.24113303e-01 3.12022626e-01
1.58957079e-01 5.69952190e-01 -3.77664238e-01 1.13376975e+00
1.56991422e-01 8.45848978e-01 -1.12194255e-01 -4.49557334e-01
4.86968815e-01 7.27583542e-02 8.28749657e-01 -5.06371498e-01
-5.64345658e-01 -4.42932583e-02 5.06064236e-01 -1.06102800e+00
-3.91074449e-01 -9.95405734e-01 -1.55579960e+00 -2.69066691e-01
-1.11439005e-01 2.27485999e-01 7.25441337e-01 1.20171845e+00
5.55328429e-01 6.84606016e-01 6.39827371e-01 -1.05436087e+00
-1.03244233e+00 -7.37955511e-01 -6.76203892e-02 1.61031559e-01
3.50445330e-01 -1.06787086e+00 -7.74246380e-02 -9.96715873e-02] | [4.1694536209106445, 2.019341230392456] |
7832eb73-43de-4859-8415-80c8ca8fabf1 | stationary-diffusion-state-neural-estimation | 2112.01334 | null | https://arxiv.org/abs/2112.01334v1 | https://arxiv.org/pdf/2112.01334v1.pdf | Stationary Diffusion State Neural Estimation for Multiview Clustering | Although many graph-based clustering methods attempt to model the stationary diffusion state in their objectives, their performance limits to using a predefined graph. We argue that the estimation of the stationary diffusion state can be achieved by gradient descent over neural networks. We specifically design the Stationary Diffusion State Neural Estimation (SDSNE) to exploit multiview structural graph information for co-supervised learning. We explore how to design a graph neural network specially for unsupervised multiview learning and integrate multiple graphs into a unified consensus graph by a shared self-attentional module. The view-shared self-attentional module utilizes the graph structure to learn a view-consistent global graph. Meanwhile, instead of using auto-encoder in most unsupervised learning graph neural networks, SDSNE uses a co-supervised strategy with structure information to supervise the model learning. The co-supervised strategy as the loss function guides SDSNE in achieving the stationary state. With the help of the loss and the self-attentional module, we learn to obtain a graph in which nodes in each connected component fully connect by the same weight. Experiments on several multiview datasets demonstrate effectiveness of SDSNE in terms of six clustering evaluation metrics. | ['Kun Zhan', 'Yixuan Ma', 'Zhuolin Liao', 'Chenghua Liu'] | 2021-12-02 | null | null | null | null | ['multiview-learning'] | ['computer-vision'] | [-1.84893444e-01 4.36394304e-01 -3.97705615e-01 -3.90545070e-01
-2.67076701e-01 -2.98209608e-01 4.68825519e-01 1.46679938e-01
1.47045171e-02 2.03221917e-01 7.51524121e-02 7.98464119e-02
-3.08483660e-01 -7.55236804e-01 -7.14661300e-01 -8.38709593e-01
-1.22708879e-01 4.71330851e-01 1.10852636e-01 1.26903623e-01
-2.55251545e-02 1.84687257e-01 -7.90753007e-01 -8.20143297e-02
1.00086951e+00 6.68985307e-01 4.71737534e-01 6.31865144e-01
-1.58888236e-01 1.22559667e+00 -1.51992410e-01 -4.37020391e-01
2.43668213e-01 -6.21204376e-01 -5.68277836e-01 6.15971446e-01
3.56683344e-01 6.50148913e-02 -5.59502363e-01 1.49933398e+00
4.87803966e-01 1.97172165e-02 7.27204263e-01 -1.27993858e+00
-8.45627010e-01 7.54261553e-01 -7.60671437e-01 1.52149960e-01
-8.06587189e-02 7.75769725e-02 1.23774135e+00 -6.42922640e-01
1.01614928e+00 1.09410870e+00 4.16872591e-01 4.48408097e-01
-1.42391479e+00 -3.28624129e-01 7.10535169e-01 2.49785781e-01
-1.41355526e+00 -3.45204592e-01 1.52790344e+00 -5.79691052e-01
6.51705384e-01 -3.35177958e-01 7.41368234e-01 8.01832557e-01
3.38134378e-01 6.72009230e-01 7.28877604e-01 -1.50533110e-01
2.67213851e-01 6.52716458e-02 2.50405103e-01 1.40400720e+00
1.41487390e-01 -3.67376208e-01 -4.90720242e-01 8.96515027e-02
6.04049921e-01 8.42244402e-02 -2.41572812e-01 -1.13302732e+00
-9.31661665e-01 1.00425005e+00 8.77622426e-01 2.59048134e-01
-5.94475687e-01 1.61204398e-01 4.88204837e-01 4.91906494e-01
6.98223650e-01 -1.99468508e-02 -1.90023601e-01 6.74256146e-01
-6.77798986e-01 -4.97648984e-01 8.08017612e-01 9.81593609e-01
1.12276506e+00 4.21578616e-01 1.33642301e-01 7.63796210e-01
5.28818607e-01 2.91828573e-01 4.18043703e-01 -1.11456311e+00
5.99210560e-01 1.00554693e+00 -6.48500621e-01 -1.26006079e+00
-2.09562033e-01 -7.95909286e-01 -1.44111621e+00 8.17829818e-02
-2.78273430e-02 -3.55345786e-01 -8.46751750e-01 1.91352880e+00
3.04971755e-01 1.13469362e-01 1.00559250e-01 7.64448345e-01
5.23702443e-01 5.26105404e-01 -1.64837465e-01 -5.42283118e-01
6.07217431e-01 -1.20954347e+00 -6.79517925e-01 -1.19095795e-01
6.92392826e-01 3.00027709e-02 6.82941556e-01 3.57822984e-01
-1.10466039e+00 -5.32264113e-01 -1.17619824e+00 3.36698800e-01
-1.70630977e-01 3.47960554e-02 3.23342085e-01 6.60224184e-02
-1.31868899e+00 6.27039671e-01 -9.92529154e-01 -3.27246428e-01
3.61841589e-01 3.31542641e-01 -5.20812213e-01 4.60335873e-02
-9.62148726e-01 4.66810763e-01 5.35084188e-01 7.50944167e-02
-1.17205787e+00 -3.36257964e-01 -1.06302011e+00 4.68047820e-02
4.84548360e-01 -1.04304099e+00 4.23148721e-01 -1.54822266e+00
-1.39105594e+00 5.86804926e-01 -8.50124285e-02 -4.25568283e-01
3.01622808e-01 3.79965097e-01 -2.40404561e-01 5.61376631e-01
2.58503675e-01 5.95185757e-01 1.32353246e+00 -1.64318860e+00
-3.46414953e-01 -4.84880984e-01 -6.78064600e-02 5.35905361e-01
-5.11245131e-01 -5.75933397e-01 -7.41809785e-01 -4.38912958e-01
4.39928293e-01 -9.08005953e-01 -4.28570569e-01 -1.90167159e-01
-5.20510733e-01 -3.25188875e-01 1.08884287e+00 -6.32426739e-01
1.29727101e+00 -1.85167480e+00 8.91113698e-01 3.69137645e-01
8.44203532e-01 -1.96333468e-01 -1.28800869e-01 5.08835435e-01
-1.77542359e-01 -8.45743492e-02 -2.34210759e-01 -5.29936671e-01
-3.28942180e-01 2.47501954e-01 2.00577646e-01 7.20364988e-01
4.17566933e-02 8.41619015e-01 -9.54215586e-01 -6.25771463e-01
2.34667554e-01 4.31420386e-01 -5.61825871e-01 3.30653220e-01
-1.29924566e-01 2.88603276e-01 -5.85476339e-01 2.26350680e-01
3.57327759e-01 -9.78395104e-01 6.90732360e-01 -4.48232114e-01
2.10441217e-01 -2.70171821e-01 -1.16937542e+00 1.99351263e+00
-3.07859153e-01 3.91430348e-01 5.19352198e-01 -1.50441349e+00
9.35680509e-01 1.09245725e-01 5.84008515e-01 -3.47493351e-01
-1.10662170e-01 -1.90232962e-01 -4.04894277e-02 -3.46245468e-01
-5.58823906e-02 -5.10633597e-03 3.58304262e-01 2.87984014e-01
5.68549752e-01 2.81888247e-01 1.08187497e-01 8.93903255e-01
1.00166476e+00 -1.69445127e-01 1.14671245e-01 -3.69487464e-01
7.24905431e-01 -2.90549964e-01 5.51689208e-01 5.30079544e-01
-2.40825996e-01 3.63779366e-01 6.26205564e-01 -1.90759793e-01
-9.21405911e-01 -1.18352532e+00 6.59772396e-01 7.49001741e-01
2.37522423e-01 -6.43101871e-01 -6.67339981e-01 -1.18907034e+00
-3.47256541e-01 3.15323591e-01 -5.90612531e-01 -3.63888383e-01
-2.97252715e-01 -3.74346763e-01 -2.42080942e-01 2.52426326e-01
5.14698267e-01 -7.49410808e-01 1.05070673e-01 2.19201192e-01
-7.90421814e-02 -8.21066320e-01 -7.90612340e-01 3.76398742e-01
-8.65748584e-01 -1.01886618e+00 -6.98555648e-01 -9.96315420e-01
1.10574341e+00 4.37408894e-01 8.60479832e-01 4.08365577e-02
1.55827790e-01 9.02083218e-01 -1.51892871e-01 1.58017650e-01
-4.68587428e-01 3.49326015e-01 -4.31018434e-02 4.37345296e-01
-5.15924469e-02 -9.46516633e-01 -6.20202541e-01 1.97551046e-02
-7.80784488e-01 1.09641358e-01 3.72280568e-01 7.72811532e-01
6.25688612e-01 2.21979350e-01 7.10457623e-01 -9.74739194e-01
7.16764390e-01 -5.59958220e-01 -5.94924927e-01 3.86765242e-01
-1.10739470e+00 3.01670432e-01 9.20846581e-01 -3.28813136e-01
-7.86007345e-01 2.38599613e-01 3.30325335e-01 -1.22586739e+00
2.95216620e-01 8.25849533e-01 -2.15523064e-01 9.53620374e-02
3.85375679e-01 4.03970927e-01 3.96611392e-01 -4.37147208e-02
6.67316079e-01 1.16289198e-01 1.92854226e-01 -2.30336577e-01
7.67137706e-01 5.64826071e-01 1.09443054e-01 -8.48398924e-01
-8.38616490e-01 -5.07524431e-01 -8.04236710e-01 -6.25065982e-01
1.09766221e+00 -1.13815176e+00 -7.09955156e-01 3.26580077e-01
-8.21796656e-01 -2.39547938e-01 -2.07395256e-01 3.37464750e-01
-5.34726620e-01 7.86786377e-01 -6.26545012e-01 -6.21563077e-01
-4.10045505e-01 -1.03680921e+00 7.85423517e-01 -7.59632662e-02
2.03765780e-01 -1.60788870e+00 1.00417264e-01 1.76972866e-01
-1.20983044e-04 1.42706394e-01 8.57268274e-01 -5.84197760e-01
-5.82700312e-01 -3.62618417e-02 -1.00885838e-01 6.53930008e-01
3.42044652e-01 -6.40065670e-02 -5.79911411e-01 -7.09007442e-01
5.85406497e-02 -4.11781490e-01 1.03166974e+00 6.42808855e-01
9.81107891e-01 -4.68120575e-01 -3.39229614e-01 6.49977505e-01
1.79361904e+00 3.96128409e-02 9.75814834e-02 -6.22553416e-02
1.40815234e+00 8.86222064e-01 -6.57011569e-02 1.47191733e-01
8.15241277e-01 1.29125223e-01 6.66454554e-01 6.75954223e-02
-1.17222264e-01 -4.88256156e-01 4.47068810e-01 1.72293139e+00
3.85602377e-02 -3.11540693e-01 -8.23039591e-01 4.77222145e-01
-2.05298066e+00 -9.82204676e-01 9.80491638e-02 1.98636508e+00
3.33284885e-01 2.97428131e-01 2.10087355e-02 -3.25581044e-01
1.07379711e+00 7.82992005e-01 -7.27347970e-01 -1.33280978e-02
-1.28393784e-01 -6.06877506e-01 4.12689865e-01 8.65426481e-01
-9.71996069e-01 9.25286233e-01 5.12141514e+00 7.15093076e-01
-1.20193648e+00 1.02760315e-01 6.94185793e-01 2.95412719e-01
-3.73382837e-01 1.72638386e-01 -4.97050107e-01 2.84390986e-01
6.94466949e-01 -1.22789904e-01 5.61942697e-01 8.82608593e-01
2.81689465e-01 3.78839612e-01 -8.76082003e-01 1.01229739e+00
2.64905632e-01 -1.34952247e+00 4.49467093e-01 1.55349419e-01
9.38206434e-01 1.79670066e-01 -1.11402109e-01 9.87739712e-02
7.73493528e-01 -5.44061661e-01 4.84431654e-01 8.04170430e-01
4.24390614e-01 -8.35977495e-01 2.69991487e-01 5.88518202e-01
-1.49908817e+00 -4.30020988e-02 -3.93729895e-01 3.94599319e-01
1.47570252e-01 5.93191206e-01 -5.77132344e-01 9.96052444e-01
3.89847934e-01 1.55623412e+00 -5.14658034e-01 3.29935282e-01
-2.15507835e-01 5.20538747e-01 1.38458153e-02 2.43559346e-01
3.64048600e-01 -9.65522885e-01 7.91296661e-01 7.17489064e-01
-6.29271418e-02 -2.13296592e-01 5.67855895e-01 7.87165105e-01
-2.77572751e-01 5.39318696e-02 -8.74145925e-01 -2.50707924e-01
8.29580501e-02 1.41762972e+00 -7.53119111e-01 -4.52236414e-01
-5.88115454e-01 1.32131457e+00 1.10582232e+00 5.76101303e-01
-4.60827380e-01 -1.37280762e-01 1.20658584e-01 -2.30985060e-02
5.80318630e-01 -4.34843272e-01 1.47053152e-01 -1.44357991e+00
9.56617109e-03 -7.50594616e-01 4.67270225e-01 -6.44623399e-01
-1.58497524e+00 5.99299133e-01 -1.41316548e-01 -1.18542528e+00
-2.48415604e-01 -1.50749058e-01 -8.93039227e-01 5.59043407e-01
-1.30503154e+00 -1.28381526e+00 -1.97076365e-01 8.75134826e-01
5.75538516e-01 -3.34640622e-01 3.07111412e-01 3.45637351e-02
-8.81449878e-01 1.86709866e-01 8.23925585e-02 1.07444696e-01
6.03849530e-01 -1.54717016e+00 -1.00833558e-01 7.71817982e-01
1.64901003e-01 4.66240108e-01 4.65893924e-01 -9.61705983e-01
-1.49589849e+00 -1.33184397e+00 4.04439986e-01 -1.05523346e-02
9.46907580e-01 -2.09369540e-01 -9.53732073e-01 8.52628648e-01
4.83508199e-01 1.68288410e-01 2.69709915e-01 9.54512432e-02
-3.54204059e-01 -3.74581158e-01 -6.40183270e-01 5.91762125e-01
1.01595914e+00 -7.46715426e-01 -1.89022258e-01 3.14861685e-01
9.08532619e-01 1.65888757e-01 -1.06236112e+00 2.04219133e-01
1.38664752e-01 -8.44343245e-01 7.04684556e-01 -6.54654860e-01
4.54608530e-01 -2.67034262e-01 4.09255177e-02 -1.73264730e+00
-5.14649332e-01 -4.91535693e-01 -3.16009909e-01 1.29515290e+00
4.21232343e-01 -6.18065476e-01 8.60661268e-01 1.04911424e-01
-1.08011380e-01 -8.01671922e-01 -6.24479353e-01 -6.74017847e-01
-1.01940058e-01 2.85290275e-02 -1.82190388e-01 1.27026105e+00
-2.26654038e-01 9.51070070e-01 -4.74665999e-01 4.24775898e-01
1.12824774e+00 6.59167320e-02 6.24415696e-01 -1.19410634e+00
-3.24161917e-01 -3.02266300e-01 -3.68019432e-01 -1.04628313e+00
5.21135211e-01 -1.29272962e+00 -3.41883153e-01 -1.91936910e+00
2.50288278e-01 4.23536785e-02 -2.95810491e-01 2.27530953e-02
-7.04799742e-02 -4.23261255e-01 2.07813904e-01 2.82224685e-01
-1.06282067e+00 8.47398937e-01 1.41479254e+00 -5.74367881e-01
-3.37461233e-01 -1.81212723e-01 -5.30670881e-01 7.81796038e-01
6.70277059e-01 -4.36714947e-01 -9.20099854e-01 -2.72421420e-01
1.81960225e-01 3.07386518e-01 1.11797094e-01 -7.17721105e-01
7.72585809e-01 1.29211172e-01 1.80704445e-01 -5.11189401e-01
1.87309086e-01 -1.02748775e+00 1.78626820e-01 5.68751574e-01
-4.22037810e-01 1.34088263e-01 -5.30764401e-01 1.22687554e+00
-2.62516141e-01 -6.72398955e-02 8.35792959e-01 -2.84691840e-01
-5.61273456e-01 7.91745007e-01 -1.93396822e-01 5.83491959e-02
1.00389457e+00 -1.89814031e-01 -1.44902915e-01 -7.01291561e-01
-1.24529302e+00 5.79370737e-01 3.82998914e-01 2.57610351e-01
6.78146243e-01 -1.36717486e+00 -4.21891958e-01 -1.70576815e-02
2.28834469e-02 7.49277622e-02 2.90599614e-01 8.01129520e-01
-2.28858516e-01 -6.28086999e-02 6.26260266e-02 -8.67508709e-01
-1.09433007e+00 9.63158429e-01 5.60423374e-01 -5.12414455e-01
-6.80476129e-01 5.08907616e-01 4.49802816e-01 -4.59787369e-01
2.82597631e-01 1.31004125e-01 -2.98337162e-01 2.11936891e-01
-6.35888474e-03 3.97031516e-01 -3.06040674e-01 -6.89726949e-01
-9.11890417e-02 5.68667293e-01 -1.55926511e-01 8.81319195e-02
1.35717165e+00 -6.51898861e-01 -3.83976877e-01 6.98579311e-01
1.57966506e+00 -1.68122247e-01 -1.44457555e+00 -4.02683914e-01
-1.22371539e-01 2.32264206e-01 3.14591587e-01 -1.25141829e-01
-1.63207293e+00 7.53842533e-01 4.33130980e-01 5.02654850e-01
1.06937361e+00 1.47865906e-01 2.66700566e-01 5.57388723e-01
1.36309609e-01 -1.05846334e+00 4.99821484e-01 2.94900358e-01
8.07102442e-01 -1.32659805e+00 -5.23487031e-02 -4.54976797e-01
-8.59201670e-01 1.05328989e+00 7.63326705e-01 -4.57996964e-01
1.18793702e+00 -1.14584856e-01 -1.65553048e-01 -8.37687373e-01
-9.89486456e-01 -1.08071029e-01 3.59267414e-01 5.28296709e-01
1.94879360e-02 -1.27245992e-01 5.42524382e-02 1.58963993e-01
3.53659004e-01 -5.03083825e-01 3.66556495e-01 5.61533093e-01
-4.46808070e-01 -7.84101427e-01 3.37487727e-01 6.37129188e-01
-9.02227089e-02 5.06211817e-02 -5.05993903e-01 5.88864386e-01
-3.52024794e-01 8.14210713e-01 1.39575582e-02 -4.74069357e-01
9.07720849e-02 -8.92533734e-02 3.29706728e-01 -6.20640576e-01
-4.36914712e-01 2.45945930e-01 -1.04741141e-01 -5.16736507e-01
-7.31250525e-01 -4.08060700e-01 -9.37589467e-01 -1.40447274e-01
-4.24300224e-01 2.65467137e-01 3.61978918e-01 7.91977644e-01
4.14664626e-01 6.57099545e-01 1.04145181e+00 -5.40004909e-01
-3.64108056e-01 -7.15179384e-01 -8.22791874e-01 4.63856637e-01
3.97084057e-01 -3.23929429e-01 -6.77845240e-01 1.12160452e-01] | [7.51605749130249, 5.942156791687012] |
d9b92c43-f5b8-45c8-8e81-3bc8239bafee | a-real-time-wrong-way-vehicle-detection-based | 2210.10226 | null | https://arxiv.org/abs/2210.10226v1 | https://arxiv.org/pdf/2210.10226v1.pdf | A Real-Time Wrong-Way Vehicle Detection Based on YOLO and Centroid Tracking | Wrong-way driving is one of the main causes of road accidents and traffic jam all over the world. By detecting wrong-way vehicles, the number of accidents can be minimized and traffic jam can be reduced. With the increasing popularity of real-time traffic management systems and due to the availability of cheaper cameras, the surveillance video has become a big source of data. In this paper, we propose an automatic wrong-way vehicle detection system from on-road surveillance camera footage. Our system works in three stages: the detection of vehicles from the video frame by using the You Only Look Once (YOLO) algorithm, track each vehicle in a specified region of interest using centroid tracking algorithm and detect the wrong-way driving vehicles. YOLO is very accurate in object detection and the centroid tracking algorithm can track any moving object efficiently. Experiment with some traffic videos shows that our proposed system can detect and identify any wrong-way vehicle in different light and weather conditions. The system is very simple and easy to implement. | ['Muhammad Ahsan Ullah', 'Amit Mazumder Ami', 'Zillur Rahman'] | 2022-10-19 | null | null | null | null | ['video-object-tracking'] | ['computer-vision'] | [-1.73531160e-01 -5.79325140e-01 -3.15245569e-01 -1.52529553e-01
-1.37468129e-01 -5.03315628e-01 4.09285754e-01 -2.62590677e-01
-6.61880255e-01 4.57681686e-01 -2.66208291e-01 -4.68442738e-01
-4.68696654e-02 -8.17319214e-01 -5.19986749e-01 -7.32614458e-01
3.33840549e-01 1.05871201e-01 1.25572729e+00 -1.08673900e-01
5.07843733e-01 6.66696668e-01 -1.72063756e+00 -7.17917904e-02
4.89812732e-01 7.08530784e-01 3.25105041e-01 1.06136215e+00
-7.45068491e-02 8.61898601e-01 -4.11214262e-01 3.79614346e-02
3.69801909e-01 -2.03492656e-01 -1.30257010e-01 1.19380340e-01
3.95313948e-01 -5.71420610e-01 -5.82399249e-01 9.44741070e-01
1.37361854e-01 2.61294246e-01 4.26011622e-01 -1.66135859e+00
7.28260577e-02 -3.29498053e-01 -8.91733289e-01 6.90614462e-01
1.11259252e-01 1.61638305e-01 -1.11959159e-01 -6.30121291e-01
3.32098782e-01 1.30140746e+00 4.93296742e-01 4.54036325e-01
-5.57617366e-01 -1.01483822e+00 -3.09450746e-01 6.99446142e-01
-1.52384794e+00 -6.28130734e-01 4.20020014e-01 -4.54705179e-01
7.00590372e-01 1.87851369e-01 6.33306146e-01 1.61285594e-01
6.57193184e-01 4.33388889e-01 7.49862671e-01 -3.06809396e-01
-1.20397009e-01 1.54881075e-01 4.41768944e-01 7.30380714e-01
7.42503762e-01 2.86403835e-01 3.10886372e-03 9.65274274e-02
2.70406395e-01 5.51475763e-01 -6.31944800e-04 -3.78172934e-01
-1.02710605e+00 6.36245549e-01 -3.24537270e-02 1.52961597e-01
-2.33410776e-01 3.50722879e-01 4.87206340e-01 -1.94124095e-02
-2.48709805e-02 -3.84216756e-01 -1.62925050e-01 -4.13967401e-01
-8.87141526e-01 1.65111989e-01 2.30640799e-01 9.19767022e-01
5.13348222e-01 -1.35782165e-02 1.59014329e-01 5.71066022e-01
2.10429505e-01 1.11617970e+00 1.41034588e-01 -1.06474698e+00
3.95258278e-01 6.44235432e-01 1.13274567e-01 -1.09652078e+00
-3.79107088e-01 7.16021582e-02 -4.22031522e-01 6.43185794e-01
4.64807481e-01 -2.29707286e-01 -8.98493767e-01 1.02292132e+00
5.31009495e-01 3.38486850e-01 -1.06556237e-01 7.15570807e-01
7.15229630e-01 1.09636402e+00 9.27354470e-02 -2.86040783e-01
1.52459049e+00 -7.97901988e-01 -8.83777678e-01 -1.67168200e-01
5.95911801e-01 -8.65910709e-01 2.98478216e-01 3.13229352e-01
-5.74465811e-01 -6.87471032e-01 -7.49022841e-01 1.69190273e-01
-6.65905058e-01 1.52810499e-01 1.41382307e-01 7.85275400e-01
-5.40699184e-01 -2.78376669e-01 -5.36814690e-01 -5.92477143e-01
2.69769132e-01 2.30537474e-01 -5.31548500e-01 -4.96719539e-01
-8.60436380e-01 1.08413124e+00 3.08700949e-01 9.41070095e-02
-6.13443851e-01 -2.13056147e-01 -6.97812557e-01 -2.15452343e-01
6.69258416e-01 -3.45567882e-01 9.15333748e-01 -7.26377606e-01
-7.42536128e-01 7.71737933e-01 -6.79034531e-01 -3.17650229e-01
4.85154241e-01 -7.53483772e-02 -8.83932173e-01 1.86244845e-01
4.64245826e-01 4.59156424e-01 8.06165218e-01 -8.70946527e-01
-1.47864664e+00 -2.51229972e-01 -2.95917362e-01 -2.38417219e-02
4.72034067e-01 3.55473161e-01 -3.55995625e-01 1.41525552e-01
-2.59118646e-01 -1.12518144e+00 4.95744906e-02 -1.80975378e-01
1.58414375e-02 -4.60836142e-01 1.72837698e+00 -2.64349133e-01
1.33491302e+00 -2.37100792e+00 -8.30730796e-01 1.48066536e-01
1.96781307e-01 8.13433349e-01 2.83669025e-01 2.99364656e-01
1.79603890e-01 -2.31942087e-01 2.85795659e-01 3.90708685e-01
-5.16856074e-01 3.32979232e-01 3.44017632e-02 5.69557965e-01
-1.93213329e-01 3.34705263e-01 -8.97156060e-01 -6.65311635e-01
8.04926991e-01 4.98795599e-01 3.30898985e-02 -1.07024364e-01
5.95519841e-01 1.00727350e-01 -3.77601892e-01 5.13671041e-01
1.00434780e+00 5.62425792e-01 -6.51173890e-01 2.08430681e-02
-7.63719797e-01 -3.72334838e-01 -1.18488085e+00 1.88566208e-01
-1.63431749e-01 1.39897704e+00 -8.85713566e-03 -6.53641880e-01
7.80081213e-01 2.67905414e-01 1.69794545e-01 -8.54813159e-01
2.41314903e-01 1.49212062e-01 6.59716055e-02 -1.04218996e+00
6.08978868e-01 2.34875962e-01 -5.67519292e-02 1.24127351e-01
-4.98986721e-01 4.41464007e-01 6.83293045e-01 1.29633531e-01
9.98456120e-01 -3.67591232e-01 3.86800796e-01 1.03820056e-01
7.36163139e-01 4.42902654e-01 6.95851207e-01 5.94687343e-01
-7.10744798e-01 3.98060158e-02 1.20809041e-01 -6.41605020e-01
-8.92341435e-01 -8.00934374e-01 -5.74813373e-02 9.27094996e-01
4.83807445e-01 1.31227881e-01 -8.41210544e-01 -3.03329438e-01
1.01844490e-01 5.95731258e-01 -3.24177474e-01 -1.83533177e-01
-8.21636558e-01 -2.14729458e-01 2.97907174e-01 2.39654347e-01
8.87331367e-01 -8.88553560e-01 -9.85195339e-01 4.26312089e-02
1.23074740e-01 -1.17760563e+00 -5.91083109e-01 -4.80330080e-01
-5.43762803e-01 -1.46269858e+00 -5.14952421e-01 -8.65746021e-01
7.40300477e-01 1.46107030e+00 4.51296568e-01 3.72628927e-01
-3.99225414e-01 1.14235915e-01 -1.13223508e-01 -6.27717555e-01
-4.29531991e-01 -2.91469902e-01 5.24439514e-02 1.45449698e-01
1.02741075e+00 2.57884949e-01 -7.48118699e-01 6.74481690e-01
-3.89041156e-01 -1.55903265e-01 4.12165463e-01 7.99790397e-02
2.06285894e-01 7.18136489e-01 3.76759201e-01 -6.65704548e-01
3.95044506e-01 -4.73280072e-01 -1.07397437e+00 -4.77632433e-02
-3.12312156e-01 -4.34474409e-01 3.65372002e-01 3.40522937e-02
-8.40754509e-01 1.69295624e-01 1.74908131e-01 -1.93516880e-01
-6.53447449e-01 -4.35134590e-01 -5.08191362e-02 -4.33133207e-02
1.36452809e-01 3.22100967e-01 1.45268053e-01 -8.64970461e-02
1.34980842e-01 1.01998889e+00 6.46762550e-01 5.12860179e-01
8.36376488e-01 6.37824774e-01 3.29617769e-01 -1.29803944e+00
-3.19664598e-01 -1.09132481e+00 -5.50997436e-01 -9.46943402e-01
1.18894446e+00 -7.78384924e-01 -1.14972579e+00 5.97536862e-01
-1.11964071e+00 4.00926173e-01 4.32444841e-01 7.65250683e-01
5.13272621e-02 3.66841555e-01 2.93200500e-02 -1.12073219e+00
-2.74765547e-02 -1.12757003e+00 7.81607985e-01 6.43977463e-01
1.11291081e-01 -7.76584268e-01 -1.24210693e-01 4.72344369e-01
3.02463204e-01 1.77012518e-01 2.93034136e-01 1.22540360e-02
-9.53118205e-01 -6.93996727e-01 -5.30680656e-01 1.90128922e-01
2.11992145e-01 4.15843964e-01 -6.32939339e-01 8.36520642e-02
-3.62286210e-01 7.73695111e-01 9.93608952e-01 8.15665245e-01
4.86530155e-01 9.81833860e-02 -9.34877872e-01 2.53249258e-01
1.32649779e+00 9.28898513e-01 7.71467328e-01 4.30736721e-01
6.84017420e-01 5.31157374e-01 1.02839339e+00 -1.15514271e-01
4.45496798e-01 5.01588464e-01 2.93484598e-01 -1.45365104e-01
-2.71472394e-01 -5.23834191e-02 4.21309441e-01 4.87831354e-01
-7.14693666e-02 -1.63997918e-01 -8.94502878e-01 7.46799827e-01
-1.83539951e+00 -1.66123021e+00 -1.09671021e+00 2.18570328e+00
-1.38458654e-01 3.02363783e-01 4.98950273e-01 3.25855941e-01
1.27146745e+00 -3.07731956e-01 -6.67506009e-02 -6.76475525e-01
3.34613174e-01 -5.20918965e-01 1.30761933e+00 5.43732226e-01
-1.04639018e+00 7.45064139e-01 6.06892681e+00 7.10718691e-01
-1.04156327e+00 1.42885655e-01 2.94710696e-01 1.30804926e-01
5.36651075e-01 2.32192669e-02 -1.26964331e+00 9.65665400e-01
8.69822323e-01 -1.16864026e-01 1.52219338e-02 7.91629016e-01
9.77091908e-01 -1.02416360e+00 -4.99959916e-01 1.06406128e+00
1.29518315e-01 -1.26595116e+00 -2.67390132e-01 2.11385429e-01
3.09529871e-01 -1.27624497e-01 -4.17659670e-01 1.20818712e-01
-4.81092222e-02 -4.98586923e-01 5.46969712e-01 7.52323806e-01
3.44396710e-01 -1.14785457e+00 8.34001899e-01 5.62591791e-01
-1.42705643e+00 -3.28122973e-01 -3.56725872e-01 -1.55739695e-01
5.20069718e-01 2.84088761e-01 -7.60750115e-01 -1.99581489e-01
7.18533039e-01 3.46201867e-01 -5.83981991e-01 1.60334623e+00
1.87102824e-01 6.69828534e-01 -2.44817212e-01 -1.82457864e-01
3.57695520e-01 -2.50952750e-01 6.94554627e-01 1.24699378e+00
3.73511136e-01 1.92821443e-01 1.18471123e-01 9.58283767e-02
3.62192631e-01 -3.20998013e-01 -1.13884068e+00 5.88169873e-01
5.48136473e-01 1.03114092e+00 -8.93069088e-01 -4.54802096e-01
-5.42919695e-01 3.32398176e-01 -5.17632842e-01 8.28275308e-02
-9.86529708e-01 -9.60350215e-01 4.97713953e-01 6.96680248e-01
4.66927320e-01 -3.50550026e-01 1.95159554e-01 -4.30360973e-01
-2.00331390e-01 -8.40348452e-02 2.79605418e-01 -9.72031116e-01
-4.01913017e-01 2.89046496e-01 3.55000705e-01 -1.41481495e+00
9.49960947e-02 -6.23854220e-01 -8.39561224e-01 3.75979394e-01
-1.52392709e+00 -7.93567121e-01 -4.57547128e-01 6.18864298e-01
8.57690692e-01 -3.73863906e-01 -3.80813144e-03 6.77247524e-01
-7.47877598e-01 1.53278038e-01 2.45365515e-01 2.09466144e-01
5.48030257e-01 -6.35336876e-01 5.30646779e-02 1.04024661e+00
-4.88358319e-01 2.95438528e-01 6.48513138e-01 -5.78973591e-01
-1.22400844e+00 -1.05509353e+00 1.08011878e+00 -4.82512861e-01
2.95299232e-01 -1.30797192e-01 -5.50854266e-01 2.96070278e-01
2.45555878e-01 1.00537710e-01 1.56551376e-01 -6.64591610e-01
3.22657853e-01 -6.89330280e-01 -1.09051859e+00 3.22429270e-01
5.21759689e-01 -2.53075529e-02 -5.78814864e-01 2.43735760e-01
2.84080684e-01 -8.80523399e-02 2.63071898e-02 -1.23407140e-01
7.01631129e-01 -9.18683648e-01 6.86052144e-01 -6.37193071e-03
-4.25097346e-01 -1.00227284e+00 3.02864403e-01 -7.78948247e-01
-2.92971194e-01 -3.33983958e-01 3.98147404e-01 1.17835867e+00
6.19071946e-02 -8.04250658e-01 4.54622984e-01 4.47327077e-01
-8.03073049e-02 -1.20077468e-01 -1.06811857e+00 -6.84253335e-01
-6.33762121e-01 -3.84161323e-01 3.82745117e-01 3.68517071e-01
-3.04276675e-01 3.98129374e-01 -6.72060668e-01 2.52074659e-01
8.52275908e-01 -2.29774177e-01 1.15435040e+00 -1.15709591e+00
5.29553056e-01 -1.56901285e-01 -9.15340364e-01 -7.75747418e-01
-2.54973441e-01 -1.07561365e-01 4.51355167e-02 -1.55507648e+00
3.56309026e-01 -1.07729018e-01 1.72558099e-01 1.17328718e-01
1.57634690e-02 3.75127226e-01 9.53618437e-02 1.02403231e-01
-8.09647262e-01 -2.23873854e-01 1.10475254e+00 1.23296566e-01
1.30853295e-01 3.41403097e-01 -4.18225005e-02 7.72190452e-01
8.71142209e-01 -8.26799810e-01 -4.15794849e-01 -2.31545582e-01
1.98608283e-02 8.46845806e-02 5.32130182e-01 -1.16063058e+00
4.99398887e-01 -2.83289105e-01 2.18140155e-01 -1.19204080e+00
1.77086696e-01 -1.26484275e+00 1.19893432e-01 7.66019940e-01
4.02282894e-01 3.12746197e-01 2.31372789e-01 7.42675424e-01
-7.13486895e-02 -4.07535017e-01 9.87712801e-01 2.08688304e-02
-1.30381811e+00 -6.22469485e-02 -1.30312717e+00 -3.44986588e-01
1.80016398e+00 -1.06483412e+00 -5.45392096e-01 -3.36052328e-01
-1.72964036e-01 5.24356186e-01 2.48120651e-01 6.02789998e-01
6.41921997e-01 -1.26249969e+00 -4.62200493e-01 3.25335592e-01
8.60971957e-02 -4.87505347e-01 3.14421237e-01 9.65583444e-01
-9.28193629e-01 8.97765577e-01 -3.18757981e-01 -8.07399452e-01
-1.96845353e+00 6.76812291e-01 2.41742194e-01 4.24019009e-01
-5.91881812e-01 2.58027017e-01 6.99911788e-02 2.85708904e-01
-7.80385956e-02 -2.92560369e-01 -5.44548869e-01 -2.90867507e-01
8.48560095e-01 1.07631302e+00 -2.62053818e-01 -1.48250675e+00
-6.79524124e-01 1.05576539e+00 -3.14252749e-02 2.84016490e-01
7.38123417e-01 -5.07965088e-01 1.05662271e-01 3.15054864e-01
1.04395866e+00 1.26309529e-01 -1.01026082e+00 3.42858702e-01
-2.64337420e-01 -9.07128274e-01 3.14189434e-01 -2.14240462e-01
-9.18761373e-01 8.24266016e-01 1.07754731e+00 4.85041529e-01
8.18977177e-01 -1.18446499e-01 1.17126822e+00 3.07563156e-01
3.76786321e-01 -1.28367341e+00 -4.99027342e-01 3.77449483e-01
9.82966051e-02 -1.36271071e+00 5.12545630e-02 -3.83497268e-01
-6.05096996e-01 1.05418754e+00 5.78960598e-01 -2.18546480e-01
7.08023965e-01 1.42727032e-01 1.68121576e-01 -3.12002271e-01
-6.06876612e-01 -5.54576814e-01 -2.44389232e-02 8.13780665e-01
-1.65635180e-02 1.44417673e-01 -3.19663554e-01 -3.48451018e-01
2.62537569e-01 1.20089233e-01 7.77370811e-01 9.27176058e-01
-1.31691730e+00 -8.07936668e-01 -9.13123310e-01 4.88197774e-01
-4.68664706e-01 5.72605789e-01 -1.28951492e-02 1.10178959e+00
6.28387630e-01 1.39774454e+00 2.22295389e-01 -1.04963884e-01
5.07212341e-01 -1.67064890e-01 5.76980710e-02 -1.35241777e-01
5.86699694e-02 -2.06114978e-01 1.52551576e-01 -4.39120114e-01
-4.46951181e-01 -7.31086493e-01 -1.32985330e+00 -7.28634119e-01
-5.58079541e-01 2.05089808e-01 7.65349567e-01 9.69359875e-01
2.75724947e-01 2.52488226e-01 6.35626197e-01 -5.38684607e-01
3.60532790e-01 -4.63083088e-01 -4.69058007e-01 1.58085912e-01
6.48236394e-01 -1.00659728e+00 -1.07024923e-01 2.24066079e-01] | [7.966064929962158, -1.1080280542373657] |
7018b60b-d8b4-4af2-9a9c-a883a08ee3be | bayesian-optimization-of-expensive-nested | 2306.05150 | null | https://arxiv.org/abs/2306.05150v1 | https://arxiv.org/pdf/2306.05150v1.pdf | Bayesian Optimization of Expensive Nested Grey-Box Functions | We consider the problem of optimizing a grey-box objective function, i.e., nested function composed of both black-box and white-box functions. A general formulation for such grey-box problems is given, which covers the existing grey-box optimization formulations as special cases. We then design an optimism-driven algorithm to solve it. Under certain regularity assumptions, our algorithm achieves similar regret bound as that for the standard black-box Bayesian optimization algorithm, up to a constant multiplicative term depending on the Lipschitz constants of the functions considered. We further extend our method to the constrained case and discuss several special cases. For the commonly used kernel functions, the regret bounds allow us to derive a convergence rate to the optimal solution. Experimental results show that our grey-box optimization method empirically improves the speed of finding the global optimal solution significantly, as compared to the standard black-box optimization algorithm. | ['Colin N. Jones', 'Bratislav Svetozarevic', 'Yuning Jiang', 'Wenjie Xu'] | 2023-06-08 | null | null | null | null | ['bayesian-optimization'] | ['methodology'] | [ 6.46146461e-02 3.08139771e-01 -1.78565174e-01 -5.09596705e-01
-1.16002226e+00 -5.16727149e-01 1.11588642e-01 2.71706395e-02
-4.87539709e-01 8.09529006e-01 -1.64664071e-02 -3.42294335e-01
-6.31812274e-01 -4.64688897e-01 -7.92719662e-01 -1.06045794e+00
-2.76292980e-01 5.63713431e-01 -2.05786414e-02 -1.59441799e-01
3.66713583e-01 3.54028940e-01 -9.01910961e-01 -3.49526554e-01
9.78799820e-01 1.34965909e+00 -1.11069843e-01 8.02581310e-01
1.08741589e-01 3.27751160e-01 -4.91163462e-01 -6.11937642e-01
7.08585680e-01 -2.20030978e-01 -6.23762071e-01 1.92575276e-01
1.91706896e-01 -1.62027910e-01 9.48266089e-02 1.16381991e+00
4.87725794e-01 5.72426021e-01 5.58572114e-01 -1.04942632e+00
-3.71033728e-01 6.17380440e-01 -8.10445189e-01 1.46041578e-02
-5.05443402e-02 -4.43592340e-01 1.22557616e+00 -8.13191772e-01
2.23589078e-01 1.42561758e+00 8.58692288e-01 4.92306024e-01
-1.25415683e+00 -2.20474333e-01 4.86833930e-01 -1.07921392e-01
-1.24708116e+00 -2.08533823e-01 3.68680924e-01 -2.32866913e-01
5.84400833e-01 4.20486420e-01 3.41732532e-01 4.69274014e-01
3.60376686e-01 7.91212976e-01 1.22703791e+00 -5.13348460e-01
5.26142061e-01 1.23449817e-01 2.01546520e-01 8.95069420e-01
4.29374874e-01 3.12210649e-01 -3.82146180e-01 -3.99944454e-01
4.86666530e-01 -1.94735706e-01 -3.61467898e-01 -7.61030018e-01
-9.12953556e-01 1.16056740e+00 3.28193992e-01 -2.84490343e-02
-5.39044589e-02 5.05467534e-01 5.77194989e-02 2.27051347e-01
9.85762358e-01 3.70988250e-01 -5.01438618e-01 -6.88900352e-02
-1.04085600e+00 3.31992894e-01 1.15453351e+00 8.34806979e-01
6.03844583e-01 -3.57802622e-02 -2.26856172e-01 7.38654375e-01
6.36184454e-01 5.75213552e-01 -1.22927524e-01 -9.49890792e-01
4.79379714e-01 1.36431262e-01 6.28176332e-01 -5.16264379e-01
-5.85717380e-01 -5.32883883e-01 -6.27601683e-01 5.66743731e-01
9.91183877e-01 -5.55612683e-01 -6.54502571e-01 1.46659791e+00
6.23282611e-01 -2.40007892e-01 -1.30334407e-01 1.03631234e+00
1.56723872e-01 7.74095595e-01 -5.57449162e-01 -3.58993411e-01
1.25167620e+00 -1.03772998e+00 -9.34111178e-01 -3.18916738e-01
6.39558852e-01 -6.64072216e-01 9.92294669e-01 4.72812027e-01
-1.16306818e+00 6.93763345e-02 -1.16338062e+00 7.49452636e-02
-1.66814178e-01 5.76445926e-03 6.62206292e-01 1.30154443e+00
-7.80753136e-01 7.47167706e-01 -6.20651186e-01 9.28578433e-04
3.49956393e-01 4.76927012e-01 -8.57169107e-02 1.01586081e-01
-7.61160135e-01 1.06273723e+00 3.41356575e-01 5.56415677e-01
-8.21523368e-01 -8.68704975e-01 -7.92537987e-01 1.21114403e-01
9.28375423e-01 -6.88666403e-01 1.45133984e+00 -1.01188588e+00
-1.88612843e+00 5.69527209e-01 -2.71984220e-01 -4.41606164e-01
8.54885876e-01 -4.52495098e-01 1.38032109e-01 -7.56536499e-02
-1.92719027e-01 5.12914062e-02 1.10924280e+00 -1.13448215e+00
-4.60760295e-01 -3.61122310e-01 3.33391607e-01 1.91169381e-01
7.70451054e-02 3.20723981e-01 -3.53715241e-01 -8.42783809e-01
-1.55568093e-01 -8.03974271e-01 -6.39839947e-01 3.35081011e-01
-2.12829471e-01 1.58759039e-02 3.50758523e-01 -3.92683744e-01
1.29704428e+00 -2.08393145e+00 1.67052060e-01 5.41680634e-01
-6.57043383e-02 -4.25224125e-01 1.54421851e-01 3.07065338e-01
-1.30535319e-01 7.77075887e-02 -6.82059467e-01 -4.56990063e-01
3.24264854e-01 2.52755910e-01 -2.55660772e-01 9.56449628e-01
8.28538686e-02 8.32150936e-01 -7.78645039e-01 -2.31372640e-01
-4.32552099e-02 5.79219200e-02 -5.35342157e-01 5.23334406e-02
-1.89382121e-01 1.67470261e-01 -4.26305115e-01 5.14437795e-01
6.45862579e-01 -2.32166603e-01 -3.73154953e-02 3.04408848e-01
-2.17489637e-02 -2.18612507e-01 -1.52871823e+00 1.35922754e+00
-7.06689417e-01 4.70319718e-01 7.31201231e-01 -1.05412638e+00
8.82040262e-01 -5.84600791e-02 3.82291734e-01 1.27081049e-03
2.49348566e-01 2.28275865e-01 -4.03228588e-02 -1.15641221e-01
2.73139477e-01 -8.48576844e-01 -2.03438714e-01 4.32237059e-01
-1.52134255e-01 -1.57238498e-01 6.24061339e-02 -9.13568065e-02
7.83261299e-01 1.04266107e-01 4.21726853e-01 -6.11818731e-01
3.71830255e-01 -1.33679107e-01 5.01482725e-01 1.11904502e+00
-1.34741232e-01 5.19225121e-01 7.82755792e-01 -2.73904651e-01
-8.09179842e-01 -7.21812367e-01 -4.95610118e-01 1.28750992e+00
2.73668736e-01 -2.71443576e-01 -7.52686620e-01 -8.39354634e-01
3.30230922e-01 5.37438095e-01 -9.52078402e-01 2.44976748e-02
-4.20782030e-01 -1.47303379e+00 9.85174999e-02 3.90398234e-01
3.38412046e-01 -2.06335664e-01 -6.24510944e-01 9.08472687e-02
2.88166434e-01 -7.51204669e-01 -6.70872629e-01 3.69787186e-01
-1.02768481e+00 -9.23764884e-01 -9.11531568e-01 -2.23105222e-01
6.80499911e-01 -9.25781280e-02 8.39867711e-01 -2.74761438e-01
-5.64372465e-02 3.42108369e-01 -1.61420405e-01 -4.53600287e-01
6.38642758e-02 -3.52423549e-01 -2.13101000e-01 3.28494757e-01
-2.61684388e-01 1.42564923e-01 -4.89181101e-01 6.18578851e-01
-8.15573096e-01 -3.66900712e-01 3.78610343e-01 1.04885089e+00
6.12346411e-01 1.81933552e-01 4.81590658e-01 -9.49735045e-01
7.38041937e-01 -3.53765637e-01 -1.45338988e+00 3.88724476e-01
-7.75628924e-01 5.40821493e-01 4.41359222e-01 -3.65983993e-01
-1.15863323e+00 1.58137172e-01 1.28216535e-01 -6.80121779e-02
5.62166393e-01 3.93871754e-01 -2.64191210e-01 -6.54666603e-01
4.66850877e-01 -4.09319878e-01 -9.87837166e-02 -4.75597739e-01
7.08087206e-01 4.75200176e-01 2.80084431e-01 -9.10345256e-01
8.26459348e-01 7.02351034e-01 2.97834456e-01 -5.10202885e-01
-1.18522000e+00 -4.94183540e-01 -4.10753548e-01 -2.60874599e-01
6.29363239e-01 -4.08438385e-01 -1.08224070e+00 -6.55381009e-02
-9.05963361e-01 -5.27550101e-01 -5.17440200e-01 5.00299215e-01
-9.47362602e-01 4.73856479e-01 -1.19214639e-01 -1.53439248e+00
-9.57109332e-02 -1.04582787e+00 1.21942949e+00 -8.11073638e-04
8.76629204e-02 -1.37771678e+00 2.67323077e-01 3.31595957e-01
2.26240128e-01 2.26353496e-01 6.21119857e-01 -3.80063981e-01
-4.86489445e-01 -3.06510091e-01 -1.79610685e-01 3.88085157e-01
-5.20280838e-01 -1.82398364e-01 -6.43450975e-01 -3.43123853e-01
5.93948603e-01 -2.89198291e-02 9.77656603e-01 7.20048010e-01
1.10272145e+00 -4.44648951e-01 -2.20382139e-01 9.49818909e-01
1.46905100e+00 -1.11377854e-02 2.71599948e-01 5.67869246e-01
2.27161303e-01 6.30024791e-01 9.67096984e-01 7.06569970e-01
-8.65159556e-02 6.12404644e-01 4.85530019e-01 5.35949841e-02
5.68639576e-01 1.76829979e-01 4.09125268e-01 2.01428235e-01
1.33483231e-01 -7.38938451e-02 -6.61284447e-01 3.37244183e-01
-2.27440763e+00 -5.29704928e-01 -8.42929035e-02 2.64538646e+00
6.80349052e-01 1.73148345e-02 2.55369931e-01 -1.14725858e-01
7.06114888e-01 2.01083526e-01 -6.74822450e-01 -6.81454778e-01
-1.08802639e-01 3.76433581e-01 1.20656669e+00 9.86565888e-01
-1.24321258e+00 6.28416538e-01 8.13812065e+00 1.09074819e+00
-5.05084157e-01 5.09395361e-01 6.07795835e-01 -5.61511517e-01
-1.17137983e-01 2.55224586e-01 -9.48542833e-01 3.97477537e-01
8.43752086e-01 -2.12861732e-01 6.00993514e-01 9.19393539e-01
3.98574829e-01 -3.99014980e-01 -8.86673629e-01 8.39113176e-01
-2.39038412e-02 -9.52053070e-01 -7.70814002e-01 7.37692788e-02
9.74261045e-01 -3.03352028e-01 -7.37272352e-02 -1.12875821e-02
4.91445422e-01 -1.19277012e+00 6.67233348e-01 4.97440457e-01
4.31620479e-01 -8.41340780e-01 8.10557902e-01 2.22740188e-01
-8.22211683e-01 -2.41641030e-01 -4.87773955e-01 -1.84566200e-01
4.10336912e-01 8.94035578e-01 -1.95561588e-01 5.70924759e-01
6.23619974e-01 2.48260245e-01 -5.22795878e-02 1.49143851e+00
-2.72116393e-01 5.32183766e-01 -8.11829209e-01 -3.35744470e-01
4.64241832e-01 -6.56238794e-01 6.62052989e-01 1.08416128e+00
1.67708486e-01 -4.23538424e-02 -4.04672474e-02 7.42700577e-01
1.22354664e-01 2.86410242e-01 -1.01660058e-01 2.42711633e-01
-2.86891665e-02 1.23110139e+00 -6.09838605e-01 -1.58641279e-01
-4.69277471e-01 6.80286825e-01 3.41803789e-01 5.22520959e-01
-8.18160772e-01 -6.69224679e-01 6.26680493e-01 -1.80509508e-01
6.22305810e-01 -2.38254309e-01 -6.60817862e-01 -1.14634550e+00
1.22929797e-01 -7.00489879e-01 5.31920791e-01 -4.92446661e-01
-1.33560801e+00 4.58308309e-02 3.64485860e-01 -7.15063870e-01
8.90876539e-03 -1.06062853e+00 -4.86660570e-01 7.90582240e-01
-1.42389774e+00 -5.78248143e-01 1.27141625e-01 3.74603152e-01
1.46354973e-01 8.16265196e-02 3.88323039e-01 3.35707515e-02
-7.70521104e-01 5.50646544e-01 9.29978013e-01 -3.29075009e-01
4.88499731e-01 -1.63002563e+00 -7.01453984e-02 9.06431019e-01
3.30428389e-04 5.67696154e-01 8.78500938e-01 -4.78958309e-01
-1.46617711e+00 -6.03738368e-01 4.25229669e-01 -3.85490447e-01
1.11382449e+00 -5.15582740e-01 -5.94425201e-01 6.40253365e-01
-6.01944588e-02 1.81041081e-02 7.36532152e-01 3.72603118e-01
-1.49883315e-01 -2.47673243e-01 -1.29273224e+00 5.67854583e-01
7.05699027e-01 -9.05402601e-02 -5.74625850e-01 6.40007794e-01
2.91929543e-01 -5.38133383e-01 -1.09901440e+00 4.26765472e-01
6.47232115e-01 -6.38036668e-01 8.64912868e-01 -7.06876695e-01
-1.18275277e-01 -4.30070221e-01 -1.14515550e-01 -1.14930069e+00
-8.01180676e-02 -1.35094059e+00 -4.61006999e-01 5.77035725e-01
3.58850062e-01 -9.11626518e-01 6.48716390e-01 7.98518717e-01
1.04927443e-01 -1.14880359e+00 -1.38669670e+00 -1.09294200e+00
2.82658011e-01 -5.64744830e-01 1.74772501e-01 2.74310678e-01
5.43946214e-02 9.99901071e-02 -6.49317563e-01 2.64483839e-01
8.01642060e-01 2.94036627e-01 6.74924195e-01 -9.09203351e-01
-7.26711035e-01 -6.87047839e-01 -1.89915150e-01 -1.47014570e+00
2.29327455e-01 -7.10398197e-01 2.46539220e-01 -1.20180511e+00
1.30894393e-01 -3.26826900e-01 -5.93030499e-03 1.77434638e-01
-4.01095718e-01 2.23778724e-03 2.61318296e-01 -2.06968799e-01
-4.62349474e-01 6.39496326e-01 1.23304689e+00 1.35424539e-01
-6.96872696e-02 4.21926290e-01 -8.32165122e-01 7.43630946e-01
6.82480097e-01 -6.89190507e-01 -6.83145672e-02 -1.83982387e-01
5.73005378e-01 -7.02667460e-02 1.84650630e-01 -4.94267762e-01
-3.18816379e-02 -2.57669091e-01 9.35184434e-02 -4.02173191e-01
5.12742758e-01 -8.70705605e-01 -1.43844694e-01 3.35520089e-01
-4.00160223e-01 -2.60305703e-01 7.98621699e-02 7.70903707e-01
6.15791455e-02 -1.03385866e+00 1.07295251e+00 2.33661383e-01
-7.91271254e-02 1.26983151e-01 -3.57000411e-01 1.83825120e-01
1.07698762e+00 -1.33715540e-01 -2.68505067e-01 -7.16607690e-01
-7.77970970e-01 5.73818803e-01 8.74610692e-02 -1.60923243e-01
3.45404088e-01 -1.12652862e+00 -5.66984713e-01 -2.37046421e-01
-2.26989195e-01 -4.38749455e-02 -1.20375387e-01 1.20655203e+00
-4.88147885e-01 3.23735565e-01 4.87987131e-01 -2.98533559e-01
-1.06523490e+00 6.48056626e-01 6.01163566e-01 -4.19009179e-01
-3.94571453e-01 1.01486576e+00 3.31853360e-01 -2.53946215e-01
4.05208945e-01 -5.57124138e-01 4.39905077e-01 -1.54107958e-01
4.15280640e-01 7.21224427e-01 -9.63935032e-02 -2.34680414e-01
-2.64859736e-01 8.38240325e-01 3.18659186e-01 -5.04249394e-01
1.13227427e+00 -1.36519209e-01 -1.78188682e-01 6.10566378e-01
1.12222648e+00 1.24189116e-01 -1.40399885e+00 -1.43614128e-01
1.48150235e-01 -6.72142804e-01 2.65554905e-01 -7.60624826e-01
-8.70557487e-01 8.18895042e-01 4.23656493e-01 2.77287453e-01
1.09516108e+00 -1.11650312e-02 3.00293654e-01 6.68042958e-01
2.52101630e-01 -1.23681891e+00 -2.95053929e-01 4.66957062e-01
9.46677387e-01 -1.23530376e+00 3.20176721e-01 -4.81740624e-01
-4.23101127e-01 1.25494289e+00 7.10540190e-02 -4.01522398e-01
8.03965867e-01 3.39448273e-01 -8.30526799e-02 1.63828924e-01
-5.74067414e-01 -4.73444730e-01 5.63827515e-01 2.45602414e-01
1.79579705e-01 -1.64297950e-02 -4.70309585e-01 5.10232508e-01
-1.60183325e-01 -4.60232317e-01 2.26442248e-01 7.23742247e-01
-6.98351443e-01 -1.12423956e+00 -8.83714259e-01 3.24767947e-01
-8.15312445e-01 -9.72164422e-02 -4.05802101e-01 8.87499571e-01
-4.74750072e-01 1.12431777e+00 -1.17581837e-01 3.97293895e-01
1.31355479e-01 9.49820504e-02 7.32470453e-01 -4.19118881e-01
-6.66038454e-01 4.20166135e-01 1.39433607e-01 -6.23436689e-01
-3.07603776e-01 -5.48546612e-01 -9.66257989e-01 -2.61679053e-01
-7.78930843e-01 2.95706153e-01 7.73765922e-01 1.07052135e+00
-1.45412654e-01 2.71886170e-01 5.70364177e-01 -7.03514397e-01
-1.14770353e+00 -7.69231379e-01 -7.04114377e-01 -9.58990902e-02
4.37820017e-01 -7.42262959e-01 -4.85715806e-01 -5.27439594e-01] | [6.473679065704346, 4.118503570556641] |
37deb794-3569-4104-8406-1ec708078452 | contextualized-embeddings-based-transformer | null | null | https://aclanthology.org/2020.lrec-1.676 | https://aclanthology.org/2020.lrec-1.676.pdf | Contextualized Embeddings based Transformer Encoder for Sentence Similarity Modeling in Answer Selection Task | Word embeddings that consider context have attracted great attention for various natural language processing tasks in recent years. In this paper, we utilize contextualized word embeddings with the transformer encoder for sentence similarity modeling in the answer selection task. We present two different approaches (feature-based and fine-tuning-based) for answer selection. In the feature-based approach, we utilize two types of contextualized embeddings, namely the Embeddings from Language Models (ELMo) and the Bidirectional Encoder Representations from Transformers (BERT) and integrate each of them with the transformer encoder. We find that integrating these contextual embeddings with the transformer encoder is effective to improve the performance of sentence similarity modeling. In the second approach, we fine-tune two pre-trained transformer encoder models for the answer selection task. Based on our experiments on six datasets, we find that the fine-tuning approach outperforms the feature-based approach on all of them. Among our fine-tuning-based models, the Robustly Optimized BERT Pretraining Approach (RoBERTa) model results in new state-of-the-art performance across five datasets. | ['Md Tahmid Rahman Laskar', 'Enamul Hoque', 'Jimmy Xiangji Huang'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['answer-selection'] | ['natural-language-processing'] | [-1.57437012e-01 -3.12275231e-01 -1.70730412e-01 -6.28045619e-01
-7.56256163e-01 -2.49489844e-01 4.90249068e-01 5.20743787e-01
-7.90874839e-01 3.03441286e-01 7.58207977e-01 -4.42337215e-01
-1.79909900e-01 -9.44125354e-01 -3.08500320e-01 -2.07604989e-01
3.85483831e-01 3.20535868e-01 4.89568949e-01 -6.22030616e-01
6.61925435e-01 -1.09544285e-01 -1.46173930e+00 5.31816125e-01
1.23361707e+00 1.05328953e+00 2.75989085e-01 6.21715665e-01
-7.65730917e-01 6.28613710e-01 -2.56345719e-01 -7.71704197e-01
-6.58570081e-02 -3.22813064e-01 -1.06971097e+00 -5.38082540e-01
1.54256061e-01 -2.06766561e-01 -4.02662456e-01 7.16879368e-01
7.05972195e-01 5.06769180e-01 5.62070727e-01 -8.20215523e-01
-1.31330276e+00 5.00360310e-01 -2.17027500e-01 4.59962815e-01
6.37733638e-01 -1.31296724e-01 1.55582929e+00 -1.47527099e+00
4.68289524e-01 1.46915662e+00 7.80473530e-01 3.55625957e-01
-1.06289458e+00 -3.51067394e-01 2.68762916e-01 6.87234521e-01
-1.25486565e+00 -2.37963289e-01 6.87659442e-01 -2.99148411e-01
1.62711835e+00 6.57353103e-02 3.84368718e-01 7.89063215e-01
5.90790749e-01 7.62728870e-01 6.10634029e-01 -6.77350521e-01
1.53855339e-01 3.36236715e-01 6.17095053e-01 7.24752367e-01
-4.95464727e-02 -1.32736489e-01 -4.51962888e-01 -3.65495712e-01
1.27454355e-01 2.73923784e-01 -6.28549084e-02 -1.22821331e-01
-1.08825612e+00 1.31242979e+00 5.18931329e-01 5.13192356e-01
-2.33407393e-01 -8.09924453e-02 6.35053039e-01 6.11334205e-01
6.92542732e-01 6.93064213e-01 -6.65449560e-01 -1.29470363e-01
-5.57820916e-01 2.13947475e-01 5.38083732e-01 8.40090036e-01
8.48767400e-01 -2.63183773e-01 -8.16337883e-01 1.26936877e+00
3.92246336e-01 5.07424027e-02 1.07711649e+00 -4.23536807e-01
7.09204257e-01 8.92561376e-01 -2.06272721e-01 -1.01757634e+00
-2.40177095e-01 -3.51720303e-01 -5.09597182e-01 -6.61757648e-01
3.70047651e-02 -1.66954949e-01 -5.84560573e-01 1.67944324e+00
2.76852846e-01 1.59673229e-01 1.20571256e-01 6.89947546e-01
1.20821452e+00 7.80243218e-01 1.22653596e-01 1.54112726e-01
1.51853538e+00 -1.49912572e+00 -7.59222448e-01 -3.15326393e-01
1.10479867e+00 -8.44133139e-01 1.49633527e+00 -2.24095523e-01
-7.38413751e-01 -8.58108103e-01 -1.25001812e+00 -3.94436002e-01
-5.83994865e-01 1.14625871e-01 3.70159388e-01 4.18550730e-01
-9.55963850e-01 6.30200922e-01 -4.43962127e-01 -7.75635719e-01
-7.36711035e-03 2.09667385e-01 -2.95991927e-01 -2.15295613e-01
-1.60229635e+00 1.22461414e+00 7.49486089e-02 -3.38548541e-01
-2.37479225e-01 -5.38506448e-01 -9.62874055e-01 3.27598810e-01
-1.42150998e-01 -1.06066453e+00 1.08619916e+00 -2.95695484e-01
-1.40651774e+00 6.80820704e-01 -5.97272217e-01 -2.97344238e-01
-3.78489345e-01 -2.83642709e-01 -3.26538056e-01 6.97521418e-02
2.36832619e-01 5.19825041e-01 8.01552534e-01 -7.35629022e-01
-3.95665914e-01 -2.11540669e-01 2.79204667e-01 1.99981794e-01
-1.07680798e+00 3.23384821e-01 -2.64418721e-01 -6.48617387e-01
-6.80445731e-02 -6.03877664e-01 -2.30596080e-01 -1.10697567e-01
4.41009039e-03 -8.89827013e-01 5.70141494e-01 -4.61652011e-01
1.81923532e+00 -1.95695627e+00 1.88300654e-01 -4.07822698e-01
2.11542565e-02 6.26071990e-01 -6.71390176e-01 8.43532205e-01
-1.92360207e-01 2.82641768e-01 -9.12052952e-03 -6.60368621e-01
2.41054118e-01 2.28812262e-01 -4.34464067e-01 -1.75098047e-01
5.41079402e-01 1.09150052e+00 -9.75358546e-01 -5.32361150e-01
4.78904583e-02 2.57601887e-01 -8.92690361e-01 8.12492132e-01
9.76696610e-02 -4.52707186e-02 -5.01502633e-01 3.09371352e-01
4.27336991e-01 -1.14435777e-02 5.41606918e-03 -2.44328335e-01
2.11702839e-01 7.57879794e-01 -7.85455287e-01 1.82750714e+00
-8.00392330e-01 3.00969243e-01 -4.84182179e-01 -8.09196115e-01
1.17000341e+00 3.55903864e-01 8.83230269e-02 -8.64073217e-01
1.22470945e-01 3.57000202e-01 7.95131847e-02 -9.40371037e-01
9.44108784e-01 -2.90105999e-01 -1.31166115e-01 7.79552996e-01
5.57807088e-01 -9.56217423e-02 2.54495203e-01 3.07770610e-01
1.32347131e+00 -1.65798485e-01 4.05848175e-01 -1.60665140e-01
9.16781187e-01 -4.32519913e-01 5.41528761e-01 4.84570712e-01
-3.55926037e-01 6.67968333e-01 2.61555046e-01 -3.71034592e-01
-7.23198712e-01 -1.01735806e+00 7.84831867e-03 1.38211453e+00
3.29820905e-03 -9.11828101e-01 -5.78551710e-01 -9.20095563e-01
7.55701140e-02 8.69823039e-01 -6.80376709e-01 -7.11741805e-01
-6.20010078e-01 -5.48932254e-01 2.39405110e-01 8.78315210e-01
2.40098119e-01 -9.26787615e-01 -1.14529639e-01 2.63422787e-01
-2.95830578e-01 -9.24977899e-01 -8.71158123e-01 2.00078487e-01
-8.81161511e-01 -7.34253466e-01 -4.03559923e-01 -9.54087973e-01
4.18637276e-01 3.95824611e-01 1.20233846e+00 2.92641044e-01
1.28289193e-01 4.18597162e-01 -9.39929008e-01 -1.72096252e-01
5.71045987e-02 1.98027834e-01 1.94062382e-01 2.44255260e-01
9.32359815e-01 -3.80499989e-01 -4.82547849e-01 2.41000175e-01
-9.28780615e-01 -5.86600959e-01 2.19257414e-01 1.12461793e+00
3.50686282e-01 -4.46670175e-01 1.07604122e+00 -6.94374740e-01
1.34582496e+00 -6.52303755e-01 2.23733604e-01 4.72689450e-01
-6.89052761e-01 3.65497857e-01 7.01461315e-01 -3.07118177e-01
-9.43137944e-01 -5.00027239e-01 -6.65142238e-01 -2.19239429e-01
2.18771532e-01 7.16255248e-01 6.43971423e-03 1.04054779e-01
4.53741938e-01 -6.70373738e-02 -2.79651374e-01 -6.70681953e-01
4.33439434e-01 9.60234880e-01 -4.49847393e-02 -5.85046232e-01
6.00809097e-01 -1.17322110e-01 -6.95517123e-01 -4.86691386e-01
-1.09105754e+00 -6.06166601e-01 -6.78408146e-01 1.67555824e-01
1.04414761e+00 -5.29641509e-01 4.77632508e-02 9.00253304e-04
-1.54529369e+00 2.88141042e-01 -3.42172623e-01 6.24845326e-01
-2.59556711e-01 3.62344682e-01 -6.78917527e-01 -4.92982477e-01
-6.44720376e-01 -1.28185952e+00 8.68001699e-01 2.04114482e-01
-4.58297879e-01 -1.39054668e+00 3.93712431e-01 3.48574311e-01
8.26193035e-01 -3.94395620e-01 1.47598445e+00 -1.09499514e+00
-8.99301991e-02 -3.91238719e-01 -2.64578015e-01 4.60403293e-01
1.97992086e-01 -2.79928178e-01 -1.01694131e+00 -9.04102251e-02
3.27946782e-01 -2.89778352e-01 1.14305699e+00 4.51286659e-02
1.08619463e+00 9.44276303e-02 -1.28438711e-01 3.16428721e-01
1.27518082e+00 6.52499720e-02 5.11290252e-01 3.39853287e-01
4.36086535e-01 4.82673049e-01 7.84984171e-01 3.22417438e-01
7.72587895e-01 7.43455589e-01 1.58174455e-01 3.23257387e-01
-5.38874231e-02 -4.06431109e-01 4.64004695e-01 1.67590106e+00
3.67564231e-01 -2.27744505e-01 -6.17899835e-01 6.85560226e-01
-2.01110482e+00 -8.88223588e-01 4.39504161e-02 2.10234141e+00
9.23409700e-01 2.81193033e-02 -1.71695799e-01 5.72392046e-02
6.47031307e-01 3.79695028e-01 -2.13419557e-01 -1.26853049e+00
1.82803839e-01 6.33909166e-01 -3.04859608e-01 6.51039302e-01
-1.04246283e+00 9.35548246e-01 5.92103767e+00 8.81968439e-01
-6.69152081e-01 5.03956139e-01 1.60850137e-01 1.81811646e-01
-7.39190638e-01 2.05115199e-01 -9.30966377e-01 3.23331237e-01
1.19166303e+00 -4.17997807e-01 6.53509125e-02 5.85761726e-01
2.27566063e-02 1.39942169e-01 -1.38660562e+00 7.34525919e-01
4.23405439e-01 -1.33056688e+00 1.75730392e-01 -5.34538567e-01
5.83061337e-01 -1.84538335e-01 -1.02266982e-01 9.00377512e-01
-1.14511095e-01 -8.21326435e-01 3.15435410e-01 5.59730887e-01
3.81875932e-01 -6.58095002e-01 1.13471746e+00 2.41540447e-01
-1.50232685e+00 -2.88640052e-01 -7.12802708e-01 -2.62485534e-01
3.75434607e-01 6.75080717e-01 -3.74307573e-01 6.85754776e-01
5.63088477e-01 8.61253858e-01 -8.09029639e-01 8.73907208e-01
-4.34557408e-01 6.20990574e-01 1.45423442e-01 -3.31111521e-01
2.68956900e-01 -1.30197302e-01 3.20668221e-01 1.21349704e+00
2.05174863e-01 -1.18167214e-01 8.80667865e-02 7.50658631e-01
-1.68783143e-01 3.44046533e-01 -5.61582148e-01 -6.21446315e-03
6.27086639e-01 1.26988804e+00 3.70385796e-02 -3.94102156e-01
-6.59006596e-01 9.18043852e-01 7.15865552e-01 1.20404273e-01
-7.81759858e-01 -9.76440072e-01 7.52027690e-01 -1.34660229e-01
4.01208669e-01 -1.85952380e-01 -2.92940736e-01 -1.41497612e+00
8.60532746e-02 -6.29868448e-01 4.15371954e-01 -5.73609948e-01
-1.82869327e+00 8.76637995e-01 -1.00718640e-01 -1.21148682e+00
-2.54260510e-01 -6.20782673e-01 -1.13474834e+00 1.07639098e+00
-1.74436605e+00 -9.64642942e-01 1.21174827e-01 2.87906438e-01
7.61917889e-01 -2.67708808e-01 1.20637071e+00 5.87345183e-01
-7.02440679e-01 8.67257714e-01 1.15624949e-01 -1.05588334e-02
9.22440767e-01 -1.07485092e+00 3.79627168e-01 6.41159058e-01
1.90164179e-01 1.12280130e+00 2.08719090e-01 -2.40201488e-01
-1.26400077e+00 -1.11260819e+00 1.84565496e+00 -4.65400517e-01
8.66620362e-01 -2.62084991e-01 -1.12531269e+00 5.60877979e-01
7.44415998e-01 -1.74107447e-01 1.11524808e+00 4.46378887e-01
-5.44370472e-01 -1.76440790e-01 -9.04758930e-01 4.71361995e-01
6.81704998e-01 -8.80606472e-01 -1.36719525e+00 1.66298285e-01
1.29053044e+00 8.18707868e-02 -9.83803272e-01 4.13101584e-01
2.75080562e-01 -9.45171058e-01 8.95740092e-01 -8.98681223e-01
7.10458398e-01 -1.02723300e-01 -4.10113394e-01 -1.65180862e+00
-5.04552722e-01 1.39534231e-02 -1.69172719e-01 1.48922074e+00
5.06303608e-01 -7.85367906e-01 3.04907143e-01 3.85925889e-01
-3.27489972e-01 -1.45125997e+00 -8.38338614e-01 -7.47748256e-01
5.07221580e-01 -3.25523436e-01 7.87346363e-01 7.84936786e-01
3.23437840e-01 9.13580894e-01 1.16343819e-01 -3.88567120e-01
-2.38866154e-02 1.94967017e-01 4.08344626e-01 -9.38213408e-01
-2.48304605e-01 -3.12805593e-01 -2.94027179e-01 -1.32649326e+00
4.17531252e-01 -1.11901128e+00 -2.02755705e-01 -1.90655124e+00
2.92822152e-01 -1.86161339e-01 -5.72883785e-01 2.08389252e-01
-7.62664020e-01 -7.95008689e-02 1.35887086e-01 -2.65139043e-01
-6.53816521e-01 1.11050868e+00 9.09820974e-01 -2.22312331e-01
7.76625350e-02 -2.18637541e-01 -8.02563369e-01 6.28473282e-01
8.44950497e-01 -6.57299042e-01 -5.31711638e-01 -1.02086151e+00
2.04613239e-01 -4.25417691e-01 -4.06317152e-02 -8.17143679e-01
2.45523542e-01 5.19749001e-02 -2.10170206e-02 -5.12646437e-01
4.09785897e-01 -4.74131018e-01 -8.06679368e-01 2.50980139e-01
-6.22706115e-01 5.84414542e-01 1.37575902e-03 5.10632098e-01
-5.71572363e-01 -7.22564399e-01 5.54573119e-01 3.09488308e-02
-6.47964120e-01 2.02343360e-01 -3.29939961e-01 4.73750353e-01
5.19466817e-01 -6.25079721e-02 -3.55118692e-01 -3.32204044e-01
-4.69045281e-01 4.79800850e-01 -1.15205668e-01 8.76990795e-01
1.07875276e+00 -1.70704091e+00 -5.96732378e-01 2.54075885e-01
4.96277094e-01 -4.86459613e-01 1.54082105e-01 8.32969487e-01
-7.94470608e-02 6.24433279e-01 1.76599517e-01 -2.65622169e-01
-1.16279817e+00 5.57390392e-01 2.32919991e-01 -7.57674754e-01
5.97402640e-02 1.18693566e+00 1.71398353e-02 -1.09346306e+00
6.93887323e-02 -4.60583061e-01 -5.84253669e-01 1.86462536e-01
5.03531873e-01 2.15670794e-01 2.83513874e-01 -4.17759538e-01
-5.11033893e-01 8.02616179e-01 -3.43360066e-01 6.81789964e-02
1.31886494e+00 -1.61811128e-01 -2.69293636e-01 4.86970514e-01
1.65661776e+00 -3.71814892e-02 -3.94298434e-01 -6.56217396e-01
3.30086768e-01 -6.24692500e-01 -6.89359754e-02 -3.35951924e-01
-6.94447577e-01 1.36819220e+00 2.52939105e-01 7.28869587e-02
9.91020322e-01 -2.47471362e-01 1.25146592e+00 4.26856756e-01
2.23882064e-01 -1.28513920e+00 4.33229268e-01 1.12405753e+00
9.25041914e-01 -1.12075555e+00 -1.83484584e-01 -1.83446348e-01
-6.93389058e-01 1.13709676e+00 9.62517321e-01 -2.99663514e-01
9.48415101e-01 -2.33284131e-01 -2.63596829e-02 5.67031652e-02
-1.31818199e+00 -4.10152406e-01 4.54271764e-01 3.94746512e-01
8.42353880e-01 -1.89678013e-01 -7.78071880e-01 1.00958443e+00
2.60202587e-02 -2.34857291e-01 1.73875257e-01 1.03852654e+00
-7.48898029e-01 -1.68952739e+00 1.25451004e-02 6.14038706e-01
-4.26888652e-02 -5.08000016e-01 -3.82173240e-01 1.20935939e-01
1.14149921e-01 1.35227704e+00 9.48388278e-02 -1.02525997e+00
6.64044380e-01 5.17074525e-01 3.47124547e-01 -1.11170554e+00
-1.09242892e+00 -5.93621671e-01 2.53980964e-01 -3.92597705e-01
-2.05701455e-01 -3.38941813e-01 -1.13837671e+00 -1.80895180e-01
-6.50548935e-01 4.42006648e-01 3.49924684e-01 1.21353614e+00
6.61748827e-01 6.05097651e-01 8.68105054e-01 -4.66955483e-01
-9.26660419e-01 -1.24465179e+00 -2.29809344e-01 3.02591950e-01
9.68318060e-02 -6.07958078e-01 -3.18270445e-01 -5.49089193e-01] | [10.958894729614258, 8.546529769897461] |
43416a15-83aa-4427-9dd5-bedb88a1542d | adversarially-learned-one-class-classifier | 1802.09088 | null | http://arxiv.org/abs/1802.09088v2 | http://arxiv.org/pdf/1802.09088v2.pdf | Adversarially Learned One-Class Classifier for Novelty Detection | Novelty detection is the process of identifying the observation(s) that
differ in some respect from the training observations (the target class). In
reality, the novelty class is often absent during training, poorly sampled or
not well defined. Therefore, one-class classifiers can efficiently model such
problems. However, due to the unavailability of data from the novelty class,
training an end-to-end deep network is a cumbersome task. In this paper,
inspired by the success of generative adversarial networks for training deep
models in unsupervised and semi-supervised settings, we propose an end-to-end
architecture for one-class classification. Our architecture is composed of two
deep networks, each of which trained by competing with each other while
collaborating to understand the underlying concept in the target class, and
then classify the testing samples. One network works as the novelty detector,
while the other supports it by enhancing the inlier samples and distorting the
outliers. The intuition is that the separability of the enhanced inliers and
distorted outliers is much better than deciding on the original samples. The
proposed framework applies to different related applications of anomaly and
outlier detection in images and videos. The results on MNIST and Caltech-256
image datasets, along with the challenging UCSD Ped2 dataset for video anomaly
detection illustrate that our proposed method learns the target class
effectively and is superior to the baseline and state-of-the-art methods. | ['Mohammad Sabokrou', 'Mahmood Fathy', 'Ehsan Adeli', 'Mohammad Khalooei'] | 2018-02-25 | adversarially-learned-one-class-classifier-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Sabokrou_Adversarially_Learned_One-Class_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Sabokrou_Adversarially_Learned_One-Class_CVPR_2018_paper.pdf | cvpr-2018-6 | ['one-class-classifier'] | ['methodology'] | [ 1.60203815e-01 6.70837760e-02 1.60102576e-01 -3.20683926e-01
-5.31792521e-01 -2.65586823e-01 4.90419269e-01 1.20617807e-01
-2.73555428e-01 3.78977090e-01 -1.51794612e-01 1.64882224e-02
3.96642275e-02 -5.24824500e-01 -1.10549402e+00 -9.57060099e-01
-1.26681328e-01 5.66386640e-01 1.14169024e-01 -2.60928161e-02
2.18447238e-01 4.08254534e-01 -1.67792249e+00 4.57484394e-01
1.00560784e+00 1.35019052e+00 -3.74308616e-01 5.21132886e-01
1.44645140e-01 8.93397689e-01 -7.00732529e-01 -2.27428779e-01
5.77857912e-01 -6.49680436e-01 -2.27815196e-01 3.78304988e-01
7.42157996e-01 -4.38855022e-01 -5.45886040e-01 1.34254277e+00
3.78982782e-01 1.75144091e-01 7.38909066e-01 -1.63486660e+00
-6.70649111e-01 2.73905337e-01 -4.79767412e-01 3.10050815e-01
1.44729719e-01 1.19490013e-01 7.75340736e-01 -1.21584427e+00
4.22210991e-01 9.47710574e-01 6.95932209e-01 4.43191409e-01
-9.73282278e-01 -6.68211520e-01 2.43287340e-01 2.50178874e-01
-1.36832428e+00 -4.89931554e-01 1.00031793e+00 -5.80328941e-01
4.38768923e-01 1.48190230e-01 4.58732247e-01 1.46128142e+00
2.20266119e-01 1.02243900e+00 6.18029833e-01 -1.37100607e-01
3.78987521e-01 1.64729774e-01 -1.37506291e-01 3.77246976e-01
3.37950528e-01 3.37409198e-01 -4.36395079e-01 -3.22927833e-01
4.27454025e-01 6.15790606e-01 -4.39510673e-01 -6.73150420e-01
-9.88875806e-01 8.72979581e-01 5.24734497e-01 1.12952895e-01
-4.20091629e-01 2.06322074e-02 5.33955216e-01 7.02737033e-01
4.97811258e-01 3.62089604e-01 -2.56222934e-01 -4.17689420e-02
-1.21209919e+00 2.04897463e-01 6.47557676e-01 7.81509995e-01
3.89446437e-01 3.66077662e-01 -7.63878450e-02 5.32650590e-01
3.12116683e-01 1.87829852e-01 8.01592886e-01 -4.78294432e-01
2.55686522e-01 6.26410663e-01 1.44711673e-01 -1.00394201e+00
-1.81437269e-01 -7.83783734e-01 -9.75726724e-01 6.35914147e-01
6.32685483e-01 1.15409404e-01 -1.17692971e+00 1.60198259e+00
2.85804659e-01 6.23693109e-01 1.29913852e-01 1.03363764e+00
5.62661290e-01 5.55217922e-01 -3.14215153e-01 5.25271632e-02
7.49821424e-01 -1.03860545e+00 -4.22753632e-01 -2.84811735e-01
6.55483902e-01 -7.28704929e-01 8.00822139e-01 7.17958033e-01
-7.43669927e-01 -6.76059902e-01 -1.15395045e+00 2.14143828e-01
-3.60115588e-01 1.92357272e-01 1.70588762e-01 3.14487964e-01
-6.69730008e-01 7.28774786e-01 -8.96323979e-01 -2.22648621e-01
7.01360166e-01 2.18770251e-01 -5.15337527e-01 -1.04194224e-01
-7.87413836e-01 5.27150452e-01 3.92718077e-01 3.36527556e-01
-1.16583943e+00 -4.68865454e-01 -7.77306616e-01 4.16914150e-02
4.11418557e-01 -3.89307588e-01 8.40690076e-01 -1.65593660e+00
-1.02952635e+00 6.72146857e-01 3.31333399e-01 -7.77203262e-01
1.01532364e+00 -4.35079902e-01 -5.41839123e-01 4.81434679e-03
3.86451520e-02 3.11605155e-01 1.48302424e+00 -1.14433956e+00
-7.58293629e-01 -3.76205474e-01 -2.38933295e-01 -4.88222465e-02
-9.27854180e-02 -2.99893200e-01 -2.91349348e-02 -9.77422595e-01
5.64214528e-01 -8.76311481e-01 3.56301665e-02 2.78897464e-01
-5.70131063e-01 9.56653524e-03 1.29804265e+00 -5.01197159e-01
6.85194373e-01 -2.71827888e+00 -8.69762450e-02 4.42224324e-01
4.03365344e-01 2.86450773e-01 -1.70187019e-02 2.48348787e-01
-5.43575406e-01 -2.57183552e-01 -2.56742179e-01 -3.34368110e-01
-1.52701352e-04 2.54922628e-01 -5.96027136e-01 9.05907989e-01
4.26256269e-01 5.63799560e-01 -9.70335424e-01 2.05927361e-02
1.15676142e-01 2.66126096e-01 -5.42990327e-01 3.31666440e-01
4.28844476e-03 6.51446819e-01 -1.51120469e-01 8.57169569e-01
6.96885943e-01 6.57215267e-02 -4.53971237e-01 -3.17473197e-03
3.64614874e-01 -5.93153387e-02 -1.55350518e+00 1.45782602e+00
1.34543017e-01 7.15962589e-01 -1.68580353e-01 -1.27567029e+00
8.69205594e-01 3.45501781e-01 3.59266549e-01 -3.34946036e-01
1.19014926e-01 5.50103366e-01 3.75264078e-01 -5.95261037e-01
2.59623706e-01 -4.00861539e-02 2.50660509e-01 2.35486552e-01
3.07963043e-01 3.18675935e-01 7.56264627e-02 5.33032119e-02
1.16394961e+00 6.96010217e-02 2.06471458e-01 2.33851615e-02
3.29965830e-01 -3.72699648e-01 8.89168799e-01 9.57460403e-01
-4.10748303e-01 9.09781277e-01 6.08839571e-01 -6.98691964e-01
-1.09568548e+00 -1.32634723e+00 -1.10189281e-01 6.44479871e-01
1.40028760e-01 8.92002136e-02 -3.80430460e-01 -1.07280755e+00
2.31487513e-01 6.68028951e-01 -8.06458473e-01 -5.60536027e-01
-3.47252458e-01 -3.28747064e-01 6.16266906e-01 5.36617637e-01
4.63227093e-01 -1.00386715e+00 -4.42291141e-01 1.34697571e-01
5.55350743e-02 -9.40230191e-01 -3.32403183e-01 1.31921932e-01
-9.15491164e-01 -1.32535756e+00 -4.78378922e-01 -5.96950054e-01
1.04657614e+00 6.46570027e-02 1.00948763e+00 3.20315808e-01
-3.63596939e-02 7.11023435e-02 -4.55794334e-01 -6.35591209e-01
-4.93692398e-01 -2.92001992e-01 2.82883137e-01 5.01052678e-01
4.79928434e-01 -6.74003959e-01 -5.61963201e-01 3.37249875e-01
-1.26210868e+00 -5.10779977e-01 6.49977982e-01 1.06866980e+00
5.73751748e-01 3.66561323e-01 5.96285462e-01 -1.05124176e+00
8.59080479e-02 -6.42978191e-01 -3.86485279e-01 -1.45152286e-01
-3.73280853e-01 -1.32265389e-01 8.68046224e-01 -6.34809732e-01
-4.97215539e-01 1.97400302e-02 -1.40820846e-01 -9.65556204e-01
-4.35433805e-01 3.40236098e-01 -2.25360900e-01 -4.14608940e-02
7.34930813e-01 2.88760334e-01 2.21976321e-02 -3.44971895e-01
-6.32774234e-02 3.39705646e-01 7.26236284e-01 -3.01587850e-01
1.09128582e+00 6.58052504e-01 -1.10571161e-01 -8.37625921e-01
-7.94360459e-01 -4.44000572e-01 -5.86958706e-01 -8.15748796e-02
4.27301705e-01 -9.67081845e-01 9.25289169e-02 7.33533978e-01
-8.96722436e-01 8.99450481e-02 -8.70679438e-01 5.86908817e-01
-4.38182980e-01 4.28490371e-01 -2.22981259e-01 -6.78474307e-01
-1.60076246e-01 -1.21807480e+00 9.06107068e-01 1.75628588e-01
-1.38314310e-02 -7.73846686e-01 -8.35137740e-02 7.39212558e-02
2.02702463e-01 6.30192757e-01 7.21702814e-01 -1.49864531e+00
-5.71035981e-01 -8.92427206e-01 2.02486143e-01 7.62002587e-01
5.19252801e-03 2.18725592e-01 -1.10152674e+00 -6.01627529e-01
3.21868002e-01 -2.46365726e-01 9.74439085e-01 2.06764609e-01
1.20655704e+00 -3.32045048e-01 -1.62356898e-01 7.81289339e-01
1.30016685e+00 1.39225349e-01 8.89810205e-01 4.28099900e-01
7.12537825e-01 2.75392890e-01 5.57960331e-01 3.18197787e-01
-5.27152494e-02 3.34764928e-01 8.75506461e-01 -2.14942709e-01
2.22974524e-01 -2.07333028e-01 5.63961029e-01 3.58030319e-01
3.72180283e-01 -3.81401122e-01 -7.32712626e-01 6.20856166e-01
-1.98287606e+00 -1.00414002e+00 -5.35052549e-03 2.38421750e+00
3.66578817e-01 4.90547031e-01 1.26231328e-01 4.30672646e-01
6.63888454e-01 5.51169626e-02 -8.47706139e-01 -1.19055204e-01
-2.89428085e-01 -2.59150006e-02 1.32932544e-01 -6.85611516e-02
-1.34574127e+00 5.34174860e-01 4.85955906e+00 6.75829351e-01
-1.27311659e+00 -4.34742384e-02 6.38241172e-01 -2.60511637e-01
2.95234829e-01 2.67440714e-02 -3.25240552e-01 6.62879884e-01
6.72399879e-01 2.06794456e-01 1.70044675e-01 1.12602282e+00
2.50600628e-03 1.97803155e-02 -1.61562800e+00 9.56232131e-01
3.45473021e-01 -9.15030837e-01 2.11251810e-01 -5.66918589e-02
8.40997696e-01 1.62862122e-01 1.09688699e-01 4.38989639e-01
-2.55283803e-01 -9.06287789e-01 9.59208310e-01 6.44907653e-01
2.20249355e-01 -8.19618583e-01 1.12690747e+00 5.11613071e-01
-6.56645060e-01 -3.22211623e-01 -5.27876318e-01 -1.77835149e-03
-2.45827615e-01 8.18573356e-01 -6.45488322e-01 4.50571299e-01
8.39917123e-01 8.26259136e-01 -6.86961830e-01 1.43089449e+00
-2.76519001e-01 6.05043530e-01 -3.49842876e-01 6.02542818e-01
3.21374834e-01 -2.13673458e-01 1.00365341e+00 8.10074985e-01
4.52889740e-01 -3.03773552e-01 3.07846367e-01 9.20708716e-01
-2.77301520e-01 -2.13194758e-01 -8.54197204e-01 -1.16982691e-01
1.73851490e-01 1.05744946e+00 -7.09767640e-01 -3.89641970e-01
-4.47411060e-01 1.08235180e+00 1.58714622e-01 3.96161050e-01
-8.95630717e-01 -6.06115699e-01 6.98674083e-01 1.55993789e-01
5.75479209e-01 2.98504792e-02 -1.73366308e-01 -1.38559651e+00
4.08426642e-01 -1.21744299e+00 4.02744055e-01 -5.80837429e-01
-1.56254399e+00 6.28415287e-01 -4.78836745e-01 -1.83050585e+00
-2.17396468e-01 -6.69020951e-01 -1.00086701e+00 4.71684068e-01
-1.28852689e+00 -8.88318598e-01 -4.55966622e-01 4.96728212e-01
4.87131804e-01 -4.53832209e-01 4.73971456e-01 4.54662412e-01
-5.24779558e-01 8.80201638e-01 5.35646558e-01 5.10367990e-01
8.45314384e-01 -1.24196517e+00 2.11579025e-01 1.42448735e+00
3.47366810e-01 3.65371644e-01 8.09841990e-01 -5.79290688e-01
-9.89222884e-01 -1.35704434e+00 4.69039708e-01 -3.64034146e-01
6.10800683e-01 -3.13716322e-01 -1.26772130e+00 7.84192085e-01
-1.71851560e-01 4.57765460e-01 4.67661470e-01 -3.29134554e-01
-3.79933685e-01 -1.42344639e-01 -1.30548680e+00 4.74179238e-01
8.69648874e-01 -4.07667726e-01 -7.01132357e-01 4.35076416e-01
3.71266633e-01 -6.23752773e-01 -4.24812734e-01 3.88731539e-01
3.05804878e-01 -1.18723142e+00 8.88569295e-01 -7.95493782e-01
5.84414065e-01 -4.46807444e-01 -2.03732252e-01 -1.41363657e+00
1.77248586e-02 -2.81526327e-01 -4.64738995e-01 1.07483220e+00
1.57091051e-01 -7.69326389e-01 7.98950613e-01 2.95222521e-01
-4.56810415e-01 -7.84498751e-01 -1.13354385e+00 -9.70110536e-01
-1.82537317e-01 -3.36740792e-01 4.96362776e-01 1.11342442e+00
-6.28886878e-01 -2.06007473e-02 -4.58834946e-01 6.08893275e-01
6.19220257e-01 -2.23185997e-02 9.92018163e-01 -1.33912051e+00
-2.69985706e-01 -2.48604134e-01 -1.08901763e+00 -8.02568316e-01
1.24344990e-01 -8.07320535e-01 1.11634985e-01 -9.34045613e-01
-1.90183893e-01 -2.51593232e-01 -4.64152277e-01 3.57421756e-01
-1.40406832e-01 2.51436859e-01 -8.46592803e-03 3.42191458e-01
-5.42749882e-01 7.81261504e-01 7.16968298e-01 -3.26147676e-01
-1.43565238e-01 2.68455237e-01 -4.95855033e-01 1.04900634e+00
6.33568347e-01 -8.27434778e-01 -3.26781809e-01 -3.64787519e-01
1.66023880e-01 -4.11109567e-01 6.96905255e-01 -1.44016469e+00
2.45763004e-01 2.50660330e-01 8.01941335e-01 -7.25031018e-01
1.58548102e-01 -1.25823450e+00 -1.07292317e-01 4.81805623e-01
-2.19279408e-01 6.44124895e-02 -5.94006777e-02 1.04060066e+00
-4.32837546e-01 -3.91665399e-01 9.40580904e-01 3.04841343e-02
-5.98150313e-01 5.25284886e-01 -1.82006836e-01 3.22156429e-01
1.15501809e+00 -4.55065399e-01 -1.79625168e-01 -4.60197926e-01
-7.81985343e-01 8.25640038e-02 6.47041321e-01 4.78428543e-01
8.63146365e-01 -1.47458827e+00 -7.48267114e-01 7.70250797e-01
3.61482501e-01 2.41701692e-01 1.43318564e-01 9.69431579e-01
-4.49932754e-01 -2.76147604e-01 -1.60049036e-01 -9.38880444e-01
-8.61253619e-01 5.94419479e-01 5.39963484e-01 -2.70893462e-02
-6.78334713e-01 8.31149280e-01 2.31383517e-01 -3.34357917e-01
3.90073657e-01 -4.44908679e-01 8.14645737e-02 -1.49326131e-01
5.08259833e-01 3.51332664e-01 2.13329732e-01 -6.29959345e-01
-2.52299935e-01 6.17304184e-02 -2.74648577e-01 4.30153787e-01
1.31871104e+00 1.32876962e-01 -5.98745272e-02 5.57253718e-01
1.21337867e+00 -9.29077193e-02 -1.30432105e+00 -4.55601186e-01
1.11010246e-01 -7.73558497e-01 -1.48356080e-01 -4.07681286e-01
-1.27357352e+00 7.53911316e-01 9.25056100e-01 2.25688756e-01
1.16886389e+00 -3.41613889e-01 6.35344565e-01 5.37115753e-01
2.84884889e-02 -1.13362420e+00 2.19484165e-01 4.32249397e-01
8.99062395e-01 -1.65658021e+00 -1.41832292e-01 -1.87148061e-02
-5.11525035e-01 1.17422378e+00 8.07128787e-01 -6.45106912e-01
6.23277366e-01 -1.84500143e-01 8.36737305e-02 -2.49372438e-01
-4.95268255e-01 2.05091774e-01 4.62305903e-01 4.91222888e-01
8.09689462e-02 -2.55095005e-01 1.53921515e-01 5.54526389e-01
-4.61763926e-02 -2.30385512e-01 5.46560943e-01 9.69610274e-01
-3.30677330e-01 -6.36839271e-01 -5.00773489e-01 8.23909938e-01
-4.63166922e-01 5.85596822e-02 -2.77308613e-01 8.63331676e-01
3.96978706e-01 7.61928678e-01 3.27771455e-01 -3.88418257e-01
4.14285451e-01 1.76706463e-01 2.89989752e-03 -5.33463657e-01
-4.12850022e-01 -3.01964041e-02 -3.18947375e-01 -6.59623146e-01
-1.13107376e-01 -6.28674090e-01 -1.11322665e+00 -6.49184287e-02
-4.47581738e-01 1.54469818e-01 4.17152643e-01 1.08162057e+00
3.69125307e-01 5.21956682e-01 8.92103076e-01 -8.56391013e-01
-8.35896850e-01 -9.39139187e-01 -7.33622372e-01 8.14607024e-01
7.62895167e-01 -8.28686118e-01 -8.42390835e-01 -8.15838948e-02] | [7.685258388519287, 2.2852025032043457] |
3d1c7394-94eb-4d51-a22d-200c6b1478a9 | absformer-transformer-based-model-for | 2306.04787 | null | https://arxiv.org/abs/2306.04787v1 | https://arxiv.org/pdf/2306.04787v1.pdf | Absformer: Transformer-based Model for Unsupervised Multi-Document Abstractive Summarization | Multi-document summarization (MDS) refers to the task of summarizing the text in multiple documents into a concise summary. The generated summary can save the time of reading many documents by providing the important content in the form of a few sentences. Abstractive MDS aims to generate a coherent and fluent summary for multiple documents using natural language generation techniques. In this paper, we consider the unsupervised abstractive MDS setting where there are only documents with no groundtruh summaries provided, and we propose Absformer, a new Transformer-based method for unsupervised abstractive summary generation. Our method consists of a first step where we pretrain a Transformer-based encoder using the masked language modeling (MLM) objective as the pretraining task in order to cluster the documents into semantically similar groups; and a second step where we train a Transformer-based decoder to generate abstractive summaries for the clusters of documents. To our knowledge, we are the first to successfully incorporate a Transformer-based model to solve the unsupervised abstractive MDS task. We evaluate our approach using three real-world datasets from different domains, and we demonstrate both substantial improvements in terms of evaluation metrics over state-of-the-art abstractive-based methods, and generalization to datasets from different domains. | ['Huseyin Uzunalioglu', 'Mohamed Trabelsi'] | 2023-06-07 | null | null | null | null | ['abstractive-text-summarization', 'multi-document-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 5.86611211e-01 4.75104511e-01 6.39488548e-02 -4.14073110e-01
-1.56417894e+00 -5.14806449e-01 1.01204693e+00 5.65538585e-01
-1.77206486e-01 8.67597938e-01 9.33647752e-01 2.39006169e-02
1.88392609e-01 -4.31379884e-01 -7.05902100e-01 -4.57535774e-01
3.03715676e-01 9.35576439e-01 1.19343802e-01 -1.42022058e-01
5.00576675e-01 5.41434661e-02 -1.44176161e+00 6.23170614e-01
1.40518284e+00 5.93437910e-01 4.45368916e-01 8.11888218e-01
-3.49486232e-01 9.88510609e-01 -1.30005002e+00 -4.76624429e-01
-2.05132470e-01 -9.36139882e-01 -9.58022833e-01 4.55163151e-01
5.51863909e-01 -4.79962409e-01 -7.08767101e-02 8.20827365e-01
5.65159678e-01 3.52289349e-01 8.50052953e-01 -9.94037867e-01
-7.92070508e-01 1.18797326e+00 -4.49679911e-01 -1.91693399e-02
4.90291476e-01 -2.96118915e-01 1.16069019e+00 -9.73144770e-01
5.64070284e-01 1.43495870e+00 3.02519575e-02 7.21403360e-01
-1.16117060e+00 -1.95441708e-01 2.77147442e-01 -1.92084685e-01
-9.87070262e-01 -7.01754987e-01 6.65201783e-01 -1.74614504e-01
1.06566215e+00 3.34588468e-01 1.98240519e-01 9.86297250e-01
2.48698115e-01 1.13219023e+00 4.37252581e-01 -5.51322699e-01
4.54860091e-01 -5.47243953e-02 2.51958460e-01 5.56469738e-01
3.60113531e-01 -7.25449204e-01 -7.26112962e-01 -1.73606321e-01
7.71685988e-02 -1.91098571e-01 -6.49964213e-02 1.90218985e-01
-1.47806382e+00 9.51627076e-01 2.26456746e-01 3.05376917e-01
-5.47821641e-01 1.61221072e-01 5.13015747e-01 6.16260692e-02
8.77320409e-01 6.15827382e-01 3.73710655e-02 1.27727866e-01
-1.48716271e+00 5.63161194e-01 8.43229234e-01 1.07784450e+00
4.44848239e-01 6.17144369e-02 -6.49572134e-01 7.45814443e-01
1.22987971e-01 4.03214723e-01 7.89863586e-01 -9.98434782e-01
9.92811859e-01 5.94920099e-01 1.38421044e-01 -7.16461837e-01
-1.32422164e-01 -3.77827376e-01 -8.92771304e-01 -3.50392967e-01
-2.24083722e-01 -2.09116921e-01 -8.59651327e-01 1.67286265e+00
-9.73529369e-02 -7.40722641e-02 5.71411490e-01 5.95131874e-01
1.10461915e+00 1.16130841e+00 -2.58905321e-01 -4.09276843e-01
1.15198696e+00 -1.33010089e+00 -7.77686834e-01 -4.16327745e-01
5.24407625e-01 -6.34818137e-01 1.09622431e+00 3.37182194e-01
-1.56968713e+00 -3.86675209e-01 -1.33510399e+00 -4.00864124e-01
5.13352780e-03 5.13150156e-01 8.02996531e-02 8.98719579e-02
-1.19716036e+00 5.39810538e-01 -8.98686707e-01 -5.12508452e-01
2.83912539e-01 1.19905986e-01 -2.10172325e-01 4.58638230e-03
-8.77381206e-01 7.82767892e-01 6.06169939e-01 -2.96265125e-01
-9.87465739e-01 -5.42859435e-01 -1.03942943e+00 2.91117877e-01
3.29329908e-01 -1.21316111e+00 1.64623630e+00 -7.06540823e-01
-1.35974705e+00 6.12151504e-01 -6.19776726e-01 -5.92773438e-01
2.37185806e-01 -2.56825149e-01 -5.05802222e-02 4.71051633e-01
6.52326941e-01 7.29504526e-01 6.02549136e-01 -1.37439716e+00
-7.00842261e-01 -3.05581659e-01 -1.24185674e-01 4.76229399e-01
-4.26170588e-01 1.30716205e-01 -3.61890405e-01 -9.16030049e-01
-3.82396840e-02 -6.13925219e-01 -1.98542327e-01 -6.03535950e-01
-9.27777231e-01 -4.06702369e-01 6.58634305e-01 -8.63324881e-01
1.33418536e+00 -1.76370084e+00 6.15338743e-01 -4.41050828e-01
2.87598640e-01 1.27620935e-01 -2.99573511e-01 8.03353131e-01
2.43433103e-01 1.28827050e-01 -6.94228470e-01 -1.15454936e+00
2.60488719e-01 6.72000349e-02 -8.13201547e-01 -1.51808977e-01
3.56120408e-01 9.00266349e-01 -1.01181352e+00 -6.50107384e-01
-1.24367759e-01 1.11978248e-01 -4.25042301e-01 4.28233206e-01
-6.42043352e-01 2.65065908e-01 -4.49259996e-01 1.38163209e-01
3.78962785e-01 -2.78022856e-01 -5.11917323e-02 4.83288094e-02
-1.10633358e-01 7.33407497e-01 -8.32479060e-01 2.12299109e+00
-4.13721770e-01 6.52947843e-01 -3.09922487e-01 -1.11506391e+00
8.22019517e-01 3.91893595e-01 1.38839573e-01 -4.42237020e-01
-4.35933806e-02 3.82916927e-01 -5.00290573e-01 -1.90139204e-01
1.19804358e+00 -2.42387310e-01 -6.48601711e-01 9.37243521e-01
1.52867362e-01 -5.86004794e-01 7.09847689e-01 8.99201632e-01
1.03253484e+00 -3.16649914e-01 2.72630841e-01 -1.26543581e-01
4.14433360e-01 2.97535639e-02 1.94786221e-01 8.88070941e-01
5.95364749e-01 9.94038939e-01 6.42679095e-01 1.69702172e-01
-9.89425361e-01 -1.09003365e+00 5.16248822e-01 8.35183024e-01
6.51433272e-03 -7.25202262e-01 -9.44833636e-01 -7.50234723e-01
-3.41086000e-01 1.37552750e+00 -3.96425217e-01 -3.04702282e-01
-6.00737751e-01 -4.53491688e-01 5.72538972e-01 5.48061371e-01
5.48832357e-01 -9.92988050e-01 -5.83824158e-01 2.94825882e-01
-6.25347674e-01 -1.09295821e+00 -8.48733127e-01 1.01593263e-01
-8.19408476e-01 -5.31818569e-01 -8.85704339e-01 -9.06669140e-01
6.64519012e-01 2.61075050e-01 1.16979492e+00 -7.67629743e-02
2.03694060e-01 1.50580585e-01 -3.45867544e-01 -4.95298237e-01
-8.59684408e-01 3.54641199e-01 -1.10667482e-01 -1.27834111e-01
-4.81466576e-02 -4.06619668e-01 -2.13426352e-01 -3.45720768e-01
-1.29555988e+00 4.65520412e-01 4.84601825e-01 6.58938706e-01
4.93667036e-01 1.72934726e-01 9.26787555e-01 -8.40370893e-01
1.27397907e+00 -3.04718614e-01 -7.02172294e-02 4.62870091e-01
-1.98478132e-01 5.27625442e-01 9.20254290e-01 -8.59206691e-02
-1.21620023e+00 -3.11865062e-01 -3.88806351e-02 -3.68524306e-02
1.58378482e-01 7.65278816e-01 -2.37590522e-01 8.70396674e-01
6.24426603e-01 6.63852751e-01 -1.72420904e-01 -4.67978776e-01
4.42015320e-01 9.00670528e-01 8.66559744e-01 -4.86776143e-01
5.46609104e-01 3.00479561e-01 -2.84943849e-01 -7.93313384e-01
-1.20837569e+00 -2.72254288e-01 -5.97485006e-01 1.67708933e-01
8.63378108e-01 -1.02189612e+00 3.98952626e-02 3.17844272e-01
-1.54503548e+00 -1.36428460e-01 -5.85608602e-01 1.48679450e-01
-7.62223601e-01 6.35162354e-01 -3.62105757e-01 -5.02485991e-01
-9.49780166e-01 -1.02778363e+00 1.56462407e+00 2.23451972e-01
-4.61236089e-01 -8.67529273e-01 8.42890590e-02 2.61509538e-01
9.07879397e-02 3.18803430e-01 9.24550712e-01 -9.63965774e-01
-2.81646639e-01 -9.96682793e-02 2.90819146e-02 4.14888114e-01
4.13048625e-01 -1.19480789e-01 -6.30262613e-01 -2.67091125e-01
9.03187096e-02 -3.75945061e-01 1.29702306e+00 3.34841937e-01
8.61829996e-01 -8.25186431e-01 -2.78167963e-01 9.40018594e-02
1.08793902e+00 1.22734748e-01 4.69630361e-01 6.04403652e-02
5.54554522e-01 5.72976232e-01 7.06805468e-01 5.22436857e-01
7.62995601e-01 5.30314803e-01 2.84376964e-02 5.30988276e-02
-3.25417370e-01 -4.70342785e-01 5.22217393e-01 1.10974216e+00
4.61858600e-01 -1.00578737e+00 -5.82583725e-01 7.88186073e-01
-2.05249882e+00 -1.15572727e+00 1.67003140e-01 1.99951565e+00
9.70379412e-01 9.52754691e-02 6.76144660e-02 1.15172844e-02
7.86095738e-01 2.89955050e-01 -4.94807214e-01 -5.48990309e-01
-1.16534695e-01 8.80724788e-02 -2.07697555e-01 5.99390566e-01
-9.97866750e-01 9.34092104e-01 5.61062384e+00 5.37724555e-01
-9.12576199e-01 -9.90165100e-02 4.05988216e-01 -3.46239418e-01
-6.66201413e-01 1.22371176e-02 -7.21145689e-01 5.42818129e-01
1.06805980e+00 -7.65656233e-01 -1.25768222e-02 6.01363182e-01
4.16492373e-01 -1.78274423e-01 -1.29102457e+00 5.90110421e-01
6.49333537e-01 -1.53417826e+00 7.03641653e-01 -2.44723842e-01
9.32783902e-01 -2.20527381e-01 -8.82670954e-02 1.62842423e-01
3.88644010e-01 -7.37684727e-01 1.09854007e+00 4.48535353e-01
6.28218174e-01 -8.78591239e-01 6.37319267e-01 6.38060153e-01
-9.05247748e-01 1.85029164e-01 -3.53162795e-01 1.31205991e-01
4.98362660e-01 6.71992898e-01 -9.23058093e-01 9.15661037e-01
2.37964362e-01 9.81825233e-01 -5.89975715e-01 7.09127963e-01
-3.23044151e-01 3.69151533e-01 8.69918913e-02 -1.39965802e-01
2.76773065e-01 -2.78680883e-02 7.37489581e-01 1.34798288e+00
6.29223228e-01 1.07899802e-02 2.57776529e-01 8.81458044e-01
-2.62398958e-01 -5.56029230e-02 -4.08459723e-01 -3.80430818e-01
5.31718791e-01 1.02337205e+00 -6.57815516e-01 -9.18508887e-01
-7.25660771e-02 1.39720893e+00 1.80705011e-01 1.68946519e-01
-6.19752049e-01 -7.36913204e-01 1.03975095e-01 -7.12435469e-02
3.33710253e-01 -1.26344919e-01 -3.74449968e-01 -1.48900521e+00
2.74732739e-01 -1.00081682e+00 3.55405122e-01 -9.86894429e-01
-1.06766546e+00 8.64794910e-01 2.16891274e-01 -1.09593701e+00
-7.91834772e-01 1.57591537e-01 -8.75494063e-01 8.49949479e-01
-1.30276799e+00 -1.15892160e+00 -1.62911385e-01 1.23993680e-01
1.12494195e+00 -3.43857646e-01 7.93393552e-01 -1.53177917e-01
-5.61446786e-01 3.12048703e-01 2.14694321e-01 -2.81388313e-01
7.59122014e-01 -1.63778794e+00 8.03431511e-01 1.27098310e+00
-9.78937373e-03 6.40288889e-01 1.06090891e+00 -7.40685880e-01
-1.00878060e+00 -1.34243846e+00 1.47348332e+00 -3.13370198e-01
3.36781889e-01 -4.58623797e-01 -6.35231793e-01 7.82349706e-01
7.39036381e-01 -7.72116780e-01 6.47134900e-01 -4.44639415e-01
4.96742614e-02 1.29852548e-01 -9.20327902e-01 7.21235096e-01
7.54135728e-01 -1.69657871e-01 -1.11024058e+00 4.10131782e-01
1.25643992e+00 -3.40066135e-01 -4.64027286e-01 1.99089739e-02
4.27602120e-02 -7.51396000e-01 6.07460499e-01 -3.71761352e-01
1.15461636e+00 -3.02410007e-01 -1.54208019e-01 -1.77503145e+00
7.42588937e-02 -6.80745542e-01 -2.32797101e-01 1.69104850e+00
4.44613934e-01 -4.11777377e-01 4.40153807e-01 3.52086395e-01
-7.76902080e-01 -4.88368273e-01 -6.22605920e-01 -6.12712920e-01
1.43568322e-01 7.31130913e-02 7.55620897e-01 3.41770709e-01
2.18770489e-01 8.86561811e-01 -2.24067956e-01 -5.65854013e-02
6.49769604e-01 4.57391232e-01 8.77333701e-01 -1.09281147e+00
-1.82412460e-01 -4.06845391e-01 8.32608268e-02 -1.25671530e+00
4.11781788e-01 -1.00460863e+00 3.69671106e-01 -2.46537304e+00
6.66774094e-01 1.96618423e-01 3.83711725e-01 2.61739701e-01
-3.95233780e-01 -2.13071913e-01 2.63219744e-01 2.02530622e-01
-9.28200960e-01 8.73033464e-01 1.07400763e+00 -4.49450165e-01
-3.34587812e-01 4.85342629e-02 -1.45425081e+00 4.39729035e-01
6.22654259e-01 -5.23026109e-01 -7.87406385e-01 -7.06021428e-01
-7.82389343e-02 2.18140110e-01 1.56293899e-01 -9.39842582e-01
4.30231094e-01 3.48780081e-02 1.13472193e-01 -1.13585913e+00
2.56174117e-01 -1.21433407e-01 -1.85315102e-01 3.43442887e-01
-8.63313496e-01 1.33873835e-01 1.88992828e-01 3.67313623e-01
-4.50262159e-01 -3.73442560e-01 4.45398778e-01 -2.24967256e-01
-2.39181638e-01 -5.92412287e-03 -5.41222632e-01 3.31230760e-01
6.82159424e-01 -9.88220498e-02 -5.10619879e-01 -5.79925179e-01
-3.81726623e-01 3.78507107e-01 4.22296286e-01 3.33275378e-01
8.30114067e-01 -1.24491441e+00 -1.26844764e+00 -1.31988853e-01
4.64111120e-02 3.53330314e-01 7.90095553e-02 4.11217541e-01
-3.92537564e-01 7.19983637e-01 2.17300162e-01 -4.03309286e-01
-1.20812547e+00 3.95987809e-01 -8.40877667e-02 -6.42169237e-01
-7.28633821e-01 6.20439708e-01 4.28148687e-01 -2.92176694e-01
1.31669208e-01 -4.76442516e-01 -1.20706089e-01 2.39461407e-01
8.46914411e-01 3.67598444e-01 3.02572161e-01 -5.04478931e-01
-2.24788651e-01 2.38580897e-01 -3.93417418e-01 -5.66599071e-01
1.40094531e+00 -2.99373358e-01 -3.31076175e-01 4.79623824e-01
1.20031762e+00 -1.19738921e-03 -9.76380229e-01 -2.31734842e-01
1.93405792e-01 -3.26098278e-02 -1.49030358e-01 -7.87433088e-01
-5.39570034e-01 7.79118598e-01 -3.57292593e-01 2.02071458e-01
1.27699208e+00 2.92797029e-01 1.03589475e+00 4.41399336e-01
1.47330230e-02 -9.89399016e-01 6.38907433e-01 5.59288621e-01
1.31846571e+00 -9.30014849e-01 -7.89821334e-03 -5.51200211e-02
-1.03517401e+00 1.00858200e+00 3.77993554e-01 -5.09964153e-02
-7.78694153e-02 1.44495741e-01 -2.50798255e-01 -2.42263854e-01
-1.06034732e+00 3.99134718e-02 3.38982552e-01 3.76396298e-01
3.99843156e-01 -1.17590785e-01 -3.28059971e-01 9.45133686e-01
-5.81717432e-01 -1.16848119e-01 1.00107086e+00 1.00556326e+00
-7.90426850e-01 -9.90903974e-01 -1.75848886e-01 5.41172683e-01
-3.94332409e-01 -1.46049201e-01 -8.10456276e-01 3.04497600e-01
-5.29433489e-01 1.31398427e+00 2.00374231e-01 -1.55991599e-01
3.79174978e-01 3.57937142e-02 4.13943529e-01 -1.26890278e+00
-5.37742257e-01 1.76435597e-02 2.16506138e-01 -9.60508548e-03
-3.35762113e-01 -7.72216678e-01 -1.54029679e+00 -2.53535867e-01
-1.48858026e-01 5.25788128e-01 5.27112663e-01 1.16154349e+00
4.88197505e-01 7.61334240e-01 6.68581605e-01 -9.47634757e-01
-6.22464240e-01 -1.01005816e+00 -4.60519820e-01 1.71556428e-01
5.37755489e-01 -4.93371300e-02 -2.86342680e-01 3.00305009e-01] | [12.516444206237793, 9.347265243530273] |
d2b103b4-1c9b-49ac-b15d-28d38047f780 | unsupervised-keyphrase-extraction-via-1 | 2203.07640 | null | https://arxiv.org/abs/2203.07640v2 | https://arxiv.org/pdf/2203.07640v2.pdf | Unsupervised Keyphrase Extraction via Interpretable Neural Networks | Keyphrase extraction aims at automatically extracting a list of "important" phrases representing the key concepts in a document. Prior approaches for unsupervised keyphrase extraction resorted to heuristic notions of phrase importance via embedding clustering or graph centrality, requiring extensive domain expertise. Our work presents a simple alternative approach which defines keyphrases as document phrases that are salient for predicting the topic of the document. To this end, we propose INSPECT -- an approach that uses self-explaining models for identifying influential keyphrases in a document by measuring the predictive impact of input phrases on the downstream task of the document topic classification. We show that this novel method not only alleviates the need for ad-hoc heuristics but also achieves state-of-the-art results in unsupervised keyphrase extraction in four datasets across two domains: scientific publications and news articles. | ['Yulia Tsvetkov', 'Svitlana Volkova', 'Maria Glenski', 'Emily Saldanha', 'Vidhisha Balachandran', 'Rishabh Joshi'] | 2022-03-15 | null | null | null | null | ['keyphrase-extraction'] | ['natural-language-processing'] | [ 1.52304634e-01 6.07944608e-01 -6.69775426e-01 2.71211416e-01
-7.54630864e-01 -7.89241731e-01 1.16302884e+00 1.23286033e+00
-4.06661153e-01 7.33708799e-01 9.66615617e-01 -5.40839374e-01
-3.57976228e-01 -8.60855043e-01 -3.73771876e-01 -6.01297975e-01
-3.59828949e-01 5.99440515e-01 1.71725407e-01 -4.22143377e-02
1.06161761e+00 2.95749813e-01 -1.29611063e+00 3.23840261e-01
6.11752748e-01 4.66085255e-01 -1.25987241e-02 7.93582559e-01
-7.24645197e-01 9.50002074e-01 -6.75859988e-01 -3.06617081e-01
-2.37871453e-01 -3.11625212e-01 -9.87860143e-01 -9.51532796e-02
1.06616527e-01 1.86961606e-01 -2.93223083e-01 9.58281159e-01
2.15559863e-02 -8.34381729e-02 9.23230588e-01 -1.24364913e+00
-2.62688607e-01 1.29469073e+00 -7.64127553e-01 7.65786767e-01
3.74283791e-01 -4.59221452e-01 2.03498673e+00 -9.47919130e-01
9.41550016e-01 1.06854987e+00 3.52674872e-01 -2.11089790e-01
-1.36673927e+00 -3.81777465e-01 1.57532126e-01 -8.26986134e-02
-1.14185131e+00 -6.44304678e-02 1.10035312e+00 -6.33026421e-01
9.90405798e-01 3.37107688e-01 8.80582273e-01 8.76766026e-01
4.07064795e-01 8.73093784e-01 9.08217251e-01 -5.28209448e-01
2.26044223e-01 2.80304819e-01 7.57844388e-01 4.69202042e-01
7.34711111e-01 -5.83962023e-01 -9.76021171e-01 -7.85401940e-01
2.12407634e-01 -1.63217843e-01 -1.48495391e-01 -5.95185533e-02
-1.53646100e+00 1.17849803e+00 -1.21436946e-01 6.01976216e-01
-8.36236358e-01 -9.13765877e-02 4.95006561e-01 2.33516455e-01
1.02073097e+00 1.19704902e+00 -6.28618181e-01 -2.38184214e-01
-1.18832386e+00 5.55651128e-01 1.16207623e+00 7.82722116e-01
5.64344049e-01 -6.88056707e-01 -4.37393785e-01 3.01978737e-01
2.43808612e-01 -4.75382060e-02 2.03016326e-01 -4.21341211e-01
5.64121544e-01 9.78902280e-01 1.65964469e-01 -1.46201062e+00
-3.57817858e-01 -5.05200744e-01 -4.66325313e-01 -6.53398752e-01
-1.30479351e-01 -4.23767231e-02 -7.80541837e-01 1.09317160e+00
2.60939032e-01 -2.63990581e-01 -8.85203108e-02 2.68018842e-01
8.77099812e-01 9.20428872e-01 2.35258237e-01 -4.78993893e-01
1.69430745e+00 -7.60774672e-01 -7.70722389e-01 1.14366062e-01
3.59482795e-01 -1.01326931e+00 8.53002369e-01 4.38463479e-01
-8.44619751e-01 6.66283444e-02 -8.61403048e-01 -1.24713749e-01
-6.65917933e-01 -1.09995298e-01 6.78961396e-01 6.39919564e-02
-8.82207334e-01 6.01085603e-01 -3.09005827e-01 -2.55722523e-01
3.99421513e-01 -8.64651427e-03 -6.44114688e-02 2.56699592e-01
-1.10654819e+00 7.07057059e-01 6.64943516e-01 -9.00797248e-01
-6.10587060e-01 -1.23562706e+00 -3.78069848e-01 3.72300833e-01
7.40663350e-01 -5.68740189e-01 1.04391384e+00 -2.32643306e-01
-7.63496041e-01 8.60751569e-01 -3.79236996e-01 -5.26601613e-01
-7.88458809e-03 -5.28565824e-01 -2.56574899e-01 5.14947176e-01
6.04180872e-01 3.31100494e-01 1.10878217e+00 -1.15157521e+00
-1.00307763e+00 -9.46138129e-02 -1.30042620e-02 7.86631629e-02
-8.99727643e-01 1.88787684e-01 -5.18206954e-01 -1.14409530e+00
2.85055280e-01 -7.47078955e-01 -3.21698874e-01 -6.73051119e-01
-8.71417761e-01 -6.67890608e-01 8.49889934e-01 -4.62278455e-01
1.74388897e+00 -1.86394119e+00 4.06943709e-01 6.62480056e-01
9.99932289e-01 -5.64315856e-01 2.32623145e-01 9.40987349e-01
-9.25860181e-03 7.26183891e-01 1.30449399e-01 3.15772891e-02
-6.24564886e-02 -4.72438056e-03 -1.01041496e+00 2.16730222e-01
7.12942258e-02 6.75581396e-01 -1.18992758e+00 -8.67456436e-01
-2.67003328e-01 7.11649209e-02 -3.75186831e-01 3.04427147e-02
-5.17639995e-01 -5.51944971e-02 -8.43451738e-01 5.88514388e-01
-2.42379624e-02 -5.48615396e-01 6.58799931e-02 -1.39871269e-01
-2.51754791e-01 8.63719523e-01 -9.17711318e-01 1.02249491e+00
1.99571364e-02 8.59613776e-01 -1.86131909e-01 -6.82323456e-01
5.80793679e-01 3.42841327e-01 1.03668201e+00 -7.11615779e-04
1.32798910e-01 5.95729090e-02 -3.48598510e-01 -1.62160750e-02
1.04651213e+00 1.36443332e-01 -2.36891553e-01 8.56396854e-01
1.73722833e-01 -1.47638604e-01 6.89997733e-01 1.05898941e+00
1.31693578e+00 -5.44928491e-01 8.11257958e-01 -9.73488629e-01
2.66306370e-01 4.60049808e-01 1.71371270e-02 9.50858474e-01
3.14088553e-01 1.57357976e-01 1.08690631e+00 -2.47449622e-01
-1.24924576e+00 -6.58334911e-01 8.88929293e-02 1.27966142e+00
-1.41453445e-01 -1.42506933e+00 -3.30848277e-01 -8.56366754e-01
4.10409153e-01 7.66894519e-01 -9.33957338e-01 2.76957482e-01
-3.05742353e-01 -7.18932509e-01 2.32908472e-01 1.39195159e-01
-2.12922722e-01 -6.68818355e-01 -2.06726491e-01 3.02924663e-01
-3.39433849e-01 -1.13913929e+00 -2.63590693e-01 4.59039658e-01
-8.42166007e-01 -1.08345532e+00 -7.35105455e-01 -4.20598537e-01
1.02608788e+00 3.95168275e-01 1.46529353e+00 4.61083651e-03
-1.92217395e-01 3.21034700e-01 -6.12786055e-01 -6.35283887e-01
-4.60025012e-01 6.66286051e-01 2.82180905e-02 -3.51885408e-01
5.95968425e-01 -5.74633837e-01 -3.88692349e-01 -2.99065441e-01
-6.34615481e-01 1.15090303e-01 6.22727752e-01 5.65851212e-01
5.12122214e-01 6.08877838e-01 2.60606647e-01 -1.17116916e+00
1.31470287e+00 -7.01909304e-01 -4.05095190e-01 2.00838037e-02
-1.29315484e+00 4.84132022e-01 5.18086672e-01 -3.32370728e-01
-5.53811133e-01 -4.96979713e-01 4.15475488e-01 1.68841377e-01
8.93653631e-02 9.00671363e-01 5.48673391e-01 3.95427704e-01
5.74387789e-01 3.37710530e-01 -6.76044881e-01 -5.03590822e-01
6.31158054e-01 4.98385936e-01 1.64943084e-01 -7.11019039e-01
1.12374222e+00 5.10956109e-01 1.24468088e-01 -1.14938879e+00
-1.13654530e+00 -1.19041908e+00 -5.81366897e-01 -1.78341210e-01
5.74168921e-01 -8.98753107e-01 -4.84315187e-01 -3.13953608e-01
-1.11844122e+00 5.78046799e-01 -3.84282023e-01 2.70147502e-01
-4.81744371e-02 3.61330330e-01 -6.18901908e-01 -6.29214644e-01
-7.66885579e-01 -5.00595927e-01 1.09771419e+00 3.14656943e-02
-9.69817519e-01 -9.42064106e-01 3.86001587e-01 7.25886524e-02
7.85531998e-02 1.62952438e-01 1.28646386e+00 -1.00386739e+00
-4.20542359e-01 -2.78577596e-01 -2.38672987e-01 -4.11106288e-01
2.75244474e-01 2.36770585e-01 -6.79303706e-01 -1.00621052e-01
-3.67150277e-01 1.21775635e-01 1.18883181e+00 2.71254897e-01
7.98209429e-01 -9.58224893e-01 -7.49493659e-01 -8.53496194e-02
1.22025502e+00 -1.99884176e-01 7.14465007e-02 7.39754677e-01
6.79680467e-01 7.56317317e-01 6.29055262e-01 6.94249094e-01
3.03437531e-01 1.83613494e-01 -9.20015052e-02 -5.01293428e-02
4.36924249e-01 -5.80383360e-01 1.12356588e-01 9.82283354e-01
-9.75483209e-02 -2.30814368e-01 -1.05058110e+00 1.03967595e+00
-1.83502579e+00 -9.67629194e-01 -3.74434292e-01 1.40095258e+00
1.26010525e+00 6.84204638e-01 2.83381432e-01 4.12164599e-01
4.07781631e-01 5.07019103e-01 1.91387400e-01 -1.31334499e-01
1.40348971e-01 5.26348352e-02 6.30546927e-01 3.69638950e-01
-1.24025881e+00 1.14661336e+00 6.30602455e+00 6.37443781e-01
-5.10256648e-01 -1.30033121e-01 3.75822663e-01 2.00558990e-01
-5.79953194e-01 3.20452780e-01 -1.08488286e+00 1.65419862e-01
8.34557116e-01 -7.27620363e-01 -2.30382010e-01 1.16896307e+00
3.18023086e-01 -3.78057480e-01 -9.82808053e-01 4.63219762e-01
7.42557505e-03 -1.66568267e+00 4.76784527e-01 2.19979718e-01
8.30642700e-01 -1.08618386e-01 -1.29180536e-01 -2.37211362e-01
5.65427065e-01 -4.83376205e-01 4.55266505e-01 2.59245038e-01
1.81883857e-01 -8.00408125e-01 4.61518645e-01 1.65865988e-01
-1.09061062e+00 8.50955918e-02 -2.29942933e-01 -1.01971151e-02
-3.32273208e-02 1.20886075e+00 -1.37243533e+00 3.52861285e-01
7.60979056e-01 7.55593717e-01 -6.52636886e-01 6.74385309e-01
-5.21832883e-01 1.05397069e+00 -1.12412445e-01 -5.73589087e-01
5.51126897e-01 1.29142419e-01 1.06701958e+00 1.82275462e+00
-8.94145519e-02 1.11656450e-01 1.37615487e-01 6.59228384e-01
-3.31786305e-01 4.50801760e-01 -7.26405144e-01 -9.75192010e-01
4.15609956e-01 1.44371986e+00 -1.50545931e+00 -6.67596877e-01
-5.62721603e-02 5.01191914e-01 -7.52598494e-02 2.25234360e-01
-1.62748232e-01 -5.93853593e-01 4.56017137e-01 3.64969939e-01
3.33288074e-01 -5.45205176e-01 -5.60590982e-01 -1.13663208e+00
-1.29200503e-01 -8.07283044e-01 5.45626283e-01 -3.26533079e-01
-1.20257318e+00 3.29635173e-01 2.14971125e-01 -8.25579405e-01
-2.87432551e-01 -3.99511695e-01 -6.22133613e-01 5.41126192e-01
-1.38500607e+00 -8.89263213e-01 1.45705253e-01 2.74178475e-01
5.94488561e-01 -2.14802608e-01 7.76793659e-01 -3.07573944e-01
-7.70215467e-02 -2.24091873e-01 4.71576937e-02 1.42567828e-01
5.06230474e-01 -1.64087749e+00 7.04686761e-01 9.88135695e-01
6.78418696e-01 9.60834324e-01 1.28772020e+00 -8.61549199e-01
-1.30038595e+00 -7.20128000e-01 1.61095881e+00 -7.11639345e-01
1.25199330e+00 -4.03950602e-01 -6.53752446e-01 3.30322236e-01
5.06535947e-01 -7.94633508e-01 9.34365034e-01 7.41480768e-01
-3.87117505e-01 2.76720554e-01 -4.30547118e-01 8.73851597e-01
6.21737599e-01 -5.48022687e-01 -1.18961382e+00 5.34705698e-01
1.08008206e+00 2.00019777e-01 -8.91330540e-01 7.25157186e-02
1.78257167e-01 -1.74560651e-01 9.72489953e-01 -8.12635064e-01
8.42923760e-01 -1.49972513e-01 1.32882029e-01 -1.17983365e+00
-5.26925981e-01 -1.03541458e+00 -7.74288356e-01 1.28983414e+00
7.26705551e-01 -1.36596859e-01 7.51045227e-01 1.62392274e-01
5.25333464e-01 -5.91569364e-01 -3.69908363e-01 -1.67530164e-01
-2.92210490e-01 -2.90434629e-01 1.91519126e-01 1.31946540e+00
6.32118642e-01 9.43335116e-01 -8.96496028e-02 -2.73710787e-02
6.95862353e-01 3.54420632e-01 6.45442247e-01 -1.72127140e+00
2.63441186e-02 -7.64343083e-01 -1.11576319e-01 -6.26497567e-01
8.35036859e-02 -7.89170682e-01 -2.42735803e-01 -1.78512645e+00
5.36973298e-01 -1.05777137e-01 -3.61976624e-01 3.43780547e-01
-4.90614206e-01 -2.94243962e-01 -1.73295394e-01 5.96301019e-01
-6.92332089e-01 2.69174844e-01 8.81696284e-01 -3.78347695e-01
-3.98273587e-01 -1.68855246e-02 -1.28392780e+00 8.10994744e-01
6.03243768e-01 -9.65319693e-01 -3.93673450e-01 2.53496885e-01
8.42705786e-01 -4.64299589e-01 1.84141114e-01 -4.34309512e-01
4.80770499e-01 -3.03953499e-01 1.63162693e-01 -8.57089043e-01
-2.41327509e-01 -4.83884811e-01 -5.25288999e-01 2.65018582e-01
-7.57296622e-01 3.81791115e-01 1.15667969e-01 7.38015115e-01
-2.36207783e-01 -1.99796483e-01 2.62265295e-01 -2.35082626e-01
-5.29370725e-01 2.81629320e-02 -6.61235631e-01 3.11733007e-01
5.67204297e-01 4.28089917e-01 -3.66190642e-01 -4.06018555e-01
-3.65591645e-01 5.40060876e-03 1.67980324e-02 5.67799151e-01
6.87202811e-01 -8.86206686e-01 -9.04105604e-01 -2.92974353e-01
2.66071916e-01 -2.85845637e-01 -6.07365370e-01 6.99282467e-01
-2.45685160e-01 8.70798707e-01 4.01173502e-01 -2.42738202e-01
-1.38332951e+00 5.55728197e-01 -6.43119991e-01 -7.36781895e-01
-9.09275413e-01 7.31345773e-01 -1.67148009e-01 2.37704843e-01
1.15822986e-01 -3.66320252e-01 -5.87667704e-01 8.35912466e-01
5.59839845e-01 2.01424584e-01 -5.89292794e-02 -4.11071271e-01
-2.78083354e-01 2.23640218e-01 -5.67355156e-01 -3.97743791e-01
1.49271286e+00 -7.27204382e-02 -5.55436313e-01 5.62264979e-01
1.08330870e+00 3.77999693e-01 -3.89688849e-01 -4.98702228e-01
9.00194168e-01 -1.61482781e-01 4.55265909e-01 -5.08865178e-01
-2.69402921e-01 3.68287921e-01 -4.34286565e-01 8.63646448e-01
7.42696524e-01 3.97010028e-01 5.54943502e-01 6.84051573e-01
2.99104527e-02 -1.42904377e+00 2.06543267e-01 5.51476240e-01
8.14829588e-01 -1.02500689e+00 8.06322455e-01 -4.37858313e-01
-4.72839475e-01 1.16700006e+00 3.84830832e-02 6.43833354e-02
9.71544981e-01 2.31149539e-01 -2.58896917e-01 -8.38375986e-01
-9.20978189e-01 -2.40464777e-01 7.71014988e-01 -1.76230624e-01
6.21678591e-01 -6.06455579e-02 -7.70123124e-01 5.16477168e-01
-4.00677174e-01 -4.71091479e-01 4.88025486e-01 1.14421165e+00
-8.37332308e-01 -8.41253102e-01 -3.57118338e-01 7.16523528e-01
-1.01901996e+00 -5.64905882e-01 -1.16133368e+00 7.20268667e-01
-3.57591659e-01 8.11958134e-01 -2.13004500e-01 -4.00141835e-01
-1.66480839e-01 1.00820325e-01 -2.32065827e-01 -7.72660792e-01
-5.33925474e-01 4.59459662e-01 1.69588163e-01 -1.43945888e-01
-3.85097325e-01 -5.25202990e-01 -1.08395100e+00 -3.05374503e-01
