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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ec818eb5-d895-4413-9fb7-61de0d2f4984 | automatic-diagnosis-of-the-short-duration-12 | 1904.01949 | null | https://arxiv.org/abs/1904.01949v2 | https://arxiv.org/pdf/1904.01949v2.pdf | Automatic diagnosis of the 12-lead ECG using a deep neural network | The role of automatic electrocardiogram (ECG) analysis in clinical practice is limited by the accuracy of existing models. Deep Neural Networks (DNNs) are models composed of stacked transformations that learn tasks by examples. This technology has recently achieved striking success in a variety of task and there are great expectations on how it might improve clinical practice. Here we present a DNN model trained in a dataset with more than 2 million labeled exams analyzed by the Telehealth Network of Minas Gerais and collected under the scope of the CODE (Clinical Outcomes in Digital Electrocardiology) study. The DNN outperform cardiology resident medical doctors in recognizing 6 types of abnormalities in 12-lead ECG recordings, with F1 scores above 80% and specificity over 99%. These results indicate ECG analysis based on DNNs, previously studied in a single-lead setup, generalizes well to 12-lead exams, taking the technology closer to the standard clinical practice. | ['Thomas B. Schön', 'Wagner Meira Jr.', 'Peter W. Macfarlane', 'Jéssica A. Canazart', 'Gabriela M. M. Paixão', 'Antônio H. Ribeiro', 'Milton P. S. Ferreira', 'Derick M. Oliveira', 'Carl R. Andersson', 'Paulo R. Gomes', 'Manoel Horta Ribeiro', 'Antonio Luiz P. Ribeiro'] | 2019-04-02 | null | null | null | null | ['ecg-classification', 'electrocardiography-ecg'] | ['medical', 'methodology'] | [ 4.41669196e-01 3.21573883e-01 1.33230194e-01 -5.96228004e-01
-7.69972861e-01 -5.84107399e-01 -3.33753526e-02 2.16311499e-01
-4.26984042e-01 6.69612944e-01 3.38665307e-01 -9.81142402e-01
-5.61008453e-01 -4.26801562e-01 -4.58195686e-01 -4.48692054e-01
-4.06371623e-01 9.12117362e-01 -6.52033985e-01 8.98789391e-02
-3.12464923e-01 5.70148051e-01 -6.35280967e-01 6.16054296e-01
4.86306399e-01 7.87131667e-01 -5.37047207e-01 1.16811299e+00
6.02965593e-01 1.04511857e+00 -9.04074848e-01 -2.79311985e-01
3.25822890e-01 -8.51031423e-01 -7.03909397e-01 -4.04243588e-01
5.54988027e-01 -4.85009015e-01 -3.16662014e-01 6.07756734e-01
1.36775553e+00 -2.84628928e-01 4.34925288e-01 -6.71602011e-01
-4.22664106e-01 7.76565015e-01 -6.34541214e-02 9.93040323e-01
-1.65316928e-02 4.12245363e-01 6.69419765e-01 -4.53244478e-01
5.80990016e-01 7.10654616e-01 1.53086472e+00 5.22370458e-01
-1.34100509e+00 -6.06228650e-01 -5.46045542e-01 -1.69871729e-02
-1.13382697e+00 -8.96674246e-02 2.47216791e-01 -4.63617325e-01
1.00481343e+00 4.33440119e-01 1.01241755e+00 1.20611429e+00
5.89677155e-01 2.80733258e-01 9.11753595e-01 -3.66530329e-01
-1.27948865e-01 -1.24303937e-01 2.65439749e-01 3.24651241e-01
5.28328001e-01 9.11051035e-02 -4.98537213e-01 -4.25507605e-01
1.09417677e+00 1.57746851e-01 -3.46922845e-01 -1.26303583e-01
-1.63233709e+00 5.80534458e-01 1.61111936e-01 6.19747043e-01
-6.81781650e-01 1.50666669e-01 8.36670101e-01 6.36977613e-01
1.94311351e-01 9.57627773e-01 -6.99347734e-01 -5.39908588e-01
-1.01991355e+00 1.95475832e-01 7.74947226e-01 4.41482127e-01
-1.25368267e-01 3.07511717e-01 -2.15453431e-01 5.47167957e-01
-1.56044677e-01 3.25965136e-01 6.87774956e-01 -1.07452893e+00
3.93946737e-01 2.72471488e-01 -1.97714180e-01 -7.58116364e-01
-8.72996926e-01 -1.06705391e+00 -1.15524435e+00 -3.45772668e-03
5.25338948e-01 -7.15844929e-01 -1.02510655e+00 1.28873754e+00
-2.58034915e-01 2.52739608e-01 -9.45113041e-03 6.76657081e-01
7.96526074e-01 -6.83249384e-02 2.66101867e-01 -1.05788335e-02
1.26859140e+00 -1.31284803e-01 -6.42268360e-01 6.45641759e-02
1.03951848e+00 -4.62013155e-01 7.59249091e-01 7.92934060e-01
-1.06631863e+00 -7.69093335e-01 -1.06545305e+00 2.77486414e-01
2.40945294e-02 4.60513830e-02 4.45032030e-01 1.10212862e+00
-1.21173441e+00 9.74400103e-01 -1.10464334e+00 -4.89989758e-01
8.31782043e-01 6.61233544e-01 -2.59847879e-01 1.46666467e-01
-1.21704209e+00 1.19313264e+00 3.44917148e-01 2.34183177e-01
-9.72392499e-01 -1.17809451e+00 -5.96796036e-01 1.46937132e-01
-1.61798269e-01 -1.08812916e+00 1.34783816e+00 -7.02000856e-01
-7.56176472e-01 1.01440382e+00 3.24744493e-01 -9.79540467e-01
7.39796579e-01 -3.51502031e-01 -7.22650170e-01 2.86367416e-01
-6.77523762e-02 4.35383379e-01 3.66576254e-01 -6.89407289e-01
-4.51908380e-01 -4.91548479e-01 -2.06156552e-01 2.59361472e-02
-2.66512871e-01 -4.71068174e-02 3.37216467e-01 -6.72440052e-01
-2.61302979e-04 -9.98379171e-01 -5.08599997e-01 -3.00844312e-01
-2.26097792e-01 1.90983370e-01 4.21676278e-01 -1.08167088e+00
1.37987411e+00 -2.11120343e+00 -2.61620402e-01 4.39243585e-01
8.66753638e-01 5.61151028e-01 2.06914246e-01 1.64404422e-01
-7.29738295e-01 3.34208459e-01 -2.54092336e-01 8.97741839e-02
-2.82358855e-01 2.23685727e-01 5.36109246e-02 4.17267054e-01
1.07050918e-01 1.29228544e+00 -8.31344843e-01 -2.12629750e-01
3.53631765e-01 4.13600385e-01 -3.12959701e-01 7.51720294e-02
6.46904111e-01 7.79209673e-01 -1.36323214e-01 4.57567930e-01
2.87446469e-01 -5.23398578e-01 6.49974704e-01 -3.66845615e-02
3.68548363e-01 2.17444688e-01 -7.18711317e-01 1.79145777e+00
3.55704017e-02 7.51727223e-01 -3.84919196e-01 -1.12257349e+00
8.07029426e-01 8.77894640e-01 8.53947878e-01 -3.39423597e-01
1.82363585e-01 2.38759696e-01 1.00753450e+00 -8.26520801e-01
-3.70958149e-01 -5.68547547e-01 2.21348852e-01 4.59598124e-01
1.41994968e-01 -7.77070671e-02 -1.88338250e-01 3.40098515e-02
1.58899117e+00 -6.04356527e-02 3.79960507e-01 -2.94024527e-01
-5.33016250e-02 6.45696372e-02 6.44992948e-01 1.26690710e+00
-5.28704226e-01 1.18610394e+00 4.94677275e-01 -1.19900298e+00
-1.13357580e+00 -1.12904668e+00 -4.91525143e-01 3.94541591e-01
-7.48884439e-01 -3.84121537e-01 -7.27650464e-01 -6.92587554e-01
-1.86780337e-02 3.53890717e-01 -8.40525627e-01 -3.76722068e-01
-8.39599788e-01 -1.09956217e+00 1.27723110e+00 1.12094283e+00
2.50508070e-01 -1.38047481e+00 -1.13834977e+00 5.94953239e-01
1.74929053e-02 -9.07071888e-01 2.20461693e-02 4.70415950e-01
-1.42856348e+00 -1.32884622e+00 -1.12487113e+00 -8.01659405e-01
2.16710716e-01 -6.64020419e-01 1.60911798e+00 -1.61152914e-01
-7.52452791e-01 4.36589450e-01 6.35134464e-04 -1.02968538e+00
-5.84316730e-01 3.34013790e-01 1.15443580e-01 -1.82247490e-01
6.32311046e-01 -8.80738437e-01 -9.39549685e-01 -2.85142064e-01
-5.58768868e-01 -2.46409193e-01 8.69165301e-01 9.42309260e-01
3.94212395e-01 -5.32112539e-01 1.00397515e+00 -1.41569078e+00
1.05443370e+00 -3.39352846e-01 2.01339215e-01 7.55745992e-02
-9.07825947e-01 -6.38571143e-01 3.32233906e-01 -1.40760884e-01
-6.12656355e-01 -8.34072903e-02 -1.51593015e-01 -5.28463185e-01
-5.45241416e-01 5.91700077e-01 4.53363478e-01 1.41534045e-01
1.17818081e+00 -4.15975526e-02 9.82836783e-02 -3.59232187e-01
-2.23765805e-01 4.89700645e-01 7.69295812e-01 -3.08862984e-01
4.29109037e-01 2.75893450e-01 3.80720556e-01 -5.68115413e-01
-6.09812200e-01 -2.30629817e-01 -8.12058032e-01 -1.78267762e-01
1.07511485e+00 -7.69719541e-01 -5.82060158e-01 1.06655158e-01
-8.25578392e-01 -3.21891576e-01 -8.77793729e-01 7.32165456e-01
-2.91617453e-01 1.76778078e-01 -8.20275843e-01 -5.46893179e-01
-7.46815681e-01 -7.90022492e-01 7.72482574e-01 -1.18799262e-01
-9.46054637e-01 -1.22442770e+00 1.62446991e-01 4.02458683e-02
5.20171046e-01 7.63526797e-01 1.20699692e+00 -9.92071509e-01
7.80549198e-02 -4.03832853e-01 4.07652967e-02 8.44895542e-01
2.54105091e-01 -4.28631186e-01 -9.48765576e-01 -2.08401650e-01
2.42444217e-01 -9.65442061e-02 4.91917521e-01 8.28798890e-01
1.45702708e+00 3.66956770e-01 3.25749815e-02 7.00940490e-01
9.27318931e-01 5.82746148e-01 8.34331751e-01 2.91349173e-01
7.69872904e-01 4.69893143e-02 -1.47298902e-01 1.45960271e-01
1.90913081e-01 -1.66284114e-01 1.76123567e-02 -7.18484640e-01
-4.34240326e-02 1.77727774e-01 -2.23980397e-01 8.31952274e-01
-5.08028328e-01 1.73317075e-01 -1.53738034e+00 4.98784959e-01
-1.60222781e+00 -7.36984968e-01 -3.97030473e-01 1.97717381e+00
5.97157538e-01 3.47675025e-01 1.66544765e-01 3.91874939e-01
5.03214300e-01 -2.08594665e-01 -7.04358757e-01 -5.93042552e-01
-1.73398241e-01 8.90351295e-01 2.63252020e-01 -1.08392164e-01
-9.66155827e-01 -6.04474545e-03 7.74947548e+00 -9.37975720e-02
-1.02257967e+00 1.46205768e-01 1.24208081e+00 -1.57431021e-01
1.57955676e-01 -4.65341806e-01 8.71901121e-03 1.31409466e-01
1.43737054e+00 6.89383000e-02 -9.21192169e-02 3.82842124e-01
4.13958549e-01 4.20936465e-01 -1.49317765e+00 1.27341068e+00
6.10369481e-02 -1.53717899e+00 -8.25247616e-02 1.30393550e-01
6.43653512e-01 1.68686524e-01 2.22876012e-01 2.70153344e-01
-3.81040648e-02 -1.47612536e+00 3.97394672e-02 6.27488852e-01
1.27580333e+00 -5.81857562e-01 1.21633804e+00 7.38305971e-02
-3.88226062e-01 -9.86682326e-02 5.65730408e-02 -1.46353662e-01
-6.22203909e-02 3.73877257e-01 -1.17743421e+00 3.78225863e-01
9.02663887e-01 1.03442490e+00 -5.58669388e-01 1.05580401e+00
9.59817618e-02 1.29466760e+00 -4.06080764e-03 5.77058017e-01
1.68747962e-01 1.61932990e-01 3.94893348e-01 1.34293020e+00
3.37657183e-01 2.47193769e-01 -1.03637516e-01 5.82986891e-01
-1.80972725e-01 5.34267873e-02 -7.55893767e-01 1.02733135e-01
-2.58638449e-02 1.16847658e+00 -6.07713044e-01 -7.20359385e-01
-1.95365846e-01 4.58228290e-01 -3.88958603e-01 4.06267196e-01
-7.51098812e-01 -4.57361490e-01 2.91437179e-01 4.25459713e-01
-2.17197552e-01 4.96509463e-01 -9.93000269e-01 -9.22373354e-01
6.19695708e-02 -1.48485196e+00 6.74389303e-01 -8.30500305e-01
-1.21787167e+00 7.03809142e-01 -2.30980560e-01 -1.46871161e+00
-3.92264307e-01 -6.98470533e-01 -4.32269990e-01 1.20470166e+00
-1.09398913e+00 -6.91714466e-01 -2.71112621e-01 3.28579098e-01
2.52182662e-01 -3.24508041e-01 1.48155642e+00 6.50385678e-01
-3.19499969e-01 6.21050000e-01 -1.68783680e-01 6.08743191e-01
7.88253725e-01 -1.59360468e+00 5.86691320e-01 4.39623326e-01
1.23121679e-01 9.40848708e-01 3.23417634e-01 -3.26242387e-01
-6.93298042e-01 -1.00807154e+00 1.00426912e+00 -9.35236871e-01
1.62267670e-01 1.88928708e-01 -8.85514677e-01 1.00947237e+00
3.19736004e-01 5.47220781e-02 1.14503324e+00 2.02028945e-01
1.08460620e-01 -2.39283308e-01 -1.22675252e+00 2.49553263e-01
9.32552934e-01 -5.27137220e-01 -8.49465668e-01 3.39423478e-01
-1.50577426e-02 -7.40315914e-01 -1.39785779e+00 8.06783259e-01
9.70891237e-01 -7.90925920e-01 8.95476401e-01 -1.31140304e+00
5.10092616e-01 1.48667678e-01 3.15441489e-01 -1.36998558e+00
-4.00904387e-01 -8.94492328e-01 -8.54074359e-02 3.79975647e-01
5.20861089e-01 -7.97651350e-01 8.16635668e-01 5.76057076e-01
-4.01855707e-01 -7.91579306e-01 -7.44048536e-01 -2.10320204e-01
1.34638846e-01 -4.66916203e-01 4.70457375e-01 1.28862119e+00
-1.81178659e-01 3.12006056e-01 -1.71225175e-01 9.21791047e-03
4.42678869e-01 -3.41470838e-01 4.00717348e-01 -1.72205377e+00
-4.90744710e-01 -2.72127092e-01 -8.24454010e-01 -2.80499279e-01
-4.30230349e-01 -9.98649716e-01 -4.58823919e-01 -1.60616016e+00
2.24417284e-01 -2.77740628e-01 -1.05128491e+00 5.29851019e-01
-1.90550864e-01 5.70179582e-01 7.73327798e-02 -1.92968417e-02
-1.73555732e-01 -5.35517871e-01 1.16043103e+00 -1.53333703e-02
-2.43706331e-01 5.71760386e-02 -1.06196511e+00 8.69561434e-01
8.82918358e-01 -6.83681726e-01 -3.66976500e-01 -5.28419912e-01
1.78585917e-01 3.30999225e-01 3.24569046e-01 -1.36351776e+00
-1.52299702e-01 6.34421885e-01 1.10846925e+00 -1.97994754e-01
-2.70790774e-02 -7.29609728e-01 4.03816432e-01 8.31928551e-01
-6.28567517e-01 8.18670392e-01 3.92809927e-01 1.70734704e-01
-2.17717022e-01 1.46677062e-01 5.47379375e-01 -5.14850974e-01
-1.31946221e-01 1.46298811e-01 -6.08130872e-01 3.84704024e-01
6.41745269e-01 -3.29599261e-01 1.06960721e-01 -4.68736023e-01
-1.62821615e+00 -9.90348384e-02 -3.42050940e-01 1.97863147e-01
5.10704458e-01 -1.10660017e+00 -1.08128130e+00 2.10808635e-01
-1.13567738e-02 6.99798285e-04 4.23506171e-01 1.17265129e+00
-1.07443070e+00 4.75328326e-01 -3.67861807e-01 -9.27672267e-01
-1.21604145e+00 -7.70176873e-02 9.92905080e-01 -4.52682942e-01
-1.09804809e+00 6.73402905e-01 -1.76178679e-01 -3.29193443e-01
2.46601850e-01 -6.72722220e-01 -2.38820389e-02 -7.50156790e-02
4.73995626e-01 4.33801770e-01 5.81939936e-01 -5.47041707e-02
-2.67822713e-01 9.64457467e-02 -9.77391452e-02 8.22001770e-02
1.57665575e+00 4.38346416e-01 1.17529683e-01 5.52619517e-01
8.14035594e-01 -6.46153986e-01 -5.63349426e-01 2.02086076e-01
-5.95923327e-02 7.17100725e-02 -7.50562847e-02 -1.26124072e+00
-1.03866315e+00 1.34293985e+00 1.25003195e+00 1.97090939e-01
1.03335905e+00 -5.69561660e-01 7.16946065e-01 5.36206543e-01
-6.65828958e-02 -8.53319943e-01 -7.04307556e-02 9.07634571e-02
3.73213500e-01 -8.79123449e-01 -8.25001150e-02 2.14491934e-01
-8.18055749e-01 1.26410306e+00 2.40996227e-01 -2.59157091e-01
7.69606709e-01 2.41005227e-01 7.36000657e-01 -5.37638724e-01
-3.80238652e-01 4.80355471e-01 5.47022223e-02 7.51835763e-01
7.36407757e-01 3.11547428e-01 -2.02812865e-01 6.49694920e-01
-9.33301896e-02 4.75852281e-01 6.42532647e-01 9.85518873e-01
1.85331374e-01 -9.91271794e-01 -1.48592383e-01 1.06304562e+00
-1.23116612e+00 -3.56857389e-01 -2.98386604e-01 9.28636968e-01
4.11344528e-01 6.60086751e-01 -6.80163056e-02 -2.16197759e-01
7.38964736e-01 6.67073369e-01 5.03017306e-01 -9.16485310e-01
-1.39735651e+00 8.33821595e-02 1.50287792e-01 -1.95605308e-01
-2.75433809e-01 -7.58895040e-01 -9.59217072e-01 1.04583114e-01
2.97399014e-01 4.77823578e-02 2.91332930e-01 6.50091887e-01
6.93119943e-01 1.31156683e+00 2.03301683e-02 -4.18115288e-01
-7.36781955e-01 -1.29300094e+00 -7.80515969e-01 4.77752239e-01
5.53282499e-01 5.16930856e-02 -1.16873503e-01 3.38490933e-01] | [14.36816692352295, 3.339888334274292] |
0b0f3a9d-5853-42e7-a8c6-cfc325945137 | uncertainty-sensitive-learning-and-planning-1 | 1912.09996 | null | https://arxiv.org/abs/1912.09996v3 | https://arxiv.org/pdf/1912.09996v3.pdf | Uncertainty-sensitive Learning and Planning with Ensembles | We propose a reinforcement learning framework for discrete environments in which an agent makes both strategic and tactical decisions. The former manifests itself through the use of value function, while the latter is powered by a tree search planner. These tools complement each other. The planning module performs a local \textit{what-if} analysis, which allows to avoid tactical pitfalls and boost backups of the value function. The value function, being global in nature, compensates for inherent locality of the planner. In order to further solidify this synergy, we introduce an exploration mechanism with two distinctive components: uncertainty modelling and risk measurement. To model the uncertainty we use value function ensembles, and to reflect risk we use propose several functionals that summarize the implied by the ensemble. We show that our method performs well on hard exploration environments: Deep-sea, toy Montezuma's Revenge, and Sokoban. In all the cases, we obtain speed-up in learning and boost in performance. | ['Łukasz Kuciński', 'Piotr Miłoś', 'Konrad Czechowski', 'Maciek Klimek', 'Piotr Kozakowski'] | 2019-12-19 | null | null | null | null | ['montezumas-revenge'] | ['playing-games'] | [-1.90717384e-01 5.30584872e-01 -9.82921794e-02 -2.49421801e-02
-5.64203262e-01 -7.26209998e-01 8.89309108e-01 2.16257587e-01
-7.36400008e-01 1.15774775e+00 3.12172174e-01 -4.13898855e-01
-7.07728863e-01 -9.77872372e-01 -5.73294103e-01 -8.66400063e-01
-4.75116968e-01 6.34438097e-01 9.20046344e-02 -3.76175910e-01
4.23258364e-01 3.87755454e-01 -1.33745337e+00 -2.46170029e-01
9.43915606e-01 1.09386683e+00 2.07944617e-01 3.71687502e-01
9.50772688e-02 1.02726090e+00 -5.85422456e-01 -1.23869084e-01
5.30148447e-01 -2.19184682e-01 -6.73741102e-01 -3.16986084e-01
-5.33013046e-01 -4.35880154e-01 5.21387085e-02 9.70543146e-01
3.67534250e-01 2.28156254e-01 8.74394178e-01 -1.07405722e+00
1.88384861e-01 1.03094089e+00 -3.21057916e-01 1.11549594e-01
2.62842417e-01 4.95421916e-01 1.06770515e+00 -2.15283364e-01
4.99887705e-01 1.19955683e+00 4.77848083e-01 7.38169700e-02
-1.11155450e+00 -3.20908040e-01 1.67959765e-01 8.16534553e-03
-1.00045848e+00 -2.17822373e-01 5.50625741e-01 -5.35712957e-01
7.55988717e-01 1.59156755e-01 8.48777294e-01 9.22845900e-01
5.54561436e-01 6.10501051e-01 1.42472589e+00 -4.73113716e-01
9.50225234e-01 -1.14023447e-01 -4.51205075e-01 5.00450194e-01
2.35977516e-01 8.84407818e-01 -5.40426552e-01 -2.01469898e-01
7.41663575e-01 -2.97162116e-01 -1.10372014e-01 -5.80790162e-01
-9.43002760e-01 8.95361960e-01 3.29329014e-01 5.76590188e-02
-5.38204610e-01 3.52040112e-01 1.86355591e-01 5.13849437e-01
