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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d7d6f83d-880d-4ea6-9930-2a1422f1eae3 | zusammenqa-data-augmentation-with-specialized | 2205.14981 | null | https://arxiv.org/abs/2205.14981v1 | https://arxiv.org/pdf/2205.14981v1.pdf | ZusammenQA: Data Augmentation with Specialized Models for Cross-lingual Open-retrieval Question Answering System | This paper introduces our proposed system for the MIA Shared Task on Cross-lingual Open-retrieval Question Answering (COQA). In this challenging scenario, given an input question the system has to gather evidence documents from a multilingual pool and generate from them an answer in the language of the question. We devised several approaches combining different model variants for three main components: Data Augmentation, Passage Retrieval, and Answer Generation. For passage retrieval, we evaluated the monolingual BM25 ranker against the ensemble of re-rankers based on multilingual pretrained language models (PLMs) and also variants of the shared task baseline, re-training it from scratch using a recently introduced contrastive loss that maintains a strong gradient signal throughout training by means of mixed negative samples. For answer generation, we focused on language- and domain-specialization by means of continued language model (LM) pretraining of existing multilingual encoders. Additionally, for both passage retrieval and answer generation, we augmented the training data provided by the task organizers with automatically generated question-answer pairs created from Wikipedia passages to mitigate the issue of data scarcity, particularly for the low-resource languages for which no training data were provided. Our results show that language- and domain-specialization as well as data augmentation help, especially for low-resource languages. | ['Simone Paolo Ponzetto', 'Goran Glavaš', 'Marco Bombieri', 'Sotaro Takeshita', 'Tornike Tsereteli', 'Robert Litschko', 'Tommaso Green', 'Chia-Chien Hung'] | 2022-05-30 | null | https://aclanthology.org/2022.mia-1.8 | https://aclanthology.org/2022.mia-1.8.pdf | naacl-mia-2022-7 | ['passage-retrieval'] | ['natural-language-processing'] | [-4.80222479e-02 7.51418024e-02 1.42049477e-01 -2.32942164e-01
-1.95739245e+00 -9.43027198e-01 8.10394526e-01 4.42341447e-01
-1.07161570e+00 1.01895738e+00 3.73638749e-01 -4.27132279e-01
-5.64488843e-02 -5.47078252e-01 -8.49940240e-01 -2.06745237e-01
2.12190315e-01 9.31442022e-01 2.94688284e-01 -7.82561243e-01
5.18865101e-02 1.00950152e-02 -1.24173892e+00 7.67590761e-01
1.29747438e+00 7.21585572e-01 3.77495050e-01 9.00589049e-01
-4.65089023e-01 1.07488644e+00 -6.25048637e-01 -6.83828056e-01
1.49307027e-01 -4.75467414e-01 -1.29463661e+00 -5.00453591e-01
7.44791031e-01 -1.69089109e-01 -6.21458218e-02 5.89663148e-01
6.08127415e-01 3.17688316e-01 6.45989895e-01 -6.96317315e-01
-9.50135052e-01 6.98857009e-01 -4.06137146e-02 2.75602311e-01
7.28354573e-01 3.56927104e-02 1.38298559e+00 -1.11274564e+00
9.38704729e-01 1.26078331e+00 3.78217012e-01 7.27946103e-01
-1.28107309e+00 -3.26756775e-01 -5.04840128e-02 2.67106891e-01
-1.27822900e+00 -5.56957185e-01 5.79875708e-01 -1.62149608e-01
1.32477617e+00 2.47457579e-01 -1.84818879e-01 1.01655734e+00
-1.07940935e-01 9.96493697e-01 8.24961841e-01 -9.39924777e-01
-8.46997797e-02 5.72300971e-01 2.38579676e-01 5.22482753e-01
-3.48495364e-01 -2.99262762e-01 -3.40809077e-01 -3.81505579e-01
8.24247375e-02 -6.94458008e-01 -4.66296643e-01 -1.00663662e-01
-1.17689526e+00 9.22505319e-01 4.22960162e-01 5.58148861e-01
-2.87835658e-01 -2.13155940e-01 4.74802405e-01 8.87514830e-01
5.76276481e-01 9.05847907e-01 -8.41825008e-01 3.55766863e-01
-9.00309265e-01 4.84402865e-01 1.01898372e+00 7.07193017e-01
9.76520836e-01 -3.84483129e-01 -7.73102164e-01 1.24612582e+00
2.64065176e-01 5.76006234e-01 7.62877703e-01 -6.36332095e-01
1.01571500e+00 5.74705422e-01 2.46520683e-01 -6.72877014e-01
-1.57102376e-01 -3.13787371e-01 -4.35574979e-01 -4.65481460e-01
5.09575963e-01 -2.30476871e-01 -6.20883465e-01 2.05232739e+00
1.68589085e-01 -3.06973845e-01 5.64004421e-01 7.62001455e-01
1.06310916e+00 8.42268288e-01 1.95765391e-01 8.39596689e-02
1.38483775e+00 -1.39475119e+00 -6.56115532e-01 -2.25222558e-01
9.26434577e-01 -9.89919901e-01 1.52906728e+00 -1.10727986e-02
-1.09338486e+00 -8.92627597e-01 -9.24680471e-01 -6.39484704e-01
-6.18654668e-01 3.96140277e-01 -9.03602131e-03 2.01850653e-01
-1.28746784e+00 4.81280647e-02 -1.14512913e-01 -3.40940088e-01
-3.12166035e-01 7.59201944e-02 -3.48468184e-01 -4.32279319e-01
-1.76982927e+00 1.14051270e+00 2.44388267e-01 -6.58331497e-04
-8.56486380e-01 -6.27587438e-01 -8.90570402e-01 -9.09880027e-02
2.06542626e-01 -9.48566616e-01 1.22527051e+00 -9.77036536e-01
-1.40752053e+00 1.21992683e+00 -1.18924044e-01 -6.69372380e-01
3.36165249e-01 -5.19635677e-01 -4.84275520e-01 2.89544016e-01
3.65482718e-01 7.82536566e-01 7.27065146e-01 -1.15480721e+00
-3.37703824e-01 -1.75818443e-01 4.45168376e-01 3.84838909e-01
-2.47347951e-01 1.26257077e-01 -5.09174526e-01 -4.39680785e-01
-4.37990665e-01 -6.94614887e-01 -2.17074350e-01 -6.14368141e-01
-1.41246244e-01 -5.62110841e-01 2.23604560e-01 -1.21297252e+00
1.28719759e+00 -1.81798005e+00 2.35734209e-01 5.90799563e-02
-2.23401442e-01 3.86038572e-01 -8.73969138e-01 6.90899730e-01
1.48529574e-01 3.57405245e-02 -2.46545002e-01 -4.14010584e-01
-1.56232417e-02 1.70498751e-02 -5.87165475e-01 -5.31783625e-02
5.23670733e-01 9.81030107e-01 -1.13224280e+00 -5.58335066e-01
-3.02643001e-01 3.05815279e-01 -6.62039042e-01 5.40673494e-01
-7.62701511e-01 6.02708101e-01 -3.73018444e-01 1.97456583e-01
3.12876403e-01 -1.27836158e-02 -2.40281783e-02 8.47070739e-02
2.65391488e-02 8.32554877e-01 -9.20661509e-01 2.13364959e+00
-1.07774186e+00 3.33677143e-01 1.07436292e-01 -6.21211052e-01
9.50071990e-01 6.12063110e-01 4.41137888e-02 -1.05883265e+00
-1.60149902e-01 7.28275955e-01 -1.43665776e-01 -6.80444956e-01
9.16944444e-01 -3.41756940e-02 -1.43487751e-01 6.48634732e-01
5.40222287e-01 -2.41066575e-01 5.76166868e-01 4.54076678e-01
1.12883198e+00 2.34948292e-01 -3.43343057e-02 -2.97759563e-01
1.40198112e+00 1.69602767e-01 -1.31287113e-01 8.64007711e-01
1.75950199e-01 5.98456562e-01 -1.12852892e-02 1.80591375e-01
-8.47059011e-01 -1.05277836e+00 -8.13097283e-02 1.50150800e+00
-3.59584153e-01 -4.39094663e-01 -7.08664298e-01 -8.42501223e-01
-1.20745793e-01 7.25158215e-01 -2.54231542e-01 -1.40749022e-01
-8.80696654e-01 -4.14503396e-01 6.87738538e-01 3.88694294e-02
2.64290482e-01 -1.36527956e+00 -6.98414445e-02 4.22580183e-01
-7.49414802e-01 -1.15963161e+00 -5.27281940e-01 7.03791380e-02
-6.13228321e-01 -9.95481789e-01 -1.17372143e+00 -1.17677414e+00
4.02791113e-01 -1.19714610e-01 1.78913236e+00 1.29468009e-01
1.30320698e-01 7.42290080e-01 -4.51923102e-01 -1.55468211e-01
-7.44329810e-01 5.37366688e-01 -1.98430628e-01 2.98146140e-02
3.73135686e-01 -1.21810481e-01 -4.49863613e-01 -1.68433692e-02
-9.77448106e-01 -3.92190158e-01 5.13176739e-01 9.02669966e-01
5.83386779e-01 -6.38895810e-01 1.23719621e+00 -8.10134530e-01
1.25365841e+00 -5.97069085e-01 -4.53004301e-01 8.19957674e-01
-9.45681408e-02 5.00640035e-01 7.41435707e-01 -3.47305775e-01
-1.26477671e+00 -4.61696357e-01 -5.05756676e-01 -2.57076677e-02
1.16484374e-01 6.62361443e-01 -6.88838586e-02 2.92329013e-01
1.05423772e+00 2.88614593e-02 -2.50042379e-01 -7.44582355e-01
8.35291386e-01 7.31659830e-01 4.09584343e-01 -9.77273345e-01
6.22608960e-01 -1.20528124e-01 -6.93971455e-01 -6.96269453e-01
-1.03471243e+00 -8.04491937e-01 -5.68037808e-01 1.42692300e-02
9.25492048e-01 -1.00299656e+00 -1.31979719e-01 1.15135513e-01
-1.43914819e+00 -3.52976173e-01 -4.60546017e-01 3.71091813e-01
-4.38567370e-01 1.45558581e-01 -7.74953485e-01 -6.07273042e-01
-7.83807039e-01 -9.88624871e-01 1.01804423e+00 -9.04795676e-02
-1.09197818e-01 -1.24650943e+00 6.67105556e-01 7.04814374e-01
6.66147888e-01 -2.40065336e-01 1.32492268e+00 -1.14005208e+00
-4.93808895e-01 -2.46131584e-01 2.07925197e-02 7.60960042e-01
-1.03607655e-01 -4.65133756e-01 -1.10120296e+00 -2.93157697e-01
2.92335749e-02 -1.08922601e+00 8.87174129e-01 -2.33596861e-01
5.33445299e-01 -3.73384565e-01 1.71363890e-01 -1.14933506e-01
1.24498296e+00 -4.48516339e-01 3.57084274e-01 3.91983420e-01
4.57647294e-01 9.69218314e-01 4.21712846e-01 -2.80716538e-01
7.40627289e-01 5.69371760e-01 -1.96687296e-01 4.39911149e-03
-4.13822383e-01 -3.75469565e-01 4.84742880e-01 1.35272372e+00
4.63507921e-01 -2.56024718e-01 -8.70536029e-01 9.45120990e-01
-1.55467629e+00 -5.28824866e-01 -2.41843149e-01 2.44138598e+00
1.37279820e+00 -2.11933881e-01 -1.06133729e-01 -2.54305393e-01
2.70005882e-01 9.78050679e-02 -2.68369585e-01 -4.19687748e-01
-3.67149979e-01 5.84316373e-01 -4.93277572e-02 9.03765440e-01
-8.60379457e-01 1.11975121e+00 5.42717695e+00 8.71935964e-01
-6.93486631e-01 4.94645029e-01 3.05225909e-01 1.58360839e-01
-6.41870201e-01 3.61740887e-02 -8.76269877e-01 1.46274446e-02
1.20929158e+00 -1.70989588e-01 2.89358318e-01 4.58942443e-01
-2.45240286e-01 -1.11648895e-01 -1.26016533e+00 5.06743014e-01
4.21839714e-01 -1.02495992e+00 3.26770753e-01 -6.07851088e-01
9.06528592e-01 4.75450605e-01 -5.52715771e-02 9.20296848e-01
3.39405000e-01 -6.66585267e-01 4.58577037e-01 6.21329963e-01
5.05338192e-01 -5.51692426e-01 8.22765708e-01 5.48725545e-01
-9.48001266e-01 1.09530546e-01 -3.74122262e-01 2.62119830e-01
2.68238664e-01 4.21463251e-01 -6.91505373e-01 8.70871484e-01
3.32570136e-01 2.42185786e-01 -9.97170269e-01 7.34308541e-01
-5.04169762e-01 4.65501696e-01 -3.15458953e-01 5.88246323e-02
5.00772119e-01 -8.77007842e-02 4.62909669e-01 1.27553606e+00
1.11675533e-02 -3.75254363e-01 1.61058202e-01 6.86852038e-01
-4.18590993e-01 8.00821662e-01 -5.67112744e-01 1.70663372e-01
1.57963827e-01 1.14430261e+00 6.75602704e-02 -4.69332486e-01
-4.64098603e-01 1.07373512e+00 8.12268257e-01 5.08167565e-01
-2.51450747e-01 -4.10359472e-01 1.17907738e-02 -1.23419382e-01
-4.97288927e-02 -8.96247327e-02 3.64315689e-01 -1.50510168e+00
3.35396200e-01 -1.27628434e+00 7.84394324e-01 -5.86841226e-01
-1.63846433e+00 1.06265211e+00 -1.99495345e-01 -1.01705360e+00
-7.76646554e-01 -4.21801925e-01 -2.82452911e-01 1.38671041e+00
-2.06163049e+00 -1.22134125e+00 4.58043605e-01 7.83842862e-01
6.54442787e-01 -2.79087782e-01 1.02336085e+00 7.86159754e-01
-1.10758789e-01 7.42440522e-01 1.59725845e-01 5.89815192e-02
1.25002742e+00 -1.23455071e+00 9.80452634e-03 6.44444168e-01
5.59207618e-01 8.42770100e-01 3.93688530e-01 -2.96513408e-01
-1.11585891e+00 -9.46522832e-01 1.64161837e+00 -7.79882729e-01
9.82107639e-01 -4.44723368e-01 -1.30757010e+00 5.33573508e-01
6.53279781e-01 -3.12842578e-01 6.89517260e-01 3.65212917e-01
-3.23784053e-01 -1.01315245e-01 -9.77771759e-01 6.13833427e-01
3.58895272e-01 -1.11927044e+00 -9.84584630e-01 6.79244220e-01
8.48756492e-01 -8.43777359e-02 -6.98825955e-01 3.95664275e-01
1.01132289e-01 -4.48841661e-01 9.01014507e-01 -8.75832438e-01
3.23221207e-01 -1.72241896e-01 -3.10020447e-01 -1.25566423e+00
1.75175026e-01 -2.12765962e-01 -1.17485505e-02 1.56303442e+00
8.96164954e-01 -4.02481228e-01 7.47173727e-02 2.81247348e-01
-4.40244973e-02 -4.83118266e-01 -8.91205966e-01 -5.67713141e-01
6.99787796e-01 -2.95284897e-01 2.44821474e-01 6.62943006e-01
-1.33413911e-01 1.11475921e+00 -1.17088117e-01 -1.55203436e-02
1.85949892e-01 -3.82145345e-02 8.22932065e-01 -8.41078579e-01
-3.18149954e-01 -2.42980331e-01 4.53743935e-01 -1.22803128e+00
7.10572064e-01 -1.48583782e+00 2.53064096e-01 -1.60906601e+00
-4.43710126e-02 -4.76373255e-01 -4.36124086e-01 6.52933568e-02
-6.07997239e-01 9.42456573e-02 1.49313450e-01 1.60734326e-01
-9.02557611e-01 7.59461164e-01 1.15514016e+00 -1.90713063e-01
-1.99852139e-01 -1.02414489e-01 -5.83177626e-01 4.18701410e-01
5.11561573e-01 -5.27567148e-01 -4.37607855e-01 -9.04400647e-01
5.52755237e-01 2.39852935e-01 2.19836831e-01 -8.83492708e-01
1.85579076e-01 4.75629508e-01 -1.52995571e-01 -4.67718810e-01
1.87483549e-01 -6.18864894e-01 -6.18859529e-01 2.72538126e-01
-8.52079928e-01 2.21595481e-01 3.53231609e-01 2.73965240e-01
-6.42101765e-01 -5.97866952e-01 3.68632376e-01 -2.79513091e-01
-5.11888862e-01 -7.69059584e-02 -3.27971607e-01 6.23302281e-01
2.30611563e-01 5.66871464e-01 -3.36135179e-01 -5.01749516e-01
-5.25753975e-01 6.85830653e-01 1.46919275e-02 7.83337355e-01
3.72419834e-01 -1.33259320e+00 -1.23783469e+00 -8.68670363e-03
5.20937085e-01 -2.51753420e-01 3.15915287e-01 6.01332009e-01
-3.07376474e-01 9.81730998e-01 2.29356691e-01 -2.81296283e-01
-7.57891178e-01 4.78843987e-01 3.89585495e-01 -1.16123450e+00
-1.22088902e-01 9.08756673e-01 4.91974726e-02 -1.32307971e+00
1.68533131e-01 -1.86525151e-01 -6.22861505e-01 2.54616648e-01
5.97593546e-01 7.61272907e-02 5.09576261e-01 -6.13471448e-01
-1.54315889e-01 4.37727869e-01 -2.40204394e-01 -6.05072618e-01
9.00726736e-01 -3.30248207e-01 -2.57076472e-01 5.12415349e-01
1.40867400e+00 3.33230346e-01 -4.56534415e-01 -6.51671469e-01
4.69730973e-01 1.13025807e-01 -2.53424227e-01 -1.10429692e+00
-5.15370488e-01 8.86013031e-01 3.93860489e-01 -1.34413093e-01
8.90204191e-01 2.29722679e-01 8.45570385e-01 8.73083413e-01
3.92869890e-01 -1.14356291e+00 3.24547291e-01 1.18841147e+00
1.26392317e+00 -1.32689261e+00 -5.03271341e-01 2.52805799e-01
-5.17152965e-01 6.88796341e-01 5.49707770e-01 -1.19452223e-01
4.73826289e-01 -3.92667383e-01 5.24591148e-01 -1.05120927e-01
-9.08626139e-01 -4.25336450e-01 7.81538069e-01 3.34879994e-01
8.27734053e-01 -2.40056351e-01 -6.69048309e-01 6.36000395e-01
-2.97624469e-01 -1.70014322e-01 9.89414379e-02 7.77886510e-01
-2.30768502e-01 -1.43266630e+00 -2.61001348e-01 2.04527482e-01
-6.55027628e-01 -5.98464191e-01 -3.02550405e-01 4.83343989e-01
-5.18998504e-02 1.20990074e+00 -1.27401888e-01 3.85598689e-02
4.41650629e-01 5.46084821e-01 4.12738264e-01 -9.02918518e-01
-1.02974868e+00 -2.91886270e-01 4.86792892e-01 -2.11488962e-01
-4.47589189e-01 -3.67250830e-01 -8.90062392e-01 5.00595748e-01
-3.49157006e-01 6.70903981e-01 6.51692927e-01 1.09552133e+00
2.85313666e-01 4.77699637e-01 5.30991197e-01 -2.79353112e-01
-7.30677307e-01 -1.26930916e+00 -8.10906589e-02 5.31317353e-01
4.62611347e-01 -5.12916893e-02 -4.50684905e-01 -4.91802506e-02] | [11.366140365600586, 8.264545440673828] |
cbf2ec9a-86d1-460e-bd49-a13d9166d067 | interleaved-multitask-learning-for-audio | 1908.05182 | null | https://arxiv.org/abs/1908.05182v1 | https://arxiv.org/pdf/1908.05182v1.pdf | Interleaved Multitask Learning for Audio Source Separation with Independent Databases | Deep Neural Network-based source separation methods usually train independent models to optimize for the separation of individual sources. Although this can lead to good performance for well-defined targets, it can also be computationally expensive. The multitask alternative of a single network jointly optimizing for all targets simultaneously usually requires the availability of all target sources for each input. This requirement hampers the ability to create large training databases. In this paper, we present a model that decomposes the learnable parameters into a shared parametric model (encoder) and independent components (decoders) specific to each source. We propose an interleaved training procedure that optimizes the sub-task decoders independently and thus does not require each sample to possess a ground truth for all of its composing sources. Experimental results on MUSDB18 with the proposed method show comparable performance to independently trained models, with less trainable parameters, more efficient inference, and an encoder transferable to future target objectives. The results also show that using the proposed interleaved training procedure leads to better Source-to-Interference energy ratios when compared to the simultaneous optimization of all training objectives, even when all composing sources are available. | ['Olumide Okubadejo', 'Clement S. J. Doire'] | 2019-08-14 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 2.98075199e-01 3.03633325e-03 -1.92566454e-01 -5.42306483e-01
-1.34358358e+00 -4.53041703e-01 4.73292261e-01 1.37460023e-01
-3.94895554e-01 9.09208596e-01 6.64126873e-02 1.33275315e-01
-2.19132125e-01 -4.87140805e-01 -9.48311865e-01 -9.95784879e-01
-1.62265107e-01 7.38262236e-01 7.31783062e-02 2.90488482e-01
-3.35216671e-01 3.46047848e-01 -1.39658999e+00 -4.32172380e-02
8.18409324e-01 1.24600434e+00 7.49386907e-01 6.26465380e-01
3.78438942e-02 9.25405502e-01 -9.40096200e-01 -1.26491159e-01
2.00599104e-01 -6.62811100e-01 -2.62363940e-01 -8.60273018e-02
4.56515439e-02 1.03791161e-02 -1.09071098e-01 1.03676081e+00
7.70720005e-01 1.70136780e-01 7.47906625e-01 -1.22998309e+00
-4.44494545e-01 8.53256822e-01 -4.99499172e-01 1.62764311e-01
-4.64312255e-01 -2.69858420e-01 9.27711189e-01 -5.62427461e-01
3.19337845e-02 8.00439477e-01 3.53199273e-01 3.92110378e-01
-1.33579528e+00 -9.02684093e-01 9.97487977e-02 4.56103683e-01
-1.39683688e+00 -9.79076922e-01 8.68280232e-01 -1.73012942e-01
8.82496059e-01 7.91218057e-02 1.99822679e-01 1.15436232e+00
-2.54919022e-01 7.63275921e-01 9.51027155e-01 -3.83671045e-01
3.77336025e-01 5.35475016e-01 1.18255340e-01 4.84559417e-01
3.61152500e-01 -3.59854028e-02 -7.98244119e-01 1.57523051e-01
2.10319325e-01 -5.10012329e-01 -3.24935377e-01 -1.84956998e-01
-1.08464348e+00 7.34056771e-01 1.74194172e-01 4.48775291e-01
-5.03691256e-01 2.18592569e-01 1.54174030e-01 -1.39464084e-02
6.45440996e-01 1.48776740e-01 -5.41843474e-01 1.61970422e-01
-1.39458179e+00 -2.41507202e-01 8.95627499e-01 1.01253486e+00
8.32376957e-01 5.43143630e-01 -4.02022064e-01 9.79370594e-01
2.51122862e-01 7.98240960e-01 3.77198368e-01 -6.80036545e-01
6.01300538e-01 1.22144952e-01 4.56541963e-02 -8.08252275e-01
-4.64439809e-01 -1.14129651e+00 -9.58306491e-01 3.81623730e-02
4.19069260e-01 -5.04520714e-01 -9.14876521e-01 2.11463833e+00
1.13146380e-01 2.97739893e-01 3.08434814e-01 8.03331494e-01
6.34508967e-01 1.01590681e+00 -2.80507025e-03 -4.57862616e-01
1.08793747e+00 -1.18724906e+00 -8.04944754e-01 -5.75513184e-01
1.78483203e-01 -6.76736653e-01 5.04143834e-01 4.85130340e-01
-1.08673012e+00 -5.85158706e-01 -1.25463045e+00 1.06165200e-01
-1.04391761e-01 7.50410795e-01 4.30983186e-01 9.53446984e-01
-8.21764886e-01 2.19579309e-01 -8.21648479e-01 1.93356514e-01
4.26050484e-01 6.98587418e-01 1.47700924e-02 2.23822877e-01
-1.08136272e+00 1.04017937e+00 6.16877735e-01 -5.45098335e-02
-1.48323750e+00 -7.89809763e-01 -6.79463744e-01 5.11519611e-01
5.70360184e-01 -4.23971683e-01 1.18525147e+00 -1.44465292e+00
-1.76943016e+00 3.22602838e-01 -2.67689764e-01 -6.40204191e-01
9.82218906e-02 -8.34786296e-02 -4.69914019e-01 -4.20438871e-03
-1.27075866e-01 3.64421278e-01 9.91984725e-01 -1.25173628e+00
-6.23760939e-01 -9.47939306e-02 -2.64466137e-01 4.59569097e-01
-7.16438770e-01 9.60052609e-02 -7.13747621e-01 -6.30043626e-01
-3.77944261e-01 -6.54292285e-01 1.01728432e-01 -2.48038277e-01
-6.60287499e-01 5.53390421e-02 4.37254488e-01 -8.55003536e-01
9.65228915e-01 -1.97540891e+00 5.26795626e-01 1.52783602e-01
-4.73617390e-02 1.03137828e-01 -2.68703729e-01 -2.46283016e-03
-1.25194401e-01 -3.64751011e-01 -3.25801671e-01 -7.86633015e-01
3.49252932e-02 1.74146459e-01 1.95033193e-01 6.33620918e-01
7.69823343e-02 5.08051634e-01 -6.07995927e-01 -4.70748633e-01
5.24186678e-02 6.84215486e-01 -2.23621801e-01 3.40649366e-01
-3.40903670e-01 5.47557950e-01 -1.96013883e-01 3.44658881e-01
4.49647784e-01 -3.24649453e-01 2.29431078e-01 -4.88888383e-01
1.64239645e-01 3.21886182e-01 -1.35312736e+00 1.74282217e+00
-7.90143907e-01 7.81410456e-01 5.22882164e-01 -1.56733596e+00
7.38467515e-01 5.69187999e-01 6.46394014e-01 -7.12644517e-01
2.07825661e-01 2.52217561e-01 1.91738859e-01 -6.65671676e-02
-7.30178356e-02 -2.88829297e-01 -7.17335716e-02 5.11586010e-01
8.21119010e-01 3.32772046e-01 3.29209238e-01 3.76439691e-02
8.30388427e-01 -6.18670061e-02 1.08921811e-01 -1.97432712e-01
5.58393002e-01 -3.00273180e-01 6.88112438e-01 5.60225010e-01
3.17780256e-01 2.77019888e-01 1.53127328e-01 2.98889548e-01
-8.71888757e-01 -9.77252245e-01 1.64253265e-01 9.88263965e-01
9.42054540e-02 -5.51907904e-02 -6.52933598e-01 -4.71474379e-01
-4.12021488e-01 9.92925167e-01 -4.20129985e-01 -5.05714044e-02
-5.18180132e-01 -1.28957713e+00 5.49029887e-01 1.61753401e-01
4.04604614e-01 -5.71904540e-01 -6.96320653e-01 1.88006148e-01
-5.26425838e-01 -1.20465064e+00 -4.02424604e-01 9.64292586e-01
-5.30640543e-01 -7.45582700e-01 -8.79622459e-01 -4.16100919e-01
5.99783421e-01 6.88524991e-02 9.83502805e-01 -3.99346143e-01
7.15310797e-02 1.97015584e-01 9.94475652e-03 -6.84962571e-01
-5.17802536e-01 1.38978913e-01 2.75840759e-02 3.81837726e-01
-4.01102081e-02 -5.56258559e-01 -1.06130250e-01 1.42742053e-01
-6.97069943e-01 2.62567967e-01 7.73779094e-01 5.74788809e-01
7.12149858e-01 4.61952269e-01 8.24828863e-01 -5.85163832e-01
2.29386657e-01 -7.44992733e-01 -7.75821567e-01 4.19046044e-01
-6.87830925e-01 5.14635623e-01 8.18606377e-01 -4.98699993e-01
-1.44527316e+00 2.58940935e-01 2.96633895e-02 -4.00885344e-01
-9.38451812e-02 3.32712680e-01 -5.82557201e-01 5.20604737e-02
5.80307484e-01 3.82289022e-01 -3.42647165e-01 -7.10534215e-01
2.74525136e-01 4.90626067e-01 5.64243436e-01 -4.71595049e-01
9.47075903e-01 7.27356076e-02 5.66535927e-02 -7.31241167e-01
-9.40784335e-01 -3.15037638e-01 -2.46584758e-01 -3.00198406e-01
7.99503982e-01 -1.18887579e+00 -3.00572276e-01 4.41590488e-01
-1.20062256e+00 -2.46148422e-01 -2.53319204e-01 7.96220839e-01
-4.38847750e-01 1.08331554e-02 -1.16059147e-01 -8.31536949e-01
-3.14515084e-01 -1.12388420e+00 8.05945396e-01 2.32791692e-01
2.00192437e-01 -1.05107331e+00 -4.77738455e-02 2.21180275e-01
3.85008454e-01 -3.13946828e-02 8.55220675e-01 -1.18290401e+00
-5.71000934e-01 -1.00019231e-01 -4.52800840e-02 6.60294116e-01
1.94672152e-01 -4.26327616e-01 -1.31453621e+00 -3.05833165e-02
5.52458942e-01 -3.84779692e-01 8.38553846e-01 5.22133052e-01
1.11709487e+00 -4.26981091e-01 -3.51633817e-01 5.83059311e-01
1.55646253e+00 2.70885110e-01 2.15422019e-01 -3.59215550e-02
5.66035867e-01 4.60008502e-01 9.86203030e-02 3.82354409e-01
2.35040560e-01 8.02183270e-01 3.93170089e-01 -2.70218253e-01
-4.30600733e-01 2.29881674e-01 6.98125601e-01 7.63854563e-01
6.11045510e-02 -7.95792997e-01 -5.29881597e-01 5.61943352e-01
-1.49374712e+00 -9.62978005e-01 -1.68847486e-01 2.39405251e+00
1.03135955e+00 -4.82251681e-02 1.46663263e-01 2.90892750e-01
6.66361034e-01 1.79855604e-04 -6.48341477e-01 8.49375129e-03
-3.20581049e-01 4.11292553e-01 7.59567559e-01 6.49448812e-01
-1.04643285e+00 2.70643950e-01 5.78602362e+00 1.19953930e+00
-1.08199990e+00 7.33619928e-01 6.00088239e-01 -6.97134376e-01
-2.44061783e-01 -3.01317334e-01 -8.61683190e-01 6.15749776e-01
1.40072405e+00 -2.61530221e-01 6.86722934e-01 4.93307561e-01
1.54161111e-01 -3.96056265e-01 -1.26979005e+00 9.99113262e-01
4.14115995e-01 -9.71593440e-01 -2.97253579e-01 -1.17642313e-01
6.32499456e-01 1.88252121e-01 -1.33806348e-01 2.01088026e-01
1.61903843e-01 -9.58546758e-01 1.10470271e+00 5.37563682e-01
4.12207544e-01 -8.81914318e-01 4.86965865e-01 8.76162946e-01
-9.26173806e-01 -1.27669096e-01 -1.66267246e-01 3.21509987e-01
1.67757258e-01 1.03991652e+00 -6.44985974e-01 8.96707714e-01
4.48660910e-01 3.19018066e-01 -3.78816426e-01 1.00650918e+00
-2.90227354e-01 7.81969190e-01 -6.21504426e-01 3.34480293e-02
-1.83007568e-01 -1.14789449e-01 7.27806330e-01 1.23796439e+00
6.42346799e-01 -2.85378456e-01 1.20241225e-01 8.16186428e-01
-1.93830550e-01 4.28033918e-02 -9.40499976e-02 -3.02883442e-02
3.15716863e-01 1.37761819e+00 -7.05711424e-01 -4.63613153e-01
-4.47363913e-01 1.04989934e+00 3.33509624e-01 4.97906864e-01
-1.25592351e+00 -2.87183315e-01 2.59921879e-01 -3.83289814e-01
4.15714771e-01 -1.20718077e-01 -3.28070134e-01 -1.06671321e+00
5.58006018e-03 -7.42902517e-01 2.83781022e-01 -5.56790650e-01
-9.60205078e-01 6.02496505e-01 2.28628665e-01 -8.94322872e-01
-3.64124268e-01 -3.18767846e-01 -4.24164146e-01 1.27612782e+00
-1.70896506e+00 -1.01391387e+00 -3.94205116e-02 7.42172480e-01
6.24777794e-01 -2.63800174e-01 9.15770829e-01 7.39587605e-01
-8.44800711e-01 5.98498583e-01 1.99907169e-01 -1.05246633e-01
7.26800740e-01 -1.10314846e+00 -3.59134704e-01 1.08566463e+00
5.96547782e-01 -8.51913318e-02 8.41314852e-01 -2.87115902e-01
-1.29379785e+00 -1.05378962e+00 7.11044610e-01 -9.91595462e-02
3.33698392e-01 -5.18762350e-01 -6.28775954e-01 4.41467762e-01
5.14861584e-01 -2.43803978e-01 9.34743226e-01 9.76506472e-02
-6.00232333e-02 -5.54220557e-01 -8.79803836e-01 -7.25194663e-02
5.65332651e-01 -2.17006966e-01 -2.62507468e-01 5.30772328e-01
3.37686062e-01 -2.24864975e-01 -5.92438161e-01 1.98694855e-01
2.40420893e-01 -9.68510985e-01 1.00873613e+00 9.73955356e-03
6.02836534e-02 -2.95503736e-01 -2.21828163e-01 -1.53393853e+00
-1.76128864e-01 -2.51183748e-01 -4.93934989e-01 1.46776628e+00
8.62597227e-01 -3.46892118e-01 4.96024936e-01 4.57947969e-01
-3.20353448e-01 -4.01988864e-01 -1.06393600e+00 -1.00715089e+00
-3.37129384e-01 -6.79608166e-01 2.91984469e-01 5.91631055e-01
-4.04707491e-01 6.69595242e-01 -8.11422706e-01 6.11912131e-01
9.64445591e-01 1.18927032e-01 4.29739833e-01 -1.02590656e+00
-9.82186556e-01 -3.42311144e-01 2.25474298e-01 -9.14041519e-01
2.98433125e-01 -1.19652772e+00 5.24335861e-01 -1.53683150e+00
2.51555592e-01 -5.68056881e-01 -5.24721146e-01 6.53370321e-01
-6.81995079e-02 1.01357818e-01 1.13364078e-01 4.13348898e-02
-4.63575065e-01 6.00219965e-01 7.42002428e-01 -4.54317927e-01
-1.23700187e-01 2.61303842e-01 -7.94502795e-01 5.12455106e-01
8.77580225e-01 -9.56047177e-01 -8.38354647e-01 -7.31980860e-01
-3.73564027e-02 7.43827671e-02 3.25896770e-01 -1.43918145e+00
3.66378188e-01 -6.28161430e-03 6.66472614e-01 -2.99782425e-01
6.60017610e-01 -1.04947352e+00 5.80010176e-01 2.80681521e-01
-3.69061530e-01 -6.52860284e-01 3.71806651e-01 6.39966488e-01
-1.36774302e-01 -6.06822252e-01 8.51182282e-01 1.82113841e-01
-3.25925320e-01 1.32583842e-01 -4.13123280e-01 -1.28604919e-01
1.08655727e+00 1.79202408e-01 1.25012115e-01 -4.47171211e-01
-7.90523410e-01 1.50043935e-01 -1.64097026e-01 2.72128791e-01
2.11954549e-01 -1.11936986e+00 -7.97217488e-01 -1.28717972e-02
-3.74571890e-01 -1.98096901e-01 1.95774466e-01 8.47627461e-01
2.64615536e-01 4.78092402e-01 7.33990595e-02 -4.23448414e-01
-1.12966800e+00 6.67465031e-01 5.07241309e-01 -4.81372744e-01
-9.98548120e-02 9.04325545e-01 2.90051639e-01 -6.84312657e-02
5.35572231e-01 -6.72688857e-02 -2.08980381e-01 3.28901231e-01
3.63920331e-01 3.82499129e-01 2.25657120e-01 -6.36595786e-01
-3.32804531e-01 2.56913662e-01 4.31777894e-01 -2.79773682e-01
1.41842401e+00 -1.70074195e-01 3.38728353e-02 5.73387325e-01
1.20522106e+00 1.74136728e-01 -1.16037834e+00 -3.54603738e-01
-1.57619655e-01 7.32724816e-02 5.82045674e-01 -1.19231892e+00
-1.42357326e+00 9.48547244e-01 6.70566022e-01 -3.86800915e-02
1.69833481e+00 -2.63801236e-02 5.20325720e-01 4.31430608e-01
2.66972065e-01 -1.05956531e+00 -1.53779805e-01 1.98043600e-01
6.12220943e-01 -1.02176785e+00 -1.47795022e-01 -2.67285913e-01
-4.20339882e-01 9.97079492e-01 3.50355119e-01 1.90756276e-01
6.66824937e-01 5.60269415e-01 -2.87936479e-01 -1.98220778e-02
-7.60650337e-01 -2.92211384e-01 6.12743735e-01 5.09930551e-01
1.86920226e-01 -3.95608172e-02 1.22735821e-01 9.44462180e-01
2.70274818e-01 -1.31041080e-01 1.27552047e-01 4.32902902e-01
-4.80011851e-01 -1.27740502e+00 -5.22397101e-01 5.32227278e-01
-6.59460247e-01 -1.81747571e-01 4.05119732e-02 3.71839523e-01
3.27335864e-01 1.17744863e+00 -1.72930688e-01 -9.29836705e-02
1.30693674e-01 2.09268600e-01 6.72354639e-01 -6.92632854e-01
-5.00651658e-01 4.62941796e-01 4.54402089e-01 -2.35238299e-01
-8.17258000e-01 -5.37805259e-01 -1.04236150e+00 1.42417520e-01
-5.51735282e-01 3.21147472e-01 1.00785124e+00 1.05548835e+00
1.94177330e-01 9.61041212e-01 6.98084831e-01 -9.78869498e-01
-4.60238904e-01 -9.68311071e-01 -5.72873890e-01 -2.28587568e-01
4.54103261e-01 -7.00540483e-01 -2.83029437e-01 3.52003455e-01] | [15.303679466247559, 5.61536169052124] |
9e47d28b-4ada-489c-bf56-1a151e3b945c | can-eye-movement-data-be-used-as-ground-truth | 1804.08749 | null | http://arxiv.org/abs/1804.08749v1 | http://arxiv.org/pdf/1804.08749v1.pdf | Can Eye Movement Data Be Used As Ground Truth For Word Embeddings Evaluation? | In recent years a certain success in the task of modeling lexical semantics
was obtained with distributional semantic models. Nevertheless, the scientific
community is still unaware what is the most reliable evaluation method for
these models. Some researchers argue that the only possible gold standard could
be obtained from neuro-cognitive resources that store information about human
cognition. One of such resources is eye movement data on silent reading. The
goal of this work is to test the hypothesis of whether such data could be used
to evaluate distributional semantic models on different languages. We propose
experiments with English and Russian eye movement datasets (Provo Corpus, GECO
and Russian Sentence Corpus), word vectors (Skip-Gram models trained on
national corpora and Web corpora) and word similarity datasets of Russian and
English assessed by humans in order to find the existence of correlation
between embeddings and eye movement data and test the hypothesis that this
correlation is language independent. As a result, we found that the validity of
the hypothesis being tested could be questioned. | ['Amir Bakarov'] | 2018-04-23 | null | null | null | null | ['embeddings-evaluation'] | ['natural-language-processing'] | [-1.79035634e-01 6.76237270e-02 -4.39171419e-02 -3.73950362e-01
1.46465935e-02 -4.76446301e-01 9.15177643e-01 6.33257449e-01
-1.17859483e+00 5.18900335e-01 4.03538853e-01 -5.53689361e-01
-2.81077445e-01 -7.06540167e-01 -1.96411833e-01 -2.74277449e-01
3.63847226e-01 4.42648977e-01 4.77038682e-01 -3.71107429e-01
9.11332607e-01 2.75798887e-01 -1.83822763e+00 1.41195536e-01
8.40488911e-01 5.07634342e-01 5.31800032e-01 3.46825927e-01
-4.54013288e-01 5.17672896e-01 -4.75996524e-01 -3.52760792e-01
-1.85492833e-03 -6.24658108e-01 -1.04696870e+00 -3.02662969e-01
2.37752438e-01 2.25159436e-01 2.34705601e-02 1.40244079e+00
6.25516772e-01 1.93218216e-01 7.69574106e-01 -8.93273830e-01
-1.28467834e+00 4.77375001e-01 -5.36816828e-02 6.93676114e-01
4.62058634e-01 -2.31800154e-02 1.09186876e+00 -7.72081912e-01
9.35322762e-01 1.22438860e+00 2.16857210e-01 5.85308015e-01
-9.74666655e-01 -1.97949842e-01 -8.74822363e-02 6.49362147e-01
-1.45910609e+00 -2.30771929e-01 4.79768366e-01 -4.03093189e-01
1.05025780e+00 1.38942510e-01 5.97409666e-01 1.27083218e+00
1.05642878e-01 1.91517800e-01 1.79233575e+00 -9.87476468e-01
3.08626741e-01 8.25615823e-01 5.42541564e-01 4.20691073e-01
5.54280698e-01 1.73566341e-01 -8.85985851e-01 1.15291908e-01
3.34264040e-01 -5.63986123e-01 -4.72226024e-01 -2.39835814e-01
-1.05618131e+00 8.98608148e-01 1.46097258e-01 9.49350238e-01
-2.53593475e-01 -4.90997821e-01 4.18332219e-01 6.31232500e-01
5.41464567e-01 3.64388555e-01 -5.56768239e-01 -2.66105324e-01
-6.99751019e-01 1.99821740e-01 7.02501833e-01 2.39291564e-01
3.77939731e-01 -1.11770540e-01 6.70843795e-02 8.32258463e-01
8.62504721e-01 4.68395501e-01 1.24345803e+00 -2.23219812e-01
2.79246747e-01 6.28967464e-01 -1.90162975e-02 -9.77346003e-01
-6.25476897e-01 1.22365758e-01 -8.19869619e-03 2.50183940e-01
5.60478151e-01 -1.62505452e-02 -4.77048248e-01 1.81292725e+00
1.37403637e-01 -3.11184675e-01 2.50223547e-01 9.32679057e-01
9.55337703e-01 2.20120430e-01 2.50685662e-01 -2.16174826e-01
1.61042547e+00 -2.65470564e-01 -7.52323866e-01 -2.54408777e-01
1.02455080e+00 -7.17598021e-01 1.23387182e+00 1.01095147e-01
-6.65445328e-01 -6.55119419e-01 -1.05457318e+00 -3.38532656e-01
-9.04934227e-01 -1.15887132e-02 3.30476433e-01 9.01630342e-01
-1.04870653e+00 3.96677017e-01 -5.81470132e-01 -1.01489651e+00
1.41472341e-02 1.94052774e-02 -5.22102177e-01 2.08847523e-01
-1.27487648e+00 1.61598015e+00 6.57433152e-01 -1.57862186e-01
-2.59819388e-01 -1.56794667e-01 -8.06720614e-01 -2.19738990e-01
4.62145470e-02 -5.16910434e-01 6.75163865e-01 -1.35388625e+00
-1.30994785e+00 1.57324314e+00 -2.63949394e-01 -5.96145034e-01
1.75451666e-01 4.54666056e-02 -6.78236425e-01 -3.05241272e-02
9.03005376e-02 4.49181587e-01 4.33066100e-01 -7.32042670e-01
-2.98752725e-01 -9.19564962e-01 -1.55686319e-01 2.55041361e-01
-3.87071431e-01 4.46812391e-01 2.96518385e-01 -4.79887247e-01
-5.95568419e-02 -9.39779997e-01 4.14961755e-01 -2.07801685e-01
4.73607099e-03 -7.79572785e-01 3.32942247e-01 -7.16110170e-01
1.15370810e+00 -2.11072636e+00 8.63246340e-03 4.86306027e-02
-3.53027247e-02 3.29127789e-01 -1.30904257e-01 3.88871521e-01
-2.19110101e-01 2.86840320e-01 -1.29305208e-02 6.10363074e-02
1.15473583e-01 1.08478867e-01 -2.36442000e-01 6.73280358e-01
-5.75329326e-02 9.00232255e-01 -6.60695851e-01 -4.37880009e-01
1.56842843e-01 3.87148112e-01 -4.43614393e-01 -1.73223078e-01
-7.82177523e-02 2.53316671e-01 -3.30890000e-01 1.27780437e-01
3.04850042e-01 7.61308447e-02 1.31494537e-01 5.42349890e-02
-5.53587556e-01 6.31398618e-01 -1.06129277e+00 1.47368848e+00
-3.30177844e-02 1.10558367e+00 -6.11960173e-01 -1.04343414e+00
9.78248417e-01 2.48532891e-01 -1.90412134e-01 -1.30255520e+00
5.87816179e-01 3.20666730e-01 5.61871588e-01 -7.65258014e-01
6.26328409e-01 -4.36356127e-01 2.65061885e-01 3.41896564e-01
1.89130887e-01 2.58947045e-01 2.39164397e-01 -3.07988048e-01
6.47416830e-01 2.31427804e-01 4.32760179e-01 -7.67504275e-01
7.61056602e-01 8.78535807e-02 3.41978222e-02 3.45500052e-01
-2.07635462e-01 4.71939743e-01 5.63419998e-01 -2.43106782e-01
-6.97380364e-01 -1.01268530e+00 -5.50647378e-01 9.81206834e-01
9.73792374e-02 -4.51301306e-01 -9.74194646e-01 -2.74028838e-01
-3.66086274e-01 1.32251358e+00 -7.17002690e-01 -1.83414951e-01
-1.17358662e-01 -7.69897997e-01 2.32099131e-01 1.84709832e-01
1.95599213e-01 -1.28335679e+00 -1.06594205e+00 -2.54301816e-01
1.68818496e-02 -9.50474381e-01 2.10373193e-01 9.90890712e-02
-9.80237842e-01 -1.19410646e+00 -4.42846715e-01 -6.66413844e-01
3.62787426e-01 1.33152753e-01 9.10365164e-01 2.69024782e-02
-1.79620326e-01 3.76227349e-01 -4.98803675e-01 -7.81206787e-01
-3.42787534e-01 -4.64718603e-02 1.95920840e-01 -2.98296481e-01
1.25085735e+00 -1.55608267e-01 -2.74337739e-01 4.61878674e-03
-9.35950875e-01 -4.86102849e-01 3.09847891e-01 4.07102376e-01
2.00561404e-01 -4.25651431e-01 5.13482332e-01 -6.37704015e-01
1.05532801e+00 -6.01713777e-01 -5.00028908e-01 2.76823878e-01
-7.89129794e-01 3.28726679e-01 9.01343152e-02 -3.42359155e-01
-9.00609553e-01 -8.39605033e-01 -2.21238628e-01 1.20594054e-01
-5.27040601e-01 6.41216159e-01 2.04537567e-02 7.21629485e-02
9.85270441e-01 2.52083063e-01 2.14411423e-01 -5.10952115e-01
2.29351982e-01 8.32462788e-01 2.33498886e-02 -2.51004487e-01
4.24951643e-01 2.47496858e-01 -2.58734554e-01 -1.46226120e+00
-6.91431284e-01 -3.32173586e-01 -6.10786676e-01 6.83368966e-02
1.29646719e+00 -6.99887395e-01 -5.58253706e-01 1.59848094e-01
-1.14706290e+00 -5.68987317e-02 -2.01351330e-01 7.17553794e-01
-4.88429666e-01 3.95441383e-01 -4.90608700e-02 -7.34646201e-01
7.73994136e-04 -7.90924609e-01 5.41825533e-01 1.44444630e-01
-7.31345057e-01 -1.34884202e+00 3.94279361e-01 3.29819143e-01
3.56348604e-01 -2.91665256e-01 1.06674838e+00 -1.19053149e+00
9.19781253e-03 6.36849105e-02 -1.21814765e-01 3.93892854e-01
-6.17034137e-02 -1.76294550e-01 -8.50530863e-01 2.03810543e-01
3.96240562e-01 -2.74189800e-01 7.45609701e-01 2.81729490e-01
6.05717003e-01 2.43135929e-01 -3.21179368e-02 -6.65736618e-03
1.60471988e+00 3.36012915e-02 7.61296809e-01 5.73150218e-01
3.47811252e-01 1.01307762e+00 4.49449390e-01 8.94579142e-02
5.32935441e-01 7.77514040e-01 -3.10337190e-02 5.89401782e-01
-1.50352344e-01 -7.40830153e-02 5.22981107e-01 7.81541646e-01
-1.38095662e-01 -3.13773364e-01 -1.01121140e+00 6.91762269e-01
-1.56312382e+00 -1.03192723e+00 -4.03719872e-01 2.47836828e+00
5.80975950e-01 1.14137776e-01 1.36846885e-01 9.59490687e-02
4.45233703e-01 1.07515231e-01 4.86200079e-02 -7.51759529e-01
-2.26415902e-01 1.57524794e-01 2.48283282e-01 4.27102178e-01
-6.42516911e-01 9.15672958e-01 5.89148760e+00 4.61261392e-01
-1.16265655e+00 1.29824847e-01 1.53380424e-01 2.72792652e-02
-3.88492376e-01 -4.48473282e-02 -8.38095009e-01 5.24683297e-01
1.47010422e+00 -2.75980115e-01 1.00165851e-01 4.05128896e-01
3.11912388e-01 -4.13760304e-01 -8.47187400e-01 9.06375051e-01
6.86406136e-01 -9.74089324e-01 -8.23623464e-02 6.41029514e-03
2.45266020e-01 4.04377311e-01 2.76509952e-02 1.97872952e-01
-2.66522288e-01 -9.90934074e-01 8.05875003e-01 7.14209139e-01
1.06895015e-01 -3.98643374e-01 7.96302080e-01 5.81128776e-01
-3.45213622e-01 1.94894060e-01 -5.72456181e-01 -2.35594526e-01
-3.70635488e-03 -8.93205851e-02 -5.33522427e-01 1.37087569e-01
6.09630167e-01 3.43295842e-01 -9.41233873e-01 8.92760038e-01
-3.26273680e-01 5.14736772e-01 -1.65417671e-01 -5.95876813e-01
2.21361831e-01 -3.29733342e-01 5.05999207e-01 9.27798450e-01
4.46513057e-01 -2.01366216e-01 -3.92861873e-01 1.01610672e+00
3.73100162e-01 8.26322258e-01 -8.71872187e-01 -3.90783221e-01
2.84253120e-01 8.81044745e-01 -9.42945421e-01 1.18056528e-01
-8.84912133e-01 1.02147543e+00 5.73389947e-01 5.92320561e-02
-5.80049634e-01 -2.30194479e-01 5.07938504e-01 4.02591139e-01
1.65660307e-01 -3.01426142e-01 -2.90767401e-01 -1.10508990e+00
-2.81762779e-02 -6.31576359e-01 1.31564468e-01 -9.90149856e-01
-1.20367098e+00 6.14946544e-01 9.35510695e-02 -8.34777296e-01
-1.70801058e-01 -1.27260244e+00 -2.51704901e-01 1.28224182e+00
-1.40872931e+00 -6.06733561e-01 8.05946533e-03 5.41701853e-01
4.22551543e-01 -4.12530154e-01 9.90788996e-01 -1.54717444e-02
-1.79060459e-01 1.36537865e-01 -1.49710357e-01 -4.57214601e-02
8.29225838e-01 -1.16838491e+00 5.58894128e-03 5.96160650e-01
7.30335474e-01 9.28708732e-01 9.66284811e-01 -6.27716124e-01
-8.05507898e-01 -3.48731965e-01 1.57164395e+00 -8.54109883e-01
8.44384313e-01 4.86039035e-02 -1.03079689e+00 3.34790587e-01
6.43320739e-01 -1.94100574e-01 9.94927108e-01 1.92659408e-01
-3.53960484e-01 5.09618759e-01 -9.66406703e-01 6.57841325e-01
8.84289026e-01 -7.40827978e-01 -1.44231927e+00 1.91610128e-01
3.06207746e-01 6.22768551e-02 -5.46818435e-01 -4.03835364e-02
1.57795399e-01 -1.10465777e+00 7.61917353e-01 -7.73183942e-01
3.13080609e-01 -2.16879040e-01 -2.32998326e-01 -1.41318142e+00
1.01696365e-01 9.20684412e-02 4.39903617e-01 1.11417234e+00
4.63179678e-01 -7.87390947e-01 3.81564528e-01 4.55765545e-01
1.72177196e-01 -2.07194954e-01 -1.08716977e+00 -6.93315566e-01
4.39508438e-01 -5.14458895e-01 1.04527906e-01 1.04851913e+00
1.23358153e-01 7.08552361e-01 2.22868621e-01 -7.63625130e-02
1.72987446e-01 -2.08916575e-01 3.74321342e-01 -1.67021620e+00
6.87878802e-02 -5.86487770e-01 -6.93730533e-01 -3.01561654e-01
7.24103570e-01 -1.15313888e+00 -2.55061835e-01 -1.59746063e+00
9.09131765e-02 1.27684355e-01 -3.38427454e-01 1.22198351e-01
-1.77694708e-01 8.37445855e-02 2.55247653e-01 -1.04590952e-02
-2.40383402e-01 4.55483764e-01 8.46268713e-01 4.42166090e-01
-3.27116512e-02 -3.87892276e-01 -8.68275344e-01 9.05000806e-01
9.81491268e-01 -5.12781084e-01 -6.61419690e-01 -3.39619994e-01
6.43967390e-01 -4.55891132e-01 6.63783848e-01 -8.27224433e-01
9.94074196e-02 -7.31087336e-03 5.68604283e-02 -1.30320936e-01
4.06689756e-02 -7.61529863e-01 -4.39323425e-01 2.86297828e-01
-3.73973399e-01 4.87606108e-01 2.41324008e-01 4.95645702e-01
-3.12236071e-01 -8.84881616e-01 6.13096952e-01 -1.05657414e-01
-9.77781475e-01 -2.78754920e-01 -4.25667644e-01 4.60070223e-01
9.62097704e-01 -5.18486321e-01 -3.82787853e-01 -4.53770570e-02
-5.93342960e-01 -2.54573464e-01 4.97717291e-01 9.59399104e-01
4.05210495e-01 -1.06140387e+00 -6.27891779e-01 1.28381521e-01
3.80945325e-01 -1.03627527e+00 1.22932553e-01 1.00338757e+00
-5.34640133e-01 7.06524670e-01 -4.22838330e-01 -1.29665613e-01
-1.44599938e+00 8.35041463e-01 1.82359472e-01 3.23430210e-01
-3.19793820e-01 6.31122828e-01 -2.58298993e-01 -3.80256772e-01
-1.58226565e-01 -1.51673004e-01 -6.79571688e-01 4.94527489e-01
6.16497636e-01 2.92219341e-01 7.35555217e-02 -1.07023335e+00
-4.23028529e-01 3.45651001e-01 1.20568380e-01 -3.89530659e-01
1.29868066e+00 -3.54795486e-01 -3.35202485e-01 1.05890393e+00
1.11094224e+00 2.61139184e-01 -2.81350344e-01 -5.27496636e-02
4.64476913e-01 -2.74289638e-01 1.49885252e-01 -7.04779863e-01
-4.47374076e-01 1.04014421e+00 9.68260527e-01 3.90857249e-01
4.78617847e-01 1.17433637e-01 4.80224565e-02 3.01538557e-01
2.77873576e-01 -1.54225826e+00 -3.30573261e-01 6.27714992e-01
7.30782747e-01 -1.39330900e+00 -2.87334442e-01 -2.38139853e-02
-8.23058307e-01 1.02846920e+00 4.10444051e-01 -1.35464355e-01
7.85766065e-01 -3.19179446e-01 1.03762083e-01 -4.69771445e-01
-5.96146047e-01 -5.95627785e-01 5.34351230e-01 6.94139242e-01
1.15475702e+00 -2.40209457e-02 -1.25169587e+00 5.98129988e-01
-4.17302489e-01 1.04898848e-02 4.93090123e-01 6.80089712e-01
-7.14469373e-01 -1.28352678e+00 -2.88426727e-01 2.34376624e-01
-5.93389869e-01 -3.03073972e-01 -5.71531713e-01 9.74929988e-01
9.48103070e-02 9.51393783e-01 1.76241636e-01 1.50330573e-01
1.75517350e-01 5.83644390e-01 7.48346567e-01 -8.37024093e-01
-4.52023625e-01 -4.59359050e-01 1.45605117e-01 -2.81606317e-01
-6.69630468e-01 -8.48994732e-01 -1.13183975e+00 -6.28756136e-02
-2.30105981e-01 2.07524300e-01 8.25439274e-01 1.23378658e+00
1.80557296e-01 3.13633621e-01 -2.05031723e-01 -7.04907849e-02
-5.32002449e-01 -1.16769958e+00 -5.30589700e-01 6.65833771e-01
-1.56340420e-01 -7.64011323e-01 -3.76800954e-01 -5.29797003e-02] | [10.530095100402832, 9.304298400878906] |
31c88216-66fd-48bc-9bd1-4e708ad7de62 | adversarial-embedding-a-robust-and-elusive | 1912.01487 | null | https://arxiv.org/abs/1912.01487v1 | https://arxiv.org/pdf/1912.01487v1.pdf | Adversarial Embedding: A robust and elusive Steganography and Watermarking technique | We propose adversarial embedding, a new steganography and watermarking technique that embeds secret information within images. The key idea of our method is to use deep neural networks for image classification and adversarial attacks to embed secret information within images. Thus, we use the attacks to embed an encoding of the message within images and the related deep neural network outputs to extract it. The key properties of adversarial attacks (invisible perturbations, nontransferability, resilience to tampering) offer guarantees regarding the confidentiality and the integrity of the hidden messages. We empirically evaluate adversarial embedding using more than 100 models and 1,000 messages. Our results confirm that our embedding passes unnoticed by both humans and steganalysis methods, while at the same time impedes illicit retrieval of the message (less than 13% recovery rate when the interceptor has some knowledge about our model), and is resilient to soft and (to some extent) aggressive image tampering (up to 100% recovery rate under jpeg compression). We further develop our method by proposing a new type of adversarial attack which improves the embedding density (amount of hidden information) of our method to up to 10 bits per pixel. | ['Salah Ghamizi', 'Mike Papadakis', 'Yves Le Traon', 'Maxime Cordy'] | 2019-11-14 | null | null | null | null | ['steganalysis'] | ['computer-vision'] | [ 8.06611538e-01 5.05438745e-01 7.52262771e-02 1.66191593e-01
-5.17964900e-01 -1.04447782e+00 7.11539924e-01 -1.46469548e-01
-4.65069294e-01 3.98444861e-01 3.58440094e-02 -6.50718570e-01
5.21032393e-01 -1.05068624e+00 -1.16167819e+00 -7.52158344e-01
-7.59179115e-01 -4.55047280e-01 2.70270109e-01 -3.51194561e-01
1.96108118e-01 5.29695332e-01 -9.55511212e-01 3.38250548e-01
3.12533587e-01 9.91697311e-01 -4.91097987e-01 1.17886698e+00
5.31888127e-01 1.05345786e+00 -8.20662022e-01 -7.49555349e-01
7.19052017e-01 -3.26255679e-01 -6.07695401e-01 -1.12793408e-01
9.68947411e-02 -1.01624250e+00 -9.43853557e-01 1.37937129e+00
3.43679339e-01 -7.35230386e-01 4.14148718e-01 -1.28641343e+00
-8.81022155e-01 7.13432968e-01 -3.64769191e-01 -6.19635880e-02
1.71625778e-01 5.54805815e-01 5.19610226e-01 -2.63911694e-01
6.38795793e-01 9.81976151e-01 5.56983232e-01 6.95834696e-01
-9.26455915e-01 -9.91741359e-01 -2.70692527e-01 9.70100462e-02
-1.32956374e+00 -7.81658769e-01 6.60090327e-01 -4.16622832e-02
5.25249124e-01 5.76204598e-01 4.22365665e-01 1.13990211e+00
7.59167790e-01 4.70379919e-01 1.09412098e+00 -4.51064199e-01
8.01692382e-02 4.88738596e-01 -5.61468124e-01 6.70337319e-01
3.48968387e-01 5.63968003e-01 -9.60320830e-02 -3.53002101e-01
6.98477447e-01 -8.11419412e-02 -5.63460410e-01 -4.00474705e-02
-1.24961805e+00 9.63714123e-01 5.93306005e-01 1.29781827e-01
-2.73347497e-01 7.65527427e-01 4.35168117e-01 7.96070993e-01
1.40344352e-02 2.51471996e-01 8.38089362e-03 3.33346695e-01
-4.61811125e-01 -2.17951592e-02 9.58443522e-01 7.26527214e-01
3.99513155e-01 2.95440435e-01 3.03799540e-01 3.50488201e-02
4.07097846e-01 9.85792637e-01 4.86234069e-01 -1.00159061e+00
4.05557215e-01 4.06048633e-02 -8.08110833e-02 -1.79559720e+00
3.45646530e-01 -2.13929024e-02 -9.67292845e-01 6.56944752e-01
1.06581829e-01 -1.74724519e-01 -7.92839944e-01 1.71181023e+00
6.25243690e-03 1.21137671e-01 4.94745135e-01 5.77745676e-01
4.27019298e-01 8.93568039e-01 -2.02722684e-01 3.85064483e-02
1.23607659e+00 -5.91919482e-01 -6.49446845e-01 -4.49626476e-01
3.34233254e-01 -7.90679693e-01 3.92051756e-01 3.67934555e-02
-1.11767876e+00 -2.52931923e-01 -1.55068445e+00 2.96643734e-01
-4.69933093e-01 -7.85001993e-01 2.63893634e-01 1.20001280e+00
-1.01022804e+00 4.41881984e-01 -5.55095196e-01 4.30394918e-01
4.78605807e-01 7.84303188e-01 -7.94498265e-01 -2.19784737e-01
-1.60745132e+00 5.48730910e-01 5.04282594e-01 -9.48094428e-02
-1.12903833e+00 -2.35573724e-01 -1.12339830e+00 3.55042703e-02
2.83400416e-02 -4.61999893e-01 6.14404023e-01 -1.56432498e+00
-1.24594152e+00 8.65006804e-01 4.25928622e-01 -9.43497717e-01
6.00896478e-01 2.55276084e-01 -6.85091019e-01 7.08930135e-01
-4.48281616e-01 7.05243766e-01 1.52262318e+00 -1.39663208e+00
-1.25499353e-01 -1.33644566e-01 3.51837605e-01 -3.37042689e-01
-6.99615479e-01 6.03328422e-02 -1.15684181e-01 -9.62555408e-01
-2.78787259e-02 -1.11109900e+00 -8.81191269e-02 4.13899451e-01
-4.79372054e-01 8.48744512e-01 1.07208323e+00 -8.98889661e-01
8.16927195e-01 -2.45894432e+00 -1.12475514e-01 5.50859213e-01
5.24550438e-01 5.92677891e-01 -4.39970851e-01 6.52429104e-01
-1.25208154e-01 6.53193176e-01 -5.36506772e-01 8.82190838e-02
-5.50951622e-02 2.95471787e-01 -6.06479108e-01 9.49790418e-01
5.78735098e-02 1.22384942e+00 -7.43506014e-01 -3.51422012e-01
1.79524809e-01 9.09370780e-01 -4.86435324e-01 1.00590110e-01
1.85794502e-01 -1.87592078e-02 -1.58266231e-01 5.38992524e-01
1.03814876e+00 -8.82617831e-02 3.71094435e-01 -7.26766139e-02
4.94613707e-01 -5.17418161e-02 -9.96623576e-01 6.61229134e-01
-1.71907216e-01 9.71223652e-01 2.09889442e-01 -5.98704159e-01
5.10921359e-01 5.56024253e-01 1.01969637e-01 -4.46191221e-01
2.91305989e-01 2.00369343e-01 2.19624247e-02 -4.65108037e-01
3.75613809e-01 1.56489611e-02 -1.78306371e-01 7.98061907e-01
-2.80973047e-01 3.10308754e-01 -5.73065996e-01 2.64994115e-01
1.35478473e+00 -5.38326085e-01 1.71329290e-01 1.71011522e-01
4.54859287e-01 -5.54130018e-01 1.79842323e-01 8.57740223e-01
-2.32492134e-01 2.13495672e-01 4.14701849e-01 -3.05736333e-01
-1.37486720e+00 -7.57098675e-01 2.81602949e-01 5.39718866e-01
3.39011639e-01 -1.87923253e-01 -9.07121480e-01 -7.30240583e-01
6.74667284e-02 1.86127782e-01 -6.53050542e-01 -6.87916636e-01
-7.12679386e-01 -3.27463955e-01 1.23387778e+00 2.40589842e-01
1.02954733e+00 -1.11698067e+00 -5.13779879e-01 4.21228074e-02
-1.93868414e-01 -1.41185558e+00 -3.83771926e-01 -3.20700556e-01
-6.06416523e-01 -1.18733585e+00 -6.15535796e-01 -7.56821990e-01
9.27656651e-01 1.85355902e-01 5.00214756e-01 6.48000538e-01
-5.85126579e-02 3.50910902e-01 -3.48519534e-01 -2.12574393e-01
-1.41088760e+00 -3.81230444e-01 -1.64100766e-01 1.51444867e-01
-1.64654523e-01 -3.91431034e-01 -6.57601655e-01 2.41464853e-01
-1.58468008e+00 -2.32098445e-01 6.25866413e-01 6.87609255e-01
-2.19101384e-02 5.48894107e-01 1.69081137e-01 -8.83756161e-01
3.69090647e-01 -4.08206075e-01 -5.72470129e-01 -8.79322365e-03
-5.52540660e-01 -6.78095147e-02 6.34735763e-01 -7.72707224e-01
-2.77338028e-01 -2.58497804e-01 -3.38401586e-01 -3.66387635e-01
1.06200337e-01 2.13934138e-01 -9.59655046e-02 -1.09241927e+00
4.80759323e-01 5.65100551e-01 4.35615063e-01 1.31440699e-01
4.74202454e-01 7.26797938e-01 5.91399133e-01 2.59082019e-02
1.51850414e+00 8.43037426e-01 -4.04328369e-02 -6.19985104e-01
1.75443485e-01 4.54456657e-01 -2.18623221e-01 -7.32905045e-02
5.22021532e-01 -7.38596618e-01 -1.03164756e+00 1.03556514e+00
-1.07527566e+00 -1.56922787e-01 2.00968720e-02 7.54209161e-02
-3.10900658e-01 9.92352486e-01 -9.33779478e-01 -7.41385460e-01
-4.57180679e-01 -1.23053670e+00 4.86642599e-01 -3.72122318e-01
1.75101817e-01 -1.03625166e+00 -4.65649039e-01 2.84324080e-01
6.96326852e-01 9.59735513e-01 8.40585768e-01 -5.22281528e-01
-9.75115240e-01 -8.41716230e-01 -1.04822241e-01 7.17047215e-01
1.03760503e-01 -1.08965918e-01 -9.13868785e-01 -7.39001930e-01
3.40433389e-01 -2.38383859e-01 7.80424953e-01 -8.14873055e-02
9.83255506e-01 -1.29936028e+00 -8.21086615e-02 8.20987880e-01
1.56327558e+00 2.17186168e-01 1.34361327e+00 6.05233908e-01
4.66443658e-01 4.94455695e-01 -8.81456435e-02 2.08796680e-01
9.19644013e-02 2.49909952e-01 1.00466073e+00 -2.04586923e-01
1.44356251e-01 -1.73550248e-01 8.39717865e-01 7.16803670e-01
2.44160876e-01 -5.70646167e-01 -4.90693778e-01 4.21519309e-01
-1.19649267e+00 -1.03958571e+00 2.54922062e-01 2.14176726e+00
8.93283427e-01 4.16286826e-01 -2.35650748e-01 5.73536277e-01
7.59068608e-01 5.37068605e-01 -3.37406337e-01 -5.68983972e-01
-8.99112821e-02 5.99029288e-02 1.26831329e+00 6.35891199e-01
-1.11634862e+00 8.52869749e-01 6.87068558e+00 5.37370026e-01
-1.25848591e+00 -5.89031279e-02 5.55525422e-01 2.97660232e-01
-3.54791522e-01 -2.52530072e-02 -2.49624908e-01 7.30281770e-01
1.11770880e+00 -7.67860794e-03 7.12738156e-01 3.04195583e-01
-3.49516004e-01 2.93552667e-01 -7.31617391e-01 6.15005076e-01
2.48136073e-01 -1.56389999e+00 1.50954604e-01 6.13249123e-01
4.44050312e-01 -3.24133664e-01 5.31290114e-01 -2.72776991e-01
2.97783852e-01 -1.07796168e+00 6.76645815e-01 1.56866625e-01
9.35953140e-01 -9.30680513e-01 8.43520641e-01 2.01345310e-01
-8.25736344e-01 -2.52911925e-01 -3.11860651e-01 2.87273914e-01
3.72126140e-02 8.86954963e-02 -6.59883678e-01 2.67262638e-01
3.56661826e-01 3.09586316e-01 -5.19977987e-01 2.72208512e-01
-3.61588389e-01 6.53741777e-01 -2.93379605e-01 2.65916020e-01
4.19797838e-01 4.62158024e-01 7.62333870e-01 1.07344937e+00
8.29649270e-02 5.79675920e-02 -2.54959643e-01 6.60736263e-01
-5.81879079e-01 -3.60887766e-01 -1.06833196e+00 -2.41820544e-01
6.19859815e-01 6.48785114e-01 -5.35961151e-01 -2.61390686e-01
4.96343710e-03 1.39673996e+00 -5.55475414e-01 4.41855580e-01
-7.93755949e-01 -7.14282274e-01 5.68489254e-01 4.60875295e-02
4.86486733e-01 -2.48944536e-01 -2.11744147e-04 -8.62103879e-01
3.73816802e-05 -1.47984254e+00 -1.08775189e-02 -4.73727822e-01
-8.20145786e-01 4.84956056e-01 -4.51916963e-01 -1.10356474e+00
-2.83013254e-01 -4.49832618e-01 -4.38145012e-01 6.20721757e-01
-1.84371686e+00 -1.18219221e+00 9.61485803e-02 8.24304879e-01
-5.61634265e-02 -3.85260463e-01 1.03439200e+00 1.53586969e-01
-9.89376158e-02 9.39511895e-01 1.67978942e-01 7.76558518e-01
4.19802189e-01 -7.00150371e-01 8.85717452e-01 1.08135509e+00
-2.18102381e-01 6.56919658e-01 6.64928198e-01 -5.64891160e-01
-1.76318789e+00 -1.08136797e+00 6.45505548e-01 -2.92060133e-02
6.55974627e-01 -5.21815717e-01 -8.58134270e-01 9.21353817e-01
4.40777898e-01 1.79027036e-01 6.40804112e-01 -1.00621951e+00
-7.48227537e-01 6.84244409e-02 -1.67687547e+00 5.76940060e-01
3.56110573e-01 -9.64136362e-01 -1.70989171e-01 1.94452718e-01
1.12243915e+00 -3.51995826e-01 -8.90032887e-01 1.63604736e-01
8.28362525e-01 -9.54514503e-01 1.32584453e+00 -2.21259385e-01
5.94954669e-01 -1.38093770e-01 -4.09385145e-01 -7.56486535e-01
-1.25802532e-02 -9.92890060e-01 -4.52263653e-01 8.29800308e-01
1.19791470e-01 -1.07211781e+00 5.49601138e-01 3.18320870e-01
5.27222872e-01 -2.29764253e-01 -8.98500264e-01 -8.01552117e-01
1.16581760e-01 -2.99951494e-01 7.59375811e-01 1.12619579e+00
-2.81714916e-01 -6.26974106e-01 -9.06719029e-01 7.23428488e-01
9.37777817e-01 -5.20816207e-01 8.31633210e-01 -4.45443302e-01
-2.92385399e-01 3.73850140e-04 -9.40113306e-01 -7.19651997e-01
2.38435879e-01 -6.99023545e-01 -3.73089492e-01 -8.08575809e-01
-2.72539109e-01 -2.63739884e-01 -2.49659956e-01 5.15996695e-01
2.55607426e-01 9.18396056e-01 4.06169564e-01 3.96824270e-01
7.78809860e-02 7.99704939e-02 9.59697187e-01 -3.97606105e-01
4.23890531e-01 -1.78600863e-01 -7.43220389e-01 5.92880964e-01
7.08353400e-01 -8.14109921e-01 -1.87878668e-01 -3.39061469e-01
3.57069105e-01 9.70644224e-03 9.11108673e-01 -9.16950226e-01
8.09298232e-02 7.59674758e-02 2.93115109e-01 1.41246557e-01
3.04859936e-01 -1.24917996e+00 3.31059396e-01 1.18746281e+00
-4.68286335e-01 -1.81339651e-01 1.71705052e-01 6.55973077e-01
-3.33764330e-02 -2.80762583e-01 9.83882546e-01 -1.71659067e-01
-5.49088299e-01 1.16854914e-01 -5.75797677e-01 -3.33566487e-01
1.17339206e+00 -3.56637746e-01 -5.15633404e-01 -8.07966471e-01
-6.13079071e-01 -2.24294767e-01 8.53657961e-01 2.37988129e-01
1.04899526e+00 -1.25028884e+00 -6.89048588e-01 6.58502400e-01
-5.62897958e-02 -7.25266159e-01 -4.10947949e-02 2.36763343e-01
-9.48919952e-01 -4.51220535e-02 -3.22163850e-01 -2.18450814e-01
-1.49991035e+00 8.34015906e-01 5.51065445e-01 -1.86647847e-02
-7.40531743e-01 5.41504264e-01 -1.43140912e-01 1.06368437e-01
2.27584407e-01 1.96741298e-01 2.34875396e-01 -5.28899133e-01
9.73625124e-01 2.26839781e-01 -3.16381216e-01 -6.47432387e-01
-1.86561644e-01 3.75258386e-01 -2.01858521e-01 -1.43046319e-01
9.69798386e-01 -1.94272548e-01 -3.16953808e-01 -2.34501213e-01
1.73855376e+00 2.77643144e-01 -1.01879835e+00 -2.31423080e-01
-5.23490727e-01 -7.60623872e-01 1.42050937e-01 -5.49645543e-01
-1.19947922e+00 6.84407532e-01 7.96888649e-01 6.50557518e-01
1.03163111e+00 -4.01294798e-01 1.32755995e+00 3.92004877e-01
2.43534222e-01 -5.59208691e-01 3.71699478e-03 1.02797858e-01
6.30329788e-01 -1.10142088e+00 -5.05009703e-02 -3.87678295e-01
-3.25296909e-01 1.02458930e+00 -1.55211002e-01 -2.49043450e-01
7.15926409e-01 5.46598494e-01 1.96181253e-01 -2.63009127e-02
-3.96045119e-01 5.27145326e-01 -9.54553410e-02 8.21419716e-01
-3.15833628e-01 -5.73871210e-02 2.19108209e-01 -1.96226835e-01
-4.61155288e-02 -1.94747120e-01 8.51520360e-01 1.23113763e+00
-4.03046042e-01 -1.12263811e+00 -7.10027874e-01 -1.81875914e-01
-9.12490845e-01 -1.99858353e-01 -4.09829348e-01 7.82842040e-01
-6.80998489e-02 9.25167024e-01 -1.82394400e-01 -6.22879803e-01
-1.64820790e-01 -3.14281195e-01 1.49255976e-01 2.60743243e-03
-6.68796480e-01 -1.36631310e-01 -2.11437359e-01 -5.17311573e-01
-4.27604139e-01 4.13035974e-03 -9.40601885e-01 -1.04919958e+00
-3.17145318e-01 -2.49585956e-01 8.88083756e-01 6.36830747e-01
3.54543477e-01 2.65364200e-01 1.29864085e+00 -4.88191217e-01
-6.20850623e-01 -4.42210197e-01 -4.01012778e-01 2.78691798e-01
9.30685520e-01 1.05722293e-01 -8.16357076e-01 3.85000110e-01] | [4.481238842010498, 8.001784324645996] |
8161dc44-440f-4a75-944a-ccab66041e5d | exploring-multimodal-sentiment-analysis-via | 2303.14708 | null | https://arxiv.org/abs/2303.14708v1 | https://arxiv.org/pdf/2303.14708v1.pdf | Exploring Multimodal Sentiment Analysis via CBAM Attention and Double-layer BiLSTM Architecture | Because multimodal data contains more modal information, multimodal sentiment analysis has become a recent research hotspot. However, redundant information is easily involved in feature fusion after feature extraction, which has a certain impact on the feature representation after fusion. Therefore, in this papaer, we propose a new multimodal sentiment analysis model. In our model, we use BERT + BiLSTM as new feature extractor to capture the long-distance dependencies in sentences and consider the position information of input sequences to obtain richer text features. To remove redundant information and make the network pay more attention to the correlation between image and text features, CNN and CBAM attention are added after splicing text features and picture features, to improve the feature representation ability. On the MVSA-single dataset and HFM dataset, compared with the baseline model, the ACC of our model is improved by 1.78% and 1.91%, and the F1 value is enhanced by 3.09% and 2.0%, respectively. The experimental results show that our model achieves a sound effect, similar to the advanced model. | ['chunming Ma', 'Dan Yang', 'Zenyu Ren', 'Xiuhong Li', 'Huiru Wang'] | 2023-03-26 | null | null | null | null | ['multimodal-sentiment-analysis', 'multimodal-sentiment-analysis'] | ['computer-vision', 'natural-language-processing'] | [ 1.45580634e-01 -3.49752069e-01 1.02547169e-01 -6.46526337e-01
-6.89118147e-01 -2.01973811e-01 4.05843318e-01 -4.31848429e-02
-7.52658188e-01 3.79290730e-01 4.99601722e-01 2.45367885e-01
3.48045111e-01 -4.77069348e-01 -4.52026308e-01 -8.55380714e-01
5.05220115e-01 -3.67033005e-01 1.70502067e-01 -4.55693364e-01
1.85014904e-01 -2.53072351e-01 -1.31823421e+00 7.38749027e-01
9.85852838e-01 1.29079258e+00 4.58252788e-01 4.18480873e-01
-4.16647911e-01 7.44089842e-01 -5.26271284e-01 -6.67239428e-01
-3.33794624e-01 -6.10640943e-01 -3.48943979e-01 1.45007089e-01
2.76411697e-02 -3.78705502e-01 -5.07822275e-01 1.20495689e+00
7.52568066e-01 9.96659100e-02 4.44591284e-01 -1.10012710e+00
-6.36379838e-01 9.02419567e-01 -9.63243484e-01 2.43303981e-02
1.95122018e-01 1.55214220e-01 8.95408690e-01 -1.10786259e+00
1.28649205e-01 1.49100220e+00 3.65101188e-01 3.40506911e-01
-7.39290595e-01 -7.28727281e-01 4.38658863e-01 4.95294541e-01
-1.26893020e+00 -4.53508168e-01 8.42081368e-01 6.52658194e-02
7.15656042e-01 1.80813208e-01 4.70977515e-01 1.09522951e+00
2.54597366e-01 1.33845007e+00 8.17253590e-01 -2.94459254e-01
-5.20229101e-01 1.58799767e-01 1.06490985e-01 7.20108867e-01
-1.08194992e-01 -4.16498482e-01 -6.91069543e-01 4.27518845e-01
2.89795637e-01 4.09370244e-01 -2.36648366e-01 3.05582017e-01
-1.44638824e+00 6.91741467e-01 5.98972142e-01 4.06553626e-01
-1.50508523e-01 1.92253087e-02 7.45665491e-01 1.72321528e-01
3.50651085e-01 -9.19013843e-02 -4.82374519e-01 -3.79676431e-01
-4.17251825e-01 -9.86380577e-02 1.01275183e-01 7.94869006e-01
7.14061797e-01 7.52515048e-02 -2.79941112e-01 9.77441251e-01
6.23154759e-01 9.13585305e-01 8.59539390e-01 -5.19976258e-01
8.93893659e-01 8.56393933e-01 -4.23927844e-01 -1.36852121e+00
-5.31333745e-01 -2.83476859e-01 -1.20128787e+00 -5.13767421e-01
-2.09780291e-01 -4.03957427e-01 -9.30701494e-01 1.64507663e+00
2.89285369e-02 -1.46736279e-01 2.42004007e-01 9.96655285e-01
1.14724028e+00 1.05395675e+00 -1.60137340e-02 -3.64354700e-01
1.44973624e+00 -1.12288868e+00 -1.04165399e+00 -1.66201934e-01
6.45162404e-01 -1.06066787e+00 1.11634338e+00 4.07823384e-01
-7.93388844e-01 -8.64897549e-01 -1.03617299e+00 -1.99153632e-01
-2.22156063e-01 3.24995607e-01 3.24442208e-01 2.56986588e-01
-4.88496840e-01 2.26329371e-01 -6.10139608e-01 -6.68678954e-02
5.27386308e-01 4.07403231e-01 -3.32817525e-01 -2.40199819e-01
-1.43412614e+00 6.77391529e-01 5.32958448e-01 6.42430902e-01
-3.21193010e-01 -1.98387355e-01 -1.00156796e+00 1.02509275e-01
3.36454749e-01 -4.42795128e-01 1.21925330e+00 -1.06539977e+00
-1.41199458e+00 2.39588544e-01 -4.89602298e-01 6.09397553e-02
1.54261291e-01 -2.94980377e-01 -6.94495618e-01 1.52473718e-01
-3.25501710e-02 8.06360900e-01 7.10838497e-01 -9.51921999e-01
-9.53796685e-01 -4.27794486e-01 -3.04528745e-03 4.73963439e-01
-6.89446568e-01 4.73837927e-02 -6.63664937e-01 -5.91335058e-01
4.02964801e-02 -7.65619874e-01 -7.97858909e-02 -5.91242313e-01
-4.74342585e-01 -5.59539795e-02 8.71346176e-01 -8.78960133e-01
1.32221329e+00 -2.36559224e+00 4.24508005e-01 2.46201009e-02
6.01854548e-02 1.80298641e-01 -4.19471294e-01 2.36335322e-01
1.50789350e-01 2.65538484e-01 -3.24907213e-01 -5.37938535e-01
-4.55674157e-02 1.20143734e-01 -1.18198454e-01 -1.74435601e-02
4.91174996e-01 1.09399831e+00 -4.69591618e-01 -6.70961380e-01
1.83507949e-01 5.60865581e-01 -3.25769305e-01 9.94853154e-02
8.28308389e-02 5.44230282e-01 -7.29502738e-01 5.10157764e-01
8.53524625e-01 -7.27787763e-02 -1.60102069e-01 -6.61141396e-01
-9.88776758e-02 3.32222395e-02 -1.02652836e+00 1.96637130e+00
-5.85273862e-01 5.58929384e-01 -1.57447740e-01 -8.83656561e-01
8.95951927e-01 2.32986569e-01 1.50562719e-01 -1.11358893e+00
5.41220486e-01 1.18480101e-02 1.21363975e-01 -8.45339775e-01
5.65683722e-01 -9.24783871e-02 -2.85237193e-01 3.63129564e-02
6.04836643e-02 2.80943722e-01 6.56912997e-02 1.82848990e-01
5.02232492e-01 -1.45159336e-03 -9.06744376e-02 1.28240764e-01
1.02208388e+00 -3.33474636e-01 6.80515826e-01 6.94359541e-02
5.80455139e-02 5.61579525e-01 5.04821360e-01 -1.79732591e-01
-6.96173310e-01 -3.79385114e-01 1.88926961e-02 1.00569689e+00
5.09733021e-01 -5.07689178e-01 -5.55805326e-01 -7.73685575e-01
-3.50290895e-01 3.12241346e-01 -6.25912428e-01 -6.56833410e-01
-4.79683101e-01 -9.91186619e-01 3.18028122e-01 6.13507509e-01
1.08150983e+00 -1.08778894e+00 -1.22602247e-02 1.06716581e-01
-6.59393311e-01 -1.09092557e+00 -6.60645366e-01 -2.31302217e-01
-7.55963027e-01 -5.98117650e-01 -7.17761219e-01 -8.70004654e-01
5.60599864e-01 3.45950544e-01 5.55318654e-01 2.45564654e-01
1.15993835e-01 -9.81098488e-02 -8.29446256e-01 -3.98098797e-01
5.12258187e-02 2.68747300e-01 -3.02490681e-01 4.24260795e-01
3.70884985e-01 -2.50657648e-01 -5.59646070e-01 1.80441275e-01
-1.03608811e+00 3.35531443e-01 9.83837545e-01 1.14811492e+00
3.42942625e-01 -4.19389755e-02 6.68394387e-01 -5.29711664e-01
5.13436437e-01 -2.80344486e-01 -2.34210968e-01 4.87836003e-02
-3.21425855e-01 -4.11119461e-02 7.91501641e-01 -4.22486931e-01
-1.49465942e+00 1.12094581e-01 -3.41618776e-01 -1.68950662e-01
-9.42211505e-03 1.02670944e+00 -7.82837212e-01 2.92743742e-01
-1.30124435e-01 4.94253844e-01 -5.49338292e-03 -5.47810793e-01
2.60643423e-01 8.70028675e-01 3.14443886e-01 -4.77196649e-02
5.05490959e-01 2.31715009e-01 -1.10236488e-01 -5.96658409e-01
-9.94534969e-01 -1.88723698e-01 -5.69254875e-01 -4.55436967e-02
1.02210450e+00 -1.07700241e+00 -1.03383958e+00 7.81840503e-01
-1.40966821e+00 2.83720821e-01 2.07755163e-01 7.69879401e-01
-1.62873119e-01 6.73156083e-01 -6.91575348e-01 -6.36215448e-01
-4.66809958e-01 -1.45451283e+00 1.07389891e+00 5.86313784e-01
2.05213770e-01 -6.66514218e-01 -2.63681084e-01 5.09910882e-01
2.83020347e-01 -4.10873555e-02 7.05419183e-01 -6.05505586e-01
-3.14388305e-01 -2.49838650e-01 -3.95025134e-01 6.15733743e-01
9.58179384e-02 2.53948141e-02 -9.74362254e-01 2.06421856e-02
6.80048615e-02 -2.94153035e-01 1.42070663e+00 9.66660902e-02
1.23684132e+00 -1.07049339e-01 -2.07015932e-01 4.29926634e-01
9.87699568e-01 3.57120991e-01 6.79798543e-01 2.75962830e-01
1.03408194e+00 5.59469283e-01 8.37328792e-01 3.80064964e-01
6.37075424e-01 3.77854228e-01 4.95657444e-01 -1.25013337e-01
9.71273854e-02 -1.50593802e-01 7.03112364e-01 1.48842299e+00
9.46549624e-02 -2.85215795e-01 -6.43620670e-01 2.99333453e-01
-2.18620372e+00 -8.40605259e-01 -4.14994299e-01 1.72508466e+00
7.20961213e-01 9.79146585e-02 -1.51820704e-01 4.26785573e-02
7.83099353e-01 3.03627193e-01 -4.20493335e-01 -7.07766935e-02
-4.56673115e-01 -3.15994710e-01 3.27170528e-02 2.31559634e-01
-1.25186110e+00 7.93218255e-01 4.87536240e+00 1.00870705e+00
-1.17727661e+00 -2.43266374e-02 7.17702091e-01 -1.78568304e-01
-2.03974247e-01 -2.43863165e-01 -8.10405254e-01 9.52957809e-01
7.79743135e-01 3.78351063e-02 1.70843482e-01 4.45365280e-01
1.53850496e-01 2.62254905e-02 -6.35926843e-01 1.26966584e+00
3.78488988e-01 -9.50132310e-01 2.36059994e-01 -4.10740450e-02
5.61925113e-01 -2.40224227e-01 2.35658213e-01 5.14478147e-01
-2.49752462e-01 -8.62601936e-01 5.93962967e-01 7.64161170e-01
4.15020555e-01 -1.31130600e+00 1.52925622e+00 2.60584980e-01
-1.36686718e+00 -1.09097145e-01 -4.37775224e-01 2.67103054e-02
3.98757935e-01 6.50447607e-01 -1.92231089e-01 1.01001072e+00
6.94178104e-01 1.18485665e+00 -7.92952418e-01 8.66742134e-01
-1.83858603e-01 4.27466810e-01 -8.38177353e-02 -3.08030009e-01
2.96187520e-01 -1.71574816e-01 1.28310189e-01 1.32857203e+00
3.85660261e-01 5.01182452e-02 1.93629503e-01 3.15800518e-01
-2.94994473e-01 3.80052656e-01 -2.07156137e-01 -1.68247253e-01
1.18513905e-01 1.67925584e+00 -3.39786440e-01 -4.41482693e-01
-6.34010851e-01 1.15648651e+00 -5.94920304e-04 2.95615047e-01
-1.01424396e+00 -8.16414475e-01 3.22621673e-01 -7.10755467e-01
5.30179203e-01 -1.56804081e-02 -6.92270696e-02 -1.46940637e+00
1.95856571e-01 -9.33841527e-01 2.67637581e-01 -9.61030900e-01
-1.24932170e+00 6.03988230e-01 -5.17633021e-01 -1.26739514e+00
1.89821739e-02 -5.89024842e-01 -5.87271035e-01 9.58601832e-01
-1.54326522e+00 -1.52411151e+00 -3.31271321e-01 5.43385029e-01
5.57909131e-01 -2.00864509e-01 4.35735464e-01 5.88646650e-01
-8.75143051e-01 6.52510941e-01 9.25035626e-02 3.77116114e-01
8.85074556e-01 -7.41201639e-01 3.78310829e-02 8.72631311e-01
-1.00754112e-01 4.91509646e-01 1.72136262e-01 -4.15162832e-01
-1.53558457e+00 -1.14701641e+00 6.73704863e-01 -2.63730407e-01
4.88429844e-01 -2.99492121e-01 -9.60596025e-01 5.84674120e-01
5.74963808e-01 -3.59049410e-01 7.63693392e-01 -3.07268202e-02
-3.13427299e-01 -3.90516222e-01 -6.28761411e-01 6.86694562e-01
6.52505755e-01 -4.61719155e-01 -5.42924643e-01 -1.27472132e-01
1.23144948e+00 -1.77551985e-01 -8.42314005e-01 6.36608064e-01
5.42763889e-01 -7.41212606e-01 5.75056732e-01 -4.21007812e-01
7.55586267e-01 -4.87634480e-01 -3.80910456e-01 -1.33607769e+00
-2.87259310e-01 -1.42204568e-01 9.58369300e-02 1.60906267e+00
5.68889499e-01 -4.32548404e-01 2.31695488e-01 8.29772577e-02
-3.39391977e-01 -7.50885308e-01 -8.05505693e-01 -2.67780244e-01
-1.70621529e-01 -4.53516036e-01 6.96786106e-01 9.02229488e-01
1.60465732e-01 1.02606142e+00 -6.17518485e-01 -8.11610669e-02
8.48135203e-02 2.61827737e-01 6.65321946e-01 -1.04590166e+00
2.87710689e-02 -5.73670328e-01 -2.69115210e-01 -1.26648283e+00
1.17608525e-01 -7.86275983e-01 8.96042958e-03 -1.43355358e+00
5.64134717e-01 3.10136914e-01 -5.44648647e-01 4.96016592e-01
-5.64378083e-01 2.45110273e-01 3.77835631e-01 -1.23049915e-01
-9.47539568e-01 1.21388900e+00 1.60076118e+00 -2.68723577e-01
5.75150885e-02 -3.98819029e-01 -9.32130039e-01 7.97734439e-01
7.80008018e-01 -1.35918155e-01 -2.01328114e-01 -6.56772554e-01
4.64725405e-01 -9.72762182e-02 1.14983782e-01 -6.35265231e-01
3.85571748e-01 -1.25209391e-02 8.32702756e-01 -9.41568375e-01
5.26343346e-01 -8.63778651e-01 -4.55846280e-01 3.64108980e-01
-3.65884632e-01 2.17731908e-01 3.51988316e-01 4.76410866e-01
-6.51747346e-01 -1.48781344e-01 2.81013310e-01 1.24723122e-01
-8.36530685e-01 3.40617865e-01 -3.38761002e-01 -2.48941585e-01
5.88475883e-01 1.54335648e-01 -4.72409368e-01 -4.62517381e-01
-5.20720780e-01 6.81979537e-01 7.03475773e-02 6.90071702e-01
8.85679901e-01 -1.67793882e+00 -7.00786650e-01 1.79518968e-01
1.86252609e-01 7.77134523e-02 8.35631788e-01 1.12401116e+00
7.51459673e-02 2.44041800e-01 -4.65684533e-02 -5.22638261e-01
-1.40080452e+00 4.04833406e-01 4.96408530e-02 2.67220032e-03
-1.66824147e-01 6.62036479e-01 2.73062319e-01 -4.45092738e-01
-1.06330499e-01 -2.65320778e-01 -6.22661531e-01 3.24630350e-01
7.83052802e-01 1.43188298e-01 -1.31451353e-01 -8.93283129e-01
-3.57519239e-01 8.42085183e-01 -3.44218731e-01 -7.30651543e-02
1.15609050e+00 -4.94758070e-01 -4.07179534e-01 6.82616472e-01
1.45521522e+00 1.10803097e-01 -8.76718700e-01 -2.96424627e-01
-3.57784241e-01 -2.49438152e-01 1.43306896e-01 -7.69423246e-01
-1.36607623e+00 1.36646152e+00 4.70432818e-01 5.69966324e-02
1.31065965e+00 -1.42308876e-01 1.08077562e+00 5.21400094e-01
-2.91614354e-01 -1.03727639e+00 3.29866141e-01 7.60929346e-01
8.17551076e-01 -1.53107083e+00 -2.22289771e-01 -3.44157934e-01
-1.07453692e+00 1.13781786e+00 7.52601802e-01 1.72758073e-01
3.27985793e-01 -1.44290421e-02 1.67602569e-01 5.81723042e-02
-5.90406120e-01 -2.24579945e-01 4.80767310e-01 1.76637992e-01
4.74892884e-01 -2.65420705e-01 -3.19940865e-01 1.32212722e+00
-8.91857967e-02 -2.56864756e-01 3.60771954e-01 7.32646883e-01
-6.18502736e-01 -9.50337470e-01 -1.57691658e-01 4.23063099e-01
-4.94598746e-01 -4.04961795e-01 -2.00457752e-01 6.05718255e-01
1.88753679e-01 1.24012423e+00 1.14615239e-01 -9.55311179e-01
4.55132842e-01 5.91861606e-02 1.78906381e-01 -2.71222830e-01
-3.86603773e-01 5.35884261e-01 -1.33414874e-02 -3.32162440e-01
-7.10175991e-01 -4.98343587e-01 -1.56306314e+00 -1.12817965e-01
-5.55860996e-01 1.41149879e-01 6.26797020e-01 1.11355698e+00
5.51416814e-01 9.38858569e-01 9.97231066e-01 -5.98880172e-01
-1.59095153e-01 -1.15850830e+00 -4.45587158e-01 3.59883487e-01
3.10787499e-01 -5.00993669e-01 -2.15789586e-01 1.75452784e-01] | [13.133956909179688, 5.016683101654053] |
22aaaa74-e222-4448-bd31-95d2baa6695a | on-the-importance-of-sign-labeling-the | 2302.10768 | null | https://arxiv.org/abs/2302.10768v2 | https://arxiv.org/pdf/2302.10768v2.pdf | On the Importance of Sign Labeling: The Hamburg Sign Language Notation System Case Study | Labeling is the cornerstone of supervised machine learning, which has been exploited in a plethora of various applications, with sign language recognition being one of them. However, such algorithms must be fed with a huge amount of consistently labeled data during the training process to elaborate a well-generalizing model. In addition, there is a great need for an automated solution that works with any nationally diversified sign language. Although there are language-agnostic transcription systems, such as the Hamburg Sign Language Notation System (HamNoSys) that describe the signer's initial position and body movement instead of the glosses' meanings, there are still issues with providing accurate and reliable labels for every real-world use case. In this context, the industry relies heavily on manual attribution and labeling of the available video data. In this work, we tackle this issue and thoroughly analyze the HamNoSys labels provided by various maintainers of open sign language corpora in five sign languages, in order to examine the challenges encountered in labeling video data. We also investigate the consistency and objectivity of HamNoSys-based labels for the purpose of training machine learning models. Our findings provide valuable insights into the limitations of the current labeling methods and pave the way for future research on developing more accurate and efficient solutions for sign language recognition. | ['Jakub Nalepa', 'Milena Olech', 'Agnieszka Mikołajczyk-Bareła', 'Alicja Kwaśniwska', 'Marta Plantykow', 'Sylwia Majchrowska', 'Maria Ferlin'] | 2023-01-19 | null | null | null | null | ['sign-language-recognition'] | ['computer-vision'] | [ 1.63661346e-01 -3.33310604e-01 -4.86915171e-01 -5.40099859e-01
-5.59917688e-01 -7.38052905e-01 5.21747231e-01 -5.16738057e-01
-5.14477074e-01 5.15723467e-01 4.33149815e-01 -1.87810495e-01
-1.29207864e-01 -2.31703475e-01 -2.46708721e-01 -6.43354714e-01
3.19050848e-01 4.17379618e-01 1.65193424e-01 -1.61804736e-01
1.59676626e-01 4.49700177e-01 -1.84819770e+00 2.17261106e-01
5.75491667e-01 7.80588150e-01 -1.66029707e-02 4.49895173e-01
-3.61446530e-01 1.00123155e+00 -6.99898303e-01 -5.70653021e-01
3.13427895e-01 -6.23244941e-01 -8.19393516e-01 4.01532829e-01
7.96282768e-01 -3.51041228e-01 -1.70179710e-01 8.72517884e-01
3.19800347e-01 -1.10286124e-01 6.83190882e-01 -1.32920837e+00
-4.31857556e-01 5.05579948e-01 -1.54764324e-01 -2.83093572e-01
2.28467271e-01 2.52699733e-01 1.05865085e+00 -3.03009510e-01
1.08942604e+00 6.42902136e-01 5.46901643e-01 7.23929524e-01
-5.46552956e-01 -8.90346348e-01 1.94380149e-01 3.92410696e-01
-1.29021740e+00 -2.86129057e-01 6.93419278e-01 -8.12378824e-01
3.93215239e-01 2.62596875e-01 7.43866682e-01 1.14417207e+00
-4.76023614e-01 8.60770047e-01 1.26778269e+00 -7.89286971e-01
1.57875437e-02 3.27218026e-02 3.86841387e-01 6.63927674e-01
2.76827335e-01 -1.57983035e-01 -6.05468571e-01 4.55454066e-02
7.63894796e-01 -1.98071539e-01 -3.69704396e-01 -4.41362023e-01
-1.06498075e+00 4.49169546e-01 5.27636744e-02 7.50806153e-01
-2.32913598e-01 1.65694177e-01 6.83967471e-01 2.71827400e-01
-1.46663234e-01 3.76615882e-01 -3.83388221e-01 -7.30259895e-01
-9.37329054e-01 6.69302642e-02 8.35878253e-01 1.02908349e+00
4.09795195e-01 -1.39698107e-02 1.67416826e-01 1.01720846e+00
4.27791178e-01 4.63732600e-01 5.99375188e-01 -1.17477417e+00
2.46068582e-01 7.44076729e-01 1.11421034e-01 -8.53170216e-01
-2.42946371e-01 -1.72531575e-01 -4.26727861e-01 3.68143797e-01
9.44272995e-01 2.43016705e-03 -1.20073617e+00 1.65546465e+00
-8.67348444e-03 -2.06654761e-02 -9.47009996e-02 1.09270155e+00
7.33994186e-01 2.06571639e-01 2.83686161e-01 8.86666477e-02
1.27350569e+00 -8.77061009e-01 -8.34998131e-01 -3.26451696e-02
9.52563763e-01 -8.41905892e-01 1.20204973e+00 5.50810039e-01
-4.43670809e-01 -2.32117221e-01 -8.76309812e-01 -5.76712377e-02
-3.97039205e-01 5.39043903e-01 6.55779183e-01 8.82794857e-01
-5.84683955e-01 1.22537278e-01 -7.98834741e-01 -8.35150123e-01
2.30905280e-01 1.43231213e-01 -5.26220381e-01 -2.99430609e-01
-8.50904346e-01 1.19485247e+00 1.72741055e-01 3.67517829e-01
-3.57404023e-01 -6.73408732e-02 -8.16029429e-01 -5.49554884e-01
3.54671121e-01 -7.05095530e-02 1.52464294e+00 -1.23541451e+00
-1.50337243e+00 1.25354350e+00 -9.33520123e-02 -3.05565167e-02
6.23955727e-01 -1.35172561e-01 -6.65734470e-01 -9.03669298e-02
-1.83009487e-02 3.40878665e-01 6.74781382e-01 -1.26569045e+00
-8.61641824e-01 -7.75660798e-02 8.30100700e-02 -1.12982363e-01
-1.93998680e-01 4.84288901e-01 -5.57544291e-01 -6.49130821e-01
2.63948917e-01 -1.04052532e+00 1.31532326e-01 2.31846049e-01
-1.63725894e-02 -1.14071950e-01 8.36225867e-01 -1.07458448e+00
1.41007566e+00 -2.18447733e+00 5.55586182e-02 3.08625132e-01
-1.12097310e-02 7.26540744e-01 -4.49005291e-02 5.10248542e-01
2.25485981e-01 -8.38322341e-02 -3.46839428e-01 -7.96973035e-02
1.90168172e-01 7.81709731e-01 -3.57888103e-01 4.26906645e-01
-1.74095437e-01 6.89718723e-01 -9.97359395e-01 -5.68935990e-01
3.86125803e-01 5.32880008e-01 -2.54226744e-01 1.51782827e-02
-1.04594141e-01 6.77761436e-01 -2.15060934e-01 8.87000024e-01
5.50784208e-02 -1.75164729e-01 3.46252561e-01 -1.37470156e-01
-2.16325045e-01 1.63590118e-01 -1.24591434e+00 1.54588342e+00
-4.42399323e-01 9.84171987e-01 -6.90322593e-02 -8.02260876e-01
8.11367989e-01 5.26324630e-01 5.47677755e-01 -3.43055367e-01
3.57831270e-01 7.42349803e-01 8.18851963e-02 -9.57849324e-01
4.78003740e-01 -1.39118195e-01 -1.35575384e-01 5.72495222e-01
6.18500262e-02 -3.13094221e-02 5.41208684e-01 -1.88892648e-01
8.52817595e-01 4.11981821e-01 3.55892479e-01 3.15960467e-01
5.37141383e-01 2.44879633e-01 4.72829878e-01 3.56782347e-01
-4.98691142e-01 6.88556790e-01 9.84698161e-02 -4.37124997e-01
-9.05947924e-01 -5.90604007e-01 -6.32482693e-02 1.01236141e+00
-1.80432320e-01 -5.28715789e-01 -6.29345953e-01 -6.89156830e-01
-2.94877529e-01 4.87391740e-01 -2.18837976e-01 2.66918391e-01
-7.66240001e-01 -1.38204202e-01 9.54368830e-01 6.21521533e-01
5.46707749e-01 -1.15038204e+00 -1.04418790e+00 2.82245018e-02
-3.98903251e-01 -1.18498254e+00 -3.52591246e-01 -8.57297182e-02
-4.93024766e-01 -1.42212653e+00 -8.69784415e-01 -1.05520117e+00
7.98295796e-01 -2.33767126e-02 5.73699892e-01 2.12670505e-01
-4.82195765e-02 6.00166976e-01 -9.57964003e-01 -2.52105951e-01
-6.66255713e-01 1.51167959e-02 -1.78363129e-01 6.78685680e-02
6.46260619e-01 -1.83938503e-01 5.29046915e-02 6.63761437e-01
-1.01801717e+00 9.86844972e-02 5.42699397e-01 7.84806907e-01
2.12340862e-01 -2.02655122e-01 2.27752507e-01 -5.93678474e-01
4.55231637e-01 2.83991173e-02 -5.90669036e-01 5.35170078e-01
-2.83083200e-01 5.32582738e-02 3.56263459e-01 -5.02860188e-01
-8.56949985e-01 2.06368566e-01 -1.93954438e-01 -1.00061603e-01
-4.72613484e-01 6.82844102e-01 -4.10885029e-02 -3.22773695e-01
5.81043363e-01 2.57843643e-01 -1.98051147e-02 -5.30543327e-01
5.31518340e-01 1.24322021e+00 7.87693262e-01 -4.81376350e-01
5.72570443e-01 4.18409288e-01 -1.77843377e-01 -1.07704306e+00
-6.65464699e-01 -8.46385777e-01 -7.48765528e-01 -7.46188581e-01
6.94668829e-01 -4.73225117e-01 -3.06133002e-01 8.44009101e-01
-9.53760862e-01 -3.11380029e-01 -3.09406281e-01 6.78380609e-01
-6.82153046e-01 6.35656655e-01 -3.65536243e-01 -7.77841210e-01
1.62466139e-01 -1.11432087e+00 8.28564882e-01 5.55427670e-02
-7.85032332e-01 -5.64832807e-01 9.88123491e-02 8.15302551e-01
3.02901745e-01 1.43727750e-01 6.32200480e-01 -4.02522802e-01
-4.25031245e-01 -5.51358104e-01 -1.42328739e-01 7.81756401e-01
4.43062246e-01 1.82663232e-01 -8.61333311e-01 5.69745339e-02
-3.87779236e-01 -4.28723812e-01 4.82981414e-01 2.29851469e-01
5.04671216e-01 -1.59800932e-01 1.08193688e-01 3.44957918e-01
1.07726538e+00 2.12174237e-01 5.05015194e-01 3.73950422e-01
6.45115852e-01 8.08183491e-01 5.81022441e-01 2.16321394e-01
4.20311213e-01 7.53751814e-01 5.37608527e-02 2.71521956e-01
-5.29922724e-01 -4.55353022e-01 2.92435646e-01 1.10279250e+00
-6.27901971e-01 -6.52213544e-02 -1.06855452e+00 5.74908376e-01
-1.91291070e+00 -1.11382663e+00 -1.81004465e-01 2.10380292e+00
6.65205598e-01 -4.58424121e-01 1.42457664e-01 6.07573450e-01
5.71224153e-01 5.84405474e-02 -1.44826278e-01 -2.00451031e-01
-1.61393940e-01 1.78429913e-02 5.55772245e-01 2.41585791e-01
-1.21243823e+00 1.05853248e+00 6.42078304e+00 3.95764887e-01
-1.54334438e+00 5.08050695e-02 -1.16581470e-01 2.08708242e-01
2.94100165e-01 -8.96674022e-02 -5.72107196e-01 5.24578810e-01
5.48806250e-01 2.20455736e-01 2.76088417e-01 8.91886771e-01
2.98602700e-01 -1.05509974e-01 -9.33756948e-01 1.27639771e+00
3.66584361e-01 -9.45873916e-01 -3.76946814e-02 6.48696497e-02
8.04960549e-01 -5.97302057e-02 -3.47675145e-01 1.46472439e-01
2.08144128e-01 -7.14681804e-01 9.95476484e-01 3.84519517e-01
8.45878482e-01 -3.03561036e-02 8.22542429e-01 2.55444676e-01
-1.06162262e+00 -7.50905648e-02 1.44651651e-01 -3.32685977e-01
6.80600584e-01 2.75541488e-02 -5.93366206e-01 3.02637637e-01
4.33902234e-01 6.40832186e-01 -2.38887176e-01 1.43327153e+00
-7.32927263e-01 8.49990606e-01 -5.25775015e-01 -1.11305162e-01
2.74365813e-01 -2.94107348e-01 2.08897904e-01 1.14731753e+00
3.17622125e-01 4.95798327e-02 3.37956250e-02 2.49465361e-01
1.84042647e-01 3.70441973e-01 -4.13202971e-01 -5.03919899e-01
1.80263370e-01 7.25693762e-01 -7.87922680e-01 -3.05706888e-01
-6.62101150e-01 9.35904086e-01 -1.33689672e-01 3.81627232e-01
-6.32483959e-01 -5.32766618e-02 5.96428156e-01 2.07688674e-01
1.45595729e-01 -6.16958022e-01 -2.39493251e-01 -1.16415298e+00
3.52254122e-01 -1.05911911e+00 5.01011908e-01 -6.90578401e-01
-1.18175387e+00 3.45779121e-01 5.28735816e-02 -1.68642783e+00
-4.90349233e-01 -9.33960617e-01 -6.02186806e-02 4.49084610e-01
-1.52539313e+00 -1.56253910e+00 -5.51047981e-01 4.44501132e-01
3.77830267e-01 -4.97624353e-02 8.58211100e-01 7.25964248e-01
-2.54884183e-01 4.27166790e-01 -1.28646731e-01 6.06246948e-01
8.90370131e-01 -7.01244950e-01 -9.55455974e-02 8.24204206e-01
6.28574610e-01 5.71850657e-01 6.72882199e-01 -5.96797645e-01
-9.77560699e-01 -7.31786668e-01 1.30324125e+00 -4.46815968e-01
7.69814551e-01 1.62909821e-01 -5.26833415e-01 8.57685328e-01
-1.58087879e-01 -1.46064043e-01 8.17203999e-01 -2.97131464e-02
-5.65346003e-01 1.54627517e-01 -7.80528724e-01 5.96927285e-01
1.20877695e+00 -6.76600873e-01 -8.83659244e-01 2.54830599e-01
-2.08937228e-01 -4.80543435e-01 -5.04813552e-01 1.35848626e-01
1.09244001e+00 -7.19581783e-01 3.76489431e-01 -6.04354322e-01
2.61948496e-01 -5.70655346e-01 -2.49456003e-01 -8.67601812e-01
2.09790334e-01 -3.05842191e-01 2.67920464e-01 1.15028191e+00
2.23606884e-01 -5.80306530e-01 8.57149124e-01 1.02235985e+00
-2.61142135e-01 -2.66343623e-01 -9.93412137e-01 -1.07557845e+00
-3.82267982e-01 -9.32218075e-01 3.14938754e-01 1.11717021e+00
1.89203113e-01 -1.66906729e-01 -7.01459944e-01 -1.05940759e-01
3.32167834e-01 -6.59867749e-02 1.08118749e+00 -1.14440095e+00
-1.01941943e-01 -6.04332328e-01 -9.96850193e-01 -1.06852281e+00
1.40452877e-01 -7.64376700e-01 6.16013929e-02 -1.61826205e+00
-1.98058769e-01 -3.50417644e-01 -1.74242053e-02 8.19278121e-01
3.99586588e-01 3.94895315e-01 4.62329984e-01 4.84346122e-01
-3.97605151e-01 9.80279893e-02 1.01918936e+00 -1.07693195e-01
-1.83308259e-01 8.24550807e-04 -3.10779423e-01 1.01107574e+00
7.46151924e-01 -2.64597178e-01 -1.40661672e-01 -6.46965444e-01
-2.30137329e-03 -6.04112566e-01 4.17530924e-01 -9.74521518e-01
3.92476678e-01 -2.79162318e-01 -3.76561344e-01 -3.08213413e-01
2.54785538e-01 -1.16738391e+00 3.34479362e-01 2.65782595e-01
-2.15896070e-01 -3.84622872e-01 -2.56120741e-01 2.27384996e-02
-5.59142590e-01 -4.84787405e-01 6.76267982e-01 -1.18991517e-01
-1.37987769e+00 -8.76508057e-02 -4.91428733e-01 7.82634914e-02
1.04423153e+00 -7.20920503e-01 -1.11867614e-01 -5.78912437e-01
-4.93803263e-01 5.40810637e-02 6.53351307e-01 5.48301816e-01
2.69592255e-01 -1.25708473e+00 -3.82600844e-01 2.88897634e-01
5.79415619e-01 -3.16167712e-01 4.75368202e-02 6.35814309e-01
-9.09548819e-01 5.01663446e-01 -3.19340855e-01 -4.68622148e-01
-1.39084256e+00 -6.14918359e-02 1.51779056e-01 8.41988921e-02
-6.68495178e-01 5.24104714e-01 -5.09304166e-01 -3.60318512e-01
5.11591434e-01 -6.54643655e-01 -1.67154804e-01 9.63221863e-02
3.82628500e-01 3.27604532e-01 -1.83959410e-01 -1.31655419e+00
-1.97356373e-01 9.19062912e-01 3.88572216e-01 -3.07095230e-01
1.02135277e+00 1.58723086e-01 -1.92375109e-02 5.85427582e-01
6.94350719e-01 3.16403434e-02 -9.62950826e-01 -1.29049450e-01
2.82724857e-01 -6.32532418e-01 -1.95784286e-01 -9.84475672e-01
-9.94914412e-01 6.21199012e-01 4.90765184e-01 -2.33170420e-01
1.01706862e+00 6.99714571e-02 8.42456937e-01 5.01903057e-01
8.16895008e-01 -1.42420006e+00 -5.09956777e-01 5.10333538e-01
8.98443997e-01 -1.22237027e+00 -1.98147655e-01 -2.95132756e-01
-8.08231473e-01 1.00113130e+00 2.25221574e-01 2.38849133e-01
5.60283840e-01 1.55039698e-01 9.58255410e-01 -1.52342677e-01
1.18343599e-01 -4.35813844e-01 2.98591405e-01 6.61362231e-01
4.99518067e-01 2.37489104e-01 -9.47934330e-01 5.20015061e-01
-3.07363302e-01 4.81308103e-01 3.78936827e-01 1.32151866e+00
-2.49499261e-01 -1.53387892e+00 -4.86930043e-01 3.97358119e-01
-7.57353231e-02 3.58246386e-01 -6.40020251e-01 1.01149368e+00
4.65472639e-01 8.57562900e-01 -4.45024699e-01 -4.37998444e-01
5.21328211e-01 5.48549533e-01 5.97020686e-01 -5.78265250e-01
-3.91331881e-01 -2.62617975e-01 3.38563353e-01 -3.05896252e-01
-9.94796276e-01 -8.53438079e-01 -1.24234891e+00 1.15957879e-01
-3.07249635e-01 -1.35782734e-01 8.37707520e-01 1.19523668e+00
-1.45386159e-01 -8.20282027e-02 -1.11934349e-01 -5.89461505e-01
-5.25135517e-01 -8.66779923e-01 -6.69698298e-01 8.77359092e-01
-2.32606032e-03 -8.16432357e-01 -2.72018373e-01 5.42577207e-01] | [9.124751091003418, -6.426894187927246] |
1ef430ed-3932-4eee-8edc-352f03ae4b46 | a-digital-swedish-yiddish-yiddish-swedish | null | null | https://aclanthology.org/2022.eurali-1.14 | https://aclanthology.org/2022.eurali-1.14.pdf | A Digital Swedish-Yiddish/Yiddish-Swedish Dictionary: A Web-Based Dictionary that is also Available Offline | Yiddish is one of the national minority languages of Sweden, and one of the languages for which the Swedish Institute for Language and Folklore is responsible for developing useful language resources. We here describe the web-based version of a Swedish-Yiddish/Yiddish-Swedish dictionary. The single search field of the web-based dictionary is used for incrementally searching all three components of the dictionary entries (the word in Swedish, the word in Yiddish with Hebrew characters and the transliteration in Latin script). When the user accesses the dictionary in an online mode, the dictionary is saved in the web browser, which makes it possible to also use the dictionary offline. | ['Rickard Domeij', 'Maria Skeppstedt', 'Gunnar Eriksson', 'Jean Hessel', 'Magnus Ahltorp'] | null | null | null | null | eurali-lrec-2022-6 | ['transliteration'] | ['natural-language-processing'] | [-1.02473944e-01 -3.53524864e-01 -6.99583828e-01 2.16762871e-02
-4.77130651e-01 -1.04413748e+00 4.85796094e-01 1.72423333e-01
-9.15938377e-01 9.00663674e-01 4.83089417e-01 -1.09052408e+00
-8.35171267e-02 -7.65917182e-01 1.76684260e-01 -4.46289212e-01
5.32363296e-01 6.24890804e-01 3.03933144e-01 -5.69151044e-01
2.67498255e-01 5.92233360e-01 -1.22619069e+00 -1.33146822e-01
7.76167870e-01 3.77150506e-01 4.57226187e-01 4.52442288e-01
-5.88098109e-01 3.25154603e-01 -4.43875492e-01 -5.45968235e-01
1.96337670e-01 -7.64924705e-01 -7.38477349e-01 -4.37136739e-01
8.77402946e-02 -3.49992156e-01 -4.42480296e-01 1.04589629e+00
3.83621067e-01 -2.30009124e-01 3.68647069e-01 -5.25279224e-01
-3.58646631e-01 7.64103949e-01 -1.44433543e-01 1.86491624e-01
7.44005442e-01 -2.77432144e-01 7.78824449e-01 -8.40754926e-01
1.00319517e+00 1.14187288e+00 6.28688693e-01 2.21526042e-01
-7.32463658e-01 -8.77479672e-01 -1.45599052e-01 1.87595174e-01
-1.67695355e+00 -5.79896688e-01 4.87117410e-01 -1.28358513e-01
1.09655631e+00 5.20134985e-01 9.86147165e-01 8.09440374e-01
6.42491430e-02 9.83821228e-02 1.07605207e+00 -1.02427411e+00
9.35917869e-02 4.44137454e-01 2.93470412e-01 5.30870020e-01
7.15189993e-01 -9.69739854e-02 -5.95251441e-01 -4.37618494e-01
7.46311903e-01 -2.02197179e-01 -3.50323096e-02 4.58154902e-02
-1.16885912e+00 7.30360985e-01 -4.01671976e-01 7.39264667e-01
-1.94636658e-01 -3.49471033e-01 6.45756662e-01 5.01562357e-01
9.07646399e-03 8.34179446e-02 -6.10086620e-01 -5.93897820e-01
-9.96032715e-01 -4.75308970e-02 1.12506831e+00 9.80556250e-01
4.66521144e-01 2.30262764e-02 7.59578168e-01 1.07041407e+00
5.14520943e-01 6.37858152e-01 6.73508704e-01 -7.23550439e-01
8.28727782e-02 7.11123168e-01 7.07170069e-02 -5.41332245e-01
-2.91710913e-01 -1.25570148e-01 -3.00388724e-01 -1.77997187e-01
6.30654514e-01 8.36725980e-02 -5.77983022e-01 1.13864815e+00
4.22948182e-01 -7.26386845e-01 1.33031264e-01 5.59840381e-01
1.22229838e+00 4.40487295e-01 1.02566056e-01 -6.69007301e-01
1.55149186e+00 -6.21274829e-01 -9.73202169e-01 1.67653023e-03
6.91055357e-01 -1.21392310e+00 1.07415617e+00 3.41621876e-01
-1.06515121e+00 -2.18186677e-01 -1.16729128e+00 -2.98099637e-01
-7.92179585e-01 -7.28461519e-02 5.08934557e-01 1.04994452e+00
-1.00641441e+00 2.76187003e-01 -7.23666012e-01 -9.84610379e-01
-3.07436615e-01 2.52869695e-01 -5.00070393e-01 5.29994741e-02
-1.38507068e+00 1.12862873e+00 6.44817352e-01 -3.33089709e-01
-1.96559533e-01 -2.04144329e-01 -1.03982496e+00 -4.74333227e-01
2.20724806e-01 -3.43151167e-02 8.30198050e-01 -8.42123866e-01
-1.37537038e+00 1.32058227e+00 -1.42765850e-01 1.56282619e-01
4.39777136e-01 2.33372167e-01 -9.68939483e-01 2.70666420e-01
4.57358897e-01 6.44173920e-02 3.49894971e-01 -7.28735328e-01
-7.26138890e-01 -4.75215822e-01 -2.01094374e-01 1.57736793e-01
-2.74617314e-01 7.55451858e-01 -8.02854717e-01 -8.94782305e-01
3.35715771e-01 -6.49413288e-01 2.29897961e-01 -3.30986351e-01
1.04709305e-01 -2.44686067e-01 4.34906304e-01 -1.43082440e+00
1.86626339e+00 -2.46105051e+00 -3.24862540e-01 5.09913683e-01
-1.11102365e-01 1.66612566e-01 2.64190048e-01 9.32757795e-01
3.04593565e-03 1.75423995e-01 6.54451475e-02 3.70398939e-01
5.16499616e-02 6.25913620e-01 -2.54686060e-03 6.12156272e-01
-6.97258711e-01 5.52266717e-01 -1.09245503e+00 -4.92332667e-01
1.25564635e-01 3.85129213e-01 -2.94115096e-02 -5.29082477e-01
3.06597531e-01 2.70786524e-01 -2.49351576e-01 7.63779461e-01
5.67214191e-01 3.39703530e-01 7.49065399e-01 3.25451583e-01
-6.41394854e-01 8.67963016e-01 -1.48114824e+00 1.53898549e+00
-3.90780509e-01 2.86163181e-01 6.12077527e-02 -3.33528250e-01
8.53538990e-01 4.86159116e-01 1.67740300e-01 -6.63959324e-01
-1.21233962e-01 1.11754298e+00 1.03403881e-01 -2.47642681e-01
8.00957203e-01 -2.12917477e-02 -2.54126817e-01 6.64809465e-01
-4.16856706e-02 -1.37458205e-01 6.76228225e-01 1.96379304e-01
6.12421691e-01 3.52238595e-01 1.00157428e+00 -7.31328905e-01
7.66596198e-01 2.26091504e-01 4.45230126e-01 5.05134761e-01
1.64575383e-01 3.58113311e-02 1.63702607e-01 -5.19286275e-01
-1.21984982e+00 -9.36072409e-01 -7.77724862e-01 1.28124356e+00
-1.34062633e-01 -8.87320757e-01 -5.29408038e-01 -1.90561742e-01
-1.93359956e-01 8.37915063e-01 -8.34562257e-02 1.72316089e-01
-6.92116141e-01 4.32757139e-02 6.23748481e-01 1.64278522e-01
6.15952134e-01 -1.11824644e+00 -6.07033849e-01 3.36529344e-01
-2.41948962e-01 -5.71246266e-01 -5.28628290e-01 3.55580360e-01
-6.81981862e-01 -8.46698225e-01 -6.56062186e-01 -1.16635990e+00
2.19605818e-01 -6.52295500e-02 8.38586330e-01 3.21533710e-01
-5.70056476e-02 2.24242359e-01 -3.68044049e-01 -1.69642285e-01
-7.88710237e-01 1.77489489e-01 2.22664736e-02 -8.77130389e-01
7.25812316e-01 -4.67697531e-01 -2.27681413e-01 -2.80319955e-02
-7.84763873e-01 -1.89325586e-01 -1.42745674e-01 5.90670407e-01
3.83706480e-01 -1.21361382e-01 3.60321730e-01 -9.32577074e-01
2.52647579e-01 -2.42015839e-01 -7.04309344e-01 1.27057925e-01
-6.79795682e-01 -3.11050475e-01 4.57800865e-01 -1.97217107e-01
-5.57892323e-01 -1.98874727e-01 -8.27546120e-01 6.22717679e-01
-1.49600267e-01 7.56294608e-01 -4.31891799e-01 -1.00780264e-01
3.31256330e-01 6.42958939e-01 -2.37942353e-01 -7.57006228e-01
8.46430734e-02 1.28908229e+00 5.43281496e-01 -2.19048023e-01
3.72263223e-01 1.55563205e-01 -4.21607077e-01 -1.10511577e+00
-6.36311397e-02 -7.62327313e-01 -7.62028873e-01 3.27348411e-02
5.44470191e-01 -9.53360081e-01 -3.34821343e-01 3.53408933e-01
-8.74349058e-01 -1.63083211e-01 -4.80137229e-01 5.44404685e-01
-3.48158896e-01 5.58730721e-01 -4.73629296e-01 -7.09622979e-01
-1.09772399e-01 -8.15573633e-01 5.92425704e-01 1.89605653e-01
-6.94947004e-01 -1.12527013e+00 1.70390040e-01 -5.52693903e-02
4.87734191e-02 -5.79228364e-02 1.45741558e+00 -1.08151078e+00
1.77004278e-01 -1.31648704e-01 6.40240386e-02 -4.07394543e-02
2.91654259e-01 -1.19882450e-01 -6.53232113e-02 -1.90561712e-01
3.92452255e-02 8.91674906e-02 4.59255934e-01 -1.02565072e-01
2.03674376e-01 -3.18400025e-01 -1.76007256e-01 5.47430158e-01
1.63400686e+00 5.90631366e-01 5.98124981e-01 1.07972085e+00
1.56250060e-01 1.33619040e-01 3.21243703e-01 4.71437991e-01
6.45333350e-01 6.16771698e-01 -3.34822208e-01 2.00911731e-01
-3.37698013e-01 -3.25073302e-01 3.42745334e-01 1.08031952e+00
3.94196063e-02 2.46197566e-01 -1.06035709e+00 8.68099034e-01
-1.42463958e+00 -7.16684937e-01 -3.75002623e-01 2.48134995e+00
1.19753587e+00 -3.07473131e-02 4.35310066e-01 3.20614576e-01
4.67214614e-01 -1.90498885e-02 -1.09131262e-01 -9.03568208e-01
-2.05853984e-01 7.21790314e-01 6.60183489e-01 9.98577178e-01
-6.98461413e-01 1.23244584e+00 7.57193422e+00 7.21240103e-01
-9.71624851e-01 1.83773860e-01 -1.21226087e-01 1.56103924e-01
-5.46383977e-01 1.86155409e-01 -1.02468455e+00 4.95729446e-01
9.27649438e-01 -4.59461510e-01 5.62238991e-01 5.84864140e-01
3.77865821e-01 -5.99598467e-01 -5.39961278e-01 9.15035844e-01
6.18028790e-02 -1.14421582e+00 1.27612531e-01 3.24627936e-01
3.75009537e-01 -5.54358028e-03 -2.26761669e-01 6.95704222e-02
3.71705860e-01 -7.58590817e-01 1.07559335e+00 -3.34261581e-02
1.40994835e+00 -8.14849555e-01 4.91725892e-01 3.04679722e-01
-9.68182325e-01 2.34486446e-01 -3.99056375e-01 -1.73142835e-01
5.71718365e-02 2.47472137e-01 -4.05161530e-01 4.34057951e-01
5.70092916e-01 3.21584105e-01 -5.26539624e-01 8.73045385e-01
-1.73382148e-01 8.48321795e-01 -5.35281599e-01 -1.16547540e-01
3.32635880e-01 -6.90284789e-01 7.86588788e-01 1.47701776e+00
4.10352200e-01 2.81784292e-02 2.38886207e-01 4.36109379e-02
3.05709571e-01 6.44544184e-01 -4.16096509e-01 -1.92980990e-01
7.77456641e-01 8.62187505e-01 -9.51605499e-01 -3.20024192e-01
-4.86189932e-01 9.50984716e-01 -2.61701316e-01 3.61339748e-01
-2.54375964e-01 -7.83520758e-01 4.91695106e-01 3.62696022e-01
4.37068045e-01 -6.47304595e-01 -2.87256807e-01 -7.61990607e-01
-1.73210248e-01 -1.50324249e+00 5.90737641e-01 -3.53427440e-01
-5.05674720e-01 6.46633625e-01 3.98714803e-02 -6.60872877e-01
-5.05877316e-01 -6.36316240e-01 6.88486174e-02 1.17590177e+00
-1.02801156e+00 -1.09792519e+00 3.12781751e-01 7.73950517e-01
3.57725441e-01 -3.54373574e-01 1.37551582e+00 4.32503432e-01
-2.79838175e-01 5.55654943e-01 6.80569708e-01 2.00756595e-01
7.32876301e-01 -1.13050950e+00 2.91282862e-01 5.83811343e-01
2.21890196e-01 9.33930695e-01 7.76919127e-01 -1.01308644e+00
-1.11758339e+00 -3.82756770e-01 1.86801744e+00 -9.72304046e-02
9.55309451e-01 -4.16269034e-01 -5.78752160e-01 6.93663299e-01
5.21080017e-01 -7.72083640e-01 1.11605966e+00 -1.56861067e-01
-2.36232787e-01 1.46129027e-01 -9.94228005e-01 9.46215868e-01
9.45343018e-01 -5.13176858e-01 -6.50589406e-01 4.93208826e-01
4.43897247e-01 -3.16231817e-01 -7.91597545e-01 -3.27239156e-01
9.66395438e-01 -7.10020602e-01 6.35418236e-01 -1.65132061e-01
-1.24992412e-02 -2.32035130e-01 -3.38275462e-01 -8.82636666e-01
-2.58382529e-01 -8.88239384e-01 3.39215815e-01 9.86426711e-01
5.40453315e-01 -1.02492285e+00 4.17732716e-01 1.35617986e-01
1.75062623e-02 3.27948630e-02 -1.31446588e+00 -7.83355474e-01
1.28356427e-01 -2.63445467e-01 6.22539997e-01 8.83809030e-01
4.91223633e-01 1.26056805e-01 -3.71663272e-02 -4.19181347e-01
2.23670840e-01 -1.67549714e-01 4.22662079e-01 -1.01013267e+00
-1.44768715e-01 -4.99062985e-01 -3.32989752e-01 -8.51129830e-01
-6.08514622e-02 -1.04088902e+00 -4.69646633e-01 -1.58106160e+00
-5.84903546e-02 -3.99282873e-01 1.11335717e-01 4.72211719e-01
1.67949960e-01 4.62063879e-01 1.55782878e-01 4.22119141e-01
-1.18665919e-01 -2.33710662e-01 7.52690017e-01 2.53396302e-01
-4.49802309e-01 -3.40089858e-01 -8.40653002e-01 8.14362764e-01
6.31586671e-01 -6.11403763e-01 3.12711000e-02 -2.70548373e-01
5.55716693e-01 -3.14812452e-01 -1.13459565e-01 -6.54875875e-01
1.57992169e-01 -3.21139902e-01 2.31255561e-01 -8.69527698e-01
-7.55535513e-02 -8.76369596e-01 5.47037899e-01 8.04697573e-01
1.16148651e-01 2.91931897e-01 2.71686792e-01 -1.63757771e-01
-8.25652108e-02 -7.34291375e-01 6.87274218e-01 -4.48204041e-01
-7.60014117e-01 -2.21463591e-01 -1.06700063e+00 -1.11772992e-01
7.83543706e-01 -8.33678603e-01 -6.87328056e-02 -4.10745591e-01
-5.83635747e-01 -2.19808534e-01 1.01185560e+00 6.56239465e-02
4.36770707e-01 -1.24982178e+00 -6.60492420e-01 5.41105926e-01
2.56032735e-01 -8.36819470e-01 -2.62609303e-01 6.66851699e-01
-1.33849847e+00 4.82581556e-01 -2.09732562e-01 2.95683760e-02
-1.13673985e+00 4.44212049e-01 6.22454807e-02 -3.12602252e-01
-9.16668773e-01 3.75898629e-01 -3.51018041e-01 -5.03368378e-01
1.56194150e-01 2.90875614e-01 -3.13322425e-01 1.51882514e-01
7.33793437e-01 5.01049340e-01 2.41087154e-01 -1.15121508e+00
-7.14363813e-01 3.95591110e-01 -3.43324952e-02 -7.83716023e-01
1.19930685e+00 -5.45529544e-01 -6.22082710e-01 7.54258692e-01
8.54540408e-01 7.70530939e-01 -1.45938560e-01 -1.51138783e-01
1.94775134e-01 -3.52406263e-01 -3.57576869e-02 -1.00369668e+00
-2.65018970e-01 2.16302916e-01 4.20059055e-01 -5.83387390e-02
1.04546392e+00 -3.58139239e-02 8.72580767e-01 1.63911805e-01
4.34187710e-01 -1.20194089e+00 -7.96328664e-01 8.93287182e-01
6.94755852e-01 -5.02243042e-01 7.62976930e-02 -3.04570824e-01
-4.53581363e-01 1.35808837e+00 -3.19141775e-01 1.01055115e-01
8.93776655e-01 3.94584686e-01 5.57963014e-01 1.16226315e-01
-2.25259796e-01 -3.36275160e-01 -1.17487766e-01 8.34896863e-01
7.29952514e-01 4.39132862e-02 -1.60376024e+00 6.59590483e-01
-8.01152587e-01 -1.67940333e-01 4.42808360e-01 1.01862693e+00
-4.86642420e-01 -1.70929325e+00 -6.81831300e-01 3.76761705e-01
-7.42699742e-01 -6.76370144e-01 -6.23417139e-01 9.38853145e-01
1.98160410e-01 9.11015749e-01 7.28085488e-02 1.17179580e-01
1.58466905e-01 3.70403409e-01 4.78521645e-01 -5.32805979e-01
-1.10031748e+00 6.80423856e-01 4.17112470e-01 -1.04505628e-01
-1.13882691e-01 -7.77435601e-01 -1.38361251e+00 -6.13688350e-01
-8.06947891e-03 5.44613421e-01 7.76478052e-01 7.68755615e-01
-3.20455104e-01 -1.65847614e-01 1.19732089e-01 -4.70814817e-02
-1.49229333e-01 -8.03671122e-01 -1.10464025e+00 -1.63126886e-01
-1.96271148e-02 -2.66360849e-01 -1.15047418e-01 -2.45874487e-02] | [10.346799850463867, 10.287995338439941] |
6dd3a6c5-a30d-4511-8332-1ecbe7571dea | parsing-to-noncrossing-dependency-graphs | null | null | https://aclanthology.org/Q15-1040 | https://aclanthology.org/Q15-1040.pdf | Parsing to Noncrossing Dependency Graphs | We study the generalization of maximum spanning tree dependency parsing to maximum acyclic subgraphs. Because the underlying optimization problem is intractable even under an arc-factored model, we consider the restriction to noncrossing dependency graphs. Our main contribution is a cubic-time exact inference algorithm for this class. We extend this algorithm into a practical parser and evaluate its performance on four linguistic data sets used in semantic dependency parsing. We also explore a generalization of our parsing framework to dependency graphs with pagenumber at most k and show that the resulting optimization problem is NP-hard for k {\mbox{$\geq$}} 2. | ['Peter Jonsson', 'Marco Kuhlmann'] | 2015-01-01 | null | null | null | tacl-2015-1 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [ 2.29492486e-01 7.44363844e-01 -3.66858184e-01 -5.90771139e-01
-9.53421593e-01 -9.74968791e-01 -1.08803310e-01 2.02830061e-01
-2.90025055e-01 9.10248280e-01 -1.28370672e-01 -8.95281613e-01
-2.76343137e-01 -1.06605387e+00 -7.80205250e-01 -3.92173439e-01
-4.98053104e-01 7.42915213e-01 6.09629869e-01 -1.48183927e-01
-2.69794948e-02 5.68611979e-01 -1.03936255e+00 1.33842779e-02
8.42486560e-01 3.06859523e-01 6.15296736e-02 1.11521673e+00
-4.24604088e-01 2.84278333e-01 -6.46333098e-01 -7.73689210e-01
1.76018059e-01 -3.03144544e-01 -1.52359962e+00 2.02458315e-02
4.23392773e-01 2.57248897e-02 -3.15262862e-02 1.24336100e+00
6.60116524e-02 -1.46640837e-01 1.83015451e-01 -1.31756425e+00
-3.61319393e-01 1.20951283e+00 -3.96700919e-01 5.23302495e-01
3.23835939e-01 -7.35456407e-01 1.59035003e+00 -1.52534768e-01
8.10743392e-01 1.38319194e+00 3.85035992e-01 4.74618405e-01
-1.27999282e+00 -4.34521317e-01 4.85658616e-01 3.94572467e-02
-8.76008749e-01 -6.61964342e-02 4.60155249e-01 -3.95738799e-03
1.60526395e+00 3.00156683e-01 1.74393848e-01 4.18691278e-01
2.91750170e-02 4.97977227e-01 1.35121107e+00 -8.35726976e-01
4.32885177e-02 -5.04231274e-01 8.81467283e-01 1.02626359e+00
7.79158831e-01 -2.42113203e-01 -2.42028028e-01 -3.74999009e-02
6.08976543e-01 -8.17678332e-01 1.57980010e-01 1.84724867e-01
-5.95789373e-01 1.10198641e+00 6.06048517e-02 4.02952164e-01
4.15506363e-01 3.04576993e-01 2.28332788e-01 4.85673785e-01
2.48555213e-01 2.27772638e-01 -8.48411024e-01 1.49532363e-01
-3.35698634e-01 1.23861961e-01 1.54006696e+00 1.26600134e+00
7.80497730e-01 -2.30550155e-01 5.57320297e-01 5.72234929e-01
-8.16452354e-02 6.30231798e-01 -2.80740708e-01 -1.17938375e+00
9.73650753e-01 2.90888041e-01 -6.91136941e-02 -5.07473946e-01
-8.43714952e-01 -1.32857025e-01 -5.05829230e-02 -7.08251223e-02
6.58039033e-01 -3.96986008e-01 -7.66956866e-01 1.96733642e+00
4.15641516e-01 -4.92441237e-01 1.69736370e-01 3.91598523e-01
4.28221434e-01 6.20817900e-01 4.04815018e-01 -5.30404747e-01
1.65736592e+00 -8.61132562e-01 -6.36656165e-01 -4.70367372e-01
1.15614676e+00 -5.38215756e-01 6.08595967e-01 3.77514750e-01
-1.12055433e+00 2.40693279e-02 -9.58130002e-01 -3.35889757e-01
-2.55313456e-01 -3.36360276e-01 1.31799924e+00 1.09567606e+00
-1.20141947e+00 5.91606736e-01 -8.79182577e-01 -3.47418159e-01
1.83805022e-02 6.91288710e-01 -6.02664053e-01 -2.22090676e-01
-1.20001531e+00 9.35047448e-01 8.02121997e-01 1.44400254e-01
-6.17549457e-02 3.30428742e-02 -1.05473244e+00 -6.54579923e-02
6.16471469e-01 -3.81072193e-01 1.36449182e+00 -6.85448408e-01
-1.18017042e+00 1.11518788e+00 -3.85560542e-01 -2.40187675e-01
-6.47212565e-02 -5.94855845e-02 -4.51611429e-01 3.22621137e-01
3.56190026e-01 2.87046313e-01 -2.37356797e-02 -8.84608328e-01
-9.10759449e-01 -7.25397646e-01 8.19008291e-01 5.93798496e-02
-3.13992575e-02 5.87954223e-01 -5.82106113e-01 -4.50137079e-01
5.64184427e-01 -1.17057288e+00 -3.85124981e-01 -7.72191584e-01
-6.92748249e-01 -5.49165606e-01 2.89117336e-01 -7.91133046e-01
1.39061713e+00 -1.64179945e+00 2.65715182e-01 3.52318048e-01
-3.16846110e-02 -1.47636056e-01 -5.96406534e-02 5.82983971e-01
-1.67628527e-01 4.53320146e-01 -5.44031799e-01 -1.59080606e-02
2.00798064e-01 9.97841120e-01 8.22027773e-03 3.48738134e-01
2.04769492e-01 8.56139660e-01 -6.86807036e-01 -9.61667717e-01
-2.61145443e-01 -2.26855740e-01 -6.69637501e-01 -5.75430058e-02
-4.98574197e-01 -1.46996334e-01 -8.78221571e-01 6.85595274e-01
9.59112942e-01 -4.05437723e-02 1.00563753e+00 3.57031614e-01
-1.59915075e-01 5.42912900e-01 -1.23054242e+00 1.78488648e+00
-5.40519357e-01 2.75316983e-01 4.16942298e-01 -1.06851053e+00
5.43830991e-01 -5.54414205e-02 1.22309610e-01 -3.81156534e-01
3.19643691e-02 1.65578827e-01 -5.95197342e-02 -5.34459531e-01
3.85854751e-01 -3.29060644e-01 -8.64455879e-01 4.60533738e-01
4.93307151e-02 -3.34614627e-02 6.32433057e-01 4.07741725e-01
1.33725512e+00 2.24403188e-01 8.19241554e-02 -7.01404154e-01
3.17189306e-01 3.06593716e-01 7.69225240e-01 6.94133461e-01
6.99344799e-02 -4.14962061e-02 1.38627720e+00 -2.58149743e-01
-7.95706987e-01 -1.01048803e+00 -5.18114567e-01 1.44231617e+00
9.19185057e-02 -5.00180542e-01 -1.12609446e+00 -9.84702528e-01
-3.07384491e-01 7.30748057e-01 -4.30976093e-01 3.83227110e-01
-1.20601130e+00 -1.17655015e+00 6.44691706e-01 9.03189301e-01
3.14897746e-01 -8.56844783e-01 -5.62276959e-01 4.76713926e-01
-1.87853083e-01 -1.74993086e+00 1.93091901e-03 7.76328564e-01
-1.17498171e+00 -1.26271665e+00 1.30381405e-01 -1.28879857e+00
7.15927601e-01 -4.22874898e-01 1.31887090e+00 2.39505723e-01
-1.13464206e-01 1.64499402e-01 -5.45524180e-01 -1.00385062e-01
-5.05993545e-01 3.90449256e-01 -4.64422971e-01 -9.77129281e-01
4.32807952e-01 -4.11083549e-01 2.66576651e-02 5.69395423e-02
-8.47561419e-01 -2.40436226e-01 2.11007521e-01 6.53098643e-01
5.46200812e-01 3.49966705e-01 2.84910232e-01 -1.60141480e+00
3.22888225e-01 -2.62982070e-01 -1.06162715e+00 5.23789644e-01
-3.25153589e-01 6.52204156e-01 6.54934466e-01 2.99070299e-01
-1.14815497e+00 4.12436813e-01 -5.55477679e-01 5.89704216e-01
2.58148182e-02 5.11261344e-01 -6.49049044e-01 5.31751104e-02
4.38579312e-03 -4.15577322e-01 -7.25565970e-01 -5.72150528e-01
3.87443781e-01 2.22544223e-01 6.65457249e-01 -1.00407326e+00
6.74783826e-01 2.85926819e-01 6.51514471e-01 -6.94745660e-01
-9.22310233e-01 -1.80510059e-01 -7.89059579e-01 4.64864731e-01
1.09243596e+00 -5.06527901e-01 -6.53777599e-01 3.88994999e-02
-1.46460593e+00 -4.37946618e-01 1.32808089e-01 2.71816760e-01
-4.25013334e-01 9.32837546e-01 -9.71390426e-01 -7.62936294e-01
-1.04640789e-01 -8.87430131e-01 9.90103900e-01 -7.73090795e-02
6.91553056e-02 -1.08741128e+00 2.56212503e-01 2.67800808e-01
-3.97398740e-01 3.29166442e-01 1.48729181e+00 -8.40379894e-01
-5.03548443e-01 4.24442478e-02 -2.64511883e-01 6.36499971e-02
-2.07418591e-01 -1.14348367e-01 -5.61416388e-01 1.33842081e-01
-3.53770673e-01 -1.11173078e-01 8.54118228e-01 1.56602472e-01
9.22804177e-01 -1.46273836e-01 -6.40510798e-01 4.37217414e-01
1.68920183e+00 -9.27221030e-02 4.46468830e-01 3.80007178e-01
5.03501713e-01 8.47100496e-01 6.64340675e-01 -1.25574663e-01
5.01824915e-01 4.11826819e-01 3.88356298e-01 1.64382964e-01
2.19865739e-01 -1.40746742e-01 1.87256917e-01 8.68806660e-01
-3.09234440e-01 -5.24718285e-01 -7.67574012e-01 5.74295998e-01
-1.53703547e+00 -4.32699949e-01 -7.27603137e-01 1.91572523e+00
8.21868062e-01 3.78683597e-01 6.31180182e-02 2.85319626e-01
9.36770856e-01 -2.14541689e-01 2.21488010e-02 -1.33292162e+00
-2.09852770e-01 1.06582296e+00 9.71144676e-01 9.92318630e-01
-1.17818987e+00 1.44309652e+00 7.48441076e+00 3.37555975e-01
-2.64736444e-01 3.15834761e-01 2.27311417e-01 3.37245762e-01
-4.70997751e-01 4.19088036e-01 -1.06503272e+00 4.58027832e-02
1.20823967e+00 2.38906920e-01 2.53711641e-01 7.77488708e-01
-4.74008471e-01 -2.94936031e-01 -8.41184020e-01 1.61911741e-01
-3.67202133e-01 -6.84014916e-01 -4.49855983e-01 4.28157076e-02
4.21829879e-01 -1.12315297e-01 -4.33788031e-01 1.45707205e-01
8.30939054e-01 -9.26729381e-01 3.07986975e-01 -5.39735436e-01
8.43233526e-01 -9.29063201e-01 5.82113087e-01 1.32252976e-01
-1.43159771e+00 -1.08892031e-01 -4.86758143e-01 -1.12036891e-01
6.06229007e-01 5.51895738e-01 -4.25925374e-01 8.22357059e-01
6.78806305e-01 1.25316568e-02 -2.10090920e-01 4.27617669e-01
-7.95587063e-01 5.91278315e-01 -7.46813476e-01 6.82172328e-02
1.82203621e-01 -1.78510174e-01 3.39222610e-01 1.65114081e+00
4.44792099e-02 6.65241122e-01 1.89855933e-01 2.24215165e-01
-1.60410125e-02 -1.77761295e-03 -2.67049402e-01 -2.92120893e-02
4.48645473e-01 1.08894324e+00 -1.48704636e+00 -4.94235344e-02
-6.48236930e-01 9.51597393e-01 7.99936354e-01 -3.59964073e-02
-7.32248962e-01 -7.06732988e-01 4.11423266e-01 -3.23901355e-01
7.46534288e-01 -6.13938451e-01 -1.93826884e-01 -1.00676012e+00
2.56691813e-01 -4.82819319e-01 1.03234160e+00 -3.73407751e-01
-9.12031412e-01 6.99118376e-01 4.94521052e-01 -2.67832577e-02
-3.66846710e-01 -1.13254225e+00 -2.36710414e-01 7.50091136e-01
-1.72840929e+00 -1.00502443e+00 2.47272402e-01 4.77243692e-01
2.40573466e-01 6.72694802e-01 1.21367466e+00 5.60916252e-02
-4.32640374e-01 5.57173729e-01 -3.63730490e-01 1.65871620e-01
-5.82730658e-02 -1.75534093e+00 7.29983270e-01 1.32345939e+00
1.31393924e-01 2.91387469e-01 7.07533181e-01 -6.59417391e-01
-1.78886223e+00 -8.71678591e-01 1.48574543e+00 -3.94005299e-01
8.92323911e-01 -4.86767411e-01 -7.03042805e-01 1.23084080e+00
1.03152856e-01 6.84775710e-02 5.17477512e-01 4.07278508e-01
-6.31090760e-01 2.50648052e-01 -1.00263298e+00 1.64168432e-01
1.75889659e+00 -2.63930827e-01 -7.59513021e-01 4.23843294e-01
9.69623208e-01 -6.32674813e-01 -1.32369339e+00 3.83355618e-01
3.83036643e-01 -6.38305664e-01 6.58395112e-01 -8.88981819e-01
2.05058202e-01 1.24185160e-02 -2.60484725e-01 -8.38048220e-01
-2.32032537e-01 -6.71117485e-01 1.80687338e-01 1.11201286e+00
9.33028340e-01 -8.65668356e-01 8.14705074e-01 6.45972073e-01
-2.17684716e-01 -4.33024317e-01 -1.21419704e+00 -8.01911712e-01
6.80214584e-01 -5.81862330e-01 5.27192712e-01 7.24257946e-01
4.51165318e-01 4.91194278e-01 1.03904463e-01 6.40702844e-01
6.30433679e-01 6.44662499e-01 2.31162518e-01 -1.14839423e+00
-7.77227879e-01 -8.45616832e-02 -2.19540060e-01 -1.02596676e+00
6.68606222e-01 -1.09754825e+00 1.04178265e-01 -1.67704237e+00
1.62695348e-01 -7.75336564e-01 2.24868972e-02 5.85603356e-01
-8.10981169e-02 -8.58035963e-03 -8.00599232e-02 -3.57387960e-01
-5.67503333e-01 -4.65701252e-01 1.08827937e+00 2.76175857e-01
1.02253608e-01 1.40189575e-02 -6.70439661e-01 6.41620874e-01
7.94083476e-01 -9.18702364e-01 -1.97654977e-01 -7.87413001e-01
6.51574492e-01 4.82285887e-01 1.62491873e-02 -3.47184479e-01
-9.90728661e-02 -2.61625439e-01 -3.35555196e-01 -4.08027560e-01
-4.58205454e-02 -6.35371745e-01 -4.00333200e-03 4.31700259e-01
-2.21380904e-01 4.91826326e-01 2.50248760e-01 4.02187109e-01
7.58058727e-02 -7.79130101e-01 3.62074405e-01 -2.02397630e-01
-7.15175927e-01 2.17602681e-02 -2.49894544e-01 4.52702165e-01
8.60612750e-01 -2.49833260e-02 -5.34564316e-01 1.26006365e-01
-1.14059782e+00 2.93288678e-01 3.73634696e-01 3.15535739e-02
4.44372505e-04 -5.42932034e-01 -5.63606858e-01 -1.76239684e-01
-4.01799865e-02 1.26985282e-01 -6.78272024e-02 4.33977664e-01
-8.14266145e-01 5.46804011e-01 1.17234424e-01 -1.79215297e-01
-1.55200529e+00 6.51365638e-01 5.63561954e-02 -6.93534970e-01
-5.59898317e-01 8.77320409e-01 -2.45209355e-02 -3.57558489e-01
-1.05407692e-01 -5.33232987e-01 -1.01559237e-01 -4.09686446e-01
6.23903200e-02 2.86677539e-01 1.81103289e-01 -6.48051798e-01
-4.89227980e-01 5.55212498e-01 7.13305101e-02 -1.80432573e-01
1.28357530e+00 -2.00835001e-02 -5.85662603e-01 -1.32948473e-01
1.19191241e+00 2.72404283e-01 -6.51511908e-01 9.45030078e-02
4.73875493e-01 -8.87594000e-02 -3.59726399e-01 -4.62145865e-01
-9.22497690e-01 5.00453413e-01 -1.23221561e-01 5.70003808e-01
1.22023201e+00 6.22525394e-01 7.55626976e-01 7.38567948e-01
6.72080696e-01 -1.15570545e+00 -7.30119824e-01 6.34509861e-01
2.06526458e-01 -7.19390810e-01 1.87691629e-01 -1.30026245e+00
-2.04608187e-01 1.14633787e+00 3.06447536e-01 -1.75036415e-01
6.04296505e-01 8.06824207e-01 -3.11709940e-01 -1.25667930e-01
-5.41527748e-01 -4.32158023e-01 -4.43795115e-01 5.43732643e-01
3.87706935e-01 3.73345464e-01 -9.23988640e-01 7.29836047e-01
-6.58131301e-01 -2.60241240e-01 6.14817142e-01 1.33172154e+00
-5.01456678e-01 -1.80987895e+00 -1.20939225e-01 5.63234948e-02
-1.16056287e+00 -2.32792705e-01 -3.70558470e-01 1.28909421e+00
-1.00074545e-03 1.13226807e+00 -4.00362676e-03 4.44250032e-02
2.77100891e-01 6.56111166e-02 1.17904079e+00 -8.12616408e-01
-5.15242636e-01 -5.42651825e-02 9.87518072e-01 -3.34962338e-01
-6.28323078e-01 -6.43188119e-01 -1.76600552e+00 -3.65963012e-01
-7.09662676e-01 4.92323279e-01 7.01868415e-01 1.05899441e+00
-1.57949001e-01 2.55116135e-01 3.66238743e-01 -7.02406093e-03
-3.99460852e-01 -5.77355742e-01 -7.42235482e-01 -7.18315095e-02
-1.52506366e-01 -4.26967949e-01 -2.30471879e-01 -8.27437788e-02] | [10.312551498413086, 9.604239463806152] |
e67a8878-5ccb-4430-b80f-62feb86eb34d | projective-urban-texturing | 2201.10938 | null | https://arxiv.org/abs/2201.10938v2 | https://arxiv.org/pdf/2201.10938v2.pdf | Projective Urban Texturing | This paper proposes a method for automatic generation of textures for 3D city meshes in immersive urban environments. Many recent pipelines capture or synthesize large quantities of city geometry using scanners or procedural modeling pipelines. Such geometry is intricate and realistic, however the generation of photo-realistic textures for such large scenes remains a problem. We propose to generate textures for input target 3D meshes driven by the textural style present in readily available datasets of panoramic photos capturing urban environments. Re-targeting such 2D datasets to 3D geometry is challenging because the underlying shape, size, and layout of the urban structures in the photos do not correspond to the ones in the target meshes. Photos also often have objects (e.g., trees, vehicles) that may not even be present in the target geometry. To address these issues we present a method, called Projective Urban Texturing (PUT), which re-targets textural style from real-world panoramic images to unseen urban meshes. PUT relies on contrastive and adversarial training of a neural architecture designed for unpaired image-to-texture translation. The generated textures are stored in a texture atlas applied to the target 3D mesh geometry. To promote texture consistency, PUT employs an iterative procedure in which texture synthesis is conditioned on previously generated, adjacent textures. We demonstrate both quantitative and qualitative evaluation of the generated textures. | ['Evangelos Kalogerakis', 'Tom Kelly', 'Melinos Averkiou', 'Yiangos Georgiou'] | 2022-01-25 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 7.41123259e-01 2.85141706e-01 5.65480709e-01 -1.46506101e-01
-6.36852682e-01 -7.17969358e-01 8.54285419e-01 -3.92089635e-01
3.04326355e-01 5.71297348e-01 1.46960104e-02 -5.64277582e-02
3.30590278e-01 -1.46163964e+00 -1.26311576e+00 -4.91178572e-01
2.41787851e-01 8.95596504e-01 3.23874623e-01 -4.87630188e-01
-1.02183176e-03 9.43956196e-01 -1.84622061e+00 3.15886676e-01
6.62271440e-01 6.56152129e-01 2.43858770e-01 9.26096022e-01
-9.72688645e-02 3.23010892e-01 -4.62183267e-01 -6.19282305e-01
7.87929654e-01 -4.73048836e-01 -5.85841835e-01 5.44222236e-01
1.04810679e+00 -3.05689394e-01 -4.02495235e-01 1.06769276e+00
1.61249548e-01 1.63833722e-02 8.77761185e-01 -1.10987461e+00
-5.49823761e-01 7.51165971e-02 -3.98004800e-01 -4.94922489e-01
3.53882551e-01 3.35243165e-01 4.51295167e-01 -8.44407499e-01
1.28814542e+00 1.44150555e+00 5.49696803e-01 4.10395294e-01
-1.56837702e+00 -5.51838815e-01 -3.31440836e-01 -3.50381762e-01
-1.45637774e+00 -4.35523301e-01 1.12423372e+00 -6.51425183e-01
3.34268898e-01 7.00281620e-01 8.81149411e-01 1.46696997e+00
3.52863610e-01 2.93588877e-01 1.52226353e+00 -3.85651588e-01
2.47457504e-01 9.70595703e-02 -9.59562600e-01 7.12269366e-01
-1.07710838e-01 4.19702053e-01 -1.24755614e-01 2.38113161e-02
1.35331511e+00 -2.38518491e-01 -1.18382126e-01 -6.24261379e-01
-1.27541244e+00 5.98407924e-01 4.10317153e-01 -7.31065720e-02
-4.73748386e-01 2.08582327e-01 3.57722752e-02 3.56455326e-01
7.64945447e-01 4.09324288e-01 1.13317221e-02 8.23816657e-02
-9.51177955e-01 5.10052502e-01 4.71078932e-01 1.19028318e+00
1.05294859e+00 3.79275173e-01 1.64656445e-01 6.62711978e-01
-6.50518909e-02 1.05200481e+00 1.15358399e-03 -9.86673832e-01
4.03562218e-01 4.64457422e-01 2.95036826e-02 -1.30366433e+00
1.00199029e-01 1.62802458e-01 -9.76241112e-01 5.78380346e-01
3.84127587e-01 1.85140610e-01 -1.15334642e+00 1.38212407e+00
7.92154014e-01 1.89252384e-02 -1.01945400e-01 7.52134025e-01
7.82125771e-01 6.35220349e-01 -2.46216714e-01 4.09020066e-01
9.74503338e-01 -5.55732369e-01 -3.56606066e-01 -2.02276736e-01
9.48506743e-02 -1.10852206e+00 1.37136495e+00 -2.76744328e-02
-1.06390643e+00 -4.70709324e-01 -8.29332530e-01 -1.39519989e-01
-4.04544026e-01 -1.73041075e-01 2.21304715e-01 7.36535430e-01
-9.52432454e-01 3.45469922e-01 -4.72803831e-01 -5.53872108e-01
3.61781031e-01 -8.59485492e-02 -5.32403588e-01 -1.69475693e-02
-9.18763101e-01 7.57405102e-01 1.50996000e-01 2.37629041e-02
-8.24091554e-01 -6.74582481e-01 -1.03491759e+00 -2.60252923e-01
1.11874796e-01 -7.53297687e-01 8.40334296e-01 -1.16720390e+00
-1.86367977e+00 1.38131082e+00 1.73065528e-01 1.75131653e-02
1.02894652e+00 3.28404993e-01 -3.79207879e-01 3.91158462e-02
2.35534236e-01 8.35216463e-01 1.04624760e+00 -1.83846474e+00
-2.50782877e-01 -4.12858352e-02 -2.84550730e-02 1.77756995e-01
4.77795511e-01 -4.09831524e-01 -4.66183692e-01 -9.68874156e-01
3.41258526e-01 -9.89702404e-01 -3.21716905e-01 3.85669738e-01
-8.02563131e-01 7.88764358e-01 1.06365335e+00 -8.28998148e-01
4.05026972e-01 -2.04313946e+00 6.23052455e-02 7.22411811e-01
-1.38502626e-03 -3.39685708e-01 -2.11637497e-01 4.73674893e-01
-3.72607969e-02 1.38160944e-01 -3.43523949e-01 -3.16163808e-01
1.16436318e-01 4.06044543e-01 -6.30499840e-01 4.08922195e-01
1.75256059e-01 9.69309151e-01 -7.39634335e-01 -4.88639832e-01
7.08171010e-01 4.12066489e-01 -5.63354790e-01 1.77697793e-01
-4.61834401e-01 9.07601595e-01 -3.44148636e-01 7.15140760e-01
8.85418236e-01 1.51094034e-01 1.65987555e-02 -1.06245801e-01
-2.15062182e-02 -9.05745253e-02 -1.31326818e+00 1.44400752e+00
-6.10149264e-01 7.23003983e-01 -3.82446870e-02 -5.42589247e-01
1.16215885e+00 6.67457953e-02 2.69727945e-01 -9.14610207e-01
1.55980408e-01 3.43394637e-01 -5.31397820e-01 -3.95305037e-01
9.89329219e-01 -3.25084209e-01 -3.47294837e-01 3.03782761e-01
-5.34195304e-01 -1.34518635e+00 -1.41865373e-01 -1.57037139e-01
7.79673278e-01 2.84007818e-01 -2.13732049e-01 -2.21550271e-01
2.28950828e-01 3.16471666e-01 1.10518448e-01 6.64429009e-01
3.99129003e-01 1.29506087e+00 3.07055026e-01 -7.11061835e-01
-1.94794011e+00 -1.46234632e+00 -2.45881602e-01 2.30670169e-01
1.44191176e-01 4.90037017e-02 -8.74290764e-01 -3.08732152e-01
-1.91790551e-01 5.84641993e-01 -9.67021823e-01 1.03784725e-01
-6.20372236e-01 -1.37295187e-01 5.44209361e-01 -9.90297571e-02
6.07751191e-01 -1.09898353e+00 -6.07554674e-01 2.50678897e-01
-3.01296502e-01 -1.24514079e+00 -3.10355335e-01 -4.21122193e-01
-6.17904425e-01 -1.05729818e+00 -6.51839793e-01 -7.41432607e-01
9.54877377e-01 -1.09841071e-01 1.26656783e+00 -1.02047727e-01
-2.03328103e-01 3.25197875e-01 -2.12313265e-01 -2.82410473e-01
-1.00289941e+00 -1.99634701e-01 -2.63896763e-01 4.31150645e-01
-5.00625730e-01 -7.75194883e-01 -3.44885230e-01 5.91141701e-01
-1.26998580e+00 8.51346374e-01 1.97092608e-01 6.77947998e-01
9.82617497e-01 2.49062464e-01 -2.38645121e-01 -1.16642010e+00
1.24516174e-01 -2.33252004e-01 -7.32477784e-01 4.36943397e-02
1.65194273e-01 -1.39971778e-01 6.99827492e-01 -6.08274043e-01
-9.90064263e-01 2.02239826e-01 -1.46001682e-01 -7.26742625e-01
-4.26313341e-01 1.65896833e-01 -2.88705915e-01 -1.76936746e-01
8.97660315e-01 3.03602427e-01 -1.61732927e-01 -2.38896012e-01
3.88211071e-01 3.04035991e-01 6.85661197e-01 -8.71918201e-01
1.56454051e+00 9.27091897e-01 1.42716452e-01 -9.75555182e-01
-2.81704158e-01 3.81312340e-01 -6.09993398e-01 -5.04826248e-01
8.05305958e-01 -7.06449330e-01 1.49854552e-03 6.55219972e-01
-9.75732863e-01 -9.22976673e-01 -6.79139197e-01 5.28481081e-02
-1.02689171e+00 6.12582192e-02 -3.08158815e-01 -3.92710268e-01
-4.86738794e-02 -1.27458537e+00 1.55985141e+00 -2.54799902e-01
-2.69215435e-01 -8.38418365e-01 8.00920278e-02 3.02604347e-01
5.66243470e-01 1.10094678e+00 1.07844961e+00 5.26667237e-01
-9.29197252e-01 -1.62827685e-01 7.31444336e-04 -8.44791979e-02
2.48439521e-01 4.50199366e-01 -9.50828314e-01 8.64700153e-02
-3.83051246e-01 -1.89773887e-01 1.19028665e-01 2.90043741e-01
1.04479790e+00 -3.84166658e-01 -3.51016074e-02 6.82485878e-01
1.38839853e+00 -1.21151567e-01 1.10192847e+00 4.58002478e-01
1.06570530e+00 8.57786238e-01 3.91543478e-01 1.13938101e-01
2.29558095e-01 9.70215738e-01 3.16904783e-01 -3.41299027e-01
-4.16632414e-01 -6.88835442e-01 1.85527101e-01 5.14756382e-01
-1.31732449e-01 -2.23260507e-01 -9.34180319e-01 5.17445982e-01
-1.30392170e+00 -1.07046497e+00 -3.90948445e-01 2.30990577e+00
7.49424219e-01 1.58759415e-01 -2.76678443e-01 1.37460098e-01
8.29778373e-01 1.12037569e-01 -3.46361369e-01 -5.67186534e-01
-5.78608692e-01 5.44190407e-01 5.81024051e-01 6.66318178e-01
-9.81545448e-01 1.22025251e+00 5.76056671e+00 9.49311614e-01
-1.58056247e+00 -1.12264911e-02 8.52788627e-01 2.15431288e-01
-6.84760749e-01 2.59782877e-02 -3.20281014e-02 3.50824028e-01
5.97603023e-01 -5.97457476e-02 5.28435111e-01 7.52088904e-01
4.00274009e-01 -9.91162360e-02 -8.01415384e-01 9.87640798e-01
-9.84660164e-02 -1.70852137e+00 4.43373770e-01 2.25140005e-01
1.40365064e+00 -9.09561813e-02 2.83721715e-01 -1.55037984e-01
5.61180294e-01 -1.15210831e+00 1.37117600e+00 6.25102401e-01
1.52879846e+00 -6.04108751e-01 1.43173441e-01 1.35820031e-01
-1.21614444e+00 6.56956911e-01 -2.54527390e-01 1.91842183e-01
2.49748230e-01 6.16676748e-01 -6.89295828e-01 5.66706359e-01
6.64145172e-01 3.31398755e-01 -6.01314247e-01 7.87478447e-01
-1.83030516e-01 2.29113117e-01 -5.17301023e-01 4.42887396e-01
2.00158522e-01 -3.93794239e-01 6.44215822e-01 7.71556556e-01
5.79204857e-01 -1.72608927e-01 1.59812570e-02 1.18848741e+00
-5.96943498e-02 2.35208154e-01 -1.27647686e+00 1.38593584e-01
3.38431537e-01 9.92480457e-01 -9.82767284e-01 -3.85930717e-01
2.63311956e-02 1.16597545e+00 -2.46837754e-02 4.13605332e-01
-7.15385318e-01 -7.59360716e-02 5.06830931e-01 6.23473823e-01
1.53277129e-01 -3.12591553e-01 -5.78674734e-01 -1.10279799e+00
9.91545320e-02 -1.05435681e+00 -3.56493264e-01 -1.23167276e+00
-1.10272026e+00 6.83536470e-01 -8.24178457e-02 -1.56255686e+00
-4.26352024e-02 -1.47712618e-01 -6.83440983e-01 9.27887917e-01
-8.83153737e-01 -1.68383026e+00 -5.57109773e-01 6.08192027e-01
5.36752343e-01 1.10816836e-01 8.59842896e-01 2.74931669e-01
-1.20347023e-01 2.99468100e-01 2.46316418e-01 1.18627056e-01
4.40190643e-01 -1.01523650e+00 8.89726520e-01 9.07649696e-01
5.69648072e-02 -9.44367796e-02 9.21064734e-01 -8.04996610e-01
-1.20623887e+00 -1.57012331e+00 6.95281208e-01 -5.99425137e-01
4.05107021e-01 -7.23484933e-01 -7.13320911e-01 7.23914385e-01
-1.44357756e-01 -1.90813035e-01 -8.68663415e-02 -6.45478845e-01
-1.35855079e-01 1.86004534e-01 -1.46360052e+00 1.11035120e+00
1.10452449e+00 -6.05393767e-01 -2.31471851e-01 3.25372428e-01
2.19336629e-01 -1.04361284e+00 -7.57152736e-01 3.18893164e-01
4.44012254e-01 -9.09565926e-01 9.65472639e-01 -1.25456944e-01
9.61098790e-01 -4.52456772e-01 -3.59893680e-01 -1.43833661e+00
-1.33765275e-02 -6.11938119e-01 4.82995272e-01 9.68198061e-01
1.72634065e-01 -4.14286464e-01 9.30339575e-01 6.74582481e-01
-2.01399922e-01 -2.08789825e-01 -1.08250093e+00 -6.21868908e-01
1.85316220e-01 -4.76373941e-01 1.01822603e+00 1.23730898e+00
-8.30254197e-01 -1.22243322e-01 -4.26217914e-01 2.99019843e-01
6.79685295e-01 3.63578111e-01 1.38286197e+00 -1.05223012e+00
1.23795848e-02 -3.88997257e-01 -5.80762565e-01 -6.42645359e-01
1.17767297e-01 -7.14576721e-01 4.57278307e-04 -1.25034094e+00
-3.49107265e-01 -9.59458947e-01 8.08831453e-01 1.25931039e-01
2.52754182e-01 8.38418245e-01 -8.80485028e-02 1.58560932e-01
3.28480631e-01 8.43909442e-01 1.76138246e+00 -3.73995483e-01
-1.26501322e-01 -3.15060556e-01 -3.68043184e-02 5.69932699e-01
7.87398994e-01 -4.16334510e-01 -2.99602151e-01 -4.81101334e-01
1.51316360e-01 4.85240407e-02 6.33590281e-01 -1.16867507e+00
-3.35483342e-01 -4.44966555e-01 4.29049969e-01 -4.93493229e-01
5.70117831e-01 -9.57691848e-01 1.14862585e+00 2.71930963e-01
2.17821877e-02 1.29473597e-01 2.15956613e-01 4.37259734e-01
-4.92188856e-02 3.45631421e-01 1.01343215e+00 -2.23347440e-01
-5.84497392e-01 4.55059022e-01 -3.56939763e-01 8.06395113e-02
9.60247159e-01 -7.47490287e-01 -1.23740956e-01 -4.36682016e-01
-7.42816627e-01 -4.25366968e-01 1.32997417e+00 5.18446267e-01
7.34256983e-01 -1.62682188e+00 -6.94347680e-01 6.08880043e-01
1.84550777e-01 3.53799492e-01 4.53554451e-01 1.90870821e-01
-1.30761957e+00 -1.74519435e-01 -5.79642296e-01 -5.80910385e-01
-1.08191419e+00 2.14857116e-01 5.44077218e-01 1.67470612e-02
-9.82665062e-01 4.01433408e-01 5.89409411e-01 -9.44012105e-01
-4.36223596e-01 -3.06096166e-01 3.03102672e-01 -4.98582244e-01
5.45521341e-02 -1.04897857e-01 9.01325792e-02 -1.05060971e+00
2.85788715e-01 7.89272606e-01 5.25059938e-01 -3.53893131e-01
1.31650448e+00 -6.54284582e-02 1.09891377e-01 3.89233112e-01
8.87647390e-01 3.79633069e-01 -1.41301572e+00 -1.83502465e-01
-5.71345687e-01 -9.79473233e-01 -1.07378267e-01 -4.20190334e-01
-1.26527822e+00 6.57952607e-01 5.81091464e-01 -1.35643721e-01
7.55990624e-01 -1.81758642e-01 7.68860638e-01 4.68321927e-02
7.79782534e-01 -1.05685222e+00 -7.91785643e-02 4.32008266e-01
1.20380521e+00 -8.99621010e-01 -3.54651630e-01 -7.61267006e-01
-3.20091188e-01 1.03498268e+00 4.51496482e-01 -3.19421023e-01
5.12791455e-01 1.15087524e-01 2.41485894e-01 -4.99440342e-01
-2.04712272e-01 -2.26120315e-02 2.16533318e-01 7.83484936e-01
-3.16840678e-01 3.33814144e-01 3.10159147e-01 -3.51585269e-01
-9.68196332e-01 -3.60610306e-01 8.09453785e-01 7.93808877e-01
1.64770290e-01 -1.05006754e+00 -9.65070903e-01 3.47660124e-01
1.21861510e-01 -9.76804760e-04 -4.28935647e-01 8.72406423e-01
2.37335265e-01 5.88854551e-01 4.40901607e-01 -4.77648824e-01
5.46815515e-01 -3.55981410e-01 6.60625815e-01 -5.18682718e-01
-2.53188640e-01 -9.08327103e-02 2.43853003e-01 -4.56692338e-01
-3.19886357e-01 -5.76020896e-01 -6.20751321e-01 -8.44484329e-01
2.17339352e-01 -3.46957564e-01 6.02268875e-01 6.47738636e-01
1.70969039e-01 2.08500043e-01 7.34645307e-01 -1.42687941e+00
3.44354391e-01 -5.48965991e-01 -7.40696251e-01 9.14098799e-01
3.42966467e-02 -7.95593202e-01 -1.08485915e-01 4.60586667e-01] | [9.216387748718262, -3.341254711151123] |
042a9288-2d58-443a-bd26-7b460aff4a03 | nilc_usp-aspect-extraction-using-semantic | null | null | https://aclanthology.org/S14-2075 | https://aclanthology.org/S14-2075.pdf | NILC\_USP: Aspect Extraction using Semantic Labels | null | ['Pedro Balage Filho', 'Thiago Pardo'] | 2014-08-01 | null | null | null | semeval-2014-8 | ['aspect-extraction'] | ['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] | [-7.323763847351074, 3.8498995304107666] |
abf998e7-7f1e-4ed1-a5fd-927d35200cef | collaborative-video-object-segmentation-by-1 | 2010.06349 | null | https://arxiv.org/abs/2010.06349v2 | https://arxiv.org/pdf/2010.06349v2.pdf | Collaborative Video Object Segmentation by Multi-Scale Foreground-Background Integration | This paper investigates the principles of embedding learning to tackle the challenging semi-supervised video object segmentation. Unlike previous practices that focus on exploring the embedding learning of foreground object (s), we consider background should be equally treated. Thus, we propose a Collaborative video object segmentation by Foreground-Background Integration (CFBI) approach. CFBI separates the feature embedding into the foreground object region and its corresponding background region, implicitly promoting them to be more contrastive and improving the segmentation results accordingly. Moreover, CFBI performs both pixel-level matching processes and instance-level attention mechanisms between the reference and the predicted sequence, making CFBI robust to various object scales. Based on CFBI, we introduce a multi-scale matching structure and propose an Atrous Matching strategy, resulting in a more robust and efficient framework, CFBI+. We conduct extensive experiments on two popular benchmarks, i.e., DAVIS and YouTube-VOS. Without applying any simulated data for pre-training, our CFBI+ achieves the performance (J&F) of 82.9% and 82.8%, outperforming all the other state-of-the-art methods. Code: https://github.com/z-x-yang/CFBI. | ['Yi Yang', 'Yunchao Wei', 'Zongxin Yang'] | 2020-10-13 | null | null | null | null | ['one-shot-visual-object-segmentation'] | ['computer-vision'] | [ 1.38017982e-01 -1.80335119e-01 -3.12087238e-01 -1.72297031e-01
-5.62370837e-01 -3.96273226e-01 3.66809189e-01 -7.80845582e-02
-3.63801420e-01 3.67996216e-01 3.25879455e-02 -7.94074386e-02
2.48175085e-01 -6.11192882e-01 -7.30074525e-01 -9.23091650e-01
9.53976437e-02 4.36798520e-02 8.29548359e-01 2.84163773e-01
1.73923999e-01 2.70547271e-01 -1.28383803e+00 2.42022738e-01
9.01409149e-01 1.08756340e+00 3.55880022e-01 6.18202448e-01
-2.29517400e-01 6.97330475e-01 -5.60024977e-01 -4.95525837e-01
4.59343493e-01 -4.00924802e-01 -7.83606350e-01 3.08851719e-01
5.77994227e-01 -3.91508877e-01 -5.33842802e-01 1.11880374e+00
3.66180718e-01 2.10062176e-01 5.60598850e-01 -1.29754436e+00
-6.33553147e-01 2.36143917e-01 -9.87412333e-01 5.66170871e-01
5.70867807e-02 3.02294821e-01 8.86522532e-01 -9.71134782e-01
5.41766226e-01 1.22229576e+00 3.84896368e-01 5.82471132e-01
-1.14030266e+00 -6.52984977e-01 5.24917006e-01 3.58843029e-01
-1.36229169e+00 -2.80031264e-01 1.01824212e+00 -5.73096752e-01
5.55774748e-01 2.91584313e-01 6.82939947e-01 7.24863648e-01
-5.88325113e-02 1.43514228e+00 8.84962082e-01 -2.43169188e-01
1.63385309e-02 -5.32479491e-03 2.28685588e-01 6.25009894e-01
2.14242667e-01 -6.85299560e-02 -2.50905514e-01 3.13253552e-02
8.96464646e-01 2.07066581e-01 -5.86899877e-01 -4.52396482e-01
-1.29586625e+00 5.48984408e-01 5.61706424e-01 1.93410173e-01
-3.56411815e-01 3.03419083e-01 5.12258172e-01 -1.58702880e-01
4.34873551e-01 -1.81627348e-01 -3.39400381e-01 1.40452728e-01
-1.15609813e+00 1.40887529e-01 5.01977324e-01 9.83997941e-01
7.33496547e-01 8.43636990e-02 -3.51070225e-01 7.26662874e-01
6.30325615e-01 2.09106147e-01 3.76398742e-01 -1.01526523e+00
4.54483509e-01 6.11835599e-01 8.42991918e-02 -1.04708219e+00
1.62298217e-01 -3.76978040e-01 -6.04578376e-01 9.05098170e-02
3.97309333e-01 -2.95737125e-02 -9.62099910e-01 1.29691005e+00
7.76778460e-01 7.24140823e-01 -6.36995435e-02 1.05635536e+00
9.34005976e-01 9.75225449e-01 2.04821795e-01 -4.08013389e-02
1.27242684e+00 -1.61332643e+00 -6.63012266e-01 -3.32277507e-01
3.91852468e-01 -7.14387059e-01 9.09215987e-01 2.78533399e-01
-1.24693608e+00 -8.54512751e-01 -9.19038355e-01 -5.74469119e-02
-5.29225618e-02 1.96348786e-01 4.71446991e-01 4.61410493e-01
-8.37410808e-01 5.69181919e-01 -9.55066025e-01 -4.03090492e-02
7.94552207e-01 2.03512430e-01 -1.86938107e-01 -8.79854336e-02
-8.70615959e-01 4.21041429e-01 6.08773232e-01 2.35095799e-01
-9.97639477e-01 -6.18819118e-01 -7.77766883e-01 4.64152023e-02
6.45310879e-01 -5.36576211e-01 1.04825759e+00 -1.26979816e+00
-1.33556557e+00 8.69332552e-01 -2.40800172e-01 -2.04151914e-01
6.10032678e-01 -4.52634573e-01 -1.16099656e-01 4.44996446e-01
4.84432280e-02 9.03133512e-01 9.48407531e-01 -1.49222696e+00
-8.02447975e-01 -8.72932971e-02 7.90015608e-02 7.94991404e-02
-3.37022305e-01 1.95751593e-01 -1.13687086e+00 -7.59997725e-01
4.07680050e-02 -6.51824236e-01 -1.64393723e-01 3.29753935e-01
-4.19343472e-01 -3.15864563e-01 1.08198011e+00 -9.55114841e-01
1.37073350e+00 -2.38447523e+00 2.01390803e-01 -1.95551440e-01
3.97523046e-01 6.05598152e-01 -1.48176163e-01 -2.20217220e-02
4.49796766e-02 1.98168397e-01 -2.42186800e-01 -3.55914980e-01
-1.03188105e-01 2.03676835e-01 1.73377655e-02 6.44327283e-01
4.59595442e-01 1.10392976e+00 -8.48120809e-01 -9.15688097e-01
2.92339355e-01 4.19811845e-01 -6.61158383e-01 3.70072126e-01
-1.84505969e-01 3.92549127e-01 -4.33378190e-01 9.89030421e-01
7.75943875e-01 -3.97590578e-01 -1.09411709e-01 -3.01744968e-01
4.20565456e-02 -3.28495242e-02 -1.25832474e+00 1.45667064e+00
-2.99380682e-02 6.34453595e-01 3.52228880e-01 -9.98835087e-01
8.27753723e-01 9.12035480e-02 5.64808249e-01 -5.39356887e-01
1.43848434e-01 1.80244148e-01 5.63046746e-02 -4.83501345e-01
2.75668740e-01 2.64173329e-01 3.69967103e-01 1.02713108e-01
3.03627923e-02 2.76760101e-01 4.22180295e-01 3.46654773e-01
7.57733285e-01 4.95375812e-01 3.45598310e-02 -3.25689971e-01
7.23709583e-01 -2.59260714e-01 9.89716947e-01 4.87958968e-01
-7.54670382e-01 5.95039248e-01 4.70245272e-01 -1.90902531e-01
-8.20804834e-01 -1.03450215e+00 2.90381592e-02 1.00203919e+00
7.71711946e-01 -2.11522460e-01 -8.92045259e-01 -8.64123821e-01
-5.73737593e-03 4.20901716e-01 -5.98175824e-01 1.79534145e-02
-7.33995795e-01 -5.38928270e-01 1.80863410e-01 7.78907776e-01
6.98317170e-01 -1.10107899e+00 -5.69631338e-01 2.22281650e-01
-2.57769287e-01 -1.15189672e+00 -8.67406964e-01 -9.41731781e-02
-9.65699673e-01 -1.00200081e+00 -1.06272280e+00 -1.02365613e+00
6.03472650e-01 4.93404061e-01 1.03878450e+00 3.50569904e-01
-2.70925343e-01 3.46847177e-01 -4.15125340e-01 -1.07578337e-01
-1.62178338e-01 -1.42605662e-01 -3.65700364e-01 1.89067110e-01
1.86582595e-01 -2.51760185e-01 -1.01704240e+00 4.56721395e-01
-9.31566238e-01 2.72881627e-01 6.54674888e-01 8.03931236e-01
7.68690705e-01 -1.61876202e-01 3.30752462e-01 -6.32895827e-01
1.99029911e-02 -4.17076409e-01 -5.45482635e-01 3.13539237e-01
-3.39476466e-01 -5.30774057e-01 2.49170601e-01 -7.24484801e-01
-8.60983372e-01 2.80505382e-02 -1.01341799e-01 -7.11598992e-01
-1.66082025e-01 3.76804098e-02 -4.93837059e-01 -8.08081105e-02
1.03961594e-01 4.52692270e-01 -1.85131758e-01 -4.66291130e-01
2.13177651e-01 6.87312126e-01 5.50984025e-01 -4.74739790e-01
6.76911414e-01 4.48107213e-01 -5.26226640e-01 -7.62788713e-01
-7.40651965e-01 -8.11524093e-01 -6.70768321e-01 -3.97482812e-01
1.10089946e+00 -8.42047274e-01 -3.94971311e-01 4.72956717e-01
-1.25008357e+00 -5.69098115e-01 -2.18362242e-01 3.42087448e-01
-4.67498720e-01 7.24306345e-01 -7.72600174e-01 -7.09964991e-01
-3.35886210e-01 -1.33525932e+00 1.15953839e+00 5.05966783e-01
5.43247536e-02 -8.86390448e-01 -7.85650313e-02 5.65164089e-01
1.87204853e-01 2.58787751e-01 4.85641658e-01 -5.96825719e-01
-9.19116855e-01 -1.96526647e-02 -5.90046644e-01 4.88365322e-01
3.77758592e-02 2.94491738e-01 -1.00104058e+00 -2.81174302e-01
-1.36395425e-01 -2.60042213e-02 9.93967354e-01 4.06086624e-01
1.17532766e+00 -2.24239424e-01 -4.64882553e-01 7.87690282e-01
1.42443800e+00 3.99101704e-01 7.87377596e-01 4.91973698e-01
9.23297167e-01 4.64870811e-01 8.07716668e-01 1.65370584e-01
3.38038355e-01 6.72509193e-01 4.77236986e-01 -3.73494029e-01
-2.64002174e-01 2.87595578e-02 4.84584332e-01 7.43566036e-01
4.91611063e-02 -3.65156233e-01 -7.56381869e-01 7.90568352e-01
-1.92619038e+00 -8.25606525e-01 -1.44355625e-01 1.87777984e+00
8.65280211e-01 2.71037579e-01 2.59761482e-01 4.81371097e-02
9.20781493e-01 3.01259696e-01 -5.49656749e-01 -7.59873167e-02
3.46702784e-02 -3.37831117e-03 3.64943266e-01 3.09960127e-01
-1.47812390e+00 9.91593838e-01 5.18205452e+00 1.01572275e+00
-1.07310867e+00 3.22714746e-01 9.13148165e-01 -2.83263586e-02
-4.46079150e-02 -6.98588639e-02 -7.93546379e-01 7.86508203e-01
5.49738109e-01 -5.62422462e-02 1.67139798e-01 8.81377459e-01
1.00562930e-01 -1.07192397e-01 -9.06856298e-01 9.33868051e-01
-3.76669243e-02 -1.31622553e+00 -4.31815162e-02 -1.14525869e-01
8.63073766e-01 -1.70311853e-02 -2.66377926e-01 3.08306813e-01
-1.86222807e-01 -7.09125161e-01 9.35762346e-01 3.28215897e-01
4.07842815e-01 -5.32657921e-01 6.73874021e-01 1.66660756e-01
-1.46234334e+00 1.35494666e-02 -2.37851888e-01 3.95527512e-01
2.36801952e-01 4.55142319e-01 -1.47660315e-01 6.61887467e-01
9.54875112e-01 8.48729730e-01 -5.80463707e-01 1.30900323e+00
-1.71689659e-01 7.88277745e-01 -1.28898770e-01 1.71809465e-01
3.72309387e-01 -2.46470243e-01 4.06792670e-01 1.45480847e+00
-5.29331565e-02 5.70601672e-02 4.96442586e-01 8.88442397e-01
-1.57248676e-02 1.78594723e-01 -3.19614746e-02 -1.68825746e-01
2.68980891e-01 1.28703809e+00 -1.07589662e+00 -6.06157482e-01
-6.59992456e-01 1.21772492e+00 3.76515463e-02 5.08363724e-01
-1.15916765e+00 -3.94853294e-01 6.32270455e-01 1.44688964e-01
8.49584699e-01 -1.32762015e-01 -1.54915899e-02 -1.14016759e+00
5.76613098e-02 -7.63002634e-01 2.73648679e-01 -5.53617418e-01
-1.11645555e+00 4.82550442e-01 -5.34257516e-02 -1.07236767e+00
3.61476243e-01 -6.37191176e-01 -8.19754720e-01 7.37885833e-01
-1.74669850e+00 -1.06897056e+00 -4.76302207e-01 3.80549669e-01
8.04629922e-01 2.10857078e-01 8.77593979e-02 6.76313996e-01
-8.66875350e-01 5.19063830e-01 3.71776298e-02 4.53267574e-01
5.01570880e-01 -1.17472565e+00 3.60841393e-01 1.07150722e+00
1.57608762e-01 4.54180807e-01 3.84536386e-01 -6.86814070e-01
-1.16712236e+00 -1.25880432e+00 4.47943985e-01 -2.17385978e-01
6.13849163e-01 -2.52154171e-01 -1.24634957e+00 4.81276304e-01
1.71236709e-01 3.93968672e-01 5.39094687e-01 -5.35027862e-01
-2.71961577e-02 -7.38621727e-02 -9.18739796e-01 6.52235091e-01
9.82260406e-01 -2.43577912e-01 -4.36549723e-01 2.28361398e-01
8.51524889e-01 -4.19217765e-01 -7.95171082e-01 4.29795325e-01
4.41503823e-01 -9.71099257e-01 1.08372808e+00 -3.49957407e-01
5.16464651e-01 -5.82002938e-01 -8.72877091e-02 -6.23209476e-01
-3.24634343e-01 -5.07728815e-01 -5.84126711e-01 1.39335299e+00
-5.40525541e-02 -4.85976964e-01 8.71939301e-01 3.87109220e-01
-1.36282533e-01 -1.18722248e+00 -7.34506488e-01 -7.73266315e-01
-1.38142362e-01 -2.65366942e-01 2.75671005e-01 7.04269707e-01
-6.59881115e-01 -1.16074167e-01 -2.57122040e-01 1.75588280e-01
6.23203278e-01 3.54990095e-01 9.16910291e-01 -8.46435547e-01
-4.43235844e-01 -5.99766135e-01 -4.86015260e-01 -1.43521047e+00
3.25292796e-02 -7.58475304e-01 4.28702235e-02 -1.57541239e+00
3.66526365e-01 -2.91918695e-01 -4.19110656e-01 3.56128901e-01
-6.14435673e-01 4.25939739e-01 4.37901258e-01 3.62335712e-01
-8.57522249e-01 6.53025270e-01 1.55326033e+00 -2.56292552e-01
-1.23762116e-01 -6.28041923e-02 -5.29718459e-01 8.15320253e-01
7.89644599e-01 -4.75952148e-01 -1.15059361e-01 -3.57851326e-01
-4.71757114e-01 -2.06043571e-01 4.47546601e-01 -7.90820062e-01
3.10934298e-02 -2.46035114e-01 3.81073177e-01 -6.03843153e-01
1.54941022e-01 -7.38619924e-01 -2.32104659e-02 6.30112410e-01
-9.77150276e-02 -1.86484665e-01 2.81533211e-01 7.11945653e-01
-3.25203598e-01 -3.77194524e-01 8.31017911e-01 -2.12730281e-03
-1.09697914e+00 5.12461662e-01 -1.80781364e-01 1.53072119e-01
1.42753720e+00 -5.36416173e-01 -1.51495308e-01 -2.52262391e-02
-5.81358433e-01 4.52260971e-01 5.86266756e-01 3.62213373e-01
6.59948587e-01 -1.18805742e+00 -5.50303936e-01 1.09232351e-01
-4.25159559e-02 2.41269469e-01 2.76980162e-01 1.14671683e+00
-7.60972500e-01 1.77034557e-01 6.19314276e-02 -8.19496155e-01
-1.41292548e+00 6.40434504e-01 2.58491129e-01 -1.02353014e-01
-6.68338299e-01 9.89518285e-01 6.54810965e-01 -8.60604495e-02
3.84982824e-01 -3.07927549e-01 -1.17633581e-01 -1.15230698e-02
5.58758080e-01 5.09863317e-01 -3.96136910e-01 -7.49518871e-01
-4.71729964e-01 6.20660424e-01 -1.82300821e-01 2.84195870e-01
1.07874274e+00 -1.64379835e-01 2.49556564e-02 2.98238814e-01
1.19154823e+00 -1.55249052e-02 -1.75445998e+00 -3.24327439e-01
9.36953798e-02 -7.46392250e-01 2.24263668e-02 -4.49580073e-01
-1.49929500e+00 9.43918347e-01 7.18886614e-01 -3.72915938e-02
1.26204002e+00 4.94982302e-02 1.03442359e+00 -1.30522028e-01
5.97397611e-02 -9.23903942e-01 3.89722705e-01 1.41868249e-01
6.43459499e-01 -1.39042258e+00 1.63964763e-01 -5.79380512e-01
-7.38621950e-01 8.92263651e-01 7.93447495e-01 -2.53093719e-01
4.72519577e-01 1.55269176e-01 1.18489280e-01 9.10440832e-03
-5.53128123e-01 -2.88306534e-01 5.95865130e-01 3.10966909e-01
3.49534243e-01 -2.06199393e-01 -2.65158892e-01 5.36283016e-01
6.36618316e-01 -7.97149539e-02 1.76742002e-01 1.04261899e+00
-4.33299273e-01 -8.87337923e-01 -3.52416933e-01 3.45743597e-01
-6.22803569e-01 -3.19515467e-02 -2.45882466e-01 8.70695353e-01
2.81385005e-01 6.25364840e-01 -7.26190116e-03 -2.12819606e-01
4.40481976e-02 -9.98734385e-02 3.62494349e-01 -5.69643259e-01
-5.53779960e-01 5.17129600e-01 -3.55599612e-01 -6.81668460e-01
-6.52453423e-01 -6.83459580e-01 -1.31985652e+00 6.29611034e-03
-6.94218040e-01 -1.05275050e-01 4.18721020e-01 6.29469812e-01
2.50080526e-01 7.20623195e-01 4.93844032e-01 -1.19264579e+00
-2.09427908e-01 -6.30030453e-01 -3.33802164e-01 3.90732527e-01
2.20572084e-01 -6.95231378e-01 -2.08293945e-01 2.09950745e-01] | [9.294143676757812, -0.13594751060009003] |
0f95a327-1eef-467a-a6d2-86845dc1af34 | preventing-zero-shot-transfer-degradation-in | 2303.06628 | null | https://arxiv.org/abs/2303.06628v1 | https://arxiv.org/pdf/2303.06628v1.pdf | Preventing Zero-Shot Transfer Degradation in Continual Learning of Vision-Language Models | Continual learning (CL) can help pre-trained vision-language models efficiently adapt to new or under-trained data distributions without re-training. Nevertheless, during the continual training of the Contrastive Language-Image Pre-training (CLIP) model, we observe that the model's zero-shot transfer ability significantly degrades due to catastrophic forgetting. Existing CL methods can mitigate forgetting by replaying previous data. However, since the CLIP dataset is private, replay methods cannot access the pre-training dataset. In addition, replaying data of previously learned downstream tasks can enhance their performance but comes at the cost of sacrificing zero-shot performance. To address this challenge, we propose a novel method ZSCL to prevent zero-shot transfer degradation in the continual learning of vision-language models in both feature and parameter space. In the feature space, a reference dataset is introduced for distillation between the current and initial models. The reference dataset should have semantic diversity but no need to be labeled, seen in pre-training, or matched image-text pairs. In parameter space, we prevent a large parameter shift by averaging weights during the training. We propose a more challenging Multi-domain Task Incremental Learning (MTIL) benchmark to evaluate different methods, where tasks are from various domains instead of class-separated in a single dataset. Our method outperforms other methods in the traditional class-incremental learning setting and the MTIL by 9.7% average score. Our code locates at https://github.com/Thunderbeee/ZSCL. | ['Yang You', 'Xiangyu Yue', 'Ziheng Qin', 'Kai Wang', 'Mingyuan Ma', 'Zangwei Zheng'] | 2023-03-12 | null | null | null | null | ['class-incremental-learning'] | ['computer-vision'] | [ 2.43353352e-01 -2.43514761e-01 -1.44783139e-01 -3.01505774e-01
-7.92997360e-01 -4.74612683e-01 8.02360117e-01 -2.05602020e-01
-7.42663622e-01 7.57456601e-01 -1.26529694e-01 -1.38384283e-01
2.83010483e-01 -3.00344795e-01 -1.08573496e+00 -7.19015300e-01
4.35445637e-01 4.14316654e-01 6.69553936e-01 5.02168685e-02
5.50793931e-02 2.38246039e-01 -1.63013864e+00 4.90911245e-01
1.00648773e+00 6.36859536e-01 7.56859839e-01 5.51169395e-01
-2.44180396e-01 7.19681144e-01 -4.88234103e-01 -2.22203240e-01
3.32422167e-01 -4.34883147e-01 -5.38597941e-01 2.38870963e-01
8.36545587e-01 -5.32425880e-01 -5.04198372e-01 1.12114668e+00
5.42348504e-01 2.87756294e-01 5.05927086e-01 -1.49845600e+00
-1.03401923e+00 4.04516935e-01 -5.39989352e-01 2.68826168e-02
-1.33589059e-01 5.63391507e-01 3.34502548e-01 -1.46430993e+00
7.33450949e-01 1.08076608e+00 6.46233857e-01 9.57468212e-01
-1.29272151e+00 -8.38773668e-01 4.85603631e-01 3.87742549e-01
-1.37561488e+00 -8.18779826e-01 6.16447031e-01 -4.19998080e-01
8.12892854e-01 -3.08096647e-01 4.70499694e-01 1.49124002e+00
1.29288942e-01 1.06853759e+00 1.03575683e+00 -4.71498460e-01
3.00662577e-01 4.68617201e-01 2.20946044e-01 7.72080600e-01
2.46995673e-01 1.11358605e-01 -8.67331862e-01 -2.16759313e-02
4.07810360e-01 3.26839268e-01 -3.07050288e-01 -6.28036261e-01
-1.16324449e+00 4.21185911e-01 2.35920489e-01 2.11758882e-01
-1.23474590e-01 3.13714370e-02 3.58592093e-01 7.13054061e-01
4.19067085e-01 1.58683851e-01 -6.53068721e-01 1.42793030e-01
-1.01049256e+00 3.89040587e-03 3.38214129e-01 1.03015852e+00
1.03035450e+00 1.06176160e-01 -4.84357059e-01 9.58105385e-01
-2.77904365e-02 5.77609301e-01 8.07063520e-01 -8.31159174e-01
3.85713190e-01 3.84127855e-01 -3.69021632e-02 -2.51489341e-01
5.34010725e-03 -3.23156297e-01 -7.92257845e-01 2.34285131e-01
4.36755389e-01 -1.24909982e-01 -1.31760442e+00 1.94485104e+00
2.77788460e-01 4.56719279e-01 2.48156562e-01 5.17716050e-01
5.47277451e-01 6.77966237e-01 6.47745281e-02 -3.96260202e-01
8.86704266e-01 -1.50234294e+00 -4.69390482e-01 -4.72249210e-01
5.57006776e-01 -6.80731714e-01 1.72216916e+00 3.65949541e-01
-7.78091609e-01 -8.38050008e-01 -9.68848646e-01 -1.14111423e-01
-3.28896463e-01 7.61363953e-02 8.95055160e-02 1.84651926e-01
-1.19941092e+00 4.94516343e-01 -8.57415438e-01 -3.44968826e-01
5.85457742e-01 -1.74052101e-02 -4.06764656e-01 -5.99178851e-01
-1.00690985e+00 7.14295864e-01 4.14245516e-01 -4.55362231e-01
-1.31325185e+00 -1.14816296e+00 -5.59562206e-01 -1.20628744e-01
6.52390301e-01 -7.16684878e-01 1.37646866e+00 -1.33406794e+00
-1.40055299e+00 7.21825063e-01 -3.17778796e-01 -5.46901166e-01
7.70579755e-01 -4.19910312e-01 -1.96534693e-01 -1.73242256e-01
5.59599437e-02 1.00832105e+00 1.41676092e+00 -1.36427438e+00
-5.73260188e-01 -2.55940259e-01 -1.51554406e-01 3.89610350e-01
-4.88542199e-01 -5.42973459e-01 -6.09873891e-01 -5.69671512e-01
-1.58070773e-01 -1.06085861e+00 1.84161082e-01 3.96413088e-01
-4.23576050e-02 -5.96004650e-02 1.11725652e+00 -3.87386769e-01
9.91850495e-01 -2.41478634e+00 -2.89840549e-02 -6.90485001e-01
1.04674831e-01 6.52837813e-01 -5.66482127e-01 6.91541433e-02
9.80671719e-02 -1.41779944e-01 -3.86955738e-01 -8.32532167e-01
-3.96086931e-01 3.55642915e-01 -5.73815048e-01 1.39033124e-01
1.24442108e-01 1.00325537e+00 -9.42496359e-01 -2.41582766e-01
7.27539212e-02 3.50638330e-01 -4.77238894e-01 1.62343696e-01
-5.70443273e-01 4.05907929e-01 8.99885818e-02 4.69780028e-01
8.41295958e-01 -3.24683070e-01 -1.36151984e-01 -2.57434845e-02
2.22568214e-02 -1.91148028e-01 -7.44644284e-01 2.11269593e+00
-4.78442699e-01 6.34633720e-01 -3.58012080e-01 -8.91750157e-01
6.08804882e-01 4.84961458e-02 1.43801972e-01 -9.80728209e-01
-2.60869324e-01 3.63964103e-02 -6.96076751e-02 -2.95858234e-01
4.33071822e-01 -4.51377481e-02 1.59695342e-01 5.05067527e-01
4.19605702e-01 3.44692692e-02 -3.23509015e-02 3.01854193e-01
9.68771756e-01 3.26777175e-02 5.13052866e-02 5.77737801e-02
3.44434798e-01 1.51728345e-02 5.80989718e-01 1.09844267e+00
-4.92552072e-01 6.17434382e-01 8.27545673e-02 -4.05095160e-01
-1.13781893e+00 -1.28189778e+00 -1.65634081e-02 1.38140464e+00
7.21690953e-02 -9.44720134e-02 -4.70343113e-01 -9.99661922e-01
9.52581167e-02 1.14327943e+00 -5.33566773e-01 -7.78019786e-01
-3.02594155e-01 -4.85050172e-01 3.49139303e-01 2.16412172e-01
8.24731529e-01 -8.67291987e-01 -5.40514231e-01 9.78697911e-02
-9.39935446e-02 -1.06657135e+00 -7.95371532e-01 1.37981817e-01
-9.35968995e-01 -8.38072956e-01 -1.05497742e+00 -8.22843671e-01
8.27975154e-01 8.82100761e-01 1.00727832e+00 -3.28625739e-02
-1.74870104e-01 7.31532216e-01 -3.04963201e-01 -3.68075818e-01
-5.60412645e-01 7.13177174e-02 2.82654107e-01 2.55422145e-02
4.00008500e-01 -3.76542956e-01 -5.58468163e-01 2.76626199e-01
-9.14251864e-01 3.48201931e-01 5.05375326e-01 1.25822127e+00
7.49087989e-01 -1.80985928e-01 6.42186463e-01 -8.51361036e-01
3.84792805e-01 -5.31543970e-01 -5.23610473e-01 6.31400645e-01
-9.57436740e-01 1.62640467e-01 6.14931941e-01 -1.09550428e+00
-1.30698550e+00 8.69243816e-02 2.83784896e-01 -1.01137936e+00
5.63518070e-02 1.70389608e-01 6.83310926e-02 -1.74396057e-02
8.08904052e-01 6.59833491e-01 1.50082558e-01 -4.65096831e-01
5.35032213e-01 4.53832000e-01 5.72479069e-01 -5.20451665e-01
7.45877624e-01 4.55478996e-01 -5.51844835e-01 -6.62702739e-01