-3.12947422e-01 8.53379071e-01 5.26615679e-01 9.73653555e-01
5.39686382e-01 7.15993345e-01 4.99431014e-01 -4.76930261e-01
-7.14193508e-02 -1.03544295e+00 -3.90802979e-01 8.46900269e-02
3.40149522e-01 -3.56619000e-01 -3.44803840e-01 2.09625676e-01] | [12.182626724243164, 8.90673828125] |
359dc3cd-55bd-4405-be6a-4f7e54b0d32d | juno-jump-start-reinforcement-learning-based | 2205.08422 | null | https://arxiv.org/abs/2205.08422v1 | https://arxiv.org/pdf/2205.08422v1.pdf | JUNO: Jump-Start Reinforcement Learning-based Node Selection for UWB Indoor Localization | Ultra-Wideband (UWB) is one of the key technologies empowering the Internet of Thing (IoT) concept to perform reliable, energy-efficient, and highly accurate monitoring, screening, and localization in indoor environments. Performance of UWB-based localization systems, however, can significantly degrade because of Non Line of Sight (NLoS) connections between a mobile user and UWB beacons. To mitigate the destructive effects of NLoS connections, we target development of a Reinforcement Learning (RL) anchor selection framework that can efficiently cope with the dynamic nature of indoor environments. Existing RL models in this context, however, lack the ability to generalize well to be used in a new setting. Moreover, it takes a long time for the conventional RL models to reach the optimal policy. To tackle these challenges, we propose the Jump-start RL-based Uwb NOde selection (JUNO) framework, which performs real-time location predictions without relying on complex NLoS identification/mitigation methods. The effectiveness of the proposed JUNO framework is evaluated in term of the location error, where the mobile user moves randomly through an ultra-dense indoor environment with a high chance of establishing NLoS connections. Simulation results corroborate the effectiveness of the proposed framework in comparison to its state-of-the-art counterparts. | ['Arash Mohammadi', 'Ming Hou', 'Zohreh Hajiakhondi-Meybodi'] | 2022-05-06 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-5.19110076e-02 -1.17588721e-01 -1.71798140e-01 3.98536436e-02
-8.07951331e-01 -4.37518597e-01 2.43722931e-01 4.33306098e-01
-3.63841861e-01 1.22298717e+00 -3.29434127e-01 -5.65875351e-01
-4.87108558e-01 -1.01921535e+00 -4.98741925e-01 -9.94367957e-01
-3.83351296e-01 9.54077318e-02 2.33711123e-01 2.34048087e-02
-6.27264455e-02 5.38960755e-01 -9.11573768e-01 -8.75320792e-01
1.00272286e+00 1.38917232e+00 1.71104580e-01 5.28021336e-01
-4.06792574e-03 1.58003032e-01 -6.78661466e-01 1.20793261e-01
-3.24583389e-02 -1.62289172e-01 -1.04182847e-01 -4.75443691e-01
-2.68590689e-01 -1.28035113e-01 -1.89607009e-01 6.38574660e-01
1.12176013e+00 1.91118896e-01 3.91871780e-01 -1.26892793e+00
-1.56227991e-01 4.56973046e-01 -4.07958269e-01 2.27271467e-01
5.70046425e-01 -2.95436621e-01 4.76328820e-01 -4.53298450e-01
2.36516461e-01 7.65069723e-01 1.06873715e+00 3.00115973e-01
-7.91696131e-01 -8.00470173e-01 -1.08706184e-01 -9.84226912e-03
-1.50012362e+00 -7.10136116e-01 6.39563859e-01 3.40128630e-01
3.01786780e-01 1.69483304e-01 5.61235845e-01 1.33464837e+00
4.04818296e-01 2.73986816e-01 6.99667096e-01 -6.20768964e-01
8.67081106e-01 -8.11824352e-02 -3.05456579e-01 4.86344010e-01
6.18947804e-01 -2.10486412e-01 -2.73591578e-01 -3.40448022e-01
4.21500206e-01 -9.11834091e-02 8.60719457e-02 -2.09880531e-01
-9.56549823e-01 3.59002590e-01 7.76028514e-01 6.33720338e-01
-6.66377604e-01 6.37555122e-01 -1.51303604e-01 -7.64658253e-05
9.95239466e-02 1.51754379e-01 -1.77424639e-01 -2.13623941e-01
-9.67938185e-01 -3.79379690e-01 7.67823935e-01 1.10090172e+00
3.20331752e-01 1.73258081e-01 -1.36941195e-01 4.97448772e-01
7.37460732e-01 7.87255883e-01 8.70678201e-02 -8.37266743e-01
3.58161181e-01 -1.90247055e-02 6.64598942e-01 -8.27664614e-01
-9.29017842e-01 -1.14039683e+00 -1.03848004e+00 -3.99197340e-01
2.38888457e-01 -7.72923172e-01 -5.30981421e-01 1.79061246e+00
6.97249174e-01 6.51956975e-01 4.51789707e-01 3.80365193e-01
3.97812814e-01 7.15821028e-01 3.20652097e-01 -5.67799985e-01
9.69123423e-01 -4.54637527e-01 -7.72842646e-01 -2.51505107e-01
5.01455605e-01 -5.57478786e-01 3.79947782e-01 1.05610766e-01
-7.81751394e-01 -2.45514065e-01 -1.29100072e+00 7.34021544e-01
-3.40193033e-01 -8.14325660e-02 4.96564478e-01 1.20139718e+00
-8.76550853e-01 -1.25002358e-02 -8.96699071e-01 -7.92575240e-01
4.23310131e-01 5.02583683e-01 2.58177668e-01 -3.56500834e-01
-1.32756472e+00 4.31056172e-01 1.04747213e-01 2.38817200e-01
-5.41669071e-01 -6.42234862e-01 -5.21715522e-01 -1.70723046e-03
4.43250418e-01 -6.83186531e-01 1.11052573e+00 -3.46928895e-01
-1.40503561e+00 -4.76872027e-01 -1.85071275e-01 -5.96959531e-01
5.64796209e-01 6.63966537e-02 -6.93556905e-01 6.65770844e-02
1.55583963e-01 2.90321503e-02 4.31670189e-01 -1.55320382e+00
-7.12894380e-01 -1.77616477e-01 7.38539323e-02 -3.79523337e-02
-4.61341560e-01 -5.98268867e-01 -7.47014284e-02 -3.41801077e-01
2.91673601e-01 -7.94219017e-01 -6.40526831e-01 7.51273334e-02
-4.59414303e-01 -1.24328129e-01 7.86101699e-01 -1.23102613e-01
1.13224852e+00 -1.99187434e+00 -7.58452892e-01 6.09588683e-01
-3.13569695e-01 1.46986380e-01 1.19674549e-01 7.64788866e-01
7.71373451e-01 4.81399931e-02 2.10102886e-01 -5.67452490e-01
-1.07466960e-02 3.60026270e-01 9.34248641e-02 6.92060173e-01
-4.92438972e-01 2.66910404e-01 -1.25371253e+00 -6.21761501e-01
2.79305309e-01 6.64479196e-01 -1.58664703e-01 6.48194999e-02
1.71909958e-01 9.23607647e-01 -1.19857740e+00 8.51357996e-01
7.19460011e-01 3.60864438e-02 1.03782475e-01 -1.00630485e-01
-3.26035559e-01 -1.96999356e-01 -1.42400324e+00 1.58803844e+00
-1.15179813e+00 -4.11201790e-02 1.13977946e-01 -8.97925913e-01
9.12994981e-01 5.40070176e-01 8.89612257e-01 -1.03584135e+00
2.16174692e-01 6.42695963e-01 -2.93982953e-01 -4.57661688e-01
1.61405317e-02 3.15248109e-02 -2.92049319e-01 2.38345921e-01
3.85718979e-02 4.25677568e-01 -2.06842143e-02 -7.62958685e-03
1.51211452e+00 -1.71863943e-01 4.90830779e-01 -1.24594606e-01
7.57208228e-01 -6.35880709e-01 6.87416434e-01 1.30407238e+00
-3.12626332e-01 -1.44458354e-01 -1.48918957e-01 1.19114369e-02
-4.64763016e-01 -1.30773413e+00 -2.26740167e-01 7.14597940e-01
6.83151364e-01 1.02938041e-01 -4.37547952e-01 -3.76715004e-01
-3.73607092e-02 8.89862597e-01 -7.46275410e-02 -2.01714009e-01
-1.81229427e-01 -7.92669356e-01 1.05497634e+00 7.77752027e-02
9.15757418e-01 -4.85766679e-01 -8.36874425e-01 7.24106908e-01
-2.16367438e-01 -1.48979056e+00 2.64708161e-01 4.69872266e-01
-5.14064193e-01 -6.12179518e-01 -6.67601228e-01 -4.06181097e-01
4.87353295e-01 3.85604858e-01 4.25448030e-01 1.73058003e-01
-5.60150892e-02 8.56397986e-01 -5.70811927e-01 -4.42709148e-01
-1.43942609e-01 2.74094582e-01 5.21272659e-01 2.42797554e-01
-2.88784444e-01 -8.35680246e-01 -1.14534712e+00 4.09525067e-01
-4.88799095e-01 -7.45249093e-01 7.28286624e-01 4.50319558e-01
3.24735701e-01 4.97068763e-01 1.44580185e+00 -1.56424254e-01
3.51246178e-01 -1.09609592e+00 -5.34572244e-01 4.30856854e-01
-5.85395157e-01 -1.90357357e-01 7.43740261e-01 -1.36676788e-01
-1.11571276e+00 1.00871816e-01 -5.50914705e-01 5.57034612e-01
-1.32807493e-01 3.16114664e-01 -2.85734773e-01 -7.31990933e-01
3.48238975e-01 5.60333654e-02 -8.37917745e-01 -2.59962887e-01
2.57646024e-01 8.60429466e-01 4.37061012e-01 -6.31642997e-01
1.02884507e+00 5.58477223e-01 4.91155267e-01 -1.33346951e+00
-6.18826210e-01 -5.84398210e-01 -3.32342684e-01 -3.47238928e-01
4.64810461e-01 -8.07110429e-01 -7.76638925e-01 -3.14308866e-03
-9.37302172e-01 -1.13112316e-01 1.33919626e-01 5.45556486e-01
-3.23869795e-01 5.35430193e-01 -1.25101600e-02 -1.54323220e+00
-2.98684657e-01 -8.51560473e-01 6.52723074e-01 6.88669741e-01
1.24894902e-01 -1.07091033e+00 -9.97075811e-02 7.85330534e-02
7.52415597e-01 4.90615010e-01 7.22231925e-01 -5.16858101e-01
-4.93773580e-01 -5.19859314e-01 -7.25893602e-02 -3.34600061e-01
1.23934053e-01 -4.30733025e-01 -7.94373989e-01 -2.84456253e-01
-2.46120438e-01 -1.61227182e-01 2.56370634e-01 4.21881825e-01
8.65491509e-01 -3.17998946e-01 -7.67271578e-01 4.56503898e-01
1.65049314e+00 2.44911686e-01 4.51606572e-01 3.71063262e-01
1.65887162e-01 -1.22336298e-01 7.39200294e-01 7.77784705e-01
5.27576447e-01 6.98080361e-01 9.37860250e-01 -5.55359870e-02
1.67246059e-01 -2.06797555e-01 8.77271518e-02 2.30373219e-01
1.41050890e-01 -9.91262615e-01 -6.81670845e-01 3.13397616e-01
-1.65356803e+00 -5.35562098e-01 -9.16435663e-03 2.35976672e+00
2.78195351e-01 2.26832867e-01 -1.53951585e-01 4.70720977e-01
4.23588425e-01 1.06173120e-01 -4.22497392e-01 -9.04802233e-02
2.51187891e-01 1.21906204e-02 9.76087153e-01 3.90127033e-01
-9.61796105e-01 3.86306196e-01 4.73974943e+00 8.40413749e-01
-9.15324152e-01 3.76394391e-01 1.48169637e-01 5.19440651e-01
-1.03046633e-01 -3.53610098e-01 -5.54815769e-01 4.77928072e-01
1.17288685e+00 2.28432655e-01 8.39227960e-02 6.49026155e-01
6.88880622e-01 -5.48643231e-01 -5.32177448e-01 7.33418047e-01
-2.69714773e-01 -9.67038870e-01 -5.96887648e-01 2.41290573e-02
6.72183096e-01 -1.78813711e-02 -2.84901485e-02 7.49298632e-02
5.64341433e-02 -5.01632214e-01 5.43148935e-01 5.71726143e-01
4.40456241e-01 -9.15922761e-01 9.28986132e-01 5.89859188e-01
-1.43920887e+00 -7.32459486e-01 -4.08909589e-01 2.35396937e-01
6.08255446e-01 9.90915835e-01 -7.78035104e-01 9.73968506e-01
5.17872751e-01 6.84736893e-02 -3.71791929e-01 1.85518301e+00
-1.36185840e-01 6.03644907e-01 -7.95442045e-01 -3.89024884e-01
6.00304782e-01 3.10044110e-01 6.94036305e-01 9.53388989e-01
1.09963071e+00 -4.02454846e-02 3.30770910e-01 3.67470384e-01
-1.56090438e-01 -2.01389771e-02 -5.20331919e-01 5.02239168e-01
1.19840777e+00 1.25402033e+00 -1.08199835e+00 1.84042230e-01
-1.84209391e-01 6.89172149e-01 -3.91148537e-01 6.55283511e-01
-9.62525785e-01 -4.02579308e-01 9.72979218e-02 -5.20991441e-03
1.58572197e-01 -6.02894962e-01 -2.68485129e-01 -4.66941416e-01
-1.23798557e-01 -2.77873557e-02 7.31822178e-02 -3.11412543e-01
-1.00895536e+00 2.36108869e-01 -2.75990665e-01 -1.19842052e+00
2.08970979e-01 1.84184819e-01 -6.95264697e-01 1.93201318e-01
-1.84238505e+00 -1.23269260e+00 -5.20154417e-01 4.69748020e-01
3.55452150e-01 2.19607741e-01 9.73190308e-01 7.90733755e-01
-5.27890801e-01 7.65440702e-01 6.42133713e-01 -3.01208556e-01
5.06692111e-01 -9.39968288e-01 -2.09282160e-01 8.47123504e-01
-7.83405676e-02 4.95794296e-01 9.56701100e-01 -6.13692164e-01
-1.70803428e+00 -1.17888474e+00 5.97098708e-01 2.36238450e-01
4.23127294e-01 -2.82486796e-01 -9.98253673e-02 1.09775871e-01
-7.44196251e-02 6.19324207e-01 7.42070138e-01 -3.33668977e-01
3.72253656e-01 -5.72835982e-01 -1.55761182e+00 4.41460788e-01
8.94038975e-01 1.53357133e-01 9.31535885e-02 3.03251177e-01
3.58498544e-01 -7.68732280e-03 -8.50735307e-01 4.34010535e-01
8.80172849e-01 -6.95399165e-01 1.35117209e+00 3.13701183e-01
-8.64805162e-01 -2.84014672e-01 -3.67238730e-01 -1.17091572e+00
2.18936950e-01 -8.62420559e-01 -2.77257890e-01 1.67774606e+00
3.53591800e-01 -7.03361034e-01 5.77575326e-01 5.07242419e-02
1.24425754e-01 -6.45096779e-01 -1.65049684e+00 -9.35877264e-01
-5.41530788e-01 -6.64028049e-01 6.03135049e-01 2.76014239e-01
-1.79672658e-01 2.05351189e-01 -4.29399997e-01 9.04450834e-01
9.34567213e-01 -6.65953457e-01 5.07069886e-01 -1.31746185e+00
5.21615148e-02 9.19622853e-02 -2.89161026e-01 -1.04696560e+00
-2.44337574e-01 -2.64727175e-01 5.10889769e-01 -2.01530576e+00
-6.93215907e-01 -1.39681673e+00 -5.66087484e-01 1.40464365e-01
3.63457829e-01 4.14913684e-01 -1.45641133e-01 -1.92185253e-01
-1.14336836e+00 6.66909456e-01 5.66663980e-01 -4.16638218e-02
-3.36514622e-01 8.44879031e-01 -2.58320242e-01 7.31155038e-01
8.69925976e-01 -7.17774510e-01 -5.42506635e-01 -1.71045497e-01
2.75812775e-01 4.45433736e-01 4.12070155e-01 -1.78280556e+00
4.85432625e-01 4.13883552e-02 5.70234597e-01 -3.03430140e-01
3.87868613e-01 -1.49125159e+00 -9.58052836e-03 8.12581837e-01
1.03404678e-01 -3.10498863e-01 -1.50643721e-01 1.19563234e+00
5.48801959e-01 -3.39302540e-01 5.94661832e-01 2.17667297e-01
-5.45163274e-01 2.68764406e-01 -7.29759216e-01 -4.10447776e-01
1.33910394e+00 -2.18689576e-01 -2.07179397e-01 -6.40782177e-01
-3.91777605e-01 4.45546120e-01 -8.95846337e-02 6.64012283e-02
3.42278421e-01 -1.19245851e+00 4.91694398e-02 -2.51350701e-01
-6.54929578e-02 -3.29495609e-01 2.36621067e-01 1.13853467e+00
-2.63339400e-01 3.95915627e-01 1.07277200e-01 -3.62356186e-01
-6.90417826e-01 1.53673828e-01 4.29515362e-01 -2.35192642e-01
-3.41823399e-01 4.93422776e-01 -7.56203413e-01 -2.39562258e-01
7.27833152e-01 -1.60418615e-01 -2.03265816e-01 -9.01628658e-02
2.41439894e-01 7.26718962e-01 9.15974081e-02 -3.69087100e-01
-6.96859837e-01 6.67244256e-01 8.01696062e-01 -6.61645010e-02
1.22117031e+00 -8.64432752e-01 3.47762436e-01 3.26309025e-01
6.19384170e-01 3.40547889e-01 -9.75783467e-01 -2.44780779e-02
1.82753429e-01 -4.64858636e-02 1.66220486e-01 -7.56115437e-01
-5.83229005e-01 4.94674861e-01 1.16820490e+00 4.32683349e-01
8.16363752e-01 -2.73894638e-01 1.14419246e+00 6.88231945e-01
9.72834051e-01 -1.04203784e+00 1.60009488e-01 2.37653274e-02
6.85941279e-02 -1.02853572e+00 -8.96505453e-03 -1.90962479e-01
2.26924330e-01 1.22060704e+00 3.01000983e-01 1.32098973e-01
8.70336115e-01 2.74599403e-01 -7.03696162e-02 2.02491283e-01
3.19363587e-02 -3.74828011e-01 -3.99027556e-01 9.15366232e-01
1.41493371e-02 -2.12106600e-01 -3.12806219e-01 6.62282407e-01
2.65312970e-01 -2.93065365e-02 3.67972791e-01 1.16676021e+00
-7.79137433e-01 -1.05350363e+00 -5.63419938e-01 1.76692501e-01
-5.84143579e-01 4.29189801e-01 4.68215019e-01 7.39751577e-01
4.17411268e-01 1.50748265e+00 -5.13419092e-01 -1.29679024e-01
2.61540383e-01 -2.74201483e-01 2.11588249e-01 -5.47819920e-02
7.43954778e-02 9.75140706e-02 1.33357674e-01 -3.78292114e-01
-5.02463579e-01 -3.73744488e-01 -1.46985579e+00 9.32867527e-02
-4.52264935e-01 4.19910282e-01 1.02913249e+00 1.38051128e+00
1.11616738e-01 8.23202193e-01 1.04696524e+00 -7.43159831e-01
-4.78390485e-01 -6.52499020e-01 -3.59366328e-01 -6.57106876e-01
7.07696736e-01 -9.87900674e-01 -1.92532822e-01 -8.49647582e-01] | [6.260490894317627, 1.1292407512664795] |
b0d239f5-7282-46df-9169-7a7bc3105101 | quo-vadis-is-trajectory-forecasting-the-key | 2210.07681 | null | https://arxiv.org/abs/2210.07681v2 | https://arxiv.org/pdf/2210.07681v2.pdf | Quo Vadis: Is Trajectory Forecasting the Key Towards Long-Term Multi-Object Tracking? | Recent developments in monocular multi-object tracking have been very successful in tracking visible objects and bridging short occlusion gaps, mainly relying on data-driven appearance models. While we have significantly advanced short-term tracking performance, bridging longer occlusion gaps remains elusive: state-of-the-art object trackers only bridge less than 10% of occlusions longer than three seconds. We suggest that the missing key is reasoning about future trajectories over a longer time horizon. Intuitively, the longer the occlusion gap, the larger the search space for possible associations. In this paper, we show that even a small yet diverse set of trajectory predictions for moving agents will significantly reduce this search space and thus improve long-term tracking robustness. Our experiments suggest that the crucial components of our approach are reasoning in a bird's-eye view space and generating a small yet diverse set of forecasts while accounting for their localization uncertainty. This way, we can advance state-of-the-art trackers on the MOTChallenge dataset and significantly improve their long-term tracking performance. This paper's source code and experimental data are available at https://github.com/dendorferpatrick/QuoVadis. | ['Laura Leal-Taixé', 'Aljoša Ošep', 'Vladimir Yugay', 'Patrick Dendorfer'] | 2022-10-14 | null | null | null | null | ['trajectory-forecasting'] | ['computer-vision'] | [-5.34234226e-01 -3.98883224e-01 -1.49671182e-01 4.23502736e-02
-4.32277292e-01 -9.26436305e-01 6.30744815e-01 -1.00543164e-01
-4.11274970e-01 8.18977416e-01 -1.29582420e-01 -4.14983392e-01
-7.88576305e-02 -3.02070707e-01 -7.13216662e-01 -5.16852558e-01
-2.08549008e-01 5.37554741e-01 8.66506338e-01 3.81805636e-02
-1.29226312e-01 5.86501658e-01 -1.73398399e+00 -2.62547165e-01
5.35226524e-01 6.82502866e-01 1.46026045e-01 9.91303444e-01
2.93781459e-01 6.30613029e-01 -6.01804137e-01 -2.28140175e-01
4.29479629e-01 4.25402597e-02 -2.14951560e-01 -1.71804547e-01
8.60472560e-01 -4.62213725e-01 -4.80799496e-01 9.44758415e-01
4.41125005e-01 1.44725278e-01 1.47965357e-01 -1.72671413e+00
-4.57713157e-01 1.80065960e-01 -4.64884609e-01 5.23110867e-01
3.49200994e-01 7.04662383e-01 6.78245425e-01 -6.71663523e-01
7.89593279e-01 1.26422334e+00 8.64177644e-01 6.80433333e-01
-1.09593272e+00 -7.91626036e-01 6.36261523e-01 1.29738182e-01
-1.33746660e+00 -5.48991084e-01 3.23647112e-01 -7.85093963e-01
9.14845049e-01 4.86693889e-01 8.63975585e-01 9.83610451e-01
3.02766144e-01 7.33544767e-01 8.16791296e-01 2.79258359e-02
-1.55610982e-02 1.89259201e-01 4.51441556e-02 6.02985442e-01
6.17078066e-01 7.78930366e-01 -4.93942052e-01 -2.48847827e-01
6.30046725e-01 6.15267269e-02 -1.10159449e-01 -7.30161428e-01
-1.36611223e+00 5.70099235e-01 4.53388989e-01 1.29326075e-01
-1.39430732e-01 4.60641325e-01 6.72092512e-02 1.87393472e-01
5.44763923e-01 1.82475284e-01 -3.59121263e-01 -2.48836368e-01
-8.64011049e-01 7.43142903e-01 6.00707889e-01 1.18914044e+00
4.56550837e-01 -4.17782031e-02 -3.88294458e-01 9.93752554e-02
5.41330695e-01 8.75145078e-01 -1.77304983e-01 -1.01685357e+00
2.35566661e-01 4.13502514e-01 6.48898244e-01 -7.37160623e-01
-5.58145046e-01 -6.20819092e-01 1.31889340e-02 7.61399388e-01
8.11314821e-01 -5.19683480e-01 -8.24340403e-01 1.80666268e+00
8.38237345e-01 3.47019643e-01 -3.76005143e-01 9.94174123e-01
4.75995034e-01 2.32271567e-01 6.27559423e-02 2.97432020e-02
1.39260459e+00 -1.01464534e+00 -5.10219574e-01 -5.26015878e-01
5.53442121e-01 -8.89568925e-01 3.98228317e-01 1.55187696e-01
-1.07238436e+00 -6.33739889e-01 -8.02997172e-01 2.30464816e-01
-3.54909062e-01 2.09665179e-01 6.87816203e-01 7.39352107e-01
-1.19700408e+00 3.56784880e-01 -1.13758564e+00 -7.22506106e-01
4.16159034e-01 3.51151228e-01 -1.06326386e-01 1.06279060e-01
-7.97688246e-01 1.24638307e+00 1.00866459e-01 4.16690633e-02
-9.83470619e-01 -7.74944842e-01 -5.83529651e-01 -2.62938887e-01
6.06148243e-01 -8.75589311e-01 1.34061754e+00 -2.52632499e-01
-9.80009913e-01 5.33121169e-01 -2.67827362e-01 -7.84830689e-01
9.11213756e-01 -5.50155997e-01 -3.42732817e-01 -2.81400859e-01
5.49643897e-02 9.15306687e-01 6.65863156e-01 -1.30867577e+00
-1.00468612e+00 -4.61682916e-01 1.15005568e-01 2.79977202e-01
3.81939188e-02 1.72364816e-01 -6.59564674e-01 -3.96283180e-01
-1.87455222e-01 -1.36395514e+00 -2.18637004e-01 6.58681750e-01
1.07793793e-01 -2.35339254e-01 1.05877578e+00 -3.76283795e-01
1.09039164e+00 -1.85567129e+00 2.93850191e-02 -2.56626427e-01
3.62204581e-01 3.37220937e-01 1.23217078e-02 3.96580487e-01
4.41430628e-01 -3.31500590e-01 5.08739352e-01 -5.71397305e-01
2.43309885e-01 -7.70769715e-02 -4.88327712e-01 7.13580251e-01
-1.18281119e-01 1.05783367e+00 -1.10586178e+00 -2.55599469e-01
6.94417477e-01 4.81537610e-01 -2.76788533e-01 -5.58324046e-02
-5.61987281e-01 7.16029584e-01 -4.39729035e-01 5.81134439e-01
7.16456056e-01 -2.85132766e-01 -1.84242502e-01 7.56042227e-02
-5.86762428e-01 -6.76072622e-03 -1.23572052e+00 1.61466229e+00
1.09589975e-02 9.52697396e-01 2.85316352e-02 -2.01766551e-01
6.40284538e-01 -3.24199162e-02 7.34887064e-01 -4.98068780e-01
4.90567163e-02 -7.34222233e-02 1.14055373e-01 -1.37508363e-01
6.30411923e-01 2.57140964e-01 2.63506711e-01 2.39593908e-01
-4.59714532e-01 1.76366791e-01 4.51439351e-01 2.81209856e-01
1.20072961e+00 6.59661472e-01 -1.75451506e-02 5.87477349e-03
2.03088224e-01 4.74486798e-01 6.10247731e-01 9.61461782e-01
-6.51761532e-01 2.14477196e-01 -1.66889623e-01 -6.52002037e-01
-8.98660779e-01 -1.09077120e+00 -1.80101261e-01 1.01132607e+00
6.07437670e-01 -3.92033935e-01 -4.13835794e-01 -5.24966717e-01
4.05123591e-01 5.82781196e-01 -5.32104671e-01 1.04117215e-01
-5.85641563e-01 -5.75586557e-01 4.82672095e-01 6.92445099e-01
4.94029783e-02 -7.90744185e-01 -1.23500621e+00 2.57841080e-01
1.47819713e-01 -1.19261241e+00 -4.19512331e-01 -2.90715069e-01
-6.72872066e-01 -1.17121518e+00 -6.44723594e-01 -1.04820892e-01
5.02817035e-01 7.83702075e-01 1.11393833e+00 2.23867849e-01
-5.55979490e-01 6.09617889e-01 -1.68496028e-01 -6.42731786e-01
-2.10115150e-01 -3.78338367e-01 2.41406530e-01 -5.26943266e-01
4.99861032e-01 -1.09667994e-01 -7.40968406e-01 6.15286887e-01
-9.27983075e-02 1.08725384e-01 3.22397619e-01 3.72717381e-01
2.98305094e-01 -1.64427564e-01 -3.10618579e-02 -2.23463163e-01
1.16747670e-01 -4.03318614e-01 -1.34130168e+00 2.89729893e-01
-4.52430218e-01 -6.09465986e-02 7.94561878e-02 -7.59787500e-01
-9.44013476e-01 5.57198338e-02 3.17651898e-01 -8.41337502e-01
-1.66769683e-01 -4.68135215e-02 5.25965393e-01 -2.90269315e-01
6.11711740e-01 -4.49155010e-02 -1.35697782e-01 -3.36177707e-01
4.37741250e-01 6.19342588e-02 3.74824136e-01 -4.91853237e-01
1.07920635e+00 9.34466839e-01 1.35263905e-01 -5.05077481e-01
-5.40690839e-01 -7.12243617e-01 -4.92634535e-01 -7.25392938e-01
8.66464436e-01 -1.21424377e+00 -1.25489700e+00 2.96424508e-01
-1.11729133e+00 -5.29250622e-01 -1.91648945e-01 6.22694910e-01
-3.31400812e-01 2.00660422e-01 -1.98153764e-01 -1.17515588e+00
-5.78873903e-02 -1.04495943e+00 1.09975767e+00 3.64572287e-01
-2.11660191e-01 -9.21463132e-01 4.99037147e-01 2.48692185e-01
4.51081991e-01 2.04866275e-01 -5.99859878e-02 -3.12358201e-01
-1.29346395e+00 -1.90041840e-01 -2.99794048e-01 -4.89887178e-01
-2.02140212e-01 1.48519307e-01 -8.37860286e-01 -7.80272245e-01
-3.61507177e-01 9.58541185e-02 9.00282085e-01 6.18470669e-01
4.76387084e-01 1.95927173e-01 -1.05262983e+00 4.43269938e-01
1.17589509e+00 2.36877874e-01 2.01693118e-01 4.54532772e-01
6.20940924e-01 3.11900020e-01 1.11517763e+00 3.92692506e-01
7.06911802e-01 1.18805301e+00 7.19781578e-01 1.19918555e-01
-5.68144500e-01 -1.94551587e-01 4.08320934e-01 4.95992973e-02
-6.88336417e-02 -4.29809570e-01 -1.02236497e+00 7.27581859e-01
-2.33044529e+00 -1.20194602e+00 -3.48004758e-01 2.33414531e+00
3.71507734e-01 3.03158402e-01 4.52731848e-01 -4.74727452e-01
6.69649839e-01 3.48072685e-02 -8.51661563e-01 3.97266388e-01
1.84818357e-01 -6.06830657e-01 8.51391077e-01 7.92321980e-01
-1.28177106e+00 9.88948107e-01 5.93473864e+00 4.88595307e-01
-9.01981175e-01 1.70851320e-01 -1.54190764e-01 -5.54871619e-01
1.42547220e-01 3.10574144e-01 -1.53144491e+00 5.09034336e-01
9.16815221e-01 -9.06061605e-02 3.28966737e-01 8.16069484e-01
1.15923427e-01 -4.67433304e-01 -1.11329651e+00 8.51849437e-01
-2.79540457e-02 -1.31288469e+00 -5.52642226e-01 4.59310293e-01
6.25098467e-01 6.56944394e-01 1.77203491e-01 2.31351420e-01
8.02029371e-01 -7.11938620e-01 1.08557916e+00 6.12055540e-01
2.76037395e-01 -1.20843679e-01 3.32318217e-01 4.54577357e-01
-1.64416039e+00 -1.09851174e-01 -4.33580160e-01 -1.26457721e-01
3.35588038e-01 2.02901527e-01 -7.52445638e-01 4.84753788e-01
7.85748661e-01 7.47750819e-01 -8.21698487e-01 1.64071393e+00
2.41371151e-02 1.24710426e-01 -6.54956698e-01 -6.78664148e-02
1.52007371e-01 8.81805718e-02 1.02052116e+00 8.18648875e-01
3.08489859e-01 -1.22675188e-01 4.31941986e-01 7.60614932e-01
3.63749683e-01 -6.60600483e-01 -6.25701725e-01 2.22805843e-01
7.55903780e-01 1.03581417e+00 -6.95888698e-01 -3.67354214e-01
-5.65332353e-01 4.49592412e-01 3.64682049e-01 3.21305633e-01
-1.40193677e+00 3.67697001e-01 1.00303698e+00 -4.78180349e-02
5.75236201e-01 -3.41457099e-01 -1.75466657e-01 -1.23214972e+00
1.17224567e-01 -4.52186763e-01 3.13597947e-01 -8.48932385e-01
-9.41456795e-01 4.63491142e-01 1.07139811e-01 -1.50003374e+00
-1.41637921e-01 -4.60992157e-01 -3.76037031e-01 6.92202985e-01
-1.43768609e+00 -1.44606435e+00 -4.61571842e-01 3.33201438e-01
6.20762289e-01 1.91335694e-03 4.97344732e-01 3.65460753e-01
-4.29273069e-01 2.16093570e-01 1.91967383e-01 -2.66530514e-01
8.11970949e-01 -1.02188551e+00 8.14886868e-01 1.09213710e+00
3.56965810e-01 5.80599070e-01 1.20162189e+00 -9.95816946e-01
-1.59795666e+00 -9.09476280e-01 5.47444522e-01 -1.30341530e+00
8.23645115e-01 -3.78009230e-01 -6.04915798e-01 1.05598712e+00
6.29648864e-02 3.14967722e-01 1.02233462e-01 2.22216561e-01
-2.72297770e-01 1.23289973e-01 -7.93765485e-01 7.84976363e-01
1.29346812e+00 -1.69928461e-01 -3.07979733e-01 4.10835892e-01
4.08378839e-01 -7.85488188e-01 -6.28172576e-01 3.91404033e-01
9.58186328e-01 -9.96778667e-01 1.28312838e+00 -4.25555646e-01
-4.83397633e-01 -7.69038677e-01 1.48055911e-01 -9.44835365e-01
-3.55719447e-01 -7.01192021e-01 -5.67147017e-01 8.72826159e-01
1.42129868e-01 -5.28047740e-01 8.90462399e-01 7.08259463e-01
-2.02966668e-02 -4.97402072e-01 -9.15205836e-01 -1.16347945e+00
-1.82176813e-01 -3.76181066e-01 4.38510805e-01 4.37581599e-01
-2.89144486e-01 -9.71067622e-02 -5.86727500e-01 5.08672893e-01
1.06685174e+00 2.99775481e-01 1.32628334e+00 -1.37163985e+00
-1.06325664e-01 -2.89541036e-01 -3.92956883e-01 -1.20892143e+00
-7.83414766e-02 -3.47451746e-01 1.98627084e-01 -1.43222868e+00
2.72783697e-01 -7.02553928e-01 -9.81197134e-02 2.94297278e-01
-1.77317560e-01 1.46868378e-01 6.75150514e-01 2.46082127e-01
-9.81275380e-01 3.63137484e-01 1.03970361e+00 -3.72170061e-02
-1.74329467e-02 1.47160769e-01 -1.58500314e-01 6.85964763e-01
6.07538640e-01 -6.28867745e-01 -3.40930194e-01 -6.74882233e-01
8.50030854e-02 2.39363797e-02 9.25773323e-01 -1.27339447e+00
7.43844450e-01 -2.74612427e-01 5.14838994e-01 -1.14653289e+00
7.27746308e-01 -1.01346195e+00 6.67673528e-01 7.37008631e-01
4.84481491e-02 7.76489899e-02 7.59006679e-01 7.40879834e-01
5.04029214e-01 2.08018228e-01 4.44014251e-01 -2.24345084e-02
-8.45110536e-01 4.36230093e-01 -1.72211424e-01 -1.97067782e-02
1.43359494e+00 -3.50455254e-01 -8.20452511e-01 -2.74166346e-01
-7.64852703e-01 5.41221142e-01 8.55124414e-01 8.01927149e-01
2.60231525e-01 -1.18587673e+00 -5.84325135e-01 1.80533133e-03
6.23088814e-02 -4.25120562e-01 2.52475500e-01 9.68814313e-01
-1.43752545e-01 8.93000603e-01 -2.24285737e-01 -9.41182733e-01
-1.63319826e+00 6.00492656e-01 2.51557261e-01 -1.14047490e-01
-7.65885711e-01 8.80232513e-01 2.46098757e-01 -2.66784690e-02
3.51877570e-01 -2.59990305e-01 8.73192772e-02 -1.58264384e-01
5.15815020e-01 5.81829906e-01 -3.69966924e-01 -7.60714471e-01
-7.58224487e-01 8.54175270e-01 -1.29302636e-01 -1.16037935e-01
1.09070003e+00 -4.50371176e-01 4.02405918e-01 1.37035742e-01
4.06376928e-01 7.27779195e-02 -2.08428860e+00 -8.68686140e-02
-9.95812938e-02 -9.18564320e-01 5.37897795e-02 -7.25293994e-01
-7.08756745e-01 6.28285229e-01 8.52692783e-01 2.14482725e-01
5.34354448e-01 2.16233671e-01 5.55748105e-01 2.20198214e-01
7.70281553e-01 -5.98714113e-01 -1.58570305e-01 3.50340396e-01
5.79785228e-01 -1.47562718e+00 3.81554902e-01 -2.90010929e-01
-4.22471553e-01 8.04924488e-01 8.92272234e-01 1.46179497e-01
3.18357766e-01 4.26831543e-01 8.44594017e-02 -3.95563543e-01
-1.06670094e+00 -6.24284387e-01 3.02379757e-01 7.23936379e-01
7.98554793e-02 -3.31179276e-02 1.38694629e-01 -1.57346353e-01
6.82713240e-02 -2.39095110e-02 1.90726236e-01 8.80470037e-01
-5.57746291e-01 -8.96448970e-01 -6.61533833e-01 -4.91902232e-02
-9.51133296e-02 2.22419277e-01 -1.11497879e-01 9.58954811e-01
4.63823415e-02 1.09339654e+00 6.65101930e-02 -7.16780424e-02
2.42952466e-01 -2.68010885e-01 7.83002734e-01 -4.23539966e-01
-4.25353140e-01 1.28971949e-01 2.02540189e-01 -7.70513594e-01
-4.62899178e-01 -1.05438030e+00 -1.01573241e+00 -5.97915888e-01
-5.80230832e-01 -1.65157408e-01 6.95526361e-01 8.25239360e-01
5.33571362e-01 5.34959018e-01 -3.16045322e-02 -1.06807399e+00
-4.31643099e-01 -6.93926513e-01 1.59236595e-01 8.73788148e-02
6.37013257e-01 -1.11696792e+00 -1.26063854e-01 -1.03341982e-01] | [6.385349750518799, -2.048044204711914] |
82d8e87b-2292-46a9-987b-ed19e749361b | video-saliency-detection-by-3d-convolutional | 1807.04514 | null | http://arxiv.org/abs/1807.04514v1 | http://arxiv.org/pdf/1807.04514v1.pdf | Video Saliency Detection by 3D Convolutional Neural Networks | Different from salient object detection methods for still images, a key
challenging for video saliency detection is how to extract and combine spatial
and temporal features. In this paper, we present a novel and effective approach
for salient object detection for video sequences based on 3D convolutional
neural networks. First, we design a 3D convolutional network (Conv3DNet) with
the input as three video frame to learn the spatiotemporal features for video
sequences. Then, we design a 3D deconvolutional network (Deconv3DNet) to
combine the spatiotemporal features to predict the final saliency map for video
sequences. Experimental results show that the proposed saliency detection model
performs better in video saliency prediction compared with the state-of-the-art
video saliency detection methods. | ['Yuming Fang', 'Guanqun Ding'] | 2018-07-12 | null | null | null | null | ['video-saliency-detection'] | ['computer-vision'] | [ 1.81107685e-01 -3.93548965e-01 -1.46475881e-01 -2.23813150e-02
-2.35945433e-01 4.55164611e-02 2.43207023e-01 -3.09089512e-01
-2.83166319e-01 6.07431114e-01 4.49774712e-01 -2.97227083e-03
2.09500864e-01 -2.65289247e-01 -8.48053694e-01 -4.69153345e-01
-1.99174017e-01 -4.77568358e-01 1.25621676e+00 -8.62183049e-02
5.88173032e-01 2.22677931e-01 -1.58782637e+00 2.01932237e-01
7.90718436e-01 1.26272607e+00 1.02528477e+00 5.17017543e-01
-3.68209630e-02 1.20331550e+00 -2.06984311e-01 4.51512992e-01
1.64551169e-01 -5.52054107e-01 -7.42523372e-01 2.29505584e-01
3.63495499e-01 -7.39896059e-01 -6.05111957e-01 1.11032450e+00
3.42392594e-01 2.24415153e-01 3.89796168e-01 -1.37910938e+00
-7.61237562e-01 4.76493478e-01 -6.76610827e-01 8.98473561e-01
4.80472952e-01 -1.34855315e-01 5.91359437e-01 -1.37769258e+00
6.26491368e-01 1.23958373e+00 4.69152808e-01 5.66348851e-01
-6.39309287e-01 -3.28784674e-01 3.44674945e-01 7.85847723e-01
-1.20814896e+00 -2.55525053e-01 1.12360418e+00 -3.93727630e-01
4.96389925e-01 1.46106258e-01 8.98944616e-01 8.60395193e-01
2.63776451e-01 1.63226378e+00 6.75758123e-01 -1.46429405e-01
1.68712929e-01 -2.20156118e-01 9.12115201e-02 7.32938528e-01
1.40267843e-02 1.32312149e-01 -8.95528793e-01 1.64290577e-01
1.20633793e+00 4.98257488e-01 -3.63909513e-01 -6.43316329e-01
-1.41030741e+00 5.45687675e-01 6.88209355e-01 2.75644660e-01
-6.62847817e-01 1.99803069e-01 2.57028371e-01 -3.94550376e-02
4.78363842e-01 6.90077096e-02 -2.98341930e-01 2.55161852e-01
-1.33508492e+00 4.87853736e-01 2.37770304e-02 1.10006845e+00
7.67540812e-01 5.43850005e-01 -5.95507801e-01 3.43235642e-01
4.07074630e-01 4.96998668e-01 6.32048488e-01 -8.01213562e-01
5.52886724e-02 5.60654998e-01 4.13082033e-01 -1.08369339e+00
-3.78178716e-01 -2.23958954e-01 -5.01879334e-01 2.53698498e-01
-3.14047076e-02 -1.05737776e-01 -1.03585410e+00 1.29653251e+00
2.74627984e-01 8.46406639e-01 1.63923249e-01 1.73869026e+00
1.44018555e+00 6.86572134e-01 2.35188175e-02 -7.40175471e-02
1.01843286e+00 -1.30303884e+00 -6.22430086e-01 -2.29250193e-01
-5.10952584e-02 -6.46609902e-01 6.04502141e-01 -2.79027939e-01
-1.26523483e+00 -8.60787630e-01 -9.70843434e-01 -2.69965947e-01
-9.78638139e-03 1.34173721e-01 3.96552384e-01 -1.47010341e-01
-1.25977397e+00 2.49593228e-01 -7.91151464e-01 -2.92926311e-01
6.66385233e-01 2.46315956e-01 -3.43195908e-02 1.51786163e-01
-1.33628285e+00 9.53798831e-01 7.11041152e-01 6.12456873e-02
-1.68811834e+00 -6.49118006e-01 -1.11386216e+00 2.71437049e-01
4.31624442e-01 -5.71299911e-01 1.27622831e+00 -1.27327669e+00
-1.04503059e+00 5.96611381e-01 -5.64332604e-01 -6.98071182e-01
1.16197519e-01 -3.72041464e-01 -1.28288776e-01 6.18253708e-01
3.64558846e-01 1.05766737e+00 1.19626260e+00 -1.10957444e+00
-9.97490108e-01 1.56624377e-01 4.19912722e-05 4.26213831e-01
-2.35699862e-02 4.97034937e-01 -3.97132337e-01 -7.70224333e-01
2.03973144e-01 -4.99933332e-01 -2.88993388e-01 2.03784078e-01
-1.84046239e-01 -8.46896768e-02 1.56388414e+00 -8.90613139e-01
9.42575157e-01 -2.13615894e+00 3.52829635e-01 -4.22300816e-01
4.99139369e-01 4.76370573e-01 -8.02694932e-02 -1.71765700e-01
3.14105377e-02 -3.35530490e-01 -2.66346425e-01 -1.46114498e-01
-2.91911334e-01 -3.21067423e-01 -4.85411167e-01 2.32311681e-01
6.23901784e-01 1.23642194e+00 -1.23755777e+00 -7.85389662e-01
5.76214194e-01 4.34726834e-01 -5.32966137e-01 2.91164070e-01
-2.56161898e-01 3.61750931e-01 -7.78587997e-01 8.82664442e-01
7.64072895e-01 -2.51527071e-01 -3.46554607e-01 4.10691537e-02
-4.88948494e-01 1.58395112e-01 -9.19099927e-01 1.78845274e+00
3.16372037e-01 1.03604305e+00 -5.88854589e-02 -1.03675342e+00
9.87923205e-01 2.51405180e-01 4.41886336e-01 -4.84711498e-01
2.00347364e-01 2.32026324e-01 -4.60572839e-01 -6.10172331e-01
6.66486681e-01 1.11905448e-01 1.26985162e-01 7.28555098e-02
4.66089845e-01 8.32132399e-02 -1.21923395e-01 2.69104540e-01
7.76692033e-01 3.15782428e-01 1.20484471e-01 -4.82081741e-01
6.45891547e-01 1.40086055e-01 7.91541159e-01 5.76866269e-01
-6.47574484e-01 9.99986053e-01 3.74241620e-01 -6.28565192e-01
-9.04154539e-01 -1.13202739e+00 3.78704488e-01 9.35330153e-01
1.04280198e+00 1.35680977e-02 -8.14348161e-01 -7.14805543e-01
-1.64446488e-01 3.42715323e-01 -6.19392455e-01 -2.06685990e-01
-6.18425727e-01 -1.21033780e-01 -9.81909707e-02 6.39746010e-01
8.50636125e-01 -1.43101096e+00 -1.15048945e+00 1.40343949e-01
-3.33876401e-01 -1.05352366e+00 -8.84236991e-01 -1.67972460e-01
-8.68038714e-01 -9.84204054e-01 -1.04297483e+00 -1.57125974e+00
4.80276644e-01 1.21066856e+00 7.86930561e-01 3.28769684e-01
-8.84071887e-02 1.51506856e-01 -4.60272759e-01 -6.09353006e-01
2.55553931e-01 -1.00556634e-01 1.35328770e-01 -2.72833649e-02
4.88945276e-01 -3.38097572e-01 -7.32113242e-01 2.22802237e-01
-9.28614378e-01 5.25410056e-01 5.13686299e-01 6.90628886e-01
5.34280717e-01 -3.20788056e-01 6.76235557e-01 4.57868632e-03
2.60417074e-01 -5.79156458e-01 -6.45682335e-01 1.25340283e-01
1.57011092e-01 -2.21476614e-01 2.62930483e-01 -3.50026160e-01
-9.62619424e-01 3.63953292e-01 1.85328275e-01 -1.07631814e+00
-1.65998891e-01 4.03108716e-01 1.17635921e-01 -2.64656752e-01
2.87067831e-01 8.24166000e-01 -4.04085442e-02 -4.55010444e-01
-3.49156037e-02 5.17901957e-01 5.67435741e-01 8.44993070e-02
7.57636070e-01 3.64908338e-01 -2.30954468e-01 -8.48467350e-01
-1.18534720e+00 -5.25249302e-01 -9.44849491e-01 -5.18598616e-01
1.21366405e+00 -1.30044413e+00 -3.66203457e-01 6.67800128e-01
-1.23399174e+00 -2.87745923e-01 -7.91171566e-03 6.60292327e-01
-7.02538908e-01 3.78280580e-01 -5.58749139e-01 -5.56881189e-01
-4.51708734e-01 -1.19189775e+00 1.27244163e+00 7.08406925e-01
1.54809222e-01 -9.39511418e-01 -3.68650585e-01 4.27821577e-02
3.95663798e-01 2.84084320e-01 2.70025104e-01 -2.32400268e-01
-1.19071913e+00 3.65754187e-01 -5.93110681e-01 7.57175311e-02
2.53719091e-01 -2.71464199e-01 -5.74175000e-01 2.76250736e-04
2.14848876e-01 -5.76943792e-02 1.14257407e+00 9.14403200e-01
1.06328511e+00 -1.01741023e-01 -5.63864410e-01 5.55157423e-01
1.24452055e+00 2.07707405e-01 5.30871928e-01 3.64558995e-01
8.36831391e-01 1.99500144e-01 1.02445734e+00 2.38337010e-01
3.67291868e-01 5.41444242e-01 5.51613331e-01 -3.75111878e-01
-3.00399750e-01 -4.66842353e-01 4.12113130e-01 5.79175532e-01
4.23723869e-02 2.92235255e-01 -5.67721248e-01 9.90011454e-01
-2.23036838e+00 -1.22021341e+00 -9.18946564e-02 1.71086586e+00
6.52484953e-01 2.38222539e-01 2.13070780e-01 -1.51521802e-01
1.04688978e+00 2.99025059e-01 -7.21861064e-01 9.08094421e-02
-3.56072426e-01 -1.22002609e-01 2.36382529e-01 3.38214695e-01
-1.49120748e+00 1.40546405e+00 6.90073538e+00 5.99803150e-01
-1.31124961e+00 1.59596846e-01 3.02427858e-01 -1.05035678e-01
-1.85257196e-01 -7.61756673e-02 -7.09834337e-01 7.07162261e-01
3.22436154e-01 -2.58090943e-01 1.14786670e-01 9.49280202e-01
4.80394393e-01 -2.50938624e-01 -7.30951071e-01 9.70280886e-01
3.87152612e-01 -1.71342874e+00 1.57634690e-01 -5.99756420e-01
1.03647542e+00 1.06451571e-01 8.70139152e-02 9.33131762e-03
2.77197063e-02 -7.45623469e-01 1.17345226e+00 7.60181010e-01
2.31404752e-01 -5.21420956e-01 6.31740510e-01 2.55511194e-01
-1.26463568e+00 -1.43847346e-01 -5.99685967e-01 -2.73726493e-01
4.63773578e-01 3.71917158e-01 -7.47182012e-01 1.97796211e-01
1.00970614e+00 1.34845519e+00 -5.11412323e-01 1.72433734e+00
-2.18576163e-01 4.57901716e-01 1.32605523e-01 -2.90006906e-01
4.81259942e-01 1.98717974e-02 8.81481826e-01 1.02894747e+00
4.22868282e-01 2.20371813e-01 3.15746546e-01 1.03194988e+00
9.94573459e-02 -2.61900008e-01 -5.30313611e-01 2.64837325e-01
2.83732414e-01 1.04246330e+00 -6.85083926e-01 -5.18673897e-01
-5.81763029e-01 1.14584839e+00 1.57644153e-01 4.87115443e-01
-9.43209887e-01 -3.67343068e-01 7.73029089e-01 1.14730131e-02
9.78351951e-01 -3.12141627e-01 -4.87284772e-02 -1.33353794e+00
-6.68360814e-02 -3.94361764e-01 1.07960589e-01 -1.40793955e+00
-7.12996185e-01 3.66412431e-01 -4.84380312e-02 -1.45684552e+00
3.89594249e-02 -4.69404221e-01 -8.12929630e-01 7.97890902e-01
-1.91181016e+00 -1.14643204e+00 -3.56012583e-01 7.82708466e-01
1.08088875e+00 -2.46380240e-01 7.65376687e-02 -1.21138850e-02
-2.57512808e-01 -1.16302080e-01 -6.62864894e-02 2.66536444e-01
3.14348876e-01 -9.00427580e-01 6.15132451e-01 1.23653996e+00
-1.92968454e-02 1.85267374e-01 7.33200252e-01 -7.29697645e-01
-1.28803909e+00 -1.34758329e+00 8.30811679e-01 -3.11631441e-01
4.18255031e-01 -2.82439083e-01 -9.28198695e-01 6.66489542e-01
4.06379253e-01 1.31685525e-01 1.18320696e-02 -5.91372967e-01
2.14744166e-01 1.59797892e-01 -9.68932927e-01 7.31360018e-01
1.28589582e+00 -3.79144460e-01 -1.10875010e+00 -3.83194983e-02
1.34883678e+00 -5.71938097e-01 -2.41264194e-01 4.83490795e-01
2.22466901e-01 -8.36266279e-01 1.08593214e+00 -5.44769466e-01
6.50267959e-01 -8.99774551e-01 -2.50850748e-02 -1.11883068e+00
-5.14768779e-01 -3.81010205e-01 -3.31458747e-01 7.03522265e-01
-6.92289919e-02 -4.47394177e-02 7.66968668e-01 2.02652663e-01
-5.17037153e-01 -5.66981196e-01 -7.60281503e-01 -3.23886812e-01
-5.95548272e-01 -5.39347646e-04 3.79463047e-01 6.20199621e-01
-1.28687704e-02 2.51402318e-01 -7.12451875e-01 2.22344831e-01
7.17749655e-01 5.71942747e-01 3.23989183e-01 -1.16446030e+00
3.90331775e-01 -3.77009183e-01 -7.81630933e-01 -1.52719951e+00
2.51575351e-01 -5.68951905e-01 2.81428128e-01 -1.58076811e+00
5.38563609e-01 1.11542545e-01 -6.94862068e-01 3.31236959e-01
-6.47101104e-01 2.92977750e-01 3.72462541e-01 -2.28554127e-03
-1.26144266e+00 9.36043859e-01 1.49160492e+00 -2.54028827e-01
-3.28228861e-01 -1.26946628e-01 -6.64070368e-01 5.44598341e-01
6.84741259e-01 -1.76541626e-01 -2.91990101e-01 -5.57761669e-01
-5.39611816e-01 2.06726506e-01 8.74202430e-01 -1.14448762e+00
2.86134094e-01 -3.70953798e-01 7.75803208e-01 -1.22697115e+00
2.24474370e-01 -6.37923956e-01 -3.13857973e-01 4.02315229e-01
-2.68748075e-01 -1.96799070e-01 3.01350564e-01 3.88180047e-01
-4.83033955e-01 -1.67228833e-01 5.74258089e-01 -2.09756374e-01
-1.52640200e+00 4.35911804e-01 -6.55290961e-01 -1.10197298e-01
1.09404683e+00 -3.79430532e-01 -1.65568769e-01 -2.37803012e-01
-5.68186939e-01 4.83352661e-01 4.13942635e-01 8.86824667e-01
1.39035869e+00 -1.59322917e+00 -7.01918244e-01 2.51136422e-01
-2.09293500e-01 7.74284359e-03 3.97105545e-01 8.22811007e-01
-5.00628471e-01 5.82675874e-01 -7.35560298e-01 -7.91621506e-01
-1.25599265e+00 9.88040805e-01 3.32689136e-01 5.81559181e-01
-5.47812521e-01 8.89167130e-01 5.95266998e-01 7.55422190e-02
1.84553310e-01 -4.77667570e-01 -5.86040854e-01 -2.88800091e-01
8.06383610e-01 7.52203465e-02 -5.35171986e-01 -7.83103049e-01
-5.27867794e-01 4.58133161e-01 7.44583681e-02 2.73148775e-01
1.44690311e+00 -3.87526572e-01 -1.56679600e-01 2.96118110e-01
1.01430619e+00 -5.78046799e-01 -1.86902833e+00 -4.00795490e-01
5.23220468e-03 -7.07499743e-01 1.62144884e-01 -2.32329100e-01
-1.14947069e+00 8.10277939e-01 6.89364552e-01 -3.29792313e-02
1.15603650e+00 1.01247527e-01 9.98517632e-01 -4.28901874e-02
4.15466219e-01 -9.12233472e-01 2.50539243e-01 5.98379552e-01
1.01733649e+00 -1.51591158e+00 -1.21070504e-01 -4.21190351e-01
-8.18604231e-01 9.23928499e-01 8.60064626e-01 -4.86179739e-01
7.94054866e-01 -2.88187295e-01 -6.15548529e-02 -1.36363015e-01
-6.13197982e-01 -6.09025300e-01 5.39045811e-01 6.60339952e-01
1.28985003e-01 -1.85344011e-01 -1.61840051e-01 6.23365879e-01
3.39046746e-01 2.78206319e-01 6.52081192e-01 8.55250001e-01
-1.03663564e+00 -4.58741665e-01 -3.26137006e-01 3.17847401e-01
-3.08855295e-01 -3.68174881e-01 -2.30832502e-01 2.92210162e-01
3.29234041e-02 7.71069527e-01 1.99323684e-01 -4.08096731e-01
7.33885989e-02 -2.58136004e-01 2.35031351e-01 -6.15691006e-01
-1.29882500e-01 1.45507351e-01 -4.27241176e-01 -4.60996389e-01
-8.25540662e-01 -8.97715747e-01 -1.46563232e+00 2.18892008e-01
-2.78134435e-01 1.29869148e-01 3.02828819e-01 9.60306644e-01
4.55819368e-01 5.58581233e-01 6.54984415e-01 -1.44488871e+00
6.35103509e-02 -7.98038661e-01 -6.74723744e-01 2.06799105e-01
9.32741284e-01 -1.11849892e+00 -4.73241024e-02 3.14641774e-01] | [9.741941452026367, -0.3159821629524231] |
a259b78a-f04c-498b-b3bb-f609acbc0e04 | hierarchical-neural-program-synthesis | 2303.06018 | null | https://arxiv.org/abs/2303.06018v1 | https://arxiv.org/pdf/2303.06018v1.pdf | Hierarchical Neural Program Synthesis | Program synthesis aims to automatically construct human-readable programs that satisfy given task specifications, such as input/output pairs or demonstrations. Recent works have demonstrated encouraging results in a variety of domains, such as string transformation, tensor manipulation, and describing behaviors of embodied agents. Most existing program synthesis methods are designed to synthesize programs from scratch, generating a program token by token, line by line. This fundamentally prevents these methods from scaling up to synthesize programs that are longer or more complex. In this work, we present a scalable program synthesis framework that instead synthesizes a program by hierarchically composing programs. Specifically, we first learn a task embedding space and a program decoder that can decode a task embedding into a program. Then, we train a high-level module to comprehend the task specification (e.g., input/output pairs or demonstrations) from long programs and produce a sequence of task embeddings, which are then decoded by the program decoder and composed to yield the synthesized program. We extensively evaluate our proposed framework in a string transformation domain with input/output pairs. The experimental results demonstrate that the proposed framework can synthesize programs that are significantly longer and more complex than the programs considered in prior program synthesis works. Website at https://thoughtp0lice.github.io/hnps_web/ | ['Shao-Hua Sun', 'Joseph J. Lim', 'Jesse Zhang', 'Ryan Lindeborg', 'Linghan Zhong'] | 2023-03-09 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 4.59784269e-01 6.13517910e-02 -1.44433603e-01 -4.54002976e-01
-4.72403824e-01 -9.23659980e-01 5.59591234e-01 4.75374088e-02
-1.76526442e-01 3.40593815e-01 6.14652038e-02 -5.90877235e-01
4.27138358e-01 -1.03674936e+00 -1.21817899e+00 -4.20581490e-01
2.16804847e-01 1.14088237e-01 1.23261034e-01 -1.09051019e-01
2.29423076e-01 7.52932718e-03 -1.63257992e+00 2.38756657e-01
1.08272421e+00 2.71788478e-01 5.43496490e-01 1.06665325e+00
-1.38665602e-01 8.99876714e-01 -5.04231691e-01 -3.12409818e-01
-1.00642465e-01 -5.15572131e-01 -7.64560163e-01 -1.80344984e-01
9.52078700e-02 -6.47205591e-01 -4.44283634e-01 1.23805201e+00
1.77526116e-01 1.12153515e-02 3.41778547e-01 -1.56408489e+00
-1.14935255e+00 9.83912706e-01 1.78531762e-02 -6.04005814e-01
6.76964879e-01 8.14858377e-01 1.09092605e+00 -7.27236748e-01
4.58537847e-01 1.28683484e+00 2.66461641e-01 8.83000076e-01
-1.34968400e+00 -6.87060833e-01 4.02253158e-02 -3.74761581e-01
-9.64693427e-01 -2.03556314e-01 6.90109968e-01 -7.92845786e-01
1.08989704e+00 2.30165184e-01 6.20682538e-01 1.21960771e+00
2.95838773e-01 9.53964770e-01 7.59324551e-01 -1.36711210e-01
1.94351450e-01 2.75263116e-02 4.82079536e-01 1.06522226e+00
1.51404247e-01 2.94618905e-01 -1.43131554e-01 -3.77504319e-01
9.77699459e-01 3.41827013e-02 -2.62855113e-01 -4.84890848e-01
-1.90475833e+00 7.58395612e-01 4.62731451e-01 7.50555843e-02
-2.15962753e-01 8.29986095e-01 8.45498323e-01 5.50579786e-01
-2.31183156e-01 5.29205203e-01 -1.54050499e-01 -1.60593763e-01
-5.86679220e-01 8.67230952e-01 1.10338962e+00 1.47555327e+00
6.67770803e-01 3.21976036e-01 -3.07381183e-01 3.99572015e-01
2.92709231e-01 7.40650177e-01 3.89637619e-01 -8.00555289e-01
7.28989542e-01 8.04386675e-01 2.06003860e-01 -9.84091401e-01
-2.71218330e-01 1.13594241e-01 -6.49264157e-01 2.25055799e-01
7.94937760e-02 -4.03651834e-01 -6.49951816e-01 1.91187131e+00
2.13688344e-01 -1.26887247e-01 3.39744210e-01 8.70541215e-01
1.01598585e+00 1.06668305e+00 -1.76174901e-02 3.05844516e-01
1.35352600e+00 -1.57661104e+00 -4.90762264e-01 -2.14888334e-01
9.83174920e-01 -4.55672264e-01 1.68722224e+00 5.56486286e-02
-1.02814305e+00 -8.54784906e-01 -1.04288411e+00 -1.86951280e-01
-1.59160316e-01 3.22170019e-01 7.96989858e-01 3.36891532e-01
-1.05491114e+00 4.31868851e-01 -1.05942941e+00 -2.51923561e-01
1.98248789e-01 4.46225554e-01 -1.79050460e-01 1.45375982e-01
-7.30472565e-01 5.07285953e-01 8.09388697e-01 -1.11774355e-01
-1.50601315e+00 -5.96938491e-01 -1.36706424e+00 1.63307846e-01
2.51438677e-01 -1.07936800e+00 1.63644707e+00 -1.08270288e+00
-1.78560019e+00 5.73206425e-01 4.41211946e-02 -2.34207347e-01
3.45147371e-01 8.75387788e-02 -2.03500152e-01 -3.02369624e-01
-3.84401381e-02 6.66007519e-01 7.25949883e-01 -1.03703070e+00
-5.44043422e-01 7.16470182e-02 6.44499242e-01 -5.06019965e-03
-2.67469943e-01 2.56435692e-01 -1.38869062e-01 -6.24712348e-01
-4.23049986e-01 -1.22589684e+00 -1.41159445e-01 1.21692613e-01
-6.08578563e-01 -3.29940498e-01 6.54188871e-01 -5.43535054e-01
1.07876909e+00 -2.27704668e+00 8.25110674e-01 -1.96175814e-01
3.32217753e-01 9.32165086e-02 -4.83846456e-01 7.28844762e-01
1.68056469e-02 9.58838165e-02 -5.03475666e-01 -3.05627584e-02
5.54740548e-01 2.51841754e-01 -4.31083590e-01 2.62479067e-01
3.50829303e-01 1.17036450e+00 -1.28487980e+00 -2.53641903e-01
8.10300931e-03 2.02299327e-01 -1.01823854e+00 6.58967853e-01
-9.06770587e-01 5.61763942e-01 -7.79671907e-01 3.58017027e-01
2.90284723e-01 -1.92260653e-01 -4.81094122e-02 -8.09974521e-02
-2.37254634e-01 3.99545610e-01 -8.45665574e-01 1.98062420e+00
-8.46269667e-01 6.08548462e-01 -9.03760791e-02 -8.78617287e-01
1.04094172e+00 3.91601235e-01 -1.88087806e-01 -1.00346550e-01
1.03449628e-01 2.32331261e-01 1.12317041e-01 -9.19190228e-01
4.78465647e-01 1.57532707e-01 -6.66328311e-01 7.39656031e-01
1.81949846e-02 -8.29496861e-01 2.91713119e-01 7.98457190e-02
1.16796494e+00 5.99385202e-01 2.82474667e-01 -8.21832567e-02
6.00882590e-01 1.93034023e-01 2.49022588e-01 6.77719653e-01
1.41213089e-01 2.78145462e-01 8.99668455e-01 -4.82577115e-01
-1.51985168e+00 -8.76342177e-01 5.41504145e-01 1.33183527e+00
1.70067966e-03 -4.76147801e-01 -9.03077662e-01 -5.79470098e-01
-1.70685612e-02 8.92279029e-01 -4.15752023e-01 -2.70762891e-01
-8.66112828e-01 -1.28199413e-01 8.35866928e-01 5.66875160e-01
5.03433526e-01 -1.46370590e+00 -9.07060266e-01 2.12964788e-01
2.45499257e-02 -8.46049607e-01 -8.88767838e-01 -1.39301121e-01
-6.82102382e-01 -8.95795166e-01 -6.14787042e-01 -1.47110808e+00
1.10588706e+00 6.80941576e-03 9.31185544e-01 3.10878545e-01
2.02285707e-01 1.15872264e-01 -2.11684585e-01 -4.60626073e-02
-9.22714114e-01 2.41474539e-01 -9.00543183e-02 -2.49999493e-01
-2.19718501e-01 -6.10561192e-01 -2.90958107e-01 -3.08207963e-02
-1.17159235e+00 7.78990328e-01 5.71104765e-01 1.03866327e+00
1.98541149e-01 -7.17109516e-02 5.00003517e-01 -1.02250254e+00
1.04352546e+00 -2.73837328e-01 -9.95761931e-01 2.29572594e-01
1.18560297e-02 4.12730247e-01 1.21292710e+00 -7.31821120e-01
-8.74214351e-01 2.84362674e-01 -6.24669306e-02 -2.64551848e-01
-1.28611878e-01 6.31949008e-01 -1.44713208e-01 4.34250683e-02
8.05832326e-01 7.20115423e-01 -6.08888641e-02 1.47674531e-02
5.68163335e-01 6.99660659e-01 6.16904914e-01 -1.17286444e+00
1.20299661e+00 -1.15879759e-01 -4.83750761e-01 -2.65941650e-01
-8.04501250e-02 2.91623712e-01 -4.18733954e-01 2.34314412e-01
7.67166734e-01 -6.67330027e-01 -8.95258546e-01 5.91397405e-01
-1.49814379e+00 -9.77506936e-01 -3.47541906e-02 2.12431923e-01
-9.22475934e-01 7.08479341e-03 -6.31463289e-01 -3.16055745e-01
-4.64218944e-01 -1.78203917e+00 1.18283153e+00 2.76052892e-01
-4.73922580e-01 -7.82987356e-01 1.49441496e-01 -1.99651137e-01
3.94067705e-01 4.51687902e-01 1.23904860e+00 -5.26893914e-01
-7.29028344e-01 -1.33235529e-01 -1.43864855e-01 3.07544380e-01
2.18417004e-01 3.85514617e-01 -4.08300668e-01 -2.93918282e-01
-2.54669517e-01 -4.04425204e-01 1.06272206e-01 -2.50239670e-01
1.23230386e+00 -7.91676104e-01 -2.56086797e-01 7.77268052e-01
1.15339923e+00 3.57344896e-01 3.87403160e-01 5.75039163e-02
8.71098220e-01 3.91352922e-01 2.91174144e-01 2.52482176e-01
5.71148753e-01 4.71355051e-01 3.04084718e-01 2.45777234e-01
7.96669070e-03 -7.18365908e-01 8.10913086e-01 9.53492224e-01
3.64177763e-01 7.67081007e-02 -1.01989841e+00 6.33084118e-01
-1.89369881e+00 -6.71680927e-01 1.35436729e-01 1.77537537e+00
1.18731475e+00 3.12263817e-02 2.43196145e-01 -2.15393245e-01
5.79215944e-01 1.43639177e-01 -7.14389980e-01 -7.32215166e-01
5.63562334e-01 1.97927937e-01 1.39211327e-01 4.16295648e-01
-8.00364435e-01 1.16635072e+00 5.44728899e+00 3.25214863e-01
-1.23211551e+00 -8.76031891e-02 -9.18888580e-03 1.22335851e-01
-7.17058361e-01 1.50126383e-01 -3.81464154e-01 6.06039584e-01
8.05922270e-01 -8.67118537e-01 8.17894101e-01 1.04624784e+00
4.92370911e-02 4.28461850e-01 -1.81037915e+00 6.85263157e-01
-1.41925901e-01 -1.35991526e+00 9.25239399e-02 -5.28554797e-01
7.23828793e-01 -2.89790839e-01 8.14817101e-02 8.30201805e-01
7.17775345e-01 -9.72310483e-01 1.02622676e+00 1.99480072e-01
7.85434246e-01 -5.36856055e-01 2.82008052e-01 5.21484613e-01
-1.23155987e+00 -1.24239117e-01 -7.99333900e-02 -1.51549652e-01
-9.61298719e-02 5.23523390e-02 -6.48830116e-01 2.54887760e-01
2.26761520e-01 7.14036584e-01 -4.57953751e-01 7.89910972e-01
-6.23987496e-01 3.14486712e-01 1.93697773e-02 -6.16871536e-01
2.61851996e-01 -5.03657609e-02 4.03543979e-01 1.20736432e+00
4.93317455e-01 1.14021294e-01 2.71329552e-01 1.62952840e+00
-2.20557079e-01 -6.02755286e-02 -9.78225887e-01 -6.38936877e-01
4.38786000e-01 1.14689469e+00 -3.07114452e-01 -6.13952041e-01
-6.25245154e-01 1.03414154e+00 3.53914648e-01 4.83737767e-01
-1.36826622e+00 -9.85097229e-01 5.81562161e-01 -4.00213301e-01
5.41388318e-02 -5.91485798e-01 -2.34983444e-01 -1.45073676e+00
1.45435914e-01 -1.38947105e+00 -1.72120050e-01 -9.24433231e-01
-6.23051405e-01 6.80874825e-01 7.66191110e-02 -1.07034767e+00
-2.48853356e-01 -3.91761720e-01 -9.20677006e-01 8.06063354e-01
-9.26984310e-01 -1.27582979e+00 -3.40383589e-01 2.98479259e-01
6.95150077e-01 -1.43092528e-01 9.32938397e-01 2.15433866e-01
-7.72306681e-01 7.36606598e-01 -3.14265013e-01 3.56071562e-01
1.69684604e-01 -1.14529347e+00 1.00612509e+00 8.10507596e-01
-3.15303504e-01 9.74167407e-01 7.79913008e-01 -4.95018780e-01
-2.17099190e+00 -1.34246254e+00 5.13177454e-01 -3.47719431e-01
9.98395026e-01 -6.89685404e-01 -7.93655932e-01 1.11914623e+00
2.25083366e-01 -1.21801466e-01 1.05041429e-01 -5.29290140e-01
-6.50648594e-01 2.06672415e-01 -9.56856787e-01 1.03886712e+00
1.07331383e+00 -8.69497895e-01 -6.99818015e-01 4.12447661e-01
1.42403769e+00 -8.78938973e-01 -9.13237393e-01 1.03473425e-01
5.85427046e-01 -3.73146147e-01 7.06507862e-01 -6.26164794e-01
9.93309975e-01 -5.61604321e-01 1.28717162e-02 -1.45229185e+00
2.09466950e-03 -7.65885830e-01 2.09434275e-02 1.21046233e+00
4.50487226e-01 -8.30860376e-01 4.72496927e-01 7.45628953e-01
-4.61464733e-01 -5.52548349e-01 -1.63927481e-01 -6.56844795e-01
2.57659018e-01 -1.43694088e-01 1.06988013e+00 7.70942748e-01
3.74111712e-01 2.74751067e-01 -1.75434381e-01 2.63657182e-01
3.76516968e-01 6.93414748e-01 1.21796620e+00 -5.62221348e-01
-7.04234838e-01 -6.32539451e-01 -6.27823696e-02 -1.25794101e+00
5.82430542e-01 -1.36525059e+00 3.21738690e-01 -1.40699399e+00
1.81118593e-01 -4.21399891e-01 4.00940895e-01 7.42592037e-01
-1.64681986e-01 -5.61683655e-01 2.56166935e-01 9.02429149e-02
-2.45667726e-01 5.81323802e-01 1.51282406e+00 -5.29109895e-01
-2.28702947e-01 -2.06024393e-01 -6.44690454e-01 4.24798816e-01
8.49793613e-01 -6.05423152e-01 -5.23284316e-01 -1.00826204e+00
3.55489105e-01 4.87562895e-01 5.66068947e-01 -9.24345911e-01
3.09287757e-01 -5.50489366e-01 -4.44302976e-01 -8.73729587e-02
-1.65173352e-01 -7.27362096e-01 3.44806820e-01 8.09917152e-01
-8.24956536e-01 3.19489270e-01 2.42040217e-01 1.29608780e-01
-1.10546604e-01 -3.91080827e-01 4.90150750e-01 -1.15147896e-01
-6.07861459e-01 3.30776840e-01 -2.60233253e-01 -2.17108741e-01
1.13988400e+00 2.35227749e-01 -5.81456661e-01 -1.33689925e-01
-5.24275362e-01 4.90211695e-01 8.18088293e-01 6.81860149e-01
6.02537155e-01 -1.52652740e+00 -6.57019079e-01 3.48140717e-01
3.67157221e-01 2.40239277e-01 -5.81491701e-02 5.33250391e-01
-9.04131293e-01 3.40142846e-01 -3.66409004e-01 -4.47970092e-01
-1.17184961e+00 8.03700268e-01 2.41136625e-01 2.85413992e-02
-6.23411298e-01 5.43963373e-01 4.01930571e-01 -1.20824575e+00
2.15247542e-01 -1.13836622e+00 2.51117021e-01 -7.12704837e-01
5.24500072e-01 -1.29863814e-01 -5.70641279e-01 -2.34586716e-01
-6.69112206e-02 5.37315965e-01 1.61837950e-01 2.01958697e-02
1.35744417e+00 5.81713974e-01 -5.14996827e-01 2.56646872e-01
1.37830579e+00 -4.05343831e-01 -1.17952061e+00 -2.36238107e-01
-1.13354988e-01 -2.78365076e-01 -4.72204477e-01 -2.20775992e-01
-9.35556531e-01 8.76568675e-01 2.86754575e-02 1.87616169e-01
8.60991299e-01 -8.44525769e-02 9.64967191e-01 7.63908744e-01
5.31454980e-01 -4.90503818e-01 3.89692008e-01 6.96205199e-01
1.09083462e+00 -9.47109044e-01 -4.43761945e-01 -2.80013174e-01
-5.99133492e-01 1.22510231e+00 6.95588946e-01 -5.07996619e-01
1.24802768e-01 6.29475713e-01 -5.70149064e-01 -9.57078487e-02
-8.86421382e-01 1.92078888e-01 -9.52622071e-02 7.13713944e-01
5.24352849e-01 3.23955804e-01 -1.57624722e-01 7.58127093e-01
-4.79288876e-01 4.55961764e-01 7.57962584e-01 1.11262870e+00
-2.52687931e-01 -1.18783557e+00 -9.48070139e-02 3.38267952e-01
1.11162402e-01 -1.46452989e-02 -1.14023924e-01 6.17035747e-01
-1.58318639e-01 5.01838565e-01 -1.53376728e-01 -6.96804941e-01
4.93303597e-01 1.93955265e-02 5.01574159e-01 -1.05893636e+00
-8.17492545e-01 -5.80945313e-01 1.57901585e-01 -4.93391067e-01
-6.94944710e-02 -3.81848335e-01 -1.59442770e+00 -4.97858673e-01
1.02715053e-01 -6.55153170e-02 4.66479450e-01 4.01838958e-01
2.08870247e-01 8.67677808e-01 5.10458112e-01 -9.26633835e-01
-6.92690790e-01 -4.70780283e-01 5.69742359e-02 4.55586046e-01
5.28097391e-01 -1.82106897e-01 -6.58480152e-02 4.80430871e-01] | [7.8884758949279785, 7.711095333099365] |
a79ada20-ea00-4171-a7cc-042dfa743e17 | shape-from-texture-using-locally-scaled-point | 1311.7401 | null | http://arxiv.org/abs/1311.7401v1 | http://arxiv.org/pdf/1311.7401v1.pdf | Shape from Texture using Locally Scaled Point Processes | Shape from texture refers to the extraction of 3D information from 2D images
with irregular texture. This paper introduces a statistical framework to learn
shape from texture where convex texture elements in a 2D image are represented
through a point process. In a first step, the 2D image is preprocessed to
generate a probability map corresponding to an estimate of the unnormalized
intensity of the latent point process underlying the texture elements. The
latent point process is subsequently inferred from the probability map in a
non-parametric, model free manner. Finally, the 3D information is extracted
from the point pattern by applying a locally scaled point process model where
the local scaling function represents the deformation caused by the projection
of a 3D surface onto a 2D image. | ['Christoph Schnörr', 'Alex Lenkoski', 'Eva-Maria Didden', 'Thordis Linda Thorarinsdottir'] | 2013-11-28 | null | null | null | null | ['shape-from-texture'] | ['computer-vision'] | [ 6.65476024e-01 4.31481212e-01 9.85279828e-02 -2.82948494e-01
-5.44908702e-01 -2.25166619e-01 9.09998775e-01 -1.70754820e-01
-4.98550273e-02 2.12059796e-01 6.56474903e-02 2.74913460e-01
-1.22777581e-01 -1.01218748e+00 -7.44166195e-01 -1.17720950e+00
2.34462991e-01 1.02295864e+00 1.54685616e-01 1.25081360e-01
3.84358972e-01 8.39633405e-01 -1.49291861e+00 2.10652292e-01
2.30843857e-01 8.99088323e-01 1.09904252e-01 4.62468266e-01
-3.73596698e-01 -1.24609314e-01 -1.58083871e-01 -1.91749901e-01
3.32503855e-01 -2.00446144e-01 -3.32020253e-01 6.88671589e-01
4.22137529e-01 -5.79717197e-02 -3.46492790e-02 1.03393757e+00
-1.09264448e-01 1.10644416e-03 1.38887978e+00 -4.26892966e-01
-4.86961156e-01 -2.90365756e-01 -8.31942320e-01 -4.01132792e-01
1.80905208e-01 -3.35126489e-01 7.31937349e-01 -1.20544600e+00
9.10608292e-01 1.69867885e+00 3.03479552e-01 2.17114151e-01
-1.80072057e+00 1.82468340e-01 -4.95313734e-01 -5.42035043e-01
-1.23016036e+00 -1.05391957e-01 1.18415511e+00 -7.23641276e-01
4.64913130e-01 7.82206208e-02 5.58925748e-01 1.03939414e+00
5.68258584e-01 6.39810383e-01 1.47028315e+00 -7.83471942e-01
3.42045158e-01 -2.01320574e-01 -1.31453738e-01 6.65458024e-01
6.26298711e-02 2.88908720e-01 -5.58653891e-01 -4.72912550e-01
1.55275905e+00 -7.74717554e-02 3.97021264e-01 -7.06504941e-01
-9.30022359e-01 4.25531030e-01 -7.69880712e-02 4.74519022e-02
-5.82770646e-01 2.16002353e-02 -1.97859168e-01 3.05986274e-02
1.15729332e+00 -7.93910325e-02 -1.71371713e-01 -9.83516276e-02
-7.87122488e-01 -1.91505048e-02 9.08039451e-01 5.18143296e-01
1.07154322e+00 -1.52558267e-01 1.51220411e-01 5.92118621e-01
6.81772470e-01 1.25096691e+00 -1.63059801e-01 -1.06712937e+00
-6.73426837e-02 7.15608478e-01 6.31615669e-02 -1.12294281e+00
1.08857028e-01 1.91618696e-01 -7.15352058e-01 6.62649930e-01
5.11468112e-01 5.18131912e-01 -1.16041279e+00 1.29209828e+00
3.04613888e-01 2.50228029e-02 -1.91342831e-01 5.12594342e-01
1.54159546e-01 5.66364765e-01 -4.64300245e-01 -1.42426968e-01
1.27991009e+00 4.43329141e-02 -5.97828567e-01 1.41701415e-01
-2.45865941e-01 -8.83018672e-01 9.05002236e-01 3.26966792e-01
-9.74479675e-01 -4.17694241e-01 -6.03315711e-01 1.58504713e-02
7.55402222e-02 1.19461074e-01 -1.33690134e-01 3.07289213e-01
-5.31933367e-01 5.59155047e-01 -1.35671926e+00 -3.22756857e-01
4.89005834e-01 1.06838733e-01 -5.86550474e-01 2.14340180e-01
-3.67275059e-01 8.54461670e-01 -1.96742460e-01 -1.38323279e-02
-5.00220239e-01 -2.15963498e-01 -7.93218911e-01 -1.91288546e-01
5.17013706e-02 -6.44930303e-01 6.12151682e-01 -7.23742902e-01
-1.88018537e+00 1.37942064e+00 -6.93958342e-01 1.46990791e-01
4.64778751e-01 -1.18884191e-01 2.00926900e-01 2.32663959e-01
1.23990074e-01 1.96261898e-01 1.59571433e+00 -1.73547328e+00
-7.51892626e-02 -6.95095301e-01 -6.86680079e-01 2.57094055e-01
3.80720913e-01 -4.06528562e-01 -4.99134928e-01 -6.74405217e-01
9.65272009e-01 -9.82439756e-01 1.10592879e-01 -3.59530114e-02
-4.56323653e-01 -1.81320272e-02 9.42232549e-01 -5.19778967e-01
4.80196834e-01 -2.49696326e+00 4.28583771e-01 6.76811218e-01
2.46953279e-01 -6.90344453e-01 -2.15819050e-02 1.92471504e-01
1.97423369e-01 -5.87277077e-02 -5.32463968e-01 -4.54565078e-01
9.74158123e-02 3.06158185e-01 -3.78491908e-01 6.87207639e-01
4.89177614e-01 6.54239118e-01 -6.19916558e-01 -2.04842925e-01
3.71825665e-01 5.70049763e-01 -2.05644608e-01 2.21983232e-02
-3.11637342e-01 5.88525414e-01 -6.33980215e-01 5.79872966e-01
8.49980354e-01 -2.76442338e-02 -8.82903486e-02 -2.81703293e-01
-2.60519862e-01 -1.81947142e-01 -1.01720381e+00 1.25728989e+00
-2.67554939e-01 3.97678941e-01 -1.18352078e-01 -6.27283156e-01
1.57936561e+00 3.93664807e-01 6.18614197e-01 -2.39766598e-01
4.37617972e-02 3.38824958e-01 -3.81127834e-01 -3.47931921e-01
2.54952788e-01 -5.00326574e-01 -8.97965506e-02 5.51879883e-01
-7.97915757e-02 -8.35537314e-01 -4.72742468e-01 -4.42344755e-01
7.84452498e-01 6.45004809e-01 9.87970643e-03 -3.89103949e-01
2.85701603e-01 -2.16676071e-01 9.75239277e-02 5.76525986e-01
3.99026841e-01 7.41936266e-01 6.66221797e-01 -7.79491067e-01
-1.53064692e+00 -1.67020190e+00 -5.54678619e-01 1.61923721e-01
7.38288388e-02 1.22458600e-01 -7.62437820e-01 -2.74584770e-01
1.68549836e-01 4.62439895e-01 -9.65386331e-01 2.67336313e-02
-6.23391867e-01 -7.19190955e-01 -5.50902449e-02 1.06617510e-01
4.04021055e-01 -7.60213137e-01 -4.99904275e-01 3.04428395e-02
1.33478800e-02 -8.90658975e-01 -3.26402694e-01 1.07390925e-01
-1.26212323e+00 -6.94845200e-01 -5.41500628e-01 -6.08521104e-01
1.14056432e+00 -2.36439332e-01 8.60229671e-01 -4.66816723e-01
-2.81729519e-01 4.72940713e-01 6.82018772e-02 -3.41818035e-01
-5.15589535e-01 -4.88413960e-01 1.14755042e-01 7.76225090e-01
5.30506194e-01 -3.95755440e-01 -1.07792616e-01 4.46395457e-01
-6.96314931e-01 1.52880698e-01 5.10854006e-01 7.37831295e-01
1.38082826e+00 1.60169423e-01 -3.02264094e-01 -5.10917068e-01
1.92875162e-01 -1.96941532e-02 -7.14249790e-01 -2.45642029e-02
9.53884646e-02 4.11620736e-01 -2.14267045e-01 -4.52692151e-01
-1.38171840e+00 4.03969675e-01 2.89521873e-01 -5.38712144e-01
-4.48201030e-01 3.13359648e-01 -1.50594324e-01 1.92630872e-01
6.47597432e-01 3.13440830e-01 5.12456179e-01 -7.10171223e-01
2.20248342e-01 3.52561802e-01 3.36765796e-01 -8.17430437e-01
1.06036341e+00 1.31878424e+00 6.02731526e-01 -1.23879826e+00
-6.82571709e-01 -1.41853005e-01 -1.25888062e+00 -2.28150964e-01
1.09307778e+00 -4.06914204e-01 -4.72270608e-01 5.80195308e-01
-9.61718738e-01 -6.17174804e-01 -6.80331886e-01 5.93622208e-01
-1.12014806e+00 2.80353934e-01 -4.94410813e-01 -8.87753725e-01
1.20007463e-01 -9.59916890e-01 1.56297445e+00 -7.04343319e-02
-3.13024133e-01 -1.14267826e+00 4.17621255e-01 1.44768149e-01
3.62379849e-02 4.06724483e-01 1.21993279e+00 8.05316120e-02
-6.81530654e-01 -4.00068939e-01 -3.50825377e-02 3.82259786e-01
5.98307490e-01 2.88747430e-01 -8.91198754e-01 2.94780523e-01
6.57089114e-01 1.73464119e-01 4.69669402e-01 8.44020009e-01
7.60979772e-01 -1.33242290e-02 -2.43373871e-01 5.07294893e-01
1.36672020e+00 -1.82562917e-01 3.85265350e-01 1.53808445e-01
6.50217295e-01 7.61130750e-01 3.50591183e-01 1.44277781e-01
-1.76577792e-01 7.37981260e-01 1.15874834e-01 2.43472867e-02
-1.80703655e-01 -4.33145970e-01 2.01587766e-01 7.81728387e-01
-6.16390049e-01 1.70699283e-01 -9.32939887e-01 5.24326935e-02
-1.54083657e+00 -8.38747025e-01 -2.97355741e-01 2.43404531e+00
5.05984604e-01 2.12599829e-01 -3.82898301e-01 -6.39063939e-02
8.09345007e-01 -2.78397888e-01 -3.74231189e-01 -2.52455652e-01
-5.20491540e-01 3.57902735e-01 2.78395236e-01 9.40187395e-01
-8.12336922e-01 1.03251171e+00 6.59433460e+00 7.84376442e-01
-1.10459030e+00 -1.79829463e-01 4.78545427e-01 3.85315895e-01
-4.89440262e-01 1.31249398e-01 -5.79089165e-01 3.92522156e-01
5.83599627e-01 -2.80093011e-02 6.39210939e-02 4.37571973e-01
4.29949343e-01 -3.10834378e-01 -9.07800496e-01 8.53049457e-01
-2.44567338e-02 -9.45938945e-01 2.81029195e-01 7.54801333e-01
7.30294168e-01 -1.53075373e-02 2.85368949e-01 -4.72252905e-01
7.29054138e-02 -7.57150054e-01 7.84476519e-01 1.26693833e+00
1.12654352e+00 -2.88505226e-01 5.99965811e-01 4.19960737e-01
-7.82007992e-01 4.86418247e-01 -3.87439191e-01 -9.63891894e-02
3.44859570e-01 9.55916464e-01 -6.70751512e-01 1.18543588e-01
5.85967183e-01 5.22526562e-01 -1.45247594e-01 5.32753050e-01
-3.41143280e-01 5.64054966e-01 -7.29139686e-01 3.78509909e-01
-1.52200118e-01 -9.12809014e-01 9.73802984e-01 6.37964189e-01
5.86621881e-01 8.85775089e-02 -1.36056006e-01 1.07810330e+00
1.26969859e-01 -5.88227287e-02 -8.41407239e-01 3.06069374e-01
1.12876013e-01 1.17232478e+00 -1.30472612e+00 -1.43747628e-01
-1.37240008e-01 9.72993135e-01 -9.69825611e-02 3.82741451e-01
-1.01728663e-01 1.98751196e-01 1.20994247e-01 4.80756730e-01
2.71917671e-01 -6.14264965e-01 -7.59282947e-01 -9.84205425e-01
-4.53465059e-02 -2.78082997e-01 -3.56791586e-01 -1.02009273e+00
-1.56936443e+00 3.43264252e-01 1.85430184e-01 -9.69781876e-01
3.43055129e-02 -9.05766428e-01 -4.72340584e-01 1.38398349e+00
-7.02878177e-01 -1.01484418e+00 -1.67040959e-01 2.11921737e-01
2.12380543e-01 -4.14841529e-03 1.00392473e+00 -6.52791321e-01
1.32296488e-01 -3.12396139e-01 3.83450061e-01 -1.25281811e-01
5.61110735e-01 -1.55757666e+00 6.94794893e-01 4.59970087e-01
7.33879358e-02 5.35321474e-01 6.42375767e-01 -1.09977579e+00
-1.32663620e+00 -5.42309701e-01 8.69648337e-01 -1.14040017e+00
5.23244917e-01 -6.08905315e-01 -9.95031953e-01 4.59344894e-01
-4.55417067e-01 1.10216267e-01 4.46201116e-01 -1.39694437e-01
-4.32569645e-02 1.50847346e-01 -1.03052747e+00 4.30597335e-01
6.72868013e-01 -8.03651452e-01 -9.20142293e-01 -1.27044857e-01
-6.39015436e-02 -3.94690782e-01 -1.03842688e+00 4.06288743e-01
5.67225933e-01 -4.91241962e-01 8.57293189e-01 -6.16153143e-02
4.26361293e-01 -2.85143197e-01 -3.94297302e-01 -1.03277040e+00
-5.41230142e-01 -5.60196042e-01 -8.13047960e-02 6.54986262e-01
9.08052921e-02 -3.89430285e-01 1.11356664e+00 4.50332850e-01
3.19377631e-02 -5.58201730e-01 -1.24619114e+00 -6.02153420e-01
-1.30738029e-02 -2.46293917e-01 3.16122949e-01 5.29116631e-01
-4.91316825e-01 5.71596064e-02 -5.68119064e-02 1.52956650e-01
1.18257642e+00 1.97008714e-01 7.29031980e-01 -1.73894536e+00
-5.01648411e-02 -1.45113349e-01 -5.03950298e-01 -9.73614275e-01
1.34055123e-01 -7.91222811e-01 2.60750473e-01 -1.16888750e+00
2.55414933e-01 -3.47881615e-01 2.07308814e-01 6.57424480e-02
6.80418685e-02 3.37843508e-01 -2.17198923e-01 7.53005624e-01
3.20150137e-01 6.16415799e-01 1.54871213e+00 1.78519443e-01
-3.23777020e-01 1.99388757e-01 3.12094600e-03 1.13123596e+00
4.11675513e-01 -4.61744189e-01 -6.45141602e-02 -2.43023992e-01
2.33597100e-01 1.42434061e-01 5.52200079e-01 -4.00839955e-01
-2.22337902e-01 -2.46407568e-01 4.96166676e-01 -7.16628313e-01
5.30825615e-01 -8.21863294e-01 2.32603297e-01 9.58273858e-02
2.57253479e-02 -2.37184241e-01 -5.55839799e-02 8.40673625e-01
-1.41121730e-01 -1.25455633e-01 7.27994978e-01 -8.02040845e-02
-1.42827416e-02 4.35904801e-01 -3.46904457e-01 -3.55441242e-01
6.88842177e-01 -5.29236138e-01 1.66435152e-01 -5.19524626e-02
-1.08305812e+00 -8.07685316e-01 1.03910792e+00 4.69337665e-02
6.79496348e-01 -1.34267867e+00 -5.99004388e-01 7.18477190e-01
-8.51859599e-02 3.13154668e-01 2.55352527e-01 7.02576578e-01
-5.86904943e-01 -4.17684242e-02 -3.52134198e-01 -1.21031022e+00
-1.09473026e+00 1.04414031e-01 2.76745290e-01 1.37911424e-01
-9.52897310e-01 4.80168164e-01 5.67629695e-01 -5.16582668e-01
-4.17965025e-01 -2.59217024e-01 1.91818222e-01 -1.39256820e-01
1.27867416e-01 2.40045950e-01 3.81143689e-02 -1.04417527e+00
1.21129990e-01 1.26768327e+00 3.58933836e-01 -7.28851974e-01
1.43289435e+00 -1.45891011e-01 -4.13920224e-01 9.96606469e-01
1.10752380e+00 2.05224231e-01 -1.86999452e+00 -3.79295260e-01
-9.44244396e-03 -6.08578444e-01 3.68483923e-02 -2.63738155e-01
-4.59207267e-01 7.78857827e-01 4.62556243e-01 1.11688025e-01
6.24244690e-01 4.02653694e-01 3.95794101e-02 2.99188141e-02
3.19768608e-01 -1.02762353e+00 2.41731316e-01 6.59053564e-01
9.78439629e-01 -9.35757041e-01 1.02434300e-01 -5.66022158e-01
-2.15945870e-01 1.49280143e+00 -8.14324543e-02 -5.12910783e-01
1.14263284e+00 3.27272743e-01 4.97149527e-02 -6.68151438e-01
-3.55831712e-01 1.76586494e-01 6.05163395e-01 4.91898447e-01
5.53615764e-02 2.97117203e-01 2.04561397e-01 3.97694074e-02
-3.10991347e-01 -1.35652870e-01 4.49317515e-01 8.00139368e-01
-3.35456401e-01 -1.05673683e+00 -7.69994199e-01 2.60651559e-01
-2.41565555e-01 2.71379411e-01 -1.55298561e-01 3.68387759e-01
7.05365092e-03 1.32470757e-01 7.24246442e-01 -2.57469025e-02
3.57118845e-01 2.62368262e-01 9.81204212e-01 -7.64625549e-01
5.40138721e-01 7.04462647e-01 -5.08369565e-01 -2.91957438e-01
-4.64037716e-01 -9.90999341e-01 -9.80916142e-01 1.36896104e-01
-9.35395211e-02 -2.01371536e-01 1.01515388e+00 8.67812216e-01
1.30190536e-01 -1.11075796e-01 5.36419094e-01 -1.28939903e+00
-2.65498191e-01 -8.06156754e-01 -8.71016264e-01 5.23989558e-01
1.43129798e-02 -9.06646729e-01 -6.85853183e-01 5.31348646e-01] | [8.875306129455566, -3.3635849952697754] |
0fc40938-435e-43ee-84df-709ec59bcf66 | learning-semantic-representations-for-the | 1312.0482 | null | http://arxiv.org/abs/1312.0482v1 | http://arxiv.org/pdf/1312.0482v1.pdf | Learning Semantic Representations for the Phrase Translation Model | This paper presents a novel semantic-based phrase translation model. A pair
of source and target phrases are projected into continuous-valued vector
representations in a low-dimensional latent semantic space, where their
translation score is computed by the distance between the pair in this new
space. The projection is performed by a multi-layer neural network whose
weights are learned on parallel training data. The learning is aimed to
directly optimize the quality of end-to-end machine translation results.
Experimental evaluation has been performed on two Europarl translation tasks,
English-French and German-English. The results show that the new semantic-based
phrase translation model significantly improves the performance of a
state-of-the-art phrase-based statistical machine translation sys-tem, leading
to a gain of 0.7-1.0 BLEU points. | ['Wen-tau Yih', 'Xiaodong He', 'Li Deng', 'Jianfeng Gao'] | 2013-11-28 | null | null | null | null | ['learning-semantic-representations'] | ['methodology'] | [ 3.25968027e-01 -5.82190380e-02 -6.55170500e-01 -3.79722953e-01
-1.43747449e+00 -4.78066772e-01 8.48079622e-01 -4.73603494e-02
-5.01923978e-01 9.23621118e-01 4.67660546e-01 -4.25986201e-01
3.15024883e-01 -6.85180724e-01 -8.72679770e-01 -6.53220654e-01
4.02186841e-01 1.01478720e+00 -3.22897494e-01 -2.66012341e-01
1.46477476e-01 1.49029106e-01 -8.08837533e-01 4.23976570e-01
6.80117309e-01 4.20720845e-01 2.71852136e-01 4.55476463e-01
-2.10987002e-01 -4.34904575e-01 -3.07842851e-01 -5.79474270e-01
3.73409182e-01 -4.98189896e-01 -7.65862823e-01 -2.94315696e-01
8.56661052e-02 1.27808467e-01 -2.25986112e-02 1.19918203e+00
4.76018786e-01 -1.75378069e-01 7.12383151e-01 -8.63434970e-01
-1.04706752e+00 5.84358394e-01 -2.80941010e-01 6.67350441e-02
3.59229296e-01 -1.98865682e-01 1.41236103e+00 -1.48381567e+00
5.64358652e-01 1.23218679e+00 4.70478475e-01 5.49389124e-01
-1.61815965e+00 -4.28837746e-01 -3.94697815e-01 7.24406540e-03
-1.33591437e+00 -2.70995259e-01 4.95710254e-01 -2.05622345e-01
1.54206955e+00 -2.33520884e-02 3.59697104e-01 1.32324517e+00
8.63996565e-01 5.50405681e-01 1.07939982e+00 -7.74435043e-01
1.26164228e-01 3.65889877e-01 -2.10850179e-01 5.02480865e-01
1.41455783e-02 1.70457110e-01 -7.74042130e-01 -2.42984563e-01
6.35314584e-01 -2.63375998e-01 -1.69720799e-02 -5.77646196e-01
-1.60142541e+00 1.05451930e+00 3.21524233e-01 4.89175260e-01
-6.36127949e-01 -3.55057642e-02 3.10765475e-01 5.58859646e-01
9.02302027e-01 6.28071070e-01 -6.58624470e-01 -3.54188621e-01
-8.66396904e-01 6.80822507e-02 8.31888795e-01 7.50231683e-01
5.35019398e-01 -1.69319406e-01 -1.89513579e-01 9.05113101e-01
5.23669839e-01 1.00406826e+00 6.40236855e-01 -3.78145754e-01
1.03907561e+00 4.32476908e-01 2.28279442e-01 -8.91749918e-01
-2.57506259e-02 -4.72619057e-01 -5.19873261e-01 -1.62617058e-01
-9.44555625e-02 -2.45298728e-01 -6.58329070e-01 1.77961051e+00
-1.68220222e-01 6.53571039e-02 4.66515601e-01 9.70407128e-01
2.75351733e-01 1.37139988e+00 -9.14173573e-03 -4.64184225e-01
1.08123195e+00 -1.37332654e+00 -5.77062607e-01 -2.62424707e-01
3.49250555e-01 -1.01965475e+00 1.32086051e+00 -1.50107563e-01
-1.35271394e+00 -6.53958082e-01 -1.10061789e+00 1.31499246e-01
-2.56219149e-01 2.47135624e-01 1.24070153e-01 3.89001966e-01
-1.19813979e+00 7.50566900e-01 -1.03631842e+00 -5.07359266e-01
-7.44947270e-02 6.37454271e-01 -4.37390387e-01 4.08944078e-02
-1.39350700e+00 1.20076585e+00 3.55600297e-01 -1.73385650e-01
-4.97805774e-01 -5.66919982e-01 -5.66347539e-01 3.55038434e-01
-3.54781330e-01 -9.04970288e-01 1.20983076e+00 -9.47674394e-01
-1.98635459e+00 1.01042283e+00 -4.43932444e-01 -2.76871473e-01
2.94322789e-01 -3.76321405e-01 -5.64889789e-01 -5.93904592e-02
5.14314175e-01 5.98465860e-01 7.77779281e-01 -9.84703720e-01
-5.49414635e-01 -3.50607783e-01 -5.61727643e-01 6.88702703e-01
-5.24818718e-01 4.71340925e-01 -5.92445135e-01 -5.34204066e-01
1.71340823e-01 -1.12299001e+00 -2.14712769e-01 -2.97357351e-01
1.59265414e-01 -1.44660622e-01 4.70274121e-01 -7.74341524e-01
9.96980429e-01 -1.87274349e+00 9.19181943e-01 4.83113807e-03
-4.94256258e-01 3.18200618e-01 -6.34477437e-01 6.08494759e-01
-1.64220467e-01 -1.56589985e-01 -2.86699742e-01 -5.70030510e-01
-5.95846921e-02 2.21552402e-01 -5.25515199e-01 2.30125055e-01
2.85340428e-01 1.27263331e+00 -9.30331111e-01 -5.31203263e-02
4.08168808e-02 5.17213702e-01 -1.55768722e-01 3.47592264e-01
-2.58667588e-01 3.64414990e-01 -3.01361740e-01 2.79515058e-01
3.89096469e-01 -5.61231375e-02 2.19433814e-01 2.97953784e-01
6.41079992e-02 7.39882410e-01 -4.40731972e-01 2.26881504e+00
-7.02398121e-01 4.68649268e-01 -4.59944248e-01 -7.41991818e-01
1.17399979e+00 6.05818987e-01 2.56207079e-01 -8.03621233e-01
-1.78283527e-02 4.69932914e-01 -1.49975598e-01 1.68209448e-02
5.85678339e-01 -5.65012813e-01 -3.49611074e-01 6.12443566e-01
2.43729815e-01 -7.77475461e-02 5.88060580e-02 -3.40951890e-01
5.77366114e-01 4.49512988e-01 6.22208491e-02 -4.13339853e-01
7.04751194e-01 1.28656358e-01 5.45436561e-01 2.06950083e-01
1.07817113e-01 5.19620240e-01 3.61891538e-01 -4.47489351e-01
-1.45466316e+00 -1.43392897e+00 1.23151086e-01 1.13387239e+00
4.76651751e-02 -2.79085815e-01 -7.65108764e-01 -6.98738217e-01
-2.68518895e-01 1.11216712e+00 -1.53621554e-01 -2.92095542e-01
-8.08672905e-01 -9.12597537e-01 3.58491033e-01 6.12961054e-01
1.59827158e-01 -9.34445858e-01 -2.58337613e-02 5.67581654e-01
-5.60897529e-01 -1.05311489e+00 -5.89855909e-01 5.76017797e-02
-1.26475787e+00 -2.31634855e-01 -8.70861769e-01 -1.17675817e+00
5.85943222e-01 1.39236376e-01 1.19882965e+00 -6.53818190e-01
3.49489599e-01 -1.63431421e-01 -3.02352250e-01 -1.56866655e-01
-7.68095136e-01 4.66231287e-01 6.07933462e-01 -3.72203961e-02
1.10264492e+00 -5.19222498e-01 -1.18771970e-01 3.64677757e-01
-3.11472058e-01 2.23105147e-01 6.82943165e-01 1.18068171e+00
7.59884000e-01 -6.02275550e-01 3.50508451e-01 -2.42809430e-01
1.04907811e+00 -2.90654421e-01 -6.28958642e-01 4.34669852e-01
-8.85951698e-01 3.93370628e-01 7.18887985e-01 -4.21496421e-01
-7.94968188e-01 -3.70610319e-02 -1.15911163e-01 -2.91487306e-01
8.52454156e-02 5.31761527e-01 -7.02772737e-02 3.19506228e-01
6.82295680e-01 5.21361947e-01 -7.25120306e-02 -3.65310729e-01
6.20338559e-01 9.88885701e-01 4.19674695e-01 -5.45043230e-01
9.26914811e-01 -7.34825209e-02 -2.45643899e-01 -2.15764478e-01
-4.97011662e-01 -6.33275390e-01 -9.32066023e-01 3.80228698e-01
9.23053861e-01 -1.07936752e+00 1.70833617e-01 5.61213717e-02
-1.31202567e+00 -4.15331777e-03 -1.04400523e-01 9.24875438e-01
-1.04142427e+00 3.08467653e-02 -7.96491861e-01 -1.23650424e-01
-7.62423217e-01 -1.28056657e+00 1.21026993e+00 -1.29050523e-01
-3.34592700e-01 -1.06678057e+00 5.52446127e-01 3.19273949e-01
4.40891922e-01 -3.85919601e-01 1.21464276e+00 -7.12432206e-01
-3.05246562e-01 -3.20558250e-01 -1.47906289e-01 4.07229364e-01
7.32380077e-02 -4.93689328e-01 -5.03796875e-01 -6.26154542e-01
6.90992130e-03 -7.74550363e-02 4.69416112e-01 1.49331465e-01
1.89104863e-02 -1.11375198e-01 -3.84098381e-01 5.39615154e-01
1.72389197e+00 2.82017868e-02 2.66450703e-01 2.94862390e-01
4.19368923e-01 1.32844850e-01 6.23255253e-01 -2.34409079e-01
-4.71108444e-02 1.09822416e+00 -1.02941774e-01 1.20489364e-02
-6.99820444e-02 -5.33291578e-01 8.09378684e-01 1.40865088e+00
-3.21326628e-02 -1.31994933e-01 -9.82862353e-01 5.99614501e-01
-1.99124086e+00 -5.88743985e-01 1.55732766e-01 2.20380092e+00
8.94068718e-01 3.00443947e-01 -1.45000383e-01 -3.69651884e-01
7.60715008e-01 5.50172888e-02 -1.30285561e-01 -9.83050585e-01
1.21281333e-01 5.03875256e-01 5.18686831e-01 7.30535567e-01
-8.82637322e-01 1.55421889e+00 6.55984879e+00 5.58497727e-01
-1.34670413e+00 4.60986733e-01 2.62555867e-01 -1.93740651e-02
-1.86224982e-01 -1.49998784e-01 -7.56603897e-01 2.70702928e-01
1.57612407e+00 -4.82668966e-01 4.93426025e-01 7.42005885e-01
3.09800804e-01 4.91803735e-01 -1.33273029e+00 8.62787724e-01
2.58656502e-01 -1.39134884e+00 3.29721123e-01 -1.53844031e-02
1.01314545e+00 4.70020622e-01 2.58118421e-01 5.16977012e-01
-7.13662943e-03 -9.19586003e-01 3.02203149e-01 3.05687100e-01
1.14626729e+00 -8.57638240e-01 9.61822867e-01 3.37274611e-01
-7.70518780e-01 2.17758909e-01 -7.97832549e-01 -7.89763406e-02
4.10876453e-01 6.90776780e-02 -1.01416767e+00 6.44639194e-01
2.93316871e-01 6.21525943e-01 3.20543163e-02 5.34123182e-01
-6.13716841e-01 7.44071305e-01 1.89179480e-02 -2.16310292e-01
5.05145729e-01 -4.88403738e-01 6.86137438e-01 1.27895808e+00
5.43722570e-01 -3.60843569e-01 6.09110408e-02 9.03297305e-01
-3.29907924e-01 4.48097676e-01 -5.58023870e-01 -3.05871338e-01
5.49221747e-02 9.47659791e-01 -3.60576063e-01 -3.59461188e-01
-5.56888461e-01 1.53453362e+00 2.62922823e-01 3.43899637e-01
-7.23911226e-01 -1.19684607e-01 7.87111878e-01 -1.46527290e-01
2.91949987e-01 -4.46257859e-01 -5.52314878e-01 -1.28482080e+00
1.02138408e-01 -7.00630367e-01 -2.90942878e-01 -6.42165422e-01
-1.29638350e+00 8.25774789e-01 -2.85195619e-01 -1.31301582e+00
-5.95478296e-01 -6.33461297e-01 -3.92135948e-01 1.62383008e+00
-1.26841915e+00 -1.18426669e+00 7.15022326e-01 2.46780679e-01
8.25429976e-01 -9.82929289e-01 1.70792103e+00 8.71876329e-02
-1.53639823e-01 7.88267136e-01 7.62493432e-01 4.42424044e-02
7.41142750e-01 -1.08181560e+00 1.20827425e+00 8.21601987e-01
4.92400438e-01 6.74566150e-01 6.68766499e-01 -5.69533765e-01
-1.47524035e+00 -1.02291167e+00 1.87168908e+00 -7.63100803e-01
7.33103335e-01 -4.56344962e-01 -7.25978911e-01 4.90421474e-01
4.04055774e-01 -2.66068876e-01 8.53489518e-01 3.92468542e-01
-4.32444185e-01 1.04951747e-01 -9.86465096e-01 7.39144564e-01
7.08459795e-01 -7.12022364e-01 -1.28574908e+00 5.81904471e-01
1.10219622e+00 -2.36796454e-01 -7.84749568e-01 2.43492559e-01
4.85313833e-01 -2.44253621e-01 9.56580758e-01 -1.11838949e+00
7.87173927e-01 -9.00392607e-02 -4.53453809e-01 -1.74307072e+00
-6.32466972e-01 -3.91961575e-01 -2.92783752e-02 5.12658060e-01
9.82595623e-01 -3.66263956e-01 8.62008512e-01 2.04548806e-01
1.72542796e-01 -9.30283129e-01 -1.13905430e+00 -8.71628821e-01
5.23766935e-01 -4.53440696e-02 4.02595788e-01 7.05641568e-01
2.92373687e-01 1.09558952e+00 -4.19153959e-01 1.15593098e-01
2.89269358e-01 2.94560909e-01 4.95486856e-01 -1.03819096e+00
-4.65358883e-01 -3.93226177e-01 -3.78657520e-01 -1.01614356e+00
4.14940894e-01 -1.38974237e+00 2.36014239e-02 -1.69988239e+00
2.78310806e-01 8.46243352e-02 -6.14223480e-01 3.79094124e-01
-3.12869251e-01 3.50958347e-01 1.84380874e-01 5.06418049e-01
-1.66476220e-01 7.80451655e-01 8.77505958e-01 -1.50341630e-01
-3.41156840e-01 1.53649151e-01 -4.42519277e-01 2.80887246e-01
9.68744457e-01 -7.89183855e-01 -2.24473074e-01 -8.40648234e-01
1.33965224e-01 3.43948305e-01 -2.90561557e-01 -7.20467508e-01
-1.16060644e-01 -1.01765826e-01 2.19444439e-01 -3.95122975e-01
4.94171649e-01 -6.74887061e-01 -1.69906646e-01 5.73809981e-01
-5.09926319e-01 4.96692330e-01 1.23942815e-01 5.26128769e-01
-4.37876225e-01 3.02661005e-02 5.81063032e-01 -2.33121384e-02
-2.95124233e-01 1.61784351e-01 -1.18802011e-01 -3.64495516e-01
8.60086858e-01 7.94581547e-02 2.45709345e-01 -1.32545248e-01
-5.94415665e-01 -2.15684220e-01 3.76719952e-01 1.00668502e+00
5.33558846e-01 -1.67877066e+00 -1.21121860e+00 5.08150280e-01
2.98134029e-01 -8.18708837e-01 -4.90372479e-01 5.52636504e-01
-3.84707540e-01 8.28925431e-01 -5.46899617e-01 -6.77850902e-01
-1.25465977e+00 5.62061310e-01 -2.91025676e-02 -5.29481113e-01
-5.67098796e-01 9.26505923e-01 -3.79214913e-01 -9.33067024e-01
-2.50838064e-02 5.26667805e-03 9.43441987e-02 -3.96144003e-01
3.28559339e-01 -1.06344204e-02 1.00159034e-01 -9.70810831e-01
-2.83438802e-01 6.42433584e-01 -1.81546345e-01 -7.32626200e-01
1.35429513e+00 1.78456807e-03 -2.82549471e-01 5.00949621e-01
1.51873398e+00 -1.59363285e-01 -8.24566603e-01 -6.58435345e-01
1.11385860e-01 -3.55647802e-01 -1.43434733e-01 -1.02326524e+00
-3.66164356e-01 1.10186315e+00 6.39613390e-01 -6.06353223e-01
8.12947214e-01 -1.99160755e-01 9.86656368e-01 5.79265893e-01
5.83984673e-01 -1.17150009e+00 -9.44829583e-02 8.28004062e-01
9.03949976e-01 -1.23568082e+00 -2.65116453e-01 -4.62669060e-02
-5.33237159e-01 1.13859415e+00 6.02987260e-02 -1.69428229e-01
2.61077940e-01 -9.08752158e-02 3.47519368e-01 3.13560307e-01
-7.89747775e-01 4.26177144e-01 4.00946647e-01 3.60531688e-01
6.05305910e-01 4.95659322e-01 -8.90861034e-01 3.84509087e-01
-2.54486352e-01 -3.21793631e-02 -2.29610167e-02 6.31148994e-01
-4.45331335e-01 -1.77324700e+00 -3.53827626e-01 -1.07664511e-01
-4.02158976e-01 -4.45703536e-01 -2.92176008e-01 1.60502687e-01
-3.28689992e-01 8.18607628e-01 1.46937398e-02 -4.01400864e-01
2.32075453e-01 6.20829403e-01 3.63335341e-01 -8.77353728e-01
-5.17248571e-01 1.86262697e-01 7.80734047e-03 -4.44607735e-01
-1.79572210e-01 -4.95480001e-01 -1.17391574e+00 -2.63011549e-02
-9.25446376e-02 4.63381618e-01 1.21161878e+00 8.36786687e-01
5.13759911e-01 2.14198932e-01 8.47394824e-01 -6.02488220e-01
-1.16724300e+00 -1.05699027e+00 -1.87260273e-03 2.34319299e-01
-1.87683538e-01 -2.61053234e-01 7.96448588e-02 -2.23290306e-02] | [11.615616798400879, 10.278844833374023] |
357ac170-6cab-44e2-a8ab-6f1a6f989582 | monocular-3d-multi-person-pose-estimation-by | 2104.01797 | null | https://arxiv.org/abs/2104.01797v2 | https://arxiv.org/pdf/2104.01797v2.pdf | Monocular 3D Multi-Person Pose Estimation by Integrating Top-Down and Bottom-Up Networks | In monocular video 3D multi-person pose estimation, inter-person occlusion and close interactions can cause human detection to be erroneous and human-joints grouping to be unreliable. Existing top-down methods rely on human detection and thus suffer from these problems. Existing bottom-up methods do not use human detection, but they process all persons at once at the same scale, causing them to be sensitive to multiple-persons scale variations. To address these challenges, we propose the integration of top-down and bottom-up approaches to exploit their strengths. Our top-down network estimates human joints from all persons instead of one in an image patch, making it robust to possible erroneous bounding boxes. Our bottom-up network incorporates human-detection based normalized heatmaps, allowing the network to be more robust in handling scale variations. Finally, the estimated 3D poses from the top-down and bottom-up networks are fed into our integration network for final 3D poses. Besides the integration of top-down and bottom-up networks, unlike existing pose discriminators that are designed solely for single person, and consequently cannot assess natural inter-person interactions, we propose a two-person pose discriminator that enforces natural two-person interactions. Lastly, we also apply a semi-supervised method to overcome the 3D ground-truth data scarcity. Our quantitative and qualitative evaluations show the effectiveness of our method compared to the state-of-the-art baselines. | ['Robby T. Tan', 'Bo Yang', 'Bo wang', 'Yu Cheng'] | 2021-04-05 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Cheng_Monocular_3D_Multi-Person_Pose_Estimation_by_Integrating_Top-Down_and_Bottom-Up_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Cheng_Monocular_3D_Multi-Person_Pose_Estimation_by_Integrating_Top-Down_and_Bottom-Up_CVPR_2021_paper.pdf | cvpr-2021-1 | ['3d-multi-person-pose-estimation-absolute', '3d-multi-person-pose-estimation-root-relative', '3d-multi-person-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.16946436e-01 -1.72122649e-03 3.45585532e-02 -2.10840479e-01
-5.87116361e-01 -4.29696560e-01 4.24619555e-01 -3.14912319e-01
-4.83724624e-01 5.19925535e-01 4.08317715e-01 5.46895146e-01
2.71138340e-01 -6.91345513e-01 -6.23260915e-01 -3.08598787e-01
8.25794507e-03 5.54906964e-01 6.03788614e-01 -2.01005444e-01
-3.42346758e-01 4.30836588e-01 -1.47717512e+00 1.19585074e-01
6.35893941e-01 6.56822622e-01 -3.68622869e-01 6.66458309e-01
5.93033016e-01 1.13697015e-01 -7.83978224e-01 -4.89735395e-01
6.11388981e-01 -3.58456939e-01 -3.03246111e-01 2.66392678e-01
1.12763679e+00 -6.59901321e-01 -5.30458212e-01 9.09348249e-01
9.02630031e-01 1.33554012e-01 4.93894309e-01 -1.44068158e+00
-1.67220727e-01 2.89021507e-02 -1.00328875e+00 -2.71183908e-01
1.04200840e+00 3.25756192e-01 7.50698507e-01 -1.12460995e+00
6.94713771e-01 1.94613087e+00 1.02628613e+00 6.42607152e-01
-1.31937933e+00 -7.06866145e-01 4.23713118e-01 -2.59788930e-01
-1.50941908e+00 -2.83394009e-01 6.55460477e-01 -4.93886411e-01
6.25791252e-01 1.31176844e-01 1.05092299e+00 1.38545394e+00
-1.63141042e-02 9.44316447e-01 9.06293988e-01 -3.60982537e-01
-1.68655545e-01 -2.56028563e-01 -2.35829614e-02 7.09016204e-01
7.70012259e-01 2.15913489e-01 -7.46874273e-01 -1.27915785e-01
1.19105566e+00 1.06571712e-01 -8.87392014e-02 -9.92515743e-01
-1.43276703e+00 5.89529097e-01 6.22996628e-01 -1.24751225e-01
-3.02528948e-01 2.35033259e-01 3.77369732e-01 -1.31380841e-01
3.98686439e-01 4.82955910e-02 -1.75554290e-01 1.57612160e-01
-1.05057299e+00 8.64504576e-01 5.41626692e-01 1.04960144e+00
4.83515412e-01 -3.11173975e-01 -4.66034234e-01 6.97850525e-01
4.28744942e-01 4.24405456e-01 -7.28387991e-03 -9.03804004e-01
5.37489951e-01 8.47062230e-01 3.73795599e-01 -1.19714260e+00
-7.75245786e-01 -4.50574815e-01 -6.51943803e-01 3.93577516e-01
8.19910169e-01 -1.87676802e-01 -1.09844232e+00 1.89423692e+00
6.98020697e-01 -3.21618497e-01 -4.96435195e-01 1.34736228e+00
6.06652319e-01 4.71238233e-03 -1.86650991e-03 3.13377440e-01
1.61965573e+00 -1.29498875e+00 -5.38051248e-01 -6.37521565e-01
2.34995216e-01 -5.10541975e-01 8.62740755e-01 3.13191563e-01
-1.12991047e+00 -7.88361967e-01 -1.02316809e+00 -8.61274526e-02
-2.21288696e-01 3.29508543e-01 4.84035790e-01 7.03602612e-01
-6.85667515e-01 4.97892529e-01 -7.55608678e-01 -7.75168300e-01
1.71898276e-01 3.00985575e-01 -6.72659695e-01 -8.98344368e-02
-1.11088634e+00 9.65315938e-01 1.71524137e-01 4.59086925e-01
-5.16175687e-01 -2.63371110e-01 -9.94640350e-01 -3.78485590e-01
7.52927244e-01 -1.26987922e+00 1.02068722e+00 -5.39123535e-01
-1.18444848e+00 8.16138148e-01 -1.11768894e-01 -1.63947985e-01
1.30784321e+00 -8.10899258e-01 -1.89470407e-02 2.06659093e-01
2.83515662e-01 9.36187148e-01 6.74786091e-01 -1.35252178e+00
-5.09606421e-01 -5.88555217e-01 6.47685006e-02 5.09732127e-01
-1.60489321e-01 -1.71833143e-01 -1.16845477e+00 -8.26695740e-01
4.18714195e-01 -1.20279980e+00 -2.89274424e-01 5.08657575e-01
-7.57059991e-01 -2.28263792e-02 6.69119298e-01 -6.46428883e-01
9.55517828e-01 -1.73283422e+00 3.92733634e-01 1.22131012e-01
3.31893027e-01 -8.35893303e-03 -1.19400352e-01 3.12778860e-01
1.53896183e-01 -1.17246047e-01 1.00007728e-01 -8.53199244e-01
1.58430353e-01 3.75610478e-02 4.72702384e-01 7.59804428e-01
1.29280329e-01 7.77045429e-01 -7.58514762e-01 -6.67524576e-01
4.60889697e-01 6.61514580e-01 -5.95669568e-01 1.73219696e-01
1.00949705e-01 5.08583784e-01 -2.08406538e-01 8.59884977e-01
5.67128003e-01 -6.27892390e-02 1.08471578e-02 -3.79798472e-01
1.58521220e-01 -4.70112786e-02 -1.73499537e+00 1.74958396e+00
2.04674564e-02 1.94605470e-01 1.24564581e-01 -2.75321037e-01
4.57962036e-01 3.29546601e-01 4.40435082e-01 -2.39628464e-01
1.02267876e-01 1.10944167e-01 -1.39280751e-01 -3.07881343e-03
5.28372407e-01 1.08074762e-01 -3.11226398e-01 1.36313528e-01
1.61072996e-03 1.12427950e-01 3.63419294e-01 2.21840531e-01
7.57421374e-01 8.21049571e-01 3.08839291e-01 1.57731324e-01
3.50722075e-01 -3.66091162e-01 9.22112167e-01 8.35600615e-01
-6.43720031e-01 1.12589419e+00 3.05814892e-01 -6.52185738e-01
-1.08516026e+00 -1.20933807e+00 1.64683759e-01 9.73717093e-01
2.94253618e-01 -4.58637476e-01 -8.82074654e-01 -7.06086755e-01
3.02974999e-01 -1.02513410e-01 -6.93285823e-01 1.08233116e-05
-6.31725669e-01 -4.71958697e-01 6.66005552e-01 7.96735823e-01
5.78116655e-01 -4.52806115e-01 -6.70114219e-01 1.55514032e-01
-5.27823865e-01 -1.52542651e+00 -7.33074486e-01 -3.86681318e-01
-6.13448739e-01 -1.29569387e+00 -1.23433161e+00 -4.20859665e-01
7.60682583e-01 2.93043494e-01 9.84904706e-01 -5.91742136e-02
-2.71448642e-01 4.83357638e-01 -1.65407613e-01 -2.58815467e-01
-3.04888058e-02 6.34143054e-02 6.92297101e-01 -6.51702583e-02
3.81824166e-01 -5.05055428e-01 -1.02117300e+00 6.33395374e-01
-2.65071779e-01 1.75758615e-01 4.93501991e-01 6.70301735e-01
3.21784675e-01 -2.60538638e-01 1.54122576e-01 -4.22690779e-01
2.50993669e-01 2.71773398e-01 -3.69970351e-01 7.42142871e-02
-1.37984619e-01 -4.58637357e-01 1.50692508e-01 -5.05667150e-01
-7.51740754e-01 5.09416103e-01 1.78288341e-01 -4.87311125e-01
-2.80253828e-01 -2.60120928e-01 -3.68281990e-01 -1.74032435e-01
9.54164565e-01 -2.15643361e-01 5.87171689e-02 -4.18849349e-01
4.93429482e-01 3.49354684e-01 8.38295579e-01 -6.03589833e-01
1.11101842e+00 6.15965962e-01 -9.20384303e-02 -5.56961119e-01
-1.00004160e+00 -5.83283603e-01 -1.14273524e+00 -6.64957643e-01
1.11822462e+00 -1.33117294e+00 -7.73969591e-01 6.83657885e-01
-1.28573537e+00 5.48216254e-02 -5.18208593e-02 5.43062389e-01
-4.30704802e-01 6.07052565e-01 -8.15486372e-01 -1.05604899e+00
-2.03800067e-01 -1.02044225e+00 1.64670467e+00 1.21623889e-01
-6.51874244e-01 -5.52183032e-01 -9.13788900e-02 4.98497576e-01
-1.65495187e-01 7.85348713e-01 3.49151040e-03 -2.84883052e-01
-4.04399425e-01 -6.33138359e-01 -1.90039858e-01 1.53193995e-01
4.09942400e-03 -1.36992455e-01 -9.04973269e-01 -6.25341058e-01
-6.13491476e-01 -3.02047580e-01 7.50206888e-01 3.72660100e-01
3.36250007e-01 -2.58007739e-02 -5.15194654e-01 3.74706060e-01
8.77613485e-01 -6.52458549e-01 4.84348059e-01 4.90470529e-01
9.90484416e-01 9.20047879e-01 5.88974118e-01 3.90152514e-01
5.62931836e-01 1.12352192e+00 2.15539217e-01 -4.10208255e-01
-3.72232050e-01 -5.27661264e-01 4.13099349e-01 2.86197990e-01
-5.12232482e-01 -3.30585316e-02 -7.81923234e-01 3.98611665e-01
-2.05524921e+00 -8.80950809e-01 -1.56399682e-01 2.35419941e+00
5.84038734e-01 4.10808861e-01 8.35512221e-01 1.20649934e-02
9.15263295e-01 1.14791647e-01 -4.94513154e-01 5.00581622e-01
-1.02485411e-01 -3.91382337e-01 4.70492154e-01 3.00750971e-01
-1.40550411e+00 8.18299472e-01 6.28876114e+00 4.19716984e-01
-4.61108565e-01 -1.00021698e-01 1.61547735e-01 -5.27384520e-01
3.23395252e-01 -1.18057303e-01 -9.97207105e-01 2.60097474e-01
1.51198745e-01 4.94082391e-01 -6.77639544e-02 7.49971151e-01
3.03996414e-01 -3.06654185e-01 -1.27884662e+00 1.12630546e+00
9.35009122e-02 -5.97842932e-01 -8.29512440e-03 1.73299804e-01
6.18163168e-01 -4.22598392e-01 -3.26075196e-01 2.17364371e-01
2.72458345e-01 -7.54949272e-01 1.08022285e+00 4.93069082e-01
5.35960972e-01 -7.59810269e-01 6.93660617e-01 3.74992579e-01
-1.56055641e+00 6.85675889e-02 -1.42559066e-01 -1.79909393e-01
5.35832763e-01 6.73624873e-01 -3.25682253e-01 5.67640007e-01
8.29766214e-01 4.69969660e-01 -6.80640161e-01 9.63743687e-01
-4.26548362e-01 -2.07300201e-01 -5.44749141e-01 2.50879079e-01
-1.94618672e-01 1.03330277e-02 7.92345405e-01 1.27230978e+00
8.92978609e-02 -1.85140982e-01 8.03379714e-01 7.58200347e-01
1.31824642e-01 -2.19178900e-01 -4.13257658e-01 5.13254464e-01
4.44686472e-01 1.15744805e+00 -7.13555813e-01 -3.06137800e-01
-3.90197396e-01 1.46450138e+00 2.66827226e-01 3.57945174e-01
-7.54181862e-01 -1.75287917e-01 8.02896380e-01 4.42080349e-01
5.76279908e-02 -4.92465317e-01 -1.68191791e-01 -1.43402386e+00
3.91118765e-01 -9.61955249e-01 3.78460348e-01 -6.65391743e-01
-1.43574989e+00 3.07058722e-01 3.07443976e-01 -1.31820071e+00
-3.19012702e-01 -6.05604172e-01 -1.84174493e-01 7.43851721e-01
-8.84867668e-01 -1.53870106e+00 -5.01883686e-01 5.21927834e-01
4.31264907e-01 1.42268345e-01 5.66119730e-01 4.19895232e-01
-6.36842608e-01 8.30018103e-01 -8.10065091e-01 5.77995062e-01
1.10187268e+00 -1.21569097e+00 7.17098713e-01 1.14879203e+00
-7.24731609e-02 1.00514364e+00 8.14501524e-01 -1.00517893e+00
-1.19242120e+00 -9.15346742e-01 7.10062981e-01 -8.72610509e-01
1.81603283e-01 -9.30890560e-01 -5.22959888e-01 7.02644229e-01
-3.86755973e-01 2.15729013e-01 4.01158124e-01 3.84448379e-01
-6.13361180e-01 9.61035416e-02 -1.11687934e+00 9.03778911e-01
1.48534954e+00 -3.99008483e-01 -5.61541736e-01 2.94971824e-01
5.07992089e-01 -7.65306354e-01 -7.05190957e-01 4.26254541e-01
1.04307628e+00 -1.13695967e+00 1.38629091e+00 -2.48064190e-01
1.53722361e-01 -5.73590219e-01 7.90506378e-02 -9.83639657e-01
-4.63693947e-01 -5.24001479e-01 -3.36454570e-01 9.17094827e-01
6.31137341e-02 -3.73852938e-01 1.08901620e+00 7.47765362e-01
2.90583044e-01 -4.41390008e-01 -9.00520384e-01 -9.98260081e-01
-1.64808929e-01 -3.10890734e-01 2.62446105e-01 5.79472899e-01
-7.37448707e-02 3.58598143e-01 -8.94734621e-01 3.10875475e-01
1.03712904e+00 -2.36249879e-01 1.49789548e+00 -1.30550504e+00
-4.95318621e-01 -2.27649137e-01 -7.05994308e-01 -1.25253189e+00
-3.95424604e-01 -1.20643228e-01 2.91706175e-01 -1.58739042e+00
5.30610561e-01 3.66703831e-02 1.25080958e-01 5.23976028e-01
-4.50881034e-01 7.95906007e-01 4.88856822e-01 2.63838708e-01
-5.85573077e-01 4.71253961e-01 1.31470156e+00 2.03976303e-01
-1.78530812e-01 -6.16710298e-02 -4.96298194e-01 1.07844746e+00
2.58870482e-01 -3.25500965e-01 -6.01654574e-02 -1.85321912e-01
6.74100369e-02 -2.51159489e-01 9.58419263e-01 -1.41220713e+00
2.73019195e-01 1.05837308e-01 9.49111223e-01 -7.32542634e-01
6.78789079e-01 -5.70097446e-01 1.05915539e-01 5.39608598e-01
1.09536253e-01 -4.54970971e-02 -4.34006676e-02 6.92787290e-01
7.77525902e-02 3.14968258e-01 6.96062386e-01 -4.03942704e-01
-3.77430409e-01 4.37200010e-01 5.45569621e-02 -1.33083716e-01
8.51874530e-01 -7.21391976e-01 3.13214473e-02 -6.27683818e-01
-7.60180652e-01 4.07155991e-01 7.16690958e-01 6.83733582e-01
4.06902254e-01 -1.48337793e+00 -7.08793819e-01 -4.54732776e-02
2.05881774e-01 2.35420436e-01 2.20576435e-01 8.32889259e-01
-3.38291943e-01 1.40243143e-01 -1.13221020e-01 -8.61872673e-01
-1.44621527e+00 2.86999017e-01 5.07904708e-01 -2.55044430e-01
-7.03445911e-01 6.80402875e-01 3.31770271e-01 -7.41840899e-01
3.61297905e-01 4.05632444e-02 1.05516806e-01 9.13521722e-02
4.60209966e-01 5.36053061e-01 -2.75564790e-01 -9.47299063e-01
-5.57184935e-01 1.03709531e+00 1.98507607e-02 -3.63975048e-01
9.58273232e-01 -2.15507299e-01 2.12721586e-01 2.47388616e-01
8.26248884e-01 1.46923825e-01 -1.65995622e+00 -1.81648701e-01
-4.39785033e-01 -5.56811333e-01 -5.05852818e-01 -6.73663676e-01
-7.69455731e-01 7.17633426e-01 6.14267647e-01 -2.30343521e-01
7.48496234e-01 -1.28865242e-01 7.10914969e-01 1.43128097e-01
6.46250904e-01 -1.39692974e+00 3.53757977e-01 3.08074176e-01
8.95837486e-01 -1.35480702e+00 5.84402621e-01 -8.74463856e-01
-3.10275078e-01 9.62293863e-01 1.00154638e+00 -1.87168017e-01
-7.34803639e-03 1.16948351e-01 8.80856737e-02 -1.84813410e-01
-1.37474686e-01 -3.13023925e-01 7.09643841e-01 7.20358491e-01
3.96334887e-01 -4.30700667e-02 -1.54380038e-01 4.54187006e-01
-2.13353038e-01 3.93117517e-02 1.47732973e-01 1.13927054e+00
-2.55524307e-01 -1.04222775e+00 -9.30550098e-01 3.55187505e-02
-2.32890636e-01 4.56173897e-01 -7.28996813e-01 1.20198393e+00
2.46101126e-01 8.52419794e-01 -2.06429303e-01 -4.73321110e-01
9.47232306e-01 -1.51122250e-02 6.66705728e-01 -5.49116254e-01
-5.48099339e-01 4.27426279e-01 2.02503547e-01 -9.11744416e-01
-5.26387930e-01 -7.63602614e-01 -1.03493834e+00 -2.74007469e-01
-4.55145806e-01 -4.97305304e-01 2.27876365e-01 7.35615313e-01
2.41024405e-01 4.45493519e-01 8.66298079e-02 -1.59015012e+00
-5.01548886e-01 -1.02691674e+00 -2.86925375e-01 6.45251215e-01
2.52716243e-01 -1.07994115e+00 -2.35594615e-01 -1.91963971e-01] | [7.0492448806762695, -0.8966325521469116] |
3c73c170-84df-474b-bca6-ebbec9dc49d7 | characterization-of-graphs-for-protein | 1407.8033 | null | http://arxiv.org/abs/1407.8033v5 | http://arxiv.org/pdf/1407.8033v5.pdf | Characterization of graphs for protein structure modeling and recognition of solubility | This paper deals with the relations among structural, topological, and
chemical properties of the E.Coli proteome from the vantage point of the
solubility/aggregation propensity of proteins. Each E.Coli protein is initially