5.85794076e-02 7.94206619e-01 -3.37008953e-01 -4.13877249e-01
-8.05156231e-01 3.79107624e-01 9.88947630e-01 3.30923408e-01
6.76791012e-01 -2.56636105e-02 5.36825731e-02 2.76623398e-01
4.81126517e-01 2.75623977e-01 2.98472136e-01 -1.16043973e+00
3.42878640e-01 3.98769349e-01 5.13686240e-01 -5.97202778e-01
-6.97974086e-01 -5.15449286e-01 -3.57650280e-01 9.81974781e-01
6.72097802e-01 -4.63814795e-01 -6.46594703e-01 1.83379781e+00
3.66570979e-01 -9.88479853e-02 3.10504079e-01 6.31686509e-01
-1.61402807e-01 3.19826365e-01 -2.66701356e-02 -2.82401174e-01
1.09857011e+00 -6.04852855e-01 -5.16775489e-01 -2.68741488e-01
5.40808976e-01 -1.87435195e-01 7.75135279e-01 6.02704525e-01
-1.07998073e+00 1.48001257e-02 -1.15479696e+00 5.24867773e-01
-2.50075519e-01 -4.36033577e-01 6.67512417e-01 7.12395787e-01
-8.48286211e-01 1.03103507e+00 -8.89602721e-01 -5.36675053e-03
1.48268387e-01 1.93874091e-01 -3.11520863e-02 5.81172287e-01
-1.39195919e+00 1.35555756e+00 7.05067456e-01 1.50492340e-01
-1.29734731e+00 -3.55598986e-01 -7.02997088e-01 3.18181694e-01
8.67367923e-01 -4.18277293e-01 1.31443024e+00 -7.35792518e-01
-1.92157495e+00 1.65380180e-01 6.87382877e-01 -7.50339329e-01
1.01108873e+00 -2.57136464e-01 7.16415793e-02 -3.39662395e-02
-3.99684679e-04 2.72614300e-01 6.21467233e-01 -1.22637713e+00
-5.63477099e-01 -2.78849870e-01 2.44346961e-01 4.29006040e-01
2.36647695e-01 -2.64521241e-01 3.88451576e-01 -3.70152056e-01
-3.59058976e-02 -8.31156790e-01 -6.04196429e-01 -4.33892310e-01
-2.19349489e-01 1.51487038e-01 3.22790854e-02 -2.59505183e-01
1.32099879e+00 -1.88960350e+00 3.19996059e-01 6.52479529e-01
-1.52022056e-02 -2.12516069e-01 1.74267963e-01 9.03614938e-01
2.40328103e-01 7.80269057e-02 -4.45182532e-01 -1.06836691e-01
3.60599041e-01 4.50637192e-01 -4.19067770e-01 3.59909743e-01
-1.18725404e-01 6.72718167e-01 -8.70692372e-01 -3.12088966e-01
1.87992752e-01 -1.82419330e-01 -5.30313551e-01 1.14990875e-01
-4.60357428e-01 4.96411502e-01 -7.58730054e-01 4.76871610e-01
4.09086496e-01 2.51832902e-01 5.67281902e-01 5.17704725e-01
-5.62007308e-01 3.26460510e-01 -1.53218985e+00 1.56133521e+00
-3.22998643e-01 -6.33168593e-03 3.99732620e-01 -7.44667947e-01
7.92088747e-01 -2.79879831e-02 4.73847032e-01 -7.33184218e-01
1.89114109e-01 4.21640694e-01 1.23354197e-01 -3.11075002e-01
4.79594439e-01 -4.32625234e-01 -3.11632067e-01 6.82911813e-01
-4.73340638e-02 -4.53828871e-01 -3.66326161e-02 -1.14789968e-02
1.13841927e+00 5.81653416e-01 7.51637101e-01 -7.56981790e-01
2.70895451e-01 -9.09063816e-02 5.55288553e-01 9.28850651e-01
-2.75532603e-01 -5.91054000e-02 1.17367995e+00 -5.14179230e-01
-7.04215527e-01 -1.14549446e+00 -1.16001092e-01 8.75568330e-01
3.13823879e-01 -2.66632169e-01 -7.31660306e-01 -7.81569242e-01
1.94289684e-01 1.00995493e+00 -8.75044525e-01 -1.62957817e-01
-4.42217082e-01 -5.99795699e-01 3.89876187e-01 3.69947135e-01
4.55547154e-01 -1.04175854e+00 -1.51010311e+00 1.91893548e-01
1.14052705e-01 -3.62163991e-01 -2.32546758e-02 7.60635734e-01
-8.59173238e-01 -9.80111957e-01 -1.52141243e-01 9.61979106e-02
6.54387549e-02 -4.08006072e-01 1.06269443e+00 -2.64036030e-01
1.32067919e-01 2.61671901e-01 -2.43664145e-01 -5.86830318e-01
-4.06481057e-01 -2.03537285e-01 2.53204733e-01 -2.59413987e-01
-1.48336738e-01 -8.57683957e-01 -4.69060838e-01 1.89619660e-01
-8.13240707e-01 -1.48378268e-01 6.68183148e-01 1.05628037e+00
2.80076534e-01 2.79166162e-01 3.92789871e-01 -4.42175418e-01
6.41241789e-01 -6.08024478e-01 -1.12797368e+00 2.10019842e-01
-7.02519834e-01 6.98197901e-01 2.60796487e-01 -1.19642429e-01
-1.28822768e+00 -8.55103135e-03 6.41933754e-02 1.55637443e-01
-6.55219238e-03 6.85317695e-01 -1.52499795e-01 1.89453959e-02
6.97167397e-01 -4.52418961e-02 8.56733024e-02 -4.14104223e-01
3.34758699e-01 3.56469333e-01 2.34625489e-01 -1.07598543e+00
7.53076136e-01 3.34082842e-01 1.70052603e-01 -5.61034203e-01
-3.85963559e-01 2.69814670e-01 -5.16072452e-01 -4.18100893e-01
7.12330461e-01 -5.50208569e-01 -9.65498388e-01 1.37510255e-01
-7.69058168e-01 -6.08668447e-01 -8.55838656e-01 5.07933319e-01
-1.06498849e+00 1.67358994e-01 -2.41305009e-01 -1.24185133e+00
1.01561092e-01 -1.32592320e+00 6.98176920e-01 3.78412038e-01
3.33686359e-02 -9.72167671e-01 4.39812720e-01 -1.45524219e-01
4.75333720e-01 5.58197320e-01 7.87612915e-01 -7.07427680e-01
-7.18388736e-01 2.59084642e-01 2.17118382e-01 4.17977124e-02
-1.89436525e-01 -3.29976499e-01 -9.27251041e-01 -1.54353559e-01
2.92066872e-01 -3.01868916e-01 7.79511094e-01 2.90574610e-01
6.76200986e-01 -4.03059334e-01 -1.19997319e-02 5.06503880e-01
1.40014362e+00 5.04602313e-01 6.29911244e-01 8.79415154e-01
-7.15163499e-02 8.61484647e-01 7.36815870e-01 9.03232753e-01
3.72500569e-01 7.17817426e-01 1.04775405e+00 4.87491280e-01
5.37150502e-01 -3.82988006e-01 5.07062197e-01 4.01783660e-02
-4.63372469e-01 8.36984441e-02 -1.04663539e+00 1.69106618e-01
-2.06444263e+00 -1.06775415e+00 5.19054234e-01 2.38244724e+00
7.47546196e-01 4.23496932e-01 2.03190804e-01 -1.10440277e-01
7.80160502e-02 2.97773451e-01 -4.45267439e-01 -4.89095718e-01
1.30013809e-01 -6.06267713e-03 5.88779926e-01 8.68102252e-01
-8.65278900e-01 7.16692030e-01 6.71896791e+00 8.38898838e-01
-7.91053355e-01 -9.58405510e-02 5.07399321e-01 -1.29294172e-01
-5.53510487e-01 2.60194033e-01 -2.98355401e-01 5.12727499e-01
8.69243205e-01 -8.57434422e-02 7.04930663e-01 9.34601009e-01
1.59053147e-01 -8.00199330e-01 -1.04267299e+00 3.47881526e-01
-5.01119673e-01 -1.27009153e+00 -4.78163809e-01 4.65207040e-01
3.43719363e-01 1.54732456e-02 -1.64751500e-01 3.86979789e-01
8.52144420e-01 -1.13202322e+00 1.16046047e+00 7.28366196e-01
2.92350948e-01 -9.63652968e-01 7.46993363e-01 6.82654142e-01
-8.36967587e-01 -4.95647877e-01 4.52408046e-02 -4.38914359e-01
2.06898943e-01 4.10743684e-01 -8.16686392e-01 6.83636010e-01
3.57839584e-01 -1.26232207e-01 -1.00293368e-01 8.30211461e-01
-3.16221803e-01 2.73865283e-01 -6.41724825e-01 -1.09647274e-01
5.78224838e-01 -6.26187921e-01 7.24567711e-01 8.08286250e-01
2.83881187e-01 4.46169488e-02 2.75823236e-01 1.08586264e+00
4.26493466e-01 -1.92582607e-01 -6.48830950e-01 5.62441871e-02
5.30894697e-01 9.33566511e-01 -6.35753751e-01 8.35639760e-02
1.88649014e-01 4.02863562e-01 3.00310850e-01 1.70162350e-01
-7.77251005e-01 -2.39021644e-01 6.80785537e-01 -2.48000011e-01
6.69152513e-02 -3.49355906e-01 -3.69810760e-01 -1.01507735e+00
-3.88728306e-02 -8.61047983e-01 2.39045531e-01 -2.57303327e-01
-7.09657133e-01 3.69058728e-01 3.95429045e-01 -9.61722076e-01
-7.43203163e-01 -6.82328105e-01 -6.93719745e-01 7.86191285e-01
-1.49221218e+00 -8.51269960e-01 9.08909068e-02 2.23591805e-01
2.58379668e-01 -2.59002537e-01 7.42900848e-01 -3.39719355e-01
-5.03739178e-01 3.06721386e-02 2.09262997e-01 -4.39562798e-01
1.92855924e-01 -1.61038589e+00 1.40047386e-01 6.51402950e-01
-1.17894053e-01 4.29709613e-01 9.36330736e-01 -7.13901162e-01
-1.28472376e+00 -2.51000345e-01 3.85365576e-01 -5.30510604e-01
9.55424547e-01 -1.08850777e-01 -4.53033715e-01 5.76136827e-01
1.63153172e-01 -4.73733395e-01 3.29276025e-01 2.35830754e-01
-3.78354430e-01 6.14116006e-02 -1.21522892e+00 5.25520861e-01
7.58794904e-01 -1.11614972e-01 -8.48501861e-01 7.63179502e-03
3.40052933e-01 -3.70688051e-01 -7.32302189e-01 3.10920537e-01
7.22589135e-01 -1.47832918e+00 7.34174013e-01 -5.69983006e-01
2.96889871e-01 -2.72488952e-01 -3.90608251e-01 -1.67521775e+00
-1.18377008e-01 -1.04563510e+00 -9.98119935e-02 8.38999391e-01
3.21879834e-01 -8.60789955e-01 5.82129180e-01 5.52991152e-01
-1.04074433e-01 -7.15691984e-01 -1.29364204e+00 -9.59559381e-01
3.54481220e-01 -4.00665671e-01 7.57944942e-01 6.05747223e-01
3.11990589e-01 -1.66291967e-01 -3.16020697e-01 4.77709733e-02
8.18763912e-01 1.49215795e-02 4.39464182e-01 -1.12821901e+00
-7.95368612e-01 -6.96578145e-01 2.79324073e-02 -5.70811689e-01
-4.54790182e-02 -3.85345846e-01 4.65583801e-02 -1.15232074e+00
-2.02949256e-01 -6.84187770e-01 -2.98907787e-01 4.23848659e-01
3.22033584e-01 -5.83427489e-01 4.20964807e-01 1.21767864e-01
-4.03948665e-01 6.47114038e-01 9.46771204e-01 1.90514863e-01
-4.46350545e-01 -1.51827065e-02 -5.61183989e-01 1.03053844e+00
8.51139605e-01 -3.18333358e-01 -4.04605836e-01 -1.61982402e-01
5.89527249e-01 4.89745200e-01 1.96697980e-01 -1.03802550e+00
1.66785970e-01 -7.05916405e-01 7.53420070e-02 -1.31296277e-01
2.43300036e-01 -7.82977462e-01 2.71397680e-01 7.83375859e-01
-3.59055519e-01 -9.00224596e-02 1.01633169e-01 6.69871807e-01
-3.40097323e-02 -5.61174870e-01 7.98342109e-01 -2.64324665e-01
-6.22989535e-01 -2.24823862e-01 -4.36336488e-01 -6.67052940e-02
1.06134999e+00 4.29526111e-03 -2.70986915e-01 -5.05297601e-01
-6.45682573e-01 6.39238656e-01 4.34770793e-01 -7.42167160e-02
2.41932571e-01 -1.01100612e+00 -4.91568506e-01 1.61512479e-01
-1.38591960e-01 -1.49675861e-01 2.36669496e-01 8.17461491e-01
-5.16767740e-01 1.17995955e-01 -5.74788749e-01 -5.25618792e-02
-4.01144296e-01 4.45327848e-01 7.88725913e-01 -6.75149798e-01
-3.91093969e-01 7.13460565e-01 1.80876940e-01 -3.84384871e-01
2.47445583e-01 -2.27934733e-01 -2.36367479e-01 4.12749499e-01
4.28530723e-01 4.90670860e-01 -1.95943847e-01 -7.23912567e-02
-3.01721960e-01 2.74922013e-01 3.40697348e-01 -7.05081820e-01
1.32915056e+00 -1.33535981e-01 8.67365673e-02 3.66057515e-01
2.48631597e-01 2.06511214e-01 -1.71476984e+00 8.63208473e-02
4.57177460e-01 -2.89861321e-01 1.04564168e-01 -1.29260135e+00
-4.04182315e-01 6.64734066e-01 3.25497210e-01 4.92859781e-01
1.06980336e+00 -3.45925659e-01 8.08423012e-02 5.81031501e-01
8.91182184e-01 -1.46797132e+00 -1.41956165e-01 6.34768903e-01
1.02613962e+00 -9.93179381e-01 1.80245154e-02 1.48526222e-01
-8.39619756e-01 1.10623181e+00 3.23223144e-01 -9.75767672e-02
3.62100452e-01 4.97700989e-01 -1.11697905e-01 -6.82568774e-02
-7.92656898e-01 -3.90084088e-01 -2.42348149e-01 5.30654907e-01
-2.08163694e-01 3.74517590e-01 -2.48563021e-01 7.11023331e-01
-2.84052461e-01 -2.29631886e-01 5.71242929e-01 9.93224800e-01
-8.56092095e-01 -1.20820117e+00 -5.12466609e-01 9.81589481e-02
-1.89044595e-01 5.63322790e-02 -2.42128313e-01 1.02551901e+00
1.59248784e-01 6.45311177e-01 -1.11446716e-01 -3.65889370e-01
2.14828193e-01 1.14610046e-01 3.75646800e-01 -1.72983438e-01
-6.94830298e-01 1.42933846e-01 3.95634919e-01 -8.60814154e-01
9.75004286e-02 -8.06164980e-01 -1.17146087e+00 -1.30805941e-02
-3.48404013e-02 6.06734097e-01 8.22791576e-01 1.01904583e+00
1.81964226e-02 4.75711703e-01 7.05072939e-01 -1.08724236e+00
-1.50461268e+00 -8.31022203e-01 -8.64666700e-01 -2.77007610e-01
4.30507034e-01 -1.04297435e+00 -5.27065635e-01 -7.26770163e-01] | [4.136070251464844, 2.2246451377868652] |
efee7040-fc46-4e2e-908e-80418d53aed3 | deformernet-a-deep-learning-approach-to-3d | 2107.08067 | null | https://arxiv.org/abs/2107.08067v1 | https://arxiv.org/pdf/2107.08067v1.pdf | DeformerNet: A Deep Learning Approach to 3D Deformable Object Manipulation | In this paper, we propose a novel approach to 3D deformable object manipulation leveraging a deep neural network called DeformerNet. Controlling the shape of a 3D object requires an effective state representation that can capture the full 3D geometry of the object. Current methods work around this problem by defining a set of feature points on the object or only deforming the object in 2D image space, which does not truly address the 3D shape control problem. Instead, we explicitly use 3D point clouds as the state representation and apply Convolutional Neural Network on point clouds to learn the 3D features. These features are then mapped to the robot end-effector's position using a fully-connected neural network. Once trained in an end-to-end fashion, DeformerNet directly maps the current point cloud of a deformable object, as well as a target point cloud shape, to the desired displacement in robot gripper position. In addition, we investigate the problem of predicting the manipulation point location given the initial and goal shape of the object. | ['Tucker Hermans', 'Alan Kuntz', 'Bao Thach'] | 2021-07-16 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [-1.08699828e-01 1.33167610e-01 -1.93454772e-01 -2.10495576e-01
-2.35162482e-01 -8.79611969e-01 3.32617819e-01 -1.97088137e-01
-2.53879666e-01 1.21081583e-01 -3.58815998e-01 2.65620160e-03
-2.93137968e-01 -6.69498920e-01 -1.05575716e+00 -4.57672685e-01
6.10790886e-02 1.21975076e+00 2.56003112e-01 -3.73611808e-01
4.96701300e-01 1.23198009e+00 -1.03107560e+00 -7.28105083e-02
3.86004001e-01 1.06571913e+00 4.17974889e-01 8.43891263e-01
2.20901579e-01 5.29300459e-02 -2.76119292e-01 3.62966567e-01
6.13972366e-01 2.55469620e-01 -6.63062274e-01 1.83809921e-01
2.92077392e-01 -6.48375332e-01 -5.06058216e-01 8.74272346e-01
2.20358968e-01 3.82707626e-01 7.69506514e-01 -1.07544374e+00
-8.20346892e-01 6.05160110e-02 2.23930199e-02 -3.28067660e-01
1.87899679e-01 4.11945701e-01 4.42624927e-01 -7.28540838e-01
8.99768054e-01 1.42950022e+00 5.12914777e-01 6.68227077e-01
-8.71684253e-01 -2.01181769e-01 2.15588018e-01 -4.13428783e-01
-9.38706875e-01 1.43875135e-02 9.50210094e-01 -7.78257966e-01
9.84432876e-01 -1.52480140e-01 7.84018695e-01 6.96344256e-01
7.59424627e-01 4.17250693e-01 5.25084019e-01 -1.78488061e-01
8.87960345e-02 -4.75855917e-01 -1.06311075e-01 7.85205185e-01
-8.38060006e-02 3.12809139e-01 1.45057797e-01 -5.97399957e-02
1.53014529e+00 4.47896540e-01 1.83760989e-02 -1.07713962e+00
-1.28523815e+00 5.91377556e-01 9.38292205e-01 -4.15736698e-02
-5.23410141e-01 4.89423990e-01 1.60916746e-02 1.47621274e-01
1.82672754e-01 7.56289482e-01 -7.43877172e-01 -6.42198697e-02
-3.42905149e-02 7.75107980e-01 8.67925465e-01 1.09392285e+00
5.80761671e-01 1.50335627e-02 -1.52672932e-01 9.98822600e-02
5.57007611e-01 7.23681033e-01 9.30094197e-02 -1.21489704e+00
4.56057429e-01 9.45119321e-01 5.24152219e-01 -8.72124314e-01
-4.79143530e-01 1.90990530e-02 -3.11847389e-01 8.91452014e-01
2.92901516e-01 -2.80627936e-01 -1.47375751e+00 1.45295358e+00
5.56310356e-01 -2.58716106e-01 -1.12655260e-01 1.30546165e+00
2.38888428e-01 6.32302940e-01 -4.29829866e-01 4.40150201e-01
6.78442657e-01 -6.64219677e-01 -3.91206741e-01 -1.64677858e-01
2.42283151e-01 -2.31896028e-01 6.19155288e-01 1.73489414e-02
-1.33122134e+00 -5.94128251e-01 -1.01645827e+00 -3.19302440e-01
-4.89456415e-01 7.85983801e-02 4.02735114e-01 -4.35802668e-01
-8.21045935e-01 9.94849026e-01 -1.44368601e+00 -1.12434678e-01
3.21461737e-01 8.64298403e-01 -4.13225740e-01 2.01909944e-01
-5.80888510e-01 1.58364356e+00 4.48685586e-01 4.57849830e-01
-1.11139286e+00 -4.06232506e-01 -8.79643261e-01 -6.06466606e-02
3.61333013e-01 -6.83525503e-01 1.28863323e+00 -4.15112376e-01
-1.86872232e+00 8.26641440e-01 2.95066416e-01 -5.75730763e-02
4.31422442e-01 -4.92181271e-01 3.52811158e-01 -4.45085578e-02
-1.05580881e-01 6.04561567e-01 9.28098083e-01 -1.41128027e+00
-1.40890464e-01 -8.35366309e-01 2.88660437e-01 3.40541810e-01
3.84369761e-01 -4.01157796e-01 -2.69426465e-01 -3.26903999e-01
6.40207231e-01 -1.36961079e+00 -3.81996214e-01 5.99892080e-01
-4.44850504e-01 -4.14395362e-01 1.28993893e+00 -2.18962938e-01
1.34120911e-01 -2.11912012e+00 8.42618704e-01 2.08190665e-01
1.28544837e-01 5.22992611e-01 -2.82089680e-01 3.92533273e-01
1.25303164e-01 -1.56347439e-01 2.31889077e-02 2.72019744e-01
1.28987804e-01 3.85604829e-01 -3.47569793e-01 4.84566152e-01
6.50042057e-01 1.44645834e+00 -1.07553864e+00 2.46041745e-01
3.02527666e-01 5.15235186e-01 -7.28451014e-01 5.38290679e-01
-8.14700723e-01 6.24087334e-01 -1.19570553e+00 4.36917335e-01
4.73232269e-01 -5.55917853e-03 -1.01159774e-01 -2.08989933e-01
-3.84866893e-01 4.63230908e-02 -8.60567987e-01 1.99525714e+00
-2.86961824e-01 1.03726551e-01 2.91934073e-01 -9.11511183e-01
1.36787975e+00 2.66661167e-01 5.68844795e-01 -4.39720303e-02
4.35629249e-01 1.97234362e-01 -6.55062636e-03 -6.07636273e-01
2.56793320e-01 -9.42013934e-02 -2.48851240e-01 1.82702079e-01
-1.28509616e-02 -9.29137468e-01 -6.07436121e-01 -5.10159910e-01
9.12753940e-01 7.37437129e-01 -1.02120735e-01 3.08804959e-02
-1.36042116e-02 3.19192499e-01 4.81366329e-02 4.24001455e-01
4.11847122e-02 5.91063797e-01 2.82634467e-01 -6.22123599e-01
-1.38497758e+00 -1.34908307e+00 1.28892750e-01 4.53785956e-01
4.32875484e-01 3.40175569e-01 -4.47781771e-01 -4.91171420e-01
7.91802585e-01 2.66190559e-01 -6.10420525e-01 -4.19896185e-01
-9.88996327e-01 3.05658221e-01 -1.91190735e-01 6.07514024e-01
2.08481088e-01 -1.14143014e+00 -1.13924003e+00 3.58565718e-01
3.97866875e-01 -9.15893257e-01 -3.95018280e-01 3.76312077e-01
-1.05273736e+00 -1.21343029e+00 -3.32543582e-01 -1.10530972e+00
9.83547270e-01 -3.58570814e-02 3.86875540e-01 -8.45436081e-02
-2.81903505e-01 4.73402590e-01 -3.45542252e-01 -5.27328074e-01
-4.90866035e-01 2.53063366e-02 1.80739284e-01 -4.50487196e-01