-9.98167515e-01 -2.26918265e-01 -6.43962204e-01 -1.53264523e-01
3.47741097e-01 -1.24325943e+00 1.36139125e-01 9.82563674e-01
-1.06442440e+00 -8.67142797e-01 -5.80457389e-01 2.80524999e-01
-3.94041598e-01 3.10436368e-01 -2.10695535e-01 -5.92058659e-01
-2.30282664e-01 -1.01505613e+00 8.16772163e-01 2.16460943e-01
2.89572597e-01 -8.22979450e-01 8.96764845e-02 1.06755167e-01
4.92490798e-01 -4.43509609e-01 9.20828998e-01 -6.45550549e-01
-5.24358451e-01 1.62356794e-01 -2.22532645e-01 7.45742559e-01
2.92557001e-01 -2.41089791e-01 -1.05853987e+00 -7.90189028e-01
1.59941331e-01 -7.61202097e-01 1.37653852e+00 1.87549964e-01
1.09137678e+00 -3.25783372e-01 -3.24933082e-01 5.56674063e-01
1.36972558e+00 2.29706332e-01 5.59752584e-01 2.03366071e-01
6.36573315e-01 3.30510348e-01 6.89814866e-01 3.87287110e-01
3.78355175e-01 3.35317880e-01 1.41245693e-01 1.30505651e-01
-7.10521638e-01 -5.01872718e-01 8.56527984e-01 6.70218825e-01
4.88613725e-01 -2.50385940e-01 -9.44013119e-01 7.11552680e-01
-1.88712072e+00 -8.07809949e-01 4.58251864e-01 2.57535172e+00
1.03056657e+00 2.54638374e-01 -2.81644821e-01 -3.95050824e-01
6.48032427e-01 -4.62771952e-02 -1.19346857e+00 9.50177163e-02
-2.29896590e-01 -8.60012248e-02 4.01131690e-01 6.80092514e-01
-9.17604983e-01 1.41343677e+00 5.29528761e+00 9.04985070e-01
-1.41360271e+00 5.98732114e-01 6.72172785e-01 -5.74370980e-01
-2.19550446e-01 6.96578948e-03 -1.04316390e+00 5.72939277e-01
9.11908150e-01 -4.26409841e-01 5.35546720e-01 9.39002931e-01
-2.18068883e-01 -3.25604707e-01 -1.06799889e+00 1.08031476e+00
3.67794544e-01 -1.30033839e+00 4.55834836e-01 -3.95338327e-01
9.75659251e-01 5.44667721e-01 5.23504794e-01 8.34604919e-01
3.04613650e-01 -5.08757055e-01 6.02110803e-01 7.32682943e-01
1.18800950e+00 -3.79129648e-01 2.77926385e-01 5.86839139e-01
-7.29053438e-01 -2.52046019e-01 -5.59485376e-01 2.37734810e-01
-5.56649081e-02 5.25340199e-01 -1.05826628e+00 2.01988012e-01
7.23389745e-01 6.69442654e-01 -8.72340322e-01 8.96915376e-01
-4.21691388e-02 5.80724359e-01 -1.08842820e-01 4.04324710e-01
6.46694377e-02 1.34522915e-01 5.75865686e-01 8.89339685e-01
5.04123747e-01 -3.12037647e-01 1.62140176e-01 6.73889935e-01
-1.79485232e-01 -2.46078730e-01 -8.02786410e-01 7.60194436e-02
5.61501563e-01 6.98125541e-01 -3.01584780e-01 -6.52801991e-01
-6.07984424e-01 1.56562662e+00 5.46252668e-01 6.36626899e-01
-7.71699250e-01 -1.81826707e-02 6.34656191e-01 1.61586538e-01
2.21746668e-01 -2.70744979e-01 5.61693758e-02 -1.46107054e+00
5.02747558e-02 -7.38518476e-01 2.32964531e-01 -8.11789334e-01
-1.33449531e+00 5.40205240e-01 -1.81176662e-02 -1.15050173e+00
-1.98539004e-01 -3.09394389e-01 -3.32618386e-01 6.07513130e-01
-1.77890873e+00 -1.12541366e+00 -4.02377993e-01 9.45677936e-01
1.19995344e+00 -6.78084254e-01 5.57866335e-01 2.51328945e-01
-4.95255738e-01 9.63000536e-01 4.34827685e-01 -2.74664253e-01
1.37241745e+00 -8.34168613e-01 4.04688030e-01 9.12342250e-01
2.37921417e-01 4.34406042e-01 4.16697443e-01 -6.94922566e-01
-1.25905669e+00 -1.49883735e+00 6.90224826e-01 -4.92047012e-01
5.06247938e-01 -4.92313445e-01 -1.42224824e+00 7.85634875e-01
1.11786492e-01 2.95975149e-01 4.20323431e-01 -2.34629437e-01
-7.44553864e-01 -2.99561948e-01 -1.09795475e+00 8.19597006e-01
1.20963776e+00 -5.89192510e-01 -5.66867828e-01 4.20716375e-01
1.19439816e+00 -1.37081668e-01 -1.45451218e-01 2.53655493e-01
4.14974064e-01 -8.53587925e-01 8.68213594e-01 -5.37391663e-01
1.56110659e-01 -2.22600743e-01 -1.59713864e-01 -1.33457851e+00
-2.91122288e-01 -4.13963020e-01 -2.82947034e-01 1.13423181e+00
2.50307500e-01 -6.98173106e-01 5.40329576e-01 3.66679728e-01
-9.56568494e-02 -4.06606019e-01 -1.00258148e+00 -1.13902175e+00
2.45064840e-01 -3.56322080e-01 2.38044769e-01 8.76754224e-01
-5.96350491e-01 6.07823789e-01 -4.97504264e-01 3.40285450e-02
6.74376845e-01 -2.28611588e-01 8.79294217e-01 -1.04783690e+00
-2.86561221e-01 -1.80433437e-01 1.50945097e-01 -1.11129951e+00
2.34457910e-01 -1.02152956e+00 7.79775158e-02 -1.18355060e+00
4.15732145e-01 -5.20098448e-01 -6.09209776e-01 7.94708729e-01
-2.27647379e-01 3.09161725e-04 4.58928674e-01 5.87887347e-01
-8.80505621e-01 9.01031613e-01 1.29279041e+00 -3.74260545e-01
-4.52406079e-01 -1.23885930e-01 -5.13755679e-01 5.33268809e-01
7.76100397e-01 -6.42822862e-01 -9.29733336e-01 -7.93902636e-01
-1.13083676e-01 -3.11973542e-01 4.78783041e-01 -1.15874994e+00
5.31247854e-01 -1.37099981e-01 3.20222259e-01 -4.94015247e-01
2.09516212e-01 -6.94062591e-01 -3.84755991e-02 5.44162214e-01
-2.62194037e-01 -1.74639270e-01 4.94159460e-01 7.94500649e-01
-4.57983278e-02 -2.63137847e-01 1.11658728e+00 -2.37100989e-01
-1.10326850e+00 5.00185668e-01 -2.16412455e-01 1.33720830e-01
1.02578282e+00 -1.35845736e-01 -5.79575956e-01 -7.44365975e-02
-7.95370638e-01 3.57286721e-01 7.57231474e-01 7.13455319e-01
8.29422414e-01 -1.20963097e+00 -6.12838387e-01 3.41153473e-01
4.69524086e-01 1.49468392e-01 6.62098110e-01 6.94110811e-01
-1.51280791e-03 5.37027419e-02 -3.53417546e-01 -7.50931144e-01
-1.16634560e+00 8.09166491e-01 2.92738646e-01 -1.68277889e-01
-6.17763281e-01 1.04628861e+00 4.45124656e-01 -4.74997312e-01
4.21201915e-01 -5.33041637e-03 2.28637367e-01 2.01694220e-01
7.14154124e-01 1.06120393e-01 -5.33932894e-02 -9.36323702e-02
-1.18485987e-01 3.15113127e-01 -7.72687137e-01 -1.99647725e-01
1.06005204e+00 -4.12795395e-01 2.33647466e-01 9.53493357e-01
1.09479797e+00 -4.09323812e-01 -1.87032092e+00 -8.77369702e-01
-2.05010101e-01 -3.60643536e-01 -2.63096392e-02 -1.00615513e+00
-8.64883304e-01 9.55613852e-01 1.00118124e+00 -5.18233538e-01
1.05624890e+00 -2.00677104e-02 7.92888820e-01 7.67845929e-01
4.23566073e-01 -1.23996866e+00 7.31122971e-01 8.77858341e-01
1.08769000e+00 -1.53263152e+00 -2.41190389e-01 2.74877936e-01
-1.03028047e+00 7.59431422e-01 9.78090465e-01 6.44959882e-02
7.25105643e-01 1.23788178e-01 3.53565924e-02 3.21735680e-01
-1.20832002e+00 -1.20499842e-01 2.31469162e-02 7.34876692e-01
-1.14011429e-01 -2.22757861e-01 2.04558030e-01 4.81158197e-01
2.10793182e-01 2.36850321e-01 4.47039247e-01 9.29397225e-01
-6.08700275e-01 -9.43187773e-01 -8.46993625e-02 2.60638058e-01
2.02324599e-01 -3.25657815e-01 -2.10204795e-01 5.78252733e-01
2.44126976e-01 5.57486832e-01 1.87463507e-01 -4.10545141e-01
2.47053802e-01 5.60766578e-01 5.46057045e-01 -7.73884475e-01
-6.17235787e-02 -9.85604823e-02 -4.91796404e-01 -3.39974999e-01
-8.85220915e-02 -5.91053426e-01 -1.01695740e+00 -1.70695052e-01
-9.63415354e-02 -2.14003637e-01 3.59324962e-01 7.11819410e-01
6.95373833e-01 4.66057450e-01 6.34396851e-01 -6.05205119e-01
-7.13605702e-01 -8.40421021e-01 -3.08926940e-01 3.22513461e-01
6.11124218e-01 -8.14596534e-01 -4.15266335e-01 2.69915611e-01] | [9.932376861572266, 3.220952033996582] |
19e28185-2738-4b35-8022-9430c12cb4a7 | industrial-scene-change-detection-using-deep | 2212.14278 | null | https://arxiv.org/abs/2212.14278v1 | https://arxiv.org/pdf/2212.14278v1.pdf | Industrial Scene Change Detection using Deep Convolutional Neural Networks | Finding and localizing the conceptual changes in two scenes in terms of the presence or removal of objects in two images belonging to the same scene at different times in special care applications is of great significance. This is mainly due to the fact that addition or removal of important objects for some environments can be harmful. As a result, there is a need to design a program that locates these differences using machine vision. The most important challenge of this problem is the change in lighting conditions and the presence of shadows in the scene. Therefore, the proposed methods must be resistant to these challenges. In this article, a method based on deep convolutional neural networks using transfer learning is introduced, which is trained with an intelligent data synthesis process. The results of this method are tested and presented on the dataset provided for this purpose. It is shown that the presented method is more efficient than other methods and can be used in a variety of real industrial environments. | ['Hassan Shahbazi', 'Kiavash azimi', 'Ehsan Rahnama', 'Ali Atghaei'] | 2022-12-29 | null | null | null | null | ['scene-change-detection', 'change-detection'] | ['computer-vision', 'computer-vision'] | [ 4.70925868e-01 -5.39237380e-01 5.88741481e-01 -4.63569850e-01
3.20396096e-01 -3.43369335e-01 4.64886069e-01 2.55339444e-01
-2.58962721e-01 5.44649482e-01 -2.16317236e-01 -6.55533820e-02
-2.98083723e-01 -8.14543366e-01 -5.09409785e-01 -7.35979140e-01
1.53796300e-01 1.00366540e-01 3.95727277e-01 -4.06263798e-01
5.25813758e-01 9.41113234e-01 -1.87570465e+00 3.95806819e-01
3.23483944e-01 8.09462190e-01 8.31757009e-01 4.14625853e-01
-2.62328088e-01 6.64012253e-01 -9.55778301e-01 2.15146571e-01
3.54353994e-01 -5.36744773e-01 -4.51741040e-01 4.57660377e-01
4.94349271e-01 -1.66101262e-01 2.80912250e-01 1.05339956e+00
5.93870938e-01 3.52944076e-01 4.20649260e-01 -9.10997033e-01
-2.38041699e-01 2.74631791e-02 -4.12579954e-01 2.61555254e-01
2.78047919e-01 -5.05893491e-02 1.23869859e-01 -5.61034024e-01
2.73905486e-01 1.15691507e+00 3.95333558e-01 6.77388906e-02
-1.09815228e+00 -4.28009331e-01 -1.23484023e-01 5.42177916e-01
-1.08157933e+00 -1.97318032e-01 1.16159427e+00 -6.38678014e-01
6.86214268e-01 4.03286099e-01 4.28843409e-01 7.03903258e-01
6.45235240e-01 9.82567519e-02 1.36390460e+00 -9.29812670e-01
3.74279827e-01 5.51946819e-01 6.33438875e-04 4.07680750e-01
3.31301302e-01 -1.64280176e-01 -7.97262937e-02 3.67011696e-01
5.62783599e-01 2.17182115e-01 -3.64163876e-01 -4.76907372e-01
-8.94570589e-01 5.00171185e-01 3.66541356e-01 8.39578032e-01
-3.03161919e-01 -4.02572118e-02 4.42876965e-01 2.79132098e-01
2.05428496e-01 5.44455409e-01 -5.69113791e-01 9.51198339e-02
-5.44960916e-01 2.49419827e-02 6.74591243e-01 5.16590953e-01
7.87647009e-01 -1.33364692e-01 2.08436534e-01 6.57462239e-01
1.08930416e-01 2.00044960e-01 5.55608690e-01 -5.55180550e-01
7.24742413e-02 7.17698276e-01 2.85575893e-02 -1.54725707e+00
-3.99393529e-01 -3.16066295e-01 -6.26287401e-01 9.35256958e-01
2.96949714e-01 6.28209189e-02 -8.83992612e-01 1.30335379e+00
5.24181902e-01 1.37478337e-01 7.17903301e-02 7.10056245e-01
8.31160843e-01 7.19505370e-01 -1.36656091e-01 -2.50657916e-01
1.28668308e+00 -8.15148056e-01 -1.17018461e+00 -9.95161980e-02
1.27484277e-01 -1.21343100e+00 8.83097649e-01 5.00727177e-01
-6.17887855e-01 -1.06439149e+00 -1.13497305e+00 8.11856464e-02
-8.67323339e-01 3.28458846e-01 3.22069913e-01 5.92578232e-01
-7.68374026e-01 7.58220017e-01 -4.44333673e-01 -8.17201197e-01
-2.27483079e-01 3.82904142e-01 -7.81650916e-02 1.04278445e-01
-8.72536361e-01 1.10431492e+00 5.79694092e-01 5.70466697e-01
-4.06652987e-01 -3.30144107e-01 -6.03792608e-01 3.92645225e-03
3.64412218e-01 -8.40248764e-02 1.02906311e+00 -1.52922380e+00
-1.50818801e+00 6.94731832e-01 1.75496325e-01 -1.64347321e-01
6.83140635e-01 -2.56836057e-01 -5.85588515e-01 -1.41328603e-01
-1.33138597e-01 -1.85948741e-02 8.98925960e-01 -1.38537693e+00
-7.50987947e-01 -3.42690080e-01 2.92908885e-02 1.80423290e-01
-1.29487619e-01 1.91463679e-01 -2.71494120e-01 -4.19339448e-01
-7.93808848e-02 -7.33662724e-01 -6.21759146e-02 -4.55320142e-02
-1.67401388e-01 -4.54035327e-02 1.35155511e+00 -6.83368683e-01
7.83701003e-01 -2.12375164e+00 -1.77299708e-01 1.71054155e-01
-3.72949570e-01 5.81072032e-01 2.34291539e-01 4.46276933e-01
-4.74308103e-01 -2.65287995e-01 -1.61106184e-01 1.89907234e-02
-4.34869796e-01 2.96543807e-01 3.17418948e-02 4.18516189e-01
7.28350952e-02 2.26655547e-02 -4.89384353e-01 -3.81671607e-01
8.37336123e-01 4.42704588e-01 7.61490986e-02 3.43728334e-01
-5.60836727e-03 7.77473152e-01 -2.42097050e-01 2.42215827e-01
7.84529686e-01 4.98510838e-01 1.84288591e-01 -5.00955284e-01
-4.27670509e-01 -2.61056215e-01 -1.53836679e+00 1.31557167e+00
-4.64326292e-01 7.32415974e-01 -1.57302856e-01 -1.06125021e+00
1.10694289e+00 3.45937163e-01 3.61377060e-01 -9.06963587e-01
5.06141782e-01 2.16627970e-01 -6.75634108e-03 -1.05112040e+00
3.71548474e-01 -7.23635778e-03 3.67402256e-01 1.69365689e-01
-4.00160134e-01 -3.69239569e-01 3.63859624e-01 -5.37072539e-01
7.62639642e-01 2.69030958e-01 4.17355955e-01 -2.08752796e-01
9.36797917e-01 -2.05976442e-01 6.13039851e-01 2.93755203e-01
-1.56210721e-01 3.87972802e-01 1.72668248e-01 -8.32526267e-01
-9.04232740e-01 -4.29673344e-01 1.93272289e-02 7.17847109e-01
4.86207485e-01 2.35128552e-01 -6.20698035e-01 -3.72611672e-01
-2.46134534e-01 8.47049952e-01 -5.62636495e-01 -1.99502945e-01
-6.89819276e-01 -5.65720320e-01 -3.16884995e-01 1.18789949e-01
5.16274691e-01 -1.52125335e+00 -1.36863530e+00 1.42230406e-01
1.18688099e-01 -1.05009425e+00 1.85059786e-01 3.72311294e-01
-8.25893342e-01 -1.21623921e+00 -4.23813671e-01 -9.99815404e-01
8.18684042e-01 5.22903860e-01 8.01009893e-01 1.28101975e-01
-6.62490785e-01 1.63685784e-01 -6.38309479e-01 -8.03537428e-01
-7.92469740e-01 -2.32376426e-01 -2.33747274e-01 1.59913942e-01
3.91237587e-01 -2.56090522e-01 -4.05964106e-01 3.82086158e-01
-1.24075782e+00 1.28118530e-01 3.94100875e-01 4.39217418e-01
2.37977207e-01 8.34723592e-01 2.85626203e-01 -7.86202788e-01
6.64626777e-01 -1.22034810e-01 -8.63661587e-01 3.67865473e-01
-4.78605151e-01 -6.38887137e-02 6.73782587e-01 -3.17762822e-01
-1.40519345e+00 2.24581495e-01 2.41795242e-01 -4.72709946e-02
-8.27971578e-01 1.30636618e-01 -3.42759639e-01 -6.80844337e-02
6.05401576e-01 -2.05399245e-01 -4.35030870e-02 -6.11952543e-01
-5.41891940e-02 7.61509597e-01 2.65483618e-01 -7.73542672e-02
5.98683238e-01 4.31596845e-01 2.00236827e-01 -1.02029276e+00
-3.81562889e-01 -5.03735006e-01 -8.42361152e-01 -6.27904773e-01
1.00108171e+00 -2.71981627e-01 -2.52153993e-01 7.12611556e-01
-1.28824782e+00 -1.57314733e-01 -2.28699982e-01 4.32455927e-01
-3.84198099e-01 4.87665772e-01 -1.29906252e-01 -7.54646361e-01
-1.13880768e-01 -1.29241478e+00 5.26320636e-01 4.25781369e-01
6.71152025e-02 -7.93071330e-01 1.95364486e-02 1.68283418e-01
3.31100166e-01 5.00243902e-01 9.70420897e-01 -2.96001881e-01
-3.24541301e-01 -1.60615593e-01 -1.09133311e-02 7.69286633e-01
9.73673284e-01 3.83430630e-01 -9.21343148e-01 -1.43019915e-01
4.85619396e-01 1.85324237e-01 4.44021016e-01 3.27451229e-01
1.07237911e+00 1.52736664e-01 -1.95516914e-01 2.43357033e-01
1.81217241e+00 9.74046767e-01 7.95842528e-01 6.36148274e-01
5.39953828e-01 8.66417527e-01 8.33127916e-01 4.12278920e-01
-4.15628791e-01 8.03838372e-01 6.70618892e-01 -6.06968880e-01
-3.12143415e-01 4.88749892e-01 1.75996482e-01 5.31467438e-01
-1.03483766e-01 -2.22979024e-01 -6.58301950e-01 3.85783404e-01
-1.77221060e+00 -9.50532377e-01 -4.64022219e-01 2.19983387e+00
4.46204901e-01 4.02578264e-02 -3.94293964e-01 5.41843235e-01
1.09659791e+00 -8.15994516e-02 -2.57408351e-01 -9.49848771e-01
9.27339941e-02 2.30284736e-01 3.72655004e-01 1.71686247e-01
-1.19380832e+00 5.27260721e-01 5.43216038e+00 3.66688251e-01
-1.44727504e+00 -1.15676984e-01 5.24791121e-01 4.91101205e-01
4.15701270e-01 -2.89260507e-01 -2.01015681e-01 5.71212351e-01
5.06588817e-01 3.03250968e-01 2.74529874e-01 7.99590826e-01
4.91922528e-01 -6.07050180e-01 -8.94329131e-01 9.53147411e-01
4.42721188e-01 -6.48125529e-01 -3.06944370e-01 -4.24176395e-01
6.96149409e-01 -5.14905095e-01 -4.12966050e-02 -2.95074612e-01
-1.54643863e-01 -7.12735116e-01 6.28819823e-01 5.45002639e-01
1.58850849e-01 -9.17261839e-01 1.09510458e+00 1.31192893e-01
-9.02979970e-01 -1.24123424e-01 -3.83314341e-01 -2.46113554e-01
1.83759611e-02 6.35098875e-01 -1.20816183e+00 4.65333819e-01
8.57134044e-01 1.51941434e-01 -6.87044024e-01 1.30199301e+00
-3.09137553e-01 1.71435580e-01 -1.49205938e-01 -2.42815062e-01
-1.70343801e-01 -3.00220788e-01 2.50513971e-01 1.17171597e+00
5.71162939e-01 -2.81199247e-01 -1.77639071e-02 7.48526216e-01
4.02032763e-01 3.34286600e-01 -8.54692400e-01 2.64828414e-01
-1.28411263e-01 1.33383262e+00 -1.09321797e+00 -2.45657295e-01
-4.32043076e-01 1.22680867e+00 -2.43824080e-01 3.59183609e-01
-8.57687771e-01 -5.73316813e-01 2.56609648e-01 1.62088439e-01
2.37459645e-01 -1.95709884e-01 -5.25226593e-02 -3.96771044e-01
-5.13046868e-02 -7.30705738e-01 4.96411510e-02 -1.04142618e+00
-1.00604010e+00 5.41222453e-01 8.15095976e-02 -1.15222228e+00
-7.48692229e-02 -8.39380622e-01 -4.85287368e-01 7.85955131e-01
-1.36550939e+00 -8.75514269e-01 -7.03559399e-01 5.52520454e-01
9.24061537e-01 -1.26310930e-01 7.10187376e-01 4.00474131e-01
-5.19999385e-01 -1.32164553e-01 1.86775088e-01 -1.35402977e-01
8.42445970e-01 -1.09352338e+00 -8.94025117e-02 9.79173541e-01
-1.53119147e-01 1.85130015e-01 1.11478388e+00 -5.28852999e-01
-8.92268896e-01 -9.40786660e-01 6.85760736e-01 4.67476547e-02
1.61761329e-01 -1.39779508e-01 -8.31818819e-01 4.18218136e-01
5.77304244e-01 -2.83260137e-01 4.08079267e-01 -3.04945797e-01
2.45143667e-01 -5.39608300e-01 -1.28651357e+00 4.97147501e-01
3.20930570e-01 -1.10097662e-01 -8.15365434e-01 3.73622298e-01
3.89673561e-01 -3.23898166e-01 -3.39971125e-01 4.31422055e-01
3.15920174e-01 -1.38874674e+00 6.63408220e-01 -1.46438450e-01
2.56737202e-01 -5.70661426e-01 7.45708495e-02 -1.33520424e+00
-1.17082901e-01 -2.62664884e-01 5.73444307e-01 1.14935839e+00
3.85075994e-02 -5.57947516e-01 4.03426588e-01 4.58000988e-01
2.60123294e-02 -2.97527872e-02 -5.71390748e-01 -5.14801919e-01
-5.18532872e-01 -9.55493227e-02 4.91196096e-01 9.19343829e-01
-6.03853166e-01 1.16477609e-01 -1.38900295e-01 2.66363442e-01
1.60583332e-01 2.07425281e-01 9.92667496e-01 -1.41727626e+00
-1.02701068e-01 -9.38496366e-02 -6.71038449e-01 -1.61251109e-02
-1.66777268e-01 -1.76319569e-01 4.16960329e-01 -1.61013293e+00
1.51805421e-02 -2.97133654e-01 -3.93958479e-01 3.20851505e-01
7.74680972e-02 7.09431395e-02 2.57037222e-01 -6.68877512e-02
-1.47127202e-02 2.37610579e-01 1.08351433e+00 -6.95954859e-02
-3.01498204e-01 1.16754755e-01 -3.28040794e-02 9.26096022e-01
1.27881265e+00 -3.32396358e-01 -5.16492009e-01 -4.26339865e-01
1.17980517e-01 -4.22728896e-01 2.28932396e-01 -1.60346460e+00
6.52479902e-02 -1.81869298e-01 4.50912058e-01 -6.18635118e-01
2.68535078e-01 -1.59285617e+00 4.33838129e-01 6.24945700e-01
-5.91713786e-02 3.56641650e-01 3.00951183e-01 2.95042783e-01
-1.59989402e-01 -7.10877776e-01 8.39509785e-01 -3.58978122e-01
-1.15493786e+00 -5.44900775e-01 -4.76759821e-01 -8.59409273e-01
1.47889614e+00 -5.44603646e-01 -3.10975611e-02 -2.62574553e-01
-4.19176549e-01 -3.74344915e-01 2.26034001e-01 5.72986782e-01
4.56964910e-01 -9.22068238e-01 -2.46092498e-01 4.36559081e-01
1.48590803e-01 -1.09972030e-01 3.54731768e-01 5.02615809e-01
-1.07698989e+00 2.02793017e-01 -7.45627046e-01 -3.63647312e-01
-1.86059451e+00 8.08459759e-01 3.41801912e-01 1.25611633e-01
-3.91924262e-01 2.82924950e-01 1.32048383e-01 -2.23282084e-01
2.25539953e-01 -6.15705431e-01 -5.77630579e-01 1.67486798e-02
4.65208262e-01 5.41158319e-01 4.52395141e-01 -6.03738666e-01
-8.76806229e-02 7.44459748e-01 2.59723783e-01 1.79483682e-01
1.26470804e+00 -1.84437454e-01 -3.99618179e-01 5.89223444e-01
1.04610002e+00 1.67076692e-01 -9.71351802e-01 6.26964495e-02
-1.06652930e-01 -6.81770384e-01 2.88320091e-02 -9.19018686e-01
-1.14515400e+00 8.64320159e-01 1.39353669e+00 3.99126530e-01
1.46080911e+00 -3.92113775e-01 3.89476866e-01 3.90717775e-01
3.13500494e-01 -1.40582550e+00 8.94595906e-02 2.66671687e-01
1.04622865e+00 -1.17373300e+00 1.12162732e-01 -5.20327449e-01
-3.33094090e-01 1.53407705e+00 5.75218379e-01 -2.02347506e-02
5.67753732e-01 2.46222228e-01 3.49650353e-01 -2.07856759e-01
2.96443626e-02 -1.68605044e-01 -3.66195366e-02 6.66166902e-01
3.84338051e-01 -3.33318524e-02 -6.90324366e-01 -2.64012396e-01
1.62020683e-01 1.83569789e-01 6.12554848e-01 1.27786887e+00
-6.24329031e-01 -1.14329207e+00 -9.28364873e-01 -1.29022270e-01
-3.68605912e-01 3.47566813e-01 -4.53894526e-01 8.52089226e-01
9.09443498e-01 1.19480348e+00 1.00810476e-01 -2.56607473e-01
4.17766362e-01 -2.59546191e-01 4.12418246e-01 -4.70960081e-01
-5.63426375e-01 4.73670401e-02 -1.71084225e-01 -4.60468382e-01
-7.92086840e-01 -4.20244753e-01 -1.18229818e+00 1.32658542e-03
-3.51771981e-01 -1.03900827e-01 1.26118338e+00 6.83292329e-01
1.87838916e-02 9.99837935e-01 6.35977149e-01 -7.41858482e-01
-1.33536890e-01 -7.98169732e-01 -5.89333534e-01 7.68964171e-01
1.93696678e-01 -6.78776979e-01 -2.43211672e-01 4.82717067e-01] | [9.764660835266113, -1.9996412992477417] |
158fe6fa-25a1-4039-912a-e0cd550890ff | searching-for-robust-neural-architectures-via | 2203.03128 | null | https://arxiv.org/abs/2203.03128v2 | https://arxiv.org/pdf/2203.03128v2.pdf | $A^{3}D$: A Platform of Searching for Robust Neural Architectures and Efficient Adversarial Attacks | The robustness of deep neural networks (DNN) models has attracted increasing attention due to the urgent need for security in many applications. Numerous existing open-sourced tools or platforms are developed to evaluate the robustness of DNN models by ensembling the majority of adversarial attack or defense algorithms. Unfortunately, current platforms do not possess the ability to optimize the architectures of DNN models or the configuration of adversarial attacks to further enhance the robustness of models or the performance of adversarial attacks. To alleviate these problems, in this paper, we first propose a novel platform called auto adversarial attack and defense ($A^{3}D$), which can help search for robust neural network architectures and efficient adversarial attacks. In $A^{3}D$, we employ multiple neural architecture search methods, which consider different robustness evaluation metrics, including four types of noises: adversarial noise, natural noise, system noise, and quantified metrics, resulting in finding robust architectures. Besides, we propose a mathematical model for auto adversarial attack, and provide multiple optimization algorithms to search for efficient adversarial attacks. In addition, we combine auto adversarial attack and defense together to form a unified framework. Among auto adversarial defense, the searched efficient attack can be used as the new robustness evaluation to further enhance the robustness. In auto adversarial attack, the searched robust architectures can be utilized as the threat model to help find stronger adversarial attacks. Experiments on CIFAR10, CIFAR100, and ImageNet datasets demonstrate the feasibility and effectiveness of the proposed platform, which can also provide a benchmark and toolkit for researchers in the application of automated machine learning in evaluating and improving the DNN model robustnesses. | ['Xiaoqian Chen', 'Wen Yao', 'Chao Li', 'Tingsong Jiang', 'Jialiang Sun'] | 2022-03-07 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [-1.94794476e-01 -4.84353125e-01 5.10194123e-01 -1.02781370e-01
-3.23849201e-01 -1.10662293e+00 4.56608534e-01 -4.17390287e-01
-3.36404353e-01 3.83141071e-01 -1.31381437e-01 -4.61490363e-01
-3.30609143e-01 -9.22181487e-01 -6.18538380e-01 -9.22696650e-01
1.74939111e-02 -5.44091836e-02 2.12105885e-01 -5.70421219e-01
1.50557876e-01 9.02240872e-01 -1.15381026e+00 -1.89501420e-01
7.16119468e-01 1.13938189e+00 -8.43833387e-02 5.26230097e-01
4.26016785e-02 4.55656350e-01 -1.06039834e+00 -5.39231241e-01
6.85804248e-01 -2.30721369e-01 -4.34136480e-01 -6.68640912e-01
1.15933254e-01 -3.99403840e-01 -7.22148299e-01 1.63399494e+00
8.38444054e-01 7.88318217e-02 2.89781988e-01 -1.50624287e+00
-5.50508440e-01 7.80754268e-01 -1.35642029e-02 3.67329776e-01
-1.00305900e-01 5.34339488e-01 5.36022067e-01 -3.35139632e-01
5.01735471e-02 1.50937879e+00 3.70773196e-01 8.52355063e-01
-6.99868619e-01 -1.27707887e+00 4.62448865e-01 1.55180544e-01
-1.24427855e+00 -3.47437501e-01 1.00177121e+00 -2.10192755e-01
6.51319504e-01 3.91136229e-01 2.47708708e-01 1.46525800e+00
2.28723869e-01 4.79929447e-01 7.61813939e-01 4.31173220e-02
2.72570163e-01 -2.58201733e-02 1.07932419e-01 5.48872292e-01
3.07241440e-01 5.14224052e-01 2.02717066e-01 -2.53981739e-01
7.68774688e-01 2.08771061e-02 -2.59838969e-01 2.38368616e-01
-6.53204620e-01 6.76434457e-01 7.66116202e-01 1.56541824e-01
-1.66093949e-02 4.87266630e-02 6.97209597e-01 4.16857034e-01
-1.85994670e-01 7.07241476e-01 -5.18632650e-01 2.27678791e-01
-1.50625139e-01 2.69976944e-01 4.73436534e-01 7.70618856e-01
5.14905989e-01 7.58094370e-01 8.94618258e-02 7.85424232e-01
5.06476462e-01 7.04237342e-01 4.64445233e-01 -7.29591012e-01
5.64763129e-01 7.54623473e-01 -3.64391774e-01 -1.26103628e+00
-4.37957108e-01 -5.03356516e-01 -1.09536946e+00 4.99068946e-01
-7.28482828e-02 -5.21557510e-01 -8.40052128e-01 1.86134875e+00
2.95688540e-01 4.28136855e-01 5.13503611e-01 7.17178941e-01
9.42559361e-01 6.21048748e-01 -9.94498953e-02 -7.18466118e-02
1.10580075e+00 -7.53336847e-01 -5.42834640e-01 -1.91776335e-01
2.05097571e-01 -7.60429442e-01 8.50657284e-01 4.10410166e-01
-8.28721404e-01 -6.85557902e-01 -1.30618751e+00 5.25845706e-01
-4.71181095e-01 -2.95074075e-01 2.77552754e-01 1.01289952e+00
-7.87291884e-01 4.72692609e-01 -7.18487203e-01 2.02469721e-01
4.01832521e-01 5.01775205e-01 -2.32413739e-01 6.25569150e-02
-1.66723955e+00 8.95704806e-01 6.27355933e-01 3.77746105e-01
-1.48236382e+00 -6.34158552e-01 -6.97936416e-01 -1.02043040e-01
1.02559336e-01 -3.54078412e-01 7.48934805e-01 -8.62227976e-01
-1.50284529e+00 2.36683324e-01 8.50464225e-01 -4.53108132e-01
3.31067055e-01 -1.74888849e-01 -9.74634707e-01 -2.01033130e-02
-4.62338209e-01 1.51942909e-01 7.07659900e-01 -1.15619445e+00
-1.75489470e-01 -3.01021338e-01 4.66282755e-01 -9.50724632e-02
-7.07312047e-01 4.44514245e-01 -2.75357604e-01 -8.99805546e-01
-7.62321725e-02 -9.41417813e-01 -4.50617224e-01 -2.45645508e-01
-5.16547740e-01 1.76448554e-01 1.08534825e+00 -3.77168387e-01
1.27987909e+00 -2.27340937e+00 1.17771581e-01 5.87821543e-01
1.09929167e-01 9.79257286e-01 -5.09712040e-01 7.91262016e-02
-2.06979468e-01 4.33240920e-01 -1.66644856e-01 2.12344781e-01
4.28962111e-02 3.47178280e-01 -5.60198307e-01 2.16711223e-01
2.77457863e-01 6.42867804e-01 -5.23066401e-01 -1.79676563e-01
2.04518259e-01 4.80144799e-01 -5.06842792e-01 3.47787708e-01
-6.18356885e-03 2.22441912e-01 -8.91221285e-01 6.83966339e-01
9.13585246e-01 4.39051598e-01 -1.83656856e-01 -4.13996577e-01
2.21561402e-01 -3.71039599e-01 -1.43899024e+00 9.15502250e-01
-3.80795568e-01 2.81137615e-01 2.15173975e-01 -8.92080426e-01
1.28055847e+00 2.55028754e-01 1.91237316e-01 -5.24823308e-01
6.04034126e-01 1.37738913e-01 3.54022235e-01 -3.79522204e-01
-1.92702040e-01 2.01204479e-01 -2.05417126e-01 1.56515330e-01
7.99704567e-02 6.95165619e-02 -2.23866060e-01 2.86583472e-02
1.36789179e+00 -3.66045982e-01 -1.00580111e-01 -2.13961810e-01
9.63105857e-01 -5.02094686e-01 7.12292373e-01 6.34155691e-01
-3.79333109e-01 3.46313506e-01 4.08509284e-01 -6.34239554e-01
-7.74520099e-01 -1.26134360e+00 -1.25202015e-01 8.07123363e-01
1.65620521e-01 -1.19802222e-01 -8.53161156e-01 -9.62409973e-01
-3.64483207e-01 4.07890201e-01 -4.49316561e-01 -7.32676327e-01
-7.09205389e-01 -8.56264532e-01 1.38640285e+00 5.42592764e-01
1.00395286e+00 -1.14014995e+00 -1.18446007e-01 3.57511938e-02
-2.84547340e-02 -9.28183138e-01 -2.49846965e-01 1.01783127e-01
-6.44863963e-01 -1.13562214e+00 -2.38299519e-01 -6.76478684e-01
5.36520898e-01 -3.40204500e-02 7.39649475e-01 4.23002452e-01
-1.75250798e-01 2.19024215e-02 -3.52817953e-01 -5.57898343e-01
-6.59145176e-01 -3.43653411e-01 6.59833729e-01 -1.51293457e-01
-2.19867468e-01 -6.63426995e-01 -4.05421227e-01 6.48215532e-01
-1.25379395e+00 -8.18990707e-01 5.46605229e-01 6.92676961e-01
4.10747737e-01 4.92888331e-01 7.95495212e-01 -4.42032069e-01
9.26098287e-01 -5.34280241e-01 -9.36470747e-01 3.94948930e-01
-5.02195895e-01 1.70385689e-01 1.34768820e+00 -7.64683485e-01
-7.49861896e-01 -2.16688514e-01 -7.27937400e-01 -1.00267732e+00
-1.48118258e-01 4.25543219e-01 -9.52035367e-01 -5.53153634e-01
9.10648286e-01 9.46520492e-02 -1.00091591e-01 -4.69199449e-01
2.12493718e-01 2.22849101e-01 6.06376886e-01 -6.78391278e-01
1.29762578e+00 -7.11617842e-02 1.70861110e-01 -3.02243650e-01
-4.41480994e-01 2.41525292e-01 -1.04236081e-01 -1.81077629e-01
6.77033246e-01 -5.82554340e-01 -8.48243296e-01 8.07889163e-01
-1.08175635e+00 -1.54463127e-02 2.23101601e-01 2.77056605e-01
-1.27573431e-01 4.95445490e-01 -5.27669013e-01 -6.79737568e-01
-6.58717275e-01 -1.68921041e+00 3.41730952e-01 4.81508344e-01
4.86321718e-01 -8.65654349e-01 -4.60862704e-02 7.50836506e-02
5.80679238e-01 4.94710684e-01 9.70430672e-01 -1.08083844e+00
-4.94429290e-01 -4.07307625e-01 2.25649215e-03 9.54111874e-01
-7.84411281e-02 3.80331337e-01 -8.93940985e-01 -2.73699045e-01
2.19622239e-01 -3.10613334e-01 4.94214952e-01 9.25119668e-02
1.40618980e+00 -7.58607805e-01 -1.01806529e-01 9.47971702e-01
1.37849820e+00 7.02519834e-01 9.58192229e-01 5.73222697e-01
6.19029880e-01 2.35894740e-01 3.49102199e-01 2.59323001e-01
-2.99860686e-01 4.40430045e-01 1.04191971e+00 3.09108030e-02
4.74665791e-01 3.23022567e-02 5.99513888e-01 7.07002580e-01
-4.67547327e-02 -3.74374360e-01 -8.91175389e-01 -4.13815975e-02
-1.52109551e+00 -1.00072742e+00 1.78475678e-01 1.91969228e+00
5.75107574e-01 4.35992539e-01 -1.31588995e-01 3.22498858e-01
6.40829086e-01 9.62235034e-02 -7.29808509e-01 -4.43760991e-01
-2.98451066e-01 2.19693467e-01 3.24242026e-01 2.32654735e-01
-1.18891275e+00 8.58463049e-01 6.06564331e+00 1.00085127e+00
-1.17068565e+00 -1.85764611e-01 5.33069253e-01 2.95538753e-02
-2.77478933e-01 -2.16611728e-01 -8.31713438e-01 5.65029860e-01
7.95628846e-01 -9.94267166e-02 6.16625726e-01 1.13684952e+00
-8.96816701e-03 8.26050341e-01 -8.85956943e-01 8.73723328e-01
-1.14829592e-01 -1.21492457e+00 3.32334012e-01 -1.15074113e-01
6.62258744e-01 1.74919888e-02 3.09929222e-01 4.43645060e-01
6.31385505e-01 -1.01843762e+00 4.12395328e-01 4.52748030e-01
3.52150917e-01 -1.08322275e+00 1.02389324e+00 2.42068052e-01
-1.21350741e+00 -3.92250896e-01 -5.20384789e-01 2.62766123e-01
-2.21612871e-01 3.88799645e-02 -2.16153517e-01 7.29177296e-01
8.87760282e-01 1.05226070e-01 -6.60882115e-01 8.62757921e-01
-2.70361751e-01 5.19208610e-01 -2.89536357e-01 -6.86871037e-02
2.88654655e-01 -7.84243271e-02 6.45833015e-01 8.25091720e-01
2.11536780e-01 1.80442005e-01 1.77907914e-01 7.57080138e-01
-2.03521997e-01 -5.83572388e-02 -5.79190373e-01 3.61755453e-02
8.52874696e-01 1.28378594e+00 -4.03128743e-01 8.72434154e-02
-4.43275608e-02 4.97844368e-01 6.50765076e-02 4.04158652e-01
-1.24653041e+00 -7.19610512e-01 1.06479490e+00 -5.43135703e-01
-1.99677765e-01 -9.16937590e-02 -1.64206058e-01 -9.54133213e-01
-9.80434567e-02 -1.49709129e+00 4.94660169e-01 -4.69150454e-01
-1.50943625e+00 1.08964753e+00 -2.60824841e-02 -1.28879881e+00
1.38763577e-01 -8.49335253e-01 -1.14661980e+00 9.17337418e-01
-1.04962099e+00 -9.76576090e-01 -2.40434289e-01 1.08857989e+00
2.74587393e-01 -8.90973747e-01 7.32691526e-01 4.49929148e-01
-1.18818378e+00 1.14230895e+00 2.45666616e-02 4.57941204e-01
4.42365319e-01 -5.86450577e-01 4.63006675e-01 1.39581943e+00
-7.98252448e-02 9.06116724e-01 6.95792317e-01 -3.45564097e-01
-1.37380719e+00 -1.43603837e+00 -1.13810070e-01 -2.52020270e-01
7.70466864e-01 -1.94715723e-01 -8.91371310e-01 4.64791864e-01
-6.32207692e-02 1.66944146e-01 6.22825444e-01 -2.67035395e-01
-5.71569741e-01 -5.99264979e-01 -1.39069390e+00 9.44783211e-01
8.73788536e-01 -3.30477059e-01 -3.46953899e-01 2.06608102e-01
1.23444808e+00 -2.78891146e-01 -8.25024068e-01 7.23199368e-01
3.19929332e-01 -7.88522542e-01 1.38835216e+00 -7.28131771e-01
3.25261146e-01 -3.55347484e-01 -3.71100247e-01 -1.21043408e+00
-2.84906328e-01 -5.91737330e-01 4.82502487e-03 1.52573657e+00
4.09375906e-01 -9.88707781e-01 4.36812609e-01 6.42814159e-01
-4.35343206e-01 -9.36609387e-01 -9.66659307e-01 -9.46735799e-01
1.70888633e-01 -5.27386487e-01 1.07858646e+00 8.38629782e-01
-6.53454185e-01 -2.34943315e-01 -1.94809139e-01 7.26812720e-01
5.00402153e-01 -6.38954520e-01 7.28976846e-01 -9.59360719e-01
-1.76525936e-01 -8.76472473e-01 -7.14409888e-01 -5.44238806e-01
3.43004555e-01 -5.80893815e-01 -1.08599350e-01 -9.76994574e-01
-3.37754607e-01 -6.66195512e-01 -8.84903073e-01 6.58554375e-01
-2.52532870e-01 3.60327475e-02 2.08983079e-01 7.71946162e-02
-2.09021986e-01 5.78268826e-01 9.21350241e-01 -2.69589812e-01
-8.58137663e-03 2.07290426e-01 -8.81453156e-01 9.08705831e-01
9.73207712e-01 -3.88934165e-01 -6.19519532e-01 -6.13459885e-01
9.50074941e-02 -1.94845617e-01 5.06437302e-01 -1.27992666e+00
1.68623358e-01 -3.51359904e-01 3.12512398e-01 -2.67469227e-01
6.26444593e-02 -1.07270133e+00 2.24307030e-01 5.55101573e-01
-1.56838298e-01 4.45705235e-01 5.27479887e-01 4.35912371e-01
-1.72013968e-01 -2.76579887e-01 1.01811123e+00 3.81600000e-02
-6.34006262e-01 6.80272877e-01 5.94782783e-03 2.67519653e-02
1.08572125e+00 -5.76272840e-03 -7.51150489e-01 3.87389865e-03
-4.69202787e-01 4.43930477e-01 3.51659432e-02 6.99038744e-01
8.81333888e-01 -1.48775113e+00 -5.67935169e-01 4.66988891e-01
-2.33963951e-01 -1.10701792e-01 5.44911981e-01 1.92863882e-01
-6.74677074e-01 -1.36210136e-02 -6.71881080e-01 -1.60685584e-01
-1.27390206e+00 9.72962677e-01 7.42624104e-01 -2.21998259e-01
-5.28572164e-02 1.04079747e+00 3.33239049e-01 -5.22670984e-01
4.50563133e-01 -4.87635545e-02 -3.10332417e-01 -4.79253739e-01
5.31760633e-01 4.16583002e-01 7.64031187e-02 -3.91592741e-01
-5.43781400e-01 3.88278246e-01 -5.10022640e-02 1.56669900e-01
1.06239986e+00 1.70440912e-01 -1.71705380e-01 -2.35576615e-01
1.10588384e+00 -1.44593880e-01 -1.03753781e+00 -1.59583949e-02
-4.69011635e-01 -1.57823280e-01 -3.29757817e-02 -8.67384911e-01
-1.49542594e+00 8.33039105e-01 8.86114597e-01 2.74957210e-01
1.65127325e+00 -4.37754899e-01 6.95380807e-01 6.25614166e-01
2.27616206e-01 -7.65637457e-01 2.88316727e-01 6.53615594e-01
1.10133016e+00 -9.39110875e-01 -2.63121516e-01 1.13394521e-02
-4.16765183e-01 1.13523340e+00 1.11885917e+00 -3.80834997e-01
8.43700707e-01 5.32592356e-01 3.46477270e-01 1.01367280e-01
-4.28133965e-01 2.81494915e-01 2.35315442e-01 6.95926547e-01
-2.43146628e-01 -2.30848745e-01 -2.96244379e-02 1.09220791e+00
-3.35646510e-01 -7.62124300e-01 1.57463208e-01 5.20376205e-01
-3.71033549e-01 -1.18091011e+00 -5.14262021e-01 1.29034549e-01
-5.01098812e-01 -1.32170543e-01 -3.51256877e-01 6.07681394e-01
4.11614329e-01 1.13645256e+00 -4.78405535e-01 -1.12171304e+00
5.57316303e-01 -2.39570260e-01 -2.25362182e-03 -1.12540267e-01
-9.77719963e-01 -2.76567101e-01 -2.48917654e-01 -5.68259716e-01
8.10407773e-02 -1.49430007e-01 -9.95489180e-01 -4.44465518e-01
-4.08804268e-01 5.16231917e-02 6.32565320e-01 8.92116547e-01
2.29154974e-01 6.71170592e-01 1.15385807e+00 -5.37859499e-01
-9.18808401e-01 -6.72834098e-01 -1.23936713e-01 2.35663772e-01
1.49030700e-01 -5.86842835e-01 -5.30331075e-01 -4.39567447e-01] | [5.524845600128174, 7.941572666168213] |
9e6af9df-5a05-4bff-b8db-e241c8ca4d2b | shiro-soft-hierarchical-reinforcement | 2212.12786 | null | https://arxiv.org/abs/2212.12786v1 | https://arxiv.org/pdf/2212.12786v1.pdf | SHIRO: Soft Hierarchical Reinforcement Learning | Hierarchical Reinforcement Learning (HRL) algorithms have been demonstrated to perform well on high-dimensional decision making and robotic control tasks. However, because they solely optimize for rewards, the agent tends to search the same space redundantly. This problem reduces the speed of learning and achieved reward. In this work, we present an Off-Policy HRL algorithm that maximizes entropy for efficient exploration. The algorithm learns a temporally abstracted low-level policy and is able to explore broadly through the addition of entropy to the high-level. The novelty of this work is the theoretical motivation of adding entropy to the RL objective in the HRL setting. We empirically show that the entropy can be added to both levels if the Kullback-Leibler (KL) divergence between consecutive updates of the low-level policy is sufficiently small. We performed an ablative study to analyze the effects of entropy on hierarchy, in which adding entropy to high-level emerged as the most desirable configuration. Furthermore, a higher temperature in the low-level leads to Q-value overestimation and increases the stochasticity of the environment that the high-level operates on, making learning more challenging. Our method, SHIRO, surpasses state-of-the-art performance on a range of simulated robotic control benchmark tasks and requires minimal tuning. | ['Omer Eldar', 'Mathew Strong', 'Kandai Watanabe'] | 2022-12-24 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 1.28511023e-02 2.81596035e-01 -4.26865041e-01 9.32790190e-02
-4.91720796e-01 -5.22480190e-01 6.39154613e-01 2.92670727e-01
-9.32133794e-01 1.20525956e+00 -6.88758492e-02 -2.30172291e-01
-3.45793456e-01 -7.09451199e-01 -7.42438138e-01 -1.07858610e+00
-5.01922190e-01 3.46539348e-01 2.16358975e-01 -3.37277293e-01
3.18949342e-01 3.47361565e-01 -1.84396183e+00 -1.56930000e-01
7.89436936e-01 9.80644643e-01 4.99273896e-01 4.92725164e-01
3.26373667e-01 6.15486503e-01 -5.88094831e-01 2.88067877e-01
5.25449634e-01 -4.95625377e-01 -8.05916250e-01 -3.97663862e-01
-1.07602634e-01 -3.40403706e-01 -2.75219721e-03 8.18068743e-01
5.62124312e-01 4.09736544e-01 4.80423957e-01 -9.64296103e-01
-1.02176622e-01 7.85992324e-01 -3.93786937e-01 1.01946769e-02
1.31744212e-02 5.74567199e-01 9.29385483e-01 -2.60068059e-01
7.32254565e-01 1.29604840e+00 2.46165961e-01 6.19270802e-01
-1.41357636e+00 -5.51225305e-01 2.77676195e-01 -3.51069160e-02
-1.02999651e+00 -8.82358178e-02 5.39907932e-01 -1.91577598e-01
1.24823284e+00 -1.30896971e-01 9.88839626e-01 8.56422603e-01
6.66263998e-01 5.12542903e-01 1.61593843e+00 -4.38008368e-01
1.02908862e+00 -4.48776856e-02 -5.15897155e-01 5.88149607e-01
3.68063509e-01 9.49721992e-01 -5.61145723e-01 -4.86314250e-03
6.59978807e-01 -4.52950716e-01 8.28001425e-02 -8.55400741e-01
-9.09484744e-01 9.18273509e-01 5.55159867e-01 3.27730775e-01
-5.89842439e-01 4.59931254e-01 2.05974624e-01 4.03880626e-01
-1.65509939e-01 1.27411020e+00 -4.61603701e-01 -5.42554975e-01
-6.44168973e-01 4.53095168e-01 7.09486425e-01 3.81165475e-01
6.66259229e-01 3.51115838e-02 -3.57386023e-01 4.97684389e-01
7.01327100e-02 1.23886079e-01 5.25848627e-01 -1.45680344e+00
2.74407446e-01 3.55781972e-01 3.77519280e-01 -3.03952634e-01
-5.78261256e-01 -7.93702781e-01 -3.03480864e-01 9.88335073e-01
4.96994883e-01 -5.40073872e-01 -6.64002061e-01 2.02743387e+00
2.92239875e-01 -5.65915406e-01 3.08496892e-01 1.00271261e+00
-2.05900863e-01 5.39043963e-01 1.78073689e-01 -5.27488708e-01
8.44562948e-01 -8.09375942e-01 -6.71499372e-01 -3.00164312e-01
6.49232864e-01 -2.62573808e-01 1.30760479e+00 4.88126695e-01
-1.22926605e+00 -2.78553903e-01 -1.32708514e+00 5.01438022e-01
-2.62538612e-01 -3.21493149e-01 6.21272147e-01 3.77301753e-01
-9.92951810e-01 1.26380670e+00 -8.49812150e-01 -2.76867479e-01
1.91877231e-01 5.24594605e-01 1.11445218e-01 4.08293933e-01
-1.34743893e+00 1.33360934e+00 8.85989606e-01 -3.22122395e-01
-1.03087437e+00 -4.07316446e-01 -4.99728054e-01 4.05209549e-02
7.52426505e-01 -3.46572757e-01 1.52823079e+00 -6.51077449e-01
-2.06012821e+00 2.46577442e-01 3.85866344e-01 -7.41999030e-01
5.54317653e-01 -1.39673561e-01 2.83238888e-01 -8.63476992e-02
-1.59384251e-01 1.13194251e+00 8.52686226e-01 -1.23414254e+00
-4.74402308e-01 -3.02290857e-01 2.03569502e-01 6.17729962e-01
-2.64705837e-01 -7.28203535e-01 1.61036640e-01 -3.65479618e-01
-1.37327567e-01 -1.25484920e+00 -5.20775735e-01 -4.53309745e-01
1.88928440e-01 -2.36050755e-01 5.70286453e-01 -4.71267812e-02
1.21430838e+00 -1.91594303e+00 4.15824294e-01 1.38784423e-01
-1.90196365e-01 -3.45718823e-02 -8.43228400e-02 6.24956727e-01
2.86683202e-01 8.18752274e-02 -2.03849018e-01 6.53578788e-02
1.34451613e-01 4.36049104e-01 -2.27193207e-01 3.67254049e-01
3.96593846e-02 7.76387274e-01 -1.13148332e+00 -3.45068753e-01
1.30272120e-01 3.54331620e-02 -8.59912217e-01 2.22360026e-02
-4.78165239e-01 5.34484267e-01 -3.84891868e-01 2.78427929e-01
9.28833857e-02 -1.20689580e-02 2.86421269e-01 3.38893205e-01
-5.20076752e-01 3.45074117e-01 -9.25287008e-01 1.64774966e+00
-4.70839083e-01 2.70136505e-01 -4.91554104e-02 -6.28774107e-01
8.36253464e-01 -8.76054242e-02 5.98861098e-01 -1.09654272e+00
1.50691837e-01 2.16043845e-01 3.83398950e-01 -1.67424098e-01
5.22695243e-01 -2.66347617e-01 -3.01260829e-01 3.15305114e-01
-1.11365706e-01 -7.03954756e-01 3.64885867e-01 -1.82003736e-01
1.07850516e+00 5.44958174e-01 5.24594307e-01 -5.68793595e-01
2.01100364e-01 -6.35856912e-02 4.84034836e-01 1.04270816e+00
-3.33655179e-01 -2.95347452e-01 7.99204528e-01 -1.92105800e-01
-1.09357095e+00 -1.00433159e+00 -8.91280994e-02 1.20132208e+00
2.96143502e-01 -2.81427562e-01 -7.42103934e-01 -4.99753654e-01
2.74433732e-01 1.07251954e+00 -6.69006824e-01 -4.60139275e-01
-5.72104454e-01 -6.50695801e-01 2.65450567e-01 2.70820886e-01
4.86940056e-01 -1.29452336e+00 -1.57259810e+00 3.13900113e-01
3.19759965e-01 -7.39296317e-01 -1.41492516e-01 8.40837359e-01
-1.06488538e+00 -6.23173118e-01 -3.44568759e-01 -2.94923782e-01
3.35188091e-01 -4.87504840e-01 9.00569081e-01 -3.52242351e-01
-3.47588748e-01 2.45446905e-01 -1.80552617e-01 -3.75731826e-01
-3.97412479e-01 2.66463131e-01 2.85378098e-01 -7.27345884e-01
-5.75696379e-02 -5.30947804e-01 -6.20991230e-01 2.44295657e-01
-7.17009008e-01 -1.88631192e-01 8.23775709e-01 9.94026005e-01
5.57422698e-01 2.99675077e-01 7.04501569e-01 -1.89937398e-01
9.73729491e-01 -9.90619808e-02 -1.07134569e+00 -4.94957753e-02
-1.11187065e+00 1.01351368e+00 5.83522737e-01 -6.13139451e-01
-9.48642015e-01 2.18966290e-01 2.39929229e-01 -2.28464574e-01
1.83572248e-01 2.74286836e-01 3.50162357e-01 -2.57013645e-02
7.32927918e-01 7.59658068e-02 1.84080526e-01 -1.28237933e-01
3.64060014e-01 2.25364044e-01 1.60589106e-02 -8.92799318e-01
4.38453883e-01 6.43617362e-02 4.82137412e-01 -5.81735849e-01
-5.58669508e-01 9.96189564e-02 -3.66338789e-01 -3.96086782e-01
6.96693718e-01 -4.34288055e-01 -1.17657447e+00 1.43187165e-01
-5.90623260e-01 -9.31728601e-01 -8.70406687e-01 5.23746908e-01
-1.16837692e+00 -6.86553493e-02 -3.71285379e-01 -1.18619084e+00
-3.95302698e-02 -1.38350224e+00 7.56151915e-01 3.51232916e-01
-2.54160553e-01 -5.66562295e-01 1.74554393e-01 -2.20724821e-01
5.26228487e-01 2.84055352e-01 9.27533209e-01 -1.20571919e-01
-5.30727386e-01 3.51709068e-01 3.76420498e-01 9.17221904e-02
-1.64898336e-01 -3.88856590e-01 -6.92479193e-01 -6.74495578e-01
-1.50441658e-02 -8.24274063e-01 9.81584787e-01 4.61400092e-01
1.05503380e+00 -4.09887582e-01 -7.91200846e-02 2.11358652e-01
1.45715332e+00 5.45076430e-01 3.96013021e-01 7.94406891e-01
-1.20101288e-01 7.50678360e-01 9.24445093e-01 7.97512591e-01
-9.09147039e-02 7.50034094e-01 7.34268069e-01 2.97643602e-01
2.42815763e-01 -3.82062435e-01 6.97462142e-01 2.97822297e-01
4.34600934e-02 6.05924986e-02 -6.81370258e-01 2.24571213e-01
-1.91404927e+00 -7.96605468e-01 7.45302975e-01 2.57435441e+00
1.08297408e+00 5.13252914e-01 2.99638987e-01 -1.19992092e-01
2.97746956e-01 1.48427412e-01 -1.14835167e+00 -7.03260839e-01
9.64226797e-02 1.54050127e-01 9.14524674e-01 6.60747826e-01
-8.09360385e-01 1.02155113e+00 6.76316166e+00 7.44546652e-01
-1.12788332e+00 -1.53981477e-01 3.80888581e-01 -6.66101754e-01
8.77603237e-03 1.44900478e-04 -6.58719003e-01 3.64881098e-01
9.09938395e-01 -1.61090642e-01 9.55824733e-01 9.26206231e-01
2.70843089e-01 -5.81374466e-01 -1.14621949e+00 5.93011558e-01
-6.37501478e-01 -1.06971359e+00 -3.39207709e-01 3.89485508e-01
7.51553595e-01 1.21443085e-01 3.12214971e-01 6.30012631e-01
5.69637299e-01 -1.00589252e+00 8.18631351e-01 3.20176691e-01
5.79187214e-01 -1.06242561e+00 5.63910186e-01 5.38883865e-01
-8.44069421e-01 -6.45735502e-01 -1.49909794e-01 -3.20098311e-01
-1.25965372e-01 1.72157809e-01 -9.30949688e-01 -1.04223024e-02
6.12031400e-01 6.80037662e-02 -3.60120118e-01 6.68590903e-01
-3.32055330e-01 1.63740084e-01 -4.89581317e-01 -6.14477336e-01
8.07360351e-01 -2.48624325e-01 6.21751726e-01 7.31777251e-01
1.17627189e-01 -8.59274939e-02 2.99924403e-01 8.75349641e-01
1.49085119e-01 -9.63787064e-02 -6.45053744e-01 -1.98735923e-01
4.22227353e-01 8.46200824e-01 -5.54999173e-01 1.54166743e-02
3.31360549e-01 6.07426167e-01 5.32848179e-01 2.46942908e-01
-6.85795128e-01 -2.96828747e-01 5.11557519e-01 -7.56536573e-02
3.42257887e-01 -4.29131001e-01 -2.99103200e-01 -6.48573756e-01
-7.38458782e-02 -6.44115388e-01 2.84060150e-01 -2.48453066e-01
-7.50056744e-01 3.12980622e-01 2.24879161e-01 -9.57343578e-01
-7.67487109e-01 -2.89856821e-01 -6.23979531e-02 5.83962083e-01
-1.46789634e+00 -9.44967195e-02 7.58447796e-02 1.68178514e-01
5.12598038e-01 -9.18841511e-02 6.46041930e-01 -2.95748502e-01
-3.17641854e-01 4.56929356e-01 4.39285100e-01 -6.05596960e-01
4.51054692e-01 -1.43994176e+00 -1.39029294e-01 2.86554664e-01
-4.75780517e-01 3.96907747e-01 1.04947078e+00 -6.69279754e-01
-1.42907083e+00 -7.05736756e-01 1.74427837e-01 -6.24190345e-02
5.89332521e-01 -1.96138963e-01 -5.03233612e-01 9.50111672e-02
2.27246568e-01 -3.92631978e-01 1.58756003e-01 6.87859505e-02
1.45848289e-01 -1.09687515e-01 -1.32234037e+00 9.31692600e-01
8.46288025e-01 -1.99460387e-01 -5.26470065e-01 2.47664973e-02
7.98022449e-01 -3.75017285e-01 -9.92632151e-01 6.46764934e-01
7.47865856e-01 -1.05853164e+00 7.27487981e-01 -3.95714283e-01
3.61406058e-01 -1.77117109e-01 -5.26590459e-02 -1.49505806e+00
-2.16534004e-01 -8.50725174e-01 -4.07785773e-01 4.50651050e-01
3.62347692e-01 -6.99059784e-01 7.32745945e-01 2.47304931e-01
1.79333538e-01 -1.34696269e+00 -1.10272002e+00 -1.23531485e+00
4.49054211e-01 4.65143919e-02 2.98705012e-01 4.06078607e-01
4.61562932e-01 2.36339971e-01 -1.94358796e-01 -1.77371874e-01
7.36997902e-01 2.98705846e-02 2.71704257e-01 -9.40698087e-01
-6.02410257e-01 -8.44027340e-01 1.19502567e-01 -8.61785352e-01
6.35089502e-02 -5.56469977e-01 4.09110487e-01 -1.26082623e+00
-1.65514112e-01 -6.26717687e-01 -3.77522141e-01 3.38370055e-01
1.05239086e-01 -4.33141857e-01 5.22322834e-01 1.90315261e-01
-5.77421784e-01 9.53568578e-01 1.36921859e+00 6.29503373e-03
-8.47846806e-01 -2.19042882e-01 -4.07620460e-01 5.12335181e-01
1.17953336e+00 -4.20577884e-01 -5.93803942e-01 8.99031311e-02
3.24031532e-01 1.99198201e-01 -1.13709323e-01 -1.13390422e+00
7.14527667e-02 -3.90611351e-01 2.97870636e-01 -3.73844504e-01
5.42991698e-01 -6.09225333e-01 -1.78372070e-01 1.08906913e+00
-9.30546820e-01 7.56243914e-02 2.77027249e-01 6.19358659e-01
2.68456012e-01 -3.92902166e-01 1.09334302e+00 -3.10174286e-01
-5.01583278e-01 -1.24840945e-01 -5.96567869e-01 5.57833649e-02
1.28327322e+00 1.61825530e-02 -6.52943328e-02 -2.44862571e-01
-4.69516546e-01 4.75377947e-01 7.23115087e-01 4.46166009e-01
3.43581587e-01 -1.01185560e+00 -3.12442482e-01 -6.47742748e-02
-1.60489365e-01 -1.63062289e-01 -2.34386444e-01 6.22228265e-01
-2.45793745e-01 6.84310853e-01 -5.83579779e-01 -4.84681100e-01
-7.33592808e-01 6.71197176e-01 5.73082328e-01 -7.42293656e-01
-4.68372911e-01 5.15858948e-01 -2.24063769e-01 -3.12597901e-01
4.99936640e-01 -4.08068568e-01 -2.93575246e-02 1.46298975e-01
1.08446948e-01 5.02214074e-01 -7.81656727e-02 -2.33833585e-02
-2.80778825e-01 4.76697981e-01 -5.01781069e-02 -6.86922371e-01
1.09157181e+00 1.46407396e-01 1.76572978e-01 5.34422219e-01
8.55572522e-01 -4.71723258e-01 -1.79709935e+00 2.22981244e-01
1.87301308e-01 -2.54827976e-01 4.25333768e-01 -9.21950877e-01
-4.73094791e-01 6.30461156e-01 8.06791306e-01 1.57348320e-01
9.25722063e-01 -5.72981089e-02 4.07902896e-01 8.26411009e-01
7.79186845e-01 -1.75549376e+00 3.80682886e-01 6.80617213e-01
8.87726426e-01 -9.93510306e-01 1.89482167e-01 3.91296446e-01
-8.50083470e-01 9.78960514e-01 8.25407028e-01 -9.50845629e-02
1.91919595e-01 2.92270958e-01 -4.81638610e-01 2.17015669e-02
-1.16053355e+00 -3.75974029e-01 -2.48302072e-01 3.69708151e-01
7.32430071e-02 1.95600539e-01 -5.68581223e-01 -1.51914075e-01
-4.32988614e-01 -1.09886304e-01 4.52112883e-01 1.35004938e+00
-8.55300248e-01 -1.09841299e+00 -2.68546522e-01 3.64899427e-01
-2.75431365e-01 1.96709976e-01 -2.35976741e-01 8.17698896e-01
1.89663880e-02 7.41638184e-01 -3.13190632e-02 -1.23696752e-01
1.74724504e-01 2.76896749e-02 8.50332499e-01 -2.81331331e-01
-6.12265825e-01 -6.34811968e-02 1.11630475e-02 -9.01789963e-01
-5.52062318e-02 -8.01600575e-01 -1.58902454e+00 1.15376689e-01
-1.54772654e-01 3.67966443e-01 8.32444906e-01 7.86682248e-01
3.28835398e-01 6.41031921e-01 6.59604609e-01 -9.24841106e-01
-1.20877504e+00 -6.57006800e-01 -4.91518259e-01 9.69703421e-02
5.21310389e-01 -1.12175763e+00 -3.88870507e-01 -6.20126128e-01] | [4.166051864624023, 2.0854110717773438] |
fa8591a7-aa66-4201-906d-9cb298d5a612 | boosting-breast-ultrasound-video | 2306.06877 | null | https://arxiv.org/abs/2306.06877v1 | https://arxiv.org/pdf/2306.06877v1.pdf | Boosting Breast Ultrasound Video Classification by the Guidance of Keyframe Feature Centers | Breast ultrasound videos contain richer information than ultrasound images, therefore it is more meaningful to develop video models for this diagnosis task. However, the collection of ultrasound video datasets is much harder. In this paper, we explore the feasibility of enhancing the performance of ultrasound video classification using the static image dataset. To this end, we propose KGA-Net and coherence loss. The KGA-Net adopts both video clips and static images to train the network. The coherence loss uses the feature centers generated by the static images to guide the frame attention in the video model. Our KGA-Net boosts the performance on the public BUSV dataset by a large margin. The visualization results of frame attention prove the explainability of our method. The codes and model weights of our method will be made publicly available. | ['LiWei Wang', 'Dong Wang', 'Yuting Dai', 'Meng Lei', 'Zhao Zhang', 'AnLan Sun'] | 2023-06-12 | null | null | null | null | ['video-classification'] | ['computer-vision'] | [ 2.55921632e-01 2.81734049e-01 -2.78339088e-01 -5.61893642e-01
-6.95341527e-01 -1.54912114e-01 2.73814410e-01 -3.58059794e-01
-1.03924036e-01 3.70778620e-01 3.44070256e-01 -4.22182173e-01
-1.72842279e-01 -4.67052013e-01 -1.04131854e+00 -7.99016178e-01
-4.08519566e-01 -9.37890559e-02 1.49757698e-01 1.28590062e-01
-1.56346619e-01 2.61644814e-02 -1.09381628e+00 6.65607631e-01
6.03779674e-01 1.22518253e+00 4.21507180e-01 7.84892201e-01
3.22490156e-01 1.36595654e+00 -2.30752930e-01 -4.17940736e-01
4.41381708e-02 -4.90278631e-01 -8.74844193e-01 5.86185046e-02
6.20607793e-01 -7.99168289e-01 -9.85524178e-01 9.57181573e-01
7.44023681e-01 8.90421271e-02 3.85272294e-01 -1.22758818e+00
-6.08605325e-01 8.73474717e-01 -4.71126437e-01 7.07368553e-01
1.97286457e-01 1.98083624e-01 8.55555475e-01 -7.65854895e-01
7.43036807e-01 1.24819207e+00 6.90375388e-01 6.27526581e-01
-7.31564939e-01 -7.37327993e-01 1.38577908e-01 7.65666902e-01
-1.30078292e+00 -5.73327124e-01 7.42099941e-01 -2.37581477e-01
7.01339900e-01 2.92962223e-01 8.66116822e-01 1.39442933e+00
2.54195511e-01 1.21115506e+00 4.93028104e-01 -2.76193619e-01
-4.67730798e-02 -1.94838300e-01 2.70256341e-01 1.11607850e+00
5.16721345e-02 -3.30666192e-02 -4.96231794e-01 1.55616850e-01
8.03466678e-01 1.53913528e-01 -5.71080983e-01 -3.18958640e-01
-9.86930490e-01 7.74171889e-01 5.28734922e-01 1.55547038e-01
-3.20818216e-01 5.73320568e-01 4.03385997e-01 1.59871444e-01
4.93920565e-01 2.90413588e-01 -9.81420130e-02 -3.20845902e-01
-7.07162738e-01 -7.53304586e-02 4.28451687e-01 9.64147151e-01
1.36045396e-01 -9.62663218e-02 -5.26706278e-01 7.11614013e-01
1.71840847e-01 3.92862380e-01 3.49874467e-01 -1.17378354e+00
3.74908656e-01 1.47164911e-01 -3.47180963e-01 -1.33097720e+00
-5.79825759e-01 -5.21730244e-01 -8.58510733e-01 -5.07756472e-01
2.26276800e-01 -2.41940185e-01 -1.03602540e+00 1.68077004e+00
1.53003126e-01 9.78646398e-01 -5.34552820e-02 1.10997999e+00
9.87499177e-01 6.65659964e-01 -9.69038457e-02 -2.70101994e-01
1.29131877e+00 -1.23552930e+00 -9.05468404e-01 1.75253853e-01
8.48405421e-01 -4.08996522e-01 7.99767137e-01 1.98449045e-01
-1.21419513e+00 -6.17355108e-01 -7.52812624e-01 7.37523958e-02
4.75691289e-01 1.63706660e-01 7.41063297e-01 3.67722273e-01
-1.07005978e+00 4.27974254e-01 -1.42394555e+00 -3.10678542e-01
9.04911280e-01 3.45763296e-01 -1.77002177e-01 -6.11958265e-01
-1.29496372e+00 4.94444668e-01 1.71144575e-01 2.90901423e-01
-1.02660275e+00 -9.25122917e-01 -9.95768964e-01 3.26244771e-01
3.76709163e-01 -5.77070236e-01 1.33544278e+00 -1.06728864e+00
-1.00961328e+00 5.96553624e-01 -9.79279354e-02 -8.30932319e-01
6.18485153e-01 -1.27743194e-02 -2.32214838e-01 7.62772322e-01
-6.07728139e-02 9.00583982e-01 7.84633934e-01 -8.76503110e-01
-6.95439458e-01 4.26192395e-02 4.57016855e-01 -1.62792295e-01
-4.44101661e-01 -2.49205977e-01 -9.19116676e-01 -7.96369672e-01
-4.33085859e-02 -1.10305846e+00 -1.11036159e-01 -4.28345390e-02
-3.02210182e-01 -1.23041712e-01 8.00291002e-01 -8.08607399e-01
1.39970744e+00 -2.36193156e+00 2.74634156e-02 7.07498565e-02
4.89255816e-01 1.96705297e-01 -3.53820473e-01 -1.72068521e-01
-2.51806170e-01 1.48088560e-01 1.71122268e-01 3.69636714e-02
-4.53730434e-01 5.91261089e-02 3.11845867e-03 4.96073276e-01
3.40222031e-01 1.02402914e+00 -1.05350149e+00 -6.54623210e-01
4.99982201e-02 4.56233531e-01 -1.08054686e+00 2.88325727e-01
1.51536629e-01 3.52433324e-01 -5.36849380e-01 6.70578480e-01
5.07439375e-01 -7.48170197e-01 2.86343694e-01 -8.44477475e-01
4.45713490e-01 -2.23359261e-02 -5.56349695e-01 1.87704670e+00
-2.12802052e-01 1.02180707e+00 -1.67049035e-01 -1.18329430e+00
2.43615240e-01 4.95604753e-01 1.00946355e+00 -7.70330369e-01
2.62438148e-01 -1.50216475e-01 3.89076710e-01 -1.15006781e+00
2.00247720e-01 4.53420728e-02 4.35233921e-01 1.58315390e-01
2.40342841e-01 5.42411149e-01 3.43024254e-01 3.93795192e-01
1.41603577e+00 -1.33945897e-01 -1.29828885e-01 -1.19612180e-01
1.86769113e-01 -2.66940266e-01 6.12417340e-01 9.68329012e-01
-3.50189358e-01 8.92509401e-01 6.36060417e-01 -5.55263042e-01
-8.95329177e-01 -7.16248393e-01 -4.72585738e-01 6.96370721e-01
2.50233382e-01 -4.92832363e-01 -6.88662767e-01 -9.04072046e-01
-1.29771279e-02 4.48815912e-01 -1.02461290e+00 -5.02450049e-01
-6.19485676e-01 -6.81536078e-01 3.39299560e-01 7.79439330e-01
4.38863426e-01 -7.58673012e-01 -3.83538961e-01 1.67555228e-01
-6.62460148e-01 -1.28161418e+00 -6.15286052e-01 -3.96823585e-01
-6.89628243e-01 -1.21860492e+00 -8.63145351e-01 -8.27213049e-01
8.04056287e-01 5.59797168e-01 8.89029443e-01 3.70271951e-01
-4.83001083e-01 6.24490678e-01 -6.97341740e-01 -2.32927695e-01
-3.33287627e-01 1.23628341e-01 -8.96486044e-02 6.85789883e-02
1.19601719e-01 -3.23849380e-01 -8.52287352e-01 2.18437552e-01
-9.46446776e-01 4.65354204e-01 3.65727812e-01 9.99558270e-01
3.14568043e-01 -7.06496760e-02 3.48621994e-01 -9.05128121e-01
9.27157998e-02 -7.03879297e-01 -2.54496783e-01 1.26471311e-01
-1.30357653e-01 -2.60120600e-01 2.74852008e-01 -5.12540579e-01
-7.83211946e-01 -1.37962759e-01 -2.09254771e-01 -1.02565312e+00
4.24722850e-01 7.93482423e-01 2.42069274e-01 -1.28511205e-01
2.41442144e-01 3.66475433e-02 1.25145569e-01 -1.69779271e-01
3.91318984e-02 4.54737633e-01 3.37676585e-01 -1.80136710e-01
4.62554634e-01 5.10751247e-01 -1.14989810e-01 -8.17416549e-01
-9.48037684e-01 -2.63863385e-01 -8.39302391e-02 -5.92711747e-01
9.59964514e-01 -1.05402839e+00 -8.03715587e-01 1.65931582e-01
-9.78282690e-01 -1.17397524e-01 4.82314005e-02 7.80185878e-01
-5.06627262e-01 3.03144366e-01 -8.73308718e-01 -5.40205717e-01
-8.23329315e-02 -1.40208244e+00 8.53472948e-01 2.27573320e-01
3.95279974e-02 -7.87046254e-01 -3.48011136e-01 5.76180995e-01
2.75120884e-01 -3.64915691e-02 7.20726430e-01 -4.54204828e-01
-8.62042069e-01 -9.56820231e-03 -4.70608920e-01 3.33063781e-01
2.23805994e-01 4.70034145e-02 -1.07130420e+00 -6.32130325e-01
-5.31207360e-02 -4.56344277e-01 1.16385651e+00 1.02698243e+00
2.00610042e+00 -3.36863488e-01 -4.17827725e-01 1.04706168e+00
1.19429040e+00 4.08921659e-01 6.89724386e-01 1.85199663e-01
8.94789994e-01 3.91009986e-01 6.28180981e-01 3.75081271e-01
3.34700882e-01 4.95981961e-01 4.98983771e-01 -3.49511743e-01
-3.30368459e-01 -1.44235313e-01 1.62049234e-01 1.15038967e+00
-2.22745627e-01 -4.52880293e-01 -9.97038484e-01 4.96759355e-01
-1.89430130e+00 -1.20661819e+00 2.40376398e-01 1.48416293e+00
6.13183916e-01 3.66472616e-03 -2.22380981e-01 -3.36118579e-01
5.34430563e-01 1.88267648e-01 -4.12707865e-01 2.78463006e-01
5.44953384e-02 -2.57654876e-01 4.45248872e-01 4.20936972e-01
-1.43201232e+00 4.83713835e-01 6.24146271e+00 9.11575794e-01
-1.20945394e+00 1.07871488e-01 1.03620434e+00 -5.16752601e-01
-1.54737383e-01 -3.70765448e-01 -5.60308993e-01 6.93062782e-01
8.85278940e-01 -1.44932628e-01 6.36009946e-02 6.37406051e-01
4.41589117e-01 8.80915225e-02 -1.21104407e+00 1.11791790e+00
3.06671530e-01 -1.79822767e+00 1.04324460e-01 -1.04308724e-01
6.81449533e-01 5.15918620e-02 1.02136455e-01 3.48894894e-01
-5.40093593e-02 -9.40822601e-01 5.49842119e-01 5.42594671e-01
9.65738356e-01 -4.84732687e-01 1.02896440e+00 -1.35208368e-01
-9.42700148e-01 -2.41377369e-01 -3.71897966e-01 3.15296680e-01
1.33290261e-01 3.29542696e-01 -8.23938012e-01 5.18364966e-01
8.68300378e-01 1.22387516e+00 -7.13875055e-01 1.35195959e+00
3.51916522e-01 1.08385694e+00 5.24012651e-03 2.20206857e-01
4.02721763e-01 1.42730683e-01 3.52914780e-01 1.22126043e+00
4.89450693e-01 4.08448219e-01 1.89919189e-01 4.38699752e-01
-3.45182955e-01 -1.30684868e-01 -6.10371590e-01 -7.72045776e-02
1.63581610e-01 1.12695122e+00 -6.44672394e-01 -3.11795503e-01
-7.28487968e-01 5.72280884e-01 7.10516497e-02 3.76588136e-01
-1.18645263e+00 -1.68314189e-01 4.19980913e-01 3.31362002e-02
4.66177374e-01 2.71446913e-01 3.18243027e-01 -1.39383101e+00
-1.17959559e-01 -9.77789462e-01 4.08357084e-01 -9.99491274e-01
-1.19673777e+00 4.87722367e-01 -5.06883115e-02 -1.34562278e+00
-8.31105933e-02 -5.51174164e-01 -3.39600027e-01 2.28782848e-01
-1.44760132e+00 -9.78155673e-01 -6.28121912e-01 3.76195520e-01
6.69930518e-01 -1.59441143e-01 3.72869104e-01 6.91039860e-01
-7.62923837e-01 8.39038551e-01 1.00041695e-01 5.78702629e-01
5.78055978e-01 -7.42120981e-01 5.10773510e-02 7.47684717e-01
-1.02187924e-01 3.30827355e-01 6.22424185e-01 -4.47125345e-01
-1.51057184e+00 -1.15250814e+00 2.64963239e-01 -2.47362271e-01
6.50668621e-01 -1.27229154e-01 -1.00881410e+00 7.77764440e-01
3.11102390e-01 2.76030630e-01 6.88273907e-01 -1.69087097e-01
6.45488948e-02 -2.22087219e-01 -7.85669565e-01 4.53811735e-01
1.10291684e+00 -3.10726017e-01 -1.94152519e-01 4.74854618e-01
1.04170847e+00 -6.07356966e-01 -9.28275824e-01 5.42986453e-01
6.19864225e-01 -6.01988256e-01 9.76763248e-01 -9.05038178e-01
1.04813910e+00 8.34509656e-02 -1.07826538e-01 -1.00408840e+00
-4.67535079e-01 -2.54954934e-01 -2.17880607e-01 9.25464809e-01
3.49631131e-01 -1.35967776e-01 9.22012210e-01 5.80247223e-01
-2.68239021e-01 -9.49494183e-01 -9.27784204e-01 -4.25117433e-01
-3.04772407e-01 -5.20657539e-01 3.09705079e-01 1.11709249e+00
2.17898209e-02 1.03555523e-01 -6.08199656e-01 1.48596108e-01
5.73634624e-01 -4.70066676e-03 2.97927856e-01 -7.34890580e-01
-4.98321623e-01 -2.86288202e-01 -5.56808114e-01 -1.00553977e+00
1.85044035e-01 -1.12300920e+00 -1.72982767e-01 -1.31063259e+00
7.15959132e-01 -3.52709621e-01 -3.43888849e-01 4.88893628e-01
-3.40927452e-01 2.98828781e-01 3.79193902e-01 6.51258379e-02
-9.10966754e-01 5.59253395e-01 1.58164728e+00 -5.62272549e-01
1.20628878e-01 -6.67104572e-02 -5.82216322e-01 7.37910688e-01
6.17608607e-01 -4.17757124e-01 -5.13725281e-01 -7.18547881e-01
-8.83256868e-02 4.92022246e-01 3.29050183e-01 -8.69926035e-01
2.11587816e-01 3.22627276e-02 5.46595871e-01 -4.44311291e-01
2.79164433e-01 -7.84977615e-01 -4.03675102e-02 7.57342398e-01
-5.94564676e-01 -1.11549638e-01 9.05850232e-02 7.47908235e-01
-2.45395407e-01 -2.33953252e-01 6.59886420e-01 2.07829736e-02
-5.72172225e-01 7.14767039e-01 -3.75330180e-01 1.24078523e-02
8.98048937e-01 -4.51681316e-02 -2.40361422e-01 -6.03449166e-01
-8.74260664e-01 5.74876130e-01 9.68992338e-02 5.16196370e-01
8.01642954e-01 -1.52712655e+00 -8.02348554e-01 2.85354257e-01
2.46699043e-02 -7.40887672e-02 7.05756068e-01 1.26318777e+00
-7.56289363e-01 5.28229594e-01 -1.65278822e-01 -9.08338547e-01
-1.47764540e+00 5.20117998e-01 5.39697945e-01 7.09504336e-02
-8.13927770e-01 9.43422794e-01 6.07149899e-01 9.72521007e-02
5.26223719e-01 -5.08489847e-01 -4.74598289e-01 -8.95181075e-02
7.48293698e-01 2.53665149e-01 -1.46695614e-01 -4.36799049e-01
-3.12326789e-01 2.05158830e-01 -3.50156575e-01 3.90305102e-01
1.42016602e+00 -1.58367589e-01 9.59387571e-02 3.22642997e-02
1.63397861e+00 -4.48617816e-01 -1.40857697e+00 -2.25869820e-01
-3.99510920e-01 -6.36042714e-01 2.77034551e-01 -4.58831608e-01
-1.68999541e+00 9.02899683e-01 8.69988084e-01 1.21851817e-01
1.30557275e+00 5.89362420e-02 7.19332695e-01 4.02572155e-01
2.05979437e-01 -7.17911899e-01 2.38578111e-01 2.14654252e-01
6.57438099e-01 -1.32882118e+00 -2.33358443e-01 -4.41262752e-01
-5.29225349e-01 9.93600965e-01 7.27242291e-01 5.85560277e-02
7.17750430e-01 1.42186537e-01 3.75678204e-02 -1.45526856e-01
-9.65346038e-01 1.47392988e-01 2.47425720e-01 4.18522477e-01
4.96015280e-01 -2.27589488e-01 -2.96931207e-01 7.52859294e-01
2.00550810e-01 2.84137130e-01 7.27837682e-01 7.16468096e-01
-1.50720865e-01 -5.63025892e-01 -1.20267704e-01 8.44370961e-01
-7.74709880e-01 -4.95202132e-02 3.56086910e-01 5.72803736e-01
-1.11466825e-01 7.72063732e-01 1.07480541e-01 -5.01482546e-01
2.59815633e-01 -2.89071739e-01 5.44038594e-01 -2.44819567e-01
-1.17678314e-01 2.89922446e-01 1.57483622e-01 -8.53317916e-01
-6.55750930e-01 -7.16399014e-01 -1.16329181e+00 -3.60591441e-01
-3.10575217e-01 1.64095089e-02 1.78438842e-01 7.52499998e-01
2.85599709e-01 1.05387747e+00 8.28881145e-01 -7.87116766e-01
-3.57315540e-01 -8.46996903e-01 -4.22418535e-01 5.35188973e-01
4.98290718e-01 -6.39057517e-01 -1.97455436e-01 3.19708407e-01] | [8.945693969726562, 0.3117155134677887] |
41702ffa-f7c4-4165-aff4-8ea1c0953f4e | active-semantic-localization-with-graph | 2305.06141 | null | https://arxiv.org/abs/2305.06141v3 | https://arxiv.org/pdf/2305.06141v3.pdf | Active Semantic Localization with Graph Neural Embedding | Semantic localization, i.e., robot self-localization with semantic image modality, is critical in recently emerging embodied AI applications such as point-goal navigation, object-goal navigation and vision language navigation. However, most existing works on semantic localization focus on passive vision tasks without viewpoint planning, or rely on additional rich modalities (e.g., depth measurements). Thus, the problem is largely unsolved. In this work, we explore a lightweight, entirely CPU-based, domain-adaptive semantic localization framework, called graph neural localizer.Our approach is inspired by two recently emerging technologies: (1) Scene graph, which combines the viewpoint- and appearance- invariance of local and global features; (2) Graph neural network, which enables direct learning/recognition of graph data (i.e., non-vector data). Specifically, a graph convolutional neural network is first trained as a scene graph classifier for passive vision, and then its knowledge is transferred to a reinforcement-learning planner for active vision. Experiments on two scenarios, self-supervised learning and unsupervised domain adaptation, using a photo-realistic Habitat simulator validate the effectiveness of the proposed method. | ['Daiki Iwata', 'Ryogo Yamamoto', 'Kanji Tanaka', 'Mitsuki Yoshida'] | 2023-05-10 | null | null | null | null | ['vision-language-navigation', 'unsupervised-domain-adaptation'] | ['computer-vision', 'methodology'] | [ 2.52252907e-01 4.44006063e-02 -1.76846221e-01 -2.96295494e-01
-4.24039572e-01 -5.22947729e-01 5.62589109e-01 2.42351711e-01
-5.35843551e-01 5.44001102e-01 3.13448869e-02 -4.03796509e-02
1.45708872e-02 -7.26208746e-01 -8.33892643e-01 -7.99222589e-01
-5.08760884e-02 3.00389528e-01 4.76452500e-01 -2.32974529e-01
4.93096530e-01 2.69936383e-01 -1.51790905e+00 -3.64596635e-01
9.60594118e-01 1.16630852e+00 7.80072629e-01 5.12414396e-01
-2.06165746e-01 1.03797543e+00 -7.33258948e-02 4.30199802e-01
5.74037172e-02 -4.30984169e-01 -6.94449961e-01 1.29888386e-01
8.02128166e-02 -1.40776455e-01 -4.11539227e-01 1.43069625e+00
2.48284623e-01 5.42837203e-01 4.90996808e-01 -1.42763209e+00
-6.87931418e-01 1.09990530e-01 -2.28733167e-01 -6.01947941e-02
6.48037612e-01 2.79338181e-01 6.63863719e-01 -7.85212159e-01
6.27938569e-01 1.11784863e+00 5.23426414e-01 5.89971006e-01
-7.89162099e-01 -3.89105797e-01 2.92095810e-01 5.67287743e-01
-1.46931159e+00 -5.83900988e-01 1.21448421e+00 -4.08961862e-01
1.01997113e+00 -4.43008810e-01 9.15299416e-01 1.00644231e+00
4.75613058e-01 5.40441871e-01 1.09902084e+00 -4.43743736e-01
9.41168547e-01 -5.46654314e-02 -1.04860820e-01 1.11959219e+00
2.02589989e-01 2.49781996e-01 -7.44207561e-01 -5.31449802e-02
8.65669012e-01 6.62153885e-02 -2.67462879e-01 -1.21258104e+00
-1.22966957e+00 8.70119631e-01 8.78533781e-01 9.58732292e-02
-4.71899986e-01 3.33631247e-01 2.32381776e-01 1.27989203e-01
2.86730111e-01 1.19079150e-01 -2.26498663e-01 1.23563215e-01
-3.45290393e-01 -1.10363968e-01 8.87397170e-01 1.34232438e+00
1.22875488e+00 2.95329273e-01 3.43518585e-01 5.27635515e-01
7.78369129e-01 8.89453530e-01 6.79554760e-01 -7.48947084e-01
2.17335358e-01 8.34666371e-01 -5.24032675e-03 -1.34498131e+00
-7.51941979e-01 -3.26307304e-02 -6.61583185e-01 3.87461483e-01
7.23064467e-02 -5.19673750e-02 -9.24331725e-01 1.90090096e+00
2.85782635e-01 2.22407922e-01 4.26327169e-01 1.22332275e+00
9.75766718e-01 3.84656549e-01 1.82423502e-01 -8.44605640e-02
9.73635256e-01 -1.28878725e+00 -4.87360537e-01 -8.03983688e-01
5.83340645e-01 -1.70346975e-01 8.47067833e-01 -6.50900528e-02
-5.86768091e-01 -5.17612517e-01 -1.10060167e+00 1.07699186e-01
-7.46835887e-01 -1.88872442e-01 7.40137756e-01 1.57512188e-01
-1.39849317e+00 7.94293359e-02 -9.52521086e-01 -1.16894674e+00
1.82259351e-01 4.08508569e-01 -7.27996707e-01 -2.18692079e-01
-9.57806408e-01 7.84719050e-01 8.02294135e-01 -1.75632611e-02
-1.23249614e+00 9.64144841e-02 -1.47160661e+00 -1.41874269e-01
5.33126533e-01 -7.81388819e-01 9.40732896e-01 -1.16619718e+00
-1.83263195e+00 7.24335909e-01 -3.76014821e-02 -4.17379349e-01
4.43251617e-02 4.32255073e-03 -1.19572528e-01 2.30777711e-01
3.34330022e-01 6.58074319e-01 8.15732598e-01 -1.47413158e+00
-5.78324318e-01 -7.07727611e-01 2.63950497e-01 8.55077028e-01
-1.28512472e-01 -6.13377333e-01 -2.93329448e-01 3.07733957e-02
5.40003717e-01 -9.10173893e-01 -3.94566566e-01 1.22830629e-01
-1.94683641e-01 -1.28350899e-01 7.96431005e-01 -4.70218509e-01
3.48810732e-01 -2.03574300e+00 3.27537835e-01 1.37755573e-01
2.57724285e-01 -1.89674214e-01 -1.14714243e-01 4.44474041e-01
2.90895045e-01 -6.15769327e-01 -2.31810287e-01 -2.58429021e-01
-9.81594995e-02 3.51932138e-01 7.90848862e-03 7.62853742e-01
-1.33863702e-01 1.03356564e+00 -1.42208517e+00 -5.37606239e-01
6.05358422e-01 1.66756094e-01 -4.53331113e-01 4.30665135e-01
-2.12694377e-01 8.15968871e-01 -6.78873837e-01 8.64997208e-01
5.45897722e-01 -3.11820179e-01 1.47102162e-01 -1.76715508e-01
-3.74185950e-01 -2.16388226e-01 -9.16893899e-01 2.57646894e+00
-5.31112611e-01 4.11493361e-01 3.05598021e-01 -1.41357839e+00
1.10014462e+00 -1.75020650e-01 3.52455378e-01 -1.10023379e+00
2.30607793e-01 2.02912241e-01 -4.74091679e-01 -5.73697209e-01
4.74968910e-01 2.64069110e-01 -1.38009489e-01 2.90316921e-02
4.37839597e-01 -2.58736134e-01 -2.71109968e-01 -1.51339332e-02
1.34169912e+00 4.68931347e-01 6.94105685e-01 -3.74882013e-01
5.46604872e-01 5.27285516e-01 2.51177400e-01 8.07954729e-01
-4.29304212e-01 3.73185694e-01 -2.54332155e-01 -1.95442900e-01
-6.55447841e-01 -1.17826080e+00 5.21537185e-01 1.09110653e+00
1.12898040e+00 -1.51933715e-01 -6.78515375e-01 -6.17085814e-01
-1.99320048e-01 4.28906441e-01 -3.45158219e-01 -5.28299034e-01
-3.74736577e-01 -1.20783880e-01 3.16964276e-02 4.98752743e-01
8.89412463e-01 -1.21510017e+00 -9.63910222e-01 2.56819904e-01
-2.40654941e-03 -1.22478354e+00 -1.58909768e-01 3.55254799e-01
-7.66822755e-01 -1.10075688e+00 -4.29758906e-01 -1.15788054e+00
8.75043511e-01 7.64271796e-01 6.89081848e-01 -1.67586118e-01
-1.13111595e-02 1.20326006e+00 -5.76893091e-01 -1.37915134e-01
6.93705529e-02 -6.61477745e-02 2.60561377e-01 3.82645540e-02
2.34172940e-01 -5.47525406e-01 -7.20068693e-01 1.52688041e-01
-5.15637100e-01 6.35545477e-02 8.10646236e-01 1.01113570e+00
6.72139168e-01 -7.34730139e-02 3.25960726e-01 -4.80246395e-01
4.61943358e-01 -5.95520794e-01 -7.44436562e-01 2.74373978e-01
-4.01343733e-01 -1.82575565e-02 7.00461447e-01 -3.59566152e-01
-9.04765069e-01 5.30468047e-01 5.69085442e-02 -3.89302701e-01
-3.75464797e-01 7.17161775e-01 -4.18833762e-01 -7.37800479e-01
6.93238497e-01 7.44380236e-01 1.03732452e-01 1.36341631e-01
6.38821661e-01 4.34731275e-01 6.98592603e-01 -2.64130831e-01
8.19183290e-01 6.51564837e-01 -4.51254956e-02 -1.05984473e+00
-2.81662494e-01 -7.86443949e-01 -7.30291724e-01 -3.38824779e-01
9.20696497e-01 -1.10551238e+00 -6.77553535e-01 5.90641737e-01
-1.08805132e+00 -7.17405677e-01 -1.70901299e-01 7.59858072e-01
-1.11764669e+00 3.53459895e-01 2.29926268e-03 -7.98686743e-01
-1.67081878e-01 -1.02236247e+00 1.19542885e+00 6.40218616e-01
1.58753768e-01 -1.35066879e+00 2.14454621e-01 2.04152074e-02
4.86662269e-01 2.05991894e-01 5.45622349e-01 -6.49569750e-01
-8.41795981e-01 -1.04122639e-01 -3.91104966e-01 -8.67347345e-02
1.44673467e-01 -8.33253384e-01 -8.65792155e-01 -4.93858576e-01
5.08586057e-02 -5.62458694e-01 5.14441311e-01 3.35129380e-01
6.42814994e-01 -1.25582144e-01 -7.06490636e-01 8.53360832e-01
1.60021222e+00 5.05932689e-01 3.24939638e-01 3.90671790e-01
9.10884500e-01 4.38799411e-01 7.25256026e-01 3.78937274e-01
8.22951376e-01 5.35800278e-01 8.06585610e-01 -1.26197711e-01
-6.12061843e-02 -5.00964642e-01 2.70861268e-01 8.31799686e-01
1.10454224e-01 -2.76839495e-01 -1.08258641e+00 3.81715238e-01
-2.31490564e+00 -4.92476612e-01 4.44856435e-01 1.94162238e+00
6.28221184e-02 -7.76306987e-02 -3.75706553e-01 -3.00788164e-01
6.83351934e-01 1.18852220e-01 -9.21608508e-01 1.91782892e-01
-3.99574861e-02 -1.51431426e-01 6.92737103e-01 3.98605347e-01
-1.19357550e+00 1.42527604e+00 4.96637297e+00 4.74472553e-01
-1.13646674e+00 2.81978458e-01 -1.27210855e-01 7.68128991e-01
5.18518034e-03 2.23119691e-01 -2.90117294e-01 1.78577915e-01
3.68908495e-01 -1.61617786e-01 7.35613167e-01 1.48252678e+00
-3.83705050e-02 -3.54738355e-01 -9.82981861e-01 1.35600829e+00
6.26440823e-01 -1.03580046e+00 -1.72310904e-01 -2.97976106e-01
4.18827981e-01 1.45300359e-01 -2.20614046e-01 4.78912026e-01
3.67382020e-01 -6.82230592e-01 8.44692469e-01 5.80823600e-01
7.10196257e-01 -4.31709975e-01 6.92867219e-01 5.43393373e-01
-1.52201915e+00 -2.41615430e-01 -4.36688513e-01 -1.85251936e-01
1.21530086e-01 -6.29764125e-02 -6.54905677e-01 6.98215961e-01
6.29335046e-01 1.09382272e+00 -5.23293555e-01 1.10534692e+00
-3.12407225e-01 1.45443290e-01 -2.20518559e-01 -3.56085867e-01
2.50727206e-01 -2.29663968e-01 6.47444427e-01 7.39925385e-01
2.62660712e-01 1.40354365e-01 6.52612567e-01 8.92940879e-01
4.06203240e-01 1.16365142e-01 -1.14016986e+00 9.30949897e-02
3.76200974e-01 1.20602012e+00 -1.06884873e+00 -5.50047681e-02
-5.45752645e-01 1.28667176e+00 5.59202552e-01 6.15857184e-01
-6.11938059e-01 -4.70607519e-01 2.19795972e-01 -3.64055037e-01
2.00951435e-02 -6.90742671e-01 -2.48984788e-02 -1.23370492e+00
-2.45680302e-01 -2.59825230e-01 1.24447301e-01 -1.00826979e+00
-1.05561638e+00 5.11857569e-01 -2.87098885e-02 -1.14636326e+00
-3.66940945e-01 -6.98796332e-01 -6.62417650e-01 6.43654287e-01
-1.65105236e+00 -1.48965263e+00 -1.10956371e+00 1.00966012e+00
4.75816131e-01 -4.14486378e-01 8.59373569e-01 -9.70381424e-02
-8.28699321e-02 3.38058829e-01 8.83637741e-02 1.28679961e-01
2.94824600e-01 -1.08165598e+00 2.61129379e-01 7.71131873e-01
-7.23867072e-03 4.79438484e-01 4.21218276e-01 -7.21483052e-01
-2.17403698e+00 -1.20428002e+00 1.30345970e-01 -1.58246621e-01
6.69519424e-01 -5.76545835e-01 -5.20011425e-01 6.80869818e-01
-2.04473641e-02 1.58681795e-01 7.15023577e-02 -3.08500886e-01
-6.06786646e-02 -6.04520999e-02 -1.14917731e+00 6.23805642e-01
1.42377090e+00 -6.20198846e-01 -5.44588983e-01 4.64895964e-01
9.15884852e-01 -6.22559786e-01 -3.15389991e-01 3.09297144e-01
3.99061024e-01 -8.03191364e-01 9.55836654e-01 -1.42392144e-01
-1.76808104e-01 -3.63856792e-01 -5.60832441e-01 -1.43999720e+00
-3.88323843e-01 -2.03732312e-01 -1.12717159e-01 9.66148496e-01
-8.88876244e-02 -9.67692912e-01 9.14514422e-01 1.18383124e-01
-5.17874539e-01 -2.30304554e-01 -1.10566163e+00 -8.34007740e-01
-5.51828325e-01 -1.24343023e-01 2.06747860e-01 8.39308143e-01
7.62870163e-02 4.56909508e-01 -3.76877375e-02 4.34761375e-01
8.27149689e-01 5.56872860e-02 8.46828759e-01 -1.01844859e+00
1.01439707e-01 -2.12953448e-01 -1.06023848e+00 -1.37910366e+00
5.33526123e-01 -8.87164891e-01 6.15791202e-01 -1.95074880e+00
-5.52997477e-02 -3.95734161e-01 -1.76310897e-01 5.18859267e-01
2.75024801e-01 -8.37303996e-02 -1.18726268e-01 1.79228634e-01
-1.15803254e+00 9.74905431e-01 9.48457122e-01 -3.84844750e-01
-3.66328657e-01 -2.69459069e-01 -3.07211876e-01 8.05873215e-01
7.61581242e-01 -2.67415792e-01 -8.18259358e-01 -4.18715239e-01
-3.65108810e-02 2.60899633e-01 7.33370900e-01 -1.45002782e+00
9.76190686e-01 -4.30615425e-01 1.10571273e-01 -3.93088728e-01
5.48266351e-01 -1.01589000e+00 -2.14240149e-01 6.63507462e-01
7.18288636e-03 -2.46270355e-02 5.80350161e-02 1.08909786e+00
-1.89839035e-01 -9.47219580e-02 6.55309796e-01 -5.11855543e-01
-1.71351886e+00 4.52205122e-01 -3.35278153e-01 6.86365589e-02
1.12396240e+00 -4.49621439e-01 -4.45001751e-01 -3.78792822e-01
-4.70010221e-01 2.63907403e-01 7.07519233e-01 4.70019370e-01
1.05094826e+00 -1.42840886e+00 -8.63702819e-02 3.34136724e-01
7.87357628e-01 2.74619937e-01 2.74361670e-01 7.16723263e-01
-6.90248132e-01 3.97103608e-01 -3.47235531e-01 -1.03743339e+00
-7.41436422e-01 7.40604877e-01 2.64993519e-01 4.67105031e-01
-7.32257605e-01 7.55986750e-01 5.38149297e-01 -7.68538654e-01
2.22914070e-01 -1.87894061e-01 -2.49747843e-01 -3.98987085e-01
2.44683325e-01 1.24448389e-01 -2.30467677e-01 -9.25807834e-01
-6.56799316e-01 8.93893361e-01 4.38367993e-01 -4.88853864e-02
1.04938293e+00 -5.15398622e-01 -1.63499251e-01 5.59037447e-01
1.04791558e+00 -5.25825322e-01 -1.29695427e+00 -4.54068989e-01
4.05810736e-02 -2.22919032e-01 6.39546365e-02 -4.25846487e-01
-8.01191747e-01 6.14653885e-01 8.70202005e-01 3.62075418e-02
1.02711296e+00 1.16688833e-01 3.77795935e-01 8.46398294e-01
1.19234073e+00 -9.54966724e-01 4.85909373e-01 7.89301515e-01
7.13422298e-01 -1.49396861e+00 -2.83588350e-01 -1.34845078e-01
-5.22272706e-01 1.07470381e+00 9.51738119e-01 -1.69887409e-01
6.45596683e-01 -1.21033236e-01 1.42344400e-01 -3.05365503e-01
-3.28262120e-01 -5.20415902e-01 5.29845394e-02 1.10646510e+00
-3.35771024e-01 3.89008634e-02 2.60829747e-01 4.12160695e-01
-2.93772202e-02 -7.46239796e-02 1.67326197e-01 1.16322947e+00
-7.68555999e-01 -3.00275415e-01 -1.33945644e-01 -6.75615331e-04
6.08587503e-01 -3.48043405e-02 -2.75543630e-01 8.20188999e-01
1.62196547e-01 1.04977643e+00 1.02617452e-02 -5.88976145e-01
2.71434873e-01 -3.32871914e-01 4.96980101e-01 -7.58451521e-01
1.83942467e-01 -3.62692446e-01 -3.81455302e-01 -8.04686308e-01
-5.03844559e-01 -3.64681035e-01 -1.53522193e+00 1.08424082e-01
-3.87348980e-01 9.77704749e-02 1.06955051e+00 1.03008437e+00
2.41360083e-01 3.28584850e-01 4.67128694e-01 -1.13395715e+00
-1.66707903e-01 -7.60307133e-01 -5.49534857e-01 2.58395076e-01
3.16417605e-01 -1.12785518e+00 -3.92597288e-01 -1.00838304e-01] | [7.474506855010986, -1.9533249139785767] |
94fb5380-26a8-4ddb-b7b2-9866246438b2 | dynamic-clustering-and-cluster-contrastive | 2303.06810 | null | https://arxiv.org/abs/2303.06810v1 | https://arxiv.org/pdf/2303.06810v1.pdf | Dynamic Clustering and Cluster Contrastive Learning for Unsupervised Person Re-identification | Unsupervised Re-ID methods aim at learning robust and discriminative features from unlabeled data. However, existing methods often ignore the relationship between module parameters of Re-ID framework and feature distributions, which may lead to feature misalignment and hinder the model performance. To address this problem, we propose a dynamic clustering and cluster contrastive learning (DCCC) method. Specifically, we first design a dynamic clustering parameters scheduler (DCPS) which adjust the hyper-parameter of clustering to fit the variation of intra- and inter-class distances. Then, a dynamic cluster contrastive learning (DyCL) method is designed to match the cluster representation vectors' weights with the local feature association. Finally, a label smoothing soft contrastive loss ($L_{ss}$) is built to keep the balance between cluster contrastive learning and self-supervised learning with low computational consumption and high computational efficiency. Experiments on several widely used public datasets validate the effectiveness of our proposed DCCC which outperforms previous state-of-the-art methods by achieving the best performance. | ['Fei Su', 'Zhicheng Zhao', 'Yunhao Du', 'Mengjia Xue', 'Ziqi He'] | 2023-03-13 | null | null | null | null | ['person-re-identification', 'unsupervised-person-re-identification'] | ['computer-vision', 'computer-vision'] | [-2.58014619e-01 -3.97152722e-01 -3.30257386e-01 -8.17070723e-01
-6.18495643e-01 -3.53069752e-01 5.93293905e-01 2.74079084e-01
-4.51676607e-01 2.63686091e-01 1.56863004e-01 2.46750310e-01
-3.43073994e-01 -4.51837689e-01 -9.21848565e-02 -1.06229687e+00
7.10231811e-02 4.18760061e-01 2.25996375e-01 4.07189280e-01
3.82199794e-01 3.90030980e-01 -1.65653026e+00 -7.98365101e-03
1.11329401e+00 8.55165899e-01 3.25540960e-01 3.33262533e-02
-3.66437435e-01 4.90487546e-01 -3.85963947e-01 5.83049953e-02
2.23000184e-01 -4.60777611e-01 -6.01864576e-01 3.29182744e-01
-3.68760154e-02 1.86688751e-01 -2.23360350e-03 1.17722559e+00
6.21152461e-01 4.04242575e-01 9.22152102e-01 -1.33464503e+00
-5.20918071e-01 6.00087047e-01 -9.53469276e-01 1.91175982e-01
-2.19403759e-01 -9.50300395e-02 8.75736058e-01 -9.33754146e-01
3.71653348e-01 1.24571061e+00 5.53993702e-01 4.44068372e-01
-1.32528412e+00 -9.09807324e-01 4.18012500e-01 2.74966210e-01
-1.77660847e+00 -3.50113064e-01 8.77410233e-01 -4.26892817e-01
4.55485880e-01 1.30972803e-01 1.61892012e-01 3.17759514e-01
-2.04224929e-01 6.79429352e-01 9.84778047e-01 -5.84083617e-01
2.87571341e-01 3.15665692e-01 2.33942911e-01 6.32137060e-01
8.95980746e-03 -2.44247317e-01 -1.52453363e-01 -2.53937617e-02
4.11194652e-01 3.57163697e-01 9.47051197e-02 -6.64295495e-01
-1.08602548e+00 8.15096676e-01 3.99651051e-01 4.41445053e-01
-6.29758984e-02 -2.57428974e-01 4.77094382e-01 6.67142496e-02
4.25696194e-01 1.74119305e-02 -3.71829927e-01 1.33930802e-01
-7.34203756e-01 -7.06758276e-02 2.05588937e-01 8.51343215e-01
1.00648081e+00 -2.64389485e-01 -2.95025468e-01 1.24956119e+00
5.60750663e-01 3.12995225e-01 8.30969274e-01 -5.45756936e-01
3.76933306e-01 8.30667078e-01 -1.86377868e-01 -1.15810168e+00
-5.43840647e-01 -4.58812356e-01 -8.07222188e-01 -1.56106234e-01
9.11919996e-02 -7.03485012e-02 -7.57353008e-01 1.64756799e+00
6.29557073e-01 3.33937973e-01 -2.98316833e-02 8.01940322e-01
6.50439918e-01 5.81866860e-01 3.22683275e-01 -5.28122723e-01
9.62506711e-01 -8.89285862e-01 -7.04875946e-01 1.96918063e-02
6.96758091e-01 -7.69852817e-01 1.04816151e+00 9.74570364e-02
-7.12657988e-01 -8.29229474e-01 -1.01225638e+00 3.37696970e-01
-1.63688630e-01 4.01083946e-01 4.08558190e-01 5.35612285e-01
-5.40105760e-01 3.72052968e-01 -8.20799053e-01 -1.90999582e-01
4.05129284e-01 5.25468051e-01 -1.07669666e-01 -3.57056595e-02
-7.73092091e-01 3.81862223e-01 7.06845760e-01 2.74206083e-02
-5.13111055e-01 -6.62626147e-01 -5.96830010e-01 7.27697983e-02
2.50944883e-01 9.87193920e-03 8.68058622e-01 -1.06549740e+00
-1.28089559e+00 7.47273326e-01 -9.12369192e-02 3.31576243e-02
1.65169865e-01 9.14352760e-02 -7.32901394e-01 -8.67978930e-02
1.64944381e-01 6.35279715e-01 6.25384569e-01 -1.53606761e+00
-9.58918929e-01 -4.66819078e-01 -5.74606597e-01 5.58012426e-01
-6.37448013e-01 1.58140450e-04 -7.48981953e-01 -5.56517899e-01
4.57974613e-01 -9.01120245e-01 -2.79632121e-01 -3.95269543e-01
-3.40698034e-01 -5.70077896e-01 1.02511060e+00 -2.43736297e-01
1.63273001e+00 -2.53210139e+00 -7.72090554e-02 6.24157429e-01
1.45753384e-01 4.02613938e-01 -4.47512604e-02 1.38936520e-01
-1.70889571e-01 -8.19214359e-02 -1.00611031e-01 -2.47242823e-01
-4.37572971e-02 8.35928246e-02 2.61552870e-01 5.46149433e-01
-6.49811476e-02 5.19938290e-01 -8.44629288e-01 -8.88858914e-01
3.40090275e-01 2.32610554e-01 -4.61156607e-01 3.12619358e-01
-1.84580926e-02 3.44324231e-01 -4.91712570e-01 4.54020172e-01
9.07834411e-01 -4.11531687e-01 4.64235574e-01 -4.40705359e-01
-1.81890622e-01 -1.24603599e-01 -1.59853637e+00 1.32346368e+00
-1.44676656e-01 -5.02960160e-02 -1.21152528e-01 -1.24212587e+00
1.18877423e+00 -1.03110135e-01 7.64744639e-01 -6.78635895e-01
1.17895640e-01 8.39930959e-03 -1.17018603e-01 -4.36518371e-01
1.54481709e-01 -3.38562466e-02 5.52005954e-02 4.92001176e-01
-1.28692448e-01 4.93294507e-01 -1.64522696e-02 6.09881990e-02
6.31659150e-01 -1.23979174e-01 2.01596230e-01 -5.69653869e-01
7.84613431e-01 -1.92116216e-01 9.04298782e-01 3.06473404e-01
-2.77913004e-01 5.30059159e-01 3.19556177e-01 -2.71734327e-01
-7.82415211e-01 -1.00278962e+00 -1.84237048e-01 1.18519652e+00
3.89124274e-01 -2.41627365e-01 -8.65679145e-01 -9.91372645e-01
-2.96834242e-02 4.79828268e-01 -4.81114686e-01 -2.78978616e-01
-4.13647681e-01 -1.07131732e+00 1.25610650e-01 4.11821872e-01
5.78125954e-01 -8.68379653e-01 -3.62151861e-02 1.75868914e-01
-1.19867019e-01 -6.48825049e-01 -7.96677113e-01 1.99934244e-01
-7.40058303e-01 -9.87155735e-01 -3.81179661e-01 -1.36464429e+00
1.11163962e+00 4.71817642e-01 6.66942894e-01 2.59248525e-01
-2.86704421e-01 -1.81981735e-02 -5.50563037e-01 -2.66289823e-02
-1.75976276e-01 3.61131340e-01 2.21566975e-01 2.43056774e-01
7.86167264e-01 -3.63917053e-01 -7.03369200e-01 6.14268541e-01
-9.40327287e-01 -2.11443022e-01 4.70994502e-01 8.23574543e-01
8.73969316e-01 3.83266151e-01 7.89039135e-01 -9.62457895e-01
5.93685746e-01 -6.31203890e-01 -5.43741941e-01 4.45530981e-01
-1.24916434e+00 1.85235262e-01 7.38045812e-01 -5.39228022e-01
-1.14697886e+00 3.18433523e-01 1.48991570e-01 -4.03336018e-01
-9.15840268e-02 4.35148388e-01 -5.09740233e-01 1.79201856e-01
4.21226978e-01 2.92448610e-01 1.21794768e-01 -5.06867349e-01
3.72450948e-01 1.02797806e+00 3.94781530e-01 -4.78537112e-01
8.58277738e-01 4.78261709e-01 -2.34514445e-01 -2.03629777e-01
-8.58866036e-01 -7.77152121e-01 -8.87878895e-01 -7.83313736e-02
6.27637267e-01 -9.33304727e-01 -6.70302927e-01 3.98666173e-01
-4.02324229e-01 1.99781694e-02 -7.27649555e-02 6.21669233e-01
-2.09992066e-01 5.29711902e-01 -3.12763005e-01 -6.08868361e-01
-3.18968505e-01 -1.09309232e+00 5.83783388e-01 4.93971914e-01
1.80223987e-01 -1.02987921e+00 1.63137510e-01 7.85421208e-02
1.47743419e-01 8.60539824e-02 9.79752541e-01 -7.42048383e-01
-2.84639914e-02 -2.23991461e-02 -4.04305905e-01 4.33037668e-01
6.23666465e-01 -1.57899375e-03 -7.87928164e-01 -4.66617942e-01
-3.44463766e-01 -1.68994918e-01 6.69256866e-01 2.53426075e-01
1.28829658e+00 -3.09091657e-01 -6.79056048e-01 6.01412892e-01
1.54343891e+00 4.52781558e-01 3.98808569e-01 2.54513353e-01
9.21648920e-01 5.63339233e-01 1.01895845e+00 6.91082537e-01
4.93284613e-01 5.84796906e-01 1.61980718e-01 -1.98618993e-01
1.11095816e-01 -3.63349199e-01 -2.21860744e-02 1.11544836e+00
2.05112383e-01 1.08962514e-01 -8.19545388e-01 5.01355231e-01
-1.87386811e+00 -8.52160633e-01 -5.46972267e-02 2.27879667e+00
8.49003434e-01 1.34323493e-01 3.06335658e-01 1.45157441e-01
1.14479578e+00 -1.08072244e-01 -6.76711679e-01 -5.18925525e-02
1.47843570e-01 -2.37013608e-01 4.41969663e-01 3.10733229e-01
-1.34072864e+00 9.10489202e-01 5.78092432e+00 1.20319128e+00
-1.14136446e+00 4.12192941e-02 6.52178586e-01 1.38400048e-01
-3.13169718e-01 -8.01343024e-02 -9.60158885e-01 8.15688431e-01
5.41694820e-01 -3.61770652e-02 3.84934634e-01 9.75586832e-01
2.29478806e-01 1.27309784e-01 -7.45672703e-01 1.12317789e+00
4.83829789e-02 -9.41109002e-01 -1.81221873e-01 -4.75144833e-02
8.63356233e-01 -1.17608964e-01 1.49627996e-03 2.59573996e-01
4.31789875e-01 -6.95998192e-01 3.43611509e-01 4.23996270e-01
6.41513348e-01 -1.24811757e+00 6.74016356e-01 3.15768450e-01
-1.53725612e+00 -3.47071469e-01 -4.75878388e-01 4.76785660e-01
-2.41840184e-01 6.67890072e-01 -6.49790764e-01 4.44057763e-01
7.65677512e-01 7.64860570e-01 -7.89142966e-01 9.58532631e-01
1.66322902e-01 5.37748933e-01 -1.71801567e-01 -3.89773548e-02
1.92855552e-01 -2.62560368e-01 5.95818348e-02 1.24581409e+00
8.01887810e-02 9.86742205e-04 5.98059118e-01 5.13124824e-01
6.07750155e-02 4.83232170e-01 1.36056244e-02 2.58776695e-01
1.05025280e+00 1.44974756e+00 -8.87887239e-01 -2.51944095e-01
-3.24413925e-01 7.92618513e-01 5.58593631e-01 1.50267363e-01
-7.84185708e-01 -4.43613857e-01 4.18074340e-01 5.42121083e-02
4.71561402e-01 -7.35299736e-02 -4.73836847e-02 -9.45622027e-01
-1.29849344e-01 -7.17346072e-01 7.34465897e-01 -1.05789013e-01
-1.66603816e+00 3.30929786e-01 9.97153893e-02 -1.67552578e+00
-5.15561812e-02 -3.26524749e-02 -5.48870862e-01 4.63713288e-01
-1.49423683e+00 -1.13484824e+00 -3.31731051e-01 7.86727846e-01
4.78703767e-01 -3.29703897e-01 5.05038857e-01 4.45383638e-01
-8.75376523e-01 9.66697574e-01 5.97086549e-01 5.21460064e-02
9.46494877e-01 -1.04329622e+00 -1.38485044e-01 6.63941085e-01
-1.58369899e-01 5.94563127e-01 3.69458228e-01 -4.38813001e-01
-9.88384962e-01 -1.32202506e+00 6.96711838e-01 7.50068277e-02
2.22676426e-01 -2.72808492e-01 -9.52458560e-01 3.43985498e-01
-3.19010794e-01 2.97362775e-01 9.51118171e-01 -1.23819918e-03
-5.19764900e-01 -6.44530833e-01 -1.43062878e+00 3.89096767e-01
8.61227155e-01 -1.30556658e-01 -1.75872877e-01 2.07240552e-01
5.50490916e-01 5.31251281e-02 -1.12004614e+00 3.75389487e-01
4.60358828e-01 -7.70155847e-01 8.52702796e-01 -3.01445603e-01
-8.32948536e-02 -7.04530418e-01 -6.31138310e-02 -1.23701656e+00
-7.78145015e-01 -1.73160821e-01 1.87921137e-01 1.95681500e+00
2.12722749e-01 -5.03709197e-01 7.05782950e-01 5.21195889e-01
2.43124217e-02 -5.50983906e-01 -5.60620070e-01 -7.34454691e-01
-1.50145786e-02 1.17621139e-01 7.36331642e-01 1.37480879e+00
2.44728088e-01 3.08356613e-01 -1.42310917e-01 1.37491718e-01
8.13644350e-01 1.13090426e-01 5.04615247e-01 -1.32402730e+00
-9.76373535e-03 -3.75786632e-01 -2.71950930e-01 -7.06940293e-01
2.15510800e-01 -1.07916057e+00 1.56789511e-01 -1.19623709e+00
5.59095681e-01 -1.11601233e+00 -7.21550703e-01 5.74805915e-01
-5.65579236e-01 4.54890542e-03 4.02736515e-02 7.08037078e-01
-1.08354294e+00 4.75044221e-01 9.32052732e-01 5.56707419e-02
-4.96467650e-01 -1.48992226e-01 -7.37684906e-01 5.70612133e-01
9.59903955e-01 -6.99726999e-01 -6.33816421e-01 -1.83260098e-01
-3.44685912e-01 -2.99979806e-01 -2.26197734e-01 -9.62741077e-01
3.82862836e-01 -3.75806838e-01 5.44244587e-01 -6.77558959e-01
-3.93563300e-01 -8.67694438e-01 1.12011008e-01 4.69374061e-01
-4.58404511e-01 4.33170907e-02 -2.17258468e-01 5.33012986e-01
-1.50308669e-01 -2.68558651e-01 1.08072698e+00 -3.99957672e-02
-6.14788175e-01 3.87861431e-01 -1.34872913e-01 8.94092619e-02
1.18892610e+00 -8.12785253e-02 -1.28124449e-02 1.71622813e-01
-4.03136492e-01 5.30394733e-01 5.18850505e-01 4.20089155e-01
4.67897207e-01 -1.52826130e+00 -5.84815741e-01 3.83550197e-01
4.07605946e-01 6.34070486e-02 3.89917105e-01 5.41602314e-01
-3.03233445e-01 1.34026870e-01 -4.94563133e-02 -6.62371218e-01
-1.28596199e+00 7.65875638e-01 1.19346872e-01 -3.04238707e-01
-5.19362688e-01 5.62909007e-01 5.44192549e-03 -7.14576602e-01
4.36699390e-01 2.77491063e-01 -4.85100001e-01 1.68546006e-01
5.13198793e-01 5.01780212e-01 -4.51086834e-02 -6.84744537e-01
-5.00810027e-01 6.93628013e-01 -4.57909882e-01 2.20216423e-01
1.14981115e+00 -5.67732871e-01 -6.90236837e-02 2.45074481e-01
1.49568069e+00 -1.13846637e-01 -1.45500171e+00 -5.10774672e-01
1.56930566e-01 -3.95993382e-01 4.45647053e-02 -4.72216666e-01
-1.24547589e+00 4.38068122e-01 1.09373677e+00 4.57041599e-02
1.20306849e+00 5.19319661e-02 5.81806898e-01 1.62365705e-01
-4.10239957e-03 -1.55856740e+00 2.99156338e-01 1.27422586e-01
1.04183532e-01 -1.31315231e+00 1.23903573e-01 -2.85066545e-01
-8.71234357e-01 7.87380099e-01 8.03895652e-01 -6.70096576e-02
9.44902301e-01 1.46110862e-01 2.94674754e-01 -5.29584698e-02
-4.28747892e-01 -1.70468733e-01 1.79477453e-01 5.97974181e-01
3.55413646e-01 8.46032575e-02 -5.92694163e-01 5.73212564e-01
1.66929871e-01 -2.46906772e-01 -8.10098425e-02 8.20839763e-01
-6.08699977e-01 -1.24966252e+00 -2.60143906e-01 4.46086854e-01
-3.97461981e-01 2.42415711e-01 3.58985513e-02 6.29203618e-01
4.37527567e-01 1.07441831e+00 2.89433599e-01 -6.08990133e-01
4.41264398e-02 -6.73849434e-02 9.33476239e-02 -5.31892240e-01
-4.95551586e-01 4.03210104e-01 -4.43474889e-01 -1.09373942e-01
-6.71645045e-01 -7.91905999e-01 -1.65694225e+00 -1.67205352e-02
-6.64148688e-01 3.71289372e-01 5.47482848e-01 8.48913610e-01
3.58432740e-01 2.97763616e-01 1.46517384e+00 -4.22805935e-01
-5.11525095e-01 -8.66563559e-01 -8.23720753e-01 6.64325416e-01
8.79153260e-04 -7.46421337e-01 -4.16762084e-01 2.25647733e-01] | [14.860363006591797, 1.220374584197998] |
8622021e-1c64-468d-9ee2-e3682a3cbd1b | finding-lookalike-customers-for-e-commerce | 2301.03147 | null | https://arxiv.org/abs/2301.03147v2 | https://arxiv.org/pdf/2301.03147v2.pdf | Finding Lookalike Customers for E-Commerce Marketing | Customer-centric marketing campaigns generate a large portion of e-commerce website traffic for Walmart. As the scale of customer data grows larger, expanding the marketing audience to reach more customers is becoming more critical for e-commerce companies to drive business growth and bring more value to customers. In this paper, we present a scalable and efficient system to expand targeted audience of marketing campaigns, which can handle hundreds of millions of customers. We use a deep learning based embedding model to represent customers and an approximate nearest neighbor search method to quickly find lookalike customers of interest. The model can deal with various business interests by constructing interpretable and meaningful customer similarity metrics. We conduct extensive experiments to demonstrate the great performance of our system and customer embedding model. | ['Wei Shen', 'Changzheng Liu', 'Yang Peng'] | 2023-01-09 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [-7.50886917e-01 -1.73110381e-01 -4.23680484e-01 -1.05760658e+00
-7.71901488e-01 -4.58153069e-01 1.97546825e-01 6.01300299e-01
-2.24294186e-01 1.23532534e-01 3.69748384e-01 -3.49195868e-01
-1.44889727e-01 -1.35805821e+00 -2.74662673e-01 -1.33497849e-01
-9.03532431e-02 1.03105187e+00 -1.50533214e-01 -7.91953504e-01
3.23520929e-01 4.26374048e-01 -1.28525269e+00 6.08709097e-01
5.13051271e-01 1.32370412e+00 -1.33841500e-01 3.25022608e-01
-5.26217520e-01 3.89904678e-01 -2.40048215e-01 -9.27352786e-01
4.80013132e-01 -4.47239242e-02 -5.26408076e-01 8.15059617e-02
2.32967809e-01 -5.01459837e-01 -2.95655310e-01 7.27621078e-01
5.33041179e-01 5.59418239e-02 2.75397480e-01 -1.42595196e+00
-1.22834802e+00 1.06057703e+00 -7.37689495e-01 1.87589392e-01
-2.04882100e-02 -1.00771569e-01 1.57702065e+00 -9.04316604e-01
4.53586310e-01 1.36490595e+00 5.24553239e-01 1.24286816e-01
-1.07856750e+00 -5.46409130e-01 3.43406975e-01 3.82746637e-01
-9.63689804e-01 1.66361824e-01 9.71727610e-01 8.71354043e-02
7.73442268e-01 4.16551143e-01 1.00129211e+00 8.51488471e-01
-1.25287160e-01 9.70701516e-01 4.06739622e-01 -3.53370458e-02
2.27320865e-01 5.79840839e-01 5.21566570e-01 3.02948859e-02
3.96531880e-01 -1.06876716e-01 -2.82406360e-01 -1.44836381e-01
6.12677991e-01 7.09849477e-01 4.98173922e-01 -3.99807632e-01
-7.45497406e-01 1.75852191e+00 8.88776422e-01 1.50511473e-01
-5.22638142e-01 2.55811960e-01 5.92688143e-01 3.18070322e-01
5.06046951e-01 5.08875430e-01 -4.61045682e-01 -1.93461716e-01
-3.89739394e-01 4.77263957e-01 9.05575335e-01 1.10803878e+00
5.33376873e-01 -1.67164132e-01 4.98421878e-01 8.65836561e-01
4.05157030e-01 5.02277553e-01 5.36069930e-01 -8.07843626e-01
2.92901218e-01 1.07947910e+00 -1.42483056e-01 -1.38107610e+00
-2.00716063e-01 -8.58065665e-01 -4.22778577e-01 -5.35563469e-01
2.46175770e-02 2.88518846e-01 -3.22180152e-01 1.01173782e+00
4.02329266e-01 -3.06790322e-01 -1.49091870e-01 9.42337036e-01
6.32580340e-01 7.57549226e-01 -2.39509679e-02 2.29403123e-01
1.42991579e+00 -1.02258337e+00 -3.74177575e-01 -3.55588555e-01
7.83389091e-01 -9.26657498e-01 1.28033316e+00 2.11988047e-01
-9.25533056e-01 -3.60651881e-01 -6.40090942e-01 -2.82660186e-01
-5.76074481e-01 -5.08016825e-01 1.26427209e+00 3.45802814e-01
-4.84625965e-01 4.17642891e-01 -4.81524855e-01 -3.93750191e-01
5.98201931e-01 4.94491160e-01 8.43842402e-02 -2.81639338e-01
-1.26415205e+00 5.10322869e-01 1.62605435e-01 -9.53076333e-02
-1.56128392e-01 -5.70506155e-01 -7.84228802e-01 3.51735473e-01
-1.19371057e-01 -4.30090725e-01 1.38609517e+00 -6.62380517e-01
-5.94646215e-01 6.78791463e-01 9.77507159e-02 -5.79402983e-01
1.43224508e-01 -1.95386365e-01 -6.70670271e-01 -2.56861925e-01
6.62046894e-02 6.23194873e-01 2.60164559e-01 -1.10359824e+00
-1.20181382e+00 -6.46289706e-01 -5.58254123e-02 2.74028871e-02
-8.71944606e-01 4.99280915e-03 -3.16269904e-01 -2.17000782e-01
3.90191287e-01 -7.38048613e-01 -4.86788183e-01 -1.87699452e-01
6.34494983e-03 -4.43656832e-01 8.69417727e-01 -6.05160952e-01
1.15337300e+00 -2.06084585e+00 -2.87119299e-01 3.96718651e-01
5.40082455e-01 -9.87672359e-02 -4.32652086e-01 7.34896243e-01
1.89800262e-01 7.06280097e-02 4.73143101e-01 2.14922175e-01
4.06125247e-01 2.01775134e-01 -4.58756089e-01 7.90596846e-03
-5.05550019e-02 1.43738031e+00 -8.64894211e-01 -2.05707446e-01
1.79728314e-01 6.05434060e-01 -7.84189165e-01 -4.37681228e-02
-9.16100219e-02 -3.34078938e-01 -8.87751877e-01 8.15981030e-01
7.10344195e-01 -4.53475952e-01 1.55674508e-02 -3.53613079e-01
3.94788653e-01 2.61956453e-01 -7.59229839e-01 9.77489412e-01
-6.26323342e-01 5.41758895e-01 -1.27576590e-01 -9.44669783e-01
1.21578813e+00 -3.88484001e-01 7.04351783e-01 -1.13456404e+00
4.94157702e-01 2.28823185e-01 -4.60875481e-02 -1.82767302e-01
8.37308228e-01 -1.85669199e-01 -4.10168350e-01 6.12072229e-01
-5.33420444e-01 3.77723992e-01 1.67487949e-01 2.43911818e-01
8.03483605e-01 -7.08188176e-01 1.00230433e-01 5.30105736e-03
1.83784083e-01 3.62083584e-01 4.09936935e-01 9.94552299e-02
-3.32356155e-01 3.37796696e-02 2.92809039e-01 -1.08615983e+00
-1.29616725e+00 -1.15722334e+00 -9.05825198e-02 1.43346417e+00
2.05511317e-01 -1.56217963e-01 -3.52453142e-01 -6.12691164e-01
7.20592558e-01 1.08805203e+00 -2.68029571e-01 -3.27091306e-01
-6.96865857e-01 -6.84342504e-01 -1.93921641e-01 8.47815096e-01
3.79764408e-01 -7.87397146e-01 3.77529152e-02 4.98944640e-01
-2.43259072e-01 -8.75382662e-01 -6.25774980e-01 -2.60616839e-01
-1.06423640e+00 -8.49935830e-01 -2.69434392e-01 -1.35307348e+00
5.65179527e-01 5.90574265e-01 1.27555335e+00 -3.47700447e-01
-4.01551634e-01 1.78425997e-01 -5.46864688e-01 -4.11674857e-01
-2.68318534e-01 2.65445620e-01 -1.18670398e-02 8.31924193e-03
1.58343065e+00 -3.63548756e-01 -9.59681988e-01 5.84545076e-01
-8.08746219e-01 -5.01539946e-01 5.61011612e-01 5.71783245e-01
5.94683349e-01 1.89702839e-01 1.02472842e+00 -1.01388872e+00
1.22557867e+00 -8.91784787e-01 -3.95538241e-01 -2.49284431e-02
-1.05627179e+00 -1.23179883e-01 5.49755156e-01 -5.98372400e-01
-3.45473289e-01 -2.88325906e-01 -2.05494240e-01 -1.09478988e-01
3.89396608e-01 6.80760503e-01 -5.53879179e-02 3.23122710e-01
5.35182178e-01 1.65096939e-01 -7.28637388e-04 -6.33074462e-01
9.03472304e-01 1.21441340e+00 2.44810238e-01 -5.20988777e-02
6.30660474e-01 3.29468459e-01 -3.77235919e-01 -3.31216097e-01
-5.66824555e-01 -8.75094414e-01 -2.49884591e-01 4.78658341e-02
4.71090198e-01 -7.84022510e-01 -1.11734724e+00 -2.00691327e-01
-9.05497611e-01 4.86829132e-01 -5.75398743e-01 5.51271737e-01
-2.71460980e-01 7.43668750e-02 -8.90539944e-01 -4.21880603e-01
-4.47716475e-01 -8.20839643e-01 8.49811435e-01 9.25740525e-02
-2.06182897e-01 -1.07424426e+00 -5.02479449e-02 7.66903996e-01
7.17779934e-01 -3.92943658e-02 1.00610602e+00 -1.38006425e+00
-7.22310483e-01 -9.71885979e-01 -4.72581983e-01 3.65946919e-01
1.17678391e-02 -5.77517033e-01 -3.89058530e-01 -2.72498667e-01
-1.14352711e-01 -3.77668403e-02 7.44337678e-01 3.05855274e-01
9.75597620e-01 -4.46819723e-01 -3.55787665e-01 3.68641555e-01
1.42165756e+00 3.03953379e-01 4.74650472e-01 6.04883850e-01
5.75868249e-01 8.87509644e-01 7.95365691e-01 7.17772245e-01
7.64376998e-01 2.53516823e-01 6.13390565e-01 -1.49162533e-02
5.26525140e-01 -5.23158729e-01 -6.25153854e-02 1.03360057e+00
2.90108532e-01 -1.47910357e-01 -6.50816679e-01 5.78953564e-01
-1.96216285e+00 -1.02706873e+00 -8.29571951e-03 2.00077605e+00
3.40768158e-01 2.07535386e-01 3.07945997e-01 -1.07148588e-01
8.37253034e-01 -1.84401169e-01 -8.84409666e-01 -8.31956029e-01
1.36631086e-01 -1.63883343e-01 6.24884248e-01 2.03819945e-01
-7.33744085e-01 9.55985308e-01 6.37597132e+00 6.63782656e-01
-9.82425511e-01 -4.46096659e-02 1.08498514e+00 -4.75202024e-01
-8.79286289e-01 -2.42775500e-01 -1.04935205e+00 4.00958508e-01
1.04894471e+00 -5.19963086e-01 4.30582434e-01 1.51355362e+00
1.28265679e-01 7.60392249e-01 -1.22209060e+00 1.27445626e+00
3.09590921e-02 -1.58662820e+00 1.75671086e-01 5.68307519e-01
5.18852413e-01 -5.73724099e-02 3.62154752e-01 6.03214085e-01
5.30823946e-01 -8.21753502e-01 5.50515354e-02 -4.47974578e-02
1.65803671e-01 -1.16465592e+00 9.49284673e-01 1.31623559e-02
-1.13907123e+00 -5.25589585e-01 -6.94132209e-01 -1.21871926e-01
6.19140208e-01 5.90600252e-01 -8.41747224e-01 -2.89401591e-01
6.81995809e-01 6.33494318e-01 -3.17593813e-01 7.81869531e-01
4.39326018e-01 4.04979259e-01 -2.50748992e-01 -6.17297769e-01
5.02095640e-01 -3.81623685e-01 1.30500406e-01 7.74132192e-01
3.12811673e-01 -1.58790514e-01 1.03875972e-01 7.16986179e-01
-3.94187570e-01 6.47716165e-01 -4.93082494e-01 -5.90667844e-01
6.59780264e-01 1.47976720e+00 -5.51557243e-01 -2.29135767e-01
-5.79899251e-01 8.46500635e-01 2.19979420e-01 2.27681864e-02
-1.01290882e+00 -5.65032303e-01 9.12154138e-01 7.66970396e-01
4.40538138e-01 -1.25743195e-01 -2.03207344e-01 -7.37875760e-01
8.08486529e-03 -8.32160771e-01 4.35258836e-01 -4.42633510e-01
-1.96977413e+00 4.96126920e-01 -4.34431255e-01 -1.08850896e+00
-3.06917220e-01 -5.94428897e-01 -5.26936531e-01 6.30887926e-01
-1.35759926e+00 -1.42985106e+00 -6.06800206e-02 2.28748932e-01
5.72069287e-01 -4.06757414e-01 5.50711751e-01 4.93536621e-01
-2.30373636e-01 7.97097683e-01 5.42170346e-01 2.35078231e-01
4.36203808e-01 -8.61751318e-01 7.12045372e-01 -2.77267992e-02
7.28938282e-02 9.29933071e-01 4.97624665e-01 -4.14668828e-01
-1.57031953e+00 -1.15494251e+00 1.47700691e+00 -3.66748244e-01
1.04167163e+00 -4.22283322e-01 -5.85577965e-01 8.48948419e-01
4.54102680e-02 -3.88990343e-01 1.23506165e+00 6.49559617e-01
-5.43820083e-01 -7.64237344e-01 -1.11863792e+00 4.49111015e-01
6.76641107e-01 -2.24060312e-01 -4.41579789e-01 6.00572407e-01
1.23149800e+00 2.57341713e-01 -1.13169897e+00 -9.12724733e-02
5.41529596e-01 -7.62924194e-01 1.23074031e+00 -9.59894896e-01
4.89108890e-01 3.58278245e-01 -5.05102277e-01 -1.06722820e+00
-7.31769919e-01 -2.22624943e-01 -2.20881347e-02 1.19502366e+00
6.28055155e-01 -7.94328630e-01 1.23334086e+00 1.06414187e+00
1.61614895e-01 -1.12269437e+00 -5.35233200e-01 -6.59785688e-01
7.75967687e-02 -3.83561820e-01 1.23010170e+00 8.99584055e-01
1.82493508e-01 6.78905487e-01 -1.38140008e-01 -2.14434713e-01
6.04820073e-01 7.69036412e-01 6.93463922e-01 -1.36555779e+00
-3.06034118e-01 -5.27565300e-01 -7.66971827e-01 -1.27976477e+00
-4.20317322e-01 -1.04092932e+00 -6.32998049e-01 -1.68903041e+00
2.80906111e-01 -6.10340476e-01 -5.55226743e-01 1.87847510e-01
3.23628813e-01 3.17090780e-01 -5.94511367e-02 1.46320239e-01
-6.82367325e-01 4.62432563e-01 1.22051764e+00 -2.54399657e-01
-2.60380238e-01 2.12886348e-01 -1.62348592e+00 3.94715309e-01
1.00420427e+00 -1.47567019e-01 -4.81433421e-01 -4.75226939e-01
8.79859328e-01 -3.34128022e-01 -3.25673670e-02 -3.02890152e-01
-2.35893717e-03 -1.23549350e-01 3.87642264e-01 -7.99062908e-01
2.77546853e-01 -1.02619302e+00 -2.53518105e-01 4.88237560e-01
-5.43138802e-01 4.14451212e-01 -1.15959547e-01 6.87749207e-01
-4.05487120e-01 2.07915567e-02 3.48784089e-01 3.15266214e-02
-7.65231013e-01 6.04683757e-01 2.40798723e-02 -1.04426272e-01
1.18960381e+00 -2.40633473e-01 -2.70234972e-01 -7.46859610e-01
-6.82724178e-01 5.42580187e-01 1.45247087e-01 7.99188614e-01
7.67643154e-01 -1.71936846e+00 -7.70899832e-01 2.91446239e-01
4.36020404e-01 -3.43262225e-01 1.05412252e-01 1.93977326e-01
-6.42511904e-01 5.35218596e-01 -1.23540252e-01 -3.11999500e-01
-9.74771917e-01 8.48909020e-01 -2.38205776e-01 -1.99031472e-01
-4.15374666e-01 8.99939775e-01 -5.54991476e-02 -7.10476518e-01
1.35939792e-01 -2.38980770e-01 -1.62623942e-01 2.08076403e-01
6.52450562e-01 6.46480620e-01 -3.27635497e-01 -6.29523933e-01
-2.17053175e-01 3.06117088e-01 -6.40855074e-01 2.01591447e-01
1.71912968e+00 -2.00522184e-01 5.70989922e-02 -8.24628212e-03
1.84590578e+00 -1.89363524e-01 -8.80609751e-01 -3.24727118e-01
-1.52571052e-01 -8.36623013e-01 2.51657873e-01 -5.50132334e-01
-1.39200211e+00 9.20448422e-01 7.09583580e-01 5.18915534e-01
9.48854804e-01 4.03831393e-01 1.83414268e+00 4.58251297e-01
3.61543447e-01 -1.19888127e+00 1.73546478e-01 4.11378220e-03
5.68553448e-01 -1.53326106e+00 -2.24910080e-01 -3.13165545e-01
-8.70568514e-01 1.01901448e+00 3.46548855e-01 -1.89353287e-01
9.43217278e-01 -7.13600218e-02 2.43635610e-01 -3.37903202e-01
-5.60295224e-01 -5.80704324e-02 -1.05693109e-01 5.27583301e-01
3.68234068e-01 3.01150918e-01 -5.00333071e-01 1.03244519e+00
-4.75216478e-01 -3.37537169e-01 3.15664038e-02 7.46134937e-01
-9.91749287e-01 -1.48213995e+00 1.63307920e-01 7.16942668e-01
-9.08346996e-02 -3.31794202e-01 -3.20251346e-01 7.68557608e-01
-1.91491261e-01 9.61968064e-01 3.99833053e-01 -7.14904606e-01
4.49179322e-01 -2.10688815e-01 -3.73333469e-02 -2.91369706e-01
-4.74633634e-01 1.47214755e-01 -5.08735739e-02 -4.44522649e-01
2.81985760e-01 -4.73701179e-01 -1.42482245e+00 -1.17415380e+00
-3.11585695e-01 2.49236450e-01 1.10393393e+00 2.65454531e-01
7.83600569e-01 3.35101523e-02 1.15206134e+00 -1.87162772e-01
-1.05135453e+00 -6.59713209e-01 -9.06695962e-01 7.74086297e-01
-4.27554160e-01 -1.63176626e-01 -2.73283273e-01 -2.38077492e-01] | [9.968194007873535, 6.013302326202393] |
ca36329e-2e96-49a0-b6f3-4fdcbe405ead | local-and-global-point-cloud-reconstruction | 2112.06389 | null | https://arxiv.org/abs/2112.06389v1 | https://arxiv.org/pdf/2112.06389v1.pdf | Local and Global Point Cloud Reconstruction for 3D Hand Pose Estimation | This paper addresses the 3D point cloud reconstruction and 3D pose estimation of the human hand from a single RGB image. To that end, we present a novel pipeline for local and global point cloud reconstruction using a 3D hand template while learning a latent representation for pose estimation. To demonstrate our method, we introduce a new multi-view hand posture dataset to obtain complete 3D point clouds of the hand in the real world. Experiments on our newly proposed dataset and four public benchmarks demonstrate the model's strengths. Our method outperforms competitors in 3D pose estimation while reconstructing realistic-looking complete 3D hand point clouds. | ['Angela Yao', 'Shicheng Chen', 'Linlin Yang', 'Ziwei Yu'] | 2021-12-13 | null | null | null | null | ['3d-point-cloud-reconstruction', 'point-cloud-reconstruction', '3d-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-4.46850002e-01 -5.14488995e-01 -1.80620909e-01 -1.03501290e-01
-8.59080076e-01 -8.53061974e-01 1.99732363e-01 -7.24873126e-01
-2.48049483e-01 1.07283182e-01 1.41991541e-01 -4.63667372e-03
1.32725090e-01 -2.80513525e-01 -6.83892429e-01 -4.40118283e-01
3.03183436e-01 1.42307317e+00 2.57068247e-01 -9.70072001e-02
3.26717436e-01 1.36667132e+00 -1.36864519e+00 2.17912234e-02
-3.26984972e-02 7.31381118e-01 3.12500179e-01 6.48343027e-01
1.15883797e-01 7.61574060e-02 -4.84825492e-01 -2.72373915e-01
7.89031982e-01 4.01084721e-01 -9.88021433e-01 3.51650268e-01
8.27498078e-01 -9.41521704e-01 -3.61997873e-01 4.47555095e-01
1.05151224e+00 -1.92352623e-01 4.74261403e-01 -1.15391588e+00
-3.14931609e-02 -1.00141913e-01 -7.29313612e-01 -3.99620265e-01
1.00005472e+00 4.43111956e-01 8.64982069e-01 -1.25748992e+00
1.08991027e+00 1.54783678e+00 7.46721506e-01 5.61206758e-01
-9.71765399e-01 -4.96116281e-01 2.68400878e-01 4.03644284e-03
-1.68669868e+00 1.54555375e-02 9.17260528e-01 -4.65551466e-01
1.33122528e+00 2.69980550e-01 1.24151742e+00 1.50950181e+00
6.02683872e-02 1.02905583e+00 1.25421107e+00 -4.42195177e-01
-1.01345815e-01 -4.65336680e-01 -6.59238100e-02 7.16928720e-01
-4.70703430e-02 2.73301154e-02 -7.94663608e-01 -3.74692410e-01
1.59591067e+00 4.24273193e-01 -1.00629583e-01 -9.31613922e-01
-1.42249274e+00 7.79917464e-02 6.92676187e-01 -1.96191907e-01
-6.50759399e-01 3.38535845e-01 -9.76514965e-02 -2.36563370e-01
2.61070728e-01 -1.23284169e-01 -7.73248911e-01 -4.11130458e-01
-7.95619667e-01 6.54534459e-01 7.56379128e-01 1.47619092e+00
1.87515154e-01 -5.39140940e-01 -1.68246552e-02 1.64817408e-01
6.44498765e-01 9.04995978e-01 -1.20677903e-01 -1.12448812e+00
7.97956407e-01 7.30243623e-01 3.83747399e-01 -5.69961786e-01
-5.37975132e-01 -4.03560489e-01 -2.92003453e-01 5.05623817e-01
5.38230240e-01 5.39188445e-01 -1.00078273e+00 9.64114904e-01
4.41257477e-01 -1.73069522e-01 -7.36816585e-01 1.34110844e+00
6.37313247e-01 8.05492699e-02 -4.53879267e-01 2.80767649e-01
1.36206901e+00 -7.81517863e-01 -4.46417898e-01 -2.32234627e-01
-2.09114760e-01 -7.22234666e-01 1.18987441e+00 7.59109199e-01
-1.09463084e+00 -3.93274367e-01 -5.00338674e-01 -4.42598552e-01
1.17893312e-02 1.64089173e-01 4.61755961e-01 5.04667759e-01
-8.51493180e-01 4.16868031e-01 -1.38537753e+00 -2.24157244e-01
4.48329955e-01 4.07720655e-01 -5.94940543e-01 -4.81319055e-02
-2.14049309e-01 9.67998147e-01 7.22325817e-02 2.83240438e-01
-6.47079825e-01 -4.48728949e-01 -3.93949062e-01 -4.91967618e-01
1.81529224e-01 -1.14364004e+00 1.25796235e+00 3.25858653e-01
-1.72279263e+00 1.25895357e+00 -4.24976528e-01 2.89200395e-01
1.03809726e+00 -6.83750927e-01 6.05925679e-01 1.85598344e-01
-1.56944171e-01 3.87197673e-01 9.67282832e-01 -1.61194205e+00
-6.61674812e-02 -1.27866232e+00 -2.22488046e-01 2.45815888e-01
3.70774776e-01 1.25203095e-02 -9.36502755e-01 -6.47753060e-01
9.77657557e-01 -1.24264717e+00 -8.90541896e-02 2.31707335e-01
-4.60261613e-01 -3.26259494e-01 7.38897681e-01 -9.23366964e-01
3.03071111e-01 -1.51076937e+00 6.74997389e-01 5.32276571e-01
2.41639420e-01 -1.87400490e-01 1.39513969e-01 2.39204675e-01
1.92336217e-01 -2.30873168e-01 2.76860416e-01 -8.60514462e-01
1.94078639e-01 3.93165231e-01 -4.23300177e-01 6.13235950e-01
-2.82097518e-01 1.23893595e+00 -7.51040518e-01 -5.73559821e-01
3.92505258e-01 1.06909597e+00 -5.70316374e-01 3.77365589e-01
-2.30681986e-01 8.75731289e-01 -5.21447718e-01 1.16619706e+00
6.84569716e-01 -2.82712668e-01 1.68672670e-02 -3.60764533e-01
7.56201670e-02 2.99882859e-01 -1.33360577e+00 2.67381859e+00
-2.92356193e-01 -7.89002031e-02 -3.62506360e-02 1.56753451e-01
5.29564738e-01 6.23358250e-01 5.99649310e-01 7.30684698e-02
3.81247401e-01 2.68528640e-01 -6.05259001e-01 -1.11135468e-01
3.92757170e-02 1.30574675e-02 2.93011427e-01 8.33669603e-01
1.56554118e-01 -6.04535580e-01 -8.90909016e-01 -1.83411360e-01
6.90020740e-01 9.51785505e-01 1.78655013e-01 2.94584513e-01
4.99724522e-02 -1.59679443e-01 -6.13555163e-02 4.41536248e-01
-9.87985507e-02 1.11932826e+00 2.14189850e-03 -6.93996727e-01
-1.08162212e+00 -1.61065400e+00 -2.93894559e-02 8.67642999e-01
-1.34017423e-01 -4.70320880e-01 -5.27133405e-01 -6.43614531e-01
5.31442285e-01 6.01754114e-02 -4.46889162e-01 7.54455388e-01
-8.92255485e-01 1.02102891e-01 1.62460715e-01 1.09458661e+00
2.14022443e-01 -7.62617469e-01 -1.07327795e+00 -2.96069115e-01
-3.60457838e-01 -1.12664378e+00 -6.13100708e-01 -2.02313051e-01
-1.17657185e+00 -1.18413436e+00 -1.06024432e+00 -5.88219285e-01
6.97323978e-01 3.18942636e-01 1.09594464e+00 -1.59868464e-01
-4.75783020e-01 9.05547738e-01 -3.21255565e-01 -6.38520539e-01
2.93322623e-01 1.66212007e-01 5.07726967e-01 -5.78387856e-01
2.72196889e-01 -7.87095189e-01 -8.40693533e-01 5.43813288e-01
2.28948630e-02 -2.29792163e-01 4.24641997e-01 3.90049040e-01
1.05970550e+00 -7.96444058e-01 -3.32277447e-01 -3.45342904e-02
3.90599430e-01 3.92954409e-01 -6.53295398e-01 2.81689346e-01
-2.41734222e-01 -5.08086868e-02 -2.81671137e-01 -3.37362826e-01
-9.59921837e-01 8.69922459e-01 -3.93684447e-01 -1.07940066e+00
-2.26346761e-01 -1.71675876e-01 -2.17673466e-01 -2.04496920e-01
6.93209469e-01 2.68180311e-01 1.77267894e-01 -1.19320750e+00
5.60484648e-01 7.61277735e-01 6.65178895e-01 -7.21056163e-01
8.99463832e-01 9.79280889e-01 1.36193469e-01 -2.04629719e-01
-4.84215558e-01 -6.80755496e-01 -1.92364812e+00 -2.49521747e-01
6.51692450e-01 -8.87023687e-01 -1.39738297e+00 5.95489740e-01
-1.83857274e+00 -1.17900334e-01 -2.19928160e-01 5.04404604e-01
-1.14461482e+00 4.33683366e-01 -3.90106261e-01 -1.02931845e+00
-6.66573644e-01 -1.37682700e+00 2.06568837e+00 -6.87527597e-01
-4.28618044e-01 -4.75303859e-01 1.73560143e-01 3.35599452e-01
-3.75109047e-01 4.37560588e-01 2.13261902e-01 4.52741683e-02
-8.25932205e-01 -5.92680454e-01 1.21415846e-01 -4.69819494e-02
1.02485962e-01 -2.78723508e-01 -1.15884614e+00 -6.06577456e-01
-4.64625619e-02 -1.80363908e-01 2.76652813e-01 3.28484237e-01
1.13691318e+00 5.81373982e-02 -3.83070379e-01 7.90748477e-01
1.05293608e+00 -4.21464711e-01 2.19189987e-01 2.43816301e-01
9.96543407e-01 4.12089407e-01 5.17611086e-01 7.01092184e-01
4.70669121e-01 1.06580341e+00 8.31962943e-01 4.18900877e-01
-3.00003201e-01 -5.73251009e-01 -9.11896601e-02 7.37241089e-01
-1.11600506e+00 4.57345217e-01 -1.24843884e+00 7.50185475e-02
-1.50967252e+00 -6.14381433e-01 -6.92724809e-02 2.10453844e+00
6.97623014e-01 -1.92371160e-01 7.67061830e-01 4.50605065e-01
2.17850253e-01 -1.66847959e-01 -7.35691249e-01 4.41900641e-01
2.01550111e-01 5.87384403e-01 4.21758890e-01 4.83711779e-01
-6.66234016e-01 1.17804909e+00 7.05031061e+00 -4.13630195e-02
-8.29548657e-01 2.67085940e-01 -7.28163481e-01 -5.43670893e-01
1.20160694e-03 -2.33037889e-01 -8.27059627e-01 -1.96176559e-01
-9.98245701e-02 4.02175307e-01 6.03440523e-01 9.54718709e-01
-1.09893359e-01 3.82103413e-01 -1.35957301e+00 1.76062548e+00
1.95207894e-01 -9.12266254e-01 2.17257574e-01 3.50281239e-01
3.72703135e-01 2.80611306e-01 5.25744744e-02 -3.81430119e-01
4.81677726e-02 -9.93998945e-01 1.16532087e+00 6.88995719e-01
9.38304842e-01 -5.38894415e-01 3.52239847e-01 8.57688785e-01
-1.09363472e+00 3.06515634e-01 -1.27418742e-01 -1.96982116e-01
1.96178705e-01 -4.93467599e-03 -1.07267952e+00 4.12139803e-01
1.01976717e+00 6.68287754e-01 -5.62878668e-01 8.48148644e-01
-4.60955143e-01 -1.13847658e-01 -5.94218135e-01 3.23945194e-01
-2.93823719e-01 1.50452256e-01 7.00233698e-01 7.92628527e-01
5.90242930e-02 4.19964403e-01 -5.43235708e-03 9.90073919e-01
2.70598978e-01 -1.54288754e-01 -4.43621933e-01 4.80565488e-01
4.94033217e-01 9.10177588e-01 -7.28934288e-01 1.50589049e-01
2.03104153e-01 1.45774019e+00 2.34752715e-01 2.33483180e-01
-3.13389421e-01 9.04668774e-03 5.44969678e-01 2.43804321e-01
1.60655424e-01 -1.00808573e+00 -5.57898462e-01 -1.34201777e+00
5.38582325e-01 -6.23467565e-01 1.87716505e-03 -1.21749711e+00
-1.29693615e+00 4.92548794e-01 1.52333423e-01 -1.39315140e+00
-4.65934724e-01 -1.11814404e+00 5.58529869e-02 1.49193585e+00
-1.12419426e+00 -1.64471698e+00 -7.83876300e-01 1.15887022e+00
5.02262533e-01 -8.12595561e-02 1.11047542e+00 -4.22734559e-01
2.82982290e-01 4.37541991e-01 -4.60039943e-01 1.54082283e-01
6.40131056e-01 -1.26424670e+00 1.10543597e+00 2.05170646e-01
5.32883883e-01 1.04766333e+00 2.98896313e-01 -7.56869495e-01
-2.20207143e+00 -2.68119454e-01 3.41732740e-01 -1.85640538e+00
-4.08277027e-02 -5.74283481e-01 -4.42110658e-01 1.27712667e+00
-3.31570387e-01 2.64500320e-01 3.79172355e-01 2.29284123e-01
-9.45441842e-01 3.24638993e-01 -1.44339526e+00 2.28677377e-01
1.88961661e+00 -7.01138496e-01 -9.23466623e-01 5.80555797e-01
2.10903719e-01 -1.27290475e+00 -1.08870828e+00 1.52438596e-01
1.47914541e+00 -6.58940852e-01 1.67916262e+00 -7.88362682e-01
-2.78729238e-02 -1.77286729e-01 -2.62780339e-01 -1.17788458e+00
-2.51120985e-01 -5.92830956e-01 -6.18291080e-01 3.72315675e-01
-4.51769143e-01 -3.47801983e-01 1.34268153e+00 3.61881852e-01
3.82070810e-01 -6.53839827e-01 -1.23043251e+00 -9.49424982e-01
8.04160312e-02 -7.72216082e-01 8.19814384e-01 2.72123426e-01
-1.21448867e-01 -7.15280697e-02 -1.35026708e-01 4.43816066e-01
1.19893324e+00 4.43308264e-01 1.25443840e+00 -1.58826923e+00
-1.09523565e-01 -2.43938342e-01 -6.32358968e-01 -1.47873235e+00
2.83056587e-01 -8.38273883e-01 -1.28044188e-01 -1.62183642e+00
2.00859606e-01 -2.81149149e-01 2.10609436e-01 4.84152913e-01
1.77079886e-01 4.13296938e-01 5.49047589e-01 8.65956366e-01
-1.11645967e-01 1.05765268e-01 1.56919575e+00 1.16506271e-01
-1.98805943e-01 4.26007837e-01 4.05044295e-02 8.28402877e-01
4.28424865e-01 -2.47295991e-01 -5.58228344e-02 -8.49744737e-01
1.54173836e-01 2.36096792e-02 9.64223325e-01 -5.16783297e-01
1.89422712e-01 -1.88868314e-01 7.32285321e-01 -1.54422700e+00
8.75713289e-01 -1.25558162e+00 1.61282733e-01 5.62777340e-01
1.80567041e-01 2.65634567e-01 -1.00118525e-01 3.85623842e-01
5.04370809e-01 3.20234179e-01 3.39334995e-01 -5.54808617e-01
-3.41482729e-01 6.36629522e-01 4.75029349e-01 -3.06796551e-01
8.02683890e-01 -5.48147798e-01 2.54738271e-01 -1.45538300e-01
-9.90270853e-01 -8.16572532e-02 7.44451225e-01 6.83636069e-01
1.08076251e+00 -1.45430076e+00 -6.14206195e-01 4.46842074e-01
1.56510502e-01 5.87505162e-01 -1.51303336e-01 5.72576761e-01
-8.62575471e-01 8.27454746e-01 -4.44944531e-01 -1.22833228e+00
-1.56714499e+00 4.58581209e-01 3.48494738e-01 3.35635394e-01
-9.33775008e-01 9.46771502e-01 -4.62292910e-01 -1.05678546e+00
6.64227724e-01 -6.72422707e-01 4.52965766e-01 -3.96598577e-01
4.00549680e-01 6.49189651e-01 3.40695262e-01 -9.18208480e-01
-7.34819949e-01 1.30053461e+00 3.81672949e-01 -3.57279897e-01
1.41606724e+00 1.99076384e-01 -1.63998947e-01 5.39769173e-01
1.10940063e+00 1.79073364e-01 -1.43039298e+00 -3.81975442e-01
-3.35296035e-01 -1.18320501e+00 -8.97254720e-02 -9.29551363e-01
-8.79352748e-01 1.27206326e+00 7.44307399e-01 -6.49136126e-01
7.09922194e-01 6.76256955e-01 7.13047922e-01 6.71071708e-01
1.27349496e+00 -7.16526210e-01 1.94766238e-01 5.05737901e-01
1.68946826e+00 -9.00156200e-01 3.62286210e-01 -6.87904060e-01
-3.42868626e-01 1.25433719e+00 2.98303753e-01 -1.42534748e-01
6.91680253e-01 2.68461615e-01 8.10576454e-02 -3.52629155e-01
-2.82044206e-02 -3.34382094e-02 5.87364316e-01 8.70168149e-01
1.27225742e-01 2.16163307e-01 2.46571317e-01 3.99976224e-01
-6.01928830e-01 2.25980446e-01 -2.49488518e-01 1.44867885e+00
1.07439131e-01 -1.23651922e+00 -9.71747518e-01 -7.92935118e-02
-5.62763885e-02 4.17991757e-01 -7.87069559e-01 7.80071437e-01
-1.19278438e-01 2.19672114e-01 8.16631019e-02 -5.89807570e-01
9.44676518e-01 3.67963403e-01 1.71710348e+00 -7.31289983e-01
-5.33272862e-01 2.49454349e-01 -7.88887382e-01 -9.87758219e-01
-4.02380139e-01 -8.64046872e-01 -1.13630736e+00 -2.46690646e-01
-1.64017752e-01 -6.24255002e-01 9.96625543e-01 7.24239767e-01
3.15794557e-01 -7.73795065e-04 2.75707662e-01 -1.91248918e+00
-1.11003840e+00 -8.86083841e-01 -8.38827848e-01 3.07866544e-01
3.23691905e-01 -1.21039069e+00 5.13794906e-02 -3.19417529e-02] | [6.538787841796875, -0.8778291344642639] |
44a17003-29ff-4719-8426-522f6d7ee0c5 | controlling-styles-in-neural-machine | 2212.08909 | null | https://arxiv.org/abs/2212.08909v2 | https://arxiv.org/pdf/2212.08909v2.pdf | Controlling Styles in Neural Machine Translation with Activation Prompt | Controlling styles in neural machine translation (NMT) has attracted wide attention, as it is crucial for enhancing user experience. Earlier studies on this topic typically concentrate on regulating the level of formality and achieve some progress in this area. However, they still encounter two major challenges. The first is the difficulty in style evaluation. The style comprises various aspects such as lexis, syntax, and others that provide abundant information. Nevertheless, only formality has been thoroughly investigated. The second challenge involves excessive dependence on incremental adjustments, particularly when new styles are necessary. To address both challenges, this paper presents a new benchmark and approach. A multiway stylized machine translation (MSMT) benchmark is introduced, incorporating diverse categories of styles across four linguistic domains. Then, we propose a method named style activation prompt (StyleAP) by retrieving prompts from stylized monolingual corpus, which does not require extra fine-tuning. Experiments show that StyleAP could effectively control the style of translation and achieve remarkable performance. | ['Mingxuan Wang', 'Weiguo Zheng', 'Shanbo Cheng', 'Zewei Sun', 'Yifan Wang'] | 2022-12-17 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 3.46680522e-01 -4.50052649e-01 -2.66753167e-01 -5.31274080e-01
-8.40920091e-01 -9.20259595e-01 7.12444365e-01 -1.70377746e-01
-4.83142823e-01 7.73003757e-01 1.71187714e-01 -4.61006463e-01
3.05092752e-01 -5.28887093e-01 -4.20163661e-01 -5.27527153e-01
6.79873645e-01 4.94763821e-01 4.13321741e-02 -6.03013813e-01
4.28520054e-01 2.94947863e-01 -9.65697348e-01 2.78447658e-01
1.34130919e+00 6.53797567e-01 3.47737402e-01 1.33988217e-01
-7.24574149e-01 8.03069994e-02 -9.52222645e-01 -7.02337980e-01
1.18032530e-01 -7.42309213e-01 -8.28461647e-01 -4.05100472e-02
1.60379857e-01 -3.08906916e-03 2.12154552e-01 1.22112596e+00
6.16636574e-01 -8.86241645e-02 6.84109807e-01 -1.00859237e+00
-8.87186766e-01 8.71650457e-01 -5.57421982e-01 1.07541434e-01
1.70324266e-01 8.49538669e-02 9.32659328e-01 -9.26072121e-01
5.34933209e-01 1.39110327e+00 2.10260615e-01 8.82532418e-01
-1.18213582e+00 -6.66288018e-01 4.17794645e-01 -9.35878232e-03
-1.03561854e+00 -3.43481809e-01 9.58230853e-01 -2.97895461e-01
5.54871678e-01 2.58279860e-01 3.90754968e-01 1.30000114e+00
4.19778116e-02 1.04063070e+00 1.58889651e+00 -7.03788519e-01
1.32251933e-01 3.88892472e-01 7.73191676e-02 3.66452813e-01
2.89719477e-02 -2.12163076e-01 -5.87455988e-01 1.15222149e-01
7.72953451e-01 -4.17998105e-01 -2.02449039e-01 1.48006603e-02
-1.57344151e+00 6.03809416e-01 -1.55196339e-02 5.64277470e-01
1.29109412e-01 -4.30312276e-01 4.77657586e-01 6.54648960e-01
4.09625918e-01 8.33413601e-01 -5.71529210e-01 -5.69671690e-01
-7.80668378e-01 3.01030036e-02 5.99106133e-01 1.21625030e+00
4.66029942e-01 1.52772680e-01 -3.22116822e-01 1.38794971e+00
-7.73699284e-02 7.87376642e-01 6.18419290e-01 -5.18962264e-01
8.13810229e-01 7.38339663e-01 -1.81245208e-02 -6.59415662e-01
-1.66581973e-01 -3.92387420e-01 -8.59952211e-01 -8.23046714e-02
4.82214898e-01 -3.69942605e-01 -6.76262438e-01 1.96481991e+00
2.10827626e-02 -6.14462435e-01 -2.07836270e-01 9.37749565e-01
6.52929127e-01 7.84670234e-01 1.04856929e-02 -3.83923233e-01
1.47963393e+00 -1.08108568e+00 -8.59063625e-01 -3.27170730e-01
5.53337157e-01 -1.35847914e+00 1.90725219e+00 5.33503294e-01
-1.20804811e+00 -5.00310063e-01 -9.83385801e-01 -9.38755050e-02
-2.84692943e-01 4.90105331e-01 5.50312757e-01 6.44537270e-01
-9.26339507e-01 4.11537856e-01 -5.56658924e-01 -4.55998719e-01
6.65074810e-02 3.56019914e-01 -2.71001849e-02 2.73890913e-01
-1.35099924e+00 8.76346409e-01 1.03085846e-01 -1.48889637e-02
-1.85263321e-01 -4.04781014e-01 -5.68811595e-01 -1.75640687e-01
3.02373528e-01 -6.08124673e-01 1.60384095e+00 -1.32230592e+00
-2.16206837e+00 8.61141801e-01 -4.19524968e-01 1.45412788e-01
7.18332350e-01 -3.91894013e-01 -5.35417378e-01 -3.39781880e-01
1.25536531e-01 6.33249879e-01 7.47428536e-01 -1.11428952e+00
-6.76717758e-01 -3.11434120e-01 4.40114662e-02 4.48748112e-01
-7.73468196e-01 4.62230235e-01 -7.32056618e-01 -1.24870992e+00
5.42488769e-02 -1.15459764e+00 2.03430299e-02 -4.64696974e-01
-3.84800494e-01 -4.40429091e-01 4.40806478e-01 -3.48359436e-01
1.67633867e+00 -2.05086613e+00 5.16425490e-01 -1.38255641e-01
-1.35164022e-01 3.53101820e-01 -1.25104576e-01 3.08710992e-01
2.92458445e-01 2.76145786e-01 -2.12973624e-01 -3.07991207e-01
5.06226979e-02 1.85823068e-01 -3.50346655e-01 -1.94487914e-01
3.10504377e-01 1.01176858e+00 -7.94301331e-01 -5.33047974e-01
-2.38018200e-01 3.69019434e-02 -4.70751971e-01 2.85347164e-01
-3.61354560e-01 6.71267569e-01 -5.67070127e-01 7.41278112e-01
4.55509484e-01 -1.46651240e-02 1.58621609e-01 -6.36305064e-02
-3.10694754e-01 5.05152702e-01 -9.90740478e-01 1.82560730e+00
-6.65075004e-01 4.28280294e-01 -1.69789061e-01 -4.97745335e-01
1.07277071e+00 1.70852020e-01 9.88668390e-03 -7.50004232e-01
3.14291686e-01 4.87316310e-01 1.33921579e-01 -3.99277687e-01
6.62583053e-01 -1.15076534e-01 -3.66240621e-01 6.31655335e-01
-1.23202279e-01 -3.09008360e-01 3.15696090e-01 -3.55133675e-02
6.27486885e-01 4.15943414e-01 3.03616673e-01 -5.35231113e-01
6.15088105e-01 1.22958183e-01 8.01434338e-01 3.84641171e-01
-2.94183522e-01 5.24766684e-01 7.08686888e-01 -1.75586134e-01
-8.63682210e-01 -8.74316812e-01 5.93451597e-02 1.41413713e+00
1.62283987e-01 -4.73656923e-01 -1.16567779e+00 -8.04791152e-01
-4.90350783e-01 5.90772212e-01 -5.11764765e-01 4.19096909e-02
-9.92474258e-01 -8.14956903e-01 5.28064609e-01 5.31020999e-01
5.98991752e-01 -1.26997316e+00 -1.75918803e-01 1.59772068e-01
-4.41725641e-01 -1.02497065e+00 -1.03089345e+00 -4.08200510e-02
-1.01131916e+00 -4.56018358e-01 -8.97680461e-01 -9.11808133e-01
6.42411709e-01 3.44954044e-01 1.27408230e+00 -1.62198216e-01
3.72028530e-01 -2.12405860e-01 -3.11122835e-01 -5.63005030e-01
-4.71633822e-01 7.06056178e-01 1.55616343e-01 -5.24285063e-02
5.44355690e-01 -4.38277215e-01 -3.84400964e-01 6.92283988e-01
-7.97672391e-01 3.51501048e-01 9.22087550e-01 7.98519194e-01
5.40473223e-01 -4.24036264e-01 7.33089626e-01 -1.20075333e+00
1.09178317e+00 -6.90631866e-02 -5.62626839e-01 4.15808588e-01
-6.73330903e-01 2.56466985e-01 9.17015910e-01 -5.65084815e-01
-1.33670509e+00 -2.01092064e-01 -8.46209601e-02 9.17886719e-02
-2.43347287e-01 3.87320727e-01 -6.02587759e-01 7.29130208e-02
5.93348742e-01 1.81604654e-01 -2.88373172e-01 -6.31413043e-01
2.71930605e-01 8.15186739e-01 3.34791720e-01 -1.20397401e+00
8.74372780e-01 -5.05031720e-02 -3.21085572e-01 -6.16319835e-01
-8.61414552e-01 -1.13962553e-01 -7.04025328e-01 5.49975671e-02
4.59493518e-01 -5.75964034e-01 -3.44903260e-01 4.79210198e-01
-1.26343715e+00 -3.46655667e-01 1.73741043e-01 4.36824918e-01
-4.17132586e-01 3.63983303e-01 -8.25746298e-01 -3.76701444e-01
-4.69072551e-01 -1.55160916e+00 9.76111233e-01 2.12043390e-01
-6.23317540e-01 -9.37478840e-01 8.13680515e-02 1.79814398e-01
5.83486140e-01 -1.11750215e-01 1.31039011e+00 -3.18881899e-01
-3.44012648e-01 3.06356728e-01 -2.83447802e-01 2.35211521e-01
4.47478861e-01 -5.63808307e-02 -8.28390002e-01 -8.48556161e-02
-5.85480668e-02 -2.97512919e-01 4.09912348e-01 4.39776555e-02
1.00426376e+00 -1.99430168e-01 9.69947577e-02 6.24125004e-01
9.80220973e-01 5.00716567e-01 2.52199054e-01 5.95897913e-01
7.71361947e-01 7.03909695e-01 7.15316117e-01 3.25880535e-02
3.30054194e-01 9.42856908e-01 -2.01555610e-01 -1.61529481e-01
6.92913076e-03 -1.59269959e-01 6.54410362e-01 1.34262383e+00
-2.77966529e-01 -7.74818212e-02 -7.03186393e-01 2.68913984e-01
-1.64151728e+00 -5.20503581e-01 -2.82382723e-02 2.01174736e+00
1.41554034e+00 4.17458296e-01 3.13961685e-01 4.40657884e-02
6.51197255e-01 9.26363915e-02 -5.51726639e-01 -7.74433851e-01
-3.02097648e-01 4.58144657e-02 7.14280689e-03 4.46732700e-01
-8.67130399e-01 1.51435232e+00 5.89523363e+00 9.88460600e-01
-1.37174821e+00 -1.42069524e-02 5.81713140e-01 -7.13873133e-02
-4.78850484e-01 -2.93619752e-01 -1.02363634e+00 6.48438036e-01
5.58358967e-01 -2.22108111e-01 6.51101470e-01 7.45427132e-01
3.87488186e-01 2.05473319e-01 -1.08615780e+00 9.93690848e-01
-2.83732750e-02 -7.50791490e-01 3.59496355e-01 -1.62895679e-01
7.94005752e-01 -2.88922966e-01 2.79971600e-01 4.96460825e-01
-2.65476014e-03 -7.06924260e-01 8.46822143e-01 1.03394344e-01
1.09614098e+00 -8.52891922e-01 5.44113457e-01 2.38337085e-01
-9.58068192e-01 3.41732144e-01 -2.58513600e-01 -5.60609698e-02
1.22896120e-01 4.73512322e-01 -4.40916002e-01 3.49406034e-01
4.98174846e-01 4.68054622e-01 -7.70729721e-01 5.95916569e-01
-5.55974543e-01 8.61216843e-01 -7.18673617e-02 -3.87715906e-01
3.11043918e-01 -4.65522528e-01 5.37553906e-01 1.53339005e+00
3.20135772e-01 -3.59515212e-02 9.48326364e-02 6.90398097e-01
-1.07578844e-01 5.89961231e-01 -3.86892200e-01 -1.65849969e-01
4.80390251e-01 1.30286002e+00 -7.79625773e-01 -3.24795157e-01
-4.36164021e-01 1.12468445e+00 3.88199240e-01 4.08552051e-01
-7.16950357e-01 -3.09165746e-01 7.37651706e-01 -1.01241633e-01
-2.42539793e-01 -3.65709007e-01 -8.00561309e-01 -1.54473031e+00
2.88005859e-01 -1.47105849e+00 1.33350313e-01 -3.68463337e-01
-1.15094209e+00 1.01571012e+00 -1.18173994e-01 -1.36360121e+00
-1.83346733e-01 -6.79432333e-01 -5.28011560e-01 1.07412577e+00
-1.29318690e+00 -1.04796731e+00 -9.14269686e-02 3.90143067e-01
1.07534039e+00 -3.23292255e-01 7.61082768e-01 3.81619394e-01
-8.22118342e-01 1.06011868e+00 -7.45662004e-02 -1.76537540e-02
1.29616427e+00 -1.55928266e+00 7.46026576e-01 8.52823794e-01
4.70617674e-02 1.00292397e+00 7.15012372e-01 -4.82716411e-01
-1.29217827e+00 -9.48188126e-01 1.23691070e+00 -5.98726392e-01
6.03890121e-01 -5.19539237e-01 -8.93176854e-01 3.53402078e-01
4.70076770e-01 -6.06415808e-01 6.58406198e-01 2.73773730e-01
-2.47945279e-01 -2.07524467e-02 -5.91912568e-01 1.22847342e+00
1.33373451e+00 -2.97883958e-01 -5.72352707e-01 -1.36490795e-03
6.77268922e-01 -4.80901837e-01 -5.60681403e-01 4.25801933e-01
4.20514196e-01 -8.13542366e-01 4.39561933e-01 -5.78665376e-01
5.13829708e-01 -3.61906916e-01 -2.36583222e-02 -1.79503727e+00
-3.84304076e-01 -1.04768670e+00 1.40230969e-01 1.64585304e+00
6.40694559e-01 -4.03202504e-01 5.68915188e-01 3.81899536e-01
-2.10924998e-01 -1.01182032e+00 -4.10579771e-01 -7.70397007e-01
4.05141771e-01 -1.62823737e-01 8.89624715e-01 1.01175249e+00
2.74166204e-02 9.33994353e-01 -4.48581129e-01 -3.39779884e-01
2.12288871e-01 2.97014058e-01 8.02056193e-01 -1.06570244e+00
-3.55146289e-01 -9.76927400e-01 1.94292128e-01 -1.19791555e+00
4.59731407e-02 -7.97628880e-01 -5.79473935e-02 -1.27832663e+00
-3.21981609e-02 -4.50998276e-01 -2.29660958e-01 3.90152663e-01
-6.34072721e-01 1.51343733e-01 1.90385684e-01 3.01549941e-01
-4.88552123e-01 5.69308579e-01 1.57085216e+00 -8.90328437e-02
-3.70147824e-01 1.97344229e-01 -9.70468402e-01 7.36449242e-01
1.24259138e+00 -1.98661730e-01 -5.05197287e-01 -9.16130304e-01
3.72653872e-01 -8.23135599e-02 -3.80701661e-01 -5.79428136e-01
-8.07261020e-02 -5.34527540e-01 1.75189167e-01 -4.50116426e-01
-2.56723892e-02 -5.01470983e-01 -3.20520997e-01 5.66034690e-02
-4.45534050e-01 7.63182700e-01 1.73366427e-01 7.56109282e-02
-2.96160519e-01 -1.44982770e-01 8.29127371e-01 -9.06445161e-02
-5.19361079e-01 2.68644780e-01 -3.45555007e-01 2.81822085e-01
6.30333364e-01 1.09609984e-01 -1.15509905e-01 -8.88303369e-02
-3.45522732e-01 1.61663577e-01 6.63747132e-01 8.19719851e-01
1.72754630e-01 -1.53208649e+00 -6.25617504e-01 2.89697468e-01
3.55275929e-01 -1.94173187e-01 -2.58910805e-01 6.24572158e-01
-3.29553068e-01 4.50798631e-01 -2.65843093e-01 -4.50121224e-01
-1.07386994e+00 4.82948303e-01 7.31389597e-02 -3.07454020e-01
-2.56992161e-01 6.28017247e-01 3.41367394e-01 -5.37630618e-01
2.80744433e-01 -3.91688019e-01 -3.39317590e-01 9.77266729e-02
5.11877477e-01 2.20930606e-01 1.47076339e-01 -4.67113256e-01
-5.90390526e-04 8.00149918e-01 -4.14571136e-01 -3.74114662e-01
8.54976773e-01 -3.01897347e-01 -3.01223308e-01 6.91248953e-01
8.28203857e-01 3.21141988e-01 -1.03019285e+00 -5.63187718e-01
3.92045945e-01 -3.45953345e-01 -3.32425773e-01 -1.04727948e+00
-7.71593928e-01 1.13321137e+00 2.34812930e-01 1.11599602e-01
1.16956210e+00 -4.14834201e-01 1.08415186e+00 5.42309999e-01
5.12464225e-01 -1.40034437e+00 -3.38101462e-02 9.22268987e-01
8.72857511e-01 -1.29412067e+00 -4.49471802e-01 -3.70253623e-01
-8.91775727e-01 1.14908779e+00 9.88218009e-01 2.93882906e-01
2.40135834e-01 1.39691472e-01 6.12748146e-01 2.76221365e-01
-5.78798115e-01 8.39330629e-03 4.24421102e-01 2.26029575e-01
9.90548313e-01 1.83269799e-01 -8.11198533e-01 6.99590802e-01
-5.35207450e-01 -2.64158696e-01 1.37937084e-01 6.97333753e-01
-3.19094956e-01 -1.74059546e+00 -3.32901895e-01 2.11997494e-01
-6.72440946e-01 -2.89754540e-01 -7.73100197e-01 5.18163502e-01
-2.52910294e-02 9.43681002e-01 -3.81506205e-01 -3.55389118e-01
5.65327644e-01 3.31325941e-02 5.24616301e-01 -7.20540166e-01
-7.06210315e-01 3.55971456e-01 7.23916888e-02 -3.76472652e-01
-2.12001041e-01 -6.41791523e-01 -8.13087285e-01 -3.33069503e-01
-1.48210749e-01 4.39315110e-01 5.86508334e-01 9.90136087e-01
1.57465950e-01 4.06501859e-01 7.10986435e-01 -6.16424799e-01
-6.23966932e-01 -1.11682963e+00 -2.81869143e-01 3.20378572e-01
-1.56447082e-03 -6.07107401e-01 -1.01880357e-01 1.38554588e-01] | [11.641361236572266, 10.048783302307129] |
92a7dccb-2893-475e-9ee0-6f01e2ab86bc | supercon-supervised-contrastive-learning-for | 2202.05685 | null | https://arxiv.org/abs/2202.05685v1 | https://arxiv.org/pdf/2202.05685v1.pdf | SuperCon: Supervised Contrastive Learning for Imbalanced Skin Lesion Classification | Convolutional neural networks (CNNs) have achieved great success in skin lesion classification. A balanced dataset is required to train a good model. However, due to the appearance of different skin lesions in practice, severe or even deadliest skin lesion types (e.g., melanoma) naturally have quite small amount represented in a dataset. In that, classification performance degradation occurs widely, it is significantly important to have CNNs that work well on class imbalanced skin lesion image dataset. In this paper, we propose SuperCon, a two-stage training strategy to overcome the class imbalance problem on skin lesion classification. It contains two stages: (i) representation training that tries to learn a feature representation that closely aligned among intra-classes and distantly apart from inter-classes, and (ii) classifier fine-tuning that aims to learn a classifier that correctly predict the label based on the learnt representations. In the experimental evaluation, extensive comparisons have been made among our approach and other existing approaches on skin lesion benchmark datasets. The results show that our two-stage training strategy effectively addresses the class imbalance classification problem, and significantly improves existing works in terms of F1-score and AUC score, resulting in state-of-the-art performance. | ['J. Morris Chang', 'Di Zhuang', 'Keyu Chen'] | 2022-02-11 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 5.62448740e-01 -2.30675623e-01 -5.18255472e-01 -5.54223418e-01
-6.31081164e-01 -3.51328254e-02 2.13475779e-01 3.59976590e-01
-3.27336699e-01 7.20219493e-01 -1.01644963e-01 -1.82585001e-01
-3.28114897e-01 -9.55016613e-01 -2.25627750e-01 -9.27825212e-01
1.93470702e-01 5.13834395e-02 1.82128400e-01 -1.05664946e-01
2.52138674e-01 4.43789095e-01 -1.52461624e+00 6.69495165e-01
1.01640475e+00 1.17865431e+00 -2.26593882e-01 5.35131335e-01
-3.19213986e-01 9.76112843e-01 -5.53397477e-01 -2.62586862e-01
1.04970392e-02 -3.96494538e-01 -8.66203368e-01 7.54950494e-02
2.98090070e-01 -3.94191258e-02 -1.19972788e-01 1.01179087e+00
6.37170553e-01 -4.05092329e-01 7.83352792e-01 -1.07913768e+00
-3.27705234e-01 1.15236834e-01 -1.01928091e+00 2.73198158e-01
-1.84172541e-01 1.53578515e-03 5.43497860e-01 -4.87700731e-01
4.12531585e-01 7.63537407e-01 1.01037657e+00 7.45461881e-01
-7.74509728e-01 -6.85936689e-01 -7.43864477e-02 3.18465203e-01
-1.35601604e+00 -8.00557360e-02 6.20413423e-01 -4.03191686e-01
5.20855010e-01 5.72619379e-01 5.78873336e-01 9.12772298e-01
3.03387463e-01 7.81889677e-01 1.19431937e+00 -4.55602676e-01
8.97063166e-02 4.33034785e-02 3.06615531e-01 4.70500439e-01
2.41318002e-01 -2.10328817e-01 -3.02587003e-01 -9.10474046e-05
5.91398835e-01 7.18381777e-02 -1.73041195e-01 -1.48723841e-01
-7.91846693e-01 7.19887972e-01 8.83643270e-01 4.59574193e-01
-5.37732720e-01 -1.13612581e-02 8.00712228e-01 1.54058903e-01
5.07623672e-01 6.01908155e-02 -3.61638576e-01 1.00711271e-01
-6.62267327e-01 7.51989260e-02 3.94766480e-01 2.21879572e-01
3.10308397e-01 -3.38934064e-01 -3.14316958e-01 1.25976193e+00
-1.25504928e-02 -4.98817302e-02 8.65836442e-01 -4.39244062e-02
4.57777947e-01 1.05127156e+00 -4.03742254e-01 -1.01870084e+00
-6.17892206e-01 -7.12548614e-01 -1.47850275e+00 1.03221193e-01
4.42146480e-01 5.05341291e-02 -1.24927950e+00 1.50279355e+00
3.73369694e-01 1.24005549e-01 8.65725055e-02 8.89457822e-01
9.86212254e-01 3.82338911e-01 3.65787357e-01 -1.71988830e-01
1.33939898e+00 -8.74840558e-01 -7.36693263e-01 -1.53204933e-01
7.48902082e-01 -6.33399487e-01 8.10577095e-01 4.22866166e-01
-6.33280277e-01 -5.11338472e-01 -1.06650186e+00 3.01614162e-02
-4.86085415e-01 3.09264660e-01 8.80339921e-01 6.58704281e-01
-7.98080683e-01 5.01791954e-01 -5.69745064e-01 -5.74445188e-01
1.01552641e+00 4.46402073e-01 -6.78753793e-01 -3.12083095e-01
-1.12341881e+00 7.48916924e-01 4.07020509e-01 3.17647070e-01
-6.49818480e-01 -7.05277085e-01 -4.80620503e-01 -1.41991630e-01
1.98126398e-03 -3.72497320e-01 1.08503187e+00 -1.56237984e+00
-1.03159916e+00 1.11241639e+00 1.44727573e-01 -2.48018607e-01
4.97265100e-01 1.77562371e-01 -5.32801867e-01 -7.61912018e-02
-1.74829289e-01 3.70167017e-01 5.29505134e-01 -9.81803477e-01
-7.16242313e-01 -6.56313002e-01 -9.09283832e-02 2.62737244e-01
-6.66506946e-01 -5.16204201e-02 -1.89283401e-01 -3.70271385e-01
1.70933232e-01 -5.47330976e-01 -3.78188282e-01 2.88901955e-01
-6.03460431e-01 -4.47475195e-01 6.83037102e-01 -6.86745226e-01
1.33238482e+00 -1.98475361e+00 -1.88610882e-01 3.42209846e-01
2.12091282e-01 8.06719244e-01 -1.20178632e-01 3.39753509e-01
-5.58162987e-01 2.70035505e-01 -1.51107222e-01 3.98349576e-02
-3.62675548e-01 9.39339865e-03 2.40231305e-01 4.99862522e-01
5.69864988e-01 6.78682983e-01 -6.20412290e-01 -5.12354910e-01
1.59527004e-01 4.87497628e-01 -4.66947258e-02 2.13065833e-01
6.70008585e-02 1.97500840e-01 -4.50583279e-01 9.85073388e-01
9.53620136e-01 -3.59905660e-01 2.11913317e-01 -3.75796169e-01
1.41224325e-01 -3.16659696e-02 -1.00412226e+00 1.43587255e+00
-3.67557675e-01 4.10993487e-01 -2.57406682e-01 -1.18643415e+00
8.77714992e-01 1.64161727e-01 4.06774998e-01 -9.09238398e-01
2.90849239e-01 3.45418751e-01 2.14590579e-01 -9.70878184e-01
-4.40698154e-02 -2.52489030e-01 2.49656588e-01 2.39392393e-03
-1.42038941e-01 3.43611807e-01 1.93621367e-01 -2.74795026e-01
1.13682163e+00 -3.17736685e-01 6.30342424e-01 -1.22658886e-01
7.63544798e-01 2.91054789e-02 7.23713100e-01 2.49297544e-01
-4.08875734e-01 6.20890856e-01 6.85587943e-01 -8.80288720e-01
-8.57763469e-01 -6.47857070e-01 -5.59215307e-01 6.67003274e-01
9.47480276e-02 3.01182503e-03 -9.22741771e-01 -9.36506510e-01
-6.34200945e-02 -1.27752041e-02 -1.13763773e+00 -2.90865481e-01
-3.60273749e-01 -1.43975675e+00 7.55234957e-01 7.26946235e-01
8.43937457e-01 -1.00347865e+00 -2.07290798e-01 -6.50165007e-02
9.03521292e-03 -4.92539942e-01 6.58380687e-02 3.41175735e-01
-7.42118061e-01 -1.57934403e+00 -7.34164476e-01 -1.05686331e+00
1.01453185e+00 2.69506097e-01 8.32189918e-01 4.65435803e-01
-8.43543291e-01 -5.20949662e-01 -4.20549512e-01 -5.48830926e-01
-1.62639782e-01 2.33689576e-01 -4.15479422e-01 3.22876692e-01
7.50247419e-01 -1.98232293e-01 -7.27789700e-01 1.89365342e-01
-1.26962900e+00 1.95445083e-02 9.70039666e-01 1.07950914e+00
6.06555641e-01 3.32186520e-01 9.03873444e-01 -1.11180770e+00
5.03430724e-01 -6.02523148e-01 7.27531360e-03 4.53328341e-01
-3.17071319e-01 -4.70859110e-01 6.90922201e-01 -2.15447828e-01
-8.56328785e-01 1.06672965e-01 -4.26493943e-01 6.96083009e-02
-3.44395250e-01 6.74455643e-01 -1.19386427e-01 -2.65796483e-01
1.01497948e+00 1.57858491e-01 2.51110464e-01 -3.30780894e-01
-2.73007274e-01 1.21435285e+00 1.91351756e-01 -3.78754213e-02
5.41055977e-01 3.27353030e-01 8.83502737e-02 -6.81339145e-01
-1.06593227e+00 -5.33269107e-01 -5.99747598e-01 -1.80313930e-01
7.24235892e-01 -8.37915421e-01 -4.43816036e-01 1.12307334e+00
-6.71496511e-01 -1.97250068e-01 -1.13719985e-01 1.81985229e-01
-1.36379227e-01 2.28920624e-01 -4.73548383e-01 -6.60859585e-01
-5.41926682e-01 -9.66630518e-01 8.78087044e-01 7.69638121e-01
-4.56430353e-02 -1.02807426e+00 7.75707737e-02 4.30457175e-01
6.46370232e-01 6.44326389e-01 1.05567181e+00 -6.74693942e-01
-1.06147043e-01 -6.35695279e-01 -5.64483702e-01 5.78721702e-01
3.56384605e-01 2.21620202e-01 -1.19045079e+00 -4.14745510e-01
-3.43185008e-01 -6.84298575e-01 1.00701737e+00 3.06795329e-01
1.86956084e+00 4.63804640e-02 -4.62394357e-01 7.36016214e-01
1.72697556e+00 1.06803633e-01 1.07666266e+00 3.84797126e-01
4.62635785e-01 6.60361707e-01 6.58979535e-01 2.75864284e-02
1.84924439e-01 3.75091374e-01 6.74949527e-01 -6.59517705e-01
-4.12728578e-01 -1.48739487e-01 -3.39607805e-01 6.71520829e-01
-9.75500494e-02 -2.09899724e-01 -9.06128407e-01 6.88705862e-01
-1.62663996e+00 -6.05100513e-01 -2.28406236e-01 2.03233624e+00
9.94670451e-01 -1.21161854e-03 3.15752104e-02 6.70002460e-01
9.20380116e-01 7.87839815e-02 -6.95923626e-01 -3.56961846e-01
-7.13716000e-02 3.88171434e-01 3.09843749e-01 -3.72517183e-02
-1.43664765e+00 5.35656810e-01 5.91639853e+00 1.26736069e+00
-1.53251421e+00 -1.10823095e-01 1.13962233e+00 1.92230225e-01
1.05797425e-01 -5.51148534e-01 -5.16473889e-01 5.71954727e-01
4.17516589e-01 1.11973502e-01 -9.28672701e-02 6.52622402e-01
-7.14363307e-02 -6.12020977e-02 -7.36047506e-01 9.12431300e-01
1.83877125e-01 -1.20232308e+00 9.98667348e-03 -2.97770295e-02
6.76745296e-01 -2.28405759e-01 1.08492941e-01 1.49181679e-01
-2.49314606e-01 -1.52778113e+00 -7.19373077e-02 4.85450268e-01
1.12497401e+00 -9.21834886e-01 1.37164509e+00 3.47978711e-01
-8.88036609e-01 -1.64547607e-01 -5.53827286e-01 1.78545690e-03
-5.47641754e-01 9.27235484e-01 -8.14112782e-01 7.63540089e-01
6.21791005e-01 8.01083565e-01 -7.60552764e-01 1.47151482e+00
-1.26833871e-01 5.85407853e-01 3.86056900e-02 -2.91558415e-01
1.30117595e-01 2.48367295e-01 -1.34530723e-01 1.08621490e+00
2.30271164e-02 1.15390960e-02 3.18969898e-02 3.33350003e-01
-8.01475346e-02 4.82155412e-01 -3.40119123e-01 7.22532868e-02
1.20862521e-01 1.56074476e+00 -6.42537773e-01 -5.97096123e-02
-6.39883354e-02 6.29924536e-01 3.65657955e-01 -1.00116292e-02
-7.08289266e-01 -6.51595831e-01 6.07482433e-01 1.75771341e-01
-9.45525467e-02 6.11103773e-01 -4.17708665e-01 -8.46376359e-01
1.90924898e-01 -9.00044799e-01 5.73084116e-01 -3.45589310e-01
-1.61317813e+00 7.35719442e-01 -6.94091797e-01 -1.32129753e+00
1.05762877e-01 -7.83469617e-01 -8.91027331e-01 8.63724947e-01
-1.98372722e+00 -1.43059540e+00 -9.69325483e-01 4.59850430e-01
4.01074290e-01 -1.50025353e-01 1.03239429e+00 4.87472475e-01
-9.38499153e-01 9.36889648e-01 4.33339477e-02 3.84597510e-01
7.65322685e-01 -1.16466784e+00 -5.18430993e-02 4.60119963e-01
-4.60202187e-01 1.95482597e-01 -1.50916930e-02 -3.84528905e-01
-9.01751518e-01 -1.26631117e+00 8.30728352e-01 6.55828714e-02
2.66153574e-01 -3.06904633e-02 -8.71665776e-01 2.20637068e-01
-9.42729320e-03 4.13634777e-01 1.21790445e+00 1.25368372e-01
-4.12341833e-01 -5.20023942e-01 -1.46999443e+00 1.70574963e-01
7.48066545e-01 -1.55133158e-01 -8.56306627e-02 6.13963008e-01
1.81808710e-01 -5.13701200e-01 -8.86082709e-01 8.17962050e-01
6.29286766e-01 -1.09004295e+00 7.93619573e-01 -1.00962353e+00
8.82138669e-01 -2.94479370e-01 1.17719332e-02 -1.36373007e+00
-3.27057421e-01 2.27288574e-01 2.34304681e-01 1.06044137e+00
2.07771510e-01 -4.95279491e-01 9.77166474e-01 1.01663232e-01
-6.87826099e-03 -1.50824142e+00 -9.00823057e-01 -3.54083776e-01
2.29729488e-01 8.07932988e-02 6.42218828e-01 1.26480079e+00
-2.57423759e-01 -2.67019887e-02 -2.01552317e-01 5.46182580e-02
4.90276158e-01 -8.36318806e-02 5.55154026e-01 -1.23183084e+00
1.74448475e-01 -3.57427716e-01 -8.56384218e-01 -3.17173600e-01
-2.45323583e-01 -8.58147025e-01 -8.06422979e-02 -1.60609555e+00
5.65500379e-01 -8.17611277e-01 -7.99684286e-01 7.48868644e-01
-4.13186431e-01 6.58642113e-01 -2.24122018e-01 2.04509515e-02
-4.87578750e-01 7.38982037e-02 1.40868211e+00 -3.15749049e-01
1.61167383e-01 1.19745553e-01 -9.80692327e-01 6.76365376e-01
9.45867896e-01 -3.09953958e-01 -3.06787163e-01 -4.18742776e-01
1.01961203e-01 -1.62128150e-01 3.34514797e-01 -1.27789164e+00
2.56020963e-01 -2.54395634e-01 9.77927864e-01 -5.49566746e-01
6.40477613e-03 -5.73348820e-01 -8.26925337e-02 7.27641940e-01
-4.77838635e-01 -4.51258212e-01 1.15012921e-01 3.06596547e-01
-5.11364102e-01 -3.53783071e-01 1.18493772e+00 -3.63760330e-02
-6.79993153e-01 5.23933232e-01 1.93937588e-02 -2.19886243e-01
1.38160908e+00 -3.09446007e-01 -6.26512289e-01 1.47449002e-01
-5.41450620e-01 1.17180966e-01 2.46643513e-01 4.16246742e-01
4.75860029e-01 -1.51680386e+00 -8.74931931e-01 3.34953576e-01
4.79551077e-01 1.91710263e-01 6.70234621e-01 7.88450837e-01
-8.41484010e-01 1.80107638e-01 -5.78686237e-01 -7.43269622e-01
-1.34735131e+00 6.19176365e-02 6.98588431e-01 -5.78327000e-01
-8.20869952e-02 1.16371524e+00 3.61442417e-02 -5.87958097e-01
3.62741977e-01 -1.07566997e-01 -5.82086623e-01 1.16776578e-01
8.93650472e-01 1.65806413e-01 5.09665012e-01 -4.19776052e-01
-4.14455175e-01 7.12491035e-01 -5.17749190e-01 8.56963575e-01
1.38172317e+00 4.07755703e-01 -4.40171093e-01 1.17516175e-01
1.38227808e+00 -4.11719263e-01 -8.81905913e-01 -3.70947272e-02
-1.99688181e-01 -6.80965364e-01 7.49056786e-02 -1.29860246e+00
-1.36672568e+00 1.03097570e+00 1.15037954e+00 3.33780199e-01
1.54393268e+00 -4.81170177e-01 8.78195882e-01 -6.41199702e-04
6.54330030e-02 -1.10769105e+00 1.53284132e-01 1.58088684e-01
6.27866209e-01 -1.28509521e+00 -5.65002300e-02 -7.38159835e-01
-3.92344415e-01 1.26646340e+00 9.94272590e-01 -1.68535322e-01
5.77126205e-01 2.02155188e-01 3.80204290e-01 -1.43803939e-01
-6.43601477e-01 -1.03724927e-01 2.12151945e-01 8.76665235e-01
5.39107382e-01 2.04724833e-01 -5.67981422e-01 7.12371111e-01
1.92808270e-01 3.77500981e-01 2.96554029e-01 7.80616641e-01
-4.30308431e-01 -1.23289394e+00 -1.55230209e-01 9.63170767e-01
-6.84038043e-01 1.89344987e-01 -5.30481875e-01 8.37426424e-01
6.00603998e-01 6.54051423e-01 1.88505515e-01 -5.38874745e-01
4.12095696e-01 -1.34048164e-01 2.91454554e-01 -4.45674270e-01
-7.57782102e-01 -3.75615805e-01 5.34486659e-02 -5.18930614e-01
-4.07153070e-01 -2.13758752e-01 -9.01288927e-01 -1.23303689e-01
-4.43633318e-01 -1.40405819e-01 8.34315360e-01 7.68342912e-01
8.07219595e-02 8.52960169e-01 8.93138528e-01 -1.72962844e-01
-7.12378979e-01 -1.16895938e+00 -6.59290135e-01 5.33808291e-01
2.68527567e-01 -5.76821864e-01 -2.32037008e-01 -2.95252830e-01] | [15.575234413146973, -2.908388137817383] |
285e6738-0533-4ca7-bced-48fceeb9269e | dynamic-few-shot-visual-learning-without | 1804.09458 | null | http://arxiv.org/abs/1804.09458v1 | http://arxiv.org/pdf/1804.09458v1.pdf | Dynamic Few-Shot Visual Learning without Forgetting | The human visual system has the remarkably ability to be able to effortlessly
learn novel concepts from only a few examples. Mimicking the same behavior on
machine learning vision systems is an interesting and very challenging research
problem with many practical advantages on real world vision applications. In
this context, the goal of our work is to devise a few-shot visual learning
system that during test time it will be able to efficiently learn novel
categories from only a few training data while at the same time it will not
forget the initial categories on which it was trained (here called base
categories). To achieve that goal we propose (a) to extend an object
recognition system with an attention based few-shot classification weight
generator, and (b) to redesign the classifier of a ConvNet model as the cosine
similarity function between feature representations and classification weight
vectors. The latter, apart from unifying the recognition of both novel and base
categories, it also leads to feature representations that generalize better on
"unseen" categories. We extensively evaluate our approach on Mini-ImageNet
where we manage to improve the prior state-of-the-art on few-shot recognition
(i.e., we achieve 56.20% and 73.00% on the 1-shot and 5-shot settings
respectively) while at the same time we do not sacrifice any accuracy on the
base categories, which is a characteristic that most prior approaches lack.