represented according to its known folded 3D shape. This step consists in
representing the available E.Coli proteins in terms of graphs. We first analyze
those graphs by considering pure topological characterizations, i.e., by
analyzing the mass fractal dimension and the distribution underlying both
shortest paths and vertex degrees. Results confirm the general architectural
principles of proteins. Successively, we focus on the statistical properties of
a representation of such graphs in terms of vectors composed of several
numerical features, which we extracted from their structural representation. We
found that protein size is the main discriminator for the solubility, while
however there are other factors that help explaining the solubility degree. We
finally analyze such data through a novel one-class classifier, with the aim of
discriminating among very and poorly soluble proteins. Results are encouraging
and consolidate the potential of pattern recognition techniques when employed
to describe complex biological systems. | ['Alessandro Giuliani', 'Alireza Sadeghian', 'Lorenzo Livi'] | 2014-07-30 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [ 1.47007808e-01 1.28565744e-01 -1.67766362e-02 -1.41405314e-01
3.30809802e-02 -7.74261594e-01 6.45364642e-01 8.25178802e-01
-3.05750847e-01 7.89360166e-01 -5.44566065e-02 -3.77259284e-01
-5.05835176e-01 -7.73806810e-01 -5.41226149e-01 -1.11446762e+00
-5.94421387e-01 6.97095037e-01 3.03283483e-01 -3.17974865e-01
2.07034782e-01 9.40472901e-01 -1.24387074e+00 7.68161491e-02
8.57225776e-01 7.69338667e-01 2.97418028e-01 6.32073462e-01
-3.06987882e-01 5.06114841e-01 -4.88921970e-01 -1.78793550e-01
-1.56030238e-01 -2.71885604e-01 -7.08260417e-01 1.87043697e-01
-9.57736894e-02 6.57133579e-01 -1.26407191e-01 8.67757857e-01
2.32809559e-02 -1.74832225e-01 1.22116554e+00 -1.05386734e+00
-4.80545223e-01 1.72825262e-01 -1.57044664e-01 1.08631022e-01
5.24873435e-01 -9.57638696e-02 1.01312351e+00 -6.55799150e-01
1.09343553e+00 1.06738961e+00 5.14452815e-01 1.57182347e-02
-1.61260986e+00 3.41536969e-01 -4.31984030e-02 1.94363281e-01
-1.05423844e+00 -4.85029910e-03 7.23047376e-01 -9.83807087e-01
9.01232541e-01 2.90604204e-01 5.81298649e-01 5.93469977e-01
8.39055240e-01 1.69659480e-01 1.23768139e+00 -5.53861976e-01
4.91017491e-01 -2.75547169e-02 5.47803640e-01 8.11957479e-01
6.71146274e-01 -3.09516400e-01 -9.14685130e-02 -3.86121035e-01
2.79757231e-01 -2.04110537e-02 -2.35233679e-01 -1.20463002e+00
-1.08516538e+00 7.41309643e-01 2.45413661e-01 7.41213739e-01
-3.89658719e-01 -3.61958504e-01 3.86304289e-01 3.02389324e-01
2.55310655e-01 6.53858423e-01 -6.26193166e-01 2.43549764e-01
-4.52666521e-01 1.31558388e-01 9.60643232e-01 6.60193026e-01
8.62719357e-01 -5.19900799e-01 4.03248608e-01 4.85648304e-01
1.72572106e-01 3.94287497e-01 3.51457328e-01 -5.02691209e-01
1.12275086e-01 1.02832127e+00 2.64380798e-02 -1.27426314e+00
-7.86366582e-01 -4.56887335e-02 -6.05050266e-01 8.28897879e-02
8.56769919e-01 2.74760395e-01 -4.32592958e-01 1.54998946e+00
1.27920836e-01 -6.28157139e-01 1.73665300e-01 5.24528205e-01
6.45961344e-01 4.66479301e-01 2.86820918e-01 -5.40842116e-01
1.53252792e+00 -3.46631289e-01 -4.80396837e-01 3.37009370e-01
8.25117350e-01 -4.16456223e-01 6.29637957e-01 3.54386359e-01
-8.47343564e-01 -2.68043816e-01 -1.07547057e+00 2.14442849e-01
-8.16607416e-01 3.42054427e-01 5.02207339e-01 4.27408785e-01
-8.01669061e-01 9.54302490e-01 -9.27457929e-01 -7.10201263e-01
-1.08533144e-01 3.90997529e-01 -7.64701843e-01 2.63336301e-01
-8.39462698e-01 1.01577842e+00 3.85973036e-01 -2.38843650e-01
6.69329613e-02 -5.30336797e-02 -6.72564626e-01 -1.34410998e-02
5.27596474e-03 -3.13412100e-01 5.47328711e-01 -5.20068467e-01
-1.15666592e+00 1.02512348e+00 -2.44666740e-01 -2.75491714e-01
1.64215371e-01 5.44814169e-01 -3.91846120e-01 4.39686805e-01
-1.22798935e-01 -2.35082190e-02 1.21594816e-01 -1.24459636e+00
-5.90577386e-02 -5.71756840e-01 -1.11873522e-01 -2.73745418e-01
2.28313170e-02 -1.70401186e-01 1.88973635e-01 -2.33089715e-01
5.67482233e-01 -9.30591226e-01 -4.03639108e-01 -1.09728709e-01
-4.65287387e-01 -2.79248506e-01 3.96592170e-01 -5.03313959e-01
9.20943558e-01 -1.84681273e+00 6.15459442e-01 6.88716412e-01
7.65730560e-01 1.05389245e-01 8.23931843e-02 8.52970600e-01
-5.28306901e-01 9.80196148e-02 -3.06420565e-01 4.58158344e-01
-1.08770497e-01 7.80940652e-02 3.50297503e-02 8.89195025e-01
3.58004361e-01 8.71322870e-01 -8.83271277e-01 -3.75325590e-01
2.04567909e-01 3.42578918e-01 -2.18055904e-01 -5.84204830e-02
-3.19772899e-01 1.76646993e-01 -8.49114597e-01 4.14654374e-01
7.24898756e-01 -5.11138439e-01 1.05367804e+00 -2.97713101e-01
-2.81622618e-01 1.67476475e-01 -7.02651262e-01 1.09243536e+00
2.21367821e-01 4.81632292e-01 -7.15921968e-02 -1.26581109e+00
1.27718461e+00 2.60145385e-02 7.81647861e-01 -4.06364709e-01
3.12348306e-01 3.49329531e-01 2.67627180e-01 -3.37702155e-01
1.98176876e-01 1.31692395e-01 7.08836019e-02 1.42200753e-01
3.20717171e-02 3.46163630e-01 5.24461806e-01 2.00315475e-01
1.26733446e+00 1.46720991e-01 7.03194082e-01 -8.73660028e-01
7.39816368e-01 1.97828680e-01 1.02095254e-01 1.67544752e-01
-2.70848069e-02 3.31242323e-01 1.32135069e+00 -6.62833631e-01
-1.21622014e+00 -1.06874156e+00 -3.00310344e-01 7.22397566e-01
1.11920185e-01 -5.56630552e-01 -8.47939789e-01 -2.95927733e-01
2.48818696e-01 8.02843124e-02 -7.27434218e-01 -8.05152506e-02
-4.64673817e-01 -1.10376465e+00 5.59518822e-02 9.42529552e-03
-3.16012502e-01 -7.41903424e-01 -6.84829473e-01 2.43969381e-01
5.79312593e-02 -8.72686028e-01 1.53376997e-01 7.49972999e-01
-1.08175302e+00 -1.61711669e+00 -5.49839675e-01 -5.88730752e-01
8.43116462e-01 1.60241574e-01 1.03908443e+00 1.85344834e-02
-4.01500344e-01 1.02071285e-01 -3.08201164e-01 -1.45603761e-01
-6.09221697e-01 -2.36191358e-02 2.96348512e-01 -1.65050089e-01
5.02501905e-01 -7.75704920e-01 -4.54619437e-01 4.02155876e-01
-8.50686789e-01 -3.19302201e-01 5.33089578e-01 5.33573687e-01
7.81979024e-01 -4.20789458e-02 1.75816044e-01 -9.78734076e-01
6.63564920e-01 -5.31039178e-01 -4.18482184e-01 4.59346056e-01
-4.91420954e-01 4.58272517e-01 7.39077866e-01 -2.97720492e-01
-3.83657485e-01 2.31309161e-01 2.84700036e-01 3.92145813e-02
-2.21447349e-01 3.53980780e-01 -3.91276598e-01 -2.92240530e-01
6.28808558e-01 4.27334398e-01 3.70936215e-01 -4.56926465e-01
1.90817893e-01 3.58880490e-01 1.18353717e-01 -6.88077152e-01
4.56793576e-01 2.96923697e-01 5.82717121e-01 -1.18598723e+00
7.49067590e-02 -4.51928198e-01 -9.45956349e-01 -9.99838784e-02
7.38897085e-01 -3.02100718e-01 -1.16950202e+00 2.69432008e-01
-1.13527715e+00 2.73527116e-01 -1.20153889e-01 3.16533059e-01
-9.10597980e-01 8.50765705e-01 -5.78267217e-01 -5.76375246e-01
2.67576367e-01 -1.01618338e+00 7.33482361e-01 -2.40234300e-01
-2.80553401e-01 -1.03057635e+00 5.28206587e-01 -7.67000318e-02
2.72431467e-02 7.19614267e-01 1.53390574e+00 -9.45499122e-01
-3.42278153e-01 -7.04419762e-02 -4.52034958e-02 -2.19999671e-01
2.13284984e-01 2.69040495e-01 -4.32068020e-01 -2.79536277e-01
-2.89131254e-02 1.96455538e-01 8.68339241e-01 3.73301536e-01
6.68873668e-01 -1.95801079e-01 -6.50852323e-01 3.21228802e-01
1.70312881e+00 4.05606866e-01 4.40830946e-01 4.02667791e-01
3.71903002e-01 9.37067866e-01 3.89579266e-01 1.72207579e-01
-1.47828877e-01 7.19959915e-01 2.05492422e-01 8.59775022e-02
2.96310067e-01 -1.38703123e-01 1.01660199e-01 7.68937171e-01
-4.05191928e-01 -3.01502109e-01 -1.04150593e+00 8.20425972e-02
-1.75103199e+00 -7.74068415e-01 -4.27424818e-01 2.06117511e+00
4.68509257e-01 1.64519116e-01 3.77377152e-01 2.94818133e-01
6.67794108e-01 3.26366350e-02 -3.80465239e-01 -5.22912681e-01
-5.53174257e-01 2.73848455e-02 6.60841346e-01 3.56653154e-01
-8.72611165e-01 3.62029880e-01 6.90105534e+00 3.83207947e-01
-1.01293588e+00 -5.64836979e-01 5.06772757e-01 4.27830875e-01
-2.10312590e-01 -2.76038740e-02 -3.63298029e-01 4.89049584e-01
1.03041005e+00 -1.95213914e-01 1.33694708e-01 7.90636122e-01
-2.88786087e-02 -3.00413907e-01 -8.88234854e-01 5.38777649e-01
-8.42945278e-02 -1.27009726e+00 8.29525515e-02 4.25489634e-01
3.14485401e-01 -2.20127210e-01 -3.54228169e-01 -4.21745360e-01
-1.13621011e-01 -8.92281413e-01 3.74811023e-01 8.85055482e-01
6.59280300e-01 -5.52976072e-01 6.87660158e-01 2.43988022e-01
-1.17971277e+00 7.94693157e-02 -5.53119361e-01 -3.76691520e-02
-1.24561504e-01 6.25952959e-01 -8.30727875e-01 5.99560320e-01
6.05616570e-02 4.89739835e-01 -5.21096170e-01 7.78873861e-01
3.45124841e-01 5.55090010e-02 -2.32859224e-01 -4.06188935e-01
-2.17070687e-03 -8.55586171e-01 5.02298474e-01 1.20826745e+00
2.18648925e-01 2.95699090e-01 8.69067479e-03 7.34914780e-01
3.08559418e-01 5.86569846e-01 -8.79428744e-01 -5.00484526e-01
1.11656031e-02 1.06659174e+00 -1.25606155e+00 -2.16806948e-01
-4.34931785e-01 4.31496859e-01 4.19298321e-01 2.28325576e-01
-2.34536633e-01 -4.28513944e-01 7.19161749e-01 4.73973721e-01
4.20831412e-01 -5.51952600e-01 1.00171305e-02 -1.12910593e+00
1.03167603e-02 -6.17894351e-01 7.42669916e-03 -4.96776313e-01
-1.10728955e+00 7.81618237e-01 -5.74877448e-02 -9.93813992e-01
-1.27088815e-01 -1.28607738e+00 -3.00145566e-01 9.01016951e-01
-1.09869766e+00 -6.91018760e-01 1.91872120e-01 6.35676309e-02
-6.76716641e-02 -1.87158167e-01 1.10179043e+00 -2.91070521e-01
-4.24606383e-01 -1.16171092e-01 8.64775360e-01 -2.85980612e-01
2.72212207e-01 -1.42732441e+00 3.95428330e-01 2.86993086e-01
-4.90754694e-02 8.60366821e-01 1.13554525e+00 -6.69555008e-01
-1.53902960e+00 -5.53725541e-01 1.26131153e+00 -6.12561464e-01
8.90383244e-01 -4.78072137e-01 -9.00850356e-01 1.90990090e-01
-2.44105756e-01 -1.40052289e-01 6.38206244e-01 4.50077616e-02
-2.21515521e-01 2.68620849e-01 -1.08049476e+00 3.45188558e-01
9.54243541e-01 -3.84550095e-01 -6.09447181e-01 4.59475279e-01
4.03787822e-01 2.22609520e-01 -1.13351262e+00 3.54851395e-01
4.49504316e-01 -1.10368419e+00 9.33090746e-01 -1.01608276e+00
3.35447490e-01 -3.31083745e-01 -1.47124305e-01 -1.08378124e+00
-8.10126126e-01 -3.18150192e-01 4.11174148e-02 4.52960610e-01
4.73739684e-01 -6.50773883e-01 9.48649228e-01 2.28635684e-01
3.15117866e-01 -1.02587402e+00 -8.49376142e-01 -7.73572147e-01
1.25900283e-01 3.69992137e-01 3.70630026e-01 7.30245709e-01
6.51643276e-01 3.94887358e-01 5.53880781e-02 -2.64010221e-01
4.80947137e-01 2.91214675e-01 3.66863400e-01 -1.64811409e+00
-1.19875707e-01 -3.01628113e-01 -1.02665460e+00 -7.13821888e-01
2.98385471e-02 -9.55920458e-01 -2.71257401e-01 -1.20135891e+00
2.49965876e-01 -2.09112078e-01 -2.67707825e-01 -1.22422881e-01
2.47690260e-01 -2.49502599e-01 -1.49913570e-02 3.54154527e-01
-5.14502943e-01 9.12591442e-02 1.11881411e+00 -2.19130120e-03
-1.68137446e-01 -1.39325663e-01 -3.89949352e-01 5.24410069e-01
9.39604700e-01 -1.42590389e-01 -2.29475304e-01 3.69726747e-01
2.97485977e-01 2.48389184e-01 -2.61325166e-02 -9.02352750e-01
-3.21487859e-02 -2.30366826e-01 3.50540519e-01 -3.84835005e-01
3.07563275e-01 -9.37152863e-01 5.07463574e-01 8.02631915e-01
-2.22323865e-01 1.95209563e-01 -8.12456086e-02 8.46404612e-01
-5.84077984e-02 -1.18773870e-01 7.92692244e-01 -1.37031868e-01
-3.31939340e-01 -1.16927093e-02 -6.47410333e-01 -3.31217736e-01
1.20097327e+00 -5.56623399e-01 -3.33978534e-01 2.49307051e-01
-1.10435855e+00 -2.95551091e-01 1.03005898e+00 -3.48046497e-02
2.91448712e-01 -1.00229645e+00 -3.04598689e-01 2.66499519e-01
3.64413023e-01 -8.94669116e-01 1.48544712e-02 7.79296339e-01
-1.03910148e+00 9.76981461e-01 -6.35550082e-01 -5.64065218e-01
-1.58957875e+00 1.15060973e+00 2.68608838e-01 -3.60856861e-01
-3.43201697e-01 1.17070213e-01 2.87529826e-01 -2.04683095e-01
-1.88874021e-01 -5.33953488e-01 -5.85804939e-01 1.28564745e-01
1.38841212e-01 3.30843747e-01 1.79027632e-01 -9.31970477e-01
-5.54545522e-01 8.99141967e-01 1.13709718e-01 5.43139994e-01
1.44023156e+00 -3.13021950e-02 -5.86155713e-01 4.54751343e-01
1.21366715e+00 2.26102266e-02 -7.01207578e-01 -2.66162038e-01
4.78593439e-01 -1.60157993e-01 -7.44665205e-01 -4.10185039e-01
-5.63775599e-01 5.71229279e-01 3.44498605e-01 6.83548450e-01
7.83104956e-01 2.52831578e-01 2.04708129e-01 6.06899381e-01
6.23251379e-01 -6.36079431e-01 -2.45154217e-01 5.32926679e-01
5.98246396e-01 -7.52278924e-01 5.30896112e-02 -9.33550119e-01
-2.59341449e-01 1.42358375e+00 -1.48124129e-01 -2.43995100e-01
6.68318808e-01 1.44741312e-01 -2.64789760e-01 -4.50473070e-01
-7.90478110e-01 -4.44860803e-03 2.19985694e-01 5.87924302e-01
7.79394627e-01 4.63138908e-01 -1.12659597e+00 4.64636624e-01
-8.22855458e-02 -4.06554103e-01 5.02773643e-01 7.93542385e-01
-9.38518226e-01 -1.45416033e+00 -2.37108767e-01 4.45370495e-01
-3.33964378e-01 1.78499192e-01 -8.39181840e-01 7.83887446e-01
-7.31871501e-02 6.63005710e-01 -1.60025179e-01 -4.24192488e-01
4.57203329e-01 2.93894976e-01 5.29071152e-01 -2.18349561e-01
-9.93938297e-02 -3.08075935e-01 2.19043076e-01 -3.80272746e-01
-3.60948771e-01 -6.44221127e-01 -1.13615394e+00 -2.69971430e-01
-1.47439301e-01 7.25368917e-01 7.50073910e-01 5.92715681e-01
2.89244264e-01 3.06437820e-01 5.62807798e-01 -6.65964365e-01
-3.39151561e-01 -6.14379466e-01 -1.14392531e+00 5.63361466e-01
2.18389168e-01 -7.93572903e-01 -3.24783444e-01 8.21611434e-02] | [4.989155292510986, 5.3171515464782715] |
f072e057-9546-4f02-a2df-7a5b9325328e | learning-integrable-dynamics-with-action | 2211.15338 | null | https://arxiv.org/abs/2211.15338v1 | https://arxiv.org/pdf/2211.15338v1.pdf | Learning Integrable Dynamics with Action-Angle Networks | Machine learning has become increasingly popular for efficiently modelling the dynamics of complex physical systems, demonstrating a capability to learn effective models for dynamics which ignore redundant degrees of freedom. Learned simulators typically predict the evolution of the system in a step-by-step manner with numerical integration techniques. However, such models often suffer from instability over long roll-outs due to the accumulation of both estimation and integration error at each prediction step. Here, we propose an alternative construction for learned physical simulators that are inspired by the concept of action-angle coordinates from classical mechanics for describing integrable systems. We propose Action-Angle Networks, which learn a nonlinear transformation from input coordinates to the action-angle space, where evolution of the system is linear. Unlike traditional learned simulators, Action-Angle Networks do not employ any higher-order numerical integration methods, making them extremely efficient at modelling the dynamics of integrable physical systems. | ['Shirley Ho', 'Tess Smidt', 'Miles Cranmer', 'Arthur Kosmala', 'Ameya Daigavane'] | 2022-11-24 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-3.26851398e-01 9.87416059e-02 3.33173424e-02 8.23753476e-02
-3.82156342e-01 -4.07948941e-01 6.39507830e-01 -1.45728484e-01
-3.35665375e-01 1.32425928e+00 -7.09978998e-01 -2.15086132e-01
-3.83349121e-01 -9.95159209e-01 -8.49165499e-01 -8.96016419e-01
-4.98355478e-01 4.88301039e-01 8.41217339e-02 -7.71832168e-01
-4.10933644e-02 7.04943120e-01 -8.99108887e-01 -4.00506467e-01
1.05097497e+00 6.24842584e-01 -2.31872216e-01 9.40431893e-01
2.17012689e-01 6.68751061e-01 -4.79200035e-01 1.28566682e-01
1.96371958e-01 -5.97144723e-01 -4.51511770e-01 -4.52001601e-01
6.81563094e-02 8.54373258e-03 -1.01751161e+00 1.07174051e+00
2.42352903e-01 3.76260132e-01 7.56371677e-01 -1.21367681e+00
-3.70673269e-01 3.74628603e-01 2.39162996e-01 -1.27128452e-01
-3.39365341e-02 6.53687954e-01 7.65467942e-01 -4.96587425e-01
6.47924185e-01 1.00674713e+00 1.05911446e+00 4.10929531e-01
-1.49403369e+00 -5.64872146e-01 -3.46667200e-01 2.94351548e-01
-1.45030606e+00 1.84440851e-01 7.96886384e-01 -5.28961241e-01
9.09945905e-01 2.86211241e-02 1.00475252e+00 8.74040902e-01
1.11167765e+00 -4.38179448e-02 9.91327465e-01 -3.00684005e-01
6.93693280e-01 -9.88437161e-02 5.35488129e-04 7.13458478e-01
2.64358491e-01 5.95358670e-01 -2.51222908e-01 -2.81268269e-01
1.07355762e+00 -1.99767590e-01 -1.59543961e-01 -8.53505194e-01
-9.11142826e-01 9.58409011e-01 5.12206912e-01 2.08485007e-01
-4.18212473e-01 5.01601756e-01 2.61762351e-01 4.55649614e-01
7.81602412e-02 9.51209247e-01 -5.50680935e-01 -2.88086474e-01
-6.22086167e-01 7.55757034e-01 9.42020893e-01 4.46524471e-01
8.15979779e-01 5.60963213e-01 3.61737579e-01 8.04548115e-02
-1.20317772e-01 5.81491649e-01 1.69529229e-01 -1.13329375e+00
6.18666708e-02 5.51900446e-01 3.99500817e-01 -7.62162983e-01
-6.60126388e-01 -5.10202050e-01 -1.40233600e+00 6.38113320e-01
2.61525661e-01 -3.40502501e-01 -4.74676162e-01 1.66147041e+00
3.06650043e-01 8.65151063e-02 1.14763014e-01 5.56926727e-01
6.73669763e-03 8.24838281e-01 -2.65868813e-01 -3.01784754e-01
5.93928158e-01 -5.97679555e-01 -6.25201702e-01 3.74728650e-01
8.36574078e-01 -1.04284259e-02 8.84467661e-01 3.71226877e-01
-1.23715270e+00 -5.31497121e-01 -1.40628040e+00 2.93825030e-01
-3.85078609e-01 -2.51545042e-01 4.20658916e-01 8.12465474e-02
-1.01469898e+00 1.74160147e+00 -1.05701530e+00 1.99154764e-01
-2.83708866e-03 6.52634025e-01 -4.15329844e-01 7.67573774e-01
-1.47945035e+00 1.24700308e+00 2.48954788e-01 -4.45826836e-02
-7.89203405e-01 -1.09001875e+00 -6.87392175e-01 1.13353148e-01
2.34950498e-01 -6.86104178e-01 1.29582393e+00 -4.93766010e-01
-2.06471944e+00 -3.74602526e-02 2.33432874e-01 -7.05473661e-01
4.97349828e-01 -4.52427231e-02 -3.21516305e-01 -5.99733852e-02
-3.74796480e-01 3.76017764e-02 8.68887722e-01 -8.61070931e-01
1.58405095e-01 4.52480279e-02 2.15062663e-01 -1.95043966e-01
-1.26046613e-01 -7.74854243e-01 5.82146883e-01 -2.73476154e-01
-9.05134231e-02 -1.03830564e+00 -7.65539467e-01 2.54425466e-01
-1.40272230e-01 -9.70883220e-02 9.11273599e-01 -3.54887635e-01
9.79681909e-01 -1.74324203e+00 5.76724768e-01 1.47962183e-01
4.64819312e-01 5.51817119e-01 1.52562425e-01 9.62142885e-01
-1.96327686e-01 1.28621636e-02 -3.46574515e-01 9.88205373e-02
3.56034264e-02 2.21406698e-01 -5.71500897e-01 3.56877804e-01
3.92154127e-01 9.47152793e-01 -8.97935152e-01 -1.99505538e-01
2.10033417e-01 7.88273990e-01 -5.99172533e-01 1.45219356e-01
-4.24672425e-01 9.89034593e-01 -5.79101562e-01 -1.64802462e-01
4.29527640e-01 -2.89362580e-01 6.77599460e-02 -1.83625832e-01
-4.52567279e-01 2.94411778e-01 -9.78783965e-01 1.02104199e+00
-5.98442912e-01 4.54641283e-01 2.34588347e-02 -1.44721949e+00
8.02796960e-01 4.05126214e-01 6.81751311e-01 -7.52722979e-01
1.52711526e-01 2.03887299e-01 4.54038203e-01 -2.50809878e-01
1.63269833e-01 -3.14324617e-01 -3.33213627e-01 2.00358912e-01
2.09970132e-01 -7.87213266e-01 1.35665268e-01 -2.36758892e-03
1.27597821e+00 2.46806026e-01 6.15727782e-01 -2.77115196e-01
8.31018686e-01 2.18696594e-01 5.98303020e-01 8.00087392e-01
-5.01511768e-02 6.38349280e-02 6.36226833e-01 -6.94445252e-01
-1.50514424e+00 -1.13723433e+00 -1.89571187e-01 1.75508782e-01
-8.34027585e-03 -5.20873010e-01 -8.23883235e-01 -5.80666773e-02
1.94927365e-01 4.94377255e-01 -6.82463169e-01 -6.51872039e-01
-1.02454233e+00 -4.51447904e-01 5.04290283e-01 2.74417490e-01
3.75814289e-01 -1.08470988e+00 -6.72758400e-01 5.64481854e-01
4.54284877e-01 -8.64278734e-01 -4.92715575e-02 2.38933608e-01
-1.02531421e+00 -1.12600625e+00 -3.75564575e-01 -3.08224976e-01
3.06103021e-01 -4.10448581e-01 7.81277001e-01 6.38003051e-02
-5.64589679e-01 2.10300833e-01 2.83947319e-01 9.38428566e-02
-7.16117501e-01 4.12534252e-02 6.08397603e-01 -4.72407579e-01
-4.33144331e-01 -1.07216322e+00 -4.51765031e-01 2.27439582e-01
-6.67402565e-01 8.46007839e-02 3.97044927e-01 1.04802406e+00
4.56552237e-01 1.26573622e-01 6.62667930e-01 -1.93309113e-01
6.04947150e-01 -2.84161955e-01 -1.01580274e+00 1.94780082e-02
-3.11806202e-01 3.92407119e-01 1.47533739e+00 -6.91460431e-01
-5.26106596e-01 -4.43278216e-02 -1.02095552e-01 -2.69770026e-01
3.06352794e-01 3.72700840e-01 2.13673070e-01 -6.82794571e-01
5.87660909e-01 4.30390567e-01 2.17275426e-01 -2.94430882e-01
3.08631629e-01 7.90717378e-02 5.49397588e-01 -6.87997162e-01
1.19347012e+00 1.81314334e-01 7.88217604e-01 -9.56400990e-01
-3.98590356e-01 2.74994612e-01 -9.07197595e-01 -1.53733000e-01
6.18839502e-01 -3.11816692e-01 -1.34764028e+00 7.11323142e-01
-1.05558670e+00 -5.77438951e-01 -7.12624967e-01 5.95793128e-01
-9.53439176e-01 1.95622459e-01 -7.14285076e-01 -9.27453697e-01
-1.92187786e-01 -9.63207603e-01 5.18091023e-01 2.93641835e-01
-2.24579841e-01 -1.11623406e+00 5.76270998e-01 -6.73935533e-01
7.67784297e-01 4.23506498e-01 1.33717501e+00 -1.97671309e-01
-6.49863839e-01 -3.20135683e-01 3.65610868e-01 1.74653023e-01
-1.84743047e-01 2.53015280e-01 -4.37248409e-01 -2.95135587e-01
2.90743619e-01 -2.06240281e-01 3.11196148e-01 1.93767563e-01
9.33183551e-01 -4.39293414e-01 -3.51236641e-01 5.80297351e-01
1.29781234e+00 3.68118376e-01 3.51028949e-01 -5.60370879e-03
5.93007863e-01 1.82772174e-01 4.86320220e-02 2.61708617e-01
2.75340937e-02 9.14524436e-01 3.05705518e-01 1.55942664e-01
3.09331030e-01 -8.50711316e-02 3.86923254e-01 1.28973246e+00
-1.70409098e-01 3.37628573e-01 -8.12608242e-01 8.38946179e-02
-1.67209852e+00 -8.33028615e-01 -5.18334545e-02 2.37472129e+00
8.83762121e-01 2.62962312e-01 9.99254361e-02 1.06076561e-01
3.38561803e-01 -1.55186430e-01 -8.93296838e-01 -5.01216650e-01
-2.52240859e-02 4.78562266e-01 2.43318960e-01 7.37587988e-01
-6.67037249e-01 6.66561842e-01 7.01506329e+00 6.70881987e-01
-1.49302387e+00 1.54918209e-02 1.68317035e-01 2.30408654e-01
1.19003423e-01 3.55638772e-01 -5.91241360e-01 3.60147446e-01
1.23730326e+00 -6.24898136e-01 6.10509276e-01 7.73753285e-01
3.46521020e-01 4.62759584e-02 -1.00554264e+00 4.99382019e-01
-7.07204521e-01 -1.50318909e+00 -1.91093445e-01 2.22359866e-01
8.15036714e-01 -2.75070310e-01 -2.86905225e-02 5.78820348e-01
2.90782541e-01 -1.12472415e+00 3.71743798e-01 9.48217154e-01
5.02292275e-01 -6.98100030e-01 4.56565589e-01 7.17714846e-01
-1.09503245e+00 -5.80529729e-03 -3.98016542e-01 -5.16417742e-01
3.96593422e-01 4.05307770e-01 -4.10833895e-01 5.11055052e-01
1.87176436e-01 6.03658438e-01 -1.68434426e-01 1.00147188e+00
1.68759286e-01 4.45860624e-01 -4.04779941e-01 -4.53075111e-01
2.77696580e-01 -8.13907385e-01 5.56074679e-01 4.01436359e-01
5.64818025e-01 1.17877468e-01 -1.13218157e-02 1.11355352e+00
2.45722830e-01 -2.59725302e-01 -8.91781926e-01 -2.19325498e-01
1.74745560e-01 9.73445654e-01 -3.69402766e-01 -2.75639862e-01
-2.51037218e-02 3.46434176e-01 2.96814770e-01 4.18146551e-01
-1.05943120e+00 -4.20256108e-01 7.76036382e-01 1.21671043e-01
4.12700057e-01 -9.17369962e-01 1.86424255e-02 -1.03577685e+00
-8.60560834e-02 -6.71429038e-01 -4.30899292e-01 -6.23008192e-01
-1.08944869e+00 4.26571190e-01 7.52870142e-02 -1.35236013e+00
-5.39555728e-01 -7.33048856e-01 -7.87265718e-01 6.57300889e-01
-8.52090895e-01 -4.50544715e-01 9.82555673e-02 3.57038349e-01
-2.64049694e-02 -1.08692147e-01 1.11665344e+00 -1.09714463e-01
-3.95860702e-01 2.93419689e-01 6.85910702e-01 -1.96671888e-01
2.35635892e-01 -1.32468987e+00 4.80603456e-01 1.24263979e-01
-2.15247855e-01 6.56108201e-01 1.09016430e+00 -5.94025373e-01
-1.59991682e+00 -7.09503174e-01 4.41434145e-01 -1.14995763e-01
1.08675361e+00 -3.94470811e-01 -1.14642167e+00 3.07541460e-01
4.94653098e-02 1.72799021e-01 1.28178030e-01 -1.70656189e-01
-1.00079700e-01 -2.86246061e-01 -1.05281663e+00 8.25661063e-01
8.15495372e-01 -5.23881018e-01 -2.72966802e-01 3.36275280e-01
7.50206947e-01 -2.55469263e-01 -1.10083151e+00 5.19825518e-01
6.02992833e-01 -7.06961691e-01 9.86772001e-01 -9.22852755e-01
3.11389923e-01 -1.53048530e-01 2.53759474e-01 -1.63367999e+00
-2.14351505e-01 -1.06749523e+00 -5.69706321e-01 4.23159897e-01
9.48053449e-02 -9.78375196e-01 5.58090985e-01 9.84704942e-02
7.73100406e-02 -1.06912446e+00 -1.27547801e+00 -1.12016380e+00
7.73390412e-01 -6.98299613e-03 3.66214365e-01 7.65342474e-01
2.95968592e-01 1.42118677e-01 -3.15834165e-01 1.58256546e-01
6.81104064e-01 -5.60428463e-02 7.83763051e-01 -1.16070402e+00
-5.64374328e-01 -5.52322745e-01 -7.09371686e-01 -7.46641517e-01
2.97636241e-01 -4.44557011e-01 -3.80680524e-02 -9.59882438e-01
-3.10653478e-01 -1.81462064e-01 -1.03806734e-01 -8.81117061e-02
2.25725189e-01 -5.31451628e-02 8.71913787e-03 1.47883296e-01
-2.57008851e-01 1.03679848e+00 1.39846814e+00 1.69802338e-01
-3.45049828e-01 2.03961924e-01 1.30667254e-01 8.84922326e-01
9.34929311e-01 -5.61111033e-01 -3.26102734e-01 3.37979853e-01
3.99047256e-01 5.03568470e-01 6.92663789e-01 -1.51184905e+00
3.35689455e-01 -3.98620188e-01 1.72787890e-01 -6.29220366e-01
7.37826049e-01 -4.11571413e-01 3.20355594e-01 8.84617090e-01
-2.66675204e-01 3.50618690e-01 1.21250942e-01 4.17964548e-01
-2.38055944e-01 -2.99029667e-02 1.01600111e+00 -6.94330260e-02
-6.07789494e-02 3.60038280e-01 -8.12729716e-01 3.24076079e-02
8.21537316e-01 2.10158959e-01 -2.65414834e-01 -5.59217930e-01
-9.78756309e-01 2.20902450e-02 5.07481456e-01 -1.36757344e-01
8.63461718e-02 -1.33270657e+00 -2.50132024e-01 1.63913861e-01
-4.08114225e-01 -1.82476237e-01 3.85164887e-01 8.23679864e-01
-7.96241164e-01 5.97308099e-01 -4.26834196e-01 -5.31158984e-01
-5.90390682e-01 3.12902719e-01 9.40801919e-01 -7.94242442e-01
-7.73906410e-01 1.83543622e-01 6.64448366e-02 -8.41554463e-01
-2.85273224e-01 -3.24671030e-01 2.52419949e-01 -5.80925167e-01
1.96298644e-01 3.46825689e-01 -8.13877434e-02 -4.14302111e-01
4.31730039e-03 8.14188242e-01 2.60289133e-01 -2.31098697e-01
1.42817426e+00 4.30270940e-01 -1.36539936e-01 7.67783940e-01
1.05777323e+00 -4.01439786e-01 -1.56016421e+00 -8.46858844e-02
-1.06954411e-01 1.39307708e-01 -2.63992041e-01 -3.89328599e-01
-5.16103506e-01 1.11192465e+00 3.37413102e-01 4.53533649e-01
4.64552283e-01 -4.89765376e-01 7.53928006e-01 1.04145956e+00
5.74765861e-01 -1.11477673e+00 1.74429074e-01 1.08984804e+00
9.00138259e-01 -6.41758800e-01 8.00047163e-03 -1.74157530e-01
9.85846892e-02 1.34779394e+00 6.14357233e-01 -8.57020617e-01
8.75618041e-01 4.29539889e-01 -3.51255387e-01 6.13124594e-02
-8.36132109e-01 4.42968398e-01 8.53491798e-02 5.73707342e-01
8.47280324e-02 6.49017375e-03 -3.33537370e-01 4.50696081e-01
-4.08841848e-01 -4.01181281e-02 7.47336328e-01 8.17500234e-01
-5.23378670e-01 -1.29962170e+00 -2.16963410e-01 -3.14571895e-03
1.22618362e-01 2.41846144e-01 -1.94042385e-01 1.40187168e+00
-1.79943636e-01 2.17567310e-01 -8.74780938e-02 -2.67747521e-01
3.98066491e-01 3.24944168e-01 7.81657338e-01 -3.34932834e-01
-4.92834449e-01 -4.75880086e-01 -2.91861832e-01 -5.56814492e-01
1.56508133e-01 -4.45237279e-01 -1.32504940e+00 -4.58285034e-01
-1.44333318e-01 4.82138634e-01 6.00697994e-01 1.16468668e+00
3.81593853e-01 7.40120351e-01 6.85047388e-01 -1.15500927e+00
-1.31134510e+00 -8.78619790e-01 -5.30031621e-01 -5.01351207e-02
6.80908978e-01 -8.71376991e-01 -4.90142047e-01 -3.23287845e-01] | [6.531064510345459, 3.466583251953125] |
e43ef4f1-0af7-49ab-8bad-e74b6b26506d | predictive-inference-with-feature-conformal | 2210.00173 | null | https://arxiv.org/abs/2210.00173v4 | https://arxiv.org/pdf/2210.00173v4.pdf | Predictive Inference with Feature Conformal Prediction | Conformal prediction is a distribution-free technique for establishing valid prediction intervals. Although conventionally people conduct conformal prediction in the output space, this is not the only possibility. In this paper, we propose feature conformal prediction, which extends the scope of conformal prediction to semantic feature spaces by leveraging the inductive bias of deep representation learning. From a theoretical perspective, we demonstrate that feature conformal prediction provably outperforms regular conformal prediction under mild assumptions. Our approach could be combined with not only vanilla conformal prediction, but also other adaptive conformal prediction methods. Apart from experiments on existing predictive inference benchmarks, we also demonstrate the state-of-the-art performance of the proposed methods on large-scale tasks such as ImageNet classification and Cityscapes image segmentation.The code is available at \url{https://github.com/AlvinWen428/FeatureCP}. | ['Yang Yuan', 'Yang Gao', 'Yoshua Bengio', 'Dinghuai Zhang', 'Chuan Wen', 'Jiaye Teng'] | 2022-10-01 | null | null | null | null | ['prediction-intervals'] | ['miscellaneous'] | [ 8.59366357e-02 4.32090491e-01 -3.77831787e-01 -7.60421515e-01
-9.94458973e-01 -5.45425177e-01 1.03595054e+00 1.17317416e-01
7.05744475e-02 7.09203601e-01 3.50096017e-01 -4.43734884e-01
-4.66678053e-01 -1.09284747e+00 -8.65088820e-01 -5.47300696e-01
3.86767127e-02 7.62596369e-01 2.38734603e-01 1.14277713e-01
2.11454138e-01 1.88063845e-01 -1.34939039e+00 2.56845862e-01
1.05943763e+00 9.57279146e-01 -2.68144906e-01 4.21201378e-01
3.60589102e-02 7.21857131e-01 1.14360198e-01 -8.94370258e-01
5.65030165e-02 -1.18825227e-01 -7.75283039e-01 -6.16675615e-01
3.92154038e-01 -1.50708571e-01 -3.32354307e-01 9.54534113e-01
1.39358506e-01 1.61760107e-01 1.04479468e+00 -1.42962146e+00
-1.12520504e+00 8.08295667e-01 -4.25989479e-01 1.92447528e-01
8.35574046e-02 -3.13302219e-01 1.14906967e+00 -9.41365302e-01
3.53528529e-01 1.06888020e+00 1.11197007e+00 3.39873701e-01
-1.32128108e+00 -9.23005879e-01 7.30797276e-02 3.23787719e-01
-1.43004310e+00 -1.90406650e-01 5.50451756e-01 -6.05797172e-01
6.04325056e-01 2.97371566e-01 5.78033328e-01 1.32786834e+00
3.46698076e-01 7.30844200e-01 1.20267081e+00 -3.51652175e-01
2.06456453e-01 1.31024420e-01 1.91901907e-01 6.77695274e-01
1.09212557e-02 3.56015414e-01 -3.27021852e-02 -4.51553524e-01
6.27350628e-01 2.35119984e-01 -7.67136291e-02 -3.50060970e-01
-9.91114557e-01 1.24032915e+00 5.22687852e-01 8.75520706e-02
4.27226983e-02 5.80590248e-01 1.25469401e-01 1.12007065e-02
9.50324476e-01 -3.25293615e-02 -4.50547636e-01 -7.92529806e-02
-8.23581636e-01 4.16170627e-01 5.99688411e-01 1.14680505e+00
5.34624398e-01 -2.97777921e-01 -4.33420628e-01 6.29787564e-01
3.64485741e-01 5.55131793e-01 1.28150076e-01 -9.99374390e-01
2.12608159e-01 3.33639503e-01 1.46159960e-03 -8.77005219e-01
-4.31929588e-01 -2.98033148e-01 -7.09881306e-01 -1.50284588e-01
4.40573186e-01 8.18274915e-03 -7.19370067e-01 1.67138529e+00
2.96172142e-01 5.68611920e-01 -1.53983399e-01 6.15537643e-01
6.33087933e-01 6.33373141e-01 4.20509160e-01 2.84778178e-01
1.06469750e+00 -1.00242138e+00 -2.43680999e-01 2.74669230e-01
5.57983458e-01 -6.63994372e-01 1.18915617e+00 1.19939715e-01
-7.62069881e-01 -4.18917269e-01 -8.10149610e-01 -2.31236324e-01
-4.12781268e-01 -4.27308753e-02 9.25775230e-01 5.88272095e-01
-1.00781643e+00 7.59182990e-01 -9.00335252e-01 -4.09333348e-01
7.86536753e-01 7.00809434e-02 -1.94632590e-01 -8.82765129e-02
-1.22488010e+00 6.83663130e-01 -5.12435660e-02 -3.54009867e-01
-5.69206178e-01 -1.03223395e+00 -7.31583834e-01 2.27506876e-01
-6.23904951e-02 -7.59923339e-01 1.28974962e+00 -8.02126586e-01
-1.13723588e+00 5.24883509e-01 -9.97251645e-02 -5.82688570e-01
6.37478411e-01 -3.10415596e-01 -3.78749937e-01 -2.00400367e-01
1.88899264e-01 6.49562299e-01 4.88600105e-01 -1.02132630e+00
-6.32355273e-01 -4.28030133e-01 1.52416945e-01 -5.06015830e-02
-3.99412096e-01 -1.55587822e-01 -3.59193355e-01 -6.19326591e-01
-2.40632713e-01 -1.18264019e+00 -4.23963577e-01 3.09261605e-02
-4.70609933e-01 -7.58569419e-01 3.90140533e-01 -4.05322552e-01
1.14530325e+00 -2.05804276e+00 -6.62486702e-02 4.73253876e-01
-3.06651840e-04 -2.76085407e-01 1.02013513e-01 3.47608238e-01
-5.32244891e-02 2.47165978e-01 -3.41534078e-01 -5.74138284e-01
4.39228445e-01 5.96196428e-02 -7.67277896e-01 6.52291894e-01
2.96397950e-03 1.20617807e+00 -7.66044080e-01 -5.58965206e-01
1.67392731e-01 9.04264987e-01 -6.08904600e-01 -3.36266577e-01
-2.14141414e-01 4.23217356e-01 -6.37973547e-01 4.33959395e-01
5.52366495e-01 -4.90359843e-01 -3.51931602e-02 1.94605112e-01
1.41117111e-01 1.53144583e-01 -7.37162530e-01 1.64859283e+00
-5.34240186e-01 5.48735380e-01 -8.90361309e-01 -1.03991425e+00
8.66417170e-01 1.51302516e-01 4.01615709e-01 -7.16112435e-01
-1.84711833e-02 -1.05029613e-01 -5.01868486e-01 6.31914660e-02
4.43643451e-01 -5.13459109e-02 -3.01103562e-01 3.74563962e-01
-4.66545299e-02 5.35777323e-02 -3.13447803e-01 6.33396581e-02
9.23417926e-01 6.68730259e-01 1.78484976e-01 -7.53007174e-01
2.19532743e-01 -1.68713499e-02 4.86693650e-01 7.38594294e-01
4.71681468e-02 6.19088829e-01 5.31843901e-01 -7.09203243e-01
-1.19978595e+00 -1.60318887e+00 -8.41763198e-01 1.15159523e+00
1.13799430e-01 -5.60692012e-01 -8.08627009e-01 -1.01840353e+00
1.81849450e-01 1.08871627e+00 -1.12130475e+00 -3.21123227e-02
-1.35276005e-01 -6.65230751e-01 5.99510312e-01 8.55548799e-01
1.33966014e-01 -6.04198217e-01 -1.70091391e-01 -1.47967964e-01
3.67847993e-03 -7.67256677e-01 -3.04715991e-01 2.13734452e-02
-7.59640813e-01 -9.41987634e-01 -3.99344683e-01 -4.48226362e-01
6.27999425e-01 -2.01425731e-01 1.21267748e+00 -9.56348050e-03
2.96386555e-02 5.24758101e-01 -2.69385874e-01 -6.00426018e-01
-1.88669592e-01 2.49702930e-01 -7.69062117e-02 -1.72370002e-01
7.01495767e-01 -6.29341483e-01 -8.07338476e-01 3.20037633e-01
-5.74576795e-01 2.16015339e-01 3.28953341e-02 7.03933656e-01
1.02226245e+00 -5.90988435e-02 8.17690551e-01 -1.29306257e+00
3.45957041e-01 -8.40887189e-01 -6.42113507e-01 2.04085082e-01
-8.26729774e-01 -6.67077005e-02 6.16757452e-01 -1.75887793e-01
-1.04132617e+00 -7.14834854e-02 -1.80615842e-01 -4.37559694e-01
-3.03029865e-01 2.71261096e-01 1.81185052e-01 3.39148194e-01
6.56006813e-01 4.63419594e-02 -2.88701206e-01 -3.34612668e-01
5.29375434e-01 4.93147314e-01 1.68066233e-01 -6.10243261e-01
8.19891095e-01 7.24976122e-01 3.08312126e-03 -2.91464001e-01
-1.18874836e+00 -8.94609187e-03 -6.15227222e-01 -2.23318525e-02
9.59573746e-01 -9.29782510e-01 -4.01460618e-01 -2.90688753e-01
-7.76304722e-01 -5.60414791e-01 -4.09681767e-01 3.83079231e-01
-8.46974254e-01 2.65697204e-02 -4.65710789e-01 -4.86930668e-01
-3.69569659e-01 -9.89122570e-01 1.10752153e+00 2.15920120e-01
-4.02528912e-01 -1.37761700e+00 2.85903782e-01 2.67751336e-01
4.05394226e-01 1.92188904e-01 1.01935720e+00 -8.91172588e-01
-3.80774498e-01 -2.32725501e-01 -2.05553383e-01 2.75292881e-02
-1.88655302e-01 3.58280018e-02 -1.22237957e+00 -2.54728477e-02
-5.70626915e-01 -5.86632714e-02 9.96732652e-01 5.65380931e-01
1.83605206e+00 -2.11650982e-01 -6.52059376e-01 6.92200243e-01
1.40397263e+00 -1.67904407e-01 8.13741803e-01 2.85624504e-01
5.94797790e-01 4.82341290e-01 6.12260044e-01 4.99594003e-01
9.02433574e-01 8.21134508e-01 1.84824616e-01 3.25185359e-02
-3.68098319e-02 -6.43610001e-01 -1.01632021e-01 2.65945226e-01
-3.53052437e-01 -4.00004759e-02 -9.03783560e-01 5.33360541e-01
-2.00030684e+00 -1.08689058e+00 -2.43554235e-01 2.17848969e+00
6.75862491e-01 -5.00965863e-02 -3.39739360e-02 -1.44396946e-01
5.18623471e-01 -2.30136290e-01 -4.20505047e-01 -5.96316993e-01
1.38276502e-01 3.41546506e-01 4.61689204e-01 5.01626253e-01
-1.51567471e+00 8.14263761e-01 5.58170700e+00 1.07384777e+00
-6.62970185e-01 5.14009953e-01 1.07403755e+00 -9.68803093e-02
-9.61040378e-01 1.44743904e-01 -9.00895715e-01 6.58718824e-01
1.08350086e+00 -3.13607603e-02 2.87809014e-01 1.10319984e+00
-3.54149967e-01 3.32700580e-01 -1.08086538e+00 5.97811162e-01
-1.21518530e-01 -1.47893596e+00 -1.68030560e-01 2.25753784e-01
9.31216836e-01 3.12638491e-01 3.91677588e-01 5.05029917e-01
5.73332429e-01 -1.24185443e+00 7.01890707e-01 8.84477794e-01
8.40214908e-01 -1.00092149e+00 5.05750954e-01 1.14456348e-01
-1.10403907e+00 -1.17717035e-01 -5.84559560e-01 1.13699347e-01
-8.41002017e-02 7.00915158e-01 -5.03377795e-01 4.92201805e-01
1.06584370e+00 9.56658423e-01 -6.38538480e-01 8.84643018e-01
-2.99077839e-01 1.00978708e+00 -1.84897557e-01 1.75936267e-01
6.06218949e-02 -1.56401694e-01 1.64703667e-01 1.19799566e+00
7.07932353e-01 -1.49593074e-02 -7.11323842e-02 1.03017795e+00
2.98346747e-02 2.56242573e-01 -1.07784438e+00 4.20107454e-01
5.41951120e-01 1.20737994e+00 -6.69207454e-01 -1.43382356e-01
-6.93877816e-01 7.58747399e-01 6.36560023e-01 2.48827949e-01
-1.29850435e+00 -4.70291413e-02 5.56408644e-01 2.51628011e-01
5.02037942e-01 -4.06013727e-02 -5.26901782e-01 -9.24305499e-01
-1.44723862e-01 -5.07061407e-02 5.44455171e-01 -6.59952581e-01
-1.65647542e+00 5.32589614e-01 2.21217468e-01 -1.20087683e+00
-2.45328262e-01 -4.39744830e-01 -6.95122242e-01 6.74572051e-01
-1.59002244e+00 -1.51994038e+00 -1.58388689e-01 6.15741670e-01
2.96470553e-01 -1.44626483e-01 1.20158851e+00 2.41788879e-01
-1.70417637e-01 7.36744583e-01 4.50089395e-01 -9.42433625e-02
6.13324761e-01 -1.37357605e+00 5.47857285e-01 4.79764014e-01
3.95699322e-01 4.60829705e-01 4.61102873e-01 -5.82190514e-01
-8.44555557e-01 -1.44312000e+00 1.00519025e+00 -8.91219854e-01
7.72119343e-01 -3.81055027e-01 -7.02489853e-01 1.07142591e+00
6.81409761e-02 2.47424603e-01 1.13857317e+00 5.01247466e-01
-5.37874401e-01 -1.43003678e-02 -9.98728395e-01 5.28902113e-01
1.17253602e+00 -3.39396417e-01 -2.66593844e-01 4.94047701e-01
8.29636812e-01 -3.01778138e-01 -1.32512486e+00 5.51019132e-01
7.26772010e-01 -8.24649274e-01 1.10815465e+00 -6.52619362e-01
7.94805646e-01 1.22934971e-02 -4.91776377e-01 -1.14286327e+00
-6.73226833e-01 -3.42879854e-02 -9.30135399e-02 1.23139524e+00
7.18163729e-01 -9.19787109e-01 6.49400115e-01 7.16006637e-01
-2.23767027e-01 -1.18246448e+00 -1.13425064e+00 -6.13387465e-01
8.69839847e-01 -7.12475836e-01 1.03521609e+00 1.00752950e+00
-2.77418047e-02 1.11736935e-02 -2.48124406e-01 3.13449770e-01
7.45065987e-01 5.57798862e-01 5.87038338e-01 -1.51487875e+00
-4.20136005e-01 -3.50640982e-01 -4.93187249e-01 -6.81832194e-01
5.88763773e-01 -1.31748128e+00 -3.78205717e-01 -1.62653005e+00
4.37443137e-01 -1.00746131e+00 -4.21589762e-01 5.06305754e-01
3.57995182e-02 5.77570438e-01 -6.97963824e-03 1.66212529e-01
-6.35436356e-01 6.95680261e-01 1.14747632e+00 -4.61360961e-02
2.71330148e-01 2.64852732e-01 -8.85219097e-01 8.00999999e-01
1.20085800e+00 -4.06559199e-01 -7.22978413e-01 -4.61258203e-01
3.90778869e-01 -1.95747867e-01 7.57460713e-01 -8.41972589e-01
2.35681348e-02 -9.66211036e-02 7.86935329e-01 -3.41662705e-01
2.61732131e-01 -7.23564744e-01 2.39608780e-01 9.49321911e-02
-5.31169832e-01 9.25909951e-02 2.00661585e-01 9.03666139e-01
1.93153620e-01 -3.76088731e-02 4.95276034e-01 3.17473948e-01
-4.85703558e-01 5.94335377e-01 -8.71817488e-03 6.83965981e-02
1.19754505e+00 -2.24172622e-02 -6.57234132e-01 -3.69019598e-01
-8.19779217e-01 2.23375812e-01 5.04254341e-01 5.37961423e-01
5.79549730e-01 -1.60661161e+00 -7.51697659e-01 1.78082526e-01
2.40330771e-01 -1.80223659e-01 2.72841245e-01 9.72701848e-01
-1.50648668e-01 4.88072008e-01 1.61345601e-01 -6.58688605e-01
-7.66778171e-01 5.23903430e-01 2.96853393e-01 -3.10506709e-02
-1.04409242e+00 6.82334900e-01 5.27886569e-01 -8.28880131e-01
2.47072950e-02 -2.17174873e-01 -2.61847615e-01 -4.39707041e-01
4.66444850e-01 2.57508039e-01 -1.22823663e-01 -4.87617671e-01
-2.54240215e-01 6.03256643e-01 5.43697961e-02 1.19300067e-01
1.34504306e+00 -1.17191404e-01 1.60890222e-01 5.18978357e-01
9.93576765e-01 -1.18207194e-01 -1.30733848e+00 -1.78373098e-01
-1.46787494e-01 -5.01646221e-01 4.81684729e-02 -7.99346089e-01
-8.87769997e-01 8.80157530e-01 6.75190032e-01 1.75005466e-01
8.61372530e-01 4.89858508e-01 4.91884291e-01 9.55429301e-02
5.71573853e-01 -8.69876623e-01 -4.95892465e-01 2.92755038e-01
8.27826619e-01 -1.31342995e+00 -8.31799656e-02 -2.16336876e-01
-6.67709887e-01 8.47079217e-01 3.94362152e-01 -4.33899611e-01
1.30855513e+00 2.88295239e-01 -3.83362383e-01 -1.16778351e-01
-8.69465232e-01 3.31548937e-02 4.93621647e-01 6.13322616e-01
6.25039279e-01 5.20272493e-01 -1.46512538e-01 1.04319763e+00
-5.28623879e-01 5.15083559e-02 1.68145880e-01 3.16997111e-01
-5.65984808e-02 -1.05822086e+00 -1.33002298e-02 7.99054265e-01
-4.42516357e-01 -2.80095011e-01 1.12045631e-01 7.46010602e-01
-1.92297231e-02 6.95828497e-01 5.39459169e-01 -3.42630476e-01
9.81741250e-02 7.78201222e-02 5.72919846e-01 -2.96785802e-01
-3.82657275e-02 -3.13799322e-01 1.29136369e-02 -6.44898295e-01
-2.73448288e-01 -1.03473806e+00 -1.28432512e+00 -5.84017217e-01
-3.96301709e-02 3.82016227e-02 5.46303451e-01 8.37578058e-01
5.16545534e-01 3.61773551e-01 3.86418879e-01 -6.01132691e-01
-5.37442088e-01 -9.95476007e-01 -4.28303063e-01 2.51059532e-01
-8.04305673e-02 -9.10325885e-01 -1.60173938e-01 -9.90248621e-02] | [7.956416606903076, 4.150622367858887] |
c8eae135-2cbe-43a1-a83b-5c96e5e51570 | meci-a-multilingual-dataset-for-event | null | null | https://aclanthology.org/2022.coling-1.206 | https://aclanthology.org/2022.coling-1.206.pdf | MECI: A Multilingual Dataset for Event Causality Identification | Event Causality Identification (ECI) is the task of detecting causal relations between events mentioned in the text. Although this task has been extensively studied for English materials, it is under-explored for many other languages. A major reason for this issue is the lack of multilingual datasets that provide consistent annotations for event causality relations in multiple non-English languages. To address this issue, we introduce a new multilingual dataset for ECI, called MECI. The dataset employs consistent annotation guidelines for five typologically different languages, i.e., English, Danish, Spanish, Turkish, and Urdu. Our dataset thus enable a new research direction on cross-lingual transfer learning for ECI. Our extensive experiments demonstrate high quality for MECI that can provide ample research challenges and directions for future research. We will publicly release MECI to promote research on multilingual ECI. | ['Thien Huu Nguyen', 'Franck Dernoncourt', 'Minh Van Nguyen', 'Amir Pouran Ben Veyseh', 'Viet Dac Lai'] | null | null | null | null | coling-2022-10 | ['event-causality-identification'] | ['natural-language-processing'] | [-3.52641463e-01 1.81898556e-03 -5.43472886e-01 -2.11128861e-01
-7.03858733e-01 -6.12627685e-01 9.97314811e-01 2.71036446e-01
-3.71944696e-01 1.22896266e+00 6.33573115e-01 -4.11498934e-01
1.66534763e-02 -7.24192977e-01 -8.73268068e-01 -1.58227131e-01
-1.24365263e-01 3.83088112e-01 1.31403193e-01 1.06546640e-01
-3.88109200e-02 1.73838258e-01 -1.20159912e+00 5.14369190e-01
8.12269330e-01 2.75940210e-01 2.00570934e-02 1.00212023e-01
9.63659137e-02 1.15487313e+00 -4.91234481e-01 -7.09629416e-01
-4.11128968e-01 -6.00869238e-01 -1.13601172e+00 -6.61428392e-01
2.28704557e-01 1.32843807e-01 -3.91035646e-01 9.21383202e-01
3.34833413e-01 -2.35122189e-01 7.07591593e-01 -1.63992357e+00
-7.81882823e-01 1.25009120e+00 -5.22889674e-01 4.11121428e-01
5.54393470e-01 -4.04272348e-01 1.28285158e+00 -9.71173286e-01
1.20398462e+00 1.52031744e+00 4.40374941e-01 1.90706030e-01
-6.21128857e-01 -1.11871541e+00 2.86509275e-01 8.20285618e-01
-1.33420694e+00 -6.82352483e-02 4.92922634e-01 -4.49234754e-01
1.00222838e+00 3.76707017e-02 2.58408010e-01 1.56790113e+00
4.39695984e-01 9.81766701e-01 1.13371885e+00 -6.28117442e-01
-1.97781831e-01 3.22379731e-02 8.89237598e-02 4.31142300e-01
3.45149428e-01 1.42471448e-01 -8.98671806e-01 1.66292325e-01
4.35088336e-01 -5.35779536e-01 -1.73093349e-01 2.91652888e-01
-1.71314085e+00 8.41047585e-01 2.59107761e-02 6.47840798e-01
-1.32084519e-01 7.37463832e-02 8.66698384e-01 3.30507427e-01
4.88530636e-01 3.04792020e-02 -8.01936924e-01 -2.51617461e-01
-2.21009463e-01 3.47289056e-01 8.23885977e-01 1.00146282e+00
3.59311908e-01 -2.50600815e-01 1.02447510e-01 8.10162723e-01
3.99665028e-01 3.45429987e-01 1.22783102e-01 -4.96855497e-01
8.14898014e-01 6.34235919e-01 1.30471885e-01 -1.01918519e+00
-5.45131207e-01 2.81828493e-02 -6.48271024e-01 -3.27889502e-01
6.45840287e-01 -3.58656824e-01 -7.47613385e-02 2.03087234e+00
3.57883960e-01 3.01123053e-01 1.93639234e-01 7.50456154e-01
1.08896267e+00 6.98887229e-01 4.81513172e-01 -2.49785170e-01
1.47726285e+00 -6.71636581e-01 -1.04538333e+00 -1.69582218e-01
7.57975519e-01 -8.17042530e-01 1.07936311e+00 3.42530906e-01
-7.07393229e-01 -2.05417439e-01 -7.72020936e-01 -8.86711329e-02
-8.17194521e-01 3.64171833e-01 9.32102501e-01 2.72943974e-01
-6.28051043e-01 -1.55860171e-01 -6.37462914e-01 -6.48552477e-01
1.69482734e-02 -2.00340122e-01 -4.92582172e-01 3.79256941e-02
-2.17237496e+00 9.82421815e-01 8.50835860e-01 -1.04515972e-02
-9.21581030e-01 -9.76598024e-01 -1.00801122e+00 -2.44625583e-01
5.16109586e-01 -5.46697266e-02 1.15562856e+00 -5.23514509e-01
-7.81732857e-01 1.14404774e+00 -2.36469507e-01 -2.31915578e-01
2.73010284e-01 -3.69105846e-01 -1.08745170e+00 -2.84176618e-01
6.13839030e-01 1.98075250e-01 1.95320860e-01 -1.13315213e+00
-1.12626719e+00 -4.66084152e-01 -3.70382518e-02 -4.03293297e-02
-3.70854616e-01 9.51225042e-01 -6.10983968e-01 -9.55528498e-01
-4.86086577e-01 -9.71080303e-01 3.35711598e-01 -6.81727409e-01
-3.99303973e-01 -8.33058834e-01 8.47373664e-01 -8.34379554e-01
1.63946831e+00 -1.93418515e+00 -1.11917377e-01 -3.96329969e-01
-1.53405279e-01 -1.86030373e-01 7.36235082e-02 6.27958834e-01
-7.21108377e-01 4.79574740e-01 2.08806098e-01 1.57367423e-01
-6.08560070e-02 3.00414920e-01 -5.63942254e-01 3.45824420e-01
3.01729202e-01 9.79529917e-01 -1.45136261e+00 -7.59409308e-01
-2.49078721e-02 2.58363634e-01 -3.39529634e-01 -8.93756971e-02
-1.92747727e-01 7.15433240e-01 -4.10642207e-01 4.98273551e-01
1.55155122e-01 -1.74673483e-01 4.81507987e-01 -2.69817084e-01
-6.94709063e-01 7.30665803e-01 -1.27424562e+00 1.48289502e+00
-5.27154505e-01 8.95678401e-01 -2.84726471e-01 -1.01928616e+00
4.68388706e-01 9.57849562e-01 5.88629067e-01 -5.69468975e-01
-1.26299813e-01 4.66139078e-01 2.61962134e-02 -4.92343485e-01
3.23550820e-01 1.83228791e-01 -5.96106946e-01 3.70718569e-01
4.37875763e-02 4.94267017e-01 7.03141570e-01 2.26762027e-01
7.09967017e-01 3.07751238e-01 6.08626604e-01 -2.82375991e-01
5.82994461e-01 1.59765959e-01 1.12419832e+00 4.35409635e-01
-3.21983814e-01 -1.27350003e-03 5.86176991e-01 -3.71338189e-01
-6.57979965e-01 -8.36550772e-01 -4.13488626e-01 9.59051788e-01
-5.80186024e-02 -6.87052131e-01 -4.27545011e-01 -8.72310162e-01
-2.83474505e-01 8.95081401e-01 -6.44226015e-01 3.26008171e-01
-8.46762121e-01 -9.48869765e-01 7.15684116e-01 5.36717415e-01
5.92739284e-01 -1.30833411e+00 -1.72552750e-01 3.31959963e-01
-8.94929886e-01 -1.38731551e+00 -3.82804602e-01 -8.92916620e-02
-2.44429812e-01 -1.24185729e+00 -1.81640998e-01 -8.89021277e-01
1.34605598e-02 -1.06427826e-01 1.47879839e+00 -6.30224586e-01
-1.01957120e-01 3.27480555e-01 -4.29075003e-01 -8.16636622e-01
-6.78207934e-01 1.03601657e-01 2.09143251e-01 -3.09528977e-01
6.74089611e-01 -1.91907898e-01 -1.84406370e-01 4.66784775e-01
-5.93537569e-01 -5.38833328e-02 1.05167469e-02 5.46133101e-01
5.00127614e-01 4.44584042e-01 1.05963361e+00 -9.95796859e-01
2.81368285e-01 -8.72778952e-01 -4.75151479e-01 5.86844981e-01
-4.54283208e-01 -1.73577234e-01 5.05528927e-01 -2.55835950e-01
-1.72474813e+00 -5.90206981e-01 1.72834262e-01 3.08615983e-01
-2.92523116e-01 1.05980515e+00 -3.63091975e-01 5.87205112e-01
3.79733592e-01 -3.21188986e-01 -1.02067149e+00 -1.80852786e-01
4.04394984e-01 5.65289378e-01 6.13546968e-01 -8.60152245e-01
4.03951705e-01 3.09689194e-01 -4.10409093e-01 -7.26461887e-01
-1.02159822e+00 -2.11499751e-01 -5.82312703e-01 -4.44666535e-01
1.15464199e+00 -1.28737712e+00 -6.14723921e-01 4.29091334e-01
-1.50591838e+00 -4.29593563e-01 2.96943307e-01 7.93397188e-01
-8.55125189e-02 7.54155293e-02 -8.27830851e-01 -4.68039900e-01
-1.12647213e-01 -8.97294462e-01 8.45966220e-01 3.45296897e-02
-6.28781855e-01 -1.44543910e+00 4.05408472e-01 7.01106340e-02
-2.67443866e-01 3.62462878e-01 1.14267898e+00 -4.92736787e-01
-2.25295722e-01 1.42711163e-01 -4.32052195e-01 -2.23743290e-01
2.42969945e-01 3.51815224e-01 -6.08669281e-01 -2.27719755e-03
-3.28887939e-01 -5.98500848e-01 6.23478770e-01 2.16486424e-01
8.99897516e-01 -2.39092380e-01 -6.74201071e-01 -3.16306539e-02
1.35068917e+00 5.27517617e-01 5.58336735e-01 6.02683723e-01
8.38370264e-01 7.37117290e-01 8.81130695e-01 3.79734486e-01
1.11133897e+00 5.75701416e-01 -7.79311685e-03 -1.02946736e-01
-3.55410010e-01 -4.37630594e-01 5.50048470e-01 9.81765151e-01
-2.44085535e-01 -4.33510810e-01 -1.35857797e+00 1.05743158e+00
-1.89988303e+00 -1.04370809e+00 -7.99301744e-01 1.95341575e+00
1.30250287e+00 -7.82586038e-02 -1.62466958e-01 -2.11114287e-02
6.24682128e-01 -1.03810474e-01 -5.78186698e-02 -2.54139662e-01
-2.81646281e-01 2.74329316e-02 3.18481565e-01 2.54940927e-01
-1.42020190e+00 1.19119930e+00 6.31198215e+00 7.88116634e-01
-9.46880639e-01 4.45773363e-01 4.05497074e-01 4.66809988e-01
-3.00735295e-01 2.11569101e-01 -1.18984568e+00 5.24880171e-01
1.07637501e+00 -3.37410241e-01 8.30283761e-02 5.05688608e-01
5.10434628e-01 4.61749062e-02 -1.16982055e+00 8.52509856e-01
-2.23374173e-01 -1.25816762e+00 -1.94193989e-01 -7.63277635e-02
9.88926351e-01 2.72286355e-01 -3.24947715e-01 4.95364338e-01
7.99841225e-01 -7.50070512e-01 8.95353675e-01 1.86556458e-01
8.78262281e-01 -5.96062779e-01 4.54140991e-01 -5.86505271e-02
-1.40181470e+00 1.24031045e-01 -5.77042252e-02 -1.05120189e-01
1.79228023e-01 5.76078057e-01 -5.35708308e-01 9.49299812e-01
1.09017897e+00 1.11705625e+00 -5.45736432e-01 5.17593741e-01
-7.97402263e-01 1.31933510e+00 -1.22040279e-01 1.57532439e-01
4.07477282e-02 -6.17360808e-02 4.14031088e-01 1.59615791e+00
2.38381550e-01 -2.47949392e-01 2.99819291e-01 7.50832140e-01
-2.46968456e-02 2.59382725e-01 -7.38047481e-01 -2.55248040e-01
6.24657035e-01 9.40161824e-01 -6.87329352e-01 -1.65995851e-01
-9.84991729e-01 8.59117746e-01 5.52602708e-01 4.27189708e-01
-1.23351073e+00 -9.18633193e-02 4.41432536e-01 -3.78629863e-01
-2.59648979e-01 -3.52629036e-01 -3.27260494e-01 -1.54450464e+00
-2.55081773e-01 -8.81419897e-01 1.09575295e+00 -6.20090604e-01
-1.49745476e+00 1.13542452e-01 4.18117732e-01 -8.39453399e-01
-3.78320217e-01 -6.22465432e-01 -5.18534362e-01 6.54988348e-01
-1.59385431e+00 -1.36700308e+00 3.72610837e-02 6.20237231e-01
4.47955966e-01 -7.27175027e-02 8.11302662e-01 8.69612455e-01
-1.09174371e+00 4.43817347e-01 -2.57504165e-01 3.09614539e-01
1.40609860e+00 -1.30452943e+00 3.50677073e-01 1.03251851e+00
2.92068303e-01 5.67156434e-01 4.66416210e-01 -1.04153097e+00
-1.01804113e+00 -1.23707640e+00 1.80334163e+00 -6.50396764e-01
1.26485479e+00 -2.75191754e-01 -8.02588999e-01 1.25897729e+00
6.26347661e-01 -5.17053843e-01 8.07108402e-01 6.15366340e-01
-4.39721137e-01 8.25659633e-02 -4.89032030e-01 9.08756554e-01
1.05127192e+00 -4.40580517e-01 -5.69818974e-01 3.10021579e-01
7.44234264e-01 -2.09882528e-01 -1.22713649e+00 3.32805693e-01
1.75096214e-01 -4.18573320e-01 8.09434533e-01 -5.70480943e-01
8.44262064e-01 -3.34822804e-01 1.51037216e-01 -1.27916026e+00
-1.49493903e-01 -3.87102276e-01 7.67771378e-02 1.86414993e+00
7.76388586e-01 -7.16481507e-01 3.70987840e-02 4.72258389e-01
-8.07541907e-02 -1.68350428e-01 -8.11811030e-01 -7.80888975e-01
4.64726090e-01 -1.03648627e+00 3.85252386e-01 1.60530174e+00
3.07517439e-01 5.40890813e-01 -3.18730056e-01 3.32442820e-01
5.28127968e-01 2.22004503e-01 5.12983620e-01 -1.22212672e+00
2.99595743e-01 -3.48757118e-01 2.57924646e-01 -2.10477531e-01
5.82891285e-01 -1.23428726e+00 -2.11823761e-01 -1.50424278e+00
4.04725790e-01 -5.16138434e-01 -1.41114235e-01 9.54658628e-01
-5.96787035e-01 7.93938860e-02 -3.88383418e-01 2.29542792e-01
-4.58467394e-01 3.42467487e-01 9.95032191e-01 5.58948405e-02
-1.24823168e-01 -6.68300331e-01 -4.63257104e-01 9.06955898e-01
9.01610494e-01 -5.90184748e-01 -3.35968792e-01 -5.26409149e-01
7.29844630e-01 -1.19816866e-02 3.97517234e-01 -5.65539598e-01
1.38631418e-01 -4.58992392e-01 9.38707441e-02 -7.73205101e-01
-6.54312313e-01 -4.27753240e-01 2.31878176e-01 9.31641087e-02
-3.36107284e-01 1.49247736e-01 4.38624442e-01 4.42388773e-01
-5.93256950e-01 -1.23702347e-01 3.66827399e-01 -5.34249432e-02
-1.14834547e+00 1.57564387e-01 -5.77995181e-01 5.48135459e-01
9.28743541e-01 4.78463560e-01 -3.55053395e-01 -5.15281782e-02
-3.31204981e-01 4.19261426e-01 -3.37826386e-02 8.32176924e-01
1.67191610e-01 -1.59624732e+00 -1.22249269e+00 -3.51365924e-01
4.74696010e-01 -3.65554303e-01 -1.84007213e-02 7.44518399e-01
2.46971790e-02 8.64745140e-01 1.27652049e-01 -1.80567861e-01
-9.52896178e-01 6.01428866e-01 -4.55006622e-02 -8.03403437e-01
-5.96983731e-01 3.92859757e-01 3.77522022e-01 -5.12745380e-01
1.26842737e-01 -2.79100925e-01 -4.30146903e-01 2.90993094e-01
5.52845955e-01 3.64045382e-01 -2.28535999e-02 -5.59605718e-01
-6.32933080e-01 9.61724203e-03 1.59500930e-02 -2.56865233e-01
1.02747810e+00 -2.00172991e-01 -6.41020656e-01 8.68969083e-01
9.89354730e-01 2.44833440e-01 -5.09316325e-01 -1.66344032e-01
5.64987957e-01 5.78423671e-04 -7.71269351e-02 -1.10721028e+00
-6.45411849e-01 5.09412408e-01 1.22940451e-01 -9.11462121e-04
9.42616940e-01 2.61885703e-01 6.34900153e-01 -5.36508188e-02
5.66840053e-01 -1.10784531e+00 -2.17146754e-01 9.15659428e-01
1.28587949e+00 -1.27037966e+00 -4.06059057e-01 -5.79584062e-01
-5.94289124e-01 8.21688831e-01 6.18280590e-01 3.82745504e-01
7.17164099e-01 4.72818881e-01 1.95554551e-02 -4.23716217e-01
-7.90974736e-01 -3.05767775e-01 2.28583768e-01 3.69820803e-01
1.25535524e+00 2.69766033e-01 -9.46033180e-01 8.15605879e-01
-1.45254597e-01 1.48816511e-01 3.55889112e-01 4.33092535e-01
4.70476836e-01 -1.27098155e+00 -4.63140458e-01 -4.70523387e-02
-8.34613323e-01 -2.28260443e-01 -2.97739714e-01 1.26906490e+00
3.11769187e-01 1.21250653e+00 -1.57129586e-01 6.22013919e-02
2.53212750e-01 2.59456217e-01 4.11825746e-01 -5.01291871e-01
-3.98650348e-01 -1.79101497e-01 4.76473451e-01 -4.26799625e-01
-7.24995434e-01 -1.00626814e+00 -1.42923367e+00 4.40865904e-02
6.06788360e-02 -9.30617657e-03 4.10668403e-01 9.99505162e-01
8.93797874e-02 6.88823521e-01 2.08754078e-01 -1.79272756e-01
3.79850686e-01 -8.09361696e-01 -1.91431805e-01 4.73312467e-01
-3.10440749e-01 -9.30390954e-01 -1.89879075e-01 3.68566662e-01] | [9.132208824157715, 9.210675239562988] |
e9825a58-1a99-4a2a-aa1a-3b2107a344ff | exploiting-intrinsic-stochasticity-of-real | 2304.06056 | null | https://arxiv.org/abs/2304.06056v1 | https://arxiv.org/pdf/2304.06056v1.pdf | Exploiting Intrinsic Stochasticity of Real-Time Simulation to Facilitate Robust Reinforcement Learning for Robot Manipulation | Simulation is essential to reinforcement learning (RL) before implementation in the real world, especially for safety-critical applications like robot manipulation. Conventionally, RL agents are sensitive to the discrepancies between the simulation and the real world, known as the sim-to-real gap. The application of domain randomization, a technique used to fill this gap, is limited to the imposition of heuristic-randomized models. We investigate the properties of intrinsic stochasticity of real-time simulation (RT-IS) of off-the-shelf simulation software and its potential to improve the robustness of RL methods and the performance of domain randomization. Firstly, we conduct analytical studies to measure the correlation of RT-IS with the occupation of the computer hardware and validate its comparability with the natural stochasticity of a physical robot. Then, we apply the RT-IS feature in the training of an RL agent. The simulation and physical experiment results verify the feasibility and applicability of RT-IS to robust RL agent design for robot manipulation tasks. The RT-IS-powered robust RL agent outperforms conventional RL agents on robots with modeling uncertainties. It requires fewer heuristic randomization and achieves better generalizability than the conventional domain-randomization-powered agents. Our findings provide a new perspective on the sim-to-real problem in practical applications like robot manipulation tasks. | ['Homayoun Najjaran', 'Dean Richert', 'Zengjie Zhang', 'Amir M. Soufi Enayati', 'Ram Dershan'] | 2023-04-12 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [-3.52103449e-02 1.41361773e-01 -1.41778171e-01 1.53857470e-01
-3.92652780e-01 -4.51023281e-01 5.82080007e-01 -1.69711798e-01
-6.14434600e-01 9.70579803e-01 -4.41465497e-01 -5.08615732e-01
-6.60494626e-01 -7.10973263e-01 -9.00097787e-01 -8.14311087e-01
-6.32068872e-01 7.03275025e-01 2.68283635e-01 -6.82111263e-01
5.28756678e-02 6.98244274e-01 -1.69800150e+00 -3.73489439e-01
6.88928127e-01 6.34333253e-01 7.13808239e-01 4.79152769e-01
4.37721342e-01 7.83489823e-01 -8.29224467e-01 7.69788027e-01
4.50956911e-01 -3.09766352e-01 -2.46753007e-01 -1.29424810e-01
-7.30603755e-01 -3.11516434e-01 -3.41749460e-01 9.59997892e-01
6.18233383e-01 3.93733978e-01 4.85902905e-01 -1.83163345e+00
2.68576201e-02 7.27745235e-01 -4.98136848e-01 -1.41468555e-01
4.02053922e-01 6.39933646e-01 -3.14398482e-02 -1.48503616e-01
5.61407804e-01 1.57322025e+00 4.72869039e-01 4.11203563e-01
-8.75446737e-01 -7.75199175e-01 8.89853537e-02 -6.44308031e-02
-1.45017278e+00 -6.35683239e-02 5.19433796e-01 -4.47220415e-01
9.46994364e-01 -2.47588068e-01 3.66887778e-01 1.15380740e+00
8.65185797e-01 3.62616032e-01 1.16504622e+00 -4.01567668e-01
7.99256682e-01 -7.39610493e-02 -5.74267745e-01 4.40906644e-01
4.70003426e-01 8.96173775e-01 -1.72962993e-01 -9.35824364e-02
8.95889580e-01 -4.56052214e-01 -1.35489509e-01 -6.11532986e-01
-1.35759640e+00 6.95570171e-01 1.99476644e-01 9.43077877e-02
-4.58942831e-01 4.99290019e-01 6.81653321e-01 5.55432737e-01
-3.28627437e-01 1.03446078e+00 -5.79485476e-01 -3.47510844e-01
-2.02526718e-01 2.64579773e-01 9.19015110e-01 1.24502659e+00
3.04020762e-01 6.19712591e-01 2.34040350e-01 2.53821194e-01
2.89139271e-01 7.05364466e-01 7.73525894e-01 -1.23381197e+00
2.21856624e-01 3.12658876e-01 5.52274406e-01 -6.28343463e-01
-9.04411972e-01 -1.88855290e-01 -5.53523898e-01 7.45025933e-01
3.29455137e-01 -4.99301106e-01 -5.33898592e-01 1.64205897e+00
4.33257520e-01 1.49965677e-02 4.86452520e-01 9.28997219e-01
-3.67082283e-02 5.59192479e-01 -2.96505481e-01 -4.60762948e-01
8.86900783e-01 -5.16753018e-01 -6.63835526e-01 -2.25233316e-01
6.11380339e-01 -4.74846661e-01 1.13159502e+00 5.22879422e-01
-7.47422934e-01 -5.14351368e-01 -1.41650200e+00 1.00166845e+00
-1.15172066e-01 -1.29338607e-01 4.26215053e-01 6.11831903e-01
-6.75024092e-01 7.61214316e-01 -9.32040274e-01 -3.04176718e-01
-2.77989864e-01 5.83495259e-01 -1.90494135e-01 2.44272038e-01
-1.24489605e+00 1.54088473e+00 5.18725812e-01 8.14262927e-02
-1.13930917e+00 -5.47802925e-01 -9.22858417e-01 -4.66010094e-01
6.82835877e-01 -2.72276253e-01 1.57578814e+00 -6.44239068e-01
-2.29353714e+00 -8.91792472e-04 4.01141793e-01 -4.88624603e-01
7.60041893e-01 -4.15677950e-02 -2.41792515e-01 4.85159233e-02
9.43058431e-02 5.29749431e-02 1.00735259e+00 -1.52100182e+00
-5.41635752e-01 1.83617800e-01 8.03088471e-02 5.49667850e-02
4.91034210e-01 -2.17649072e-01 4.14917648e-01 -3.56481552e-01
-1.11379676e-01 -1.30374551e+00 -6.64741516e-01 -4.81188923e-01
1.73955619e-01 -1.79493397e-01 8.07555437e-01 1.39216185e-02
5.05620062e-01 -2.06981468e+00 4.65150736e-03 3.75532389e-01
-2.77734548e-01 1.13230303e-01 -2.81922132e-01 8.57587457e-01
-1.08325653e-01 -2.41705701e-01 2.03421727e-01 4.37591255e-01
1.65582106e-01 4.88994420e-01 -4.05697167e-01 9.49604988e-01
1.44050375e-01 3.56694728e-01 -1.22501576e+00 -2.52469480e-01
4.65240836e-01 -9.34042335e-02 -2.77113020e-01 4.37505931e-01
-3.48115623e-01 6.18530571e-01 -7.35389590e-01 2.05768481e-01
4.72273916e-01 3.72735441e-01 3.17726105e-01 8.85788277e-02
-4.59003985e-01 1.50306463e-01 -1.28416741e+00 1.06590116e+00
-8.23080540e-01 4.23668444e-01 3.11826348e-01 -8.88395727e-01
1.07767940e+00 2.39257440e-01 5.67380846e-01 -9.02477086e-01
5.05436063e-01 4.25590515e-01 4.47178543e-01 -6.82659686e-01
5.20297587e-01 -1.21222474e-01 -4.00692672e-01 3.85231137e-01
-2.66623199e-01 -9.64785039e-01 2.24632487e-01 -2.24294275e-01
1.07950032e+00 4.34882164e-01 4.63360548e-01 -5.07542849e-01
2.92207599e-01 2.44321674e-01 6.11664057e-01 8.19647908e-01
-3.54149818e-01 2.28677802e-02 6.04240000e-01 5.50475866e-02
-1.11546063e+00 -8.27297509e-01 2.54078269e-01 7.19157875e-01
5.58305621e-01 7.46582970e-02 -4.43838626e-01 -3.57597649e-01
3.18915486e-01 1.16292441e+00 -4.74810004e-01 -6.65265262e-01
-7.41010129e-01 -4.49851960e-01 4.60115582e-01 4.19380605e-01
1.41088203e-01 -9.05126214e-01 -1.34934258e+00 4.81365442e-01
2.08239123e-01 -1.17948043e+00 -2.45163869e-02 6.30462885e-01
-5.71527302e-01 -1.04196346e+00 -5.07001132e-02 -4.49824482e-01
5.70252180e-01 2.72012293e-01 5.69992959e-01 4.66335788e-02
-1.96246013e-01 5.06602287e-01 -6.00401223e-01 -5.80233753e-01
-1.08910000e+00 -4.19530064e-01 7.10898817e-01 -7.04070687e-01
-4.10246700e-01 -3.36379945e-01 -2.66569346e-01 9.03104484e-01
-7.03138351e-01 -3.51125568e-01 6.48794293e-01 8.92436922e-01
9.99868438e-02 7.31357336e-01 7.65005291e-01 -1.17046516e-02
9.14743066e-01 -2.89961308e-01 -1.09878898e+00 -5.52442158e-03
-4.81544912e-01 4.16230470e-01 7.58878648e-01 -1.01068056e+00
-7.04402566e-01 2.07054898e-01 3.12339932e-01 -3.13426852e-01
1.46333441e-01 5.88345110e-01 -7.16995150e-02 -2.08144069e-01
6.98591709e-01 -9.91619378e-02 6.43502176e-01 2.36659244e-01
7.65197873e-02 5.25150657e-01 2.05747500e-01 -9.57292080e-01
9.78752792e-01 2.09117636e-01 4.19951588e-01 -9.87775564e-01
-1.64126325e-02 -4.93877046e-02 -1.93733454e-01 -4.78849858e-01
2.99237549e-01 -6.49911642e-01 -1.09527826e+00 3.47918719e-01
-9.30462122e-01 -1.07206488e+00 -5.06675482e-01 8.41039479e-01
-1.18281090e+00 2.42635310e-01 -2.05502674e-01 -1.25716174e+00
2.36236513e-01 -1.55835605e+00 7.44548678e-01 2.20157817e-01
-1.78218022e-01 -5.57873428e-01 -7.52417818e-02 -2.85564721e-01
3.47151339e-01 3.53585541e-01 7.58211374e-01 -3.83530378e-01
-2.79630929e-01 -2.76547581e-01 1.92785814e-01 8.14302117e-02
-9.71621349e-02 1.38226435e-01 -4.46342438e-01 -6.46797001e-01
3.01993489e-01 -4.33209360e-01 2.33773496e-02 3.84409845e-01
4.31265742e-01 -5.04371300e-02 -2.52260685e-01 -1.31692246e-01
1.27656400e+00 7.03344047e-01 6.28533304e-01 8.67637575e-01
-5.00556268e-02 7.10877538e-01 1.61827552e+00 8.65378022e-01
-1.43946626e-03 6.08977199e-01 7.89144814e-01 1.13152325e-01
2.83717006e-01 -2.87052453e-01 8.28365743e-01 5.56288123e-01
1.57899857e-01 -9.90955010e-02 -8.78203630e-01 4.09433782e-01
-2.03098249e+00 -6.93126261e-01 9.58215818e-02 2.44529915e+00
5.70592999e-01 2.69077152e-01 8.85355473e-02 2.15443537e-01
8.38218153e-01 -4.03236479e-01 -4.95763749e-01 -5.71543694e-01
2.93281525e-01 -2.55485237e-01 9.59249377e-01 4.21993285e-01
-6.29141212e-01 6.71732247e-01 5.90806055e+00 9.52991903e-01
-1.07269895e+00 -1.39524773e-01 -2.02298656e-01 2.31438085e-01
2.20454320e-01 3.15787904e-02 -4.45034534e-01 3.71445000e-01
1.06301749e+00 -5.29937267e-01 5.62397182e-01 1.08452046e+00
9.15592432e-01 -7.20517993e-01 -1.27927756e+00 5.26937008e-01
-3.79554242e-01 -7.68338203e-01 -4.63761270e-01 -5.52817434e-03
4.88245517e-01 -8.07967931e-02 -1.14611350e-02 6.23345494e-01
7.33702481e-01 -9.80584800e-01 1.14672494e+00 1.21966660e-01
4.69893575e-01 -1.05632222e+00 1.07324517e+00 6.99026227e-01
-1.09505582e+00 -4.36831474e-01 -2.58265167e-01 -3.38571221e-01
2.72708088e-01 2.38540217e-01 -1.40172207e+00 5.99392176e-01
4.25151348e-01 -2.05731522e-02 -1.43697346e-02 9.43836391e-01
-1.84076577e-01 6.39097020e-02 -5.42637229e-01 -5.84187627e-01
3.16433638e-01 -1.29145801e-01 6.88089013e-01 7.66833723e-01
2.60827422e-01 -2.41578773e-01 5.62735558e-01 5.99612057e-01
7.30101883e-01 -3.96755010e-01 -7.77761579e-01 -1.01919092e-01
7.37641215e-01 6.89042985e-01 -8.00059617e-01 9.41525325e-02
1.28921255e-01 3.21266621e-01 -6.77932799e-02 1.93155482e-01
-1.17859221e+00 -4.41429913e-01 5.13407767e-01 -7.33792186e-02
1.28177285e-01 -8.23983312e-01 -7.66554102e-02 -2.68128902e-01
-1.86164811e-01 -1.24559593e+00 -2.52160043e-01 -7.94015884e-01
-1.10813379e+00 4.43612278e-01 3.87757331e-01 -1.66290379e+00
-6.70629025e-01 -6.65487766e-01 -2.91680336e-01 4.77522910e-01
-1.32102859e+00 -6.94142759e-01 -7.45149851e-02 3.58034939e-01
5.96402943e-01 -4.13167983e-01 5.15939295e-01 -2.39007056e-01
-4.41704363e-01 1.57594472e-01 3.45084548e-01 -3.86817247e-01
7.60700524e-01 -7.92023957e-01 6.81807250e-02 4.46973354e-01
-9.75214839e-01 6.21959448e-01 1.40882850e+00 -9.24802125e-01
-2.00639701e+00 -9.61798012e-01 -1.30158424e-01 -1.23762339e-01
1.13491857e+00 -1.15361117e-01 -3.93616885e-01 3.47434521e-01
-8.66042301e-02 -3.54756624e-01 -1.23215735e-01 -4.15478736e-01
1.51873395e-01 7.56385131e-03 -1.30961442e+00 9.21674967e-01
4.94157463e-01 -8.27920139e-02 -6.87340677e-01 2.23490521e-01
9.76853490e-01 -4.98881817e-01 -8.61438990e-01 7.26773858e-01
5.80280423e-01 -5.29924631e-01 8.13099921e-01 -1.49775431e-01
1.68102346e-02 -5.93591511e-01 1.54489803e-03 -1.66916931e+00
-1.35521758e-02 -9.41921353e-01 3.32081586e-01 7.78660834e-01
2.14047939e-01 -9.06055152e-01 4.16993290e-01 2.87119180e-01
-1.59230515e-01 -4.59737837e-01 -1.21750963e+00 -1.62133121e+00
6.10223636e-02 -4.07974422e-01 4.45813328e-01 7.07551539e-01
3.43989879e-01 -2.05516860e-01 -6.94118962e-02 5.20982206e-01
5.31356752e-01 -2.06360474e-01 1.02769029e+00 -7.88357317e-01
-4.10414010e-01 -2.52332598e-01 -3.60446632e-01 -5.17289221e-01
4.73748893e-01 -1.93485096e-01 8.14890206e-01 -1.26780105e+00
-3.93125951e-01 -7.46632755e-01 1.96826294e-01 6.27966523e-02
2.91508645e-01 -6.28279567e-01 3.17617893e-01 6.42785951e-02
-4.76409376e-01 9.14436340e-01 1.21203613e+00 4.82459106e-02
-5.62003791e-01 1.83795199e-01 1.01589970e-01 8.00639689e-01
1.08722317e+00 -6.57793105e-01 -6.55683160e-01 7.20252991e-02
2.44896725e-01 4.90569383e-01 2.06715077e-01 -1.27262437e+00
9.60670486e-02 -7.21536160e-01 -3.65468293e-01 -3.85072380e-01
1.04363210e-01 -1.24241471e+00 7.79499561e-02 1.07024288e+00
-1.94875941e-01 2.42703319e-01 5.23571432e-01 7.09230244e-01
2.13542119e-01 -5.60144842e-01 8.51278126e-01 -1.14186898e-01
-6.62490666e-01 -4.57531989e-01 -1.10175133e+00 -3.19593400e-03
1.45020163e+00 -5.25012277e-02 -4.01348501e-01 -3.86612922e-01
-4.94678132e-02 4.59990680e-01 4.77918297e-01 5.19743919e-01
5.79694569e-01 -7.40271688e-01 -3.51329893e-01 5.74779930e-03
-1.39566422e-01 -8.66717547e-02 -1.23823144e-01 8.12511265e-01
-5.81103623e-01 2.59572208e-01 -4.88084733e-01 -5.55156589e-01
-9.35379922e-01 9.03989732e-01 4.20320809e-01 -1.69211447e-01
-4.95695382e-01 2.06935197e-01 -8.66048262e-02 -5.57358265e-01
2.63005942e-01 -4.74032730e-01 2.39897177e-01 -3.71543854e-01
1.10871546e-01 5.90964019e-01 2.93325470e-03 -7.87963346e-02
-4.07770783e-01 2.76080936e-01 4.14720476e-01 -5.39987981e-01
1.08958256e+00 -9.97391045e-02 -2.58042458e-02 5.55718958e-01
4.92792457e-01 -2.37476692e-01 -1.54254174e+00 4.16222274e-01
1.74398065e-01 -2.16399953e-01 -3.46055478e-02 -5.41561663e-01
-4.75973815e-01 2.80718088e-01 6.25517428e-01 1.18165895e-01
5.51863670e-01 -3.51491332e-01 1.85483932e-01 6.59355462e-01
1.14272892e+00 -1.60753119e+00 4.16918576e-01 8.32932711e-01
1.19832289e+00 -9.50770259e-01 2.85672218e-01 -3.35723519e-01
-8.98078740e-01 1.10886276e+00 7.62446284e-01 -5.18165827e-01
4.06981885e-01 9.19445455e-01 -1.23000897e-01 2.62381375e-01
-6.51666939e-01 -2.67486513e-01 -5.54627240e-01 1.06146169e+00
-1.69008255e-01 1.33326247e-01 -2.20605448e-01 4.24846590e-01
-2.00632393e-01 1.29554108e-01 9.35146451e-01 1.40670574e+00
-5.65074623e-01 -9.18743074e-01 -6.93121374e-01 -2.40500391e-01
2.90360719e-01 6.49041235e-01 9.73072723e-02 1.43357193e+00
-8.78251642e-02 1.33521736e+00 -1.52758896e-01 -4.98853385e-01
6.23750508e-01 -3.54466021e-01 6.63767099e-01 -3.70210469e-01
-5.54276586e-01 -5.38638718e-02 3.98914069e-01 -6.36082530e-01
-1.65301800e-01 -6.23889923e-01 -1.79062009e+00 -2.45658562e-01
-4.70494896e-01 2.70925820e-01 1.05246770e+00 9.41553771e-01
1.89140841e-01 8.53169918e-01 9.17220473e-01 -1.39389765e+00
-1.28637111e+00 -6.07891560e-01 -5.57828903e-01 -1.71833083e-01
4.13659453e-01 -1.24066901e+00 -5.33927202e-01 -4.22588617e-01] | [4.758749008178711, 1.720940351486206] |
3c8846aa-b29d-4161-8981-309e502e51b0 | multimodal-fusion-via-teacher-student-network | null | null | https://ojs.aaai.org/index.php/AAAI/article/view/16430 | https://ojs.aaai.org/index.php/AAAI/article/view/16430/16237 | Multimodal Fusion via Teacher-Student Network for Indoor Action Recognition | Indoor action recognition plays an important role in modern
society, such as intelligent healthcare in large mobile cabin
hospitals. With the wide usage of depth sensors like Kinect,
multimodal information including skeleton and RGB modalities
brings a promising way to improve the performance.
However, existing methods are either focusing on a single
data modality or failed to take the advantage of multiple data
modalities. In this paper, we propose a Teacher-Student Multimodal
Fusion (TSMF) model that fuses the skeleton and
RGB modalities at the model level for indoor action recognition.
In our TSMF, we utilize a teacher network to transfer
the structural knowledge of the skeleton modality to a
student network for the RGB modality. With extensive experiments
on two benchmarking datasets: NTU RGB+D and
PKU-MMD, results show that the proposed TSMF consistently
performs better than state-of-the-art single modal and
multimodal methods. It also indicates that our TSMF could
not only improve the accuracy of the student network but also
significantly improve the ensemble accuracy. | ['Keith C.C. Chan', 'Yan Liu', 'Bruce X.B. Yu'] | 2021-05-18 | null | null | null | association-for-the-advancement-of-artificial-3 | ['action-recognition-in-videos-2'] | ['computer-vision'] | [ 2.53410161e-01 -1.67087048e-01 -1.26914129e-01 -4.16027576e-01
-8.32502246e-01 -5.72962500e-02 5.09438396e-01 -2.27502808e-01
-3.71541023e-01 7.20835745e-01 2.74425119e-01 -2.73736089e-01
-1.57337874e-01 -7.64629602e-01 -6.03380144e-01 -1.01900578e+00
4.66666073e-01 1.10899679e-01 3.61520082e-01 -1.06171407e-01
-1.05704367e-01 2.69596577e-01 -1.70015156e+00 4.06980306e-01
9.23576534e-01 1.28900313e+00 2.06022169e-02 4.88139242e-01
-1.61982596e-01 9.21488464e-01 -3.71791065e-01 -2.50883669e-01
1.14270687e-01 -4.32691544e-01 -7.59684741e-01 1.39391422e-01
2.67232835e-01 -4.61677819e-01 -4.45595354e-01 6.99325323e-01
8.18566680e-01 3.78394872e-01 2.77451873e-01 -1.22197068e+00
-6.88759759e-02 2.95171261e-01 -5.53853750e-01 -2.53041476e-01
7.02265501e-01 -1.28113374e-01 4.88537818e-01 -5.84803343e-01
3.63654017e-01 1.00180471e+00 5.00971496e-01 6.59557521e-01
-6.54388785e-01 -5.02149820e-01 2.74910986e-01 5.12417138e-01
-1.18115318e+00 -2.39635631e-01 9.27218080e-01 -6.90130293e-02
7.70539343e-01 3.65181863e-01 8.22999954e-01 1.17779291e+00
-1.03676155e-01 1.29439306e+00 1.29358983e+00 -3.75162512e-01
1.45640716e-01 -1.48847297e-01 -1.64802372e-01 8.61595929e-01
-1.65111363e-01 -2.58248270e-01 -8.72072756e-01 1.52659714e-01
8.74763608e-01 4.93562311e-01 -3.68070275e-01 -4.85844016e-01
-1.43011737e+00 3.86712581e-01 6.04588866e-01 3.78534704e-01
-4.15700853e-01 2.12328434e-01 1.44916937e-01 -1.21575594e-01
3.43961239e-01 -7.41247237e-02 -5.22904158e-01 -6.44904375e-01
-6.99966371e-01 -2.83895195e-01 5.17813504e-01 4.54350621e-01
6.66341662e-01 -3.76304895e-01 1.63570717e-01 8.30579519e-01
6.81567073e-01 7.05877244e-01 6.32181466e-01 -1.03397417e+00
6.24331832e-01 9.79626000e-01 -1.76076040e-01 -7.38919735e-01
-3.95174503e-01 4.21105437e-02 -7.49742329e-01 6.71398491e-02
3.12941283e-01 4.64539416e-02 -1.16112792e+00 1.35405731e+00
6.94183290e-01 5.41359484e-01 1.85199276e-01 9.41168487e-01
1.31985462e+00 2.67615139e-01 -9.93998721e-03 6.82488754e-02
9.82088923e-01 -1.15174568e+00 -5.88919461e-01 -1.80172827e-02
7.23673701e-01 -5.86906314e-01 8.78046870e-01 5.11497855e-01
-7.31853306e-01 -6.06708944e-01 -8.67716730e-01 -2.96792411e-03
-3.82577330e-01 3.12837958e-01 7.22966552e-01 8.79219174e-01
-7.57591128e-01 4.51318026e-01 -1.35276270e+00 -6.45130575e-01
4.27621543e-01 5.18457055e-01 -8.26294601e-01 -4.89030182e-01
-8.70247304e-01 6.93066657e-01 3.17729324e-01 2.14935124e-01
-7.87611961e-01 -3.31550777e-01 -8.55438530e-01 -3.10742140e-01
3.04780215e-01 -7.22743571e-01 1.10497642e+00 -7.12786436e-01
-1.87468624e+00 4.69945401e-01 -2.04391092e-01 -4.99486513e-02
4.87847984e-01 -3.71467918e-01 -1.79196134e-01 3.50495428e-01
-3.75130326e-01 7.12526500e-01 5.19389629e-01 -1.43317688e+00
-7.80942678e-01 -7.48707414e-01 3.24613094e-01 4.66062397e-01
-5.87284267e-01 -4.94069457e-01 -4.92638379e-01 -1.32000685e-01
6.40778542e-01 -8.98488462e-01 -1.16660915e-01 -2.25110680e-01
-3.84121567e-01 -2.27892280e-01 1.04762566e+00 -6.05556607e-01
1.03535783e+00 -2.05445075e+00 4.34557766e-01 1.43237710e-01
5.10981679e-02 1.97841495e-01 1.25600815e-01 2.51722306e-01
9.08592939e-02 -2.49378860e-01 -2.03436241e-01 -7.24233925e-01
-1.05286269e-02 7.29533255e-01 2.37874120e-01 4.46586579e-01
-4.01351750e-01 7.68661141e-01 -8.96455467e-01 -7.73861170e-01
6.93425000e-01 7.93029785e-01 -2.45940417e-01 3.16940159e-01
1.59146160e-01 9.04756844e-01 -5.80521762e-01 1.16059673e+00
4.17694837e-01 -2.24742949e-01 2.38825276e-01 -2.84554601e-01
2.28778332e-01 6.27310053e-02 -1.26487243e+00 2.39310026e+00
-4.56405044e-01 2.36546606e-01 -1.06049925e-01 -9.05550599e-01
6.48616612e-01 4.72127616e-01 7.93253243e-01 -8.14491093e-01
2.28390396e-01 3.64540815e-01 -2.43771330e-01 -8.08505893e-01
2.79724568e-01 -8.62951577e-02 1.45825475e-01 2.81790465e-01
1.20236874e-01 6.07058182e-02 -1.75845414e-01 -5.01317978e-02
1.18254220e+00 6.63152337e-01 8.01144466e-02 5.13296306e-01
6.93630457e-01 -4.09343600e-01 5.06380796e-01 3.80654305e-01
-4.30817306e-01 7.71184921e-01 -5.58501109e-02 -3.21210533e-01
-2.79821724e-01 -1.00996315e+00 7.42295459e-02 1.04439294e+00
3.85885775e-01 -3.39288563e-01 -6.62704408e-01 -1.07176399e+00
-1.77149668e-01 2.80975074e-01 -5.73977649e-01 -1.07518472e-01
-3.55264515e-01 -7.46044755e-01 5.70806861e-01 9.43014443e-01
1.06831980e+00 -8.35287571e-01 -7.90530622e-01 -1.88644230e-01
-6.17173731e-01 -1.21679378e+00 2.00228199e-01 2.19136342e-01
-1.22480857e+00 -1.17448545e+00 -8.51845264e-01 -3.57474804e-01
7.89810240e-01 3.10065925e-01 5.88286996e-01 4.97302823e-02
9.77017209e-02 9.54527557e-01 -6.23110533e-01 -1.78698644e-01
5.21878600e-02 2.03132600e-01 5.30153289e-02 2.53308296e-01
2.12524667e-01 -7.32689977e-01 -6.39656007e-01 3.20068687e-01
-9.43396449e-01 1.51818946e-01 6.51511014e-01 5.55620372e-01
5.30028105e-01 -7.37151653e-02 9.27406624e-02 -4.71414059e-01
3.76412384e-02 -3.03098917e-01 -4.65616380e-04 4.59460586e-01
-3.77449363e-01 -2.04448909e-01 1.08412251e-01 -3.39809835e-01
-1.26706779e+00 3.74158114e-01 -4.23367769e-01 -4.51490611e-01
-5.80281258e-01 5.15128374e-01 -5.37307262e-01 -3.15970749e-01
2.69002140e-01 2.27678403e-01 -6.60267770e-02 -7.23276794e-01
1.13037571e-01 7.81529069e-01 3.90296191e-01 -7.94602156e-01
5.44043064e-01 6.53560042e-01 1.21754438e-01 -7.21573532e-01
-5.90176642e-01 -5.66546023e-01 -8.73732865e-01 -5.89407504e-01
1.01104379e+00 -9.18020129e-01 -9.34919536e-01 8.19735289e-01
-8.26086998e-01 -3.39026093e-01 -1.23792447e-01 7.90075898e-01
-4.44353431e-01 4.23660219e-01 -4.41601634e-01 -8.62411380e-01
1.51818246e-01 -1.38572025e+00 1.34844351e+00 6.19878232e-01
1.80404052e-01 -1.10845435e+00 7.28979781e-02 1.10163295e+00
2.68321425e-01 4.36736643e-01 5.93654037e-01 -4.80773807e-01
-5.82290292e-01 -3.30843210e-01 -4.01222743e-02 3.65759134e-01
4.32484984e-01 -2.07217067e-01 -1.20389616e+00 8.50055441e-02
-1.52012691e-01 -5.05766332e-01 9.59855318e-01 2.07190096e-01
1.16184711e+00 3.13962191e-01 -4.01603818e-01 6.10719204e-01
1.13461494e+00 3.10638845e-01 8.74953866e-01 5.68507731e-01
1.22562373e+00 4.36248779e-01 6.07429981e-01 4.21294421e-01
8.02184701e-01 4.94723648e-01 8.36625040e-01 -2.00813890e-01
1.72075573e-02 -1.26804411e-01 4.40273106e-01 9.54394400e-01
-7.56954074e-01 2.83466149e-02 -1.08955777e+00 3.46129358e-01
-2.18398356e+00 -6.49242342e-01 -5.05788773e-02 2.16362047e+00
6.39236152e-01 -1.75106376e-01 6.74668401e-02 5.52128255e-01
1.73902094e-01 1.24734871e-01 -3.02180290e-01 1.35692343e-01
1.12384036e-02 1.37510255e-01 2.99993247e-01 1.50295541e-01
-1.32002687e+00 5.07491231e-01 5.45327330e+00 7.23852694e-01
-1.09486735e+00 3.21922451e-01 4.11900938e-01 -1.38456538e-01
-2.93324254e-02 -3.50230604e-01 -4.83647972e-01 3.75012130e-01
8.40772033e-01 6.89011157e-01 2.57475644e-01 8.84459138e-01
-1.00090140e-02 -6.03251457e-01 -1.00921559e+00 1.14842200e+00
2.29677081e-01 -8.83766055e-01 -2.15711504e-01 2.50455618e-01
8.50976348e-01 -1.33884363e-02 -1.01955190e-01 2.47028500e-01
5.75323356e-03 -1.05574918e+00 2.94180393e-01 8.38450193e-01
2.98416048e-01 -7.14586675e-01 1.13205326e+00 4.67766374e-01
-1.21133113e+00 1.14367045e-02 1.38247758e-01 8.70547444e-02
6.37161285e-02 3.52336526e-01 -5.51024675e-01 1.09080493e+00
9.37567115e-01 9.40998077e-01 -7.45029628e-01 1.10736787e+00
-5.03711522e-01 4.28769410e-01 -4.31519061e-01 1.77294061e-01
5.36627881e-02 -9.33120176e-02 7.23624974e-02 6.76398456e-01
5.33261359e-01 1.34158552e-01 3.06456625e-01 7.91031793e-02
2.78448742e-02 -5.58863468e-02 -4.68543112e-01 -1.59409121e-02
-4.43754792e-02 1.20788860e+00 -7.47442484e-01 -3.18171918e-01
-5.66842496e-01 1.10776079e+00 -9.26251486e-02 3.69768530e-01
-9.55415845e-01 -3.03505510e-02 6.14097774e-01 -2.22782373e-01
2.45997131e-01 -3.48736167e-01 -2.96711922e-01 -1.26505291e+00
2.09855800e-03 -8.47521901e-01 2.94038028e-01 -8.29885185e-01
-8.08232725e-01 3.05981100e-01 -6.91583306e-02 -1.37012696e+00
5.01190871e-02 -7.46316552e-01 -3.50806624e-01 3.86807561e-01
-1.42554748e+00 -1.63276458e+00 -7.53358126e-01 9.89967763e-01
2.95633048e-01 3.30220126e-02 9.59156632e-01 5.79871118e-01
-8.06824088e-01 3.09827268e-01 -1.45601071e-02 1.21464573e-01
6.65451884e-01 -1.15350318e+00 -4.83345211e-01 6.51177108e-01
2.26676971e-01 5.02587795e-01 2.34417409e-01 -4.21669006e-01
-1.59207082e+00 -7.17016280e-01 4.62784857e-01 -4.96668339e-01
2.16359168e-01 1.87099695e-01 -5.09813547e-01 5.12973785e-01
9.80042592e-02 2.04734296e-01 9.61587846e-01 6.90452456e-02
-9.52970535e-02 -4.03205216e-01 -1.18709600e+00 1.54792741e-01
9.11763906e-01 -5.70682704e-01 -3.89405221e-01 1.75695971e-01
3.75868410e-01 -8.31550539e-01 -1.06117129e+00 8.01438510e-01
8.48267555e-01 -1.24949145e+00 1.05370831e+00 -3.27672303e-01
6.32566512e-01 -5.32773435e-01 -6.27267361e-01 -1.13304996e+00
3.64323080e-01 2.14639261e-01 -4.62012470e-01 1.17840517e+00
8.78027752e-02 -4.97756600e-01 9.31852162e-01 7.08185017e-01
-1.03626095e-01 -1.02524662e+00 -1.33628142e+00 -4.40044045e-01
-4.41001892e-01 -6.89322770e-01 6.23970270e-01 8.20413709e-01
-7.43214563e-02 2.06370349e-03 -3.45624596e-01 1.78513125e-01
4.64585543e-01 -5.21785654e-02 9.39668953e-01 -1.26566398e+00
-2.49532640e-01 -1.92304924e-01 -6.70472443e-01 -1.03611028e+00
-4.56610695e-02 -5.26388347e-01 -2.96366382e-02 -2.10410476e+00
2.12569818e-01 -1.95108578e-01 -7.29908586e-01 7.89529622e-01
-2.05674216e-01 5.19288301e-01 1.50096327e-01 -7.87744299e-04
-1.04546618e+00 7.90582418e-01 1.36617994e+00 -1.10557370e-01
-1.27019256e-01 1.30073488e-01 -3.57463062e-01 9.08873796e-01
5.38189054e-01 -2.13864595e-01 -4.58312124e-01 -5.44759333e-01
1.00708287e-02 4.63137329e-02 4.88897294e-01 -1.36330211e+00
3.51237595e-01 -1.81446731e-01 6.68852925e-01 -6.71693683e-01
7.66786575e-01 -1.21569347e+00 -8.99441242e-02 4.27715719e-01
2.08635956e-01 -3.15286338e-01 1.27539784e-02 5.15387475e-01
-3.65942895e-01 1.30354837e-01 3.98370355e-01 -1.62761211e-01
-8.82286012e-01 7.21686259e-02 -1.47165701e-01 -5.15028715e-01
1.09987259e+00 -5.07187724e-01 -3.35407734e-01 -3.48603606e-01
-7.40572453e-01 2.29144976e-01 3.85899156e-01 3.87604207e-01
8.53259742e-01 -1.49925959e+00 -3.12801218e-03 2.05002606e-01
7.04786628e-02 2.71560192e-01 4.44887638e-01 1.48752534e+00
-4.05394554e-01 3.10945451e-01 -2.05437601e-01 -8.77789140e-01
-1.47947836e+00 -1.53691739e-01 4.72686410e-01 -3.10171753e-01
-2.71404088e-01 1.00552166e+00 -1.91776291e-01 -9.15528357e-01
6.29420638e-01 -4.25172746e-01 -1.88415855e-01 1.13678776e-01
2.67308444e-01 6.39887869e-01 1.65267467e-01 -7.88590729e-01
-6.79931760e-01 7.74010777e-01 4.71766025e-01 -1.47136867e-01
1.36197901e+00 -1.23043470e-01 -1.16805688e-01 6.79754734e-01
1.01939285e+00 -3.11415225e-01 -1.07980371e+00 -1.57333076e-01
-3.08295459e-01 -5.51842868e-01 6.11112267e-02 -8.70126188e-01
-1.31565046e+00 1.06373346e+00 9.02066588e-01 -5.26927002e-02
1.51308477e+00 -7.27015585e-02 1.06465077e+00 4.95244533e-01
6.30962372e-01 -1.02543259e+00 3.69723201e-01 4.24999714e-01
3.62089276e-01 -1.59339046e+00 9.74172801e-02 -2.28126228e-01
-6.26229048e-01 1.19712818e+00 8.25145125e-01 4.17437851e-01
6.20706975e-01 -8.11479911e-02 3.43333483e-01 1.61975771e-02
-2.04464793e-01 -4.31616098e-01 4.38140810e-01 4.80356961e-01
4.82937187e-01 2.88267713e-02 1.01035908e-01 5.26612639e-01
1.36349753e-01 2.48885199e-01 5.06361909e-02 1.43840957e+00
-4.20984417e-01 -1.35940027e+00 -6.21927977e-01 1.43430173e-01
-2.49553964e-01 3.38222235e-01 -3.35205972e-01 7.77986407e-01
5.57296038e-01 1.08595610e+00 -3.04020554e-01 -8.87945771e-01
3.55546832e-01 3.60445350e-01 8.18674386e-01 -3.24838340e-01
-6.00467384e-01 1.43471986e-01 -7.49039501e-02 -9.36433673e-01
-1.20599556e+00 -5.66217959e-01 -1.41746545e+00 -2.18038514e-01
-1.46437600e-01 -1.23247713e-01 9.35997963e-01 1.39657307e+00
1.22959621e-01 8.25678825e-01 2.66358882e-01 -1.04184723e+00
-5.19598797e-02 -9.54090416e-01 -4.72286761e-01 2.22895771e-01
2.62224138e-01 -8.79007459e-01 -1.07534230e-01 3.01380572e-03] | [7.871856212615967, 0.4184073209762573] |
8ebad10d-5d18-4616-b840-5cb584ae2205 | discriminating-known-from-unknown-objects-via | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_Discriminating_Known_From_Unknown_Objects_via_Structure-Enhanced_Recurrent_Variational_AutoEncoder_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_Discriminating_Known_From_Unknown_Objects_via_Structure-Enhanced_Recurrent_Variational_AutoEncoder_CVPR_2023_paper.pdf | Discriminating Known From Unknown Objects via Structure-Enhanced Recurrent Variational AutoEncoder | Discriminating known from unknown objects is an important essential ability for human beings. To simulate this ability, a task of unsupervised out-of-distribution object detection (OOD-OD) is proposed to detect the objects that are never-seen-before during model training, which is beneficial for promoting the safe deployment of object detectors. Due to lacking unknown data for supervision, for this task, the main challenge lies in how to leverage the known in-distribution (ID) data to improve the detector's discrimination ability. In this paper, we first propose a method of Structure-Enhanced Recurrent Variational AutoEncoder (SR-VAE), which mainly consists of two dedicated recurrent VAE branches. Specifically, to boost the performance of object localization, we explore utilizing the classical Laplacian of Gaussian (LoG) operator to enhance the structure information in the extracted low-level features. Meanwhile, we design a VAE branch that recurrently generates the augmentation of the classification features to strengthen the discrimination ability of the object classifier. Finally, to alleviate the impact of lacking unknown data, another cycle-consistent conditional VAE branch is proposed to synthesize virtual OOD features that deviate from the distribution of ID features, which improves the capability of distinguishing OOD objects. In the experiments, our method is evaluated on OOD-OD, open-vocabulary detection, and incremental object detection. The significant performance gains over baselines show the superiorities of our method. The code will be released at https://github.com/AmingWu/SR-VAE. | ['Cheng Deng', 'Aming Wu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['object-localization'] | ['computer-vision'] | [-8.09450224e-02 -1.74047187e-01 -7.40944222e-02 -1.14122041e-01
-6.22219801e-01 -2.30045885e-01 4.76687670e-01 -3.74060631e-01
-1.04537308e-01 2.48165831e-01 -8.29032958e-02 -1.40803844e-01
2.51813401e-02 -5.48983455e-01 -6.35547042e-01 -8.77165556e-01
1.24381170e-01 1.09801263e-01 3.74756396e-01 5.06137311e-02
-8.62964466e-02 4.04477149e-01 -1.88301814e+00 1.32541403e-01
8.65244746e-01 1.10496473e+00 6.34656787e-01 3.14693123e-01
-7.49490112e-02 5.23991466e-01 -4.34843898e-01 6.88048229e-02
3.10173184e-01 -2.77268589e-01 -6.03122450e-02 2.08562404e-01
1.87187105e-01 -6.04447842e-01 -5.31219602e-01 1.17465615e+00
6.29709661e-01 9.03888866e-02 9.57052052e-01 -1.40074921e+00
-9.34818923e-01 1.47683352e-01 -6.28930748e-01 4.52810109e-01
-1.04812242e-01 3.42376888e-01 9.48790610e-01 -1.24240398e+00
3.04205000e-01 1.30720353e+00 3.69866610e-01 5.53351820e-01
-1.12008035e+00 -8.52074206e-01 1.88274950e-01 1.63965255e-01
-1.61446643e+00 -2.99696863e-01 7.91602612e-01 -6.29703760e-01
6.06259704e-01 1.01917945e-01 4.39549059e-01 1.28843331e+00
6.34949878e-02 1.16424966e+00 7.76935041e-01 -2.95697212e-01
3.77273001e-02 3.71986598e-01 3.08476150e-01 7.78007805e-01
5.54410219e-01 2.46026739e-01 -4.68172997e-01 2.33104099e-02
6.06745839e-01 4.00597781e-01 -2.37103596e-01 -5.19802690e-01
-1.00303733e+00 6.90869510e-01 5.50124764e-01 1.43504173e-01
-4.33589131e-01 -1.46663398e-01 1.42037779e-01 -3.74382101e-02
4.92343098e-01 2.00532824e-02 -3.14301997e-01 1.93050459e-01
-5.06448805e-01 2.39773765e-02 3.39478642e-01 9.79806960e-01
7.81093836e-01 1.06761292e-01 -6.23004913e-01 8.17914069e-01
6.88666105e-01 7.35990703e-01 5.55584490e-01 -5.63522995e-01
3.56069148e-01 8.58761549e-01 8.33033249e-02 -8.23913693e-01
-1.32216677e-01 -8.09666753e-01 -8.22045267e-01 7.09421784e-02
9.61447433e-02 8.17691907e-02 -1.20365989e+00 1.63235450e+00
6.57256424e-01 3.44161630e-01 4.84691076e-02 9.64761317e-01
9.12825644e-01 6.49219871e-01 -6.51271716e-02 -6.03179187e-02
1.54701746e+00 -7.84627199e-01 -6.79401338e-01 -2.19846040e-01
5.74138165e-01 -4.24114555e-01 1.18479264e+00 7.20577314e-02
-3.74563217e-01 -8.55902016e-01 -1.08982277e+00 -4.06180955e-02
-2.81873852e-01 5.00073314e-01 5.39848506e-01 5.43178856e-01
-4.47199345e-01 6.27753586e-02 -8.37580204e-01 3.59982699e-02
7.25768268e-01 1.58298492e-01 -6.23663142e-02 -1.25232846e-01
-1.15792501e+00 4.20609653e-01 4.82947618e-01 2.74790227e-01
-1.28757000e+00 -6.63834810e-01 -8.85355592e-01 4.91015501e-02
5.29757261e-01 -4.63422924e-01 1.04958034e+00 -6.36659980e-01
-1.06430471e+00 6.20158553e-01 -2.77916282e-01 -3.02457005e-01
5.49457788e-01 -1.58770993e-01 -3.36205781e-01 -5.88425994e-02
3.44439536e-01 7.15376616e-01 1.29751706e+00 -1.25336552e+00
-7.66970038e-01 -4.70520794e-01 -1.93604663e-01 1.84038609e-01
-6.02361321e-01 -2.89593458e-01 -6.96669042e-01 -6.51854813e-01
2.27058366e-01 -9.53573823e-01 1.55245304e-01 1.02235220e-01
-4.80650872e-01 -6.73269987e-01 9.01963651e-01 -6.08336806e-01
1.16653311e+00 -2.73682332e+00 -2.49118835e-01 7.51115084e-02
3.72579634e-01 2.19021350e-01 -1.43548176e-01 -1.29171401e-01
1.99021727e-01 -6.83653653e-02 -7.86644146e-02 -3.60573143e-01
1.74204811e-01 1.72617584e-01 -3.81449401e-01 3.88437748e-01
5.81461966e-01 9.05296981e-01 -7.65215337e-01 -5.47039807e-01
8.33060667e-02 5.96833110e-01 -3.46193552e-01 3.09145749e-01
-1.36780679e-01 4.17831123e-01 -8.65722835e-01 9.14869547e-01
7.63548195e-01 -4.67905134e-01 -2.14858815e-01 -2.48489425e-01
4.97379899e-03 1.00222766e-01 -1.37252927e+00 1.18272257e+00
-1.92538694e-01 4.41113859e-01 -1.12091929e-01 -7.89115965e-01
8.61906409e-01 1.09787598e-01 2.51520336e-01 -6.88363433e-01
2.94849843e-01 1.64633632e-01 1.14545345e-01 -4.50402617e-01
3.59668165e-01 3.73972028e-01 9.92310345e-02 8.76112506e-02
6.03029653e-02 3.35315824e-01 1.22876741e-01 2.82083839e-01
8.43460560e-01 4.16578762e-02 1.21493541e-01 -1.36139721e-01
4.91320431e-01 -1.84917703e-01 8.26191723e-01 8.90339732e-01
-4.62351561e-01 4.62921530e-01 2.51272202e-01 8.86344910e-03
-7.59041131e-01 -1.21350241e+00 -4.26870614e-01 9.91943538e-01
4.52508539e-01 -1.73068538e-01 -5.75484574e-01 -8.13938916e-01
1.34429902e-01 6.63906693e-01 -6.61479175e-01 -4.39846843e-01
-2.12961018e-01 -8.54843438e-01 3.26485217e-01 6.81613922e-01
6.98389113e-01 -9.52607751e-01 -5.85879147e-01 -2.27371939e-02
-1.72351807e-01 -8.67345035e-01 -5.56914389e-01 1.72523648e-01
-5.96248150e-01 -9.25503194e-01 -8.50353956e-01 -7.87034154e-01
6.44488752e-01 7.74739325e-01 5.64553738e-01 3.06076296e-02
-3.38641137e-01 3.36423069e-01 -3.33461791e-01 -7.07516432e-01
-1.48994386e-01 -7.50533352e-03 3.13154489e-01 3.72591794e-01
6.38808310e-01 -2.96159267e-01 -5.80762386e-01 4.19467896e-01
-7.82604456e-01 -1.73567042e-01 9.68026340e-01 1.00036156e+00
9.15136933e-01 2.52970546e-01 4.68722701e-01 -4.95351225e-01
3.98296803e-01 -6.07883275e-01 -5.93138754e-01 1.16454467e-01
-6.32471442e-01 1.54417127e-01 3.60103041e-01 -9.57134843e-01
-1.13100719e+00 9.93295982e-02 1.33641213e-01 -8.80370677e-01
-5.80701455e-02 2.57182777e-01 -4.25281703e-01 2.27450326e-01
4.59227920e-01 4.69139099e-01 -7.65148997e-02 -5.38570642e-01
1.73294753e-01 9.04568017e-01 3.99355114e-01 -3.59372109e-01
9.61181819e-01 5.14884889e-01 -2.85645306e-01 -1.00037909e+00
-9.00344551e-01 -7.27257192e-01 -3.63447487e-01 -8.97338241e-02
8.74957621e-01 -1.22830701e+00 -4.97139037e-01 6.15327358e-01
-9.38986421e-01 -2.05453902e-01 -9.42328423e-02 7.07663238e-01
-8.88425205e-03 3.38125497e-01 -3.00448239e-01 -1.06356442e+00
-2.59156555e-01 -1.10525179e+00 1.21989989e+00 5.19894302e-01
3.86129200e-01 -4.47700828e-01 -2.80500680e-01 2.91545838e-01
9.94471014e-02 -2.00469539e-01 5.42154491e-01 -8.05330276e-01
-8.40306878e-01 -1.82818532e-01 -4.28279430e-01 5.46176732e-01
1.74715489e-01 -2.84144908e-01 -1.13504577e+00 -4.20317054e-01
-1.17933294e-02 -2.57086575e-01 1.06661379e+00 3.46964151e-01
1.15057564e+00 7.87593126e-02 -6.19279325e-01 4.49714899e-01
8.59532237e-01 9.86009985e-02 4.55116123e-01 1.70275167e-01
7.84418285e-01 3.57255578e-01 8.17104936e-01 4.17475909e-01
3.66681010e-01 5.38390815e-01 3.18943650e-01 6.45381138e-02
-3.38639647e-01 -4.09796417e-01 6.17131233e-01 6.49161100e-01
1.74167201e-01 -2.05056354e-01 -6.30743206e-01 7.00425863e-01
-1.63314426e+00 -8.00225735e-01 -1.71071216e-01 2.09655833e+00
4.86379832e-01 2.83901364e-01 -3.09148226e-02 5.55790141e-02
8.64396036e-01 6.75226301e-02 -8.32184255e-01 3.24725151e-01
-2.25065853e-02 -1.66508421e-01 3.03922713e-01 -3.11740651e-03
-1.26314509e+00 7.66701877e-01 4.60886717e+00 1.11327970e+00
-1.01342857e+00 2.75515437e-01 2.65991539e-01 2.45891213e-01
-6.29490763e-02 -2.27132037e-01 -1.31232607e+00 7.03580856e-01
4.01295066e-01 1.13748930e-01 1.95781723e-01 1.08198500e+00
1.49728090e-01 -2.20270967e-03 -8.96964669e-01 9.27512228e-01
7.86463544e-02 -7.13836074e-01 -1.78243369e-02 1.62081808e-01
4.75283384e-01 1.25055509e-02 3.46736670e-01 7.53524125e-01
-2.07797866e-02 -4.85726029e-01 7.36222267e-01 4.22804087e-01
6.35905147e-01 -4.10997719e-01 6.34554386e-01 6.47833765e-01
-1.49413085e+00 -2.75431037e-01 -5.28948426e-01 2.84138918e-01
-1.30206063e-01 7.16455519e-01 -9.08103883e-01 4.65909839e-01
9.86065328e-01 6.26014054e-01 -7.87715137e-01 9.05489206e-01
-3.99625748e-01 7.66313195e-01 -4.26025689e-01 -8.42239931e-02
1.45229930e-02 -3.26521061e-02 7.60456979e-01 9.01167929e-01
3.12986881e-01 7.32481778e-02 3.21251988e-01 1.15104961e+00
-1.36650935e-01 -1.23598866e-01 -5.22232711e-01 -8.58846605e-02
7.03816056e-01 1.14379919e+00 -5.24699271e-01 -2.87472665e-01
-4.45302635e-01 8.86724472e-01 2.62862414e-01 3.51473063e-01
-1.04772520e+00 -4.37302798e-01 5.64954162e-01 -4.15483601e-02
7.23738134e-01 -1.48827702e-01 6.72097653e-02 -1.25680637e+00
3.72302771e-01 -7.99693704e-01 4.65725839e-01 -6.51559532e-01
-1.48680615e+00 4.13498789e-01 -3.87131870e-02 -1.39984655e+00
1.90615535e-01 -6.30354047e-01 -5.01780927e-01 7.48376667e-01
-1.57983088e+00 -1.19722378e+00 -5.43064654e-01 4.77563947e-01
6.11669183e-01 -1.12144858e-01 2.05454737e-01 4.66401160e-01
-8.85979891e-01 7.39966154e-01 1.33627117e-01 2.59125292e-01
5.71503341e-01 -8.84615958e-01 3.70780081e-02 1.00796795e+00
2.39280507e-01 5.88400483e-01 4.86270100e-01 -8.45650852e-01
-1.30401313e+00 -1.36254752e+00 3.74228507e-01 -3.95984471e-01
4.00117815e-01 -5.62389553e-01 -1.15170074e+00 6.13205194e-01
-4.86084849e-01 3.20818722e-01 3.28579187e-01 -1.14770293e-01
-3.85838151e-01 -1.76829755e-01 -8.19399714e-01 4.43872005e-01
1.10127366e+00 -6.11257493e-01 -8.20956469e-01 2.20738992e-01
8.17876756e-01 -3.02445382e-01 -4.90386665e-01 5.67136645e-01
4.00639564e-01 -6.75981164e-01 1.12285554e+00 -3.13471466e-01
4.90389355e-02 -7.86195934e-01 -2.49368370e-01 -9.18854356e-01
-3.71928513e-01 -2.13210672e-01 -5.60173631e-01 1.41132987e+00
2.01951683e-01 -7.26449847e-01 5.69624901e-01 1.82122126e-01
-1.98783889e-01 -6.35235608e-01 -7.20987737e-01 -1.06616974e+00
-3.19999993e-01 -3.93411815e-01 4.14646715e-01 7.19822645e-01
-7.13139713e-01 3.42568606e-01 -3.22933525e-01 7.51453161e-01
7.45192409e-01 1.81152418e-01 9.44491625e-01 -1.22671020e+00
-3.60543013e-01 -6.68651313e-02 -4.03371900e-01 -1.53260767e+00
6.01418726e-02 -9.74561155e-01 2.16542333e-01 -1.24376774e+00
4.10367161e-01 -4.04494345e-01 -5.65985620e-01 4.41296220e-01
-5.37011206e-01 1.34820700e-01 1.16656020e-01 4.89234984e-01
-7.30160356e-01 1.01762617e+00 1.18785870e+00 -3.48348528e-01
-3.71286452e-01 1.23635054e-01 -5.76963365e-01 5.39446950e-01
6.37235463e-01 -7.77942538e-01 -4.21324700e-01 -2.54945308e-01
-2.67814010e-01 -4.44114059e-01 7.50498950e-01 -9.40359116e-01
1.05805635e-01 8.50354955e-02 5.89118421e-01 -9.74924266e-01
3.28949481e-01 -5.24226487e-01 -1.59739763e-01 6.04936063e-01
9.25826374e-03 -4.52417374e-01 2.27206554e-02 1.01261234e+00
-1.45154461e-01 -1.38973475e-01 6.09628975e-01 2.04384089e-01
-1.02627277e+00 4.97871935e-01 -2.02763483e-01 1.87185500e-02
1.07892311e+00 -1.38834134e-01 -1.59234315e-01 1.83424947e-03
-3.88409704e-01 5.29209316e-01 1.88534439e-01 5.55337965e-01
6.96083128e-01 -1.26397264e+00 -6.55620337e-01 3.69329184e-01
4.99558210e-01 2.43241087e-01 4.04904932e-01 9.75728393e-01
1.33514069e-02 2.83864021e-01 1.57073349e-01 -8.83938253e-01
-1.09818232e+00 7.28065193e-01 3.06488603e-01 -5.29483333e-02
-6.86039925e-01 8.22139502e-01 7.20196068e-01 -3.92878205e-01
3.60795200e-01 -2.05659926e-01 -2.25159422e-01 5.28185517e-02
5.66107810e-01 3.18364441e-01 -1.24327861e-01 -4.55888689e-01
-4.20186669e-01 2.27260128e-01 -2.62803018e-01 2.61737585e-01
1.07812321e+00 -2.52521247e-01 1.74914092e-01 4.74739850e-01
1.02918959e+00 2.84197833e-02 -1.54567337e+00 -4.01472479e-01
-6.05305396e-02 -4.95788276e-01 2.65157670e-01 -3.93527716e-01
-8.99247289e-01 1.01996267e+00 1.09002173e+00 5.93631417e-02
9.59231317e-01 2.86788285e-01 6.89414203e-01 5.27089298e-01
1.87595636e-01 -8.78592253e-01 3.73626798e-01 3.47868949e-01
6.67611659e-01 -1.40329993e+00 -2.22300753e-01 -4.48363483e-01
-6.01924241e-01 7.51023889e-01 8.29046428e-01 3.28022726e-02
6.51744306e-01 -1.69619787e-02 -6.08546399e-02 -2.77812243e-01
-4.06626165e-01 -6.44485891e-01 5.51132441e-01 5.75686336e-01
-8.16153735e-02 -3.33219953e-02 -2.66727828e-03 9.47773635e-01
3.13681006e-01 -1.94528222e-01 1.27193287e-01 8.35992873e-01
-4.81131673e-01 -6.62476778e-01 -3.70243698e-01 6.08661115e-01
-1.12797879e-01 -1.15781538e-01 -1.58408493e-01 7.01129198e-01
3.46162856e-01 8.33163440e-01 3.71369831e-02 -3.58251303e-01
2.33962402e-01 3.93623076e-02 9.63204503e-02 -5.33240438e-01
1.26187131e-01 1.05596431e-01 -3.31056774e-01 -3.42481226e-01
-1.87904164e-01 -6.38729334e-01 -1.26029706e+00 2.48710588e-01
-8.27362478e-01 -4.12135012e-02 4.51179266e-01 8.17674577e-01
6.27051532e-01 7.39122987e-01 7.53307641e-01 -7.18881845e-01
-9.84153509e-01 -9.28037703e-01 -6.65186524e-01 3.28400731e-01
4.20051247e-01 -1.27341950e+00 -6.03500128e-01 -1.38799235e-01] | [9.351404190063477, 1.487152099609375] |