-9.90860537e-02 -3.14061791e-01 -4.80242610e-01 2.15513766e-01
-6.15831196e-01 -2.44802132e-01 5.68873286e-01 5.79996467e-01
7.44344175e-01 -1.92596883e-01 3.33165079e-01 -1.51041627e-01
3.66506100e-01 -5.27252145e-02 -8.31116855e-01 -7.21240044e-02
-2.88998224e-02 2.66960889e-01 4.57005352e-01 -8.77131045e-01
-5.27378619e-01 7.94376314e-01 -6.49187248e-03 -1.10333228e+00
-6.58027306e-02 5.31980217e-01 5.15564792e-02 -7.88271725e-02
4.71002907e-01 -8.34674388e-02 5.86353004e-01 -6.24164641e-01
3.66999418e-01 4.96352434e-01 5.64805508e-01 -6.46589458e-01
9.60428476e-01 3.64417881e-01 4.13881838e-01 -1.78244352e-01
-6.86782956e-01 -1.24354519e-01 -1.44737303e+00 -1.46936357e-01
9.61806893e-01 -4.45458949e-01 -1.25337732e+00 3.97889227e-01
-1.38034904e+00 -6.79585338e-01 -4.55763549e-01 4.52308446e-01
-1.11832440e+00 -2.50807464e-01 -3.17493260e-01 -6.17304504e-01
-3.64616185e-01 -1.23073530e+00 1.56520832e+00 -1.22033410e-01
-1.28773108e-01 -6.05503857e-01 -9.79465470e-02 -1.84756473e-01
3.08038324e-01 9.23822641e-01 1.08072972e+00 -3.66338015e-01
-7.18379021e-01 -9.42587376e-01 1.56313881e-01 1.65021867e-01
2.98088521e-01 -3.40803787e-02 -3.15975189e-01 -4.63772327e-01
1.03379861e-01 -3.22627485e-01 2.71740615e-01 6.17825806e-01
1.06244171e+00 -3.13940912e-01 -6.39120996e-01 4.13748682e-01
1.30140805e+00 4.12678450e-01 1.92791849e-01 1.11222848e-01
7.70295560e-01 3.33073825e-01 7.78054297e-01 2.49081805e-01
1.85413465e-01 9.09142137e-01 1.14783144e+00 2.28743523e-01
1.42458573e-01 -3.83232743e-01 1.92486927e-01 3.61557364e-01
-2.53120720e-01 -2.08647233e-02 -8.49604964e-01 3.60147238e-01
-1.90190613e+00 -5.33874273e-01 1.47829637e-01 2.18800426e+00
6.95016563e-01 2.00528458e-01 -1.71678543e-01 -2.70329177e-01
5.94138503e-01 -1.82925716e-01 -1.14608860e+00 -4.12083745e-01
6.62710965e-01 5.61601445e-02 3.80695999e-01 5.44405937e-01
-9.83106673e-01 1.11955976e+00 5.80374050e+00 -1.70621872e-01
-1.62603593e+00 -4.49258775e-01 -2.56152540e-01 2.51985937e-02
2.29417771e-01 -9.08913910e-02 -8.74079585e-01 1.20926119e-01
3.41119915e-01 -4.10582498e-02 5.89172125e-01 8.64654243e-01
1.38564885e-01 4.36003357e-01 -1.60699892e+00 5.15705466e-01
-1.14191145e-01 -1.34854102e+00 -2.52825003e-02 1.20134838e-01
5.12830973e-01 1.00492857e-01 1.53045043e-01 2.04403892e-01
3.60374838e-01 -8.77480388e-01 1.06144238e+00 8.51181865e-01
7.53696203e-01 -3.70141447e-01 2.26062670e-01 7.93844640e-01
-8.98923993e-01 -3.53742480e-01 -4.21600699e-01 -2.68347800e-01
3.22049916e-01 -1.13382816e-01 -1.10200691e+00 1.55667275e-01
4.60145205e-01 5.90909302e-01 1.61965445e-01 8.19847822e-01
-3.29046659e-02 -1.14287607e-01 -3.81769121e-01 -1.80974364e-01
4.06599164e-01 3.87991108e-02 9.31241870e-01 3.89843553e-01
2.71710724e-01 3.78875226e-01 5.64325750e-01 1.21722484e+00
3.34982201e-02 -6.90015554e-01 -9.52801704e-01 -1.70456007e-01
2.42906198e-01 1.03222561e+00 -5.34584343e-01 -1.44652408e-02
3.02393317e-01 7.87661374e-01 4.76639450e-01 2.13711292e-01
-4.95421350e-01 -4.19638157e-01 6.77226424e-01 2.70079523e-01
6.05169058e-01 -1.02621174e+00 7.64241219e-02 -8.39017808e-01
1.71630025e-01 -3.79580408e-01 -4.90314484e-01 -1.23324978e+00
-9.59557652e-01 1.65946037e-01 -4.79514189e-02 -1.22385609e+00
-2.65454710e-01 -1.14921963e+00 -3.99223626e-01 9.25675511e-01
-1.19397295e+00 -1.14323199e+00 -2.29782864e-01 3.19076955e-01
4.50978994e-01 2.54942507e-01 9.27007973e-01 -5.32336593e-01
8.68771672e-02 -9.33445469e-02 -1.95777509e-02 3.85624439e-01
2.90835291e-01 -1.12651992e+00 6.03639007e-01 1.08698010e-01
-4.26306695e-01 6.45909667e-01 4.05746222e-01 -8.37768853e-01
-2.25103593e+00 -1.22760069e+00 3.10058057e-01 -1.09633851e+00
5.17556965e-01 -4.64673817e-01 -8.90888155e-01 1.16020453e+00
-4.85636741e-01 4.61053669e-01 -5.23174047e-01 -3.84676218e-01
1.94964260e-02 2.20953345e-01 -1.29627454e+00 5.73517084e-01
1.13716280e+00 -2.33251482e-01 -8.09267581e-01 5.80255806e-01
8.53637397e-01 -1.42573380e+00 -1.13895750e+00 6.23843312e-01
7.68615603e-01 4.10680845e-03 1.07737505e+00 -1.23204255e+00
6.06201947e-01 -3.08291197e-01 -4.46136370e-02 -1.51807535e+00
-5.31405032e-01 -5.32100439e-01 -2.60995626e-01 2.70424992e-01
1.32660270e-01 -4.41386759e-01 1.03343701e+00 6.50600314e-01
-5.89484572e-01 -1.10887992e+00 -1.10374033e+00 -7.63657212e-01
4.27329838e-01 1.31300166e-01 6.24245346e-01 6.66094780e-01
-1.80992950e-02 -4.72364165e-02 1.49576738e-01 7.64881611e-01
2.06859067e-01 6.15080297e-01 7.81270027e-01 -1.36873186e+00
1.60876647e-01 -2.50653207e-01 -8.03970218e-01 -1.42630839e+00
5.19280195e-01 -1.20628428e+00 5.94656408e-01 -1.72976530e+00
-2.23300934e-01 -6.75231814e-01 1.84305474e-01 7.15488791e-01
1.35248557e-01 -2.03303322e-01 3.24067682e-01 1.33487508e-01
-1.63789243e-01 4.94867831e-01 2.05657077e+00 -2.29695156e-01
-3.95311207e-01 1.47796497e-01 -8.06637779e-02 4.52391714e-01
6.83310091e-01 -2.08760679e-01 -3.59544232e-02 -9.51453209e-01
-1.96879879e-01 5.28405964e-01 7.08806455e-01 -6.49002194e-01
2.84584403e-01 -5.31040251e-01 5.75470984e-01 -6.55910850e-01
7.33918905e-01 -1.31584334e+00 3.54122184e-02 8.40080023e-01
-5.22462666e-01 6.98976591e-02 1.59735695e-01 4.53801364e-01
2.37783894e-01 -1.47359192e-01 6.67300820e-01 -2.42635354e-01
-2.51577884e-01 8.13745618e-01 -5.50121590e-02 -3.74724299e-01
1.18966937e+00 -2.73113638e-01 -1.35937659e-02 8.82881433e-02
-1.02361465e+00 3.68605196e-01 7.75157690e-01 8.09489131e-01
8.77174854e-01 -1.30521953e+00 -4.55880135e-01 4.92137253e-01
-1.45119518e-01 8.75712097e-01 -1.24890357e-01 2.23331824e-01
-4.67374802e-01 3.97821218e-01 -3.98595005e-01 -9.28598940e-01
-7.55126595e-01 6.83014512e-01 7.58236587e-01 2.65986323e-01
-7.99311459e-01 5.90131104e-01 -1.85046032e-01 -9.42323208e-01
-5.43016195e-03 -8.86399090e-01 8.04083943e-02 -7.14603126e-01
-2.49272287e-01 -6.12179935e-02 -5.26659936e-02 -6.12848103e-01
-1.40327901e-01 1.03173614e+00 1.66099578e-01 2.01427564e-01
1.68411601e+00 4.03081715e-01 -1.58786684e-01 3.61644179e-01
1.34406948e+00 -5.24350882e-01 -1.89493060e+00 -4.25691865e-02
-1.46730274e-01 -4.66225177e-01 -1.18868127e-01 -7.78352916e-01
-9.72673893e-01 7.37968147e-01 4.47346747e-01 6.79531172e-02
3.02907914e-01 2.72433400e-01 6.91162407e-01 8.19304585e-01
4.49575752e-01 -7.07568765e-01 3.54148746e-01 1.02941346e+00
1.49140441e+00 -1.08515465e+00 -9.54864100e-02 -2.37864748e-01
-1.86914802e-01 1.39882851e+00 5.99058032e-01 -8.33991587e-01
8.81412804e-01 3.21183532e-01 6.44796267e-02 -5.10295868e-01
-5.26033282e-01 2.49696940e-01 4.02929932e-01 7.52863526e-01
-1.24169230e-01 5.21801673e-02 5.29537678e-01 1.41875371e-01
-1.20541096e-01 2.37841323e-01 1.97739750e-01 1.52222657e+00
-8.41352046e-01 -7.48317778e-01 -1.98050916e-01 2.34713897e-01
4.49683256e-02 6.88890636e-01 -4.67662245e-01 7.78860152e-01
-3.87349203e-02 3.09389949e-01 4.12138879e-01 -2.85841912e-01
8.91455293e-01 5.71078621e-02 9.68383789e-01 -1.08171713e+00
-4.07689303e-01 -2.52699584e-01 -5.60480535e-01 -5.74111104e-01
-7.53135085e-02 -5.47632992e-01 -1.80396295e+00 1.69233069e-01
-3.93045783e-01 -2.51689613e-01 1.10878217e+00 8.90925229e-01
5.25325119e-01 3.04646254e-01 8.45952332e-01 -1.83166444e+00
-1.28206229e+00 -6.50235891e-01 -3.09615225e-01 2.66827881e-01
7.77183056e-01 -9.82360244e-01 -1.11330390e-01 -5.87386042e-02] | [5.011669158935547, 0.1887478083372116] |
2d0d0713-7bfe-4b9e-818d-a669a3b4f13d | usr-an-unsupervised-and-reference-free | 2005.00456 | null | https://arxiv.org/abs/2005.00456v1 | https://arxiv.org/pdf/2005.00456v1.pdf | USR: An Unsupervised and Reference Free Evaluation Metric for Dialog Generation | The lack of meaningful automatic evaluation metrics for dialog has impeded open-domain dialog research. Standard language generation metrics have been shown to be ineffective for evaluating dialog models. To this end, this paper presents USR, an UnSupervised and Reference-free evaluation metric for dialog. USR is a reference-free metric that trains unsupervised models to measure several desirable qualities of dialog. USR is shown to strongly correlate with human judgment on both Topical-Chat (turn-level: 0.42, system-level: 1.0) and PersonaChat (turn-level: 0.48 and system-level: 1.0). USR additionally produces interpretable measures for several desirable properties of dialog. | ['Shikib Mehri', 'Maxine Eskenazi'] | 2020-05-01 | usr-an-unsupervised-and-reference-free-1 | https://aclanthology.org/2020.acl-main.64 | https://aclanthology.org/2020.acl-main.64.pdf | acl-2020-6 | ['dialogue-evaluation', 'open-domain-dialog'] | ['natural-language-processing', 'natural-language-processing'] | [-3.42289448e-01 6.67451143e-01 2.06977166e-02 -8.57708216e-01
-8.69059622e-01 -1.11216259e+00 1.22352469e+00 3.58244777e-01
-2.57241547e-01 1.09476221e+00 7.72556841e-01 -4.52894688e-01
-7.18674436e-02 -4.06470418e-01 4.50171679e-01 -7.64272064e-02
1.00955687e-01 9.09608126e-01 8.55448768e-02 -9.45032656e-01
3.32804888e-01 2.89218217e-01 -7.95344174e-01 2.14236394e-01
7.30458796e-01 4.83289093e-01 -1.18307866e-01 1.20507574e+00
-2.15756997e-01 1.23527777e+00 -1.20553553e+00 -4.25417542e-01
-1.77983552e-01 -8.23754668e-01 -1.49198961e+00 5.33605134e-03
2.38377169e-01 -5.05382776e-01 -1.40186638e-01 8.65728080e-01
2.93331891e-01 3.20984006e-01 9.43378985e-01 -1.34830379e+00
-5.94060898e-01 9.05209243e-01 2.63698041e-01 2.08554596e-01
1.05463123e+00 3.63307148e-01 1.45452452e+00 -4.80265409e-01
7.87852108e-01 1.75090337e+00 2.97186732e-01 9.02362525e-01
-1.32052314e+00 -1.35679230e-01 -2.70369887e-01 -6.41307175e-01
-6.84942663e-01 -2.98493832e-01 2.61913151e-01 -5.88159680e-01
1.12454367e+00 4.37812120e-01 -5.78527525e-02 1.05865741e+00
2.77718902e-01 3.92805219e-01 1.54571176e+00 -4.28293407e-01
1.66073397e-01 6.65782094e-01 8.35109353e-01 4.63192940e-01
-3.65108609e-01 -4.97489842e-03 -3.69871318e-01 -3.42252582e-01
8.50074470e-01 -6.46565616e-01 -9.39734094e-03 2.82827497e-01
-1.14704895e+00 1.11887336e+00 1.74612105e-01 6.71624422e-01
1.13070179e-02 -3.36432964e-01 4.04487491e-01 9.13173258e-01
2.72073478e-01 1.17574894e+00 -4.41737026e-01 -6.67024970e-01
-2.33463898e-01 5.03209114e-01 1.67247856e+00 1.14362776e+00
5.39852798e-01 5.69130033e-02 -6.22569919e-01 1.09198797e+00
2.80750662e-01 2.46180743e-01 3.69935006e-01 -1.43947506e+00
2.45393306e-01 8.69255304e-01 5.78384519e-01 -5.84315062e-01
-6.78966284e-01 4.09579605e-01 -2.61932462e-01 3.14157009e-01
8.88100743e-01 -5.30462027e-01 -3.16355795e-01 1.81632912e+00
-3.91323119e-02 -1.03786910e+00 5.65533757e-01 7.31394291e-01
1.36871946e+00 5.81840634e-01 2.71920651e-01 -3.23694199e-01
1.33063066e+00 -1.00507998e+00 -1.08022153e+00 -1.96404383e-01
9.07318532e-01 -1.03620863e+00 1.58859313e+00 1.95090607e-01
-1.13659394e+00 -4.73379880e-01 -8.61411095e-01 -4.25908417e-02
-3.67803961e-01 -3.46997857e-01 4.30293679e-01 1.02951407e+00
-1.29735541e+00 5.00005007e-01 -2.01066166e-01 -5.32540739e-01
-5.59029043e-01 2.34073818e-01 -1.55885622e-01 5.56337118e-01
-1.24653256e+00 1.61603022e+00 8.70722458e-02 -5.62864840e-01
-8.43786061e-01 -2.80387074e-01 -7.49085307e-01 -4.84217366e-04
1.73160017e-01 -1.94192231e-01 2.17616940e+00 -5.28344393e-01
-1.96833611e+00 9.50189888e-01 1.51561230e-01 -4.46549535e-01
3.56203139e-01 -2.31497929e-01 -5.16140699e-01 2.44289599e-02
8.74526501e-02 5.71989536e-01 1.27849624e-01 -1.47021127e+00
-4.54438567e-01 -6.61358535e-02 7.38104522e-01 5.02212644e-01
-3.67370099e-01 4.08442855e-01 3.54431272e-01 -1.95912853e-01
-3.60777676e-01 -7.57693410e-01 -3.32053304e-01 -5.61075509e-01
-3.65104258e-01 -7.95769215e-01 7.24487424e-01 -6.05052710e-01
1.37480879e+00 -1.50498152e+00 -1.80007681e-01 -1.37723312e-01
5.36798239e-01 1.90791428e-01 3.27230571e-03 7.76218951e-01
2.13364810e-01 1.05928563e-01 -8.78891721e-02 -9.98886526e-02
3.57729524e-01 1.41481563e-01 -8.47175047e-02 -1.63019538e-01
1.38116583e-01 7.49530792e-01 -1.04457664e+00 -6.23932719e-01
5.03467202e-01 -4.08702344e-02 -3.99624228e-01 8.84638667e-01
-5.54219544e-01 4.77806091e-01 -2.58989453e-01 2.78218627e-01
1.21368460e-01 -1.20731525e-01 3.33985537e-01 3.86639744e-01
-3.19692552e-01 6.70785248e-01 -5.94486892e-01 1.27820396e+00
-5.01170397e-01 7.58720398e-01 5.63337058e-02 -7.38982856e-02
1.40110135e+00 6.09803379e-01 7.69482851e-02 -4.58031535e-01
5.61191678e-01 4.79837805e-02 3.78159463e-01 -3.33973140e-01
8.57640445e-01 -3.39069366e-01 -5.74447870e-01 9.50150788e-01
3.69769722e-01 -6.38871729e-01 2.45653257e-01 7.54623294e-01
1.09818876e+00 -5.78216434e-01 4.24854547e-01 -6.81145191e-01
8.43405306e-01 1.18784443e-01 -1.18000969e-01 7.16859281e-01
-6.87262058e-01 1.73910543e-01 8.19602132e-01 1.50854336e-02
-9.31657374e-01 -1.24713755e+00 -7.99828619e-02 1.41681433e+00
2.48696432e-01 -6.76243246e-01 -1.12818825e+00 -7.00009882e-01
-3.33530664e-01 1.29018068e+00 -2.90793270e-01 4.71748263e-02
-3.52175832e-01 -1.55560538e-01 7.81406045e-01 4.86195356e-01
3.58100086e-01 -1.16472459e+00 -3.48138064e-01 1.64975837e-01
-4.59160864e-01 -1.25938570e+00 -5.79122543e-01 1.50687486e-01
-6.72866285e-01 -8.25785935e-01 -4.28718895e-01 -7.39691317e-01
9.55317020e-02 5.50257340e-02 1.53774095e+00 8.40025097e-02
2.44179115e-01 7.36634254e-01 -5.30736804e-01 -2.00288698e-01
-1.27327144e+00 -1.08930975e-01 7.43475780e-02 -6.91546142e-01
6.27008080e-01 -2.48795580e-02 -1.30262211e-01 7.66085029e-01
-5.29483616e-01 -4.64991242e-01 2.52977386e-02 9.36560929e-01
-5.38069665e-01 -5.29472351e-01 9.18731987e-01 -7.84319401e-01
1.91539872e+00 -2.19740674e-01 -2.96241373e-01 3.06189895e-01
-7.12471247e-01 -1.26624117e-02 6.07973218e-01 -2.99398273e-01
-1.42986119e+00 -3.65357995e-01 -8.83273706e-02 3.73454243e-01
-5.88897169e-01 2.28089839e-01 -4.73447470e-03 1.93380624e-01
1.10567093e+00 -3.17095548e-01 3.03616881e-01 -2.10409045e-01
5.21918058e-01 1.12901449e+00 4.63981777e-01 -6.85830712e-01
5.21049201e-01 -2.45547295e-01 -6.41211390e-01 -1.22729504e+00
-7.36751199e-01 -6.13630414e-01 -6.01259530e-01 -5.69455624e-01
1.18368387e+00 -6.89286530e-01 -9.61682975e-01 1.89399019e-01
-1.11664152e+00 -7.10359693e-01 -1.07929133e-01 3.29209924e-01
-6.40478671e-01 4.80572909e-01 -9.19596553e-01 -1.24714720e+00
-4.87525821e-01 -1.04379392e+00 6.95121408e-01 4.60475504e-01
-1.24670196e+00 -1.36335433e+00 3.32997680e-01 5.43608963e-01
3.03345948e-01 6.42913282e-02 5.26336908e-01 -1.24723184e+00
9.91732776e-02 -5.81558496e-02 -1.69194520e-01 4.94171143e-01
4.22196805e-01 -7.96557665e-02 -1.02824736e+00 -3.90518121e-02
2.64087409e-01 -9.81199503e-01 -3.62279154e-02 -6.67122528e-02
7.85638914e-02 -5.46687007e-01 1.35652676e-01 -6.45625770e-01
7.52004921e-01 6.43478930e-01 3.24301809e-01 6.25823438e-02
5.03187031e-02 9.15071070e-01 8.29120994e-01 4.96868134e-01
4.00558829e-01 6.47670746e-01 -1.41490966e-01 1.32927865e-01
-6.91045076e-02 -1.03398934e-01 6.17179871e-01 9.00596917e-01
7.18015283e-02 -3.81981581e-01 -1.08873796e+00 4.12701726e-01
-1.51237535e+00 -8.01821053e-01 -4.11501348e-01 1.85241318e+00
1.08049750e+00 4.17773336e-01 5.95329404e-01 -2.85638034e-01
6.51156008e-01 2.24728793e-01 -5.41337468e-02 -1.02586973e+00
-4.48981673e-02 1.67072833e-01 -1.80555150e-01 1.32251561e+00
-8.53210568e-01 1.27786553e+00 7.29958868e+00 1.99571684e-01
-2.21041441e-01 7.87690952e-02 5.59968412e-01 4.20098662e-01
-1.81412056e-01 2.57216334e-01 -7.76793778e-01 1.69765875e-01
1.41364622e+00 -6.36622429e-01 2.04740494e-01 9.16020274e-01
1.74659878e-01 -1.76344693e-01 -1.53378344e+00 5.53278983e-01
-4.67722043e-02 -8.14263642e-01 -1.43450916e-01 -1.51626132e-02
6.38944626e-01 -4.75307584e-01 -2.60418355e-01 6.27466083e-01
1.07029128e+00 -1.08618081e+00 2.29791537e-01 7.39379078e-02
5.62200844e-01 -5.00805616e-01 8.33696783e-01 2.49144077e-01
-7.59128749e-01 2.33367547e-01 -2.71872789e-01 -3.73545170e-01
4.71685052e-01 -5.04479855e-02 -1.57699597e+00 1.18944675e-01
2.45766476e-01 2.00469062e-01 -3.10118109e-01 1.36149094e-01
-2.23816767e-01 3.60378355e-01 -4.86172102e-02 -6.82535470e-01
6.00420475e-01 -3.86657506e-01 7.03958452e-01 1.56692541e+00
-3.25658381e-01 3.90222102e-01 4.10179585e-01 8.24172258e-01
1.42125234e-01 9.21850353e-02 -8.32513750e-01 -2.87861705e-01
5.90207338e-01 1.28014314e+00 -3.95144135e-01 -3.39502513e-01
-4.12033051e-02 6.51969492e-01 6.38382286e-02 -1.01444550e-01
-7.42167771e-01 -4.74090219e-01 6.01621330e-01 -3.47526729e-01
-5.88920057e-01 -4.88185734e-01 -4.89591867e-01 -7.60322750e-01
-5.55174887e-01 -1.32415128e+00 4.83699799e-01 -5.22291183e-01
-1.55267847e+00 1.10078323e+00 1.04882531e-01 -1.05172634e+00
-7.62839317e-01 -7.44882584e-01 -9.60164785e-01 8.51210296e-01
-5.20944417e-01 -8.13828766e-01 -4.33245420e-01 5.99245489e-01
9.55073059e-01 -5.92910707e-01 1.30161011e+00 -2.82785922e-01