Finally, we apply our approach on the recently introduced few-shot benchmark of
Bharath and Girshick [4] where we also achieve state-of-the-art results. The
code and models of our paper will be published on:
https://github.com/gidariss/FewShotWithoutForgetting | ['Spyros Gidaris', 'Nikos Komodakis'] | 2018-04-25 | dynamic-few-shot-visual-learning-without-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Gidaris_Dynamic_Few-Shot_Visual_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Gidaris_Dynamic_Few-Shot_Visual_CVPR_2018_paper.pdf | cvpr-2018-6 | ['novel-concepts'] | ['reasoning'] | [ 1.86810315e-01 -9.67711657e-02 1.67538628e-01 -3.30640852e-01
-3.16469550e-01 -3.59529644e-01 9.54432905e-01 1.43571664e-02
-6.14830196e-01 5.05388618e-01 -2.44778410e-01 -2.60940902e-02
-1.31631293e-03 -7.78286695e-01 -6.92628086e-01 -7.03052402e-01
1.20417468e-01 2.37409100e-01 6.08907342e-01 -3.22646081e-01
1.87970802e-01 3.85743558e-01 -2.15892768e+00 7.23329037e-02
6.37989759e-01 1.07456946e+00 3.33796859e-01 8.05187941e-01
1.63223997e-01 7.45600700e-01 -4.13814217e-01 -3.95518064e-01
2.84664243e-01 -4.57170278e-01 -6.82893157e-01 1.89574704e-01
6.43050194e-01 -8.45548883e-02 -3.77364159e-01 1.12773275e+00
3.98967355e-01 5.49386919e-01 7.28873789e-01 -1.19580197e+00
-8.54590178e-01 3.19891274e-01 -4.10497457e-01 4.45153505e-01
-7.05740601e-02 5.31257033e-01 9.27578092e-01 -1.16643417e+00
7.25122452e-01 7.56024420e-01 4.17953789e-01 9.61328328e-01
-1.07759762e+00 -4.42038566e-01 8.27927068e-02 7.73014426e-01
-1.28165853e+00 -6.38024151e-01 5.91059625e-01 -4.89658386e-01
1.03634119e+00 9.15133879e-02 5.45380592e-01 1.36474812e+00
3.87497284e-02 5.57884336e-01 9.01462436e-01 -5.57833910e-01
5.45519471e-01 2.52575010e-01 5.91011584e-01 6.27057135e-01
2.52706051e-01 2.04350099e-01 -1.67623371e-01 2.50130802e-01
2.40230873e-01 5.49083889e-01 -4.07851398e-01 -6.87296569e-01
-8.63791168e-01 9.53906655e-01 6.53479993e-01 7.23048627e-01
-2.86925167e-01 1.66553557e-01 4.19439077e-01 4.27948773e-01
2.21313044e-01 2.86981285e-01 -1.64550409e-01 -8.24703872e-02
-8.17603707e-01 4.76465188e-02 8.06804419e-01 9.23996925e-01
8.63476813e-01 1.89928964e-01 -3.37814718e-01 1.04167235e+00
-1.77177027e-01 2.84883589e-01 9.68697488e-01 -4.73082483e-01
-1.40009463e-01 2.88638622e-01 -1.58206165e-01 -5.91390491e-01
-1.77827969e-01 -5.95656574e-01 -8.81489038e-01 3.91263396e-01
2.41892412e-01 2.08525248e-02 -1.41271889e+00 1.69004297e+00
-1.35475714e-02 5.39504647e-01 6.20310865e-02 7.74566829e-01
8.32902133e-01 6.06585503e-01 5.93025126e-02 -1.85628936e-01
1.34059072e+00 -1.07264948e+00 -3.01681250e-01 -3.91744137e-01
4.43561703e-01 -5.79162955e-01 1.17476845e+00 3.60794663e-01
-7.15340674e-01 -9.08089936e-01 -1.17111790e+00 2.19297856e-01
-7.03465283e-01 -1.15307808e-01 5.40321052e-01 6.36501491e-01
-9.82542336e-01 7.86683142e-01 -8.15102994e-01 -8.13299596e-01
6.53852344e-01 -5.65500297e-02 -2.82549232e-01 -3.80781502e-01
-8.51165771e-01 1.12833798e+00 5.79257667e-01 -4.28987056e-01
-1.22277761e+00 -8.03785503e-01 -8.37335348e-01 3.37857455e-01
5.63493729e-01 -6.23415291e-01 1.28143919e+00 -8.85785639e-01
-1.28416157e+00 8.47463429e-01 1.26979664e-01 -6.77017689e-01
3.87610972e-01 -8.70213378e-03 -4.47594553e-01 6.47697672e-02
-1.78515411e-03 6.63610876e-01 8.87633085e-01 -1.05667841e+00
-7.86120832e-01 -1.49636522e-01 2.14157160e-02 -1.21868744e-01
-6.88843608e-01 -2.86610574e-01 -2.22279847e-01 -5.86389363e-01
-2.24024802e-01 -8.30319941e-01 -7.20209405e-02 1.29119247e-01
5.02828322e-03 -3.78187507e-01 8.18867147e-01 -1.34214506e-01
7.67586172e-01 -2.25887752e+00 1.66789755e-01 -2.47324213e-01
2.75203556e-01 7.74543166e-01 -2.96754777e-01 2.76504785e-01
-2.60453641e-01 -1.80465952e-01 -2.84914881e-01 -2.78310835e-01
-2.58755181e-02 2.31287599e-01 -4.25085127e-01 5.65282226e-01
2.54092723e-01 1.00488019e+00 -9.11694288e-01 6.02921732e-02
5.04796445e-01 3.02465200e-01 -4.76329595e-01 1.85344607e-01
-4.80515659e-02 -6.94346353e-02 1.16735488e-01 5.47485709e-01
4.98434365e-01 -2.58256763e-01 -2.65815735e-01 5.43470792e-02
9.60736070e-04 -2.09649846e-01 -1.10572946e+00 1.62336767e+00
-4.31405634e-01 6.02171302e-01 -5.73531747e-01 -1.41421556e+00
9.55458164e-01 1.43610582e-01 1.13094598e-02 -6.00086987e-01
3.72424632e-01 7.68676028e-02 2.75707692e-01 -4.06996220e-01
2.29641825e-01 -5.95650375e-01 8.25782120e-02 2.93485492e-01
8.06002080e-01 5.04332408e-02 3.93292904e-01 5.85080832e-02
1.24650645e+00 -2.61916608e-01 5.81681311e-01 -9.75063741e-02
4.10986006e-01 -1.75099954e-01 3.42985809e-01 1.10832834e+00
-3.20152313e-01 7.25888133e-01 1.48442239e-01 -5.72733462e-01
-1.16624284e+00 -1.17480659e+00 -2.16499537e-01 1.20411325e+00
-8.81279036e-02 -1.00246280e-01 -4.28665757e-01 -6.87109411e-01
-1.19308122e-01 1.14489222e+00 -9.00987744e-01 -6.03564680e-01
-9.80535746e-02 -5.32686472e-01 3.35071713e-01 4.67722356e-01
4.10810441e-01 -1.22810292e+00 -9.13938284e-01 1.19931541e-01
3.61020297e-01 -8.30796897e-01 -1.66098550e-01 4.24639016e-01
-6.29452407e-01 -1.26821399e+00 -1.01806641e+00 -7.42094159e-01
4.10863340e-01 5.70112944e-01 9.12625194e-01 4.70533185e-02
-6.95747495e-01 4.72350746e-01 -7.74025261e-01 -5.89025617e-01
-1.86205477e-01 -3.68873887e-02 1.34530246e-01 3.16834807e-01
7.23612130e-01 -7.21777320e-01 -4.73844558e-01 1.55894950e-01
-1.02195346e+00 -1.40436724e-01 5.18877029e-01 1.18265450e+00
3.04447055e-01 -1.98076814e-01 6.18941009e-01 -8.20628881e-01
2.50322610e-01 -7.22111225e-01 -4.21291828e-01 2.56250024e-01
-6.34540141e-01 -8.79013017e-02 8.86483252e-01 -8.31207633e-01
-7.73965418e-01 -3.11986431e-02 -1.34229928e-01 -7.71042109e-01
-4.42653120e-01 1.68714315e-01 2.14513376e-01 -2.06977904e-01
9.83756185e-01 5.70153475e-01 -1.81177899e-01 -5.94889522e-01
5.17857969e-01 6.56758487e-01 4.59872752e-01 -1.78078666e-01
9.13860917e-01 4.71296459e-01 -2.94120342e-01 -9.80951369e-01
-8.00184071e-01 -7.21532404e-01 -6.45914137e-01 -1.33384421e-01
5.94051063e-01 -6.76904142e-01 -5.08250535e-01 4.51253593e-01
-9.25709963e-01 -1.83646679e-01 -6.86412811e-01 4.56967175e-01
-6.86089873e-01 2.71603018e-01 -2.53423065e-01 -6.72768474e-01
-2.44721338e-01 -7.56268501e-01 5.79306364e-01 5.21181703e-01
-1.27774756e-02 -7.90597677e-01 2.05389172e-01 -9.57502574e-02
5.87704718e-01 9.39245820e-02 8.87786269e-01 -9.74869013e-01
-3.30190867e-01 -2.54185766e-01 -2.32823640e-01 5.87996960e-01
4.47740704e-02 -3.08971941e-01 -1.23819828e+00 -5.15068173e-01
2.08212331e-01 -4.98767406e-01 1.44454217e+00 9.64839086e-02
9.36617136e-01 5.66795468e-02 -1.04440093e-01 6.24395847e-01
1.76033580e+00 2.14484587e-01 7.22098410e-01 1.27897739e-01
4.04404283e-01 3.73403668e-01 4.53828871e-01 5.32368958e-01
-3.73496637e-02 7.07523227e-01 4.11743671e-01 3.45889688e-01
-4.61171776e-01 4.84352349e-04 9.84499231e-02 7.01812685e-01
-2.07060397e-01 -1.51199207e-01 -8.89975131e-01 8.70801806e-01
-1.87075281e+00 -1.24424100e+00 3.25586557e-01 2.20331788e+00
6.32426858e-01 2.35829845e-01 -2.05327477e-02 7.35690892e-02
6.95541084e-01 1.73974156e-01 -6.46566451e-01 -4.15023983e-01
1.48224428e-01 7.18118727e-01 2.30881214e-01 1.38287336e-01
-1.04992354e+00 9.75216568e-01 4.79504633e+00 9.42453563e-01
-1.10686910e+00 4.27235156e-01 3.88233870e-01 -3.09860826e-01
2.11750463e-01 2.24968023e-03 -8.64622295e-01 3.31469744e-01
9.06442225e-01 -5.46757102e-01 3.97463471e-01 1.19922662e+00
-3.43252182e-01 -3.81155172e-03 -1.23859501e+00 1.08163393e+00
5.80883980e-01 -1.29094720e+00 1.58019245e-01 -2.29394689e-01
7.38408387e-01 2.69751400e-01 4.74737957e-02 8.25441360e-01
1.72060654e-01 -9.33449984e-01 5.84869444e-01 7.15255082e-01
8.28365386e-01 -5.74996412e-01 7.52389610e-01 4.37920928e-01
-9.25655186e-01 -4.47866917e-01 -9.07029092e-01 -1.92193612e-01
-1.53295904e-01 3.82594615e-01 -7.76401639e-01 3.75085056e-01
6.19765282e-01 7.62749314e-01 -7.85741270e-01 1.57140911e+00
-1.78756922e-01 4.81133759e-01 -1.97814982e-02 -2.85810322e-01
3.95299017e-01 2.91331500e-01 4.87801850e-01 1.05668569e+00
4.88220990e-01 1.99132413e-01 6.28155172e-02 7.63966620e-01
-2.25757241e-01 -1.60707962e-02 -8.60118508e-01 1.15304366e-01
1.76388517e-01 1.25898898e+00 -7.41219163e-01 -6.37718976e-01
-6.14720702e-01 9.19893146e-01 5.78106046e-01 2.79765189e-01
-6.72131240e-01 -7.45282590e-01 6.21526062e-01 1.36111360e-02
7.97323108e-01 2.22634450e-02 2.36901678e-02 -1.46526599e+00
-1.30580381e-01 -6.25038385e-01 4.27597255e-01 -7.21458912e-01
-1.50629568e+00 6.78924322e-01 -1.06224022e-03 -1.27554345e+00
-3.70661110e-01 -7.13920236e-01 -8.18977654e-01 6.85336471e-01
-1.48879719e+00 -9.08656418e-01 -5.46054244e-01 6.37431920e-01
9.15685654e-01 -5.09513676e-01 9.80922341e-01 2.03678504e-01
-2.06870988e-01 6.36550486e-01 1.87988788e-01 1.26227632e-01
6.74134493e-01 -1.04131758e+00 4.80237633e-01 9.23488736e-01
4.99150455e-01 4.70078498e-01 8.36131275e-01 -2.99304098e-01
-1.15276575e+00 -1.17358398e+00 7.58390725e-01 -4.42975640e-01
7.32358038e-01 -4.16873276e-01 -1.19188380e+00 6.12333596e-01
9.26339850e-02 3.48237097e-01 4.23621416e-01 7.51059726e-02
-6.30713463e-01 -1.54953882e-01 -1.11012053e+00 6.35976017e-01
1.08486223e+00 -2.94995278e-01 -1.05796242e+00 3.17920208e-01
5.42586327e-01 1.01215169e-01 -3.19991469e-01 2.29200393e-01
4.05384988e-01 -1.13312101e+00 8.35704029e-01 -7.67903209e-01
4.34758812e-01 -1.61100045e-01 -3.21956784e-01 -1.54055715e+00
-5.76143086e-01 -5.97683229e-02 -2.89342016e-01 1.00891340e+00
6.45873845e-02 -7.30942130e-01 5.21116436e-01 2.47719213e-01
-1.98772639e-01 -6.94553852e-01 -1.08182740e+00 -1.19834089e+00
-1.68529283e-02 -4.03652668e-01 1.81892693e-01 8.47493172e-01
-1.97613969e-01 3.84144425e-01 -4.90245908e-01 -2.08327904e-01
7.27928519e-01 1.31617606e-01 6.28610075e-01 -1.24768841e+00
-4.30882215e-01 -4.76179540e-01 -8.90726030e-01 -5.17862678e-01
-8.94193072e-03 -1.04093456e+00 9.49421972e-02 -1.19418085e+00
5.54441333e-01 -1.04380310e-01 -7.28726923e-01 7.29699314e-01
-1.45017938e-03 5.70409596e-01 6.09777391e-01 7.12897927e-02
-7.23734200e-01 7.33412623e-01 8.12237561e-01 -3.04291070e-01
-1.04974583e-02 5.11943176e-02 -7.56554842e-01 5.77096045e-01
6.60391569e-01 -6.03403807e-01 -4.31327492e-01 -3.93254682e-02
-3.71073097e-01 -3.55725884e-01 7.12927878e-01 -1.43611825e+00
4.33459103e-01 -1.04568964e-02 4.21724468e-01 -1.50542304e-01
3.93434197e-01 -6.32878244e-01 -4.25449312e-02 7.40166247e-01
-1.55146435e-01 -4.88119900e-01 1.03714623e-01 6.90080523e-01
-3.20678428e-02 -7.83586442e-01 1.28114402e+00 -2.75964499e-01
-1.26501358e+00 4.18653995e-01 -1.64464310e-01 6.08682185e-02
1.34169340e+00 -2.08534896e-01 -5.63711882e-01 -2.30856776e-01
-8.56153309e-01 -2.58838162e-02 5.36784708e-01 6.86158955e-01
7.83893824e-01 -1.26835930e+00 -6.75382137e-01 2.98905790e-01
6.28371358e-01 -6.24498725e-01 4.30779725e-01 6.08785927e-01
-1.76917896e-01 3.78332824e-01 -6.35215580e-01 -5.21188974e-01
-1.18141150e+00 9.45676506e-01 2.30164215e-01 8.08469951e-02
-8.11544120e-01 8.11230063e-01 1.50914490e-01 -9.72321108e-02
3.14883143e-01 7.04420954e-02 -3.79593492e-01 2.01211154e-01
9.57738042e-01 2.02288762e-01 1.44847244e-01 -3.91587973e-01
-2.96737045e-01 4.71970290e-01 -3.94778341e-01 2.41393954e-01
1.66767943e+00 1.63712755e-01 4.04468536e-01 6.97226048e-01
1.28160203e+00 -6.04673326e-01 -1.18819070e+00 -3.85117650e-01
-7.24051148e-02 -5.73001981e-01 -1.24380782e-01 -8.06534648e-01
-8.20349097e-01 1.09440112e+00 1.05333817e+00 2.62876660e-01
1.02145088e+00 1.28166974e-01 6.18021965e-01 7.00877964e-01
4.62170780e-01 -9.73348677e-01 2.14955479e-01 5.88104010e-01
8.16664994e-01 -1.38493538e+00 -1.72833845e-01 1.37231350e-01
-6.56838238e-01 1.15984762e+00 4.84643430e-01 -5.45524716e-01
6.84229493e-01 -2.60255188e-01 -2.95484453e-01 -1.82155102e-01
-1.08698797e+00 -6.79023445e-01 2.73374349e-01 8.34550381e-01
1.87330041e-02 -9.43600982e-02 -3.21708411e-01 7.12696373e-01
8.09192955e-02 2.11392805e-01 7.19744027e-01 7.50969410e-01
-8.28028560e-01 -7.87526250e-01 3.13922316e-02 6.67054355e-01
-1.23902373e-01 -1.80705056e-01 1.67649966e-02 7.21079409e-01
1.76017553e-01 8.16124558e-01 5.77291697e-02 -5.04421473e-01
4.69261289e-01 4.24616784e-01 3.43738526e-01 -1.14235711e+00
-2.92679191e-01 -3.69078428e-01 -2.31523231e-01 -3.79837930e-01
-1.67904168e-01 -4.26523328e-01 -7.50568628e-01 -2.13998556e-01
-1.77261934e-01 -6.91264421e-02 5.65605342e-01 9.69280481e-01
1.82813242e-01 5.70190191e-01 6.74733639e-01 -8.84247005e-01
-8.92243564e-01 -1.24828899e+00 -6.64961874e-01 6.51547313e-01
1.87220797e-01 -9.60617363e-01 -5.38349569e-01 8.91996399e-02] | [9.932404518127441, 2.7950448989868164] |
79dba9a9-ff59-4b35-a5d0-874fb3774fc5 | technical-outlier-detection-via-convolutional | 2305.12068 | null | https://arxiv.org/abs/2305.12068v1 | https://arxiv.org/pdf/2305.12068v1.pdf | Technical outlier detection via convolutional variational autoencoder for the ADMANI breast mammogram dataset | The ADMANI datasets (annotated digital mammograms and associated non-image datasets) from the Transforming Breast Cancer Screening with AI programme (BRAIx) run by BreastScreen Victoria in Australia are multi-centre, large scale, clinically curated, real-world databases. The datasets are expected to aid in the development of clinically relevant Artificial Intelligence (AI) algorithms for breast cancer detection, early diagnosis, and other applications. To ensure high data quality, technical outliers must be removed before any downstream algorithm development. As a first step, we randomly select 30,000 individual mammograms and use Convolutional Variational Autoencoder (CVAE), a deep generative neural network, to detect outliers. CVAE is expected to detect all sorts of outliers, although its detection performance differs among different types of outliers. Traditional image processing techniques such as erosion and pectoral muscle analysis can compensate for the poor performance of CVAE in certain outlier types. We identify seven types of technical outliers: implant, pacemaker, cardiac loop recorder, improper radiography, atypical lesion/calcification, incorrect exposure parameter and improper placement. The outlier recall rate for the test set is 61% if CVAE, erosion and pectoral muscle analysis each select the top 1% images ranked in ascending or descending order according to image outlier score under each detection method, and 83% if each selects the top 5% images. This study offers an overview of technical outliers in the ADMANI dataset and suggests future directions to improve outlier detection effectiveness. | ['Davis J. McCarthy', 'Susan Wei', 'Carlos A. Pena Solorzano', 'Hui Li'] | 2023-05-20 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection', 'outlier-detection'] | ['knowledge-base', 'medical', 'methodology'] | [ 3.39671135e-01 3.93838495e-01 1.13030292e-01 -4.26681936e-01
-1.04388082e+00 -2.49010697e-01 1.53293326e-01 6.08235896e-01
-3.71067584e-01 1.27383590e-01 4.36718166e-01 -6.85188472e-01
-5.13171136e-01 -6.36861265e-01 -8.54659081e-01 -5.66370666e-01
-1.56822845e-01 7.46875465e-01 1.92842647e-01 2.70037264e-01
-3.49139533e-04 4.44266796e-01 -1.11849058e+00 7.00299263e-01
4.96553004e-01 9.04602349e-01 -6.55423701e-02 1.08794045e+00
2.18330920e-01 7.74226248e-01 -3.61730814e-01 -5.92584252e-01
2.63816237e-01 -6.44127250e-01 -2.40481079e-01 1.79464057e-01
5.59676588e-01 -6.37615204e-01 -6.18497431e-02 8.67082059e-01
9.87608194e-01 -4.22668427e-01 9.06187534e-01 -7.30078161e-01
-4.98224944e-01 6.03649914e-01 -7.22068429e-01 6.37314916e-01
-3.14096153e-01 3.74117970e-01 3.46847534e-01 -8.38648021e-01
6.23470068e-01 8.14715147e-01 1.32158267e+00 5.46209574e-01
-1.12673914e+00 -3.39953631e-01 -6.82329714e-01 -1.84508681e-01
-9.13896680e-01 -4.21082795e-01 2.64252603e-01 -5.41689813e-01
8.68233442e-01 5.97445011e-01 7.48937726e-01 1.03861499e+00
8.54060292e-01 4.07819331e-01 4.46332842e-01 -2.41779372e-01
2.61428922e-01 -1.77973583e-01 1.24966549e-02 5.88056862e-01
8.34419250e-01 1.28860414e-01 -4.85272445e-02 -7.49450088e-01
5.79186618e-01 2.49620736e-01 2.24764585e-01 -2.74130225e-01
-1.01289320e+00 8.87414873e-01 3.64496291e-01 2.56824583e-01
-6.88357651e-01 2.36233026e-01 8.54631782e-01 2.37340733e-01
1.49032027e-01 1.85791880e-01 -1.76951751e-01 9.12206620e-02
-1.35697651e+00 1.65031031e-02 4.35174286e-01 4.32272553e-01
-3.38140994e-01 1.99437484e-01 -1.49930716e-01 9.54507053e-01
6.72552168e-01 2.87745565e-01 9.95867491e-01 -1.07745922e+00
-5.71984472e-03 6.15137637e-01 -1.74034089e-01 -1.13274086e+00
-7.83039570e-01 -1.46582723e-01 -9.49365377e-01 1.64193556e-01
3.39023888e-01 6.71589747e-02 -1.50298285e+00 9.53386724e-01
1.43505722e-01 -1.66174352e-01 3.60389262e-01 5.95507681e-01
1.05738270e+00 1.33382529e-01 2.49983713e-01 -2.64671057e-01
1.28712475e+00 -6.82758018e-02 -4.86898810e-01 -1.09845631e-01
9.31852520e-01 -6.30032659e-01 5.80143631e-01 4.67163503e-01
-1.20686483e+00 -4.26601380e-01 -1.02567351e+00 4.31802794e-02
-9.31725204e-02 2.05377743e-01 4.84342158e-01 9.03488398e-01
-8.35663021e-01 6.22249782e-01 -1.64660668e+00 -3.53811651e-01
8.11563492e-01 6.41715825e-01 -3.84800851e-01 -2.36097619e-01
-6.77494049e-01 8.07952285e-01 2.69715339e-01 3.68374079e-01
-8.67682457e-01 -7.57570744e-01 -1.06548941e+00 -3.38351458e-01
-5.94953150e-02 -4.69151765e-01 1.03914642e+00 -1.14434528e+00
-5.99680662e-01 1.16134238e+00 1.09033577e-01 -8.64325643e-01
8.74585927e-01 -8.06650445e-02 -6.40403390e-01 1.38285216e-02
2.09789485e-01 4.36519623e-01 7.59265721e-01 -9.92007077e-01
-5.39625585e-01 -7.40332901e-01 -1.26723158e+00 -1.54342398e-01
7.37121031e-02 1.59351274e-01 -2.90874958e-01 -5.82008839e-01
5.18973172e-01 -9.87826288e-01 -5.65120637e-01 7.12698698e-02
-1.28471002e-01 3.17761973e-02 6.36443496e-01 -1.32253921e+00
1.06750011e+00 -2.36996055e+00 -3.23068082e-01 4.82999563e-01
4.92976040e-01 7.49627501e-02 3.90671998e-01 -3.10306013e-01
-3.49646419e-01 2.13685110e-01 -4.50866282e-01 1.50240539e-02
-3.43828797e-01 3.27881783e-01 5.47282815e-01 8.15876245e-01
4.62133855e-01 9.02443707e-01 -7.22056091e-01 -5.72783172e-01
3.13865334e-01 3.58856052e-01 -5.51884592e-01 -3.45190376e-01
2.00095937e-01 5.64641841e-02 -1.07919201e-01 1.13639259e+00
6.14305198e-01 1.62661020e-02 3.73135996e-03 -1.72753319e-01
1.69536963e-01 -5.41406512e-01 -1.06339455e+00 1.02706754e+00
1.87619612e-01 6.90236926e-01 -1.00270681e-01 -8.59763682e-01
7.33246624e-01 3.23969454e-01 7.74361312e-01 -5.26392341e-01
1.34575456e-01 7.36990452e-01 6.97558701e-01 -9.38455164e-01
2.48072669e-01 -2.43608877e-01 2.16597378e-01 -3.91994566e-02
-1.65069893e-01 2.69587606e-01 -9.07951780e-03 2.81587839e-01
1.43996394e+00 -3.34325969e-01 4.59172577e-01 -1.50715008e-01
-1.08681574e-01 4.31910545e-01 8.07273090e-01 9.69435155e-01
-6.06290102e-01 1.22883356e+00 5.90218723e-01 -5.73753476e-01
-1.32342923e+00 -1.11627638e+00 -7.53503382e-01 5.13840437e-01
-5.23054123e-01 4.80447710e-02 -1.88251019e-01 -4.44209129e-01
4.33532417e-01 5.57823896e-01 -8.51004124e-01 -4.72881883e-01
-6.61041379e-01 -1.38603747e+00 6.82287574e-01 8.48789394e-01
5.44981509e-02 -1.28048754e+00 -1.04743659e+00 4.61040109e-01
3.69040430e-01 -4.00344729e-01 1.28061518e-01 3.83262247e-01
-1.11900890e+00 -1.33899271e+00 -1.03217995e+00 -7.33216763e-01
7.87350476e-01 -5.19246519e-01 1.08531153e+00 2.17293888e-01
-9.88416076e-01 3.18921745e-01 -9.42967981e-02 -7.32883155e-01
-9.68267143e-01 -6.62155032e-01 -1.04373537e-01 -3.80769879e-01
7.39982009e-01 2.07907837e-02 -6.07880533e-01 -1.54545516e-01
-1.19519305e+00 -4.74103779e-01 9.23783302e-01 1.04442704e+00
1.02311468e+00 8.92886147e-02 1.56648323e-01 -1.10309744e+00
3.55960578e-01 -9.23037648e-01 -2.48372480e-01 -1.54214695e-01
-3.11410517e-01 -4.23764974e-01 -8.79213065e-02 -3.48982394e-01
-7.44863570e-01 1.40419006e-01 -1.71507344e-01 -3.14363420e-01
-1.92830503e-01 7.41660714e-01 2.26289824e-01 2.10407019e-01
1.09534776e+00 -1.84241265e-01 3.41000408e-01 -1.41362280e-01
-4.35349733e-01 6.93675697e-01 8.11001360e-01 1.27110928e-01
2.91312724e-01 7.05406785e-01 4.49809013e-03 -8.87170970e-01
-2.81631380e-01 -5.50175011e-01 -4.32064116e-01 1.16769932e-02
9.30101395e-01 -7.50903428e-01 -3.41544226e-02 4.16752934e-01
-4.84569103e-01 -2.58259475e-01 -4.35309142e-01 7.49543190e-01
-3.47611457e-01 1.59053504e-01 -4.91282851e-01 -7.13067949e-01
-4.94204730e-01 -1.41925919e+00 1.05663717e+00 2.37405285e-01
-5.46904564e-01 -6.25449538e-01 -1.50737479e-01 2.36333206e-01
2.76980370e-01 9.15559173e-01 1.03706229e+00 -9.57214534e-01
5.17632663e-02 -5.66427290e-01 -2.28578046e-01 3.64362627e-01
3.15403193e-02 3.86153072e-01 -3.32520992e-01 -2.39628986e-01
2.04236805e-01 5.75937144e-02 1.08179355e+00 1.29989409e+00
1.08000493e+00 1.61382750e-01 -6.27878904e-01 5.86320817e-01
1.51361871e+00 5.78870714e-01 7.12011039e-01 5.42671740e-01
5.55302501e-01 1.88729987e-01 1.56861141e-01 2.59362668e-01
-9.35082659e-02 -8.69611427e-02 4.93209779e-01 -2.64031827e-01
-4.47896048e-02 2.06297085e-01 1.93898544e-01 1.76992267e-01
3.73117067e-02 1.19032092e-01 -1.39636803e+00 1.01902390e+00
-1.46756923e+00 -8.80120456e-01 -6.84918642e-01 1.98382688e+00
3.56263638e-01 5.16848505e-01 1.05760641e-01 3.33109617e-01
6.21266484e-01 -4.89244968e-01 -6.68959916e-01 -5.87939858e-01
-3.82293537e-02 2.25770116e-01 9.60674882e-01 3.38770561e-02
-9.20966864e-01 3.52339357e-01 6.27800703e+00 6.79572746e-02
-9.55558002e-01 8.14753696e-02 9.68537629e-01 -2.73170382e-01
-1.66817539e-04 -3.37041169e-01 -5.32142162e-01 5.83771765e-01
1.28343368e+00 3.98920625e-01 -4.26383793e-01 7.92410016e-01
3.00634682e-01 -5.60637474e-01 -9.01026845e-01 5.37829280e-01
1.37856022e-01 -1.27572966e+00 -2.15770140e-01 1.09753527e-01
4.02018011e-01 2.63102621e-01 4.34234180e-02 1.75379127e-01
1.87918782e-01 -1.25515950e+00 2.83685148e-01 7.19742775e-01
6.58032775e-01 -8.04926217e-01 1.31382751e+00 -1.74819246e-01
-2.53794879e-01 -2.13947177e-01 -5.88128626e-01 4.33199644e-01
-2.45812923e-01 7.29195774e-01 -9.38430727e-01 2.80644417e-01
1.36652172e+00 2.85167783e-01 -9.20581937e-01 1.21566141e+00
8.11955750e-01 1.08542430e+00 -7.10164309e-01 2.78928846e-01
6.52582049e-02 2.24414483e-01 4.99212921e-01 1.01488352e+00
4.99348640e-01 2.38912538e-01 -3.70274156e-01 5.12393475e-01
1.69421926e-01 -3.85809108e-03 -4.43450361e-01 -4.85457014e-03
1.52316734e-01 8.53152096e-01 -1.00039089e+00 -1.88174844e-01
-5.77007949e-01 8.78810704e-01 -4.71993029e-01 -2.25633755e-02
-5.90506911e-01 -1.46648258e-01 -8.49425420e-02 3.57398003e-01
1.80800706e-01 5.24793744e-01 -5.78596354e-01 -5.43378294e-01
1.02661669e-01 -1.14242864e+00 8.97632241e-01 -5.11823893e-01
-1.13638985e+00 9.85998288e-02 -2.03235328e-01 -8.32066417e-01
-3.00300717e-01 -6.62440836e-01 -5.80772996e-01 5.63876450e-01
-7.09321439e-01 -8.51983607e-01 -4.25292522e-01 1.72830313e-01
5.61964273e-01 -3.62361550e-01 7.34814346e-01 6.59948885e-02
-5.03682852e-01 7.25678027e-01 6.36503935e-01 4.85386848e-01
6.27835512e-01 -1.39627540e+00 4.55997437e-02 5.80603719e-01
-3.43782455e-01 2.36745819e-01 6.23938382e-01 -1.14894474e+00
-1.15930665e+00 -1.23103452e+00 4.24461663e-01 -6.49148464e-01
2.76999414e-01 4.37866211e-01 -9.21731532e-01 1.00587225e+00
-3.58242959e-01 3.71107996e-01 1.09750700e+00 -3.92958730e-01
4.29311067e-01 2.35802233e-01 -1.42763412e+00 4.16880071e-01
3.58449608e-01 4.01398063e-01 -5.89973450e-01 2.66293198e-01
6.44824579e-02 -8.33376586e-01 -1.19407320e+00 8.58181834e-01
7.16564715e-01 -1.16572845e+00 8.29380453e-01 -6.88845634e-01
6.93457723e-01 7.12075010e-02 -1.82741597e-01 -7.14828491e-01
-2.43870854e-01 -3.01397979e-01 -9.44296718e-02 8.31297219e-01
5.10063946e-01 -2.23876774e-01 1.16901529e+00 7.38207757e-01
-3.99597168e-01 -7.40480244e-01 -1.00909758e+00 -1.69042394e-01
2.04238102e-01 -6.35169804e-01 4.45402004e-02 7.34807134e-01
-5.92024088e-01 -6.88896477e-01 1.07055381e-01 3.49150002e-01
5.53764462e-01 -5.55334985e-01 3.71493995e-01 -1.22480559e+00
-3.08330953e-01 -4.08771574e-01 -9.94666338e-01 3.31953287e-01
-7.30221033e-01 -8.53201926e-01 -8.11627582e-02 -1.52355170e+00
3.12574536e-01 -2.17075184e-01 -4.24027652e-01 1.91548645e-01
-3.49617124e-01 4.61308539e-01 -2.57130980e-01 2.53255725e-01
5.52380867e-02 -2.06856817e-01 7.50679016e-01 -1.41681120e-01
-4.41033870e-01 2.09646434e-01 -7.96106696e-01 1.05451822e+00
6.55412734e-01 -7.93790281e-01 1.57551900e-01 -4.42056328e-01
1.41920507e-01 1.06937364e-01 7.05031276e-01 -1.25764406e+00
-9.64901447e-02 1.33156508e-01 1.52192545e+00 -8.87091756e-01
-2.20010981e-01 -9.07055974e-01 4.58372682e-01 1.05629647e+00
-1.95662096e-01 1.14749253e-01 4.90712434e-01 6.32487476e-01
1.08499877e-01 -3.18458945e-01 1.07247186e+00 -4.27544534e-01
-4.57581341e-01 -1.56028057e-02 -6.07746780e-01 -1.35473818e-01
1.25902045e+00 -7.61022806e-01 2.60421455e-01 -4.29835655e-02
-1.27725625e+00 1.86441377e-01 3.17429006e-01 1.49892583e-01
6.57428741e-01 -1.24510419e+00 -1.17053962e+00 2.59886682e-01
1.76837683e-01 3.73842984e-01 4.66554850e-01 1.16040146e+00
-1.12427986e+00 -8.06962550e-02 -1.21988147e-01 -9.61893022e-01
-1.42110622e+00 1.37171939e-01 5.83678603e-01 -1.83622107e-01
-9.99806821e-01 1.06579185e+00 -3.72095615e-01 8.45332339e-04
2.56919712e-01 -3.33616525e-01 5.85323051e-02 -5.13152294e-02
4.00048316e-01 4.39426154e-01 4.33025271e-01 -6.14745915e-01
-3.12578559e-01 4.95990328e-02 -4.05830204e-01 1.75062776e-01
1.41680253e+00 3.63586724e-01 1.04929864e-01 3.09322536e-01
1.01432347e+00 -2.66762584e-01 -8.46869051e-01 2.13679701e-01
3.01254362e-01 -1.95823871e-02 3.58244121e-01 -6.41276658e-01
-1.04992819e+00 3.92857403e-01 1.43336165e+00 1.63934335e-01
1.00051272e+00 1.37966633e-01 5.74707985e-01 7.68896416e-02
-6.13838792e-01 -1.08913350e+00 -1.56609401e-01 -3.16713572e-01
8.80489469e-01 -1.60189223e+00 2.72057116e-01 2.15494901e-01
-7.87318468e-01 1.20411503e+00 5.33560514e-01 -3.52051914e-01
6.38722718e-01 3.96779269e-01 3.04309487e-01 -5.09803116e-01
-2.92036742e-01 3.64900321e-01 2.02386722e-01 8.13262641e-01
5.85035801e-01 8.86440575e-02 -4.43303674e-01 8.79027486e-01
-1.72948405e-01 4.34756689e-02 5.21874189e-01 1.10945499e+00
-4.36518818e-01 -3.55110139e-01 -6.78265989e-01 1.75498140e+00
-9.94123340e-01 4.79723029e-02 -3.58944803e-01 1.08430552e+00
4.50651586e-01 3.64047825e-01 4.06569183e-01 2.63285726e-01
4.19047266e-01 9.30102542e-02 8.58639702e-02 -4.28087175e-01
-6.66796446e-01 4.95979220e-01 -1.40886098e-01 -4.19299752e-01
-1.19277775e-01 -1.10677803e+00 -1.28904033e+00 1.21791057e-01
-4.72209573e-01 -2.96054125e-01 6.92193031e-01 6.60387933e-01
1.98943049e-01 7.67340958e-01 -2.20242053e-01 -5.05666614e-01
-5.70780098e-01 -1.05332708e+00 -4.91029829e-01 4.33843017e-01
4.69527006e-01 -4.24930632e-01 -4.67224896e-01 2.31421009e-01] | [15.246638298034668, -2.4819483757019043] |
24cfab08-0ab5-4f04-b872-5c61edba70e1 | measuring-and-improving-compositional-1 | 2205.02054 | null | https://arxiv.org/abs/2205.02054v1 | https://arxiv.org/pdf/2205.02054v1.pdf | Measuring and Improving Compositional Generalization in Text-to-SQL via Component Alignment | In text-to-SQL tasks -- as in much of NLP -- compositional generalization is a major challenge: neural networks struggle with compositional generalization where training and test distributions differ. However, most recent attempts to improve this are based on word-level synthetic data or specific dataset splits to generate compositional biases. In this work, we propose a clause-level compositional example generation method. We first split the sentences in the Spider text-to-SQL dataset into sub-sentences, annotating each sub-sentence with its corresponding SQL clause, resulting in a new dataset Spider-SS. We then construct a further dataset, Spider-CG, by composing Spider-SS sub-sentences in different combinations, to test the ability of models to generalize compositionally. Experiments show that existing models suffer significant performance degradation when evaluated on Spider-CG, even though every sub-sentence is seen during training. To deal with this problem, we modify a number of state-of-the-art models to train on the segmented data of Spider-SS, and we show that this method improves the generalization performance. | ['Matthew Purver', 'Qiuping Huang', 'Xinyun Chen', 'Yujian Gan'] | 2022-05-04 | null | https://aclanthology.org/2022.findings-naacl.62 | https://aclanthology.org/2022.findings-naacl.62.pdf | findings-naacl-2022-7 | ['text-to-sql'] | ['computer-code'] | [ 6.96201503e-01 3.19279075e-01 -5.19896112e-02 -7.76011407e-01
-9.19967055e-01 -8.40537429e-01 4.93921310e-01 6.15189262e-02
-2.68924683e-01 8.92996728e-01 1.32597148e-01 -5.74699342e-01
3.86927962e-01 -8.73106480e-01 -1.01960647e+00 -3.57536107e-01
8.55855867e-02 9.18733776e-01 5.39364278e-01 -3.82489353e-01
1.47915885e-01 2.55069464e-01 -1.44006968e+00 1.04942966e+00
1.17100632e+00 7.18549430e-01 1.59335211e-01 5.80273986e-01
-3.90397370e-01 5.00681877e-01 -9.36793447e-01 -6.57161891e-01
3.30294371e-01 -8.88582170e-01 -9.20512557e-01 1.07882498e-02
8.36508751e-01 3.34606618e-02 2.41549999e-01 9.50460255e-01
2.89576203e-01 -1.20445520e-01 6.51743710e-01 -1.21504641e+00
-4.99092847e-01 1.20163977e+00 -1.41478524e-01 -7.56165013e-02
2.73158282e-01 5.67649454e-02 1.19863343e+00 -8.71792614e-01
9.26955223e-01 1.45539463e+00 9.27417338e-01 9.74487185e-01
-1.96181703e+00 -6.45331740e-01 1.50686741e-01 -3.36128734e-02
-1.03854191e+00 -1.49570227e-01 6.78046882e-01 -8.30634162e-02
1.02445602e+00 3.95589709e-01 6.40220284e-01 1.60850191e+00
-4.84241284e-02 1.04167783e+00 1.37525821e+00 -4.67531830e-01
4.68735695e-01 2.49084696e-01 1.11865193e-01 3.20104957e-01
2.26250559e-01 -1.00270644e-01 -4.53005135e-01 -4.41941321e-02
2.35355690e-01 -4.85121369e-01 -2.38578796e-01 -3.98408294e-01
-1.11245966e+00 9.39610362e-01 4.74903941e-01 2.83896357e-01
3.28496918e-02 -1.46200776e-01 7.54582226e-01 4.01629955e-01
5.33101737e-01 7.24260986e-01 -6.34647965e-01 1.11968637e-01
-1.23876131e+00 1.01909173e+00 1.10928047e+00 1.12659228e+00
6.71204329e-01 -1.17594562e-01 -1.13194071e-01 1.08243597e+00
-2.23892450e-01 8.42974633e-02 8.63203049e-01 -6.93584979e-01
8.17181706e-01 5.75462162e-01 -4.68208998e-01 -3.71395171e-01
-1.92877308e-01 -5.38017750e-01 -6.38709545e-01 -1.45556992e-02
3.67850572e-01 -1.28334954e-01 -1.02435088e+00 2.00161099e+00
1.55460328e-01 -3.19494516e-01 3.00172478e-01 4.64640290e-01
5.06573081e-01 6.52928948e-01 3.62294205e-02 -2.14089695e-02
9.50527847e-01 -9.99526262e-01 -3.01409721e-01 -6.07001901e-01
7.79029369e-01 -3.62013161e-01 1.78220570e+00 3.63137484e-01
-1.06094027e+00 -5.69321573e-01 -1.27588511e+00 1.80240840e-01
-5.73405206e-01 -5.20206504e-02 1.18619062e-01 7.32296288e-01
-8.80684257e-01 6.79693401e-01 -6.43890440e-01 -5.22629738e-01
4.48075145e-01 1.48812041e-01 -1.66737705e-01 -8.82720053e-02
-1.10538340e+00 8.53478789e-01 8.30137074e-01 -5.22313178e-01
-7.71839380e-01 -9.10924971e-01 -8.95188093e-01 9.64882970e-02
5.03609657e-01 -7.63206303e-01 1.61215186e+00 -1.07281888e+00
-1.16686416e+00 8.74807596e-01 -1.60254315e-01 -7.50663936e-01
5.55223823e-01 6.35078223e-03 -7.77385160e-02 -2.10702106e-01
3.10126215e-01 1.09874821e+00 7.10769236e-01 -1.69195890e+00
-5.54199278e-01 -4.37656671e-01 -5.46326414e-02 3.27793328e-04
-2.43064195e-01 -2.39836037e-01 6.91315606e-02 -1.03922820e+00
9.67571735e-02 -1.09445441e+00 -1.64186046e-01 -6.07412636e-01
-7.88100660e-01 -1.92189157e-01 7.37143576e-01 -3.30840409e-01
1.23119020e+00 -1.94762278e+00 2.45840773e-01 -8.05657879e-02
-3.62489790e-01 3.09579253e-01 -3.10855925e-01 6.20462298e-01
-2.93926269e-01 4.18871224e-01 -8.12848628e-01 -5.86087883e-01
1.24299236e-01 4.35783863e-01 -6.20810449e-01 -3.28229278e-01
5.58192730e-01 8.46396327e-01 -7.43325770e-01 -3.87112796e-01
-3.03102434e-01 -1.59605443e-01 -9.14974570e-01 7.00730905e-02
-9.58409607e-01 1.45564169e-01 -8.73438912e-05 3.82386506e-01
5.36081910e-01 1.24708660e-01 2.27335542e-01 -2.77681947e-02
3.24301094e-01 7.89586246e-01 -8.78748715e-01 1.86652565e+00
-5.73273063e-01 5.46382964e-01 -3.37505430e-01 -1.21510875e+00
8.30354273e-01 -1.65322557e-01 -2.29370683e-01 -5.32696605e-01
-2.97444046e-01 5.71857095e-01 2.51235694e-01 -4.85437244e-01
7.34811604e-01 -6.01562798e-01 -2.87973195e-01 4.69576091e-01
2.07765445e-01 -6.64930761e-01 7.97345817e-01 3.12704951e-01
1.02751815e+00 4.20319289e-01 -6.13116622e-02 -2.98342139e-01
4.99426693e-01 2.38987774e-01 3.36263776e-01 7.61117041e-01
3.52940202e-01 9.21675682e-01 7.95508742e-01 -3.15683663e-01
-1.34315968e+00 -1.40416121e+00 -1.16294742e-01 1.11697769e+00
-3.80072683e-01 -5.51696241e-01 -1.34505546e+00 -1.19238460e+00
4.40059155e-02 1.66880095e+00 -7.16713130e-01 -2.80046105e-01
-7.97320545e-01 -9.01336789e-01 6.86257243e-01 5.84504962e-01
3.77742380e-01 -1.37126398e+00 -6.24641359e-01 2.59420127e-01
-9.66782123e-02 -1.04976046e+00 -2.91825771e-01 5.87093592e-01
-9.00695682e-01 -6.94008648e-01 -4.75410372e-01 -9.19717491e-01
5.49678743e-01 -3.94378215e-01 1.52813339e+00 -3.47980559e-01
-3.05569619e-01 -2.31312275e-01 -4.48452443e-01 -4.82427448e-01
-1.09075177e+00 5.62453866e-01 -3.52047563e-01 -3.64497840e-01
2.75939405e-01 -5.89391351e-01 -7.54067078e-02 1.71854824e-01
-1.49212730e+00 1.23387769e-01 6.22387052e-01 9.99657273e-01
4.05781060e-01 -1.47044480e-01 9.85854685e-01 -1.40124094e+00
1.07023239e+00 -2.47957379e-01 -1.47727147e-01 2.30259955e-01
-4.94391948e-01 3.30875993e-01 8.89380932e-01 -8.48735452e-01
-8.84628177e-01 -1.72397010e-02 -2.18069598e-01 -1.56585604e-01
-2.74658501e-01 6.80003762e-01 -3.42068791e-01 4.93420094e-01
1.15759933e+00 4.55756992e-01 4.11938168e-02 -3.27147096e-01
6.39363050e-01 7.00827956e-01 4.96776670e-01 -8.13236535e-01
5.98199725e-01 -6.30468056e-02 -1.53666779e-01 -6.84932292e-01
-1.13090920e+00 2.28523672e-01 -3.79956871e-01 3.21132690e-01
6.04344964e-01 -4.09745246e-01 3.93650085e-02 2.58494884e-01
-1.11623001e+00 -7.98404038e-01 -6.48712277e-01 7.45119201e-03
-6.28667891e-01 1.68773577e-01 -5.40013552e-01 -5.02237380e-01
-2.20344335e-01 -1.13476622e+00 1.08227992e+00 -1.50915563e-01
-6.14659011e-01 -7.88948596e-01 2.03458667e-01 1.65043220e-01
4.28701818e-01 1.43185660e-01 1.53339612e+00 -1.22794855e+00
-1.77447379e-01 -1.41498521e-01 -3.91878299e-02 8.14047813e-01
5.04008271e-02 -1.57519385e-01 -7.58309782e-01 -1.84602365e-01
9.90339071e-02 -7.59044707e-01 8.23218822e-01 -5.92096336e-02
1.39575839e+00 -4.30036008e-01 -2.40112618e-01 3.49971145e-01
1.19548309e+00 6.13198392e-02 4.80402946e-01 4.27254677e-01
2.78315455e-01 8.37308466e-01 5.30512154e-01 -1.36391625e-01
1.94042519e-01 5.36456943e-01 3.31927538e-01 2.20003083e-01
-1.41040593e-01 -5.32918096e-01 3.58920366e-01 3.81373644e-01
6.21472359e-01 -5.02368391e-01 -8.79094183e-01 7.20166028e-01
-1.59288979e+00 -8.89126599e-01 4.86535169e-02 2.21035647e+00
1.29537487e+00 5.02770483e-01 2.38877267e-01 3.12811017e-01
3.26757014e-01 1.69945080e-02 -4.85459954e-01 -4.63941753e-01
-2.67496377e-01 5.25907278e-01 1.06585801e-01 4.13313240e-01
-8.93126905e-01 1.16417348e+00 6.24807787e+00 1.02336538e+00
-1.11232138e+00 -1.98413715e-01 6.92807436e-01 -2.98765093e-01
-6.04049265e-01 -7.10485280e-02 -8.55909765e-01 5.42604148e-01
1.05827713e+00 6.93824589e-02 4.21625495e-01 7.94318795e-01
-2.34031811e-01 1.05393899e-03 -1.63816392e+00 4.34004873e-01
5.21999002e-01 -1.27269900e+00 6.55467927e-01 -3.54015738e-01
8.09141219e-01 -1.27588049e-01 2.38480791e-02 6.96685076e-01
1.20101705e-01 -1.03905296e+00 8.70754421e-01 -1.19085588e-01
6.00308418e-01 -6.84334755e-01 6.84822440e-01 6.03602767e-01
-6.31311715e-01 3.34118865e-02 -3.04325819e-01 2.76239991e-01
7.85615668e-02 3.74371648e-01 -1.11799049e+00 4.96674389e-01
6.29985392e-01 2.30371147e-01 -9.99316037e-01 6.61710978e-01
-3.23193252e-01 7.87205279e-01 -2.89449275e-01 -3.81355703e-01
2.75120556e-01 -1.97457328e-01 5.12762785e-01 1.47893286e+00
3.54267478e-01 -3.96520585e-01 -7.12036388e-03 1.36541307e+00
-1.99537992e-01 6.45828396e-02 -7.15216815e-01 -1.63960472e-01
3.48998129e-01 9.37893271e-01 -5.35559654e-01 -6.68403804e-01
-2.84901917e-01 1.08681822e+00 6.15846634e-01 4.22362596e-01
-7.07831025e-01 -5.81689239e-01 2.42492050e-01 4.31937218e-01
3.98694783e-01 4.11621779e-02 -4.76863950e-01 -9.99717712e-01
3.63451570e-01 -1.43839908e+00 1.96236581e-01 -8.04912746e-01
-1.53265798e+00 8.77911448e-01 4.38728333e-01 -8.42399538e-01
-6.89894438e-01 -6.28523290e-01 -5.90504229e-01 9.22254324e-01
-8.08637202e-01 -1.30043173e+00 -1.38749376e-01 1.55463710e-01
9.42998707e-01 -2.61056721e-01 7.90971220e-01 -2.99173910e-02
-3.14713001e-01 7.47248888e-01 -2.44789243e-01 6.98398892e-03
7.26408005e-01 -1.59375632e+00 1.03970790e+00 5.96272767e-01
3.10987264e-01 7.97916949e-01 9.60412502e-01 -6.61319077e-01
-8.18423569e-01 -1.19591582e+00 9.55180109e-01 -3.60710680e-01
7.24402905e-01 -9.42575634e-01 -1.17409503e+00 9.11361694e-01
4.07279074e-01 -3.43042046e-01 6.66956902e-01 5.59265092e-02
-5.41316271e-01 -7.09926113e-02 -1.13700783e+00 9.58945870e-01
1.18910110e+00 -4.52205837e-01 -1.06441295e+00 3.69928896e-01
1.00512791e+00 -4.41288710e-01 -4.36397791e-01 7.19744146e-01
2.48254821e-01 -1.20609963e+00 6.30944610e-01 -9.48817372e-01
8.51185679e-01 -1.51912332e-01 -2.92201310e-01 -1.67834306e+00
2.99900711e-01 -2.71493256e-01 2.88870603e-01 1.30071557e+00
9.90834355e-01 -6.46079957e-01 1.18743920e+00 1.43367752e-01
-4.75184947e-01 -1.06736553e+00 -8.28797460e-01 -1.03729331e+00
5.39704204e-01 -3.87973547e-01 5.73624253e-01 7.35010445e-01
1.26678452e-01 7.78884113e-01 2.80504733e-01 -3.50640208e-01
3.73806179e-01 2.03132868e-01 9.54720378e-01 -8.56029332e-01
-5.03644466e-01 -5.39854348e-01 -9.09773186e-02 -1.15257895e+00
3.17515343e-01 -1.11680305e+00 2.11554617e-01 -1.31123698e+00
2.69530714e-01 -3.48664820e-01 4.39448267e-01 3.52527082e-01
-3.86000156e-01 1.31447554e-01 1.26940534e-01 -2.49789521e-01
-1.01993665e-01 5.46085060e-01 9.47516203e-01 -2.71881908e-01
-2.44563714e-01 -3.47238630e-02 -4.47035640e-01 4.13513631e-01
8.55736375e-01 -4.29079682e-01 -7.13445187e-01 -2.56207675e-01
1.99483201e-01 -3.08013767e-01 2.59738773e-01 -1.04980743e+00
-2.56206542e-01 1.04231589e-01 1.81259051e-01 -8.62716556e-01
1.93145752e-01 -5.82214713e-01 7.15580434e-02 5.19652426e-01
-7.94133306e-01 2.06494689e-01 4.35215592e-01 3.54260355e-01
-3.96038711e-01 -7.62594223e-01 7.45118141e-01 -1.95062622e-01
-3.41162495e-02 -1.98322937e-01 -2.77643263e-01 5.52381039e-01
9.90425169e-01 -6.40401393e-02 -3.15612376e-01 -2.45588765e-01
-6.97611690e-01 4.76625003e-02 7.43795395e-01 3.31004053e-01
3.59510988e-01 -1.16998625e+00 -8.33489656e-01 3.96097958e-01
2.49826029e-01 4.66125220e-01 -1.13433972e-01 4.29546624e-01
-7.06480086e-01 2.83270478e-01 -2.08433103e-02 -7.32834697e-01
-1.11299086e+00 6.73112214e-01 3.80903155e-01 -2.92504072e-01
-2.87594318e-01 9.05942500e-01 3.13873261e-01 -9.63703454e-01
1.69271797e-01 -7.78018594e-01 2.98886627e-01 -4.69093248e-02
1.49811208e-01 2.88589448e-02 2.60436624e-01 -1.24303766e-01
5.87632740e-03 2.56237537e-01 -3.03020865e-01 -6.16413951e-01
1.25425124e+00 3.62441063e-01 -6.52607307e-02 6.63756609e-01
1.20382988e+00 5.42533770e-02 -8.21587801e-01 -5.11300527e-02
3.23168367e-01 -2.02323288e-01 -7.87480533e-01 -1.00707233e+00
-6.10104144e-01 7.96261787e-01 1.24568924e-01 5.39455116e-01
9.34596777e-01 6.16692454e-02 7.73135483e-01 4.29906547e-01
1.91700459e-01 -8.30580652e-01 2.32613668e-01 5.21045446e-01
1.18203723e+00 -9.68456507e-01 -2.21179739e-01 -5.06556153e-01
-7.62358308e-01 1.00858879e+00 8.03269923e-01 -2.09512338e-01
7.47971162e-02 2.43198618e-01 2.06783097e-02 1.22683637e-01
-8.65775883e-01 1.59721687e-01 9.17093456e-02 4.81416434e-01
5.36523938e-01 -7.36722872e-02 -2.62404591e-01 5.19337416e-01
-8.14142644e-01 -2.92796105e-01 4.25893664e-01 9.40573454e-01
-1.92411616e-01 -1.57515788e+00 -1.09896883e-01 6.27766013e-01
-2.77377039e-01 -3.81948739e-01 -9.54119325e-01 1.07589841e+00
1.10052831e-01 6.51220083e-01 5.05125113e-02 -4.39200729e-01
4.86815155e-01 6.60017431e-01 6.01848304e-01 -9.93897736e-01
-8.02533031e-01 -3.02119702e-01 4.15678799e-01 -3.16378564e-01
-1.07415766e-02 -8.54595840e-01 -1.22404218e+00 7.23685846e-02
-3.12108267e-02 2.03417003e-01 7.36808717e-01 7.55690694e-01
2.48691160e-02 6.74891353e-01 2.88811684e-01 -9.21780467e-01
-1.15043080e+00 -1.41045511e+00 -2.60456145e-01 8.39437008e-01
1.90087259e-01 -1.40353516e-01 -5.30505896e-01 5.52427955e-02] | [11.071906089782715, 8.96293830871582] |
49a24163-0cb3-428d-9e68-ecfc933d57f0 | visithers-visible-thermal-infrared-stereo | 2304.11291 | null | https://arxiv.org/abs/2304.11291v1 | https://arxiv.org/pdf/2304.11291v1.pdf | VisiTherS: Visible-thermal infrared stereo disparity estimation of human silhouette | This paper presents a novel approach for visible-thermal infrared stereoscopy, focusing on the estimation of disparities of human silhouettes. Visible-thermal infrared stereo poses several challenges, including occlusions and differently textured matching regions in both spectra. Finding matches between two spectra with varying colors, textures, and shapes adds further complexity to the task. To address the aforementioned challenges, this paper proposes a novel approach where a high-resolution convolutional neural network is used to better capture relationships between the two spectra. To do so, a modified HRNet backbone is used for feature extraction. This HRNet backbone is capable of capturing fine details and textures as it extracts features at multiple scales, thereby enabling the utilization of both local and global information. For matching visible and thermal infrared regions, our method extracts features on each patch using two modified HRNet streams. Features from the two streams are then combined for predicting the disparities by concatenation and correlation. Results on public datasets demonstrate the effectiveness of the proposed approach by improving the results by approximately 18 percentage points on the $\leq$ 1 pixel error, highlighting its potential for improving accuracy in this task. The code of VisiTherS is available on GitHub at the following link https://github.com/philippeDG/VisiTherS. | ['Wassim Bouachir', 'Guillaume-Alexandre Bilodeau', 'Philippe Duplessis-Guindon', 'Noreen Anwar'] | 2023-04-22 | null | null | null | null | ['disparity-estimation'] | ['computer-vision'] | [ 5.08551419e-01 -3.96170169e-01 -4.34803106e-02 -1.87901482e-01
-8.62977147e-01 -4.24597353e-01 2.38306940e-01 5.00126556e-03
-2.70252883e-01 4.12891626e-01 1.98571280e-01 -6.48726150e-02
-8.51464923e-03 -8.33050072e-01 -5.06537557e-01 -8.33601534e-01
1.29843310e-01 -2.53716350e-01 1.10381424e-01 -1.47334486e-01
2.28457704e-01 6.48027241e-01 -1.87585497e+00 4.28678572e-01
7.81162500e-01 1.44155324e+00 2.54711747e-01 6.52522445e-01
2.11272895e-01 4.00593102e-01 -1.28629804e-01 -2.81244189e-01
9.55468714e-01 -1.29715111e-02 -3.49024117e-01 -3.77293713e-02
1.11807489e+00 -6.40181661e-01 -3.70731145e-01 1.02144992e+00
6.49039745e-01 1.44521236e-01 3.58012289e-01 -7.73678243e-01
-3.73942345e-01 -1.70754775e-01 -7.35073864e-01 2.06723452e-01
5.90535223e-01 3.35996300e-01 1.00809467e+00 -8.56800139e-01
3.41258615e-01 9.40896988e-01 7.61149704e-01 1.46928236e-01
-1.13060117e+00 -8.53508592e-01 -1.99840516e-01 3.07399690e-01
-1.33920968e+00 -4.77215946e-01 8.13300133e-01 -4.50495690e-01
8.05545628e-01 4.50112134e-01 6.61893964e-01 9.54929054e-01
8.94920826e-02 3.99630398e-01 1.42927217e+00 -3.09513956e-01
-1.31419674e-01 5.53497747e-02 -1.51815891e-01 6.91380382e-01
1.62102625e-01 5.24879158e-01 -7.63032973e-01 -1.61307827e-01
7.64921248e-01 2.68373132e-01 -4.48386312e-01 -1.92779034e-01
-1.12583482e+00 4.77445334e-01 6.03913665e-01 6.48446754e-02
-3.26149136e-01 -9.04728547e-02 9.90655199e-02 2.46223152e-01
7.49798775e-01 3.07089359e-01 -2.58786142e-01 2.38732815e-01
-7.50073850e-01 6.11427948e-02 3.19192022e-01 6.92266166e-01
9.75141287e-01 -3.45768332e-01 -2.71329470e-02 9.91842508e-01
1.08851537e-01 8.14067304e-01 3.28714103e-02 -9.01360691e-01
5.20736277e-01 5.34550250e-01 1.57792315e-01 -1.25080037e+00
-4.48525965e-01 -3.97200704e-01 -8.74448359e-01 4.55720901e-01
3.95740122e-01 -1.40388384e-01 -7.32276917e-01 1.26401162e+00
6.80637419e-01 7.80554116e-02 -2.47306004e-01 1.26904702e+00
8.95902812e-01 5.66007614e-01 -2.57617891e-01 1.67260185e-01
1.48351836e+00 -7.12407053e-01 -1.79766342e-01 -1.52234107e-01
1.71171248e-01 -1.16176891e+00 7.31589258e-01 3.45470548e-01
-9.13063169e-01 -7.64639854e-01 -9.45480466e-01 -3.05551499e-01
-3.86811644e-01 2.94092715e-01 3.92244041e-01 5.57835996e-01
-1.11530244e+00 6.71992898e-01 -4.91970181e-01 -4.68051612e-01
2.00275809e-01 3.05095136e-01 -1.99172705e-01 -2.85063297e-01
-9.74788368e-01 4.89529014e-01 6.02854565e-02 3.71115863e-01
-7.85917640e-02 -7.93527246e-01 -9.15854692e-01 -3.58450823e-02
2.04214633e-01 -6.41316533e-01 6.82964981e-01 -1.04925323e+00
-1.38134313e+00 8.24451864e-01 -3.53802204e-01 -2.63625775e-02
5.51284015e-01 -1.57222182e-01 -4.93228853e-01 4.18932617e-01
1.12132169e-01 5.06177485e-01 9.88835514e-01 -1.05416381e+00
-9.25536513e-01 -5.52586496e-01 4.02288288e-02 3.97355050e-01
-2.98718184e-01 -2.56883800e-02 -5.27457952e-01 -6.16422653e-01
2.51950651e-01 -7.78721213e-01 6.57628104e-02 2.75942296e-01
-3.54670078e-01 1.60647109e-01 6.19727969e-01 -8.96600604e-01
7.50433087e-01 -2.23235750e+00 -2.99265236e-01 3.03723752e-01
3.33334208e-01 1.89181268e-01 -1.39371082e-01 3.55922610e-01
-3.00813038e-02 -3.04892600e-01 -3.83029208e-02 -1.77923232e-01
-1.52298436e-01 -5.79559505e-01 -3.08542177e-02 7.97432542e-01
1.10525839e-01 6.27077162e-01 -6.15624428e-01 -2.64758110e-01
5.99756718e-01 7.47921765e-01 -9.64318067e-02 1.35471523e-01
3.79720628e-02 5.79374075e-01 -2.49701366e-01 8.76358211e-01
1.10504913e+00 -1.24040455e-01 1.21749394e-01 -5.18553019e-01
-4.47920173e-01 5.98590001e-02 -1.21137500e+00 1.36317706e+00
-4.57368851e-01 7.50783861e-01 1.11125447e-02 -6.44283473e-01
1.15930319e+00 7.22024068e-02 7.09066749e-01 -1.20522892e+00
2.11262666e-02 2.92281330e-01 -5.38379848e-01 -4.22882408e-01
5.31122625e-01 7.80329332e-02 2.70583123e-01 3.40010583e-01
-5.08611500e-01 7.46118948e-02 -1.01257928e-01 -3.87946427e-01
7.56985307e-01 2.35551015e-01 1.40021652e-01 -9.04914886e-02
6.06289029e-01 -1.63929188e-03 5.03085256e-01 5.72125375e-01
-1.53091609e-01 7.90557444e-01 1.19346380e-02 -7.40192235e-01
-1.13306487e+00 -9.14158165e-01 -2.08084285e-01 9.54808652e-01
4.67266023e-01 2.24443469e-02 -3.61478418e-01 -1.86171606e-01
2.43259639e-01 8.77724290e-02 -5.37994802e-01 2.43038431e-01
-4.30619597e-01 -3.44926566e-01 1.45108506e-01 4.29524243e-01
8.34110081e-01 -6.55140698e-01 -7.36347675e-01 -2.29117602e-01
-3.12026918e-01 -1.21245944e+00 -4.68134910e-01 -2.09815130e-01
-8.28900397e-01 -1.25739110e+00 -8.33144784e-01 -5.98724008e-01
4.82249618e-01 9.37952459e-01 7.55010128e-01 1.36468798e-01
-7.08027601e-01 2.96382725e-01 -2.07065821e-01 -1.70993149e-01
1.23587184e-01 4.66604494e-02 -3.75786602e-01 2.77766466e-01
4.34282809e-01 -4.73955005e-01 -1.22124767e+00 3.83054942e-01
-7.40021408e-01 2.83662528e-01 4.52661216e-01 6.77185774e-01
5.60282707e-01 -6.17142394e-02 -1.35827467e-01 -4.59890902e-01
1.39215767e-01 -2.95822442e-01 -9.72307980e-01 4.31179553e-02
-4.74258274e-01 -2.32189372e-01 6.77083194e-01 6.36677723e-03
-1.18282986e+00 2.09417820e-01 1.60737246e-01 -4.11590755e-01
-3.91029894e-01 1.20198406e-01 3.06718230e-01 -3.19492161e-01
6.50210857e-01 1.58881113e-01 7.57444799e-02 -5.64706624e-01
1.08866468e-01 8.44781935e-01 5.23221076e-01 -2.31229797e-01
9.84341145e-01 9.72022533e-01 1.05707735e-01 -9.88392770e-01
-7.26594508e-01 -1.00997806e+00 -6.88427150e-01 -4.73421544e-01
8.41595888e-01 -1.10998917e+00 -7.94258356e-01 5.53501904e-01
-8.79277527e-01 -2.53972441e-01 1.62768826e-01 5.94235897e-01
-2.64135122e-01 5.10040522e-01 -6.05202854e-01 -7.68694937e-01
-5.81240118e-01 -9.40425813e-01 1.45822120e+00 2.50445157e-01
2.26295106e-02 -9.20946002e-01 1.47089846e-02 7.38858640e-01
5.33364892e-01 4.31883186e-01 6.43607378e-01 5.70975952e-02
-8.49607170e-01 -2.72132158e-01 -6.19012594e-01 1.41588777e-01
2.20180839e-01 5.32295147e-04 -1.35606289e+00 -4.28136855e-01
-1.39323175e-01 -1.44557819e-01 9.57197845e-01 4.62576866e-01
9.19203997e-01 1.00489818e-01 -1.11330837e-01 1.04581892e+00
1.78189015e+00 -9.24110040e-02 6.69353187e-01 6.36468709e-01
7.71257997e-01 7.96399772e-01 7.62463033e-01 5.71428835e-01
2.17750669e-01 8.76941621e-01 4.32325572e-01 -6.21611834e-01
-2.22675800e-01 1.21962540e-01 2.12808937e-01 2.83714533e-01
-2.83654243e-01 1.20247707e-01 -9.01881158e-01 3.59223425e-01
-1.58789837e+00 -1.01821804e+00 -2.66964406e-01 2.52048945e+00
3.79161119e-01 -3.64330739e-01 1.03860050e-01 -2.93189865e-02
9.58019316e-01 4.48985457e-01 -5.24699092e-01 -4.82975915e-02
-2.07046613e-01 3.91919672e-01 8.60295951e-01 4.48850155e-01
-1.13515091e+00 6.47508204e-01 5.46330976e+00 5.28523386e-01
-1.60752857e+00 -1.97824806e-01 7.19207346e-01 -2.28918716e-01
-1.04864679e-01 -1.75172284e-01 -5.79606056e-01 3.58385175e-01
4.75313753e-01 1.28535256e-01 6.22887552e-01 3.41855735e-01
5.43070734e-01 -4.68767643e-01 -6.16960227e-01 1.17165792e+00
1.92779422e-01 -1.09669363e+00 -3.05472553e-01 1.47643656e-01
7.74258196e-01 2.38276318e-01 4.18653339e-01 -4.71315920e-01
-9.52469334e-02 -9.49210346e-01 5.79561949e-01 5.74664056e-01
1.07501924e+00 -8.31699789e-01 7.03950286e-01 -1.10153578e-01
-1.52756703e+00 -1.53596878e-01 -4.51198220e-01 1.00881636e-01
-3.02555233e-01 7.31683254e-01 -6.92155302e-01 8.52110744e-01
9.57946837e-01 8.07228506e-01 -5.76017857e-01 1.05144322e+00
1.58975899e-01 -1.77483819e-02 -2.71210521e-01 2.25170836e-01
-1.86678097e-01 -4.00999725e-01 3.41615468e-01 1.17918730e+00
4.93712544e-01 -6.79256991e-02 2.06385583e-01 8.75534952e-01
1.14728354e-01 1.59173235e-01 -6.05283141e-01 3.30787033e-01
2.67153651e-01 1.46958959e+00 -5.08760393e-01 -1.77214846e-01
-5.77531695e-01 9.75278556e-01 1.29131645e-01 5.21096110e-01
-7.69784272e-01 -4.86731946e-01 7.77791083e-01 1.63807869e-01
3.77488881e-01 -3.06368500e-01 -3.72229278e-01 -1.11694241e+00
2.63583422e-01 -8.55381906e-01 4.07714963e-01 -8.61591101e-01
-1.28007770e+00 5.23115754e-01 -3.15545082e-01 -1.55577362e+00
2.21535757e-01 -6.69322073e-01 -3.51135194e-01 1.23633099e+00
-1.90274000e+00 -1.13116324e+00 -1.01278603e+00 7.14206398e-01
3.72415036e-01 1.75131649e-01 4.80455160e-01 2.67393708e-01
-5.01558185e-01 3.12585145e-01 4.00377065e-01 -1.64563805e-02
9.47827399e-01 -1.00295496e+00 4.88337487e-01 8.43031883e-01
-2.82579541e-01 3.33051413e-01 4.54255164e-01 -5.28199136e-01
-1.48876357e+00 -1.14395022e+00 6.34478748e-01 -1.01855993e-01
3.09714168e-01 -2.27904946e-01 -6.85478151e-01 1.31909460e-01
1.91735420e-02 -8.94201745e-04 5.50934672e-01 -1.24039181e-01
-5.52215815e-01 -4.49609220e-01 -1.11393130e+00 5.12857437e-01
9.38852191e-01 -9.16123748e-01 6.33955970e-02 2.82169372e-01
2.62099117e-01 -5.15441895e-01 -1.00728047e+00 3.62029076e-01
8.96885216e-01 -1.53615630e+00 1.06276667e+00 1.96052700e-01
5.63814461e-01 -2.94457078e-01 -1.40169784e-01 -9.69061077e-01
-3.36805552e-01 -4.83340114e-01 3.42723608e-01 7.31410801e-01
2.66882539e-01 -1.02273202e+00 7.51828134e-01 5.30049503e-01
2.34289199e-01 -6.37528300e-01 -6.97967708e-01 -6.15096569e-01
-4.23790693e-01 -3.15253675e-01 5.61227798e-01 8.39488447e-01
-3.03868622e-01 -6.48976713e-02 -3.14358115e-01 5.56384742e-01
8.46382260e-01 4.46267366e-01 7.72859573e-01 -1.12244344e+00
-1.66317433e-01 -3.61171603e-01 -3.65560502e-01 -9.53512490e-01
-9.19097885e-02 -7.90330708e-01 -2.93405801e-02 -1.35098362e+00
2.31017709e-01 -3.95695806e-01 -1.07160732e-01 1.57647818e-01
-2.66375035e-01 6.54099405e-01 2.13691577e-01 3.77613723e-01
-2.45462924e-01 2.55936712e-01 1.11156785e+00 -1.17110968e-01
-1.58235952e-01 -7.14479908e-02 -4.72208709e-01 3.79760802e-01
1.06093299e+00 -6.91780448e-02 -9.38830599e-02 -5.13565183e-01
-1.60736740e-02 -1.16039207e-02 6.80683315e-01 -1.28514993e+00
2.29703546e-01 -1.38536870e-01 6.87397540e-01 -4.73297954e-01
5.03781855e-01 -9.40557003e-01 3.52316529e-01 3.70935529e-01
-2.21022531e-01 -2.32898965e-02 1.49427831e-01 5.85382164e-01
-7.67310560e-02 2.99672157e-01 9.24520612e-01 -9.73554030e-02
-6.85758948e-01 3.67626548e-01 1.44618630e-01 -4.26088870e-01
8.92404556e-01 -5.17697632e-01 -5.38733006e-01 -4.03799742e-01
-2.71918714e-01 -1.51391923e-02 8.67826343e-01 2.21506715e-01
6.24721766e-01 -1.07516634e+00 -6.13149464e-01 4.06548858e-01
2.89727241e-01 -2.54087687e-01 5.85086107e-01 1.01437175e+00
-7.88972020e-01 3.86132926e-01 -3.08191746e-01 -7.14408398e-01
-1.61588168e+00 2.08704367e-01 6.01177275e-01 1.34445116e-01
-6.52605176e-01 5.82678735e-01 1.83642045e-01 -2.59510934e-01
1.44812286e-01 -2.49988317e-01 -1.40151724e-01 -7.54820034e-02
5.53870320e-01 6.72154069e-01 1.91393197e-01 -6.91240013e-01
-4.85009626e-02 9.24075186e-01 2.02727541e-01 7.78511465e-02
1.37269807e+00 -4.07336831e-01 -2.32575879e-01 1.90959111e-01
1.41127360e+00 1.40466228e-01 -1.29466629e+00 -4.50539351e-01
-3.97169203e-01 -1.00690520e+00 2.47256055e-01 -4.47603911e-01
-1.23768985e+00 8.56856287e-01 9.45303321e-01 1.21272437e-01
1.40898800e+00 -2.31973618e-01 1.00992811e+00 5.95429204e-02
2.21290603e-01 -9.85315204e-01 -1.09582216e-01 1.95740014e-01
7.08465755e-01 -1.32494724e+00 1.51700974e-01 -6.83123648e-01
-3.72707188e-01 1.45089209e+00 5.10982335e-01 4.37769964e-02
3.01791906e-01 -6.51374236e-02 3.37967366e-01 -3.21558207e-01
-4.04564291e-01 -4.56904322e-01 4.62004393e-01 6.50965154e-01
5.76189041e-01 -4.63438481e-02 6.60024434e-02 -5.27240574e-01
-1.98826101e-02 -1.78719148e-01 1.57448113e-01 7.55277276e-01
-4.73656654e-01 -8.08109581e-01 -7.08160400e-01 4.55899119e-01
-2.72900790e-01 -3.23076934e-01 -3.42859536e-01 4.61301237e-01
8.38127583e-02 1.00792480e+00 9.96238291e-02 -5.33103824e-01
3.22404355e-01 -2.99516588e-01 2.64496505e-01 -2.49053538e-01
-5.30929446e-01 2.41816938e-01 1.64339915e-01 -8.90661716e-01
-5.90466857e-01 -4.90604520e-01 -6.28269136e-01 -5.89248657e-01
-1.42594725e-02 -3.90635341e-01 6.50179923e-01 4.27765310e-01
5.09991705e-01 4.35630493e-02 1.03946662e+00 -1.11131132e+00
-3.66844147e-01 -6.10757351e-01 -7.45696425e-01 5.79231381e-01
6.17915094e-01 -4.46297646e-01 -4.70143527e-01 -1.26906708e-01] | [10.150654792785645, -2.457443952560425] |
b8c969e5-0500-402b-919d-08a32b6f2a0e | high-dynamic-range-imaging-with-context-aware | 2304.04416 | null | https://arxiv.org/abs/2304.04416v4 | https://arxiv.org/pdf/2304.04416v4.pdf | High Dynamic Range Imaging with Context-aware Transformer | Avoiding the introduction of ghosts when synthesising LDR images as high dynamic range (HDR) images is a challenging task. Convolutional neural networks (CNNs) are effective for HDR ghost removal in general, but are challenging to deal with the LDR images if there are large movements or oversaturation/undersaturation. Existing dual-branch methods combining CNN and Transformer omit part of the information from non-reference images, while the features extracted by the CNN-based branch are bound to the kernel size with small receptive field, which are detrimental to the deblurring and the recovery of oversaturated/undersaturated regions. In this paper, we propose a novel hierarchical dual Transformer method for ghost-free HDR (HDT-HDR) images generation, which extracts global features and local features simultaneously. First, we use a CNN-based head with spatial attention mechanisms to extract features from all the LDR images. Second, the LDR features are delivered to the Hierarchical Dual Transformer (HDT). In each Dual Transformer (DT), the global features are extracted by the window-based Transformer, while the local details are extracted using the channel attention mechanism with deformable CNNs. Finally, the ghost free HDR image is obtained by dimensional mapping on the HDT output. Abundant experiments demonstrate that our HDT-HDR achieves the state-of-the-art performance among existing HDR ghost removal methods. | ['Zhenming Fu', 'Dan Zhang', 'Fangfang Zhou'] | 2023-04-10 | null | null | null | null | ['deblurring'] | ['computer-vision'] | [ 1.21395960e-01 -1.50439948e-01 8.81092623e-02 1.93866089e-01
-5.22648752e-01 -2.36266702e-01 3.38468999e-01 -6.26026928e-01
4.71317507e-02 7.23293602e-01 3.13307971e-01 6.74911961e-02
2.69778907e-01 -6.93776608e-01 -5.48080385e-01 -1.21962845e+00
9.87372026e-02 -1.67174116e-01 6.01579487e-01 -4.91794646e-01
1.82911217e-01 8.58775854e-01 -1.50578403e+00 3.08461845e-01
1.08572090e+00 1.13774204e+00 4.40054029e-01 7.17669606e-01
1.18182413e-01 1.04047978e+00 -6.91087544e-01 2.88131148e-01
3.91380966e-01 -3.96498203e-01 -4.22621429e-01 -1.10985547e-01
3.96306038e-01 -8.73251796e-01 -1.06593716e+00 9.23447669e-01
9.15701091e-01 2.54439652e-01 5.59011996e-01 -9.88156021e-01
-1.21355927e+00 1.60314992e-01 -9.04905319e-01 5.09653628e-01
3.05972081e-02 5.35521746e-01 2.99564809e-01 -1.07295859e+00
6.40311301e-01 1.36526537e+00 3.50511491e-01 6.42116725e-01
-1.10553169e+00 -1.08434343e+00 1.11831147e-02 2.26936370e-01
-1.53451312e+00 -5.88602960e-01 7.57724881e-01 -1.04301006e-01
1.17883575e+00 3.79789799e-01 5.96313536e-01 6.53464079e-01
4.55128998e-01 4.80739206e-01 1.26862335e+00 -5.38218096e-02
-2.41379708e-01 -2.99597025e-01 -2.67016981e-02 5.38801908e-01
7.51142055e-02 5.49315095e-01 -4.47444588e-01 2.88161397e-01
1.30198157e+00 1.40155226e-01 -7.15837657e-01 1.09471112e-01
-1.28262973e+00 3.78479004e-01 6.55796885e-01 1.96289286e-01
-1.62170067e-01 8.26913342e-02 1.96173653e-01 3.15158874e-01
6.20299757e-01 1.94596916e-01 -1.66667998e-01 5.04733205e-01
-9.53878403e-01 -2.89432122e-03 1.83530539e-01 9.90857959e-01
8.28673303e-01 2.66239941e-01 -6.38462901e-01 9.42064404e-01
4.12097983e-02 7.67730415e-01 2.71595359e-01 -7.95470715e-01
3.36641163e-01 4.70295966e-01 6.23304769e-02 -1.00850904e+00
-2.53271073e-01 -2.54378080e-01 -1.62012327e+00 3.66616696e-01
-5.82315624e-02 2.34468758e-01 -1.34392691e+00 1.23788583e+00
2.29691580e-01 1.21799432e-01 1.27786309e-01 1.30935800e+00
1.25751984e+00 1.23114049e+00 -1.94082603e-01 -5.28910398e-01
1.13361335e+00 -1.02831733e+00 -1.03570712e+00 3.37032788e-02
8.84637833e-02 -8.26501667e-01 7.58551300e-01 2.45535329e-01
-1.05535412e+00 -7.46279597e-01 -1.15680659e+00 -5.94378352e-01
-3.59194249e-01 2.00332657e-01 1.85591087e-03 1.42725781e-01
-1.17090404e+00 5.29347122e-01 -2.78808504e-01 2.20581125e-02
3.58100593e-01 3.86744022e-01 -2.50475794e-01 -4.84035999e-01
-1.50097299e+00 8.32066059e-01 2.00840294e-01 6.36390269e-01
-1.14890265e+00 -7.55639434e-01 -7.38729596e-01 4.34645377e-02
1.47216886e-01 -4.10286725e-01 3.97754163e-01 -6.98423207e-01
-1.54473269e+00 7.15276778e-01 9.25757512e-02 -3.56948338e-02
4.33033437e-01 4.38502841e-02 -5.63536644e-01 1.00170709e-01
-1.17138341e-01 7.56758094e-01 1.36463773e+00 -1.36735284e+00
-5.21427631e-01 -1.85886443e-01 -2.00711221e-01 3.88714075e-01
-7.44294599e-02 9.64299515e-02 -4.50080961e-01 -8.68869305e-01
1.51119038e-01 -7.30437279e-01 -1.30336406e-02 7.43838772e-03
-5.89731276e-01 1.76803797e-01 1.44138539e+00 -9.79755521e-01
1.45703697e+00 -2.32959342e+00 2.16997996e-01 -1.05131283e-01
7.29546010e-01 5.44319808e-01 -3.09094101e-01 4.31189388e-02
-3.97858590e-01 1.00729287e-01 -1.01667427e-01 1.66798130e-01
-3.86899352e-01 -1.19649805e-01 -6.81318164e-01 7.41883099e-01
2.50525445e-01 1.00561726e+00 -6.20308995e-01 -5.88581383e-01
7.27032721e-01 7.11629033e-01 5.64848026e-03 4.97914314e-01
1.04521766e-01 5.50760567e-01 -2.32999668e-01 6.61703289e-01
1.21043575e+00 -1.19418800e-01 -1.91362068e-01 -9.95385647e-01
-5.12188077e-01 -1.96109235e-01 -8.81771207e-01 1.07616830e+00
-2.91999161e-01 8.61530423e-01 4.40410078e-02 -5.10840058e-01
1.03107762e+00 8.43012854e-02 3.94093126e-01 -1.22756886e+00
-1.11484025e-02 2.27799401e-01 -2.31652856e-01 -4.59757388e-01
5.63135028e-01 6.73469063e-03 5.81866913e-02 3.66261840e-01
-1.95809707e-01 -5.18785566e-02 -2.13108838e-01 1.78256065e-01
1.03513336e+00 -1.34726301e-01 1.07843079e-01 -3.72965664e-01
6.33667350e-01 -3.21420312e-01 6.26833916e-01 5.09792924e-01
-1.45117536e-01 9.68591869e-01 2.36098826e-01 -6.91048145e-01
-1.48738396e+00 -1.07775128e+00 -5.68569861e-02 6.82404816e-01
6.23665154e-01 -1.10321771e-02 -3.76751125e-01 -3.80205035e-01
-2.74190485e-01 2.73860246e-01 -5.49768090e-01 -4.19826329e-01
-9.86713648e-01 -9.64509130e-01 4.73266810e-01 2.69994020e-01
1.00124657e+00 -1.24300635e+00 -4.19271946e-01 2.66117871e-01
-4.07330900e-01 -1.04501247e+00 -8.47038686e-01 9.24693123e-02
-7.61289358e-01 -8.71664464e-01 -1.26871407e+00 -7.36594141e-01
6.49523318e-01 7.33425796e-01 9.04211700e-01 2.85742462e-01
-6.13251686e-01 -2.94295490e-01 -3.29757690e-01 2.79388368e-01
-2.77667969e-01 -1.57290921e-01 -1.85678050e-01 -5.89412823e-02
-1.14217095e-01 -3.57584625e-01 -9.58904862e-01 5.15782714e-01
-1.07240415e+00 4.23729718e-01 7.62784719e-01 9.91944849e-01
6.35756493e-01 4.15686071e-01 2.86542267e-01 -3.56256872e-01
5.22117615e-01 -7.16903508e-02 -5.71335852e-01 1.83235228e-01
-4.16956276e-01 -5.62620796e-02 8.29119682e-01 -7.76181817e-01
-1.24163210e+00 -1.89686462e-01 2.05195650e-01 -9.07204986e-01
1.04082488e-01 -4.03771728e-01 -3.62657428e-01 -4.64822888e-01
4.21016425e-01 6.78891718e-01 -1.51181981e-01 -1.91642582e-01
3.10656846e-01 7.13190377e-01 5.22951901e-01 -2.37678766e-01
1.08779979e+00 6.02913380e-01 -1.15770318e-01 -1.05351961e+00
-5.38068116e-01 -1.91576451e-01 -3.26891124e-01 -4.61121649e-01
1.15182352e+00 -9.82838452e-01 -6.15338862e-01 1.16921020e+00
-1.10702145e+00 -7.09813714e-01 -2.11973414e-01 9.82598290e-02
-3.64551187e-01 3.97415727e-01 -1.03414202e+00 -6.01306200e-01
-6.82677031e-01 -1.36312819e+00 1.11787152e+00 5.22507608e-01
4.41990346e-01 -5.23900628e-01 -2.56570101e-01 1.34625837e-01
8.54931056e-01 4.12841849e-02 1.14362824e+00 8.93284976e-02
-1.21668065e+00 3.88101310e-01 -9.01215315e-01 3.38583946e-01
2.25382179e-01 -9.99377221e-02 -1.01893461e+00 -4.90733445e-01
-9.54098403e-02 -1.51393399e-01 1.04092801e+00 6.59732044e-01
1.16120231e+00 -3.56010944e-01 -3.36995661e-01 1.02152467e+00
1.48869681e+00 2.73620963e-01 1.40711963e+00 3.07962000e-01
1.08314276e+00 2.90761739e-01 7.51699626e-01 1.81642160e-01
5.96426427e-02 8.71055543e-01 2.64892906e-01 -8.82733405e-01
-8.48586380e-01 6.67581037e-02 5.84357083e-01 8.27185035e-01
-1.11148776e-02 -3.52407664e-01 -4.03118819e-01 3.70854527e-01
-1.51601577e+00 -9.02769625e-01 -2.62396961e-01 2.15618658e+00
8.48765314e-01 -2.56744385e-01 -3.28361332e-01 -1.90000892e-01
1.08003211e+00 5.56047916e-01 -7.64627635e-01 -2.47278437e-02
-5.33612967e-01 -6.96953833e-02 8.72919440e-01 3.97878528e-01
-8.44777286e-01 9.82839048e-01 6.02380085e+00 1.10156310e+00
-1.26733661e+00 6.49852008e-02 8.43331039e-01 -8.27487782e-02
-1.73280582e-01 -1.97481260e-01 -8.79596472e-01 6.34984851e-01
4.01881188e-01 2.29605049e-01 7.53204346e-01 1.48529232e-01
3.49668711e-01 -1.31700128e-01 -5.15571296e-01 1.32155311e+00
2.23054394e-01 -1.12706757e+00 1.04749732e-01 1.31110877e-01
9.90617156e-01 -1.49445504e-01 3.96835148e-01 1.65287137e-01
2.28142336e-01 -1.13673890e+00 5.43808222e-01 5.69334924e-01
1.48747396e+00 -7.93858290e-01 6.06296778e-01 -2.49582902e-02
-1.53758156e+00 -2.01127395e-01 -6.40317440e-01 6.98283255e-01
1.43784404e-01 9.37156856e-01 -1.06658801e-01 4.55083996e-01
1.06316292e+00 7.65999734e-01 -5.43463886e-01 7.19706774e-01
2.17184424e-02 1.36796702e-02 -5.61022796e-02 5.24069965e-01
-1.12707943e-01 -2.18783230e-01 7.43076921e-01 1.09419906e+00
4.30423498e-01 4.70237345e-01 -1.78817749e-01 1.18014944e+00
-1.27731249e-01 -2.50719666e-01 -6.37081802e-01 2.39684463e-01
2.83803463e-01 1.41146231e+00 -6.15610838e-01 -4.80602711e-01
-2.00522766e-01 1.40036607e+00 -2.44554933e-02 7.34898388e-01
-9.81320083e-01 -7.74066746e-01 5.11000752e-01 2.21898243e-01
3.42667848e-01 8.06830004e-02 1.12760797e-01 -1.34067535e+00
-1.94077998e-01 -9.46504474e-01 6.58834055e-02 -1.31921113e+00
-1.34438479e+00 6.51112795e-01 -1.81840286e-01 -1.41231155e+00
4.40587670e-01 -3.27730536e-01 -5.21873236e-01 1.10166693e+00
-1.98711574e+00 -1.08935356e+00 -8.22078884e-01 7.71430373e-01
5.83575845e-01 3.36898752e-02 1.29941195e-01 5.42023659e-01
-7.53629684e-01 2.60156959e-01 6.62989840e-02 8.60233903e-02
1.15537286e+00 -8.71927381e-01 2.78853118e-01 9.04286385e-01
-9.09681261e-01 3.77507210e-01 4.83241707e-01 -1.01395369e+00
-1.35127985e+00 -1.46134186e+00 3.82901549e-01 8.46264660e-02
1.55314818e-01 -3.18524927e-01 -1.20738125e+00 3.98984879e-01
1.57743573e-01 4.31981653e-01 -8.58082324e-02 -7.04826832e-01
-3.03110898e-01 -4.31328803e-01 -1.25990701e+00 5.24483681e-01
1.01405513e+00 -5.91153681e-01 -1.76263630e-01 -6.49314327e-03
1.00026643e+00 -6.77472830e-01 -1.03635502e+00 3.78470033e-01
4.65499938e-01 -1.04607487e+00 1.22158289e+00 1.60772994e-01
5.28659165e-01 -8.71082962e-01 3.59564796e-02 -1.25008714e+00
-5.99795759e-01 -6.52121365e-01 -2.85029292e-01 1.28480577e+00
-4.68525700e-02 -7.19734490e-01 2.77217805e-01 2.25567922e-01
-3.42512816e-01 -4.13092613e-01 -7.37890422e-01 -6.75043821e-01
-9.83722433e-02 3.92122418e-01 6.62696540e-01 9.47579622e-01
-6.02542102e-01 2.30310500e-01 -8.35406959e-01 2.95264035e-01
9.45885956e-01 8.70639831e-02 3.82447988e-01 -8.55448961e-01
1.41070127e-01 -1.48440063e-01 -1.54664397e-01 -1.14494014e+00
-5.15707694e-02 -5.53474307e-01 3.65133673e-01 -1.49865282e+00
4.53921288e-01 -5.36904693e-01 -1.94720104e-01 3.15117657e-01
-1.70801297e-01 5.31983674e-01 2.93847203e-01 4.90014136e-01
-4.37894851e-01 6.90327644e-01 1.86145151e+00 -3.11484307e-01
-4.54755962e-01 -6.46867692e-01 -2.89334953e-01 1.91842347e-01
6.25153124e-01 -2.71255851e-01 -2.12829322e-01 -2.92493880e-01
-3.61739174e-02 2.69942343e-01 6.19700551e-01 -1.09750521e+00
2.29523122e-01 -1.05627567e-01 1.03819609e+00 -8.70344162e-01
2.30692074e-01 -6.87387824e-01 3.99384290e-01 4.02128965e-01
-1.40140623e-01 -1.13784462e-01 1.10740952e-01 4.05830026e-01
-1.30125135e-01 4.49637175e-01 1.36833632e+00 -7.14048371e-02
-7.38625348e-01 7.20861316e-01 -3.70056480e-01 -1.03563547e-01
1.05750644e+00 -3.10872793e-01 -9.79409337e-01 -2.36393675e-01
-4.29377407e-01 1.14788085e-01 6.63375616e-01 4.31608170e-01
1.09139562e+00 -1.53734827e+00 -6.32858753e-01 5.93953967e-01
-8.64531472e-02 1.48133576e-01 9.51622427e-01 1.00975430e+00
-6.07270360e-01 3.69271457e-01 -5.51784396e-01 -3.49029005e-01
-1.17550170e+00 9.05194759e-01 6.72837377e-01 -4.00059491e-01
-1.16280997e+00 3.37627232e-01 9.26912785e-01 6.50230981e-03
-8.95978287e-02 -1.17844440e-01 -2.99438417e-01 -1.14905894e-01
8.26502144e-01 5.50150990e-01 -1.30021870e-01 -7.88386285e-01
-1.20326482e-01 8.78939211e-01 -2.91601449e-01 1.75688282e-01
1.22432506e+00 -6.27260804e-01 -3.30948085e-01 -6.21894896e-02
1.26846898e+00 -1.71044752e-01 -1.51094759e+00 -3.13281491e-02
-9.73365486e-01 -7.33428121e-01 3.40008914e-01 -6.25195086e-01
-1.69744945e+00 9.79532897e-01 9.94432986e-01 3.33667062e-02
1.51053917e+00 -1.31214827e-01 1.23040378e+00 -2.36589879e-01
-2.10639779e-02 -9.41273570e-01 3.02436978e-01 3.42222810e-01
1.07249010e+00 -8.94694984e-01 9.24541503e-02 -4.14112985e-01
-4.71865416e-01 1.28972208e+00 8.97176087e-01 -1.83389619e-01
4.12948668e-01 4.41506356e-01 -1.65763367e-02 -1.32935435e-01
-5.81947684e-01 -1.57987073e-01 1.32676408e-01 9.66130912e-01
-1.47480229e-02 -2.98170090e-01 1.65721878e-01 8.44822824e-02
4.91971999e-01 -3.05002220e-02 6.33107424e-01 4.36596960e-01
-5.26447296e-01 -5.82587063e-01 -5.16532123e-01 4.55621988e-01
-2.77413189e-01 -4.55097944e-01 3.43621336e-02 6.40124321e-01
2.04057381e-01 6.39038503e-01 2.02322498e-01 -5.47957301e-01
3.00097972e-01 -6.64933562e-01 4.18304592e-01 -2.16521069e-01
-4.22015995e-01 3.71490955e-01 -3.12031865e-01 -8.38262200e-01
-2.39588425e-01 -6.15136400e-02 -1.12619758e+00 -6.09278023e-01
-3.76387805e-01 -4.65297222e-01 1.30569667e-01 5.64287007e-01
3.93750340e-01 8.34595025e-01 8.06364179e-01 -1.12741160e+00
1.81388285e-03 -8.05126786e-01 -8.70103240e-01 2.58698344e-01
8.47218633e-01 -5.65210104e-01 -6.17484033e-01 1.67689711e-01] | [10.90865421295166, -2.1398561000823975] |
66ae0abc-adf2-40ec-83fa-3c915fe084f6 | human-trajectory-prediction-using-spatially | 1705.09436 | null | http://arxiv.org/abs/1705.09436v1 | http://arxiv.org/pdf/1705.09436v1.pdf | Human Trajectory Prediction using Spatially aware Deep Attention Models | Trajectory Prediction of dynamic objects is a widely studied topic in the
field of artificial intelligence. Thanks to a large number of applications like
predicting abnormal events, navigation system for the blind, etc. there have
been many approaches to attempt learning patterns of motion directly from data
using a wide variety of techniques ranging from hand-crafted features to
sophisticated deep learning models for unsupervised feature learning. All these
approaches have been limited by problems like inefficient features in the case
of hand crafted features, large error propagation across the predicted
trajectory and no information of static artefacts around the dynamic moving
objects. We propose an end to end deep learning model to learn the motion
patterns of humans using different navigational modes directly from data using
the much popular sequence to sequence model coupled with a soft attention
mechanism. We also propose a novel approach to model the static artefacts in a
scene and using these to predict the dynamic trajectories. The proposed method,
tested on trajectories of pedestrians, consistently outperforms previously
proposed state of the art approaches on a variety of large scale data sets. We
also show how our architecture can be naturally extended to handle multiple
modes of movement (say pedestrians, skaters, bikers and buses) simultaneously. | ['G. Srinivasaraghavan', 'Daksh Varshneya'] | 2017-05-26 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [-1.13708358e-02 -3.31917971e-01 1.42312184e-01 -4.90662575e-01
-5.47974885e-01 -3.59569490e-01 1.12990975e+00 6.74729422e-02
-6.74551129e-01 5.66837013e-01 3.55007589e-01 -2.93208957e-01
-3.62302721e-01 -7.44474232e-01 -7.59106457e-01 -7.49537766e-01
-3.03868562e-01 3.15185159e-01 9.23846066e-01 -4.13355678e-01
4.42365468e-01 5.05773902e-01 -1.80272949e+00 3.84291112e-01
4.81750429e-01 7.27596521e-01 3.09086651e-01 1.03380716e+00
5.60856052e-03 1.11504066e+00 -3.00909251e-01 -3.30544740e-01
2.20682070e-01 -3.19713593e-01 -8.35688412e-01 -2.68194433e-02
2.37267867e-01 -7.88699612e-02 -6.31249011e-01 6.21565819e-01
4.41644162e-01 6.31000102e-01 8.07256639e-01 -1.25754964e+00
-2.16804653e-01 1.86463192e-01 -3.36419284e-01 5.05601466e-01
3.89762819e-01 3.16456288e-01 6.67486012e-01 -6.22766852e-01
6.73596442e-01 1.28522110e+00 7.67016530e-01 4.26100045e-01
-7.89391875e-01 -3.05537935e-02 2.92289168e-01 1.06772268e+00
-1.01664650e+00 -3.95841986e-01 5.74066401e-01 -8.73492122e-01
1.16327035e+00 3.44397038e-01 5.72365165e-01 1.22356081e+00
2.12414816e-01 1.11640203e+00 6.04089320e-01 -2.31117308e-01
1.68654665e-01 3.97740863e-02 3.19377273e-01 6.08042300e-01
-3.00711632e-01 3.08169276e-01 -3.14239860e-02 4.92833517e-02
2.79424697e-01 2.80849785e-01 -9.19848830e-02 -2.40590185e-01
-1.40671790e+00 6.83847308e-01 5.01377463e-01 2.95671135e-01
-3.53544235e-01 1.55778497e-01 5.31097293e-01 1.71201169e-01
8.93262699e-02 -4.45960611e-02 -5.68814993e-01 -4.82040346e-01
-8.61696124e-01 6.34473562e-01 6.06474459e-01 9.38959002e-01
6.37125790e-01 -6.77816868e-02 -3.16889644e-01 5.54185629e-01
3.56382012e-01 2.57731736e-01 8.53386104e-01 -6.07735813e-01
5.04710972e-01 5.55134833e-01 4.62600529e-01 -1.18596268e+00
-7.40998328e-01 -5.55081964e-02 -6.71519697e-01 2.75618851e-01
8.42055142e-01 -1.44103914e-01 -9.23594832e-01 1.42337441e+00
3.71234208e-01 3.79594028e-01 -1.40241534e-01 9.48714793e-01
5.67995429e-01 9.13295627e-01 9.49986428e-02 2.01166883e-01
1.02548742e+00 -1.38139606e+00 -3.06936204e-01 -1.86258256e-01
7.87041664e-01 -4.74073350e-01 9.00881231e-01 4.67657536e-01
-8.17691267e-01 -8.34240377e-01 -6.58782423e-01 -1.19589709e-01
-7.73206711e-01 1.49738654e-01 3.50785762e-01 5.57519376e-01
-9.60478961e-01 8.35629404e-01 -1.11282778e+00 -6.75739408e-01
5.21355331e-01 4.14735526e-01 -2.69557595e-01 8.98027048e-03
-7.94671893e-01 1.00528014e+00 4.44319993e-01 2.43831441e-01
-8.12779129e-01 -2.10770711e-01 -8.03686202e-01 -4.85861637e-02
8.02750215e-02 -7.17636228e-01 1.06365788e+00 -1.06351495e+00
-1.39545059e+00 4.94274795e-01 -2.57444769e-01 -6.31183743e-01
8.43729019e-01 -5.38070023e-01 -5.97436011e-01 -2.87601650e-01
-6.95538567e-03 4.75802988e-01 8.46420050e-01 -7.19689310e-01
-1.01628613e+00 -2.55841076e-01 -1.98243782e-01 -3.07971202e-02
-1.29718006e-01 1.01530932e-01 -2.28404894e-01 -5.70923984e-01
-3.83139968e-01 -1.06204391e+00 -5.42960882e-01 -1.88979894e-01
-4.28012401e-01 -3.63573611e-01 1.01351357e+00 -6.74495637e-01
1.16578889e+00 -1.89085138e+00 3.23999226e-01 3.32973227e-02
-2.15418875e-01 5.68125308e-01 -1.07011199e-01 6.19862258e-01
1.26597956e-01 -4.51552987e-01 -2.33713418e-01 -2.04758972e-01
1.04266256e-01 1.04722328e-01 -1.60212100e-01 4.99751955e-01
2.26029143e-01 9.40019906e-01 -1.13030076e+00 -9.58198532e-02
5.29597163e-01 4.81711864e-01 -3.48927826e-01 2.24537045e-01
-1.26920596e-01 7.69735336e-01 -4.00005430e-01 1.99126050e-01
3.76836091e-01 -1.29049450e-01 -3.32120627e-01 2.54764527e-01
-2.61016518e-01 1.16109006e-01 -1.13305151e+00 1.59628141e+00
-3.75714988e-01 9.87405121e-01 -5.59347272e-01 -1.18782020e+00
6.43761218e-01 1.52669400e-01 3.11025769e-01 -4.77865428e-01
7.26740062e-02 -3.44281197e-02 1.42128438e-01 -1.06091464e+00
6.59122825e-01 1.35595398e-02 -1.07629420e-02 1.86690778e-01
9.79059860e-02 6.00995004e-01 2.34388411e-01 -5.41891977e-02
1.36372042e+00 4.14730400e-01 2.68400073e-01 -5.33502139e-02
8.42890799e-01 2.33979508e-01 3.49323332e-01 7.97970414e-01
-2.99136847e-01 5.44120431e-01 6.93331808e-02 -1.10854685e+00
-1.16270745e+00 -7.39029229e-01 2.15528786e-01 1.22350907e+00
1.32181197e-01 -3.04878414e-01 -6.17453158e-01 -6.84501767e-01
-2.71648347e-01 6.18244708e-01 -5.28115869e-01 -7.52919987e-02
-8.89602304e-01 -7.78111160e-01 2.87546664e-01 7.15505719e-01
4.02568281e-01 -1.42876565e+00 -9.83070910e-01 4.38352764e-01
-2.18995102e-02 -1.00832963e+00 -2.45465815e-01 -2.27600381e-01
-6.75191760e-01 -1.04524052e+00 -7.17429996e-01 -8.49051654e-01
3.66268426e-01 8.86433870e-02 7.83148706e-01 7.88077116e-02
-3.64564389e-01 3.76109958e-01 -3.98553222e-01 -3.31479013e-01
-2.79014975e-01 -7.33512864e-02 2.40916051e-02 3.88189048e-01
5.19705832e-01 -4.08945262e-01 -8.83538425e-01 4.14764076e-01
-6.76686943e-01 -3.94450836e-02 4.70718950e-01 8.22842181e-01
1.78022772e-01 1.69888362e-01 5.29496610e-01 -7.12497592e-01
2.83355206e-01 -8.27950656e-01 -4.60797280e-01 1.26400828e-01
-3.70715000e-02 1.17308855e-01 8.64794672e-01 -5.90824544e-01
-9.65720117e-01 3.92726332e-01 -5.52894592e-01 -1.08192474e-01
-7.69310892e-01 1.11127540e-01 -1.77943662e-01 7.93870315e-02
6.11914217e-01 5.44774473e-01 -2.94272244e-01 -4.29349631e-01
4.74640340e-01 5.94754815e-01 5.59979141e-01 -1.41683996e-01
6.97861731e-01 6.32076919e-01 2.35896893e-02 -1.05507708e+00
-4.38324898e-01 -7.64246702e-01 -1.00844038e+00 -3.94974887e-01
9.32858765e-01 -5.95518470e-01 -7.15938568e-01 7.53911078e-01
-1.14769387e+00 -6.22431338e-01 -7.40252510e-02 6.55632019e-01
-8.75281274e-01 5.06764650e-01 -3.73232275e-01 -9.34781015e-01
-3.81346159e-02 -1.15781212e+00 9.96772885e-01 2.45965928e-01
-2.44297951e-01 -1.19230354e+00 2.18281135e-01 4.83029112e-02
3.33973229e-01 4.72820282e-01 7.40039825e-01 -8.03950608e-01
-7.82113850e-01 -4.08279598e-01 3.58237885e-02 2.10067537e-02
2.80148033e-02 -1.61855742e-02 -7.27047205e-01 -2.14487761e-01
-1.98974922e-01 -3.60119902e-02 9.92775619e-01 4.98611063e-01
9.03026879e-01 -5.13371646e-01 -7.84842670e-01 4.53680038e-01
1.23251963e+00 2.59464651e-01 7.77378440e-01 6.86370254e-01
8.29604030e-01 7.85936892e-01 6.06278598e-01 4.66334343e-01
5.16055167e-01 7.44264245e-01 4.49811876e-01 3.86058122e-01
1.21869721e-01 -1.37289330e-01 4.63952571e-01 4.71034527e-01
-3.08065981e-01 -4.64398891e-01 -9.06383395e-01 1.03581297e+00
-2.36269164e+00 -1.44742692e+00 -5.13051271e-01 2.15502882e+00
-8.41642693e-02 1.61062524e-01 7.34308541e-01 2.53142834e-01
5.49583733e-01 -6.26163334e-02 -3.33253443e-01 -3.55083019e-01
3.20697844e-01 -3.02731127e-01 3.26533943e-01 3.79927069e-01
-1.39077175e+00 9.43603158e-01 5.82286596e+00 5.79331160e-01
-9.27564323e-01 6.28449693e-02 3.14510047e-01 -3.14270891e-02
1.40315533e-01 -1.77364603e-01 -9.03687835e-01 8.22089016e-01
1.15433145e+00 1.53512254e-01 1.43132105e-01 7.24124372e-01
4.87116247e-01 1.56294331e-02 -1.21090388e+00 9.44981277e-01
-1.60150483e-01 -1.29204786e+00 -1.44894883e-01 -1.06688514e-01
6.54187441e-01 1.49153233e-01 3.84283289e-02 4.31463361e-01
2.76849002e-01 -1.01206958e+00 7.77000964e-01 9.34986234e-01
-4.17895094e-02 -7.98457742e-01 6.18764579e-01 8.51291299e-01
-1.09293795e+00 -5.22534549e-01 -3.48351747e-01 -2.63583660e-01
5.01366079e-01 1.31390095e-01 -8.59846234e-01 3.82854164e-01
7.21065819e-01 9.96859252e-01 -5.98165512e-01 1.61372840e+00
1.77140206e-01 6.22835517e-01 -3.35948020e-01 -2.62185752e-01
7.74363101e-01 -6.95993484e-04 6.75927281e-01 1.56358838e+00
5.08270860e-01 -4.52669077e-02 3.24859610e-03 4.75707710e-01
5.49435735e-01 6.80585653e-02 -9.27248359e-01 4.17608202e-01
-7.64750913e-02 1.08036029e+00 -7.82224000e-01 -3.43230516e-01
-6.57355607e-01 9.66131628e-01 3.81354481e-01 2.91049898e-01
-7.81001985e-01 -3.00843686e-01 5.82811952e-01 2.81551927e-01
6.23618782e-01 -4.12061751e-01 1.82437867e-01 -9.72467661e-01
1.96581278e-02 -4.31126714e-01 4.19151455e-01 -5.17719209e-01
-1.17177415e+00 7.97166467e-01 -7.74171352e-02 -1.55503535e+00
-5.79959750e-01 -7.06421018e-01 -9.97071445e-01 5.42068243e-01
-1.35242689e+00 -1.09277725e+00 -2.75607169e-01 8.61136317e-01
9.89976168e-01 -3.82451057e-01 4.51179326e-01 3.53015929e-01
-3.66541088e-01 3.57415259e-01 4.14637834e-01 2.14718401e-01
4.11276698e-01 -1.24675906e+00 8.14447522e-01 1.03150046e+00
3.15622956e-01 2.78836042e-01 7.02346444e-01 -4.89614964e-01
-1.19649613e+00 -1.37621200e+00 7.60562360e-01 -8.95677328e-01
6.61317110e-01 -1.59291178e-01 -8.70847583e-01 8.90247166e-01
6.34774640e-02 1.97906360e-01 4.57223326e-01 -2.02538714e-01
1.27472565e-01 1.57268912e-01 -8.64938080e-01 6.27187490e-01
1.04957986e+00 3.46153043e-02 -5.87432981e-01 3.94772381e-01
2.92542577e-01 -2.58782685e-01 -5.43588519e-01 1.60858572e-01
5.89672863e-01 -1.14664412e+00 1.09017694e+00 -1.14237809e+00
2.84613281e-01 -4.45823729e-01 1.05125673e-01 -1.25878251e+00
-5.62400877e-01 -6.02126956e-01 -3.79046142e-01 8.25370252e-01
2.86229968e-01 -2.05072820e-01 8.03702176e-01 5.00731528e-01
-2.47127697e-01 -5.89147031e-01 -9.20457661e-01 -7.51312137e-01
-1.11793600e-01 -6.09052598e-01 5.24808109e-01 5.44815361e-01
-4.69167382e-02 7.37626404e-02 -8.37973714e-01 2.43026808e-01
4.00469661e-01 6.10885955e-02 9.52936709e-01 -1.11077964e+00
-2.69619256e-01 -4.95876163e-01 -1.15995157e+00 -1.25440800e+00
9.55803171e-02 -8.38404834e-01 1.92631796e-01 -1.54298019e+00
-9.70064327e-02 -2.53527075e-01 -1.16612829e-01 8.64089429e-02
-2.25046098e-01 -8.34195316e-02 1.78437799e-01 1.38010561e-01
-9.13197458e-01 4.47530746e-01 1.02269447e+00 -1.10359937e-01
-3.37078691e-01 6.73906386e-01 -1.22972853e-01 9.71134424e-01
5.88936985e-01 -3.81665647e-01 -3.53731245e-01 -4.90820557e-01
9.88254398e-02 7.75082130e-03 7.23785758e-01 -1.30789924e+00
4.62496758e-01 -8.64620954e-02 4.09422904e-01 -7.38563776e-01
3.70783031e-01 -9.46004152e-01 1.23029891e-02 4.22003180e-01
-3.57035697e-01 1.13374978e-01 1.26480073e-01 1.10442126e+00
-8.42335448e-02 -1.14725046e-01 5.66579044e-01 -2.62651622e-01
-1.31956112e+00 3.73068869e-01 -8.11457336e-01 -1.67672798e-01
1.26070833e+00 -3.67292762e-01 -4.44508791e-02 -4.06155467e-01
-1.11846900e+00 2.73081034e-01 1.72719046e-01 9.07766879e-01
6.33835852e-01 -1.33057213e+00 -6.01576626e-01 2.02898070e-01
1.00346111e-01 -1.34586066e-01 3.37388039e-01 8.94886136e-01
-5.16979992e-01 5.15201867e-01 -3.78055751e-01 -7.67544687e-01
-1.09948504e+00 8.99881899e-01 1.90991834e-01 -2.32116431e-01
-1.02044177e+00 6.97544217e-01 -5.06527834e-02 -4.06996369e-01
2.09627405e-01 -4.19405371e-01 -4.67321008e-01 -1.68511376e-01
8.61327350e-01 8.19845438e-01 -7.59452768e-03 -1.09205818e+00
-3.48899305e-01 5.18781424e-01 -4.73434739e-02 1.03079386e-01
1.51798332e+00 -2.72736877e-01 3.57128769e-01 4.96486664e-01
1.19553554e+00 -4.85524654e-01 -1.56482387e+00 -1.32121608e-01
5.46021700e-01 -4.80883062e-01 -3.80577028e-01 -6.39219940e-01
-7.65920699e-01 1.09775722e+00 7.69769132e-01 2.38617286e-01
8.53496552e-01 -1.07499458e-01 9.94883120e-01 5.71683884e-01
4.55524117e-01 -9.65671241e-01 -7.32210353e-02 4.88161266e-01
5.65330684e-01 -1.52756119e+00 -4.80915129e-01 2.44381413e-01
-8.72081220e-01 1.26036787e+00 4.11219686e-01 -3.82657677e-01
7.83770859e-01 -3.93033680e-03 -1.78501129e-01 8.05676822e-03
-7.17636049e-01 -3.31792206e-01 4.60019469e-01 8.18574131e-01
1.72784582e-01 -4.92412634e-02 -9.40759182e-02 5.02679884e-01
-5.72795384e-02 1.13647217e-02 5.53902030e-01 8.16859782e-01
-5.89955986e-01 -9.43785906e-01 -3.45349133e-01 5.05845249e-01
-2.85195976e-01 2.88913846e-01 -6.34859502e-02 7.77467489e-01
2.84528613e-01 9.10511553e-01 6.74255565e-02 -3.84549052e-01
2.99611419e-01 1.86823100e-01 3.51271302e-01 -2.77579278e-01
-5.50816119e-01 -3.02526444e-01 3.40925455e-02 -7.44296670e-01
-5.46104133e-01 -1.05813015e+00 -9.60240543e-01 -1.53839275e-01
-1.52796954e-02 -2.71555454e-01 4.23280597e-01 1.24298024e+00
3.38285625e-01 2.96214640e-01 4.00507599e-01 -1.15013540e+00
-2.32979715e-01 -9.50889230e-01 -3.27815682e-01 6.80429280e-01
5.62724769e-01 -5.94932675e-01 9.56144035e-02 2.97740549e-01] | [6.488801956176758, 0.5494167804718018] |
2c82b09e-5e43-4097-9ccd-17ea0bdb2180 | robust-3d-scene-segmentation-through | 2111.08434 | null | https://arxiv.org/abs/2111.08434v1 | https://arxiv.org/pdf/2111.08434v1.pdf | Robust 3D Scene Segmentation through Hierarchical and Learnable Part-Fusion | 3D semantic segmentation is a fundamental building block for several scene understanding applications such as autonomous driving, robotics and AR/VR. Several state-of-the-art semantic segmentation models suffer from the part misclassification problem, wherein parts of the same object are labelled incorrectly. Previous methods have utilized hierarchical, iterative methods to fuse semantic and instance information, but they lack learnability in context fusion, and are computationally complex and heuristic driven. This paper presents Segment-Fusion, a novel attention-based method for hierarchical fusion of semantic and instance information to address the part misclassifications. The presented method includes a graph segmentation algorithm for grouping points into segments that pools point-wise features into segment-wise features, a learnable attention-based network to fuse these segments based on their semantic and instance features, and followed by a simple yet effective connected component labelling algorithm to convert segment features to instance labels. Segment-Fusion can be flexibly employed with any network architecture for semantic/instance segmentation. It improves the qualitative and quantitative performance of several semantic segmentation backbones by upto 5% when evaluated on the ScanNet and S3DIS datasets. | ['Sreenivas Subramoney', 'Om J Omer', 'Prashant Laddha', 'Benjamin Ummenhofer', 'Anirud Thyagharajan'] | 2021-11-16 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 6.10770881e-01 7.53426790e-01 -2.40049794e-01 -7.84363091e-01
-6.60872161e-01 -3.95818889e-01 3.77439886e-01 4.91060436e-01
-1.22729465e-01 3.40509266e-01 -2.98934877e-01 -2.44119272e-01
-3.46342593e-01 -8.32077026e-01 -7.52522051e-01 -4.97506022e-01
1.46086693e-01 8.81232917e-01 8.50074053e-01 -2.08298430e-01
3.96529108e-01 7.66002417e-01 -2.03285193e+00 1.47398114e-01
1.10192525e+00 1.31762397e+00 3.05703789e-01 3.09165657e-01
-7.33612478e-01 3.96358401e-01 -4.86948401e-01 -9.52516273e-02
3.41165423e-01 -2.38902941e-01 -1.25486565e+00 4.25332427e-01
5.17997086e-01 1.24696523e-01 1.31079033e-01 1.25193644e+00
2.03669488e-01 2.45068356e-01 5.43365121e-01 -1.68327403e+00
-2.28379399e-01 6.37348354e-01 -4.46092308e-01 -2.38190964e-02
2.47787207e-01 -1.35572523e-01 8.12894166e-01 -5.81119776e-01
4.96245533e-01 1.46844554e+00 6.86035633e-01 3.55943829e-01
-9.56437707e-01 -4.37998772e-01 4.81994897e-01 4.20883417e-01
-1.25188398e+00 6.75310520e-03 8.05881739e-01 -4.07010913e-01
1.07626855e+00 2.77442217e-01 9.41508114e-01 4.41680461e-01
-3.36527415e-02 9.29878831e-01 9.38210487e-01 -1.82574451e-01
3.18881631e-01 4.12006043e-02 5.61428905e-01 8.04709017e-01
2.88203090e-01 -2.60225177e-01 -1.06586009e-01 2.98936427e-01
6.88371003e-01 2.38974571e-01 1.84225842e-01 -7.14480519e-01
-9.31522727e-01 8.72973442e-01 9.17391419e-01 2.56138057e-01
-4.10342008e-01 2.09211469e-01 2.29015321e-01 -8.91214088e-02
5.77811062e-01 3.32075536e-01 -6.24584854e-01 3.80739450e-01
-1.11849248e+00 2.04353869e-01 7.25700974e-01 1.30251741e+00
1.28200638e+00 -1.08493455e-01 -2.12430015e-01 7.43000329e-01
4.90183353e-01 2.80780196e-01 3.47099215e-01 -1.17760384e+00
1.72583058e-01 1.30398893e+00 -3.43991280e-01 -8.47545207e-01
-7.88916886e-01 -3.70439500e-01 -4.76850569e-01 2.89167166e-01
3.88049404e-03 2.66245186e-01 -1.90614486e+00 1.35123920e+00
7.02327251e-01 2.61171192e-01 -7.41134733e-02 1.03805816e+00
1.19274449e+00 1.57470837e-01 3.91414404e-01 3.70307148e-01
1.46915305e+00 -1.17024589e+00 -5.29032528e-01 -4.14916933e-01
6.61535501e-01 -4.48076904e-01 5.69346011e-01 3.32320035e-02
-1.10978770e+00 -6.95181668e-01 -1.19546795e+00 -3.10134292e-01
-9.10426080e-01 -2.65054107e-01 6.48434997e-01 6.39389932e-01
-1.19009399e+00 6.27147675e-01 -8.68454099e-01 -5.32304227e-01
8.79922867e-01 6.22345626e-01 -3.82739812e-01 -2.20465377e-01
-8.43792140e-01 8.92667174e-01 9.40181315e-01 1.29661262e-01
-8.77512515e-01 -5.03407836e-01 -1.26231217e+00 1.22207642e-01
6.61284685e-01 -8.34738553e-01 1.13219297e+00 -1.02772331e+00
-1.15314019e+00 1.00089169e+00 -2.42698826e-02 -6.54818833e-01
3.23844820e-01 -9.07290131e-02 -1.53987616e-01 3.64442319e-01
4.23375994e-01 1.22108090e+00 8.74154627e-01 -1.66287637e+00
-9.75197852e-01 -6.36312723e-01 1.14063039e-01 4.62778360e-01
5.99903882e-01 -5.38660169e-01 -6.42989218e-01 -3.21873903e-01
9.30136502e-01 -7.47882307e-01 -5.94479620e-01 -2.71343619e-01
-6.21673644e-01 -4.03681964e-01 1.30340362e+00 -5.70122421e-01
6.58711731e-01 -1.85787201e+00 1.92171216e-01 4.71113354e-01
3.20347220e-01 2.63335347e-01 -1.48607809e-02 -5.80149926e-02
1.46961048e-01 1.82832897e-01 -8.31309319e-01 -3.62205535e-01
-1.34650646e-02 4.32062984e-01 2.60767072e-01 1.81341842e-01
3.55769485e-01 9.68804121e-01 -1.01532209e+00 -6.95469737e-01
7.06534564e-01 3.78849864e-01 -5.45285881e-01 1.28288805e-01
-4.49784726e-01 3.38513225e-01 -5.45600176e-01 8.72744381e-01
6.92179382e-01 -5.45319989e-02 -2.35531643e-01 -1.78113222e-01
7.13398084e-02 1.61748514e-01 -1.15441382e+00 2.00618196e+00
-8.15182254e-02 3.52989763e-01 3.63175236e-02 -1.43985212e+00
1.02072942e+00 -2.97334380e-02 6.69699252e-01 -6.58268869e-01
3.76496047e-01 2.82103270e-01 -3.70961070e-01 -4.36443120e-01
7.34289050e-01 -1.18056312e-02 -3.31175655e-01 -3.17690149e-02
3.23157191e-01 -6.56252146e-01 1.75592452e-01 9.80988890e-02
9.42136228e-01 3.79967302e-01 -5.97805046e-02 -2.51231194e-01
6.64151073e-01 5.83902121e-01 5.81204295e-01 7.30421484e-01
-2.60753661e-01 8.23970377e-01 2.44541883e-01 -2.51514494e-01
-8.28797519e-01 -9.49767470e-01 -1.88306607e-02 9.70280588e-01
8.01623940e-01 -1.94737315e-02 -1.25868690e+00 -9.15200412e-01
2.07378358e-01 8.54291856e-01 -4.58219111e-01 -4.04489338e-01
-3.43507439e-01 -3.34281474e-01 2.40850568e-01 6.78186715e-01
8.14867079e-01 -1.24680078e+00 -7.23618269e-01 3.04834723e-01
-9.74509642e-02 -1.16307306e+00 5.51042259e-02 5.04003763e-01
-9.09334958e-01 -1.25820315e+00 -4.32204634e-01 -9.69639540e-01
8.26702774e-01 5.13385296e-01 1.02370465e+00 3.39149982e-01
-4.10830528e-01 3.65031302e-01 -4.35616076e-01 -4.90558714e-01
-3.24160665e-01 3.46684992e-01 -4.80487704e-01 -8.23258087e-02
3.41385216e-01 -1.54259294e-01 -5.49220443e-01 4.45869982e-01
-8.99734437e-01 1.05616562e-01 6.36563659e-01 4.17945415e-01
1.05042994e+00 9.99391526e-02 5.13028979e-01 -1.17435706e+00
1.00686766e-01 -4.92876232e-01 -4.97644097e-01 1.44706905e-01
-5.82710683e-01 1.55047448e-02 -1.01676760e-02 1.83321118e-01
-1.00457752e+00 3.66646588e-01 -1.94615692e-01 -5.30618191e-01
-6.80662632e-01 2.97856659e-01 -4.03531641e-01 -1.88927367e-01
3.06997716e-01 -1.15774870e-01 1.86994427e-03 -4.64548916e-01
7.66486287e-01 6.43453181e-01 6.61662519e-01 -2.63353437e-01
4.21314269e-01 5.60070395e-01 9.87395421e-02 -7.94676244e-01
-1.01325536e+00 -1.02256858e+00 -9.67435062e-01 -2.07744747e-01
1.33881366e+00 -8.09691727e-01 -3.25330704e-01 6.10336483e-01
-9.02395904e-01 -3.84870827e-01 -4.91682559e-01 1.10699438e-01
-8.77287984e-01 2.81442940e-01 -1.73866615e-01 -5.20407677e-01
-1.30185544e-01 -1.34630930e+00 1.56852472e+00 5.09103119e-01
5.80371879e-02 -8.49475682e-01 -4.86255616e-01 7.68957436e-01
3.61865498e-02 3.94271791e-01 9.12587285e-01 -8.75540853e-01
-6.92430139e-01 -2.04549283e-01 -6.00667119e-01 3.02426636e-01
-5.04874922e-02 -2.69676685e-01 -8.91512811e-01 -3.07462998e-02
-2.88598061e-01 5.55301365e-03 1.02385747e+00 4.75583106e-01
1.22398126e+00 -5.59970848e-02 -7.17742741e-01 5.19173861e-01
1.44525468e+00 3.51695389e-01 4.94240046e-01 1.22369289e-01
1.23223078e+00 7.19214737e-01 7.17380524e-01 -1.04212776e-01
5.50418317e-01 3.76284122e-01 8.53848219e-01 -4.14828569e-01
-3.50426733e-01 -1.54346809e-01 -2.80415803e-01 3.10975879e-01
1.00817084e-01 -2.42338166e-01 -9.42249477e-01 7.57726431e-01
-2.01715636e+00 -5.60826719e-01 -3.58990431e-01 1.87155271e+00
1.80399314e-01 3.33370417e-01 1.16423838e-01 4.17793214e-01
9.95490909e-01 -1.94469333e-01 -6.97777152e-01 -3.96317303e-01
4.28901315e-02 2.25810006e-01 9.60425675e-01 4.09641594e-01
-1.21136618e+00 1.40157557e+00 5.54340124e+00 7.47839272e-01
-7.38492727e-01 1.56639472e-01 5.37231028e-01 4.72883105e-01
-2.04317734e-01 9.43080708e-02 -6.81793094e-01 1.01499401e-01
6.21548235e-01 3.56790453e-01 2.08161492e-02 9.30212080e-01
-6.88439459e-02 -4.65673596e-01 -8.71874511e-01 6.98358476e-01
1.71383426e-01 -1.08979344e+00 1.37453049e-01 -1.72066763e-01
8.07470620e-01 1.13602728e-01 -3.50637168e-01 3.07915241e-01
4.15294707e-01 -9.67001915e-01 9.47982848e-01 5.08107841e-01
4.04547065e-01 -8.51689637e-01 9.89600360e-01 2.18358085e-01
-1.32439435e+00 -6.22959137e-02 -2.41728246e-01 2.54603654e-01
2.74150282e-01 3.56048048e-01 -8.64913344e-01 7.79356956e-01
9.20530319e-01 6.86759531e-01 -6.35446310e-01 1.35359812e+00
-2.53291905e-01 2.39831626e-01 -4.45812970e-01 2.19185278e-01
6.31866336e-01 -2.41107628e-01 4.87698168e-01 9.56040680e-01
1.54955193e-01 1.27932847e-01 5.32756448e-01 9.00949895e-01
7.10732713e-02 -1.79199576e-01 -3.95983905e-01 1.22950077e-01
3.24635714e-01 1.21300030e+00 -1.58010542e+00 -6.73901141e-01
-1.75173566e-01 1.00334775e+00 9.38521400e-02 3.05305123e-01
-6.97416067e-01 -5.50741255e-01 4.86302793e-01 1.75871775e-01
5.71980596e-01 -1.22314952e-01 -7.32977390e-01 -4.53360200e-01
-1.54774666e-01 -1.96888208e-01 4.24442440e-01 -8.24095070e-01
-8.97502005e-01 5.51951468e-01 3.80949914e-01 -8.46972644e-01
1.19292010e-02 -3.85462523e-01 -3.35393220e-01 5.67364573e-01
-1.55857146e+00 -1.33342254e+00 -7.11876750e-01 5.74227095e-01
7.90134370e-01 2.16745347e-01 3.81116033e-01 1.27754956e-01
-4.70921129e-01 1.28454417e-01 -5.00454187e-01 -4.63318005e-02
-8.91611353e-02 -1.39005947e+00 4.59291905e-01 6.37535572e-01
-2.06569955e-02 3.19605991e-02 5.62290430e-01 -8.87733638e-01
-7.16085136e-01 -1.44941556e+00 6.82198584e-01 -3.30453724e-01
1.89902246e-01 -2.15998888e-01 -1.01206541e+00 5.90817273e-01
-1.27704859e-01 1.14615776e-01 2.11214080e-01 -2.78690577e-01
-6.78665414e-02 1.82611942e-01 -1.61977971e+00 2.52614558e-01
1.42508733e+00 -2.78599590e-01 -8.23168933e-01 2.81758487e-01
1.16886950e+00 -5.38194716e-01 -7.51745284e-01 7.41329551e-01
1.62459701e-01 -1.01660812e+00 1.00693631e+00 -3.85770798e-01
-7.31267557e-02 -6.15820885e-01 -6.82335272e-02 -1.11617053e+00
-2.91508973e-01 -1.60026625e-01 1.58660308e-01 1.08114791e+00
3.42609614e-01 -4.98695016e-01 9.32226121e-01 4.77894962e-01
-8.55111837e-01 -5.99923968e-01 -9.54472661e-01 -6.35750115e-01
-2.42040396e-01 -5.67335606e-01 9.21198010e-01 6.57269359e-01
-4.57567930e-01 1.68188348e-01 4.79521990e-01 3.29786927e-01
6.73528969e-01 1.66918382e-01 4.62078989e-01 -1.54743671e+00
4.47922885e-01 -8.22525203e-01 -1.05751324e+00 -7.10745394e-01
1.82845429e-01 -1.16041279e+00 4.10352945e-01 -2.27737570e+00
-2.28227407e-01 -7.06006289e-01 -2.86240041e-01 6.86105311e-01
-4.62882407e-02 4.25495118e-01 2.18671247e-01 -3.54570746e-02
-8.29118490e-01 4.11776483e-01 1.14840972e+00 -3.82347077e-01
-4.52037871e-01 8.94674733e-02 -4.95677799e-01 7.48153090e-01
8.00731301e-01 -5.15922248e-01 -5.51485777e-01 -1.06657073e-01
-2.19242170e-01 -5.73178008e-02 5.94737828e-01 -1.29290640e+00
1.78354606e-01 1.25353739e-01 1.17894612e-01 -1.06513679e+00
2.80041069e-01 -1.15628481e+00 1.86027631e-01 3.58862191e-01
-1.00646265e-01 -3.09425622e-01 2.53687620e-01 6.50728583e-01
-2.41504893e-01 -2.29518101e-01 7.42393851e-01 -4.51925188e-01
-1.34559965e+00 3.82360786e-01 -1.83221102e-01 1.15782199e-02
1.32375038e+00 -7.48404801e-01 -1.43234702e-02 2.32795581e-01
-1.02891445e+00 5.19927442e-01 4.11707073e-01 7.00131714e-01
6.06047630e-01 -1.13516760e+00 -3.60354453e-01 4.31786954e-01
1.96873575e-01 7.73408532e-01 4.61040407e-01 7.44582295e-01
-7.34696865e-01 4.72358525e-01 -2.97647774e-01 -1.10283720e+00
-1.03615975e+00 5.33754349e-01 4.55133140e-01 1.71586081e-01
-5.65073967e-01 9.08701897e-01 7.64214769e-02 -6.41761661e-01
1.38776109e-01 -6.93480194e-01 -3.28271508e-01 1.68045580e-01
-1.92955986e-01 5.44103563e-01 3.41719568e-01 -1.06679058e+00
-4.35883641e-01 8.88835609e-01 1.45074785e-01 1.83786318e-01
9.68190610e-01 -3.19432467e-01 6.32411987e-03 3.23023558e-01
1.11976302e+00 -9.75182772e-01 -1.35988355e+00 -9.85619798e-02
1.72574624e-01 -7.01339096e-02 2.17913941e-01 -7.46434331e-01
-1.28200901e+00 7.83392847e-01 5.85197747e-01 4.51866627e-01
1.04193413e+00 4.43212301e-01 8.41719687e-01 7.90742859e-02
5.15677273e-01 -1.28786218e+00 -3.63407999e-01 3.82421851e-01
5.86577713e-01 -1.31729984e+00 -3.88792343e-02 -9.15024817e-01
-4.68468249e-01 7.01623380e-01 7.01677740e-01 -1.35866046e-01
7.11364627e-01 -2.61039793e-01 1.30934948e-02 -6.67990744e-01
-2.43330784e-02 -8.13679993e-01 4.97460544e-01 8.42475533e-01
-1.98914811e-01 2.20670536e-01 -1.17422938e-01 3.55306685e-01
-1.33360744e-01 -3.15996051e-01 1.08231299e-01 1.07323349e+00
-9.35590088e-01 -6.45793200e-01 -2.42708325e-01 6.20485485e-01
-2.05596820e-01 3.50966901e-01 -4.19909805e-01 9.34198737e-01
6.35254741e-01 1.02223480e+00 3.99498552e-01 -3.61357301e-01
4.49222714e-01 1.01181500e-01 2.86760360e-01 -8.04000139e-01
-6.09160304e-01 -4.85818014e-02 -2.58532446e-02 -8.55564475e-01
-7.65819609e-01 -6.43783927e-01 -1.91316509e+00 1.73351333e-01
-4.77658927e-01 -2.47846115e-02 1.00844908e+00 1.26899397e+00
3.85278910e-01 7.76095748e-01 3.75453055e-01 -1.20486248e+00
1.85802728e-01 -7.34166265e-01 -4.36188221e-01 5.16621232e-01
2.10438028e-01 -1.05233943e+00 -3.02412808e-01 -9.90544111e-02] | [8.154186248779297, -2.9281740188598633] |
5117af7a-a4b2-41a5-848d-b04cbdba47fc | makeup-extraction-of-3d-representation-via | 2302.13279 | null | https://arxiv.org/abs/2302.13279v1 | https://arxiv.org/pdf/2302.13279v1.pdf | Makeup Extraction of 3D Representation via Illumination-Aware Image Decomposition | Facial makeup enriches the beauty of not only real humans but also virtual characters; therefore, makeup for 3D facial models is highly in demand in productions. However, painting directly on 3D faces and capturing real-world makeup are costly, and extracting makeup from 2D images often struggles with shading effects and occlusions. This paper presents the first method for extracting makeup for 3D facial models from a single makeup portrait. Our method consists of the following three steps. First, we exploit the strong prior of 3D morphable models via regression-based inverse rendering to extract coarse materials such as geometry and diffuse/specular albedos that are represented in the UV space. Second, we refine the coarse materials, which may have missing pixels due to occlusions. We apply inpainting and optimization. Finally, we extract the bare skin, makeup, and an alpha matte from the diffuse albedo. Our method offers various applications for not only 3D facial models but also 2D portrait images. The extracted makeup is well-aligned in the UV space, from which we build a large-scale makeup dataset and a parametric makeup model for 3D faces. Our disentangled materials also yield robust makeup transfer and illumination-aware makeup interpolation/removal without a reference image. | ['Yoshihiro Kanamori', 'Takafumi Taketomi', 'Xingchao Yang'] | 2023-02-26 | null | null | null | null | ['inverse-rendering'] | ['computer-vision'] | [ 4.43067610e-01 5.39049171e-02 6.32953048e-02 -4.39103037e-01
-3.30451101e-01 -4.93625224e-01 4.66668934e-01 -6.48973048e-01
3.90899539e-01 5.83814144e-01 -5.65610491e-02 2.85762906e-01
3.60360742e-01 -1.02637529e+00 -6.59443796e-01 -7.84125209e-01
4.09889489e-01 3.42302561e-01 -1.84018314e-01 -3.06353897e-01
-7.60026723e-02 1.03923619e+00 -1.56118286e+00 1.80263028e-01
1.02008939e+00 9.84977424e-01 -1.56489715e-01 5.62281251e-01
-5.68355918e-01 3.53568763e-01 -5.20306528e-01 -9.04497504e-01
6.12625122e-01 -4.75190312e-01 7.80833699e-03 4.90346223e-01
9.44495022e-01 -6.59669340e-01 -2.17263892e-01 9.24396992e-01
3.24106991e-01 -1.49631247e-01 8.21157634e-01 -1.26177561e+00
-9.20044839e-01 -2.47586071e-01 -1.23926246e+00 -8.37826073e-01
3.11131686e-01 2.61233717e-01 4.78074104e-01 -1.05295897e+00
8.67110848e-01 1.63138390e+00 6.48840010e-01 7.90748060e-01
-1.41552758e+00 -1.16941929e+00 -2.44605448e-03 -5.00564516e-01
-1.44977379e+00 -6.61819875e-01 1.44945347e+00 -4.32648897e-01
1.88544586e-01 7.08513021e-01 1.03673553e+00 9.06905890e-01
4.99181077e-02 5.08004844e-01 1.67127252e+00 -3.40071380e-01
-1.06841326e-01 2.01291919e-01 -3.22700828e-01 1.05676270e+00
5.96309565e-02 6.98240893e-03 -5.74356616e-01 -2.75590420e-01
1.41354942e+00 2.10601538e-01 -3.41921955e-01 -2.55659133e-01
-7.66261816e-01 3.47058773e-01 2.92409927e-01 -4.03946966e-01
-2.37443671e-01 9.97387618e-02 -4.15687948e-01 5.14847450e-02
8.98289084e-01 2.68174767e-01 -1.46638989e-01 1.41747266e-01
-1.09942651e+00 2.90502757e-01 5.43251216e-01 9.78882909e-01
1.30717897e+00 9.49784964e-02 3.43946844e-01 1.11844289e+00
4.16537821e-01 1.12132442e+00 -3.04201812e-01 -1.28990066e+00
2.46385396e-01 6.72558367e-01 1.47066683e-01 -1.05467582e+00
9.89314988e-02 9.58231017e-02 -7.97575831e-01 6.27404988e-01
1.34973541e-01 -4.89642471e-02 -1.13314569e+00 1.50776601e+00
8.53384197e-01 1.36988297e-01 -3.63135993e-01 9.54565465e-01
8.91650081e-01 5.19156039e-01 -4.27297890e-01 -2.29425907e-01
1.23802149e+00 -6.73063934e-01 -7.97607720e-01 -1.01984888e-01
-1.93364486e-01 -9.84161019e-01 9.76542950e-01 4.16894615e-01
-1.24708390e+00 -3.50768149e-01 -9.66355920e-01 -4.85298276e-01
2.90733397e-01 6.94718808e-02 6.07878625e-01 7.88734019e-01
-6.30225956e-01 5.82117617e-01 -6.38353527e-01 3.94182876e-02
4.88753825e-01 1.74739867e-01 -5.84983230e-01 -1.34009019e-01
-6.95063174e-01 5.06614268e-01 -4.23909783e-01 1.81968912e-01
-6.72167242e-01 -1.04376411e+00 -8.94387066e-01 -4.59864974e-01
2.03650296e-01 -7.49971211e-01 6.31813109e-01 -1.36268687e+00
-1.91059804e+00 1.24955320e+00 -3.48340303e-01 4.63501811e-01
5.41534185e-01 -1.00157827e-01 -2.89045155e-01 9.56960097e-02
-2.82380521e-01 5.41685760e-01 1.42130423e+00 -1.85128450e+00
1.96472019e-01 -5.86449146e-01 6.95775216e-03 2.91456103e-01
-6.62659332e-02 -1.08283766e-01 -5.68825543e-01 -7.64125168e-01
4.68234718e-01 -7.89439976e-01 1.51119664e-01 8.01981628e-01
-5.45357704e-01 6.53627932e-01 7.99748361e-01 -1.02715886e+00
7.01709688e-01 -2.05827022e+00 1.45693317e-01 3.94504130e-01
3.94736260e-01 -9.95700359e-02 -3.79025131e-01 3.32352549e-01
-9.85493138e-02 2.15371177e-01 -3.94669861e-01 -6.34769320e-01
4.51237373e-02 2.39093363e-01 -2.35879943e-01 3.51353317e-01
3.71108353e-01 6.74667776e-01 -4.93120909e-01 -6.55767620e-01
1.76386744e-01 9.77239728e-01 -4.04016644e-01 1.48969740e-01
-3.33868027e-01 7.36435115e-01 -4.92346138e-01 9.77393091e-01
1.36359692e+00 2.96712905e-01 -6.67127743e-02 -2.42054075e-01
2.46395660e-03 -1.22179441e-01 -1.09709692e+00 1.66288793e+00
-6.20331168e-01 3.65830451e-01 7.37929940e-01 9.92218554e-02
1.35903251e+00 -3.04119233e-02 4.76891071e-01 -4.85936671e-01
1.08481854e-01 1.90717265e-01 -4.56495583e-01 -2.80581921e-01
3.61805648e-01 -5.38857043e-01 4.68317538e-01 3.57965320e-01
-3.52884293e-01 -9.64196205e-01 -4.01333898e-01 -9.13479999e-02
4.64854181e-01 5.52292883e-01 -1.97017670e-01 -6.37040958e-02
2.73807496e-01 -3.58631313e-01 6.67970061e-01 -3.37250412e-01
2.89711744e-01 9.73770916e-01 5.97412467e-01 -4.17412847e-01
-1.16813183e+00 -1.16693711e+00 -2.27970153e-01 4.14855927e-01
1.89860553e-01 -3.54921103e-01 -1.10819268e+00 -2.52177835e-01
1.75482020e-01 4.27436233e-01 -7.66256094e-01 9.76898521e-02
-5.54855406e-01 -5.46122968e-01 2.88107663e-01 9.34442729e-02
7.33359277e-01 -4.46770370e-01 -1.71358198e-01 -3.90218973e-01
2.34003831e-02 -9.55960393e-01 -7.59052157e-01 -7.23116219e-01
-8.18853557e-01 -1.20185649e+00 -8.54298532e-01 -2.87171990e-01
9.51950192e-01 3.11505795e-01 1.15424156e+00 2.93753803e-01
-3.67584020e-01 2.38003880e-01 7.46736079e-02 -6.43609464e-01
-3.68140221e-01 -6.80414677e-01 -3.20092067e-02 7.34655559e-01
-7.39149153e-02 -9.43886280e-01 -7.11216867e-01 4.83016372e-01
-6.82369769e-01 7.88316250e-01 2.04543695e-01 4.38088238e-01
8.96209717e-01 -3.44935566e-01 -4.78837900e-02 -9.39171851e-01
1.68288797e-01 -9.95634776e-03 -6.73860729e-01 6.87089562e-02
-1.73841909e-01 -2.02652827e-01 1.10229358e-01 -5.09641111e-01
-1.57494366e+00 9.38994903e-03 -1.56500787e-01 -7.97553539e-01
-3.37405726e-02 -3.25036883e-01 -5.63844204e-01 -3.34544986e-01
6.83159292e-01 -6.27270639e-02 3.22716296e-01 -6.50420129e-01
4.32493627e-01 5.11310101e-01 2.00352609e-01 -6.84395432e-01
1.29956889e+00 1.01504457e+00 3.09673697e-01 -1.38822639e+00
-6.18677974e-01 2.11515620e-01 -6.97763503e-01 -3.95961851e-01
7.58669376e-01 -8.90187979e-01 -5.59878230e-01 6.49428308e-01
-1.17363596e+00 -4.67667460e-01 -3.80412340e-01 1.22000739e-01
-3.42731774e-01 3.51310849e-01 -6.61803365e-01 -8.67037654e-01
-2.85157919e-01 -9.42339003e-01 1.40808892e+00 2.71983862e-01
-5.72462603e-02 -6.34176850e-01 6.09333254e-02 6.54015779e-01
5.46963699e-02 9.08613026e-01 8.30746591e-01 8.19174945e-01
-9.70837057e-01 3.46928020e-03 -2.72550464e-01 4.23241615e-01
3.52510929e-01 5.24474084e-01 -1.44534302e+00 1.90884650e-01
-7.88889378e-02 -7.04343766e-02 5.80006003e-01 1.46547750e-01
1.01355135e+00 -2.87693828e-01 -8.49395692e-02 1.28292382e+00
1.29045033e+00 -9.95802656e-02 6.02186918e-01 -2.17935994e-01
1.17660785e+00 1.02983236e+00 5.66141903e-01 5.00313342e-01
1.74759597e-01 6.54879808e-01 6.12588525e-01 -3.63957465e-01
-6.03918493e-01 -6.42826557e-01 3.45326424e-01 7.62237191e-01
-6.25112116e-01 3.34295213e-01 -4.60264146e-01 2.96106050e-03
-1.21373892e+00 -5.78898668e-01 -4.61385280e-01 2.33036256e+00
9.08364177e-01 -5.49249768e-01 -1.93559989e-01 -1.81991518e-01
5.88087320e-01 1.25635073e-01 -6.80530787e-01 -3.34853858e-01
-3.96998018e-01 4.56366539e-01 3.09475452e-01 6.64465845e-01
-5.24693072e-01 8.37283790e-01 5.53310108e+00 7.35531032e-01
-1.18411613e+00 1.00976966e-01 7.05512941e-01 -4.30522054e-01
-1.11211431e+00 3.99417244e-02 -4.92055118e-01 3.50904703e-01
3.00726920e-01 2.20163524e-01 4.69626337e-01 5.72194457e-01
2.31951401e-01 -1.00041315e-01 -7.97763765e-01 1.42796504e+00
2.45979056e-01 -1.11840749e+00 1.58855528e-01 3.60766947e-01
1.12331951e+00 -5.42478502e-01 1.01354718e-01 -3.99367839e-01
9.13089439e-02 -1.11914182e+00 9.49283183e-01 7.85351217e-01
1.40899634e+00 -6.27426565e-01 -1.00926541e-01 2.92577624e-01
-9.78315592e-01 5.23484468e-01 -2.85626918e-01 -6.96457038e-03
7.06987828e-02 9.12344038e-01 -4.18132186e-01 3.71982455e-01
3.53893340e-01 5.43869615e-01 -2.29910329e-01 4.71996814e-01
-7.07256496e-01 2.11753786e-01 -4.97922391e-01 2.38168597e-01
-3.56578022e-01 -9.28967357e-01 5.00414848e-01 7.26119459e-01
4.04253453e-01 3.37020159e-01 -1.70771375e-01 1.30923295e+00
-4.54724371e-01 1.72731221e-01 -5.94709694e-01 1.35341555e-01
2.73800611e-01 1.41576755e+00 -4.53324467e-01 -5.31096831e-02
-9.91970003e-02 1.32405639e+00 6.72653839e-02 6.20977581e-01
-7.62002170e-01 -6.58171177e-02 9.97117937e-01 5.63185096e-01
-3.23371232e-01 -4.53918874e-01 -4.96696085e-01 -1.35389256e+00
2.22527727e-01 -6.59370899e-01 -4.44721788e-01 -1.16845334e+00
-1.09873044e+00 3.56757462e-01 -1.22941531e-01 -9.67418194e-01
3.12491745e-01 -5.20304978e-01 -5.99054456e-01 1.33287144e+00
-1.74694896e+00 -1.56789196e+00 -6.24633610e-01 5.37324965e-01
3.61250401e-01 4.04503882e-01 8.13130498e-01 1.41098499e-01
-6.13848865e-01 3.78179371e-01 -2.14842428e-02 -6.36169165e-02
8.29670072e-01 -9.41725254e-01 4.03739542e-01 4.90080863e-01
5.33848368e-02 5.89031935e-01 3.85397404e-01 -7.21970260e-01
-1.94083035e+00 -8.47390711e-01 3.42934430e-01 -5.55843771e-01
1.19821206e-01 -6.90151691e-01 -8.34935367e-01 5.18410206e-01
-2.00320259e-02 1.00394757e-02 6.37314379e-01 -4.63460796e-02
-6.17964745e-01 -3.35688829e-01 -1.39549315e+00 7.64997900e-01
1.23572505e+00 -5.46409547e-01 -1.79422989e-01 3.27085435e-01
3.34958375e-01 -6.40435278e-01 -9.29661155e-01 4.41790260e-02
8.70652795e-01 -1.26438355e+00 1.02180600e+00 -6.63086027e-02
5.61824083e-01 -3.19257617e-01 3.61549854e-02 -1.20742595e+00
6.55583814e-02 -1.20257735e+00 -3.29355374e-02 1.54652464e+00
1.44043908e-01 -5.42209923e-01 1.00439239e+00 1.12876976e+00
-2.50949748e-02 -6.50702059e-01 -5.33047557e-01 -4.19059783e-01
-4.92602922e-02 -3.63728076e-01 1.02983320e+00 1.11338174e+00
-8.91118467e-01 5.34100905e-02 -5.47064304e-01 1.68724302e-02
1.04858637e+00 5.41156709e-01 1.21420217e+00 -1.24812961e+00
-1.14724047e-01 -1.36267707e-01 5.30946292e-02 -9.36209500e-01
1.56953573e-01 -7.27301836e-01 -2.27852687e-01 -1.23785150e+00
1.58788189e-01 -8.95425379e-01 5.36972880e-01 3.14576119e-01
1.02865420e-01 5.52030921e-01 2.81587839e-01 2.44706601e-01
4.66812193e-01 7.70751059e-01 1.92104554e+00 -1.61742181e-01
-2.58658916e-01 -2.43177950e-01 -6.62871242e-01 1.18834984e+00
5.26387095e-01 -1.31807342e-01 -3.64089012e-01 -6.50324047e-01
3.50476086e-01 -1.54571915e-02 5.33309579e-01 -4.92700636e-01
-4.67722625e-01 -5.70002377e-01 7.62601078e-01 -3.12912256e-01
1.09467983e+00 -7.65848696e-01 6.10935330e-01 -7.40984231e-02
2.69984454e-01 -3.82649601e-01 1.14959083e-01 2.99980104e-01
1.27558812e-01 1.09063517e-02 1.09016383e+00 -2.25089237e-01
-3.21082026e-01 7.93990314e-01 4.20410097e-01 -1.76109865e-01
8.20697904e-01 -6.55838668e-01 -1.10100970e-01 -3.11431289e-01
-4.11832154e-01 -1.61122307e-01 1.33601236e+00 1.02345243e-01
8.63629758e-01 -1.43062270e+00 -8.27887714e-01 6.56938612e-01
-1.74175531e-01 2.29760513e-01 6.20637894e-01 5.61063051e-01
-1.03600478e+00 -4.01347756e-01 -2.75603920e-01 -3.94710511e-01
-1.61273098e+00 2.41223514e-01 3.93367261e-01 4.66308594e-01
-6.13387346e-01 1.02426803e+00 7.40793109e-01 -4.82823610e-01
-2.74462163e-01 -1.65553004e-01 3.24083507e-01 -4.87048179e-02
5.68537116e-01 2.60630906e-01 -2.77208835e-01 -8.89316678e-01
2.38371827e-02 1.33390403e+00 4.97228652e-01 -1.26732260e-01
1.28240764e+00 -1.07348375e-01 -5.32558441e-01 3.78465056e-01
1.10662401e+00 6.60929084e-01 -1.74284172e+00 1.03967272e-01
-8.91284525e-01 -1.17539716e+00 -2.30162032e-02 -5.11252761e-01
-1.31524932e+00 1.28864312e+00 -4.31502052e-02 -3.23253065e-01
1.28806925e+00 -2.50832707e-01 9.17281330e-01 -1.48801833e-01
3.70062560e-01 -8.38951409e-01 -1.74052864e-01 5.47985919e-02
1.27837563e+00 -8.67753088e-01 2.10599318e-01 -1.13293266e+00
-4.10668463e-01 1.14677489e+00 5.65924585e-01 -1.62970558e-01
6.85071230e-01 2.70540595e-01 1.46185905e-01 -1.41041696e-01
-4.11880881e-01 1.63819790e-01 4.81552184e-01 7.02684641e-01
1.98307008e-01 1.54993325e-01 1.49049222e-01 3.94603372e-01
-3.72957915e-01 -2.82772809e-01 4.44604427e-01 6.59312367e-01
-1.78927422e-01 -1.11501026e+00 -8.14447939e-01 2.27017105e-01
-9.26359892e-02 -1.25039294e-02 -7.67676950e-01 7.13009179e-01
4.86467272e-01 5.21684229e-01 1.58983450e-02 -2.41476700e-01
4.82928276e-01 -3.14079262e-02 9.89790022e-01 -5.70639431e-01
-1.57310009e-01 3.48949790e-01 9.12922919e-02 -6.81896269e-01
-3.18325788e-01 -3.95503193e-01 -1.03475261e+00 -6.15239561e-01
-1.80273995e-01 -3.00128549e-01 1.00046170e+00 3.17423999e-01
3.01681727e-01 1.69199586e-01 8.24337661e-01 -1.19874144e+00
1.58982098e-01 -6.36329830e-01 -1.06995022e+00 5.03835380e-01
3.37129533e-01 -8.34879398e-01 -1.58302337e-01 2.46570528e-01] | [12.752662658691406, -0.3261476755142212] |
5075f5b4-10df-4b89-a32c-f045ebf58390 | understanding-pure-clip-guidance-for-voxel | 2209.15172 | null | https://arxiv.org/abs/2209.15172v1 | https://arxiv.org/pdf/2209.15172v1.pdf | Understanding Pure CLIP Guidance for Voxel Grid NeRF Models | We explore the task of text to 3D object generation using CLIP. Specifically, we use CLIP for guidance without access to any datasets, a setting we refer to as pure CLIP guidance. While prior work has adopted this setting, there is no systematic study of mechanics for preventing adversarial generations within CLIP. We illustrate how different image-based augmentations prevent the adversarial generation problem, and how the generated results are impacted. We test different CLIP model architectures and show that ensembling different models for guidance can prevent adversarial generations within bigger models and generate sharper results. Furthermore, we implement an implicit voxel grid model to show how neural networks provide an additional layer of regularization, resulting in better geometrical structure and coherency of generated objects. Compared to prior work, we achieve more coherent results with higher memory efficiency and faster training speeds. | ['Angel X. Chang', 'Han-Hung Lee'] | 2022-09-30 | null | null | null | null | ['text-to-3d'] | ['computer-vision'] | [ 3.31754714e-01 4.58414048e-01 3.38011056e-01 6.00032061e-02
-8.23110461e-01 -8.44532490e-01 8.10352981e-01 -2.59629846e-01
-4.61824611e-02 6.04012609e-01 4.01126981e-01 -1.69257477e-01
2.53613412e-01 -8.94000769e-01 -1.12900567e+00 -4.83604252e-01
-8.54729414e-02 3.32911789e-01 2.42849976e-01 -2.12325633e-01
1.93755969e-01 5.88090181e-01 -1.17677975e+00 2.66379863e-01
8.96764398e-01 7.32411981e-01 -3.09665680e-01 9.05063033e-01
1.70236856e-01 6.30272508e-01 -8.77649426e-01 -6.24870777e-01
8.01680326e-01 -2.78831035e-01 -6.09566808e-01 -1.12320306e-02
9.73056734e-01 -5.53297937e-01 -3.85462582e-01 7.22335756e-01
6.93935454e-01 7.85706192e-02 9.95461464e-01 -1.09007657e+00
-8.57836306e-01 5.76966166e-01 -5.98157406e-01 -1.72682688e-01
3.05874586e-01 5.66279769e-01 5.69172800e-01 -6.44853890e-01
8.68958890e-01 1.29333854e+00 8.13366175e-01 8.56839299e-01
-1.48848999e+00 -5.05550802e-01 9.64376703e-02 -4.67923284e-01
-1.05797577e+00 -3.20931196e-01 8.21135998e-01 -4.32909399e-01
6.78543389e-01 4.67507005e-01 6.21857345e-01 1.31552923e+00
3.29703540e-01 6.89508140e-01 8.65790546e-01 -3.07700574e-01
1.22309722e-01 -1.15092518e-02 -2.34925687e-01 3.96223426e-01
2.14214027e-01 5.26581287e-01 -3.03160548e-01 -1.85255200e-01
1.18480611e+00 -4.89470810e-01 -2.99123079e-01 -3.93768758e-01
-8.65973949e-01 8.48812819e-01 6.55378878e-01 -2.07808033e-01
-1.52156115e-01 7.74567306e-01 2.78931737e-01 -1.81774013e-02
6.95538104e-01 9.11800742e-01 -4.18361202e-02 1.70155331e-01
-9.42767322e-01 9.64001298e-01 6.83916152e-01 1.12051320e+00
2.73147166e-01 5.32919049e-01 -4.23531085e-01 5.74427783e-01
8.59152824e-02 5.99383414e-01 -1.65578142e-01 -1.46679378e+00
3.82826358e-01 2.57292110e-02 1.36067927e-01 -1.05835378e+00
-3.86670858e-01 -5.21829605e-01 -6.79141104e-01 9.50241566e-01
5.90911865e-01 -5.15934467e-01 -1.17910981e+00 1.87891936e+00
1.34082958e-01 3.96607995e-01 -2.31125101e-01 9.51849818e-01
6.85051799e-01 5.67387998e-01 -7.44648743e-03 3.09638202e-01
8.92255247e-01 -9.72012162e-01 -2.45742366e-01 -2.22412720e-01
5.02254725e-01 -9.47974265e-01 9.55425501e-01 3.34477216e-01
-1.61166251e+00 -3.46483886e-01 -1.10780907e+00 -2.35449329e-01
-1.41945228e-01 -3.95461857e-01 9.37390149e-01 5.78945696e-01
-1.10209370e+00 9.57478404e-01 -8.13449323e-01 2.07429938e-02
7.64691591e-01 1.94123983e-01 -1.16772108e-01 8.14969391e-02
-1.01101315e+00 8.60676110e-01 -1.06697962e-01 -2.34260246e-01
-1.03051472e+00 -1.40477490e+00 -8.88077617e-01 -2.55423278e-01
1.27022147e-01 -1.11579883e+00 1.25308836e+00 -7.97811389e-01
-1.31986344e+00 7.91664720e-01 3.50920320e-01 -6.97711885e-01
9.59160328e-01 -4.43384677e-01 2.50336558e-01 3.35593708e-02
-2.57233620e-01 1.26897550e+00 9.58209038e-01 -1.67778921e+00
1.62915699e-02 4.19704877e-02 4.23416376e-01 2.41677284e-01
-3.35404240e-02 -1.94693103e-01 -4.51998591e-01 -1.05520093e+00
-3.72755140e-01 -9.72038507e-01 -6.11843705e-01 3.65916729e-01
-5.89079201e-01 2.83042133e-01 9.28878307e-01 -4.70223427e-01
8.92583966e-01 -1.91269302e+00 8.90307054e-02 1.77813724e-01
3.46046835e-01 1.75037757e-01 -3.93677860e-01 1.86701193e-01
-4.52840850e-02 7.38424540e-01 -4.34865892e-01 -6.08581126e-01
8.86734575e-02 7.16175064e-02 -6.51163816e-01 1.51211485e-01
4.21956122e-01 1.15393066e+00 -5.51735580e-01 -2.75959551e-01
2.66614079e-01 7.21847355e-01 -1.27501500e+00 -6.35133386e-02
-5.41355431e-01 5.52117050e-01 -4.12293226e-01 5.12091219e-01
9.93736744e-01 1.06357560e-01 -4.00109172e-01 -3.88624258e-02
1.05613492e-01 -2.12938562e-02 -8.47262740e-01 1.81150091e+00
-4.60723877e-01 5.50137937e-01 3.78948003e-01 -4.24906194e-01
6.66797340e-01 1.21614218e-01 2.12358162e-01 -4.24173951e-01
1.25877976e-01 -3.46669629e-02 1.65963154e-02 -1.11177869e-01
6.78709090e-01 -1.47075385e-01 -7.89831504e-02 5.60561419e-01
-3.58859628e-01 -9.24328566e-01 1.11299427e-03 5.52803934e-01
1.13251495e+00 5.64844489e-01 -5.09118199e-01 -2.10015580e-01
-2.13930011e-01 8.38361084e-02 2.77170569e-01 1.01745033e+00
2.04841390e-01 1.31705904e+00 5.33811927e-01 -2.31794670e-01
-1.49133646e+00 -1.32387304e+00 -2.05698937e-01 6.05063796e-01
1.97147205e-01 -4.30692285e-01 -9.73775089e-01 -5.89687645e-01
3.06831270e-01 8.68641734e-01 -7.36001492e-01 -8.71617645e-02
-9.16776717e-01 -6.32026553e-01 6.97115123e-01 6.54494703e-01
1.64447874e-01 -1.15048313e+00 -4.93951619e-01 -4.21555638e-02
2.54770547e-01 -8.73729229e-01 -6.48918629e-01 -1.49410293e-01
-9.16282892e-01 -8.48488629e-01 -7.94107020e-01 -4.95418072e-01
6.92744076e-01 -1.02328241e-01 1.56812406e+00 3.73151571e-01
-5.37903309e-01 4.39344317e-01 -1.87428549e-01 -6.77973747e-01
-7.66238391e-01 -2.04679687e-02 -1.69110551e-01 -7.75552094e-01
-5.39690733e-01 -8.92897904e-01 -7.71097243e-01 6.76291585e-02
-1.11053634e+00 2.95617700e-01 3.04119855e-01 4.90112722e-01
5.02141416e-01 -2.73366094e-01 2.87087619e-01 -9.89376962e-01
6.91692412e-01 -2.14849442e-01 -5.93719840e-01 -3.34921002e-01
-3.12095016e-01 1.15795843e-01 7.00209498e-01 -4.67678040e-01
-9.37069297e-01 -1.09025873e-01 -1.66288406e-01 -6.71391189e-01
-5.15009984e-02 1.15807496e-01 -4.23766039e-02 -3.41283023e-01
9.98165369e-01 -3.14824641e-01 8.01748037e-02 -3.81676584e-01
7.80506730e-01 -1.27650991e-01 6.30515933e-01 -9.71083879e-01
1.18949676e+00 5.48163235e-01 1.62498191e-01 -3.28883708e-01
-6.33598268e-01 4.44828242e-01 -4.39397216e-01 -1.06549151e-01
9.21551108e-01 -6.61217988e-01 -4.10577059e-01 4.89399672e-01
-1.23635530e+00 -8.48719060e-01 -5.86228132e-01 -3.81296128e-03
-8.40673208e-01 2.47771844e-01 -8.55724752e-01 -6.22634411e-01
-4.02989417e-01 -1.10751092e+00 1.12202787e+00 5.29762544e-02
-2.90359735e-01 -1.08254981e+00 -8.71644393e-02 2.48125922e-02
6.56526923e-01 8.72189462e-01 8.55395317e-01 -1.02609187e-01
-9.76053596e-01 -2.14761794e-01 -1.43788189e-01 3.27401340e-01
-2.32411727e-01 2.07548678e-01 -1.00145423e+00 -2.86039487e-02
-2.33033001e-01 -3.24670613e-01 8.96950006e-01 6.27120793e-01
1.29308951e+00 -4.56177950e-01 -4.85568464e-01 9.39500630e-01
1.12469780e+00 -1.73162982e-01 1.03314269e+00 4.53595556e-02
9.77512360e-01 4.03547049e-01 2.74523944e-01 1.92162931e-01
-1.55828744e-01 7.30280221e-01 6.01912260e-01 -4.05113637e-01
-6.32731557e-01 -4.06040668e-01 1.92458108e-01 2.47973502e-01
-9.11672413e-02 -5.13954937e-01 -6.77123070e-01 3.71324331e-01
-1.62055612e+00 -8.74104559e-01 -7.25829676e-02 2.06236267e+00
7.94770837e-01 4.44248319e-01 -5.94293177e-02 -1.80727527e-01
3.79955322e-01 1.86499387e-01 -5.60959160e-01 -5.83419561e-01
-7.69925267e-02 5.82077146e-01 6.00442052e-01 7.65044451e-01
-1.30871511e+00 1.02339840e+00 7.55985165e+00 8.86938751e-01
-9.77643609e-01 -1.58213466e-01 8.11345637e-01 -6.11655295e-01
-8.43375504e-01 -7.54279867e-02 -4.76035893e-01 3.30530137e-01
4.20865268e-01 -2.17368215e-01 3.07676941e-01 7.68689454e-01
2.34033301e-01 7.67734870e-02 -1.23547792e+00 5.71326613e-01
7.83262625e-02 -1.65968728e+00 3.23025584e-01 1.15742892e-01
1.03115022e+00 -1.68210879e-01 3.16881835e-01 1.27043262e-01
7.50759959e-01 -1.35699463e+00 1.03465605e+00 7.12354302e-01
8.20536017e-01 -8.42695296e-01 2.94419467e-01 1.09617449e-01
-7.76691914e-01 3.56991529e-01 -3.02239090e-01 7.08036572e-02
5.67211866e-01 5.24162054e-01 -6.42976403e-01 4.31549072e-01
4.20394480e-01 4.81637001e-01 -5.85028470e-01 1.06365013e+00
-3.51941377e-01 5.73823810e-01 -4.46523875e-01 3.73515159e-01
8.83025974e-02 -3.28107662e-02 1.05660498e+00 1.04811513e+00
2.38763794e-01 -8.35771579e-03 1.91111237e-01 1.31200659e+00
-1.71370521e-01 -3.19336861e-01 -8.46061468e-01 2.60048419e-01
3.45143348e-01 1.05437684e+00 -6.09567165e-01 -1.03106827e-01
2.50521395e-02 1.08657038e+00 1.63695857e-01 4.32964563e-01
-1.04261172e+00 -2.41237044e-01 8.25568497e-01 6.14336491e-01
2.04111606e-01 -3.55059355e-01 -9.20856893e-01 -7.96444893e-01
-9.94686112e-02 -6.82308972e-01 -3.31945531e-02 -1.15221417e+00
-1.42723072e+00 3.87923926e-01 1.47996396e-01 -1.03349483e+00
-3.06541890e-01 -5.30303061e-01 -9.96118069e-01 9.80136454e-01
-1.11584449e+00 -1.29942620e+00 -2.30090365e-01 1.82179451e-01
4.15286958e-01 2.95951307e-01 7.70356119e-01 2.91230436e-02
-2.06945956e-01 8.93202901e-01 -3.18247885e-01 3.30663957e-02
8.33323240e-01 -1.33759022e+00 8.61957312e-01 9.77452099e-01
-6.86929896e-02 5.33277333e-01 8.37776124e-01 -8.66129279e-01
-1.16991985e+00 -1.22663701e+00 1.96165834e-02 -1.04050910e+00
3.95360887e-01 -4.46390301e-01 -7.94125497e-01 6.32884324e-01
2.83716440e-01 -6.81806430e-02 2.09479824e-01 -9.59875956e-02
-4.20893997e-01 3.76350462e-01 -1.23060417e+00 1.04181969e+00
1.48498034e+00 -6.14570118e-02 -1.26472846e-01 3.23495239e-01
1.23438966e+00 -9.19982553e-01 -7.47276008e-01 4.76850957e-01
4.12224323e-01 -9.92199719e-01 1.32539356e+00 -6.06347024e-01
1.01729000e+00 -2.48668745e-01 1.08400472e-01 -1.43152308e+00
-3.93645138e-01 -7.31976509e-01 -1.93159252e-01 1.12253797e+00
6.24357879e-01 -3.71006936e-01 1.01763165e+00 8.95519793e-01
-5.97924352e-01 -9.21930909e-01 -5.19321859e-01 -7.72471905e-01
7.54630744e-01 -6.29038930e-01 5.42921245e-01 6.44030571e-01
-4.12944347e-01 -9.86221358e-02 -4.27967101e-01 6.17718324e-02
7.22704172e-01 -8.72379262e-03 9.91397083e-01 -7.81128943e-01
-3.91037941e-01 -6.87279046e-01 -2.75171131e-01 -9.83387947e-01
-6.17007278e-02 -1.06799328e+00 1.27933165e-02 -1.58696175e+00
1.98482901e-01 -8.16695869e-01 2.26390630e-01 4.09510136e-01
-3.73954564e-01 7.89260983e-01 4.93538171e-01 1.87639613e-02
-1.70219198e-01 4.40972984e-01 1.68096626e+00 -3.50804999e-02
-2.55276799e-01 -1.85554206e-01 -9.89691913e-01 8.01147282e-01
6.90469086e-01 -3.13091189e-01 -4.72015560e-01 -9.05230701e-01
1.80365041e-01 -3.40933442e-01 7.86462843e-01 -1.11306834e+00
-2.15367347e-01 7.59009346e-02 6.74078941e-01 -3.76891792e-01
4.00382370e-01 -4.14149582e-01 2.17044026e-01 2.17027947e-01
-3.39335859e-01 -2.57614464e-01 7.01653183e-01 3.37715983e-01
3.31362367e-01 -2.30316684e-01 8.43397200e-01 -1.33321494e-01
-6.31715730e-02 5.53098440e-01 -2.33316660e-01 3.30135912e-01
8.80700052e-01 -1.71881184e-01 -3.49467427e-01 -5.93819082e-01
-8.44320536e-01 1.66878521e-01 9.18178737e-01 5.00753462e-01
4.15795058e-01 -1.48017895e+00 -9.29621160e-01 5.42230420e-02
-3.80842894e-01 4.80836511e-01 2.65473634e-01 3.35453928e-01
-7.50144720e-01 -1.24965169e-01 1.23349927e-01 -6.20287538e-01
-1.05899262e+00 4.92543399e-01 3.70347828e-01 -2.80914366e-01
-8.37189019e-01 1.10517275e+00 6.20706499e-01 -3.73185366e-01
1.49477944e-01 -1.31323546e-01 3.55417252e-01 -3.31949919e-01
3.28413010e-01 1.57807782e-01 -1.48504257e-01 -3.53388935e-02
-7.37551302e-02 6.09543741e-01 -1.40290946e-01 -3.12253445e-01
1.21001315e+00 2.88778275e-01 2.55696982e-01 -9.08619165e-02
7.91166961e-01 4.30173039e-01 -1.78721952e+00 4.10390884e-01
-6.95866466e-01 -4.22334224e-01 -1.30839229e-01 -9.14308727e-01
-1.21385229e+00 8.38880777e-01 3.85944128e-01 2.10939869e-01
8.29999566e-01 -4.87704873e-02 8.48002017e-01 -1.85181603e-01
3.03402692e-01 -7.98319757e-01 2.87107617e-01 5.50398767e-01
1.47012722e+00 -7.75013804e-01 2.00142205e-01 -7.53650844e-01
-5.59050977e-01 7.02752113e-01 8.03693056e-01 -6.15728796e-01
5.50787807e-01 9.14498448e-01 1.65071756e-01 -5.67682795e-02
-5.48072755e-01 1.51739538e-01 2.29024827e-01 9.56201136e-01
2.44305268e-01 -1.27311662e-01 1.39805544e-02 4.75007832e-01
-8.51317465e-01 -1.44242540e-01 5.10022819e-01 8.17491889e-01
-8.05050209e-02 -1.21892071e+00 -4.17314053e-01 4.19924438e-01
-5.42288661e-01 -3.35097849e-01 -4.61153716e-01 8.73194098e-01
1.45765781e-01 6.71521962e-01 1.77106321e-01 -2.73966312e-01
2.88207948e-01 -1.74104765e-01 9.21080053e-01 -6.59633636e-01
-6.99469030e-01 4.58320230e-02 1.07138619e-01 -7.23501861e-01
8.26595351e-02 -3.97289604e-01 -1.21418560e+00 -5.04584610e-01
-2.82062918e-01 -1.86836451e-01 4.53848273e-01 4.89296824e-01
6.27260625e-01 7.75524497e-01 4.61663514e-01 -1.44289470e+00
-2.01343566e-01 -6.99650884e-01 -1.37656644e-01 6.20150864e-01
1.23417154e-01 -4.44743454e-01 -5.76502442e-01 1.81951150e-01] | [11.316457748413086, -0.4792816936969757] |
539a82b7-bdea-40b9-ae5a-3786bdc932a8 | s3lam-structured-scene-slam | 2109.07339 | null | https://arxiv.org/abs/2109.07339v2 | https://arxiv.org/pdf/2109.07339v2.pdf | S3LAM: Structured Scene SLAM | We propose a new SLAM system that uses the semantic segmentation of objects and structures in the scene. Semantic information is relevant as it contains high level information which may make SLAM more accurate and robust. Our contribution is twofold: i) A new SLAM system based on ORB-SLAM2 that creates a semantic map made of clusters of points corresponding to objects instances and structures in the scene. ii) A modification of the classical Bundle Adjustment formulation to constrain each cluster using geometrical priors, which improves both camera localization and reconstruction and enables a better understanding of the scene. We evaluate our approach on sequences from several public datasets and show that it improves camera pose estimation with respect to state of the art. | ['Jérôme Royan', 'Amine Kacete', 'Eric Marchand', 'Mathieu Gonzalez'] | 2021-09-15 | null | null | null | null | ['camera-localization'] | ['computer-vision'] | [ 7.26785064e-02 -1.08530104e-01 5.45320846e-02 -5.51051259e-01
-4.72985655e-01 -7.56597698e-01 7.05560684e-01 2.18832627e-01
-3.72635543e-01 3.64933789e-01 6.90224990e-02 1.52237996e-01
-1.76747650e-01 -7.58019209e-01 -9.56663966e-01 -2.98358738e-01
2.52679497e-01 1.01408803e+00 7.73492157e-01 -1.80488363e-01
5.22680402e-01 7.90191114e-01 -1.58803558e+00 -7.89436325e-02
5.58971345e-01 7.77996063e-01 8.78558517e-01 3.77574086e-01
-1.40258297e-01 4.64950770e-01 -1.37405515e-01 -1.56110615e-01
5.94184577e-01 -1.38242915e-01 -9.66899276e-01 6.17976427e-01
7.16328263e-01 9.75948200e-02 -6.65620267e-02 1.18909919e+00
-1.13263696e-01 6.96936771e-02 4.41808760e-01 -1.22166240e+00
3.58804643e-01 -2.67337402e-03 -5.15183389e-01 -3.70643109e-01
6.90770686e-01 -1.45497620e-01 9.77010608e-01 -7.58443415e-01
1.00043893e+00 1.13455176e+00 1.01141632e+00 -9.11576152e-02
-1.28518236e+00 -2.56727695e-01 -1.50978506e-01 4.22427565e-01
-1.72523570e+00 -4.10300076e-01 7.92248249e-01 -5.30832708e-01
9.49452400e-01 3.59415151e-02 9.22701240e-01 6.33568704e-01
1.17686823e-01 5.69922447e-01 9.79228795e-01 -5.33164680e-01
3.20078671e-01 1.13583423e-01 -1.22697465e-02 6.58555567e-01
3.65969300e-01 -1.30616561e-01 -6.42141402e-01 -4.45022173e-02
9.67558384e-01 1.41012117e-01 4.94040065e-02 -1.38828790e+00
-1.51252568e+00 8.51880550e-01 6.57293677e-01 6.31401837e-02
-6.02856219e-01 2.54668713e-01 -4.47884053e-02 -2.67054945e-01
3.77750732e-02 4.46766347e-01 -2.85569370e-01 -9.65647865e-03
-1.04319406e+00 1.60913542e-01 6.61989212e-01 1.36396241e+00
1.49810767e+00 -5.54256618e-01 7.36972809e-01 4.64901000e-01
6.64914846e-01 7.10420191e-01 -9.55991223e-02 -1.37916493e+00
2.85396188e-01 7.63330221e-01 2.09447339e-01 -1.26195693e+00
-5.92612982e-01 9.45687760e-03 -1.93816140e-01 8.21458027e-02
3.46653759e-02 5.08538127e-01 -8.32270980e-01 1.11838341e+00
4.29344684e-01 3.81638974e-01 -1.22523367e-01 9.26129520e-01
6.18265450e-01 3.25782776e-01 -4.36169297e-01 3.46489817e-01
1.37490106e+00 -1.00588751e+00 -5.99855840e-01 -5.97872257e-01
4.77912426e-01 -1.01279366e+00 6.06511772e-01 3.07040155e-01
-6.09100878e-01 -4.63667512e-01 -1.08157086e+00 -9.68431681e-02
-3.28508526e-01 -4.41608689e-04 8.54952276e-01 4.55376387e-01
-1.14422035e+00 4.81013298e-01 -9.69087422e-01 -1.06154549e+00
9.56686735e-02 3.41229141e-01 -6.65581763e-01 -2.35868365e-01
-5.33728302e-01 1.15662384e+00 6.89386308e-01 -1.04268193e-01
-7.69553840e-01 -2.58246362e-01 -1.32967162e+00 -4.19363499e-01
5.59690893e-01 -7.96870708e-01 9.76070523e-01 -5.40664613e-01
-1.37088490e+00 1.20580554e+00 -3.20843041e-01 -4.57604975e-01
5.62946498e-01 -3.55540812e-01 4.63250317e-02 2.84960240e-01
4.47547317e-01 9.54637110e-01 5.11971831e-01 -1.56005883e+00
-7.74246633e-01 -4.00116712e-01 -7.27733672e-02 4.56397086e-01
6.31638885e-01 -9.98890102e-02 -9.41089034e-01 -3.86346653e-02
9.95003879e-01 -1.30030751e+00 -4.78691310e-01 -2.97027379e-01
-3.93276066e-01 2.25128099e-01 7.66322374e-01 -6.00659430e-01
4.96754706e-01 -1.88149738e+00 3.68879169e-01 5.20881474e-01
5.16339727e-02 -4.13412571e-01 -2.56514847e-02 6.61240935e-01
1.80718735e-01 -2.40670905e-01 -2.54205972e-01 -6.43277109e-01
-7.25230575e-02 7.45800972e-01 -2.56915428e-02 9.46187556e-01
-3.04451764e-01 7.09792376e-01 -7.38008082e-01 -6.10703051e-01
1.13878787e+00 2.66621739e-01 -5.26592851e-01 1.30558133e-01
-9.11814719e-02 6.76846802e-01 -2.14341432e-01 6.22590482e-01
9.16266799e-01 -1.50274998e-02 -5.80867082e-02 -6.36207759e-02
-4.30326104e-01 3.55681658e-01 -1.72814572e+00 2.42941928e+00
-2.21233338e-01 5.47128618e-01 5.99850789e-02 -6.18391573e-01
9.92354393e-01 -8.00656974e-02 6.62513494e-01 -2.05588102e-01
1.04433291e-01 -3.76429618e-03 -6.08998120e-01 -1.82545215e-01
8.61931145e-01 9.87546965e-02 -1.50259823e-01 -1.49200991e-01
2.64524966e-01 -9.39964354e-01 -3.52491960e-02 3.35531205e-01
9.11646068e-01 5.71826577e-01 5.71309328e-01 -5.94076574e-01
4.99214232e-01 5.16716480e-01 6.07177854e-01 7.63013959e-01
9.69268288e-03 8.09614658e-01 6.41952008e-02 -5.28226316e-01
-1.12476540e+00 -1.22492278e+00 -1.61730528e-01 3.26243520e-01
8.61029983e-01 -6.43787920e-01 -5.99644780e-01 -4.84996170e-01
1.28849253e-01 5.58208466e-01 -3.61981750e-01 3.26214284e-01
-2.67383486e-01 -3.74882579e-01 -1.97178751e-01 2.11125076e-01
5.72631657e-01 -8.10858309e-01 -7.53239989e-01 7.74998665e-02
-5.44653118e-01 -1.55141878e+00 6.35497784e-03 2.26859041e-02
-9.50450480e-01 -1.45738876e+00 6.42127618e-02 -5.06467700e-01
7.87466466e-01 6.11064494e-01 1.17243922e+00 -1.13979965e-01
-3.03045124e-01 7.26625860e-01 -6.25640154e-01 -3.80571455e-01
-3.21400642e-01 -2.65508369e-02 2.27575675e-01 -3.85245048e-02
4.21223730e-01 -3.22859347e-01 -1.34604484e-01 6.37371719e-01
-4.38837141e-01 2.72145778e-01 4.75451112e-01 2.14314520e-01
8.22073042e-01 2.12549530e-02 -5.40706038e-01 -7.62202144e-01
-3.26264948e-01 -1.26368433e-01 -1.19468653e+00 -2.65071541e-02
-4.89964366e-01 3.07639781e-02 -8.27365071e-02 3.73868912e-01
-8.10238540e-01 9.53592062e-01 -1.14368591e-02 -3.55521083e-01
-4.61611062e-01 1.38926163e-01 -3.42335522e-01 -5.72113335e-01
4.72686023e-01 1.02516904e-01 4.43739183e-02 -6.98349416e-01
5.61261237e-01 3.52227956e-01 7.47966826e-01 -3.44057232e-01
9.90755022e-01 1.02237427e+00 4.55758601e-01 -1.08372962e+00
-7.66676009e-01 -1.42448223e+00 -1.60421658e+00 -2.41875023e-01
9.72314477e-01 -1.02245092e+00 -6.05969071e-01 2.21354857e-01
-1.22088015e+00 -6.10158257e-02 -1.64261237e-01 8.92104805e-01
-1.07059157e+00 6.49194837e-01 -2.80935884e-01 -5.21539450e-01
3.82446885e-01 -1.41976798e+00 1.49025404e+00 -9.75137278e-02
-2.42968023e-01 -9.26702619e-01 2.63442934e-01 4.99069929e-01
-2.16310710e-01 4.25445884e-01 1.50137525e-02 -2.54883766e-01
-1.12944019e+00 -1.49539024e-01 -8.65089595e-02 1.40782043e-01
-9.18985903e-02 -5.80305345e-02 -8.38334084e-01 -1.43123403e-01
3.76290381e-02 2.74171114e-01 7.07997799e-01 3.66764039e-01
4.62716818e-01 2.33880848e-01 -7.10287333e-01 9.05480802e-01
1.82647467e+00 -1.65291354e-01 8.32753062e-01 8.62743437e-01
9.55231488e-01 6.90923870e-01 9.24541950e-01 4.34736490e-01
6.03700221e-01 1.09340370e+00 9.00345504e-01 -4.36631516e-02
9.55481306e-02 -4.14155126e-01 1.18610281e-02 7.44615912e-01
-1.02113374e-01 2.97146320e-01 -1.20136690e+00 4.94537503e-01
-2.11153364e+00 -5.58441401e-01 -7.41274953e-01 2.20084929e+00
5.95691353e-02 -2.24583760e-01 8.26488808e-03 -8.57336968e-02
5.88990986e-01 5.13946265e-02 -2.38518585e-02 1.25976935e-01
-1.97181165e-01 -1.80391863e-01 1.24495709e+00 9.14738655e-01
-1.34574413e+00 1.37089896e+00 6.75939035e+00 3.23485136e-01
-4.69813526e-01 1.25540614e-01 -5.78041911e-01 4.70741868e-01
2.27242619e-01 6.62397087e-01 -1.01181436e+00 1.40216365e-01
5.71633160e-01 2.05935523e-01 3.59177798e-01 1.07216525e+00
5.37270345e-02 -6.95582747e-01 -1.09864151e+00 1.32384193e+00
3.92871857e-01 -1.33935630e+00 -1.93133235e-01 5.34144223e-01
8.05185556e-01 6.07500017e-01 -6.36369765e-01 -3.05559307e-01
5.36103845e-01 -5.16955912e-01 1.09350431e+00 5.21636665e-01
1.73963696e-01 -7.06564486e-01 9.91107047e-01 5.59333622e-01
-1.33835018e+00 3.16587776e-01 -6.47612810e-01 -2.39314631e-01
3.69653076e-01 5.17844439e-01 -1.48902774e+00 8.95413637e-01
8.76791298e-01 1.13097548e+00 -9.70538199e-01 1.32007790e+00
-6.61129653e-01 1.22523345e-01 -6.47566974e-01 4.00109321e-01
2.68283069e-01 -6.96327388e-01 7.62234867e-01 1.14143252e+00
3.54051560e-01 -2.05274209e-01 5.60989082e-01 7.62211561e-01
3.96391809e-01 3.74989584e-02 -9.26263690e-01 6.29650116e-01
4.30095077e-01 1.27456820e+00 -1.17667019e+00 -2.59390444e-01
-2.67874479e-01 1.16654909e+00 -3.49153168e-02 -2.18614005e-02
-5.76665759e-01 2.68947612e-02 7.33617067e-01 2.18457028e-01
2.67617702e-01 -7.96612084e-01 -2.80721128e-01 -1.24124730e+00
-1.02263547e-01 -3.52379888e-01 3.47502321e-01 -1.11181939e+00
-8.84663463e-01 2.46651590e-01 4.35626134e-02 -1.49744177e+00
-2.18903914e-01 -5.90795398e-01 4.10985984e-02 7.21205711e-01
-1.32673812e+00 -1.38207555e+00 -7.81654358e-01 4.94490594e-01
4.95508403e-01 1.57584980e-01 6.53598368e-01 -7.43831471e-02
4.66361977e-02 -3.87820154e-01 2.20865473e-01 -2.79851034e-02
5.84438860e-01 -1.25166106e+00 4.20309901e-01 1.14153814e+00
6.46371186e-01 3.95678610e-01 8.63083959e-01 -8.10982585e-01
-1.35934520e+00 -1.01045561e+00 7.93810487e-01 -1.19980991e+00
4.73024964e-01 -7.33329475e-01 -4.99240726e-01 1.10987818e+00
-4.20649983e-02 -2.36078247e-01 2.44747818e-01 1.27866060e-01
-7.16777146e-02 3.98943983e-02 -1.02254713e+00 1.80104628e-01
1.06771505e+00 -5.14803290e-01 -9.00219440e-01 5.01302540e-01
5.27845025e-01 -6.04402661e-01 -6.21108890e-01 4.86308426e-01
2.39971459e-01 -1.26893222e+00 1.17289257e+00 1.75330549e-01
-2.65935361e-01 -8.43154013e-01 -6.77011371e-01 -1.26978016e+00
-2.07963735e-01 -4.75328922e-01 4.31551695e-01 9.33446467e-01
-1.52955130e-01 -5.86069226e-01 9.82429206e-01 2.45342329e-01
-3.26627702e-01 2.30241388e-01 -1.01637232e+00 -1.03573179e+00
-7.74355054e-01 -8.15848410e-01 6.10750973e-01 9.75515783e-01
-3.41441721e-01 1.87382609e-01 -2.34368667e-01 6.45138323e-01
1.03980708e+00 1.60257265e-01 1.31906509e+00 -1.62316763e+00
1.07731119e-01 -1.61676526e-01 -1.20768762e+00 -1.05254650e+00
2.21969336e-01 -9.07147527e-01 2.97071695e-01 -1.83293641e+00
1.13375835e-01 -4.67301935e-01 2.63983011e-01 3.23791981e-01
4.66763973e-01 3.72330308e-01 2.18737766e-01 5.41849673e-01
-1.06767726e+00 3.96490544e-01 5.65941632e-01 2.11564943e-01
-3.36944126e-03 -1.60096586e-01 2.95470692e-02 1.21765006e+00
3.10759753e-01 -5.77761054e-01 1.59681529e-01 -4.18361753e-01
1.63787365e-01 -1.77419320e-01 6.17097557e-01 -1.38374031e+00
2.56690979e-01 -9.53452736e-02 2.27440253e-01 -1.26395273e+00
6.32497072e-01 -1.33903551e+00 6.13089263e-01 5.28866351e-01
3.72416943e-01 2.61398517e-02 -1.88745558e-02 6.25990093e-01
-2.31477097e-01 -5.31137109e-01 7.51561582e-01 -4.22200859e-01
-1.48705864e+00 -1.45776020e-02 -3.56431380e-02 -6.06287181e-01
1.15539610e+00 -4.59979266e-01 2.76758134e-01 -4.07260329e-01
-6.24290049e-01 2.01351479e-01 1.46464145e+00 5.11600733e-01
5.04763365e-01 -1.25181854e+00 -4.12004620e-01 3.72504175e-01
5.99873185e-01 3.70829701e-01 -5.72076403e-02 8.01392674e-01
-1.32135701e+00 6.68914914e-01 -1.96373254e-01 -1.43895113e+00
-1.36656451e+00 5.41390538e-01 1.69171751e-01 3.90100509e-01
-7.30381072e-01 5.29872596e-01 2.81101137e-01 -8.51937830e-01
4.91012074e-02 -2.20597535e-01 -1.37464134e-02 -1.36808842e-01
2.00923473e-01 5.04889429e-01 8.61453861e-02 -1.35541856e+00
-8.26883197e-01 1.16529143e+00 6.55708015e-01 -2.54663467e-01
1.32453537e+00 -7.19287992e-01 -6.49732113e-01 5.26019454e-01
9.48496044e-01 1.41200915e-01 -1.24197280e+00 -3.10453802e-01
3.83745581e-01 -9.45236504e-01 7.97239318e-02 -4.03590530e-01
-5.14468133e-01 7.08391130e-01 5.01588643e-01 -1.59617648e-01
6.94173634e-01 3.44575644e-01 3.07874680e-01 4.99525577e-01
1.21223402e+00 -1.06433189e+00 -1.77687198e-01 6.80926621e-01
6.52635694e-01 -1.37583411e+00 3.49731058e-01 -8.10216248e-01
-6.55960977e-01 9.44620967e-01 1.31168634e-01 -1.63499549e-01
3.99801254e-01 6.06327876e-02 2.10653469e-01 -3.90596986e-01
1.08007200e-01 -5.62929094e-01 3.09491366e-01 7.21054316e-01
-2.15319723e-01 1.16595395e-01 8.55043977e-02 -1.55355632e-01
-3.86877954e-01 -2.81719357e-01 5.39290190e-01 1.01653469e+00
-7.26984859e-01 -1.10219181e+00 -8.49586010e-01 -1.43430337e-01
1.76233277e-01 1.26342207e-01 -4.25971925e-01 9.21652496e-01
3.05230916e-01 9.97839391e-01 1.59956500e-01 -3.91663730e-01
3.80461603e-01 1.53214037e-02 6.77327812e-01 -8.15742850e-01
-2.12296052e-03 3.25175405e-01 -1.06618656e-02 -1.08151460e+00
-7.56063581e-01 -1.00328767e+00 -1.23260880e+00 -1.26907885e-01
-3.82930785e-01 1.63200781e-01 1.47567105e+00 8.21046472e-01
2.88442910e-01 2.91848518e-02 4.99209255e-01 -1.28727937e+00
5.64039163e-02 -8.30654502e-01 -9.07507241e-01 6.10433459e-01
7.96410739e-02 -9.78081226e-01 -1.46051660e-01 3.01416993e-01] | [7.348358154296875, -2.2714362144470215] |
f568405f-350d-4f57-a3fe-3586fa1f504b | active-detection-and-localization-of | 1603.07022 | null | http://arxiv.org/abs/1603.07022v1 | http://arxiv.org/pdf/1603.07022v1.pdf | Active Detection and Localization of Textureless Objects in Cluttered Environments | This paper introduces an active object detection and localization framework
that combines a robust untextured object detection and 3D pose estimation
algorithm with a novel next-best-view selection strategy. We address the
detection and localization problems by proposing an edge-based registration
algorithm that refines the object position by minimizing a cost directly
extracted from a 3D image tensor that encodes the minimum distance to an edge
point in a joint direction/location space. We face the next-best-view problem
by exploiting a sequential decision process that, for each step, selects the
next camera position which maximizes the mutual information between the state
and the next observations. We solve the intrinsic intractability of this
solution by generating observations that represent scene realizations, i.e.
combination samples of object hypothesis provided by the object detector, while
modeling the state by means of a set of constantly resampled particles.
Experiments performed on different real world, challenging datasets confirm the
effectiveness of the proposed methods. | ['Alberto Pretto', 'Marco Imperoli'] | 2016-03-22 | null | null | null | null | ['active-object-detection'] | ['computer-vision'] | [ 2.21765548e-01 -1.26236096e-01 1.61078006e-01 -9.15410072e-02
-6.75225317e-01 -4.93951261e-01 6.21274889e-01 1.79738566e-01
-8.14505160e-01 3.80015761e-01 -2.14654356e-01 3.78723472e-01
-4.61170465e-01 -5.58595479e-01 -8.05437803e-01 -1.08957303e+00
1.94022909e-01 1.03474808e+00 5.19569337e-01 2.50865519e-01
5.05010903e-01 7.13622570e-01 -1.50514722e+00 -3.46706480e-01
4.79721576e-01 1.01686990e+00 8.35462630e-01 7.72323430e-01
4.23789948e-01 3.29483837e-01 -2.34623358e-01 3.43332104e-02
3.88390094e-01 -1.81981605e-02 -1.44857332e-01 8.32289279e-01
2.27234080e-01 -1.93633839e-01 2.18467891e-01 1.13255429e+00
4.85161185e-01 4.34760064e-01 5.56522429e-01 -1.00897133e+00
-7.19826370e-02 -1.62480772e-01 -5.16033888e-01 2.64303893e-01
3.90505672e-01 2.91197538e-01 8.32853973e-01 -1.13279033e+00
8.37633371e-01 1.02304935e+00 2.89304882e-01 2.16322169e-01
-1.31148195e+00 -1.22465521e-01 2.84789801e-01 1.39723167e-01
-1.46954286e+00 -3.26310545e-01 1.15279269e+00 -6.77206874e-01
6.16243184e-01 1.31342202e-01 8.22633684e-01 8.34463775e-01
3.68394017e-01 4.57048178e-01 1.14596498e+00 -3.64524633e-01
5.72480202e-01 2.07956180e-01 1.20008744e-01 8.81555438e-01
4.87470806e-01 2.58954763e-01 -5.90796828e-01 -4.70803291e-01
7.27014005e-01 2.27232859e-01 -1.58161208e-01 -1.17820656e+00
-1.37059784e+00 6.33567929e-01 1.53424069e-01 -1.28884137e-01
-7.57957518e-01 -1.16119318e-01 -3.80009890e-01 -2.71694124e-01
5.12157857e-01 1.63049310e-01 -2.63980240e-01 2.27935746e-01
-5.65264821e-01 2.47615099e-01 5.86662948e-01 7.59178221e-01
8.64538491e-01 -2.98168391e-01 1.27040539e-02 2.74566054e-01
8.67174387e-01 8.48605633e-01 -4.57153767e-02 -8.67039979e-01
2.32793644e-01 7.28062451e-01 7.20613480e-01 -9.98693347e-01
-2.84554660e-01 -6.64441526e-01 -3.44149798e-01 5.53046882e-01
2.71582067e-01 -1.07106820e-01 -6.63562298e-01 1.54024339e+00
1.12054694e+00 3.64845753e-01 -5.57896495e-02 1.22804987e+00
1.93468317e-01 4.85207707e-01 -3.14571261e-01 -4.73650515e-01
1.21843457e+00 -6.29724920e-01 -3.19055200e-01 -2.53024131e-01
-5.12847751e-02 -7.08409369e-01 2.83177346e-01 4.40851867e-01
-1.15626514e+00 -6.36374652e-01 -9.78666306e-01 4.36369121e-01
-7.07457587e-02 2.19113171e-01 1.12722479e-01 2.17124805e-01
-6.97545350e-01 2.72955149e-01 -1.18199706e+00 -3.07497084e-01
-5.31020239e-02 5.48477709e-01 -3.98253709e-01 1.87179729e-01
-3.07396024e-01 9.03970063e-01 4.13466841e-01 3.61382127e-01
-1.13746703e+00 -3.37549716e-01 -5.39876580e-01 -3.44969898e-01
8.82119954e-01 -9.72410142e-01 7.91912675e-01 -5.15405655e-01
-1.55719876e+00 6.80520892e-01 -3.07324320e-01 -4.14907187e-01
6.43349051e-01 -2.74996817e-01 1.98813789e-02 1.89868823e-01
2.13610709e-01 2.27366120e-01 9.78324950e-01 -1.56728649e+00
-6.79102123e-01 -9.04359579e-01 -1.81984171e-01 6.63491547e-01
3.47192347e-01 -3.56040627e-01 -6.74216926e-01 -6.00053445e-02
7.86155760e-01 -1.46671081e+00 -6.08347535e-01 6.07302524e-02
-4.15431112e-01 -7.38100484e-02 9.54908788e-01 -2.60827780e-01
4.46720660e-01 -2.03915286e+00 6.03120744e-01 2.79538244e-01
1.72626138e-01 -2.00325966e-01 1.13220312e-01 2.18558967e-01
4.22566593e-01 -4.16081220e-01 -6.32413030e-02 -7.71965563e-01
-3.06730390e-01 -8.65391514e-04 3.21441926e-02 1.08652961e+00
4.46917675e-02 5.84426165e-01 -9.77187872e-01 -4.24176306e-01
6.41506314e-01 5.84684789e-01 -4.70082372e-01 3.82294208e-01
-3.11992705e-01 8.95860672e-01 -8.36296856e-01 2.99774498e-01
7.80706823e-01 -2.93044865e-01 -7.60837719e-02 -3.39018434e-01
-5.15610158e-01 -6.80579618e-02 -1.76298273e+00 1.72033918e+00
-3.67677838e-01 9.85597670e-02 3.20673019e-01 -6.45788014e-01
8.97013128e-01 1.23847529e-01 8.11464131e-01 -3.79579514e-02
2.73875922e-01 -1.53803648e-02 -1.90000430e-01 -5.49878061e-01
6.03024840e-01 4.61565033e-02 4.77405302e-02 2.49543220e-01
-5.60556129e-02 -1.41553611e-01 -1.03233829e-01 -3.73365320e-02
8.72863948e-01 3.50800812e-01 3.38492155e-01 -5.06480373e-02
5.32465577e-01 -7.44807497e-02 4.86876190e-01 9.99855459e-01
-1.47423074e-01 5.79538941e-01 -7.52862990e-02 -2.48212621e-01
-7.92115867e-01 -1.24179685e+00 -4.65242937e-02 3.51418138e-01
7.20540404e-01 1.75886407e-01 -4.59340543e-01 -7.35166430e-01
5.06460071e-02 5.35097241e-01 -6.60497069e-01 -9.74772722e-02
-6.03027642e-01 -7.81248093e-01 -6.10726297e-01 -2.33238004e-02
2.52694011e-01 -6.98451877e-01 -1.31244540e+00 3.81657898e-01
-1.66116804e-01 -1.15060556e+00 -2.83776760e-01 2.17959464e-01
-8.60459387e-01 -9.44915473e-01 -2.78790385e-01 -5.32771111e-01
9.71661329e-01 3.27712089e-01 7.92649269e-01 -2.00790331e-01
-2.24757776e-01 7.36579239e-01 -1.12261385e-01 -6.50088564e-02
-2.93083996e-01 -2.19714046e-01 3.33203256e-01 6.97743654e-01
-8.75925273e-02 -2.05743134e-01 -6.79668307e-01 4.07140225e-01
-3.53879750e-01 1.75906777e-01 6.40898347e-01 6.45963252e-01
1.02195144e+00 1.51667476e-01 2.13395897e-02 -4.43134993e-01
7.21911564e-02 -2.96990395e-01 -1.28558254e+00 1.61670908e-01
-3.78942847e-01 6.76629841e-02 -4.33203345e-03 -5.88451803e-01
-1.08059776e+00 7.41230488e-01 4.67947304e-01 -4.25058872e-01
-2.65083730e-01 1.85635999e-01 -1.96901366e-01 -2.55622357e-01
1.98826060e-01 4.21492308e-01 6.03850260e-02 -4.99324590e-01
3.58937174e-01 2.98878193e-01 3.71563494e-01 -3.62117350e-01
1.01944280e+00 9.94492590e-01 3.71288925e-01 -9.54449654e-01
-8.10413122e-01 -7.66395211e-01 -9.17469740e-01 -7.32974648e-01
1.04802632e+00 -7.13816822e-01 -1.00451386e+00 3.80452216e-01
-1.16384518e+00 4.88455780e-02 -3.58663201e-01 9.24160659e-01
-6.34876311e-01 4.43059742e-01 1.25398738e-02 -1.31578267e+00
-6.82402104e-02 -1.21461892e+00 1.53240728e+00 1.24357946e-01
1.41449437e-01 -9.19792950e-01 2.18241140e-01 4.11127627e-01
-9.61930454e-02 4.46371615e-01 2.66480327e-01 -3.14430475e-01
-1.28440046e+00 -3.96088392e-01 1.94039822e-01 3.51281688e-02
-1.18728913e-01 -1.41285146e-02 -5.76616466e-01 -3.81416529e-01
6.04304492e-01 2.66017646e-01 7.21467018e-01 6.56528473e-01
4.43252176e-01 -1.26257867e-01 -5.89678824e-01 4.55242813e-01
1.70069230e+00 2.81821579e-01 -1.04234599e-01 1.51836574e-01
5.25613904e-01 3.78484696e-01 9.23854530e-01 6.78366125e-01
3.40639591e-01 1.07520604e+00 8.81039619e-01 2.12137982e-01
6.34734333e-02 -6.57049417e-02 2.75420278e-01 6.71153486e-01
-1.65745452e-01 -4.26419377e-01 -6.46395266e-01 3.56433958e-01
-1.92197776e+00 -6.93031967e-01 -4.51115035e-02 2.61148977e+00
1.45914197e-01 4.43115473e-01 -7.36721903e-02 -2.25481302e-01
7.58620203e-01 -5.42590469e-02 -7.23668694e-01 7.10990965e-01
2.16330891e-03 -2.39156440e-01 6.70494914e-01 9.17491198e-01
-1.03528464e+00 6.27313256e-01 4.96874142e+00 1.32708237e-01
-1.07011247e+00 1.62510902e-01 -6.42279088e-02 -6.23951592e-02
2.13260785e-01 3.19942683e-01 -1.21417463e+00 4.48450506e-01
3.70345592e-01 2.27144305e-02 1.25194609e-01 6.28865659e-01
3.84728938e-01 -5.57055950e-01 -9.86167312e-01 8.62328410e-01
3.39892536e-01 -1.17979467e+00 -7.73237869e-02 3.37242097e-01
6.49542987e-01 2.97119468e-01 1.20788932e-01 -4.56638962e-01
1.12153955e-01 1.16302311e-01 1.18683171e+00 8.61195087e-01
-5.26675470e-02 -3.49367440e-01 3.28012228e-01 7.29681373e-01
-1.19945681e+00 -2.51436651e-01 -9.10646319e-02 3.02474853e-02
5.69029391e-01 6.44168019e-01 -1.10941374e+00 5.23426831e-01
3.60390574e-01 7.09466636e-01 -5.43157101e-01 1.31226349e+00
-7.46163446e-03 1.61903709e-01 -6.82048976e-01 -1.38839826e-01
-2.04957183e-02 -5.81356525e-01 1.47237766e+00 6.60243928e-01
3.18793684e-01 2.51792055e-02 5.62167704e-01 1.04407525e+00
5.48775256e-01 -2.59129167e-01 -4.66507435e-01 6.14450574e-01
3.72786760e-01 1.38233590e+00 -1.26432741e+00 -2.12833941e-01
-2.31554553e-01 8.99491549e-01 1.42690659e-01 3.68762761e-01
-7.09520400e-01 3.37632716e-01 1.85661554e-01 1.86023414e-01
6.35310352e-01 -6.59116626e-01 1.86188936e-01 -1.32704628e+00
1.80162266e-01 -1.90615624e-01 1.43938050e-01 -8.60120237e-01
-9.72068369e-01 5.01070559e-01 2.58311599e-01 -1.42385089e+00
-2.06136659e-01 -4.27671373e-01 -2.67970532e-01 6.46882534e-01
-1.19209468e+00 -8.82085919e-01 -2.44853348e-01 3.92810166e-01
6.10933244e-01 2.26182967e-01 3.11653376e-01 -2.27391236e-02
-5.10069191e-01 -3.33087981e-01 8.68424550e-02 -2.44813353e-01
1.17390886e-01 -1.08613217e+00 5.41183837e-02 9.57520545e-01
2.52230018e-01 3.93572271e-01 8.18143845e-01 -8.19699407e-01
-1.84241986e+00 -9.62796688e-01 4.80493009e-01 -8.10574889e-01
4.61522818e-01 -6.79999471e-01 -5.14095962e-01 5.71260393e-01
-3.23356837e-01 1.45570546e-01 -7.24597275e-02 -3.71897489e-01
2.64694959e-01 -1.18741676e-01 -1.05818021e+00 2.92646289e-01
8.44612420e-01 -2.26441562e-01 -4.60661352e-01 3.70313227e-01
3.03678870e-01 -5.64306796e-01 -4.75746661e-01 4.37149554e-01
3.63689125e-01 -7.84017324e-01 1.11671197e+00 -7.36353546e-02
-5.17539442e-01 -7.42188215e-01 -7.56936744e-02 -1.03789711e+00
-3.51377100e-01 -5.67244947e-01 -1.37676924e-01 8.22426438e-01
1.69096678e-01 -5.50359368e-01 1.00012362e+00 2.16231942e-01
4.83808368e-02 -6.07173145e-01 -1.33237505e+00 -7.34470963e-01
-7.99904287e-01 -2.84449250e-01 4.09808792e-02 3.62425685e-01
-6.99316382e-01 5.53766072e-01 -2.70754486e-01 8.67473364e-01
1.42041743e+00 4.49589431e-01 8.87557447e-01 -1.36111212e+00
-4.27632064e-01 8.64346512e-03 -7.11703539e-01 -1.22871470e+00
5.37147857e-02 -6.15732312e-01 3.49322110e-01 -1.28502512e+00
3.31426501e-01 -3.00279617e-01 -8.64804536e-02 -2.62842745e-01
-2.26509254e-02 2.31154375e-02 2.15307578e-01 1.73755646e-01
-9.05760586e-01 4.95073080e-01 1.11800003e+00 1.23728067e-01
-2.76705444e-01 3.04793566e-01 3.75529341e-02 9.35558558e-01
1.96411878e-01 -6.71717405e-01 -3.03033292e-01 -4.61107045e-01
1.40658423e-01 4.42129582e-01 8.29980552e-01 -1.12831759e+00
4.03645098e-01 -3.64748597e-01 2.86510438e-01 -1.14105499e+00
8.53766263e-01 -1.25817764e+00 5.73857903e-01 5.07015646e-01
-2.08354607e-01 1.34153785e-02 -3.17294717e-01 1.07150352e+00
2.67179430e-01 -3.37456435e-01 9.88130271e-01 -1.64166421e-01
-5.81440806e-01 3.45268250e-01 -2.59243935e-01 -2.78386623e-01
1.27760768e+00 -2.65222400e-01 1.80199414e-01 -1.35327980e-01
-1.11129284e+00 2.53779721e-02 6.71828032e-01 2.48404413e-01
5.81192076e-01 -1.04597831e+00 -7.03197241e-01 4.09900576e-01
7.55722672e-02 2.11231634e-01 9.01005790e-02 9.28308845e-01
-1.42207310e-01 1.31941780e-01 2.56859630e-01 -1.22343695e+00
-1.31846213e+00 5.54878831e-01 4.91214573e-01 -1.65250003e-01
-6.48236334e-01 6.07761025e-01 8.23531002e-02 -1.38724610e-01
-7.71214589e-02 -3.10943693e-01 -1.80063084e-01 -1.05323829e-01
1.71014532e-01 5.47635198e-01 1.70597062e-01 -1.19847882e+00
-3.74362141e-01 8.89796734e-01 9.50961262e-02 -4.91775602e-01
1.32318902e+00 -5.73367000e-01 5.32840565e-02 6.26642168e-01
9.99056458e-01 1.25751331e-01 -1.73659885e+00 -3.88548225e-01
-1.31964656e-02 -5.42466700e-01 2.23357111e-01 -5.96555531e-01
-8.92490685e-01 3.31555843e-01 1.01872098e+00 1.82468876e-01
6.38129056e-01 3.38881761e-01 2.10904568e-01 3.43701929e-01
5.18047452e-01 -7.94295967e-01 1.43260524e-01 3.05828005e-01
6.86770976e-01 -1.20142925e+00 2.22327873e-01 -4.09973770e-01
-2.48653576e-01 7.91600585e-01 4.95815307e-01 -4.23877448e-01
7.93185949e-01 1.36592552e-01 -1.15914442e-01 -2.90175140e-01
-7.43570745e-01 -2.44367614e-01 3.51484865e-01 3.03271800e-01
-3.44443917e-01 1.08446248e-01 4.65726294e-02 -3.38150421e-04
2.24546880e-01 -1.73528567e-01 3.55277866e-01 1.08304930e+00
-5.60076892e-01 -8.29853892e-01 -6.93271160e-01 3.94975916e-02
-2.33661346e-02 5.16030371e-01 -1.79518431e-01 5.96284866e-01
2.70446897e-01 7.17964292e-01 3.18628788e-01 -4.86755855e-02
3.79172921e-01 -1.43386513e-01 7.96627402e-01 -7.60250807e-01
-1.87062681e-01 4.91496861e-01 -3.24542791e-01 -4.13434863e-01
-6.32647038e-01 -1.17849123e+00 -9.96536732e-01 5.77494860e-01
-8.90952468e-01 2.05591589e-01 9.28088725e-01 9.72055256e-01
3.95647347e-01 9.75390598e-02 8.64352465e-01 -1.32271326e+00
-6.44658566e-01 -6.52380884e-01 -5.66933513e-01 3.73952448e-01
6.31433845e-01 -9.68939483e-01 -6.17828190e-01 -1.60895486e-03] | [7.023499488830566, -2.2542011737823486] |
045942b8-fea3-4c81-826d-e26f5023c322 | adaptive-fusion-for-rgb-d-salient-object | 1901.01369 | null | http://arxiv.org/abs/1901.01369v2 | http://arxiv.org/pdf/1901.01369v2.pdf | Adaptive Fusion for RGB-D Salient Object Detection | RGB-D salient object detection aims to identify the most visually distinctive
objects in a pair of color and depth images. Based upon an observation that
most of the salient objects may stand out at least in one modality, this paper
proposes an adaptive fusion scheme to fuse saliency predictions generated from
two modalities. Specifically, we design a two-streamed convolutional neural
network (CNN), each of which extracts features and predicts a saliency map from
either RGB or depth modality. Then, a saliency fusion module learns a switch
map that is used to adaptively fuse the predicted saliency maps. A loss
function composed of saliency supervision, switch map supervision, and
edge-preserving constraints is designed to make full supervision, and the
entire network is trained in an end-to-end manner. Benefited from the adaptive
fusion strategy and the edge-preserving constraint, our approach outperforms
state-of-the-art methods on three publicly available datasets. | ['Ningning Wang', 'Xiaojin Gong'] | 2019-01-05 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 5.84715843e-01 9.52980220e-02 -1.30497038e-01 -5.48137009e-01
-6.12212777e-01 -2.64800042e-02 4.10672486e-01 1.10864937e-01
-3.12921286e-01 4.27840739e-01 2.06601337e-01 4.35489565e-02
1.11127250e-01 -4.61994410e-01 -8.20601404e-01 -5.98798931e-01
8.12506899e-02 -3.02982718e-01 8.21617723e-01 -1.27529338e-01
3.92626852e-01 3.02131563e-01 -1.98361576e+00 2.65591532e-01
1.03633785e+00 1.56370771e+00 7.53491461e-01 3.17322969e-01
6.64781546e-03 9.21549916e-01 1.15927413e-01 -1.08220078e-01
4.03011829e-01 -2.75864989e-01 -7.43412375e-01 2.49524027e-01
5.73854506e-01 -4.13760811e-01 -3.54967952e-01 1.10848927e+00
4.56821680e-01 1.34352192e-01 1.56330496e-01 -1.41465425e+00
-7.07077801e-01 4.15851057e-01 -7.99560070e-01 4.21758413e-01
3.22279900e-01 2.13294819e-01 1.01898658e+00 -1.15320027e+00
3.44306737e-01 1.08379042e+00 2.52771169e-01 4.03636187e-01
-1.05625927e+00 -4.85489815e-01 3.83928955e-01 1.43498152e-01
-1.18744564e+00 -3.62519026e-01 1.30356753e+00 3.57057899e-02
5.68434477e-01 4.67011370e-02 8.08061421e-01 8.19987476e-01
3.29760350e-02 1.45657218e+00 1.04958045e+00 -1.92063063e-01
3.07096869e-01 1.29012957e-01 -2.65209407e-01 7.69810379e-01
-7.52404612e-03 2.78201308e-02 -1.00872123e+00 1.07865885e-01
7.85142541e-01 4.68157351e-01 -2.25860924e-01 -8.97830307e-01
-1.30848134e+00 4.76085752e-01 1.07927096e+00 1.98828373e-02
-5.59580028e-01 -2.57191844e-02 1.00370094e-01 -1.80973545e-01
4.06805009e-01 1.38390288e-01 -3.67079198e-01 3.06405336e-01
-9.97096181e-01 1.94350049e-01 5.38239330e-02 9.44999278e-01
1.17442691e+00 -9.54314321e-02 -2.88241416e-01 5.00309944e-01
5.30589819e-01 5.95786870e-01 4.49809045e-01 -7.36658931e-01
3.38988066e-01 9.99011993e-01 1.50019065e-01 -9.62042511e-01
-3.86633039e-01 -3.84072840e-01 -5.97201288e-01 4.29184198e-01
-5.74281327e-02 1.63978189e-02 -1.17612720e+00 1.74416995e+00
4.94287521e-01 4.86325651e-01 8.33070278e-02 1.48767257e+00
1.00295901e+00 2.27679566e-01 1.38289496e-01 2.22949237e-01
1.07574892e+00 -1.14686489e+00 -3.96243304e-01 -4.70348358e-01
5.94032742e-02 -6.05414569e-01 1.05862355e+00 -1.64360881e-01
-1.44084537e+00 -6.85585916e-01 -1.19595242e+00 -4.53200221e-01
-3.51900756e-01 1.33009434e-01 6.19885325e-01 7.85345286e-02
-1.10636151e+00 3.41056764e-01 -8.61382127e-01 -1.57630160e-01
6.12308502e-01 3.22930783e-01 -1.51612744e-01 1.05850965e-01
-1.15831220e+00 7.21940815e-01 4.92191702e-01 3.50160040e-02
-9.58524287e-01 -5.89592218e-01 -1.18325090e+00 1.90795958e-01
2.30418846e-01 -7.80060947e-01 1.09824216e+00 -1.39597750e+00
-1.35784197e+00 9.89605010e-01 -2.96284109e-01 -4.37773407e-01
3.95512283e-01 -7.89803118e-02 -2.76588529e-01 4.93183136e-01
4.47637051e-01 1.05110919e+00 1.12103045e+00 -1.37252259e+00
-1.26841068e+00 -4.04037714e-01 5.76215610e-02 5.30170262e-01
-2.55869985e-01 1.95382014e-02 -5.59363902e-01 -5.49777508e-01
4.26172346e-01 -6.09622061e-01 -2.53267288e-01 1.18749410e-01
-6.07047379e-01 3.67159173e-02 9.51249003e-01 -4.05803025e-01
8.75083029e-01 -2.21242285e+00 4.01970237e-01 7.08966479e-02
4.10410255e-01 -7.25685582e-02 -6.59309477e-02 -2.27937743e-01
-1.24807730e-01 -3.24641705e-01 -4.83675838e-01 -8.11228037e-01
-1.60968497e-01 -2.94184774e-01 -4.22810584e-01 3.11809927e-01
6.77823305e-01 9.74482179e-01 -1.26022243e+00 -5.87044418e-01
4.48805034e-01 5.28166831e-01 -1.73880637e-01 4.03532863e-01
-1.28609031e-01 3.67593318e-01 -6.07595980e-01 1.00028086e+00
8.37829411e-01 -5.00415146e-01 -2.82388717e-01 -1.23561300e-01
-2.75471717e-01 3.20469409e-01 -1.01057279e+00 2.21294498e+00
-9.45443586e-02 4.52024788e-01 -1.51213199e-01 -6.47437215e-01
9.85191941e-01 -9.18328241e-02 3.48740667e-01 -8.70910466e-01
1.97540954e-01 3.51205260e-01 -4.40956861e-01 -6.95762038e-02
6.07417047e-01 1.51484460e-01 -6.52417168e-02 3.33982021e-01
1.14871167e-01 -5.43030761e-02 -1.84685424e-01 2.74629384e-01
8.13191593e-01 1.52509421e-01 -2.55262833e-02 -7.49586746e-02
5.67626178e-01 -1.33850187e-01 6.84548318e-01 5.49401999e-01
-5.31620145e-01 9.72536981e-01 2.09180951e-01 -4.13621128e-01
-8.25985968e-01 -1.36079955e+00 8.57472420e-02 1.30403316e+00
1.03918827e+00 3.06221768e-02 -4.23479497e-01 -8.01557958e-01
1.59016967e-01 3.80870730e-01 -9.73947763e-01 -4.41687882e-01
-2.85639107e-01 -1.15666345e-01 -1.18165024e-01 6.54379487e-01
7.88179636e-01 -1.17903173e+00 -1.16070139e+00 -6.72664866e-02
-3.41713056e-02 -9.20630276e-01 -6.94910944e-01 3.42725724e-01
-8.10791850e-01 -8.46718669e-01 -8.45301926e-01 -1.13282740e+00
6.88492179e-01 1.06144392e+00 9.15733933e-01 1.20558895e-01
-4.57994118e-02 6.78038076e-02 -3.42616737e-01 -5.41417003e-01
1.81027874e-01 1.84509024e-01 -1.68072045e-01 3.14323485e-01
3.74883145e-01 -5.78443468e-01 -9.69899714e-01 1.77020594e-01
-9.70300078e-01 6.50574684e-01 7.83183992e-01 5.52201331e-01
7.17395961e-01 -4.05306041e-01 5.32566488e-01 -2.41132408e-01
7.38520622e-02 -5.39741457e-01 -5.86032748e-01 3.35876584e-01
-3.80293965e-01 6.92471191e-02 2.86271811e-01 -6.80958182e-02
-1.00759089e+00 6.11836076e-01 3.47123802e-01 -9.07082617e-01
-1.76101953e-01 2.71235019e-01 -1.81498647e-01 -1.33778006e-01
2.11218506e-01 4.65984374e-01 4.89196219e-02 -4.17178780e-01
3.80623788e-01 5.34288883e-01 8.12896073e-01 2.55794208e-02
8.55582714e-01 6.86759114e-01 -2.96554446e-01 -2.95523852e-01
-1.22399652e+00 -4.76964951e-01 -6.67808712e-01 -2.77153701e-01
9.32612538e-01 -1.24210072e+00 -4.00493652e-01 6.71354473e-01
-9.31619048e-01 -1.03515856e-01 -2.12081239e-01 2.92980880e-01
-6.00595295e-01 2.78060902e-02 -2.25998044e-01 -6.67162538e-01
-3.06853473e-01 -1.15374053e+00 1.56186211e+00 9.61276650e-01
3.30768734e-01 -6.45026982e-01 -2.03687757e-01 1.23109773e-01
4.21656430e-01 3.64762962e-01 4.24730301e-01 -2.87986636e-01
-9.01815176e-01 3.92200910e-02 -6.96071446e-01 8.33192617e-02
4.42640603e-01 -2.12746993e-01 -1.06350172e+00 -2.03285471e-01
-1.86631963e-01 -4.56595719e-01 1.18695903e+00 4.30344880e-01
1.21839643e+00 8.33654627e-02 -4.24781382e-01 7.24948049e-01
1.42779911e+00 -4.58630286e-02 3.47203583e-01 4.35530007e-01
8.51315618e-01 3.87005478e-01 6.45053506e-01 4.18915778e-01
7.44988322e-01 4.58385050e-01 8.86432827e-01 -5.46025932e-01
-1.11778162e-01 -4.56391037e-01 1.67769328e-01 1.41521230e-01
3.49492490e-01 3.89968455e-01 -7.18336999e-01 8.85055900e-01
-1.98843074e+00 -8.72570515e-01 3.86723965e-01 2.23942184e+00
1.05781841e+00 2.95958132e-01 4.34026301e-01 -6.75681159e-02
9.49753940e-01 3.54513407e-01 -1.00972354e+00 1.14842989e-01
-5.36708891e-01 -9.58446637e-02 4.13828820e-01 2.14584827e-01
-1.39136958e+00 9.48293209e-01 5.66241741e+00 3.89721483e-01
-1.61907291e+00 -1.22118704e-01 7.32740760e-01 -2.76909173e-01
-5.19956172e-01 2.59819943e-02 -4.99236822e-01 6.62106633e-01
2.53518850e-01 -1.94243118e-01 1.67023093e-01 8.96360159e-01
4.24539894e-02 -3.43604743e-01 -8.86590898e-01 8.04367721e-01
2.18454927e-01 -1.32621551e+00 -1.13500096e-01 -3.88684928e-01
9.57798660e-01 3.07663530e-01 4.59807366e-01 -1.44258067e-01
2.64737099e-01 -7.44195163e-01 1.17731690e+00 6.27391756e-01
4.39077348e-01 -7.26508439e-01 6.21276259e-01 1.04965769e-01
-1.27833748e+00 -2.90401787e-01 -1.96087241e-01 4.24866565e-03
2.21512154e-01 5.12803137e-01 -5.72896361e-01 5.03283918e-01
9.64279234e-01 1.14743888e+00 -8.19563091e-01 1.33260429e+00
-3.61947626e-01 1.21929601e-01 -1.80999085e-01 3.20181362e-02
3.98583114e-01 2.00999513e-01 5.37838101e-01 8.93961132e-01
1.31121904e-01 1.10789444e-02 2.51942247e-01 1.05753541e+00
4.20259777e-03 -2.90150762e-01 -3.28456253e-01 4.25759971e-01
6.02254331e-01 1.38898504e+00 -7.25068808e-01 -2.58036762e-01
-5.81558287e-01 1.17861593e+00 4.13476437e-01 2.57209063e-01
-8.47030997e-01 -4.87108856e-01 7.05475986e-01 -7.53811821e-02
7.83597946e-01 1.77880019e-01 -6.52290404e-01 -1.16306126e+00
1.23350285e-01 -2.49134600e-01 2.56442457e-01 -1.22931361e+00
-1.14039993e+00 5.17519414e-01 -2.44552135e-01 -1.53138888e+00
5.15963398e-02 -3.12881678e-01 -8.18486691e-01 1.07379222e+00
-2.21285176e+00 -1.36367035e+00 -6.41105533e-01 6.68616116e-01
3.52385908e-01 4.16314602e-02 2.55822957e-01 -1.05297796e-01
-6.47773743e-01 3.44900280e-01 -1.87391505e-01 6.09468073e-02
5.57736397e-01 -1.39824080e+00 3.24900299e-01 1.03993380e+00
-1.99882239e-01 3.09342414e-01 5.75665534e-01 -5.24389446e-01
-1.19394243e+00 -1.23583782e+00 7.23133445e-01 -2.92733163e-01
4.47411984e-01 -2.45348647e-01 -9.19283390e-01 4.43036586e-01
2.70126998e-01 4.40050125e-01 2.52771586e-01 -3.03026140e-01
-3.61705780e-01 -2.11551622e-01 -1.01421237e+00 5.57873905e-01
9.61048961e-01 -6.34330094e-01 -7.86627293e-01 -1.33983754e-02
1.09883404e+00 -7.61017025e-01 -4.01772112e-01 3.97999644e-01
4.10905749e-01 -1.16057539e+00 1.12705624e+00 -3.90165508e-01
7.88760126e-01 -7.05673695e-01 -9.37602073e-02 -1.13815415e+00
-3.36500078e-01 -3.21058512e-01 -1.78221509e-01 9.03202832e-01
3.62647444e-01 -2.44852319e-01 7.92800426e-01 7.43602395e-01
-3.83334875e-01 -8.66186380e-01 -9.36879516e-01 -3.06152374e-01
-5.08537769e-01 -2.61028893e-02 7.50896454e-01 6.83883548e-01
-1.28876241e-02 2.77333707e-01 -2.93593794e-01 3.47298861e-01
7.91883886e-01 7.31273830e-01 4.39243853e-01 -1.18112111e+00
1.65445074e-01 -6.02507651e-01 -4.54620451e-01 -1.03117633e+00
1.54575363e-01 -6.88132346e-01 3.64350706e-01 -1.29115033e+00
3.85993034e-01 -3.49640459e-01 -1.03062439e+00 8.16193938e-01
-6.81257069e-01 3.61758262e-01 3.61541241e-01 1.27909079e-01
-1.07237017e+00 8.98714066e-01 1.24014294e+00 -2.35826984e-01
-4.17546839e-01 -1.58045143e-01 -1.02232563e+00 6.92550659e-01
6.64153755e-01 -2.00672254e-01 -4.41543579e-01 -5.64988613e-01
-1.08994171e-01 -7.31984302e-02 7.80878246e-01 -1.06964123e+00
4.73380089e-01 -3.21828872e-01 7.31813192e-01 -8.76579940e-01
3.14836591e-01 -7.80113876e-01 -4.56358075e-01 1.89657584e-01
-5.12630880e-01 -1.62755549e-01 9.88987237e-02 4.92949098e-01
-3.02369416e-01 3.43518376e-01 8.97238672e-01 1.28491789e-01
-1.12585151e+00 5.06420791e-01 2.51621693e-01 -8.01733807e-02
1.09828568e+00 -4.23383892e-01 -2.21576363e-01 -9.31992605e-02
-3.37612897e-01 5.19180715e-01 7.83349931e-01 7.32273519e-01
1.00680196e+00 -1.57994974e+00 -5.27257979e-01 4.55661893e-01
4.34402972e-01 3.06012034e-01 4.04501081e-01 8.89990091e-01
2.80225533e-03 2.25331932e-01 -6.26202047e-01 -8.10965836e-01
-9.38995540e-01 6.41156316e-01 2.67797858e-01 1.60186499e-01
-3.77925634e-01 1.18190384e+00 2.31017813e-01 -8.93195197e-02
2.74058789e-01 -4.47640210e-01 -2.12017477e-01 -1.79909870e-01
7.18737423e-01 -3.62369344e-02 -7.85846859e-02 -8.28234851e-01
-5.66931665e-01 5.01223505e-01 -2.29572624e-01 2.46298313e-02
1.36549962e+00 -4.41388458e-01 2.96897870e-02 4.32305425e-01
1.18696845e+00 -3.31725448e-01 -1.93436778e+00 -7.35308528e-01
-2.77072281e-01 -7.85476029e-01 3.74034703e-01 -7.76999831e-01
-1.31247556e+00 6.49148047e-01 7.25027025e-01 5.61552308e-02
1.72059381e+00 1.65534303e-01 9.52897251e-01 -1.21523619e-01
2.97228545e-01 -1.00980663e+00 2.29058146e-01 3.31369907e-01
7.95820415e-01 -1.74033749e+00 -2.36087918e-01 -3.01427394e-01
-8.92912805e-01 6.54045522e-01 8.12125802e-01 -2.78632730e-01
5.72712421e-01 -7.23462179e-02 -1.17614210e-01 -1.21509120e-01
-6.95649743e-01 -6.53543591e-01 4.92014349e-01 4.65638131e-01
-1.79338269e-02 -6.62399456e-02 1.73695087e-01 6.34329677e-01
1.01706557e-01 8.76159444e-02 2.29900196e-01 1.10192037e+00
-7.76813269e-01 -5.04199386e-01 -1.75655112e-01 4.26494837e-01
-1.47551559e-02 -1.38818830e-01 -4.19965982e-01 2.63657659e-01
1.67976737e-01 8.00449550e-01 2.58531272e-01 -5.47081232e-01
2.59152710e-01 -3.24102521e-01 8.62850174e-02 -5.08524537e-01
-4.55810308e-01 -1.36408821e-01 -5.59609056e-01 -7.64724910e-01
-7.35500276e-01 -8.16014230e-01 -1.47465479e+00 2.37734709e-02
-2.58427531e-01 -1.44597515e-01 4.23349172e-01 9.57353234e-01
6.94420993e-01 4.27799642e-01 1.04084194e+00 -1.28076625e+00
-1.58539012e-01 -6.66059852e-01 -6.89405143e-01 5.23210764e-01
8.73041928e-01 -9.20165777e-01 -2.67530322e-01 5.86715378e-02] | [9.760245323181152, -0.6138543486595154] |
535fff05-9c47-499e-b41c-84880b44eaf2 | merging-classification-predictions-with | 2210.00834 | null | https://arxiv.org/abs/2210.00834v1 | https://arxiv.org/pdf/2210.00834v1.pdf | Merging Classification Predictions with Sequential Information for Lightweight Visual Place Recognition in Changing Environments | Low-overhead visual place recognition (VPR) is a highly active research topic. Mobile robotics applications often operate under low-end hardware, and even more hardware capable systems can still benefit from freeing up onboard system resources for other navigation tasks. This work addresses lightweight VPR by proposing a novel system based on the combination of binary-weighted classifier networks with a one-dimensional convolutional network, dubbed merger. Recent work in fusing multiple VPR techniques has mainly focused on increasing VPR performance, with computational efficiency not being highly prioritized. In contrast, we design our technique prioritizing low inference times, taking inspiration from the machine learning literature where the efficient combination of classifiers is a heavily researched topic. Our experiments show that the merger achieves inference times as low as 1 millisecond, being significantly faster than other well-established lightweight VPR techniques, while achieving comparable or superior VPR performance on several visual changes such as seasonal variations and viewpoint lateral shifts. | ['Shoaib Ehsan', 'Klaus D. McDonald-Maier', 'Michael Milford', 'Bruno Ferrarini', 'Bruno Arcanjo'] | 2022-10-03 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 1.20232612e-01 -4.04444225e-02 -2.99466670e-01 -3.47437441e-01
-3.96268576e-01 -5.06999671e-01 6.71766043e-01 1.36790663e-01
-1.02408516e+00 4.38094854e-01 -2.43798167e-01 -6.33613408e-01
-1.19252943e-01 -7.82411814e-01 -8.20221663e-01 -4.84636605e-01
-2.84999460e-01 1.81876585e-01 7.25058496e-01 -3.01000029e-01
5.92527151e-01 6.92404509e-01 -2.11687922e+00 -2.12434590e-01
5.48317432e-01 1.13174260e+00 3.89570951e-01 8.64804506e-01
1.33604303e-01 4.90309834e-01 -4.57764059e-01 -4.03774977e-02
5.02971113e-01 5.47515094e-01 -4.00904238e-01 -4.66490835e-01
6.12546206e-01 -1.21006727e-01 -2.68407464e-01 7.29599953e-01
5.89156628e-01 2.87129670e-01 4.18863744e-01 -1.53258705e+00
-3.34295452e-01 2.24185467e-01 -4.91022557e-01 1.73007727e-01
2.83706516e-01 5.06176315e-02 9.31371927e-01 -7.45848060e-01
5.14187396e-01 9.47998345e-01 1.03459275e+00 3.91506314e-01
-1.00689852e+00 -5.29512465e-01 1.91622391e-01 4.22941864e-01
-1.53538311e+00 -5.02283573e-01 5.98509312e-01 -2.00609148e-01
1.58524954e+00 2.35365883e-01 6.20768607e-01 8.45133185e-01
2.96449095e-01 7.59860039e-01 7.73255646e-01 -2.91119933e-01
5.58775246e-01 -1.41731262e-01 1.30457684e-01 8.44796419e-01
6.27987027e-01 -5.44781275e-02 -5.86583734e-01 8.98132324e-02
6.94071949e-01 1.11637339e-01 -7.19465464e-02 -6.25256479e-01
-1.32352960e+00 5.77618420e-01 6.90202832e-01 1.51232034e-01
-2.37925380e-01 5.89279056e-01 3.39423269e-01 3.38617004e-02
1.15719311e-01 3.23662937e-01 -5.87672591e-01 -5.24393916e-01
-9.14441824e-01 2.77708828e-01 7.30957210e-01 1.15265203e+00
8.51682365e-01 -6.62713498e-02 1.82904437e-01 6.98325813e-01
4.69271868e-01 6.04115188e-01 6.67695582e-01 -9.50759828e-01
3.11997831e-01 5.35714149e-01 1.70460299e-01 -1.04581213e+00
-7.45389581e-01 2.17438899e-02 -5.39395928e-01 6.13360167e-01
3.24901611e-01 -8.30180496e-02 -1.11441302e+00 1.46335721e+00
2.03749791e-01 7.73247629e-02 1.22133315e-01 8.95500422e-01
7.66466677e-01 4.11242902e-01 -9.87282172e-02 3.37276816e-01
1.39323115e+00 -1.12410021e+00 -3.75131994e-01 -5.56071877e-01
5.05742669e-01 -5.88883340e-01 8.93226445e-01 4.03741658e-01
-6.51725650e-01 -2.86710858e-01 -1.68605149e+00 -5.42018831e-01
-7.82899022e-01 1.09394476e-01 9.36643720e-01 6.58110559e-01
-1.28406036e+00 4.85976070e-01 -9.91799057e-01 -6.62077129e-01
3.44966799e-01 6.53994083e-01 -5.40655851e-01 4.21904735e-02
-7.95944214e-01 1.28483224e+00 9.78893191e-02 1.16962299e-01
-3.66550803e-01 -4.67509598e-01 -1.09053671e+00 1.44334827e-02
3.75581145e-01 -2.87928641e-01 1.51487350e+00 -3.26240510e-01
-1.75792670e+00 5.54740489e-01 -2.65152067e-01 -8.50990355e-01
5.77094734e-01 -1.71876460e-01 -1.71113864e-01 -1.98146775e-01
-1.58678651e-01 8.92997026e-01 6.17645442e-01 -8.41794610e-01
-8.84242117e-01 -2.57354140e-01 2.68038303e-01 2.37283230e-01
-2.73521423e-01 -4.03880209e-01 -4.70829993e-01 -1.81649446e-01
3.12208295e-01 -1.07663536e+00 -3.92193615e-01 4.78988051e-01
2.06180736e-01 -1.53596342e-01 1.03740931e+00 -1.63483962e-01
9.27623928e-01 -2.09220123e+00 -1.69257522e-01 1.50604814e-01
2.00680450e-01 3.44838768e-01 7.37273842e-02 7.11104309e-04
2.93457925e-01 -1.62605301e-01 -5.11124209e-02 -2.69367903e-01
6.69696331e-02 5.53917348e-01 -2.49597058e-01 7.68898726e-01
5.29688895e-02 6.77373528e-01 -9.80428159e-01 -3.86732072e-01
5.23704529e-01 4.94025946e-01 -5.10643721e-01 -3.01932752e-01
1.60116460e-02 -2.73998797e-01 -2.15179622e-02 8.09417248e-01
5.72746515e-01 -1.69818804e-01 1.63414657e-01 -7.19472095e-02
-5.38123131e-01 1.59699798e-01 -1.23968816e+00 1.67463291e+00
-6.96030617e-01 1.18238652e+00 -3.86779495e-02 -8.71601045e-01
1.11364329e+00 -1.79991856e-01 2.40997851e-01 -9.46461856e-01
2.85470188e-01 3.39230329e-01 -1.19382583e-01 -2.89396763e-01
1.17643452e+00 3.30273658e-01 -1.23918809e-01 -6.40964434e-02
1.59269199e-02 1.26574963e-01 2.21355353e-02 -1.71793461e-01
1.21278954e+00 4.31261957e-01 5.08489847e-01 -2.07802251e-01
2.93648839e-01 1.53921261e-01 5.07560134e-01 7.92582870e-01
-6.83398545e-01 4.79659885e-01 7.41938055e-02 -5.22169948e-01
-9.01666760e-01 -9.04177725e-01 -2.46293634e-01 1.01652610e+00
7.49021947e-01 -4.64452922e-01 -4.73727226e-01 -3.33605230e-01
2.70450234e-01 3.58033180e-01 -5.49820244e-01 2.30326310e-01
-5.53989828e-01 -6.00080013e-01 6.95049644e-01 8.55805993e-01
7.83121109e-01 -7.61117220e-01 -1.60208666e+00 1.45060286e-01
1.45477289e-02 -1.15690672e+00 1.12416275e-01 5.49069226e-01
-8.01927865e-01 -8.56602132e-01 -7.07460165e-01 -7.00192094e-01
5.13729393e-01 8.75742555e-01 7.33345449e-01 1.46887228e-02
-3.78325194e-01 2.66290218e-01 -3.84225368e-01 -7.30490685e-01
4.67942983e-01 3.00923288e-01 4.03056443e-01 -3.39953065e-01
5.55142224e-01 -5.36432147e-01 -5.05692303e-01 2.75487423e-01
-4.82608259e-01 8.07979256e-02 7.30177581e-01 7.13748276e-01
4.87785578e-01 -3.76947701e-01 9.52679366e-02 -4.43836868e-01
1.58555552e-01 -3.94646883e-01 -9.69500899e-01 1.48161590e-01
-8.98017645e-01 3.54885548e-01 3.22491229e-01 -6.72986150e-01
-8.64364445e-01 1.77674055e-01 -1.37447342e-01 -2.60092258e-01
-1.23118423e-01 2.65378892e-01 2.04819351e-01 -6.57352984e-01
8.09010446e-01 1.28774017e-01 1.23888440e-03 -5.46371415e-02
4.65455115e-01 8.60252142e-01 5.50919116e-01 -3.68782818e-01
7.28250742e-01 7.16317773e-01 1.13661312e-01 -1.09329510e+00
-1.78886652e-01 -6.73716068e-01 -4.05703545e-01 -3.96546394e-01
6.38024032e-01 -8.55923235e-01 -9.87340093e-01 3.13291579e-01
-9.21986520e-01 -5.27887404e-01 1.41162917e-01 5.56605756e-01
-6.90860868e-01 1.88917264e-01 -1.94611415e-01 -8.37624252e-01
-1.55539915e-01 -1.18328643e+00 1.19214821e+00 3.66356730e-01
-1.09762594e-01 -6.33882403e-01 9.51494947e-02 5.91108203e-02
6.74169660e-01 1.18193692e-02 2.92078674e-01 -5.18569171e-01
-7.32405841e-01 -4.99074340e-01 -4.11914319e-01 -3.92197549e-01
-2.77281582e-01 5.36137708e-02 -1.18811047e+00 -1.55793846e-01
-4.95918810e-01 3.04580182e-02 8.29334855e-01 3.23014647e-01
9.43656921e-01 1.47710070e-01 -6.41520858e-01 6.79674208e-01
1.54614067e+00 2.48000070e-01 7.08650887e-01 8.97575438e-01
5.04352093e-01 3.45856816e-01 7.61637270e-01 2.62151957e-01
6.17217541e-01 9.54662025e-01 6.12375081e-01 -1.35480966e-02
-1.08645231e-01 4.80456203e-02 2.95230448e-01 4.91081476e-01
-3.78964454e-01 -8.12137965e-03 -1.24613547e+00 6.08771682e-01
-2.26180506e+00 -7.80016482e-01 -6.81691663e-03 2.24616909e+00
3.30745727e-01 2.34628230e-01 -1.82569563e-01 2.05198124e-01
5.40642202e-01 3.02785546e-01 -5.27662039e-01 -5.78168571e-01
1.63977131e-01 7.68920481e-02 1.29760635e+00 3.82481337e-01
-1.18823147e+00 8.70604515e-01 6.62894773e+00 5.23522973e-01
-1.41675687e+00 2.68931761e-02 -1.25242785e-01 -1.77516922e-01
2.08385095e-01 -5.79244494e-02 -9.47951198e-01 3.12701285e-01
1.10489249e+00 -2.80110016e-02 3.32866162e-01 1.60249388e+00
-2.00570688e-01 -4.44454283e-01 -9.15550470e-01 1.32830811e+00
2.55858779e-01 -1.46678865e+00 -3.82376313e-01 2.55121022e-01
3.20410579e-01 4.63380963e-01 3.24621014e-02 4.57538962e-01
2.42026344e-01 -7.78462052e-01 9.44205821e-01 2.48574883e-01
6.72127724e-01 -6.59017742e-01 8.52726519e-01 1.56532556e-01
-1.47031140e+00 -3.07740092e-01 -4.39376175e-01 -4.75921154e-01
-9.35111120e-02 2.79090375e-01 -7.69052863e-01 2.86917329e-01
1.08350539e+00 6.61827505e-01 -6.61855936e-01 1.30865586e+00
-1.26756623e-01 3.63647826e-02 -5.80195725e-01 -6.19899094e-01
2.80385137e-01 2.25646168e-01 3.66539478e-01 1.23541737e+00
3.48651528e-01 -2.64867306e-01 -1.61490858e-01 1.36755511e-01
1.85900152e-01 -1.53500915e-01 -9.80399370e-01 4.13346112e-01
5.74488997e-01 1.37498927e+00 -1.10837770e+00 -2.69273967e-01
-5.66002786e-01 8.24309826e-01 4.02018994e-01 -3.70574445e-02
-1.02845895e+00 -7.41593063e-01 1.04198349e+00 -2.68488407e-01
6.67663276e-01 -8.60694349e-01 -6.00063384e-01 -1.05706120e+00
9.28449333e-02 -2.46191904e-01 -1.49305403e-01 -7.16954470e-01
-6.96479023e-01 3.86710137e-01 -1.44234478e-01 -1.58858728e+00
5.46507388e-02 -1.12201381e+00 -2.01763779e-01 3.88402104e-01
-1.86077893e+00 -8.54642272e-01 -5.45362711e-01 3.89760286e-01
6.35529101e-01 2.32973397e-02 9.43781495e-01 3.89205694e-01
-4.90187168e-01 7.70456851e-01 -4.76506818e-03 -8.12857002e-02
6.65505350e-01 -1.07299101e+00 4.98219043e-01 9.50370491e-01
1.15368657e-01 7.25732684e-01 6.34809136e-01 -4.08043653e-01
-1.86491501e+00 -9.34192598e-01 9.50760365e-01 -3.60332757e-01
3.97428989e-01 -5.25648117e-01 -5.90585709e-01 6.18870676e-01
1.52068753e-02 2.50165015e-01 5.28185308e-01 3.43502253e-01
-5.95242441e-01 -2.12903976e-01 -1.24495816e+00 1.02964497e+00
1.28197765e+00 -3.74365300e-01 -6.63699448e-01 8.99724439e-02
5.34942627e-01 -7.31585085e-01 -3.74364257e-01 3.37444633e-01
1.00151420e+00 -7.42623627e-01 1.03889441e+00 8.12170445e-04
1.90731913e-01 -7.03908384e-01 -5.35733700e-01 -8.34635794e-01
-2.11049855e-01 -2.92205602e-01 -2.85440892e-01 6.58498287e-01
4.44991052e-01 -7.58837461e-01 8.79062772e-01 5.64787924e-01
-1.14773624e-01 -5.66431105e-01 -1.10140789e+00 -1.04226875e+00
-5.15005827e-01 -7.76716411e-01 3.36175472e-01 6.76095128e-01
1.48233861e-01 1.93963900e-01 -1.95257992e-01 5.00910938e-01
5.03334343e-01 -2.10336708e-02 9.24814105e-01 -1.33315003e+00
2.18296424e-03 -5.89612365e-01 -1.09976017e+00 -1.00514603e+00
-3.58824730e-02 -6.39675200e-01 6.94814563e-01 -1.63221943e+00
-2.40987420e-01 -4.03946161e-01 -1.80236086e-01 7.27210462e-01
3.37067395e-01 4.50776517e-01 1.29297495e-01 1.05806746e-01
-7.18847990e-01 2.57881492e-01 4.95249480e-01 -1.61146924e-01
-1.23880208e-01 -3.47676396e-01 -4.20112848e-01 8.48358214e-01
7.26295292e-01 -2.39669293e-01 -3.43008310e-01 -6.00435972e-01
2.82102674e-01 -4.24585253e-01 2.17394561e-01 -1.63483667e+00
7.74571300e-01 -1.44152194e-01 3.35027993e-01 -5.21113276e-01
4.65717942e-01 -1.09807658e+00 -5.15470691e-02 5.13539016e-01
-4.76286784e-02 2.09370658e-01 4.09940988e-01 6.52739227e-01
1.74721126e-02 4.31671506e-03 5.80475271e-01 1.23319430e-02
-1.48498249e+00 2.81439777e-02 -7.32674956e-01 -5.12758493e-01
1.28008699e+00 -7.07369685e-01 -4.55384284e-01 -6.90065175e-02
-2.67767012e-01 2.13274002e-01 4.51615006e-01 6.45556509e-01
7.03222752e-01 -1.14813232e+00 2.21996784e-01 2.23050505e-01
3.95245045e-01 5.97265102e-02 5.21828644e-02 6.95717037e-01
-9.45161343e-01 6.78558350e-01 -3.97447407e-01 -7.69880295e-01
-1.34114027e+00 4.69554156e-01 4.53229137e-02 1.86169699e-01
-8.09127390e-01 8.00466776e-01 -4.77346212e-01 -5.96284986e-01
5.07585883e-01 -5.04944801e-01 -1.89691007e-01 1.36984348e-01
5.61099708e-01 5.72472274e-01 3.16148460e-01 -4.80659693e-01
-7.25647628e-01 6.60898030e-01 9.51194540e-02 -2.16202840e-01
1.29853630e+00 -1.12941429e-01 3.21457773e-01 4.53740865e-01
9.84938145e-01 -1.97047800e-01 -1.41505730e+00 -1.09102823e-01
1.62962839e-01 -3.74163955e-01 2.96360880e-01 -5.18115222e-01
-7.08410680e-01 8.27879667e-01 9.03458118e-01 -1.08652666e-01
6.93393826e-01 -3.60399038e-01 6.83783293e-01 1.01348996e+00
1.13460302e+00 -1.23074436e+00 -1.76357269e-01 9.31150615e-01
5.28572679e-01 -1.32454181e+00 9.60099995e-02 -2.07340047e-01
-2.92517096e-01 1.15172923e+00 7.17578530e-01 -4.42389965e-01
6.21022224e-01 5.30952632e-01 1.48107290e-01 1.42069412e-02
-4.77444142e-01 -4.28312957e-01 -5.97650297e-02 7.61601925e-01
6.80391639e-02 8.46499205e-02 -3.18253487e-01 2.33158708e-01
-2.89490610e-01 9.25966054e-02 5.45251131e-01 1.53934443e+00
-5.76080739e-01 -7.90409207e-01 -2.01941699e-01 2.72243887e-01
-8.59747455e-02 -2.85468018e-03 3.87004875e-02 8.11777472e-01
5.77148274e-02 7.80105412e-01 3.70187283e-01 -7.35081911e-01
3.78270656e-01 -8.80105235e-03 3.68277162e-01 -3.13851297e-01
-3.55876178e-01 -5.12442350e-01 1.32720724e-01 -9.91967797e-01
-5.08302450e-01 -5.80847263e-01 -1.27223575e+00 -4.22949046e-01
-3.00057590e-01 -3.97758693e-01 1.26309931e+00 8.87915552e-01
5.48500299e-01 4.96786714e-01 1.39585942e-01 -1.51555526e+00
-1.95371360e-01 -6.83067739e-01 -2.22018883e-01 -3.10921013e-01
5.00910342e-01 -1.21167469e+00 -2.22417817e-01 -2.59843856e-01] | [7.646856307983398, -1.859878659248352] |
d5ffc57b-9a8b-4d47-8b68-b975e727f745 | transfer-dynamics-in-emergent-evolutionary | 2203.10941 | null | https://arxiv.org/abs/2203.10941v1 | https://arxiv.org/pdf/2203.10941v1.pdf | Transfer Dynamics in Emergent Evolutionary Curricula | PINSKY is a system for open-ended learning through neuroevolution in game-based domains. It builds on the Paired Open-Ended Trailblazer (POET) system, which originally explored learning and environment generation for bipedal walkers, and adapts it to games in the General Video Game AI (GVGAI) system. Previous work showed that by co-evolving levels and neural network policies, levels could be found for which successful policies could not be created via optimization alone. Studied in the realm of Artificial Life as a potentially open-ended alternative to gradient-based fitness, minimal criteria (MC)-based selection helps foster diversity in evolutionary populations. The main question addressed by this paper is how the open-ended learning actually works, focusing in particular on the role of transfer of policies from one evolutionary branch ("species") to another. We analyze the dynamics of the system through creating phylogenetic trees, analyzing evolutionary trajectories of policies, and temporally breaking down transfers according to species type. Furthermore, we analyze the impact of the minimal criterion on generated level diversity and inter-species transfer. The most insightful finding is that inter-species transfer, while rare, is crucial to the system's success. | ['L. B. Soros', 'Julian Togelius', 'Amy K Hoover', 'Aaron Dharna'] | 2022-03-03 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 1.98690742e-02 9.52689871e-02 8.84859264e-02 4.52749580e-01
2.83639848e-01 -6.53618872e-01 5.45738935e-01 5.42949922e-02
-7.39400327e-01 1.38035345e+00 -1.62700787e-01 -4.70307559e-01
-5.79498410e-01 -1.05370235e+00 -7.40430057e-01 -9.93137717e-01
-5.24417400e-01 4.42436159e-01 4.66972798e-01 -1.19394159e+00
5.49437463e-01 2.01469794e-01 -2.08721471e+00 -3.72355133e-01
1.09604609e+00 1.91652507e-01 2.00575575e-01 7.81013012e-01
-1.29892617e-01 5.99699855e-01 -7.81278074e-01 -3.32864404e-01
3.87689948e-01 -9.66009676e-01 -5.92651010e-01 -4.60240036e-01
-4.48149443e-01 2.16430411e-01 1.38215765e-01 9.05985653e-01
6.94277585e-01 1.83580548e-01 6.25956357e-01 -1.27833772e+00
-3.99964929e-01 8.40340793e-01 -3.19988996e-01 1.40462443e-01
9.78306979e-02 5.36275685e-01 7.40876973e-01 -1.02131717e-01
1.01619816e+00 1.20078480e+00 7.60755897e-01 8.50220561e-01
-1.11314142e+00 -4.76817995e-01 -1.14387371e-01 3.83924812e-01
-1.20821023e+00 1.31976992e-01 4.71735746e-01 -4.33091938e-01
7.80408204e-01 5.06675184e-01 1.67140508e+00 1.00669944e+00
5.87508559e-01 3.37238610e-01 1.03565443e+00 -7.25685954e-01
7.92005122e-01 -1.87002376e-01 -4.07748818e-01 5.26473999e-01
6.90215528e-01 7.56288946e-01 -6.02150977e-01 -1.61245968e-02
7.76919425e-01 -8.75161827e-01 -1.01387188e-01 -5.84379971e-01
-7.08524644e-01 9.83184576e-01 5.69207728e-01 5.01671731e-01
-4.69305307e-01 1.45292327e-01 2.51197308e-01 4.80393052e-01
1.18690683e-02 1.09443247e+00 -6.03262722e-01 -5.20809710e-01
-5.49721003e-01 3.22494745e-01 8.56972873e-01 4.56178039e-02
6.85362160e-01 3.88244003e-01 -1.09855654e-02 6.59843981e-01
1.52353495e-01 3.99704427e-02 7.27527261e-01 -9.40489233e-01
-3.51326346e-01 6.27284288e-01 -1.97084486e-01 -7.66242623e-01
-3.32582533e-01 -7.24556267e-01 -2.98381865e-01 8.64379466e-01
5.99715829e-01 -6.10442102e-01 -5.30674577e-01 2.04548168e+00
4.71154362e-01 7.96766765e-03 -5.04729059e-03 5.44325769e-01
2.18428582e-01 6.28447413e-01 1.85302142e-02 -2.84364760e-01
7.84707963e-01 -7.42226064e-01 -4.04555291e-01 -5.06663956e-02
6.53847873e-01 1.21914998e-01 8.90412509e-01 3.51241827e-01
-8.96589816e-01 -1.95372269e-01 -1.09591007e+00 9.09151495e-01
-7.36580968e-01 -7.91856945e-01 5.02572119e-01 9.87450540e-01
-1.39191592e+00 9.91300285e-01 -5.55100679e-01 -7.38101840e-01
7.57529885e-02 4.32928771e-01 6.06801398e-02 7.27743208e-01
-1.48206151e+00 1.24175107e+00 4.98024046e-01 -3.31913114e-01
-7.14966595e-01 -3.98775458e-01 -2.13297725e-01 1.30430281e-01
6.20958030e-01 -8.97694647e-01 7.66771972e-01 -1.41851521e+00
-1.98452735e+00 4.94743764e-01 5.28325319e-01 -5.30546486e-01
5.71429014e-01 5.61464727e-01 8.26752484e-02 -2.79524237e-01
-2.18304247e-02 9.80583310e-01 3.47600251e-01 -1.41976225e+00
-6.86137557e-01 -4.90596503e-01 4.18284833e-02 5.31139910e-01
-3.65604639e-01 -1.18875213e-01 1.24900542e-01 -5.25473535e-01
-4.94759113e-01 -1.05312479e+00 -5.19635081e-01 -2.36716926e-01
2.51613349e-01 -1.82425439e-01 2.24757850e-01 -3.30938339e-01
1.28481138e+00 -1.64296734e+00 6.40368342e-01 2.27764428e-01
-9.15340483e-02 4.68744338e-01 -1.55503139e-01 6.94801867e-01
1.61104292e-01 2.23798856e-01 -3.30135405e-01 5.98641276e-01
-1.32935479e-01 5.67755640e-01 3.88728857e-01 -2.22520754e-02
-7.55487084e-02 7.80352533e-01 -1.12301564e+00 -2.67365098e-01
-5.05431332e-02 2.32907921e-01 -8.14491153e-01 -8.78590494e-02
-3.13537776e-01 6.37762249e-01 -7.96370655e-02 3.37184668e-01
3.25234644e-02 3.58967990e-01 3.28576028e-01 7.46830881e-01
-6.47279322e-01 -3.94234896e-01 -7.87300408e-01 1.25813985e+00
-1.26316100e-01 6.53883815e-01 4.97646704e-02 -9.74904239e-01
1.10592330e+00 -1.92303270e-01 5.18858314e-01 -8.28372598e-01
5.41052520e-01 4.82939422e-01 8.23770106e-01 -3.57811064e-01
4.44322199e-01 -4.05971855e-02 3.09469085e-02 7.91575611e-02
2.80516297e-01 -3.35358411e-01 5.61933696e-01 -3.63889277e-01
1.17387021e+00 4.51901495e-01 5.02256453e-01 -7.24091768e-01
4.21178192e-01 4.11620170e-01 7.36807406e-01 8.86970520e-01
-4.90984082e-01 7.02678412e-02 7.02831805e-01 -2.21603185e-01
-9.54697371e-01 -8.84403467e-01 7.21034855e-02 1.35856628e+00
2.62317419e-01 -1.47627369e-01 -9.96320307e-01 -8.16917494e-02
-1.11978546e-01 9.38039422e-01 -1.01652408e+00 -6.68431342e-01
-5.58941126e-01 -9.91334975e-01 7.27825403e-01 -8.18857774e-02
3.18678856e-01 -1.46194291e+00 -1.37790668e+00 4.24846143e-01
1.96676284e-01 7.49814436e-02 1.78103790e-01 4.73775029e-01
-7.88302362e-01 -1.02133131e+00 -7.87051439e-01 -6.30771279e-01
1.97602481e-01 -3.27851415e-01 1.00779271e+00 2.53131241e-01
-4.11095232e-01 1.93267480e-01 -4.85509068e-01 -6.43782735e-01
-8.02544475e-01 1.36242464e-01 1.42240971e-01 -4.36690778e-01
-1.39943108e-01 -8.67047668e-01 -4.18836653e-01 4.08030540e-01
-7.71692157e-01 -8.72176588e-02 3.01946700e-01 1.02623570e+00
5.84085621e-02 1.41349584e-01 8.34919751e-01 -2.82992512e-01
1.09317780e+00 -5.08535445e-01 -4.29488450e-01 5.05262494e-01
-6.99016511e-01 1.64271697e-01 3.80281985e-01 -6.16786540e-01
-8.90209317e-01 -3.59775633e-01 -2.46996060e-01 1.23880334e-01
1.76034287e-01 3.87539595e-01 4.72709835e-02 -4.44250762e-01
1.11707819e+00 3.48724186e-01 3.63538504e-01 1.67424962e-01
4.20741379e-01 2.93793023e-01 3.04813623e-01 -8.06837618e-01
4.22749311e-01 -5.26190782e-03 7.54253119e-02 -9.45866525e-01
2.66294241e-01 3.32086444e-01 -5.60163200e-01 -9.20898438e-01
8.32559764e-01 -2.76635528e-01 -9.67340529e-01 7.03841329e-01
-6.05755568e-01 -7.73874462e-01 -8.85353744e-01 1.76863760e-01
-8.29182923e-01 -5.68717234e-02 -4.38056849e-02 -1.00815225e+00
-1.16356678e-01 -8.84669781e-01 1.38352975e-01 6.92500353e-01
-1.74601763e-01 -1.02078426e+00 7.09625423e-01 -1.41916066e-01
6.27573431e-01 5.41955173e-01 8.93363535e-01 -4.81987745e-01
3.91499214e-02 4.20826137e-01 6.21906042e-01 -8.76279324e-02
-2.78664585e-02 4.14431185e-01 -1.76426739e-01 -1.91570818e-01
-1.02545366e-01 -1.24129117e-01 4.66354042e-01 4.93026495e-01
3.08586299e-01 -2.17463568e-01 -1.23053871e-01 5.53604484e-01
1.34048474e+00 1.04309309e+00 6.59466565e-01 1.19164944e+00
2.01771766e-01 9.80987787e-01 4.45073038e-01 4.30349648e-01
1.98093906e-01 6.09861672e-01 6.30191028e-01 1.84244558e-01
3.54101025e-02 -1.57726631e-01 3.68522882e-01 6.88118577e-01
-5.07930517e-01 -2.90294409e-01 -1.05796611e+00 3.91366750e-01
-1.96435153e+00 -1.08295739e+00 9.81574208e-02 2.12975669e+00
8.69920194e-01 1.62720561e-01 5.64649880e-01 6.54635876e-02
8.43850970e-01 -4.33742225e-01 -7.87969172e-01 -8.46033752e-01
-4.57196534e-01 2.92024046e-01 4.73646343e-01 4.36101288e-01
-1.77011743e-01 9.91049588e-01 6.65304089e+00 8.70142996e-01
-1.18221664e+00 -2.13858128e-01 4.33300227e-01 -3.38236913e-02
-1.95994303e-01 1.62389606e-01 -3.21684450e-01 6.64413810e-01
9.81418967e-01 -7.01273143e-01 7.92300045e-01 4.66285855e-01
2.74986893e-01 -3.84585112e-01 -4.56174493e-01 3.63979399e-01
-1.42150432e-01 -1.34251249e+00 -7.47192800e-02 2.90099978e-01
8.26707780e-01 -2.92448878e-01 -2.09109411e-01 5.44365287e-01
7.23272622e-01 -7.60708570e-01 9.37909603e-01 4.71257746e-01
3.41185480e-01 -1.13164437e+00 4.46000516e-01 5.59267223e-01
-7.39884257e-01 -5.14012575e-01 -6.61509335e-02 -5.63987970e-01
2.18574852e-02 -3.05180579e-01 -4.39627975e-01 4.06671762e-01
6.90260530e-01 -3.37653682e-02 -8.02478611e-01 1.23363519e+00
-2.01190174e-01 5.37217438e-01 -1.78980798e-01 -7.37282574e-01
3.74512821e-01 -5.84346771e-01 1.00018239e+00 6.47291064e-01
4.75456595e-01 6.36048019e-02 -4.30993676e-01 7.21219838e-01
5.37720025e-01 1.97412506e-01 -8.35451305e-01 6.75895959e-02
4.68453377e-01 7.62275398e-01 -1.07405257e+00 5.63639402e-03
3.86454940e-01 6.39560342e-01 2.09844053e-01 2.26956699e-03
-8.70276868e-01 -4.08412695e-01 7.06572890e-01 3.45454924e-02
3.52470338e-01 -6.72866777e-02 -3.83983552e-01 -7.63235986e-01
-7.99508750e-01 -1.00884509e+00 2.55655497e-01 -5.95662951e-01
-1.00886548e+00 4.68648940e-01 3.18845622e-02 -7.01367021e-01
-4.56797630e-01 -3.66096437e-01 -7.05915809e-01 4.58921820e-01
-6.53813124e-01 -6.59357369e-01 1.83965005e-02 4.37955856e-01
3.27180803e-01 -5.21269321e-01 4.56953168e-01 -2.37631291e-01
-6.53108776e-01 5.05843520e-01 6.91703081e-01 -5.45932949e-01
1.43335804e-01 -1.14944696e+00 1.61007941e-01 6.67107463e-01
-1.18934788e-01 3.78803492e-01 8.76221955e-01 -7.82784760e-01
-1.17040944e+00 -5.40642083e-01 1.74054265e-01 5.96889369e-02
7.47047663e-01 -1.82528168e-01 -5.72454870e-01 -1.27297148e-01
4.02337223e-01 -8.08648586e-01 4.69894588e-01 1.09465804e-03
4.19589937e-01 1.26985878e-01 -1.18632543e+00 1.22252905e+00
1.24076879e+00 2.71280229e-01 -3.73008162e-01 -3.29905868e-01
6.27728224e-01 -6.02242053e-02 -3.71801049e-01 3.60621661e-01
7.52190650e-01 -1.15622497e+00 9.23769176e-01 -7.83664584e-01
4.18856353e-01 -1.48704007e-01 1.09264098e-01 -1.79190540e+00
-5.70498228e-01 -6.27146780e-01 3.94540757e-01 1.03463137e+00
3.43641311e-01 -9.07206774e-01 7.40317404e-01 6.51423037e-02
1.11623353e-03 -9.40398812e-01 -1.11853516e+00 -1.15761423e+00
8.43808472e-01 2.03035846e-01 5.48768818e-01 9.72685993e-01
1.56219155e-01 2.14343518e-01 -3.36535424e-01 -5.01178384e-01
3.60845923e-01 -5.50426722e-01 7.88264573e-01 -1.21796024e+00
-6.32564545e-01 -1.06108689e+00 -5.41966319e-01 1.72356352e-01
-8.88057575e-02 -5.75697064e-01 2.59748429e-01 -1.22510087e+00
-2.40815252e-01 -4.19204354e-01 -5.45430966e-02 1.85213000e-01
-2.36031830e-01 -2.97845397e-02 5.22147238e-01 3.94758135e-02
-1.49859101e-01 6.43725574e-01 1.17507315e+00 2.84512252e-01
-7.34143078e-01 -1.99461371e-01 -8.98581266e-01 6.19605005e-01
1.05731690e+00 -4.64806229e-01 -4.45152730e-01 4.03601490e-02
6.07316732e-01 2.40008235e-01 -2.95293517e-02 -1.44002461e+00
2.34979883e-01 -4.68139291e-01 -1.08956069e-01 1.19125813e-01
7.07528740e-02 -3.85272235e-01 5.58566213e-01 1.29156947e+00
-2.66163081e-01 1.69547141e-01 3.97051275e-01 5.24106622e-01
2.24587187e-01 -5.40534198e-01 8.31579030e-01 -3.11554819e-01
-7.30291307e-01 -2.78203368e-01 -1.09278333e+00 1.57487199e-01
1.45123649e+00 -8.56936812e-01 -2.46193230e-01 -2.17193112e-01
-7.13173389e-01 2.29870409e-01 6.89665616e-01 2.55408406e-01
4.61125299e-02 -8.53450298e-01 -8.22740555e-01 -1.73674196e-01
-3.17032129e-01 -5.36261380e-01 2.75973622e-02 4.10145760e-01
-9.77230191e-01 -4.16949764e-02 -1.17852139e+00 -3.41889232e-01
-1.16949105e+00 3.07400376e-01 8.29232991e-01 -2.48948261e-01
-8.85382369e-02 1.04121327e+00 1.20764067e-02 -6.97338641e-01
2.05663331e-02 3.70343387e-01 -4.77322280e-01 2.86050826e-01
-3.83688249e-02 6.32176816e-01 -2.52561778e-01 -4.10300523e-01
-3.93712074e-01 5.48665941e-01 6.76360071e-01 -6.56324208e-01
1.43840563e+00 -4.65267617e-03 -2.33127132e-01 5.97085118e-01
3.54348511e-01 -3.80694568e-01 -1.19740951e+00 5.53458452e-01
-3.34059671e-02 -3.10448438e-01 -3.20630699e-01 -1.00302064e+00
-6.66313529e-01 4.43780601e-01 6.37785256e-01 3.05872172e-01
1.08826745e+00 -5.35427570e-01 1.41824469e-01 3.61827612e-01
6.90501153e-01 -1.23744464e+00 2.60225356e-01 8.01985681e-01
7.06928849e-01 -6.56358659e-01 -3.77581567e-01 1.72913492e-01
-5.74490011e-01 1.09447920e+00 9.13071930e-01 -2.09889814e-01
3.46966654e-01 1.75042212e-01 -1.32358834e-01 4.96648513e-02
-9.08083200e-01 -4.07160729e-01 -3.37317854e-01 1.05536163e+00
7.31598213e-02 1.09127052e-01 -1.15250480e+00 1.29698262e-01
-6.49671972e-01 -8.88287202e-02 8.25975120e-01 1.11313772e+00
-7.58071303e-01 -1.29095995e+00 -5.42895675e-01 1.67789787e-01
-4.18141373e-02 1.85485810e-01 -8.32760870e-01 9.95137632e-01
6.84115112e-01 8.21208239e-01 8.12124610e-02 -4.86795425e-01
6.26044795e-02 -1.70715898e-01 7.03370154e-01 -2.00296193e-01
-1.15361416e+00 -4.78507012e-01 -4.50084992e-02 -2.26355717e-02
-1.90406993e-01 -7.82209098e-01 -1.01105022e+00 -4.28642184e-01
-3.61999929e-01 4.97608393e-01 8.28030229e-01 5.57199299e-01
1.98862895e-01 7.52568662e-01 4.79460031e-01 -7.65865803e-01
-2.39819422e-01 -5.43283641e-01 -5.24976432e-01 -7.01674074e-02
-2.94852644e-01 -7.97188580e-01 -3.07322115e-01 -3.37512255e-01] | [5.548349857330322, 3.9319703578948975] |
dea3e681-25e7-46e6-9b47-c2f1bc853ce1 | policy-gradient-methods-in-the-presence-of | 2305.05666 | null | https://arxiv.org/abs/2305.05666v1 | https://arxiv.org/pdf/2305.05666v1.pdf | Policy Gradient Methods in the Presence of Symmetries and State Abstractions | Reinforcement learning on high-dimensional and complex problems relies on abstraction for improved efficiency and generalization. In this paper, we study abstraction in the continuous-control setting, and extend the definition of MDP homomorphisms to the setting of continuous state and action spaces. We derive a policy gradient theorem on the abstract MDP for both stochastic and deterministic policies. Our policy gradient results allow for leveraging approximate symmetries of the environment for policy optimization. Based on these theorems, we propose a family of actor-critic algorithms that are able to learn the policy and the MDP homomorphism map simultaneously, using the lax bisimulation metric. Finally, we introduce a series of environments with continuous symmetries to further demonstrate the ability of our algorithm for action abstraction in the presence of such symmetries. We demonstrate the effectiveness of our method on our environments, as well as on challenging visual control tasks from the DeepMind Control Suite. Our method's ability to utilize MDP homomorphisms for representation learning leads to improved performance, and the visualizations of the latent space clearly demonstrate the structure of the learned abstraction. | ['Doina Precup', 'David Meger', 'Rosie Zhao', 'Sahand Rezaei-Shoshtari', 'Prakash Panangaden'] | 2023-05-09 | null | null | null | null | ['policy-gradient-methods', 'continuous-control'] | ['methodology', 'playing-games'] | [-1.86234787e-01 1.84258789e-01 -4.91455913e-01 1.64163157e-01