2ca69883-9763-45c3-9a61-7e9ddf7ff5f2 | do-llms-understand-user-preferences | 2305.06474 | null | https://arxiv.org/abs/2305.06474v1 | https://arxiv.org/pdf/2305.06474v1.pdf | Do LLMs Understand User Preferences? Evaluating LLMs On User Rating Prediction | Large Language Models (LLMs) have demonstrated exceptional capabilities in generalizing to new tasks in a zero-shot or few-shot manner. However, the extent to which LLMs can comprehend user preferences based on their previous behavior remains an emerging and still unclear research question. Traditionally, Collaborative Filtering (CF) has been the most effective method for these tasks, predominantly relying on the extensive volume of rating data. In contrast, LLMs typically demand considerably less data while maintaining an exhaustive world knowledge about each item, such as movies or products. In this paper, we conduct a thorough examination of both CF and LLMs within the classic task of user rating prediction, which involves predicting a user's rating for a candidate item based on their past ratings. We investigate various LLMs in different sizes, ranging from 250M to 540B parameters and evaluate their performance in zero-shot, few-shot, and fine-tuning scenarios. We conduct comprehensive analysis to compare between LLMs and strong CF methods, and find that zero-shot LLMs lag behind traditional recommender models that have the access to user interaction data, indicating the importance of user interaction data. However, through fine-tuning, LLMs achieve comparable or even better performance with only a small fraction of the training data, demonstrating their potential through data efficiency. | ['Derek Zhiyuan Cheng', 'Ed Chi', 'Lichan Hong', 'Maheswaran Sathiamoorthy', 'Nikhil Mehta', 'Jianmo Ni', 'Wang-Cheng Kang'] | 2023-05-10 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [-1.97808787e-01 -5.01571894e-01 -4.70081538e-01 -6.20746732e-01
-4.47508335e-01 -6.16209805e-01 6.33007884e-01 1.72457755e-01
-5.46676040e-01 4.13307041e-01 3.01989764e-01 -4.74095255e-01
-3.84447902e-01 -5.95810413e-01 -1.93514913e-01 -1.88483268e-01
-1.13184921e-01 4.76724654e-01 2.71387339e-01 -6.82416260e-01
5.04325151e-01 1.62410229e-01 -1.71206200e+00 5.53884029e-01
9.77772713e-01 9.56480563e-01 4.63339239e-01 5.51411748e-01
-1.80379674e-01 5.00951111e-01 -4.81012404e-01 -6.82711422e-01
1.77346110e-01 -1.02787256e-01 -6.01317585e-01 -3.01369071e-01
2.51067787e-01 -2.77894795e-01 -1.09018557e-01 6.30666256e-01
4.95768100e-01 8.16642761e-01 5.94102919e-01 -8.79666388e-01
-1.02031541e+00 6.10241771e-01 -2.64876783e-01 4.18973178e-01
6.09374762e-01 -2.73179859e-01 1.39783621e+00 -1.20117331e+00
3.30384612e-01 1.04315948e+00 5.63732743e-01 5.00650287e-01
-1.18011296e+00 -5.37137210e-01 5.84648132e-01 1.74220398e-01
-1.22618234e+00 -4.06326741e-01 2.95451820e-01 -4.50585335e-01
1.12515008e+00 3.11030567e-01 3.85088652e-01 9.32412446e-01
1.45505160e-01 5.78508556e-01 8.07460904e-01 -4.64525580e-01
3.77113640e-01 5.84375441e-01 4.96223807e-01 4.79201525e-02
9.98875424e-02 8.52026790e-03 -6.09327078e-01 -5.03009796e-01
4.86627519e-01 4.63762075e-01 -8.07645470e-02 -2.21452117e-01
-7.65360773e-01 1.06118977e+00 8.29871669e-02 3.25284153e-01
-1.89645052e-01 -5.38443565e-01 1.88839540e-01 7.66061366e-01
6.90292060e-01 7.31315732e-01 -8.67667496e-01 -2.67362118e-01
-8.97895753e-01 9.21462774e-02 1.04986107e+00 8.53055179e-01
4.66739416e-01 -2.01479301e-01 -2.05544844e-01 1.33542037e+00
2.57776588e-01 2.59935081e-01 8.92743349e-01 -7.13989496e-01
2.44773164e-01 4.27340716e-01 4.81730402e-01 -8.71043980e-01
-2.36608282e-01 -5.45089960e-01 -6.45710111e-01 -3.22540253e-02
2.57624239e-01 -1.53560683e-01 -6.06125593e-01 1.62573743e+00
1.21290080e-01 -5.51648065e-02 -1.23380013e-01 7.84789443e-01
6.05386972e-01 6.31429255e-01 1.30442068e-01 -4.80333805e-01
1.05631828e+00 -1.03974342e+00 -3.34034830e-01 -3.54897976e-01
6.62691832e-01 -7.77090549e-01 1.41756833e+00 8.15214097e-01
-9.15760159e-01 -8.49737704e-01 -7.30264783e-01 6.75093159e-02
-3.41967791e-01 9.46169794e-02 1.00337362e+00 9.08415020e-01
-9.85570729e-01 8.77368391e-01 -3.23058218e-01 -4.47092623e-01
-3.92093882e-02 4.86261666e-01 -1.14558622e-01 -2.47154176e-01
-1.42423320e+00 9.38291728e-01 -2.13728026e-01 -3.36990595e-01
-3.32895577e-01 -7.11615562e-01 -4.19484735e-01 2.96342164e-01
3.47628921e-01 -5.06672978e-01 1.47656417e+00 -8.08501363e-01
-1.63358045e+00 1.81880832e-01 -1.91290602e-01 -2.96978742e-01
8.61507729e-02 -3.98993671e-01 -8.65219772e-01 -4.52399790e-01
-2.73388267e-01 1.57745808e-01 5.62829494e-01 -8.48786891e-01
-8.75071228e-01 -1.90137982e-01 4.34271961e-01 2.82661706e-01
-7.47270286e-01 1.68480992e-01 -3.46885085e-01 -6.26393676e-01
-1.95413411e-01 -8.92811537e-01 -4.41374183e-01 -4.79803503e-01
3.62737477e-01 -3.91734391e-01 1.67100415e-01 -4.86260913e-02
1.94073880e+00 -2.04551148e+00 -9.59845036e-02 2.41177067e-01
-1.65075716e-02 4.67135608e-01 -2.08169714e-01 7.31237054e-01
1.94943607e-01 1.15996920e-01 6.10827446e-01 -4.36553568e-01
-5.98175917e-03 1.47381514e-01 -3.24513406e-01 8.41893023e-04
-5.61354041e-01 8.26364696e-01 -8.33975792e-01 -6.22387789e-02
8.05413276e-02 2.16705441e-01 -9.26648140e-01 3.05789471e-01
-5.17707877e-02 1.59280226e-01 -3.31182152e-01 3.66564751e-01
3.19547772e-01 -6.42762601e-01 4.62234557e-01 -6.64851740e-02
-1.51052997e-01 3.72827321e-01 -1.15578139e+00 1.60660923e+00
-8.08359981e-01 8.27353522e-02 -2.03950688e-01 -7.53197670e-01
8.45478892e-01 2.84591109e-01 5.03726006e-01 -8.02030027e-01
1.33063113e-02 2.87703034e-02 9.45230201e-02 -3.37217808e-01
7.80405045e-01 -1.94932655e-01 -9.80473161e-02 7.57536948e-01
8.88339952e-02 4.50990915e-01 4.22032207e-01 3.54214430e-01
8.57786834e-01 -3.85259062e-01 5.09498715e-01 -7.45171830e-02
3.09275985e-01 -2.55907238e-01 2.73826689e-01 1.12845588e+00
1.66492119e-01 2.80715883e-01 -2.08431736e-01 -2.16664940e-01
-6.99339449e-01 -8.16735566e-01 -5.37255630e-02 1.99908471e+00
1.03799939e-01 -7.85834968e-01 -1.93759143e-01 -6.14694178e-01
1.55388162e-01 8.84757221e-01 -5.68966985e-01 -1.97901011e-01
6.02676533e-02 -8.20530057e-01 -4.16254289e-02 4.68151152e-01
-1.17264548e-03 -8.23248863e-01 -2.46853739e-01 2.57616460e-01
8.26047212e-02 -6.58918679e-01 -7.73700237e-01 1.28006905e-01
-9.81642306e-01 -5.95128357e-01 -6.28154397e-01 -3.70375097e-01
4.26405638e-01 7.80270576e-01 1.33362150e+00 8.90901908e-02
2.16624215e-01 4.13274288e-01 -8.27436268e-01 -2.08618641e-01
-1.12004213e-01 1.65641278e-01 5.68917930e-01 -4.12855633e-02
5.61311901e-01 -7.71389961e-01 -7.76232064e-01 7.86112368e-01
-5.83789766e-01 -3.16838503e-01 7.07321405e-01 6.64281428e-01
1.98277444e-01 -1.47075653e-01 8.36007178e-01 -1.29697478e+00
1.10721219e+00 -8.92520189e-01 -1.35125503e-01 3.58621269e-01
-1.27166069e+00 -1.40406072e-01 8.83116066e-01 -9.76525486e-01
-1.21585572e+00 -3.52213770e-01 -2.92570114e-01 -1.65498108e-01
9.13859997e-03 7.99081266e-01 2.04543337e-01 1.01024896e-01
8.54829848e-01 -2.76121907e-02 -2.43711650e-01 -9.70915258e-01
6.67717576e-01 9.37253475e-01 1.68553531e-01 -5.05018950e-01
3.51482570e-01 1.47490188e-01 -7.17914224e-01 -4.20203179e-01
-1.22530091e+00 -1.03566444e+00 -4.90442097e-01 -4.51154634e-02
2.62037456e-01 -7.61616468e-01 -6.37324035e-01 -9.94578749e-02
-4.86239642e-01 -9.81045291e-02 -2.95240164e-01 7.54614830e-01
-1.96896553e-01 3.70374501e-01 -7.78143525e-01 -9.60758805e-01
-3.52046013e-01 -7.02264369e-01 4.54523712e-01 2.24636167e-01
-5.43220162e-01 -1.04745889e+00 2.13765219e-01 4.77434605e-01
7.23788977e-01 -7.08399892e-01 1.00127602e+00 -1.07384813e+00
1.13570832e-01 -4.11960691e-01 1.35292202e-01 3.62852633e-01
2.95732468e-01 -3.02507311e-01 -7.74489403e-01 -6.10275984e-01
-1.06611503e-02 -3.14635187e-01 6.53117657e-01 2.49984771e-01
1.00244582e+00 -1.52139276e-01 -2.68406391e-01 2.04515532e-01
1.27423322e+00 3.85587484e-01 3.13766479e-01 -5.73098660e-03
2.27517873e-01 3.71507972e-01 8.94042850e-01 7.40535140e-01
2.67468154e-01 7.31173217e-01 1.12941064e-01 1.34665489e-01
2.54648834e-01 -2.84967542e-01 4.30049956e-01 1.08016539e+00
-3.07043552e-01 -4.69346017e-01 -2.62577802e-01 9.73659977e-02
-1.88911939e+00 -9.64289904e-01 1.06021017e-01 2.68511295e+00
6.60117209e-01 2.37861082e-01 3.35090280e-01 -4.70428914e-02
4.91461784e-01 -1.23728000e-01 -6.85946107e-01 -6.03381276e-01
1.80794477e-01 3.11107695e-01 2.71766663e-01 3.00935626e-01
-6.53307021e-01 6.98540211e-01 7.22932816e+00 9.46120203e-01
-8.23142231e-01 2.00790107e-01 4.54949230e-01 -5.07319570e-01
-3.17228973e-01 -1.32089093e-01 -9.71047044e-01 5.20440638e-01
1.31922960e+00 -3.75113755e-01 7.27707505e-01 1.01951230e+00
2.72942275e-01 -1.45528153e-01 -1.14246500e+00 9.19370949e-01
9.27200317e-02 -1.18149710e+00 1.82784021e-01 1.20100416e-01
8.82573426e-01 1.55675218e-01 6.73805848e-02 8.08460534e-01
6.16765022e-01 -9.37616169e-01 4.21232045e-01 6.09955788e-01
8.18135381e-01 -6.46914661e-01 6.34747922e-01 8.08309019e-01
-1.07631052e+00 -6.05275691e-01 -7.24339485e-01 -5.32940388e-01
1.76597804e-01 5.69519222e-01 -3.38835865e-01 4.70451295e-01
7.37206459e-01 6.52981639e-01 -5.91131032e-01 9.68561828e-01
2.75110245e-01 7.68696070e-01 -5.37702814e-02 -1.02513246e-01
-4.53115674e-03 -3.18582237e-01 -1.01097316e-01 1.03466058e+00
6.58907235e-01 3.96493971e-01 4.34274495e-01 1.84754848e-01
5.43492772e-02 5.09012878e-01 -2.19638452e-01 -2.31056567e-02
5.89186132e-01 1.28209054e+00 -3.23074520e-01 -3.83499831e-01
-9.22598541e-01 8.10507476e-01 3.93498689e-01 2.03543067e-01
-6.29535139e-01 -9.60396081e-02 8.93379927e-01 2.84569532e-01
2.47749418e-01 -1.86410904e-01 2.40531489e-02 -1.24715781e+00
-3.73076290e-01 -8.64445210e-01 5.79981387e-01 -6.75056279e-01
-1.68229580e+00 5.51146924e-01 -1.81105852e-01 -1.33135271e+00
-3.23400855e-01 -5.63735008e-01 -5.44093013e-01 8.26371372e-01
-1.16622043e+00 -6.97903693e-01 -7.19214603e-02 6.41166687e-01
7.90220201e-01 -3.08904290e-01 9.89515841e-01 5.93420208e-01
-2.64157623e-01 7.26397038e-01 6.09035730e-01 -4.59144592e-01
1.15996945e+00 -9.98100758e-01 2.93897003e-01 1.82927102e-01
4.44013178e-01 1.18651164e+00 6.19666457e-01 -4.57342386e-01
-1.21729481e+00 -6.85889065e-01 1.00919712e+00 -7.10783958e-01
5.59994698e-01 -3.35162461e-01 -9.81330514e-01 4.17827904e-01
-1.58560053e-01 -1.32513493e-01 1.30370128e+00 8.85815144e-01
-4.11194235e-01 -1.52352611e-02 -8.73998582e-01 4.94796246e-01
1.14905620e+00 -4.27836388e-01 -6.44101799e-01 1.84663534e-01
6.47843122e-01 7.68793151e-02 -1.22096634e+00 1.87301576e-01
9.43604410e-01 -1.14471900e+00 1.06573796e+00 -9.50031459e-01
2.08158404e-01 -9.42331390e-04 -3.00157160e-01 -1.57185566e+00
-9.59718525e-01 -4.77063119e-01 -1.51546806e-01 1.08321249e+00
6.51567280e-01 -5.27549028e-01 6.69033766e-01 8.75102162e-01
-1.09357983e-02 -9.79750872e-01 -2.78442472e-01 -7.40939915e-01
4.45763916e-02 -5.85547805e-01 6.09304845e-01 9.09670532e-01
3.93982470e-01 7.45355964e-01 -8.20697665e-01 -2.18112200e-01
2.12195180e-02 3.91839623e-01 6.21474087e-01 -1.68485689e+00
-7.38729298e-01 -3.65476966e-01 2.67352432e-01 -1.48685288e+00
-1.45059094e-01 -8.57293665e-01 -3.90113980e-01 -1.40858114e+00
3.29538852e-01 -5.40105402e-01 -8.68845761e-01 9.56908166e-02
-2.34206438e-01 2.47348204e-01 1.07502997e-01 5.18332422e-01
-9.08695400e-01 1.57905862e-01 1.06199181e+00 2.18168750e-01
-4.65501070e-01 5.85271478e-01 -1.11117709e+00 6.18158400e-01
6.56585217e-01 -2.38505304e-01 -8.99425983e-01 -2.14669317e-01
7.05622911e-01 1.80687204e-01 -3.87127399e-01 -5.81656814e-01
3.89969975e-01 -5.11894643e-01 3.94829541e-01 -2.56762773e-01
2.98369676e-01 -5.93612909e-01 1.89252138e-01 -1.46543756e-01
-6.49197102e-01 1.97975963e-01 -9.22383517e-02 7.89673805e-01
-5.17104752e-02 -2.89443731e-01 4.87508476e-01 -2.16847673e-01
-8.82375062e-01 4.54506367e-01 -5.31626225e-01 -1.90117344e-01
5.04922748e-01 -2.70892680e-01 -8.77011046e-02 -6.56886101e-01
-1.14349759e+00 1.31604224e-01 3.47459614e-01 7.66484857e-01
4.53663856e-01 -1.18497443e+00 -3.06048006e-01 2.15766445e-01
3.44716012e-01 -7.87398696e-01 5.18479407e-01 6.23782039e-01
4.95579273e-01 5.34246266e-01 -7.36058727e-02 -4.90996465e-02
-1.33416343e+00 8.55700076e-01 -2.17349991e-01 -4.01406109e-01
-2.20428064e-01 8.06662560e-01 1.66189700e-01 -3.98907721e-01
2.05700621e-01 -1.32845268e-01 -7.57300317e-01 2.79078275e-01
6.31609917e-01 4.25647318e-01 1.45050287e-01 -3.27679962e-01
-9.54713002e-02 3.02776843e-01 -3.85084182e-01 1.01932101e-01
1.13931501e+00 -5.23620844e-01 3.03169429e-01 7.72922575e-01
8.10172021e-01 2.87926435e-01 -9.66449618e-01 -5.20902455e-01
6.98523670e-02 -6.26071095e-01 4.16493975e-03 -1.04931808e+00
-7.89173603e-01 5.60740173e-01 3.80970061e-01 4.04948384e-01
1.02469671e+00 -1.31367873e-02 8.02540541e-01 5.08780539e-01
7.75939465e-01 -1.19960308e+00 1.36546135e-01 5.99879324e-01
4.48964894e-01 -1.29941857e+00 1.08198643e-01 -4.44733016e-02
-9.77960408e-01 7.69114435e-01 5.16314924e-01 4.80660722e-02
1.04871976e+00 -4.18882594e-02 3.13330144e-02 1.44456625e-01
-1.41494107e+00 -1.59917027e-01 5.16622126e-01 2.78737515e-01
9.02680397e-01 1.37357756e-01 -4.59509075e-01 1.17002869e+00
-1.77526683e-01 1.51929334e-01 2.62867391e-01 6.50207996e-01
-7.67415345e-01 -1.43534565e+00 8.82591009e-02 1.00915825e+00
-4.81451780e-01 -3.09365511e-01 -2.24081129e-01 2.77971029e-01
2.56109294e-02 1.45092964e+00 -1.90763190e-01 -7.16837764e-01
3.99724841e-01 1.18409082e-01 2.03047112e-01 -1.04642153e+00
-7.97146320e-01 3.07033816e-03 1.13041244e-01 -5.15895486e-01
-1.62724257e-01 -4.35190707e-01 -7.75240600e-01 -4.12204266e-01
-7.61375189e-01 6.34556353e-01 4.75130439e-01 9.08638120e-01
5.85184753e-01 1.34560525e-01 8.48593593e-01 -7.32446849e-01
-1.10633039e+00 -1.13372040e+00 -9.53187525e-01 5.85647345e-01
-1.30610064e-01 -7.55254149e-01 -3.51752549e-01 -2.89764285e-01] | [10.117950439453125, 5.697439193725586] |
9d8f8cd5-9e00-4eed-8ee0-769f07a22606 | saac-safe-reinforcement-learning-as-an | 2204.09424 | null | https://arxiv.org/abs/2204.09424v2 | https://arxiv.org/pdf/2204.09424v2.pdf | SAAC: Safe Reinforcement Learning as an Adversarial Game of Actor-Critics | Although Reinforcement Learning (RL) is effective for sequential decision-making problems under uncertainty, it still fails to thrive in real-world systems where risk or safety is a binding constraint. In this paper, we formulate the RL problem with safety constraints as a non-zero-sum game. While deployed with maximum entropy RL, this formulation leads to a safe adversarially guided soft actor-critic framework, called SAAC. In SAAC, the adversary aims to break the safety constraint while the RL agent aims to maximize the constrained value function given the adversary's policy. The safety constraint on the agent's value function manifests only as a repulsion term between the agent's and the adversary's policies. Unlike previous approaches, SAAC can address different safety criteria such as safe exploration, mean-variance risk sensitivity, and CVaR-like coherent risk sensitivity. We illustrate the design of the adversary for these constraints. Then, in each of these variations, we show the agent differentiates itself from the adversary's unsafe actions in addition to learning to solve the task. Finally, for challenging continuous control tasks, we demonstrate that SAAC achieves faster convergence, better efficiency, and fewer failures to satisfy the safety constraints than risk-averse distributional RL and risk-neutral soft actor-critic algorithms. | ['Debabrota Basu', 'Yannis Flet-Berliac'] | 2022-04-20 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-1.13426045e-01 6.76208138e-01 -2.75505632e-01 2.21598342e-01
-9.56664920e-01 -7.78324425e-01 5.59433579e-01 1.66757286e-01
-7.81955361e-01 1.10817254e+00 -1.18639395e-02 -6.61071777e-01
-4.53678876e-01 -6.08262002e-01 -4.52211589e-01 -1.11767459e+00
-3.73573869e-01 3.98904562e-01 -1.58168197e-01 -3.79127860e-01
1.86569244e-02 1.65456355e-01 -7.31944740e-01 -5.03077507e-01
8.33300292e-01 1.09571505e+00 -3.03954214e-01 4.76518512e-01
5.12921810e-01 1.21214759e+00 -7.53386855e-01 -2.96587437e-01
6.21834457e-01 -2.76011437e-01 -5.99108636e-01 -4.66971636e-01
-1.92568555e-01 -4.84858125e-01 -1.58517644e-01 1.34090781e+00
5.44542432e-01 4.93226945e-01 5.92339814e-01 -1.55009949e+00
-1.87110707e-01 7.16101229e-01 -5.55186987e-01 -1.01919994e-01
2.20740616e-01 6.73236370e-01 1.04254293e+00 -2.76351161e-03
2.76056916e-01 1.33562171e+00 3.08125079e-01 1.01544166e+00
-1.29069984e+00 -5.57398617e-01 6.47964716e-01 -5.32273114e-01
-1.02996576e+00 -1.56263217e-01 5.48781574e-01 -3.96627426e-01
8.01589847e-01 3.36889803e-01 4.99583989e-01 1.42471194e+00
6.69724107e-01 7.96215177e-01 1.15186608e+00 -1.18113972e-01
9.60197270e-01 -4.00926135e-02 -4.32020485e-01 3.42689961e-01
3.14427465e-01 9.73589301e-01 8.64386633e-02 -6.52918994e-01
3.74754697e-01 -2.00872794e-01 -1.37539998e-01 -4.66963977e-01
-7.61128068e-01 1.15599442e+00 2.70000011e-01 -1.77870035e-01
-5.48661292e-01 5.29870033e-01 4.72159892e-01 5.35013437e-01
2.61813521e-01 9.88658726e-01 -3.83655041e-01 -4.70825844e-02
-3.86097997e-01 7.20921457e-01 8.01698744e-01 5.77705085e-01
-5.52210547e-02 7.97531128e-01 -3.56340051e-01 1.76968783e-01
5.30624509e-01 5.85295856e-01 9.93025899e-02 -1.30587780e+00
4.95210111e-01 -8.29417184e-02 6.20706797e-01 -6.98879480e-01
-4.37351078e-01 -5.29712558e-01 -5.15381098e-01 1.13012838e+00
5.64057529e-01 -8.30221534e-01 -6.59845352e-01 2.22619224e+00
2.30827257e-01 -5.45089729e-02 4.68719989e-01 8.54409814e-01
-2.07708910e-01 5.92317164e-01 2.30702505e-01 -7.90687144e-01
8.53772044e-01 -5.69366515e-01 -7.81674027e-01 -4.81912881e-01
3.70237350e-01 -1.28734097e-01 8.49467099e-01 6.26969159e-01
-1.31640267e+00 2.55820990e-01 -1.04111850e+00 7.57352591e-01
6.34290501e-02 -7.97670245e-01 3.53941828e-01 6.87108636e-01
-7.89457977e-01 4.90703791e-01 -7.92046666e-01 2.88152725e-01
3.96528721e-01 4.28930938e-01 1.70667440e-01 6.52200699e-01
-1.49904108e+00 1.21611881e+00 4.50499982e-01 -1.27253950e-01
-1.60370815e+00 -5.58356106e-01 -8.23594630e-01 4.00550440e-02
1.24533939e+00 -4.79963481e-01 1.48109972e+00 -1.03538454e+00
-1.97885609e+00 1.78124860e-01 7.84725666e-01 -8.71813536e-01
1.00096691e+00 -3.30223560e-01 -2.01132670e-02 -1.18242525e-01
-1.50306284e-01 1.88985988e-01 1.00908601e+00 -1.41803741e+00
-4.88790065e-01 -7.00147301e-02 4.67130303e-01 5.27009606e-01
1.39152437e-01 -1.46645218e-01 3.74967217e-01 -7.49003351e-01
-7.43420005e-01 -1.19145572e+00 -8.20183635e-01 -2.42001578e-01
-3.98993075e-01 -2.78372783e-03 5.78737319e-01 -2.23254621e-01
1.22144175e+00 -2.05526781e+00 2.86151558e-01 4.90914494e-01
1.18436120e-01 3.42875183e-01 -3.93735431e-02 4.17813063e-01
8.56833905e-03 2.43549690e-01 -3.96977186e-01 -2.35334933e-01
2.99494445e-01 3.07603985e-01 -7.29562879e-01 7.61472762e-01
8.74555334e-02 7.24810004e-01 -1.09617913e+00 -5.10627702e-02
2.82345638e-02 8.46420228e-03 -6.70355201e-01 3.50487143e-01
-4.89787400e-01 4.54442412e-01 -9.09199595e-01 2.99046576e-01
2.55277008e-01 4.29852247e-01 2.73997009e-01 6.78603411e-01
-5.48831299e-02 -1.46971673e-01 -1.14321399e+00 9.26055670e-01
-4.55643833e-01 3.94369438e-02 6.43616140e-01 -8.50031257e-01
5.63454330e-01 2.97651321e-01 5.14559507e-01 -5.62584221e-01
4.16772425e-01 -2.50858068e-02 6.90753907e-02 -1.40758857e-01
2.94917047e-01 -5.85560501e-01 -5.48833311e-01 6.53252304e-01
-2.40010709e-01 -3.99357259e-01 -3.92859370e-01 2.39678755e-01
1.04519808e+00 2.51931660e-02 6.16656065e-01 -5.10599554e-01
2.97003895e-01 -2.94614613e-01 9.23958898e-01 9.87606168e-01
-5.62013090e-01 -4.27805185e-02 9.64748621e-01 -2.13604167e-01
-6.67837918e-01 -1.19362330e+00 2.55461156e-01 9.06796098e-01
1.26967445e-01 -2.93642860e-02 -6.01182163e-01 -1.08287060e+00
4.36418980e-01 1.28305328e+00 -7.72516608e-01 -5.76122642e-01
-4.38878536e-01 -5.40937185e-01 5.44421911e-01 2.22647786e-01
2.67630309e-01 -8.79911602e-01 -8.97827923e-01 1.28454849e-01
2.89398402e-01 -3.69586140e-01 -7.35984623e-01 4.53356475e-01
-2.29680613e-01 -9.64731753e-01 -2.42851600e-01 -2.15883809e-03
5.27244329e-01 -3.11425298e-01 7.60139167e-01 -3.11190963e-01
7.83704892e-02 5.71747780e-01 5.55890948e-02 -8.36279452e-01
-5.20913303e-01 -5.00813186e-01 5.45774281e-01 -1.50868937e-01
-3.76508743e-01 -2.60252178e-01 -4.86562461e-01 2.38772124e-01
-8.55869472e-01 -6.10903144e-01 6.08613249e-03 1.02988470e+00
4.63986754e-01 3.89367014e-01 8.46540093e-01 -6.92941546e-01
1.07675242e+00 -6.82169139e-01 -1.18232739e+00 1.35890305e-01
-8.84851992e-01 3.05175841e-01 9.15546596e-01 -8.23754609e-01
-1.07950974e+00 1.34979030e-02 1.25270039e-01 -5.89997828e-01
3.59023124e-01 2.71722436e-01 -2.51546890e-01 -1.03813589e-01
7.13823497e-01 -8.81566852e-02 3.70480925e-01 4.49905321e-02
2.65964359e-01 2.40113169e-01 2.71999031e-01 -1.06115377e+00
1.01171231e+00 1.46511093e-01 4.69170064e-02 -2.11691007e-01
-8.29585195e-01 3.96052003e-01 1.78120881e-01 -3.09952766e-01
8.56475949e-01 -7.96836197e-01 -1.27945912e+00 1.80914909e-01
-6.22305870e-01 -7.18202293e-01 -8.69390190e-01 3.52875203e-01
-1.00869811e+00 1.39673427e-01 -2.15230927e-01 -1.58158100e+00
-2.20531896e-01 -1.24304211e+00 3.96279722e-01 2.61551470e-01
-9.40978229e-02 -1.19369948e+00 3.17107856e-01 -2.34363712e-02
4.78087783e-01 7.45662391e-01 5.66392601e-01 -7.80290008e-01
-2.36689687e-01 -6.61163628e-02 6.68178439e-01 4.11170483e-01
-1.61808446e-01 -1.54647171e-01 -6.52568817e-01 -8.25641811e-01
4.92943496e-01 -7.33632922e-01 5.69861174e-01 4.46376652e-01
8.95902038e-01 -1.08265924e+00 1.36978356e-02 5.55287480e-01
1.20372963e+00 8.70364726e-01 2.47267589e-01 5.53834498e-01
2.04146683e-01 6.29509211e-01 8.49957228e-01 8.15761685e-01
1.63530603e-01 4.87739325e-01 1.06801558e+00 2.91818351e-01
9.37993467e-01 -4.28467304e-01 9.79253888e-01 -7.01182485e-02
7.91516751e-02 -4.64739859e-01 -7.40326405e-01 -1.92514211e-02
-2.16239190e+00 -1.20971656e+00 6.36707842e-01 2.53903174e+00
8.77787232e-01 5.39669752e-01 3.16990554e-01 -3.56921762e-01
2.99087971e-01 3.33438098e-01 -1.10713148e+00 -8.44574928e-01
7.27350730e-03 -2.46461913e-01 8.22445929e-01 1.01492655e+00
-1.18055642e+00 8.59343648e-01 6.74362564e+00 9.48611856e-01
-9.44903612e-01 6.74762204e-03 8.06931019e-01 -6.55239582e-01
-4.00424391e-01 4.26702723e-02 -4.68584806e-01 4.71532047e-01
8.21217954e-01 -6.11426890e-01 9.24463272e-01 1.02559614e+00
2.44598448e-01 -2.21422955e-01 -1.07079625e+00 4.65658575e-01
-3.56348544e-01 -8.03845286e-01 -4.87745821e-01 2.62019753e-01
6.78865969e-01 -1.81115180e-01 4.64306027e-01 6.29720688e-01
1.28886175e+00 -1.31032479e+00 1.12963140e+00 3.05899501e-01
6.26251101e-01 -1.28289783e+00 6.15886509e-01 6.11735404e-01
-7.06723392e-01 -7.20045269e-01 1.63281392e-02 -3.06182474e-01
7.98119828e-02 1.07871525e-01 -3.67045552e-01 1.89966768e-01
2.42860034e-01 1.23774400e-02 2.21380308e-01 3.01473796e-01
-5.03199935e-01 3.73529166e-01 -3.43646407e-01 -6.49144035e-03
7.72275567e-01 -3.01845700e-01 9.18138981e-01 6.50957882e-01
-1.81740355e-02 3.08192879e-01 7.48277187e-01 9.08985317e-01
3.28963846e-01 -2.92907238e-01 -1.00484252e+00 -8.38030055e-02
4.84644026e-01 9.21298325e-01 -2.36571550e-01 8.70010480e-02
1.46833379e-02 3.98087800e-01 2.83649981e-01 5.42831063e-01
-1.00056136e+00 -2.54857451e-01 1.12578142e+00 -4.03293014e-01
-2.71922857e-01 -1.01498120e-01 -2.59293109e-01 -8.53974402e-01
-3.05962890e-01 -1.25149119e+00 7.40273714e-01 -2.52611727e-01
-1.27886331e+00 3.90549839e-01 2.54835542e-02 -1.03144431e+00
-5.70352435e-01 -4.10588741e-01 -9.04741704e-01 9.25384760e-01
-1.10129237e+00 -6.70265079e-01 7.30616152e-01 7.37884879e-01
2.93614179e-01 -4.18068796e-01 6.59362316e-01 -4.34687495e-01
-9.53973174e-01 6.31196022e-01 3.02201778e-01 -2.10462496e-01
4.67608005e-01 -1.45025253e+00 8.44370853e-03 8.13671768e-01
-5.55807114e-01 4.63654310e-01 1.00060415e+00 -7.18501151e-01
-1.52192807e+00 -1.10874319e+00 -1.45676792e-01 -4.21169698e-01
9.70715284e-01 -2.63440698e-01 -5.72275698e-01 7.44107306e-01
1.62160665e-01 9.35765281e-02 1.49040192e-01 -1.10189222e-01
-2.79824317e-01 -7.14890845e-03 -1.51170337e+00 9.77514982e-01
5.78898728e-01 -3.12862098e-01 -5.93705297e-01 2.27844059e-01
8.66354287e-01 -4.65266943e-01 -6.34474516e-01 2.21966341e-01
1.76666006e-01 -4.61780518e-01 9.56608236e-01 -1.02362061e+00
1.88197587e-02 -1.87257752e-01 -1.09016292e-01 -1.79032612e+00
-3.76058877e-01 -1.46882653e+00 -2.80630827e-01 6.41932070e-01
3.37853134e-01 -9.28557634e-01 5.18801093e-01 8.77026975e-01
-2.71946676e-02 -8.56862962e-01 -1.39846885e+00 -1.19855499e+00
7.12946892e-01 -2.73240805e-01 2.92165935e-01 9.03334856e-01
3.27350408e-01 -1.19848229e-01 -7.70105541e-01 4.26833868e-01
8.97814512e-01 -2.88848400e-01 3.78349036e-01 -6.27201974e-01
-6.78272247e-01 -7.23331630e-01 2.90979207e-01 -3.74204069e-01
5.40159583e-01 -5.63887000e-01 4.28119987e-01 -8.67955446e-01
-2.13964418e-01 -5.18126667e-01 -5.71821749e-01 6.55590773e-01
-2.43317977e-01 -6.65798962e-01 6.33038878e-01 -2.27085948e-01
-7.18198180e-01 7.44944036e-01 1.15763497e+00 -3.24325502e-01
-6.24906063e-01 2.79835045e-01 -9.26269770e-01 8.86611223e-01
9.48782921e-01 -4.63475108e-01 -8.53518307e-01 9.24041867e-03
4.84903336e-01 6.20820165e-01 1.66699857e-01 -5.10424793e-01
6.39834553e-02 -1.14891136e+00 -1.69958726e-01 8.34588036e-02
1.79286182e-01 -8.36978912e-01 1.43231362e-01 1.00446796e+00
-7.47779429e-01 3.06353420e-02 1.93041757e-01 7.89614677e-01
2.11095959e-01 -1.75156131e-01 1.20912027e+00 -1.24660723e-01
-7.05745956e-03 3.13531429e-01 -7.72741556e-01 5.83010793e-01
1.51570392e+00 5.03770411e-01 -5.02991736e-01 -7.60510862e-01
-6.69295788e-01 1.05321562e+00 3.18270564e-01 3.04626077e-01
5.24132967e-01 -1.06254780e+00 -5.83147109e-01 -1.70521274e-01
-3.10217887e-01 6.32142797e-02 1.78139478e-01 5.05987108e-01
2.30967198e-02 1.02125160e-01 -9.77018550e-02 7.89768100e-02
-8.51814091e-01 9.65700328e-01 8.59559894e-01 -6.09453022e-01
-3.78148019e-01 7.25638509e-01 5.68648994e-01 -1.78884253e-01
6.40132904e-01 -6.17252151e-03 4.77236276e-03 1.17404172e-02
5.73573589e-01 3.93489242e-01 -6.02165878e-01 -1.30299360e-01
-3.24340463e-01 1.02001734e-01 -5.92753850e-02 -5.28332472e-01
1.15343714e+00 -4.13615294e-02 4.16029781e-01 2.97399074e-01
4.88034010e-01 1.07206777e-01 -1.86143136e+00 -1.02340970e-02
7.28119314e-02 -3.02972108e-01 2.21491367e-01 -1.14419413e+00
-9.80263412e-01 5.27775347e-01 3.56279403e-01 3.89100403e-01
9.25922811e-01 -5.23125648e-01 3.34844917e-01 4.44606811e-01
5.83857715e-01 -1.40087867e+00 1.47491425e-01 6.73024178e-01
1.12270772e+00 -1.09297419e+00 9.17494521e-02 3.46511662e-01
-1.31652033e+00 6.30357683e-01 7.46887445e-01 -3.05851191e-01
5.04068613e-01 6.03542030e-01 1.35900797e-02 2.27724716e-01
-1.12445152e+00 5.81877530e-02 -1.29731715e-01 6.49033487e-01
-4.68844533e-01 4.13811564e-01 -5.02321720e-02 6.28906131e-01
4.71397899e-02 -4.89753217e-01 6.06696904e-01 1.08056319e+00
-4.99056309e-01 -9.01932001e-01 -4.82582599e-01 2.15928674e-01
-6.33248925e-01 1.60646856e-01 -3.21526706e-01 8.52659345e-01
-1.51908532e-01 8.85131955e-01 -1.98007062e-01 -2.56277442e-01
2.51714259e-01 -1.47448167e-01 4.28413935e-02 -3.52635831e-01
-8.67197931e-01 1.32220820e-01 1.60741419e-01 -9.31550622e-01
2.44848177e-01 -6.76779389e-01 -1.27901447e+00 -2.72300094e-01
-1.03736706e-01 3.43908161e-01 2.94283599e-01 8.10640156e-01
-1.16529698e-02 4.84276921e-01 1.19521010e+00 -4.97256815e-01
-1.69134581e+00 -4.35746282e-01 -6.45083606e-01 -5.70853539e-02
7.79044390e-01 -7.05863535e-01 -7.27647543e-01 -7.58551776e-01] | [4.337841510772705, 2.394789457321167] |
e2f34e17-acba-46e4-b23a-4f295680e23c | objectseeker-certifiably-robust-object | 2202.01811 | null | https://arxiv.org/abs/2202.01811v2 | https://arxiv.org/pdf/2202.01811v2.pdf | ObjectSeeker: Certifiably Robust Object Detection against Patch Hiding Attacks via Patch-agnostic Masking | Object detectors, which are widely deployed in security-critical systems such as autonomous vehicles, have been found vulnerable to patch hiding attacks. An attacker can use a single physically-realizable adversarial patch to make the object detector miss the detection of victim objects and undermine the functionality of object detection applications. In this paper, we propose ObjectSeeker for certifiably robust object detection against patch hiding attacks. The key insight in ObjectSeeker is patch-agnostic masking: we aim to mask out the entire adversarial patch without knowing the shape, size, and location of the patch. This masking operation neutralizes the adversarial effect and allows any vanilla object detector to safely detect objects on the masked images. Remarkably, we can evaluate ObjectSeeker's robustness in a certifiable manner: we develop a certification procedure to formally determine if ObjectSeeker can detect certain objects against any white-box adaptive attack within the threat model, achieving certifiable robustness. Our experiments demonstrate a significant (~10%-40% absolute and ~2-6x relative) improvement in certifiable robustness over the prior work, as well as high clean performance (~1% drop compared with undefended models). | ['Prateek Mittal', 'Saeed Mahloujifar', 'Alexander Valtchanov', 'Chong Xiang'] | 2022-02-03 | null | null | null | null | ['robust-object-detection'] | ['computer-vision'] | [ 4.73389417e-01 2.66784191e-01 -7.25167468e-02 2.03098834e-01
-9.73941863e-01 -1.38482201e+00 4.84598666e-01 8.21010023e-02
-2.25910500e-01 2.90058076e-01 -5.49680650e-01 -7.83515453e-01
3.52790684e-01 -8.84093881e-01 -1.24800694e+00 -8.60963643e-01
-4.35847878e-01 -2.00875744e-01 7.78134644e-01 -1.55837595e-01
2.15584747e-02 8.41119230e-01 -1.32867992e+00 1.09190144e-01
4.81658578e-01 1.07801104e+00 -3.35844040e-01 1.00277638e+00
7.68691242e-01 4.27506357e-01 -1.10320747e+00 -2.83788979e-01
8.33040893e-01 6.78485446e-03 -1.17514171e-01 -1.39405981e-01
7.82956243e-01 -5.66052794e-01 -6.43237770e-01 1.44264543e+00
3.15918885e-02 -4.77957517e-01 1.56780347e-01 -1.72398961e+00
-5.84215701e-01 3.16156089e-01 -5.64027965e-01 1.61294997e-01
4.81519066e-02 6.68902278e-01 4.86183345e-01 -3.98840487e-01
1.66290209e-01 1.21977186e+00 3.27365607e-01 8.32874119e-01
-1.04600704e+00 -1.21846592e+00 5.90831265e-02 -7.83592612e-02
-1.72947872e+00 -7.14928567e-01 3.80917519e-01 -2.59464562e-01
6.95626497e-01 7.12082088e-01 -1.75012797e-01 7.24002957e-01
5.93621671e-01 2.85138398e-01 1.08404899e+00 -1.44407377e-01
2.02213272e-01 3.38918597e-01 -7.84351379e-02 5.49716949e-01
8.86298656e-01 9.52836096e-01 -4.21849675e-02 -6.17878139e-01
2.86562562e-01 -1.71391051e-02 -3.43803078e-01 -3.99042405e-02
-1.00640404e+00 5.92746258e-01 6.96072698e-01 -3.58416945e-01
1.04111597e-01 6.31559312e-01 2.46188730e-01 4.53010887e-01
-1.36247531e-01 2.83026725e-01 -2.05552951e-01 6.15286708e-01
-1.14908621e-01 3.15304488e-01 4.99073416e-01 1.09004056e+00
6.89264953e-01 3.92017514e-01 -2.94929355e-01 -2.89082587e-01
3.79510403e-01 1.41977334e+00 -4.14153069e-01 -6.57617509e-01
4.37008530e-01 1.78887397e-01 2.49644458e-01 -1.08701098e+00
3.09768111e-01 -1.43486664e-01 -4.96187419e-01 1.04113841e+00
-6.93662837e-02 -3.28904033e-01 -1.08798683e+00 1.78106809e+00
5.78573942e-01 3.89336854e-01 4.13593739e-01 6.64006472e-01
5.34347832e-01 5.96010745e-01 3.25503945e-02 -1.47716552e-02
1.50075960e+00 -4.56269771e-01 -4.55977708e-01 -4.80017930e-01
1.30231038e-01 -6.04135275e-01 4.20799106e-01 1.37954414e-01
-6.42886221e-01 -4.65831578e-01 -1.93121505e+00 5.69663048e-01
-2.58853585e-01 -3.46068531e-01 1.77368477e-01 1.37619245e+00
-7.31368840e-01 -1.27313718e-01 -7.55942941e-01 4.20029581e-01
6.67500794e-01 6.29205406e-01 -5.16815722e-01 3.95230688e-02
-1.00635934e+00 8.17872643e-01 2.55242586e-01 -5.90306930e-02
-1.72328150e+00 -6.37561738e-01 -9.64379489e-01 -1.80656388e-01
6.51600063e-01 -7.63065815e-02 1.16129506e+00 -7.05756843e-01
-8.19314122e-01 7.01277673e-01 1.11483149e-02 -1.02808380e+00
4.16437775e-01 4.14140634e-02 -7.76547611e-01 4.43080693e-01
6.60484936e-03 6.01089776e-01 1.46614158e+00 -1.75839865e+00
-6.09338403e-01 -2.25403354e-01 4.09997344e-01 -5.48008800e-01
-1.77375674e-01 2.77876437e-01 -6.17643520e-02 -6.62855148e-01
-2.85419315e-01 -1.16634011e+00 -2.82126993e-01 3.66200328e-01
-5.29935181e-01 5.89199781e-01 1.64538920e+00 -3.96186203e-01
6.26017869e-01 -2.83539224e+00 -8.14222693e-01 3.13877076e-01
4.61105704e-01 5.99307716e-01 -2.20421419e-01 1.07374988e-01
1.42279774e-01 3.42255741e-01 -2.70132899e-01 6.91490397e-02
-1.92908719e-02 1.56236917e-01 -1.12219512e+00 1.18556464e+00
5.31344473e-01 8.56083930e-01 -5.91635942e-01 4.34129946e-02
4.38845977e-02 4.69338357e-01 -4.59565759e-01 -4.25221361e-02
-3.48396361e-01 5.33102900e-02 -4.28506643e-01 7.64858723e-01
1.44092596e+00 3.21133673e-01 -9.18444432e-03 1.45835564e-01
1.43188238e-01 4.55754213e-02 -1.12628031e+00 3.70011002e-01
-1.17423600e-02 8.25664818e-01 4.78272110e-01 -7.59122893e-02
6.89592302e-01 4.41362530e-01 -2.26048738e-01 -6.13526762e-01
1.70334876e-01 1.93556957e-02 2.07404047e-01 8.49909633e-02
5.31836092e-01 -4.89116423e-02 -5.62014341e-01 4.50698882e-01
-5.12894332e-01 2.39329576e-01 -5.14435351e-01 3.96614820e-01
1.35366857e+00 -7.21471071e-01 -1.07849494e-01 -2.89760500e-01
3.92593175e-01 1.57194640e-02 6.42713785e-01 1.03595662e+00
-3.35606605e-01 1.95304230e-01 2.86032796e-01 -3.78722958e-02
-7.08532393e-01 -1.31169689e+00 -5.70157096e-02 5.95291138e-01
8.20073366e-01 -1.38369605e-01 -7.80432999e-01 -9.70320702e-01
4.55125749e-01 5.20558536e-01 -7.03466833e-01 -6.16760612e-01
-4.86595422e-01 -2.30726019e-01 1.17530203e+00 5.73789060e-01
6.36947811e-01 -6.28759205e-01 -7.39999592e-01 -3.05425197e-01
5.52461445e-01 -1.32597709e+00 -6.42473936e-01 -3.76601927e-02
-2.70116031e-01 -1.21282268e+00 2.22764999e-01 -6.73979044e-01
1.11367965e+00 9.74529326e-01 3.34094614e-01 7.19859481e-01
-2.86328763e-01 2.23107994e-01 -2.32313529e-01 -7.65122652e-01
-1.04390323e+00 -7.45840788e-01 3.57795060e-01 2.62022972e-01
-1.87565573e-02 -4.48551960e-02 -5.12575090e-01 5.62627316e-01
-1.20886195e+00 -7.37750053e-01 4.70989913e-01 2.46114999e-01
3.31817567e-01 5.92628956e-01 5.52836597e-01 -5.50187349e-01
1.37959542e-02 -3.68673474e-01 -1.34503269e+00 1.57753727e-03
-3.00602376e-01 -1.24562770e-01 5.22028625e-01 -8.75648737e-01
-5.06332099e-01 -1.78900398e-02 1.34225011e-01 -5.52753448e-01
-1.08561248e-01 -3.50940198e-01 -8.12289298e-01 -9.87998188e-01
7.80436933e-01 9.06339139e-02 8.75426531e-02 9.28697884e-02
4.16477263e-01 3.89103204e-01 8.29176605e-01 -3.54002118e-01
1.93762791e+00 1.02041399e+00 1.24336764e-01 -4.67334807e-01
-1.38447970e-01 -6.96305037e-02 2.13519603e-01 -1.85496628e-01
5.64885914e-01 -1.07155550e+00 -1.15570843e+00 5.26033938e-01
-1.11259067e+00 -1.70072094e-01 9.02451426e-02 -1.00697361e-01
5.41554354e-02 4.00149137e-01 -2.40718544e-01 -9.54794049e-01
-1.51039481e-01 -1.41006362e+00 1.04978573e+00 7.90455341e-02
4.06728119e-01 -3.44613642e-01 -4.03411537e-01 1.60201445e-01
4.67904031e-01 6.05188191e-01 7.35497713e-01 -6.66007519e-01
-1.25729823e+00 -8.27328742e-01 -7.13472143e-02 4.09843922e-01
5.65939061e-02 -1.36747018e-01 -1.20209754e+00 -8.13327909e-01
2.28985399e-01 3.80386375e-02 8.30331683e-01 -2.66836852e-01
7.47899771e-01 -9.32813406e-01 -5.24911702e-01 4.50229079e-01
1.38717222e+00 4.65875000e-01 9.44194317e-01 2.20892578e-01
4.38081086e-01 1.25183001e-01 5.94150484e-01 2.95179635e-02
-1.79359525e-01 5.29984832e-01 1.00463939e+00 6.31271526e-02
-7.18032643e-02 -2.79107213e-01 8.49955380e-01 -3.83734584e-01
8.09909821e-01 -4.35733855e-01 -3.56127948e-01 5.14623642e-01
-1.20320308e+00 -9.71924186e-01 -1.17008306e-01 2.40205693e+00
4.67406124e-01 6.40308678e-01 -3.11451219e-02 2.47801021e-01
7.72065699e-01 2.56965488e-01 -6.30722940e-01 -2.61315942e-01
-1.00802749e-01 3.29460651e-02 1.23536372e+00 7.14078546e-01
-1.28260994e+00 8.63771379e-01 6.36445427e+00 8.04737687e-01
-1.12009978e+00 3.02299947e-01 3.19610596e-01 2.87404656e-02
-1.82434812e-01 3.00371289e-01 -9.68402922e-01 3.90287459e-01
7.83260226e-01 -1.90909028e-01 2.13799730e-01 7.62363613e-01
-1.67812586e-01 8.02096501e-02 -9.69377160e-01 1.81470722e-01
2.34504983e-01 -1.43920159e+00 -1.09885536e-01 5.39664030e-01
6.05579317e-01 -1.52830839e-01 5.89319170e-01 8.96917284e-02
3.94962192e-01 -9.60350752e-01 1.08463395e+00 -1.44372210e-01
8.88670385e-01 -8.84655118e-01 4.72913384e-01 3.07808042e-01
-1.27906525e+00 -2.95360446e-01 -2.64369011e-01 1.60753906e-01
1.82362366e-02 1.62439302e-01 -8.75190079e-01 3.23900998e-01
6.42696083e-01 -2.84162462e-01 -4.98895675e-01 8.03299248e-01
-4.54213142e-01 8.14370990e-01 -3.01688969e-01 3.23221028e-01
5.73885590e-02 6.51762426e-01 1.17597806e+00 8.70917976e-01
8.00781026e-02 1.65202439e-01 2.98062176e-01 8.78861129e-01
-1.96412683e-01 -6.96193635e-01 -7.99198031e-01 1.16219893e-01
8.35050285e-01 9.18308079e-01 -6.40603900e-01 -2.33521953e-01
-1.99552357e-01 7.37439632e-01 -4.98372734e-01 3.65069419e-01
-1.20843017e+00 -5.56225419e-01 1.18573630e+00 1.15796983e-01
8.88060153e-01 -3.92179459e-01 -4.29586053e-01 -5.55538952e-01
2.91699260e-01 -1.28639066e+00 1.61105126e-01 -1.83955044e-01
-9.46042955e-01 5.04220545e-01 -1.72290906e-01 -1.51462603e+00
3.51185024e-01 -5.79929054e-01 -8.20940614e-01 6.14502192e-01
-1.21571290e+00 -1.05751300e+00 2.20384374e-02 5.07152379e-01
-1.03479162e-01 -2.54077941e-01 5.31137168e-01 1.06770121e-01
-3.62373739e-01 1.13228595e+00 -1.89842090e-01 3.82364422e-01
3.48404199e-01 -6.43515468e-01 9.23702538e-01 1.59647894e+00
-2.59909662e-03 7.67714143e-01 7.68281162e-01 -9.51637149e-01
-2.02499843e+00 -1.54306936e+00 2.11632058e-01 -6.66020751e-01
7.91553855e-01 -6.93226874e-01 -8.90062809e-01 7.65594482e-01
1.99305847e-01 3.73762012e-01 2.71784574e-01 -7.30970502e-01
-1.23253369e+00 -3.63054127e-01 -1.76289308e+00 6.73148751e-01
4.47066993e-01 -5.85915864e-01 -3.56633216e-01 1.24248885e-01
1.30497038e+00 -4.52032745e-01 -2.20584482e-01 6.75577164e-01
3.30486149e-01 -4.69831288e-01 1.11445284e+00 -2.97959179e-01
-3.80614907e-01 -1.09659016e+00 -4.17136639e-01 -4.21252847e-01
3.95619459e-02 -1.21464825e+00 -3.07009786e-01 1.06602335e+00
3.41549546e-01 -1.11593270e+00 4.30688530e-01 2.03215703e-01
1.46128060e-02 4.64485101e-02 -1.21038616e+00 -1.33006239e+00
-8.68810788e-02 -6.22530639e-01 7.61022985e-01 5.14426827e-01
-3.06626409e-01 -2.95256406e-01 -3.93580317e-01 1.70008326e+00
9.59668517e-01 -1.12560049e-01 1.01323307e+00 -4.85541224e-01
-4.57166314e-01 -1.82643086e-01 -8.93277943e-01 -8.76944482e-01
-1.31507115e-02 -6.13724291e-01 4.99131233e-01 -5.77832460e-01
1.72904417e-01 -4.52264190e-01 -1.75911009e-01 6.87580228e-01
-2.17015430e-01 8.14185321e-01 4.01055336e-01 -1.03139549e-01
-4.20698792e-01 -3.03226337e-02 5.93342304e-01 -5.66970527e-01
1.81924701e-01 1.72909111e-01 -9.59358275e-01 3.16615939e-01
6.84670568e-01 -1.01179874e+00 -3.00830454e-01 -3.32425803e-01
-1.04240447e-01 -1.58372104e-01 1.13755953e+00 -1.05861139e+00
1.02542728e-01 -3.05236369e-01 -3.81147489e-02 -3.32796454e-01
2.66097844e-01 -1.05904722e+00 2.24000499e-01 8.48263502e-01
1.28904760e-01 -1.59380168e-01 8.96286488e-01 9.38703775e-01
1.06798030e-01 2.57430784e-02 8.93583238e-01 4.30984914e-01
-4.72006440e-01 4.01520401e-01 -5.99695206e-01 -3.03635031e-01
1.60519671e+00 -2.77894735e-01 -1.00317395e+00 -1.17147587e-01
-1.26794785e-01 9.50092003e-02 8.03086758e-01 4.21525031e-01
8.00912499e-01 -1.13245487e+00 -7.30104744e-01 6.40434325e-01
2.72325665e-01 -4.04389590e-01 3.73037845e-01 3.22787285e-01
-3.60721380e-01 6.65192679e-02 9.22897980e-02 -4.79132026e-01
-1.53280628e+00 1.18895233e+00 4.36597794e-01 4.89878118e-01
-6.79733634e-01 8.08106005e-01 5.68496108e-01 1.50448069e-01
2.41673201e-01 -1.92141861e-01 5.97388327e-01 -8.42957854e-01
9.87359107e-01 1.55128479e-01 -7.49640465e-02 -7.72722125e-01
-7.20083177e-01 3.27547461e-01 -2.34708801e-01 -1.57969654e-01
4.93656397e-01 2.39005983e-01 -7.72481710e-02 -6.03279352e-01
1.05842769e+00 7.93899775e-01 -1.39895594e+00 -6.01148885e-03
-3.24232101e-01 -7.28750288e-01 -8.10939521e-02 -6.31168902e-01
-1.13624048e+00 3.71358305e-01 6.37150764e-01 4.03172761e-01
9.67554152e-01 1.17275044e-01 8.62706304e-01 2.90407121e-01
6.65974617e-01 -1.92156047e-01 -4.31475118e-02 9.21688080e-02
7.43433952e-01 -9.95642781e-01 -3.07498034e-02 -5.91288745e-01
-3.51491839e-01 4.11491126e-01 5.17635465e-01 -4.18266445e-01
5.85634768e-01 9.13255453e-01 9.34645683e-02 -1.01052798e-01
-7.46352732e-01 8.21411759e-02 1.92687318e-01 8.49076331e-01
-9.44817483e-01 3.04888904e-01 4.36908156e-01 4.77645934e-01
3.67014147e-02 -8.24230611e-01 6.11161172e-01 1.25681269e+00
-7.87372768e-01 -8.91388059e-01 -8.91566694e-01 -2.05094114e-01
-5.17151833e-01 2.52984874e-02 -5.47816753e-01 6.62147343e-01
2.50025541e-01 1.45502520e+00 -2.00532615e-01 -7.18634546e-01
1.78065047e-01 -5.99841416e-01 1.18487976e-01 -4.01766509e-01
-3.91194403e-01 -9.62168053e-02 1.08197993e-02 -4.68645275e-01
9.80164409e-02 -3.58098775e-01 -1.42992997e+00 -3.92695934e-01
-4.54055309e-01 5.19929752e-02 5.24476230e-01 8.45245838e-01
5.07639706e-01 2.38243967e-01 1.12346590e+00 -4.54144508e-01
-6.82675064e-01 -2.28079677e-01 -2.20201612e-01 -1.89660341e-01
1.13641906e+00 -5.10978103e-01 -7.73742676e-01 4.81147394e-02] | [5.555655002593994, 7.919543743133545] |
3e521974-8391-42eb-ad23-6447152a2ee3 | a-technical-report-for-iccv-2021-vipriors-re | 2109.15164 | null | https://arxiv.org/abs/2109.15164v1 | https://arxiv.org/pdf/2109.15164v1.pdf | A Technical Report for ICCV 2021 VIPriors Re-identification Challenge | Person re-identification has always been a hot and challenging task. This paper introduces our solution for the re-identification track in VIPriors Challenge 2021. In this challenge, the difficulty is how to train the model from scratch without any pretrained weight. In our method, we show use state-of-the-art data processing strategies, model designs, and post-processing ensemble methods, it is possible to overcome the difficulty of data shortage and obtain competitive results. (1) Both image augmentation strategy and novel pre-processing method for occluded images can help the model learn more discriminative features. (2) Several strong backbones and multiple loss functions are used to learn more representative features. (3) Post-processing techniques including re-ranking, automatic query expansion, ensemble learning, etc., significantly improve the final performance. The final score of our team (ALONG) is 96.5154% mAP, ranking first in the leaderboard. | ['Yue Lin', 'Yunbo Peng', 'Cen Liu'] | 2021-09-30 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [-3.79932225e-02 -4.76972550e-01 8.10359567e-02 -4.12564844e-01
-9.66295183e-01 -4.62676257e-01 5.66707194e-01 -1.26101628e-01
-9.41256642e-01 6.81088448e-01 2.24726379e-01 2.66545951e-01
-5.30923605e-02 -3.02832842e-01 -4.52630788e-01 -3.17299426e-01
2.55936123e-02 6.44033730e-01 1.06293842e-01 -3.40524703e-01
3.79948527e-01 3.77216190e-01 -1.73638093e+00 3.42504799e-01
8.68148386e-01 8.63687098e-01 6.30525351e-02 7.22087026e-01
7.65303373e-02 3.88226926e-01 -6.21828794e-01 -9.13380206e-01
5.36300123e-01 1.08635090e-01 -8.93943191e-01 -2.23791331e-01
8.99411559e-01 -2.13016734e-01 -4.18638349e-01 9.33049977e-01
1.11010396e+00 3.09007585e-01 5.74185669e-01 -1.09296417e+00
-5.12269080e-01 2.49823719e-01 -7.05810070e-01 3.13811064e-01
4.50465530e-01 -9.57740396e-02 7.25613296e-01 -1.19645214e+00
3.72454554e-01 1.10366380e+00 1.00745904e+00 9.08690989e-01
-9.51266766e-01 -9.13062632e-01 1.10911056e-01 5.52353323e-01
-1.61041224e+00 -5.82132280e-01 4.60691303e-01 -3.24416965e-01
9.18477893e-01 5.49688876e-01 4.80574995e-01 1.26968277e+00
-4.19197917e-01 9.46864963e-01 9.20167863e-01 -4.60761279e-01
-3.32528859e-01 5.31715095e-01 3.42855036e-01 4.04791683e-01
1.62079453e-01 6.10419363e-02 -6.59855366e-01 -9.38818902e-02
4.91911829e-01 1.32849813e-01 4.78383899e-02 -5.84298447e-02
-1.00675845e+00 5.05551159e-01 3.84535789e-01 2.12293714e-01
-4.09774110e-02 -7.41403876e-03 4.30479765e-01 2.93722957e-01
2.86096752e-01 5.96880555e-01 -4.59827214e-01 -4.12735008e-02
-9.95492578e-01 5.25323749e-01 4.69268709e-01 8.31787407e-01
5.88958561e-01 -3.26083511e-01 -3.91289979e-01 1.16865230e+00
2.05951314e-02 5.07474422e-01 6.38239145e-01 -7.33606994e-01
6.56725764e-01 5.27213693e-01 2.29070067e-01 -8.13160717e-01
-4.98566359e-01 -6.99052751e-01 -8.95430982e-01 -5.29745780e-02
4.04743165e-01 -2.87143916e-01 -9.09505248e-01 1.56814945e+00
9.31148827e-02 3.78003091e-01 -1.36726499e-01 6.93749011e-01
8.72844338e-01 3.10855836e-01 1.97354957e-01 7.01935440e-02
1.40691698e+00 -1.36643755e+00 -4.40792114e-01 -2.45599538e-01
6.70243025e-01 -8.02527785e-01 7.41501272e-01 4.64649856e-01
-9.41160858e-01 -1.10969245e+00 -9.52288628e-01 -3.26770023e-02
-5.64391553e-01 5.82608879e-01 5.25583148e-01 8.47794712e-01
-1.22990310e+00 3.84585589e-01 -2.48529404e-01 -5.55972934e-01
1.74273759e-01 7.45313406e-01 -7.84027159e-01 -2.07580522e-01
-1.10627615e+00 9.52929795e-01 2.68053263e-01 4.01789695e-02
-6.12128437e-01 -8.20284367e-01 -5.61106682e-01 -1.09148867e-01
3.25368047e-01 -9.30628419e-01 9.99813557e-01 -7.75698662e-01
-1.24237883e+00 1.20466602e+00 -2.43651375e-01 -4.63637292e-01
7.15359211e-01 -7.32888043e-01 -6.50062263e-01 -1.90289080e-01
1.13888577e-01 9.14317071e-01 7.21345842e-01 -1.13176084e+00
-8.86282265e-01 -5.52280188e-01 -4.02511984e-01 4.36356157e-01
-7.38405585e-01 3.64146501e-01 -8.84790778e-01 -6.49041235e-01
-3.58237743e-01 -9.18951452e-01 -3.47923517e-01 -5.12800157e-01
-2.86800325e-01 -4.34351563e-01 5.63629687e-01 -9.37990129e-01
1.24011528e+00 -2.16818237e+00 -1.52440369e-02 2.80497342e-01
1.91338450e-01 4.71925557e-01 -1.75471485e-01 2.19213635e-01
-1.55180410e-01 1.71340451e-01 1.80950567e-01 -9.87834156e-01
-8.61328691e-02 -2.90109038e-01 -2.37222999e-01 1.99587807e-01
-4.38606068e-02 8.94735157e-01 -4.95959342e-01 -5.17231703e-01
2.14725152e-01 4.65201080e-01 -4.02140439e-01 2.92399913e-01
4.79302049e-01 5.29656112e-01 -2.17891395e-01 5.52184880e-01
8.49608123e-01 -1.27967326e-02 -1.43662676e-01 -4.40421462e-01
-2.24191889e-01 4.79222760e-02 -1.40843713e+00 1.80566609e+00
-2.43587956e-01 3.71076822e-01 -1.16687924e-01 -1.02314425e+00
9.42616820e-01 -4.37854789e-02 5.82532406e-01 -7.88873553e-01
-8.32149163e-02 9.42451581e-02 -3.01619172e-01 -4.95286614e-01
6.82444036e-01 4.37980056e-01 -2.52504032e-02 2.33314633e-01
4.03759591e-02 5.78014314e-01 2.02254027e-01 7.06318021e-02
7.20038593e-01 1.35117367e-01 -1.03140123e-01 -2.06681073e-01
7.89945483e-01 -1.30969837e-01 6.49934411e-01 9.12022591e-01
-3.51836771e-01 8.88410509e-01 -2.56562084e-01 -6.81981862e-01
-1.25290775e+00 -7.51682520e-01 -1.57779604e-02 1.41732705e+00
1.06155246e-01 -6.12118900e-01 -7.21971750e-01 -6.95910215e-01
3.00584547e-02 3.19414735e-01 -6.60807729e-01 -7.27235004e-02
-7.72407889e-01 -1.11331177e+00 7.05418587e-01 6.57135427e-01
8.67510617e-01 -6.82260633e-01 6.78002909e-02 8.68990794e-02
-4.02118355e-01 -1.21269834e+00 -6.43316150e-01 -4.29463595e-01
-6.38246596e-01 -8.38865340e-01 -1.08752203e+00 -9.43000615e-01
6.67404711e-01 5.23884416e-01 1.08480024e+00 2.93979257e-01
-5.42786181e-01 5.18927097e-01 -3.02846074e-01 -3.70828509e-01
3.23712677e-01 1.69700608e-01 4.67086315e-01 2.27826923e-01
5.39044023e-01 -2.71789581e-01 -7.63502598e-01 5.77201247e-01
-3.69746238e-01 -1.17592588e-01 5.59543431e-01 8.38804901e-01
5.91908932e-01 -3.77684906e-02 3.95523757e-01 -7.68123329e-01
5.40382206e-01 4.19439450e-02 -2.48789340e-01 7.12839365e-01
-7.27661133e-01 -1.25194624e-01 1.96731403e-01 -4.71813858e-01
-9.70651448e-01 2.50017583e-01 -2.31783062e-01 -1.66431740e-01
-1.36787474e-01 -1.69534087e-02 -3.43298875e-02 -3.33360910e-01
7.35073805e-01 3.85042906e-01 -1.39501736e-01 -9.32563424e-01
2.52911717e-01 7.70615280e-01 7.00903058e-01 -4.59854275e-01
8.43871772e-01 3.22781473e-01 -2.97762752e-01 -7.54273713e-01
-6.28970742e-01 -9.29188132e-01 -7.50635743e-01 -7.70885497e-02
8.61814916e-01 -1.45687413e+00 -8.92689288e-01 5.66936731e-01
-9.76937711e-01 2.25915127e-02 9.40646604e-03 5.88007927e-01
-1.48574814e-01 5.70028186e-01 -2.56782085e-01 -8.81487727e-01
-6.47058427e-01 -8.90469193e-01 9.07053590e-01 6.13669634e-01
-4.09146361e-02 -6.08835459e-01 2.59951532e-01 8.64538491e-01
5.80918133e-01 -1.85809121e-01 9.28925201e-02 -9.71704721e-01
-3.88389856e-01 -5.04581749e-01 -3.66477937e-01 3.23707670e-01
-1.96245626e-01 -2.51545250e-01 -1.14916909e+00 -6.61876380e-01
-4.72152382e-01 -3.02309155e-01 1.14988005e+00 1.74745053e-01
1.44714975e+00 -3.35598625e-02 -6.32877707e-01 7.81616986e-01
9.97528553e-01 -2.28307620e-01 7.34572172e-01 5.63543320e-01
8.03080678e-01 6.79230988e-01 5.48140228e-01 3.33286583e-01
8.02234769e-01 1.05623174e+00 -2.75360234e-02 -1.71223938e-01
-1.84651896e-01 -2.66046494e-01 4.12645489e-01 6.44497633e-01
-6.55490279e-01 7.14254603e-02 -8.28053236e-01 3.58166158e-01
-2.00961709e+00 -1.09601736e+00 -6.17727749e-02 2.48720384e+00
4.60813373e-01 -2.34898031e-01 6.78100765e-01 -5.55275343e-02
9.03378665e-01 -3.31249118e-01 -3.90097767e-01 9.51677635e-02
-4.04253155e-01 -5.92727959e-02 7.04358459e-01 5.17099798e-01
-1.47649670e+00 1.04200363e+00 6.94914722e+00 9.77067709e-01
-7.72723436e-01 7.97650963e-02 6.72998250e-01 -1.65936843e-01
1.66667283e-01 -2.98795223e-01 -1.48058355e+00 5.38254917e-01
6.79789543e-01 -9.95323341e-03 4.22163248e-01 8.36049378e-01
-3.88276935e-01 1.66736841e-01 -9.67585564e-01 1.80417144e+00
6.38666153e-01 -1.10050178e+00 -8.10527876e-02 8.04990977e-02
6.35683358e-01 2.12819930e-02 2.37772927e-01 6.66285694e-01
1.92031726e-01 -1.26636362e+00 2.82768279e-01 7.94797838e-01
7.21033752e-01 -7.68088698e-01 1.01368093e+00 1.91193059e-01
-1.17608929e+00 -3.70827645e-01 -5.01183808e-01 1.54352948e-01
2.58985572e-02 2.26281792e-01 -5.41820705e-01 8.34522784e-01
1.10405803e+00 6.38648927e-01 -1.14676595e+00 1.55701566e+00
1.52811646e-01 2.47372657e-01 -4.32436377e-01 1.74786210e-01
-3.42724890e-01 6.96325600e-02 3.40059280e-01 1.46931636e+00
2.89312333e-01 -6.95556477e-02 3.37157845e-01 1.76546678e-01
-9.78250280e-02 2.18404636e-01 -3.04892600e-01 3.04862410e-01
2.28262350e-01 1.41073716e+00 -1.41235545e-01 -3.04531068e-01
-2.83786565e-01 1.22972238e+00 5.15904069e-01 3.85294586e-01
-7.28967488e-01 -2.26561815e-01 6.32171869e-01 1.11208996e-02
1.61405548e-03 -1.56387985e-01 -2.86812633e-01 -1.38023102e+00
3.28540206e-02 -8.74438643e-01 7.69178629e-01 -4.08106208e-01
-1.59488368e+00 7.81318128e-01 -1.71520531e-01 -1.14294338e+00
-2.90929914e-01 -7.14517236e-01 -5.15195489e-01 8.75674725e-01
-1.40790391e+00 -1.29948175e+00 -4.31813449e-01 8.71829391e-01
4.41448003e-01 -8.92728209e-01 9.03586566e-01 8.77993345e-01
-9.91422117e-01 1.47233939e+00 1.39266491e-01 2.51580626e-01
1.19639492e+00 -1.18159842e+00 3.48098189e-01 8.17761838e-01
3.31613719e-01 6.17738187e-01 4.73819166e-01 -4.20777798e-01
-1.19122291e+00 -8.47518861e-01 9.74041581e-01 -8.05576146e-01
2.56496519e-01 -4.95095372e-01 -6.63046658e-01 4.82618392e-01
6.15137480e-02 -8.44390243e-02 9.90374625e-01 6.03845477e-01
-4.56679285e-01 -5.75671017e-01 -1.00681961e+00 3.74252558e-01
1.12195969e+00 -4.44400489e-01 -3.58755410e-01 5.07934034e-01
2.61218220e-01 -1.56119168e-01 -6.76563799e-01 5.46335578e-01
7.97056675e-01 -7.02099919e-01 1.53276682e+00 -9.43309247e-01
-2.90316164e-01 -3.37137341e-01 -4.78377603e-02 -9.89458919e-01
-6.31009579e-01 -6.98566794e-01 -2.54547503e-02 1.47151244e+00
4.91825879e-01 -4.41397756e-01 9.84697580e-01 9.78548408e-01
2.20459759e-01 -3.96544605e-01 -8.69200051e-01 -7.33561456e-01
-2.35139400e-01 -2.12266207e-01 4.28016305e-01 7.84371436e-01
-2.76423723e-01 4.40139323e-01 -9.80948627e-01 -2.68859472e-02
8.58652830e-01 -2.18841597e-01 1.22287095e+00 -1.29111040e+00
-2.43387133e-01 -4.10492152e-01 -4.06841218e-01 -1.18925881e+00
-1.55130699e-01 -6.87486589e-01 -4.26581532e-01 -1.28190756e+00
5.60860455e-01 -6.72648191e-01 -6.51442111e-01 4.87827003e-01
-5.54622889e-01 5.63377261e-01 5.33957601e-01 4.53080833e-01
-1.05339837e+00 4.14272338e-01 8.90144229e-01 -1.65600449e-01
-1.70893341e-01 2.31276497e-01 -9.13866103e-01 4.38250244e-01
9.54045177e-01 -2.31145605e-01 -9.19118077e-02 -5.66612959e-01
1.91688135e-01 -6.75141335e-01 5.09609044e-01 -1.21777332e+00
6.32807612e-01 1.99451387e-01 9.78823960e-01 -7.12544024e-01
6.70789003e-01 -6.03766561e-01 1.41103148e-01 2.67919004e-01
-1.06305435e-01 3.96481574e-01 1.80455297e-01 3.92890334e-01
-1.71899945e-01 -2.58935213e-01 4.98822778e-01 -1.05696425e-01
-9.67021227e-01 4.64846998e-01 2.54833907e-01 -7.91544840e-02
9.12511468e-01 -2.28719682e-01 -2.29742512e-01 -2.86407083e-01
-9.18191254e-01 5.42657554e-01 1.83260351e-01 6.15252793e-01
4.43997711e-01 -1.41610336e+00 -1.18920934e+00 1.40310973e-01
1.19361550e-01 -4.23681915e-01 5.95974803e-01 7.76566744e-01
-3.80540758e-01 2.97575980e-01 -2.38499105e-01 -4.65719521e-01
-1.54213929e+00 5.16301394e-01 2.44389191e-01 -5.15381455e-01
-4.35594618e-01 1.24612367e+00 2.57099718e-02 -6.33236408e-01
5.30598879e-01 6.50397658e-01 -6.28329158e-01 1.27397701e-01
1.14130068e+00 6.07650459e-01 3.70647246e-03 -7.72736549e-01
-5.89356005e-01 8.90083611e-01 -6.24769866e-01 -8.80244076e-02
1.27832294e+00 -1.86967388e-01 1.32232353e-01 -3.74945812e-02
1.08304846e+00 5.95367253e-02 -8.63205135e-01 -2.76892275e-01
1.36471093e-02 -5.18877387e-01 -1.90806210e-01 -1.02725220e+00
-8.96428883e-01 7.84625411e-01 1.10365725e+00 -3.73789184e-02
1.00539398e+00 -1.78942651e-01 7.67912567e-01 5.09222567e-01
2.46374771e-01 -1.53173041e+00 6.64103776e-02 4.20342714e-01
8.80175471e-01 -1.48139024e+00 1.52161673e-01 -4.02353555e-01
-8.00850034e-01 8.43629539e-01 8.68375540e-01 -1.11860856e-01
5.16934395e-01 -1.64336488e-01 1.21318720e-01 2.23221973e-01
-4.96188372e-01 -3.69360656e-01 6.11584544e-01 7.49231696e-01
2.46452495e-01 9.82284267e-03 -2.54067093e-01 9.05279458e-01
-3.17991555e-01 -1.92633554e-01 -1.82685450e-01 3.58467817e-01
-2.06133232e-01 -1.40353978e+00 -4.62628454e-01 3.07519883e-01
-5.76438844e-01 -9.56271887e-02 -4.24542755e-01 6.09983146e-01
4.28032666e-01 8.86321247e-01 -3.78297456e-02 -8.58652115e-01
5.44837296e-01 -2.01077107e-02 6.11087203e-01 -2.51146942e-01
-7.81063259e-01 -1.67928353e-01 1.61803097e-01 -5.61816216e-01
-2.74778664e-01 -6.43971443e-01 -6.21349454e-01 -3.04755479e-01
-2.92175144e-01 2.22583205e-01 7.39471912e-01 7.31098175e-01
7.41367638e-01 1.56015232e-01 6.37410104e-01 -7.13586688e-01
-6.35364056e-01 -1.15636396e+00 -2.31610000e-01 7.18975961e-01
1.13996621e-02 -7.31742918e-01 -9.94240865e-02 2.15974599e-02] | [14.718932151794434, 0.9128021597862244] |
4fab06ee-738d-48a9-9592-4a6e6488c019 | assessing-the-coherence-modeling-capabilities | null | null | https://openreview.net/forum?id=NkOTih-TP7T | https://openreview.net/pdf?id=NkOTih-TP7T | Assessing the Coherence Modeling Capabilities of Pretrained Transformer-based Language Models | The task of ordering a shuffled set of sentences into a coherent text is used to evaluate the capacity of a model to understand causal and temporal relations between entities and events. Recent approaches rely on pretrained Transformer-based models, but it remains unknown whether the differences between them, such as size, pretraining data and objectives, affect their coherence modeling capacity. We present a simple architecture for sentence ordering that relies exclusively on pretrained Transformer-based encoder-only models. This allows us to compare the coherence modeling capabilities of the monolingual and multilingual versions of BERT, RoBERTa, and DistilBERT. We show that RoBERTa-based models outperform BERT-based models and are more robust when ordering longer documents with more than 10 sentences. Thus, the intuitive advantage offered by sentence-based objectives such as Next Sentence Prediction used in BERT is effectively compensated by the higher amount and diversity of the training data used in RoBERTa. However, the difference between multilingual versions of BERT and RoBERTa is narrower. This suggests that exposure to different languages partially makes up for the benefits of larger and more diverse training data. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['sentence-ordering'] | ['natural-language-processing'] | [-3.71205062e-01 3.52575451e-01 -1.02570906e-01 -7.64007688e-01
-6.55263901e-01 -6.09562397e-01 9.91631925e-01 5.37059486e-01
-6.08038068e-01 8.39436650e-01 8.98126304e-01 -5.39237440e-01
-2.05120057e-01 -6.18867338e-01 -6.97523654e-01 -5.55772111e-02
-3.58263820e-01 6.72240436e-01 8.26042518e-02 -6.08517945e-01
-8.82651284e-02 1.23483099e-01 -1.18160641e+00 5.88097036e-01
9.11787331e-01 3.96345675e-01 5.30340612e-01 5.60831070e-01
-6.13230132e-02 9.14149225e-01 -5.68216801e-01 -6.16564095e-01
-7.09630325e-02 -5.55344701e-01 -1.05643463e+00 -2.00197548e-01
4.03004825e-01 -4.17240411e-01 -5.35907269e-01 4.94805634e-01
2.60891676e-01 1.08417198e-01 6.19887710e-01 -7.89921343e-01
-5.63532114e-01 1.39424169e+00 1.41789215e-02 4.03602362e-01
4.24548626e-01 3.45399044e-02 1.49105811e+00 -6.53088570e-01
8.64199996e-01 1.35119689e+00 6.32262230e-01 2.40143031e-01
-1.37547624e+00 -2.98692554e-01 5.26938617e-01 3.57078016e-01
-1.12887228e+00 -7.30239987e-01 4.94315386e-01 -5.74108541e-01
1.40480697e+00 3.59041214e-01 8.41932118e-01 1.09287357e+00
5.23075700e-01 6.47532761e-01 9.85601544e-01 -3.52365434e-01
-1.70008689e-01 3.85112017e-01 3.01657677e-01 4.66355950e-01
2.24275097e-01 3.20557877e-02 -8.67142558e-01 7.73288384e-02
2.68675834e-01 -2.85923898e-01 -3.95619631e-01 -9.19184536e-02
-1.32976913e+00 6.64253652e-01 4.77641404e-01 6.26442075e-01
-1.58417806e-01 -1.69896156e-01 5.45829773e-01 6.64016724e-01
5.16609550e-01 1.00842118e+00 -6.64358675e-01 -1.47707969e-01
-1.08228433e+00 1.72359258e-01 8.31289172e-01 7.86607563e-01
6.49845123e-01 -5.25041632e-02 -2.43270606e-01 6.36550546e-01
3.70952904e-01 2.68876702e-01 5.38362205e-01 -7.83966184e-01
6.89715087e-01 3.33849490e-01 1.17959213e-02 -9.84663904e-01
-4.85888809e-01 -5.54499745e-01 -2.46741191e-01 -4.38849866e-01
4.42554295e-01 -1.13590583e-01 -4.37558860e-01 2.06053209e+00
-2.06328049e-01 -4.96233255e-01 1.17798708e-01 7.05079973e-01
5.81719518e-01 6.13796711e-01 8.51953477e-02 -2.51303613e-01
1.10841870e+00 -8.19551528e-01 -3.97266954e-01 -5.73259234e-01
9.25140083e-01 -6.84966445e-01 1.17158055e+00 -6.57792836e-02
-1.07031524e+00 -5.07256329e-01 -1.07980025e+00 -4.10806656e-01
-2.38956586e-01 3.78924347e-02 5.92012286e-01 3.47848356e-01
-1.09044182e+00 8.65052521e-01 -8.53903472e-01 -6.79831624e-01
-4.69156131e-02 2.04585448e-01 -3.60232115e-01 1.42573506e-01
-1.60105705e+00 1.39146721e+00 4.49328005e-01 -7.41378590e-02
-6.54141545e-01 -7.09293664e-01 -9.40155029e-01 3.91251773e-01
-1.70212671e-01 -6.40740514e-01 1.29766572e+00 -7.27488637e-01
-1.17889321e+00 8.45812500e-01 -3.52606535e-01 -6.81276858e-01
2.93112457e-01 -2.90040523e-01 -3.50950986e-01 -1.00368083e-01
3.53726715e-01 5.64174950e-01 2.08479375e-01 -9.08044398e-01
-3.89242798e-01 -3.59543324e-01 2.78969735e-01 4.73513365e-01
-4.33684826e-01 8.57007354e-02 -2.21661795e-02 -3.27173471e-01
-1.40371174e-01 -5.97967565e-01 4.26277183e-02 -7.53114820e-01
-4.59645271e-01 -2.12145641e-01 7.88705572e-02 -8.35152626e-01
1.43131149e+00 -1.94237721e+00 9.71095040e-02 -1.97565645e-01
8.67944434e-02 -3.27262580e-01 -3.28449786e-01 1.01860631e+00
-1.81277320e-01 4.49380696e-01 2.12171048e-01 -4.55907136e-01
3.82189127e-03 9.34715122e-02 -5.03613830e-01 2.31785327e-01
2.28429645e-01 7.94236541e-01 -9.33493912e-01 -4.99833107e-01
1.29017709e-02 1.17894746e-01 -5.13823152e-01 -4.99999784e-02
-2.11702555e-01 4.75059599e-01 -1.48687854e-01 1.10426687e-01
1.13192059e-01 -2.81165332e-01 6.70530021e-01 -1.26671478e-01
-3.52409214e-01 1.26381767e+00 -5.27349114e-01 1.52541673e+00
-7.52085388e-01 1.09384906e+00 -2.18825087e-01 -7.42823601e-01
5.97502530e-01 5.05263269e-01 1.49809793e-01 -9.36222494e-01
-6.59979433e-02 2.87597030e-01 5.28446019e-01 -5.00094950e-01
7.97045171e-01 -4.10518944e-01 -1.06641926e-01 4.86520827e-01
6.43028393e-02 -2.53259361e-01 8.10142100e-01 5.67487836e-01
8.62672925e-01 -3.36417444e-02 8.69586691e-02 -5.33324659e-01
1.63021073e-01 8.65933374e-02 6.19175494e-01 7.28569448e-01
6.71394616e-02 3.29012394e-01 6.11657679e-01 -3.67085665e-01
-1.27232265e+00 -9.92251933e-01 -4.59455669e-01 1.06262970e+00
-4.90178093e-02 -8.97196651e-01 -1.07252903e-01 -5.82865179e-01
-1.31935952e-02 1.26570344e+00 -5.07194936e-01 -9.43255574e-02
-7.03397274e-01 -5.84197402e-01 4.22581702e-01 4.51560825e-01
1.93567321e-01 -5.80806136e-01 -4.61878985e-01 2.96261221e-01
-6.21812165e-01 -1.04142511e+00 -4.32564557e-01 2.05537394e-01
-9.76612389e-01 -6.90640926e-01 -2.79507220e-01 -5.59020996e-01
3.93870413e-01 -1.19222060e-01 1.58364737e+00 -2.88650841e-01
3.51315916e-01 2.25973129e-01 -2.53131598e-01 -2.59341657e-01
-6.08325422e-01 3.88808936e-01 3.72718759e-02 -4.36730325e-01
9.74721313e-02 -4.15459543e-01 -3.16266686e-01 2.38980260e-02
-5.86454272e-01 3.19413483e-01 6.08701050e-01 9.75272715e-01
-1.37435362e-01 -8.63232091e-02 5.95745742e-01 -9.96892452e-01
6.68968856e-01 -5.66059053e-01 -2.04555780e-01 4.54267472e-01
-7.04344809e-01 3.97008210e-01 6.38816178e-01 -1.75948843e-01
-1.06382513e+00 -5.50181329e-01 -1.23644300e-01 1.45552471e-01
2.38247648e-01 1.03462064e+00 1.76792830e-01 6.80662453e-01
6.23869777e-01 -4.68943901e-02 -2.63509601e-01 -3.35406363e-01
3.41812879e-01 4.15843278e-01 7.49560371e-02 -5.20396888e-01
3.76643330e-01 1.36038452e-01 -4.58797663e-01 -7.30077744e-01
-9.08565044e-01 -1.54261008e-01 -5.85400283e-01 -9.78568941e-02
7.82280922e-01 -1.08938730e+00 -2.96796888e-01 1.12918481e-01
-1.37657464e+00 -5.87001801e-01 -3.69496018e-01 7.35906720e-01
-3.01955521e-01 7.48891979e-02 -8.24054182e-01 -5.76362908e-01
-1.22452274e-01 -9.29604709e-01 8.16190183e-01 -1.47587702e-01
-7.27533221e-01 -1.30471683e+00 1.04913190e-01 2.62418538e-01
3.69049817e-01 -1.52790040e-01 1.27443123e+00 -8.49238813e-01
-4.47534233e-01 -6.12544008e-02 1.64400741e-01 1.83353141e-01
1.88317165e-01 8.69419500e-02 -7.02310443e-01 -3.52392346e-01
9.27954167e-02 -3.81950319e-01 9.03867900e-01 1.99625716e-01
4.05608177e-01 -6.12209976e-01 -3.42561543e-01 1.83410004e-01
1.18557060e+00 1.35872141e-01 2.33998984e-01 4.33610290e-01
3.54101866e-01 8.35039973e-01 4.07749385e-01 1.02475278e-01
9.38995481e-01 6.25994265e-01 -5.82432486e-02 1.04511909e-01
-2.33633947e-02 -4.82911050e-01 7.12138772e-01 1.31819546e+00
3.11425686e-01 -3.30645293e-01 -1.07937312e+00 8.38509977e-01
-1.69905198e+00 -1.15223432e+00 1.72123208e-03 2.25280309e+00
1.28989255e+00 4.14019257e-01 -1.45526618e-01 -1.52872115e-01
2.95608819e-01 5.10153472e-01 -1.95899829e-01 -5.79567254e-01
-3.26260298e-01 -1.43140942e-01 1.80761606e-01 8.01396668e-01
-8.22179258e-01 6.67479634e-01 7.19698668e+00 4.98913497e-01
-1.23791420e+00 -7.48765916e-02 7.48936951e-01 -3.70324165e-01
-7.83132017e-01 2.53451407e-01 -7.91394234e-01 4.79554653e-01
1.43685365e+00 -4.87913668e-01 1.68007508e-01 2.90498406e-01
4.67267305e-01 -1.90979853e-01 -1.69128454e+00 2.55090266e-01
7.31112808e-03 -1.42379189e+00 3.21466401e-02 -1.78656951e-01
6.44228518e-01 3.62784535e-01 -5.91739416e-02 6.03646278e-01
4.34363902e-01 -9.79218841e-01 1.16006064e+00 3.00158024e-01
4.44771022e-01 -2.61267930e-01 7.01846302e-01 4.97714996e-01
-8.96216512e-01 -5.39298356e-02 -2.79133826e-01 -4.86195803e-01
2.63966382e-01 7.51736283e-01 -9.25554276e-01 6.54081881e-01
5.35586298e-01 7.80087352e-01 -8.22778046e-01 5.07202744e-01
-2.91555673e-01 7.68565297e-01 -2.28209034e-01 -1.45395368e-01
1.25316292e-01 -8.97592958e-03 5.40975332e-01 1.20248878e+00
5.96001484e-02 -2.55255640e-01 1.38398007e-01 7.18450129e-01
4.78706992e-04 2.32213765e-01 -6.34241104e-01 -3.96097332e-01
6.52022183e-01 8.39285433e-01 -3.31590116e-01 -3.44665766e-01
-5.70366204e-01 6.39575303e-01 6.37186110e-01 3.10196400e-01
-6.61056221e-01 -5.41807003e-02 2.93810844e-01 2.94595510e-01
8.75948817e-02 -5.13773501e-01 -5.07262886e-01 -1.42299628e+00
6.19955780e-03 -9.07183826e-01 2.40868315e-01 -7.66953588e-01
-1.01587069e+00 7.70748615e-01 1.23370171e-01 -7.49704361e-01
-4.90866810e-01 -2.21001118e-01 -6.02454484e-01 1.07212949e+00
-1.21920431e+00 -1.07522547e+00 2.24780962e-01 6.49672300e-02
6.99953020e-01 6.15065806e-02 6.85691416e-01 1.96276665e-01
-6.05737686e-01 3.53420466e-01 2.35745952e-01 1.92564040e-01
9.53425527e-01 -1.41518235e+00 3.73229712e-01 1.02932608e+00
4.29653108e-01 1.07723367e+00 9.57080007e-01 -6.10918045e-01
-8.25920761e-01 -8.01812589e-01 1.88324356e+00 -7.45964944e-01
8.42629015e-01 -3.97262841e-01 -6.44510150e-01 1.03332424e+00
7.49748647e-01 -8.66590023e-01 6.31966531e-01 9.53197360e-01
-4.83569890e-01 -2.92467207e-01 -5.03983200e-01 8.21739733e-01
7.84816146e-01 -7.99606860e-01 -7.95091629e-01 3.62182677e-01
9.26226020e-01 -2.36393109e-01 -1.09073937e+00 3.30357730e-01
3.05497229e-01 -1.08170629e+00 5.80235779e-01 -7.06514299e-01
8.61607969e-01 6.99350908e-02 -8.33927020e-02 -1.46573675e+00
-5.15437186e-01 -2.66116202e-01 3.29360336e-01 1.26149178e+00
1.03482246e+00 -7.21915901e-01 2.30711773e-01 6.58882439e-01
-1.90962866e-01 -8.85102272e-01 -7.40648389e-01 -8.31322968e-01
4.97449607e-01 -2.35680029e-01 4.75209475e-01 1.01270998e+00
4.04502720e-01 9.18305814e-01 2.71174777e-02 -1.03910856e-01
1.54818803e-01 2.80589074e-01 2.13669747e-01 -9.69152987e-01
-4.46759850e-01 -3.16010892e-01 -3.86746228e-03 -1.03508961e+00
2.48379841e-01 -1.11770177e+00 -4.61176895e-02 -1.64257431e+00
3.12880188e-01 -5.69990754e-01 1.04064131e-02 3.83536011e-01
-3.02145809e-01 -3.07415158e-01 4.69188005e-01 4.16927546e-01
-3.40914249e-01 6.31295025e-01 9.23178971e-01 -1.62618324e-01
-2.68754900e-01 -1.32135674e-01 -7.92718053e-01 5.60229838e-01
6.93838000e-01 -3.56138229e-01 -6.73054099e-01 -1.02187192e+00
5.50812781e-01 9.42659378e-02 2.20874175e-01 -6.28075242e-01
3.21140140e-01 3.33724692e-02 1.85981274e-01 -4.25016195e-01
1.76629812e-01 -2.99418211e-01 1.75119951e-01 3.82878095e-01
-8.44375312e-01 3.68935168e-01 2.79596061e-01 2.77310401e-01
-4.72321033e-01 -2.22417802e-01 5.13589919e-01 -2.88301200e-01
-4.11092520e-01 -2.06923455e-01 -5.79820812e-01 2.87418842e-01
5.10687530e-01 3.25837806e-02 -3.98016483e-01 -5.54984987e-01
-8.86120796e-01 2.62609303e-01 3.00923616e-01 5.13693392e-01
2.13579655e-01 -1.06535530e+00 -9.88533020e-01 -5.85494637e-02
2.46334285e-03 -2.99226552e-01 9.35779139e-02 1.06702721e+00
-3.87147099e-01 9.35376823e-01 5.75094819e-02 -4.64319855e-01
-1.07791591e+00 2.60083824e-01 3.39830667e-01 -6.98832631e-01
-3.07905614e-01 8.20822775e-01 2.43647099e-01 -4.12938565e-01
-1.18435606e-01 -3.81554693e-01 -8.60103741e-02 2.14104280e-01
9.12635494e-03 1.33710712e-01 1.32594734e-01 -4.47974682e-01
-4.67065871e-01 4.09965590e-02 -3.83588880e-01 -4.52065319e-01
1.24107146e+00 -4.16059673e-01 -2.07506821e-01 8.96136403e-01
9.82561946e-01 2.90096819e-01 -7.84361959e-01 -1.05816387e-01
1.43906116e-01 -1.61680788e-01 -1.49489552e-01 -8.16076577e-01
-5.04268527e-01 8.49901319e-01 -8.03018883e-02 2.67380774e-01
7.76203573e-01 1.60525456e-01 4.92810458e-01 3.53228509e-01
2.80161262e-01 -9.15340245e-01 -4.52740006e-02 8.05904448e-01
1.13588989e+00 -1.12830603e+00 -3.18768136e-02 -9.25701261e-02
-7.23363876e-01 9.35965419e-01 5.15663445e-01 2.34258249e-01
4.22365099e-01 7.31604323e-02 -2.00978648e-02 -1.62568968e-02
-1.35070288e+00 -9.90118161e-02 1.27422646e-01 1.68341056e-01
9.53084528e-01 -1.38242235e-02 -5.18342137e-01 4.40206110e-01
-7.52955019e-01 -4.70879495e-01 5.92379212e-01 5.47398388e-01
-3.21924448e-01 -1.10619724e+00 -3.77719291e-02 5.35537720e-01
-2.48183832e-01 -6.67959094e-01 -3.43616277e-01 9.11214948e-01
1.55221662e-02 1.10043180e+00 3.29180807e-01 -3.63478512e-01
1.85310021e-01 3.28303754e-01 4.98910099e-01 -9.50549185e-01
-6.86030209e-01 -1.32052600e-01 7.57850468e-01 -2.67320186e-01
-2.79921323e-01 -9.74456608e-01 -9.85394418e-01 -3.98838550e-01
-3.27125221e-01 3.47399712e-01 2.97460020e-01 1.07660329e+00
4.68579322e-01 6.28428936e-01 5.32035232e-01 -3.06696206e-01
-6.86755061e-01 -1.17793477e+00 -3.88859004e-01 3.77855599e-01
3.87890756e-01 -3.05886090e-01 -2.65714049e-01 1.04319178e-01] | [11.525938987731934, 9.132550239562988] |
fda500cc-af4e-4c95-8beb-9f3396739def | trahgr-few-shot-learning-for-hand-gesture | 2203.16336 | null | https://arxiv.org/abs/2203.16336v2 | https://arxiv.org/pdf/2203.16336v2.pdf | TraHGR: Transformer for Hand Gesture Recognition via ElectroMyography | Deep learning-based Hand Gesture Recognition (HGR) via surface Electromyogram (sEMG) signals has recently shown significant potential for development of advanced myoelectric-controlled prosthesis. Existing deep learning approaches, typically, include only one model as such can hardly maintain acceptable generalization performance in changing scenarios. In this paper, we aim to address this challenge by capitalizing on the recent advances of hybrid models and transformers. In other words, we propose a hybrid framework based on the transformer architecture, which is a relatively new and revolutionizing deep learning model. The proposed hybrid architecture, referred to as the Transformer for Hand Gesture Recognition (TraHGR), consists of two parallel paths followed by a linear layer that acts as a fusion center to integrate the advantage of each module and provide robustness over different scenarios. We evaluated the proposed architecture TraHGR based on the commonly used second Ninapro dataset, referred to as the DB2. The sEMG signals in the DB2 dataset are measured in the real-life conditions from 40 healthy users, each performing 49 gestures. We have conducted extensive set of experiments to test and validate the proposed TraHGR architecture, and have compared its achievable accuracy with more than five recently proposed HGR classification algorithms over the same dataset. We have also compared the results of the proposed TraHGR architecture with each individual path and demonstrated the distinguishing power of the proposed hybrid architecture. The recognition accuracies of the proposed TraHGR architecture are 86.18%, 88.91%, 81.44%, and 93.84%, which are 2.48%, 5.12%, 8.82%, and 4.30% higher than the state-ofthe-art performance for DB2 (49 gestures), DB2-B (17 gestures), DB2-C (23 gestures), and DB2-D (9 gestures), respectively. | ['Arash Mohammadi', 'Amir Asif', 'Elahe Rahimian', 'Soheil Zabihi'] | 2022-03-28 | null | null | null | null | ['hand-gesture-recognition', 'gesture-recognition'] | ['computer-vision', 'computer-vision'] | [ 8.59606266e-02 -1.28580704e-01 1.65732354e-02 5.70487268e-02
-6.65144086e-01 -2.28355467e-01 4.02008295e-01 -7.84392118e-01
-5.54827988e-01 7.07555950e-01 8.80073309e-02 1.01604275e-02
-3.61405224e-01 -3.32446694e-01 -5.01693666e-01 -9.68333662e-01
-3.42757225e-01 2.91992635e-01 -5.47007620e-02 -1.70646027e-01
7.14931041e-02 5.02836287e-01 -1.41498721e+00 2.99798280e-01
7.15506256e-01 1.16118872e+00 1.76398396e-01 5.11041939e-01
2.09484234e-01 3.16010475e-01 -7.36816883e-01 -1.58984736e-02
3.31831902e-01 -2.80988514e-01 -4.46667731e-01 -4.28355426e-01
5.34694269e-02 -6.83383882e-01 -5.14206588e-01 6.56054974e-01
1.19308496e+00 9.89139453e-02 6.33591175e-01 -9.24628973e-01
-3.15148234e-01 7.82659233e-01 -5.40777743e-01 -5.35457358e-02
2.89078206e-01 4.12296981e-01 5.74061930e-01 -5.52589834e-01
6.68626189e-01 1.03244793e+00 7.14049637e-01 1.11393523e+00
-8.31288874e-01 -9.79008496e-01 7.80699626e-02 3.13678175e-01
-1.44922018e+00 -4.95072901e-02 8.97841573e-01 -4.65548456e-01
1.17847931e+00 4.30954695e-01 8.16752255e-01 1.64772153e+00
1.05775945e-01 1.01047611e+00 1.25245345e+00 -3.47029328e-01
7.08789974e-02 -1.80138394e-01 4.11794990e-01 2.73909926e-01
3.58221263e-01 2.45426446e-01 -5.17267466e-01 2.14997098e-01
9.40255880e-01 -5.76667441e-03 -6.33221030e-01 1.74156725e-01
-1.06350553e+00 1.32180229e-01 5.49763680e-01 8.71753871e-01
-7.19771206e-01 1.67084038e-01 3.38868797e-01 2.27477729e-01
-1.90292031e-01 8.91769081e-02 -2.36336783e-01 -6.36318445e-01
-6.43573344e-01 2.63476878e-01 8.64461899e-01 8.70614529e-01
-1.81722805e-01 2.92370439e-01 -1.42819837e-01 9.04466212e-01
4.44904953e-01 5.21477401e-01 1.00660551e+00 -2.57873923e-01
6.34861767e-01 5.14797270e-01 1.92568954e-02 -8.42574298e-01
-7.54647970e-01 -5.18631399e-01 -9.65124726e-01 3.99245322e-01
4.67023164e-01 -3.07528675e-01 -1.15779638e+00 1.72090018e+00
-1.56885490e-01 -9.47405398e-02 9.71987247e-02 1.27093613e+00
1.06417882e+00 3.06357712e-01 1.63297072e-01 -6.46536122e-04
9.55847859e-01 -4.39472795e-01 -7.32396901e-01 2.33520910e-01
1.74864665e-01 -4.25148219e-01 1.11318839e+00 8.97027373e-01
-8.11051846e-01 -6.94168210e-01 -1.15185547e+00 4.41513389e-01
-8.61361474e-02 5.12169480e-01 4.61129844e-01 8.12902629e-01
-8.48751187e-01 7.72297561e-01 -1.14652801e+00 -4.97519732e-01
1.04612656e-01 7.73748457e-01 -3.70336384e-01 4.66570437e-01
-1.07112360e+00 8.00122678e-01 4.11911458e-01 7.65907645e-01
-7.12162852e-01 -1.97146282e-01 -1.22437328e-01 -8.09784457e-02
-1.16349319e-02 -1.51559070e-01 8.89232576e-01 -4.74152893e-01
-1.97312975e+00 5.61483085e-01 1.91272929e-01 -1.18542932e-01
8.97513986e-01 -6.31018937e-01 -5.15218794e-01 -8.62720311e-02
-6.56379700e-01 2.08141103e-01 5.09179771e-01 -1.06920958e+00
-3.69682014e-01 -8.08355391e-01 -1.69804797e-01 1.34107709e-01
-4.72586274e-01 -1.58156067e-01 -7.40208179e-02 -7.94882476e-01
3.06105673e-01 -9.82210577e-01 3.12404126e-01 -3.53250176e-01
-4.02018994e-01 -2.99695224e-01 7.21140623e-01 -1.08642137e+00
1.23051071e+00 -2.00069451e+00 3.74420017e-01 3.92490178e-01
9.27907974e-02 7.59160399e-01 -1.39571637e-01 2.85519898e-01
2.18818747e-02 -1.63680464e-01 1.37568749e-02 8.90535191e-02
5.12634255e-02 8.24341699e-02 -2.30558794e-02 2.04649761e-01
-1.39384359e-01 8.91530454e-01 -5.08959055e-01 2.16163299e-03
4.27526832e-01 6.15908742e-01 -2.35653058e-01 3.67994964e-01
3.00651163e-01 6.51596606e-01 -5.10209501e-01 8.38321209e-01
4.99032021e-01 2.36663774e-01 4.25825477e-01 -4.72979397e-01
-3.98036130e-02 1.57179832e-02 -1.37140262e+00 1.56119025e+00
-3.36624533e-01 6.85763776e-01 -1.65252164e-02 -9.08489704e-01
1.23682892e+00 6.86115563e-01 6.89486504e-01 -6.93597257e-01
6.10280097e-01 5.96131086e-01 4.56821203e-01 -8.94201696e-01
-9.18717235e-02 -1.32058412e-01 3.48975152e-01 2.57867277e-01
2.44896233e-01 5.49713790e-01 -2.87743390e-01 -4.45412874e-01
1.03368676e+00 4.27306324e-01 1.22503728e-01 -1.07558101e-01
5.43406427e-01 -3.83168787e-01 3.06734294e-01 7.61835992e-01
-3.27805728e-01 5.34967840e-01 5.83785065e-02 -2.75136977e-01
-6.15662515e-01 -1.11363447e+00 2.95183286e-02 4.90598291e-01
1.41332105e-01 4.83591259e-02 -6.15957379e-01 -3.90391022e-01
3.07737309e-02 2.29542032e-01 -4.11503017e-01 -3.74537194e-03
-1.06354797e+00 -8.85596633e-01 9.21798766e-01 1.06304657e+00
9.39244092e-01 -1.33078349e+00 -8.23720038e-01 3.15668762e-01
-1.07378893e-01 -8.28952730e-01 9.59774330e-02 9.72099975e-02
-9.09452915e-01 -1.17614901e+00 -1.14591181e+00 -7.75030196e-01
1.17824748e-01 -3.12446624e-01 3.46327335e-01 -1.61594167e-01
-1.64648667e-01 4.08017933e-01 -7.10850060e-01 -2.59883285e-01
-1.38752714e-01 2.49297559e-01 3.11671674e-01 3.46891060e-02
4.10332471e-01 -8.41460049e-01 -4.77601141e-01 3.63404989e-01
-3.81128460e-01 -2.60825723e-01 9.82702672e-01 1.03541112e+00
2.37264782e-01 -4.83088046e-01 7.66946971e-01 -2.12393776e-01
8.74941826e-01 -3.33001278e-02 -1.91530868e-01 3.94892544e-01
-5.61367095e-01 -1.03282139e-01 3.96528095e-01 -8.13298821e-01
-1.13881040e+00 5.03338687e-03 -5.35745919e-01 -3.95177811e-01
-2.04565451e-01 3.56031030e-01 -4.80425805e-01 -2.63604820e-01
5.01554549e-01 5.22135556e-01 -5.71830980e-02 -9.48959231e-01
8.69324878e-02 1.53545487e+00 7.27883160e-01 -6.23339117e-01
4.98670757e-01 7.12497765e-03 -1.87650949e-01 -9.84400332e-01
3.97211879e-01 -2.84771711e-01 -6.31429136e-01 -5.72458506e-01
6.75329804e-01 -5.91426194e-01 -1.26312971e+00 1.29010606e+00
-1.07222962e+00 -4.59104687e-01 9.63826627e-02 1.02113450e+00
-5.18367887e-01 2.09470019e-01 -5.71252763e-01 -1.18579042e+00
-9.49853301e-01 -1.13853133e+00 8.52166593e-01 3.60790104e-01
-3.02162945e-01 -3.86549473e-01 -1.17290050e-01 1.89562306e-01
7.13921964e-01 6.00296080e-01 5.94476044e-01 -7.93943405e-01
-2.02668130e-01 -3.67380291e-01 -2.23018825e-02 4.02080864e-01
2.55057901e-01 -4.20063883e-01 -1.13485551e+00 -5.23117959e-01
8.43419041e-03 -2.88226217e-01 4.69233245e-01 1.98854104e-01
8.66481602e-01 -5.53461127e-02 -3.05244356e-01 3.74735743e-01
1.26293600e+00 1.07342529e+00 9.47615325e-01 2.60658920e-01
7.43666768e-01 2.94273108e-01 1.64752066e-01 1.72450408e-01
-1.55491844e-01 9.95254338e-01 1.80359334e-01 2.58832723e-01
-2.93184429e-01 9.54978541e-03 4.19853002e-01 9.46424603e-01
-9.81558561e-01 -1.59065172e-01 -8.22318316e-01 2.71676600e-01
-1.62548220e+00 -7.71704018e-01 -2.45506689e-03 1.96618009e+00
4.98867720e-01 -6.30074739e-03 2.77595490e-01 5.44992626e-01
5.59558034e-01 -3.16246971e-02 -6.78196013e-01 -2.23512217e-01
5.86481169e-02 8.17634344e-01 3.16593975e-01 1.28406197e-01
-1.05543590e+00 5.98502994e-01 5.37668467e+00 5.92973530e-01
-1.53925729e+00 1.24874413e-02 -3.03420871e-01 -1.94183990e-01
4.20072496e-01 -7.25536108e-01 -6.38314068e-01 4.70886886e-01
5.71758449e-01 1.20389342e-01 6.08975410e-01 8.09105754e-01
1.89652383e-01 3.82923186e-01 -1.01564479e+00 1.35940528e+00
4.54588756e-02 -7.14242995e-01 9.35864598e-02 2.51629874e-02
2.80200779e-01 -1.38282195e-01 -6.16984628e-02 4.03837830e-01
-4.11617965e-01 -1.01666880e+00 4.74370599e-01 5.56275070e-01
8.55057359e-01 -2.58320808e-01 9.89910960e-01 2.55868852e-01
-1.36447322e+00 -2.34395713e-01 2.07727462e-01 -9.30332765e-02
9.42400247e-02 -2.92028755e-01 -4.71219242e-01 8.23592365e-01
7.79277802e-01 5.98809361e-01 -1.39825195e-01 8.84986758e-01
-2.73298919e-01 5.80954075e-01 -4.08403367e-01 -5.36012232e-01
8.52719024e-02 1.35992795e-01 8.75977039e-01 1.01768982e+00
3.16280514e-01 3.53477985e-01 -4.41981107e-01 8.19332421e-01
1.27009392e-01 4.20931727e-02 -2.98957974e-01 -1.39945313e-01
3.11010301e-01 9.00196970e-01 -2.67804384e-01 -2.00762495e-01
-6.25392348e-02 9.63757455e-01 -1.00623071e-01 3.79915416e-01
-6.64403379e-01 -7.29875565e-01 6.24497414e-01 -3.14548910e-01
7.58834705e-02 -2.84971476e-01 -3.59383851e-01 -1.25836873e+00
5.38655162e-01 -1.02026486e+00 2.07620576e-01 -5.32820761e-01
-1.13832748e+00 8.35541248e-01 -3.09450999e-02 -1.41686463e+00
-4.95538980e-01 -1.30216539e+00 -5.85550308e-01 1.08052337e+00
-8.26545060e-01 -1.09383047e+00 -8.38103712e-01 6.48601294e-01
3.62111539e-01 -3.61651331e-01 8.75600338e-01 6.36942089e-01
-5.63313186e-01 1.18908978e+00 1.50129125e-01 2.44777724e-01
2.47717828e-01 -1.02895725e+00 1.25717670e-01 7.41273284e-01
-1.32808492e-01 9.79168415e-01 2.61472225e-01 -5.41292071e-01
-1.63707292e+00 -5.64686477e-01 2.40916982e-01 -5.48171811e-02
1.72816917e-01 -2.73563504e-01 -8.61371458e-01 6.68207705e-01
-1.33160353e-01 -4.56487954e-01 3.02443653e-01 -1.22940883e-01
-8.81566256e-02 -2.60025591e-01 -1.36567688e+00 4.63925332e-01
1.53081965e+00 -2.15412170e-01 -7.84714282e-01 -2.42479026e-01
1.14149731e-02 -5.25903165e-01 -1.15523291e+00 6.97760463e-01
1.64337432e+00 -5.65291226e-01 6.36480272e-01 -6.23333395e-01
-9.26265940e-02 -4.12843376e-02 -4.05528724e-01 -1.17385221e+00
-1.28091807e-02 -4.57135767e-01 -2.78836995e-01 1.09550893e+00
5.01259714e-02 -8.84658754e-01 8.80979180e-01 6.03311718e-01
-1.82914108e-01 -8.80298376e-01 -1.09552622e+00 -1.18393195e+00
-8.03414956e-02 -3.97527575e-01 5.54318488e-01 5.75913966e-01
3.75786692e-01 -6.83259368e-02 -5.77477217e-01 -4.34557982e-02
5.21312058e-01 -1.21390998e-01 8.38529944e-01 -1.23521471e+00
-4.50719476e-01 -6.44358099e-01 -8.07512283e-01 -1.02302802e+00
-3.67670506e-01 -6.72545969e-01 -2.19763666e-02 -1.57129419e+00
-1.27300378e-02 -3.83820951e-01 -5.92752993e-01 5.39866865e-01
1.32545814e-01 2.27142587e-01 4.30781484e-01 3.08455557e-01
3.24037999e-01 4.54610825e-01 1.12970805e+00 -2.25610003e-01
-6.46577001e-01 6.38644248e-02 -1.98058605e-01 4.76238668e-01
8.45748842e-01 9.03553590e-02 -1.49999589e-01 -3.98069084e-01
-7.62044430e-01 -8.12137946e-02 2.97102153e-01 -1.39354408e+00
-3.01285945e-02 2.58926958e-01 4.68514442e-01 -4.13058639e-01
4.66208309e-01 -8.78930092e-01 5.09578109e-01 9.21237886e-01
-1.93889603e-01 -3.91773134e-01 1.29021779e-01 1.11590341e-01
-7.28582963e-02 3.15532058e-01 5.88982761e-01 1.03337295e-01
-9.17171717e-01 -4.10028845e-02 -2.46509537e-01 -5.10299504e-01
9.32799518e-01 -5.21140695e-01 -1.72658682e-01 -1.74213275e-01
-1.16812170e+00 -7.75904134e-02 -4.05535758e-01 7.12589145e-01
7.38015831e-01 -1.32039547e+00 -4.43036854e-01 3.41113001e-01
-9.09612477e-02 -4.62096035e-01 3.98520559e-01 9.05212343e-01
-3.53016615e-01 6.16753817e-01 -8.35956752e-01 -4.33568448e-01
-1.33467209e+00 -3.67660522e-02 6.51768327e-01 -1.98513076e-01
-9.83987570e-01 5.64225316e-01 -3.16781551e-01 -2.61146337e-01
7.74589002e-01 -5.32731771e-01 -4.29190338e-01 -1.57586619e-01
4.33455974e-01 8.34969580e-01 1.82556033e-01 -5.50731003e-01
-6.89895272e-01 9.14641619e-01 1.17341191e-01 -2.21311361e-01
1.31188643e+00 4.44796175e-01 4.14417267e-01 3.62169087e-01
9.77124631e-01 -4.33958441e-01 -8.19045484e-01 1.28585100e-01
-2.62961853e-02 -2.72027493e-01 -3.69167656e-01 -1.46387744e+00
-1.20911896e+00 1.08125401e+00 1.45780277e+00 -2.82756805e-01
1.38152659e+00 -4.06152427e-01 1.02444100e+00 4.18931127e-01
8.07573915e-01 -7.53423274e-01 -1.42568871e-01 3.75831947e-02
1.21370506e+00 -6.98113441e-01 -4.83191878e-01 5.78945801e-02
-2.96937793e-01 1.30094719e+00 8.03962946e-01 -3.50897402e-01
4.90687162e-01 2.26850897e-01 1.95597932e-01 -2.53731627e-02
-8.69695395e-02 -3.95201370e-02 3.76508027e-01 7.09702671e-01
4.33234990e-01 4.46086168e-01 -1.21861541e+00 1.22063744e+00
-6.68897182e-02 6.65945888e-01 2.51334477e-02 1.03154767e+00
-8.59153345e-02 -1.01559925e+00 -2.86138028e-01 5.49939275e-01
-3.07245612e-01 5.04658639e-01 -4.13008481e-01 1.33637226e+00
1.20712638e-01 8.23200881e-01 -4.11216795e-01 -1.37459886e+00
7.64790833e-01 2.91330665e-01 8.48279655e-01 3.74077037e-02
-7.85374820e-01 1.11391492e-01 -1.25732839e-01 -6.32481039e-01
-4.48118985e-01 -4.26080257e-01 -1.20286632e+00 3.17931101e-02
-2.05684155e-01 -1.09459296e-01 7.51472533e-01 9.83823776e-01
2.07907096e-01 5.72491646e-01 1.76067248e-01 -1.05083036e+00
-1.03239751e+00 -1.68799722e+00 -8.97784770e-01 3.11074436e-01
1.01205483e-01 -1.04914129e+00 -2.09055528e-01 -3.04930508e-01] | [6.837733268737793, 0.12616433203220367] |
70e37572-3613-47bc-9bf5-e1e97dcf0e74 | cine-cardiac-mri-motion-artifact-reduction | 2006.12700 | null | https://arxiv.org/abs/2006.12700v1 | https://arxiv.org/pdf/2006.12700v1.pdf | Cine Cardiac MRI Motion Artifact Reduction Using a Recurrent Neural Network | Cine cardiac magnetic resonance imaging (MRI) is widely used for diagnosis of cardiac diseases thanks to its ability to present cardiovascular features in excellent contrast. As compared to computed tomography (CT), MRI, however, requires a long scan time, which inevitably induces motion artifacts and causes patients' discomfort. Thus, there has been a strong clinical motivation to develop techniques to reduce both the scan time and motion artifacts. Given its successful applications in other medical imaging tasks such as MRI super-resolution and CT metal artifact reduction, deep learning is a promising approach for cardiac MRI motion artifact reduction. In this paper, we propose a recurrent neural network to simultaneously extract both spatial and temporal features from under-sampled, motion-blurred cine cardiac images for improved image quality. The experimental results demonstrate substantially improved image quality on two clinical test datasets. Also, our method enables data-driven frame interpolation at an enhanced temporal resolution. Compared with existing methods, our deep learning approach gives a superior performance in terms of structural similarity (SSIM) and peak signal-to-noise ratio (PSNR). | ['Debiao Li', 'Qing Lyu', 'Hongming Shan', 'Ge Wang', 'Yibin Xie'] | 2020-06-23 | null | null | null | null | ['metal-artifact-reduction'] | ['medical'] | [ 4.59343404e-01 -4.79442656e-01 3.08563281e-02 -2.48373926e-01
-1.02700055e+00 -5.02584204e-02 1.18282743e-01 -7.28987083e-02
-4.52979654e-01 5.89149952e-01 2.99688697e-01 -1.09850526e-01
-1.82923034e-01 -3.64974350e-01 -1.13982186e-01 -9.35407937e-01
-3.88602406e-01 -3.82902250e-02 2.36394092e-01 2.35479251e-01
3.72305922e-02 4.60679621e-01 -7.93801367e-01 1.10225946e-01
8.95609021e-01 9.14502680e-01 5.56194723e-01 4.90538120e-01
2.69128889e-01 1.17465782e+00 -2.83883780e-01 1.21552065e-01
1.19604953e-01 -8.71417761e-01 -8.63384426e-01 8.86023939e-02
1.25965089e-01 -5.57202518e-01 -6.90892637e-01 1.10017848e+00
9.70638096e-01 2.07135573e-01 1.70026809e-01 -4.18460369e-01
-4.71077919e-01 4.47467029e-01 -9.32337284e-01 9.25701916e-01
-6.25493079e-02 1.83903202e-01 1.84292287e-01 -9.25965190e-01
6.28541470e-01 5.03197789e-01 7.31571555e-01 5.15079796e-01
-1.37267673e+00 -5.08256435e-01 -4.00530517e-01 3.68133962e-01
-1.12451458e+00 -2.65232384e-01 9.42933679e-01 -3.15345347e-01
4.16437477e-01 2.58030117e-01 6.10472381e-01 7.90617168e-01
5.68926394e-01 6.31150901e-01 1.32396472e+00 -2.23711446e-01
-8.09779391e-03 -4.02184814e-01 -2.75804281e-01 4.12968487e-01
-3.12097911e-02 1.87791526e-01 -1.61097765e-01 -1.42099019e-02
1.33401620e+00 2.68563449e-01 -7.06051707e-01 -3.32270294e-01
-1.62468123e+00 4.83421654e-01 5.97902715e-01 8.00205886e-01
-6.23106003e-01 2.59855270e-01 7.15605974e-01 -7.07087014e-03
3.88445407e-01 3.09691012e-01 1.18238786e-02 -1.97757661e-01
-1.04439878e+00 -1.21070363e-01 8.22596624e-02 2.57385433e-01
-3.51550803e-02 4.91554946e-01 -3.04041266e-01 1.05118155e+00
-3.26330774e-02 4.19831663e-01 9.71996486e-01 -1.26370931e+00
1.31736442e-01 -4.00406905e-02 -3.59969027e-02 -1.07905233e+00
-4.52621609e-01 -7.05480099e-01 -1.57970142e+00 1.72959939e-01
2.21872553e-01 4.56252918e-02 -7.32340097e-01 1.48507249e+00
2.47953549e-01 4.55769598e-01 -3.04242402e-01 1.51320124e+00
7.70019591e-01 4.60429549e-01 1.95825741e-01 -9.03962016e-01
1.29385805e+00 -7.57788777e-01 -1.10890293e+00 -2.96248067e-02
4.55661654e-01 -7.24473774e-01 9.30922270e-01 2.24964574e-01
-1.45552325e+00 -5.13338149e-01 -9.17124927e-01 1.66128263e-01
6.70527697e-01 -1.30314693e-01 5.41699767e-01 4.17120934e-01
-8.64830971e-01 8.62031817e-01 -1.24751222e+00 2.13762686e-01
5.97995043e-01 2.99717039e-01 -1.51492774e-01 -1.87655568e-01
-1.20275939e+00 7.93073595e-01 -2.98159327e-02 2.60477602e-01
-7.72436082e-01 -1.01184285e+00 -7.03481674e-01 -1.30899891e-01
2.08892852e-01 -6.68685138e-01 1.13164759e+00 -6.94910705e-01
-1.44155216e+00 7.24707544e-01 -1.33663729e-01 -4.20348555e-01
5.79279006e-01 -1.78512782e-02 -5.87803125e-01 4.86353666e-01
1.50071472e-01 3.94291222e-01 9.17066991e-01 -9.20578241e-01
-2.02449515e-01 -3.54154378e-01 -4.47715223e-01 1.12289667e-01
-3.94789316e-02 3.02785128e-01 -3.30208361e-01 -1.12964070e+00
6.32461488e-01 -9.24033642e-01 -6.71869874e-01 6.12107553e-02
1.54943421e-01 4.83417034e-01 5.86810231e-01 -1.10631013e+00
1.39776695e+00 -1.99935186e+00 -9.84031800e-03 -1.18246883e-01
5.57110548e-01 4.05231059e-01 1.49225816e-01 -2.91230112e-01
-4.34838444e-01 5.25513012e-03 -5.72030663e-01 2.46933132e-01
-8.81157815e-01 -1.13790505e-01 7.44699910e-02 6.55347407e-01
-4.63207774e-02 1.00943279e+00 -9.94801223e-01 -5.99938393e-01
4.32337463e-01 7.88390398e-01 -3.30165923e-01 1.79758713e-01
5.26279449e-01 1.36356556e+00 -4.21526879e-01 3.12819004e-01
7.66827106e-01 -6.14113748e-01 3.21961999e-01 -3.83445591e-01
-7.47876093e-02 1.48540497e-01 -7.86638439e-01 1.94227779e+00
-5.86751759e-01 5.86164594e-01 2.33266745e-02 -8.59216750e-01
5.97956240e-01 7.73120463e-01 1.12530243e+00 -1.34018910e+00
2.16373101e-01 2.29653865e-01 2.44198009e-01 -7.77690291e-01
-6.53953925e-02 -5.32625675e-01 3.66594613e-01 5.09991825e-01
-6.00291967e-01 1.83896683e-02 -7.24113807e-02 -7.83962011e-02
9.40346241e-01 -1.14102714e-01 3.84138711e-02 -2.97987282e-01
6.84864700e-01 -3.63435209e-01 8.40147018e-01 4.73501235e-01
-4.76668715e-01 9.03898060e-01 -4.25816467e-03 -8.25575709e-01
-1.25165451e+00 -9.77482140e-01 -3.20728391e-01 3.45062762e-01
2.79249161e-01 4.04076539e-02 -6.31933689e-01 -3.48924279e-01
-5.96449018e-01 2.03777194e-01 -3.11283708e-01 -8.22078064e-02
-1.22857392e+00 -9.23381567e-01 2.79877126e-01 6.59443676e-01
5.78475654e-01 -9.94200945e-01 -9.92255211e-01 5.94205439e-01
-7.72523820e-01 -1.11436319e+00 -8.15285504e-01 -2.52741307e-01
-1.49079025e+00 -7.58173168e-01 -1.36203682e+00 -7.89339125e-01
5.85983694e-01 5.40362179e-01 1.05982101e+00 4.00028192e-02
-7.15976357e-01 -2.00920820e-01 -1.20775394e-01 1.08324133e-01
-4.75491703e-01 -3.48837346e-01 8.15187022e-03 -1.88611016e-01
-2.02626556e-01 -6.91328049e-01 -1.17094731e+00 3.21407855e-01
-1.15639877e+00 2.19216347e-01 5.71694493e-01 1.16294670e+00
8.52519274e-01 2.39812955e-02 4.97732401e-01 -6.99673891e-01
5.47024190e-01 -8.56088474e-02 -4.84942943e-01 3.03407311e-02
-5.55030763e-01 -6.54211268e-03 4.79963571e-01 -3.89317632e-01
-1.16179657e+00 -1.02934957e-01 -1.41623467e-01 -5.06355882e-01
1.02800056e-01 5.18351495e-01 3.91054928e-01 -2.34718353e-01
5.62933981e-01 4.59046543e-01 1.95338547e-01 -2.53437996e-01
-3.29774380e-01 1.57228023e-01 6.72811687e-01 -1.20623007e-01
5.21384299e-01 5.83866715e-01 2.46379331e-01 -6.49829686e-01
-6.50065958e-01 -2.57641047e-01 -5.18860161e-01 -3.48380685e-01
9.94845212e-01 -8.80036652e-01 -6.48876429e-01 4.18777585e-01
-1.03839922e+00 2.82627475e-02 -1.73671350e-01 9.74422276e-01
-4.03345942e-01 7.91314781e-01 -1.12861991e+00 -4.73603010e-01
-7.17408478e-01 -1.47605431e+00 4.79179889e-01 1.08225621e-01
-6.24163635e-02 -1.00755584e+00 -2.48816106e-02 4.08191234e-01
8.04579914e-01 6.08947158e-01 9.03631449e-01 1.77050740e-01
-7.07069695e-01 9.57406014e-02 -3.31530124e-01 4.40349370e-01
3.77250284e-01 -7.28075027e-01 -6.31961405e-01 -5.39561808e-01
6.60846233e-01 7.59168901e-03 5.67525566e-01 1.10874248e+00
1.50637567e+00 5.97591773e-02 6.78734258e-02 9.00568604e-01
1.47154248e+00 6.08261108e-01 7.37426877e-01 2.38568723e-01
6.71726227e-01 2.95871764e-01 4.41137165e-01 5.34780145e-01
-1.41628653e-01 6.97191298e-01 2.77761579e-01 -6.30397737e-01
-2.75676608e-01 2.15488806e-01 -1.32396176e-01 1.13436472e+00
-3.71191382e-01 4.40322936e-01 -9.71817374e-01 6.10100746e-01
-1.58000398e+00 -1.09749341e+00 -3.76863688e-01 2.30722141e+00
9.27457154e-01 -1.94451272e-01 -9.93209332e-02 1.77848876e-01
7.93834329e-01 1.49540007e-01 -7.37631440e-01 3.49163227e-02
1.82289686e-02 3.03190887e-01 4.77518201e-01 3.43069673e-01
-1.02749574e+00 1.92668930e-01 6.43627596e+00 5.94485879e-01
-1.41693664e+00 4.80409503e-01 1.08160090e+00 -1.91407368e-01
-1.39064416e-01 -3.65113914e-01 1.36183016e-02 6.09108627e-01
8.57457876e-01 -1.00464664e-01 3.86213362e-01 4.29221988e-01
7.46354640e-01 -5.01209982e-02 -7.08565533e-01 1.41937327e+00
-1.12717949e-01 -1.53573537e+00 -2.37574294e-01 1.58949960e-02
7.27663040e-01 -7.62777627e-02 2.56634980e-01 -2.59890795e-01
-3.66824120e-01 -1.10619605e+00 2.81201273e-01 3.84955943e-01
1.19074392e+00 -8.29384685e-01 7.95595169e-01 1.79419935e-01
-9.03666079e-01 1.75832018e-01 -1.30064055e-01 2.29832843e-01
4.03031945e-01 8.81730318e-01 -4.91558969e-01 4.78718281e-01
7.46972203e-01 6.31804824e-01 -1.99412823e-01 1.14080560e+00
5.96953183e-02 4.90694702e-01 2.61665165e-01 5.87845087e-01
5.99439964e-02 -3.35963577e-01 6.22231603e-01 8.68403077e-01
3.31882864e-01 5.24314106e-01 1.44780070e-01 7.40189254e-01
1.20251030e-01 -4.91929911e-02 -2.93538898e-01 3.62285167e-01
1.52541548e-01 1.13908648e+00 -7.82152355e-01 -4.19792682e-01
-3.84050667e-01 9.54789042e-01 -3.08742136e-01 3.14266831e-01
-9.40900505e-01 -1.83724090e-01 1.33404121e-01 4.66613024e-01
-6.82814941e-02 -1.65494785e-01 -3.86531383e-01 -1.24664450e+00
9.36183035e-02 -9.84062374e-01 2.34902844e-01 -5.89613080e-01
-9.49622273e-01 8.19784880e-01 -2.82165855e-01 -1.55147970e+00
-3.14449191e-01 4.45685536e-02 -3.61070365e-01 1.04301620e+00
-1.51888645e+00 -7.53147006e-01 -4.00779545e-01 4.87800628e-01
5.81612527e-01 2.01141253e-01 6.16087973e-01 7.26034403e-01
-2.67682254e-01 2.44839445e-01 6.03064187e-02 2.07104385e-01
7.13906944e-01 -1.01568735e+00 1.35233983e-01 8.96403193e-01
-3.48743826e-01 5.50873935e-01 5.80615222e-01 -4.69241828e-01
-1.13320708e+00 -9.91736352e-01 5.88784337e-01 1.19517222e-01
3.98158401e-01 3.98445487e-01 -1.17282486e+00 2.01835603e-01
1.20647155e-01 5.71029782e-01 5.13135076e-01 -6.14135861e-01
2.50333250e-01 -2.60354280e-01 -1.18001938e+00 4.58845407e-01
6.84093475e-01 -3.70862037e-01 -2.64444143e-01 1.69562578e-01
4.85336840e-01 -7.51133382e-01 -1.21991765e+00 5.78170300e-01
4.54085618e-01 -1.06194973e+00 1.07239866e+00 -1.57795683e-01
6.69947028e-01 -3.55225146e-01 4.19714808e-01 -1.15042865e+00
-7.55698502e-01 -5.75428665e-01 -2.70781051e-02 7.00607955e-01
-8.58136043e-02 -3.24194074e-01 8.03534925e-01 5.40972829e-01
-1.03390947e-01 -7.33518362e-01 -1.01938081e+00 -6.25722885e-01
-4.44235057e-02 -2.02119842e-01 1.27660796e-01 1.12871420e+00
-2.47888833e-01 1.97586734e-02 -5.87534547e-01 1.07015237e-01
9.38899577e-01 1.24175802e-01 -1.65309384e-02 -8.89010966e-01
-2.95150369e-01 -2.69637108e-01 -2.43878886e-01 -8.96207035e-01
-2.20068008e-01 -8.30203295e-01 -8.07035118e-02 -1.37562811e+00
5.24065495e-01 -4.40262556e-01 -6.44721031e-01 -6.88269213e-02
-3.09524953e-01 5.55389941e-01 6.29911758e-03 4.80781704e-01
-2.60216951e-01 2.62769997e-01 1.98091483e+00 -1.23512879e-01
-1.76546112e-01 -8.99025872e-02 -4.18669522e-01 6.96424007e-01
5.96617520e-01 -4.06439275e-01 -2.91072935e-01 -6.46039188e-01
-1.08639516e-01 9.40775454e-01 3.10413986e-01 -1.25061750e+00
8.23271796e-02 5.08210063e-02 6.82629824e-01 -4.52949464e-01
1.02131367e-01 -7.41018951e-01 3.90711010e-01 9.49633956e-01
-4.26904291e-01 2.03238145e-01 6.27020746e-02 2.75693059e-01
-5.50419986e-01 -1.17148109e-01 1.39675152e+00 -4.15118098e-01
-4.93017286e-01 5.74970424e-01 -4.90256935e-01 4.45478112e-02
8.32566559e-01 -1.73656657e-01 2.37881541e-01 -3.76828223e-01
-9.12504256e-01 -1.66181594e-01 1.64271787e-01 1.23201735e-01
1.01991582e+00 -1.41409731e+00 -8.54297698e-01 1.47141695e-01
-1.74440816e-01 -6.56030327e-02 9.47309554e-01 1.61069369e+00
-8.13542664e-01 3.09138864e-01 -2.49192998e-01 -9.36112106e-01
-1.22013819e+00 5.06635845e-01 5.19795656e-01 -6.51393890e-01
-1.04368198e+00 6.22030973e-01 3.15760404e-01 2.28015348e-01
-5.68914600e-02 -3.25639665e-01 3.93983424e-02 -4.75804090e-01
8.80713522e-01 3.76982927e-01 1.82709873e-01 -6.03779137e-01
-2.71981657e-01 6.42593443e-01 -1.83461636e-01 -2.58077625e-02
1.36045516e+00 -3.36515546e-01 -6.57468885e-02 2.09320992e-01
1.10417569e+00 -2.31301039e-01 -1.43453979e+00 -3.05904686e-01
2.40320824e-02 -7.58877516e-01 5.65310538e-01 -7.36814916e-01
-1.45251524e+00 1.09643793e+00 1.22457051e+00 -7.82174021e-02
1.43256283e+00 -4.22999829e-01 1.36911309e+00 -3.25222850e-01
3.37486625e-01 -8.25710535e-01 2.21756548e-01 1.46082401e-01
7.43593693e-01 -1.38269269e+00 1.37281850e-01 -3.24972719e-01
-7.95652747e-01 1.01364505e+00 2.47471303e-01 -1.17849179e-01
3.38040531e-01 3.89866590e-01 3.10079813e-01 1.70534272e-02
-4.14919466e-01 3.24519098e-01 8.18400010e-02 5.18597126e-01
9.15719390e-01 -1.69271022e-01 -4.60100383e-01 2.96044704e-02
4.70767856e-01 2.72633046e-01 5.36984086e-01 8.18389595e-01
-2.68986821e-01 -9.47331250e-01 -4.88930285e-01 3.92918080e-01
-1.08994007e+00 -1.94554418e-01 4.99012589e-01 3.71422708e-01
-1.23505674e-01 6.87839389e-01 -2.25567386e-01 1.61897391e-01
1.30243689e-01 -3.62105608e-01 5.95038831e-01 -3.33696812e-01
-6.13352180e-01 5.58256805e-01 -3.08775932e-01 -5.95767856e-01
-6.70469761e-01 -5.30896783e-01 -1.40569103e+00 -1.36328042e-01
-1.09058574e-01 -4.41365801e-02 6.20987058e-01 6.55613720e-01
2.14798152e-01 1.05307722e+00 9.27863479e-01 -6.55863225e-01
-2.87466437e-01 -7.12583601e-01 -6.65681660e-01 6.72624171e-01
4.09063727e-01 -2.58540124e-01 -2.43168194e-02 2.90364176e-01] | [13.65608024597168, -2.465764045715332] |