-2.02564657e-01 6.93599284e-01 -3.36087763e-01 2.86502481e-01
9.73067939e-01 -1.76847756e+00 3.59947056e-01 1.51424468e-01
-1.24433637e-01 9.60389256e-01 1.07725966e+00 -5.58127701e-01
-8.23853672e-01 -4.78980899e-01 1.11477172e+00 -1.16421103e+00
1.08526480e+00 -1.00467905e-01 -7.71862447e-01 5.69125712e-01
9.13983047e-01 -1.27690578e+00 7.70662844e-01 4.77950931e-01
-4.81647849e-01 3.54914933e-01 -1.23656833e+00 8.18241060e-01
5.70677519e-01 -8.19546640e-01 -1.22139084e+00 6.65302515e-01
8.51594746e-01 -2.89518744e-01 -1.40046549e+00 -6.28308058e-02
3.67977023e-01 -1.10425746e+00 7.50290871e-01 -8.17088842e-01
5.26738584e-01 2.99807668e-01 -2.01812088e-01 -1.21184182e+00
-1.09973714e-01 -1.15198660e+00 3.46271247e-02 1.43148565e+00
8.48313391e-01 -6.67506576e-01 5.80652595e-01 1.27842879e+00
-3.00619483e-01 -1.61736667e-01 -6.51433229e-01 -8.82747948e-01
4.81495529e-01 6.89080581e-02 2.45669767e-01 1.02689409e+00
1.02210402e+00 1.31782341e+00 -2.98320353e-01 -2.66224444e-01
3.27983230e-01 -3.95102054e-02 1.07770717e+00 -1.26607347e+00
7.16619194e-02 -6.43068016e-01 -9.37248841e-02 -1.25347424e+00
1.38086870e-01 -4.28426296e-01 1.52074769e-01 -1.45822763e+00
-1.02749854e-01 -2.29402617e-01 2.65173942e-01 1.66391432e-01
-1.71024352e-01 -4.35473770e-01 2.16378465e-01 1.27708405e-01
-5.12882769e-01 5.31731486e-01 1.18095410e+00 -4.72723506e-02
-5.29772103e-01 -4.41697314e-02 -7.39334345e-01 7.70453751e-01
1.44004571e+00 -6.40949160e-02 -5.72935104e-01 8.89416039e-02
-4.52895164e-01 4.03887749e-01 -2.76998002e-02 -9.92342234e-01
9.48703662e-02 -3.31815451e-01 -2.97241420e-01 -3.57807517e-01
6.21571779e-01 -4.71294612e-01 -4.66904163e-01 4.39600885e-01
-9.16940570e-01 1.89876080e-01 -9.80870798e-02 2.48245448e-01
-3.50853115e-01 -5.72863281e-01 8.08821142e-01 -2.40194827e-01
-4.38281983e-01 -4.31211144e-01 -7.77512729e-01 4.19455558e-01
8.78986776e-01 -1.28680468e-01 -6.64785683e-01 -1.06304681e+00
-5.86710334e-01 5.33735394e-01 2.29628980e-01 6.28099859e-01
4.46563303e-01 -9.44615066e-01 -6.62518799e-01 -3.53785038e-01
1.27977833e-01 -6.48360372e-01 -1.91443682e-01 3.65971863e-01
-5.49510241e-01 6.84017360e-01 -2.85665363e-01 -4.64274198e-01
-1.60578811e+00 1.39450654e-01 2.98479408e-01 -5.41111052e-01
-2.23320901e-01 7.44216263e-01 5.48199080e-02 -8.22437167e-01
4.50476527e-01 -1.44073740e-01 -4.72419202e-01 2.95810029e-02
5.48663139e-01 4.70544189e-01 -3.21562588e-01 -5.78646004e-01
-6.42563924e-02 -2.24022523e-01 -2.84744740e-01 -7.66602218e-01
7.59849966e-01 -1.93497360e-01 -1.14880770e-01 8.26326370e-01
8.60292792e-01 -2.64656246e-01 -5.77392578e-01 -9.65558365e-02
4.85587388e-01 -2.18311623e-01 -5.49813271e-01 -1.12578094e+00
2.27874349e-04 6.44967079e-01 3.89265805e-01 9.28317189e-01
4.47899342e-01 1.16214067e-01 5.78093946e-01 9.36288834e-01
3.35512906e-01 -1.48463166e+00 7.27676094e-01 1.24819660e+00
1.19308984e+00 -1.47106314e+00 -1.67886794e-01 -4.88855988e-01
-1.22040105e+00 1.02025723e+00 1.26949036e+00 2.65545845e-01
4.18364584e-01 2.19948560e-01 6.48702741e-01 -4.67812747e-01
-8.92309427e-01 -3.17758322e-03 -1.44520076e-03 5.45226634e-01
1.04663336e+00 2.74703711e-01 -5.59598327e-01 4.83093739e-01
-7.48824418e-01 -5.96695840e-01 7.06324160e-01 9.09334838e-01
-8.52344751e-01 -1.26571226e+00 -3.38663310e-01 2.59019524e-01
-7.61061013e-02 1.16729289e-01 -1.53925550e+00 9.49736357e-01
-9.19469178e-01 1.86793089e+00 -2.90834635e-01 -5.76273263e-01
4.38527733e-01 3.46171558e-01 1.72446430e-01 -8.37057769e-01
-1.25913489e+00 -5.10068983e-02 1.01841486e+00 -1.69078887e-01
-4.58542138e-01 -2.15732917e-01 -1.21283340e+00 -5.19233108e-01
-5.72011232e-01 7.84800291e-01 7.87635222e-02 7.18078375e-01
-1.49740085e-01 1.82031870e-01 8.36263418e-01 -3.76552314e-01
-1.15660203e+00 -1.70384705e+00 -4.32503641e-01 6.06994212e-01
-2.92153731e-02 -5.46058953e-01 -4.66626018e-01 3.25795151e-02] | [12.911748886108398, 8.076598167419434] |
01199ff2-bfb0-4eca-b6ab-360462faa128 | synonymous-generalization-in-sequence-to | 2003.06658 | null | https://arxiv.org/abs/2003.06658v5 | https://arxiv.org/pdf/2003.06658v5.pdf | Revisit Systematic Generalization via Meaningful Learning | Humans can systematically generalize to novel compositions of existing concepts. Recent studies argue that neural networks appear inherently ineffective in such cognitive capacity, leading to a pessimistic view and a lack of attention to optimistic results. We revisit this controversial topic from the perspective of meaningful learning, an exceptional capability of humans to learn novel concepts by connecting them with known ones. We reassess the compositional skills of sequence-to-sequence models conditioned on the semantic links between new and old concepts. Our observations suggest that models can successfully one-shot generalize to novel concepts and compositions through semantic linking, either inductively or deductively. We demonstrate that prior knowledge plays a key role as well. In addition to synthetic tests, we further conduct proof-of-concept experiments in machine translation and semantic parsing, showing the benefits of meaningful learning in applications. We hope our positive findings will encourage excavating modern neural networks' potential in systematic generalization through more advanced learning schemes. | ['Zhouhan Lin', 'Xiangyu Liu', 'Wei Wang', 'Boxin Wang', 'Ning Shi'] | 2020-03-14 | from-scan-to-real-data-systematic | https://openreview.net/forum?id=9qKAGxS1Tq2 | https://openreview.net/pdf?id=9qKAGxS1Tq2 | null | ['systematic-generalization', 'novel-concepts'] | ['reasoning', 'reasoning'] | [ 7.20053494e-01 5.99225700e-01 -2.87164897e-01 -6.11113369e-01
-4.61429954e-01 -8.69367242e-01 7.07850456e-01 2.71304250e-01
-3.97866070e-01 1.07629478e+00 3.21921259e-01 -7.32512772e-01
-6.66468292e-02 -8.24520171e-01 -1.14074945e+00 -2.41832972e-01
-2.49046504e-01 6.34732008e-01 1.55674696e-01 -5.05564451e-01
4.04333085e-01 3.89549553e-01 -1.39434814e+00 3.62831771e-01
1.00464487e+00 3.05281848e-01 3.22012752e-01 4.73014832e-01
-4.90580529e-01 6.65957689e-01 -3.69875103e-01 -5.98178864e-01
-1.15644960e-02 -7.05778420e-01 -1.10279405e+00 -2.47378975e-01
4.72906649e-01 -1.97962031e-01 1.30009040e-01 1.01571953e+00
7.84053877e-02 2.57968068e-01 4.12403405e-01 -1.07928681e+00
-1.27558315e+00 1.14385724e+00 -1.02110922e-01 2.31525198e-01
2.53046244e-01 3.07263851e-01 1.16485691e+00 -6.92097783e-01
7.12080419e-01 1.26486909e+00 9.58754182e-01 1.13569272e+00
-1.27953660e+00 -7.37057149e-01 4.93737161e-01 3.75346214e-01
-7.75751173e-01 -3.96903247e-01 6.40514672e-01 -3.23046893e-01
1.15021932e+00 6.69286102e-02 8.20588112e-01 1.18504751e+00
7.79811246e-03 6.54200673e-01 1.11458886e+00 -6.88856065e-01
3.24429721e-01 -1.52675007e-02 3.46942320e-02 7.71939516e-01
5.70382893e-01 4.86316592e-01 -7.49614120e-01 1.73194885e-01
4.44950581e-01 -2.29664281e-01 -5.75147979e-02 -2.02025145e-01
-1.09039140e+00 8.59613955e-01 4.36539173e-01 4.81259525e-01
-3.69281210e-02 3.57350290e-01 3.98808897e-01 3.68746102e-01
1.54036567e-01 1.17255282e+00 -8.99738729e-01 1.18994161e-01
-5.60375690e-01 2.28864610e-01 7.51041174e-01 1.15751505e+00
7.68962920e-01 1.69553056e-01 4.68751609e-01 3.93411547e-01
-1.34378612e-01 2.89483279e-01 6.64265156e-01 -1.28338647e+00
-5.43445610e-02 3.33586872e-01 -2.06807435e-01 -5.91444552e-01
-3.23780090e-01 -5.27459323e-01 -4.62765872e-01 -1.37287542e-01
3.55980396e-01 -2.49209836e-01 -6.72377467e-01 2.33652329e+00
-4.86617796e-02 2.04443142e-01 2.63712674e-01 3.61194938e-01
3.75274956e-01 4.18013155e-01 5.89599133e-01 -4.55494016e-01
1.03052151e+00 -5.62970161e-01 -2.86150336e-01 -5.06996334e-01
7.92004585e-01 -4.32803810e-01 1.33341157e+00 3.94753039e-01
-1.26481402e+00 -6.38080060e-01 -1.12011504e+00 -1.24431223e-01
-5.20988286e-01 -5.97227693e-01 1.35467875e+00 7.26982951e-01
-1.01401556e+00 9.41825688e-01 -6.12834930e-01 -7.00178683e-01
7.28013813e-01 3.72581571e-01 -1.13924146e-01 4.17749435e-02
-1.15233064e+00 1.04798186e+00 9.63614345e-01 -1.86994225e-01
-8.56509924e-01 -9.75928962e-01 -5.47434807e-01 1.32798404e-01
6.35687053e-01 -1.15089190e+00 1.55535614e+00 -1.33936095e+00
-1.21556509e+00 9.28537071e-01 -1.21310577e-01 -7.77354777e-01
8.66437256e-02 -4.66940142e-02 -2.58818090e-01 2.04932824e-01
5.03723808e-02 1.09190559e+00 4.77803200e-01 -1.49150610e+00
-6.84990823e-01 -2.80963361e-01 4.41218197e-01 5.60983755e-02
-4.30435836e-01 -2.37063468e-01 4.49321002e-01 -6.66107893e-01
3.00493211e-01 -9.30070281e-01 -1.62502512e-01 -6.32828251e-02
2.30399352e-02 -4.46626216e-01 1.23073459e-01 -4.01065916e-01
8.31612825e-01 -1.66423225e+00 8.16618577e-02 -2.00329706e-01
1.45699725e-01 1.89604998e-01 -1.71058163e-01 3.21073443e-01
-2.99531996e-01 4.56223071e-01 -3.24998319e-01 2.16650680e-01
1.66121364e-01 4.36292678e-01 -5.84882200e-01 -2.32195914e-01
3.16856354e-01 1.36191475e+00 -1.22540498e+00 -2.21698761e-01
-3.60020250e-01 -9.35246050e-02 -6.79632485e-01 -6.06392100e-02
-7.47212648e-01 3.63877475e-01 -2.48177022e-01 4.32520688e-01
2.31587410e-01 -3.35612535e-01 6.88836038e-01 7.12160021e-02
2.99961418e-01 4.93658513e-01 -4.29127395e-01 2.06096578e+00
-4.80890870e-01 5.24598122e-01 -4.67026323e-01 -1.39002633e+00
7.39982188e-01 1.67978451e-01 -1.42884582e-01 -5.82921147e-01
7.03620687e-02 3.12586635e-01 4.54469889e-01 -6.96213722e-01
4.86991793e-01 -1.15769458e+00 -1.43297408e-02 6.40118122e-01
1.82984397e-01 -3.08312595e-01 1.85553282e-01 3.27663600e-01
6.44069970e-01 4.64371055e-01 2.65468717e-01 -4.60129529e-01
2.82449424e-01 2.83578396e-01 4.49798167e-01 9.74433661e-01
-4.56836075e-02 8.47819448e-02 3.96092355e-01 -5.99645317e-01
-1.30584455e+00 -1.36129367e+00 1.48477092e-01 1.48935390e+00
-7.85170794e-02 -1.62842736e-01 -7.05617368e-01 -7.64785290e-01
-2.10319415e-01 1.29901016e+00 -6.98022246e-01 -5.52324891e-01
-8.87689948e-01 -6.24605894e-01 6.45017445e-01 9.05306578e-01
1.86231628e-01 -1.18680203e+00 -5.78616023e-01 2.43848413e-01
4.28197999e-03 -8.24008763e-01 6.88908622e-02 3.17517072e-01
-1.51450348e+00 -8.68782938e-01 -3.83598536e-01 -1.17736220e+00
7.31186986e-01 -2.28677914e-02 1.26420653e+00 3.07692707e-01
5.54568321e-02 4.15272564e-01 -2.78299451e-01 -4.68843013e-01
-9.58515823e-01 2.73984015e-01 2.61354327e-01 -9.21142638e-01
7.29532599e-01 -1.12081826e+00 -2.43117452e-01 2.70551536e-02
-9.91756201e-01 2.71818131e-01 5.57004750e-01 9.20527577e-01
3.48226987e-02 -1.57296330e-01 1.21452534e+00 -1.13654566e+00
8.07885170e-01 -4.76185292e-01 -2.80394197e-01 5.09927511e-01
-8.14192891e-01 4.41905558e-01 6.67284071e-01 -6.16018355e-01
-1.31970966e+00 -2.61832207e-01 -7.32000098e-02 3.09476137e-01
-2.91585565e-01 6.30793631e-01 4.89621470e-03 1.70146480e-01
1.16440928e+00 2.94162959e-01 -9.17275995e-02 -2.81954944e-01
8.46191108e-01 1.59671500e-01 7.48117149e-01 -1.41733503e+00
8.01332772e-01 3.66692632e-01 2.82871015e-02 -6.02856934e-01
-1.03792882e+00 3.18767011e-01 -7.61133075e-01 3.07527691e-01
5.76451480e-01 -7.35552371e-01 -9.13342714e-01 -6.66137934e-02
-1.14143896e+00 -3.72515172e-01 -5.53246260e-01 3.94947469e-01
-7.78152764e-01 4.51465487e-01 -6.15552306e-01 -4.47746426e-01
-2.09469169e-01 -7.01879382e-01 3.23244691e-01 4.33761567e-01
-6.99224412e-01 -1.14898086e+00 -2.09878922e-01 3.00719768e-01
3.76675099e-01 -7.68066347e-02 1.66369176e+00 -9.82018173e-01
-6.84209108e-01 7.54655153e-02 1.11637143e-02 3.99546653e-01
-1.06917340e-02 -4.60064828e-01 -9.85995650e-01 -2.03983068e-01
-2.79718153e-02 -7.18566597e-01 9.10938978e-01 -2.92609038e-04
1.02095330e+00 -3.17886144e-01 -3.84359688e-01 2.58378953e-01
1.31960332e+00 4.30089206e-01 4.07366008e-01 3.13768506e-01
4.62476611e-01 8.60885680e-01 6.05103150e-02 -6.69324473e-02
2.45597765e-01 1.95543505e-02 -1.52581617e-01 4.42623168e-01
-2.56366640e-01 -5.91716349e-01 8.53646323e-02 8.44908595e-01
-1.25302747e-01 -6.58097491e-02 -9.76609468e-01 5.64960063e-01
-1.48129582e+00 -1.00929534e+00 4.40991789e-01 1.98433495e+00
1.36959720e+00 5.72473943e-01 -1.06726862e-01 -1.56414732e-01
5.85708857e-01 -2.89542228e-01 -6.84322774e-01 -5.11075258e-01
-1.82020366e-01 6.66530848e-01 -2.89825927e-02 5.12141824e-01
-4.70875114e-01 1.36822057e+00 7.09596920e+00 6.67442918e-01
-7.51146197e-01 7.97496736e-02 4.67065036e-01 8.92390311e-02
-8.58268261e-01 3.59960705e-01 -5.64926147e-01 2.81848405e-02
1.03961265e+00 -3.75543147e-01 5.99055290e-01 6.56418800e-01
-3.07919443e-01 -9.58966166e-02 -1.59772730e+00 3.07218105e-01
1.23835906e-01 -1.67723787e+00 5.81793666e-01 -3.84522408e-01
9.17831600e-01 -3.28434914e-01 3.69464345e-02 4.87229735e-01
5.56761324e-01 -1.11434293e+00 5.81329763e-01 4.00554240e-01
5.41679919e-01 -7.72212267e-01 2.88580805e-01 3.03307116e-01
-5.62664628e-01 -1.48333251e-01 -5.16142309e-01 -5.32403171e-01
1.44972235e-01 -3.23152281e-02 -1.11417150e+00 1.40980959e-01
2.32892618e-01 5.55411816e-01 -6.48162007e-01 6.00685418e-01
-6.60700679e-01 5.92238128e-01 1.05372913e-01 -3.13832700e-01
1.94758013e-01 1.95995703e-01 2.06897855e-01 9.97338533e-01
2.39520863e-01 3.53716910e-01 -1.01222426e-01 1.13106561e+00
-2.62041956e-01 -1.63803875e-01 -5.75976372e-01 -4.58501846e-01
4.88311619e-01 6.33445084e-01 -9.87878025e-01 -7.27473319e-01
-3.38906646e-01 6.85005367e-01 3.79323989e-01 4.07414705e-01
-3.88216913e-01 -1.41612634e-01 2.99338549e-01 -7.27918968e-02
2.48553291e-01 -3.62369657e-01 -8.41571987e-01 -1.10093164e+00
1.35734997e-04 -9.38080728e-01 2.61372298e-01 -1.01767731e+00
-1.41720843e+00 3.55451554e-01 1.95643693e-01 -5.17526269e-01
-3.49345177e-01 -7.17799306e-01 -4.48500693e-01 7.47104466e-01
-1.22560668e+00 -1.21193254e+00 9.62461531e-02 1.63910761e-01
7.28096902e-01 -3.72841448e-01 1.00918198e+00 -3.44969362e-01
1.00030430e-01 5.89711428e-01 -1.22022368e-01 -8.06925297e-02
3.64183396e-01 -9.70388055e-01 9.15037453e-01 7.82173157e-01
2.68833846e-01 1.24158978e+00 8.96717787e-01 -7.10842967e-01
-1.15868843e+00 -5.15708745e-01 1.03913105e+00 -7.13673055e-01
7.47372806e-01 -3.67143035e-01 -1.08341432e+00 9.72684741e-01
2.96970069e-01 -6.21408641e-01 8.85944188e-01 5.78688979e-01
-7.35075772e-01 7.69681856e-02 -8.68608057e-01 9.12967622e-01
1.64012730e+00 -5.84569156e-01 -1.38591003e+00 3.75163108e-01
1.29839313e+00 7.74260759e-02 -6.20634258e-01 4.85764712e-01
7.51271904e-01 -8.88103902e-01 1.06186974e+00 -1.30198193e+00
7.72624552e-01 4.14405130e-02 -2.66370088e-01 -1.34489584e+00
-2.18226209e-01 -4.63195443e-01 1.47123739e-01 9.80841637e-01
7.26126730e-01 -5.76480269e-01 8.26370299e-01 5.63079178e-01
-4.34183002e-01 -5.07352352e-01 -5.68572879e-01 -1.11430013e+00
8.54436159e-01 -6.49019361e-01 7.13952661e-01 1.19033897e+00
5.62523484e-01 7.09140480e-01 1.25705898e-01 -1.86089948e-01
4.26637173e-01 1.17929861e-01 3.63091826e-01 -1.35591447e+00
-4.90990996e-01 -6.51324868e-01 -4.63042706e-02 -8.58680964e-01
6.22774959e-01 -1.20650804e+00 6.37229383e-02 -1.22431767e+00
3.87249917e-01 -4.45739836e-01 -1.19229481e-01 4.32149440e-01
-2.35934108e-01 1.15427785e-01 1.99488938e-01 1.80429276e-02
-4.31268543e-01 2.60614038e-01 1.29888797e+00 -2.44010076e-01
-7.97699019e-02 -3.60391259e-01 -1.40267491e+00 9.38030243e-01
1.06378579e+00 -4.54434007e-01 -8.56252134e-01 -5.87172508e-01
7.16398537e-01 -3.00511420e-01 3.53533387e-01 -9.08696115e-01
8.84356275e-02 -4.21129137e-01 4.84990954e-01 4.51452360e-02
6.46030381e-02 -4.85342354e-01 -1.74297184e-01 6.92156672e-01
-8.56142640e-01 7.53669888e-02 4.62336689e-01 5.22304475e-01
2.57601380e-01 -4.80761111e-01 6.57062113e-01 -8.35689604e-01
-1.13637877e+00 -1.73112363e-01 -2.31288984e-01 4.44180220e-01
7.46823967e-01 -2.65016109e-01 -4.11998004e-01 -2.77793974e-01
-1.10878336e+00 -6.04824163e-03 5.42659760e-01 4.13488030e-01
6.26787543e-01 -1.03270984e+00 -5.74322402e-01 6.80860728e-02
2.29490906e-01 -2.75941491e-01 1.46708503e-01 2.86371231e-01
-4.82471764e-01 4.57042605e-01 -3.94445449e-01 -2.91613609e-01
-7.27813721e-01 1.08875656e+00 1.84508771e-01 3.31194699e-01
-4.64758277e-01 1.04046178e+00 3.63609225e-01 -5.14210999e-01
1.79294989e-01 -3.87847334e-01 1.83818609e-01 -1.06736347e-01
3.74753654e-01 1.66797176e-01 -2.56184816e-01 -6.03769273e-02
-1.28941208e-01 4.75654066e-01 -3.29316795e-01 -2.61941344e-01
1.27600956e+00 -1.46232232e-01 -8.86560678e-02 4.54707056e-01
9.05550063e-01 -2.68200785e-01 -1.01746798e+00 -3.65577936e-01
4.32975024e-01 -1.55664653e-01 -7.43993402e-01 -1.23208535e+00
-2.11525530e-01 1.10008752e+00 2.16467753e-01 -9.14717764e-02
9.79784787e-01 3.12886119e-01 8.52251589e-01 9.93731856e-01
5.19762397e-01 -1.05829215e+00 3.97596806e-01 4.97682869e-01
7.29004800e-01 -1.03726614e+00 1.26165554e-01 -1.86934397e-01
-4.70295429e-01 1.14306140e+00 6.31354213e-01 3.76060940e-02