-4.38337266e-01 -7.94352472e-01 9.98447001e-01 -1.48095816e-01
-1.93028301e-01 7.38557100e-01 6.57011688e-01 -4.44325417e-01
-2.28075668e-01 -4.39869761e-01 -7.44935751e-01 -6.65524840e-01
-5.51800907e-01 2.46191531e-01 -1.38643980e-01 -2.62837976e-01
1.70813248e-01 7.21136272e-01 -1.38533854e+00 -4.70001474e-02
5.86289942e-01 5.73925793e-01 -1.16343364e-01 7.91152060e-01
3.23485553e-01 7.69300878e-01 -3.42983752e-01 4.06196415e-01
8.34467888e-01 -4.45690602e-01 -6.55406058e-01 2.29059666e-01
6.06435537e-01 -6.04750335e-01 -4.77872998e-01 7.41760314e-01
2.87687749e-01 5.08018196e-01 5.41941166e-01 -1.63171017e+00
-6.56821132e-01 3.89843225e-01 -3.70440304e-01 -4.19893488e-02
1.15961768e-01 6.77441418e-01 1.18747592e+00 -5.25130093e-01
8.53587747e-01 1.65152693e+00 3.29799592e-01 7.87578642e-01
-1.78862715e+00 -3.75853479e-01 5.92091799e-01 -1.54115111e-01
-8.71071041e-01 -4.51670140e-01 4.76748884e-01 -6.33752525e-01
1.13798535e+00 -1.10482477e-01 8.00781310e-01 1.22427022e+00
1.96258798e-01 7.84839272e-01 1.25853133e+00 -2.80567229e-01
6.58537328e-01 -9.61555466e-02 -2.76962668e-01 1.03630054e+00
8.03825725e-03 7.81510890e-01 -4.94563818e-01 -3.09254557e-01
1.11610973e+00 -8.55358597e-03 -1.40176445e-01 -1.25705647e+00
-1.26181483e+00 9.26512182e-01 3.02290708e-01 -2.68078595e-01
-2.81155586e-01 7.19562352e-01 4.45192248e-01 4.90679562e-01
9.80978087e-02 8.73199165e-01 -1.72454655e-01 -3.45566630e-01
-4.01703894e-01 6.40882492e-01 9.43979442e-01 1.01086867e+00
4.61685956e-01 4.87991959e-01 -4.21996355e-01 2.54421353e-01
3.42139490e-02 4.68683898e-01 5.83102070e-02 -1.81511497e+00
3.95334303e-01 1.72124192e-01 6.10602438e-01 -7.62239754e-01
-3.43125314e-01 -3.88294458e-01 -4.35297191e-01 1.03950441e+00
4.01911646e-01 -1.86065763e-01 -8.22418869e-01 2.24445510e+00
3.22541416e-01 1.21863425e-01 1.26041204e-01 7.60239542e-01
-6.04648173e-01 6.23982608e-01 1.42722763e-02 -1.55027300e-01
7.96367288e-01 -9.16454136e-01 -5.19744694e-01 -3.79035845e-02
6.74743652e-01 -6.53385790e-03 1.56619811e+00 1.90941453e-01
-1.20193946e+00 -3.06610376e-01 -1.14460766e+00 -1.54364863e-02
-7.66069591e-02 -1.55109242e-01 5.67911804e-01 8.36564675e-02
-1.14289987e+00 1.00238323e+00 -1.41385520e+00 -3.88154775e-01
4.19544339e-01 3.22521776e-01 -2.41598889e-01 2.87481844e-01
-8.28216493e-01 1.12431848e+00 2.59632140e-01 -3.63148361e-01
-1.58860040e+00 -9.85367715e-01 -7.77449429e-01 3.03601056e-01
6.25509739e-01 -8.17513525e-01 1.59897149e+00 -5.75145721e-01
-1.83586919e+00 3.31442654e-01 2.56446987e-01 -6.80805981e-01
6.53479815e-01 -1.25461787e-01 1.54073730e-01 2.26900071e-01
4.85464968e-02 5.65012693e-01 1.00603902e+00 -1.05451095e+00
-6.81002498e-01 -2.62855113e-01 5.83514929e-01 4.79051650e-01
-3.51335794e-01 -4.28707063e-01 2.92425841e-01 -3.47025633e-01
-3.96366417e-01 -1.15670156e+00 -4.13088411e-01 3.68804902e-01
-1.53706908e-01 -1.14780620e-01 9.56992567e-01 -2.06712469e-01
9.27872002e-01 -2.24239993e+00 7.38237143e-01 2.52169281e-01
3.50589514e-01 -7.90156946e-02 -1.69132128e-01 7.27237523e-01
-1.15557157e-01 1.62025169e-01 -2.33991355e-01 -3.55106384e-01
6.90598726e-01 4.93609697e-01 -8.67650330e-01 7.11852908e-01
1.76998600e-01 8.05369198e-01 -1.01424360e+00 -2.46578276e-01
2.17225686e-01 8.54613930e-02 -9.13226485e-01 2.17805326e-01
-6.25747323e-01 5.49140930e-01 -5.28642237e-01 1.20310701e-01
8.95292386e-02 1.34414406e-02 3.55835795e-01 2.28558898e-01
-3.91236305e-01 2.90640861e-01 -1.22312331e+00 1.86143982e+00
-5.67176521e-01 2.89425284e-01 4.60206568e-01 -8.54289114e-01
4.80476379e-01 2.01230139e-01 6.38200462e-01 -5.58734536e-01
-3.53241622e-01 -9.29923579e-02 -1.15046345e-01 -2.00359255e-01
3.17446321e-01 -1.88018173e-01 -6.23520277e-02 8.42675090e-01
-2.71991957e-02 -2.83903301e-01 2.83915967e-01 4.02061254e-01
1.12588286e+00 6.18799329e-01 2.01022342e-01 -5.93451023e-01
8.09179768e-02 -5.07939160e-02 4.05511886e-01 1.04603064e+00
-2.73046821e-01 -1.22700490e-01 1.22209203e+00 -4.25664693e-01
-1.51593542e+00 -1.38688588e+00 1.29375651e-01 1.15957272e+00
-3.27971011e-01 -6.66352868e-01 -4.11645353e-01 -6.45176530e-01
5.48449934e-01 7.37261653e-01 -8.88973117e-01 -3.86520892e-01
-6.69042289e-01 -1.38944099e-02 3.28505337e-01 6.59357846e-01
3.22869927e-01 -7.94023633e-01 -1.01945448e+00 1.58875600e-01
5.14426291e-01 -9.10282373e-01 -5.76618612e-01 -3.15982252e-02
-8.29064131e-01 -9.20255542e-01 -3.79621267e-01 -4.13710773e-01
3.42063159e-01 -1.15968540e-01 8.00564229e-01 -2.50166029e-01
-2.47114748e-01 1.14951861e+00 2.35840544e-01 -2.49215409e-01
-6.89420879e-01 -2.26047114e-02 3.86565983e-01 -3.35035056e-01
-3.36443514e-01 -9.18628812e-01 -4.97311860e-01 1.02971792e-01
-7.34155595e-01 2.56244481e-01 1.15762427e-01 9.79928017e-01
3.70603442e-01 -4.51746792e-01 2.17747897e-01 -4.31352615e-01
1.02543008e+00 -3.32028061e-01 -1.27296174e+00 3.71513702e-02
-6.22520685e-01 8.58831465e-01 6.94890380e-01 -4.32811618e-01
-8.45354557e-01 2.80872639e-02 5.44886827e-01 -7.15458810e-01
2.41774991e-01 1.66739181e-01 2.10591629e-01 1.25123769e-01
8.25500548e-01 6.42158613e-02 5.20016372e-01 -3.85312140e-01
8.34500551e-01 4.57610860e-02 3.84768188e-01 -1.45330632e+00
7.27894187e-01 7.36049056e-01 5.52452028e-01 -5.54765105e-01
-6.51589036e-01 1.51236296e-01 -2.66372502e-01 1.21959798e-01
7.60811388e-01 -7.09727407e-01 -1.16260111e+00 -5.51038720e-02
-7.60577738e-01 -9.00275052e-01 -8.13466072e-01 4.02508527e-01
-1.29882264e+00 1.55574918e-01 -5.92732668e-01 -7.96561241e-01
2.00411931e-01 -1.26977432e+00 9.07397151e-01 -2.12169483e-01
-1.60641391e-02 -9.63946760e-01 4.64966595e-01 -6.30005002e-01
5.35111368e-01 4.72095430e-01 1.21752715e+00 -2.00813606e-01
-7.68809557e-01 3.80546749e-01 1.71653047e-01 3.12601745e-01
-8.16050731e-03 -4.33332548e-02 -5.99040747e-01 -6.89411640e-01
-1.94577649e-02 -4.95943069e-01 6.77337408e-01 2.25609317e-01
1.23476088e+00 -5.19621313e-01 -5.87639622e-02 7.63240755e-01
1.23676336e+00 -1.33288959e-02 2.24586233e-01 3.78955811e-01
6.36199117e-01 4.44200993e-01 4.02603567e-01 7.16440022e-01
7.21354932e-02 7.54343748e-01 6.35944963e-01 2.89949346e-02
1.69506863e-01 -6.32683098e-01 6.48262620e-01 1.69432968e-01
-2.90883556e-02 3.88130307e-01 -7.63194263e-01 3.74325156e-01
-2.13094807e+00 -1.05762053e+00 8.58489156e-01 2.24618912e+00
7.89226532e-01 5.82808629e-02 5.17581284e-01 -4.68633652e-01
4.21741664e-01 3.13437015e-01 -9.85613942e-01 -7.76908576e-01
3.59104723e-01 4.01024818e-01 5.96493542e-01 8.15448344e-01
-9.19633687e-01 8.99609983e-01 6.95601749e+00 3.27144504e-01
-9.57772255e-01 -4.78641577e-02 1.04394197e-01 -3.99572700e-01
-2.14492887e-01 1.77151054e-01 -6.31482363e-01 2.49596059e-01
8.19213212e-01 -4.99959141e-01 1.05840755e+00 1.00099301e+00
3.73160303e-01 2.84618199e-01 -1.60286331e+00 6.23513520e-01
-5.30889034e-01 -1.55238509e+00 -1.09926760e-01 4.22830969e-01
8.76984656e-01 -1.19826235e-01 5.14114439e-01 5.97590983e-01
9.41263199e-01 -1.05967617e+00 6.46233499e-01 1.70748234e-01
7.36958086e-01 -7.54927516e-01 -3.36966515e-01 2.58761734e-01
-9.72307026e-01 -5.53202808e-01 -2.31598213e-01 -2.99498290e-01
-1.05373293e-01 -3.24685067e-01 -7.69997537e-01 7.45410472e-02
1.60214573e-01 9.65212345e-01 -3.06509994e-02 4.83450413e-01
-4.82024819e-01 2.78081536e-01 -4.65769917e-01 1.43172741e-01
6.41171873e-01 -3.40155393e-01 6.80890560e-01 8.69366229e-01
-9.79819102e-04 -8.33183229e-02 5.31429172e-01 1.18183398e+00
2.42184266e-03 -3.61891448e-01 -1.00594032e+00 -2.57048011e-01
2.93001205e-01 8.86969686e-01 -1.79540366e-01 -2.95127332e-01
-1.77309781e-01 5.72066128e-01 6.69757485e-01 7.87750661e-01
-8.51040959e-01 -8.91328305e-02 1.47051585e+00 -3.33808325e-02
4.87975746e-01 -8.87050152e-01 6.11794293e-02 -1.29015517e+00
-1.50021538e-03 -1.16397750e+00 3.30868334e-01 -4.21614408e-01
-9.11688030e-01 -2.03829646e-01 5.13736367e-01 -1.04679143e+00
-6.86144829e-01 -6.50474012e-01 -4.22184497e-01 8.16042304e-01
-1.12705231e+00 -7.37764835e-01 2.02098131e-01 7.34581470e-01
2.39002004e-01 -1.03093319e-01 9.22926247e-01 -2.27737069e-01
-4.20135528e-01 2.90169358e-01 4.00442064e-01 -1.65746316e-01
4.02311265e-01 -1.64218271e+00 5.08030534e-01 8.17045450e-01
6.15025451e-03 7.79171646e-01 8.32663536e-01 -3.31162989e-01
-1.80868375e+00 -8.22986007e-01 -2.43182525e-01 -4.94875520e-01
1.02062654e+00 -5.96276462e-01 -5.92711329e-01 1.14639461e+00
2.81900764e-01 2.24383175e-02 6.69323951e-02 2.69761831e-01
-5.98045826e-01 -1.39125645e-01 -9.93924499e-01 1.07275999e+00
1.30759907e+00 -7.26306856e-01 -5.92381358e-01 4.15329576e-01
9.44387257e-01 -6.01480067e-01 -8.45474482e-01 -1.27937809e-01
5.76159775e-01 -7.02353477e-01 1.04376209e+00 -1.43700075e+00
5.00502110e-01 -2.08100736e-01 -1.23520777e-01 -1.72160411e+00
-3.73566180e-01 -1.04895449e+00 -5.91824710e-01 5.53467751e-01
6.86379969e-02 -7.87810743e-01 5.04065752e-01 6.58186615e-01
-1.97128713e-01 -8.43925536e-01 -9.76464212e-01 -1.00451398e+00
4.44250494e-01 -3.63832898e-03 5.86769581e-01 7.72216380e-01
2.94178158e-01 1.14140928e-01 -4.70678031e-01 1.90702841e-01
8.15169275e-01 2.55487919e-01 8.41907978e-01 -5.17876208e-01
-8.98384333e-01 -5.49134731e-01 -2.69410938e-01 -1.08677495e+00
4.47553068e-01 -9.28446472e-01 -3.09841245e-01 -1.23462594e+00
-1.12879552e-01 -3.11024815e-01 -2.86487073e-01 6.51121497e-01
3.78778815e-01 -5.72634339e-01 6.43202662e-01 1.45572752e-01
-4.56827372e-01 9.70561385e-01 1.43230188e+00 -1.35204867e-02
-4.20877188e-01 -2.26637796e-01 -5.12509644e-01 5.64109981e-01
8.98424268e-01 -3.76926273e-01 -8.93827260e-01 -4.69469249e-01
7.42083043e-02 5.56823201e-02 5.82954764e-01 -8.71482670e-01
-7.78201818e-02 -6.49000347e-01 1.23062715e-01 1.34681627e-01
4.75990534e-01 -6.50341570e-01 -2.68361628e-01 7.38095105e-01
-8.56266201e-01 3.79361629e-01 3.01690459e-01 6.81321323e-01
3.87353957e-01 3.95428717e-01 9.10046339e-01 -1.70199484e-01
-5.39485514e-01 4.39428359e-01 -2.77558684e-01 4.85714108e-01
1.04330111e+00 2.58420914e-01 -4.44923222e-01 -4.55924392e-01
-9.07362580e-01 5.29099762e-01 7.52692342e-01 2.65449405e-01
3.69485825e-01 -1.41199207e+00 -4.22868699e-01 2.73857921e-01
-6.09584376e-02 -3.60142380e-01 -2.65572608e-01 7.01336205e-01
-2.77527004e-01 2.22452402e-01 -6.60128951e-01 -5.33268452e-01
-9.20928836e-01 7.81070888e-01 6.27483249e-01 -3.16399723e-01
-9.59040463e-01 1.54588297e-01 3.46003681e-01 -1.85060039e-01
4.06410545e-01 -7.46837437e-01 4.90887374e-01 -4.31008548e-01
4.80672538e-01 2.97771156e-01 -5.44850171e-01 2.14704633e-01
-4.08976674e-02 2.41095051e-01 -8.73536170e-02 -7.54112601e-01
1.32713819e+00 1.14160568e-01 1.45649552e-01 3.49809289e-01
1.07665527e+00 -2.59637922e-01 -2.08872819e+00 -8.14285278e-02
-1.43056467e-01 -6.15572691e-01 -2.18837231e-01 -5.57812393e-01
-6.03101015e-01 9.52083826e-01 4.27256525e-01 2.30396539e-01
7.04723179e-01 -1.63075313e-01 1.26024991e-01 7.66158700e-01
3.86743844e-01 -1.26593864e+00 2.03075886e-01 6.10024989e-01
1.12321806e+00 -6.51510358e-01 7.65446573e-02 2.90340900e-01
-6.62407517e-01 1.01413178e+00 5.52711666e-01 -5.33375680e-01
3.40332568e-01 3.30605716e-01 -3.97295684e-01 -3.98141630e-02
-1.11242628e+00 -6.90816939e-02 -3.34770083e-01 6.94232404e-01
-1.32971078e-01 2.55867150e-02 6.51741773e-02 -3.59581798e-01
-7.04703331e-02 -1.52023181e-01 6.87969625e-01 1.33273315e+00
-4.01786000e-01 -1.22292554e+00 2.03471910e-03 6.76456317e-02
-4.02704021e-03 2.77123988e-01 -7.35842064e-02 1.15385675e+00
-4.76163268e-01 3.12495857e-01 3.90184261e-02 4.08433974e-02
4.55296397e-01 1.56610057e-01 8.74230087e-01 -6.96630538e-01
-2.47975573e-01 -2.70648837e-01 -5.28173670e-02 -1.19081557e+00
-1.10114522e-01 -6.78822458e-01 -1.23183334e+00 -4.75450158e-01
5.49497306e-01 1.34006843e-01 7.03396142e-01 6.11716270e-01
5.84129155e-01 3.94023865e-01 6.90881729e-01 -7.39086926e-01
-1.76879263e+00 -4.77227867e-01 -6.58583045e-01 4.43807721e-01
8.86290491e-01 -9.09489810e-01 -4.00082916e-01 -2.00559378e-01] | [4.126101016998291, 1.893982172012329] |
e5e60996-cbd6-4970-a120-acebb2c82d24 | multiple-instance-learning-via-iterative-self | 2210.09452 | null | https://arxiv.org/abs/2210.09452v1 | https://arxiv.org/pdf/2210.09452v1.pdf | Multiple Instance Learning via Iterative Self-Paced Supervised Contrastive Learning | Learning representations for individual instances when only bag-level labels are available is a fundamental challenge in multiple instance learning (MIL). Recent works have shown promising results using contrastive self-supervised learning (CSSL), which learns to push apart representations corresponding to two different randomly-selected instances. Unfortunately, in real-world applications such as medical image classification, there is often class imbalance, so randomly-selected instances mostly belong to the same majority class, which precludes CSSL from learning inter-class differences. To address this issue, we propose a novel framework, Iterative Self-paced Supervised Contrastive Learning for MIL Representations (ItS2CLR), which improves the learned representation by exploiting instance-level pseudo labels derived from the bag-level labels. The framework employs a novel self-paced sampling strategy to ensure the accuracy of pseudo labels. We evaluate ItS2CLR on three medical datasets, showing that it improves the quality of instance-level pseudo labels and representations, and outperforms existing MIL methods in terms of both bag and instance level accuracy. | ['Carlos Fernandez-Granda', 'Krzysztof J. Geras', 'Narges Razavian', 'Sheng Liu', 'Yiqiu Shen', 'Weicheng Zhu', 'Kangning Liu'] | 2022-10-17 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Liu_Multiple_Instance_Learning_via_Iterative_Self-Paced_Supervised_Contrastive_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Liu_Multiple_Instance_Learning_via_Iterative_Self-Paced_Supervised_Contrastive_Learning_CVPR_2023_paper.pdf | cvpr-2023-1 | ['multiple-instance-learning'] | ['methodology'] | [ 6.65453196e-01 3.28435451e-02 -8.14127684e-01 -5.51568031e-01
-1.31313813e+00 -3.31876367e-01 3.97615343e-01 7.04108894e-01
-2.32288435e-01 1.06567097e+00 4.36965078e-02 1.39844745e-01
-3.43624353e-01 -8.27598333e-01 -6.51646495e-01 -8.54349196e-01
-5.91956638e-02 7.04706967e-01 -1.16959445e-01 9.63795558e-02
9.88277867e-02 2.43988857e-01 -1.81043553e+00 8.25244427e-01
9.88808155e-01 9.75271463e-01 -2.37087145e-01 2.41491869e-01
-4.91172642e-01 9.82119262e-01 -7.06700385e-01 -3.63798179e-02
2.00892121e-01 -5.49225450e-01 -7.97171712e-01 1.80521160e-01
6.46311760e-01 1.66198581e-01 1.88375652e-01 7.32791245e-01
4.13574755e-01 8.84570032e-02 9.15647984e-01 -1.34183335e+00
-2.64845490e-01 6.07668340e-01 -6.65878892e-01 1.51040688e-01
3.43535133e-02 -1.39539525e-01 1.05685151e+00 -6.39279306e-01
3.86451364e-01 1.17686820e+00 7.77496457e-01 7.21796334e-01
-1.55848396e+00 -8.89028907e-01 3.90103787e-01 6.80302009e-02
-1.32043767e+00 -2.92040527e-01 8.81824374e-01 -4.02357519e-01
3.54535490e-01 3.95572543e-01 3.40782285e-01 1.10096610e+00
-6.36755899e-02 1.14250934e+00 1.46692967e+00 -4.81156468e-01
4.94802922e-01 4.19108450e-01 4.46476519e-01 3.88610005e-01
3.94835263e-01 -9.41000506e-02 -6.81102931e-01 -4.49142456e-01
6.23716414e-01 2.25589916e-01 -7.42526129e-02 -6.34784162e-01
-1.18909943e+00 7.88017333e-01 5.04848599e-01 4.35404271e-01
-3.85083079e-01 1.51078533e-02 4.57254708e-01 1.57631531e-01
9.20016825e-01 4.63188231e-01 -5.33556759e-01 9.33820754e-02
-1.01554644e+00 2.19088033e-01 4.36157882e-01 7.87117660e-01
8.17480803e-01 -2.68635035e-01 -5.55697560e-01 1.20982826e+00
-9.80262384e-02 1.76098228e-01 9.21441317e-01 -4.03841972e-01
5.13728440e-01 9.25539792e-01 -1.27264306e-01 -7.52889752e-01
-3.54392141e-01 -8.90928864e-01 -1.03524756e+00 1.09973280e-02
2.86355108e-01 1.98792428e-01 -8.41002584e-01 1.78983033e+00
4.40795600e-01 3.79691660e-01 1.59600869e-01 4.91549492e-01
1.08672607e+00 3.70665580e-01 3.36622864e-01 -5.27205884e-01
1.10885358e+00 -9.40371931e-01 -6.74402535e-01 -3.35019976e-01
8.42845619e-01 -8.90473723e-02 8.80011320e-01 3.36655527e-01
-8.19448054e-01 -6.17098927e-01 -1.06939960e+00 4.44777399e-01
-3.44192684e-01 8.40580389e-02 7.17335343e-01 5.39720953e-01
-4.57992703e-01 7.20539570e-01 -6.11832857e-01 2.30875999e-01
8.68170321e-01 1.55604467e-01 -4.71294075e-01 -3.49217057e-01
-1.08564413e+00 4.39695120e-01 5.36551297e-01 -1.58072785e-01
-6.27020955e-01 -9.69796419e-01 -8.89174700e-01 -3.64506841e-02
3.47577751e-01 -2.96448857e-01 9.93996322e-01 -1.24776196e+00
-9.05518174e-01 1.14313185e+00 -1.63971439e-01 -4.51204181e-01
6.40247822e-01 -5.08320052e-03 -3.28971505e-01 -1.98267549e-02
2.67814398e-01 6.82087719e-01 9.23662484e-01 -1.43489945e+00
-5.34339666e-01 -4.53978986e-01 -1.29488394e-01 2.56384343e-01
-3.71085554e-01 -6.23477817e-01 6.12852303e-03 -7.61918426e-01
3.02260637e-01 -6.16279662e-01 -2.48378024e-01 -2.49800816e-01
-2.97277033e-01 -6.37441635e-01 2.79770881e-01 -2.19554752e-01
1.16154742e+00 -2.19642687e+00 -1.01918459e-01 5.73026426e-02
3.27803135e-01 2.77334154e-01 -2.29302660e-01 6.32645339e-02
-4.05240268e-01 -4.34817821e-02 -3.96457434e-01 -4.84064490e-01
-3.09543639e-01 3.86929154e-01 -1.54120639e-01 2.53671259e-01
4.20610875e-01 7.20736802e-01 -1.34877205e+00 -6.10631943e-01
5.35452962e-02 2.98207760e-01 -3.80793184e-01 3.09023142e-01
-1.74886540e-01 6.17891669e-01 -3.01493496e-01 6.78264320e-01
6.13339126e-01 -5.17006397e-01 2.42059290e-01 -2.22024113e-01
5.84443808e-01 2.91530192e-01 -1.25233030e+00 1.49693227e+00
-6.52883589e-01 2.41848916e-01 -6.22183561e-01 -1.41738701e+00
1.22130692e+00 3.21689218e-01 7.50229180e-01 -7.93917239e-01
-1.92544311e-01 2.54469752e-01 -3.82799596e-01 -1.41451448e-01
7.48446286e-02 -3.38317007e-01 -3.09574045e-02 3.69112581e-01
4.78116982e-02 8.23625457e-03 2.05780834e-01 8.50385502e-02
8.99889588e-01 9.15127695e-02 6.30332887e-01 -1.32155016e-01
4.65296000e-01 -8.89295116e-02 8.12843323e-01 8.97469938e-01
-1.84158862e-01 8.04644287e-01 4.15765524e-01 -6.29341781e-01
-5.68633735e-01 -1.06044269e+00 -5.35683811e-01 1.08399165e+00
8.66140276e-02 -4.27336574e-01 -6.23244822e-01 -1.28408003e+00
1.96668193e-01 7.19397545e-01 -9.16310310e-01 -3.97363037e-01
-3.84697258e-01 -1.14261723e+00 7.16647059e-02 5.82729399e-01
2.19152823e-01 -1.17052889e+00 -4.01525378e-01 4.03664500e-01
-3.15833181e-01 -8.91754091e-01 -1.54850334e-01 5.35790503e-01
-1.18406582e+00 -1.31485379e+00 -6.74651146e-01 -5.74514925e-01
1.02399707e+00 1.46382868e-01 1.42331421e+00 9.35489461e-02
-5.30290663e-01 2.08241448e-01 -4.97358799e-01 -3.99090201e-01
-5.03115892e-01 2.36528978e-01 -1.29273176e-01 3.61915380e-01
3.89086157e-01 -4.18106556e-01 -4.59954023e-01 1.77067906e-01
-9.34467196e-01 1.25074372e-01 5.67864597e-01 1.13944268e+00
9.69428182e-01 8.53487924e-02 1.11283374e+00 -1.42635620e+00
4.37561810e-01 -6.19722962e-01 -1.94615945e-01 3.84035349e-01
-8.34387541e-01 1.54467434e-01 5.84042788e-01 -6.49357796e-01
-5.98724961e-01 -2.81479228e-02 6.25763088e-02 -3.30313951e-01
-2.54026473e-01 3.32244515e-01 1.35102123e-01 2.19412148e-01
9.65272009e-01 2.91049361e-01 1.40389159e-01 -4.24509674e-01
1.63605884e-01 8.32570672e-01 2.57599264e-01 -6.65974140e-01
4.12881166e-01 2.60834664e-01 -6.84057176e-02 -4.06918734e-01
-1.60890949e+00 -6.05291009e-01 -5.64132988e-01 -1.20331503e-01
2.57353902e-01 -1.02391946e+00 -3.71535152e-01 2.71250904e-01
-5.31859040e-01 -3.25436682e-01 -7.89608538e-01 3.54563266e-01
-6.07794762e-01 -5.17183170e-02 -3.06229353e-01 -8.16956997e-01
-4.53923047e-01 -1.17138755e+00 1.28860927e+00 1.88137993e-01
-2.91394055e-01 -1.03925276e+00 1.54033005e-01 5.29650509e-01
2.73628771e-01 5.40463269e-01 9.47542012e-01 -9.84306991e-01
-1.06147870e-01 -4.20623779e-01 -5.49797341e-02 4.71539021e-01
6.09326541e-01 -6.70602262e-01 -1.10731447e+00 -6.07225299e-01
-3.29846561e-01 -7.55541384e-01 9.79666471e-01 2.99123734e-01
1.66315687e+00 -2.62914807e-01 -4.13574338e-01 3.50866020e-01
1.34888375e+00 -1.22009866e-01 4.51036602e-01 2.24200025e-01
5.42018652e-01 6.06190562e-01 7.92759240e-01 4.55801100e-01
1.92060828e-01 6.27562821e-01 3.14717472e-01 -2.06983298e-01
-2.79729068e-01 -1.82253540e-01 -1.25159964e-01 5.93552709e-01
2.03486726e-01 1.00584529e-01 -7.31603444e-01 6.61561370e-01
-1.87451565e+00 -7.05047250e-01 2.19290584e-01 2.55707312e+00
1.25225711e+00 1.74612954e-01 8.41058567e-02 5.44940472e-01
7.58214772e-01 2.28716768e-02 -7.80282915e-01 -6.90307794e-03
1.01384565e-01 3.77691954e-01 3.39659303e-01 1.61879644e-01
-1.26081717e+00 5.01405478e-01 6.25255680e+00 1.05562198e+00
-9.97168839e-01 2.03496590e-01 1.18358517e+00 -1.15731589e-01
-2.21880510e-01 -3.58354002e-01 -9.24425364e-01 5.59645414e-01
7.38789618e-01 3.43269445e-02 7.02488869e-02 9.91698623e-01
-2.53564805e-01 6.46693185e-02 -1.30066299e+00 1.26221418e+00
2.81926870e-01 -1.41007102e+00 2.25598872e-01 -1.49948657e-01
9.35147524e-01 -3.58099639e-01 3.24865244e-02 6.24038994e-01
1.77730918e-01 -1.03762233e+00 2.45497122e-01 3.92365158e-01
1.00351882e+00 -9.02967393e-01 7.80817926e-01 4.46094334e-01
-1.03354299e+00 -1.14446521e-01 -2.88444936e-01 2.46318668e-01
-2.70343274e-01 9.25385118e-01 -9.98604596e-01 6.94574714e-01
5.54604471e-01 9.46666837e-01 -6.62692845e-01 9.48709488e-01
-4.10736352e-02 6.91560507e-01 1.11559898e-01 1.48241416e-01
6.21418878e-02 5.56592904e-02 1.35577038e-01 1.02631879e+00
-1.35969818e-01 -2.46142045e-01 3.74790311e-01 6.52020752e-01
-1.00430265e-01 2.96340466e-01 -2.87391484e-01 1.36118025e-01
4.80545700e-01 1.19001853e+00 -6.72304094e-01 -5.01479089e-01
1.68492217e-02 6.96933985e-01 5.20405710e-01 -5.33005111e-02
-5.93950510e-01 -1.50134657e-02 4.52950031e-01 1.22578748e-01
-1.70440767e-02 4.43951875e-01 -2.33818978e-01 -1.01944888e+00
3.21733765e-02 -9.51401889e-01 8.54049325e-01 -3.62496853e-01
-1.78974700e+00 5.61338902e-01 5.32567203e-02 -1.65797067e+00
-2.87323117e-01 -2.87998915e-01 -1.77396134e-01 5.16046226e-01
-1.70734739e+00 -1.14099336e+00 -4.73192602e-01 3.88416976e-01
6.21946335e-01 -1.51516259e-01 1.07848632e+00 2.22602561e-01
-4.82543737e-01 1.13018405e+00 2.68622667e-01 1.04204319e-01
1.01588809e+00 -1.28367281e+00 -9.66502875e-02 9.97052938e-02
3.77256453e-01 2.90837586e-01 3.51860940e-01 -4.29672599e-01
-8.42458427e-01 -1.44095671e+00 6.89436138e-01 -2.79003382e-01
1.67972688e-02 -2.47709572e-01 -1.06437540e+00 4.77798790e-01
-4.74854857e-01 6.71672344e-01 1.07438517e+00 3.15336913e-01
-5.35492957e-01 -5.30697286e-01 -1.52126658e+00 2.70899892e-01
9.37225759e-01 -2.39681914e-01 -4.96695757e-01 7.60807812e-01
5.88764012e-01 -3.71106088e-01 -1.02261066e+00 7.17003524e-01
3.28731656e-01 -7.59572685e-01 1.07416523e+00 -8.06192279e-01
3.94663930e-01 -1.65762603e-01 -8.68354887e-02 -1.47749591e+00
-1.91858232e-01 4.09633592e-02 -2.74246514e-01 1.22759438e+00
3.04848790e-01 -7.95444965e-01 8.81083608e-01 2.47944057e-01
2.28262290e-01 -1.11798799e+00 -8.59564662e-01 -8.50758374e-01
1.89858284e-02 -2.38887459e-01 6.97855592e-01 1.27988267e+00
-1.04130574e-01 2.66569912e-01 -3.32818121e-01 -3.79513539e-02
8.23809326e-01 3.66420418e-01 5.86843252e-01 -1.58822405e+00
-3.21108669e-01 -1.97094321e-01 -6.24468625e-01 -5.26307523e-01
1.37822673e-01 -1.22427166e+00 3.23646255e-02 -1.41844189e+00
5.43112993e-01 -1.17407882e+00 -8.29502344e-01 7.62362599e-01
-6.05053186e-01 4.31037873e-01 -1.46386027e-01 2.35332161e-01
-8.18394721e-01 4.37166512e-01 1.15806854e+00 -3.80650967e-01
-2.11831599e-01 1.12887464e-01 -7.95896471e-01 3.22223365e-01
7.44228899e-01 -7.15939224e-01 -4.56502765e-01 1.02597155e-01
-1.56372730e-02 2.27539451e-03 5.11874743e-02 -1.15967464e+00
-1.14179157e-01 -1.84366077e-01 5.76552510e-01 -6.56956911e-01
1.58020169e-01 -5.91805220e-01 4.47535552e-02 4.77807075e-01
-8.75464737e-01 -4.83767986e-01 8.49169791e-02 6.66632116e-01
-2.98353881e-01 -2.95093060e-01 9.78872001e-01 -2.57179737e-01
-4.77072358e-01 3.69806528e-01 1.76759705e-01 3.57092887e-01
1.08006704e+00 -1.88508004e-01 -1.74929604e-01 -7.56915361e-02
-7.47508228e-01 1.92984968e-01 8.73833224e-02 4.29584980e-01
5.21109819e-01 -1.57092714e+00 -9.31733906e-01 3.50679636e-01
7.98715353e-01 4.63168561e-01 3.89648467e-01 6.67577684e-01
1.27953989e-03 9.35786888e-02 5.03761061e-02 -9.66535211e-01
-1.21957171e+00 5.38739383e-01 2.69154459e-01 -6.81561172e-01
-5.03905654e-01 7.92243659e-01 1.92338064e-01 -7.97827482e-01
3.54081124e-01 -9.23015084e-03 -3.94110352e-01 3.18389028e-01
8.50940049e-01 3.35468054e-01 3.63924414e-01 -3.02346051e-01
-2.52781510e-01 3.91190141e-01 -4.62518036e-01 4.78991508e-01
1.45894015e+00 2.99938709e-01 7.81426132e-02 9.88300025e-01
1.24157572e+00 -3.53817314e-01 -1.21180236e+00 -5.56102872e-01
-1.24676274e-02 -5.20113051e-01 -7.86609575e-02 -8.85017633e-01
-1.05406678e+00 6.38374388e-01 8.13390791e-01 -7.51782730e-02
1.01103091e+00 -5.63707873e-02 4.73015577e-01 2.37212166e-01
5.88442266e-01 -1.04677856e+00 4.64484096e-01 4.15313952e-02
6.81887448e-01 -1.47800028e+00 4.88441288e-02 -4.24708784e-01
-5.38044274e-01 9.56504166e-01 6.95724487e-01 -1.54241621e-01
3.45272243e-01 2.22991612e-02 1.00544682e-02 8.68254453e-02
-6.82280660e-01 -8.34569056e-03 3.97101462e-01 5.75514793e-01
5.36837876e-01 3.28568280e-01 -2.54696250e-01 4.64725465e-01
2.18007341e-01 -9.00820270e-02 1.17034271e-01 1.01393664e+00
-1.95121348e-01 -1.28560793e+00 -4.19531196e-01 9.43120539e-01
-2.69007564e-01 -1.84811950e-02 -5.03796414e-02 4.62744772e-01
1.74089193e-01 7.16409445e-01 2.56944686e-01 -1.34001479e-01
1.30716681e-01 2.38575161e-01 6.08700454e-01 -8.05779278e-01
-5.93367159e-01 -3.34790736e-01 -5.81804514e-02 -4.00711268e-01
-5.50055742e-01 -4.42695558e-01 -1.10041642e+00 2.31764898e-01
-4.11369920e-01 2.76070148e-01 4.16988015e-01 9.95602965e-01
3.12913209e-01 7.95372009e-01 9.69098628e-01 -6.38222814e-01
-6.57464981e-01 -9.49704945e-01 -6.09950542e-01 8.96967947e-01
4.56330061e-01 -8.91166449e-01 -5.42029321e-01 -1.54119223e-01] | [9.525864601135254, 3.4253644943237305] |
b23dc5a2-6a8c-48bb-8325-ab74aff9ac91 | balanced-districting-on-grid-graphs-with | 2102.05028 | null | https://arxiv.org/abs/2102.05028v1 | https://arxiv.org/pdf/2102.05028v1.pdf | Balanced Districting on Grid Graphs with Provable Compactness and Contiguity | Given a graph $G = (V,E)$ with vertex weights $w(v)$ and a desired number of parts $k$, the goal in graph partitioning problems is to partition the vertex set V into parts $V_1,\ldots,V_k$. Metrics for compactness, contiguity, and balance of the parts $V_i$ are frequent objectives, with much existing literature focusing on compactness and balance. Revisiting an old method known as striping, we give the first polynomial-time algorithms with guaranteed contiguity and provable bicriteria approximations for compactness and balance for planar grid graphs. We consider several types of graph partitioning, including when vertex weights vary smoothly or are stochastic, reflecting concerns in various real-world instances. We show significant improvements in experiments for balancing workloads for the fire department and reducing over-policing using 911 call data from South Fulton, GA. | ['Yao Xie', 'Swati Gupta', 'Shixiang Zhu', 'Cyrus Hettle'] | 2021-02-09 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [-3.41133356e-01 3.00542802e-01 -4.31942195e-01 -3.71958539e-02
-2.92089611e-01 -7.20880151e-01 -6.14582479e-01 4.71859008e-01
-2.96567883e-02 8.49037826e-01 -3.26130778e-01 -7.09792554e-01
-8.66914153e-01 -1.10511518e+00 -4.45683807e-01 -2.70278245e-01
-8.10557306e-01 1.21022213e+00 3.23190898e-01 1.47130108e-03
2.13119790e-01 7.58388877e-01 -9.75581765e-01 -1.48838580e-01
7.07947850e-01 7.24961042e-01 3.59394052e-03 7.39587069e-01
-1.92407370e-01 3.07472587e-01 -5.11284649e-01 -4.09142584e-01
6.56693399e-01 -2.01034471e-01 -1.30495143e+00 6.44074976e-01
-1.33370236e-01 2.80374497e-01 -4.25936967e-01 6.69875264e-01
2.12292716e-01 6.06844537e-02 5.54915965e-01 -1.82662559e+00
-3.33735198e-02 1.20889318e+00 -1.47047794e+00 4.60357577e-01
8.12362954e-02 6.33855956e-03 1.25626028e+00 1.12840328e-02
6.09041512e-01 1.02462268e+00 6.13501310e-01 -3.48376743e-02
-1.74405789e+00 -6.38282239e-01 2.39270821e-01 -2.63915777e-01
-1.85048509e+00 -9.96490195e-02 4.99665439e-01 -3.38801175e-01
1.11795366e+00 7.74018466e-01 5.16921937e-01 -3.42251092e-01
4.53321747e-02 3.17143172e-01 6.95788085e-01 -3.16169858e-01
1.96379289e-01 -1.59668364e-02 4.44009155e-01 7.16227889e-01
9.81953561e-01 -2.94477254e-01 -2.25301292e-02 -5.82871556e-01
5.82887530e-01 -2.22703949e-01 -2.73713082e-01 -8.64463270e-01
-6.43675983e-01 7.84465134e-01 2.22310856e-01 4.95179109e-02
-1.53112739e-01 5.73083878e-01 4.45237488e-01 1.61036819e-01
1.95108891e-01 4.91872609e-01 -4.80510354e-01 9.74353254e-02
-1.19050443e+00 1.87596738e-01 9.96373475e-01 1.79928493e+00
7.59195268e-01 -1.14891574e-01 -9.51251388e-02 6.07223332e-01
-7.28536770e-02 4.09114867e-01 -6.52704597e-01 -9.85660732e-01
8.21327448e-01 7.69898593e-01 1.04816638e-01 -1.13710403e+00
-6.32390320e-01 -2.94967055e-01 -6.90764010e-01 -2.83584267e-01
3.58716637e-01 -1.29450247e-01 -7.80867875e-01 1.83591354e+00
2.20470309e-01 -3.33859235e-01 -5.86783588e-01 4.31304842e-01
1.91553980e-01 5.14099538e-01 -1.02689780e-01 -8.21870029e-01
9.96720314e-01 -7.09290028e-01 -2.30529308e-01 -1.90457255e-01
5.90152442e-01 -6.84475839e-01 6.58340156e-01 2.99397290e-01
-1.73801684e+00 -8.30233023e-02 -5.50996304e-01 5.11882424e-01
-4.09173071e-02 -4.26172197e-01 5.16107559e-01 1.08986032e+00
-1.45216775e+00 3.64632040e-01 -5.55526555e-01 -4.56105888e-01
3.94507319e-01 6.07205272e-01 -9.06091481e-02 -3.33729446e-01
-3.58985662e-01 4.04316485e-01 2.94317424e-01 -4.53307867e-01
-4.38137829e-01 -6.33697450e-01 -6.22853518e-01 4.07506883e-01
7.29555249e-01 -4.17877465e-01 6.39970958e-01 -2.09717035e-01
-2.76839972e-01 8.44746172e-01 9.49257612e-02 -1.67739913e-01
1.97642177e-01 8.72874022e-01 -2.12964594e-01 1.52972192e-01
1.55446604e-01 2.71415174e-01 2.16810063e-01 -1.44068277e+00
-7.23315775e-01 -4.54425037e-01 2.25100935e-01 3.45797390e-02
5.49226953e-03 6.18871786e-02 -4.70072895e-01 -2.32794672e-01
2.17214197e-01 -9.81407881e-01 -5.41606247e-01 -7.34525442e-01
-8.07554543e-01 -1.55335709e-01 3.91235828e-01 -3.11809063e-01
1.95838332e+00 -1.94443870e+00 3.64401750e-02 1.06885028e+00
7.06770062e-01 -1.49384722e-01 -9.73525494e-02 8.30725968e-01
-1.40965596e-01 5.14719963e-01 -1.26674190e-01 2.30884030e-01
1.48242682e-01 2.80834794e-01 2.36399509e-02 7.10597694e-01
-3.98953617e-01 4.57787961e-01 -5.63328981e-01 -4.21378046e-01
-2.48699427e-01 -4.34034616e-01 -8.70254040e-01 -1.20010138e-01
-4.38581780e-02 -3.60251695e-01 -9.86172557e-02 6.12500489e-01
1.11115682e+00 -3.24768215e-01 7.71880925e-01 2.95348048e-01
-4.20513051e-03 9.24895983e-03 -1.51699233e+00 1.02119064e+00
2.46838816e-02 9.60853547e-02 8.61612976e-01 -1.17942607e+00
7.12728977e-01 -2.16765538e-01 9.53374624e-01 -3.79685849e-01
3.23924243e-01 1.58502445e-01 7.19417930e-02 -7.46522695e-02
6.98054731e-01 -1.49928242e-01 -4.88675177e-01 1.00935256e+00
-4.15069193e-01 -3.48871648e-01 7.96076536e-01 7.95649290e-01
1.70470047e+00 -6.85951710e-01 1.92434371e-01 -9.86822665e-01
-2.50694066e-01 4.20584321e-01 4.88953412e-01 5.38686574e-01
-4.33519423e-01 3.67007941e-01 1.11446786e+00 -2.00535864e-01
-1.20414007e+00 -1.09595478e+00 1.26985416e-01 8.82413685e-01
5.72491229e-01 -4.30330902e-01 -9.46744084e-01 -3.79902333e-01
3.28812540e-01 5.25627196e-01 -5.52179456e-01 3.02153435e-02
-4.98146713e-01 -6.87977374e-01 1.31454989e-01 4.42287177e-01
-1.85145602e-01 -5.95907569e-01 -6.26822054e-01 4.13069427e-01
-2.62841046e-01 -9.97812748e-01 -1.18676198e+00 4.16100919e-01
-9.41909552e-01 -1.49273682e+00 -4.70597520e-02 -7.73492575e-01
9.88842905e-01 7.35558450e-01 1.55164576e+00 2.68448681e-01
-7.40084827e-01 4.92873192e-01 -1.04096547e-01 -1.27983645e-01
2.10487440e-01 1.94345191e-01 -1.34390563e-01 -6.36124730e-01
1.18385121e-01 -5.20745993e-01 -5.98456144e-01 6.40808165e-01
-7.87756324e-01 -2.58830070e-01 9.41111520e-02 3.94797444e-01
8.47785711e-01 7.19265282e-01 2.43924916e-01 -1.16153860e+00
6.84451878e-01 -6.78899646e-01 -4.58299339e-01 3.19368839e-01
-7.48211205e-01 -1.59510687e-01 3.96923840e-01 1.47175035e-02
-2.36516684e-01 -3.71902019e-01 4.42982674e-01 -3.36109757e-01
3.80453497e-01 4.08561915e-01 -2.01930538e-01 9.89687890e-02
3.49138409e-01 -3.02262902e-01 4.72267345e-02 3.79604138e-02
2.29459032e-01 2.13817671e-01 1.90222844e-01 -8.40950906e-01
5.37236094e-01 3.91805112e-01 2.14187607e-01 -4.63591963e-01
7.20657483e-02 -6.58616185e-01 -2.16004267e-01 -1.76648516e-02
1.11107789e-01 -3.92247826e-01 -1.16991365e+00 -2.60061204e-01
-5.87659359e-01 -5.92362642e-01 -5.16776741e-01 -6.09129556e-02
-8.29578459e-01 3.90566975e-01 -5.11066735e-01 -8.83088291e-01
-2.00739935e-01 -8.49123657e-01 5.38800061e-01 -3.00646983e-02
-3.09476942e-01 -6.63710654e-01 1.56593442e-01 1.49964496e-01
2.62302607e-01 2.46781796e-01 1.41319597e+00 -4.24654663e-01
-5.91958404e-01 -9.69949514e-02 -5.71079910e-01 -3.47779602e-01
-1.82858075e-03 4.33244593e-02 5.29715344e-02 -8.95770431e-01
-5.22550881e-01 4.85789366e-02 4.77428168e-01 8.30054343e-01
1.40460896e+00 -5.61635852e-01 -9.85368133e-01 4.50827599e-01
1.80092609e+00 4.90700334e-01 5.79424560e-01 -1.84511781e-01
3.02425146e-01 6.44938886e-01 6.77464545e-01 8.43439519e-01
3.65428269e-01 2.40561470e-01 4.96656090e-01 -8.62409323e-02
1.61157504e-01 2.05843627e-01 -4.23212826e-01 4.89816457e-01
-4.48462330e-02 -9.34382856e-01 -1.14495814e+00 9.91946101e-01
-1.66115808e+00 -8.86809111e-01 -1.94973961e-01 2.36843061e+00
3.63196850e-01 2.65735030e-01 7.57056475e-01 3.35031360e-01
1.15437388e+00 -6.66362494e-02 -3.05448174e-01 -9.09374356e-01
2.45951965e-01 4.92431402e-01 1.19081306e+00 5.76453924e-01
-4.16716516e-01 4.36728269e-01 7.13835192e+00 1.03888440e+00
-4.09475297e-01 -1.15089104e-01 1.19160306e+00 -5.91265261e-01
-7.05537677e-01 -1.95694491e-02 -5.36885619e-01 4.60601360e-01
1.00573993e+00 -4.44425493e-01 8.16135108e-01 7.51359165e-01
1.34785160e-01 -4.10434693e-01 -9.10558462e-01 1.00721192e+00
-2.89177537e-01 -1.27224529e+00 -3.63957226e-01 5.78101814e-01
8.15283179e-01 -3.61238033e-01 -9.59114060e-02 2.44477257e-01
9.03108120e-01 -1.21288383e+00 4.04044151e-01 -1.09172821e-01
1.11771381e+00 -1.62540007e+00 2.68012762e-01 3.82061005e-01
-1.69201946e+00 -2.68460572e-01 -4.18485940e-01 1.65553942e-01
2.90106416e-01 6.61076248e-01 -6.40242875e-01 4.87586647e-01
8.24468851e-01 -3.69140416e-01 1.25933662e-01 9.46830034e-01
7.92161882e-01 3.73537779e-01 -5.98233879e-01 2.46082947e-01
2.43424386e-01 -4.41478729e-01 3.68137330e-01 1.03582573e+00
3.12879175e-01 7.41550863e-01 5.16076684e-01 6.84067845e-01
-1.50582999e-01 3.24768662e-01 -6.66318655e-01 -2.10394204e-01
8.23986888e-01 1.12264097e+00 -1.53640616e+00 -5.97077832e-02
-1.37728736e-01 1.63995698e-01 3.90110493e-01 1.81754261e-01
-1.15376508e+00 -7.76686966e-01 8.15565407e-01 9.09459174e-01
4.21856254e-01 -4.62655246e-01 -6.32456481e-01 -3.37486356e-01
-1.96507722e-01 -6.30421519e-01 8.38777184e-01 -2.62143701e-01
-9.79953468e-01 6.61457956e-01 2.46486023e-01 -5.84370852e-01
4.20530289e-01 -2.89861709e-01 -5.40572941e-01 7.86806107e-01
-7.65503407e-01 -6.08357012e-01 -4.00381275e-02 4.81795341e-01
1.72600478e-01 1.65972978e-01 1.07594743e-01 1.49353191e-01
-4.95079279e-01 5.53940475e-01 1.89116020e-02 -3.05982918e-01
8.80048648e-02 -1.08486784e+00 3.41173440e-01 8.86706054e-01
-3.38351995e-01 4.87739712e-01 7.81989515e-01 -7.11519718e-01
-1.39664674e+00 -1.06042814e+00 6.74207032e-01 -7.62275830e-02
3.67360294e-01 -2.94637412e-01 -4.25645798e-01 7.80918062e-01
2.02053890e-01 3.66837829e-02 6.95313632e-01 2.14769080e-01
-2.95418594e-02 -1.51956901e-01 -1.67386544e+00 4.63048398e-01
1.60484040e+00 1.53120175e-01 3.68235350e-01 2.35784575e-01
5.54627359e-01 -4.39521134e-01 -7.79029071e-01 4.16301280e-01
2.21192062e-01 -9.63101447e-01 8.28550160e-01 -6.57367885e-01
-1.87910542e-01 -6.16465732e-02 -1.85856894e-01 -1.17999923e+00
-8.90445769e-01 -8.80977929e-01 3.42292786e-01 9.03217673e-01
5.84942877e-01 -5.00457466e-01 1.09159565e+00 9.56634700e-01
-1.49428532e-01 -9.47108567e-01 -8.33408535e-01 -9.58547831e-01
1.24454759e-01 -2.50931025e-01 9.98267889e-01 8.31729293e-01
5.62707543e-01 1.80215701e-01 -2.39301063e-02 5.82528785e-02
7.64348984e-01 6.58853173e-01 7.34386563e-01 -1.08137846e+00
-3.95083129e-01 -7.20255077e-01 -1.37144268e-01 -7.81479359e-01
-1.34906068e-01 -9.46931362e-01 -5.98957315e-02 -1.77710438e+00
6.49232268e-01 -1.23295009e+00 1.10453255e-01 4.44284737e-01
1.56934083e-01 1.73586354e-01 -3.11890710e-03 -1.88468248e-02
-7.84281373e-01 -1.14822336e-01 1.09867525e+00 -1.38548627e-01
-3.35279316e-01 -3.75779867e-02 -1.08686221e+00 1.97717935e-01
8.07745814e-01 -3.55071127e-01 -7.94583917e-01 -2.76063859e-01
2.59117573e-01 7.19652534e-01 -5.05875051e-01 -5.04052639e-01
-4.29314114e-02 -7.85769165e-01 -2.42198423e-01 -7.77486384e-01
-9.58792865e-02 -9.20052052e-01 9.06334639e-01 6.86080992e-01
-1.04998998e-01 5.33945382e-01 3.32291037e-01 6.07717991e-01
1.53927475e-01 7.76328519e-02 1.00849307e+00 -6.47058040e-02
-1.37974009e-01 9.46029365e-01 -2.58801758e-01 5.68503022e-01
1.71920907e+00 -6.76808894e-01 -5.80914080e-01 -2.20263630e-01
-8.28866899e-01 7.11685181e-01 7.25788772e-01 -2.27610543e-01
3.44231129e-01 -1.08205819e+00 -5.85303843e-01 -4.78526056e-02
3.05450168e-02 5.40957153e-02 4.69980568e-01 9.73470449e-01
-1.06067669e+00 3.70081455e-01 -1.49766371e-01 -3.65961373e-01
-1.34990799e+00 1.05819178e+00 7.28314370e-02 -5.49065292e-01
-2.46078387e-01 1.05156839e+00 2.34232694e-01 2.51809899e-02
-2.23766472e-02 -2.53305733e-01 5.33475399e-01 -1.50780454e-01
1.13552045e-02 9.61505771e-01 8.83219168e-02 -4.16832864e-01
-5.40051281e-01 1.89878076e-01 4.42652255e-02 3.14113833e-02
1.10170615e+00 -4.63860273e-01 -4.73530084e-01 -3.16748321e-01
8.45626593e-01 2.81530529e-01 -7.01856613e-01 1.87848136e-02
-7.03669712e-02 -8.76993418e-01 -5.56100130e-01 -1.68533087e-01
-1.56461668e+00 2.27695286e-01 -1.45830572e-01 1.04168248e+00
1.17450666e+00 2.60857373e-01 7.41810203e-01 -2.32024565e-01
1.08185875e+00 -1.25168324e+00 -2.36724168e-01 2.41196334e-01
5.03025115e-01 -3.44017029e-01 1.40419528e-01 -8.85755777e-01
-4.31522399e-01 7.43308485e-01 6.37364328e-01 -1.68824688e-01
9.43610072e-01 6.86662853e-01 -6.52442932e-01 -5.66136718e-01
-8.04860592e-01 -2.77931303e-01 -5.51370025e-01 5.20825863e-01
7.86425322e-02 6.68677390e-01 -7.53943324e-01 5.38666427e-01
-1.92359239e-01 -5.46072125e-01 6.45130813e-01 1.02676284e+00
-8.58833253e-01 -1.05737352e+00 -3.95601720e-01 1.10361922e+00
-3.32965642e-01 6.49864078e-02 -2.07031474e-01 9.67808962e-01
1.02392212e-02 1.19729555e+00 4.54234868e-01 -3.26640069e-01
3.06891203e-01 -4.40316200e-01 6.38413012e-01 -8.51562798e-01
-4.18067843e-01 1.53715864e-01 4.31028694e-01 -5.02232373e-01
1.95643380e-01 -5.65484524e-01 -1.35209870e+00 -1.35680330e+00
-5.74781299e-01 6.61268950e-01 2.31799394e-01 3.57328385e-01
4.48844433e-01 5.12016058e-01 8.48015010e-01 -3.10070723e-01
-1.69942915e-01 -5.09017169e-01 -1.57980609e+00 1.13328151e-01
-3.16521019e-01 -6.31899118e-01 -2.88357496e-01 -3.41132909e-01] | [6.903982162475586, 5.1224822998046875] |
bcd3319e-da6f-4d2b-8172-09539b03a767 | semi-uformer-semi-supervised-uncertainty | 2210.16057 | null | https://arxiv.org/abs/2210.16057v1 | https://arxiv.org/pdf/2210.16057v1.pdf | Semi-UFormer: Semi-supervised Uncertainty-aware Transformer for Image Dehazing | Image dehazing is fundamental yet not well-solved in computer vision. Most cutting-edge models are trained in synthetic data, leading to the poor performance on real-world hazy scenarios. Besides, they commonly give deterministic dehazed images while neglecting to mine their uncertainty. To bridge the domain gap and enhance the dehazing performance, we propose a novel semi-supervised uncertainty-aware transformer network, called Semi-UFormer. Semi-UFormer can well leverage both the real-world hazy images and their uncertainty guidance information. Specifically, Semi-UFormer builds itself on the knowledge distillation framework. Such teacher-student networks effectively absorb real-world haze information for quality dehazing. Furthermore, an uncertainty estimation block is introduced into the model to estimate the pixel uncertainty representations, which is then used as a guidance signal to help the student network produce haze-free images more accurately. Extensive experiments demonstrate that Semi-UFormer generalizes well from synthetic to real-world images. | ['Mingqiang Wei', 'Xuefeng Yan', 'Peng Cui', 'Yongzhen Wang', 'Ming Tong'] | 2022-10-28 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 1.50385454e-01 1.48786724e-01 9.10175890e-02 -5.85692644e-01
-6.80840313e-01 -1.40652731e-01 3.71539026e-01 -2.69162714e-01
-7.92075098e-02 8.09354961e-01 -3.29040028e-02 -2.44200021e-01
-9.75704491e-02 -1.03679109e+00 -1.01098633e+00 -9.17024076e-01
5.17140388e-01 2.15688929e-01 3.51144820e-01 -1.59984127e-01
-2.38475259e-02 -1.70027643e-01 -1.62366521e+00 9.33214575e-02
1.96582079e+00 1.14413631e+00 3.90220284e-01 5.15555918e-01
7.08643720e-02 1.00701392e+00 -6.78452790e-01 -4.24896687e-01
2.99066335e-01 -2.29302660e-01 -2.87591845e-01 3.15455139e-01
6.22675836e-01 -7.21338272e-01 -5.42461634e-01 1.66233516e+00
3.29156399e-01 2.22423196e-01 6.85140431e-01 -1.40620673e+00
-1.21955717e+00 6.19885147e-01 -6.42927885e-01 1.05608709e-01
-2.43855745e-01 5.22643566e-01 6.55838668e-01 -9.93158162e-01
4.82722372e-02 1.31306243e+00 3.14404786e-01 3.60024571e-01
-8.77636552e-01 -9.93500054e-01 2.66860962e-01 4.83699232e-01
-1.35476100e+00 -2.22463906e-01 9.48852658e-01 -3.02333713e-01
2.95526031e-02 -1.70267820e-01 5.98812103e-01 8.43215406e-01
7.54178539e-02 1.27368045e+00 1.33605504e+00 -2.12320685e-01
2.16252446e-01 3.75987321e-01 -1.41098559e-01 4.94862020e-01
4.24414128e-01 3.97114068e-01 -4.63642240e-01 4.76023614e-01
6.53978705e-01 1.71072602e-01 -6.83627248e-01 -2.34271750e-01
-8.15482259e-01 7.79911160e-01 8.46366286e-01 -2.53216207e-01
-3.07374001e-01 9.82770473e-02 -2.80905604e-01 3.08575869e-01
7.39572108e-01 5.14932513e-01 -1.61550567e-01 3.27737421e-01
-1.11023855e+00 3.18101078e-01 2.87594914e-01 1.05893707e+00
1.16517770e+00 4.50194359e-01 -2.44562864e-01 7.81071663e-01
6.05613470e-01 7.66795635e-01 1.64873406e-01 -1.04394376e+00
3.47632080e-01 4.91224706e-01 2.03048512e-01 -8.65361691e-01
3.13161433e-01 -4.96178299e-01 -8.59036267e-01 5.32351613e-01
1.81337088e-01 -1.22237436e-01 -1.54722631e+00 1.50657094e+00
3.81888151e-01 7.94558346e-01 7.71054104e-02 1.08786571e+00
8.11379075e-01 1.05415630e+00 -3.01599443e-01 -5.78651354e-02
1.06709778e+00 -1.32071614e+00 -8.56058657e-01 -3.14975798e-01
5.46698421e-02 -4.28210676e-01 9.58437264e-01 7.21884847e-01
-9.86681759e-01 -4.88341093e-01 -1.33961308e+00 -3.76503309e-03
-3.41510534e-01 -2.36617699e-01 3.47384363e-01 6.19223416e-01
-8.32736313e-01 4.56698984e-01 -6.85610235e-01 4.50913727e-01
8.59281182e-01 -6.03050739e-02 7.40829781e-02 -7.82575369e-01
-1.52596021e+00 8.84121060e-01 7.21200466e-01 3.43158692e-01
-1.48374724e+00 -1.07376814e+00 -1.17053545e+00 -1.09839223e-01
7.60310292e-01 -7.05766380e-01 1.13761270e+00 -9.76224959e-01
-1.55757511e+00 2.88551718e-01 3.19588780e-01 -4.94258642e-01
4.81428206e-01 -3.72188747e-01 -5.33557892e-01 2.63690323e-01
7.38011524e-02 8.62189710e-01 1.54482698e+00 -1.78844309e+00
-7.49090314e-01 -1.86813831e-01 2.72599846e-01 2.66359746e-01
-2.48336941e-01 -7.36234367e-01 -4.44927812e-01 -1.05305099e+00
-1.59308225e-01 -5.38514853e-01 -2.76977152e-01 2.72560894e-01
-3.82399410e-01 -2.19371468e-02 8.53076041e-01 -5.16338527e-01
1.10996056e+00 -1.96401346e+00 -4.95280093e-03 1.62115648e-01
4.12885278e-01 4.94389266e-01 -1.64473429e-01 -1.59970120e-01
5.26334271e-02 -7.26496950e-02 -5.80840945e-01 -8.82640034e-02
-5.29103912e-02 5.93279898e-01 -5.30799747e-01 3.00323963e-01
6.17874384e-01 9.25869524e-01 -1.28570127e+00 -6.11789823e-01
4.87541169e-01 5.48620105e-01 -4.48943287e-01 5.32208383e-01
-6.11193776e-01 4.53508377e-01 -5.11137307e-01 7.56871700e-01
1.14984572e+00 -1.02660269e-01 -7.03795135e-01 -1.64082319e-01
1.57558545e-01 -1.04283854e-01 -9.86237824e-01 1.68811762e+00
-6.39541566e-01 5.21898746e-01 6.99789673e-02 -7.46390283e-01
7.77840555e-01 5.11239506e-02 -9.95212793e-02 -5.40632010e-01
1.47419661e-01 5.92844337e-02 -2.57938981e-01 -4.14119333e-01
3.82511348e-01 -2.79992342e-01 5.63321114e-01 2.74844408e-01
1.30541265e-01 -9.46411431e-01 -4.32378538e-02 3.27898175e-01
4.36478257e-01 2.79464394e-01 -1.61129132e-01 -2.57700384e-01
5.57903290e-01 -2.14930087e-01 7.88867593e-01 6.96401775e-01
-3.54342580e-01 1.00109828e+00 3.20174396e-02 -3.33114415e-02
-6.64265156e-01 -1.32170463e+00 -2.08873808e-01 6.34607255e-01
5.68126142e-01 -1.07709788e-01 -9.80068803e-01 -7.63430953e-01
-1.28453150e-01 1.07883477e+00 -8.56801033e-01 -6.39872134e-01
1.33975804e-01 -6.08019173e-01 1.09881386e-01 2.47577861e-01
8.91881049e-01 -5.43147027e-01 -2.19627857e-01 2.64775585e-02
-1.72335207e-01 -9.48351562e-01 -5.58089316e-01 -4.18617502e-02
-5.97683251e-01 -9.38446164e-01 -8.57311606e-01 -3.54154229e-01
6.64235473e-01 6.40253305e-01 1.09247971e+00 9.58999544e-02
2.55334768e-02 5.80884479e-02 -4.31019038e-01 -1.08258271e+00
-3.10827315e-01 -4.63100165e-01 -1.22855911e-02 1.89197123e-01
3.20030749e-01 -7.31408656e-01 -8.02029014e-01 3.67062002e-01
-1.41508842e+00 5.81448935e-02 7.19057739e-01 1.07759809e+00
5.27236521e-01 7.72469401e-01 3.93351495e-01 -8.64044964e-01
2.62698829e-01 -7.12226152e-01 -8.21244955e-01 2.54693687e-01
-1.08723187e+00 2.12730765e-01 4.35275197e-01 -4.92956936e-01
-1.60356283e+00 -3.17970425e-01 8.78449231e-02 -1.04548442e+00
5.67934513e-02 4.72146571e-01 -4.92544889e-01 -1.22852527e-01
5.85341513e-01 3.30059946e-01 1.38690388e-02 -1.12781778e-01
5.71384728e-01 6.06019557e-01 9.10324156e-01 -7.41134763e-01
1.53824508e+00 6.38768733e-01 -3.62611860e-01 -5.66431582e-01
-1.55137968e+00 -3.02142918e-01 -2.51135916e-01 -1.77562162e-01
7.89261580e-01 -1.42635524e+00 -1.16031647e-01 9.01146114e-01
-8.98606122e-01 -2.80301243e-01 -3.78399730e-01 4.08514351e-01
-3.35819274e-01 4.16258365e-01 -2.15945050e-01 -9.85798717e-01
-1.54273659e-01 -1.21064913e+00 9.13938463e-01 6.98429167e-01
9.05564964e-01 -9.91770864e-01 -1.79280564e-01 5.94526291e-01
2.97985464e-01 1.01214886e-01 7.55379081e-01 -3.04736402e-02
-9.40799534e-01 1.61397547e-01 -5.60481489e-01 9.96852100e-01
1.36464328e-01 1.44587815e-01 -1.23810577e+00 3.22582386e-02
2.38055721e-01 -6.74840927e-01 1.36455274e+00 4.52541709e-01
1.38112032e+00 -1.61386803e-01 1.67581588e-01 8.06799471e-01
1.13335478e+00 -9.73345563e-02 7.27831900e-01 3.22438538e-01
8.35159957e-01 6.98907375e-01 9.07149017e-01 2.85285205e-01
6.72621310e-01 1.52397618e-01 1.01581216e+00 1.41887777e-02
-3.08101684e-01 -4.68220800e-01 3.00249517e-01 7.22841859e-01
2.56027102e-01 -4.27725315e-01 -6.42195106e-01 7.22489953e-01
-1.79781580e+00 -6.14254534e-01 1.61938742e-01 1.87316692e+00
1.13828540e+00 2.09888220e-01 -4.00717765e-01 -4.09541875e-02
6.02484405e-01 3.84473562e-01 -8.55625331e-01 2.05848053e-01
-1.60221696e-01 -9.40049738e-02 3.33627313e-01 4.54411417e-01
-1.14669430e+00 9.83449101e-01 5.16434860e+00 1.04169619e+00
-8.54676485e-01 1.16894743e-03 7.53980637e-01 9.10344496e-02
-8.90687227e-01 -6.35000691e-02 -3.78377289e-01 6.55029714e-01
6.58791840e-01 -1.65620759e-01 4.60630536e-01 8.06632400e-01
1.38465613e-01 -4.01507884e-01 -8.34862292e-01 1.10044754e+00
1.84489816e-01 -1.26136184e+00 3.91120225e-01 -7.96294138e-02
1.31137657e+00 -1.37331098e-01 5.06964862e-01 4.48244959e-01
8.19227099e-01 -1.24099624e+00 8.38503182e-01 5.62777340e-01
6.07159555e-01 -9.64570165e-01 8.58172536e-01 4.44476992e-01
-7.98213303e-01 -5.08466735e-02 -7.29007602e-01 2.05597252e-01
1.43677145e-01 1.16553247e+00 -5.82542479e-01 8.80131423e-01
7.45046794e-01 8.80392075e-01 -6.72022462e-01 1.08374655e+00
-8.40592504e-01 8.48737359e-01 -1.91982538e-01 6.18819594e-01
4.77812320e-01 -5.04289448e-01 4.24745321e-01 7.45332360e-01
3.76739800e-01 3.79205644e-01 1.14715591e-01 1.19443583e+00
-5.46263009e-02 -6.20600045e-01 -2.95853972e-01 -1.07967779e-01
5.46210647e-01 9.68001783e-01 -2.10810244e-01 -3.52903754e-01
-3.84185940e-01 7.97211409e-01 1.40790269e-01 5.55040359e-01
-9.26281810e-01 -5.51750958e-01 7.91164875e-01 -8.73912200e-02
3.06716561e-01 1.54916018e-01 -2.30760753e-01 -1.32674980e+00
-2.14127168e-01 -9.72946167e-01 1.42001018e-01 -1.32941926e+00
-1.67900884e+00 4.24390912e-01 1.90457985e-01 -1.20718384e+00
1.17380954e-02 -6.76187098e-01 -8.63916159e-01 8.99662733e-01
-2.35828471e+00 -1.29439545e+00 -6.95647597e-01 7.54478574e-01
7.21055448e-01 5.21129891e-02 1.06289357e-01 1.08783126e-01
-7.86891639e-01 4.94187087e-01 1.13829024e-01 -3.59278098e-02
9.23591733e-01 -1.53039312e+00 6.07215576e-02 1.29466581e+00
1.03694104e-01 4.77536082e-01 9.27850962e-01 -5.02150416e-01
-9.99799967e-01 -1.63204455e+00 -1.47124186e-01 -6.64474368e-01
7.25452244e-01 -9.99023393e-02 -1.27041316e+00 4.03386921e-01
3.24587077e-01 3.60517114e-01 2.69759953e-01 -3.97588104e-01
-7.84804344e-01 -3.33610207e-01 -1.06562102e+00 6.18536532e-01
6.44298851e-01 -5.72614014e-01 -8.92209113e-01 2.77509868e-01
1.36898994e+00 -5.53646564e-01 -5.58645368e-01 4.99235809e-01
1.05191588e-01 -1.17421651e+00 9.03357923e-01 -2.03536332e-01
8.06481779e-01 -6.09879196e-01 -6.06229762e-03 -1.92464054e+00
7.07592815e-02 -6.81774259e-01 -4.62426543e-01 1.14232039e+00
3.25787485e-01 -5.50187528e-01 7.67429173e-01 3.51704150e-01
-2.97601551e-01 -7.18953550e-01 -5.07004380e-01 -7.68092394e-01
2.26736411e-01 -6.34443164e-01 9.68385696e-01 9.33607280e-01
-7.09110618e-01 -4.12718244e-02 -5.58866978e-01 6.28713191e-01
1.13777208e+00 3.95933092e-02 6.75936878e-01 -1.10050702e+00
-2.72697777e-01 -6.58622757e-02 -2.98869461e-01 -1.22018766e+00
2.27770463e-01 -4.33951348e-01 5.44601917e-01 -1.36795378e+00
4.91185337e-02 -3.47228020e-01 -5.28151870e-01 2.62942404e-01
-8.70563507e-01 1.58092111e-01 -1.54505856e-02 2.72048600e-02
-3.66581321e-01 1.15058875e+00 1.76500380e+00 -6.33066595e-01
2.71892361e-02 1.18874520e-01 -8.77517581e-01 8.06180298e-01
5.41659832e-01 -5.18200874e-01 -1.13047349e+00 -7.29293466e-01
1.43542573e-01 -2.93013483e-01 4.21956629e-01 -9.71955955e-01
2.28563368e-01 -5.69488406e-01 3.29187214e-01 -6.26248598e-01
2.39732042e-01 -8.08409572e-01 -3.48424584e-01 -4.02071103e-02
-3.95319797e-03 -5.39591730e-01 -1.16049424e-01 1.06579065e+00
-5.45069277e-01 -2.97939271e-01 9.65622723e-01 -1.63953498e-01
-8.73661757e-01 7.60062158e-01 4.07910123e-02 3.10162276e-01
1.00760818e+00 -3.52686256e-01 -5.81522167e-01 -7.13445008e-01
-2.97120273e-01 7.99750566e-01 5.33888221e-01 3.36189568e-01
9.89229500e-01 -1.22002983e+00 -6.50844514e-01 2.30789766e-01
4.70120877e-01 1.03333890e+00 5.45183539e-01 5.68184018e-01
-4.06664103e-01 -1.43923506e-01 -2.01186561e-03 -7.03798056e-01
-6.02728248e-01 8.97836030e-01 2.60740042e-01 1.57306179e-01
-4.61405843e-01 1.18343973e+00 6.94726944e-01 -4.70983446e-01
2.03645289e-01 -4.44975793e-01 4.45170663e-02 -2.08726615e-01
8.65656793e-01 9.72074568e-02 -1.43958390e-01 -2.28818610e-01
1.19938381e-01 2.37883210e-01 -1.33335263e-01 -1.11593232e-01
1.16926551e+00 -4.42094475e-01 -1.20051131e-01 3.54123533e-01
7.19268382e-01 -1.73008129e-01 -1.99986601e+00 -5.71472347e-01
-2.26125509e-01 -8.14965963e-01 6.58519387e-01 -8.36021066e-01
-1.30787182e+00 1.18693519e+00 4.45435822e-01 -2.76191115e-01
1.37961197e+00 -3.11973959e-01 9.13835943e-01 5.20108402e-01
3.20406914e-01 -9.42518890e-01 4.14302230e-01 3.09230536e-01
6.88862920e-01 -1.60836291e+00 1.09387666e-01 -5.82833827e-01
-7.97786057e-01 7.40192294e-01 1.02716541e+00 -8.87567401e-02
8.89747739e-01 1.16789661e-01 3.72441351e-01 -1.44133687e-01
-6.92202210e-01 -2.08873048e-01 4.01693106e-01 1.05313993e+00
-2.91701317e-01 -1.84171468e-01 5.95027447e-01 6.27802372e-01
1.67377647e-02 4.40486670e-02 7.73082197e-01 6.88710570e-01
-8.09510291e-01 -5.37426829e-01 -4.96583313e-01 3.03419352e-01
5.80509454e-02 -3.57901990e-01 -8.13764799e-03 4.95186538e-01
3.80181998e-01 1.19995451e+00 -2.22983122e-01 -3.89118552e-01
2.68603731e-02 -3.10208142e-01 3.94318283e-01 -7.72740841e-01
-4.27648425e-02 -1.30057752e-01 -3.92901272e-01 -3.93673539e-01
-3.78222078e-01 -5.57258613e-02 -1.02863586e+00 -1.31070778e-01
-6.60800934e-01 4.40738738e-01 4.52934653e-01 1.07261217e+00
1.13105394e-01 8.13234329e-01 8.41672480e-01 -7.12937951e-01
-7.62022197e-01 -8.24547291e-01 -8.32800329e-01 6.87677637e-02
6.88849568e-01 -9.35083687e-01 -5.46677291e-01 3.40977684e-02] | [10.938675880432129, -3.1439783573150635] |
7c11e5b7-398a-43d0-ad60-0b6900afc3f4 | local-group-invariant-representations-via | 1612.01988 | null | http://arxiv.org/abs/1612.01988v2 | http://arxiv.org/pdf/1612.01988v2.pdf | Local Group Invariant Representations via Orbit Embeddings | Invariance to nuisance transformations is one of the desirable properties of
effective representations. We consider transformations that form a \emph{group}
and propose an approach based on kernel methods to derive local group invariant
representations. Locality is achieved by defining a suitable probability
distribution over the group which in turn induces distributions in the input
feature space. We learn a decision function over these distributions by
appealing to the powerful framework of kernel methods and generate local
invariant random feature maps via kernel approximations. We show uniform
convergence bounds for kernel approximation and provide excess risk bounds for
learning with these features. We evaluate our method on three real datasets,
including Rotated MNIST and CIFAR-10, and observe that it outperforms competing
kernel based approaches. The proposed method also outperforms deep CNN on
Rotated-MNIST and performs comparably to the recently proposed
group-equivariant CNN. | ['Bernhard Schölkopf', 'P. Thomas Fletcher', 'Anant Raj', 'Youssef Mroueh', 'Abhishek Kumar'] | 2016-12-06 | null | null | null | null | ['rotated-mnist'] | ['computer-vision'] | [-6.27965033e-02 1.44388154e-01 -1.33598760e-01 -5.40073633e-01
-9.91263151e-01 -6.73648000e-01 8.08050692e-01 -1.94196120e-01
-6.58218682e-01 5.66668093e-01 4.46654022e-01 -5.16074784e-02
-4.86275136e-01 -7.05115795e-01 -9.83256280e-01 -9.60518181e-01
-2.88543195e-01 2.64074385e-01 1.35674447e-01 1.48290411e-01
2.29157414e-02 8.26237917e-01 -1.16427398e+00 1.83206871e-01
7.87327468e-01 6.33335233e-01 -3.26550305e-01 7.07497776e-01
5.38142920e-01 8.76440585e-01 -2.05633819e-01 -1.85756758e-01
4.19612080e-01 -2.19077706e-01 -8.45401168e-01 1.30189303e-03
7.45682955e-01 -1.74066618e-01 -5.47033966e-01 1.35403848e+00
3.29648316e-01 7.42565334e-01 1.39317417e+00 -1.02568340e+00
-1.25529671e+00 6.53462291e-01 -2.50505656e-01 3.95310879e-01
-1.61240786e-01 -3.07074636e-01 1.11034632e+00 -1.10079932e+00
3.65225732e-01 1.38757658e+00 6.49875045e-01 3.52520972e-01
-1.76094365e+00 -5.51115215e-01 -8.42682421e-02 -3.42784040e-02
-1.51888335e+00 -3.33355278e-01 3.87107939e-01 -4.96386081e-01
7.94692338e-01 3.82022224e-02 8.66657123e-02 1.16047013e+00
3.55190933e-01 7.90270388e-01 9.55609381e-01 -5.84715128e-01
4.91513073e-01 6.52879775e-02 4.08003628e-01 8.80627275e-01
2.09125265e-01 7.16353953e-02 -2.37562388e-01 -6.13800824e-01
1.17329776e+00 1.14056811e-01 -3.82330328e-01 -8.68625939e-01
-1.35423100e+00 1.29082727e+00 5.79281688e-01 9.82328057e-02
-2.18226805e-01 5.20317674e-01 4.63239342e-01 3.34401757e-01
6.19083881e-01 2.47246623e-01 -1.59028664e-01 3.31522286e-01
-5.63585162e-01 3.33373487e-01 7.98032939e-01 8.67487490e-01
9.26442385e-01 2.67550230e-01 -2.70523518e-01 8.12669635e-01
9.69704315e-02 6.56247556e-01 4.10032481e-01 -8.48220825e-01
2.24489331e-01 7.73511976e-02 -3.19597982e-02 -8.41226757e-01
-3.55481148e-01 -3.20408940e-01 -9.78033066e-01 2.17547387e-01
5.23191631e-01 5.36885783e-02 -9.59811389e-01 1.95519531e+00
3.59265842e-02 2.56252140e-01 2.04869434e-01 6.03701413e-01
1.38613760e-01 6.85825467e-01 4.59949896e-02 3.05376261e-01
9.49181437e-01 -6.93139732e-01 -4.80891764e-01 3.53540361e-01
7.45317280e-01 -4.92243558e-01 9.80553687e-01 1.76908046e-01
-6.87758327e-01 -4.78503823e-01 -9.60901082e-01 -5.41715184e-03
-4.20390666e-01 3.77527297e-01 8.93648386e-01 6.72416985e-01
-1.27997100e+00 8.34278047e-01 -9.99812424e-01 -3.89629960e-01
4.57956463e-01 5.46363592e-01 -8.35376680e-01 5.84676526e-02
-1.04553282e+00 6.94695055e-01 4.88316715e-01 -1.78058594e-01
-9.72347200e-01 -1.03165376e+00 -1.12496138e+00 -1.63659956e-02
-2.20606282e-01 -5.14709949e-01 9.46431577e-01 -9.59673703e-01
-1.54219437e+00 6.03980541e-01 2.38590956e-01 -6.68397784e-01
3.07886183e-01 -3.72872710e-01 -1.70402825e-01 3.62032413e-01
3.98578905e-02 3.84261459e-01 1.41220379e+00 -5.99180222e-01
-4.15446728e-01 -1.36190340e-01 -6.80213794e-02 -6.68456331e-02
-5.12296557e-01 1.59182191e-01 1.35777220e-01 -7.80228853e-01
-1.37699425e-01 -1.10028613e+00 -2.72208363e-01 -9.36375782e-02
-4.34723884e-01 -4.05867010e-01 6.79417491e-01 -5.17076969e-01
6.65582359e-01 -2.34553123e+00 1.74156979e-01 5.68174899e-01
1.51451305e-01 -1.47518832e-02 -2.16341108e-01 2.01152444e-01
-5.84326982e-01 -2.14104336e-02 -5.33678345e-02 -2.60471795e-02
3.38895917e-01 -2.80687809e-02 -8.43980849e-01 1.19056785e+00
1.40397549e-01 9.97341037e-01 -6.83175743e-01 4.52729575e-02
2.71473438e-01 8.72948945e-01 -7.30467379e-01 -1.42713767e-02
3.40070635e-01 -6.11101054e-02 -4.80705589e-01 1.75270155e-01
6.62234843e-01 -1.70237750e-01 -2.80213743e-01 -2.96569288e-01
2.40687221e-01 -3.45646329e-02 -9.99340892e-01 1.43243349e+00
-4.09474552e-01 5.54139376e-01 -3.08327287e-01 -1.27796412e+00
8.46534431e-01 2.19715774e-01 7.22311884e-02 6.46523386e-02
3.49312760e-02 2.36853380e-02 -3.13575983e-01 1.13950528e-01
1.80581331e-01 -1.30912572e-01 -2.64551848e-01 5.51613450e-01
5.93818665e-01 2.94449568e-01 -4.37207609e-01 3.06279987e-01
1.25478387e+00 3.49585563e-02 4.25365537e-01 -9.96126354e-01
2.96573669e-01 -6.22093737e-01 1.81219012e-01 1.02780306e+00
-8.88208151e-02 5.69909036e-01 6.29822791e-01 -6.62120402e-01
-1.10543323e+00 -1.67805457e+00 -3.61747503e-01 1.15031600e+00
-3.16429734e-01 -1.71234101e-01 -8.06464136e-01 -9.81507540e-01
2.09710330e-01 6.41833067e-01 -1.23447621e+00 -6.56715512e-01
-2.83452660e-01 -7.67808080e-01 7.90159166e-01 8.21462810e-01
4.90182221e-01 -8.48137081e-01 -5.48654608e-02 -6.20414801e-02
4.01268005e-01 -9.20652390e-01 -7.49465108e-01 5.05920410e-01
-7.86173105e-01 -6.76995873e-01 -7.36278415e-01 -6.77454889e-01
9.26256180e-01 1.10194147e-01 6.22325659e-01 -6.41296506e-01
-4.02578652e-01 6.55795753e-01 -1.35308027e-01 -1.45315289e-01
-3.02343428e-01 3.76725852e-01 5.32530189e-01 4.77802038e-01
2.55347222e-01 -5.83424151e-01 -4.12809074e-01 2.74460673e-01
-1.03546560e+00 -4.45988476e-01 6.81064308e-01 9.00144517e-01
4.91579384e-01 1.49705335e-02 6.40218794e-01 -9.55576897e-01
7.04950452e-01 -3.57361794e-01 -7.58570433e-01 2.65093267e-01
-2.33683825e-01 6.81437790e-01 7.10795343e-01 -7.18206406e-01
-9.66981590e-01 2.77627796e-01 5.41477382e-01 -6.85431242e-01
-1.30969241e-01 1.85625240e-01 -2.23095700e-01 -3.22198123e-01
1.07868564e+00 1.78879917e-01 -5.36782779e-02 -3.75987738e-01
6.73516929e-01 2.51615196e-01 6.81974053e-01 -9.66255665e-01
1.29160082e+00 8.26680183e-01 1.25203608e-02 -9.54553962e-01
-7.91986823e-01 -4.07017916e-01 -8.59341800e-01 3.08614463e-01
9.13750768e-01 -9.73956227e-01 -4.12869483e-01 2.95838654e-01
-8.45430017e-01 -4.02474135e-01 -5.92532456e-01 8.47878158e-01
-9.75151956e-01 1.32489368e-01 -6.74882948e-01 -4.39859003e-01
-1.89400196e-01 -7.91495085e-01 1.01324153e+00 3.12584415e-02
-5.98430969e-02 -1.45392478e+00 2.75497884e-01 -3.50926816e-01
5.80356359e-01 2.50172526e-01 9.64816093e-01 -9.74053621e-01
-5.24831176e-01 -5.31338394e-01 -3.78298759e-01 8.27420413e-01
3.40356648e-01 -1.21161439e-01 -1.04224980e+00 -4.54423875e-01
-3.22411537e-01 -4.30129111e-01 1.21083343e+00 5.27848005e-01
1.20404267e+00 -4.06344235e-01 -1.41642079e-01 1.03232205e+00
1.27705383e+00 -4.61298764e-01 6.02672637e-01 1.05442472e-01
8.42423379e-01 3.25758487e-01 2.00499162e-01 2.55562991e-01
-2.03802902e-02 4.68216568e-01 -1.40996799e-01 -4.64759097e-02
1.92472890e-01 -2.34574929e-01 8.41964543e-01 4.72180843e-01
-2.78802723e-01 4.07234997e-01 -4.21069205e-01 5.00968218e-01
-2.08616877e+00 -1.00282586e+00 2.90283650e-01 2.28864670e+00
6.89026952e-01 -1.95779592e-01 -1.19838528e-02 -3.31048191e-01
7.68162727e-01 9.03802738e-02 -2.97941476e-01 -3.76502275e-01
-1.50231004e-01 7.16980159e-01 8.27982306e-01 6.29428744e-01
-1.65891504e+00 1.23640609e+00 6.94412136e+00 1.09189296e+00
-9.47234750e-01 7.31146932e-02 4.00419295e-01 1.59956858e-01
-1.61632985e-01 -6.38351068e-02 -8.68846059e-01 -1.48122564e-01
9.69974995e-01 -4.48326647e-01 4.67657596e-01 1.20412350e+00
-1.66360259e-01 3.75196993e-01 -1.39871001e+00 8.52443755e-01
1.21967427e-01 -1.32151198e+00 2.70898342e-01 4.27048713e-01
8.55594397e-01 1.65588498e-01 4.63286757e-01 4.48563337e-01
9.01217997e-01 -1.22066915e+00 4.91466850e-01 7.42244482e-01
7.17214942e-01 -1.28028345e+00 3.50373030e-01 -7.03042969e-02
-1.09616518e+00 2.25862473e-01 -1.16606998e+00 3.31378698e-01
-5.04596651e-01 6.25535786e-01 -8.33381951e-01 3.74017835e-01
5.57286859e-01 8.01615119e-01 -6.38100207e-01 7.69431591e-01
-2.25319311e-01 8.09042692e-01 -4.16111231e-01 3.26420128e-01
2.17431307e-01 -2.19524831e-01 3.36819321e-01 1.37072480e+00
2.40210786e-01 -2.95868844e-01 7.42352679e-02 1.09316516e+00
-2.37753764e-01 2.03723341e-01 -1.09748340e+00 1.10082023e-01
9.59684700e-02 1.46721184e+00 -7.80165434e-01 -1.68913141e-01
-3.71761918e-01 1.27415740e+00 8.70517492e-01 6.59013927e-01
-8.69493365e-01 -7.21401036e-01 8.38017106e-01 -3.39280218e-01
6.05562389e-01 -2.66878724e-01 4.99328703e-01 -1.49426973e+00
-1.42892495e-01 -6.20949388e-01 5.34842074e-01 -3.54541361e-01
-1.51270485e+00 5.11724412e-01 2.90464699e-01 -9.89357650e-01
-2.56189644e-01 -9.32578146e-01 -6.48816407e-01 9.19208825e-01
-1.20694435e+00 -1.33391082e+00 2.13533431e-01 1.02159679e+00
-9.58399009e-03 -3.52259010e-01 1.12850606e+00 -1.11730181e-01
-3.40331227e-01 9.19950247e-01 4.61638689e-01 5.02698779e-01
8.49693179e-01 -1.51771784e+00 5.33236325e-01 7.96368062e-01
4.92153943e-01 1.03997338e+00 3.64596307e-01 -2.76159883e-01
-1.27357721e+00 -1.55105662e+00 3.03130209e-01 -7.81004250e-01
9.40705955e-01 -6.64265275e-01 -9.18863833e-01 1.19475126e+00
-1.25594869e-01 8.08237672e-01 8.05674911e-01 1.76353976e-01
-1.15310252e+00 -9.13257524e-02 -9.85906184e-01 4.84555334e-01
8.32432806e-01 -9.87922788e-01 -5.74977279e-01 4.20221478e-01
7.31111884e-01 1.21308863e-02 -8.83337379e-01 2.48372689e-01
4.07338709e-01 -6.55762076e-01 1.07870138e+00 -1.02407253e+00
-2.92536497e-01 -1.95948839e-01 -4.51402873e-01 -1.58365405e+00
-8.05818796e-01 -7.97295690e-01 8.82623419e-02 9.85916495e-01
3.76172423e-01 -9.69338298e-01 4.42707479e-01 4.84108210e-01
1.25904292e-01 -2.78921217e-01 -9.19130921e-01 -1.20179069e+00
5.07490695e-01 -3.09032172e-01 3.67251873e-01 1.03560042e+00
-1.29860312e-01 6.05120370e-03 -2.25236639e-01 5.18564522e-01
1.02666926e+00 -1.45992532e-01 7.27586448e-01 -1.10738921e+00
-3.43773544e-01 -2.20134392e-01 -9.50153410e-01 -5.98299205e-01
9.50717509e-01 -1.40459681e+00 -3.24099451e-01 -8.49629641e-01
3.19301188e-01 -1.74291447e-01 -7.24393249e-01 6.27970695e-01
1.29906863e-01 2.80551761e-01 -1.88458964e-01 2.01021209e-02
-6.94178104e-01 5.96054256e-01 7.60101855e-01 2.58666985e-02
-4.47379611e-02 2.81602386e-02 -7.05375373e-01 9.00812924e-01
7.16702461e-01 -3.45763713e-01 -5.05499721e-01 7.82028735e-02
1.22511245e-01 -6.06110990e-01 6.54942214e-01 -1.03746355e+00
2.35582069e-02 -9.28123817e-02 6.80150807e-01 -3.21412534e-02
8.89070623e-04 -4.88597184e-01 -3.24043334e-01 1.25520781e-01
-7.95533299e-01 -1.77111819e-01 5.62474318e-02 8.61605287e-01
4.55314033e-02 5.10726757e-02 1.18521047e+00 2.74468124e-01
-4.43344831e-01 5.03946841e-01 -3.78276199e-01 1.75803512e-01
8.86103988e-01 1.66131660e-01 -1.36960134e-01 -4.52318698e-01
-8.50878179e-01 -3.94155651e-01 3.57751995e-01 4.28515673e-01
6.64759994e-01 -1.74512506e+00 -7.62666881e-01 5.21212757e-01
3.10533047e-01 -2.24794388e-01 -1.02405012e-01 7.86514819e-01
-5.11319757e-01 3.97341728e-01 -1.82006538e-01 -5.39653480e-01
-7.63545036e-01 5.89972258e-01 4.18895572e-01 -1.43076181e-01
-9.53799248e-01 7.69697845e-01 9.52947259e-01 -6.05371833e-01
1.04668990e-01 -7.73428500e-01 1.24343164e-01 -2.35136166e-01
8.26475084e-01 2.04998374e-01 9.52062979e-02 -6.17851138e-01
-2.75788248e-01 4.42530632e-01 -5.15231669e-01 -2.54564762e-01
1.27176178e+00 3.49688351e-01 -1.76172629e-01 3.68674278e-01
1.75415325e+00 -2.74934480e-03 -1.33995163e+00 -5.32715857e-01
1.29528090e-01 -3.07446808e-01 1.20418184e-02 -6.47003651e-02
-8.68622661e-01 6.57009184e-01 5.42027175e-01 -1.09252103e-01
6.26817882e-01 2.96359062e-01 1.27664089e-01 9.22819078e-01
3.44133705e-01 -9.70273972e-01 9.70215499e-02 6.34800017e-01
9.09303188e-01 -7.93279409e-01 -1.12972282e-01 -1.77918628e-01
-4.15587842e-01 1.18335748e+00 1.40827417e-01 -9.70754445e-01
1.07298148e+00 1.06682971e-01 -1.08378693e-01 -2.50751860e-02
-5.43411136e-01 -5.51015697e-02 6.48777306e-01 8.52202833e-01
2.13989571e-01 2.02451363e-01 2.44301051e-01 5.04806519e-01
-2.51494378e-01 -3.70754182e-01 3.53942096e-01 5.44979572e-01
-4.74329472e-01 -7.85914600e-01 -3.43745440e-01 4.24932778e-01
-6.14360034e-01 -1.75658748e-01 -9.61811319e-02 9.06443536e-01
-5.24700463e-01 2.20717981e-01 -1.77006721e-02 -1.47953331e-01
2.68612076e-02 3.21490824e-01 6.18991256e-01 -6.91910446e-01
1.34086609e-03 -1.08093135e-01 -2.98724741e-01 -7.74060071e-01
-1.53514475e-01 -8.15559924e-01 -1.30833089e+00 -1.36761367e-01
-2.51151085e-01 1.45042643e-01 3.42513412e-01 8.42574358e-01
2.35191435e-01 2.01203421e-01 8.33504021e-01 -9.65497017e-01
-1.12521684e+00 -7.43587077e-01 -1.06755209e+00 5.40470362e-01
5.11601329e-01 -6.72864974e-01 -8.47961128e-01 1.31630823e-01] | [8.948630332946777, 2.555530309677124] |
c7f48eb3-ad6c-4d89-8445-268a627019bb | cd-tools-condensed-detachment-and-structure | 2207.08453 | null | https://arxiv.org/abs/2207.08453v1 | https://arxiv.org/pdf/2207.08453v1.pdf | CD Tools -- Condensed Detachment and Structure Generating Theorem Proving (System Description) | CD Tools is a Prolog library for experimenting with condensed detachment in first-order ATP, which puts a recent formal view centered around proof structures into practice. From the viewpoint of first-order ATP, condensed detachment offers a setting that is relatively simple but with essential features and serious applications, making it attractive as a basis for developing and evaluating novel techniques. CD Tools includes specialized provers based on the enumeration of proof structures. We focus here on one of these, SGCD, which permits to blend goal- and axiom-driven proof search in particularly flexible ways. In purely goal-driven configurations it acts similarly to a prover of the clausal tableaux or connection method family. In blended configurations its performance is much stronger, close to state-of-the-art provers, while emitting relatively short proofs. Experiments show characteristics and application possibilities of the structure generating approach realized by that prover. For a historic problem often studied in ATP it produced a new proof that is much shorter than any known one. | ['Christoph Wernhard'] | 2022-07-18 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 1.60338625e-01 1.00376737e+00 -1.93686098e-01 2.39989936e-01
-5.97278059e-01 -9.28221166e-01 1.03230727e+00 3.83550644e-01
-8.13688114e-02 1.31856894e+00 -2.75506467e-01 -1.03650558e+00
-6.13402128e-01 -1.05015349e+00 -6.51132643e-01 -5.65110445e-01
-4.14022803e-01 9.16657865e-01 6.39741361e-01 -6.16713226e-01
4.68682349e-01 4.53043401e-01 -1.41220129e+00 1.11744396e-01
8.60921621e-01 3.83814692e-01 3.02371055e-01 7.66733110e-01
-1.88638434e-01 9.40461099e-01 -3.11355561e-01 -7.41248310e-01
8.33644122e-02 -6.22553408e-01 -1.36825943e+00 -3.96045953e-01
-1.18738478e-02 -4.90371808e-02 -1.35681123e-01 1.15587735e+00
2.97478978e-02 -6.57286167e-01 3.71495068e-01 -1.78658664e+00
-6.46814555e-02 1.25962520e+00 1.09138943e-01 -9.53633264e-02
9.01408434e-01 1.67791918e-02 1.13309014e+00 -2.47562423e-01
9.84717906e-01 1.29008234e+00 5.78215361e-01 6.56410098e-01
-1.50141728e+00 -1.71373382e-01 -2.79249042e-01 5.89225590e-01
-1.21608925e+00 -3.07183206e-01 5.49102783e-01 -1.81534976e-01
9.83983934e-01 7.46932447e-01 7.67281532e-01 8.14778149e-01
5.20854652e-01 6.67337656e-01 1.23582482e+00 -8.39618564e-01
2.06445813e-01 4.19901878e-01 1.02352537e-01 8.47887039e-01
7.05131769e-01 8.05132836e-02 -2.91148156e-01 -2.59669483e-01
5.61629951e-01 -6.89121366e-01 -2.84029663e-01 -7.19107866e-01
-1.58029068e+00 7.62123883e-01 -1.56750903e-01 6.95872664e-01
1.63393915e-01 1.86420605e-01 6.84213579e-01 3.81468832e-01
-3.31045598e-01 6.49018407e-01 -5.24713397e-01 -4.44294006e-01
-7.30752945e-01 9.45064723e-01 1.51355076e+00 8.99925768e-01
4.56870705e-01 -3.05008054e-01 -1.20720059e-01 -1.45124450e-01
3.03111017e-01 4.03942704e-01 -1.62964582e-01 -1.41384196e+00
4.10331160e-01 6.31055832e-01 1.92527294e-01 -6.59779429e-01
-2.52844453e-01 -1.08333774e-01 -4.48935419e-01 4.89594609e-01
5.28004885e-01 1.65606037e-01 -1.84987918e-01 1.69021416e+00
2.88837880e-01 -1.62574887e-01 4.39652950e-01 4.49536175e-01
5.47221243e-01 7.12615132e-01 -4.39246505e-01 -8.44869375e-01
1.08333528e+00 -7.18202174e-01 -7.29746580e-01 4.83550280e-01
9.53208804e-01 -3.88938576e-01 5.64284921e-01 1.11490393e+00
-1.66793692e+00 1.05515920e-01 -1.16193533e+00 -1.22385994e-01
-4.90493745e-01 -5.19397020e-01 9.47123051e-01 5.13434470e-01
-1.39302766e+00 8.08716595e-01 -3.61456513e-01 -2.80434132e-01
-7.64369741e-02 4.04911071e-01 -2.98268944e-01 -9.20250341e-02
-1.24350119e+00 1.21175957e+00 8.48866642e-01 6.58363402e-02
-7.96747684e-01 -5.79799771e-01 -1.01365674e+00 1.48194864e-01
8.49501014e-01 -7.53942788e-01 1.67313814e+00 -9.99218598e-02
-1.75869739e+00 8.25559020e-01 1.04427129e-01 -8.97413969e-01
9.02755976e-01 4.94174689e-01 -2.69081265e-01 1.88955903e-01
3.47611368e-01 5.27844906e-01 2.57011473e-01 -1.02627230e+00
-4.49665010e-01 2.93468814e-02 6.89946532e-01 -1.64615184e-01
4.86877143e-01 5.60676120e-02 2.54340798e-01 1.55148417e-01
-3.23595516e-02 -5.89573920e-01 -1.67674363e-01 -1.00663766e-01
-2.75029480e-01 -6.71889305e-01 5.55240631e-01 -3.61340269e-02
1.32475734e+00 -1.67762303e+00 4.65842068e-01 2.02304780e-01
2.83749282e-01 3.36853892e-01 2.82880783e-01 1.17655158e+00
-8.29298049e-02 2.77057290e-01 -1.36505663e-01 2.91658700e-01
8.20974290e-01 6.07474208e-01 -2.92476594e-01 2.98682719e-01
3.53734158e-02 9.24708903e-01 -1.25299466e+00 -1.06724215e+00
2.97214836e-01 -3.03114295e-01 -7.67011523e-01 -1.49348468e-01
-7.41352856e-01 -1.65421516e-01 -4.03839409e-01 3.93017620e-01
5.91698468e-01 -3.21342833e-02 7.07089305e-01 3.84964734e-01
-5.85143268e-01 6.05464339e-01 -1.35149992e+00 1.86416161e+00
-2.71567553e-01 2.53506303e-01 2.68453032e-01 -9.74692345e-01
6.88196778e-01 5.08543193e-01 1.85549095e-01 -4.60657328e-01
1.92601770e-01 4.94553715e-01 3.41050684e-01 -3.71964991e-01
3.36615205e-01 -3.53535920e-01 -1.20268613e-01 3.14712197e-01
6.59348369e-02 -6.70339584e-01 1.15341914e+00 5.52961051e-01
1.46366036e+00 4.70095366e-01 7.06542671e-01 -6.43997252e-01
1.01331878e+00 2.20196322e-01 5.37934840e-01 7.29248345e-01
2.08457634e-01 -5.29524535e-02 1.11506438e+00 -3.57867450e-01
-1.06712008e+00 -1.10363710e+00 -2.61799037e-01 2.31092706e-01
1.42395824e-01 -1.39646900e+00 -7.14840531e-01 -7.42651880e-01
-1.13639325e-01 7.78507173e-01 -2.68426776e-01 1.45106420e-01
-5.65843761e-01 1.00595141e-02 7.88093030e-01 8.60040188e-02
4.84327197e-01 -1.19749463e+00 -7.70280004e-01 3.72293890e-01
-2.16674298e-01 -8.94482791e-01 5.97354770e-01 3.52440983e-01
-7.12178469e-01 -1.55440319e+00 -1.83018953e-01 -6.17314219e-01
3.89050037e-01 -3.12758714e-01 1.07550347e+00 2.24339992e-01
1.21790729e-01 1.85969025e-02 -3.70955408e-01 -5.34543753e-01
-1.10878491e+00 4.50220257e-02 -9.34740156e-02 -7.81305790e-01
-2.05587059e-01 -7.64583349e-01 1.80877760e-01 -2.65529305e-02
-7.92862773e-01 1.33524850e-01 4.23996657e-01 8.40710402e-01
-1.95572060e-02 4.38765496e-01 3.51733193e-02 -8.61447811e-01
7.13799238e-01 -1.20224536e-03 -1.03898358e+00 2.30751052e-01
-7.86033750e-01 2.89000452e-01 9.16380286e-01 1.76879063e-01
-6.58408403e-01 -2.97522217e-01 -9.07856897e-02 2.76198953e-01
-6.97311386e-02 5.49374282e-01 -2.26053849e-01 -3.04957572e-02
6.09493136e-01 1.81980222e-01 6.17045499e-02 -1.07925281e-01
2.70440698e-01 1.19914785e-01 6.70987189e-01 -1.10412359e+00
1.05644262e+00 -9.21327770e-02 8.37830961e-01 -4.18168813e-01
-8.83501917e-02 2.04317775e-02 -3.34051460e-01 -6.05695099e-02
2.97333717e-01 -2.73297787e-01 -1.31228995e+00 1.17478788e-01
-1.46642065e+00 -5.78640521e-01 -5.94940305e-01 5.83878532e-03
-1.06411862e+00 6.05618954e-01 -5.55198252e-01 -9.85936642e-01
-2.16737352e-02 -1.06113672e+00 9.50288951e-01 -1.31796524e-01
-4.70908523e-01 -9.72288489e-01 4.42143947e-01 -5.88534772e-02
-1.93871826e-01 3.84813726e-01 1.13199270e+00 -5.29006720e-01
-6.29223943e-01 6.87963516e-03 -1.35086790e-01 2.59069085e-01
-2.77793437e-01 2.56786287e-01 -4.34756696e-01 5.62910475e-02
-2.53303021e-01 -1.42666861e-01 1.39107089e-02 -8.52259770e-02
6.03906929e-01 -5.85057616e-01 -5.28354645e-01 -1.41659006e-01
1.59732068e+00 1.62442587e-02 9.93075013e-01 6.01702631e-01
-1.44626662e-01 2.37198427e-01 8.66622388e-01 6.20302930e-02
2.98330635e-01 8.58764887e-01 4.82521176e-01 3.64000887e-01
2.53384471e-01 -2.69618660e-01 2.23528072e-01 4.68533695e-01
-2.52571166e-01 4.82512750e-02 -9.75364208e-01 5.80106556e-01
-2.13012195e+00 -1.34925818e+00 -6.39128745e-01 1.86306417e+00
1.26339412e+00 7.22162724e-01 3.17322075e-01 1.08610320e+00
3.81001770e-01 -1.31530717e-01 4.93323028e-01 -1.16733181e+00
1.56465545e-01 6.17832720e-01 1.84707165e-01 8.70609164e-01
-5.50142288e-01 7.51245379e-01 6.92006302e+00 7.05459833e-01
-5.08704364e-01 -2.00426131e-01 -5.28896213e-01 5.01784563e-01
-4.07314211e-01 5.27183473e-01 -5.44584095e-01 3.61433983e-01
9.88164663e-01 -6.61649227e-01 5.48783123e-01 8.36217999e-01
1.21897422e-01 -5.19046068e-01 -1.51644278e+00 4.36697453e-01
-3.09540898e-01 -1.80929375e+00 -1.30722642e-01 2.24574611e-01
2.91090876e-01 -5.18952012e-01 -5.96699536e-01 3.49679798e-01
3.06544811e-01 -8.18047464e-01 1.04059482e+00 1.61999553e-01
4.83425647e-01 -7.40720391e-01 9.68397915e-01 7.26147115e-01
-8.44999433e-01 -2.11068362e-01 -7.95851052e-02 -8.26729298e-01
1.64013997e-01 3.98695469e-01 -1.06569099e+00 1.20757186e+00
1.77564383e-01 4.70601439e-01 -2.04241917e-01 1.17166471e+00
-6.27588034e-01 2.97874749e-01 -4.62693244e-01 -4.48517799e-01
4.42536533e-01 -4.00727570e-01 7.59484708e-01 1.37052310e+00
4.12357636e-02 1.30833253e-01 -1.23924069e-01 1.03122735e+00
5.39909780e-01 -3.43723238e-01 -9.84296381e-01 1.91721648e-01
2.15045840e-01 1.12781858e+00 -6.84371114e-01 -4.22961086e-01
1.51126951e-01 3.82790446e-01 7.85098746e-02 -2.56335139e-01
-1.21420193e+00 -4.35115635e-01 4.48823832e-02 2.56175816e-01
2.22062245e-01 -3.75835776e-01 -6.62316084e-02 -9.84005809e-01
2.89450854e-01 -9.72946107e-01 1.32775515e-01 -9.84063089e-01
-6.35873854e-01 2.72056311e-01 6.76466405e-01 -7.66340733e-01
-7.87186921e-01 -7.75515854e-01 -5.89421630e-01 6.04061127e-01
-1.43983281e+00 -9.67956364e-01 2.94606745e-01 4.03193742e-01
3.02230030e-01 3.87993395e-01 1.23537326e+00 -1.17082382e-02
-1.19667351e-01 1.92231908e-01 -3.62147212e-01 -3.22432935e-01
1.80698872e-01 -1.74120581e+00 4.94507365e-02 1.08163369e+00
9.46351215e-02 6.38877809e-01 1.21199334e+00 -3.22681248e-01
-1.80092013e+00 -4.66157109e-01 1.64138150e+00 -3.46587598e-01
1.16180718e+00 -3.79939407e-01 -4.09685820e-01 6.01253510e-01
3.61098558e-01 -4.72597629e-01 -2.02379912e-01 2.07393959e-01
-7.06900284e-02 -1.81265980e-01 -1.18942237e+00 8.22785079e-01
1.29926646e+00 -3.23244363e-01 -1.17223179e+00 5.02445161e-01
3.56411248e-01 -6.44593656e-01 -7.86322832e-01 4.10041362e-01
4.36510026e-01 -1.21526706e+00 5.76150715e-01 -2.87188947e-01
5.74901342e-01 -5.80569088e-01 2.16618236e-02 -9.04168069e-01
-3.62833261e-01 -1.44822836e+00 -3.47280860e-01 1.31684291e+00
5.33570528e-01 -8.84672284e-01 3.81500840e-01 2.41532356e-01
-2.94362098e-01 -7.52819657e-01 -1.13344824e+00 -1.03518045e+00
2.45894402e-01 -3.41383696e-01 3.93466204e-01 6.70796275e-01
1.24917459e+00 4.88554925e-01 2.92001218e-01 -9.51907411e-02
4.70558643e-01 1.44716769e-01 9.14607465e-01 -1.37847245e+00
-4.34706002e-01 -6.14360154e-01 -6.13675594e-01 -5.41856647e-01
3.50216419e-01 -1.19438410e+00 7.06913322e-02 -1.70357013e+00
-1.45744875e-01 -4.29493636e-01 8.83735865e-02 6.65272951e-01
5.42725921e-01 5.79850227e-02 3.26737314e-01 -4.43243384e-01
-7.48097539e-01 -6.23382516e-02 1.23139155e+00 -9.54716131e-02
1.54836059e-01 -1.70382299e-02 -5.02778888e-01 7.51014292e-01
7.16272295e-01 -4.64212000e-01 -4.63643253e-01 3.09915185e-01
8.82144094e-01 3.48482937e-01 5.99213004e-01 -1.08807671e+00
2.38654971e-01 -2.98385948e-01 -5.74280858e-01 -5.14442742e-01
-6.77673444e-02 -9.27251816e-01 5.99131048e-01 6.94431782e-01
-7.76029229e-02 1.00241855e-01 2.17820868e-01 -6.02825247e-02
-9.54382569e-02 -7.58316934e-01 3.08010280e-01 -3.90839249e-01
-5.95357060e-01 -2.51371622e-01 -5.63738763e-01 -2.96272095e-02
1.15213799e+00 -3.51095378e-01 -4.30384398e-01 -2.02834071e-03
-5.95609426e-01 2.50763059e-01 7.06241548e-01 -2.38925308e-01
2.29478374e-01 -9.74070311e-01 -5.53033471e-01 -2.71142513e-01
-7.90268555e-02 3.16570252e-02 -3.26039821e-01 1.14451778e+00
-9.39183593e-01 7.99258173e-01 -3.23574156e-01 -4.59948838e-01
-1.31953347e+00 9.09066141e-01 1.81786299e-01 -7.72257388e-01
-9.08576429e-01 1.32699534e-01 -1.90663680e-01 -2.30824888e-01
-1.93070490e-02 -7.46127665e-01 2.06546545e-01 -4.00118440e-01
4.26664829e-01 4.05675352e-01 6.46726117e-02 1.67756692e-01
-6.65751994e-01 3.09054136e-01 3.43808442e-01 -2.75178134e-01
1.25087619e+00 2.29937658e-01 -7.43307650e-01 4.07183468e-01
4.61816370e-01 1.80202007e-01 -4.88074780e-01 5.59899867e-01
1.64873321e-02 4.02051620e-02 -4.04923886e-01 -1.02004242e+00
1.82906091e-02 4.77289349e-01 -4.11711216e-01 1.08726919e+00
8.10344994e-01 6.56624660e-02 2.98553348e-01 7.13653564e-01
1.01836479e+00 -7.76082456e-01 -6.18548393e-01 5.26209652e-01
8.73368859e-01 -4.20847476e-01 2.98629731e-01 -8.01133037e-01
-6.81315064e-02 1.36004579e+00 1.02775106e-02 -2.44881094e-01
1.90545563e-02 8.97120595e-01 -4.48014289e-01 -4.37822863e-02
-9.47202802e-01 -1.93738833e-01 -3.82163674e-01 7.97045052e-01
-1.03654422e-01 -3.45872417e-02 -1.14009333e+00 2.44620681e-01
-5.20461977e-01 5.46160758e-01 1.00261796e+00 1.28217781e+00
-3.71077389e-01 -1.85025716e+00 -4.90471065e-01 -2.96183527e-01
-2.49790251e-01 -1.16499767e-01 -5.35972357e-01 1.71125376e+00
2.96929598e-01 9.52753663e-01 -6.07892275e-01 1.54590867e-02
1.80268943e-01 1.84348121e-01 1.52626193e+00 -5.95598578e-01
-6.18911207e-01 -3.97407442e-01 9.82214689e-01 -5.86961508e-01
-5.38885236e-01 -6.78040028e-01 -1.33046782e+00 -7.76519537e-01
-3.70005786e-01 9.39332724e-01 4.93648976e-01 1.17769539e+00
-2.21956342e-01 4.19444948e-01 3.37134719e-01 -8.67214262e-01
-6.21878207e-01 -4.92398798e-01 -7.45276928e-01 -3.59172493e-01
1.49738982e-01 -7.04559922e-01 -8.87628719e-02 -4.32785675e-02] | [8.753703117370605, 6.891692638397217] |
f84238cc-dc93-4ba0-96f4-9b5c5dd8b81b | a-survey-on-facial-image-deblurring | 2302.05017 | null | https://arxiv.org/abs/2302.05017v2 | https://arxiv.org/pdf/2302.05017v2.pdf | A survey on facial image deblurring | When a facial image is blurred, it significantly affects high-level vision tasks such as face recognition. The purpose of facial image deblurring is to recover a clear image from a blurry input image, which can improve the recognition accuracy, etc. However, general deblurring methods do not perform well on facial images. Therefore, some face deblurring methods have been proposed to improve performance by adding semantic or structural information as specific priors according to the characteristics of the facial images. In this paper, we survey and summarize recently published methods for facial image deblurring, most of which are based on deep learning. First, we provide a brief introduction to the modeling of image blurring. Next, we summarize face deblurring methods into two categories: model-based methods and deep learning-based methods. Furthermore, we summarize the datasets, loss functions, and performance evaluation metrics commonly used in the neural network training process. We show the performance of classical methods on these datasets and metrics and provide a brief discussion on the differences between model-based and learning-based methods. Finally, we discuss the current challenges and possible future research directions. | ['Quan Zheng', 'Fanjiang Xu', 'Bingnan Wang'] | 2023-02-10 | null | null | null | null | ['deblurring'] | ['computer-vision'] | [ 1.55851662e-01 -4.37583089e-01 -8.16807970e-02 -4.93383706e-01
-2.35897258e-01 -4.66070212e-02 3.48048925e-01 -9.17922139e-01
-1.41368397e-02 7.65594661e-01 4.88095790e-01 2.44337022e-01
-2.86668800e-02 -1.18763879e-01 -5.12191415e-01 -1.05202353e+00
2.50273794e-01 -2.88999379e-01 -4.37211365e-01 2.42807552e-01
2.24934042e-01 6.14066184e-01 -1.64855576e+00 2.44059190e-01
7.47501433e-01 1.03431809e+00 4.34243772e-03 5.67783356e-01
2.06182122e-01 9.20001328e-01 -8.00377548e-01 -3.11895370e-01
-1.31402677e-02 -6.34714961e-01 -8.34934294e-01 3.80040735e-01
1.00164568e+00 -7.99072444e-01 -8.36963356e-01 1.39700925e+00
6.03479922e-01 6.24561161e-02 7.19660997e-01 -1.20955324e+00
-1.28999984e+00 1.22237958e-01 -7.61472464e-01 4.21814293e-01
1.80990081e-02 -3.44297551e-02 2.13531688e-01 -1.26610482e+00
2.36747324e-01 1.56643486e+00 7.37806141e-01 1.02550960e+00
-8.71914506e-01 -7.64441729e-01 -5.55344746e-02 8.80283296e-01
-1.38456857e+00 -1.04544520e+00 9.14106905e-01 -3.87410492e-01
3.70858490e-01 1.89794451e-01 2.08123490e-01 1.03855681e+00
1.05379999e-01 8.86173606e-01 1.09596288e+00 -2.94628441e-01
6.13005906e-02 -3.31108093e-01 2.17134923e-01 6.95934355e-01
2.97590166e-01 1.86001256e-01 -6.00614250e-01 -1.66019186e-01
7.64673948e-01 -4.17363793e-02 -7.36652613e-01 -5.59499860e-02
-5.96198857e-01 4.43629593e-01 2.85375834e-01 1.96939930e-01
-3.28063518e-01 2.28425011e-01 2.22979426e-01 1.57166079e-01
7.88594067e-01 1.64712314e-02 -3.14644724e-01 2.13650703e-01
-1.31692231e+00 1.30109534e-01 3.22046131e-01 6.41141832e-01
6.22853100e-01 3.00475419e-01 -3.21513206e-01 1.35131764e+00
4.05498683e-01 5.02360821e-01 5.72254956e-01 -1.09031665e+00
-1.71892419e-01 -1.73458546e-01 1.10178903e-01 -1.16818404e+00
-1.64006814e-01 -1.59445196e-01 -1.26602244e+00 2.19523668e-01
1.66597649e-01 -1.83736339e-01 -1.19031036e+00 1.62311614e+00
5.52453436e-02 7.47643650e-01 -6.19539842e-02 1.21268487e+00
1.07492864e+00 4.87529308e-01 -2.79008448e-01 -6.00574970e-01
1.39780891e+00 -1.16163862e+00 -1.24604213e+00 -1.66092977e-01
-9.59502682e-02 -1.03475964e+00 4.71753836e-01 2.84884363e-01
-1.19055200e+00 -7.23576128e-01 -8.64162385e-01 -1.39772937e-01
2.52575371e-02 5.10223210e-01 4.16909963e-01 8.30780268e-01
-1.38924563e+00 4.37561065e-01 -6.67562962e-01 -2.92683154e-01
8.86721730e-01 2.13843405e-01 -2.73240864e-01 -5.46470821e-01
-1.21765268e+00 1.13200092e+00 -2.72744358e-01 5.62966764e-01
-1.19580293e+00 -8.09702158e-01 -9.65462565e-01 1.25810668e-01
-1.89310014e-01 -7.07245886e-01 1.36855590e+00 -1.14028168e+00
-1.54843116e+00 8.84568930e-01 -6.69295669e-01 -1.85472220e-01
3.77660692e-01 -5.05503833e-01 -5.76760352e-01 2.03383401e-01
-3.77381593e-01 6.91902578e-01 1.63865888e+00 -1.26696265e+00
-2.18917623e-01 -4.07760113e-01 -2.88616508e-01 -3.32324058e-02
-3.32048744e-01 3.62872094e-01 -4.17500466e-01 -8.47200274e-01
-2.34474868e-01 -6.33876622e-01 2.07690656e-01 1.73128992e-01
-4.27549966e-02 8.34220834e-03 1.31979895e+00 -9.74417150e-01
1.18756413e+00 -2.15316868e+00 8.65766704e-02 -5.41816294e-01
2.85983175e-01 7.59935677e-01 -4.00600493e-01 7.76413381e-02
-5.21340072e-01 -1.73636615e-01 -2.40693361e-01 -5.23463249e-01
-4.21753854e-01 -2.33649034e-02 -3.87396812e-01 9.52716708e-01
4.16707098e-02 7.95980692e-01 -6.69407189e-01 -1.37231886e-01
2.17787981e-01 8.19983721e-01 -4.04534191e-01 3.87947947e-01
1.59086406e-01 4.78186816e-01 1.39765844e-01 7.44459689e-01
1.42463231e+00 -8.15083086e-02 -3.11818391e-01 -6.45695210e-01
1.17396668e-01 -2.77582169e-01 -5.25094628e-01 1.37629616e+00
-4.93856609e-01 1.12357891e+00 5.01632988e-01 -8.55651200e-01
6.01394951e-01 3.85179847e-01 4.63043272e-01 -8.14895779e-02
1.85161218e-01 6.61496744e-02 -2.97631502e-01 -9.17873859e-01
3.82877290e-01 -1.83522806e-01 7.32812941e-01 4.23784018e-01
-1.04628794e-01 1.34976164e-01 -2.77757794e-01 -2.55816966e-01
7.62491286e-01 -1.16792284e-01 1.40868530e-01 -1.81802809e-01
9.04374003e-01 -6.25984490e-01 3.65110248e-01 3.40773791e-01
-5.62235892e-01 1.00663412e+00 1.70996010e-01 -6.58753037e-01
-7.06602931e-01 -5.16917169e-01 -3.21798086e-01 7.43663430e-01
3.18265945e-01 1.94322970e-02 -1.27457011e+00 -5.93931556e-01
-2.35193465e-02 3.53265494e-01 -8.33182752e-01 -5.63005686e-01
-4.08610642e-01 -1.25038326e+00 7.41758764e-01 2.38537982e-01
9.38087702e-01 -8.28146100e-01 -1.36896849e-01 -4.58251894e-01
-4.72630620e-01 -1.05876875e+00 -9.69464421e-01 -6.81530356e-01
-8.56008530e-01 -1.01623261e+00 -1.35290778e+00 -1.01948237e+00
1.02795291e+00 8.54824901e-01 7.06655324e-01 3.11205059e-01
-3.41049373e-01 3.64920199e-01 -7.08959103e-02 6.19106293e-02
-3.47216278e-01 -5.14491081e-01 2.65759796e-01 4.38936383e-01
4.65407252e-01 -3.70672792e-01 -7.12109327e-01 3.48858297e-01
-1.05148065e+00 -1.30466238e-01 5.29231369e-01 1.06171989e+00
-9.63137001e-02 1.55751616e-01 5.23548245e-01 -3.58350486e-01
7.36726224e-01 -3.47900718e-01 -3.66635472e-01 2.22451821e-01
-5.86832762e-01 4.29916941e-02 2.09558755e-01 -4.98825699e-01
-1.40218699e+00 -2.99588740e-01 -1.06192000e-01 -8.72812212e-01
-2.30026320e-01 2.78132886e-01 -3.41713876e-01 -4.04744655e-01
5.26583314e-01 4.17627454e-01 4.42086488e-01 -8.05940509e-01
3.39906961e-01 1.02090442e+00 5.54871142e-01 -7.64853507e-02
5.06690502e-01 5.30226827e-01 -1.54307097e-01 -1.18222868e+00
-9.25578892e-01 -2.66389161e-01 -5.25060236e-01 -3.24871123e-01
6.27604425e-01 -9.20129716e-01 -5.62349319e-01 1.38719678e+00
-1.60176909e+00 -1.43230841e-01 3.73963952e-01 4.18676913e-01
-5.05245447e-01 7.46903062e-01 -9.26839650e-01 -5.43624699e-01
-3.43627155e-01 -1.27512670e+00 9.36289847e-01 5.36440969e-01
1.85048938e-01 -1.03930056e+00 9.22183320e-03 5.42074442e-01
7.86632299e-01 -3.36531430e-01 6.50813282e-01 2.40818299e-02
-3.40950996e-01 -1.52073139e-02 -6.50180042e-01 8.48796725e-01
8.21047366e-01 -2.82286942e-01 -1.41766477e+00 -5.28848290e-01
3.83955866e-01 -1.01684779e-01 1.14015222e+00 8.68290007e-01
1.53593409e+00 -6.43483222e-01 -3.13706100e-01 8.71884227e-01
1.08316469e+00 -1.06563149e-02 9.23642039e-01 -8.23818147e-02
7.14258015e-01 5.32883048e-01 2.36326128e-01 1.97959453e-01
9.37757567e-02 7.19562173e-01 5.08980095e-01 -7.79843852e-02
-7.62146533e-01 1.27588585e-01 5.66871226e-01 4.76751208e-01
-9.68722105e-02 -8.56259614e-02 -2.23330781e-01 4.34849948e-01
-1.77259588e+00 -1.04204679e+00 7.49347508e-02 1.98959029e+00
1.11794460e+00 -8.23627949e-01 -2.89771050e-01 -2.60062814e-01
1.23817408e+00 4.96323109e-01 -6.58652782e-01 1.47445465e-03
-1.54313505e-01 -6.83522224e-02 3.01127017e-01 7.84348488e-01
-1.24403787e+00 1.17616451e+00 7.33111715e+00 6.51698053e-01
-1.28254592e+00 3.09923768e-01 8.81318152e-01 -1.08690120e-01
1.68875858e-01 -3.86411846e-01 -7.84199893e-01 6.22293770e-01
7.50419497e-01 -2.98559833e-02 7.05265880e-01 7.44888484e-01
5.99498034e-01 -3.54531072e-02 -8.87246013e-01 1.57776189e+00
7.01243699e-01 -1.17207587e+00 8.69456455e-02 -1.11901067e-01
9.53501701e-01 -2.53197789e-01 3.65001440e-01 -1.42403826e-01
-1.79181337e-01 -1.34424615e+00 3.93323302e-01 7.04754531e-01
9.84303653e-01 -4.62197900e-01 8.39680970e-01 -2.63270307e-02
-5.43329418e-01 1.48930445e-01 -6.96745217e-01 -5.72834350e-02
-6.60414714e-03 8.70575845e-01 -2.42803663e-01 2.82741427e-01
8.05746078e-01 1.10675335e+00 -3.91237348e-01 1.19524145e+00
-2.51251966e-01 4.50293750e-01 3.31594080e-01 4.82574433e-01
-2.00741202e-01 -1.20532945e-01 5.55657744e-01 1.19222963e+00
4.14276123e-01 1.92453265e-01 -6.18081391e-01 8.59044433e-01
-3.50369424e-01 -4.23029989e-01 -2.25954950e-01 2.27304742e-01
2.51850814e-01 1.20518911e+00 -2.51534022e-03 -3.28135759e-01
-4.92493451e-01 1.34736347e+00 -5.51336817e-02 6.59758151e-01
-9.41692829e-01 -2.98282683e-01 1.41447031e+00 -1.34523228e-01
2.57955432e-01 -1.19849369e-01 -8.52639452e-02 -1.37439477e+00
-1.32415220e-01 -9.22279060e-01 7.02875033e-02 -1.05187404e+00
-1.31608033e+00 5.89340270e-01 1.03960752e-01 -9.48472083e-01
1.27776638e-01 -6.29162729e-01 -4.35026795e-01 1.15024972e+00
-2.04658556e+00 -9.33177710e-01 -7.36701667e-01 3.60685587e-01
9.36740339e-01 -2.24850416e-01 2.99756646e-01 4.43858385e-01
-7.60641098e-01 5.19515455e-01 1.48754522e-01 8.63212571e-02
1.17008150e+00 -6.53784513e-01 2.94712424e-01 1.09262776e+00
-3.25892478e-01 6.78382695e-01 8.11696768e-01 -4.80048656e-01
-1.18721247e+00 -1.29095113e+00 6.88642502e-01 -9.47571769e-02
2.21677527e-01 1.94682956e-01 -1.20076573e+00 4.49349433e-01
6.11517191e-01 3.84215921e-01 4.10403371e-01 -3.58726501e-01
-4.33992416e-01 -2.99395889e-01 -1.26813424e+00 4.85103160e-01
9.08751428e-01 -4.10878360e-01 -5.73849976e-01 1.80748150e-01
3.29913229e-01 -3.19604009e-01 -3.67314816e-01 4.66319203e-01
6.38426244e-01 -1.03755057e+00 9.93988693e-01 -6.94010317e-01
5.81926644e-01 -3.26768279e-01 2.27738246e-01 -1.66587603e+00
-6.59879446e-01 -8.22600365e-01 -4.63540703e-01 1.06984699e+00
-3.65102947e-01 -5.97629726e-01 7.21085727e-01 5.39880216e-01
-4.03706394e-02 -3.52269977e-01 -6.27961874e-01 -6.73470438e-01
7.27365073e-03 6.20459467e-02 6.37220919e-01 8.75611901e-01
-2.65641302e-01 1.04603879e-01 -9.07428861e-01 3.83265436e-01
9.42584693e-01 5.04635870e-02 3.88667822e-01 -8.26053262e-01
2.83950299e-01 -6.19282961e-01 -2.71920741e-01 -1.43432105e+00
5.03859222e-01 -5.29360652e-01 1.32069185e-01 -1.27690852e+00
4.74747062e-01 9.56529900e-02 -1.46274477e-01 4.30819571e-01
-4.39523876e-01 5.16637266e-01 -2.50865221e-01 5.33387482e-01
-6.95324764e-02 7.29134738e-01 1.40623975e+00 -3.22069496e-01
1.60140097e-01 7.53201172e-02 -8.52238357e-01 7.56653190e-01
7.16254771e-01 -1.39664561e-01 -2.61100382e-01 -8.82547736e-01
-3.34264189e-01 -1.81674466e-01 4.60696280e-01 -7.49022663e-01
3.68544489e-01 -3.42115164e-01 5.89615464e-01 -3.31975013e-01
3.97951782e-01 -6.97475910e-01 2.18146760e-03 3.40199083e-01
-3.70585680e-01 -2.49011114e-01 2.61986434e-01 4.82446343e-01
-4.37050819e-01 -3.46050352e-01 1.51963043e+00 1.91783626e-02
-5.80666959e-01 4.96461123e-01 -4.55630332e-01 -3.56863678e-01
7.87086189e-01 -5.53613668e-03 -5.11461020e-01 -6.64308369e-01
-5.60306787e-01 -2.06410646e-01 4.18328792e-01 6.36593461e-01
9.09849823e-01 -1.33522308e+00 -8.44587684e-01 4.59610462e-01
-2.96760142e-01 -4.21782196e-01 5.73323488e-01 9.86736953e-01
-4.95939016e-01 5.47421455e-01 -4.42919970e-01 -3.48665655e-01
-1.75816524e+00 7.41045952e-01 6.72124445e-01 5.88239431e-01
-1.06830768e-01 1.09563100e+00 6.50694370e-01 2.31679276e-01
2.95189887e-01 -2.11092442e-01 -3.89753222e-01 -1.89628646e-01
1.23507643e+00 5.76717317e-01 2.80976862e-01 -8.95663440e-01
-4.03748423e-01 7.03679442e-01 -1.56470373e-01 2.46892676e-01
1.12817454e+00 -5.53168833e-01 -6.88509405e-01 -3.06061625e-01
1.39202833e+00 -1.09547161e-01 -1.46884549e+00 -3.52620304e-01
-2.98898160e-01 -8.47687840e-01 4.67525303e-01 -6.68172777e-01
-1.72498214e+00 1.04869413e+00 6.93161130e-01 -1.82417884e-01
1.63137949e+00 -3.14775929e-02 7.12600827e-01 -1.36184022e-01
-4.41535888e-03 -6.28847480e-01 3.09795529e-01 4.70856935e-01
1.17079306e+00 -1.08520925e+00 8.64016339e-02 -4.33400512e-01
-2.91966885e-01 1.24937308e+00 5.28586149e-01 6.73408657e-02
1.00331378e+00 8.81273970e-02 1.37439236e-01 8.98042843e-02
-6.66707337e-01 -8.62170663e-03 3.26973319e-01 7.25390434e-01
3.28086138e-01 -4.56611603e-01 -6.41253144e-02 4.19930786e-01
1.27505720e-01 3.57634455e-01 6.15314722e-01 3.82873535e-01
-4.92767006e-01 -8.19114089e-01 -8.02856207e-01 6.27653375e-02
-5.72248280e-01 -1.59856156e-01 -2.86636472e-01 4.05111983e-02
-2.40674019e-02 1.07956350e+00 3.65391225e-02 -2.31921092e-01
-1.26589566e-01 -9.90019441e-02 8.65994751e-01 -3.83667558e-01
2.58605629e-01 -7.87232965e-02 -3.79263937e-01 -3.16513211e-01
-6.80889666e-01 -6.13137424e-01 -5.19075215e-01 -5.69263816e-01
-3.88590395e-01 -1.71151944e-02 5.49829483e-01 9.35238123e-01
5.64049602e-01 2.27534980e-01 6.58054709e-01 -1.08539796e+00
-5.43132663e-01 -1.30454075e+00 -5.01275837e-01 6.68512434e-02
1.00932038e+00 -7.68602610e-01 -6.09557867e-01 7.19236076e-01] | [11.643357276916504, -2.6472318172454834] |
bb8180f0-309a-4165-a590-d7d571bb618d | agmi-attention-guided-multi-omics-integration | 2112.08366 | null | https://arxiv.org/abs/2112.08366v2 | https://arxiv.org/pdf/2112.08366v2.pdf | AGMI: Attention-Guided Multi-omics Integration for Drug Response Prediction with Graph Neural Networks | Accurate drug response prediction (DRP) is a crucial yet challenging task in precision medicine. This paper presents a novel Attention-Guided Multi-omics Integration (AGMI) approach for DRP, which first constructs a Multi-edge Graph (MeG) for each cell line, and then aggregates multi-omics features to predict drug response using a novel structure, called Graph edge-aware Network (GeNet). For the first time, our AGMI approach explores gene constraint based multi-omics integration for DRP with the whole-genome using GNNs. Empirical experiments on the CCLE and GDSC datasets show that our AGMI largely outperforms state-of-the-art DRP methods by 8.3%--34.2% on four metrics. Our data and code are available at https://github.com/yivan-WYYGDSG/AGMI. | ['Jian Wu', 'Ji Cao', 'Danny Z. Chen', 'Minshan Lai', 'Yufeng Xie', 'Ruiwei Feng'] | 2021-12-15 | null | null | null | null | ['drug-response-prediction'] | ['medical'] | [ 7.85357431e-02 1.16598804e-03 -6.76285625e-01 -6.89152302e-03
-6.47128880e-01 -2.97364622e-01 2.59778172e-01 6.26734078e-01
3.03119212e-01 9.53263402e-01 2.55251497e-01 -3.97120714e-01
-5.87052643e-01 -1.07912040e+00 -4.25518304e-01 -6.84375763e-01
7.07760230e-02 8.00275803e-01 3.44885588e-02 -1.70045540e-01
2.24164680e-01 4.33493584e-01 -5.91559470e-01 2.70686984e-01
1.19002891e+00 4.77867872e-01 -1.43464491e-01 4.85622853e-01
7.23301768e-02 4.15798277e-01 1.20710939e-01 -2.84354866e-01
-2.45992355e-02 -5.47289789e-01 -6.45060301e-01 -4.99332160e-01
-1.57491028e-01 4.08934236e-01 -5.11433125e-01 9.67466474e-01
8.47131014e-01 -2.09168836e-01 6.42784178e-01 -1.19388759e+00
-8.33876729e-01 5.19567788e-01 -7.70278573e-01 6.96068406e-02
5.15409827e-01 1.92344621e-01 1.16114187e+00 -7.06298828e-01
1.03113270e+00 1.17753100e+00 5.28282046e-01 3.92834306e-01
-1.32285821e+00 -5.00332296e-01 -9.18980241e-02 2.40170121e-01
-1.64546442e+00 -1.23094104e-01 3.51512939e-01 -5.26674986e-01
1.29433703e+00 5.02024114e-01 6.29717886e-01 7.41946161e-01
6.73977673e-01 5.72504818e-01 8.81488919e-01 -3.63093875e-02
2.72794127e-01 -6.22997046e-01 4.32905912e-01 9.12051320e-01
5.46189070e-01 -1.85667500e-02 -3.90724182e-01 -7.24420905e-01
2.82611281e-01 4.12704051e-01 -4.85437870e-01 2.09686160e-01
-1.12280798e+00 7.29772270e-01 3.20312411e-01 2.77190596e-01
-5.64435780e-01 3.31429951e-02 2.11991683e-01 1.44813418e-01
8.09722424e-01 6.46586299e-01 -6.89450324e-01 1.08507916e-01
-4.93473202e-01 -1.00302488e-01 9.33158755e-01 8.78820539e-01
4.56187159e-01 -2.71109879e-01 -3.47507715e-01 7.68964171e-01
1.64938867e-01 2.92611271e-01 2.85746068e-01 -3.42336386e-01
1.00954331e-01 1.03906429e+00 -3.32034349e-01 -1.20891726e+00
-8.89432609e-01 -5.41939914e-01 -1.10436833e+00 -5.81028938e-01
1.22358665e-01 -1.31086156e-01 -8.86405468e-01 1.40581918e+00
7.44693100e-01 5.65977216e-01 -1.60369605e-01 6.67758703e-01
1.11131179e+00 5.12674630e-01 3.58598202e-01 -2.81060308e-01
1.41555095e+00 -8.99220228e-01 -6.03883088e-01 1.48740634e-01
9.42760885e-01 -2.67055005e-01 6.01060629e-01 3.30750108e-01
-7.74390876e-01 -6.98575191e-03 -5.51059783e-01 -3.69535126e-02
-4.11201179e-01 -1.47708997e-01 6.75169110e-01 2.71910757e-01
-8.46271396e-01 6.51312649e-01 -6.47786081e-01 -4.78199393e-01
6.36435330e-01 4.50426251e-01 -5.81787467e-01 -6.22717857e-01
-1.07389367e+00 6.18781567e-01 3.50363165e-01 -3.49232018e-01
-7.66114354e-01 -1.34231222e+00 -6.05666280e-01 7.91853443e-02
4.40002263e-01 -1.29483461e+00 3.50743115e-01 -1.59233242e-01
-1.45314682e+00 5.44407070e-01 -3.27028275e-01 -1.08186930e-01
8.72710161e-03 9.35559943e-02 -5.40646791e-01 3.27592403e-01
-1.70447960e-01 3.76213640e-01 -4.14509214e-02 -6.12698019e-01
-1.70582920e-01 -8.54093075e-01 -3.43497127e-01 2.86647640e-02
-3.64171639e-02 5.82905933e-02 -4.25655276e-01 -4.30558860e-01
7.69392475e-02 -1.02945578e+00 -4.82438505e-01 -4.42434579e-01
-8.90705526e-01 -2.34551594e-01 1.93327129e-01 -7.56074131e-01
1.50735331e+00 -1.68874812e+00 4.86467183e-01 3.35335284e-01
7.08689570e-01 1.13887697e-01 -6.08839869e-01 9.06513751e-01
-3.90261024e-01 4.15056765e-01 -1.80903062e-01 4.75028940e-02
-2.28627264e-01 -9.51699838e-02 2.82097906e-01 7.03794956e-01
2.08206192e-01 1.37955511e+00 -1.00560439e+00 -1.37055069e-01
-2.00647131e-01 5.77876449e-01 -6.60055459e-01 -2.37490594e-01
-5.29338419e-01 6.57683730e-01 -8.28116834e-01 1.22772002e+00
7.63739824e-01 -7.98879087e-01 7.39262283e-01 3.02071646e-02
8.61661136e-02 -4.08727191e-02 -5.42247474e-01 1.66709280e+00
2.36033142e-01 -9.61622670e-02 -5.38564980e-01 -9.55370784e-01
7.30393171e-01 2.82086760e-01 9.67854857e-01 -5.49096942e-01
1.99588925e-01 3.51066232e-01 3.05893183e-01 -3.38542700e-01
-8.33740681e-02 1.66285440e-01 1.22038022e-01 1.60672978e-01
-1.40190646e-01 3.86706531e-01 1.18824057e-01 3.34645152e-01
1.69608676e+00 -1.76480845e-01 6.37736678e-01 -6.56903803e-01
4.81201351e-01 1.30801857e-01 1.00089157e+00 3.30245018e-01
7.55809098e-02 5.83230078e-01 9.14366603e-01 -3.48944724e-01
-7.07751155e-01 -4.25114512e-01 -1.79597557e-01 5.63246131e-01
-1.66594073e-01 -6.51225090e-01 -3.01362276e-01 -5.69188118e-01
5.22529900e-01 4.23418045e-01 -7.21799016e-01 -1.31721020e-01
-1.48235917e-01 -1.49014664e+00 3.17666203e-01 7.54681081e-02
-4.32141451e-03 -3.26429009e-01 4.04195189e-01 3.06310177e-01
3.69628519e-02 -8.04223657e-01 -4.46869463e-01 -1.06830269e-01
-7.87361324e-01 -1.45614922e+00 -4.86749530e-01 -2.10363537e-01
8.28760862e-01 1.09452039e-01 1.04544103e+00 3.22751433e-01
-7.01988578e-01 -3.40856835e-02 -4.23329085e-01 -4.42577809e-01
-1.14673324e-01 1.08738787e-01 1.49570024e-02 -1.11105427e-01
6.61910117e-01 -6.74424827e-01 -9.07661796e-01 1.34882867e-01
-5.23442566e-01 2.11191669e-01 4.79915559e-01 7.80955374e-01
1.20319474e+00 -2.94350713e-01 9.78723168e-01 -1.41804624e+00
5.37249982e-01 -1.08258808e+00 -6.32266402e-01 4.19626594e-01
-9.31069553e-01 -1.58177927e-01 5.95662951e-01 -1.55938059e-01
-3.57763171e-01 3.47179025e-02 -3.70934576e-01 -1.05731033e-01
-6.91449642e-02 9.68136668e-01 -1.98775753e-01 -3.13945621e-01
5.13080239e-01 -2.48894673e-02 -3.41358222e-02 -5.35714328e-01
2.30313912e-01 5.83020031e-01 -1.88350499e-01 -2.77254969e-01
1.62624851e-01 3.63667667e-01 7.19745874e-01 -4.78164762e-01
-6.27090633e-01 -6.48380101e-01 -3.37223083e-01 2.60728836e-01
8.46596241e-01 -9.85869527e-01 -1.08442056e+00 3.00210983e-01
-6.44627631e-01 -2.90988237e-01 2.03564912e-01 5.93821108e-01
-2.43261725e-01 5.00259221e-01 -9.57140267e-01 -4.07673270e-01
-7.84174919e-01 -9.46016490e-01 8.56871903e-01 1.60971470e-02
-1.46199122e-01 -1.07368803e+00 5.33333302e-01 3.97158116e-01
1.73619613e-01 7.17753410e-01 1.27497590e+00 -9.93899643e-01
-4.52003360e-01 -1.87380239e-01 -3.70234400e-01 -5.01450479e-01
2.64111012e-01 3.84431928e-02 -6.84891045e-01 -1.05808392e-01
-5.47309339e-01 1.00563228e-01 9.83579040e-01 7.50523567e-01
1.36208391e+00 -1.61335707e-01 -9.38178837e-01 8.03933799e-01
1.89772904e+00 2.36038923e-01 8.23990226e-01 -3.19734663e-02
8.05182874e-01 1.40563995e-01 4.49177265e-01 5.87127745e-01
4.60485846e-01 5.17910957e-01 3.32274228e-01 -1.43558189e-01
2.16215085e-02 -1.66206107e-01 -6.53292239e-02 8.56244743e-01
-7.45895728e-02 -7.89452612e-01 -1.14846122e+00 4.89862144e-01
-2.15976763e+00 -4.76661980e-01 -5.90122044e-01 1.84906924e+00
8.84590626e-01 -5.88368714e-01 1.77373528e-01 -4.69292134e-01
5.44895768e-01 -3.86176646e-01 -8.69255662e-01 -2.95549542e-01
-3.32806647e-01 3.56966347e-01 4.78001684e-01 4.68715757e-01
-6.77572966e-01 7.58198023e-01 5.38504362e+00 9.26886559e-01
-5.96423805e-01 5.75279184e-02 1.03741658e+00 -2.42822282e-02
-4.26182806e-01 8.75106007e-02 -8.48938465e-01 5.64284146e-01
1.26325905e+00 -4.99408871e-01 2.22132951e-01 3.39164346e-01
4.60957795e-01 -1.89573467e-01 -9.11149144e-01 7.53205657e-01
-2.26501152e-01 -1.67893302e+00 8.06121007e-02 4.44361180e-01
8.49662602e-01 5.62080562e-01 -2.23004028e-01 -1.26260504e-01
6.20642960e-01 -1.12421370e+00 -5.84934652e-01 8.46292794e-01
8.96382451e-01 -8.25688720e-01 6.18568778e-01 1.95955694e-01
-9.51357365e-01 -2.67293267e-02 -6.13089025e-01 3.73247802e-01
-6.81686550e-02 1.34780574e+00 -7.46770144e-01 1.43699956e+00
2.02798977e-01 1.16914666e+00 -3.10999185e-01 1.20612454e+00
-3.57301831e-02 6.52926803e-01 -1.58728361e-01 4.02070992e-02
-1.28759354e-01 -3.66871119e-01 5.33034265e-01 1.15773535e+00
5.81918776e-01 6.62092030e-01 2.01956883e-01 7.16268420e-01
-2.73433059e-01 4.82798636e-01 -4.71343011e-01 -5.63908696e-01
4.44693089e-01 1.47829223e+00 -4.68407094e-01 -1.88167423e-01
-2.82794148e-01 6.89715505e-01 3.63762110e-01 1.86975881e-01
-7.10179687e-01 -2.84980774e-01 9.53917444e-01 -3.88312414e-02
-7.75689706e-02 2.20995560e-01 -2.37188041e-01 -1.28541958e+00
-6.04528666e-01 -9.75868046e-01 9.17611837e-01 -6.78669274e-01
-1.57121336e+00 3.59580636e-01 -4.26385343e-01 -9.87884939e-01
2.11944848e-01 -5.90795636e-01 -3.49522859e-01 9.68604505e-01
-1.73686099e+00 -1.02455020e+00 -1.51664913e-01 2.98039943e-01
4.18668650e-02 -2.61648409e-02 1.14342713e+00 3.67876589e-01
-1.09973729e+00 4.35798526e-01 4.60673898e-01 -2.71383345e-01
4.78382200e-01 -8.93313527e-01 3.83385032e-01 4.13896948e-01
-3.68912697e-01 5.25728345e-01 3.66374403e-01 -1.22270131e+00
-2.05966163e+00 -1.41770089e+00 9.22495604e-01 -3.10280621e-01
6.94025338e-01 -1.20313242e-01 -9.54371870e-01 5.58076203e-01
1.43524736e-01 1.84536334e-02 1.39589882e+00 2.26221696e-01
-3.12880099e-01 1.72969460e-01 -1.21977174e+00 5.50370514e-01
1.33722901e+00 -1.06165037e-01 1.42404333e-01 9.32176411e-01
8.27916980e-01 -2.96205342e-01 -1.64363289e+00 4.61729854e-01
3.25398058e-01 -3.31455976e-01 9.71176565e-01 -1.15196979e+00
5.35908818e-01 -4.06839043e-01 -4.58562523e-02 -1.57317173e+00
-9.78554070e-01 -5.91996074e-01 -2.13143885e-01 6.05538368e-01
6.56343102e-01 -1.14646947e+00 6.56657398e-01 3.94999027e-01
-2.84863144e-01 -1.17552650e+00 -6.90511107e-01 -6.82555676e-01
1.92433700e-01 6.47576228e-02 8.09429288e-01 1.45658028e+00
5.53771794e-01 4.19561863e-01 -2.30052322e-01 1.72389552e-01
4.23905730e-01 2.95196533e-01 4.51129407e-01 -1.41827321e+00
-5.88806808e-01 -3.68916452e-01 -5.44436455e-01 -3.62159550e-01
-7.12015806e-03 -1.49497104e+00 -7.08478451e-01 -1.82765162e+00
7.73357987e-01 -1.94828317e-01 -4.77817506e-01 7.76902258e-01
-3.81841123e-01 7.59696960e-02 -1.92660540e-01 -3.41294892e-02
-5.12564898e-01 4.76029485e-01 1.28542006e+00 -2.89769825e-02
-1.06734984e-01 -3.90890449e-01 -1.05434966e+00 2.26986378e-01
1.21770298e+00 -5.66431165e-01 -3.49260479e-01 -1.97734132e-01
4.82319504e-01 2.62337238e-01 1.12086274e-01 -4.85456735e-01
3.39964837e-01 -4.71340239e-01 4.29168403e-01 -4.44863141e-01
1.32349744e-01 -3.17508280e-01 1.04682946e+00 8.68816793e-01
-1.87010821e-02 4.06234562e-02 3.44976395e-01 8.30147326e-01
7.42642730e-02 2.47195601e-01 4.88293469e-01 -1.12741813e-01
-1.64784566e-01 7.74926901e-01 3.08247674e-02 -6.56482875e-02
9.57508206e-01 1.34011611e-01 -1.14350176e+00 1.45164415e-01
-7.06335783e-01 5.66116691e-01 2.58869618e-01 -3.31977904e-02
6.64808035e-01 -1.14427602e+00 -1.28514814e+00 -6.95322976e-02
3.02664459e-01 -3.51619095e-01 5.56721807e-01 1.39145720e+00
-4.57137972e-01 6.63611472e-01 1.08839467e-01 -3.28867108e-01
-1.37418532e+00 7.38287985e-01 3.50140631e-01 -5.80883801e-01
-6.79789186e-01 7.13772833e-01 3.74506265e-01 -4.10127968e-01
-4.61230129e-01 1.59056738e-01 -1.69279009e-01 -1.39148086e-01
4.18064952e-01 4.05277550e-01 1.12944208e-01 -2.68796474e-01
-5.60270548e-01 5.58121860e-01 -1.53106913e-01 6.96301341e-01
1.68885660e+00 1.86197579e-01 -7.23883152e-01 -3.62249650e-03
1.20626605e+00 -4.79893237e-02 -3.53053689e-01 -1.41900480e-01
1.38647899e-01 -5.50825536e-01 -8.29083025e-02 -1.17723262e+00
-1.09419429e+00 4.74031776e-01 8.83219764e-02 -2.08644718e-01
1.31090462e+00 -2.74402872e-02 6.55439734e-01 2.21692488e-01
1.81242377e-01 -8.29577804e-01 -3.80402237e-01 3.68362844e-01
9.08852577e-01 -8.15289199e-01 3.13273013e-01 -9.03398991e-01
-4.54587162e-01 9.36765254e-01 3.79424959e-01 8.78325626e-02
1.02741432e+00 -2.76431255e-03 -5.79692602e-01 -6.22204244e-01
-1.21567380e+00 -6.30987138e-02 3.33361506e-01 3.68739605e-01
5.57735682e-01 3.96325350e-01 -7.61506557e-01 6.00064516e-01
3.66812557e-01 3.58996749e-01 2.63168126e-01 5.94155729e-01
-3.19055259e-01 -1.44193053e+00 1.97230846e-01 8.27182591e-01
-5.83677232e-01 -5.64461350e-01 -7.04680502e-01 4.60148662e-01
-2.47984111e-01 7.82272279e-01 -3.21286291e-01 -5.81079900e-01
1.38503060e-01 -4.21353914e-02 2.19855636e-01 -6.60183728e-01
-5.93670845e-01 3.46450299e-01 2.40755647e-01 -6.79760396e-01
-1.98774725e-01 -6.04871094e-01 -1.41611052e+00 -6.09361410e-01
-3.10177356e-01 7.11965635e-02 4.72592413e-01 3.79379928e-01
1.39817560e+00 7.04979658e-01 6.39198363e-01 1.28556356e-01
-9.42425057e-02 -8.13463867e-01 -6.17247701e-01 1.33188965e-03
-1.52599856e-01 -2.87387967e-01 -1.29344299e-01 -5.78387797e-01] | [5.758615016937256, 5.738428592681885] |
e458c0f2-e3dd-428f-8015-3d0d9312adc9 | repere-premiers-resultats-dun-defi-autour-de | null | null | https://aclanthology.org/F12-1063 | https://aclanthology.org/F12-1063.pdf | REPERE : premiers r\'esultats d'un d\'efi autour de la reconnaissance multimodale des personnes (REPERE : preliminary results of a multimodal person recognition challenge) [in French] | null | ["Matthieu Carr{\\'e}", 'Ludovic Quintard', 'Juliette Kahn', 'Olivier Galibert', 'Aude Giraudel'] | 2012-06-01 | repere-premiers-resultats-dun-defi-autour-de-1 | https://aclanthology.org/F12-1063 | https://aclanthology.org/F12-1063.pdf | jeptalnrecital-2012-6 | ['person-recognition'] | ['computer-vision'] | [-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] | [-7.418349742889404, 3.6206932067871094] |
dc1b4b51-9425-45f7-8326-63d9517d9135 | usr-unsupervised-separated-3d-garment-and | 2302.10518 | null | https://arxiv.org/abs/2302.10518v3 | https://arxiv.org/pdf/2302.10518v3.pdf | USR: Unsupervised Separated 3D Garment and Human Reconstruction via Geometry and Semantic Consistency | Dressed people reconstruction from images is a popular task with promising applications in the creative media and game industry. However, most existing methods reconstruct the human body and garments as a whole with the supervision of 3D models, which hinders the downstream interaction tasks and requires hard-to-obtain data. To address these issues, we propose an unsupervised separated 3D garments and human reconstruction model (USR), which reconstructs the human body and authentic textured clothes in layers without 3D models. More specifically, our method proposes a generalized surface-aware neural radiance field to learn the mapping between sparse multi-view images and geometries of the dressed people. Based on the full geometry, we introduce a Semantic and Confidence Guided Separation strategy (SCGS) to detect, segment, and reconstruct the clothes layer, leveraging the consistency between 2D semantic and 3D geometry. Moreover, we propose a Geometry Fine-tune Module to smooth edges. Extensive experiments on our dataset show that comparing with state-of-the-art methods, USR achieves improvements on both geometry and appearance reconstruction while supporting generalizing to unseen people in real time. Besides, we also introduce SMPL-D model to show the benefit of the separated modeling of clothes and the human body that allows swapping clothes and virtual try-on. | ['Jingyi Chai', 'Wenjun Zhang', 'Bingbing Ni', 'Yuxuan Xiong', 'Yue Shi'] | 2023-02-21 | null | null | null | null | ['virtual-try-on'] | ['computer-vision'] | [ 5.43927066e-02 7.44437873e-02 3.17086279e-01 -3.02037865e-01
-1.94139406e-01 -3.61142546e-01 1.18059188e-01 -5.95254123e-01
1.22397058e-01 2.06204280e-01 1.03622206e-01 4.80678052e-01
1.49723470e-01 -9.26416636e-01 -8.86396587e-01 -6.21503949e-01
4.86585051e-01 6.17201507e-01 4.37935621e-01 -4.68089789e-01
-3.34342271e-01 3.66326094e-01 -1.52305555e+00 1.62138849e-01
8.00767601e-01 1.05481029e+00 2.84555793e-01 3.44599426e-01
1.99910626e-02 2.63487011e-01 6.16324646e-03 -8.01363826e-01
7.20880449e-01 -4.24049169e-01 -3.04722786e-01 7.37414122e-01
6.38377488e-01 -5.10791183e-01 -5.27870953e-01 9.76194263e-01
4.75008339e-01 1.52052924e-01 5.25658548e-01 -8.16315711e-01
-8.45270336e-01 1.76520515e-02 -1.01834798e+00 -7.35931933e-01
5.63995004e-01 1.37621284e-01 5.86086035e-01 -7.99420416e-01
6.54075027e-01 1.30594814e+00 1.03089726e+00 6.74305081e-01
-1.18819368e+00 -5.74042618e-01 4.13923860e-01 -1.52682364e-01
-1.40600789e+00 -2.58084804e-01 1.37805080e+00 -3.40163857e-01
2.74374574e-01 3.14850867e-01 1.14505851e+00 1.21856844e+00
-1.31567225e-01 8.71573925e-01 1.00621593e+00 -3.24270010e-01
3.62147315e-04 1.83311075e-01 -3.88653219e-01 1.22993588e+00
1.57699227e-01 -1.06478874e-02 -4.06260669e-01 1.24661252e-01
1.37957859e+00 3.18246603e-01 -4.00901496e-01 -7.22783267e-01
-9.27100480e-01 4.03446883e-01 4.27344620e-01 -3.03477585e-01
-5.62214673e-01 3.71663235e-02 3.42656597e-02 -2.14018837e-01
8.33235264e-01 -2.39620879e-01 -2.65508026e-01 5.45370102e-01
-8.61689091e-01 2.40708470e-01 5.53461969e-01 1.08936810e+00
6.71462893e-01 9.42578837e-02 3.12920660e-02 1.04364371e+00
5.10206521e-01 7.55378246e-01 -1.84478149e-01 -8.59892368e-01
3.44215423e-01 8.09030414e-01 2.23365590e-01 -1.25072384e+00
-2.76427716e-01 -3.41442615e-01 -1.13760245e+00 1.69297382e-01
3.38365227e-01 1.39872450e-02 -1.00379741e+00 1.41910863e+00
9.25579190e-01 3.66269529e-01 -3.15345496e-01 1.53614759e+00
9.77054834e-01 4.32294607e-01 -1.79481715e-01 2.13923335e-01
1.48053014e+00 -1.17591584e+00 -4.51080024e-01 -4.48802173e-01
-2.24014789e-01 -6.79652393e-01 9.76856887e-01 4.96168017e-01
-1.33334279e+00 -8.26063931e-01 -7.41967916e-01 -4.15574402e-01
1.13435298e-01 2.83591509e-01 7.37319589e-01 5.48080206e-01
-7.00028718e-01 6.21003449e-01 -7.68200099e-01 -3.64425987e-01
3.76087636e-01 9.96343121e-02 -4.72635090e-01 -4.39410299e-01
-8.99307489e-01 5.17367661e-01 3.36729065e-02 6.05015934e-01
-7.54506648e-01 -5.49621105e-01 -1.10826612e+00 -2.95551032e-01
4.77034837e-01 -1.14578617e+00 6.59024417e-01 -1.06360292e+00
-1.69731796e+00 1.16807711e+00 2.02569425e-01 2.08377928e-01
9.53266382e-01 -2.15577036e-01 -2.98843712e-01 2.02340335e-01
-7.26038143e-02 2.98145741e-01 1.07008028e+00 -1.74129939e+00
-2.18658939e-01 -7.52783060e-01 -6.24254160e-02 4.50909376e-01
-4.96962145e-02 -3.67592871e-01 -1.07985866e+00 -8.94528866e-01
5.68815947e-01 -8.38801146e-01 -2.29539230e-01 6.67994857e-01
-4.79803860e-01 3.57004762e-01 4.34131116e-01 -1.31040883e+00
7.53114820e-01 -2.06843758e+00 6.74738109e-01 2.67109305e-01
1.81992143e-01 -1.29237220e-01 -1.37832388e-01 -7.76757076e-02
2.82237470e-01 -4.54002887e-01 -2.94699788e-01 -8.89746845e-01
1.97517872e-01 3.06778967e-01 9.22907218e-02 8.49604130e-01
-6.39695153e-02 8.80252540e-01 -7.56409943e-01 -4.85810459e-01
4.82902259e-01 8.95447791e-01 -5.34179151e-01 4.04174477e-01
-2.02512071e-01 5.75439990e-01 -5.40937543e-01 9.07928646e-01
1.12126887e+00 -3.19488704e-01 3.81438464e-01 -6.59490705e-01
1.92890838e-01 -2.57005870e-01 -1.51936722e+00 2.18073893e+00
-5.28180778e-01 -2.01340899e-01 6.48828924e-01 -7.97217906e-01
9.36187506e-01 2.86446452e-01 4.57541108e-01 -7.65206635e-01
1.25396252e-01 -9.59955901e-03 -8.11982095e-01 -4.95799839e-01
2.67865092e-01 -4.90334779e-01 9.14055854e-03 1.45747244e-01
-1.51046753e-01 -1.85040861e-01 -5.61878443e-01 -1.40453950e-01
4.16633219e-01 8.95545483e-01 -1.32197663e-01 7.47695640e-02
2.78572351e-01 -2.19000325e-01 5.97735226e-01 1.84373066e-01
1.36391327e-01 1.24659371e+00 -1.17199823e-01 -6.93857193e-01
-9.82381821e-01 -1.23896277e+00 2.86677748e-01 9.50193286e-01
7.72945404e-01 -2.65544236e-01 -9.24974978e-01 -6.98994935e-01
1.51958808e-01 3.13931286e-01 -7.16229439e-01 8.69770423e-02
-5.44427216e-01 -6.28177285e-01 6.43472895e-02 5.07109165e-01
7.14341342e-01 -6.89935267e-01 -4.05138910e-01 7.59720383e-03
-5.74576795e-01 -1.23135459e+00 -7.42311537e-01 -3.15384328e-01
-6.27839446e-01 -1.11947072e+00 -9.64749932e-01 -7.99902797e-01
8.41445506e-01 4.52923506e-01 1.12703693e+00 1.84286475e-01
-3.65473092e-01 4.72643226e-01 -2.31278732e-01 7.50386119e-02
-6.28554076e-02 -3.73078525e-01 4.01479229e-02 4.27139342e-01
-1.57989815e-01 -7.76256442e-01 -9.01351452e-01 6.16641402e-01
-6.34435415e-01 7.71014810e-01 4.68986541e-01 4.72306430e-01
9.16005373e-01 1.52817562e-01 1.53153185e-02 -8.08631539e-01
-1.15367949e-01 -3.65745932e-01 -3.59952301e-01 4.62208748e-01
-2.02523068e-01 -1.49342015e-01 4.94260103e-01 -5.53299069e-01
-1.18099320e+00 4.16545540e-01 -2.33423620e-01 -8.53923678e-01
-2.38728300e-01 -3.28892946e-01 -4.98360902e-01 -8.27891231e-02
2.65246153e-01 3.68712246e-01 1.01021826e-01 -9.42605317e-01
3.72912824e-01 4.23152655e-01 5.01682758e-01 -6.80301428e-01
9.13272321e-01 8.41647267e-01 -5.37626445e-02 -6.70517206e-01
-1.00323391e+00 -3.64018828e-01 -5.93339443e-01 -5.59194088e-01
1.07890141e+00 -1.20377600e+00 -6.88886106e-01 6.49724841e-01
-1.10784280e+00 -4.76956934e-01 -2.82980233e-01 1.69101536e-01
-4.79091972e-01 6.21980369e-01 -8.26505125e-01 -8.93672705e-01
-4.08856660e-01 -9.12074566e-01 1.66871727e+00 1.24859139e-01
1.28234312e-01 -8.38440955e-01 -2.81139255e-01 8.51116955e-01
7.96045884e-02 6.34606779e-01 4.11897033e-01 3.63313794e-01
-5.75613558e-01 1.07795924e-01 -2.62475640e-01 4.16629136e-01
8.01723152e-02 -5.65226316e-01 -1.02103853e+00 -1.18976898e-01
4.17344421e-02 -2.58044958e-01 6.92868412e-01 4.61239725e-01
1.02772999e+00 -2.40098163e-01 -2.20066026e-01 9.77214456e-01
1.43739784e+00 -5.72302341e-01 6.16925955e-01 -5.71256168e-02
1.30977404e+00 9.41240549e-01 3.75335306e-01 6.33451581e-01
7.17431962e-01 8.94389033e-01 4.87661242e-01 -4.89573658e-01
-5.41264951e-01 -6.78695202e-01 3.75100970e-01 8.54138672e-01
-6.70593798e-01 1.82631165e-02 -2.02291608e-01 2.22758070e-01
-1.84420097e+00 -6.56307697e-01 -3.17291021e-01 2.27803254e+00
6.03913486e-01 -1.07869834e-01 2.91427732e-01 -2.34456912e-01
7.91881919e-01 1.26699284e-01 -5.91273725e-01 1.79018736e-01
-1.59101024e-01 2.16526836e-01 4.31904584e-01 5.46341240e-01
-8.66714358e-01 9.23344254e-01 5.38426065e+00 8.70476961e-01
-7.88704157e-01 1.84771419e-01 4.99307185e-01 -3.13732028e-02
-5.41593075e-01 -3.20972562e-01 -5.89745343e-01 4.66804385e-01
-4.49922718e-02 7.02347279e-01 9.00802970e-01 8.08685839e-01
1.65317029e-01 2.19310433e-01 -8.51894975e-01 1.38058233e+00
3.88489485e-01 -1.01544285e+00 8.74515399e-02 4.47437204e-02
7.75960267e-01 -4.25396860e-01 -2.25301743e-01 5.44486344e-02
1.55408025e-01 -7.29628980e-01 1.33609569e+00 9.58864093e-01
1.02169085e+00 -5.44757783e-01 3.75582159e-01 3.64255458e-01
-1.53017068e+00 2.37453371e-01 -3.22465599e-01 -1.35968393e-02
3.61084312e-01 6.39679492e-01 -1.45156011e-01 9.68451023e-01
7.93973625e-01 7.30156422e-01 -4.37188387e-01 5.56453228e-01
-3.45756710e-01 1.71499610e-01 -3.92939895e-01 4.49962854e-01
-2.83148915e-01 -6.45366549e-01 3.45669359e-01 9.79269564e-01
3.04867238e-01 3.65174860e-01 4.90751445e-01 1.19557381e+00
5.28618582e-02 -8.25711861e-02 -2.41795987e-01 3.43748808e-01
-1.01619206e-01 1.38807607e+00 -8.74020994e-01 -8.44371915e-02
-3.34706098e-01 1.85368741e+00 2.37922281e-01 4.59657550e-01
-1.00480461e+00 2.01661676e-01 4.87608194e-01 7.16805041e-01
3.30599099e-01 -2.77451515e-01 -2.31301904e-01 -1.50811398e+00
3.36860150e-01 -6.05290830e-01 1.02324247e-01 -9.66152847e-01
-1.61271548e+00 3.87190908e-01 -2.29456201e-01 -1.18016422e+00
3.82151932e-01 -5.16125500e-01 -2.09750071e-01 8.78602028e-01
-1.44312513e+00 -1.77231669e+00 -6.95608974e-01 7.33787894e-01
6.23557985e-01 3.86087924e-01 6.33572400e-01 4.08599406e-01
-4.94983494e-01 3.29630166e-01 -2.38873228e-01 1.29223302e-01
6.31799698e-01 -1.01262593e+00 4.70382184e-01 5.52959681e-01
-3.40696089e-02 2.89371252e-01 4.40480590e-01 -9.29562628e-01
-1.84379816e+00 -1.23834336e+00 1.66110635e-01 -3.91004175e-01
-1.30174130e-01 -7.60973930e-01 -7.58994281e-01 5.43250680e-01
-3.11742693e-01 2.21152276e-01 2.99612761e-01 -1.54219531e-02
-3.30749661e-01 -1.33210212e-01 -1.43253994e+00 6.03226602e-01
1.58827496e+00 -2.76345819e-01 -3.24091077e-01 2.39788055e-01
4.78914380e-01 -8.01351726e-01 -7.02364683e-01 3.52113575e-01
9.49420154e-01 -1.12200201e+00 1.45246172e+00 -1.47983611e-01
4.04010624e-01 -4.47883844e-01 -3.78701925e-01 -1.05971050e+00
-3.87893736e-01 -4.59641784e-01 -4.23515737e-01 1.07501292e+00
-3.09904963e-01 -2.40685940e-01 9.48762119e-01 8.21447849e-01
-5.78649901e-02 -7.11488843e-01 -4.99521703e-01 -4.46465313e-01
-4.73895967e-01 -4.23214227e-01 5.43282568e-01 9.69886124e-01
-6.63020432e-01 2.18030214e-01 -9.94175851e-01 4.94072795e-01
1.09849679e+00 4.77330774e-01 8.90975535e-01 -1.32964110e+00
-7.11381376e-01 1.04629423e-03 -2.08113581e-01 -1.37026083e+00
-1.24623356e-02 -7.34627426e-01 1.47762299e-01 -1.82334459e+00
3.03450614e-01 -5.52790940e-01 1.05110824e-01 3.51955086e-01
-1.59228295e-01 6.68196619e-01 1.07546179e-02 -1.44749388e-01
-6.88396215e-01 8.27764630e-01 1.67641044e+00 -1.03472546e-01
-1.43772021e-01 2.14183424e-03 -7.00913489e-01 1.01481342e+00
2.58849651e-01 -1.12594068e-01 -2.95270741e-01 -6.28658175e-01
1.93683714e-01 -1.37100434e-02 9.36764002e-01 -7.52845764e-01
-3.52394544e-02 -3.18823248e-01 8.25036108e-01 -4.23129410e-01
7.09937096e-01 -1.02387667e+00 7.59696066e-01 6.13575466e-02
2.06524596e-01 -4.21709508e-01 2.34336555e-02 8.33468378e-01
4.55131412e-01 1.60287112e-01 8.29623640e-01 -5.08051336e-01
-5.91778994e-01 5.28003991e-01 3.76170367e-01 -9.28289369e-02
8.33531141e-01 -4.12304312e-01 3.91187489e-01 -1.52567357e-01
-8.35321069e-01 7.03403205e-02 9.98472571e-01 3.35579306e-01
9.58831787e-01 -1.39754272e+00 -7.20606327e-01 4.99123335e-01
-1.38550594e-01 5.79846740e-01 9.11181331e-01 6.24656379e-01
-6.44811988e-01 -4.15513366e-01 -8.64508897e-02 -5.38866937e-01
-1.20724320e+00 5.96224904e-01 5.06103337e-01 8.28657448e-02
-1.29937363e+00 7.02051342e-01 6.05521679e-01 -7.49931097e-01
2.71926105e-01 -2.77450144e-01 3.23302567e-01 -3.71915251e-01
4.08596694e-01 3.69035214e-01 -1.46832183e-01 -7.89247990e-01
-1.94265977e-01 1.12855327e+00 3.86080176e-01 9.58222970e-02
1.45963860e+00 -4.62938011e-01 -5.90138994e-02 2.87855595e-01
7.95356989e-01 2.18978778e-01 -1.74938893e+00 -2.44858176e-01
-7.89460242e-01 -7.65557110e-01 2.90820142e-03 -7.08546460e-01
-1.42864358e+00 7.39037693e-01 5.38240850e-01 -1.41697958e-01
1.18503273e+00 4.21001501e-02 1.21029758e+00 -2.31608197e-01
6.95186496e-01 -1.06226480e+00 9.58871022e-02 -1.06313959e-01
9.70302939e-01 -1.10709727e+00 9.39295217e-02 -8.99520218e-01
-7.57890999e-01 9.05775905e-01 5.95170379e-01 -3.10644209e-01
5.52385688e-01 3.53145637e-02 -1.83864579e-01 -4.09275502e-01
2.39605252e-02 -1.39994562e-01 6.07930362e-01 7.03051865e-01
-2.57142028e-03 2.06324160e-01 3.03586304e-01 9.84752893e-01
1.02821112e-01 -4.28963676e-02 -2.18957104e-03 5.37782192e-01
-7.69832637e-03 -8.43756497e-01 -5.44951022e-01 8.02675355e-03
-1.36747239e-02 7.16057606e-03 -2.41402760e-01 5.12642801e-01
6.14253223e-01 6.40759766e-01 -2.11475864e-01 -4.80129689e-01
6.13996029e-01 -1.30241871e-01 9.34232235e-01 -6.21817410e-01
-4.16136086e-01 3.78355920e-01 -5.19306175e-02 -6.54706001e-01
-5.22018433e-01 -2.53699362e-01 -9.72443402e-01 -3.63214433e-01
-3.63589406e-01 -2.68033415e-01 5.11004388e-01 7.99960077e-01
2.93647200e-01 4.81896639e-01 5.16790211e-01 -1.38781965e+00
-1.26242280e-01 -6.02678895e-01 -1.13245869e+00 7.00818539e-01
1.87660828e-01 -9.22300458e-01 -9.83533487e-02 1.68744117e-01] | [7.2929840087890625, -1.3059332370758057] |
e63fc1ff-412d-48c4-a00b-aa4b4edb731e | secure-data-sharing-with-flow-model | 2009.11762 | null | https://arxiv.org/abs/2009.11762v1 | https://arxiv.org/pdf/2009.11762v1.pdf | Secure Data Sharing With Flow Model | In the classical multi-party computation setting, multiple parties jointly compute a function without revealing their own input data. We consider a variant of this problem, where the input data can be shared for machine learning training purposes, but the data are also encrypted so that they cannot be recovered by other parties. We present a rotation based method using flow model, and theoretically justified its security. We demonstrate the effectiveness of our method in different scenarios, including supervised secure model training, and unsupervised generative model training. Our code is available at https://github.com/ duchenzhuang/flowencrypt. | ['Chenwei Wu', 'Yang Yuan', 'Chenzhuang Du'] | 2020-09-24 | null | null | null | null | ['privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['methodology', 'natural-language-processing'] | [ 2.03621045e-01 9.94082242e-02 -1.97464347e-01 -1.89069703e-01
-1.07566035e+00 -1.49464393e+00 6.34563267e-01 -2.02088699e-01
-3.12987417e-01 7.99657643e-01 -1.81147084e-02 -6.33691967e-01
1.17620997e-01 -1.17209828e+00 -8.34778607e-01 -1.16664743e+00
-3.52212526e-02 6.71870887e-01 -3.81814837e-01 1.80334389e-01
1.73323564e-02 4.26183045e-01 -6.20239615e-01 4.13202137e-01
1.34412184e-01 5.29354870e-01 -3.87780577e-01 1.09045839e+00
5.80593765e-01 6.93744838e-01 -4.00344402e-01 -9.43039238e-01
8.84877264e-01 -3.90795350e-01 -1.29039061e+00 -3.34957242e-01
-1.55963257e-01 -4.91855383e-01 -9.03886020e-01 1.14903939e+00
3.39303464e-01 -3.23806524e-01 4.04597342e-01 -1.69819641e+00
-6.00504398e-01 8.40163887e-01 -5.45828044e-02 -3.82168919e-01
1.31609272e-02 3.05415392e-01 1.06084216e+00 -3.87642562e-01
9.16824102e-01 9.82128918e-01 1.14567563e-01 6.30681038e-01
-1.37650919e+00 -7.89478540e-01 -6.54951930e-01 9.38354209e-02
-1.44991136e+00 -3.04155141e-01 8.45300019e-01 -9.94135961e-02
3.67360353e-01 5.30463815e-01 3.42016071e-01 9.82649624e-01
2.40331233e-01 7.74202406e-01 1.60544014e+00 -3.55163634e-01
2.76330914e-02 5.10639369e-01 -3.31732212e-03 4.58911479e-01
3.82190645e-01 4.88170981e-01 -1.94093719e-01 -9.46192205e-01
4.41930592e-01 2.40776613e-01 -5.73365271e-01 -2.28077859e-01
-1.17766762e+00 1.07207966e+00 2.28619009e-01 2.16166094e-01
-2.60849260e-02 6.20639145e-01 1.33209050e-01 8.94896328e-01
2.20442697e-01 -1.44804180e-01 -7.09956110e-01 4.08839956e-02
-3.94128293e-01 3.93329501e-01 1.15942049e+00 9.88284051e-01
1.11716402e+00 -3.62442136e-01 1.10933013e-01 -1.41247034e-01
3.45990419e-01 8.17966700e-01 1.60956636e-01 -9.36060846e-01
6.87617004e-01 -2.09831111e-02 -8.54226295e-03 -5.48544765e-01
3.05840760e-01 8.98203924e-02 -1.02285624e+00 -1.00998186e-01
4.46300328e-01 -5.14298856e-01 -5.69286942e-01 1.74060607e+00
3.41818035e-01 2.03575179e-01 6.86693728e-01 5.79685986e-01
3.94789994e-01 6.86180472e-01 -2.51954466e-01 -1.86894238e-02
1.24848866e+00 -7.16956615e-01 -5.75654864e-01 3.51575404e-01
8.30435455e-01 -8.29588950e-01 3.78272116e-01 2.64945060e-01
-9.67112720e-01 7.12483823e-02 -6.23485565e-01 -3.37032348e-01
-3.18683952e-01 -1.23587593e-01 1.05595696e+00 1.10078311e+00
-1.04275298e+00 7.16450572e-01 -8.65794837e-01 3.96188080e-01
5.99233985e-01 8.30066741e-01 -8.53728294e-01 -1.08067483e-01
-1.37446773e+00 3.09426934e-01 6.19011819e-01 -1.29888326e-01
-1.10211730e+00 -2.44295120e-01 -6.99207962e-01 1.28967106e-01
1.79193869e-01 -9.18038309e-01 1.35558069e+00 -6.01522148e-01
-1.39920545e+00 8.69609892e-01 -3.49724591e-01 -4.95586514e-01
5.62445462e-01 4.55720693e-01 -1.51036456e-01 1.34399831e-01
-4.96168345e-01 -3.26394439e-02 7.52838492e-01 -1.45708561e+00
-4.41852987e-01 -5.45213640e-01 3.51283193e-01 -2.71182477e-01
6.15094639e-02 1.48777008e-01 -5.76898195e-02 -3.58112693e-01
-1.21056819e-02 -1.60996521e+00 -3.68428916e-01 -3.80486220e-01
-6.21027410e-01 2.25095928e-01 1.06850147e+00 -5.62930405e-01
8.25907111e-01 -1.99493992e+00 -1.24692582e-01 7.70048082e-01
2.79125929e-01 3.00469965e-01 1.44713661e-02 9.14030731e-01
1.61415070e-01 6.49515688e-01 -4.08452123e-01 -4.93634015e-01
1.30378678e-01 3.66967112e-01 -7.39566088e-01 6.96172535e-01
-3.01067412e-01 1.26895225e+00 -7.35710502e-01 -1.42569035e-01
-2.01342851e-01 6.54229641e-01 -5.26943207e-01 1.13288261e-01
4.80765067e-02 8.40957999e-01 -4.83964205e-01 4.95553166e-01
1.15244114e+00 -5.13299704e-01 7.11658299e-01 2.74439193e-02
4.39906955e-01 2.43161559e-01 -1.39375210e+00 1.30974102e+00
-4.18858439e-01 3.80558133e-01 9.91618857e-02 -3.95908803e-01
3.52348685e-01 9.45509613e-01 2.18139932e-01 2.49376595e-01
5.45232892e-01 1.83178857e-01 1.18169099e-01 -1.95971176e-01
4.21427071e-01 -2.26199105e-01 -2.25772157e-01 1.41104543e+00
6.28969148e-02 -1.65970907e-01 -4.40713435e-01 4.47030336e-01
8.80503476e-01 -1.50512844e-01 3.55550408e-01 1.18385926e-01
3.09155792e-01 -2.68238008e-01 4.91691828e-01 6.27065778e-01
7.23448023e-02 4.74350154e-01 7.98434556e-01 -4.25469488e-01
-1.26492798e+00 -7.58470118e-01 1.23429358e-01 6.17333293e-01
1.42921403e-01 -7.28035748e-01 -6.90629601e-01 -1.03571427e+00
1.94793820e-01 3.40359509e-01 -4.24091756e-01 1.16291620e-01
-4.54989791e-01 -8.25762928e-01 8.89786005e-01 1.13664262e-01
5.72429001e-01 -6.39366388e-01 -2.47508381e-02 -1.82410508e-01
-3.06445837e-01 -1.13639271e+00 -5.52045584e-01 7.06931949e-02
-9.10629153e-01 -1.31935310e+00 -2.33593047e-01 -5.23346901e-01
1.11621451e+00 3.86932880e-01 5.47556102e-01 3.96587223e-01
7.73367658e-02 3.28022480e-01 -9.12867337e-02 -3.76337856e-01
-9.60900128e-01 1.62334934e-01 -1.36756808e-01 3.45584393e-01
3.33440937e-02 -5.51688075e-01 -7.26014614e-01 1.03634804e-01
-1.45904863e+00 5.89516421e-04 4.31697875e-01 8.30804348e-01
5.11963248e-01 1.01225004e-01 -8.39263648e-02 -1.68923688e+00
5.27599514e-01 -4.46620435e-01 -7.34380364e-01 4.52699542e-01
-5.41015148e-01 3.32262665e-01 8.67176056e-01 -3.62285197e-01
-7.93879092e-01 1.57533973e-01 2.67301146e-02 -3.13453674e-01
-5.68852425e-02 4.72284667e-03 -3.13028574e-01 -5.43051004e-01
2.32550308e-01 3.61830205e-01 1.04107201e-01 -3.62555981e-01
4.29739028e-01 8.68010640e-01 1.24827541e-01 -7.85583496e-01
1.44974995e+00 8.61429691e-01 3.94268513e-01 -1.14444718e-02
-8.80946219e-02 6.28397837e-02 -6.17799878e-01 4.29328680e-01
2.77058423e-01 -8.02830696e-01 -1.27512348e+00 7.75957763e-01
-1.31631124e+00 -3.69387865e-01 -1.59625083e-01 4.54357922e-01
-5.12393534e-01 6.36394024e-01 -8.78829479e-01 -7.16866791e-01
-6.40309334e-01 -1.21547914e+00 7.42035925e-01 1.48970932e-01
3.22922587e-01 -1.31032372e+00 2.07411230e-01 5.18466353e-01
4.14942652e-01 1.88035205e-01 6.80201709e-01 -1.08465457e+00
-1.14050710e+00 -4.66073483e-01 7.67999589e-02 4.59311724e-01
6.63898792e-03 -4.64074090e-02 -1.34545183e+00 -7.81703889e-01
1.61746308e-01 -3.11354369e-01 5.52150071e-01 -3.29190463e-01
1.31295991e+00 -1.03471708e+00 -1.93110362e-01 1.00865936e+00
1.68491411e+00 -8.26702863e-02 7.85275698e-01 -1.37904555e-01
4.16329950e-01 1.46708593e-01 -1.05496891e-01 4.96884018e-01
5.96094668e-01 4.09385324e-01 3.44117373e-01 6.90432787e-02
4.21970457e-01 -5.19391179e-01 1.91804186e-01 5.99077582e-01
-5.05284011e-01 -3.10850114e-01 -5.04229128e-01 1.99986175e-01
-1.65422511e+00 -1.06915069e+00 -1.24698699e-01 2.23130155e+00
9.57887828e-01 -4.21487898e-01 -3.80106747e-01 -3.64119895e-02
6.01585746e-01 1.11323871e-01 -3.69898789e-02 -4.76886541e-01
-2.07732052e-01 7.77604103e-01 1.10600185e+00 7.62934744e-01
-1.04197192e+00 9.43997443e-01 5.59202862e+00 5.52094519e-01
-1.01459968e+00 6.44771636e-01 8.45297754e-01 2.03404143e-01
-8.69944751e-01 8.79981220e-01 -6.50390506e-01 4.94746566e-01
9.92249429e-01 -6.19937778e-01 8.22606385e-01 7.33223021e-01
-2.80159265e-01 1.06744304e-01 -1.11609852e+00 8.09474170e-01
-3.84536058e-01 -1.38551366e+00 -2.50370175e-01 4.69739199e-01
9.48428273e-01 -1.98552534e-01 3.37381750e-01 1.29588977e-01
6.46245420e-01 -8.14067662e-01 1.81925371e-01 4.40477341e-01
8.54403079e-01 -8.19616139e-01 7.33251929e-01 5.13477623e-01
-7.77042389e-01 1.28612548e-01 -3.45633805e-01 2.15950236e-01
4.76853065e-02 2.38449603e-01 -9.10804033e-01 1.13863528e+00
2.14721989e-02 -4.19611707e-02 -3.15711141e-01 6.91959143e-01
-7.62509644e-01 8.63833129e-01 -4.16644841e-01 1.49611428e-01
8.64616558e-02 -3.92802536e-01 3.89318079e-01 9.15600598e-01
4.90179449e-01 4.42678660e-01 -1.15980223e-01 7.81102836e-01
-6.49241030e-01 -8.91381130e-02 -7.38049030e-01 1.39354080e-01
6.05896831e-01 1.32215106e+00 -3.51125032e-01 -4.22066033e-01
-1.12466551e-01 1.20934188e+00 1.18271320e-03 4.82761979e-01
-6.15478098e-01 -3.37672442e-01 6.59415841e-01 -2.53016740e-01
1.78444043e-01 -3.15975010e-01 -8.17920715e-02 -1.70469475e+00
-4.85348590e-02 -8.31743419e-01 5.03465354e-01 -2.46169627e-01
-1.06143808e+00 4.32308793e-01 -1.31985471e-01 -1.28838682e+00
-4.17291373e-01 -3.23880762e-01 -8.40204537e-01 1.12318206e+00
-1.66776073e+00 -1.63497996e+00 2.81602442e-01 1.12924218e+00
-7.16948032e-01 -1.57258719e-01 1.30071414e+00 -8.11658427e-02
-1.02391422e-01 8.32697809e-01 4.82969642e-01 4.30542260e-01
5.60556293e-01 -1.13271439e+00 4.85172153e-01 1.01240635e+00
5.50533235e-01 1.04756534e+00 3.50763470e-01 -4.57770020e-01
-1.98018479e+00 -1.09857059e+00 1.17840672e+00 -4.21162009e-01
5.54429650e-01 -6.06566966e-01 -6.46661937e-01 1.34564459e+00
4.16611314e-01 3.93929631e-01 1.03910410e+00 -3.87449026e-01
-6.62612915e-01 -5.23741059e-02 -1.52890980e+00 2.48988688e-01
4.81695771e-01 -9.25243974e-01 3.64449583e-02 6.87509716e-01
5.40175200e-01 -6.91727400e-01 -8.58221769e-01 -4.11246531e-02
4.72093731e-01 -5.64982533e-01 5.64772964e-01 -9.35185611e-01
5.79993427e-02 -2.85661191e-01 -2.42359623e-01 -1.05901921e+00
1.67409137e-01 -1.41877222e+00 -4.93907809e-01 8.93953979e-01
4.81123716e-01 -1.31760013e+00 7.77602136e-01 9.43755150e-01
7.96180665e-01 -2.97941238e-01 -1.20205903e+00 -5.38286686e-01
3.33040982e-01 -4.11775082e-01 1.12062740e+00 9.13269579e-01
-8.81806463e-02 -1.61399007e-01 -8.00974727e-01 5.68778753e-01
8.46513271e-01 5.42938232e-01 1.26486886e+00 -7.65735567e-01
-6.95347846e-01 2.15976626e-01 -2.64215142e-01 -9.21435654e-01
2.21900597e-01 -1.36932659e+00 -6.71254456e-01 -9.39887166e-01
5.43623984e-01 -5.72553456e-01 -1.17540024e-01 8.02482724e-01
9.76615250e-02 3.36139679e-01 6.62198484e-01 4.72770512e-01
-7.26089478e-02 2.43777588e-01 1.26625228e+00 3.28985602e-02
2.44560614e-01 7.10440159e-01 -9.77891982e-01 1.66972622e-01
1.21078348e+00 -8.22995126e-01 -2.26603195e-01 -3.34609330e-01
2.26859108e-01 3.71567369e-01 7.69908607e-01 -4.90328074e-01
4.01041120e-01 -1.79687634e-01 3.02524567e-02 -2.31960997e-01
1.94614440e-01 -1.04528272e+00 8.63860428e-01 8.06851923e-01
-2.81548768e-01 7.38146231e-02 -3.88094693e-01 3.46334398e-01
-8.87725800e-02 -3.29800367e-01 7.70290256e-01 -8.03314298e-02
-2.57246830e-02 9.69277203e-01 -1.95625629e-02 -2.63548762e-01
1.13639569e+00 5.15527844e-01 -7.05101728e-01 -7.66791821e-01
-9.57256198e-01 -7.06662312e-02 7.58399189e-01 -8.10112432e-02
5.47703028e-01 -1.26416838e+00 -8.07620704e-01 5.57936192e-01
-2.74506271e-01 -8.94717127e-02 3.72530371e-01 7.76327670e-01
-4.99798954e-01 5.10574996e-01 1.47084802e-01 -7.63532966e-02
-1.64135611e+00 5.61886966e-01 7.97270834e-02 -6.36464894e-01
-3.51244271e-01 3.22253019e-01 -1.60529152e-01 -5.72922349e-01
-3.30703676e-01 -7.04867095e-02 6.56706452e-01 -4.28028226e-01
5.56343317e-01 2.44141996e-01 -7.59413019e-02 -6.11499369e-01
1.80535957e-01 1.93890929e-01 2.80488492e-03 -4.58850801e-01
1.23405457e+00 -1.25112208e-02 -5.64682961e-01 -2.69141167e-01
1.73582602e+00 3.95002306e-01 -6.48728073e-01 -6.73622131e-01
-5.37704349e-01 -8.37682068e-01 -2.81876028e-01 -3.97377729e-01
-1.39496374e+00 8.54851782e-01 3.79019082e-01 2.46013001e-01
9.59135413e-01 2.24927608e-02 1.01967573e+00 4.78579700e-01
9.62348580e-01 -4.89664823e-01 -7.59446502e-01 4.50805753e-01
2.42359921e-01 -1.18248093e+00 -6.50353655e-02 -3.83319378e-01
-7.80010939e-01 1.16145384e+00 -2.33476296e-01 -5.23266979e-02
9.30720627e-01 4.91270930e-01 7.31134415e-02 2.42546633e-01
-1.00929236e+00 1.41244948e-01 -2.09902465e-01 6.29683018e-01
-2.25052387e-01 4.47066426e-01 -1.36374310e-01 5.85926235e-01
-3.59224230e-01 1.23586869e-02 8.55531812e-01 1.07274771e+00
3.56235653e-01 -2.13104963e+00 -4.45648670e-01 -4.80822101e-02
-1.06130576e+00 -2.15613693e-01 -2.08646014e-01 7.05053449e-01
-4.35156971e-02 5.99454701e-01 -5.03336489e-01 -4.60827738e-01
-3.32177758e-01 1.60408005e-01 3.54045212e-01 -3.63420069e-01
-8.78705084e-01 -8.62683505e-02 -1.80543020e-01 -2.35483274e-01
-4.98987287e-01 -5.47528028e-01 -1.05078852e+00 -1.03911531e+00
-2.73453742e-01 4.13859606e-01 8.25250685e-01 4.91951048e-01
4.60860163e-01 -3.39230448e-01 1.66144788e+00 -2.40528747e-01
-7.17917204e-01 -6.36848986e-01 -6.33210778e-01 3.72325897e-01
3.77151430e-01 2.98634589e-01 -5.99220991e-01 3.25816035e-01] | [5.836306571960449, 6.857388496398926] |
b631c7cb-6920-4c83-8ed8-4f5bbb7a23db | the-hardware-impact-of-quantization-and | 2302.04174 | null | https://arxiv.org/abs/2302.04174v1 | https://arxiv.org/pdf/2302.04174v1.pdf | The Hardware Impact of Quantization and Pruning for Weights in Spiking Neural Networks | Energy efficient implementations and deployments of Spiking neural networks (SNNs) have been of great interest due to the possibility of developing artificial systems that can achieve the computational powers and energy efficiency of the biological brain. Efficient implementations of SNNs on modern digital hardware are also inspired by advances in machine learning and deep neural networks (DNNs). Two techniques widely employed in the efficient deployment of DNNs -- the quantization and pruning of parameters, can both compress the model size, reduce memory footprints, and facilitate low-latency execution. The interaction between quantization and pruning and how they might impact model performance on SNN accelerators is currently unknown. We study various combinations of pruning and quantization in isolation, cumulatively, and simultaneously (jointly) to a state-of-the-art SNN targeting gesture recognition for dynamic vision sensor cameras (DVS). We show that this state-of-the-art model is amenable to aggressive parameter quantization, not suffering from any loss in accuracy down to ternary weights. However, pruning only maintains iso-accuracy up to 80% sparsity, which results in 45% more energy than the best quantization on our architectural model. Applying both pruning and quantization can result in an accuracy loss to offer a favourable trade-off on the energy-accuracy Pareto-frontier for the given hardware configuration. | ['Siddharth Joshi', 'Mark Horeni', 'Pooria Taheri', 'Clemens JS Schaefer'] | 2023-02-08 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 5.88891268e-01 5.69521263e-02 -3.72576411e-03 -1.49236664e-01
-8.31978321e-02 -2.83107460e-01 6.50312424e-01 -1.17402487e-02
-9.77026939e-01 3.36158454e-01 -2.12897882e-01 -3.76924068e-01
-1.03878275e-01 -8.40167761e-01 -5.98290801e-01 -8.95255268e-01
-9.84017998e-02 2.50834346e-01 6.39474571e-01 -1.95892411e-03
2.25856498e-01 8.61638665e-01 -2.05243707e+00 7.56486580e-02
3.29854399e-01 1.38311052e+00 2.80067652e-01 8.12464833e-01
-7.98753835e-03 3.33573848e-01 -6.09680951e-01 -2.45133564e-01
5.47095537e-01 -3.30913782e-01 -1.91920802e-01 -5.74400663e-01
4.06802624e-01 -2.54382819e-01 -4.59186226e-01 9.53692377e-01
6.39900506e-01 -1.32723406e-01 4.18053985e-01 -1.11035311e+00
-2.60473276e-03 5.15909672e-01 -2.18317807e-01 4.76629794e-01
-4.13701743e-01 5.34522593e-01 4.17160779e-01 -2.25190058e-01
4.44158554e-01 9.67947841e-01 7.15386331e-01 8.40820670e-01
-1.17417169e+00 -7.58014679e-01 -2.38038778e-01 1.89865142e-01
-1.14340210e+00 -8.63371968e-01 3.43335629e-01 8.29280987e-02
1.96761560e+00 1.21590324e-01 1.28951108e+00 9.42592323e-01
4.20497030e-01 2.08845049e-01 8.01249862e-01 -4.24051076e-01
8.68710160e-01 -5.25996864e-01 1.25068471e-01 7.14150310e-01
7.83607900e-01 3.58910769e-01 -9.64050174e-01 -2.15236680e-03
7.02234387e-01 -1.53160840e-01 -6.26043007e-02 -1.26904175e-01
-8.35600078e-01 4.35521871e-01 4.52330172e-01 3.11697721e-01
-3.58192086e-01 9.18714643e-01 5.64193070e-01 -7.80160204e-02
-2.32911423e-01 5.69868028e-01 -4.72158045e-01 -6.34410024e-01
-1.09423697e+00 2.51682103e-01 8.66351008e-01 6.47897720e-01
2.08293632e-01 4.33501571e-01 1.83773220e-01 4.76251662e-01
2.26406351e-01 5.08669734e-01 6.67594671e-01 -1.03548872e+00
1.35640904e-01 7.07279205e-01 -3.14776331e-01 -2.96428412e-01
-7.89415956e-01 -3.40039134e-01 -8.67630780e-01 6.79857314e-01
3.72184694e-01 -4.62397002e-02 -1.19863582e+00 1.70295978e+00
1.03806786e-01 3.45530026e-02 9.66016352e-02 8.13708305e-01
4.35873687e-01 5.05997062e-01 3.07527512e-01 -1.25597537e-01
1.45643771e+00 -5.30771911e-01 -3.70189250e-01 -4.03798848e-01
5.27026892e-01 -3.38859260e-01 6.80184066e-01 5.10593832e-01
-1.30776429e+00 -2.06355140e-01 -1.68865395e+00 -2.17616513e-01
-4.25908715e-01 -1.04505591e-01 6.52104139e-01 1.07781506e+00
-1.30011177e+00 9.19749916e-01 -1.67863727e+00 -4.93239999e-01
6.72364295e-01 1.03965974e+00 1.66447774e-01 7.83412382e-02
-5.82468927e-01 1.20183122e+00 4.78995889e-01 -3.94630916e-02
-5.76592743e-01 -8.36040437e-01 -3.92053425e-01 3.97153676e-01
2.64679398e-02 -8.15962851e-01 1.23526371e+00 -8.02690148e-01
-1.64399600e+00 5.29513717e-01 -2.46696658e-02 -1.04023719e+00
2.48901173e-02 1.68465942e-01 -2.39348650e-01 5.70188239e-02
-6.97263002e-01 1.28458118e+00 5.81998765e-01 -5.23180604e-01
-4.51854169e-01 -5.49678922e-01 1.75051228e-03 4.33936976e-02
-6.85475707e-01 -4.92812432e-02 -1.86667904e-01 -2.92512119e-01
9.32734758e-02 -9.81547534e-01 -4.11694735e-01 4.71279800e-01
1.53224871e-01 -1.28325760e-01 9.63054657e-01 -2.11532474e-01
9.64795470e-01 -1.89427400e+00 4.74524386e-02 8.43064114e-02
1.05156876e-01 6.34578645e-01 -2.07893867e-02 -1.72805116e-01
4.07316327e-01 -8.47205892e-02 -1.84165746e-01 -2.21331045e-01
-9.20705572e-02 6.55327260e-01 2.95128617e-02 4.76862937e-01
1.16933607e-01 8.34015250e-01 -4.11553472e-01 -1.15185857e-01
2.62637824e-01 7.95813978e-01 -6.95744455e-01 -2.54128605e-01
-2.11357832e-01 -2.44887039e-01 -5.28413244e-02 7.36331105e-01
2.84235626e-01 -9.88990664e-02 3.88622671e-01 -4.72074658e-01
-2.12330431e-01 2.75255233e-01 -8.45556498e-01 1.77142417e+00
-1.80535033e-01 8.89508009e-01 1.36864945e-01 -8.57358336e-01
7.65589178e-01 -2.31092572e-02 3.35480839e-01 -1.09240615e+00
4.83527035e-01 5.29593825e-01 4.58751589e-01 -4.16670330e-02
3.86929393e-01 -5.71680218e-02 2.86945760e-01 1.24043599e-01
3.93083453e-01 4.15219255e-02 2.42435426e-01 -3.08260381e-01
1.40176249e+00 1.11384235e-01 1.81762367e-01 -7.45760143e-01
-8.63484368e-02 2.68755537e-02 3.32250416e-01 5.79209447e-01
-2.84660369e-01 1.74412765e-02 4.47037518e-01 -5.45511484e-01
-1.39416313e+00 -7.95701802e-01 -2.03501388e-01 8.71900082e-01
1.37903526e-01 -3.22391748e-01 -1.09561634e+00 1.89949840e-01
-3.28830153e-01 5.47354519e-01 -3.08126748e-01 -4.33285773e-01
-7.48327732e-01 -1.00463891e+00 1.06155181e+00 6.78520977e-01
6.04932368e-01 -8.45647931e-01 -2.16196752e+00 3.01700741e-01
6.04644477e-01 -1.15374362e+00 2.15837777e-01 1.22751820e+00
-1.36464119e+00 -8.97353470e-01 -2.88808614e-01 -4.26432341e-01
4.79869038e-01 -1.74369246e-01 1.06235111e+00 2.22148612e-01
-5.66915870e-01 6.89614341e-02 1.05179427e-02 -7.21651793e-01
-2.02092603e-01 2.38073424e-01 2.58194268e-01 -8.36090386e-01
5.64694047e-01 -9.51120377e-01 -8.92509937e-01 -8.61978382e-02
-9.01557922e-01 1.66290328e-01 6.16585553e-01 6.47802532e-01
6.06376112e-01 -8.17475989e-02 1.07191205e-01 -1.82648093e-01
1.98970184e-01 -8.55642110e-02 -1.01747167e+00 6.91401288e-02
-9.03786898e-01 5.85150480e-01 8.01410079e-01 -5.56098163e-01
-5.18714726e-01 4.03765976e-01 -2.54287094e-01 -3.96758348e-01
1.83974660e-03 -9.27255824e-02 -6.57615066e-02 -6.75926626e-01
7.81244159e-01 6.10860661e-02 -9.13218036e-02 -1.89668342e-01
7.49276727e-02 2.59462506e-01 7.09954858e-01 -3.61832768e-01
1.03184029e-01 5.62968910e-01 7.38231480e-01 -9.04204905e-01
3.33007053e-02 -3.39471299e-04 -2.90482283e-01 -1.29050121e-01
7.56325603e-01 -5.82125306e-01 -9.20896411e-01 5.39052546e-01
-1.01232171e+00 -4.91384983e-01 -4.61475909e-01 2.86800534e-01
-4.76287156e-01 -5.82794473e-02 -4.63759482e-01 -8.03390265e-01
-5.93463302e-01 -1.29048586e+00 9.05923247e-01 6.05699897e-01
-1.81071714e-01 -5.11796534e-01 -2.77097493e-01 -2.11806446e-01
8.17509651e-01 8.36899504e-02 1.04093206e+00 -4.39851463e-01
-6.31544173e-01 2.42003828e-01 -2.18651667e-01 -8.15118384e-03
-4.44167346e-01 -7.63133913e-02 -1.12700462e+00 -2.85102457e-01
1.91180594e-02 -3.39338988e-01 1.05115616e+00 7.52284765e-01
1.05398107e+00 -1.52013451e-01 -4.36300516e-01 9.31019247e-01
1.78875196e+00 4.25097287e-01 6.71202242e-01 3.42510343e-01
3.54369670e-01 1.45980092e-02 -2.30740547e-01 4.49204862e-01
-1.73572570e-01 7.60442197e-01 8.57261121e-01 2.40274385e-01
-3.22957754e-01 1.52845860e-01 1.78860828e-01 7.15464473e-01
-2.31863901e-01 -3.11148763e-01 -1.00883639e+00 4.17540938e-01
-1.45217717e+00 -5.91785908e-01 1.97883248e-01 2.19571090e+00
6.00963414e-01 4.76601750e-01 -2.15375833e-02 3.50862026e-01
1.85357690e-01 -4.60507572e-02 -1.01976418e+00 -7.96092987e-01
-1.90961331e-01 7.69124866e-01 1.27935815e+00 -5.15572242e-02
-7.24508464e-01 4.58105564e-01 6.61131620e+00 8.26248527e-01
-1.23715782e+00 1.34838864e-01 4.44422901e-01 -9.07176793e-01
5.60367107e-02 -2.05152780e-01 -1.11847496e+00 5.65642655e-01
1.81997383e+00 1.12375990e-01 8.13033342e-01 9.57730532e-01
-1.45125628e-01 -4.00270045e-01 -1.11678624e+00 1.23494613e+00
-1.14840657e-01 -1.74706209e+00 5.70247043e-03 3.35005909e-01
3.26737404e-01 4.35644954e-01 -9.91024897e-02 5.84956631e-02
1.97800457e-01 -1.05183721e+00 8.90437424e-01 1.43164843e-01
9.63220119e-01 -9.69929636e-01 6.21444106e-01 3.44488740e-01
-1.15028274e+00 -4.41651702e-01 -6.60781384e-01 -1.87169209e-01
1.51537016e-01 3.91452968e-01 -2.68783450e-01 -1.54990643e-01
1.02659404e+00 -3.86665203e-02 -4.10833091e-01 1.02120459e+00
4.17847455e-01 3.48738939e-01 -9.80475426e-01 -6.21041000e-01
4.21550483e-01 5.61715290e-02 2.58636117e-01 1.15798652e+00
4.62109894e-01 4.75478739e-01 -4.68761176e-01 7.04415441e-01
-2.35989198e-01 -5.58828712e-01 -3.22259545e-01 -1.25841677e-01
7.01340735e-01 9.18927848e-01 -1.19797623e+00 -3.28300774e-01
-1.17122203e-01 7.05744922e-01 2.41320968e-01 -1.09550290e-01
-6.90247238e-01 -2.40956813e-01 9.36990559e-01 1.18874215e-01
3.85731548e-01 -4.36388403e-01 -7.93692768e-01 -5.92415452e-01
1.01865651e-02 -4.96593684e-01 -1.26584068e-01 -5.23312688e-01
-4.14386362e-01 6.56678319e-01 -1.43252328e-01 -6.42702281e-01
-7.59442821e-02 -1.04227698e+00 -2.90149838e-01 5.43170691e-01
-1.31680441e+00 -7.96796739e-01 -4.47336078e-01 2.06741720e-01
4.84438211e-01 -1.04652576e-01 9.29366946e-01 2.93842703e-01
-3.87164265e-01 7.01058507e-01 1.17384329e-01 -4.10661459e-01
-1.33137420e-01 -7.44851351e-01 6.81165934e-01 9.53414083e-01
1.94378704e-01 3.54886293e-01 8.20953727e-01 -3.30560833e-01
-1.92415416e+00 -8.77328515e-01 4.48010772e-01 2.62945425e-02
3.87012810e-01 -4.16319132e-01 -6.71965957e-01 3.84771302e-02
1.90045357e-01 5.41592017e-02 3.88526589e-01 -4.36614543e-01
-1.44404799e-01 -2.69582327e-02 -1.32502913e+00 8.28756213e-01
1.50565934e+00 -1.50918007e-01 -2.11444139e-01 4.90346961e-02
4.93038028e-01 -3.97518694e-01 -5.83467603e-01 4.23170656e-01
8.94558191e-01 -1.10950398e+00 9.00033355e-01 -2.03384474e-01
2.27749765e-01 -1.30291834e-01 -3.52735519e-01 -9.36972260e-01
-9.88752022e-02 -4.10680115e-01 -5.93766034e-01 6.39094234e-01
6.22625984e-02 -4.31251138e-01 1.20688367e+00 7.91600764e-01
-2.68741816e-01 -1.12095773e+00 -1.64223027e+00 -9.68300819e-01
-2.03825042e-01 -4.57895428e-01 4.88589913e-01 1.95813283e-01
-6.78227618e-02 -1.14360124e-01 1.76296607e-01 -5.60280085e-02
7.19771624e-01 -4.54700857e-01 1.20410286e-01 -1.11931825e+00
-2.20710576e-01 -1.08849120e+00 -9.44416046e-01 -9.48992550e-01
-1.98887482e-01 -5.24281502e-01 -1.44234532e-02 -1.28948426e+00
-1.95725132e-02 -4.23272729e-01 -2.24000156e-01 5.95185101e-01
4.32002783e-01 3.47124517e-01 2.37912372e-01 1.83623478e-01
-3.78385007e-01 3.20055425e-01 3.98013115e-01 8.55774283e-02
-1.04890794e-01 -6.88888609e-01 -2.96452135e-01 7.56368220e-01
6.45723939e-01 -6.49160922e-01 -4.66569394e-01 -6.34087026e-01
1.17909901e-01 -3.12602907e-01 3.94886196e-01 -1.66461313e+00
7.15491474e-01 1.11175522e-01 4.86722946e-01 -3.00393969e-01
6.85634613e-01 -9.75864708e-01 3.68533760e-01 1.12504220e+00
-8.95698369e-02 1.46791533e-01 5.70390761e-01 3.99829447e-01
3.66471589e-01 -2.69184113e-01 1.11315155e+00 -1.03922728e-02
-7.92578459e-01 1.76120549e-02 -6.12141490e-01 -3.13661039e-01
7.95702934e-01 -7.80735910e-01 -5.37187755e-01 3.34054321e-01
-2.46663570e-01 -2.45255038e-01 5.04461765e-01 1.09447978e-01
3.85779649e-01 -1.02528846e+00 -8.84340778e-02 3.39144886e-01
-3.89570981e-01 -6.57036453e-02 1.28849313e-01 5.43429136e-01
-8.86485219e-01 6.82415903e-01 -8.16148102e-01 -6.44890964e-01
-1.27461159e+00 3.43498915e-01 6.18192852e-01 -1.33963704e-01
-6.37812316e-01 1.12434900e+00 -4.55331892e-01 2.34456018e-01
4.89433914e-01 -7.65756011e-01 2.95206577e-01 -3.35469335e-01
4.63689238e-01 5.95592260e-01 4.50235456e-01 -7.11109266e-02
-5.89180946e-01 4.39713389e-01 2.51597315e-01 3.47117633e-02
1.34286320e+00 4.49941009e-01 9.65320095e-02 -3.45767215e-02
8.64256263e-01 -9.05234337e-01 -1.43837607e+00 3.08443576e-01
9.58262831e-02 -3.81953642e-03 5.84092915e-01 -7.65620232e-01
-1.19679153e+00 8.34274232e-01 1.14112711e+00 -3.60905826e-02
1.45322096e+00 -3.03538680e-01 1.02839649e+00 6.39396489e-01
6.11750185e-01 -1.17950690e+00 -2.66770780e-01 2.72430271e-01
2.61548936e-01 -5.93123496e-01 1.87878281e-01 1.16663426e-02
4.82427254e-02 1.29130650e+00 6.97270036e-01 -3.54510516e-01
5.83421290e-01 1.24180937e+00 -5.13811707e-01 -2.20946088e-01
-9.83113706e-01 6.64668307e-02 -2.18193606e-01 6.60042644e-01
-4.37603630e-02 1.98786729e-03 -2.63559580e-01 4.44191784e-01
-2.34624475e-01 3.00390244e-01 2.95444638e-01 1.18794656e+00
-6.44388914e-01 -8.28853309e-01 -1.13776974e-01 8.35271299e-01
-3.04060638e-01 -1.74619108e-01 -4.92864512e-02 5.56802690e-01
3.60091031e-01 5.99808991e-01 4.65453118e-01 -5.34999311e-01
2.86131591e-01 1.24722719e-01 9.07380164e-01 -8.16432536e-02
-9.43531334e-01 -1.40450507e-01 1.51159897e-01 -7.92165399e-01
-3.31619382e-01 -3.60264242e-01 -1.45518947e+00 -6.31726921e-01
-3.26966107e-01 -4.24240977e-01 1.40588403e+00 9.52715099e-01
5.36975741e-01 7.53135979e-01 -3.01401287e-01 -1.33431244e+00
-5.39302826e-01 -5.53574979e-01 -3.33430231e-01 -3.79717499e-01
2.29021627e-02 -6.61517143e-01 -2.31507480e-01 -8.61856863e-02] | [8.317597389221191, 2.5746407508850098] |
22c25b79-88ef-4067-8635-e89af62ffe63 | medfuse-multi-modal-fusion-with-clinical-time | 2207.07027 | null | https://arxiv.org/abs/2207.07027v2 | https://arxiv.org/pdf/2207.07027v2.pdf | MedFuse: Multi-modal fusion with clinical time-series data and chest X-ray images | Multi-modal fusion approaches aim to integrate information from different data sources. Unlike natural datasets, such as in audio-visual applications, where samples consist of "paired" modalities, data in healthcare is often collected asynchronously. Hence, requiring the presence of all modalities for a given sample is not realistic for clinical tasks and significantly limits the size of the dataset during training. In this paper, we propose MedFuse, a conceptually simple yet promising LSTM-based fusion module that can accommodate uni-modal as well as multi-modal input. We evaluate the fusion method and introduce new benchmark results for in-hospital mortality prediction and phenotype classification, using clinical time-series data in the MIMIC-IV dataset and corresponding chest X-ray images in MIMIC-CXR. Compared to more complex multi-modal fusion strategies, MedFuse provides a performance improvement by a large margin on the fully paired test set. It also remains robust across the partially paired test set containing samples with missing chest X-ray images. We release our code for reproducibility and to enable the evaluation of competing models in the future. | ['Farah E. Shamout', 'Krzysztof J. Geras', 'Nasir Hayat'] | 2022-07-14 | null | null | null | null | ['phenotype-classification'] | ['medical'] | [ 3.04211706e-01 -1.94556087e-01 -7.28939474e-02 -4.03599054e-01
-1.69759572e+00 -4.53697234e-01 4.33029324e-01 5.64589739e-01
-4.71743077e-01 7.93325245e-01 5.73326170e-01 -3.76685917e-01
-3.31703782e-01 -4.24686074e-01 -4.62860525e-01 -6.79363489e-01
-1.57384679e-01 5.06664813e-01 -2.34921306e-01 2.78145343e-01
-5.36151111e-01 1.99316621e-01 -1.43286729e+00 1.01484275e+00
3.62255514e-01 1.11427605e+00 1.78093404e-01 1.06320906e+00
2.11631879e-01 8.85685325e-01 -5.24596691e-01 -1.78005919e-01
-8.07839334e-02 -6.67747557e-01 -6.92526281e-01 -1.08316325e-01
4.01628405e-01 -2.04697713e-01 -4.78717029e-01 5.42690992e-01
1.27455640e+00 7.63534978e-02 5.15485764e-01 -1.23820865e+00
-1.33324638e-01 6.07629836e-01 -3.38919461e-01 3.33436459e-01
6.43215537e-01 1.35917068e-01 7.03488052e-01 -7.51037478e-01
5.20877182e-01 9.69202161e-01 9.13635135e-01 5.24402618e-01
-1.37405133e+00 -3.97626221e-01 -1.39900193e-01 1.60278112e-01
-1.23393142e+00 -5.72504997e-01 4.58333224e-01 -4.71536934e-01
8.18996489e-01 6.02849483e-01 4.89163816e-01 1.40921581e+00
1.43957511e-01 8.28330517e-01 1.11572230e+00 -2.11588353e-01
1.72491565e-01 -6.25189170e-02 -6.20173253e-02 4.95469809e-01
-3.69119972e-01 1.94510907e-01 -9.69553530e-01 -8.01833928e-01
3.78792256e-01 4.27252650e-01 -7.04538882e-01 -1.75277993e-01
-1.90370047e+00 5.49805820e-01 2.63593614e-01 2.84544110e-01
-6.72801852e-01 -2.72817146e-02 5.45097172e-01 5.67785978e-01
3.46006602e-01 3.88976820e-02 -5.67147791e-01 -3.38975370e-01
-1.31784904e+00 1.95501167e-02 7.21808255e-01 4.18637753e-01
1.31758347e-01 -1.81308970e-01 -5.11569500e-01 9.61543143e-01
5.49177937e-02 6.17269039e-01 8.10680807e-01 -1.11640465e+00
4.47741300e-01 3.03407967e-01 -1.09504029e-01 -4.56092983e-01
-7.34724104e-01 -1.80434808e-01 -1.20537913e+00 7.81434029e-02
4.94149178e-01 -3.73512745e-01 -9.11547959e-01 1.87286401e+00
2.78692067e-01 4.30644602e-01 2.03338608e-01 8.59165549e-01
1.38831198e+00 3.49353433e-01 1.03332207e-01 -5.45056939e-01
1.38010538e+00 -3.51220042e-01 -7.54972041e-01 1.35563985e-01
7.18751609e-01 -8.28376591e-01 8.06742907e-01 4.38187748e-01
-1.21103060e+00 -3.58815819e-01 -7.12512374e-01 2.85267830e-01
-9.90355983e-02 1.08197354e-01 1.96087450e-01 3.44626427e-01
-1.09921646e+00 6.95909381e-01 -1.01201403e+00 -3.27489436e-01
3.96101445e-01 3.99490565e-01 -7.94085443e-01 -1.79070622e-01
-1.24633420e+00 7.07974970e-01 2.39509240e-01 -9.41096172e-02
-6.89568162e-01 -9.92600024e-01 -8.79095316e-01 -1.67049598e-02
1.68984532e-01 -1.14115107e+00 1.34581614e+00 -6.46260798e-01
-1.06105208e+00 7.64840126e-01 -1.98602363e-01 -4.71392035e-01
7.36544251e-01 5.83407618e-02 -5.55246174e-01 3.10600519e-01
-2.11293072e-01 7.70283997e-01 6.14641011e-01 -1.13317037e+00
-5.69847822e-01 -3.35434705e-01 -2.34665781e-01 2.34075353e-01
-5.06143235e-02 2.61000115e-02 -8.70981291e-02 -7.22135663e-01
-1.41980693e-01 -7.88960576e-01 -2.08572641e-01 -6.79733083e-02
-3.55411232e-01 1.68498874e-01 6.74411714e-01 -7.39984870e-01
1.15184534e+00 -2.42000484e+00 1.19572788e-01 2.25899275e-02
4.58281219e-01 -3.04333004e-03 -1.23201191e-01 3.78587574e-01
-4.61353153e-01 -9.56808403e-02 -3.15785617e-01 -7.48821437e-01
-1.96150765e-01 2.13528976e-01 -4.27995324e-02 5.11980474e-01
1.18482381e-01 8.81596088e-01 -6.87223732e-01 -7.86383867e-01
4.05180305e-01 6.42732143e-01 -3.82455528e-01 2.33421326e-01
2.10037813e-01 6.79946423e-01 1.91669185e-02 7.54980862e-01
2.48586908e-01 -6.53792381e-01 -2.10102350e-02 -2.15513200e-01
3.27351004e-01 4.11088355e-02 -1.10089993e+00 2.13103127e+00
-4.46318656e-01 5.69743276e-01 2.24948689e-01 -7.78530478e-01
3.79374981e-01 1.08232152e+00 8.92727256e-01 -5.65429211e-01
1.87268093e-01 8.61121416e-02 2.33096957e-01 -5.02499282e-01
8.31977502e-02 -4.53343570e-01 -9.10751373e-02 5.11489153e-01
1.63211524e-01 1.81299508e-01 9.28149500e-04 1.73675716e-01
1.48900199e+00 -3.13217789e-01 1.49005115e-01 1.22862875e-01
2.67612368e-01 -2.11401448e-01 4.96961951e-01 9.66884673e-01
-3.17615360e-01 1.20216036e+00 2.16735095e-01 -1.45593137e-01
-5.95231533e-01 -1.27310884e+00 -3.43420267e-01 9.02788162e-01
-3.00781727e-01 -4.48074818e-01 -2.27506757e-01 -5.66755593e-01
-2.74384003e-02 4.17769492e-01 -7.06079245e-01 -6.39430806e-02
-2.96423942e-01 -9.47892010e-01 7.14373350e-01 6.76866889e-01
-2.48805135e-02 -9.99295056e-01 -8.59449029e-01 2.51448750e-01
-5.54378152e-01 -9.11239147e-01 -2.77505070e-01 4.62144107e-01
-8.41108382e-01 -1.23142087e+00 -1.10515118e+00 -4.53710467e-01
1.07404754e-01 -1.85929582e-01 1.23114812e+00 -2.28522360e-01
-4.83236283e-01 8.06477070e-01 -2.11309031e-01 -3.61045480e-01
-5.05074561e-01 -1.67520568e-01 -1.47623122e-01 7.04721808e-02
3.68973948e-02 -4.86038685e-01 -6.59212530e-01 -5.33747301e-02
-8.90257418e-01 1.58603996e-01 4.29414064e-01 1.17626369e+00
6.64838493e-01 -3.13901603e-01 7.82934427e-01 -6.46937609e-01
5.89958072e-01 -7.34523177e-01 8.28988627e-02 3.39360356e-01
-1.26326814e-01 -2.56368905e-01 4.76697266e-01 -5.71162760e-01
-8.64440620e-01 9.64974388e-02 -1.35055020e-01 -8.20046484e-01
-3.92899245e-01 8.64879251e-01 1.82796761e-01 4.25820053e-01
6.25026584e-01 1.08111098e-01 8.17661881e-02 -5.78732014e-01
1.47511095e-01 7.30070531e-01 8.65806758e-01 -3.65503311e-01
1.54073656e-01 3.94623548e-01 2.64173318e-02 -4.76810634e-01
-3.95242125e-01 -3.72305363e-01 -3.12445909e-01 -2.45185509e-01
7.59038329e-01 -1.11805391e+00 -8.86835575e-01 3.72369736e-01
-9.58681047e-01 -3.10213357e-01 -5.78073502e-01 9.71031070e-01
-6.53823793e-01 1.38608575e-01 -7.53327787e-01 -6.08672380e-01
-4.11875933e-01 -1.27891505e+00 1.27596557e+00 -7.82896504e-02
-4.74903822e-01 -1.08847773e+00 2.75945544e-01 1.73011750e-01
2.92135775e-01 5.81391215e-01 8.49461436e-01 -7.56777525e-01
1.74040738e-02 -2.54118353e-01 7.42332265e-02 1.56130791e-01
3.62596124e-01 -4.63893525e-02 -1.23956203e+00 -3.68867546e-01
8.06257129e-02 -5.76113403e-01 1.00725031e+00 6.33426011e-01
1.26171339e+00 2.53517538e-01 -3.01381052e-01 3.87610972e-01
1.01448917e+00 3.24621171e-01 1.35421142e-01 -2.69001395e-01
4.55393374e-01 4.72584784e-01 3.98019493e-01 6.49747610e-01
4.83579755e-01 2.83968598e-01 2.67506719e-01 -5.27393997e-01
-6.72165677e-02 1.64456740e-01 1.37885854e-01 8.68490219e-01
3.92024368e-01 -4.80214387e-01 -1.15653431e+00 5.83750606e-01
-1.86776638e+00 -9.12732720e-01 -2.83456426e-02 2.34780765e+00
9.39599037e-01 -2.56096482e-01 2.37285182e-01 3.96548241e-01
7.61198699e-01 -2.57063024e-02 -4.44722503e-01 -2.44878270e-02
-2.80763388e-01 3.04344177e-01 1.35668427e-01 3.59764248e-01
-1.15439606e+00 -1.61096558e-01 6.74811554e+00 6.37818217e-01
-1.35943305e+00 4.02440131e-01 7.45810211e-01 -7.91884959e-01
-1.88125700e-01 -4.58916724e-01 4.45058607e-02 6.53449297e-01
1.29824233e+00 3.67507301e-02 1.66742086e-01 1.22206450e-01
1.74672872e-01 -3.39811087e-01 -1.53097904e+00 1.44553316e+00
-1.59368008e-01 -1.36317718e+00 -4.85834599e-01 -1.57626212e-01
4.63887304e-01 5.34417212e-01 1.44650370e-01 1.32113248e-01
1.09133795e-01 -1.14868879e+00 5.01713991e-01 7.33658314e-01
1.08834815e+00 -5.72064757e-01 7.86554396e-01 3.79865170e-01
-8.64677310e-01 -2.36924980e-02 2.63208121e-01 2.56377280e-01
4.08220559e-01 6.57645047e-01 -5.73630810e-01 9.60600138e-01
8.77090096e-01 7.40607560e-01 -6.13267183e-01 1.15528715e+00
6.13809586e-01 6.14967823e-01 -4.14961934e-01 5.97262561e-01
-9.86143425e-02 4.91639346e-01 3.86743367e-01 1.10687482e+00
5.74894190e-01 5.53026497e-02 2.40960687e-01 3.45534593e-01
-7.22401822e-03 -5.55992462e-02 -6.66805923e-01 4.33260240e-02
4.58932847e-01 1.14132202e+00 -4.05469179e-01 -6.51346684e-01
-7.64331222e-01 5.83939552e-01 -9.35845524e-02 3.52273852e-01
-8.75582457e-01 4.61061448e-02 3.00134540e-01 -6.22883290e-02
-4.95636463e-02 3.70449632e-01 -1.68457001e-01 -1.07434583e+00
-1.61835372e-01 -1.00817513e+00 9.42912996e-01 -9.25062954e-01
-1.39226186e+00 7.17369139e-01 5.05670486e-03 -1.47024775e+00
-6.64541304e-01 -2.77296692e-01 -4.06519830e-01 9.64732885e-01
-1.07987463e+00 -9.30689454e-01 -3.11662912e-01 9.75319505e-01
1.03893019e-01 -1.33550793e-01 1.25128007e+00 7.16064990e-01
-4.50529277e-01 4.80240494e-01 1.55958533e-01 -6.86427578e-02
1.14432395e+00 -1.22907245e+00 -2.16848552e-01 4.49161053e-01
9.92610529e-02 2.82738835e-01 6.40287817e-01 -4.86459136e-01
-1.14253521e+00 -9.92662251e-01 6.75909042e-01 -4.91507530e-01
4.07676280e-01 -4.09836583e-02 -9.60281670e-01 5.54679155e-01
4.10932481e-01 3.39678109e-01 1.22821498e+00 2.86719780e-02
-7.41339773e-02 -2.98148468e-02 -1.18759120e+00 1.81223556e-01
6.49792016e-01 -7.24613369e-01 -5.53506196e-01 2.68543810e-01
4.54493105e-01 -5.31303465e-01 -1.48536372e+00 8.29163194e-01
4.05582786e-01 -9.73583400e-01 8.31015348e-01 -5.29743433e-01
3.16493452e-01 -1.89181969e-01 -3.55295062e-01 -1.42842615e+00
1.67507976e-02 -4.76072729e-01 -2.64612883e-01 1.09498215e+00
3.75264585e-01 -6.43941224e-01 5.49720466e-01 4.64874864e-01
-5.32612018e-02 -7.29128242e-01 -1.33462071e+00 -3.08138579e-01
-7.25252554e-02 -7.56061554e-01 4.30493742e-01 1.01605523e+00
1.41451463e-01 1.44752949e-01 -4.64671254e-01 -2.77716387e-02
3.55199903e-01 2.02552021e-01 3.04211646e-01 -1.06880438e+00
-7.14569926e-01 -3.74663651e-01 -2.20968872e-01 -2.48836040e-01
-2.12833956e-02 -9.76763070e-01 -1.11507595e-01 -1.45903027e+00
3.87827486e-01 -3.47865880e-01 -7.22805321e-01 7.35700905e-01
-2.34315231e-01 5.38166642e-01 1.93167329e-01 9.10373628e-02
-5.08568048e-01 4.26854461e-01 1.06853116e+00 -1.87452599e-01
-1.48602337e-01 -2.83292457e-02 -4.92987663e-01 6.13507032e-01
5.45699477e-01 -5.72924852e-01 -5.39717913e-01 -2.50249147e-01
-1.05286404e-01 1.07801759e+00 6.87610686e-01 -1.04446089e+00
1.95067048e-01 1.58995926e-01 6.77695930e-01 -8.25955331e-01
6.95451140e-01 -8.07968259e-01 6.19620621e-01 4.92789805e-01
-4.87372726e-01 3.73812377e-01 3.26446891e-01 5.05880713e-01
-6.08864367e-01 2.93388546e-01 6.05662227e-01 -3.06852758e-02
-1.63988203e-01 1.89014807e-01 -4.85029608e-01 2.07265556e-01
8.70683789e-01 -3.38740461e-02 -4.05821651e-01 -5.19199610e-01
-1.32486761e+00 3.43319774e-01 1.50448918e-01 4.15920079e-01
6.90632522e-01 -1.47909665e+00 -9.42727327e-01 2.26742372e-01
2.71914840e-01 4.09716833e-03 7.19760835e-01 1.48161209e+00
-2.60045584e-02 1.71414107e-01 -6.29870370e-02 -9.81061697e-01
-1.65436780e+00 5.34284055e-01 4.45300639e-01 -1.81648538e-01
-5.58282316e-01 6.25608444e-01 3.43940467e-01 -3.61246645e-01
3.62704277e-01 -4.49181676e-01 1.08626513e-02 3.48310918e-01
5.91504693e-01 2.28345066e-01 4.04683411e-01 -7.23755002e-01
-5.79236686e-01 -3.16172913e-02 1.15768455e-01 -3.49919140e-01
1.22047496e+00 -1.22849435e-01 9.71239880e-02 9.85778511e-01
1.27905083e+00 -4.12851542e-01 -7.91547477e-01 -4.02378142e-01
-3.66206437e-01 -1.44742325e-01 4.17288542e-02 -1.22159457e+00
-1.17193496e+00 1.05306232e+00 9.80666757e-01 1.56843886e-01
1.46559119e+00 9.72808823e-02 4.79595214e-01 1.87677965e-02
7.58957267e-02 -5.83020627e-01 -1.96343482e-01 6.68667704e-02
7.33743548e-01 -1.36906397e+00 -2.72740632e-01 1.97224572e-01
-7.18976557e-01 8.34787846e-01 1.47968337e-01 3.45125705e-01
8.42376292e-01 5.88716865e-01 4.09010738e-01 -1.43890530e-01
-1.31410933e+00 -8.60850960e-02 4.00277108e-01 3.85442615e-01
6.45984650e-01 1.31331459e-01 7.95771778e-02 5.80493569e-01
1.23563774e-01 2.36151591e-01 2.50716805e-01 1.17993224e+00
2.36036435e-01 -1.05913329e+00 -6.53157532e-01 9.17872250e-01
-8.47876310e-01 -1.07411101e-01 -1.18690617e-01 6.35303617e-01
1.75735846e-01 1.02415884e+00 8.50507095e-02 -4.07538772e-01
2.61921078e-01 3.63918543e-01 5.14770567e-01 -3.63001078e-01
-8.45737100e-01 3.94709349e-01 8.12673792e-02 -6.56525254e-01
-6.76095009e-01 -8.58475864e-01 -1.11773407e+00 -2.61128098e-01
-8.56840387e-02 -7.76384771e-02 3.42186421e-01 8.44881654e-01
5.00718236e-01 9.32433367e-01 3.85151714e-01 -8.09345305e-01
-3.72887790e-01 -9.58657086e-01 -5.05534828e-01 4.90837336e-01
8.62893403e-01 -5.87143242e-01 -1.78675309e-01 7.85251111e-02] | [15.021893501281738, -1.806209921836853] |
05b53574-f325-4341-b2ad-bd5f411e011c | learning-rich-features-for-image-manipulation | 1805.04953 | null | http://arxiv.org/abs/1805.04953v1 | http://arxiv.org/pdf/1805.04953v1.pdf | Learning Rich Features for Image Manipulation Detection | Image manipulation detection is different from traditional semantic object
detection because it pays more attention to tampering artifacts than to image
content, which suggests that richer features need to be learned. We propose a
two-stream Faster R-CNN network and train it endto- end to detect the tampered
regions given a manipulated image. One of the two streams is an RGB stream
whose purpose is to extract features from the RGB image input to find tampering
artifacts like strong contrast difference, unnatural tampered boundaries, and
so on. The other is a noise stream that leverages the noise features extracted
from a steganalysis rich model filter layer to discover the noise inconsistency
between authentic and tampered regions. We then fuse features from the two
streams through a bilinear pooling layer to further incorporate spatial
co-occurrence of these two modalities. Experiments on four standard image
manipulation datasets demonstrate that our two-stream framework outperforms
each individual stream, and also achieves state-of-the-art performance compared
to alternative methods with robustness to resizing and compression. | ['Peng Zhou', 'Xintong Han', 'Larry S. Davis', 'Vlad I. Morariu'] | 2018-05-13 | learning-rich-features-for-image-manipulation-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Zhou_Learning_Rich_Features_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Zhou_Learning_Rich_Features_CVPR_2018_paper.pdf | cvpr-2018-6 | ['image-manipulation-detection', 'steganalysis'] | ['computer-vision', 'computer-vision'] | [ 1.07377005e+00 -3.55124652e-01 -8.23296607e-02 -1.54627666e-01
-9.20876622e-01 -3.62421840e-01 4.59335834e-01 -5.45512363e-02
-2.88562089e-01 -8.98310244e-02 1.14265606e-01 -8.52368400e-02
4.31152463e-01 -8.14685404e-01 -9.89912927e-01 -6.94246948e-01
-1.03730805e-01 -5.97712040e-01 4.30520803e-01 -3.18908751e-01
5.88195145e-01 2.25769535e-01 -1.74149883e+00 7.89514244e-01
4.14690077e-01 1.65290916e+00 8.88208821e-02 8.55695665e-01
2.58503854e-01 1.19623828e+00 -5.28618574e-01 -3.38974334e-02
5.37267268e-01 -3.61352593e-01 -4.94862229e-01 1.32836923e-01
4.67862129e-01 -8.48558009e-01 -7.08741486e-01 1.26915145e+00
3.17541569e-01 -1.01832198e-02 2.11463630e-01 -1.28718686e+00
-9.29516673e-01 4.41581279e-01 -8.39059949e-01 3.09342176e-01
4.26419705e-01 5.97441375e-01 3.99814308e-01 -8.04675341e-01
5.94631135e-01 1.28672874e+00 7.64873266e-01 2.81871766e-01
-9.71371770e-01 -9.25412416e-01 -1.49526283e-01 3.71481061e-01
-9.93688464e-01 -5.08884072e-01 9.83742952e-01 7.07818312e-04
8.40331972e-01 1.15096644e-01 2.46046618e-01 1.28312325e+00
1.34057358e-01 8.33487391e-01 1.23615563e+00 -2.51200557e-01
5.61569519e-02 -1.80854529e-01 -3.29699457e-01 8.89790952e-01
7.46379718e-02 1.97944462e-01 -6.83914065e-01 6.93145469e-02
8.86390567e-01 2.70037979e-01 -5.50612092e-01 -1.39124334e-01
-1.39347839e+00 3.85894328e-01 8.32266867e-01 1.56675279e-01
-3.43215525e-01 5.81072390e-01 3.50076824e-01 4.07639652e-01
4.98168059e-02 4.36222970e-01 -3.75034660e-01 -1.11286528e-01
-9.98162270e-01 -4.74891812e-02 2.72680461e-01 8.90984952e-01
8.71705890e-01 -2.41909847e-02 6.30936772e-02 4.88002002e-01
1.75124392e-01 5.29062808e-01 4.06319946e-01 -9.79212105e-01
7.22681642e-01 6.71316266e-01 -4.25692126e-02 -1.26689887e+00
-1.00934930e-01 4.18636240e-02 -8.40422273e-01 5.42464256e-01
4.16819900e-01 2.92051226e-01 -1.28783035e+00 1.29655862e+00
1.48918748e-01 3.01391333e-01 -6.37289360e-02 1.08424544e+00
6.82679832e-01 4.80936408e-01 -1.20705314e-01 3.59475911e-01
1.47736657e+00 -9.24233139e-01 -5.88460863e-01 -3.49952579e-01
3.19997996e-01 -8.34484994e-01 8.76751840e-01 4.26021874e-01
-1.16119146e+00 -5.45270383e-01 -1.37135112e+00 -5.23848295e-01
-8.20201993e-01 -1.18438788e-02 4.80508953e-01 5.51879525e-01
-8.18127573e-01 7.73730814e-01 -7.09690630e-01 -1.08931728e-01
7.69995332e-01 3.20912302e-01 -4.40338463e-01 -4.51229423e-01
-1.08385479e+00 7.38438070e-01 5.31443477e-01 2.72641063e-01
-9.40808833e-01 -5.91145575e-01 -1.21302831e+00 4.30382900e-02
4.69946265e-01 -2.08335072e-01 7.93915987e-01 -1.30910170e+00
-1.36666226e+00 7.26876318e-01 -5.60526224e-03 -1.75804108e-01
6.16878390e-01 4.21538539e-02 -3.93169582e-01 6.19277716e-01
1.15785018e-01 6.19492590e-01 1.56228757e+00 -1.49260652e+00
-6.02068841e-01 -3.80065769e-01 -1.78513899e-01 -1.82969779e-01
-1.98871911e-01 3.23235810e-01 -6.72158301e-01 -8.19121718e-01
3.11029136e-01 -5.59911847e-01 5.78244664e-02 4.68841016e-01
-4.38052773e-01 2.80117214e-01 1.46108985e+00 -9.01273608e-01
9.08625007e-01 -2.18457770e+00 -8.40537548e-02 3.50662112e-01
1.85809284e-01 3.31223488e-01 -5.14932513e-01 5.81586733e-02
-2.69716501e-01 1.48028687e-01 -3.52842003e-01 -1.75462872e-01
-2.87870854e-01 -3.73492360e-01 -4.57745492e-01 8.47767949e-01
7.71184325e-01 1.13244045e+00 -1.14461792e+00 -3.33464563e-01
5.93630135e-01 6.44715846e-01 -1.61698416e-01 1.25453249e-01
-1.51315257e-01 2.76306659e-01 -3.08646798e-01 1.22990763e+00
1.02744806e+00 -2.64472425e-01 -1.97566345e-01 -5.17944098e-01
1.09995641e-01 9.12051052e-02 -8.90784562e-01 1.77409959e+00
-2.95778394e-01 8.10155213e-01 1.67781606e-01 -6.74346626e-01
7.19235182e-01 3.30742523e-02 4.53588009e-01 -9.74307895e-01
6.57918334e-01 2.97985584e-01 -4.77227181e-01 -9.06471133e-01
6.81447864e-01 3.28381985e-01 -1.59139797e-01 4.25324947e-01
-3.59229259e-02 -1.16459370e-01 -3.67173135e-01 6.40109703e-02
1.31125820e+00 2.80099750e-01 -6.33047447e-02 2.86947578e-01
2.69964814e-01 -3.63247126e-01 3.09508145e-01 1.02110910e+00
-4.53038216e-01 1.01906121e+00 4.64293987e-01 -3.50136906e-01
-1.13418961e+00 -9.01762962e-01 3.13490361e-01 9.36487138e-01
7.74759293e-01 -1.89408064e-01 -5.95242858e-01 -7.52909720e-01
6.34274110e-02 3.08112919e-01 -9.06325996e-01 -3.88403684e-01
-8.28326166e-01 -2.96642065e-01 6.34189665e-01 6.43702805e-01
1.01942015e+00 -1.07894456e+00 -1.11677921e+00 -7.09865168e-02
-4.05052185e-01 -1.49182010e+00 -4.05625671e-01 4.87702817e-01
-4.67250884e-01 -1.35835004e+00 -5.08468211e-01 -7.79853582e-01
5.90308726e-01 6.39106035e-01 6.96549952e-01 3.63429576e-01
-5.71796536e-01 2.63523281e-01 -6.17860198e-01 -3.30061764e-01
-3.18209589e-01 -3.78106922e-01 -3.27860236e-01 1.30991951e-01
3.90706360e-01 -1.24413289e-01 -8.66556287e-01 3.27368587e-01
-1.53813696e+00 -4.77198511e-02 5.15830696e-01 6.53439879e-01
2.01442420e-01 4.57355142e-01 -2.85118213e-03 -3.04978818e-01
2.05005020e-01 -4.76920307e-01 -4.59853142e-01 1.60551772e-01
-2.21673042e-01 -8.60577729e-03 6.50375247e-01 -6.74838722e-01
-7.83608735e-01 1.09106287e-01 3.13868821e-01 -7.26032972e-01
-1.89597905e-01 -3.06656677e-03 -1.44829661e-01 -4.13605183e-01
4.93270189e-01 2.96070129e-01 9.81560946e-02 -7.02214167e-02
3.71346444e-01 8.10918510e-01 9.67019081e-01 -1.90721378e-01
1.24292612e+00 1.02039635e+00 -2.05679819e-01 -8.00016105e-01
-3.24596763e-01 -4.98152763e-01 -3.59888077e-01 -3.78157914e-01
8.72197211e-01 -6.87377632e-01 -8.36289465e-01 9.94455278e-01
-1.29161322e+00 -3.61133099e-01 -1.18351802e-01 7.53564835e-02
-6.51953340e-01 5.96362412e-01 -6.48517907e-01 -7.99371839e-01
-3.10446054e-01 -1.33102334e+00 1.53001153e+00 2.82387406e-01
3.17493021e-01 -3.91354829e-01 -8.03179085e-01 5.01803458e-01
6.91465974e-01 5.63541293e-01 8.05105925e-01 -1.73626333e-01
-9.15437818e-01 -4.33174938e-01 -6.94489717e-01 3.88675332e-01
1.88508600e-01 -2.32453048e-01 -1.14908421e+00 -2.05273405e-01
1.77714840e-01 -6.40520751e-01 1.37468517e+00 9.18383803e-03
1.68276024e+00 -1.24136485e-01 -1.02918200e-01 9.40698504e-01
1.53573799e+00 1.42379906e-02 1.17235422e+00 5.90644002e-01
8.61134648e-01 4.99510556e-01 4.37215537e-01 1.79178953e-01
1.59752537e-02 3.52212161e-01 8.68101776e-01 -4.58909929e-01
-2.35545054e-01 -3.23679984e-01 7.34082878e-01 4.46648858e-02
2.20709980e-01 -1.93639457e-01 -6.47563100e-01 3.73658508e-01
-1.60746396e+00 -9.15644705e-01 1.01782359e-01 2.01904011e+00
5.84302485e-01 2.44689398e-02 -1.89473808e-01 3.80523890e-01
7.41677999e-01 5.18369734e-01 -7.23433435e-01 -8.63979012e-02
-3.32116872e-01 2.18517795e-01 1.06543469e+00 -4.73530479e-02
-1.33300090e+00 9.92286026e-01 5.83127880e+00 6.69550657e-01
-1.43365276e+00 -5.30001372e-02 6.63370848e-01 3.64819318e-02
-1.40150944e-02 -4.90859561e-02 -1.40582263e-01 5.60194969e-01
4.44654971e-01 7.33295798e-01 7.68214881e-01 5.28932929e-01
8.56199190e-02 -4.30170745e-01 -8.91834617e-01 1.07367516e+00
5.43671131e-01 -1.20226192e+00 -2.58871108e-01 -1.12374194e-01
6.71624839e-01 7.29283243e-02 4.45311725e-01 -1.04334518e-01
1.44311190e-02 -1.18271625e+00 1.02703834e+00 3.70242119e-01
1.03218448e+00 -4.17414814e-01 7.52193749e-01 -9.57624093e-02
-1.20978618e+00 -4.38645363e-01 -1.03064790e-01 3.53666961e-01
-1.24579892e-01 2.11574078e-01 -6.39112890e-01 2.40145236e-01
1.00414228e+00 8.11324656e-01 -6.33706808e-01 7.31865823e-01
-2.55723685e-01 3.19091320e-01 -3.13651294e-01 2.61809081e-01
2.49506474e-01 3.44053149e-01 3.58154088e-01 1.12802505e+00
7.83217251e-02 -4.42267209e-02 -1.64791077e-01 1.09709144e+00
-3.62600356e-01 -4.56987113e-01 -6.56470060e-01 -4.40534987e-02
2.98326045e-01 1.12969947e+00 -1.00462890e+00 -3.02662879e-01
-5.27767062e-01 1.54595900e+00 -1.67479068e-01 5.30618072e-01
-8.92899930e-01 -8.86746764e-01 5.57042658e-01 1.24654934e-01
8.53152156e-01 -6.76636994e-02 -3.42585236e-01 -1.21410561e+00
2.05817059e-01 -1.14544797e+00 2.09559426e-01 -1.02853596e+00
-1.10441589e+00 1.12878844e-01 -4.68522519e-01 -1.27085769e+00
2.95677543e-01 -8.95030975e-01 -4.32350963e-01 6.59214139e-01
-2.05886483e+00 -1.46582055e+00 -8.78994882e-01 6.33159041e-01
5.32083809e-01 1.94665909e-01 4.23917234e-01 1.01506509e-01
-4.52076495e-01 6.27979398e-01 -1.67111769e-01 7.11481750e-01
7.61337042e-01 -8.71234298e-01 6.67336524e-01 1.20338464e+00
-4.62561250e-01 4.56959307e-01 3.08760524e-01 -7.74085104e-01
-1.84705174e+00 -1.27360272e+00 1.73455060e-01 -3.87510628e-01
6.10417187e-01 -4.45451230e-01 -9.75637019e-01 4.62216318e-01
8.60886946e-02 4.63480055e-01 5.10911923e-03 -6.98005974e-01
-1.08689606e+00 3.43987271e-02 -1.33586121e+00 2.97629118e-01
8.14578474e-01 -8.92783225e-01 -1.90068036e-01 -9.51804221e-02
6.41095996e-01 -4.76963639e-01 -6.71007037e-01 4.86252785e-01
6.94466949e-01 -9.98556197e-01 1.13464093e+00 -3.69763374e-01
8.80223811e-01 -5.58588922e-01 -2.59752810e-01 -7.97869146e-01
-3.04671954e-02 -8.47368181e-01 -7.43276775e-02 8.10854316e-01
-2.15964820e-02 -3.14044476e-01 6.76487982e-01 3.53886068e-01
2.26097256e-01 -5.98266363e-01 -7.30991602e-01 -5.33874750e-01
-2.06368357e-01 -7.31693864e-01 7.84713149e-01 1.02407861e+00
-1.23611107e-01 -2.24802047e-01 -5.69285691e-01 3.97854060e-01
7.05254316e-01 -9.83708724e-02 6.65496945e-01 -5.25725067e-01
-6.08512983e-02 -4.39166725e-01 -7.35220671e-01 -9.57301259e-01
-9.48287919e-02 -6.77876234e-01 3.73296231e-01 -1.08470643e+00
1.76517647e-02 -2.83422291e-01 -4.78165269e-01 6.88591123e-01
-3.50033790e-01 7.96736062e-01 1.73699990e-01 -1.44459466e-02
-7.12774694e-01 3.32136810e-01 9.76553977e-01 -4.52806979e-01
-1.44500509e-01 -7.53751159e-01 -7.73576200e-01 6.30174935e-01
6.48306310e-01 -3.84073138e-01 1.57068312e-01 -6.56731248e-01
1.73000112e-01 -9.23441723e-02 8.59999716e-01 -1.13875031e+00
2.02783138e-01 8.92007910e-03 9.63727057e-01 -4.62881446e-01
2.44039625e-01 -1.01937985e+00 -4.88589585e-01 3.54819834e-01
-4.28138942e-01 -1.62408128e-01 7.05085322e-02 8.46011579e-01
6.83156997e-02 1.92128211e-01 1.06134987e+00 -1.94248363e-01
-7.95277834e-01 9.18233246e-02 -2.19222367e-01 -2.27914199e-01
1.00180435e+00 -6.27844572e-01 -6.52611196e-01 -4.43244725e-01
-1.63994394e-02 4.38080281e-02 8.26655328e-01 8.24865460e-01
1.07508087e+00 -1.12931311e+00 -4.55166042e-01 7.15447187e-01
3.67228448e-01 9.08336267e-02 2.18404561e-01 7.75773883e-01
-5.66972554e-01 -9.94858444e-02 -1.71965688e-01 -6.12101257e-01
-1.19284809e+00 6.43050849e-01 3.75056714e-01 1.46492749e-01
-6.40948176e-01 8.59937429e-01 -3.57093304e-01 6.57013804e-02
4.48912024e-01 -6.09398782e-01 3.46072257e-01 -2.58080930e-01
8.78724337e-01 4.72753376e-01 5.14149107e-02 -7.34404206e-01
-2.76281893e-01 5.31934977e-01 -7.40333423e-02 4.30424809e-02
1.33472085e+00 -8.06209370e-02 -2.13193193e-01 2.82257758e-02
1.75287509e+00 -4.18338627e-01 -1.40096688e+00 -5.02926826e-01
-1.66184932e-01 -1.02023458e+00 2.94791043e-01 -8.26867342e-01
-1.34564507e+00 7.18195379e-01 8.28442574e-01 1.30758792e-01
1.58616781e+00 -1.08768910e-01 1.22377205e+00 1.66993544e-01
2.68624663e-01 -1.20344937e+00 4.94639426e-01 2.30279848e-01
7.08468735e-01 -1.47470140e+00 -3.21724615e-03 -3.49124610e-01
-3.43541116e-01 1.47684455e+00 4.63877201e-01 -2.53284842e-01
3.86176407e-01 4.64549124e-01 1.45847559e-01 -2.25302935e-01
-2.00384140e-01 -2.04547957e-01 -4.11210163e-03 7.25994945e-01
-1.56683370e-01 -2.76197970e-01 5.01257658e-01 1.24050369e-02
5.01368493e-02 9.64478105e-02 4.22745705e-01 1.26812160e+00
-2.47038946e-01 -6.08550727e-01 -7.14053094e-01 5.32006979e-01
-4.24295932e-01 -9.39254090e-02 -3.80712211e-01 4.52569693e-01
4.81613100e-01 1.25344086e+00 1.02522045e-01 -6.73068404e-01
1.78223938e-01 -2.46376052e-01 5.33706248e-01 -1.43342584e-01
-8.18261385e-01 -1.33075401e-01 -3.90097260e-01 -1.34847903e+00
-3.08425695e-01 -4.43101227e-01 -1.02132213e+00 -1.86793551e-01
-6.04181409e-01 -7.14629710e-01 1.16974843e+00 7.19198287e-01
2.59950697e-01 5.28426707e-01 8.92118752e-01 -1.29466045e+00
-4.16701317e-01 -7.24945366e-01 -3.30435455e-01 7.87190497e-01
1.07431591e+00 -3.72186035e-01 -5.43968499e-01 1.91953868e-01] | [12.281764030456543, 0.927366316318512] |
ab021239-4c50-43ca-b36f-9feccc35aa71 | towards-open-set-text-recognition-via-label | 2203.05179 | null | https://arxiv.org/abs/2203.05179v3 | https://arxiv.org/pdf/2203.05179v3.pdf | Towards Open-Set Text Recognition via Label-to-Prototype Learning | Scene text recognition is a popular topic and extensively used in the industry. Although many methods have achieved satisfactory performance for the close-set text recognition challenges, these methods lose feasibility in open-set scenarios, where collecting data or retraining models for novel characters could yield a high cost. For example, annotating samples for foreign languages can be expensive, whereas retraining the model each time when a novel character is discovered from historical documents costs both time and resources. In this paper, we introduce and formulate a new open-set text recognition task which demands the capability to spot and recognize novel characters without retraining. A label-to-prototype learning framework is also proposed as a baseline for the proposed task. Specifically, the framework introduces a generalizable label-to-prototype mapping function to build prototypes (class centers) for both seen and unseen classes. An open-set predictor is then utilized to recognize or reject samples according to the prototypes. The implementation of rejection capability over out-of-set characters allows automatic spotting of unknown characters in the incoming data stream. Extensive experiments show that our method achieves promising performance on a variety of zero-shot, close-set, and open-set text recognition datasets | ['Cheng-Lin Liu', 'Xu-Cheng Yin', 'Xiaobin Zhu', 'Hai-Bo Qin', 'Chun Yang', 'Chang Liu'] | 2022-03-10 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 5.56669235e-01 -4.74773467e-01 -1.98855370e-01 -7.53698409e-01
-8.68052840e-01 -5.61604559e-01 6.68112040e-01 2.27580398e-01
-4.68036979e-01 3.03228527e-01 -2.82658041e-01 -5.80007806e-02
2.04154506e-01 -5.36360741e-01 -4.63327289e-01 -6.88045144e-01
3.50000769e-01 8.70358944e-01 3.46813023e-01 -1.65304095e-02
4.29546446e-01 5.14705420e-01 -1.89540398e+00 5.76361179e-01
9.38363194e-01 1.02315855e+00 3.61084372e-01 6.66953921e-01
-5.92466593e-01 2.74188071e-01 -7.45373607e-01 -2.94462502e-01
2.98654318e-01 -2.34479055e-01 -5.50013542e-01 6.47614002e-01
5.22851110e-01 -1.21513821e-01 -2.40837514e-01 8.57025921e-01
4.60479826e-01 4.96832013e-01 7.68208683e-01 -1.16841686e+00
-7.03990519e-01 4.75600451e-01 -4.62332398e-01 1.14255317e-01
3.38323206e-01 -6.96241185e-02 7.12609589e-01 -1.39103425e+00
3.57360840e-01 9.12893593e-01 5.15062869e-01 7.46519029e-01
-9.52142060e-01 -6.03010654e-01 5.82064152e-01 4.52133656e-01
-1.50572324e+00 -7.50467539e-01 5.59640765e-01 -2.82808870e-01
7.30903864e-01 5.71405709e-01 2.95101345e-01 1.28548729e+00
-2.82961756e-01 1.23805952e+00 6.57539308e-01 -6.68891966e-01
3.81676495e-01 5.94970703e-01 6.20910168e-01 3.79902035e-01
6.84834421e-02 -4.49775040e-01 -4.65582192e-01 -1.14018060e-01
1.19812742e-01 4.94215935e-01 -3.72224659e-01 -2.36717418e-01
-9.78373051e-01 6.62233651e-01 -8.32783896e-03 1.67222410e-01
-1.24984467e-02 -5.47729731e-01 5.48509896e-01 1.82429001e-01
6.00368381e-01 -6.43640682e-02 -1.80115506e-01 -1.11349612e-01
-1.04199982e+00 -2.34542280e-01 9.06919897e-01 1.19773817e+00
6.82550609e-01 -5.72700379e-03 -3.45144033e-01 1.29457676e+00
5.60597330e-02 4.74430412e-01 8.35296750e-01 -4.98112813e-02
5.36540568e-01 6.41208291e-01 -3.96024138e-02 -6.46527708e-01
-1.19098976e-01 -3.31746668e-01 -6.74008250e-01 -2.93360949e-01
1.14361674e-01 1.04893014e-01 -1.04262054e+00 1.03051484e+00
4.38222438e-01 4.26774770e-01 2.94475734e-01 8.74720216e-01
6.34871244e-01 1.07017183e+00 -2.94113725e-01 -4.00235116e-01
1.11538410e+00 -1.14572990e+00 -6.33299232e-01 -4.37737435e-01
5.68100989e-01 -7.53850996e-01 1.04843807e+00 3.96074474e-01
-4.15672034e-01 -4.81354386e-01 -9.10903037e-01 1.56109422e-01
-5.24369657e-01 2.30270699e-01 1.29237413e-01 7.14700818e-01
-5.10257304e-01 2.23068863e-01 -4.95419413e-01 -6.77700043e-01
3.26961845e-01 1.01005010e-01 -6.50171041e-02 -3.76571089e-01
-6.82916760e-01 5.71071684e-01 5.03767073e-01 2.81399399e-01
-7.25010514e-01 -2.49303535e-01 -7.42355168e-01 1.46121994e-01
5.59026659e-01 -7.33314306e-02 1.23061955e+00 -1.27688050e+00
-1.42307317e+00 7.48519182e-01 -1.93496823e-01 -2.74936467e-01
5.61888754e-01 -3.76293696e-02 -8.96468222e-01 -4.03837487e-02
3.91386449e-02 3.31385553e-01 1.27276611e+00 -1.09727323e+00
-8.93975794e-01 -3.47602576e-01 -5.75194418e-01 4.34858471e-01
-9.02670920e-01 5.61685227e-02 -7.51666665e-01 -7.78179049e-01
3.63766909e-01 -7.88158953e-01 1.04544841e-01 -3.10118805e-04
-4.65640277e-01 -4.20488536e-01 1.25812948e+00 -2.49549985e-01
1.18037868e+00 -2.21404386e+00 -2.36663520e-01 1.70628309e-01
-2.57394999e-01 4.36479807e-01 -1.66305155e-01 3.18074822e-01
1.20520636e-01 -2.55947083e-01 -1.69527084e-01 -4.05993521e-01
4.18237746e-02 2.27982208e-01 -8.01824272e-01 3.75201732e-01
1.81144118e-01 5.08625746e-01 -6.37597024e-01 -3.74645025e-01
1.81648627e-01 1.05412222e-01 -2.26807997e-01 2.62248844e-01
-3.27942640e-01 -8.57390538e-02 -3.38682473e-01 6.88789427e-01
5.76101840e-01 -3.20024192e-01 -3.80515270e-02 2.52482623e-01
-1.10842682e-01 -3.02799523e-01 -1.54454660e+00 1.39452410e+00
-6.26718327e-02 6.11562073e-01 -2.66327888e-01 -1.15471768e+00
1.28647268e+00 7.43778497e-02 9.07543153e-02 -4.88651514e-01
6.15336820e-02 2.14290738e-01 -2.36522719e-01 -5.70816696e-01
8.95944953e-01 1.58091541e-02 -1.70158476e-01 6.23139679e-01
-8.57748650e-03 2.00176984e-01 3.57251763e-01 1.06105410e-01
8.10583949e-01 -3.94585952e-02 -3.26026380e-02 1.14911586e-01
7.14455366e-01 2.57146537e-01 5.35751343e-01 9.31040168e-01
-1.79114968e-01 6.81861281e-01 -1.62297204e-01 -7.05219746e-01
-8.73695612e-01 -7.03116536e-01 -3.57530206e-01 1.46503496e+00
2.55157053e-01 -1.38298690e-01 -5.37164986e-01 -8.05401564e-01
-3.92688112e-03 7.34217048e-01 -4.14514959e-01 -6.27912804e-02
-3.00444305e-01 -5.34818530e-01 4.56851900e-01 3.08714449e-01
1.43176332e-01 -1.00311220e+00 -4.88527596e-01 1.91197500e-01
-2.29158893e-01 -9.29344594e-01 -7.81015694e-01 3.63653988e-01
-6.02010071e-01 -9.32944238e-01 -7.86838889e-01 -1.25517535e+00
9.11171138e-01 6.25705302e-01 3.85914594e-01 -8.63626450e-02
-5.17458916e-01 3.67845207e-01 -7.84710765e-01 -5.12954235e-01
-4.61889952e-01 1.67522900e-04 2.46329188e-01 8.27357233e-01
6.43991470e-01 2.84748584e-01 -9.47454646e-02 6.89756572e-01
-9.71597135e-01 1.60171375e-01 2.65280068e-01 1.00570714e+00
7.36167312e-01 -7.75014013e-02 6.77161396e-01 -8.18360150e-01
5.18110812e-01 -3.81327808e-01 -4.89847511e-01 6.08877301e-01
-6.04671061e-01 -1.80247590e-01 9.18335080e-01 -1.00070441e+00
-1.04193628e+00 1.76943213e-01 2.07512811e-01 -7.39433348e-01
-3.32298636e-01 3.60403180e-01 -4.29308526e-02 1.89622819e-01
6.71249449e-01 7.45087147e-01 -2.15550303e-01 -4.77851510e-01
1.58411697e-01 1.40824831e+00 5.17792642e-01 -4.16149855e-01
7.10158348e-01 5.79154253e-01 -7.89354503e-01 -1.10033715e+00
-9.04331088e-01 -9.35879350e-01 -7.05840230e-01 -2.84678906e-01
3.31108958e-01 -8.84001434e-01 -2.28590518e-01 5.94286859e-01
-8.58972013e-01 -3.77347544e-02 -5.55273235e-01 3.76524895e-01
-2.45538622e-01 5.74750364e-01 -2.70982563e-01 -8.86816204e-01
-5.33124506e-01 -1.03916800e+00 1.05734468e+00 3.88487965e-01
1.04346201e-02 -7.27446258e-01 -1.04642719e-01 2.39558324e-01
1.16363548e-01 -3.63990873e-01 6.94316745e-01 -1.45739150e+00
-2.99106568e-01 -6.62155569e-01 2.61658500e-03 3.23384732e-01
7.38692135e-02 -5.58703244e-02 -1.13564217e+00 -5.13857007e-01
-5.59968203e-02 -5.93171179e-01 8.05106580e-01 -8.66794586e-02
1.25012541e+00 -2.35008121e-01 -4.69424903e-01 5.86066604e-01
1.01423192e+00 4.73450840e-01 3.68701965e-01 2.65863985e-01
6.07991278e-01 5.60216129e-01 7.99575090e-01 1.02136159e+00
7.84113035e-02 4.61168826e-01 -1.63801789e-01 2.79837865e-02
2.28928924e-01 -2.24031776e-01 3.64663213e-01 8.12416196e-01
8.23001921e-01 -7.67814100e-01 -9.26814318e-01 4.63860542e-01
-1.91628814e+00 -9.49485898e-01 -2.04868078e-01 2.31433511e+00
7.10544467e-01 1.12849027e-01 -1.95316046e-01 2.49970034e-01
1.24020016e+00 -5.81025928e-02 -9.32797909e-01 -2.17692211e-01
-3.20239395e-01 -4.93794344e-02 7.52571449e-02 9.67594534e-02
-1.13079441e+00 9.69052374e-01 5.69768715e+00 9.19181168e-01
-1.25547647e+00 -1.04113698e-01 6.55975819e-01 -1.69172212e-01
4.95314971e-02 2.96655744e-02 -1.27256417e+00 4.46877390e-01
7.42105544e-01 -2.41926730e-01 2.05825746e-01 9.31194425e-01
-8.89751874e-03 8.55582207e-03 -1.44758987e+00 1.36506808e+00
9.14268076e-01 -1.18773103e+00 2.86731392e-01 -3.55166048e-01
6.11620069e-01 -5.11450395e-02 4.40739021e-02 6.04184389e-01
-1.69614717e-01 -5.38121521e-01 7.94938505e-01 4.90189165e-01
8.06134045e-01 -6.41595542e-01 3.67251486e-01 9.00908291e-01
-1.01397848e+00 -2.34761238e-01 -6.36508107e-01 7.08602369e-02
-9.06430632e-02 3.44093949e-01 -1.16976047e+00 2.10662201e-01
4.87936229e-01 7.90541410e-01 -6.71965778e-01 1.21468139e+00
7.52705410e-02 6.29402459e-01 -4.10245478e-01 -4.15448636e-01
1.64785460e-01 -3.24941814e-01 4.06564295e-01 1.34401071e+00
2.67694741e-01 8.98717344e-02 6.17435694e-01 4.91560489e-01
-1.04216225e-01 4.42201108e-01 -4.67600197e-01 5.61017962e-03
5.11816263e-01 1.19129026e+00 -9.03284311e-01 -6.96143389e-01
-4.22263503e-01 1.37276185e+00 2.56493628e-01 3.58234614e-01
-7.45571256e-01 -6.70582712e-01 2.78519005e-01 -1.24117829e-01
4.42129195e-01 2.47728512e-01 -7.87243098e-02 -1.57231200e+00
2.64690697e-01 -7.59165466e-01 7.40503132e-01 -6.04636133e-01
-1.45937490e+00 4.27009046e-01 -3.20844948e-01 -1.43269837e+00
-1.57708645e-01 -5.37884295e-01 -6.23440444e-01 6.50784075e-01
-1.36453700e+00 -7.90092707e-01 -4.35227454e-01 7.21526086e-01
1.36893880e+00 -4.34071720e-01 7.70267963e-01 2.15192899e-01
-1.03726196e+00 8.27632785e-01 8.06265652e-01 2.88815677e-01
9.30690229e-01 -7.70068526e-01 3.89567554e-01 1.00128698e+00
3.73651147e-01 2.59436250e-01 4.49853301e-01 -6.99086487e-01
-1.69165313e+00 -1.35872972e+00 7.49080956e-01 -3.53798270e-01
5.20900071e-01 -7.88690209e-01 -1.23770082e+00 6.61880076e-01
-2.23468825e-01 3.19441736e-01 8.30118418e-01 2.14632768e-02
-3.55660647e-01 -7.99103379e-02 -9.24958050e-01 5.92997491e-01
6.88274682e-01 -3.01661789e-01 -5.91641665e-01 6.14518464e-01
4.16312397e-01 -2.96020746e-01 -3.43831331e-01 -1.03811614e-01
4.24755663e-01 -5.30570924e-01 5.18215895e-01 -5.60230255e-01
-1.65389359e-01 -3.02936524e-01 -3.53359580e-01 -1.06626999e+00
-2.38491222e-01 -4.93591934e-01 -9.84228030e-02 1.45863068e+00
3.20597261e-01 -5.39233029e-01 7.93308675e-01 6.13820493e-01
-3.37491006e-01 -4.39951509e-01 -9.67756867e-01 -1.00187624e+00
-2.91795462e-01 -5.47414124e-01 2.73164928e-01 1.08414102e+00
1.72644213e-01 5.17094433e-01 -3.85219693e-01 2.24315211e-01
5.11109889e-01 4.34367329e-01 9.65912879e-01 -1.25054538e+00
-1.25611499e-01 -1.84539080e-01 -4.23249811e-01 -1.40371203e+00
3.35391641e-01 -1.13259923e+00 4.79310691e-01 -1.27142763e+00
4.51465458e-01 -7.29326546e-01 -2.01366097e-01 4.83115882e-01
-1.39685169e-01 -8.24773535e-02 2.65275121e-01 5.96724093e-01
-9.50007737e-01 7.06001878e-01 6.36806250e-01 -4.40234691e-01
-3.28936964e-01 1.76983193e-01 -4.08833772e-01 6.45997703e-01
4.75547075e-01 -5.33342540e-01 -2.56913424e-01 -4.44469362e-01
-3.43502820e-01 1.77772008e-02 -3.00610941e-02 -1.01194835e+00
7.07421184e-01 -2.73656905e-01 5.41376710e-01 -9.55006301e-01
3.71782899e-01 -9.02867198e-01 -7.26677850e-02 2.41574228e-01
-4.47169244e-01 -2.42965862e-01 7.72795007e-02 8.98971498e-01
1.59978028e-02 -7.16242909e-01 7.53123701e-01 2.21816927e-01
-1.11122668e+00 4.49625373e-01 -4.32406813e-01 8.63942876e-02
1.38508332e+00 -7.98051775e-01 -2.83483386e-01 -6.81207776e-02
-5.15467882e-01 3.92193466e-01 3.80128205e-01 8.86795759e-01
9.51370180e-01 -9.86765027e-01 -7.87484109e-01 6.06846452e-01
7.47197211e-01 -7.29747564e-02 2.33768508e-01 3.78674179e-01
-2.36544967e-01 2.42425159e-01 3.25714290e-01 -9.74220753e-01
-1.29216790e+00 7.99196541e-01 1.38930157e-01 6.90828338e-02
-7.82858253e-01 8.42047274e-01 2.39562109e-01 -5.54209650e-01
6.58935428e-01 -6.40441403e-02 -2.07556784e-01 1.72947332e-01
7.97733665e-01 2.43165076e-01 2.17378870e-01 -7.56535888e-01
-2.04614773e-01 2.24306747e-01 -6.43776715e-01 1.15596831e-01
1.19863403e+00 -2.12638915e-01 3.30556333e-01 9.73885357e-01
9.94146764e-01 -4.42870915e-01 -1.27294910e+00 -6.96966112e-01
5.75430281e-02 -5.28871238e-01 -2.21020415e-01 -6.67527318e-01
-4.30590928e-01 7.51596510e-01 8.68585169e-01 -2.05610581e-02
8.49868596e-01 -2.14449316e-01 7.67177343e-01 9.29320633e-01
3.50363314e-01 -1.46879768e+00 1.94086835e-01 8.59138966e-01
5.73995590e-01 -1.45870519e+00 -1.84459418e-01 -1.90822780e-01
-7.59286106e-01 1.31378269e+00 8.17975283e-01 3.21423203e-01
5.24944961e-01 8.01938623e-02 1.09866783e-01 1.97600186e-01
-8.80605698e-01 3.23430002e-02 1.52021930e-01 4.51270491e-01
2.01509036e-02 1.00019947e-01 1.95195064e-01 4.53294724e-01
2.78860629e-01 -3.24224144e-01 6.23890460e-01 1.27826273e+00
-8.88230681e-01 -8.67499590e-01 -7.23994851e-01 7.89251268e-01
2.31105667e-02 -3.42077576e-02 -3.24506223e-01 2.49412864e-01
-1.01563685e-01 1.02060807e+00 4.87499207e-01 -2.95831084e-01
4.57731515e-01 4.52222973e-01 4.54992838e-02 -1.04489386e+00
-4.15524453e-01 1.92395166e-01 -1.75773934e-01 1.80480391e-01
-3.75176147e-02 -7.98969626e-01 -1.26142836e+00 9.57081169e-02
-9.14695144e-01 2.10123673e-01 7.32365608e-01 9.30641711e-01
3.00669163e-01 1.27245009e-01 9.86119747e-01 -6.46801293e-01
-8.36303055e-01 -8.55439544e-01 -7.96105802e-01 6.23375773e-01
9.53208059e-02 -3.29983979e-01 -1.97538853e-01 4.65943128e-01] | [10.219867706298828, 3.432596445083618] |
67fe9bd0-f495-41e6-9de4-5b84656d3256 | recurrent-color-constancy | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Qian_Recurrent_Color_Constancy_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Qian_Recurrent_Color_Constancy_ICCV_2017_paper.pdf | Recurrent Color Constancy | We introduce a novel formulation of temporal color constancy which considers multiple frames preceding the frame for which illumination is estimated. We propose an end-to-end trainable recurrent color constancy network -- the RCC-Net -- which exploits convolutional LSTMs and a simulated sequence to learn compositional representations in space and time. We use a standard single frame color constancy benchmark, the SFU Gray Ball Dataset, which can be adapted to a temporal setting. Extensive experiments show that the proposed method consistently outperforms single-frame state-of-the-art methods and their temporal variants.