51a70ed8-782f-4ce4-9ac7-efbf7f462aee | path-neural-networks-expressive-and-accurate | 2306.05955 | null | https://arxiv.org/abs/2306.05955v1 | https://arxiv.org/pdf/2306.05955v1.pdf | Path Neural Networks: Expressive and Accurate Graph Neural Networks | Graph neural networks (GNNs) have recently become the standard approach for learning with graph-structured data. Prior work has shed light into their potential, but also their limitations. Unfortunately, it was shown that standard GNNs are limited in their expressive power. These models are no more powerful than the 1-dimensional Weisfeiler-Leman (1-WL) algorithm in terms of distinguishing non-isomorphic graphs. In this paper, we propose Path Neural Networks (PathNNs), a model that updates node representations by aggregating paths emanating from nodes. We derive three different variants of the PathNN model that aggregate single shortest paths, all shortest paths and all simple paths of length up to K. We prove that two of these variants are strictly more powerful than the 1-WL algorithm, and we experimentally validate our theoretical results. We find that PathNNs can distinguish pairs of non-isomorphic graphs that are indistinguishable by 1-WL, while our most expressive PathNN variant can even distinguish between 3-WL indistinguishable graphs. The different PathNN variants are also evaluated on graph classification and graph regression datasets, where in most cases, they outperform the baseline methods. | ['Michalis Vazirgiannis', 'Johannes Lutzeyer', 'Giannis Nikolentzos', 'Gaspard Michel'] | 2023-06-09 | null | null | null | null | ['graph-regression', 'graph-classification'] | ['graphs', 'graphs'] | [ 3.74228418e-01 4.96771365e-01 -4.75092769e-01 -1.27718270e-01
-2.05065534e-01 -9.86839533e-01 3.81130159e-01 3.17552388e-01
-1.91530645e-01 7.87255704e-01 -1.19977511e-01 -9.42959607e-01
-6.54486179e-01 -1.27677095e+00 -8.89679968e-01 -5.73297262e-01
-1.05491614e+00 8.00541222e-01 3.12025905e-01 -2.60781109e-01
-1.63710400e-01 7.99087346e-01 -1.20838118e+00 -5.51540889e-02
6.51726246e-01 6.33708954e-01 -5.06134391e-01 1.05223429e+00
-2.16276720e-01 6.84427202e-01 -3.27827722e-01 -4.84958172e-01
5.00138402e-01 -5.07854342e-01 -9.50960159e-01 -4.88534540e-01
7.28155792e-01 -2.29466602e-01 -8.79922271e-01 1.30225754e+00
1.55551836e-01 1.20754495e-01 6.35897875e-01 -1.75418377e+00
-9.32286203e-01 1.32508147e+00 -3.59040380e-01 1.49134472e-01
3.03741187e-01 -1.99227110e-01 1.51475286e+00 -2.31925055e-01
6.24572217e-01 1.29713702e+00 1.19133842e+00 6.72615469e-01
-1.30223203e+00 -6.62656486e-01 2.06424057e-01 2.23017275e-01
-1.29619896e+00 7.08106607e-02 7.26056874e-01 -1.86563000e-01
8.98566186e-01 5.50364614e-01 5.42898118e-01 1.12859213e+00
1.80753410e-01 5.70423841e-01 8.64743769e-01 -3.92499596e-01
4.46739607e-02 -3.26012254e-01 6.72308862e-01 1.12032199e+00
8.58142018e-01 3.51860493e-01 -4.83783931e-02 -2.64997125e-01
7.89607823e-01 -1.24338940e-01 -4.37241793e-01 -7.22490013e-01
-8.58487844e-01 1.03336942e+00 7.26978898e-01 5.21839857e-01
9.99390706e-02 4.64520812e-01 3.65169764e-01 7.45711684e-01
1.83053926e-01 5.70106268e-01 -1.94985375e-01 5.15540242e-01
-4.31710690e-01 7.62898251e-02 1.19096422e+00 9.59718823e-01
7.92279124e-01 1.82091877e-01 -6.67695105e-02 1.59412146e-01
-4.42623086e-02 1.30076706e-01 3.13881665e-01 -7.21814036e-01
2.80144542e-01 7.19998121e-01 -4.85501289e-01 -1.30024195e+00
-8.61653447e-01 -5.10612249e-01 -1.33220005e+00 -5.07909320e-02
5.50220609e-01 -7.84896314e-02 -9.49098885e-01 2.04858232e+00
1.00183435e-01 3.90534610e-01 2.00633593e-02 3.38836104e-01
1.03883016e+00 4.16099906e-01 -1.89366028e-01 5.60614765e-02
7.17872024e-01 -9.21912014e-01 -3.02454919e-01 1.22873761e-01
1.13204694e+00 2.21055448e-01 8.54814112e-01 1.15944438e-01
-9.99347508e-01 -2.81046838e-01 -1.11374652e+00 8.76877755e-02
-7.09659398e-01 -5.76500952e-01 1.07404613e+00 1.04944515e+00
-1.67531741e+00 1.20821345e+00 -5.71091056e-01 -4.05455530e-01
3.99285734e-01 6.04903400e-01 -7.80580938e-01 -6.65872023e-02
-1.37513924e+00 4.41426307e-01 7.91573107e-01 5.09372950e-02
-6.87325358e-01 -4.73488390e-01 -1.23171604e+00 3.02933753e-01
6.42060459e-01 -6.38916790e-01 9.71700609e-01 -9.36136067e-01
-1.02510917e+00 6.67941749e-01 3.52867156e-01 -9.54906404e-01
3.26261669e-01 5.32314897e-01 -6.36307597e-01 1.81307882e-01
-2.10997462e-01 3.84000778e-01 3.19292963e-01 -1.10462880e+00
-3.76223803e-01 -2.79852778e-01 5.80364704e-01 -8.12665224e-02
-4.40690637e-01 -3.30484688e-01 -1.32916182e-01 -4.49969679e-01
3.17833275e-02 -1.10176277e+00 -2.54798472e-01 1.28212348e-01
-1.03038955e+00 -5.16100407e-01 6.40357256e-01 -1.66224018e-01
1.16721153e+00 -1.73292267e+00 9.72743984e-03 7.40551293e-01
1.07888019e+00 3.71715456e-01 -6.94602489e-01 5.36118150e-01
-2.20038697e-01 5.97940207e-01 -4.20273721e-01 -5.27233221e-02
3.15754563e-01 5.34036219e-01 2.76195537e-02 6.10626221e-01
-1.10908553e-01 1.13674533e+00 -1.08035040e+00 -2.80126661e-01
-1.72341421e-01 2.45907143e-01 -3.37219626e-01 -1.11142077e-01
-2.29714945e-01 -9.76464823e-02 -1.19001418e-01 4.87363905e-01
5.98209023e-01 -6.84073627e-01 6.40368581e-01 1.81023851e-01
5.48347056e-01 7.47953802e-02 -1.14505875e+00 1.15629208e+00
-1.23786833e-02 7.38530517e-01 -9.88931507e-02 -1.24038863e+00
7.16699839e-01 6.48524892e-03 2.36199960e-01 -5.30570209e-01
-1.26789603e-03 4.05544825e-02 3.27702343e-01 -5.47501445e-02
4.39264208e-01 3.72197889e-02 -1.02526702e-01 6.94578648e-01
1.11553051e-01 4.51050282e-01 5.87351680e-01 5.39570689e-01
1.57579267e+00 -2.65637428e-01 4.24405277e-01 -3.09473813e-01
4.51703250e-01 -3.52672279e-01 3.42559040e-01 1.35152078e+00
-3.12184453e-01 2.56727725e-01 1.02584553e+00 -6.33309722e-01
-7.30918467e-01 -1.54021966e+00 3.92837524e-01 1.11796308e+00
2.09004119e-01 -6.12909794e-01 -5.80876648e-01 -1.08956432e+00
1.35883957e-01 4.02421385e-01 -8.84137928e-01 -4.38839376e-01
-6.59875035e-01 -3.57440442e-01 1.05403566e+00 7.11207926e-01
2.20736712e-01 -9.85738277e-01 -3.61211523e-02 -1.00976057e-01
2.88599759e-01 -9.97733533e-01 -4.42857832e-01 2.27542132e-01
-8.00627410e-01 -1.52799773e+00 -4.44122344e-01 -1.05979264e+00
7.73199856e-01 2.40443081e-01 1.34711969e+00 5.62845051e-01
7.24764541e-02 4.40411329e-01 -2.83689499e-01 -4.80668582e-02
-7.20955074e-01 4.54402179e-01 9.73397344e-02 -3.14113945e-01
1.30799100e-01 -9.33373094e-01 -2.74044961e-01 1.24298252e-01
-9.11158323e-01 -1.34119198e-01 5.67441046e-01 5.66958904e-01
5.10889769e-01 3.13152969e-01 4.89778668e-01 -1.51849747e+00
9.76010144e-01 -6.53167486e-01 -6.17887616e-01 3.57928425e-01
-9.22095895e-01 3.78991872e-01 1.23310804e+00 -3.91026616e-01
-2.93560207e-01 -2.49045476e-01 2.97296774e-02 -3.06720942e-01
-1.41499981e-01 6.91459596e-01 -1.88179333e-02 -5.74548721e-01
6.64584339e-01 1.58676818e-01 -2.05559239e-01 -1.52187422e-01
5.39610684e-01 1.54940672e-02 6.77074909e-01 -6.45876050e-01
1.12837088e+00 3.61524075e-01 6.17268682e-01 -7.15594172e-01
-4.38540906e-01 -1.17961660e-01 -5.66977322e-01 -6.18258417e-02
4.93691802e-01 -1.60581380e-01 -8.52707863e-01 3.90769601e-01
-9.14521992e-01 -6.05211556e-01 -2.71616727e-01 1.91380471e-01
-2.72412270e-01 8.85774016e-01 -9.16610301e-01 -5.43437779e-01
-4.63963181e-01 -6.33671463e-01 3.08527619e-01 6.52197599e-02
-1.47963434e-01 -1.44466209e+00 2.26689935e-01 -5.41979432e-01
3.11674058e-01 7.35226631e-01 1.56698346e+00 -1.34081054e+00
-4.07591790e-01 -1.78278372e-01 -3.51349056e-01 1.26825556e-01
2.37251446e-02 -7.78614655e-02 -4.74484116e-01 -6.04601860e-01
-6.81878269e-01 -3.00158262e-01 9.09510255e-01 2.09361851e-01
1.39761055e+00 -6.56723738e-01 -5.31276822e-01 8.07449102e-01
1.50310993e+00 3.36613990e-02 4.41968232e-01 6.97765425e-02
1.03855598e+00 3.77506733e-01 -5.26070595e-01 -2.57284105e-01
4.45158303e-01 2.99877906e-03 6.06637895e-01 5.53732328e-02
-8.14315379e-02 -4.25476134e-01 3.06030840e-01 8.06195617e-01
-1.90893933e-01 -7.86564589e-01 -8.24403167e-01 4.08451408e-01
-1.69143283e+00 -9.32343781e-01 -4.60828155e-01 2.14191031e+00
3.63628745e-01 4.28767890e-01 2.72799194e-01 3.21866751e-01
8.79983246e-01 4.23934221e-01 -6.94284618e-01 -7.65061915e-01
-3.19335163e-01 5.85797966e-01 7.39371181e-01 5.17183721e-01
-1.02280951e+00 7.21205771e-01 6.92019510e+00 5.04703343e-01
-5.88693798e-01 -3.52377862e-01 3.88946533e-01 2.69823372e-01
-5.90269864e-01 -3.47895212e-02 -2.70720690e-01 1.38566181e-01
1.16914475e+00 -4.70586598e-01 5.38847864e-01 8.63138199e-01
-6.78051531e-01 5.17122388e-01 -1.63790369e+00 6.89042747e-01
5.19990735e-02 -1.47237825e+00 3.27629119e-01 6.90720677e-02
7.97417223e-01 8.26976374e-02 -7.37385675e-02 6.27799273e-01
1.01099002e+00 -1.38727736e+00 1.81526497e-01 2.42042750e-01
7.29418039e-01 -8.65560472e-01 6.31992519e-01 2.81817138e-01
-1.58832932e+00 -4.88551818e-02 -2.47915223e-01 -8.67132396e-02
-2.37134233e-01 3.33469123e-01 -6.63188934e-01 9.49530244e-01
4.95597929e-01 5.67087233e-01 -6.53941154e-01 8.81109655e-01
-2.41673902e-01 5.10897577e-01 -3.23195785e-01 -1.84687272e-01
4.78507996e-01 -2.03070506e-01 6.09786153e-01 1.10773814e+00
1.82836026e-01 -1.04297422e-01 3.31981629e-01 9.60557759e-01
-6.44637883e-01 -3.17969710e-01 -9.77487028e-01 -4.21880931e-01
6.00711703e-01 1.14771867e+00 -9.35063124e-01 -1.99325755e-01
-2.37314656e-01 8.75515759e-01 7.64812589e-01 3.21618617e-01
-5.46460807e-01 -9.00403380e-01 4.70442384e-01 -7.56088793e-02
2.89363027e-01 -2.24263906e-01 1.99348450e-01 -7.83991873e-01
-7.91839510e-02 -7.70407200e-01 1.02331603e+00 -5.05369127e-01
-1.61365175e+00 8.81141126e-01 -3.22242044e-02 -6.44718111e-01
-2.58465230e-01 -1.00294411e+00 -7.48745084e-01 4.83331978e-01
-1.35697615e+00 -1.04158795e+00 -2.20610023e-01 8.50336432e-01
-3.45662326e-01 1.06769428e-01 1.05342066e+00 5.49759753e-02
-4.58935410e-01 1.03983545e+00 1.19213827e-01 5.97131789e-01
-9.14426520e-03 -1.66106403e+00 8.42296541e-01 8.40788782e-01
6.20866477e-01 5.36414027e-01 4.41984981e-01 -5.64998567e-01
-1.55306673e+00 -1.15380836e+00 6.82771564e-01 -2.29167923e-01
8.49343777e-01 -3.68724555e-01 -1.07636690e+00 1.20559824e+00
-9.65448990e-02 2.71937728e-01 6.53617680e-01 2.91423976e-01
-8.49927306e-01 1.13934711e-01 -1.17882693e+00 9.04551387e-01
1.70068920e+00 -6.76640630e-01 -3.46379369e-01 2.61250257e-01
1.07376933e+00 -2.09670797e-01 -8.04179668e-01 5.18399060e-01
5.57283282e-01 -1.22315001e+00 9.26148117e-01 -1.12535620e+00
8.78194049e-02 6.42472953e-02 1.52385067e-02 -1.37417531e+00
-5.83022237e-01 -8.92464876e-01 -5.54477096e-01 7.00572848e-01
4.49518859e-01 -1.10992181e+00 1.15410662e+00 7.65224248e-02
-4.26674671e-02 -7.91200399e-01 -8.89193296e-01 -1.38073552e+00
4.36064839e-01 -3.79973203e-01 9.06084120e-01 1.06832838e+00
2.39396398e-03 3.53167742e-01 -2.16878697e-01 3.82832706e-01
8.42336357e-01 1.46072894e-01 6.81541681e-01 -1.75877452e+00
-4.15677518e-01 -1.01107442e+00 -8.16818476e-01 -8.18320274e-01
7.00503409e-01 -1.52110636e+00 -3.34146082e-01 -1.80083048e+00
1.20242067e-01 -2.60437310e-01 -5.49761057e-01 5.63556731e-01
1.51151046e-01 1.16065212e-01 -1.47481650e-01 -2.69639701e-01
-6.10234499e-01 1.27415314e-01 1.02647233e+00 -3.31486225e-01
-1.85583264e-01 2.75039976e-03 -8.85982335e-01 7.04623699e-01
1.00799417e+00 -5.06248891e-01 -6.99782372e-01 -2.18473673e-01
4.71363395e-01 -2.22427770e-01 4.83118892e-01 -9.59057987e-01
3.87038618e-01 1.57313332e-01 1.55943394e-01 -4.57613915e-01
-8.75271931e-02 -6.37596369e-01 1.48419067e-01 8.12937737e-01
-6.18174791e-01 3.18673104e-01 1.05480112e-01 9.56454396e-01
3.66998643e-01 -8.43907297e-02 4.76341158e-01 -1.90679133e-01
-6.73068523e-01 8.89089823e-01 -1.42096311e-01 2.66757309e-01
8.53545189e-01 -4.68351215e-01 -6.93719387e-01 -6.53208733e-01
-6.68526232e-01 2.80105650e-01 1.85397387e-01 1.76213890e-01
5.26854277e-01 -1.47094083e+00 -6.24640167e-01 2.63044387e-01
1.50532022e-01 -1.08211990e-02 7.51162544e-02 6.32471621e-01
-5.81433296e-01 3.33053648e-01 -2.39141196e-01 -2.70660043e-01
-1.27527726e+00 8.64970207e-01 5.01303434e-01 -6.08006597e-01
-8.17510247e-01 8.78234982e-01 2.42732063e-01 -8.39623094e-01
3.80645037e-01 -4.70158249e-01 8.70637149e-02 -5.02355874e-01
3.75135899e-01 6.36278093e-01 -1.82231531e-01 -4.88122404e-01
-1.85785174e-01 4.02863622e-01 -3.26798290e-01 3.29300821e-01
1.23114991e+00 3.19821954e-01 -2.76763529e-01 2.40953982e-01
1.39359474e+00 -1.91132650e-01 -7.72289157e-01 -4.34219509e-01
2.24729851e-01 -7.90118650e-02 -3.95702034e-01 -4.16400224e-01
-1.05299866e+00 6.03198826e-01 1.39172509e-01 9.82811213e-01
9.77563679e-01 1.97335154e-01 9.44776773e-01 8.63017678e-01
3.80427837e-01 -6.17304146e-01 -1.51148841e-01 6.12566054e-01
4.56690818e-01 -8.34326148e-01 -1.52026534e-01 -3.24041843e-01
1.37464345e-01 1.29617274e+00 5.55633307e-01 -3.38018686e-01
5.79154253e-01 1.01527639e-01 -5.07513881e-01 -3.80758524e-01
-5.29239833e-01 -2.77308941e-01 3.63395333e-01 1.04751503e+00
-1.21909104e-01 3.26673299e-01 2.07998678e-02 5.62334239e-01
-3.40853274e-01 -3.84160340e-01 7.09720373e-01 6.21412933e-01
-3.49452406e-01 -9.96524036e-01 1.53628424e-01 6.79549873e-01
-1.97291598e-01 -1.57543555e-01 -1.02674210e+00 1.26510382e+00
-1.65220320e-01 7.48035252e-01 9.18029696e-02 -7.95434475e-01
1.22059286e-01 -1.97991788e-01 7.50216663e-01 -3.69599283e-01
-3.37460786e-01 -8.57619345e-01 3.77590984e-01 -7.31524408e-01
-1.63281128e-01 1.39915636e-02 -1.18812215e+00 -9.53309357e-01
-3.12704057e-01 2.02434942e-01 1.53557450e-01 7.12307990e-01
2.98025489e-01 4.34208155e-01 5.98040104e-01 -4.27268684e-01
-6.23853922e-01 -4.92146730e-01 -9.12769377e-01 2.21552670e-01
5.33869267e-01 -2.63167799e-01 -7.66568482e-01 -6.99408770e-01] | [6.863509178161621, 6.24064826965332] |
10b455b6-cd35-4f53-84f6-ff183a633d03 | on-the-compactness-efficiency-and | 1707.01992 | null | http://arxiv.org/abs/1707.01992v1 | http://arxiv.org/pdf/1707.01992v1.pdf | On the Compactness, Efficiency, and Representation of 3D Convolutional Networks: Brain Parcellation as a Pretext Task | Deep convolutional neural networks are powerful tools for learning visual
representations from images. However, designing efficient deep architectures to
analyse volumetric medical images remains challenging. This work investigates
efficient and flexible elements of modern convolutional networks such as
dilated convolution and residual connection. With these essential building
blocks, we propose a high-resolution, compact convolutional network for
volumetric image segmentation. To illustrate its efficiency of learning 3D
representation from large-scale image data, the proposed network is validated
with the challenging task of parcellating 155 neuroanatomical structures from
brain MR images. Our experiments show that the proposed network architecture
compares favourably with state-of-the-art volumetric segmentation networks
while being an order of magnitude more compact. We consider the brain
parcellation task as a pretext task for volumetric image segmentation; our
trained network potentially provides a good starting point for transfer
learning. Additionally, we show the feasibility of voxel-level uncertainty
estimation using a sampling approximation through dropout. | ['Lucas Fidon', 'Guotai Wang', 'Tom Vercauteren', 'Sebastien Ourselin', 'Wenqi Li', 'M. Jorge Cardoso'] | 2017-07-06 | null | null | null | null | ['3d-medical-imaging-segmentation', 'volumetric-medical-image-segmentation'] | ['medical', 'medical'] | [ 2.38091081e-01 8.19709957e-01 -1.72980458e-01 -5.72981775e-01
-7.77540743e-01 -1.11643687e-01 2.68409520e-01 5.26940897e-02
-7.51988292e-01 9.02053297e-01 6.19183481e-03 -4.62086529e-01
1.40308693e-01 -5.49741089e-01 -9.96365547e-01 -6.30793691e-01
-5.48786402e-01 8.10816526e-01 2.11512014e-01 1.55128837e-01
-1.34722376e-02 9.79643166e-01 -1.03797686e+00 3.79067302e-01
7.67587781e-01 1.18945980e+00 1.34191766e-01 5.03163338e-01
-2.24000096e-01 8.94415081e-01 -4.35178101e-01 -5.41577600e-02
1.36029691e-01 -2.64390528e-01 -1.31737077e+00 4.09343988e-02
4.78519619e-01 -6.12197101e-01 -2.65347719e-01 7.79947996e-01
5.59487820e-01 9.88424644e-02 9.66415882e-01 -6.66721225e-01
-7.34883666e-01 8.26856196e-01 -4.01356071e-01 7.54178643e-01
-4.15210873e-01 1.43053845e-01 7.60863006e-01 -7.06427932e-01
7.22325027e-01 1.05487311e+00 8.12131941e-01 6.34846091e-01
-1.42671466e+00 -3.20349813e-01 -1.98116288e-01 -1.03404380e-01
-1.23652244e+00 -2.24400431e-01 6.52619302e-01 -7.20997214e-01
1.13612354e+00 1.62905842e-01 8.05013418e-01 7.72826493e-01
3.70724171e-01 1.00067079e+00 1.07224369e+00 -4.31314632e-02
2.16281757e-01 -6.35362556e-03 2.50304312e-01 9.52197671e-01
1.83958218e-01 5.77601120e-02 2.39628494e-01 -8.07248354e-02
1.40045071e+00 -1.24921009e-01 -2.94540673e-01 -5.26489496e-01
-1.00375628e+00 1.09363830e+00 1.30504584e+00 4.86218184e-01
-3.47961664e-01 5.63596785e-01 5.13724446e-01 -2.31221635e-02
5.81061482e-01 4.80214775e-01 -1.61450908e-01 3.48536462e-01
-1.42030108e+00 -7.39084883e-03 3.47122550e-01 6.50927007e-01
3.38954419e-01 3.31360728e-01 -3.60115558e-01 7.67265916e-01
2.11681947e-01 7.47432336e-02 5.04484832e-01 -7.84207404e-01
9.44502652e-02 3.77641201e-01 -3.67744803e-01 -3.66346538e-01
-9.80324209e-01 -5.34338832e-01 -1.12937248e+00 4.20394868e-01
5.18182218e-01 -4.52148169e-02 -1.46317565e+00 1.23308110e+00
1.56886935e-01 -9.77934822e-02 -2.49465942e-01 9.88169312e-01
1.20793462e+00 2.97148526e-01 2.28824750e-01 -7.15874285e-02
1.19954932e+00 -8.12638938e-01 -2.94768155e-01 -9.86562967e-02
7.36668766e-01 -2.65005231e-01 7.00268507e-01 1.63735971e-01
-1.39659655e+00 -4.56675738e-01 -9.97690082e-01 -3.37286115e-01
-1.68174922e-01 1.02313377e-01 9.39683616e-01 7.24692643e-01
-1.23186898e+00 8.93151879e-01 -1.14915097e+00 9.66403037e-02
1.45129967e+00 6.13261342e-01 -1.94446489e-01 1.68383479e-01
-8.42331588e-01 1.02465880e+00 6.78124487e-01 1.20135680e-01
-1.07404613e+00 -1.24395108e+00 -8.65075946e-01 1.34633750e-01
-6.73024356e-02 -6.82882607e-01 1.07543981e+00 -9.55162168e-01
-1.25236392e+00 1.38183939e+00 3.64724994e-01 -8.49460781e-01
7.77231932e-01 1.17844395e-01 1.36111423e-01 7.02166200e-01
-2.13077441e-01 1.18458140e+00 8.47346127e-01 -1.06014836e+00
-1.09161753e-02 -2.85538644e-01 -1.96849659e-01 -2.32403621e-01
2.37448767e-01 -7.57079720e-02 -1.42719569e-02 -6.34770393e-01
2.55251527e-02 -5.49170136e-01 -5.54439664e-01 2.95024604e-01
-4.22134787e-01 -3.32330167e-02 4.45310742e-01 -8.56105328e-01
5.05298257e-01 -1.74824917e+00 1.30260572e-01 4.16562051e-01
7.51506865e-01 1.64243951e-01 2.36541568e-03 -4.65774924e-01
-4.39987928e-01 1.74745634e-01 -6.69324219e-01 -2.49046624e-01
-3.77975315e-01 1.15105594e-02 1.62651777e-01 7.82980025e-01
5.20705402e-01 1.51880765e+00 -7.07565486e-01 -6.38473809e-01
3.78269136e-01 7.95968175e-01 -6.58138275e-01 2.23978058e-01
-1.77614719e-01 9.47074354e-01 -3.45360398e-01 5.43616831e-01
6.11944318e-01 -5.40536463e-01 -1.96056459e-02 -2.53358096e-01
1.48044154e-01 -6.61855936e-02 -4.52176332e-01 1.80052173e+00
-4.09371704e-01 7.03919709e-01 2.11848974e-01 -1.63916564e+00
8.50776315e-01 2.83516765e-01 7.45038092e-01 -7.38979459e-01
7.01738060e-01 2.55155623e-01 1.35030240e-01 -3.79781365e-01
9.02912989e-02 -4.79266971e-01 1.88941523e-01 4.00615185e-01
5.16652822e-01 -4.31167275e-01 4.74092588e-02 9.80785638e-02
9.12002146e-01 -1.20588563e-01 1.22449122e-01 -6.49563670e-01
4.05867517e-01 -1.94710717e-01 2.33498719e-02 5.86469650e-01
-4.04789716e-01 8.69366944e-01 8.14952910e-01 -8.48068178e-01
-1.43465543e+00 -1.28203440e+00 -6.05665028e-01 5.77005267e-01
-3.88719797e-01 2.59525508e-01 -9.53169286e-01 -6.55118465e-01
-1.01470627e-01 3.37352365e-01 -1.10613596e+00 1.22412138e-01
-7.77121007e-01 -8.10874164e-01 6.46143019e-01 8.92381370e-01
4.20949012e-01 -1.27388585e+00 -9.08187449e-01 2.37620234e-01
1.54889241e-01 -1.01901186e+00 -1.57686830e-01 4.64093924e-01
-1.29466224e+00 -1.20648336e+00 -1.27158070e+00 -9.94460821e-01
8.83833408e-01 -2.90884018e-01 1.46376848e+00 4.05041188e-01
-8.95040274e-01 2.52357572e-01 1.62150979e-01 -7.47740045e-02
-4.55425829e-01 3.88831049e-01 -4.15462524e-01 -5.26068330e-01
-1.10854343e-01 -8.18462729e-01 -9.07898128e-01 -1.46822602e-01
-9.73126173e-01 9.13660973e-02 7.28556216e-01 9.30849195e-01
8.75491142e-01 -5.51316440e-01 5.31924844e-01 -1.17636681e+00
5.91126025e-01 -4.82213914e-01 -5.91395319e-01 1.39567435e-01
-3.63511652e-01 2.89020091e-01 4.89212662e-01 -2.29899824e-01
-8.48285913e-01 1.69136241e-01 -4.95825887e-01 -4.03980047e-01
-1.88527495e-01 3.37404370e-01 5.16283393e-01 -3.58134776e-01
8.81170213e-01 -2.08792165e-02 3.47305208e-01 -2.63423622e-01
6.67957246e-01 2.28096440e-01 5.71790814e-01 -6.39403403e-01
2.66511649e-01 7.09588349e-01 1.94321439e-01 -8.61359358e-01
-5.26008725e-01 -2.60418177e-01 -1.13465381e+00 -2.07963839e-01
1.22461629e+00 -6.33738458e-01 -6.16646588e-01 7.70568661e-03
-1.09530115e+00 -5.89842618e-01 -5.55016458e-01 3.01428616e-01
-8.49493384e-01 2.01346472e-01 -9.06618237e-01 -2.41337761e-01
-7.37831175e-01 -1.58632994e+00 9.96887982e-01 1.32916152e-01
-1.21482968e-01 -1.21106255e+00 -4.02815342e-02 3.70772988e-01
5.82504034e-01 6.11189365e-01 1.15420103e+00 -5.89906096e-01
-5.33998728e-01 6.80177435e-02 -6.01878047e-01 4.51999843e-01
-1.77332595e-01 -3.97364914e-01 -1.10940325e+00 -2.94105679e-01
-2.04286486e-01 -8.78074765e-01 1.28955305e+00 1.05444300e+00
1.82596481e+00 7.66984299e-02 -2.19763905e-01 1.08227241e+00
1.24220645e+00 -2.29798436e-01 7.69699037e-01 1.87324464e-01
8.03631067e-01 4.71529067e-01 -2.54838318e-01 1.07028730e-01
-6.65110722e-02 1.60226956e-01 4.18840975e-01 -6.49509788e-01
-4.63556826e-01 4.59956706e-01 -4.48342532e-01 7.40839958e-01
-3.71467322e-01 5.41659355e-01 -9.80432451e-01 4.37526077e-01
-1.34171760e+00 -7.51210809e-01 9.77073610e-02 1.71089149e+00
8.91155362e-01 1.53912172e-01 1.45805612e-01 -2.60827035e-01
3.35527509e-01 1.62451923e-01 -5.84645867e-01 -5.89689434e-01
2.81282485e-01 8.00099611e-01 5.74394286e-01 4.07933354e-01
-1.29052389e+00 7.76748419e-01 6.80997276e+00 8.14980865e-01
-1.19139791e+00 2.91944116e-01 1.33489084e+00 -1.47054002e-01
-1.62409589e-01 -6.36570930e-01 -2.21286014e-01 7.79328495e-02
9.64061499e-01 4.16360110e-01 3.06771725e-01 7.70426571e-01
-8.48183930e-02 -5.43960743e-02 -1.21530747e+00 9.78201687e-01
-2.23324560e-02 -1.85940039e+00 -4.62811813e-02 -6.93212450e-02
8.31963837e-01 4.23127592e-01 1.89722329e-01 2.91987807e-01
1.52132481e-01 -1.75952661e+00 5.87755620e-01 4.68927026e-01
1.07647109e+00 -8.61556232e-01 6.76690221e-01 -1.37420222e-01
-9.43089187e-01 2.65881777e-01 -5.00678837e-01 4.18047428e-01
1.48452759e-01 5.30400991e-01 -8.66177142e-01 2.80064821e-01
6.73182607e-01 2.99660116e-01 -5.37977874e-01 1.25648201e+00
-1.28258262e-02 4.94345605e-01 -2.03140065e-01 2.65188992e-01
5.51325440e-01 -1.41165346e-01 -4.95108627e-02 1.41115105e+00
1.23691298e-02 9.28870961e-02 1.04872197e-01 1.50127053e+00
-5.07585645e-01 1.08352743e-01 -4.26424742e-01 1.03063555e-02
-2.13454396e-01 1.35056341e+00 -1.42788219e+00 -4.71926570e-01
-2.09510297e-01 7.05578566e-01 8.14369619e-01 2.94948250e-01
-7.65263438e-01 -1.21151470e-01 2.53581494e-01 9.61927772e-02
4.88238841e-01 -4.63953950e-02 -6.58454776e-01 -8.16377640e-01
-3.11156362e-01 -5.15632033e-01 1.98886067e-01 -6.27845705e-01
-1.15466392e+00 7.96387613e-01 1.50049282e-02 -5.63576460e-01
-2.45327055e-01 -8.53950560e-01 -5.43124437e-01 8.48387003e-01
-1.62244391e+00 -1.09684443e+00 -1.90911502e-01 5.09552956e-01
3.41640294e-01 7.52228964e-03 6.44185781e-01 1.45660460e-01
-8.02031681e-02 3.42087239e-01 1.08662201e-02 4.23737049e-01
1.18764881e-02 -1.21716297e+00 3.55490088e-01 4.39655364e-01
-6.59590811e-02 4.85222697e-01 3.26249570e-01 -5.52542031e-01
-8.41578960e-01 -1.16849363e+00 1.51529521e-01 -2.60077357e-01
5.73664546e-01 -4.11009550e-01 -1.08887899e+00 8.49349916e-01
1.16048247e-01 6.57303035e-01 5.21774232e-01 -6.04596995e-02
-1.70920104e-01 4.37348455e-01 -1.50710952e+00 3.18657637e-01
8.45528781e-01 -5.02835393e-01 -7.81421125e-01 5.53811491e-01
6.92291081e-01 -7.55372345e-01 -1.09034324e+00 2.76743650e-01
2.96736389e-01 -8.55155289e-01 1.41262710e+00 -5.54499984e-01
7.30192184e-01 3.72231275e-01 3.50594580e-01 -1.28193438e+00
-2.91917443e-01 -1.17821962e-01 -7.74331912e-02 4.88160372e-01
3.36199969e-01 -3.25480431e-01 1.10324621e+00 6.24831796e-01
-4.36788738e-01 -1.09200501e+00 -1.24588954e+00 -5.15652895e-01
8.25705171e-01 -4.05233741e-01 2.16483548e-01 7.54873693e-01
-3.13858598e-01 1.57839246e-02 1.17658386e-02 -2.84463376e-01
8.86942148e-01 -1.09753385e-01 3.92040834e-02 -1.21744275e+00
-1.21421501e-01 -8.44968975e-01 -4.85754848e-01 -7.50585914e-01
3.42962086e-01 -1.34377491e+00 -1.21477328e-01 -1.64798439e+00
3.96829665e-01 -4.12417442e-01 -3.67489398e-01 2.36578122e-01
2.50645190e-01 5.74721634e-01 4.65645036e-03 6.62833005e-02
-3.21508884e-01 4.50745195e-01 1.77843642e+00 -5.98447561e-01
-1.35994881e-01 -8.89045745e-02 -3.12578022e-01 5.99798501e-01
9.01888311e-01 -4.00954574e-01 -4.16235954e-01 -2.42551640e-01
-2.78872520e-01 9.09860246e-04 6.00416481e-01 -8.83568823e-01
-1.68117180e-01 3.33288431e-01 9.85235751e-01 -6.01299107e-01
2.16268733e-01 -6.42372727e-01 -2.77077645e-01 7.39069045e-01
-5.42020738e-01 -3.44629139e-01 4.72254366e-01 8.24493989e-02
8.22713003e-02 -3.12646776e-01 1.13514066e+00 -6.41194224e-01
-4.81205016e-01 6.12246037e-01 -3.92670155e-01 4.00939256e-01
8.34906936e-01 -2.88630962e-01 -1.21176057e-01 1.65326893e-01
-1.14450276e+00 2.53412966e-02 7.63966218e-02 -1.90483183e-02
8.86240602e-01 -1.12341261e+00 -6.42462909e-01 1.51500627e-01
-2.70108998e-01 1.99253321e-01 4.06806916e-01 9.31822300e-01
-1.05323708e+00 6.27911627e-01 -7.20485806e-01 -8.87563169e-01
-9.07233417e-01 4.81332153e-01 7.64733613e-01 -4.64452535e-01
-1.19681597e+00 1.12575138e+00 3.20365965e-01 -4.47788030e-01
2.87244141e-01 -1.01036644e+00 -2.60867625e-01 -2.17984945e-01
5.37127018e-01 -5.62018864e-02 3.08251768e-01 -4.70745504e-01
-2.80627578e-01 1.90357104e-01 -1.89446315e-01 1.36632055e-01
1.76468778e+00 3.08462411e-01 -1.27335876e-01 2.04287067e-01
1.52626872e+00 -8.59457195e-01 -1.44651341e+00 -1.39052182e-01
-2.50059497e-02 -4.97648716e-02 4.52266276e-01 -5.95482647e-01
-1.58677828e+00 1.13550222e+00 8.86547804e-01 -1.44627824e-01
7.16471076e-01 4.09208447e-01 6.19817257e-01 2.91552454e-01
2.64064163e-01 -1.02366054e+00 1.66638836e-01 2.85091847e-01
1.02202702e+00 -1.43142450e+00 1.63288981e-01 -1.52441740e-01
-3.38871896e-01 1.24930811e+00 4.62154448e-01 -5.14925003e-01
6.39716625e-01 3.41477245e-01 -1.19533598e-01 -7.12524235e-01
-3.24925855e-02 -1.12287953e-01 6.23847723e-01 9.02121663e-01
7.05645978e-01 2.12729722e-01 -1.01411007e-01 3.12302679e-01
-4.46019582e-02 4.54755165e-02 4.04338151e-01 7.52174258e-01
-5.23496151e-01 -6.06475174e-01 -1.79314613e-01 8.55466127e-01
-6.01147950e-01 -2.76262254e-01 6.08331785e-02 1.01224124e+00
-1.24859147e-01 -5.33382073e-02 2.81576753e-01 3.51740777e-01
-5.79604395e-02 2.29722820e-02 1.01682210e+00 -5.56200206e-01
-7.47001350e-01 -5.29912785e-02 -2.21077710e-01 -7.46032417e-01
-3.57570976e-01 -3.35299134e-01 -1.67872894e+00 -1.30131885e-01
1.50099874e-01 -2.18714491e-01 4.85056132e-01 1.04474854e+00
-1.01604246e-01 1.12305260e+00 4.58384603e-02 -1.34723186e+00
-4.09068972e-01 -9.16483462e-01 -6.78555310e-01 4.32395965e-01
3.37460011e-01 -7.84184098e-01 8.38441625e-02 -4.97567430e-02] | [14.518417358398438, -2.5107359886169434] |
154e19db-7129-45d7-a8ff-d0097c78d81c | constrained-multi-task-learning-for-automated | null | null | https://aclanthology.org/P16-1075 | https://aclanthology.org/P16-1075.pdf | Constrained Multi-Task Learning for Automated Essay Scoring | null | ['Ted Briscoe', 'Ronan Cummins', 'Meng Zhang'] | 2016-08-01 | null | null | null | acl-2016-8 | ['automated-essay-scoring'] | ['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] | [-7.400345802307129, 3.7519710063934326] |
398e6e48-c603-4e10-8322-4f4471858f2b | noise-to-norm-reconstruction-for-industrial | 2307.02836 | null | https://arxiv.org/abs/2307.02836v1 | https://arxiv.org/pdf/2307.02836v1.pdf | Noise-to-Norm Reconstruction for Industrial Anomaly Detection and Localization | Anomaly detection has a wide range of applications and is especially important in industrial quality inspection. Currently, many top-performing anomaly-detection models rely on feature-embedding methods. However, these methods do not perform well on datasets with large variations in object locations. Reconstruction-based methods use reconstruction errors to detect anomalies without considering positional differences between samples. In this study, a reconstruction-based method using the noise-to-norm paradigm is proposed, which avoids the invariant reconstruction of anomalous regions. Our reconstruction network is based on M-net and incorporates multiscale fusion and residual attention modules to enable end-to-end anomaly detection and localization. Experiments demonstrate that the method is effective in reconstructing anomalous regions into normal patterns and achieving accurate anomaly detection and localization. On the MPDD and VisA datasets, our proposed method achieved more competitive results than the latest methods, and it set a new state-of-the-art standard on the MPDD dataset. | ['Jun Gong', 'Ruiyan Zhuang', 'Zhiyu Sun', 'Shiqi Deng'] | 2023-07-06 | null | null | null | null | ['anomaly-detection'] | ['methodology'] | [-6.94213361e-02 -4.87210661e-01 2.86526352e-01 -2.70094723e-01
-5.47478497e-01 8.80649537e-02 2.95247585e-01 3.72560620e-01
6.22153049e-03 6.21737540e-02 -2.84680188e-01 -2.35692188e-01
-2.23087177e-01 -7.82949984e-01 -4.05007422e-01 -9.48725998e-01
-1.15347490e-01 -1.18519170e-02 4.42698210e-01 -1.45302653e-01
4.22150612e-01 7.43993402e-01 -1.58415282e+00 1.71424210e-01
9.13703263e-01 1.33288598e+00 -1.36161432e-01 2.81566441e-01
-2.46350586e-01 5.46719670e-01 -7.21246183e-01 -7.22761005e-02
2.54207969e-01 -3.78950864e-01 -2.02911094e-01 2.59541184e-01
5.12977898e-01 -2.44797826e-01 -6.06018126e-01 1.29594481e+00
5.72775662e-01 2.33227402e-01 4.41357851e-01 -1.25356758e+00
-9.13752139e-01 1.97565883e-01 -9.73788679e-01 6.03769243e-01
2.68809974e-01 2.70959675e-01 9.13337290e-01 -1.13369298e+00
4.70115468e-02 1.10068202e+00 6.92389369e-01 2.51823515e-01
-1.07136524e+00 -5.73802292e-01 4.27838504e-01 7.06620038e-01
-1.27459741e+00 -2.50387117e-02 1.17098582e+00 -3.38125587e-01
7.87989736e-01 2.66806483e-01 4.77929950e-01 1.09621060e+00
5.47338426e-01 8.42041194e-01 5.76378286e-01 -3.34885687e-01
1.99052498e-01 -3.77932072e-01 2.80515969e-01 5.73084354e-01
5.05955458e-01 -1.08232766e-01 -3.89609933e-01 -2.75113463e-01
6.43302381e-01 7.74263799e-01 -3.95629376e-01 -3.20733726e-01
-1.27641451e+00 7.77255654e-01 5.11025250e-01 3.28018874e-01
-7.29462206e-01 -8.75541195e-02 7.27727115e-01 3.68691891e-01
6.22388601e-01 2.88183391e-01 -2.99074143e-01 8.99261981e-02
-6.16782784e-01 5.91618307e-02 2.06334561e-01 4.87988889e-01
3.81084889e-01 5.27325213e-01 -2.28022829e-01 8.75858247e-01
5.94437361e-01 3.48572731e-01 5.59238374e-01 -6.63728952e-01
3.67883861e-01 9.30299759e-01 -2.24726558e-01 -1.54266977e+00
-3.63300174e-01 -6.15890622e-01 -1.13761818e+00 4.66922700e-01
2.91696787e-01 2.22752139e-01 -9.07875955e-01 1.32101798e+00
4.92321074e-01 8.71805787e-01 -7.17587620e-02 8.27462494e-01
3.86417836e-01 6.44954264e-01 -2.40934461e-01 -1.17365211e-01
1.04311097e+00 -7.57637382e-01 -1.03152943e+00 -6.47096112e-02
5.36503315e-01 -6.55753553e-01 1.03205097e+00 7.24174917e-01
-6.66617453e-01 -6.25530005e-01 -1.13940191e+00 4.24555749e-01
-4.10308123e-01 1.80726841e-01 3.35105032e-01 3.43249947e-01
-6.21639609e-01 8.09593558e-01 -1.00468838e+00 -2.70935267e-01
5.21081030e-01 -1.12326838e-01 -4.13317859e-01 -2.13331655e-01
-9.01239336e-01 5.43420851e-01 3.17697346e-01 5.62755227e-01
-7.98867106e-01 -5.12682378e-01 -9.61724579e-01 -4.43812013e-02
4.41281378e-01 -2.48988673e-01 8.71402681e-01 -4.07233626e-01
-1.35751045e+00 3.84328723e-01 1.06527582e-01 -5.57268500e-01
2.93808252e-01 -5.20585060e-01 -1.05869818e+00 8.01594034e-02
8.58369768e-02 -3.51837605e-01 1.06884074e+00 -9.11413312e-01
-7.83870339e-01 -5.98344326e-01 -4.52802956e-01 -3.23234946e-01
-4.50937837e-01 -2.45236501e-01 -1.48488268e-01 -7.82739341e-01
8.34345818e-01 -4.80421841e-01 -3.42829853e-01 2.40516230e-01
-5.99298894e-01 -3.93632740e-01 1.50963318e+00 -7.45238841e-01
1.33372879e+00 -2.43310785e+00 -2.18419641e-01 4.62592483e-01
1.55473977e-01 4.08880174e-01 -2.57661082e-02 2.99742252e-01
-2.84837931e-01 -3.78854603e-01 -5.41299164e-01 -2.53314912e-01
9.26283225e-02 3.66935469e-02 -5.52554786e-01 1.00782275e+00
3.42793405e-01 2.98410237e-01 -8.06930840e-01 -3.12560171e-01
6.88116610e-01 4.16050047e-01 -5.02556205e-01 1.93434775e-01
-3.00477934e-03 5.11931837e-01 -4.22470540e-01 9.34028983e-01
8.04571509e-01 4.78413068e-02 -3.80892128e-01 -2.10227549e-01
2.95484625e-02 -1.21152535e-01 -1.44823265e+00 1.97695816e+00
-4.13226873e-01 5.36927760e-01 1.31437164e-02 -1.20204592e+00
1.16679907e+00 8.08634162e-02 7.34359801e-01 -5.30491531e-01
-8.51383880e-02 3.90928090e-01 -3.90792564e-02 -6.28455758e-01
2.29050562e-01 4.19066936e-01 6.73903152e-02 1.05139747e-01
-4.61562537e-02 3.34538281e-01 1.04837649e-01 -8.12645182e-02
1.45162511e+00 3.10339648e-02 8.56023356e-02 -1.05981156e-01
8.15697312e-01 -3.90252322e-01 9.37845469e-01 4.96754080e-01
-3.14731568e-01 6.00961804e-01 4.45865184e-01 -7.08372414e-01
-5.64126551e-01 -1.27491093e+00 -3.34865779e-01 5.21865010e-01
2.99077541e-01 -3.42773765e-01 -3.46149266e-01 -1.15670872e+00
2.85212398e-01 9.90890563e-01 -4.03803766e-01 -6.74956024e-01
-5.30934751e-01 -9.24121559e-01 3.78704339e-01 5.61902344e-01
7.43936241e-01 -1.00098693e+00 -3.77549171e-01 3.94967914e-01
-9.38876867e-02 -7.67621636e-01 -3.40683371e-01 -4.90145758e-02
-1.04526734e+00 -1.23102176e+00 -4.37149048e-01 -5.72559297e-01
8.44625235e-01 1.77415341e-01 9.44314420e-01 3.39277834e-02
-6.85965955e-01 2.77065188e-01 -3.95286590e-01 -5.14472067e-01
-2.26319164e-01 -4.03062999e-01 2.13291869e-01 5.54222286e-01
6.10320985e-01 -6.51027977e-01 -7.24665284e-01 4.76480454e-01
-9.31940198e-01 -9.74068820e-01 7.26779163e-01 9.05326784e-01
8.79071295e-01 2.62847066e-01 7.70550489e-01 -6.44656658e-01
5.25152028e-01 -6.77500784e-01 -6.11446202e-01 -1.10524341e-01
-8.84226382e-01 -2.28208572e-01 7.19940484e-01 -2.38422617e-01
-8.52126658e-01 -1.65802464e-01 -4.88925666e-01 -8.72550130e-01
-4.79158521e-01 3.22878093e-01 -3.32360268e-01 1.23668481e-02
5.89035213e-01 3.48161548e-01 8.18822682e-02 -9.44760799e-01
-5.54968193e-02 5.42731881e-01 6.63766682e-01 -2.72963792e-01
8.08564365e-01 4.52921838e-01 1.00956582e-01 -8.96867692e-01
-7.20606387e-01 -6.61347687e-01 -4.81231779e-01 -1.00301333e-01
6.37188196e-01 -7.76521862e-01 -5.34784019e-01 7.47828603e-01
-9.89614666e-01 2.91425526e-01 -7.39422321e-01 5.70509791e-01
-1.82136178e-01 8.55120063e-01 -6.57626033e-01 -7.44003534e-01
-2.77425349e-01 -9.94397879e-01 9.69356537e-01 -9.45286825e-03
2.52749622e-01 -7.28293061e-01 1.41809195e-01 -1.70781836e-01
5.41022599e-01 5.72148502e-01 8.59430432e-01 -9.31306362e-01
-4.51456040e-01 -6.92464590e-01 -7.52783194e-03 6.95195794e-01
4.14771020e-01 9.76434276e-02 -9.34042275e-01 -4.21910763e-01
2.17596844e-01 3.12281460e-01 1.01519668e+00 3.34938318e-01
1.73088813e+00 -8.46210718e-02 -2.69759297e-01 5.90876579e-01
1.39831114e+00 1.26586020e-01 8.05959165e-01 3.88724148e-01
9.04576957e-01 3.31916004e-01 7.88555443e-01 5.70173323e-01
-6.82522655e-02 5.45896709e-01 1.07073283e+00 -6.07299395e-02
1.78048745e-01 1.44229382e-01 4.86297041e-01 9.09178257e-01
1.83873326e-01 -1.34853080e-01 -7.70822763e-01 7.16310322e-01
-2.05142903e+00 -1.16232967e+00 -3.36377203e-01 2.23933887e+00
1.22839361e-01 3.44325274e-01 1.26196304e-02 5.34508169e-01
7.57487655e-01 3.25035065e-01 -7.56716430e-01 -1.77048802e-01
-2.88761337e-03 -1.10311665e-01 9.16297883e-02 -1.86916813e-01
-1.31964910e+00 3.08325082e-01 5.81005239e+00 7.87875175e-01
-8.64106238e-01 1.04981691e-01 3.22656512e-01 9.63542387e-02
1.20532587e-01 -4.87531185e-01 -5.16966403e-01 7.77093947e-01
8.01442862e-01 4.20364529e-01 -2.00273648e-01 1.17622066e+00
8.93477127e-02 2.75044888e-01 -9.05009151e-01 1.01043844e+00
1.54505685e-01 -1.08191645e+00 -8.40134770e-02 -1.30968884e-01
5.17384231e-01 -7.05426931e-02 9.64362770e-02 3.50176752e-01
-9.16409940e-02 -5.61847508e-01 3.69763345e-01 5.84212959e-01
2.51385480e-01 -9.48450685e-01 1.16255271e+00 7.25891665e-02
-1.22836196e+00 -4.40126538e-01 -5.78133702e-01 2.65324026e-01
1.40700251e-01 1.18992269e+00 -3.35214823e-01 7.60214031e-01
1.02752674e+00 1.12576735e+00 -5.15245497e-01 1.26850295e+00
-2.07220301e-01 7.69183397e-01 -2.14845389e-01 3.97275448e-01
1.03299372e-01 -4.96953130e-01 1.00115979e+00 9.96131122e-01
6.97007120e-01 -5.19682229e-01 3.97046566e-01 9.81619298e-01
2.09654450e-01 7.33775496e-02 -8.81505430e-01 3.30396533e-01
1.80611521e-01 1.21497834e+00 -5.63764334e-01 7.59574547e-02
-5.74047625e-01 1.05404043e+00 -1.27712741e-01 1.85057878e-01
-9.50431168e-01 -8.35340738e-01 1.09838808e+00 4.14374024e-02
4.58355397e-01 -1.50124773e-01 -4.60899360e-02 -1.16128910e+00
3.58145028e-01 -8.75654936e-01 6.28617465e-01 -3.09984028e-01
-1.74742103e+00 4.45653707e-01 -1.96861193e-01 -1.76959383e+00
-8.18379298e-02 -7.54696012e-01 -1.20202088e+00 5.01444340e-01
-1.47738838e+00 -8.78756702e-01 -4.23809111e-01 6.43827260e-01
7.43884742e-01 -4.63561535e-01 6.96433723e-01 6.25457466e-01
-1.16967309e+00 5.33291936e-01 3.63783181e-01 2.53890187e-01
6.31971121e-01 -1.37065148e+00 3.67357492e-01 1.51441073e+00
2.73145676e-01 4.04708534e-01 5.34321487e-01 -5.72472870e-01
-1.20372045e+00 -1.45277512e+00 1.33896932e-01 -2.12732509e-01
5.74273288e-01 1.07923478e-01 -1.41016555e+00 6.31931126e-01
7.75624588e-02 6.60174489e-01 4.51534539e-01 -3.32631133e-02
-2.08726451e-01 -4.32222605e-01 -1.41244841e+00 4.32801574e-01
9.23774123e-01 -2.92619288e-01 -5.19137800e-01 2.97681451e-01
6.94603384e-01 -4.16347563e-01 -8.16117525e-01 6.08787954e-01
-1.30913645e-01 -9.40583408e-01 1.10203695e+00 -4.03260469e-01
1.24271795e-01 -7.76116014e-01 -1.81590736e-01 -1.38378167e+00
-5.50194681e-01 -3.15449625e-01 -7.05248117e-01 1.35612416e+00
1.23746805e-01 -1.05871284e+00 4.59820569e-01 -7.59175122e-02
-6.77730381e-01 -9.71167922e-01 -1.07953501e+00 -8.56056273e-01
-3.37398022e-01 -6.42455518e-01 7.44257331e-01 9.34284151e-01
-3.97006124e-01 -1.58622086e-01 -9.53805819e-02 7.32895136e-01
7.94157445e-01 2.20250040e-01 4.21848625e-01 -1.51551151e+00
-3.87372486e-02 -3.07085365e-01 -1.06858695e+00 -5.95271707e-01
1.56137019e-01 -8.13526511e-01 6.15159683e-02 -1.37347603e+00
-4.92576480e-01 -6.67609274e-02 -1.03621495e+00 3.90153170e-01
-2.14877784e-01 3.18330884e-01 -2.84939885e-01 8.19051787e-02
-6.75705433e-01 7.99521983e-01 7.84553468e-01 -3.64362538e-01
-1.48495913e-01 2.56282777e-01 -3.27674478e-01 9.33450162e-01
9.54422772e-01 -3.78326654e-01 -2.64660567e-01 -1.78104684e-01
-5.03081381e-02 -5.09355128e-01 5.71158826e-01 -1.47205710e+00
7.97968805e-02 1.51757509e-01 5.44800580e-01 -1.01685739e+00
9.54167247e-02 -1.27811849e+00 -2.95042694e-01 5.07932842e-01
-1.76419318e-02 5.71779609e-01 2.21520916e-01 1.25943291e+00
-4.53557283e-01 -9.25796852e-02 7.98604906e-01 1.75965279e-01
-1.08049691e+00 4.68531102e-01 -3.19585502e-01 -1.51153907e-01
1.34233153e+00 -1.19687743e-01 -1.52647853e-01 -2.92698983e-02
-5.54960489e-01 1.47193179e-01 1.99778929e-01 5.74995518e-01
1.15867066e+00 -1.70508134e+00 -7.27448761e-01 7.61437833e-01
3.50565642e-01 3.14953715e-01 4.15555894e-01 1.08770382e+00
-5.41363716e-01 -1.13855503e-01 -1.21739656e-01 -1.10170734e+00
-1.01042438e+00 6.35908306e-01 2.84100890e-01 -1.80554479e-01
-1.18364573e+00 5.72919965e-01 -1.59088880e-01 -4.44755197e-01
4.02893990e-01 -3.74805778e-01 -1.78186312e-01 -1.84623674e-01
6.72473192e-01 7.78262973e-01 2.74173349e-01 -4.92268741e-01
-4.33961093e-01 6.66012168e-01 -9.21593010e-02 4.72358197e-01
1.31260145e+00 -4.90538403e-02 -1.48386806e-01 6.58290803e-01
1.04282165e+00 -5.72216474e-02 -1.18159783e+00 -3.85986179e-01
1.31645545e-01 -8.31688821e-01 2.91857660e-01 -4.20628220e-01
-1.67762768e+00 1.00842655e+00 1.13985360e+00 3.69120747e-01
1.28801823e+00 -3.89492333e-01 1.03290081e+00 4.76819277e-01
2.10131258e-01 -1.03702295e+00 3.59854251e-01 4.02958810e-01
8.21454287e-01 -1.46248078e+00 3.57329883e-02 -1.80679888e-01
-3.99793863e-01 1.18994439e+00 8.79828513e-01 -2.67054260e-01
7.99919426e-01 -4.11526859e-02 9.71285701e-02 -3.75145912e-01
-3.40302944e-01 8.82468894e-02 4.55595076e-01 5.59835315e-01
1.32352620e-01 -2.49029204e-01 9.13673714e-02 3.78032029e-01
5.48714817e-01 -5.83187640e-01 2.12597057e-01 9.45116520e-01
-4.25093889e-01 -7.91437089e-01 -5.31160772e-01 7.38053620e-01
-6.80141330e-01 3.21143299e-01 1.73117220e-01 6.88418388e-01
1.70839682e-01 1.05555177e+00 2.93497741e-01 -5.41695237e-01
6.63379014e-01 2.26752490e-01 -6.01404607e-02 -3.27420473e-01
-3.10224116e-01 -1.45764709e-01 -4.40349728e-01 -1.10257256e+00
-7.53167272e-02 -7.72655547e-01 -1.22760069e+00 -1.10719807e-01
-6.30226791e-01 -3.99412438e-02 5.94534099e-01 7.29190588e-01
7.20664322e-01 1.03808880e+00 9.23922062e-01 -8.09393704e-01
-8.65211368e-01 -9.37346220e-01 -9.44001973e-01 7.03427792e-01
5.62128901e-01 -7.29573607e-01 -7.15892315e-01 -3.70372981e-01] | [7.545282363891602, 2.0608460903167725] |
5932251a-0154-4c72-8eb6-951d78c9d965 | flow-per-instance-personalized-federated | 2211.15281 | null | https://arxiv.org/abs/2211.15281v1 | https://arxiv.org/pdf/2211.15281v1.pdf | Flow: Per-Instance Personalized Federated Learning Through Dynamic Routing | Personalization in Federated Learning (FL) aims to modify a collaboratively trained global model according to each client. Current approaches to personalization in FL are at a coarse granularity, i.e. all the input instances of a client use the same personalized model. This ignores the fact that some instances are more accurately handled by the global model due to better generalizability. To address this challenge, this work proposes Flow, a fine-grained stateless personalized FL approach. Flow creates dynamic personalized models by learning a routing mechanism that determines whether an input instance prefers the local parameters or its global counterpart. Thus, Flow introduces per-instance routing in addition to leveraging per-client personalization to improve accuracies at each client. Further, Flow is stateless which makes it unnecessary for a client to retain its personalized state across FL rounds. This makes Flow practical for large-scale FL settings and friendly to newly joined clients. Evaluations on Stackoverflow, Reddit, and EMNIST datasets demonstrate the superiority in prediction accuracy of Flow over state-of-the-art non-personalized and only per-client personalized approaches to FL. | ['Hui Guan', 'Sunav Choudhary', 'Kunjal Panchal'] | 2022-11-28 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [-3.86300117e-01 -3.15952450e-02 -5.94769716e-01 -7.51554549e-01
-6.72330499e-01 -7.36420453e-01 4.04443413e-01 -1.43522888e-01
-2.24024445e-01 7.33632982e-01 4.05070394e-01 -1.67940453e-01
-2.01110974e-01 -6.54807627e-01 -6.81190729e-01 -6.69802368e-01
-1.45127410e-02 7.23601639e-01 5.35736978e-01 1.63229525e-01
-5.80732748e-02 4.81338739e-01 -1.60133970e+00 8.64052355e-01
6.31325006e-01 7.51762569e-01 9.96326506e-02 6.20256066e-01
-4.52586204e-01 6.90826535e-01 -5.20584285e-01 -5.90998411e-01
3.52129698e-01 -9.12168771e-02 -1.07114911e+00 3.34005393e-02
8.97091925e-01 -5.21925211e-01 -2.08976910e-01 5.88752568e-01
3.62917602e-01 3.44979107e-01 -3.29672769e-02 -1.36572576e+00
-6.64910436e-01 1.39496922e+00 -2.81981379e-01 3.61500055e-01
-1.91806063e-01 2.43479997e-01 1.31927609e+00 -4.01696473e-01
4.41525787e-01 1.13623393e+00 6.78452849e-01 8.06069732e-01
-1.54933786e+00 -4.21379417e-01 9.03227866e-01 1.89447731e-01
-8.69974077e-01 -6.82316303e-01 2.77404249e-01 -1.21182591e-01
1.02714193e+00 6.60557210e-01 2.29131430e-01 8.54137003e-01
-9.40252692e-02 7.72698283e-01 7.86641836e-01 -1.82597078e-02
2.74005234e-01 3.91034812e-01 5.04591465e-01 3.75218302e-01
1.19632088e-01 -3.52171153e-01 -8.53868127e-01 -5.32771468e-01
3.33701193e-01 4.00752693e-01 -3.98739547e-01 -5.19243777e-01
-8.55998456e-01 5.35389364e-01 1.77530348e-01 2.61290789e-01
-4.97635484e-01 2.33496994e-01 5.22006869e-01 3.79542857e-01
3.77882123e-01 4.93802011e-01 -1.28374302e+00 -4.30040747e-01
-1.00978923e+00 2.72401810e-01 1.19653881e+00 1.02935910e+00
1.28752351e+00 -2.64000714e-01 -5.58092654e-01 6.26639426e-01
2.01762319e-01 -9.87401158e-02 8.47204387e-01 -1.37182724e+00
3.70478541e-01 8.34063351e-01 -1.11145250e-01 -1.32665247e-01
-1.56524658e-01 -7.92488337e-01 -5.23144081e-02 4.34171893e-02
1.79408282e-01 -1.89949960e-01 -3.81456733e-01 1.89125192e+00
6.37039959e-01 3.45184386e-01 -1.81148559e-01 5.79829693e-01
3.31353217e-01 3.80739093e-01 2.95912564e-01 -1.66777074e-01
9.44865942e-01 -1.59519100e+00 -1.05284438e-01 -1.00650582e-02
6.52534723e-01 -5.05444348e-01 1.25492406e+00 2.35291958e-01
-9.58185911e-01 -1.84810326e-01 -4.54403967e-01 2.94217169e-01
-1.19949728e-01 -4.91240144e-01 8.23713958e-01 8.04333448e-01
-1.17596745e+00 9.53097224e-01 -9.82427239e-01 -5.25825977e-01
5.23864388e-01 4.78931814e-01 -2.88285017e-01 1.01854160e-01
-5.79839826e-01 3.97612214e-01 2.68172979e-01 -6.32826865e-01
-7.36919820e-01 -1.47582638e+00 -1.45291016e-01 5.91833949e-01
3.38674903e-01 -8.91019344e-01 1.90269840e+00 -1.03302312e+00
-1.71801758e+00 4.26417530e-01 -3.35476130e-01 -3.51580709e-01
7.82831132e-01 -2.20234543e-01 -2.87062407e-01 -3.82007390e-01
-1.80695996e-01 1.17594473e-01 7.65386522e-01 -1.21634722e+00
-1.14439654e+00 -4.62269872e-01 4.28575337e-01 3.44032273e-02
-7.39406645e-01 -2.08125472e-01 -8.88833165e-01 -8.14020559e-02
-2.06902236e-01 -8.12741995e-01 -1.66161969e-01 -1.57433778e-01
-1.23670116e-01 -5.17264128e-01 1.09681654e+00 -1.99144796e-01
1.41196525e+00 -2.04693699e+00 -1.81697652e-01 7.65899494e-02
3.13671768e-01 4.01558816e-01 -4.17708427e-01 5.64371705e-01
1.21092647e-01 2.43686825e-01 2.75190562e-01 -8.26723576e-01
4.09939706e-01 5.16788661e-01 -2.51512647e-01 2.04284638e-01
-5.27800977e-01 6.61868930e-01 -8.03548455e-01 -3.67875189e-01
3.94773334e-02 5.16150415e-01 -1.19024897e+00 3.77043664e-01
-6.74150109e-01 4.35459346e-01 -5.78878045e-01 2.85117537e-01
7.00819075e-01 -5.97839534e-01 4.19009387e-01 -2.17082322e-01
-1.38953790e-01 6.35835171e-01 -1.21386802e+00 1.38210118e+00
-7.82675862e-01 -1.87719986e-02 2.68081307e-01 -4.31439310e-01
5.91824472e-01 2.59502143e-01 7.08018422e-01 -3.91082942e-01
-3.37922394e-01 -3.46561745e-02 -5.45672357e-01 -2.79900193e-01
5.16423523e-01 3.68043929e-01 2.39149302e-01 1.17103922e+00
-4.94413218e-03 8.23577106e-01 2.55986899e-01 3.14087778e-01
1.29247701e+00 2.83259362e-01 -8.73461738e-02 -2.19951272e-01
4.63239133e-01 -3.90300125e-01 1.13222027e+00 1.05494428e+00
-1.42552435e-01 2.46462196e-01 3.55807275e-01 -4.91734922e-01
-5.98780155e-01 -8.83919239e-01 4.81156409e-02 2.28026247e+00
-2.11649716e-01 -7.79609442e-01 -7.78845549e-01 -1.33576941e+00
3.58272493e-01 9.40580845e-01 -5.51363885e-01 8.62535015e-02
-6.32904887e-01 -5.44269264e-01 7.80001804e-02 3.33388865e-01
5.10150313e-01 -7.71643162e-01 -4.56081986e-01 5.69751322e-01
-1.55397290e-02 -7.54448354e-01 -1.05821908e+00 1.66387390e-02
-1.00066197e+00 -9.96749222e-01 -9.36821848e-02 -2.42961586e-01
4.68106598e-01 3.50954115e-01 1.32789302e+00 8.67307857e-02
1.24528967e-01 6.88017964e-01 -4.63129163e-01 1.97433978e-01
-4.12474543e-01 7.73460507e-01 -1.86420187e-01 4.09564257e-01
3.84722948e-01 -9.16381598e-01 -7.98732340e-01 6.29733622e-01
-8.47521186e-01 -1.61876976e-01 2.00480565e-01 5.49894154e-01
5.07251680e-01 -1.60694316e-01 7.01330364e-01 -1.64052558e+00
3.86545241e-01 -7.32324004e-01 -4.48516846e-01 6.95569575e-01
-1.21888459e+00 1.57485113e-01 9.74516511e-01 -4.57539588e-01
-1.58476591e+00 -7.76794404e-02 1.87403738e-01 -4.28521335e-01
1.42986104e-02 1.33118406e-01 -2.95895875e-01 5.80790304e-02
7.33888268e-01 4.16339710e-02 -1.71198383e-01 -1.13911688e+00
5.60247362e-01 7.64223516e-01 4.57271576e-01 -9.63801444e-01
5.61643660e-01 2.16852486e-01 -4.52135742e-01 -4.16731760e-02
-7.75409758e-01 -4.09329593e-01 -3.58061194e-01 2.84463149e-02
1.11670360e-01 -6.28266633e-01 -1.16233921e+00 3.15526873e-01
-8.02091479e-01 -7.42083967e-01 -5.38536191e-01 1.60606414e-01
-3.67411345e-01 2.54553854e-01 -6.46944523e-01 -2.65509158e-01
-7.27657855e-01 -1.13177156e+00 6.29204988e-01 4.74185973e-01
-2.25497305e-01 -1.12119603e+00 2.28054971e-01 5.32165051e-01
1.11900175e+00 -4.14229125e-01 7.44386256e-01 -9.26050186e-01
-8.07455897e-01 -4.40536998e-02 -5.94590157e-02 2.75301546e-01
4.04884875e-01 2.21571133e-01 -1.14769435e+00 -7.38098025e-01
-3.36492926e-01 3.83442752e-02 5.91054440e-01 -4.08639520e-01
1.50682247e+00 -8.80689025e-01 -3.80560964e-01 8.78993988e-01
1.40622032e+00 -8.56333449e-02 4.39510345e-02 4.72891390e-01
6.44568384e-01 2.31748402e-01 1.42396003e-01 9.73798871e-01
6.39780283e-01 7.38885283e-01 3.92261297e-01 4.36128914e-01
-2.75618732e-01 -2.12454781e-01 3.88539940e-01 3.56650770e-01
2.48028964e-01 -2.85651833e-01 -6.55547738e-01 5.16444325e-01
-2.06335926e+00 -8.98953676e-01 1.88411370e-01 2.29383636e+00
9.57321942e-01 -2.80564547e-01 4.13726792e-02 -4.46504682e-01
7.41603434e-01 1.18268110e-01 -1.03501356e+00 -4.20061976e-01
1.54423878e-01 1.51975900e-01 6.62754834e-01 3.16052794e-01
-6.83390915e-01 1.00140393e+00 5.52169991e+00 5.00574827e-01
-1.47664690e+00 4.88642037e-01 4.55472559e-01 -4.51348960e-01
-6.03324890e-01 3.25653329e-02 -1.24941850e+00 6.70617044e-01
1.02293301e+00 -6.29472017e-01 9.31119442e-01 1.27265525e+00
5.06048463e-02 2.01946124e-01 -1.52495611e+00 5.83083212e-01
-3.21132928e-01 -1.54075241e+00 2.56373405e-01 -1.05569892e-01
9.53476667e-01 6.71248853e-01 8.66973847e-02 5.02745032e-01
8.40030015e-01 -4.12756294e-01 5.70574224e-01 6.21241093e-01
3.73149753e-01 -7.09706783e-01 2.92969257e-01 3.59920204e-01
-9.82494473e-01 -2.78170407e-01 -1.88854471e-01 3.96108001e-01
-1.26080185e-01 4.69536543e-01 -9.88654554e-01 4.76675093e-01
1.12452543e+00 3.19459856e-01 -6.52859211e-01 1.00126755e+00
1.99771777e-01 8.84322703e-01 -4.40925598e-01 4.88742948e-01
5.55355623e-02 1.90303832e-01 3.43035907e-01 1.22829735e+00
2.74620336e-02 -2.93872386e-01 1.54939547e-01 5.52669108e-01
-4.10335273e-01 2.11712867e-01 3.84113044e-01 2.67741799e-01
9.57936645e-01 1.42285895e+00 1.52352098e-02 -4.07184273e-01
-5.80315828e-01 8.13288629e-01 7.29452729e-01 4.05603677e-01
-6.65297568e-01 1.07408576e-01 1.34386969e+00 2.48631701e-01
6.36433840e-01 2.62658238e-01 -1.15454115e-01 -1.26492548e+00
-5.66810146e-02 -9.49105203e-01 8.85508358e-01 -1.41071111e-01
-1.62807918e+00 7.96418905e-01 -3.47662181e-01 -6.46653235e-01
-1.93064362e-01 4.43769768e-02 -8.66865337e-01 9.89280939e-01
-1.63414025e+00 -1.20033944e+00 -4.33072418e-01 9.55792248e-01
3.52057964e-01 -3.84128504e-02 7.88787246e-01 4.78426754e-01
-9.04363692e-01 1.32330728e+00 4.78842676e-01 -4.77834672e-01
1.17251873e+00 -1.10116506e+00 4.35136080e-01 7.20294058e-01
-2.31065527e-01 9.21652019e-01 7.06490099e-01 -3.87115687e-01
-1.48710382e+00 -1.75428152e+00 9.45985198e-01 -5.16100824e-01
4.98122722e-01 1.42844483e-01 -1.23056388e+00 1.26208806e+00
8.48421752e-02 2.91939408e-01 8.43350649e-01 4.19144064e-01
-7.38197684e-01 -7.96603560e-01 -1.47035110e+00 4.30342942e-01
1.18580616e+00 -3.79504591e-01 -6.75372779e-02 3.14047456e-01
7.24991739e-01 -1.66499332e-01 -1.25196576e+00 1.28849804e-01
4.52764839e-01 -1.25710988e+00 4.97960985e-01 -8.09042037e-01
-2.96613634e-01 -1.67359889e-01 -2.48197943e-01 -1.31141067e+00
-6.33202016e-01 -1.15319836e+00 -5.05105615e-01 1.78614676e+00
4.57338065e-01 -1.28734303e+00 1.18776417e+00 1.38223767e+00
-2.41952151e-01 -6.99917376e-01 -7.68128932e-01 -7.52203584e-01
-1.23363495e-01 -1.76845878e-01 1.64974141e+00 9.65113461e-01
-8.18788111e-02 4.01720181e-02 -7.18319342e-02 3.01549792e-01
7.64103770e-01 4.42197919e-01 1.11944270e+00 -1.15665352e+00
-9.64224637e-01 -4.39240277e-01 2.50079334e-01 -1.18325722e+00
2.20345274e-01 -1.02303421e+00 -4.36473876e-01 -1.26482356e+00
2.75885165e-01 -1.04148316e+00 -6.09221280e-01 9.40318644e-01
-2.38686562e-01 -1.22875676e-01 3.82448852e-01 6.09519541e-01
-9.74386573e-01 3.10231239e-01 9.68912959e-01 2.20655397e-01
-5.41556716e-01 3.88332725e-01 -1.02390337e+00 2.04860911e-01
8.77286911e-01 -4.46962684e-01 -4.66816008e-01 -6.79934442e-01
-5.36082052e-02 -1.25752300e-01 -7.63998553e-02 -7.30866015e-01
3.92734259e-01 -2.94249982e-01 -2.75665941e-03 -2.36084238e-01
-1.40996441e-01 -8.32946002e-01 6.72009528e-01 -8.99086893e-02
-4.34080720e-01 -5.54663548e-03 -7.19734803e-02 3.90514314e-01
1.92779407e-01 -1.88759379e-02 9.73421156e-01 -1.90926880e-01
-5.74748099e-01 9.57468927e-01 5.98557256e-02 -5.26348054e-02
1.10262811e+00 3.29358540e-02 -7.02780485e-01 -2.00247422e-01
-7.19411671e-01 4.90421981e-01 7.85534441e-01 4.63548392e-01
-3.13539326e-01 -8.48238707e-01 -5.59915066e-01 1.59951717e-01
1.75474361e-01 -3.30906287e-02 5.88652790e-01 7.43574858e-01
-1.09279074e-01 2.64867842e-01 6.65719435e-02 -3.88026148e-01
-1.43567610e+00 5.71502924e-01 6.41356707e-01 -4.31963414e-01
-6.17649972e-01 8.58337402e-01 2.96668231e-01 -6.48181856e-01
4.54256654e-01 1.19997092e-01 1.84394881e-01 -1.58935249e-01
6.66794181e-01 7.10199296e-01 3.67372245e-01 -3.43759477e-01
-1.57052010e-01 4.76635732e-02 -6.47855937e-01 3.61482561e-01
1.52421355e+00 -4.22485292e-01 -3.46860349e-01 2.59110332e-01
1.12268043e+00 1.18201897e-01 -1.77578628e+00 -9.14288998e-01
-1.69419125e-01 -7.51584530e-01 1.23368196e-01 -1.06939900e+00
-1.60318732e+00 2.23613083e-01 2.65091658e-01 4.60151993e-02
9.77758527e-01 6.96675703e-02 8.02909732e-01 3.94363493e-01
5.63104033e-01 -8.12441349e-01 -4.26447153e-01 3.83789957e-01
4.53507632e-01 -7.29805708e-01 -1.72927946e-01 -5.69765158e-02
-5.62738240e-01 1.06375003e+00 1.03747308e+00 3.08251143e-01
5.97687662e-01 4.69672121e-02 1.60739467e-01 2.94752359e-01
-1.47046006e+00 3.51433456e-01 -2.58586586e-01 2.44946912e-01
2.69718766e-01 1.92785397e-01 -3.97919491e-02 6.61959589e-01
-9.45191383e-02 3.78375612e-02 3.15009356e-01 9.13789153e-01
-2.33274579e-01 -1.60694647e+00 -2.02036068e-01 5.97736239e-01
-6.15388215e-01 1.73058763e-01 1.96075499e-01 2.95679331e-01
-2.64745438e-03 8.42657685e-01 1.02404788e-01 -2.90940911e-01
3.50058496e-01 3.09130251e-01 1.37550071e-01 -6.73853576e-01
-1.28990912e+00 -2.74774760e-01 1.59317590e-02 -9.81797457e-01
8.64137709e-02 -7.78196156e-01 -1.12914276e+00 -1.06895900e+00
-1.11988895e-01 6.02338433e-01 6.56894803e-01 5.94530761e-01
1.02078879e+00 4.39523280e-01 1.15082097e+00 -7.43508875e-01
-8.47760260e-01 -6.99514747e-01 -5.86705267e-01 4.11154568e-01
2.62643874e-01 -2.50691503e-01 -4.93134439e-01 1.70899704e-02] | [5.821394920349121, 6.261359214782715] |
14a985c1-2c39-49ac-8f3c-9c14a4c1ec29 | provably-doubly-accelerated-federated | 2210.13277 | null | https://arxiv.org/abs/2210.13277v3 | https://arxiv.org/pdf/2210.13277v3.pdf | Provably Doubly Accelerated Federated Learning: The First Theoretically Successful Combination of Local Training and Communication Compression | In federated learning, a large number of users are involved in a global learning task, in a collaborative way. They alternate local computations and two-way communication with a distant orchestrating server. Communication, which can be slow and costly, is the main bottleneck in this setting. To reduce the communication load and therefore accelerate distributed gradient descent, two strategies are popular: 1) communicate less frequently; that is, perform several iterations of local computations between the communication rounds; and 2) communicate compressed information instead of full-dimensional vectors. We propose the first algorithm for distributed optimization and federated learning, which harnesses these two strategies jointly and converges linearly to an exact solution in the strongly convex setting, with a doubly accelerated rate: our algorithm benefits from the two acceleration mechanisms provided by local training and compression, namely a better dependency on the condition number of the functions and on the dimension of the model, respectively. | ['Ivan Agarský', 'Peter Richtárik', 'Laurent Condat'] | 2022-10-24 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-4.50230002e-01 -2.84855068e-01 -1.81483403e-01 -1.45345941e-01
-8.17614675e-01 -6.46015286e-01 3.60060990e-01 5.16821921e-01
-7.32129872e-01 6.27314270e-01 1.50662974e-01 -4.43114519e-01
-2.00508416e-01 -9.08106923e-01 -6.11032963e-01 -9.03674185e-01
-2.49935046e-01 5.97455442e-01 -9.33374986e-02 9.58235040e-02
-1.89844087e-01 4.16217834e-01 -1.30578017e+00 5.19858301e-03
6.44643486e-01 1.17148709e+00 1.09530300e-01 7.32769549e-01
-3.77481967e-01 8.85837138e-01 -3.54533553e-01 -4.61265177e-01
5.53779125e-01 -3.89656097e-01 -6.38168573e-01 5.57022206e-02
8.37176479e-03 -3.46463323e-01 -1.49810715e-02 9.15506005e-01
6.07201576e-01 3.88123721e-01 5.26603237e-02 -1.00195670e+00
2.39130631e-01 8.27724457e-01 -5.30988991e-01 -4.30059712e-03
-7.50940517e-02 7.17870444e-02 9.89985228e-01 -7.37268806e-01
3.99999410e-01 6.75981879e-01 6.81890726e-01 2.28273198e-01
-1.15533102e+00 -6.08528554e-01 2.54322916e-01 8.27500373e-02
-1.08811057e+00 -5.70088267e-01 6.18692219e-01 -1.96172073e-01
5.70682108e-01 3.97918820e-01 6.90093100e-01 5.04819393e-01
-2.36422181e-01 5.96661568e-01 9.35047925e-01 -3.27108383e-01
7.19277382e-01 2.62896657e-01 -1.55297562e-01 7.48533487e-01
7.23449588e-02 -1.63270250e-01 -7.45667696e-01 -6.53339922e-01
3.78329754e-01 3.55944514e-01 -4.08244163e-01 -3.26299787e-01
-9.19487953e-01 9.14388299e-01 3.13284248e-01 2.91284740e-01
-7.41847694e-01 2.71546036e-01 7.39538312e-01 8.93947840e-01
7.27515697e-01 -9.79648810e-03 -5.94734311e-01 -4.24082190e-01
-8.86882901e-01 1.51279330e-01 1.32355034e+00 3.88208866e-01
1.12541187e+00 -3.78110170e-01 7.50519410e-02 5.73697865e-01
1.29260467e-02 4.11244363e-01 5.54515660e-01 -7.95792937e-01
7.65138805e-01 5.09262681e-01 2.33858407e-01 -9.50603962e-01
-2.12079093e-01 -5.92883646e-01 -1.06225121e+00 1.77599445e-01
5.37891030e-01 -6.03069007e-01 1.61509141e-02 1.67267311e+00
1.08491325e+00 8.27698708e-02 -1.84669957e-01 1.05789936e+00
6.94765570e-03 5.44041812e-01 -3.46808225e-01 -3.49332184e-01
1.04033470e+00 -1.29085553e+00 -4.16808248e-01 1.05638109e-01
1.18484879e+00 -5.93475401e-01 8.80284190e-01 3.62297684e-01
-1.11712360e+00 2.69200057e-02 -7.29070306e-01 5.04149087e-02
-1.74185410e-01 4.16503325e-02 8.72055829e-01 6.29287660e-01
-9.74056363e-01 9.10452425e-01 -1.04360104e+00 2.86044270e-01
3.92730474e-01 4.00364190e-01 -4.02705669e-01 -1.30745083e-01
-5.30764401e-01 2.27047443e-01 -3.63289565e-02 2.58850008e-02
-6.74516141e-01 -8.82034957e-01 -2.62572408e-01 4.45415407e-01
3.47143650e-01 -8.61719012e-01 9.67939556e-01 -9.92092133e-01
-1.58956909e+00 3.93250644e-01 -5.63355237e-02 -4.26770806e-01
1.10375261e+00 -2.42251679e-01 4.14791584e-01 -1.59980673e-02
-2.58798033e-01 -3.59881282e-01 8.58084321e-01 -6.80391848e-01
-7.35323191e-01 -8.86314332e-01 -6.93152249e-02 3.10442597e-01
-1.00977063e+00 -1.96756750e-01 -4.71242934e-01 -2.72381008e-01
-1.67228088e-01 -7.62727559e-01 -5.24094760e-01 1.93459123e-01
-2.37261225e-02 -3.79322737e-01 7.67200053e-01 -5.21321714e-01
1.00124419e+00 -2.23088670e+00 2.78865844e-01 5.04164755e-01
5.78620553e-01 1.68756172e-01 1.68448978e-03 5.89042962e-01
3.84192050e-01 -2.01078698e-01 7.70449787e-02 -1.02750623e+00
6.85323775e-02 2.50825465e-01 -2.35065892e-01 7.92925060e-01
-7.92164445e-01 5.60661137e-01 -9.32501972e-01 -2.36368582e-01
-2.35049561e-01 2.63507485e-01 -6.40925825e-01 3.27633172e-01
-1.12913810e-01 5.66231310e-01 -7.13813186e-01 1.07610971e-02
6.86637223e-01 -5.76210499e-01 5.37321806e-01 1.25533193e-01
-2.42055357e-01 2.63300091e-01 -1.60570884e+00 1.77894688e+00
-1.08785343e+00 1.60398781e-01 9.78652894e-01 -1.21617866e+00
4.86258656e-01 4.90029186e-01 6.77995443e-01 -3.40551108e-01
8.13654885e-02 4.23994631e-01 -6.60346568e-01 -1.13879263e-01
-4.20763791e-02 1.30893692e-01 3.68281960e-01 1.28245866e+00
-1.53122351e-01 3.49264801e-01 8.09186399e-02 3.16759467e-01
1.29422021e+00 -3.82473022e-01 -9.77912843e-02 -2.71675140e-01
4.10913855e-01 -3.85894179e-01 6.59224570e-01 8.76117468e-01
1.83208257e-01 -8.40402320e-02 5.02248824e-01 -8.05648983e-01
-7.99638927e-01 -5.79340518e-01 4.26175356e-01 1.57549179e+00
-1.28437191e-01 -7.65851438e-01 -6.40669882e-01 -8.69177818e-01
1.15132213e-01 1.64700970e-01 -4.50949401e-01 1.33735552e-01
-6.14329934e-01 -4.81141984e-01 8.81517231e-02 2.26588637e-01
2.79543549e-01 -3.98986489e-01 -6.68382883e-01 3.49883020e-01
1.26114264e-01 -7.37773955e-01 -7.06891179e-01 4.57180172e-01
-1.14311373e+00 -1.05708408e+00 -7.00088978e-01 -3.66154402e-01
7.92438507e-01 5.87767184e-01 8.79466236e-01 3.40024114e-01
-1.18970245e-01 3.60280156e-01 -2.38952950e-01 -6.96672499e-02
-9.83653590e-02 9.15831923e-02 -7.39686713e-02 4.47021812e-01
-2.23357737e-01 -1.12011492e+00 -9.18933332e-01 9.81019884e-02
-9.01904523e-01 2.06190974e-01 3.26222599e-01 1.02598429e+00
4.65539068e-01 5.85756404e-03 2.51020938e-01 -9.50103939e-01
6.79579079e-01 -6.38217390e-01 -9.62873340e-01 2.16740385e-01
-7.63692379e-01 1.31382659e-01 1.28917086e+00 -3.04169804e-01
-7.30886400e-01 2.83323824e-01 1.14735611e-01 -5.73852062e-01
3.18454236e-01 5.90097249e-01 2.31649309e-01 -3.78682882e-01
5.65418780e-01 3.08927327e-01 1.91577300e-01 -7.96134055e-01
6.27358913e-01 5.82174957e-01 2.67177075e-02 -8.07462275e-01
5.59263289e-01 7.08966792e-01 4.05802131e-02 -5.70048094e-01
-5.24458647e-01 -5.67538142e-01 -1.81694493e-01 1.20079117e-02
7.08699897e-02 -8.61974657e-01 -1.01858974e+00 4.06730264e-01
-1.03196096e+00 -5.09536386e-01 -5.69697917e-01 6.31995380e-01
-1.76788166e-01 4.65176016e-01 -7.65655518e-01 -7.17301011e-01
-8.58538687e-01 -8.90610993e-01 5.84029555e-01 -2.36874130e-02
4.07280833e-01 -1.03547239e+00 2.99059600e-01 3.21697742e-01
9.61519420e-01 2.93853972e-02 6.21282756e-01 -6.89639211e-01
-5.38736999e-01 -4.53722358e-01 -1.63894191e-01 2.25150853e-01
3.76948640e-02 -5.59879184e-01 -5.33664823e-01 -7.83813894e-01
3.03008616e-01 -4.93031949e-01 5.94454467e-01 -2.36380860e-01
1.32014525e+00 -9.26016390e-01 -1.29311562e-01 8.47267270e-01
1.32073140e+00 -7.30879664e-01 -2.95067519e-01 -1.81274712e-01
5.01725316e-01 4.28525746e-01 -5.06689996e-02 1.26302505e+00
4.20742929e-01 5.15393138e-01 3.93632680e-01 -4.13320251e-02
1.04655050e-01 -1.54474959e-01 1.89931408e-01 1.26479805e+00
-2.44172394e-01 1.98250398e-01 -7.33313918e-01 3.31611961e-01
-2.13624120e+00 -5.52168965e-01 2.61501558e-02 2.54843378e+00
1.09425116e+00 -4.48004216e-01 2.49827698e-01 -1.75252538e-02
4.56062555e-01 8.83091688e-02 -6.51965141e-01 -1.46933511e-01
1.21190079e-01 3.08899343e-01 6.11279130e-01 2.72865415e-01
-6.20441973e-01 5.41801810e-01 5.30256891e+00 8.90798748e-01
-1.38512349e+00 7.25745559e-01 7.60282397e-01 -6.91593468e-01
-2.09574729e-01 1.73435330e-01 -4.39836800e-01 6.45319045e-01
9.78734076e-01 -2.61346251e-01 1.14466751e+00 1.21033800e+00
-4.48428914e-02 1.39217645e-01 -1.04945529e+00 1.25313127e+00
-4.63376522e-01 -1.56886899e+00 -3.49495858e-01 3.72722924e-01
9.98263836e-01 5.10165334e-01 -2.79525101e-01 -2.84978002e-02
6.19731307e-01 -5.99467635e-01 3.74562502e-01 1.99966013e-01
3.60086501e-01 -8.95103931e-01 4.77405995e-01 8.12810361e-01
-1.18871951e+00 -3.72927904e-01 -3.71683747e-01 -1.85849339e-01
-2.55608018e-02 1.12986660e+00 5.36876125e-03 4.66064274e-01
7.52312481e-01 1.23062760e-01 -6.81035966e-02 8.90305758e-01
3.15391161e-02 7.31811464e-01 -9.18907344e-01 -1.44140586e-01
1.32943437e-01 -6.13135815e-01 3.96718770e-01 9.91487086e-01
2.86164999e-01 -1.25753105e-01 3.82074744e-01 4.62136179e-01
-4.35128927e-01 6.23885751e-01 -1.62089318e-01 1.43987462e-01
4.62036848e-01 1.61414611e+00 -3.26898575e-01 -2.75880247e-01
-7.20252037e-01 1.06416667e+00 8.75493884e-01 2.54979253e-01
-4.82411981e-01 -4.82950598e-01 6.19168937e-01 -8.02891031e-02
6.04756057e-01 -4.79643404e-01 -1.30787775e-01 -1.29903686e+00
5.88479221e-01 -8.95435452e-01 5.62322557e-01 2.52012640e-01
-1.30283093e+00 4.35990036e-01 -7.65724897e-01 -5.77368319e-01
-1.77605361e-01 3.46639962e-03 -7.69105017e-01 7.46525526e-01
-1.43359888e+00 -8.95148337e-01 -2.95695812e-01 1.10370719e+00
-6.34971857e-02 9.58770290e-02 9.73509073e-01 5.74196219e-01
-5.16916513e-01 8.07038546e-01 7.95948327e-01 -6.67602196e-02
5.06425381e-01 -1.03088307e+00 -1.72603086e-01 5.37147164e-01
2.28095472e-01 5.36918342e-01 3.10229599e-01 -8.48194361e-02
-2.15132475e+00 -9.36716735e-01 9.27413523e-01 2.83520639e-01
8.22886646e-01 -6.24263763e-01 -6.19979799e-01 3.66831928e-01
-1.09410644e-01 4.58257318e-01 9.35679197e-01 5.02340078e-01
-4.33947712e-01 -6.14472032e-01 -1.08556759e+00 3.59381557e-01
8.23279202e-01 -5.08511722e-01 3.37555528e-01 8.83729339e-01
4.08192664e-01 -2.63959020e-01 -9.05360639e-01 -2.83272326e-01
2.70713419e-01 -9.55907047e-01 6.06000543e-01 -6.18191242e-01
-3.29685472e-02 -1.38087511e-01 9.23823565e-03 -1.27612603e+00
-4.64629196e-02 -1.33952975e+00 -7.43488729e-01 9.26230013e-01
1.63998857e-01 -1.07412052e+00 9.43777978e-01 5.93825042e-01
3.30365568e-01 -1.09664249e+00 -1.35364521e+00 -6.83891833e-01
-4.14876528e-02 -3.27483922e-01 7.39164352e-01 7.97005892e-01
4.33242060e-02 1.76639348e-01 -4.12078738e-01 -7.15628788e-02
6.22448444e-01 6.12831056e-01 1.14457440e+00 -1.02650571e+00
-9.31875527e-01 -3.64102393e-01 1.69508737e-02 -1.36872625e+00
2.63995561e-03 -8.55882823e-01 -2.64857113e-01 -8.58389080e-01
1.32726178e-01 -9.08111155e-01 -1.84298694e-01 6.50076032e-01
4.60510999e-02 -1.09224521e-01 2.90328532e-01 5.43497384e-01
-8.48459542e-01 6.34978056e-01 7.85471618e-01 1.44388214e-01
-3.46237421e-01 2.55612284e-01 -5.04473507e-01 3.80028397e-01
4.73651230e-01 -5.46987653e-01 -1.24273509e-01 -7.19274282e-01
6.43416345e-01 3.99530590e-01 1.29818797e-01 -7.94085562e-01
8.07354450e-01 -2.29439195e-02 5.71751483e-02 -5.40416315e-03
8.38009492e-02 -9.59551871e-01 -3.42911892e-02 5.91269433e-01
-2.92675495e-01 8.78456142e-03 -3.91420126e-01 7.99379110e-01
-1.15109451e-01 -1.76620651e-02 6.37213886e-01 -1.54915854e-01
2.36240372e-01 6.54911160e-01 -8.87542367e-02 -5.09723462e-02
8.72530758e-01 4.26897824e-01 1.34928320e-02 -5.57669878e-01
-6.17587924e-01 4.10824120e-01 3.71276677e-01 -4.63068150e-02
1.34156272e-01 -1.16487718e+00 -6.05693698e-01 8.39123055e-02
-4.75241363e-01 1.45843089e-01 2.53887117e-01 1.34618092e+00
-2.58197784e-01 7.39943013e-02 4.76502627e-01 -3.28867316e-01
-1.09470284e+00 3.78213465e-01 1.65868044e-01 -6.42961204e-01
-8.65441918e-01 8.78981948e-01 -2.82863259e-01 -4.26858693e-01
6.29468977e-01 2.03530401e-01 4.83504653e-01 -5.04837595e-02
8.54074299e-01 6.66070819e-01 4.88780707e-01 5.98668568e-02
-9.94411558e-02 9.72762182e-02 -7.31008127e-02 3.69112939e-02
1.64848673e+00 -2.11303875e-01 -4.91875499e-01 1.42833203e-01
1.47390151e+00 2.87352622e-01 -1.21285605e+00 -7.10964322e-01
-2.52752662e-01 -6.73859656e-01 4.32534695e-01 -4.28028464e-01
-1.59828722e+00 4.52285111e-01 5.22260725e-01 2.49484867e-01
1.12753618e+00 -2.28147581e-02 1.16778803e+00 6.87433302e-01
5.41428089e-01 -1.21591938e+00 -1.60037488e-01 3.85290354e-01
4.27713156e-01 -8.59244764e-01 1.61662158e-02 5.71499765e-03
-1.16656840e-01 1.31340396e+00 9.47759673e-02 -1.02128036e-01
7.56857276e-01 3.69841099e-01 -2.77330559e-02 -4.49375734e-02
-1.03949177e+00 2.91784197e-01 -2.14078411e-01 -7.18945563e-02
2.22133547e-01 6.73455223e-02 -4.78174418e-01 6.50166810e-01
3.04454733e-02 -3.66294906e-02 -1.72911823e-01 9.71039832e-01
-2.28039503e-01 -1.33566749e+00 -1.82404518e-01 3.71815830e-01
-5.31883121e-01 3.72491814e-02 -4.06540111e-02 5.77365272e-02
-1.47759452e-01 9.40336645e-01 -7.83139542e-02 -6.28829002e-02
-1.68974668e-01 -6.39801323e-02 1.98500231e-01 -1.59433618e-01
-1.00014436e+00 -1.05018020e-01 -2.25771487e-01 -9.48217273e-01
1.95565578e-02 -3.19720030e-01 -1.08350122e+00 -7.76057243e-01
-4.01208341e-01 7.23672509e-01 1.18935168e+00 9.45860982e-01
9.52046990e-01 -2.43066773e-02 1.33913171e+00 -8.12419474e-01
-1.34503424e+00 -5.68722606e-01 -6.12610340e-01 4.05066103e-01
2.11156547e-01 1.24171041e-01 -7.61126935e-01 -3.83480757e-01] | [6.20983362197876, 5.107778072357178] |
0813f877-921d-4692-8b37-10114d77bd64 | know-thy-corpus-robust-methods-for-digital | 2003.06389 | null | https://arxiv.org/abs/2003.06389v1 | https://arxiv.org/pdf/2003.06389v1.pdf | Know thy corpus! Robust methods for digital curation of Web corpora | This paper proposes a novel framework for digital curation of Web corpora in order to provide robust estimation of their parameters, such as their composition and the lexicon. In recent years language models pre-trained on large corpora emerged as clear winners in numerous NLP tasks, but no proper analysis of the corpora which led to their success has been conducted. The paper presents a procedure for robust frequency estimation, which helps in establishing the core lexicon for a given corpus, as well as a procedure for estimating the corpus composition via unsupervised topic models and via supervised genre classification of Web pages. The results of the digital curation study applied to several Web-derived corpora demonstrate their considerable differences. First, this concerns different frequency bursts which impact the core lexicon obtained from each corpus. Second, this concerns the kinds of texts they contain. For example, OpenWebText contains considerably more topical news and political argumentation in comparison to ukWac or Wikipedia. The tools and the results of analysis have been released. | ['Serge Sharoff'] | 2020-03-13 | know-thy-corpus-robust-methods-for-digital-1 | https://aclanthology.org/2020.lrec-1.298 | https://aclanthology.org/2020.lrec-1.298.pdf | lrec-2020-5 | ['genre-classification'] | ['computer-vision'] | [-6.93240315e-02 2.17664167e-01 -2.00862840e-01 3.17962840e-03
-1.14269722e+00 -9.96285021e-01 1.21308231e+00 7.13243067e-01
-5.62387884e-01 7.90479362e-01 6.70653999e-01 -3.34803522e-01
-4.65390831e-01 -6.30276680e-01 -5.07430315e-01 -7.23487854e-01
1.18409969e-01 7.46754587e-01 3.10207069e-01 -3.42713565e-01
5.38727939e-01 -6.73872605e-02 -1.82379985e+00 3.16031426e-01
8.13465595e-01 5.72300911e-01 5.28235316e-01 4.53919411e-01
-8.52609932e-01 3.94122809e-01 -8.17279398e-01 -4.85969841e-01
-3.14743876e-01 -3.44509304e-01 -1.02189338e+00 2.01771289e-01
1.94475561e-01 2.66277820e-01 4.16936763e-02 1.01620448e+00
1.12681121e-01 -2.88905442e-01 1.01527309e+00 -6.18519783e-01
1.02131948e-01 1.23898137e+00 -1.85692564e-01 2.94513613e-01
5.67211747e-01 -3.43897432e-01 1.21464097e+00 -4.64558065e-01
1.08991778e+00 9.99769092e-01 3.59678984e-01 -1.36577189e-01
-1.13461554e+00 -1.54440761e-01 -2.04910740e-01 -1.49320904e-02
-1.30052602e+00 -2.33053491e-01 6.29959106e-01 -6.82595134e-01
6.76276386e-01 2.93033779e-01 5.37507355e-01 1.33289242e+00
1.12523496e-01 3.54751289e-01 1.26699448e+00 -1.00429773e+00
1.92156747e-01 7.28672683e-01 3.98329496e-01 3.76922637e-02
5.11307895e-01 -6.18977427e-01 -4.94764090e-01 -4.21451837e-01
1.33470699e-01 -7.14387000e-01 -2.56015122e-01 -1.55982226e-01
-1.01817334e+00 9.04698133e-01 -4.73026156e-01 1.16197264e+00
-4.31718528e-01 -3.67733806e-01 8.77569616e-01 1.94457337e-01
8.54699969e-01 2.56087780e-01 -5.91208816e-01 -3.59523505e-01
-8.81286621e-01 3.84986252e-01 1.35150361e+00 9.05418634e-01
5.61541677e-01 -5.25644422e-01 2.90348142e-01 1.06788802e+00
3.06580633e-01 3.67267191e-01 7.55960763e-01 -5.25692940e-01
7.79649079e-01 6.64613307e-01 1.41832247e-01 -1.11185956e+00
-4.38489974e-01 -3.12348157e-01 -1.53247088e-01 -3.72601837e-01
8.52412701e-01 -5.46019971e-02 -3.95082057e-01 1.41486859e+00
4.26081657e-01 -6.35155737e-01 1.15841493e-01 2.54489988e-01
6.78092241e-01 7.32997835e-01 2.78486729e-01 -5.45274973e-01
1.67346287e+00 -1.13880448e-01 -9.34336483e-01 9.22737494e-02
6.87262416e-01 -1.15261829e+00 1.07796121e+00 6.50681496e-01
-8.29052329e-01 -2.42936105e-01 -8.56777310e-01 2.05210268e-01
-8.01412880e-01 1.01162277e-01 4.79987353e-01 7.91378975e-01
-6.41953826e-01 4.16145086e-01 -5.64103067e-01 -8.11463952e-01
6.43574353e-03 -1.59507051e-01 -1.56124383e-01 3.47238630e-01
-1.17367136e+00 9.90500271e-01 7.45590270e-01 -4.18539733e-01
-3.59379113e-01 -3.96508306e-01 -5.61186969e-01 -1.69133350e-01
5.56789219e-01 2.18008384e-01 1.10023189e+00 -9.90510583e-01
-1.22329855e+00 1.14592099e+00 1.16575934e-01 -5.11611640e-01
4.92343545e-01 -7.14498712e-03 -5.27312517e-01 2.47246727e-01
4.17260796e-01 -1.98940128e-01 6.03956282e-01 -1.15584075e+00
-1.04928529e+00 -1.84767172e-01 -1.04914725e-01 -5.37605658e-02
-6.05774224e-01 3.91675234e-01 -6.00864232e-01 -7.72030175e-01
4.85751033e-02 -5.92066884e-01 4.34138067e-02 -9.45632637e-01
-3.65885735e-01 -6.37370586e-01 2.84600049e-01 -9.71081436e-01
1.72933650e+00 -1.94356656e+00 4.23156992e-02 4.30075586e-01
1.00110918e-01 -2.84219056e-01 3.30058634e-01 9.40977991e-01
1.68066993e-01 3.48205388e-01 -4.37187925e-02 5.45703322e-02
3.31565350e-01 1.78482562e-01 -4.51872468e-01 4.94739711e-01
-2.98959076e-01 1.12996586e-01 -7.44859457e-01 -7.52298594e-01
-9.24446508e-02 3.27656895e-01 -1.95557505e-01 -2.14653254e-01
-5.24495065e-01 -9.19257775e-02 -5.87112606e-01 3.48425895e-01
2.59917319e-01 -4.46894877e-02 6.38764679e-01 -2.21802033e-02
-6.53311193e-01 9.02390838e-01 -1.04518890e+00 1.48936462e+00
-3.97912830e-01 9.35715318e-01 -9.98324901e-02 -8.91762853e-01
7.79534698e-01 6.15377724e-01 6.41496420e-01 -5.07027626e-01
4.63210821e-01 5.08977711e-01 1.21727604e-02 -5.21032214e-01
8.09127748e-01 1.15826175e-01 -3.61544758e-01 7.05879569e-01
4.13833618e-01 -1.66754331e-02 1.00857139e+00 3.82389992e-01
7.34223962e-01 1.92129269e-01 4.66903418e-01 -8.32332969e-01
7.11892426e-01 4.32116270e-01 7.80780613e-02 4.94980633e-01
2.83966959e-01 3.28457683e-01 8.32484841e-01 -2.61390597e-01
-1.16960192e+00 -6.05749011e-01 -7.32179880e-01 8.59266162e-01
-2.33351663e-01 -9.50416327e-01 -1.06869149e+00 -5.38321257e-01
-2.06944898e-01 7.20711291e-01 -5.49206257e-01 4.70919013e-01
-5.15070736e-01 -9.45583642e-01 2.84125865e-01 -3.97849739e-01
4.70003635e-02 -1.10475564e+00 -4.56735790e-01 3.92931849e-01
-7.30330169e-01 -1.06385589e+00 -7.07011819e-02 2.99223393e-01
-6.27108991e-01 -1.43975425e+00 -4.38128740e-01 -6.40169442e-01
3.89127553e-01 -1.96758658e-01 1.11700737e+00 -2.73685187e-01
-2.85087168e-01 5.43519676e-01 -7.21107721e-01 -6.79953575e-01
-1.00929892e+00 6.00784957e-01 -7.78772011e-02 -1.03971481e-01
5.78845263e-01 -4.31333750e-01 9.02094617e-02 3.67745422e-02
-1.14185464e+00 -5.25140166e-01 3.22854072e-01 2.24110812e-01
2.11870596e-01 3.51885349e-01 4.94792312e-01 -1.17656016e+00
9.80557501e-01 -7.27014422e-01 -5.16172230e-01 2.18514428e-01
-6.50494516e-01 -7.13241771e-02 3.32126886e-01 -3.98394197e-01
-1.14161122e+00 -4.20468658e-01 -1.47856951e-01 6.19459748e-01
-3.25593442e-01 9.74502087e-01 8.03016052e-02 4.41327065e-01
9.29700732e-01 2.53752172e-01 -1.92728236e-01 -9.32825625e-01
1.70504451e-02 1.04132879e+00 2.68646359e-01 -7.22794950e-01
8.32853913e-01 1.99428275e-01 -5.17181993e-01 -1.39844656e+00
-7.03279793e-01 -7.94478118e-01 -7.96603024e-01 -3.51023465e-01
6.26133561e-01 -8.17163348e-01 -2.37264618e-01 2.88945526e-01
-1.07463586e+00 -9.72615033e-02 -2.16311157e-01 5.14573216e-01
-4.50193614e-01 5.57878375e-01 -3.91678363e-01 -9.91327286e-01
-2.67149478e-01 -5.53367496e-01 5.51386833e-01 -5.03653288e-02
-7.04261482e-01 -1.22060335e+00 3.91972065e-01 3.06682944e-01
1.44028738e-01 2.74020672e-01 1.15191579e+00 -8.94078255e-01
-9.48328450e-02 -2.11240023e-01 1.97245821e-01 1.71829432e-01
1.09284624e-01 3.94180149e-01 -8.06357205e-01 -9.36261863e-02
2.47987825e-02 -1.77849337e-01 3.99264216e-01 2.64349729e-01
2.86858171e-01 -4.86270547e-01 -3.46918553e-01 -4.14820686e-02
1.66491818e+00 3.48114967e-03 5.32118261e-01 1.00980747e+00
2.31793616e-02 9.61850464e-01 5.77603757e-01 4.80786115e-01
1.95670038e-01 6.50111735e-01 -8.42160173e-03 6.05435550e-01
-4.91054244e-02 -1.70775071e-01 5.47048569e-01 1.32247150e+00
-9.15864334e-02 -2.03175977e-01 -1.03810632e+00 7.92634428e-01
-1.55854285e+00 -8.29741061e-01 -4.02129292e-01 2.22856474e+00
1.12397814e+00 5.13788581e-01 3.20820183e-01 2.43841261e-01
5.79871774e-01 1.26270384e-01 3.17978710e-01 -2.43748903e-01
-2.77568132e-01 1.27116814e-01 3.95075023e-01 4.99937117e-01
-9.50387716e-01 6.27552509e-01 6.54239893e+00 1.10595584e+00
-5.73776066e-01 1.88671559e-01 1.91942200e-01 1.25936776e-01
-3.61350596e-01 9.42636132e-02 -1.09908354e+00 7.23060250e-01
1.31191754e+00 -4.58948165e-01 2.89265513e-02 7.44651079e-01
2.20037267e-01 -6.20631814e-01 -5.29878557e-01 5.92237234e-01
2.71974355e-01 -1.16320181e+00 6.34346157e-02 3.83590609e-01
5.76121867e-01 1.66723892e-01 -4.99571532e-01 2.64826268e-01
1.90855011e-01 -1.75983295e-01 1.16459060e+00 3.18500608e-01
5.03630340e-01 -7.30322659e-01 7.69586563e-01 4.53938305e-01
-7.25959837e-01 2.03794390e-01 -4.66936439e-01 2.36581355e-01
-4.05139066e-02 7.25827932e-01 -6.30206883e-01 6.28643453e-01
6.82583451e-01 3.08204889e-01 -7.10492074e-01 8.59277725e-01
-1.69195622e-01 9.63033795e-01 -4.70106632e-01 -3.67566496e-01
1.69849843e-01 -3.99680406e-01 8.58804584e-01 1.41176331e+00
3.09455186e-01 -3.89439583e-01 3.27019878e-02 6.49131715e-01
8.40972960e-02 8.78724158e-01 -2.82330364e-01 -5.12414515e-01
3.66900891e-01 1.33609068e+00 -1.16150594e+00 -3.68653357e-01
-4.37289000e-01 3.38180780e-01 1.99863434e-01 2.61010021e-01
-5.82835972e-01 -5.10599315e-01 2.47586533e-01 4.59819674e-01
1.99287221e-01 -2.61233926e-01 -1.85442623e-02 -1.25453889e+00
4.05444093e-02 -9.96618032e-01 4.93319631e-01 -2.95651168e-01
-1.10080600e+00 6.94996774e-01 3.11113000e-01 -9.61625993e-01
-4.00559902e-01 -6.59749508e-01 -3.28568190e-01 7.36146688e-01
-1.14414752e+00 -8.22498500e-01 1.26496151e-01 3.30473751e-01
6.31642997e-01 -3.03315729e-01 8.55751097e-01 2.87345529e-01
-2.98757374e-01 5.66133447e-02 4.76520032e-01 3.95054296e-02
7.69886136e-01 -1.33800709e+00 -9.67853591e-02 6.14806473e-01
3.34967613e-01 5.96168458e-01 1.22088814e+00 -7.30811536e-01
-1.07424939e+00 -3.44962031e-01 1.54512131e+00 -5.67873359e-01
1.36090827e+00 -4.47419375e-01 -6.74914122e-01 2.58827031e-01
5.39061010e-01 -9.98387992e-01 7.53402412e-01 5.50676107e-01
-1.90011695e-01 7.03965649e-02 -6.36693180e-01 4.89044100e-01
4.78164017e-01 -6.04985833e-01 -8.31032872e-01 6.09974623e-01
4.12325323e-01 -8.81968290e-02 -9.53678191e-01 -2.16704205e-01
4.24151838e-01 -8.65862429e-01 5.27173519e-01 -2.00146377e-01
5.11546195e-01 1.28698826e-01 -1.22679502e-01 -9.60147679e-01
-9.98217091e-02 -6.81723237e-01 2.25836590e-01 1.83019340e+00
6.78769052e-01 -5.84665716e-01 5.92854261e-01 2.33132616e-01
9.75561589e-02 -2.16280460e-01 -7.94575214e-01 -5.33096611e-01
1.53551877e-01 -6.45142257e-01 1.10392012e-01 1.03390980e+00
5.48372567e-01 2.87355393e-01 1.39925286e-01 -3.22896123e-01
6.33767962e-01 -6.26242012e-02 8.02001178e-01 -1.52621210e+00
-1.97734106e-02 -6.53921723e-01 3.94798303e-03 -2.37781197e-01
-8.44089538e-02 -8.14715207e-01 -1.23673923e-01 -1.26250052e+00
2.66267031e-01 -3.42828840e-01 2.11493120e-01 -1.74454097e-02
1.58724770e-01 1.11088045e-02 -1.45748600e-01 5.35327196e-01
-2.63450712e-01 -1.02426797e-01 6.59502506e-01 -1.74112879e-02
-2.68564790e-01 -4.07624282e-02 -6.64506853e-01 8.08126986e-01
7.70472527e-01 -6.04682326e-01 -2.11927220e-01 8.53408277e-02
7.66721070e-01 -3.47953796e-01 -8.15219805e-02 -8.77240419e-01
-7.56349089e-03 1.05386376e-01 2.47816797e-02 -6.12750173e-01
-2.49354780e-01 -7.70469308e-01 1.81813776e-01 3.50494236e-01
-3.19799185e-01 -1.62087798e-01 2.44657964e-01 4.52364147e-01
-3.37842911e-01 -8.23945761e-01 3.89776677e-01 -2.16833055e-01
-3.30002755e-01 -2.16088593e-01 -8.99210751e-01 3.49231213e-01
7.14037061e-01 -6.42639250e-02 -1.21322282e-01 -2.29708746e-01
-8.58767211e-01 -2.90588886e-01 2.49384522e-01 3.44539076e-01
-1.97177097e-01 -9.63834226e-01 -8.84971082e-01 -1.65847450e-01
1.14557125e-01 -4.09201920e-01 -2.24272572e-02 8.59695971e-01
-6.38438106e-01 8.39829743e-01 -2.98036896e-02 -2.69246906e-01
-1.23472488e+00 3.12729031e-01 -2.00139210e-01 -7.03118503e-01
-5.57269335e-01 1.53929219e-01 -1.61664829e-01 -2.04684943e-01
3.54495078e-01 -2.64380872e-01 -6.95387244e-01 8.68260860e-01
4.31957364e-01 2.13939622e-01 1.81509808e-01 -7.73204327e-01
2.98137087e-02 2.79746354e-01 -1.74043253e-01 -3.91701907e-01
1.55034077e+00 -5.01773536e-01 -4.14273411e-01 9.27415967e-01
1.07928371e+00 3.49819750e-01 -6.25101030e-01 -3.80149305e-01
7.02159286e-01 -2.26307157e-02 1.71948806e-03 -7.50960290e-01
-4.27506298e-01 2.80228168e-01 -1.33212069e-02 1.03124404e+00
7.39619553e-01 1.54038966e-01 2.22312793e-01 1.93049252e-01
4.73011047e-01 -1.67322409e+00 -5.07507503e-01 6.59422278e-01
9.75707829e-01 -8.11943114e-01 2.29902819e-01 -5.32945096e-01
-2.32802272e-01 1.38591039e+00 -1.55761734e-01 1.31966010e-01
8.90560985e-01 1.23474061e-01 2.86966376e-02 -4.10778403e-01
-5.98542631e-01 -3.48191708e-01 3.41640472e-01 2.90823400e-01
6.67622745e-01 -2.19795838e-01 -1.32134449e+00 7.22606182e-01
-4.11841720e-01 -4.69397962e-01 6.73860967e-01 7.08875716e-01
-5.52943766e-01 -1.38620436e+00 -5.79537749e-01 3.72989297e-01
-1.07422829e+00 -1.01401918e-01 -4.65277076e-01 1.26285887e+00
2.19335165e-02 7.85191953e-01 -5.47171617e-03 -3.51304673e-02
7.29536414e-02 4.45897192e-01 3.86194944e-01 -5.44795394e-01
-7.63426483e-01 7.32266843e-01 6.89321816e-01 -1.11038409e-01
-6.76450968e-01 -1.12126231e+00 -8.13223124e-01 -1.17647611e-01
-6.10489905e-01 8.42674315e-01 1.06427705e+00 1.01146662e+00
-2.26830408e-01 3.35700363e-01 3.32381040e-01 -7.53211617e-01
-3.85784179e-01 -1.37980759e+00 -9.34179306e-01 4.06245679e-01
-2.63879776e-01 -3.59737098e-01 -7.53915966e-01 4.13043618e-01] | [9.85293960571289, 9.648719787597656] |
0e33eb82-778c-498f-9f98-b35acc1c5f19 | hdr-plenoxels-self-calibrating-high-dynamic | 2208.06787 | null | https://arxiv.org/abs/2208.06787v2 | https://arxiv.org/pdf/2208.06787v2.pdf | HDR-Plenoxels: Self-Calibrating High Dynamic Range Radiance Fields | We propose high dynamic range (HDR) radiance fields, HDR-Plenoxels, that learn a plenoptic function of 3D HDR radiance fields, geometry information, and varying camera settings inherent in 2D low dynamic range (LDR) images. Our voxel-based volume rendering pipeline reconstructs HDR radiance fields with only multi-view LDR images taken from varying camera settings in an end-to-end manner and has a fast convergence speed. To deal with various cameras in real-world scenarios, we introduce a tone mapping module that models the digital in-camera imaging pipeline (ISP) and disentangles radiometric settings. Our tone mapping module allows us to render by controlling the radiometric settings of each novel view. Finally, we build a multi-view dataset with varying camera conditions, which fits our problem setting. Our experiments show that HDR-Plenoxels can express detail and high-quality HDR novel views from only LDR images with various cameras. | ['Tae-Hyun Oh', 'Moon Ye-Bin', 'Kim Yu-Ji', 'Kim Jun-Seong'] | 2022-08-14 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 2.92957760e-02 -4.99446422e-01 2.37472430e-01 -3.77499968e-01
-7.89225757e-01 -8.64118636e-01 3.75900865e-01 -7.52772331e-01
-7.41242170e-02 4.09022689e-01 5.34400493e-02 -1.03927054e-01
1.58523858e-01 -1.04876292e+00 -8.10182571e-01 -4.48608339e-01
2.39958227e-01 5.40002346e-01 2.97734022e-01 -7.34487623e-02
5.88297322e-02 8.89975667e-01 -1.53898454e+00 2.53424123e-02
5.46127260e-01 8.55702102e-01 4.74725813e-01 1.41716540e+00
3.03417653e-01 1.19535983e+00 -4.19067532e-01 1.94886159e-02
9.72702265e-01 -1.01467580e-01 -3.39130670e-01 2.52576411e-01
9.50176477e-01 -1.22185099e+00 -9.24871683e-01 6.03103399e-01
4.98547196e-01 1.67421803e-01 2.61947960e-01 -8.74613643e-01
-6.16117954e-01 -2.11060137e-01 -9.62925375e-01 2.69918621e-01
8.18746150e-01 5.59507251e-01 6.53454721e-01 -7.69288778e-01
7.99527526e-01 1.10246384e+00 4.49412078e-01 2.68493325e-01
-1.40859973e+00 -6.22903109e-01 -1.46738663e-01 -2.87094086e-01
-1.27967942e+00 -2.48402864e-01 8.11561108e-01 -3.85555834e-01
8.83329868e-01 5.70311964e-01 1.02070332e+00 8.68769109e-01
4.58745211e-01 -3.02911699e-02 1.48763001e+00 -7.37429084e-03
1.10309601e-01 -3.44831534e-02 -6.60284579e-01 5.96454024e-01
-1.91018209e-01 5.07452190e-01 -7.23393619e-01 -2.11323693e-01
1.63595629e+00 2.03047171e-01 -8.00079763e-01 -4.00480658e-01
-1.46252191e+00 4.72113103e-01 3.50809574e-01 -4.53706235e-01
2.07038131e-02 3.60753119e-01 -1.95510760e-01 3.78608435e-01
5.26136875e-01 5.03743708e-01 -6.39472067e-01 1.56881973e-01
-6.60045326e-01 3.82601321e-02 3.63691956e-01 1.13074315e+00
1.04771376e+00 5.54108396e-02 2.63876349e-01 7.61092067e-01
2.78950274e-01 1.15241146e+00 -1.05916522e-01 -1.50235164e+00
2.98472017e-01 1.30980417e-01 3.56015675e-02 -8.17069292e-01
-4.25123602e-01 3.24014283e-04 -7.11533964e-01 6.64783120e-01
3.13348509e-02 8.78334939e-02 -8.21888268e-01 1.19438100e+00
6.69913650e-01 -3.85737270e-02 -7.80099556e-02 1.26791906e+00
9.04761791e-01 9.41574216e-01 -4.25566226e-01 -9.71919149e-02
1.13301563e+00 -5.95559239e-01 -3.02313656e-01 -3.56261611e-01
1.22315750e-01 -9.35458183e-01 1.38711286e+00 6.41369700e-01
-1.19412875e+00 -3.67764056e-01 -9.57852304e-01 -7.68878102e-01
1.72248408e-01 -1.84969902e-01 7.54939198e-01 4.04973358e-01
-1.20525992e+00 2.93576956e-01 -5.99995136e-01 5.24860620e-02
-5.45877358e-03 6.50916472e-02 -3.65978599e-01 -5.97589791e-01
-7.72301137e-01 6.84453189e-01 -1.11085899e-01 -1.56925023e-01
-1.02680290e+00 -1.64246571e+00 -9.58356559e-01 -2.40584970e-01
4.13466811e-01 -1.11432159e+00 8.50252092e-01 -5.43537796e-01
-1.56002271e+00 1.11985064e+00 4.15461570e-01 2.90745735e-01
6.77122474e-01 7.12319463e-02 -2.73775190e-01 5.65287769e-01
-3.66970673e-02 5.32539964e-01 5.57872057e-01 -1.49995899e+00
-2.19269603e-01 -5.06369293e-01 2.89456010e-01 6.56657219e-01
4.98637229e-01 -8.14224109e-02 -5.71222544e-01 -4.40430313e-01
2.29825422e-01 -7.69031882e-01 -1.57639906e-01 8.37387860e-01
-2.23022565e-01 9.16852891e-01 7.48992324e-01 -5.75373292e-01
4.84564871e-01 -2.05042458e+00 6.28558546e-02 -6.68845177e-02
5.55497646e-01 -4.48492616e-01 6.24233335e-02 -1.19153932e-02
-1.49928629e-01 -3.22771013e-01 -6.14836030e-02 -1.70405507e-01
-3.86166066e-01 7.87253492e-03 -4.38904494e-01 7.54059553e-01
-3.46041888e-01 6.58683360e-01 -8.34583938e-01 -4.24797773e-01
1.13343430e+00 9.54839230e-01 -5.98835766e-01 6.46130443e-01
-3.67679149e-01 8.71853769e-01 -2.02418968e-01 6.33132875e-01
1.33070207e+00 -2.22972557e-01 1.91377588e-02 -6.45231426e-01
-3.26875538e-01 -2.84903020e-01 -1.03673196e+00 2.07402778e+00
-9.85358715e-01 8.98141503e-01 1.67377800e-01 4.08553630e-01
8.81564200e-01 -2.00079814e-01 7.93055058e-01 -1.08006954e+00
-3.61971296e-02 -6.06869198e-02 -7.29308069e-01 -3.63874525e-01
7.26060629e-01 -3.26251745e-01 8.32183659e-02 5.29424548e-01
-3.06360185e-01 -1.01102293e+00 -4.55844045e-01 2.64190465e-01
1.03485930e+00 3.41675669e-01 3.01976383e-01 2.08565854e-02
2.02231497e-01 -1.78757921e-01 1.86083585e-01 4.21742022e-01
2.59236157e-01 1.44816041e+00 4.77269366e-02 -7.44251132e-01
-1.55433297e+00 -1.56704056e+00 -3.26157629e-01 7.51903892e-01
2.25979254e-01 -1.66298732e-01 -7.76227713e-02 6.66773915e-02
-1.82239026e-01 4.98622864e-01 -5.70017278e-01 3.25445145e-01
-5.31808913e-01 -7.23760664e-01 9.41208005e-02 1.33867130e-01
7.40383029e-01 -4.39672530e-01 -1.14141214e+00 -2.49333799e-01
-1.51063606e-01 -1.42345214e+00 -6.71297669e-01 1.21850364e-01
-6.99574590e-01 -1.24702358e+00 -6.51061714e-01 1.35771353e-02
3.94283652e-01 7.98408568e-01 1.62715900e+00 -2.02180311e-01
-8.26558232e-01 6.98888600e-01 -1.60486072e-01 2.91017205e-01
-3.39098483e-01 -5.25309205e-01 -3.59086603e-01 -3.38787645e-01
-3.25655967e-01 -7.62766540e-01 -1.10497522e+00 5.39744020e-01
-1.26110494e+00 6.55508935e-01 1.40790358e-01 2.36969903e-01
1.08833241e+00 -4.36993055e-02 -5.50942898e-01 -8.06878030e-01
-3.06919515e-01 -2.22269565e-01 -1.18406022e+00 3.13011289e-01
-2.44117156e-01 -1.22532606e-01 7.76433647e-01 -2.54204631e-01
-1.30199301e+00 1.15365267e-01 1.39763057e-01 -9.20473874e-01
-4.84060608e-02 -4.44747984e-01 -3.51503640e-01 -4.39908713e-01
7.30013847e-01 3.16882908e-01 -3.77659470e-01 -1.76856190e-01
5.60783386e-01 2.99781561e-01 8.21783781e-01 -7.13626742e-01
9.12744701e-01 1.04780829e+00 1.53478712e-01 -7.81433761e-01
-7.43079007e-01 -3.59427869e-01 -4.49839115e-01 -5.07469773e-01
1.04324651e+00 -1.44563317e+00 -8.51855099e-01 5.38995564e-01
-9.69150662e-01 -8.30820620e-01 -2.94600606e-01 4.03762788e-01
-7.46640921e-01 2.26094261e-01 -5.90534031e-01 -3.69273096e-01
-1.60000637e-01 -1.18116808e+00 1.62842667e+00 2.68860519e-01
3.91729414e-01 -1.05358565e+00 2.50853539e-01 4.20058131e-01
4.08783466e-01 6.20416939e-01 6.65460587e-01 9.75282371e-01
-1.40390408e+00 5.07856071e-01 -4.35641348e-01 -6.27027452e-02
-3.64485476e-03 2.71861609e-02 -1.41796982e+00 -2.78283477e-01
2.97989100e-01 -1.92326188e-01 3.60583872e-01 6.42552733e-01
1.48787153e+00 -3.00135743e-03 2.13355720e-01 1.85571659e+00
2.13474464e+00 -9.51305702e-02 1.05539548e+00 3.98683250e-01
1.13163579e+00 3.92398328e-01 5.80317438e-01 6.15461886e-01
6.12516761e-01 1.02088809e+00 6.60127163e-01 -4.27288800e-01
-4.68042910e-01 -1.70278847e-01 1.51564270e-01 4.29260403e-01
-4.29137945e-02 -3.28525275e-01 -6.16258144e-01 -1.33919835e-01
-1.05834055e+00 -8.44431758e-01 -3.75109732e-01 2.26319909e+00
7.08805084e-01 -5.12693346e-01 -2.28343144e-01 -4.10233200e-01
3.02361548e-01 5.44967949e-01 -1.03949034e+00 -1.83700621e-01
-4.17425066e-01 -1.79609153e-02 1.04057097e+00 7.21364141e-01
-4.95400906e-01 6.16026878e-01 6.55276012e+00 2.17097566e-01
-1.19603503e+00 -1.39928674e-02 7.96794415e-01 -6.02552593e-01
-9.82446611e-01 -6.83601655e-04 -4.04312342e-01 3.08332771e-01
5.51027179e-01 3.12663764e-02 1.08445358e+00 6.99352384e-01
1.76017299e-01 -5.69833636e-01 -9.96542811e-01 1.66938090e+00
4.65825558e-01 -1.18820751e+00 -7.29687065e-02 2.26409361e-01
7.77373731e-01 3.86015981e-01 2.77084202e-01 -3.99540663e-01
5.70018768e-01 -8.44158173e-01 6.82498455e-01 5.61710179e-01
1.59070158e+00 -6.47843480e-01 -2.05650002e-01 -3.68087478e-02
-9.69953597e-01 9.28084105e-02 -6.97495580e-01 4.71988738e-01
5.06870627e-01 7.17434585e-01 -5.72280467e-01 5.53742111e-01
9.67226446e-01 6.40455902e-01 -6.00718975e-01 5.78344703e-01
-4.07934524e-02 -2.93309450e-01 -2.28288144e-01 9.17788982e-01
-3.84221703e-01 -5.08813381e-01 3.16581786e-01 6.59427166e-01
4.02274609e-01 6.15021527e-01 -2.37686187e-01 1.05998516e+00
1.96665209e-02 -4.10722911e-01 -7.90137529e-01 8.04763019e-01
2.75704831e-01 1.39004886e+00 -5.36408901e-01 1.17432691e-01
-4.04423684e-01 1.33402658e+00 2.04982795e-02 4.32391018e-01
-9.54219937e-01 4.11193147e-02 7.57302523e-01 5.97016096e-01
-4.48537543e-02 -3.78726393e-01 -1.56674147e-01 -1.65233600e+00
-5.86224161e-02 -8.04450393e-01 1.12758636e-01 -1.94787073e+00
-1.12945104e+00 6.50107563e-01 2.21273735e-01 -1.42959917e+00
-4.18989845e-02 -6.50435746e-01 -1.74562991e-01 8.28924239e-01
-1.87297428e+00 -1.12416017e+00 -1.04105115e+00 8.00300121e-01
6.36006474e-01 2.54056126e-01 6.87060714e-01 2.94718266e-01
-5.79504073e-02 1.30177047e-02 8.89348388e-02 -2.28964761e-01
8.81702721e-01 -1.25829053e+00 4.01957214e-01 6.53733969e-01
-2.38054812e-01 2.93374866e-01 6.26661897e-01 -3.31724167e-01
-2.01124215e+00 -1.01482821e+00 -2.60362923e-01 -7.88409233e-01
2.42439181e-01 -5.75422168e-01 -6.12139761e-01 8.02487314e-01
-5.02951853e-02 5.96913517e-01 6.71107113e-01 -4.00755227e-01
-6.49124682e-01 -2.93604523e-01 -1.52111530e+00 4.30888772e-01
1.06052101e+00 -7.35211432e-01 2.31022120e-01 3.93565446e-01
9.23845053e-01 -1.18645680e+00 -1.22273755e+00 -6.32514954e-02
6.81129396e-01 -1.54815173e+00 1.53061974e+00 3.45930904e-01
7.43467033e-01 -6.54643536e-01 -5.29358268e-01 -1.38717473e+00
4.26253378e-02 -5.59464574e-01 6.72967881e-02 8.03931594e-01
-2.80673206e-01 -5.27896523e-01 4.03411806e-01 9.77481246e-01
-2.56750565e-02 -3.10221225e-01 -5.71173191e-01 -3.69931012e-01
-1.48619562e-01 -4.18150336e-01 8.60904336e-01 9.12995934e-01
-8.87466073e-01 -5.96030243e-03 -5.73154569e-01 4.28028554e-01
1.13223898e+00 3.51392239e-01 1.09908998e+00 -9.01343822e-01
-7.86965132e-01 2.03263476e-01 -2.64709085e-01 -1.12838674e+00
-1.11018442e-01 -5.39725006e-01 -1.69308260e-01 -1.27749300e+00
4.76736993e-01 -7.66584456e-01 5.40481985e-01 -1.84108824e-01
1.05154671e-01 5.89447558e-01 1.26452520e-01 2.36368299e-01
-3.12374651e-01 3.61071646e-01 1.65653479e+00 1.15048625e-01
-1.17596075e-01 -6.20104313e-01 -5.00803947e-01 5.77310860e-01
3.01498294e-01 -2.29454130e-01 -6.28370285e-01 -1.10333347e+00
6.46007001e-01 6.91279352e-01 6.67119443e-01 -9.44980025e-01
6.60114083e-03 -4.55279708e-01 9.91796672e-01 -6.89199150e-01
6.78354740e-01 -9.70434248e-01 8.96311224e-01 -2.56616622e-01
-4.82592434e-02 2.10303694e-01 -7.20694214e-02 3.80720794e-01
4.45495427e-01 5.08963406e-01 1.14071524e+00 -4.83466655e-01
-5.40149033e-01 7.87178576e-01 3.45281571e-01 1.46085799e-01
9.07333910e-01 -3.36310059e-01 -7.27388680e-01 -4.67388242e-01
-1.05781935e-01 7.47463293e-03 1.43092167e+00 1.77081838e-01
8.68225396e-01 -1.17883933e+00 -5.64756691e-01 4.79114801e-01
2.86987036e-01 6.27305686e-01 9.06454742e-01 1.66517556e-01
-1.49012101e+00 -1.19629562e-01 -3.08005959e-01 -8.14588070e-01
-1.12016737e+00 5.95216334e-01 6.64809108e-01 2.31321324e-02
-1.22866869e+00 4.51850235e-01 7.44879782e-01 -6.51234150e-01
-2.39431188e-01 -2.12542266e-01 4.34033930e-01 -5.93823671e-01
7.61605501e-01 3.29363704e-01 -3.35304104e-02 -6.34020209e-01
-1.29185200e-01 1.16332650e+00 2.36729622e-01 -3.04871619e-01
1.45954812e+00 -6.71422839e-01 2.33759143e-04 4.33299661e-01
1.42057872e+00 1.37028158e-01 -1.93917477e+00 1.73025236e-01
-1.22306871e+00 -1.35009241e+00 5.58706105e-01 -7.90380478e-01
-1.54466581e+00 6.76704943e-01 7.03722656e-01 -4.24962252e-01
1.39263511e+00 2.14947015e-03 6.81295276e-01 -2.44824905e-02
8.30609322e-01 -6.97081625e-01 -2.69254241e-02 1.06307857e-01
8.60626757e-01 -1.26409209e+00 5.54652214e-01 -5.79938948e-01
-6.27578378e-01 1.33375967e+00 6.82160020e-01 -3.47917043e-02
4.66138482e-01 6.92146242e-01 3.84562880e-01 -3.09256375e-01
-6.94414139e-01 3.43425632e-01 -4.50295620e-02 9.47875381e-01
1.70775399e-01 -8.37007388e-02 6.24205112e-01 -4.87368435e-01
-3.81657660e-01 -2.49146551e-01 1.09693325e+00 4.13442224e-01
-9.61598307e-02 -6.40399277e-01 -6.28688812e-01 1.58796728e-01
1.35467676e-02 -5.17310835e-02 1.97618797e-01 6.39684498e-01
-1.58174530e-01 6.95316732e-01 3.52683157e-01 -2.49595508e-01
3.96692604e-01 -9.20302689e-01 1.01156306e+00 -5.00484467e-01
-3.67275268e-01 6.15003668e-02 -3.30509096e-01 -1.14403212e+00
-3.57123822e-01 -4.04457361e-01 -9.62434053e-01 -7.33458221e-01
8.28678831e-02 -6.09251380e-01 8.09902072e-01 2.02875391e-01
1.05141431e-01 4.76696193e-01 1.27432096e+00 -1.08373201e+00
9.09387320e-02 -3.03269148e-01 -1.00614810e+00 2.56860286e-01
5.95639765e-01 -3.40847373e-01 -7.35625625e-01 5.48666157e-02] | [9.60166072845459, -3.0587897300720215] |
59022517-ab47-425e-8e45-600c4c00642c | low-rank-factorization-of-determinantal-point | 1602.05436 | null | http://arxiv.org/abs/1602.05436v1 | http://arxiv.org/pdf/1602.05436v1.pdf | Low-Rank Factorization of Determinantal Point Processes for Recommendation | Determinantal point processes (DPPs) have garnered attention as an elegant
probabilistic model of set diversity. They are useful for a number of subset
selection tasks, including product recommendation. DPPs are parametrized by a
positive semi-definite kernel matrix. In this work we present a new method for
learning the DPP kernel from observed data using a low-rank factorization of
this kernel. We show that this low-rank factorization enables a learning
algorithm that is nearly an order of magnitude faster than previous approaches,
while also providing for a method for computing product recommendation
predictions that is far faster (up to 20x faster or more for large item
catalogs) than previous techniques that involve a full-rank DPP kernel.
Furthermore, we show that our method provides equivalent or sometimes better
predictive performance than prior full-rank DPP approaches, and better
performance than several other competing recommendation methods in many cases.