1.34720787e-01 -1.17286421e-01 -1.93186775e-01 4.55840304e-02
-9.01964068e-01 -1.15007743e-01 -3.18337791e-02 9.39283907e-01
6.09214187e-01 1.40008458e-03 -4.18328941e-01 5.06858528e-01
-5.86190224e-01 1.66396558e-01 5.91462553e-01 9.40997005e-01
-7.97311962e-01 -1.30200434e+00 -7.66557157e-02 1.10973462e-01
-9.39741656e-02 -4.95945960e-01 -3.80198628e-01 8.91522586e-01
4.88867968e-01 7.18110085e-01 -2.62241419e-02 -1.93232208e-01
-1.30383307e-02 6.76824629e-01 9.66083288e-01 -8.85606945e-01
-4.00680959e-01 -3.90137285e-01 2.33359501e-01 -5.36342412e-02
-4.77542132e-01 -6.24682486e-01 -1.51273227e+00 -4.00628299e-01
5.17135719e-03 2.41511643e-01 3.90993804e-01 1.19466126e+00
1.37363657e-01 4.26673234e-01 6.26172721e-02 -2.67416835e-01
-7.99116850e-01 -7.88037062e-01 -2.48789564e-01 4.55856323e-01
1.00193888e-01 -3.43845606e-01 -1.73845947e-01 4.39875662e-01] | [9.685881614685059, 7.335309982299805] |
07992888-b598-445d-9901-a4af39f2d76c | a-comparison-of-latent-semantic-analysis-and | 2108.06197 | null | https://arxiv.org/abs/2108.06197v4 | https://arxiv.org/pdf/2108.06197v4.pdf | A comparison of latent semantic analysis and correspondence analysis of document-term matrices | Latent semantic analysis (LSA) and correspondence analysis (CA) are two techniques that use a singular value decomposition (SVD) for dimensionality reduction. LSA has been extensively used to obtain low-dimensional representations that capture relationships among documents and terms. In this article, we present a theoretical analysis and comparison of the two techniques in the context of document-term matrices. We show that CA has some attractive properties as compared to LSA, for instance that effects of margins, i.e. sums of row elements and column elements, arising from differing document-lengths and term-frequencies are effectively eliminated, so that the CA solution is optimally suited to focus on relationships among documents and terms. A unifying framework is proposed that includes both CA and LSA as special cases. We empirically compare CA to various LSA based methods on text categorization in English and authorship attribution on historical Dutch texts, and find that CA performs significantly better. We also apply CA to a long-standing question regarding the authorship of the Dutch national anthem Wilhelmus and provide further support that it can be attributed to the author Datheen, amongst several contenders. | ['Peter G. M. van der Heijden', 'Tejaswini Deoskar', 'David J. Hessen', 'Qianqian Qi'] | 2021-07-25 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [ 1.68964192e-01 -1.94290921e-01 -2.89793223e-01 -2.76966870e-01
-3.31707865e-01 -8.13557088e-01 1.03983104e+00 3.65892738e-01
-3.78302217e-01 1.84260949e-01 9.37083066e-01 -3.89409631e-01
-8.20125163e-01 -4.95086670e-01 4.04458195e-02 -5.97480536e-01
-3.99329662e-02 4.68381673e-01 -2.28443205e-01 -3.80914122e-01
7.82008648e-01 5.12009740e-01 -1.77371347e+00 1.84482828e-01
6.79205120e-01 5.32667756e-01 6.61768904e-03 4.58489239e-01
-5.49660742e-01 5.67103088e-01 -5.77263117e-01 -5.84211707e-01
1.27436444e-01 -3.43372643e-01 -1.07831430e+00 2.20813453e-01
3.58649611e-01 -1.81828253e-02 -2.23583505e-01 1.01541722e+00
1.91155046e-01 9.78482142e-03 1.17858660e+00 -1.20406258e+00
-8.20306718e-01 6.41530752e-01 -3.60710800e-01 1.80383027e-01
4.63652968e-01 -5.69936693e-01 1.34993374e+00 -1.07385528e+00
8.41462910e-01 1.68960488e+00 7.39352524e-01 2.12056831e-01
-1.31387699e+00 -2.79494047e-01 7.88380206e-02 2.65230596e-01
-1.38081086e+00 -5.46907246e-01 8.54704440e-01 -6.62282765e-01
7.97898114e-01 5.13977468e-01 3.26184243e-01 1.07463408e+00
2.48732075e-01 6.27629876e-01 1.04581594e+00 -9.75564480e-01
1.97421402e-01 2.76817083e-01 6.14845037e-01 2.86892831e-01
4.33976352e-01 -2.36680031e-01 -4.20980901e-01 -6.09746873e-01
3.24365348e-01 -2.01805815e-01 1.23445340e-01 -5.64894617e-01
-1.26403475e+00 1.18822730e+00 -5.97935878e-02 8.15033853e-01
-3.33766222e-01 -1.36274278e-01 5.82750916e-01 2.31555030e-01
3.43271583e-01 4.75075215e-01 -6.88977167e-02 -2.15558320e-01
-9.57438946e-01 2.92670876e-01 8.84372711e-01 7.29668200e-01
3.27757925e-01 -1.51304349e-01 -4.57944460e-02 1.12917721e+00
5.56736469e-01 3.45787138e-01 9.02945578e-01 -1.07914364e+00
4.47965026e-01 6.05587780e-01 6.33266848e-03 -1.70130837e+00
-3.58604103e-01 -1.16604507e-01 -6.08249485e-01 -1.95536196e-01
2.96135366e-01 3.06067824e-01 -4.65226531e-01 1.56051683e+00
-5.44681847e-02 -5.59067428e-01 2.42481485e-01 6.56785607e-01
5.65143108e-01 5.35515189e-01 -1.55570153e-02 -5.41648924e-01
1.69716561e+00 -5.72290957e-01 -1.02534616e+00 -1.93490416e-01
7.67248571e-01 -1.08002484e+00 8.87389481e-01 2.91083127e-01
-9.87474561e-01 -2.77171642e-01 -9.12554204e-01 -3.01748961e-01
-5.64805448e-01 1.16263144e-01 6.88132107e-01 8.17612469e-01
-7.81330109e-01 7.15823352e-01 -5.13555884e-01 -7.21558690e-01
-7.30240569e-02 1.39576063e-01 -2.54685760e-01 2.43834525e-01
-1.30746448e+00 8.75701010e-01 9.62035209e-02 4.38740999e-02
-7.11745545e-02 -2.67522335e-01 -6.89973652e-01 -1.32115753e-02
2.39796653e-01 -2.58720338e-01 7.98477292e-01 -8.07337880e-01
-1.19653714e+00 1.18782628e+00 -5.42220414e-01 -2.38657057e-01
3.46754700e-01 5.46037890e-02 -6.94742084e-01 2.35281676e-01
4.21343774e-01 9.44099426e-02 8.24448884e-01 -1.15849483e+00
-3.43525976e-01 -6.44708455e-01 -2.48627439e-01 7.98598304e-02
-8.54944646e-01 3.29257280e-01 -2.83964902e-01 -1.11007488e+00
7.15919435e-01 -1.07719457e+00 2.60546088e-01 -2.00064033e-01
-2.89219797e-01 -5.46878219e-01 5.16595602e-01 -7.73815811e-01
1.84503472e+00 -2.20750690e+00 4.71823663e-01 6.21689081e-01
3.44105154e-01 -6.45857602e-02 4.94215749e-02 9.21994150e-01
-2.21833393e-01 2.21830696e-01 -2.20437154e-01 -2.56672412e-01
2.92152524e-01 3.89695704e-01 -4.68085289e-01 6.52600527e-01
-3.52562726e-01 6.42933965e-01 -6.86483324e-01 -6.69224799e-01
7.08683655e-02 4.00090039e-01 -1.47982165e-01 -4.27273810e-01
2.52338499e-01 -2.12922156e-01 -2.54673034e-01 4.75807816e-01
5.06084204e-01 -1.98500663e-01 5.76620579e-01 -1.64347380e-01
-4.24300879e-01 2.69769877e-01 -1.34207630e+00 1.37392712e+00
-1.65554866e-01 8.89462590e-01 1.17200114e-01 -8.73811781e-01
9.83184934e-01 2.14546651e-01 3.36184800e-01 -4.63974863e-01
3.46363336e-01 3.35788280e-01 -9.41243470e-02 -3.63785833e-01
9.08842146e-01 -1.79099604e-01 -2.95523345e-01 5.93257725e-01
-1.65720075e-01 2.08040193e-01 4.41584855e-01 6.81961775e-01
6.15388393e-01 -3.68871808e-01 5.10231018e-01 -7.84260452e-01
7.97399104e-01 4.04835120e-02 2.20528498e-01 5.07072389e-01
-1.35766119e-01 2.41578266e-01 7.11121023e-01 -2.54630655e-01
-1.21880925e+00 -7.18172252e-01 -5.79856813e-01 1.10323370e+00
1.07401730e-02 -8.35887253e-01 -7.28274167e-01 -1.94586784e-01
2.83568412e-01 9.13050830e-01 -7.95825839e-01 -3.94327082e-02
-3.27753276e-01 -6.60636246e-01 4.11612540e-01 4.49943215e-01
-4.88335919e-03 -6.40041769e-01 -5.19245684e-01 -1.09471520e-02
-1.46504864e-01 -8.21460962e-01 -2.31565088e-01 4.04725000e-02
-8.63193095e-01 -8.97604465e-01 -7.75265157e-01 -7.45032072e-01
5.58409452e-01 4.59600180e-01 6.75145626e-01 -6.25226349e-02
-1.73313707e-01 6.04074001e-01 -5.53960085e-01 -9.44807529e-02
-6.88233137e-01 7.70820975e-02 4.00348783e-01 -1.56304054e-02
7.31821954e-01 -3.71004313e-01 -1.76290944e-01 3.65007281e-01
-8.76552761e-01 -2.99533457e-01 2.49581486e-01 7.99193144e-01
1.61977306e-01 1.46682709e-01 2.73017973e-01 -8.85694861e-01
1.09693420e+00 -3.94471854e-01 -3.81679624e-01 3.05616111e-01
-9.77190197e-01 1.72056690e-01 2.66279489e-01 -2.41322264e-01
-9.65054333e-01 -3.80100965e-01 2.19830632e-01 -2.87079871e-01
5.71848676e-02 5.49870968e-01 1.73097271e-02 5.88031597e-02
7.28402376e-01 2.39118040e-01 1.71878934e-01 -6.81025624e-01
4.53923136e-01 1.21887422e+00 3.51001412e-01 -3.97172928e-01
6.35117292e-01 7.17870116e-01 1.22754961e-01 -1.04684269e+00
-6.00651324e-01 -6.33411288e-01 -8.97228479e-01 -8.38582870e-04
6.41719818e-01 -6.28568590e-01 -5.71807027e-01 1.22379780e-01
-1.03082073e+00 5.18863738e-01 -1.68015212e-01 5.59621692e-01
-4.94492441e-01 7.12774515e-01 -4.69668448e-01 -9.06177521e-01
-3.85390557e-02 -8.18862021e-01 7.38613784e-01 -3.73239398e-01
-8.48764062e-01 -1.32069850e+00 7.76406005e-02 4.04005438e-01
2.54343450e-01 1.41666502e-01 1.38788891e+00 -6.96331561e-01
4.28131461e-01 -2.93864220e-01 -6.18596077e-02 2.00548738e-01
2.77257711e-01 1.26805514e-01 -7.61410773e-01 -4.75146055e-01
2.21628472e-01 2.22246364e-01 7.17162669e-01 2.02225149e-01
6.72737360e-01 -3.81858408e-01 -4.50331181e-01 1.75554633e-01
1.23960245e+00 2.08321929e-01 4.23192203e-01 6.57963932e-01
7.24117517e-01 1.14780557e+00 3.60949606e-01 5.46793461e-01
3.36858213e-01 7.50079036e-01 -9.17781368e-02 4.50759023e-01
1.13233645e-02 -2.85182670e-02 2.65789956e-01 1.34721768e+00
7.98086002e-02 -1.92918003e-01 -1.08497047e+00 6.45063579e-01
-1.79847717e+00 -9.16978776e-01 -6.16371691e-01 2.08400416e+00
4.20413762e-01 5.10937124e-02 2.44926348e-01 5.59257507e-01
8.97411227e-01 2.23676518e-01 -1.22401647e-01 -6.45257473e-01
-4.19914454e-01 -2.44333327e-01 4.53312844e-01 6.01193786e-01
-9.25619602e-01 8.47998261e-01 7.60035563e+00 9.92101610e-01
-5.59239209e-01 -4.35889587e-02 1.61627278e-01 5.10940282e-03
-5.02184629e-01 1.68921366e-01 -7.43197978e-01 3.91094297e-01
9.09683347e-01 -4.94835526e-01 2.76327491e-01 6.42909586e-01
6.83121905e-02 -5.64192161e-02 -9.68605757e-01 1.04365122e+00
4.16764289e-01 -1.18414402e+00 4.07520741e-01 2.76579112e-01
6.97412550e-01 -6.42754376e-01 1.24236368e-01 -1.55778259e-01
-6.64795786e-02 -6.92315161e-01 8.34346354e-01 3.48999977e-01
4.85128403e-01 -7.95977771e-01 5.69429755e-01 7.52630979e-02
-9.55256879e-01 -3.49535823e-01 -5.92621624e-01 -2.20432952e-01
-2.00547390e-02 5.02191603e-01 -3.61194849e-01 4.37796265e-01
4.69074309e-01 7.96172559e-01 -6.30145252e-01 3.94123346e-01
2.33309139e-02 3.92423391e-01 2.89251357e-02 -9.91053656e-02
3.80961835e-01 -4.15079892e-01 8.85087073e-01 1.16699445e+00
1.31366193e-01 -5.39209954e-02 -3.38264346e-01 7.24331737e-01
3.21802735e-01 5.42705953e-01 -4.75309968e-01 -4.39962000e-01
6.08661890e-01 9.44408894e-01 -1.06584251e+00 -3.69827956e-01
-4.19351608e-01 9.92733479e-01 -1.56547874e-02 1.27709240e-01
-1.27572447e-01 -5.78809738e-01 5.21939695e-01 3.03186476e-01
2.18128830e-01 -3.13068271e-01 -3.31156194e-01 -1.11017215e+00
-4.39067977e-03 -7.77898908e-01 6.74908221e-01 -4.94409353e-01
-1.49053466e+00 5.52128613e-01 2.47393101e-01 -1.24221492e+00
-3.61004233e-01 -7.10939229e-01 -5.93887866e-02 8.04750025e-01
-8.95435095e-01 -9.09548044e-01 1.53660074e-01 5.59687495e-01
4.72657979e-01 -4.56112146e-01 8.99839580e-01 7.44023845e-02
-5.14208317e-01 5.18945277e-01 8.44766080e-01 -1.57057066e-02
6.28228009e-01 -1.10777903e+00 3.97806495e-01 6.87641740e-01
2.43303820e-01 1.11293387e+00 8.44411194e-01 -7.72823095e-01
-1.40844607e+00 -3.67358834e-01 1.55108857e+00 -6.03391588e-01
1.04915380e+00 -4.54692811e-01 -9.58720684e-01 5.59595704e-01
7.78315663e-02 -7.05780685e-01 9.44474578e-01 3.39938223e-01
-5.11798322e-01 2.75056958e-01 -9.54988062e-01 6.73251569e-01
8.19588482e-01 -7.08375275e-01 -1.08580482e+00 4.43643570e-01
4.71712768e-01 1.84772789e-01 -9.35855865e-01 -1.92305863e-01
6.97981775e-01 -8.41180861e-01 1.07794166e+00 -6.72969818e-01
3.12756777e-01 -1.90291762e-01 -3.86047751e-01 -9.81331348e-01
-6.86347544e-01 -5.22199512e-01 3.06632090e-02 1.39038944e+00
4.55301553e-02 -6.97081447e-01 4.65505809e-01 6.59245312e-01
2.58276343e-01 -2.26405799e-01 -1.10137594e+00 -1.05731988e+00
3.74510348e-01 -3.29320312e-01 3.61398637e-01 1.44415164e+00
4.68191355e-01 5.09751260e-01 -4.59890574e-01 -3.15413058e-01
6.72661424e-01 9.14499685e-02 3.44345957e-01 -1.69790792e+00
1.41722038e-01 -6.38849318e-01 -5.32966018e-01 -6.09130085e-01
3.51919860e-01 -1.04693460e+00 -4.64051038e-01 -1.41082656e+00
3.42612714e-02 -2.52203524e-01 8.25361684e-02 1.25605255e-01
2.21417263e-01 -5.21475486e-02 2.05444232e-01 9.46370482e-01
-2.08534092e-01 3.71119618e-01 7.89641798e-01 1.31211191e-01
-2.96274573e-01 -2.53557324e-01 -8.97918820e-01 7.82929242e-01
5.77019870e-01 -5.47817230e-01 -2.47617036e-01 -2.80516446e-01
5.04365385e-01 -2.14166448e-01 2.05222771e-01 -5.44796586e-01
2.85519630e-01 1.67964455e-02 3.55952561e-01 -5.46790600e-01
1.85403302e-01 -7.37361073e-01 2.01026320e-01 4.15068209e-01
-5.89872956e-01 2.49099374e-01 4.74951304e-02 6.74854040e-01
-2.13401794e-01 -7.37404644e-01 4.52578902e-01 -1.04275540e-01
-7.18031943e-01 -5.20806968e-01 -8.55573177e-01 -1.51581898e-01
7.91220069e-01 -5.21042824e-01 -1.44535810e-01 -2.83672601e-01
-7.29179800e-01 8.49456619e-03 4.35451537e-01 6.50968134e-01
3.39510560e-01 -1.62398553e+00 -6.86713457e-01 1.04939558e-01
3.13036561e-01 -9.01142895e-01 2.27307186e-01 7.38330185e-01
-4.97946650e-01 7.92240381e-01 -1.12278193e-01 -2.54531324e-01
-1.52349591e+00 8.74764085e-01 -3.14304352e-01 8.54686350e-02
-7.02680826e-01 5.88518798e-01 -1.15454376e-01 -2.13038713e-01
1.89259529e-01 1.05737947e-01 -5.74621797e-01 8.39931905e-01
5.64476907e-01 9.06673193e-01 -1.16386086e-01 -1.44171488e+00
-6.55756116e-01 9.27225888e-01 -1.56062901e-01 -3.41373324e-01
1.15522373e+00 -5.66534102e-01 -5.48600793e-01 8.03128362e-01
1.52231121e+00 1.78132102e-01 -3.98659497e-01 -4.51058209e-01
4.47378516e-01 -6.88356221e-01 2.71830052e-01 -3.95431072e-01
-4.43254143e-01 7.86452532e-01 5.38461864e-01 5.56722224e-01
7.49692798e-01 9.82273370e-02 3.62329572e-01 2.76614934e-01
-3.91265377e-02 -1.49196315e+00 -8.20778236e-02 4.03614938e-01
1.01663685e+00 -8.02284896e-01 4.33394223e-01 -4.52983439e-01
-6.81592345e-01 1.34440637e+00 -1.84876874e-01 3.25922444e-02
5.55980504e-01 -2.44179070e-01 2.40767817e-03 -4.15829301e-01
-4.48566675e-01 -1.17253223e-02 4.11819816e-01 2.30826810e-01
5.02965748e-01 5.10015599e-02 -9.90999758e-01 5.51366210e-01
-4.86096680e-01 -5.27574301e-01 6.19006395e-01 1.01597905e+00
-3.20180595e-01 -1.29788971e+00 -8.96522462e-01 4.40740705e-01
-3.83552760e-01 -9.01786014e-02 -7.66842484e-01 7.23953485e-01
-2.25465834e-01 9.22593355e-01 2.94171691e-01 -3.44171613e-01
1.54711157e-01 3.56652915e-01 1.88890561e-01 -3.13608944e-01
-5.04465759e-01 3.68103027e-01 5.77275455e-02 -1.59092337e-01
-6.55452371e-01 -1.21719623e+00 -9.27340031e-01 -5.72029650e-01
-4.37156945e-01 4.67523128e-01 8.49519789e-01 8.54537964e-01
2.11204305e-01 1.91774502e-01 5.66455305e-01 -5.00926912e-01
-5.46523333e-01 -8.72726023e-01 -9.90081906e-01 4.49949205e-01
1.66813061e-01 -7.87195086e-01 -8.78055811e-01 -2.67213620e-02] | [10.193150520324707, 7.4711012840271] |
319f00bd-0970-4d11-86a2-75b7abd6453e | variational-psom-deep-probabilistic | 1910.01590 | null | https://arxiv.org/abs/1910.01590v3 | https://arxiv.org/pdf/1910.01590v3.pdf | DPSOM: Deep Probabilistic Clustering with Self-Organizing Maps | Generating interpretable visualizations from complex data is a common problem in many applications. Two key ingredients for tackling this issue are clustering and representation learning. However, current methods do not yet successfully combine the strengths of these two approaches. Existing representation learning models which rely on latent topological structure such as self-organising maps, exhibit markedly lower clustering performance compared to recent deep clustering methods. To close this performance gap, we (a) present a novel way to fit self-organizing maps with probabilistic cluster assignments (PSOM), (b) propose a new deep architecture for probabilistic clustering (DPSOM) using a VAE, and (c) extend our architecture for time-series clustering (T-DPSOM), which also allows forecasting in the latent space using LSTMs. We show that DPSOM achieves superior clustering performance compared to current deep clustering methods on MNIST/Fashion-MNIST, while maintaining the favourable visualization properties of SOMs. On medical time series, we show that T-DPSOM outperforms baseline methods in time series clustering and time series forecasting, while providing interpretable visualizations of patient state trajectories and uncertainty estimation. | ['Gunnar Rätsch', 'Matthias Hüser', 'Laura Manduchi', 'Julia Vogt', 'Vincent Fortuin'] | 2019-10-03 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [-2.70342380e-01 2.83086121e-01 3.17284048e-01 -3.31622303e-01
-5.23080885e-01 -4.64364767e-01 8.59590828e-01 4.47289437e-01
1.69885978e-01 4.17464495e-01 7.72399843e-01 -5.70476651e-01
-7.60886192e-01 -5.47856331e-01 -3.32098961e-01 -9.26249683e-01
-7.99413741e-01 9.37960625e-01 -2.13983059e-01 2.14066878e-01
8.63150805e-02 3.62694979e-01 -1.46060038e+00 7.27373183e-01