| ['Joni-Kristian Kamarainen', 'Yanlin Qian', 'Ke Chen', 'Jiri Matas', 'Jarno Nikkanen'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['color-constancy'] | ['computer-vision'] | [ 7.15970919e-02 -6.68920815e-01 -1.28010795e-01 -4.56062824e-01
-6.45676792e-01 -4.36664194e-01 6.29948974e-01 -6.72697127e-01
-6.34107769e-01 6.55662298e-01 -2.51371376e-02 -3.58858138e-01
4.92778182e-01 -2.64139920e-01 -9.82938945e-01 -8.63966167e-01
-1.99946329e-01 -3.88176918e-01 4.48383182e-01 -1.87227920e-01
1.69403017e-01 3.14375818e-01 -1.37187982e+00 2.84906715e-01
4.53727484e-01 1.23767376e+00 5.48400059e-02 1.10032320e+00
1.52890265e-01 1.40103889e+00 -3.02285463e-01 -1.97225422e-01
4.32835937e-01 -9.92336631e-01 -5.79100907e-01 1.10245705e-01
9.27002549e-01 -5.07539570e-01 -7.23841310e-01 8.83967102e-01
4.24937278e-01 6.72291517e-01 7.08580017e-02 -1.19231510e+00
-1.29847789e+00 6.84446096e-02 -5.98497689e-01 6.00802779e-01
1.11616842e-01 3.95140380e-01 9.45571542e-01 -7.48337507e-01
7.51440406e-01 1.32943988e+00 6.06650472e-01 1.00443864e+00
-1.11597311e+00 -2.95806438e-01 5.22831857e-01 5.17050087e-01
-7.53496826e-01 -5.80537736e-01 7.76968122e-01 5.45014516e-02
1.15483975e+00 1.54759571e-01 9.22247887e-01 1.07022536e+00
4.53719079e-01 1.00562930e+00 1.53988278e+00 -3.89051229e-01
2.62201548e-01 -8.97166014e-01 -2.70027965e-01 1.03077936e+00
-2.56503612e-01 5.65280557e-01 -8.18485022e-01 4.29579377e-01
1.16789150e+00 -1.75817192e-01 -3.80176276e-01 -3.61671060e-01
-1.36786759e+00 3.53202820e-01 8.71887088e-01 1.60307780e-01
-1.20850861e-01 1.19871640e+00 5.88883936e-01 3.57949406e-01
8.46574068e-01 -9.69150104e-03 -5.34523368e-01 -3.06057811e-01
-9.25543904e-01 5.92217669e-02 2.10868955e-01 9.29668903e-01
6.21164382e-01 8.40311408e-01 -5.53830683e-01 4.75710064e-01
1.54889254e-02 3.89903426e-01 5.50731897e-01 -1.39896381e+00
-9.51885283e-02 -1.03500240e-01 3.50045145e-01 -4.39767331e-01
-3.32760155e-01 -1.10110398e-02 -6.55864656e-01 5.85585773e-01
3.48778903e-01 1.61010679e-02 -1.43291688e+00 1.98162997e+00
2.47585386e-01 9.27579165e-01 -7.73169100e-03 1.33747518e+00
7.83503234e-01 5.73810816e-01 1.85639888e-01 -3.83552074e-01
8.89753580e-01 -1.36689210e+00 -8.76587212e-01 -1.77916333e-01
-1.77669115e-02 -7.36202955e-01 1.02296901e+00 2.37129882e-01
-1.43484294e+00 -7.47339189e-01 -1.11178720e+00 -4.56372768e-01
-3.26911271e-01 3.52949166e-04 8.67914796e-01 5.85173905e-01
-1.66430950e+00 7.63342857e-01 -1.13252807e+00 -1.64647982e-01
2.01712236e-01 -2.37846628e-01 -9.90048349e-02 -2.26830706e-01
-1.13907373e+00 9.56762314e-01 1.94317281e-01 5.65779626e-01
-1.10785282e+00 -6.89864457e-01 -1.03634453e+00 -4.10728782e-01
2.38195993e-02 -9.09031451e-01 1.65748453e+00 -1.86224544e+00
-2.07287955e+00 8.83996665e-01 -6.48806751e-01 -5.51558673e-01
8.00872028e-01 -3.07057410e-01 -4.72057253e-01 1.65175349e-01
-9.82853398e-02 6.49665892e-01 1.10157919e+00 -1.30537629e+00
-7.09120154e-01 1.01111934e-01 2.20712870e-01 1.38459504e-01
4.90081012e-01 2.44235486e-01 -7.79054284e-01 -8.28778923e-01
7.92845562e-02 -1.02457654e+00 -6.55186027e-02 5.41805208e-01
-2.10573465e-01 -4.81471457e-02 9.29063082e-01 -5.86855114e-01
6.40498817e-01 -1.89461970e+00 1.89315513e-01 -3.10589433e-01
4.91768755e-02 -5.43515906e-02 -4.35326964e-01 -2.91322142e-01
-3.78666162e-01 -2.97710896e-01 -8.43292624e-02 -8.07422161e-01
1.13375550e-02 4.59204406e-01 -3.11091602e-01 9.75507140e-01
1.43245429e-01 1.17279315e+00 -1.19603753e+00 -3.38006228e-01
5.78365862e-01 6.29235923e-01 -2.10215315e-01 4.01823223e-01
-6.35883391e-01 4.73765612e-01 2.30120212e-01 7.16857374e-01
9.09598529e-01 -1.59703687e-01 -3.88414897e-02 -3.15727353e-01
-4.87164855e-01 8.89091939e-02 -6.66993976e-01 2.25925326e+00
-4.73388165e-01 1.27981794e+00 -4.30821270e-01 -5.48556864e-01
3.86897504e-01 1.68498933e-01 4.55458879e-01 -1.26388812e+00
1.21480495e-01 7.59205446e-02 -2.85358310e-01 -2.53113240e-01
7.96934247e-01 -9.44743454e-02 1.78319022e-01 3.50305319e-01
7.04658180e-02 -3.47893000e-01 3.47018778e-01 -1.82063114e-02
7.45089531e-01 1.10371256e+00 -2.17008833e-02 -1.43569425e-01
2.74000734e-01 -3.75772864e-01 6.86193466e-01 6.36957765e-01
-7.76328444e-01 8.52205932e-01 2.16274291e-01 -7.91673541e-01
-1.10157073e+00 -1.31890345e+00 2.03327194e-01 1.52881992e+00
4.73440617e-01 1.11845978e-01 -3.34529638e-01 -2.26843506e-01
-3.48059863e-01 7.47801244e-01 -1.17460454e+00 -1.87631156e-02
-6.71571732e-01 -4.53443140e-01 2.20717236e-01 6.23064995e-01
8.71600807e-01 -9.98273909e-01 -1.16223049e+00 6.56485483e-02
-1.90195084e-01 -1.28292632e+00 -8.97072613e-01 2.88916200e-01
-5.96845090e-01 -1.07612491e+00 -8.92350316e-01 -7.42564499e-01
1.47967666e-01 5.92167974e-01 1.46481943e+00 -7.54271448e-02
-2.17888638e-01 5.96348107e-01 -3.96854699e-01 4.16618213e-03
-7.16757700e-02 -5.32990575e-01 -3.10718656e-01 5.26994094e-02
-9.21537131e-02 -1.11834161e-01 -1.11072695e+00 1.10921644e-01
-1.01716948e+00 4.35768902e-01 6.73085675e-02 8.84485066e-01
7.55320430e-01 -6.78837180e-01 4.45702719e-03 -4.96818215e-01
1.34921372e-01 1.42209157e-01 -7.11838365e-01 6.09457314e-01
-3.12898844e-01 2.35483989e-01 3.97894859e-01 -4.63514954e-01
-1.38791800e+00 1.28896028e-01 2.73955882e-01 -8.46626222e-01
3.79775584e-01 2.63879187e-02 4.02649939e-01 -4.53383684e-01
4.69037205e-01 4.70495284e-01 -2.99460977e-01 2.53661424e-01
9.63613153e-01 -2.99890280e-01 1.26341963e+00 -6.43575072e-01
5.98016143e-01 8.61964583e-01 1.06777094e-01 -3.46337318e-01
-9.63015914e-01 -3.09420943e-01 -7.41355896e-01 -6.05518937e-01
1.01661956e+00 -1.07959425e+00 -8.98363769e-01 9.40862894e-01
-1.38506758e+00 -1.12158406e+00 -2.00424910e-01 1.87815443e-01
-8.99856508e-01 3.63468707e-01 -8.64519715e-01 -6.87638164e-01
-3.93181860e-01 -8.56754959e-01 1.08191931e+00 6.07457578e-01
3.59497637e-01 -1.12903440e+00 2.91138053e-01 -4.78919029e-01
6.82227492e-01 6.37972593e-01 4.69720513e-01 6.01101220e-01
-7.92797804e-01 3.46607238e-01 -6.29334807e-01 2.52525926e-01
3.26821983e-01 4.55404907e-01 -1.14207482e+00 -3.64761800e-01
-3.38957936e-01 -5.23892522e-01 1.35253501e+00 8.52694988e-01
1.13204861e+00 2.69439910e-02 3.40565771e-01 1.04483044e+00
1.68597639e+00 1.33529976e-01 8.82494390e-01 5.34114540e-01
8.43058109e-01 -1.53143391e-01 5.79465330e-01 4.31765169e-01
4.61044192e-01 5.44529915e-01 6.26335204e-01 -6.12956107e-01
-6.43485308e-01 4.47600372e-02 5.26016831e-01 3.87634337e-01
-3.91436428e-01 -1.31601587e-01 -1.63206175e-01 5.06513178e-01
-2.08525729e+00 -1.32647097e+00 8.55778903e-02 1.84062898e+00
1.11944199e+00 -1.89415485e-01 9.63032395e-02 -2.18530402e-01
5.50944626e-01 7.55287111e-01 -9.17309046e-01 -5.60166061e-01
-6.02245808e-01 3.35012496e-01 7.63728917e-01 6.09518349e-01
-1.27720666e+00 1.35477841e+00 7.38185883e+00 2.60016829e-01
-1.53079164e+00 2.23107308e-01 9.58954871e-01 -3.69807482e-01
-3.15843821e-01 -1.38698280e-01 2.73729246e-02 2.61006445e-01
9.65111732e-01 2.00920273e-02 8.14692020e-01 6.16681755e-01
4.96530920e-01 -3.45162898e-01 -1.16330290e+00 1.08714437e+00
4.57586110e-01 -1.39506364e+00 -2.06257164e-01 -8.83285642e-01
1.16140389e+00 2.96810299e-01 6.46241128e-01 4.14855443e-02
6.58525348e-01 -8.36255431e-01 1.35977638e+00 7.89325058e-01
1.25265467e+00 -4.02096212e-01 -5.08533530e-02 -7.21852064e-01
-1.53083873e+00 2.62724429e-01 -4.10162657e-01 2.25234464e-01
3.00744444e-01 1.90360457e-01 -6.78250268e-02 3.81508768e-01
1.01271546e+00 1.19339192e+00 -6.95043504e-01 1.02206409e+00
-3.93567115e-01 2.03949809e-01 4.18343432e-02 9.26882774e-02
5.31752825e-01 -6.59674928e-02 -1.95037983e-02 1.42980886e+00
1.87515661e-01 4.05583763e-03 -1.39816418e-01 9.09201086e-01
-1.42387440e-02 -7.16462195e-01 -1.43855348e-01 3.21181923e-01
7.05925748e-02 1.06570387e+00 -5.51523685e-01 -3.89107198e-01
-7.40837276e-01 1.89514565e+00 2.97588944e-01 1.00638223e+00
-1.25729907e+00 -1.83488816e-01 9.26713526e-01 -7.79887259e-01
4.36731040e-01 -5.42298317e-01 -1.80367514e-01 -1.40977263e+00
-3.17701012e-01 -5.60323238e-01 4.09766883e-01 -1.56230891e+00
-1.23490584e+00 6.89647794e-01 -3.43410790e-01 -1.21802676e+00
-4.06837195e-01 -7.76433587e-01 -5.88972211e-01 8.31384122e-01
-2.25793242e+00 -1.51064551e+00 -6.82420850e-01 1.07311296e+00
6.33503139e-01 2.74327904e-01 5.30487061e-01 1.17245372e-02
-5.37594378e-01 5.38725913e-01 2.38596469e-01 8.43444616e-02
9.94066358e-01 -1.47731173e+00 9.18467283e-01 1.56093478e+00
1.82588294e-01 2.05513597e-01 9.74739671e-01 -1.57011613e-01
-1.57524788e+00 -1.32691646e+00 5.24474084e-01 -3.28063071e-01
8.05448472e-01 -1.95608735e-01 -6.25608146e-01 9.45443571e-01
6.38413370e-01 7.93424249e-01 9.76750031e-02 -3.39150697e-01
-7.79157758e-01 -1.54631227e-01 -8.10201585e-01 8.94433558e-01
1.17116010e+00 -8.10230792e-01 -2.09500745e-01 1.63102120e-01
1.04179239e+00 -1.04936504e+00 -4.42247093e-01 -2.27584876e-02
7.59328127e-01 -1.22070134e+00 9.02139246e-01 -5.14318466e-01
5.13501406e-01 -3.70569944e-01 -2.81612664e-01 -1.56303370e+00
-4.24965382e-01 -9.37101066e-01 -8.52581710e-02 6.72958910e-01
3.97185944e-02 -4.14606363e-01 5.64865351e-01 9.59015429e-01
-2.80108005e-01 -3.32681298e-01 -1.32196307e+00 -6.62585795e-01
6.21442646e-02 -4.39331323e-01 2.12671235e-01 7.15869725e-01
-5.17235696e-01 -2.94166684e-01 -7.50955641e-01 -1.14673115e-01
6.98318422e-01 5.00408888e-01 4.33637410e-01 -4.40052450e-01
-2.84861773e-01 -5.13902366e-01 -1.31473601e-01 -1.13942862e+00
5.26518166e-01 -1.86324492e-01 4.62234378e-01 -1.33580959e+00
7.14347959e-02 8.15974623e-02 -7.19087899e-01 5.08355081e-01
-3.80001366e-01 5.89940667e-01 2.22407416e-01 -2.11695686e-01
-1.13366032e+00 8.39369357e-01 1.54036784e+00 -3.05337101e-01
-3.18324305e-02 -4.39413458e-01 -9.69131216e-02 1.88400269e-01
5.72893798e-01 1.57653391e-01 -3.11942399e-01 -7.48373330e-01
-1.61637254e-02 6.51033670e-02 5.76647401e-01 -6.94992423e-01
1.56895325e-01 -5.52484214e-01 6.79039180e-01 -3.99735332e-01
5.10749400e-01 -5.00233352e-01 -2.02394754e-01 4.05797184e-01
-3.86054605e-01 6.47244692e-01 5.06799281e-01 6.39385104e-01
-5.69612421e-02 5.38312078e-01 1.20086312e+00 -1.31658435e-01
-1.50041175e+00 6.11247778e-01 -2.06489608e-01 1.58089325e-01
8.10122371e-01 -1.30161807e-01 -8.54641140e-01 -5.34175098e-01
-2.98239112e-01 -1.19432859e-01 6.88683867e-01 4.54091370e-01
9.08295572e-01 -1.58801889e+00 -6.69221699e-01 -5.75228967e-02
1.19526394e-01 -4.95105714e-01 3.87708575e-01 6.86244905e-01
-8.70529830e-01 1.82500824e-01 -5.24709702e-01 -6.11161292e-01
-1.07373512e+00 6.89669669e-01 9.85391200e-01 3.79252374e-01
-7.56356955e-01 1.08824611e+00 4.58048955e-02 2.49370828e-01
7.10236058e-02 -9.21641648e-01 4.39976245e-01 -4.82933283e-01
4.53425616e-01 8.58178921e-03 -2.13933676e-01 -6.28426969e-01
-3.04838717e-01 5.34051955e-01 4.43384200e-01 -3.77305716e-01
1.20739019e+00 -5.00445604e-01 -3.27657610e-01 8.74972403e-01
1.25729859e+00 -6.74417913e-01 -2.26184821e+00 -4.28201497e-01
-2.91418612e-01 -8.93193722e-01 1.42585978e-01 -7.97514021e-01
-1.58931017e+00 5.72788060e-01 1.00782943e+00 -2.17887849e-01
1.52675617e+00 -3.36033106e-01 6.05517626e-01 2.03580230e-01
1.96241722e-01 -1.12695491e+00 4.89320070e-01 6.23799682e-01
8.24647784e-01 -1.32937694e+00 -6.09791949e-02 2.34365270e-01
-7.55271018e-01 1.25635052e+00 6.51920140e-01 -4.38270837e-01
3.47319216e-01 4.60678823e-02 3.83550256e-01 3.34833741e-01
-1.02267885e+00 -5.52299678e-01 4.14574564e-01 6.07789874e-01
7.42067397e-01 -9.97829810e-03 1.13361217e-01 -3.44998151e-01
1.92576393e-01 7.86809698e-02 6.14345014e-01 6.42430782e-01
-1.24060720e-01 -6.83906019e-01 -3.32132950e-02 -2.38900721e-01
-2.90369660e-01 -2.92105168e-01 -1.79416448e-01 6.56869709e-01
-6.02876283e-02 9.06257033e-01 3.94782990e-01 -8.33010226e-02
-1.73996983e-03 -2.05116823e-01 1.06112576e+00 1.91646323e-01
-3.50201249e-01 1.60986871e-01 -4.54839095e-02 -1.31767309e+00
-1.16222751e+00 -6.56263053e-01 -1.14102221e+00 -5.86456180e-01
2.76551664e-01 -5.18851101e-01 4.39451784e-01 7.10476279e-01
-1.24667622e-01 8.61585855e-01 6.76372886e-01 -1.40913153e+00
1.23707995e-01 -7.38302469e-01 -6.25328124e-01 5.91554224e-01
1.02588260e+00 -4.01753753e-01 -1.50276914e-01 4.77326214e-01] | [10.801830291748047, -1.5128265619277954] |
94d23bea-db56-4dce-bf40-b29be5b70e42 | leveraging-modality-specific-representations | 2212.05301 | null | https://arxiv.org/abs/2212.05301v2 | https://arxiv.org/pdf/2212.05301v2.pdf | Leveraging Modality-specific Representations for Audio-visual Speech Recognition via Reinforcement Learning | Audio-visual speech recognition (AVSR) has gained remarkable success for ameliorating the noise-robustness of speech recognition. Mainstream methods focus on fusing audio and visual inputs to obtain modality-invariant representations. However, such representations are prone to over-reliance on audio modality as it is much easier to recognize than video modality in clean conditions. As a result, the AVSR model underestimates the importance of visual stream in face of noise corruption. To this end, we leverage visual modality-specific representations to provide stable complementary information for the AVSR task. Specifically, we propose a reinforcement learning (RL) based framework called MSRL, where the agent dynamically harmonizes modality-invariant and modality-specific representations in the auto-regressive decoding process. We customize a reward function directly related to task-specific metrics (i.e., word error rate), which encourages the MSRL to effectively explore the optimal integration strategy. Experimental results on the LRS3 dataset show that the proposed method achieves state-of-the-art in both clean and various noisy conditions. Furthermore, we demonstrate the better generality of MSRL system than other baselines when test set contains unseen noises. | ['Eng Siong Chng', 'Beier Zhu', 'Heqing Zou', 'Qiang Zhang', 'Yuchen Hu', 'Chen Chen'] | 2022-12-10 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 1.93796307e-01 -4.25340980e-01 -3.33016887e-02 -1.64488167e-01
-1.19436526e+00 -3.53039950e-01 6.16182327e-01 -3.25690895e-01
-4.13481176e-01 4.40044075e-01 4.34404194e-01 -3.53272319e-01
1.15992635e-01 -2.83895284e-01 -6.83756709e-01 -9.60306466e-01
4.01326984e-01 -3.39361221e-01 -5.22961430e-02 -2.00986579e-01
1.56332925e-01 3.34980845e-01 -1.68142438e+00 3.25462192e-01
6.73198342e-01 1.12832904e+00 5.71815193e-01 6.74256325e-01
-8.76136124e-03 8.53598893e-01 -6.59580588e-01 3.41350702e-03
6.71846196e-02 -3.91855776e-01 -4.32750434e-01 2.04499260e-01
2.55464822e-01 -9.28990394e-02 -7.04132736e-01 1.07933331e+00
7.68394351e-01 4.50270981e-01 6.18670762e-01 -1.29792058e+00
-8.36954474e-01 3.72497559e-01 -6.15259171e-01 4.83998030e-01
4.56975460e-01 5.66429317e-01 1.13829315e+00 -1.12470973e+00
2.99730301e-01 1.51827967e+00 5.56216426e-02 6.56007588e-01
-1.17800176e+00 -6.79856062e-01 5.73822796e-01 6.56288624e-01
-1.34491599e+00 -9.68995273e-01 9.33085740e-01 -1.32715017e-01
9.92293477e-01 3.48225892e-01 4.01200384e-01 1.59479094e+00
-7.36965314e-02 1.11081457e+00 1.14215255e+00 -4.13096696e-01
1.95829690e-01 -1.01171480e-02 -1.87005654e-01 1.32598057e-01
-2.58062899e-01 3.09024870e-01 -9.65790808e-01 1.41746536e-01
5.73627949e-01 -1.29999101e-01 -5.06545186e-01 -2.90475816e-01
-1.12515485e+00 5.61988235e-01 3.50324333e-01 1.72139123e-01
-3.63082409e-01 1.08427599e-01 4.48183656e-01 4.77246940e-01
2.60303080e-01 5.42945378e-02 -2.86885202e-01 -1.88983440e-01
-6.25010788e-01 -3.71522941e-02 3.19642648e-02 5.94294429e-01
2.80587077e-01 6.70630217e-01 -5.24970949e-01 1.35340905e+00
6.73616648e-01 5.99140823e-01 6.01386487e-01 -7.00609505e-01
6.96052492e-01 2.39730462e-01 -1.59268081e-01 -5.87912798e-01
-5.65429293e-02 -5.61404586e-01 -7.67861903e-01 2.71896422e-01
-2.41387896e-02 2.03664765e-01 -1.21548450e+00 1.95248818e+00
2.01128244e-01 3.96305859e-01 4.09864634e-01 1.13939214e+00
9.92524922e-01 8.21577549e-01 3.48976374e-01 -4.52744931e-01
1.23874879e+00 -9.68873858e-01 -1.00343204e+00 -4.56084400e-01
3.71595509e-02 -6.09091341e-01 1.45480943e+00 3.28343749e-01
-9.81458783e-01 -6.35013580e-01 -1.00434947e+00 1.66930452e-01
-7.47810304e-02 4.94813621e-02 6.10171780e-02 5.50218999e-01
-1.00335562e+00 -3.16059701e-02 -7.19496310e-01 -1.47481203e-01
3.09248120e-01 1.25025734e-01 -5.06698310e-01 -1.06198050e-01
-1.22005343e+00 7.31691420e-01 2.42187411e-01 2.89815307e-01
-1.25349438e+00 -3.53611708e-01 -9.91507709e-01 -3.44769210e-02
6.36669338e-01 -5.75395882e-01 1.26490724e+00 -9.27132905e-01
-1.75802982e+00 5.18895566e-01 -3.73094559e-01 -3.44352007e-01
1.95264608e-01 -1.49787879e-02 -7.99972951e-01 3.00192207e-01
-1.78376660e-01 5.81412137e-01 1.49077964e+00 -1.49880588e+00
-5.47019303e-01 -2.42313311e-01 -4.52417172e-02 6.14295900e-01
-4.82388645e-01 8.76327083e-02 -5.96912026e-01 -9.26953554e-01
1.24013543e-01 -5.79163730e-01 1.24814376e-01 -1.56156883e-01
-2.07608715e-01 -2.97594130e-01 8.58420610e-01 -7.21448362e-01
1.17795634e+00 -2.52847934e+00 4.16661054e-01 -1.12826880e-02
2.95548458e-02 5.24696350e-01 -5.63348651e-01 3.30813199e-01
-5.34802936e-02 1.13041671e-02 -1.06974967e-01 -4.33828861e-01
-1.03672706e-01 2.18184680e-01 -4.59488869e-01 4.14543569e-01
4.42706496e-01 7.74034917e-01 -7.81978726e-01 -3.97659779e-01
3.54448944e-01 8.26003969e-01 -3.75382543e-01 3.55417013e-01
-1.15985926e-02 6.96427584e-01 -2.66994357e-01 9.17092979e-01
5.18999994e-01 4.78503816e-02 -8.01446140e-02 -3.23055536e-01
3.72953303e-02 2.58887559e-01 -1.11618638e+00 1.66139746e+00
-6.05480909e-01 6.56731427e-01 1.77761793e-01 -1.04637372e+00
8.97805750e-01 4.79545087e-01 4.08569425e-02 -1.05677855e+00
2.64966227e-02 -6.72231195e-03 -6.60561770e-02 -5.64038336e-01
2.67008603e-01 -8.02988857e-02 2.19572768e-01 1.05485991e-01
2.31912285e-01 5.26758246e-02 -1.71849936e-01 2.09920168e-01
1.00640094e+00 2.75476240e-02 3.36113781e-01 4.20997918e-01
7.66763866e-01 -6.55288100e-01 5.91971278e-01 7.34485626e-01
-5.44238567e-01 6.59899473e-01 2.46158823e-01 1.50724575e-01
-6.12080514e-01 -1.37182391e+00 8.76747891e-02 1.31981337e+00
2.23537490e-01 -2.55799174e-01 -5.17511547e-01 -5.75003445e-01
-2.15959311e-01 7.60308862e-01 -3.70361149e-01 -4.32847112e-01
-3.09629440e-01 -3.30523551e-01 6.65133595e-01 6.10432327e-01
4.14088279e-01 -1.25522733e+00 -2.57324755e-01 -1.41090481e-03
-5.32837033e-01 -1.27865100e+00 -4.34655070e-01 4.39081974e-02
-5.47987819e-01 -7.63915241e-01 -7.72185206e-01 -6.77326560e-01
4.57058430e-01 7.66622961e-01 5.91742396e-01 -1.71477810e-01
-2.74047293e-02 7.66623616e-01 -5.04020751e-01 -1.49763957e-01
-4.29449648e-01 -2.69900888e-01 1.65834323e-01 4.39373285e-01
1.02787964e-01 -5.38635135e-01 -5.78941882e-01 2.81770229e-01
-8.66964638e-01 -2.29548499e-01 6.62927091e-01 1.05836320e+00
6.56046271e-01 -1.27477258e-01 8.69030178e-01 -8.16266537e-02
7.51848698e-01 -3.78011942e-01 -3.75148982e-01 3.93989593e-01
-4.09470737e-01 -7.54023939e-02 4.95200813e-01 -7.74826169e-01
-1.27478182e+00 -1.73530877e-02 -1.29214749e-01 -8.26105833e-01
-1.57100394e-01 5.25825441e-01 -4.76562649e-01 4.22924384e-03
4.47978020e-01 5.89489043e-01 -5.50670996e-02 -4.88774002e-01
4.58026052e-01 9.67028975e-01 6.86184824e-01 -3.32233071e-01
8.23297381e-01 2.03325287e-01 -4.25944179e-01 -1.06225383e+00
-4.55519557e-01 -5.36423922e-01 -9.04825404e-02 -5.56742728e-01
6.35302722e-01 -1.33291256e+00 -7.04583764e-01 4.43914354e-01
-9.62344229e-01 -8.40008855e-02 -7.03768879e-02 5.68216562e-01
-5.48646271e-01 4.65772063e-01 -4.16179508e-01 -1.25592899e+00
-1.42053440e-01 -1.43161607e+00 1.10811269e+00 2.49770761e-01
9.54829827e-02 -4.76469904e-01 -9.51974317e-02 6.35639369e-01
3.24423492e-01 -2.80242801e-01 6.74898624e-01 -4.91904169e-01
-4.91153330e-01 8.91837627e-02 -2.61729360e-01 5.37804306e-01
7.60520473e-02 -5.60399219e-02 -1.50457764e+00 -4.11851674e-01
-1.19427582e-02 -5.12551427e-01 1.19550669e+00 3.27823818e-01
1.22048736e+00 -1.95754424e-01 6.29372895e-02 4.71475720e-01
1.02418053e+00 4.02411312e-01 7.58178830e-01 3.66020918e-01
6.80951893e-01 4.62038845e-01 6.96474850e-01 3.17643106e-01
4.12843555e-01 8.75817001e-01 5.77849329e-01 -4.39106338e-02
-4.63781029e-01 -3.86654109e-01 7.70466805e-01 8.92569780e-01
4.97962721e-02 -4.59070295e-01 -6.62391722e-01 4.46793616e-01
-1.84773135e+00 -9.45237279e-01 4.40482676e-01 2.15947556e+00
8.72274578e-01 2.00805347e-02 1.39319792e-01 4.53106821e-01
6.76477969e-01 4.85635072e-01 -5.03090143e-01 -1.46177679e-01
-4.69475448e-01 1.26161650e-02 -3.08445022e-02 4.40665483e-01
-1.05590165e+00 1.01942933e+00 5.95848894e+00 1.11255419e+00
-1.33212698e+00 1.84945926e-01 4.36218888e-01 -2.83643723e-01
-3.53843838e-01 -3.26827258e-01 -5.16551256e-01 3.72400761e-01
9.05034423e-01 9.72127169e-02 7.48400688e-01 5.62977493e-01
5.48820972e-01 7.50771165e-02 -8.09689462e-01 1.22484517e+00
6.96437284e-02 -8.90141666e-01 1.40008405e-01 -1.08141273e-01
2.89223969e-01 -3.86341438e-02 5.04171968e-01 4.76202220e-01
-4.53453958e-02 -1.09271336e+00 9.82607722e-01 4.12502617e-01
8.66221189e-01 -7.01878965e-01 3.51284295e-01 1.83313012e-01
-1.31855297e+00 -4.12647128e-01 -1.59173414e-01 2.35904917e-01
1.99599296e-01 6.92836894e-03 -7.76417851e-01 3.56641144e-01
7.56344855e-01 6.12682879e-01 -4.79141474e-01 8.89576435e-01
-4.22233582e-01 6.92806542e-01 3.45067233e-02 2.75587648e-01
7.65583590e-02 1.11582160e-01 8.32849920e-01 1.17944825e+00
1.83870807e-01 -4.34389059e-03 1.09071657e-02 5.41421890e-01
-7.86738545e-02 1.40773788e-01 -6.94525480e-01 -2.11915955e-01
6.71599329e-01 9.63286936e-01 -2.59084761e-01 3.54042239e-02
-5.12427926e-01 7.66314328e-01 4.07458514e-01 8.83586705e-01
-7.59924412e-01 -1.31895557e-01 8.21256876e-01 -3.11499327e-01
5.39969265e-01 -1.97768077e-01 -1.46013916e-01 -1.12195754e+00
1.83597744e-01 -1.24139988e+00 2.85568565e-01 -1.01782548e+00
-1.08819973e+00 7.61547685e-01 -2.00213760e-01 -1.46147621e+00
-3.33085835e-01 -4.50792611e-01 -4.23166186e-01 7.88756490e-01
-1.76800990e+00 -1.17990911e+00 -6.09064884e-02 9.01984870e-01
9.53527868e-01 -4.79126960e-01 5.60419440e-01 3.52278531e-01
-8.28859746e-01 8.89872253e-01 -8.66982788e-02 -7.35307410e-02
7.67421842e-01 -8.74732196e-01 2.19610572e-01 1.02448308e+00
4.67288256e-01 4.12468582e-01 7.81352341e-01 -5.35935163e-01
-1.63928628e+00 -9.70602214e-01 4.60354209e-01 -4.69469130e-02
5.84766448e-01 -3.01081866e-01 -1.05904353e+00 3.06914032e-01
1.87338725e-01 2.33369723e-01 5.22215784e-01 -1.77475646e-01
-7.70085037e-01 -3.13802510e-01 -8.39943528e-01 7.58028924e-01
9.83427346e-01 -1.07577026e+00 -5.35169661e-01 -8.74343514e-02
9.86355126e-01 -3.00585866e-01 -4.31144416e-01 4.00687695e-01
4.39545333e-01 -6.82223439e-01 1.27700520e+00 -6.33231163e-01
8.31703693e-02 -5.60054064e-01 -5.72814643e-01 -1.44195831e+00
-1.62621364e-01 -4.66960967e-01 -3.72741014e-01 1.30539846e+00
3.70950609e-01 -4.43794906e-01 6.62190691e-02 8.06920677e-02
-1.99026078e-01 -4.96930808e-01 -1.32900250e+00 -8.22501302e-01
-4.51824516e-01 -7.35099316e-01 3.35284024e-01 7.09101200e-01
-1.77788511e-01 3.81119311e-01 -6.97597444e-01 4.75700289e-01
4.83576089e-01 -3.69704127e-01 5.43919504e-01 -7.73540199e-01
-3.85095507e-01 -3.75530571e-01 -3.60008359e-01 -9.72342193e-01
3.63768429e-01 -6.49839461e-01 3.01663846e-01 -1.32335770e+00
-4.49751038e-03 -1.10016372e-02 -8.71806383e-01 6.02339566e-01
-3.09410334e-01 2.61855423e-01 3.82626683e-01 2.06366152e-01
-7.98155367e-01 1.01191616e+00 1.27282035e+00 -4.17452604e-01
-3.07038814e-01 -1.25906849e-02 -7.30844736e-01 4.51490045e-01
6.98754251e-01 -1.67382970e-01 -5.64877272e-01 -4.64894742e-01
-1.40885442e-01 3.04626673e-01 4.88090634e-01 -6.33256674e-01
2.03983054e-01 -3.74513805e-01 2.07329422e-01 -5.22136867e-01
7.35425651e-01 -6.46567583e-01 -1.52934179e-01 4.76004630e-02
-5.28973341e-01 -2.01931760e-01 2.59203970e-01 1.01496375e+00
-5.44548213e-01 3.69874686e-02 7.19607413e-01 2.04163522e-01
-8.15163970e-01 6.15133643e-02 -4.96829033e-01 8.91720690e-03
6.55276954e-01 -2.31130883e-01 -3.30419958e-01 -6.14745855e-01
-6.42323852e-01 1.13043040e-01 7.33033009e-03 8.73952389e-01
1.08280146e+00 -1.48944855e+00 -7.15373278e-01 3.39086980e-01
3.89219761e-01 -4.09806371e-01 5.20250499e-01 7.82557189e-01
1.59053281e-01 2.36957207e-01 -2.80842558e-02 -7.78668284e-01
-1.45282590e+00 5.11685252e-01 2.78617412e-01 1.43023387e-01
-5.34294963e-01 8.61687899e-01 4.00321484e-01 -5.17864712e-02
8.54071736e-01 -6.58560991e-02 -4.82363671e-01 4.44967411e-02
7.28032112e-01 3.14496756e-01 9.44037214e-02 -7.99359977e-01
-5.38127840e-01 2.98506469e-01 -1.27957970e-01 -4.74074185e-01
1.18928003e+00 -4.27598953e-01 4.59042996e-01 5.72339177e-01
1.02872729e+00 -1.08273819e-01 -1.50738060e+00 -5.26762486e-01
-3.54614139e-01 -5.30776381e-01 5.04598081e-01 -8.86668265e-01
-1.24075723e+00 1.08973134e+00 8.88525724e-01 3.23590599e-02
1.40111232e+00 -3.07159424e-02 4.17816907e-01 2.08633453e-01
-1.80022744e-03 -1.08768284e+00 4.47119981e-01 3.93760502e-01
1.20140207e+00 -1.35973763e+00 -3.28402907e-01 -1.95529580e-01
-1.04242063e+00 8.63350630e-01 5.95094502e-01 3.99022311e-01
3.01971644e-01 2.79359780e-02 3.44356924e-01 2.80038863e-01
-9.87783968e-01 -4.82112408e-01 5.63203931e-01 9.31960344e-01
1.84849203e-01 9.66253877e-03 1.64883539e-01 6.48221612e-01
1.59461156e-01 -3.93527716e-01 7.17145205e-02 7.49566913e-01
-4.33484346e-01 -9.10582900e-01 -6.88125193e-01 1.59951225e-02
-3.01639646e-01 -1.88356429e-01 -3.00804496e-01 3.54734451e-01
-3.28046054e-01 1.41985631e+00 -1.99938312e-01 -5.76127410e-01
4.83115464e-01 3.72107923e-02 3.92486036e-01 -2.72499055e-01
-3.95434052e-01 6.70718074e-01 -1.26555473e-01 -5.02914548e-01
-3.91650379e-01 -6.06629491e-01 -1.01390612e+00 1.32397816e-01
-3.16071302e-01 -2.49134347e-01 6.29399419e-01 9.90616322e-01
3.83691370e-01 7.80266345e-01 8.22067857e-01 -8.06302309e-01
-6.85159326e-01 -9.23630476e-01 -3.86170447e-01 2.84372836e-01
7.23540306e-01 -8.97955775e-01 -5.26637971e-01 -1.58059046e-01] | [14.288662910461426, 5.188167095184326] |
9c46a9b1-4f8a-41ef-84a6-341ca3e9d6f4 | an-adaptive-cm-array-preconditioner-for-blind | 1807.09692 | null | http://arxiv.org/abs/1807.09692v1 | http://arxiv.org/pdf/1807.09692v1.pdf | An Adaptive CM Array Preconditioner for Blind Multi-User Separation | The family of constant-modulus algorithms is widely used in wireless
communication systems and in radar. The classical constant-modulus adaptive
(CMA) algorithm, however, fails to lock onto a single mode when used in
conjunction with an antenna array. Instead, it equalizes the entire spatial
spectrum. In this paper, we describe in full detail our recently proposed
approach for the separation of multiple users in a radio system with frequency
reuse, such as a cellular network, making use of the CMA algorithm. Based on
the observation that the differential filter weights resemble a superposition
of the array steering vectors, we cast the original task to a
direction-of-arrival estimation problem. With rigorous theoretical analysis of
the array response based on the discrete-space Fourier transform we elaborate a
solution that solves the problem by finding the roots of a polynomial equation.
We provide a numerical example to demonstrate the validity of the approach
under high-SNR conditions. In addition, we propose a more general preprocessor
for the CMA array which allows the modulated signals to differ in amplitude. As
a byproduct, the preprocessor yields a low-cost estimate of the number of
concurrent users, i.e. the model order, by simply counting the roots with the
strongest response. | [] | 2018-07-25 | null | null | null | null | ['direction-of-arrival-estimation'] | ['audio'] | [ 5.92423022e-01 -1.01875611e-01 3.70635629e-01 1.23640738e-01
-4.58979309e-01 -6.79703057e-01 3.65419269e-01 1.33020068e-02
-4.31912839e-01 4.60127294e-01 -3.28973114e-01 -6.36326373e-01
-7.07800984e-01 -7.77336478e-01 -1.99002296e-01 -1.07492065e+00
-3.79868895e-01 7.11659193e-02 -1.40635923e-01 -1.67877719e-01
7.24802837e-02 7.32692719e-01 -1.02129912e+00 -3.22697103e-01
7.36602545e-01 1.20086765e+00 6.72032163e-02 8.26130509e-01
2.64502287e-01 6.66789189e-02 -7.60541260e-01 -8.59540477e-02
3.95501912e-01 -6.60884619e-01 1.45082278e-02 1.45442605e-01
4.18329611e-04 2.67341500e-03 -7.77853057e-02 9.10825193e-01
6.04002178e-01 -1.51992068e-01 7.42198288e-01 -1.05349541e+00
1.80644885e-01 3.33276749e-01 -4.66415793e-01 2.23088041e-01
1.52584985e-01 -5.74505091e-01 6.26082480e-01 -1.04603648e+00
7.80553743e-02 7.69787550e-01 7.97026157e-01 -1.51547208e-01
-1.18941915e+00 -7.64691532e-01 -4.86770391e-01 -3.76960114e-02
-1.63436651e+00 -6.59821928e-01 6.28813565e-01 -1.53023094e-01
3.45866084e-01 5.71737289e-01 5.48064411e-01 2.49140680e-01
7.70437643e-02 1.08764403e-01 9.37675655e-01 -1.00370562e+00
3.23170394e-01 -2.43079498e-01 1.14482962e-01 6.19527400e-01
8.51384699e-01 1.01622321e-01 -2.35083953e-01 -7.16418386e-01
7.65608728e-01 -2.90704012e-01 -5.98463833e-01 -5.37855327e-01
-1.05515385e+00 6.41464531e-01 -1.09956197e-01 7.94269204e-01
-3.09800059e-01 1.76256001e-01 -4.51881796e-01 4.43411231e-01
2.00529978e-01 5.25716364e-01 -2.69144718e-02 4.83403027e-01
-1.00674725e+00 1.64448693e-01 1.06532049e+00 6.64383829e-01
6.08519733e-01 2.29911655e-01 1.17567472e-01 4.92615730e-01
4.53334510e-01 9.54488516e-01 -9.18220580e-02 -6.61662102e-01
7.49717578e-02 -1.67331949e-01 2.65275985e-01 -1.05605316e+00
-7.02996254e-01 -1.40948725e+00 -8.99720669e-01 1.83956489e-01
9.24422383e-01 -5.69781303e-01 -4.05937582e-01 1.62031102e+00
1.56921864e-01 4.61574614e-01 3.88062596e-01 6.17212474e-01
-1.25375807e-01 5.60241818e-01 -5.37411869e-01 -7.48644590e-01
1.28890443e+00 -2.68881083e-01 -7.22156405e-01 -2.70249903e-01
2.52279073e-01 -8.99941862e-01 -2.20859319e-01 7.38217771e-01
-1.02895784e+00 -2.48665258e-01 -1.40123379e+00 8.43295097e-01
2.18433127e-01 3.41031939e-01 1.96814463e-01 1.14884973e+00
-1.00837624e+00 2.52939492e-01 -3.77503723e-01 -1.70121536e-01
-2.10429296e-01 1.93258241e-01 6.35291710e-02 1.78235546e-01
-9.26771820e-01 5.72989941e-01 -5.42306863e-02 2.51596332e-01
-8.35600495e-03 -5.81138253e-01 -4.78104502e-01 2.99764395e-01
2.75217623e-01 -5.41532516e-01 1.16680717e+00 -8.39507043e-01
-1.24936140e+00 2.80534476e-01 -3.42267275e-01 -4.59037870e-01
2.55003333e-01 8.37429538e-02 -6.83218181e-01 3.19067121e-01
-1.02116987e-01 -5.59792399e-01 1.06048906e+00 -1.00706983e+00
-6.43793046e-01 -2.76299626e-01 -3.58640462e-01 -1.43428743e-01
-6.01712428e-02 -3.28795403e-01 -8.82832035e-02 -7.17525899e-01
5.40025294e-01 -9.92738247e-01 -6.10396802e-01 -2.43644461e-01
-5.92657067e-02 4.13341582e-01 2.14724153e-01 -2.88145036e-01
1.34831357e+00 -2.59821296e+00 1.27049312e-01 1.05219769e+00
1.37368977e-01 2.87828259e-02 1.42327756e-01 8.29850912e-01
-2.71773696e-01 -3.86073083e-01 -2.95061588e-01 1.68289736e-01
-2.93525696e-01 -2.17119470e-01 -3.47456664e-01 9.41198349e-01
-4.91077527e-02 1.12316817e-01 -8.35296214e-01 -4.32121307e-02
-2.37384558e-01 9.48713943e-02 -5.10490000e-01 -1.26051798e-01
4.17973071e-01 2.37613782e-01 -5.23861647e-01 3.32116604e-01
1.00944901e+00 -1.81950301e-01 6.08864009e-01 -3.37387502e-01
-5.67081630e-01 -3.52426380e-01 -1.55214822e+00 1.18575311e+00
-4.89558846e-01 6.42794430e-01 7.22291648e-01 -1.22867012e+00
7.79983938e-01 4.86284673e-01 6.87441409e-01 -5.27538419e-01
1.91245049e-01 6.93661034e-01 2.80337006e-01 -5.73962331e-02
2.87763983e-01 -3.00114840e-01 -2.83195704e-01 3.86378378e-01
-1.29723158e-02 2.21031159e-01 1.59306332e-01 7.46523729e-03
1.13683820e+00 -5.75025141e-01 8.95334959e-01 -5.86665988e-01
7.42872655e-01 -3.33114654e-01 2.25382596e-01 8.84509206e-01
2.66000092e-01 2.49054193e-01 3.45324099e-01 1.25283301e-01
-8.43038142e-01 -1.04370260e+00 -2.75532812e-01 4.32172745e-01
2.09765270e-01 -1.93806499e-01 -5.07392585e-01 2.28613615e-01
5.92406467e-02 5.83992779e-01 -1.09142758e-01 1.07925899e-01
-5.64909458e-01 -1.09847748e+00 3.85086924e-01 -2.46267334e-01
2.99766213e-01 6.74441382e-02 -7.79489815e-01 4.28496033e-01
1.03826579e-02 -1.20857370e+00 -7.72843063e-02 3.32171977e-01
-4.54754293e-01 -9.85123932e-01 -8.96596789e-01 -5.07190824e-01
6.34320498e-01 6.86729193e-01 5.41831195e-01 -9.15567502e-02
-3.99316959e-02 7.87845373e-01 -2.55479574e-01 -4.24639255e-01
-3.05424839e-01 -3.96890819e-01 1.59692511e-01 5.17297328e-01
-3.54486704e-01 -7.74779916e-01 -4.78586882e-01 4.30427760e-01
-6.89103544e-01 -2.48911783e-01 8.70184541e-01 5.77743828e-01
-1.85184199e-02 3.50145549e-01 6.93429768e-01 -3.73902291e-01
6.80999219e-01 -5.37665248e-01 -7.81207681e-01 1.70460165e-01
-3.64299655e-01 1.00703470e-01 5.81317544e-01 -3.52866858e-01
-8.27473640e-01 1.77682683e-01 -1.42342076e-01 3.22212487e-01
4.15740788e-01 7.17043221e-01 -4.45486829e-02 -6.74047172e-01
5.90116382e-01 3.72107774e-01 -6.09612800e-02 -2.91677415e-01
2.05459356e-01 5.49132347e-01 7.35380232e-01 -1.56802267e-01
1.33788371e+00 5.60810804e-01 6.60076678e-01 -1.41360402e+00
-4.41599309e-01 -6.22501910e-01 -1.02276213e-01 -4.05630529e-01
1.62237987e-01 -6.72670543e-01 -6.49990261e-01 4.24294591e-01
-1.13491905e+00 9.93505642e-02 1.39106423e-01 8.69081199e-01
-3.34555775e-01 6.19131267e-01 8.90792012e-02 -1.32620060e+00
-1.75502405e-01 -3.59847099e-01 4.55932379e-01 1.27075911e-01
-1.04590520e-01 -5.91266155e-01 1.32435128e-01 -5.28477907e-01
6.35577500e-01 1.90142676e-01 8.10363531e-01 -5.40180385e-01
-4.50029403e-01 -4.04236197e-01 1.98361069e-01 -1.00242764e-01
-1.60779878e-02 -4.15487379e-01 -6.16490006e-01 -3.42929304e-01
4.76896942e-01 7.45604336e-01 4.97390211e-01 6.50498927e-01
5.21762550e-01 -4.37361039e-02 -6.10762179e-01 6.27456963e-01
1.65242302e+00 3.71741414e-01 5.12614846e-01 7.77626857e-02
-8.06578025e-02 2.17209414e-01 4.82492447e-01 7.23232925e-01
-3.35499495e-01 8.29985738e-01 2.32181266e-01 4.41809744e-02
2.78658003e-01 5.51961243e-01 1.48809716e-01 7.10138321e-01
3.23318541e-02 -5.05821943e-01 -4.90165025e-01 1.68059021e-01
-1.62775409e+00 -9.30343747e-01 -4.37198430e-01 2.53206253e+00
2.92502344e-01 3.71354185e-02 -2.24876069e-02 5.06992459e-01
5.58074355e-01 -1.01945050e-01 -8.31027851e-02 -7.93899372e-02
-1.66481540e-01 4.87071246e-01 1.06329107e+00 6.31872177e-01
-8.67468834e-01 -2.06022948e-01 6.14722729e+00 9.46950555e-01
-1.01785219e+00 1.03317341e-02 -7.64104351e-02 3.23070765e-01
-2.44243026e-01 -2.04172954e-02 -5.29321432e-01 3.91481489e-01
9.26410019e-01 -5.61556280e-01 2.09834546e-01 3.77896309e-01
-9.02868155e-03 -4.64349955e-01 -7.81821191e-01 1.02928185e+00
-1.04563572e-02 -9.57443655e-01 -4.43721384e-01 3.55900884e-01
2.11325243e-01 -6.26790822e-01 1.29363433e-01 -1.44013509e-01
-4.87363547e-01 -6.37315452e-01 6.99432909e-01 7.05760121e-01
4.71304595e-01 -6.88599527e-01 4.93525296e-01 6.17462873e-01
-1.30011570e+00 -3.42268795e-01 3.38787846e-02 -2.72395104e-01
4.56952214e-01 1.16844642e+00 -8.50121081e-01 9.94075060e-01
-1.50973260e-01 -1.51769072e-01 -6.06443472e-02 1.57492614e+00
3.03444654e-01 5.96520543e-01 -7.24462628e-01 -5.28468862e-02
7.18566552e-02 -4.17953849e-01 7.98855364e-01 1.21715009e+00
1.26013112e+00 4.71541166e-01 -9.80870426e-02 2.08481789e-01
3.36239964e-01 5.06476723e-02 -2.74449110e-01 2.03477904e-01
5.99911988e-01 1.11375785e+00 -7.80839384e-01 -7.46841729e-02
-5.14718652e-01 5.69042087e-01 -4.59163964e-01 5.74275792e-01
-4.28170741e-01 -6.64486289e-01 4.28334355e-01 1.69048712e-01
5.95779598e-01 -8.96898866e-01 -2.28229329e-01 -6.27795756e-01
-1.02890551e-01 -7.26084590e-01 1.79697499e-02 -3.34588200e-01
-8.13914061e-01 4.29876149e-01 1.59144737e-02 -1.61556351e+00
-5.48388302e-01 -5.94980478e-01 -4.03106838e-01 1.04433048e+00
-1.34792686e+00 -4.26047385e-01 1.23934643e-02 2.79894531e-01
-1.43870085e-01 -2.08686471e-01 1.01557755e+00 6.27821624e-01
-1.41235054e-01 3.81476939e-01 5.14413238e-01 -1.34861499e-01
3.89720976e-01 -8.19636405e-01 7.78829083e-02 1.10580480e+00
-4.54486981e-02 6.56298041e-01 1.25892305e+00 -3.89468133e-01
-1.27561581e+00 -4.61071908e-01 7.69670069e-01 2.49817163e-01
6.63037896e-01 -2.36699745e-01 -4.22783285e-01 9.71850231e-02
1.95365667e-01 -1.08440153e-01 9.89614427e-01 -2.74463266e-01
2.22170681e-01 -2.68234432e-01 -8.20893049e-01 3.63722712e-01
5.49514472e-01 2.58535892e-03 -2.20365033e-01 1.21095724e-01
-2.75061149e-02 -3.37512493e-01 -5.34519553e-01 4.18828279e-01
8.39473546e-01 -9.65413928e-01 1.11460769e+00 1.29376218e-01
-3.55238974e-01 -5.49276173e-01 -3.27091783e-01 -1.24493408e+00
-6.66340888e-01 -1.00142062e+00 -3.16188373e-02 6.87223613e-01
3.32744360e-01 -8.36602211e-01 4.76772040e-01 -2.84751445e-01
1.26653984e-01 -6.47030115e-01 -1.18310738e+00 -9.22008753e-01
-5.14835417e-01 -3.07820559e-01 1.03956491e-01 5.90138912e-01
8.58826861e-02 4.07651275e-01 -4.64338064e-01 9.41594601e-01
1.03537810e+00 1.16196513e-01 5.43564856e-01 -1.36973763e+00
-9.96298850e-01 -3.04239511e-01 -4.50012356e-01 -1.34905374e+00
-2.96814829e-01 -5.96131086e-01 2.39077825e-02 -1.14923549e+00
-4.34435695e-01 -6.85438931e-01 -5.55914454e-02 -3.83202046e-01
1.46094844e-01 2.65972495e-01 8.47739428e-02 -1.81765836e-02
1.14770140e-03 -5.50298132e-02 7.00850189e-01 1.90184757e-01
3.46445628e-02 8.19401324e-01 -4.62785959e-01 6.72146857e-01
7.02520609e-01 -3.55652899e-01 -3.12942147e-01 -1.25756994e-01
6.24423802e-01 3.88567299e-01 2.91111141e-01 -1.45186102e+00
4.59987760e-01 1.61778480e-01 3.61676544e-01 -4.43095952e-01
4.89738673e-01 -1.11900222e+00 6.26417577e-01 8.34226191e-01
1.87659919e-01 -2.16677561e-01 5.33346497e-02 7.89447188e-01
7.81640559e-02 -6.96476400e-01 5.91853201e-01 2.65676558e-01
-3.03757071e-01 -2.45683953e-01 -9.67092931e-01 -4.00756538e-01
9.33751106e-01 -1.36850327e-01 -8.28653276e-02 -7.34328985e-01
-7.42297411e-01 -3.88765447e-02 -2.43145619e-02 -5.30532420e-01
3.21318299e-01 -1.15270638e+00 -7.99885809e-01 3.10276181e-01
-3.09302211e-01 -6.11060739e-01 2.39783973e-01 1.09327173e+00
-3.40493768e-01 6.44749403e-01 -9.59754735e-02 -4.88303751e-01
-1.29153359e+00 1.96805000e-01 5.40450037e-01 -1.23048425e-01
-9.23165865e-03 4.30700719e-01 1.07188977e-01 4.17678475e-01
-1.02377012e-01 -1.82812631e-01 -1.29144683e-01 1.00984670e-01
7.12676883e-01 2.88951635e-01 1.44159704e-01 -3.77083182e-01
-4.40506637e-01 9.03487504e-01 5.87064207e-01 -5.90653896e-01
9.76251125e-01 -1.86747506e-01 -1.80036783e-01 1.79895476e-01
9.75186288e-01 1.02381396e+00 -5.81860662e-01 -3.53662193e-01
-7.54188523e-02 -3.56957287e-01 -1.98122069e-01 -3.26137125e-01
-5.91271937e-01 3.51983219e-01 6.71214700e-01 7.40478098e-01
1.41308486e+00 -1.77132666e-01 3.70823354e-01 6.27570152e-01
6.45111680e-01 -5.64813197e-01 -1.16210490e-01 3.45131934e-01
6.02075398e-01 -2.45608717e-01 3.13913912e-01 -7.36441493e-01
3.14077377e-01 1.47067034e+00 -4.10234809e-01 -2.29872242e-01
8.14119637e-01 5.63327253e-01 -3.60810035e-03 1.07030839e-01
-2.17195183e-01 -2.54051179e-01 7.44938403e-02 4.89058107e-01
2.50440240e-01 7.67204314e-02 -8.28555167e-01 5.42239070e-01
-1.60606965e-01 -1.85983881e-01 8.70674968e-01 8.84784281e-01
-9.85766053e-01 -1.30468965e+00 -9.47748959e-01 3.15845937e-01
-3.69822532e-01 -3.36404927e-02 7.06255957e-02 4.81633693e-01
-4.97568138e-02 1.03778517e+00 2.65469719e-02 -9.57915336e-02
5.86060166e-01 -9.54502076e-03 6.21522427e-01 -2.25751683e-01
6.06468543e-02 4.22514230e-01 2.78673649e-01 3.97727825e-02
-2.54057944e-01 -7.39119530e-01 -9.89731610e-01 -5.02791628e-02
-4.31074888e-01 5.31685889e-01 6.06635809e-01 8.35326552e-01
2.29837820e-02 5.81424534e-01 9.15167451e-01 -6.95515931e-01
-6.95462108e-01 -5.29049456e-01 -9.22193468e-01 -3.71926069e-01
5.90797663e-01 -4.00802910e-01 -4.62739617e-01 -4.10794735e-01] | [6.4805707931518555, 1.334424614906311] |
b15ddcb0-3757-4d37-82f1-5dba0b023c6e | losdd-leave-out-support-vector-data | 2212.13626 | null | https://arxiv.org/abs/2212.13626v1 | https://arxiv.org/pdf/2212.13626v1.pdf | LOSDD: Leave-Out Support Vector Data Description for Outlier Detection | Support Vector Machines have been successfully used for one-class classification (OCSVM, SVDD) when trained on clean data, but they work much worse on dirty data: outliers present in the training data tend to become support vectors, and are hence considered "normal". In this article, we improve the effectiveness to detect outliers in dirty training data with a leave-out strategy: by temporarily omitting one candidate at a time, this point can be judged using the remaining data only. We show that this is more effective at scoring the outlierness of points than using the slack term of existing SVM-based approaches. Identified outliers can then be removed from the data, such that outliers hidden by other outliers can be identified, to reduce the problem of masking. Naively, this approach would require training N individual SVMs (and training $O(N^2)$ SVMs when iteratively removing the worst outliers one at a time), which is prohibitively expensive. We will discuss that only support vectors need to be considered in each step and that by reusing SVM parameters and weights, this incremental retraining can be accelerated substantially. By removing candidates in batches, we can further improve the processing time, although it obviously remains more costly than training a single SVM. | ['Erich Schubert', 'Thomas Liebig', 'Daniel Boiar'] | 2022-12-27 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 1.12459235e-01 -1.14012875e-01 -1.30505264e-02 -3.96159381e-01
-3.79799873e-01 -4.58988130e-01 2.00804085e-01 5.45804977e-01
-5.18194556e-01 7.18479156e-01 -4.00712550e-01 -4.41467673e-01
-3.82559584e-03 -4.78183389e-01 -4.49434608e-01 -9.39183474e-01
-2.09641680e-01 4.13325071e-01 6.62485421e-01 -1.75676316e-01
5.69598675e-01 6.74623907e-01 -1.89411151e+00 3.19716036e-01
1.13959646e+00 8.39925230e-01 -1.63823321e-01 5.49136519e-01
-1.33053735e-01 4.97274667e-01 -1.01891494e+00 -1.56771883e-01
4.65201050e-01 -3.98396939e-01 -5.46770453e-01 2.32718825e-01
5.19849360e-01 -2.47016251e-01 -1.70578375e-01 9.53181565e-01
1.64967433e-01 3.84052217e-01 4.68434602e-01 -1.57575822e+00
-9.27378014e-02 1.97303414e-01 -3.72278720e-01 2.92782843e-01
2.61312395e-01 -1.55410329e-02 8.01792443e-01 -9.86760139e-01
3.90138477e-01 7.21424162e-01 8.21664751e-01 4.14382815e-01
-1.23052788e+00 -5.25833011e-01 2.93204397e-01 1.99331596e-01
-1.14494932e+00 -4.44174379e-01 6.26738846e-01 -3.48651916e-01
1.05989182e+00 7.52701938e-01 5.98275244e-01 8.15298498e-01
1.84898488e-02 7.08016455e-01 9.46362555e-01 -4.66212034e-01
4.29755360e-01 3.97681624e-01 4.74251539e-01 6.18179083e-01
5.96745551e-01 -5.27225807e-02 -2.75869220e-01 -5.46859622e-01
3.87656450e-01 3.37631285e-01 -1.41536698e-01 -5.61298966e-01
-8.88943374e-01 1.01434255e+00 2.79834509e-01 2.39608228e-01
-3.15031447e-02 -3.64181131e-01 6.90076828e-01 7.55183458e-01
6.06110215e-01 5.37330866e-01 -5.97172081e-01 -1.59132063e-01
-1.32634401e+00 1.67146057e-01 7.72582889e-01 6.14721358e-01
6.30586863e-01 1.98758736e-01 2.72539943e-01 1.01409090e+00
-1.55743822e-01 2.46057719e-01 6.06294513e-01 -5.36374211e-01
5.22457600e-01 8.09437692e-01 1.75470054e-01 -8.67894292e-01
-4.30821151e-01 -1.80124462e-01 -7.04717040e-01 7.60925114e-01
5.64547956e-01 -5.00698686e-02 -1.27758431e+00 1.05141115e+00
2.25521728e-01 1.32988542e-01 8.03720392e-03 6.37098491e-01
3.26716632e-01 4.75292981e-01 -3.20864946e-01 -3.33401144e-01
6.84748232e-01 -1.00124371e+00 -4.46247071e-01 -2.96969742e-01
9.51522470e-01 -7.27968931e-01 9.75215018e-01 9.41651106e-01
-7.82282591e-01 -3.57166559e-01 -1.29112983e+00 3.97046506e-01
-6.61608517e-01 1.76746324e-01 5.08844852e-01 6.90032959e-01
-5.95590651e-01 1.07742417e+00 -1.08128381e+00 -2.45593980e-01
3.36075634e-01 6.12551570e-01 -5.58952212e-01 7.82916695e-03
-9.38100874e-01 1.18389118e+00 3.45364004e-01 2.33194843e-01
-3.31507146e-01 -2.59188980e-01 -9.87138629e-01 -7.77233690e-02
4.28061873e-01 1.09087624e-01 8.07533979e-01 -1.04551864e+00
-1.08260667e+00 5.14848948e-01 -5.25842607e-01 -4.46159631e-01
7.20917881e-01 -9.37292278e-02 -5.73203027e-01 -1.03340536e-01
-7.94418752e-02 -6.73881024e-02 1.04375696e+00 -1.26984143e+00
-6.39111280e-01 -3.96106720e-01 -3.50678444e-01 -5.99506758e-02
-3.29234481e-01 2.58512288e-01 -2.27759361e-01 -5.95663846e-01
6.26718521e-01 -1.22354889e+00 -4.65786666e-01 -2.38698438e-01
-3.19447249e-01 -1.03639908e-01 1.05924332e+00 -6.63473070e-01
1.38220716e+00 -2.49068975e+00 -2.46792600e-01 8.07417393e-01
8.54080468e-02 6.79673433e-01 1.75516561e-01 3.03688020e-01
-3.17627728e-01 -2.11657137e-02 -3.91760856e-01 -3.83559227e-01
-3.27097058e-01 4.37279671e-01 -4.46605682e-01 7.02716529e-01
2.87996531e-01 3.01495254e-01 -7.56178021e-01 -2.28243202e-01
4.07621980e-01 4.40567099e-02 -5.07737160e-01 -1.97232693e-01
2.45973796e-01 -3.88243310e-02 9.82806981e-02 6.96347654e-01
8.05933595e-01 -4.97828797e-03 -2.02890649e-01 2.33862773e-01
-2.10469943e-02 -2.40652729e-02 -1.83071506e+00 8.16823602e-01
-3.40690821e-01 8.05800259e-01 -5.51029742e-02 -1.44224465e+00
1.05963933e+00 2.16240138e-01 3.92330348e-01 -1.92135036e-01
-1.77848876e-01 7.06027210e-01 8.10884759e-02 -4.27459568e-01
3.73877078e-01 -2.74984818e-02 -2.29669791e-02 6.07603937e-02
-1.61850601e-01 8.57965276e-02 2.97517598e-01 -1.36983618e-01
1.13001621e+00 -2.09760264e-01 1.82269230e-01 7.64095485e-02
4.36164081e-01 3.62685978e-01 6.79602146e-01 8.23792219e-01
-3.07426184e-01 6.88994884e-01 5.75414777e-01 -5.32342315e-01
-8.36580098e-01 -8.40856850e-01 -4.10726488e-01 9.91606414e-01
-1.47968195e-02 -3.15138698e-01 -4.51013774e-01 -1.03034186e+00
5.76645195e-01 7.47274756e-01 -6.14366233e-01 -2.51796901e-01
-7.12205112e-01 -9.53705132e-01 1.99825957e-01 5.63947618e-01
-1.48827493e-01 -9.90751028e-01 -5.74586868e-01 3.07939529e-01
3.96518469e-01 -5.55692792e-01 -8.41444135e-02 9.89183128e-01
-1.16515398e+00 -1.06644690e+00 -3.80846173e-01 -5.69464087e-01
1.31809747e+00 3.54915351e-01 4.64807957e-01 4.50302362e-01
-3.33646983e-01 -3.96849215e-01 -4.58803833e-01 -3.93826544e-01
-3.26390624e-01 -2.24087253e-01 4.35582608e-01 -2.23280698e-01
4.47431862e-01 -1.94997787e-01 -1.43782973e-01 5.52894354e-01
-8.19120169e-01 -5.04117131e-01 3.00394654e-01 1.24610901e+00
2.51820654e-01 4.05968279e-01 4.91518945e-01 -1.03782988e+00
4.66795087e-01 -3.34883332e-01 -6.09466016e-01 3.64538729e-02
-6.67258382e-01 -5.79376332e-02 1.17920005e+00 -5.18684268e-01
-6.58654630e-01 5.64526096e-02 -7.40282610e-02 -5.96501231e-01
-2.63010114e-01 2.60555744e-01 1.03059761e-01 -4.86663461e-01
8.54354680e-01 8.50587711e-02 2.23570645e-01 -6.03641391e-01
-1.59574434e-01 9.86331522e-01 1.84458807e-01 -1.44835144e-01
8.52380753e-01 3.77790928e-01 -6.50408268e-02 -9.19820845e-01
-6.71558201e-01 -8.95607531e-01 -7.72519767e-01 1.38432741e-01
8.49413276e-02 -5.96145749e-01 -2.65197515e-01 3.26126009e-01
-7.90413618e-01 6.77077891e-03 -3.93764973e-01 4.77283329e-01
-1.65952370e-01 4.49626267e-01 -3.29415411e-01 -9.04858649e-01
7.80951083e-02 -1.19455802e+00 7.22244561e-01 -4.18296410e-03
-5.16111612e-01 -7.38613009e-01 -4.34378088e-02 1.19831711e-01
3.67612094e-01 1.93768635e-01 6.70513213e-01 -1.48667133e+00
1.10827744e-01 -9.18185115e-01 7.39006624e-02 8.58152628e-01
3.07703286e-01 3.93986195e-01 -7.95823276e-01 -6.32882833e-01
1.02504171e-01 5.14937975e-02 9.87748265e-01 -3.77033204e-02
1.20263958e+00 -2.74499327e-01 -5.09855986e-01 4.72652674e-01
1.27611542e+00 3.44427466e-01 3.69098306e-01 5.74433208e-01
5.84974468e-01 3.91616166e-01 8.03052664e-01 4.15492356e-01
-3.15966606e-01 5.35788953e-01 2.69729048e-01 -4.02944297e-01
3.46314192e-01 2.05477938e-01 5.27538419e-01 5.90223372e-01
-4.63614166e-02 1.18366808e-01 -9.10647154e-01 5.67025840e-01
-1.86166787e+00 -1.08452415e+00 -4.69486445e-01 2.66154075e+00
5.59062183e-01 5.35912156e-01 2.53917307e-01 7.87275195e-01
6.49303019e-01 -7.24586397e-02 -4.74320650e-01 -7.69700050e-01
-2.82253791e-02 3.62256914e-01 5.67537129e-01 5.82396269e-01
-1.28893149e+00 6.67959988e-01 6.66967010e+00 6.88426256e-01
-1.34898615e+00 -3.11673611e-01 3.21025968e-01 -5.52764118e-01
5.05046546e-03 2.67746747e-01 -8.32223713e-01 7.73088694e-01
7.92054653e-01 8.54092017e-02 1.83097675e-01 1.15336120e+00
2.18541309e-01 -5.46446383e-01 -1.03669882e+00 7.43265986e-01
2.08453685e-01 -8.83958280e-01 -2.36841112e-01 -6.89544976e-02
6.59364343e-01 5.23838587e-02 -3.52712363e-01 3.20993662e-01
4.10458632e-02 -8.38654876e-01 3.50909680e-01 1.41141206e-01
1.24008715e-01 -8.42254281e-01 1.11393249e+00 4.24665570e-01
-6.43640041e-01 -3.66194248e-01 -6.53264105e-01 -1.51974186e-01
-1.16169393e-01 1.04682422e+00 -7.65210092e-01 5.01674652e-01
7.71395683e-01 3.97259206e-01 -7.31198013e-01 1.42099869e+00
1.36188194e-01 4.91976678e-01 -7.05748439e-01 -5.53670786e-02
1.49699822e-01 -3.08797508e-01 6.11529529e-01 1.09962201e+00
5.03405333e-01 -2.09986374e-01 3.84031057e-01 9.67535302e-02
6.12227023e-01 1.75142929e-01 -6.72968507e-01 4.82554406e-01
2.27521494e-01 9.23160851e-01 -8.96053195e-01 -6.31510556e-01
-4.53006029e-01 1.19901133e+00 9.28663835e-02 1.80101946e-01
-7.89516389e-01 -9.33540821e-01 5.24399281e-01 1.46610454e-01
5.99145532e-01 -2.69451857e-01 -4.10338163e-01 -1.29311001e+00
4.66103852e-01 -8.83949578e-01 4.92178351e-01 -3.60012174e-01
-1.32774663e+00 5.91967523e-01 -2.20941886e-01 -1.66873252e+00
-1.45899832e-01 -7.33988881e-01 -6.82646036e-01 7.94271708e-01
-1.16161275e+00 -2.71681190e-01 -2.06760079e-01 5.76773286e-01
3.36771101e-01 1.29834488e-01 7.69525826e-01 1.14079162e-01
-7.85081685e-01 6.10610664e-01 6.13320351e-01 2.67862380e-02
9.97821748e-01 -1.34029615e+00 2.85737216e-01 1.06565654e+00
2.49482274e-01 7.63699353e-01 9.23550546e-01 -7.43119240e-01
-8.59631240e-01 -8.56470108e-01 1.10114479e+00 -4.34899896e-01
6.92921281e-01 -4.24520105e-01 -1.36502373e+00 4.74211812e-01
-3.76959771e-01 3.80994797e-01 8.31514120e-01 2.76124984e-01
-1.21086583e-01 -1.19713664e-01 -1.32041264e+00 5.81203699e-01
7.13022649e-01 -1.99446589e-01 -7.21757770e-01 4.61497992e-01
2.70482272e-01 -4.06628251e-01 -6.68750107e-01 3.85349423e-01
2.53122330e-01 -1.14899147e+00 7.01596856e-01 -7.84810424e-01
-1.34753302e-01 -6.94974184e-01 1.32817462e-01 -1.48835349e+00
9.95362699e-02 -4.24146831e-01 -9.06289741e-02 7.18773246e-01
6.28838062e-01 -9.76007700e-01 9.65258360e-01 7.20653474e-01
-1.39317438e-01 -8.23611021e-01 -1.12389505e+00 -1.32129383e+00
-2.80105770e-01 -2.85491168e-01 3.44830155e-01 1.10472095e+00
4.98007923e-01 -3.53788137e-01 -3.27382952e-01 2.23665237e-01
3.36938441e-01 -8.94193072e-03 7.69279420e-01 -1.50544500e+00
-1.29100949e-01 -2.69911885e-01 -5.90606391e-01 -5.72740018e-01
-7.60064647e-02 -7.53263712e-01 2.99657494e-01 -1.05000770e+00
-3.23146880e-01 -6.69551253e-01 -4.27836061e-01 9.24416542e-01
-4.36259061e-01 3.09195429e-01 1.20101348e-01 2.96508580e-01
-2.53488958e-01 1.03553921e-01 6.87269151e-01 1.94615796e-01
-6.04226053e-01 3.01074386e-01 -3.45619172e-01 8.85544062e-01
7.79134631e-01 -6.51993811e-01 -1.00932665e-01 4.60951552e-02
-1.57156318e-01 -1.31628305e-01 3.60213816e-01 -1.18817317e+00
1.09248951e-01 -1.94493175e-01 7.63807476e-01 -4.68080074e-01
1.22314818e-01 -1.10816658e+00 -1.21688232e-01 6.33190930e-01
-4.86338846e-02 2.03369305e-01 2.23086476e-01 3.93119901e-01
-3.01580608e-01 -7.63897836e-01 7.32758045e-01 2.14719892e-01
-3.58765185e-01 2.06692107e-02 -4.79808033e-01 -4.10990477e-01
1.37108827e+00 -7.38182783e-01 -1.20683759e-01 6.91704126e-03
-9.82994139e-01 3.13973986e-02 7.65674651e-01 1.90914616e-01
5.09828210e-01 -1.10180950e+00 -3.09753209e-01 6.91712856e-01
1.56334952e-01 -2.21824069e-02 -7.73506463e-02 9.55189228e-01
-4.93543029e-01 2.05081090e-01 -1.58636242e-01 -6.75620735e-01
-1.53569794e+00 6.54766679e-01 -8.51430818e-02 -6.99897781e-02
-7.70987213e-01 7.91864038e-01 -6.75825119e-01 -2.30262816e-01
3.25297922e-01 -4.64887202e-01 -1.26141116e-01 2.98019588e-01
3.32768410e-01 5.70553660e-01 6.76085532e-01 -3.43568057e-01
-5.75228989e-01 3.32467824e-01 -5.16848385e-01 2.29599088e-01
1.37513900e+00 3.30710262e-01 -1.54587567e-01 5.18759549e-01
1.18725467e+00 3.85183901e-01 -9.91093993e-01 1.55897141e-01
3.30980778e-01 -8.43722641e-01 -3.04908961e-01 -3.93997252e-01
-8.89439702e-01 6.53275788e-01 3.61195415e-01 6.56991601e-01
1.08410227e+00 -3.65133733e-01 5.99108040e-01 4.69003052e-01
2.04307377e-01 -1.35300219e+00 -2.58775145e-01 5.00920773e-01
4.66371059e-01 -1.36740148e+00 2.11512581e-01 -4.52307761e-01
-7.02951670e-01 1.31392300e+00 6.02973878e-01 -4.62243915e-01
4.95132685e-01 1.47824064e-01 6.05529472e-02 9.02673751e-02
-4.80385870e-01 3.65684442e-02 1.08974546e-01 4.20153618e-01
7.85874799e-02 -1.04868561e-01 -2.58819103e-01 3.25536638e-01
-1.97793931e-01 -1.88962176e-01 6.93948686e-01 1.23218131e+00
-7.88697779e-01 -1.24579310e+00 -6.01804733e-01 1.09043252e+00
-2.12086916e-01 -6.75309151e-02 -2.11415395e-01 7.55080044e-01
2.39680126e-01 9.75860834e-01 1.89422131e-01 -2.99694061e-01
6.97218001e-01 4.18153316e-01 1.03290007e-01 -5.72568834e-01
-5.76337218e-01 -8.05208907e-02 1.68596819e-01 -6.07013047e-01
1.36870250e-01 -8.49314034e-01 -1.15983689e+00 -3.09588760e-01
-7.72768497e-01 3.75156343e-01 6.13245010e-01 9.75036621e-01
2.24325925e-01 1.89286292e-01 9.39138830e-01 -7.12597072e-01
-1.00632012e+00 -5.60581267e-01 -6.03335261e-01 4.74507362e-01
5.73561966e-01 -8.25161695e-01 -1.16251075e+00 -1.19140230e-01] | [7.70510721206665, 2.6490941047668457] |
d5f89c14-4eaf-4ac3-8deb-6db41c6538cd | unsupervised-video-analysis-based-on-a | 1503.06917 | null | http://arxiv.org/abs/1503.06917v1 | http://arxiv.org/pdf/1503.06917v1.pdf | Unsupervised Video Analysis Based on a Spatiotemporal Saliency Detector | Visual saliency, which predicts regions in the field of view that draw the
most visual attention, has attracted a lot of interest from researchers. It has
already been used in several vision tasks, e.g., image classification, object
detection, foreground segmentation. Recently, the spectrum analysis based
visual saliency approach has attracted a lot of interest due to its simplicity
and good performance, where the phase information of the image is used to
construct the saliency map. In this paper, we propose a new approach for
detecting spatiotemporal visual saliency based on the phase spectrum of the
videos, which is easy to implement and computationally efficient. With the
proposed algorithm, we also study how the spatiotemporal saliency can be used
in two important vision task, abnormality detection and spatiotemporal interest
point detection. The proposed algorithm is evaluated on several commonly used
datasets with comparison to the state-of-art methods from the literature. The
experiments demonstrate the effectiveness of the proposed approach to
spatiotemporal visual saliency detection and its application to the above
vision tasks | ['Yilin Wang', 'Qiang Zhang', 'Baoxin Li'] | 2015-03-24 | null | null | null | null | ['interest-point-detection', 'foreground-segmentation'] | ['computer-vision', 'computer-vision'] | [ 5.45308530e-01 -4.45298076e-01 -3.07972968e-01 -8.16604570e-02
-1.85395285e-01 -1.03395492e-01 4.68158156e-01 4.69780356e-01
-2.88169235e-01 6.42955422e-01 1.13202278e-02 8.96730796e-02
-1.78453967e-01 -3.87114942e-01 -4.88551736e-01 -7.86683381e-01
6.91154748e-02 -3.30438972e-01 1.28249407e+00 5.02470769e-02
1.07094598e+00 3.66634190e-01 -2.15525699e+00 -2.43662938e-01
1.05588710e+00 1.11093259e+00 8.19843948e-01 2.92109728e-01
-3.06992847e-02 8.88463438e-01 -3.65511745e-01 1.98550314e-01
-1.08506270e-01 -5.05173683e-01 -6.85060382e-01 2.08494186e-01
1.51098385e-01 1.48387253e-01 -3.91778126e-02 1.25294173e+00
3.00894588e-01 4.03766751e-01 5.53598881e-01 -1.34079945e+00
-5.17404675e-01 1.22178502e-01 -1.00620258e+00 9.80608523e-01
1.78888083e-01 -3.06784183e-01 7.72385418e-01 -9.69185948e-01
5.02487719e-01 9.68300223e-01 2.19949976e-01 1.21960945e-01
-7.07881570e-01 -3.25575948e-01 1.01041600e-01 9.59599674e-01
-1.17488182e+00 -4.27596495e-02 1.27939785e+00 -3.94937724e-01
4.31478083e-01 3.46953034e-01 7.98477709e-01 3.58716935e-01
2.68794805e-01 1.09945571e+00 1.29097807e+00 -5.06473184e-01
2.47890890e-01 1.58569083e-01 3.34717959e-01 5.94979465e-01
3.25378716e-01 -1.23487934e-01 -6.11894608e-01 7.30918050e-02
6.74357891e-01 1.87491298e-01 -3.44240218e-01 -6.03517413e-01
-1.24912953e+00 7.54204154e-01 7.39761531e-01 6.02264345e-01
-4.67008889e-01 -1.48653984e-03 2.36841127e-01 -4.34812486e-01
5.90018094e-01 2.38180071e-01 3.07304747e-02 8.92899632e-02
-1.18463194e+00 2.36067742e-01 4.12648804e-02 7.17292845e-01
5.73870659e-01 2.12026685e-01 -2.16049969e-01 8.60700071e-01
3.11880827e-01 4.86889780e-01 6.71659708e-01 -6.15047634e-01
8.82471539e-03 7.04817116e-01 2.71860927e-01 -1.50055945e+00
-3.47824007e-01 -3.37524414e-01 -5.81220388e-01 1.31611779e-01
1.83803752e-01 2.35668659e-01 -1.00208557e+00 1.40090263e+00
4.88943696e-01 7.41549075e-01 -2.61997129e-03 1.30493534e+00
1.06348228e+00 7.61699021e-01 1.65418703e-02 -3.17609936e-01
1.35015869e+00 -9.59118724e-01 -6.69330955e-01 -3.67142349e-01
-1.56947687e-01 -1.08006060e+00 7.36847699e-01 6.80125058e-02
-9.45113003e-01 -6.98026717e-01 -1.09338307e+00 6.40134364e-02
-3.12377304e-01 2.61125028e-01 4.86050159e-01 2.85316855e-01
-9.91328478e-01 1.71589091e-01 -7.78073847e-01 -8.34377110e-01
4.09252733e-01 3.02347898e-01 1.16263099e-01 1.61530241e-01
-1.07691777e+00 9.66464341e-01 7.69970238e-01 -2.35037897e-02
-7.87097812e-01 -2.16497540e-01 -8.08363080e-01 1.67728469e-01
4.53414351e-01 -2.96275228e-01 9.29492533e-01 -1.11067843e+00
-1.01418686e+00 7.25463867e-01 -6.00835562e-01 -5.03625214e-01
1.06815234e-01 2.74844235e-04 -2.86397487e-01 5.61363220e-01
4.10341710e-01 6.03724301e-01 1.13178444e+00 -9.61544335e-01
-1.12803853e+00 -2.83716887e-01 -2.04906791e-01 3.63832712e-01
-1.19815208e-01 2.52740890e-01 -3.68089169e-01 -8.41233313e-01
3.64572704e-01 -8.59911680e-01 -1.56461313e-01 -2.81308174e-01
-4.01385874e-01 -4.21280384e-01 1.30127740e+00 -5.87917984e-01
1.09617281e+00 -2.02488875e+00 1.15479566e-01 1.35099098e-01
7.54249766e-02 4.31974918e-01 4.31983650e-01 1.60524517e-01
7.70504996e-02 -1.24670088e-01 -4.06841785e-01 2.52773583e-01
-6.03501201e-01 -2.69954205e-01 -2.12499544e-01 5.77504337e-01
3.30571413e-01 7.02383101e-01 -8.80975127e-01 -9.31018412e-01
6.91596687e-01 3.01855624e-01 -7.89562017e-02 2.06265762e-01
7.23658726e-02 4.85749453e-01 -7.29194880e-01 7.29897857e-01
6.19761050e-01 -2.81479299e-01 -4.14291859e-01 -1.95210859e-01
-5.78918278e-01 -8.60617384e-02 -1.07050765e+00 1.24806333e+00
1.50925219e-01 1.05573595e+00 -2.81231284e-01 -1.19471908e+00
9.80099678e-01 7.65749887e-02 5.88811278e-01 -6.39094889e-01
1.27443165e-01 3.22659314e-01 1.35576099e-01 -5.71281135e-01
6.60921395e-01 1.54617861e-01 2.91321635e-01 1.97531223e-01
-8.04899111e-02 3.58511694e-03 3.64989936e-01 -2.76663229e-02
5.62986910e-01 -2.32183635e-02 5.74621439e-01 -4.35992092e-01
1.05245113e+00 3.36300164e-01 4.62547868e-01 4.32335407e-01
-4.54817116e-01 5.04216611e-01 -6.67160526e-02 -1.41605869e-01
-8.58534932e-01 -9.11797285e-01 -8.34025741e-02 8.17174613e-01
1.15310681e+00 1.05246589e-01 -7.88548112e-01 -3.63699198e-01
-2.07430884e-01 4.81937051e-01 -5.76016307e-01 -1.01123624e-01
-6.13912940e-01 -5.46503067e-01 -1.01789594e-01 4.24098402e-01
9.55441892e-01 -1.49619853e+00 -1.41079342e+00 4.13956307e-02
-2.74725139e-01 -9.99452829e-01 -2.97232568e-01 -2.51379877e-01
-9.98837888e-01 -1.10829425e+00 -1.18427551e+00 -1.34665763e+00
6.41328514e-01 1.16590488e+00 6.12016678e-01 1.66737646e-01
-4.68704045e-01 4.21535313e-01 -4.39769447e-01 -5.77384472e-01
1.55931205e-01 1.37903631e-01 -2.06566930e-01 4.25020695e-01
3.28000575e-01 -1.87845111e-01 -7.88908839e-01 5.07051885e-01
-1.01659036e+00 2.12451488e-01 5.93120515e-01 7.24747717e-01
6.08661771e-01 1.24106362e-01 8.07763696e-01 -6.12070322e-01
4.37642097e-01 -4.63266313e-01 -7.82355905e-01 1.73925340e-01
-3.00460339e-01 -8.54586139e-02 2.42680505e-01 -2.35458523e-01
-1.12391663e+00 -3.50623904e-03 2.24460095e-01 -3.47201079e-01
-9.52504650e-02 5.73705733e-01 8.16143975e-02 -4.24801409e-01
2.16276616e-01 6.82885826e-01 -8.71470645e-02 -2.98708200e-01
1.89763661e-02 5.91102779e-01 6.29247963e-01 1.41541407e-01
6.75049424e-01 4.71801609e-01 1.85085177e-01 -1.12049484e+00
-7.32983232e-01 -1.00381136e+00 -5.28269172e-01 -4.69963700e-01
1.05978107e+00 -6.16772830e-01 -5.09585679e-01 5.83599031e-01
-1.10944116e+00 4.85969007e-01 1.48084775e-01 5.88760138e-01
-6.57867432e-01 5.08262694e-01 -4.05402295e-02 -1.11213267e+00
-3.97423625e-01 -1.28216934e+00 1.01709521e+00 9.08894837e-01
1.46855339e-01 -9.28388596e-01 -3.65830630e-01 1.28007352e-01
3.78155589e-01 3.22822779e-01 7.71569908e-01 -4.08123583e-01
-1.03251827e+00 1.96697474e-01 -4.74876404e-01 7.36461673e-03
3.64926070e-01 -1.37648672e-01 -7.94698596e-01 -2.68871207e-02
2.15745404e-01 6.30268008e-02 8.42029154e-01 8.36158693e-01
9.96037483e-01 1.53660551e-01 -6.77496552e-01 2.11088553e-01
1.57301652e+00 5.95809400e-01 5.63397706e-01 4.83331382e-01
6.56429589e-01 6.45389616e-01 1.39060903e+00 2.55319536e-01
3.73659102e-04 7.56487131e-01 5.53530633e-01 -3.90453219e-01
-2.72044539e-02 -4.98951115e-02 -4.89062890e-02 7.37205863e-01
-1.51068911e-01 3.45645249e-02 -7.79942930e-01 1.00568330e+00
-2.14674711e+00 -1.08857989e+00 -3.17546695e-01 2.13335419e+00
4.05968070e-01 6.62687197e-02 1.44627854e-01 3.61354589e-01
1.18025124e+00 3.28034520e-01 -5.99647820e-01 -1.57769755e-01
-9.32873040e-02 -7.81591143e-03 4.60261941e-01 1.98787630e-01
-1.38543081e+00 1.04513264e+00 5.86774540e+00 8.54588389e-01
-1.51473355e+00 1.16709188e-01 6.27990365e-01 4.67168778e-01
1.91383556e-01 -4.62578200e-02 -4.95970130e-01 7.88507164e-01
3.52095991e-01 -4.30657208e-01 3.87587696e-02 8.46474051e-01
4.39457685e-01 -8.17760289e-01 -5.50553620e-01 1.14444983e+00
3.93086672e-01 -1.30531406e+00 -2.27936544e-02 -3.37718159e-01
7.42368460e-01 -2.90578425e-01 3.43416452e-01 -4.65330720e-01
-4.46803868e-01 -8.05517197e-01 7.17501760e-01 3.91568691e-01
3.09882969e-01 -8.34407210e-01 8.25891078e-01 3.65595102e-01
-1.40960658e+00 -1.83710411e-01 -3.68488312e-01 6.52368553e-03
2.48112693e-01 4.10692960e-01 -1.02679038e+00 3.66164118e-01
8.43597591e-01 1.05793858e+00 -8.26937139e-01 1.80583084e+00
-1.92915991e-01 6.66377842e-01 -7.92203844e-02 -5.07943273e-01
2.94229716e-01 -2.07907125e-01 8.09079587e-01 8.46060991e-01
4.21841174e-01 -2.48872268e-04 2.10549593e-01 9.09526944e-01
4.72971529e-01 4.25590247e-01 -6.24852955e-01 8.59593451e-02
3.09240609e-01 1.06352365e+00 -1.37566566e+00 -4.08176243e-01
-3.22294861e-01 8.05363536e-01 -1.51049733e-01 2.22777545e-01
-8.12514424e-01 -6.14945590e-01 1.90734148e-01 1.61820814e-01
5.94894946e-01 -1.09650984e-01 -2.34013066e-01 -9.49035704e-01
-3.53841074e-02 -3.71579289e-01 2.08427623e-01 -1.04645777e+00
-7.48021424e-01 3.66611212e-01 3.52164477e-01 -1.45862806e+00
-1.45322248e-01 -3.28691691e-01 -7.73298800e-01 9.07660604e-01
-1.72496355e+00 -1.01685739e+00 -4.98918027e-01 6.99457586e-01
9.97745395e-01 -1.10690355e-01 -2.66453829e-02 6.57336563e-02
-3.67069066e-01 -2.89296880e-02 -8.97781923e-03 -8.20284337e-02
3.92272800e-01 -1.10970581e+00 2.31281519e-01 1.34802282e+00
4.33900282e-02 5.11988461e-01 9.78470981e-01 -7.57270098e-01
-1.10215902e+00 -1.08272552e+00 8.09350848e-01 6.11638352e-02
2.86025435e-01 1.13319039e-01 -8.75395596e-01 1.99097455e-01
3.12392443e-01 -5.35560250e-02 1.95817471e-01 -6.40591025e-01
5.28981447e-01 -1.06166452e-01 -1.13078964e+00 4.96054769e-01
6.49636090e-01 -1.39763847e-01 -8.07191312e-01 -4.42375615e-02
4.97278780e-01 -3.23640347e-01 -1.35764495e-01 4.46427226e-01
3.14109445e-01 -9.85166132e-01 9.94633317e-01 9.75221395e-02
1.95455745e-01 -8.39715600e-01 2.09643409e-01 -1.19205379e+00
-3.55256349e-01 -2.48889640e-01 5.67897074e-02 1.08339524e+00
-1.30957291e-01 -6.10369444e-01 5.89586794e-01 -6.70038089e-02
-1.84531901e-02 -6.44719958e-01 -1.00437915e+00 -2.79422194e-01
-8.08434784e-01 5.18290214e-02 2.03365862e-01 5.35297334e-01
-1.60857022e-01 4.03067052e-01 -4.17621583e-01 1.84132934e-01
7.08572447e-01 5.69367826e-01 4.12971348e-01 -1.48896825e+00
1.14741519e-01 -3.81820440e-01 -8.79580021e-01 -9.98072863e-01
-1.61139265e-01 -5.85369706e-01 1.24656774e-01 -1.64163113e+00
3.27192128e-01 -2.25710586e-01 -5.84843040e-01 3.94507274e-02
-5.65456212e-01 4.88974690e-01 3.75876695e-01 3.17702353e-01
-7.32568204e-01 5.13268888e-01 1.27938259e+00 -2.20309436e-01
-2.81653106e-01 2.09917590e-01 -6.73350632e-01 7.79361129e-01
7.93662488e-01 -3.18701357e-01 -6.54702365e-01 -1.06582716e-02
-4.11427498e-01 -1.16924457e-01 4.64203298e-01 -1.22434783e+00
3.18345696e-01 -3.07572693e-01 4.83955920e-01 -9.44334984e-01
3.30420047e-01 -7.73948729e-01 -1.52089119e-01 5.44021010e-01
-1.09342054e-01 1.12224348e-01 9.53258276e-02 6.23199821e-01
-5.91538489e-01 -4.11823183e-01 1.11191654e+00 -8.45969990e-02
-1.37296617e+00 1.69423707e-02 -3.52588475e-01 -7.82859996e-02
1.49324763e+00 -5.30616701e-01 -2.07603186e-01 -1.90411314e-01
-3.51237357e-01 3.78415138e-02 3.45390856e-01 5.86524129e-01
1.12678993e+00 -1.29093909e+00 -3.54431629e-01 1.83534145e-01
8.20191205e-02 -3.40988457e-01 2.57062584e-01 1.16786587e+00
-6.26862645e-01 5.79496503e-01 -5.28615177e-01 -1.16535330e+00
-1.55761909e+00 8.49982679e-01 -1.12757891e-01 2.36023262e-01
-6.06290460e-01 4.30318266e-01 3.83983701e-01 6.25186622e-01
8.46240297e-02 -6.34721458e-01 -8.85761917e-01 -1.92990556e-01
5.24409413e-01 5.59551179e-01 -2.70601124e-01 -1.15778756e+00
-6.25088930e-01 1.15047312e+00 2.01161012e-01 1.37891527e-02
8.98805559e-01 -3.32788825e-01 -1.34754881e-01 6.66322470e-01
8.28237593e-01 4.51669842e-03 -9.57237005e-01 -1.89715207e-01
3.28161746e-01 -8.21722031e-01 1.77854910e-01 -4.33777392e-01
-1.00757420e+00 1.02432179e+00 9.38420951e-01 5.27145922e-01
1.34607768e+00 1.93926003e-02 8.08904767e-01 -2.39085183e-01
6.34208739e-01 -1.03989983e+00 2.23416731e-01 2.31841266e-01
6.84172809e-01 -1.43091381e+00 5.63073196e-02 -7.80962527e-01
-7.46882498e-01 8.03610623e-01 5.75006366e-01 -3.58810663e-01
6.66984499e-01 -3.49940360e-01 -6.67648837e-02 -2.68245135e-02
-1.94568872e-01 -6.86058044e-01 6.44046128e-01 7.03742266e-01
2.32254267e-01 -1.19520910e-01 -7.14389980e-01 1.56955719e-01
2.27478966e-01 6.28655627e-02 5.24127781e-01 1.07664335e+00
-1.03261757e+00 -6.88973427e-01 -6.25045836e-01 4.31741416e-01
-6.17786527e-01 5.88010624e-02 -2.74965435e-01 5.93820155e-01
1.54163614e-01 1.09542179e+00 -4.39744517e-02 4.44168448e-02
4.28855047e-02 -3.49646509e-01 2.63196886e-01 -5.10363698e-01
-2.09829733e-01 1.74954697e-01 -4.98833716e-01 -2.92810172e-01
-1.07248783e+00 -6.80691361e-01 -1.36113942e+00 4.08677846e-01
-4.32646155e-01 4.09245819e-01 5.46126962e-01 8.76540840e-01
2.88518310e-01 4.66731131e-01 6.14324093e-01 -1.01388705e+00
1.72179207e-01 -9.46964562e-01 -6.07014298e-01 4.30562556e-01
4.45794851e-01 -1.30736947e+00 -8.00252706e-02 7.88776577e-02] | [9.769124031066895, -0.46672648191452026] |
a8cbaa56-b26b-4372-a2f4-416dcc18d17d | chaotic-variational-auto-encoder-based-one | 2212.07802 | null | https://arxiv.org/abs/2212.07802v1 | https://arxiv.org/pdf/2212.07802v1.pdf | Chaotic Variational Auto Encoder based One Class Classifier for Insurance Fraud Detection | Of late, insurance fraud detection has assumed immense significance owing to the huge financial & reputational losses fraud entails and the phenomenal success of the fraud detection techniques. Insurance is majorly divided into two categories: (i) Life and (ii) Non-life. Non-life insurance in turn includes health insurance and auto insurance among other things. In either of the categories, the fraud detection techniques should be designed in such a way that they capture as many fraudulent transactions as possible. Owing to the rarity of fraudulent transactions, in this paper, we propose a chaotic variational autoencoder (C-VAE to perform one-class classification (OCC) on genuine transactions. Here, we employed the logistic chaotic map to generate random noise in the latent space. The effectiveness of C-VAE is demonstrated on the health insurance fraud and auto insurance datasets. We considered vanilla Variational Auto Encoder (VAE) as the baseline. It is observed that C-VAE outperformed VAE in both datasets. C-VAE achieved a classification rate of 77.9% and 87.25% in health and automobile insurance datasets respectively. Further, the t-test conducted at 1% level of significance and 18 degrees of freedom infers that C-VAE is statistically significant than the VAE. | ['Vadlamani Ravi', 'Yelleti Vivek', 'B. Akhil Kumar', 'K. S. N. V. K. Gangadhar'] | 2022-12-15 | null | null | null | null | ['one-class-classifier', 'one-class-classification'] | ['methodology', 'miscellaneous'] | [-3.70405436e-01 3.39204329e-03 7.68847540e-02 -1.81075603e-01
-5.11832535e-01 -2.18281433e-01 3.71629715e-01 -1.37647197e-01
-3.70647758e-01 8.96982789e-01 2.19392285e-01 -4.15163696e-01
1.24852382e-01 -9.39446211e-01 -7.20668137e-01 -6.00949168e-01
2.93468926e-02 2.05473796e-01 -2.68106610e-01 -4.49219882e-01
3.35311651e-01 1.21310502e-01 -1.49912190e+00 3.50510210e-01
1.00678384e+00 1.08734810e+00 -1.55485258e-01 3.84058207e-01
6.06686659e-02 9.41967010e-01 -6.29963160e-01 -1.01216924e+00
2.15292096e-01 -6.03024602e-01 -4.34572846e-01 -9.39942673e-02
-3.63232851e-01 -6.36151791e-01 -2.69866109e-01 1.12712753e+00
1.54724922e-02 -1.52317971e-01 1.04914308e+00 -1.38184226e+00
-8.45645010e-01 3.58976841e-01 -6.72302604e-01 1.30047619e-01
1.31477222e-01 -1.87976182e-01 8.42279732e-01 -8.33585978e-01
5.62594950e-01 8.28564644e-01 9.38811779e-01 4.74075764e-01
-1.00439990e+00 -6.81699753e-01 -5.76987207e-01 9.50739626e-03
-1.19274521e+00 -1.11722648e-02 9.07147288e-01 -6.48368061e-01
1.02058887e+00 1.09991416e-01 7.71383584e-01 1.28623605e+00
6.53781474e-01 7.01522946e-01 1.00631785e+00 -3.70231092e-01
2.27862358e-01 4.83910561e-01 -8.66130553e-03 6.80897117e-01
8.22053015e-01 2.85148382e-01 -1.80320963e-01 -5.62197685e-01
8.44826937e-01 4.78186429e-01 -3.78075242e-02 -8.11469927e-02
-1.08634627e+00 1.32858348e+00 1.49201468e-01 3.27108711e-01
-8.54889393e-01 1.91001847e-01 7.03566790e-01 5.39508998e-01
1.84097514e-01 5.52288406e-02 -2.12705553e-01 -2.58482397e-01
-8.89060199e-01 3.29436511e-01 8.50093186e-01 6.68231905e-01
2.34066471e-01 3.45628768e-01 1.30193874e-01 6.78018689e-01
4.26682472e-01 1.89170107e-01 8.99214685e-01 -6.81463122e-01
6.58309340e-01 6.94592178e-01 2.17486143e-01 -1.24077392e+00
2.46080831e-01 -1.43055797e-01 -1.16388643e+00 2.10638121e-01
2.78480023e-01 -1.72219679e-01 -6.13198400e-01 1.37224638e+00
-1.63740143e-01 4.41928674e-03 5.59398770e-01 7.98846006e-01
2.10554689e-01 4.34874624e-01 -1.08146280e-01 -3.46203297e-01
1.32840145e+00 -4.21989620e-01 -1.01968181e+00 4.45641696e-01
3.82128030e-01 -6.23167813e-01 8.76592398e-01 4.96363521e-01
-8.83274615e-01 -3.04735452e-01 -1.00481856e+00 1.07366778e-01
-2.03675449e-01 1.49944499e-01 7.09612012e-01 9.08643126e-01
-5.16510665e-01 5.03187180e-01 -9.29175913e-01 -6.46377057e-02
6.32432222e-01 8.31770897e-02 -1.84019446e-01 1.75513968e-01
-1.37456596e+00 7.48949826e-01 1.79040700e-01 5.65708503e-02
-6.75933301e-01 3.05256210e-02 -6.90842271e-01 -1.99943245e-03
-1.00083351e-01 -4.97570962e-01 9.75437582e-01 -9.15649652e-01
-9.45663214e-01 8.41669798e-01 -1.90157872e-02 -1.04177630e+00
1.03293252e+00 -1.05282545e-01 -4.62272018e-01 6.99101537e-02
1.31666869e-01 1.43401727e-01 8.07579279e-01 -9.66008186e-01
-4.80504155e-01 -3.65601689e-01 -3.35667759e-01 2.13323836e-03
-2.72250146e-01 -2.60494471e-01 2.29655802e-01 -1.04079723e+00
2.50599563e-01 -8.23457539e-01 2.56529123e-01 -3.38501424e-01
-1.64674193e-01 -1.86326638e-01 7.84251213e-01 -1.04605281e+00
1.42511666e+00 -2.00145173e+00 -2.46003568e-01 1.02483317e-01
-2.58561149e-02 6.92387596e-02 8.14429224e-01 6.42519414e-01
-1.80215146e-02 7.84066617e-02 -6.33355498e-01 -7.78187737e-02
-1.55167800e-05 3.82559091e-01 -6.51639104e-01 3.57661128e-01
3.28032523e-01 8.86386812e-01 -5.40471911e-01 -5.54818273e-01
5.85304797e-02 4.25523043e-01 -7.64792204e-01 5.90944327e-02
1.55962914e-01 2.53345165e-02 -4.74565238e-01 1.21636343e+00
6.70999706e-01 -1.29995733e-01 1.27512947e-01 -7.82003701e-02
1.26630217e-01 -2.27772415e-01 -9.28840160e-01 1.04634857e+00
-4.51150574e-02 4.92173612e-01 -2.59153217e-01 -1.20690250e+00
9.66503322e-01 7.20822453e-01 6.05884135e-01 -5.41348696e-01
3.05427253e-01 6.25062466e-01 -1.67469949e-01 -7.14365602e-01
5.17859221e-01 -5.35355330e-01 -3.03931803e-01 3.31353873e-01
-2.49603927e-01 4.18440580e-01 -3.07226479e-01 7.02358484e-02
7.61811078e-01 3.48548964e-02 5.45704842e-01 -3.64533290e-02
3.85897636e-01 1.89960912e-01 7.61982501e-01 5.10044456e-01
-6.98366106e-01 6.52263939e-01 6.76576555e-01 -4.89586681e-01
-1.03635216e+00 -9.63246346e-01 -3.33040297e-01 2.58239806e-01
6.10697530e-02 3.92992169e-01 -8.22799861e-01 -6.94404721e-01
5.89639187e-01 8.12652886e-01 -6.78124309e-01 -1.30714312e-01
-4.60068375e-01 -1.01270688e+00 8.54170382e-01 4.07810509e-01
8.43340158e-01 -1.31816149e+00 -1.00958884e+00 2.46022552e-01
-3.36117417e-01 -6.87074363e-01 -2.52357394e-01 -1.76772222e-01
-1.07632327e+00 -1.30569839e+00 -9.28193390e-01 -7.94871628e-01
4.89062250e-01 -6.69249073e-02 7.59811938e-01 -1.79539248e-02
-1.90380007e-01 -7.17694014e-02 -3.16204518e-01 -5.61288655e-01
-6.49129272e-01 -4.75689471e-01 2.49811321e-01 2.60368675e-01
9.76677716e-01 -3.40901613e-01 -5.65305769e-01 1.36051655e-01
-9.14822459e-01 -3.07435721e-01 4.88016963e-01 1.32001209e+00
5.09156704e-01 3.09507877e-01 8.59920025e-01 -7.41912425e-01
1.03291869e+00 -8.20935965e-01 -4.18564945e-01 1.29440233e-01
-9.68182266e-01 -8.25376958e-02 2.61280894e-01 -2.61250347e-01
-8.22322190e-01 -3.47182542e-01 -1.32798791e-01 -7.55809486e-01
1.54156983e-01 5.72660744e-01 2.89487690e-01 3.88496339e-01
3.87361497e-01 2.78306752e-01 1.32536471e-01 -3.65972966e-01
-3.66077095e-01 1.16596878e+00 5.51901162e-01 -1.71479777e-01
2.13484824e-01 5.46282530e-01 -2.77957737e-01 -6.58167660e-01
-7.30105788e-02 -1.60319880e-01 -1.29729033e-01 1.44273164e-02
1.09352863e+00 -8.71021688e-01 -1.04208696e+00 8.58670652e-01
-1.06065488e+00 4.14313912e-01 -1.96792260e-01 7.56006181e-01
-5.33074319e-01 4.21454847e-01 -8.41481566e-01 -1.42124557e+00
-5.63380539e-01 -1.07572913e+00 4.53130484e-01 -3.23909298e-02
-1.98404774e-01 -8.88544559e-01 -1.19176202e-01 4.59899426e-01
2.47278184e-01 9.90168691e-01 8.21782231e-01 -5.32667458e-01
-2.40323052e-01 -5.59758425e-01 -1.18422508e-01 6.37497842e-01
6.48504943e-02 -1.69636190e-01 -9.76460934e-01 -1.70661598e-01
6.58372223e-01 -2.25616559e-01 9.47590411e-01 4.72859383e-01
7.98183799e-01 -4.49484140e-01 -2.89406925e-02 3.37260932e-01
1.73741007e+00 6.72547996e-01 9.13132310e-01 5.32825768e-01
2.69934297e-01 4.80482757e-01 6.74883008e-01 5.36484361e-01
4.53617781e-01 3.94933164e-01 4.90369886e-01 1.59709573e-01
6.37779832e-01 -5.04210532e-01 3.99018258e-01 9.69603240e-01
-1.84006006e-01 6.48258626e-02 -8.12021911e-01 7.81536102e-01
-1.81000412e+00 -1.33150136e+00 -3.16122085e-01 2.24591231e+00
6.35644615e-01 2.63345093e-01 8.32297504e-02 5.03631592e-01
6.64983630e-01 -2.89504409e-01 -6.99142694e-01 -7.16301739e-01
-2.06984356e-01 -1.43685296e-01 5.98451912e-01 1.65192395e-01
-1.07579911e+00 3.85169923e-01 5.73850012e+00 4.95521873e-01
-7.95095742e-01 2.07416236e-01 7.98835278e-01 1.62851453e-01
-2.33594552e-01 -1.39345884e-01 -3.56479824e-01 1.01052737e+00
8.49144638e-01 -1.94241498e-02 2.67471015e-01 7.98681259e-01
5.06384298e-02 -1.38531610e-01 -5.17432034e-01 9.93569791e-01
-1.96054205e-01 -1.12700164e+00 2.97189923e-03 4.41870064e-01
6.06370389e-01 -1.24808513e-01 -1.06261067e-01 4.78239655e-01
6.43341169e-02 -9.61365879e-01 7.27441788e-01 8.46913576e-01
8.14139366e-01 -1.06628597e+00 1.29213655e+00 5.20722806e-01
-7.98006475e-01 -3.04651678e-01 -5.14382899e-01 -2.20848639e-02
2.25042015e-01 5.53416371e-01 -4.35060292e-01 4.99137878e-01
7.97885418e-01 4.75375444e-01 5.11480086e-02 4.13219959e-01
3.49172443e-01 8.39008629e-01 -6.94364831e-02 1.45388678e-01
1.34505138e-01 -6.37527883e-01 4.51480955e-01 1.01858842e+00
5.36003470e-01 5.33734001e-02 -3.70627671e-01 1.01996541e+00
-1.07032306e-01 -2.33449638e-01 -7.89913476e-01 -2.80459404e-01
2.74361223e-01 5.05813718e-01 -3.42533827e-01 -3.94739538e-01
-4.02589202e-01 1.27900159e+00 -1.70605078e-01 2.76789606e-01
-9.34615433e-01 -7.20780015e-01 7.42930055e-01 5.08536957e-02
4.63738739e-01 1.68349028e-01 -5.02861202e-01 -1.44736838e+00
2.85910100e-01 -7.87342906e-01 3.94356281e-01 -4.21547800e-01
-1.23186958e+00 5.84258795e-01 -2.14016959e-01 -1.41633332e+00
-5.64654112e-01 -4.21767563e-01 -3.26423347e-01 8.51131618e-01
-1.45268369e+00 -7.11650312e-01 -1.38991505e-01 6.40912175e-01
2.50029147e-01 -6.89813614e-01 9.02802765e-01 3.82452786e-01
-4.40336823e-01 7.70220995e-01 4.97256488e-01 4.23887908e-01
2.42710620e-01 -1.13499975e+00 2.94421434e-01 4.80089545e-01
-4.20213677e-02 7.14209616e-01 6.76934719e-01 -8.16888273e-01
-1.11021519e+00 -8.06292832e-01 1.26404333e+00 -4.25128937e-01
3.79133910e-01 -1.01239994e-01 -1.09051144e+00 6.10241890e-01
-5.29509690e-03 -2.85759389e-01 4.56588507e-01 -5.32545865e-01
-1.97558910e-01 1.20122738e-01 -1.86647356e+00 2.40894154e-01
3.92907232e-01 -6.24911845e-01 -8.29826474e-01 2.13203579e-02
5.54907024e-01 -2.59723961e-01 -1.00125861e+00 3.15298140e-01
8.91484380e-01 -1.48972917e+00 8.73960674e-01 -3.79306942e-01
7.91376591e-01 1.10245675e-01 -3.78286481e-01 -5.69822967e-01
2.37030029e-01 -2.36271307e-01 -4.23813462e-01 9.77872252e-01
-2.26147939e-03 -1.02560294e+00 8.04546952e-01 4.36593533e-01
2.99210012e-01 -1.11947453e+00 -1.01948559e+00 -6.44295037e-01
4.54234779e-02 -4.37939554e-01 5.87053061e-01 1.16749573e+00
-1.86026037e-01 -3.17681879e-01 -6.04341090e-01 -1.83215275e-01
6.97028100e-01 7.74662569e-02 4.26824152e-01 -1.08893895e+00
-4.45191830e-01 -1.97876006e-01 -4.88087088e-01 -3.91907036e-01
-3.51706207e-01 -6.52175009e-01 -6.07329726e-01 -1.12557840e+00
1.03136748e-01 -1.75951406e-01 -5.26125669e-01 1.39524966e-01
3.11192833e-02 2.23620772e-01 -2.05140948e-01 7.13527381e-01
2.71397650e-01 5.16635895e-01 8.16255748e-01 2.67107808e-03
-2.47020155e-01 2.80228525e-01 -7.03706861e-01 5.39020121e-01
7.50215888e-01 -5.08520126e-01 -1.44754633e-01 -2.28785463e-02
6.21170513e-02 4.91521001e-01 4.91714031e-01 -7.60367334e-01
-2.38673434e-01 -3.37593295e-02 2.80208558e-01 -7.30674505e-01
2.05261886e-01 -7.15761125e-01 2.43036389e-01 8.60271931e-01
-8.88729095e-02 3.50547463e-01 -1.26782492e-01 9.56946731e-01
-6.59194350e-01 -2.77997434e-01 5.22742867e-01 -1.48890242e-01
-3.31981361e-01 -2.13146120e-01 -2.78228253e-01 -6.89130649e-02
1.00055683e+00 -6.06666386e-01 1.46914363e-01 -4.75520164e-01
-3.51816267e-01 6.16571354e-03 3.98029268e-01 2.34574407e-01
8.12183082e-01 -1.47902393e+00 -8.53514373e-01 5.55587173e-01
4.70207147e-02 -3.10816646e-01 1.34713441e-01 9.03258502e-01
-7.81086624e-01 5.18723488e-01 -1.83928788e-01 -2.90839463e-01
-1.08202577e+00 3.99871647e-01 3.46546501e-01 -4.25813138e-01
-7.02115417e-01 5.84121883e-01 -2.61630327e-01 -1.50814116e-01
1.59810081e-01 -4.63472694e-01 -3.20114374e-01 1.90358549e-01
7.79569149e-02 6.07935369e-01 -1.21407650e-01 -8.07393849e-01
-2.34101638e-01 2.00042337e-01 -1.06108710e-01 -2.19382152e-01
1.28652573e+00 1.05386168e-01 1.63596272e-02 4.97680485e-01
1.33721828e+00 -1.12221800e-01 -1.03380942e+00 2.73752630e-01
2.38195226e-01 -4.13813293e-01 -1.18914552e-01 -6.02348089e-01
-9.40383792e-01 8.08233082e-01 7.54219949e-01 5.21790385e-01
1.06875145e+00 -6.55765057e-01 1.10634851e+00 1.83559224e-01
6.25965118e-01 -1.22728598e+00 -3.37376118e-01 1.51696652e-01
8.84051442e-01 -1.39573228e+00 -1.58971041e-01 2.21785143e-01
-1.23906767e+00 8.47346187e-01 -7.23345205e-02 -6.36506975e-01
7.55695522e-01 -1.15340889e-01 1.46212608e-01 -3.77057523e-01
-5.64675272e-01 3.42238814e-01 1.94872245e-02 2.41433501e-01
5.19215763e-01 2.05886811e-01 -7.46339083e-01 8.41784835e-01
-2.60592908e-01 5.00633419e-01 5.74753642e-01 1.09432566e+00
-1.48492664e-01 -7.17212796e-01 -5.16889274e-01 7.08915770e-01
-8.95262539e-01 2.16831669e-01 -2.47310959e-02 8.57475877e-01
1.38079733e-01 9.07353520e-01 2.19193816e-01 -4.61317509e-01
2.47803017e-01 3.07625860e-01 -1.61516428e-01 1.58054411e-01
-8.49237323e-01 5.71524054e-02 -2.32652366e-01 -2.74135768e-01
-2.74293870e-01 -8.15098524e-01 -1.24488103e+00 -4.12353456e-01
-3.64602894e-01 2.28022963e-01 7.84160316e-01 6.76818728e-01
2.33464420e-01 2.72714853e-01 6.47297859e-01 -1.53215945e-01
-1.11097598e+00 -9.90295768e-01 -9.44986343e-01 7.82662094e-01
4.89599943e-01 -6.83478713e-01 -5.02001643e-01 4.68151234e-02] | [7.802088737487793, 5.184650897979736] |
4412e11a-88d7-42b2-8d0d-f12a7de8a07a | unsupervised-domain-adaptation-for-point | 2208.04510 | null | https://arxiv.org/abs/2208.04510v1 | https://arxiv.org/pdf/2208.04510v1.pdf | Unsupervised Domain Adaptation for Point Cloud Semantic Segmentation via Graph Matching | Unsupervised domain adaptation for point cloud semantic segmentation has attracted great attention due to its effectiveness in learning with unlabeled data. Most of existing methods use global-level feature alignment to transfer the knowledge from the source domain to the target domain, which may cause the semantic ambiguity of the feature space. In this paper, we propose a graph-based framework to explore the local-level feature alignment between the two domains, which can reserve semantic discrimination during adaptation. Specifically, in order to extract local-level features, we first dynamically construct local feature graphs on both domains and build a memory bank with the graphs from the source domain. In particular, we use optimal transport to generate the graph matching pairs. Then, based on the assignment matrix, we can align the feature distributions between the two domains with the graph-based local feature loss. Furthermore, we consider the correlation between the features of different categories and formulate a category-guided contrastive loss to guide the segmentation model to learn discriminative features on the target domain. Extensive experiments on different synthetic-to-real and real-to-real domain adaptation scenarios demonstrate that our method can achieve state-of-the-art performance. | ['Jin Xie', 'Jianjun Qian', 'Le Hui', 'Yikai Bian'] | 2022-08-09 | null | null | null | null | ['graph-matching'] | ['graphs'] | [ 8.09401423e-02 -8.93690437e-02 -1.75291210e-01 -7.03301847e-01
-6.80536985e-01 -4.80023354e-01 1.62212685e-01 1.37243673e-01
-1.83942422e-01 3.78593862e-01 -2.23904043e-01 2.45266691e-01
-2.50870079e-01 -9.19338048e-01 -5.93695045e-01 -7.57777393e-01
3.17749828e-01 7.57100523e-01 5.88580012e-01 -6.09792769e-02
2.54664183e-01 4.41375464e-01 -1.14728403e+00 -8.98140520e-02
1.35079157e+00 1.15330088e+00 5.31209648e-01 -1.39080912e-01
-5.39259553e-01 6.02225587e-02 -3.19259912e-01 -7.94980153e-02
4.66160595e-01 -4.35663104e-01 -8.23936760e-01 6.49910986e-01
2.24823326e-01 1.02866881e-01 -3.08119297e-01 1.37592554e+00
3.36290866e-01 3.78875762e-01 7.25995123e-01 -1.32359242e+00
-5.29477060e-01 -7.53642544e-02 -6.78615332e-01 1.30692258e-01
1.62856847e-01 2.10472390e-01 8.73282969e-01 -7.71906495e-01
6.55665517e-01 1.21961737e+00 3.24259341e-01 3.37734193e-01
-9.77491260e-01 -7.71641731e-01 6.62633061e-01 4.17024136e-01
-1.42725992e+00 -2.20353231e-02 1.16603172e+00 -4.75157440e-01
2.92682439e-01 -2.64715582e-01 6.48212492e-01 4.74800527e-01
-1.74447060e-01 6.52803600e-01 9.80286837e-01 -3.00012141e-01
2.49298453e-01 1.47983238e-01 7.35605136e-02 7.58869529e-01
-3.24083716e-02 -1.10961132e-01 -4.79383677e-01 2.95394361e-02
8.33200276e-01 5.52570149e-02 -2.30028600e-01 -1.05422199e+00
-9.96374607e-01 8.12922418e-01 7.82821953e-01 1.91700220e-01
-2.01029718e-01 -4.71473396e-01 2.59550601e-01 1.31876260e-01
4.08478498e-01 1.81018680e-01 -4.70797300e-01 4.09892321e-01
-4.76928115e-01 3.84577364e-02 4.88488257e-01 1.19755125e+00
1.35122299e+00 -2.65180737e-01 -3.90864499e-02 1.07562780e+00
5.83402693e-01 3.90993059e-01 5.55075347e-01 -7.00208426e-01
7.55412459e-01 1.02359271e+00 -1.99011475e-01 -1.18924928e+00
-1.79405987e-01 -4.08186287e-01 -6.35197699e-01 -1.49847671e-01
4.21538562e-01 1.15520500e-01 -1.10734189e+00 1.65861309e+00
7.28006780e-01 5.07153630e-01 4.86161076e-02 1.21275747e+00
6.27656639e-01 4.43875730e-01 8.60847309e-02 -9.91478097e-03
1.07683277e+00 -9.66739118e-01 -2.11540520e-01 -6.10902607e-01
4.71211642e-01 -4.36590046e-01 1.16084099e+00 -2.15099439e-01
-5.57122469e-01 -6.39239252e-01 -1.02362287e+00 1.86631352e-01
-3.12748373e-01 -2.03095973e-01 3.23073804e-01 2.76499540e-01
-4.98062342e-01 4.80560690e-01 -7.07545459e-01 -4.97837216e-01
5.74183822e-01 3.52677852e-01 -2.36614540e-01 -3.43649507e-01
-1.16302037e+00 3.60391796e-01 7.34929919e-01 -8.25612172e-02
-5.93463898e-01 -4.84143555e-01 -9.06878531e-01 -9.87280533e-03
4.82741714e-01 -6.38244212e-01 7.96880543e-01 -1.11469126e+00
-1.36255395e+00 9.52391863e-01 -6.27317876e-02 -8.89468472e-03
3.90201598e-01 2.24831313e-01 -4.06870753e-01 1.70089439e-01
5.23890972e-01 6.94669664e-01 7.62317717e-01 -1.23301589e+00
-9.27219391e-01 -5.98880649e-01 -1.68051854e-01 7.32492447e-01
-3.66215795e-01 -4.21142131e-01 -8.50317717e-01 -5.36992013e-01
8.29730630e-01 -9.36678171e-01 -2.64779329e-01 2.02363394e-02
-2.97242343e-01 -2.16772318e-01 8.17951798e-01 -5.01248419e-01
7.12451696e-01 -2.40155101e+00 2.89412886e-01 6.00541413e-01
1.32453784e-01 -5.24756350e-02 -1.82234764e-01 -8.25478286e-02
1.22138567e-01 -1.55159473e-01 -7.45001018e-01 6.89924834e-03
-2.91474819e-01 4.65145051e-01 -7.20631778e-02 4.54327077e-01
2.51414180e-01 6.86368048e-01 -9.73028421e-01 -7.56935358e-01
1.50649652e-01 -7.68121183e-02 -5.79904199e-01 4.89960164e-01
-2.21561790e-01 8.83683801e-01 -1.04871953e+00 5.28834283e-01
9.21260297e-01 -3.30547094e-01 1.15197256e-01 -1.46888375e-01
4.24599379e-01 1.60917174e-02 -1.22330618e+00 1.98664999e+00
-2.76607960e-01 1.76364332e-01 -5.74153587e-02 -1.31166804e+00
1.23786652e+00 -2.59704441e-01 4.92123753e-01 -6.94108129e-01
-2.22163051e-02 4.05346930e-01 -1.46383196e-01 -2.87009299e-01
5.42806052e-02 -2.10881606e-01 -2.32825220e-01 -6.25268277e-03
1.62208930e-01 -3.64206642e-01 -1.39592543e-01 -1.26514465e-01
6.68148994e-01 6.25463426e-02 7.67261684e-02 -2.54317909e-01
8.91387105e-01 1.15218312e-01 8.67747307e-01 2.56878853e-01
-2.57264704e-01 7.45842814e-01 1.40373662e-01 -2.79661000e-01
-6.15721583e-01 -1.21150053e+00 1.84071790e-02 8.34384143e-01
9.95718360e-01 1.79072618e-01 -8.40909243e-01 -1.22071016e+00
3.55091356e-02 4.91105974e-01 -2.48902902e-01 -5.52249491e-01
-5.81030071e-01 -3.66872400e-01 -7.24364519e-02 3.93664479e-01
9.79199469e-01 -1.00171673e+00 -2.13509291e-01 2.16923922e-01
-2.06554428e-01 -1.17635107e+00 -7.21059263e-01 1.83065441e-02
-1.01455688e+00 -1.14503193e+00 -6.37581468e-01 -1.26504648e+00
1.08863544e+00 3.15383673e-01 8.61106098e-01 4.85669896e-02
1.62188470e-01 4.19153988e-01 -3.24620992e-01 5.86097017e-02
-1.62022501e-01 2.86995828e-01 -1.97758824e-01 3.21139812e-01
4.88468468e-01 -6.14942968e-01 -5.64120352e-01 6.99037254e-01
-8.49545181e-01 -1.03077054e-01 5.14251232e-01 8.08509111e-01
9.84418690e-01 2.42043789e-02 5.36070585e-01 -8.87256086e-01
3.40207994e-01 -5.66359699e-01 -6.90144777e-01 3.27078551e-01
-5.62752008e-01 2.17094451e-01 5.64518631e-01 -4.52327967e-01
-1.06481421e+00 2.98439622e-01 2.01952502e-01 -6.77963138e-01
-2.07943335e-01 6.34342194e-01 -8.70429039e-01 -1.70100555e-01
3.74995440e-01 3.43915224e-01 -1.33199170e-01 -4.07837451e-01
4.06569183e-01 6.00468397e-01 5.45605779e-01 -8.69425178e-01
1.04958224e+00 4.30693239e-01 -1.44079641e-01 -4.66304988e-01
-7.97243595e-01 -6.18792772e-01 -8.76651049e-01 -1.04970239e-01
8.75774086e-01 -8.43097925e-01 2.62453593e-02 6.04918003e-01
-9.27114069e-01 -2.26802900e-01 -3.79605204e-01 3.83453429e-01
-6.01903796e-01 5.02009034e-01 -5.85744157e-02 -2.29867950e-01
1.16060443e-01 -1.27858067e+00 1.13194025e+00 6.57845736e-01
2.58831710e-01 -1.06808710e+00 1.55057723e-03 2.10032061e-01
-8.45523775e-02 1.39939025e-01 1.02717996e+00 -8.54720116e-01
-7.30593204e-01 8.16526264e-02 -5.35182357e-01 3.42690319e-01
3.57134849e-01 -5.00343859e-01 -5.92341542e-01 -3.58552217e-01
1.15738980e-01 -7.35799596e-02 5.42816579e-01 2.47708455e-01
1.18254697e+00 -6.89790025e-02 -6.95908964e-01 8.42943728e-01
1.26994240e+00 2.94119865e-01 4.02475506e-01 3.48914802e-01
9.71493423e-01 7.33078301e-01 1.14904749e+00 2.21465871e-01
4.86957461e-01 7.18065083e-01 2.94689894e-01 1.19364308e-02
-9.70926657e-02 -4.79821682e-01 -3.87186781e-02 8.98884237e-01
4.91462380e-01 -4.78234068e-02 -9.92594242e-01 7.58064151e-01
-1.91915441e+00 -2.14926139e-01 2.97179759e-01 2.21574926e+00
5.85736275e-01 3.18525255e-01 -6.26698434e-02 -4.64762866e-01
1.25671160e+00 8.72728825e-02 -8.92329395e-01 1.19000368e-01
1.32647634e-01 -7.41409361e-02 5.52789271e-01 2.52360970e-01
-1.11602521e+00 1.07847083e+00 4.91537142e+00 1.05612636e+00
-1.26984227e+00 4.35506478e-02 5.31917930e-01 3.33337545e-01
-4.00052518e-01 1.79411858e-01 -5.67607820e-01 6.72409952e-01
3.00014675e-01 -3.66064847e-01 3.80408347e-01 1.05983567e+00
-1.76925167e-01 2.71372020e-01 -1.01747847e+00 9.61228311e-01
-9.80265588e-02 -8.91038477e-01 1.55704513e-01 8.18587746e-03
7.50112593e-01 -2.73732748e-02 -3.12506892e-02 2.83382177e-01
9.40574110e-02 -5.72841287e-01 4.34400171e-01 3.13726276e-01
7.44253337e-01 -8.15306544e-01 5.79146624e-01 3.94826949e-01
-1.55144048e+00 5.54069616e-02 -6.54732585e-01 4.36959535e-01
-4.99655586e-03 4.94364917e-01 -8.47088456e-01 8.19818974e-01
6.86190009e-01 8.86411130e-01 -5.82957149e-01 1.06940389e+00
-2.42942914e-01 3.34499717e-01 -3.74788910e-01 3.00341815e-01
1.18263252e-01 -5.86493433e-01 7.32679307e-01 7.15052664e-01
3.68890136e-01 1.29871652e-01 6.58987820e-01 8.02745938e-01
-9.37804505e-02 2.67500937e-01 -4.16476846e-01 1.26339927e-01
7.81253159e-01 9.88641500e-01 -9.12490785e-01 -1.75756380e-01
-4.01183814e-01 1.11548138e+00 6.17548406e-01 4.42237824e-01
-7.62380421e-01 -4.04344082e-01 6.95799232e-01 8.49074498e-02
2.78425932e-01 -1.44234046e-01 -3.13691944e-01 -1.18874717e+00
1.35106280e-01 -5.89613020e-01 5.63482046e-01 -4.57770824e-01
-1.56109464e+00 4.64178294e-01 1.63120434e-01 -1.54420817e+00
-6.17208034e-02 -3.07164520e-01 -6.64482832e-01 9.92480099e-01
-1.50154269e+00 -1.10355937e+00 -6.61546707e-01 8.25424671e-01
5.52201927e-01 -2.78459549e-01 4.12237018e-01 3.13133717e-01
-4.01108116e-01 6.63510799e-01 3.21697742e-02 2.53284156e-01
7.01921821e-01 -9.84882534e-01 3.33635509e-01 7.77315974e-01
1.03740588e-01 2.79408604e-01 3.31654817e-01 -7.17726290e-01
-8.20526183e-01 -1.33411145e+00 2.47900411e-01 -9.56945643e-02
3.14565688e-01 -1.75191104e-01 -1.39614093e+00 5.46356320e-01
-3.16150635e-01 2.89424419e-01 3.69453460e-01 -1.34512007e-01
-3.06823194e-01 -2.45689258e-01 -1.35451150e+00 4.88790095e-01
1.33089805e+00 -4.22734231e-01 -6.14943147e-01 3.56260628e-01
9.24437284e-01 -5.79522908e-01 -7.30512083e-01 5.33255100e-01
1.27313405e-01 -6.74233317e-01 7.86396086e-01 -4.69383001e-01
1.62978992e-01 -4.98451829e-01 -1.72215834e-01 -1.52588606e+00
-3.79494399e-01 -1.29403165e-02 4.10940051e-01 1.46241260e+00
2.67760634e-01 -1.02547169e+00 1.01709104e+00 5.18268168e-01
-2.10856944e-01 -5.88461101e-01 -1.05259132e+00 -8.37365806e-01
1.13156624e-01 -8.06208402e-02 8.69919002e-01 1.02736270e+00
-3.74416113e-01 4.29112971e-01 2.80071944e-01 5.23920178e-01
5.07274687e-01 5.39161980e-01 8.02508593e-01 -1.36871696e+00
-2.14685097e-01 -3.26755822e-01 -7.02454746e-01 -1.23629069e+00
4.32982177e-01 -1.09853828e+00 2.70180166e-01 -1.49700880e+00
1.45084485e-01 -8.54914427e-01 -5.38279355e-01 3.64710569e-01
-3.93080890e-01 -2.38228470e-01 1.21858291e-01 4.36983526e-01
-6.03359878e-01 7.89056718e-01 1.41823280e+00 -2.94301957e-01
-3.96294683e-01 -4.18129973e-02 -5.72712004e-01 5.95199883e-01
7.73364425e-01 -6.77027643e-01 -5.30002356e-01 -3.85592997e-01
-3.18769634e-01 -9.59927440e-02 2.19880760e-01 -1.12076426e+00
1.71960622e-01 -5.19530356e-01 1.84377119e-01 -2.93155849e-01
2.11494206e-03 -1.09658337e+00 -1.12601712e-01 8.86126906e-02
-1.06690191e-01 -4.15703833e-01 1.40866771e-01 8.88722539e-01
-4.40975696e-01 -2.30158448e-01 9.90757227e-01 -7.15030283e-02
-1.13155639e+00 6.65795445e-01 3.20630640e-01 4.30786461e-01
1.12489760e+00 -3.52880836e-01 2.29467954e-02 -1.82548240e-01
-6.71971917e-01 6.29050910e-01 8.30053210e-01 5.42265415e-01
6.72186732e-01 -1.39096129e+00 -5.42014599e-01 4.52586979e-01
4.69223142e-01 6.48535967e-01 3.48992854e-01 4.57587481e-01
-3.41807246e-01 -4.37065884e-02 -3.86558831e-01 -1.05967677e+00
-9.77507591e-01 5.52807927e-01 4.16753560e-01 -2.73757845e-01
-4.18842256e-01 8.19288015e-01 7.45640814e-01 -8.30798805e-01
-1.58976287e-01 3.25377695e-02 -1.93572670e-01 -1.47884578e-01
-1.37454912e-01 1.35909095e-01 -4.92206737e-02 -6.72258079e-01
-5.89661539e-01 1.01534736e+00 -1.41909897e-01 2.10089460e-02
9.90176618e-01 -3.79168153e-01 4.17145677e-02 1.50655940e-01
1.40944850e+00 -1.17183216e-01 -1.54816544e+00 -5.93583941e-01
4.20963876e-02 -7.00554013e-01 -2.33654127e-01 -4.28611964e-01
-1.43406343e+00 7.95284331e-01 8.06501865e-01 -4.72693220e-02
1.26590848e+00 2.50297159e-01 8.27641547e-01 6.16817288e-02
4.84168351e-01 -1.16209972e+00 1.82699904e-01 4.16796535e-01
5.88284612e-01 -1.24355686e+00 -2.07050130e-01 -8.11746478e-01
-7.34260261e-01 8.33462715e-01 1.02904117e+00 -2.45446533e-01
5.82298875e-01 -4.90350872e-01 1.17726199e-01 -2.67364562e-01
-3.65684591e-02 -2.48050898e-01 5.14239490e-01 9.04045343e-01
-2.18995497e-01 1.30087972e-01 -9.23479944e-02 5.71229994e-01
-4.28221142e-03 -2.28976756e-01 -4.82338555e-02 7.80703247e-01
-4.88473922e-01 -1.20502806e+00 -1.50574327e-01 4.15910244e-01
1.75873473e-01 2.54298210e-01 -2.96660066e-01 7.06474721e-01
1.52498350e-01 8.48245800e-01 8.40569213e-02 -4.18132067e-01
5.86717308e-01 4.10912879e-04 3.27739984e-01 -8.78611505e-01
-1.06545419e-01 9.86328274e-02 -4.11457747e-01 -4.43710268e-01
-2.63106674e-01 -6.06302798e-01 -1.66266608e+00 2.34685957e-01
-3.09167206e-01 2.73421198e-01 3.61654580e-01 1.07628930e+00
5.29373944e-01 4.50530469e-01 7.86046088e-01 -6.50077045e-01
-3.65387887e-01 -6.53579652e-01 -7.01322734e-01 6.83736563e-01
3.06229126e-02 -9.76037920e-01 -3.08247387e-01 -1.92120552e-01] | [9.662360191345215, 1.4967410564422607] |
1cd6ad0e-156d-4c6c-8431-ebf42d96d427 | consensus-based-phase-connectivity | 2301.03938 | null | https://arxiv.org/abs/2301.03938v1 | https://arxiv.org/pdf/2301.03938v1.pdf | Consensus based phase connectivity identification for distribution network with limited observability | The mitigation of distribution network (DN) unbalance and the use of single-phase flexibility for congestion mitigation requires accurate phase connection information, which is often not available. For a large DN, the naive phase identification proposed in the majority of the prior works using a single voltage reference does not scale well for a multi-feeder DN. We present a consensus algorithm-based phase identification mechanism which uses multiple three-phase reference points to improve the prediction of phases. Due to the absence of real measurements for a real-suburban German DN, the algorithms are developed and evaluated over synthetic data using a digital twin. To utilize strongly correlated measurements, the DN is clustered into zones. We observe those reference measurements located in the same zone as the single-phase consumer leads to accurate prediction of DN phases. Four consensus algorithms are developed and compared. Using numerical results, we recommend the most robust phase identification mechanism. In our evaluation, measurement error, and the impact of the neutral conductor are also assessed. We assume limited DN observability and apply our findings to a German DN without smart meters, but only less than 8% of nodes have measurement boxes along with single-phase consumers with a home energy management system. Voltage time series for 1 month (hourly sampled) is utilized. The numerical results indicate that for 1% accuracy class measurement, the phase connectivity of 308 out of 313 single-phase consumers in a German DN can be identified. Further, we also propose metrics quantifying the goodness of the phase identification. The phase identification framework based on consensus algorithms for DN zones is scalable for large DN and robust towards measurement errors as the estimation is not dependent on a single measurement point. | ['Dirk Van Hertem', 'Arpan Koirala', 'Rickard Lundholm', 'David Brummund', 'Md Umar Hashmi'] | 2023-01-10 | null | null | null | null | ['energy-management'] | ['time-series'] | [-2.03353196e-01 -2.95889564e-02 -6.26530945e-02 2.08333060e-02
-6.40511930e-01 -9.69391763e-01 2.59428054e-01 6.21739030e-01
2.06203878e-01 1.05949402e+00 -1.14159800e-01 -3.06451499e-01
-7.15553403e-01 -9.94192123e-01 -9.72639546e-02 -9.20425534e-01
-3.78272563e-01 6.11425519e-01 6.93173409e-02 -3.61429542e-01
4.89834920e-02 7.06429124e-01 -8.21668029e-01 -1.64684489e-01
1.05361640e+00 9.86517668e-01 -2.45093912e-01 1.82953000e-01
5.21020353e-01 3.64425004e-01 -1.25825715e+00 2.96395659e-01
4.64892149e-01 -3.01399291e-01 -3.96988869e-01 -1.49473146e-01
-4.43065196e-01 -3.91724825e-01 -1.65995255e-01 1.10271347e+00
9.01349783e-01 -2.13925660e-01 6.08016253e-01 -1.81815517e+00
3.25179607e-01 9.64363277e-01 -6.26481533e-01 3.20425540e-01
5.63999593e-01 1.86382309e-01 9.30659831e-01 -2.26799011e-01
3.31102312e-01 5.12923896e-01 8.18849504e-01 -3.89798731e-01
-1.60775101e+00 -6.50578976e-01 -2.63248142e-02 4.68614042e-01
-1.78342211e+00 1.44646257e-01 1.02290452e+00 -2.64571905e-01
1.06788886e+00 2.03429580e-01 1.21265650e+00 2.82723814e-01
8.37506279e-02 1.59234032e-01 1.04232502e+00 -3.03916633e-01
7.37838030e-01 -2.07424402e-01 2.76404596e-03 -3.74105215e-01
6.76245272e-01 -1.59218326e-01 -5.25342114e-02 -3.25443864e-01
2.39657059e-01 -5.46290219e-01 -6.88878119e-01 -6.40214860e-01
-8.29759598e-01 7.86922157e-01 4.13932085e-01 5.42854905e-01
-4.75850195e-01 -3.59912634e-01 4.09679234e-01 3.77840132e-01
3.85122955e-01 2.75854439e-01 -7.10085511e-01 -1.94915608e-01
-1.29755843e+00 6.53901994e-02 1.16050100e+00 9.55763102e-01
6.61926568e-01 7.20009267e-01 4.15609181e-01 1.60733327e-01
3.18032861e-01 8.93712759e-01 2.24563271e-01 -8.44340205e-01
2.98669487e-01 7.13227808e-01 2.68932939e-01 -6.74183309e-01
-8.82432342e-01 -7.85836041e-01 -8.74983847e-01 4.34253216e-01
5.48978269e-01 -6.99625611e-01 -1.63155705e-01 1.44141364e+00
1.39597267e-01 -1.22837767e-01 1.74407884e-01 5.30818343e-01
-1.21537987e-02 5.74800193e-01 -5.01035810e-01 -7.70286322e-01
1.06329608e+00 -2.13235259e-01 -8.98707032e-01 2.90815115e-01
8.04620683e-01 -8.25161576e-01 -4.32662107e-02 7.79998064e-01
-9.52028394e-01 -7.03068897e-02 -1.48631680e+00 9.49489534e-01
-2.95466661e-01 -2.13268816e-01 -3.03151440e-02 8.22135389e-01
-1.10455585e+00 6.41244054e-01 -6.77044868e-01 -4.11957026e-01
-4.02175896e-02 3.68767530e-01 -1.33749291e-01 1.00294530e-01
-1.43644917e+00 1.41688037e+00 2.98472166e-01 4.64426756e-01
-2.73309141e-01 -1.10096467e+00 -6.44381106e-01 1.30045637e-01
2.46906863e-03 -4.08689559e-01 1.04114246e+00 -8.70077431e-01
-1.18627489e+00 -2.04095855e-01 2.75757760e-01 -8.66915107e-01
6.14086151e-01 2.08940014e-01 -1.01890039e+00 5.18743813e-01
4.78224866e-02 -2.82085687e-01 2.08655417e-01 -1.07383275e+00
-6.34646297e-01 4.23569679e-02 -1.68775678e-01 6.14363402e-02
2.20353276e-01 -3.90087932e-01 7.75386214e-01 -3.10215950e-01
3.39251548e-01 -4.94425356e-01 -4.01975930e-01 -3.55717629e-01
-4.24411625e-01 1.69529498e-01 1.04363632e+00 -7.46314526e-01
1.41050804e+00 -1.50780368e+00 -3.10444921e-01 1.30371153e+00
4.89538051e-02 1.43823177e-01 4.76847887e-01 1.21041036e+00
-5.23648918e-01 -3.31696421e-02 -2.57636070e-01 3.52918655e-01
3.95818025e-01 2.98362464e-01 6.57140336e-05 9.39846635e-01
-1.34651706e-01 2.85984486e-01 -7.77049959e-01 2.17533141e-01
7.63787627e-01 3.47444952e-01 -4.53158259e-01 -2.45894372e-01
2.95000851e-01 2.37589031e-01 -4.02853228e-02 3.52454364e-01
1.03371108e+00 -1.19761564e-01 8.02183986e-01 -7.34482884e-01
-2.79196054e-01 3.75222355e-01 -2.05842519e+00 1.25220263e+00
-3.93154204e-01 7.12891936e-01 2.89547861e-01 -1.18769240e+00
8.39356780e-01 5.93451142e-01 1.07064509e+00 -9.47728038e-01
-1.50549740e-01 3.75109822e-01 5.28224468e-01 1.57355540e-03
-9.64593068e-02 2.44871855e-01 4.70012277e-02 7.67255068e-01
9.86451954e-02 -3.80954921e-01 7.67042160e-01 1.78294837e-01
1.06799412e+00 -4.33355898e-01 6.42337143e-01 -9.88340199e-01
5.36873639e-01 -1.56668663e-01 1.13635623e+00 -1.54526085e-02
-2.48452634e-01 3.54878098e-01 8.28204155e-01 1.03401817e-01
-9.11261559e-01 -1.22010171e+00 -4.66450453e-01 -3.17513466e-01
1.30784139e-01 -5.53525329e-01 -4.42905098e-01 -2.58404940e-01
1.58246651e-01 1.16731465e+00 -1.44119918e-01 -2.92390846e-02
-5.54900110e-01 -1.12890410e+00 2.60537982e-01 4.90432501e-01
3.20719451e-01 -1.70168981e-01 -3.12419206e-01 6.71854317e-01
-1.44435778e-01 -8.69271338e-01 2.04177588e-01 5.04342496e-01
-5.15867829e-01 -1.57874739e+00 -4.40940499e-01 -3.61123621e-01
7.93774366e-01 -3.07546467e-01 1.04640746e+00 -1.11084238e-01
8.93628597e-02 2.74625987e-01 -1.31256998e-01 1.15411721e-01
-3.67897987e-01 -1.72947690e-01 4.05921459e-01 -2.82006949e-01
3.50329280e-01 -1.18323672e+00 -8.02020371e-01 5.75572789e-01
-4.84783471e-01 -3.24016184e-01 1.83252394e-01 6.19552493e-01
3.41219418e-02 6.80103302e-01 9.66485858e-01 -4.50989306e-01
3.03679287e-01 -6.61224484e-01 -1.20270383e+00 -2.43279692e-02
-1.11579943e+00 -2.17357576e-01 7.97487617e-01 -7.65615925e-02
-5.77650011e-01 -2.82824367e-01 7.53212254e-03 1.15503624e-01
-9.16989446e-02 5.04466295e-01 -5.51760316e-01 -1.97411522e-01
2.51509845e-01 -3.23919594e-01 -2.36521155e-01 -4.99198675e-01
2.54199305e-03 4.43493456e-01 9.84590006e-05 -4.10022944e-01
1.28947020e+00 3.66140246e-01 2.68174976e-01 -7.75396883e-01
1.70302521e-02 -3.48570138e-01 -7.47832358e-01 -2.55842984e-01
1.76557541e-01 -1.05431950e+00 -1.07917225e+00 4.63853717e-01
-8.72184515e-01 -2.95705974e-01 -4.94327247e-01 7.07465470e-01
-3.06661010e-01 5.02297997e-01 -5.67098558e-01 -6.30402267e-01
-3.70761305e-01 -1.23973715e+00 2.42520303e-01 2.96914965e-01
-5.95254421e-01 -1.35603011e+00 6.87188432e-02 -1.41806796e-01
4.48843092e-01 4.49498951e-01 8.49731088e-01 -9.29441869e-01
-4.80085224e-01 3.16116912e-03 1.24600478e-01 5.64485490e-01
3.14716995e-01 3.83689046e-01 -6.67205334e-01 -8.25092137e-01
-7.33316466e-02 5.30937016e-01 -4.42239940e-01 4.18696731e-01
2.77917475e-01 -2.99952567e-01 -3.63641769e-01 3.37858349e-01
1.96997237e+00 4.28308487e-01 4.75544155e-01 4.82211053e-01
1.62519053e-01 2.86738396e-01 4.57104295e-01 7.10869610e-01
5.47918379e-01 3.82257193e-01 4.18185860e-01 -2.66125590e-01
4.52116877e-01 2.27686331e-01 2.21597001e-01 1.06285977e+00
1.82892010e-01 -4.29634064e-01 -8.92610729e-01 6.68321311e-01
-1.51972425e+00 -8.63911927e-01 -4.45792913e-01 2.14584422e+00
3.45608324e-01 5.24275422e-01 8.49120989e-02 1.03616500e+00
7.42648721e-01 -1.37613997e-01 -1.37056947e-01 -3.01000476e-01
-4.96296018e-01 3.55468094e-01 1.05720460e+00 5.79088092e-01
-5.42473853e-01 -4.47108358e-01 5.78576088e+00 5.98417521e-01
-9.79545891e-01 -1.75501302e-01 1.40322059e-01 1.89142868e-01
-2.82066971e-01 3.78462464e-01 -5.42778015e-01 6.96288347e-01
1.32052529e+00 -9.48847711e-01 8.20292309e-02 4.92688656e-01
9.31757033e-01 -4.98306006e-01 -1.03483319e+00 6.74285173e-01
-3.60010147e-01 -9.51002121e-01 -3.98549408e-01 1.18634447e-01
9.99453604e-01 8.40740502e-02 -6.30583763e-01 -8.98080692e-02
3.63408059e-01 -4.87236857e-01 6.48871422e-01 2.24753127e-01
1.53333738e-01 -1.15049970e+00 1.05016565e+00 2.32652754e-01
-1.48418820e+00 -2.80577004e-01 1.59838468e-01 -1.36804372e-01
8.09772909e-01 1.02908659e+00 -7.88821578e-01 1.34002769e+00
5.74109256e-01 7.07276165e-01 -2.59202749e-01 1.52371132e+00
-5.69172740e-01 9.51712549e-01 -1.01162243e+00 4.05198753e-01
-1.43845066e-01 -5.62271416e-01 5.10881782e-01 6.78588986e-01
4.66790468e-01 -4.12622653e-02 2.22926259e-01 5.50233424e-01
4.51402366e-01 -9.02688503e-02 -1.06246516e-01 6.94253385e-01
9.97467518e-01 1.24194789e+00 -5.84033430e-01 -1.38341308e-01
-4.87125397e-01 1.11626633e-01 -5.91496110e-01 7.94133902e-01
-6.49165273e-01 -4.45841283e-01 8.33535492e-01 2.74014175e-01
1.48938313e-01 -7.90590271e-02 -5.35235226e-01 -8.34957480e-01
1.74782142e-01 -8.07009220e-01 3.58111948e-01 -7.64156997e-01
-1.36410105e+00 3.04686576e-01 2.73910165e-01 -1.66574013e+00
-1.13860738e+00 -5.41601144e-02 -9.27158773e-01 1.16685271e+00
-1.50868559e+00 -7.75945902e-01 3.13786007e-02 2.95303106e-01
5.88860773e-02 1.51418492e-01 8.21248293e-01 7.84981787e-01
-7.19101429e-01 2.17936516e-01 5.97345293e-01 2.06644669e-01
2.38042101e-01 -1.46064115e+00 -3.49456668e-02 1.26462233e+00
-4.86394107e-01 1.12755820e-01 1.03086078e+00 -5.25231361e-01
-1.01285970e+00 -5.12193263e-01 7.47282624e-01 3.85654569e-01
1.03849232e+00 -3.18859071e-01 -6.30910039e-01 7.46762991e-01
8.38449299e-01 -1.48467645e-01 5.62763333e-01 -4.54354048e-01
3.97804052e-01 -6.79319680e-01 -1.50530803e+00 3.29501927e-01
1.12439699e-01 -3.90504569e-01 -5.17225742e-01 1.63429871e-01
-1.86607197e-01 -1.60006225e-01 -1.60009563e+00 5.26718080e-01
2.77066585e-02 -7.24234164e-01 7.72039711e-01 3.06173354e-01
-7.04465210e-01 -9.69037771e-01 -1.64697692e-01 -1.96768212e+00
5.48650837e-03 -9.77414131e-01 -7.22817779e-02 1.40239513e+00
3.69702667e-01 -1.32170200e+00 5.62686026e-01 2.93162912e-01
1.35066867e-01 -2.81987637e-01 -1.33378220e+00 -6.92864299e-01
1.78641960e-01 -2.64447570e-01 1.02650297e+00 1.13369203e+00
5.71031392e-01 1.36337042e-01 3.43759567e-01 7.93187797e-01
1.11555195e+00 1.33388311e-01 2.66218811e-01 -1.37259901e+00
-1.47518958e-03 -3.92599791e-01 -5.80276191e-01 -5.17839134e-01
-1.12357609e-01 -4.01885480e-01 -5.01663625e-01 -1.81420457e+00
-6.10521078e-01 -4.60289717e-01 -1.96751654e-01 1.75646231e-01
5.10329485e-01 -2.26527173e-02 6.72009066e-02 -2.67828405e-01
1.73099741e-01 3.03720713e-01 3.44091386e-01 -2.54171550e-01
3.94101858e-01 3.84999603e-01 6.33526221e-02 7.60788739e-01
1.04096234e+00 -2.29700521e-01 -5.82573414e-01 1.99517369e-01
1.67684555e-01 4.01270181e-01 7.42093800e-03 -1.37889612e+00
5.06869435e-01 3.49884093e-01 4.24777061e-01 -1.11680937e+00
8.57322142e-02 -1.44727278e+00 8.00601900e-01 9.40650165e-01
4.59354162e-01 6.45585537e-01 1.70771942e-01 1.19972721e-01
-2.16918811e-01 -1.14825450e-01 7.15803146e-01 3.89139801e-01
-5.98374009e-01 -1.90382347e-01 -8.78929555e-01 -2.77080797e-02
1.32123661e+00 -3.29612732e-01 -5.75044572e-01 -5.84682345e-01
-7.90981174e-01 8.11312199e-01 5.34811258e-01 -1.07402340e-01
-1.35774082e-02 -1.14838910e+00 -6.67812586e-01 2.13245019e-01
-3.61422926e-01 -1.43574819e-01 2.39285126e-01 9.73505080e-01
-7.81049848e-01 4.64189321e-01 -1.33466691e-01 -8.20581198e-01
-8.01081777e-01 1.17384009e-01 7.75456786e-01 -2.81819522e-01
-2.40807682e-01 -2.77075589e-01 -5.86782277e-01 -1.86426058e-01
-2.12748170e-01 -5.78628659e-01 -1.66716829e-01 7.40253687e-01
2.92700380e-01 9.34011102e-01 7.92545557e-01 -5.23205519e-01
-4.51473117e-01 5.22144437e-01 4.68773097e-01 1.90902259e-02
1.30369413e+00 -6.12415373e-01 -2.21770257e-02 3.14699113e-01
1.08123994e+00 4.48531248e-02 -9.64076877e-01 2.79012084e-01
1.13314819e-02 8.57015327e-02 -1.37271494e-01 -8.34897876e-01
-1.51205707e+00 4.01597172e-01 5.93382478e-01 6.64621830e-01
1.20894516e+00 -9.97920752e-01 3.79558355e-01 -6.35329708e-02
7.06013620e-01 -1.22777390e+00 -8.33565176e-01 1.44924223e-01
2.33762592e-01 -5.57919502e-01 2.35445082e-01 -3.95354182e-01
1.06950346e-02 1.22688639e+00 2.50729233e-01 -1.44367427e-01
1.06064665e+00 8.64516914e-01 1.73977941e-01 2.47460380e-01
-7.73096979e-01 4.89793383e-02 -3.08325350e-01 8.38237941e-01
-4.01761243e-03 8.12013373e-02 -6.39072299e-01 3.38003129e-01
-3.40539068e-01 -1.29173905e-01 1.26039040e+00 1.06976259e+00
-3.34076546e-02 -1.16068673e+00 -4.27390814e-01 4.48472589e-01
-3.90768319e-01 2.02055305e-01 6.00561678e-01 1.18787432e+00
-1.17475502e-02 1.22830737e+00 4.01084334e-01 1.41329959e-01
7.73985326e-01 -1.69089347e-01 1.23208843e-01 3.05954125e-02
-7.85770476e-01 5.80464825e-02 3.41965139e-01 -5.46444058e-01
-3.64424735e-01 -9.60389256e-01 -1.45632887e+00 -9.26303983e-01
-5.47452629e-01 7.81528890e-01 5.80125809e-01 9.09143031e-01
-1.79784298e-01 5.96021891e-01 1.04275775e+00 -5.59872985e-01
-7.93757975e-01 -7.95106709e-01 -1.20330560e+00 -5.35854287e-02
1.00825563e-01 -5.45053959e-01 -9.71951783e-01 -5.77999890e-01] | [5.882043838500977, 2.5610387325286865] |
b4a7eef0-eac0-4aa9-8244-3f3b271a557f | a-low-shot-object-counting-network-with | 2211.08217 | null | https://arxiv.org/abs/2211.08217v1 | https://arxiv.org/pdf/2211.08217v1.pdf | A Low-Shot Object Counting Network With Iterative Prototype Adaptation | We consider low-shot counting of arbitrary semantic categories in the image using only few annotated exemplars (few-shot) or no exemplars (no-shot). The standard few-shot pipeline follows extraction of appearance queries from exemplars and matching them with image features to infer the object counts. Existing methods extract queries by feature pooling, but neglect the shape information (e.g., size and aspect), which leads to a reduced object localization accuracy and count estimates. We propose a Low-shot Object Counting network with iterative prototype Adaptation (LOCA). Our main contribution is the new object prototype extraction module, which iteratively fuses the exemplar shape and appearance queries with image features. The module is easily adapted to zero-shot scenario, enabling LOCA to cover the entire spectrum of low-shot counting problems. LOCA outperforms all recent state-of-the-art methods on FSC147 benchmark by 20-30% in RMSE on one-shot and few-shot and achieves state-of-the-art on zero-shot scenarios, while demonstrating better generalization capabilities. | ['Matej Kristan', 'Vitjan Zavrtanik', 'Alan Lukezic', 'Nikola Djukic'] | 2022-11-15 | null | null | null | null | ['object-counting'] | ['computer-vision'] | [ 1.9036461e-01 -3.0108559e-01 -9.0323187e-02 -3.7161541e-01
-8.5602784e-01 -5.6296355e-01 7.4192548e-01 4.3017662e-01
-7.6277548e-01 3.6406386e-01 -2.6544085e-01 3.9982179e-01
9.1452308e-02 -7.7739364e-01 -9.0430099e-01 -4.5575652e-01
1.9891690e-01 6.3699394e-01 9.4774395e-01 1.7830627e-01
4.5038944e-01 4.2379844e-01 -2.0881319e+00 2.8806174e-01
3.9715439e-01 1.2220253e+00 3.7132311e-01 8.3824927e-01
-4.8931012e-01 8.6410934e-01 -6.4845401e-01 -4.3308362e-01
1.3852464e-01 -1.4574306e-01 -4.7623745e-01 -7.7595703e-02
1.0013796e+00 -5.0365770e-01 -1.2882751e-01 1.0601048e+00
5.4203469e-01 1.3632235e-01 9.4339657e-01 -1.1960304e+00
-7.7084589e-01 3.1642791e-02 -9.4848490e-01 5.1784027e-01
1.5120491e-01 4.3654054e-01 9.6061403e-01 -1.2229688e+00
4.9243638e-01 1.2491401e+00 6.6789138e-01 4.1528314e-01
-1.0665063e+00 -6.7560583e-01 1.4184949e-02 1.7686309e-01
-1.5973943e+00 -3.2182023e-01 1.8790188e-01 -3.2112440e-01
1.2109669e+00 1.3945596e-01 7.2150737e-01 7.0248646e-01
-1.9508544e-01 8.4929705e-01 9.8542881e-01 -4.1215372e-01
6.0448992e-01 9.6810848e-02 4.6048784e-01 9.0854943e-01
6.8207103e-01 -2.7621606e-01 -5.4664171e-01 -1.4847833e-01
7.2574103e-01 3.2053632e-01 4.0785277e-01 -4.5743719e-01
-1.1031839e+00 7.7792525e-01 4.4651112e-01 3.3545531e-02
-2.0325954e-01 5.1188403e-01 4.8872417e-01 -1.2981483e-01
2.4156117e-01 2.6355857e-01 -2.2389123e-01 3.4932118e-02
-1.0693264e+00 2.8070790e-01 7.9060680e-01 1.6063907e+00
8.7760854e-01 -1.2427431e-01 -9.3289953e-01 9.9434000e-01
-2.5253284e-01 8.0419034e-01 1.6438515e-01 -9.7258615e-01
2.0839533e-01 6.8152660e-01 3.6795244e-01 -6.7078239e-01
-3.4499359e-01 -2.4814744e-01 -4.0672025e-01 -1.2817070e-01
4.7758308e-01 3.0757302e-01 -1.2876499e+00 1.3268183e+00
4.6228439e-01 3.4201676e-01 -3.4617344e-01 9.1666627e-01
1.1711867e+00 3.9625239e-01 3.7848660e-01 -2.6623091e-02
1.9634560e+00 -1.0231316e+00 -2.0002389e-01 -4.7131997e-01
3.7365630e-01 -5.8963597e-01 1.2401527e+00 1.0648620e-01
-1.0155619e+00 -5.8313543e-01 -1.3394254e+00 -3.3752441e-01
-7.3975289e-01 4.0334794e-01 7.3537195e-01 6.4726675e-01
-5.7826096e-01 5.2795851e-01 -6.4255983e-01 -6.6644591e-01
9.2591369e-01 3.7785026e-01 -1.5879743e-02 -2.9853079e-01
-7.2923827e-01 7.0396173e-01 6.3918662e-01 -6.3454533e-01
-8.8828498e-01 -9.5275438e-01 -8.6271793e-01 3.0477479e-01
7.4523145e-01 -6.0471106e-01 1.3388511e+00 -4.5481420e-01
-1.1176846e+00 1.0497199e+00 -5.3167652e-02 -3.1501484e-01
4.2003033e-01 -4.4246264e-02 -8.5694946e-02 2.4512975e-01
4.2929274e-01 1.0688282e+00 8.7315685e-01 -9.4800043e-01
-1.0378666e+00 -3.3687434e-01 1.4813742e-01 -6.2042411e-02
-3.1173411e-01 5.5501398e-02 -7.9810870e-01 -2.4730414e-01
-1.4902383e-01 -6.0597217e-01 2.2680162e-01 4.3088943e-01
1.0807724e-02 -5.0104815e-01 7.3170334e-01 3.6572378e-02
9.9064815e-01 -2.0199354e+00 -4.8112032e-01 -4.2051059e-01
1.6304137e-01 2.7387947e-01 -1.9924740e-01 1.6731097e-01
4.0706697e-01 -3.3367604e-01 -6.0480729e-02 -5.6508952e-01
3.9426692e-02 2.2182959e-01 4.8393807e-03 6.5302598e-01
6.2264204e-01 1.1594695e+00 -1.2182480e+00 -1.1665694e+00
6.5387732e-01 2.7818891e-01 -4.8790252e-01 1.1257551e-02
-2.9273230e-01 -1.0557585e-02 -2.7853578e-01 1.0647761e+00
9.0522802e-01 -5.6654602e-01 -3.6523503e-01 -4.0601900e-01
-1.8029644e-01 -5.5616903e-01 -1.2503140e+00 1.6371105e+00
-5.2036732e-01 3.3080354e-01 -4.6818626e-01 -7.0019752e-01
7.4967778e-01 -1.8978961e-01 3.6567718e-01 -6.4347851e-01
4.5190096e-01 2.3693471e-01 -1.8670970e-01 -4.2044783e-01
4.6051162e-01 -2.8259398e-02 -3.2914191e-01 1.6243228e-01
7.9657972e-01 -2.0945351e-01 7.9702640e-01 3.7039858e-01
1.0908340e+00 -1.0986297e-02 8.8018090e-01 -2.6898107e-01
3.0813441e-01 2.0089369e-02 3.6977702e-01 1.2458622e+00
-5.2654713e-01 1.0099440e+00 2.1224770e-01 -4.7183728e-01
-1.2874792e+00 -1.0740634e+00 -2.8229371e-01 1.4749181e+00
3.9627618e-01 -2.0048444e-01 -7.0072412e-01 -5.2113795e-01
1.9286305e-01 8.6719429e-01 -8.2638603e-01 8.5569441e-02
-3.2795101e-01 -6.9729382e-01 3.9681774e-01 8.5610497e-01
4.2075720e-01 -1.0463610e+00 -1.1479989e+00 1.6418898e-01
1.5954772e-01 -1.4490106e+00 -5.8874905e-01 1.8704720e-01
-4.5784190e-01 -1.1890044e+00 -8.9354432e-01 -6.0832322e-01
5.5006534e-01 4.6141514e-01 1.2206290e+00 9.1231942e-02
-1.1970439e+00 5.6142819e-01 -3.7691763e-01 -8.0471575e-01
1.5397631e-01 -1.2286585e-01 -1.8731387e-01 -1.1237773e-01
8.8291091e-01 -1.8133844e-01 -9.8035282e-01 3.0049920e-01
-6.7202222e-01 -2.4994007e-01 7.7485400e-01 7.2386140e-01
8.4788263e-01 -4.7588363e-01 6.1011636e-01 -9.0865380e-01
1.4698379e-01 -4.4499740e-01 -7.2472125e-01 4.5127264e-01
-2.8508890e-01 -9.0027988e-02 6.6082108e-01 -7.5804013e-01
-1.0386302e+00 4.4245654e-01 4.5538160e-01 -6.0169792e-01
-2.4727203e-01 -5.0896877e-01 1.5512191e-01 -8.1803590e-02
9.1729963e-01 2.0455746e-01 -4.1343415e-01 -2.5655147e-01
6.4441198e-01 6.1332303e-01 7.3396617e-01 -4.4467878e-01
4.3054330e-01 9.6442890e-01 1.0107698e-01 -9.9241722e-01
-1.3750272e+00 -1.4688344e+00 -1.0032014e+00 -2.5572008e-01
9.1767675e-01 -9.6550059e-01 -6.5297914e-01 3.5281235e-01
-1.1937069e+00 1.4876425e-01 -5.9267324e-01 2.9953092e-01
-8.6955547e-01 1.8907303e-01 -5.5796844e-01 -1.0775741e+00
-6.4385915e-01 -7.5501370e-01 1.6963339e+00 5.0927323e-01
7.4210204e-02 -3.2031468e-01 -2.7041575e-01 -3.5260282e-03
2.6333839e-01 5.7753846e-02 7.5356781e-01 -7.3489028e-01
-6.9953531e-01 -6.1986601e-01 -8.1696671e-01 9.7159939e-03
-3.6324975e-01 -1.5035942e-01 -1.2076137e+00 -2.7390596e-01
-3.6363646e-01 -6.9076097e-01 1.1137675e+00 5.7903486e-01
1.3311094e+00 1.1663760e-01 -3.3896238e-01 4.9823686e-01
1.8865489e+00 -3.3823568e-01 4.8848772e-01 -8.9283049e-02
3.9499947e-01 1.9946297e-01 8.7673622e-01 6.9515771e-01
-2.4526361e-02 4.9582830e-01 4.0553898e-01 1.7540045e-01
-3.8796496e-01 -9.4901249e-02 -2.6833206e-01 4.6761775e-01
-1.7450872e-01 -7.6650120e-02 -6.2498444e-01 6.5801400e-01
-1.8668678e+00 -8.3703685e-01 -4.1758120e-02 2.1280866e+00
4.0239140e-01 2.9842520e-01 3.5421106e-01 -2.6756597e-01
9.9888873e-01 7.4866056e-02 -7.7957994e-01 -1.3999167e-02
-1.6845111e-02 3.3592784e-01 1.0069531e+00 -1.7262356e-01
-1.2488769e+00 1.0093241e+00 5.9687676e+00 1.1058847e+00
-4.4932827e-01 5.8061624e-01 5.4232782e-01 -4.4073939e-01
7.0478445e-01 -3.7330028e-02 -1.2784982e+00 4.2639056e-01
5.7919371e-01 -1.2854351e-01 3.7622407e-01 1.2617127e+00
-3.6697689e-01 -4.5065209e-01 -1.2713389e+00 1.2763070e+00
5.8768749e-01 -1.2131827e+00 -1.6185194e-02 -4.1678759e-01
7.6339668e-01 1.4738601e-01 -2.9662207e-01 8.1332725e-01
1.9515266e-01 -7.7293748e-01 9.0650308e-01 6.5349931e-01
1.3244717e+00 -5.5947870e-01 7.5798905e-01 4.6737444e-01
-1.5708034e+00 -4.4888791e-01 -1.1373976e+00 1.9012829e-02
1.6794722e-01 2.5910565e-01 -4.8228571e-01 1.0312799e-01
7.3183393e-01 4.2828351e-01 -9.2887157e-01 1.3863124e+00
1.5203680e-01 1.7681819e-01 -5.2079654e-01 -4.9768966e-01
1.3968977e-01 2.9092568e-01 1.1961006e-01 1.4920976e+00
2.8668144e-01 3.3243084e-01 3.1690621e-01 1.0996752e+00
-1.9022010e-01 1.8674324e-01 -3.4055555e-01 8.3559215e-02
8.9285994e-01 1.4544994e+00 -1.3293092e+00 -8.5028762e-01
-7.5316083e-01 9.6666151e-01 5.2382892e-01 -8.7741278e-02
-1.0020845e+00 -6.4419007e-01 -4.6451621e-02 2.5426519e-01
7.3011738e-01 2.3438157e-01 -2.8659865e-02 -8.8321459e-01
-1.5855803e-01 -2.2751768e-01 5.7982451e-01 -6.8445194e-01
-1.5898192e+00 5.2966598e-02 4.6309915e-01 -1.0913237e+00
2.8506406e-02 -8.3379728e-01 -6.7095453e-01 3.2155782e-01
-1.2492063e+00 -1.4205559e+00 -8.1727952e-01 2.2957458e-01
1.1172704e+00 -1.0600713e-01 6.7542398e-01 2.3544286e-01
-2.0279506e-01 5.9722716e-01 -8.9065209e-03 8.1704423e-02
5.4461765e-01 -1.2135254e+00 5.1163393e-01 3.7620565e-01
2.6553330e-01 4.3314913e-01 4.7530887e-01 -4.6441156e-01
-1.3070555e+00 -1.3575667e+00 2.0148876e-01 -7.0233649e-01
5.4110277e-01 -6.3899547e-01 -6.7353076e-01 1.6226183e-01
-3.7154451e-01 8.8539976e-01 2.5162971e-01 -2.0661400e-01
-5.9797436e-01 5.7464186e-02 -1.2298727e+00 2.5822139e-01
1.2214861e+00 -2.8688446e-01 -4.3571031e-01 4.8491079e-01
7.0585769e-01 -2.4315707e-02 -5.8316052e-01 3.2051444e-01
6.8304211e-01 -9.4143134e-01 1.1943904e+00 -2.8243044e-01
1.7256367e-01 -1.9529165e-01 -2.5001815e-01 -5.5659825e-01
-5.2957362e-01 1.7872488e-01 -4.8399436e-01 8.8680249e-01
4.1063828e-03 -2.1665429e-01 7.6586062e-01 6.6028535e-02
5.7452037e-03 -7.0957947e-01 -9.6296000e-01 -1.2013655e+00
-5.7434704e-02 -1.8391891e-01 5.7303303e-01 4.1443923e-01
-3.6277965e-01 4.5736158e-01 -2.5981951e-01 -2.0622543e-01
1.0067507e+00 1.2475737e-01 8.7667477e-01 -1.2936045e+00
-3.7212691e-01 -3.3974421e-01 -7.2878343e-01 -8.6849356e-01
-2.1597640e-01 -7.2987270e-01 2.9558000e-01 -1.4368037e+00
1.0087609e+00 2.3305642e-03 -2.7761188e-01 3.6448413e-01
-3.1138235e-01 7.4251819e-01 3.8779444e-01 2.5453806e-01
-1.3902464e+00 3.0686024e-01 9.6523649e-01 -2.4554834e-01
1.0509248e-01 -2.6944396e-01 -1.9077551e-01 7.5940251e-01
5.0075561e-01 -6.4009011e-01 -1.3097416e-01 -5.8195285e-02
-3.8130090e-02 -2.9796916e-01 6.4058900e-01 -1.4682232e+00
3.6735895e-01 -2.3434853e-02 6.0150641e-01 -9.4599134e-01
5.9316176e-01 -3.6425531e-01 -4.4350457e-01 3.9053896e-01
4.1030060e-02 -5.4906958e-01 4.3904610e-02 9.9266452e-01
2.1322848e-01 -6.5293735e-01 1.2017956e+00 -5.6370014e-01
-1.0482821e+00 3.2457918e-01 2.0094728e-01 2.4966782e-01
1.1998544e+00 -5.9045368e-01 -5.1267153e-01 1.4580095e-01
-4.6990705e-01 1.4740904e-01 4.5282066e-01 2.7709958e-01
3.2091591e-01 -1.1723204e+00 -6.3154018e-01 2.8102176e-02
8.6058372e-01 9.8207437e-02 3.1384799e-01 8.8520676e-01
-4.8506585e-01 5.8062440e-01 -1.6532056e-01 -1.0069870e+00
-1.2001897e+00 9.7210199e-01 1.1758803e-01 1.0661737e-01
-7.1666265e-01 9.8314595e-01 3.6473176e-01 -1.3537428e-01
3.5708052e-01 -2.8576517e-01 -7.4288204e-02 8.2784928e-02
7.6876318e-01 5.8498251e-01 9.7903693e-03 -5.3458756e-01
-4.1320035e-01 7.3346132e-01 -1.5864600e-01 4.3607101e-01
1.1836632e+00 8.4504411e-02 4.1509074e-01 8.5519618e-01
1.1633050e+00 -5.0695020e-01 -1.3373504e+00 -2.8196430e-01
-9.6533587e-03 -7.1928376e-01 -1.0366258e-01 -5.9886992e-01
-7.1062595e-01 9.2561281e-01 7.7483761e-01 -2.7340513e-01
6.6184682e-01 5.0919777e-01 4.4567057e-01 5.6981605e-01
7.5782716e-01 -1.4737134e+00 6.5958053e-01 4.3668127e-01
5.7707930e-01 -1.5691786e+00 4.1588774e-01 -4.3491289e-01
-5.1411521e-01 9.2131853e-01 8.1270677e-01 -3.8862967e-01
4.7217107e-01 2.9278591e-01 -6.0217422e-01 -5.5228847e-01
-5.7788581e-01 -6.5194416e-01 3.3266133e-01 4.8548993e-01
1.1117107e-01 -9.6283071e-03 -2.3534591e-01 6.1520016e-01
3.3118939e-01 7.3886369e-03 2.2972906e-01 1.0350294e+00
-1.0484419e+00 -1.8383820e-01 -3.4658858e-01 1.0390580e+00
-1.8860945e-01 -1.0009716e-01 -2.7539796e-01 1.0366498e+00
4.4849506e-01 6.2003505e-01 6.2978190e-01 2.4079843e-01
3.5782662e-01 5.1261481e-02 7.1553481e-01 -1.0297146e+00
7.1343535e-04 -8.7815545e-02 -2.3144205e-01 -6.0266101e-01
-2.9656234e-01 -6.3314623e-01 -1.2184190e+00 4.0943101e-03
-9.9383378e-01 -5.4444718e-01 6.3882774e-01 6.7577285e-01
6.4882807e-02 5.0241870e-01 6.9903679e-02 -1.0021966e+00
-8.5281783e-01 -1.3202170e+00 -7.2848821e-01 6.9722468e-01
-9.3448482e-02 -1.1637630e+00 -3.8162914e-01 -5.0905839e-02] | [9.00898265838623, 0.5478692650794983] |
53394e68-280b-42f5-be92-44079c29d69e | generating-pertinent-and-diversified-comments | 2005.04396 | null | https://arxiv.org/abs/2005.04396v1 | https://arxiv.org/pdf/2005.04396v1.pdf | Generating Pertinent and Diversified Comments with Topic-aware Pointer-Generator Networks | Comment generation, a new and challenging task in Natural Language Generation (NLG), attracts a lot of attention in recent years. However, comments generated by previous work tend to lack pertinence and diversity. In this paper, we propose a novel generation model based on Topic-aware Pointer-Generator Networks (TPGN), which can utilize the topic information hidden in the articles to guide the generation of pertinent and diversified comments. Firstly, we design a keyword-level and topic-level encoder attention mechanism to capture topic information in the articles. Next, we integrate the topic information into pointer-generator networks to guide comment generation. Experiments on a large scale of comment generation dataset show that our model produces the valuable comments and outperforms competitive baseline models significantly. | ['Kang Xu', 'Junheng Huang', 'Weihua Peng', 'Lu Pan', 'Fayuan Li'] | 2020-05-09 | null | null | null | null | ['comment-generation'] | ['natural-language-processing'] | [-5.67040555e-02 3.36707830e-01 -3.81768733e-01 -1.62879050e-01
-8.34430695e-01 -1.81872129e-01 9.51448143e-01 -3.01108453e-02
1.46294639e-01 1.26380551e+00 1.04231381e+00 -2.47983202e-01
4.68371809e-01 -9.81756985e-01 -4.45831209e-01 -3.83134693e-01
4.08557564e-01 2.92734057e-01 2.34805495e-01 -5.55839539e-01
5.69147348e-01 -3.23007643e-01 -1.20143282e+00 3.82472008e-01
1.45624554e+00 6.32296979e-01 4.94512618e-01 4.97026771e-01
-5.43829322e-01 6.92560673e-01 -9.18653727e-01 -4.47426230e-01
-1.41027570e-01 -8.74284327e-01 -7.14514852e-01 -6.88181147e-02
1.25468791e-01 -4.23613369e-01 -1.31046996e-01 8.08591604e-01
8.77802730e-01 1.51361421e-01 8.76784384e-01 -1.32975590e+00
-1.38244188e+00 1.33609390e+00 -4.43621576e-01 1.39689997e-01
3.55095416e-01 2.00926960e-01 1.23524427e+00 -9.66482103e-01
8.67490053e-01 1.40444064e+00 3.01880509e-01 8.70345175e-01
-8.25628579e-01 -7.54612446e-01 4.47251827e-01 -1.17457561e-01
-9.68312383e-01 -7.20129758e-02 1.02699244e+00 -4.04011786e-01
6.44785941e-01 -1.69918090e-01 6.39104784e-01 1.39447677e+00
2.09862649e-01 1.11745095e+00 5.42379320e-01 -2.72947907e-01
4.71506529e-02 2.81619161e-01 -2.39669472e-01 2.68361270e-01
3.12601089e-01 -2.23660991e-01 -6.21348798e-01 -1.25796318e-01
6.12826109e-01 2.46260390e-02 -4.77132052e-01 2.64740527e-01
-1.37663150e+00 1.33713460e+00 4.77016360e-01 3.13414857e-02
-5.86514115e-01 9.79659408e-02 4.61450905e-01 -9.50148180e-02
9.38018978e-01 7.93421149e-01 -1.50135830e-02 -1.63001567e-01
-8.84360135e-01 7.59263873e-01 8.59168649e-01 1.42526329e+00
5.91688454e-01 2.66047806e-01 -1.05168164e+00 7.23225832e-01
4.09923822e-01 5.47316432e-01 9.29964066e-01 -3.37424546e-01
8.13034534e-01 7.36113489e-01 1.82241857e-01 -1.02939844e+00
5.25034741e-02 -5.12609720e-01 -1.05361235e+00 -4.71035570e-01
-2.91553885e-01 -8.61697733e-01 -6.84334099e-01 1.57693899e+00
-4.26374450e-02 -6.99732006e-02 1.28978938e-01 7.23852277e-01
1.14217412e+00 1.22671473e+00 1.26985610e-01 -1.84067652e-01
1.02791655e+00 -1.32106209e+00 -7.61508048e-01 -1.15739614e-01
2.34297186e-01 -8.06062877e-01 1.07757485e+00 -7.06806779e-02
-1.35485876e+00 -6.86852872e-01 -8.12183797e-01 -9.86122862e-02
-1.40754834e-01 2.87730694e-01 5.60084522e-01 1.08280517e-01
-1.03037941e+00 1.04127929e-01 -3.68547618e-01 -2.72776246e-01
5.05997002e-01 -4.84927818e-02 3.10164660e-01 1.99194491e-01
-1.59700465e+00 2.77533203e-01 4.83728558e-01 -1.09411016e-01
-7.96519637e-01 -8.26991796e-01 -7.27590919e-01 1.64240018e-01
1.89918205e-01 -1.06163025e+00 1.63735986e+00 -5.56191444e-01
-1.60842860e+00 1.26573652e-01 -4.31416929e-01 -5.71470201e-01
3.45787138e-01 -1.13915116e-01 -3.23524237e-01 -3.33185843e-03
3.73909593e-01 1.23192549e+00 7.04048991e-01 -1.19197392e+00
-8.63117099e-01 3.82762671e-01 1.71168506e-01 3.81738961e-01
-4.87322927e-01 -1.71000361e-01 -2.44000524e-01 -1.18087971e+00
-4.59824115e-01 -6.31820142e-01 -4.30159926e-01 -4.92029816e-01
-8.91739786e-01 -7.09605694e-01 6.61163628e-01 -4.87106770e-01
1.69709253e+00 -1.78136730e+00 7.06480443e-02 -2.75490314e-01
1.84036463e-01 1.70359462e-01 -1.27786115e-01 9.89278555e-01
2.89714932e-01 5.23589969e-01 6.60213679e-02 -3.09177876e-01
2.72393078e-01 -2.86766708e-01 -9.87341166e-01 -5.51463366e-01
3.59258771e-01 1.30198121e+00 -1.11368418e+00 -4.66258138e-01
-4.47079241e-01 1.91002309e-01 -8.95278215e-01 5.78275323e-01
-9.02138531e-01 2.19656795e-01 -8.64253700e-01 3.27353448e-01
3.94193918e-01 -5.26079834e-01 -4.64370728e-01 6.72816485e-02
-4.69176024e-02 7.90846169e-01 -4.85788971e-01 1.67200553e+00
-4.94742811e-01 4.59656894e-01 -5.45491159e-01 -2.55337000e-01
1.22541416e+00 4.90365684e-01 4.63364683e-02 -4.60504860e-01
1.40092587e-02 1.83583334e-01 2.74542086e-02 -2.47842729e-01
1.22814274e+00 -2.17693582e-01 -2.46126890e-01 1.20558619e+00
-7.82651305e-02 -2.43057251e-01 5.57929456e-01 7.53973305e-01
6.76709712e-01 -1.13501169e-01 9.41321552e-02 -1.62427232e-01
4.69396740e-01 3.95375155e-02 2.40094200e-01 6.28699601e-01
3.51582199e-01 8.33984196e-01 6.25397682e-01 4.24024872e-02
-9.63185608e-01 -7.85897672e-01 3.69494706e-01 9.65346575e-01
-3.32375355e-02 -5.59072733e-01 -5.89933813e-01 -7.46756732e-01
-1.26228958e-01 8.38631809e-01 -5.89280546e-01 -3.14936638e-01
-5.17634869e-01 -6.34770453e-01 9.63791534e-02 6.93646610e-01
5.19161820e-01 -1.60677004e+00 -1.46781400e-01 5.23883224e-01
-5.48519075e-01 -6.82929277e-01 -1.08986306e+00 -3.53098750e-01
-7.66812265e-01 -3.75872940e-01 -1.29456604e+00 -9.18326795e-01
8.23206782e-01 2.96830535e-01 1.35387588e+00 -3.39916013e-02
1.83417022e-01 -1.51811600e-01 -8.12333763e-01 -7.96510994e-01
-4.52746272e-01 8.71050656e-01 -4.18047160e-01 -1.94387630e-01
2.56021738e-01 -4.04994547e-01 -7.65636265e-01 3.47513445e-02
-8.96489561e-01 3.57429892e-01 6.99244499e-01 1.00827825e+00
1.91009238e-01 -1.96871370e-01 1.32597184e+00 -1.09088838e+00
1.61851847e+00 -8.25228691e-01 -2.94592857e-01 -1.91118987e-03
-6.91409171e-01 1.15551218e-01 8.28354895e-01 -4.02800113e-01
-1.36092639e+00 -7.23496795e-01 -1.07579999e-01 6.48572370e-02
1.22124121e-01 8.91412199e-01 -2.68376708e-01 7.31763780e-01
4.85392898e-01 4.32295561e-01 -1.62270710e-01 -2.85145998e-01
4.76317853e-01 7.82026529e-01 2.70100892e-01 -5.52362204e-01
8.29150319e-01 1.08183004e-01 -5.39614320e-01 -2.12831885e-01
-1.00361180e+00 -3.26660901e-01 -1.61244482e-01 -5.83450776e-03
6.94289923e-01 -1.09520984e+00 -2.00927332e-01 4.22347128e-01
-1.54208910e+00 -3.07385117e-01 -4.30120230e-01 2.12152973e-01
-2.69677669e-01 5.21294959e-02 -5.51638663e-01 -6.10882938e-01
-9.59897280e-01 -9.91621912e-01 1.28072369e+00 6.24733686e-01
-2.67506361e-01 -1.14391351e+00 1.50420994e-01 2.15519089e-02
7.20266521e-01 -4.72467430e-02 7.76707768e-01 -6.81239903e-01
-9.52860594e-01 -7.98705965e-02 -4.48514998e-01 1.67323366e-01
3.08926046e-01 -1.36939704e-01 -4.78743106e-01 -2.24833190e-02
-4.75507140e-01 -3.21652323e-01 1.03497028e+00 2.12784231e-01
1.06641185e+00 -8.15355241e-01 -4.17037874e-01 3.12042922e-01
9.88183618e-01 1.46144047e-01 5.12579501e-01 5.75065874e-02
7.89098978e-01 4.59639192e-01 6.00552619e-01 7.22121179e-01
8.71393263e-01 4.08492208e-01 1.93633854e-01 -8.86391252e-02
-2.50088215e-01 -8.71963799e-01 3.34211051e-01 1.20482087e+00
1.78497881e-01 -9.22747016e-01 -4.69942868e-01 8.69604588e-01
-1.87318850e+00 -1.14761519e+00 -7.35235438e-02 1.52694678e+00
1.13312936e+00 1.27086043e-01 2.47373387e-01 -4.01084691e-01
7.65850365e-01 3.75820398e-01 -4.25533026e-01 -4.09650713e-01
-4.30767201e-02 -8.85910615e-02 -3.66251096e-02 2.34368324e-01
-6.40243053e-01 1.05796409e+00 6.06246519e+00 7.26477504e-01
-1.06066358e+00 -1.35301903e-01 7.05367565e-01 9.11099538e-02
-1.01851022e+00 -1.39118433e-01 -1.35462070e+00 1.00025332e+00
7.81680048e-01 -1.08673215e+00 -2.08679393e-01 9.49369073e-01
4.92560983e-01 3.26544404e-01 -8.38750601e-01 4.65177000e-01
5.05504310e-01 -1.63628626e+00 8.03142548e-01 1.51174897e-02
1.37967122e+00 -3.57154846e-01 2.34161302e-01 5.23008883e-01
7.58706570e-01 -9.28440750e-01 5.07411003e-01 6.53109789e-01
6.38618290e-01 -8.42134237e-01 9.29762900e-01 3.30467254e-01
-1.03347158e+00 1.42048225e-01 -5.89261949e-01 -8.27244371e-02
8.26760828e-01 7.06420302e-01 -1.22858763e+00 5.13173878e-01
1.04362413e-01 9.42842782e-01 -4.24176872e-01 1.24037611e+00
-6.65538132e-01 8.49148035e-01 2.55658805e-01 -7.13992178e-01
5.91739118e-01 2.92804576e-02 3.45648408e-01 1.13004434e+00
6.81285560e-01 5.48895858e-02 3.06565374e-01 1.20856154e+00
-4.94419873e-01 2.21934244e-01 -3.19854885e-01 -2.76299059e-01
5.38922012e-01 1.19629240e+00 -3.61553282e-01 -4.62311774e-01
-1.82391718e-01 7.94326663e-01 1.13668278e-01 4.22753513e-01
-6.97568178e-01 -6.75983787e-01 3.15762550e-01 1.37235284e-01
4.53132927e-01 1.13244362e-01 -1.15772337e-01 -1.18713415e+00
-1.09555922e-01 -8.31945479e-01 1.55842572e-01 -1.05873096e+00
-1.58040619e+00 7.96562791e-01 1.02809623e-01 -1.28120697e+00
-7.83583820e-01 -6.38459027e-02 -1.29576635e+00 1.40605569e+00
-1.67553353e+00 -1.23334563e+00 -3.86833847e-01 1.70564339e-01
1.06161261e+00 -3.75654757e-01 3.79769623e-01 3.16517334e-03
-2.82334208e-01 6.89010143e-01 -1.92088813e-01 1.56397849e-01
7.86087155e-01 -1.35256886e+00 9.73608971e-01 7.85633266e-01
-2.29256302e-01 9.19217765e-01 5.42787850e-01 -9.45308685e-01
-7.34744012e-01 -1.52481186e+00 1.46701908e+00 -1.58550009e-01
6.71703577e-01 -3.32034469e-01 -7.70333946e-01 5.39742887e-01
7.81293094e-01 -8.16793323e-01 6.78672969e-01 -7.59072881e-03
-9.22583789e-02 1.74798116e-01 -5.00919640e-01 9.20327783e-01
8.66327405e-01 -2.52385616e-01 -7.36604273e-01 3.74586999e-01
1.39668620e+00 -3.23086023e-01 -2.92078435e-01 6.13213219e-02
1.90881416e-01 -5.75735569e-01 6.08522296e-01 -4.88175750e-01
1.10648429e+00 -1.94742084e-01 4.54286128e-01 -1.89666033e+00
-3.64052981e-01 -9.90827203e-01 -1.61058083e-01 1.71786880e+00
8.32500935e-01 -4.05328274e-01 8.55617166e-01 2.23367885e-01
-5.41941822e-01 -8.83798659e-01 -1.36332691e-01 -3.62055421e-01
2.15635255e-01 1.68875545e-01 1.05598998e+00 5.02630830e-01
2.84379363e-01 9.34204876e-01 -5.12747586e-01 -5.35302937e-01
-1.14430692e-02 3.93542469e-01 9.66505885e-01 -1.17362559e+00
-1.30806670e-01 -7.15777218e-01 1.57908633e-01 -1.56034720e+00
2.18672678e-01 -9.62373435e-01 3.57882529e-01 -2.12149954e+00
3.51286501e-01 -3.55896443e-01 2.88703978e-01 2.29671955e-01
-8.65220189e-01 -1.75200533e-02 1.00135669e-01 1.81740880e-01
-6.17717981e-01 1.10214579e+00 1.67082822e+00 -1.06896073e-01
-1.39230207e-01 1.80269316e-01 -1.40487945e+00 3.26721251e-01
8.07407677e-01 -1.73071280e-01 -7.95108676e-01 -4.91028935e-01
5.72768211e-01 -4.04743813e-02 -1.29685644e-02 -7.44382799e-01
3.29624563e-01 -2.26347759e-01 1.97866485e-02 -1.03577876e+00
-1.42920269e-02 -4.79837731e-02 -4.13241178e-01 1.10609800e-01
-8.87593091e-01 4.24621046e-01 -1.31268963e-01 6.57168508e-01
-4.54401046e-01 -2.24280044e-01 2.12422594e-01 -2.93863326e-01
-3.60450983e-01 7.32599378e-01 -3.98606539e-01 4.77168262e-01
6.37363672e-01 1.45305216e-01 -6.49375975e-01 -7.63064921e-01
1.16101746e-03 5.29212415e-01 2.04359442e-01 7.25806117e-01
6.77667677e-01 -1.58263397e+00 -1.19129622e+00 -3.92814204e-02
3.04684967e-01 2.64322430e-01 1.95586547e-01 3.12572360e-01
-1.37940422e-01 6.95151150e-01 -2.77314037e-02 -5.92057966e-02
-5.59538722e-01 4.25707161e-01 -2.68289238e-01 -5.14580548e-01
-4.82806325e-01 1.10037839e+00 3.70484859e-01 -2.09361359e-01
1.82531271e-02 -3.94459337e-01 -4.85453129e-01 2.46924862e-01
7.86644340e-01 7.43915364e-02 -3.28812152e-01 -3.89275342e-01
3.15501302e-01 7.76114613e-02 -5.52883208e-01 -2.61331260e-01
1.06864500e+00 -2.28964433e-01 4.12322022e-03 3.13961327e-01
9.82632101e-01 6.82530031e-02 -1.11472046e+00 -1.10477887e-01
-1.63716421e-01 -2.48781219e-01 -1.61060974e-01 -8.74193966e-01
-9.48287547e-01 1.09060216e+00 -3.80663693e-01 3.38670313e-01
8.08983684e-01 2.41916962e-02 1.48077118e+00 1.10852458e-01
6.47677779e-02 -8.61343384e-01 5.37857056e-01 7.62943566e-01
1.11633146e+00 -1.07079184e+00 -2.92685777e-01 -2.88359374e-01
-9.36364055e-01 8.35846126e-01 1.11289155e+00 -1.10198699e-01
4.20555174e-01 -7.47334212e-02 9.44305509e-02 -3.98957692e-02
-1.28570890e+00 -1.35097384e-01 4.07838106e-01 8.06923509e-01
7.58669615e-01 -2.32625127e-01 -6.30616248e-01 8.08009148e-01
-6.63258493e-01 4.80622835e-02 8.55543733e-01 5.34340918e-01
-7.15323985e-01 -1.28715181e+00 2.86021009e-02 6.60628080e-01
-3.27016532e-01 -6.80196702e-01 -3.76486242e-01 4.09827858e-01
-1.57484412e-02 8.43517721e-01 1.35406479e-02 -2.53459692e-01
-4.11193855e-02 -4.10751365e-02 -3.23499531e-01 -1.15037811e+00
-7.64831364e-01 1.22644350e-01 7.95259699e-02 6.29740879e-02
-1.90508276e-01 -6.13748610e-01 -1.07587004e+00 -1.36698365e-01
-4.64078516e-01 8.14723372e-01 3.26162040e-01 6.21638358e-01
7.62528062e-01 8.64502370e-01 6.31713867e-01 -6.26441479e-01
-5.09245753e-01 -1.40593052e+00 -4.46474850e-01 2.25665078e-01
1.37183666e-01 -3.59833568e-01 -3.34312797e-01 2.33839750e-01] | [12.150935173034668, 9.01282024383545] |
bdc8f0f4-2cd0-4345-8f55-ac9801a4d801 | a-signed-subgraph-encoding-approach-via | 2305.09869 | null | https://arxiv.org/abs/2305.09869v1 | https://arxiv.org/pdf/2305.09869v1.pdf | A Signed Subgraph Encoding Approach via Linear Optimization for Link Sign Prediction | In this paper, we consider the problem of inferring the sign of a link based on limited sign data in signed networks. Regarding this link sign prediction problem, SDGNN (Signed Directed Graph Neural Networks) provides the best prediction performance currently to the best of our knowledge. In this paper, we propose a different link sign prediction architecture call SELO (Subgraph Encoding via Linear Optimization), which obtains overall leading prediction performances compared the state-of-the-art algorithm SDGNN. The proposed model utilizes a subgraph encoding approach to learn edge embeddings for signed directed networks. In particular, a signed subgraph encoding approach is introduced to embed each subgraph into a likelihood matrix instead of the adjacency matrix through a linear optimization method. Comprehensive experiments are conducted on six real-world signed networks with AUC, F1, micro-F1, and Macro-F1 as the evaluation metrics. The experiment results show that the proposed SELO model outperforms existing baseline feature-based methods and embedding-based methods on all the six real-world networks and in all the four evaluation metrics. | ['Yaonan Wang', 'Shaolin Tan', 'Zhihong Fang'] | 2023-05-17 | null | null | null | null | ['link-sign-prediction'] | ['graphs'] | [ 1.70114249e-01 5.85247815e-01 -6.54728353e-01 -6.42770886e-01
6.14324026e-02 -2.98509926e-01 6.05487585e-01 2.56838696e-03
4.81949486e-02 7.17783868e-01 1.41042277e-01 -2.38789126e-01
-7.85804451e-01 -8.56861889e-01 -4.35793310e-01 -2.33440876e-01
-7.49118984e-01 4.40151900e-01 5.49580753e-01 -1.46059811e-01
1.47517055e-01 2.60176122e-01 -9.39572275e-01 -1.09818801e-01
1.00835836e+00 1.17638433e+00 -6.20592833e-01 3.14299017e-01
3.03693563e-02 8.31292748e-01 -8.69592205e-02 -1.01073301e+00
4.02020693e-01 -4.49762970e-01 -3.77155751e-01 -4.06246156e-01
7.86795676e-01 -2.26129934e-01 -7.84487188e-01 1.14705586e+00
3.53432655e-01 -2.65364468e-01 7.42327571e-01 -1.81624866e+00
-7.30985284e-01 9.33954835e-01 -3.51606578e-01 -2.00371638e-01
2.68505245e-01 -7.55395666e-02 1.80965543e+00 -7.98584282e-01
1.05651855e+00 1.14757848e+00 8.47330332e-01 1.22598141e-01
-9.19202149e-01 -6.28436446e-01 2.26225749e-01 4.69843775e-01
-1.16741526e+00 3.24409269e-02 1.25138938e+00 -2.16909155e-01
6.28108621e-01 -3.43345515e-02 8.16349685e-01 7.13861942e-01
1.84366584e-01 8.43013465e-01 1.05759859e+00 -5.80450930e-02
-1.35503054e-01 -1.46843448e-01 3.76040637e-01 1.38198686e+00
8.11989307e-01 1.57537252e-01 -7.29445755e-01 -4.14072983e-02
3.62265170e-01 -1.75157160e-01 -4.86045539e-01 -8.33732545e-01
-1.14305890e+00 7.77505696e-01 1.16272402e+00 -3.22654359e-02
-2.92610139e-01 4.20600444e-01 3.74756575e-01 4.80652452e-01
4.16975796e-01 5.99638820e-02 -4.41482693e-01 2.04414144e-01
-6.81622028e-01 -1.71867862e-01 1.30826819e+00 7.29101717e-01
3.71803850e-01 6.19112328e-02 -8.74550194e-02 6.13842487e-01
9.66533899e-01 2.30414629e-01 1.45451769e-01 -5.07560134e-01
6.81278050e-01 1.07474172e+00 -5.42214870e-01 -1.64640772e+00
-3.74619693e-01 -7.44577229e-01 -7.73686528e-01 9.45750028e-02
1.86734065e-01 -1.46635622e-01 -8.96846771e-01 1.57721961e+00
1.95215166e-01 6.75472260e-01 -4.49405946e-02 9.29482281e-01
1.18291962e+00 3.52166831e-01 -2.83168197e-01 1.20407552e-01
7.98401475e-01 -1.15003502e+00 -5.95890284e-01 -1.03052787e-01
5.68260133e-01 -3.49617571e-01 2.85891950e-01 1.77729994e-01
-5.61916769e-01 -1.16144754e-01 -1.53306317e+00 1.79026917e-01
-7.16784835e-01 2.08993211e-01 1.06356049e+00 4.98826414e-01
-1.07275081e+00 9.27999496e-01 -5.47373474e-01 -5.22942841e-01
3.69746447e-01 5.00796139e-01 -6.78711772e-01 -3.47407788e-01
-1.19966984e+00 5.22869468e-01 4.31320786e-01 4.91203636e-01
-3.51296633e-01 -5.45186043e-01 -9.73561704e-01 1.15160018e-01
3.06569517e-01 -4.98646438e-01 4.58656728e-01 -5.40667593e-01
-1.21289742e+00 6.98742628e-01 1.86357349e-01 -3.94292861e-01
6.69281840e-01 1.22034043e-01 -7.68663585e-01 6.83485195e-02
-2.40625426e-01 5.54313779e-01 6.32436693e-01 -1.28566074e+00
-1.97777092e-01 -7.23908003e-03 5.88883385e-02 -2.27387324e-01
-6.03189468e-01 -3.88875037e-01 -3.93655241e-01 -6.65471017e-01
6.16767228e-01 -9.15170372e-01 2.34572038e-01 6.43049538e-01
-5.77719033e-01 -2.97925681e-01 1.09766424e+00 -6.68130994e-01
1.43768919e+00 -1.84861338e+00 1.06533386e-01 9.47416246e-01
7.03503728e-01 4.47320282e-01 -4.27029103e-01 4.52626348e-01
-2.09954798e-01 1.03955910e-01 -1.52371600e-01 -1.96108207e-01
4.47003067e-01 2.74686068e-01 2.88820833e-01 4.91860688e-01
7.46210292e-02 1.22535682e+00 -1.10190225e+00 -7.07480550e-01
-1.71393216e-01 6.00773275e-01 -3.94694835e-01 -1.22258201e-01
9.41103771e-02 -2.38827690e-01 -3.08178008e-01 9.20160413e-01
7.80656457e-01 -4.77936834e-01 7.98367679e-01 -7.13838637e-01
4.41642195e-01 1.06241843e-02 -1.28454137e+00 1.21017504e+00
7.66013563e-02 7.95508087e-01 -1.75380930e-01 -8.41953874e-01
1.34763157e+00 3.13657150e-02 4.45170462e-01 -3.06385070e-01
2.06320867e-01 6.16954505e-01 3.63234699e-01 -3.22746962e-01
1.29798904e-01 3.16825539e-01 1.80912361e-01 4.28684682e-01
3.08034331e-01 2.77975112e-01 5.61998188e-01 5.78045487e-01
1.48580480e+00 -2.12534834e-02 1.63832024e-01 1.86119363e-01
6.60636008e-01 -6.06250346e-01 7.39781380e-01 3.07755619e-01
-5.42056262e-01 3.85988235e-01 1.09266829e+00 -7.28241503e-02
-4.44870949e-01 -1.24461722e+00 7.79539570e-02 6.05368257e-01
2.05307856e-01 -5.81878960e-01 -2.23028153e-01 -1.27396870e+00
6.21676207e-01 1.55801699e-01 -6.74697220e-01 -2.99953580e-01
-4.53651547e-01 -5.11138558e-01 5.71441054e-01 5.67596376e-01
5.39717853e-01 -8.66905808e-01 6.58412218e-01 -1.42060481e-02
2.03858271e-01 -1.14093924e+00 -5.87646484e-01 -1.06009059e-01
-8.73022914e-01 -1.58500206e+00 -4.40449059e-01 -1.21151185e+00
1.12191498e+00 -7.57991076e-02 6.92604959e-01 4.07510489e-01
-1.36375785e-01 1.88501388e-01 -4.43305373e-01 1.74061581e-01
2.32803375e-02 -6.80695707e-03 -1.34675354e-01 4.89111125e-01
8.84970725e-02 -6.78677917e-01 -6.42632067e-01 2.93854952e-01
-5.58647215e-01 -2.72600055e-01 9.38184559e-01 9.51895356e-01
2.78520197e-01 -2.50309527e-01 5.67428708e-01 -9.87276375e-01
8.85199904e-01 -4.15981442e-01 -5.57676315e-01 5.45068085e-01
-1.05969763e+00 2.30048984e-01 4.33126748e-01 -1.12774462e-01
-5.68191051e-01 -2.92452812e-01 4.56955403e-01 -8.73708948e-02
6.26622558e-01 9.46357071e-01 -1.08707502e-01 -8.18387508e-01
1.11847676e-01 -5.79805113e-03 9.73488092e-02 -4.87231225e-01
2.70245492e-01 4.39353049e-01 3.69184047e-01 -1.57952473e-01
1.03568256e+00 2.00073615e-01 6.10770166e-01 -2.76792854e-01
-5.90343297e-01 -2.54002094e-01 -7.56892562e-01 -5.09571135e-01
3.27674657e-01 -4.91629064e-01 -8.86144936e-01 4.86233532e-01
-9.37007010e-01 1.76799633e-02 2.20637038e-01 6.00523472e-01
-9.60462317e-02 5.90088487e-01 -6.54738724e-01 -3.66537869e-01
-3.21758866e-01 -7.31357396e-01 6.30776346e-01 1.21679351e-01
-9.76589024e-02 -1.28205585e+00 5.38259745e-02 2.95259655e-01
4.49543893e-01 6.17800832e-01 8.97018492e-01 -7.73257613e-01
-8.51360857e-01 -7.69279301e-01 -8.83424759e-01 3.95665437e-01
1.97371155e-01 1.59541994e-01 -4.84084159e-01 -2.75111884e-01
-1.08198798e+00 -1.78231001e-01 1.14439833e+00 -2.30992511e-02
7.69168019e-01 -2.69106507e-01 -4.68787760e-01 7.77193725e-01
1.53672898e+00 -3.09406877e-01 4.14377332e-01 6.50018901e-02
8.89146507e-01 3.60414505e-01 6.22738600e-01 2.46115357e-01
6.72550738e-01 3.76756608e-01 7.58522213e-01 2.04860032e-01
-5.76227188e-01 -7.35685527e-01 3.75498682e-01 1.22992110e+00
-1.46732047e-01 -5.85066855e-01 -6.25542641e-01 5.18247306e-01
-2.09635615e+00 -7.61721134e-01 -4.12925392e-01 1.78343463e+00
3.51875275e-01 5.58922291e-01 -2.57143140e-01 3.25854272e-01
6.71553493e-01 6.76972091e-01 -3.52475554e-01 -1.71847090e-01
-1.60186827e-01 3.71250361e-02 6.36308372e-01 6.31658792e-01
-1.06576562e+00 7.16942012e-01 5.58214188e+00 3.21030855e-01
-1.15678728e+00 -1.22714870e-01 3.80715691e-02 3.60529661e-01
-3.20040047e-01 1.96942672e-01 -5.02126694e-01 5.51285267e-01
4.07404840e-01 -2.72618353e-01 3.43567878e-01 5.73671043e-01
2.55357251e-02 4.87341374e-01 -1.02885890e+00 7.50089824e-01
4.54817951e-01 -1.01118183e+00 3.19766589e-02 1.01724595e-01
8.93957496e-01 1.34936482e-01 -2.07889415e-02 2.02287942e-01
2.51848340e-01 -7.72457659e-01 2.29785174e-01 8.41699243e-01
6.21204674e-01 -3.45410079e-01 1.15986943e+00 -2.65019089e-01
-1.59877646e+00 1.23274848e-02 1.34835482e-01 2.94950575e-01
4.06128675e-01 6.98545456e-01 -8.78675461e-01 8.53470445e-01
2.08148941e-01 1.50290632e+00 -8.32545102e-01 1.56833327e+00
-8.20191562e-01 7.85262644e-01 -4.50116754e-01 -4.82100517e-01
2.45808005e-01 -2.49589443e-01 5.65871596e-01 9.76707995e-01
2.88186401e-01 -3.21132064e-01 1.24357685e-01 5.01803815e-01
-6.29890323e-01 1.13904312e-01 -6.04723215e-01 -3.03750396e-01
2.67227948e-01 1.34522140e+00 -5.68701625e-01 -5.60816117e-02
-3.73706609e-01 9.33406651e-01 3.36743385e-01 4.05086070e-01
-9.13704216e-01 -9.55086231e-01 3.57362896e-01 2.75606979e-02
4.32758212e-01 -1.51864067e-01 2.74772532e-02 -1.06468308e+00
4.06190872e-01 -4.51630265e-01 4.66996551e-01 -6.54465616e-01
-1.41729164e+00 2.69080043e-01 -2.18583360e-01 -1.48398972e+00
4.27333303e-02 -1.03127372e+00 -9.10842836e-01 2.65719086e-01
-1.87563360e+00 -1.51273131e+00 -5.86171746e-01 8.99151489e-02
-4.16829854e-01 -3.63749206e-01 5.09245217e-01 6.33309603e-01
-4.75389004e-01 9.86668050e-01 1.47083133e-01 6.52665675e-01
7.48118281e-01 -1.24939847e+00 1.70470938e-01 7.16439366e-01
1.50577322e-01 4.29329544e-01 2.43365899e-01 -9.27404761e-01
-1.47580230e+00 -1.11428988e+00 1.35705960e+00 2.53791809e-01
1.13229167e+00 -7.23604765e-03 -5.11746824e-01 7.93240309e-01
-1.99176878e-01 7.78315902e-01 5.78539371e-01 1.03857897e-01
-5.89370012e-01 -4.95650381e-01 -1.35101998e+00 3.63054574e-01
1.58707511e+00 -4.09829795e-01 -1.63499519e-01 9.55736414e-02
4.80434507e-01 -1.87026143e-01 -1.17257309e+00 8.25992823e-01
9.60593700e-01 -7.16450870e-01 8.67736816e-01 -7.75604725e-01
2.53637284e-01 -5.25584817e-01 -1.62735149e-01 -1.32439899e+00
-2.64506280e-01 -4.59818453e-01 -7.34476745e-01 1.24763739e+00
6.92195714e-01 -1.11442029e+00 1.12392926e+00 4.72312756e-02
1.30771577e-01 -1.13849413e+00 -8.73480201e-01 -9.79418099e-01
-7.01110125e-01 -1.31967813e-01 6.83813751e-01 1.08146286e+00
8.37439969e-02 1.12528838e-01 -3.68596554e-01 1.95698440e-01
1.11098826e+00 5.33011742e-02 5.92696488e-01 -1.58578479e+00
-2.73424506e-01 -5.31200230e-01 -1.36735153e+00 -9.50154603e-01
4.14819121e-01 -1.48473155e+00 -3.32464963e-01 -1.90537572e+00
-7.30855241e-02 -3.52164686e-01 -6.41271532e-01 5.60213506e-01
7.54004046e-02 3.24390650e-01 4.45755512e-01 -2.54378229e-01
-8.85733306e-01 7.16530502e-01 1.37139153e+00 -5.96728086e-01
1.31202757e-01 -1.31496698e-01 -4.65428591e-01 6.65178239e-01
5.43229938e-01 -4.43160206e-01 -4.49304789e-01 -3.00266176e-01
3.51871192e-01 -1.88158393e-01 4.83305961e-01 -9.43712711e-01
3.89422119e-01 2.51119494e-01 3.59949879e-02 -5.02723813e-01
3.61210585e-01 -9.39897418e-01 -1.49679065e-01 7.03355670e-01
-2.42936522e-01 -2.87675947e-01 -4.52787668e-01 9.96513605e-01
-4.24253702e-01 -5.04831001e-02 3.73672277e-01 6.51497781e-01
-1.01340473e+00 5.67939937e-01 6.15269601e-01 -5.06510399e-03
9.62884247e-01 -3.43585134e-01 -7.21715033e-01 -5.29591203e-01
-7.94307470e-01 6.63236737e-01 -5.39590605e-02 5.88270843e-01
9.68191147e-01 -1.74365819e+00 -6.71374500e-01 1.26233473e-01
4.63416666e-01 -6.20893061e-01 -2.43234470e-01 1.02923036e+00
-7.19140410e-01 3.50458890e-01 1.29615948e-01 -2.30188340e-01
-1.48801327e+00 -2.02330351e-01 3.13126415e-01 -6.40500784e-01
-2.35587254e-01 9.08909738e-01 -7.55012274e-01 -9.12492752e-01
4.63486165e-01 -2.37646922e-01 -1.13977939e-01 1.07509330e-01
-1.39184043e-01 6.10462785e-01 -9.13757086e-02 -7.01943457e-01
-5.92194498e-01 7.81025469e-01 5.71153127e-02 4.04278278e-01
1.53541946e+00 4.13211703e-01 -3.78801435e-01 -5.81246130e-02
1.71399891e+00 -7.12186024e-02 -9.11054969e-01 -3.57284367e-01
9.29713696e-02 -5.39896190e-01 -3.75351794e-02 -9.67875540e-01
-1.50775826e+00 4.91741866e-01 5.20940006e-01 1.39307420e-04
6.34019554e-01 -6.18994422e-02 8.80058527e-01 4.54372257e-01
3.26115668e-01 -8.23969662e-01 1.65045425e-01 4.65572774e-01
8.18254650e-01 -1.48236430e+00 2.99053848e-01 -8.91032755e-01
-2.58127898e-01 1.27058387e+00 7.62930095e-01 -3.17078501e-01
1.40567911e+00 -2.89392531e-01 -1.43185228e-01 -4.93616849e-01
-5.63198924e-01 -1.95677608e-01 8.38680267e-01 5.26253819e-01
4.04755503e-01 2.07942799e-01 -8.49481940e-01 3.18662494e-01
-8.74405727e-02 1.69711515e-01 2.38152504e-01 8.07853818e-01
-3.98560345e-01 -1.35351837e+00 1.44491807e-01 9.29225504e-01
4.31534573e-02 3.78038809e-02 -1.00638199e+00 8.52354705e-01
-1.32970763e-02 6.56029165e-01 -3.38264495e-01 -6.98327363e-01
3.96289945e-01 -9.31866243e-02 4.03180957e-01 -2.80586183e-01
-1.21830679e-01 -7.44926155e-01 5.34571350e-01 -5.66478789e-01
-3.33878696e-01 -5.00036895e-01 -1.35685897e+00 -4.11705613e-01
-5.06857634e-01 -1.80806577e-01 5.28467715e-01 6.15018070e-01
3.10897052e-01 5.45018256e-01 7.46295869e-01 -5.32549143e-01
-5.27470648e-01 -7.55425155e-01 -8.29336524e-01 4.01244074e-01
1.68704927e-01 -7.72423744e-01 -5.69181383e-01 -6.28453195e-01] | [7.170010566711426, 6.267354965209961] |
0d20a148-1a32-479a-bbfd-ebd8c6dedaac | multi-view-gait-recognition-based-on-siamese | 2210.10421 | null | https://arxiv.org/abs/2210.10421v1 | https://arxiv.org/pdf/2210.10421v1.pdf | Multi-view Gait Recognition based on Siamese Vision Transformer | While the Vision Transformer has been used in gait recognition, its application in multi-view gait recognition is still limited. Different views significantly affect the extraction and identification accuracy of the characteristics of gait contour. To address this, this paper proposes a Siamese Mobile Vision Transformer (SMViT). This model not only focuses on the local characteristics of the human gait space but also considers the characteristics of long-distance attention associations, which can extract multi-dimensional step status characteristics. In addition, it describes how different perspectives affect gait characteristics and generate reliable perspective feature relationship factors. The average recognition rate of SMViT on the CASIA B data set reached 96.4%. The experimental results show that SMViT can attain state-of-the-art performance compared to advanced step recognition models such as GaitGAN, Multi_view GAN, Posegait and other gait recognition models. | ['Feiyan Cheng', 'Ruoyu Li', 'Lijun Yun', 'Yanchen Yang'] | 2022-10-19 | null | null | null | null | ['gait-recognition'] | ['computer-vision'] | [-3.24833304e-01 -7.38353014e-01 -2.62925714e-01 4.73372936e-02
-5.90013862e-01 -3.04951500e-02 3.51520509e-01 -9.04565811e-01
-1.22258335e-01 5.75743139e-01 2.85100639e-01 4.92240161e-01
2.16710702e-01 -8.12439322e-01 6.10287301e-02 -9.95006382e-01
-7.61773363e-02 4.42090631e-01 2.13862941e-01 -5.12103796e-01
8.34693462e-02 1.49790168e-01 -1.57536113e+00 -4.62592617e-02
8.02264988e-01 8.39719415e-01 1.23208873e-01 5.67078948e-01
3.21138442e-01 5.91566980e-01 -4.85972792e-01 -4.85349566e-01
-4.74776663e-02 -5.01100540e-01 -2.13377103e-01 1.53825149e-01
3.68209720e-01 -3.08990330e-01 -6.32058978e-01 9.32779074e-01
8.73675108e-01 -6.91747665e-02 8.22179556e-01 -1.44035041e+00
-7.73748040e-01 6.43120706e-02 -1.09509575e+00 7.59048462e-01
4.55577731e-01 2.87645370e-01 7.40371525e-01 -9.48525488e-01
4.60730255e-01 1.20658338e+00 8.27340841e-01 4.42290574e-01
-5.04198790e-01 -6.55704558e-01 -1.96912065e-01 7.17371047e-01
-1.13048923e+00 -2.73492247e-01 8.92933905e-01 -3.62535357e-01
9.69934881e-01 1.05795234e-01 1.07084727e+00 1.29321218e+00
7.19312906e-01 9.25351202e-01 1.20704806e+00 -2.21656710e-01
-2.39767119e-01 -7.18991041e-01 2.75806308e-01 1.15776825e+00
3.73835742e-01 1.27838671e-01 -7.53188372e-01 3.21512550e-01
8.11065316e-01 -2.08265081e-01 -1.80744559e-01 -2.09534064e-01
-1.14998901e+00 6.62759006e-01 1.83027029e-01 4.68089551e-01
-3.51332277e-01 1.70623735e-01 5.35499036e-01 4.69032973e-02
5.05479276e-01 -3.94480228e-01 1.43811017e-01 -7.28109956e-01
-1.05942869e+00 -2.08021179e-01 1.70690104e-01 7.11782992e-01
4.39629287e-01 8.29983652e-01 -1.69730797e-01 8.52044344e-01
4.08143729e-01 1.22318220e+00 7.62482107e-01 -6.82323992e-01
6.97661996e-01 5.45893967e-01 -3.98422629e-01 -1.29166341e+00
-2.69949883e-01 -3.12301159e-01 -1.02313864e+00 -1.99376419e-02
2.58186519e-01 -5.94383478e-02 -1.14410961e+00 1.21905816e+00
-1.82805046e-01 1.43509641e-01 -4.11791466e-02 8.94045770e-01
8.41416776e-01 4.33329672e-01 4.53119923e-04 1.57653466e-02
1.48253238e+00 -9.58742499e-01 -8.34756613e-01 -5.38473845e-01
3.33307058e-01 -5.72524488e-01 9.03205693e-01 1.53386682e-01
-1.01883686e+00 -8.20499778e-01 -1.11030579e+00 2.26306587e-01
-3.81886482e-01 7.78483808e-01 4.47623432e-01 1.15949392e+00
-1.08782947e+00 1.61401436e-01 -1.24919343e+00 -7.55823195e-01
4.36196327e-01 1.63944021e-01 -4.51403111e-01 -3.93999457e-01
-1.12882090e+00 1.08325374e+00 2.15825468e-01 2.03383744e-01
-5.21615922e-01 -2.46519268e-01 -9.62523699e-01 3.13611254e-02
-9.31832101e-03 -1.05575502e+00 4.76001829e-01 -1.42086402e-01
-1.28715813e+00 8.33679736e-01 -2.68199801e-01 -2.55831420e-01
5.66001654e-01 -6.98266849e-02 -7.49484062e-01 3.60633343e-01
5.36743879e-01 3.44444484e-01 1.07626009e+00 -7.14393795e-01
-5.55693030e-01 -7.89926171e-01 -5.59478104e-01 2.91044325e-01
-2.88335294e-01 -1.11341670e-01 -8.65896642e-01 -7.00402021e-01
1.27458116e-02 -1.02070141e+00 2.49457225e-01 -3.71119469e-01
-2.17931092e-01 2.95871925e-02 1.34326017e+00 -1.05361557e+00
1.19770014e+00 -1.93910289e+00 2.37305030e-01 6.42977878e-02
6.03646189e-02 4.33064371e-01 9.41553339e-02 2.85762489e-01
2.07138538e-01 -2.40042582e-01 -8.02410468e-02 -2.17024460e-01
-6.99908286e-02 2.82962888e-01 1.24634482e-01 5.66006124e-01
-1.24766760e-01 1.32203436e+00 -6.08029127e-01 -6.60111964e-01
6.48300588e-01 4.51990396e-01 -4.37846184e-01 -1.92268908e-01
6.53426588e-01 3.32272768e-01 -6.06538236e-01 1.15909076e+00
7.26297617e-01 -2.34154165e-01 1.08944625e-02 -6.41523361e-01
2.43415549e-01 -6.96579158e-01 -7.19519258e-01 1.44992208e+00
-2.55939722e-01 5.88744342e-01 -4.54053432e-01 -1.20679355e+00
1.05792773e+00 1.05479665e-01 8.54511619e-01 -9.01941061e-01
1.73451707e-01 1.36630520e-01 -1.50075197e-01 -7.35407889e-01
4.63130772e-01 9.54823568e-03 -2.53832281e-01 4.09033932e-02
3.72322172e-01 3.57918650e-01 4.12291795e-01 -2.17915967e-01
7.69077361e-01 3.56154948e-01 3.29827040e-01 -1.54908504e-02
7.34717309e-01 -3.86407338e-02 6.76777840e-01 2.06289858e-01
-6.72404945e-01 7.67917275e-01 -2.89173350e-02 -1.70631126e-01
-8.02299738e-01 -1.34664392e+00 2.88769156e-01 4.96260196e-01
2.95387864e-01 -2.69171119e-01 -5.98264098e-01 -6.00566447e-01
-1.59334362e-01 3.50676537e-01 -5.64402223e-01 -4.68854308e-01
-8.51357698e-01 -1.11883795e+00 9.91731822e-01 1.07658923e+00
1.61883056e+00 -8.09478760e-01 -7.46932983e-01 1.40977517e-01
-9.05136645e-01 -1.49538231e+00 -8.48941088e-01 -5.28190255e-01
-8.20285439e-01 -1.24350464e+00 -1.21174765e+00 -7.14781702e-01
3.61576855e-01 3.87831807e-01 9.17048395e-01 -2.61213571e-01
-2.76790977e-01 8.41830432e-01 -3.90948385e-01 1.69471845e-01
1.13690905e-01 -6.76895455e-02 1.71839058e-01 7.18413517e-02
6.54118538e-01 -6.25370920e-01 -5.40292025e-01 5.75961053e-01
-2.29965910e-01 -2.36509323e-01 5.29613078e-01 1.04016054e+00
3.82925540e-01 4.28431928e-02 4.75121707e-01 -8.14981461e-02
5.76889634e-01 -1.86190143e-01 -5.39362095e-02 4.72968012e-01
-5.79539537e-01 -2.52361059e-01 2.56558686e-01 -1.09894164e-01
-1.07148385e+00 -2.86262244e-01 -1.07113510e-01 -8.26555967e-01
-7.98236132e-02 6.36274934e-01 -3.21817815e-01 -2.73514867e-01
1.44180402e-01 9.14287567e-01 3.24587643e-01 -1.36997327e-01
9.33769047e-02 3.85072052e-01 6.17280722e-01 -3.19510490e-01
7.92075753e-01 5.23757577e-01 1.29821554e-01 -1.28248358e+00
-6.23080619e-02 -5.71506739e-01 -7.20529497e-01 -7.84654617e-01
1.21362162e+00 -9.96284068e-01 -6.09813929e-01 1.11049795e+00
-5.38503408e-01 2.17338905e-01 -8.52271989e-02 5.31163990e-01
-9.67489421e-01 7.05251634e-01 -7.21445322e-01 -5.15990496e-01
-5.40705740e-01 -1.31136894e+00 1.17242014e+00 3.00842404e-01
-4.23395149e-02 -1.10497200e+00 -4.29081246e-02 1.05187941e+00
3.59701425e-01 3.15368295e-01 5.88914633e-01 4.93260510e-02
-3.13730747e-01 -1.21017165e-01 -7.08637461e-02 4.03516693e-03
4.57150280e-01 -4.22978960e-02 -8.97991359e-01 -3.24769199e-01
-2.00648323e-01 7.08729699e-02 1.05547178e+00 7.62738824e-01
4.21603978e-01 1.13439701e-01 -5.47018647e-01 6.58137560e-01
1.24773097e+00 5.63364506e-01 1.07867360e+00 4.68140721e-01
9.87352014e-01 1.26125693e-01 9.21766281e-01 2.49607667e-01
6.38752997e-01 9.71799135e-01 1.22542620e-01 5.84858991e-02
-4.58450288e-01 -3.40987116e-01 8.89492273e-01 1.35479891e+00
-7.25317419e-01 -1.76468238e-01 -9.70818579e-01 5.12397587e-01
-1.82084990e+00 -1.37906909e+00 -8.15109760e-02 1.53838801e+00
-1.82498693e-01 -1.57325730e-01 4.59816247e-01 3.72818679e-01
7.90187478e-01 4.11357015e-01 -5.40474474e-01 -2.47548167e-02
-3.49538833e-01 1.86489925e-01 5.49720049e-01 3.65283452e-02
-1.29235864e+00 9.35685158e-01 6.28870535e+00 1.04502833e+00
-1.19948733e+00 1.07259966e-01 3.83597344e-01 2.64681667e-01
3.97067070e-01 -7.14501202e-01 -7.31738865e-01 6.42906964e-01
7.13101327e-01 -3.36760074e-01 3.67918871e-02 5.74338973e-01
2.35222563e-01 -7.34711215e-02 -3.77986640e-01 1.45267951e+00
5.30282438e-01 -1.06469142e+00 1.45310864e-01 4.10806000e-01
3.13971460e-01 -7.20283017e-02 4.05034035e-01 2.80601621e-01
-6.25378191e-02 -8.90585661e-01 4.09874737e-01 4.65786844e-01
1.01552939e+00 -8.22157443e-01 7.61636555e-01 1.69704691e-01
-1.89249122e+00 -4.05317210e-02 -1.40488341e-01 2.58782506e-01
3.91547769e-01 1.90995827e-01 -4.63292778e-01 1.14064074e+00
7.88365543e-01 1.51668823e+00 -8.56482506e-01 9.79982734e-01
7.18573555e-02 7.19287813e-01 -1.54985577e-01 1.79706559e-01
1.52722508e-01 -6.22477889e-01 5.60517788e-01 9.35255051e-01
9.74983931e-01 -1.51396513e-01 4.90550548e-02 5.71710765e-01
5.69197655e-01 -2.43540585e-01 -8.68497312e-01 -3.75229977e-02
7.49617023e-03 9.78499413e-01 -9.23534811e-01 -1.59027159e-01
-5.32683909e-01 1.13163662e+00 -2.55637795e-01 2.99254537e-01
-9.25777078e-01 2.08585367e-01 7.67606795e-01 -7.73529336e-02
7.20146418e-01 -4.88777131e-01 -1.91736475e-01 -1.60780025e+00
1.31764367e-01 -6.90044641e-01 5.46557367e-01 -9.53719378e-01
-1.24755824e+00 2.64535546e-01 4.31541242e-02 -1.83160019e+00
-4.54965681e-01 -6.66555643e-01 -7.00840354e-01 6.18992865e-01
-1.17878532e+00 -1.77952945e+00 -4.48574275e-01 8.57471645e-01
7.12252676e-01 -7.20419526e-01 7.38350630e-01 6.63728714e-01
-6.65572882e-01 7.80467808e-01 1.25699490e-02 4.61945713e-01
3.73488724e-01 -8.66705835e-01 7.47733176e-01 1.05672181e+00
1.79152768e-02 2.07587093e-01 4.20603245e-01 -7.87952960e-01
-1.65336239e+00 -1.06305408e+00 4.21272546e-01 -2.96976924e-01
3.66976142e-01 3.27302486e-01 -4.21438009e-01 6.00130975e-01
8.00916031e-02 -7.02008754e-02 5.01968145e-01 -3.65839213e-01
5.09972088e-02 -8.55434909e-02 -1.24109650e+00 5.44368804e-01
1.30209351e+00 -3.73866588e-01 -4.79511768e-01 -3.66165906e-01
-3.82674575e-01 -3.37263614e-01 -9.64273512e-01 3.81636977e-01
7.22023904e-01 -9.80590403e-01 1.52743495e+00 -2.66915709e-01
2.79818267e-01 -3.61351073e-01 -3.95043463e-01 -1.61193120e+00
-4.75753397e-01 1.59170672e-01 -2.97244847e-01 1.15261459e+00
-3.33658725e-01 -7.75523007e-01 1.01528132e+00 -2.93447133e-02
-7.03957006e-02 -5.39472640e-01 -1.26545608e+00 -9.17510450e-01
2.35296972e-02 -1.41657203e-01 3.22384387e-01 9.26695645e-01
-5.77031150e-02 2.90543884e-01 -7.65105784e-01 -3.21333408e-02
1.06001008e+00 1.28858119e-01 5.59033215e-01 -1.14513171e+00
1.64835259e-01 -3.22563648e-01 -1.26490355e+00 -7.05498576e-01
3.34846005e-02 -7.35909700e-01 -4.91837204e-01 -1.67543614e+00
9.78925750e-02 1.18573688e-01 -1.05276771e-01 1.78348362e-01
-1.97301731e-01 6.05947435e-01 4.28843856e-01 2.04367548e-01
-2.92896479e-01 8.99279177e-01 1.68466151e+00 -5.28316438e-01
2.82821745e-01 -5.87192290e-02 -3.00718457e-01 7.36012042e-01
9.82400298e-01 2.38617644e-01 -4.93037313e-01 -1.21616893e-01
-5.40171206e-01 3.80116254e-01 3.79436791e-01 -1.54574060e+00
2.45339632e-01 -3.32472124e-03 6.39694035e-01 -1.00499344e+00
8.12152505e-01 -5.71438730e-01 1.89030737e-01 7.38929033e-01
5.46037316e-01 5.70893347e-01 1.21527724e-01 5.76423347e-01
-4.07957554e-01 4.07538772e-01 6.57648444e-01 -2.56687135e-01
-1.41812789e+00 4.55519944e-01 -6.69701338e-01 2.29253113e-01
1.01428986e+00 -1.01630819e+00 -4.91768152e-01 -3.85895908e-01
-8.96383047e-01 1.87484279e-01 2.93720603e-01 6.59141958e-01
1.04007185e+00 -1.82285416e+00 -6.17227316e-01 4.18355584e-01
1.25260964e-01 -8.01654994e-01 4.14904714e-01 1.03752410e+00
-4.35420156e-01 5.26872039e-01 -8.45710576e-01 -8.52675617e-01
-1.54957151e+00 1.17856194e-03 5.60293436e-01 -4.91337001e-01
-9.21722770e-01 2.34989151e-01 -1.47807166e-01 -1.50633439e-01
-4.18444991e-01 6.21896191e-03 -7.46672869e-01 3.99805844e-01
1.58287227e-01 8.26106668e-01 7.88962618e-02 -1.38255107e+00
-5.88482141e-01 1.39651310e+00 1.60086975e-01 -9.97649133e-02
1.13032711e+00 -3.19843322e-01 3.02499175e-01 3.67026359e-01
1.03592718e+00 -5.49358428e-01 -9.65948284e-01 2.09012836e-01
-2.75057971e-01 -6.33688033e-01 -4.43331450e-02 -5.36771834e-01
-1.68555307e+00 1.05036199e+00 1.24238467e+00 -2.37003192e-01
1.22741067e+00 -5.57487249e-01 1.15345037e+00 -5.40255941e-02
8.83872807e-01 -1.10005450e+00 4.18187052e-01 6.22458518e-01
5.83258212e-01 -1.10266292e+00 -1.71068043e-01 -5.34583211e-01
-8.50236118e-01 1.16451538e+00 7.56231844e-01 -8.49420950e-02
6.03431642e-01 2.95712292e-01 9.11104232e-02 -3.03174615e-01
-2.51559764e-01 -3.58397037e-01 2.70924360e-01 1.21350062e+00
1.88839331e-01 1.46856979e-01 -1.57670081e-01 2.85551995e-01
-3.24748725e-01 2.87440181e-01 4.45616990e-01 7.88027823e-01
-1.90728128e-01 -8.18319499e-01 -4.58100408e-01 5.79255998e-01
-3.33312780e-01 2.26274624e-01 6.84188120e-03 7.80959666e-01
-5.49040101e-02 8.78043950e-01 -1.13679841e-01 -8.47585917e-01
3.87745887e-01 1.33387953e-01 8.32498372e-01 6.78501800e-02
-1.84990972e-01 -2.88023874e-02 3.39071095e-01 -5.51622629e-01
-7.98059285e-01 -8.94585788e-01 -9.59379196e-01 -5.99719226e-01
-1.13325596e-01 -7.02995136e-02 9.84332487e-02 1.01495600e+00
1.84398562e-01 6.79636538e-01 4.39153403e-01 -1.07875872e+00
-2.63705641e-01 -7.63006926e-01 -7.49745429e-01 2.32058704e-01
7.52779990e-02 -1.25930035e+00 -4.67290543e-03 2.14120343e-01] | [14.262801170349121, 1.4392391443252563] |
77dfae19-f61a-4ff5-aedb-c1726e68c376 | kosmos-2-grounding-multimodal-large-language | 2306.14824 | null | https://arxiv.org/abs/2306.14824v2 | https://arxiv.org/pdf/2306.14824v2.pdf | Kosmos-2: Grounding Multimodal Large Language Models to the World | We introduce Kosmos-2, a Multimodal Large Language Model (MLLM), enabling new capabilities of perceiving object descriptions (e.g., bounding boxes) and grounding text to the visual world. Specifically, we represent refer expressions as links in Markdown, i.e., ``[text span](bounding boxes)'', where object descriptions are sequences of location tokens. Together with multimodal corpora, we construct large-scale data of grounded image-text pairs (called GrIT) to train the model. In addition to the existing capabilities of MLLMs (e.g., perceiving general modalities, following instructions, and performing in-context learning), Kosmos-2 integrates the grounding capability into downstream applications. We evaluate Kosmos-2 on a wide range of tasks, including (i) multimodal grounding, such as referring expression comprehension, and phrase grounding, (ii) multimodal referring, such as referring expression generation, (iii) perception-language tasks, and (iv) language understanding and generation. This work lays out the foundation for the development of Embodiment AI and sheds light on the big convergence of language, multimodal perception, action, and world modeling, which is a key step toward artificial general intelligence. Data, demo, and pretrained models are available at https://aka.ms/kosmos-2. | ['Furu Wei', 'Shuming Ma', 'Shaohan Huang', 'Yaru Hao', 'Li Dong', 'Wenhui Wang', 'Zhiliang Peng'] | 2023-06-26 | null | null | null | null | ['referring-expression-generation', 'visual-grounding', 'referring-expression', 'image-captioning', 'phrase-grounding'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'natural-language-processing'] | [ 2.78550982e-01 2.17589185e-01 -1.56649977e-01 -3.40951145e-01
-8.17983091e-01 -8.25290024e-01 9.36388791e-01 1.98633775e-01