We conduct an extensive experimental evaluation using several real-world
datasets in the domain of product recommendation to demonstrate the utility of
our method, along with its limitations. | ['Mike Gartrell', 'Noam Koenigstein', 'Ulrich Paquet'] | 2016-02-17 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [ 8.41476768e-03 -3.66899043e-01 -7.78268695e-01 -3.52996707e-01
-1.00517261e+00 -9.35472906e-01 7.09371507e-01 1.89721510e-01
-1.40664214e-02 4.80160117e-01 4.58348215e-01 -4.13140059e-01
-4.61677551e-01 -5.38864195e-01 -7.35885918e-01 -7.63717890e-01
-1.85054913e-01 6.88660860e-01 3.20950985e-01 -2.07322910e-01
4.94898587e-01 2.42090985e-01 -1.57015812e+00 4.21233535e-01
7.28597343e-01 9.23579037e-01 -7.98735842e-02 4.28909570e-01
3.04769397e-01 6.08802557e-01 -1.35135269e-02 -6.53634191e-01
4.45531815e-01 1.60837248e-01 -6.18394971e-01 -2.09561847e-02
5.50179899e-01 -9.82049480e-02 -8.27883109e-02 8.49675775e-01
1.55672714e-01 3.70909840e-01 1.15213454e+00 -1.02963197e+00
-7.87181735e-01 6.78890407e-01 -7.49736965e-01 1.24101736e-01
3.53578627e-01 -3.16264570e-01 1.72138870e+00 -1.07482457e+00
1.83586612e-01 1.19537401e+00 8.23485613e-01 1.12204393e-02
-1.73861313e+00 -5.79669654e-01 2.51705915e-01 -2.66537488e-01
-1.31785452e+00 -1.06199957e-01 3.75819266e-01 -6.37907207e-01
6.65104866e-01 4.28458214e-01 3.82186830e-01 9.10175562e-01
4.33315098e-01 1.11034489e+00 1.00290060e+00 -2.03979924e-01
3.56566846e-01 1.95013717e-01 7.45473385e-01 5.42163908e-01
5.59539795e-01 8.93524364e-02 -8.09281349e-01 -1.13510561e+00
8.54360521e-01 1.82494193e-01 -2.09125593e-01 -6.37986183e-01
-1.27986383e+00 1.25125968e+00 -1.59543768e-01 -1.16248325e-01
-4.23397750e-01 2.24975184e-01 3.00481141e-01 2.51665741e-01
7.93816805e-01 4.77718025e-01 -5.65980971e-01 -7.84735978e-02
-8.52703393e-01 4.27296788e-01 1.25711477e+00 7.25829065e-01
6.26767993e-01 -3.26570779e-01 -3.18736196e-01 9.59164143e-01
4.21766371e-01 5.12963116e-01 1.86573431e-01 -8.61867249e-01
3.36374015e-01 2.58791625e-01 5.71637332e-01 -8.41639042e-01
-1.50357425e-01 -4.17982399e-01 -7.43199229e-01 4.08896841e-02
4.85960156e-01 1.57767147e-01 -5.76959312e-01 1.48828876e+00
2.03724265e-01 3.49836886e-01 -1.29394442e-01 7.79630661e-01
1.84520364e-01 8.56348813e-01 -6.09872527e-02 -3.52589279e-01
1.25111854e+00 -8.65483880e-01 -2.72531301e-01 2.90681392e-01
5.13474166e-01 -8.98745596e-01 1.20213628e+00 9.44431841e-01
-6.69438124e-01 -3.46009403e-01 -8.60351980e-01 2.33050063e-01
1.48730218e-01 2.37317070e-01 1.34784818e+00 9.18710113e-01
-9.34297979e-01 7.45508015e-01 -6.88827634e-01 -2.47614294e-01
1.35136366e-01 4.03004408e-01 -1.60618708e-01 -1.27199024e-01
-8.09203327e-01 3.23687702e-01 -5.24444729e-02 -2.58850664e-01
-5.57051599e-01 -1.18311119e+00 -5.61047018e-01 1.78994894e-01
1.93825960e-01 -7.18513727e-01 1.44762337e+00 -6.21024430e-01
-1.73445940e+00 1.94699824e-01 -1.29642516e-01 -4.01994228e-01
-1.80960950e-02 -5.73235869e-01 -3.43364209e-01 -2.60107011e-01
-9.36470777e-02 1.88562006e-01 1.00556326e+00 -1.00641549e+00
-7.95516074e-01 -2.45679334e-01 3.62851215e-03 2.06805050e-01
-4.21153694e-01 3.24043818e-02 -2.96290547e-01 -8.29648137e-01
2.47939266e-02 -1.27624106e+00 -5.26104271e-01 -3.46536100e-01
-2.59951413e-01 -5.43297231e-01 2.82009512e-01 -2.93881208e-01
1.32625353e+00 -2.22077775e+00 -4.11537988e-03 5.18949628e-01
1.87571213e-01 -3.28928456e-02 -1.51138291e-01 6.75110281e-01
1.19747482e-01 7.78350085e-02 1.62826136e-01 -2.49990985e-01
2.49934793e-01 1.18546456e-01 -8.73738229e-01 7.21769154e-01
-1.04669228e-01 6.71569288e-01 -8.67354274e-01 8.66480172e-02
-3.72218601e-02 3.22396308e-01 -6.74840510e-01 -3.02467406e-01
-3.03390622e-01 5.73834814e-02 -4.37532574e-01 3.83109540e-01
5.33369064e-01 -5.75286210e-01 1.76461622e-01 -1.19054683e-01
3.30366902e-02 3.66504639e-01 -1.54677618e+00 1.25349784e+00
-2.37197548e-01 3.02979499e-01 -3.87054503e-01 -5.12580574e-01
6.80515409e-01 6.27650991e-02 6.91128194e-01 1.53592736e-01
-2.26716653e-01 2.00385273e-01 -1.93083003e-01 3.47664565e-01
8.24895084e-01 -8.63019228e-02 -1.93287075e-01 8.09770286e-01
4.91125183e-03 2.54072458e-01 2.70591766e-01 2.75368482e-01
1.17912269e+00 -1.36047080e-01 2.05685601e-01 -7.12337554e-01
3.47681850e-01 -1.19189687e-01 6.10101044e-01 9.84722018e-01
2.89401621e-01 4.06387150e-01 5.44741631e-01 -5.23674011e-01
-8.34142983e-01 -1.19601321e+00 -2.24346071e-01 1.43278027e+00
1.94175288e-01 -7.64374971e-01 -5.20121790e-02 -6.32573545e-01
5.79885483e-01 6.75944209e-01 -5.14083147e-01 5.12813032e-02
-1.28896415e-01 -1.16251802e+00 3.38567823e-01 6.14486337e-01
8.18699449e-02 -4.02005762e-01 3.77271026e-01 2.19962716e-01
2.70981282e-01 -8.29998732e-01 -9.76974607e-01 1.03243165e-01
-1.09684503e+00 -8.96929145e-01 -7.46274412e-01 -6.76857233e-01
4.26399231e-01 6.25138342e-01 1.15235996e+00 -5.06084383e-01
1.83111832e-01 5.62878609e-01 -3.20768386e-01 -2.14982852e-01
-3.49026620e-01 1.34270206e-01 6.07840717e-01 1.83970749e-01
4.58050281e-01 -5.17807841e-01 -5.54592192e-01 6.66382730e-01
-7.80131876e-01 -1.66648626e-01 6.39273405e-01 9.83884752e-01
7.55630851e-01 2.67943323e-01 4.57037717e-01 -1.26782107e+00
9.67440486e-01 -5.08434355e-01 -8.26079428e-01 2.92291880e-01
-9.67579842e-01 3.61269891e-01 4.28163379e-01 -7.70134628e-01
-1.09874451e+00 1.88096926e-01 7.40153790e-02 -1.62482545e-01
1.53180063e-01 6.31806731e-01 3.27669889e-01 -8.84076655e-02
8.16447079e-01 3.50555778e-02 -1.92772925e-01 -6.95822895e-01
6.36989772e-01 4.68571752e-01 3.94735597e-02 -6.61502421e-01
7.57317781e-01 4.49393511e-01 7.47415004e-03 -5.16046703e-01
-9.49532926e-01 -9.67411876e-01 -3.91458303e-01 3.30923975e-01
3.56199294e-01 -9.95309114e-01 -7.93383360e-01 1.67082489e-01
-5.71574867e-01 -2.25659505e-01 -2.22927436e-01 8.82102609e-01
-5.71328282e-01 3.77911091e-01 -1.08211231e+00 -8.36854339e-01
-3.58323306e-01 -6.73233151e-01 1.09610558e+00 -3.65141593e-02
-1.24419376e-01 -9.36829925e-01 5.09365082e-01 2.59141862e-01
7.26890564e-02 -4.51832443e-01 1.09313536e+00 -7.97224224e-01
-3.75257850e-01 -1.28912404e-01 -3.18326727e-02 2.88286895e-01
3.56855500e-03 5.47365798e-03 -5.42936862e-01 -3.83921891e-01
-2.46935338e-01 -8.94912053e-03 1.10129237e+00 5.11195242e-01
7.57157087e-01 -1.53609931e-01 -4.87851471e-01 3.10505569e-01
1.34265018e+00 -2.68810689e-01 3.35796714e-01 -6.01725699e-03
6.75783873e-01 2.50661373e-01 7.99580395e-01 5.99479318e-01
2.95558900e-01 7.50587821e-01 -1.60314500e-01 1.03162602e-01
2.46040821e-01 -4.01835352e-01 6.10564291e-01 8.25993657e-01
-1.86394542e-01 -2.17666864e-01 -5.52568197e-01 2.64586300e-01
-2.45593953e+00 -1.10058403e+00 -3.04319531e-01 2.69612217e+00
6.36182964e-01 4.02689129e-02 5.21152616e-01 -2.33151704e-01
5.85204124e-01 -1.16676316e-01 -5.92814684e-01 -3.20612013e-01
-4.83127236e-02 1.55985773e-01 9.20975447e-01 4.63118136e-01
-1.32846653e+00 7.94294238e-01 7.65393066e+00 9.80622411e-01
-6.41678154e-01 6.26723692e-02 5.62954068e-01 -6.15211800e-02
-4.98830169e-01 -2.64927316e-02 -1.16565418e+00 3.31980020e-01
8.30056608e-01 -9.91736427e-02 3.60113025e-01 1.10820723e+00
-4.75668050e-02 -1.59823105e-01 -1.30135870e+00 1.13561106e+00
-2.10058447e-02 -1.41419029e+00 7.23289102e-02 4.47463840e-01
1.12408388e+00 1.35815591e-01 5.26603639e-01 3.92838806e-01
8.84827673e-01 -7.25564420e-01 3.62646848e-01 3.16988915e-01
4.34499681e-01 -9.56969619e-01 6.39154255e-01 4.67027932e-01
-1.12226880e+00 -2.22825453e-01 -8.43195438e-01 -8.13496187e-02
1.75412521e-01 1.07137132e+00 -7.60998189e-01 3.32506895e-01
5.45098364e-01 8.13360751e-01 -2.27899075e-01 1.27365959e+00
-3.00701484e-02 1.12036741e+00 -5.11505067e-01 -9.66519564e-02
-7.44495913e-02 -3.96713138e-01 4.66140032e-01 1.24986601e+00
4.33680981e-01 9.14429352e-02 1.93615317e-01 3.85965616e-01
-9.96367857e-02 5.02715826e-01 -4.40146685e-01 -8.59449357e-02
1.74715459e-01 1.13944399e+00 -6.13926172e-01 -1.45591293e-02
-6.01663113e-01 9.70474124e-01 3.92413974e-01 3.81535560e-01
-7.05973685e-01 1.66014269e-01 9.90510345e-01 1.33578330e-01
6.63972497e-01 -3.41296881e-01 -1.19270496e-01 -1.40363741e+00
-2.28126749e-01 -7.54332542e-01 3.46826822e-01 -2.37711862e-01
-2.05934620e+00 1.42650634e-01 1.01698011e-01 -1.33712685e+00
-3.29995811e-01 -9.11403120e-01 -2.79571831e-01 7.59705186e-01
-1.16291189e+00 -9.33129728e-01 5.46416879e-01 3.79496932e-01
3.77551556e-01 -3.56574208e-01 9.60725069e-01 1.42629325e-01
-3.83693814e-01 5.36576748e-01 7.92100012e-01 -3.85883301e-01
9.30186093e-01 -1.45407104e+00 4.04361606e-01 5.68194211e-01
7.82550216e-01 9.17269289e-01 6.94259226e-01 -6.83355927e-01
-1.76838827e+00 -8.33130956e-01 5.18485785e-01 -7.36242950e-01
8.38928282e-01 -4.14340109e-01 -5.56920648e-01 6.94425166e-01
-2.06606686e-01 -1.08390570e-01 1.39490938e+00 1.04611158e+00
-7.10274398e-01 -1.88901395e-01 -7.31915236e-01 6.32618546e-01
8.45634341e-01 -3.42172146e-01 -4.14740562e-01 5.63547671e-01
6.10819161e-01 -1.01964712e-01 -1.01393449e+00 3.41286570e-01
8.74007285e-01 -7.48039484e-01 9.60829258e-01 -6.72742724e-01
5.71769886e-02 -4.33061898e-01 -3.44878435e-01 -1.26923215e+00
-1.19474006e+00 -8.73372674e-01 -4.92442936e-01 1.01262152e+00
7.45868802e-01 -6.05211079e-01 9.34144676e-01 6.85766101e-01
1.85492903e-01 -8.24834287e-01 -5.39377570e-01 -8.25221062e-01
-1.09759783e-02 -5.64319193e-01 3.73918146e-01 6.89826369e-01
8.41437355e-02 7.51154363e-01 -6.22276962e-01 4.21122283e-01
6.62738621e-01 5.45003355e-01 7.83326209e-01 -1.57943523e+00
-8.87964964e-01 -2.68597424e-01 -4.36017424e-01 -1.72556126e+00
1.17702357e-01 -8.55426610e-01 -1.79351375e-01 -1.16542721e+00
6.13046169e-01 -6.76616251e-01 -6.43991947e-01 1.55755609e-01
-1.48071006e-01 4.37790811e-01 -1.50791675e-01 5.86356223e-01
-7.47949481e-01 3.19085777e-01 9.31169927e-01 8.43290687e-02
-4.16290075e-01 5.25954306e-01 -1.07839477e+00 7.75176287e-01
6.97891295e-01 -4.79782999e-01 -6.52547956e-01 9.85679217e-03
6.50853336e-01 -2.23744616e-01 -6.38950169e-02 -6.52202487e-01
1.13386512e-01 -2.84387082e-01 3.58627975e-01 -7.65907228e-01
3.01443100e-01 -4.09445405e-01 3.23669314e-01 7.06670806e-02
-3.45355928e-01 -8.80147666e-02 4.91589587e-03 1.13179350e+00
-1.96791608e-02 -2.21755505e-01 3.20402622e-01 2.40037918e-01
-5.77889681e-01 4.25255120e-01 -4.11023647e-01 -3.94381136e-01
7.75808930e-01 1.15409397e-01 -2.11069688e-01 -5.23842812e-01
-5.82679093e-01 -1.14936739e-01 4.13806081e-01 3.62593949e-01
3.79056782e-01 -1.48325372e+00 -6.92337155e-01 -1.09379124e-02
2.93552458e-01 -6.99931920e-01 -1.80877205e-02 8.08063745e-01
-1.53279707e-01 5.05300224e-01 4.04848814e-01 -5.98666131e-01
-1.19775712e+00 7.96215355e-01 -2.03469202e-01 -6.84264243e-01
-5.23971260e-01 8.69009495e-01 4.89771336e-01 -3.34463090e-01
6.79791570e-02 -3.73628050e-01 -7.62450099e-02 -2.38562018e-01
8.42564821e-01 3.50749820e-01 -1.70199737e-01 -4.70749855e-01
-6.99215084e-02 3.94388318e-01 -4.50097144e-01 -2.69162863e-01
1.27741385e+00 2.07012929e-02 3.02040819e-02 6.63927794e-01
8.76703143e-01 5.32568753e-01 -1.22743249e+00 -4.06612068e-01
4.88267280e-02 -6.18741274e-01 -1.49645256e-02 -5.53643346e-01
-6.00886583e-01 2.17892408e-01 3.35027337e-01 3.55027407e-01
7.19883263e-01 2.28835955e-01 6.78055167e-01 5.95917821e-01
8.81286919e-01 -8.30430806e-01 4.07918654e-02 4.91135836e-01
6.39258981e-01 -1.21989429e+00 2.65093207e-01 -8.77215147e-01
-8.06201339e-01 9.57214653e-01 6.97375014e-02 -4.92266119e-01
1.10438395e+00 1.81272492e-01 -4.72208083e-01 -4.47062514e-04
-1.00241041e+00 -4.98328470e-02 7.47453511e-01 3.97777230e-01
4.65944201e-01 3.24091166e-01 -5.17741919e-01 9.70403016e-01
-2.92612277e-02 -5.60757332e-02 2.36667618e-01 6.36448324e-01
-6.61282957e-01 -1.53333533e+00 -2.55420983e-01 9.70033050e-01
-5.69172144e-01 -2.94526994e-01 -8.38815272e-02 3.59528750e-01
-2.91457117e-01 1.00680852e+00 -1.11867376e-01 -6.22083187e-01
1.34780928e-01 -3.43223359e-03 5.64935684e-01 -8.85888100e-01
-4.29324329e-01 2.55017132e-01 1.97391927e-01 -5.22309899e-01
-1.82001546e-01 -9.95445609e-01 -7.18514025e-01 -5.45176208e-01
-6.04274452e-01 3.67042750e-01 3.63199323e-01 6.51985884e-01
5.21559954e-01 -2.27713734e-01 7.44886279e-01 -7.26684809e-01
-1.01770175e+00 -7.78486907e-01 -1.11810279e+00 2.79374033e-01
-1.41213238e-01 -7.96292424e-01 -3.15297008e-01 -1.03343986e-02] | [9.520500183105469, 5.40716552734375] |
0778d480-2d75-4c3c-93b4-222f46ff3418 | electramed-a-new-pre-trained-language | 2104.09585 | null | https://arxiv.org/abs/2104.09585v1 | https://arxiv.org/pdf/2104.09585v1.pdf | ELECTRAMed: a new pre-trained language representation model for biomedical NLP | The overwhelming amount of biomedical scientific texts calls for the development of effective language models able to tackle a wide range of biomedical natural language processing (NLP) tasks. The most recent dominant approaches are domain-specific models, initialized with general-domain textual data and then trained on a variety of scientific corpora. However, it has been observed that for specialized domains in which large corpora exist, training a model from scratch with just in-domain knowledge may yield better results. Moreover, the increasing focus on the compute costs for pre-training recently led to the design of more efficient architectures, such as ELECTRA. In this paper, we propose a pre-trained domain-specific language model, called ELECTRAMed, suited for the biomedical field. The novel approach inherits the learning framework of the general-domain ELECTRA architecture, as well as its computational advantages. Experiments performed on benchmark datasets for several biomedical NLP tasks support the usefulness of ELECTRAMed, which sets the novel state-of-the-art result on the BC5CDR corpus for named entity recognition, and provides the best outcome in 2 over the 5 runs of the 7th BioASQ-factoid Challange for the question answering task. | ['Carlotta Orsenigo', 'Giulio Mantoan', 'Giacomo Miolo'] | 2021-04-19 | null | null | null | null | ['drug-drug-interaction-extraction', 'medical-named-entity-recognition'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.83536962e-01 2.88739800e-01 -6.28940910e-02 -4.20451999e-01
-8.93528044e-01 -2.09949717e-01 7.43211150e-01 5.82249582e-01
-9.35586691e-01 1.02003062e+00 4.71963845e-02 -2.41904512e-01
-3.67940575e-01 -5.62707782e-01 -5.86855531e-01 -6.05741918e-01
-6.93831369e-02 9.14711893e-01 2.30163828e-01 -2.16743127e-01
3.66039462e-02 7.08575249e-01 -1.10864937e+00 5.71521878e-01
8.41725647e-01 8.53619933e-01 3.08618188e-01 3.82314503e-01
-5.98071396e-01 5.83165169e-01 -6.05045915e-01 -5.69819033e-01
-2.31691152e-01 -3.58725280e-01 -1.07798505e+00 -3.57108712e-01
-5.66733666e-02 3.97533149e-01 -4.23862152e-02 6.10474527e-01
8.16069186e-01 -1.12820677e-01 5.93595028e-01 -6.91887856e-01
-2.35640943e-01 6.80498123e-01 -7.76197240e-02 3.53256643e-01
1.93367705e-01 -1.59783006e-01 9.74348009e-01 -9.51110363e-01
1.05605769e+00 9.34359312e-01 5.78378677e-01 8.15333605e-01
-1.21811318e+00 -3.83407056e-01 -2.15187892e-01 2.18665183e-01
-1.44678354e+00 -4.66146708e-01 4.05728102e-01 -3.72741193e-01
1.31282806e+00 1.88681364e-01 1.94720134e-01 1.39627743e+00
3.03189129e-01 8.80615890e-01 1.12863243e+00 -5.18366694e-01
4.64897215e-01 3.54628384e-01 1.57078221e-01 3.27073187e-01
3.75842303e-01 -1.54641166e-01 -4.60523129e-01 -4.11306709e-01
1.82233989e-01 -4.41001356e-01 -2.69298673e-01 -3.94499391e-01
-1.27372003e+00 7.20533907e-01 3.66467517e-03 1.09359694e+00
-6.67900920e-01 -3.33967507e-01 8.18703175e-01 2.42869332e-01
6.02537394e-01 7.96445727e-01 -8.80808532e-01 -3.64080630e-02
-1.08467531e+00 1.64369971e-01 1.10712171e+00 5.61196268e-01
2.62277305e-01 -3.01429778e-01 -1.51099026e-01 9.57638562e-01
1.33485928e-01 2.23342732e-01 8.01436841e-01 -2.31797233e-01
3.87499034e-01 7.29196310e-01 -1.23331718e-01 -7.56558061e-01
-8.33166063e-01 -6.46291256e-01 -1.01953268e+00 -3.11149329e-01
6.65960729e-01 -1.94397777e-01 -7.52828062e-01 1.72793722e+00
3.54209304e-01 6.85392618e-02 5.57369173e-01 5.08231282e-01
9.13350165e-01 5.74984074e-01 4.52005863e-01 -1.73200533e-01
1.47219336e+00 -4.37111259e-01 -7.73064733e-01 -2.06779525e-01
7.97564745e-01 -6.67674363e-01 4.98759389e-01 4.91127640e-01
-7.93805540e-01 -4.02016938e-01 -8.36672723e-01 3.31754535e-02
-8.40270221e-01 2.69317150e-01 3.67642075e-01 6.40928626e-01
-9.84248936e-01 5.25251389e-01 -8.33148241e-01 -7.48035848e-01
4.27308053e-01 4.32634205e-01 -5.77756584e-01 -1.74871162e-01
-1.50140321e+00 1.26321852e+00 9.77056861e-01 2.04648882e-01
-5.16355753e-01 -7.59909928e-01 -4.53415632e-01 1.76400878e-02
3.52417946e-01 -6.33797228e-01 7.40788281e-01 -6.83508992e-01
-1.37506294e+00 1.30043387e+00 2.34762087e-01 -8.63387406e-01
5.14146328e-01 -8.80797058e-02 -5.87576509e-01 2.50441402e-01
-1.16100326e-01 5.82658291e-01 3.94453645e-01 -8.49349856e-01
-3.64022642e-01 -4.51269150e-01 -2.27446213e-01 -1.68635637e-01
-4.84544337e-01 2.19521876e-02 -2.81276554e-01 -4.82134283e-01
-5.12468040e-01 -6.96526945e-01 -4.60398644e-01 -3.14111501e-01
-2.84258902e-01 -6.19986951e-01 2.45426252e-01 -6.59178078e-01
1.23945951e+00 -2.28670287e+00 2.11510777e-01 7.44645000e-02
-3.02388053e-02 6.97438061e-01 -2.51218617e-01 7.01297164e-01
-4.93776709e-01 -1.07066035e-02 -3.44163030e-01 -1.38504207e-01
-1.01084650e-01 3.45226496e-01 -9.88077074e-02 4.11671638e-01
3.65183711e-01 7.76721776e-01 -7.90384471e-01 -6.92132771e-01
-6.31048083e-02 4.60143089e-01 -3.93465847e-01 2.20094398e-01
-3.60090345e-01 3.53640378e-01 -8.23794007e-01 1.65857762e-01
5.53164959e-01 -4.23998386e-01 4.19099987e-01 -6.85283393e-02
-1.24780007e-01 3.96928519e-01 -9.54096019e-01 1.90675092e+00
-3.09476852e-01 1.44472092e-01 1.38149364e-02 -1.53903759e+00
9.95519400e-01 7.57498085e-01 7.90856540e-01 -6.17256522e-01
1.36393160e-01 5.68258405e-01 1.03281192e-01 -7.68898070e-01
3.27120274e-02 -3.45324546e-01 -3.32596563e-02 5.98196499e-02
4.51037347e-01 9.64206457e-02 4.13053423e-01 4.09669168e-02
1.28409278e+00 1.09082848e-01 8.10221910e-01 -6.65004790e-01
1.02940547e+00 1.49132058e-01 2.95186043e-01 5.74638844e-01
-3.24351229e-02 4.23333049e-01 4.63885367e-01 -5.37679613e-01
-6.83693647e-01 -6.75995111e-01 -6.18030250e-01 8.80663216e-01
-5.18610537e-01 -3.57601583e-01 -7.52997756e-01 -7.62322545e-01
-2.24192873e-01 7.40203559e-01 -4.60287988e-01 6.55504018e-02
-6.95398510e-01 -1.10350049e+00 9.23824489e-01 2.65390407e-02
3.03244263e-01 -1.35705292e+00 -6.85999870e-01 6.74403727e-01
-1.81603685e-01 -1.34400523e+00 1.74699068e-01 4.48101163e-01
-9.37412083e-01 -1.04264736e+00 -1.06707633e+00 -6.49056971e-01
1.83628157e-01 -6.38305783e-01 1.27344298e+00 -2.43494198e-01
-4.58728999e-01 3.07507545e-01 -4.58574772e-01 -6.13719046e-01
-6.95414484e-01 6.31279767e-01 -2.37055555e-01 7.68612251e-02
5.97140133e-01 -2.82755435e-01 -2.57885545e-01 -4.80148643e-02
-1.05487764e+00 -1.41032815e-01 9.87300158e-01 1.00741744e+00
5.44551492e-01 -2.18868390e-01 1.16260874e+00 -1.11992049e+00
5.26686728e-01 -6.46810532e-01 -3.96063626e-01 4.76184875e-01
-5.05277812e-01 3.47685963e-01 8.18761647e-01 -2.62145221e-01
-9.96652007e-01 7.26877525e-02 -8.06732059e-01 1.45866409e-01
-4.71541405e-01 8.94157231e-01 -2.55210191e-01 1.96782231e-01
9.10192847e-01 4.08491731e-01 -2.35119328e-01 -7.58832037e-01
1.66113630e-01 7.00898588e-01 2.57775605e-01 -5.01995265e-01
1.94362223e-01 2.55231798e-01 1.04690485e-01 -1.19832182e+00
-8.84792745e-01 -7.71622121e-01 -4.65295047e-01 2.69574940e-01
1.14605105e+00 -7.48698175e-01 -4.54647750e-01 3.74863923e-01
-1.34103334e+00 -5.39025441e-02 -3.05881828e-01 5.79870343e-01
-3.49278063e-01 3.99494588e-01 -4.28619057e-01 -5.47565341e-01
-6.63339436e-01 -9.37291086e-01 9.24190164e-01 -1.45464689e-01
-2.90455014e-01 -1.12809944e+00 3.77328813e-01 2.29741737e-01
3.53628516e-01 2.32591167e-01 1.24424171e+00 -1.48584843e+00
-9.61371660e-02 -2.54624814e-01 -1.00171650e-02 4.03089464e-01
-5.62066436e-02 -5.79016984e-01 -1.01292765e+00 -1.29037842e-01
7.36796185e-02 -4.74257231e-01 7.80285835e-01 5.83092682e-02
1.01547027e+00 6.91854628e-03 -5.88628173e-01 1.09931201e-01
1.47263265e+00 9.00117084e-02 6.44413769e-01 3.36250484e-01
7.25371465e-02 6.70081556e-01 4.53980863e-01 3.36151183e-01
1.36032611e-01 6.90634429e-01 7.86156505e-02 -2.01755300e-01
1.55410931e-01 1.45298243e-01 -2.86981072e-02 7.41107583e-01
-3.56208868e-02 -4.60376889e-01 -1.22498953e+00 7.22525954e-01
-1.52718747e+00 -6.82553887e-01 -4.48430143e-03 1.97846115e+00
1.22025907e+00 1.93299472e-01 -1.08540900e-01 -5.39134890e-02
4.12644565e-01 -3.45257372e-02 -3.02068889e-01 -3.74761760e-01
-1.63543478e-01 7.26014316e-01 1.70620784e-01 3.01820710e-02
-1.09043694e+00 6.94083750e-01 5.77610207e+00 1.12400711e+00
-1.09678853e+00 2.18273222e-01 5.21583378e-01 2.83799946e-01
1.26639828e-01 -4.89444435e-01 -9.25704420e-01 3.23874444e-01
1.64045548e+00 -2.02297479e-01 -1.05903052e-01 7.66039252e-01
5.57078831e-02 7.11304173e-02 -1.25381804e+00 9.49013174e-01
1.81760073e-01 -1.37886953e+00 6.90709949e-02 5.04115946e-04
3.43212605e-01 3.81440073e-01 -3.64506304e-01 4.68482614e-01
-2.11856231e-01 -9.79870319e-01 2.54384518e-01 5.58932960e-01
5.74946880e-01 -4.50770676e-01 1.08879542e+00 6.23851657e-01
-7.26792753e-01 2.07956821e-01 -4.52645093e-01 5.61885476e-01
4.31375913e-02 9.19902265e-01 -1.09876180e+00 1.07060635e+00
4.69614297e-01 5.55940568e-01 -4.63819176e-01 1.09196556e+00
1.01404525e-01 7.00117588e-01 -3.90696555e-01 -1.96239278e-01
2.99991220e-01 1.46247104e-01 5.19429982e-01 1.75620937e+00
1.88821524e-01 4.24749665e-02 -9.22048688e-02 6.45734727e-01
-2.06586331e-01 7.29384720e-01 -4.62576419e-01 -4.54376101e-01
-2.08800044e-02 1.21724319e+00 -5.85104167e-01 -4.89285946e-01
-3.63051832e-01 6.46646380e-01 4.40859258e-01 1.48903457e-02
-6.16109014e-01 -3.56360823e-01 1.57054722e-01 8.00813958e-02
2.83955067e-01 -2.80934405e-02 -7.89732784e-02 -1.03687668e+00
-1.46006986e-01 -1.19702220e+00 7.32404172e-01 -2.84056276e-01
-1.49233389e+00 9.55885828e-01 5.59544601e-02 -9.39383388e-01
-3.35528463e-01 -9.37842906e-01 -3.65493074e-02 7.85724223e-01
-1.68842852e+00 -9.74605918e-01 3.21105152e-01 5.43153226e-01
4.12876189e-01 -2.93336540e-01 1.27843726e+00 6.85252130e-01
-2.90734649e-01 3.42929572e-01 3.40482652e-01 9.29259807e-02
9.20004249e-01 -1.10257673e+00 -1.77617855e-02 3.63052696e-01
1.13238700e-01 6.62803173e-01 7.18116999e-01 -3.15680295e-01
-1.23359859e+00 -9.02766585e-01 1.67785907e+00 -4.33391541e-01
6.29742682e-01 -3.97132427e-01 -1.23707235e+00 3.11564893e-01
2.34876961e-01 -7.64844045e-02 8.76756728e-01 1.43516704e-01
-6.92584962e-02 -1.14640988e-01 -1.10725987e+00 1.42294273e-01
6.31461799e-01 -3.49326313e-01 -1.02116144e+00 4.91888046e-01
2.05300182e-01 -2.15584621e-01 -1.23818421e+00 4.21426684e-01
2.17515424e-01 -5.50480485e-01 9.81141031e-01 -7.85374641e-01
3.56270343e-01 1.74407959e-02 1.43137231e-01 -1.16452825e+00
6.22438528e-02 -2.61568636e-01 1.95153549e-01 1.11105072e+00
7.42923915e-01 -6.88808978e-01 5.46649337e-01 3.37742805e-01
-6.90248385e-02 -8.74475658e-01 -1.34721470e+00 -6.86886549e-01
2.67484903e-01 -2.56926388e-01 2.65117377e-01 1.03615260e+00
1.14636846e-01 6.19456351e-01 -6.70427307e-02 -1.35428339e-01
1.78811699e-01 -6.48938492e-03 4.07744378e-01 -1.50558162e+00
-2.98240304e-01 -1.54887781e-01 -4.63754326e-01 -7.03870118e-01
3.30628932e-01 -1.19410324e+00 7.28628635e-02 -1.54394662e+00
1.84675276e-01 -3.82639617e-01 -5.53295255e-01 6.25160575e-01
-6.71447963e-02 -8.27560797e-02 -7.31706545e-02 -6.77061975e-02
-6.20005906e-01 2.75519043e-01 8.99749875e-01 -8.71674642e-02
-1.01773903e-01 -8.92851278e-02 -4.45719063e-01 5.63076973e-01
6.45013332e-01 -7.56137490e-01 -1.80691138e-01 -2.20285892e-01
2.62298703e-01 7.93094486e-02 6.49574324e-02 -1.09318590e+00
2.03176335e-01 4.61655334e-02 1.94810465e-01 -5.25511682e-01
1.83263108e-01 -8.64151418e-01 1.32776171e-01 7.22594738e-01
-3.93176675e-01 -2.30283335e-01 3.65420938e-01 4.17447448e-01
-3.16113561e-01 -4.21993613e-01 9.54222381e-01 -1.52378738e-01
-6.57448411e-01 1.16904430e-01 -4.29178357e-01 2.73213208e-01
8.28411102e-01 1.37127787e-01 -1.44565869e-02 3.94811541e-01
-9.65606928e-01 -1.56475864e-02 -1.59338266e-01 3.47561598e-01
2.75799125e-01 -7.29468822e-01 -1.04865229e+00 -9.81003568e-02
3.43887746e-01 -3.09106827e-01 3.96778435e-01 8.69457424e-01
-4.21170384e-01 1.05244577e+00 -1.27788752e-01 -5.07730722e-01
-1.01391923e+00 7.43924737e-01 4.07009751e-01 -1.01592040e+00
-5.80664873e-01 7.23165512e-01 2.14618891e-01 -5.36390722e-01
-6.23431280e-02 -3.61272007e-01 -5.86478531e-01 2.28509232e-01
5.29558778e-01 1.08762212e-01 5.51944792e-01 -4.45754945e-01
-7.50569820e-01 1.66945651e-01 -1.10741705e-01 6.96161166e-02
1.59802115e+00 3.10589463e-01 -1.77548885e-01 3.89719516e-01
1.20798147e+00 -1.60393849e-01 -3.89715254e-01 -2.98210859e-01
7.40143180e-01 2.65597850e-01 -7.50342384e-02 -1.17940366e+00
-7.40358949e-01 8.84616911e-01 5.38885176e-01 7.97061399e-02
1.03384161e+00 4.18760888e-02 5.64115167e-01 7.24359393e-01
5.06053269e-01 -1.01210272e+00 -2.67719984e-01 5.19376636e-01
9.46636379e-01 -9.97309446e-01 -5.18336855e-02 -2.27142811e-01
-4.50063884e-01 1.16990876e+00 2.12308034e-01 7.84064457e-02
6.31107271e-01 3.09707582e-01 -2.85451952e-02 -3.18899781e-01
-7.28454828e-01 -1.13170870e-01 2.80319303e-01 3.61744195e-01
7.48214841e-01 -6.49412796e-02 -9.77371633e-01 9.18931901e-01
2.53431618e-01 5.77985108e-01 6.48525283e-02 8.16616416e-01
-1.85598716e-01 -1.51371157e+00 -1.25068486e-01 3.48459661e-01
-9.95153308e-01 -2.21276611e-01 -3.99693489e-01 1.02891004e+00
3.04129511e-01 8.14109623e-01 -4.93460834e-01 3.03233117e-01
5.62105000e-01 5.81864893e-01 3.60020012e-01 -6.21146262e-01
-9.36987162e-01 -7.60044530e-02 3.21264327e-01 -4.68707919e-01
-7.22352207e-01 -5.25631070e-01 -1.34779894e+00 2.50694305e-01
-1.70711562e-01 5.75582504e-01 7.53236592e-01 1.22389627e+00
5.07066905e-01 5.21441996e-01 3.38922366e-02 -1.58497095e-01
-6.63088202e-01 -1.11001325e+00 -5.00915825e-01 4.09543633e-01
5.67631982e-02 -4.02803987e-01 -4.50054780e-02 1.39606342e-01] | [8.570206642150879, 8.70952033996582] |
973dcc40-a952-48a6-9524-db9aa6a69e91 | to-adapt-or-to-fine-tune-a-case-study-on | 2208.14559 | null | https://arxiv.org/abs/2208.14559v1 | https://arxiv.org/pdf/2208.14559v1.pdf | To Adapt or to Fine-tune: A Case Study on Abstractive Summarization | Recent advances in the field of abstractive summarization leverage pre-trained language models rather than train a model from scratch. However, such models are sluggish to train and accompanied by a massive overhead. Researchers have proposed a few lightweight alternatives such as smaller adapters to mitigate the drawbacks. Nonetheless, it remains uncertain whether using adapters benefits the task of summarization, in terms of improved efficiency without an unpleasant sacrifice in performance. In this work, we carry out multifaceted investigations on fine-tuning and adapters for summarization tasks with varying complexity: language, domain, and task transfer. In our experiments, fine-tuning a pre-trained language model generally attains a better performance than using adapters; the performance gap positively correlates with the amount of training data used. Notably, adapters exceed fine-tuning under extremely low-resource conditions. We further provide insights on multilinguality, model convergence, and robustness, hoping to shed light on the pragmatic choice of fine-tuning or adapters in abstractive summarization. | ['Pinzhen Chen', 'Zheng Zhao'] | 2022-08-30 | null | https://aclanthology.org/2022.ccl-1.73 | https://aclanthology.org/2022.ccl-1.73.pdf | ccl-2022-10 | ['pretrained-multilingual-language-models'] | ['natural-language-processing'] | [ 2.01743454e-01 1.61702916e-01 -2.95578748e-01 -1.77049696e-01
-1.22161329e+00 -6.70532525e-01 7.31636524e-01 4.10744846e-01
-7.96976388e-01 7.85748720e-01 7.30025947e-01 -5.57366788e-01
1.47601172e-01 -3.23833674e-01 -6.64514959e-01 -2.37858489e-01
2.15670481e-01 4.41039920e-01 -6.19432069e-02 -3.51619840e-01
4.22306299e-01 1.21477313e-01 -1.29191148e+00 2.44310528e-01
1.28582084e+00 4.11039770e-01 3.59573722e-01 6.93997264e-01
-1.81704551e-01 4.65475887e-01 -8.43592405e-01 -5.87599397e-01
-2.41283923e-02 -5.17955124e-01 -9.98306274e-01 -1.61258996e-01
5.70026338e-01 -2.93558747e-01 -1.75580606e-02 7.01796651e-01
7.26792216e-01 -3.60443033e-02 7.64594674e-01 -7.48450875e-01
-5.52163541e-01 8.32485855e-01 -3.70705783e-01 4.03437674e-01
3.44648778e-01 2.52869129e-01 1.16992545e+00 -4.86566484e-01
5.05083859e-01 1.03083670e+00 8.31046939e-01 5.56205392e-01
-1.29078019e+00 -3.72233242e-01 1.93486378e-01 -1.39458910e-01
-1.09007132e+00 -1.01254010e+00 2.84042597e-01 -2.96953440e-01
1.39920580e+00 3.96041363e-01 3.30373257e-01 1.16478348e+00
2.01440006e-01 7.31963277e-01 7.70087957e-01 -4.36717868e-01
-2.88996138e-02 2.76964039e-01 1.38552375e-02 5.32593012e-01
6.12851083e-01 -4.11216319e-01 -6.56982720e-01 -2.77263075e-01
3.73800129e-01 -5.74899733e-01 -3.04773450e-01 -1.87296886e-02
-1.29673576e+00 9.03036594e-01 1.62717342e-01 3.41808647e-01
-3.37582141e-01 -3.37758958e-02 8.81590486e-01 6.10832691e-01
6.83261037e-01 1.07948470e+00 -6.95523202e-01 -4.29426253e-01
-1.32329321e+00 1.53127328e-01 9.92700577e-01 9.76651073e-01
5.46470225e-01 1.28593460e-01 -2.00644240e-01 9.94526863e-01
-1.39675096e-01 2.65844554e-01 6.72672153e-01 -8.21359932e-01
8.95030916e-01 4.05526996e-01 -1.00860469e-01 -4.97815758e-01
-4.66281354e-01 -5.36601007e-01 -7.52248824e-01 -4.28760409e-01
4.46357161e-01 -4.52006251e-01 -5.52378714e-01 1.84421408e+00
-1.28214151e-01 -3.53657007e-01 2.27489293e-01 5.76113880e-01
7.73168504e-01 6.29067183e-01 1.24613427e-01 -3.34010750e-01
1.35574460e+00 -1.15806413e+00 -3.85011554e-01 -5.90741873e-01
1.14570034e+00 -1.05709708e+00 1.54799080e+00 1.82657540e-01
-1.38827431e+00 -2.85521775e-01 -1.03019071e+00 -2.40531951e-01
3.14044915e-02 1.75846592e-01 5.81915319e-01 6.61871672e-01
-1.09447813e+00 7.55804479e-01 -8.19038332e-01 -7.12785780e-01
6.25219420e-02 2.51769722e-01 -4.05942589e-01 1.43008143e-01
-8.83225977e-01 1.26707792e+00 4.08002406e-01 -2.90802062e-01
-2.83882409e-01 -8.25030088e-01 -8.26927543e-01 1.66951582e-01
2.82965899e-01 -1.34380078e+00 1.51183867e+00 -9.50424612e-01
-1.66784716e+00 8.18993807e-01 -1.24034844e-01 -5.48365176e-01
5.91164291e-01 -3.76730770e-01 -2.17523705e-02 -1.90643847e-01
1.26334473e-01 4.48716730e-01 6.43414199e-01 -7.62374461e-01
-4.47771102e-01 -8.49034339e-02 9.70580652e-02 5.64104676e-01
-5.32674968e-01 2.45901659e-01 -2.82455325e-01 -6.54957294e-01
-3.50024283e-01 -8.29651713e-01 -2.34042034e-01 -6.32940412e-01
-3.81843597e-01 -2.53495514e-01 2.65933037e-01 -7.78644264e-01
1.43667316e+00 -1.88832545e+00 1.78066686e-01 -4.37026739e-01
-4.46931981e-02 5.35814822e-01 -4.23044235e-01 8.60785067e-01
3.69721860e-01 2.34252185e-01 -3.25999498e-01 -4.29407001e-01
3.12602557e-02 4.13960703e-02 -3.52271169e-01 1.64228871e-01
2.30195552e-01 9.18113768e-01 -8.84922683e-01 -5.38354278e-01
-2.03994825e-01 2.25176096e-01 -8.98633659e-01 1.26390249e-01
-2.58280367e-01 3.57560813e-01 -4.12855476e-01 3.66510719e-01
2.78272897e-01 -1.77437648e-01 2.28425011e-01 4.79254872e-02
-1.92116514e-01 8.98693502e-01 -7.15379953e-01 1.88941312e+00
-7.99463630e-01 6.44661665e-01 9.82029364e-02 -9.36155498e-01
6.09288573e-01 3.14873397e-01 5.87357134e-02 -7.11911619e-01
-1.55179435e-02 4.92499650e-01 1.70368016e-01 -6.04943395e-01
1.00300527e+00 -3.18413436e-01 -2.22858191e-01 6.50598407e-01
1.24446474e-01 -3.29916775e-01 3.65494102e-01 3.04587483e-01
1.14024544e+00 1.60113275e-02 4.05018210e-01 -4.50494647e-01
2.95857221e-01 4.62424047e-02 1.97585493e-01 7.86630988e-01
1.42749310e-01 5.02453566e-01 4.90256608e-01 8.43037516e-02
-1.12283635e+00 -7.75222778e-01 -2.85334396e-03 1.44490588e+00
-3.01778346e-01 -7.21782446e-01 -9.40667868e-01 -6.42714322e-01
-1.41103461e-01 1.06476927e+00 -1.01146430e-01 -3.32318246e-01
-6.94407940e-01 -8.95157218e-01 1.01782370e+00 4.49752748e-01
2.88159698e-01 -9.31663275e-01 -6.47100925e-01 2.10132569e-01
-5.04535913e-01 -1.07605493e+00 -7.71842480e-01 3.69289666e-02
-1.21170139e+00 -5.91933846e-01 -7.92228222e-01 -6.06936157e-01
4.67257559e-01 3.97471279e-01 1.41398621e+00 2.06091385e-02
2.91472703e-01 2.94974118e-01 -2.12302089e-01 -4.74760920e-01
-7.48745859e-01 1.01474917e+00 7.40069821e-02 -6.19329929e-01
1.13848247e-01 -5.96994638e-01 -5.18003523e-01 -2.71305186e-03
-8.42531383e-01 7.82170817e-02 1.02226794e+00 7.85005867e-01
-1.92318987e-02 -4.90501791e-01 9.56916869e-01 -9.88145113e-01
1.23780918e+00 -4.02938843e-01 -5.09999208e-02 3.41987967e-01
-7.08967209e-01 4.94726658e-01 8.02412033e-01 -3.16802830e-01
-1.12803483e+00 -4.18948084e-01 -2.16270238e-01 1.54890582e-01
8.26539621e-02 8.22115004e-01 1.08189769e-01 3.65810335e-01
8.16100359e-01 1.87038034e-01 9.11486372e-02 -6.39227152e-01
5.72289824e-01 7.52356708e-01 3.86843801e-01 -8.07509542e-01
4.32359010e-01 4.35570441e-02 -6.00370824e-01 -1.00714481e+00
-8.95369112e-01 -3.31430525e-01 -4.00377780e-01 3.05545807e-01
4.84040290e-01 -1.10247362e+00 -1.55259445e-02 1.01305842e-01
-1.11349249e+00 -5.37680030e-01 -2.23483935e-01 3.98508459e-01
-6.40437186e-01 5.81311226e-01 -7.03106046e-01 -3.03304940e-01
-8.82919729e-01 -1.02283239e+00 1.08719528e+00 -4.30239290e-02
-7.96086550e-01 -1.12814963e+00 1.19441405e-01 6.01308942e-01
6.58767223e-01 -2.01980218e-01 1.14211214e+00 -8.22453916e-01
-2.54070699e-01 -1.80459797e-01 -2.43248224e-01 4.02505606e-01
3.96669321e-02 -2.65236315e-03 -7.43594825e-01 -5.33218980e-01
-4.42969762e-02 -4.68191743e-01 8.15934241e-01 3.21664393e-01
7.10581958e-01 -6.67734563e-01 -7.33479112e-02 5.10852993e-01
1.13739812e+00 -4.29732233e-01 3.75418007e-01 4.70930517e-01
4.67315257e-01 6.50482178e-01 4.08291340e-01 2.93108404e-01
5.52330315e-01 6.23103321e-01 -9.09955725e-02 1.58745632e-01
-2.24869132e-01 -3.39991838e-01 5.39343834e-01 1.37178326e+00
-1.91620022e-01 -2.44857416e-01 -8.76961410e-01 5.59554458e-01
-1.66106868e+00 -8.31359565e-01 1.45211935e-01 2.34960628e+00
1.10666168e+00 2.31914908e-01 2.46841714e-01 -2.24067107e-01
3.81719738e-01 3.11210245e-01 -3.68092120e-01 -8.39103401e-01
-7.20557570e-02 7.25670159e-03 4.27287817e-01 4.62332934e-01
-5.69944441e-01 1.07090747e+00 6.69536018e+00 8.58315289e-01
-1.42933989e+00 1.55189723e-01 3.04690361e-01 -3.45867038e-01
-4.80815113e-01 7.89650455e-02 -7.72734463e-01 4.53917474e-01
1.30278146e+00 -4.74548459e-01 3.81152123e-01 6.23384655e-01
2.68866926e-01 -3.81506570e-02 -1.29323256e+00 5.18421352e-01
2.45185137e-01 -1.27771366e+00 2.49340698e-01 1.09447362e-02
6.68944180e-01 4.24742937e-01 -1.36193007e-01 6.31910682e-01
1.14320286e-01 -8.87374103e-01 6.93377614e-01 8.37096423e-02
8.00977170e-01 -5.59370577e-01 6.60366297e-01 7.09982395e-01
-6.25215530e-01 1.40856177e-01 -3.08381975e-01 -2.26546958e-01
3.17749530e-01 3.18840921e-01 -9.47984159e-01 6.29964054e-01
2.11299971e-01 3.46983790e-01 -6.96323693e-01 7.59135008e-01
2.45712269e-02 7.29766011e-01 -2.19979897e-01 -1.35250734e-02
4.84807462e-01 -1.20192215e-01 6.05931997e-01 1.73959124e+00
2.66776115e-01 -1.83500379e-01 -5.50813004e-02 4.22097951e-01
-1.60630345e-01 4.39008057e-01 -6.43834889e-01 -1.85613737e-01
5.00549614e-01 1.04351342e+00 -3.51071656e-01 -4.96853918e-01
-4.82612044e-01 9.27430689e-01 6.78784013e-01 2.41936177e-01
-6.76227331e-01 -2.75353223e-01 4.22572762e-01 2.64151573e-01
1.28631383e-01 -4.49979901e-01 -4.82030481e-01 -1.52666938e+00
1.13306813e-01 -1.18237400e+00 4.06312168e-01 -3.99472982e-01
-1.09465826e+00 6.76461637e-01 1.67395115e-01 -9.63293493e-01
-4.47891176e-01 -1.68779612e-01 -7.51288295e-01 7.91535974e-01
-1.45589554e+00 -1.05136740e+00 5.32195382e-02 1.87050015e-01
8.58719587e-01 -5.59305027e-02 7.82794774e-01 3.04065466e-01
-7.56013155e-01 8.79479825e-01 3.52250397e-01 -3.01154405e-01
1.09769046e+00 -1.03633571e+00 5.52563965e-01 7.44277358e-01
1.04707256e-01 8.16854298e-01 9.37486231e-01 -4.22762990e-01
-1.29898822e+00 -9.75577116e-01 1.30438280e+00 -6.18856847e-01
7.21980155e-01 -1.39798597e-01 -9.42822635e-01 7.55518377e-01
5.67927718e-01 -7.33884752e-01 6.91232622e-01 5.61586201e-01
-4.29851323e-01 1.08436674e-01 -7.51403451e-01 8.90516639e-01
1.07394242e+00 -5.68336427e-01 -8.72399807e-01 3.07044983e-01
7.80928910e-01 -2.11959347e-01 -9.42886829e-01 2.76678860e-01
5.32449782e-01 -8.93025339e-01 8.15751195e-01 -7.29450405e-01
5.61133862e-01 2.32275277e-01 -7.34860525e-02 -1.66133320e+00
-2.27156490e-01 -6.36699200e-01 5.06183617e-02 1.33626986e+00
7.11058557e-01 -7.58145273e-01 5.39676011e-01 5.51696718e-01
-4.77437526e-01 -8.07268441e-01 -6.20488107e-01 -9.47266102e-01
5.24243712e-01 -1.07055150e-01 4.14814532e-01 9.20428693e-01
2.52007633e-01 1.00850308e+00 -1.87137485e-01 -2.28574306e-01
2.64993876e-01 -4.11927328e-02 9.81240034e-01 -8.89301360e-01
-3.05184066e-01 -7.78309166e-01 1.08597472e-01 -1.18399036e+00
2.34902665e-01 -1.04187536e+00 -1.09144032e-01 -1.68561423e+00
2.89718330e-01 -3.94875139e-01 2.00570989e-02 2.62918502e-01
-4.75283593e-01 1.91723946e-02 2.98442096e-01 4.31582004e-01
-6.92121685e-01 5.35766602e-01 9.99276638e-01 -5.64677007e-02
-4.38335389e-01 1.03002355e-01 -1.21205187e+00 6.45793736e-01
1.01034331e+00 -2.95814365e-01 -5.37532628e-01 -1.06531382e+00
1.67462021e-01 1.55470431e-01 -1.62405580e-01 -7.09437430e-01
1.96864620e-01 5.35676442e-02 -2.27838635e-01 -1.97459236e-01
1.65252581e-01 -2.09270269e-01 -5.93184717e-02 3.42913479e-01
-5.22716999e-01 3.29378992e-01 3.42938185e-01 3.44804645e-01
-1.19347885e-01 -4.62791026e-01 5.97803414e-01 -1.76500171e-01
-1.33582830e-01 -8.56565312e-02 -4.94009435e-01 5.44723928e-01
4.72998291e-01 -2.49662027e-01 -3.40490758e-01 -3.59458566e-01
-2.67889440e-01 7.39507154e-02 6.58982158e-01 5.07992685e-01
9.22161564e-02 -8.59333098e-01 -1.05022514e+00 7.13425595e-03
1.54039741e-01 -5.62940687e-02 1.79856017e-01 1.01212299e+00
-4.42923278e-01 7.21145630e-01 -1.08179532e-01 -4.89955753e-01
-1.19416714e+00 3.18189323e-01 4.04650578e-03 -7.01413214e-01
-5.84933579e-01 6.72145545e-01 8.24165866e-02 -5.90561688e-01
1.35686815e-01 -2.76550770e-01 1.05931774e-01 2.32087612e-01
2.71546632e-01 4.36343253e-01 3.98005754e-01 -3.68835688e-01
-2.25163221e-01 3.07915866e-01 -4.48538095e-01 -1.65947124e-01
1.16933572e+00 -1.90854505e-01 -1.41069368e-02 3.78193378e-01
1.02281654e+00 2.02719435e-01 -1.06838846e+00 -3.20357710e-01
1.72291100e-01 -1.00557506e-01 -1.28243729e-01 -6.77938581e-01
-5.47599554e-01 7.33953416e-01 -2.56878674e-01 2.23170027e-01
9.37847733e-01 2.17786636e-02 9.48340893e-01 5.74447215e-01
2.14338318e-01 -1.14373064e+00 3.50655466e-02 6.35104239e-01
1.04967201e+00 -1.12888932e+00 2.61084527e-01 -3.47434368e-04
-8.86023521e-01 8.01728487e-01 4.69055831e-01 1.84702110e-02
9.67490152e-02 -1.44061819e-02 3.84790823e-02 -1.78830903e-02
-1.10851610e+00 2.70952471e-02 1.99398294e-01 2.97693938e-01
8.23176444e-01 2.78093927e-02 -5.79770327e-01 5.61322451e-01
-7.65748084e-01 -2.33476639e-01 5.28391600e-01 6.84503317e-01
-4.89300609e-01 -9.67007935e-01 -1.50057718e-01 7.08816528e-01
-5.66308796e-01 -4.35343862e-01 -3.14347595e-01 7.83529878e-01
-3.32736015e-01 9.48778808e-01 2.04699133e-02 -1.95684023e-02
5.41155398e-01 1.79083288e-01 6.82716727e-01 -8.64834130e-01
-8.57115269e-01 -2.05862135e-01 6.90191805e-01 -1.78383932e-01
-2.10391209e-01 -7.48078048e-01 -7.88761616e-01 -6.15027070e-01
-4.92699832e-01 2.85763025e-01 5.80953360e-01 9.97164547e-01
6.89004898e-01 3.63594413e-01 2.59471595e-01 -7.80689359e-01
-1.12958264e+00 -1.25128412e+00 -6.50359690e-02 1.84134901e-01
2.15672791e-01 -1.19104117e-01 -2.22629979e-01 -9.77444649e-03] | [11.624166488647461, 9.06070613861084] |
a79d64fd-2b7d-429a-b8d0-d079e0dc73f1 | decoder-only-or-encoder-decoder-interpreting | 2304.04052 | null | https://arxiv.org/abs/2304.04052v1 | https://arxiv.org/pdf/2304.04052v1.pdf | Decoder-Only or Encoder-Decoder? Interpreting Language Model as a Regularized Encoder-Decoder | The sequence-to-sequence (seq2seq) task aims at generating the target sequence based on the given input source sequence. Traditionally, most of the seq2seq task is resolved by the Encoder-Decoder framework which requires an encoder to encode the source sequence and a decoder to generate the target text. Recently, a bunch of new approaches have emerged that apply decoder-only language models directly to the seq2seq task. Despite the significant advancements in applying language models to the seq2seq task, there is still a lack of thorough analysis on the effectiveness of the decoder-only language model architecture. This paper aims to address this gap by conducting a detailed comparison between the encoder-decoder architecture and the decoder-only language model framework through the analysis of a regularized encoder-decoder structure. This structure is designed to replicate all behaviors in the classical decoder-only language model but has an encoder and a decoder making it easier to be compared with the classical encoder-decoder structure. Based on the analysis, we unveil the attention degeneration problem in the language model, namely, as the generation step number grows, less and less attention is focused on the source sequence. To give a quantitative understanding of this problem, we conduct a theoretical sensitivity analysis of the attention output with respect to the source input. Grounded on our analysis, we propose a novel partial attention language model to solve the attention degeneration problem. Experimental results on machine translation, summarization, and data-to-text generation tasks support our analysis and demonstrate the effectiveness of our proposed model. | ['Nigel Collier', 'Zhiyuan Liu', 'Shengding Hu', 'Anthony Man-Cho So', 'Qian Yu', 'Wai Lam', 'Zihao Fu'] | 2023-04-08 | null | null | null | null | ['data-to-text-generation'] | ['natural-language-processing'] | [ 6.36879563e-01 4.92430955e-01 -1.32037133e-01 -7.77618140e-02
-7.45847821e-01 -4.67159867e-01 6.88874185e-01 -1.17023075e-02
-8.07165578e-02 8.28935266e-01 5.91797650e-01 -5.56702137e-01
3.66129756e-01 -6.32536352e-01 -8.11784863e-01 -4.62398142e-01
4.59247589e-01 3.46680582e-01 -9.02231559e-02 -5.54772079e-01
2.76725322e-01 -5.29762730e-02 -1.34274375e+00 4.36385274e-01
1.10094106e+00 5.21408737e-01 7.60866582e-01 7.18543708e-01
-3.74657422e-01 9.07227755e-01 -7.42119312e-01 -4.84661192e-01
1.12610832e-01 -1.17134988e+00 -8.79841447e-01 -2.08251923e-01
-7.37179145e-02 -2.91498393e-01 -1.59991115e-01 1.04850411e+00
7.31469870e-01 -1.90818414e-01 5.89904308e-01 -1.09594655e+00
-1.08393216e+00 9.93277907e-01 -3.06098998e-01 2.53616542e-01
5.39153576e-01 1.98820412e-01 1.11456728e+00 -8.10166657e-01
6.23870432e-01 1.25271964e+00 3.82839829e-01 7.44809151e-01
-8.64481211e-01 -3.16804498e-01 1.13971390e-01 -3.76517922e-02
-1.11330426e+00 -4.05276239e-01 5.32439530e-01 -3.30906659e-01
1.20412362e+00 1.89288795e-01 5.06145597e-01 1.23884666e+00
5.53584397e-01 7.79500544e-01 6.72424138e-01 -6.48056448e-01
2.95198150e-03 1.49413213e-01 6.00413308e-02 2.98993856e-01
2.00654626e-01 1.51288882e-01 -4.63275015e-01 2.82451779e-01
6.68563426e-01 -4.05877024e-01 -3.24609965e-01 8.12979415e-02
-1.07404518e+00 7.99731195e-01 1.79811448e-01 3.91785800e-01
-4.31868136e-01 9.67583507e-02 5.83225667e-01 3.46010000e-01
5.61179638e-01 5.07443130e-01 -2.05293611e-01 -3.90299588e-01
-1.04075706e+00 2.16755942e-01 8.02882969e-01 1.21765006e+00
4.87559289e-01 3.05382162e-01 -5.58917820e-01 6.22504532e-01
1.29722863e-01 3.94371867e-01 9.31373596e-01 -3.64735156e-01
8.59816015e-01 6.33603573e-01 -3.03954184e-02 -7.00565636e-01
-8.32073763e-02 -5.47965884e-01 -8.79183650e-01 -2.07055926e-01
1.24505460e-01 -3.83245498e-01 -7.38173664e-01 1.91054201e+00
-1.13338619e-01 -4.38488647e-02 3.32762152e-01 8.94442260e-01
7.11075544e-01 9.46849644e-01 -1.84792414e-01 -4.20763314e-01
1.15742707e+00 -1.17133045e+00 -9.20804739e-01 -2.16767818e-01
8.64755750e-01 -6.58082008e-01 1.20473301e+00 -8.21446627e-02
-1.32199717e+00 -7.18146145e-01 -9.95387435e-01 -3.21619481e-01
-5.99153601e-02 6.70250505e-02 -6.40860759e-04 5.44962943e-01
-1.16077209e+00 4.20982897e-01 -5.82477450e-01 -4.30963069e-01
-8.20049495e-02 1.25588432e-01 -7.36660808e-02 1.94978952e-01
-1.50337827e+00 9.70151246e-01 5.92959940e-01 1.95327684e-01
-9.03079510e-01 -6.24808729e-01 -8.05524409e-01 3.28056335e-01
1.71209857e-01 -1.00363958e+00 1.48842764e+00 -1.42674577e+00
-1.76307929e+00 4.65188891e-01 -3.25378090e-01 -5.98389447e-01
4.64922249e-01 -2.94644773e-01 -1.22555390e-01 -1.07904874e-01
2.25267857e-02 4.06756729e-01 5.91008365e-01 -1.02729034e+00
-5.52830040e-01 -2.14409791e-02 4.77872379e-02 3.98180157e-01
-1.83486626e-01 1.15856834e-01 -3.74528617e-01 -8.95532250e-01
-6.07960463e-01 -8.03245842e-01 -1.41058341e-01 -7.15820909e-01
-6.09016120e-01 -5.98626025e-02 4.43417668e-01 -7.44594455e-01
1.72966886e+00 -1.97665763e+00 5.20659447e-01 -2.63789117e-01
-6.76983446e-02 4.96897280e-01 -5.33449233e-01 9.98113751e-01
-3.20173264e-01 2.67347336e-01 -4.46690977e-01 -3.22346210e-01
-1.88206136e-01 7.07162246e-02 -7.11383045e-01 -5.01278155e-02
4.26994652e-01 1.19676721e+00 -1.04659379e+00 -2.36462936e-01
-1.94170371e-01 3.41786087e-01 -6.28782690e-01 5.47073960e-01
-2.86797941e-01 3.30597550e-01 -3.37437510e-01 2.35065654e-01
3.24028492e-01 -2.31936842e-01 2.44079694e-01 6.30031899e-02
-3.77208352e-01 6.11348987e-01 -6.84525192e-01 1.60658181e+00
-4.26330119e-01 6.17402077e-01 -3.28214914e-02 -9.95625794e-01
9.54274416e-01 5.36819041e-01 9.72739831e-02 -7.75304914e-01
1.57066926e-01 2.86492229e-01 3.75265449e-01 -5.76116920e-01
7.87680268e-01 -4.24754649e-01 -1.27361104e-01 6.29000902e-01
6.11922927e-02 5.47888204e-02 4.65641707e-01 2.51060188e-01
9.82931674e-01 2.53033966e-01 5.28978288e-01 -2.78354049e-01
6.78634644e-01 6.70572137e-03 4.30570006e-01 6.92992389e-01
1.67486519e-01 7.01013803e-01 6.95410192e-01 -2.06328817e-02
-1.32463586e+00 -6.92852020e-01 4.03141141e-01 1.03517282e+00
3.75021249e-02 -6.47809982e-01 -1.15797484e+00 -5.95499873e-01
-3.86316359e-01 9.97393310e-01 -5.05070925e-01 -4.03421342e-01
-6.70281947e-01 -7.44824111e-01 8.97079885e-01 4.84266520e-01
3.01567256e-01 -1.21275198e+00 -7.35894024e-01 3.79607886e-01
-4.21265721e-01 -9.11581695e-01 -7.59098589e-01 -1.25739753e-01
-7.72613645e-01 -7.67561674e-01 -8.58760953e-01 -7.79055595e-01
4.88597572e-01 9.89882797e-02 1.05262911e+00 -1.42975468e-02
9.00782123e-02 1.08748376e-01 -6.18033886e-01 -4.79788065e-01
-1.13320470e+00 5.30772388e-01 -6.23593070e-02 8.78801290e-03
2.59310585e-02 -4.08720046e-01 -3.40834111e-01 9.16509330e-03
-1.18269253e+00 6.45493627e-01 9.54955757e-01 7.68564105e-01
1.81392610e-01 -2.81093448e-01 1.02279592e+00 -7.13237762e-01
1.21446502e+00 -6.97034180e-01 -3.63169014e-01 3.22791219e-01
-6.26134098e-01 4.32356268e-01 1.07080770e+00 -2.72720128e-01
-1.04645562e+00 -2.19334915e-01 -3.84519666e-01 -1.67288065e-01
2.04806089e-01 8.27751338e-01 -2.73983836e-01 5.99970996e-01
3.70961070e-01 6.84428751e-01 2.71530360e-01 -5.09159029e-01
4.23041821e-01 9.23854530e-01 3.93770725e-01 -3.78303528e-01
5.46466231e-01 -9.86182243e-02 -3.34736794e-01 -7.21014678e-01
-5.50451696e-01 1.42132957e-02 -4.87312794e-01 -6.60601631e-02
7.88495541e-01 -7.65144765e-01 -3.90517741e-01 3.13552886e-01
-1.57744431e+00 -4.58391130e-01 -3.36135119e-01 2.44053185e-01
-7.79302835e-01 4.36115175e-01 -5.43503344e-01 -8.40662122e-01
-6.08498335e-01 -1.29809499e+00 1.06103849e+00 -2.55403612e-02
-3.61826003e-01 -1.08684087e+00 2.29367480e-01 8.38426594e-03
5.11599243e-01 -4.94818278e-02 1.15002251e+00 -7.76484728e-01
-4.40686643e-01 1.41841918e-01 -1.55896708e-01 4.34956670e-01
1.92612857e-01 -2.44925082e-01 -6.06229067e-01 -1.85569152e-01
2.20550492e-01 -3.72679904e-02 7.14841425e-01 8.89953524e-02
8.05026293e-01 -6.56085968e-01 -8.25095475e-02 4.06437755e-01
1.33682859e+00 2.59929359e-01 7.61460960e-01 8.28098878e-02
6.69983089e-01 6.37371838e-01 3.24482113e-01 3.01265806e-01
4.42944467e-01 7.29044557e-01 3.52812439e-01 -4.98024374e-03
-3.76425803e-01 -6.94282532e-01 8.89748275e-01 1.28286433e+00
9.30181965e-02 -7.67813265e-01 -8.64820063e-01 6.52867019e-01
-1.81666756e+00 -1.01380217e+00 -2.07633585e-01 2.16725802e+00
1.06848049e+00 -4.07212637e-02 5.62061518e-02 1.28198266e-02
7.96990693e-01 1.53947964e-01 -3.44981164e-01 -8.69891226e-01
-2.46779574e-03 4.13854383e-02 2.96657801e-01 7.03137696e-01
-4.52435613e-01 9.29019272e-01 7.04280853e+00 8.47980976e-01
-1.27057767e+00 9.61125456e-03 3.88521671e-01 -1.02108561e-01
-5.56822240e-01 3.18956608e-03 -8.52821052e-01 7.68814564e-01
1.33985603e+00 -7.87620962e-01 4.78455067e-01 4.48621541e-01
6.83017373e-01 1.29537567e-01 -1.46095002e+00 6.76208913e-01
2.70624369e-01 -1.28665423e+00 7.20276237e-01 3.15137245e-02
5.62512696e-01 -1.88377276e-01 -1.46671265e-01 4.33975905e-01
8.31519142e-02 -1.11538303e+00 8.88354361e-01 4.12416846e-01
9.12977099e-01 -5.58297098e-01 7.43628621e-01 7.66661406e-01
-1.00188029e+00 -2.22491637e-01 -1.74193993e-01 -4.27350760e-01
6.18381321e-01 4.53319311e-01 -9.78443801e-01 1.01160550e+00
8.48285854e-02 6.57783091e-01 -2.02226579e-01 5.80070853e-01
-2.46774450e-01 6.01146460e-01 2.58579701e-01 2.27569649e-03
3.14459592e-01 -2.59461880e-01 7.11856067e-01 1.41390800e+00
6.89730406e-01 -1.15838982e-01 -1.30297050e-01 1.10939848e+00
-1.37337834e-01 2.28256643e-01 -7.43726671e-01 -4.15144116e-01
2.67892212e-01 8.20586681e-01 -2.76254594e-01 -4.13899750e-01
-3.86291832e-01 1.13412249e+00 3.39617223e-01 4.66799229e-01
-8.69647563e-01 -3.79506946e-01 5.83503425e-01 2.55706638e-01
2.20318422e-01 -1.89417377e-01 -3.97402823e-01 -1.01678944e+00
-7.40482509e-02 -1.16396070e+00 -1.12556681e-01 -7.77730465e-01
-9.68905151e-01 7.80985773e-01 8.04706067e-02 -1.12426817e+00
-7.34874606e-01 -2.02408403e-01 -6.60535932e-01 1.27105904e+00
-1.52526891e+00 -8.32329392e-01 7.28018805e-02 1.45793483e-01
9.70901191e-01 -1.41454831e-01 6.36991382e-01 2.90840000e-01
-6.83789253e-01 7.23773003e-01 6.11315370e-02 6.50427416e-02
3.94240648e-01 -9.24481213e-01 7.98684418e-01 1.02528083e+00
-2.94458091e-01 7.82412231e-01 7.86515057e-01 -7.71902978e-01
-1.49322975e+00 -1.28848815e+00 1.33896267e+00 -4.46901858e-01
4.75706875e-01 -4.71070230e-01 -8.71708393e-01 6.61962926e-01
6.32305920e-01 -7.31990874e-01 5.40544152e-01 -4.53472823e-01
8.51612315e-02 1.69017002e-01 -5.82437575e-01 7.36412823e-01
1.10849929e+00 -4.48974520e-01 -8.72447312e-01 -4.38624844e-02
1.05159724e+00 -3.82976741e-01 -4.73676205e-01 1.99881598e-01
3.99087369e-01 -8.12556863e-01 4.96647090e-01 -6.02412164e-01
1.09878063e+00 -2.41846964e-01 1.56899728e-02 -1.55638385e+00
-3.26531619e-01 -8.77552509e-01 -1.80379063e-01 1.26429272e+00
5.43045342e-01 -4.60835278e-01 2.87486553e-01 1.26826867e-01
-4.94461745e-01 -7.12092280e-01 -8.96570623e-01 -9.39198911e-01
2.98360825e-01 -1.87596664e-01 6.60606980e-01 5.46845555e-01
1.59915268e-01 8.04975867e-01 -5.30990481e-01 -1.78086877e-01
5.10653034e-02 5.45885451e-02 6.94731057e-01 -7.62946665e-01
-3.86560172e-01 -5.70133984e-01 1.36246219e-01 -1.47242403e+00
1.39389545e-01 -1.23409510e+00 2.30593190e-01 -1.85688663e+00
3.46252382e-01 8.69912878e-02 3.28218825e-02 3.05497795e-01
-4.17336047e-01 -9.20858905e-02 4.61332023e-01 3.58879536e-01
-2.09717035e-01 8.70009005e-01 1.22488451e+00 9.50242579e-02
-2.43007824e-01 -1.84125662e-01 -1.12893879e+00 1.89823478e-01
7.26006687e-01 -3.84916782e-01 -5.57595611e-01 -7.20370829e-01
5.37363887e-01 2.99957901e-01 -2.68991035e-03 -6.40944541e-01
9.46923718e-03 -9.67679694e-02 -2.77740687e-01 -3.15626830e-01
-8.00140426e-02 -4.92806464e-01 7.63732865e-02 5.48801064e-01
-7.29885936e-01 4.15678173e-01 1.52943999e-01 3.73902977e-01
-3.86371225e-01 -2.95859545e-01 4.82873857e-01 -5.80809601e-02
-3.45409751e-01 -6.36799037e-02 -7.07426727e-01 2.44901508e-01
8.83548617e-01 -1.89496338e-01 -2.99228042e-01 -5.71622729e-01
-3.44715327e-01 1.94020063e-01 4.03199524e-01 7.06243634e-01
4.82453763e-01 -1.38650441e+00 -1.06595337e+00 4.04937148e-01
-7.95852691e-02 -1.87199429e-01 5.55799082e-02 7.97865331e-01
-3.74661952e-01 8.58291209e-01 -2.07240075e-01 -2.26998553e-01
-9.80218351e-01 7.04666555e-01 3.06090832e-01 -5.18543661e-01
-3.93107444e-01 4.23184425e-01 5.29662013e-01 -2.42217228e-01
-9.14708897e-02 -4.61951822e-01 -1.72399685e-01 -6.86322898e-02
5.76134026e-01 3.88040662e-01 -1.28734544e-01 -7.03651726e-01
-1.05601676e-01 2.20183074e-01 -9.50308982e-03 -2.17897117e-01
9.38024402e-01 -2.55894095e-01 -2.05879942e-01 5.26803732e-01
1.01271057e+00 -9.18835867e-03 -1.00253296e+00 1.00140855e-01
-2.92353239e-02 -5.58082387e-03 -3.15202236e-01 -7.11571693e-01
-6.67086899e-01 1.09347165e+00 -4.56466136e-04 3.50066185e-01
1.08319330e+00 -2.56952196e-01 8.91229093e-01 -6.22590631e-02
-3.46832015e-02 -1.08964217e+00 -8.34714738e-04 1.02813494e+00
1.15172935e+00 -6.94152415e-01 -4.57450539e-01 -3.02290380e-01
-7.60522723e-01 1.03941321e+00 5.61086535e-01 1.30176023e-01
1.30425230e-01 3.97207350e-01 -1.32139251e-01 1.95922405e-01
-1.09095836e+00 -9.72486883e-02 1.23254068e-01 4.23407823e-01
8.93378794e-01 -1.63807020e-01 -7.51515746e-01 7.98033357e-01
-4.25671309e-01 2.86271602e-01 7.95140147e-01 6.97207510e-01
-3.96852821e-01 -1.42156184e+00 -2.66766310e-01 7.02744350e-02
-4.72053826e-01 -5.57660639e-01 -6.02828324e-01 5.46090961e-01
-1.73369065e-01 9.87900972e-01 1.92182854e-01 -3.11949402e-01
3.70227277e-01 3.27391893e-01 2.30156586e-01 -8.44741821e-01
-8.72280598e-01 -1.27237082e-01 5.89904003e-02 -3.04927170e-01
-2.00201511e-01 -2.93098211e-01 -1.16720915e+00 -1.77614927e-01
-1.83352843e-01 2.93255299e-01 5.15732467e-01 8.96406114e-01
8.07104826e-01 9.11895871e-01 6.00496531e-01 -6.67415142e-01
-8.51815462e-01 -1.10950196e+00 -1.87230766e-01 2.38564610e-01
3.75149339e-01 -6.28921166e-02 -1.11680046e-01 2.02006295e-01] | [11.948740005493164, 9.133220672607422] |
5c12bfe3-d567-4da3-aded-66a004ca061d | style-controllable-speech-driven-gesture | null | null | https://diglib.eg.org/handle/10.1111/cgf13946 | https://onlinelibrary.wiley.com/doi/epdf/10.1111/cgf.13946 | Style-Controllable Speech-Driven Gesture Synthesis Using Normalising Flows | Automatic synthesis of realistic gestures promises to transform the fields of animation, avatars and communicative agents. In off‐line applications, novel tools can alter the role of an animator to that of a director, who provides only high‐level input for the desired animation; a learned network then translates these instructions into an appropriate sequence of body poses. In interactive scenarios, systems for generating natural animations on the fly are key to achieving believable and relatable characters. In this paper we address some of the core issues towards these ends. By adapting a deep learning‐based motion synthesis method called MoGlow, we propose a new generative model for generating state‐of‐the‐art realistic speech‐driven gesticulation. Owing to the probabilistic nature of the approach, our model can produce a battery of different, yet plausible, gestures given the same input speech signal. Just like humans, this gives a rich natural variation of motion. We additionally demonstrate the ability to exert directorial control over the output style, such as gesture level, speed, symmetry and spacial extent. Such control can be leveraged to convey a desired character personality or mood. We achieve all this without any manual annotation of the data. User studies evaluating upper‐body gesticulation confirm that the generated motions are natural and well match the input speech. Our method scores above all prior systems and baselines on these measures, and comes close to the ratings of the original recorded motions. We furthermore find that we can accurately control gesticulation styles without unnecessarily compromising perceived naturalness. Finally, we also demonstrate an application of the same method to full‐body gesticulation, including the synthesis of stepping motion and stance. | ['Taras Kucherenko', 'Gustav Eje Henter', 'Simon Alexanderson', 'Jonas Beskow'] | 2020-05-25 | null | null | null | computer-graphics-forum-2020-5 | ['normalising-flows', 'gesture-generation'] | ['methodology', 'robots'] | [ 6.62104860e-02 4.64737833e-01 -3.83492038e-02 -3.54250163e-01
-5.47130585e-01 -7.58830845e-01 1.10005915e+00 -8.69272709e-01
-1.59215461e-02 5.14840126e-01 6.65879965e-01 -7.87421390e-02
3.48493367e-01 -6.40404642e-01 -4.73908722e-01 -5.57181895e-01
9.38665420e-02 5.41374147e-01 1.84929855e-02 -7.13278294e-01
-4.34687734e-02 7.29164004e-01 -1.61997581e+00 3.53610933e-01
3.78301442e-01 3.80336195e-01 1.06793828e-01 1.27009475e+00
1.57829016e-01 6.57701671e-01 -7.78567672e-01 -1.90217614e-01
1.30601004e-01 -7.64912963e-01 -6.34342432e-01 1.49703816e-01
4.81262445e-01 -7.00383902e-01 -3.77288669e-01 4.78286147e-01
7.60821104e-01 2.98112154e-01 7.50748336e-01 -1.13168752e+00
-2.65986741e-01 7.61042416e-01 -2.66051918e-01 -4.19487298e-01
8.29573393e-01 8.31648290e-01 8.42478275e-01 -5.53894043e-01
9.82999921e-01 1.53068435e+00 5.47965288e-01 1.20316792e+00
-1.36742008e+00 -7.09647477e-01 -1.41255736e-01 -4.16350901e-01
-9.70785201e-01 -8.02922487e-01 6.61636949e-01 -4.20994192e-01
6.47366345e-01 6.64300859e-01 9.86272216e-01 1.76186216e+00
1.01113409e-01 6.53974593e-01 8.97127748e-01 -3.56246769e-01
4.29614745e-02 -2.95035005e-01 -6.65390670e-01 5.05061448e-01
-4.08715308e-01 3.27925056e-01 -7.09724247e-01 -3.80481631e-02
1.29857230e+00 -5.76238930e-01 -2.18850747e-01 -1.48054436e-01
-1.50107336e+00 4.70241964e-01 4.90945540e-02 2.58041888e-01
-3.66805524e-01 6.63878024e-01 1.85212642e-01 -5.41060604e-02
-1.31898031e-01 8.36747229e-01 -1.34351194e-01 -7.29609430e-01
-9.02436912e-01 7.78419495e-01 1.07000732e+00 9.44337726e-01
1.50575219e-02 5.97430825e-01 -4.71686482e-01 5.21389902e-01
2.54486531e-01 7.03350723e-01 3.39671165e-01 -1.56033850e+00
1.14805192e-01 4.92197983e-02 2.78562665e-01 -1.08521593e+00
-5.08081436e-01 -1.73128203e-01 -5.66807449e-01 8.70960832e-01
6.18163347e-01 -4.85271186e-01 -9.25892234e-01 2.04679990e+00
3.77911866e-01 -1.51223347e-01 -1.32251918e-01 1.07632756e+00
7.65502274e-01 6.40297115e-01 3.72001052e-01 9.78653431e-02
1.23705924e+00 -6.32288575e-01 -8.97378325e-01 -2.21666619e-01
3.55930358e-01 -9.47886884e-01 1.55833960e+00 4.49485540e-01
-1.42620325e+00 -5.45196533e-01 -9.48122263e-01 1.52727291e-01
3.31715643e-01 1.35304078e-01 7.49479651e-01 6.96772337e-01
-9.49232876e-01 8.84824157e-01 -9.96186316e-01 -2.52484888e-01
-1.35291085e-01 3.47191781e-01 -3.05490524e-01 6.27998233e-01
-9.84929264e-01 9.28395391e-01 -8.32529366e-03 -1.02212027e-01
-6.60136282e-01 -5.00278533e-01 -8.73977542e-01 -1.45150825e-01
7.54699809e-03 -8.82645845e-01 1.61971843e+00 -1.05552685e+00
-2.22847009e+00 6.60044134e-01 1.50467947e-01 1.57579593e-02
8.95058215e-01 -2.14880481e-01 -2.08420381e-01 1.04946665e-01
-1.08683519e-01 1.21245122e+00 8.89465511e-01 -1.38883877e+00
-2.49572977e-01 1.82278037e-01 2.57118586e-02 5.40897429e-01
-1.32714421e-01 -5.25647514e-02 -4.45109397e-01 -1.03643012e+00
-1.80614948e-01 -1.34604323e+00 -1.39993757e-01 2.13677034e-01
-4.90579069e-01 7.94077218e-02 1.06660247e+00 -6.90750301e-01
1.14354372e+00 -1.90423644e+00 5.48805594e-01 2.58565336e-01
1.03925243e-02 -4.14536484e-02 -3.67348641e-02 5.71830630e-01
3.68968621e-02 1.75592959e-01 2.08525583e-02 -4.38859433e-01
2.40758806e-01 2.11514473e-01 -2.20120132e-01 2.26524919e-01
-7.77158886e-02 1.04017103e+00 -9.28100944e-01 -4.21617985e-01
2.92538643e-01 6.60869896e-01 -6.49281323e-01 4.47566897e-01
-4.56487715e-01 9.97025669e-01 -2.51912773e-01 2.91567594e-01
1.28194662e-02 1.22566581e-01 3.58686954e-01 -2.09500223e-01
1.87288579e-02 3.34576339e-01 -1.21294749e+00 1.75130498e+00
-6.39857531e-01 7.90648580e-01 3.03284734e-01 -1.32880464e-01
8.90744388e-01 5.05513012e-01 4.21826035e-01 -2.97782242e-01
1.31128758e-01 4.80631590e-02 3.60044956e-01 -6.81838691e-01
5.83254933e-01 -2.84604162e-01 -3.41491252e-01 7.60918856e-01
-2.49618396e-01 -8.98826718e-01 -2.42571607e-01 2.91790534e-02
9.60157931e-01 8.57520401e-01 1.50241017e-01 -2.05407888e-02
-1.55119479e-01 -1.31125944e-02 2.99834590e-02 5.55996597e-01
1.24339059e-01 9.13289070e-01 3.59949857e-01 -1.64461285e-01
-1.33069301e+00 -9.83374298e-01 5.40314555e-01 1.30215287e+00
-2.29269832e-01 -3.22148561e-01 -8.94031346e-01 -1.79962993e-01
-2.76365519e-01 8.55248570e-01 -5.48850596e-01 -8.50455537e-02
-1.12016010e+00 -1.71891823e-01 1.00787616e+00 4.96979117e-01
2.04737559e-01 -1.49007118e+00 -1.05475891e+00 1.90598607e-01
-3.53213191e-01 -1.00772762e+00 -7.90398300e-01 -3.24370563e-01
-5.13773382e-01 -5.16863048e-01 -7.40916073e-01 -5.84187925e-01
2.33222261e-01 -2.32442632e-01 1.16216230e+00 1.58227444e-01
-5.05228452e-02 3.30008954e-01 -3.84751618e-01 -1.48120806e-01
-1.14785767e+00 -2.81732791e-04 1.91625684e-01 -3.14623177e-01
-4.28360969e-01 -9.23658371e-01 -7.33538032e-01 3.55629981e-01
-7.50889063e-01 6.31861329e-01 4.26672250e-01 7.00137734e-01
9.95870903e-02 -6.50030077e-01 3.24607819e-01 -4.00997579e-01
9.11860585e-01 -9.38821733e-02 1.19474977e-01 -3.34658802e-01
-7.40471333e-02 1.33961797e-01 6.56554282e-01 -9.99029517e-01
-1.06260610e+00 3.05489928e-01 -3.21754038e-01 -2.84916639e-01
-3.42438161e-01 -4.21622619e-02 -2.65417278e-01 1.52981997e-01
8.50204527e-01 -5.19104227e-02 2.60827065e-01 -6.04270771e-02
8.51023912e-01 5.27795136e-01 1.00924659e+00 -8.58139098e-01
8.55229080e-01 4.72440392e-01 1.55884866e-02 -8.62721920e-01
-7.81307369e-03 3.61975938e-01 -5.70097864e-01 -6.12622499e-01
9.03859437e-01 -5.43388784e-01 -1.07674789e+00 5.11315942e-01
-1.24177945e+00 -9.08696592e-01 -4.43661869e-01 4.31530893e-01
-1.04007912e+00 1.90129504e-01 -6.69979751e-01 -7.33710170e-01
-2.71963447e-01 -1.36512411e+00 1.32576180e+00 2.10005283e-01
-1.31630230e+00 -7.28637755e-01 -3.32867913e-02 1.70769230e-01
4.46207047e-01 8.60049427e-01 6.55183136e-01 -2.49986783e-01
-1.96341932e-01 -1.12053923e-01 5.09155393e-01 -1.06307358e-01
3.51812065e-01 4.53892142e-01 -7.88816035e-01 -6.33786470e-02
-4.87344027e-01 -4.24535722e-01 1.55107588e-01 4.23792005e-01
6.14070237e-01 -5.81627011e-01 -1.45210803e-01 6.24309003e-01
4.32236582e-01 3.49107027e-01 6.83499873e-01 8.10689405e-02
8.65751565e-01 8.66867185e-01 5.71346045e-01 5.74234128e-01
2.21165210e-01 1.16650295e+00 2.65376210e-01 -1.89041048e-01
-5.88579118e-01 -5.63384771e-01 4.72509116e-01 3.93702447e-01
-5.99996328e-01 -2.33934373e-01 -7.83513486e-01 1.97037369e-01
-1.51644492e+00 -1.37911808e+00 -3.97371650e-02 1.89640915e+00
1.16920292e+00 9.30067822e-02 4.39782828e-01 -2.60818340e-02
3.50907505e-01 2.79064059e-01 -4.62898761e-01 -8.33056927e-01
1.44555151e-01 3.55552644e-01 2.43359700e-01 6.46686256e-01
-8.80515516e-01 1.24010062e+00 6.65443850e+00 6.52614772e-01
-1.45698905e+00 -3.39982033e-01 6.07837379e-01 -2.59427786e-01
-6.08803034e-01 -2.17274681e-01 -4.33305413e-01 2.67478079e-01
7.07991362e-01 5.68781979e-02 6.17253542e-01 6.41686141e-01
9.81122971e-01 1.77236855e-01 -1.33775687e+00 6.52956009e-01
-1.64193317e-01 -1.47343159e+00 3.69587317e-02 7.00972378e-02
5.76005459e-01 -7.20980167e-01 2.58132577e-01 1.05386190e-01
7.33496189e-01 -1.60408568e+00 1.03217971e+00 5.72783411e-01
1.39252162e+00 -6.69813573e-01 -3.26245055e-02 3.72094065e-01
-1.01859069e+00 3.23662460e-01 4.70053136e-01 -1.65934056e-01
6.12179637e-01 -4.17068064e-01 -1.00123680e+00 -8.86144415e-02
3.19746345e-01 4.07839902e-02 1.02717511e-01 2.86184400e-01
-5.14248669e-01 5.49052119e-01 -3.78035992e-01 -3.23086143e-01
1.12661816e-01 9.74799320e-02 8.78834486e-01 1.44596207e+00
2.13341326e-01 4.47326303e-01 2.14111451e-02 9.10763383e-01
1.36188656e-01 -1.09490171e-01 -7.33996928e-01 -1.18407376e-01
6.38071120e-01 1.04433358e+00 -5.20323813e-01 -1.94917515e-01
2.47912928e-01 1.17109358e+00 -1.91280156e-01 3.42357695e-01
-8.90127957e-01 -1.53618082e-01 9.33565140e-01 3.24517041e-01
7.35035539e-03 -6.54049337e-01 -5.66414595e-01 -7.11369693e-01
-3.10478151e-01 -1.47980177e+00 -3.30276161e-01 -9.69274640e-01
-6.25567555e-01 5.37993610e-01 1.82594880e-01 -1.28400588e+00
-1.10857129e+00 -2.09179834e-01 -9.02076602e-01 8.21617305e-01
-3.11938167e-01 -1.42035604e+00 -4.40257221e-01 2.22833559e-01
8.30846906e-01 7.56117031e-02 9.28186297e-01 -3.72535586e-02
6.04129396e-02 5.83049297e-01 -6.62132323e-01 2.46122807e-01
6.93783343e-01 -1.15552032e+00 9.25360203e-01 6.62154317e-01
1.62732005e-01 5.99216938e-01 1.53178763e+00 -6.60383224e-01
-1.45650876e+00 -4.86150205e-01 3.68444175e-01 -5.38714707e-01
5.08584559e-01 -2.93296576e-01 -4.97564048e-01 5.70453823e-01
1.47991031e-01 -4.95743573e-01 3.64739865e-01 -3.15414667e-01
-5.17225526e-02 4.49165493e-01 -9.71695602e-01 1.30395973e+00
1.30189621e+00 -1.84564725e-01 -4.92233187e-01 -3.15741710e-02
6.67980015e-01 -8.99947524e-01 -8.28802586e-01 2.41717875e-01
1.27001357e+00 -1.03488469e+00 1.15871382e+00 -7.33724594e-01
8.06120217e-01 -6.17681257e-02 9.25801545e-02 -1.47694933e+00
-2.79010624e-01 -1.24011338e+00 -5.22918627e-02 1.06973684e+00
2.97748327e-01 5.43830171e-02 9.56154108e-01 8.87631238e-01
-1.36127487e-01 -5.59743285e-01 -6.49155140e-01 -5.06162286e-01
2.22821772e-01 -5.96517801e-01 5.66477478e-01 8.28712821e-01
2.41005838e-01 2.74022222e-01 -7.90476859e-01 -2.66769272e-03
4.18053329e-01 -9.28063095e-02 1.44145489e+00 -7.68602133e-01
-7.90827155e-01 -7.36141384e-01 -1.78245753e-01 -1.20967829e+00
2.36716792e-02 -5.11167526e-01 2.50931948e-01 -1.45328701e+00
-2.85313010e-01 -4.09600139e-01 8.66227567e-01 4.82764274e-01
-3.10384631e-02 2.02083379e-01 5.75606346e-01 2.71869332e-01
2.44239897e-01 4.61854368e-01 1.85458291e+00 1.80085957e-01
-6.78895772e-01 8.42117220e-02 -4.97342736e-01 9.70326066e-01
7.77991593e-01 -5.01455814e-02 -4.74411875e-01 -2.67980278e-01
-5.49666286e-02 4.76375818e-01 3.81175697e-01 -9.76775646e-01
1.46762254e-02 -6.43137932e-01 4.31359261e-01 -5.42993248e-02
7.80635238e-01 -4.42357302e-01 6.18167877e-01 5.16872644e-01
-5.71896911e-01 9.98505205e-02 8.10008869e-02 1.40226573e-01
4.09847438e-01 3.06181103e-01 7.57641733e-01 -1.79204881e-01
-4.21235114e-01 2.58351415e-02 -7.18066335e-01 -1.16830386e-01
8.55351806e-01 -4.37304646e-01 -3.87300737e-02 -1.18019044e+00
-1.03722656e+00 -7.26759136e-02 6.64845288e-01 6.30876660e-01
5.34329772e-01 -1.39478993e+00 -7.40769923e-01 -1.95166599e-02
-3.18234563e-01 -1.46066591e-01 -1.45131573e-01 4.86074716e-01
-7.96888769e-01 -9.95018035e-02 -3.93111855e-01 -5.66906512e-01
-1.46173298e+00 -1.49030089e-01 4.49704438e-01 2.13453025e-02
-8.35212767e-01 5.94704747e-01 -1.91617990e-03 -2.44620696e-01
1.06797963e-01 -2.20150858e-01 -1.50291577e-01 -2.15675786e-01
3.61718625e-01 1.91650391e-01 -5.02455175e-01 -7.83884823e-01
-8.80516842e-02 4.48113054e-01 5.22033036e-01 -9.81295049e-01
1.11189651e+00 1.38217628e-01 4.28232640e-01 3.29917759e-01
6.79805279e-01 4.56964582e-01 -1.86266923e+00 6.00734234e-01
-4.53618705e-01 -4.13260370e-01 -5.18158615e-01 -7.97612786e-01
-7.56878436e-01 6.54420793e-01 1.60367548e-01 -4.02108952e-02
6.54445767e-01 5.62177189e-02 8.61933529e-01 2.30108351e-01
4.17268544e-01 -9.09442544e-01 3.92649621e-01 3.13835621e-01
1.28500998e+00 -7.06068873e-01 -2.35715583e-01 -2.37303630e-01
-1.02455473e+00 1.18158174e+00 6.98942423e-01 -9.15321335e-02
1.35004207e-01 8.77732694e-01 3.92336935e-01 -7.85843432e-02
-6.67704582e-01 1.57694802e-01 3.32261890e-01 8.94230366e-01
7.23788559e-01 3.12934726e-01 -3.71734768e-01 2.60332316e-01
-1.07479417e+00 -1.87007517e-01 6.16340637e-01 6.95946276e-01
-3.86580825e-01 -1.05780697e+00 -5.30159950e-01 -2.40355711e-02
-3.28319430e-01 1.47509903e-01 -5.55490911e-01 1.05455482e+00
-1.52649894e-01 9.15377200e-01 -7.34735727e-02 -6.77583337e-01
4.12597567e-01 -4.56866547e-02 6.47891164e-01 -6.43599808e-01
-6.99744284e-01 1.59224153e-01 5.13638198e-01 -6.79655254e-01
-1.25384256e-01 -7.37856686e-01 -1.68199229e+00 -5.94651282e-01
1.45555332e-01 -2.00855017e-01 6.25720739e-01 6.87273324e-01
1.77708909e-01 5.92450857e-01 4.25618023e-01 -1.73832059e+00
-4.43506867e-01 -1.04448247e+00 -3.08845360e-02 7.92580366e-01
8.40611979e-02 -4.49126601e-01 -2.39247337e-01 3.85696828e-01] | [5.613417148590088, -0.08190403878688812] |
35506d7c-a87a-4666-b2bc-fc8d63c62568 | finrl-a-deep-reinforcement-learning-library | 2011.09607 | null | https://arxiv.org/abs/2011.09607v2 | https://arxiv.org/pdf/2011.09607v2.pdf | FinRL: A Deep Reinforcement Learning Library for Automated Stock Trading in Quantitative Finance | As deep reinforcement learning (DRL) has been recognized as an effective approach in quantitative finance, getting hands-on experiences is attractive to beginners. However, to train a practical DRL trading agent that decides where to trade, at what price, and what quantity involves error-prone and arduous development and debugging. In this paper, we introduce a DRL library FinRL that facilitates beginners to expose themselves to quantitative finance and to develop their own stock trading strategies. Along with easily-reproducible tutorials, FinRL library allows users to streamline their own developments and to compare with existing schemes easily. Within FinRL, virtual environments are configured with stock market datasets, trading agents are trained with neural networks, and extensive backtesting is analyzed via trading performance. Moreover, it incorporates important trading constraints such as transaction cost, market liquidity and the investor's degree of risk-aversion. FinRL is featured with completeness, hands-on tutorial and reproducibility that favors beginners: (i) at multiple levels of time granularity, FinRL simulates trading environments across various stock markets, including NASDAQ-100, DJIA, S&P 500, HSI, SSE 50, and CSI 300; (ii) organized in a layered architecture with modular structure, FinRL provides fine-tuned state-of-the-art DRL algorithms (DQN, DDPG, PPO, SAC, A2C, TD3, etc.), commonly-used reward functions and standard evaluation baselines to alleviate the debugging workloads and promote the reproducibility, and (iii) being highly extendable, FinRL reserves a complete set of user-import interfaces. Furthermore, we incorporated three application demonstrations, namely single stock trading, multiple stock trading, and portfolio allocation. The FinRL library will be available on Github at link https://github.com/AI4Finance-LLC/FinRL-Library. | ['Christina Dan Wang', 'Bowen Xiao', 'Liuqing Yang', 'Runjia Zhang', 'Qian Chen', 'Hongyang Yang', 'Xiao-Yang Liu'] | 2020-11-19 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-1.27289474e+00 -3.70286882e-01 -3.34993631e-01 -1.49827018e-01
-8.00978959e-01 -9.16840315e-01 3.30113590e-01 -1.44176707e-01
-4.31210279e-01 9.11787808e-01 -4.25853759e-01 -8.81486654e-01
-1.14790499e-01 -9.93890584e-01 -5.45934379e-01 -3.30540925e-01
-3.65049571e-01 7.51843870e-01 1.63263068e-01 -5.06908953e-01
4.69792694e-01 5.67925274e-01 -1.19135714e+00 -1.42275944e-01
7.70551860e-01 1.37644649e+00 -2.08224952e-01 5.05309880e-01
-1.47831306e-01 1.28395212e+00 -7.07239091e-01 -7.28177547e-01
7.54864454e-01 -1.61751598e-01 -3.61068219e-01 -3.13400030e-01
-1.70052156e-01 -8.17893982e-01 -1.76722072e-02 9.63901103e-01
4.48054492e-01 -2.60346204e-01 3.01353216e-01 -1.61398709e+00
-6.32129490e-01 8.75531793e-01 -7.55482495e-01 3.52396756e-01
-1.31221756e-01 6.65805340e-01 1.29542887e+00 -6.49195135e-01
1.79836452e-01 1.08903873e+00 4.23203856e-01 3.70148599e-01
-9.23363924e-01 -1.12106514e+00 1.34031639e-01 -1.56154141e-01
-6.46598279e-01 -3.22968252e-02 5.77056110e-01 -5.23566842e-01
1.11272979e+00 1.06656618e-01 7.64556885e-01 8.54722261e-01
4.74468619e-01 9.39522982e-01 1.18551397e+00 2.11395919e-02
6.41535401e-01 2.59763420e-01 4.35006283e-02 5.02778947e-01
2.79696494e-01 5.70835650e-01 -2.32555792e-01 -3.25879842e-01
1.29425466e+00 2.52051633e-02 3.84023413e-02 -3.63860905e-01
-1.00523615e+00 1.16445029e+00 1.57617435e-01 1.08706914e-01
-3.53601575e-01 3.15698385e-01 4.54347283e-01 9.54049647e-01
3.10154617e-01 6.84897423e-01 -9.03694272e-01 -5.42929828e-01
-7.42351890e-01 7.30980515e-01 1.18103909e+00 8.48643482e-01
4.37397331e-01 3.69792342e-01 -6.76534325e-02 5.12380481e-01
4.23049033e-01 5.04059017e-01 6.35866880e-01 -1.46592271e+00
5.41817605e-01 3.62568170e-01 5.75680375e-01 -4.21538353e-01
-3.86232525e-01 -5.57847798e-01 -5.66599488e-01 9.94024992e-01
5.59790909e-01 -5.56395411e-01 -3.05519700e-01 1.50169182e+00
1.50171900e-02 -7.36717880e-02 1.66138545e-01 7.86645651e-01
1.24123357e-01 4.77266490e-01 -2.92308599e-01 -2.68129826e-01
1.23658204e+00 -1.03718412e+00 -1.89930469e-01 1.43055558e-01
3.91853899e-01 -5.55583954e-01 1.27367938e+00 7.09883153e-01
-1.53004813e+00 -2.49512151e-01 -9.45204318e-01 4.33343858e-01
-4.10883754e-01 -3.77409816e-01 8.40648353e-01 6.90103292e-01
-1.14337850e+00 9.97181118e-01 -9.53563035e-01 5.22655308e-01
2.51716942e-01 3.83981675e-01 3.36274058e-01 5.75267017e-01
-1.51499581e+00 7.06751704e-01 1.32636264e-01 -9.96369794e-02
-9.34972465e-01 -1.01261020e+00 -5.80558598e-01 3.61102641e-01
4.46692556e-01 -4.53669071e-01 1.79468083e+00 -8.27108264e-01
-1.88933694e+00 4.32391912e-01 8.82168770e-01 -8.28012049e-01
1.16754413e+00 -2.00360075e-01 -2.64068097e-01 -1.35914572e-02
1.22696832e-01 3.97802114e-01 4.57510531e-01 -5.92631042e-01
-6.55421078e-01 -2.75854543e-02 2.50263751e-01 2.38152668e-01
-3.75006460e-02 1.55493334e-01 -1.19100071e-01 -8.59456718e-01
-5.78910053e-01 -5.06384134e-01 -1.55292943e-01 -2.26528585e-01
7.46086091e-02 -1.53292030e-01 4.83124077e-01 -4.40816402e-01
1.13669372e+00 -2.13155484e+00 -3.88837993e-01 2.56781965e-01
8.62710029e-02 -3.30672450e-02 2.05899011e-02 4.65550601e-01
-4.73179482e-02 3.59580576e-01 -6.20958358e-02 4.09333184e-02
7.82762349e-01 -1.08957589e-01 -4.50132996e-01 1.34302765e-01
9.14701372e-02 9.41296399e-01 -7.12319493e-01 -1.86991498e-01
1.83897480e-01 -2.53766119e-01 -7.21946359e-01 4.34676170e-01
-6.29528940e-01 6.60407022e-02 -5.95870793e-01 9.36053097e-01
6.85357153e-01 -4.50468004e-01 2.14184877e-02 4.23515618e-01
-4.49811488e-01 3.12474877e-01 -1.44852638e+00 9.25332546e-01
-4.99526620e-01 8.97162631e-02 6.03999756e-02 -6.60411119e-01
9.10497248e-01 2.26699024e-01 4.96131986e-01 -1.03044999e+00
3.22731249e-02 5.66387773e-01 -6.75494820e-02 9.21758935e-02
2.61364192e-01 -1.30971134e-01 -2.83640474e-01 1.11238253e+00
-1.08898640e-01 -9.92229283e-02 5.51278830e-01 2.69179065e-02
1.11198342e+00 1.35714009e-01 1.00167461e-01 -3.19711894e-01
1.12256847e-01 -1.12407252e-01 6.48495078e-01 6.80269599e-01
-3.92964184e-01 3.69030833e-02 1.15680254e+00 -4.89463240e-01
-1.03392160e+00 -1.35985124e+00 -8.66069347e-02 1.06884181e+00
-2.33933687e-01 1.28658459e-01 -5.65737426e-01 -6.16382062e-01
6.29164279e-01 8.32979620e-01 -4.48048532e-01 3.15505087e-01
-4.27273452e-01 -6.63430810e-01 4.18370962e-01 5.35177290e-01
1.00570953e+00 -1.55005753e+00 -8.79127979e-01 3.20340276e-01
5.19105732e-01 -3.91757905e-01 -7.01298118e-01 3.41994166e-01
-8.01402748e-01 -1.12878215e+00 -8.34802687e-01 -4.30095702e-01
1.03288054e-01 -1.48738444e-01 1.48067641e+00 5.25752380e-02
-6.76802099e-02 2.14620799e-01 -8.70124325e-02 -4.06277537e-01
-2.46709034e-01 -9.01046395e-02 -8.04055929e-02 -4.52181429e-01
1.29386753e-01 -4.50987220e-01 -7.04249442e-01 5.23227274e-01
-7.10855007e-01 -3.10236216e-01 4.23173726e-01 9.97592151e-01
3.43480587e-01 1.11008093e-01 9.26152766e-01 -1.04304290e+00
8.54337990e-01 -4.71794963e-01 -1.68613064e+00 3.34683567e-01
-1.16074479e+00 2.54239459e-02 6.85705543e-01 -2.49041870e-01
-9.98154700e-01 -6.59333408e-01 -1.67450905e-01 -4.10940140e-01
3.64057332e-01 5.44722855e-01 -5.26170842e-02 2.29349762e-01
3.45914394e-01 2.17764918e-02 2.23634303e-01 -4.36152428e-01
1.23407811e-01 4.20003116e-01 1.00315899e-01 -8.36085975e-01
7.19197333e-01 -2.92448729e-01 -5.68905890e-01 5.73748201e-02
-4.34032202e-01 1.83266088e-01 -4.63597737e-02 6.02315255e-02
5.10162115e-01 -9.45555508e-01 -1.19680905e+00 9.08937395e-01
-4.80304718e-01 -1.09439576e+00 -3.97163242e-01 4.69272703e-01
-6.49211943e-01 -2.61343662e-02 -1.34834135e+00 -7.54684091e-01
-4.40203398e-01 -1.28424597e+00 2.47028068e-01 4.61088628e-01
-1.11445822e-02 -1.19194698e+00 1.84658587e-01 2.05747947e-01
6.79484785e-01 2.13924125e-01 8.89823496e-01 -8.69898438e-01
-7.88341224e-01 9.63302422e-03 -1.73167050e-01 5.92999399e-01
-3.54472697e-02 2.69089907e-01 -5.62264144e-01 -4.68629062e-01
-1.31374765e-02 -7.40732551e-01 3.55870575e-01 4.52948779e-01
1.01730025e+00 -5.38148105e-01 3.60613346e-01 5.89794517e-01
1.17177629e+00 7.28600860e-01 4.15462434e-01 1.03381050e+00
-2.68617235e-02 3.23717773e-01 8.83157551e-01 8.64570498e-01
3.15481186e-01 2.26386964e-01 4.46497113e-01 -7.28949308e-02
7.81842411e-01 -4.06635329e-02 6.81847274e-01 5.67306995e-01
7.26797357e-02 8.83861333e-02 -6.61651015e-01 -8.92406031e-02
-1.81551898e+00 -1.01370573e+00 3.41408521e-01 2.15755415e+00
1.16707301e+00 7.55809009e-01 6.05471253e-01 -3.10087830e-01
2.76974320e-01 3.04938015e-02 -1.11831951e+00 -5.33569038e-01
-7.87179670e-05 3.82341057e-01 7.02463508e-01 4.60290283e-01
-8.10825765e-01 8.62207472e-01 5.79440022e+00 7.99196064e-01
-1.19003892e+00 -1.50697440e-01 1.22072363e+00 -2.06793100e-01
-4.88136619e-01 -1.26254991e-01 -8.15495133e-01 6.36667848e-01
1.02731562e+00 -6.31266713e-01 7.40608633e-01 1.23359704e+00
1.53604895e-01 1.61990091e-01 -9.97201502e-01 6.44663811e-01
-8.56377780e-01 -1.65552032e+00 -4.17365700e-01 1.89053968e-01
5.07825077e-01 9.95706320e-02 3.87147218e-01 9.81536806e-01
1.00415528e+00 -9.44093585e-01 9.88275766e-01 2.85263419e-01
4.84517276e-01 -1.22772121e+00 9.44244742e-01 1.26093492e-01
-9.39204633e-01 -1.59991458e-01 -3.19055021e-01 -1.17326781e-01
-1.83517307e-01 3.30535322e-01 -3.65792722e-01 3.37541491e-01
7.99959064e-01 5.39966583e-01 -4.72474843e-01 7.36359894e-01
-1.91953331e-01 5.11379063e-01 -1.69277489e-01 -2.02968776e-01
4.09622014e-01 -4.96156782e-01 -1.71592623e-01 8.58862162e-01
3.18067402e-01 1.21764950e-01 1.53073907e-01 1.21846342e+00
-9.89292935e-02 -2.80858316e-02 -1.12617977e-01 -2.14111149e-01
5.51394641e-01 1.19906294e+00 -6.07400179e-01 -2.58278072e-01
-4.55222726e-01 4.88763958e-01 5.28775901e-02 3.17470223e-01
-1.08711219e+00 -4.54106659e-01 8.85548711e-01 9.41812918e-02
2.94361949e-01 -1.31661594e-02 -2.68613100e-01 -1.09959388e+00
-3.16057242e-02 -1.33629012e+00 5.48389971e-01 -5.57380974e-01
-1.41616535e+00 4.13463116e-01 -1.03791542e-01 -1.14520311e+00
-5.79853475e-01 -7.36324251e-01 -7.64865875e-01 9.30317342e-01
-1.67337441e+00 -3.13702077e-01 2.32536033e-01 5.73443294e-01
4.07561958e-01 -7.46182263e-01 4.64492977e-01 3.60085040e-01
-7.60087192e-01 7.10702479e-01 2.70625472e-01 5.05767345e-01
4.97577459e-01 -1.71645820e+00 6.95388496e-01 3.10609341e-01
-1.96926132e-01 5.46461701e-01 4.45665509e-01 -5.95052958e-01
-1.24284756e+00 -8.53067935e-01 1.92458078e-01 -1.83813587e-01
1.41566885e+00 -9.31903273e-02 -8.18377972e-01 6.95775449e-01
2.94315219e-01 -1.69747844e-01 5.09444475e-01 -2.00244024e-01
-2.54406393e-01 -4.07305211e-01 -1.22821832e+00 6.63116693e-01
2.63557464e-01 -1.77392855e-01 -3.59868884e-01 1.98367923e-01
7.83619106e-01 -6.05994463e-01 -1.04887044e+00 -5.66896833e-02
4.49380219e-01 -1.46422195e+00 7.41189122e-01 -2.61338502e-01
3.77855986e-01 3.59372608e-03 1.21278301e-01 -1.15538204e+00
-7.27049112e-02 -9.35703993e-01 8.32541101e-03 1.19039524e+00
6.70300901e-01 -1.18065012e+00 7.95839846e-01 8.80300939e-01
8.14066231e-02 -9.35569525e-01 -6.91257238e-01 -1.01392436e+00
6.40559256e-01 -1.70273170e-01 9.52909470e-01 1.02788305e+00
3.74988057e-02 -1.93057314e-01 -1.84488192e-01 -1.47465438e-01
8.51424336e-01 4.39809293e-01 6.27835572e-01 -9.38727200e-01
-1.00979316e+00 -1.02236950e+00 3.98914218e-01 -8.12763095e-01
-7.04449713e-02 -5.56095600e-01 -4.98976856e-01 -8.46602678e-01
-1.68537974e-01 -7.03955412e-01 -7.24187315e-01 5.73047936e-01
2.78334022e-01 -2.69291133e-01 3.81690383e-01 2.64602780e-01
-4.91046607e-01 3.23108435e-01 1.18276024e+00 -1.37858465e-02
-3.55259836e-01 4.01986569e-01 -7.78186202e-01 5.08558810e-01
1.16334558e+00 -2.17261508e-01 -2.84865260e-01 1.63452793e-02
4.27323133e-01 7.45009482e-01 2.47563541e-01 -6.01091504e-01
-2.30157878e-02 -5.77866197e-01 3.69517833e-01 -5.04470229e-01
-2.43639946e-01 -3.86117101e-01 2.02555016e-01 8.43755305e-01
-2.88888425e-01 6.76289618e-01 1.81087986e-01 -1.18671574e-01
-3.70531827e-01 -2.61989772e-01 8.45466554e-01 -4.98062789e-01
-4.20631886e-01 4.75315988e-01 -2.19248474e-01 5.54620028e-01
1.17538655e+00 2.32183561e-01 -4.62392211e-01 -5.24028778e-01
-4.64060843e-01 9.22461033e-01 5.95660627e-01 1.80168942e-01
3.98199052e-01 -1.20046890e+00 -5.80644429e-01 1.50112659e-01
-3.56806427e-01 -1.31298080e-01 7.15979701e-03 5.65917730e-01
-1.03589296e+00 1.79605857e-01 -4.63216364e-01 5.35296090e-02
-5.97147524e-01 5.37496030e-01 7.37020671e-01 -8.51470172e-01
-4.38496858e-01 7.56948411e-01 1.22833997e-01 -4.46742415e-01
3.62971127e-01 -5.34503639e-01 1.96452409e-01 1.55533120e-01
6.16623402e-01 3.88763785e-01 -1.17195167e-01 4.99440730e-01
-1.67669073e-01 1.59241855e-01 -9.25282165e-02 -4.12095040e-01
1.65555966e+00 3.01105946e-01 4.85189706e-02 5.76779783e-01
7.15946019e-01 -2.72582881e-02 -1.88882947e+00 1.36654630e-01
3.21657449e-01 -2.07686156e-01 -2.19419956e-01 -1.03604555e+00
-1.44007373e+00 7.18668163e-01 4.58579183e-01 5.21680474e-01
9.72667158e-01 -4.71835226e-01 6.96677685e-01 3.28130186e-01
6.74354732e-01 -1.33009911e+00 2.15152860e-01 4.79407191e-01
8.69078815e-01 -1.18076825e+00 -7.50121251e-02 5.23529172e-01
-9.92581427e-01 1.28800297e+00 6.47111475e-01 -4.43771929e-01
8.36380064e-01 7.32723951e-01 3.93066376e-01 2.76123341e-02
-1.11617160e+00 2.64982700e-01 -4.02189642e-01 1.17911175e-01
3.49154204e-01 1.59471020e-01 -2.32221223e-02 8.32394838e-01
-4.72970575e-01 1.18727330e-02 7.40316868e-01 9.66198266e-01
-4.11154211e-01 -1.31958985e+00 -2.94377744e-01 3.63234401e-01
-7.16997683e-01 -1.38473511e-01 -9.31579098e-02 1.30947363e+00
-4.70336527e-01 4.59369361e-01 2.17481047e-01 3.76643091e-02
3.26038808e-01 -1.77820593e-01 1.38685554e-02 -3.58535737e-01
-1.18209720e+00 3.63661170e-01 -1.14116348e-01 -5.59202731e-01
2.90795892e-01 -7.11114824e-01 -1.43797791e+00 -7.96527565e-01
8.57658833e-02 5.36190152e-01 2.93433309e-01 5.08926868e-01
1.24355480e-01 3.20876122e-01 1.07656658e+00 -6.31524444e-01
-1.44209087e+00 -7.29698479e-01 -1.24053073e+00 -4.26073782e-02
2.03550547e-01 -6.75185800e-01 -6.15551889e-01 -5.49950182e-01] | [4.428737640380859, 3.945425510406494] |
d3ab078b-5930-4eec-844e-41df931a713a | masakhapos-part-of-speech-tagging-for | 2305.13989 | null | https://arxiv.org/abs/2305.13989v1 | https://arxiv.org/pdf/2305.13989v1.pdf | MasakhaPOS: Part-of-Speech Tagging for Typologically Diverse African Languages | In this paper, we present MasakhaPOS, the largest part-of-speech (POS) dataset for 20 typologically diverse African languages. We discuss the challenges in annotating POS for these languages using the UD (universal dependencies) guidelines. We conducted extensive POS baseline experiments using conditional random field and several multilingual pre-trained language models. We applied various cross-lingual transfer models trained with data available in UD. Evaluating on the MasakhaPOS dataset, we show that choosing the best transfer language(s) in both single-source and multi-source setups greatly improves the POS tagging performance of the target languages, in particular when combined with cross-lingual parameter-efficient fine-tuning methods. Crucially, transferring knowledge from a language that matches the language family and morphosyntactic properties seems more effective for POS tagging in unseen languages. | ['Dietrich Klakow', 'Muhammad Abdullahi', 'Aliyu Yusuf', 'Chinedu Uchechukwu', 'Seydou Traore', 'Apelete Agbolo', 'Patrick Mizha', 'Kudzai Gotosa', 'Ester Chimhenga', 'Emile Niyomutabazi', 'Théogène Musabeyezu', 'Marien Nahimana', 'Olanrewaju Samuel', 'Idris Akinade', 'Tolulope Adelani', 'Gratien Atindogbe', 'Ikechukwu Onyenwe', 'Mboning Tchiaze Elvis', 'Tajuddeen Gwadabe', 'Vukosi Marivate', 'Tebogo Macucwa', 'Godson Kalipe', 'Amelia Taylor', 'Fatoumata Ouoba Kabore', 'Victoire Memdjokam Koagne', 'Edwin Munkoh-Buabeng', 'Allahsera Auguste Tapo', 'Rooweither Mabuya', 'Andiswa Bukula', 'Bonaventure F. P. Dossou', 'Blessing Sibanda', 'Jonathan Mukiibi', 'Derguene Mbaye', 'Catherine Gitau', 'Anuoluwapo Aremu', 'Perez Ogayo', 'Chris Chinenye Emezue', 'Shamsuddeen Hassan Muhammad', 'Happy Buzaaba', 'Thapelo Sindane', 'Jesujoba Alabi', 'Peter Nabende', 'David Adelani', 'Cheikh M. Bamba Dione'] | 2023-05-23 | null | null | null | null | ['part-of-speech-tagging', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing'] | [-3.25847328e-01 2.75902753e-03 -4.98047382e-01 -3.29900712e-01
-1.12140512e+00 -1.18616247e+00 3.66429478e-01 2.43892953e-01
-6.59580231e-01 1.18428707e+00 3.46223384e-01 -5.69662571e-01
2.91688681e-01 -3.97125214e-01 -7.53482521e-01 -5.11171162e-01
3.65095586e-02 1.02942538e+00 3.25044781e-01 -3.46409440e-01
-1.08083747e-02 3.79996866e-01 -7.66148269e-01 2.96848714e-01
9.79570806e-01 6.03691153e-02 5.78204393e-01 4.59683090e-01
-2.33429536e-01 4.26606745e-01 -4.26470399e-01 -9.17046845e-01
1.21720374e-01 -1.64976299e-01 -1.19377768e+00 -5.23210406e-01
4.64227289e-01 4.34871882e-01 1.70258522e-01 9.42927718e-01
5.39295971e-01 -1.50355518e-01 6.16778791e-01 -6.78913355e-01
-1.05110204e+00 1.26442099e+00 -2.44016394e-01 6.73492014e-01
2.19146207e-01 5.31072021e-02 1.31038487e+00 -8.47027957e-01
1.00333309e+00 1.41640425e+00 8.78272116e-01 4.23019499e-01
-1.32826674e+00 -4.69961643e-01 -1.38323128e-01 1.12658069e-02
-1.31116533e+00 -2.98147440e-01 4.62273717e-01 -5.22716105e-01
1.38259768e+00 -3.00022125e-01 9.71510708e-02 1.20646715e+00
1.86823025e-01 6.52539134e-01 1.35728621e+00 -1.13994694e+00
-1.48731068e-01 1.74902216e-01 -9.06674340e-02 4.81399894e-01
2.71494359e-01 -1.09171569e-01 -4.77894962e-01 -2.05737069e-01
5.21361291e-01 -7.52461076e-01 1.33767277e-01 7.05276951e-02
-1.31378150e+00 6.77394927e-01 6.29382879e-02 9.53723431e-01
-2.18242213e-01 -1.35799393e-01 6.07587039e-01 2.63269573e-01
5.88222027e-01 4.91228521e-01 -1.39276171e+00 -2.32268915e-01
-3.68664443e-01 -2.46829793e-01 8.73540163e-01 9.62342322e-01
8.09067011e-01 -5.65617830e-02 1.71371214e-02 1.41774559e+00
1.60071760e-01 1.00390124e+00 3.80588323e-01 -5.93957007e-01
7.62597144e-01 -2.81197410e-02 1.48417473e-01 -3.53831440e-01
-1.52812362e-01 1.10738955e-01 1.68828778e-02 -6.34624422e-01
6.55058146e-01 -6.35975599e-01 -1.11482346e+00 1.90061975e+00
2.57924259e-01 -2.67342031e-01 5.37525177e-01 2.60089517e-01
3.39031607e-01 7.81698048e-01 1.00235391e+00 -1.22728422e-01
1.75861526e+00 -8.28717232e-01 -4.57788676e-01 -6.35610402e-01
9.52527404e-01 -1.25988138e+00 1.02973568e+00 -6.55279532e-02
-9.25055623e-01 -5.74490905e-01 -5.48118472e-01 -1.23699471e-01
-6.46959662e-01 3.16741556e-01 6.72327876e-01 6.14733219e-01
-1.18688667e+00 3.91287893e-01 -9.25351739e-01 -7.37064540e-01
-7.78495222e-02 3.16907376e-01 -6.56157136e-01 -1.77666783e-01
-1.57623446e+00 1.14434981e+00 9.43649471e-01 -2.11575568e-01
-4.97506738e-01 -5.71608782e-01 -1.08934021e+00 -1.81606740e-01
-7.69909769e-02 -1.13631964e-01 1.36333370e+00 -1.06529164e+00
-1.29499185e+00 1.36722851e+00 -1.08873531e-01 -2.95095354e-01
-1.04173750e-01 -3.57199818e-01 -7.66119659e-01 7.42996410e-02
5.30286968e-01 9.35018063e-01 4.95252758e-01 -9.03871536e-01
-9.36845362e-01 -6.22556508e-02 -2.32749686e-01 1.53681710e-01
-7.94454813e-02 8.92638028e-01 -2.81421155e-01 -6.97950423e-01
-4.04117644e-01 -1.20902288e+00 -2.18272403e-01 -8.94148827e-01
3.32490914e-02 -6.93796873e-01 1.90241024e-01 -1.18465996e+00
8.99464309e-01 -2.09703565e+00 -1.59604207e-01 -8.93782675e-02
-8.46991420e-01 5.99742115e-01 -2.63809443e-01 7.73506880e-01
-1.23379812e-01 2.23453507e-01 8.09097216e-02 -9.34173316e-02
-2.68669389e-02 9.02042687e-01 -2.81466842e-01 1.50646150e-01
7.36420512e-01 9.20098960e-01 -1.15044284e+00 -6.93777204e-01
-2.79573537e-03 5.31907380e-01 -3.25455099e-01 -7.26556256e-02
-1.39784291e-01 7.15935469e-01 -2.30699509e-01 7.41786659e-01
5.31464219e-01 2.59611219e-01 7.30538726e-01 3.53918552e-01
-2.45208785e-01 1.04497111e+00 -4.66202825e-01 1.56720042e+00
-8.00229847e-01 2.73914456e-01 -1.69872522e-01 -7.05610096e-01
8.81446183e-01 5.23854733e-01 3.45021151e-02 -5.22669613e-01
-1.76481292e-01 8.64042461e-01 3.67432296e-01 -3.92635286e-01
4.90200460e-01 -3.26193333e-01 -4.32320237e-01 2.62902472e-02
7.91881144e-01 1.99331030e-01 4.46946740e-01 -4.07668769e-01
7.32047021e-01 5.55368960e-01 7.38032699e-01 -8.07349086e-01
4.85612690e-01 3.85545641e-01 9.23114657e-01 3.87025595e-01
-2.49133065e-01 1.51986748e-01 3.44374418e-01 -2.27975741e-01
-1.09928286e+00 -1.12242484e+00 -7.43193984e-01 1.61939716e+00
-4.91397411e-01 -1.05481975e-01 -5.22467315e-01 -1.00308406e+00
-1.53179124e-01 7.32098401e-01 -4.13749993e-01 6.57942712e-01
-1.19622457e+00 -7.12276757e-01 8.68455470e-01 6.39115512e-01
-1.45138741e-01 -1.59349203e+00 3.23112488e-01 4.96602029e-01
-1.11722223e-01 -1.39952445e+00 -4.30258632e-01 6.75191045e-01
-4.11651343e-01 -7.37256408e-01 -8.75958622e-01 -1.58892822e+00
5.29103458e-01 -1.36311784e-01 1.25526905e+00 -2.90871292e-01
3.63470823e-01 2.70844437e-03 -7.40288019e-01 -4.41709012e-01
-7.47465491e-01 4.59635675e-01 1.27860740e-01 -5.14777899e-01
8.92756045e-01 -3.42482746e-01 2.77599126e-01 1.19249851e-01
-4.80997890e-01 -5.52388191e-01 5.92899561e-01 6.35007262e-01
5.97399950e-01 -3.40333343e-01 5.10673940e-01 -1.12738729e+00
-4.37859558e-02 -6.19396210e-01 -6.03256226e-01 4.34575409e-01
4.17855084e-02 9.93256643e-02 7.76424408e-01 -4.56238210e-01
-1.44562626e+00 5.56933396e-02 -5.73457778e-01 2.43849248e-01
-4.78901058e-01 4.56029415e-01 -2.66588062e-01 6.47405386e-02
7.34075010e-01 -2.26057217e-01 -8.36091995e-01 -6.89880908e-01
4.10825968e-01 7.84738600e-01 4.90436524e-01 -1.04707921e+00
5.46489060e-01 -5.94071224e-02 -3.74412537e-01 -9.29902673e-01
-9.59391117e-01 -6.35213196e-01 -1.30904150e+00 4.96854454e-01
1.09887350e+00 -1.26107466e+00 -2.39130650e-02 3.65300745e-01
-1.36216867e+00 -7.13523865e-01 2.80363765e-02 6.38392270e-01
-1.42113224e-01 3.36186171e-01 -9.42556620e-01 -4.58645403e-01
-1.39069065e-01 -8.91463101e-01 1.04289985e+00 1.38248503e-01
-3.06123495e-01 -1.65095460e+00 7.46576011e-01 1.16459623e-01
-6.07161783e-02 -1.54882506e-01 9.80903387e-01 -1.13589334e+00
6.11213222e-02 3.35785717e-01 9.31806192e-02 4.88856256e-01
8.73809457e-02 1.91635508e-02 -8.40928614e-01 -1.97803557e-01
-7.55195737e-01 -5.11312425e-01 3.87909025e-01 1.87383473e-01
4.99559492e-02 -2.11075231e-01 -2.69435108e-01 2.75667280e-01
1.49846876e+00 3.39503825e-01 4.00837302e-01 3.80577475e-01
7.65578270e-01 7.04520285e-01 7.34811664e-01 -3.03745121e-01
6.51650310e-01 4.46318507e-01 -5.99916697e-01 1.58222377e-01
-2.07932681e-01 -6.36802256e-01 9.94649768e-01 1.44346988e+00
1.50804356e-01 -2.09034309e-01 -1.61191118e+00 1.26561344e+00
-1.59550810e+00 -5.22541046e-01 -2.22329095e-01 1.85138559e+00
1.37943232e+00 2.85026580e-02 -1.63207948e-02 -5.62105834e-01
9.11538959e-01 -1.17454916e-01 3.53807300e-01 -9.92296040e-01
-4.11516190e-01 9.40928817e-01 7.13467002e-01 8.39715660e-01
-1.30068362e+00 1.82483363e+00 6.49016333e+00 9.71289873e-01
-1.04999518e+00 4.87988412e-01 1.42685488e-01 9.23820198e-01
-1.39526129e-01 3.08560938e-01 -1.47259068e+00 4.95994359e-01
1.48975277e+00 2.05273017e-01 1.68566346e-01 1.00113654e+00
-2.66989738e-01 -6.53275624e-02 -9.57373917e-01 4.91787225e-01
-3.92663525e-03 -8.17860126e-01 -2.64054924e-01 -8.20976123e-02
9.52214062e-01 8.89470220e-01 -4.02793407e-01 5.78703701e-01
9.81294036e-01 -6.44937277e-01 6.12745702e-01 -2.84311146e-01
8.24475050e-01 -7.39435077e-01 1.05395389e+00 1.14193082e-01
-1.11448538e+00 5.11206269e-01 -6.94466472e-01 9.61117744e-02
3.39651108e-01 3.31334203e-01 -1.03653920e+00 6.13226175e-01
7.28608191e-01 5.03455460e-01 -6.55593157e-01 6.11960828e-01
-9.97314036e-01 1.18046010e+00 -4.65787023e-01 3.23123597e-02
3.93864483e-01 -6.68717921e-02 4.40658152e-01 1.86951160e+00
2.38628760e-01 -2.49209642e-01 4.12918478e-01 5.86366281e-02
-4.09716740e-02 5.12549520e-01 -5.19050300e-01 -2.46403813e-01
5.68318546e-01 1.24886262e+00 -7.69358575e-01 -3.05199087e-01
-5.59116185e-01 8.75145376e-01 8.76504838e-01 1.61158636e-01
-4.15733546e-01 -2.52227694e-01 5.54062843e-01 -2.04150677e-01
8.16505849e-01 -4.72593874e-01 2.55354866e-02 -1.19158423e+00
-1.68158919e-01 -6.43323243e-01 6.76570296e-01 -5.87541461e-01
-1.74578440e+00 7.41474628e-01 1.99823633e-01 -6.64402246e-01
-3.92021656e-01 -1.02826583e+00 -3.87698889e-01 1.19902897e+00
-1.84473240e+00 -1.68004477e+00 7.04982758e-01 2.59363502e-01
2.31068671e-01 -7.50021543e-03 1.15910947e+00 3.69738400e-01
-3.01907510e-01 4.91534054e-01 7.37279877e-02 5.86146057e-01
1.06949651e+00 -1.45670879e+00 8.76369178e-01 9.93849695e-01
4.55124676e-01 5.85156858e-01 5.19189775e-01 -8.46046746e-01
-7.40694821e-01 -9.65956032e-01 2.01769495e+00 -8.14202428e-01
1.26248157e+00 -4.36271459e-01 -1.00638282e+00 1.25968134e+00
6.96353137e-01 2.00400241e-02 9.47710514e-01 6.44876719e-01
-3.13048154e-01 3.11263680e-01 -9.81554627e-01 3.78910661e-01
9.00711000e-01 -7.39571333e-01 -1.26811349e+00 5.20950675e-01
7.68176436e-01 -2.50481844e-01 -1.22907960e+00 8.83269608e-02
4.27549779e-02 -2.25948736e-01 5.98036528e-01 -9.24323142e-01
4.15500924e-02 -2.25967884e-01 -1.93028599e-01 -1.72938049e+00
-4.51205999e-01 -7.54971325e-01 8.72131944e-01 1.85703802e+00
8.67641568e-01 -6.62563145e-01 2.10108057e-01 1.25408471e-01
-3.85758698e-01 -3.29519548e-02 -8.74890327e-01 -9.71675217e-01
6.05266392e-01 -3.27332169e-01 2.41167963e-01 1.42767870e+00
1.94917738e-01 7.62664258e-01 -1.16302855e-01 4.13979143e-01
2.03550130e-01 -3.41554403e-01 4.04193342e-01 -1.09667516e+00
-4.79838133e-01 1.88763723e-01 -2.80358434e-01 -4.47719544e-01
6.53303683e-01 -1.19805849e+00 3.59275043e-01 -1.16587889e+00
-8.19001943e-02 -9.98075902e-01 -4.02050227e-01 1.17045760e+00
-3.99069220e-01 2.07608432e-01 1.38151988e-01 1.31349608e-01
-2.79156834e-01 -1.03721336e-01 8.43760669e-01 4.16476995e-01
-2.90866613e-01 -4.20751661e-01 -3.62702101e-01 7.66004324e-01
9.15639758e-01 -9.41864371e-01 2.16510966e-01 -8.21001410e-01
3.58530551e-01 -2.40244552e-01 -3.00884753e-01 -5.69081724e-01
-2.78735608e-01 -4.55081105e-01 2.17045277e-01 -2.97213852e-01
-1.23867758e-01 -4.88963753e-01 -1.36631265e-01 3.71206939e-01
1.32552966e-01 2.58627087e-01 6.21179402e-01 -6.72742948e-02
-1.66852593e-01 -6.29412532e-01 9.01568890e-01 -1.85783863e-01
-1.07242167e+00 6.14584684e-02 -3.04628462e-01 6.88066065e-01
8.48104000e-01 3.46689612e-01 -4.13921207e-01 4.85758305e-01
-5.73376536e-01 4.53317650e-02 5.28393090e-01 5.25180519e-01
-5.41721702e-01 -1.21557200e+00 -7.91048527e-01 2.06878796e-01
2.90821880e-01 -4.07116383e-01 -2.37090811e-02 5.99557281e-01
-7.47254014e-01 8.40342700e-01 -4.80847120e-01 -4.62811887e-01
-8.85249615e-01 4.93576825e-01 -1.37219861e-01 -5.75340807e-01
-5.72446287e-02 8.47772360e-01 -1.73791554e-02 -1.08297312e+00
-3.10817361e-01 -1.68900609e-01 -2.06710085e-01 2.08081603e-02
1.86480448e-01 -8.79989639e-02 1.44770682e-01 -1.08649480e+00
-7.48116612e-01 3.56234074e-01 -1.01816297e-01 -3.36080670e-01
1.30019712e+00 -1.97460242e-02 -1.46870434e-01 4.34923083e-01
9.09960806e-01 8.96081626e-01 -9.11924124e-01 -3.56787443e-01
6.03107154e-01 -7.10476413e-02 -4.05529946e-01 -9.56264794e-01
-2.44442284e-01 8.29206586e-01 -7.43613765e-02 -7.84553736e-02
4.61000949e-01 3.71777028e-01 8.45858276e-01 3.66471618e-01
6.25852466e-01 -1.39999747e+00 -5.81064165e-01 1.28199160e+00
3.23889792e-01 -1.17318189e+00 -4.91422057e-01 -6.06320679e-01
-1.02228940e+00 7.51674652e-01 3.74405175e-01 -2.27901444e-01
7.22298920e-01 2.90803492e-01 5.10843039e-01 1.06944308e-01
-5.03433526e-01 -5.27412534e-01 2.94785231e-01 1.09342694e+00
8.63493323e-01 5.61515152e-01 -6.94813550e-01 3.77004921e-01
-5.20428896e-01 -3.64059716e-01 1.39174789e-01 9.21524286e-01
-3.81300837e-01 -1.92064297e+00 -3.26331049e-01 -1.98717192e-01
-1.12743008e+00 -6.91419780e-01 -3.26646268e-01 1.10684752e+00
5.60475171e-01 6.20158374e-01 2.33591452e-01 2.13899210e-01
7.88580775e-02 4.83749121e-01 6.60440326e-01 -1.13149619e+00
-7.66006887e-01 2.51582742e-01 6.84876680e-01 1.47180364e-01
-7.18943417e-01 -1.03020716e+00 -1.26475370e+00 5.88130672e-03
-1.27451748e-01 4.43948984e-01 5.36465168e-01 1.08131862e+00
1.30620003e-01 -1.87738284e-01 3.03315938e-01 -6.65859222e-01
-2.40226060e-01 -1.19536304e+00 -4.90784734e-01 2.00911567e-01
-2.81368047e-01 -4.96023923e-01 -5.93756326e-02 2.71026105e-01] | [10.473433494567871, 9.968659400939941] |
3c121926-49ff-472c-b6f1-d259e4a22ecc | counterfactual-mean-embedding-a-kernel-method | 1805.08845 | null | https://arxiv.org/abs/1805.08845v4 | https://arxiv.org/pdf/1805.08845v4.pdf | Counterfactual Mean Embeddings | Counterfactual inference has become a ubiquitous tool in online advertisement, recommendation systems, medical diagnosis, and econometrics. Accurate modeling of outcome distributions associated with different interventions -- known as counterfactual distributions -- is crucial for the success of these applications. In this work, we propose to model counterfactual distributions using a novel Hilbert space representation called counterfactual mean embedding (CME). The CME embeds the associated counterfactual distribution into a reproducing kernel Hilbert space (RKHS) endowed with a positive definite kernel, which allows us to perform causal inference over the entire landscape of the counterfactual distribution. Based on this representation, we propose a distributional treatment effect (DTE) that can quantify the causal effect over entire outcome distributions. Our approach is nonparametric as the CME can be estimated under the unconfoundedness assumption from observational data without requiring any parametric assumption about the underlying distributions. We also establish a rate of convergence of the proposed estimator which depends on the smoothness of the conditional mean and the Radon-Nikodym derivative of the underlying marginal distributions. Furthermore, our framework allows for more complex outcomes such as images, sequences, and graphs. Our experimental results on synthetic data and off-policy evaluation tasks demonstrate the advantages of the proposed estimator. | ['Sorawit Saengkyongam', 'Sanparith Marukatat', 'Motonobu Kanagawa', 'Krikamol Muandet'] | 2018-05-22 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 2.13207424e-01 2.18821570e-01 -3.37225884e-01 -1.57517463e-01
-5.22700548e-01 -3.10035825e-01 5.36915481e-01 4.74879332e-02
-3.41826916e-01 1.00020814e+00 4.76416320e-01 -5.29319048e-01
-4.55826223e-01 -9.21994925e-01 -9.74559665e-01 -7.76403487e-01
-5.54912388e-01 -1.29169868e-02 -3.85392278e-01 2.93060601e-01
8.59505236e-02 2.28616878e-01 -1.16945207e+00 -3.01916629e-01
1.16401792e+00 7.13597834e-01 -4.61225770e-02 4.30838078e-01
9.13946703e-02 4.23597783e-01 -3.11621368e-01 -5.17014086e-01
1.02992237e-01 -3.84464622e-01 -4.77608919e-01 -2.35986635e-02
-1.25213385e-01 -2.77857333e-01 -2.12655604e-01 1.27584970e+00
4.60598320e-01 5.06944582e-02 1.02201283e+00 -1.36854053e+00
-5.61610579e-01 5.26383817e-01 -7.55709887e-01 -4.49746065e-02
6.65567666e-02 4.17995863e-02 9.24648345e-01 -5.16906559e-01
3.90087783e-01 1.22299385e+00 6.10388398e-01 2.41587147e-01
-1.54555500e+00 -6.90649450e-01 -1.67384848e-01 5.63122928e-02
-1.27489448e+00 -4.22156192e-02 9.27962482e-01 -5.54747462e-01
-1.96564287e-01 2.35388577e-01 6.62349343e-01 1.02596438e+00
5.08595884e-01 7.15437412e-01 1.39628375e+00 -2.80823946e-01
4.94255990e-01 1.36572629e-01 -2.50854284e-01 6.16631746e-01
5.25972068e-01 3.20566535e-01 1.02861682e-02 -5.23760796e-01
9.35070932e-01 1.24499515e-01 -7.02874482e-01 -7.93179750e-01
-1.23935795e+00 1.11860049e+00 1.90212026e-01 8.94355401e-02
-9.32584167e-01 1.64732412e-01 3.25790465e-01 3.65488208e-03
6.14641190e-01 7.94337541e-02 -7.54708201e-02 9.18533206e-02
-5.47825694e-01 3.41593862e-01 8.36042106e-01 4.60344970e-01
3.49712312e-01 -2.35210553e-01 -4.16171312e-01 4.92610574e-01
1.88323349e-01 8.80421042e-01 2.44724557e-01 -8.93659890e-01
2.98988849e-01 9.69148353e-02 5.87395549e-01 -9.28856552e-01
-9.89350304e-02 -4.07570988e-01 -1.09022117e+00 -9.35109928e-02
6.61979377e-01 -3.82372350e-01 -1.81410134e-01 2.10351181e+00
6.88320458e-01 7.00956523e-01 -9.00900885e-02 7.64821410e-01
-6.59818426e-02 4.84410912e-01 2.31587425e-01 -6.99760854e-01
1.08895624e+00 -1.63201708e-02 -8.81889284e-01 6.08243048e-01
5.93125463e-01 -3.85323048e-01 1.22032642e+00 5.85407093e-02
-7.22514570e-01 -9.48491693e-02 -8.08005750e-01 6.63126051e-01
2.18606740e-03 -9.98123437e-02 4.81119305e-01 7.50560462e-01
-6.67082429e-01 7.01548278e-01 -3.86951268e-01 -5.39895184e-02
4.72204000e-01 4.96744104e-02 -2.56936938e-01 -7.32941404e-02
-1.33287454e+00 5.39056838e-01 2.63674498e-01 -1.88157856e-01
-7.99947083e-01 -1.12961257e+00 -7.84487963e-01 2.20962882e-01
4.82140094e-01 -8.64926457e-01 1.01611686e+00 -7.71992743e-01
-1.45898461e+00 3.56875330e-01 1.76920146e-01 -6.81154430e-01
7.71862388e-01 1.83539957e-01 -4.93943393e-01 1.82141867e-02
1.59024999e-01 -5.66884354e-02 1.14223444e+00 -9.05659139e-01
-3.44349653e-01 -5.37844181e-01 6.93739802e-02 1.31933689e-01
-1.78618759e-01 -4.83559728e-01 3.19026321e-01 -7.57631361e-01
-3.25320095e-01 -7.74106026e-01 -2.92000383e-01 7.15078926e-03
-4.69803661e-01 -1.13399580e-01 5.01771867e-01 -6.20921016e-01
1.14994240e+00 -2.31717086e+00 -4.65559810e-02 1.67120486e-01
-1.26921413e-02 -3.01206261e-01 9.87496376e-02 2.97833741e-01
-2.80473709e-01 7.65051171e-02 -4.55022842e-01 2.26648927e-01
2.07766876e-01 -1.96213245e-01 -5.15738249e-01 1.06803834e+00
1.25602409e-01 6.05585337e-01 -9.37395632e-01 -4.71605748e-01
2.99023509e-01 4.64052141e-01 -4.77840155e-01 2.16683865e-01
-1.48007825e-01 3.68701130e-01 -6.38238668e-01 -6.50707334e-02
9.11708653e-01 -2.52685517e-01 3.35767478e-01 -1.39696911e-01
-1.27805978e-01 5.14619052e-02 -1.25438929e+00 1.36702144e+00
-6.70021176e-01 1.67735294e-01 1.86154414e-02 -1.46719635e+00
4.86907840e-01 3.82945627e-01 5.72808862e-01 -4.22362894e-01
1.03876434e-01 1.21353805e-01 -3.47901918e-02 -5.21501243e-01
-2.14539841e-02 -7.91456759e-01 -1.76096529e-01 5.34027576e-01
-2.66860664e-01 2.03546733e-01 -1.47617266e-01 3.11707966e-02
8.01131546e-01 -1.15020424e-01 7.71312654e-01 -6.89681113e-01
4.82301801e-01 -3.72350186e-01 5.17259657e-01 6.10766470e-01
-2.22813293e-01 9.20938775e-02 7.31281698e-01 -3.62004079e-02
-9.07402992e-01 -1.37000489e+00 -4.84858513e-01 2.91666776e-01
6.75017908e-02 4.37251061e-01 -8.44388127e-01 -9.50904131e-01
3.65624428e-01 1.10947335e+00 -8.64851773e-01 -3.27606350e-01
-1.12886854e-01 -8.55800509e-01 2.18640372e-01 2.44865209e-01
6.11765265e-01 -6.67380452e-01 -4.41819161e-01 1.60377979e-01
-2.54944235e-01 -6.43142998e-01 -6.47899508e-01 -5.19187748e-01
-9.21950400e-01 -1.28928864e+00 -1.03505158e+00 -5.92729747e-02
2.93095976e-01 2.24379953e-02 7.52408981e-01 -7.80975103e-01
-1.44419059e-01 6.58098876e-01 1.61045089e-01 -3.89161110e-01
-4.69063669e-01 -6.19129717e-01 2.12019339e-01 6.21881068e-01
-1.03439085e-01 -6.10875845e-01 -9.00041819e-01 1.05784468e-01
-9.79503930e-01 -1.71404448e-03 5.61643124e-01 9.15591478e-01
4.87133205e-01 3.55857372e-01 9.43054676e-01 -1.00679696e+00
7.40712345e-01 -8.90161037e-01 -8.27321470e-01 3.10151637e-01
-5.53531587e-01 4.27449733e-01 6.72044337e-01 -5.83953083e-01
-1.35530066e+00 -4.51312304e-01 3.86336982e-01 -2.93491900e-01
-7.74936453e-02 6.60242200e-01 -3.57245147e-01 5.81247866e-01
3.96869868e-01 1.88969523e-01 2.22009867e-01 -4.52468961e-01
6.91206634e-01 6.52496517e-01 5.83815992e-01 -4.84841794e-01
4.97330636e-01 9.65080500e-01 3.52647781e-01 -8.50104868e-01
-7.40766585e-01 -3.14957500e-01 -1.02804393e-01 -5.41196801e-02
7.53786564e-01 -5.81432283e-01 -1.13032484e+00 1.27496272e-01
-9.42877769e-01 -5.94356917e-02 -5.74283123e-01 1.07681286e+00
-9.22726691e-01 3.61828774e-01 -1.91869214e-01 -1.30758965e+00
-8.85460600e-02 -7.52089500e-01 7.06531942e-01 -1.99299473e-02
1.12329282e-01 -1.38040555e+00 2.56094038e-01 -1.13910124e-01
2.31199771e-01 7.36931145e-01 1.15561521e+00 -1.73608795e-01
-2.41025358e-01 -3.31384569e-01 -3.36619347e-01 2.25973189e-01
2.85669476e-01 -2.89389849e-01 -6.88880205e-01 -1.87458172e-01
3.09569418e-01 1.79585889e-01 6.05922043e-01 1.03500104e+00
1.21344078e+00 -7.98725069e-01 -3.80599797e-01 5.44186905e-02
1.45454872e+00 -4.21871133e-02 5.48750222e-01 -2.30999157e-01
2.69674927e-01 8.21068168e-01 8.19024563e-01 8.15089285e-01
2.67058760e-01 5.27476192e-01 4.62163299e-01 1.30465984e-01
3.35384488e-01 -5.51426768e-01 3.55771720e-01 3.50710660e-01
1.75660297e-01 -1.49584413e-02 -4.33573037e-01 8.11836123e-01
-1.78799248e+00 -1.22066844e+00 -1.05302930e-01 2.87040210e+00
7.34984994e-01 -2.12457493e-01 3.57785285e-01 8.15209746e-02
1.18196118e+00 -9.33184624e-02 -6.34794593e-01 -2.15418980e-01
1.47333756e-01 1.41299516e-01 7.19578028e-01 4.18902814e-01
-9.43905056e-01 5.44641651e-02 5.32638454e+00 9.76759851e-01
-1.00263119e+00 2.07211420e-01 5.90379000e-01 1.70078337e-01
-5.80603004e-01 3.66124138e-02 -1.24850884e-01 7.96236694e-01
1.04312241e+00 -6.95349157e-01 2.78746665e-01 7.26507783e-01
7.19787121e-01 -1.09344460e-01 -8.92351925e-01 7.92306364e-01
-4.68245536e-01 -1.36209357e+00 -2.02727854e-01 5.67506075e-01
6.12300575e-01 -5.18800318e-01 4.14781660e-01 1.23427108e-01
3.96745503e-01 -7.57125974e-01 6.44532442e-01 7.02680647e-01
1.05995500e+00 -9.39264894e-01 6.59656346e-01 4.54820931e-01
-7.88419843e-01 -1.55427456e-01 -2.42143184e-01 3.11654937e-02
1.53183669e-01 1.06377614e+00 -8.85124326e-01 7.11806476e-01
1.47670284e-01 4.90165502e-01 1.53209910e-01 9.86793101e-01
-6.94741309e-02 8.10819387e-01 -2.36828938e-01 -6.72040880e-02
-1.72320589e-01 -4.05546784e-01 6.66582644e-01 8.92876327e-01
5.66324592e-01 1.14900507e-01 -2.17532977e-01 9.86300230e-01
-1.61367685e-01 3.03768128e-01 -9.00798261e-01 -2.21649140e-01
4.58094656e-01 7.68290102e-01 -3.39522302e-01 -2.45445430e-01
-3.42446893e-01 7.70032585e-01 5.96376359e-02 5.03751278e-01
-1.02301860e+00 -2.09077328e-01 7.11385012e-01 1.64980456e-01
1.72174975e-01 2.14044645e-01 -1.27279028e-01 -1.20279539e+00
6.01377636e-02 -5.85472703e-01 6.22836292e-01 -1.81044370e-01
-1.33917022e+00 -2.38954827e-01 3.45363587e-01 -1.06649613e+00
-3.70166749e-01 -3.30741793e-01 -7.02990592e-01 8.75657856e-01
-1.31922793e+00 -6.51620567e-01 3.28286618e-01 5.69684446e-01
-7.32348561e-02 3.04857731e-01 7.53940463e-01 2.68727839e-02
-5.30193150e-01 2.65699506e-01 4.72139299e-01 -1.55597255e-01
3.29373986e-01 -1.33909369e+00 -3.23908240e-01 5.09810150e-01
-1.13971435e-01 4.23407406e-01 1.01737046e+00 -6.16604805e-01
-1.14937365e+00 -1.09516287e+00 5.91904163e-01 1.69600427e-01
8.29492033e-01 -1.79484878e-02 -5.93880832e-01 3.91661733e-01
8.11360683e-03 1.84281338e-02 6.04615211e-01 1.44783109e-01
-1.93476006e-01 -2.11174041e-01 -1.34825873e+00 7.42049158e-01
6.82440400e-01 -3.49727243e-01 -3.77289295e-01 3.45395327e-01
7.10176110e-01 2.09775969e-01 -8.84539545e-01 4.10373151e-01
5.92032313e-01 -8.67436111e-01 8.64586890e-01 -8.15124571e-01
4.04664963e-01 -1.59864381e-01 -1.92370787e-01 -1.41785812e+00
-1.13257661e-01 -5.83003938e-01 -1.18420765e-01 1.02031100e+00
7.57045746e-02 -8.91131461e-01 3.48440528e-01 4.49591756e-01
4.95794892e-01 -6.68308854e-01 -1.08933175e+00 -9.16741431e-01
3.16970833e-02 -5.99121928e-01 8.24277759e-01 1.05637062e+00
7.09179267e-02 7.04449937e-02 -4.31099832e-01 3.84216219e-01
1.13527644e+00 3.48176420e-01 6.04981422e-01 -1.07789445e+00
-6.30574048e-01 -2.07191125e-01 -3.25485706e-01 -4.66836780e-01
3.18069398e-01 -6.52327299e-01 -2.13024378e-01 -1.06817496e+00
4.09708798e-01 -3.58381271e-01 -3.61569136e-01 -2.58089691e-01
-1.33345097e-01 -2.90389985e-01 -1.85677841e-01 -6.98618293e-02
-2.03568846e-01 1.04121375e+00 1.23725212e+00 1.32253677e-01
4.49252427e-02 2.91606963e-01 -4.17938650e-01 7.09160149e-01
6.07456326e-01 -4.00935113e-01 -6.18075192e-01 4.57574278e-01
-1.49451330e-01 6.43104196e-01 7.56560504e-01 -4.14395034e-01
-4.35000569e-01 -3.83830339e-01 -1.04763597e-01 -2.46986181e-01
-1.91000048e-02 -7.24191487e-01 4.64590907e-01 6.96926415e-01
-3.64697337e-01 -3.89606684e-01 -6.56973273e-02 1.16133332e+00
6.17085658e-02 -1.52470276e-01 8.23612809e-01 2.53265679e-01
-1.43132761e-01 2.47881129e-01 -3.90656702e-02 3.13302279e-01
1.10002553e+00 2.12123796e-01 -7.57812709e-02 -6.83517337e-01
-6.81939185e-01 7.37168565e-02 2.85832196e-01 3.23274247e-02
5.14704168e-01 -1.61822677e+00 -7.76908100e-01 -1.58032015e-01
-2.25363635e-02 -6.21675849e-01 6.15525544e-01 1.17562759e+00
1.19668962e-02 4.64764267e-01 1.67059511e-01 -3.53181660e-01
-6.83056653e-01 8.55284035e-01 2.73883373e-01 -4.32720840e-01
-4.57573593e-01 1.24541253e-01 7.28916585e-01 -2.71963775e-01
-2.14709669e-01 -2.04783857e-01 -1.64587557e-01 -1.90625414e-01
5.23045719e-01 6.03471398e-01 -4.32917148e-01 -2.34655872e-01
-1.53314725e-01 -4.69401143e-02 2.16932699e-01 -4.30717200e-01
1.10957420e+00 -2.87172079e-01 6.28723353e-02 5.91643751e-01
1.40606475e+00 2.65668016e-02 -1.34758329e+00 -1.46122590e-01
-3.83394435e-02 -5.69965541e-01 1.46415174e-01 -4.22929168e-01
-6.00047767e-01 7.88527608e-01 6.41534805e-01 4.39230680e-01
1.05246830e+00 -2.61035580e-02 3.52458835e-01 -4.06131983e-01
3.36880028e-01 -8.48634064e-01 -3.80805701e-01 -4.30026948e-01
9.25568938e-01 -9.32751715e-01 -2.67987043e-01 -4.84886497e-01
-5.08149207e-01 6.18829429e-01 -2.23586500e-01 -1.48802772e-01
1.06051481e+00 -1.99718803e-01 -4.77692902e-01 -2.63973314e-04
-5.09728193e-01 -1.33453742e-01 3.49243879e-01 5.41158378e-01
4.73517418e-01 7.04382539e-01 -7.62902141e-01 5.43107212e-01
-2.68166568e-02 2.18581751e-01 5.68335056e-01 3.54938060e-01
5.49808890e-02 -6.97281539e-01 -4.30546492e-01 5.88610649e-01
-6.72520876e-01 1.45921092e-02 2.81366765e-01 9.05391991e-01
-1.49546638e-01 7.16115892e-01 4.35908325e-02 1.55945837e-01
2.67057240e-01 1.01382639e-02 4.75473583e-01 -3.05863976e-01
5.30214489e-01 7.28600547e-02 -1.49727538e-01 -5.64216673e-01
-4.12633032e-01 -9.18222845e-01 -9.04167533e-01 -3.78401041e-01
-4.30965036e-01 4.42939073e-01 6.73311174e-01 8.85577619e-01
2.24519506e-01 2.50877351e-01 9.95885551e-01 -2.10096583e-01
-1.33247590e+00 -9.06581163e-01 -1.18467438e+00 4.65911090e-01
3.32480758e-01 -8.07493269e-01 -4.17485744e-01 -1.45920128e-01] | [8.019953727722168, 5.302900791168213] |
e3733395-ae0b-4ba4-9420-f8f5dc0d6e54 | unsupervised-learning-of-parsimonious-general | 1704.03507 | null | http://arxiv.org/abs/1704.03507v2 | http://arxiv.org/pdf/1704.03507v2.pdf | Unsupervised Learning of Parsimonious General-Purpose Embeddings for User and Location Modelling | Many social network applications depend on robust representations of
spatio-temporal data. In this work, we present an embedding model based on
feed-forward neural networks which transforms social media check-ins into dense
feature vectors encoding geographic, temporal, and functional aspects for
modelling places, neighborhoods, and users. We employ the embedding model in a
variety of applications including location recommendation, urban functional
zone study, and crime prediction. For location recommendation, we propose a
Spatio-Temporal Embedding Similarity algorithm (STES) based on the embedding
model.