9.38841641e-01 6.82883084e-01 1.28129926e-02 3.74609500e-01
-2.45535016e-01 7.61463225e-01 -6.84638381e-01 3.34868461e-01
-2.63108104e-01 -2.99916118e-01 -6.36058033e-01 -1.80867538e-01
-3.72393169e-02 2.95560151e-01 -2.41481826e-01 7.07881153e-01
4.89007443e-01 4.33518440e-01 1.06794131e+00 -1.50994766e+00
-9.12781656e-01 8.98231566e-01 -4.67297494e-01 5.16778380e-02
-5.69967143e-02 -8.98548439e-02 4.82743114e-01 -5.25232196e-01
7.85448670e-01 1.60529184e+00 8.45373571e-01 5.10784328e-01
-1.70395923e+00 -3.05508286e-01 1.72132745e-01 2.12509677e-01
-1.16196299e+00 -1.32319883e-01 8.14011872e-01 -7.95201302e-01
1.00565422e+00 4.44131881e-01 6.22583926e-01 1.33702266e+00
4.36850369e-01 7.04095304e-01 1.38122809e+00 -1.28619105e-01
7.97567904e-01 -9.53531340e-02 2.73337454e-01 3.94091785e-01
-2.13049173e-01 6.96195588e-02 -5.08878946e-01 -3.21482480e-01
6.06535554e-01 5.66360176e-01 -5.46206627e-03 -7.10189581e-01
-1.82356811e+00 8.84380996e-01 7.24589705e-01 5.82225025e-01
-5.39193094e-01 5.45624971e-01 5.74651837e-01 2.37549216e-01
5.89437306e-01 5.74107289e-01 7.33810067e-02 -1.88705772e-01
-1.32320404e+00 -1.43441543e-01 5.23589075e-01 5.95726967e-01
2.09953114e-01 2.74548054e-01 -5.45078218e-01 4.61807519e-01
2.41291359e-01 1.58268124e-01 8.72114837e-01 -1.09343314e+00
5.24141192e-02 6.58758402e-01 -8.63562673e-02 -1.44277251e+00
-9.74608839e-01 -2.19791368e-01 -1.44502366e+00 4.29614186e-01
8.90258420e-03 -1.52910441e-01 -1.30542266e+00 1.61705494e+00
-1.45155117e-02 2.56420910e-01 1.75545350e-01 5.96944094e-01
7.39471972e-01 9.41150427e-01 9.74779502e-02 -3.03188950e-01
1.10180104e+00 -6.92865789e-01 -1.15483570e+00 3.67630452e-01
5.04144013e-01 -1.40919790e-01 8.97846699e-01 3.26122105e-01
-9.31644857e-01 -3.87708306e-01 -8.16389441e-01 8.45752731e-02
-8.32511067e-01 5.35611100e-02 6.11432135e-01 4.25130308e-01
-1.77842355e+00 8.44238937e-01 -1.68245018e+00 -4.67267960e-01
4.95955229e-01 3.94638389e-01 -1.69312611e-01 4.50991422e-01
-1.03692853e+00 8.58969986e-01 6.18924260e-01 2.12623447e-01
-7.06576586e-01 -6.45653903e-01 -7.80772865e-01 3.06162238e-01
-2.84925163e-01 -7.85232604e-01 5.56467772e-01 -3.05525899e-01
-1.30379486e+00 2.52524495e-01 -2.15214059e-01 -7.36457348e-01
4.99138683e-01 8.93361345e-02 -4.54698145e-01 7.13173971e-02
-1.82976630e-02 1.07158017e+00 7.25633264e-01 -1.40625310e+00
-1.79109260e-01 -1.35633931e-01 -6.38982117e-01 1.56848043e-01
-6.11549854e-01 -3.75058234e-01 -3.53064761e-02 -6.90185785e-01
3.28437775e-01 -8.05209279e-01 -7.54177034e-01 -2.21632034e-01
-8.47036719e-01 -4.42232668e-01 1.25689900e+00 -5.83981395e-01
1.39626539e+00 -2.16261315e+00 4.95192051e-01 5.78195632e-01
7.65162170e-01 -8.92256852e-03 2.27276325e-01 7.03767955e-01
-4.39404547e-01 1.33005336e-01 -5.69410205e-01 -8.01745057e-01
3.93558234e-01 3.72034788e-01 -6.26918733e-01 1.99101970e-01
2.48270240e-02 1.16238666e+00 -1.03572357e+00 -5.27410626e-01
6.23801112e-01 8.38160217e-01 -2.22876891e-01 -2.69262195e-01
-1.47531196e-01 5.61916649e-01 9.08503905e-02 2.31817201e-01
4.06166017e-01 -6.24981284e-01 3.93827319e-01 9.97810289e-02
-2.14322463e-01 1.78139150e-01 -1.10316992e+00 1.84240305e+00
-1.21458650e-01 1.03834188e+00 -4.51818556e-01 -9.87152934e-01
8.95688772e-01 5.43053985e-01 8.32437158e-01 -3.93069357e-01
-2.35662118e-01 -9.31305438e-02 -9.68238637e-02 -2.02289775e-01
2.67998666e-01 4.35316749e-02 1.62640028e-02 7.43613422e-01
-2.17276681e-02 4.83734757e-01 6.62840083e-02 3.72825056e-01
1.21479857e+00 -3.01327463e-02 -1.28091767e-01 -6.66300178e-01
-1.88088268e-01 1.38893381e-01 2.02432722e-01 6.68146372e-01
8.27096701e-02 7.48312175e-01 8.37968111e-01 -7.70212650e-01
-9.76646602e-01 -1.47784221e+00 -2.38787487e-01 8.09279144e-01
-4.26701605e-01 -6.68912828e-01 -7.74768174e-01 -4.48036075e-01
3.13960202e-02 8.54776561e-01 -1.17915738e+00 -8.86599198e-02
-4.26252723e-01 -1.09534109e+00 3.47062677e-01 6.58595324e-01
-2.56877411e-02 -1.30771315e+00 -8.42514634e-01 3.82726043e-01
-2.25158766e-01 -1.63289174e-01 -1.05879709e-01 4.87411171e-01
-1.14395785e+00 -5.32940269e-01 -6.54705524e-01 -6.48358941e-01
7.42202282e-01 -5.06679714e-02 1.12819684e+00 -3.95678401e-01
-2.75452971e-01 3.55407774e-01 -1.12176174e-02 -4.25682306e-01
-4.59163636e-01 1.62926167e-02 3.05364400e-01 -1.96427926e-01
4.61072952e-01 -1.10378647e+00 -6.79204047e-01 1.34217963e-01
-9.55680311e-01 3.59274894e-01 3.28595430e-01 7.12617934e-01
6.56109631e-01 3.16412836e-01 6.90999031e-01 -7.85737753e-01
9.02289987e-01 -7.84656286e-01 -2.19005957e-01 1.50056064e-01
-7.94685543e-01 2.25298390e-01 4.96475726e-01 -4.15885955e-01
-6.12272322e-01 1.96004599e-01 3.41167837e-01 -1.06785488e+00
-2.77121395e-01 7.63023198e-01 4.08108890e-01 7.47690380e-01
8.41388047e-01 2.77822673e-01 2.49353707e-01 -5.37698209e-01
7.80954540e-01 3.41019988e-01 7.47013032e-01 -3.11489850e-01
5.09828150e-01 9.99177158e-01 1.40559310e-02 -5.98077476e-01
-1.10302053e-01 -3.23319077e-01 -1.01373923e+00 -1.04528956e-01
1.08452833e+00 -5.38702190e-01 -8.61115694e-01 -9.02671665e-02
-9.50793743e-01 -3.67309242e-01 -5.22939086e-01 3.87767047e-01
-7.49392092e-01 2.34233975e-01 -5.91303945e-01 -9.39685702e-01
-3.43098611e-01 -8.75813484e-01 1.06113458e+00 -3.76966208e-01
-7.65588999e-01 -1.60562980e+00 2.75836974e-01 -4.85558480e-01
6.37299418e-01 1.27853417e+00 1.22258210e+00 -5.37248731e-01
-1.95895344e-01 3.04329753e-01 -3.60011719e-02 -2.98214197e-01
2.79709250e-01 4.88073975e-02 -9.90719259e-01 -4.21681821e-01
4.47811782e-02 2.49305800e-01 1.19916987e+00 9.73553836e-01
1.29377198e+00 -5.59127867e-01 -9.47885990e-01 5.43590009e-01
9.32199478e-01 2.96372235e-01 5.67761242e-01 1.60958603e-01
9.15648103e-01 1.03776574e+00 -1.21588245e-01 2.88467377e-01
5.90096354e-01 3.67029697e-01 4.09247994e-01 -4.01292503e-01
4.09978516e-02 -2.05184013e-01 1.77349880e-01 1.23115265e+00
-2.35409644e-02 -2.55853795e-02 -1.43904948e+00 7.11086690e-01
-2.62503457e+00 -1.01818144e+00 -2.36003891e-01 1.88203824e+00
4.87894982e-01 -3.12136598e-02 1.99887574e-01 2.68395990e-01
7.09187865e-01 1.29576147e-01 -5.85503876e-01 -3.72141093e-01
-4.51023839e-02 -2.61515975e-01 -9.73737314e-02 2.52930492e-01
-1.32710445e+00 3.93913478e-01 6.76226187e+00 6.30241871e-01
-9.72708106e-01 9.27168801e-02 8.70296538e-01 -3.50029349e-01
-5.28819442e-01 -4.45439905e-01 4.53207940e-02 7.32885361e-01
1.49111927e+00 -3.66042033e-02 2.27339953e-01 6.82389319e-01
3.67373347e-01 5.20421267e-01 -1.44844186e+00 1.25781095e+00
-2.09838077e-01 -1.83091080e+00 1.77240387e-01 2.79870987e-01
7.91157067e-01 8.62104297e-02 5.80733538e-01 2.29806036e-01
8.42691600e-01 -1.58661985e+00 4.84434515e-01 8.67098033e-01
6.00277662e-01 -8.38521242e-01 3.94961983e-01 1.63993552e-01
-1.07202411e+00 -1.81856170e-01 -3.66622597e-01 1.68689951e-01
1.80962890e-01 5.75263500e-01 -8.06016445e-01 5.95935762e-01
1.18675888e+00 1.15103865e+00 -6.27354741e-01 1.02683032e+00
2.16393650e-01 6.40699685e-01 -4.86338586e-01 2.60175109e-01
4.97674465e-01 -2.93763638e-01 5.69365799e-01 1.39111626e+00
4.18643475e-01 -3.76850963e-01 -5.48901744e-02 1.36116767e+00
3.46096992e-01 -3.99455696e-01 -7.93934941e-01 -1.35481849e-01
4.48437274e-01 1.18451250e+00 -1.23217201e+00 -5.78292370e-01
2.21767634e-01 8.82511199e-01 3.28447849e-01 5.27601063e-01
-5.10924280e-01 -2.06965506e-01 4.80554193e-01 1.61698628e-02
1.50393173e-01 -4.23408270e-01 -6.05176926e-01 -7.90918291e-01
-1.59797743e-01 -4.81204480e-01 9.03068781e-01 -7.95907080e-01
-1.44827175e+00 8.41067314e-01 8.42231438e-02 -1.59318852e+00
-5.88018179e-01 -3.72824639e-01 -7.76010513e-01 7.00578392e-01
-9.00544763e-01 -1.12263215e+00 9.12236329e-03 5.95357478e-01
1.55631557e-01 -2.01761737e-01 1.26180184e+00 -3.66511904e-02
-3.30935925e-01 4.91610169e-02 1.02923381e+00 -1.19569197e-01
4.03837740e-01 -2.06619430e+00 7.93817103e-01 6.11273706e-01
4.64053512e-01 1.04448164e+00 6.64615154e-01 -6.10558927e-01
-9.26038206e-01 -1.40595496e+00 5.54015458e-01 -1.10628796e+00
5.54346502e-01 -7.56415188e-01 -1.20539784e+00 6.94846869e-01
3.96781832e-01 -3.85233045e-01 9.45482552e-01 3.90836895e-01
-2.38766506e-01 2.94981033e-01 -9.39155161e-01 7.07504511e-01
6.40084803e-01 -4.07691598e-01 -7.74605572e-01 6.22198939e-01
8.66885364e-01 -2.36807823e-01 -1.00397706e+00 6.44116029e-02
2.65022278e-01 -9.93055344e-01 1.01587749e+00 -6.45147145e-01
4.11621660e-01 -5.31337678e-01 1.91646084e-01 -1.78282058e+00
-6.24410570e-01 -8.77887070e-01 -7.34213173e-01 9.89241600e-01
2.47132748e-01 -4.84720230e-01 6.74915552e-01 5.62649488e-01
-4.41939205e-01 -7.89895713e-01 -1.00719297e+00 -7.11302698e-01
2.90975332e-01 -2.52987415e-01 6.98775768e-01 1.45807719e+00
4.66922909e-01 -8.54444280e-02 -3.04582626e-01 1.53212890e-01
6.25345588e-01 2.34615088e-01 1.64241791e-01 -1.62695003e+00
4.57277410e-02 -7.11789727e-01 -4.34746861e-01 -3.12727571e-01
-7.81930611e-02 -1.12741756e+00 -1.32114813e-01 -2.24993920e+00
-2.47610286e-02 -2.87827641e-01 -8.40480149e-01 7.48438418e-01
1.24947578e-01 1.81528553e-01 1.03606939e-01 7.86761582e-01
-8.28726530e-01 5.42227447e-01 6.85783744e-01 -1.44460589e-01
-5.81977248e-01 -3.47361773e-01 -5.24398267e-01 5.57191372e-01
6.42994344e-01 -3.99069726e-01 -5.94434381e-01 -3.68564218e-01
5.75123467e-02 1.60057619e-01 4.02584910e-01 -1.19864333e+00
5.52334547e-01 2.03391150e-01 7.78844833e-01 -1.10963702e+00
3.43448818e-01 -7.00693727e-01 4.61809814e-01 6.82078063e-01
-4.54282850e-01 4.92239773e-01 4.02067989e-01 9.34129059e-01
-2.63162553e-01 6.38344646e-01 3.92833054e-01 1.64450184e-02
-4.46727663e-01 6.73038960e-02 -8.02441597e-01 -5.26968539e-01
8.01656902e-01 -2.70157695e-01 -5.13728082e-01 -4.63979453e-01
-1.33590090e+00 7.10734546e-01 1.95982680e-01 6.19695127e-01
6.73904896e-01 -1.68936133e+00 -5.53586543e-01 4.33861315e-02
-1.22231670e-01 7.41603151e-02 4.72808778e-01 8.75294209e-01
-4.22030151e-01 6.21691942e-01 -4.04081792e-01 -1.15965497e+00
-8.51422250e-01 1.07530081e+00 6.83460087e-02 -2.11915299e-01
-1.18393290e+00 3.98890913e-01 3.55028898e-01 -6.87146366e-01
4.40231055e-01 -5.09615958e-01 -4.76908803e-01 4.58230555e-01
5.12533367e-01 5.85977733e-01 2.52679251e-02 -2.96437353e-01
-3.70008230e-01 1.14004076e-01 6.39550462e-02 -1.92284048e-01
1.68489909e+00 -1.27959117e-01 -3.64966959e-01 1.09142542e+00
9.49278235e-01 -7.83892155e-01 -1.29583538e+00 7.25101605e-02
5.26539922e-01 9.60514098e-02 -1.25414908e-01 -8.80615056e-01
-7.82602131e-01 1.35682559e+00 8.72925937e-01 6.39014959e-01
9.26471770e-01 -3.88703458e-02 3.92492861e-01 5.01941919e-01
-5.41698039e-02 -9.43576097e-01 1.33764043e-01 1.09032586e-01
8.76965702e-01 -9.99088824e-01 -2.80337602e-01 2.00611442e-01
-7.32903421e-01 1.25028789e+00 1.42431214e-01 -8.98874775e-02
8.00368607e-01 1.29218563e-01 3.78245771e-01 -8.23970795e-01
-9.53167319e-01 1.54857069e-01 5.86408436e-01 9.01681781e-01
3.19998562e-01 3.59920382e-01 4.00290698e-01 2.69773722e-01
-4.73344266e-01 -4.33952898e-01 3.89838934e-01 6.71784282e-01
-3.79415303e-02 -6.01227999e-01 -4.26080137e-01 7.12448418e-01
1.81550439e-02 -1.40816402e-02 -2.07234934e-01 5.53362608e-01
-2.19344899e-01 6.90574467e-01 5.95714033e-01 -3.99863839e-01
-1.06709078e-01 3.22267771e-01 -2.87267298e-01 -4.81679440e-01
-5.40501773e-01 2.34599724e-01 -4.09001231e-01 -6.06440485e-01
-4.36465830e-01 -6.87794924e-01 -1.22493696e+00 -3.47824663e-01
3.36386949e-01 3.81877661e-01 6.73895836e-01 4.42347944e-01
8.61621082e-01 1.03591180e+00 2.51200110e-01 -1.12426758e+00
3.03153358e-02 -9.97182488e-01 -6.55988455e-01 4.91417408e-01
4.22434568e-01 -7.39309907e-01 -1.26355574e-01 1.61814421e-01] | [8.108152389526367, 3.945328950881958] |
b174b573-6235-4d8a-8d6e-4a94a4b4835f | variational-learning-for-the-inverted-beta | 2112.14375 | null | https://arxiv.org/abs/2112.14375v1 | https://arxiv.org/pdf/2112.14375v1.pdf | Variational Learning for the Inverted Beta-Liouville Mixture Model and Its Application to Text Categorization | The finite invert Beta-Liouville mixture model (IBLMM) has recently gained some attention due to its positive data modeling capability. Under the conventional variational inference (VI) framework, the analytically tractable solution to the optimization of the variational posterior distribution cannot be obtained, since the variational object function involves evaluation of intractable moments. With the recently proposed extended variational inference (EVI) framework, a new function is proposed to replace the original variational object function in order to avoid intractable moment computation, so that the analytically tractable solution of the IBLMM can be derived in an elegant way. The good performance of the proposed approach is demonstrated by experiments with both synthesized data and a real-world application namely text categorization. | ['Yuping Lai', 'Heping Song', 'Qiang Ruan', 'Wenbo Guan', 'Yongfa Ling'] | 2021-12-29 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [ 4.51163985e-02 9.38084163e-03 -1.64751440e-01 -2.71517247e-01
-9.79743719e-01 -1.80594534e-01 8.85232985e-01 -9.73198935e-02
-5.19419491e-01 8.25789869e-01 -4.64683175e-01 -5.18595278e-01
-3.86622161e-01 -3.90333265e-01 -4.46521729e-01 -1.10844064e+00
5.14035881e-01 6.27902269e-01 -1.15909286e-01 2.48299330e-01
3.23684633e-01 1.85677037e-01 -1.59229589e+00 -3.75023037e-01
1.36782169e+00 1.11637533e+00 2.68384844e-01 4.07554209e-01
-4.08371955e-01 4.85912353e-01 -5.29435337e-01 -4.04558629e-01
-1.63407531e-02 -2.78284103e-01 -4.74311560e-01 1.84986573e-02
1.98965333e-02 -1.84238523e-01 9.11920592e-02 1.36878562e+00
2.07960531e-01 6.10608101e-01 1.24589169e+00 -1.47063720e+00
-4.82177109e-01 1.90058351e-01 -8.00756693e-01 9.18430611e-02
1.67623814e-02 -4.36840177e-01 1.01321054e+00 -1.09056878e+00
3.10167342e-01 1.29526377e+00 4.08106983e-01 3.35319072e-01
-1.33802724e+00 -3.01995069e-01 2.16458693e-01 2.90252805e-01
-1.75014710e+00 -2.98104048e-01 9.65334356e-01 -3.93040478e-01
6.04431689e-01 2.60677040e-01 4.98233467e-01 1.05990005e+00
1.64891183e-01 1.12905931e+00 9.41179574e-01 -3.73083442e-01
4.21000093e-01 5.03152251e-01 2.84087509e-01 4.51365918e-01
2.86462843e-01 -4.22839522e-01 -1.25706926e-01 -6.09555304e-01
5.53826690e-01 2.15000883e-02 -7.43700713e-02 -4.29335535e-01
-7.58343875e-01 9.96803880e-01 -8.66626129e-02 2.07884595e-01
-4.56607699e-01 1.06718145e-01 1.92761287e-01 -1.99417844e-01
8.56471062e-01 -1.57400876e-01 -1.28021851e-01 -1.60442978e-01
-1.22093964e+00 3.11900973e-01 8.02005231e-01 6.86831951e-01
4.77621883e-01 2.95532495e-01 -2.95010120e-01 1.03409719e+00
1.07701588e+00 7.62076139e-01 2.43542507e-01 -9.76124823e-01
3.10018122e-01 3.71972769e-01 5.64909875e-01 -8.76524329e-01
-3.43990028e-02 -4.42581296e-01 -1.13314879e+00 3.08165587e-02
6.03538811e-01 4.16312255e-02 -5.69233894e-01 1.69065177e+00
6.53573215e-01 1.96159944e-01 2.23195881e-01 6.50183022e-01
7.97219932e-01 1.00840724e+00 8.33424777e-02 -5.64328969e-01
1.08672571e+00 -2.80017972e-01 -1.04061556e+00 9.78807807e-02
1.58237904e-01 -7.07247913e-01 9.50305045e-01 5.86076975e-01
-1.07348192e+00 -2.72891700e-01 -9.58966494e-01 -1.19047306e-01
-1.40639797e-01 1.01385333e-01 4.73862708e-01 5.93161881e-01
-7.15544105e-01 2.47677818e-01 -9.10574019e-01 -3.33696753e-02
2.26487041e-01 3.30342501e-01 2.96822358e-02 2.32398614e-01
-1.03294468e+00 6.41008019e-01 3.04126650e-01 6.36429965e-01
-7.38309741e-01 -6.79743469e-01 -7.10228205e-01 8.78539905e-02
2.95641541e-01 -6.48713648e-01 1.21015012e+00 -6.44393265e-01
-1.85156751e+00 4.91155475e-01 -5.71710348e-01 -4.03029472e-02
8.50596786e-01 1.12243086e-01 -4.91673164e-02 1.02593012e-01
-8.31846595e-02 2.11126134e-01 1.35126340e+00 -1.38190520e+00
-3.07142586e-01 -2.92770743e-01 -2.15863138e-01 -1.74085665e-02
-4.61265981e-01 -3.28272164e-01 -3.65117252e-01 -6.54814005e-01
2.27500811e-01 -6.84648097e-01 -6.58382028e-02 -2.47562230e-01
-3.38640600e-01 -5.12634516e-01 8.90753806e-01 -7.01609671e-01
1.22018158e+00 -1.98412621e+00 3.07973504e-01 3.70669484e-01
6.17739782e-02 2.00930685e-01 1.89610034e-01 3.02847475e-01
2.42273107e-01 -1.38914748e-03 -5.21486163e-01 -7.45130777e-01
1.82311937e-01 3.26082051e-01 -4.46291476e-01 6.46989882e-01
-5.01846941e-03 6.41136587e-01 -8.00686836e-01 -7.34312832e-01
3.85565728e-01 7.05999494e-01 -8.30201626e-01 1.16877705e-01
-2.53724158e-01 4.91870284e-01 -6.11991227e-01 3.67139399e-01
1.08300292e+00 -4.16262895e-01 3.58579338e-01 -6.00247793e-02
-1.50102347e-01 -2.55749524e-01 -1.49604702e+00 1.26095855e+00
-3.79590273e-01 4.84490395e-01 4.00738508e-01 -1.22506797e+00
8.22997510e-01 4.45561588e-01 6.49508059e-01 -1.36261255e-01
3.50772053e-01 2.03989714e-01 -2.74277300e-01 -3.23801219e-01
5.84820688e-01 -3.03542733e-01 7.79786706e-02 2.54496008e-01