-2.45662734e-01 2.83709556e-01 8.02098095e-01 -4.06401038e-01
3.51818830e-01 -8.94153535e-01 -9.33589935e-01 -3.78600895e-01
1.46165833e-01 3.47601533e-01 -3.44919294e-01 -4.91272122e-01
1.87536284e-01 1.74346656e-01 -1.36220872e+00 8.03007543e-01
7.25416541e-01 1.10295510e+00 3.81351292e-01 7.08417416e-01
-5.18040776e-01 1.15387201e+00 -6.06835365e-01 -6.26006246e-01
-4.91341203e-01 -4.78701711e-01 -8.05464804e-01 -1.36721358e-02
4.87401724e-01 -4.86646712e-01 -3.28314155e-01 8.74894798e-01
3.07914972e-01 3.05901796e-01 4.91303653e-01 -1.50550723e+00
-1.61336601e+00 9.61329103e-01 -2.07157224e-01 -3.41916621e-01
9.74195659e-01 4.31813091e-01 9.88921702e-01 -1.06321919e+00
7.55159914e-01 1.87177980e+00 1.59235030e-01 8.93122971e-01
-1.21039474e+00 -3.31285536e-01 2.96720177e-01 -2.22912314e-03
-1.35495579e+00 -5.84649205e-01 3.99759382e-01 -4.40820366e-01
1.19519424e+00 3.41197520e-01 5.97016335e-01 1.51375353e+00
4.17190716e-02 1.20836711e+00 9.26447988e-01 -4.14306521e-01
2.20590889e-01 -1.13695845e-01 -5.88677488e-02 7.95438826e-01
-3.95078480e-01 9.14081652e-03 -1.04558313e+00 2.44256854e-01
8.94921839e-01 -1.55101478e-01 -1.07014805e-01 6.28399700e-02
-1.97411764e+00 6.67004645e-01 6.84472144e-01 1.62092626e-01
-2.88729399e-01 6.65631235e-01 3.51527661e-01 -4.46470976e-02
2.10041761e-01 4.20553774e-01 1.00172587e-01 -2.46759847e-01
-3.68351698e-01 2.65489489e-01 4.19227868e-01 1.36905444e+00
7.03477919e-01 1.02726132e-01 -4.21200454e-01 8.03950012e-01
4.86167461e-01 1.06952381e+00 2.50981361e-01 -1.10972691e+00
7.66784787e-01 7.27767885e-01 6.20395131e-02 -8.54781151e-01
-4.32859898e-01 3.40530336e-01 -6.63112223e-01 -2.84220666e-01
2.06953436e-01 -1.74089313e-01 -9.27610517e-01 2.32897496e+00
-1.24355674e-01 -6.47974983e-02 4.46058810e-01 9.99854982e-01
1.49156713e+00 1.12064075e+00 7.12033689e-01 1.98767319e-01
1.63443244e+00 -8.00096869e-01 -8.53373647e-01 -4.27668244e-01
9.32964683e-01 -3.88344139e-01 1.61117029e+00 7.01572895e-02
-1.25650072e+00 -7.18056858e-01 -6.99274004e-01 -6.59204125e-01
-9.68039930e-01 1.07046001e-01 8.72514904e-01 7.74844810e-02
-1.26980245e+00 -1.11870758e-01 -6.09327257e-01 -7.42764711e-01
4.19543236e-01 1.84443295e-02 -7.67195284e-01 -2.18609422e-01
-1.17461121e+00 8.71089518e-01 5.38995624e-01 2.91817009e-01
-1.02889812e+00 -4.17173386e-01 -1.55973887e+00 -2.10941076e-01
3.45853120e-01 -1.04602301e+00 1.25135159e+00 -9.58140314e-01
-1.21009731e+00 1.25232577e+00 -3.48135144e-01 -2.79259801e-01
1.78863946e-02 -4.81421292e-01 -5.61567128e-01 2.32472867e-01
5.60027845e-02 1.47657454e+00 5.17274976e-01 -1.58553553e+00
-2.92288631e-01 -3.47417742e-01 5.28258741e-01 4.06568378e-01
-1.42225757e-01 1.52027443e-01 -4.71439600e-01 -5.25689960e-01
6.34227172e-02 -6.55627310e-01 1.42601788e-01 -1.20829372e-03
-8.20237756e-01 -3.99730235e-01 5.25180399e-01 -6.92504287e-01
1.10373962e+00 -2.32991433e+00 6.53444111e-01 -1.54981717e-01
3.16970199e-01 -2.27756307e-01 -6.29218936e-01 9.90723372e-01
1.48620671e-02 5.39816201e-01 -6.99678287e-02 -6.10700786e-01
5.68979323e-01 4.07117188e-01 -7.94255674e-01 3.91476080e-02
4.32479531e-01 1.72742093e+00 -1.01770258e+00 -5.30250549e-01
3.72089446e-01 3.95840019e-01 -3.63057584e-01 4.13027823e-01
-7.25253999e-01 5.30326366e-01 -3.59394938e-01 8.65744352e-01
2.08605826e-01 -1.13624558e-01 -1.28963187e-01 -4.16418701e-01
-2.01558322e-01 1.35741040e-01 -7.01105118e-01 2.26534986e+00
-7.38549113e-01 7.82306850e-01 -1.63045645e-01 -5.13571084e-01
5.97322702e-01 3.11463267e-01 -1.22913800e-01 -9.38221574e-01
4.54161391e-02 -1.84840783e-01 -2.28322729e-01 -8.72701228e-01
7.41250336e-01 -2.98034042e-01 -6.41231835e-01 4.44316238e-01
1.88655302e-01 -5.34150243e-01 3.12389135e-01 7.19182670e-01
5.53553879e-01 5.58992624e-01 3.33620757e-01 1.75574213e-01
1.29191279e-01 2.92462874e-02 -1.70206934e-01 6.23131812e-01
1.53428078e-01 3.99854034e-01 5.98206937e-01 -1.19920976e-01
-7.08308637e-01 -1.57695246e+00 1.92339778e-01 1.58265162e+00
3.11496198e-01 -6.94972456e-01 -8.75982642e-01 -3.04386690e-02
-6.72891503e-03 1.15962982e+00 -7.42167592e-01 -3.01092923e-01
-4.33785617e-01 -1.65181607e-01 8.53735030e-01 9.46065128e-01
5.42824805e-01 -1.53838563e+00 -6.61042452e-01 -1.17613241e-01
-5.37259161e-01 -1.35605764e+00 -3.94575000e-01 -1.93721220e-01
-5.46009064e-01 -5.64243555e-01 -4.36116606e-01 -7.61136174e-01
5.70140421e-01 2.65572090e-02 1.57925892e+00 5.30388206e-02
4.35662232e-02 9.30657566e-01 -4.94165838e-01 -4.47845578e-01
-3.45966130e-01 -4.16940331e-01 -9.39747915e-02 -2.28417590e-01
2.17816830e-01 -1.80418387e-01 -3.26417118e-01 1.34488165e-01
-1.24091113e+00 6.35429442e-01 6.13263190e-01 5.03158331e-01
5.18012166e-01 -7.99394429e-01 4.77875501e-01 -4.09676105e-01
8.84872794e-01 -5.60720325e-01 -1.88934013e-01 5.09506941e-01
3.35051537e-01 -2.23624438e-01 2.76676685e-01 -6.25712991e-01
-1.15029728e+00 -5.17974675e-01 -9.72652435e-02 -2.01296821e-01
-5.13172328e-01 8.22030604e-01 -5.42980969e-01 4.07223970e-01
5.59625804e-01 2.96441555e-01 -2.41195232e-01 -1.07859127e-01
1.38122976e+00 5.71330845e-01 9.91374612e-01 -1.07574975e+00
5.72328985e-01 4.08875555e-01 -1.09293126e-01 -8.28780532e-01
-6.98189616e-01 -1.19644761e-01 -6.34800076e-01 -2.57903844e-01
1.45608652e+00 -9.47256029e-01 -1.06433535e+00 2.56324053e-01
-1.33338702e+00 -8.06976020e-01 -3.46324444e-01 5.87784424e-02
-1.00247538e+00 7.68161863e-02 -7.67177343e-01 -8.35370243e-01
-9.72928107e-02 -8.87587905e-01 1.67272353e+00 3.39840680e-01
-6.07137442e-01 -1.17423785e+00 -2.83749282e-01 4.68807310e-01
2.14233920e-01 5.01412272e-01 1.26518083e+00 -2.57408589e-01
-6.74011290e-01 7.30571374e-02 -4.42584068e-01 2.81153154e-02
-2.10947424e-01 -4.66868393e-02 -8.67151737e-01 1.35883421e-01
-6.63908005e-01 -1.07547736e+00 6.25733078e-01 3.20739187e-02
1.12835288e+00 -3.30723256e-01 -3.29479426e-01 7.01023936e-01
1.16468155e+00 3.44153255e-01 8.70404065e-01 -1.17550241e-02
9.12674189e-01 7.96379566e-01 6.53581023e-01 3.44025463e-01
8.57677281e-01 5.35953760e-01 6.02024078e-01 -3.08676630e-01
-1.69135094e-01 -8.12235773e-01 5.39085388e-01 5.99122405e-01
-1.00599378e-01 -7.16329813e-01 -1.15902472e+00 3.87312293e-01
-1.80913472e+00 -1.10365999e+00 2.49988034e-01 1.65440702e+00
7.65874565e-01 -4.48419005e-01 -1.12649418e-01 -5.59098065e-01
3.04086357e-01 3.43578845e-01 -5.07336199e-01 -5.82852244e-01
-5.41562200e-01 -2.32943892e-02 -3.93942952e-01 4.60160911e-01
-8.75586450e-01 1.53472579e+00 5.97608423e+00 5.09452462e-01
-1.03010476e+00 1.24857901e-03 3.41585964e-01 5.25809405e-03
-6.31750703e-01 -4.38497141e-02 -4.13201064e-01 1.51716411e-01
8.67059708e-01 7.35582635e-02 8.02340627e-01 3.65756005e-01
2.20828757e-01 -2.28706658e-01 -1.54651344e+00 1.30792344e+00
4.20579612e-01 -1.26783717e+00 7.86838651e-01 -2.38095805e-01
3.69773269e-01 -2.10133716e-01 2.74824500e-01 5.33977151e-01
3.02165002e-01 -1.44479060e+00 1.32867539e+00 7.80043066e-01
1.14574242e+00 -2.94618517e-01 4.01458949e-01 2.62175292e-01
-1.16645324e+00 -6.83956817e-02 -4.83690314e-02 -1.95740256e-02
6.54689431e-01 -2.43117034e-01 -2.23903060e-01 5.84420979e-01
4.82685328e-01 7.93334246e-01 -4.26759154e-01 2.96433598e-01
-5.88677287e-01 2.50610173e-01 -4.26320769e-02 -1.20074086e-01
2.31039390e-01 -2.35413745e-01 3.81306410e-01 1.38004339e+00
3.95346552e-01 4.54651415e-01 7.20945150e-02 1.51269400e+00
-2.22418725e-01 3.42085287e-02 -8.35385442e-01 -7.02927768e-01
4.84547645e-01 1.05893290e+00 -3.56755316e-01 -5.01618803e-01
-4.64845985e-01 1.08768833e+00 3.11387599e-01 9.10125732e-01
-8.54045272e-01 -1.26408041e-01 5.66295445e-01 -1.22113086e-01
-2.86358625e-01 -5.67560494e-01 -1.71638653e-01 -1.13114154e+00
-2.98884124e-01 -9.23980415e-01 2.56884009e-01 -1.85070670e+00
-1.23619151e+00 5.11290431e-01 4.79989946e-01 -8.12565327e-01
-4.32994187e-01 -9.65152562e-01 -5.75301647e-01 9.03793275e-01
-9.04674470e-01 -1.93195450e+00 -5.47046483e-01 7.49079704e-01
4.96066540e-01 1.20268449e-01 1.06662524e+00 -1.20219864e-01
-4.79498893e-01 2.23751292e-01 -7.04669237e-01 1.98250696e-01
5.86787879e-01 -1.23780525e+00 2.99915552e-01 6.10328615e-01
4.08020347e-01 1.25306046e+00 4.14170295e-01 -5.72201848e-01
-1.74133229e+00 -7.79417694e-01 6.92243338e-01 -8.65022779e-01
9.31479335e-01 -7.69711912e-01 -7.21796393e-01 1.21564746e+00
6.19503140e-01 -3.83024782e-01 9.88675117e-01 4.01851274e-02
-4.73070323e-01 2.99729586e-01 -7.50601113e-01 1.13582623e+00
1.43911242e+00 -1.01423180e+00 -8.19028676e-01 3.70867461e-01
1.05865026e+00 -6.22276068e-01 -9.05253768e-01 7.44629577e-02
5.23883879e-01 -7.10987449e-01 1.04786170e+00 -1.02636933e+00
8.24139655e-01 -2.32713997e-01 -8.14094782e-01 -1.31300175e+00
4.44983989e-02 -6.33668244e-01 -2.26566866e-01 1.42903411e+00
4.61027682e-01 -4.61183280e-01 2.57565677e-02 7.12777317e-01
-4.97863442e-01 -6.07829034e-01 -8.21364641e-01 -3.89584094e-01
1.85282484e-01 -8.16986740e-01 6.49245918e-01 7.52162457e-01
4.39226210e-01 4.04267311e-01 -1.33592874e-01 8.09861571e-02
1.87864512e-01 1.07560970e-01 8.42096925e-01 -6.37829304e-01
2.74810214e-02 -4.75628197e-01 -2.01676115e-01 -1.34083593e+00
5.11528134e-01 -9.36460376e-01 1.08836092e-01 -2.06493855e+00
1.85820371e-01 -3.60313654e-02 1.96465813e-02 8.78050923e-01
-2.66037490e-02 2.21631423e-01 6.24600053e-01 9.92237963e-03
-9.48869526e-01 5.81974626e-01 1.44859958e+00 -2.44184002e-01
-2.53385250e-02 -8.10961545e-01 -9.26099479e-01 7.43962288e-01
6.53075457e-01 2.87757993e-01 -5.09614289e-01 -9.12646234e-01
6.61328197e-01 1.03600740e-01 8.99008512e-01 -4.95218635e-01
1.38807856e-02 -4.07668591e-01 2.80846745e-01 -5.68234742e-01
7.40516722e-01 -5.14340818e-01 -1.94896877e-01 -1.23923235e-01
-5.43052316e-01 4.20270115e-01 5.44181883e-01 2.01901406e-01
-3.10545444e-01 9.33200046e-02 1.44785061e-01 -2.21810192e-01
-1.35032690e+00 -3.02045140e-02 -3.11822712e-01 8.04462880e-02
8.24129879e-01 -1.87208414e-01 -1.02986503e+00 -7.51580119e-01
-8.57302189e-01 4.20505911e-01 5.64087689e-01 7.19548821e-01
1.01455534e+00 -1.50439930e+00 -3.68892342e-01 -8.28827918e-02
6.87395632e-01 5.96514996e-03 3.27023357e-01 6.20443463e-01
-2.31927380e-01 4.83186632e-01 -1.00355022e-01 -6.78984225e-01
-9.04391110e-01 7.91257501e-01 1.44599333e-01 4.72111940e-01
-3.50334495e-01 9.27671194e-01 6.57285571e-01 -3.35516989e-01
9.64789987e-02 -6.49326980e-01 -1.29740909e-01 -8.34599510e-03
6.58329129e-01 -3.96301374e-02 -7.57828355e-01 -9.79747415e-01
-2.53072560e-01 8.70462000e-01 4.64672238e-01 -6.00152612e-01
7.63397455e-01 -3.67371649e-01 -5.36869764e-01 9.44290400e-01
9.06081975e-01 -8.61497521e-02 -9.75098073e-01 -1.73986349e-02
-1.82890058e-01 -6.60922304e-02 -3.30639899e-01 -1.17531288e+00
-4.23504353e-01 1.26064289e+00 2.42571369e-01 -5.25234640e-03
1.13621533e+00 7.57997394e-01 6.52948618e-01 5.29426098e-01
4.14554596e-01 -7.36637831e-01 5.38565576e-01 5.43257535e-01
1.56326830e+00 -1.27964747e+00 -4.46085125e-01 -2.70033687e-01
-9.77625191e-01 9.60468888e-01 8.26599836e-01 4.67975408e-01
-9.31112990e-02 1.85144678e-01 2.78301388e-01 -3.31075281e-01
-8.35211277e-01 -5.33346891e-01 4.04920459e-01 9.65364039e-01
5.17031431e-01 3.32137644e-01 3.74328554e-01 6.19362295e-01
-2.91786969e-01 -1.25640884e-01 4.54949774e-02 7.31083691e-01
-2.42184862e-01 -6.98691726e-01 -4.81456161e-01 -4.04925570e-02
6.23165406e-02 -4.60517794e-01 -7.26339638e-01 1.00732636e+00
4.54502970e-01 1.17281687e+00 3.72811496e-01 -2.94962615e-01
5.06707549e-01 1.77232116e-01 5.90033114e-01 -7.16274321e-01
-2.51303524e-01 -2.20447958e-01 3.32313836e-01 -7.72086561e-01
-4.95409608e-01 -2.44814321e-01 -1.72621870e+00 -3.01465780e-01
2.81433851e-01 -2.96820670e-01 5.33189297e-01 9.54194069e-01
2.69957006e-01 5.03447592e-01 -2.59835899e-01 -8.96169424e-01
-1.10053243e-02 -1.13504231e+00 -2.66275018e-01 7.81836212e-01
2.31902018e-01 -4.90371495e-01 -2.43279506e-02 3.79086792e-01] | [10.773225784301758, 1.6455872058868408] |
6424a281-bdd9-402b-bc4f-c0f1451d04e1 | segdensenet-iris-segmentation-for-pre-and | 1801.10100 | null | http://arxiv.org/abs/1801.10100v2 | http://arxiv.org/pdf/1801.10100v2.pdf | SegDenseNet: Iris Segmentation for Pre and Post Cataract Surgery | Cataract is caused due to various factors such as age, trauma, genetics,
smoking and substance consumption, and radiation. It is one of the major common
ophthalmic diseases worldwide which can potentially affect iris-based biometric
systems. India, which hosts the largest biometrics project in the world, has
about 8 million people undergoing cataract surgery annually. While existing
research shows that cataract does not have a major impact on iris recognition,
our observations suggest that the iris segmentation approaches are not well
equipped to handle cataract or post cataract surgery cases. Therefore, failure
in iris segmentation affects the overall recognition performance. This paper
presents an efficient iris segmentation algorithm with variations due to
cataract and post cataract surgery. The proposed algorithm, termed as
SegDenseNet, is a deep learning algorithm based on DenseNets. The experiments
on the IIITD Cataract database show that improving the segmentation enhances
the identification by up to 25% across different sensors and matchers. | ['Mayank Vatsa', 'Rohit Keshari', 'Pavani Tripathi', 'Richa Singh', 'Aditya Lakra'] | 2018-01-30 | null | null | null | null | ['iris-segmentation'] | ['medical'] | [-1.55819237e-01 -2.28974134e-01 -1.07282229e-01 -1.81771070e-01
2.38493811e-02 -3.40996653e-01 1.70432478e-01 -9.74072292e-02
-5.24093568e-01 4.87499148e-01 3.70710939e-01 -3.93565029e-01
-1.28593773e-01 -4.67546642e-01 -4.15004462e-01 -6.25719666e-01
4.96070832e-02 4.07488555e-01 -3.00234199e-01 3.39563221e-01
4.87144023e-01 7.88068831e-01 -1.57840753e+00 -3.93337816e-01
1.19961894e+00 7.44629383e-01 -3.53984058e-01 5.82598925e-01
4.09787744e-02 -5.56929745e-02 -6.84066474e-01 -4.47389454e-01
7.50217140e-01 -4.20585543e-01 -5.30369103e-01 1.91310763e-01
9.54859436e-01 -6.48870349e-01 -1.47958333e-02 1.24451256e+00
7.71439850e-01 -2.83120006e-01 4.21712488e-01 -5.05706131e-01
-4.18915033e-01 3.92859697e-01 -8.32282126e-01 1.68677509e-01
2.25631133e-01 2.25838423e-01 1.06528163e-01 -5.93596041e-01
1.71178609e-01 1.04631054e+00 7.34080017e-01 7.17024505e-01
-9.21052873e-01 -7.97184527e-01 -5.68445683e-01 -3.20844501e-01
-1.20017636e+00 -3.47857296e-01 8.79312232e-02 -4.14511353e-01
5.42963386e-01 3.47787172e-01 1.12760270e+00 2.92628676e-01
2.68184453e-01 3.17546755e-01 1.71095085e+00 -4.90991741e-01
-1.46526396e-01 -5.07820249e-02 -5.52819893e-02 8.89930904e-01
6.96226358e-01 7.18822539e-01 -1.96884900e-01 4.65052240e-02
1.06496072e+00 6.27298728e-02 -2.02709720e-01 -4.62642424e-02
-1.14313149e+00 4.19935346e-01 4.83197659e-01 1.57451257e-01
-1.58581302e-01 -2.00069904e-01 1.33365229e-01 1.50955573e-01
-5.26647083e-03 7.92130113e-01 -3.38942677e-01 -2.44032145e-01
-1.07995844e+00 -1.42142639e-01 6.67765141e-01 5.88109374e-01
4.13025290e-01 -1.11158742e-02 -2.95124259e-02 6.83362305e-01
5.73577881e-01 7.22724855e-01 4.04250026e-01 -5.64479887e-01
-2.03852922e-01 9.21612322e-01 -6.73367921e-03 -7.40009725e-01
-4.17279422e-01 -6.32714689e-01 -9.33230042e-01 4.19461191e-01
7.24147856e-01 -5.28977752e-01 -1.73898470e+00 9.32120979e-01
1.08487830e-01 4.97463912e-01 -2.00051188e-01 1.14830375e+00
1.04668784e+00 -1.78034946e-01 -1.96577124e-02 -1.07547455e-01
1.34119689e+00 -5.82847714e-01 -6.69468641e-01 -2.87709385e-02
1.43388331e-01 -1.27000976e+00 7.43222952e-01 3.28149348e-01
-1.00392938e+00 -3.26043308e-01 -8.72512877e-01 7.32556656e-02
-4.53292608e-01 3.29842567e-01 8.09332788e-01 1.32436264e+00
-9.85399902e-01 1.90806866e-01 -8.49483073e-01 -9.76415575e-01
5.73271930e-01 1.06991065e+00 -1.81919456e-01 -3.37938927e-02
-4.12304223e-01 8.77707720e-01 1.84438992e-02 2.16175303e-01
-3.42791289e-01 -5.57037294e-01 -5.54862916e-01 -3.71152163e-01
9.56113860e-02 -6.91261351e-01 7.63798952e-01 -7.96073556e-01
-1.47474658e+00 1.07387888e+00 -4.77380604e-01 -3.86982322e-01
2.56856829e-01 -2.20519975e-01 -4.37444180e-01 -2.21669644e-01
-3.28165799e-01 6.08546793e-01 5.49607933e-01 -8.21564078e-01
-6.74292207e-01 -6.68876350e-01 -1.59144476e-01 1.30185038e-01
1.08250260e-01 4.29083407e-01 -6.38681114e-01 -4.73445952e-01
2.72255749e-01 -1.19298697e+00 -3.38782310e-01 -1.63995057e-01
-6.92366481e-01 -1.10858738e-01 5.07179379e-01 -7.48070717e-01
9.40522432e-01 -2.00638270e+00 -1.70952648e-01 6.32442296e-01
4.11808729e-01 6.16673648e-01 2.27469310e-01 -1.10812724e-01
8.79449621e-02 5.30819774e-01 4.24465276e-02 -9.49776471e-02
-5.11435032e-01 2.89648622e-01 3.86470973e-01 7.63221443e-01
-2.90601611e-01 5.95404029e-01 -5.95593631e-01 -2.71534234e-01
5.52484930e-01 4.01049167e-01 -2.41413623e-01 -1.30610261e-02
3.06048840e-01 6.10257924e-01 -8.74618664e-02 1.40340805e+00
8.73951554e-01 -2.49082744e-02 -2.45045647e-01 -1.88694060e-01
-3.72002095e-01 -1.86133653e-01 -1.17363751e+00 1.44550180e+00
3.40846300e-01 6.54937565e-01 -2.14822263e-01 -4.14108783e-01
6.56476438e-01 3.80111039e-01 3.14353436e-01 -6.08284950e-01
4.06794727e-01 5.80195427e-01 5.87733090e-01 -5.03716648e-01
2.83333004e-01 -3.17445882e-02 4.95894849e-01 9.15523097e-02
-1.98181987e-01 2.57530898e-01 1.39373198e-01 -2.16436163e-01
5.65870464e-01 -1.73034817e-01 1.44605860e-01 -3.29400241e-01
3.51628959e-01 -7.60617405e-02 8.53151560e-01 6.77755356e-01
-4.55354422e-01 6.75049961e-01 1.93552688e-01 -6.60607278e-01
-1.01802623e+00 -8.53916883e-01 -7.19557285e-01 -3.54261771e-02
3.45579773e-01 2.54346747e-02 -8.60679328e-01 -3.53898108e-01
2.09644467e-01 -2.33564407e-01 -4.00456607e-01 2.23321274e-01
-2.01495498e-01 -1.02705395e+00 6.24994636e-01 2.94702530e-01
9.75442350e-01 -5.71291149e-01 -3.33123386e-01 -1.88639641e-01
2.15574756e-01 -5.73879063e-01 -4.86166030e-01 -4.98488545e-01
-1.03228307e+00 -1.53694808e+00 -8.36339891e-01 -8.19880068e-01
1.23342395e+00 -1.19162537e-01 7.65974343e-01 5.61812043e-01
-8.26554537e-01 2.39175841e-01 6.22202903e-02 -9.00590479e-01
8.68998561e-03 -1.47653252e-01 3.28497142e-01 6.70043752e-02
1.10313654e+00 -7.48381838e-02 -9.24967468e-01 4.37940545e-02
-3.98985267e-01 -3.40236753e-01 8.68296385e-01 6.51615262e-01
5.51635683e-01 1.15577467e-01 -1.60488904e-01 -9.22249198e-01
4.91855979e-01 -2.54776441e-02 -8.88881981e-01 3.82085554e-02
-9.36641335e-01 -2.64843851e-01 -3.17969054e-01 -4.15404849e-02
-6.58263922e-01 2.97733873e-01 2.61707544e-01 -1.40097573e-01
-6.34129703e-01 4.05796558e-01 1.40124172e-01 -8.50384474e-01
5.37645996e-01 -1.22209005e-01 2.56572127e-01 -6.21687412e-01
-3.36312979e-01 8.78626227e-01 5.54020345e-01 -1.41029924e-01
9.88316953e-01 3.61014187e-01 2.41819158e-01 -9.72832859e-01
-4.44258571e-01 -6.82726383e-01 -5.93111098e-01 -2.55326331e-01
1.15480793e+00 -8.71651828e-01 -1.06391144e+00 1.17374730e+00
-7.89887905e-01 1.44575417e-01 2.05648139e-01 1.01237524e+00
2.52826512e-01 2.44541600e-01 -4.72419113e-01 -7.74419427e-01
-2.47631952e-01 -1.37328613e+00 5.72052300e-01 1.19005513e+00
3.40869501e-02 -8.80163312e-01 -2.76523411e-01 5.35590768e-01
4.44710225e-01 3.46768558e-01 6.84707999e-01 -2.93885946e-01
-1.00085402e+00 -4.19727452e-02 -3.52019250e-01 3.48178178e-01
5.88043034e-01 3.24551642e-01 -7.68339574e-01 -5.27703881e-01
-4.27902639e-01 1.62994459e-01 6.63957953e-01 8.44614267e-01
9.57933605e-01 5.77181727e-02 -4.73735541e-01 9.88759696e-01
1.85682440e+00 4.95266944e-01 1.08039880e+00 3.47505838e-01
6.82254374e-01 1.58912182e-01 2.27531180e-01 1.12486273e-01
1.71988666e-01 1.53638989e-01 3.58410954e-01 -9.16047633e-01
-3.56093615e-01 1.22282915e-01 -3.68464857e-01 5.28234065e-01
-7.67250240e-01 9.66110677e-02 -1.36374843e+00 7.15666890e-01
-1.11398697e+00 -6.85295284e-01 -3.73115420e-01 2.76764274e+00
8.80063653e-01 -3.33754689e-01 7.37402290e-02 -1.50012910e-01
5.92906475e-01 -6.42106295e-01 -6.19423270e-01 -4.32386249e-01
-7.31162205e-02 7.75893927e-01 1.06902254e+00 3.94562989e-01
-1.07379186e+00 9.08005655e-01 6.91009092e+00 -2.20042750e-01
-1.30247974e+00 -4.42380339e-01 5.69271147e-01 -1.00986212e-01
3.17359060e-01 1.07422628e-01 -8.34888160e-01 6.19876027e-01
6.25005901e-01 -1.53145660e-02 4.89880085e-01 7.89154917e-02
2.03261197e-01 -6.14669085e-01 -7.65483081e-01 1.10605979e+00
2.52834290e-01 -1.05119288e+00 -5.34108393e-02 5.79013646e-01
8.86501551e-01 1.78652242e-01 2.63202220e-01 -3.79124135e-01
1.31959245e-01 -1.47647643e+00 -2.44747207e-01 9.34614003e-01
9.49164569e-01 -8.41606855e-01 1.22958136e+00 -3.01681548e-01
-6.88508928e-01 1.12714902e-01 -6.99304640e-02 -9.81128886e-02
-2.60902435e-01 4.78729337e-01 -8.95691991e-01 8.34425390e-02
8.99087548e-01 4.98785466e-01 -8.71150613e-01 2.19224763e+00
-8.36956054e-02 8.72793674e-01 -3.74356210e-01 2.54276365e-01
-1.10107467e-01 -4.38751042e-01 7.81138957e-01 7.41285324e-01
3.53633940e-01 2.06619561e-01 4.98797297e-02 5.40256083e-01
-1.47901207e-01 4.46678884e-02 -4.45446074e-01 -1.07634500e-01
4.72632229e-01 6.93841577e-01 -7.21116722e-01 7.53075629e-02
-8.28338802e-01 6.48857772e-01 -4.88184184e-01 5.34703672e-01
-1.62738949e-01 -5.14088690e-01 1.00624096e+00 2.08552212e-01
-7.13377073e-02 -1.02799185e-01 -6.90716863e-01 -9.15224016e-01
-1.10090956e-01 -1.20445585e+00 -1.00367710e-01 -2.52836734e-01
-9.36348021e-01 5.85566908e-02 -6.40867412e-01 -1.15159094e+00
2.35579401e-01 -6.44181252e-01 -4.59132761e-01 1.42415476e+00
-1.55170584e+00 -1.30578411e+00 -3.82223576e-01 6.21645153e-01
6.03892840e-02 -5.48453510e-01 8.85187864e-01 2.64192075e-01
-8.37283254e-01 6.63614750e-01 -7.36072138e-02 6.62888169e-01
9.38724518e-01 -1.55267632e+00 8.16731304e-02 1.20401597e+00
-1.91456273e-01 1.24185884e+00 5.06542146e-01 -8.40263247e-01
-1.45531464e+00 -7.12046385e-01 9.80784595e-01 -3.12620848e-01
-2.23278757e-02 1.71359926e-01 -4.43404913e-01 6.60204589e-01
4.64091331e-01 -1.93717584e-01 1.10567236e+00 5.51737309e-01
1.29298583e-01 -2.36420885e-01 -1.41430831e+00 6.09841347e-01
6.56425595e-01 -4.22754079e-01 -4.42494959e-01 2.41819203e-01
-7.77833164e-02 -1.13262844e+00 -1.01702046e+00 5.43370247e-01
6.76643074e-01 -1.11000454e+00 6.75497234e-01 -3.63309443e-01
1.63626954e-01 -5.26541352e-01 4.54240531e-01 -9.60243106e-01
-1.29636675e-01 -6.57432258e-01 1.31205708e-01 1.12251246e+00
3.50434303e-01 -9.10147429e-01 8.72715354e-01 1.08333778e+00
7.67336041e-02 -4.63429421e-01 -6.98221982e-01 -4.22573656e-01
-1.50472790e-01 1.96399793e-01 7.58894682e-01 9.76741374e-01
-4.57662731e-01 -3.65764707e-01 -1.74741283e-01 6.77283525e-01
9.30054069e-01 5.48774526e-02 7.54403591e-01 -1.74250901e+00
1.77040145e-01 -3.38764846e-01 -9.30450201e-01 -6.36682868e-01
-4.76328373e-01 -3.60098988e-01 -3.59922230e-01 -1.66827846e+00
2.16073602e-01 -5.61016500e-01 -3.30685884e-01 6.18065238e-01
-4.04936485e-02 6.06407106e-01 -1.35614902e-01 7.48445392e-02
4.26198572e-01 -4.80872601e-01 1.68051064e+00 -1.15388028e-01
-3.55034739e-01 2.17880622e-01 -7.50317991e-01 6.14377975e-01
9.00997818e-01 -5.85107245e-02 -1.12002097e-01 -5.44165850e-01
6.56354055e-02 -3.49997610e-01 2.72062093e-01 -1.25743330e+00
5.38282812e-01 -2.11276906e-03 8.80406797e-01 -4.64834332e-01
-1.86784379e-02 -8.48092914e-01 6.32049292e-02 5.88821471e-01
3.31226349e-01 -3.08077216e-01 5.14456093e-01 2.86718518e-01
-3.11941206e-01 2.63428122e-01 9.39126790e-01 -1.68162718e-01
-5.43385506e-01 5.49804628e-01 -2.58448213e-01 -2.99582124e-01
8.34943354e-01 -7.89579391e-01 -2.59998024e-01 1.18392527e-01
-6.88504696e-01 2.49239370e-01 7.67664850e-01 3.80684018e-01
4.44874913e-01 -8.95986319e-01 -7.00414956e-01 7.55278289e-01
-5.40817380e-02 -4.49507572e-02 -9.76946801e-02 1.10394239e+00
-8.04997921e-01 4.99600977e-01 -1.71801776e-01 -5.66784620e-01
-1.80937457e+00 -1.37043726e-02 7.79747963e-01 3.57313812e-01
-5.57679236e-01 1.12044752e+00 -5.52596450e-01 -1.87325403e-01
6.08223915e-01 -5.94886482e-01 -3.92494410e-01 -2.59097040e-01
4.05423194e-01 3.66139382e-01 6.18678378e-03 -4.74853605e-01
-1.75680488e-01 1.01519954e+00 -4.20156687e-01 2.91301042e-01
7.30111063e-01 -2.54038982e-02 -6.18463039e-01 -4.22312580e-02
4.15451586e-01 2.56946653e-01 -6.43463612e-01 -4.33436520e-02
-9.76931453e-02 -9.11539555e-01 4.61111635e-01 -1.49305570e+00
-1.38000405e+00 5.10542095e-01 1.40726209e+00 -2.30837643e-01
1.25252795e+00 -2.94764400e-01 8.26013267e-01 5.81909716e-02
4.11536992e-01 -9.83185291e-01 -7.34630167e-01 2.60871470e-01
3.89656156e-01 -1.46557999e+00 1.03404604e-01 -2.49042079e-01
-2.01635845e-02 1.05450511e+00 6.78811073e-01 1.08684480e-01
8.76463115e-01 3.35562766e-01 6.60124004e-01 -2.49126285e-01
3.44992816e-01 -4.00570482e-01 6.32429123e-01 6.63871944e-01
7.18411326e-01 3.50883156e-01 -6.34913027e-01 -1.72973067e-01
-2.88030356e-01 2.48301819e-01 6.14398956e-01 6.30505383e-01
-1.96337134e-01 -1.50779545e+00 -4.48581964e-01 1.20532584e+00
-7.93534100e-01 -9.49288532e-02 -5.26074886e-01 6.16439223e-01
6.71841264e-01 9.02461290e-01 2.88453996e-01 -1.83531165e-01
1.86794996e-01 -1.19407862e-01 4.53547239e-01 -6.75477326e-01
-8.40942800e-01 1.63324699e-01 -1.17062733e-01 -6.03765965e-01
-9.60239172e-01 -7.62359977e-01 -1.14154673e+00 -6.47573173e-01
-3.20320219e-01 -2.06898734e-01 8.69512618e-01 9.87322211e-01
3.66908997e-01 3.43222529e-01 1.24301657e-01 -1.74211159e-01
1.40501484e-01 -9.67031181e-01 -1.04588425e+00 -6.27086312e-02
7.83003151e-01 -4.94911641e-01 -2.98047751e-01 1.11209534e-01] | [3.74393367767334, -3.6304168701171875] |
fd02f543-43e0-43ca-87f6-f20ebb5d367b | raid-g-robust-estimation-of-approximate | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Wang_RAID-G_Robust_Estimation_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Wang_RAID-G_Robust_Estimation_CVPR_2016_paper.pdf | RAID-G: Robust Estimation of Approximate Infinite Dimensional Gaussian With Application to Material Recognition | Infinite dimensional covariance descriptors can provide richer and more discriminative information than their low dimensional counterparts. In this paper, we propose a novel image descriptor, namely, robust approximate infinite dimensional Gaussian (RAID-G). The challenges of RAID-G mainly lie on two aspects: (1) description of infinite dimensional Gaussian is difficult due to its non-linear Riemannian geometric structure and the infinite dimensional setting, hence effective approximation is necessary; (2) traditional maximum likelihood estimation (MLE) is not robust to high (even infinite) dimensional covariance matrix in Gaussian setting. To address these challenges, explicit feature mapping (EFM) is first introduced for effective approximation of infinite dimensional Gaussian induced by additive kernel function, and then a new regularized MLE method based on von Neumann divergence is proposed for robust estimation of covariance matrix. The EFM and proposed regularized MLE allow a closed-form of RAID-G, which is very efficient and effective for high dimensional features. We extend RAID-G by using the outputs of deep convolutional neural networks as original features, and apply it to material recognition. Our approach is evaluated on five material benchmarks and one fine-grained benchmark. It achieves 84.9% accuracy on FMD and 86.3% accuracy on UIUC material database, which are much higher than state-of-the-arts. | ['WangMeng Zuo', 'Peihua Li', 'Lei Zhang', 'Qilong Wang'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['material-recognition'] | ['computer-vision'] | [-2.09289387e-01 -2.58123070e-01 -7.54754692e-02 -1.12239495e-01
-8.57567132e-01 -2.64657557e-01 5.92481256e-01 -3.74302387e-01
-1.58643067e-01 4.48368549e-01 2.21774966e-01 1.98322773e-01
-4.79119003e-01 -7.25197911e-01 -7.08265543e-01 -9.49212193e-01
-2.74020672e-01 2.88985848e-01 1.27235323e-01 -2.63069179e-02
1.04953080e-01 7.58594155e-01 -1.50181460e+00 -2.38033608e-01
8.18312824e-01 1.40772593e+00 1.86829641e-01 3.39438409e-01
6.11708984e-02 4.65141684e-01 -2.20000014e-01 -2.55287796e-01
3.49506289e-01 1.42057344e-01 -5.36974847e-01 2.20007345e-01
5.54441452e-01 -4.77997601e-01 -6.66523337e-01 1.09062624e+00
5.36980748e-01 3.42317581e-01 1.57885039e+00 -1.17568493e+00
-1.41430366e+00 1.44656301e-01 -5.13844073e-01 -3.03883553e-01
1.85318902e-01 -1.97775707e-01 8.37990880e-01 -1.25421250e+00
4.61594075e-01 1.46941555e+00 6.78572655e-01 5.48624396e-01
-9.84384894e-01 -4.08237934e-01 1.97746456e-04 1.84125397e-02
-1.93306673e+00 -9.24170539e-02 9.01932299e-01 -7.00796843e-01
7.38840044e-01 1.48187980e-01 1.88791946e-01 9.45702791e-01
4.74420488e-01 8.52867305e-01 9.34525073e-01 -9.89435986e-02
1.19481049e-01 -5.65915219e-02 2.83093423e-01 8.87074172e-01
2.15234444e-01 1.22480877e-01 -6.98434561e-02 -2.36593127e-01
9.37332690e-01 1.63467377e-01 -2.24377930e-01 -6.83963776e-01
-1.18180013e+00 8.49666417e-01 6.07266963e-01 1.03806727e-01
-4.74586874e-01 8.86817724e-02 4.40603226e-01 2.48400867e-03
4.02592480e-01 1.56556293e-01 1.76441297e-02 -2.40163490e-01
-4.53352153e-01 2.84371942e-01 6.66917026e-01 1.13127375e+00
6.30557775e-01 8.62939507e-02 -2.87899286e-01 1.25272071e+00
5.25495112e-01 8.52887154e-01 4.51100469e-01 -8.20901573e-01
3.48813593e-01 4.62567359e-01 -9.16190371e-02 -1.35883832e+00
-4.27752763e-01 -3.10656488e-01 -1.48472178e+00 3.25139165e-01
1.56134382e-01 2.05820918e-01 -8.15815568e-01 1.64382803e+00
1.95143417e-01 7.24306181e-02 3.81265543e-02 1.10116220e+00
7.81602502e-01 4.93611634e-01 -3.63711506e-01 1.42412171e-01
1.09857249e+00 -7.02704430e-01 -6.24605834e-01 2.34509379e-01
5.65445840e-01 -7.15929031e-01 1.13678718e+00 3.29905421e-01
-7.19132483e-01 -5.79825878e-01 -1.42927718e+00 -1.48361981e-01
-4.03736591e-01 3.52634639e-01 6.42774701e-01 5.61411977e-01
-8.66715133e-01 8.72107565e-01 -7.32648611e-01 6.42951578e-03
4.86667991e-01 5.25580108e-01 -5.48504472e-01 -2.64252275e-01
-1.01534081e+00 4.49230611e-01 3.48267779e-02 3.05844456e-01
-5.14195859e-01 -5.29830515e-01 -1.00849342e+00 -1.92358688e-01
1.27405971e-01 -4.36091900e-01 6.18075311e-01 -2.90439636e-01
-1.63098013e+00 6.44760072e-01 1.22988261e-01 -5.54151945e-02
4.21266556e-01 -3.72500509e-01 -4.02456284e-01 5.84977418e-02
9.53782201e-02 2.79374301e-01 9.82370853e-01 -1.00653672e+00
9.12299752e-02 -4.70195532e-01 -1.55081064e-01 1.09148972e-01
-5.09328187e-01 -2.18205467e-01 -5.41937828e-01 -7.59234548e-01
2.53348351e-01 -1.16747844e+00 -2.09274162e-02 2.33307272e-01
-4.50697690e-01 -3.88339400e-01 1.00061631e+00 -5.44763505e-01
1.26426601e+00 -2.25874710e+00 2.65910387e-01 2.10893095e-01
4.29187596e-01 3.50413203e-01 -9.89394188e-02 7.48284310e-02
8.67488086e-02 -8.50717872e-02 -1.22665517e-01 -3.28688413e-01
4.59074020e-01 7.82646835e-02 -8.49759877e-02 7.11985350e-01
3.47495466e-01 7.97871709e-01 -5.95962763e-01 -3.67862225e-01
4.85258460e-01 8.64314616e-01 -4.33778048e-01 1.12715036e-01
2.92672336e-01 -7.41537511e-02 -7.14776933e-01 6.19760334e-01
1.29333365e+00 -3.05625141e-01 -4.97234583e-01 -7.24886477e-01
4.29346897e-02 -2.57171214e-01 -1.37461793e+00 1.65351868e+00
-4.47726309e-01 3.77026439e-01 -2.06701487e-01 -8.12371790e-01
1.17076528e+00 -1.32891551e-01 5.26620984e-01 -4.61946547e-01
2.98250049e-01 3.55809987e-01 -1.37455478e-01 -2.82753617e-01
4.66254115e-01 2.93126225e-01 -1.59748584e-01 1.95744246e-01
9.60651562e-02 -3.51083912e-02 -2.06718929e-02 2.65569806e-01
9.06952143e-01 1.58114642e-01 -4.61820252e-02 -6.44360125e-01
8.26623201e-01 -7.51092255e-01 3.90664160e-01 5.37743568e-01
-1.86789826e-01 8.70632887e-01 2.81063676e-01 -3.19160193e-01
-9.50994015e-01 -1.37059295e+00 -5.50765991e-01 4.77348834e-01
3.74167621e-01 -4.66500401e-01 -7.44855642e-01 -6.41520202e-01
2.29544342e-01 2.42441639e-01 -6.99279547e-01 -5.61342120e-01
-4.02699828e-01 -7.38723874e-01 3.38206530e-01 7.18271554e-01
7.65210032e-01 -5.18395603e-01 2.12260723e-01 5.49033843e-02
3.26866210e-01 -1.17673028e+00 -9.45145547e-01 -1.83186308e-01
-4.88230169e-01 -9.05809939e-01 -9.66487467e-01 -6.45456851e-01
6.26532018e-01 1.97534487e-01 7.44050384e-01 -2.50618398e-01
-4.44477379e-01 4.38868791e-01 -1.92807138e-01 -4.40727621e-02
-2.89562315e-01 -1.46210538e-02 5.02295196e-01 2.53106833e-01
4.81444031e-01 -4.90092874e-01 -7.06381559e-01 6.41743541e-01
-8.72278035e-01 -8.28012899e-02 8.84924173e-01 9.28892255e-01
8.41026187e-01 -1.78794578e-01 5.29175162e-01 -4.04266715e-01
8.03599954e-01 -3.17902565e-01 -5.35599649e-01 1.72841072e-01
-6.34031057e-01 4.84932333e-01 7.27778435e-01 -6.39711499e-01
-8.14803660e-01 -6.91663474e-02 -2.53503233e-01 -8.39914024e-01
2.92186569e-02 2.05651328e-01 -3.93464535e-01 -3.41479838e-01
5.70584774e-01 1.68856323e-01 1.90119624e-01 -6.89291477e-01
4.30115759e-01 1.09455824e+00 5.50589085e-01 -9.71894741e-01
8.46568704e-01 4.31493759e-01 5.60785890e-01 -1.04653513e+00
-7.90081739e-01 -2.82108873e-01 -8.92915070e-01 -1.70811698e-01
8.67895782e-01 -7.13733256e-01 -8.48222136e-01 8.81530285e-01
-9.72137213e-01 2.24102214e-02 -1.19145341e-01 7.58184791e-01
-8.18935752e-01 7.21047819e-01 -8.22888970e-01 -5.91809154e-01
-5.68626583e-01 -1.46378970e+00 1.26536310e+00 7.90609270e-02
2.07576752e-01 -1.03435874e+00 -7.18120709e-02 9.39039662e-02
5.03904104e-01 4.10426348e-01 8.14405680e-01 -4.12752956e-01
-5.07457554e-01 -5.42999387e-01 -6.41465485e-01 9.18252885e-01
4.32152092e-01 -1.01931401e-01 -7.82443404e-01 -3.49881649e-01
-1.20435074e-01 -3.96255374e-01 5.56206703e-01 3.47470075e-01
1.23058963e+00 -5.71470112e-02 -1.24368623e-01 8.03243577e-01
1.36463118e+00 -1.73583359e-01 6.05858326e-01 1.05882332e-01
1.00212932e+00 1.80701241e-01 7.37615108e-01 4.60020483e-01
2.81502217e-01 9.02797043e-01 2.58633584e-01 9.54246968e-02
-3.82936239e-01 -5.09609766e-02 3.89475405e-01 1.31478846e+00
-1.26516655e-01 8.36739838e-02 -5.65410316e-01 1.90193281e-01
-1.89460921e+00 -5.15663326e-01 -2.57015079e-01 2.28240824e+00
3.50559890e-01 5.70806116e-02 -3.66710387e-02 -1.47481993e-01
7.15152264e-01 7.91205317e-02 -7.48365045e-01 -2.25840315e-01
-3.19458723e-01 4.68647033e-02 4.23971683e-01 2.87643939e-01
-1.30263066e+00 6.06288731e-01 5.64048290e+00 1.37542129e+00
-8.93846691e-01 -5.60527034e-02 2.96902329e-01 1.97606862e-01
1.20991647e-01 -4.74686772e-01 -9.65564132e-01 5.02886117e-01
6.13114655e-01 -2.35317677e-01 3.50393295e-01 1.11546898e+00
-2.34112427e-01 3.45768571e-01 -1.12497759e+00 1.46552503e+00
3.01940411e-01 -1.30131161e+00 1.91613510e-01 4.85657901e-01
7.21293628e-01 3.91989499e-02 2.80250669e-01 4.31495935e-01
5.36793880e-02 -9.26446259e-01 4.55244899e-01 7.77133942e-01
9.65158284e-01 -9.01351929e-01 6.60727322e-01 5.89664690e-02
-1.48992217e+00 3.41987401e-01 -8.20486605e-01 3.84266317e-01
7.93963764e-03 5.25174856e-01 -2.50351459e-01 7.02313304e-01
4.49680746e-01 9.81505930e-01 -5.85101604e-01 9.49444413e-01
2.65533388e-01 1.11323684e-01 -5.11466861e-01 -2.88771421e-01
2.95111537e-01 -5.57999611e-01 6.60616755e-01 1.09254932e+00
4.37989056e-01 -3.57738644e-01 1.78304166e-01 9.82842863e-01
-1.23800248e-01 2.48810202e-01 -6.36571348e-01 7.21415132e-03
1.72333658e-01 1.32295156e+00 -3.28209758e-01 -3.73839103e-02
-5.44647455e-01 1.29423785e+00 4.40494746e-01 2.36111462e-01
-7.06923544e-01 -6.87355757e-01 9.13037479e-01 -3.44012119e-02
3.49364609e-01 -3.84743005e-01 2.93827266e-01 -1.18906653e+00
3.46569806e-01 -5.46137333e-01 1.53002322e-01 -6.25471175e-01
-1.83555853e+00 6.44264996e-01 -5.54872714e-02 -1.49341595e+00
-3.27014893e-01 -1.20609653e+00 -2.84622967e-01 8.51257741e-01
-1.08210897e+00 -1.28530407e+00 -3.64940733e-01 6.11175954e-01
2.64602125e-01 -3.27325046e-01 8.51655304e-01 5.89112997e-01
-5.66412926e-01 8.35645854e-01 6.30168498e-01 3.45659643e-01
5.54501593e-01 -1.32032788e+00 3.00281376e-01 3.70417923e-01
-2.13370472e-01 6.09081268e-01 3.54702413e-01 -5.10454416e-01
-1.91245615e+00 -1.25834608e+00 8.33195150e-02 -3.50670516e-01
8.78399611e-01 -6.24080002e-01 -9.67547357e-01 3.06054264e-01
-4.04785335e-01 3.37758750e-01 5.60845613e-01 -1.02917887e-01
-4.79801536e-01 -5.42385578e-02 -9.87705112e-01 5.94352782e-01
1.09379292e+00 -6.61827981e-01 -1.65455207e-01 5.12176156e-01
8.38588238e-01 -2.91245043e-01 -1.46686375e+00 6.04314446e-01
6.76172674e-01 -5.77552199e-01 1.12283242e+00 -4.51259613e-01
8.15051943e-02 -2.59016842e-01 -7.39075840e-01 -1.04963517e+00
-5.62519789e-01 -5.18430471e-01 -4.43507165e-01 1.29781342e+00
1.25946999e-01 -5.98989725e-01 6.72191203e-01 6.04821503e-01
-3.32110494e-01 -1.29726017e+00 -9.75871623e-01 -1.32225072e+00
4.05844420e-01 -5.98112702e-01 5.82390308e-01 6.06402218e-01
-4.50878054e-01 9.12878811e-02 -6.14098489e-01 8.90684873e-02
9.56930101e-01 -3.06486674e-02 8.46449137e-01 -1.24343002e+00
-2.77941167e-01 -3.81779701e-01 -1.05200124e+00 -1.38401520e+00
3.32304180e-01 -9.49019372e-01 -1.42880097e-01 -1.31127501e+00
2.81110197e-01 -4.23294276e-01 -3.46739709e-01 -1.21968098e-01
1.54819250e-01 1.67586625e-01 7.10158497e-02 1.30249307e-01
-6.65111661e-01 1.18328929e+00 1.28991878e+00 -2.55491018e-01
3.70478183e-02 -1.40300719e-02 -3.09593529e-01 7.44855523e-01
3.93636703e-01 -8.45760256e-02 -2.32862547e-01 -6.18401393e-02
-1.79938987e-01 -4.21889991e-01 2.64337927e-01 -1.18134201e+00
6.37032986e-02 1.41688555e-01 2.91823775e-01 -7.73873627e-01
6.81567907e-01 -6.69585109e-01 -5.33609577e-02 9.81995743e-03
-1.90717205e-02 -2.25330472e-01 -1.16284899e-02 6.65687740e-01
-1.89686999e-01 -3.00486952e-01 7.77883351e-01 2.95166731e-01
-5.91413856e-01 8.61631393e-01 -5.08894064e-02 -1.49201443e-02
8.11833680e-01 -2.18189061e-01 -2.81022012e-01 -1.83562458e-01
-5.83313167e-01 -8.30070078e-02 4.93321091e-01 6.65953577e-01
8.82431448e-01 -2.17186975e+00 -5.68212569e-01 3.13710004e-01
2.47358859e-01 -1.41724378e-01 4.62515712e-01 8.96083057e-01
-3.38362366e-01 3.36545467e-01 2.89329607e-03 -6.92580223e-01
-1.06998456e+00 7.28379548e-01 4.06261981e-01 -2.10389629e-01
-7.67740488e-01 5.75974286e-01 5.72472692e-01 -5.23638427e-01
1.95262596e-01 -2.07000121e-01 1.10950261e-01 -4.32619095e-01
6.12438202e-01 4.51063842e-01 7.87721053e-02 -1.19290829e+00
-3.69591832e-01 1.16122770e+00 -2.78251767e-01 2.92867780e-01
1.31399763e+00 -9.51202810e-02 -6.47581518e-02 3.74172211e-01
2.00284958e+00 -2.12925926e-01 -1.29588068e+00 -4.99485254e-01
-2.51234561e-01 -5.02609968e-01 3.64453107e-01 -1.07426688e-01
-9.97551680e-01 9.56122875e-01 9.64062572e-01 7.83881694e-02
8.91265452e-01 1.39704749e-01 9.12155449e-01 6.17956698e-01
3.38269413e-01 -1.19782591e+00 2.52410382e-01 4.95060891e-01
1.30114400e+00 -1.07095742e+00 1.45230114e-01 -6.27869010e-01
-3.64004105e-01 1.21419322e+00 5.98314583e-01 -4.78481650e-01
1.15363550e+00 1.49770036e-01 -3.20816875e-01 -3.49966943e-01
-2.00991943e-01 -2.40097642e-02 8.38248789e-01 6.28934622e-01
2.88062394e-01 9.46364254e-02 -1.77323386e-01 7.40086138e-01
-1.52997747e-02 -2.58096904e-01 7.55130723e-02 7.24617243e-01
-2.35396311e-01 -9.25364614e-01 -3.41978639e-01 7.08738387e-01
-2.25414827e-01 2.26575509e-01 -1.77599847e-01 7.70325363e-01
-1.78581640e-01 6.10756695e-01 6.48608506e-02 -6.98324800e-01
3.26560527e-01 -3.57822329e-01 6.38429403e-01 -1.79674149e-01
2.74651468e-01 2.55431570e-02 -1.24857321e-01 -7.57032633e-01
-1.57516882e-01 -7.12936342e-01 -1.17895508e+00 -2.47911662e-01
-5.04117548e-01 -8.90772641e-02 7.54456699e-01 7.81055868e-01
5.41775107e-01 2.30528638e-01 8.04725349e-01 -1.08919930e+00
-9.68668938e-01 -9.06242967e-01 -1.09688056e+00 8.35256755e-01
1.51718199e-01 -1.24795508e+00 -5.29341638e-01 -4.99743432e-01] | [8.928760528564453, 2.0362493991851807] |
7481ec9a-b40b-4343-9844-1872729d16a4 | learning-to-compile-smartly-for-program-size | 2301.05104 | null | https://arxiv.org/abs/2301.05104v2 | https://arxiv.org/pdf/2301.05104v2.pdf | Learning Compiler Pass Orders using Coreset and Normalized Value Prediction | Finding the optimal pass sequence of compilation can lead to a significant reduction in program size and/or improvement in program efficiency. Prior works on compilation pass ordering have two major drawbacks. They either require an excessive budget (in terms of compilation steps) at compile time or fail to generalize to unseen programs. In this paper, for code-size reduction tasks, we propose a novel pipeline to find program-dependent pass sequences within 45 compilation calls. It first identifies a coreset of 50 pass sequences via greedy optimization of a submodular function, and then learns a policy with Graph Neural Network (GNN) to pick the optimal sequence by predicting the normalized values of the pass sequences in the coreset. Despite its simplicity, our pipeline outperforms the default -Oz flag by an average of 4.7% over a large collection (4683) of unseen code repositories from diverse domains across 14 datasets. In comparison, previous approaches like reinforcement learning on the raw pass sequence space may take days to train due to sparse reward, and may not generalize well in held-out ones from different domains. Our results demonstrate that existing human-designed compiler flags can be improved with a simple yet effective technique that transforms the raw action space into a small one with denser rewards. | ['Yuandong Tian', 'Hugh Leather', 'Xiaomeng Yang', 'Pengtao Xie', 'Benoit Steiner', 'Jiadong Guo', 'Mostafa Elhoushi', 'Chris Cummins', 'Ali Shameli', 'Kevin Stone', 'Youwei Liang'] | 2023-01-09 | null | null | null | null | ['compiler-optimization', 'value-prediction'] | ['computer-code', 'computer-code'] | [ 5.92779629e-02 -1.87926486e-01 -5.80393493e-01 -2.53803670e-01
-9.83815908e-01 -9.65905786e-01 3.15177031e-02 1.78941697e-01
-1.61873087e-01 6.62752867e-01 1.42865032e-01 -6.19979620e-01
-4.87968698e-02 -8.85412455e-01 -1.22677636e+00 -2.41389453e-01
-3.24932277e-01 3.81342173e-01 2.54744917e-01 -3.85539889e-01
5.40851057e-01 -3.52868401e-02 -1.22237360e+00 2.31920034e-01
1.22008133e+00 6.03650153e-01 4.11427945e-01 7.61660516e-01
6.06866665e-02 8.99184883e-01 -7.47251034e-01 -1.26820177e-01
4.90550309e-01 -4.27420169e-01 -9.52498138e-01 -9.63876545e-02
3.73595178e-01 -4.19741094e-01 -3.11566770e-01 1.26550603e+00
2.81215012e-01 2.17047155e-01 1.21689945e-01 -1.07349956e+00
-7.71099150e-01 1.09602582e+00 -6.66758239e-01 3.99884701e-01
4.85929489e-01 5.15010417e-01 1.15050936e+00 -2.21864924e-01
6.72673523e-01 1.01118112e+00 5.82087040e-01 5.12668073e-01
-1.54577565e+00 -4.80306685e-01 9.03295353e-02 -5.77669330e-02
-1.01647317e+00 -1.00953877e-01 5.31163394e-01 -6.32161379e-01
1.42124188e+00 2.88907409e-01 6.43800259e-01 9.82114494e-01
2.23280475e-01 7.09894598e-01 1.01021075e+00 -2.48925924e-01
3.52776974e-01 -1.64205298e-01 5.65174744e-02 9.70963776e-01
2.49451622e-01 3.36362362e-01 -4.59973276e-01 -3.82398516e-01
3.28252643e-01 -1.81522481e-02 -3.56886417e-01 -4.21614349e-01
-1.31263840e+00 1.01019371e+00 6.26367748e-01 6.96350113e-02
-4.02466953e-02 5.23670077e-01 7.50919461e-01 4.89092708e-01
-6.85152737e-03 1.21235883e+00 -5.48465431e-01 -5.69206476e-01
-9.79380071e-01 4.54850972e-01 1.15487862e+00 1.01273930e+00
1.17141342e+00 1.97192773e-01 -3.13194424e-01 6.34629786e-01
-9.71449837e-02 3.71367216e-01 5.76027334e-01 -8.04058969e-01
7.91749001e-01 9.74469304e-01 -1.72367215e-01 -9.51908112e-01
-3.54660153e-01 -4.56120908e-01 -3.38741034e-01 2.17511386e-01
2.89166987e-01 -4.73358154e-01 -6.91566527e-01 1.54964221e+00
6.76675960e-02 -2.08806619e-01 -2.50336111e-01 9.62167978e-01
4.51374918e-01 6.00297570e-01 -1.87078625e-01 2.58034378e-01
1.02317953e+00 -1.23173022e+00 -7.86314905e-02 -7.15474486e-01
1.04403865e+00 -5.43722928e-01 1.43078649e+00 4.10387844e-01
-7.69911706e-01 -2.90487617e-01 -1.20595205e+00 1.48384094e-01
-1.67316571e-01 1.48829445e-01 1.03819263e+00 7.53821373e-01
-1.13666964e+00 7.87613213e-01 -9.34348762e-01 -6.04749992e-02
3.33426982e-01 5.84756434e-01 8.61972003e-05 -1.33147985e-01
-6.45593286e-01 5.30778825e-01 6.52554512e-01 -3.09786111e-01
-1.36511552e+00 -1.01749623e+00 -9.32500601e-01 2.23572195e-01
9.01393950e-01 -6.31119132e-01 1.24138546e+00 -9.94540870e-01
-1.33396244e+00 5.80027997e-01 2.23256454e-01 -7.00267494e-01
3.31802368e-01 -4.28572625e-01 7.94502273e-02 -3.10786575e-01
1.87244490e-01 3.47345710e-01 6.56321824e-01 -8.63776922e-01
-5.62598407e-01 -9.59032401e-02 6.47413552e-01 -2.00406052e-02
-3.86190385e-01 -1.05360873e-01 -3.40911895e-01 -3.80711824e-01
-4.25628364e-01 -1.08748424e+00 -4.48448539e-01 -6.49998605e-01
-4.74310845e-01 -1.39508769e-01 4.78956372e-01 -6.47892416e-01
1.49077618e+00 -2.10782218e+00 3.44359696e-01 -2.61548292e-02
2.93815047e-01 -5.22897989e-02 -2.43641511e-01 4.32218909e-01
1.94450170e-01 2.37142548e-01 -3.67580801e-01 2.39772663e-01
-5.54529717e-03 1.91496089e-01 -2.61663646e-01 5.20761192e-01
1.18335769e-01 8.17397535e-01 -1.28098428e+00 -1.18276544e-01
-9.93355215e-02 -2.69887805e-01 -1.22779119e+00 2.73373157e-01
-7.60813296e-01 1.95450187e-01 -3.62053007e-01 6.73885703e-01
2.80145347e-01 -5.13594747e-01 1.87285498e-01 4.55446482e-01
1.46307063e-03 5.17449737e-01 -8.88208568e-01 1.83202410e+00
-6.15304708e-01 6.01419628e-01 -2.25274771e-01 -1.07415688e+00
1.03134966e+00 -2.73410618e-01 2.14228779e-01 -7.91819632e-01
-1.27364129e-01 3.44358951e-01 3.16752344e-01 -5.21388233e-01
7.83643663e-01 3.36516082e-01 -5.51733017e-01 4.62387741e-01
5.50822914e-03 -2.10463360e-01 6.84939981e-01 5.73318005e-02
1.75889242e+00 2.31054142e-01 3.34472358e-01 -4.01185393e-01
3.84858578e-01 4.71412808e-01 7.12289691e-01 9.93164062e-01
-3.92527804e-02 3.92061502e-01 1.12190700e+00 -6.43559456e-01
-1.11344802e+00 -5.53070664e-01 4.19893593e-01 1.58723664e+00
-7.03815967e-02 -5.10971248e-01 -7.80177116e-01 -1.10614836e+00
1.00938819e-01 7.69077063e-01 -6.83789849e-01 -3.38567168e-01
-9.65271235e-01 -7.26184070e-01 4.60424930e-01 5.21029592e-01
3.80370140e-01 -1.09630620e+00 -6.68068528e-01 2.63571054e-01
1.45347983e-01 -8.54857266e-01 -9.71842527e-01 5.25750995e-01
-8.48733008e-01 -1.22015738e+00 -3.68766069e-01 -7.42389977e-01
8.07892084e-01 1.06638506e-01 1.60164559e+00 4.30168599e-01
-2.77567297e-01 4.87053953e-02 -3.57717127e-01 -2.51194518e-02
-7.12837815e-01 5.54870188e-01 -3.64162982e-01 -6.90476894e-01
2.74733931e-01 -2.32491106e-01 -6.25508368e-01 2.65513927e-01
-6.19183123e-01 -3.41228515e-01 4.84568328e-01 1.11190832e+00
3.31628352e-01 -1.47580519e-01 3.45731616e-01 -1.31693470e+00
8.22649121e-01 -6.95936024e-01 -1.30390990e+00 1.66813716e-01
-6.77223384e-01 5.07725298e-01 1.26254070e+00 -3.41688693e-01
-6.93313777e-01 2.55961567e-01 2.22567841e-01 -3.35407346e-01
1.76823780e-01 5.35474122e-01 3.65635425e-01 -2.11467862e-01
1.33183146e+00 3.82579625e-01 -1.99455380e-01 -1.72581658e-01
1.49296120e-01 1.56706095e-01 4.04853076e-01 -1.04987454e+00
7.92312443e-01 -4.46464829e-02 -1.79313749e-01 -9.50118750e-02
-8.13448906e-01 -3.71619940e-01 -2.39611819e-01 1.90255761e-01
5.00590384e-01 -8.43776762e-01 -8.25084627e-01 -7.91650172e-03
-8.05715740e-01 -7.57981718e-01 -5.15781268e-02 1.76298380e-01
-6.04737043e-01 4.16630447e-01 -5.73453605e-01 -3.19418877e-01
-4.50353533e-01 -1.60198462e+00 9.74660039e-01 3.58541638e-01
-3.44438940e-01 -7.99457967e-01 2.00133830e-01 2.82481372e-01
4.15836722e-01 2.45491982e-01 1.10715413e+00 -4.95618463e-01
-9.71924663e-01 -2.84288730e-02 8.06950592e-03 4.61772591e-01
-1.09324798e-01 -1.06076442e-01 -4.16662961e-01 -6.04051352e-01
-3.01892161e-01 -4.69256312e-01 8.34694505e-01 2.03262165e-01
1.38074625e+00 -5.14491975e-01 -1.88634142e-01 1.01506233e+00
1.77370274e+00 2.64130086e-01 3.39394689e-01 8.29530716e-01
6.71207428e-01 -2.61495020e-02 7.78193712e-01 5.94682217e-01
1.22744158e-01 2.65849203e-01 8.35047483e-01 1.52098283e-01
-1.15125766e-02 -3.92705351e-01 7.78065681e-01 7.14063346e-01
6.78838342e-02 6.60933042e-03 -1.16366994e+00 1.01605058e+00
-1.66456437e+00 -6.19331598e-01 2.59112537e-01 2.28783560e+00
9.12146926e-01 2.68209726e-01 3.90068084e-01 -3.21504653e-01
5.22535980e-01 9.52199176e-02 -7.48151839e-01 -8.22613657e-01
2.63833165e-01 1.83316767e-01 9.71331537e-01 3.15129429e-01
-9.25129354e-01 9.70020711e-01 6.30087614e+00 7.03669250e-01
-1.27190709e+00 -1.35448560e-01 5.62809289e-01 -7.74901286e-02
-3.29800785e-01 4.12452787e-01 -8.78998220e-01 5.23355246e-01
1.01502943e+00 -5.33088088e-01 1.13492405e+00 1.50563419e+00
-1.04909360e-01 -1.74243614e-01 -1.41758192e+00 6.38148010e-01
-1.20355435e-01 -1.53107071e+00 -3.90411675e-01 -7.31171072e-02
1.24298990e+00 3.69284123e-01 -1.00045442e-03 8.12746167e-01
8.74372423e-01 -1.09142137e+00 5.72122276e-01 -4.85524274e-02
5.37433326e-01 -8.84718597e-01 5.01540422e-01 5.25039792e-01
-8.41604829e-01 -3.40541661e-01 -5.90661407e-01 -1.94004491e-01
-5.59727848e-01 4.14965272e-01 -1.44222724e+00 2.45934904e-01
7.71787107e-01 5.70633531e-01 -8.16499829e-01 1.26308715e+00
-1.93899140e-01 8.27046454e-01 -1.17128722e-01 -6.22744799e-01
4.60158736e-01 1.79044735e-02 5.68344533e-01 1.32059205e+00
2.58165956e-01 -4.41174626e-01 5.52731097e-01 1.22982037e+00
-3.67337078e-01 8.89851071e-04 -6.81461155e-01 -2.72932440e-01
3.58835489e-01 1.22684228e+00 -6.49974525e-01 -1.36776373e-01
-4.74290788e-01 7.04950690e-01 6.60087347e-01 1.91897333e-01
-9.56279576e-01 -5.95703244e-01 4.63187724e-01 -1.50362318e-02
4.93019849e-01 -2.80841529e-01 -4.33502048e-01 -9.98003662e-01
2.52869666e-01 -1.39761972e+00 4.09464210e-01 -4.23195995e-02
-9.24819112e-01 6.04073882e-01 -2.55659640e-01 -1.16383910e+00
-3.61653835e-01 -5.10975897e-01 -5.63451171e-01 7.23927975e-01
-1.37699664e+00 -3.57391536e-01 -1.77701622e-01 2.49269083e-01
6.64289117e-01 -3.00187767e-01 6.69739246e-01 1.87666133e-01
-5.91327727e-01 7.50189066e-01 2.53748387e-01 7.74968043e-02
4.83335227e-01 -1.69746029e+00 6.44061863e-01 9.22985077e-01
-1.48158655e-01 8.81641626e-01 7.36154318e-01 -6.16929471e-01
-1.87053788e+00 -1.09906793e+00 2.30193645e-01 -3.63710403e-01
8.64678979e-01 -4.41160023e-01 -8.45994413e-01 7.31472373e-01
1.89902484e-01 1.03808925e-01 3.68747026e-01 3.60058993e-01
-3.89685631e-01 -7.36514926e-02 -8.66027594e-01 5.30761719e-01
9.48250175e-01 -2.83735752e-01 -3.24290335e-01 5.73329926e-01
9.16244626e-01 -9.84730065e-01 -9.92005467e-01 7.25972950e-02
6.44622510e-03 -8.57910216e-01 6.47854805e-01 -7.53696263e-01
1.04953778e+00 -2.76665479e-01 -6.01229519e-02 -1.58304060e+00
-3.57204884e-01 -1.02081525e+00 -6.40028268e-02 8.59565973e-01
6.58045590e-01 -4.19064939e-01 9.65943217e-01 3.07149261e-01
-2.89927870e-01 -1.03534412e+00 -4.58931416e-01 -8.74369323e-01
-2.58318651e-02 -1.26073688e-01 8.28870833e-01 7.69673228e-01
7.59707391e-02 3.79141271e-01 -4.23073798e-01 8.56012106e-02
3.81678134e-01 6.36884391e-01 1.01022112e+00 -5.85430086e-01
-9.74868238e-01 -4.62875903e-01 -1.72163844e-01 -1.21104240e+00
2.08887950e-01 -1.14979446e+00 3.35771352e-01 -1.09775054e+00
4.26999539e-01 -4.39711481e-01 -7.44523406e-02 6.55899346e-01
-3.85740697e-01 -3.71356666e-01 6.29892647e-02 -1.62051003e-02
-8.50712359e-01 3.43742430e-01 1.18893182e+00 -2.99815148e-01
-3.43526036e-01 -2.02954169e-02 -9.04330909e-01 5.03411591e-01
6.44214809e-01 -7.22528398e-01 -2.59214461e-01 -7.35015154e-01
6.59235120e-01 4.58064288e-01 1.43314376e-01 -1.04322147e+00
1.62120327e-01 -3.30559254e-01 3.37533839e-02 -3.30748767e-01
-4.54971880e-01 -4.46405590e-01 -1.93415001e-01 7.96549141e-01
-4.20779049e-01 2.57805973e-01 4.12839115e-01 6.47658467e-01
-1.03627212e-01 -6.83231950e-01 5.51647723e-01 -3.79506201e-01
-9.59066272e-01 7.73342550e-02 -1.47589251e-01 3.89679074e-01
1.09535480e+00 -4.42002900e-02 -5.66289067e-01 1.03721514e-01
-1.94366515e-01 4.71757472e-01 6.95004284e-01 6.33492768e-01
1.70165896e-01 -8.59951079e-01 -6.01126373e-01 -5.84186688e-02
1.21065453e-01 6.62711114e-02 2.05322336e-02 5.30721843e-01
-8.60426605e-01 3.86138022e-01 -2.91084200e-01 -7.54227102e-01
-1.03841424e+00 7.05018818e-01 2.61917949e-01 -7.32568383e-01
-6.50592804e-01 8.37748408e-01 4.48084511e-02 -6.66058242e-01
3.47966701e-02 -4.35346246e-01 2.44031563e-01 -4.27191079e-01
1.68786585e-01 3.91671389e-01 2.34397978e-01 -1.17287580e-02
-2.89871544e-01 1.75390288e-01 -3.23129088e-01 4.89863306e-01
1.64104295e+00 4.57071185e-01 -2.05858901e-01 4.87674661e-02
1.41180825e+00 1.76550239e-01 -1.32281876e+00 1.86567032e-03
1.98242277e-01 -7.81892121e-01 -3.56996246e-02 -7.20771432e-01
-1.19427061e+00 4.24791843e-01 3.23371619e-01 2.49762043e-01
1.06348348e+00 -1.25138447e-01 7.64210165e-01 6.52780354e-01
5.31875074e-01 -1.04499722e+00 1.55030474e-01 6.27530158e-01
7.34448195e-01 -1.30007994e+00 -6.04669796e-03 -2.83441275e-01
-5.00233412e-01 1.33243060e+00 1.08459008e+00 -5.15154660e-01
-7.48498961e-02 4.28162932e-01 -4.84027743e-01 -2.43755266e-01
-8.75906110e-01 2.00159371e-01 1.08115144e-01 4.73970324e-01
5.93640268e-01 1.90794274e-01 -2.91684240e-01 3.55428308e-01
-5.54100335e-01 -9.59722698e-03 8.81658673e-01 9.36991990e-01
-3.64025772e-01 -1.01962721e+00 -3.75943661e-01 8.65197539e-01
-5.74497104e-01 -7.92583227e-02 -8.62546936e-02 6.22424245e-01
-1.63676575e-01 6.04288399e-01 -3.12693149e-01 -5.02909124e-01
4.19961214e-01 -3.41363639e-01 3.97696912e-01 -1.04103982e+00
-1.09987891e+00 -4.18761194e-01 1.32061735e-01 -8.42250645e-01
2.86896467e-01 -6.56952381e-01 -1.20574498e+00 -3.51116300e-01
-3.79047245e-01 -9.54809487e-02 4.25964296e-01 4.70302492e-01
4.46068972e-01 9.05987918e-01 6.85803115e-01 -5.32725573e-01
-1.04752982e+00 -5.61026692e-01 -2.74822831e-01 3.44096094e-01
3.87028933e-01 -4.37568635e-01 -2.84827381e-01 -9.11201015e-02] | [7.878328800201416, 7.615054607391357] |
c8e2adee-1e45-4011-ab0d-1da717ef8e6a | multiresolution-deep-implicit-functions-for | 2109.05591 | null | https://arxiv.org/abs/2109.05591v2 | https://arxiv.org/pdf/2109.05591v2.pdf | Multiresolution Deep Implicit Functions for 3D Shape Representation | We introduce Multiresolution Deep Implicit Functions (MDIF), a hierarchical representation that can recover fine geometry detail, while being able to perform global operations such as shape completion. Our model represents a complex 3D shape with a hierarchy of latent grids, which can be decoded into different levels of detail and also achieve better accuracy. For shape completion, we propose latent grid dropout to simulate partial data in the latent space and therefore defer the completing functionality to the decoder side. This along with our multires design significantly improves the shape completion quality under decoder-only latent optimization. To the best of our knowledge, MDIF is the first deep implicit function model that can at the same time (1) represent different levels of detail and allow progressive decoding; (2) support both encoder-decoder inference and decoder-only latent optimization, and fulfill multiple applications; (3) perform detailed decoder-only shape completion. Experiments demonstrate its superior performance against prior art in various 3D reconstruction tasks. | ['Danhang Tang', 'Thomas Funkhouser', 'Cem Keskin', 'Ruofei Du', 'Christian Haene', 'Sofien Bouaziz', 'Sean Fanello', 'Kyle Genova', 'yinda zhang', 'Zhang Chen'] | 2021-09-12 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Chen_Multiresolution_Deep_Implicit_Functions_for_3D_Shape_Representation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Chen_Multiresolution_Deep_Implicit_Functions_for_3D_Shape_Representation_ICCV_2021_paper.pdf | iccv-2021-1 | ['3d-shape-representation'] | ['computer-vision'] | [ 6.80951104e-02 3.01166266e-01 4.31724899e-02 -1.22948162e-01
-1.26834273e+00 -6.44543946e-01 5.95755756e-01 -1.10567071e-01
2.95303226e-01 4.20161188e-01 6.49622262e-01 -2.14261100e-01
3.70962560e-01 -9.82983470e-01 -8.19995463e-01 -3.81735504e-01
1.40696950e-02 9.47416961e-01 2.46330649e-02 -3.16430442e-02
2.45968287e-04 9.19208050e-01 -1.23478186e+00 6.96855903e-01
9.45317924e-01 8.33766460e-01 4.88571256e-01 6.02842271e-01
-4.74128276e-02 5.16868293e-01 -9.70100835e-02 -2.41416827e-01
2.25427523e-01 -9.68283042e-03 -5.94407320e-01 3.04116040e-01
5.08587897e-01 -9.46119368e-01 -3.51451784e-01 6.29295647e-01
4.98412311e-01 -1.99848682e-01 7.85568714e-01 -4.39673364e-01
-8.34391713e-01 1.37082428e-01 -5.64577162e-01 -4.22224402e-01
3.69510412e-01 2.65049368e-01 9.37090278e-01 -1.45244396e+00
6.43439770e-01 1.56326604e+00 6.28012538e-01 3.10215235e-01
-1.59243929e+00 -5.49989522e-01 1.25106990e-01 -5.07323146e-01
-1.51037812e+00 -6.73170269e-01 8.70220959e-01 -5.48282444e-01
1.12241054e+00 5.70603497e-02 6.93437636e-01 8.54309380e-01
2.00290367e-01 8.16150725e-01 7.97555089e-01 1.39242068e-01
2.00549275e-01 -6.15674078e-01 -5.13154030e-01 8.94936562e-01
-7.16362670e-02 -5.24407215e-02 -4.71185893e-01 -3.15062493e-01
1.65693069e+00 -1.10240102e-01 -3.51426572e-01 -2.29213446e-01
-1.29726732e+00 7.03901649e-01 2.87042022e-01 -1.29129007e-01
-4.99676853e-01 5.33114552e-01 9.97093767e-02 1.54362619e-01
7.86818027e-01 2.54714072e-01 -4.07303989e-01 3.33667956e-02
-1.17997861e+00 3.79111767e-01 7.08788216e-01 1.24646306e+00
9.28353786e-01 4.64482129e-01 -2.70142287e-01 6.81445122e-01
3.86674106e-01 6.57701910e-01 -2.56766021e-01 -1.26921892e+00
5.68359554e-01 4.27453607e-01 1.35996833e-01 -8.87579441e-01
-1.18970096e-01 -6.28295720e-01 -1.06822574e+00 4.34721261e-01
-8.64925385e-02 3.47044408e-01 -1.02764678e+00 1.52316523e+00
2.61089802e-01 2.96688735e-01 -1.33038640e-01 8.41185093e-01
9.61192250e-01 8.97402227e-01 -2.72229910e-01 -9.18870866e-02
1.13878715e+00 -8.25596809e-01 -4.99073684e-01 2.96050869e-02
4.39569831e-01 -9.62173939e-01 1.04033089e+00 4.68502194e-01
-1.59160089e+00 -5.68738520e-01 -1.02768743e+00 -7.29580641e-01
3.80742908e-01 2.22811505e-01 7.63209999e-01 1.10538259e-01
-1.47313678e+00 5.74212551e-01 -1.32331002e+00 4.32315350e-01
5.23227751e-01 2.85734266e-01 -3.03004831e-01 -3.36288542e-01
-6.01520538e-01 3.69024873e-01 -1.30715638e-01 -5.44109456e-02
-1.33275747e+00 -9.52891707e-01 -1.01288259e+00 6.98961690e-02
1.13207690e-01 -1.21151721e+00 9.58882213e-01 -4.12765533e-01
-1.54389894e+00 8.38341534e-01 -5.04082263e-01 8.74056853e-03
3.74249816e-01 -1.42923906e-01 2.55393684e-01 2.02353656e-01
1.38749719e-01 8.39865923e-01 8.67485046e-01 -1.39931226e+00
3.39788049e-02 -3.57065916e-01 1.66228525e-02 3.78679544e-01
-2.84435065e-03 -5.01487553e-01 -8.90543103e-01 -1.06266952e+00
6.20995283e-01 -6.04206085e-01 -3.27912569e-01 4.31772619e-01
-2.81355262e-01 -3.15635675e-03 7.44228601e-01 -1.11222398e+00
9.92603362e-01 -2.06175113e+00 5.50326467e-01 8.82106367e-03
5.28734088e-01 -2.45453313e-01 -2.80433536e-01 4.11239743e-01
1.34355783e-01 1.90280363e-01 -5.54979920e-01 -1.14657676e+00
-7.12490827e-02 4.84037608e-01 -4.86086011e-01 3.48687708e-01
1.96057200e-01 1.34549928e+00 -5.94041467e-01 -2.14722514e-01
2.25110114e-01 8.31357241e-01 -1.20430994e+00 2.33239278e-01
-4.37128663e-01 8.41862261e-01 -6.44631565e-01 7.25956261e-01
8.70307446e-01 -4.87931192e-01 4.79756184e-02 -6.25030041e-01
-1.76625878e-01 5.56636751e-01 -1.11233771e+00 2.27735686e+00
-6.61150277e-01 2.73235232e-01 5.81226707e-01 -5.37935734e-01
9.14033473e-01 5.20462573e-01 4.60362196e-01 -4.86116201e-01
-3.62519711e-01 3.11611116e-01 -5.90269506e-01 1.95417866e-01
6.95623696e-01 -4.85088862e-02 1.38371428e-02 3.47179413e-01
-2.20784247e-01 -5.78818560e-01 -3.56291294e-01 1.67504326e-01
9.34129238e-01 4.93956029e-01 -6.96308836e-02 -3.75282079e-01
3.32207441e-01 -4.14749235e-01 6.43256187e-01 3.80214542e-01
5.68812490e-01 1.05621982e+00 3.82984877e-01 -4.17343259e-01
-1.52082717e+00 -1.45886278e+00 -9.97240916e-02 5.51502526e-01
-2.60380246e-02 -6.00473464e-01 -2.80443430e-01 -1.00949772e-01
3.38653810e-02 7.19314039e-01 -2.48344541e-01 2.69761026e-01
-8.25075448e-01 -1.88946515e-01 5.11591554e-01 5.03115356e-01
4.10988331e-01 -6.89767003e-01 -1.63203984e-01 3.60761911e-01
-2.97555536e-01 -1.13518667e+00 -7.20557690e-01 -1.98951572e-01
-1.37055528e+00 -7.03842342e-01 -7.58793414e-01 -7.15167165e-01
8.70241463e-01 8.24441835e-02 1.38375902e+00 3.26886594e-01
1.94583982e-02 4.11513478e-01 -1.56706125e-02 4.77691144e-01
-6.42989159e-01 8.12355950e-02 -2.94539988e-01 -2.25105599e-01
-6.84432566e-01 -1.37281799e+00 -6.55795634e-01 1.54160768e-01
-9.15434539e-01 8.16654384e-01 5.28970182e-01 7.08913743e-01
1.05764890e+00 -1.45783469e-01 1.77691564e-01 -5.26341677e-01
2.75022596e-01 -9.77309644e-02 -6.01022661e-01 -6.51296042e-03
-2.41289020e-01 3.31331998e-01 5.40572941e-01 7.82250613e-02
-1.06729341e+00 -5.53671047e-02 -6.62201107e-01 -7.49575734e-01
9.92677435e-02 4.12566841e-01 -3.76417547e-01 3.27458270e-02
2.90966749e-01 3.58608246e-01 -1.69176668e-01 -1.05030084e+00
3.86350155e-01 1.70638457e-01 4.21818763e-01 -1.00346947e+00
7.19871819e-01 6.28967583e-01 3.25620890e-01 -7.50572979e-01
-5.79421580e-01 2.08632462e-02 -6.77095890e-01 9.02142897e-02
8.13059509e-01 -1.35479581e+00 -6.63152516e-01 3.22761834e-01
-1.47937167e+00 -5.51464617e-01 -2.46642947e-01 8.77549201e-02
-7.27152944e-01 6.25086188e-01 -9.77778375e-01 -5.93533397e-01
-5.39073765e-01 -1.36685026e+00 1.81449211e+00 -4.65339959e-01
-8.15399922e-03 -8.70818317e-01 -3.29042226e-01 2.14953944e-01
2.49537215e-01 4.50710833e-01 1.07110512e+00 4.36070353e-01
-1.32225227e+00 2.75545031e-01 -1.94411874e-01 1.20334774e-01
1.18022740e-01 -3.48057032e-01 -8.04125309e-01 -6.02731884e-01
-3.77364196e-02 -2.51769334e-01 9.20014203e-01 4.15963084e-01
1.26040804e+00 -4.27889675e-01 -1.54063866e-01 1.22899818e+00
1.31750655e+00 -3.37927848e-01 8.32692027e-01 -3.55139434e-01
1.05883121e+00 1.63278088e-01 1.32094696e-01 5.80883265e-01
7.21074104e-01 9.10207331e-01 6.09811902e-01 -1.99023187e-01
-6.33929610e-01 -7.21156836e-01 3.87836605e-01 1.45509899e+00
-2.30197459e-01 -4.63294238e-02 -8.71913135e-01 4.09214407e-01
-1.65432858e+00 -5.65769255e-01 -1.88449159e-01 2.04426241e+00
8.47224653e-01 -8.58570486e-02 -3.64058167e-01 -2.29573429e-01
2.57033378e-01 3.62407327e-01 -6.55489087e-01 -1.81833714e-01
-1.28640890e-01 3.69706482e-01 2.96665847e-01 9.63662922e-01
-6.30774856e-01 1.14586449e+00 6.05001259e+00 1.02609491e+00
-8.06374013e-01 2.30870619e-01 8.54998231e-01 -4.02996913e-02
-1.16680443e+00 1.68627486e-01 -6.29015267e-01 1.00594595e-01
2.47145012e-01 2.19012827e-01 6.50704265e-01 6.50868773e-01
1.92201793e-01 2.74323761e-01 -1.11605012e+00 1.07691252e+00
-1.27012581e-01 -1.63082945e+00 6.22512996e-01 4.08272892e-01
9.66436863e-01 2.71880720e-02 4.39959243e-02 1.55566305e-01
1.81732059e-01 -1.23411953e+00 8.73781264e-01 6.65041268e-01
1.50723660e+00 -7.84132183e-01 -1.81580447e-02 6.24679148e-01
-1.72081876e+00 3.87040824e-01 -4.87985879e-01 -3.18377689e-02
5.90639651e-01 7.40745723e-01 -4.24185425e-01 8.40640604e-01
2.62352765e-01 9.80328143e-01 -1.12243466e-01 5.16232610e-01
-2.91228235e-01 4.30080563e-01 -3.81313950e-01 6.84212804e-01
3.18718068e-02 -3.01453054e-01 8.02363336e-01 7.77096629e-01
6.67433858e-01 2.69560814e-01 5.29303908e-01 1.17207813e+00
-1.05906062e-01 -5.64883240e-02 -5.20061672e-01 5.07217906e-02
4.57161188e-01 8.05077851e-01 -6.46624506e-01 -1.90829530e-01
-2.91075021e-01 1.34228206e+00 5.55471420e-01 5.04842281e-01
-5.72711587e-01 4.88361090e-01 7.97663927e-01 4.06902701e-01
4.86661315e-01 -9.60635841e-01 -7.79379010e-01 -1.45570874e+00
1.52087456e-03 -5.61932683e-01 -5.47979064e-02 -9.66571093e-01
-9.73026693e-01 4.41723883e-01 -1.60039529e-01 -1.08827209e+00
-9.77494344e-02 -3.35091114e-01 -1.65867209e-01 1.15917611e+00
-1.52204275e+00 -1.47754157e+00 -9.56304893e-02 5.24123967e-01
7.13033020e-01 2.52812300e-02 9.95280027e-01 3.98925245e-01
1.47504613e-01 4.28931028e-01 -1.95807278e-01 -3.82264033e-02
1.14071012e-01 -9.76916552e-01 8.35447609e-01 7.08624363e-01
-7.99250379e-02 6.03714287e-01 2.56868571e-01 -1.16869903e+00
-1.79396617e+00 -1.14211833e+00 5.77668190e-01 -3.64214599e-01
-2.12461557e-02 -4.18206811e-01 -8.91069114e-01 7.79042006e-01
-2.44493097e-01 -2.12364510e-01 3.37419927e-01 -4.92791347e-02
-3.80833209e-01 2.50283420e-01 -9.57290828e-01 7.70685673e-01
1.41620553e+00 -7.78475404e-01 -2.71015584e-01 1.18634753e-01
9.74757731e-01 -9.02230322e-01 -1.13497913e+00 7.34112084e-01
4.73015547e-01 -9.52567458e-01 1.35908306e+00 2.05573142e-02
8.25963795e-01 -4.17104512e-01 -4.72891331e-01 -8.63896430e-01
-6.27909839e-01 -7.94415951e-01 -5.90400994e-01 9.34139252e-01
1.49550170e-01 -3.75993311e-01 8.51116240e-01 4.24891889e-01
-7.05150664e-01 -1.13735342e+00 -1.01803732e+00 -4.12722647e-01
1.82490677e-01 -6.59045160e-01 9.52744544e-01 6.92390144e-01
-8.01132739e-01 4.19915840e-02 -5.36885619e-01 4.33290362e-01
8.40740502e-01 5.19378603e-01 7.53645003e-01 -1.03090835e+00
-6.56737268e-01 -4.14013833e-01 5.86722093e-03 -2.18720293e+00
-9.59567502e-02 -1.29429770e+00 -2.35868305e-01 -2.00855947e+00
2.05755718e-02 -5.74534714e-01 4.34145272e-01 5.00463784e-01
1.52289659e-01 4.83759969e-01 1.85237929e-01 3.90714020e-01
-1.67672545e-01 1.11745501e+00 1.90558779e+00 5.33837900e-02
-1.02070138e-01 -3.11298728e-01 -6.66836023e-01 7.14725971e-01
3.21113884e-01 -1.53974175e-01 -3.54130954e-01 -9.62508380e-01
4.64699179e-01 6.92117035e-01 4.81994271e-01 -6.23645663e-01
1.07569667e-02 9.86579806e-03 6.18396401e-01 -1.00790262e+00
7.44303226e-01 -6.47902608e-01 5.26439488e-01 1.78304121e-01
-1.61396340e-02 1.95946991e-02 2.95193613e-01 4.97383982e-01
-8.11433569e-02 2.37477005e-01 7.04831719e-01 -3.11154157e-01
-2.32249260e-01 8.87579322e-01 -1.58074930e-01 -3.88955884e-02
3.56485397e-01 -2.94296741e-01 1.44798160e-01 -5.46313167e-01
-9.06619728e-01 8.10186118e-02 9.22373533e-01 1.31971881e-01
9.47777271e-01 -1.69776106e+00 -9.87823308e-01 4.35067356e-01
-2.85962015e-01 6.12947583e-01 3.34125370e-01 5.65594494e-01
-7.32260585e-01 1.46736294e-01 3.03702503e-01 -8.34186077e-01
-8.71833563e-01 9.34257582e-02 1.88993067e-01 -5.91716766e-01
-1.17381895e+00 7.22945094e-01 9.00666475e-01 -5.21014631e-01
-1.71384905e-02 -2.59612799e-01 2.02713847e-01 -3.14833552e-01
3.01132500e-01 1.61402911e-01 2.67729070e-02 -6.36552393e-01
2.60602012e-02 8.54773462e-01 1.60404503e-01 -1.65671557e-01
1.50600576e+00 -1.29476190e-01 -3.22178990e-01 1.79478243e-01
1.24603617e+00 1.96756124e-01 -1.69926417e+00 -1.09111086e-01
-6.17983818e-01 -5.99873662e-01 2.34193698e-01 -5.27516663e-01
-1.02210093e+00 9.68892753e-01 -3.04875858e-02 -3.41952920e-01
1.03823936e+00 -5.95919881e-03 1.07945955e+00 9.11688283e-02
7.17330754e-01 -6.52791142e-01 2.12214515e-01 7.70678699e-01
1.38608515e+00 -6.11729622e-01 2.02988565e-01 -8.71111870e-01
-2.92669863e-01 1.27397251e+00 3.13060343e-01 -3.44268054e-01
6.51002407e-01 6.62335753e-01 -6.03823304e-01 -3.12590450e-01
-9.20632243e-01 -5.51775619e-02 6.19851410e-01 4.20299381e-01
5.69498360e-01 2.08087098e-02 -5.00684604e-02 4.53035474e-01
-1.57881752e-01 -1.44867927e-01 1.90188646e-01 4.90830481e-01
-2.05636799e-01 -1.24237943e+00 -2.30627835e-01 3.72055352e-01
-4.89790589e-02 -3.26933116e-01 7.79332742e-02 2.93916464e-01
-1.03899948e-01 4.91489530e-01 5.32107465e-02 -5.98356538e-02
2.38599718e-01 -1.42955214e-01 6.91734314e-01 -7.90682316e-01
-2.82401383e-01 5.82337618e-01 1.38953358e-01 -7.88194776e-01
1.28706858e-01 -5.06503642e-01 -1.26722121e+00 -4.81353521e-01
1.85119823e-01 -2.22484827e-01 3.77476543e-01 7.30111957e-01
6.99084759e-01 7.01847494e-01 5.31003594e-01 -1.35939932e+00
-2.67569244e-01 -6.33429468e-01 -3.76016349e-01 1.88766867e-01
3.34164053e-01 -6.16445959e-01 -1.27389312e-01 -3.51906731e-03] | [8.984600067138672, -3.499633312225342] |
72edddfc-7f01-47ec-a74e-4facd223e298 | m2fnet-multi-modal-fusion-network-for-emotion | 2206.02187 | null | https://arxiv.org/abs/2206.02187v1 | https://arxiv.org/pdf/2206.02187v1.pdf | M2FNet: Multi-modal Fusion Network for Emotion Recognition in Conversation | Emotion Recognition in Conversations (ERC) is crucial in developing sympathetic human-machine interaction. In conversational videos, emotion can be present in multiple modalities, i.e., audio, video, and transcript. However, due to the inherent characteristics of these modalities, multi-modal ERC has always been considered a challenging undertaking. Existing ERC research focuses mainly on using text information in a discussion, ignoring the other two modalities. We anticipate that emotion recognition accuracy can be improved by employing a multi-modal approach. Thus, in this study, we propose a Multi-modal Fusion Network (M2FNet) that extracts emotion-relevant features from visual, audio, and text modality. It employs a multi-head attention-based fusion mechanism to combine emotion-rich latent representations of the input data. We introduce a new feature extractor to extract latent features from the audio and visual modality. The proposed feature extractor is trained with a novel adaptive margin-based triplet loss function to learn emotion-relevant features from the audio and visual data. In the domain of ERC, the existing methods perform well on one benchmark dataset but not on others. Our results show that the proposed M2FNet architecture outperforms all other methods in terms of weighted average F1 score on well-known MELD and IEMOCAP datasets and sets a new state-of-the-art performance in ERC. | ['Naoyuki Onoe', 'Pankaj Wasnik', 'Nirmesh Shah', 'Ashish Gudmalwar', 'Purbayan Kar', 'Vishal Chudasama'] | 2022-06-05 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 2.17139423e-01 -3.50603878e-01 8.32327455e-02 -3.69980872e-01
-1.20128548e+00 -1.92082182e-01 5.38575232e-01 -6.58023804e-02
-4.83836144e-01 5.53590059e-01 5.20688951e-01 3.05658996e-01
-3.10759321e-02 -5.58813252e-02 -3.56051773e-01 -7.33331680e-01
1.80031881e-01 2.26902161e-02 -3.70310158e-01 -1.28451943e-01
8.81635025e-03 3.10789704e-01 -1.96973646e+00 7.52055824e-01
6.93525672e-01 1.62400770e+00 -1.83947459e-01 7.65844047e-01
-2.36159176e-01 8.00016642e-01 -5.07101595e-01 -6.20761991e-01
-2.15527222e-01 -5.17622471e-01 -6.27879739e-01 1.86290532e-01
-8.69658403e-03 2.69111041e-02 -9.68221277e-02 6.87989354e-01
8.21484029e-01 4.75342542e-01 7.53072858e-01 -1.61894238e+00
-2.56939232e-01 2.01150924e-01 -5.30175090e-01 -5.51111400e-02
6.14075780e-01 -7.02463835e-02 1.00459170e+00 -1.25428391e+00
5.02148509e-01 1.25947106e+00 4.65992540e-01 7.04324484e-01
-7.81836927e-01 -6.12091303e-01 1.52349487e-01 4.73752499e-01
-1.29508936e+00 -6.62496626e-01 1.14426053e+00 -2.70221412e-01
1.05917907e+00 3.44413072e-01 4.60259795e-01 1.46001887e+00
1.96612030e-01 1.19464064e+00 1.09001684e+00 -5.11909962e-01
8.49775213e-04 5.07866621e-01 -1.54458180e-01 4.54135954e-01
-7.64970839e-01 -1.18792906e-01 -1.21196103e+00 -5.83760142e-02
-1.16031826e-01 7.24527985e-02 -4.00634021e-01 -4.50140759e-02
-9.50583100e-01 7.80792594e-01 1.14849128e-01 4.30407315e-01
-5.47131836e-01 -5.33707738e-02 8.56716871e-01 3.53174716e-01
7.57146180e-01 -1.20713629e-01 -4.45536315e-01 -6.46785021e-01
-7.62048721e-01 -6.57096058e-02 6.24846101e-01 4.49781537e-01
2.93861479e-01 3.42823588e-03 -2.38609582e-01 1.11631048e+00
5.01828074e-01 3.26372355e-01 5.78159213e-01 -7.14178324e-01
6.12896323e-01 6.81117892e-01 -1.82668224e-01 -1.17742538e+00
-4.69031245e-01 -1.33515358e-01 -8.88016164e-01 7.72368237e-02
-1.71977684e-01 -4.15747881e-01 -5.20219386e-01 1.84576917e+00
4.20008987e-01 2.09679514e-01 4.05388862e-01 8.90613496e-01
1.29494798e+00 7.64891744e-01 2.07417250e-01 -3.75948250e-01
1.28987777e+00 -1.07927513e+00 -1.11603463e+00 5.25771976e-02
4.56612527e-01 -8.51982117e-01 7.65139222e-01 5.39660037e-01
-9.34841096e-01 -4.75905776e-01 -8.91284049e-01 8.86214338e-03
-6.20255411e-01 3.33389789e-01 3.58685702e-01 5.39241076e-01
-7.46066272e-01 1.81473747e-01 -5.16834736e-01 -3.44407856e-01
1.14327073e-01 2.14792207e-01 -8.11093509e-01 1.86182912e-02
-1.25310612e+00 9.03247178e-01 3.32263857e-01 4.52520162e-01
-6.56548977e-01 -2.63808131e-01 -9.48151708e-01 1.34351730e-01
3.09502512e-01 -3.77644032e-01 9.00926650e-01 -1.35604227e+00
-1.77198970e+00 6.66045785e-01 -2.81599224e-01 -1.54218584e-01
2.92799324e-01 -3.45823139e-01 -8.50591719e-01 4.19172853e-01
-3.78238648e-01 7.16640413e-01 9.68122840e-01 -1.23531270e+00
-7.68701553e-01 -3.49571854e-01 -2.53957957e-01 5.48044086e-01
-6.68832541e-01 3.43842506e-01 -4.62004900e-01 -5.66068828e-01
-2.50015229e-01 -8.44979763e-01 3.08104247e-01 -2.30342746e-01
-2.64864177e-01 -3.78137857e-01 1.22656500e+00 -6.01552606e-01
1.23113537e+00 -2.54541826e+00 4.91028160e-01 -2.81222127e-02
9.66467634e-02 1.57101706e-01 -2.50601739e-01 4.55594331e-01
-1.60077646e-01 -7.84313679e-02 5.21106124e-02 -9.20233190e-01
3.18710804e-01 -1.64642036e-01 2.61748787e-02 3.47566277e-01
3.42362732e-01 7.64222920e-01 -4.36601520e-01 -7.07345605e-01
4.09749418e-01 1.01883399e+00 -3.28052312e-01 3.67536038e-01
1.49660438e-01 4.90784883e-01 -2.10379317e-01 7.40275323e-01
3.82236600e-01 -4.57004383e-02 -1.10426813e-01 -4.04024899e-01
5.64563926e-03 -2.51808017e-01 -1.14052236e+00 1.77342308e+00
-7.67213881e-01 9.00095820e-01 1.64929613e-01 -1.16475976e+00
8.60542238e-01 1.03061056e+00 6.64071321e-01 -4.74653155e-01
5.25322855e-01 1.46448836e-01 -2.61801362e-01 -7.26424396e-01
4.65823233e-01 -3.65896910e-01 -3.59814495e-01 2.94042647e-01
5.10999918e-01 2.02114880e-01 -8.97079185e-02 7.76377395e-02
7.27436960e-01 -1.25084966e-01 1.66882619e-01 4.35011268e-01
8.60596955e-01 -5.04554927e-01 5.49177170e-01 2.57348776e-01
-5.11364937e-01 4.25844461e-01 2.98849642e-01 -9.14036483e-02
-4.41371858e-01 -6.30596101e-01 3.61857489e-02 1.34398031e+00
-7.94689581e-02 -5.39860010e-01 -4.80488688e-01 -8.01509798e-01
-1.92373648e-01 4.25684780e-01 -7.11613715e-01 -4.17012632e-01
-1.11630587e-02 -4.84004706e-01 5.47623038e-01 4.48817343e-01
4.09230471e-01 -1.35536528e+00 -4.41433698e-01 1.64667353e-01
-7.29063511e-01 -1.34291589e+00 -4.19692129e-01 2.47710854e-01
-3.33279788e-01 -6.79989994e-01 -6.67446017e-01 -5.45419991e-01
1.59149796e-01 2.81323586e-02 7.79874146e-01 -2.52412140e-01
-3.41554970e-01 9.41529393e-01 -7.31392384e-01 -5.29167175e-01
-3.19348387e-02 -8.79311711e-02 -1.39891192e-01 8.24114978e-01
4.97282267e-01 -3.17719758e-01 -3.79669189e-01 1.06727995e-01
-8.42021406e-01 -2.78939661e-02 4.16373491e-01 8.98467481e-01
3.09956253e-01 1.30686432e-01 7.98523724e-01 -2.96009898e-01
6.74658537e-01 -6.51373863e-01 1.62238047e-01 2.96023071e-01
-7.08051994e-02 -2.94154406e-01 4.95294780e-01 -5.63862443e-01
-1.43099904e+00 1.09317414e-01 -2.12033778e-01 -6.68949604e-01
-3.07803035e-01 8.07239532e-01 -3.69910091e-01 7.80064464e-02
3.22972238e-02 8.17515180e-02 -1.56782717e-01 -2.16093600e-01
2.67302513e-01 1.09746337e+00 3.88851762e-01 -4.93873745e-01
1.20309740e-01 2.72221476e-01 -1.79354131e-01 -9.71122801e-01
-6.53276443e-01 -6.49329305e-01 -1.77549243e-01 -8.24327350e-01
1.02462828e+00 -1.03727353e+00 -1.00226593e+00 5.87834179e-01
-1.11803734e+00 2.56512046e-01 3.98827307e-02 8.78458202e-01
-5.43565750e-01 1.68395370e-01 -5.39315164e-01 -1.24022770e+00
-4.80100155e-01 -1.19829977e+00 1.32877028e+00 2.93046176e-01
-2.88761735e-01 -1.01874506e+00 1.05473222e-02 7.07673728e-01
3.87730777e-01 4.40024376e-01 5.41358650e-01 -7.83192754e-01
-2.73723658e-02 -4.33173001e-01 -9.18221995e-02 5.20012558e-01
-8.88351817e-03 5.26450109e-04 -1.30572248e+00 -1.83221251e-01
-7.42665434e-04 -8.92669916e-01 1.00159431e+00 1.97489917e-01
1.07063496e+00 1.02151847e-02 -6.04523011e-02 3.16141307e-01
1.13437963e+00 2.63750732e-01 4.26056385e-01 3.22581939e-02
5.20629525e-01 8.15435171e-01 6.64940417e-01 7.90598691e-01
4.13899183e-01 7.14941442e-01 4.29938585e-01 -6.44346848e-02
2.12956920e-01 2.01610491e-01 6.16451323e-01 1.05161154e+00
-3.07593234e-02 -5.86934805e-01 -5.65634489e-01 5.05040228e-01
-2.03242493e+00 -1.05699241e+00 2.18158662e-01 1.60521615e+00
6.38725698e-01 -1.10912830e-01 1.21285550e-01 3.32933635e-01
6.79861963e-01 1.79650053e-01 -4.12224531e-01 -6.13398671e-01
-3.01592827e-01 8.90044719e-02 -3.08432847e-01 2.14927897e-01
-1.31573892e+00 5.11557996e-01 5.02927208e+00 9.97391105e-01
-1.29635227e+00 2.36128628e-01 5.15477717e-01 -4.35924828e-01
-1.57928407e-01 -5.29048502e-01 -6.18149161e-01 4.18732077e-01
1.14144993e+00 9.12066549e-02 2.95342118e-01 7.21600354e-01
1.67005256e-01 -8.78179073e-03 -8.66778433e-01 1.50517380e+00
5.58822989e-01 -8.28076243e-01 -2.54047662e-01 -1.68615475e-01
3.82923007e-01 -2.72415400e-01 2.29066670e-01 6.20075643e-01
-3.60764563e-01 -1.07819283e+00 5.72898090e-01 7.27721751e-01
6.26516700e-01 -1.07085645e+00 9.89683449e-01 1.90782547e-01
-1.34549356e+00 -1.20691776e-01 2.94131875e-01 2.29138300e-01
3.20986211e-01 3.96568388e-01 -2.28944764e-01 6.92432284e-01
8.57605636e-01 7.54066229e-01 -1.71614394e-01 7.16878474e-01
1.64489716e-01 4.27992493e-01 -2.60139167e-01 -3.67855243e-02
1.68155581e-01 9.30777639e-02 4.44411635e-01 1.49302924e+00
4.59210187e-01 2.93181725e-02 6.80207536e-02 2.79503971e-01
-1.91313758e-01 4.99305487e-01 -5.58329761e-01 -2.39068240e-01
2.24231735e-01 1.63512790e+00 -4.14455414e-01 -2.87665814e-01
-6.48970723e-01 1.12261665e+00 2.93266535e-01 2.44509205e-01
-9.88638818e-01 -5.76519370e-01 4.62837130e-01 -8.46432269e-01
3.87182772e-01 2.29338363e-01 2.14863241e-01 -1.22043848e+00
1.27782360e-01 -9.96029913e-01 4.86383647e-01 -7.65897095e-01
-1.35535061e+00 8.91285896e-01 -2.36691579e-01 -1.31556463e+00
-2.88369060e-01 -5.98632872e-01 -5.33284843e-01 6.15685463e-01
-1.50046659e+00 -1.30156195e+00 -3.09303850e-01 9.17553246e-01
6.13754094e-01 -3.58524829e-01 9.53345954e-01 6.15625083e-01
-6.89049721e-01 7.82465756e-01 4.04184647e-02 -1.08510517e-02
9.96342003e-01 -1.01148343e+00 -6.32827580e-01 4.23314005e-01
2.32100748e-02 2.38038599e-01 5.06301522e-01 -1.61719576e-01
-1.38742614e+00 -8.49913597e-01 8.77264559e-01 -3.72146107e-02
4.04513508e-01 -3.01621765e-01 -6.88249648e-01 4.31219667e-01
6.26982808e-01 1.26284957e-01 1.19598818e+00 1.84152186e-01
-3.21352780e-01 -9.25906450e-02 -1.10817647e+00 3.15546691e-01
4.90663230e-01 -9.00856435e-01 -3.88842553e-01 -1.19501360e-01
6.15835786e-01 -2.06300586e-01 -1.06742525e+00 5.85912347e-01
9.15581703e-01 -9.07316923e-01 7.88565576e-01 -5.98603129e-01
5.24218738e-01 -8.40558391e-03 -5.82266688e-01 -1.46347213e+00
2.07765654e-01 -5.40869176e-01 -2.38396019e-01 1.65082359e+00
2.64415503e-01 -3.34887266e-01 3.41736525e-01 4.31211889e-01
-1.59141630e-01 -7.58789718e-01 -1.07782590e+00 -2.75415808e-01
-3.62060696e-01 -7.19865799e-01 2.91998118e-01 9.40489233e-01
3.57624620e-01 7.06406116e-01 -8.51978779e-01 -1.30246460e-01
2.35468835e-01 4.21092398e-02 5.72703898e-01 -1.16658437e+00
-9.16843414e-02 -3.97728950e-01 -3.44072342e-01 -3.71677160e-01
6.16324961e-01 -7.06396043e-01 1.31689653e-01 -1.20126402e+00
2.01528952e-01 1.78189814e-01 -6.25801504e-01 2.96814084e-01
-1.77118272e-01 4.28690165e-01 5.03162801e-01 -2.76396036e-01
-9.90820348e-01 1.11563396e+00 9.50473547e-01 -2.17284679e-01
-1.03892982e-01 -2.84062743e-01 -4.12606001e-01 7.71891475e-01
7.22056985e-01 -3.09317440e-01 -2.45117813e-01 9.30077955e-02
1.44934759e-01 3.93787563e-01 2.81046957e-01 -9.57046568e-01
2.51728565e-01 1.81368981e-02 4.37495887e-01 -7.82425284e-01
1.01865923e+00 -1.06854725e+00 -6.83615636e-03 -1.63591892e-01
-4.98504043e-01 -7.46188611e-02 3.57902378e-01 5.50697565e-01
-7.79866576e-01 -2.34689210e-02 6.04626954e-01 1.75129890e-01
-7.11423576e-01 1.60377212e-02 -5.42230129e-01 -3.38055901e-02
9.74548757e-01 -7.83154890e-02 -1.79027453e-01 -5.60408115e-01
-9.97677445e-01 2.19876915e-01 -1.97783977e-01 6.75348938e-01
8.31165969e-01 -1.66458726e+00 -6.88562751e-01 6.87274477e-03
4.64440554e-01 -6.43326640e-01 8.15212905e-01 1.13588810e+00
2.87218183e-01 3.67935777e-01 -3.01491320e-01 -4.44944054e-01
-1.67620754e+00 3.51280957e-01 2.16642216e-01 -3.66343558e-01
-1.19021900e-01 7.93379903e-01 4.52467985e-02 -5.64323604e-01
6.08890295e-01 1.89848348e-01 -5.60764015e-01 5.38103282e-01
5.40162325e-01 3.18962812e-01 1.41131297e-01 -1.13029063e+00
-4.81145650e-01 6.24730110e-01 6.13424070e-02 -2.92705268e-01
1.34362495e+00 -3.58493388e-01 7.31264278e-02 8.65009367e-01
1.66140795e+00 -1.32917851e-01 -8.04094970e-01 -3.16016108e-01
-2.16278717e-01 -1.68338686e-01 2.49370128e-01 -8.53984594e-01
-1.16301858e+00 1.09218335e+00 8.71413827e-01 5.13942093e-02
1.55177891e+00 -1.29048035e-01 7.92302191e-01 1.08078174e-01
-2.10185558e-01 -1.42578232e+00 4.77615207e-01 4.49291706e-01
1.01128471e+00 -1.50494361e+00 -4.46278393e-01 -3.76683548e-02
-1.11968708e+00 1.06786764e+00 6.34507537e-01 4.03539717e-01
8.43701184e-01 1.71494856e-01 1.92573175e-01 -2.22297639e-01
-1.15477896e+00 -3.83660018e-01 6.69370949e-01 2.83284217e-01
6.89823449e-01 -7.05283582e-02 -2.62485981e-01 9.35435593e-01
2.09629148e-01 7.57895261e-02 7.33879134e-02 9.31512535e-01
-1.25203893e-01 -9.65555668e-01 -4.52136129e-01 2.23932087e-01
-7.55880177e-01 1.21819995e-01 -3.71995807e-01 5.68794727e-01
1.21959284e-01 1.40694857e+00 -1.03894085e-01 -6.12270474e-01
2.59650916e-01 5.87914944e-01 2.02250436e-01 -4.46997769e-02
-7.21698582e-01 2.72713393e-01 2.26529390e-01 -5.56628883e-01
-9.97245014e-01 -6.18237495e-01 -1.01841819e+00 -9.71241742e-02
-4.32404101e-01 2.83839136e-01 9.14503098e-01 1.18120146e+00
5.18574357e-01 7.79746115e-01 7.88656116e-01 -9.94343996e-01
-9.03592706e-02 -1.11015487e+00 -6.20793521e-01 5.53152382e-01
5.04120946e-01 -9.06928957e-01 -5.22049129e-01 1.96157433e-02] | [13.262347221374512, 5.127523899078369] |
2fe79d72-627d-42b2-992f-c1329f79c8cd | open-set-rf-fingerprinting-via-improved | 2306.13895 | null | https://arxiv.org/abs/2306.13895v1 | https://arxiv.org/pdf/2306.13895v1.pdf | Open-Set RF Fingerprinting via Improved Prototype Learning | Deep learning has been widely used in radio frequency (RF) fingerprinting. Despite its excellent performance, most existing methods only consider a closed-set assumption, which cannot effectively tackle signals emitted from those unknown devices that have never been seen during training. In this letter, we exploit prototype learning for open-set RF fingerprinting and propose two improvements, including consistency-based regularization and online label smoothing, which aim to learn a more robust feature space. Experimental results on a real-world RF dataset demonstrate that our proposed measures can significantly improve prototype learning to achieve promising open-set recognition performance for RF fingerprinting. | ['Lu Gan', 'Hongshu Liao', 'Weidong Wang'] | 2023-06-24 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [ 5.81187069e-01 -2.60899097e-01 -6.24523103e-01 -9.07268286e-01
-6.78520381e-01 -6.35103285e-01 2.55131155e-01 -3.78692925e-01
-2.19547004e-01 7.90242910e-01 -1.72628880e-01 -5.14692426e-01
-7.13411510e-01 -9.16879952e-01 -8.82225454e-01 -5.03234625e-01
-2.34813318e-01 1.37941778e-01 -3.89046699e-01 3.66083831e-01
2.11388124e-05 4.05374229e-01 -1.15095830e+00 1.28089100e-01
7.85400033e-01 1.35813522e+00 -7.09781349e-02 1.58255935e-01
-7.54642785e-02 5.64023674e-01 -5.61088502e-01 -5.75382113e-02
3.11434329e-01 4.98687662e-03 -6.46271229e-01 -5.23104727e-01
7.28740513e-01 -3.31864834e-01 -5.85619688e-01 8.74798119e-01
8.11030447e-01 8.68022293e-02 5.27077377e-01 -8.04985702e-01
-8.50343049e-01 7.69594193e-01 -5.13039768e-01 -6.12098500e-02
4.46172029e-01 -5.21339476e-01 6.81979179e-01 -4.45230305e-01
3.18830788e-01 1.09562600e+00 1.21810627e+00 6.18745148e-01
-1.21030104e+00 -1.44202232e+00 -2.16502044e-02 -4.53784838e-02
-1.68523896e+00 -4.49958295e-01 8.12861383e-01 4.31760401e-02
4.80871379e-01 2.56397337e-01 9.86560881e-02 1.53400779e+00
7.67791178e-03 5.98992407e-01 1.17156661e+00 -1.35843948e-01
2.26653423e-02 1.07020430e-01 1.54677154e-02 5.83393514e-01
4.07403588e-01 6.37032747e-01 -1.69519141e-01 -6.16289973e-01
9.87662673e-01 4.23658162e-01 -2.07173467e-01 -4.57621038e-01
-1.18980050e+00 6.57930851e-01 6.52204096e-01 5.62690973e-01
-1.87363878e-01 3.44701350e-01 1.82581261e-01 4.98638123e-01
1.54649720e-01 2.48631403e-01 -3.94679785e-01 2.47119561e-01
-9.71870005e-01 1.32415798e-02 7.84184813e-01 8.72380912e-01
8.87136817e-01 -5.75510524e-02 -4.05242562e-01 1.09667552e+00
4.23516244e-01 1.34645808e+00 2.94646025e-01 -6.88145697e-01
3.95781100e-01 1.95868701e-01 2.52661318e-01 -1.08489275e+00
-6.75145507e-01 -1.03326237e+00 -1.04536200e+00 -5.61681509e-01
1.10508539e-01 -2.76856512e-01 -8.93807650e-01 1.73821819e+00
8.88610557e-02 9.42436337e-01 -1.09756991e-01 6.74144208e-01
9.01917636e-01 2.25106612e-01 -8.05300847e-02 4.43324596e-02
9.22165394e-01 -4.99517232e-01 -4.52960193e-01 -3.54356393e-02
5.93818963e-01 -5.45437038e-01 8.03549886e-01 3.42466414e-01
-1.84781954e-01 -7.97651350e-01 -1.39727044e+00 4.52930301e-01
-1.17726430e-01 8.65161791e-02 1.25066233e+00 1.33047163e+00
-5.19542873e-01 6.18908882e-01 -3.35462302e-01 -5.68316169e-02
8.56323957e-01 7.39725292e-01 -2.21362576e-01 -3.28222483e-01
-1.28416145e+00 2.64927447e-01 7.47397020e-02 3.99634689e-01
-7.15007484e-01 -5.73311388e-01 -5.54443479e-01 -1.32658482e-02
3.91525626e-01 -6.56157792e-01 1.07809222e+00 -5.01295328e-01
-1.58070064e+00 5.11535347e-01 -2.17831191e-02 -2.81780392e-01
2.81111926e-01 -1.12103214e-02 -1.12506080e+00 -3.03429186e-01
-1.63099930e-01 7.77562931e-02 9.69984114e-01 -1.20309412e+00
-3.44326228e-01 -3.97175670e-01 -3.51684652e-02 -5.06052911e-01
-4.94614124e-01 -4.17032182e-01 -4.43473682e-02 -6.30294800e-01
3.63717735e-01 -9.86663342e-01 -1.44073620e-01 -3.85464430e-01
-2.58992076e-01 -8.58184993e-02 7.84122348e-01 -1.49314031e-01
1.28248084e+00 -1.89083815e+00 -7.40218997e-01 9.41033542e-01
-6.86898828e-02 4.65578109e-01 -2.82164782e-01 -7.22606480e-02
1.82171494e-01 8.23493898e-02 -1.17621168e-01 -5.12800589e-02
1.08736880e-01 1.61355466e-01 -6.13720775e-01 5.84250271e-01
-2.78475225e-01 9.91402268e-01 -9.37335849e-01 -3.14012617e-01
1.76098317e-01 6.90178454e-01 -2.30905369e-01 -5.07482141e-02
1.11060143e-02 7.28279173e-01 -9.53677535e-01 7.17339277e-01
1.24992871e+00 -3.48571718e-01 2.33637542e-01 -2.86216676e-01
2.71758854e-01 8.73275623e-02 -1.31243622e+00 1.74103248e+00
-9.10565138e-01 1.68205604e-01 -1.79230571e-01 -1.10661066e+00
1.28486335e+00 9.50609967e-02 6.96961582e-01 -1.27085507e+00
2.51869429e-02 3.57089639e-01 -2.38153994e-01 -4.45846140e-01
2.09984347e-01 1.73016693e-02 -3.88177514e-01 6.70195282e-01
-4.10814211e-02 4.24714446e-01 -4.53461140e-01 -4.29750711e-01
1.10668135e+00 4.27764803e-02 -2.46758252e-01 -2.41841725e-03
7.65948415e-01 -6.69770896e-01 7.06358612e-01 1.60940337e+00
-4.34787236e-02 3.19005400e-01 -1.64400086e-01 -5.19127905e-01
-5.37967622e-01 -9.90722120e-01 -7.92701900e-01 1.02769029e+00
4.59868580e-01 1.18008658e-01 -2.58599222e-01 -1.04587150e+00
4.21302974e-01 1.24953434e-01 -7.12874234e-01 -1.38180777e-01
-7.99946249e-01 -6.65802121e-01 1.19798613e+00 5.04044235e-01
5.19612849e-01 -8.27071667e-01 1.25307739e-01 2.78474152e-01
-1.95887610e-01 -8.66861761e-01 -3.95173371e-01 6.93973452e-02
-6.98044717e-01 -9.10612762e-01 -7.95245767e-01 -9.28909600e-01
5.60206831e-01 5.17124534e-01 7.15116084e-01 1.53148517e-01
-2.35308573e-01 2.63597697e-01 -2.09741086e-01 -3.18969965e-01
-5.01996651e-02 2.77452618e-01 1.76816851e-01 3.41245472e-01
3.19141716e-01 -5.66919327e-01 -6.55957222e-01 7.57073164e-01
-5.87389112e-01 -8.26915860e-01 1.02922571e+00 1.01113915e+00
6.21618330e-01 -6.38864487e-02 9.61526394e-01 -1.13433158e+00
4.85669225e-01 -5.23600936e-01 -4.35782194e-01 6.88802600e-01
-7.97982395e-01 2.96552598e-01 7.36138701e-01 -5.78041375e-01
-1.00597763e+00 -1.19229011e-01 -2.33078837e-01 -5.78709394e-02
-5.39654978e-02 3.82888287e-01 -1.23973675e-01 -9.44500864e-01
6.91832840e-01 2.94220805e-01 -5.36094964e-01 -5.68469584e-01
2.43838206e-01 1.05401385e+00 5.44465959e-01 -8.52894962e-01
8.42388272e-01 7.32167661e-01 8.13501850e-02 -6.70784414e-01
-9.56338167e-01 -6.81497037e-01 -1.48956627e-01 1.28580019e-01
5.60540371e-02 -8.24489653e-01 -1.17922604e+00 1.47933692e-01
-7.01710761e-01 -1.39947578e-01 1.16897590e-01 7.15528190e-01
-4.20789450e-01 5.27324855e-01 -8.03184584e-02 -7.61053205e-01
-6.54920280e-01 -4.82235014e-01 1.00831103e+00 3.79544973e-01
2.26352319e-01 -8.15311491e-01 4.12569553e-01 8.23592767e-02
7.73592174e-01 -5.36152581e-03 4.51783597e-01 -4.74945009e-01
-4.04508948e-01 -4.65356231e-01 -4.60462302e-01 2.53794879e-01
2.61016160e-01 -5.88506937e-01 -1.21060169e+00 -5.84520400e-01
-2.37120226e-01 -2.91361600e-01 1.05903256e+00 2.66130656e-01
1.62841797e+00 -1.03645846e-01 -9.17358398e-01 1.31633055e+00
1.30135763e+00 1.45912543e-01 6.59422696e-01 -8.31391886e-02
6.26879334e-01 9.38271806e-02 7.37774789e-01 5.40353417e-01
5.81203364e-02 6.53619409e-01 6.68206066e-02 -1.75912380e-01
2.23530307e-01 -5.09275377e-01 -1.41095459e-01 1.23980574e-01
7.11934566e-02 -1.46730781e-01 -4.75086212e-01 1.68894660e-02
-1.86137378e+00 -1.04297304e+00 1.99426353e-01 2.44319916e+00
3.13083529e-01 1.14684708e-01 -6.59934208e-02 -3.77156697e-02
8.01617146e-01 1.97629541e-01 -7.58910298e-01 -5.95686100e-02
-1.63552478e-01 6.94309831e-01 9.38864470e-01 2.33035296e-01
-1.44983029e+00 4.27779406e-01 6.81482172e+00 8.04569781e-01
-1.40655494e+00 1.55072495e-01 1.99935019e-01 1.47439942e-01
-5.97949028e-01 -1.38111666e-01 -8.13018918e-01 5.50068378e-01
7.60396779e-01 1.63132712e-01 5.38907886e-01 7.81761289e-01
-1.82415381e-01 4.70156461e-01 -9.71818984e-01 1.35582292e+00
-1.97608292e-01 -1.30176771e+00 -2.71120220e-01 1.14717439e-01
8.03750515e-01 7.89101273e-02 4.26997930e-01 7.26401210e-01
-6.31099120e-02 -1.09112871e+00 3.17836553e-01 7.17986763e-01
8.97628903e-01 -8.19566667e-01 7.58663416e-01 7.82864764e-02
-1.25979531e+00 -3.78209323e-01 -6.93661571e-01 2.94005036e-01
-2.05716088e-01 7.35555947e-01 -7.69447863e-01 8.16060960e-01
5.67310810e-01 5.04550040e-01 -2.83489317e-01 1.18094563e+00
-8.67970660e-02 7.09260046e-01 -5.97363055e-01 -1.49636805e-01
-5.77314794e-02 1.49552494e-01 1.27671763e-01 1.17326367e+00
3.62240076e-01 -2.75355786e-01 1.87383875e-01 8.65665376e-01
-3.06831867e-01 -8.11201613e-03 -5.88165283e-01 -2.83444556e-03
9.36748087e-01 1.16846418e+00 -3.22811574e-01 1.14870943e-01
-4.56516981e-01 7.58837342e-01 5.29827811e-02 5.16032457e-01
-1.11323559e+00 -7.72599399e-01 4.48098540e-01 -3.27909999e-02
4.95835334e-01 -2.15223521e-01 -5.33926114e-02 -1.00673795e+00
1.59109354e-01 -4.52192068e-01 3.84978712e-01 3.69115137e-02
-1.48249590e+00 2.61339873e-01 -4.77702051e-01 -1.32189536e+00
-1.51040852e-01 -3.38633597e-01 -4.50058788e-01 6.12491488e-01
-1.56344390e+00 -1.43984580e+00 -4.51099575e-01 6.56471074e-01
-4.94801044e-01 -5.65174520e-02 1.08020818e+00 8.64671171e-01
-4.21859622e-01 1.41268718e+00 7.95620859e-01 1.45196378e-01
8.88391972e-01 -8.19400728e-01 5.16532362e-01 1.81525320e-01
4.14963692e-01 1.11427939e+00 7.65840486e-02 -5.76932788e-01
-1.69385993e+00 -1.18571591e+00 3.68095547e-01 -1.70164466e-01
2.35329270e-01 -6.04942739e-01 -6.62066698e-01 5.47964871e-01
-8.17415893e-01 6.35781586e-01 9.51600611e-01 5.67553997e-01
-7.57432282e-01 -7.21467853e-01 -1.36224937e+00 -7.69463181e-02
1.34145987e+00 -8.08014393e-01 -1.21416010e-01 2.42811456e-01
2.98466623e-01 -1.66966006e-01 -1.17654395e+00 8.34461272e-01
1.32217610e+00 -4.25109148e-01 1.39580870e+00 -4.21424568e-01
-6.27506673e-01 -2.86845088e-01 -1.54552743e-01 -7.06573129e-01
-4.41217184e-01 -7.35847652e-01 -3.21416974e-01 1.26107562e+00
3.10359120e-01 -9.60682929e-01 1.02610242e+00 -4.92636934e-02
2.27509141e-01 -4.32621211e-01 -8.52693677e-01 -1.32607555e+00
-2.59669602e-01 -5.16402364e-01 1.01531374e+00 1.10615849e+00
-2.42350638e-01 5.61918458e-03 -8.65118921e-01 1.82590574e-01
1.09357858e+00 4.89550740e-01 7.61748016e-01 -1.52218354e+00
-5.91429174e-01 -8.36887583e-02 -3.13939184e-01 -1.33953643e+00
3.48552495e-01 -1.00551438e+00 7.66755044e-02 -1.28056550e+00
1.64489776e-01 -1.31954205e+00 -9.92043614e-01 4.71672773e-01
1.44642845e-01 5.45067906e-01 -4.89568204e-01 1.36810333e-01
-7.96489239e-01 3.44218791e-01 1.00184548e+00 -5.18508792e-01
-1.85325190e-01 5.12386084e-01 -8.40821505e-01 3.97957742e-01
6.15670204e-01 -6.53088272e-01 -3.07582319e-01 -3.96666437e-01
1.71381339e-01 -7.56399557e-02 3.56451213e-01 -1.31224573e+00
-1.21524716e-02 -1.36123806e-01 8.29929054e-01 -2.96494722e-01
-4.14316133e-02 -8.17083240e-01 3.62541139e-01 4.64091301e-01
-3.10874790e-01 -9.18696344e-01 2.78255413e-03 9.25004542e-01
1.42163754e-01 -4.34070416e-02 5.11943340e-01 3.59038800e-01
-6.08767271e-01 7.29770541e-01 2.92563915e-01 -2.60027051e-01
5.30425191e-01 -1.72794208e-01 -4.01211739e-01 -2.45506749e-01
-4.06056702e-01 8.11036155e-02 1.71437383e-01 6.75998271e-01
5.29696941e-01 -1.54000711e+00 -5.46052754e-01 4.45375651e-01
3.07592452e-01 -3.91266972e-01 5.69942951e-01 5.37765920e-01
-2.27482449e-02 7.31122553e-01 -9.33354255e-03 -6.18024349e-01
-8.17805648e-01 6.30775273e-01 4.00480956e-01 -1.86273575e-01
-4.84803557e-01 8.77889097e-01 -1.00106820e-01 -9.56037223e-01
5.31750619e-01 -1.47687078e-01 -3.29054594e-02 -5.23376942e-01
5.85969329e-01 2.97027081e-01 1.25872105e-01 -3.66255105e-01
-7.65479207e-01 1.09464145e+00 -1.23065904e-01 3.48302484e-01
1.06368673e+00 2.37377677e-02 2.64684170e-01 4.92062308e-02
1.51318085e+00 1.51801258e-01 -9.00681853e-01 -5.92330635e-01
2.16562022e-02 -6.48868859e-01 6.03213906e-02 -8.52625310e-01
-7.77520359e-01 4.94084477e-01 1.24117160e+00 3.38389818e-03
8.04687142e-01 -1.52672097e-01 1.29005361e+00 1.09449828e+00
9.86500442e-01 -9.75202382e-01 -3.17168951e-01 4.10235465e-01
3.82655501e-01 -1.35723531e+00 1.54663578e-01 -3.10118556e-01
2.49997556e-01 9.54733968e-01 2.35618100e-01 1.65913627e-02
8.11837196e-01 2.71654129e-02 -8.21981393e-03 1.73580766e-01
8.77045691e-02 -5.18161990e-02 4.38544333e-01 8.61635506e-01
4.32240397e-01 3.64341915e-01 -1.51431859e-01 1.01128447e+00
5.46457954e-02 3.06901395e-01 -2.42921099e-01 9.58771229e-01
-4.47904795e-01 -1.48035836e+00 -3.42174023e-01 7.81065345e-01
-5.01820028e-01 3.81505281e-01 -1.44993514e-01 2.63099194e-01
2.92042375e-01 1.01104665e+00 -3.30268681e-01 -8.67348135e-01
3.22361261e-01 -4.95821893e-01 1.09262633e+00 -2.44238406e-01
-3.17681253e-01 -3.55306745e-01 -2.65541226e-02 -7.59017825e-01
-4.30879414e-01 -2.86694467e-01 -1.18171501e+00 -2.44715035e-01
-7.51129150e-01 1.79605961e-01 9.20814395e-01 8.96021903e-01
4.59555566e-01 3.14293057e-01 1.25495243e+00 -4.38259631e-01
-9.44590628e-01 -8.15948009e-01 -7.50539541e-01 2.19289839e-01
3.69080186e-01 -8.70578945e-01 -1.67917296e-01 -6.73129201e-01] | [6.502023220062256, 0.9273337125778198] |
0c9a2932-4e38-45da-b0f7-18f0f1a24278 | meta-curriculum-learning-for-domain | 2103.02262 | null | https://arxiv.org/abs/2103.02262v1 | https://arxiv.org/pdf/2103.02262v1.pdf | Meta-Curriculum Learning for Domain Adaptation in Neural Machine Translation | Meta-learning has been sufficiently validated to be beneficial for low-resource neural machine translation (NMT). However, we find that meta-trained NMT fails to improve the translation performance of the domain unseen at the meta-training stage. In this paper, we aim to alleviate this issue by proposing a novel meta-curriculum learning for domain adaptation in NMT. During meta-training, the NMT first learns the similar curricula from each domain to avoid falling into a bad local optimum early, and finally learns the curricula of individualities to improve the model robustness for learning domain-specific knowledge. Experimental results on 10 different low-resource domains show that meta-curriculum learning can improve the translation performance of both familiar and unfamiliar domains. All the codes and data are freely available at https://github.com/NLP2CT/Meta-Curriculum. | ['Lidia S. Chao', 'Derek F. Wong', 'Xuebo Liu', 'Runzhe Zhan'] | 2021-03-03 | null | null | null | null | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 2.23173007e-01 -8.34950283e-02 -6.03511274e-01 -3.54308605e-01
-1.06176925e+00 -7.34147906e-01 4.33386683e-01 -2.34761894e-01
-3.68886411e-01 1.03717792e+00 5.70864901e-02 -4.89070892e-01
1.94705620e-01 -4.69411820e-01 -1.07801044e+00 -5.34212649e-01
3.87293369e-01 8.61594975e-01 -1.07418820e-01 -5.24433076e-01
-5.22655286e-02 -4.84625921e-02 -8.98534715e-01 6.47975743e-01
1.43616700e+00 1.56126946e-01 5.68597734e-01 2.38355994e-01
-2.58303106e-01 1.25232995e-01 -7.44067311e-01 -6.09134436e-01
3.26856792e-01 -6.95537508e-01 -8.83325517e-01 -4.19577271e-01
5.79144955e-01 -1.54579803e-01 -2.35507026e-01 1.24974227e+00
6.73662901e-01 1.22651003e-01 6.36221588e-01 -9.81349707e-01
-1.10677898e+00 8.42438936e-01 -4.41868335e-01 2.59017646e-01
-4.99222279e-02 1.20215595e-01 7.02293634e-01 -9.98331547e-01
7.94380546e-01 1.16835618e+00 5.32141507e-01 1.10835600e+00
-1.07564604e+00 -1.07950521e+00 1.81993544e-01 2.01978296e-01
-1.18475306e+00 -2.85360366e-01 6.70272231e-01 -5.81158958e-02
1.08367598e+00 -1.25530988e-01 2.74537861e-01 1.64622533e+00
2.10819900e-01 8.69890273e-01 1.20027173e+00 -4.32070315e-01
9.95482132e-02 2.91297585e-01 -1.39516667e-01 4.57405418e-01
2.12428287e-01 2.20128372e-01 -4.44797695e-01 -1.62109613e-01
6.06085956e-01 -3.16902280e-01 -7.26243779e-02 -1.99660465e-01
-1.41504228e+00 5.16933560e-01 2.35418335e-01 4.99137372e-01
-2.00910315e-01 -1.25715196e-01 4.41177964e-01 7.92996526e-01
7.11473107e-01 7.61545122e-01 -1.08787346e+00 -2.21012712e-01
-6.58376932e-01 8.92878622e-02 4.59066689e-01 1.29270446e+00
7.60258973e-01 -1.30936950e-02 4.99843769e-02 1.15588129e+00
-2.53465891e-01 8.15349996e-01 7.95945406e-01 -7.51068354e-01
1.22350097e+00 6.17298186e-01 -1.00155594e-02 -3.46497595e-01
-6.14052564e-02 -7.21045077e-01 -8.40061367e-01 -2.38632455e-01
4.10630286e-01 -6.21022701e-01 -1.04320455e+00 1.98723114e+00
2.80440211e-01 2.36371785e-01 3.85227084e-01 7.54069030e-01
8.33925307e-01 9.20803487e-01 7.41612762e-02 -2.07251102e-01
1.11828434e+00 -1.24704409e+00 -6.82562470e-01 -4.92300749e-01
1.01775050e+00 -8.82851064e-01 1.22390819e+00 1.73753008e-01
-1.18490195e+00 -8.36329997e-01 -9.41826701e-01 -2.43179128e-02
-4.82375562e-01 1.23001702e-01 2.38715872e-01 3.51203054e-01
-6.68969452e-01 6.21929228e-01 -6.95990443e-01 -2.97028333e-01
4.24465477e-01 3.67921352e-01 -2.68455058e-01 -3.50111634e-01
-1.60178375e+00 1.14286613e+00 6.93887353e-01 -2.84900576e-01
-1.00677776e+00 -9.83658254e-01 -4.30639654e-01 -1.41222745e-01
2.29236618e-01 -8.30443799e-01 1.55875921e+00 -1.53379583e+00
-1.53627372e+00 7.11093843e-01 -1.64719462e-01 -2.46604532e-01
6.49962544e-01 -3.14803869e-01 -4.43680495e-01 -2.64663011e-01
-1.42985089e-02 7.53785551e-01 7.86317945e-01 -1.06117666e+00
-5.84300756e-01 -1.72940701e-01 -1.36773527e-01 5.32032967e-01
-5.11435211e-01 2.32455358e-02 -3.73379767e-01 -7.87325263e-01
-4.78457026e-02 -1.17423248e+00 -8.24305117e-02 -7.23105311e-01
1.54486656e-01 -3.23970854e-01 5.74700475e-01 -7.70938933e-01
9.46928084e-01 -1.94890034e+00 5.65708518e-01 -2.69987375e-01
-1.87596217e-01 6.55094206e-01 -7.18316972e-01 3.86798769e-01
-1.63555384e-01 1.43528402e-01 -1.73475116e-01 -9.98655707e-02
-3.85520339e-01 2.94355810e-01 -2.22447336e-01 1.80295840e-01
3.47202897e-01 1.13364244e+00 -1.15076756e+00 -3.50250363e-01
-1.73963472e-01 3.45311940e-01 -5.50165534e-01 1.59266114e-01
-5.36808372e-01 7.82099485e-01 -5.25737345e-01 5.48442602e-01
6.67787015e-01 -2.97256142e-01 4.52612907e-01 1.76293239e-01
1.66619390e-01 4.76797402e-01 -6.10958159e-01 2.23816729e+00
-5.84925652e-01 5.76718330e-01 -3.98760021e-01 -9.77764368e-01
8.88809383e-01 5.13260424e-01 3.45840156e-01 -1.18797731e+00
4.56131296e-03 5.19804657e-01 1.87038004e-01 -3.06283295e-01
6.01646423e-01 -9.02731717e-02 -4.76706447e-03 4.22931403e-01
2.93597639e-01 2.05747068e-01 -6.06677532e-02 -2.18699157e-01
6.69925511e-01 3.70706648e-01 -7.92086963e-03 -1.40892386e-01
2.68949121e-01 5.95594943e-01 8.41303587e-01 5.04998744e-01
-2.09101573e-01 3.28989834e-01 -4.50286865e-02 -2.68743485e-01
-1.39022195e+00 -1.01619673e+00 3.93555164e-02 1.40422094e+00
9.51892659e-02 -1.88718155e-01 -9.38759685e-01 -8.85395825e-01
-1.13696784e-01 8.06442440e-01 -3.29968929e-01 -5.11797190e-01
-9.96837556e-01 -8.45989227e-01 4.53930229e-01 4.47231710e-01
5.96428633e-01 -8.85656059e-01 -7.52577707e-02 3.07643324e-01
-7.27418005e-01 -1.01222014e+00 -4.86211807e-01 2.51516849e-01
-1.45126748e+00 -5.06780088e-01 -1.14846289e+00 -1.12522531e+00
8.66011918e-01 4.05625999e-01 1.11698067e+00 -1.38245299e-01
3.44316125e-01 -2.46061742e-01 -4.52549666e-01 -3.16552788e-01
-9.58803713e-01 7.17807174e-01 3.41072857e-01 -6.50430441e-01
6.09032750e-01 -5.30690789e-01 -2.30188534e-01 5.86020172e-01
-5.70395589e-01 2.41234511e-01 7.89926052e-01 1.09749532e+00
5.80287516e-01 1.11499257e-01 7.73654521e-01 -8.99595022e-01
7.48599589e-01 -6.32264495e-01 -5.02776086e-01 5.62375963e-01
-6.87742591e-01 1.09660216e-01 8.67107928e-01 -1.07342887e+00
-1.20312297e+00 -3.06340903e-01 1.99841797e-01 -5.66789329e-01
1.41670648e-02 5.34030974e-01 -2.39274174e-01 1.24081880e-01
9.00569379e-01 4.08340037e-01 -2.73091167e-01 -7.19678581e-01
3.77356887e-01 6.97865546e-01 3.87066394e-01 -9.52237964e-01
9.18039382e-01 -1.07351325e-01 -4.40579295e-01 -4.30883616e-01
-6.44831240e-01 -2.31449813e-01 -7.88769186e-01 8.12884197e-02
4.29868400e-01 -1.25367403e+00 1.06578045e-01 3.38389426e-01
-1.11753190e+00 -7.30095685e-01 2.68406235e-02 5.12618959e-01
-4.90189791e-01 1.44157499e-01 -3.63174021e-01 -1.53712064e-01
-3.51417422e-01 -1.18836081e+00 6.40952051e-01 2.40634069e-01
-1.69856220e-01 -1.06991005e+00 3.79184127e-01 5.49302816e-01
3.76655191e-01 -2.27318153e-01 1.12004411e+00 -7.61639178e-01
-3.68870884e-01 2.02820390e-01 5.04498072e-02 4.65032250e-01
2.51298428e-01 -2.10243866e-01 -8.35661471e-01 -6.45386815e-01
-2.85088360e-01 -3.22946221e-01 5.68054259e-01 1.84845462e-01
6.98846519e-01 -3.87506098e-01 -4.14171845e-01 7.93966353e-01
1.25429380e+00 2.23061070e-01 3.69547576e-01 7.17374504e-01
4.86106694e-01 4.74112332e-01 9.20794845e-01 -1.84016541e-01
2.18260810e-01 7.26186275e-01 4.10697050e-03 -5.79544231e-02
-3.40118587e-01 -4.56575096e-01 6.94831312e-01 1.19068778e+00
-9.12266225e-02 -1.54453978e-01 -1.20289958e+00 7.66568005e-01
-1.97669816e+00 -7.97271192e-01 2.10322216e-01 1.99085224e+00
1.29706216e+00 7.61878192e-02 1.01457223e-01 -6.41546309e-01
7.57047534e-01 -2.71898448e-01 -7.65288711e-01 -6.44330323e-01
-2.61063248e-01 1.52266040e-01 5.99414408e-01 3.08088750e-01
-9.21432674e-01 1.55297422e+00 6.02544880e+00 8.50232005e-01
-1.07888782e+00 6.13190353e-01 3.03951532e-01 -2.39966601e-01
-3.17630619e-01 -2.33262777e-02 -8.79417241e-01 5.06910861e-01
1.30418348e+00 -4.42541242e-01 5.05708337e-01 9.14962590e-01
2.24167109e-02 5.28995335e-01 -9.68079150e-01 5.11535108e-01
-2.94590324e-01 -1.23313248e+00 2.50585288e-01 8.34698975e-02
1.35295641e+00 4.37780261e-01 3.94094437e-01 8.12375128e-01
4.78874624e-01 -7.26330459e-01 3.71820420e-01 3.19465362e-02
7.60222614e-01 -8.90702307e-01 6.04445815e-01 5.87775826e-01
-5.73520720e-01 -3.61098275e-02 -7.96496391e-01 2.08815888e-01
-3.16849262e-01 1.68853372e-01 -1.21713901e+00 7.22934544e-01
5.17088771e-01 6.11938894e-01 -5.31394958e-01 7.29274333e-01
-2.37945616e-01 6.02163315e-01 8.63386877e-03 -8.23057815e-02
3.16822976e-01 -1.30891800e-01 5.56589365e-01 1.24616492e+00
3.29126030e-01 -1.35205999e-01 4.00393130e-03 7.00204551e-01
-3.79410058e-01 2.42699221e-01 -5.72859049e-01 -2.87442446e-01
5.97917914e-01 6.73429608e-01 -7.97309652e-02 -4.80436623e-01
-4.56245154e-01 1.28402472e+00 5.38145125e-01 6.42896235e-01
-9.02468085e-01 -6.71790689e-02 1.06232107e+00 -2.00785577e-01
1.33373678e-01 -3.24570000e-01 -3.51442754e-01 -1.48197854e+00
5.05698957e-02 -1.42433178e+00 4.31365341e-01 -6.31538212e-01
-1.15251732e+00 5.65108061e-01 1.99539736e-01 -1.55792642e+00
-3.81061971e-01 -6.04961693e-01 -2.73881108e-01 1.11485279e+00
-1.59886086e+00 -1.21452463e+00 2.02020690e-01 5.08495986e-01
8.61606598e-01 -5.90694308e-01 8.23041618e-01 3.01855475e-01
-5.91565847e-01 1.17649126e+00 8.56546342e-01 1.92457452e-01
1.23052049e+00 -9.79742289e-01 9.15357709e-01 8.09832573e-01
-2.61180941e-03 8.73013616e-01 6.52672410e-01 -8.94050002e-01
-1.45295703e+00 -1.25771117e+00 1.00336385e+00 -6.90897763e-01
4.88656193e-01 -2.47922480e-01 -1.12739611e+00 7.01693773e-01
3.47251415e-01 -4.62559283e-01 6.91321909e-01 3.70527595e-01
-6.46496236e-01 7.07652643e-02 -1.02044725e+00 8.44014645e-01
1.21810436e+00 -5.30782044e-01 -9.16108191e-01 5.55663526e-01
8.99933696e-01 -6.23703718e-01 -9.37982082e-01 4.53044266e-01
4.32803512e-01 -2.12098613e-01 9.92721319e-01 -1.22175276e+00
6.37884796e-01 -1.27403764e-02 -1.25362724e-01 -1.92842650e+00
-4.65697527e-01 -4.85250860e-01 -1.91125169e-01 1.11816239e+00
7.77576208e-01 -7.03152657e-01 6.86655104e-01 1.21535003e-01
-2.26404577e-01 -4.89790857e-01 -9.87334609e-01 -1.25034022e+00
1.01506042e+00 -1.29811035e-03 7.75459528e-01 1.49413824e+00
4.59257327e-02 4.65965867e-01 -3.35068285e-01 3.41470242e-01
5.80010891e-01 2.22165465e-01 8.29689443e-01 -9.21113610e-01
-3.83718997e-01 -4.73778754e-01 2.43747711e-01 -9.79010165e-01
4.18909043e-01 -1.28539777e+00 -2.17950866e-01 -1.37490153e+00
2.92809695e-01 -3.58294070e-01 -4.65758085e-01 6.61849320e-01
-4.98169035e-01 -5.82649298e-02 1.06108403e-02 5.14346242e-01
-3.50152731e-01 4.11924660e-01 1.71977401e+00 -4.07327622e-01
-2.87555605e-01 -4.60414737e-02 -6.36521518e-01 2.16670945e-01
1.38158751e+00 -8.46160054e-01 -5.21853864e-01 -1.05166733e+00
2.04665542e-01 -1.45780891e-01 -1.12874456e-01 -6.99295104e-01
4.33162041e-02 -4.90961671e-01 4.52342570e-01 -5.11767745e-01
1.21654615e-01 -7.52113998e-01 3.19846086e-02 4.63552982e-01
-3.79926056e-01 3.28368545e-01 6.46779060e-01 2.47080863e-01
6.44662827e-02 -2.88871676e-01 8.17544580e-01 -3.04752767e-01
-7.41622150e-01 1.96343929e-01 -1.35382399e-01 2.96982586e-01
6.26796722e-01 -9.13392454e-02 -5.58431864e-01 -6.74959496e-02
-5.25299132e-01 4.39446211e-01 6.97696447e-01 8.65062654e-01
3.41879100e-01 -1.45243633e+00 -1.04768741e+00 1.86330289e-01
2.80564874e-01 -9.88068581e-02 2.37974033e-01 5.29648364e-01
-1.18696317e-01 4.78514194e-01 -5.39842129e-01 -6.75001502e-01
-1.28026414e+00 6.19006574e-01 5.17718315e-01 -1.88433543e-01
-3.77786607e-01 7.94973433e-01 6.45414367e-02 -1.04422545e+00
1.57280594e-01 -2.98345774e-01 2.69952327e-01 -1.62650406e-01
3.93518984e-01 2.37579614e-01 1.25986859e-01 -4.39702481e-01
-2.25706697e-01 5.23141325e-01 -6.66236997e-01 -2.17532799e-01
1.19175529e+00 -7.01594129e-02 1.88661829e-01 2.44746104e-01
1.23411274e+00 -3.50965828e-01 -1.08515596e+00 -7.21185386e-01
1.40151829e-01 -3.36814702e-01 -1.87927201e-01 -1.23937130e+00
-6.86626494e-01 9.67781901e-01 6.24558270e-01 -6.07278168e-01
1.08125556e+00 -2.86104321e-01 1.01960576e+00 7.77669430e-01
4.12418783e-01 -1.35992205e+00 1.04814976e-01 9.19005692e-01
6.31545663e-01 -1.26366103e+00 -1.83057323e-01 9.88443196e-02
-6.35995328e-01 9.89621580e-01 8.74296904e-01 1.87828466e-01
2.65622616e-01 -1.03760473e-01 3.78297925e-01 2.97253311e-01
-9.69754815e-01 1.79060340e-01 3.38968277e-01 7.10519135e-01
5.31572163e-01 1.96665347e-01 -1.71346202e-01 3.81067306e-01
-2.26964414e-01 -9.61088389e-02 2.04008698e-01 9.25836444e-01
-4.48152274e-01 -1.54260623e+00 -2.23389581e-01 -1.41824931e-01
-3.37013543e-01 -2.64026344e-01 -8.06168258e-01 7.40493774e-01
1.36989400e-01 7.11112320e-01 -1.29580334e-01 -5.67817092e-01
5.11558771e-01 5.18542409e-01 7.08392918e-01 -7.73890018e-01
-8.77000749e-01 -8.47172644e-03 5.89108802e-02 -2.36874476e-01
-2.07637146e-01 -6.11203015e-01 -1.08301306e+00 -4.13113415e-01
-1.53291747e-01 3.66994560e-01 7.38020182e-01 7.33329415e-01
5.10793626e-01 4.20778483e-01 5.68819880e-01 -2.62615293e-01
-8.61838877e-01 -1.16528809e+00 1.85388818e-01 2.67183602e-01
3.57438833e-01 -4.20662701e-01 -1.75854634e-03 -8.21533892e-03] | [11.640252113342285, 10.278436660766602] |
b0bb8a61-e2de-4140-965b-5ae0dcfb1c04 | invariant-3d-shape-recognition-using | 2005.11558 | null | https://arxiv.org/abs/2005.11558v1 | https://arxiv.org/pdf/2005.11558v1.pdf | Invariant 3D Shape Recognition using Predictive Modular Neural Networks | In this paper PREMONN (PREdictive MOdular Neural Networks) model/architecture is generalized to functions of two variables and to non-Euclidean spaces. It is presented in the context of 3D invariant shape recognition and texture recognition. PREMONN uses local relation, it is modular and exhibits incremental learning. The recognition process can start at any point on a shape or texture, so a reference point is not needed. Its local relation characteristic enables it to recognize shape and texture even in presence of occlusion. The analysis is mainly mathematical. However, we present some experimental results. The methods presented in this paper can be applied to many problems such as gesture recognition, action recognition, dynamic texture recognition etc. | ['Vasileios Petridis'] | 2020-05-23 | null | null | null | null | ['3d-shape-recognition', 'dynamic-texture-recognition'] | ['computer-vision', 'computer-vision'] | [ 4.11261678e-01 -1.63946897e-01 -3.29701960e-01 -1.63199306e-01
1.83276415e-01 -2.75427073e-01 6.95333064e-01 -1.37314528e-01
-3.18013608e-01 4.96669710e-01 -3.12711895e-01 -2.09284350e-01
-5.34701347e-01 -9.12991464e-01 -3.64327133e-01 -9.09219146e-01
-5.92575148e-02 6.96905494e-01 5.84191024e-01 -1.55757263e-01
3.70537072e-01 1.14479399e+00 -1.51100409e+00 4.01323140e-01
1.67132840e-02 9.41940963e-01 1.46111175e-01 9.35797751e-01
-1.43712014e-01 7.30527103e-01 -3.07615012e-01 4.70208302e-02
2.49089062e-01 -2.17531502e-01 -8.89611065e-01 -2.97196698e-03
1.69194981e-01 2.14858979e-01 1.93130691e-02 8.27237189e-01
4.84899670e-01 2.65694618e-01 1.25129521e+00 -1.06332994e+00
-5.53670466e-01 1.14255629e-01 -1.64371744e-01 1.19295731e-01
8.30968395e-02 -3.01740557e-01 4.41270739e-01 -1.02154338e+00
6.94864392e-01 1.39709437e+00 8.52034092e-01 3.22072446e-01
-1.13210416e+00 -1.73654761e-02 -1.91398129e-01 6.28864944e-01
-1.36358035e+00 -1.01006173e-01 8.01385939e-01 -4.00925249e-01
1.05563283e+00 5.21598279e-01 4.05749768e-01 7.44050086e-01
8.14031601e-01 7.09840834e-01 1.20529199e+00 -7.75431275e-01
1.21564448e-01 -2.36864582e-01 9.49847028e-02 6.82818055e-01
-1.09933458e-01 -1.54890548e-02 -2.51651376e-01 9.98083130e-02
1.20289612e+00 2.84790933e-01 2.84319341e-01 -5.02573788e-01
-9.77120996e-01 4.77545142e-01 1.86913311e-01 1.03166902e+00
-2.62060076e-01 -1.26876935e-01 1.87710166e-01 6.93162739e-01
5.47524542e-02 1.90895513e-01 -5.35250902e-01 -2.31211632e-01
-6.59874320e-01 -1.65023386e-01 7.79940188e-01 4.99657005e-01
3.44694853e-01 3.25624377e-01 4.06762213e-01 7.53486753e-01
2.80723125e-01 5.72433412e-01 7.93553650e-01 -5.40227592e-01
-9.06874314e-02 8.81165683e-01 -2.10697889e-01 -1.25420809e+00
-8.30780149e-01 1.21173859e-01 -1.00470853e+00 7.39840627e-01
3.56771767e-01 1.03933476e-01 -8.77201140e-01 1.01704824e+00
1.87959924e-01 -1.39951035e-01 4.71515432e-02 6.32618964e-01
1.00693583e+00 3.92735213e-01 -2.99397647e-01 7.19772503e-02
1.10223341e+00 -6.85043931e-01 -4.58232492e-01 3.27576965e-01
1.27089947e-01 -9.42868352e-01 4.48612928e-01 5.19395590e-01
-8.30189884e-01 -7.83053219e-01 -1.06757426e+00 5.28702736e-02
-8.42204392e-01 4.04581189e-01 3.68582129e-01 6.15770638e-01
-1.02317548e+00 7.80021369e-01 -9.79537845e-01 -6.99471712e-01
-4.34670597e-03 7.86964774e-01 -6.02245033e-01 4.28693742e-01
-8.20484161e-01 1.16974139e+00 6.17471099e-01 2.85369545e-01
-3.95842224e-01 4.32924896e-01 -6.45647168e-01 -1.94378436e-01
-1.05915524e-01 -3.88209075e-01 9.33095515e-01 -1.20438278e+00
-1.77089810e+00 6.42049789e-01 -3.68096381e-01 -1.41248956e-01
4.66787636e-01 1.39950410e-01 -4.28525120e-01 7.46394321e-02
-4.91726100e-01 3.68078321e-01 1.12956965e+00 -7.26676881e-01
-2.09830135e-01 -3.51158082e-01 -9.54974517e-02 1.93079025e-01
1.23584699e-02 8.73919427e-02 -2.01756731e-02 -5.23213565e-01
8.43002319e-01 -6.99946761e-01 7.55619109e-02 -1.05865628e-01
8.05111080e-02 -3.91348302e-01 1.23250985e+00 -4.94064450e-01
7.50717521e-01 -2.01733565e+00 3.54672015e-01 5.85635781e-01
-9.93015021e-02 2.28743732e-01 1.72585342e-02 4.65245247e-01
-2.61921585e-01 -4.32211794e-02 -9.56072509e-02 1.25934118e-02
-1.59964293e-01 4.53684390e-01 1.84598938e-01 5.00132918e-01
1.02071024e-01 9.27963376e-01 -3.83586854e-01 -6.27275705e-01
3.51954311e-01 5.73709607e-01 4.23203409e-02 -2.82613009e-01
8.35379288e-02 2.61205435e-01 -4.61936355e-01 1.04477513e+00
5.01199484e-01 1.49418965e-01 1.54066488e-01 -2.15259165e-01
-1.72827169e-01 -4.29824889e-01 -1.49902344e+00 1.10578012e+00
-4.89948988e-01 9.08051252e-01 -1.82078466e-01 -1.36856854e+00
1.49182260e+00 6.27050459e-01 4.50278640e-01 -4.60651755e-01
4.74643648e-01 1.58581138e-01 8.58989283e-02 -7.65867233e-01
3.54675978e-01 -2.89236039e-01 3.87500107e-01 4.47076380e-01
1.37509733e-01 1.96708158e-01 2.73389891e-02 -6.07851446e-01
6.92389548e-01 3.64160120e-01 4.80500251e-01 -3.08676571e-01
9.45009112e-01 -3.81457955e-01 2.93328673e-01 5.55871606e-01
-7.03318045e-02 4.31719601e-01 3.89629692e-01 -8.69028747e-01
-1.09251690e+00 -1.16614366e+00 -4.06856477e-01 1.04326725e+00
5.81864603e-02 4.87675257e-02 -2.71773040e-01 -3.89585674e-01
-1.75478935e-01 4.28778306e-02 -8.31371129e-01 2.08127797e-01
-6.80361331e-01 -5.46163082e-01 4.80583131e-01 6.42852366e-01
6.25245988e-01 -1.39755023e+00 -7.46577859e-01 2.48533472e-01
2.62815326e-01 -6.10079587e-01 1.32738829e-01 2.98055559e-01
-1.49499202e+00 -1.28020132e+00 -5.11767209e-01 -1.14780545e+00
8.01164508e-01 -1.45054936e-01 7.52324402e-01 -3.29576433e-02
-2.77795494e-01 6.66831374e-01 -3.55327010e-01 -3.91590059e-01
-2.54063249e-01 8.42608139e-02 1.03201330e-01 2.03466088e-01
4.99912143e-01 -7.21837223e-01 -2.04382300e-01 7.74100602e-01
-9.14652407e-01 -2.98911154e-01 5.95876038e-01 1.01050925e+00
5.55198789e-01 2.46351600e-01 2.58670092e-01 -4.37380999e-01
4.56218660e-01 8.30121115e-02 -4.25239742e-01 4.17829692e-01
-3.91495705e-01 1.57403067e-01 4.59859878e-01 -6.32096052e-01
-8.41382444e-01 2.50278413e-01 1.21427206e-02 -1.19163774e-01
-5.25364757e-01 3.44861180e-01 -1.17957734e-01 -4.77192640e-01
7.12390065e-01 4.74934787e-01 3.22566479e-01 -8.13204229e-01
-4.00636531e-02 5.38827956e-01 3.15362841e-01 -4.95990485e-01
5.90700805e-01 3.30915958e-01 7.20534146e-01 -1.25582778e+00
2.85350353e-01 -2.02318668e-01 -1.40300405e+00 -4.22157675e-01
6.90317988e-01 -1.68876097e-01 -8.24300647e-01 5.36170065e-01
-1.01292896e+00 -4.05154489e-02 -9.21030715e-02 3.96340430e-01
-6.78439081e-01 3.81583124e-01 -4.58754390e-01 -1.01543462e+00
-3.51362497e-01 -1.04424870e+00 6.54306173e-01 1.89698786e-01
-1.98824689e-01 -1.49737155e+00 1.14528559e-01 -1.97834939e-01
4.46691483e-01 2.36355662e-01 9.16320026e-01 -6.56323910e-01
-3.73843491e-01 -3.85253519e-01 1.56608716e-01 3.12037826e-01
2.05405712e-01 2.34168008e-01 -7.04221308e-01 -4.43655811e-02
1.13842867e-01 -3.58408852e-03 7.87029743e-01 2.18576804e-01
7.80086994e-01 -1.64553493e-01 -2.20259547e-01 3.66485924e-01
1.49035203e+00 7.37279892e-01 7.39870906e-01 6.27226412e-01
3.60377997e-01 4.93518114e-01 2.23846287e-01 8.22136253e-02
-9.82259214e-02 6.68554246e-01 1.94925085e-01 7.22479895e-02
3.78645770e-02 2.22545967e-01 3.01366329e-01 7.80512929e-01
-7.28094161e-01 1.20026685e-01 -1.01222587e+00 2.62125373e-01
-1.86466157e+00 -1.07748973e+00 -5.92568107e-02 2.10854602e+00
4.42128897e-01 2.99035423e-02 1.53521255e-01 3.99871558e-01
8.28935921e-01 -7.33089596e-02 -1.74488470e-01 -1.02885234e+00
-4.78275806e-01 4.79182035e-01 1.81899369e-01 6.55155420e-01
-1.34367442e+00 8.83137286e-01 7.19524240e+00 8.17704916e-01
-1.67888904e+00 1.48627102e-01 3.00410062e-01 2.11983964e-01
2.74165839e-01 -3.73310864e-01 -3.95655721e-01 2.02597640e-02
6.02012634e-01 2.26098880e-01 2.16863111e-01 7.02235460e-01
-1.90345570e-01 -1.98052272e-01 -9.49595332e-01 1.17051005e+00
1.52883962e-01 -9.17719483e-01 5.55360615e-02 -3.50023448e-01
3.55599999e-01 -2.06145525e-01 1.48343205e-01 1.33767247e-01
-3.07131529e-01 -1.34293735e+00 2.97101408e-01 1.02182734e+00
4.97460574e-01 -7.38896430e-01 9.93952870e-01 4.01922584e-01
-1.36640382e+00 -2.13797744e-02 -4.91265625e-01 -3.44153702e-01
-2.62984455e-01 -2.60103475e-02 -9.08668458e-01 7.39347517e-01
3.80548090e-01 5.64889073e-01 -6.71039760e-01 1.15184772e+00
-4.82983515e-02 2.30457738e-01 -5.53986371e-01 -5.81171334e-01
8.50766748e-02 -2.25215226e-01 5.79092681e-01 1.22655416e+00
3.60385001e-01 7.08173066e-02 -5.36776334e-02 4.26383913e-01
6.28110230e-01 6.18189573e-01 -8.45794559e-01 2.66931891e-01
-7.55902603e-02 9.62614179e-01 -1.07312477e+00 -1.66727781e-01
-1.98269427e-01 9.31924880e-01 -1.43754452e-01 3.21321458e-01
-3.08024138e-01 -4.57302302e-01 8.44792724e-02 -1.81487232e-01
4.17768598e-01 -4.47313279e-01 -5.22153676e-01 -8.78984809e-01
-1.18716192e-02 -5.43120921e-01 1.89231291e-01 -7.09728062e-01
-8.60238791e-01 6.81005597e-01 1.53706238e-01 -1.35315168e+00
-3.23223203e-01 -1.29444468e+00 -7.11166561e-01 7.95193195e-01
-6.52364194e-01 -1.16503048e+00 1.19061582e-01 7.45854795e-01
3.21180582e-01 -7.99045384e-01 1.05386865e+00 6.45239875e-02
-3.86483967e-01 4.71065015e-01 3.22720557e-01 2.55913436e-01
3.85419577e-01 -1.17197299e+00 -1.41254783e-01 5.45319140e-01
2.10609376e-01 5.61099052e-01 4.55598950e-01 -5.53786337e-01
-1.26438558e+00 -4.94292259e-01 1.14698064e+00 -4.35577989e-01
4.74668205e-01 8.26828331e-02 -6.39501095e-01 5.07878602e-01
4.32432517e-02 -2.36842722e-01 5.35166681e-01 1.17536187e-01
-2.06985667e-01 -2.87491649e-01 -1.24817884e+00 5.56544006e-01
6.58968270e-01 -5.28381288e-01 -8.72879446e-01 1.37936667e-01
-1.76451996e-01 -3.33374500e-01 -9.62430179e-01 4.98149872e-01
1.09664845e+00 -9.45531845e-01 9.39048767e-01 -6.07273579e-01
-1.03713699e-01 -3.78578991e-01 -2.08619773e-01 -6.66112423e-01
-5.35746455e-01 -4.49084073e-01 2.66892351e-02 6.46702886e-01
4.13200617e-01 -7.18341351e-01 7.53259420e-01 1.73918977e-01
3.45906675e-01 -7.97946572e-01 -1.40743935e+00 -1.01530194e+00
-1.58031434e-02 -5.65885365e-01 3.35231513e-01 7.28333294e-01
2.35954430e-02 2.99882412e-01 -3.70127857e-01 -2.11983219e-01
3.62082273e-01 2.67422169e-01 5.82128525e-01 -1.54695022e+00
-2.23794058e-01 -7.77305722e-01 -1.27649629e+00 -7.89302468e-01
-1.53801829e-01 -8.39519560e-01 -3.73471588e-01 -1.24159884e+00
-1.29050732e-01 -3.05317432e-01 -4.95166183e-01 7.09446311e-01
4.87799376e-01 5.12593448e-01 2.16375023e-01 1.67913079e-01
-4.51398551e-01 2.64047414e-01 1.22403276e+00 -2.60097563e-01
-2.08349764e-01 2.50779152e-01 8.93813297e-02 8.44036400e-01
1.31653070e+00 -1.80798873e-01 -2.02770531e-01 -1.36810392e-01
1.15399934e-01 -1.47919372e-01 3.36975545e-01 -1.14134860e+00
3.30380410e-01 -1.74262911e-01 7.30015695e-01 -6.54053092e-01
3.90381575e-01 -1.26956534e+00 3.36270869e-01 5.98376870e-01
1.68650612e-01 4.69581038e-01 1.27027720e-01 1.27187729e-01
-2.81413853e-01 -6.39392912e-01 6.99117720e-01 -2.31797472e-02
-8.36070836e-01 2.23521143e-02 -5.67495108e-01 -7.24368691e-01
9.98449683e-01 -9.06514883e-01 2.93886550e-02 -2.80816853e-01
-1.47154129e+00 -4.99391019e-01 2.78474927e-01 3.10572952e-01
8.45038295e-01 -1.56684148e+00 -3.80466074e-01 5.05466640e-01
-1.27951443e-01 -4.20621008e-01 2.09629368e-02 1.07786143e+00
-9.21543598e-01 7.03073502e-01 -6.75831914e-01 -8.15856099e-01
-1.64864933e+00 4.03123111e-01 5.45243859e-01 -4.69927443e-03
-5.35383701e-01 4.45554376e-01 -2.92928159e-01 -3.86059195e-01
2.80188292e-01 -2.72684723e-01 -6.27799213e-01 -6.75662514e-03
4.79406536e-01 6.02415681e-01 3.46547775e-02 -9.30045545e-01
-2.72097588e-01 1.19297481e+00 3.41317952e-01 -2.15617001e-01
1.36507487e+00 2.65957534e-01 -5.89143932e-01 9.14991081e-01
1.05250883e+00 -2.20406696e-01 -7.43999124e-01 -3.24048966e-01
1.76364481e-01 -1.52124599e-01 -2.62612432e-01 -7.24153280e-01
-7.56355166e-01 9.23192322e-01 1.06786287e+00 2.86176652e-01
1.18458509e+00 -2.33392879e-01 1.90270334e-01 9.29518104e-01
4.16742027e-01 -1.14647818e+00 -9.51633751e-02 1.04114962e+00
1.17272437e+00 -9.06802833e-01 -3.47649567e-02 6.28831238e-03
-2.39561722e-01 1.98262978e+00 3.39284092e-01 -2.43488312e-01
1.01037848e+00 3.92362952e-01 3.08758132e-02 4.55020629e-02
-6.71584725e-01 -9.83611271e-02 6.61103427e-01 8.73837233e-01
6.03311002e-01 2.41439372e-01 -5.76547801e-01 2.52428919e-01
-6.47897124e-02 -6.83763996e-02 1.67967707e-01 1.11657631e+00
-6.33317232e-01 -1.44545674e+00 -8.88430536e-01 1.44264132e-01
-2.95945138e-01 1.30225375e-01 -5.82815111e-01 9.29693222e-01
4.47782695e-01 6.16480231e-01 2.45464053e-02 -5.83136976e-01
2.80899763e-01 6.34131849e-01 8.80082548e-01 -1.82982296e-01
-4.50304180e-01 2.62896538e-01 -1.13194823e-01 -2.48030514e-01
-6.75289035e-01 -1.03654635e+00 -1.08949113e+00 -2.00008467e-01
-5.60879447e-02 -1.07589357e-01 6.88327134e-01 8.71396184e-01
-4.40808199e-02 3.48414093e-01 3.17722380e-01 -8.83114219e-01
-1.08926468e-01 -1.25616252e+00 -6.49026930e-01 -2.02675387e-02
1.68085560e-01 -9.38944697e-01 4.11260650e-02 1.59308389e-01] | [10.147196769714355, -0.515471339225769] |
5e98aa00-0e0d-4546-a2bf-b57d261871da | solving-seismic-wave-equations-on-variable | 2209.12340 | null | https://arxiv.org/abs/2209.12340v3 | https://arxiv.org/pdf/2209.12340v3.pdf | Solving Seismic Wave Equations on Variable Velocity Models with Fourier Neural Operator | In the study of subsurface seismic imaging, solving the acoustic wave equation is a pivotal component in existing models. The advancement of deep learning enables solving partial differential equations, including wave equations, by applying neural networks to identify the mapping between the inputs and the solution. This approach can be faster than traditional numerical methods when numerous instances are to be solved. Previous works that concentrate on solving the wave equation by neural networks consider either a single velocity model or multiple simple velocity models, which is restricted in practice. Instead, inspired by the idea of operator learning, this work leverages the Fourier neural operator (FNO) to effectively learn the frequency domain seismic wavefields under the context of variable velocity models. We also propose a new framework paralleled Fourier neural operator (PFNO) for efficiently training the FNO-based solver given multiple source locations and frequencies. Numerical experiments demonstrate the high accuracy of both FNO and PFNO with complicated velocity models in the OpenFWI datasets. Furthermore, the cross-dataset generalization test verifies that PFNO adapts to out-of-distribution velocity models. Moreover, PFNO has robust performance in the presence of random noise in the labels. Finally, PFNO admits higher computational efficiency on large-scale testing datasets than the traditional finite-difference method. The aforementioned advantages endow the FNO-based solver with the potential to build powerful models for research on seismic waves. | ['Youzuo Lin', 'Xiu Yang', 'Hanchen Wang', 'Bian Li'] | 2022-09-25 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [ 4.87725884e-02 -3.37755263e-01 3.43013644e-01 -3.50393355e-02
-8.61113727e-01 -1.81990325e-01 -1.01396674e-02 -3.57335120e-01
-3.12549442e-01 5.05036414e-01 -1.15425386e-01 -3.57890874e-01
-6.61501944e-01 -1.02142453e+00 -8.52109492e-01 -1.02423525e+00
-6.09071076e-01 2.44750947e-01 1.29715592e-01 -2.93890208e-01
1.23063996e-01 5.13212562e-01 -1.35915470e+00 5.98145574e-02
8.06973577e-01 1.34261107e+00 -2.05034241e-02 5.30695379e-01
-5.11271134e-03 7.39858747e-01 -4.24089700e-01 3.13852340e-01
3.80635798e-01 -1.55204415e-01 -8.68037522e-01 -2.95753509e-01
8.90457481e-02 -4.62478906e-01 -3.12671989e-01 8.71687889e-01
8.02390337e-01 4.92366225e-01 5.51959097e-01 -1.07016945e+00
-4.62281793e-01 5.85237324e-01 -5.87562919e-01 3.25058818e-01
2.83208400e-01 5.08535057e-02 9.30542886e-01 -1.01787698e+00
2.26962954e-01 7.87805498e-01 1.42880607e+00 1.90203905e-01
-1.07814753e+00 -6.86435938e-01 -2.57301241e-01 2.14500889e-01
-1.38512874e+00 4.00204025e-03 1.04809093e+00 -7.45990396e-01
8.37792695e-01 1.77253813e-01 5.68258941e-01 8.37721646e-01
-1.41520761e-02 5.37959754e-01 6.98518515e-01 -2.44226873e-01
3.03188384e-01 -5.30476153e-01 2.33863309e-01 5.79476297e-01
2.42900718e-02 2.65847534e-01 -6.76817477e-01 -2.97896892e-01
9.95462418e-01 -2.21451208e-01 -7.41889775e-01 7.54290894e-02
-9.99863505e-01 9.79836166e-01 3.03390563e-01 2.67091602e-01
-4.31495011e-01 2.98822552e-01 4.59140360e-01 3.13489288e-01
6.06623650e-01 7.35228062e-01 -5.16762555e-01 -8.98295417e-02
-1.12740564e+00 5.00696421e-01 8.94702852e-01 3.53542119e-01
9.89871860e-01 6.43244267e-01 2.47662425e-01 7.55405247e-01
1.93250746e-01 6.76079094e-01 5.10720193e-01 -1.24107087e+00
1.86873764e-01 1.73800066e-01 -3.14158089e-02 -1.09644949e+00
-7.37722993e-01 -6.69828534e-01 -1.07134891e+00 2.94425458e-01
5.38292766e-01 -5.02594650e-01 -7.22281992e-01 1.48794878e+00
3.50248337e-01 8.94630432e-01 3.33461761e-01 9.50462222e-01
8.62171173e-01 8.63820136e-01 -2.59368390e-01 -1.07232340e-01
7.75204539e-01 -6.21557415e-01 -5.84852040e-01 -1.17906881e-02
8.76647711e-01 -3.49398345e-01 8.82428050e-01 4.21367556e-01
-8.71489942e-01 -3.90574783e-01 -8.80604863e-01 2.97304273e-01
-4.01728638e-02 -2.68216908e-01 8.55743647e-01 3.72413814e-01
-1.09287691e+00 8.17072034e-01 -1.00515306e+00 3.59762192e-01
3.49113286e-01 2.71230400e-01 -1.73049793e-01 -1.60709932e-03
-1.57551920e+00 4.13787901e-01 1.09885722e-01 5.56303799e-01
-8.42425704e-01 -1.45067978e+00 -8.23085546e-01 2.91763455e-01
1.55392259e-01 -4.56179231e-01 1.20099187e+00 -7.20597029e-01
-1.38009393e+00 9.16702673e-02 4.83505465e-02 -3.00673455e-01
5.77921271e-01 -7.80140162e-02 -1.68804020e-01 4.41208422e-01
2.15167135e-01 -4.48170528e-02 6.56963944e-01 -1.20765352e+00
-3.95581514e-01 1.34058848e-01 1.77196879e-02 -1.89529568e-01
-4.36206073e-01 -3.89664114e-01 1.28371879e-01 -6.59009278e-01
5.78019619e-01 -4.49743748e-01 -3.22330147e-01 -7.68525600e-02
-1.21497013e-01 -1.81528121e-01 6.74601436e-01 -7.61627316e-01
1.13191760e+00 -2.08823705e+00 -2.47424487e-02 6.39459014e-01
1.82394519e-01 1.04870513e-01 1.25769712e-02 4.17578816e-01
-3.02645892e-01 1.18196495e-02 -7.91717649e-01 1.55436099e-02
-1.75851703e-01 2.47847527e-01 -4.05494720e-01 4.38832700e-01
2.07618296e-01 5.73335946e-01 -8.44230056e-01 -1.50292456e-01
-2.79104322e-01 5.10132313e-01 -8.50254178e-01 4.65860777e-02
1.01016648e-01 6.71779931e-01 -5.90396047e-01 5.43745279e-01
1.00036466e+00 -2.99658507e-01 -2.27002263e-01 -2.36715719e-01
-3.13336849e-01 -2.52546400e-01 -1.53884697e+00 1.40774751e+00
-7.12649465e-01 4.34010655e-01 3.24025095e-01 -1.62561321e+00
9.00144160e-01 6.78123891e-01 8.67114842e-01 -5.52390575e-01
-5.81857264e-02 7.02645659e-01 -1.42227840e-02 -1.11848390e+00
1.90959144e-02 -3.82042885e-01 2.50633836e-01 4.68299717e-01
1.35123298e-01 -5.39237596e-02 1.28706411e-01 -1.66331440e-01
1.13284624e+00 4.31434028e-02 -2.57585734e-01 -4.15858239e-01
4.91850823e-01 -5.91571964e-02 6.74509704e-01 9.69573617e-01
7.52972588e-02 7.68150985e-01 3.81120145e-01 -6.38849974e-01
-5.74371219e-01 -7.11392224e-01 -5.83211958e-01 8.70934427e-01
-1.85057875e-02 2.58597493e-01 -6.07992649e-01 -9.43590775e-02
-2.98842769e-02 3.12793463e-01 -5.83050668e-01 1.09891491e-02
-9.51820970e-01 -1.00982654e+00 9.45115805e-01 7.90372133e-01
6.70039535e-01 -8.98589253e-01 -7.50668406e-01 2.69617558e-01
-3.62412095e-01 -1.03729665e+00 1.20400012e-01 2.70057946e-01
-7.77740121e-01 -9.20685351e-01 -8.00436020e-01 -9.05907750e-01
2.50871509e-01 -1.06808178e-01 6.88946366e-01 2.95036137e-01
-1.17213726e-01 4.78878438e-01 -3.37411702e-01 -1.76196203e-01
-1.87057078e-01 1.52138053e-02 -9.95516106e-02 3.39391530e-01
-1.75522596e-01 -9.67238367e-01 -4.67881620e-01 6.44125417e-02
-1.11889637e+00 -3.09637457e-01 1.40990943e-01 1.10433757e+00
2.82251626e-01 3.84152383e-01 6.84217036e-01 -7.22006619e-01
4.70372140e-01 -8.33743632e-01 -4.29036796e-01 -1.38710474e-03
-1.73891574e-01 1.44553778e-03 5.67162871e-01 -5.28773367e-01
-1.15610301e+00 -1.21181719e-01 -6.45343840e-01 -3.09930384e-01
4.33861390e-02 1.28349364e+00 2.43372276e-01 -5.22307456e-01
7.42720425e-01 2.14528233e-01 -4.16752249e-02 -5.02897739e-01
-3.69278044e-01 2.33415663e-01 5.59136093e-01 -8.92442942e-01
8.39670837e-01 6.75346255e-01 3.38944823e-01 -1.11365879e+00
-7.90924847e-01 -5.40449262e-01 -3.55769396e-01 -3.51575315e-01
7.74089992e-01 -6.70119405e-01 -9.70386803e-01 1.02087855e+00
-1.23208547e+00 -5.61774850e-01 -1.78155929e-01 7.27956831e-01
-2.24927649e-01 4.76564795e-01 -7.63351798e-01 -9.29998338e-01
-5.14408052e-02 -1.18396008e+00 8.57131779e-01 1.21308960e-01
1.04336590e-01 -1.37822175e+00 -3.12293824e-02 -7.22706988e-02
5.45552552e-01 5.77335238e-01 8.04467380e-01 -2.99749792e-01
-3.60311419e-01 -2.61210233e-01 -4.05041827e-03 2.35631555e-01
-2.49971315e-01 -7.91945681e-02 -1.16478300e+00 -4.84128557e-02
6.62133336e-01 -3.61494273e-01 9.32838023e-01 7.90565073e-01
1.20384848e+00 -7.08148479e-02 1.79236252e-02 1.31225300e+00
1.53889060e+00 1.77857324e-01 4.44035381e-01 4.25660551e-01
7.38566279e-01 5.05864441e-01 7.00532794e-02 6.88825727e-01
1.75415859e-01 2.01892540e-01 3.88658792e-01 -3.56902152e-01
2.50702262e-01 2.20003873e-01 -2.50851493e-02 6.97986364e-01
-4.45137471e-01 -6.54750243e-02 -1.38990390e+00 4.69757915e-01
-1.55161321e+00 -9.78195250e-01 -3.66310507e-01 1.67880809e+00
6.88846409e-01 -7.32164830e-02 -3.56610686e-01 7.17814445e-01
3.48836392e-01 9.60355625e-02 -3.57291549e-01 -1.05845239e-02
-2.20996246e-01 6.10279799e-01 4.61824805e-01 7.37078190e-01
-1.18863046e+00 2.75201231e-01 6.17385197e+00 6.45668387e-01
-1.60924947e+00 2.32332513e-01 4.53977972e-01 3.08674067e-01
-3.97413671e-01 -3.73838425e-01 -4.23237145e-01 2.15460777e-01
8.43877137e-01 1.69499740e-01 3.77631485e-01 5.78504145e-01
3.01268816e-01 -1.07816132e-02 -7.94445395e-01 7.73397386e-01
-1.64395854e-01 -1.64833117e+00 -2.18293741e-01 -2.78779447e-01
7.31913567e-01 3.88688184e-02 -4.96343598e-02 1.42279044e-01
-1.71008080e-01 -1.09922826e+00 8.40290189e-01 6.46925747e-01
7.22215354e-01 -6.44864619e-01 8.75544310e-01 4.39573109e-01
-1.33098209e+00 -2.00146496e-01 -2.02375904e-01 -4.23660994e-01
5.09426773e-01 5.96642077e-01 -3.50713193e-01 8.39178026e-01
1.06869471e+00 6.51994109e-01 4.59708124e-02 1.04128933e+00
-3.63486856e-02 1.14702928e+00 -5.42058170e-01 2.96766996e-01
5.16985714e-01 -1.79572999e-01 4.43751454e-01 1.00724828e+00
6.48337722e-01 3.62336665e-01 3.46771061e-01 7.75252342e-01
3.69256049e-01 -8.85656625e-02 -2.94071078e-01 3.74101490e-01
1.42606765e-01 9.24239516e-01 -7.31966853e-01 7.27067664e-02
-4.28314149e-01 1.65220201e-01 6.18668646e-02 9.13585603e-01
-7.01790631e-01 -3.87930900e-01 5.25384963e-01 2.22111329e-01
3.70900869e-01 -4.19461280e-01 -4.99245316e-01 -7.69159079e-01
6.58302978e-02 -5.88637471e-01 4.24697816e-01 -5.17493248e-01
-1.23320317e+00 6.31599367e-01 1.77425399e-01 -1.20063794e+00
-4.04319465e-02 -8.89970362e-01 -9.10936475e-01 9.70098436e-01
-1.78225291e+00 -8.17655742e-01 -4.13820952e-01 4.97963578e-01
1.52413011e-01 5.35081625e-02 6.59819782e-01 6.11252666e-01
-4.60763305e-01 4.22338575e-01 1.64375037e-01 5.35510659e-01
1.51374236e-01 -8.85928154e-01 2.58852132e-02 8.73296559e-01
-2.88487107e-01 4.36305761e-01 7.32701540e-01 -5.08725762e-01
-1.28805661e+00 -7.74560094e-01 5.30327797e-01 2.13365510e-01
9.54497516e-01 -7.83161074e-02 -1.66846120e+00 3.73047382e-01
-1.28801629e-01 5.81863105e-01 5.91900289e-01 -1.56136841e-01
-1.37810424e-01 -9.90341306e-02 -9.38562989e-01 -1.85786560e-02
8.41108680e-01 -3.74204248e-01 -1.84535384e-01 3.88671279e-01
6.58596337e-01 -6.91742599e-01 -9.92249727e-01 7.11045921e-01
3.65727127e-01 -9.42433536e-01 1.12769163e+00 -4.23459083e-01
8.20741415e-01 -1.46213710e-01 -1.34579927e-01 -1.15227675e+00
-1.53697446e-01 -6.35881066e-01 -9.52581912e-02 8.46023381e-01
3.99211735e-01 -1.04797971e+00 6.90049469e-01 4.03714389e-01
-6.72989607e-01 -1.00263453e+00 -1.29618180e+00 -7.62474954e-01
3.91394407e-01 -1.02031946e+00 5.72329462e-01 1.10504615e+00
-2.87361443e-01 -3.16288978e-01 -2.54813313e-01 8.18859041e-01
6.16645098e-01 -5.38593344e-02 1.85144514e-01 -1.35430980e+00
-6.37589753e-01 -3.83838952e-01 -2.32685000e-01 -1.02762258e+00
3.64953279e-01 -9.93615568e-01 3.69663149e-01 -1.36320591e+00
-4.96545672e-01 -7.33848751e-01 -2.38554865e-01 4.31176931e-01
-1.19058095e-01 4.19808835e-01 -3.76088798e-01 1.33484498e-01
2.24487588e-01 4.91198361e-01 1.24947655e+00 -5.34493104e-02
-7.12258145e-02 1.13599047e-01 -1.33976892e-01 9.57147539e-01
5.30538142e-01 -4.71388996e-01 -3.35766762e-01 -8.57834220e-01
4.30390626e-01 3.82032484e-01 6.03754818e-01 -1.27601111e+00
4.55024093e-01 -1.55194432e-01 1.24577358e-01 -3.08359295e-01
7.95461796e-03 -5.99993587e-01 3.82386334e-02 5.87952852e-01
-1.75409347e-01 5.87051548e-02 2.52886057e-01 3.60190004e-01
-5.08639455e-01 -6.27520680e-01 6.36399150e-01 -1.00413084e-01
-7.82038867e-01 3.73875499e-01 -5.77805281e-01 4.54356730e-01
6.08503878e-01 -3.43711734e-01 9.88477618e-02 -3.58034760e-01
-7.52114832e-01 1.21750258e-01 -3.83164465e-01 -3.35361004e-01
6.94351077e-01 -1.05971360e+00 -7.64191568e-01 6.08184755e-01
-3.89385432e-01 4.34423149e-01 7.00301170e-01 1.12565100e+00
-1.03714979e+00 -2.38959447e-01 2.09453329e-01 -8.26057792e-01
-4.44816738e-01 1.47070866e-02 8.55099559e-01 -1.61635369e-01
-7.50183344e-01 1.34455681e+00 2.39513904e-01 -5.80392838e-01
2.06166551e-01 -4.60727155e-01 -1.39726698e-01 8.12249854e-02
4.96242404e-01 7.05426633e-01 1.62853450e-01 -4.19150233e-01
-1.61255747e-01 7.34602928e-01 6.31376743e-01 -1.27879858e-01
1.77261531e+00 3.53681654e-01 -1.53291821e-01 4.30405825e-01
1.33804905e+00 -1.91816539e-01 -1.39291084e+00 -2.73623019e-01
-1.69827286e-02 -1.27517089e-01 2.51768827e-01 -3.55460167e-01
-1.37981832e+00 1.02313960e+00 3.48646909e-01 2.62130469e-01
1.19731247e+00 -2.86364198e-01 9.58133698e-01 4.86895114e-01
5.16859554e-02 -9.21092927e-01 1.65969878e-01 8.45781088e-01
8.55951130e-01 -1.05588913e+00 -2.33621135e-01 -4.85337883e-01
-2.61920333e-01 1.44364202e+00 3.79788429e-01 -1.77364856e-01
1.17340600e+00 8.52056861e-01 4.61873591e-01 -2.58541405e-01
-1.79900378e-01 2.23557100e-01 2.34776929e-01 4.39465493e-01
3.96719068e-01 -3.61087054e-01 -4.99551035e-02 7.07224011e-01
-2.12584376e-01 -1.33504886e-02 5.63426912e-01 9.90816057e-01
-2.63662905e-01 -6.47410095e-01 -5.46319127e-01 2.99153835e-01
-3.82442802e-01 -1.16102785e-01 5.83150446e-01 7.53777027e-01
1.84257776e-01 6.48204744e-01 -2.96359994e-02 -1.19669311e-01
2.62609899e-01 1.83700919e-01 6.50854548e-03 -3.44300687e-01
-6.23156250e-01 4.47062366e-02 -2.91492760e-01 -3.96300614e-01
-5.80749094e-01 -7.63191581e-01 -1.68391705e+00 4.40818034e-02
-2.64583439e-01 3.52976769e-01 2.62590587e-01 1.21592510e+00
-1.78896695e-01 7.96241164e-01 4.85141218e-01 -1.23502851e+00
-7.91215897e-01 -7.16443360e-01 -7.63814390e-01 2.90192485e-01
6.79732263e-01 -8.77884865e-01 -7.57329047e-01 -9.23422948e-02] | [6.875374794006348, 2.6072983741760254] |
febcb26e-8721-4575-ab39-edc11692e73b | learning-multiple-stock-trading-patterns-with | 2106.12950 | null | https://arxiv.org/abs/2106.12950v2 | https://arxiv.org/pdf/2106.12950v2.pdf | Learning Multiple Stock Trading Patterns with Temporal Routing Adaptor and Optimal Transport | Successful quantitative investment usually relies on precise predictions of the future movement of the stock price. Recently, machine learning based solutions have shown their capacity to give more accurate stock prediction and become indispensable components in modern quantitative investment systems. However, the i.i.d. assumption behind existing methods is inconsistent with the existence of diverse trading patterns in the stock market, which inevitably limits their ability to achieve better stock prediction performance. In this paper, we propose a novel architecture, Temporal Routing Adaptor (TRA), to empower existing stock prediction models with the ability to model multiple stock trading patterns. Essentially, TRA is a lightweight module that consists of a set of independent predictors for learning multiple patterns as well as a router to dispatch samples to different predictors. Nevertheless, the lack of explicit pattern identifiers makes it quite challenging to train an effective TRA-based model. To tackle this challenge, we further design a learning algorithm based on Optimal Transport (OT) to obtain the optimal sample to predictor assignment and effectively optimize the router with such assignment through an auxiliary loss term. Experiments on the real-world stock ranking task show that compared to the state-of-the-art baselines, e.g., Attention LSTM and Transformer, the proposed method can improve information coefficient (IC) from 0.053 to 0.059 and 0.051 to 0.056 respectively. Our dataset and code used in this work are publicly available: https://github.com/microsoft/qlib/tree/main/examples/benchmarks/TRA. | ['Jiang Bian', 'Weiqing Liu', 'Dong Zhou', 'Hengxu Lin'] | 2021-06-24 | null | null | null | null | ['stock-prediction'] | ['time-series'] | [-4.48137373e-01 -4.80898887e-01 -5.60767651e-01 -5.16757786e-01
-6.77920997e-01 -5.82396567e-01 5.35624921e-01 2.29111444e-02
-3.86924654e-01 8.61547291e-01 -3.11156898e-03 -5.14829695e-01
-2.09770367e-01 -9.65410829e-01 -8.60839725e-01 -5.79028368e-01
-2.94889361e-02 6.02296114e-01 5.07085681e-01 -3.63318771e-01
6.53240442e-01 2.92428553e-01 -1.31935871e+00 1.64542362e-01
8.57036054e-01 1.64104152e+00 1.76908940e-01 3.62794921e-02
-5.60963035e-01 1.26533639e+00 -5.06691039e-01 -7.22798049e-01
6.08392179e-01 -1.97768852e-01 -4.24103707e-01 -3.18362087e-01
2.01047063e-01 -6.95506513e-01 -2.98956513e-01 1.03650010e+00
3.60321641e-01 -7.54537731e-02 5.67311466e-01 -1.33440864e+00
-5.02758920e-01 9.44620430e-01 -9.05123651e-01 6.03716254e-01
-4.88055497e-01 4.67334278e-02 1.48991871e+00 -8.43061090e-01
1.58093959e-01 8.76504898e-01 4.51575726e-01 2.81854153e-01
-9.70336735e-01 -1.10962045e+00 3.56870681e-01 3.28564376e-01
-9.65174198e-01 -2.53373891e-01 8.80002737e-01 -3.55322152e-01
7.33412802e-01 9.80346575e-02 6.98206842e-01 7.37294495e-01
2.92127311e-01 1.09312880e+00 1.09359479e+00 9.15820524e-02
7.39771128e-03 5.14400899e-01 1.73908293e-01 5.39631307e-01
1.53541833e-01 7.16093415e-03 -6.99996710e-01 1.33334368e-01
7.55634665e-01 2.09588483e-01 -4.66540530e-02 -1.24228604e-01
-1.01350677e+00 9.33476150e-01 6.83993578e-01 2.82373309e-01
-4.90909845e-01 3.09855998e-01 4.41349506e-01 6.02625132e-01
7.28179753e-01 4.15535331e-01 -7.69962788e-01 -2.10940495e-01
-1.05004907e+00 2.74871528e-01 6.70762599e-01 8.50300968e-01
7.11433172e-01 3.24014574e-01 -1.49410451e-03 6.50684655e-01
2.28318229e-01 3.56234640e-01 5.59447408e-01 -6.84214652e-01
9.08921480e-01 6.36575639e-01 2.55231023e-01 -9.52835619e-01
-1.79026067e-01 -1.01926243e+00 -6.42591953e-01 3.69434282e-02
3.68218482e-01 1.77749991e-02 -5.36843777e-01 1.41769528e+00
1.15588523e-01 4.69792336e-01 -8.58795121e-02 7.92926371e-01
1.53923199e-01 8.99961889e-01 -2.18621790e-01 2.44771373e-02
9.57996488e-01 -1.17720306e+00 -4.58967179e-01 -9.46445093e-02
6.16933703e-01 -7.10715711e-01 1.02108419e+00 5.05296350e-01
-1.09367836e+00 -3.26555073e-01 -1.05701995e+00 2.52896607e-01
-5.13618946e-01 -4.24446613e-02 7.04161704e-01 3.71719152e-01
-8.06265652e-01 9.13697243e-01 -8.11454654e-01 4.09848452e-01
5.47863841e-01 4.64015961e-01 1.57886803e-01 1.82486609e-01
-1.32130027e+00 9.62466300e-01 3.49717319e-01 3.29955578e-01
-7.38204420e-01 -1.09857857e+00 -3.03324133e-01 1.28957078e-01
4.62648541e-01 -3.06977242e-01 1.42055357e+00 -1.16977251e+00
-1.54215896e+00 2.50397801e-01 2.82032013e-01 -7.89741695e-01
7.30843723e-01 -4.49719399e-01 -2.56816000e-01 -1.74707279e-01
3.62844840e-02 3.84862453e-01 5.78041315e-01 -8.36209834e-01
-1.12240362e+00 -1.84784815e-01 1.34122789e-01 -4.88724560e-02
-5.60417295e-01 9.32550132e-02 -4.07937884e-01 -7.62030482e-01
-2.35591799e-01 -4.81206328e-01 -7.71909729e-02 -1.09363504e-01
-2.99767792e-01 -1.40974686e-01 4.00521457e-01 -6.88457668e-01
1.33364034e+00 -1.73393786e+00 -6.74801646e-03 2.45104611e-01
-8.80001672e-03 1.70350894e-01 -4.43786420e-02 4.21393603e-01
2.08276808e-01 -5.89744262e-02 -1.25986680e-01 -5.16204298e-01
2.84857899e-01 -4.68974300e-02 -7.05830216e-01 3.79301220e-01
2.90173054e-01 8.79849255e-01 -8.09098601e-01 -2.70084620e-01
1.25612959e-01 2.81905055e-01 -4.73239392e-01 1.90918982e-01
-4.30757433e-01 1.75567865e-01 -6.77692771e-01 7.92444289e-01
5.33454359e-01 -3.58609170e-01 -9.38733369e-02 5.30917756e-02
-4.37556565e-01 6.01610124e-01 -9.87612426e-01 1.27358019e+00
-5.89686692e-01 3.52763593e-01 -2.92180061e-01 -1.20793903e+00
1.09758687e+00 6.65393993e-02 5.31970918e-01 -9.54705894e-01
1.95733681e-01 6.97287977e-01 -6.73322240e-03 1.72494218e-01
5.39704382e-01 -3.15729707e-01 -1.05584309e-01 5.13680875e-01
-1.80701941e-01 4.00520921e-01 3.38249207e-01 6.55632615e-02
9.36158180e-01 6.85146078e-02 -1.75113350e-01 -2.30092719e-01
5.45957088e-01 -4.54181917e-02 9.61400449e-01 3.62926006e-01
-6.36347532e-02 2.26670772e-01 6.09470427e-01 -5.92660844e-01
-9.96885836e-01 -9.65941966e-01 -1.17168888e-01 8.47471416e-01
1.67497352e-03 1.11989761e-02 -2.60571599e-01 -7.05799043e-01
3.07793468e-01 7.83195972e-01 -4.95578259e-01 -1.59854710e-03
-5.79122901e-01 -8.39793265e-01 2.37018958e-01 6.69405460e-01
6.82008564e-01 -1.04719436e+00 -4.79575843e-01 5.02668798e-01
1.22998394e-01 -8.33213449e-01 -5.35835922e-01 1.35609563e-02
-8.78250837e-01 -8.55317056e-01 -8.62985253e-01 -2.90011168e-01
4.49583977e-01 3.55623923e-02 1.13361573e+00 2.05053434e-01
1.15970962e-01 -9.75435302e-02 -3.28421205e-01 -5.71898758e-01
8.11912343e-02 4.13642943e-01 -3.62756364e-02 1.70663193e-01
2.59616971e-01 -5.84598660e-01 -7.80569077e-01 4.32945460e-01
-7.36321688e-01 2.12738495e-02 8.34501147e-01 9.17264521e-01
4.84212518e-01 2.36688435e-01 9.54882860e-01 -1.08000064e+00
3.37389141e-01 -8.02694976e-01 -1.16662669e+00 3.22925061e-01
-1.03077984e+00 1.82583869e-01 8.84058535e-01 -2.74586886e-01
-9.50737238e-01 -4.13778186e-01 5.38680889e-02 -6.16912544e-01
5.91115832e-01 7.75110602e-01 8.43826383e-02 1.42200351e-01
-9.32825282e-02 5.08252978e-01 -1.14706896e-01 -7.18039215e-01
1.49936266e-02 4.99719977e-01 1.06447972e-01 -4.47141409e-01
9.35009062e-01 8.16010013e-02 -7.19543695e-02 -7.92303756e-02
-1.02631462e+00 -2.77511686e-01 -4.71344471e-01 -5.28415293e-02
3.03797215e-01 -8.33800495e-01 -5.34674585e-01 6.32192791e-01
-8.37427378e-01 -4.32896972e-01 -5.37463352e-02 4.56142634e-01
-4.30206299e-01 -1.18161000e-01 -8.04952204e-01 -9.30448711e-01
-6.08553827e-01 -1.09495509e+00 6.45411670e-01 5.95265850e-02
2.52872556e-01 -1.08807659e+00 -1.02813773e-01 3.85528773e-01
8.33390713e-01 5.04778745e-03 7.95179844e-01 -7.52378285e-01
-1.11410570e+00 -3.18987697e-01 -2.49283254e-01 6.07853353e-01
1.34499490e-01 -8.29912350e-02 -6.66164100e-01 -2.77958095e-01
-1.34504428e-02 -3.26062351e-01 1.05818868e+00 2.28835493e-01
9.56452370e-01 -4.81044590e-01 -7.40471035e-02 4.61116254e-01
1.57980907e+00 3.79537016e-01 5.21233141e-01 7.55128324e-01
5.27833402e-01 7.73811877e-01 1.01542819e+00 6.02075040e-01
4.84640002e-01 5.80190718e-01 3.12122256e-01 1.91153228e-01
3.44597459e-01 -4.46175486e-01 5.73267519e-01 9.94384348e-01
1.84904814e-01 -1.31360516e-01 -8.61544371e-01 4.07742888e-01
-1.84375775e+00 -8.63879502e-01 3.35565865e-01 2.16953325e+00
9.50670242e-01 6.52610183e-01 1.99074924e-01 1.69246033e-01
4.40751940e-01 2.52818793e-01 -7.55817652e-01 -1.02897370e-02
1.05300292e-01 2.40371991e-02 9.49579120e-01 3.56599748e-01
-9.92173970e-01 9.83097553e-01 4.66821623e+00 1.01050723e+00
-1.36304760e+00 3.57011624e-04 9.79231000e-01 -3.33491355e-01
-4.31335241e-01 -1.41859576e-01 -1.22670949e+00 9.33796525e-01
1.11241913e+00 -3.88543218e-01 2.92389601e-01 7.03349710e-01
1.37849346e-01 3.15414101e-01 -8.67275417e-01 5.60910642e-01
-3.29659909e-01 -1.59325850e+00 1.94278285e-01 -2.39391048e-02
5.25279641e-01 1.07425094e-01 3.86360049e-01 4.73474950e-01
2.30410814e-01 -7.07081020e-01 1.10512638e+00 6.37533784e-01
3.16991776e-01 -9.91395235e-01 9.90178704e-01 3.99185926e-01
-1.17348647e+00 -3.33463281e-01 -5.56975305e-01 -3.17829363e-02
1.95962340e-01 6.08500302e-01 -5.26652515e-01 6.89201474e-01
7.20036328e-01 1.11830747e+00 -3.37200403e-01 1.00936973e+00
-7.70283565e-02 6.25147283e-01 -3.29464406e-01 -2.43887365e-01
5.97337544e-01 -3.61913025e-01 1.20086208e-01 6.42881811e-01
5.45945227e-01 -9.83406305e-02 -4.54347245e-02 8.25200558e-01
-2.60614276e-01 2.56079763e-01 -3.33232731e-01 -2.74093330e-01
5.29356956e-01 1.02477574e+00 -5.74761450e-01 -2.38559812e-01
-7.51655221e-01 4.82817382e-01 4.08816785e-01 1.18539438e-01
-9.59892511e-01 -3.81555587e-01 5.73178232e-01 2.16073647e-01
5.18780828e-01 8.30965340e-02 -1.75282240e-01 -1.11265934e+00
2.45286614e-01 -7.87083268e-01 3.61719191e-01 -2.68142790e-01
-1.51860261e+00 6.80201888e-01 -1.46523401e-01 -1.48695290e+00
-5.78852035e-02 -6.53538704e-01 -7.34214008e-01 8.99700522e-01
-2.10119176e+00 -9.71784294e-01 3.02797318e-01 2.47585863e-01
5.69787562e-01 -5.55552006e-01 1.60390109e-01 7.13665128e-01
-8.68878126e-01 7.28502154e-01 2.09385052e-01 1.39294565e-01
4.02509779e-01 -1.35774052e+00 4.59981710e-01 7.30746448e-01
1.69668719e-01 3.61682475e-01 5.37590623e-01 -5.27090967e-01
-1.35397494e+00 -1.13259900e+00 6.23459518e-01 -2.39985585e-01
1.27224898e+00 -2.18389690e-01 -9.62018311e-01 8.26860309e-01
1.94765732e-03 -1.28280282e-01 4.65371192e-01 -1.28452644e-01
-2.89114654e-01 -7.70466149e-01 -8.96230280e-01 2.68235356e-01
6.58884645e-01 -1.63649946e-01 -2.57803857e-01 1.19595364e-01
7.38834202e-01 -6.44317344e-02 -9.52065825e-01 3.26155394e-01
5.20515323e-01 -1.04438245e+00 8.92521501e-01 -4.40981448e-01
5.00697196e-01 -6.85573518e-02 -1.03124969e-01 -1.11934340e+00
-9.82702225e-02 -5.74496448e-01 -2.59630352e-01 1.40367281e+00
8.12263250e-01 -1.12616026e+00 9.35334206e-01 7.12985277e-01
-1.75608695e-01 -1.32071650e+00 -9.68180776e-01 -8.71863127e-01
3.40884566e-01 -2.58736461e-01 8.55160058e-01 8.06985140e-01
-2.31168002e-01 1.36114866e-01 -5.90696394e-01 4.42613438e-02
7.04634428e-01 5.05887389e-01 5.58675945e-01 -1.09185767e+00
-5.12517869e-01 -8.40807438e-01 -1.91254526e-01 -1.19121909e+00
3.22045475e-01 -7.21992671e-01 -2.53622830e-01 -1.26258290e+00
-3.86092998e-02 -1.01650000e+00 -9.81477499e-01 3.92270207e-01
-4.07436341e-02 -6.28777668e-02 2.75720119e-01 4.12503272e-01
-4.55667794e-01 8.47758651e-01 1.06540263e+00 -2.73461521e-01
-3.80533040e-02 4.54611391e-01 -6.34154022e-01 4.64527160e-01
1.34363592e+00 -5.54752290e-01 -3.82433116e-01 -5.07593930e-01
3.30744088e-01 1.60069019e-01 -6.64267391e-02 -6.85096323e-01
3.47987115e-01 -1.04565740e-01 2.48378515e-01 -8.31156492e-01
4.16881442e-01 -7.40016997e-01 -4.23374660e-02 7.10717857e-01
-2.90435076e-01 4.14723903e-01 1.00688986e-01 4.56467599e-01
-5.18258214e-01 -2.20770597e-01 5.30552268e-01 -1.34124592e-01
-7.15099514e-01 7.05133677e-01 1.61577970e-01 -6.54101595e-02
9.42973077e-01 1.95338145e-01 -4.07232463e-01 -2.08991930e-01
-5.67784160e-03 7.15673089e-01 1.37918338e-01 4.36255008e-01
5.53599179e-01 -1.33912146e+00 -6.58399761e-01 -3.20385620e-02
-2.93896914e-01 -2.04723179e-01 1.65694151e-02 9.14397895e-01
-6.11335337e-01 6.30582869e-01 -1.65676087e-01 -2.25799214e-02
-7.18607843e-01 5.05160153e-01 4.83327299e-01 -6.30818605e-01
-4.75875914e-01 8.72281313e-01 1.00435168e-01 -1.40549704e-01
3.87641549e-01 -5.11815131e-01 -1.59410104e-01 2.28967026e-01
6.92070186e-01 5.02786756e-01 -3.05302050e-02 -3.72202039e-01
-1.41463742e-01 5.50863862e-01 -3.18831652e-01 -2.47412100e-02
1.74276376e+00 2.76340600e-02 -1.48238793e-01 8.32862079e-01
1.13833082e+00 -2.94470072e-01 -1.41449261e+00 -4.44051474e-01
5.31190395e-01 -6.48409247e-01 2.47170791e-01 -7.75228679e-01
-1.67423260e+00 1.08200824e+00 3.12248647e-01 2.31275082e-01
1.06715131e+00 -3.35022420e-01 1.20945191e+00 2.77699828e-01
5.90907693e-01 -9.56548035e-01 3.41174216e-03 2.46129826e-01
8.10786545e-01 -1.27137816e+00 -5.15889488e-02 -1.09222062e-01
-6.34150147e-01 8.65987778e-01 6.52899802e-01 -2.29505405e-01
8.18325043e-01 3.05325508e-01 1.62049904e-01 -1.30646989e-01
-1.19373071e+00 1.12602405e-01 2.98370928e-01 -3.02107573e-01
2.93574005e-01 -1.80651367e-01 -1.25705078e-01 7.19074786e-01
-1.91921234e-01 3.27718854e-02 3.71936679e-01 7.64498591e-01
-4.81075406e-01 -1.32955849e+00 1.34828493e-01 8.16978157e-01
-8.29569578e-01 -1.25671744e-01 3.63108143e-03 8.09342146e-01
-3.50313306e-01 5.14938235e-01 3.63425553e-01 -3.75698388e-01
4.28118616e-01 -1.05008528e-01 8.91289935e-02 -5.08924127e-01
-7.41500974e-01 -2.32865550e-02 -7.70648569e-02 -4.95579839e-01
-2.17791006e-01 -6.72983587e-01 -1.15468657e+00 -5.82523823e-01
-2.60613084e-01 2.49313056e-01 3.27285022e-01 6.29940808e-01
1.08058698e-01 4.28097665e-01 1.08012259e+00 -5.42992949e-01
-1.13606513e+00 -9.38434541e-01 -9.78809953e-01 6.24656193e-02
3.68401051e-01 -8.90427291e-01 -4.64607805e-01 -4.34189826e-01] | [4.406030178070068, 4.2534894943237305] |
3e23d0e1-0025-4801-bfa7-bac5e1753ffc | cracking-the-black-box-distilling-deep-sports | 2006.04551 | null | https://arxiv.org/abs/2006.04551v4 | https://arxiv.org/pdf/2006.04551v4.pdf | Cracking the Black Box: Distilling Deep Sports Analytics | This paper addresses the trade-off between Accuracy and Transparency for deep learning applied to sports analytics. Neural nets achieve great predictive accuracy through deep learning, and are popular in sports analytics. But it is hard to interpret a neural net model and harder still to extract actionable insights from the knowledge implicit in it. Therefore, we built a simple and transparent model that mimics the output of the original deep learning model and represents the learned knowledge in an explicit interpretable way. Our mimic model is a linear model tree, which combines a collection of linear models with a regression-tree structure. The tree version of a neural network achieves high fidelity, explains itself, and produces insights for expert stakeholders such as athletes and coaches. We propose and compare several scalable model tree learning heuristics to address the computational challenge from datasets with millions of data points. | ['Xiangyu Sun', 'Oliver Schulte', 'Jack Davis', 'Guiliang Liu'] | 2020-06-04 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [-2.37696916e-01 8.28849018e-01 -8.21435153e-01 -5.72587669e-01
-2.65698016e-01 -5.81662118e-01 1.99734896e-01 1.51486889e-01
-5.72756119e-02 6.20823026e-01 5.76921940e-01 -4.55129296e-01
-3.28897476e-01 -9.56299126e-01 -1.18318725e+00 1.73538309e-02
-2.29073286e-01 7.27027118e-01 -1.96561351e-01 -1.89576641e-01
-3.34531143e-02 1.15256533e-01 -1.36686862e+00 8.37785482e-01
8.53787005e-01 1.09302127e+00 -5.50129890e-01 5.95848680e-01
-6.11479133e-02 1.72990119e+00 -3.81067574e-01 -8.75075459e-01
6.97694957e-01 -2.65089516e-02 -9.29997802e-01 -2.99266398e-01
4.73791480e-01 -4.07744676e-01 -4.40406471e-01 6.10507250e-01
-3.04045901e-02 -2.51909494e-01 3.48668128e-01 -1.63069761e+00
-7.48053074e-01 1.21777105e+00 -2.75896460e-01 1.24117218e-01
-1.80915654e-01 7.91175902e-01 1.25034606e+00 -3.11721176e-01
4.56101567e-01 1.25479341e+00 1.10219681e+00 5.46300292e-01
-1.31235373e+00 -8.16098332e-01 2.56500185e-01 3.37966979e-01
-7.23746061e-01 -2.95791030e-01 5.02563179e-01 -8.36726964e-01
8.27814221e-01 3.51002902e-01 1.32511652e+00 1.07925928e+00
3.93169552e-01 1.09229970e+00 1.04950571e+00 2.87957136e-02
3.21249962e-01 -1.34139322e-02 5.50864220e-01 6.62800610e-01
5.91102004e-01 5.52786827e-01 -1.19116604e+00 1.17471546e-01
8.67693126e-01 2.55251855e-01 3.10706079e-01 -3.20222050e-01
-6.69814706e-01 8.45035017e-01 7.61544883e-01 -4.07521963e-01
-5.04666209e-01 7.59008110e-01 4.81577307e-01 3.60614359e-01
5.50421774e-01 7.92767763e-01 -6.68404400e-01 -5.44815540e-01
-8.21259797e-01 5.73881328e-01 1.11105442e+00 7.43532300e-01
6.31353140e-01 2.97972172e-01 -5.66349700e-02 2.35812724e-01
6.00323714e-02 6.03290014e-02 -1.75063591e-02 -1.51174629e+00
3.09367955e-01 1.03430450e+00 -1.41153440e-01 -1.05457473e+00
-3.97789240e-01 -9.44254875e-01 -8.28722179e-01 4.46944267e-01
4.37209994e-01 -2.89229810e-01 -8.46566021e-01 1.42437720e+00
1.05278589e-01 8.61961320e-02 5.54086939e-02 1.09718180e+00
1.00472534e+00 3.49225223e-01 9.40453634e-02 4.64040071e-01
9.89946365e-01 -1.05996656e+00 -5.38602173e-01 -7.44178653e-01
5.07340729e-01 2.27616310e-01 1.07249272e+00 8.20683122e-01
-1.24481714e+00 -4.29167420e-01 -9.84304011e-01 -4.20768499e-01
-2.62352169e-01 -2.70642579e-01 1.15179646e+00 3.08923930e-01
-8.53966415e-01 7.39046574e-01 -9.36730921e-01 2.10805327e-01
8.89522135e-01 5.00265718e-01 -2.08614647e-01 1.26322389e-01
-1.19477487e+00 8.90532374e-01 6.65361047e-01 1.97403848e-01
-8.86336267e-01 -1.18882096e+00 -6.44854426e-01 5.08249342e-01
7.42533147e-01 -9.79336143e-01 1.58856642e+00 -1.40411675e+00
-1.38646162e+00 7.77682900e-01 1.83711424e-01 -1.03098500e+00
7.87875950e-01 -7.08632827e-01 1.15223274e-01 -4.56912458e-01
-8.07569325e-02 4.59548533e-01 4.44995731e-01 -1.24511623e+00
-5.91690540e-01 -3.78560036e-01 4.50435013e-01 -2.32416809e-01
-5.64519763e-02 -2.29610816e-01 -1.92909583e-01 -4.29168522e-01
-1.91882655e-01 -6.74661219e-01 -4.97177392e-01 1.24637112e-01
-3.46828043e-01 -1.79007828e-01 2.46269733e-01 -8.32241297e-01
1.42682731e+00 -1.83245051e+00 -6.19398952e-02 3.60869288e-01
1.14618564e+00 1.16060317e-01 5.53338155e-02 2.89797157e-01
1.59331374e-02 5.71693480e-01 1.53844893e-01 1.05432525e-01
1.43595070e-01 6.79627240e-01 -3.83633912e-01 -1.56717956e-01
2.71522373e-01 1.39178634e+00 -9.51885462e-01 -4.45398539e-01
-1.36076510e-01 -7.13408142e-02 -9.41640615e-01 1.90377474e-01
-6.00683630e-01 -2.33585052e-02 -4.38440293e-01 6.41720533e-01
2.89042592e-01 -3.26934874e-01 4.60970759e-01 3.93412933e-02
-1.11395970e-01 4.09417242e-01 -9.56872344e-01 1.25573361e+00
-1.74707979e-01 9.13230836e-01 4.20813076e-02 -1.08531618e+00
8.91348302e-01 -7.97034949e-02 4.12046343e-01 -7.44906664e-01
-1.36614546e-01 -8.96184742e-02 2.55069673e-01 -7.39839494e-01
3.83947045e-01 -1.33494943e-01 -2.05587789e-01 5.88129997e-01
-2.63273325e-02 1.22180484e-01 -4.13884036e-02 1.10050455e-01
1.15869415e+00 2.12056458e-01 4.38035965e-01 -1.42943859e-01
-5.10491669e-01 5.79289973e-01 1.02543664e+00 1.20044279e+00
7.74915516e-02 2.18747005e-01 1.10386801e+00 -1.66716325e+00
-1.00831819e+00 -9.90594566e-01 3.74708205e-01 1.37018788e+00
-3.55385870e-01 -8.48547637e-01 -5.87103963e-01 -5.79256833e-01
4.69998568e-01 4.46932524e-01 -1.00050330e+00 -3.41631502e-01
-5.88461101e-01 -2.40713954e-01 7.47336566e-01 9.98803139e-01
4.53580469e-01 -8.52071404e-01 -7.53282726e-01 2.84190595e-01
-1.38727665e-01 -8.50706756e-01 1.15843855e-01 4.44871098e-01
-1.03332281e+00 -1.10058558e+00 4.77834642e-01 -8.06536898e-02
2.91354023e-02 -2.49430016e-01 1.79593241e+00 2.26706401e-01
6.15957950e-04 -3.75387743e-02 -9.96408090e-02 -1.25809360e+00
-4.88616168e-01 3.75585675e-01 1.20471083e-01 -1.89511076e-01
7.20656455e-01 -5.25974214e-01 -4.60094690e-01 -7.95757249e-02
-6.37637436e-01 5.16587496e-01 7.65062273e-01 8.76780808e-01
5.06384313e-01 -1.93849713e-01 3.68196279e-01 -1.11981535e+00
7.45602190e-01 -5.66436648e-01 -3.71341944e-01 2.97827780e-01
-9.57502067e-01 3.69624376e-01 5.85999489e-01 -3.27820152e-01
-5.69158673e-01 1.39993101e-01 2.83292770e-01 -4.08204645e-01
2.11536922e-02 9.96548235e-01 9.88496989e-02 3.14744800e-01
1.07208979e+00 -1.09223396e-01 1.81379691e-01 -6.31255388e-01
5.47311008e-01 5.45172036e-01 8.59668136e-01 -8.55469644e-01
6.31896496e-01 2.26886585e-01 2.90860664e-02 -1.85544804e-01
-1.28824675e+00 1.01272717e-01 -5.71174026e-01 -4.78471041e-01
6.48059070e-01 -1.13782084e+00 -1.23496366e+00 9.87835750e-02
-1.00778413e+00 -7.61151791e-01 -9.53195930e-01 1.38888314e-01
-6.07570708e-01 -2.94614553e-01 -4.86730874e-01 -6.20923042e-01
-4.57233310e-01 -4.96767551e-01 5.35519838e-01 2.01182887e-01
-6.29383922e-01 -8.67782593e-01 7.56077282e-03 6.45498157e-01
3.51116061e-01 5.78441441e-01 9.83812153e-01 -8.17073643e-01
-8.87317479e-01 -1.60486549e-01 -2.41147846e-01 4.03390408e-01
-3.49677801e-01 9.35049206e-02 -9.58609104e-01 2.37417981e-01
-3.87962013e-01 -5.56213975e-01 8.75920653e-01 6.14050567e-01
1.69239128e+00 -9.95626032e-01 1.56822968e-02 7.94864833e-01
1.02595425e+00 -3.54305834e-01 5.40110588e-01 7.55402505e-01
8.72028649e-01 6.97306573e-01 3.57402146e-01 3.51165146e-01
6.99916661e-01 2.83401996e-01 6.66949511e-01 -3.56338352e-01
1.80958211e-01 -7.41072118e-01 1.68233991e-01 4.21582878e-01
-7.07050443e-01 1.96709186e-01 -1.33671474e+00 4.18277115e-01
-2.49860644e+00 -1.18245566e+00 -2.66822755e-01 1.70570600e+00
9.75618124e-01 5.12804866e-01 4.49683189e-01 -4.81014140e-03
2.74975151e-02 -4.36249077e-02 -1.07064128e+00 -9.34500217e-01
6.50706142e-02 2.24875540e-01 5.68267524e-01 3.13815951e-01
-7.52307296e-01 1.06858850e+00 7.81734562e+00 5.03964663e-01
-8.85557532e-01 -3.09263095e-02 7.82007813e-01 -7.00122595e-01
-3.84834617e-01 1.62851691e-01 -3.99911195e-01 1.45731628e-01
1.11926305e+00 -3.65567029e-01 4.51750785e-01 1.24329293e+00
4.37065631e-01 2.85901338e-01 -1.53147101e+00 8.31178963e-01
-4.46679384e-01 -2.05703926e+00 2.10800767e-01 3.04306984e-01
6.86823487e-01 2.13608101e-01 7.80047104e-02 5.33332229e-01
1.17760170e+00 -1.67735207e+00 1.00085926e+00 9.78748262e-01
4.68461782e-01 -6.52429283e-01 5.48787296e-01 6.89045787e-01
-7.20587730e-01 -8.03859413e-01 -2.16200963e-01 -1.08960819e+00
-4.01569039e-01 5.24879515e-01 -7.59327590e-01 1.73571885e-01
1.06315660e+00 1.06522501e+00 -6.03923380e-01 8.76765251e-01
-1.52567565e-01 1.08394635e+00 -1.91689193e-01 8.86215195e-02
3.30477417e-01 -1.60246909e-01 2.27735370e-01 1.13601196e+00
-2.55076051e-01 7.56698102e-02 1.50937930e-01 1.29370880e+00
-2.87623525e-01 -3.15162897e-01 -7.24024415e-01 -1.76492587e-01
2.31700182e-01 9.13597405e-01 9.30984616e-02 -4.64862287e-01
-1.80169478e-01 -1.13787338e-01 7.85147309e-01 3.41726094e-01
-4.43910688e-01 3.29190105e-01 1.12653852e+00 5.35768449e-01
-1.35253817e-01 -3.76314446e-02 -1.17321956e+00 -1.09453940e+00
1.22473367e-01 -1.43973494e+00 5.29977798e-01 -8.38947952e-01
-1.36066115e+00 1.65336147e-01 -5.15101664e-02 -1.02460563e+00
-6.06271505e-01 -6.22128963e-01 -6.48872435e-01 6.33618593e-01
-1.12929153e+00 -1.41336882e+00 -3.46123785e-01 1.91285029e-01
4.07438815e-01 -1.78668946e-01 4.15293127e-01 -2.96207219e-01
-5.28302848e-01 4.83822852e-01 -5.31140640e-02 4.86217976e-01
1.61716640e-01 -1.41011345e+00 6.88486516e-01 5.98722160e-01
-4.83230650e-02 6.66398108e-01 8.14071953e-01 -5.81417561e-01
-1.35309720e+00 -1.08329403e+00 7.38117993e-01 -1.00972831e+00
7.24493444e-01 -3.48087549e-01 -8.33431423e-01 1.19205356e+00
7.95558169e-02 -2.40908593e-01 8.65696907e-01 7.60648966e-01
-5.57053924e-01 -5.07007480e-01 -7.24792957e-01 4.95327979e-01
1.17203569e+00 -3.97182941e-01 -7.31055915e-01 5.41372485e-02
7.52660990e-01 -5.30476093e-01 -1.03341496e+00 3.08260441e-01
1.01123130e+00 -1.06472480e+00 8.97479832e-01 -1.61077082e+00
1.21753812e+00 1.33585691e-01 2.23829255e-01 -1.35311127e+00
-5.15899062e-01 -6.68182194e-01 -8.04357648e-01 6.38149142e-01
4.12434369e-01 -3.83746564e-01 1.12506890e+00 1.28209412e+00
-1.98515519e-01 -1.21671069e+00 -5.96659243e-01 -6.70559824e-01
3.23126137e-01 -7.36687183e-01 1.05978251e+00 1.08617604e+00
7.40751997e-02 1.73454106e-01 -5.52420497e-01 -4.11741696e-02
8.44899178e-01 2.20933408e-01 1.31211352e+00 -1.65391636e+00
-5.26376128e-01 -5.43735325e-01 -4.07998651e-01 -9.86923516e-01
-6.71764538e-02 -8.05684686e-01 -3.35934222e-01 -1.65406466e+00
4.48981106e-01 -3.90976638e-01 -2.51289338e-01 1.14457273e+00
7.14370385e-02 -1.13427386e-01 3.26983243e-01 2.27270693e-01
-5.61339736e-01 -7.71644805e-03 1.07353127e+00 -2.96642601e-01
-2.36981690e-01 4.21531647e-02 -1.40824044e+00 1.06097817e+00
7.49667466e-01 -6.46369338e-01 -1.07024007e-01 -8.43813598e-01
9.10948455e-01 -2.19941847e-02 8.97220433e-01 -7.86515355e-01
2.66761005e-01 -7.81004608e-01 5.03140032e-01 -3.31960231e-01
3.78245153e-02 -6.15547478e-01 4.72533256e-01 5.75862408e-01
-8.39847028e-01 -2.43165046e-02 3.45477834e-02 4.41381246e-01
-2.96464562e-01 2.46647432e-01 2.01660529e-01 -4.42921817e-01
-6.86043978e-01 3.82662505e-01 -1.42789423e-01 2.02785358e-01
6.21598363e-01 -4.97900307e-01 -5.07603884e-01 -6.71024799e-01
-9.68228281e-01 7.34120250e-01 1.28492609e-01 5.42647243e-01
5.34375787e-01 -1.21133673e+00 -8.73106420e-01 6.46069124e-02
1.01241082e-01 3.88467014e-01 -8.51340592e-02 3.77375841e-01
-6.35669172e-01 2.15915889e-01 -5.73219419e-01 -4.64278430e-01
-9.47541177e-01 3.58714968e-01 6.64439976e-01 -5.76781154e-01
-5.50614715e-01 5.55125833e-01 -2.65763197e-02 -6.00637615e-01
1.91315264e-01 -8.39162707e-01 -1.04627125e-02 -2.68438131e-01
4.45196927e-01 4.45543945e-01 -3.60868067e-01 8.68698657e-02
4.66293246e-02 1.34000286e-01 1.33415490e-01 2.52259195e-01
1.85204828e+00 2.88335145e-01 9.05314926e-03 6.14496171e-01
4.70890701e-01 -4.82620597e-01 -1.68090951e+00 -3.64643246e-01
9.64229628e-02 -4.44507629e-01 6.46514073e-02 -1.10273612e+00
-1.13159764e+00 1.00345230e+00 1.74382165e-01 2.98608601e-01
7.97764659e-01 6.75813928e-02 6.52809799e-01 6.68202937e-01
6.63124919e-02 -1.40217495e+00 6.45915791e-02 4.37679499e-01
8.95487428e-01 -1.34944999e+00 2.18883008e-01 -3.06934793e-03
-8.35959971e-01 1.25197029e+00 1.06416976e+00 -1.60401449e-01
4.86954510e-01 4.79077011e-01 1.96138740e-01 -5.70357203e-01
-1.27349865e+00 1.67066026e-02 3.91806334e-01 6.56714737e-01
9.94820297e-02 3.97342235e-01 2.46170610e-01 1.29919100e+00
-8.90431046e-01 7.62574553e-01 4.10182565e-01 6.12416863e-01
-3.22115541e-01 -7.19527483e-01 -1.14843756e-01 9.00373101e-01
-4.27675217e-01 -1.24685150e-02 -8.11185181e-01 8.53087485e-01
4.17542785e-01 7.10010648e-01 1.10370502e-01 -9.25448298e-01
4.16659862e-01 3.34236883e-02 1.26245776e-02 -5.01499176e-01
-9.90698516e-01 -4.96037871e-01 3.75201315e-01 -9.47880387e-01
-6.73269853e-02 -3.14902663e-01 -1.05488920e+00 -8.78019333e-01
2.61317641e-01 6.20053057e-03 5.11972308e-01 1.03418314e+00
4.97458369e-01 5.23095965e-01 2.60718912e-01 -4.82549280e-01
-5.63065052e-01 -7.03806698e-01 -3.54674965e-01 3.23078126e-01
3.30670238e-01 -3.96539778e-01 3.18199843e-02 1.84287757e-01] | [8.979409217834473, 6.14611291885376] |
865f6356-8a52-46a1-82d7-3116afb58b41 | activity-detection-for-grant-free-noma-in | 2301.01274 | null | https://arxiv.org/abs/2301.01274v1 | https://arxiv.org/pdf/2301.01274v1.pdf | Activity Detection for Grant-Free NOMA in Massive IoT Networks | Recently, grant-free transmission paradigm has been introduced for massive Internet of Things (IoT) networks to save both time and bandwidth and transmit the message with low latency. In order to accurately decode the message of each device at the base station (BS), first, the active devices at each transmission frame must be identified. In this work, first we investigate the problem of activity detection as a threshold comparing problem. We show the convexity of the activity detection method through analyzing its probability of error which makes it possible to find the optimal threshold for minimizing the activity detection error. Consequently, to achieve an optimum solution, we propose a deep learning (DL)-based method called convolutional neural network (CNN)-activity detection (AD). In order to make it more practical, we consider unknown and time-varying activity rate for the IoT devices. Our simulations verify that our proposed CNN-AD method can achieve higher performance compared to the existing non-Bayesian greedy-based methods. This is while existing methods need to know the activity rate of IoT devices, while our method works for unknown and even time-varying activity rates | ['Masoud Ardakani', 'Mostafa Mohammadkarimi', 'Mehrtash Mehrabi'] | 2022-12-23 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [ 2.90802777e-01 1.10392787e-01 -5.41670263e-01 -2.10498497e-01
-3.46185833e-01 -2.92825729e-01 2.42342371e-02 -1.44270569e-01