In a range of experiments on real life data collected from Foursquare, we
demonstrate our model's effectiveness at characterizing places and people and
its applicability in aforementioned problem domains. Finally, we select eight
major cities around the globe and verify the robustness and generality of our
model by porting pre-trained models from one city to another, thereby
alleviating the need for costly local training. | ['Eickhoff Carsten', 'Yang Jing'] | 2018-01-22 | null | null | null | null | ['crime-prediction'] | ['miscellaneous'] | [-3.60861301e-01 -2.90373266e-01 -2.68660754e-01 -3.66063893e-01
-2.91153669e-01 -1.84276938e-01 7.79764593e-01 5.20539761e-01
-3.33027661e-01 3.88599426e-01 7.88220227e-01 -5.23563623e-01
-3.25813890e-01 -1.10996270e+00 -2.82181889e-01 -2.40667596e-01
-4.64806348e-01 -2.28363588e-01 1.16503514e-01 -3.12632680e-01
1.85593799e-01 3.42542708e-01 -1.22502148e+00 1.34836540e-01
7.41426885e-01 7.08935082e-01 -5.62953250e-03 3.01742047e-01
1.62394181e-01 6.09917879e-01 -1.52349412e-01 -3.42582345e-01
1.87821686e-01 4.19394560e-02 -4.89590794e-01 -1.28798142e-01
-3.58382538e-02 -3.03694338e-01 -8.72559488e-01 5.45794547e-01
3.83393973e-01 6.39502287e-01 8.67950916e-01 -1.19886982e+00
-1.40848649e+00 3.86687130e-01 -3.92840385e-01 4.27125037e-01
3.40646744e-01 -7.89713711e-02 1.05257881e+00 -1.02721798e+00
2.60452718e-01 8.77147734e-01 1.10590076e+00 -9.56953131e-03
-8.12197983e-01 -4.29551601e-01 2.87285686e-01 1.76333338e-01
-1.75289679e+00 -3.84361267e-01 7.77884364e-01 -5.03540099e-01
1.18102944e+00 1.85887262e-01 5.61139882e-01 9.36014533e-01
9.56862420e-02 6.09986186e-01 4.26528096e-01 -5.73293604e-02
2.45583668e-01 2.04805687e-01 1.29344724e-02 6.35548830e-01
2.27714270e-01 -6.71893507e-02 -2.59326100e-01 -5.39320290e-01
5.03378928e-01 6.57520652e-01 5.29865846e-02 -1.66571304e-01
-1.18278885e+00 1.01432753e+00 9.04438913e-01 4.19750422e-01
-4.52041030e-01 2.03563735e-01 1.93927392e-01 3.88230272e-02
9.34853792e-01 1.31982192e-01 -1.24447815e-01 1.79319844e-01
-9.96474564e-01 6.11524619e-02 3.18629205e-01 7.49272466e-01
7.31729448e-01 -1.16980374e-02 -2.16577321e-01 8.81603956e-01
5.44485331e-01 4.86095637e-01 5.30504942e-01 -5.32813132e-01
5.54032505e-01 8.54450524e-01 9.83374268e-02 -1.94591570e+00
-4.61560190e-01 -2.90998638e-01 -1.14617085e+00 -6.74006343e-01
-1.51136085e-01 -1.27685159e-01 -5.09570122e-01 1.47752023e+00
3.23497981e-01 7.86308646e-01 -1.41787767e-01 4.84027237e-01
7.63118684e-01 8.41137409e-01 1.18320026e-01 3.37061614e-01
9.27464843e-01 -8.74355197e-01 -3.26364994e-01 -1.28865093e-01
9.55446005e-01 -2.21534714e-01 7.04794228e-01 -4.31042254e-01
-7.35213339e-01 -6.21416450e-01 -5.56193829e-01 -2.00424775e-01
-9.83211040e-01 1.97602451e-01 7.96564281e-01 3.12762618e-01
-1.15547085e+00 5.62714875e-01 -7.79948950e-01 -6.85345769e-01
2.97876835e-01 2.09296986e-01 -5.45581341e-01 4.83083576e-02
-1.36678505e+00 4.05312598e-01 1.25449359e-01 5.05951457e-02
1.97121948e-02 -6.30833745e-01 -1.23332417e+00 3.89892280e-01
-3.76335472e-01 -4.62642640e-01 5.01839578e-01 -2.73655891e-01
-8.09176266e-01 2.96908975e-01 -3.98184985e-01 -4.13157672e-01
-5.89632727e-02 1.54649422e-01 -9.65139210e-01 -2.31674537e-01
4.81812209e-01 4.43198949e-01 4.16214287e-01 -6.72369659e-01
-5.91294229e-01 -1.78760946e-01 1.50028214e-01 -1.09252386e-01
-1.00775671e+00 5.10024372e-03 -3.71829063e-01 -8.14354599e-01
1.64822385e-01 -6.90694809e-01 -4.27866012e-01 7.34073296e-02
-1.67720005e-01 -3.25720012e-01 7.36876011e-01 -7.81934798e-01
1.65204060e+00 -2.44631243e+00 -3.65740836e-01 6.18994296e-01
2.77986288e-01 1.20668694e-01 -4.98066843e-01 8.70012462e-01
1.04085334e-01 2.64166266e-01 -6.62717894e-02 -3.90137732e-01
2.29838297e-01 -8.38716030e-02 -3.02125782e-01 6.84608698e-01
2.63271779e-01 1.20529568e+00 -1.07716644e+00 -2.54224956e-01
3.44272286e-01 7.88507164e-01 -7.51699090e-01 -1.79524511e-01
3.79522473e-01 5.73788732e-02 -4.94682074e-01 4.96569157e-01
3.58961433e-01 -5.00098169e-01 -8.04984476e-03 8.70446637e-02
-9.35318097e-02 3.66765231e-01 -9.54224646e-01 1.18004560e+00
-7.07755208e-01 7.79882729e-01 -5.46828866e-01 -7.75172949e-01
8.97980392e-01 2.08089381e-01 6.10672474e-01 -9.19887006e-01
1.60789732e-02 -8.08594152e-02 -3.26148599e-01 -5.28280914e-01
8.33675027e-01 5.33714116e-01 -2.24823833e-01 8.75847757e-01
-2.10967034e-01 5.78371942e-01 -2.43627597e-02 1.89041972e-01
1.19857705e+00 -4.74147826e-01 3.88539314e-01 -1.62057996e-01
5.52492976e-01 -4.50476557e-01 2.67622560e-01 5.05919337e-01
-4.10838872e-01 3.78756166e-01 6.58195540e-02 -7.87233055e-01
-9.80837286e-01 -8.57572615e-01 -5.10469042e-02 1.07853162e+00
-9.12086591e-02 -5.19029617e-01 -8.18893909e-02 -4.96765733e-01
2.50358254e-01 4.64115024e-01 -8.91045451e-01 -2.93596298e-01
-2.95652509e-01 -6.51422262e-01 4.65270489e-01 7.41561234e-01
5.39841712e-01 -6.85809910e-01 -8.99620503e-02 1.40371382e-01
-2.32013330e-01 -8.88952553e-01 -5.32265961e-01 -5.19570410e-01
-3.93493205e-01 -8.02379072e-01 -7.30976760e-01 -8.91732097e-01
6.65101469e-01 8.83797467e-01 6.99289739e-01 2.06134737e-01
2.00726599e-01 3.28928709e-01 -4.02214795e-01 8.44691545e-02
2.63848484e-01 4.69628751e-01 2.84548342e-01 3.87653470e-01
7.48608828e-01 -8.72996092e-01 -7.34542012e-01 3.81369531e-01
-6.66917145e-01 -1.89085558e-01 1.07853860e-01 5.57525039e-01
4.83153090e-02 1.53687522e-01 7.68867731e-01 -5.40540695e-01
8.76765847e-01 -1.35341322e+00 -2.47785866e-01 1.90802962e-01
-3.46463293e-01 -2.55839139e-01 4.37810719e-01 -2.89083958e-01
-4.94822919e-01 -2.41641685e-01 -5.54767400e-02 -4.61744256e-02
6.39701113e-02 8.04170191e-01 2.33483016e-01 3.86321619e-02
6.26199424e-01 3.84007305e-01 -4.20167476e-01 -3.42407197e-01
4.30240721e-01 1.10709715e+00 1.06921099e-01 -9.27220508e-02
9.53322589e-01 5.61107337e-01 -4.26072955e-01 -9.68291223e-01
-4.31029439e-01 -8.43220472e-01 -4.95316356e-01 8.23842958e-02
5.88840604e-01 -1.06945813e+00 -6.51660383e-01 6.24175183e-02
-9.49892998e-01 -2.02675223e-01 7.18036816e-02 5.69332480e-01
-1.43932298e-01 2.09084272e-01 -5.27073264e-01 -7.28086710e-01
-1.23366155e-01 -5.21855712e-01 1.00204515e+00 -1.64498165e-02
-2.87457854e-01 -1.59093332e+00 4.20690924e-01 8.42481479e-02
7.52224088e-01 1.59305796e-01 7.64584363e-01 -5.21205902e-01
-4.28714842e-01 -4.03230906e-01 -5.77369630e-01 -1.33902475e-01
4.63676244e-01 -1.24989271e-01 -8.53855371e-01 -1.99512482e-01
-6.22555256e-01 6.85784444e-02 9.53282356e-01 2.69554824e-01
1.29561663e+00 -4.23641235e-01 -5.64748347e-01 6.45357132e-01
1.28889465e+00 -1.15437791e-01 4.19419974e-01 3.89813274e-01
7.26162434e-01 3.51420254e-01 1.47629797e-01 6.32468700e-01
8.99145424e-01 4.83467788e-01 1.15569793e-01 -2.17934772e-01
2.48261079e-01 -5.29219389e-01 1.85054511e-01 7.79532313e-01
-1.14628360e-01 -3.87942165e-01 -9.77160633e-01 1.22979665e+00
-2.07383537e+00 -1.31135380e+00 1.46650791e-01 1.77610075e+00
1.78703874e-01 -2.23562047e-01 2.16652945e-01 4.59718406e-02
4.82868373e-01 5.96825480e-01 -2.66737372e-01 -2.91705638e-01
1.19607359e-01 -1.32221892e-01 4.94085073e-01 4.01447892e-01
-1.24443150e+00 7.85688937e-01 6.48941994e+00 5.40665865e-01
-1.19361317e+00 9.29642469e-02 5.86259544e-01 -1.44518316e-02
-6.00752890e-01 -4.13441420e-01 -4.06871498e-01 6.55424953e-01
1.18803930e+00 -1.99283257e-01 5.04266441e-01 8.17991316e-01
6.56442523e-01 4.14816469e-01 -7.62741923e-01 8.14481437e-01
2.17810515e-02 -1.66601968e+00 -5.45600429e-02 1.89021051e-01
8.75908136e-01 2.75268048e-01 4.08631116e-01 4.60573643e-01
1.87793955e-01 -1.02856350e+00 3.58310401e-01 5.02400279e-01
6.87524736e-01 -6.97127044e-01 6.53378427e-01 2.72717953e-01
-1.53429973e+00 -3.18264902e-01 -3.90916854e-01 -2.81157851e-01
1.67078435e-01 4.15538400e-01 -8.47395837e-01 4.23781931e-01
5.63772321e-01 1.31303227e+00 -6.99088573e-01 9.21987534e-01
1.68979485e-02 5.79788446e-01 -2.28098944e-01 -1.01090200e-01
4.97496516e-01 -5.55965714e-02 -5.21493964e-02 1.20688200e+00
5.54210424e-01 1.11251161e-01 -2.97769140e-02 7.97823548e-01
-1.93209052e-01 1.89552978e-01 -1.08657694e+00 -9.36562270e-02
6.69785976e-01 9.75380540e-01 -4.40342277e-01 3.40345851e-03
-6.83520377e-01 8.25068235e-01 6.31016016e-01 5.57545602e-01
-1.00243449e+00 -5.00384033e-01 7.93082774e-01 3.18333119e-01
4.04826492e-01 -4.50129300e-01 -1.27451882e-01 -1.23994720e+00
-5.65364920e-02 -2.16261685e-01 1.25698550e-02 -6.62087381e-01
-1.47039056e+00 4.55506086e-01 -4.53093909e-02 -1.27643681e+00
-2.88602442e-01 -3.74034137e-01 -9.70283508e-01 7.50515461e-01
-1.54084599e+00 -1.24086702e+00 3.58180776e-02 7.13944733e-01
2.21452817e-01 -3.59499842e-01 8.15626204e-01 7.94210970e-01
-6.91316307e-01 7.31267571e-01 4.79849428e-01 5.80985963e-01
2.01019377e-01 -7.09384620e-01 1.03452170e+00 8.07719231e-01
2.73956925e-01 9.48128343e-01 6.48344457e-02 -4.98980582e-01
-9.45676386e-01 -1.70423639e+00 1.74230635e+00 -3.52002203e-01
9.70162511e-01 -4.46736753e-01 -6.30100131e-01 9.50783134e-01
-3.95953618e-02 2.77609795e-01 1.14121258e+00 4.86205131e-01
-3.03774118e-01 5.91114275e-02 -1.16469085e+00 8.20417702e-01
1.17874777e+00 -9.17041481e-01 -3.32561821e-01 4.45864111e-01
6.18296504e-01 2.96966940e-01 -8.52712452e-01 4.95843310e-03
5.48662901e-01 -7.06396401e-01 1.15705848e+00 -7.31229484e-01
4.87583756e-01 -1.11611567e-01 -4.53895062e-01 -1.25678730e+00
-1.01362085e+00 -7.25326613e-02 -1.97890043e-01 1.01590705e+00
3.95531863e-01 -7.93582976e-01 6.06837571e-01 7.68832922e-01
1.88677594e-01 -8.54593635e-01 -7.84410238e-01 -4.97616202e-01
-1.74051642e-01 -5.91550827e-01 1.14790750e+00 1.31122911e+00
1.58322588e-01 1.67802691e-01 -4.44665343e-01 4.67126369e-01
1.17881857e-01 -4.46250290e-02 6.28929913e-01 -1.08606565e+00
-6.62944838e-02 -2.34445021e-01 -7.69991517e-01 -9.05520320e-01
3.53017956e-01 -8.15822840e-01 -5.45784593e-01 -1.68779647e+00
1.37361512e-01 -7.19636858e-01 -5.52940726e-01 5.98594427e-01
-3.31513695e-02 6.30678713e-01 -1.47948503e-01 8.91647339e-02
-7.41896331e-01 7.55708992e-01 5.22108078e-01 -4.78125393e-01
-3.27240199e-01 1.32280037e-01 -7.55517662e-01 5.98775506e-01
8.38601053e-01 -2.53820240e-01 -7.16310620e-01 -7.63343394e-01
5.38832843e-01 -3.42296213e-01 4.37706500e-01 -9.70105827e-01
2.14430466e-01 -3.11358154e-01 3.58404726e-01 -5.51043749e-01
4.95300740e-01 -9.07701015e-01 -1.04334153e-01 1.31791547e-01
-2.65491009e-01 2.84248948e-01 3.98659334e-02 8.43835533e-01
-1.21786118e-01 3.97049963e-01 1.22309767e-01 3.04284036e-01
-8.83708417e-01 6.72612190e-01 -4.10120249e-01 -2.99307019e-01
1.02177715e+00 -3.50330144e-01 -1.19805336e-01 -6.25348866e-01
-6.77289665e-01 2.37276644e-01 3.44649106e-01 5.55447698e-01
8.48907650e-01 -1.78440320e+00 -6.25498414e-01 6.15015149e-01
4.39152688e-01 -6.65375113e-01 2.92472005e-01 6.65266037e-01
-3.57073903e-01 6.56180322e-01 1.38133109e-01 -1.44897059e-01
-8.10633004e-01 5.98811686e-01 1.64960131e-01 -6.66991770e-02
-4.74719971e-01 4.45829213e-01 -1.06544599e-01 -7.10676849e-01
3.80045101e-02 -5.32393813e-01 -4.22078311e-01 -2.35660955e-01
7.10524142e-01 5.13496876e-01 -1.39522403e-01 -8.85313094e-01
-4.90867019e-01 2.55755246e-01 2.78427720e-01 3.08841635e-02
1.66927207e+00 -3.05457532e-01 1.59730151e-01 3.99786800e-01
1.48138428e+00 7.51526793e-04 -7.97395825e-01 -4.95798141e-01
-7.10297078e-02 -5.84935188e-01 1.66257903e-01 -1.59036830e-01
-9.62549567e-01 7.90967643e-01 4.21245366e-01 4.33543056e-01
7.77717173e-01 -7.14780614e-02 1.05604959e+00 5.42774737e-01
2.49343783e-01 -8.05132866e-01 -2.18900621e-01 6.50584221e-01
4.77341294e-01 -1.25382054e+00 -1.25958771e-01 4.48898226e-02
-4.10159111e-01 7.03010559e-01 1.54077441e-01 -3.35029632e-01
1.40335119e+00 -3.86879891e-01 -3.11639190e-01 -1.64560407e-01
-6.11885846e-01 -2.62445211e-01 5.05251169e-01 5.96006632e-01
3.92514408e-01 1.88559502e-01 -8.25512856e-02 4.77129966e-01
-1.01726867e-01 -5.59097454e-02 1.04789920e-01 8.08265448e-01
-4.02573556e-01 -7.37457454e-01 4.07380164e-02 4.76608783e-01
1.85800549e-02 -3.05062771e-01 -1.54828057e-01 6.18584871e-01
2.63636410e-01 9.53624845e-01 3.67484421e-01 -7.27928936e-01
-5.14347199e-03 -1.88481271e-01 -2.15093493e-01 -5.84958196e-01
-4.46466058e-01 -5.12151659e-01 -1.90936983e-01 -4.21523809e-01
-4.13895398e-01 -5.68353117e-01 -8.63770187e-01 -8.71232867e-01
2.61485334e-02 4.61782105e-02 4.47643876e-01 8.76531303e-01
9.19929028e-01 1.56239137e-01 1.00751424e+00 -7.64233410e-01
-1.20607421e-01 -9.39353526e-01 -6.11966372e-01 3.16642672e-01
4.17586863e-01 -5.49421906e-01 -1.05245002e-01 -3.22674483e-01] | [6.6155686378479, 2.09202241897583] |
af539d5e-fe8e-4a43-87c7-b75564f7b6f1 | making-a-spiking-net-work-robust-brain-like | 2208.01204 | null | https://arxiv.org/abs/2208.01204v2 | https://arxiv.org/pdf/2208.01204v2.pdf | Making a Spiking Net Work: Robust brain-like unsupervised machine learning | The surge in interest in Artificial Intelligence (AI) over the past decade has been driven almost exclusively by advances in Artificial Neural Networks (ANNs). While ANNs set state-of-the-art performance for many previously intractable problems, the use of global gradient descent necessitates large datasets and computational resources for training, potentially limiting their scalability for real-world domains. Spiking Neural Networks (SNNs) are an alternative to ANNs that use more brain-like artificial neurons and can use local unsupervised learning to rapidly discover sparse recognizable features in the input data. SNNs, however, struggle with dynamical stability and have failed to match the accuracy of ANNs. Here we show how an SNN can overcome many of the shortcomings that have been identified in the literature, including offering a principled solution to the dynamical "vanishing spike problem", to outperform all existing shallow SNNs and equal the performance of an ANN. It accomplishes this while using unsupervised learning with unlabeled data and only 1/50th of the training epochs (labeled data is used only for a simple linear readout layer). This result makes SNNs a viable new method for fast, accurate, efficient, explainable, and re-deployable machine learning with unlabeled data. | ['Tara J. Hamilton', 'Allen Cheung', 'Chip Essam', 'Andrew Wabnitz', 'Peter G. Stratton'] | 2022-08-02 | null | null | null | null | ['machine-learning', 'machine-learning'] | ['methodology', 'miscellaneous'] | [ 2.33720616e-01 -6.28239885e-02 2.09240824e-01 -2.06201568e-01
5.70147000e-02 -5.08148909e-01 4.87071902e-01 4.40447070e-02
-6.47526443e-01 9.74219799e-01 -3.57466519e-01 -2.44335189e-01
-7.03494623e-02 -5.84684491e-01 -5.21719873e-01 -1.00146914e+00
1.04124121e-01 5.05907655e-01 4.77512449e-01 -3.17980796e-01
3.97894651e-01 7.35448718e-01 -1.77438498e+00 2.08203942e-01
5.92915058e-01 1.16645992e+00 2.64696985e-01 3.22557926e-01
-2.91961551e-01 7.51607239e-01 -4.17937905e-01 1.34508267e-01
2.19877198e-01 -5.58472157e-01 -4.72926885e-01 -3.90786886e-01
-5.15860766e-02 1.43506110e-01 -4.49325740e-01 7.57036686e-01
6.56913459e-01 -1.35522157e-01 6.27352417e-01 -1.04135334e+00
-5.65305889e-01 3.71815294e-01 -1.74795017e-01 5.54142296e-01
-1.67724818e-01 3.40687364e-01 5.47030628e-01 -9.46326852e-01
6.14935398e-01 6.88882113e-01 9.97788489e-01 1.06633890e+00
-1.47943103e+00 -5.71329772e-01 -1.69514775e-01 -9.34689268e-02
-1.18265009e+00 -7.30652928e-01 5.31436026e-01 -2.50915885e-01
1.45156634e+00 -2.82594580e-02 1.10766244e+00 1.01116061e+00
4.16716278e-01 4.69044656e-01 1.09513950e+00 -3.46219361e-01
7.42727935e-01 -1.10482769e-02 8.47307369e-02 4.15386975e-01
6.22376442e-01 2.35212311e-01 -6.93761349e-01 1.11924767e-01
1.04794323e+00 6.11307397e-02 9.66935158e-02 -2.94835001e-01
-1.11950660e+00 5.55493832e-01 4.12567258e-01 6.90675557e-01
-3.76541853e-01 2.81084836e-01 2.27132261e-01 3.58243793e-01
1.50135651e-01 8.31537008e-01 -6.04880631e-01 -3.44511241e-01
-1.11832571e+00 1.08567104e-01 8.63130927e-01 3.26056302e-01
6.47769809e-01 8.05519283e-01 4.95194823e-01 5.89995503e-01
9.22536030e-02 4.36328828e-01 1.11911750e+00 -8.01046431e-01
-1.23844877e-01 1.09066308e+00 -1.42365992e-01 -7.83814192e-01
-7.78269589e-01 -6.81584775e-01 -1.15472388e+00 5.18255770e-01
5.35299599e-01 -2.00376123e-01 -1.14787698e+00 1.54171515e+00
-1.28335297e-01 9.34992954e-02 3.00411463e-01 7.33438432e-01
6.65702164e-01 5.47656059e-01 -1.50928915e-01 -1.27785623e-01
6.58327520e-01 -6.50117397e-01 -4.23072696e-01 -6.85327530e-01
5.79319000e-01 -2.71618426e-01 7.43039548e-01 3.56522173e-01
-1.08239794e+00 -2.48305053e-01 -1.25431776e+00 1.65930450e-01
-8.20528507e-01 -1.97822571e-01 9.12192106e-01 6.30724192e-01
-1.34098196e+00 7.50768781e-01 -1.19960070e+00 -6.41729772e-01
5.82969904e-01 1.02128911e+00 -3.20530981e-01 2.68362999e-01
-7.04494298e-01 1.24913204e+00 5.21578610e-01 2.32107297e-01
-5.75593352e-01 -5.89337885e-01 -4.66703832e-01 -1.52103931e-01
-2.65871167e-01 -4.95012403e-01 9.51476872e-01 -1.09007001e+00
-1.60185790e+00 8.42087984e-01 -4.40211236e-01 -9.37137425e-01
-1.46933064e-01 2.14475125e-01 -3.75949502e-01 8.45173281e-03
-2.66757309e-01 8.43662560e-01 5.44064343e-01 -9.65090692e-01
-1.05882585e-01 -2.89909333e-01 -6.20270848e-01 -1.91595033e-01
-5.38554430e-01 -7.98382089e-02 3.15265477e-01 -3.89116526e-01
6.23315930e-01 -8.95812333e-01 -3.87075186e-01 1.81446169e-02
9.12595019e-02 1.24634445e-01 1.08094740e+00 -1.64243937e-01
9.15755510e-01 -2.00958133e+00 -1.16935093e-02 1.86991692e-01
3.50908697e-01 6.71929240e-01 -2.97024529e-02 4.49908763e-01
1.06802946e-02 -1.44313890e-02 -5.22374868e-01 1.17789976e-01
-4.35513347e-01 4.77887034e-01 -2.75689781e-01 3.06752473e-01
4.91864145e-01 1.25934553e+00 -7.94329286e-01 -5.81682958e-02
1.28574491e-01 4.58192110e-01 -2.61933416e-01 -8.65107402e-02
-7.87581727e-02 7.85355449e-01 -8.94137025e-02 6.87325954e-01
2.46745497e-01 -6.33140028e-01 1.33602589e-01 1.38726383e-01
-4.08065677e-01 1.88883096e-01 -1.01058435e+00 1.50828898e+00
-1.22968949e-01 9.84896719e-01 -2.03310266e-01 -1.50884473e+00
1.29260933e+00 2.74748415e-01 6.38357580e-01 -1.03945673e+00
1.12119511e-01 8.65780652e-01 4.69183981e-01 -2.89438844e-01
-3.06858629e-01 -2.27585047e-01 3.40272427e-01 4.33728725e-01
4.40509081e-01 8.69601406e-03 1.07722200e-01 -9.98275951e-02
1.36421967e+00 8.54726285e-02 1.47373542e-01 -7.04314470e-01
2.05851778e-01 1.25075132e-01 5.32024622e-01 7.52925634e-01
-3.39521579e-02 6.94707096e-01 2.08787933e-01 -1.15615010e+00
-1.26566541e+00 -8.96756947e-01 -2.27386743e-01 7.10390866e-01
1.09278001e-01 2.38911454e-02 -6.61835432e-01 1.37104914e-01
-7.38446489e-02 2.76335627e-01 -5.51026523e-01 -2.48114660e-01
-5.67543805e-01 -1.19389307e+00 7.61007309e-01 5.81252098e-01
6.13370717e-01 -1.46301734e+00 -1.24752474e+00 7.39271402e-01
2.35460624e-01 -1.06384647e+00 6.32832706e-01 9.29482698e-01
-1.42476714e+00 -7.06908822e-01 -6.53013527e-01 -8.92500877e-01
8.69143784e-01 -1.45687595e-01 1.01842117e+00 1.94440171e-01
-7.56246984e-01 3.90600376e-02 -3.50324102e-02 -6.79514110e-01
-7.96165094e-02 1.89143300e-01 5.88456154e-01 -1.99522138e-01
5.63619971e-01 -1.17377925e+00 -6.44287944e-01 1.61809355e-01
-8.70585918e-01 8.83966610e-02 7.25244164e-01 9.99130428e-01
5.65054059e-01 -2.13216558e-01 1.06241119e+00 -6.51184082e-01
4.06547308e-01 -4.00947124e-01 -3.73234659e-01 -3.82146090e-02
-8.98376882e-01 4.64102387e-01 9.19269025e-01 -5.90930820e-01
-6.26130879e-01 2.43428260e-01 -1.22643366e-01 -1.30362526e-01
-1.74769253e-01 4.95819658e-01 5.69463432e-01 -7.06512749e-01
1.05639386e+00 5.90402722e-01 2.68859148e-01 -3.41677487e-01
-3.26654941e-01 3.38089168e-01 5.87456584e-01 -2.21413180e-01
7.15598285e-01 6.21013284e-01 8.87519047e-02 -1.02370167e+00
-6.27576292e-01 -2.53202289e-01 -8.37115228e-01 -1.26186728e-01
5.10509014e-01 -5.23951530e-01 -5.85244715e-01 7.42495835e-01
-9.59244311e-01 -4.78614450e-01 -6.81266129e-01 4.00052816e-01
-3.82117063e-01 -2.18871921e-01 -6.11589193e-01 -1.08303046e+00
-4.56615865e-01 -6.98581755e-01 4.60920811e-01 5.69297016e-01
-3.89264166e-01 -1.04985940e+00 1.60292566e-01 -2.52635181e-01
1.09090436e+00 3.20533216e-01 7.70700753e-01 -7.89790809e-01
-2.54202306e-01 -3.90070528e-01 -1.52725771e-01 1.35264978e-01
-6.34156261e-03 -3.03206202e-02 -1.15652430e+00 -2.86937773e-01
1.48378924e-01 -6.31939411e-01 9.51264441e-01 4.23366189e-01
8.43174875e-01 -1.32250085e-01 -3.67378801e-01 5.48878372e-01
1.58663034e+00 3.74096513e-01 7.03723550e-01 4.81159091e-01
2.96442598e-01 2.98676759e-01 -3.87552679e-01 1.56182289e-01
-1.58093065e-01 9.20044780e-02 4.68898118e-01 -2.25468576e-01
-4.41805199e-02 2.86931451e-03 2.21758932e-01 1.18855977e+00
-3.51370513e-01 1.38189852e-01 -1.27188110e+00 5.50717771e-01
-1.88796771e+00 -9.48218942e-01 1.04311377e-01 2.18906522e+00
5.84073067e-01 3.87128681e-01 -4.08459455e-02 3.83348763e-01
4.13670957e-01 -3.23410273e-01 -1.24486697e+00 -4.98163283e-01
-6.26289129e-01 5.93323469e-01 4.40102756e-01 -7.50177503e-02
-7.24044502e-01 8.43732953e-01 7.41162634e+00 2.72691756e-01
-1.60703659e+00 -2.09888011e-01 5.32433510e-01 -6.80295676e-02
-9.47729498e-02 3.09723220e-03 -8.34864378e-01 4.08220202e-01
1.40080237e+00 8.61974955e-02 7.54544079e-01 4.71962661e-01
7.89870396e-02 -1.84477061e-01 -9.42335725e-01 1.08268833e+00
-1.70070440e-01 -1.75543118e+00 -3.41123454e-02 -3.09584420e-02
9.15265918e-01 5.58401942e-01 2.00943872e-01 1.06515460e-01
7.98238441e-02 -1.39934170e+00 4.13196743e-01 5.34277737e-01
6.55331492e-01 -3.12094361e-01 8.00152540e-01 5.97090542e-01
-9.38107669e-01 -4.06706810e-01 -6.58110559e-01 -6.87571108e-01
-2.68125981e-01 6.09863877e-01 -5.04326582e-01 -1.82144254e-01
7.83521414e-01 5.19120157e-01 -6.19174778e-01 1.34425485e+00
4.26269978e-01 6.59644723e-01 -8.72288167e-01 -5.56151211e-01
5.13351560e-01 -1.14917777e-01 3.47265810e-01 9.34172511e-01
3.42839777e-01 1.48035854e-01 -3.54000062e-01 1.00784481e+00
3.23461629e-02 -1.31696418e-01 -8.15034568e-01 -4.95530993e-01
4.61644202e-01 1.03814411e+00 -1.14535117e+00 -1.13570362e-01
-2.66618252e-01 6.69757247e-01 3.58120054e-01 3.97956014e-01
-1.57130122e-01 -3.76117975e-01 1.24180503e-01 1.73964992e-01
2.63083905e-01 -3.60515833e-01 -7.14830518e-01 -1.07572222e+00
-5.83791398e-02 -6.02541566e-01 -1.29471153e-01 -6.45034254e-01
-1.23502076e+00 6.86899781e-01 -6.18316770e-01 -1.11280346e+00
-5.57867885e-01 -9.47952271e-01 -7.34258711e-01 6.23512447e-01
-1.22232509e+00 -8.66507173e-01 -1.46387652e-01 4.47972745e-01
2.84485161e-01 -3.77764761e-01 1.42355680e+00 -1.30216749e-02
-4.71943796e-01 2.91239828e-01 6.30459964e-01 2.64468212e-02
1.97070569e-01 -9.08164740e-01 4.81193066e-01 6.40630662e-01
3.32014650e-01 6.41851306e-01 7.56560564e-01 -3.28200817e-01
-1.57392216e+00 -5.97458661e-01 8.15845490e-01 -3.06174248e-01
7.37865627e-01 -4.36270714e-01 -9.89340246e-01 4.28047150e-01
-3.06150429e-02 3.77333760e-01 5.49375355e-01 -1.09809279e-01
-1.92776710e-01 -1.90132976e-01 -1.11304247e+00 4.77562994e-01
9.15477514e-01 -3.32893223e-01 -4.60362285e-01 4.54890579e-02
1.38533756e-01 -1.45413399e-01 -5.14250219e-01 5.20076573e-01
8.10237467e-01 -1.01319981e+00 6.42464638e-01 -6.42184556e-01
2.60534286e-01 -1.72998399e-01 2.72009611e-01 -1.03083658e+00
-2.56805420e-01 -5.80474675e-01 -3.28876436e-01 6.72308207e-01
7.23903298e-01 -1.04021513e+00 1.01931334e+00 8.25228870e-01
8.75760987e-02 -1.00319946e+00 -1.04814851e+00 -9.13114369e-01
1.82532609e-01 -5.20056933e-02 6.94516301e-02 7.84329951e-01
2.21365854e-01 1.97936013e-01 4.70760465e-02 -3.06917340e-01
5.77373862e-01 4.49071042e-02 1.00721300e-01 -1.69378054e+00
2.12321114e-02 -6.75076663e-01 -9.05443430e-01 -6.85427606e-01
-1.43056363e-01 -7.30728447e-01 5.53789325e-02 -1.45236230e+00
-3.57766934e-02 -4.66802537e-01 -4.77866918e-01 7.48112023e-01
4.93694544e-01 6.80839658e-01 -1.55436605e-01 6.46041155e-01
-4.38433707e-01 7.89284483e-02 7.67918825e-01 1.31543344e-02
-3.52507591e-01 -2.19294056e-01 -5.56968808e-01 8.33193481e-01
1.12324834e+00 -6.04573190e-01 -2.46863440e-01 -4.35533851e-01
4.24823821e-01 -4.11080509e-01 4.48070347e-01 -1.67431855e+00
7.61307418e-01 1.66152015e-01 8.47152114e-01 -1.47612810e-01
2.84269452e-01 -7.52663136e-01 1.31358504e-01 8.31574619e-01
-2.01140568e-01 1.03145950e-01 5.48877478e-01 3.08628470e-01
-2.44614109e-01 -3.58441800e-01 8.83943379e-01 -3.14692408e-01
-6.89684272e-01 1.55995190e-01 -9.03346062e-01 6.25690853e-04
8.69261742e-01 -7.44091272e-01 -3.29195708e-01 -3.55215147e-02
-7.20527411e-01 -1.77924588e-01 4.04844522e-01 -6.88294396e-02
5.87229848e-01 -9.72297847e-01 -3.78581017e-01 5.60401618e-01
-2.70732850e-01 7.66909346e-02 -7.47261420e-02 7.80408025e-01
-7.25347042e-01 7.06363678e-01 -7.18839467e-01 -5.91404557e-01
-4.71055120e-01 2.98635066e-01 4.38006222e-01 -3.28231826e-02
-7.81692684e-01 8.85880291e-01 -3.01200211e-01 -4.29217458e-01
8.77555758e-02 1.33441426e-02 -1.10303476e-01 -2.30219737e-01
3.79969984e-01 1.17448099e-01 3.34447116e-01 -2.24165708e-01
-4.30151910e-01 4.46971804e-01 7.56616443e-02 -1.97193071e-01
1.74599230e+00 2.28484213e-01 -2.46574104e-01 9.51810300e-01
8.44243884e-01 -5.94606400e-01 -1.20779586e+00 -4.35649790e-02
2.30773866e-01 3.44998181e-01 2.93322280e-02 -8.65172446e-01
-9.36921895e-01 1.20255029e+00 7.39770412e-01 4.04385090e-01
9.91862357e-01 -2.77772576e-01 8.64717543e-01 9.72160578e-01
4.02361810e-01 -1.19883597e+00 2.91605238e-02 7.51359105e-01
4.69322652e-01 -1.08302212e+00 -2.50070304e-01 2.53197253e-01
-2.47575670e-01 1.47122669e+00 7.33545542e-01 -5.39668977e-01
5.77947855e-01 7.91763246e-01 6.93202540e-02 -2.41804868e-01
-8.04630339e-01 3.35763134e-02 -1.83117211e-01 6.31477058e-01
4.17886108e-01 -3.45957041e-01 -8.46319348e-02 4.20853019e-01
-1.07326962e-01 2.33054549e-01 2.08915591e-01 1.23114479e+00
-7.46041954e-01 -8.30955923e-01 -1.20281968e-02 8.63826752e-01
-3.57202470e-01 -3.37101519e-01 -3.82101476e-01 5.80040038e-01
-7.42423087e-02 5.28046370e-01 2.60161459e-01 -3.86869848e-01
-5.64773893e-03 4.51033205e-01 5.36603928e-01 -3.35473567e-01
-6.64556026e-01 -3.31826687e-01 -3.94729525e-01 -3.28180820e-01
-4.31052655e-01 -4.77891415e-01 -1.64026523e+00 -2.44146928e-01
-2.76280701e-01 -1.64487641e-02 8.90714228e-01 1.29743552e+00
5.74054837e-01 3.07486653e-01 3.59907717e-01 -1.12398672e+00
-4.11742806e-01 -7.18679130e-01 -4.79145914e-01 1.42317768e-02
1.73414066e-01 -5.86821735e-01 -5.37272811e-01 2.06385612e-01] | [8.109894752502441, 2.6797778606414795] |
a3624c1d-8c73-4ead-b56b-02d8a410affa | weakly-supervised-temporal-action-3 | 2103.16155 | null | https://arxiv.org/abs/2103.16155v1 | https://arxiv.org/pdf/2103.16155v1.pdf | Weakly Supervised Temporal Action Localization Through Learning Explicit Subspaces for Action and Context | Weakly-supervised Temporal Action Localization (WS-TAL) methods learn to localize temporal starts and ends of action instances in a video under only video-level supervision. Existing WS-TAL methods rely on deep features learned for action recognition. However, due to the mismatch between classification and localization, these features cannot distinguish the frequently co-occurring contextual background, i.e., the context, and the actual action instances. We term this challenge action-context confusion, and it will adversely affect the action localization accuracy. To address this challenge, we introduce a framework that learns two feature subspaces respectively for actions and their context. By explicitly accounting for action visual elements, the action instances can be localized more precisely without the distraction from the context. To facilitate the learning of these two feature subspaces with only video-level categorical labels, we leverage the predictions from both spatial and temporal streams for snippets grouping. In addition, an unsupervised learning task is introduced to make the proposed module focus on mining temporal information. The proposed approach outperforms state-of-the-art WS-TAL methods on three benchmarks, i.e., THUMOS14, ActivityNet v1.2 and v1.3 datasets. | ['Gang Hua', 'Nanning Zheng', 'Junsong Yuan', 'Wei Tang', 'Le Wang', 'Ziyi Liu'] | 2021-03-30 | null | null | null | null | ['weakly-supervised-temporal-action'] | ['computer-vision'] | [ 3.52350116e-01 -3.50506604e-01 -7.18845069e-01 -2.89013773e-01
-6.83215141e-01 -3.35499763e-01 6.75476730e-01 7.98392147e-02
-3.66870642e-01 5.38887084e-01 5.39401650e-01 3.03482622e-01
-7.47158192e-04 -2.88248986e-01 -5.96404970e-01 -9.01904941e-01
-3.21195304e-01 -1.81237057e-01 5.74211299e-01 3.90100241e-01
2.86361754e-01 1.33081570e-01 -1.58425367e+00 8.98673952e-01
5.61824560e-01 1.38407350e+00 2.20526308e-01 3.53933007e-01
-2.09796831e-01 1.55652833e+00 -4.67978507e-01 3.03994328e-01
2.05878153e-01 -7.06521392e-01 -6.62734330e-01 4.16528970e-01
5.25984287e-01 -3.65476429e-01 -5.07081449e-01 6.32241249e-01
7.08218887e-02 3.10857713e-01 4.93091166e-01 -1.57432067e+00
-1.67894974e-01 3.88236225e-01 -6.82340622e-01 5.13078928e-01
5.21441042e-01 4.18761164e-01 1.02180123e+00 -9.14074659e-01
6.47564292e-01 1.09182668e+00 5.73883057e-01 3.89723718e-01
-8.97392988e-01 -4.83607948e-01 6.59450173e-01 7.29130149e-01
-1.27626121e+00 -5.58971465e-01 8.25164080e-01 -7.10256100e-01
7.64474869e-01 3.66526172e-02 6.56954527e-01 1.38946915e+00
1.41638378e-02 1.29665101e+00 1.01255858e+00 5.68011962e-03
4.36303407e-01 -3.17964017e-01 -3.67068835e-02 5.71188569e-01
-3.09599847e-01 -8.39232504e-02 -9.11138296e-01 5.08907139e-02
5.78085184e-01 5.29297531e-01 -1.02202252e-01 -4.15677786e-01
-1.51502562e+00 3.54330182e-01 1.84271514e-01 3.46856833e-01
-5.04556715e-01 2.38693058e-01 7.51190960e-01 -1.34790838e-01
2.86848783e-01 -7.78756291e-02 -5.46084106e-01 -5.57802737e-01
-8.82227659e-01 -6.32982934e-03 2.96461314e-01 8.88576090e-01
7.58204043e-01 -6.57792911e-02 -6.26681805e-01 5.44424891e-01
1.17595710e-01 2.06131592e-01 6.73818946e-01 -7.95161247e-01
7.02761948e-01 8.93684566e-01 1.76750839e-01 -8.95445943e-01
-2.84034073e-01 -2.14621663e-01 -5.60368538e-01 -1.29971147e-01
3.89635444e-01 7.37690479e-02 -1.04183185e+00 1.58567500e+00
4.74650681e-01 9.17914748e-01 1.85805298e-02 9.99919653e-01
7.25455225e-01 5.43222070e-01 3.88765424e-01 -3.49870443e-01
1.09216583e+00 -1.17869735e+00 -8.75504613e-01 -3.28855306e-01
8.56090546e-01 -3.88021410e-01 1.10450852e+00 1.68860927e-01
-6.74406230e-01 -6.33023739e-01 -8.57885540e-01 1.78719774e-01
-1.70271412e-01 4.47014272e-01 5.43230116e-01 -1.83448493e-02
-3.98892701e-01 6.20363712e-01 -1.24118459e+00 -4.16006982e-01
7.67708361e-01 -4.52910513e-02 -5.47834635e-01 -4.42555211e-02
-9.46303606e-01 4.66296375e-01 3.91244203e-01 -9.81942937e-03
-1.22924733e+00 -4.89103496e-01 -8.31641614e-01 -1.25580847e-01
8.33981216e-01 -9.03401524e-03 1.16642642e+00 -1.30604732e+00
-1.13873923e+00 5.66833913e-01 -4.37150300e-01 -5.78525126e-01
5.56314468e-01 -3.27198178e-01 -4.83703107e-01 3.78388107e-01
4.00872678e-01 3.36774290e-01 8.44247401e-01 -9.96115744e-01
-1.15294528e+00 -3.81273121e-01 2.01832533e-01 2.38923088e-01
-3.24711293e-01 -7.09486604e-02 -5.86243868e-01 -6.64177120e-01
1.86754286e-01 -7.59260178e-01 5.57442708e-03 1.21173203e-01
-3.67567956e-01 -3.69689703e-01 1.06393254e+00 -5.13705432e-01
1.41694546e+00 -2.52699327e+00 5.03633134e-02 -1.75256848e-01
3.83329503e-02 1.18072294e-01 -9.41485912e-02 4.52054381e-01
-3.16111557e-02 -1.37251258e-01 1.62925962e-02 -4.37541455e-01
-1.57442138e-01 3.34985435e-01 -2.18340516e-01 6.18606567e-01
2.15984598e-01 6.60629809e-01 -1.15395355e+00 -7.12680519e-01
3.38596076e-01 9.17242095e-02 -3.81464034e-01 3.32818747e-01
-2.99243003e-01 8.04394960e-01 -6.32405400e-01 8.72803509e-01
1.77087888e-01 -9.43121091e-02 3.13850753e-02 -3.49382281e-01
-1.62723675e-01 2.91880906e-01 -1.23771214e+00 1.94447231e+00
-2.94563323e-01 4.74401772e-01 -3.21169704e-01 -1.22501874e+00
5.02304792e-01 4.62614715e-01 1.09079468e+00 -7.60126591e-01
-1.64057195e-01 -2.44776607e-02 -1.37926430e-01 -8.62658978e-01
-2.57214177e-02 1.28712922e-01 -1.44830495e-02 3.44126105e-01
2.89057106e-01 9.03235435e-01 3.41465294e-01 2.22503185e-01
1.44213486e+00 5.78843355e-01 3.44043314e-01 1.15727596e-01
5.80321193e-01 -7.22994655e-02 1.20637107e+00 6.47341907e-01
-6.81880713e-01 4.44752276e-01 7.15139389e-01 -6.06216729e-01
-5.07283092e-01 -9.55050230e-01 3.21555942e-01 1.31324625e+00
3.97888929e-01 -7.91554809e-01 -4.91579980e-01 -1.22939801e+00
-1.32373095e-01 4.49578375e-01 -8.74338150e-01 -2.16299400e-01
-5.95366180e-01 -2.54848927e-01 3.25700492e-01 8.01261127e-01
5.88100016e-01 -1.14532042e+00 -8.15078914e-01 1.66276485e-01
-4.00289714e-01 -1.51805735e+00 -6.48997188e-01 2.56382465e-01
-7.35212028e-01 -1.42341650e+00 -2.72476375e-01 -4.81118023e-01
4.44939584e-01 3.83777142e-01 7.72370279e-01 -1.44763961e-01
-2.77999848e-01 5.04903555e-01 -7.07212210e-01 -7.62616694e-02
6.54012561e-02 -2.42934912e-01 1.17568530e-01 5.87499440e-01
6.63253307e-01 -6.71314001e-01 -9.04685557e-01 6.66679859e-01
-7.40414262e-01 1.98646426e-01 6.08698189e-01 6.23226643e-01
8.07794869e-01 4.51987870e-02 3.96986961e-01 -4.83590037e-01
-1.05429016e-01 -6.76171601e-01 -9.33468640e-02 2.09194809e-01
-1.16604112e-01 -1.29941911e-01 5.86176217e-01 -6.28541529e-01
-9.19336498e-01 4.97803092e-01 4.17343408e-01 -8.44220340e-01
-4.99276131e-01 4.39057142e-01 -4.07049179e-01 4.22769368e-01
4.24780458e-01 5.13874412e-01 -4.77299988e-01 -3.97950888e-01
9.08812359e-02 3.69941115e-01 4.61771488e-01 -3.63959163e-01
5.58510005e-01 6.94146812e-01 -1.33653000e-01 -6.42054260e-01
-1.13509500e+00 -8.29047322e-01 -8.49344790e-01 -4.96623963e-01
1.17598403e+00 -1.03758204e+00 -3.13503116e-01 4.36312646e-01
-9.72351849e-01 -3.46422255e-01 -3.31531614e-01 6.36201024e-01
-7.33540475e-01 3.23753923e-01 -3.08933526e-01 -8.39447498e-01
3.05064201e-01 -1.00054729e+00 1.16303790e+00 -1.64532959e-02
-1.71536267e-01 -6.93540454e-01 4.04001446e-03 3.52265149e-01
-1.29690602e-01 4.52735811e-01 4.96154040e-01 -7.02787936e-01
-6.94699109e-01 -2.03679562e-01 -2.87191179e-02 2.58399427e-01
5.35928786e-01 -1.23064511e-01 -9.34089303e-01 -1.55387493e-02
-8.88260901e-02 -4.40834403e-01 9.76799369e-01 2.91040331e-01
1.37110102e+00 -4.11879569e-01 -4.00466174e-01 4.77717131e-01
1.05782211e+00 3.72945338e-01 4.44533646e-01 1.83508694e-01
9.39795554e-01 5.12697101e-01 1.18515420e+00 7.51140594e-01
2.61069536e-01 8.79592955e-01 5.59331357e-01 1.35167748e-01
-6.12441078e-02 -5.32503963e-01 6.58529520e-01 2.85366565e-01
-2.02177539e-01 -6.04557358e-02 -8.10511231e-01 5.47719479e-01
-2.35120273e+00 -1.26618874e+00 9.05005336e-02 2.04748178e+00
6.62742257e-01 2.14480683e-01 3.66037399e-01 1.65723443e-01
7.05294907e-01 6.18242204e-01 -7.89390981e-01 3.72500092e-01
1.03041336e-01 -3.48376334e-01 3.97881031e-01 1.06351087e-02
-1.62036455e+00 8.77943277e-01 4.97672415e+00 8.86879146e-01
-9.88098919e-01 2.52109021e-01 4.43996936e-01 -4.79054332e-01
3.02408159e-01 2.95572430e-02 -6.45609021e-01 7.42809415e-01
6.14756465e-01 1.95203424e-01 1.08504593e-01 9.16151047e-01
6.65803671e-01 -1.90585032e-01 -1.47070456e+00 1.20990837e+00
2.24682212e-01 -1.08465087e+00 -1.00529693e-01 -1.74029127e-01
5.93514144e-01 -1.66066557e-01 -1.70865923e-01 5.78295052e-01
-1.60864741e-01 -8.60688865e-01 8.73684227e-01 6.36762977e-01
5.60012817e-01 -4.90738243e-01 5.11381805e-01 4.50258553e-01
-1.63096762e+00 -5.43224335e-01 4.47420031e-02 -1.80093482e-01
1.80561274e-01 3.14472169e-01 -3.57838750e-01 4.51321900e-01
7.67679572e-01 1.51192343e+00 -4.75410819e-01 9.54712272e-01
-2.86176026e-01 6.50467217e-01 -1.89392585e-02 3.27724606e-01
5.22309661e-01 8.54168981e-02 4.56990510e-01 1.09498203e+00
2.03111157e-01 1.51574925e-01 8.33701372e-01 4.47588682e-01
2.09031120e-01 -5.54404147e-02 -4.93861735e-01 -2.74278820e-01
3.22517931e-01 1.03674531e+00 -6.62868917e-01 -3.17745537e-01
-6.45449936e-01 1.11979997e+00 1.99606806e-01 4.39467937e-01
-1.18293607e+00 6.40341267e-02 7.97765255e-01 2.19903290e-01
3.89528155e-01 -2.39788070e-01 3.81451361e-02 -1.40784919e+00
2.93860167e-01 -7.51522601e-01 4.82675433e-01 -5.25598168e-01
-1.05766273e+00 1.47091255e-01 5.20257838e-02 -1.85691655e+00
-9.41213891e-02 -3.51831943e-01 -7.85748124e-01 3.08240861e-01
-1.23923254e+00 -1.18044770e+00 -5.61280251e-01 8.44237387e-01
9.38246608e-01 -1.58365443e-01 4.50664759e-01 4.00552720e-01
-7.89970934e-01 4.74778801e-01 -2.43993193e-01 3.39603096e-01
8.57401907e-01 -1.11491203e+00 -1.19941331e-01 8.89276028e-01
3.08918059e-01 1.90607145e-01 4.75813419e-01 -6.67288065e-01
-1.45174932e+00 -1.38219917e+00 6.00980222e-01 -4.42157418e-01
8.13440263e-01 -4.05274808e-01 -7.35924780e-01 7.71297753e-01
-1.72589153e-01 5.11448979e-01 6.86862886e-01 -1.68596189e-02
-3.51737708e-01 -2.36543864e-01 -7.46639550e-01 5.61319709e-01
1.50540173e+00 -6.92377031e-01 -5.32956898e-01 4.20285344e-01
5.25083959e-01 -1.67649463e-01 -6.49146676e-01 3.51377875e-01
5.82657754e-01 -1.06779599e+00 8.09170306e-01 -7.89618075e-01
6.64894283e-01 -4.99154866e-01 -3.08178037e-01 -9.03507113e-01
-2.31245175e-01 -3.35340410e-01 -5.53206205e-01 1.18921804e+00
-2.32173931e-02 -1.01168938e-01 8.59165430e-01 2.30516449e-01
-2.51006722e-01 -8.50393891e-01 -1.07744610e+00 -9.07624960e-01
-6.12971127e-01 -5.67062318e-01 3.35706919e-01 9.54315305e-01
7.79811963e-02 2.60961771e-01 -7.05347776e-01 9.55412388e-02
4.10433918e-01 1.02076843e-01 7.66247153e-01 -9.11002159e-01
-2.88844526e-01 -2.04066724e-01 -8.46702039e-01 -1.19265485e+00
2.19088510e-01 -5.55209696e-01 2.19613239e-01 -1.32588100e+00
3.14701825e-01 -1.97754085e-01 -8.05141389e-01 7.76121438e-01
-1.50514960e-01 2.75350804e-03 1.10986091e-01 3.79722506e-01
-1.30509973e+00 7.75785863e-01 1.01656568e+00 -1.73117727e-01
-2.10405543e-01 -4.47889566e-02 -1.76169410e-01 8.96977484e-01
7.35481381e-01 -4.28926706e-01 -7.20086098e-01 -3.04382086e-01
-2.11456463e-01 1.81821629e-01 5.64116597e-01 -1.32412183e+00
2.50358641e-01 -6.38733625e-01 4.16077614e-01 -8.09978426e-01
3.71326655e-01 -9.54846621e-01 -2.29822442e-01 2.03923196e-01
-6.19204223e-01 -3.35749179e-01 -7.12193325e-02 1.04024065e+00
-4.41291958e-01 1.14952467e-01 6.41174197e-01 -1.75382167e-01
-1.16437817e+00 6.86157763e-01 -4.48844850e-01 1.09416798e-01
1.47917688e+00 -4.13905710e-01 -2.37360835e-01 -2.69601822e-01
-9.11733866e-01 3.34169298e-01 2.78148502e-01 5.66076934e-01
5.61820269e-01 -1.54622173e+00 -4.13885087e-01 7.76407658e-04
5.83054900e-01 -1.40047669e-01 5.69698274e-01 1.32876527e+00
5.32998852e-02 2.71021217e-01 -1.64275244e-01 -8.20613086e-01
-1.33875453e+00 6.21605039e-01 2.81265199e-01 -2.28808686e-01
-6.81172729e-01 5.64760327e-01 5.03573716e-01 1.56520769e-01
5.83576083e-01 -4.24269825e-01 -4.09486562e-01 1.58553988e-01
7.31264651e-01 3.40892971e-01 -3.14178020e-01 -7.41002023e-01
-6.14020526e-01 3.13880384e-01 2.69379038e-02 1.21514574e-01
1.07078016e+00 -1.06175490e-01 2.51343131e-01 8.88623774e-01
1.16996956e+00 -2.80763984e-01 -1.75957286e+00 -4.17575300e-01
2.53420949e-01 -6.23857677e-01 -1.11517876e-01 -6.38632059e-01
-1.08693361e+00 9.56189454e-01 6.75562084e-01 -1.60676271e-01
1.20463538e+00 4.89726178e-02 7.25482941e-01 2.37220451e-01
3.38211000e-01 -1.27760601e+00 5.67623556e-01 4.10496205e-01
6.84773624e-01 -1.36981809e+00 -5.08652739e-02 -2.77228206e-01
-9.24324870e-01 8.00709844e-01 9.67903197e-01 -4.61555906e-02
5.23002148e-01 1.52681367e-02 -1.16941571e-01 -4.82636876e-02
-8.89332056e-01 -3.42995614e-01 2.71625042e-01 5.21798491e-01
2.40803048e-01 -6.86543658e-02 -2.52751917e-01 7.38085270e-01
6.63651288e-01 1.56011522e-01 1.51751861e-01 1.29186702e+00
-3.86770934e-01 -7.84872115e-01 -3.01537570e-02 4.92028892e-01
-4.57098424e-01 1.80924073e-01 -3.84851247e-01 5.59625208e-01
5.40238976e-01 9.18075740e-01 9.63223130e-02 -7.33952284e-01
3.73956501e-01 8.30327645e-02 1.31253332e-01 -6.39920056e-01
-2.55485028e-01 1.74497411e-01 1.19157583e-01 -1.28266823e+00
-8.00861418e-01 -8.65713656e-01 -1.29862463e+00 2.56964236e-01
7.24598765e-02 -3.92441787e-02 1.28821194e-01 1.21334910e+00
3.92250329e-01 5.77284575e-01 6.99472368e-01 -8.54417503e-01
-4.04063225e-01 -9.59267199e-01 -6.94185257e-01 6.81327999e-01
4.21723843e-01 -9.54360783e-01 -4.43051726e-01 3.06952387e-01] | [8.420928955078125, 0.6522144079208374] |
8bfa4017-5a16-4037-bdae-e3702eafe822 | hierarchical-adaptive-pooling-by-capturing | 2104.05960 | null | https://arxiv.org/abs/2104.05960v1 | https://arxiv.org/pdf/2104.05960v1.pdf | Hierarchical Adaptive Pooling by Capturing High-order Dependency for Graph Representation Learning | Graph neural networks (GNN) have been proven to be mature enough for handling graph-structured data on node-level graph representation learning tasks. However, the graph pooling technique for learning expressive graph-level representation is critical yet still challenging. Existing pooling methods either struggle to capture the local substructure or fail to effectively utilize high-order dependency, thus diminishing the expression capability. In this paper we propose HAP, a hierarchical graph-level representation learning framework, which is adaptively sensitive to graph structures, i.e., HAP clusters local substructures incorporating with high-order dependencies. HAP utilizes a novel cross-level attention mechanism MOA to naturally focus more on close neighborhood while effectively capture higher-order dependency that may contain crucial information. It also learns a global graph content GCont that extracts the graph pattern properties to make the pre- and post-coarsening graph content maintain stable, thus providing global guidance in graph coarsening. This novel innovation also facilitates generalization across graphs with the same form of features. Extensive experiments on fourteen datasets show that HAP significantly outperforms twelve popular graph pooling methods on graph classification task with an maximum accuracy improvement of 22.79%, and exceeds the performance of state-of-the-art graph matching and graph similarity learning algorithms by over 3.5% and 16.7%. | ['Hongzuo Xu', 'Zhiquan Lai', 'Yiming Zhang', 'Dongsheng Li', 'Songlei Jian', 'Ning Liu'] | 2021-04-13 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [ 2.05156058e-02 1.96908221e-01 -5.38326085e-01 -2.01378003e-01
-3.50641876e-01 -3.80246550e-01 4.31775481e-01 7.27104902e-01
-9.19695273e-02 2.93142378e-01 1.81432471e-01 -3.23075280e-02
-3.11528832e-01 -1.38272965e+00 -7.20573246e-01 -7.20523894e-01
-6.20594680e-01 1.82053611e-01 3.95675153e-01 -2.37566426e-01
3.61332074e-02 5.24141133e-01 -1.28650498e+00 3.00835550e-01
9.23799634e-01 9.05576766e-01 2.40845352e-01 4.41130757e-01
-3.25682461e-01 6.73979104e-01 -3.71607363e-01 -3.34905416e-01
1.77997962e-01 -2.26861835e-01 -6.51000619e-01 -1.08901821e-01
7.69016504e-01 4.59321849e-02 -7.93114901e-01 1.33942175e+00
4.60124880e-01 3.67006570e-01 3.56413424e-01 -1.19310737e+00
-1.28354990e+00 8.84252131e-01 -6.71229541e-01 4.15903866e-01
3.20095301e-01 4.57454696e-02 1.58026397e+00 -6.09601855e-01
6.37224853e-01 1.33265257e+00 7.56560028e-01 3.55660349e-01
-1.03703797e+00 -5.78118563e-01 5.61114430e-01 1.47259340e-01
-1.46038759e+00 1.92213595e-01 1.04302359e+00 -2.14434251e-01
1.40332925e+00 8.54167119e-02 6.98782325e-01 7.04412043e-01
5.45986891e-01 5.57285607e-01 5.51165402e-01 -1.41701296e-01
-2.16148868e-01 -4.47874397e-01 6.07395947e-01 1.27227068e+00
6.45386398e-01 -1.35642365e-01 -4.40081149e-01 5.12459502e-02
8.77209008e-01 2.21781969e-01 -4.20753121e-01 -2.82037020e-01
-9.67493832e-01 8.60797107e-01 1.52146065e+00 6.72188342e-01
-3.90033036e-01 2.96331972e-01 4.20684308e-01 4.27499533e-01
4.22098666e-01 4.73836005e-01 -2.30248179e-02 5.39501131e-01
-2.60528743e-01 -2.66777491e-03 4.92629766e-01 1.13732982e+00
1.20645440e+00 1.38695762e-01 -5.01945555e-01 8.40750337e-01
2.60602623e-01 -7.38908648e-02 5.94644904e-01 -7.23970458e-02
6.83260798e-01 1.56728435e+00 -8.21926117e-01 -1.56609178e+00
-5.71998417e-01 -7.35350788e-01 -1.32106125e+00 -2.60385424e-01
5.13362959e-02 2.72545069e-01 -1.16136444e+00 1.77981365e+00
5.39672337e-02 3.08072478e-01 -2.52709240e-01 5.43505132e-01
1.45196688e+00 6.18871748e-01 2.28140265e-01 1.52298689e-01
1.28064930e+00 -1.13281071e+00 -4.75969225e-01 -4.19326991e-01
7.32271552e-01 -1.05087839e-01 1.27416694e+00 -1.43899754e-01
-9.67727900e-01 -7.97924221e-01 -1.29863274e+00 -2.85223305e-01
-7.28586555e-01 -4.46851969e-01 8.44624341e-01 3.97020042e-01
-1.26177120e+00 9.09017205e-01 -5.51142275e-01 -5.75340092e-01
5.99376261e-01 4.05300200e-01 -6.77304149e-01 -4.07578617e-01
-1.30086565e+00 4.78045940e-01 7.37092674e-01 -3.81344068e-03
-3.25632602e-01 -9.63876367e-01 -1.22709119e+00 5.41478753e-01
3.02187175e-01 -5.61420321e-01 4.59765732e-01 -7.13088095e-01
-1.01918638e+00 8.57340753e-01 9.89326239e-02 -3.00244540e-01
-1.07799545e-01 2.20763668e-01 -4.61853087e-01 2.72855937e-01
6.21315651e-02 6.20881557e-01 6.89772725e-01 -7.74230480e-01
-2.56786942e-01 -5.75829208e-01 1.32768750e-01 3.07234734e-01
-6.89932883e-01 -2.28291452e-01 -8.11247110e-01 -8.09726000e-01
3.15117002e-01 -5.70064485e-01 -1.93616360e-01 -1.67068347e-01
-2.48638064e-01 -6.19872093e-01 7.20159411e-01 -4.58787352e-01
1.57699060e+00 -1.94572043e+00 1.73435405e-01 3.90434355e-01
7.73473203e-01 3.21871132e-01 -4.94716108e-01 6.79875016e-01
-1.58275411e-01 3.71673554e-01 -1.71709597e-01 -4.84802164e-02
7.44505413e-03 1.95915252e-01 2.77416091e-02 3.77498180e-01
5.03703296e-01 1.48514211e+00 -9.97746766e-01 -4.01949763e-01
4.32040542e-02 3.70202214e-01 -6.43988907e-01 1.89053431e-01
6.64473176e-02 -1.54395998e-01 -5.35623729e-01 6.89934075e-01
6.28351748e-01 -7.09098399e-01 3.47571075e-01 -2.70274520e-01
4.77702677e-01 1.28663257e-01 -9.81956065e-01 1.75894213e+00
-2.13450581e-01 3.75007808e-01 -7.67193660e-02 -1.13001788e+00
1.17116964e+00 -1.70473740e-01 2.75832474e-01 -7.36563802e-01
1.35305360e-01 -1.35251895e-01 2.18279436e-01 -1.81572452e-01
3.07218522e-01 2.70374119e-01 -2.95855731e-01 1.05159484e-01
3.82382125e-01 2.51131445e-01 2.57105768e-01 4.51831013e-01
1.48622799e+00 -4.26868170e-01 5.67736626e-01 -6.50731385e-01
6.11873329e-01 -5.66767275e-01 5.35071254e-01 7.79901087e-01
-2.45384932e-01 5.38568974e-01 6.37934268e-01 -5.92544019e-01
-5.49281836e-01 -1.03101182e+00 2.55887330e-01 1.26677740e+00
4.27656859e-01 -8.98940563e-01 -6.08237386e-01 -7.59640515e-01
2.39718720e-01 6.17165230e-02 -8.98991525e-01 -7.56331086e-01
-7.40869820e-01 -7.02454686e-01 3.61718565e-01 7.53938735e-01
6.93817794e-01 -1.30196702e+00 1.64051622e-01 3.00828397e-01
1.62979439e-01 -1.13453889e+00 -7.75839329e-01 1.53445657e-02
-7.73169935e-01 -1.15276754e+00 -2.87770659e-01 -1.25774097e+00
9.39598382e-01 5.37207603e-01 1.39559710e+00 8.31258297e-01
-3.55722010e-01 7.42042586e-02 -3.91606450e-01 7.68807232e-02
-6.97716558e-03 4.78436261e-01 -3.15797597e-01 -8.30335263e-03
3.68676901e-01 -8.46777320e-01 -3.76097709e-01 6.98813200e-02
-7.40980685e-01 -1.52138397e-01 7.18469024e-01 9.41104412e-01
6.22251809e-01 1.77785203e-01 6.32018805e-01 -9.99974310e-01
1.09329176e+00 -4.45393413e-01 -3.60191137e-01 3.97179455e-01
-6.64985001e-01 1.49779871e-01 9.78507817e-01 -2.79252410e-01
-4.83581364e-01 -3.35097462e-01 9.30336192e-02 -4.95614767e-01
1.30024448e-01 8.21700811e-01 -5.04656792e-01 -4.55355376e-01
4.98963773e-01 2.21062943e-01 -1.28055617e-01 -2.28089929e-01
3.44305247e-01 1.10021159e-01 5.26787341e-01 -5.58054745e-01
8.55144501e-01 1.29635513e-01 1.90826148e-01 -6.86212003e-01
-5.50660372e-01 -5.89749515e-01 -6.17303729e-01 -5.91925345e-02
7.79545963e-01 -7.07994819e-01 -5.84201813e-01 4.86204505e-01
-8.31210494e-01 -3.17125827e-01 5.62493987e-02 6.88161254e-02
-1.13064900e-01 6.61296666e-01 -8.82716894e-01 -1.63719207e-01
-6.92082465e-01 -8.44052613e-01 9.10427809e-01 3.10535014e-01
1.72821999e-01 -1.14027858e+00 -1.67790547e-01 -1.12927116e-01
4.16162163e-01 5.22607267e-01 1.19467926e+00 -6.02637529e-01
-7.99966574e-01 -2.92514950e-01 -6.65140808e-01 -6.62154751e-03
2.21942902e-01 -5.27039729e-02 -5.31485856e-01 -6.86586499e-01
-7.37111032e-01 -2.00032830e-01 1.15824258e+00 1.50318071e-01
1.39061403e+00 -2.86254525e-01 -5.05568504e-01 7.51523674e-01
1.55233848e+00 -1.88666210e-01 5.44228017e-01 -4.55767028e-02
1.24375916e+00 4.34281319e-01 6.61062747e-02 -8.18759054e-02
3.96398664e-01 4.45689887e-01 5.63921571e-01 -4.35329750e-02
-4.33205366e-01 -6.38571262e-01 1.40829414e-01 9.97966886e-01
6.96676075e-02 -4.78435189e-01 -8.28401387e-01 4.59992796e-01
-1.98908293e+00 -7.39325047e-01 -6.05214499e-02 1.84797370e+00
4.38853502e-01 3.61002803e-01 -5.44738695e-02 -1.52761295e-01
1.01604152e+00 6.22638702e-01 -4.10273463e-01 -3.72852802e-01
-6.75651357e-02 4.21387196e-01 5.41540921e-01 4.80257809e-01
-1.10921299e+00 1.25965178e+00 5.04906511e+00 8.76835525e-01
-8.84597778e-01 -2.64955163e-01 2.68734634e-01 5.34830809e-01
-3.63197714e-01 -2.96157032e-01 -4.62489486e-01 3.30105126e-01
5.74562848e-01 -3.30810994e-01 5.16402364e-01 8.17995369e-01
-4.96037066e-01 4.80947763e-01 -1.01061559e+00 9.57564652e-01
2.10269034e-01 -1.42596042e+00 5.46461880e-01 -5.91571908e-03
8.03378463e-01 1.37304276e-01 -2.18854055e-01 6.86242402e-01
3.17771822e-01 -1.19901037e+00 5.18511608e-02 3.79910171e-01
6.30538642e-01 -7.79734790e-01 7.50744879e-01 1.65155694e-01
-2.08269405e+00 -1.28393754e-01 -5.94223857e-01 -1.01508200e-01
-1.99474379e-01 3.72932523e-01 -5.72956145e-01 1.01847327e+00
7.36271381e-01 1.15330660e+00 -9.50416505e-01 9.47124064e-01
-4.22535330e-01 3.51848453e-01 -9.63234231e-02 -3.03498566e-01
6.68478489e-01 -2.21816078e-01 4.26686227e-01 1.13777471e+00
1.18542336e-01 7.71807926e-03 4.91354704e-01 7.66908586e-01
-6.05719566e-01 2.72139251e-01 -7.53803909e-01 -2.68938452e-01
4.97038364e-01 1.45691991e+00 -9.00405347e-01 -1.65301129e-01
-4.14594084e-01 8.73980701e-01 1.12484026e+00 2.89510787e-01
-6.78508580e-01 -8.23144853e-01 7.32041538e-01 6.53392598e-02
3.81153375e-01 -2.76353300e-01 2.63730455e-02 -1.06184542e+00
1.14519902e-01 -6.30145907e-01 9.48152184e-01 -5.02853274e-01
-1.71122706e+00 8.42661560e-01 -2.02685013e-01 -8.47345769e-01
2.40660131e-01 -8.37867618e-01 -1.15472031e+00 7.53965437e-01
-1.54403710e+00 -1.54524505e+00 -8.21885884e-01 7.41951883e-01
1.88331783e-01 -1.64315730e-01 7.16167867e-01 3.01713288e-01
-7.06769049e-01 1.04063869e+00 -3.44963282e-01 3.94136369e-01
4.16016102e-01 -1.36701548e+00 6.40819252e-01 1.01411891e+00
2.73972481e-01 8.38980377e-01 5.67770265e-02 -6.44103527e-01
-1.65437150e+00 -1.47539616e+00 7.42127120e-01 -6.07966669e-02
7.95962334e-01 -6.81282461e-01 -1.43496954e+00 6.36485636e-01
8.25015903e-02 4.91300881e-01 3.76186848e-01 2.93759525e-01
-8.06411386e-01 -4.32231486e-01 -1.05601811e+00 6.14959180e-01
1.61795223e+00 -7.11001039e-01 -4.51661170e-01 1.98245585e-01
1.00453842e+00 -2.10458383e-01 -1.07503855e+00 5.29309809e-01
1.83689237e-01 -7.04140127e-01 9.29685652e-01 -9.07611668e-01
1.31851882e-01 -2.20007390e-01 -3.62855271e-02 -1.28123748e+00
-1.13052285e+00 -8.06944191e-01 -1.75344691e-01 1.17105711e+00
2.67719507e-01 -9.18434441e-01 8.81509840e-01 1.45618781e-01
-4.30545688e-01 -8.99466813e-01 -7.34470248e-01 -7.43498504e-01
1.09812371e-01 -1.07268386e-01 8.58601511e-01 1.30032957e+00
2.22524405e-01 4.45572823e-01 -2.48063318e-02 3.91058087e-01
5.16472518e-01 3.79227489e-01 6.67981923e-01 -1.29399061e+00
-2.93068327e-02 -1.04313338e+00 -1.02279103e+00 -8.72955680e-01
4.62078989e-01 -1.56598532e+00 -3.04285020e-01 -1.93120837e+00
1.74021497e-01 -1.24507561e-01 -8.23936462e-01 6.55275464e-01
-6.06969476e-01 1.87448755e-01 -1.11613059e-04 -1.01798005e-01
-6.31901681e-01 5.56303084e-01 1.46972609e+00 -4.28556263e-01
-1.10489830e-01 -3.62413675e-01 -9.28069711e-01 3.35743934e-01
9.17250931e-01 -1.55835479e-01 -6.64477706e-01 -3.80850047e-01
8.29410553e-03 -4.20551449e-01 3.24997365e-01 -9.43243742e-01
3.66954952e-01 1.23058692e-01 2.35913336e-01 -3.51301819e-01
-1.62608996e-01 -5.48388183e-01 -7.58631304e-02 5.46880186e-01
-2.38871872e-01 2.05530524e-01 2.26205632e-01 8.62222672e-01
-2.59711087e-01 1.48273930e-01 6.94460332e-01 -1.17362395e-01
-1.07223606e+00 1.09492970e+00 1.73933908e-01 1.53096139e-01
8.66159081e-01 -3.59248728e-01 -5.39097071e-01 -1.88187540e-01
-5.50821722e-01 4.83267307e-01 2.77359396e-01 8.20841491e-01
5.19983172e-01 -1.64033139e+00 -6.72659874e-01 4.22344744e-01
5.40094376e-01 -3.89160775e-02 4.18487251e-01 3.07621151e-01
-3.39921534e-01 1.28262803e-01 -3.80413294e-01 -4.03592169e-01
-1.06456327e+00 7.09656537e-01 3.68659556e-01 -6.90816104e-01
-8.98155868e-01 1.14957619e+00 3.90895247e-01 -4.82488364e-01
1.41842663e-01 -6.62211359e-01 -2.55498290e-01 -1.12387650e-01
3.62519443e-01 1.17567338e-01 1.44022480e-01 -6.29026473e-01
-3.95791173e-01 6.74658597e-01 -2.55113155e-01 9.93206978e-01
1.16226554e+00 1.79605886e-01 -4.09423083e-01 -3.81100215e-02
1.57271624e+00 -1.71444491e-01 -9.38227713e-01 -4.39327955e-01
1.42855108e-01 -4.65105832e-01 -8.40501711e-02 -4.01869923e-01
-1.26622915e+00 8.62324119e-01 7.45288208e-02 5.87741673e-01
1.12477291e+00 2.15154439e-01 7.85915375e-01 3.97415221e-01
2.13078141e-01 -5.94260395e-01 2.71222532e-01 4.23939019e-01
1.12259984e+00 -1.13623321e+00 3.44274864e-02 -7.75919616e-01
-1.13448165e-01 1.05403602e+00 1.06519556e+00 -5.81064939e-01
7.70065308e-01 -4.84153889e-02 -6.02931321e-01 -6.01791739e-01
-5.56034982e-01 -2.99742341e-01 8.69310319e-01 7.32678294e-01
3.58397841e-01 2.38680154e-01 -1.26045570e-01 6.56758308e-01
-1.03917643e-02 -6.60479665e-01 1.23728558e-01 6.03402853e-01
-5.05336106e-01 -9.64743376e-01 3.94289583e-01 6.37873054e-01
-1.79075330e-01 -2.32961521e-01 -6.50690198e-01 1.01536846e+00
-6.26948923e-02 6.34262800e-01 1.29762948e-01 -5.95073700e-01
4.14885908e-01 -2.87965149e-01 3.81385088e-01 -7.97208548e-01
-7.95633912e-01 -3.10538888e-01 -1.95626132e-02 -8.66828740e-01
-5.28495535e-02 -1.07251659e-01 -1.37708473e+00 -4.17422861e-01
-2.63171554e-01 2.84362771e-02 -1.70818180e-01 4.46490109e-01
6.78271174e-01 7.74458408e-01 5.03751934e-01 -6.91379964e-01
-1.55032799e-01 -9.10133362e-01 -6.86318636e-01 6.55479133e-01
1.42701745e-01 -6.69125199e-01 -2.01504529e-01 -6.05103135e-01] | [7.098781585693359, 6.285186767578125] |
1247aef0-80c9-43ca-a468-e23025f749b0 | foodx-251-a-dataset-for-fine-grained-food | 1907.06167 | null | https://arxiv.org/abs/1907.06167v1 | https://arxiv.org/pdf/1907.06167v1.pdf | FoodX-251: A Dataset for Fine-grained Food Classification | Food classification is a challenging problem due to the large number of categories, high visual similarity between different foods, as well as the lack of datasets for training state-of-the-art deep models. Solving this problem will require advances in both computer vision models as well as datasets for evaluating these models. In this paper we focus on the second aspect and introduce FoodX-251, a dataset of 251 fine-grained food categories with 158k images collected from the web. We use 118k images as a training set and provide human verified labels for 40k images that can be used for validation and testing. In this work, we outline the procedure of creating this dataset and provide relevant baselines with deep learning models. The FoodX-251 dataset has been used for organizing iFood-2019 challenge in the Fine-Grained Visual Categorization workshop (FGVC6 at CVPR 2019) and is available for download. | ['Serge Belongie', 'Ajay Divakaran', 'Karan Sikka', 'Weijun Wang', 'Parneet Kaur'] | 2019-07-14 | null | null | null | null | ['fine-grained-visual-categorization'] | ['computer-vision'] | [ 7.53813535e-02 -4.41978157e-01 -3.16258520e-01 -7.41015017e-01
-5.09355485e-01 -9.48306620e-01 5.79521596e-01 9.32330012e-01
-3.77367884e-01 1.31970003e-01 1.89671367e-01 -2.20300108e-02
-2.53570336e-03 -7.33319223e-01 -8.46918523e-01 -3.42506737e-01
-3.10647041e-01 3.36507231e-01 -3.14417556e-02 -1.24742039e-01
-2.26734728e-02 6.92187855e-03 -1.88474226e+00 9.26118016e-01
4.97702330e-01 1.31366873e+00 2.23425418e-01 8.60501289e-01
1.85878668e-02 6.44250810e-01 -3.38598013e-01 -2.13927031e-01
3.87621135e-01 -3.04429352e-01 -8.12058389e-01 -1.88186299e-02
1.02384806e+00 -4.57860082e-01 -5.42256124e-02 1.26065207e+00
3.47469360e-01 4.37935255e-02 1.06072426e+00 -1.46747673e+00
-1.12625766e+00 8.78014028e-01 -5.26454329e-01 3.12292248e-01
1.18142217e-01 1.26466483e-01 8.37945163e-01 -6.04770839e-01
4.35430735e-01 1.43645716e+00 1.00278580e+00 3.73293608e-01
-1.32170928e+00 -6.24370277e-01 1.25193253e-01 5.34193873e-01
-1.20183349e+00 -3.80921513e-01 3.04397553e-01 -8.16333115e-01
1.13564026e+00 1.43821746e-01 7.64431417e-01 1.15422761e+00
-1.35374278e-01 7.83977926e-01 1.22234941e+00 -7.05340922e-01
2.60224044e-01 2.27988735e-02 9.19499159e-01 6.74192250e-01
4.78018820e-01 5.36933601e-01 -1.89204440e-01 1.80234507e-01
2.89983243e-01 -1.75531462e-01 1.99717749e-02 -4.72519338e-01
-1.14753771e+00 1.32980204e+00 1.06984484e+00 1.66633785e-01
-3.86696279e-01 9.53903124e-02 6.92994118e-01 4.29374486e-01
3.96506488e-01 2.34673232e-01 -6.81831598e-01 4.62771237e-01
-5.95469475e-01 4.59483117e-01 7.65334845e-01 7.48454750e-01
2.67424554e-01 -2.24826127e-01 -4.71336991e-02 1.02447629e+00
5.29506624e-01 7.28083849e-01 5.38683414e-01 -7.20564187e-01
2.27544621e-01 4.10866022e-01 1.22617446e-02 -1.18375099e+00
-6.46632254e-01 -1.95907094e-02 -7.30057955e-01 4.54288036e-01
5.82369566e-01 3.66905957e-01 -1.23724902e+00 1.64448905e+00
3.43804568e-01 -3.42662483e-01 1.87902242e-01 9.09475625e-01
1.52314854e+00 4.67500657e-01 6.77239001e-01 5.02540171e-01
1.71892667e+00 -1.17792726e+00 -2.68657207e-01 -2.12892905e-01
5.46269417e-01 -9.21978712e-01 1.04504526e+00 3.58105510e-01
-5.28796315e-01 -1.05084980e+00 -1.34917283e+00 -2.28067786e-01
-1.15802622e+00 9.47540179e-02 5.37914932e-01 6.15982711e-01
-9.42359507e-01 5.23677647e-01 -4.56964970e-01 -8.19148302e-01
4.77485955e-01 1.04463466e-01 -2.79410005e-01 -5.40982783e-01
-1.23687327e+00 9.35139894e-01 7.76260555e-01 -1.70954809e-01
-1.01053500e+00 -8.97619188e-01 -1.20866811e+00 -1.21215761e-01
-2.00660199e-01 -3.45019966e-01 1.47156024e+00 -1.04785061e+00
-6.34428084e-01 1.35070360e+00 7.98702836e-01 -7.28307426e-01
2.90729880e-01 3.68427224e-02 -6.42463684e-01 1.71707436e-01
4.22876805e-01 1.40714228e+00 6.53329968e-01 -1.15581608e+00
-8.06334078e-01 -3.04274380e-01 3.36495250e-01 -1.83741271e-01
2.52001490e-02 -5.46854921e-02 1.65677994e-01 -5.95655859e-01
-4.65385675e-01 -1.00260472e+00 3.03801801e-02 1.58798218e-01
-3.71100008e-01 -3.37111890e-01 4.60754097e-01 -7.99746335e-01
6.32535219e-01 -2.29430366e+00 -4.55997229e-01 -1.54545456e-01
8.94520059e-02 5.24029076e-01 -6.41738296e-01 1.04204424e-01
-3.43195558e-01 -6.84939250e-02 2.04805002e-01 9.43537280e-02
4.43473786e-01 2.78560501e-02 6.99800551e-02 4.28738117e-01
1.39438912e-01 9.32543993e-01 -1.01608717e+00 -3.16972822e-01
7.26342440e-01 4.65061635e-01 -1.02138765e-01 2.42456738e-02
-5.06211519e-01 -1.43198818e-01 -1.87477961e-01 5.56289077e-01
8.14743996e-01 -3.11734766e-01 2.98554525e-02 -7.81407058e-01
-1.27258748e-01 -1.59940675e-01 -9.21584904e-01 1.53776157e+00
-9.59175602e-02 2.97815681e-01 -3.81994955e-02 -9.69648957e-01
6.67390406e-01 -1.30966306e-01 2.56155938e-01 -9.56638753e-01
3.65771711e-01 1.26787126e-01 1.83171391e-01 -3.54528546e-01
3.24818224e-01 2.54903547e-02 -2.32338026e-01 3.68670851e-01
2.67087013e-01 3.95469852e-02 7.77570605e-01 -4.48231818e-03
6.55983031e-01 8.39532614e-02 4.81278777e-01 -8.92508388e-01
2.65962362e-01 6.77098453e-01 -1.33310882e-02 7.61413395e-01
-6.68381870e-01 5.24594605e-01 -2.20647380e-01 -1.00473583e+00
-1.18332136e+00 -9.70718682e-01 -3.40505958e-01 1.35714781e+00
6.18047602e-02 -3.46541137e-01 -8.44256282e-01 -7.47503459e-01
5.09947181e-01 5.38540363e-01 -1.26016831e+00 -3.53667408e-01
-3.01080555e-01 -8.16839159e-01 4.38169688e-01 7.73399591e-01
7.17994750e-01 -1.27587795e+00 -7.64467955e-01 1.79829329e-01
-2.46929556e-01 -8.78458619e-01 -6.39475286e-01 6.71908140e-01
-1.68060780e-01 -1.77762389e+00 -6.04709685e-01 -1.11462307e+00
1.36837885e-01 6.49509251e-01 1.78113437e+00 -2.86661573e-02
-8.21295798e-01 4.42591399e-01 -7.15179205e-01 -9.08587933e-01
-8.91850352e-01 4.64121206e-03 -2.21636206e-01 -4.69451338e-01
1.09339643e+00 2.83852160e-01 -8.23800445e-01 1.76675677e-01
-8.11848402e-01 1.42347850e-02 2.07933694e-01 6.56517148e-01
8.33659232e-01 1.15958109e-01 5.95794976e-01 -4.71864164e-01
3.15501064e-01 -5.47294199e-01 -6.29443288e-01 3.87139797e-01
-4.03738379e-01 -5.61261736e-02 6.17045641e-01 -5.08073092e-01
-6.05141103e-01 2.37614930e-01 -1.56909317e-01 -9.28593874e-02
-5.43891311e-01 2.93520778e-01 1.84696868e-01 -9.74805877e-02
9.89834726e-01 -3.05214822e-01 -1.43714398e-01 -7.32941091e-01
6.99358284e-01 6.64082944e-01 4.92274225e-01 -2.92374134e-01
1.99496701e-01 1.67026073e-01 -3.10472906e-01 -7.42157817e-01
-1.09535098e+00 -5.56752443e-01 -8.13086569e-01 -2.17445157e-02
1.24703336e+00 -1.01620197e+00 -5.02811193e-01 6.66517615e-01
-6.30842566e-01 -7.24968553e-01 -2.39146784e-01 3.37139815e-01
-3.53722513e-01 8.69436637e-02 -6.52508914e-01 -1.63882047e-01
-7.08398223e-01 -9.08583462e-01 1.00534725e+00 1.00665174e-01
-2.98246741e-01 -9.21123922e-01 1.09965302e-01 3.21886271e-01
3.37915152e-01 5.41726351e-01 1.07716942e+00 -5.22461593e-01
1.12050571e-01 7.81644322e-03 -6.43513441e-01 4.10946935e-01
1.72528774e-01 -1.81917474e-01 -9.90416646e-01 -4.30430144e-01
-4.63681191e-01 -9.89724159e-01 1.21426308e+00 7.32090175e-01
9.06583071e-01 2.15734318e-01 -2.73951411e-01 3.64636630e-01
1.64795017e+00 1.62228629e-01 5.91599271e-02 5.72796941e-01
6.91336393e-01 7.38809407e-01 6.30735755e-01 3.67100574e-02
7.17837214e-01 6.88581288e-01 4.94219273e-01 -1.77215174e-01
-8.43829036e-01 -1.57326221e-01 -4.28889953e-02 7.32430398e-01
6.05603755e-01 -1.86101452e-01 -8.37580740e-01 5.78332126e-01
-1.64096415e+00 -6.60063863e-01 -3.58369946e-01 1.86575365e+00
7.40787923e-01 -3.03486228e-01 5.23932099e-01 3.48597586e-01
7.72244096e-01 -1.82383135e-01 -7.19228685e-01 -4.94319499e-01
-2.00776666e-01 1.42749444e-01 4.92009431e-01 2.85013288e-01
-1.99361014e+00 5.63748598e-01 6.85390854e+00 6.80346429e-01
-9.59034264e-01 1.22071043e-01 8.43204439e-01 2.64960587e-01
3.73669028e-01 -1.07383263e+00 -7.71071494e-01 4.36912835e-01
1.14236438e+00 1.69023946e-01 4.51789498e-01 9.15087938e-01
-2.33588487e-01 -9.35813859e-02 -1.24561310e+00 1.01468933e+00
2.84025043e-01 -1.04511654e+00 -1.73138902e-01 -1.78892702e-01
6.76432371e-01 5.27525961e-01 -9.72558185e-03 5.53284109e-01
8.78267646e-01 -1.10413849e+00 1.15570641e+00 -1.61780208e-01
8.65627944e-01 -5.30168235e-01 5.75297713e-01 -1.33989528e-01
-1.48341727e+00 -1.44223303e-01 -8.51731896e-01 1.06158003e-01
-5.22551239e-01 3.23198676e-01 -2.99151093e-01 3.63600552e-01
1.36753261e+00 8.20215940e-01 -8.72152388e-01 9.52810168e-01
1.06723018e-01 3.52861702e-01 -4.89726037e-01 -3.25453980e-03
4.10812497e-01 1.43730685e-01 -4.26931173e-01 1.41954052e+00
3.55526619e-02 -2.08465949e-01 8.87403607e-01 5.31625390e-01
6.44117966e-02 1.36269867e-01 -4.15155858e-01 -2.03085601e-01
1.31335258e-01 1.33273482e+00 -1.14459574e+00 -5.86215019e-01
-5.29187918e-01 6.27560496e-01 3.38819206e-01 2.96539650e-03
-7.57884264e-01 -1.11558996e-01 7.64759302e-01 -1.99543968e-01
4.84827340e-01 1.67581230e-01 1.98450927e-02 -9.54407871e-01
-4.05795485e-01 -1.16170776e+00 8.24369311e-01 -6.70059502e-01
-2.01615500e+00 6.15958869e-01 1.48311764e-01 -7.90294170e-01
-4.97192815e-02 -9.15457547e-01 1.41235873e-01 6.68453038e-01
-1.64934492e+00 -1.61791503e+00 -7.62604713e-01 6.03687584e-01
5.40013075e-01 4.24076058e-02 1.27573562e+00 4.59092557e-01
-3.60592380e-02 5.04888535e-01 4.20839846e-01 2.80506164e-01
6.68308914e-01 -1.46482480e+00 7.06631839e-01 3.41117531e-01
2.83470362e-01 7.29295015e-02 7.58871794e-01 -4.21434432e-01
-1.12338960e+00 -1.55372882e+00 4.99819219e-01 -4.48572934e-01
6.40253782e-01 -4.74492818e-01 -6.27596915e-01 4.59901035e-01
3.20132047e-01 1.77106485e-01 9.48144376e-01 -1.48026317e-01
-9.24018264e-01 -1.59555301e-02 -1.45792258e+00 2.26097517e-02
8.47023308e-01 -2.25220859e-01 -5.55192769e-01 5.66122890e-01
6.73802495e-01 -1.33001491e-01 -8.90938282e-01 2.81332493e-01
7.92305470e-01 -7.55106032e-01 1.25796115e+00 -5.98374724e-01
4.24771398e-01 -2.17386574e-01 -6.91900194e-01 -1.74188650e+00
-8.12880397e-01 6.13445938e-01 1.91879824e-01 1.06908250e+00
1.30194947e-01 -1.61322713e-01 5.14903486e-01 7.86524862e-02
-6.54175952e-02 -2.66203970e-01 -3.11552376e-01 -9.40689921e-01
4.91822898e-01 -1.28704503e-01 7.99156666e-01 1.05689323e+00
-3.79347414e-01 3.99106055e-01 -8.49799514e-02 -7.59252980e-02
1.02651846e+00 5.31537890e-01 3.49356145e-01 -1.35940945e+00
-2.48343706e-01 -3.93887460e-01 -5.20600438e-01 -4.27258074e-01
-4.91991416e-02 -1.16752720e+00 3.11429232e-01 -1.68604815e+00
5.36751866e-01 -5.13511956e-01 -5.34070075e-01 7.34110057e-01
-1.05129547e-01 9.73827422e-01 4.51506257e-01 -6.38497099e-02
-5.82032442e-01 -6.13601180e-03 8.55028987e-01 -8.46027315e-01
2.04206422e-01 -6.03606105e-01 -8.09211969e-01 5.03727674e-01
1.17185831e+00 -2.42000848e-01 -5.18013418e-01 -5.96949935e-01
-3.87670398e-01 -7.03601241e-01 4.61204618e-01 -1.04498971e+00
-5.63085616e-01 -1.55919557e-02 9.56449449e-01 -6.95813060e-01
1.69183508e-01 -7.61133313e-01 2.22443298e-01 8.76111925e-01
-5.26905119e-01 1.04064725e-01 4.19606954e-01 3.95554483e-01
-3.25933173e-02 -2.50965774e-01 1.08002973e+00 -4.20652926e-01
-1.17025828e+00 3.84908080e-01 -1.78852037e-01 1.08937405e-01
1.02290225e+00 9.83958989e-02 -6.99429274e-01 2.78736472e-01
-8.69613647e-01 2.59722292e-01 4.75332618e-01 9.07376945e-01
1.08194105e-01 -1.60890496e+00 -8.90498161e-01 2.59016901e-01
6.91485226e-01 -7.64536738e-01 4.16442454e-01 9.96188521e-02
-7.80158818e-01 6.11302495e-01 -6.21853769e-01 -6.13138199e-01
-1.48904455e+00 1.34549844e+00 2.37814561e-01 -1.94841057e-01
-7.61407673e-01 7.02949464e-01 4.60091501e-01 -5.46422541e-01
2.00190768e-01 -6.16959691e-01 -4.87349868e-01 5.58816418e-02
1.14848948e+00 7.88151994e-02 2.56141514e-01 -8.01896691e-01
-4.47709143e-01 5.05268574e-01 1.05052385e-02 6.17280781e-01
1.34273851e+00 -1.50381327e-01 2.36684471e-01 3.10963720e-01
1.40829635e+00 -9.97937441e-01 -1.02294505e+00 -7.25518167e-02
-6.99939653e-02 -1.94770038e-01 4.15227152e-02 -1.47183812e+00
-1.11277449e+00 9.95068312e-01 1.51417959e+00 4.93989974e-01
7.87488759e-01 1.69552285e-02 7.68244863e-01 1.27195507e-01
3.76271099e-01 -9.27529871e-01 -6.18385896e-02 2.65428901e-01
8.30375373e-01 -1.59787965e+00 -2.42418498e-01 -4.98655289e-01
-1.36800051e-01 1.00833941e+00 4.70850766e-01 -2.17963487e-01
9.54417288e-01 1.65813938e-01 5.79888105e-01 -2.33794942e-01
-3.61812592e-01 -3.02962840e-01 4.08066124e-01 1.32042909e+00
5.14970958e-01 5.71947396e-01 -3.45102809e-02 5.04616678e-01
-1.82953686e-01 1.90279171e-01 6.55006319e-02 7.93286324e-01
-5.76031387e-01 -9.84814584e-01 -4.11485344e-01 6.07784986e-01
-2.93748438e-01 -1.63296521e-01 -3.04807752e-01 6.87155724e-01
5.56721628e-01 1.37183332e+00 4.96473424e-02 -2.90780663e-01
3.62303883e-01 -1.88985452e-01 8.62260640e-01 -3.74244094e-01
-8.23291540e-01 -1.34905428e-01 3.72818559e-01 -4.33703780e-01
-5.48559248e-01 -6.21701121e-01 -7.60830641e-01 -4.97104496e-01
4.50771302e-02 -8.11692029e-02 8.06810021e-01 4.91854191e-01
1.77326322e-01 5.04377902e-01 3.65085378e-02 -1.01315224e+00
-7.61304915e-01 -1.03826559e+00 -8.85111868e-01 1.00391066e+00
2.58502811e-01 -8.68519425e-01 -5.30857332e-02 4.34903979e-01] | [11.549955368041992, 4.374978065490723] |
5db3ab70-79a5-4d7f-9bea-45f40a2fa0b1 | arxiv4tgc-large-scale-datasets-for-temporal | 2306.04962 | null | https://arxiv.org/abs/2306.04962v1 | https://arxiv.org/pdf/2306.04962v1.pdf | arXiv4TGC: Large-Scale Datasets for Temporal Graph Clustering | Temporal graph clustering (TGC) is a crucial task in temporal graph learning. Its focus is on node clustering on temporal graphs, and it offers greater flexibility for large-scale graph structures due to the mechanism of temporal graph methods. However, the development of TGC is currently constrained by a significant problem: the lack of suitable and reliable large-scale temporal graph datasets to evaluate clustering performance. In other words, most existing temporal graph datasets are in small sizes, and even large-scale datasets contain only a limited number of available node labels. It makes evaluating models for large-scale temporal graph clustering challenging. To address this challenge, we build arXiv4TGC, a set of novel academic datasets (including arXivAI, arXivCS, arXivMath, arXivPhy, and arXivLarge) for large-scale temporal graph clustering. In particular, the largest dataset, arXivLarge, contains 1.3 million labeled available nodes and 10 million temporal edges. We further compare the clustering performance with typical temporal graph learning models on both previous classic temporal graph datasets and the new datasets proposed in this paper. The clustering performance on arXiv4TGC can be more apparent for evaluating different models, resulting in higher clustering confidence and more suitable for large-scale temporal graph clustering. The arXiv4TGC datasets are publicly available at: https://github.com/MGitHubL/arXiv4TGC. | ['Xinwang Liu', 'Sihang Zhou', 'Siwei Wang', 'Yue Liu', 'Ke Liang', 'Meng Liu'] | 2023-06-08 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [-6.59085035e-01 -1.46430403e-01 -4.71964210e-01 -1.12139970e-01
-3.87808323e-01 -8.60614061e-01 7.89628983e-01 5.36982238e-01
-2.55135335e-02 5.37222147e-01 -6.35967925e-02 -4.11327630e-01
-5.77119291e-01 -9.13376749e-01 -5.14149070e-01 -6.04018569e-01
-9.56228077e-01 4.82392550e-01 6.20805025e-01 -7.93030486e-02
-3.15297651e-03 3.92115712e-01 -1.12616718e+00 -6.78730756e-02
7.85464406e-01 6.42760396e-01 1.34982228e-01 5.47937989e-01
-1.25275040e-03 5.66194773e-01 -3.31445098e-01 -3.03039551e-01
2.11267725e-01 -4.11303878e-01 -8.40936601e-01 -4.44164798e-02
2.02829048e-01 4.28613126e-01 -9.73387361e-01 8.31260562e-01
1.89474002e-01 2.40967497e-01 4.80501115e-01 -1.94680023e+00
-6.21603906e-01 8.92943323e-01 -5.72586894e-01 3.79263818e-01
3.80210042e-01 2.54340410e-01 1.20816422e+00 -2.38951877e-01
9.00347054e-01 1.15839267e+00 6.40147388e-01 1.44789442e-01
-1.07144654e+00 -6.57130241e-01 3.65228295e-01 3.90707642e-01
-1.54769003e+00 8.21225345e-02 9.39723015e-01 -6.27493680e-01
6.78629041e-01 2.47417673e-01 9.41760480e-01 1.20992863e+00
6.74112290e-02 2.26574689e-01 1.04898942e+00 -1.62293270e-01
1.82938635e-01 -3.29544216e-01 4.29100215e-01 7.63112366e-01
-2.53058430e-02 1.04518145e-01 -4.65905935e-01 -2.18852416e-01
5.70835173e-01 1.09008141e-01 -2.18624547e-01 -4.23432499e-01
-1.41164398e+00 7.45947063e-01 7.89432228e-01 6.54180110e-01
-1.01219863e-02 4.49631155e-01 4.63523149e-01 6.29444897e-01
6.14746749e-01 1.80984363e-01 -4.30842102e-01 -2.31713578e-01
-7.79310703e-01 8.44786689e-03 7.34431207e-01 1.22401762e+00
8.26332390e-01 -1.70956790e-01 1.71033204e-01 5.40136278e-01
2.71189868e-01 2.86097467e-01 2.14924380e-01 -1.05161214e+00
4.62177783e-01 8.52122188e-01 -5.32402039e-01 -1.44697666e+00
-5.54846168e-01 -2.43628681e-01 -9.45959747e-01 -4.85536009e-01
5.81644058e-01 3.02499652e-01 -7.69793451e-01 1.76330829e+00
4.13866401e-01 6.29519165e-01 -3.59119207e-01 5.62973678e-01
9.57718611e-01 7.36987889e-01 -1.05827436e-01 -3.81872177e-01
9.34497416e-01 -8.07890117e-01 -4.39194143e-01 1.25718027e-01
1.04403222e+00 -3.88045967e-01 1.06398427e+00 6.77125081e-02
-3.98783803e-01 -2.74792522e-01 -6.14166319e-01 3.18235308e-01
-7.20460892e-01 -5.97890019e-01 1.11126363e+00 3.41005206e-01
-1.37570381e+00 9.19258237e-01 -8.55287790e-01 -9.66045678e-01
3.75083946e-02 5.20455651e-02 -3.92440617e-01 -2.02220991e-01
-1.13133335e+00 1.69716656e-01 5.59497416e-01 -1.71552241e-01
-7.06429839e-01 -7.12366641e-01 -7.78641641e-01 -1.61272898e-01
6.37064159e-01 -1.69233754e-01 8.32888246e-01 -2.63089508e-01
-7.90134251e-01 8.26788902e-01 2.66217887e-02 -3.82146001e-01
4.28635299e-01 4.76585418e-01 -7.63791621e-01 2.48311877e-01
3.70205432e-01 6.09384060e-01 4.89435047e-01 -8.69229853e-01
-9.16600153e-02 -4.28399414e-01 -1.35089522e-02 -3.98860782e-01
-5.03170133e-01 -2.19858736e-01 -1.09118545e+00 -8.10165882e-01
1.05178818e-01 -1.32154393e+00 -2.86866009e-01 -5.24492562e-01
-3.74791920e-01 -7.48941600e-01 1.03076720e+00 -1.96662202e-01
1.62098444e+00 -2.18254828e+00 1.40454710e-01 3.81537765e-01
6.40186310e-01 -3.65220845e-01 -3.46419990e-01 9.29053724e-01
-2.55383134e-01 6.17286980e-01 5.28851897e-02 -6.52062893e-02
-1.09414831e-01 1.33676931e-01 5.30347403e-04 5.84812045e-01
-4.68550652e-01 1.08028412e+00 -1.09243596e+00 -7.97609687e-01
6.16574623e-02 -1.26988500e-01 -3.19211245e-01 -1.98282480e-01
-3.41340631e-01 3.99769813e-01 -5.68192363e-01 7.56181180e-01
3.39996338e-01 -9.72051203e-01 5.05892456e-01 1.01034425e-01
-3.44383419e-02 -6.71423897e-02 -7.89861441e-01 1.74453413e+00
1.72278300e-01 7.06307888e-01 -2.75270671e-01 -9.76231456e-01
7.44622409e-01 2.35955447e-01 1.07041466e+00 -7.73716331e-01
-1.02512809e-02 1.19009547e-01 5.77437878e-02 -1.17769949e-01
3.96937400e-01 4.42325979e-01 -4.11252499e-01 6.92641497e-01
5.30098602e-02 3.35190892e-02 7.82606483e-01 1.20246196e+00
1.63486266e+00 -2.66255528e-01 -6.27615973e-02 -3.78065079e-01
1.37918768e-02 1.98407799e-01 6.40910685e-01 5.43192625e-01
-4.58080858e-01 4.54774499e-01 8.17973077e-01 -3.52877587e-01
-8.41652036e-01 -1.00676787e+00 1.79389283e-01 8.66761923e-01
2.37383246e-01 -1.30804121e+00 -2.68209785e-01 -8.30801129e-01
2.65580505e-01 1.62633270e-01 -6.89440787e-01 -1.61041453e-01
-5.62242329e-01 -6.64869606e-01 3.14981341e-01 2.37937987e-01
1.49974436e-01 -1.06543815e+00 3.88312936e-01 1.63431577e-02
-2.25423232e-01 -1.29133117e+00 -8.05031359e-01 -1.57868564e-01
-9.67014074e-01 -1.64340484e+00 -2.20815375e-01 -8.35798800e-01
4.65857893e-01 6.75330698e-01 1.31673431e+00 4.92957234e-01
-2.25997120e-01 7.58943796e-01 -7.23537743e-01 1.34409413e-01
-1.61490977e-01 4.33496386e-01 1.36663914e-01 -3.00282121e-01
1.12966560e-01 -9.99081552e-01 -4.37467575e-01 4.93809730e-01
-6.05905294e-01 -1.36516988e-01 9.64624137e-02 2.14610606e-01
4.05842334e-01 5.23397088e-01 7.27857590e-01 -1.00399125e+00
4.59316730e-01 -7.92219996e-01 -8.70004833e-01 1.82378411e-01
-8.84793162e-01 -3.85887861e-01 6.93875074e-01 -6.21395707e-01
-2.89948314e-01 -3.78195524e-01 5.58305979e-01 -7.79431880e-01
2.19491586e-01 9.50031817e-01 2.75819212e-01 -2.98716743e-02
5.25396943e-01 6.85522007e-03 -8.30884054e-02 -3.67411047e-01
5.70341289e-01 6.11169972e-02 2.79455096e-01 -7.23565161e-01
1.14722335e+00 2.61605084e-01 1.10499203e-01 -7.56446421e-01
-4.59293544e-01 -6.98714852e-01 -8.62984419e-01 -6.84047401e-01
7.06975400e-01 -7.66511559e-01 -8.46646249e-01 4.48405564e-01
-6.73781455e-01 -8.42257738e-01 -4.59293723e-02 3.92471194e-01
-1.05621919e-01 5.79747677e-01 -8.00932348e-01 -4.34111536e-01
8.57712701e-02 -6.97557211e-01 8.18304479e-01 -1.25593409e-01
-1.42759860e-01 -1.49540877e+00 8.84219110e-02 2.70176411e-01
1.37250245e-01 6.51879549e-01 1.04359734e+00 -4.66809154e-01
-1.02787149e+00 -1.88556418e-01 -6.39066026e-02 -3.90123934e-01
4.78463136e-02 4.21099335e-01 -3.65353525e-01 -6.55654132e-01
-6.03482604e-01 -3.17509770e-01 8.12389016e-01 3.20726246e-01
1.28408432e+00 -4.48147431e-02 -7.54098296e-01 4.14016575e-01
1.32317269e+00 1.37800947e-01 3.43596607e-01 1.68452159e-01
1.02180552e+00 6.76770926e-01 7.06939638e-01 3.21183532e-01
7.13273764e-01 6.42108083e-01 3.40293527e-01 2.97133774e-01
-1.27999380e-01 -2.89871752e-01 2.62570977e-01 1.47363305e+00
5.31006940e-02 -5.89697719e-01 -1.51297545e+00 6.45006359e-01
-2.20806551e+00 -9.21784222e-01 -6.68493569e-01 1.98973036e+00
5.30787945e-01 9.11782756e-02 3.80084246e-01 -9.13467705e-02
9.35332835e-01 5.32227635e-01 -3.94987136e-01 3.85120720e-01
-1.27747059e-01 -1.63754120e-01 2.71759748e-01 2.46120140e-01
-1.02556956e+00 1.10513091e+00 5.94209099e+00 9.61323023e-01
-8.84089172e-01 5.74593842e-02 5.77581406e-01 -1.07535847e-01
-3.64332438e-01 4.61902857e-01 -2.21319020e-01 6.90580845e-01
1.30218613e+00 -7.51281142e-01 6.02310538e-01 6.20621979e-01
2.02556089e-01 9.28944945e-02 -1.11105192e+00 1.09463358e+00
-4.16671008e-01 -1.39294755e+00 -3.42477024e-01 4.97786939e-01
7.68710315e-01 3.30164254e-01 -9.67165604e-02 6.55352771e-01
4.78094846e-01 -9.15127814e-01 3.57315540e-01 3.52163106e-01
7.20061302e-01 -6.13338888e-01 3.09286982e-01 3.58832836e-01
-1.82014215e+00 5.65361120e-02 -1.86846346e-01 8.29001889e-02
3.03986389e-02 8.15978229e-01 -6.71407521e-01 8.34038496e-01
1.15386796e+00 1.46461821e+00 -1.16088927e+00 8.36223900e-01
1.19439870e-01 9.26941216e-01 -3.50008667e-01 -6.82656690e-02
2.46438280e-01 -6.73875213e-01 4.45169568e-01 7.56835818e-01
3.92817885e-01 4.70487960e-02 6.37907743e-01 6.26515746e-01
-5.31068325e-01 1.95093006e-01 -9.89367604e-01 -6.43636942e-01
9.71497774e-01 1.51971102e+00 -1.17094314e+00 -1.33663550e-01
-4.51188236e-01 5.34886360e-01 6.43797398e-01 5.08834660e-01
-1.02920794e+00 -1.07695378e-01 1.91677451e-01 3.44979614e-01
-1.54859088e-02 -7.34330535e-01 3.69661838e-01 -1.21958232e+00
-4.89727855e-02 -6.50786161e-01 9.32791114e-01 -6.98054135e-01
-1.57412791e+00 3.80546629e-01 1.65979907e-01 -1.11756682e+00
1.70001891e-02 -3.24811339e-01 -7.85070419e-01 2.87081838e-01
-8.01735342e-01 -1.16085529e+00 -5.49188733e-01 9.58095074e-01
-4.09898795e-02 -1.88955918e-01 3.36975664e-01 4.42082375e-01
-6.78001285e-01 3.24592680e-01 -1.49796791e-02 1.24227516e-01
8.30978096e-01 -1.37223744e+00 5.47546923e-01 9.11641598e-01
2.19933748e-01 6.40449882e-01 4.51862097e-01 -9.56879199e-01
-1.67275000e+00 -1.54313195e+00 6.36681616e-01 -6.17312670e-01
1.33212292e+00 -6.93287909e-01 -1.03445852e+00 9.23743784e-01
1.19005881e-01 3.28976840e-01 3.53249371e-01 5.34566581e-01
-4.28829461e-01 -1.42182425e-01 -7.06475258e-01 7.23411977e-01
1.58087671e+00 -7.09705472e-01 4.54368405e-02 8.50284636e-01
1.25346839e+00 3.63561139e-02 -1.54879057e+00 3.69180709e-01
1.08882748e-01 -7.13938892e-01 7.59994924e-01 -3.63168567e-01
-9.96633172e-02 -2.46961430e-01 2.83851419e-02 -1.11617112e+00
-5.57628870e-01 -9.37366724e-01 -2.12066531e-01 1.59086943e+00
3.90551776e-01 -8.35427284e-01 1.00732934e+00 2.70708114e-01
-7.14640245e-02 -4.86410081e-01 -8.58558893e-01 -1.28835273e+00
1.24275923e-01 -5.58306694e-01 5.64679980e-01 1.55754244e+00
1.60036702e-02 4.49447960e-01 -2.31850833e-01 -2.19106041e-02
7.97685146e-01 5.13093472e-01 9.30914462e-01 -1.50988877e+00
-7.85086602e-02 -5.25973499e-01 -5.20527780e-01 -5.12341261e-01
3.54102969e-01 -1.50542474e+00 -5.01414835e-01 -1.44072664e+00
2.41754055e-01 -7.57960677e-01 -1.02839895e-01 5.30404210e-01
-3.59655581e-02 -4.02989890e-03 1.21864103e-01 4.86486048e-01
-1.10909724e+00 6.50171101e-01 1.21442986e+00 -1.32836655e-01
-2.13629141e-01 -2.48647466e-01 -3.09609264e-01 2.07134590e-01
1.05815125e+00 -5.51028907e-01 -6.91494763e-01 1.10293090e-01
2.61319220e-01 2.57333815e-01 3.64801526e-01 -7.67794132e-01
5.34636199e-01 -2.29908869e-01 -2.43798241e-01 -7.44068801e-01
-1.20292939e-02 -5.44331014e-01 5.07754147e-01 3.67954195e-01
2.32588556e-02 4.65126187e-01 -1.01674758e-01 1.02406299e+00
-2.61454701e-01 4.90281165e-01 5.51951528e-01 -1.08185977e-01
-9.07207370e-01 1.03813839e+00 -4.93512154e-02 4.74903554e-01
1.25711501e+00 1.86047070e-02 -6.23791933e-01 -5.06070971e-01
-8.03794920e-01 7.95263588e-01 6.55552030e-01 7.48301804e-01
4.23620164e-01 -1.65584600e+00 -4.56581563e-01 -3.70529413e-01
4.09652472e-01 -1.75298214e-01 2.40478620e-01 1.19707334e+00
-2.71366239e-01 3.56491506e-01 3.28280866e-01 -8.36419523e-01
-1.25500965e+00 9.78148103e-01 7.32571632e-02 -3.28073025e-01
-9.02821660e-01 4.80829030e-01 3.87712419e-02 -6.77236259e-01
1.41851142e-01 -8.94680768e-02 -5.59075661e-02 8.66632611e-02
-2.29770839e-01 3.51640612e-01 -2.06068248e-01 -4.07517612e-01
-4.38115895e-01 4.73564774e-01 1.20007813e-01 1.86117217e-01
1.40989184e+00 -1.01148017e-01 -5.53855240e-01 7.16889977e-01
1.34869015e+00 -7.13014603e-02 -6.72740936e-01 -3.57619256e-01
4.81116265e-01 -3.61703157e-01 -3.56699198e-01 -1.77959606e-01
-1.27022016e+00 3.51742595e-01 1.97162673e-01 7.67754734e-01
9.13150012e-01 5.97632110e-01 6.69168115e-01 2.70128459e-01
8.34542513e-01 -9.03081357e-01 5.90913713e-01 5.86725235e-01
7.63175488e-01 -1.32296741e+00 1.44173875e-01 -6.40681982e-01
-3.10730189e-01 7.68207490e-01 7.98086643e-01 9.06246975e-02
1.09056652e+00 -1.89182833e-01 -3.16177756e-01 -8.43529999e-01
-1.09673941e+00 -1.01262935e-01 2.34712258e-01 2.79285133e-01
3.57801825e-01 2.10629404e-01 -1.63784608e-01 2.07449198e-01
-3.46494824e-01 -5.99600196e-01 5.61160445e-01 6.93875909e-01
-9.87990107e-03 -1.31147385e+00 -3.22599597e-02 5.48383772e-01
-9.32843238e-02 8.51353183e-02 -7.13267505e-01 1.27117991e+00
-3.61518770e-01 1.11708057e+00 -1.60454735e-01 -5.55648506e-01
-7.95660913e-02 -3.00824285e-01 2.67722756e-01 -5.66815615e-01
-3.21794450e-01 -2.94496287e-02 1.08976260e-01 -8.59351575e-01
-5.26509643e-01 -7.11404324e-01 -1.34261572e+00 -9.31441903e-01
-1.34876937e-01 6.37885094e-01 1.61011741e-01 4.07887876e-01
5.66649675e-01 5.68727076e-01 7.63769805e-01 -5.58551908e-01
2.31943831e-01 -9.23582911e-01 -9.28370178e-01 6.87088072e-01
-1.80556267e-01 -5.90013385e-01 -7.09590971e-01 -2.61578232e-01] | [7.094300270080566, 6.023642063140869] |
faafec02-2804-4d01-9224-4843012b68e5 | finding-nano-otzi-semi-supervised-volume | 2104.01554 | null | https://arxiv.org/abs/2104.01554v1 | https://arxiv.org/pdf/2104.01554v1.pdf | Finding Nano-Ötzi: Semi-Supervised Volume Visualization for Cryo-Electron Tomography | Cryo-Electron Tomography (cryo-ET) is a new 3D imaging technique with unprecedented potential for resolving submicron structural detail. Existing volume visualization methods, however, cannot cope with its very low signal-to-noise ratio. In order to design more powerful transfer functions, we propose to leverage soft segmentation as an explicit component of visualization for noisy volumes. Our technical realization is based on semi-supervised learning where we combine the advantages of two segmentation algorithms. A first weak segmentation algorithm provides good results for propagating sparse user provided labels to other voxels in the same volume. This weak segmentation algorithm is used to generate dense pseudo labels. A second powerful deep-learning based segmentation algorithm can learn from these pseudo labels to generalize the segmentation to other unseen volumes, a task that the weak segmentation algorithm fails at completely. The proposed volume visualization uses the deep-learning based segmentation as a component for segmentation-aware transfer function design. Appropriate ramp parameters can be suggested automatically through histogram analysis. Finally, our visualization uses gradient-free ambient occlusion shading to further suppress visual presence of noise, and to give structural detail desired prominence. The cryo-ET data studied throughout our technical experiments is based on the highest-quality tilted series of intact SARS-CoV-2 virions. Our technique shows the high impact in target sciences for visual data analysis of very noisy volumes that cannot be visualized with existing techniques. | ['Ivan Viola', 'Timo Ropinski', 'Sai Li', 'Peter Wonka', 'Ondřej Strnad', 'Peter Mindek', 'Dominik Engel', 'Ciril Bohak', 'Ngan Nguyen'] | 2021-04-04 | null | null | null | null | ['electron-tomography'] | ['medical'] | [ 1.48274198e-01 -5.63807972e-02 4.81422842e-01 -4.80738401e-01
-5.93911827e-01 -4.19976175e-01 4.76876140e-01 1.52268589e-01
-2.81821996e-01 8.95394504e-01 -1.67375341e-01 -5.20837307e-01
1.31928697e-01 -6.00195050e-01 -4.30098385e-01 -1.01637363e+00
-4.73434180e-01 9.39679682e-01 2.88788080e-01 -9.34760645e-02
3.73052418e-01 1.03430116e+00 -1.01750195e+00 4.26854104e-01
7.91779697e-01 7.87492633e-01 4.32640523e-01 4.91022676e-01
-5.81259847e-01 2.07312912e-01 -5.90840638e-01 4.99425918e-01
2.63995498e-01 -2.44271278e-01 -6.08963788e-01 2.84303203e-02
2.08192542e-01 -2.98895240e-01 5.40230513e-01 7.85949647e-01
5.18909812e-01 -2.39508137e-01 7.95922995e-01 -9.11305249e-01
-2.33286574e-01 4.82455418e-02 -8.33324611e-01 4.34665382e-01
1.20420992e-01 4.38437372e-01 4.40088004e-01 -1.08485782e+00
1.08575201e+00 1.21485436e+00 5.12292087e-01 3.94988924e-01
-1.72070014e+00 -5.68341970e-01 -1.06166758e-01 -1.23526059e-01
-1.15526342e+00 3.01588532e-02 1.13190246e+00 -8.74999166e-01
9.09565747e-01 5.11993706e-01 8.73402059e-01 7.79563904e-01
5.54093540e-01 2.36635059e-01 1.82772017e+00 -1.54281229e-01
3.87897849e-01 3.20454776e-01 1.98023334e-01 7.11845756e-01
-5.61988214e-03 1.37912575e-02 -1.00498229e-01 -3.86094838e-01
9.25465882e-01 1.44088054e-02 -5.62812328e-01 -5.45282900e-01
-9.06657457e-01 8.60063314e-01 6.24128759e-01 4.41579133e-01
-1.73683912e-01 -1.45079419e-02 4.20097113e-01 1.64616704e-01
6.99045539e-01 5.96779168e-01 -3.12857032e-01 2.83423692e-01
-1.30742955e+00 9.12546143e-02 5.41871846e-01 4.72397447e-01
9.72241282e-01 2.08263010e-01 -5.85540272e-02 3.31199557e-01
3.87619734e-01 3.45970541e-01 1.00757383e-01 -7.55647302e-01
-9.87844914e-02 6.35608315e-01 1.27770090e-02 -6.89646065e-01
-7.99066722e-01 -3.36760074e-01 -8.49741518e-01 1.12463200e+00
3.31513107e-01 -6.99419305e-02 -1.35476017e+00 1.18801737e+00
6.48976922e-01 -3.45866233e-01 -4.88717407e-01 1.12984312e+00
9.37850118e-01 6.93842351e-01 4.20540571e-02 -4.81373310e-01
1.31600881e+00 -4.35307235e-01 -7.32152998e-01 1.78246036e-01
6.80908144e-01 -6.19534314e-01 1.19164240e+00 4.35509890e-01
-7.34693944e-01 -2.74506152e-01 -1.09913874e+00 5.91358729e-02
-4.46442395e-01 -1.28402069e-01 4.66587335e-01 6.93943799e-01
-1.14455163e+00 7.02602088e-01 -8.93399477e-01 -2.27265358e-01
6.91393971e-01 3.29114079e-01 -2.72749186e-01 4.09826189e-01
-7.53298759e-01 9.10823584e-01 7.03205392e-02 1.48794398e-01
-8.58870864e-01 -1.02466071e+00 -5.96876800e-01 -1.31697297e-01
1.31224617e-01 -6.62047267e-01 4.03332800e-01 -6.96296334e-01
-1.32131648e+00 1.12460423e+00 -7.04938546e-02 -3.30204934e-01
7.29221106e-01 7.59556070e-02 -9.15974006e-03 4.21957105e-01
-7.68167228e-02 7.02717245e-01 8.85504544e-01 -1.95182371e+00
7.75758401e-02 -5.28712869e-01 -5.47038794e-01 -2.61695683e-02
2.63828903e-01 2.50272267e-02 1.01225451e-01 -4.07129526e-01
3.03795278e-01 -4.39521402e-01 -4.65971589e-01 8.02341029e-02
-6.14531040e-01 1.64592132e-01 1.64348936e+00 -6.64448261e-01
6.86145544e-01 -1.72614980e+00 -4.11772206e-02 4.97982740e-01
6.69627190e-01 2.30036080e-01 3.40068817e-01 3.19405824e-01
-1.58339188e-01 2.43596733e-01 -6.12238884e-01 -4.64051932e-01
-2.96591014e-01 -1.02402344e-01 -3.00794780e-01 6.79774463e-01
1.35680929e-01 7.32420325e-01 -6.25515819e-01 -7.40547776e-01
4.81758744e-01 7.96742380e-01 -3.08766931e-01 3.00384402e-01
-3.50224823e-01 1.13613987e+00 -3.16481680e-01 5.83076656e-01
1.04797363e+00 -2.80839950e-01 2.90699005e-01 -3.72679889e-01
-3.77772123e-01 -9.77877825e-02 -9.04617608e-01 1.35816169e+00
1.68973184e-03 5.30150235e-01 6.86106145e-01 -8.73741269e-01
1.11993122e+00 -1.46596776e-02 5.71739376e-01 -5.91152310e-01
1.45444959e-01 1.12304248e-01 -2.55856156e-01 -5.05077839e-01
2.79595137e-01 -5.94871521e-01 3.40424746e-01 4.69201565e-01
-2.35317439e-01 -4.96315420e-01 -1.89981192e-01 2.77837992e-01
6.68788135e-01 3.16212237e-01 -2.69304384e-02 -7.85904408e-01
3.47364336e-01 1.26125842e-01 4.41772014e-01 3.20291281e-01
-2.38050252e-01 7.42849410e-01 6.30925715e-01 -7.03653634e-01
-1.23519754e+00 -1.07421446e+00 -7.09129989e-01 5.91708481e-01
-1.27979532e-01 -1.89396337e-01 -8.20469677e-01 -5.97690523e-01
-1.51149616e-01 2.81627595e-01 -6.96182311e-01 5.15495777e-01
-6.72871351e-01 -8.65949452e-01 7.70821497e-02 2.13147700e-02
1.92466900e-01 -1.16098714e+00 -1.07957578e+00 1.53881773e-01
2.83335418e-01 -6.93413019e-01 -2.34707613e-02 4.82744962e-01
-1.13113391e+00 -1.14967585e+00 -8.60186994e-01 -5.72839439e-01
7.69270778e-01 -1.54757008e-01 9.85936940e-01 2.94154197e-01
-6.67939782e-01 2.50518322e-01 -1.73046347e-02 -8.44601616e-02
-4.31706071e-01 -4.34003800e-01 -1.43427895e-02 -1.93582430e-01
1.11294221e-02 -8.56582284e-01 -7.62166858e-01 5.85618876e-02
-7.76130438e-01 2.27397338e-01 1.55340582e-01 7.69882500e-01
1.05122304e+00 -3.25467199e-01 3.52245301e-01 -1.26130676e+00
5.50163865e-01 -1.79072767e-01 -7.76323557e-01 -1.57176517e-02
-4.87280428e-01 5.74287139e-02 7.90436864e-01 -1.65437639e-01
-1.01306438e+00 -8.54252081e-04 -1.07329756e-01 -5.91477334e-01
-3.72147530e-01 2.59391963e-01 -5.64217232e-02 -3.00805539e-01
6.46162450e-01 -3.62107251e-03 2.16388702e-01 -4.97175217e-01
2.29177833e-01 4.20818657e-01 1.68977395e-01 -4.75914240e-01
6.33581460e-01 9.13867772e-01 3.22354734e-01 -1.04402566e+00
-1.18408039e-01 -2.81461030e-01 -7.95398057e-01 -3.29412878e-01
1.02446795e+00 -3.66628796e-01 -7.84799218e-01 1.66095495e-01
-1.13502085e+00 -7.14048088e-01 -2.10080951e-01 2.58415699e-01
-4.20890123e-01 6.59238279e-01 -6.57184005e-01 -8.80200922e-01
-6.79100692e-01 -1.49505472e+00 1.16033196e+00 2.15065237e-02
-1.16979323e-01 -1.11855543e+00 1.45588577e-01 1.75965905e-01
6.20871902e-01 8.31179798e-01 1.11768973e+00 -4.41871047e-01
-7.88748860e-01 2.79485911e-01 -4.38740462e-01 4.36595008e-02
-8.30619261e-02 2.40472049e-01 -1.22536922e+00 -5.08578897e-01
4.84536320e-01 -2.88199604e-01 9.55394149e-01 7.75909841e-01
1.28383410e+00 6.38046637e-02 -4.96716857e-01 9.36223388e-01
1.49558771e+00 3.14282089e-01 6.47821307e-01 2.05252111e-01
9.29172695e-01 7.86552787e-01 2.91662276e-01 3.17872912e-01
-2.11893365e-01 5.40789485e-01 4.57005024e-01 -8.14758003e-01
-6.47818893e-02 2.53143430e-01 -1.57566324e-01 5.28496802e-01
-1.32559344e-01 2.46028066e-01 -9.53459620e-01 1.75530925e-01
-1.25942826e+00 -7.26502359e-01 -5.70941329e-01 2.07566714e+00
6.33362591e-01 3.00160259e-01 8.72158557e-02 1.91079546e-02
5.47442257e-01 2.07485586e-01 -4.45664465e-01 -5.25404453e-01
-9.03023034e-02 3.45862716e-01 2.19693795e-01 7.66884565e-01
-9.99428332e-01 6.70473337e-01 6.02452707e+00 8.19580257e-01
-1.47878015e+00 1.28213882e-01 9.67807293e-01 -9.93657634e-02
-6.74872756e-01 1.23655312e-01 -5.16346514e-01 4.76135135e-01
6.76360488e-01 1.62480235e-01 5.89158200e-02 6.43735290e-01
6.36983335e-01 -5.05450368e-01 -1.03722560e+00 9.20705259e-01
-1.38194621e-01 -1.68587220e+00 1.65280625e-01 2.95450091e-01
3.71083230e-01 -8.84635895e-02 3.25470157e-02 -2.13535056e-01
-7.79604241e-02 -1.30034578e+00 3.92977387e-01 5.37109196e-01
1.02613139e+00 -7.00267434e-01 4.63498384e-01 1.99246913e-01
-1.01583886e+00 3.48456353e-01 -3.55169386e-01 2.84456581e-01
6.10567451e-01 9.86018062e-01 -1.06667066e+00 2.20949024e-01
6.47840023e-01 1.35713577e-01 -4.14025187e-01 9.61039782e-01
1.42107010e-01 4.24057394e-01 -4.45015550e-01 1.78275526e-01
1.77367643e-01 -6.48364425e-01 8.46713185e-01 1.46892750e+00
-1.74977388e-02 2.12166116e-01 3.91711473e-01 1.40349364e+00
2.03393474e-01 2.22779185e-01 -8.49089503e-01 3.07588130e-01
7.82164708e-02 1.51250720e+00 -1.44043922e+00 -2.73136795e-01
1.17490366e-01 7.77860045e-01 2.70234883e-01 4.64047909e-01
-4.79552031e-01 -2.10382000e-01 2.03563869e-01 7.17960358e-01
3.73982005e-02 -4.27139223e-01 -7.44034529e-01 -7.00931072e-01
-2.77956128e-01 -4.44683582e-01 1.44937992e-01 -9.27216232e-01
-1.10860825e+00 6.55330777e-01 8.27034414e-02 -9.56142783e-01
1.48362651e-01 -5.85685074e-01 -8.61257613e-01 1.13432956e+00
-1.23323619e+00 -1.17482877e+00 -2.28801578e-01 2.93847829e-01
2.57042944e-01 1.46584615e-01 8.02315295e-01 1.82085391e-02
-2.30610430e-01 -9.20800045e-02 -1.42501220e-02 -2.51472801e-01
4.76208508e-01 -1.49651206e+00 6.04957715e-02 6.41708314e-01
-3.08266193e-01 5.00998557e-01 9.05005991e-01 -1.07246077e+00
-1.18124533e+00 -7.19008982e-01 4.54047143e-01 -3.76003444e-01
3.67596239e-01 -7.02916384e-01 -1.42442131e+00 4.02183712e-01
2.87470043e-01 -2.98428647e-02 6.80761993e-01 -1.87641621e-01
-2.03422979e-02 1.90612257e-01 -1.57955432e+00 4.23836827e-01
4.35944110e-01 -4.70594406e-01 -5.23914576e-01 5.29416561e-01
6.05413139e-01 -3.99968952e-01 -6.99875295e-01 3.71874779e-01
1.89245805e-01 -1.29458439e+00 8.53983045e-01 -3.25802863e-01
2.04672351e-01 -5.22737801e-01 1.63076058e-01 -1.23302186e+00
-1.81192309e-01 -6.55178428e-01 3.02068800e-01 9.10986125e-01
3.18251550e-01 -6.08932018e-01 1.01497293e+00 4.92573708e-01
-3.35266978e-01 -8.90590489e-01 -1.03527379e+00 -4.99731094e-01
8.88308063e-02 -2.60905594e-01 2.43888959e-01 1.15321827e+00
-1.11487217e-01 1.44324347e-01 -1.01060368e-01 -1.65215787e-02
1.15288210e+00 3.92371625e-01 6.01890326e-01 -1.34292543e+00
-7.25816637e-02 -2.72063732e-01 -2.61915803e-01 -5.07887781e-01
-1.57605663e-01 -8.51811826e-01 -1.17765188e-01 -1.50876164e+00
3.12710494e-01 -6.89611375e-01 2.28843331e-01 2.50937343e-01
1.94181666e-01 4.37978566e-01 6.90205172e-02 3.31745952e-01
-6.36399761e-02 5.62873006e-01 1.69477963e+00 5.70510775e-02
-4.95716721e-01 -4.59250122e-01 -1.58723801e-01 7.49582946e-01
7.35274434e-01 -4.04459685e-01 -2.78128743e-01 -7.95819387e-02
-1.70513824e-01 1.58896044e-01 4.32200253e-01 -7.65821755e-01
7.54243731e-02 -3.47281359e-02 7.75606751e-01 -9.86127853e-01
3.48017305e-01 -8.18715155e-01 4.00630653e-01 5.24004877e-01
2.16163173e-01 -1.33985177e-01 2.69967556e-01 3.13981742e-01
8.14298689e-02 7.17008673e-03 1.14416373e+00 -3.95431519e-01
-2.54607499e-01 1.87771976e-01 -5.01198411e-01 -7.08502010e-02
8.82667720e-01 -5.74120462e-01 -3.53825837e-01 -1.63969606e-01
-1.14824164e+00 2.13263825e-01 8.22376311e-01 -4.52496082e-01
8.43405783e-01 -1.01158595e+00 -4.40548599e-01 4.47510153e-01
-3.52037847e-01 2.08502367e-01 5.06895900e-01 1.11239851e+00
-9.04295683e-01 2.15394363e-01 -5.60193062e-01 -1.05746746e+00
-1.33569396e+00 4.83803809e-01 4.44225609e-01 -4.13437098e-01
-1.10555196e+00 6.73258483e-01 5.04952490e-01 -3.27373505e-01
2.06867680e-02 -3.27470392e-01 -3.92179161e-01 8.54163095e-02
5.10476232e-01 1.81815729e-01 2.39076003e-01 -5.64344049e-01
-4.03272122e-01 6.07948303e-01 -9.23303962e-02 -3.96543071e-02
1.58689773e+00 -2.76515540e-02 -1.66228354e-01 5.07232964e-01
1.18854737e+00 -3.42415050e-02 -1.46902084e+00 3.47932041e-01
-3.82016413e-02 -4.27769691e-01 1.49605379e-01 -7.40562439e-01
-1.00996757e+00 1.18680656e+00 8.82970214e-01 3.16994786e-01
9.68692005e-01 -7.49032125e-02 5.66073358e-01 -4.37680408e-02
4.79922414e-01 -9.16909218e-01 1.25700310e-02 2.97758013e-01
1.00987053e+00 -1.15645981e+00 2.83586055e-01 -5.87960362e-01
-4.63197201e-01 1.26906037e+00 5.28807402e-01 -1.30498886e-01
5.54928184e-01 7.63648748e-01 3.71345073e-01 -9.11388695e-01
-4.35968906e-01 -7.13892281e-04 8.33181366e-02 9.15945590e-01
4.52194184e-01 4.80409153e-02 -3.09236467e-01 1.90963984e-01
2.62609888e-02 -1.54314488e-01 4.91073549e-01 9.40264404e-01
-7.48598456e-01 -9.16589916e-01 -6.68747246e-01 6.03838444e-01
-3.98222983e-01 -2.93492666e-03 -3.24490756e-01 7.70981133e-01
-5.80806471e-02 3.38659704e-01 2.28271753e-01 -1.32111683e-01
4.49211337e-02 8.61229822e-02 4.56749707e-01 -5.62907815e-01
-7.69849658e-01 5.55605829e-01 -9.90194082e-02 -6.18077219e-01
-3.36986929e-01 -2.45351449e-01 -1.69662738e+00 -3.06194633e-01
1.11208623e-02 9.70478281e-02 7.83492029e-01 8.07739794e-01
1.11422487e-01 4.41452086e-01 4.20205563e-01 -1.51675582e+00
1.21769130e-01 -7.63893425e-01 -8.31714630e-01 4.74469393e-01
4.25806046e-01 -6.73488438e-01 -6.31689548e-01 3.80125977e-02] | [14.150124549865723, -2.895136833190918] |