8.40425491e-03 -3.67684476e-02 1.89825669e-01 2.30334267e-01
2.26205081e-01 4.83086817e-02 2.59516299e-01 -6.20659530e-01
9.36120093e-01 -2.52391309e-01 5.06196022e-01 8.05392742e-01
-7.34904781e-02 2.61687279e-01 5.27468264e-01 1.79162487e-01
-8.80746126e-01 -1.28506696e+00 -5.71232021e-01 6.33537948e-01
1.64950699e-01 -8.80966038e-02 -7.92551160e-01 -2.55354732e-01
-6.83021918e-02 8.54234695e-01 -4.13943589e-01 5.38931042e-02
-2.03024864e-01 -1.18345320e+00 2.58683681e-01 1.65077239e-01
5.63073397e-01 -5.23300171e-01 -1.94570035e-01 2.02044055e-01
-4.82108116e-01 -9.14701879e-01 -1.44699886e-01 -2.56594062e-01
-9.32914138e-01 -7.72701800e-01 -1.05628288e+00 -3.89030933e-01
5.96964002e-01 3.38084139e-02 6.12942219e-01 -2.17225730e-01
9.02974457e-02 4.76187766e-01 -5.67132607e-02 -2.39444494e-01
-6.48658276e-01 -9.37752575e-02 1.53723523e-01 4.12101418e-01
3.16350698e-01 -3.28107417e-01 -4.71783251e-01 1.17655158e-01
-1.01747751e+00 -9.73526239e-02 1.99222594e-01 1.06555808e+00
8.07384551e-01 1.90119982e-01 7.61982858e-01 -5.96068203e-01
7.01036990e-01 -5.69810092e-01 -1.11854327e+00 4.30805326e-01
-6.93062305e-01 1.11234941e-01 4.65478361e-01 -5.73775768e-01
-1.69654477e+00 -4.14844692e-01 -3.17628682e-01 -5.78522444e-01
2.66635567e-01 6.80001616e-01 -2.35394925e-01 1.39823914e-01
-2.37862710e-02 3.97257000e-01 1.81587011e-01 -7.51189947e-01
3.40348989e-01 8.90325427e-01 6.17545724e-01 -8.72359157e-01
4.44775075e-01 6.84257627e-01 1.87862664e-01 -1.19345725e+00
-7.22789407e-01 -4.51765299e-01 -5.35789490e-01 -1.89420685e-01
8.36423874e-01 -6.74956143e-01 -1.30972040e+00 5.73040724e-01
-1.32264578e+00 2.57945322e-02 -8.28655511e-02 6.78894699e-01
-5.98896980e-01 7.69420028e-01 -4.32707042e-01 -1.64306664e+00
-3.01189303e-01 -1.19030595e+00 7.97873080e-01 1.82155624e-01
-4.29723077e-02 -1.25631642e+00 2.14272495e-02 2.37911254e-01
6.33302853e-02 9.44207534e-02 1.16551936e+00 -4.61441159e-01
-3.19954962e-01 -2.30339840e-01 -2.20827878e-01 4.88043517e-01
-2.53010958e-01 4.75581676e-01 -9.44537044e-01 -2.64415860e-01
4.18471724e-01 1.54252857e-01 6.41548157e-01 7.77781248e-01
9.77103591e-01 -1.18942119e-01 -2.12815911e-01 4.96139526e-01
1.34402883e+00 2.39853248e-01 3.71788383e-01 -1.08072110e-01
4.45074737e-01 4.38324451e-01 5.45630276e-01 7.43781924e-01
2.88300723e-01 4.26207483e-01 2.57542431e-01 3.24586391e-01
4.29934114e-01 -1.71052366e-01 2.20803380e-01 1.05179906e+00
-3.55199166e-02 -4.88984317e-01 -7.02064753e-01 3.23735595e-01
-2.10594153e+00 -9.63174105e-01 -5.26068568e-01 2.30881715e+00
4.77003962e-01 -7.61017874e-02 -6.08631968e-03 2.03019872e-01
6.20947182e-01 7.46578574e-02 -6.93939388e-01 -3.36710364e-01
4.30842116e-02 -1.83046624e-01 8.14615563e-02 7.96417594e-01
-9.62101877e-01 5.27997136e-01 7.03600883e+00 1.27777600e+00
-7.17417896e-01 3.68041694e-01 4.88667190e-01 1.13629811e-01
-5.01510084e-01 -7.90204927e-02 -8.58935058e-01 5.91603339e-01
7.39520669e-01 -1.77620143e-01 3.70599866e-01 6.95624709e-01
4.36622828e-01 -5.21935999e-01 -6.83092952e-01 1.32061005e+00
-1.36248365e-01 -9.14694548e-01 -2.28813812e-02 3.76706839e-01
6.67621374e-01 -3.37453365e-01 3.54410857e-01 2.58334786e-01
-1.07007965e-01 -4.78906721e-01 7.89925754e-01 8.90725791e-01
5.70233047e-01 -9.49049532e-01 6.52633727e-01 7.46548772e-01
-7.95188665e-01 -3.20026428e-02 -5.13299346e-01 3.50142643e-02
2.68614501e-01 1.00669026e+00 -5.64103305e-01 4.53575552e-01
3.72014463e-01 4.61138397e-01 -1.12428339e-02 8.11252058e-01
1.42473700e-02 6.50639296e-01 -5.04305542e-01 -2.80363917e-01
1.85132474e-01 -9.63450670e-01 1.08836567e+00 8.22224915e-01
4.37475324e-01 -4.89955768e-02 -1.25502363e-01 1.31165218e+00
1.07221857e-01 2.23492905e-01 -1.98202923e-01 -2.52653956e-01
2.64941692e-01 1.18701065e+00 -8.52497280e-01 -3.51666421e-01
-2.58713067e-01 7.09268630e-01 3.32267106e-01 8.52997184e-01
-8.45991194e-01 -1.00928389e-01 7.16142654e-01 -3.09305638e-01
3.30835372e-01 -3.26443136e-01 -3.46527249e-01 -1.38655841e+00
5.84962368e-02 -5.74667990e-01 3.59495699e-01 -3.80753696e-01
-1.46743739e+00 4.18311954e-01 4.42568332e-01 -9.78972971e-01
-6.93652451e-01 -5.67263126e-01 -4.96459335e-01 1.07786429e+00
-1.15099168e+00 -7.33514845e-01 1.29203379e-01 6.17615819e-01
1.15756780e-01 -1.03717215e-01 6.14243507e-01 4.15299565e-01
-1.04932880e+00 6.07195616e-01 8.81322742e-01 -4.21197623e-01
2.76563704e-01 -1.15251803e+00 -3.41225564e-01 7.11551964e-01
-1.04631171e-01 6.90781593e-01 1.12863111e+00 -5.59824109e-01
-1.46629047e+00 -5.76778710e-01 6.81121945e-01 -3.73647749e-01
6.15689814e-01 -3.71587038e-01 -1.07085061e+00 3.18391234e-01
-6.97794929e-02 -3.12566310e-01 7.49070287e-01 -7.17078075e-02
-7.34172985e-02 8.04736763e-02 -1.28698587e+00 5.00734746e-01
3.68042141e-01 -4.01601523e-01 -4.60925728e-01 1.39859200e-01
2.68956363e-01 -5.45147583e-02 -1.02924156e+00 4.06832874e-01
6.20397985e-01 -6.62036061e-01 1.00845647e+00 -4.03939128e-01
2.49263376e-01 -3.92120272e-01 -4.37426299e-01 -8.69841158e-01
1.02359867e-02 -7.39272535e-01 -5.74627697e-01 1.33180583e+00
6.27201749e-03 -9.86687183e-01 5.55855691e-01 7.84751713e-01
2.72194386e-01 -7.50624597e-01 -1.28777993e+00 -7.31043041e-01
1.42119914e-01 -8.90080929e-01 4.20049936e-01 5.52886426e-01
-2.08685905e-01 1.34288564e-01 -5.54054916e-01 2.43055238e-03
9.77360487e-01 1.83708265e-01 6.30699217e-01 -1.46768248e+00
-4.39457834e-01 -5.01989543e-01 -3.97545211e-02 -1.43552458e+00
5.43586612e-01 -9.80329752e-01 1.48328796e-01 -1.48166060e+00
3.82443607e-01 -3.23858649e-01 -3.78354728e-01 -1.68423161e-01
-3.49319994e-01 1.15324549e-01 1.20321929e-01 3.00088525e-01
-4.34406877e-01 1.11853492e+00 1.02679420e+00 -7.12562948e-02
-1.13013856e-01 2.56698400e-01 -5.47182202e-01 9.23295557e-01
3.35658848e-01 -5.72163045e-01 -6.35622263e-01 -1.55366436e-02
2.10645929e-01 1.98132530e-01 4.61222947e-01 -3.69217455e-01
5.20327054e-02 -1.14620961e-01 2.49657571e-01 -1.07447171e+00
5.89785039e-01 -6.16292477e-01 2.30317429e-01 1.05403051e-01
-4.50158753e-02 -3.49190444e-01 4.68691327e-02 8.16051066e-01
-2.60295898e-01 -7.03143716e-01 8.25381100e-01 3.91861409e-01
-2.97790051e-01 2.65743554e-01 -5.37337244e-01 7.08700111e-03
7.21741617e-01 -4.86802757e-02 1.97138190e-01 -4.33182240e-01
-7.56938100e-01 1.06990673e-01 5.51515892e-02 3.57976630e-02
5.09884953e-01 -1.38275170e+00 -6.22587085e-01 1.75212756e-01
-2.39160791e-01 -6.51016906e-02 6.22597635e-01 1.17298961e+00
-6.96001947e-02 4.43344384e-01 3.84307444e-01 -8.80404115e-01
-8.53350163e-01 5.83331943e-01 1.10837080e-01 -3.19256216e-01
-4.20123696e-01 4.33564931e-01 3.63018543e-01 -3.41373682e-01
9.64214951e-02 -1.42212987e-01 -1.59876198e-01 2.09000722e-01
5.34305036e-01 8.82652640e-01 -1.21330172e-01 -8.17381918e-01
-2.66593009e-01 2.80330032e-01 2.08084360e-01 -4.91764039e-01
9.91630614e-01 -4.55621600e-01 -3.20405364e-01 7.04990089e-01
1.42316020e+00 -2.06840485e-01 -1.13025594e+00 -1.53329745e-01
-1.35642782e-01 -4.14954752e-01 3.42441469e-01 -1.11459218e-01
-7.29719698e-01 1.02055514e+00 3.15379143e-01 3.81414413e-01
8.21596742e-01 -2.26878941e-01 6.07508063e-01 3.77781987e-01
2.06982270e-01 -1.14698124e+00 -3.22737157e-01 4.37476367e-01
7.64525831e-01 -1.28681147e+00 -3.21482904e-02 -3.15393597e-01
-3.61634940e-01 8.64828587e-01 1.61564320e-01 4.05107299e-03
1.10945451e+00 -3.65253799e-02 -1.51215881e-01 1.08077422e-01
-4.71335739e-01 1.28760383e-01 6.24678493e-01 3.83857906e-01
2.85297483e-01 4.05750424e-02 -6.05741382e-01 8.77768338e-01
2.39819437e-01 -9.01777595e-02 2.06485957e-01 4.71437901e-01
-1.29839137e-01 -9.09948945e-01 -4.63920921e-01 4.02014107e-01
-5.34498870e-01 1.02399789e-01 7.08449408e-02 7.87199378e-01
-3.91184419e-01 1.32138240e+00 2.37008616e-01 1.34948879e-01
-1.34384140e-01 2.57321119e-01 4.71533418e-01 -9.86180678e-02
7.97770321e-02 6.00069106e-01 -5.03714383e-01 -9.34878662e-02
-5.08908510e-01 -7.95389354e-01 -1.19545472e+00 -3.03994894e-01
-4.48459744e-01 5.26686668e-01 7.87106752e-01 1.16029370e+00
5.16474806e-02 3.41227740e-01 4.72052097e-01 -8.69350433e-01
-8.48222792e-01 -1.14994824e+00 -8.82412612e-01 1.64325193e-01
1.15603797e-01 -9.17448401e-01 -5.04333675e-01 -3.19842041e-01] | [7.072032928466797, 3.9644968509674072] |
d8c048b7-732b-49c3-81e5-e928506e40ae | bridging-the-gap-using-deep-acoustic | 2112.13758 | null | https://arxiv.org/abs/2112.13758v1 | https://arxiv.org/pdf/2112.13758v1.pdf | Bridging the Gap: Using Deep Acoustic Representations to Learn Grounded Language from Percepts and Raw Speech | Learning to understand grounded language, which connects natural language to percepts, is a critical research area. Prior work in grounded language acquisition has focused primarily on textual inputs. In this work we demonstrate the feasibility of performing grounded language acquisition on paired visual percepts and raw speech inputs. This will allow interactions in which language about novel tasks and environments is learned from end users, reducing dependence on textual inputs and potentially mitigating the effects of demographic bias found in widely available speech recognition systems. We leverage recent work in self-supervised speech representation models and show that learned representations of speech can make language grounding systems more inclusive towards specific groups while maintaining or even increasing general performance. | ['Cynthia Matuszek', 'Francis Ferraro', 'Edward Raff', 'Luke E. Richards', 'Gaoussou Youssouf Kebe'] | 2021-12-27 | null | null | null | null | ['language-acquisition'] | ['natural-language-processing'] | [ 3.41781735e-01 6.23695254e-01 -2.69833177e-01 -8.19042683e-01
-5.23067772e-01 -6.56609237e-01 7.99953282e-01 7.60100424e-01
-5.26469231e-01 4.92729366e-01 9.12873864e-01 -7.17464149e-01
1.59522831e-01 -7.61526167e-01 -8.12930644e-01 4.62877713e-02
-1.68875247e-01 4.33504552e-01 -1.80716410e-01 -4.60063905e-01
2.38417670e-01 3.79307002e-01 -1.57788980e+00 6.99223220e-01
7.27294207e-01 3.33265394e-01 3.30846816e-01 8.78309488e-01
-4.83881652e-01 1.41283429e+00 -8.76289904e-01 -2.23881245e-01
4.42351811e-02 -4.50868696e-01 -9.30453956e-01 2.56825909e-02
7.95157790e-01 -6.47599876e-01 -5.94998002e-01 7.63639390e-01
3.80548328e-01 4.63503838e-01 4.71558809e-01 -1.26031160e+00
-1.50075114e+00 1.24387968e+00 1.75675184e-01 2.69509792e-01
6.21709287e-01 3.30405295e-01 9.70167577e-01 -7.25470245e-01
4.69513446e-01 1.54079044e+00 5.26038826e-01 7.91626871e-01
-1.30566609e+00 -8.41741204e-01 5.09953082e-01 -3.20292234e-01
-9.54917252e-01 -1.00025451e+00 6.42478347e-01 -7.24340737e-01
1.23091722e+00 4.28635664e-02 5.90617955e-01 1.43394065e+00
-3.55093241e-01 1.02588940e+00 1.00184727e+00 -5.31741679e-01
4.45145041e-01 5.14168680e-01 4.25993145e-01 6.25299156e-01
2.96676069e-01 2.88176358e-01 -1.15151119e+00 4.22218591e-02
6.92938030e-01 -1.35432426e-02 1.26319258e-02 -2.82104164e-01
-1.35383904e+00 8.37846458e-01 8.04178834e-01 3.39362770e-01
-3.68699551e-01 2.24676386e-01 2.04214305e-01 3.35737497e-01
4.05576915e-01 6.57362759e-01 6.43796474e-02 -3.51928830e-01
-8.19064200e-01 4.80342191e-03 6.89808190e-01 1.21053672e+00
6.58292353e-01 3.20831239e-01 -2.21547961e-01 7.34552562e-01
5.89681149e-01 1.05357814e+00 4.69416559e-01 -6.13474250e-01
6.57901824e-01 7.67254829e-01 -2.33188868e-01 -8.80077660e-01
-3.51810634e-01 -1.70067653e-01 -2.49727786e-01 1.24568455e-01
1.41125068e-01 -3.33678484e-01 -1.05648601e+00 1.76997280e+00
-1.73889220e-01 -8.87143761e-02 4.52429861e-01 4.80794698e-01
1.12073159e+00 6.46455228e-01 5.95308006e-01 2.23095521e-01
1.09377348e+00 -4.71982747e-01 -5.14269888e-01 -8.44273686e-01
5.72023511e-01 -3.19149077e-01 1.39456391e+00 1.96484894e-01
-9.01987314e-01 -4.42853749e-01 -1.06285560e+00 -2.96956688e-01
-4.71241117e-01 -2.83161283e-01 8.03681016e-01 9.10858393e-01
-1.27660656e+00 2.16759428e-01 -8.33429337e-01 -7.37098992e-01
7.51639545e-01 2.79885978e-01 -4.10274684e-01 -2.79611703e-02
-1.00501394e+00 9.09762800e-01 2.07167357e-01 -2.48771206e-01
-1.22182214e+00 -6.31258786e-01 -1.24507570e+00 -1.23406179e-01
9.56835076e-02 -4.39804763e-01 1.75893438e+00 -1.10689092e+00
-1.25748658e+00 7.26305962e-01 -6.31518140e-02 -6.51347399e-01
-7.40676895e-02 -2.61754721e-01 -4.81133133e-01 8.12820047e-02
1.79859549e-01 9.88666117e-01 5.56717455e-01 -1.26020789e+00
-4.60449189e-01 -3.64156008e-01 3.36694062e-01 3.81962150e-01
-9.26910877e-01 -2.06075788e-01 3.37699026e-01 -3.69994104e-01
-2.78938115e-01 -6.20171428e-01 -1.67474091e-01 -1.05184712e-01
-1.88449383e-01 -4.62479815e-02 3.69910002e-01 -7.61957526e-01
8.61917436e-01 -2.20290875e+00 -4.37409878e-02 1.47044361e-01
4.03373480e-01 1.26275122e-01 -2.65968829e-01 7.38301992e-01
2.50734627e-01 3.22150975e-01 3.83341402e-01 -4.27335709e-01
-6.23347796e-02 2.23712608e-01 -6.88440025e-01 2.38932252e-01
2.31071532e-01 7.66841948e-01 -1.32081020e+00 -3.43666285e-01
2.89529562e-01 3.41162950e-01 -7.72326827e-01 4.04894680e-01
-2.60945231e-01 2.30648413e-01 -1.91646442e-01 4.23516572e-01
1.43394098e-01 -3.21865790e-02 6.25794977e-02 2.46485695e-01
-3.32428455e-01 8.53563607e-01 -6.94245398e-01 1.73760939e+00
-8.66247356e-01 1.01575363e+00 -4.58475715e-03 -7.39377379e-01
1.27834606e+00 2.45383993e-01 1.75690297e-02 -1.09016299e+00
7.24086314e-02 -2.94472694e-01 1.69666309e-03 -5.99014044e-01
7.55939305e-01 -5.96366644e-01 -1.11714393e-01 6.41299248e-01
2.87378043e-01 -2.36213908e-01 -5.39091825e-02 5.60577750e-01
9.10042703e-01 -3.18111122e-01 1.34297580e-01 -2.79624999e-01
-3.29667091e-01 3.71132851e-01 -4.49180827e-02 1.05798209e+00
-1.11889102e-01 3.45750690e-01 -1.13817677e-02 -1.25032589e-01
-8.10713232e-01 -1.27093625e+00 2.17686266e-01 1.91699719e+00
-6.37607127e-02 -3.61664832e-01 -4.46546704e-01 -4.38083202e-01
1.19452007e-01 1.39116204e+00 -5.69347799e-01 -5.60336411e-01
-2.21999869e-01 2.44983554e-01 7.32078671e-01 9.83897448e-01
1.57812476e-01 -1.45456207e+00 -6.98012650e-01 -8.18028953e-03
-7.07718730e-02 -1.03479052e+00 -2.51846522e-01 4.72651869e-01
-8.29601347e-01 -4.12184656e-01 -6.20946944e-01 -1.01089311e+00
7.24680781e-01 5.89875877e-01 1.35082877e+00 8.39303285e-02
1.13952495e-01 9.26304162e-01 -4.55997884e-01 -1.03531027e+00
-8.32189023e-01 3.83733541e-01 1.76324874e-01 -3.45186144e-01
5.70036709e-01 -5.15978217e-01 -1.97509527e-01 -6.55073002e-02
-7.35371232e-01 4.97273117e-01 6.00750268e-01 5.77390492e-01
-2.64569819e-01 -6.60322070e-01 7.34658957e-01 -8.52811873e-01
1.00088573e+00 -6.74057901e-01 -2.98082322e-01 2.12256134e-01
-3.39339137e-01 1.87490851e-01 3.37579757e-01 -5.16144574e-01
-1.02133632e+00 -1.58702731e-01 1.00778364e-01 -1.41990378e-01
-4.04536426e-01 6.00335062e-01 1.72640428e-01 1.95084676e-01
1.27282262e+00 7.77611285e-02 1.94324717e-01 -1.13388814e-01
6.57513916e-01 1.21085167e+00 6.26816571e-01 -8.52411330e-01
7.32447386e-01 2.28220403e-01 -7.68253148e-01 -1.37945080e+00
-7.72842050e-01 -4.20650870e-01 -5.73989213e-01 -2.63930291e-01
8.19448888e-01 -1.20435727e+00 -6.52066767e-01 -1.03533775e-01
-9.51004863e-01 -8.46533060e-01 -5.91842413e-01 3.62445772e-01
-3.85665148e-01 -2.42468536e-01 -2.92040527e-01 -1.15695500e+00
-7.43546709e-02 -7.86975205e-01 1.00676024e+00 3.29226822e-01
-7.60828555e-01 -1.09981179e+00 3.13436180e-01 5.16122282e-01
5.21993518e-01 -2.04083607e-01 6.69210494e-01 -1.04640996e+00
-5.36414981e-01 2.35481173e-01 1.10637590e-01 3.02690923e-01
2.91946858e-01 -2.96409100e-01 -1.20849001e+00 -3.49446476e-01
-5.01211941e-01 -8.57927740e-01 4.33546156e-01 -1.48557778e-02
5.55301726e-01 -4.35398698e-01 -1.49643883e-01 2.78586984e-01
8.22866499e-01 3.22255313e-01 3.53873700e-01 2.95687884e-01
7.38914847e-01 7.20775008e-01 3.02976072e-01 3.58028799e-01
6.60353363e-01 2.63498217e-01 6.64161146e-02 -3.14190358e-01
-7.29181692e-02 -9.50274169e-01 6.69895768e-01 5.86714804e-01
4.61528629e-01 -2.58104354e-01 -1.41643929e+00 9.87849891e-01
-1.54178166e+00 -1.08156419e+00 6.21420503e-01 1.94370055e+00
8.93304825e-01 1.31821081e-01 2.06026092e-01 -1.21180877e-01
3.73583794e-01 5.20188212e-02 -5.84732950e-01 -6.36660993e-01
1.21536441e-01 3.19254071e-01 3.16804111e-01 6.66249990e-01
-7.62423992e-01 9.81336653e-01 7.22646523e+00 -3.18873078e-01
-1.23619485e+00 -2.55855381e-01 4.05785501e-01 -2.72837013e-01
-8.65094423e-01 -1.22945555e-01 -5.58362365e-01 -1.66773379e-01
1.20999384e+00 -4.94611830e-01 6.28998399e-01 9.05196667e-01
9.34802517e-02 1.87711492e-01 -1.62029588e+00 9.57684219e-01
3.21899593e-01 -1.38752973e+00 5.48972070e-01 -1.06974259e-01
3.86872381e-01 2.58674234e-01 2.97856152e-01 6.59093738e-01
9.64920044e-01 -1.48310530e+00 1.13179052e+00 2.73798883e-01
7.87777126e-01 -5.20866513e-01 1.62528157e-01 4.57855493e-01