-3.97345841e-01 9.28941250e-01 -1.41998846e-02 -5.45647800e-01
-3.25358838e-01 -9.26668465e-01 -5.49090683e-01 -1.01508260e+00
-2.80976534e-01 3.16203028e-01 3.99055421e-01 6.19790733e-01
-1.92998514e-01 3.61149520e-01 -9.35180128e-01 -2.83114046e-01
5.99412978e-01 1.65119278e+00 3.10993463e-01 5.79412758e-01
1.59949377e-01 6.71816945e-01 -7.93417633e-01 3.82378288e-02
2.07874611e-01 -2.97420561e-01 -4.66268390e-01 -5.21542765e-02
-4.44121927e-01 -7.35771596e-01 -2.91587085e-01 1.02242172e+00
5.28856456e-01 -4.90048409e-01 5.55260837e-01 -1.38473213e+00
-1.61610190e-02 9.54941571e-01 -6.83767438e-01 4.41469103e-01
9.90467742e-02 -2.94479996e-01 5.81802368e-01 -8.24374631e-02
-3.09384074e-02 9.84508216e-01 3.83915514e-01 2.37666368e-01
-4.45342243e-01 -1.04742408e+00 1.06305189e-01 1.98953941e-01
-1.39632392e+00 -4.09767210e-01 7.02910066e-01 -6.94620470e-03
5.79146087e-01 6.54309914e-02 8.19959104e-01 9.45989311e-01
2.22546354e-01 8.45374644e-01 4.67487872e-01 -3.20343763e-01
4.91246998e-01 -3.43506396e-01 -2.34305248e-01 5.42813957e-01
7.73663461e-01 -1.87990874e-01 -1.55037427e-02 -1.25276566e-01
7.56992280e-01 2.15574712e-01 3.31459381e-02 7.34870806e-02
-1.40756512e+00 4.39144194e-01 3.15751016e-01 5.75944126e-01
-6.57993913e-01 8.28918099e-01 5.76444268e-02 6.07269220e-02
3.89594972e-01 -3.48402292e-01 -7.43702471e-01 -2.95898944e-01
-7.27330089e-01 -3.82079661e-01 1.03381264e+00 9.60325241e-01
4.65639532e-01 9.90564600e-02 -4.53133255e-01 3.26617777e-01
8.13387454e-01 8.22763979e-01 -7.96342716e-02 -9.44534421e-01
6.61604047e-01 4.78143632e-01 2.33766422e-01 -8.12388420e-01
-7.19946086e-01 -7.63209999e-01 -1.29357827e+00 -4.43306267e-01
2.68057376e-01 -7.10255682e-01 -5.47423065e-01 1.65501595e+00
3.63094747e-01 8.23747277e-01 5.80310300e-02 5.14172614e-01
1.27476841e-01 1.02025151e+00 1.34200066e-01 -9.23658788e-01
1.35557425e+00 -3.89996111e-01 -9.66312408e-01 -8.13298002e-02
5.83792984e-01 -4.56234694e-01 2.92257100e-01 4.76074666e-01
-1.04110992e+00 -1.08192392e-01 -1.43971694e+00 5.84206700e-01
-2.54652258e-02 4.67220247e-01 5.83587646e-01 1.19095397e+00
-6.95218146e-01 1.09815918e-01 -1.05253708e+00 -2.96150148e-01
7.17653155e-01 8.15427482e-01 5.10267198e-01 -2.30283979e-02
-9.77517128e-01 2.47170791e-01 4.20273453e-01 2.18348354e-01
-9.95571613e-01 -2.96781629e-01 -3.64650071e-01 1.79671243e-01
5.04726589e-01 -5.74878752e-01 1.37810194e+00 -6.52588725e-01
-1.58068395e+00 6.64149746e-02 -1.30702347e-01 -4.83724862e-01
3.97046089e-01 1.38610914e-01 -3.43814313e-01 1.45731822e-01
-1.90175325e-01 1.60051629e-01 5.87735295e-01 -8.19012761e-01
-1.02426136e+00 -3.42528909e-01 3.21424752e-01 -2.48303652e-01
-6.17702603e-01 2.79920492e-02 -5.81792951e-01 -2.82680809e-01
2.87294775e-01 -7.91135311e-01 -8.12915117e-02 5.63923903e-02
-4.01526034e-01 -6.90291345e-01 9.99949455e-01 -3.08553904e-01
1.20465493e+00 -1.84831297e+00 -4.33128983e-01 2.29684979e-01
3.34610671e-01 1.68279722e-01 3.21298152e-01 1.10453419e-01
6.59369051e-01 5.90931959e-02 1.81378871e-01 -2.33312756e-01
-1.39476672e-01 3.52069378e-01 3.30966055e-01 6.34128153e-01
-1.30804881e-01 7.36474693e-01 -9.49118316e-01 -6.85151756e-01
3.93287867e-01 5.80187917e-01 -3.89581352e-01 1.96045842e-02
-2.65549481e-01 6.42147839e-01 -1.01204109e+00 6.32499993e-01
9.37092841e-01 -5.82791746e-01 4.11560208e-01 -5.29203653e-01
-6.42498881e-02 1.95531294e-01 -1.28636658e+00 1.25873840e+00
-8.30257773e-01 3.31855834e-01 1.61175296e-01 -1.22979379e+00
4.43234831e-01 6.01017416e-01 1.15075624e+00 -8.32023203e-01
6.27720833e-01 1.28724813e-01 7.66210258e-03 -4.07765865e-01
-2.00927809e-01 1.01463035e-01 -1.20866656e-01 5.15679300e-01
-3.12987477e-01 8.44302297e-01 7.09758475e-02 9.91320089e-02
1.47107828e+00 -5.58983684e-01 4.28545773e-01 -1.17334016e-01
5.32391965e-01 -7.92849541e-01 8.38797271e-01 9.61175144e-01
-2.29247257e-01 -1.25824004e-01 2.98103958e-01 -2.92215198e-01
-6.27673686e-01 -1.15671229e+00 -6.97780252e-02 6.38544738e-01
4.38747197e-01 -1.55030878e-03 -6.90102100e-01 -5.08885682e-01
-3.74418199e-01 2.04020709e-01 -2.81545907e-01 1.01803034e-03
-6.11766517e-01 -1.15825355e+00 4.82058108e-01 4.86647278e-01
1.17759132e+00 -5.40255308e-01 -8.01338196e-01 6.11744881e-01
-4.23030823e-01 -1.51107812e+00 -1.17317438e-01 2.05191225e-01
-6.23447716e-01 -1.05868256e+00 -4.56372708e-01 -5.59988618e-01
4.79131371e-01 3.64007324e-01 7.04548955e-01 1.25508428e-01
1.37925014e-01 2.62529284e-01 -3.78704876e-01 -7.01859593e-01
-2.06495926e-01 3.75791848e-01 -7.24494085e-02 2.00252160e-01
4.76371825e-01 -6.73631012e-01 -1.09422004e+00 6.84393883e-01
-7.09406078e-01 -3.48146945e-01 8.87718141e-01 2.05075624e-03
2.58854479e-01 4.95702952e-01 9.19637620e-01 -1.45919010e-01
5.38769901e-01 -7.11551607e-01 -1.00381291e+00 3.38572234e-01
-5.54017663e-01 -3.62557210e-02 4.63684142e-01 -6.02356076e-01
-7.41980970e-01 2.13139370e-01 -1.86287910e-01 1.34461954e-01
2.76946127e-01 2.10414991e-01 -5.36430538e-01 3.78542356e-02
4.08623330e-02 -8.38326581e-04 -6.41599536e-01 -3.06386799e-01
-4.56238650e-02 1.18493497e+00 7.14391470e-02 -3.50195020e-01
4.69596386e-01 7.68462956e-01 2.37950757e-01 -8.15560162e-01
-8.97476792e-01 -4.07986611e-01 -2.03787953e-01 -4.21355218e-01
8.48178267e-01 -1.04268062e+00 -1.51443827e+00 4.68310028e-01
-1.46786058e+00 -1.52348429e-02 6.67463690e-02 6.84095383e-01
-1.67787150e-01 3.02964509e-01 -4.99463528e-01 -1.20483589e+00
-4.16713536e-01 -1.15812576e+00 1.13597429e+00 1.86452255e-01
2.62309641e-01 -8.13877940e-01 -2.64066637e-01 1.00235105e-03
5.53595722e-01 8.53251666e-02 6.39585793e-01 -3.89942795e-01
-1.07677817e+00 -3.12605709e-01 -5.31206369e-01 1.59664184e-01
2.85343617e-01 -3.94275278e-01 -7.09899902e-01 -3.46872002e-01
1.64237022e-01 1.95176348e-01 4.17115420e-01 8.62857401e-01
1.40826249e+00 -6.74290001e-01 -8.00574720e-01 5.91134489e-01
1.22943175e+00 6.22913897e-01 8.51975322e-01 1.07300341e-01
4.22292948e-01 -2.28490233e-02 3.92503768e-01 8.65327299e-01
3.91707957e-01 7.39931941e-01 8.08612525e-01 -2.23170798e-02
9.68513265e-02 3.71486135e-02 2.50822246e-01 7.86163330e-01
1.31020561e-01 -1.09814012e+00 -4.34220016e-01 3.73696268e-01
-1.93152273e+00 -7.37542331e-01 -1.25718176e-01 2.26471591e+00
3.52276325e-01 3.48642975e-01 1.71299547e-01 5.83798707e-01
7.44000554e-01 5.48015498e-02 -6.74356580e-01 1.36616692e-01
2.09365919e-01 -1.24586515e-01 9.26405370e-01 2.03956813e-01
-9.69260871e-01 2.50699461e-01 6.16818142e+00 8.93274426e-01
-8.46854568e-01 5.94755590e-01 4.69531685e-01 -3.85176018e-02
1.09748088e-01 -3.20616513e-01 -8.25632572e-01 9.21956241e-01
1.13588512e+00 5.67721203e-02 1.87708035e-01 6.38090611e-01
8.09569359e-01 -3.31320286e-01 -8.08557034e-01 1.19636059e+00
-3.28365952e-01 -1.01160812e+00 -3.17743957e-01 5.20050466e-01
5.04031599e-01 1.93790883e-01 -3.95926088e-01 -1.22709952e-01
4.50923592e-01 -5.13335705e-01 5.38849831e-01 2.50775963e-01
5.24761558e-01 -7.50250995e-01 1.01631391e+00 4.52783734e-01
-1.54080975e+00 -6.35650933e-01 -4.17571813e-01 -1.07187562e-01
5.74770749e-01 1.35878265e+00 -7.05151558e-01 4.37229872e-01
6.07690930e-01 5.18288136e-01 1.27630398e-01 1.20204151e+00
-1.86207563e-01 9.83339250e-01 -9.14392650e-01 -6.14284635e-01
6.37176335e-02 1.51877254e-01 2.78659999e-01 9.51491773e-01
7.34556973e-01 9.09833536e-02 2.58464903e-01 5.73297620e-01
-5.24071634e-01 -2.28477478e-01 -1.93975657e-01 2.18747601e-01
8.18185747e-01 9.71400619e-01 -1.01463783e+00 -2.70977706e-01
-4.27891821e-01 6.75272822e-01 -2.53165811e-01 2.16906667e-01
-1.01110518e+00 -1.46526426e-01 3.57267857e-01 1.94370523e-01
3.16864818e-01 -4.52879757e-01 -2.49141112e-01 -8.36440980e-01
1.42963648e-01 -1.69588670e-01 1.99086279e-01 -5.34268856e-01
-8.61044347e-01 1.69790105e-03 1.64320111e-01 -1.26020169e+00
-5.68606779e-02 -3.21468413e-01 -4.80756372e-01 1.11579552e-01
-1.51467311e+00 -1.00801444e+00 -4.57559347e-01 6.87444568e-01
3.41628879e-01 1.94566861e-01 2.02033788e-01 8.81049275e-01
-6.07315063e-01 4.00243074e-01 1.74499154e-01 1.97650358e-01
-1.57882929e-01 -6.07173860e-01 -2.91840732e-02 9.01409507e-01
-6.38190061e-02 2.06599593e-01 5.38239837e-01 -6.03888214e-01
-1.41906607e+00 -1.06519115e+00 5.55398822e-01 4.76423576e-02
4.52894270e-01 -4.19349581e-01 -1.21730208e-01 5.38097799e-01
2.49139182e-02 1.07597217e-01 4.11458164e-01 -4.44362491e-01
2.46733904e-01 -8.01123977e-01 -1.29397094e+00 2.95259178e-01
1.17750525e+00 6.65240660e-02 4.36967425e-02 4.54537332e-01
6.97236180e-01 2.03168839e-01 -5.83651960e-01 6.15301788e-01
6.68361008e-01 -5.52397728e-01 8.49158108e-01 1.73425838e-01
-5.30377567e-01 -4.99529779e-01 -8.12602267e-02 -6.47009730e-01
-2.07099710e-02 -8.67525220e-01 -6.16525114e-01 1.12637722e+00
2.21694782e-01 -6.25839770e-01 1.00907671e+00 -9.11793113e-03
2.79713392e-01 -5.27043462e-01 -1.50212538e+00 -9.91652012e-01
-4.40798342e-01 -8.58575702e-01 6.92851722e-01 3.21394712e-01
-2.32874125e-01 3.70126337e-01 -5.15556991e-01 5.12296498e-01
7.41557539e-01 -5.59791446e-01 4.41317499e-01 -1.53315282e+00
-4.54233028e-02 5.75068630e-02 -3.98596197e-01 -1.77553582e+00
-2.41196468e-01 -3.59867156e-01 9.73261595e-02 -1.80451798e+00
1.23663060e-01 -7.11399078e-01 -6.09924734e-01 4.06140774e-01
5.18781006e-01 3.74937057e-02 -1.99498966e-01 1.66613579e-01
-1.06860399e+00 3.33122581e-01 8.48787844e-01 -2.18034983e-01
-7.99912587e-02 5.93807518e-01 -5.08341253e-01 5.63864648e-01
1.10237384e+00 -8.35468888e-01 -6.75453186e-01 -5.19671142e-01
6.47591770e-01 7.42784068e-02 3.18606555e-01 -1.41137862e+00
3.42542231e-01 1.67510360e-01 2.41217747e-01 -8.97460401e-01
1.64886728e-01 -1.29925752e+00 9.50023010e-02 8.52061808e-01
1.93398908e-01 -1.15921654e-01 -2.49628320e-01 9.47096765e-01
3.69563043e-01 -2.31863737e-01 6.13693058e-01 1.49526700e-01
-1.38935760e-01 4.64875221e-01 -7.71651268e-01 -3.96041632e-01
1.16260815e+00 -2.85886437e-01 -2.97707886e-01 -6.91394627e-01
-2.20288366e-01 2.64924616e-01 -1.74907893e-01 1.25720888e-01
4.12431717e-01 -1.12982833e+00 -2.62745619e-01 -1.78994581e-01
-7.38893896e-02 -7.76936635e-02 1.18129766e-02 1.09387958e+00
-3.89854431e-01 5.87022901e-01 4.38408881e-01 -7.22656727e-01
-8.87092650e-01 3.48060578e-01 3.97431105e-01 -4.34448302e-01
-7.43158907e-02 3.42898160e-01 -2.59505749e-01 2.17582360e-01
6.44750893e-01 -7.37028062e-01 -2.76567508e-02 -2.74602622e-01
5.74411273e-01 5.49551308e-01 1.50481895e-01 -1.58230022e-01
-4.39847082e-01 6.59741282e-01 2.53294528e-01 -1.12290476e-02
1.05042255e+00 -5.58234990e-01 1.47239855e-02 -5.59621565e-02
1.29273498e+00 -7.26324692e-02 -1.14026105e+00 -2.75908671e-02
1.03276849e-01 -1.24330059e-01 4.89248097e-01 -6.66185379e-01
-1.40764153e+00 6.65487289e-01 1.02189195e+00 6.51026130e-01
1.22030830e+00 9.65419859e-02 1.04924595e+00 5.76779068e-01
7.91847825e-01 -9.26238418e-01 2.15101376e-01 1.13023318e-01
-6.14910387e-02 -9.06422257e-01 -1.65251326e-02 -5.24737656e-01
4.09353822e-01 1.04332983e+00 3.24045449e-01 4.67301846e-01
1.17392516e+00 5.84492505e-01 -3.43745381e-01 -1.98859781e-01
-4.80991334e-01 -3.54919970e-01 -3.83287966e-01 7.04995811e-01
-1.14509398e-02 -6.70572789e-03 -4.17455405e-01 3.76449585e-01
2.74665743e-01 5.14139295e-01 4.10562456e-01 1.09923446e+00
-8.01698804e-01 -1.05103433e+00 -1.65478960e-01 5.46467423e-01
-8.64646137e-01 2.46333122e-01 1.44951850e-01 4.06446666e-01
3.48088741e-01 1.56752837e+00 2.28199229e-01 -2.99479753e-01
6.20876215e-02 -6.12820327e-01 3.96398067e-01 -2.59222656e-01
2.52437979e-01 1.06309429e-01 1.67572573e-01 -4.67124134e-01
-8.07246029e-01 -2.36779287e-01 -1.23483431e+00 -3.22249979e-01
-5.40686548e-01 8.18759352e-02 1.00898826e+00 1.32671201e+00
2.31107026e-01 6.97509587e-01 9.44018364e-01 -4.76596981e-01
-1.79576188e-01 -8.48940969e-01 -4.75997150e-01 -3.47350806e-01
2.61363447e-01 -7.56733894e-01 -4.87655371e-01 -3.60018939e-01] | [6.190345764160156, 1.4026297330856323] |
6515a5c9-0a1b-4779-b214-25538bb05936 | physq-a-physics-informed-reinforcement | 2211.11830 | null | https://arxiv.org/abs/2211.11830v1 | https://arxiv.org/pdf/2211.11830v1.pdf | PhysQ: A Physics Informed Reinforcement Learning Framework for Building Control | Large-scale integration of intermittent renewable energy sources calls for substantial demand side flexibility. Given that the built environment accounts for approximately 40% of total energy consumption in EU, unlocking its flexibility is a key step in the energy transition process. This paper focuses specifically on energy flexibility in residential buildings, leveraging their intrinsic thermal mass. Building on recent developments in the field of data-driven control, we propose PhysQ. As a physics-informed reinforcement learning framework for building control, PhysQ forms a step in bridging the gap between conventional model-based control and data-intensive control based on reinforcement learning. Through our experiments, we show that the proposed PhysQ framework can learn high quality control policies that outperform a business-as-usual, as well as a rudimentary model predictive controller. Our experiments indicate cost savings of about 9% compared to a business-as-usual controller. Further, we show that PhysQ efficiently leverages prior physics knowledge to learn such policies using fewer training samples than conventional reinforcement learning approaches, making PhysQ a scalable alternative for use in residential buildings. Additionally, the PhysQ control policy utilizes building state representations that are intuitive and based on conventional building models, that leads to better interpretation of the learnt policy over other data-driven controllers. | ['Chris Develder', 'Bert Claessens', 'Gargya Gokhale'] | 2022-11-21 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [-3.53927374e-01 3.64556104e-01 -6.83996379e-01 1.69879440e-02
-6.66907310e-01 -5.90215802e-01 5.33649981e-01 3.13725889e-01
1.63020134e-01 1.03703976e+00 1.98442653e-01 -5.43321073e-01
-4.62093830e-01 -1.37361598e+00 -9.52223420e-01 -8.21115077e-01
-8.41564238e-02 5.31350315e-01 -1.51562005e-01 -4.99421239e-01
-1.84533715e-01 4.84915316e-01 -1.34654641e+00 -5.63037515e-01
1.01512194e+00 9.08422887e-01 2.51423895e-01 6.45659268e-01
2.82072961e-01 5.14637232e-01 -7.48332441e-02 5.81612170e-01
2.10107714e-01 -2.62336999e-01 -5.02080798e-01 -1.64976001e-01
-1.72706723e-01 -4.35220301e-01 -3.42339337e-01 4.66761947e-01
5.51113248e-01 5.17681539e-01 5.61209857e-01 -1.27802110e+00
-5.35780013e-01 4.66078490e-01 1.07552283e-01 2.02114936e-02
1.71629235e-01 5.90088248e-01 1.06957006e+00 -5.44379614e-02
4.41719666e-02 9.27042246e-01 5.11477470e-01 5.87529302e-01
-1.53252721e+00 -6.26317441e-01 3.89399409e-01 -4.10706252e-02
-1.12909818e+00 -2.51621306e-01 8.52330387e-01 -3.15910399e-01
1.33550096e+00 3.35618794e-01 1.46495330e+00 9.26279068e-01
2.83906460e-01 5.49177647e-01 1.35825682e+00 -4.51784909e-01
1.00003016e+00 -2.62952864e-01 -2.46959716e-01 7.08283365e-01
1.88330382e-01 6.94135845e-01 -8.91588926e-02 -1.68176830e-01
8.22534740e-01 -6.20771870e-02 -1.78846881e-01 -6.71683431e-01
-7.75989652e-01 8.13388407e-01 8.50589156e-01 1.18452355e-01
-3.82793427e-01 9.08134580e-01 -3.52088735e-03 -1.38992369e-01
2.92291269e-02 5.69405854e-01 -6.98627353e-01 -1.82898700e-01
-8.84362936e-01 3.49658072e-01 7.01849043e-01 8.95824671e-01
8.99927795e-01 4.73548740e-01 -1.24743193e-01 4.35851485e-01
6.50722146e-01 8.79045844e-01 3.44529688e-01 -1.20043254e+00
7.95749575e-02 7.02360153e-01 4.35172558e-01 -1.10756725e-01
-3.95588517e-01 -1.64606541e-01 -8.67141128e-01 2.99794137e-01
-1.44318298e-01 -1.99284941e-01 -1.05177772e+00 1.96204758e+00
4.18729573e-01 -6.38265312e-02 -1.21772125e-01 3.68074358e-01
-2.28708029e-01 9.58812356e-01 3.20716083e-01 -2.05259725e-01
1.10684526e+00 -6.95788622e-01 -5.20904779e-01 -1.55403212e-01
1.30566135e-01 -1.41579937e-03 1.13627660e+00 3.35392147e-01
-9.69841123e-01 -4.14847583e-01 -1.31086385e+00 3.17096978e-01
-5.00685453e-01 -2.49840572e-01 6.31825924e-01 9.93503213e-01
-1.23617530e+00 8.11317563e-01 -1.28803182e+00 -6.25936806e-01
2.25150079e-01 4.54205394e-01 4.80054080e-01 3.29271346e-01
-1.14711988e+00 9.72266912e-01 6.73509479e-01 -2.02631921e-01
-1.02614164e+00 -9.39805746e-01 -6.59765720e-01 2.43729636e-01
6.66228175e-01 -1.08185816e+00 1.65858734e+00 -8.50675777e-02
-2.17428589e+00 -9.32469144e-02 3.05955172e-01 -6.88334405e-01
3.73784721e-01 -1.81496158e-01 -4.24576998e-01 -9.53313783e-02
-2.36194476e-01 4.99619991e-01 7.37945676e-01 -1.12448323e+00
-5.06264925e-01 5.33050895e-02 1.87681660e-01 -7.28840306e-02
-1.35922670e-01 -9.10127044e-01 2.34198570e-01 -3.66460860e-01
-2.68144131e-01 -1.28001821e+00 -5.42104661e-01 -1.59376681e-01
-2.07905322e-01 -5.10046601e-01 9.52571929e-01 -3.22403133e-01
1.23679948e+00 -1.55794716e+00 3.40280272e-02 4.40357685e-01
-2.77489752e-01 -1.97307840e-02 4.52122003e-01 8.68685126e-01
-3.32021713e-02 3.07051808e-01 -3.31403553e-01 2.27309659e-01
5.63307106e-01 7.32597828e-01 -3.17431748e-01 -6.11674897e-02
6.88540637e-02 1.01617384e+00 -1.04828334e+00 -8.52700248e-02
9.02611732e-01 2.05128402e-01 -5.04549623e-01 1.46542221e-01
-9.10261750e-01 5.09287477e-01 -8.21989417e-01 4.86717373e-01
3.20731044e-01 -2.16266096e-01 4.57109571e-01 -2.16183513e-01
-2.59887636e-01 2.54117996e-01 -1.10378909e+00 1.55051386e+00
-1.07648253e+00 -2.94734128e-02 1.26639202e-01 -1.08800340e+00
7.65957713e-01 2.11135745e-01 7.83236623e-01 -7.58006215e-01
-2.11501077e-01 1.71312436e-01 -4.21152860e-01 -7.16274679e-02
4.58899647e-01 -5.31112313e-01 -1.72109589e-01 2.46500164e-01
-6.11334927e-02 -9.82566953e-01 9.58170593e-02 -2.38184724e-02
1.19525027e+00 5.19262135e-01 4.45907146e-01 -6.66143954e-01
3.72329146e-01 -2.16217592e-01 5.94602406e-01 3.87265801e-01
-1.37686372e-01 -3.00312728e-01 -6.93242103e-02 -7.51562789e-02
-1.00837398e+00 -1.34006310e+00 -9.12206434e-03 8.09973478e-01
-1.31212756e-01 -5.15421569e-01 -7.68039346e-01 -4.82351243e-01
2.73279816e-01 1.25739789e+00 -2.84291387e-01 -3.16467583e-01
-6.97085857e-01 -6.09175444e-01 -6.01700582e-02 6.82507873e-01
6.53634667e-01 -5.70325196e-01 -9.53410923e-01 5.16606688e-01
1.23613484e-01 -7.46425986e-01 -2.50622571e-01 6.66666746e-01
-9.94453967e-01 -1.09170914e+00 -2.27223933e-01 -6.06557913e-02
1.54886186e-01 -9.76218581e-02 1.31568551e+00 1.39664821e-02
-1.17004171e-01 1.16182637e+00 4.06350056e-03 -2.68687725e-01
-7.02823043e-01 1.61112353e-01 3.59479725e-01 -6.03638649e-01
-2.64970750e-01 -9.60344374e-01 -9.80852723e-01 1.91220880e-01
-6.84965611e-01 4.25869301e-02 4.41959709e-01 5.66025734e-01
7.55118012e-01 6.10433936e-01 6.31315947e-01 -1.90953210e-01
4.68803406e-01 -3.13491523e-01 -1.04530025e+00 3.49176019e-01
-1.33842182e+00 5.99959433e-01 8.27168524e-01 -8.25769976e-02
-1.01701283e+00 2.40250051e-01 -2.29333699e-01 -2.27435566e-02
4.40838449e-02 -4.38634492e-02 -3.07193041e-01 3.05663496e-02
2.76305169e-01 2.06092626e-01 -2.67332792e-01 -3.79990458e-01
6.99840009e-01 1.56201765e-01 2.20151782e-01 -1.28168523e+00
1.17273426e+00 8.37186724e-02 3.31291914e-01 -7.89554119e-01
-4.32819605e-01 -1.88591145e-02 -2.83089221e-01 -2.16789484e-01
8.97557616e-01 -8.81447673e-01 -1.11116600e+00 3.37178446e-02
-3.25189054e-01 -1.04533935e+00 -9.17312205e-01 2.51691461e-01
-1.21532393e+00 -1.21190369e-01 -2.21779823e-01 -1.09628749e+00
-3.90299648e-01 -9.97247517e-01 7.77676463e-01 4.81557637e-01
-7.29441121e-02 -1.02942109e+00 5.13019562e-01 3.71641889e-02
6.60880029e-01 6.22217834e-01 1.25987196e+00 4.04509902e-02
-1.02896643e+00 2.44261652e-01 4.15034175e-01 3.67050976e-01
3.82953286e-01 1.49122655e-01 -8.13629031e-01 -6.48559391e-01
-4.94215377e-02 -4.22159553e-01 4.96186554e-01 4.92297500e-01
1.19025826e+00 -4.40188348e-01 -4.39530522e-01 1.14943065e-01
1.84038138e+00 2.55011052e-01 3.17974597e-01 3.40102136e-01
3.28020811e-01 9.75808650e-02 3.98814976e-01 5.61963201e-01
6.60711527e-01 5.91400743e-01 7.07957149e-01 8.09013546e-02
8.64067823e-02 -7.21425295e-01 3.56463075e-01 6.94045424e-01
-1.08437099e-01 1.39402360e-01 -8.59888077e-01 2.61740804e-01
-1.74073958e+00 -9.26470041e-01 3.60421300e-01 2.33896065e+00
7.96688139e-01 2.16090217e-01 3.00044686e-01 1.27270877e-01
2.20744729e-01 2.70028770e-01 -1.01996160e+00 -4.62278306e-01
5.06577134e-01 5.58427334e-01 8.29292953e-01 3.52432162e-01
-7.65040278e-01 4.49083418e-01 6.94621944e+00 6.47726774e-01
-9.51289535e-01 5.31196371e-02 4.78896022e-01 7.46586500e-03
-4.47527170e-01 9.24581066e-02 -5.71911335e-01 5.42094231e-01
1.61120653e+00 -4.59687144e-01 8.44030261e-01 1.10380626e+00
8.15378428e-01 -4.09616441e-01 -1.38127363e+00 5.46345651e-01
-9.41632211e-01 -1.36600041e+00 -3.26412797e-01 5.24735749e-01
7.91278660e-01 5.09883743e-04 -1.04903750e-01 8.97798419e-01
9.87595558e-01 -7.63918638e-01 8.08312535e-01 6.10236406e-01
3.63861680e-01 -9.63405728e-01 3.95889468e-02 5.70578218e-01
-1.64054585e+00 -4.80556041e-01 -8.22210461e-02 -4.75265048e-02
1.83708981e-01 6.23883069e-01 -6.25191569e-01 8.22617352e-01
7.31487989e-01 4.53415811e-01 -3.09338629e-01 5.80979526e-01
-3.75537723e-01 9.02455628e-01 -7.20127046e-01 -1.73162907e-01
1.38353184e-01 -3.44908208e-01 2.98606120e-02 7.05295563e-01
4.57303345e-01 1.70427814e-01 6.56979918e-01 1.02218843e+00
1.07648931e-02 -4.69160736e-01 -5.16676188e-01 -7.69515708e-02
4.77161020e-01 1.16871333e+00 -5.57338357e-01 -2.92822421e-01
-1.00018941e-01 5.25730312e-01 -1.23117305e-02 3.10026288e-01
-1.07773125e+00 1.96761429e-01 7.63471246e-01 1.56156957e-01
4.63116765e-01 -5.96133351e-01 1.05016991e-01 -6.50401771e-01
-3.96689713e-01 -6.09021068e-01 -3.31333727e-02 -7.16368556e-01
-1.04979944e+00 -3.21890801e-01 4.24975157e-01 -1.01653194e+00
-5.74656606e-01 -4.77244109e-01 -6.83408022e-01 6.30076766e-01
-1.68778539e+00 -1.06555700e+00 -1.77266896e-01 4.48679864e-01
5.67284465e-01 2.96740174e-01 1.07026625e+00 -1.29135877e-01
-6.24228835e-01 2.56001260e-02 5.79188645e-01 -5.08203030e-01
1.08477123e-01 -1.75175369e+00 4.13143843e-01 5.07499635e-01
-2.97570288e-01 3.60716403e-01 8.15481186e-01 -5.83482504e-01
-2.02140355e+00 -1.06739199e+00 -2.64038652e-01 -1.91301048e-01
8.29372227e-01 -2.74346650e-01 -5.56521773e-01 4.10580039e-01
5.15557766e-01 -1.17718130e-01 4.83668506e-01 -2.48291772e-02
1.67385697e-01 -4.18036878e-01 -1.29747057e+00 4.26870525e-01
8.41962397e-01 -5.09835303e-01 -3.95419776e-01 3.20813715e-01
7.65530467e-01 -2.03433752e-01 -1.25754499e+00 3.57497811e-01
3.89599919e-01 -7.02837706e-01 1.00065351e+00 -2.04110801e-01
1.67700462e-04 -3.43627930e-01 -4.84690756e-01 -1.69193101e+00
-4.83996391e-01 -8.51152062e-01 -9.50190544e-01 1.32771182e+00
-9.68934745e-02 -5.32661617e-01 7.76584864e-01 8.45360279e-01
-1.78908050e-01 -8.36367607e-01 -1.09509099e+00 -1.24619257e+00
5.44625998e-01 -3.16346169e-01 8.01565170e-01 4.77433830e-01
-3.29442710e-01 1.30146325e-01 9.27451104e-02 3.71691585e-01
6.90998018e-01 4.49176729e-02 4.87573147e-01 -9.55133617e-01
-5.93786240e-01 -3.37541282e-01 1.06342673e-01 -8.79046202e-01
7.33137503e-02 -7.03584135e-01 1.63799852e-01 -1.97192371e+00
-8.04060847e-02 -6.20555699e-01 -5.31924129e-01 5.54892182e-01
1.72719568e-01 -4.67512965e-01 1.98598236e-01 -1.11483239e-01
-2.52538510e-02 1.16340935e+00 1.02378631e+00 -4.22939718e-01
-4.05322731e-01 2.84488678e-01 -5.58096707e-01 5.18433928e-01
1.18808699e+00 -1.21655405e-01 -8.45601737e-01 2.25686327e-01
9.49512720e-02 5.79909571e-02 4.24148053e-01 -1.34996653e+00
-1.41523257e-01 -8.11433911e-01 1.75677896e-01 -6.56923831e-01
3.06941390e-01 -1.20031059e+00 4.43314135e-01 1.01678061e+00
9.90116447e-02 -6.14906363e-02 3.30070883e-01 9.99830902e-01
4.42696869e-01 9.64792222e-02 9.10942078e-01 -3.56708407e-01
-5.59561133e-01 1.60712421e-01 -5.66731870e-01 -8.09771046e-02
1.09650552e+00 -5.02114445e-02 -1.56764388e-01 -2.40640998e-01
-7.08697915e-01 5.77642798e-01 5.77015638e-01 2.10390747e-01
1.58090204e-01 -1.53101814e+00 -5.72142079e-02 -1.35354593e-01
-2.01475888e-01 1.56306522e-03 1.17979839e-01 2.33444124e-01
-1.90593868e-01 5.28696775e-01 -9.77664292e-02 -6.13076091e-01
-5.72309494e-01 7.20950365e-01 7.46808469e-01 -6.68823600e-01
-6.81657434e-01 1.36215031e-01 -2.05064043e-01 -5.61593831e-01
-1.97718829e-01 -1.08459473e+00 6.30138338e-01 -2.02746704e-01
-7.97148645e-02 6.70463204e-01 1.51729941e-01 1.19292945e-01
-3.20679367e-01 5.44640958e-01 5.36299348e-01 -1.57112032e-01
1.46169281e+00 6.77899271e-03 2.92835891e-01 6.51142120e-01
6.16099060e-01 -1.66654050e-01 -1.65513968e+00 2.27264240e-01
7.30818436e-02 8.99998918e-02 3.95900309e-01 -1.07757974e+00
-8.81211400e-01 5.26320994e-01 1.10642779e+00 5.98914504e-01
1.28127646e+00 -2.41427824e-01 6.95118666e-01 7.22311437e-01
8.89061391e-01 -1.64503443e+00 1.93319589e-01 1.77413985e-01
8.15531433e-01 -9.79420304e-01 3.50176930e-01 1.02002949e-01
-3.11908182e-02 8.91193211e-01 5.72531283e-01 -2.52345860e-01
8.56300235e-01 1.33005351e-01 -6.74806774e-01 2.36683100e-01
-6.71010613e-01 -1.81282744e-01 -2.20775872e-01 6.18077755e-01
1.27524003e-01 4.45906699e-01 3.48495767e-02 1.28934652e-01
-2.00538248e-01 2.49219179e-01 2.81532586e-01 1.40163195e+00
-8.43872488e-01 -1.48203444e+00 -4.56664860e-01 2.69726872e-01
2.21631587e-01 2.80952483e-01 3.29265743e-01 1.10252571e+00
6.14273772e-02 9.21716154e-01 -3.28163028e-01 -4.03219610e-02
6.95869446e-01 3.12878519e-01 6.67945385e-01 -4.15566683e-01
-4.53311741e-01 -8.22386891e-02 -1.01179533e-01 -6.92470670e-01
-3.52270693e-01 -4.72116619e-01 -1.57565761e+00 -4.51519787e-01
-1.71432361e-01 6.19279034e-02 8.23373854e-01 8.44882846e-01
2.65307307e-01 1.08527708e+00 9.60587502e-01 -1.05981636e+00
-9.23212290e-01 -6.04964256e-01 -5.16542912e-01 -1.30529985e-01
4.13861692e-01 -9.44924116e-01 -6.44931495e-02 -2.55654812e-01] | [5.289761543273926, 2.3849120140075684] |
a2e66619-597a-44d2-8f21-663a38fb30fb | one-shot-object-detection-without-fine-tuning | 2005.03819 | null | https://arxiv.org/abs/2005.03819v1 | https://arxiv.org/pdf/2005.03819v1.pdf | One-Shot Object Detection without Fine-Tuning | Deep learning has revolutionized object detection thanks to large-scale datasets, but their object categories are still arguably very limited. In this paper, we attempt to enrich such categories by addressing the one-shot object detection problem, where the number of annotated training examples for learning an unseen class is limited to one. We introduce a two-stage model consisting of a first stage Matching-FCOS network and a second stage Structure-Aware Relation Module, the combination of which integrates metric learning with an anchor-free Faster R-CNN-style detection pipeline, eventually eliminating the need to fine-tune on the support images. We also propose novel training strategies that effectively improve detection performance. Extensive quantitative and qualitative evaluations were performed and our method exceeds the state-of-the-art one-shot performance consistently on multiple datasets. | ['Chi-Keung Tang', 'Yu-Wing Tai', 'Xiang Li', 'Yau Pun Chen', 'Lin Zhang'] | 2020-05-08 | null | null | null | null | ['one-shot-object-detection'] | ['computer-vision'] | [ 2.27893084e-01 2.35235199e-01 -3.05737793e-01 -3.34423095e-01
-1.03900278e+00 -3.78816307e-01 5.47882378e-01 2.45363086e-01
-6.89781070e-01 3.39148223e-01 -1.20575204e-01 2.31734477e-02
-3.20769921e-02 -7.72214353e-01 -5.95095456e-01 -3.21894288e-01
9.33862105e-02 5.62838316e-01 9.58999813e-01 -7.49829561e-02
1.45217180e-01 4.91647989e-01 -1.86054170e+00 4.62222189e-01
4.22697783e-01 1.28362429e+00 1.25065848e-01 6.76225066e-01
7.09218606e-02 7.64717519e-01 -5.05586863e-01 -6.94724798e-01
4.34134096e-01 -2.15230472e-02 -7.03566432e-01 5.88474497e-02
6.33800447e-01 -6.38656914e-01 -4.42902684e-01 1.01088488e+00
5.79940081e-01 4.07439880e-02 4.69093829e-01 -1.00756037e+00
-8.13518643e-01 5.46383500e-01 -8.24880898e-01 5.21378577e-01
5.21018393e-02 3.72721195e-01 1.25699568e+00 -1.19278932e+00
6.53458595e-01 1.12324250e+00 8.84237289e-01 5.95438898e-01
-1.18001187e+00 -6.20907307e-01 2.55588293e-02 4.90754247e-01
-1.66558444e+00 -4.35741544e-01 5.56855142e-01 -4.72863764e-01
1.03869689e+00 -1.82876632e-01 5.64548850e-01 8.62357974e-01
-3.79377395e-01 8.94996107e-01 6.22058272e-01 -4.72685426e-01
2.51363456e-01 2.33517438e-01 2.25933656e-01 7.81730533e-01
5.01346886e-01 1.86079159e-01 -2.76894867e-01 4.87071052e-02
4.20905828e-01 1.74578667e-01 2.97080845e-01 -6.73637867e-01
-1.01750493e+00 8.77774298e-01 9.19278324e-01 3.60565871e-01
-1.52154937e-01 8.25148821e-02 4.02616560e-01 1.39677543e-02
4.00557548e-01 6.04759514e-01 -3.25285196e-01 3.16663891e-01
-9.95445848e-01 7.01600388e-02 6.79025352e-01 1.03843427e+00
8.16852272e-01 -3.41832757e-01 -6.77409768e-01 8.74778986e-01
6.64510857e-03 1.44880444e-01 2.77557999e-01 -7.70671010e-01
1.79777130e-01 1.07337356e+00 2.79932648e-01 -7.01263249e-01
-6.29940450e-01 -7.67598450e-01 -3.87411565e-01 1.86808825e-01
5.09707153e-01 1.09797716e-01 -9.60251689e-01 1.37799311e+00
6.67540729e-01 3.22412252e-01 -2.55016983e-01 1.10523188e+00
9.47714448e-01 9.55072492e-02 9.40558314e-02 2.17307955e-01
1.48312938e+00 -1.20616245e+00 -2.61746019e-01 -3.11594784e-01
8.17404926e-01 -5.48809767e-01 9.81225193e-01 1.01194046e-01
-8.92632127e-01 -6.70867503e-01 -1.29297757e+00 -3.33546937e-01
-5.38942993e-01 4.40324068e-01 7.10919917e-01 5.27265131e-01
-7.81130254e-01 5.24906337e-01 -6.14000797e-01 -5.18404365e-01
1.18898618e+00 2.56020725e-01 -3.42169136e-01 -1.57924771e-01
-8.87030780e-01 1.10912573e+00 6.53083205e-01 -1.48369551e-01
-1.19200528e+00 -7.64662385e-01 -6.58209860e-01 1.71273444e-02
9.15691018e-01 -6.14319980e-01 1.27780128e+00 -4.60010648e-01
-1.00960159e+00 1.42118406e+00 1.73478127e-01 -6.38412476e-01
7.58472741e-01 -4.07424599e-01 -2.09026590e-01 2.90372938e-01
3.42249960e-01 9.47198689e-01 7.82642543e-01 -1.08570158e+00
-1.02452910e+00 -4.11400914e-01 2.89302021e-01 -9.11670476e-02
-5.28145075e-01 3.08870256e-01 -5.85821927e-01 -3.01961541e-01
-1.65946305e-01 -5.76062739e-01 -3.39358896e-01 6.64898574e-01
-2.62533814e-01 -4.85475391e-01 8.00272048e-01 -1.81232244e-01
1.01796412e+00 -1.96432257e+00 -1.39179125e-01 -2.16464505e-01
5.57981253e-01 5.54697394e-01 -3.16547155e-01 -5.92795350e-02
2.20812708e-01 -5.04206009e-02 -7.21092969e-02 -5.54107726e-01
-8.56375843e-02 5.83773069e-02 -2.54505813e-01 6.20691657e-01
7.98629940e-01 1.19299316e+00 -1.15935147e+00 -6.37185872e-01
2.35237300e-01 3.17033410e-01 -3.98592919e-01 2.82614827e-01
-1.41803414e-01 4.56103832e-02 -3.01845521e-01 1.02498877e+00
5.23491502e-01 -5.20815074e-01 -1.32997856e-01 -2.68887848e-01
-3.90043966e-02 2.41700374e-02 -1.34180200e+00 1.81379414e+00
-1.99524015e-01 4.40223008e-01 -1.81372866e-01 -1.00481510e+00
8.78884614e-01 -6.80138096e-02 2.95172572e-01 -6.26961291e-01
4.08059418e-01 2.19377384e-01 3.24024037e-02 -4.22017604e-01
3.93361270e-01 5.43544143e-02 1.12169281e-01 2.09430307e-01
3.67413849e-01 2.38827884e-01 4.22043592e-01 2.54986972e-01
1.28731251e+00 1.90556869e-01 5.14314592e-01 -5.04165292e-02
2.85182714e-01 1.62630856e-01 6.09999299e-01 1.05886781e+00
-5.42489469e-01 7.30559289e-01 2.80845463e-01 -5.60783625e-01
-1.20108771e+00 -8.10375631e-01 -3.35678428e-01 1.40733516e+00
5.22159517e-01 -2.56466478e-01 -5.97242713e-01 -9.40148115e-01
2.25028083e-01 4.72624302e-01 -1.03220558e+00 -3.02587658e-01
-5.18277526e-01 -7.40607917e-01 5.92512429e-01 8.47767472e-01
3.66196662e-01 -8.09523165e-01 -9.58468616e-01 1.64901197e-01
1.51592597e-01 -1.34875870e+00 -2.05733448e-01 2.83033997e-01
-6.56772375e-01 -1.45369315e+00 -4.66662318e-01 -6.07233703e-01
5.06767213e-01 5.84390044e-01 1.17509604e+00 3.35302413e-01
-9.87355471e-01 4.85668294e-02 -4.59355175e-01 -5.90894759e-01
-7.93984905e-02 1.33043259e-01 -6.47530854e-02 -8.30335468e-02
8.69376004e-01 -1.74700856e-01 -9.08501983e-01 5.84576488e-01
-6.20473683e-01 -1.28196925e-01 8.07385802e-01 8.02842736e-01
5.98580003e-01 -5.91600478e-01 7.49597788e-01 -6.55872285e-01
-4.14399989e-02 -4.17502582e-01 -6.33920908e-01 3.53777438e-01
-5.19845307e-01 -2.79799551e-01 1.99978799e-01 -6.73065722e-01
-7.48332083e-01 3.04060161e-01 1.51434168e-01 -6.43682659e-01
-8.21117982e-02 -1.12966344e-01 -4.92673889e-02 -2.63307363e-01
1.15339410e+00 -2.27695376e-01 -2.68854201e-01 -5.91865718e-01
6.71653867e-01 7.12446094e-01 7.52365649e-01 -1.78662553e-01
9.95575607e-01 7.64280319e-01 -1.24023736e-01 -4.95955080e-01
-1.58047879e+00 -9.79163826e-01 -1.06685770e+00 -2.12651148e-01
7.97714710e-01 -1.03897655e+00 -4.41214085e-01 2.39694610e-01
-1.12344754e+00 -1.34242266e-01 -7.79993951e-01 1.59806892e-01
-3.73805732e-01 1.43022612e-02 -5.69544017e-01 -7.55552053e-01
-4.62993056e-01 -7.35890269e-01 1.33607137e+00 2.76825309e-01
2.68033836e-02 -3.62036139e-01 1.20233539e-02 4.20246512e-01
3.19717109e-01 1.16455354e-01 1.84655815e-01 -1.05486190e+00
-7.71380067e-01 -6.78405762e-01 -8.40051174e-01 1.64539158e-01
-1.97011545e-01 -1.51672333e-01 -1.25832963e+00 -2.74496377e-01
-2.71230429e-01 -6.27050996e-01 1.26021814e+00 1.00719057e-01
1.16460848e+00 1.55993085e-02 -6.72731042e-01 6.07268631e-01
1.30101573e+00 -3.53914589e-01 4.35415238e-01 5.27077317e-01
6.85618639e-01 4.65840578e-01 9.26598370e-01 3.89327228e-01
2.14847192e-01 7.73857415e-01 6.71560884e-01 -7.25073293e-02
-6.67856097e-01 -2.47054175e-01 -2.02741593e-01 -5.72284460e-02
7.25134388e-02 1.18837776e-02 -8.53236318e-01 8.41303945e-01
-1.95279431e+00 -9.21025336e-01 -6.46035839e-03 2.07411766e+00
6.97861195e-01 4.94212121e-01 3.96419644e-01 1.07154869e-01
8.46206963e-01 -1.47343174e-01 -7.95535088e-01 3.82268697e-01
-1.17434487e-01 -5.24192862e-03 3.75964850e-01 4.37962227e-02
-1.45903170e+00 1.13692760e+00 6.30087900e+00 7.45928466e-01
-7.38316417e-01 4.68231946e-01 4.23156410e-01 -3.16053003e-01
3.34708065e-01 -4.69472893e-02 -1.12751877e+00 2.26820484e-02
6.45798743e-01 6.23961203e-02 -1.64095331e-02 1.30471361e+00
-2.20187470e-01 -4.47403863e-02 -1.36058617e+00 9.16432321e-01
7.54355490e-02 -1.59031427e+00 -1.54341638e-01 -1.81127414e-01
6.57212615e-01 3.69581431e-01 -2.38617271e-01 4.99639511e-01
1.50408536e-01 -6.73499525e-01 7.85341561e-01 3.94112140e-01
9.55487072e-01 -4.71832186e-01 6.80641115e-01 3.23539525e-01
-1.37646043e+00 -5.35231650e-01 -7.40616977e-01 4.74196114e-02
-1.77444834e-02 5.76091647e-01 -7.68125713e-01 2.43698329e-01
8.93299699e-01 7.12228179e-01 -9.38101649e-01 1.58834505e+00
-2.69800514e-01 3.12625647e-01 -3.74691367e-01 -9.34758708e-02
5.22091091e-02 4.33297545e-01 6.06992424e-01 1.35926652e+00
6.52116314e-02 1.33066803e-01 3.23289454e-01 1.04236770e+00
-3.32529604e-01 -1.42241120e-01 -4.79463696e-01 2.73183048e-01
5.95852554e-01 1.78291368e+00 -8.96604121e-01 -5.05401492e-01
-5.69989741e-01 6.54667199e-01 7.62534857e-01 -1.28444374e-01
-7.89646745e-01 -5.02977312e-01 4.86858666e-01 3.06896687e-01
6.18524075e-01 1.18469536e-01 -2.67002463e-01 -1.11444652e+00
-1.02697857e-01 -3.44956845e-01 6.48223042e-01 -5.74932516e-01
-1.47322702e+00 3.25711995e-01 -1.99887231e-01 -1.22467470e+00
3.95774469e-02 -7.04660833e-01 -5.95111668e-01 3.23514879e-01
-1.75059962e+00 -1.50833333e+00 -4.85521913e-01 3.08487415e-01
4.36338305e-01 -1.41535625e-01 5.38128912e-01 5.73605835e-01
-7.97331452e-01 8.35037351e-01 -2.14358196e-01 3.83610249e-01
6.86673582e-01 -1.20088923e+00 3.57895225e-01 9.92124438e-01
1.29266798e-01 3.79081994e-01 5.06949246e-01 -5.26737154e-01
-1.13872778e+00 -1.40187502e+00 4.20779645e-01 -8.08871627e-01
8.99085581e-01 -5.03471076e-01 -1.04794729e+00 5.13913453e-01
-3.86393934e-01 7.98330963e-01 5.12081027e-01 1.77117765e-01
-8.00034940e-01 -1.90430462e-01 -1.16400516e+00 2.06723973e-01
1.51568377e+00 -4.46993560e-01 -7.52457082e-01 4.01925206e-01
8.77268076e-01 -2.47358143e-01 -6.92516983e-01 5.75398862e-01
5.37366271e-01 -8.41629326e-01 1.07214534e+00 -8.28862250e-01
2.80330807e-01 -5.07892966e-01 -1.68230087e-01 -6.62189841e-01
-6.21996224e-01 -3.29658926e-01 -7.00943947e-01 1.15649140e+00
2.79317081e-01 -2.36624449e-01 8.25127959e-01 4.90818560e-01
-2.45220602e-01 -7.34304607e-01 -8.49057555e-01 -8.39819729e-01
-2.75113761e-01 -3.48491311e-01 3.88741642e-01 6.85344696e-01
-1.56039402e-01 5.14025867e-01 -1.95582837e-01 1.88430965e-01
8.45654964e-01 2.85227805e-01 9.13070798e-01 -1.48042381e+00
-3.88979644e-01 -4.88421917e-01 -7.40661144e-01 -8.53982627e-01
-2.70121992e-01 -9.39014614e-01 2.72356749e-01 -1.48300028e+00
7.68931746e-01 -3.13509077e-01 -6.14377022e-01 6.47147834e-01
-3.95625710e-01 7.06718147e-01 1.11793876e-01 3.36340696e-01
-1.29425561e+00 4.70342934e-01 9.17165279e-01 -1.38107166e-01
-6.92970306e-02 -8.67264345e-02 -6.90522134e-01 7.10164547e-01
3.60449404e-01 -6.58488512e-01 -1.23447319e-02 -2.35784858e-01
1.03567101e-01 -6.01726949e-01 6.78515434e-01 -1.14186561e+00
3.22050482e-01 1.64679468e-01 5.79374194e-01 -7.02097595e-01
2.19907075e-01 -5.81258893e-01 -5.13440549e-01 5.52085936e-01
-4.26782221e-01 -5.43861091e-01 1.36230156e-01 8.36362720e-01
1.78171545e-01 -2.62266725e-01 1.24620247e+00 1.07152294e-02
-9.87751007e-01 4.70185935e-01 2.70162493e-01 1.28903940e-01
1.41294181e+00 -3.16143095e-01 -5.84755480e-01 2.37167776e-01
-6.17100835e-01 4.06675309e-01 2.96095222e-01 6.59311235e-01
4.71560448e-01 -1.29382634e+00 -7.76096940e-01 -8.96928832e-02
6.51466906e-01 9.04540569e-02 6.40245378e-02 7.00446904e-01
-4.75239046e-02 2.11084217e-01 -1.02700092e-01 -7.96407998e-01
-1.12177312e+00 1.06091070e+00 4.02591765e-01 -1.74212188e-01
-7.20879436e-01 1.15562749e+00 1.75453678e-01 -2.51616240e-01
4.53387082e-01 -4.00325544e-02 -2.89424002e-01 3.35016727e-01
9.97007608e-01 7.36630261e-01 7.19268993e-02 -4.49427009e-01
-4.70782757e-01 4.55421954e-01 -4.03046012e-01 3.32728148e-01
1.38567841e+00 -5.59132360e-02 1.43767506e-01 2.98170805e-01
1.02853787e+00 -6.22969270e-01 -1.57465708e+00 -6.86888337e-01
3.31046373e-01 -5.77187717e-01 1.13808855e-01 -6.52481198e-01
-8.81271541e-01 8.27933371e-01 9.93299842e-01 3.57194367e-04
7.69743562e-01 5.10549307e-01 6.74391150e-01 5.49340487e-01
3.05241704e-01 -1.26401973e+00 6.32193208e-01 2.62449265e-01
5.98058164e-01 -1.74004817e+00 7.60340989e-02 -3.05492610e-01
-2.05854848e-01 9.47198093e-01 9.91961658e-01 -2.50721216e-01
4.91000354e-01 1.58699960e-01 -2.80220866e-01 -2.45644540e-01
-6.88461065e-01 -8.04687321e-01 6.19895399e-01 5.58708906e-01
5.36164716e-02 -2.34281972e-01 -1.29752429e-02 6.77432895e-01
2.83172131e-01 2.21847370e-01 1.39049187e-01 8.57501805e-01
-9.06458974e-01 -6.53936148e-01 -2.05141664e-01 6.33638501e-01
-3.61865520e-01 1.60120443e-01 -4.53899026e-01 7.28930295e-01
3.92052889e-01 9.73452091e-01 1.25736207e-01 -4.00039673e-01
5.37370622e-01 -1.18166111e-01 4.35128599e-01 -9.43811774e-01
-4.44216192e-01 -2.70358384e-01 -1.55758262e-02 -8.89748573e-01
-2.41536692e-01 -5.66136479e-01 -1.13574779e+00 1.45074978e-01
-8.78770590e-01 -5.01560032e-01 4.25423235e-01 9.50678408e-01
4.92890000e-01 5.20983517e-01 4.10197228e-01 -1.06340599e+00
-8.50311458e-01 -1.06567323e+00 -2.07546368e-01 4.21792984e-01
3.41056883e-01 -9.55319285e-01 -9.59506407e-02 -2.64343530e-01] | [9.26494026184082, 1.182039737701416] |
1d29c548-f9e8-426e-aa32-47b0b5a7ea65 | neurodavis-a-neural-network-model-for-data | 2304.01222 | null | https://arxiv.org/abs/2304.01222v1 | https://arxiv.org/pdf/2304.01222v1.pdf | NeuroDAVIS: A neural network model for data visualization | The task of dimensionality reduction and visualization of high-dimensional datasets remains a challenging problem since long. Modern high-throughput technologies produce newer high-dimensional datasets having multiple views with relatively new data types. Visualization of these datasets require proper methodology that can uncover hidden patterns in the data without affecting the local and global structures within the data. To this end, however, very few such methodology exist, which can realise this task. In this work, we have introduced a novel unsupervised deep neural network model, called NeuroDAVIS, for data visualization. NeuroDAVIS is capable of extracting important features from the data, without assuming any data distribution, and visualize effectively in lower dimension. It has been shown theoritically that neighbourhood relationship of the data in high dimension remains preserved in lower dimension. The performance of NeuroDAVIS has been evaluated on a wide variety of synthetic and real high-dimensional datasets including numeric, textual, image and biological data. NeuroDAVIS has been highly competitive against both t-Distributed Stochastic Neighbor Embedding (t-SNE) and Uniform Manifold Approximation and Projection (UMAP) with respect to visualization quality, and preservation of data size, shape, and both local and global structure. It has outperformed Fast interpolation-based t-SNE (Fit-SNE), a variant of t-SNE, for most of the high-dimensional datasets as well. For the biological datasets, besides t-SNE, UMAP and Fit-SNE, NeuroDAVIS has also performed well compared to other state-of-the-art algorithms, like Potential of Heat-diffusion for Affinity-based Trajectory Embedding (PHATE) and the siamese neural network-based method, called IVIS. Downstream classification and clustering analyses have also revealed favourable results for NeuroDAVIS-generated embeddings. | ['Rajat K. De', 'Dibyendu B. Seal', 'Chayan Maitra'] | 2023-04-01 | null | null | null | null | ['data-visualization', 'data-visualization'] | ['methodology', 'miscellaneous'] | [-3.65799338e-01 -1.87323630e-01 2.32078776e-01 3.18976976e-02
5.97707704e-02 -5.31436563e-01 8.93023431e-01 4.00119483e-01
-4.80545163e-01 7.02672064e-01 3.81899655e-01 -4.05245394e-01
-6.69066072e-01 -7.91408658e-01 -2.78833598e-01 -1.07486200e+00
-5.00774741e-01 5.50269902e-01 1.45500407e-01 -1.92818210e-01
3.98712754e-01 9.51660752e-01 -1.76476097e+00 1.32534131e-01
4.66385186e-01 5.30786693e-01 -4.22659256e-02 5.86649776e-01
-4.07255739e-01 -6.08752891e-02 -4.45012599e-01 -6.38502976e-03
9.68848467e-02 -3.47850680e-01 -5.39034903e-01 -5.51896274e-01
1.71064973e-01 3.02461892e-01 -3.54072809e-01 8.33272040e-01
8.32739413e-01 1.14378430e-01 1.17547178e+00 -1.21356905e+00
-9.04006481e-01 -6.79693148e-02 -7.58785725e-01 3.58066678e-01
9.10542458e-02 4.24610414e-02 4.27864403e-01 -1.28998196e+00
1.10766947e+00 1.40252018e+00 6.82781696e-01 5.01384437e-01
-1.68961179e+00 -2.60166794e-01 -3.51582497e-01 7.76020512e-02
-1.36441743e+00 2.04280481e-01 9.76614654e-01 -7.56974220e-01
1.01947892e+00 4.58345592e-01 8.02772820e-01 1.00558090e+00
4.62358981e-01 2.47592688e-01 1.27876306e+00 -3.92983794e-01
5.25984049e-01 3.43931943e-01 4.09408033e-01 3.18726957e-01
4.04355824e-01 2.66856581e-01 -4.15680140e-01 -4.07884151e-01
5.22248089e-01 1.03828557e-01 -2.07800895e-01 -8.03466439e-01
-1.38530111e+00 8.37704360e-01 4.03033525e-01 7.40409791e-01
-2.14092225e-01 -2.61952758e-01 7.42927253e-01 1.36437282e-01
6.80288851e-01 5.08232951e-01 -1.70579970e-01 -3.51619929e-01
-7.04930067e-01 2.81989634e-01 5.45781314e-01 2.78317750e-01
4.45462495e-01 2.08389778e-02 1.28216222e-01 5.87053597e-01
1.97337911e-01 2.06851214e-01 9.86771941e-01 -5.26764214e-01
2.31912628e-01 9.54541028e-01 -1.36871353e-01 -1.59916496e+00
-9.77215648e-01 -1.42953277e-01 -1.40947747e+00 7.97889292e-01
5.10165334e-01 1.91941544e-01 -7.87287056e-01 1.49472284e+00
7.09451497e-01 -1.41383573e-01 3.12308282e-01 9.31506336e-01
6.99491084e-01 9.79304492e-01 -2.19488606e-01 -1.00688979e-01
1.04743421e+00 -3.39739919e-01 -9.42852020e-01 6.72079444e-01
9.89319921e-01 -3.69258344e-01 1.29431927e+00 4.06226039e-01
-6.65663421e-01 -6.07074797e-01 -1.11304593e+00 -1.78157419e-01
-1.08047700e+00 -7.46025965e-02 2.88756102e-01 5.91361046e-01
-1.19487429e+00 7.03795910e-01 -9.06887889e-01 -7.16282189e-01
3.81491870e-01 3.39640409e-01 -8.06427777e-01 1.72180697e-01
-9.98895943e-01 9.24936116e-01 5.38560688e-01 3.81682031e-02
-3.30082178e-01 -9.29689825e-01 -6.80579603e-01 -8.39082897e-02
-4.63554621e-01 -3.73536408e-01 1.36207104e-01 -1.88508958e-01
-1.26106977e+00 7.70301104e-01 2.13537011e-02 -4.00133193e-01
4.81280297e-01 5.95204383e-02 -6.16511047e-01 4.84815054e-02
-2.94622093e-01 7.22091734e-01 5.57593524e-01 -1.36054635e+00
-4.32187766e-02 -6.59247398e-01 -5.58515310e-01 8.87052808e-03
-9.21064377e-01 -3.44306558e-01 2.43998945e-01 -5.94767511e-01
1.55846953e-01 -8.38715792e-01 -1.93768330e-02 5.74146211e-01
-3.41035336e-01 -2.39354357e-01 1.78639364e+00 -5.17958939e-01
1.42344487e+00 -2.37959981e+00 5.07711589e-01 2.48077899e-01
6.20292544e-01 8.03728163e-01 1.02781646e-01 9.11500871e-01
-5.57165682e-01 2.11269513e-01 -5.00360429e-01 -1.65926710e-01
-3.68695073e-02 2.79965073e-01 -8.41073990e-02 7.55606055e-01
7.05711022e-02 7.69096792e-01 -7.21623361e-01 -4.17130947e-01
5.66637337e-01 1.05714226e+00 -3.47470284e-01 -2.52948273e-02
1.28388062e-01 2.95744777e-01 9.26110074e-02 8.75827521e-02
7.77223706e-01 7.46501097e-03 -5.71045019e-02 -1.84682071e-01
-3.88979822e-01 -3.70437205e-01 -1.06760979e+00 1.71830666e+00
-3.21938656e-02 1.02537942e+00 -2.22800210e-01 -9.87587929e-01
1.28820515e+00 1.60736933e-01 4.34070826e-01 -7.72220969e-01
-1.78165212e-01 2.09474355e-01 1.80650145e-01 -6.13961935e-01
3.42046678e-01 -7.94582739e-02 4.10039037e-01 5.33851087e-01
-8.07487220e-02 4.63603973e-01 1.32594079e-01 3.72573227e-01
7.13186264e-01 5.75025156e-02 8.91410485e-02 -6.47132516e-01
5.55754006e-01 -2.61926893e-02 2.05279291e-02 -3.22208889e-02
-1.45136863e-01 5.97452223e-01 8.20263028e-01 -7.33597159e-01
-1.53037381e+00 -1.02333593e+00 -4.56031561e-01 3.72828573e-01
2.57621817e-02 -4.20420438e-01 -7.15001941e-01 -4.04194295e-01
2.57336080e-01 7.45559156e-01 -1.04243839e+00 -3.06715757e-01
-3.77838969e-01 -9.54677284e-01 5.32122791e-01 2.06876263e-01
-1.81282181e-02 -7.79805481e-01 -6.90713227e-01 5.25467237e-03
5.25672138e-01 -2.57329792e-01 2.94149276e-02 1.96155980e-01
-1.08542299e+00 -9.37797427e-01 -9.50519562e-01 -4.37823564e-01
7.03196645e-01 -6.19740039e-03 6.40792310e-01 -4.68641639e-01
-7.53022254e-01 -4.05320115e-02 -2.05268860e-01 -2.57197082e-01
-3.44665349e-01 -1.25758901e-01 4.17064846e-01 -6.04521707e-02
4.80809331e-01 -9.45769846e-01 -6.72878146e-01 2.91058719e-01
-1.19400656e+00 -1.44030660e-01 3.76236916e-01 8.52396429e-01
6.24310195e-01 1.73207447e-01 5.46539068e-01 -5.88368773e-01
1.09316528e+00 -5.64741969e-01 -2.59577423e-01 -1.66896656e-01
-9.38828826e-01 2.22531050e-01 8.95775020e-01 -4.68598276e-01
-4.97713357e-01 -3.24528217e-01 1.09847225e-01 -7.78202593e-01
-3.54461163e-01 4.57752258e-01 -1.45291865e-01 1.25813529e-01
9.91118133e-01 3.70693833e-01 7.22476482e-01 -8.13107789e-01
4.83041912e-01 7.67412603e-01 2.61070102e-01 -1.90537006e-01
6.38797104e-01 7.04611003e-01 4.36733812e-01 -1.22687876e+00
4.58483659e-02 -2.53161490e-01 -1.14945149e+00 -9.47032049e-02
1.00175977e+00 -1.19942702e-01 -5.67308664e-01 4.47178662e-01
-9.63372529e-01 1.03052698e-01 -5.65593839e-01 4.40982670e-01
-2.76074231e-01 4.88920152e-01 -1.39445931e-01 -8.39370489e-01
-2.07572505e-01 -9.06600416e-01 5.62465727e-01 -1.00135624e-01
-3.92397523e-01 -1.45974267e+00 6.22556031e-01 -4.11960602e-01
5.62614322e-01 9.03874815e-01 1.49534500e+00 -7.65010953e-01
2.02506915e-01 -2.91403502e-01 -3.10941547e-01 1.68320224e-01
3.11306566e-01 3.54583353e-01 -8.78323257e-01 -5.26941001e-01
-1.75993070e-01 7.46524259e-02 5.37590981e-01 3.38501185e-01
9.71185625e-01 -1.94297522e-01 -4.97515798e-01 6.06788695e-01
1.36024725e+00 2.83283442e-01 5.93451142e-01 5.23580134e-01
7.39041209e-01 9.43427444e-01 3.53814185e-01 3.22987229e-01
3.92981172e-02 5.85762680e-01 6.41658723e-01 -5.21201849e-01
-1.39807723e-02 -1.21538617e-01 -2.30328161e-02 8.90353203e-01
-1.51285407e-04 -1.00167468e-01 -1.03827631e+00 4.87018108e-01
-1.86994433e+00 -8.86121452e-01 -6.78502083e-01 2.19659066e+00
2.72249311e-01 4.20779176e-02 4.72729504e-01 6.71783388e-01
5.99692166e-01 1.92212477e-01 -7.01327562e-01 -8.46455634e-01
-3.62157404e-01 -1.45909950e-01 -1.25490921e-02 1.98971182e-01
-8.46391916e-01 3.61462533e-01 5.43372440e+00 7.65822411e-01
-1.33873856e+00 -2.10138038e-02 5.01484096e-01 -2.68373638e-01
-2.97677934e-01 -3.89204890e-01 -4.38944936e-01 6.45918548e-01
1.24471784e+00 -3.45631212e-01 4.30387817e-02 6.70756400e-01
4.68629152e-01 1.28928408e-01 -1.04766130e+00 1.10368717e+00
-1.49254665e-01 -1.49439752e+00 2.91581839e-01 5.22565663e-01
5.09923041e-01 -2.53705651e-01 3.22427511e-01 -1.22499943e-01
-2.74320304e-01 -1.21418345e+00 1.18937068e-01 7.12486565e-01
8.67398024e-01 -1.13866973e+00 7.25413561e-01 3.31668228e-01
-8.56158614e-01 1.10421270e-01 -5.43235898e-01 5.68640195e-02
2.36158706e-02 6.59661293e-01 -5.71614027e-01 5.69706500e-01
7.62911916e-01 8.30828011e-01 -6.46287143e-01 9.25148129e-01
4.45367098e-01 3.00624281e-01 -3.15726936e-01 -1.50902867e-01
5.04818678e-01 -6.15657032e-01 6.79070055e-01 1.17481387e+00
3.51655722e-01 -1.66623384e-01 -4.34709102e-01 9.04753566e-01
4.98759657e-01 3.25466037e-01 -1.09514308e+00 -3.20282072e-01
2.62697160e-01 1.20576918e+00 -7.12651849e-01 -1.71641499e-01
-1.55393168e-01 8.89264166e-01 2.09290534e-01 3.59531403e-01
-3.70855629e-01 -6.79234147e-01 8.85783672e-01 4.40732598e-01
1.30759478e-01 -5.40251911e-01 -1.52207747e-01 -6.56289279e-01
6.59899563e-02 -5.05349040e-01 5.04903793e-01 -6.72212422e-01
-1.29448974e+00 8.56494248e-01 1.31082192e-01 -1.51228917e+00
-9.30387229e-02 -9.23233807e-01 -4.47401494e-01 1.04161263e+00
-1.06957066e+00 -7.13932097e-01 -1.33062676e-01 5.69421589e-01
1.23299390e-01 -3.35248709e-01 1.18539608e+00 2.39051208e-01
-6.55318141e-01 2.93561995e-01 9.93528128e-01 -1.35462999e-01
4.83365536e-01 -1.37876558e+00 3.48376065e-01 3.67726743e-01
8.17383900e-02 6.87186539e-01 8.36529613e-01 -4.73404318e-01
-1.48767388e+00 -1.04843009e+00 4.55837220e-01 -3.32385242e-01
6.36712909e-01 -6.34212017e-01 -1.26137030e+00 1.04724705e-01
2.06802145e-01 2.01924577e-01 8.93034160e-01 -2.14129746e-01
-2.26252019e-01 9.12030563e-02 -1.36495209e+00 5.42251766e-01
7.39617646e-01 -3.84367824e-01 -5.25964916e-01 3.35610330e-01
5.53902805e-01 3.52864340e-03 -1.15536523e+00 1.66024074e-01
5.70088089e-01 -1.34217465e+00 9.51608002e-01 -7.72122562e-01
3.89195740e-01 -4.35550839e-01 2.28309631e-02 -1.44312310e+00
-3.59059542e-01 -3.87607396e-01 -4.42152351e-01 1.12746561e+00
3.93363684e-01 -9.36247528e-01 8.71875584e-01 4.78272408e-01
5.52708358e-02 -1.14807332e+00 -1.27107775e+00 -9.24838901e-01
4.29886669e-01 1.76003762e-02 4.36515123e-01 1.16598642e+00
1.27470776e-01 4.18049209e-02 -2.31276855e-01 -8.99144635e-02
5.57744503e-01 -1.30714253e-01 8.28589559e-01 -1.49319243e+00
2.98123091e-01 -5.71897030e-01 -1.13648176e+00 -2.28511482e-01
-1.55502230e-01 -9.88272786e-01 -9.45328832e-01 -1.69488609e+00
-1.59554586e-01 -2.61445016e-01 -3.99427176e-01 8.79782587e-02
2.62765467e-01 5.93055710e-02 -5.00917099e-02 4.98769939e-01
1.06067918e-01 8.18686306e-01 1.33311605e+00 6.64097220e-02
-5.54719746e-01 -5.58846474e-01 -2.33905360e-01 4.23860490e-01
7.10264862e-01 -4.17443931e-01 -5.22066653e-01 2.36043986e-02
-6.16699122e-02 -3.38140398e-01 3.39763492e-01 -1.09047902e+00
3.36246863e-02 2.47327626e-01 6.35997057e-01 -9.87877071e-01
5.49877167e-01 -9.08235788e-01 3.67882222e-01 5.61558366e-01
-2.68312424e-01 6.20772243e-01 4.51755226e-01 7.02371657e-01
-3.18863213e-01 2.55173951e-01 7.96964943e-01 2.02280119e-01
-5.13032496e-01 1.40091717e-01 -3.64469081e-01 -2.39236057e-01
1.29873002e+00 -7.65469670e-01 -5.02223194e-01 -6.17311941e-03
-7.79291689e-01 -7.75593221e-02 5.46111703e-01 4.69753951e-01
8.56242657e-01 -1.56716108e+00 -4.02453989e-01 5.19466758e-01
9.43123624e-02 -6.50870427e-02 3.94497544e-01 9.60097075e-01
-7.02372372e-01 5.64630151e-01 -5.92535794e-01 -8.24265003e-01
-1.34633768e+00 1.12060416e+00 9.59398001e-02 8.83247927e-02
-1.03394711e+00 4.67558324e-01 1.72628373e-01 -6.59267306e-01
1.22153863e-01 -3.08655888e-01 -5.15847921e-01 3.59446198e-01
5.57750285e-01 7.94619203e-01 2.20192950e-02 -8.66097569e-01
-3.14812332e-01 6.99285746e-01 -6.74984977e-02 -1.30330607e-01
1.56491852e+00 1.76266469e-02 -3.59047681e-01 8.89677942e-01
1.72800148e+00 -4.64407474e-01 -9.94199097e-01 1.19217187e-01
1.36445269e-01 -5.82121015e-01 1.23880304e-01 -6.24040842e-01
-7.71683633e-01 1.41700625e+00 1.18073881e+00 5.74200213e-01
8.78299415e-01 -3.68331373e-01 4.43877161e-01 2.17226714e-01
-5.93617931e-02 -8.70758832e-01 5.25658950e-02 1.42201588e-01
1.06430352e+00 -8.07810843e-01 -1.03507295e-01 1.09780714e-01
-4.60216284e-01 1.38000202e+00 4.71932352e-01 -6.35982901e-02
8.80959868e-01 1.56710416e-01 1.89895570e-01 -5.70902169e-01
-6.36233032e-01 3.10469449e-01 2.49482170e-01 8.96466970e-01
3.91100168e-01 -1.26840055e-01 -3.97865355e-01 -4.95530106e-02
-2.60593534e-01 -3.10764581e-01 2.44533151e-01 9.42652881e-01
-2.22962424e-01 -9.79538918e-01 -4.26921159e-01 4.18851465e-01
2.69842613e-02 1.63267866e-01 -6.73244894e-01 1.25884712e+00
-3.32783721e-02 1.77340448e-01 2.25325122e-01 -4.83784050e-01
2.66981870e-01 2.27131844e-01 -8.15736651e-02 -1.04886658e-01
-3.74987870e-01 -1.00508638e-01 -2.60812163e-01 -5.04311442e-01
-3.83904934e-01 -5.65189898e-01 -1.22011149e+00 -3.64963174e-01
-1.20467238e-01 3.48354846e-01 1.13475025e+00 4.62824523e-01
8.82986605e-01 3.61139655e-01 4.83562142e-01 -1.10904014e+00
-2.18778387e-01 -9.59488153e-01 -8.81279826e-01 4.84530479e-01
3.57817650e-01 -8.11781943e-01 -5.20042837e-01 -2.18747348e-01] | [8.000448226928711, 4.446290493011475] |
012fa89a-1cfb-443c-a7b4-db97cd67fae5 | enhancing-vision-language-pre-training-with | 2305.11769 | null | https://arxiv.org/abs/2305.11769v1 | https://arxiv.org/pdf/2305.11769v1.pdf | Enhancing Vision-Language Pre-Training with Jointly Learned Questioner and Dense Captioner | Large pre-trained multimodal models have demonstrated significant success in a range of downstream tasks, including image captioning, image-text retrieval, visual question answering (VQA), etc. However, many of these methods rely on image-text pairs collected from the web as pre-training data and unfortunately overlook the need for fine-grained feature alignment between vision and language modalities, which requires detailed understanding of images and language expressions. While integrating VQA and dense captioning (DC) into pre-training can address this issue, acquiring image-question-answer as well as image-location-caption triplets is challenging and time-consuming. Additionally, publicly available datasets for VQA and dense captioning are typically limited in scale due to manual data collection and labeling efforts. In this paper, we propose a novel method called Joint QA and DC GEneration (JADE), which utilizes a pre-trained multimodal model and easily-crawled image-text pairs to automatically generate and filter large-scale VQA and dense captioning datasets. We apply this method to the Conceptual Caption (CC3M) dataset to generate a new dataset called CC3M-QA-DC. Experiments show that when used for pre-training in a multi-task manner, CC3M-QA-DC can improve the performance with various backbones on various downstream tasks. Furthermore, our generated CC3M-QA-DC can be combined with larger image-text datasets (e.g., CC15M) and achieve competitive results compared with models using much more data. Code and dataset will be released. | ['Jing Liu', 'Xingjian He', 'Handong Li', 'Longteng Guo', 'Sihan Chen', 'Zikang Liu'] | 2023-05-19 | null | null | null | null | ['dense-captioning'] | ['computer-vision'] | [ 3.52283001e-01 -6.00052625e-02 7.21034929e-02 -5.09438217e-01
-1.47117829e+00 -8.08089197e-01 7.98147559e-01 -1.34537201e-02
-4.60899770e-01 4.52007860e-01 2.96572030e-01 -2.86327124e-01
4.50727701e-01 -4.04937476e-01 -9.99548793e-01 -3.52695197e-01
5.80269158e-01 6.29663825e-01 2.15652376e-01 -2.62837023e-01
1.21502988e-02 -7.19653144e-02 -1.45890152e+00 7.21373439e-01
8.71966302e-01 1.10612869e+00 5.57911932e-01 7.04079270e-01
-3.67200643e-01 5.99905789e-01 -2.86647022e-01 -8.31224859e-01
3.00667137e-01 -4.72997040e-01 -9.47619259e-01 3.01365167e-01
8.74749124e-01 -5.52717447e-01 -1.91755906e-01 8.00766528e-01
3.98758292e-01 -5.76783493e-02 5.84357440e-01 -1.62905681e+00
-9.01864707e-01 4.71569598e-02 -6.38537288e-01 -3.28316778e-01
3.27867210e-01 4.32240009e-01 9.02716637e-01 -1.13224983e+00
7.32453346e-01 1.35377884e+00 4.30674195e-01 7.50057280e-01
-1.22007072e+00 -5.56395113e-01 -7.43872523e-02 2.64096707e-01
-1.14624429e+00 -4.71836895e-01 6.17172599e-01 -4.23138618e-01
6.41643167e-01 1.90833658e-01 2.91502982e-01 1.28939974e+00
-2.10263208e-01 1.15554702e+00 1.07556081e+00 -3.83955210e-01
-8.61343294e-02 1.94575936e-01 -2.51476973e-01 7.84212589e-01
-9.89155844e-02 -2.72258103e-01 -4.06876087e-01 -8.23688954e-02
6.24628186e-01 -2.10854828e-01 -3.65521371e-01 -4.17224675e-01
-1.45200789e+00 9.71777916e-01 4.75214690e-01 7.20852166e-02
-2.79123098e-01 2.93039709e-01 4.13211554e-01 2.29082793e-01
1.81857988e-01 4.16602492e-01 -3.89491767e-01 -1.25149945e-02
-9.00297523e-01 2.04621121e-01 4.26141083e-01 1.04128599e+00
1.01461112e+00 -3.86670351e-01 -4.80372936e-01 9.19991314e-01
3.05966467e-01 1.03960669e+00 3.49245846e-01 -1.07756615e+00
8.97286534e-01 6.33276165e-01 5.57408072e-02 -9.31134939e-01
-1.05467193e-01 2.03732327e-01 -8.92334700e-01 -2.72467226e-01
4.51040357e-01 -2.19373349e-02 -1.41652036e+00 1.92285597e+00
2.55110294e-01 -8.31723660e-02 1.82898626e-01 1.14126253e+00
1.09985793e+00 9.11941051e-01 2.76977122e-01 1.36388808e-01
1.49321878e+00 -1.31735468e+00 -5.14027536e-01 -5.19160628e-01
6.12980247e-01 -1.00732863e+00 1.38602459e+00 -1.01016916e-01
-9.72462416e-01 -5.85634291e-01 -5.56001663e-01 -4.50170159e-01
-3.90287697e-01 2.51150995e-01 4.60516542e-01 2.95642376e-01
-1.25284040e+00 -2.77959585e-01 -4.83841598e-01 -5.95388949e-01
5.66369832e-01 1.32826060e-01 -7.32827842e-01 -6.52121484e-01
-1.05294299e+00 7.04099417e-01 1.08876638e-01 1.13683760e-01
-1.35850477e+00 -5.72917104e-01 -1.12387288e+00 -1.73547015e-01
2.43933469e-01 -9.78637874e-01 1.20246875e+00 -1.16974688e+00
-1.08802140e+00 1.06493425e+00 -3.34951699e-01 -2.63634205e-01
4.25326765e-01 -9.85770747e-02 -1.11148775e-01 5.84630549e-01
3.09023172e-01 1.54118609e+00 1.03987217e+00 -1.54649270e+00
-3.64099860e-01 -2.32436076e-01 1.54580772e-01 2.32884094e-01
-3.02747458e-01 -8.74378458e-02 -1.12745702e+00 -2.51365304e-01
-2.67319798e-01 -1.03090382e+00 -2.35389113e-01 1.00960977e-01
-3.82406950e-01 -1.73690662e-01 8.97453904e-01 -7.59174407e-01
6.82314992e-01 -1.98196840e+00 9.65423137e-02 -7.41254836e-02
1.16626300e-01 5.41971385e-01 -1.05105400e+00 5.51636517e-01
2.61989355e-01 7.63616040e-02 -3.85250002e-01 -6.17797852e-01
-7.38592222e-02 3.52986246e-01 -3.91869068e-01 1.27830938e-01
6.57965004e-01 1.61062586e+00 -9.04796243e-01 -1.00532079e+00
2.74890840e-01 4.14372683e-01 -4.75078851e-01 5.65433860e-01
-6.98920429e-01 4.83919054e-01 -4.57672119e-01 8.11941981e-01
7.00888991e-01 -6.09649837e-01 -1.00599870e-01 -5.18559515e-01
3.11445177e-01 -2.96467453e-01 -5.25486588e-01 1.88344026e+00
-7.09998846e-01 8.32580864e-01 1.36359125e-01 -8.83618712e-01
7.26489127e-01 3.21555436e-01 3.19288373e-01 -1.05485260e+00
1.70845628e-01 1.89663574e-01 -4.15190488e-01 -8.01483154e-01
5.93467951e-01 1.58735558e-01 -2.89849401e-01 3.95328313e-01
1.66717112e-01 -3.52106601e-01 3.22156399e-01 6.64524972e-01
8.92155707e-01 1.19210720e-01 -2.38829508e-01 3.38132024e-01
6.94033623e-01 4.12505686e-01 2.33410135e-01 7.27782905e-01
-3.31006378e-01 1.11845839e+00 3.28969330e-01 -1.35227880e-02
-1.28926313e+00 -1.01967168e+00 9.28242281e-02 9.51300979e-01
2.30473623e-01 -4.05902833e-01 -7.94528186e-01 -7.50610888e-01
-8.74296278e-02 3.79607469e-01 -5.60902596e-01 4.58594337e-02
-3.94226402e-01 -4.51846033e-01 5.90276837e-01 3.59306842e-01
7.24965572e-01 -1.06664777e+00 -2.76295900e-01 -8.60400200e-02
-7.88713872e-01 -1.56934273e+00 -7.10087717e-01 -3.40465784e-01
-4.82137322e-01 -1.05429924e+00 -1.01813722e+00 -1.08194709e+00
8.91988873e-01 5.48677087e-01 1.31829357e+00 1.99213356e-01
-2.84484088e-01 8.60463858e-01 -5.14569700e-01 -1.42896876e-01
-4.59515989e-01 1.34335402e-02 -4.78561044e-01 2.28132010e-01
2.65560627e-01 -8.33636522e-02 -7.97654927e-01 3.88365716e-01
-1.21801984e+00 4.40071881e-01 8.73055577e-01 9.95376706e-01
6.62447870e-01 -7.19940901e-01 6.11098945e-01 -6.64623082e-01
4.66672212e-01 -4.32634383e-01 -5.08276463e-01 6.17809951e-01
-3.73170882e-01 7.46638998e-02 3.48744482e-01 -4.48349833e-01
-1.01634741e+00 3.44796449e-01 -1.67680427e-01 -7.61280894e-01
-1.49756715e-01 4.05733526e-01 -3.55807602e-01 -1.22769997e-01
4.84117448e-01 4.25322473e-01 2.74599940e-01 -3.27272981e-01
8.98490787e-01 8.00765157e-01 8.31302881e-01 -5.44782400e-01
1.00035942e+00 5.23236990e-01 -2.20539019e-01 -5.56745708e-01
-9.71514344e-01 -6.08705163e-01 -4.30236638e-01 -1.99786305e-01
1.29482520e+00 -1.19418895e+00 -4.27461118e-01 4.28383380e-01
-1.46582687e+00 -3.25902641e-01 1.85843542e-01 1.39496624e-01
-5.80730438e-01 5.52675486e-01 -3.91427606e-01 -5.06054342e-01
-5.76026857e-01 -1.28492475e+00 1.67332852e+00 1.64590970e-01
2.55609095e-01 -7.65479803e-01 -2.57859798e-03 1.12476218e+00
3.34777474e-01 1.25808477e-01 8.64413202e-01 -3.02097291e-01
-9.25138533e-01 -1.18819430e-01 -8.18814993e-01 5.21802783e-01
-1.28020152e-01 -3.95604044e-01 -7.54797339e-01 -2.04747826e-01
-5.59642732e-01 -1.03963566e+00 8.56100261e-01 1.53802246e-01
1.09738445e+00 -1.02747321e-01 -2.51921862e-01 4.65316594e-01
1.44848037e+00 -2.64942020e-01 7.86590755e-01 2.71888167e-01
1.04213810e+00 6.93364859e-01 8.78095031e-01 -3.90597843e-02
8.85222733e-01 6.58362210e-01 6.14001930e-01 -4.45806146e-01
-4.05183733e-01 -4.19214875e-01 3.97186190e-01 6.76986575e-01
4.14107114e-01 -4.28459674e-01 -9.95799661e-01 1.00405014e+00
-1.95767152e+00 -7.29675829e-01 -2.66020447e-01 1.84802985e+00
9.62029278e-01 -5.51357210e-01 -1.34121910e-01 -5.08897960e-01
6.28071368e-01 -3.83045822e-02 -4.97355521e-01 -1.71695322e-01
-2.27565274e-01 -1.03507675e-01 3.34972858e-01 3.78212005e-01
-1.07644856e+00 1.19948208e+00 5.44822073e+00 7.65833616e-01
-1.04370689e+00 3.36550117e-01 6.78722262e-01 4.44012322e-02
-6.12563968e-01 -1.01659922e-02 -6.06295109e-01 3.22358668e-01
8.28768432e-01 1.61083117e-01 2.75443166e-01 6.50719523e-01
3.64301689e-02 -1.65707648e-01 -1.08397102e+00 1.47700524e+00
4.33911234e-01 -1.37994385e+00 5.08751869e-01 -1.12145618e-01
8.87419999e-01 3.64675343e-01 6.96317628e-02 4.14339185e-01
1.05011173e-01 -9.67325747e-01 5.43007016e-01 3.03636581e-01
1.05049074e+00 -3.15770984e-01 7.92690396e-01 1.37689531e-01
-1.02906907e+00 1.13203108e-01 -2.74412423e-01 4.67135310e-01
4.73375678e-01 4.65982974e-01 -8.68339956e-01 5.80726922e-01
7.38984466e-01 4.84840572e-01 -1.00042415e+00 8.90858233e-01
-9.69883278e-02 4.15764272e-01 -2.38789067e-01 4.86043878e-02
4.99940157e-01 -1.61921233e-01 2.32056171e-01 1.06441581e+00
2.13071287e-01 -1.23591773e-01 1.63601726e-01 7.72097766e-01
-2.89580137e-01 1.97574914e-01 -5.37094355e-01 -3.51911694e-01
2.41864294e-01 1.55185914e+00 -3.63742173e-01 -4.63339597e-01
-7.06075132e-01 1.22657466e+00 3.25384617e-01 5.51224291e-01
-8.19025099e-01 -1.58915728e-01 5.75181544e-01 -3.65759395e-02
3.97409141e-01 -1.64844662e-01 1.98923469e-01 -1.41950572e+00
-1.93752372e-03 -1.10543799e+00 3.84887695e-01 -1.31805706e+00
-1.44899976e+00 8.69522095e-01 -2.03722459e-03 -1.14660084e+00
-4.07206684e-01 -6.09192550e-01 -2.68873304e-01 7.17315197e-01
-1.75612724e+00 -1.78403807e+00 -7.49313235e-01 1.03425908e+00
5.71294129e-01 -3.01079657e-02 6.62894249e-01 4.60875303e-01
-3.11439097e-01 6.06980145e-01 -9.72427055e-02 2.86007226e-01
1.07029402e+00 -1.01334035e+00 3.43753308e-01 7.66567826e-01
4.21006322e-01 1.84818193e-01 3.79566133e-01 -4.39369619e-01
-1.55198884e+00 -1.56365597e+00 7.89992809e-01 -8.36064398e-01
4.82668221e-01 -7.37637877e-01 -8.23067248e-01 5.28583944e-01
5.48346043e-01 9.68759060e-02 4.89749789e-01 -2.50201613e-01
-4.97807533e-01 -7.60027170e-02 -8.45142543e-01 6.46077096e-01
8.75350952e-01 -7.97062516e-01 -5.31110168e-01 5.48994780e-01
9.23368692e-01 -2.79449224e-01 -7.31200695e-01 3.26040387e-01
2.96054810e-01 -6.25878692e-01 9.88195717e-01 -5.12496591e-01
6.87244236e-01 -3.64329040e-01 -3.04613739e-01 -1.10768425e+00
2.22533554e-01 -4.73756909e-01 2.88672894e-01 1.51356030e+00
5.08821964e-01 -2.03987494e-01 5.95466077e-01 5.67343712e-01
-1.25381555e-02 -5.32606661e-01 -7.79094875e-01 -4.75680172e-01
-8.51687416e-03 -5.23965359e-01 4.85090584e-01 8.66942048e-01
-4.91504878e-01 7.25375295e-01 -5.79429090e-01 3.98779623e-02
6.08408988e-01 4.35024232e-01 1.20519781e+00 -7.71765709e-01
-9.82561409e-02 -5.41377347e-03 -2.49631912e-01 -1.21888399e+00
1.98343590e-01 -1.03487313e+00 3.08442444e-01 -1.94575679e+00
4.57594723e-01 -3.13100040e-01 7.62322992e-02 6.82880759e-01
-3.22171152e-01 7.56954670e-01 3.48954678e-01 3.21412891e-01
-1.04885173e+00 8.28103900e-01 1.58159626e+00 -3.48557353e-01
5.58995120e-02 -6.28614843e-01 -5.76940060e-01 2.66118556e-01
5.02136528e-01 -3.05795372e-01 -6.07201517e-01 -8.02475929e-01
9.65999290e-02 1.37201756e-01 7.08134353e-01 -6.38601661e-01
3.33054699e-02 -7.05269203e-02 3.69427085e-01 -7.51658142e-01
5.19336283e-01 -7.30858684e-01 -1.14660405e-01 9.95787382e-02
-3.12473148e-01 1.95794031e-01 2.25554749e-01 6.71682000e-01
-5.31886518e-01 -2.60665603e-02 4.55728889e-01 -1.51508659e-01
-1.02096450e+00 5.13111591e-01 -6.48141950e-02 2.61698812e-01
9.84264970e-01 1.00319646e-01 -7.52355695e-01 -7.11191297e-01
-3.42132717e-01 7.82425940e-01 5.89599311e-01 8.51450980e-01
7.89663076e-01 -1.41711938e+00 -8.86336088e-01 -1.05911633e-03
6.76246583e-01 1.50678113e-01 3.98876488e-01 8.66837025e-01
-4.02085751e-01 6.14307761e-01 -1.42275721e-01 -1.00452566e+00
-1.28436470e+00 5.92364132e-01 4.03286032e-02 -1.75663322e-01
-3.48636568e-01 7.11719632e-01 5.60424328e-01 -5.68137705e-01
-1.20302560e-02 -6.28125891e-02 7.33926743e-02 -9.61340666e-02
4.66436177e-01 -4.44822878e-01 -2.31589258e-01 -8.57038438e-01
-3.37689638e-01 7.08903253e-01 -7.01649711e-02 -2.28696421e-01
1.02995360e+00 -4.86887336e-01 -1.81106612e-01 -7.70124346e-02
1.54560959e+00 -4.04911399e-01 -1.26129317e+00 -2.36248821e-01
-1.15923658e-01 -4.25827026e-01 -4.91362028e-02 -7.83189178e-01
-1.12740958e+00 1.08812261e+00 5.70713401e-01 -3.02160323e-01
1.30521977e+00 4.10360754e-01 1.18382990e+00 4.69036162e-01
2.86349535e-01 -9.07195389e-01 6.51155770e-01 3.67348045e-01
1.06458735e+00 -1.75159359e+00 -3.85009766e-01 -3.27690393e-01
-1.06530905e+00 7.46779919e-01 8.74721110e-01 3.22690874e-01
2.04866812e-01 -3.22536200e-01 5.43680966e-01 -9.79030579e-02
-8.88256848e-01 -5.64007878e-01 5.15707672e-01 7.90839612e-01
8.81170481e-02 -2.50072271e-01 2.97622569e-02 3.30498099e-01
7.33159184e-02 -1.41255939e-02 3.91147643e-01 7.06504941e-01
-1.51096508e-01 -1.35106075e+00 -4.14680570e-01 3.06003571e-01
-9.89047214e-02 -3.48627031e-01 -4.12655503e-01 6.55072212e-01
-2.81632459e-03 1.16384542e+00 8.58791098e-02 -3.30238521e-01
1.00721374e-01 1.67264789e-02 5.48629880e-01 -5.05803168e-01
-2.79878527e-01 -3.08635011e-02 1.91380039e-01 -7.15251088e-01
-5.49996018e-01 -4.08237338e-01 -1.07303500e+00 7.18674138e-02
-1.81575909e-01 1.34286061e-01 8.85501564e-01 9.79086399e-01
7.16567993e-01 9.51260626e-02 4.06274796e-01 -7.72514641e-01
-3.37539971e-01 -8.76670718e-01 1.51492320e-02 9.24331546e-01
4.45116669e-01 -4.24773812e-01 -1.58696830e-01 3.84236783e-01] | [10.934112548828125, 1.3757244348526] |
98ae153d-3b3b-4eb5-ba68-0bfe0195fed4 | mead-a-large-scale-audio-visual-dataset-for | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3837_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123660698.pdf | MEAD: A Large-scale Audio-visual Dataset for Emotional Talking-face Generation | The synthesis of natural emotional reactions is an essentialcriteria in vivid talking-face video generation. This criteria is nevertheless seldom taken into consideration in previous works due to the absence of a large-scale, high-quality emotional audio-visual dataset. To address this issue, we build the Multi-view Emotional Audio-visual Dataset(MEAD) which is a talking-face video corpus featuring 60 actors and actresses talking with 8 different emotions at 3 different intensity levels. High-quality audio-visual clips are captured at 7 different view angles in a strictly-controlled environment. Together with the dataset, we release an emotional talking-face generation baseline which enables the manipulation of both emotion and its intensity. Our dataset will be made public and could benefit a number of different research fields including conditional generation, cross-modal understanding and expression recognition. | ['Kaisiyuan Wang Qianyi Wu Linsen Song Zhuoqian Yang Wayne Wu Chen Qian Ran He Yu Qiao Chen Change Loy'] | null | null | null | null | eccv-2020-8 | ['talking-head-generation', 'talking-face-generation'] | ['computer-vision', 'computer-vision'] | [ 2.03685075e-01 4.71724314e-04 6.31869137e-02 -6.46334767e-01
-7.76821077e-01 -6.74855530e-01 6.89251900e-01 -4.18080032e-01
1.32638454e-01 5.04994988e-01 5.38486540e-01 3.64810228e-01
2.63460636e-01 -5.02066076e-01 -4.46249038e-01 -7.71919310e-01
6.21542819e-02 4.88451123e-02 -1.51575685e-01 -2.23775283e-01
-2.01069742e-01 5.65606654e-01 -2.24379277e+00 7.77253687e-01
1.17853940e-01 1.22803104e+00 -2.96819389e-01 8.50636482e-01
1.91119328e-01 7.02813327e-01 -7.66934216e-01 -8.32546294e-01
-1.22350957e-02 -5.55483758e-01 -5.61412334e-01 3.24989051e-01
6.04697585e-01 -1.50459841e-01 1.08244322e-01 8.20066750e-01
9.19148624e-01 -6.59266412e-02 5.19833386e-01 -1.53812325e+00
-3.77139956e-01 3.83691609e-01 -5.58495283e-01 3.10892612e-02
8.90558660e-01 1.48602143e-01 1.09588575e+00 -9.15164590e-01
1.18423283e+00 1.32424009e+00 2.62295574e-01 7.11702466e-01
-1.30036354e+00 -8.54427755e-01 -1.21910073e-01 3.81288409e-01
-1.19036937e+00 -9.83494759e-01 1.17835271e+00 -4.11093235e-01
5.67436576e-01 4.20992643e-01 9.42139745e-01 2.03884220e+00
-1.24833986e-01 5.59597850e-01 1.34311938e+00 -3.53788018e-01
2.33705938e-02 3.45607042e-01 -5.72750151e-01 1.09007545e-01
-6.87252283e-01 1.95919082e-01 -1.05708194e+00 -4.68946993e-02
3.69808584e-01 -4.96679038e-01 -4.39945489e-01 -2.35327616e-01
-1.06214190e+00 7.27360487e-01 -1.26628891e-01 5.15572488e-01
-3.80589277e-01 2.51442064e-02 6.50988400e-01 6.10419095e-01
5.33871531e-01 1.27149940e-01 -8.71226415e-02 -6.35640502e-01
-9.30739462e-01 2.22568929e-01 5.11817396e-01 6.55161321e-01
4.35123801e-01 3.88413191e-01 -1.33066580e-01 9.89429235e-01
1.24613807e-01 4.50909227e-01 -5.96955977e-03 -1.14077294e+00
1.02206387e-01 1.01570629e-01 -9.11425799e-02 -1.23026156e+00
-2.62510538e-01 1.15630306e-01 -6.98784113e-01 3.25663447e-01
1.35353819e-01 -2.93632895e-01 -1.88826501e-01 1.89252102e+00
5.02912581e-01 2.60176420e-01 3.12190549e-03 1.09501100e+00
1.23840773e+00 6.26234353e-01 -1.15857810e-01 -5.79608679e-01
1.63516808e+00 -3.46617967e-01 -1.16419947e+00 2.74731517e-01
1.42431930e-01 -1.01666915e+00 1.14925599e+00 8.80619287e-01
-1.32668293e+00 -5.37073731e-01 -8.54129732e-01 -2.39620656e-02
-2.09263846e-01 1.03744350e-01 5.81750631e-01 7.73116887e-01
-1.00963295e+00 1.47081092e-01 -2.17618451e-01 -1.74751326e-01
3.67348760e-01 -1.30212912e-02 -8.78063738e-01 1.79889143e-01
-1.31637216e+00 5.54799318e-01 -1.98682010e-01 1.33967102e-01
-8.86640668e-01 -7.14957356e-01 -8.80656838e-01 -2.82255501e-01
1.79018661e-01 -3.95602375e-01 1.14971697e+00 -1.49371147e+00
-1.83481824e+00 1.27675486e+00 -2.14323401e-01 1.03991419e-01
4.15935874e-01 -1.61384195e-01 -7.54801750e-01 6.15505934e-01
-3.20403963e-01 7.73644924e-01 1.17482376e+00 -1.41529596e+00
6.16163313e-02 -4.08288658e-01 -2.03143656e-02 -5.63122146e-03
-2.24198788e-01 5.00083685e-01 -2.50302672e-01 -8.73659730e-01
-5.00831783e-01 -8.02793741e-01 3.34814161e-01 -1.27778277e-02
-4.12808567e-01 1.78754136e-01 9.61543083e-01 -4.72904861e-01
8.80617917e-01 -2.39485240e+00 5.07788360e-01 -2.97474936e-02
1.48857251e-01 5.82132451e-02 -1.00260444e-01 5.13722956e-01
-5.41191638e-01 9.33632925e-02 2.72916675e-01 -4.86803502e-01
2.58084815e-02 -2.46507362e-01 -3.01894784e-01 4.14244264e-01
3.42597067e-01 6.57705784e-01 -7.36103833e-01 -8.89829338e-01
2.89417893e-01 9.91131902e-01 -6.40951514e-01 4.19213802e-01
-2.87880711e-02 6.67144597e-01 -1.73881516e-01 8.21960628e-01
6.11356437e-01 3.77772570e-01 -8.14521685e-02 -3.69359136e-01
-2.85348874e-02 -2.15537027e-01 -1.13422239e+00 1.67399776e+00
-6.24425650e-01 1.08102763e+00 4.90257472e-01 -5.58584571e-01
1.02655387e+00 8.96709621e-01 6.95449769e-01 -4.95974869e-01
4.74657804e-01 -1.58749118e-01 -2.93452114e-01 -8.74974430e-01
4.35016543e-01 -7.28193283e-01 -1.46637276e-01 2.74751186e-01
2.90295541e-01 -3.73696089e-01 1.60932422e-01 1.46165356e-01
8.10085356e-01 -1.23376874e-02 -8.00676718e-02 1.76130965e-01
5.52861750e-01 -6.45467520e-01 2.22796500e-01 7.79628828e-02
-2.51278311e-01 8.51764917e-01 8.45126152e-01 -1.59996599e-01
-6.58274829e-01 -1.04224861e+00 -2.15738744e-01 1.20561337e+00
-3.44991684e-01 -6.97603106e-01 -6.95647061e-01 -1.92129627e-01
-5.16432881e-01 4.21568632e-01 -7.70825922e-01 -1.27655879e-01
-2.26368196e-02 -2.80921668e-01 6.99904025e-01 1.77973688e-01
1.59547180e-01 -1.42399335e+00 -6.69896841e-01 -1.87332824e-01
-7.32971191e-01 -1.37826216e+00 -1.14496715e-01 -1.86867952e-01
-1.01094924e-01 -9.01771188e-01 -5.98027468e-01 -3.79535228e-01
1.58792719e-01 -2.00372338e-01 1.32120252e+00 -4.31048185e-01
-4.59088176e-01 6.51081502e-01 -6.00875318e-01 -4.81411070e-01
-5.60060203e-01 -4.05993253e-01 1.44314347e-02 5.69045603e-01
2.63028622e-01 -9.58898604e-01 -3.50433081e-01 2.58424282e-01
-9.06261504e-01 3.34108956e-02 3.08443662e-02 5.67010105e-01
4.53291208e-01 -2.94359386e-01 7.52552748e-01 -4.65145588e-01
7.67586529e-01 -3.78871024e-01 -2.30899587e-01 -8.96463320e-02
2.84196645e-01 -4.70410198e-01 5.62402308e-01 -7.09961176e-01
-1.27998686e+00 -2.20287457e-01 -4.10475880e-01 -8.53574157e-01
-5.65933585e-01 7.20672309e-02 -6.81716740e-01 1.89641774e-01
6.47050858e-01 -2.50613004e-01 -8.67008194e-02 -1.09548569e-02
5.40940285e-01 8.30543876e-01 7.55850136e-01 -5.03104925e-01
5.28129876e-01 5.90797961e-01 2.69625578e-02 -1.10878253e+00
-4.14359391e-01 -1.51147127e-01 -4.73378420e-01 -9.77325082e-01
9.61658657e-01 -1.04533529e+00 -9.85709608e-01 4.38595444e-01
-1.15189826e+00 -1.14856564e-01 -2.15939194e-01 2.97303438e-01
-8.75055373e-01 1.93237662e-01 -6.27706945e-01 -8.57778430e-01
-1.90795407e-01 -1.06341887e+00 1.41703749e+00 -1.42459041e-02
-5.37691355e-01 -5.71060956e-01 1.88668311e-01 5.28952062e-01
2.38030061e-01 7.64579415e-01 4.71753240e-01 7.82650784e-02
-2.61665788e-03 -1.25698566e-01 1.43168584e-01 2.57183373e-01
-7.61368945e-02 6.63945377e-01 -1.52940357e+00 8.23005959e-02
-2.99430750e-02 -1.09222186e+00 5.27978420e-01 3.62681001e-01
1.12089455e+00 -5.04189543e-02 1.01245262e-01 4.72104490e-01
9.31302786e-01 8.84385854e-02 8.62404406e-01 -3.84509504e-01
3.57655555e-01 1.14175951e+00 6.05888546e-01 7.80479074e-01
8.40542167e-02 1.08178890e+00 4.68640924e-01 -2.07326531e-01
3.11393160e-02 -3.67573015e-02 5.93735754e-01 7.26115167e-01
-2.97475517e-01 -3.83749276e-01 -5.47967136e-01 4.15233254e-01
-1.26158667e+00 -1.63293409e+00 -1.45118847e-01 1.81700909e+00
8.55720341e-01 -3.50907207e-01 3.33683133e-01 5.03356040e-01
5.32663226e-01 4.38909233e-01 -2.69616365e-01 -8.19977164e-01
-3.68891507e-01 1.98396236e-01 -6.53156757e-01 3.81389529e-01
-9.67721283e-01 6.90204799e-01 6.00371265e+00 8.06547642e-01
-1.62025082e+00 3.46595012e-02 6.53576255e-01 -7.18829095e-01
-4.39435571e-01 -3.75543714e-01 -4.52508628e-01 3.16951632e-01
1.23488808e+00 1.33071110e-01 3.61713678e-01 8.15470695e-01
5.33755004e-01 -4.25948612e-02 -1.03898704e+00 1.44446099e+00
4.01244849e-01 -9.76985991e-01 -2.66794473e-01 -8.29937309e-02
3.64564240e-01 -3.16534013e-01 2.43789583e-01 1.91917658e-01
-1.53247938e-01 -1.12704194e+00 8.50369453e-01 4.99990076e-01
1.20036829e+00 -9.12156582e-01 3.73882324e-01 -8.07502344e-02
-1.07256269e+00 2.56005645e-01 8.81164372e-02 1.40730992e-01
7.39684701e-01 5.49916208e-01 -2.39984915e-01 4.29003328e-01
1.06873393e+00 6.51245475e-01 -2.57749081e-01 2.64544874e-01
-1.00473739e-01 5.74922502e-01 -1.57367170e-01 2.04857200e-01
-2.66084641e-01 -1.75051108e-01 6.54855669e-01 1.30145919e+00
2.88875550e-01 1.75665140e-01 -3.46782595e-01 8.34327102e-01
-5.00431769e-02 2.53592521e-01 -9.27583992e-01 -2.78961271e-01
1.82018772e-01 1.72871220e+00 -5.06430328e-01 3.03086713e-02
-3.82310569e-01 8.67772639e-01 -1.12369157e-01 3.04385722e-01
-1.08895886e+00 -2.05391467e-01 8.56039166e-01 5.90430088e-02
1.99414983e-01 1.81754172e-01 2.75549263e-01 -1.27469337e+00
5.42273559e-02 -1.14668095e+00 1.42671034e-01 -1.27599001e+00
-1.19990218e+00 1.02985704e+00 2.63647288e-01 -1.22797751e+00
-6.82063997e-01 -4.61540222e-01 -4.76793557e-01 4.20599908e-01
-1.19414473e+00 -1.22073174e+00 -5.01269639e-01 9.02723968e-01
4.22875464e-01 4.03341018e-02 1.01591158e+00 5.30663133e-01
-3.97110999e-01 5.11004508e-01 -6.71691000e-01 -2.13424698e-01
1.11881602e+00 -7.28435874e-01 -3.60011041e-01 3.26611996e-01
2.29781821e-01 2.19599456e-01 1.03256452e+00 8.79714191e-02
-1.26124573e+00 -7.10515738e-01 7.66367376e-01 -5.37290215e-01
4.93386567e-01 -8.52616489e-01 -6.16079569e-01 4.24742043e-01
6.53396487e-01 1.15428373e-01 1.23311388e+00 4.18547951e-02
-3.90600294e-01 -1.15084648e-01 -1.10817897e+00 6.72662616e-01
9.67768669e-01 -7.40581036e-01 -2.69789815e-01 5.44000231e-03
3.77442926e-01 -2.20573246e-01 -1.04126883e+00 4.56705481e-01
8.84938240e-01 -1.70944643e+00 8.96306753e-01 -3.38991374e-01
7.30915248e-01 7.37474114e-02 -1.96905226e-01 -1.31073987e+00
2.84501314e-01 -8.30220044e-01 1.40795872e-01 1.75176251e+00
6.70645982e-02 -1.68249682e-01 4.43548322e-01 3.02769363e-01
1.22294493e-01 -6.80943787e-01 -9.93359029e-01 -2.93774366e-01
-1.86551407e-01 -8.84554327e-01 4.84202772e-01 9.87479985e-01
3.03018034e-01 5.55749714e-01 -7.61627972e-01 -2.97439069e-01
5.44213653e-02 2.75155604e-01 1.06298399e+00 -1.20701933e+00
-1.72166452e-01 -4.25557971e-01 -5.57202101e-01 -3.89407396e-01
5.87186635e-01 -5.57638288e-01 -2.08520174e-01 -7.82688081e-01
9.48457271e-02 2.25766584e-01 2.27087751e-01 2.51819670e-01
4.08839971e-01 5.71654618e-01 1.78939268e-01 -3.96332413e-01
-5.00290930e-01 9.03049886e-01 1.26283073e+00 4.20442782e-02
3.50745320e-02 -3.23003590e-01 -5.17177820e-01 7.35907733e-01
5.79350054e-01 -2.66954601e-01 -6.35253847e-01 1.66469991e-01
3.65437806e-01 5.05345225e-01 4.76153165e-01 -7.50616908e-01
-4.47578967e-01 -1.33921981e-01 2.95880616e-01 -4.63870376e-01
1.16040552e+00 -8.53166521e-01 5.86326838e-01 -1.75956443e-01
-3.81784230e-01 1.63126796e-01 1.60088524e-01 2.40079969e-01
-4.82484639e-01 1.93193123e-01 8.79676461e-01 -6.24092892e-02
-2.45912462e-01 2.16768488e-01 -5.35874665e-01 2.16552317e-01
1.13548326e+00 -3.51647973e-01 -1.03255048e-01 -1.01124310e+00
-8.20902467e-01 -2.02746198e-01 5.60583711e-01 6.33530736e-01
6.52575374e-01 -1.55812609e+00 -7.30210125e-01 3.27526301e-01
2.77157992e-01 -5.26260614e-01 6.56274498e-01 9.00101125e-01
-4.46888804e-02 -9.95932743e-02 -5.11439919e-01 -7.13886797e-01
-1.73388195e+00 5.23468077e-01 2.06558645e-01 7.53860921e-02
-2.96573192e-01 7.67678916e-01 3.54842432e-02 -8.35038871e-02
7.81214982e-02 1.39105216e-01 -4.69333023e-01 6.80867016e-01
7.12379038e-01 1.26241505e-01 -5.80547415e-02 -1.34944403e+00
-2.38794237e-01 5.12562394e-01 5.47437847e-01 -6.17131948e-01
1.26009297e+00 -1.51986361e-01 -8.61606095e-03 9.30405438e-01
1.34432912e+00 4.12403226e-01 -9.67192352e-01 3.80987734e-01
-5.47731280e-01 -6.73085928e-01 -6.87797219e-02 -3.16037744e-01
-1.40178168e+00 1.18901968e+00 4.91333634e-01 1.69638723e-01
1.41955459e+00 5.99768162e-02 5.50238788e-01 -8.73772949e-02
3.68529588e-01 -1.02431870e+00 6.51941121e-01 2.77841557e-02
1.36909819e+00 -1.11464393e+00 -4.11357433e-01 -4.51244712e-01
-9.70511496e-01 1.03160143e+00 5.87838471e-01 3.64193231e-01
6.89185679e-01 5.18479288e-01 4.57805455e-01 -3.46726835e-01
-1.19044101e+00 -1.00595625e-02 1.90944046e-01 8.26967299e-01
6.80254579e-01 -1.64040267e-01 -1.47555154e-02 7.62145817e-01
-7.19114900e-01 -4.90281321e-02 5.00008166e-01 5.24952412e-01
2.48677716e-01 -8.68442357e-01 -4.41222429e-01 -4.51002792e-02
-9.10359204e-01 2.69521445e-01 -7.24191487e-01 6.18054211e-01
2.17943594e-01 1.35061491e+00 -2.71437056e-02 -5.48644960e-01
3.79352272e-01 2.77950197e-01 6.14215970e-01 -2.25520030e-01
-6.29976034e-01 2.86372662e-01 3.03342164e-01 -7.52200723e-01
-9.30159032e-01 -6.98744833e-01 -8.67300868e-01 -4.39716607e-01
-1.92559183e-01 -3.57256196e-02 5.67646444e-01 5.10566711e-01
4.23771799e-01 3.70066047e-01 6.83917582e-01 -1.30406868e+00
1.99612021e-01 -9.09696281e-01 -6.78216338e-01 8.24423492e-01
2.30219245e-01 -8.81371498e-01 -5.67343533e-01 3.61455470e-01] | [13.255362510681152, -0.3889140784740448] |
990cc00e-bc0a-454f-9add-f4dcbc0199ad | phrase-localization-and-visual-relationship | 1611.06641 | null | http://arxiv.org/abs/1611.06641v4 | http://arxiv.org/pdf/1611.06641v4.pdf | Phrase Localization and Visual Relationship Detection with Comprehensive Image-Language Cues | This paper presents a framework for localization or grounding of phrases in
images using a large collection of linguistic and visual cues. We model the
appearance, size, and position of entity bounding boxes, adjectives that
contain attribute information, and spatial relationships between pairs of
entities connected by verbs or prepositions. Special attention is given to
relationships between people and clothing or body part mentions, as they are
useful for distinguishing individuals. We automatically learn weights for
combining these cues and at test time, perform joint inference over all phrases
in a caption. The resulting system produces state of the art performance on
phrase localization on the Flickr30k Entities dataset and visual relationship
detection on the Stanford VRD dataset. | ['Julia Hockenmaier', 'Christopher M. Cervantes', 'Svetlana Lazebnik', 'Bryan A. Plummer', 'Arun Mallya'] | 2016-11-21 | phrase-localization-and-visual-relationship-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Plummer_Phrase_Localization_and_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Plummer_Phrase_Localization_and_ICCV_2017_paper.pdf | iccv-2017-10 | ['visual-relationship-detection'] | ['computer-vision'] | [-1.94975108e-01 -4.08011638e-02 -1.66443005e-01 -6.64932907e-01
-5.41027069e-01 -9.15995479e-01 6.22249484e-01 8.84636045e-01
-5.72602808e-01 6.03411973e-01 4.12161767e-01 5.53591885e-02
1.93059444e-01 -6.94792747e-01 -7.52296805e-01 -3.34429502e-01
-3.88375580e-01 5.27588785e-01 3.07488233e-01 9.64900479e-02
3.28424647e-02 5.67695916e-01 -1.17687619e+00 2.77156234e-01
7.90656060e-02 1.07168758e+00 -2.40733922e-02 6.96984410e-01
-8.69579986e-02 3.25724065e-01 -4.49629277e-01 -8.10881615e-01
-1.48952782e-01 1.94459870e-01 -7.57200360e-01 3.28387588e-01
1.25983465e+00 -3.15266967e-01 -3.57998282e-01 9.39493299e-01
3.51889253e-01 1.15424201e-01 9.26469505e-01 -1.14313936e+00
-1.22976446e+00 4.35094744e-01 -7.94649899e-01 4.58873123e-01
7.43826687e-01 -2.18925610e-01 1.54598355e+00 -1.09281743e+00
7.55709291e-01 1.23420811e+00 5.46734154e-01 4.09783393e-01
-1.49063838e+00 -3.47192854e-01 4.06959027e-01 -1.85978450e-02
-1.73180997e+00 -3.38428617e-01 5.57725072e-01 -6.17710471e-01
1.14185309e+00 2.17797890e-01 5.07675588e-01 9.41983461e-01
-1.01756707e-01 8.28090489e-01 5.55159986e-01 -3.31982702e-01
-1.92334726e-02 4.51750904e-01 3.75431567e-01 9.97175276e-01
2.03178719e-01 -2.54362971e-01 -6.49445832e-01 -3.60332966e-01
9.22413588e-01 -2.03375578e-01 7.48356208e-02 -6.31406665e-01
-1.19451511e+00 5.46047568e-01 7.66238928e-01 1.30945578e-01
-2.70588189e-01 5.13088167e-01 2.42391378e-01 -2.79078126e-01
3.34132701e-01 4.32441771e-01 -3.27553451e-01 4.66442883e-01
-6.36792719e-01 4.02010113e-01 3.72549057e-01 1.23890173e+00
6.99044466e-01 -5.38088560e-01 -3.67123246e-01 8.87430429e-01
5.31696916e-01 7.30828345e-01 -1.25220686e-01 -8.01320791e-01
6.88479006e-01 5.66592991e-01 4.63570714e-01 -1.52335489e+00
-4.40165341e-01 -3.88612319e-03 -2.56868988e-01 -4.08933759e-01
3.06020856e-01 1.84000999e-01 -9.80554223e-01 1.57734001e+00
4.79668319e-01 1.12980224e-01 -9.85197797e-02 9.20681477e-01
1.48038650e+00 4.42941338e-01 5.75129151e-01 2.64987022e-01
1.96188831e+00 -7.01196849e-01 -5.50863385e-01 -4.90610540e-01
3.62726867e-01 -6.07263207e-01 7.19129503e-01 -2.06960633e-01
-7.85965502e-01 -5.95166862e-01 -5.38219333e-01 -4.57625538e-01
-8.57325613e-01 4.78277713e-01 7.26664662e-01 3.66216838e-01
-8.11113060e-01 2.31905162e-01 -6.08887792e-01 -4.81603414e-01
2.28116050e-01 5.20596862e-01 -6.75118089e-01 7.18507841e-02
-1.12280452e+00 8.13783050e-01 2.02515990e-01 2.50834882e-01
-3.03233922e-01 -3.41243297e-01 -1.52101064e+00 -1.18105881e-01
5.41748814e-02 -5.61283112e-01 8.70652795e-01 -5.70278764e-01
-5.87623298e-01 1.47561860e+00 -1.47746205e-01 -4.60328937e-01
6.36623874e-02 -2.30229422e-01 -5.79213142e-01 3.90917838e-01
3.24731052e-01 1.27673531e+00 8.01178277e-01 -1.25426209e+00
-6.81216478e-01 -4.90757614e-01 3.62430841e-01 3.43049288e-01
-3.00377816e-01 3.46695065e-01 -9.17600036e-01 -6.60458505e-01
1.70899883e-01 -9.65567708e-01 5.82264848e-02 3.34354371e-01
-6.86525643e-01 -4.94348943e-01 7.12923765e-01 -9.71163034e-01
1.00006068e+00 -2.29927063e+00 1.45541146e-01 3.19803029e-01
1.63562864e-01 -2.62913734e-01 -3.02929431e-01 -2.83560017e-03
2.10098214e-02 1.80900902e-01 2.49374449e-01 -3.64732414e-01
2.57304907e-01 2.17861667e-01 -3.52747977e-01 4.34930295e-01
3.67146283e-01 1.16690886e+00 -9.34340477e-01 -9.22103107e-01
2.40642145e-01 7.00626969e-01 -2.67745644e-01 -3.31403576e-02
-1.73416764e-01 -7.96677992e-02 -5.16298652e-01 8.73280227e-01
3.22145820e-01 -4.08850908e-01 -4.53476533e-02 -6.10676289e-01
1.80582926e-01 2.30632752e-01 -1.25721335e+00 1.49094141e+00
-1.14576459e-01 7.38149762e-01 -3.52756828e-02 -3.42394203e-01
8.05352092e-01 2.91999012e-01 2.66297132e-01 -3.63683999e-01
-4.55219522e-02 -3.33220273e-01 -5.16774178e-01 -5.90148866e-01
7.97598541e-01 3.25139165e-01 -4.22825426e-01 -5.61934002e-02
4.60655289e-03 -9.13975611e-02 3.85031343e-01 3.82494450e-01
8.41864467e-01 2.35096410e-01 3.40998501e-01 1.02375276e-01
3.94288421e-01 -1.13542013e-01 2.58030355e-01 5.94256699e-01
-1.61453217e-01 7.20545709e-01 4.06273007e-01 -4.51343328e-01
-9.60519910e-01 -1.51080751e+00 -2.16598526e-01 1.26788592e+00
4.68436927e-01 -5.89888453e-01 -1.87269095e-02 -8.85589123e-01
3.50101918e-01 5.52590609e-01 -8.58923316e-01 3.49932849e-01
-4.60779399e-01 -4.00392145e-01 2.64910191e-01 1.04618216e+00
1.97108135e-01 -9.82955635e-01 -4.33982521e-01 -1.05415136e-01
-1.91078171e-01 -1.72721529e+00 -6.73553586e-01 1.56347491e-02
-4.00719553e-01 -1.02168787e+00 -7.30833828e-01 -1.26510620e+00
9.13977325e-01 -1.15008980e-01 1.37121987e+00 4.11774181e-02
-5.81277072e-01 7.80547857e-01 -9.61937904e-02 -1.00585490e-01
2.13419095e-01 -2.48800069e-01 -8.82607978e-03 -5.78702427e-02
5.13891339e-01 3.23940739e-02 -5.22489846e-01 2.69504696e-01
-3.06810141e-01 -1.30476519e-01 2.78634429e-01 4.00318056e-01
9.53282356e-01 -8.54664221e-02 -2.81293929e-01 -5.83468974e-01
2.57001787e-01 -1.61042735e-01 -5.31302392e-01 5.52775979e-01
2.03511745e-01 -1.00522511e-01 -3.56491245e-02 -5.34576833e-01
-6.16598666e-01 4.25893575e-01 2.34038472e-01 -3.68638068e-01
-4.84941751e-01 1.46526068e-01 -7.42811412e-02 9.11627933e-02
3.60697389e-01 -4.04936858e-02 -8.50959718e-01 -1.51022434e-01
6.25234306e-01 4.86520439e-01 7.65048146e-01 -6.76628947e-01
6.71014190e-01 4.94241476e-01 3.95822525e-02 -1.07097971e+00
-9.23484206e-01 -7.68080771e-01 -9.94659424e-01 -3.19973566e-02
1.42308903e+00 -1.06517017e+00 -4.54970568e-01 -1.36813223e-01
-1.25431895e+00 8.47601593e-02 -1.76775783e-01 4.15276587e-01
-2.28036374e-01 2.30004549e-01 -6.96923733e-01 -6.08721673e-01
-3.89188938e-02 -7.69334733e-01 1.64824855e+00 2.33722880e-01
-4.13254708e-01 -8.61160159e-01 -2.52481729e-01 2.88307726e-01
-2.85081148e-01 3.20232928e-01 8.41071844e-01 -6.79724813e-01
-5.14736116e-01 -4.38521743e-01 -6.04067564e-01 8.98361672e-03
3.57728265e-02 1.63269147e-01 -5.22418916e-01 -1.03745321e-02
-1.05498147e+00 -3.88314724e-01 8.37286115e-01 3.81541550e-01
8.53535235e-01 -1.22111365e-01 -7.94873834e-01 2.30822563e-01
1.32598031e+00 -2.51591057e-01 1.96441069e-01 3.50534841e-02
8.31606627e-01 7.00714767e-01 5.68070650e-01 1.51764348e-01
6.27011001e-01 1.02454805e+00 3.28104287e-01 -1.66617423e-01
-5.96844293e-02 -2.55050123e-01 3.39646414e-02 -7.36446753e-02
9.81045514e-02 -3.85066479e-01 -1.02517140e+00 8.53954911e-01
-1.65813398e+00 -6.66296005e-01 -4.33275372e-01 1.76060987e+00
6.54539585e-01 1.82970122e-01 1.55916438e-01 -5.42899787e-01
9.45149958e-01 1.24098673e-01 -1.89694121e-01 -2.22385630e-01
-1.32781163e-01 -1.36079758e-01 6.15882814e-01 3.83170098e-01
-1.77915728e+00 1.21111965e+00 6.78240585e+00 3.01132321e-01
-6.10450566e-01 -1.56675130e-01 4.35436070e-01 -1.97908990e-02
-6.61025196e-02 -2.91244507e-01 -1.23868334e+00 1.28686696e-01
2.64716685e-01 4.15367484e-01 1.25807613e-01 7.10470140e-01
-9.77387577e-02 -1.95786506e-01 -1.35086668e+00 1.12879908e+00
3.83858263e-01 -1.12547648e+00 -1.72934905e-02 -8.34780633e-02
4.66131866e-01 -1.64787337e-01 1.85357422e-01 -8.94220620e-02
3.51746708e-01 -8.69628251e-01 7.77115285e-01 4.69612598e-01
6.00684822e-01 -5.21029770e-01 6.45167470e-01 -2.53822356e-01
-1.48103821e+00 1.60459280e-01 -4.26040202e-01 1.84038967e-01
1.46271095e-01 1.46003425e-01 -9.42230642e-01 1.59447510e-02
8.33856344e-01 6.10026300e-01 -9.20223057e-01 1.07231665e+00
-8.38409483e-01 2.48270154e-01 -5.93963206e-01 -1.99082851e-01
1.33630827e-01 1.35678932e-01 5.41413546e-01 1.36898625e+00
-1.28749952e-01 1.36017799e-01 4.97211009e-01 8.31855834e-01
-2.36473575e-01 2.12824464e-01 -5.28883398e-01 -5.11035584e-02
5.67928672e-01 1.40067852e+00 -1.10139287e+00 -2.78824836e-01
-5.19254565e-01 1.12261021e+00 3.79437000e-01 4.62742686e-01
-1.00808823e+00 -2.85329938e-01 6.14223301e-01 2.92323381e-01
7.81354249e-01 -4.08089489e-01 -1.99119723e-03 -8.95690143e-01
6.35401383e-02 -1.27604097e-01 5.64806044e-01 -1.13967264e+00
-1.38961267e+00 4.05391693e-01 9.74941626e-02 -6.94568515e-01
9.71617475e-02 -8.50613534e-01 -1.88514933e-01 7.74539471e-01
-1.18164003e+00 -1.58685362e+00 -1.60640791e-01 3.48337710e-01
2.33271927e-01 1.51867524e-01 9.66204703e-01 1.99430063e-01
-4.83582467e-01 3.97086382e-01 -3.97185206e-01 8.75936866e-01
7.72165000e-01 -1.52765155e+00 5.25739789e-01 4.84039754e-01
7.50058830e-01 7.89866209e-01 9.12709475e-01 -7.18837857e-01
-8.54439080e-01 -1.04462564e+00 1.21479726e+00 -9.00598645e-01
6.86330557e-01 -7.27262378e-01 -6.57843292e-01 8.39035749e-01
9.02202353e-02 4.01267618e-01 6.21373892e-01 6.03553414e-01
-7.97470093e-01 1.16089202e-01 -1.01257372e+00 4.55224603e-01
7.92622805e-01 -7.11463332e-01 -9.20738161e-01 6.11231685e-01
6.28146529e-01 -5.69298446e-01 -7.30567694e-01 2.18243912e-01
5.07993579e-01 -3.23010534e-01 1.49292493e+00 -7.11244702e-01
3.94720644e-01 -4.52877223e-01 -2.99598128e-01 -8.34984958e-01
-4.52449352e-01 1.09875239e-01 -1.07387640e-01 1.43743789e+00
6.59245491e-01 2.42313147e-01 8.25478256e-01 9.37489927e-01
4.15509284e-01 -4.12379891e-01 -6.94188476e-01 -4.79062885e-01
-4.73689079e-01 -1.47400871e-01 3.03210407e-01 9.39008653e-01
-1.17928416e-01 7.23395169e-01 -1.78966984e-01 6.20594561e-01
5.70321679e-01 3.24088782e-01 4.55289215e-01 -9.68280792e-01
-2.29121298e-01 -1.09622844e-01 -8.33326101e-01 -1.02170503e+00
3.81819367e-01 -6.96681440e-01 -1.06954202e-02 -1.82410669e+00
2.54393786e-01 -2.59374797e-01 -2.20609352e-01 8.13191712e-01
-1.04452096e-01 8.55099857e-01 8.59449729e-02 -2.32659250e-01
-1.08215809e+00 1.14500247e-01 8.19095671e-01 -5.05263627e-01
-2.05461800e-01 -2.59526163e-01 -4.59132105e-01 9.25416350e-01
3.63700241e-01 -3.11226815e-01 3.53535414e-02 -6.38775229e-01
3.41333926e-01 1.71629358e-02 9.00269747e-01 -6.16779208e-01
1.76808998e-01 -1.72044933e-02 9.20171440e-01 -9.21130359e-01
6.88348413e-01 -1.03927445e+00 -2.29942292e-01 -8.43366757e-02
-6.15676343e-01 2.07828730e-01 1.70233160e-01 7.74259329e-01
-2.97230929e-02 -7.35329166e-02 5.05099297e-01 -1.99657768e-01
-1.06703591e+00 1.48980662e-01 6.04049116e-02 5.08542322e-02
1.09743118e+00 -8.77954140e-02 -1.45625532e-01 -1.99400201e-01
-1.33313465e+00 5.17090619e-01 2.55882800e-01 6.87766194e-01
7.38908052e-01 -1.46040463e+00 -5.18018126e-01 -1.04026474e-01
4.18112576e-01 -2.05098003e-01 -5.77015504e-02 4.16601241e-01
-4.15803641e-01 3.03742051e-01 2.16230680e-03 -6.70225441e-01
-1.79135537e+00 6.69344902e-01 1.65395096e-01 1.65339887e-01
-4.69529808e-01 1.45385373e+00 1.95182562e-01 -2.03287467e-01
5.41892827e-01 -3.47731262e-01 -4.51634109e-01 3.44785154e-01
4.61266488e-01 -7.41742030e-02 -3.29773158e-01 -1.27913904e+00
-8.65723848e-01 7.16693401e-01 -1.43252671e-01 -5.49088754e-02
1.19960880e+00 -3.75359744e-01 -1.77652717e-01 3.55871856e-01
1.11936831e+00 2.04874888e-01 -9.42257285e-01 -4.39559996e-01
-6.31965790e-03 -4.10059303e-01 -8.83334354e-02 -7.76837409e-01
-7.14838624e-01 7.32757270e-01 4.85453188e-01 2.43212759e-01
5.48123121e-01 7.38718569e-01 4.77125764e-01 3.51703465e-01
1.54546753e-01 -1.01664615e+00 8.52898955e-02 3.47747713e-01
6.96714878e-01 -1.29467022e+00 2.69689977e-01 -8.32854986e-01
-7.93688416e-01 9.05129373e-01 7.28873610e-01 -1.75953448e-01
5.27138650e-01 1.09328628e-01 -1.53024748e-01 -3.52662444e-01
-3.97945791e-01 -4.64490950e-01 8.91679347e-01 5.99302471e-01
5.17520308e-01 2.12175429e-01 -8.89498144e-02 4.15866703e-01
-1.08238742e-01 -6.30341351e-01 -6.16187043e-02 6.74801648e-01
-3.45128387e-01 -6.44667804e-01 -4.00485039e-01 3.01316381e-01
-4.05283958e-01 -3.25169921e-01 -6.87972188e-01 7.92048573e-01
5.21695256e-01 8.00849974e-01 4.78964835e-01 -7.05283806e-02
3.60765487e-01 -3.32475640e-02 6.30236983e-01 -8.20914865e-01
-4.03621644e-01 1.48556784e-01 3.75923008e-01 -2.49737352e-01
-5.32747388e-01 -8.45189631e-01 -1.45393765e+00 2.90701181e-01
-4.44026530e-01 1.65896207e-01 4.86718476e-01 8.90528560e-01
1.88039780e-01 2.47121528e-01 -8.31020325e-02 -5.06915510e-01
1.15163073e-01 -5.21340132e-01 -6.44090891e-01 5.84541380e-01
2.84943223e-01 -6.04085326e-01 1.25183299e-01 2.75724202e-01] | [10.452692985534668, 1.5000805854797363] |
e9b165bd-a1d0-4e4e-8469-7eb48db68668 | security-and-privacy-preserving-deep-learning | 2006.12698 | null | https://arxiv.org/abs/2006.12698v2 | https://arxiv.org/pdf/2006.12698v2.pdf | Security and Privacy Preserving Deep Learning | Commercial companies that collect user data on a large scale have been the main beneficiaries of this trend since the success of deep learning techniques is directly proportional to the amount of data available for training. Massive data collection required for deep learning presents obvious privacy issues. Users personal, highly sensitive data such as photos and voice recordings are kept indefinitely by the companies that collect it. Users can neither delete it nor restrict the purposes for which it is used. So, data privacy has been a very important concern for governments and companies these days. It gives rise to a very interesting challenge since on the one hand, we are pushing further and further for high-quality models and accessible data, but on the other hand, we need to keep data safe from both intentional and accidental leakage. The more personal the data is it is more restricted it means some of the most important social issues cannot be addressed using machine learning because researchers do not have access to proper training data. But by learning how to machine learning that protects privacy we can make a huge difference in solving many social issues like curing disease etc. Deep neural networks are susceptible to various inference attacks as they remember information about their training data. In this chapter, we introduce differential privacy, which ensures that different kinds of statistical analyses dont compromise privacy and federated learning, training a machine learning model on a data to which we do not have access to. | ['Saisree Miriyala', 'Saichethan Miriyala Reddy'] | 2020-06-23 | null | null | null | null | ['privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['methodology', 'natural-language-processing'] | [-1.31657049e-01 1.30146638e-01 -7.40852803e-02 -6.05818748e-01
-2.96242177e-01 -7.35594690e-01 2.90786624e-01 5.07004499e-01
-7.90528238e-01 9.81948555e-01 -9.55463201e-03 -2.57442623e-01
-1.72981456e-01 -1.24094474e+00 -4.71596450e-01 -7.82141805e-01
6.78418875e-02 2.68973768e-01 -7.18800072e-03 -2.42197558e-01
1.16781276e-02 7.49824286e-01 -1.38330340e+00 2.47567892e-01
6.03151798e-01 1.11098599e+00 -3.81707072e-01 1.43123001e-01
-1.85799748e-01 4.81004119e-01 -5.83778560e-01 -9.64785099e-01
6.57186866e-01 8.58603492e-02 -7.51684248e-01 -5.65091789e-01
9.34845284e-02 -5.06621897e-01 -1.98209271e-01 1.34673083e+00
5.01501262e-01 -1.59441859e-01 7.94152766e-02 -1.45584989e+00
-5.89898884e-01 4.72666979e-01 -5.43239355e-01 -2.16684431e-01
-1.52158916e-01 6.81663677e-02 6.16816342e-01 -5.65979965e-02
3.70817930e-01 8.09629142e-01 6.01601958e-01 7.64804125e-01
-1.00294495e+00 -9.60075378e-01 -2.84418136e-01 5.68196587e-02
-1.14042842e+00 -4.08624351e-01 6.07491851e-01 -3.07062030e-01
3.57332498e-01 8.82764399e-01 4.88759577e-01 1.29755926e+00
1.10782169e-01 4.83653039e-01 1.05191767e+00 -7.76060438e-03
2.50783712e-01 9.67850447e-01 2.73174584e-01 3.08242202e-01
7.02245593e-01 7.16255754e-02 -8.96131545e-02 -2.77626723e-01
2.04550162e-01 4.39859092e-01 -3.40773553e-01 -3.60670328e-01
-6.58842027e-01 9.93653357e-01 2.27209985e-01 5.07925987e-01
-1.98908567e-01 -2.54313588e-01 6.59869790e-01 3.98902714e-01
1.76330209e-01 3.12113941e-01 -6.41110420e-01 4.13086973e-02
-6.63159311e-01 1.76000923e-01 8.91111851e-01 7.16087341e-01
7.29293108e-01 -3.86540115e-01 2.80764878e-01 4.99149382e-01
-4.40493524e-02 3.64747018e-01 7.67706215e-01 -6.23276293e-01
2.74676204e-01 8.03550839e-01 -8.21366012e-02 -1.36879039e+00
-1.71612352e-01 -2.84733325e-01 -1.31888676e+00 2.57515132e-01
7.02793419e-01 -4.41303313e-01 -2.32373521e-01 1.70649517e+00
3.60421121e-01 -5.40565789e-01 2.97957528e-02 6.38337791e-01
5.32704592e-01 5.73764145e-01 -2.16356777e-02 -1.78913161e-01
1.49152875e+00 -1.79557934e-01 -9.68464315e-01 2.71275565e-02
3.98691803e-01 -3.16893846e-01 9.39880371e-01 6.26192391e-01
-6.86536968e-01 -1.19873181e-01 -9.89240408e-01 -3.82639557e-01
-8.61174226e-01 -3.99777025e-01 9.04005408e-01 1.07578099e+00
-6.69994056e-01 8.12153280e-01 -6.91783071e-01 -2.73420870e-01
1.06716597e+00 7.93588817e-01 -9.24427867e-01 1.14480302e-01
-1.52552164e+00 5.65033495e-01 2.93640882e-01 9.69537273e-02
-3.68611515e-01 -4.52124268e-01 -5.01685441e-01 2.14964077e-01
3.72919619e-01 -2.74631888e-01 7.76566803e-01 -8.63424540e-01
-8.93884361e-01 1.11572790e+00 2.88931131e-01 -8.33911240e-01
8.73360157e-01 -1.37566969e-01 -6.23121023e-01 -3.10223460e-01
-2.34250829e-01 1.58066213e-01 7.19867110e-01 -9.51840639e-01
-8.00386727e-01 -9.36888576e-01 -2.65431106e-02 -3.73744339e-01
-1.07399178e+00 8.92061293e-02 7.53916427e-02 -1.87712982e-01
-3.69400233e-01 -8.37859929e-01 -1.57073870e-01 1.54592037e-01
-5.61782658e-01 -8.45550001e-03 1.48381960e+00 -8.34149659e-01
1.22416270e+00 -2.18001437e+00 -4.52967763e-01 2.49291316e-01
5.23184955e-01 8.76408339e-01 4.45970595e-01 3.56375933e-01
2.07032431e-02 7.02832878e-01 -2.51000196e-01 -2.39244983e-01
5.01671210e-02 1.34244606e-01 -2.66042978e-01 5.35990715e-01
-2.47449622e-01 6.06448174e-01 -5.37598670e-01 -2.33071387e-01
1.36781096e-01 6.94104016e-01 -4.07566786e-01 2.10169688e-01
1.39949424e-02 4.07848328e-01 -6.50312841e-01 2.94221818e-01
1.00631177e+00 -5.68841286e-02 1.51522294e-01 4.59055901e-02
-8.17165226e-02 -2.21419886e-01 -1.00071442e+00 1.14743376e+00
-2.75635809e-01 7.05933690e-01 3.54190886e-01 -9.27703321e-01
9.22020376e-01 4.55520272e-01 4.30116385e-01 -7.48812795e-01
4.68410641e-01 8.96138698e-02 -2.10223898e-01 -6.74947321e-01
3.56452286e-01 -1.66562691e-01 -2.26276621e-01 5.52448034e-01
-5.08288324e-01 5.29930711e-01 -3.31708342e-01 1.39278010e-01
9.81961071e-01 -4.36686486e-01 2.33113602e-01 -1.35526374e-01
7.21648574e-01 -4.00420040e-01 8.41369212e-01 2.72499651e-01
-4.30439234e-01 2.19306484e-01 7.24805057e-01 -7.53144801e-01
-9.76868093e-01 -5.97125053e-01 -2.53284842e-01 6.84051394e-01
-3.13655853e-01 -6.90139607e-02 -8.85209739e-01 -8.43395710e-01
2.23234475e-01 5.83871305e-01 -6.17732763e-01 -3.35886776e-01
-3.45312983e-01 -5.62671423e-01 6.32733107e-01 -6.51850775e-02
9.35836315e-01 -1.04097593e+00 -7.13797212e-01 -1.57654703e-01
-1.64600667e-02 -8.57385218e-01 -2.84411758e-01 1.35426342e-01
-6.86823606e-01 -9.67022657e-01 -3.03163916e-01 -3.97992373e-01
5.35753667e-01 4.27110605e-02 6.34912729e-01 1.21257111e-01
-3.53187114e-01 -2.31926844e-01 -3.68286669e-02 -9.88188505e-01
-4.81171966e-01 2.28190467e-01 1.98845752e-02 4.36402202e-01
8.62617493e-01 -8.03758681e-01 -4.37651008e-01 3.46300639e-02
-1.13854170e+00 -4.20681030e-01 1.99872747e-01 3.19353342e-01
1.17650993e-01 6.31174922e-01 6.26972377e-01 -1.51045084e+00
7.83982813e-01 -5.14363050e-01 -7.17443466e-01 1.99394420e-01
-6.24478936e-01 4.33829017e-02 8.51449072e-01 -1.38988361e-01
-8.29408050e-01 3.00656110e-02 -1.37926906e-01 -2.14201137e-01
-3.69012594e-01 -3.30613814e-02 -8.40911269e-01 1.37511790e-01
6.60344422e-01 -1.28367275e-01 2.26351306e-01 -5.93679130e-01
1.42867565e-01 1.19443643e+00 2.29921564e-01 -6.06115460e-02
7.44675934e-01 6.35512650e-01 5.15247546e-02 -9.95422363e-01
-5.11474729e-01 -1.83062851e-01 -4.76037413e-01 2.11305082e-01
1.00963223e+00 -4.33503985e-01 -1.26468217e+00 7.36415207e-01
-9.39719260e-01 3.77100527e-01 -4.45526838e-01 1.65798932e-01
8.18100497e-02 4.48453993e-01 -3.09025884e-01 -1.17423654e+00
-5.85685253e-01 -7.55228579e-01 2.26528138e-01 2.66173303e-01
-3.27265412e-01 -8.70684803e-01 -2.72615880e-01 5.92211843e-01
6.96823597e-01 6.45660758e-01 8.58956277e-01 -1.03365564e+00
-6.60895467e-01 -8.52901459e-01 -1.18192926e-01 7.90064335e-01
4.87607032e-01 -1.51930913e-01 -1.33256221e+00 -3.52594674e-01
6.66724324e-01 -2.46907994e-01 5.33281326e-01 7.82798976e-02
1.62625718e+00 -9.87058163e-01 -1.83839172e-01 6.80719376e-01
1.43151462e+00 1.66708544e-01 9.04952645e-01 2.45394006e-01
7.48569965e-01 1.08150744e+00 1.90565899e-01 4.74814564e-01
1.13884434e-01 2.20713541e-01 6.27601326e-01 -1.04713149e-01
7.30127096e-01 -3.32806170e-01 -8.57737958e-02 1.78617761e-01
1.04516447e-01 -1.62663519e-01 -5.58805943e-01 4.55165774e-01
-1.72817397e+00 -1.20519340e+00 -2.69048154e-01 2.84224868e+00
1.06153548e+00 -5.03205210e-02 1.39250070e-01 2.48513877e-01
6.78721130e-01 -2.27527209e-02 -6.39256239e-01 -5.94382167e-01
2.54865438e-02 7.66304955e-02 9.09258962e-01 1.28931791e-01
-1.02995944e+00 5.39955378e-01 4.30419636e+00 6.51218414e-01
-1.34333181e+00 3.70930880e-02 9.52080429e-01 -2.90258229e-01
-2.96763927e-01 -9.22708362e-02 -7.03325152e-01 7.14420855e-01
1.09064579e+00 -2.99516529e-01 4.42930490e-01 9.91027594e-01
2.53288031e-01 3.67593057e-02 -1.12361574e+00 1.08995390e+00
-5.01635611e-01 -1.23680758e+00 -2.84877926e-01 7.71171689e-01
2.38430396e-01 -2.13864401e-01 1.90692559e-01 -1.80307999e-02
3.30392241e-01 -1.17939854e+00 2.08519042e-01 5.10671616e-01
5.50474405e-01 -1.12576437e+00 7.74305642e-01 7.03456044e-01
-3.33128542e-01 -2.38709912e-01 -5.73683262e-01 1.11953750e-01
-5.15485629e-02 9.05923188e-01 -5.07742584e-01 2.31837094e-01
8.83635581e-01 2.18358144e-01 -4.08256352e-01 7.89437652e-01
-7.79151544e-02 2.89620578e-01 -3.62303138e-01 -1.38425246e-01
-1.04507826e-01 -2.65175164e-01 2.80877292e-01 8.10988843e-01
2.33659610e-01 7.44435191e-03 -5.41429281e-01 8.03741097e-01
-5.12293100e-01 1.76195040e-01 -9.95405078e-01 -4.80435103e-01
4.86861199e-01 1.31799650e+00 -3.55693161e-01 2.50498563e-01
-3.87274265e-01 9.57997501e-01 2.64286935e-01 -2.04053506e-01
-3.69018525e-01 -4.76339579e-01 8.89257848e-01 5.03046632e-01
-1.27605885e-01 5.02263904e-02 -2.77305961e-01 -1.07127428e+00
1.00997560e-01 -1.06261492e+00 5.20084798e-01 -3.99518721e-02
-1.48926616e+00 3.85088503e-01 -4.60323364e-01 -8.57465982e-01
-4.58448753e-02 -3.36121827e-01 -3.03176820e-01 9.53346729e-01
-1.27942455e+00 -8.95604193e-01 -2.50347443e-02 8.32374930e-01
-3.99398535e-01 -2.87637413e-01 1.12343717e+00 5.99303424e-01
-6.10941708e-01 8.27109933e-01 3.24580103e-01 6.47403479e-01
5.28836608e-01 -7.73075342e-01 1.80377156e-01 6.72528863e-01
1.44038692e-01 1.01738298e+00 3.97553146e-01 -4.91338581e-01
-1.37840009e+00 -8.98362577e-01 1.18528187e+00 -6.25284374e-01
3.03146660e-01 -7.41652071e-01 -1.16540289e+00 6.86042786e-01
8.21311027e-02 4.15598117e-02 1.15401363e+00 1.18063539e-01
-2.40096584e-01 -6.41209602e-01 -1.91329861e+00 3.04082364e-01
4.78816152e-01 -5.26492774e-01 -9.88430083e-02 2.59314746e-01
8.81999314e-01 1.83961987e-01 -6.75833344e-01 -1.12798773e-01
5.32633245e-01 -1.36083758e+00 6.35046959e-01 -9.08216298e-01
6.28250986e-02 5.62846176e-02 -7.62254596e-02 -7.64392912e-01
1.00653403e-01 -7.17217326e-01 9.43268239e-02 1.79255331e+00
4.16065276e-01 -9.37843919e-01 1.22223532e+00 1.65229070e+00
6.72736049e-01 -5.52934527e-01 -7.97966719e-01 -6.25176191e-01
1.42415419e-01 -3.05925936e-01 1.18501639e+00 1.43685877e+00
-1.66952446e-01 4.14887369e-02 -7.16922581e-01 8.54338109e-02
8.23117971e-01 -1.58839673e-01 8.18324447e-01 -1.46715736e+00
-2.81488430e-03 6.78136433e-03 -4.01661932e-01 -1.75563231e-01
-1.84199288e-01 -6.55249238e-01 -7.02187240e-01 -1.20400381e+00
5.59203327e-03 -5.60473025e-01 -2.81054109e-01 7.44264066e-01
3.82832289e-01 6.06118189e-03 7.89087117e-02 1.60617486e-01
3.97455208e-02 2.62054116e-01 9.56662714e-01 3.19444500e-02
-1.67968422e-01 4.90747154e-01 -1.35584557e+00 5.52325964e-01
9.84395802e-01 -6.37878776e-01 -4.97909665e-01 -2.23065764e-01
5.07767975e-01 -1.64683074e-01 3.77349347e-01 -1.05024552e+00
2.46133566e-01 -1.61289662e-01 3.99795860e-01 -2.85157233e-01
1.00277737e-01 -1.62500298e+00 3.72393221e-01 3.81959498e-01
-3.04300874e-01 -3.45715880e-01 7.93785825e-02 5.15957713e-01
2.08039545e-02 -2.79109091e-01 9.18086767e-01 -4.42493707e-01
-3.65964592e-01 7.03133404e-01 1.02402650e-01 -6.22441694e-02
1.38626885e+00 -1.78300306e-01 -3.01486075e-01 -5.48101366e-01
-7.53369749e-01 2.87460327e-01 5.68005323e-01 5.34931839e-01
2.85823584e-01 -1.02014625e+00 -4.86109912e-01 4.68798518e-01
-3.31342965e-02 7.06361085e-02 3.62942904e-01 2.99342662e-01
-2.12731391e-01 3.91604662e-01 -3.51434201e-01 -9.73853376e-03
-1.47915089e+00 9.66755748e-01 8.27490538e-02 -1.29011124e-01
-5.40328145e-01 8.51011038e-01 -9.12211463e-02 -3.67255926e-01
4.39990729e-01 -3.88744064e-02 -1.54075697e-01 3.41584831e-01
8.56283903e-01 3.52067113e-01 2.15335563e-01 -4.99129534e-01
-2.97753721e-01 -1.26780793e-01 -3.79245400e-01 2.69851893e-01
1.69908047e+00 -1.37284309e-01 -4.82233733e-01 9.57371444e-02
1.60951149e+00 3.47584873e-01 -9.76194322e-01 -8.57978389e-02
-1.20215304e-01 -7.91939139e-01 1.02220967e-01 -6.97247028e-01
-1.53076386e+00 1.18397117e+00 8.02926779e-01 7.52405941e-01
1.08073807e+00 -4.62965101e-01 1.15045142e+00 2.70031750e-01
5.67960262e-01 -1.30838394e+00 -5.54814935e-01 1.30701512e-01
4.75605905e-01 -1.39066815e+00 1.17481150e-01 -8.31041485e-02
-6.47991955e-01 8.51178050e-01 2.88132340e-01 2.84744740e-01
8.54472637e-01 3.46059561e-01 1.98206622e-02 -7.66165555e-02
-2.64464170e-01 3.94134194e-01 -2.40408346e-01 8.75454068e-01
1.59248263e-01 -7.59389577e-03 -3.96000147e-01 9.05825734e-01
-4.66408670e-01 1.07393630e-01 4.02863562e-01 7.06304967e-01
-3.58703792e-01 -1.57284474e+00 -1.51559159e-01 6.14447892e-01
-9.65913594e-01 1.91817448e-01 -5.70316792e-01 6.19553626e-01
3.09051663e-01 7.32748985e-01 -1.43300772e-01 -4.10497427e-01
2.20280856e-01 7.52997696e-02 -1.64354861e-01 -3.48709077e-01
-7.25919306e-01 -7.39017069e-01 1.73713174e-02 -4.96408373e-01
1.23065487e-01 -7.15472341e-01 -1.27734482e+00 -1.02346623e+00
-1.79960579e-01 2.85307080e-01 1.00676894e+00 5.54067254e-01
4.56016481e-01 -1.61666960e-01 8.23524594e-01 -5.06484360e-02
-7.07315207e-01 -3.34087312e-01 -7.90695727e-01 5.55147946e-01
5.03845632e-01 1.36785254e-01 -3.33552539e-01 -2.81116396e-01] | [6.056069850921631, 6.97682523727417] |
d8d33ef8-7297-4e4b-9aa5-41c6e28095c1 | deep-multi-task-learning-with-low-level-tasks | null | null | https://aclanthology.org/P16-2038 | https://aclanthology.org/P16-2038.pdf | Deep multi-task learning with low level tasks supervised at lower layers | null | ['Anders S{\\o}gaard', 'Yoav Goldberg'] | 2016-08-01 | null | null | null | acl-2016-8 | ['ccg-supertagging'] | ['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] | [-7.404977798461914, 3.7628939151763916] |
e92cb0b4-6b40-4b50-b970-33ec23663edb | deep-quaternion-networks | 1712.04604 | null | http://arxiv.org/abs/1712.04604v3 | http://arxiv.org/pdf/1712.04604v3.pdf | Deep Quaternion Networks | The field of deep learning has seen significant advancement in recent years.
However, much of the existing work has been focused on real-valued numbers.
Recent work has shown that a deep learning system using the complex numbers can
be deeper for a fixed parameter budget compared to its real-valued counterpart.
In this work, we explore the benefits of generalizing one step further into the
hyper-complex numbers, quaternions specifically, and provide the architecture
components needed to build deep quaternion networks. We develop the theoretical
basis by reviewing quaternion convolutions, developing a novel quaternion
weight initialization scheme, and developing novel algorithms for quaternion
batch-normalization. These pieces are tested in a classification model by
end-to-end training on the CIFAR-10 and CIFAR-100 data sets and a segmentation
model by end-to-end training on the KITTI Road Segmentation data set. These
quaternion networks show improved convergence compared to real-valued and
complex-valued networks, especially on the segmentation task, while having
fewer parameters | ['Chase Gaudet', 'Anthony Maida'] | 2017-12-13 | null | null | null | null | ['road-segementation'] | ['computer-vision'] | [-2.83337265e-01 2.72019804e-01 9.91775542e-02 -7.25242972e-01
-3.27797085e-01 -2.25042313e-01 3.80501807e-01 -7.97274336e-02
-1.37713552e+00 6.57409370e-01 -3.70752007e-01 -4.65403557e-01
2.95969188e-01 -6.91649497e-01 -7.32163489e-01 -4.50242460e-01
-5.22517204e-01 4.05834019e-01 9.97513300e-04 -6.07072413e-01
-1.46879954e-02 5.96360564e-01 -9.78126824e-01 -2.11952731e-01
5.33636391e-01 8.26085925e-01 -3.04777324e-01 8.49925518e-01
4.32784736e-01 5.25666595e-01 -8.29349279e-01 -7.46035278e-01
3.24151665e-01 -2.16051415e-01 -7.87415981e-01 -1.85572043e-01
6.36049211e-01 -3.79111052e-01 -2.07668856e-01 1.08614600e+00
5.96600592e-01 5.23407869e-02 5.22083044e-01 -1.38761032e+00
-4.18408394e-01 6.53093517e-01 -3.55170399e-01 3.84421825e-01
-5.45305431e-01 -1.99088916e-01 1.18650627e+00 -8.64839196e-01
3.00428659e-01 1.31612182e+00 1.14306080e+00 2.66928464e-01
-9.35099483e-01 -6.85389161e-01 9.57723632e-02 2.85733283e-01
-1.37906492e+00 1.89162135e-01 4.46738333e-01 -1.28399640e-01
1.47675061e+00 -2.30668858e-01 9.92934763e-01 6.59428000e-01
2.31819600e-01 9.98326302e-01 7.49798238e-01 -3.93525660e-01
2.40589291e-01 -1.36843577e-01 1.77997366e-01 8.46795857e-01
4.02526110e-01 -1.17209941e-01 -1.07375756e-01 4.52557415e-01
1.21484256e+00 -4.02281970e-01 1.96360145e-02 -2.15164199e-01
-1.63569105e+00 1.03430164e+00 8.70184839e-01 2.00816721e-01
-4.54579353e-01 6.19939506e-01 4.01569575e-01 3.50997418e-01
6.35113895e-01 7.51405716e-01 -5.18156826e-01 -2.29117587e-01
-9.90400314e-01 2.31496498e-01 1.03783834e+00 7.62139797e-01
8.49045157e-01 4.84752685e-01 4.28815961e-01 5.43528020e-01
2.77610064e-01 6.56438708e-01 6.22853756e-01 -8.25947285e-01
5.56277573e-01 1.13519467e-01 -7.64616281e-02 -7.86034942e-01
-9.83523846e-01 -8.02626252e-01 -1.23367536e+00 3.00927460e-01
4.96426046e-01 -5.87442458e-01 -1.12577033e+00 1.69898903e+00
8.89792219e-02 2.52431750e-01 8.16235021e-02 1.10424161e+00
5.25099874e-01 6.10712826e-01 8.37241709e-02 4.83372062e-01
1.19400084e+00 -9.72088873e-01 -5.67846477e-01 -1.59156188e-01
8.89616132e-01 -6.04774237e-01 8.70183110e-01 6.62191987e-01
-1.09151614e+00 -6.09198809e-01 -1.81756115e+00 -3.37827504e-01
-6.71461999e-01 2.46694952e-01 9.53869939e-01 9.71233845e-01
-1.21417046e+00 1.08929050e+00 -1.35591996e+00 -2.43222639e-01
3.90346110e-01 6.57491028e-01 -1.75857067e-01 5.03557026e-01
-1.56646466e+00 1.24520957e+00 5.89639962e-01 4.39818680e-01
-3.02544147e-01 -6.05171084e-01 -9.55960453e-01 9.91595909e-02
-3.68739933e-01 -6.10634744e-01 1.64566517e+00 -1.10586882e+00
-1.64857793e+00 7.22094774e-01 4.66652066e-01 -1.04947388e+00
5.35632253e-01 -2.81035215e-01 -1.64893135e-01 2.33464211e-01
-1.92401290e-01 1.24140859e+00 7.80338049e-01 -6.33442760e-01
-7.54455566e-01 -3.38632539e-02 3.29685390e-01 2.90724486e-01
-3.75955760e-01 -1.99260652e-01 -2.50434309e-01 -8.43359768e-01
2.93701649e-01 -1.10755002e+00 -3.58785003e-01 -1.52693093e-01
-2.00182796e-01 -3.02883774e-01 4.45371829e-02 -4.68317509e-01
7.51491487e-01 -2.09100080e+00 1.97681323e-01 1.76603854e-01
2.28251591e-01 5.91010034e-01 -2.11139843e-01 1.87206343e-01
-5.59117973e-01 2.55334407e-01 -3.72906253e-02 -6.18848026e-01
2.47814432e-01 3.49633455e-01 -3.26557338e-01 9.42208171e-01
7.16307580e-01 1.00574791e+00 -8.85642469e-01 -1.78351581e-01
8.65022019e-02 6.67148530e-01 -5.48837185e-01 -1.93568885e-01
1.90759286e-01 -3.45162481e-01 3.89647367e-03 2.62224406e-01
8.96974266e-01 -3.19074064e-01 -2.76623722e-02 -4.65353847e-01
-1.52012780e-01 1.84577242e-01 -1.23360562e+00 1.39183557e+00
-4.56234992e-01 9.98134077e-01 -3.85602005e-03 -1.10534263e+00
5.05706847e-01 1.14495650e-01 3.79871637e-01 -8.47884476e-01
4.08941627e-01 1.84264660e-01 3.50576341e-01 1.07664555e-01
6.58866942e-01 -4.92564142e-01 3.55204083e-02 4.78835553e-01
4.08605695e-01 -3.76692981e-01 4.75198686e-01 5.12693375e-02
5.77421963e-01 5.87664358e-02 1.76651746e-01 -2.16374509e-02
1.64569020e-01 4.03926522e-02 1.66461572e-01 4.42080528e-01
-2.75500447e-01 8.12351763e-01 4.70356911e-01 -4.67653900e-01
-1.05131376e+00 -1.18693602e+00 -1.71302795e-01 1.08155620e+00
1.00992054e-01 -3.04837853e-01 -8.63526583e-01 -9.22218189e-02
-3.17684412e-02 5.93777180e-01 -7.48760760e-01 -3.79645489e-02
-7.93336451e-01 -1.21596718e+00 1.07005584e+00 9.33604956e-01
8.76801133e-01 -8.73889089e-01 -9.24292088e-01 3.44106883e-01
1.78865209e-01 -1.17850029e+00 -6.99491948e-02 5.29724061e-01
-9.63922918e-01 -8.00662279e-01 -1.05892539e+00 -8.38411510e-01
5.01839519e-01 8.67982432e-02 1.23902392e+00 -1.07416652e-01
-3.86246741e-02 3.44023466e-01 -1.96352184e-01 -8.29759240e-01
-1.27465412e-01 4.68742371e-01 -1.85306724e-02 -3.51393461e-01
-1.67997309e-03 -2.14524165e-01 -5.90541303e-01 1.39048815e-01
-9.52298522e-01 2.05806702e-01 7.71465480e-01 9.63530660e-01
3.85624528e-01 -1.46948412e-01 6.21306300e-01 -5.38174033e-01
6.78586662e-01 -9.23275054e-02 -5.11011243e-01 -1.14690006e-01
-4.59510326e-01 1.50385454e-01 5.54366171e-01 -3.49875659e-01
-7.65098810e-01 -1.93626150e-01 -5.37279487e-01 -6.44661039e-02
3.37262630e-01 9.31509793e-01 3.58186394e-01 -2.46389791e-01
6.93852723e-01 -4.92555678e-01 4.01601056e-03 -1.62183285e-01
8.05760145e-01 3.55623662e-01 6.13940597e-01 -3.87695253e-01
7.88323641e-01 6.28208637e-01 1.04672506e-01 -9.24431264e-01
-7.09411919e-01 -4.68538702e-01 -1.01292670e+00 1.14508010e-01
9.09589231e-01 -1.20211220e+00 -6.57495201e-01 9.10008550e-01
-1.14288628e+00 -5.38834393e-01 -3.94702405e-01 9.38022971e-01
-4.57487911e-01 1.96586609e-01 -7.20626950e-01 -3.31151426e-01
-3.07133585e-01 -1.11755562e+00 1.06219614e+00 3.54277641e-01
-1.30830333e-01 -1.28341365e+00 6.98360987e-03 -2.85140336e-01
3.27575564e-01 -7.25332182e-03 7.35761940e-01 -5.12532413e-01
-4.69008893e-01 -4.79839891e-01 -4.17028248e-01 9.51433837e-01
-4.65960979e-01 5.36755398e-02 -8.44369709e-01 -3.61881316e-01
-1.81053177e-01 -5.93480289e-01 1.03816056e+00 2.52983034e-01
5.74355364e-01 1.41409174e-01 2.04288512e-01 9.68271554e-01
1.33658552e+00 -9.23734978e-02 6.24864995e-01 6.50160849e-01
7.10613191e-01 1.20213307e-01 5.40895522e-01 3.16290468e-01
8.22096169e-01 2.59342045e-01 5.30746698e-01 -8.34267616e-01
3.02559789e-02 2.09467545e-01 9.59028900e-02 9.79818821e-01
-3.38368714e-01 -8.36409405e-02 -9.37525332e-01 5.35407364e-01
-1.62967217e+00 -7.66200364e-01 1.17584005e-01 1.86817002e+00
9.52565253e-01 3.83172601e-01 1.67542621e-01 4.30052012e-01
2.82888472e-01 1.33685619e-01 -5.12830138e-01 -6.41147494e-01
-1.22534335e-01 8.85549843e-01 9.51389670e-01 2.95448244e-01
-1.52362967e+00 1.08719277e+00 6.27878952e+00 5.30900836e-01
-1.78546524e+00 -1.54497281e-01 5.74579537e-01 3.02749187e-01
1.21966630e-01 -3.96808267e-01 -7.06104815e-01 -3.64009857e-01
1.13698113e+00 4.38647084e-02 3.03585768e-01 8.35011780e-01
-7.85712004e-02 4.57368307e-02 -1.23330295e+00 1.14218795e+00
1.29099429e-01 -1.27452397e+00 -1.52667910e-01 -2.11662591e-01
4.77879971e-01 7.48751879e-01 3.05335283e-01 7.38701522e-01
2.08673462e-01 -1.40230465e+00 9.50289071e-01 2.55257308e-01
6.05664432e-01 -8.20987821e-01 9.13959980e-01 1.18697584e-01
-8.69329035e-01 2.67541051e-01 -5.89389801e-01 -3.86189789e-01
5.98588735e-02 2.02734321e-01 -1.28772151e+00 4.73407626e-01
6.01842284e-01 9.03624594e-01 -8.60331655e-01 1.21718514e+00
-4.05356318e-01 9.82922852e-01 -6.08725965e-01 -2.22355723e-01
7.59020746e-01 -3.63015860e-01 -1.10967040e-01 1.40161133e+00
8.92531220e-03 -1.20764852e-01 -2.78891951e-01 5.52123189e-01
-1.90963447e-01 -1.29038483e-01 1.81267232e-01 -2.29967430e-01
-4.78842854e-02 1.35950208e+00 -1.08189738e+00 -3.83281440e-01
-4.60251868e-01 8.20448279e-01 3.12702149e-01 6.67613029e-01
-1.22341597e+00 -1.01337504e+00 6.00586832e-01 -4.27891582e-01
4.20054823e-01 -8.31720293e-01 -4.96525377e-01 -1.24242020e+00
-3.33933055e-01 -7.02057600e-01 -5.15852384e-02 -7.33548522e-01
-9.96430635e-01 5.88190556e-01 1.13156535e-01 -1.07016945e+00
-4.38651890e-01 -1.21148992e+00 -3.81630063e-01 1.13503802e+00
-1.78917217e+00 -1.01904094e+00 -9.65769514e-02 2.17442855e-01
-1.63941294e-01 9.75237936e-02 7.16686726e-01 4.97211486e-01
-5.51256239e-01 7.83253789e-01 3.78985442e-02 5.89225829e-01
6.11270607e-01 -1.64471209e+00 1.12136412e+00 7.72326708e-01
3.31524521e-01 8.52016985e-01 4.99357790e-01 -7.42966607e-02
-1.20115530e+00 -9.26895142e-01 8.03620994e-01 8.15795809e-02
7.70685315e-01 -3.91478479e-01 -7.42320538e-01 8.60914528e-01
3.45255375e-01 -4.41255607e-02 4.21333313e-01 1.07371412e-01
-3.06941330e-01 -1.06063470e-01 -8.61500442e-01 7.79899657e-01
6.03138149e-01 -3.46599311e-01 -5.99211156e-01 3.81423384e-01
6.01927161e-01 -8.03112745e-01 -9.07504499e-01 3.32949162e-01
7.20556855e-01 -7.39126980e-01 1.26046836e+00 -8.16303790e-01
1.76257845e-02 -2.77374953e-01 5.05867414e-02 -1.65875566e+00
-2.01437056e-01 -5.56885660e-01 4.38448340e-01 6.20672584e-01
3.21270317e-01 -7.15034723e-01 8.84855449e-01 2.03342587e-01
-2.72388965e-01 -9.10333574e-01 -1.06681883e+00 -6.48436248e-01
6.42576098e-01 -6.51668429e-01 3.35166037e-01 7.44002998e-01
-3.64704758e-01 5.63160658e-01 2.97128204e-02 5.50496504e-02
3.24505717e-01 -2.74657488e-01 8.18172872e-01 -1.06274724e+00
9.05701052e-03 -6.64775133e-01 -8.63885462e-01 -1.17791069e+00
2.53586024e-01 -9.00929332e-01 -1.15927376e-01 -1.50353384e+00
-5.93031108e-01 -3.36901188e-01 -1.55651271e-01 5.59392095e-01
-9.89953503e-02 5.61712265e-01 4.22524512e-01 -1.66383281e-01
-5.52167654e-01 6.26286030e-01 1.37434459e+00 -2.10807055e-01
1.73107520e-01 8.98009818e-03 -4.00708765e-01 9.30228949e-01
9.41922545e-01 -1.14928141e-01 -5.16472995e-01 -5.78970790e-01
7.76675820e-01 -4.15777028e-01 5.55489004e-01 -1.30894148e+00
1.04541734e-01 3.00901085e-01 6.74000442e-01 -7.81035662e-01
5.34863293e-01 -5.06665587e-01 -5.85245013e-01 4.87176836e-01
-8.75397995e-02 4.22538161e-01 3.89903307e-01 2.65516222e-01
-4.30172592e-01 -2.37245798e-01 1.05938447e+00 3.98368575e-02
-9.24377322e-01 1.68692619e-01 -3.74591768e-01 3.73415530e-01
6.78762019e-01 -2.18772262e-01 -2.96740890e-01 -3.74221027e-01
-8.00539911e-01 1.58962861e-01 5.81855848e-02 1.38982937e-01
5.18369019e-01 -1.06878150e+00 -4.76438493e-01 2.79024214e-01
-3.30975801e-01 2.32757092e-01 -3.18461359e-01 7.50188887e-01
-1.24807370e+00 5.92202961e-01 -5.13487458e-01 -6.76334858e-01
-8.00861239e-01 -1.18062258e-01 8.02054524e-01 -1.53337240e-01
-2.12021396e-01 1.27046978e+00 -2.04505995e-01 -4.87814426e-01
4.92405266e-01 -1.30153489e+00 -2.59747505e-01 3.04739892e-01
2.93969512e-02 3.31979632e-01 4.87591505e-01 -6.82787180e-01
-4.03964132e-01 4.87048507e-01 3.21739316e-02 -5.36292732e-01
1.47671151e+00 2.60900348e-01 -3.72378044e-02 3.47093612e-01
1.41972578e+00 -5.62666714e-01 -1.16276431e+00 2.67209727e-02
-8.38279724e-02 2.71638840e-01 1.49687812e-01 -6.46980882e-01
-1.07179928e+00 1.42414689e+00 6.03667855e-01 3.72529812e-02
9.25462306e-01 -5.57121158e-01 7.03500867e-01 1.00786924e+00
1.48030713e-01 -1.09219956e+00 1.05031967e-01 1.14258134e+00
8.36809039e-01 -1.11240315e+00 1.89337417e-01 -1.40582412e-01
-6.23833537e-01 1.49719262e+00 3.63896459e-01 -5.46711385e-01
1.03004742e+00 1.16389588e-01 3.76296490e-01 1.45925861e-03
-2.83080727e-01 -3.84206563e-01 3.03355008e-01 5.88896990e-01
7.10073948e-01 4.37542796e-03 -1.44582698e-02 4.67998087e-01
-7.85852730e-01 4.71127704e-02 6.88925147e-01 8.68043303e-01
-2.67569691e-01 -8.58013988e-01 -2.12356851e-01 4.77384701e-02
-4.39543188e-01 -3.29850048e-01 -1.39091313e-01 1.41202223e+00
1.65250689e-01 6.12877727e-01 4.02215987e-01 -1.49614081e-01
5.59349179e-01 -1.85446456e-01 7.85689592e-01 -5.82631230e-01
-6.86331868e-01 -1.55081168e-01 5.75692505e-02 -3.02732825e-01
-7.19151914e-01 -4.37901288e-01 -1.86248589e+00 -1.67845607e-01
-1.46466404e-01 9.16726701e-03 9.79295969e-01 8.77039909e-01
-1.61296397e-01 7.31115997e-01 2.29293033e-01 -1.26320601e+00
-1.09119117e+00 -1.08259869e+00 -5.54935396e-01 1.49064764e-01
4.95212853e-01 -5.42609930e-01 -2.06432626e-01 3.67882140e-02] | [8.859745025634766, 2.765063524246216] |
3e224ad0-18f0-4ea9-83e5-186dc6defee3 | cascaded-classification-models-combining | null | null | http://papers.nips.cc/paper/3472-cascaded-classification-models-combining-models-for-holistic-scene-understanding | http://papers.nips.cc/paper/3472-cascaded-classification-models-combining-models-for-holistic-scene-understanding.pdf | Cascaded Classification Models: Combining Models for Holistic Scene Understanding | One of the original goals of computer vision was to fully understand a natural scene. This requires solving several problems simultaneously, including object detection, labeling of meaningful regions, and 3d reconstruction. While great progress has been made in tackling each of these problems in isolation, only recently have researchers again been considering the difficult task of assembling various methods to the mutual benefit of all. We consider learning a set of such classification models in such a way that they both solve their own problem and help each other. We develop a framework known as Cascaded Classification Models (CCM), where repeated instantiations of these classifiers are coupled by their input/output variables in a cascade that improves performance at each level. Our method requires only a limited âblack boxâ interface with the models, allowing us to use very sophisticated, state-of-the-art classifiers without having to look under the hood. We demonstrate the effectiveness of our method on a large set of natural images by combining the subtasks of scene categorization, object detection, multiclass image segmentation, and 3d scene reconstruction. | ['Geremy Heitz', 'Daphne Koller', 'Stephen Gould', 'Ashutosh Saxena'] | 2008-12-01 | null | null | null | neurips-2008-12 | ['3d-scene-reconstruction'] | ['computer-vision'] | [ 4.64833617e-01 -1.35592185e-02 -9.24690813e-02 -6.21910453e-01
-5.73796213e-01 -7.37971663e-01 7.38642216e-01 1.69366658e-01
-4.31588948e-01 2.18394682e-01 -3.27472359e-01 -4.52449381e-01
4.70496006e-02 -6.17397606e-01 -5.69343448e-01 -6.45497680e-01
1.65679961e-01 5.63907743e-01 6.35571897e-01 -2.21697241e-02
2.30532393e-01 7.29056358e-01 -1.80269337e+00 4.48198140e-01
6.13991439e-01 9.62041378e-01 3.06312710e-01 9.08393383e-01
-2.39315033e-01 1.06653929e+00 -3.25434595e-01 -2.65689135e-01
2.41340458e-01 -4.01709467e-01 -1.15627706e+00 7.35097408e-01
5.11168361e-01 -2.08420932e-01 2.42543414e-01 9.61690664e-01
8.38201344e-02 1.18520983e-01 6.96679890e-01 -1.29261553e+00
-2.47729957e-01 2.21644312e-01 -7.34264731e-01 1.53066725e-01
1.84123442e-01 -4.86168042e-02 1.04512739e+00 -7.06840396e-01
3.11951637e-01 1.25387871e+00 5.81976771e-01 3.83783311e-01
-1.56771982e+00 -2.46598065e-01 5.44172764e-01 6.65012524e-02
-1.04989421e+00 -5.19293308e-01 6.60239100e-01 -6.84219658e-01
1.09683156e+00 3.11219752e-01 5.88377714e-01 4.54244733e-01
-1.81467041e-01 1.02304673e+00 1.11746979e+00 -5.94763219e-01
8.63948241e-02 3.81859869e-01 4.95764315e-01 7.52746224e-01
1.24437297e-02 -1.63312688e-01 -7.04382733e-02 1.21290147e-01
6.85057580e-01 7.87720010e-02 5.96488900e-02 -5.59699118e-01
-9.50581729e-01 7.17700362e-01 6.55818939e-01 2.86125571e-01
-8.42221156e-02 9.57645383e-03 1.35265529e-01 2.31200010e-01
5.29901624e-01 4.02386159e-01 -3.66484940e-01 5.41443408e-01
-7.13073373e-01 2.62787580e-01 6.58535779e-01 8.45349133e-01
7.87487388e-01 -3.45226109e-01 3.30651760e-01 8.67729843e-01
2.14094907e-01 -8.42247382e-02 2.48705909e-01 -1.01971877e+00
1.29786298e-01 8.58959734e-01 1.24475420e-01 -7.76322305e-01
-5.88200748e-01 -4.04685348e-01 -5.35288215e-01 5.74036300e-01
7.20210493e-01 8.10399503e-02 -1.11954999e+00 1.60563183e+00
5.68753481e-01 7.13780224e-02 -9.94383842e-02 7.76409805e-01
6.48455381e-01 6.88694239e-01 2.01457232e-01 1.70780689e-01
1.32802010e+00 -1.17019320e+00 -2.03636438e-01 -5.93895435e-01
5.73005140e-01 -8.19234848e-01 7.74242818e-01 4.59819466e-01
-1.15756106e+00 -7.94968545e-01 -9.21018243e-01 -3.54901880e-01
-4.38337803e-01 -6.47583324e-03 8.19132984e-01 3.91315579e-01
-1.00238109e+00 4.53303725e-01 -8.57964158e-01 -4.73529905e-01
5.03347337e-01 3.99060637e-01 -3.78383696e-01 -1.56576023e-01
-4.30791169e-01 1.18282819e+00 2.17936367e-01 1.96780831e-01
-8.98680925e-01 -2.64780253e-01 -7.26519406e-01 -2.38621458e-01
4.10311103e-01 -7.44901597e-01 1.52900767e+00 -1.20201039e+00
-9.63260055e-01 1.23664320e+00 -1.72066063e-01 -1.98413968e-01
4.71802384e-01 -2.71838337e-01 5.92735708e-02 5.60277253e-02
8.76365080e-02 8.84802639e-01 8.65768194e-01 -1.57236135e+00
-1.01319087e+00 -6.06419981e-01 4.67016369e-01 4.38038141e-01
-7.25329891e-02 1.94476306e-01 -5.25286734e-01 -3.60505402e-01
4.05488253e-01 -6.90300226e-01 -5.66506803e-01 2.29099333e-01
-3.73193264e-01 -3.57601970e-01 8.28304112e-01 -2.99627900e-01
6.68343961e-01 -2.15878892e+00 4.79453623e-01 -1.36324644e-01
3.41427922e-01 1.13668539e-01 -1.35562837e-01 1.32110581e-01
-1.19056486e-01 1.52521461e-01 -4.59488153e-01 -5.27431548e-01
-2.82626778e-01 2.88257837e-01 -2.86775857e-01 5.35443306e-01
3.86350006e-01 8.15661371e-01 -9.48639691e-01 -5.96958280e-01
4.59043473e-01 2.62312293e-01 -5.22928596e-01 2.94403911e-01
-3.98370147e-01 4.27552432e-01 -4.27676678e-01 5.56585729e-01
3.40219557e-01 -3.70342016e-01 1.32625669e-01 -1.83893200e-02
-1.74526200e-01 1.00006595e-01 -1.35971236e+00 1.57000315e+00
-2.70798445e-01 4.97631818e-01 4.79412109e-01 -1.48487532e+00
6.04294717e-01 4.36544195e-02 5.08361101e-01 -3.05902362e-01
2.79025555e-01 1.51562423e-01 -3.29185389e-02 -5.96388876e-01
2.46983647e-01 -2.91287392e-01 2.15529073e-02 5.94664514e-01
1.65264711e-01 -5.15644968e-01 1.72628626e-01 1.38711184e-01
7.65788615e-01 3.34081471e-01 3.75691086e-01 -6.73879832e-02
3.77516836e-01 3.72800618e-01 2.83998400e-01 9.34395850e-01
-2.94539601e-01 5.41095018e-01 3.17405581e-01 -5.19993961e-01
-9.34034050e-01 -8.21519673e-01 -2.55106866e-01 1.28094256e+00
2.70600855e-01 1.15234829e-01 -8.09797525e-01 -7.05199063e-01
1.21255182e-01 4.02908385e-01 -6.23391986e-01 2.39039600e-01
-5.78970015e-01 -9.13290858e-01 1.67488188e-01 5.73567748e-01
5.11150241e-01 -9.06394660e-01 -9.81858492e-01 1.18213058e-01
5.81089742e-02 -1.26556599e+00 2.30080299e-02 6.22994423e-01
-9.77048934e-01 -1.24283719e+00 -4.53339159e-01 -1.19484818e+00
8.28777134e-01 7.54824877e-01 1.15922010e+00 2.51046181e-01
-5.25751472e-01 4.99242157e-01 -2.91849017e-01 -5.23524761e-01
-3.79096180e-01 -4.53355238e-02 -2.45165408e-01 1.00683965e-01
2.45025501e-01 -5.18410325e-01 -1.79714605e-01 3.84981722e-01
-9.88283873e-01 3.12715918e-01 5.85702002e-01 5.45787811e-01
4.88392353e-01 1.71666816e-01 2.97573507e-01 -9.71092701e-01
2.34480858e-01 -3.05280209e-01 -6.63842380e-01 4.12618637e-01
-1.45679757e-01 -2.94624478e-01 3.61741930e-01 -3.33919257e-01
-8.87632787e-01 4.78085428e-01 -1.21312685e-01 -2.90430754e-01
-7.45514691e-01 2.06553355e-01 -2.67683983e-01 -1.86204478e-01
6.51816487e-01 1.12846918e-01 -6.24845661e-02 -5.48443317e-01
7.18687952e-01 6.68545723e-01 7.10909665e-01 -4.74972516e-01
7.40863800e-01 4.68848318e-01 1.41724199e-02 -9.98416126e-01
-1.24078286e+00 -6.70866430e-01 -1.15472984e+00 -3.15402657e-01
1.02251410e+00 -8.85255933e-01 -4.73425955e-01 6.75106347e-01
-1.16034198e+00 -3.71706367e-01 -2.34979808e-01 1.12953380e-01
-5.12486339e-01 3.02669972e-01 -5.09519815e-01 -7.70924211e-01
2.18784988e-01 -1.14499879e+00 1.04264605e+00 1.55659243e-01
-1.17348753e-01 -9.22371387e-01 -2.02823460e-01 5.69561064e-01
1.00668952e-01 3.15376401e-01 1.12889826e+00 -6.22671187e-01
-5.61036468e-01 -3.77911896e-01 -3.67157817e-01 4.79392231e-01
1.52385801e-01 -1.46097451e-01 -1.25283909e+00 -3.43091227e-02
2.78628290e-01 -5.88340163e-01 1.05892837e+00 3.82033497e-01
1.25935078e+00 1.07943915e-01 -4.03188020e-01 4.70011055e-01
1.45206690e+00 2.58681864e-01 1.55333281e-01 9.81191471e-02
6.78725302e-01 9.83093023e-01 3.40501219e-01 -1.93218008e-01
4.27292436e-01 6.01154089e-01 5.14474571e-01 -5.71812153e-01
-1.92836478e-01 6.56423867e-02 -9.58917439e-02 2.95709580e-01
1.41022913e-02 -1.25033081e-01 -1.10592973e+00 4.68867958e-01
-1.97251666e+00 -8.36656749e-01 -2.62267530e-01 1.88046515e+00
5.48031271e-01 4.28416103e-01 3.09916615e-01 3.42223614e-01
6.35485649e-01 5.10111637e-02 -6.52715325e-01 -2.72298634e-01
1.16233177e-01 -1.94940940e-01 1.74942389e-01 7.55511880e-01
-1.56285524e+00 9.68967915e-01 6.99132442e+00 4.29746717e-01
-8.73516619e-01 -2.26787090e-01 8.76021624e-01 1.63786262e-01
1.02842040e-02 1.55954316e-01 -7.56358504e-01 -9.00338590e-02
3.82331699e-01 4.11563337e-01 5.22385120e-01 9.12368357e-01
-6.86712414e-02 -5.93219936e-01 -1.28285015e+00 7.44588614e-01
1.77300394e-01 -1.08737886e+00 -1.62656214e-02 -1.80762842e-01
5.00333667e-01 -1.86313898e-03 -2.78608590e-01 1.99200343e-02
5.95600724e-01 -9.54159915e-01 9.22408164e-01 1.71049669e-01
4.47138637e-01 -3.12959373e-01 2.86919683e-01 7.54690170e-01
-1.17081964e+00 -3.57870281e-01 -1.31678758e-02 -3.04022223e-01
9.24612135e-02 4.52301174e-01 -5.07901549e-01 3.31828147e-01
5.97894967e-01 7.20082462e-01 -6.28280580e-01 1.26057374e+00
-3.12030524e-01 2.56402254e-01 -3.16549838e-01 5.37967496e-02
2.69283921e-01 -1.18051805e-01 3.06464553e-01 1.14200175e+00
-2.74547398e-01 3.08138013e-01 5.27634680e-01 7.78720021e-01
1.89943179e-01 -1.92762733e-01 -5.54802299e-01 1.37622505e-01
6.34393096e-02 1.45338631e+00 -1.29782093e+00 -4.49218333e-01
-5.30633569e-01 7.66921163e-01 4.73799497e-01 2.99992114e-01
-6.27869368e-01 -2.95576453e-01 4.85181034e-01 -8.83753935e-04
2.62630522e-01 -4.66367751e-01 -6.50894105e-01 -1.18007731e+00
-1.21530369e-01 -7.47463048e-01 4.41165119e-01 -7.42084563e-01
-1.26518965e+00 5.62353671e-01 9.31306854e-02 -8.27225626e-01
-5.11988858e-03 -7.69326687e-01 -4.03355539e-01 8.34019125e-01
-1.60699379e+00 -1.21989703e+00 -3.74520361e-01 6.91827118e-01
8.59241426e-01 3.19988400e-01 8.07723761e-01 2.70875037e-01
-6.85857654e-01 -2.40825742e-01 -3.58847857e-01 1.97296962e-01
3.29804808e-01 -1.37627673e+00 3.05310994e-01 9.48615730e-01
4.27400500e-01 2.70849317e-01 6.19321465e-01 -2.74057448e-01
-1.07363725e+00 -8.82994890e-01 8.72985125e-01 -8.26504290e-01
4.76422071e-01 -7.02340901e-01 -8.11851382e-01 7.10561037e-01
-8.55242088e-02 -3.19920294e-02 5.20955026e-01 2.41114184e-01
-5.03719270e-01 -1.33318007e-01 -8.59139442e-01 3.44714046e-01
9.79149044e-01 -5.10581911e-01 -7.36239791e-01 4.55125064e-01
4.74043697e-01 -2.97660261e-01 -3.45490515e-01 3.79078746e-01
4.45784122e-01 -1.09090185e+00 1.08370101e+00 -8.66822958e-01
4.41623718e-01 -5.37859321e-01 -2.72850960e-01 -8.77563179e-01
-2.46167272e-01 -3.56432162e-02 2.01234743e-01 9.18699145e-01
4.56169188e-01 -3.75123650e-01 6.79648399e-01 7.34533191e-01
-3.07817668e-01 -6.64763272e-01 -5.58661878e-01 -4.99562591e-01
7.31529593e-02 -6.28089011e-01 2.49387801e-01 8.36004257e-01
-2.02715889e-01 7.13080645e-01 -1.33647293e-01 2.33324394e-01
6.68931127e-01 6.62542820e-01 9.19982612e-01 -1.42692184e+00
-3.86933625e-01 -6.70403659e-01 -2.87974745e-01 -1.22678304e+00
-5.52128740e-02 -9.66775775e-01 2.90898919e-01 -1.78379321e+00
4.97369796e-01 -6.37714386e-01 -1.51238501e-01 8.20177853e-01
-2.02750280e-01 3.46163958e-01 2.83032149e-01 4.49734390e-01
-7.05121636e-01 -2.95658726e-02 1.08896601e+00 -2.15248287e-01
-2.26414353e-01 1.27465338e-01 -9.33173776e-01 1.05113590e+00
5.52653134e-01 -3.48280370e-01 -4.76683319e-01 -7.99736917e-01
-1.91221550e-01 4.44281427e-03 6.71518981e-01 -1.01019585e+00
3.31188709e-01 -2.36511931e-01 6.42639041e-01 -4.04162019e-01
4.31901008e-01 -8.76067996e-01 -3.99785563e-02 3.24247777e-01
-4.31686282e-01 -8.19437578e-02 2.51024276e-01 4.03164655e-01
-1.98295951e-01 -1.03928685e-01 1.12477601e+00 -5.26912212e-01
-9.92332578e-01 1.20448377e-02 -3.66300017e-01 -2.54623622e-01
1.33043563e+00 -3.50875437e-01 -6.21839277e-02 -1.79000154e-01
-1.00469196e+00 2.85371184e-01 3.72525364e-01 4.37710911e-01
3.23242784e-01 -7.01984107e-01 -5.23424208e-01 2.94496328e-01
-9.84204933e-02 3.92796248e-01 -4.81113456e-02 6.40342653e-01
-3.64708304e-01 3.15555692e-01 -8.78735408e-02 -1.01987684e+00
-1.22842157e+00 6.56906545e-01 6.35821521e-01 -1.20628953e-01
-4.65871513e-01 1.13262415e+00 3.26191962e-01 -4.18578774e-01
4.80351895e-01 -1.85720026e-01 -3.29651475e-01 3.21085215e-01
3.54121029e-01 1.15249820e-01 7.19385594e-02 -5.28973222e-01
-3.49047065e-01 7.44695187e-01 -6.28791600e-02 -2.08424143e-02
1.46136296e+00 -1.71393767e-01 -3.18130732e-01 5.19085169e-01
9.28418398e-01 -4.69253391e-01 -1.27475858e+00 -1.90019459e-01
1.30132288e-01 -4.00519162e-01 1.38419703e-01 -8.36228728e-01
-8.55109036e-01 1.05597305e+00 3.86481822e-01 6.13435447e-01
1.28066266e+00 2.10871965e-01 3.97031128e-01 4.07612711e-01
3.73655289e-01 -7.33168244e-01 1.37882173e-01 4.32599962e-01
5.81109226e-01 -1.43079865e+00 1.64127082e-01 -5.30287802e-01
-4.69083518e-01 1.07932353e+00 5.83717108e-01 -1.62131697e-01
7.17952788e-01 3.00836295e-01 2.11950660e-01 -1.72565266e-01
-6.56519830e-01 -4.97373015e-01 2.45907828e-01 3.84733349e-01
2.90153176e-01 -1.15649424e-01 1.57795206e-01 4.56420891e-02
3.10319573e-01 -2.59871364e-01 1.84252232e-01 1.18634892e+00
-9.05800939e-01 -1.17801702e+00 -3.82784724e-01 4.18979824e-01
-2.48564497e-01 1.40748784e-01 -6.55531764e-01 7.87753880e-01
2.28237376e-01 1.20362008e+00 4.25361879e-02 -3.83338481e-01
3.51732254e-01 1.39057711e-01 5.57817221e-01 -8.70845973e-01
-6.00704730e-01 2.66999245e-01 4.54215752e-03 -4.70647424e-01
-7.53358185e-01 -7.58572698e-01 -9.80242729e-01 8.95152092e-02
-4.07284230e-01 -3.69736739e-03 8.18773568e-01 1.12581873e+00
-5.54904640e-02 3.69656950e-01 7.11265445e-01 -1.21023214e+00
-4.26778287e-01 -6.73223972e-01 -4.56822723e-01 2.76010722e-01
6.49576068e-01 -5.39620578e-01 -2.35334024e-01 2.28231370e-01] | [9.537595748901367, 0.4075257480144501] |
5a2c0f60-4406-4e53-924e-9409deae57a2 | boosting-object-representation-learning-via | 2211.09771 | null | https://arxiv.org/abs/2211.09771v1 | https://arxiv.org/pdf/2211.09771v1.pdf | Boosting Object Representation Learning via Motion and Object Continuity | Recent unsupervised multi-object detection models have shown impressive performance improvements, largely attributed to novel architectural inductive biases. Unfortunately, they may produce suboptimal object encodings for downstream tasks. To overcome this, we propose to exploit object motion and continuity, i.e., objects do not pop in and out of existence. This is accomplished through two mechanisms: (i) providing priors on the location of objects through integration of optical flow, and (ii) a contrastive object continuity loss across consecutive image frames. Rather than developing an explicit deep architecture, the resulting Motion and Object Continuity (MOC) scheme can be instantiated using any baseline object detection model. Our results show large improvements in the performances of a SOTA model in terms of object discovery, convergence speed and overall latent object representations, particularly for playing Atari games. Overall, we show clear benefits of integrating motion and object continuity for downstream tasks, moving beyond object representation learning based only on reconstruction. | ['Kristian Kersting', 'Dwarak Vittal', 'Thomas Rothenbacher', 'Wolfgang Stammer', 'Quentin Delfosse'] | 2022-11-16 | null | null | null | null | ['object-discovery', 'atari-games'] | ['computer-vision', 'playing-games'] | [ 1.15530275e-01 1.39516862e-02 -4.43478078e-02 -1.30369022e-01
-5.48009396e-01 -6.32401168e-01 8.20587575e-01 2.43961737e-01
-5.32279670e-01 3.17573339e-01 2.31539294e-01 8.14224854e-02
-1.56774819e-01 -5.78826070e-01 -9.67136860e-01 -5.69379747e-01
-1.01448067e-01 4.76128906e-01 6.44687355e-01 1.20933503e-02
3.88242215e-01 5.45104265e-01 -1.66004086e+00 3.34739894e-01
3.21141660e-01 8.15573037e-01 3.66995007e-01 8.07019234e-01
4.26806808e-02 9.54879940e-01 -2.69999176e-01 -1.47970960e-01
4.06206042e-01 -2.65220165e-01 -9.12359715e-01 3.86361122e-01
5.51179588e-01 -7.16118157e-01 -3.44736904e-01 7.80356228e-01
2.17783108e-01 3.63233179e-01 6.33860826e-01 -1.16171503e+00
-6.57318592e-01 5.82085669e-01 -7.21340179e-01 3.93623620e-01
-9.53328982e-02 5.71685195e-01 1.25688684e+00 -9.35734570e-01
6.46854997e-01 1.25337923e+00 6.11040831e-01 5.33506691e-01
-1.49496746e+00 -3.90110940e-01 5.40894747e-01 -1.11974531e-03
-1.24231172e+00 -7.55291760e-01 6.23803914e-01 -5.23418725e-01
9.84087765e-01 2.12976485e-02 5.59697211e-01 1.08458591e+00
-1.46408007e-01 1.06201494e+00 6.18704438e-01 -2.27785304e-01
2.40281358e-01 1.00015424e-01 6.22653104e-02 7.30300248e-01
4.10968781e-01 8.93175229e-02 -4.70365196e-01 1.16201229e-01
1.00265789e+00 2.19044909e-02 1.44210868e-02 -5.45861423e-01
-1.24585247e+00 6.45810008e-01 6.34913862e-01 1.60346583e-01
-3.23038131e-01 7.27220535e-01 2.00210169e-01 -1.07654959e-01
1.64285719e-01 4.48419333e-01 -1.86038688e-01 2.44763941e-02
-1.06092215e+00 5.86309135e-01 3.47203165e-01 9.74100292e-01
6.92436993e-01 1.42593943e-02 -2.62255102e-01 5.49172163e-01
4.54587936e-01 2.32375860e-01 3.55864435e-01 -1.19420493e+00
3.84234399e-01 4.36947346e-01 2.81757236e-01 -9.26371634e-01
-4.30091411e-01 -5.89901447e-01 -2.99074918e-01 4.84810919e-01
5.56268394e-01 7.27196410e-02 -1.08298910e+00 2.01687551e+00
2.80314595e-01 3.08284521e-01 -4.67836149e-02 1.09689677e+00
6.96365595e-01 5.23794413e-01 3.73699754e-01 2.67878056e-01
1.29122269e+00 -1.11309373e+00 -3.58551115e-01 -5.99716485e-01
7.43458092e-01 -4.57312256e-01 8.34043205e-01 2.98628449e-01
-1.40018833e+00 -6.14971459e-01 -1.02933455e+00 -3.73276323e-01
-9.49904099e-02 9.91385281e-02 9.03516471e-01 4.59649712e-01
-9.70702171e-01 5.96437573e-01 -1.12135303e+00 -1.36344448e-01
8.92477334e-01 3.63470465e-01 -2.74501324e-01 -3.36134695e-02
-3.91960770e-01 6.48625433e-01 4.46641445e-01 4.79311533e-02
-1.29584372e+00 -7.55095661e-01 -8.45790207e-01 2.29232535e-01
3.15884322e-01 -8.74482095e-01 1.08103871e+00 -1.13830268e+00
-1.15524662e+00 7.25797594e-01 -5.94515502e-02 -5.99209726e-01
6.91520751e-01 -4.98988271e-01 2.64057755e-01 2.76799649e-01
1.93374053e-01 1.19393229e+00 8.27927887e-01 -1.47195303e+00
-8.18226755e-01 -2.43855789e-01 2.29108885e-01 2.99955726e-01
-2.68199831e-01 -1.26507983e-01 -7.31983960e-01 -6.02903187e-01
2.09476784e-01 -9.00732219e-01 -3.28269094e-01 5.67674458e-01
-1.83465227e-01 -1.39764890e-01 9.41144049e-01 -4.42952961e-01
8.57438385e-01 -2.21451664e+00 2.42460772e-01 -1.59515113e-01
2.83932686e-01 1.23970114e-01 -4.18804973e-01 -1.66320466e-02
1.30372092e-01 9.24355686e-02 -1.17311977e-01 -7.77693927e-01
-2.23030135e-01 1.47498086e-01 -4.76396114e-01 5.24939775e-01
7.63369083e-01 1.23080099e+00 -9.84497488e-01 -3.69941443e-01
2.59518892e-01 2.63341129e-01 -1.04028249e+00 -1.52030185e-01
-4.45931941e-01 3.66907954e-01 -3.65903169e-01 5.82717478e-01
4.94284779e-01 -5.19095004e-01 2.98708025e-02 2.44538244e-02
-3.04812714e-02 3.46685261e-01 -1.35907316e+00 1.97330284e+00
-2.98765093e-01 6.43926501e-01 1.05594188e-01 -9.38269198e-01
4.61277753e-01 2.25138471e-01 6.04879558e-01 -5.44715881e-01
-4.00799550e-02 -7.51695335e-02 1.29694805e-01 -2.57627219e-01
6.69470727e-01 -6.96158037e-02 2.15333164e-01 5.89417338e-01
2.54744649e-01 1.40778080e-01 1.82316065e-01 3.84676367e-01
1.11318767e+00 5.33929646e-01 -1.32254630e-01 -2.84196764e-01
3.23975235e-02 1.33482397e-01 5.63599408e-01 1.01763725e+00
-2.59643763e-01 9.65743780e-01 3.31575215e-01 -3.36667299e-01
-1.19023812e+00 -1.30347025e+00 -9.02659073e-02 1.16569078e+00
3.86703491e-01 -2.49166131e-01 -2.61460483e-01 -6.28172934e-01
1.33111104e-01 3.72722864e-01 -6.28959894e-01 -2.58604705e-01
-9.24321353e-01 -8.11183929e-01 4.52621818e-01 8.89002323e-01
2.41016641e-01 -1.09451807e+00 -9.51146603e-01 3.99632126e-01
-2.39719629e-01 -1.18249452e+00 -3.53730291e-01 1.80342436e-01
-1.10132825e+00 -8.03686202e-01 -6.93480074e-01 -5.90751648e-01
6.06082261e-01 6.56423509e-01 1.00890017e+00 3.04803699e-01
-4.55935925e-01 4.25267935e-01 -1.52894869e-01 -2.88416475e-01
-1.03726543e-01 -5.30933356e-03 3.17401513e-02 3.79549973e-02
-1.26661649e-02 -5.95221341e-01 -8.91424417e-01 2.12542698e-01
-1.00207579e+00 1.88393712e-01 7.48903275e-01 6.05112851e-01
3.07737499e-01 -1.82391122e-01 5.52354634e-01 -5.63473880e-01
-3.80299687e-02 -4.64007765e-01 -6.48974359e-01 1.20156966e-02
-3.61678839e-01 2.25212753e-01 2.23499909e-01 -6.31892145e-01
-9.81856823e-01 2.80858010e-01 5.45058846e-02 -6.68184578e-01
-9.80273411e-02 3.47400010e-02 9.56349000e-02 3.09239719e-02
6.47529542e-01 4.69604917e-02 -5.62685020e-02 -4.65626001e-01
5.84748924e-01 2.37271562e-01 4.77001518e-01 -6.89618766e-01
6.41074657e-01 1.01844466e+00 -7.56758675e-02 -7.89219797e-01
-7.77664006e-01 -6.44342005e-01 -6.22853875e-01 -1.24295406e-01
1.01348829e+00 -1.16417599e+00 -7.45528460e-01 3.54680330e-01
-1.11141300e+00 -6.02987289e-01 -3.15313280e-01 5.43323636e-01
-5.52474558e-01 3.12889457e-01 -6.03796899e-01 -9.28751171e-01
2.56755333e-02 -1.03976917e+00 1.13866436e+00 5.68591654e-02
-9.61208045e-02 -7.37201393e-01 -2.91193277e-01 3.02916110e-01
3.09733421e-01 -2.02818066e-02 7.44952142e-01 -2.27835461e-01
-1.21649563e+00 3.48770805e-02 -4.61216599e-01 5.76873310e-02
-1.67754125e-02 -1.07585095e-01 -1.14560509e+00 -3.88901561e-01
-2.79875696e-01 -4.82613713e-01 1.30076098e+00 4.50696707e-01
1.09963429e+00 -3.04464221e-01 -5.83106756e-01 7.29172528e-01
1.54272735e+00 -1.71376109e-01 5.55523574e-01 3.62087727e-01
7.11683929e-01 6.53826177e-01 1.93928674e-01 4.25404459e-01
3.34290951e-01 7.52224445e-01 6.20387614e-01 3.89682613e-02
-6.39880717e-01 -2.29184702e-01 1.97767243e-01 1.25767469e-01
-9.32017043e-02 -3.34312528e-01 -9.13863838e-01 9.82635558e-01
-1.98998058e+00 -1.05177248e+00 -3.94263268e-02 2.04117560e+00
5.27307570e-01 4.09445524e-01 3.09723526e-01 -1.68996051e-01
3.95091057e-01 1.06824338e-01 -6.18054330e-01 1.80967689e-01
-2.58941621e-01 3.19608189e-02 4.03405279e-01 3.73229891e-01
-1.17023766e+00 1.01428831e+00 6.52807999e+00 3.68812561e-01
-1.06619251e+00 3.13898325e-01 4.66904253e-01 -5.07571518e-01
-1.97987199e-01 1.96990296e-01 -9.58273232e-01 1.81215480e-01
5.36471784e-01 4.45280313e-01 2.43128315e-01 7.26989269e-01
2.47832742e-02 -2.03519955e-01 -1.14729118e+00 7.41913736e-01
-6.83493167e-02 -1.54244483e+00 2.10720301e-01 2.00713098e-01
6.95305169e-01 1.32933214e-01 1.24149702e-01 3.23557705e-01
3.96743655e-01 -8.41082573e-01 1.22407448e+00 3.35236043e-01
2.82569200e-01 -3.44961196e-01 2.06238508e-01 2.55386829e-01
-1.11930263e+00 -3.19380164e-01 -2.98330933e-01 -1.16597988e-01
2.20866680e-01 2.20319673e-01 -4.57716137e-01 3.89010340e-01
6.67313874e-01 7.56562173e-01 -6.12531960e-01 1.19855464e+00
7.19129816e-02 5.30654132e-01 -5.00569642e-01 2.14939415e-01
5.40000081e-01 -9.42975283e-04 6.11562073e-01 1.18425846e+00
-4.74715717e-02 1.12247365e-02 1.61014333e-01 1.15549302e+00
6.13817871e-02 -3.47444892e-01 -3.47334385e-01 -2.35647522e-03
2.67655820e-01 1.13348758e+00 -1.01177108e+00 -2.63080150e-01
-4.94204789e-01 9.35283959e-01 5.45038283e-01 4.44913775e-01
-9.07705128e-01 1.14690559e-02 8.54554355e-01 3.34226996e-01
8.50918770e-01 -4.74083632e-01 -4.49750185e-01 -1.12637532e+00
1.51528075e-01 -4.31303144e-01 3.76723051e-01 -6.03700995e-01
-9.25307810e-01 1.91898108e-01 -1.02636836e-01 -9.96900558e-01
2.49406397e-02 -6.02192819e-01 -5.01297593e-01 5.05972624e-01
-1.45358694e+00 -1.35343397e+00 -1.63768291e-01 2.09763780e-01
6.27367973e-01 2.61635691e-01 3.53608638e-01 3.80019754e-01
-4.70484167e-01 3.86049360e-01 -1.80158570e-01 1.67317584e-01
4.72088218e-01 -1.09206140e+00 4.01515573e-01 1.09549713e+00
5.56296587e-01 5.84088624e-01 5.42024851e-01 -5.44768393e-01
-1.29120982e+00 -1.26570618e+00 2.64142722e-01 -8.64965618e-01
4.53252435e-01 -6.64145052e-01 -9.32678401e-01 8.74732018e-01
-2.21124336e-01 3.64085078e-01 1.36700213e-01 2.34975338e-01
-4.87899154e-01 1.32253006e-01 -7.78780222e-01 6.20324016e-01
1.32328820e+00 -3.02371562e-01 -4.26490426e-01 2.69427091e-01
6.44441068e-01 -1.78304270e-01 -4.21068788e-01 3.16588044e-01
6.60422266e-01 -8.38127136e-01 1.33023119e+00 -9.68646348e-01
5.21262228e-01 -5.05376339e-01 -1.15557402e-01 -5.84824204e-01
-6.19207084e-01 -4.40052181e-01 -3.73235583e-01 1.11394656e+00
2.40649238e-01 -1.09362863e-01 8.73706818e-01 5.74558020e-01
-2.29062065e-01 -6.24545395e-01 -7.21501052e-01 -7.57900417e-01
9.17466916e-03 -4.58867311e-01 2.03034863e-01 7.44670808e-01
-4.97659206e-01 4.09065962e-01 -2.98031092e-01 3.51212502e-01
6.91115141e-01 2.20466897e-01 7.66183853e-01 -1.03853083e+00
-6.81723595e-01 -7.48432875e-01 -4.70375299e-01 -1.49680996e+00
-1.74318850e-01 -9.61888015e-01 2.78954923e-01 -1.47948837e+00
3.29138786e-01 -8.03167403e-01 -3.62837911e-01 5.67919135e-01
-2.88862944e-01 5.50306499e-01 4.05694872e-01 5.27634323e-01
-9.59491014e-01 5.49030960e-01 1.07746804e+00 7.02280775e-02
-2.73793548e-01 -2.48718023e-01 -6.00912333e-01 7.25635231e-01
5.03025413e-01 -5.25425673e-01 -2.07980573e-01 -8.94291222e-01
8.83474946e-02 -2.46786878e-01 8.93025875e-01 -1.15818465e+00
1.18078113e-01 -8.10758863e-03 5.60341656e-01 -4.71946120e-01
6.31360233e-01 -5.29896438e-01 2.83402833e-03 6.09926879e-01
-3.32482547e-01 -1.60332799e-01 3.94010544e-01 9.23776805e-01
2.04519868e-01 -2.54552275e-01 8.64575386e-01 -2.45910928e-01
-9.34412301e-01 1.91492796e-01 -3.56355190e-01 -7.66019449e-02
9.27591622e-01 -3.74732107e-01 -2.86709666e-01 3.97690525e-03
-7.37415314e-01 2.50456423e-01 3.82625580e-01 6.11138105e-01
4.86554027e-01 -1.16058147e+00 -6.55597687e-01 1.48957968e-01
1.63050637e-01 2.51320630e-01 1.84899524e-01 8.62469673e-01
-4.07245249e-01 2.76971757e-01 -2.24993169e-01 -9.18188512e-01
-9.35041606e-01 5.40039241e-01 2.90698498e-01 1.90832447e-02
-1.00443542e+00 1.03627062e+00 7.55759239e-01 -1.00530155e-01
3.28447074e-01 -3.32715094e-01 1.75366700e-01 -1.09504960e-01
4.79946285e-01 1.57601118e-01 -3.82046402e-02 -4.14647430e-01
-3.42286021e-01 4.31265354e-01 -2.92101055e-01 -2.80398935e-01
1.37364852e+00 -2.72905707e-01 2.64092684e-01 3.44798446e-01
8.30694616e-01 -1.75307050e-01 -1.86811364e+00 -2.57027507e-01
1.51599795e-01 -4.94686604e-01 1.94924638e-01 -6.07616246e-01
-9.13710594e-01 7.89147556e-01 5.64804792e-01 -1.72442749e-01
6.92190170e-01 3.35872531e-01 5.77764213e-01 2.80705333e-01
3.74027163e-01 -7.80420363e-01 5.54405928e-01 2.69676983e-01
5.92226446e-01 -1.32171023e+00 -2.67963554e-03 -1.52366385e-01
-4.30579513e-01 9.05380070e-01 6.65536582e-01 -3.05288583e-01
4.02316898e-01 2.15096712e-01 -2.15580225e-01 -2.96171576e-01
-9.06816304e-01 -5.32729328e-01 2.25993410e-01 4.38995391e-01
2.88770169e-01 -4.18640822e-01 1.36270538e-01 3.38461161e-01
2.77426869e-01 -1.64883003e-01 4.22498941e-01 1.01315606e+00
-5.34891307e-01 -7.13311493e-01 -1.43494040e-01 2.67841101e-01
-4.26358223e-01 3.31352539e-02 -2.95153502e-02 8.59097779e-01
9.78904441e-02 7.18682528e-01 4.81137276e-01 1.16187192e-01
-1.88962054e-02 -1.61272898e-01 6.91672146e-01 -7.92978108e-01
-4.37412173e-01 9.49826241e-02 -1.76952481e-01 -6.25525475e-01
-3.29444945e-01 -7.59987056e-01 -1.22092783e+00 2.21427474e-02
-5.68332374e-01 -2.52691120e-01 6.21678829e-01 9.24975634e-01
5.13816535e-01 6.13226533e-01 1.25954986e-01 -1.08116901e+00
-5.34806907e-01 -7.85486698e-01 -2.92400301e-01 5.04240096e-01
5.47745168e-01 -7.93410778e-01 -1.68992564e-01 1.39427260e-01] | [9.494319915771484, 0.32302579283714294] |
0d504c20-f38d-4894-9b2a-da9f8cfd7f81 | subverting-machines-fluctuating-identities-re | 2205.13740 | null | https://arxiv.org/abs/2205.13740v1 | https://arxiv.org/pdf/2205.13740v1.pdf | Subverting machines, fluctuating identities: Re-learning human categorization | Most machine learning systems that interact with humans construct some notion of a person's "identity," yet the default paradigm in AI research envisions identity with essential attributes that are discrete and static. In stark contrast, strands of thought within critical theory present a conception of identity as malleable and constructed entirely through interaction; a doing rather than a being. In this work, we distill some of these ideas for machine learning practitioners and introduce a theory of identity as autopoiesis, circular processes of formation and function. We argue that the default paradigm of identity used by the field immobilizes existing identity categories and the power differentials that co$\unicode{x2010}$occur, due to the absence of iterative feedback to our models. This includes a critique of emergent AI fairness practices that continue to impose the default paradigm. Finally, we apply our theory to sketch approaches to autopoietic identity through multilevel optimization and relational learning. While these ideas raise many open questions, we imagine the possibilities of machines that are capable of expressing human identity as a relationship perpetually in flux. | ['Kevin R. McKee', 'Jackie Kay', 'Christina Lu'] | 2022-05-27 | null | null | null | null | ['relational-reasoning'] | ['natural-language-processing'] | [ 4.88209426e-02 8.36170495e-01 -1.49079695e-01 -1.99346945e-01
4.17585582e-01 -5.42639911e-01 1.40050101e+00 3.60139087e-02
-1.89759329e-01 5.52680552e-01 7.15400457e-01 -3.63883972e-01
-3.21418732e-01 -9.37185049e-01 -2.16828614e-01 -6.83998227e-01
1.33368373e-01 5.12181699e-01 -5.65767407e-01 -8.07870507e-01
3.83338958e-01 -3.46373580e-02 -1.67158067e+00 -1.78025365e-02
1.01600182e+00 3.85296613e-01 -4.69268799e-01 4.08992797e-01
3.32740657e-02 1.63445950e+00 -4.74796593e-01 -5.18712342e-01
4.27576005e-01 -1.18849397e+00 -1.12569547e+00 9.78901535e-02
2.46512115e-01 2.39839986e-01 -2.74653912e-01 8.90343308e-01
2.10888430e-01 3.01925708e-02 6.20777726e-01 -1.35914922e+00
-1.60668588e+00 1.10085452e+00 7.57621154e-02 2.91408356e-02
2.08356157e-01 5.20604432e-01 1.05650890e+00 -3.30438763e-01
5.80599129e-01 1.46184814e+00 1.04683542e+00 7.82560885e-01
-1.55200243e+00 -3.54807347e-01 -2.94276588e-02 -1.12263756e-02
-1.43722701e+00 -8.18703890e-01 5.75898290e-01 -1.20038342e+00
8.23276579e-01 4.77744877e-01 1.43250191e+00 7.56404221e-01
4.65278864e-01 2.30881229e-01 1.15243101e+00 -7.06831157e-01
1.55842200e-01 1.40104130e-01 -3.58579904e-02 8.10099304e-01
2.91565746e-01 3.49065870e-01 -6.72895014e-01 -1.77161634e-01
7.85836995e-01 -2.17599854e-01 2.06015855e-02 -1.48790091e-01
-1.43226719e+00 9.74819243e-01 1.83971390e-01 7.46735692e-01
-2.55768865e-01 4.23970133e-01 2.70601034e-01 7.66337454e-01
6.15235269e-01 1.02755892e+00 1.35421187e-01 -1.70625702e-01
-3.87581021e-01 1.69118181e-01 9.96949792e-01 4.37738568e-01
7.29171097e-01 1.52677909e-01 3.98516864e-01 2.67559767e-01
4.87172246e-01 2.12393582e-01 3.55327368e-01 -1.52417743e+00
-7.25442708e-01 6.79720819e-01 -9.24217999e-02 -1.11565161e+00
-3.28161240e-01 -3.28866959e-01 -6.83321714e-01 5.03982484e-01
5.28600276e-01 -9.41114724e-02 2.39155497e-02 1.81683910e+00
2.39754766e-01 -1.35494471e-01 2.90690422e-01 9.21389043e-01
4.56085056e-01 4.58886027e-01 2.25418717e-01 -2.35080540e-01
9.29500759e-01 -6.24871850e-01 -7.29026318e-01 1.73391439e-02
6.80169106e-01 -6.37417853e-01 1.16816950e+00 2.37178326e-01
-1.28190243e+00 -9.57001448e-02 -9.89745736e-01 -2.81419963e-01
-1.55480400e-01 -7.43012011e-01 1.25934196e+00 1.02567470e+00
-1.46841526e+00 8.34064960e-01 -6.16710782e-01 -7.25388885e-01
4.93744552e-01 2.46475965e-01 5.35752028e-02 9.97074962e-01
-1.02068079e+00 1.21490300e+00 1.85531825e-01 -2.04647169e-01
-7.44102240e-01 -6.63900793e-01 -4.62190509e-01 -2.27819473e-01
2.49993980e-01 -1.18149710e+00 9.81581986e-01 -2.15754604e+00
-1.47794807e+00 1.46959293e+00 3.12618524e-01 -4.77101028e-01
5.62591732e-01 8.28668550e-02 -3.10062438e-01 -2.00116888e-01
2.13553365e-02 4.13431585e-01 2.64787078e-01 -1.72768307e+00
-5.36103904e-01 -1.51608169e-01 3.02646041e-01 2.04870597e-01
-3.36924314e-01 1.87523484e-01 9.31647956e-01 -3.43052983e-01
1.55877009e-01 -8.48072112e-01 -5.88716231e-02 -6.76508844e-02
-7.22777024e-02 -5.94570279e-01 3.57395738e-01 -7.10015893e-02
1.31702602e+00 -1.94984508e+00 1.55911207e-01 -3.49857062e-02
8.04855406e-01 -5.21779299e-01 2.98007458e-01 8.03633928e-01
3.48683968e-02 7.90882051e-01 -4.45261300e-02 -8.81745946e-04
5.00674903e-01 2.63848573e-01 -2.18561605e-01 8.51898253e-01
-1.65167734e-01 9.64928567e-01 -1.20899749e+00 -6.09674573e-01
-6.86384877e-03 4.44324166e-01 -6.21702850e-01 -2.35562727e-01
-1.07928269e-01 5.24470389e-01 -2.55072862e-01 4.52365607e-01
-2.13933036e-01 -5.69835007e-01 6.89526975e-01 4.47515666e-01
-6.63012445e-01 3.48604113e-01 -8.09007287e-01 1.40693283e+00
-1.00605391e-01 7.04784989e-01 2.13362738e-01 -1.00151992e+00
7.68631041e-01 4.64767516e-01 8.44016731e-01 -4.11853462e-01
4.23885047e-01 2.44832281e-02 7.45621622e-01 -5.12988091e-01
4.53257591e-01 -5.75061560e-01 -1.76465556e-01 1.35958505e+00
-7.46978074e-02 -4.39006567e-01 -2.42767364e-01 1.85388193e-01
8.57818723e-01 4.30094391e-01 5.20092010e-01 -1.24052906e+00
3.46361488e-01 3.41502398e-01 7.51677752e-01 6.84768438e-01
-4.76803154e-01 -9.76761058e-03 5.04336417e-01 -1.09410226e+00
-1.50771332e+00 -1.19925261e+00 -2.28797406e-01 1.20989382e+00
2.56602854e-01 -4.43856299e-01 -7.98670173e-01 -2.61375066e-02
-5.98629974e-02 8.11629176e-01 -8.03784728e-01 -6.25870451e-02
-5.57918131e-01 -7.18802214e-01 5.03958941e-01 -4.30697292e-01
9.11297202e-02 -1.22348106e+00 -1.16839457e+00 -4.43200134e-02
-6.20829202e-02 -2.61854351e-01 -1.55785874e-01 -7.54038468e-02
-5.69409192e-01 -8.63123059e-01 4.06249166e-01 -5.68924606e-01
5.95265448e-01 -7.78423101e-02 1.39659274e+00 6.90226614e-01
-1.86449960e-01 6.92906201e-01 -2.57364750e-01 -4.76888657e-01
-1.21255016e+00 -1.47566006e-01 2.98885882e-01 -5.56446351e-02
5.64464569e-01 -1.07279932e+00 -7.59791851e-01 -1.34975657e-01
-6.83712900e-01 4.34788078e-01 -7.30326921e-02 7.08896697e-01
-6.77782595e-02 -2.19847217e-01 9.78203058e-01 -8.03872943e-01
6.56338274e-01 -7.55340159e-01 1.12542488e-01 2.53840536e-01
-1.03038478e+00 -2.15881065e-01 5.43513894e-01 -2.84556955e-01
-1.18855655e+00 -4.09112930e-01 6.57521129e-01 1.92312822e-01
1.89695001e-01 1.36739373e-01 5.59558533e-02 1.18087836e-01
1.20737171e+00 8.43778104e-02 7.85271347e-01 -2.12291386e-02
7.50189304e-01 6.67390287e-01 4.32476580e-01 -9.90411222e-01
6.19996369e-01 6.57636642e-01 -3.12397659e-01 -8.43585432e-01
-3.97226274e-01 2.72653311e-01 -6.79380655e-01 -7.20276535e-01
9.46752250e-01 -9.13115919e-01 -1.39384925e+00 4.92891707e-02
-8.56937647e-01 -5.47483563e-01 -1.12519825e+00 2.64862925e-01
-1.02254200e+00 -3.95252282e-04 -4.20614034e-01 -9.60733473e-01
-4.11981881e-01 -4.65336293e-01 2.12352872e-01 3.08012724e-01
-9.73085105e-01 -1.24352837e+00 2.46507972e-01 4.64993209e-01
6.88302994e-01 6.09219611e-01 1.00955474e+00 -6.64392292e-01
-4.00176316e-01 4.82196540e-01 5.41337252e-01 -1.04025425e-02
5.58942139e-01 1.01271570e-01 -8.66354704e-01 8.05381089e-02
4.37097639e-01 -4.34127480e-01 1.43538937e-01 -1.43479705e-01
3.39065492e-01 -8.45287383e-01 -6.75279822e-04 3.55082273e-01
1.57953680e+00 6.05123639e-01 4.54371065e-01 3.48163754e-01
4.55568343e-01 1.15772820e+00 -1.63330406e-01 6.85405791e-01
8.26480269e-01 -2.02235449e-02 -1.10699339e-02 -1.70855001e-02
-9.51491110e-03 -1.43881902e-01 2.28272632e-01 1.08569348e+00
-5.86423278e-01 2.11291239e-01 -1.41026783e+00 4.57220882e-01
-2.13908672e+00 -1.72082675e+00 -2.21745327e-01 1.83482265e+00
1.24906743e+00 -1.23053506e-01 2.59918839e-01 -2.91862875e-01
8.16802025e-01 1.25512540e-01 -3.96777421e-01 -8.93187344e-01
-3.33688527e-01 -2.85781780e-03 -8.93366411e-02 7.79322088e-01
-6.90116584e-01 1.29533708e+00 7.98138857e+00 2.36007094e-01
-8.63019049e-01 5.23158550e-01 7.73525178e-01 -1.75362453e-01
-9.63825881e-01 3.94381106e-01 1.56505257e-01 3.21157306e-01
6.44832730e-01 -8.87199044e-01 8.82349133e-01 2.57618189e-01
4.05929953e-01 -3.10292363e-01 -1.26979816e+00 4.91625071e-01
1.23618603e-01 -1.46444035e+00 -3.59132677e-01 2.57128745e-01
1.06101024e+00 -3.26709718e-01 -7.49107078e-02 -2.57680267e-01
1.27847075e+00 -1.55332100e+00 1.25456381e+00 7.30996907e-01
3.96129459e-01 -4.62399989e-01 -1.44551555e-02 2.86214650e-01
-5.15708208e-01 -3.26337636e-01 5.63491657e-02 -1.05074918e+00
8.41764559e-04 1.63882986e-01 -1.67117920e-02 -6.38741031e-02
3.10452819e-01 5.03367186e-01 -2.95281947e-01 1.59780145e-01
1.14230700e-01 2.82929748e-01 -8.94937888e-02 -8.57280940e-02
-1.30215421e-01 -6.95763171e-01 5.53770840e-01 8.79395247e-01
-1.27335146e-01 5.47662020e-01 5.96406311e-02 1.38520539e+00
3.68342161e-01 3.41233015e-02 -8.24912548e-01 -3.73600662e-01
7.66182840e-01 1.05354357e+00 -8.90963435e-01 -1.57991245e-01
-4.31069940e-01 3.84598821e-01 -5.52330613e-02 3.38973999e-02
-7.17171311e-01 3.03753704e-01 1.04646647e+00 3.03398877e-01
-6.59534276e-01 -3.72312218e-01 -7.76349604e-01 -1.16544616e+00
-6.01326048e-01 -1.21204197e+00 3.52299213e-02 -2.24322096e-01
-1.37611532e+00 1.04939483e-01 -2.29502738e-01 -4.79829043e-01
-1.04110107e-01 6.19308539e-02 -7.10308313e-01 4.18189436e-01
-4.57215607e-01 -1.53309560e+00 -6.04589544e-02 2.98488230e-01
1.79876424e-02 -3.76619637e-01 9.68434572e-01 -2.29538307e-01
-3.38725746e-01 2.99719542e-01 7.45620728e-02 -1.29697751e-02
3.37029666e-01 -1.11961877e+00 4.21430647e-01 5.57591915e-01
8.49076137e-02 9.41770434e-01 8.91229033e-01 -6.38103485e-01
-1.37069881e+00 -2.35717595e-01 1.04656816e+00 -8.04845750e-01
9.75675464e-01 -3.74197543e-01 -5.20717800e-01 9.67669070e-01
9.72612858e-01 -5.55970550e-01 1.16189015e+00 2.28933826e-01
-2.49514088e-01 1.34253651e-01 -1.13609076e+00 9.45571125e-01
1.81991827e+00 -6.76060975e-01 -8.24330628e-01 2.83686429e-01
7.73551762e-01 4.81380641e-01 -1.21497381e+00 -4.73839976e-02
8.65329266e-01 -1.22517371e+00 7.34553456e-01 -5.02300799e-01
6.28523827e-01 -1.94820672e-01 -1.06343441e-01 -9.77202117e-01
-6.23037755e-01 -1.17538929e+00 3.61086905e-01 1.38757789e+00
-1.36738596e-02 -1.21473646e+00 5.64840794e-01 8.70866835e-01
-2.33730391e-01 -4.48089629e-01 -8.92908394e-01 -4.60583806e-01
9.60379839e-01 9.13447887e-02 7.36138165e-01 1.94003868e+00
8.35428476e-01 2.89475024e-01 1.36143580e-01 -7.18401134e-01
6.18305087e-01 -4.47938181e-02 5.90707600e-01 -1.49453092e+00
-7.75093734e-02 -1.01439941e+00 -4.25747007e-01 -3.60037722e-02
2.33833835e-01 -1.26395798e+00 1.82097945e-02 -1.33520234e+00
3.56460780e-01 -8.15655291e-01 8.36614221e-02 3.28276068e-01
4.55948442e-01 -3.47814499e-03 4.29010540e-01 7.80264795e-01
-4.15313393e-01 3.81730378e-01 1.17948592e+00 1.47525042e-01
-3.45600992e-01 -9.88193989e-01 -1.54659438e+00 9.67645526e-01
9.07490492e-01 -2.96096534e-01 -6.01212859e-01 -1.57327488e-01
6.77472413e-01 -3.36708486e-01 6.88013792e-01 -6.38467908e-01
4.07063454e-01 -9.13486898e-01 2.03565340e-02 5.70291638e-01
-2.13226110e-01 -4.67072427e-01 9.49174285e-01 8.03493500e-01
-4.88100499e-01 4.44108509e-02 -4.32702273e-01 -1.76098302e-01
3.47131819e-01 1.44157290e-01 7.61064172e-01 -5.95888734e-01
-3.79567176e-01 -3.36806059e-01 -7.36907482e-01 3.30567241e-01
1.32892096e+00 -5.99792540e-01 -8.93344641e-01 -2.51827329e-01
-6.74986005e-01 1.82349011e-01 1.20766449e+00 3.51425767e-01
6.94828480e-02 -1.16806257e+00 -9.38477695e-01 -2.49175787e-01
-2.35751987e-01 -2.55290419e-01 -3.22150961e-02 6.85126185e-01
-8.87749374e-01 1.80307264e-03 -6.45771444e-01 -2.20526308e-01
-7.04585731e-01 6.00305855e-01 9.25593495e-01 6.75075710e-01
-7.58042693e-01 6.29223943e-01 4.35439020e-01 -2.81336308e-01
-3.17018658e-01 4.70719159e-01 -1.05192892e-01 3.54642719e-01
9.63194668e-02 4.11378890e-01 -9.69810843e-01 -1.00084305e+00
-2.33527735e-01 4.47304219e-01 6.46645486e-01 -3.27566803e-01
1.17224681e+00 -4.70221192e-01 -1.02039552e+00 8.16097856e-01
5.19604385e-01 -8.12320262e-02 -1.08546495e+00 5.36088198e-02
8.26584622e-02 -3.86091918e-01 -4.27789778e-01 -7.72010088e-01
-4.10895973e-01 4.14378196e-01 3.93273413e-01 7.14716792e-01
6.54447913e-01 3.11875641e-01 1.33382350e-01 1.11177824e-01
3.42279941e-01 -1.54324615e+00 2.36592606e-01 2.08807260e-01
9.70859528e-01 -8.75839591e-01 2.63837278e-01 -6.55560866e-02
-6.63537562e-01 9.55248713e-01 6.76097572e-01 -2.44670495e-01
8.37433279e-01 3.10545385e-01 2.34570310e-01 -3.63005608e-01
-1.01126790e+00 -2.68660277e-01 -4.01577950e-01 7.54563868e-01
8.51219952e-01 4.28945422e-01 -1.09232211e+00 3.69294822e-01
-8.81778955e-01 -2.20546395e-01 8.48575413e-01 8.04752171e-01
-6.93334877e-01 -1.01781261e+00 -4.33388084e-01 -8.20677057e-02
-5.87005496e-01 9.02509093e-02 -8.53727698e-01 9.27877069e-01
9.66594517e-01 1.13106036e+00 7.04031467e-01 -5.75796008e-01
-3.31056565e-01 2.40819678e-01 5.45294881e-01 -5.64160585e-01
-1.32456326e+00 -1.98263824e-01 2.08892897e-01 -1.52899042e-01
-7.66952276e-01 -1.01556325e+00 -1.51087809e+00 -1.30975163e+00
1.82934143e-02 1.70697436e-01 2.31547698e-01 8.03507328e-01
1.63766205e-01 1.88838869e-01 4.63495761e-01 -3.16235960e-01
-2.99426317e-01 -4.32569474e-01 -3.73822063e-01 4.60325539e-01
9.77324918e-02 -2.56543189e-01 -3.57180178e-01 5.39357007e-01] | [9.096700668334961, 6.32020378112793] |