f376b702-0938-40b4-9bb2-895c074e2e20 | autolaparo-a-new-dataset-of-integrated-multi | 2208.02049 | null | https://arxiv.org/abs/2208.02049v1 | https://arxiv.org/pdf/2208.02049v1.pdf | AutoLaparo: A New Dataset of Integrated Multi-tasks for Image-guided Surgical Automation in Laparoscopic Hysterectomy | Computer-assisted minimally invasive surgery has great potential in benefiting modern operating theatres. The video data streamed from the endoscope provides rich information to support context-awareness for next-generation intelligent surgical systems. To achieve accurate perception and automatic manipulation during the procedure, learning based technique is a promising way, which enables advanced image analysis and scene understanding in recent years. However, learning such models highly relies on large-scale, high-quality, and multi-task labelled data. This is currently a bottleneck for the topic, as available public dataset is still extremely limited in the field of CAI. In this paper, we present and release the first integrated dataset (named AutoLaparo) with multiple image-based perception tasks to facilitate learning-based automation in hysterectomy surgery. Our AutoLaparo dataset is developed based on full-length videos of entire hysterectomy procedures. Specifically, three different yet highly correlated tasks are formulated in the dataset, including surgical workflow recognition, laparoscope motion prediction, and instrument and key anatomy segmentation. In addition, we provide experimental results with state-of-the-art models as reference benchmarks for further model developments and evaluations on this dataset. The dataset is available at https://autolaparo.github.io. | ['Yunhui Liu', 'Qi Dou', 'Tak-Hong Cheung', 'Fangxun Zhong', 'Yonghao Long', 'Bo Lu', 'Ziyi Wang'] | 2022-08-03 | null | null | null | null | ['procedure-learning'] | ['computer-vision'] | [ 1.09173194e-01 7.12735439e-03 -5.52294254e-01 -3.37061197e-01
-7.11104631e-01 -7.10404515e-01 1.52294889e-01 3.83868873e-01
-4.37817216e-01 3.13376576e-01 1.01342574e-01 -3.67449045e-01
-3.53041708e-01 -2.33780742e-01 -6.65172338e-01 -6.52118385e-01
-1.58869043e-01 3.33201796e-01 -1.44333392e-02 -1.83748290e-01
2.52759010e-01 4.37034696e-01 -1.23910928e+00 6.32142663e-01
8.88300478e-01 1.00408494e+00 8.82304251e-01 8.89290631e-01
4.26153541e-02 5.00863135e-01 -1.79007265e-03 -1.61804736e-01
3.77033919e-01 -1.03599563e-01 -6.53118312e-01 -2.17892006e-01
1.53422326e-01 -1.61488637e-01 -2.34261096e-01 8.89345050e-01
6.38390124e-01 4.56641614e-02 3.63701373e-01 -7.86122262e-01
8.73157382e-02 4.08158720e-01 -2.91174620e-01 1.13254771e-01
3.52402270e-01 3.37408304e-01 2.28685170e-01 -5.70343018e-01
1.09362888e+00 8.07319164e-01 4.33167577e-01 4.53921080e-01
-5.60495555e-01 -6.38761103e-01 -1.54041588e-01 1.36847064e-01
-9.17198300e-01 -2.75098681e-01 6.74141467e-01 -6.87533140e-01
4.18616921e-01 2.62564451e-01 1.04885840e+00 1.04652655e+00
9.89394009e-01 9.88772810e-01 9.63075340e-01 -1.78624168e-01
-1.21367119e-01 2.51684844e-01 -3.65203768e-01 1.23306167e+00
3.04642200e-01 2.21302539e-01 -4.68761623e-01 4.14941967e-01
1.01263845e+00 3.51464808e-01 -5.54548800e-01 -1.01934564e+00
-1.73367131e+00 3.65806460e-01 5.23782790e-01 4.38051939e-01
-4.86982405e-01 -1.90042704e-01 5.99277854e-01 3.15539315e-02
-2.85466164e-01 6.88222587e-01 -5.16225278e-01 -5.21018565e-01
-4.64803487e-01 -2.12137192e-01 9.70479667e-01 1.24138188e+00
3.88611078e-01 -4.39838767e-01 -1.29014611e-01 4.36149359e-01
1.11111840e-02 1.31004199e-01 7.05628872e-01 -1.09167647e+00
5.79991221e-01 6.84262753e-01 7.71002844e-02 -1.02427936e+00
-9.24595237e-01 -4.25421119e-01 -1.02024007e+00 5.33926412e-02
1.92267865e-01 -1.04952782e-01 -1.05134904e+00 8.81099820e-01
3.32789600e-01 1.21451236e-01 -2.94753946e-02 1.03300345e+00
1.21385789e+00 2.53705621e-01 6.83640875e-03 -3.87953937e-01
1.31730807e+00 -1.45729411e+00 -9.22510982e-01 -2.19177470e-01
1.02907825e+00 -8.14290822e-01 9.69688952e-01 7.50566661e-01
-9.56449211e-01 -6.77731752e-01 -9.27465916e-01 -5.75865507e-02
-3.49110335e-01 5.77943444e-01 1.11050951e+00 3.77436489e-01
-5.97826362e-01 5.08833289e-01 -1.41702509e+00 -3.97227854e-01
6.35171533e-01 5.56641877e-01 -7.98194766e-01 -2.84907609e-01
-5.17636776e-01 1.06244707e+00 5.55185437e-01 4.66746628e-01
-1.16989553e+00 -7.63147831e-01 -1.23178661e+00 -1.73898235e-01
7.29253709e-01 -5.34481943e-01 1.35546875e+00 -4.14514393e-01
-1.46815252e+00 1.09771872e+00 2.45451421e-01 -1.69219628e-01
5.80369651e-01 -2.92347044e-01 -2.69285768e-01 2.09610552e-01
-2.97210038e-01 4.68870819e-01 5.26134551e-01 -1.36341548e+00
-6.86866224e-01 -4.19209749e-01 1.31890923e-01 4.71593112e-01
-2.46384159e-01 -3.95931363e-01 -8.39387000e-01 -2.99915701e-01
2.39013992e-02 -1.19887698e+00 -6.01143658e-01 4.24419403e-01
-1.90057158e-01 3.17745626e-01 4.51905906e-01 -7.57568181e-01
1.12884736e+00 -2.17722344e+00 3.19993913e-01 -2.77671933e-01
9.87111405e-02 4.90214288e-01 1.33470610e-01 1.26054674e-01
1.67173311e-01 -2.12081254e-01 5.21044098e-02 -1.91430241e-01
-5.76798737e-01 3.09142977e-01 3.97419780e-01 6.19692326e-01
-3.95569980e-01 8.34187329e-01 -1.17044294e+00 -1.00076246e+00
8.44557822e-01 3.98011506e-02 -7.03919053e-01 5.06356001e-01
1.36905357e-01 1.26435649e+00 -6.25892222e-01 9.95850205e-01
4.88685548e-01 -7.20073506e-02 2.56800264e-01 -6.70746803e-01
-1.90988645e-01 -3.90320480e-01 -8.36904168e-01 3.07883167e+00
-9.09914911e-01 3.51437092e-01 3.57294261e-01 -1.03836286e+00
5.81437647e-01 3.04275900e-01 7.80338526e-01 -6.27529860e-01
4.64908004e-01 3.76188904e-01 1.51139528e-01 -1.24043727e+00
2.41686597e-01 1.68763310e-01 -1.51004463e-01 -4.71041769e-01
3.01011682e-01 -4.96446252e-01 3.32649082e-01 -1.51116788e-01
8.51674497e-01 3.83140296e-01 7.00672030e-01 -2.96007991e-01
4.74890739e-01 6.53812408e-01 5.06237745e-01 6.63753986e-01
-5.67414105e-01 3.22627276e-01 6.53623864e-02 -6.75208211e-01
-5.97871482e-01 -9.51637566e-01 -1.31024450e-01 4.43699509e-01
7.58498371e-01 -1.16422229e-01 -4.11513984e-01 -7.79944420e-01
-1.58404827e-01 3.98403376e-01 -7.45544195e-01 -1.94678575e-01
-5.74129641e-01 -3.52482229e-01 -1.32640645e-01 3.60483259e-01
1.29788727e-01 -1.19326007e+00 -7.57477641e-01 1.56975538e-01
-2.76927203e-01 -1.33896005e+00 -2.99754620e-01 1.81795746e-01
-1.06566715e+00 -1.36191714e+00 -5.32063127e-01 -1.09404552e+00
7.89095521e-01 4.53847468e-01 9.61351097e-01 -2.55534470e-01
-9.99914289e-01 3.26013595e-01 -2.94652641e-01 -7.11168826e-01
-4.05812114e-01 1.01975508e-01 -2.67285287e-01 -3.47168088e-01
-2.39290342e-01 -2.18119826e-02 -1.06021130e+00 2.42497921e-01
-5.72753370e-01 5.16665220e-01 1.13166416e+00 8.31014037e-01
6.63745821e-01 -3.87549639e-01 -1.25300586e-01 -1.07497442e+00
1.12162709e-01 -5.64203858e-01 -5.73677242e-01 2.65547961e-01
-2.70566165e-01 -2.95705706e-01 7.13820517e-01 -3.54919642e-01
-1.30580807e+00 3.54945868e-01 1.52892277e-01 -7.01100051e-01
-4.79213029e-01 6.73902512e-01 2.22546488e-01 -1.90041095e-01
4.96317625e-01 3.11130695e-02 1.12745680e-01 -3.18173766e-01
1.23681195e-01 6.24219835e-01 8.44848037e-01 -2.46921137e-01
2.64440656e-01 6.23046756e-01 1.11489438e-01 -6.76223278e-01
-9.45966363e-01 -9.19450819e-01 -7.21022666e-01 -4.11922306e-01
1.02158630e+00 -9.74559486e-01 -8.92904222e-01 3.92035991e-02
-8.34220827e-01 -2.62592226e-01 4.34773006e-02 1.09039354e+00
-6.85737669e-01 6.05294928e-02 -4.92489755e-01 -4.71580654e-01
-2.76440501e-01 -1.55063474e+00 9.63401675e-01 5.08046925e-01
4.10866663e-02 -1.10306513e+00 3.56381461e-02 7.33902037e-01
3.07505459e-01 6.95413172e-01 4.70396042e-01 -4.72551674e-01
-7.66672969e-01 -5.82814276e-01 1.37058169e-01 1.17411338e-01
3.07894409e-01 -1.74026519e-01 -4.67172176e-01 -4.34492528e-01
-2.89134197e-02 -2.58714646e-01 5.05670309e-01 6.14785254e-01
1.72786045e+00 1.72113523e-01 -7.59797573e-01 1.13494134e+00
1.46200836e+00 3.42892766e-01 2.19964176e-01 3.93619090e-01
8.34231615e-01 4.70572233e-01 1.39681232e+00 4.18289840e-01
4.85666037e-01 5.20191848e-01 7.47594833e-01 -1.97911963e-01
7.81554878e-02 1.11174814e-01 -1.10098913e-01 1.20351708e+00
-3.48571628e-01 -1.07479662e-01 -1.15455830e+00 5.68468094e-01
-1.82171321e+00 -5.03229022e-01 -7.39163384e-02 2.12836432e+00
6.06048822e-01 -1.70762330e-01 -6.06605947e-01 -4.36365247e-01
3.22352558e-01 -1.28874421e-01 -3.46859574e-01 -2.77969867e-01
5.50931215e-01 -1.97629593e-02 6.06127620e-01 8.34028870e-02
-1.41820681e+00 6.79723978e-01 4.70349169e+00 6.28005803e-01
-1.21884894e+00 9.16862190e-02 2.74476111e-01 -2.03584865e-01
4.71876800e-01 -3.22072089e-01 -4.16685104e-01 3.86315465e-01
5.22733390e-01 -7.86810368e-02 1.32932916e-01 9.95455384e-01
4.45985645e-01 -3.47309709e-01 -1.34775186e+00 1.21168768e+00
1.73957720e-01 -1.51751208e+00 -1.54472813e-01 -1.87658012e-01
8.53079677e-01 -1.56947955e-01 -1.22990392e-01 3.92846495e-01
-3.25776577e-01 -9.57220256e-01 1.16928726e-01 7.25594759e-01
8.43096972e-01 -4.16986972e-01 1.09795785e+00 5.39480269e-01
-1.06871772e+00 -2.70660758e-01 -2.29389518e-02 3.70233119e-01
2.51438227e-02 -7.16256397e-03 -9.55991626e-01 9.29356158e-01
6.80907845e-01 9.28072691e-01 -2.86859423e-01 1.45078766e+00
1.39829382e-01 -8.56360197e-02 4.37197201e-02 1.12626620e-01
2.44614214e-01 -9.18674022e-02 3.49335760e-01 1.38429868e+00
5.07918119e-01 1.91097721e-01 4.98831719e-01 -3.20590995e-02
-1.73274484e-02 2.34800905e-01 -6.85600102e-01 -4.31453139e-02
6.79829419e-02 1.71812427e+00 -6.43995881e-01 -2.34126329e-01
-4.94782269e-01 9.51503813e-01 9.64529365e-02 -8.26382861e-02
-8.07029724e-01 -1.90882280e-01 3.53238881e-01 -9.36657488e-02
-7.34135509e-02 -3.57124239e-01 -1.31171912e-01 -1.28854644e+00
-5.25029413e-02 -7.10072160e-01 4.44269508e-01 -5.96804500e-01
-4.33133006e-01 3.26095939e-01 -1.53854102e-01 -1.83562458e+00
-6.13328554e-02 -9.87544656e-01 -2.51769602e-01 3.51990759e-01
-1.67504239e+00 -1.22106600e+00 -1.25254452e+00 5.29130638e-01
9.67334151e-01 3.68501693e-02 1.11358571e+00 3.23680133e-01
-5.20187020e-01 1.73699230e-01 2.45115712e-01 5.24643101e-02
1.05774701e+00 -1.20004499e+00 -4.69600409e-01 4.25461173e-01
-3.07890147e-01 4.93341029e-01 5.18770576e-01 -4.51777369e-01
-2.12765241e+00 -1.02613235e+00 -2.12549884e-02 -5.58668911e-01
4.08423156e-01 -9.64807868e-02 -4.13616240e-01 8.08717728e-01
1.79110430e-02 4.15556550e-01 7.40597129e-01 -3.96035790e-01
6.74845636e-01 -1.47605926e-01 -1.05267799e+00 5.17290890e-01
1.15033138e+00 1.69402897e-01 -3.71280402e-01 6.61362112e-01
4.64411169e-01 -1.39327824e+00 -1.25312281e+00 9.04008567e-01
8.48531842e-01 -8.91870081e-01 9.34948146e-01 -4.85765338e-01
7.85766244e-01 -8.43902677e-03 3.67836684e-01 -1.28421438e+00
1.24863178e-01 -6.20234370e-01 -2.39351913e-01 3.64569873e-01
1.55141622e-01 -2.34238744e-01 9.90189612e-01 2.49358058e-01
-6.81353331e-01 -1.16957927e+00 -5.02723992e-01 -4.73276287e-01
-4.54617798e-01 -2.16559708e-01 -5.72090484e-02 1.02175128e+00
1.73208594e-01 -3.09925348e-01 -2.15343505e-01 2.95063317e-01
6.07149541e-01 4.18228388e-01 9.77921426e-01 -9.53418851e-01
-1.21838540e-01 -2.88379729e-01 -5.69732964e-01 -8.21527839e-01
-3.30909252e-01 -9.87621248e-01 2.60165602e-01 -1.89735997e+00
2.90400326e-01 -3.52104872e-01 -3.01442891e-01 2.56479859e-01
-1.26180828e-01 2.21629012e-02 1.75778568e-01 2.19449088e-01
-7.25702643e-01 2.26714224e-01 2.11071420e+00 5.47958165e-02
-3.96064073e-01 1.40777633e-01 -2.88789183e-01 9.81636047e-01
7.62345314e-01 -3.00907850e-01 -2.73854494e-01 -4.99416023e-01
-3.03561479e-01 7.40833163e-01 4.69084643e-02 -1.06828380e+00
6.48772359e-01 -2.38161087e-01 3.98270726e-01 -5.47361493e-01
4.82983977e-01 -1.15134180e+00 2.35548839e-02 9.89078999e-01
-1.98073119e-01 -1.98422000e-01 2.91463703e-01 4.84472662e-01
-5.19532561e-01 7.75851682e-02 6.06750369e-01 -6.73378110e-01
-1.15159476e+00 6.03193104e-01 -3.58767398e-02 -1.50167704e-01
1.32390499e+00 -2.11938724e-01 -1.06740408e-01 1.48867115e-01
-1.13448787e+00 5.39341927e-01 2.30960384e-01 8.35186899e-01
7.74519682e-01 -7.80451834e-01 -5.48221111e-01 2.26576418e-01
4.76802260e-01 6.06626153e-01 7.95968771e-01 1.43927836e+00
-1.12014759e+00 6.39478385e-01 -5.61411321e-01 -7.95914948e-01
-1.17700708e+00 6.16841614e-01 3.44816118e-01 -3.63738000e-01
-4.86643881e-01 7.13215113e-01 3.96862686e-01 -6.48537874e-01
3.84231955e-01 -5.93989789e-01 -4.24849093e-01 -2.83657402e-01
1.97654814e-01 1.25345379e-01 1.66120201e-01 -1.36244252e-01
-3.90733480e-01 5.07118285e-01 -1.03588775e-01 5.51226676e-01
1.22598445e+00 -1.60479099e-01 1.88388661e-01 4.87257391e-01
1.01372039e+00 -2.51545519e-01 -1.05718923e+00 1.64129972e-01
-2.76804537e-01 -6.96348786e-01 1.92024752e-01 -9.87448156e-01
-9.32041228e-01 9.39407647e-01 7.59161055e-01 -5.03505349e-01
1.06538904e+00 -1.81993708e-01 5.75678587e-01 4.73433733e-01
7.80129492e-01 -8.26791704e-01 -1.11480244e-02 9.09490287e-02
9.60388958e-01 -1.79463077e+00 7.07641914e-02 -6.97646677e-01
-8.99461210e-01 1.17314589e+00 9.56269622e-01 1.63742632e-01
6.90731525e-01 2.93838263e-01 3.37343991e-01 -1.95632786e-01
-3.02044153e-01 1.60571709e-01 3.45095873e-01 3.16972196e-01
4.26706374e-01 2.09562138e-01 -3.46862316e-01 5.56445539e-01
-7.63949677e-02 3.48357499e-01 4.63536799e-01 1.33942294e+00
-1.87407345e-01 -6.40936494e-01 5.36482930e-02 6.36608601e-01
-4.51997697e-01 2.24708170e-01 5.10766685e-01 9.90194499e-01
2.15380788e-01 6.71993613e-01 -3.59016985e-01 3.68906818e-02
5.72974086e-01 -4.95709777e-01 7.71739066e-01 -7.47236788e-01
-6.61835909e-01 -1.14314109e-01 -5.67692891e-03 -1.09670269e+00
-3.68648022e-01 -3.94667983e-01 -1.24044096e+00 2.33928993e-01
-1.36963233e-01 -2.99645681e-03 1.31582320e+00 7.10355937e-01
2.75597095e-01 9.73020732e-01 4.57315743e-01 -1.37306130e+00
-2.74045110e-01 -7.41816282e-01 -2.28680357e-01 3.67537618e-01
4.00548011e-01 -6.43578827e-01 1.98112167e-02 2.13884816e-01] | [14.053466796875, -3.321284770965576] |
220c8a6e-b827-4982-99d4-da3eaa698950 | runner-up-solution-to-google-universal-image | 2210.08735 | null | https://arxiv.org/abs/2210.08735v2 | https://arxiv.org/pdf/2210.08735v2.pdf | 2nd Place Solution to Google Universal Image Embedding | Image representations are a critical building block of computer vision applications. This paper presents the 2nd place solution to the Google Universal Image Embedding Competition, which is part of the ECCV2022 instance-level recognition workshops. We use the instance-level fine-grained image classification method to complete this competition. We focus on data building and processing, model structure, and training strategies. Finally, the solution scored 0.713 on the public leaderboard and 0.709 on the private leaderboard. | ['Qiankun Li', 'Xiaolong Huang'] | 2022-10-17 | null | null | null | null | ['fine-grained-image-classification'] | ['computer-vision'] | [ 6.72193840e-02 -3.59586738e-02 -3.08704346e-01 -4.85268980e-01
-9.88862991e-01 -4.09507245e-01 8.22674036e-01 7.92794451e-02
-5.57110250e-01 3.86327326e-01 -1.52552705e-02 -3.58655751e-01
2.08497241e-01 -6.49674058e-01 -7.88966775e-01 -4.49914545e-01
8.62562805e-02 -8.70053470e-02 5.35748638e-02 -2.48472109e-01
3.29663277e-01 5.25174439e-01 -1.75299215e+00 9.39537585e-01
2.02150896e-01 1.68497396e+00 -1.25309139e-01 1.07628632e+00
1.25055714e-02 7.86705017e-01 -6.60636961e-01 -5.18930495e-01
2.32746810e-01 2.13284969e-01 -8.41353416e-01 -7.82055259e-02
1.09045672e+00 4.07964177e-02 -4.95226026e-01 1.22174764e+00
3.24919939e-01 -3.72890718e-02 5.44823349e-01 -1.57361996e+00
-8.29891741e-01 2.29046181e-01 -3.28852326e-01 5.19882619e-01
1.40647307e-01 1.93436034e-02 1.20490301e+00 -1.18831038e+00
4.39732373e-01 1.01182151e+00 6.68684125e-01 6.16179466e-01
-8.22452426e-01 -5.04019380e-01 3.27798218e-01 7.50530005e-01
-1.47857654e+00 -5.27471602e-01 5.36687255e-01 -3.84273648e-01
1.42137337e+00 4.47604209e-01 5.16552925e-01 8.62625659e-01
2.14427277e-01 1.00748229e+00 1.06411552e+00 -4.03284252e-01
1.29414320e-01 2.84516752e-01 7.82959044e-01 8.00235808e-01
2.99293965e-01 1.56168923e-01 -4.52969790e-01 -1.23578109e-01
4.43189085e-01 -3.81374955e-02 -1.40392616e-01 -2.28158191e-01
-1.06052744e+00 9.62231398e-01 1.04600227e+00 3.24469239e-01
-3.99429828e-01 5.25129199e-01 4.88177240e-01 4.86186236e-01
2.13049248e-01 3.51985008e-01 -3.61276746e-01 -1.46087468e-01
-7.36112297e-01 -1.41142411e-02 4.77794290e-01 7.92100906e-01
6.82929337e-01 1.26007617e-01 -4.64050137e-02 7.83108711e-01
3.49793881e-01 2.59033963e-02 6.71880603e-01 -7.56734431e-01
3.16577196e-01 6.52262866e-01 -1.76265642e-01 -8.83370757e-01
-1.28412262e-01 -3.73038739e-01 -7.57153034e-01 2.63218760e-01
-3.08212638e-01 3.11134517e-01 -1.20487452e+00 1.17477953e+00
5.37362732e-02 3.46365482e-01 -7.90361762e-02 6.96350217e-01
1.27041817e+00 8.87629986e-01 1.36783972e-01 5.28398216e-01
1.46019268e+00 -1.50126338e+00 -3.51416469e-01 -3.66511226e-01
7.08183885e-01 -5.22602499e-01 9.43752468e-01 5.44182301e-01
-9.25145805e-01 -9.66559350e-01 -1.67641950e+00 -3.19336563e-01
-8.84066224e-01 1.53140306e-01 7.81308591e-01 8.21045637e-01
-1.37486815e+00 4.69622821e-01 -6.36003077e-01 -1.69218898e-01
3.63995463e-01 4.03890431e-01 -8.07473779e-01 -3.87968481e-01
-1.02994204e+00 1.10735083e+00 3.00305545e-01 -1.93818398e-02
-7.87878990e-01 -7.03725994e-01 -9.81157899e-01 -1.27696320e-01
-4.77846086e-01 -3.26246955e-03 1.06539571e+00 -5.54067135e-01
-9.46648002e-01 1.49206364e+00 8.16823617e-02 -8.57061744e-01
8.26022997e-02 -3.16117518e-02 -7.51624167e-01 -1.47515729e-01
-1.37952000e-01 9.89335120e-01 7.69113243e-01 -7.98459530e-01
-9.17149544e-01 -3.59429061e-01 2.44993567e-01 -1.96792677e-01
-4.90872443e-01 -7.83295855e-02 -7.49372840e-01 -5.33819973e-01
-1.17250182e-01 -9.39091861e-01 -2.36680359e-01 -1.47826653e-02
-4.25557159e-02 -6.99037194e-01 9.78411496e-01 -5.90478420e-01
7.88561821e-01 -2.53781176e+00 2.94047240e-02 9.07984599e-02
1.01247586e-01 2.77274519e-01 -3.55018109e-01 1.11217111e-01
-3.28365892e-01 2.82363594e-01 3.13159138e-01 -4.21868414e-01
4.59791452e-01 6.90190867e-03 -3.34051877e-01 5.21054924e-01
1.02525681e-01 1.05336297e+00 -3.92921150e-01 -1.31782666e-01
3.67488086e-01 6.76939666e-01 -5.44438243e-01 -2.31571794e-02
2.37929448e-01 -4.80660528e-01 -1.10491067e-01 5.61585903e-01
6.32116795e-01 -3.23691458e-01 -1.04039922e-01 -7.23492682e-01
-3.24458927e-02 4.35836822e-01 -9.80832875e-01 1.84675574e+00
-5.47647476e-01 1.11075675e+00 4.09005992e-02 -1.14066231e+00
8.59954238e-01 8.08824152e-02 2.27901697e-01 -9.39034045e-01
2.89410981e-03 1.12228267e-01 -1.74932390e-01 1.49105743e-01
7.62827694e-01 3.26192588e-01 -3.21353734e-01 -2.14105230e-02
3.03614408e-01 -1.99783683e-01 6.36954010e-02 2.58575231e-01
1.20615399e+00 -2.15411276e-01 2.53022343e-01 -4.79629010e-01
5.21422744e-01 -1.87460378e-01 2.03786314e-01 6.75446749e-01
-6.53859973e-01 7.55251169e-01 1.14690006e-01 -1.03440273e+00
-8.40437829e-01 -9.83191848e-01 -2.12487653e-01 1.00646591e+00
5.00097983e-02 -9.03062701e-01 -7.50530899e-01 -8.26109588e-01
4.51083630e-02 4.61053491e-01 -9.73846018e-01 -5.99376261e-01
-1.51479900e-01 -4.19542134e-01 5.27288496e-01 6.07189953e-01
6.69206142e-01 -1.17800176e+00 -4.76628393e-01 -1.22460820e-01
2.25867145e-02 -1.06754851e+00 -3.40013862e-01 4.14707601e-01
-4.48418409e-01 -1.10303009e+00 -5.89059234e-01 -9.71089900e-01
3.68620127e-01 2.94223428e-01 1.15687621e+00 2.87842065e-01
-7.48906553e-01 4.79030579e-01 -3.41498882e-01 -4.42768753e-01
5.90004586e-02 1.92265779e-01 -9.05789435e-03 1.92040689e-02
6.46592438e-01 9.62005556e-02 -5.80959678e-01 1.87772974e-01
-6.99566126e-01 -2.27228254e-01 4.85284269e-01 9.62230921e-01
6.92818344e-01 -6.64096847e-02 7.39395618e-02 -6.42073691e-01
6.04387462e-01 -1.88355684e-01 -7.76392519e-01 4.78036910e-01
-6.19373202e-01 -1.62567079e-01 2.13502973e-01 -1.79064702e-02
-3.61513615e-01 1.10502481e-01 -4.94722694e-01 -4.77934450e-01
3.11026797e-02 3.38982105e-01 -1.46141738e-01 -6.24189794e-01
8.49773169e-01 2.65051246e-01 -5.03950655e-01 -3.06268215e-01
3.47575217e-01 9.21832860e-01 5.81170559e-01 -3.25103700e-01
5.37539363e-01 2.32887298e-01 -4.16020006e-01 -1.11854804e+00
-6.14490032e-01 -5.36516011e-01 -2.64382988e-01 1.74810916e-01
9.05682325e-01 -1.13524854e+00 -7.14544356e-01 5.22923112e-01
-1.23344445e+00 -2.10961536e-01 -3.58012706e-01 1.56340078e-01
-2.49681845e-01 1.03632309e-01 -5.67586899e-01 -2.72732317e-01
-5.20239294e-01 -1.36694038e+00 1.25823557e+00 1.35605007e-01
-4.56372499e-02 -7.42334902e-01 3.96722853e-02 6.02588117e-01
5.93814850e-01 1.12807080e-02 3.06605339e-01 -3.78672391e-01
-5.45401812e-01 -8.25358033e-01 -5.40289223e-01 7.54357874e-01
-1.46278784e-01 5.19097261e-02 -1.21167159e+00 -4.89517927e-01
-1.55938461e-01 -6.73523962e-01 1.31754005e+00 2.12355047e-01
1.54846215e+00 -5.33319972e-02 -2.71509439e-01 8.62626433e-01
1.43698919e+00 -7.20476508e-02 9.44324076e-01 5.50249398e-01
4.85410064e-01 4.01673704e-01 3.59359443e-01 1.52499720e-01
5.73328495e-01 8.87333095e-01 5.59932888e-01 4.63327840e-02
-3.68349642e-01 -2.46339813e-01 4.18179572e-01 6.55714214e-01
1.02685347e-01 1.25289828e-01 -1.04412591e+00 5.27564287e-01
-1.41457629e+00 -7.10681677e-01 -1.48344785e-01 1.79766250e+00
4.24706459e-01 3.27867836e-01 -1.75465703e-01 2.16940090e-01
4.17234689e-01 3.40455353e-01 -2.41135940e-01 -7.42196977e-01
-3.36051136e-01 5.44904411e-01 6.51072323e-01 3.86994064e-01
-1.35549068e+00 1.02468741e+00 7.08951187e+00 1.00350571e+00
-1.27745807e+00 2.62276500e-01 6.93830252e-01 -3.16152291e-04
-7.10766837e-02 -3.40525806e-01 -1.05525768e+00 4.81433362e-01
1.41911435e+00 -1.09154537e-01 2.68405944e-01 1.22280657e+00
-7.88872063e-01 3.60269755e-01 -1.06659484e+00 1.51214230e+00
2.70397782e-01 -2.08733654e+00 1.97062746e-01 2.56143212e-01
7.08167791e-01 7.32806385e-01 2.99175203e-01 7.05993354e-01
1.29136369e-01 -1.44119585e+00 7.00563073e-01 1.15716249e-01
1.00470543e+00 -6.28135085e-01 7.77352214e-01 -1.58581555e-01
-1.30403507e+00 -3.31901992e-03 -5.98313332e-01 8.22332650e-02
-4.08067763e-01 2.89125830e-01 -3.53791028e-01 1.86789393e-01
1.07284617e+00 6.89659119e-01 -8.38371515e-01 1.05314720e+00
-2.27878198e-01 4.48566854e-01 -3.84722114e-01 -7.19191656e-02
3.19708228e-01 5.11815906e-01 -1.30125403e-01 1.41683590e+00
-5.96077461e-03 8.57506006e-04 -5.72833745e-03 1.75160304e-01
-6.39109612e-01 -2.68238485e-01 -5.94573855e-01 -1.28624976e-01
3.01146895e-01 1.38242352e+00 -2.30818912e-01 -4.63380694e-01
-3.69017631e-01 1.33469045e+00 4.37648833e-01 1.13706023e-01
-8.23087871e-01 -4.80569571e-01 9.47386444e-01 -3.85173917e-01
4.78197187e-01 -2.64759194e-02 -2.45417461e-01 -1.30290234e+00
8.39970484e-02 -1.02116883e+00 5.02703130e-01 -6.42063797e-01
-1.23492491e+00 9.71265912e-01 -5.51080644e-01 -7.68158257e-01
3.76327783e-02 -1.24522567e+00 -6.50992393e-01 7.40592957e-01
-1.53975832e+00 -1.23109722e+00 -4.78403986e-01 7.84773231e-01
5.35306275e-01 -3.51195812e-01 1.22258449e+00 4.95339990e-01
-4.80208457e-01 1.28969872e+00 -2.99561694e-02 2.04123408e-01
1.62210897e-01 -1.03461301e+00 1.03733516e+00 6.54782355e-01
6.73944116e-01 4.85235423e-01 2.91967213e-01 1.31476864e-01
-1.63363624e+00 -1.17473948e+00 9.44385707e-01 -5.37205875e-01
6.44525766e-01 -4.99081671e-01 -6.51132822e-01 6.31658733e-01
2.40492642e-01 7.23129809e-01 6.36690676e-01 2.78206617e-01
-9.75761175e-01 -2.78547645e-01 -1.34572077e+00 2.07739994e-01
6.86992764e-01 -1.10022628e+00 -5.71708560e-01 3.91092122e-01
6.30651355e-01 -2.73459733e-01 -8.36864889e-01 4.33763027e-01
6.21349573e-01 -5.89051843e-01 1.14171433e+00 -8.37719679e-01
3.41362625e-01 -2.06429973e-01 -8.23139846e-01 -1.15419602e+00
-6.03000522e-01 -2.32990578e-01 -2.85043448e-01 7.21358538e-01
4.43937540e-01 -4.31109607e-01 1.28129661e+00 4.71719116e-01
-1.96469977e-01 -6.59502923e-01 -1.32145000e+00 -8.80144298e-01
6.01490922e-02 -6.74481750e-01 2.64907688e-01 7.70138741e-01
-2.28589267e-01 1.68981344e-01 -1.90687180e-01 8.90609026e-02
8.81542027e-01 -1.13660738e-01 8.04161966e-01 -9.69799459e-01
-1.75479680e-01 -5.60740888e-01 -1.38407409e+00 -9.20561850e-01
-2.39281692e-02 -1.01801753e+00 -2.98026860e-01 -1.61922669e+00
-1.14696011e-01 -3.04091394e-01 -1.14199746e+00 5.82761645e-01
5.16412184e-02 1.08993816e+00 5.99285245e-01 6.14249557e-02
-7.39275217e-01 4.24978375e-01 2.91127533e-01 -8.12710762e-01
4.94984895e-01 -3.81332994e-01 -8.87076497e-01 4.00860310e-01
1.07258260e+00 -1.15484186e-01 -1.76088229e-01 -6.47519171e-01
-1.80011969e-02 -7.24648416e-01 6.07655883e-01 -1.27548265e+00
3.73587668e-01 1.37258515e-01 3.32787275e-01 -4.59991097e-01
8.61807942e-01 -7.86088943e-01 1.33869993e-02 6.06365144e-01
-2.90983438e-01 5.66772446e-02 6.43599153e-01 1.64232627e-01
-5.74064255e-01 5.15719801e-02 8.51628482e-01 -1.86905488e-02
-1.37832689e+00 5.06993949e-01 1.79305956e-01 -2.18336314e-01
1.24265921e+00 -3.57028186e-01 -5.99324405e-01 8.98100138e-02
-5.70890427e-01 5.52449189e-02 2.42533445e-01 8.34213853e-01
8.40088010e-01 -1.62557805e+00 -7.90882528e-01 3.83017391e-01
5.77156305e-01 -7.18339324e-01 3.07806551e-01 5.22327065e-01
-5.57833672e-01 8.64574671e-01 -4.32066947e-01 -5.11594653e-01
-1.34993124e+00 4.55856919e-01 4.33929920e-01 -2.19181210e-01
-4.34179068e-01 1.59375727e+00 2.72455007e-01 -3.50360721e-01
3.96505237e-01 -1.49543673e-01 4.19471376e-02 -1.29383057e-01
1.04726052e+00 1.30059227e-01 6.28695905e-01 -5.75610459e-01
-7.88315177e-01 4.25332159e-01 -4.30735320e-01 5.48508205e-02
1.30712891e+00 3.93560022e-01 -1.58133864e-01 2.79123038e-01
1.87195003e+00 -5.20690262e-01 -9.19990480e-01 -1.18670594e-02
2.17095949e-02 -2.54576474e-01 5.84690869e-01 -7.50130653e-01
-1.19159806e+00 1.15837097e+00 9.81950760e-01 2.19920680e-01
9.24878180e-01 -7.99134970e-02 6.62958980e-01 5.07237136e-01
6.56192958e-01 -1.02289140e+00 -3.68043244e-01 5.80948889e-01
1.09512448e+00 -1.30470359e+00 2.52158847e-03 -1.85541406e-01
-3.38849545e-01 9.25596535e-01 6.85776770e-01 -5.55199921e-01
8.78977239e-01 1.06852077e-01 -2.82062516e-02 -1.57975703e-01
-1.04302776e+00 2.61986196e-01 7.47707009e-01 7.90391624e-01
5.22383094e-01 2.12183967e-01 -1.03918277e-01 4.93243098e-01
-2.85925895e-01 4.71848734e-02 1.46165907e-01 8.81740153e-01
-4.05356854e-01 -1.04735088e+00 -7.21575916e-02 4.34985936e-01
-3.13748688e-01 -3.71060073e-01 -5.95372915e-01 4.85471696e-01
1.67036742e-01 7.70842195e-01 -2.76091080e-02 -9.07733142e-01
5.71052849e-01 -9.79407355e-02 5.28261781e-01 -5.27211070e-01
-7.22213864e-01 -8.37410331e-01 7.65506700e-02 -1.00631320e+00
4.07908447e-02 -2.15174228e-01 -9.61686075e-01 -4.59763050e-01
-1.01675041e-01 4.27959085e-01 1.27874041e+00 3.21139127e-01
6.97976351e-01 4.58335668e-01 5.63413799e-01 -9.23052430e-01
-4.56177950e-01 -7.64302254e-01 -3.61208946e-01 3.30356121e-01
4.50654149e-01 -4.65018600e-01 -2.79574692e-01 -9.83594283e-02] | [9.436620712280273, 1.7329109907150269] |
fc344280-8549-491a-868b-6fc9f1648020 | semantic-role-labeling-guided-multi-turn | 2010.01417 | null | https://arxiv.org/abs/2010.01417v1 | https://arxiv.org/pdf/2010.01417v1.pdf | Semantic Role Labeling Guided Multi-turn Dialogue ReWriter | For multi-turn dialogue rewriting, the capacity of effectively modeling the linguistic knowledge in dialog context and getting rid of the noises is essential to improve its performance. Existing attentive models attend to all words without prior focus, which results in inaccurate concentration on some dispensable words. In this paper, we propose to use semantic role labeling (SRL), which highlights the core semantic information of who did what to whom, to provide additional guidance for the rewriter model. Experiments show that this information significantly improves a RoBERTa-based model that already outperforms previous state-of-the-art systems. | ['Dong Yu', 'Linqi Song', 'Haisong Zhang', 'Han Wu', 'Linfeng Song', 'Haochen Tan', 'Kun Xu'] | 2020-10-03 | null | https://aclanthology.org/2020.emnlp-main.537 | https://aclanthology.org/2020.emnlp-main.537.pdf | emnlp-2020-11 | ['dialogue-rewriting'] | ['natural-language-processing'] | [-2.36978680e-02 6.00150287e-01 -5.83201870e-02 -4.51418102e-01
-4.28080350e-01 -8.37645710e-01 7.48637795e-01 9.25977677e-02
-3.42357934e-01 8.79120469e-01 8.95904422e-01 -4.76805449e-01
3.37016523e-01 -5.45189500e-01 6.53474629e-02 -8.06081221e-02
6.57035410e-01 5.29201031e-01 4.74538118e-01 -1.31874454e+00
1.55756265e-01 1.63578749e-01 -8.57672572e-01 5.83525956e-01
9.47850943e-01 1.97721377e-01 5.52749872e-01 6.67320549e-01
-6.10345960e-01 1.65055561e+00 -1.05421972e+00 -4.75229323e-01
-2.40555793e-01 -7.38043189e-01 -1.27673149e+00 -1.91959575e-01
-1.65997103e-01 -4.88628626e-01 -4.24681723e-01 8.73128176e-01
2.88143694e-01 6.41409457e-01 2.91411966e-01 -6.90739393e-01
-8.09174061e-01 1.03931284e+00 2.06210643e-01 3.93087327e-01
8.04614663e-01 2.03255013e-01 1.24887288e+00 -7.94492304e-01
6.46862328e-01 1.87437427e+00 3.97771537e-01 1.16509449e+00
-7.48319566e-01 -8.61145109e-02 7.65499830e-01 3.42156708e-01
-8.13709557e-01 -7.37508476e-01 8.25304627e-01 1.14533789e-01
1.45164371e+00 4.70281005e-01 3.57393742e-01 1.15313697e+00
-2.79150993e-01 8.79117489e-01 9.52781737e-01 -6.90639734e-01
-2.14700028e-01 2.08472416e-01 5.18122733e-01 7.96910822e-01
-3.18046600e-01 -3.26339036e-01 -5.06944716e-01 -5.48678974e-04
4.93723094e-01 -2.61341780e-02 -3.50004286e-01 2.78167903e-01
-7.68486440e-01 7.79396653e-01 3.02625895e-01 4.55336958e-01
-1.59741081e-02 -1.81721076e-01 4.09247905e-01 5.28415322e-01
3.91739726e-01 1.01768732e+00 -5.45762599e-01 -5.42054713e-01
-7.24309385e-02 1.92945451e-01 1.29749382e+00 1.14048302e+00
4.95009631e-01 -6.84499443e-02 -4.01871562e-01 1.24367201e+00
3.48213643e-01 3.67667943e-01 2.66083390e-01 -1.44567263e+00
4.64876652e-01 1.00730538e+00 3.98801833e-01 -5.13023138e-01
-4.54819679e-01 -1.05080992e-01 -1.58206373e-01 -4.65763301e-01
3.88125300e-01 -9.67644900e-02 -6.02807641e-01 1.79271364e+00
2.19419599e-01 -3.34461510e-01 4.80755836e-01 9.27732110e-01
1.24363410e+00 6.93498194e-01 3.29744667e-01 -2.57582724e-01
1.62986445e+00 -1.43098664e+00 -1.29735899e+00 -5.92220724e-01
8.07838261e-01 -9.74981487e-01 1.60335898e+00 -3.02216951e-02
-1.00838351e+00 -3.59976947e-01 -9.50163782e-01 -6.51167154e-01
-4.64152366e-01 -7.97702074e-02 5.66592276e-01 2.78109580e-01
-1.03862250e+00 2.48877212e-01 -3.21963400e-01 -5.76267362e-01
-1.14025034e-01 -1.35492593e-01 -9.90663841e-03 -1.99934393e-01
-2.02068114e+00 1.71592271e+00 4.30861935e-02 2.49931678e-01
-6.18270397e-01 -4.30404097e-01 -1.13373148e+00 1.20324157e-01
8.41631472e-01 -5.95553696e-01 1.79496300e+00 -4.35350209e-01
-2.09608984e+00 7.35436320e-01 -5.11238158e-01 -2.75370300e-01
2.01766059e-01 -4.93965834e-01 -2.74746239e-01 1.58798948e-01
5.08435927e-02 4.97947246e-01 4.42900777e-01 -1.44092643e+00
-6.25659645e-01 -1.71847746e-01 1.05596840e+00 7.23475754e-01
8.97952728e-03 3.90582860e-01 -3.02114695e-01 -5.32003760e-01
-1.54609844e-01 -8.50828409e-01 -1.35141626e-01 -6.33420587e-01
-2.75061697e-01 -8.13088417e-01 7.73020387e-01 -8.23379397e-01
1.55941069e+00 -1.78708577e+00 3.11164856e-01 -4.99944925e-01
3.16441000e-01 6.79968476e-01 -9.81174037e-02 8.69017184e-01
3.78343493e-01 1.54330149e-01 2.67506833e-03 -5.49108982e-01
-9.60619897e-02 6.44916654e-01 -4.37147170e-01 -1.37699157e-01
9.32230428e-02 1.13663983e+00 -1.25022221e+00 -2.91312695e-01
4.84362900e-01 4.08158116e-02 -3.13739568e-01 6.65494621e-01
-5.70306420e-01 4.50569153e-01 -5.32656074e-01 4.18825269e-01
1.46145523e-01 -1.91210911e-01 4.58391666e-01 1.00237273e-01
9.46812332e-02 1.20638621e+00 -4.08135533e-01 1.63837934e+00
-8.38752031e-01 2.68111140e-01 3.71797442e-01 -5.39561987e-01
6.88790321e-01 3.06972682e-01 -1.59675911e-01 -6.80989683e-01
1.62139386e-01 -1.30404815e-01 2.57947326e-01 -4.99592870e-01
6.10020161e-01 -2.65758097e-01 -2.89210081e-01 6.18538678e-01
-8.79303664e-02 -2.79928565e-01 1.63016304e-01 7.01266825e-01
9.14539158e-01 -6.48781508e-02 5.29236734e-01 -2.50762135e-01
9.80415046e-01 -3.06201894e-02 4.46687549e-01 8.97486508e-01
-4.00845140e-01 2.59509414e-01 5.76476157e-01 -3.38781804e-01
-5.69714010e-01 -7.15169847e-01 5.14824212e-01 1.67351103e+00
4.98881042e-01 -8.37369919e-01 -9.09493089e-01 -9.57319617e-01
-3.89371157e-01 1.37244534e+00 -3.79627526e-01 -3.92969638e-01
-1.14645112e+00 -2.25027844e-01 5.93149126e-01 4.37300503e-01
4.90652889e-01 -1.20592451e+00 -3.51743013e-01 4.23252732e-01
-6.67001307e-01 -1.30874276e+00 -5.25690615e-01 -9.89105627e-02
-5.20077050e-01 -1.20975137e+00 2.37334725e-02 -5.81531107e-01
3.27066660e-01 5.42033732e-01 1.25242579e+00 5.55578828e-01
2.57581234e-01 3.63617271e-01 -7.91852713e-01 -4.91243601e-01
-8.59809160e-01 2.28910014e-01 -3.02470744e-01 -4.19400752e-01
4.65956002e-01 -2.37689406e-01 -2.11916089e-01 3.47191751e-01
-3.85341704e-01 7.31302276e-02 -4.03154865e-02 8.02227080e-01
-2.83644497e-01 -6.71784580e-01 6.74239278e-01 -1.29863930e+00
1.35661757e+00 -1.54737860e-01 2.92307530e-02 6.73402011e-01
-3.93748075e-01 1.40310571e-01 1.00182986e+00 -4.28889506e-02
-1.59265006e+00 -5.29806674e-01 -2.12277725e-01 1.35731384e-01
4.37204204e-02 3.42488110e-01 -3.33347350e-01 2.80851334e-01
5.95499337e-01 1.97118353e-02 2.36931816e-01 -6.75417960e-01
8.19035053e-01 7.51687169e-01 2.24319264e-01 -7.69763589e-01
1.47495523e-01 1.99448362e-01 -4.27923888e-01 -8.28422368e-01
-1.50118566e+00 -7.47063637e-01 -4.30851221e-01 -2.52668947e-01
5.85039258e-01 -7.64797628e-01 -7.12367237e-01 2.96276122e-01
-1.63386393e+00 -4.66289222e-01 -2.72814929e-01 -5.41597931e-03
-3.81385654e-01 5.75502992e-01 -9.22958255e-01 -9.67762470e-01
-2.58272827e-01 -1.06665266e+00 7.90625691e-01 2.28038877e-01
-6.61013901e-01 -1.25752175e+00 -2.38574803e-01 7.68609345e-01
6.41944468e-01 -5.51285267e-01 1.09607458e+00 -1.19453824e+00
-2.65694261e-01 1.17862009e-01 -1.25314519e-02 4.49408710e-01
4.56696421e-01 -3.97971869e-01 -7.91584849e-01 3.06950897e-01
6.22210026e-01 -3.99710864e-01 6.75098836e-01 -1.65144607e-01
5.76893747e-01 -6.42640293e-01 -1.66740671e-01 -9.28674489e-02
5.81717610e-01 4.80600983e-01 4.98539984e-01 -6.25032336e-02
8.90119910e-01 9.03774202e-01 8.26224029e-01 1.58075452e-01
1.07808363e+00 7.56097019e-01 2.07024470e-01 -6.33547157e-02
-4.07345831e-01 -4.10703033e-01 3.92605692e-01 1.00068080e+00
8.25802535e-02 -3.95322978e-01 -7.08865345e-01 2.87620574e-01
-2.29207993e+00 -8.01262915e-01 -1.19176038e-01 1.49060011e+00
1.35114455e+00 3.29991952e-02 -1.56287074e-01 -4.21264708e-01
7.69565284e-01 7.00804412e-01 -2.68986791e-01 -7.10030019e-01
-9.23707411e-02 -4.24479553e-03 -3.84320356e-02 1.41232872e+00
-7.82256305e-01 1.91120446e+00 6.69058132e+00 6.13155723e-01
-5.95465124e-01 4.29421157e-01 2.11262912e-01 1.03767700e-01
-4.18940008e-01 2.18623772e-01 -8.71628284e-01 1.33185044e-01
7.47152686e-01 -1.88633755e-01 7.67547846e-01 7.72723436e-01
4.10345584e-01 -2.15814382e-01 -1.15496874e+00 6.26533151e-01
4.48160917e-01 -1.06406713e+00 2.85740107e-01 -5.54425716e-01
2.33821243e-01 -5.60323656e-01 -5.30349910e-01 6.86911583e-01
8.48203719e-01 -1.00351858e+00 3.98850739e-01 4.78261381e-01
1.93490878e-01 -5.43905914e-01 8.74967813e-01 6.63211048e-01
-8.12748551e-01 -2.53630080e-03 -2.78493613e-01 -4.97315854e-01
4.04658407e-01 8.99906829e-03 -1.09509480e+00 4.28600281e-01
2.79911906e-02 6.60512745e-01 -3.46766829e-01 4.17155996e-02
-1.15073359e+00 4.67354119e-01 -9.75999236e-02 -5.44258952e-01
3.52886975e-01 -2.20597625e-01 8.19607019e-01 1.14969683e+00
-4.68866974e-01 7.25784957e-01 5.22960722e-01 5.36762834e-01
-1.94594145e-01 -1.11133642e-02 -5.65709651e-01 9.25651342e-02
7.84140825e-01 1.05303311e+00 -3.35749686e-01 -6.44125998e-01
-3.32770258e-01 1.03362346e+00 6.45758092e-01 3.16209972e-01
-5.24744570e-01 -1.44076586e-01 8.93746734e-01 -4.26237807e-02
-8.51314887e-02 -2.60741204e-01 -2.22231433e-01 -1.15624011e+00
-1.22014821e-01 -1.02675200e+00 4.92962629e-01 -8.21728468e-01
-1.26524103e+00 5.02245307e-01 -3.22583281e-02 -4.92507458e-01
-2.96020746e-01 -5.92225373e-01 -5.74310064e-01 1.03769004e+00
-1.60244906e+00 -1.26033866e+00 -9.76898894e-03 3.62420052e-01
1.23742628e+00 -4.72475402e-02 1.02825713e+00 -1.24030739e-01
-3.03154647e-01 4.51220125e-01 -6.30777836e-01 1.41769573e-01
9.56578434e-01 -1.42533052e+00 3.50207806e-01 8.05951953e-01
-1.40183792e-01 1.21848273e+00 9.01771188e-01 -6.07623279e-01
-1.25811350e+00 -6.63799942e-01 1.39748085e+00 -1.00923884e+00
7.46125281e-01 -2.99363345e-01 -1.13136375e+00 6.93963170e-01
4.93467420e-01 -4.74002630e-01 6.23894572e-01 4.10283238e-01
-3.19910914e-01 8.30183402e-02 -9.86060679e-01 9.37878311e-01
1.38559389e+00 -8.52819383e-01 -1.53760600e+00 4.57194120e-01
1.28173923e+00 -6.06267154e-01 -2.79115438e-01 9.68937054e-02
3.89781483e-02 -6.27027452e-01 8.85702550e-01 -1.14738071e+00
4.24799532e-01 -2.73124248e-01 1.18396338e-02 -1.58528697e+00
-1.68888539e-01 -8.48258913e-01 -3.74727905e-01 1.06969607e+00
4.48557377e-01 -5.38998008e-01 3.77114043e-02 9.85169590e-01
-5.46529353e-01 -4.59989578e-01 -7.48904288e-01 -4.89106119e-01
-5.79081140e-02 -9.84853134e-02 3.58448327e-01 9.77463543e-01
7.78783917e-01 1.19184208e+00 -4.26129103e-01 -1.81963488e-01
-8.68449435e-02 -1.09235249e-01 5.00713468e-01 -9.79794145e-01
5.70413247e-02 -3.54916155e-01 3.32372636e-01 -1.62989354e+00
7.08762288e-01 -7.03320026e-01 3.16509902e-01 -1.92393529e+00
-1.48232251e-01 -3.49342346e-01 -5.21920100e-02 6.60316408e-01
-8.47840250e-01 -2.40054324e-01 4.28155452e-01 8.49000961e-02
-1.05271530e+00 8.71702611e-01 1.40327930e+00 4.03829403e-02
-3.12951326e-01 5.50403520e-02 -1.24348974e+00 8.65819693e-01
6.47493958e-01 -2.73599446e-01 -7.09953487e-01 -5.34587264e-01
-2.83513013e-02 1.14727184e-01 -2.88301036e-02 -3.57068032e-01
4.74586904e-01 -3.58072758e-01 -4.60807562e-01 -9.79560912e-02
6.10881329e-01 -6.13824248e-01 -5.92050731e-01 1.84372291e-01
-8.47386599e-01 5.53845651e-02 -7.94125199e-02 4.01575387e-01
-2.31678918e-01 -6.36937320e-01 6.63940430e-01 -7.42097974e-01
-7.41482258e-01 -4.01249141e-01 -7.20437288e-01 4.53340650e-01
6.82738304e-01 2.18213454e-01 -6.67147160e-01 -8.52285743e-01
-8.32170486e-01 6.52943015e-01 2.63196468e-01 8.24948847e-01
4.80949342e-01 -8.61619174e-01 -4.76135820e-01 -2.02950880e-01
8.60592499e-02 -7.09252432e-02 1.42772943e-01 2.81205714e-01
-3.78834516e-01 6.26913249e-01 2.58639276e-01 -6.68511121e-03
-1.39140761e+00 4.29221720e-01 4.70893353e-01 -5.00183761e-01
-5.18588066e-01 9.48416114e-01 5.19266613e-02 -6.94704652e-01
2.43110210e-01 -1.89875290e-01 -7.46030867e-01 1.91158280e-01
6.66342854e-01 2.51402050e-01 7.04189986e-02 -5.40567636e-01
-4.70086306e-01 -1.35126978e-03 -5.39552033e-01 -2.95977354e-01
8.62445891e-01 -6.17865384e-01 -4.55959320e-01 6.84173107e-01
6.29375875e-01 3.05563390e-01 -8.57567191e-01 -5.76124072e-01
1.99545205e-01 -2.91141719e-01 -3.30786020e-01 -1.31367159e+00
-3.50450814e-01 9.07210171e-01 -3.81761402e-01 5.18684864e-01
5.13933957e-01 2.47891068e-01 7.78172731e-01 8.86340559e-01
4.77260053e-01 -1.39170647e+00 2.05330476e-01 1.53884077e+00
1.22861397e+00 -1.21745765e+00 -7.35597387e-02 -9.32071745e-01
-1.12352562e+00 1.08205950e+00 1.11724043e+00 2.15969011e-01
2.92990625e-01 1.24044731e-01 6.02414489e-01 -2.61404067e-01
-1.19415236e+00 -4.36361015e-01 -2.63568431e-01 4.52240348e-01
5.86637497e-01 3.19708250e-02 -5.23278534e-01 7.90438294e-01
-2.27520019e-01 -5.43791234e-01 6.48842275e-01 1.00941050e+00
-6.12990439e-01 -1.39640248e+00 9.83465761e-02 4.55694497e-02
-3.58420193e-01 -4.54595178e-01 -1.18635643e+00 6.30624413e-01
-4.85091418e-01 1.52834785e+00 -3.33139807e-01 -5.03399968e-02
8.02097678e-01 5.46665072e-01 3.02317470e-01 -1.25607955e+00
-1.10738730e+00 -2.97719598e-01 9.20457125e-01 -5.47959030e-01
-3.82962376e-01 -2.01845348e-01 -1.52230954e+00 -2.81077057e-01
-2.56904989e-01 2.70302862e-01 1.65805176e-01 1.32095206e+00
2.44059220e-01 5.88244855e-01 6.27624452e-01 -2.02470735e-01
-9.06290412e-01 -1.35422087e+00 9.33553204e-02 5.55525541e-01
2.27299333e-01 -6.11680210e-01 -4.03041214e-01 -2.23202452e-01] | [12.428603172302246, 8.0397310256958] |
fedcd317-5ecd-4567-937f-a01a5e5f0a3e | large-language-models-as-tax-attorneys-a-case | 2306.07075 | null | https://arxiv.org/abs/2306.07075v1 | https://arxiv.org/pdf/2306.07075v1.pdf | Large Language Models as Tax Attorneys: A Case Study in Legal Capabilities Emergence | Better understanding of Large Language Models' (LLMs) legal analysis abilities can contribute to improving the efficiency of legal services, governing artificial intelligence, and leveraging LLMs to identify inconsistencies in law. This paper explores LLM capabilities in applying tax law. We choose this area of law because it has a structure that allows us to set up automated validation pipelines across thousands of examples, requires logical reasoning and maths skills, and enables us to test LLM capabilities in a manner relevant to real-world economic lives of citizens and companies. Our experiments demonstrate emerging legal understanding capabilities, with improved performance in each subsequent OpenAI model release. We experiment with retrieving and utilising the relevant legal authority to assess the impact of providing additional legal context to LLMs. Few-shot prompting, presenting examples of question-answer pairs, is also found to significantly enhance the performance of the most advanced model, GPT-4. The findings indicate that LLMs, particularly when combined with prompting enhancements and the correct legal texts, can perform at high levels of accuracy but not yet at expert tax lawyer levels. As LLMs continue to advance, their ability to reason about law autonomously could have significant implications for the legal profession and AI governance. | ['Jungo Kasai', 'Jonathan H. Choi', 'Aaron Travis Lee', 'Raghav Jain', 'Meghana Bhat', 'WenTing Tao', 'Sarah B. Lawsky', 'David Karamardian', 'John J. Nay'] | 2023-06-12 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [-6.40031844e-02 5.78697979e-01 -5.07060468e-01 -2.97798485e-01
-1.16276622e+00 -8.27517867e-01 7.42924333e-01 3.77456784e-01
-4.33096886e-01 5.64111531e-01 5.86462319e-01 -1.40002859e+00
-6.97561026e-01 -7.37447500e-01 -4.47289318e-01 4.13315386e-01
2.99853414e-01 7.00560629e-01 1.22316279e-01 -6.19295776e-01
5.06835520e-01 4.01312828e-01 -9.57933664e-01 6.08896375e-01
1.54236376e+00 1.95707425e-01 -4.41098690e-01 3.98364156e-01
-3.43201071e-01 1.90135276e+00 -6.01258159e-01 -1.14973807e+00
5.39568722e-01 1.32136822e-01 -1.12803411e+00 -7.78260171e-01
1.13174260e+00 -7.55297422e-01 -3.23791921e-01 9.48537588e-01
1.66858345e-01 -2.30792880e-01 4.17162925e-01 -1.07952428e+00
-6.77831471e-01 8.90228868e-01 -2.02721432e-01 6.28306389e-01
7.03838706e-01 6.23714566e-01 1.28745115e+00 -8.69430304e-02
1.06028211e+00 1.36818194e+00 7.77750552e-01 3.61762166e-01
-1.11160219e+00 -7.90667415e-01 -1.51447639e-01 4.83352810e-01
-7.49048114e-01 -7.58009017e-01 1.31861567e-01 -7.78049588e-01
1.62690651e+00 1.22883156e-01 2.82945186e-01 6.16915405e-01
4.09184963e-01 4.29102987e-01 9.93224144e-01 -5.16757607e-01
1.35263782e-02 7.34967599e-03 6.11744583e-01 6.04855061e-01
8.34749639e-01 1.11999428e-02 -2.04979897e-01 -6.37300909e-01
1.60345718e-01 -5.12576282e-01 2.88407087e-01 3.90082240e-01
-5.63996494e-01 8.97651017e-01 5.74131794e-02 6.71225309e-01
-4.01123464e-01 1.75012708e-01 3.77008885e-01 3.69544566e-01
9.51420888e-02 9.57724452e-01 -4.93856728e-01 -6.32151306e-01
-7.34582305e-01 8.16032588e-01 1.11868596e+00 5.39698899e-01
4.52814490e-01 -2.89903820e-01 -2.29381442e-01 5.93701124e-01
3.00606549e-01 6.17597580e-01 1.14482790e-01 -1.54093421e+00
1.18623948e+00 1.20133567e+00 1.65455982e-01 -8.04722071e-01
-4.90176648e-01 -1.09940477e-01 1.54138342e-01 2.95985579e-01
8.04476202e-01 -1.82524189e-01 -5.50176203e-01 1.38528419e+00
1.03295431e-01 -6.94362223e-02 3.28203261e-01 2.47820079e-01
4.14271355e-01 2.76327699e-01 6.71819508e-01 2.30249465e-01
1.24452567e+00 -1.54525265e-02 -6.46370649e-01 -6.61311030e-01
1.40882742e+00 -5.75059474e-01 1.20290685e+00 1.64579898e-01
-1.15069020e+00 1.96058080e-02 -7.24979579e-01 -5.53672969e-01
-3.28539073e-01 -3.38357627e-01 9.80956435e-01 6.70067489e-01
-5.77917933e-01 5.21633744e-01 -4.68175739e-01 -3.01975548e-01
9.79099333e-01 -1.63832396e-01 -2.33157530e-01 -5.02774417e-01
-1.56938469e+00 1.54316318e+00 2.32546911e-01 -2.81955928e-01
1.04947448e-01 -1.04751658e+00 -1.08534312e+00 1.62849963e-01
5.62058508e-01 -4.27599996e-01 1.56464016e+00 -1.89799324e-01
-4.93423074e-01 6.27381504e-01 -2.15009078e-01 -7.78299570e-01
6.11343324e-01 -9.88038629e-02 -6.73092365e-01 1.37746632e-02
8.85170817e-01 2.74390608e-01 -1.92338049e-01 -4.58117515e-01
-7.58696556e-01 -4.29306209e-01 5.22039235e-01 -2.28217587e-01
-7.25398883e-02 6.66201830e-01 1.71541452e-01 -1.01164527e-01
-2.09045842e-01 -4.41849947e-01 -3.05031329e-01 -2.87120759e-01
7.75775090e-02 -3.77147794e-01 4.70064849e-01 -1.14198124e+00
1.55368280e+00 -1.63652730e+00 -7.66371787e-01 6.18717790e-01
1.07208878e-01 6.04070604e-01 5.91983423e-02 7.24178255e-01
2.02247426e-01 7.90935218e-01 3.15349102e-02 4.77212042e-01
2.48096421e-01 3.72046113e-01 -5.54522634e-01 7.46891797e-02
1.09806903e-01 1.35497236e+00 -6.85960114e-01 -7.62550116e-01
6.55683354e-02 -1.55686930e-01 -7.29665637e-01 -6.03009224e-01
-5.40204048e-01 2.24541444e-02 -4.19687986e-01 6.14630520e-01
1.79828420e-01 -2.85101146e-01 1.71127498e-01 3.83746624e-01
-1.93519145e-01 9.14013326e-01 -8.34153533e-01 1.31889451e+00
-3.92059028e-01 7.30973840e-01 -6.93653226e-02 -3.86650920e-01
5.54089725e-01 3.98105085e-01 1.53231487e-01 -1.41561866e+00
-2.67482568e-02 3.48629326e-01 6.76398158e-01 -1.16941905e+00
1.17694579e-01 -1.06528111e-01 -2.82028876e-03 7.34365284e-01
-5.44608235e-01 -1.10831596e-01 9.53897476e-01 6.90484464e-01
1.32539248e+00 -1.37896284e-01 3.84071082e-01 -1.08181186e-01
3.81102383e-01 7.09786654e-01 5.17065346e-01 1.02628219e+00
-1.53885260e-01 -5.14455914e-01 7.15533912e-01 -5.48248291e-01
-9.51123714e-01 -4.48116869e-01 -4.02767450e-01 8.65531325e-01
-4.50827003e-01 -6.46954656e-01 -5.65257907e-01 -6.57492757e-01
4.71983135e-01 1.43859053e+00 -3.30575466e-01 1.99796390e-02
-6.31782770e-01 -5.28013259e-02 1.00133908e+00 5.65571487e-01
5.61748683e-01 -9.75817442e-01 -9.86496747e-01 3.75243872e-01
-3.39549959e-01 -1.45962405e+00 9.62652192e-02 -5.25403023e-01
-6.30644262e-01 -1.61840868e+00 3.31959665e-01 -1.14713110e-01
1.37569532e-01 -3.80094320e-01 9.08641696e-01 5.98231494e-01
-2.58490920e-01 3.18951368e-01 -3.32726985e-02 -9.12984610e-01
-1.17618549e+00 5.68009913e-02 -4.36136156e-01 -7.98648417e-01
1.01736009e+00 -4.08706069e-01 1.10859580e-01 -2.49508321e-01
-1.05794072e+00 -7.92734176e-02 5.14103413e-01 2.15373471e-01
-4.26757872e-01 -2.43629768e-01 7.89212167e-01 -1.54003084e+00
1.35003424e+00 -4.61154491e-01 -8.30277324e-01 6.54975772e-01
-1.16350138e+00 1.93026245e-01 6.33829892e-01 1.42626807e-01
-1.13174307e+00 -7.75254846e-01 -3.18334073e-01 4.35958385e-01
-2.62932420e-01 9.49850440e-01 1.20985374e-01 2.91631706e-02
1.33633399e+00 -6.25545502e-01 6.70025349e-02 -2.27248132e-01
5.13242006e-01 8.54693949e-01 6.09333813e-01 -8.02665830e-01
9.79272187e-01 2.36719891e-01 -2.11207718e-01 -3.16328347e-01
-8.35827053e-01 -4.08796936e-01 -3.71451825e-01 -8.58149081e-02
5.11825442e-01 -5.78490496e-01 -9.11529899e-01 -2.28556156e-01
-1.02281046e+00 -4.71226692e-01 -2.75547773e-01 4.02466774e-01
-1.13420032e-01 4.89009410e-01 -7.54993498e-01 -7.72659302e-01
-3.29546511e-01 -9.64005053e-01 5.42385459e-01 1.61394566e-01
-8.45484257e-01 -9.80025887e-01 1.00260206e-01 1.19436824e+00
4.16715026e-01 -8.17945320e-03 1.70520365e+00 -1.16091192e+00
-3.34655017e-01 -5.64689159e-01 -1.31948441e-01 1.74327374e-01
-2.04836354e-01 -4.19234298e-02 -8.18009436e-01 3.51653367e-01
-3.45005184e-01 -3.52905959e-01 2.32230306e-01 -2.56876573e-02
3.66510034e-01 -7.14873135e-01 -2.06531271e-01 -1.20581836e-02
1.23017418e+00 4.77991432e-01 8.24910283e-01 7.80322313e-01
2.63710529e-01 6.88135028e-01 6.15007520e-01 1.03877030e-01
6.13506615e-01 2.62728870e-01 -1.58960372e-01 2.59886950e-01
1.54261142e-01 -3.99757653e-01 1.78222165e-01 -5.97990640e-02
2.72557288e-02 3.25949132e-01 -1.77163172e+00 4.21377867e-01
-1.95152676e+00 -1.46997297e+00 -1.90022945e-01 1.55121839e+00
6.72517836e-01 6.07857227e-01 -1.74282059e-01 8.71969163e-02
4.07758653e-02 -1.53705686e-01 -7.53186584e-01 -7.97657430e-01
-1.02246299e-01 4.11873221e-01 6.63766623e-01 9.47126985e-01
-5.50524235e-01 1.10439301e+00 6.56883478e+00 4.95371699e-01
-5.70928514e-01 1.03095397e-02 4.17203724e-01 7.73824975e-02
-8.82597506e-01 5.67660332e-01 -6.32266223e-01 3.50609422e-01
1.18766630e+00 -5.29182911e-01 5.12361944e-01 8.02165806e-01
5.24648786e-01 -2.90573597e-01 -1.03626966e+00 4.53415036e-01
-1.34182781e-01 -1.85151350e+00 1.11027345e-01 3.72019023e-01
5.84399343e-01 8.22760314e-02 -3.98182333e-01 7.00649261e-01
8.41522217e-01 -1.24275911e+00 6.79672480e-01 5.82507789e-01
4.90453422e-01 -3.54574263e-01 8.50160778e-01 9.28117275e-01
-6.12936974e-01 -8.54668319e-01 -7.17532337e-02 -6.32990360e-01
3.20935786e-01 -1.05875619e-01 -1.25260544e+00 3.96273732e-01
3.11310053e-01 5.26927054e-01 -7.90949047e-01 9.01437521e-01
-4.04040962e-01 7.14297116e-01 -3.01356107e-01 6.00769706e-02
3.36597055e-01 -1.37398079e-01 1.78543985e-01 1.12012923e+00
-2.31860951e-01 4.96700615e-01 -3.46592404e-02 8.67666006e-01
3.29458229e-02 1.20468885e-01 -7.74072349e-01 -8.39036107e-02
7.77780950e-01 7.52895474e-01 -5.53945713e-02 -7.54130602e-01
-6.14104629e-01 -1.73312888e-01 3.14084649e-01 3.97927523e-01
-5.69057941e-01 -1.50938392e-01 5.82642496e-01 6.23166382e-01
-4.20569867e-01 1.03325779e-02 -6.95196509e-01 -9.06292915e-01
1.61038220e-01 -1.43455410e+00 6.78318620e-01 -1.07209623e+00
-9.71412778e-01 4.13508601e-02 3.62176597e-01 -5.61806798e-01
-7.15876877e-01 -5.28832614e-01 -5.84079087e-01 1.00149632e+00
-1.53528965e+00 -1.17558968e+00 1.67067066e-01 3.46243858e-01
1.34824142e-01 -1.98560327e-01 6.45717740e-01 4.70791847e-01
-1.97765037e-01 3.33035260e-01 -4.38519448e-01 4.26222861e-01
6.65259361e-01 -9.45219457e-01 6.16483867e-01 1.12401021e+00
2.41904691e-01 1.18503284e+00 4.30279404e-01 -1.11153162e+00
-1.00890136e+00 -6.79717898e-01 1.35362577e+00 -1.00206685e+00
9.24918652e-01 1.24398069e-02 -9.88951862e-01 1.06547165e+00
1.04433164e-01 -6.61546886e-01 7.45077729e-01 3.54838163e-01
-5.95951319e-01 -6.47861287e-02 -1.25517774e+00 6.83365524e-01
1.06991637e+00 -9.15054619e-01 -1.24688053e+00 6.28704727e-01
3.74673218e-01 -3.57591778e-01 -9.98173773e-01 2.03612328e-01
6.51883721e-01 -8.03197205e-01 6.17896020e-01 -1.34060836e+00
3.78098518e-01 -4.11880091e-02 1.54196009e-01 -5.49350560e-01
-2.65652388e-01 -6.63374066e-01 1.40006259e-01 1.04068959e+00
1.08920467e+00 -1.09656537e+00 3.37453008e-01 1.84229672e+00
3.65514942e-02 -7.04191685e-01 -9.25988317e-01 -6.33101642e-01
4.09726501e-01 -1.18505692e+00 8.48133504e-01 1.01316118e+00
5.96683800e-01 2.74316102e-01 3.11710149e-01 3.14288348e-01
2.35815004e-01 1.32783018e-02 8.90773118e-01 -1.44079900e+00
-7.58288279e-02 -5.05133390e-01 -4.18588609e-01 -4.71900851e-01
4.01876986e-01 -1.16565919e+00 -8.37813318e-01 -2.31252408e+00
-6.49535144e-03 -5.29080570e-01 3.74269545e-01 1.00342333e+00
-3.18942964e-02 -4.80034471e-01 5.02025843e-01 2.91127801e-01
-2.70316571e-01 -5.58889627e-01 8.39118481e-01 -1.48802117e-01
2.24773888e-03 -2.31388837e-01 -1.42998827e+00 8.54889572e-01
7.19106019e-01 -3.64135802e-01 -4.48136538e-01 -5.49877226e-01
7.01480269e-01 -5.32294065e-02 3.66477221e-01 -8.51150811e-01
7.80569911e-01 -4.51599538e-01 5.67876920e-02 -1.21352591e-01
-1.56146228e-01 -7.58008301e-01 -2.50362717e-02 4.68796760e-01
-5.85942090e-01 8.10018927e-02 5.42750239e-01 -6.23993985e-02
-2.96264831e-02 -6.47614002e-01 3.31176698e-01 -2.67747939e-01
-6.63393855e-01 -2.53610939e-01 -5.53812802e-01 4.70869511e-01
9.00964916e-01 -3.77243340e-01 -1.01533926e+00 -5.42147875e-01
-2.94831604e-01 5.39694488e-01 3.58058423e-01 1.37500629e-01
1.33743718e-01 -6.40835166e-01 -7.68481791e-01 1.05195016e-01
1.85895935e-01 -2.35021725e-01 5.96125200e-02 7.14236200e-01
-5.75130701e-01 9.55152869e-01 2.12756202e-01 -6.08867295e-02
-1.03730798e+00 1.73248962e-01 3.72482568e-01 -5.47982812e-01
-5.41575551e-01 3.22100788e-01 -5.56415379e-01 -3.34293723e-01
7.44881555e-02 -7.29348183e-01 -3.43252331e-01 -2.02125639e-01
5.98286808e-01 3.34025055e-01 7.01293126e-02 -4.78284240e-01
-2.06739634e-01 3.71009588e-01 -3.89641523e-01 -2.44293585e-01
1.37456667e+00 4.25527185e-01 -2.48289958e-01 2.46884841e-02
4.87896442e-01 4.85466242e-01 -4.39449728e-01 -9.48572531e-03
6.99109018e-01 -5.96010268e-01 -2.92624354e-01 -1.45327830e+00
-3.67574602e-01 5.22725344e-01 4.03165855e-02 1.92203686e-01
4.57063496e-01 9.05723274e-02 7.50386953e-01 8.52360368e-01
2.58652836e-01 -1.21775901e+00 -6.57469392e-01 4.71163571e-01
6.74588382e-01 -1.19236279e+00 1.40572429e-01 -1.04216412e-01
-6.92232311e-01 9.73749220e-01 7.44793952e-01 4.10454094e-01
2.47815832e-01 4.78897721e-01 5.79466283e-01 -6.71154618e-01
-9.35642600e-01 1.80688109e-02 -1.27536450e-02 2.73216754e-01
3.98242027e-01 -2.65175074e-01 -5.56605577e-01 3.25228184e-01
-4.35180873e-01 3.75384241e-01 5.18844187e-01 8.21975768e-01
-4.52732921e-01 -1.31321692e+00 -4.04041559e-01 1.02464533e+00
-7.73906052e-01 -5.05936205e-01 -5.03897250e-01 1.18015361e+00
1.90459654e-01 1.17001879e+00 -1.95896342e-01 1.45732135e-01
6.42746210e-01 4.51433659e-01 4.65705432e-02 -6.38745844e-01
-7.54750788e-01 -4.98190701e-01 8.82994175e-01 -6.66204989e-01
-9.90208983e-02 -8.41035545e-01 -1.45946455e+00 -4.58174735e-01
-1.61849305e-01 3.26619178e-01 4.04737562e-01 1.43380487e+00
5.51755369e-01 -1.34471580e-02 -3.00253749e-01 5.86634040e-01
-9.31891978e-01 -1.04276395e+00 3.38542886e-04 6.78306103e-01
-4.34101522e-02 -2.81737715e-01 -9.03521013e-03 -1.98388591e-01] | [9.93082332611084, 9.135957717895508] |
79a5469a-41dd-452c-b86c-0c9376406a8b | joint-chinese-word-segmentation-and-part-of-1 | null | null | https://aclanthology.org/2020.coling-main.187 | https://aclanthology.org/2020.coling-main.187.pdf | Joint Chinese Word Segmentation and Part-of-speech Tagging via Multi-channel Attention of Character N-grams | Chinese word segmentation (CWS) and part-of-speech (POS) tagging are two fundamental tasks for Chinese language processing. Previous studies have demonstrated that jointly performing them can be an effective one-step solution to both tasks and this joint task can benefit from a good modeling of contextual features such as n-grams. However, their work on modeling such contextual features is limited to concatenating the features or their embeddings directly with the input embeddings without distinguishing whether the contextual features are important for the joint task in the specific context. Therefore, their models for the joint task could be misled by unimportant contextual information. In this paper, we propose a character-based neural model for the joint task enhanced by multi-channel attention of n-grams. In the attention module, n-gram features are categorized into different groups according to several criteria, and n-grams in each group are weighted and distinguished according to their importance for the joint task in the specific context. To categorize n-grams, we try two criteria in this study, i.e., n-gram frequency and length, so that n-grams having different capabilities of carrying contextual information are discriminatively learned by our proposed attention module. Experimental results on five benchmark datasets for CWS and POS tagging demonstrate that our approach outperforms strong baseline models and achieves state-of-the-art performance on all five datasets. | ['Fei Xia', 'Yan Song', 'Yuanhe Tian'] | 2020-12-01 | null | null | null | coling-2020-8 | ['chinese-word-segmentation'] | ['natural-language-processing'] | [ 2.05765665e-02 -3.31971377e-01 -2.61836946e-01 -4.27975982e-01
-8.08235407e-01 -5.33066273e-01 3.43310177e-01 2.84444064e-01
-8.62352967e-01 2.94398218e-01 4.45165217e-01 -6.00774646e-01
4.26196128e-01 -6.20492220e-01 -3.73268336e-01 -8.74208808e-01
5.03473543e-02 2.70101763e-02 5.13969481e-01 -1.02993473e-01
3.32298905e-01 1.58010155e-01 -1.18935478e+00 3.96616906e-01
1.13545990e+00 7.87261844e-01 8.09689760e-01 4.67658490e-01
-6.49528146e-01 2.20527738e-01 -6.11018598e-01 -1.62390903e-01
-1.84021946e-02 -4.45443213e-01 -6.30856693e-01 -3.67153347e-01
-2.19988912e-01 -2.55380832e-02 6.60682991e-02 1.20140553e+00
4.39024597e-01 3.03460568e-01 6.43839836e-01 -5.43605566e-01
-8.24282050e-01 8.12045753e-01 -3.95133346e-01 3.61795187e-01
7.42683783e-02 -8.37895498e-02 1.41059995e+00 -9.60225761e-01
2.28556897e-02 1.19335389e+00 5.23932219e-01 4.94908541e-01
-7.09940374e-01 -5.65723360e-01 7.13877618e-01 8.54376256e-02
-1.02438343e+00 9.31041688e-03 6.52185380e-01 -3.45611244e-01
1.05000842e+00 1.11041047e-01 3.50458741e-01 7.41538882e-01
2.52184838e-01 1.11587143e+00 8.49412799e-01 -7.38667488e-01
1.46823540e-01 2.79435590e-02 5.93361914e-01 2.32011139e-01
1.88712120e-01 -4.47533607e-01 -1.21638335e-01 5.93993217e-02
4.63791996e-01 1.23406604e-01 -7.59623498e-02 3.55491787e-01
-1.05954444e+00 9.73005414e-01 2.38607913e-01 8.91921163e-01
-2.36443773e-01 9.89921093e-02 3.71145993e-01 -1.34108365e-01
3.51609737e-01 3.67326230e-01 -9.13401067e-01 -1.94774345e-01
-5.27758658e-01 -1.23204194e-01 4.34468180e-01 9.52334166e-01
8.07650030e-01 -1.43937171e-01 -5.52443624e-01 9.02017951e-01
4.08441454e-01 2.28503391e-01 9.43380952e-01 -1.93969280e-01
8.06626141e-01 5.31693697e-01 9.05007198e-02 -6.02267027e-01
-3.08964670e-01 -2.02194378e-01 -6.02150202e-01 -4.63739693e-01
4.48967457e-01 -5.48077881e-01 -1.17289138e+00 1.74180055e+00
1.13125630e-01 1.95256352e-01 7.65013248e-02 8.45111907e-01
7.94104695e-01 9.60320354e-01 6.79042459e-01 -1.70040935e-01
1.71531057e+00 -1.24513865e+00 -9.39250112e-01 -7.04396009e-01
6.40103996e-01 -1.05489588e+00 1.22615182e+00 4.88359341e-03
-7.71096170e-01 -7.78519034e-01 -9.09358978e-01 -1.47762254e-01
-6.46666467e-01 4.28072900e-01 6.35947049e-01 6.58802450e-01
-6.66015863e-01 5.09460568e-01 -8.70427310e-01 -1.61875904e-01
4.15781215e-02 2.48940200e-01 3.13789956e-03 8.51704478e-02
-1.60287702e+00 6.48509920e-01 5.76678514e-01 2.39157259e-01
-3.60996902e-01 -1.32751212e-01 -8.68916094e-01 3.04852039e-01
2.33722374e-01 9.09273326e-02 1.09668124e+00 -8.21044087e-01
-1.21789944e+00 4.51202333e-01 -4.52161193e-01 -4.38484922e-02
-1.74661264e-01 -4.04489726e-01 -5.91891170e-01 3.79184112e-02
1.47765622e-01 5.03613114e-01 6.25737965e-01 -9.33417439e-01
-9.81253684e-01 -1.71262592e-01 8.80002752e-02 3.19010794e-01
-7.11947799e-01 4.77919012e-01 -7.63011634e-01 -8.19046855e-01
1.86242923e-01 -6.93436682e-01 -5.19899786e-01 -6.36211634e-01
-3.06488037e-01 -5.65767765e-01 6.93968117e-01 -8.30702364e-01
1.58347869e+00 -2.26926255e+00 -2.36510843e-01 -1.44831643e-01
-3.75317603e-01 8.37718129e-01 -4.51533377e-01 4.77190465e-01
1.40341401e-01 5.57065785e-01 -2.12269306e-01 -4.00873721e-01
-7.56624714e-02 3.63744944e-01 -2.79333070e-02 4.11657095e-02
6.70167983e-01 9.07338798e-01 -1.06434679e+00 -4.83951241e-01
-1.09394990e-01 4.33399320e-01 -2.73229092e-01 4.51512724e-01
-6.43302649e-02 3.03682327e-01 -7.45041609e-01 5.10631919e-01
6.78116262e-01 1.37571827e-01 4.31312650e-01 1.32459417e-01
-2.27005288e-01 7.07385302e-01 -1.00471663e+00 1.27738655e+00
-6.05217755e-01 1.68942928e-01 -1.09444566e-01 -1.11161339e+00
9.28654492e-01 3.61163110e-01 -1.08150356e-01 -6.28875613e-01
2.41078034e-01 3.41980964e-01 4.24494267e-01 -4.69931662e-01
6.43456995e-01 -1.82781428e-01 -3.68837208e-01 2.08819702e-01
9.91463363e-02 1.94970623e-01 2.30181426e-01 -1.09343268e-01
8.04197788e-01 1.01359859e-01 5.09604454e-01 -3.03801924e-01
7.03949869e-01 -2.98937559e-01 9.63402867e-01 6.52649939e-01
-3.60412419e-01 6.46883905e-01 5.62414229e-01 -1.61481276e-01
-6.86239362e-01 -5.80325544e-01 -7.88101479e-02 1.48125386e+00
2.20585153e-01 -3.19378555e-01 -7.18707919e-01 -1.15347624e+00
-3.81690264e-01 6.35332644e-01 -7.31222630e-01 -6.71610609e-03
-8.67915869e-01 -8.90448987e-01 4.35451448e-01 1.09457970e+00
2.12221295e-01 -1.45110750e+00 -4.46672082e-01 5.06882846e-01
-1.20052487e-01 -1.17776275e+00 -7.31595993e-01 7.08364487e-01
-7.82391489e-01 -8.67210507e-01 -8.35643470e-01 -1.44284010e+00
7.75167823e-01 6.02977216e-01 9.03396666e-01 3.84154081e-01
1.33760706e-01 -1.77529618e-01 -1.13949585e+00 -4.10378993e-01
-1.66771814e-01 2.15320662e-01 -2.23652139e-01 -1.90298911e-02
8.32424700e-01 -2.60736197e-01 -5.62336564e-01 2.48903498e-01
-1.02009499e+00 -3.03927034e-01 8.59125376e-01 9.80244935e-01
5.47714770e-01 -8.04712847e-02 4.90068108e-01 -9.13738370e-01
6.90308690e-01 -5.05604923e-01 -1.57476395e-01 3.30948204e-01
-1.77987680e-01 3.05111818e-02 9.60248947e-01 -5.60452998e-01
-1.18708634e+00 -6.44258037e-02 -5.02398729e-01 2.02670187e-01
-4.34552550e-01 8.70792925e-01 -6.27530158e-01 3.90629858e-01
-7.70161003e-02 3.91639203e-01 -4.85207677e-01 -7.66073704e-01
2.56531775e-01 9.03354824e-01 3.72496992e-02 -6.28466070e-01
4.52165753e-01 -1.52923148e-02 -4.57379043e-01 -7.17518151e-01
-8.75923276e-01 -8.54699016e-01 -8.73303175e-01 3.27183485e-01
1.21918356e+00 -8.25369120e-01 -1.02072604e-01 6.49002433e-01
-1.30284858e+00 -1.54813025e-02 2.92009652e-01 6.04778826e-01
2.43237600e-01 8.10028195e-01 -7.35359669e-01 -1.03381217e+00
-2.88903177e-01 -1.24835336e+00 9.51552629e-01 6.39043093e-01
1.01327606e-01 -1.04926240e+00 -1.34584576e-01 -4.67496701e-02
3.16308856e-01 -2.49083683e-01 1.04810476e+00 -1.23738229e+00
-1.57698795e-01 -3.01858306e-01 -2.69703031e-01 6.20862544e-01
2.05290854e-01 -6.10844605e-02 -9.95369852e-01 -6.19001016e-02
-1.44500256e-01 -7.24653602e-02 1.18178082e+00 4.49289680e-01
1.11536896e+00 -2.14812577e-01 -3.13616574e-01 2.91866541e-01
1.50174248e+00 6.93445623e-01 6.33859932e-01 2.70333499e-01
8.06924343e-01 5.59256136e-01 1.09673226e+00 2.05689251e-01
3.09344202e-01 4.46333081e-01 4.44165319e-01 3.10993809e-02
2.23428056e-01 -1.58443198e-01 4.91326004e-01 1.24049234e+00
-4.01134416e-02 -2.21498281e-01 -8.20200026e-01 8.81556749e-01
-1.79284418e+00 -4.83168513e-01 -3.26310992e-01 2.03233337e+00
9.17209208e-01 3.53102714e-01 -1.22750640e-01 -4.37971540e-02
9.88675892e-01 4.57294583e-01 -1.85584992e-01 -6.83640718e-01
-7.34765008e-02 4.32770312e-01 2.20094472e-01 3.63692194e-01
-1.34072828e+00 1.23229539e+00 5.10071468e+00 1.12598646e+00
-9.85038102e-01 1.52333304e-01 7.22632170e-01 5.00462055e-01
-4.44386095e-01 2.27547511e-01 -1.28154528e+00 7.43949413e-01
7.77435660e-01 3.12400252e-01 -5.99387400e-02 8.75924170e-01
2.44700257e-03 -2.31756285e-01 -7.05178618e-01 3.61261010e-01
-1.23648725e-01 -7.42255688e-01 -6.44717440e-02 -1.45502910e-01
7.44572818e-01 -7.11912513e-02 -1.26158729e-01 6.21519089e-01
1.87421262e-01 -8.12882721e-01 6.41866207e-01 3.91408391e-02
6.17463291e-01 -7.77525187e-01 1.24741316e+00 5.39631844e-01
-1.54964173e+00 9.02243052e-03 -6.56336963e-01 -2.49514535e-01
2.49513462e-01 6.04974806e-01 -3.54288906e-01 6.02838159e-01
5.79910398e-01 5.65247297e-01 -3.80758584e-01 9.23079073e-01
-6.82939410e-01 1.07611132e+00 -7.70122483e-02 -6.54702783e-01
7.64491498e-01 -2.92531133e-01 2.14104116e-01 1.73356140e+00
2.99166381e-01 2.51595587e-01 1.94890633e-01 5.07819474e-01
2.61369556e-01 3.99536848e-01 -8.32287036e-03 -2.10753486e-01
6.31994545e-01 1.27732444e+00 -8.61456454e-01 -3.98974121e-01
-7.40460575e-01 7.11863637e-01 4.41452801e-01 3.00449371e-01
-8.10156822e-01 -9.77906942e-01 7.28390098e-01 -2.90550143e-01
8.98375809e-01 -3.85907441e-01 -2.97938228e-01 -1.06484938e+00
2.43720487e-02 -4.33910847e-01 3.21771920e-01 -1.92390874e-01
-1.29005575e+00 8.62636387e-01 -3.01786572e-01 -1.20457637e+00
2.63571274e-02 -8.66778195e-01 -1.11021268e+00 1.33855152e+00
-1.96929204e+00 -1.18052542e+00 6.68859854e-02 1.47080615e-01
6.66552424e-01 7.01527297e-02 8.22015524e-01 1.95085362e-01
-7.23542690e-01 7.60914266e-01 1.97255567e-01 5.49364090e-01
6.22878790e-01 -1.29713595e+00 6.70057833e-01 1.22172737e+00
1.09234519e-01 9.86001730e-01 3.52289766e-01 -7.12480843e-01
-9.40060198e-01 -1.03059828e+00 1.48547757e+00 3.33472341e-02
5.84094226e-01 -4.69187975e-01 -1.12720823e+00 4.72827315e-01
1.77878186e-01 -1.53454747e-02 8.67820740e-01 3.34564149e-01
-1.99389338e-01 2.31479168e-01 -7.01483190e-01 4.57024962e-01
7.74671435e-01 -3.15208942e-01 -9.72499847e-01 -8.39651078e-02
9.90134537e-01 -7.60035068e-02 -4.44730639e-01 1.68644860e-01
2.59673744e-01 -7.58598983e-01 5.64352691e-01 -7.51060188e-01
5.30224979e-01 -3.60390455e-01 -1.74713418e-01 -1.36594534e+00
-5.77447832e-01 -3.30395341e-01 1.76546574e-01 1.59658802e+00
5.60525537e-01 -4.82892424e-01 4.44011211e-01 2.78809190e-01
-4.83405977e-01 -9.63592649e-01 -7.58808196e-01 -7.85340309e-01
3.92127275e-01 -5.66575289e-01 8.77590656e-01 8.15265834e-01
8.74348730e-02 3.25797766e-01 -3.08919400e-01 2.18509927e-01
-4.74384204e-02 9.87400487e-02 1.65992573e-01 -9.11916554e-01
-2.24503502e-01 -5.26522279e-01 -2.98857372e-02 -1.50688243e+00
1.29538402e-01 -5.53445041e-01 4.31729019e-01 -1.62774551e+00
2.49085546e-01 -6.09423757e-01 -6.51737809e-01 5.61678231e-01
-1.03005385e+00 -2.96807894e-03 1.99705258e-01 7.09822997e-02
-5.31678677e-01 4.85434026e-01 1.13362026e+00 6.91893771e-02
-2.89133519e-01 7.54559264e-02 -8.66432726e-01 7.32982874e-01
8.30869198e-01 -3.51332843e-01 1.23544864e-01 -7.19034135e-01
-1.35048693e-02 -2.23878756e-01 -2.89688855e-01 -6.91751778e-01
1.43253759e-01 -2.53861576e-01 2.53232896e-01 -5.53524196e-01
1.39827818e-01 -6.98911428e-01 -8.58731508e-01 4.16083783e-01
-1.50516465e-01 -3.50935869e-02 1.56381756e-01 5.18028557e-01
-4.44032639e-01 -7.38428712e-01 5.78290105e-01 -2.83099860e-01
-1.01373947e+00 1.94634154e-01 -5.86644650e-01 1.66221753e-01
7.52137899e-01 -2.09114879e-01 -6.94164336e-02 -6.15000576e-02
-5.69963813e-01 1.98607519e-01 -5.49086295e-02 6.14512086e-01
4.05965716e-01 -1.22308457e+00 -5.37177026e-01 2.94121623e-01
1.27911642e-01 -4.26872186e-02 2.10144103e-01 5.72960913e-01
-2.84936160e-01 6.13105178e-01 -3.18107307e-02 -1.74186081e-01
-1.10073709e+00 5.37345886e-01 -1.58304453e-01 -7.32474625e-01
-1.55547000e-02 1.11439180e+00 3.02347183e-01 -3.69865596e-01
1.73405528e-01 -8.27897608e-01 -9.04429138e-01 2.78146267e-01
7.01258779e-01 -6.23273887e-02 -2.42641196e-02 -7.79832482e-01
-4.49587703e-01 6.82437003e-01 -2.29191676e-01 1.90952018e-01
1.34157467e+00 -1.50235742e-01 -1.40446082e-01 3.40733021e-01
1.19023561e+00 2.64284074e-01 -1.18824685e+00 -3.33428830e-01
1.76106423e-01 -2.67112941e-01 -1.84024468e-01 -5.99115908e-01
-9.64772761e-01 1.27365625e+00 -5.28005708e-04 2.65698224e-01
1.08192766e+00 -3.73971811e-03 1.13630545e+00 -1.77482888e-02
1.53679878e-01 -1.31376970e+00 3.04750223e-02 1.09207892e+00
2.70776004e-01 -1.26243281e+00 -3.83857012e-01 -5.81327915e-01
-9.24909055e-01 1.14255798e+00 8.47698808e-01 -2.61514515e-01
7.79307306e-01 9.44937319e-02 2.32472241e-01 2.96154618e-01
-3.84906054e-01 -5.45015633e-01 4.67373639e-01 6.02108479e-01
7.49809265e-01 3.61432701e-01 -1.00438571e+00 1.34696162e+00
1.79378510e-01 -6.73912048e-01 2.12224051e-01 1.01283109e+00
-8.13093603e-01 -1.48220098e+00 -2.57889539e-01 4.13573325e-01
-7.81886518e-01 -4.89012897e-01 -1.22348636e-01 5.55948675e-01
3.67719412e-01 1.06211054e+00 8.66591632e-02 -4.46974725e-01
1.52428344e-01 1.31123498e-01 -9.91626754e-02 -9.98790801e-01
-7.90192902e-01 4.02622342e-01 -1.56944059e-02 -1.85451195e-01
-2.76318997e-01 -7.85384893e-01 -1.46000600e+00 3.83271575e-01
-6.93869889e-01 4.02850777e-01 6.37722909e-01 1.22014332e+00
2.39361078e-02 4.52352077e-01 7.02002823e-01 -7.68546224e-01
-5.93681753e-01 -1.13270056e+00 -6.69292390e-01 4.68511701e-01
-6.05335496e-02 -4.24339443e-01 -2.90989697e-01 -2.62087137e-01] | [10.065503120422363, 10.069745063781738] |
d09da08b-b9dd-4001-8f11-9d2e4003e43f | edge-computing-in-5g-cellular-networks-for | 2107.13767 | null | https://arxiv.org/abs/2107.13767v1 | https://arxiv.org/pdf/2107.13767v1.pdf | Edge computing in 5G cellular networks for real-time analysis of electrocardiography recorded with wearable textile sensors | Fifth-generation (5G) cellular networks promise higher data rates, lower latency, and large numbers of interconnected devices. Thereby, 5G will provide important steps towards unlocking the full potential of the Internet of Things (IoT). In this work, we propose a lightweight IoT platform for continuous vital sign analysis. Electrocardiography (ECG) is acquired via textile sensors and continuously sent from a smartphone to an edge device using cellular networks. The edge device applies a state-of-the art deep learning model for providing a binary end-to-end classification if a myocardial infarction is at hand. Using this infrastructure, experiments with four volunteers were conducted. We compare 3rd, 4th-, and 5th-generation cellular networks (release 15) with respect to transmission latency, data corruption, and duration of machine learning inference. The best performance is achieved using 5G showing an average transmission latency of 110ms and data corruption in 0.07% of ECG samples. Deep learning inference took approximately 170ms. In conclusion, 5G cellular networks in combination with edge devices are a suitable infrastructure for continuous vital sign analysis using deep learning models. Future 5G releases will introduce multi-access edge computing (MEC) as a paradigm for bringing edge devices nearer to mobile clients. This will decrease transmission latency and eventually enable automatic emergency alerting in near real-time. | ['Thomas M. Deserno', 'Valentin Purrucker', 'Arne Klingenberg', 'Nicolai Spicher'] | 2021-07-29 | null | null | null | null | ['electrocardiography-ecg'] | ['methodology'] | [-1.96768671e-01 -1.51437158e-02 -2.27825925e-01 -2.51275208e-02
-2.57983118e-01 -3.96048963e-01 -3.18640321e-01 -7.85342008e-02
-1.89149268e-02 8.78048539e-01 -2.13240966e-01 -1.00581229e+00
-1.82346120e-01 -9.14806008e-01 -2.99058765e-01 -6.18486702e-01
-5.70654035e-01 2.67893702e-01 -3.64055485e-01 4.15054977e-01
-4.08905059e-01 5.62714756e-01 -7.46480525e-01 4.30323124e-01
5.94593227e-01 1.94342744e+00 -3.28052908e-01 1.13074327e+00
5.32659769e-01 5.62196970e-01 -6.98804319e-01 -3.29453409e-01
2.13118091e-01 -2.76894331e-01 -6.13561332e-01 -3.58399808e-01
2.72339326e-03 -7.57648468e-01 -4.81567383e-01 2.55641490e-01
1.40557301e+00 -7.17901826e-01 1.09334707e-01 -1.17903006e+00
-1.58382311e-01 5.00162542e-01 2.49930602e-02 4.80816275e-01
1.40761346e-01 2.94311523e-01 2.95844883e-01 -2.42550343e-01
6.93070889e-02 3.27826887e-01 1.28543258e+00 3.67979914e-01
-9.06204820e-01 -4.89256918e-01 -7.10088730e-01 3.95177811e-01
-1.30283403e+00 -5.29687107e-01 6.93604171e-01 -2.05829978e-01
1.04983783e+00 4.72052872e-01 1.27553105e+00 1.10023272e+00
7.72428572e-01 2.13241965e-01 8.26029837e-01 -2.04932943e-01
3.43875587e-01 -1.47360206e-01 -1.84857532e-01 6.99394464e-01
6.09982729e-01 1.19027250e-01 -4.70811158e-01 -3.84506226e-01
6.70166552e-01 3.30372036e-01 -4.14815508e-02 3.97931308e-01
-1.31854653e+00 1.77452981e-01 2.75502414e-01 1.81679130e-01
-1.12744606e+00 9.43182349e-01 4.09104913e-01 3.83795410e-01
6.09447181e-01 -8.67498517e-02 -6.91148877e-01 -6.92988157e-01
-8.56656134e-01 -4.43426669e-01 9.11449790e-01 7.94772148e-01
1.21419489e-01 2.28096053e-01 -3.93882275e-01 1.92711279e-01
5.12754142e-01 9.50460553e-01 2.40124792e-01 -1.04608417e+00
2.72475839e-01 2.83149153e-01 -8.35122988e-02 -9.52516317e-01
-1.06161761e+00 -1.24696052e+00 -1.30031657e+00 -1.48195729e-01
2.38464966e-01 -1.06716430e+00 -2.40350753e-01 1.27050865e+00
1.06337227e-01 6.93649292e-01 -9.58801731e-02 4.94531780e-01
9.91180182e-01 5.09872213e-02 -1.28813624e-01 -1.84958056e-01
1.47261906e+00 -3.29874545e-01 -5.40797353e-01 5.85601367e-02
7.08919168e-01 -2.90574163e-01 6.07963979e-01 5.59806466e-01
-1.16862798e+00 -6.50622964e-01 -9.82442856e-01 5.08765876e-01
-1.71608161e-02 2.40597636e-01 7.76418269e-01 1.50446904e+00
-1.31883919e+00 7.17693150e-01 -9.15652871e-01 -4.13143963e-01
1.06593943e+00 7.08234012e-01 1.87581897e-01 1.59429833e-01
-1.16167486e+00 3.04960996e-01 -2.22302318e-01 7.81871304e-02
-4.38934654e-01 -7.89062023e-01 -1.54500857e-01 1.08726643e-01
-4.66822624e-01 -1.50157189e+00 7.44472861e-01 -2.72186339e-01
-1.56666982e+00 6.63721204e-01 2.32214838e-01 -7.60049045e-01
5.38226485e-01 -1.69287145e-01 -8.44804525e-01 3.85068893e-01
-1.69598177e-01 1.76443204e-01 6.17100537e-01 -3.69550288e-01
-3.13378274e-01 -6.56325758e-01 -1.69205576e-01 -5.89054108e-01
-4.18274254e-01 -4.41482723e-01 -9.67731476e-02 -4.32208568e-01
9.70437229e-02 -1.29950261e+00 1.88429788e-01 1.18698999e-02
-2.56965607e-01 2.46976957e-01 8.79553854e-01 -6.91016734e-01
1.21465158e+00 -1.80022931e+00 -5.73638618e-01 2.47034326e-01
6.66275918e-01 2.63214886e-01 3.13927591e-01 1.64829582e-01
2.95711815e-01 3.77297103e-01 4.23794180e-01 -2.20111027e-01
-4.02564347e-01 3.58108357e-02 3.01581442e-01 5.34469306e-01
-6.42170787e-01 1.56029701e+00 -7.23738730e-01 -5.26216961e-02
2.13092580e-01 6.38493419e-01 -4.33149278e-01 -2.48908520e-01
4.32856947e-01 8.21967304e-01 -3.69157374e-01 8.23669970e-01
4.09651726e-01 -6.73527420e-01 2.30594739e-01 -6.42514765e-01
2.96272427e-01 7.47868344e-02 -7.04919219e-01 1.46429718e+00
-8.26108932e-01 5.89743912e-01 -2.33411267e-01 -9.46652830e-01
7.14282751e-01 8.87734592e-01 1.08530164e+00 -7.80565500e-01
5.58920503e-01 2.11083159e-01 9.56688896e-02 -9.18775260e-01
-4.35642898e-01 -6.62433654e-02 -7.01737106e-02 6.00865245e-01
-2.55992770e-01 3.17239225e-01 -2.89941102e-01 -1.31510451e-01
1.41115046e+00 -3.74114007e-01 5.03746383e-02 -3.53287250e-01
-2.42068293e-03 -6.82201445e-01 3.54846001e-01 1.04372370e+00
-5.50553679e-01 2.75105834e-01 2.58048564e-01 -7.88270056e-01
-4.27132070e-01 -1.12624240e+00 1.26516536e-01 3.86135280e-01
-9.29648653e-02 -2.17350900e-01 -6.82783127e-01 -4.72199917e-01
2.34230701e-02 2.60941267e-01 -2.13729322e-01 -3.15468997e-01
-2.44595408e-01 -9.67036605e-01 1.14330649e+00 7.25571513e-01
1.04277194e+00 -6.44561529e-01 -1.04694891e+00 6.62806571e-01
-5.31134963e-01 -1.44279921e+00 -3.64409015e-02 -2.03130081e-01
-1.30697119e+00 -7.66561925e-01 -6.34977400e-01 -4.12034482e-01
1.56652480e-01 -1.30872086e-01 1.17717206e+00 1.89202800e-01
-1.96481705e-01 6.08845353e-01 -1.64578378e-01 -6.97030485e-01
-6.34663105e-02 2.81655610e-01 3.41234088e-01 2.06398875e-01
2.55719215e-01 -1.03995979e+00 -1.29061687e+00 1.29411384e-01
1.06318779e-01 -1.73688024e-01 4.70014095e-01 3.03703964e-01
3.40305537e-01 -5.70000894e-02 1.08635330e+00 -3.99984121e-01
5.92919409e-01 -6.56991065e-01 -2.87155747e-01 -6.43860698e-02
-9.15905356e-01 -5.54285109e-01 6.10600948e-01 -1.25442237e-01
-2.64089346e-01 -2.30803728e-01 -2.76323289e-01 2.96136774e-02
-2.51597781e-02 5.34274042e-01 -1.60685495e-01 -2.49333858e-01
6.12196922e-01 -1.27590731e-01 -3.96151282e-02 -1.63043618e-01
-1.95836984e-02 1.22361684e+00 4.89147574e-01 -3.07877302e-01
4.83376756e-02 7.42540002e-01 3.74720156e-01 -9.35607910e-01
-2.15685219e-01 -2.19412878e-01 -1.81831613e-01 -5.95877886e-01
6.61423087e-01 -9.44936335e-01 -1.40909278e+00 7.26063848e-01
-1.21926332e+00 -6.42083526e-01 -2.90837884e-01 7.32422233e-01
-2.48461962e-01 9.64690223e-02 -8.79229784e-01 -6.15265071e-01
-1.27555907e+00 -6.14475727e-01 9.20626223e-01 -7.17052892e-02
-3.68512869e-01 -9.92924213e-01 -3.44497025e-01 7.17496336e-01
1.07483816e+00 5.38133144e-01 5.16438246e-01 -1.65074691e-01
-4.40982997e-01 -5.23579001e-01 -1.19020030e-01 2.73586988e-01
1.65591866e-01 -6.78965390e-01 -1.05156398e+00 -4.31569844e-01
9.13699865e-02 4.24983591e-01 1.97182611e-01 1.03346658e+00
1.43673646e+00 -3.33985984e-01 -6.47032380e-01 9.53676522e-01
1.13302898e+00 3.30790490e-01 1.13603210e+00 -1.71036795e-01
6.08771265e-01 -4.52294558e-01 -2.08707422e-01 7.18689322e-01
2.35376909e-01 3.53557587e-01 5.14030874e-01 -3.37934971e-01
-2.97393054e-01 4.39072371e-01 -7.31598288e-02 7.60829091e-01
-4.91281748e-01 -6.45716548e-01 -8.84451926e-01 1.34761304e-01
-1.30297184e+00 -9.28876102e-01 -5.51975965e-01 2.17840528e+00
7.37151280e-02 3.43536556e-01 3.28310817e-01 5.60731769e-01
4.61937249e-01 -5.33963382e-01 -7.66543090e-01 -2.98989266e-01
1.69460967e-01 3.85453910e-01 7.40540624e-01 6.82397783e-02
-8.70221078e-01 1.79999813e-01 5.83425951e+00 9.88285094e-02
-1.68312359e+00 6.39976978e-01 1.18312418e+00 -1.86112136e-01
1.61529437e-01 -3.91453683e-01 -2.80017972e-01 1.01709557e+00
1.38420570e+00 3.51161212e-01 3.30655634e-01 5.48624814e-01
6.48525774e-01 1.46172285e-01 -8.45435321e-01 1.70286536e+00
-4.69973475e-01 -1.68608320e+00 -5.33130825e-01 4.21279013e-01
6.14581704e-01 5.66070795e-01 -6.49363026e-02 -1.24313906e-01
-5.96875072e-01 -6.27848387e-01 2.87271023e-01 9.89538431e-01
1.56645465e+00 -6.36816382e-01 1.25912619e+00 1.74654976e-01
-1.13428330e+00 -2.36385569e-01 2.87569880e-01 -4.50021803e-01
2.59018123e-01 1.32087851e+00 -7.20869660e-01 4.51726139e-01
8.24275672e-01 5.45023084e-01 -2.26173371e-01 1.06761754e+00
1.49862781e-01 1.00377905e+00 -5.35608351e-01 -1.08463898e-01
-4.72679853e-01 1.95160583e-01 3.59112591e-01 9.51950729e-01
7.46392310e-01 2.22635970e-01 6.29017502e-03 3.92090499e-01
-3.42268854e-01 -4.48994577e-01 -4.91543531e-01 2.82361627e-01
5.68464339e-01 1.06883085e+00 -8.70805681e-01 -5.05873084e-01
-3.64842504e-01 1.02736127e+00 -7.36043155e-01 3.46600950e-01
-9.79463756e-01 -7.79533505e-01 4.57113534e-01 7.35820115e-01
-1.31485596e-01 -2.93386936e-01 -1.03435981e+00 -8.70143473e-01
6.33687451e-02 -4.69803065e-01 1.37848631e-01 -6.64716959e-01
-1.08224463e+00 5.28126121e-01 -6.17981553e-01 -1.41501653e+00
-1.14693023e-01 -3.25001746e-01 -6.91783309e-01 5.42208374e-01
-1.00432062e+00 -1.11427224e+00 -7.88060248e-01 5.93914926e-01
-5.05321771e-02 -2.02083960e-01 1.17028272e+00 1.03902817e+00
-2.55646318e-01 7.55724251e-01 -1.36511728e-01 1.97133884e-01
-6.21118285e-02 -7.44758964e-01 6.55393124e-01 6.84134424e-01
-2.74043947e-01 2.56848872e-01 2.23557815e-01 -7.14346707e-01
-1.51534152e+00 -1.12833369e+00 8.03251624e-01 -3.19567978e-01
8.39831308e-02 -6.32671416e-02 -4.86922488e-02 3.94569069e-01
-9.75718051e-02 5.68195343e-01 1.01145756e+00 -1.09458692e-01
5.61851144e-01 -9.11347568e-01 -1.30575931e+00 3.67626697e-01
1.13642037e+00 -6.67865276e-01 4.74960536e-01 3.00484031e-01
2.37531021e-01 -1.62610292e-01 -1.20165908e+00 5.90093195e-01
1.09623039e+00 -7.94088185e-01 9.89049137e-01 -2.40657642e-01
-4.58849907e-01 3.78837995e-02 3.44968550e-02 -1.14547288e+00
-3.60771388e-01 -1.24544716e+00 -8.99336457e-01 8.41393590e-01
2.43826687e-01 -1.04113173e+00 1.17288160e+00 3.78947765e-01
-1.76825047e-01 -1.20030153e+00 -1.01815593e+00 -8.68102908e-01
-5.55759847e-01 -9.79167104e-01 7.02820957e-01 5.53412199e-01
6.79135770e-02 3.47442746e-01 -3.86922479e-01 1.23876985e-02
7.30982661e-01 -3.00032914e-01 5.48573732e-01 -1.58097732e+00
-2.79363722e-01 -1.15139298e-01 -8.06042373e-01 -9.86086428e-01
-6.40975058e-01 -8.86521041e-01 -7.68805146e-01 -1.81357360e+00
-3.03398073e-01 -7.38371193e-01 -5.33654153e-01 3.90217781e-01
3.77200484e-01 8.77637029e-01 -3.08187664e-01 -2.08478183e-01
-3.62442613e-01 -1.52819678e-01 8.50759983e-01 -1.37659356e-01
-2.86355227e-01 7.09664524e-01 -5.70373118e-01 6.02492750e-01
1.36360872e+00 -3.90542954e-01 -3.93961430e-01 -4.72444683e-01
5.31608939e-01 5.03151298e-01 7.22803414e-01 -1.63980806e+00
3.92653316e-01 6.04579568e-01 7.11671114e-01 -3.73025268e-01
2.82155275e-01 -8.82171631e-01 5.30343652e-01 1.07586205e+00
4.83977228e-01 -1.44528737e-03 1.29515186e-01 2.48269290e-01
6.16691232e-01 7.86109567e-01 4.98197317e-01 1.43115163e-01
-3.29630002e-02 6.83856487e-01 -7.51235247e-01 1.76235810e-02
8.34531248e-01 -5.19330204e-01 -8.79530683e-02 -6.24364316e-01
-1.11426258e+00 -2.67058551e-01 -1.76017046e-01 2.76917294e-02
7.69352376e-01 -1.27551782e+00 -6.87312603e-01 3.65618527e-01
-2.10728854e-01 -5.96936524e-01 5.02406955e-01 1.52466989e+00
-8.13935995e-01 4.26217526e-01 -2.90304989e-01 -8.64561439e-01
-1.15952909e+00 7.64340609e-02 8.19840491e-01 -3.35785337e-02
-6.41335428e-01 4.78992075e-01 -7.46860862e-01 3.29425991e-01
1.89099997e-01 -5.13592362e-01 3.04128736e-01 -1.12418607e-01
4.37399119e-01 1.12343049e+00 6.71307504e-01 1.28119230e-01
-6.46052063e-01 6.50954723e-01 7.53057301e-01 2.26774350e-01
1.05927932e+00 -4.99760389e-01 2.02272549e-01 2.86980476e-02
1.23891473e+00 -1.38435856e-01 -7.65213668e-01 3.64103943e-01
-4.00157750e-01 -3.23607260e-03 4.17922914e-01 -1.00691271e+00
-1.65424216e+00 8.68252754e-01 1.53331268e+00 4.47953969e-01
1.38655734e+00 -3.09606250e-02 1.64286244e+00 4.24382687e-01
7.90561974e-01 -7.25922287e-01 6.85857190e-03 4.32486236e-02
1.07575029e-01 -9.50233221e-01 -4.65864092e-02 -5.05814031e-02
1.06958590e-01 1.17499685e+00 -1.38312384e-01 2.01725304e-01
1.34978890e+00 2.86179930e-01 1.53121501e-01 -5.50822377e-01
-3.01029593e-01 -2.25957781e-02 -1.38040066e-01 9.05431509e-01
3.51725101e-01 4.53631073e-01 -1.78477719e-01 8.02974224e-01
-2.17279121e-02 7.57139742e-01 3.23399037e-01 5.60449123e-01
2.42952053e-02 -7.55273879e-01 -8.68463016e-04 9.79567826e-01
-1.12727332e+00 -1.52664855e-01 2.19770238e-01 -2.35296935e-02
6.05177581e-01 1.46418655e+00 3.55384825e-03 -6.19485855e-01
2.50095844e-01 -2.11149782e-01 4.02141303e-01 -1.22392140e-01
-6.07897103e-01 -1.93380296e-01 2.29390457e-01 -6.86251521e-01
-1.06949508e-02 -3.14353377e-01 -1.01364720e+00 -7.01740623e-01
-1.27877638e-01 -2.50886828e-01 8.85526717e-01 1.00116920e+00
1.20082426e+00 8.66603196e-01 4.70126718e-01 -5.07860363e-01
1.05376344e-03 -7.61045337e-01 -5.64255595e-01 -1.09171726e-01
3.68519604e-01 2.30193343e-02 -2.22799201e-02 2.00903580e-01] | [14.096531867980957, 3.2566916942596436] |
f3d9a4c4-c8f3-4e68-af7d-80e78219680e | license-plate-detection-and-recognition-in | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Sergio_Silva_License_Plate_Detection_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Sergio_Silva_License_Plate_Detection_ECCV_2018_paper.pdf | License Plate Detection and Recognition in Unconstrained Scenarios | Despite the large number of both commercial and academic methods for Automatic License Plate Recognition (ALPR), most existing approaches are focused on a specific license plate (LP) region (e.g. European, US, Brazilian, Taiwanese, etc.), and frequently explore datasets containing approximately frontal images. This work proposes a complete ALPR system focusing on unconstrained capture scenarios, where the LP might be considerably distorted due to oblique views. Our main contribution is the introduction of a novel Convolutional Neural Network (CNN) capable of detecting and rectifying multiple distorted license plates in a single image, which are fed to an Optical Character Recognition (OCR) method to obtain the final result. As an additional contribution, we also present manual annotations for a challenging set of LP images from different regions and acquisition conditions. Our experimental results indicate that the proposed method, without any parameter adaptation or fine tuning for a specific scenario, performs similarly to state-of-the-art commercial systems in traditional datasets, and outperforms both academic and commercial approaches in challenging datasets. | ['Sergio Montazzolli Silva', 'Claudio Rosito Jung'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['license-plate-recognition', 'license-plate-detection'] | ['computer-vision', 'computer-vision'] | [ 1.59140706e-01 -6.76212430e-01 1.11225002e-01 -3.53982151e-01
-8.65989029e-01 -9.14546430e-01 4.89408523e-01 -7.03024209e-01
-2.50436425e-01 5.47397375e-01 -3.15613061e-01 -2.21158728e-01
2.09527642e-01 -3.67077857e-01 -8.30517113e-01 -5.04462183e-01
5.54979324e-01 3.28416139e-01 4.52758193e-01 -2.46186689e-01
7.12555349e-01 1.13874149e+00 -1.40639126e+00 1.16386004e-02
9.25765097e-01 9.16163266e-01 -2.56047933e-04 5.42324305e-01
1.18724875e-01 4.70165074e-01 -8.67127180e-01 -9.29296017e-01
5.97973108e-01 1.75472513e-01 -2.06438079e-01 7.35151112e-01
8.45715344e-01 -5.39239407e-01 -6.68851852e-01 1.12143421e+00
4.09251362e-01 1.20517820e-01 6.80900216e-01 -9.96549428e-01
-1.13198602e+00 3.52764502e-02 -5.75692594e-01 1.08103707e-01
3.36600803e-02 2.21684694e-01 4.60569739e-01 -1.23240447e+00
5.41096032e-01 5.57569981e-01 9.58491027e-01 3.82044435e-01
-7.32250929e-01 -8.01752448e-01 1.13874665e-02 7.17887804e-02
-1.72946572e+00 -5.97927809e-01 9.17925715e-01 -3.99790585e-01
7.42165148e-01 6.37773871e-02 2.83695996e-01 9.45259869e-01
1.72771681e-02 9.41076458e-01 9.78031099e-01 -2.38958493e-01
-5.42090181e-03 2.50166953e-01 9.22410339e-02 3.71814191e-01
4.11131203e-01 -3.01026881e-01 -6.90605193e-02 3.81428599e-01
9.07506287e-01 2.10299864e-01 -4.42636102e-01 -1.79161653e-01
-8.79539430e-01 4.16541398e-01 -7.79349916e-03 2.16586456e-01
-4.21041772e-02 -3.63017678e-01 8.07852969e-02 -5.81700914e-02
2.49933988e-01 3.95380437e-01 -2.33499020e-01 -1.09807171e-01
-1.43606770e+00 2.73422778e-01 5.78495026e-01 1.33195508e+00
7.56406724e-01 2.97206581e-01 2.41101580e-03 1.18826318e+00
3.93884718e-01 6.55159712e-01 5.38752317e-01 -5.78438044e-01
8.69250059e-01 4.87282991e-01 4.22331750e-01 -1.04522169e+00
-2.10681915e-01 -5.67387700e-01 -7.23980188e-01 2.02465206e-01
4.18519109e-01 -9.23354626e-02 -9.28347349e-01 7.02189803e-01
-4.31364864e-01 1.29597977e-01 1.57105058e-01 1.12140775e+00
7.43178129e-01 7.38543272e-01 -7.03793466e-01 1.37816727e-01
1.14811945e+00 -1.33524644e+00 -6.24305367e-01 -4.72788721e-01
1.22063957e-01 -1.08014333e+00 8.42398226e-01 6.11774683e-01
-9.10577893e-01 -5.69608390e-01 -1.35798049e+00 -1.57815427e-01
-5.73083401e-01 1.11866844e+00 -4.97729005e-03 7.78244913e-01
-9.53394711e-01 2.59078175e-01 -2.70032406e-01 -2.79886186e-01
4.40417349e-01 2.89576679e-01 -5.65504313e-01 -3.92101556e-01
-7.60699630e-01 7.50520170e-01 6.33139834e-02 6.77817285e-01
-7.53962219e-01 -4.21323001e-01 -5.90829968e-01 1.79126374e-02
4.19455379e-01 2.53382325e-01 1.06109357e+00 -1.10534191e+00
-1.81977272e+00 8.25876653e-01 2.97798932e-01 -1.77380264e-01
7.72380054e-01 -3.31630230e-01 -8.80703449e-01 -4.20137681e-02
-2.73121059e-01 3.53850096e-01 9.44299757e-01 -1.34330499e+00
-5.10856032e-01 -3.02891999e-01 -3.00985835e-02 2.69841254e-01
-7.86204189e-02 4.35198545e-01 -1.09732044e+00 -7.12253273e-01
-6.88955337e-02 -9.79599893e-01 2.02280194e-01 -1.12922058e-01
-6.85104907e-01 1.22161910e-01 9.75562274e-01 -9.14122224e-01
8.20975006e-01 -2.37850475e+00 -4.18843299e-01 1.95458472e-01
-1.16798677e-01 9.84595239e-01 -1.41767841e-02 1.94927052e-01
-4.18902747e-02 5.16437069e-02 -3.89443308e-01 -5.28952539e-01
1.24846540e-01 -1.61648571e-01 -3.15528184e-01 8.58079791e-01
3.58461142e-01 6.54247463e-01 -6.19774684e-02 -3.02121162e-01
4.47866678e-01 4.17990923e-01 -1.17850684e-01 1.93631634e-01
2.34094143e-01 -6.06481023e-02 -1.27544388e-01 1.11495101e+00
1.31424391e+00 1.13236330e-01 -2.02217370e-01 -2.33906180e-01
-4.87331122e-01 -4.48308915e-01 -1.39514744e+00 1.11606646e+00
-2.23432630e-01 1.27949417e+00 2.30424002e-01 -7.23715961e-01
1.49900699e+00 6.16539791e-02 7.26970984e-03 -6.54413402e-01
6.98548257e-02 6.18197024e-01 -2.44416013e-01 -5.61666608e-01
9.85960603e-01 3.73400420e-01 4.65068482e-02 -9.65285115e-03
9.00739059e-03 9.87545997e-02 3.11166823e-01 -2.43278176e-01
5.52910626e-01 5.31216301e-02 -1.45435929e-01 -2.86447653e-03
1.05675840e+00 -8.80973637e-02 5.71466208e-01 6.33954227e-01
-4.91476625e-01 1.23106050e+00 2.57019639e-01 -5.61493337e-01
-1.34633613e+00 -7.96335280e-01 -1.50083423e-01 5.93196213e-01
4.90589321e-01 2.73153365e-01 -7.32102394e-01 -4.15340811e-01
-1.29801735e-01 5.28328478e-01 -2.06057370e-01 2.92487770e-01
-8.31800818e-01 -6.97513163e-01 1.05804431e+00 6.52670145e-01
1.11240041e+00 -8.34810555e-01 3.99868563e-02 -2.02501621e-02
-3.71217318e-02 -1.53491068e+00 -7.50963211e-01 -3.46180558e-01
-6.35730505e-01 -1.07959306e+00 -1.17099237e+00 -9.99971807e-01
6.33482933e-01 5.31543911e-01 8.33101094e-01 -2.19683185e-01
3.78952511e-02 1.05846971e-01 -2.28300840e-01 -3.24624270e-01
-3.22002083e-01 2.61352025e-02 -8.32263380e-02 6.43322468e-01
4.43745494e-01 -9.16425884e-02 -4.26357299e-01 8.35259855e-01
-1.06861484e+00 -3.35050941e-01 1.02702916e+00 4.67106849e-01
3.90945613e-01 5.66601418e-02 2.64277518e-01 -5.59020102e-01
7.72945940e-01 -1.43486008e-01 -1.19493008e+00 4.95985150e-01
-3.85044724e-01 -4.68652308e-01 8.67476702e-01 -3.79780173e-01
-1.22343194e+00 2.07167238e-01 -2.85201192e-01 -6.92231774e-01
-7.53430605e-01 2.41745099e-01 -3.51850033e-01 -3.86388063e-01
4.89156008e-01 8.68894815e-01 -9.41366404e-02 -5.77887058e-01
-1.84346706e-01 1.25153255e+00 8.77475917e-01 -2.12994501e-01
9.56898272e-01 2.89962262e-01 -3.70466262e-01 -1.06659770e+00
-3.28615904e-01 -6.28781855e-01 -8.45276237e-01 -6.02994263e-01
6.36222780e-01 -9.09828484e-01 -6.09969795e-01 1.27302074e+00
-1.10663414e+00 9.79925767e-02 2.45640904e-01 5.49871743e-01
-3.00360441e-01 5.88250518e-01 -4.80798662e-01 -7.01886415e-01
2.37386096e-02 -1.49287307e+00 1.26332426e+00 5.99727571e-01
6.75255120e-01 -7.20003903e-01 7.60710686e-02 8.40573251e-01
5.70877969e-01 -1.93001345e-01 1.94532409e-01 -9.79550183e-01
-7.82304764e-01 -7.98693538e-01 -5.66752195e-01 7.24503160e-01
-1.65637210e-01 5.83793342e-01 -1.02005029e+00 -8.53392854e-02
-4.09404427e-01 -2.17337042e-01 7.86731541e-01 2.17096657e-01
1.01229906e+00 -1.32906973e-01 1.72991548e-02 7.73008704e-01
1.69526803e+00 3.36865187e-01 1.25610995e+00 5.82857490e-01
7.76743591e-01 2.31725678e-01 5.98641992e-01 1.79339856e-01
1.37610152e-01 8.11074197e-01 3.68585855e-01 -1.12247095e-01
5.80429956e-02 7.62763172e-02 6.78556144e-01 5.56365967e-01
-5.41496217e-01 -5.64311206e-01 -9.47170019e-01 4.48079795e-01
-1.49018538e+00 -8.61905932e-01 -1.12679854e-01 2.12565470e+00
2.59055763e-01 -7.60370791e-02 2.31039189e-02 1.03293941e-03
1.12745392e+00 2.08459079e-01 -5.12758791e-01 -3.71577322e-01
-4.34854269e-01 -3.89828771e-01 1.02153838e+00 6.38252199e-02
-1.39605069e+00 8.84064615e-01 6.31183338e+00 8.96578014e-01
-1.45701253e+00 -2.64112234e-01 4.91259038e-01 2.84201622e-01
3.22893620e-01 -6.70018613e-01 -1.04034507e+00 4.60005611e-01
7.72725105e-01 6.52129874e-02 4.42244381e-01 8.76414716e-01
3.80285680e-02 2.38490254e-01 -7.37679124e-01 1.23432100e+00
6.65245950e-01 -1.35006166e+00 -1.19367726e-01 1.62199870e-01
7.90230870e-01 1.99348181e-01 3.54379117e-01 2.52081245e-01
-2.59778827e-01 -1.05016935e+00 8.49044442e-01 7.41267204e-01
8.25649977e-01 -9.18729484e-01 1.32512295e+00 3.11249256e-01
-8.68648708e-01 -1.83995888e-01 -6.50442362e-01 4.87431526e-01
-5.95160127e-02 2.03977585e-01 -4.88807350e-01 5.26218712e-01
4.48489845e-01 8.77071857e-01 -9.77708578e-01 1.32223642e+00
-6.77360804e-04 5.46466410e-01 -5.85329570e-02 6.57448396e-02
4.18808341e-01 -4.46214795e-01 5.27367711e-01 1.49708796e+00
5.43299317e-01 -2.37777099e-01 -2.94938922e-01 1.02535141e+00
-3.48052323e-01 1.08098708e-01 -5.27914524e-01 -4.61231805e-02
2.55256057e-01 1.30405509e+00 -5.77871859e-01 -3.90346497e-02
-7.65356004e-01 9.97939944e-01 -1.20979488e-01 5.28975070e-01
-1.09207690e+00 -6.92479908e-01 4.87237871e-01 1.35625139e-01
7.22430706e-01 -2.89180309e-01 3.47590409e-02 -1.57669330e+00
4.43442643e-01 -7.87006915e-01 -8.04515108e-02 -9.83210385e-01
-1.20962036e+00 5.43448627e-01 -3.26129586e-01 -1.90172386e+00
1.42205372e-01 -1.33770514e+00 -7.35333443e-01 8.17344129e-01
-1.85478020e+00 -1.25628650e+00 -4.00297850e-01 7.36168683e-01
9.49180245e-01 -8.16609383e-01 2.58109689e-01 6.60775542e-01
-1.01557302e+00 9.77392137e-01 8.51809680e-01 7.11316049e-01
9.30041432e-01 -9.53294635e-01 3.73396873e-01 1.35180783e+00
-1.84245080e-01 4.56437021e-01 3.17870289e-01 -3.55366886e-01
-1.34088945e+00 -1.37664413e+00 5.71453393e-01 -1.26502499e-01
3.84633541e-01 -3.48528266e-01 -7.83225715e-01 8.00994635e-01
1.46690860e-01 2.47102648e-01 5.56788325e-01 -4.28773344e-01
-3.87399912e-01 -4.42460418e-01 -1.24619877e+00 5.39446354e-01
4.32164133e-01 -3.20377380e-01 -4.37512100e-01 1.01340987e-01
1.22076124e-01 -5.18620789e-01 -7.63384283e-01 5.38621610e-03
6.40856922e-01 -1.13312876e+00 8.71583998e-01 -1.07754663e-01
5.97753763e-01 -8.22939932e-01 -1.35795742e-01 -1.08485568e+00
-1.70522258e-01 -1.86921328e-01 1.52672485e-01 1.34089351e+00
4.94430393e-01 -5.87964833e-01 6.80425882e-01 7.54745066e-01
-5.70034921e-01 -2.76731998e-01 -7.72195280e-01 -8.81361306e-01
-5.34966290e-02 -3.81019950e-01 5.44527233e-01 6.80310488e-01
-7.10568309e-01 -2.87006021e-01 -7.99300790e-01 5.71781278e-01
5.14621496e-01 1.70642510e-01 8.45082104e-01 -1.09548736e+00
5.87098524e-02 -5.05681217e-01 -8.82382333e-01 -1.08778465e+00
5.07895872e-02 -3.87668997e-01 3.50441366e-01 -1.20259500e+00
-5.56523278e-02 9.03755352e-02 -8.40070993e-02 8.82258490e-02
2.09333733e-01 8.42639863e-01 3.47655356e-01 5.57252407e-01
-7.70868361e-01 2.56622732e-01 1.00437319e+00 -5.18341839e-01
5.58878034e-02 3.51232797e-01 -5.54351449e-01 7.37600505e-01
6.99565172e-01 -7.02356175e-02 5.00747822e-02 -5.70798337e-01
1.76260434e-02 -2.19506696e-01 3.00156415e-01 -1.23310733e+00
5.66235125e-01 7.95897916e-02 6.80426776e-01 -8.83665621e-01
1.67894915e-01 -9.38414156e-01 2.65090466e-02 -2.25080177e-01
-6.94352984e-02 6.19420260e-02 2.71989107e-01 4.16060746e-01
-6.73433423e-01 -5.45160115e-01 9.29487884e-01 9.09206197e-02
-1.02128458e+00 1.89705372e-01 -6.13594651e-01 -1.14987597e-01
1.01953828e+00 -8.16479802e-01 -4.86672461e-01 -1.77880421e-01
-2.58994073e-01 -1.54800072e-01 6.48714483e-01 5.25444925e-01
7.90986717e-01 -1.26883078e+00 -8.62228692e-01 3.73515934e-01
3.29522550e-01 -1.64937284e-02 4.13817167e-01 5.94105721e-01
-1.31989193e+00 7.77187645e-01 -5.71649253e-01 -4.85819161e-01
-1.14076340e+00 3.74883384e-01 7.62218595e-01 -6.68974221e-02
-4.84854549e-01 4.20510858e-01 -1.36436298e-01 -6.73270822e-01
1.89365640e-01 -1.04450352e-01 -5.30813813e-01 -9.19671133e-02
5.99470317e-01 6.00876927e-01 5.21748304e-01 -1.10052502e+00
-3.80843222e-01 9.82644916e-01 -2.32276261e-01 1.00854762e-01
1.37240279e+00 -9.32410136e-02 2.08557531e-01 3.57697383e-02
1.10251069e+00 2.38392234e-01 -1.54440594e+00 -6.14917204e-02
-3.91263813e-01 -8.21955919e-01 -2.57484436e-01 -7.52346456e-01
-1.42209482e+00 8.38016987e-01 5.47799587e-01 -1.22503974e-01
1.09447110e+00 -4.92469072e-01 6.10436857e-01 7.79384851e-01
1.53240755e-01 -1.46731651e+00 -2.56844193e-01 6.41267836e-01
9.85284150e-01 -1.24501455e+00 -1.14535630e-01 -2.05579370e-01
-8.40689719e-01 1.73856258e+00 6.25312269e-01 -4.40447539e-01
2.37892210e-01 2.93829262e-01 2.35935941e-01 2.06556320e-01
-5.61970845e-02 2.26735398e-01 3.10013115e-01 6.66812003e-01
2.96396762e-01 -2.23628730e-01 1.45118192e-01 6.38411939e-01
1.61406890e-01 8.11195746e-02 1.17658281e+00 7.91346073e-01
-2.45787889e-01 -6.57166123e-01 -7.88342118e-01 3.14278692e-01
-5.50650299e-01 7.16959536e-02 -5.52513540e-01 1.17040813e+00
2.67352998e-01 6.96655035e-01 3.39803576e-01 -1.81469142e-01
5.40975451e-01 6.50702193e-02 8.98286179e-02 -7.69778490e-02
-4.00579214e-01 4.25338186e-02 -8.05133805e-02 -2.55635113e-01
-4.22589451e-01 -7.69851923e-01 -5.14405668e-01 -3.83658141e-01
-5.06992877e-01 -2.43703425e-01 7.11722434e-01 8.30348015e-01
2.67658025e-01 3.53179932e-01 7.11235106e-01 -1.04653573e+00
-4.71299887e-01 -9.95213687e-01 -1.06673181e+00 2.60141015e-01
2.88697183e-01 -4.87108439e-01 -3.08649421e-01 3.53594601e-01] | [9.851036071777344, -4.920660018920898] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.