c3b7ef6e-2bb0-4dcd-8249-e501e5de9fce | few-shot-generalization-for-single-image-3d | 1909.01205 | null | https://arxiv.org/abs/1909.01205v1 | https://arxiv.org/pdf/1909.01205v1.pdf | Few-Shot Generalization for Single-Image 3D Reconstruction via Priors | Recent work on single-view 3D reconstruction shows impressive results, but has been restricted to a few fixed categories where extensive training data is available. The problem of generalizing these models to new classes with limited training data is largely open. To address this problem, we present a new model architecture that reframes single-view 3D reconstruction as learnt, category agnostic refinement of a provided, category-specific prior. The provided prior shape for a novel class can be obtained from as few as one 3D shape from this class. Our model can start reconstructing objects from the novel class using this prior without seeing any training image for this class and without any retraining. Our model outperforms category-agnostic baselines and remains competitive with more sophisticated baselines that finetune on the novel categories. Additionally, our network is capable of improving the reconstruction given multiple views despite not being trained on task of multi-view reconstruction. | ['Bram Wallace', 'Bharath Hariharan'] | 2019-09-03 | few-shot-generalization-for-single-image-3d-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Wallace_Few-Shot_Generalization_for_Single-Image_3D_Reconstruction_via_Priors_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Wallace_Few-Shot_Generalization_for_Single-Image_3D_Reconstruction_via_Priors_ICCV_2019_paper.pdf | iccv-2019-10 | ['single-view-3d-reconstruction'] | ['computer-vision'] | [ 3.27180356e-01 3.40770870e-01 -1.95009246e-01 -5.63597322e-01
-8.51978719e-01 -9.27150726e-01 8.88558567e-01 -5.55702507e-01
-2.36309707e-01 4.00275618e-01 4.92265105e-01 -3.50702852e-02
2.63806999e-01 -5.03596187e-01 -1.13813138e+00 -4.79265988e-01
3.99002224e-01 9.45643663e-01 4.02609080e-01 -2.37188698e-03
1.40633360e-01 6.77144051e-01 -1.70321381e+00 6.15999818e-01
1.73646420e-01 7.90687561e-01 4.61725354e-01 6.01567686e-01
2.42328420e-01 2.43894979e-01 -2.45833784e-01 -1.06422365e-01
5.98344743e-01 -9.26498994e-02 -7.89135695e-01 6.88345909e-01
1.07951880e+00 -7.53551543e-01 -2.21892089e-01 6.46337688e-01
3.56773794e-01 1.86554864e-01 9.44417417e-01 -7.32791185e-01
-7.39485204e-01 8.02080408e-02 -5.41663527e-01 2.48715341e-01
4.10552651e-01 1.17287554e-01 9.37709153e-01 -1.48895204e+00
9.75512326e-01 1.22658348e+00 6.66824877e-01 7.69553185e-01
-1.48527753e+00 -4.38794196e-01 7.56990314e-01 -3.71969976e-02
-1.12090027e+00 -4.91812885e-01 6.63024247e-01 -5.14230728e-01
1.20838034e+00 -5.42484745e-02 6.98970854e-01 1.36475253e+00
-9.62975770e-02 9.41942096e-01 1.25206459e+00 -5.52743711e-02
6.60044998e-02 5.15710041e-02 -1.75823316e-01 3.76447767e-01
6.68922886e-02 3.17666382e-01 -3.56233954e-01 2.32653692e-01
9.94114578e-01 2.81856209e-01 -3.18124443e-01 -1.15364444e+00
-1.40510392e+00 6.47952557e-01 6.25148177e-01 -1.71345118e-02
-3.48458320e-01 9.48362350e-02 8.85101855e-02 2.41448864e-01
7.74467945e-01 3.62702757e-01 -8.39117825e-01 3.61422360e-01
-8.49280179e-01 2.56841511e-01 4.05585349e-01 1.09703410e+00
8.48541558e-01 1.13113217e-01 1.84841558e-01 7.66770542e-01
1.07697479e-01 6.89791083e-01 1.76522862e-02 -1.12964296e+00
3.49932641e-01 4.25044686e-01 1.68245360e-01 -3.87003064e-01
-4.02566910e-01 -7.41022468e-01 -5.30803084e-01 4.99256879e-01
3.13819855e-01 2.61658639e-01 -1.48288643e+00 1.74667621e+00
5.04904985e-01 1.04331993e-01 6.04775874e-03 1.03304374e+00
9.89896655e-01 3.63135308e-01 -3.21971416e-01 9.17687342e-02
9.97558117e-01 -1.11585367e+00 -9.17321071e-02 -5.99753380e-01
-7.39249662e-02 -6.89207494e-01 1.04276574e+00 6.46494567e-01
-1.26120079e+00 -7.15855062e-01 -9.60060537e-01 -2.64032125e-01
-1.43209442e-01 -1.01756006e-01 6.73912764e-01 4.36577499e-01
-1.36179078e+00 4.34951991e-01 -8.95525813e-01 -6.24929428e-01
6.83135688e-01 2.40227088e-01 -8.00188482e-01 -6.77009821e-01
-4.16897714e-01 1.00888216e+00 1.87560484e-01 -2.24097922e-01
-1.50892818e+00 -9.61632192e-01 -8.69584441e-01 -2.36927047e-01
4.81875062e-01 -1.38005555e+00 1.28704119e+00 -8.75151396e-01
-1.27938473e+00 1.29624307e+00 -9.55964178e-02 -1.90736771e-01
3.20589185e-01 -1.46192595e-01 -4.86753136e-02 3.70361447e-01
3.25915068e-01 8.40331852e-01 1.21474874e+00 -2.02480102e+00
-5.92093349e-01 -6.24763668e-01 2.98568487e-01 5.69178224e-01
1.72980487e-01 -4.12162632e-01 -4.79015440e-01 -6.13486409e-01
8.03159297e-01 -9.48329866e-01 -9.39215198e-02 2.29261577e-01
-2.31875345e-01 7.34798685e-02 7.61665165e-01 -4.19932008e-01
2.63439089e-01 -2.05439782e+00 6.22407734e-01 -1.83720738e-01
3.03363413e-01 -1.09601393e-01 -2.48204514e-01 1.44106239e-01
-2.06459448e-01 5.78229986e-02 -1.84077531e-01 -5.67136824e-01
-1.31663129e-01 3.63334358e-01 -5.22672415e-01 5.31622648e-01
1.50750026e-01 6.46668971e-01 -9.38267648e-01 3.84580083e-02
3.73203129e-01 4.07391727e-01 -9.33169723e-01 4.81419861e-01
-3.50343019e-01 7.67560244e-01 -1.01551279e-01 7.57857978e-01
9.25055325e-01 -5.58868825e-01 9.65158194e-02 -4.77764606e-01
2.72416025e-01 2.61138827e-01 -1.04264796e+00 2.26465631e+00
-6.77198470e-01 1.04840890e-01 1.69523105e-01 -1.15691233e+00
5.82932889e-01 1.36526659e-01 4.23650771e-01 -1.91963270e-01
-1.28948867e-01 1.72903672e-01 -1.78095713e-01 -3.53114754e-01
1.83620572e-01 -6.73314631e-01 6.39155433e-02 6.75563693e-01
5.93107939e-01 -5.31015694e-01 -2.96657085e-01 2.97571599e-01
9.64401424e-01 7.18419254e-01 2.46050343e-01 1.97516203e-01
6.74487576e-02 -4.89695370e-02 2.85872757e-01 7.98453152e-01
-1.35305952e-02 1.35831845e+00 -2.55052418e-01 -5.04153788e-01
-1.12470007e+00 -1.59271336e+00 -3.36443424e-01 1.05712855e+00
2.97545582e-01 -5.33611514e-02 -5.42875193e-02 -1.22650611e+00
2.51597345e-01 6.07100070e-01 -7.89592147e-01 -3.17665264e-02
-5.13524294e-01 -3.89099449e-01 -8.79229382e-02 7.02602565e-01
4.36533839e-01 -7.59678960e-01 -3.20702493e-01 2.69909091e-02
-2.93790758e-01 -1.24737513e+00 -3.21216434e-01 3.23798656e-01
-1.23654854e+00 -1.20243490e+00 -7.95203507e-01 -8.12183380e-01
8.75907660e-01 7.67092586e-01 1.59886718e+00 -9.23146214e-03
1.52829632e-01 1.03027868e+00 -2.88835973e-01 -1.40585661e-01
-2.77088195e-01 -2.15899237e-02 7.13676885e-02 -3.27404886e-02
1.18938722e-01 -9.33825612e-01 -7.67031550e-01 4.50371802e-01
-6.33129537e-01 1.92060173e-01 7.70105541e-01 8.31200123e-01
7.56262243e-01 -4.04454648e-01 5.33545732e-01 -1.06339705e+00
-3.16552252e-01 -6.20693803e-01 -5.94140626e-02 -2.50655226e-02
-2.95750678e-01 -7.66987801e-02 5.58269024e-01 -3.43511701e-01
-1.24805999e+00 3.22965592e-01 -1.50438070e-01 -8.41414571e-01
-6.51545823e-01 1.77985340e-01 -2.10574731e-01 1.54853603e-02
7.62442648e-01 2.48216018e-01 3.59678529e-02 -8.54123235e-01
5.87932885e-01 2.19185501e-01 5.72214425e-01 -4.69610959e-01
9.33993459e-01 9.54293430e-01 -6.78570196e-02 -3.99506032e-01
-1.30135870e+00 -6.39107585e-01 -9.49768722e-01 -9.05754417e-02
7.40105331e-01 -1.41076469e+00 -1.13951296e-01 2.54862607e-01
-9.40678537e-01 -3.13817620e-01 -4.78300899e-01 5.40475786e-01
-9.66423213e-01 3.40384692e-01 -2.84655422e-01 -3.28163296e-01
5.41196726e-02 -9.07974422e-01 1.47205698e+00 -3.22064310e-01
1.51911110e-01 -9.09253478e-01 -1.06432877e-01 5.51121593e-01
1.88781410e-01 1.50990307e-01 8.08634162e-01 -5.09335518e-01
-1.00350213e+00 -1.73718795e-01 -1.31970972e-01 4.92126137e-01
1.23238869e-01 -4.92742896e-01 -1.21244526e+00 -6.69280171e-01
2.25249082e-01 -6.41977012e-01 1.34196699e+00 3.81883323e-01
1.15378535e+00 3.70229036e-02 -4.61250007e-01 9.68038261e-01
1.38779366e+00 -2.97292799e-01 3.27620924e-01 2.37517461e-01
7.54364669e-01 2.91988611e-01 4.49638277e-01 1.37032613e-01
6.65346205e-01 7.01815963e-01 9.48909819e-01 -1.30093336e-01
-6.76170766e-01 -3.27219486e-01 1.94567457e-01 4.38019544e-01
-3.45698744e-01 -2.39242181e-01 -6.23086572e-01 7.75257945e-01
-1.47882700e+00 -9.82307136e-01 4.19943094e-01 2.07249236e+00
4.87280905e-01 9.24877301e-02 1.50013596e-01 -1.58432215e-01
4.04977709e-01 1.36684790e-01 -9.18877304e-01 1.57872215e-01
-6.43331409e-02 1.19109906e-01 1.75147608e-01 4.47519839e-01
-1.17841661e+00 9.09935892e-01 6.70025253e+00 3.50245386e-01
-1.02930653e+00 1.52021140e-01 3.83662879e-01 -2.71623403e-01
-5.98662138e-01 1.50175020e-01 -8.12745929e-01 -1.23142391e-01
2.21941993e-01 3.21100265e-01 5.21965921e-01 8.73894155e-01
-1.92005217e-01 -1.26683190e-01 -1.64145041e+00 9.64997292e-01
5.82072496e-01 -1.29286098e+00 4.24095392e-01 2.03601733e-01
9.67560112e-01 3.65481853e-01 3.20028394e-01 3.95474702e-01
4.87793535e-01 -9.38542604e-01 8.30465436e-01 5.31315029e-01
9.72822785e-01 -3.56697679e-01 3.95519137e-01 5.43326914e-01
-8.14426839e-01 -9.83519927e-02 -4.93705541e-01 5.66964671e-02
1.66490540e-01 2.55112827e-01 -1.04630458e+00 5.57555437e-01
8.24548185e-01 1.43787527e+00 -6.62806034e-01 8.41760516e-01
-2.62279332e-01 4.04708445e-01 -3.72407794e-01 8.33518386e-01
-1.72752403e-02 1.55487135e-01 7.18058825e-01 7.39361882e-01
3.47687632e-01 2.57867277e-01 4.01846141e-01 7.02625692e-01
-1.14704661e-01 -4.82963234e-01 -9.13498521e-01 5.53260446e-01
1.96669951e-01 1.08844757e+00 -6.53453290e-01 -3.64137948e-01
-6.54973567e-01 1.19210362e+00 6.30382478e-01 5.09529829e-01
-4.38996404e-01 4.83625501e-01 4.14136887e-01 6.18455708e-01
8.46994817e-01 -1.64056271e-01 -4.52468067e-01 -1.62654650e+00
-1.78656504e-01 -7.14382470e-01 5.26147783e-01 -1.21947849e+00
-1.59887016e+00 4.92693812e-01 2.55479962e-01 -1.46146154e+00
-3.76594901e-01 -7.06155598e-01 -8.01495984e-02 7.40140855e-01
-1.68685687e+00 -1.61267173e+00 -2.86713362e-01 7.27437913e-01
7.80712068e-01 -3.95391762e-01 9.17327940e-01 6.30034357e-02
1.35441929e-01 3.64183128e-01 -1.18093133e-01 -3.50718707e-01
9.52226460e-01 -1.43498051e+00 3.08783948e-01 8.08821559e-01
2.49750152e-01 5.34776270e-01 4.98664647e-01 -6.08381212e-01
-1.25622797e+00 -1.08130860e+00 5.35016298e-01 -1.28506851e+00
1.92171544e-01 -4.67294544e-01 -8.10606062e-01 1.11258459e+00
7.22650290e-02 5.53694129e-01 4.88242120e-01 4.32228893e-01
-7.96182930e-01 -1.83511190e-02 -1.28138661e+00 2.00366080e-01
1.65439737e+00 -5.41970909e-01 -7.05307841e-01 3.72251809e-01
5.19704282e-01 -6.98154688e-01 -9.12409544e-01 6.25383735e-01
4.94404584e-01 -1.15206087e+00 1.58569205e+00 -6.90011680e-01
5.59235036e-01 -4.63280082e-01 -5.46021342e-01 -1.51546788e+00
-5.71540236e-01 1.34687871e-01 -2.74890423e-01 6.69761777e-01
2.14582652e-01 -2.81878173e-01 7.81434655e-01 2.67248839e-01
-7.23580956e-01 -6.30643129e-01 -7.50103235e-01 -7.21106529e-01
2.88354129e-01 -2.81654269e-01 4.45963204e-01 7.54344285e-01
-6.48674965e-01 5.98026037e-01 -4.39243466e-01 3.47131282e-01
7.89250791e-01 7.45614588e-01 1.15317583e+00 -1.21973062e+00
-6.08160853e-01 -7.01605901e-02 -2.90909708e-01 -1.50950563e+00
1.32170305e-01 -1.26483977e+00 5.36974631e-02 -1.84187710e+00
6.09745800e-01 -3.42198223e-01 -4.83567864e-02 4.77855116e-01
-1.81612417e-01 6.79051340e-01 3.43076587e-01 3.04434240e-01
-6.40236735e-01 5.68506122e-01 1.72010863e+00 -7.14013427e-02
1.94030523e-01 2.23404422e-01 -9.22185779e-01 8.90723050e-01
3.50670844e-01 -3.36147100e-01 -5.70033252e-01 -7.12465465e-01
9.86688286e-02 -3.27738598e-02 6.89835668e-01 -7.38831520e-01
-1.99245632e-01 -1.27488524e-01 8.61694276e-01 -1.05248475e+00
6.17147624e-01 -1.06279564e+00 2.31841609e-01 -3.39510590e-02
-8.11039507e-02 -2.62196511e-01 7.40150288e-02 1.12302029e+00
1.03922844e-01 1.09919995e-01 9.21751142e-01 -7.27793097e-01
-9.98964190e-01 4.11330700e-01 -9.24958438e-02 1.74142033e-01
7.57317066e-01 -5.82976878e-01 -1.30898878e-01 -3.90586436e-01
-1.47393537e+00 -5.06986566e-02 9.52642381e-01 5.54874182e-01
7.44601607e-01 -1.34338117e+00 -6.99988186e-01 4.40910995e-01
3.02198499e-01 4.23559099e-01 3.26138526e-01 3.36637914e-01
-9.51325744e-02 1.38584971e-01 -2.88136423e-01 -1.11436093e+00
-8.93294454e-01 9.17934358e-01 4.01406586e-01 -1.12382911e-01
-9.60507631e-01 8.65821123e-01 8.25124919e-01 -1.04111922e+00
1.79026335e-01 -1.81807727e-01 -2.06443623e-01 -1.87390029e-01
2.10448816e-01 -3.78928743e-02 1.81413695e-01 -7.51288414e-01
-4.16843832e-01 8.84050071e-01 -1.34216815e-01 -1.10509686e-01
1.62650609e+00 -3.91717315e-01 3.53340954e-01 5.43893516e-01
1.14436221e+00 -1.41461194e-01 -1.84631920e+00 -3.64010721e-01
-6.51358962e-01 -7.25333869e-01 -1.30065111e-02 -1.19499445e+00
-1.09440625e+00 9.23225164e-01 5.48912525e-01 -4.20207053e-01
1.01899791e+00 4.37026441e-01 4.39310044e-01 3.58618230e-01
6.68397725e-01 -6.08245373e-01 6.33450031e-01 6.75806880e-01
1.21566355e+00 -1.50985074e+00 2.98699975e-01 -4.23588127e-01
-4.94914353e-01 9.21559751e-01 8.71992052e-01 -4.44761366e-01
1.01546836e+00 -1.65095896e-01 -2.84551568e-02 -4.00374323e-01
-6.79277539e-01 -3.45161468e-01 5.31493723e-01 1.06516659e+00
5.71524836e-02 -1.97057441e-01 3.86117905e-01 4.62029040e-01
-1.21454805e-01 -2.78538108e-01 5.23081303e-01 7.44657099e-01
-3.99482518e-01 -8.06273758e-01 -2.68958688e-01 5.79908252e-01
-2.06982717e-01 9.51819308e-03 -2.28952199e-01 8.37441623e-01
2.00630710e-01 6.32131517e-01 1.72909215e-01 -3.28720063e-01
4.65748489e-01 -4.10041958e-02 9.46446180e-01 -1.36538708e+00
-3.12491387e-01 3.62831086e-01 9.89419073e-02 -6.55141413e-01
-9.08983648e-01 -1.05176628e+00 -8.35538507e-01 -4.27554213e-02
-2.47804359e-01 -3.82692814e-01 3.89664799e-01 8.89819205e-01
3.61651003e-01 2.74784535e-01 7.60199666e-01 -1.50205088e+00
-4.57598746e-01 -7.08323061e-01 -5.31403244e-01 5.76391041e-01
6.40276790e-01 -1.10399687e+00 -5.48926115e-01 2.07682893e-01] | [8.41846752166748, -3.038545608520508] |
87d57028-95ad-4169-a202-12dd1fe71fb4 | learning-6-dof-object-poses-to-grasp-category | 2205.04028 | null | https://arxiv.org/abs/2205.04028v1 | https://arxiv.org/pdf/2205.04028v1.pdf | Learning 6-DoF Object Poses to Grasp Category-level Objects by Language Instructions | This paper studies the task of any objects grasping from the known categories by free-form language instructions. This task demands the technique in computer vision, natural language processing, and robotics. We bring these disciplines together on this open challenge, which is essential to human-robot interaction. Critically, the key challenge lies in inferring the category of objects from linguistic instructions and accurately estimating the 6-DoF information of unseen objects from the known classes. In contrast, previous works focus on inferring the pose of object candidates at the instance level. This significantly limits its applications in real-world scenarios.In this paper, we propose a language-guided 6-DoF category-level object localization model to achieve robotic grasping by comprehending human intention. To this end, we propose a novel two-stage method. Particularly, the first stage grounds the target in the RGB image through language description of names, attributes, and spatial relations of objects. The second stage extracts and segments point clouds from the cropped depth image and estimates the full 6-DoF object pose at category-level. Under such a manner, our approach can locate the specific object by following human instructions, and estimate the full 6-DoF pose of a category-known but unseen instance which is not utilized for training the model. Extensive experimental results show that our method is competitive with the state-of-the-art language-conditioned grasp method. Importantly, we deploy our approach on a physical robot to validate the usability of our framework in real-world applications. Please refer to the supplementary for the demo videos of our robot experiments. | ['xiangyang xue', 'Yanwei Fu', 'Haitao Lin', 'Chilam Cheang'] | 2022-05-09 | null | null | null | null | ['robotic-grasping'] | ['robots'] | [ 1.39963003e-02 -3.71355861e-02 -9.15849730e-02 -4.18773443e-01
-5.53290963e-01 -6.29143119e-01 2.27023423e-01 -5.04335277e-02
-2.38364294e-01 2.70420104e-01 -2.79696822e-01 -3.69819347e-04
-2.64054507e-01 -6.38832092e-01 -1.00251913e+00 -6.86635375e-01
2.35892758e-02 8.45669091e-01 2.14154497e-01 -1.33054003e-01
5.64405203e-01 7.52620041e-01 -1.64844823e+00 2.79886574e-01
6.39404535e-01 1.12071967e+00 9.79891241e-01 1.50200054e-01
-1.39109448e-01 4.34282839e-01 -1.42761901e-01 1.13065094e-01
3.41060460e-01 2.11457238e-01 -6.87977612e-01 4.30806935e-01
2.68261552e-01 -7.80919433e-01 -2.90920824e-01 1.15510559e+00
6.22807071e-02 2.39652134e-02 6.52944446e-01 -1.32182395e+00
-5.59387088e-01 5.50487280e-01 -3.05622697e-01 -7.56924152e-01
6.90390527e-01 9.21949819e-02 6.97631299e-01 -1.17296374e+00
4.90349174e-01 1.57737327e+00 2.05505654e-01 5.80990016e-01
-7.43889093e-01 -5.36670804e-01 4.17521745e-01 2.54208595e-01
-1.41186428e+00 -1.38146877e-01 9.88207698e-01 -6.21087611e-01
4.71747994e-01 -1.13605119e-01 4.08822805e-01 1.03061092e+00
-7.99412951e-02 1.00876510e+00 9.34684515e-01 -4.75050211e-01
4.09820646e-01 -8.93566683e-02 1.97580919e-01 7.50874043e-01
3.49271834e-01 -4.86977547e-02 -3.46130520e-01 4.33407053e-02
1.10263491e+00 4.22995836e-01 -1.41561761e-01 -1.10399175e+00
-1.60021591e+00 4.46444094e-01 7.39271939e-01 1.56508878e-01
-6.60797000e-01 2.16445491e-01 3.55250500e-02 -2.35488072e-01
-3.39756273e-02 1.19532786e-01 -5.80332935e-01 2.64933974e-01
-1.43290013e-01 3.03045899e-01 8.30600083e-01 1.80334592e+00
9.09680426e-01 -6.29801214e-01 1.02094837e-01 4.89272833e-01
5.15993416e-01 7.79654503e-01 -4.77571338e-02 -1.01244259e+00
5.75453401e-01 7.54997969e-01 4.87121493e-01 -9.47210312e-01
-2.85871297e-01 8.25374424e-02 -3.95958543e-01 1.49636149e-01
4.31227475e-01 4.37904090e-01 -1.01368237e+00 1.41884959e+00
4.60797071e-01 -1.79247901e-01 -2.36731749e-02 1.36207545e+00
6.37777269e-01 3.45801562e-01 7.10288957e-02 -6.48375647e-03
1.47522509e+00 -8.20823371e-01 -5.80396116e-01 -5.14215887e-01
1.31110042e-01 -4.99565423e-01 1.11437047e+00 4.71025139e-01
-5.83178043e-01 -6.79847240e-01 -7.43156314e-01 -2.15053648e-01
-3.86951685e-01 6.62668169e-01 8.26600552e-01 -1.47449896e-01
-5.09320080e-01 2.71978766e-01 -1.05215597e+00 -5.79337299e-01
2.19945297e-01 4.30595368e-01 -6.00547016e-01 -6.08299255e-01
-5.61999321e-01 8.92803490e-01 8.35104644e-01 4.41360623e-01
-1.02677357e+00 -6.73415810e-02 -1.02910364e+00 -2.80580193e-01
8.55878413e-01 -4.26976502e-01 1.25068533e+00 -2.21317500e-01
-1.39773953e+00 9.33257580e-01 -1.28326148e-01 1.81691021e-01
3.92337292e-01 -6.58986807e-01 3.55866045e-01 2.39475906e-01
3.37941051e-01 7.67851472e-01 8.55140626e-01 -1.88021147e+00
-5.74865580e-01 -8.34109902e-01 4.47368562e-01 1.80463433e-01
1.28968462e-01 -3.96479338e-01 -7.03972816e-01 -1.97997332e-01
1.15532374e+00 -9.87271965e-01 -1.98458973e-02 4.96699363e-01
-3.57406914e-01 -6.12349451e-01 8.83149922e-01 -4.13884193e-01
1.98205397e-01 -2.16619968e+00 4.52880859e-01 -1.27172377e-02
-7.30141774e-02 -1.75970644e-01 -3.95997874e-02 3.47772092e-01
3.98175478e-01 -3.31072241e-01 -2.31777608e-01 -2.85207152e-01
2.48785347e-01 5.93038499e-01 -5.16431153e-01 4.88639772e-01
1.99203417e-01 7.50314474e-01 -1.25752211e+00 -4.78357881e-01
4.66286689e-01 3.66143465e-01 -5.25561869e-01 6.65618241e-01
-5.26910663e-01 7.61701822e-01 -1.01831782e+00 1.11151147e+00
8.46389651e-01 1.60074323e-01 1.21029198e-01 -5.66138268e-01
-2.21047625e-01 -7.07134092e-03 -1.18643796e+00 2.33360934e+00
-4.47209090e-01 -1.25351101e-01 2.26150960e-01 -1.01731408e+00
1.05796468e+00 -5.27021650e-04 2.87847281e-01 -2.12743264e-02
2.06133708e-01 4.70372558e-01 -2.73033440e-01 -1.03181148e+00
1.91948175e-01 1.88800484e-01 -2.60297418e-01 -6.52733892e-02
5.10866828e-02 -6.52646124e-01 -1.41983837e-01 -5.57934158e-02
7.00284004e-01 9.26176190e-01 2.29879349e-01 -1.47110179e-01
3.68006408e-01 2.02786848e-01 2.16239735e-01 7.01620936e-01
-1.05410382e-01 4.66190219e-01 -2.93321814e-02 -2.90081024e-01
-7.40051806e-01 -1.23221433e+00 -3.57375247e-04 8.45731437e-01
8.05859327e-01 8.85790661e-02 -5.84967613e-01 -6.10902727e-01
2.98826873e-01 6.89690113e-01 -4.37346697e-01 1.96621969e-01
-6.47094548e-01 4.06789500e-03 -2.14086667e-01 6.19459271e-01
5.64164460e-01 -1.31418264e+00 -1.03501558e+00 1.06133521e-01
-3.26493144e-01 -1.48499763e+00 -1.24788415e-02 1.47277206e-01
-7.98108160e-01 -1.24587584e+00 -6.34030044e-01 -1.27504873e+00
1.09011757e+00 4.21534717e-01 6.07294858e-01 -9.04207379e-02
-1.62261605e-01 6.93350554e-01 -6.47060215e-01 -3.83809954e-01
-8.07686821e-02 -1.85248882e-01 2.72211522e-01 -1.51548252e-01
4.98166084e-01 -3.49497437e-01 -5.09266078e-01 3.84224355e-01
-5.93152106e-01 8.04242268e-02 9.53572512e-01 4.90326792e-01
7.13747919e-01 -1.18789598e-01 2.45586485e-01 -7.37400353e-02
1.80195853e-01 -2.30102688e-01 -6.68186069e-01 3.00266892e-01
1.09498456e-01 8.42626989e-02 2.84673482e-01 -6.70028806e-01
-8.22029114e-01 7.47412026e-01 2.48445421e-01 -6.08670294e-01
-6.31503582e-01 5.27417302e-01 -5.35176933e-01 6.58289269e-02
1.95687965e-01 1.83531061e-01 -3.91284786e-02 -7.51128018e-01
4.73046035e-01 9.37843800e-01 8.88891637e-01 -1.00940299e+00
7.28711069e-01 5.73673189e-01 1.03549995e-01 -5.18455267e-01
-9.09347653e-01 -7.23495960e-01 -1.34151900e+00 -2.77163088e-01
7.81329572e-01 -9.02055919e-01 -1.22713196e+00 4.58448350e-01
-1.67220795e+00 -1.55436829e-01 1.66428968e-01 6.72285914e-01
-1.03125799e+00 3.69121313e-01 -3.53546619e-01 -1.01061881e+00
1.32898046e-02 -1.31220376e+00 1.70834303e+00 -4.01633680e-02
2.63334215e-01 -1.87595069e-01 -6.91278040e-01 2.31037438e-01
-1.04467012e-01 2.86431819e-01 8.26001227e-01 -3.79319757e-01
-9.56478953e-01 -2.65038580e-01 -4.12345946e-01 4.10197210e-03
3.78831297e-01 -5.72593331e-01 -6.31938100e-01 -1.55227661e-01
2.28738144e-01 -3.88586015e-01 5.14015317e-01 1.44249469e-01
1.26092434e+00 -1.53229693e-02 -5.92744470e-01 1.72656938e-01
1.37548554e+00 2.34774694e-01 3.29794168e-01 1.93266645e-01
9.04949129e-01 9.55107987e-01 1.38923287e+00 3.26849222e-01
4.98094052e-01 6.86438084e-01 1.00627553e+00 5.44795334e-01
1.74065351e-01 -3.69629800e-01 1.55360609e-01 6.94732070e-01
-1.78164274e-01 -1.91760827e-02 -1.06615126e+00 5.04791975e-01
-2.06693482e+00 -4.94114220e-01 -1.17739122e-02 1.99769044e+00
5.59165597e-01 -2.07025558e-02 -3.64651650e-01 7.62420765e-04
7.00949490e-01 -4.82706875e-01 -7.04289377e-01 3.38640064e-01
3.63317847e-01 -2.31796354e-01 3.06984633e-01 3.73935699e-01
-1.11477876e+00 1.29516888e+00 4.54306602e+00 4.22371030e-01
-8.63739014e-01 -1.95785031e-01 -2.80613631e-01 5.09838343e-01
2.42709339e-01 1.22768536e-01 -7.00644374e-01 1.03761442e-01
-1.58180237e-01 3.57165694e-01 6.01393163e-01 1.20856822e+00
8.60116929e-02 -2.36422405e-01 -1.78247571e+00 1.17827272e+00
2.32341558e-01 -5.04578471e-01 9.08179581e-02 -1.52470127e-01
2.66969427e-02 -2.48960629e-01 -2.56652564e-01 3.09833229e-01
-5.42729646e-02 -5.88763654e-01 1.30511069e+00 7.10545063e-01
7.37653315e-01 -2.82786548e-01 6.62752509e-01 9.06189740e-01
-1.13732052e+00 -3.52072984e-01 -5.69885552e-01 -1.95088640e-01
2.31817603e-01 2.12725222e-01 -9.14184093e-01 6.49499655e-01
6.67510033e-01 6.17131531e-01 -1.70104370e-01 8.88712049e-01
-6.14591002e-01 -3.84214260e-02 -3.59255493e-01 -2.09610805e-01
1.04256295e-01 -2.22724438e-01 4.86946523e-01 6.26344979e-01
4.07098264e-01 6.88172996e-01 6.64277315e-01 1.13083982e+00
3.05253446e-01 -2.01593354e-01 -5.89850426e-01 9.59257632e-02
5.64701378e-01 9.38691616e-01 -7.84254432e-01 -1.99330971e-01
-6.84250444e-02 1.03284013e+00 3.72774959e-01 2.39517882e-01
-5.89767873e-01 -3.17569017e-01 2.89218009e-01 -6.53202832e-02
3.10678899e-01 -8.32157731e-01 -2.26910725e-01 -1.14132428e+00
5.50973475e-01 -5.45771718e-01 -2.07333937e-01 -1.24274123e+00
-1.26587331e+00 3.87915105e-01 4.75409538e-01 -1.44945693e+00
-1.63988411e-01 -1.22556889e+00 5.33815958e-02 6.66034281e-01
-1.38615906e+00 -1.58868599e+00 -8.21227908e-01 3.80739421e-01
8.11029971e-01 2.46572062e-01 8.27199996e-01 -1.29883111e-01
5.10432459e-02 -1.14485189e-01 -4.04732019e-01 1.91107243e-01
4.59731400e-01 -8.67555618e-01 -2.57514957e-02 4.06181306e-01
-2.00958341e-01 8.33809197e-01 6.61659956e-01 -8.19129467e-01
-2.01196790e+00 -7.82126844e-01 5.39341748e-01 -6.26713276e-01
4.66107130e-01 -7.22726464e-01 -7.39923775e-01 7.81521380e-01
-5.06564260e-01 3.75088328e-03 -2.02429533e-01 -2.73818254e-01
-2.30391353e-01 9.52926949e-02 -1.29983187e+00 4.90766197e-01
1.56825972e+00 -3.98720980e-01 -9.89120543e-01 6.39744699e-01
9.61326480e-01 -7.28697598e-01 -6.75163925e-01 8.14781427e-01
7.23698080e-01 -4.35628682e-01 1.06111658e+00 -4.03945684e-01
3.81945133e-01 -5.97954452e-01 -7.06106961e-01 -9.17218149e-01
-1.16472170e-01 -2.84185410e-02 -2.30105221e-02 1.02172434e+00
-2.18411610e-01 -4.23170716e-01 6.12272739e-01 5.47646344e-01
-3.20939928e-01 -5.33417583e-01 -6.60230398e-01 -7.74038494e-01
-1.93010896e-01 -4.59833324e-01 5.65440834e-01 5.12742162e-01
4.16014455e-02 -3.92374247e-02 -4.15605456e-02 8.40003729e-01
9.05529082e-01 6.10834002e-01 9.42224205e-01 -1.30408657e+00
1.19294912e-01 1.60841811e-02 -6.16966486e-01 -1.65678644e+00
5.01633108e-01 -6.10587180e-01 8.78696382e-01 -1.73164260e+00
2.15447545e-01 -7.62459695e-01 4.98652309e-02 7.10824430e-01
6.93169087e-02 -1.89357892e-01 3.27864915e-01 4.60711777e-01
-5.13420641e-01 6.11651838e-01 1.36615646e+00 -3.58403832e-01
-4.83376086e-02 -6.23442382e-02 -2.51401603e-01 8.41257334e-01
6.15782320e-01 -4.18014765e-01 -1.64190263e-01 -7.65932322e-01
-3.13388348e-01 6.78119734e-02 8.80699813e-01 -1.00947368e+00
3.61528873e-01 -5.53934693e-01 9.62069407e-02 -1.03899789e+00
6.38309658e-01 -1.37689579e+00 -2.36942813e-01 5.72573960e-01
-1.27228543e-01 -3.58433455e-01 -1.33505166e-02 6.01083159e-01
9.66468528e-02 -4.11957264e-01 1.55593634e-01 -4.36863333e-01
-1.16484940e+00 3.11190546e-01 7.10690916e-02 -6.14050686e-01
1.16298878e+00 -1.71751395e-01 2.49300674e-02 -1.21905535e-01
-7.52673805e-01 2.95946091e-01 5.97910941e-01 6.55112028e-01
9.23176527e-01 -1.15208352e+00 -4.22758967e-01 3.07833433e-01
5.05577862e-01 7.19697356e-01 4.98821307e-03 4.73155349e-01
-5.22348821e-01 4.63674337e-01 -1.83340460e-01 -1.14427614e+00
-8.83137584e-01 9.81180131e-01 -1.19433977e-01 6.68390274e-01
-5.03006339e-01 5.59169650e-01 4.17420298e-01 -7.54493952e-01
6.24802947e-01 -6.56861126e-01 -2.77824644e-02 -5.24977028e-01
1.81488559e-01 3.59090380e-02 -2.01275125e-01 -7.57268608e-01
-5.94126999e-01 1.02433372e+00 2.42681593e-01 7.73058087e-02
1.28063667e+00 -1.85458228e-01 -4.18476641e-01 6.26965821e-01
9.43466127e-01 -2.66264051e-01 -1.36231673e+00 -2.89042413e-01
6.82195798e-02 -5.05648553e-01 -4.03192014e-01 -7.42771208e-01
-5.99848509e-01 9.82033968e-01 4.51862216e-01 -2.85077542e-01
8.97624314e-01 5.53672552e-01 3.96121889e-01 8.87804091e-01
1.41902363e+00 -6.94505811e-01 2.15673700e-01 5.46930313e-01
1.44998431e+00 -1.44193113e+00 -4.44204174e-02 -1.05916286e+00
-2.59554714e-01 1.24339688e+00 8.28470528e-01 -1.94012612e-01
5.76442540e-01 1.14929020e-01 8.65240861e-03 -2.91887730e-01
-5.44224493e-02 -1.99048936e-01 2.01049134e-01 7.79071331e-01
-1.86417073e-01 2.85685420e-01 -7.63682202e-02 6.41123593e-01
-3.47963274e-02 5.15257902e-02 1.07912786e-01 1.38540816e+00
-7.17225432e-01 -7.28838503e-01 -5.32385290e-01 -1.01718463e-01
1.82809815e-01 2.92865336e-01 -3.06664377e-01 7.72862911e-01
3.87984037e-01 1.02267921e+00 -3.55449170e-02 -3.44324976e-01
5.59336007e-01 -5.39058335e-02 8.90421450e-01 -8.22299302e-01
1.97945088e-01 -5.12550659e-02 -4.17530060e-01 -6.36557221e-01
-6.41330481e-01 -6.32942021e-01 -1.67763042e+00 3.99189591e-01
-5.01925170e-01 -1.74604595e-01 1.13180518e+00 1.10237610e+00
1.41846702e-01 1.28953770e-01 3.42968643e-01 -1.67790413e+00
-8.55672240e-01 -1.01904237e+00 -3.49228263e-01 5.31556964e-01
2.21225172e-01 -1.29676151e+00 -2.55520970e-01 1.13965638e-01] | [5.851699352264404, -0.8962050676345825] |
c2638808-77a5-4b6f-88d7-91d4a14a7ca6 | samaug-point-prompt-augmentation-for-segment | 2307.01187 | null | https://arxiv.org/abs/2307.01187v1 | https://arxiv.org/pdf/2307.01187v1.pdf | SAMAug: Point Prompt Augmentation for Segment Anything Model | This paper introduces SAMAug, a novel visual point augmentation method for the Segment Anything Model (SAM) that enhances interactive image segmentation performance. SAMAug generates augmented point prompts to provide more information to SAM. From the initial point prompt, SAM produces the initial mask, which is then fed into our proposed SAMAug to generate augmented point prompts. By incorporating these extra points, SAM can generate augmented segmentation masks based on the augmented point prompts and the initial prompt, resulting in improved segmentation performance. We evaluate four point augmentation techniques: random selection, maximum difference entropy, maximum distance, and a saliency model. Experiments on the COCO, Fundus, and Chest X-ray datasets demonstrate that SAMAug can boost SAM's segmentation results, especially using the maximum distance and saliency model methods. SAMAug underscores the potential of visual prompt engineering to advance interactive computer vision models. | ['Xiang Li', 'Tianming Liu', 'Quanzheng Li', 'Wei Liu', 'Dajiang Zhu', 'Zihao Wu', 'Lin Zhao', 'Xiaozheng Wei', 'Peng Shu', 'Yiwei Li', 'Zhengliang Liu', 'Chong Ma', 'Haixing Dai'] | 2023-07-03 | null | null | null | null | ['prompt-engineering'] | ['natural-language-processing'] | [ 5.97737789e-01 3.15062314e-01 -3.69541913e-01 -1.97648600e-01
-6.92552626e-01 -6.46062553e-01 1.37721315e-01 2.72300094e-01
-1.07709289e-01 3.33299816e-01 8.14431533e-02 -4.99163181e-01
4.45080251e-01 -4.54413772e-01 -5.20570219e-01 -3.90472621e-01
2.01462820e-01 2.11895525e-01 7.34665096e-01 -1.51227146e-01
6.23242378e-01 3.55991721e-01 -1.22245073e+00 3.52769554e-01
1.56958044e+00 7.77328551e-01 6.24950826e-01 7.54851162e-01
-4.32602108e-01 2.31007725e-01 -6.34025514e-01 7.55733848e-02
3.02575231e-01 -3.16254526e-01 -9.49821234e-01 1.89137027e-01
3.13609183e-01 -5.15451014e-01 1.53118968e-01 8.20917487e-01
5.11100769e-01 2.55989164e-01 6.38550043e-01 -1.31368852e+00
-5.53779542e-01 4.40173060e-01 -1.04493356e+00 3.75757307e-01
4.77622628e-01 3.04217428e-01 7.62472987e-01 -8.34726453e-01
6.64467335e-01 9.63989258e-01 5.71937501e-01 4.75979745e-01
-1.15764856e+00 -3.59892428e-01 1.84116125e-01 1.45083591e-02
-8.68756413e-01 1.56797349e-01 8.32643867e-01 -3.44624937e-01
5.50880313e-01 6.85046554e-01 9.60741997e-01 5.47988653e-01
-1.29951075e-01 1.30385649e+00 1.14429104e+00 -5.53231955e-01
1.25770465e-01 1.54341117e-01 4.71951753e-01 7.68569350e-01
2.07374953e-02 7.66181201e-02 -4.34719115e-01 -1.51591659e-01
1.33705306e+00 -2.62101889e-01 -3.17564547e-01 -1.10179037e-01
-1.26957119e+00 6.83030367e-01 6.96924269e-01 3.52795720e-02
-4.87450033e-01 -1.81400403e-01 1.20651796e-01 -4.30034131e-01
4.19459701e-01 8.13249826e-01 -2.75206804e-01 6.64050877e-02
-1.09477150e+00 2.05753133e-01 2.62641191e-01 1.02308369e+00
7.17399120e-01 6.29094467e-02 -7.51232088e-01 8.09261322e-01
3.04992408e-01 5.00482142e-01 3.51962775e-01 -1.02381003e+00
1.62128896e-01 9.63717937e-01 4.00006697e-02 -1.03419483e+00
-4.10926074e-01 -4.00969505e-01 -3.92724246e-01 3.24134946e-01
5.88663928e-02 -1.69705078e-01 -1.40150893e+00 1.33832455e+00
5.20120680e-01 4.21662331e-01 -3.04028809e-01 9.50375319e-01
1.32455754e+00 6.11648798e-01 1.71530008e-01 -2.69591287e-02
1.33442008e+00 -1.25798047e+00 -6.62791431e-01 -3.23251963e-01
5.01729429e-01 -8.67083549e-01 1.54973471e+00 9.93132964e-02
-1.32503068e+00 -7.70973623e-01 -9.88920331e-01 -1.48565739e-01
-1.95574742e-02 3.37779582e-01 6.34303391e-01 4.30095404e-01
-1.24764204e+00 2.86029220e-01 -8.00527990e-01 -4.34970222e-02
6.18777454e-01 3.71988058e-01 2.71161020e-01 2.77415425e-01
-7.58535326e-01 5.66865325e-01 3.05834085e-01 -3.67717803e-01
-2.64201373e-01 -8.97491693e-01 -9.56522286e-01 9.85617787e-02
2.10506067e-01 -6.45374119e-01 1.42742383e+00 -8.91152084e-01
-1.09648049e+00 5.49144149e-01 -5.55992067e-01 -4.05952483e-01
5.17146587e-01 -8.98562819e-02 1.85796052e-01 7.10641861e-01
3.21085185e-01 1.52380919e+00 8.97627652e-01 -1.68787086e+00
-5.79199553e-01 -4.10864055e-02 -1.18078262e-01 5.59659362e-01
3.72896087e-03 -1.97231740e-01 -7.73843646e-01 -8.03499639e-01
3.35151553e-01 -9.43510056e-01 -5.85836232e-01 -6.28857762e-02
-8.83266211e-01 -3.52283269e-02 1.21789134e+00 -7.20406413e-01
1.21363187e+00 -2.21089721e+00 -2.58568019e-01 4.03071344e-01
3.94928098e-01 5.74317157e-01 -2.81712085e-01 -1.20144971e-01
-2.09935263e-01 4.01567578e-01 -3.44704092e-01 -2.87000686e-01
-5.38158238e-01 -1.01760566e-01 -1.10908516e-01 -2.57300586e-01
2.50580490e-01 1.45336294e+00 -9.15292501e-01 -9.33253765e-01
6.30024076e-01 2.69250125e-01 -6.37832224e-01 1.15066424e-01
-4.87282664e-01 5.95938683e-01 -3.96939635e-01 6.36796772e-01
7.10341692e-01 -6.58578932e-01 -3.51414561e-01 -2.52368927e-01
-1.27484903e-01 -1.12189069e-01 -8.87818098e-01 1.45696020e+00
1.54747069e-01 7.45972693e-01 -4.81169879e-01 -2.79947817e-01
9.33981836e-01 8.82389843e-02 6.23416066e-01 -4.79900241e-01
1.23970613e-01 -1.38789609e-01 -1.30191118e-01 -3.63961786e-01
8.47852468e-01 4.52506334e-01 3.38748157e-01 4.45204824e-01
-3.76872003e-01 -3.94769400e-01 8.16970784e-03 8.89678955e-01
6.56295538e-01 2.23384231e-01 2.16556221e-01 1.84078589e-02
3.25030863e-01 4.37653452e-01 2.46661037e-01 6.69502914e-01
-2.97755480e-01 1.17364931e+00 4.29649740e-01 -7.24223480e-02
-9.80584443e-01 -1.00469398e+00 1.21996246e-01 7.21736610e-01
7.40529537e-01 -3.41327906e-01 -1.04924297e+00 -9.63677108e-01
-2.87030369e-01 8.51672351e-01 -6.78822160e-01 -1.12481415e-01
-3.32306385e-01 -4.11822468e-01 1.47478595e-01 8.23680818e-01
8.55144799e-01 -1.21510804e+00 -8.55121434e-01 -2.04318352e-02
-5.74783087e-01 -8.94924283e-01 -1.07512414e+00 -3.51356715e-01
-1.02882159e+00 -1.03252530e+00 -1.01613140e+00 -9.63803411e-01
1.30046165e+00 7.33102858e-01 7.51326144e-01 4.81293321e-01
-3.04592162e-01 2.95058459e-01 -5.34062862e-01 -5.41063964e-01
-3.37462485e-01 2.37010047e-02 -4.90497917e-01 -4.17860121e-01
-1.06264748e-01 -9.77808386e-02 -9.68141377e-01 3.80894363e-01
-8.99955094e-01 8.73458028e-01 6.39392674e-01 6.46614969e-01
8.73928607e-01 -6.29390717e-01 6.03981435e-01 -9.71484780e-01
5.81468582e-01 -5.15130050e-02 -3.17374527e-01 3.32202949e-02
-5.49297094e-01 -1.99773729e-01 1.43573463e-01 -5.66462278e-01
-1.19378543e+00 1.69260234e-01 -1.37622342e-01 -3.29392463e-01
-9.39457119e-02 3.44429404e-01 2.46491536e-01 -3.39488834e-01
7.72082150e-01 2.18009889e-01 2.03280300e-01 -2.44179696e-01
5.03980935e-01 4.89459068e-01 7.91745007e-01 -2.27809921e-01
5.58491409e-01 2.61540592e-01 -3.94222677e-01 -7.79583573e-01
-5.78558385e-01 -5.63318908e-01 -7.57200956e-01 -3.66956562e-01
9.45957243e-01 -3.41205418e-01 -4.20255512e-01 2.85442501e-01
-1.17228651e+00 -3.08670551e-01 -4.76513505e-01 1.01336725e-01
-4.69006717e-01 6.67939305e-01 -5.07952571e-01 -5.86192012e-01
-5.22082806e-01 -1.46082711e+00 1.20492327e+00 8.86239290e-01
-4.55711007e-01 -9.17379439e-01 -4.58679944e-01 4.84433502e-01
1.26950800e-01 4.46822733e-01 1.00777876e+00 -5.98851979e-01
-7.94357657e-01 -8.08072090e-02 -4.89554465e-01 1.84623197e-01
9.54514295e-02 2.86409765e-01 -6.68046832e-01 7.49275833e-02
-4.34878081e-01 2.39568681e-01 5.39071202e-01 9.91679907e-01
1.35631955e+00 -1.43903479e-01 -6.44072711e-01 5.94110072e-01
1.01222944e+00 7.31136680e-01 6.17515266e-01 3.36931467e-01
7.83666730e-01 2.40989402e-01 1.03519642e+00 1.89270213e-01
4.53410953e-01 4.92655039e-01 4.08177257e-01 -9.90460932e-01
-3.72342139e-01 -2.61059523e-01 -2.04518169e-01 5.84959447e-01
-1.34696634e-02 1.28331378e-01 -9.27741587e-01 6.09647632e-01
-1.77763546e+00 -5.56448758e-01 -5.58092773e-01 1.89683509e+00
8.82973909e-01 6.23907484e-02 2.02144563e-01 -5.59574328e-02
7.49673545e-01 -1.51878530e-02 -6.69968426e-01 -3.25182647e-01
5.04508950e-02 1.68894261e-01 2.83505678e-01 5.02686381e-01
-9.82315063e-01 1.19092798e+00 6.88342237e+00 7.81783283e-01
-1.24983323e+00 -2.31516302e-01 1.10001373e+00 2.44804144e-01
-5.32935977e-01 -3.99795361e-03 -6.05225086e-01 6.01700485e-01
1.94889218e-01 -3.71895820e-01 8.50127917e-03 1.01754141e+00
4.85854328e-01 -5.79944491e-01 -7.04134405e-01 8.31558287e-01
7.61887580e-02 -1.71943033e+00 2.38708496e-01 -8.52894690e-03
1.05626023e+00 -3.66474122e-01 4.38124657e-01 -1.58067774e-02
2.05953464e-01 -8.58737767e-01 3.19745362e-01 2.71690071e-01
8.65178525e-01 -6.89391851e-01 5.78708887e-01 1.68521881e-01
-1.04305756e+00 3.10409635e-01 6.21269569e-02 4.99394119e-01
4.77249265e-01 2.39430189e-01 -1.56130135e+00 5.55933893e-01
3.53198647e-01 4.03229147e-01 -9.60193694e-01 1.51565504e+00
-1.61213726e-01 7.87716985e-01 -2.20920384e-01 7.38750352e-03
2.51781970e-01 -1.69603303e-01 7.75719583e-01 1.05451107e+00
-8.56471509e-02 2.93095231e-01 2.85251081e-01 1.02718806e+00
3.42617631e-01 2.98362345e-01 -3.70721638e-01 2.44051859e-01
6.46360278e-01 1.29731786e+00 -1.28294325e+00 -6.57740951e-01
-1.33089393e-01 1.09887087e+00 -2.75117218e-01 6.65426672e-01
-9.66280460e-01 -4.26853418e-01 1.29295439e-01 2.05364510e-01
6.69732690e-02 -1.79687604e-01 -1.24993229e+00 -5.59729040e-01
-2.64094144e-01 -7.19777405e-01 1.17216207e-01 -1.26108325e+00
-4.90794510e-01 7.36326039e-01 4.29492027e-01 -1.20274603e+00
-1.91989720e-01 -1.44527049e-03 -1.00193655e+00 9.68361735e-01
-1.16622937e+00 -1.14907146e+00 -6.55611932e-01 2.39859939e-01
9.77505863e-01 2.32190058e-01 3.65113378e-01 -1.43029362e-01
-3.98850292e-01 5.72653592e-01 -5.32246292e-01 -4.32578586e-02
3.61103892e-01 -1.52534747e+00 9.32358623e-01 8.97740781e-01
-6.02005124e-02 5.44588864e-01 6.17504656e-01 -1.04746187e+00
-5.47564864e-01 -1.06102192e+00 3.25480223e-01 -3.87778074e-01
-1.64047495e-01 3.98028582e-01 -1.09007919e+00 5.64288974e-01
2.06235766e-01 -3.24965388e-01 4.76632327e-01 -4.10842329e-01
3.56136680e-01 5.16666532e-01 -1.33739126e+00 1.06189346e+00
7.23572493e-01 4.24139202e-02 -6.60808146e-01 2.39979327e-01
1.17799592e+00 -9.46736455e-01 -2.87748218e-01 5.62416613e-01
3.31903756e-01 -7.23109245e-01 1.19625461e+00 -1.69472367e-01
5.34753382e-01 -3.92839521e-01 5.89587808e-01 -1.42663038e+00
-2.98517853e-01 -5.93194783e-01 2.32946463e-02 1.13729823e+00
5.05070329e-01 -4.15509313e-01 9.32320118e-01 8.69490266e-01
-3.69718790e-01 -9.83676016e-01 -4.29315180e-01 -2.41018951e-01
-4.87905085e-01 -4.29286212e-01 7.35121906e-01 8.17189038e-01
-9.30569097e-02 1.08965933e-01 4.66450229e-02 1.16117045e-01
4.36287791e-01 7.30738640e-02 6.95278525e-01 -1.05538368e+00
1.60562117e-02 -3.07300895e-01 -6.53730407e-02 -1.47793353e+00
-3.81518334e-01 -6.69883788e-01 -4.67660651e-02 -2.01179552e+00
1.98631331e-01 -6.00137174e-01 1.82094589e-01 6.70001805e-01
-9.84249830e-01 5.40674508e-01 3.44231606e-01 5.80287799e-02
-3.65774035e-01 3.21190089e-01 1.95450246e+00 -5.02107963e-02
-8.43707561e-01 1.44083053e-01 -9.95241821e-01 8.63709033e-01
8.49674046e-01 9.92264003e-02 -7.38965988e-01 -9.73763764e-02
-3.54634672e-01 -4.98162061e-02 2.91421354e-01 -8.20472658e-01
6.52604625e-02 -2.30171770e-01 6.21850073e-01 -1.27880001e+00
2.64053255e-01 -4.17744219e-01 -2.94376999e-01 3.77196699e-01
-3.33825678e-01 7.22331703e-02 4.70305026e-01 2.25745469e-01
3.96331213e-02 -2.56797224e-01 8.73230159e-01 -4.93140556e-02
-8.06357503e-01 2.42174834e-01 -1.98148474e-01 -5.01946546e-02
1.33523369e+00 -8.65625739e-01 -3.03478181e-01 -4.61535871e-01
-1.01337039e+00 5.47539234e-01 5.05237639e-01 3.08756948e-01
1.02834046e+00 -1.23822939e+00 -3.92974168e-01 4.91114467e-01
-1.48166716e-01 4.55856293e-01 4.22029316e-01 9.61225927e-01
-6.87211633e-01 4.54733893e-02 -2.16718808e-01 -9.61192310e-01
-1.79731882e+00 2.86224782e-01 -1.35804534e-01 3.60243134e-02
-7.93852568e-01 9.46606278e-01 4.53030080e-01 2.89795501e-03
5.90591580e-02 -7.35555947e-01 -4.60684925e-01 -2.52108604e-01
5.08257329e-01 3.83789659e-01 -3.03771615e-01 -4.27488685e-01
-2.51450241e-02 4.99748617e-01 -4.15965289e-01 -4.92394239e-01
7.30745494e-01 -1.63355827e-01 2.81799972e-01 4.68827486e-02
6.29391432e-01 8.57394189e-03 -1.33247626e+00 8.08480605e-02
-2.50378400e-01 -8.04226458e-01 -6.89474866e-02 -1.07682335e+00
-9.11974430e-01 6.69164956e-01 5.78075171e-01 2.03310266e-01
1.20547748e+00 9.66956541e-02 1.03964603e+00 -3.74010682e-01
2.41328627e-01 -8.43829155e-01 5.06568253e-01 2.58367389e-01
8.83216143e-01 -1.10926437e+00 -1.56763673e-01 -1.03929007e+00
-1.32935703e+00 7.09538102e-01 1.06805921e+00 2.76622064e-02
2.59952277e-01 1.71596259e-01 3.23506832e-01 -4.21868861e-02
-3.00687373e-01 -2.16962740e-01 7.54393339e-01 9.83343601e-01
2.21313909e-01 -8.61764550e-02 -6.41110778e-01 4.46416914e-01
-3.48669201e-01 -1.80274583e-02 5.47032952e-01 8.14529836e-01
-5.86040199e-01 -9.88303900e-01 -5.53378940e-01 8.40805590e-01
-1.02142893e-01 -2.92152762e-01 -5.69755495e-01 5.91305494e-01
7.21208304e-02 8.50619018e-01 1.85803417e-02 -4.14970815e-01
1.40715539e-01 -1.44686356e-01 1.10058673e-01 -8.43324840e-01
-6.26750767e-01 2.77413636e-01 -1.15246862e-01 -4.88505810e-01
-6.33646399e-02 -5.41773021e-01 -1.87970543e+00 1.45967156e-01
-6.15887284e-01 1.76845446e-01 6.55067384e-01 7.51655877e-01
5.27523577e-01 7.96555161e-01 4.83077705e-01 -1.03802347e+00
6.32005706e-02 -8.82761359e-01 -3.46791446e-02 3.97903264e-01
2.48135716e-01 -5.16544044e-01 7.49988332e-02 2.86487371e-01] | [9.558037757873535, -0.18225005269050598] |
fcc6518a-63d0-4949-a957-ba1a94cf6ebb | neumap-neural-coordinate-mapping-by-auto | 2211.11177 | null | https://arxiv.org/abs/2211.11177v2 | https://arxiv.org/pdf/2211.11177v2.pdf | NeuMap: Neural Coordinate Mapping by Auto-Transdecoder for Camera Localization | This paper presents an end-to-end neural mapping method for camera localization, dubbed NeuMap, encoding a whole scene into a grid of latent codes, with which a Transformer-based auto-decoder regresses 3D coordinates of query pixels. State-of-the-art feature matching methods require each scene to be stored as a 3D point cloud with per-point features, consuming several gigabytes of storage per scene. While compression is possible, performance drops significantly at high compression rates. Conversely, coordinate regression methods achieve high compression by storing scene information in a neural network but suffer from reduced robustness. NeuMap combines the advantages of both approaches by utilizing 1) learnable latent codes for efficient scene representation and 2) a scene-agnostic Transformer-based auto-decoder to infer coordinates for query pixels. This scene-agnostic network design learns robust matching priors from large-scale data and enables rapid optimization of codes for new scenes while keeping the network weights fixed. Extensive evaluations on five benchmarks show that NeuMap significantly outperforms other coordinate regression methods and achieves comparable performance to feature matching methods while requiring a much smaller scene representation size. For example, NeuMap achieves 39.1% accuracy in the Aachen night benchmark with only 6MB of data, whereas alternative methods require 100MB or several gigabytes and fail completely under high compression settings. The codes are available at https://github.com/Tangshitao/NeuMap | ['Yasutaka Furukawa', 'Ping Tan', 'Andrea Tagliasacchi', 'Sicong Tang', 'Shitao Tang'] | 2022-11-21 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tang_NeuMap_Neural_Coordinate_Mapping_by_Auto-Transdecoder_for_Camera_Localization_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tang_NeuMap_Neural_Coordinate_Mapping_by_Auto-Transdecoder_for_Camera_Localization_CVPR_2023_paper.pdf | cvpr-2023-1 | ['camera-localization'] | ['computer-vision'] | [ 2.23585851e-02 -2.91457415e-01 -2.04698026e-01 -5.15229642e-01
-1.20984709e+00 -5.34762204e-01 4.65600580e-01 -2.68192627e-02
-6.00683928e-01 7.82007799e-02 1.59945324e-01 -3.50132063e-02
5.62107675e-02 -8.00207257e-01 -1.14686966e+00 -5.89869618e-01
4.70161065e-02 4.68341827e-01 1.84492469e-01 2.70768970e-01
2.81814873e-01 6.57278717e-01 -1.50760531e+00 -7.08557665e-02
4.50127333e-01 1.28336751e+00 4.14174378e-01 8.19076300e-01
1.55855715e-02 5.87587237e-01 -2.71493584e-01 -2.64975339e-01
7.41485953e-01 1.66281044e-01 -4.01255935e-01 -9.52094793e-02
1.14795125e+00 -4.77870971e-01 -1.00946629e+00 1.09498596e+00
4.87380058e-01 5.51325344e-02 2.08083987e-01 -1.20075476e+00
-7.25304067e-01 2.17640087e-01 -5.81886172e-01 -1.79394763e-02
2.41646662e-01 1.01008102e-01 9.09711599e-01 -1.00055921e+00
3.71921748e-01 9.39120412e-01 1.15078402e+00 2.48662457e-01
-1.29219377e+00 -7.87322402e-01 -2.41287239e-02 4.94470820e-03
-1.93582344e+00 -6.02033079e-01 7.14266717e-01 -3.18812847e-01
1.42418051e+00 1.17898405e-01 6.46362484e-01 5.60444474e-01
8.09471011e-02 5.10432839e-01 4.79030490e-01 -1.00024063e-02
1.55340567e-01 -2.10154831e-01 -1.74734622e-01 1.01955986e+00
6.90077692e-02 2.28156298e-01 -8.43071997e-01 -2.11099237e-01
9.01522577e-01 3.14860582e-01 -1.94125384e-01 -7.26796091e-01
-1.30554521e+00 8.92960906e-01 1.08355463e+00 -2.09044203e-01
-3.69448125e-01 8.37890148e-01 2.85636276e-01 2.05242485e-01
3.78881842e-01 2.06596613e-01 -3.22406173e-01 -2.44266734e-01
-1.17149591e+00 7.23374411e-02 5.56093097e-01 1.38714528e+00
1.18310654e+00 -6.68561608e-02 2.01879814e-01 7.17210352e-01
2.70575374e-01 8.67320240e-01 3.57397258e-01 -1.12990987e+00
8.08489978e-01 4.56824869e-01 -1.75395623e-01 -1.39693439e+00
-3.89412224e-01 -4.47839975e-01 -7.77918696e-01 6.10211827e-02
9.66249257e-02 9.04933289e-02 -9.84706044e-01 1.58229768e+00
1.56380087e-01 3.73093158e-01 -7.89199471e-02 1.08239210e+00
6.05904818e-01 8.78189027e-01 -2.59519458e-01 5.72136700e-01
1.10879862e+00 -1.12087178e+00 2.01334488e-02 -4.93472338e-01
5.95605433e-01 -7.83779860e-01 8.80642533e-01 3.55010144e-02
-1.01095283e+00 -5.77775598e-01 -1.25007105e+00 -4.91101891e-01
-4.04468358e-01 2.77509362e-01 8.19135547e-01 3.78633231e-01
-1.31789410e+00 3.07769507e-01 -1.03316367e+00 -4.41959918e-01
4.71998423e-01 6.55392170e-01 -5.59534609e-01 -4.48489010e-01
-5.80165386e-01 5.85316181e-01 1.13666803e-01 3.96981789e-03
-7.17306495e-01 -9.71171677e-01 -1.27970624e+00 1.24246486e-01
-1.87436379e-02 -7.22182393e-01 1.23135805e+00 -5.05696714e-01
-1.21936655e+00 7.36219108e-01 -2.46235996e-01 -6.03540242e-01
3.41467410e-01 -3.13145310e-01 -8.87940526e-02 2.76855767e-01
3.40061635e-01 1.21976423e+00 7.32136786e-01 -8.08313429e-01
-6.96997821e-01 -3.21969837e-01 -2.82944083e-01 2.30660409e-01
-3.14235896e-01 -2.43215367e-01 -1.16064453e+00 -4.94403899e-01
6.91661537e-01 -1.05143440e+00 -2.32586905e-01 6.37158692e-01
-3.58927757e-01 1.68926671e-01 1.03872716e+00 -3.89062852e-01
6.01329565e-01 -2.29424620e+00 -1.34223834e-01 1.99329466e-01
3.53829890e-01 -2.67994672e-01 -3.10215622e-01 1.51281849e-01
2.58525629e-02 -2.22193062e-01 -2.13546470e-01 -6.67850077e-01
4.90386970e-02 9.98419672e-02 -4.57092583e-01 9.03978229e-01
1.06075943e-01 9.58791971e-01 -6.62382126e-01 -1.61821246e-01
4.88615751e-01 8.58786345e-01 -7.48232901e-01 -2.40653530e-02
1.54700384e-01 -6.16280325e-02 -2.37828627e-01 9.12608504e-01
8.52429211e-01 -5.22162020e-01 -3.90822947e-01 -4.16074604e-01
-2.69253016e-01 3.28571290e-01 -1.00289392e+00 2.35661960e+00
-5.37492871e-01 1.11599064e+00 -9.81502756e-02 -7.21856654e-01
1.08843803e+00 -1.43477112e-01 6.54625952e-01 -7.70585775e-01
-6.19724579e-02 8.38052556e-02 -7.28370786e-01 1.13527909e-01
7.13899136e-01 3.87307107e-01 -3.78998160e-01 1.63976610e-01
1.51943946e-02 -4.75117326e-01 -1.38478175e-01 2.18789548e-01
1.35570788e+00 6.86940476e-02 -1.59909859e-01 2.28878986e-02
1.39311343e-01 2.95412213e-01 5.96294940e-01 8.28129530e-01
1.15604334e-01 8.36566329e-01 7.89689571e-02 -7.29370773e-01
-1.21085954e+00 -1.17619383e+00 -1.23050570e-01 7.45350659e-01
4.57880735e-01 -5.53989947e-01 -4.96501803e-01 -2.36107185e-01
4.24440295e-01 2.63435096e-01 -3.24630827e-01 -1.80097058e-01
-4.98898566e-01 -4.37260062e-01 6.69418395e-01 6.87129021e-01
8.61703932e-01 -3.79720032e-01 -7.46357262e-01 1.25503689e-01
-7.72347227e-02 -1.28768098e+00 -6.09287679e-01 3.65925193e-01
-8.21609676e-01 -9.02419806e-01 -5.81398427e-01 -7.38873482e-01
9.40107107e-01 7.15768754e-01 1.10984826e+00 -3.64130214e-02
-5.24992585e-01 4.32410568e-01 -3.13442238e-02 -7.41695687e-02
1.27446651e-01 2.36629725e-01 1.41690835e-01 -3.24433565e-01
6.18493557e-01 -7.24257231e-01 -5.51350355e-01 3.72452289e-01
-5.71661949e-01 2.11252451e-01 7.09700942e-01 7.11563766e-01
1.01449656e+00 -2.10018799e-01 -2.51640648e-01 -4.53179270e-01
6.42591417e-02 -3.59525621e-01 -1.36049676e+00 -8.80062133e-02
-4.58963692e-01 2.08652973e-01 4.56533700e-01 -3.51836473e-01
-4.11609828e-01 7.49949634e-01 3.56687345e-02 -8.90550435e-01
-3.40246828e-03 2.85401016e-01 1.05037749e-01 -6.42616987e-01
6.92063749e-01 3.72733265e-01 -5.67598548e-03 -4.09327269e-01
3.76068175e-01 4.30118024e-01 1.04706156e+00 -4.24098402e-01
1.33258355e+00 4.99120533e-01 5.08872531e-02 -5.66507638e-01
-7.13914454e-01 -6.79087281e-01 -7.33077288e-01 -9.86127183e-03
7.70681381e-01 -1.51799047e+00 -6.29729271e-01 3.97926450e-01
-1.08859277e+00 -4.33029920e-01 -1.90814376e-01 6.23624444e-01
-6.93837821e-01 1.11736372e-01 -5.86938620e-01 -1.66364759e-01
-3.62648278e-01 -1.14537942e+00 1.36007500e+00 2.57434487e-01
4.50638309e-02 -5.09533823e-01 -6.72148913e-02 1.76897690e-01
5.24857163e-01 1.79478973e-01 4.80473161e-01 -3.90653200e-02
-1.27664196e+00 -6.68411911e-01 -6.34070873e-01 -4.62020934e-02
-4.77196947e-02 -2.66444892e-01 -1.06113982e+00 -3.76746207e-01
-3.44863594e-01 -2.77921379e-01 8.31218064e-01 2.77854115e-01
1.27710891e+00 -2.25343019e-01 -3.79538804e-01 1.60099709e+00
1.78523982e+00 -1.74548656e-01 3.02691907e-01 4.01291877e-01
8.62448096e-01 -1.26717808e-02 4.48837340e-01 4.37329620e-01
5.78614950e-01 8.30033123e-01 6.39557898e-01 -3.08962226e-01
-1.87905356e-01 -5.12912750e-01 2.25405321e-01 7.99614429e-01
3.46896678e-01 3.96255217e-02 -1.07977653e+00 4.21349913e-01
-1.83900166e+00 -6.90589607e-01 1.05193280e-01 2.21025348e+00
4.23325598e-01 6.11495078e-02 -2.46298507e-01 -2.83540308e-01
4.93684918e-01 2.80463129e-01 -9.62052703e-01 1.28097683e-01
-2.38139167e-01 -1.21491902e-01 1.35322988e+00 4.95497733e-01
-1.25218081e+00 1.03532410e+00 5.75494146e+00 3.62882555e-01
-1.31851709e+00 6.73058704e-02 5.15588820e-01 -5.17573655e-01
-4.27986588e-03 1.20844826e-01 -9.81527269e-01 2.90827751e-01
8.51384044e-01 6.57786988e-03 5.58652818e-01 1.30575967e+00
-2.13826552e-01 -9.05353427e-02 -1.07491243e+00 1.66813886e+00
2.61914879e-01 -1.83149099e+00 -1.76866382e-01 1.74071491e-01
6.06760859e-01 1.13558972e+00 9.44074988e-02 -2.50226222e-02
2.46389925e-01 -7.85079062e-01 9.55604494e-01 5.53262174e-01
1.08826590e+00 -7.72724450e-01 5.36486804e-01 1.25215620e-01
-1.45657480e+00 -2.27749065e-01 -9.06417489e-01 1.20847516e-01
5.19335121e-02 4.78079885e-01 -7.04290271e-01 1.63841292e-01
1.03032887e+00 9.10339355e-01 -8.02394569e-01 1.33627832e+00
1.84065867e-02 2.46218711e-01 -8.52336228e-01 2.05282792e-01
4.26842034e-01 8.21865499e-02 3.08393180e-01 1.11487865e+00
6.49291754e-01 -9.52483863e-02 1.65552914e-01 9.33884025e-01
-2.78969139e-01 -2.89437979e-01 -8.04243803e-01 3.12539369e-01
8.78480017e-01 1.14025927e+00 -5.98966300e-01 -1.81868881e-01
-6.00470364e-01 1.21642494e+00 4.22047436e-01 2.30234131e-01
-8.40534568e-01 -5.50548375e-01 8.82048845e-01 2.82422528e-02
4.66821492e-01 -6.57006204e-01 -3.81513447e-01 -1.07966411e+00
8.36667325e-03 -3.46013874e-01 1.49146751e-01 -9.86963689e-01
-9.63265300e-01 6.14733934e-01 -1.46878019e-01 -1.44078112e+00
-2.30627134e-01 -6.08307958e-01 -2.92028517e-01 8.31702232e-01
-1.60072803e+00 -1.20843792e+00 -8.75281811e-01 7.72299826e-01
4.53461766e-01 -3.12910289e-01 7.35762239e-01 4.50026989e-01
-4.21379209e-01 7.83663929e-01 1.64784670e-01 4.17910278e-01
5.67665279e-01 -1.06132972e+00 1.03163850e+00 9.27052319e-01
4.93870407e-01 4.48828429e-01 3.45517159e-01 -4.18492407e-01
-1.95860660e+00 -1.26807344e+00 8.90821218e-01 -4.65920299e-01
4.38485563e-01 -6.85373366e-01 -6.38772726e-01 6.13613129e-01
-1.62407264e-01 5.05772710e-01 4.39796627e-01 4.90990058e-02
-6.89015865e-01 -4.75921422e-01 -9.68320966e-01 4.22541410e-01
1.10927689e+00 -9.37703073e-01 -2.36635190e-02 6.00439966e-01
9.37301815e-01 -7.99523830e-01 -8.47040772e-01 1.57677636e-01
3.86031210e-01 -8.40770125e-01 1.21043003e+00 1.46091551e-01
6.94650039e-02 -5.68741441e-01 -5.84533811e-01 -7.66084909e-01
-5.59238315e-01 -4.76319283e-01 -9.51189175e-02 7.78163850e-01
3.57034475e-01 -6.26961768e-01 1.15927088e+00 9.18414116e-01
-2.88683623e-01 -6.90027475e-01 -9.72296536e-01 -6.74156964e-01
-3.71302724e-01 -6.62070811e-01 8.06462884e-01 8.25123131e-01
-4.32781816e-01 8.47402588e-02 -2.79151231e-01 6.14421546e-01
8.04085970e-01 1.73044711e-01 1.14147556e+00 -9.20969844e-01
-9.08113867e-02 -4.18645114e-01 -9.06220436e-01 -1.52918065e+00
1.00269012e-01 -8.74147356e-01 1.19552590e-01 -1.29071403e+00
4.98925522e-02 -7.90608168e-01 -1.05565339e-01 8.07774484e-01
3.95252615e-01 7.43404508e-01 3.58149320e-01 4.16031152e-01
-5.82409441e-01 4.76489514e-01 5.24346530e-01 -3.48544717e-01
-9.69440937e-02 -2.15153798e-01 -4.74694729e-01 5.47465205e-01
6.59687400e-01 -6.15392268e-01 -3.13703418e-01 -1.10611737e+00
3.00836980e-01 -6.47644848e-02 6.65356219e-01 -1.47623503e+00
9.80662644e-01 1.89161405e-01 7.19389975e-01 -1.01893365e+00
7.65981972e-01 -1.01604974e+00 3.10390264e-01 3.00580263e-01
-1.30634427e-01 5.33330142e-01 3.94718826e-01 5.44682145e-01
-2.50924945e-01 2.03368515e-01 6.05551779e-01 8.17086920e-02
-9.31216896e-01 7.00645328e-01 3.39834252e-03 -3.19204032e-01
7.10376501e-01 -5.80068529e-01 -2.36997396e-01 -4.81917769e-01
-1.26041502e-01 7.98418522e-02 7.14469373e-01 5.69232285e-01
8.16963375e-01 -1.44719696e+00 -4.41878140e-01 5.85590124e-01
2.60288924e-01 3.94214749e-01 1.82375193e-01 5.83744824e-01
-9.95856464e-01 4.91760522e-01 -1.40366003e-01 -1.02116227e+00
-1.12239039e+00 3.59299421e-01 3.82590771e-01 3.07132244e-01
-8.86922538e-01 1.18629777e+00 5.64153716e-02 -5.41750848e-01
5.00589907e-01 -4.12113786e-01 6.27108574e-01 -3.74291927e-01
4.08987969e-01 -3.22648175e-02 5.86196110e-02 -8.24847221e-01
-4.12525713e-01 9.40055847e-01 1.56467520e-02 9.77559909e-02
1.41403711e+00 -4.92309313e-03 2.13366114e-02 -1.02849361e-02
1.66526735e+00 -1.20139830e-01 -1.82400084e+00 -4.87609714e-01
-1.01028465e-01 -8.20561111e-01 4.99083757e-01 -3.75434220e-01
-1.31115413e+00 7.34537303e-01 7.61704683e-01 -4.20188576e-01
1.18759787e+00 7.14367256e-02 9.22476232e-01 8.98594558e-01
5.33653975e-01 -8.25084805e-01 -6.52136803e-02 5.85275650e-01
6.44582868e-01 -1.27745855e+00 2.23101348e-01 -1.68490991e-01
-2.81721473e-01 1.08970249e+00 6.15003467e-01 -2.96114594e-01
7.18876839e-01 5.02425432e-01 9.30580497e-02 -2.51406461e-01
-6.10804737e-01 1.03525691e-01 2.75493145e-01 5.06479442e-01
-1.36550561e-01 -1.31282166e-01 7.88944542e-01 -7.00271595e-03
-5.26430666e-01 -2.79360294e-01 -3.32115889e-02 8.45520198e-01
-3.48350048e-01 -7.18308270e-01 -3.71767938e-01 4.65342015e-01
5.09848557e-02 -2.51515597e-01 -8.15141350e-02 6.92565620e-01
-1.61195993e-01 4.44155008e-01 5.79875708e-01 -6.55555427e-01
2.30826840e-01 -3.18238229e-01 2.06995219e-01 -3.69315237e-01
-3.12161773e-01 -3.61240171e-02 -3.25629503e-01 -1.15339208e+00
-2.36536816e-01 -6.52589262e-01 -1.14976025e+00 -5.21170378e-01
-1.55360207e-01 -8.18344653e-02 1.16056550e+00 4.96659309e-01
7.95160353e-01 9.41606238e-02 5.83415627e-01 -1.15257788e+00
-2.32861847e-01 -5.55520475e-01 -1.79260254e-01 -4.42283414e-02
5.85474312e-01 -5.00729859e-01 -2.08365560e-01 -1.45479767e-02] | [7.742676734924316, -2.177781820297241] |
bbc3f4d4-3255-41ea-aff4-5aa0bfda3e36 | ultra-fine-entity-typing-with-indirect | 2202.06167 | null | https://arxiv.org/abs/2202.06167v1 | https://arxiv.org/pdf/2202.06167v1.pdf | Ultra-fine Entity Typing with Indirect Supervision from Natural Language Inference | The task of ultra-fine entity typing (UFET) seeks to predict diverse and free-form words or phrases that describe the appropriate types of entities mentioned in sentences. A key challenge for this task lies in the large amount of types and the scarcity of annotated data per type. Existing systems formulate the task as a multi-way classification problem and train directly or distantly supervised classifiers. This causes two issues: (i) the classifiers do not capture the type semantics since types are often converted into indices; (ii) systems developed in this way are limited to predicting within a pre-defined type set, and often fall short of generalizing to types that are rarely seen or unseen in training. This work presents LITE, a new approach that formulates entity typing as a natural language inference (NLI) problem, making use of (i) the indirect supervision from NLI to infer type information meaningfully represented as textual hypotheses and alleviate the data scarcity issue, as well as (ii) a learning-to-rank objective to avoid the pre-defining of a type set. Experiments show that, with limited training data, LITE obtains state-of-the-art performance on the UFET task. In addition, LITE demonstrates its strong generalizability, by not only yielding best results on other fine-grained entity typing benchmarks, more importantly, a pre-trained LITE system works well on new data containing unseen types. | ['Muhao Chen', 'Wenpeng Yin', 'Bangzheng Li'] | 2022-02-12 | null | null | null | null | ['entity-typing'] | ['natural-language-processing'] | [ 1.23409830e-01 1.95759118e-01 -5.39827287e-01 -3.89036536e-01
-5.48010826e-01 -7.42531300e-01 6.46220982e-01 3.77181977e-01
-6.32098436e-01 1.10831428e+00 1.14494145e-01 -4.22894955e-01
2.13231929e-02 -9.32880521e-01 -8.87932777e-01 -2.55064726e-01
-5.46200108e-03 8.22951913e-01 1.47225261e-01 -3.52074951e-01
1.30920544e-01 -1.32766571e-02 -1.80788565e+00 5.51576555e-01
1.27927220e+00 1.16543949e+00 1.44871265e-01 4.19318348e-01
-4.88476694e-01 6.68749571e-01 -6.69532418e-01 -6.65047586e-01
-1.87688217e-01 1.50326581e-03 -1.22618735e+00 -3.56266230e-01
6.58891916e-01 2.99612936e-02 8.81841630e-02 9.43242729e-01
1.80614457e-01 -1.55393004e-01 9.30446684e-01 -1.37950802e+00
-8.19073856e-01 5.31861365e-01 -3.21001023e-01 8.42763782e-02
8.00199449e-01 -3.35393608e-01 1.58907092e+00 -1.27615142e+00
8.86023641e-01 1.22833288e+00 9.78185236e-01 8.57887089e-01
-1.35948241e+00 -3.42384398e-01 1.07427046e-01 2.46992722e-01
-1.53498542e+00 -3.22419524e-01 1.80029437e-01 -5.68069577e-01
1.34351563e+00 6.27971649e-01 1.65379733e-01 1.21750176e+00
-8.30289871e-02 8.61573279e-01 1.18870747e+00 -4.78229403e-01
1.87546954e-01 4.54505622e-01 6.15195274e-01 6.04137123e-01
5.23712695e-01 -1.60410210e-01 -3.62295419e-01 -2.89457917e-01
4.69328538e-02 -4.14862007e-01 -1.53836742e-01 1.25250578e-01
-1.18251550e+00 5.99926353e-01 2.18903050e-01 3.79743665e-01
-1.75206617e-01 -2.93432921e-01 6.00260139e-01 3.16522062e-01
5.53220987e-01 9.71634150e-01 -1.06652713e+00 -2.20796391e-01
-5.89443624e-01 4.62151319e-01 1.28029215e+00 1.30426419e+00
1.00550985e+00 -4.24818516e-01 -3.24840188e-01 1.10535038e+00
3.05068661e-02 6.38869107e-01 3.93692642e-01 -2.49398187e-01
6.66183174e-01 9.06900167e-01 1.83024898e-01 -5.28729081e-01
-3.59617323e-01 -1.78409278e-01 -6.89072251e-01 -2.27665037e-01
5.68406105e-01 -3.29447091e-01 -7.70633399e-01 1.66970098e+00
2.83365011e-01 6.54769465e-02 -9.37785208e-03 5.87581158e-01
1.10341954e+00 4.63123351e-01 8.80524591e-02 1.03670312e-02
1.56999195e+00 -5.93882203e-01 -6.74408078e-01 -5.50756216e-01
9.47544158e-01 -5.69774985e-01 1.17135298e+00 1.69597417e-01
-6.20145977e-01 -2.33643562e-01 -8.06625426e-01 -2.54329830e-01
-1.13632751e+00 -3.43216434e-02 5.74044943e-01 6.07608199e-01
-6.88533664e-01 4.53710049e-01 -2.07463130e-01 -1.06456086e-01
1.90322667e-01 2.19379678e-01 -3.75878304e-01 -9.83144343e-02
-1.65925407e+00 8.49436760e-01 6.08333170e-01 1.63063645e-01
-1.26515493e-01 -1.16338909e+00 -9.49866891e-01 6.30821735e-02
6.98212087e-01 -9.15526092e-01 1.12621260e+00 -7.73899317e-01
-7.60457635e-01 8.53501916e-01 -5.41920364e-01 -2.73531616e-01
1.80153161e-01 -2.41719529e-01 -5.65125525e-01 -3.88703704e-01
4.81673628e-01 1.67707115e-01 2.38971338e-01 -1.41263723e+00
-1.21777427e+00 -2.53529012e-01 2.31683627e-01 5.46465293e-02
-4.74409699e-01 -1.05663925e-01 -3.08715016e-01 -8.03439796e-01
-3.29258889e-01 -9.24928010e-01 1.57474786e-01 -6.06950462e-01
-6.46217763e-01 -1.02590370e+00 7.64913321e-01 -5.04811823e-01
1.70350838e+00 -1.78537166e+00 1.07229508e-01 1.13943703e-01
4.24334198e-01 2.98849314e-01 1.41655520e-01 3.49883050e-01
-4.06503901e-02 6.46828532e-01 1.86905507e-02 -2.23134249e-01
3.69865119e-01 5.19690335e-01 -4.04267758e-01 -5.71869723e-02
1.99155003e-01 8.81492853e-01 -1.15626383e+00 -5.66014409e-01
-3.31818610e-01 -1.62602559e-01 -6.60325110e-01 3.86824101e-01
-6.16207063e-01 -1.56952724e-01 -5.40925443e-01 7.30651021e-01
4.27923471e-01 -3.01529229e-01 8.14694315e-02 -3.05828780e-01
-2.11812690e-01 7.25260437e-01 -1.11689126e+00 9.56041336e-01
-5.94810069e-01 1.90624580e-01 -2.07632989e-01 -8.21798980e-01
4.15067047e-01 2.90203124e-01 3.78886759e-01 -5.05479932e-01
-2.48830467e-01 5.74572742e-01 -1.94751620e-01 -7.97848761e-01
7.62405336e-01 -8.95451754e-02 -6.57739341e-01 2.56453335e-01
1.05658069e-01 1.42044410e-01 6.74067497e-01 5.16737998e-02
1.11527312e+00 -6.72257096e-02 2.17048481e-01 -2.84624904e-01
3.31976980e-01 1.83630258e-01 1.01030409e+00 1.03295863e+00
3.71313572e-01 3.56283188e-01 4.69765633e-01 -3.59329134e-01
-1.01222599e+00 -7.61164427e-01 -5.11064410e-01 1.22016549e+00
3.09186280e-01 -7.58040905e-01 -4.35755640e-01 -1.16398835e+00
2.62438506e-01 6.41675532e-01 -7.33333766e-01 1.61745250e-01
-5.41263461e-01 -7.11687028e-01 7.27715909e-01 5.24309158e-01
2.11788014e-01 -1.17078245e+00 8.96051973e-02 1.50160566e-01
-6.76292956e-01 -1.32453847e+00 -2.90852398e-01 2.28757992e-01
-1.03990197e-01 -9.83394921e-01 -4.14346009e-01 -9.41508710e-01
6.67197764e-01 -1.14077792e-01 1.44476795e+00 2.98747569e-01
-1.21301152e-02 -1.46598250e-01 -6.50258183e-01 -4.27741766e-01
-4.13321942e-01 5.31717122e-01 1.04716979e-01 -1.32055357e-01
6.04160786e-01 -1.24269038e-01 -5.19774221e-02 3.09353977e-01
-5.85903108e-01 -7.88858831e-02 7.12199748e-01 1.28255546e+00
3.60326707e-01 1.84903502e-01 7.04775155e-01 -1.81893933e+00
4.36811894e-01 -6.09597087e-01 -3.29339892e-01 7.02726722e-01
-9.37218368e-01 1.70661971e-01 1.00212550e+00 -2.98687398e-01
-1.23740578e+00 -4.49307114e-01 -3.61323506e-01 1.03539944e-01
-2.57101744e-01 1.02908623e+00 -4.08867180e-01 2.40804836e-01
6.91168487e-01 5.24438880e-02 -5.23871839e-01 -4.66331303e-01
1.74183145e-01 9.26760852e-01 5.08006513e-01 -1.01470709e+00
8.89587283e-01 -8.90475363e-02 -1.67718232e-01 -9.37999547e-01
-1.22809172e+00 -7.19061494e-01 -8.14949632e-01 4.45225351e-02
6.37008965e-01 -6.93824410e-01 -5.81594825e-01 5.48806071e-01
-9.38489974e-01 -3.89108747e-01 1.03044160e-01 -1.09784594e-02
-2.66901314e-01 6.00045562e-01 -6.46600842e-01 -5.92917264e-01
-2.96033949e-01 -7.33606637e-01 1.28884792e+00 4.23935726e-02
-4.74131614e-01 -1.14651573e+00 -1.80368796e-01 4.17700320e-01
9.21302661e-02 1.33528367e-01 1.42577505e+00 -7.76101291e-01
-3.83133858e-01 -2.92038590e-01 -2.23004833e-01 6.05481304e-02
1.90741524e-01 -1.46163478e-01 -8.65568399e-01 -7.31608421e-02
-5.56938946e-01 -4.92124110e-01 6.49590373e-01 -1.90348372e-01
1.04452062e+00 -9.25346076e-01 -4.67708349e-01 5.43542743e-01
1.31454754e+00 -1.50101393e-01 4.79149878e-01 4.63020414e-01
9.52511549e-01 6.57195568e-01 9.93746102e-01 4.45852071e-01
8.06285262e-01 7.57951558e-01 2.84836441e-02 -2.11783245e-01
3.44259031e-02 -4.40549672e-01 2.70589814e-02 5.69747269e-01
7.42072538e-02 -3.34410250e-01 -1.06073594e+00 5.99407315e-01
-1.94064677e+00 -1.11933315e+00 -4.02823031e-01 2.25470948e+00
1.44853985e+00 2.22802699e-01 2.37482011e-01 4.19501737e-02
5.77005744e-01 -2.58413941e-01 -2.77299762e-01 -2.47981831e-01
-1.56467900e-01 1.35497674e-01 5.29325962e-01 3.72670352e-01
-1.27203965e+00 1.02457452e+00 5.46823263e+00 1.05338287e+00
-1.04642987e+00 -2.17256978e-01 3.80048156e-01 5.01630723e-01
-5.42303622e-01 -4.76715304e-02 -1.29321742e+00 7.28981137e-01
5.98560095e-01 -2.96976149e-01 4.04047281e-01 7.73893893e-01
-2.05469146e-01 7.19506666e-02 -1.35541058e+00 5.61832488e-01
-1.68801829e-01 -1.15640771e+00 1.67593844e-02 2.60605782e-01
4.37172115e-01 -1.49070993e-01 -1.26284808e-01 7.32648969e-01
2.39131570e-01 -1.01947331e+00 9.55399334e-01 2.60803491e-01
1.12396264e+00 -4.34007019e-01 8.41396034e-01 5.75376987e-01
-1.21941936e+00 -5.41220605e-02 -3.70444387e-01 4.87497486e-02
-2.12889969e-01 7.32212245e-01 -9.97353315e-01 7.21271873e-01
6.47120476e-01 6.85530722e-01 -6.31210983e-01 8.78961444e-01
-3.41263592e-01 4.95128185e-01 -2.92547375e-01 -4.81836587e-01
3.45408358e-02 2.00700358e-01 5.06790042e-01 1.58640480e+00
-3.47935632e-02 4.41492237e-02 3.69260490e-01 8.64520729e-01
-6.04961887e-02 -1.36121839e-01 -5.09270668e-01 -3.04094218e-02
8.74566734e-01 1.28657782e+00 -1.85661986e-01 -5.93995690e-01
-6.31554604e-01 5.43696046e-01 6.73341990e-01 3.47214460e-01
-6.46817803e-01 -5.11584997e-01 6.53001487e-01 9.14756581e-02
3.36741954e-01 -3.28794122e-02 -5.67963123e-01 -1.49783230e+00
3.18475127e-01 -9.39971805e-01 7.73942530e-01 -3.88667971e-01
-1.79933250e+00 5.42598784e-01 7.00435415e-02 -1.07514060e+00
-5.92321873e-01 -8.24586451e-01 -9.87853333e-02 9.56446111e-01
-1.71343493e+00 -1.20178199e+00 -1.31613359e-01 2.70177424e-01
4.12038088e-01 3.78031805e-02 1.00333977e+00 6.89909875e-01
-8.10706675e-01 1.19742906e+00 -2.32700333e-02 5.00386596e-01
7.52331018e-01 -1.96302807e+00 2.70022750e-01 7.82630622e-01
-1.96280971e-01 9.94683146e-01 8.37349534e-01 -7.11132884e-01
-1.64925313e+00 -1.24670637e+00 1.64234126e+00 -6.42625809e-01
8.23097706e-01 -7.19395220e-01 -1.10620010e+00 6.38305426e-01
-1.23223864e-01 1.29375234e-02 6.51761115e-01 1.01882446e+00
-7.31711626e-01 7.30983391e-02 -1.04560912e+00 4.18210566e-01
1.09028459e+00 -5.84406912e-01 -9.70678449e-01 3.60865474e-01
3.62873584e-01 -6.14847362e-01 -1.09912837e+00 3.91800076e-01
7.37130940e-01 -4.71398890e-01 6.89150214e-01 -6.77994251e-01
6.96762860e-01 -1.38159692e-01 6.69398233e-02 -1.50689840e+00
-3.15521508e-01 -3.62767339e-01 -1.28134698e-01 1.70673978e+00
1.02884495e+00 -7.89788783e-01 6.01250947e-01 9.98772025e-01
-2.51007825e-01 -9.30144727e-01 -7.03497171e-01 -9.29206073e-01
5.69255166e-02 4.20085751e-02 5.12676835e-01 1.23824203e+00
3.76992911e-01 7.28298843e-01 -1.23339422e-01 2.04118565e-01
4.78451043e-01 2.97618508e-01 7.33063161e-01 -1.19432318e+00
-3.41643393e-01 -5.79904377e-01 -2.24233210e-01 -1.00899684e+00
5.43401003e-01 -9.73277032e-01 3.36953342e-01 -1.32442129e+00
2.26618975e-01 -1.23077738e+00 3.23814563e-02 7.96251655e-01
-6.99624538e-01 5.04821027e-03 -8.69130865e-02 2.39852905e-01
-4.89789903e-01 9.98298302e-02 8.26940119e-01 -2.60323644e-01
2.22263798e-01 6.14406355e-02 -8.10136318e-01 5.64050734e-01
2.17405066e-01 -3.39321822e-01 -2.39070691e-02 -3.71091157e-01
7.05932796e-01 -8.51373672e-02 1.41486287e-01 -4.07150418e-01
1.92486495e-01 -3.33121806e-01 1.06906690e-01 -5.27274072e-01
9.19314325e-02 -5.22389472e-01 -1.50013581e-01 -1.02471970e-02
-2.60698408e-01 4.70215417e-02 -1.45774588e-01 2.88620979e-01
-2.08533674e-01 -5.72064340e-01 3.18777502e-01 -1.93699580e-02
-1.32380569e+00 2.92201996e-01 -1.16120920e-01 7.32895911e-01
5.41468680e-01 -2.13400096e-01 -7.57611454e-01 2.19807863e-01
-3.44834507e-01 3.05764377e-01 6.48282230e-01 5.96687496e-01
1.61810055e-01 -1.20989931e+00 -6.26219928e-01 1.38585418e-01
6.27880454e-01 2.36366302e-01 -3.53787616e-02 5.28952777e-01
-1.19381644e-01 3.86360884e-01 1.56368375e-01 -4.86480802e-01
-1.28113747e+00 5.82759321e-01 1.99282214e-01 -5.26603401e-01
-5.87231100e-01 9.65785921e-01 3.60169291e-01 -6.33302450e-01
1.97909176e-01 -2.30352357e-01 -3.35477382e-01 2.78801203e-01
3.85563940e-01 1.99315697e-02 4.61929828e-01 -6.03890896e-01
-4.60758567e-01 1.17924385e-01 -3.41367394e-01 3.31148595e-01
1.34553838e+00 9.01824459e-02 -3.84563506e-01 4.89935875e-01
1.28946352e+00 2.38234326e-01 -5.57121694e-01 -4.45411026e-01
6.30117476e-01 -3.95441413e-01 -2.89064854e-01 -1.06645119e+00
-4.53047395e-01 3.63110632e-01 -1.27816260e-01 6.80396914e-01
7.30615616e-01 1.34892702e-01 1.07113779e+00 5.22201478e-01
4.60557222e-01 -1.19009304e+00 -3.12236965e-01 1.16904688e+00
6.29162133e-01 -1.10505688e+00 -1.28607944e-01 -7.56036997e-01
-4.34091717e-01 9.46340322e-01 7.50672996e-01 3.67017031e-01
3.55236113e-01 6.59256697e-01 -4.04975936e-02 -1.29837289e-01
-1.06497705e+00 -3.58961701e-01 7.14253724e-01 4.74295646e-01
7.80758798e-01 1.83910936e-01 -3.63914162e-01 8.03287506e-01
-4.77882057e-01 -1.51279882e-01 4.74901378e-01 8.53668213e-01
-1.81893811e-01 -1.24286819e+00 -1.38650090e-02 1.01228893e+00
-5.24506211e-01 -2.78901845e-01 -5.73482156e-01 7.39357293e-01
2.66952217e-01 9.49075341e-01 -2.45806441e-01 -4.04252857e-01
4.09360856e-01 -1.99318621e-02 3.93954188e-01 -1.04139864e+00
-6.98780596e-01 -3.40999156e-01 8.46430659e-01 -1.27308726e-01
6.65282533e-02 -6.91722691e-01 -1.00863266e+00 -3.12397003e-01
-5.90379357e-01 2.11946800e-01 2.92522013e-01 1.33710337e+00
3.25354069e-01 2.99997389e-01 7.57674992e-01 -3.64349991e-01
-6.57037437e-01 -9.43090856e-01 -4.77927655e-01 9.68742073e-01
4.40493524e-01 -9.45782423e-01 -3.24940801e-01 -1.44092083e-01] | [9.66506290435791, 8.784217834472656] |
11a03b6a-9552-434b-a9a4-066cf8092597 | mind-your-clever-neighbours-unsupervised | 2112.01839 | null | https://arxiv.org/abs/2112.01839v2 | https://arxiv.org/pdf/2112.01839v2.pdf | Mind Your Clever Neighbours: Unsupervised Person Re-identification via Adaptive Clustering Relationship Modeling | Unsupervised person re-identification (Re-ID) attracts increasing attention due to its potential to resolve the scalability problem of supervised Re-ID models. Most existing unsupervised methods adopt an iterative clustering mechanism, where the network was trained based on pseudo labels generated by unsupervised clustering. However, clustering errors are inevitable. To generate high-quality pseudo-labels and mitigate the impact of clustering errors, we propose a novel clustering relationship modeling framework for unsupervised person Re-ID. Specifically, before clustering, the relation between unlabeled images is explored based on a graph correlation learning (GCL) module and the refined features are then used for clustering to generate high-quality pseudo-labels.Thus, GCL adaptively mines the relationship between samples in a mini-batch to reduce the impact of abnormal clustering when training. To train the network more effectively, we further propose a selective contrastive learning (SCL) method with a selective memory bank update policy. Extensive experiments demonstrate that our method shows much better results than most state-of-the-art unsupervised methods on Market1501, DukeMTMC-reID and MSMT17 datasets. We will release the code for model reproduction. | ['Pingping Zhang', 'Yinjie Lei', 'Tianyu Yan', 'Xiehao Ye', 'Chenyang Yu', 'Lianjie Jia'] | 2021-12-03 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [-8.01716000e-02 -1.22379452e-01 -2.69608647e-01 -6.40949070e-01
-2.40108863e-01 -1.55535638e-01 5.18565834e-01 1.04625545e-01
-6.37728155e-01 6.33855999e-01 1.99583635e-01 1.09088883e-01
-2.02655092e-01 -6.59488380e-01 -2.29630098e-01 -6.04020715e-01
5.12442961e-02 7.90922523e-01 5.48475347e-02 2.96065450e-01
1.92095980e-01 2.71052986e-01 -1.45534551e+00 -3.44729610e-02
1.20418811e+00 4.29381877e-01 -2.29901616e-02 3.73125792e-01
3.73335481e-02 8.31810117e-01 -5.25554061e-01 -5.44867337e-01
3.45775187e-01 -5.66035032e-01 -7.74894059e-01 4.93722886e-01
8.12356323e-02 -1.76458776e-01 -4.49411243e-01 1.22216964e+00
6.39742017e-01 4.90956604e-01 8.23309839e-01 -1.35382712e+00
-5.05557060e-01 6.96493208e-01 -8.16663384e-01 1.84762940e-01
2.12708890e-01 2.70603579e-02 6.24686897e-01 -8.70662332e-01
4.06336308e-01 1.31673384e+00 5.76240778e-01 8.15852463e-01
-1.30665934e+00 -9.61476147e-01 2.42881745e-01 3.41030687e-01
-1.69340777e+00 -3.52985203e-01 7.70567894e-01 -3.60311508e-01
6.52074277e-01 5.45361312e-03 4.49307978e-01 8.58799338e-01
-3.64438832e-01 5.72622716e-01 9.61907446e-01 -5.86619318e-01
1.64558157e-01 2.80236214e-01 5.78741491e-01 7.02102602e-01
4.01219279e-01 -4.31877635e-02 -5.18768191e-01 -4.24441136e-02
6.47386968e-01 3.45630974e-01 1.72472343e-01 -1.84899062e-01
-1.08433771e+00 5.97879469e-01 2.30730847e-01 2.44135946e-01
-1.27554044e-01 -2.01480106e-01 2.56806165e-01 1.49972215e-01
4.17465478e-01 2.25255400e-01 -6.98872805e-02 6.33711889e-02
-9.80085075e-01 1.49671799e-02 5.69954097e-01 9.70897377e-01
9.30891216e-01 -2.86025435e-01 -1.51787668e-01 1.16637063e+00
4.13483977e-01 2.58791417e-01 8.46230388e-01 -8.14286113e-01
3.80889148e-01 1.01067889e+00 -1.36428982e-01 -1.13883710e+00
-5.48933804e-01 -4.77417976e-01 -1.26163721e+00 -2.74136186e-01
2.39048600e-01 -2.83932745e-01 -9.69057620e-01 1.45158172e+00
2.49264717e-01 5.76463819e-01 -2.37193368e-02 9.04404461e-01
7.21445918e-01 3.23856086e-01 1.34127051e-01 -3.79122555e-01
1.02529335e+00 -1.03306091e+00 -5.67518413e-01 -1.58579290e-01
7.12601662e-01 -2.39519775e-01 6.99733377e-01 2.23951653e-01
-7.75865674e-01 -8.98720086e-01 -9.56568420e-01 4.15231675e-01
-3.85991275e-01 2.48476654e-01 4.42436963e-01 8.52053583e-01
-1.10109973e+00 4.23188627e-01 -6.38282537e-01 -5.88739097e-01
3.40529799e-01 6.70895755e-01 -2.77566493e-01 -2.22624078e-01
-1.03520548e+00 3.56774300e-01 6.80690229e-01 4.45346124e-02
-5.02442539e-01 -3.09962660e-01 -7.29665279e-01 -3.48490775e-02
3.61918867e-01 -3.83100539e-01 8.20927501e-01 -9.37569678e-01
-1.26569998e+00 8.68442059e-01 -2.55819231e-01 -5.31428099e-01
3.33195597e-01 8.55175257e-02 -7.78294563e-01 2.02324435e-01
2.27225333e-01 7.74343371e-01 8.05924773e-01 -1.41208875e+00
-7.73888648e-01 -4.69980896e-01 -5.29465377e-01 3.55788559e-01
-7.67742753e-01 2.81660035e-02 -9.67737198e-01 -6.55705333e-01
1.78437546e-01 -1.14319944e+00 -3.83643031e-01 -8.57212543e-01
-7.17591822e-01 -4.84152168e-01 5.59670866e-01 -5.38838923e-01
1.53983998e+00 -2.21078062e+00 -1.71520099e-01 8.05220842e-01
4.44065571e-01 3.19484860e-01 -6.15861341e-02 -1.55792013e-02
-1.67151332e-01 1.88337475e-01 -2.19961062e-01 -7.86539793e-01
-1.22225970e-01 1.35316420e-02 2.52924323e-01 3.73033375e-01
-1.79370701e-01 6.49014175e-01 -7.97756910e-01 -8.17756712e-01
1.07851245e-01 8.98734853e-02 -4.62612927e-01 4.20934021e-01
2.89175093e-01 6.04791105e-01 -2.49190897e-01 5.76916218e-01
6.86037242e-01 -5.01443624e-01 5.36213279e-01 -1.92701370e-01
1.07179403e-01 -1.19703092e-01 -1.40449977e+00 1.20575821e+00
2.07233608e-01 7.45567903e-02 -2.88465321e-01 -1.18884730e+00
1.06458521e+00 1.58670191e-02 5.90857506e-01 -6.91473901e-01
5.67439273e-02 -1.83784127e-01 -1.01747416e-01 -1.73889533e-01
4.61534500e-01 2.61281818e-01 -7.88485780e-02 7.19535768e-01
-1.36487469e-01 8.10885191e-01 3.95883441e-01 6.10412717e-01
7.98557818e-01 -2.72747010e-01 -9.74582657e-02 -1.45125568e-01
8.21729720e-01 -1.28661856e-01 7.89526999e-01 9.26985145e-01
-3.28637958e-01 5.91201842e-01 4.00095917e-02 -1.65892616e-01
-8.32444012e-01 -8.99500549e-01 1.33576512e-01 9.92173195e-01
2.68728286e-01 -6.24101102e-01 -9.67251599e-01 -8.87605488e-01
-9.77632497e-03 3.90228182e-01 -4.12752509e-01 -2.85324663e-01
-5.30889332e-01 -1.07427013e+00 4.05649096e-01 5.31997383e-01
1.02643514e+00 -1.01445353e+00 1.44118845e-01 8.23306814e-02
-2.83125162e-01 -9.76239085e-01 -7.52848625e-01 -9.33535323e-02
-8.94409180e-01 -1.10373998e+00 -5.75083435e-01 -1.11296332e+00
1.29665089e+00 5.31834543e-01 8.75864506e-01 4.23431873e-01
-1.46950111e-01 5.61183631e-01 -4.11630839e-01 8.20578411e-02
-2.83905238e-01 2.51198798e-01 6.15200520e-01 2.94162154e-01
1.03916943e+00 -4.54217732e-01 -5.55931211e-01 4.83641803e-01
-5.27116060e-01 4.64316234e-02 4.57922429e-01 8.58907461e-01
6.44803822e-01 7.55483925e-01 7.86600828e-01 -1.24620020e+00
6.59599781e-01 -4.20798868e-01 -3.71401340e-01 2.63937712e-01
-1.25266683e+00 -2.66516171e-02 6.00655556e-01 -6.23963892e-01
-1.31170678e+00 2.53458589e-01 2.85852253e-01 -3.76044452e-01
-3.14150542e-01 3.99335325e-01 -3.00063640e-01 2.33629838e-01
3.33038360e-01 2.98088223e-01 -5.50267324e-02 -5.65861583e-01
3.01306218e-01 9.73139763e-01 8.25816453e-01 -3.50307673e-01
9.94589150e-01 3.46563101e-01 -4.49088693e-01 -4.91183847e-01
-5.73475480e-01 -9.49796140e-01 -1.04203916e+00 -2.69312918e-01
8.35223675e-01 -1.11628520e+00 -7.80551970e-01 6.53771341e-01
-5.58330655e-01 -3.79193008e-01 1.23199120e-01 5.99250376e-01
-8.45030248e-02 5.75750291e-01 -7.29035318e-01 -7.91633070e-01
-3.17225426e-01 -8.18790674e-01 3.58199149e-01 6.93114579e-01
-2.57548660e-01 -8.52772355e-01 -6.26318455e-02 4.68873292e-01
-6.59835637e-02 -3.16903532e-01 6.09431803e-01 -9.42505538e-01
-4.08533365e-01 -1.42708614e-01 -4.34539139e-01 1.79748818e-01
3.32703978e-01 -4.03892606e-01 -7.38843441e-01 -5.86338460e-01
-3.24696720e-01 -1.51226804e-01 9.36431646e-01 1.58405989e-01
1.19753969e+00 -3.09034765e-01 -6.98183477e-01 6.13457143e-01
1.14870477e+00 2.12785959e-01 4.62726355e-01 4.23252672e-01
1.14297664e+00 6.09872222e-01 5.92956066e-01 6.09226048e-01
7.14275956e-01 3.68443012e-01 -1.66103870e-01 -1.51949748e-01
3.63119021e-02 -3.82928073e-01 1.94082603e-01 9.01069820e-01
-9.02514607e-02 -6.29989877e-02 -8.76704812e-01 4.62166548e-01
-2.00017047e+00 -9.33025599e-01 1.93055533e-02 2.26499557e+00
8.31491411e-01 2.17134163e-01 6.23496592e-01 1.76798359e-01
1.28512692e+00 -2.46433839e-01 -6.35166407e-01 3.43200415e-01
-6.02921359e-02 -3.09021264e-01 5.59615314e-01 2.48350173e-01
-1.13214707e+00 1.05645847e+00 5.90706444e+00 6.41220152e-01
-6.18429124e-01 8.45511183e-02 1.02507710e+00 8.33810195e-02
2.31771454e-01 -1.57147005e-01 -1.17124116e+00 7.49032915e-01
1.02038515e+00 3.42049114e-02 5.29767215e-01 7.44760871e-01
2.22982690e-01 -1.38325587e-01 -9.17278171e-01 1.43035924e+00
3.62890452e-01 -9.31051850e-01 -4.40877788e-02 8.73056948e-02
9.37029481e-01 -3.20207566e-01 -4.09692898e-03 2.93987840e-01
6.53453827e-01 -8.24367762e-01 1.91166177e-01 6.79007888e-01
7.24869311e-01 -1.17878675e+00 7.11691737e-01 3.92531425e-01
-1.28857291e+00 -3.41778785e-01 -5.18029809e-01 1.99435383e-01
-1.13336958e-01 5.29213786e-01 -8.20189238e-01 4.29277122e-01
9.74911630e-01 1.02834392e+00 -1.17869782e+00 9.33738530e-01
-7.34983310e-02 8.46122980e-01 -1.78133860e-01 1.73360065e-01
-3.21853876e-01 -3.12181145e-01 8.66254568e-02 1.35179067e+00
-1.46607282e-02 2.30502725e-01 4.76932466e-01 6.68823898e-01
-1.43858239e-01 6.30521849e-02 -1.65572181e-01 -8.07547593e-04
8.12243760e-01 1.24459875e+00 -1.02529418e+00 -6.88789725e-01
-3.79848659e-01 1.14234447e+00 6.01194620e-01 4.62548405e-01
-5.68618059e-01 -1.99381188e-01 1.35091648e-01 1.61927149e-01
5.31379133e-02 -1.99518591e-01 -1.93454921e-01 -1.19090235e+00
-3.31403583e-01 -8.14157665e-01 8.83776724e-01 -3.55824471e-01
-1.72070229e+00 3.15668434e-01 3.68491001e-02 -1.23294163e+00
-3.72358829e-01 -6.00608923e-02 -5.13430834e-01 5.50150990e-01
-1.21203053e+00 -8.46003413e-01 -5.39263368e-01 9.04347956e-01
3.12483698e-01 -5.97374201e-01 4.92415369e-01 4.52177078e-01
-1.15716505e+00 1.08749580e+00 3.69723625e-02 7.12671816e-01
1.15128541e+00 -1.24058425e+00 1.85951248e-01 1.08471608e+00
-7.80732259e-02 8.46959472e-01 2.19577014e-01 -1.11626315e+00
-9.65789557e-01 -1.39258945e+00 8.50456059e-01 -1.65607750e-01
1.68778762e-01 -3.45920384e-01 -9.05390978e-01 6.37679219e-01
-1.34259790e-01 -2.99291819e-01 9.21343684e-01 1.60892963e-01
-2.60570318e-01 -3.79059315e-01 -1.07087564e+00 5.71963966e-01
1.17239058e+00 -4.13405776e-01 -3.81354511e-01 2.23445460e-01
4.51317757e-01 9.88669991e-02 -9.09846902e-01 1.75653622e-01
1.73185229e-01 -9.92528141e-01 9.20053303e-01 -1.60697967e-01
-2.40172490e-01 -5.06038725e-01 3.95074785e-01 -1.12460423e+00
-7.43336976e-01 -5.96374750e-01 -1.30628213e-01 1.80704629e+00
3.25582683e-01 -6.68489456e-01 9.50439692e-01 1.03551543e+00
3.72916490e-01 -6.15470707e-02 -4.81137842e-01 -6.93233430e-01
-5.06763577e-01 -9.82201844e-02 7.15606570e-01 1.18998539e+00
3.20025772e-01 4.63854581e-01 -4.82545882e-01 2.29858220e-01
1.04614520e+00 -2.22783327e-01 9.46849167e-01 -1.33976591e+00
-1.46137372e-01 -1.78207815e-01 -1.98084533e-01 -1.03677118e+00
2.73342013e-01 -8.99277866e-01 6.97275400e-02 -1.29318106e+00
6.54252052e-01 -7.21018970e-01 -5.20957768e-01 4.56669331e-01
-5.54681778e-01 3.83291721e-01 5.50218783e-02 7.99857438e-01
-1.13189006e+00 2.23630831e-01 7.97412097e-01 -1.29018277e-01
-6.23977125e-01 2.10797973e-02 -7.38295853e-01 5.27848244e-01
9.15772378e-01 -4.96079266e-01 -4.73725736e-01 8.95638317e-02
-2.34418571e-01 -2.59898156e-01 1.63813651e-01 -1.27794814e+00
8.28644753e-01 6.73600659e-02 8.17635357e-01 -8.54173660e-01
-5.68336807e-02 -6.21475518e-01 1.13575861e-01 3.98482114e-01
-4.89243358e-01 7.89164230e-02 -3.93577158e-01 7.73953259e-01
-4.08436619e-02 -1.31425008e-01 8.62859726e-01 -2.00229332e-01
-7.52484500e-01 5.70571482e-01 -3.89434040e-01 -8.49782750e-02
8.12482238e-01 -3.26092213e-01 -1.13255128e-01 -4.08864647e-01
-7.48812258e-01 6.14594519e-01 5.82856417e-01 3.43933374e-01
6.50622308e-01 -1.35920560e+00 -6.04260325e-01 4.51553345e-01
1.25761449e-01 4.72914353e-02 2.68942028e-01 6.46487415e-01
2.63664797e-02 2.45168522e-01 8.03912058e-02 -4.70090091e-01
-1.24532998e+00 8.51979494e-01 2.02192143e-01 -4.08441156e-01
-6.21572256e-01 6.07765317e-01 4.74122027e-03 -4.69247580e-01
5.32312453e-01 4.98404443e-01 -6.04507506e-01 -7.55713927e-03
7.44143605e-01 6.00889504e-01 -3.36066455e-01 -7.85390735e-01
-3.76805872e-01 2.71494180e-01 -5.92043161e-01 -3.08301151e-02
1.03988266e+00 -5.34433961e-01 -1.82381406e-01 1.48489192e-01
9.40231442e-01 -2.25009114e-01 -1.18901920e+00 -4.39975083e-01
3.57929468e-01 -2.09021226e-01 -2.79219836e-01 -5.51009655e-01
-1.13379467e+00 2.88811892e-01 8.92287195e-01 1.97977410e-03
1.10451865e+00 -7.36537576e-02 7.85884142e-01 4.52546746e-01
1.63821474e-01 -1.62839854e+00 4.62548882e-01 3.43286306e-01
-6.13991022e-02 -1.34221268e+00 1.54030100e-01 -3.77593249e-01
-6.77873611e-01 7.64162898e-01 1.05238986e+00 -4.61153947e-02
7.38285780e-01 -1.25340642e-02 2.74941176e-02 2.28345711e-02
-2.06263378e-01 -2.54936814e-01 1.99314237e-01 8.41266394e-01
1.53891936e-01 1.18252270e-01 -1.67664737e-01 8.14670146e-01
-6.78603202e-02 -1.28291234e-01 2.85155028e-01 6.72189415e-01
-3.08882266e-01 -1.18186474e+00 -4.64739263e-01 7.11730957e-01
-1.85982034e-01 -4.99780811e-02 -4.62346286e-01 4.14283067e-01
2.17596248e-01 1.33693469e+00 1.41704708e-01 -8.11846673e-01
3.37977521e-02 2.43945587e-02 1.07294157e-01 -6.92315698e-01
-5.28289735e-01 2.09917471e-01 2.10595485e-02 -2.82464147e-01
-7.76692688e-01 -6.59418881e-01 -1.41027117e+00 -3.84527773e-01
-3.98761302e-01 3.81389141e-01 5.63190095e-02 1.05145741e+00
4.53799605e-01 2.73452461e-01 8.63593638e-01 -6.15491509e-01
-1.48354873e-01 -1.06415069e+00 -7.97501087e-01 7.67840326e-01
-1.65386274e-01 -5.61394095e-01 -3.86741787e-01 3.73465151e-01] | [14.867118835449219, 1.1406325101852417] |
576acef7-0430-4aba-9823-f36b6ecc18f5 | polarimetric-imaging-for-perception | 2305.14787 | null | https://arxiv.org/abs/2305.14787v1 | https://arxiv.org/pdf/2305.14787v1.pdf | Polarimetric Imaging for Perception | Autonomous driving and advanced driver-assistance systems rely on a set of sensors and algorithms to perform the appropriate actions and provide alerts as a function of the driving scene. Typically, the sensors include color cameras, radar, lidar and ultrasonic sensors. Strikingly however, although light polarization is a fundamental property of light, it is seldom harnessed for perception tasks. In this work we analyze the potential for improvement in perception tasks when using an RGB-polarimetric camera, as compared to an RGB camera. We examine monocular depth estimation and free space detection during the middle of the day, when polarization is independent of subject heading, and show that a quantifiable improvement can be achieved for both of them using state-of-the-art deep neural networks, with a minimum of architectural changes. We also present a new dataset composed of RGB-polarimetric images, lidar scans, GNSS / IMU readings and free space segmentations that further supports developing perception algorithms that take advantage of light polarization. | ['Dan Levi', "Tomer Pe'er", 'Michael Baltaxe'] | 2023-05-24 | null | null | null | null | ['monocular-depth-estimation'] | ['computer-vision'] | [ 3.48452181e-01 2.11326387e-02 1.21860117e-01 -7.34967172e-01
-4.43482220e-01 -7.23620713e-01 5.94796240e-01 -1.68615401e-01
-8.01909685e-01 5.09237945e-01 -3.59471947e-01 -6.20998442e-01
-9.46702287e-02 -9.29338396e-01 -5.40228307e-01 -8.95812929e-01
2.82666951e-01 4.98447329e-01 3.40526760e-01 -6.80826068e-01
3.62999856e-01 9.77205694e-01 -2.21116877e+00 -2.08330020e-01
6.96983397e-01 1.36470509e+00 7.83637241e-02 8.41386437e-01
-3.39380167e-02 2.91278303e-01 -2.14228585e-01 -1.00097597e-01
5.10902703e-01 3.22602913e-02 -4.21238318e-02 -1.79149155e-02
7.60538757e-01 -3.53301525e-01 -3.34353507e-01 7.79107273e-01
5.37562072e-01 -6.78662807e-02 4.28414851e-01 -1.26985848e+00
-3.46430950e-02 -3.57361108e-01 -3.08849812e-01 2.90109813e-01
3.47347677e-01 6.53239787e-01 6.35272801e-01 -5.16845763e-01
3.68081599e-01 7.82821298e-01 5.45706928e-01 8.02208856e-02
-7.50302017e-01 -6.07642174e-01 -2.74830043e-01 5.08641541e-01
-9.69471931e-01 -5.67848742e-01 8.16331506e-01 -4.61180478e-01
1.06577265e+00 -1.51176443e-02 6.51777327e-01 7.69512296e-01
4.47639108e-01 1.30528092e-01 1.62989116e+00 -3.83532375e-01
3.37472051e-01 2.05286622e-01 2.08668724e-01 6.69791341e-01
6.12463593e-01 4.73800927e-01 -8.08546066e-01 6.02627933e-01
-5.87133504e-02 -2.79622942e-01 -1.39877766e-01 -6.35658562e-01
-9.08573627e-01 6.75926745e-01 4.88518566e-01 -1.62030935e-01
-5.47808647e-01 -4.79381792e-02 -3.37407701e-02 3.10791790e-01
2.77980506e-01 3.10936421e-01 -2.98542559e-01 -2.13805899e-01
-8.44561815e-01 5.34785800e-02 6.11301541e-01 6.63018584e-01
1.31821203e+00 -3.30035985e-02 4.73099828e-01 3.45727324e-01
4.87896770e-01 1.30142069e+00 -1.91044044e-02 -1.17155683e+00
2.25376114e-01 6.01180017e-01 3.04210186e-01 -7.89777815e-01
-6.92046642e-01 -1.83073461e-01 9.36237536e-03 8.95433843e-01
1.74837321e-01 -2.23952323e-01 -1.11797094e+00 1.03756285e+00
3.52923542e-01 -1.22208096e-01 3.38377178e-01 9.38853264e-01
6.28951252e-01 2.50869870e-01 -2.62914956e-01 -8.19151849e-03
1.40464365e+00 -4.14467692e-01 -5.30370772e-01 -9.42708194e-01
5.20410001e-01 -4.84311432e-01 6.53783798e-01 7.51993716e-01
-5.07255495e-01 -5.89904368e-01 -1.54948008e+00 -9.79185402e-02
-8.05863023e-01 -7.08307624e-02 5.32297015e-01 1.03747499e+00
-1.07623887e+00 3.36722843e-02 -7.18157291e-01 -5.34326792e-01
2.26494417e-01 1.02697894e-01 -4.79678541e-01 -3.02047729e-01
-1.05189550e+00 1.30049300e+00 8.32234398e-02 2.70972311e-01
-5.23182809e-01 -4.96700317e-01 -9.04658139e-01 -3.78702760e-01
3.46692026e-01 -2.99145252e-01 1.03940940e+00 -4.91660923e-01
-1.66561675e+00 1.25632274e+00 -1.72545210e-01 -7.88566113e-01
3.48136902e-01 -3.76230955e-01 -4.31190252e-01 3.97857517e-01
1.04473485e-02 7.40330994e-01 7.26424932e-01 -1.28322423e+00
-1.14207685e+00 -7.52140224e-01 2.23440722e-01 3.03341329e-01
4.01636839e-01 -2.81610668e-01 -1.17923446e-01 8.43438148e-01
3.66915822e-01 -1.27765787e+00 -9.81985554e-02 -1.84186513e-03
-2.19295546e-01 1.71227619e-01 8.73278081e-01 -3.51113766e-01
2.26047933e-01 -2.04259300e+00 -4.50195849e-01 7.29753822e-02
-1.35518974e-02 2.85835028e-01 2.19499990e-01 2.01270550e-01
1.94451153e-01 -4.41304743e-01 -3.68592143e-01 -1.88955843e-01
3.73958535e-02 6.13119662e-01 -4.53732550e-01 8.32922578e-01
8.65384117e-02 6.04106963e-01 -4.96938676e-01 3.96576114e-02
7.12210119e-01 4.36572939e-01 -1.89111248e-01 4.49527726e-02
-5.93970828e-02 6.12008929e-01 -6.81289360e-02 8.29009414e-01
1.01699948e+00 6.20254755e-01 -1.12414695e-01 -1.05073474e-01
-8.51401687e-01 4.30515260e-01 -1.00766015e+00 1.33042312e+00
-5.68922043e-01 1.26973128e+00 5.21943986e-01 -9.21328247e-01
1.09105539e+00 -2.00730488e-01 2.68946856e-01 -1.50394821e+00
2.47100443e-01 4.28105354e-01 -1.07134171e-01 -6.78363681e-01
7.41690993e-01 -2.66450852e-01 3.01389256e-03 1.53220206e-01
-2.25487605e-01 -4.70176548e-01 1.75818324e-01 -3.23840737e-01
8.59041691e-01 1.63746729e-01 -5.02479635e-02 -1.66241433e-02
4.77666944e-01 2.32126117e-01 3.41061682e-01 5.32555997e-01
-4.39104944e-01 3.43794972e-01 2.56533504e-01 -5.06300330e-01
-6.97473109e-01 -8.81926417e-01 -4.05253410e-01 9.25053000e-01
5.02299964e-01 3.59679401e-01 -4.70553607e-01 -2.19638497e-01
3.32966149e-01 1.02419245e+00 -4.39698339e-01 -7.53701553e-02
-3.47599417e-01 -7.56531239e-01 3.41349721e-01 3.68739575e-01
6.45482779e-01 -6.82467520e-01 -1.73385286e+00 -1.40538020e-02
1.15556195e-01 -1.64446497e+00 7.20768154e-01 5.05331755e-01
-5.53636849e-01 -1.41497684e+00 -5.87389283e-02 9.47815999e-02
1.91090226e-01 8.14675868e-01 9.29715395e-01 -4.34934825e-01
-2.53498316e-01 8.68900359e-01 -3.21147889e-01 -8.90941799e-01
-5.21526672e-02 -3.26773822e-01 1.44184917e-01 1.98810771e-01
8.49366069e-01 -4.91146207e-01 -7.24425197e-01 2.55272597e-01
-3.75643134e-01 -1.89258814e-01 7.58081734e-01 9.42461863e-02
4.16301548e-01 -2.81688392e-01 -3.68824512e-01 -5.81437469e-01
7.64861926e-02 -3.45262378e-01 -1.14985502e+00 -3.98783237e-01
-9.64001179e-01 -7.79388994e-02 1.88377630e-02 6.10581934e-01
-1.07228816e+00 2.42231116e-01 -1.32119387e-01 -1.23758338e-01
-8.13665628e-01 2.23566368e-01 -6.80894544e-03 -5.31527221e-01
9.29501295e-01 2.14905396e-01 1.92747504e-01 -7.57625774e-02
4.60382700e-01 9.03012037e-01 7.35205472e-01 -1.20692477e-01
8.27712417e-01 1.17059779e+00 5.73835611e-01 -1.34102166e+00
-7.32622445e-01 -6.36385441e-01 -9.27433968e-01 -4.94558752e-01
9.73835170e-01 -9.52363849e-01 -7.48865664e-01 4.26011622e-01
-8.86977971e-01 -1.93360835e-01 -9.31486487e-02 5.22728145e-01
-6.00698292e-01 2.68712312e-01 2.11888060e-01 -1.23745072e+00
5.73627986e-02 -1.12865245e+00 1.11889195e+00 5.21979213e-01
3.00850034e-01 -5.14640331e-01 2.02600136e-01 8.98153067e-01
3.64865929e-01 4.89346743e-01 3.34864289e-01 -1.67020917e-01
-8.77770841e-01 -3.93489748e-01 -2.44769439e-01 3.69720370e-01
-2.94851303e-01 8.82824585e-02 -1.54188740e+00 -4.06459626e-03
-1.10834986e-01 -1.47458792e-01 1.20035899e+00 5.08984447e-01
5.09999871e-01 2.60727882e-01 -3.25189024e-01 1.14325607e+00
1.40297353e+00 5.45628726e-01 8.64355743e-01 8.28988433e-01
5.29581904e-01 9.72319722e-01 8.13966870e-01 2.22054601e-01
8.74325156e-01 4.78954226e-01 1.01577520e+00 -9.25125927e-02
1.49783820e-01 3.22114885e-01 2.62676746e-01 9.04966891e-02
-3.48684818e-01 7.08647892e-02 -1.19496739e+00 4.87579763e-01
-1.25353992e+00 -6.53548717e-01 -6.19142830e-01 2.28838611e+00
2.14216709e-01 4.87457812e-01 -1.99865282e-01 3.20828021e-01
1.46989211e-01 1.76086575e-01 -6.05825901e-01 -7.06744254e-01
-3.17436397e-01 1.56521201e-01 1.32189846e+00 6.26821399e-01
-1.04411685e+00 7.67307460e-01 6.45001888e+00 -8.76124427e-02
-1.57220531e+00 1.63323358e-02 1.55902848e-01 -2.48913616e-02
-1.27677649e-01 8.18838626e-02 -1.09076190e+00 3.40054393e-01
1.38263559e+00 5.66317439e-01 2.56881714e-01 8.37797046e-01
1.30056933e-01 -9.20229375e-01 -7.51781762e-01 9.90281641e-01
2.65065074e-01 -9.63355482e-01 -7.83558309e-01 3.45411599e-01
3.38672161e-01 5.83244145e-01 1.88461646e-01 3.10511999e-02
-5.78706041e-02 -6.98292911e-01 8.10817301e-01 6.11551464e-01
6.89633608e-01 -7.34703362e-01 7.43289053e-01 2.82599479e-01
-8.21715176e-01 -3.00103575e-01 -3.14882040e-01 -3.78266066e-01
9.90339443e-02 6.11422122e-01 -8.45940411e-01 3.92266870e-01
8.17739904e-01 4.60863829e-01 -6.71209693e-01 8.68233442e-01
-6.52532935e-01 5.03679574e-01 -4.34888989e-01 1.55565262e-01
3.91278356e-01 -5.37544608e-01 5.91584206e-01 1.13003016e+00
3.63673806e-01 1.11303486e-01 -1.53933287e-01 6.30164325e-01
5.83196700e-01 -6.24525607e-01 -1.04002440e+00 2.21854702e-01
2.64407367e-01 1.54677784e+00 -5.19268692e-01 -3.90810855e-02
-5.22480369e-01 4.08331603e-01 -3.41294557e-01 4.80950505e-01
-6.05484366e-01 -4.55976188e-01 1.08858252e+00 4.21642274e-01
3.63327265e-01 -7.64597356e-01 -5.31568646e-01 -5.22145271e-01
6.45697117e-02 -2.91919708e-01 -1.52520642e-01 -1.22467077e+00
-5.25851488e-01 4.50173587e-01 3.19042914e-02 -1.18819010e+00
-1.24918573e-01 -1.22159433e+00 -3.28413486e-01 9.88768160e-01
-2.52785707e+00 -1.00877237e+00 -1.00383270e+00 3.99358660e-01
-1.62415542e-02 -1.26033917e-01 4.02009666e-01 7.19041899e-02
-3.52568269e-01 -1.41442940e-01 -5.68373315e-02 -3.54700953e-01
6.98165417e-01 -1.27091408e+00 2.85551250e-01 1.18027854e+00
-1.56560257e-01 1.53749153e-01 1.13665915e+00 -3.27552646e-01
-1.87848830e+00 -7.56697237e-01 6.38094723e-01 -6.04895592e-01
3.90561908e-01 -3.12635005e-01 -4.57956016e-01 5.37354648e-01
3.36915076e-01 5.98617755e-02 6.10732675e-01 -4.66106348e-02
-2.53458589e-01 -6.65568650e-01 -1.11078870e+00 2.02794775e-01
7.50050843e-01 -7.14164436e-01 -6.08336389e-01 2.25554392e-01
3.14777642e-01 -6.36572421e-01 -2.85351306e-01 5.51375389e-01
4.90701854e-01 -1.70357585e+00 9.29413140e-01 1.25658035e-01
2.78606415e-02 -5.38276315e-01 -4.22292084e-01 -1.18953085e+00
-1.50765732e-01 -3.75034302e-01 2.15341479e-01 5.81513464e-01
2.35636234e-01 -1.19320667e+00 8.04802001e-01 5.58195174e-01
-5.77898979e-01 -1.70249015e-01 -1.12034070e+00 -5.24746358e-01
-4.23342228e-01 -9.35107768e-01 2.74732828e-01 4.17129129e-01
-5.34033477e-01 2.22521931e-01 7.89691061e-02 6.34288013e-01
8.74425590e-01 1.83203295e-01 9.82546866e-01 -1.68509614e+00
2.07681730e-01 -3.30796629e-01 -7.94715166e-01 -6.99371696e-01
2.77320240e-02 -4.62484181e-01 5.08316755e-01 -1.70706701e+00
-7.22984076e-01 -3.07785153e-01 -3.36690545e-02 2.82294065e-01
2.42947817e-01 5.32542706e-01 -1.76369131e-01 -1.65064503e-02
-2.81203538e-01 2.24906474e-01 8.60168755e-01 2.28452347e-02
1.11045383e-01 -1.12175487e-01 -5.44683993e-01 6.31256819e-01
6.34961128e-01 -1.50951281e-01 -3.12652111e-01 -4.19200242e-01
8.03290367e-01 -2.30295822e-01 5.69979846e-01 -1.59585142e+00
5.16453743e-01 -7.98607841e-02 2.71837264e-01 -8.73140633e-01
8.09718609e-01 -7.38606632e-01 -2.91809380e-01 4.99445528e-01
3.84141237e-01 -8.61705914e-02 4.38380450e-01 5.58048427e-01
-1.41520336e-01 1.60047516e-01 9.69831049e-01 -1.15044758e-01
-1.44866014e+00 2.07469631e-02 -6.66590512e-01 -2.42363676e-01
1.09169888e+00 -9.35585499e-01 -6.98446572e-01 -2.96167701e-01
-1.07695490e-01 1.48955926e-01 4.65723932e-01 2.41978556e-01
5.12342334e-01 -7.67845094e-01 -4.80569869e-01 7.51099706e-01
3.95647615e-01 6.61239587e-03 1.63294315e-01 1.16909599e+00
-9.40628231e-01 9.65569556e-01 -5.99473357e-01 -9.83943760e-01
-9.90880489e-01 1.75955355e-01 5.50533712e-01 5.56546628e-01
-3.63250315e-01 7.85542667e-01 -8.79561380e-02 -3.46049994e-01
-2.23223582e-01 -4.35048044e-01 -3.10121626e-01 3.10985088e-01
3.39121729e-01 3.19383085e-01 6.16237223e-01 -9.46926951e-01
-6.11985624e-01 9.46777523e-01 5.61446786e-01 -2.68053710e-01
1.15480006e+00 -5.29854178e-01 -3.68339457e-02 5.58819413e-01
7.31774092e-01 -9.09228399e-02 -1.66553783e+00 -2.82092895e-02
-2.44690597e-01 -5.00809729e-01 4.50815380e-01 -7.69730866e-01
-8.21057558e-01 1.37934577e+00 1.02628827e+00 3.67962897e-01
1.21089184e+00 -4.07250881e-01 5.32789171e-01 6.33463144e-01
5.35364032e-01 -1.31780910e+00 -4.38193262e-01 8.15149307e-01
4.17244971e-01 -1.42775857e+00 -5.63580133e-02 -5.70902303e-02
-6.79924846e-01 1.09012437e+00 3.62088561e-01 2.84940861e-02
5.31766117e-01 2.79326230e-01 5.73998332e-01 -4.21187967e-01
-4.74424452e-01 -8.65480423e-01 1.09063961e-01 9.64447021e-01
6.73133358e-02 1.22512065e-01 1.13009311e-01 -1.31213799e-01
-4.42101836e-01 -2.91418403e-01 6.52891159e-01 1.02332819e+00
-1.14933658e+00 -6.47210479e-01 -6.21941864e-01 1.37035817e-01
-3.37125803e-03 1.04222633e-01 -3.21699381e-01 7.96451449e-01
6.20446086e-01 1.17998123e+00 3.02995205e-01 -3.98851246e-01
6.58143163e-01 2.70383090e-01 4.37011898e-01 -2.95478284e-01
8.49086791e-02 -4.67897326e-01 4.47583586e-01 -9.46779966e-01
-4.17779088e-01 -7.15802073e-01 -1.08080041e+00 -1.06775910e-01
2.64465567e-02 -4.06971008e-01 1.49456751e+00 1.21507096e+00
4.85391110e-01 4.50064778e-01 5.20484865e-01 -1.22207177e+00
-3.69383171e-02 -6.17053032e-01 -7.31296599e-01 -1.04854375e-01
7.09643662e-01 -1.04037762e+00 -4.25567120e-01 -1.69318780e-01] | [8.06129264831543, -1.8463847637176514] |
64de504b-bc3a-447f-b2b7-a2c26404a958 | auto-encoding-score-distribution-regression | 2111.11029 | null | https://arxiv.org/abs/2111.11029v2 | https://arxiv.org/pdf/2111.11029v2.pdf | Auto-Encoding Score Distribution Regression for Action Quality Assessment | The action quality assessment (AQA) of videos is a challenging vision task since the relation between videos and action scores is difficult to model. Thus, AQA has been widely studied in the literature. Traditionally, AQA is treated as a regression problem to learn the underlying mappings between videos and action scores. But previous methods ignored data uncertainty in AQA dataset. To address aleatoric uncertainty, we further develop a plug-and-play module Distribution Auto-Encoder (DAE). Specifically, it encodes videos into distributions and uses the reparameterization trick in variational auto-encoders (VAE) to sample scores, which establishes a more accurate mapping between videos and scores. Meanwhile, a likelihood loss is used to learn the uncertainty parameters. We plug our DAE approach into MUSDL and CoRe. Experimental results on public datasets demonstrate that our method achieves state-of-the-art on AQA-7, MTL-AQA, and JIGSAWS datasets. Our code is available at https://github.com/InfoX-SEU/DAE-AQA. | ['Xin Geng', 'Xu Yang', 'HUI ZHANG', 'Yinfei Xu', 'Jiayuan Chen', 'Boyu Zhang'] | 2021-11-22 | null | null | null | null | ['action-quality-assessment'] | ['computer-vision'] | [-2.64021218e-01 -2.65927970e-01 -1.42884240e-01 -5.44325173e-01
-1.22626424e+00 -4.00502712e-01 4.09459352e-01 -6.44752026e-01
-1.18106052e-01 5.81184030e-01 4.34524238e-01 1.38402075e-01
-1.21744253e-01 -5.98605931e-01 -1.01233208e+00 -8.26503038e-01
1.36332646e-01 1.69561446e-01 7.05078542e-02 2.76739120e-01
-1.29406616e-01 -2.68090457e-01 -1.14274848e+00 1.52416795e-01
1.02904654e+00 1.15765309e+00 1.17396124e-01 6.55157268e-01
1.94370538e-01 1.22732353e+00 -2.89085388e-01 -9.17385817e-01
3.83800387e-01 -5.88587165e-01 -6.98600769e-01 1.67297170e-01
5.80253422e-01 -8.30170989e-01 -9.93813515e-01 1.32496238e+00
1.59093663e-01 9.75219309e-02 8.14116359e-01 -1.69612658e+00
-9.74560916e-01 4.69997942e-01 -6.53663754e-01 6.72996044e-02
1.73984796e-01 5.75475395e-01 1.36599743e+00 -9.23954487e-01
2.38251403e-01 1.50230837e+00 4.11768347e-01 5.14343858e-01
-1.18502295e+00 -8.96495640e-01 1.38833039e-02 6.60169363e-01
-1.42071903e+00 -3.62875581e-01 5.31794608e-01 -6.72966123e-01
2.95742601e-01 -6.28483444e-02 7.86868691e-01 1.38014448e+00
2.12136820e-01 1.42370677e+00 7.87312567e-01 3.64648521e-01
2.72494733e-01 -3.98494661e-01 -2.13177085e-01 7.87460864e-01
1.54724747e-01 -1.42149165e-01 -6.92152917e-01 2.49390360e-02
9.06679511e-01 1.95181832e-01 -1.98380724e-01 -5.55622995e-01
-1.14351881e+00 7.98973858e-01 2.31346905e-01 -4.42633033e-01
-4.80727404e-01 6.95647001e-01 3.56855422e-01 2.54085094e-01
3.25703204e-01 7.59151503e-02 -2.77982056e-01 -5.42065680e-01
-7.65623808e-01 5.32363772e-01 3.07671189e-01 9.21206832e-01
6.00738466e-01 -9.22278326e-04 -5.79827845e-01 8.03625703e-01
6.78320348e-01 8.29301536e-01 2.82327682e-01 -1.73392570e+00
5.45268953e-01 2.60992855e-01 1.63444698e-01 -8.71537447e-01
2.88702101e-01 -2.97857802e-02 -7.59451032e-01 1.02661319e-01
4.34084862e-01 -1.08289950e-01 -7.56792307e-01 1.89150035e+00
3.13872486e-01 7.38636553e-01 -4.75296713e-02 1.04089320e+00
4.56000954e-01 9.98834193e-01 -2.77565401e-02 -1.60752803e-01
1.17405748e+00 -1.11025333e+00 -8.61799002e-01 1.83031876e-02
2.15644404e-01 -4.42772210e-01 1.23431695e+00 6.52601838e-01
-1.36632311e+00 -6.81019366e-01 -8.40898097e-01 -3.28560799e-01
3.55118841e-01 2.04119056e-01 4.25083756e-01 9.22814310e-02
-8.89630079e-01 6.21627450e-01 -1.21597147e+00 1.00027807e-01
9.29140270e-01 9.94011238e-02 -2.82742143e-01 -2.11992323e-01
-1.22766781e+00 4.47061718e-01 1.76117808e-01 -1.85333111e-03
-1.43049765e+00 -8.55344474e-01 -9.73882377e-01 -1.72300525e-02
7.39129961e-01 -8.54236662e-01 1.51894021e+00 -7.71385610e-01
-1.67844677e+00 4.62990403e-01 2.77607162e-02 -5.11511087e-01
5.79711497e-01 -7.07499027e-01 -2.00477332e-01 2.11222738e-01
1.76236778e-01 7.14084983e-01 9.99531865e-01 -7.86536396e-01
-5.99830449e-01 -2.01771051e-01 4.00861561e-01 1.00104101e-01
-1.70428962e-01 -1.95294634e-01 -8.34536076e-01 -8.31269741e-01
-3.17298412e-01 -8.32536519e-01 2.04535201e-01 4.93457556e-01
-1.76668525e-01 -4.34863657e-01 3.85219902e-01 -9.16576564e-01
1.45205998e+00 -2.30022144e+00 6.74457729e-01 -1.88647583e-01
1.95865437e-01 1.75343737e-01 -3.39229554e-01 2.24146575e-01
2.63679266e-01 -1.18226372e-01 -3.39041620e-01 -5.55670023e-01
4.37612444e-01 3.73019546e-01 -2.95997858e-01 4.87425059e-01
4.60482806e-01 9.46609139e-01 -1.07692409e+00 -7.22571909e-01
2.20958546e-01 5.49162090e-01 -9.48453844e-01 5.84734738e-01
-3.17020744e-01 3.44354987e-01 -6.04217887e-01 7.40328193e-01
6.89069450e-01 -3.59686702e-01 -7.61257112e-02 -3.49081159e-01
1.71898380e-01 6.79125786e-02 -1.21251261e+00 2.20382524e+00
-2.31658295e-01 5.28611958e-01 -8.36894140e-02 -7.34531581e-01
4.57365245e-01 2.10384429e-01 6.46252692e-01 -3.64886492e-01
2.31016472e-01 2.64675040e-02 -1.57421276e-01 -7.88719237e-01
2.33164877e-01 1.01955421e-01 -1.28662921e-02 2.16071799e-01
4.25528020e-01 -5.83343394e-02 3.83183450e-01 4.62810993e-01
1.05318797e+00 5.97925305e-01 2.06842363e-01 2.12377449e-03
5.07623374e-01 -4.37453926e-01 1.03794360e+00 4.67966855e-01
-4.68046069e-01 7.79231012e-01 7.79411197e-01 -8.98257643e-02
-1.05863750e+00 -1.41917527e+00 -1.03573173e-01 7.80897558e-01
4.62564006e-02 -6.56320810e-01 -9.52803612e-01 -7.08684027e-01
1.03960864e-01 5.90920269e-01 -7.23880947e-01 -3.27279061e-01
-1.30870029e-01 -3.58083159e-01 2.62183994e-01 6.51332498e-01
6.85166359e-01 -7.23053098e-01 -9.36780423e-02 1.86912194e-02
-4.53523904e-01 -8.79848361e-01 -1.00097573e+00 -7.06530571e-01
-6.04105473e-01 -1.20783079e+00 -7.79582798e-01 -1.22388288e-01
2.14498922e-01 2.41108984e-01 1.06683505e+00 -2.99765944e-01
-2.58404501e-02 6.12072587e-01 -3.78860474e-01 -2.45741561e-01
-1.69823930e-01 -4.14131254e-01 1.12439908e-01 4.14081812e-01
4.82567102e-01 -6.33590460e-01 -9.61303294e-01 2.68639773e-01
-9.90685463e-01 7.44002759e-02 6.44111514e-01 7.65751243e-01
8.76175463e-01 -2.55559146e-01 3.66360456e-01 -5.21024346e-01
2.33683437e-01 -8.19593191e-01 -7.42613018e-01 2.06208780e-01
-4.64612514e-01 1.51205435e-01 3.35281700e-01 -4.79612887e-01
-1.02684200e+00 1.17911100e-02 -1.48605853e-01 -1.22275257e+00
1.95752323e-01 3.26808065e-01 -5.76375067e-01 4.80960876e-01
1.38214426e-02 1.98350117e-01 8.61016884e-02 -3.50552231e-01
5.32442093e-01 5.90548754e-01 6.38632238e-01 -6.97329998e-01
8.20136249e-01 4.86630470e-01 -1.62617832e-01 -1.85530573e-01
-1.25481856e+00 -3.55875671e-01 -2.50230461e-01 -4.27762240e-01
1.16273761e+00 -1.37021410e+00 -7.59895563e-01 6.95448220e-01
-9.09887910e-01 -3.98615062e-01 -4.63193685e-01 7.42266715e-01
-9.61862862e-01 4.16173041e-01 -5.96230626e-01 -6.41200066e-01
-8.10296237e-02 -1.35435617e+00 1.13491058e+00 3.37806433e-01
9.74714011e-02 -6.81923389e-01 2.97898263e-01 7.33773053e-01
3.41167413e-02 1.63841531e-01 3.69825095e-01 6.44653589e-02
-8.35098743e-01 1.06895335e-01 -2.39357382e-01 9.12106454e-01
1.25017837e-02 1.32360145e-01 -8.66117060e-01 -3.05713117e-01
-7.50047565e-02 -8.06339324e-01 1.03355694e+00 6.41296387e-01
1.60840356e+00 -3.51568192e-01 3.78566086e-01 8.32684040e-01
1.15517783e+00 -1.22897811e-02 7.89484382e-01 -2.89199054e-02
7.61505187e-01 1.50494039e-01 8.96335423e-01 7.89334416e-01
7.79494166e-01 6.44842088e-01 6.28351271e-01 5.87936878e-01
-1.74232811e-01 -5.12912095e-01 9.53500032e-01 1.06412518e+00
-1.73426881e-01 -2.50310034e-01 -4.43239659e-01 4.19063658e-01
-2.18498635e+00 -1.11766946e+00 6.81366120e-03 1.93594241e+00
1.05627775e+00 -1.66151613e-01 2.10099921e-01 -2.59321332e-01
4.56187636e-01 4.48114038e-01 -9.27379072e-01 2.38560915e-01
9.30687562e-02 -1.56242251e-01 1.98799103e-01 4.95026618e-01
-1.13622475e+00 7.16694891e-01 5.16975784e+00 1.06477940e+00
-4.30200607e-01 3.49306285e-01 5.20920932e-01 -5.23718178e-01
-2.44770154e-01 -8.91043320e-02 -3.91673386e-01 9.80860233e-01
8.08018625e-01 -2.04480588e-01 6.98258638e-01 7.61258245e-01
3.03901792e-01 9.39561278e-02 -1.11718225e+00 1.14321065e+00
9.06326994e-03 -1.06326342e+00 2.29162678e-01 1.03347145e-01
7.97707021e-01 6.24948554e-02 3.25721949e-01 6.33204281e-01
5.28667033e-01 -7.25213408e-01 9.88946319e-01 9.80687320e-01
8.17372918e-01 -6.47295654e-01 5.55314183e-01 -5.81222735e-02
-1.14836121e+00 -2.44314134e-01 -7.62070239e-01 1.89532429e-01
3.34414154e-01 4.36598718e-01 3.41302231e-02 4.09896046e-01
8.24929774e-01 1.30937624e+00 -3.93830359e-01 8.76359284e-01
-4.42604423e-01 8.27644587e-01 1.24924630e-01 3.87544572e-01
2.48239428e-01 -5.86598337e-01 5.31281650e-01 7.92677701e-01
5.61382413e-01 2.65743911e-01 4.58003767e-03 1.05059683e+00
-4.03455675e-01 -2.32428834e-01 -1.00295536e-01 -2.36381099e-01
3.11317921e-01 1.08303547e+00 1.33921206e-01 -3.38400900e-01
-7.13865876e-01 1.05561399e+00 3.48820210e-01 3.66765499e-01
-1.28306293e+00 -9.67710686e-04 1.29762793e+00 -3.62862796e-02
3.78709912e-01 -9.49082822e-02 4.54229265e-01 -1.71051931e+00
1.03789479e-01 -1.02306235e+00 4.44397420e-01 -1.17587948e+00
-1.51906264e+00 -6.46198681e-03 1.61393017e-01 -1.55357945e+00
-2.60617912e-01 -5.77632725e-01 -4.81079638e-01 6.22401655e-01
-1.36321974e+00 -9.68737304e-01 -3.15980405e-01 8.43643427e-01
9.17152286e-01 -1.59369498e-01 3.40175062e-01 4.00107622e-01
-7.03899503e-01 4.70744908e-01 3.03240001e-01 3.20545137e-01
8.85076284e-01 -1.31269705e+00 2.08621860e-01 9.36729074e-01
1.92263350e-01 2.01376349e-01 5.87011635e-01 -4.77496445e-01
-1.67452681e+00 -1.19176483e+00 1.76481202e-01 -6.68365061e-01
1.05327284e+00 -2.31485397e-01 -9.95120943e-01 7.84277558e-01
3.30253541e-01 3.25133681e-01 6.93098366e-01 -2.37010241e-01
-5.12985766e-01 -3.01927596e-01 -8.00405145e-01 3.82076621e-01
1.08164775e+00 -6.65088296e-01 -4.40853149e-01 1.67476609e-01
8.85708392e-01 -3.48194569e-01 -1.04487574e+00 9.30036604e-02
6.91857517e-01 -1.06871259e+00 9.13716078e-01 -6.09117389e-01
9.33778882e-01 -4.38862056e-01 -4.32561576e-01 -1.45625865e+00
-4.55258876e-01 -5.22029698e-01 -8.71391773e-01 1.26692557e+00
-2.68526133e-02 -2.25472435e-01 4.30313915e-01 5.72789192e-01
-6.38793632e-02 -8.97877336e-01 -9.02830243e-01 -7.82587528e-01
9.04612094e-02 -5.33840954e-01 6.46210670e-01 6.17680371e-01
-3.36814135e-01 6.27292246e-02 -8.21699917e-01 2.40923941e-01
1.00049078e+00 -1.50698468e-01 8.34599495e-01 -8.95020187e-01
-7.98812449e-01 -4.81823146e-01 -4.18896079e-01 -1.18318975e+00
2.64193505e-01 -6.98957086e-01 4.99294810e-02 -1.37289870e+00
6.43262386e-01 1.55126587e-01 -5.61457455e-01 2.14093924e-01
-5.20535231e-01 8.46558586e-02 3.85038108e-01 3.03361118e-01
-1.09747934e+00 1.14560175e+00 1.41999173e+00 -2.77678311e-01
2.70940334e-01 -9.54345912e-02 -5.46802282e-01 6.62855864e-01
7.40962505e-01 -5.20585001e-01 -6.98472321e-01 -7.44227827e-01
3.91629130e-01 9.46026072e-02 4.89795953e-01 -9.69325840e-01
-2.05177162e-03 -4.75721151e-01 1.68179095e-01 -5.21097958e-01
4.92011636e-01 -5.63415647e-01 1.62316874e-01 2.10238203e-01
-4.13874298e-01 -1.32603213e-01 -2.21535563e-01 9.59747314e-01
-3.76287699e-01 -1.09463394e-01 6.76841974e-01 1.26905307e-01
-6.25265181e-01 9.96589005e-01 2.36891117e-02 4.66987312e-01
9.05502915e-01 1.98462844e-01 -1.37122273e-01 -5.61013401e-01
-4.49881852e-01 6.97452605e-01 4.09148306e-01 4.54631090e-01
7.49006629e-01 -1.82117927e+00 -9.65979457e-01 -2.55136732e-02
2.11267143e-01 3.34070802e-01 6.58465624e-01 8.44730020e-01
-3.50636274e-01 5.79437288e-03 -1.79025322e-01 -5.90516746e-01
-9.93631303e-01 4.30623800e-01 2.75900662e-01 -6.28131926e-02
-3.92209291e-01 8.48505735e-01 4.03382510e-01 -1.88983064e-02
4.21426177e-01 -3.23862582e-01 -3.44817601e-02 4.94438782e-03
7.67378747e-01 6.28278434e-01 -7.71885514e-01 -5.94223678e-01
-1.47344261e-01 3.24801266e-01 -2.03293399e-03 -1.34176403e-01
1.34966648e+00 -3.38967502e-01 2.13066742e-01 5.84874690e-01
1.31135130e+00 -2.99540490e-01 -2.24851704e+00 -4.02433872e-01
-4.93178785e-01 -8.38206768e-01 2.00465962e-01 -4.77252841e-01
-1.30783165e+00 1.10963583e+00 5.16335130e-01 -1.66740879e-01
1.13558877e+00 6.85127452e-02 8.84435296e-01 1.58607900e-01
2.47527137e-01 -1.10823202e+00 5.61456978e-01 1.89966470e-01
1.11088657e+00 -1.42786717e+00 -1.06488362e-01 -2.59454940e-02
-9.30111408e-01 7.78162122e-01 6.86702192e-01 -2.52160817e-01
6.54288590e-01 -1.20663136e-01 -1.78255379e-01 7.49831721e-02
-8.67882133e-01 1.83834396e-02 3.47524583e-01 3.03732276e-01
1.36902168e-01 2.09635034e-01 -1.02039769e-01 9.39858973e-01
1.31694704e-01 3.21531206e-01 6.11651123e-01 5.53301215e-01
-1.46742806e-01 -8.81889284e-01 5.15657179e-02 3.28719735e-01
-5.11720121e-01 4.78459662e-03 1.36059240e-01 4.59713280e-01
6.19609058e-02 7.06278026e-01 1.00146115e-01 -5.42303741e-01
2.33470589e-01 -2.63407733e-02 4.48598564e-01 -2.89233416e-01
1.13238551e-01 8.42809677e-04 -4.03529137e-01 -1.19214070e+00
-5.99493444e-01 -1.02781725e+00 -1.12928152e+00 -3.69971395e-01
-1.79388244e-02 4.25539315e-02 3.28264803e-01 6.92363679e-01
4.43153828e-01 5.62219262e-01 7.14008510e-01 -4.94363278e-01
-9.02946711e-01 -9.67727780e-01 -7.84055054e-01 5.56463838e-01
3.16920459e-01 -8.73468101e-01 -2.74506420e-01 2.37561196e-01] | [8.494704246520996, 0.7608508467674255] |
1987d4d5-9859-423a-acd8-55b8c6a2b17e | visual-relationship-detection-with-low-rank | 1911.09895 | null | https://arxiv.org/abs/1911.09895v1 | https://arxiv.org/pdf/1911.09895v1.pdf | Visual Relationship Detection with Low Rank Non-Negative Tensor Decomposition | We address the problem of Visual Relationship Detection (VRD) which aims to describe the relationships between pairs of objects in the form of triplets of (subject, predicate, object). We observe that given a pair of bounding box proposals, objects often participate in multiple relations implying the distribution of triplets is multimodal. We leverage the strong correlations within triplets to learn the joint distribution of triplet variables conditioned on the image and the bounding box proposals, doing away with the hitherto used independent distribution of triplets. To make learning the triplet joint distribution feasible, we introduce a novel technique of learning conditional triplet distributions in the form of their normalized low rank non-negative tensor decompositions. Normalized tensor decompositions take form of mixture distributions of discrete variables and thus are able to capture multimodality. This allows us to efficiently learn higher order discrete multimodal distributions and at the same time keep the parameter size manageable. We further model the probability of selecting an object proposal pair and include a relation triplet prior in our model. We show that each part of the model improves performance and the combination outperforms state-of-the-art score on the Visual Genome (VG) and Visual Relationship Detection (VRD) datasets. | ['Zhen Zhang', 'Mohammed Haroon Dupty', 'Wee Sun Lee'] | 2019-11-22 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [-2.14094520e-02 -9.51701775e-02 -2.11320654e-01 -2.56103069e-01
-6.07061148e-01 -9.06861842e-01 8.63882661e-01 3.72195661e-01
-1.20678842e-01 3.23209673e-01 3.60639915e-02 -2.40601182e-01
-2.76956767e-01 -6.42773867e-01 -8.90871108e-01 -8.10508847e-01
-2.08777532e-01 1.24370646e+00 2.32575804e-01 8.51890296e-02
-3.34582776e-02 5.19810140e-01 -1.50623190e+00 6.01883948e-01
4.23870116e-01 8.77373099e-01 2.01801941e-01 8.03290069e-01
-5.65052032e-04 6.55785918e-01 -4.14169937e-01 -5.38884759e-01
2.53824651e-01 4.24374491e-02 -8.29932153e-01 4.49016869e-01
8.03266823e-01 -1.68520510e-01 -1.26276478e-01 8.35814178e-01
6.66366890e-02 1.17126420e-01 1.33737624e+00 -1.67899299e+00
-7.05712497e-01 2.96530843e-01 -1.16907203e+00 -3.11921746e-03
3.40104729e-01 -9.52420682e-02 1.68884492e+00 -8.48344386e-01
8.35366845e-01 1.58424938e+00 2.16236025e-01 -3.64084281e-02
-1.77465475e+00 -1.10828638e-01 2.30017960e-01 2.62764007e-01
-1.43309438e+00 -1.35687947e-01 5.00162899e-01 -8.88423443e-01
8.93545508e-01 2.94859976e-01 4.82547939e-01 9.64532435e-01
-3.12293209e-02 7.92771816e-01 8.08865249e-01 -3.77097249e-01
4.48904485e-02 -1.72151700e-02 2.03734040e-01 8.33092153e-01
3.92133683e-01 -3.30828637e-01 -6.65243268e-01 -5.64676821e-01
7.51662016e-01 -2.50269417e-02 5.12676537e-02 -1.02886963e+00
-1.44403732e+00 7.23494947e-01 3.31983745e-01 2.31699310e-02
-3.20873886e-01 2.84416467e-01 1.28212750e-01 6.55147955e-02
1.82910562e-02 2.85958182e-02 -3.12871844e-01 9.74627212e-02
-4.76221412e-01 2.74688661e-01 7.58605599e-01 1.05605268e+00
9.74815905e-01 -6.93893969e-01 -3.14689547e-01 7.93596804e-01
5.79204440e-01 4.44985151e-01 -3.31988484e-01 -9.10623074e-01
5.84773362e-01 6.28878772e-01 2.11027712e-01 -1.13265216e+00
-1.70081466e-01 -1.97845846e-01 -7.29813814e-01 -3.29255946e-02
6.10526562e-01 3.61157745e-01 -9.81924951e-01 1.78841150e+00
5.56889117e-01 -1.92437872e-01 -2.32699454e-01 7.31526673e-01
6.40223384e-01 5.22873104e-01 -7.83940405e-03 3.61815803e-02
1.74400973e+00 -5.25955319e-01 -4.97451395e-01 -4.16448601e-02
5.13664961e-01 -8.14939916e-01 7.74896324e-01 3.98653626e-01
-9.60335135e-01 -2.50402093e-01 -7.03783453e-01 -3.77397001e-01
-2.59792954e-01 3.67006183e-01 8.69533539e-01 2.60273516e-01
-9.73282397e-01 1.88048244e-01 -1.11318302e+00 -3.64021122e-01
4.35608953e-01 4.48439270e-01 -6.39300287e-01 -2.54907221e-01
-5.40757179e-01 8.43517065e-01 3.26661527e-01 -8.17620680e-02
-9.52769458e-01 -3.88656557e-01 -7.58036852e-01 -5.55696292e-03
5.42004049e-01 -9.94843662e-01 7.87576973e-01 -3.46010000e-01
-6.65154397e-01 1.06179047e+00 -3.23488861e-01 4.16667759e-03
2.23079324e-01 6.81651831e-02 1.05220392e-01 4.63634282e-01
6.91360235e-02 8.31700325e-01 7.84603000e-01 -1.59402537e+00
-5.66540837e-01 -3.84716243e-01 1.03247911e-01 2.28357986e-01
-2.31105492e-01 7.71911889e-02 -7.63131499e-01 -1.84379667e-01
4.42148894e-01 -1.10835862e+00 1.37305558e-01 3.05582166e-01
-7.30864942e-01 -5.24027050e-01 6.37658536e-01 -5.22223473e-01
7.54405856e-01 -2.17003942e+00 7.23807991e-01 4.03558224e-01
5.18422067e-01 -2.83832341e-01 -3.02611738e-01 5.72699130e-01
-1.24706239e-01 1.45115256e-01 -1.25573203e-01 -6.89063549e-01
2.58207530e-01 7.86036313e-01 -1.50932252e-01 6.72484994e-01
5.13538539e-01 7.84215987e-01 -8.57649386e-01 -6.81874692e-01
-5.59140295e-02 6.59529567e-01 -5.05863369e-01 2.79905617e-01
-4.04805154e-01 3.70680898e-01 -4.68152583e-01 6.52023494e-01
8.59799504e-01 -6.14460230e-01 4.95467812e-01 -7.55104661e-01
1.66063562e-01 1.97051287e-01 -1.29342914e+00 1.49885309e+00
2.85404269e-02 4.26846772e-01 -4.86762859e-02 -8.30711424e-01
6.55538023e-01 1.76257789e-01 4.66033578e-01 -1.07771426e-01
-3.89632769e-02 -6.02190010e-02 1.64892316e-01 -5.44581532e-01
3.98066133e-01 -7.00085387e-02 1.05202839e-01 3.61994207e-01
3.32158267e-01 7.91960303e-03 4.88142103e-01 6.84947312e-01
9.34098840e-01 3.88494283e-01 8.42026919e-02 7.88180381e-02
1.98572919e-01 -2.84155756e-01 3.68811578e-01 7.42117703e-01
2.57715464e-01 5.68191588e-01 1.22721732e+00 -9.44837183e-02
-1.04148376e+00 -1.40708184e+00 -5.25744818e-02 1.04042578e+00
9.01990160e-02 -5.75658739e-01 -3.69252712e-02 -6.61880255e-01
2.21276045e-01 2.91998684e-01 -8.70994806e-01 2.08452135e-01
-4.43450600e-01 -7.73060381e-01 1.23246580e-01 4.46597904e-01
-2.18190163e-01 -3.42411995e-01 -3.83664817e-01 -3.35072279e-01
-4.30324942e-01 -1.38594234e+00 -4.01548058e-01 4.31686580e-01
-5.83775878e-01 -1.13708842e+00 -4.18762892e-01 -6.32839143e-01
9.44273531e-01 1.79475754e-01 1.34498870e+00 -9.12780464e-02
-3.53022695e-01 7.42918015e-01 -2.91122556e-01 8.81641358e-02
-2.19690770e-01 -3.07418853e-01 -7.52381906e-02 3.10156733e-01
5.31404987e-02 -6.69596791e-01 -2.70095676e-01 2.33943015e-01
-1.02365136e+00 5.30811287e-02 6.36689425e-01 9.84495163e-01
7.92948008e-01 -5.61860986e-02 -4.08353955e-01 -6.35559440e-01
1.02930248e-01 -5.11797249e-01 -6.76312685e-01 7.04186738e-01
1.50876924e-01 3.66227955e-01 1.05878145e-01 -7.30058968e-01
-7.58587360e-01 1.14517763e-01 6.83080196e-01 -7.20543146e-01
-8.46500397e-02 5.71760893e-01 -2.49162808e-01 3.35945785e-02
3.22383851e-01 -1.40071958e-01 -6.04084954e-02 -6.35192931e-01
6.46085739e-01 1.04071386e-01 4.32944030e-01 -9.57203805e-01
9.05365169e-01 7.09881663e-01 6.86997116e-01 -6.15125000e-01
-7.07606673e-01 -6.96468651e-01 -9.43192482e-01 -4.06663585e-03
8.24311972e-01 -9.12153244e-01 -1.08484161e+00 -3.34700160e-02
-1.33576715e+00 6.28290251e-02 1.11183003e-01 3.98398101e-01
-4.12447810e-01 5.78245759e-01 -5.42581081e-01 -8.69575799e-01
4.32908982e-01 -1.10165024e+00 1.44830966e+00 -1.83249384e-01
-3.31175923e-02 -9.83330369e-01 9.19576883e-02 4.11227375e-01
-2.65920520e-01 3.55600476e-01 1.39871669e+00 -4.88583595e-01
-9.18773890e-01 -1.02502078e-01 -4.46503222e-01 4.27435264e-02
3.02000996e-02 4.02466804e-01 -6.70405865e-01 -2.37085983e-01
-3.84350628e-01 -3.34544629e-01 1.05762672e+00 2.51198739e-01
7.06668556e-01 -1.38739258e-01 -5.86668015e-01 3.54379237e-01
1.34761453e+00 -3.70189369e-01 3.40865284e-01 -1.70759141e-01
1.07019031e+00 9.28386509e-01 6.71234787e-01 4.61250514e-01
6.94843173e-01 9.53361452e-01 8.42374682e-01 7.05980808e-02
2.65560262e-02 -1.39482364e-01 2.46219948e-01 3.27324390e-01
-7.32558444e-02 -5.21482170e-01 -9.97631311e-01 5.93156755e-01
-2.06937623e+00 -8.21311057e-01 -4.16770071e-01 2.36017704e+00
7.10601747e-01 -1.14353374e-01 4.25189346e-01 -1.77783117e-01
6.00246191e-01 -1.08661108e-01 -1.33111566e-01 -1.15212284e-01
-3.21907878e-01 -1.96155891e-01 2.93939501e-01 4.95981425e-01
-1.10882998e+00 5.80030262e-01 5.88576746e+00 5.55413902e-01
-5.80295384e-01 -6.67063445e-02 4.00024921e-01 -1.11568682e-02
-5.84322810e-01 3.00244182e-01 -8.03364217e-01 2.19580568e-02
4.73650455e-01 3.90497833e-01 3.72218400e-01 2.20058322e-01
-2.26031870e-01 -3.29551786e-01 -1.56672812e+00 9.61484611e-01
-5.92462122e-02 -1.02310383e+00 4.26537812e-01 5.01229584e-01
5.48578084e-01 -8.00427422e-02 3.93504322e-01 -1.45889908e-01
4.39013720e-01 -9.47883487e-01 7.92048573e-01 6.32083595e-01
4.47372586e-01 -5.65091789e-01 3.78755391e-01 1.97603136e-01
-1.14435649e+00 -1.36941625e-02 -4.32871848e-01 1.78123534e-01
1.60683692e-01 5.79348445e-01 -1.08162344e+00 8.11739206e-01
5.79296112e-01 7.51942515e-01 -5.73759437e-01 9.35017645e-01
-4.47912216e-02 3.51348698e-01 -6.95532262e-01 2.09504157e-01
7.14368671e-02 -4.53790307e-01 7.31767416e-01 9.90910470e-01
2.54936554e-02 -1.50411621e-01 4.52686965e-01 1.04483080e+00
5.86287268e-02 -6.28880486e-02 -3.68917376e-01 -2.59276956e-01
2.65695035e-01 1.34434676e+00 -7.64499605e-01 -2.06124559e-01
-4.77539510e-01 7.13737786e-01 6.43969297e-01 4.62951839e-01
-6.99363112e-01 3.96807939e-01 6.57185733e-01 -1.16404690e-01
8.60946774e-01 -5.43749452e-01 -1.78758442e-01 -1.19710159e+00
3.30647171e-01 -6.74719930e-01 6.17241323e-01 -6.85736954e-01
-1.60006809e+00 3.04045260e-01 3.89428467e-01 -1.11126733e+00
-1.29035756e-01 -8.10278177e-01 -1.71325952e-01 8.94589424e-01
-1.36790824e+00 -1.57288301e+00 -3.33798565e-02 5.64719021e-01
-2.82884330e-01 2.08115280e-01 8.23871374e-01 1.87629268e-01
-5.54588020e-01 3.28668892e-01 -1.59540787e-01 -1.80017091e-02
6.51188612e-01 -1.53875566e+00 4.35972661e-02 6.38713419e-01
4.33931917e-01 8.55014563e-01 9.02290404e-01 -7.07380950e-01
-1.57354355e+00 -7.19606161e-01 8.79804492e-01 -8.16467762e-01
7.96380818e-01 -7.07885563e-01 -6.81991279e-01 7.97168672e-01
3.93661410e-02 3.51022094e-01 7.98301041e-01 4.10572439e-01
-9.22077000e-01 6.94964230e-02 -8.85811508e-01 5.32007039e-01
8.96978796e-01 -7.27601111e-01 -3.47102344e-01 6.20218337e-01
4.12391216e-01 -2.39349708e-01 -1.00724614e+00 3.79651308e-01
5.13305366e-01 -9.95916545e-01 1.15514159e+00 -8.91728759e-01
3.72113228e-01 -6.84902132e-01 -4.36951250e-01 -8.17056000e-01
-3.31571966e-01 -4.39324826e-01 -4.53999758e-01 1.31985414e+00
3.98528308e-01 -3.11320424e-01 4.99665439e-01 5.16569138e-01
2.10104585e-01 -7.38296688e-01 -9.83110249e-01 -7.01624095e-01
-2.53122777e-01 -1.02526613e-01 2.55145580e-01 8.23457837e-01
-1.29751086e-01 5.48797965e-01 -4.79724824e-01 5.68362772e-01
7.98248827e-01 4.06571418e-01 7.69485772e-01 -1.17556572e+00
-8.79922628e-01 -2.06233740e-01 -7.28843272e-01 -1.13568521e+00
8.52140710e-02 -8.59996259e-01 -1.02681495e-01 -1.55369532e+00
7.83361495e-01 -3.76914054e-01 -1.31806985e-01 7.17283607e-01
-9.46623832e-02 3.41884077e-01 2.36377671e-01 2.56545693e-01
-8.67728710e-01 3.99117738e-01 1.27941871e+00 -3.00589859e-01
1.49211019e-01 -2.14077890e-01 -5.02082288e-01 4.76781309e-01
1.30482852e-01 -5.54505885e-01 -3.51225376e-01 -5.38055062e-01
5.52785099e-01 1.81086719e-01 7.05156386e-01 -2.01622561e-01
1.35979325e-01 -1.59475595e-01 3.90440911e-01 -8.15950692e-01
7.96248615e-01 -8.54950249e-01 3.47634524e-01 3.11873201e-02
-3.14655036e-01 -6.52561188e-02 -1.46999225e-01 7.26312459e-01
-7.46935904e-02 1.13454528e-01 4.63088065e-01 8.34619701e-02
-1.24902494e-01 3.11285138e-01 -2.16083109e-01 -1.44794822e-01
7.61512995e-01 -1.68357622e-02 -3.99419069e-01 -2.98781276e-01
-8.61463308e-01 3.34093422e-01 6.46444499e-01 2.57708818e-01
4.58415210e-01 -1.32753158e+00 -6.11657798e-01 -1.03488788e-01
4.03677046e-01 1.07343495e-01 9.18786004e-02 1.09964931e+00
-1.12104438e-01 1.08775906e-01 -1.49025872e-01 -1.09318841e+00
-1.50532663e+00 6.29908204e-01 1.50081038e-01 -3.38342994e-01
-2.20721781e-01 1.14034343e+00 5.97541690e-01 -3.46869648e-01
2.43269071e-01 -2.52464563e-01 -3.48864377e-01 3.17533463e-01
1.73690319e-01 2.05801949e-01 -2.20889956e-01 -9.78247523e-01
-6.10795021e-01 6.07966304e-01 -2.77733296e-01 -1.05740853e-01
1.36064672e+00 -7.50888810e-02 -7.07083344e-01 6.59951568e-01
1.31979930e+00 -2.15861518e-02 -1.15031910e+00 -3.09869558e-01
2.87268986e-03 -5.88558137e-01 -3.40782642e-01 -6.49765193e-01
-7.61572719e-01 8.24916780e-01 3.15376729e-01 2.01286912e-01
7.83192039e-01 6.19470298e-01 1.67335570e-01 3.35833639e-01
1.70379251e-01 -4.25786763e-01 3.83055687e-01 3.57720971e-01
7.52995193e-01 -1.08637953e+00 2.28437230e-01 -9.77241576e-01
-6.83207631e-01 1.00524700e+00 2.21389458e-01 4.09905203e-02
4.55218434e-01 1.55596569e-01 -2.80324817e-01 -4.04900372e-01
-1.07771420e+00 -4.14112300e-01 8.13594639e-01 7.63421357e-01
4.11713392e-01 1.76821113e-01 4.28218693e-02 7.47086704e-02
2.33665586e-01 -5.87762415e-01 3.56550246e-01 7.54031122e-01
-6.75773323e-02 -1.43983066e+00 -5.00722528e-01 4.50580478e-01
-3.52492869e-01 -7.99110457e-02 -4.84765977e-01 5.07233083e-01
2.66143143e-01 9.00004923e-01 1.82361856e-01 -1.15530789e-01
3.95239778e-02 -1.41564578e-01 1.02447760e+00 -6.08494759e-01
-1.69628084e-01 3.81759524e-01 2.61305958e-01 -3.76668483e-01
-5.92109203e-01 -1.02897441e+00 -1.07818305e+00 4.52362634e-02
-3.39227378e-01 1.22272279e-02 4.14864928e-01 9.78189051e-01
2.67744690e-01 1.91764504e-01 3.32950741e-01 -8.52081120e-01
-5.43636620e-01 -6.95121109e-01 -7.10276723e-01 7.07158864e-01
4.61019188e-01 -1.13265181e+00 -2.03336388e-01 1.47288635e-01] | [10.298783302307129, 1.634637713432312] |
af38b7eb-14cf-4961-9e82-86e02010c84f | object-detection-with-pixel-intensity | 1305.4537 | null | http://arxiv.org/abs/1305.4537v5 | http://arxiv.org/pdf/1305.4537v5.pdf | Object Detection with Pixel Intensity Comparisons Organized in Decision Trees | We describe a method for visual object detection based on an ensemble of
optimized decision trees organized in a cascade of rejectors. The trees use
pixel intensity comparisons in their internal nodes and this makes them able to
process image regions very fast. Experimental analysis is provided through a
face detection problem. The obtained results are encouraging and demonstrate
that the method has practical value. Additionally, we analyse its sensitivity
to noise and show how to perform fast rotation invariant object detection.
Complete source code is provided at https://github.com/nenadmarkus/pico. | ['Robert Forchheimer', 'Jörgen Ahlberg', 'Igor S. Pandžić', 'Miroslav Frljak', 'Nenad Markuš'] | 2013-05-20 | null | null | null | null | ['pico'] | ['natural-language-processing'] | [ 2.08404630e-01 -3.37019823e-02 -2.46961236e-01 -2.75454491e-01
-3.71678621e-01 -5.36580980e-01 4.99422431e-01 -1.22322261e-01
-3.83407682e-01 3.45447570e-01 -3.68715763e-01 -2.45635778e-01
3.30228925e-01 -5.78925073e-01 -3.04382205e-01 -1.02013588e+00
-1.44222707e-01 2.96693236e-01 7.86152959e-01 9.66109894e-03
3.24692398e-01 1.07250690e+00 -1.67045939e+00 5.52330613e-01
2.57559657e-01 8.65632951e-01 1.25235304e-01 9.91567969e-01
1.57113612e-01 5.17973602e-01 -5.47640383e-01 -3.21038425e-01
3.74979377e-01 -2.05588162e-01 -6.42963052e-01 2.75060236e-01
5.62508464e-01 -3.19184303e-01 6.58844113e-02 1.18233407e+00
7.01811790e-01 5.69934361e-02 9.77804959e-01 -1.17031181e+00
-1.90258697e-01 3.92774820e-01 -8.05934489e-01 6.60551906e-01
1.65980265e-01 1.55414820e-01 7.26926446e-01 -1.29753602e+00
5.98279417e-01 1.46539652e+00 5.19896209e-01 5.21232128e-01
-1.38969743e+00 -6.00444138e-01 -6.04200475e-02 3.75582278e-01
-1.56009686e+00 -6.89015925e-01 7.80810475e-01 -3.97970855e-01
8.49043071e-01 5.26246905e-01 6.31348968e-01 8.68080676e-01
2.43840143e-01 9.70026016e-01 1.34624612e+00 -7.63835371e-01
2.11841092e-01 3.30796003e-01 8.10075458e-03 1.21651554e+00
2.77747244e-01 3.85823280e-01 -3.10885489e-01 -6.99659809e-02
8.75007987e-01 -1.93195082e-02 -4.29902934e-02 -4.99437183e-01
-6.09596431e-01 9.61662471e-01 7.98260510e-01 2.45344251e-01
-1.66503444e-01 -8.35879799e-03 2.14918181e-01 1.39170021e-01
6.21921599e-01 -1.27632782e-01 -4.40120064e-02 7.12986171e-01
-9.68095720e-01 -1.49492145e-01 5.88716745e-01 5.65952420e-01
3.58545214e-01 1.13262601e-01 -6.93648905e-02 8.76532912e-01
3.49056870e-01 5.46392202e-01 1.32265329e-01 -8.85051966e-01
-1.83204532e-01 3.93141210e-01 -7.29054064e-02 -9.06445026e-01
-5.16905606e-01 -3.26335222e-01 -7.32275009e-01 1.27489638e+00
4.64708209e-01 4.50397246e-02 -1.32215226e+00 8.24776649e-01
6.83559835e-01 -1.19587742e-01 -3.78387988e-01 1.00496781e+00
9.65516746e-01 4.47376937e-01 1.21833742e-01 -8.33117589e-02
1.65403032e+00 -9.65934813e-01 -4.99937028e-01 -2.22912222e-01
2.83784688e-01 -9.29429531e-01 6.11751437e-01 5.87845266e-01
-9.98946369e-01 -7.21539319e-01 -1.03423047e+00 8.25980771e-03
-4.94383633e-01 5.55096149e-01 7.00752199e-01 9.81948078e-01
-1.11601388e+00 3.71056706e-01 -7.23496020e-01 -6.78851545e-01
6.72755241e-01 4.53114450e-01 -2.47356400e-01 1.31886080e-01
-4.33817983e-01 9.18082595e-01 4.56023306e-01 5.05647004e-01
-8.42209280e-01 3.26638371e-01 -6.16358638e-01 -2.09425643e-01
1.84208944e-01 -5.45735776e-01 1.27674139e+00 -1.18570173e+00
-1.37758291e+00 1.33876812e+00 -2.50533789e-01 -3.43471110e-01
7.75367975e-01 8.92925262e-03 -2.05592439e-01 4.50288206e-01
-1.94407299e-01 7.64105082e-01 1.40117514e+00 -1.19519842e+00
-9.47449327e-01 -4.27223742e-01 -3.63221884e-01 2.10365623e-01
-1.48176536e-01 5.88745594e-01 -6.67544425e-01 -7.33637810e-01
2.03268945e-01 -7.44406581e-01 -3.85613292e-01 3.77241284e-01
-4.38704252e-01 -3.06912869e-01 8.95940423e-01 -5.97473383e-01
9.58062053e-01 -2.12091684e+00 -1.36033908e-01 5.12665331e-01
3.75354826e-01 3.26092064e-01 -1.59251809e-01 1.24461748e-01
-1.76560804e-01 6.01119585e-02 -2.30321005e-01 -1.94993794e-01
-4.20089781e-01 -8.55558664e-02 -1.22079991e-01 8.33644927e-01
2.32571289e-01 6.09423518e-01 -4.72394496e-01 -8.31038654e-01
7.17413127e-01 5.58780968e-01 3.62417959e-02 -7.75710419e-02
2.20458895e-01 6.69224039e-02 -4.34499979e-01 1.19708645e+00
9.10792232e-01 -5.05761728e-02 1.06167920e-01 -6.38221279e-02
-1.01237811e-01 -2.93267578e-01 -1.33402395e+00 7.68141448e-01
-3.04274354e-03 7.56627142e-01 4.56190914e-01 -8.72367144e-01
9.62028146e-01 -4.46579009e-02 8.35669786e-02 -3.98442864e-01
4.73574609e-01 -3.64028551e-02 -1.83712207e-02 -3.03721994e-01
1.47352219e-01 -3.56368609e-02 3.88824344e-01 1.23083964e-01
1.23619534e-01 -2.31794670e-01 2.14900881e-01 9.00122616e-03
8.81096601e-01 1.12312332e-01 6.26159728e-01 -2.52429068e-01
7.31069565e-01 6.74494449e-03 3.32863450e-01 8.72483909e-01
-6.00453496e-01 7.45579898e-01 2.97198474e-01 -5.76922596e-01
-7.53892183e-01 -1.27077687e+00 -4.33180630e-01 1.27442789e+00
2.05903530e-01 -1.36983961e-01 -8.79256487e-01 -7.19515085e-01
-3.17190997e-02 3.60236764e-01 -8.54227245e-01 2.92180866e-01
-4.57189322e-01 -9.75032330e-01 2.10358545e-01 6.50504351e-01
3.27409416e-01 -1.15818238e+00 -9.31748748e-01 -4.35828678e-02
2.07105726e-01 -9.56967711e-01 2.02561747e-02 3.98030281e-01
-1.02229095e+00 -1.09768510e+00 -7.90889680e-01 -1.00382268e+00
9.02305424e-01 4.37521040e-01 9.47966099e-01 4.90842789e-01
-1.00395453e+00 2.40584493e-01 -3.61349940e-01 -6.10495925e-01
-2.70090312e-01 -5.28939106e-02 -9.87057388e-02 -2.89899912e-02
4.87756640e-01 -1.88151389e-01 -6.67546272e-01 6.17869556e-01
-5.25756836e-01 -3.69847752e-02 7.35230446e-01 6.64217830e-01
5.94391942e-01 7.34938085e-02 -8.73537362e-02 -8.69783878e-01
1.09982535e-01 9.66119468e-02 -1.06704116e+00 1.34695336e-01
-4.24453467e-01 -1.55883357e-01 5.29489554e-02 -2.96462119e-01
-1.08817601e+00 7.53057659e-01 -7.44087249e-02 -2.36825958e-01
-3.43654990e-01 -4.74857688e-01 2.94694781e-01 -5.34759462e-01
9.10421610e-01 -2.46406317e-01 -5.19043319e-02 -4.06867176e-01
3.62738609e-01 6.41814351e-01 3.60021919e-01 5.09216487e-02
9.70403969e-01 8.96163166e-01 2.89302375e-02 -1.20910251e+00
-3.36721480e-01 -7.41743922e-01 -8.72332752e-01 -7.48633981e-01
7.26320863e-01 -6.35990977e-01 -4.97721255e-01 4.14300472e-01
-9.98133719e-01 -3.44665229e-01 -9.50600803e-02 2.58006573e-01
-4.03824896e-01 2.95645684e-01 -6.18766069e-01 -1.24210370e+00
-5.40888071e-01 -8.35586071e-01 1.05112183e+00 5.07894754e-01
6.21119374e-03 -9.27466631e-01 -2.72701234e-01 2.79345632e-01
8.97276029e-02 2.68101066e-01 2.53341675e-01 -3.88577670e-01
-5.77135146e-01 -3.18323582e-01 -3.33336473e-01 2.54127055e-01
-1.13092072e-01 5.68830729e-01 -1.30091131e+00 -3.99958313e-01
-1.53728724e-01 -7.12250993e-02 1.47474635e+00 7.01876819e-01
9.27439153e-01 8.82788934e-03 -8.20598960e-01 3.69850248e-01
1.27252758e+00 -1.03715703e-01 5.33601224e-01 2.68791676e-01
3.92456383e-01 9.06058252e-01 6.04252040e-01 1.03622042e-01
-4.88341749e-01 6.90087497e-01 3.87829512e-01 -7.69865215e-01
-4.66753274e-01 2.90144742e-01 3.54834020e-01 9.36286077e-02
-4.15622622e-01 -7.09782839e-02 -9.59712803e-01 3.95137727e-01
-1.60906184e+00 -1.03233540e+00 -4.49017942e-01 2.10148835e+00
1.98531225e-01 2.18896866e-01 4.38031971e-01 3.49640161e-01
9.93672132e-01 4.82189804e-02 -2.25765482e-01 -7.43207693e-01
-2.23855421e-01 3.36608499e-01 6.71395600e-01 6.29886866e-01
-1.63494086e+00 1.11194360e+00 7.20333290e+00 9.51415479e-01
-1.05187738e+00 2.29959816e-01 8.90221655e-01 -8.84260610e-02
5.57458222e-01 -2.53753245e-01 -9.40220058e-01 1.54170990e-01
4.47430342e-01 1.88671052e-01 -1.31281748e-01 9.31199133e-01
1.20044902e-01 -6.75606787e-01 -5.38534641e-01 8.38070452e-01
1.79982290e-01 -9.18200076e-01 -2.03544453e-01 -7.63861015e-02
3.16741526e-01 1.54187992e-01 1.52822778e-01 -1.68240592e-01
3.31661433e-01 -1.12658763e+00 6.80135846e-01 1.67563006e-01
5.71247935e-01 -6.69481635e-01 7.31049001e-01 1.45524200e-02
-1.28122747e+00 -2.09248781e-01 -4.85214233e-01 4.70401533e-02
-2.75434375e-01 6.34741366e-01 -1.24062097e+00 1.50449902e-01
1.03432858e+00 3.23099524e-01 -1.10865426e+00 1.39289367e+00
-5.60943604e-01 5.64633429e-01 -6.93309367e-01 -1.84814557e-01
-1.33936286e-01 -1.41957805e-01 5.90659738e-01 1.57397270e+00
-2.84403469e-02 2.10816711e-02 1.18295200e-01 4.52211797e-01
3.81005198e-01 3.38284642e-01 -7.35159814e-01 5.45869887e-01
3.44837643e-02 1.80410230e+00 -1.64124238e+00 -3.04947495e-01
-1.73795417e-01 1.18590581e+00 1.91865042e-01 3.16719323e-01
-5.75403094e-01 -3.93664926e-01 2.47828722e-01 1.40042722e-01
6.42002523e-01 1.02269664e-01 -3.13104510e-01 -8.52880478e-01
-5.84923625e-02 -5.57286918e-01 7.19064116e-01 -7.79516578e-01
-9.73426759e-01 6.73250139e-01 -4.18993942e-02 -1.22804284e+00
1.40481396e-02 -1.10240078e+00 -5.81929803e-01 5.34992576e-01
-1.26223409e+00 -1.20824718e+00 -1.64817274e-01 4.36549604e-01
5.54313481e-01 -1.40852435e-02 6.29578590e-01 -4.64187413e-02
-6.69137001e-01 4.88813370e-01 6.45417050e-02 3.16865504e-01
4.70562696e-01 -1.20787024e+00 3.09595048e-01 1.07948685e+00
4.43940878e-01 1.49981618e-01 8.88510942e-01 -3.20056319e-01
-8.09073091e-01 -8.19374442e-01 5.09069681e-01 -5.29030085e-01
2.49517769e-01 -6.07387245e-01 -8.03467751e-01 4.27902162e-01
2.48127609e-01 3.20801973e-01 4.52377707e-01 -8.96341354e-03
-2.12226391e-01 -1.38882250e-01 -1.38039839e+00 4.70013082e-01
7.96656668e-01 -2.27302476e-03 -4.26250070e-01 5.67528427e-01
-2.62610137e-01 -2.77343571e-01 -2.33019993e-01 5.35896122e-01
7.15600610e-01 -1.23321033e+00 9.49698269e-01 -4.21854928e-02
-2.27381170e-01 -6.02203131e-01 2.91159928e-01 -7.30062842e-01
-4.03321356e-01 -3.22381109e-01 5.11618815e-02 9.60804522e-01
4.76198882e-01 -7.10884452e-01 9.08296525e-01 -1.29988387e-01
5.25667667e-01 -4.67269838e-01 -1.10392046e+00 -6.77642345e-01
-3.06999087e-01 -1.88938558e-01 -3.44943970e-01 4.34404910e-01
-4.82519120e-01 3.65925729e-01 1.03015509e-02 3.02835435e-01
9.51226592e-01 1.78786770e-01 5.27613223e-01 -1.33674371e+00
-5.94831854e-02 -6.71111643e-01 -7.12701142e-01 -4.90412116e-01
-6.42428026e-02 -7.11897552e-01 6.74863486e-03 -1.40473139e+00
2.11639404e-01 -1.18732765e-01 -2.41062015e-01 6.87165082e-01
-1.57335714e-01 1.01855540e+00 2.55515069e-01 1.70424923e-01
-4.94661748e-01 9.40233469e-02 7.23335445e-01 -1.50304794e-01
-2.28073206e-02 2.15528488e-01 -1.60930768e-01 1.06067681e+00
1.34559166e+00 -6.94256783e-01 2.00683266e-01 1.07853867e-01
-4.60879594e-01 -4.00654197e-01 5.96388161e-01 -1.20989275e+00
1.15171961e-01 8.01835675e-03 1.17809248e+00 -5.40367842e-01
4.87631589e-01 -7.98896790e-01 -1.05082735e-01 9.40965533e-01
5.25735021e-02 -7.92121887e-02 3.36519390e-01 4.71505791e-01
-9.06592458e-02 -4.16391253e-01 1.33524668e+00 7.33562633e-02
-1.13120830e+00 -8.33994523e-02 -7.78876781e-01 -7.02479422e-01
1.35239887e+00 -5.38168609e-01 -5.16913719e-02 -1.56170949e-01
-1.01250863e+00 1.26571834e-01 4.40451920e-01 3.01488161e-01
6.74258053e-01 -8.62592220e-01 -8.46062839e-01 4.65126097e-01
1.23685353e-01 -6.01302624e-01 1.63289867e-02 6.32241845e-01
-8.33990932e-01 9.25282091e-02 -4.41237301e-01 -8.87802958e-01
-2.21215963e+00 5.33128202e-01 6.08697653e-01 1.22813813e-01
-6.50958240e-01 1.10565782e+00 1.39881134e-01 -1.15254775e-01
2.47917280e-01 -3.35576646e-02 -4.20330971e-01 1.76830292e-01
7.58266687e-01 3.26820374e-01 6.89059868e-02 -6.22541904e-01
-6.07727706e-01 7.78065503e-01 -3.65244001e-02 -2.34925538e-01
9.81791079e-01 -7.67523795e-02 -1.59671947e-01 1.72098547e-01
7.86702275e-01 -6.32739738e-02 -1.02809298e+00 2.07241923e-02
-4.77181142e-03 -5.02605975e-01 2.52762705e-01 -6.09501302e-01
-1.17437792e+00 9.17406321e-01 1.35815775e+00 1.70102835e-01
1.49359632e+00 1.43135682e-01 -1.48896396e-01 4.21970367e-01
1.31603107e-01 -1.14728761e+00 -8.95297304e-02 1.69885084e-01
8.94035101e-01 -1.33239210e+00 5.22924304e-01 -9.70316112e-01
-3.92134219e-01 1.39704216e+00 5.23260057e-01 -2.70399213e-01
6.27525151e-01 5.80730021e-01 3.82252783e-01 -7.92515203e-02
-5.64766288e-01 -7.19890416e-01 1.81347027e-01 8.79255056e-01
3.38761568e-01 1.68751672e-01 -2.55030870e-01 -3.60775024e-01
9.29257795e-02 -2.32218519e-01 3.79792511e-01 8.95765483e-01
-7.52000451e-01 -9.37532842e-01 -8.77636969e-01 3.20553601e-01
-5.73910773e-01 1.15529507e-01 -5.77203989e-01 8.36858571e-01
6.72597904e-03 9.77127612e-01 -2.19407072e-03 -4.46687043e-02
1.11349970e-01 3.90365012e-02 6.32538915e-01 -5.75563371e-01
-5.65487564e-01 3.07767123e-01 -5.74444346e-02 -6.06583953e-01
-4.76593882e-01 -6.45660162e-01 -1.01793647e+00 -5.78974001e-02
-4.98420388e-01 -6.46176487e-02 7.01594234e-01 2.69680858e-01
-3.73579860e-02 2.30350256e-01 6.37333751e-01 -1.05730569e+00
-1.53392658e-01 -7.95219064e-01 -6.25757277e-01 -7.33992681e-02
4.09324765e-01 -7.38055170e-01 -5.48805654e-01 2.27948979e-01] | [8.632831573486328, -0.5333139300346375] |
5ff750e3-9fd4-4358-b661-6baf6ac5870b | deep-learning-on-lie-groups-for-skeleton | 1612.05877 | null | http://arxiv.org/abs/1612.05877v2 | http://arxiv.org/pdf/1612.05877v2.pdf | Deep Learning on Lie Groups for Skeleton-based Action Recognition | In recent years, skeleton-based action recognition has become a popular 3D
classification problem. State-of-the-art methods typically first represent each
motion sequence as a high-dimensional trajectory on a Lie group with an
additional dynamic time warping, and then shallowly learn favorable Lie group
features. In this paper we incorporate the Lie group structure into a deep
network architecture to learn more appropriate Lie group features for 3D action
recognition. Within the network structure, we design rotation mapping layers to
transform the input Lie group features into desirable ones, which are aligned
better in the temporal domain. To reduce the high feature dimensionality, the
architecture is equipped with rotation pooling layers for the elements on the
Lie group. Furthermore, we propose a logarithm mapping layer to map the
resulting manifold data into a tangent space that facilitates the application
of regular output layers for the final classification. Evaluations of the
proposed network for standard 3D human action recognition datasets clearly
demonstrate its superiority over existing shallow Lie group feature learning
methods as well as most conventional deep learning methods. | ['Luc van Gool', 'Zhiwu Huang', 'Chengde Wan', 'Thomas Probst'] | 2016-12-18 | deep-learning-on-lie-groups-for-skeleton-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Huang_Deep_Learning_on_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Huang_Deep_Learning_on_CVPR_2017_paper.pdf | cvpr-2017-7 | ['3d-classification', '3d-human-action-recognition'] | ['computer-vision', 'computer-vision'] | [ 5.15243001e-02 -1.21070035e-01 -4.94202793e-01 -4.87785071e-01
-3.12239498e-01 1.60822365e-02 7.42460847e-01 -5.39376378e-01
-4.26797688e-01 1.38787240e-01 5.75882971e-01 9.52085108e-03
-4.37345915e-02 -5.39096236e-01 -5.88672161e-01 -7.28167355e-01
-2.99024403e-01 1.94944575e-01 2.41005719e-01 -6.05387762e-02
3.36862504e-01 8.46059024e-01 -1.18420672e+00 2.37398863e-01
4.62191373e-01 1.13308203e+00 -3.29412818e-01 4.11498368e-01
7.27521349e-03 9.39006269e-01 -2.24295929e-01 1.65091798e-01
5.04337907e-01 -5.43060422e-01 -6.77112579e-01 3.83638322e-01
2.86748737e-01 -5.31362355e-01 -7.85791159e-01 6.28227770e-01
3.82401854e-01 3.83509994e-01 8.19632173e-01 -1.29932380e+00
-6.33350372e-01 1.92154109e-01 -6.82951391e-01 2.68238515e-01
4.90092099e-01 3.50902975e-01 8.78217816e-01 -9.92446601e-01
7.40111887e-01 1.45312762e+00 5.87044597e-01 5.88722944e-01
-8.84809196e-01 -4.47263718e-01 1.81588292e-01 7.52646506e-01
-1.13146138e+00 -3.39979023e-01 1.11153507e+00 -5.50075233e-01
1.25427496e+00 -1.33483410e-01 9.95111883e-01 1.28367198e+00
4.67092335e-01 1.25072443e+00 5.43860316e-01 -1.55773073e-01
1.14587195e-01 -7.36788332e-01 5.85269071e-02 7.58459687e-01
-2.81180441e-01 -8.60278010e-02 -6.85549378e-01 1.39009342e-01
1.01094043e+00 5.07771015e-01 -1.88938916e-01 -9.48165238e-01
-1.60079634e+00 7.26329267e-01 7.55793750e-01 3.33815634e-01
-5.24675786e-01 3.09445590e-01 4.88988817e-01 1.26168981e-01
4.98942584e-01 1.90386072e-01 -2.14600638e-02 -3.41129482e-01
-5.89032650e-01 1.47030711e-01 3.95078659e-01 6.20934367e-01
6.35851264e-01 5.18133715e-02 -2.97679722e-01 4.23505664e-01
4.76638317e-01 1.36972293e-02 7.92815924e-01 -8.58356714e-01
7.20555246e-01 1.09393477e+00 -8.78140107e-02 -1.11573470e+00
-6.22616410e-01 -1.38407290e-01 -9.63149428e-01 4.32531297e-01
3.89629036e-01 1.99675187e-01 -8.32388401e-01 1.49431956e+00
3.16283733e-01 3.38033497e-01 2.47066002e-02 1.24236190e+00
3.11116695e-01 6.18950725e-01 -7.51075242e-03 1.55032665e-01
8.90216410e-01 -1.05291677e+00 -5.02584994e-01 -3.84560563e-02
1.08176827e+00 -2.63044298e-01 1.04287350e+00 3.25345285e-02
-8.12844634e-01 -6.51264012e-01 -1.42274618e+00 -4.57325876e-01
-1.72518030e-01 3.59636635e-01 6.40935063e-01 1.75887465e-01
-7.91044772e-01 1.02951717e+00 -1.43743110e+00 -5.63518465e-01
6.48771822e-01 3.87649745e-01 -6.33377671e-01 6.24409169e-02
-1.04997027e+00 8.07142198e-01 3.69514912e-01 3.34160447e-01
-6.63433433e-01 -3.28028053e-01 -1.13754332e+00 -7.50434995e-02
1.58133134e-01 -6.46584570e-01 8.15586865e-01 -6.34295464e-01
-1.86947143e+00 8.80671859e-01 1.25310495e-01 -5.18731236e-01
7.22007275e-01 -3.59133303e-01 -2.07319170e-01 2.17053935e-01
-1.90582350e-01 6.08897626e-01 9.37194407e-01 -3.26279134e-01
-3.98161113e-01 -6.60699725e-01 1.91575084e-02 6.17847383e-01
-4.56382334e-01 -4.03925031e-02 -4.42020260e-02 -7.35673666e-01
6.74276292e-01 -9.71120179e-01 -3.46024364e-01 4.35097307e-01
-1.66184172e-01 -6.23889267e-01 1.02962089e+00 -4.76737559e-01
1.01848519e+00 -2.31184459e+00 6.42004490e-01 8.56730994e-03
1.50792733e-01 1.69459000e-01 -1.95863917e-01 2.52805531e-01
-3.53030741e-01 -2.07648784e-01 1.48459990e-02 -1.39231324e-01
1.14249833e-01 1.33517506e-02 -1.22193709e-01 1.02219105e+00
5.43648899e-01 1.05009985e+00 -6.61908150e-01 -1.65488660e-01
4.24854845e-01 4.07983303e-01 -5.51006138e-01 2.28061065e-01
-2.91844700e-02 5.51962197e-01 -9.65296209e-01 3.24685395e-01
2.88604081e-01 -1.26253322e-01 -3.34668666e-01 -2.25486293e-01
-1.01393804e-01 3.96470457e-01 -1.01925397e+00 2.20951128e+00
-1.41191319e-01 4.48004991e-01 -5.12418330e-01 -1.28526855e+00
1.38583314e+00 1.14990182e-01 1.03294635e+00 -3.10879618e-01
3.32478613e-01 4.92953695e-02 1.34647300e-03 -5.75703382e-01
2.30843455e-01 1.68141529e-01 2.50871498e-02 8.26293886e-01
-5.39911985e-02 1.13579296e-01 -2.42280975e-01 -3.86825919e-01
9.49140191e-01 5.87229788e-01 2.78968960e-01 -8.37728307e-02
9.25758898e-01 -1.51479930e-01 6.05690122e-01 7.84257799e-02
-4.80465293e-01 5.49567163e-01 6.56070590e-01 -1.06953096e+00
-1.09980845e+00 -7.36855567e-01 2.38405317e-01 8.15166473e-01
8.98509026e-02 -2.08177835e-01 -6.99103892e-01 -8.07689846e-01
-6.00374006e-02 1.95160374e-01 -6.39192939e-01 -5.95334232e-01
-9.86222267e-01 -3.62927020e-01 5.03273666e-01 7.12610364e-01
1.05638897e+00 -1.02427948e+00 -7.70929039e-01 1.72810331e-01
1.40096158e-01 -9.13841367e-01 -6.57059550e-01 -1.08428812e-02
-1.13025534e+00 -1.01505947e+00 -9.46940720e-01 -9.31114435e-01
5.67914963e-01 1.98267326e-01 2.23609924e-01 -2.89212614e-01
-1.52564675e-01 2.24488318e-01 -5.09287655e-01 3.34540047e-02
-6.16515242e-02 2.49507651e-01 4.25525486e-01 4.66986984e-01
9.39656675e-01 -7.12035477e-01 -6.39572978e-01 5.10011852e-01
-7.64469087e-01 2.30975628e-01 4.30155069e-01 8.12274933e-01
6.32694900e-01 -2.24155277e-01 2.52539426e-01 2.14816593e-02
3.10113519e-01 -2.40460172e-01 -1.39772162e-01 9.42028407e-03
-9.88561735e-02 2.90277779e-01 3.66160929e-01 -5.64171135e-01
-7.00379610e-01 4.33576882e-01 -2.28078719e-02 -6.60562932e-01
-1.53263524e-01 4.48381692e-01 -2.62112170e-01 -8.94507617e-02
8.04197013e-01 3.62495095e-01 4.33223993e-01 -6.53714299e-01
3.39935303e-01 5.82523227e-01 2.41972700e-01 -2.34336004e-01
7.62140453e-01 6.29775405e-01 2.56549239e-01 -7.42211998e-01
-8.09694290e-01 -2.83490032e-01 -1.35798860e+00 -2.90828526e-01
1.24712873e+00 -8.77583802e-01 -4.90823388e-01 8.61711085e-01
-1.03096998e+00 -3.87786746e-01 -3.44004691e-01 9.50754642e-01
-1.19297183e+00 5.56786418e-01 -6.34836078e-01 -3.55998307e-01
-1.43995643e-01 -1.08438313e+00 1.05166805e+00 6.51406199e-02
-1.80848479e-01 -6.92525208e-01 1.49790868e-01 2.70787597e-01
4.01655845e-02 5.29690921e-01 7.73140311e-01 -6.55333579e-01
-5.33741117e-01 -4.04839337e-01 -4.19027507e-02 4.21054810e-01
1.14535883e-01 -2.36970812e-01 -5.14048815e-01 -2.94086516e-01
2.89403588e-01 -3.66588265e-01 9.52796221e-01 3.62309456e-01
1.27008331e+00 -8.85219127e-02 -2.46142223e-01 8.68559480e-01
8.32102180e-01 2.12786093e-01 6.78669095e-01 5.27523220e-01
9.77685153e-01 5.75655043e-01 6.40367985e-01 5.28527141e-01
2.16688439e-01 7.40059257e-01 3.41298938e-01 -4.30237763e-02
4.99621853e-02 -4.36997682e-01 7.45124280e-01 8.32272291e-01
-3.55178475e-01 2.47303501e-01 -8.08669209e-01 1.15014538e-01
-2.12817812e+00 -9.98730779e-01 2.58976549e-01 1.87366259e+00
4.81584340e-01 1.55783683e-01 2.50800073e-01 2.30483279e-01
6.77518010e-01 6.78148448e-01 -8.46419632e-01 -1.63392842e-01
-3.70613970e-02 -1.78352326e-01 2.67405242e-01 8.25668126e-02
-1.46310198e+00 1.08252776e+00 5.73064041e+00 5.73168457e-01
-1.31312406e+00 -3.13031137e-01 3.66037369e-01 4.53475714e-02
2.25405335e-01 -2.70469069e-01 -6.39273286e-01 -5.98928854e-02
4.66964066e-01 -3.23027521e-01 1.61500752e-01 9.41118777e-01
5.13653100e-01 2.39313692e-01 -1.58724487e+00 1.24208653e+00
1.96390912e-01 -1.23663294e+00 2.57786721e-01 1.94664389e-01
5.95529437e-01 -1.07808322e-01 3.82365473e-02 1.85742468e-01
-2.15447590e-01 -1.15875912e+00 4.84962940e-01 7.92529762e-01
4.92884457e-01 -6.62998378e-01 4.31194723e-01 3.60333145e-01
-1.20135808e+00 -1.85646906e-01 -5.95807552e-01 -4.77513999e-01
2.50521421e-01 -5.70926443e-02 -6.10229552e-01 4.31836337e-01
5.55791795e-01 1.70867026e+00 -6.15569174e-01 1.00697660e+00
-1.40283212e-01 1.92633703e-01 -2.39380926e-01 -1.41523406e-01
6.48485363e-01 -3.69008452e-01 6.29449248e-01 7.62491763e-01
5.30946434e-01 2.50985801e-01 3.32199544e-01 7.38166928e-01
1.77029461e-01 1.50295153e-01 -9.37926590e-01 -1.74614817e-01
-3.14374804e-01 7.74517834e-01 -4.50596899e-01 -1.48360044e-01
-4.69486922e-01 1.11069965e+00 2.83447713e-01 3.88899207e-01
-5.35079420e-01 -3.82005006e-01 1.01661789e+00 -1.34960795e-02
3.79150808e-01 -7.36536562e-01 -8.53455737e-02 -1.33746099e+00
1.77313372e-01 -6.43979549e-01 4.73153532e-01 -7.64255404e-01
-1.04111683e+00 3.33774149e-01 -1.40075237e-01 -1.85296953e+00
-5.46458602e-01 -9.91558731e-01 -6.73838019e-01 7.25309908e-01
-1.10383821e+00 -9.86963511e-01 -1.58164769e-01 9.29229856e-01
5.14514148e-01 -4.33073193e-01 6.82465136e-01 3.48042026e-02
-5.55963755e-01 2.54529893e-01 -1.09996609e-01 3.77224714e-01
5.59886754e-01 -9.61570978e-01 6.32549942e-01 6.63529098e-01
1.24482907e-01 3.80963236e-01 7.63577297e-02 -4.46332097e-01
-1.43720353e+00 -1.10573673e+00 7.83756554e-01 -3.25935185e-01
7.94428647e-01 -3.26739520e-01 -9.21717227e-01 7.98691750e-01
-2.63879776e-01 1.11564085e-01 5.27693689e-01 -2.80595094e-01
-3.17916274e-01 -1.17003873e-01 -8.23710144e-01 8.38173568e-01
1.50068235e+00 -6.61138535e-01 -7.60942936e-01 4.59214658e-01
4.29881662e-01 -3.99519026e-01 -9.80919361e-01 4.31355476e-01
5.99930108e-01 -8.34246218e-01 7.92301059e-01 -1.23803210e+00
4.01613355e-01 -5.65290272e-01 3.93394660e-03 -1.23462105e+00
-4.99183685e-01 -6.62014425e-01 -1.57370254e-01 5.37098110e-01
-7.72906747e-03 -2.71371365e-01 1.15994513e+00 3.35859209e-01
-2.28442132e-01 -8.74778628e-01 -1.15952647e+00 -7.39784300e-01
1.12810872e-01 -2.16350377e-01 5.71256638e-01 7.62470424e-01
2.78534204e-01 2.06079304e-01 -4.21633810e-01 -1.39655545e-01
3.59217703e-01 1.01793431e-01 9.35316086e-01 -1.09431732e+00
6.05239794e-02 -5.61319172e-01 -1.29191172e+00 -1.54231846e+00
4.29040998e-01 -1.06509149e+00 -1.28809735e-01 -1.34462321e+00
-1.16614550e-01 -6.23713918e-02 -3.41611326e-01 5.78570306e-01
2.20169246e-01 5.67697361e-02 1.15841419e-01 2.59079784e-01
-5.03514409e-01 1.16315925e+00 1.62552428e+00 -1.97393626e-01
-3.42502832e-01 1.69852436e-01 -2.99634617e-02 8.45810354e-01
7.15450585e-01 -2.76193339e-02 -3.10216516e-01 -6.06379688e-01
-1.87982440e-01 -9.12944749e-02 4.08408195e-01 -1.27021825e+00
2.51081109e-01 -2.32033238e-01 5.99416494e-01 -6.50269449e-01
3.94809008e-01 -8.54814589e-01 -1.31205738e-01 5.07654428e-01
-5.42363882e-01 -5.72888702e-02 -2.89422750e-01 4.45806324e-01
-3.29473972e-01 2.30224833e-01 7.71055162e-01 2.70681977e-02
-8.92193258e-01 8.39104891e-01 -3.04124266e-01 -2.52928048e-01
1.23774922e+00 -5.92438281e-01 9.66179222e-02 -3.34642529e-01
-9.10456121e-01 1.52610868e-01 4.85801488e-01 7.42894113e-01
7.74319589e-01 -1.87031376e+00 -5.03023207e-01 4.67143327e-01
1.70764983e-01 -4.34406437e-02 2.23234057e-01 9.07510221e-01
-6.74868226e-01 3.41314793e-01 -8.93780410e-01 -8.27209055e-01
-8.90326262e-01 3.81123781e-01 6.13953292e-01 -1.92584693e-01
-1.15824735e+00 4.46695566e-01 8.89372155e-02 -3.23308945e-01
4.33148146e-01 -5.30718446e-01 -4.95581031e-01 -8.33853632e-02
4.82653409e-01 4.62102979e-01 -3.50921303e-01 -9.00133073e-01
-3.57884139e-01 9.30504262e-01 1.41428083e-01 2.44984832e-02
1.54954886e+00 -3.71325179e-03 3.59534360e-02 4.87145245e-01
1.50377810e+00 -9.77443397e-01 -1.69654799e+00 -3.54980797e-01
3.01341057e-01 -3.65300357e-01 -1.50484324e-01 1.19784139e-02
-9.95217085e-01 1.12333095e+00 5.35712123e-01 -1.41777530e-01
9.05564368e-01 -1.47595674e-01 7.15725422e-01 7.79281616e-01
2.79927522e-01 -1.01189685e+00 3.06120366e-01 7.73203313e-01
1.26743376e+00 -1.03566766e+00 3.80600840e-02 -1.59475029e-01
-5.77853203e-01 1.53503597e+00 5.27504802e-01 -7.33153105e-01
8.34009230e-01 -4.58580673e-01 1.55737609e-01 -2.34545574e-01
-4.57211047e-01 -1.14756692e-02 4.11619663e-01 5.82960367e-01
2.71553606e-01 -1.84293732e-01 -4.88500267e-01 4.05536681e-01
-2.03722328e-01 7.56833330e-02 3.31065625e-01 8.29338908e-01
-4.13775057e-01 -9.29039657e-01 -2.82282345e-02 2.39722833e-01
-1.01745613e-01 5.44448793e-01 -3.77040535e-01 7.78630435e-01
-2.13425949e-01 4.27334338e-01 2.70857781e-01 -5.45326531e-01
5.32329023e-01 2.70953923e-01 4.99210089e-01 -5.79497099e-01
-1.73695207e-01 -1.61550298e-01 -1.83485717e-01 -1.12604439e+00
-6.65698588e-01 -8.48798275e-01 -1.38456392e+00 -5.30679524e-02
3.69795620e-01 -2.75784880e-01 4.40063685e-01 1.08234644e+00
4.11696643e-01 1.98893413e-01 8.41012716e-01 -1.33573079e+00
-9.28440332e-01 -1.01725125e+00 -7.46128321e-01 4.58725303e-01
3.15711647e-01 -9.17766213e-01 -1.80355087e-01 1.15448229e-01] | [7.786080360412598, 0.3653239607810974] |
e631b7a1-d2aa-45f4-bed2-4a2ae9a36c09 | self-adapter-at-semeval-2021-task-10-entropy | null | null | https://aclanthology.org/2021.semeval-1.55 | https://aclanthology.org/2021.semeval-1.55.pdf | Self-Adapter at SemEval-2021 Task 10: Entropy-based Pseudo-Labeler for Source-free Domain Adaptation | Source-free domain adaptation is an emerging line of work in deep learning research since it is closely related to the real-world environment. We study the domain adaption in the sequence labeling problem where the model trained on the source domain data is given. We propose two methods: Self-Adapter and Selective Classifier Training. Self-Adapter is a training method that uses sentence-level pseudo-labels filtered by the self-entropy threshold to provide supervision to the whole model. Selective Classifier Training uses token-level pseudo-labels and supervises only the classification layer of the model. The proposed methods are evaluated on data provided by SemEval-2021 task 10 and Self-Adapter achieves 2nd rank performance. | ['Kyomin Jung', 'Yanghoon Kim', 'Sangwon Yoon'] | 2021-08-01 | null | null | null | semeval-2021 | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 5.91356635e-01 1.81956828e-01 -4.43261594e-01 -8.99042189e-01
-5.44045269e-01 -7.04633176e-01 7.27393985e-01 2.76755273e-01
-9.38592255e-01 1.11504602e+00 1.82402313e-01 -1.38662890e-01
2.66052246e-01 -6.37758911e-01 -6.43021584e-01 -4.99844730e-01
1.24132410e-01 8.15703452e-01 3.65321606e-01 -3.63224447e-01
2.74517089e-02 1.78270161e-01 -1.17949069e+00 8.80775988e-01
9.03892875e-01 9.87707734e-01 2.02036351e-01 7.69480109e-01
-3.98223191e-01 6.34367585e-01 -7.12601304e-01 -3.51689279e-01
3.97608504e-02 -7.01113462e-01 -1.26202846e+00 -1.21333353e-01
1.69600174e-01 1.61957607e-01 -6.41499907e-02 1.14267588e+00
5.01663625e-01 2.89020807e-01 8.66571307e-01 -8.68346393e-01
-6.62333369e-01 8.78985524e-01 -9.47296470e-02 6.40797377e-01
5.78776151e-02 -1.03259973e-01 7.51391411e-01 -8.75312567e-01
9.60129559e-01 1.11687422e+00 5.40941060e-01 9.86775637e-01
-1.30854261e+00 -6.16359234e-01 1.79156855e-01 7.05945551e-01
-1.09411061e+00 -3.77455026e-01 7.00204611e-01 -5.65137386e-01
1.32957733e+00 -1.39772341e-01 3.50745432e-02 1.55387199e+00
2.80910879e-01 7.32950270e-01 1.19348180e+00 -8.46113622e-01
6.91766977e-01 4.47988391e-01 4.19762492e-01 1.95975885e-01
-1.80982888e-01 2.17482135e-01 -6.88352644e-01 -8.25871080e-02
1.47551745e-01 -4.95824009e-01 8.85998458e-02 -2.90063649e-01
-1.01090848e+00 9.54916120e-01 3.89459580e-02 4.68601882e-01
-3.53607237e-01 -3.50776255e-01 1.05367446e+00 7.05569923e-01
9.32544112e-01 4.25528139e-01 -1.15259326e+00 -1.85735390e-01
-8.58660519e-01 1.96973741e-01 8.22288573e-01 1.08505082e+00
5.92560947e-01 5.94133846e-02 -4.45377469e-01 1.18288887e+00
-4.22274061e-02 4.58039269e-02 1.21233571e+00 -6.58967197e-01
4.90498096e-01 5.18612921e-01 2.75100698e-03 -3.38027366e-02
-4.13546205e-01 -6.80852532e-01 -7.29455292e-01 -3.54431123e-02
2.53324717e-01 -4.60448474e-01 -1.02583849e+00 1.82233739e+00
3.52945000e-01 3.37478518e-01 3.68531048e-01 4.94796425e-01
7.64421880e-01 6.75613642e-01 5.75315356e-01 -3.08823138e-01
1.12874687e+00 -1.10131168e+00 -6.89823508e-01 -5.36797881e-01
9.65396464e-01 -3.63403767e-01 1.09278965e+00 2.96129733e-01
-6.57421649e-01 -8.16739142e-01 -1.22873163e+00 1.03049830e-01
-6.92457378e-01 -5.19432761e-02 -1.90034866e-01 5.41366637e-01
-9.34085786e-01 6.46001935e-01 -4.94007349e-01 -6.75791502e-01
2.99802959e-01 2.22863957e-01 -3.08935851e-01 -4.06093113e-02
-1.86251748e+00 1.30403149e+00 1.14753735e+00 -6.44756794e-01
-8.07048798e-01 -5.46802163e-01 -7.89515257e-01 9.76017676e-03
1.41469672e-01 -4.12420362e-01 1.75301886e+00 -1.28433275e+00
-1.75361061e+00 1.23030531e+00 -6.15617111e-02 -9.01862323e-01
4.62923855e-01 -1.89888433e-01 -7.58214951e-01 -7.14693367e-02
5.99235334e-02 3.25395077e-01 7.51746237e-01 -9.59553301e-01
-8.94551873e-01 -3.15169275e-01 -3.87991279e-01 3.06722194e-01
-6.13865495e-01 -5.52399270e-02 2.70414382e-01 -6.53674662e-01
-2.60240555e-01 -8.41043770e-01 -3.71997505e-01 -6.88090742e-01
-2.55930930e-01 -5.85285902e-01 8.55733156e-01 -8.16683710e-01
1.20386600e+00 -1.97442627e+00 6.08103685e-02 -5.01845069e-02
-3.31113160e-01 7.21160054e-01 -3.47953528e-01 3.15758914e-01
-3.87108386e-01 -2.08638199e-02 -6.21289253e-01 -1.52932748e-01
-8.75899270e-02 2.03497440e-01 -5.05853556e-02 1.86503321e-01
2.75165617e-01 5.74042380e-01 -1.17751718e+00 -5.32693207e-01
-6.91964775e-02 6.10011443e-02 -3.81400675e-01 4.53098655e-01
-4.26281601e-01 4.46364015e-01 -2.93176174e-01 9.03889686e-02
5.91272533e-01 2.01861374e-02 3.86720240e-01 1.32016405e-01
1.54080153e-01 6.65130436e-01 -8.19874227e-01 1.97919023e+00
-5.47025442e-01 3.94269645e-01 -2.26710230e-01 -1.43076849e+00
1.23837566e+00 5.08202374e-01 2.97740489e-01 -7.00552762e-01
-4.18005735e-02 2.17847720e-01 -5.67914806e-02 -6.28477275e-01
3.66224557e-01 -4.51707691e-01 -3.59584689e-01 2.87562221e-01
6.79852426e-01 9.46729034e-02 2.62370348e-01 3.60815488e-02
1.31352973e+00 5.67586482e-01 8.18875968e-01 -4.04587388e-01
9.08630490e-01 2.35696465e-01 7.12099433e-01 6.67787969e-01
-4.75891322e-01 1.86533332e-01 3.03264689e-02 -5.27448893e-01
-1.31582820e+00 -9.05398190e-01 -2.89057553e-01 1.75506437e+00
-2.71289706e-01 -1.38894528e-01 -1.01752162e+00 -1.37868536e+00
-1.27745554e-01 1.30984735e+00 -6.69382572e-01 -6.48638189e-01
-6.47299826e-01 -5.44648409e-01 5.54082572e-01 5.13059914e-01
5.81997335e-01 -1.41687620e+00 -3.48913550e-01 7.19706655e-01
-4.37487036e-01 -1.15287519e+00 -4.56297755e-01 8.79019141e-01
-9.96727228e-01 -5.84725022e-01 -5.30552804e-01 -1.05577934e+00
2.82451153e-01 -5.25891066e-01 1.37448847e+00 -4.76490945e-01
2.85198450e-01 -9.47104916e-02 -7.60229766e-01 -4.50980276e-01
-1.10462725e+00 7.36394584e-01 1.41208544e-01 1.17363617e-01
1.14275289e+00 -5.25644660e-01 -9.20210332e-02 4.46279049e-02
-7.69904971e-01 -2.24693239e-01 5.48551261e-01 1.11722255e+00
5.93515575e-01 -1.66595653e-01 1.12458563e+00 -1.37754238e+00
7.53593683e-01 -7.67033696e-01 -3.63231897e-02 2.78146237e-01
-7.62395263e-01 3.64198416e-01 9.69871342e-01 -5.48390269e-01
-1.50828910e+00 5.52925825e-01 -4.70907509e-01 -6.10610247e-02
-6.71300828e-01 5.31591892e-01 -3.07490736e-01 3.85565579e-01
1.33890390e+00 4.63877916e-01 -3.41946632e-01 -8.12645376e-01
4.56653088e-01 1.11573815e+00 6.41571760e-01 -4.61261570e-01
4.88102406e-01 -9.75960568e-02 -5.42877197e-01 -6.94918215e-01
-1.13865745e+00 -5.31417072e-01 -1.29105365e+00 2.32940391e-02
6.39046609e-01 -5.84582925e-01 -2.91051492e-02 3.03992242e-01
-1.28533340e+00 -4.39504683e-01 -6.03438497e-01 2.18227550e-01
-6.55543566e-01 7.06445351e-02 -5.08767962e-01 -5.30825555e-01
-5.17104268e-01 -6.01137638e-01 7.11683273e-01 1.40949517e-01
-4.27142262e-01 -1.27033007e+00 5.01887083e-01 5.90157025e-02
3.42687607e-01 1.30157024e-01 1.06154013e+00 -1.60653722e+00
4.35769945e-01 -1.72966361e-01 2.77312636e-01 9.18887496e-01
9.36585143e-02 -5.95990717e-01 -1.35512924e+00 -2.04083055e-01
2.01603204e-01 -4.18266833e-01 9.78361130e-01 9.66207385e-02
1.12544346e+00 -3.67564559e-01 -4.82970774e-01 2.91904569e-01
1.36156631e+00 5.19090891e-01 2.74959177e-01 6.65418804e-01
3.11559677e-01 6.04060292e-01 7.02266932e-01 4.59101737e-01
2.10317612e-01 8.08064282e-01 -3.47186252e-02 2.39600733e-01
-2.57607132e-01 -4.09009129e-01 5.22338331e-01 8.18406343e-01
4.30447757e-01 -3.50036055e-01 -1.08467281e+00 6.70286953e-01
-1.82355809e+00 -1.00760055e+00 -4.97870669e-02 2.04937410e+00
1.40447712e+00 3.76904935e-01 1.62030086e-01 2.13968232e-01
9.18723285e-01 -1.91452965e-01 -9.09190118e-01 -8.91649067e-01
-1.20038494e-01 4.52263892e-01 4.68469530e-01 4.22040820e-01
-1.48246300e+00 1.21059060e+00 6.48753452e+00 1.04769564e+00
-1.02785695e+00 6.63499951e-01 6.31023049e-01 7.82456398e-02
7.73028359e-02 -2.67833799e-01 -1.04205024e+00 6.85369015e-01
1.67817152e+00 -4.23611134e-01 7.55224377e-02 1.27160692e+00
-3.61851081e-02 1.29389912e-01 -1.42185128e+00 4.74153101e-01
8.04519355e-02 -1.06265283e+00 -7.86048770e-02 -2.50087947e-01
7.59507895e-01 3.94926220e-01 -1.91545799e-01 7.48054981e-01
6.28639936e-01 -6.10104620e-01 6.63301051e-01 3.20876837e-01
9.66799140e-01 -6.98619187e-01 9.58386421e-01 6.74560606e-01
-8.16097856e-01 -2.34053612e-01 -3.61832678e-01 -7.47648105e-02
2.22014606e-01 5.76732516e-01 -1.37235188e+00 4.20484573e-01
7.26435006e-01 8.65515351e-01 -4.65926588e-01 8.55437160e-01
-2.14927614e-01 1.17886853e+00 3.05614714e-02 -5.69682345e-02
1.50248274e-01 1.23755313e-01 6.45511091e-01 1.93429458e+00
-1.04367482e-02 -1.09515093e-01 4.24350575e-02 4.23653424e-01
-2.16325119e-01 2.61629730e-01 -6.90463006e-01 1.36000097e-01
5.11356056e-01 9.35654640e-01 -3.99272442e-01 -7.13218093e-01
-2.68880665e-01 1.20375597e+00 4.86325145e-01 2.48210341e-01
-3.91024888e-01 -5.76439142e-01 6.08902752e-01 -3.36091220e-02
3.13020527e-01 3.62897757e-03 -7.33348489e-01 -1.00826955e+00
-2.42212355e-01 -8.07762623e-01 7.15697467e-01 -2.90083289e-01
-1.62847221e+00 9.47266281e-01 2.18838770e-02 -1.13187468e+00
-6.29947364e-01 -7.01618195e-01 -2.89591014e-01 8.93811226e-01
-1.72878456e+00 -8.41285527e-01 -2.76952665e-02 5.70574522e-01
1.02668548e+00 -6.01098001e-01 1.13856530e+00 3.06070387e-01
-2.28464648e-01 8.75892460e-01 7.32410908e-01 3.36409658e-01
9.04982865e-01 -1.28134239e+00 6.92777872e-01 7.32950389e-01
1.98728405e-02 1.54425964e-01 7.14168906e-01 -7.21601367e-01
-3.16885948e-01 -1.37164867e+00 1.51216948e+00 -5.88599682e-01
3.85787427e-01 -5.48476875e-01 -1.20767641e+00 6.26483381e-01
4.02589172e-01 -5.77882342e-02 7.64836967e-01 1.69047508e-02
-3.37816566e-01 -1.55759856e-01 -1.43207502e+00 5.04678376e-02
1.07761371e+00 -3.46481711e-01 -1.11326897e+00 4.79239464e-01
9.68958080e-01 -3.63966018e-01 -7.46573031e-01 3.06809992e-01
4.16004285e-02 -6.06125414e-01 7.03659534e-01 -1.03083563e+00
3.00399035e-01 7.03541338e-02 -9.83454436e-02 -1.95463979e+00
-5.18268049e-01 -1.65369987e-01 -5.76326400e-02 1.45255733e+00
6.11631989e-01 -4.63838547e-01 7.68639803e-01 2.67955422e-01
-2.56210476e-01 -3.29387248e-01 -1.23826122e+00 -1.22925162e+00
5.99951267e-01 -1.93845898e-01 6.00001752e-01 1.24349666e+00
2.07991675e-01 7.83076346e-01 -5.45923077e-02 -3.25404733e-01
4.95533437e-01 -3.82917821e-01 1.70214653e-01 -1.44294858e+00
-9.24010947e-02 -1.93676800e-01 -1.58777654e-01 -7.78363824e-01
6.70203388e-01 -1.28104508e+00 3.16612840e-01 -1.24307764e+00
-5.55437915e-02 -3.54137868e-01 -6.84003294e-01 6.41407788e-01
-2.75247376e-02 -7.50112459e-02 -1.39733985e-01 9.02431179e-03
-5.80560803e-01 6.48005784e-01 6.77117288e-01 -2.60250121e-01
-2.28896081e-01 1.39313176e-01 -3.99520695e-01 4.81292993e-01
9.69786346e-01 -8.46159816e-01 -3.80154550e-01 -1.82337299e-01
-1.67153597e-01 -2.42022723e-01 -2.70927101e-01 -9.66336966e-01
2.46525072e-02 -2.57107437e-01 3.20295841e-01 -2.47031212e-01
5.49144670e-02 -7.79454350e-01 -2.83249855e-01 4.76979405e-01
-1.06344962e+00 -2.57253200e-01 7.84125552e-02 5.40420532e-01
-2.83422291e-01 -6.54795408e-01 1.35527587e+00 -3.81499499e-01
-1.05054450e+00 7.32539669e-02 -5.69561839e-01 5.64334571e-01
9.32533503e-01 -1.08726583e-01 -6.14899062e-02 -6.28176481e-02
-9.63100433e-01 -1.05717123e-01 1.24859415e-01 7.15539277e-01
2.53127158e-01 -1.66059244e+00 -1.07720900e+00 1.03145942e-01
4.43162411e-01 -1.53303102e-01 2.59042811e-02 3.23960893e-02
-6.04754761e-02 4.52189893e-01 -2.99384087e-01 -4.31439638e-01
-1.00987411e+00 7.45274067e-01 4.69528079e-01 -4.84409690e-01
-2.83089280e-01 1.12752771e+00 -4.63917442e-02 -8.67823541e-01
1.25533357e-01 2.50408556e-02 -5.61577082e-01 1.40888412e-02
5.02450883e-01 2.81413168e-01 4.98829246e-01 -7.08769917e-01
-5.91286719e-01 -7.38510117e-02 -3.84140879e-01 -3.21876436e-01
1.37300682e+00 -1.42505676e-01 2.29734242e-01 6.13812447e-01
1.25573635e+00 -8.28316510e-01 -1.20146239e+00 -9.97365773e-01
7.89535522e-01 3.58229317e-02 -1.28229335e-01 -1.35066783e+00
-5.18495560e-01 1.00200188e+00 7.98393786e-01 1.93746481e-02
1.11911905e+00 -1.83499873e-01 6.99119389e-01 4.48616177e-01
1.41278893e-01 -1.82844090e+00 7.35034570e-02 1.05118895e+00
7.62182117e-01 -1.09958923e+00 -4.34448868e-01 9.67170894e-02
-9.60044801e-01 9.12389576e-01 1.05014396e+00 -1.25933513e-01
7.81075120e-01 1.19785964e-01 1.10812441e-01 4.24926341e-01
-9.33692455e-01 -1.42383404e-04 2.12429583e-01 9.93616045e-01
6.70930803e-01 -2.96737831e-02 -4.68351245e-01 9.39648509e-01
-6.48073256e-02 2.38169804e-01 2.83679307e-01 8.19743574e-01
-5.67475855e-01 -1.43084598e+00 -2.04445988e-01 4.45984066e-01
-3.69733900e-01 -2.26923376e-01 -5.47195852e-01 2.38525793e-01
3.45036566e-01 9.33682859e-01 -9.00166258e-02 -5.39142251e-01
3.88036638e-01 9.90253031e-01 2.53886152e-02 -1.02290058e+00
-7.40598261e-01 -3.20834994e-01 3.25616837e-01 -2.13175461e-01
-3.35973114e-01 -8.22983205e-01 -1.22159576e+00 2.24148408e-01
-6.58147708e-02 4.97286916e-01 7.39050031e-01 1.28496277e+00
5.02464771e-01 4.42849219e-01 6.65018976e-01 -4.98139173e-01
-8.96734297e-01 -1.46261573e+00 -5.15118599e-01 5.36286175e-01
3.61475080e-01 -6.30666018e-01 -7.38983750e-02 6.54591799e-01] | [10.827225685119629, 7.884324550628662] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.