-7.91221201e-01 -3.65189284e-01 1.26498670e-03 -5.04323125e-01
9.30886939e-02 -1.62980869e-01 -1.81953883e+00 -1.56436890e-01
6.02845252e-01 6.63870454e-01 -7.81112909e-01 4.62641716e-01
-1.03069320e-01 8.74465525e-01 -2.56668538e-01 -6.56627178e-01
2.47912675e-01 3.34041476e-01 2.96416610e-01 1.39592135e+00
1.20912269e-01 1.57170996e-01 4.86856252e-01 8.07627797e-01
-9.13996622e-02 1.64211243e-01 -1.39053893e+00 -7.72204578e-01
7.54870951e-01 7.74507523e-01 -5.89073658e-01 -3.15811962e-01
-5.16410172e-01 3.52626413e-01 4.03212488e-01 3.12442273e-01
-5.89292943e-02 -1.83017537e-01 7.60149240e-01 4.96467263e-01
-7.77846482e-03 -4.72403169e-01 -6.35645509e-01 -9.75815952e-01
-3.39588493e-01 -1.25321054e+00 2.76209503e-01 -9.61043835e-01
-1.15275776e+00 5.28263509e-01 2.52275765e-01 -7.93757439e-01
-5.21038055e-01 -6.19529128e-01 -3.56818318e-01 6.27677917e-01
-8.81617367e-01 -1.31716287e+00 -4.15952712e-01 6.16877139e-01
6.63882554e-01 -5.68912983e-01 8.62385750e-01 9.67410766e-03
-1.93250179e-01 7.67352939e-01 -4.75354433e-01 4.02666003e-01
5.43963969e-01 -1.22211564e+00 7.47137725e-01 7.85163045e-01
7.43087232e-01 1.28089786e+00 7.20851600e-01 -8.12072277e-01
-1.42488003e+00 -7.95069575e-01 7.45858729e-01 -8.64164889e-01
4.65793252e-01 -9.51071084e-01 -8.62579525e-01 1.13283002e+00
4.82618123e-01 -4.86332119e-01 9.80431855e-01 7.47654676e-01
-5.43691158e-01 7.07685649e-02 -1.02503133e+00 8.87770712e-01
1.30938041e+00 -1.19359684e+00 -9.34285402e-01 2.10042536e-01
1.13797677e+00 -2.02896923e-01 -5.26376903e-01 -7.17017874e-02
3.06295574e-01 -5.67607701e-01 9.72697496e-01 -9.50487494e-01
3.73518378e-01 -5.85915931e-02 -3.34773093e-01 -1.41158354e+00
-3.16362351e-01 -7.85382986e-01 1.94344059e-01 1.50107682e+00
5.67759275e-01 -3.76532495e-01 6.07502699e-01 8.05046260e-01
-2.58092105e-01 -2.52832975e-02 -5.29792905e-01 -4.56955880e-01
2.80318797e-01 -5.79263926e-01 6.42419636e-01 9.74129975e-01
5.56929767e-01 7.76037216e-01 -9.06725377e-02 3.37896556e-01
2.59185255e-01 -2.60518759e-01 1.00547421e+00 -1.17133284e+00
-1.03442863e-01 -1.92145869e-01 -4.11824524e-01 -1.18386626e+00
3.26763898e-01 -1.20194435e+00 3.78863215e-01 -1.74782658e+00
1.42909989e-01 -4.32303250e-01 -2.46085435e-01 8.02856684e-01
1.56740621e-01 -1.63218677e-01 3.23452055e-01 -2.51774698e-01
-6.74317956e-01 5.97708166e-01 8.24665308e-01 -3.10631603e-01
-4.48711842e-01 -3.48143250e-01 -1.44579399e+00 6.10815823e-01
7.62341619e-01 -3.63077939e-01 -1.02400100e+00 -7.87202299e-01
2.80228347e-01 -2.89632082e-01 4.63335276e-01 -1.22172260e+00
4.83162791e-01 -3.74439001e-01 2.66710103e-01 -4.12144750e-01
1.13321833e-01 -7.48929977e-01 -3.12756658e-01 1.34135321e-01
-1.02254653e+00 2.12257639e-01 6.21111155e-01 3.56893778e-01
-9.49876681e-02 -1.84574351e-01 2.99223274e-01 -1.49360284e-01
-8.77241313e-01 -6.54466152e-02 -6.50682390e-01 4.02084708e-01
7.82590270e-01 -1.30579352e-01 -5.29696703e-01 -8.82249534e-01
-4.44160879e-01 2.44137317e-01 4.39326733e-01 1.03215969e+00
8.59755576e-01 -1.22006810e+00 -6.07163846e-01 3.34653169e-01
6.00697875e-01 -4.40323316e-02 -1.20416671e-01 7.50924051e-02
-4.04320836e-01 1.48048058e-01 -3.10929060e-01 -6.40035272e-01
-1.18646848e+00 5.14467239e-01 1.87338352e-01 5.52418411e-01
-6.17567241e-01 8.19307625e-01 1.43197700e-01 -5.57017744e-01
5.91350555e-01 -5.62891483e-01 -2.10111126e-01 5.70643283e-02
7.30587304e-01 3.13303061e-02 -1.15339823e-01 -4.76923317e-01
-4.13104326e-01 -1.31784245e-01 -2.06144646e-01 -6.36104167e-01
1.53737903e+00 4.35894839e-02 3.58265549e-01 9.34047639e-01
9.33840513e-01 1.69907823e-01 -1.17462087e+00 -3.33549678e-01
5.36670722e-02 -2.33278751e-01 1.14384145e-01 -8.28518748e-01
-4.88128901e-01 1.14658916e+00 7.15014577e-01 1.01797476e-01
5.53917468e-01 3.59022081e-01 4.03898239e-01 9.82877374e-01
4.91835624e-01 -9.73095775e-01 5.27777076e-01 4.75990415e-01
9.95394588e-01 -1.22593677e+00 -9.35778543e-02 5.42285629e-02
-9.33432102e-01 8.72965097e-01 6.55593991e-01 1.79318279e-01
4.68867540e-01 4.83017772e-01 5.33343673e-01 -3.67972255e-01
-6.78071499e-01 -1.67603642e-01 1.17929585e-01 1.28651595e+00
7.21772015e-01 1.48711249e-01 7.53552020e-01 4.93371427e-01
-7.44036913e-01 3.81716453e-02 5.15495718e-01 1.18912113e+00
-6.29517496e-01 -7.62752950e-01 -2.57148564e-01 2.83761770e-01
1.27598524e-01 -4.42968905e-01 -8.19736362e-01 6.23595119e-01
-9.72175002e-02 1.26355040e+00 4.28236157e-01 -4.70439404e-01
6.12314165e-01 1.60971567e-01 3.08033705e-01 -1.26666689e+00
-8.49288464e-01 -7.57528424e-01 4.40583289e-01 -1.67466521e-01
-1.65703818e-01 -8.18782270e-01 -1.28494835e+00 -2.20437035e-01
3.37269530e-02 3.46358865e-02 4.94884789e-01 7.11368203e-01
4.11965251e-01 4.60522950e-01 2.83810407e-01 -7.23963261e-01
-5.81348240e-01 -1.00089121e+00 -1.46139920e-01 4.78641272e-01
7.84764051e-01 -4.38449353e-01 -1.14105187e-01 2.53855973e-01] | [9.850224494934082, 7.599193096160889] |
c46248c7-1e48-4270-8c0f-e25bf38f2b9f | robust-node-classification-on-graphs-jointly | 2208.09779 | null | https://arxiv.org/abs/2208.09779v1 | https://arxiv.org/pdf/2208.09779v1.pdf | Robust Node Classification on Graphs: Jointly from Bayesian Label Transition and Topology-based Label Propagation | Node classification using Graph Neural Networks (GNNs) has been widely applied in various real-world scenarios. However, in recent years, compelling evidence emerges that the performance of GNN-based node classification may deteriorate substantially by topological perturbation, such as random connections or adversarial attacks. Various solutions, such as topological denoising methods and mechanism design methods, have been proposed to develop robust GNN-based node classifiers but none of these works can fully address the problems related to topological perturbations. Recently, the Bayesian label transition model is proposed to tackle this issue but its slow convergence may lead to inferior performance. In this work, we propose a new label inference model, namely LInDT, which integrates both Bayesian label transition and topology-based label propagation for improving the robustness of GNNs against topological perturbations. LInDT is superior to existing label transition methods as it improves the label prediction of uncertain nodes by utilizing neighborhood-based label propagation leading to better convergence of label inference. Besides, LIndT adopts asymmetric Dirichlet distribution as a prior, which also helps it to improve label inference. Extensive experiments on five graph datasets demonstrate the superiority of LInDT for GNN-based node classification under three scenarios of topological perturbations. | ['Mohammad Al Hasan', 'Jun Zhuang'] | 2022-08-21 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [-4.33611013e-02 1.02197260e-01 -2.08763540e-01 -3.19222063e-01
-1.10900251e-03 -3.77768040e-01 6.11866832e-01 3.89429748e-01
1.31326178e-02 7.22701907e-01 -2.50021726e-01 -4.11571622e-01
-3.31192374e-01 -1.33482730e+00 -5.43173134e-01 -9.45527375e-01
-1.35384902e-01 3.86345655e-01 6.56152129e-01 -2.57000417e-01
3.47127616e-01 5.10413647e-01 -1.02447116e+00 -4.50139165e-01
9.87058520e-01 9.40566897e-01 -3.85933429e-01 2.45077059e-01
-2.50191242e-01 3.38342816e-01 -5.91810405e-01 -4.17261273e-01
3.01194757e-01 1.78842187e-01 -6.51593089e-01 -4.94161040e-01
2.49531224e-01 -8.21862146e-02 -5.87559998e-01 1.35079873e+00
5.47895730e-01 1.13866769e-01 8.68153691e-01 -1.51579392e+00
-7.39029348e-01 9.39543903e-01 -7.03034222e-01 4.03025299e-02
-1.72907665e-01 -7.09939748e-02 6.75517201e-01 -4.32062179e-01
4.93414521e-01 1.49704337e+00 1.30802929e+00 3.92576486e-01
-1.23354053e+00 -1.02778161e+00 2.27186292e-01 2.24341810e-01
-1.46720898e+00 -4.40699272e-02 1.13893247e+00 -1.26007736e-01
3.38883758e-01 -8.42526704e-02 3.47120672e-01 1.24850118e+00
2.03652665e-01 4.71795678e-01 8.79016578e-01 -7.55501091e-02
3.03566068e-01 -1.27346799e-01 -9.99493003e-02 5.47303081e-01
4.86399502e-01 -3.99697758e-02 -2.54243106e-01 -3.03069472e-01
8.00293744e-01 -3.31900418e-02 -1.73664868e-01 -2.94241488e-01
-8.52837801e-01 6.49165988e-01 9.39381838e-01 2.91699767e-01
-2.26974890e-01 2.51634657e-01 7.26128399e-01 1.91186234e-01
7.47629702e-01 5.77632003e-02 -2.61213094e-01 3.81477714e-01
-7.80519366e-01 -2.17920035e-01 7.44610190e-01 6.69237554e-01
5.76226950e-01 1.39558822e-01 -8.34161714e-02 7.36111104e-01
6.98834002e-01 5.61411917e-01 1.68677583e-01 -8.13943565e-01
2.45208338e-01 7.40981817e-01 -3.18591774e-01 -1.82077897e+00
-6.90488696e-01 -7.70908475e-01 -1.58086526e+00 2.39997804e-02
3.52378249e-01 1.34141594e-02 -1.05358112e+00 2.03286982e+00
6.58228576e-01 7.63089657e-01 -8.09700340e-02 6.18746579e-01
1.25571656e+00 5.16078413e-01 2.58226931e-01 7.24214241e-02
8.20493579e-01 -5.36512673e-01 -5.84533930e-01 1.68499902e-01
7.49922812e-01 -4.60583091e-01 3.44233632e-01 1.66433603e-01
-5.79800606e-01 -3.24229389e-01 -1.04939997e+00 4.01854336e-01
-5.55286646e-01 -3.18747103e-01 6.77048445e-01 8.75258505e-01
-1.20805085e+00 5.97346008e-01 -8.59972298e-01 -6.28965914e-01
5.21495461e-01 4.27066356e-01 -2.90438950e-01 -5.39788380e-02
-1.62242758e+00 6.24911129e-01 5.48767149e-01 3.88841808e-01
-7.71731794e-01 -5.28964579e-01 -7.69585609e-01 -8.88418779e-02
3.37606490e-01 -3.99405032e-01 6.97661221e-01 -2.27268845e-01
-1.38095570e+00 3.02653909e-01 1.84510469e-01 -2.08468989e-01
4.34262663e-01 7.45678902e-01 -3.61303210e-01 2.37471536e-01
8.33427683e-02 6.96735024e-01 6.67651355e-01 -1.37981582e+00
-1.99837953e-01 -2.99665034e-01 1.08845875e-01 1.01191968e-01
-5.87162912e-01 -4.95237261e-01 -2.43780479e-01 -8.22478473e-01
8.36019218e-01 -9.40764368e-01 -2.00869471e-01 1.30697489e-01
-6.87273145e-01 -5.84899902e-01 1.09224129e+00 -4.75688756e-01
9.57088470e-01 -1.87922859e+00 -1.13923192e-01 7.78349340e-01
4.96042281e-01 3.85420799e-01 -1.85672224e-01 3.65153491e-01
1.28385141e-01 5.32947958e-01 -1.84339061e-01 -2.52735376e-01
6.33776234e-03 2.78184295e-01 -2.79849134e-02 6.84971869e-01
-1.82060540e-01 8.84865463e-01 -8.91486406e-01 -6.07055306e-01
1.86089933e-01 6.47497535e-01 -1.47919402e-01 -3.66991103e-01
-8.72750208e-02 4.85618263e-01 -5.71663201e-01 8.67580056e-01
1.21070147e+00 -2.16364756e-01 2.04318419e-01 -2.41870672e-01
4.51851845e-01 -1.62047997e-01 -1.09841788e+00 1.13867819e+00
-6.03385754e-02 4.71569419e-01 2.79677212e-01 -1.40439248e+00
1.30504668e+00 3.00479412e-01 4.47634220e-01 -4.63255256e-01
3.04368705e-01 2.70565245e-02 -2.56191105e-01 -2.97217220e-01
1.49332270e-01 -4.55560312e-02 -1.28259212e-01 4.23874408e-01
-1.61405608e-01 2.56609619e-01 -9.51741189e-02 4.52577949e-01
1.39767945e+00 -2.92454243e-01 -2.98378974e-01 -2.71892399e-01
5.41394114e-01 -4.52564597e-01 9.95968759e-01 9.06487286e-01
-5.22797763e-01 3.00135642e-01 4.82207119e-01 -5.01772344e-01
-6.47428632e-01 -1.13552642e+00 -2.88912326e-01 9.25527871e-01
7.76413381e-01 -1.24317706e-01 -8.04582775e-01 -8.81541908e-01
1.64546028e-01 2.66180903e-01 -6.64109170e-01 -5.70495307e-01
-3.25283140e-01 -1.11667120e+00 1.32763016e+00 4.34985131e-01
1.11998141e+00 -9.24072146e-01 2.94447273e-01 1.60640925e-01
-3.71092439e-01 -9.08824980e-01 5.24428077e-02 2.78426278e-02
-1.06895459e+00 -1.16121900e+00 -3.85671467e-01 -7.75613010e-01
8.83573472e-01 2.32852235e-01 8.63572121e-01 6.45297587e-01
6.78469762e-02 7.79758394e-02 -4.34344798e-01 2.44965800e-03
-4.80561823e-01 4.21261996e-01 4.10659730e-01 -4.85198274e-02
2.25394860e-01 -9.59116876e-01 -6.47223771e-01 7.28798270e-01
-1.03322160e+00 -3.51229936e-01 5.82669199e-01 6.16666317e-01
1.26021400e-01 7.32461929e-01 9.60896790e-01 -8.57064426e-01
7.31899202e-01 -8.24963570e-01 -3.19611222e-01 3.42126846e-01
-9.11643267e-01 -1.46656856e-03 6.56972289e-01 -5.11131346e-01
-1.05242944e+00 -6.10907614e-01 -1.59249425e-01 -1.90800294e-01
-2.86147892e-01 7.66563833e-01 -1.44384503e-01 -6.81742132e-01
6.29426539e-01 -6.10005036e-02 5.21752760e-02 -3.79514337e-01
5.06625772e-02 5.57727098e-01 3.76535237e-01 -6.43793881e-01
1.20274258e+00 5.38302481e-01 6.14503086e-01 -6.62370265e-01
-3.97380888e-01 -6.23848625e-02 -4.59894508e-01 -4.74046767e-01
4.34860259e-01 -5.58323562e-01 -9.46079612e-01 9.30993319e-01
-9.41353440e-01 -1.63456291e-01 2.94405222e-01 2.11977199e-01
2.07960308e-02 7.19693005e-01 -9.50139701e-01 -6.13071561e-01
-4.37933981e-01 -1.00526595e+00 5.61749935e-01 4.41825718e-01
2.53615081e-01 -1.40044916e+00 -1.73072800e-01 2.64586002e-01
7.03594685e-01 4.65271056e-01 1.05306160e+00 -7.63882816e-01
-5.69304764e-01 -2.40408778e-01 -5.10541081e-01 2.30337754e-01
5.98568730e-02 1.74216628e-01 -6.74528778e-01 -4.50875491e-01
-2.91276783e-01 -1.52581021e-01 8.44404280e-01 4.91638362e-01
9.40497279e-01 -6.26790822e-02 -7.07517743e-01 6.79237485e-01
1.29371464e+00 -1.01015314e-01 6.13134921e-01 4.01067466e-01
9.22464609e-01 6.14042163e-01 3.48094553e-01 8.62559006e-02
4.98506963e-01 3.05165976e-01 8.21687818e-01 -8.56200829e-02
-9.45411697e-02 -2.18137562e-01 -2.68342551e-02 7.86346316e-01
2.05410376e-01 -5.36596835e-01 -1.08069682e+00 3.07540327e-01
-1.98139834e+00 -6.89361811e-01 -3.41485292e-01 1.76963186e+00
5.76684952e-01 3.79031330e-01 -1.77633807e-01 3.66352648e-01
1.28505409e+00 2.12935612e-01 -8.01272273e-01 1.09620551e-02
-9.40875113e-02 -2.81039983e-01 5.70915461e-01 1.94841295e-01
-1.08210409e+00 8.05253386e-01 6.05435324e+00 1.09196043e+00
-1.07851589e+00 8.39352235e-02 7.12968290e-01 6.40856624e-01
-2.18948945e-01 -3.16375941e-02 -6.10555410e-01 6.19020700e-01
5.20650744e-01 1.61580592e-01 1.97028652e-01 6.54416680e-01
2.39543557e-01 6.49130121e-02 -4.42184895e-01 6.49831951e-01
-4.41395976e-02 -1.08375013e+00 6.32444173e-02 1.74972042e-01
7.74567127e-01 -3.91687639e-02 1.34610552e-02 3.13842773e-01
4.81773347e-01 -8.40462089e-01 2.45630279e-01 5.23885250e-01
5.86738825e-01 -7.98011720e-01 1.14686370e+00 3.91946614e-01
-1.34356260e+00 4.23712023e-02 -4.23526883e-01 1.24599017e-01
2.05402281e-02 9.45177019e-01 -7.88307846e-01 7.07518280e-01
7.98823297e-01 7.88209975e-01 -9.22656953e-01 1.25029683e+00
-2.55848080e-01 8.59005511e-01 -8.41038406e-01 -1.16045430e-01
2.21450061e-01 -1.74498051e-01 7.51371801e-01 7.63890505e-01
2.24512249e-01 -1.98409513e-01 3.95620257e-01 8.15178394e-01
-3.54574800e-01 -1.04705714e-01 -6.07351482e-01 4.59526718e-01
9.23175335e-01 1.37132215e+00 -1.34877944e+00 -3.54775600e-02
1.54214085e-03 3.24197024e-01 2.38611206e-01 6.31267548e-01
-7.65239477e-01 -5.88098347e-01 2.47243389e-01 1.81109667e-01
4.52712290e-02 -2.85555106e-02 -3.74455184e-01 -7.46803641e-01
-4.32733148e-02 -3.99639457e-01 3.20914298e-01 -5.03541231e-01
-1.52725744e+00 3.07472348e-01 9.16251540e-02 -1.01757371e+00
4.35800821e-01 -3.52248907e-01 -1.04453492e+00 4.79290217e-01
-1.32011557e+00 -1.31504107e+00 -5.39415359e-01 2.53745139e-01
1.64542086e-02 -1.04152553e-01 3.92000824e-01 3.50870103e-01
-8.30135345e-01 7.95259833e-01 2.67632186e-01 3.56494039e-01
7.83253133e-01 -1.14432764e+00 2.89243966e-01 8.80868375e-01
-4.54207659e-01 6.22647285e-01 7.35309482e-01 -1.13185954e+00
-9.14459348e-01 -1.39510822e+00 3.27998787e-01 -1.41907513e-01
4.85886842e-01 -1.59932017e-01 -1.10879576e+00 4.21531141e-01
-1.99991599e-01 2.92251706e-01 3.74362648e-01 1.85061857e-01
-3.97971272e-01 -3.07275862e-01 -1.63269460e+00 5.02957225e-01
1.02575481e+00 -3.91234398e-01 -1.39782324e-01 1.85554340e-01
6.01520360e-01 -1.30548924e-01 -1.17550719e+00 9.92671907e-01
2.42561370e-01 -8.27610373e-01 1.01343989e+00 1.01621069e-01
-1.72981381e-01 -5.98722219e-01 1.90927491e-01 -1.29692161e+00
-4.80401158e-01 -2.67088890e-01 -7.75340572e-02 1.64363551e+00
2.59845078e-01 -1.05170870e+00 1.14887273e+00 3.02454412e-01
3.78328785e-02 -6.42543912e-01 -1.02181065e+00 -7.06651628e-01
7.85952806e-02 -3.37106913e-01 7.08260179e-01 1.11062145e+00
-2.48129770e-01 9.21271592e-02 -1.42071769e-01 3.77611727e-01
1.06360245e+00 -5.74130595e-01 5.08740187e-01 -1.94535124e+00
4.18040067e-01 -5.08258164e-01 -7.73660541e-01 -8.08410585e-01
3.88468802e-01 -1.08394003e+00 1.32295877e-01 -1.67816818e+00
-7.87983984e-02 -1.11014462e+00 -4.82429296e-01 4.87981886e-01
-6.19665757e-02 5.55643141e-01 -3.08845907e-01 3.24022353e-01
-7.88157344e-01 5.66660702e-01 9.93547678e-01 -4.22329277e-01
-8.45545530e-03 4.47972864e-02 -5.47259688e-01 8.47622216e-01
1.05169570e+00 -6.85311496e-01 -4.64431912e-01 -1.61362246e-01
4.80452120e-01 -2.32472837e-01 4.51998889e-01 -1.19215441e+00
5.28457940e-01 1.75881311e-01 4.38177675e-01 -8.18158686e-01
6.21503443e-02 -7.61811554e-01 2.25896403e-01 3.46172273e-01
-1.29205972e-01 -1.44118190e-01 6.35332381e-03 1.10205960e+00
1.94786377e-02 -2.49275237e-01 8.53946745e-01 9.95847583e-02
-4.01454538e-01 5.23266971e-01 -4.00447011e-01 -1.13268942e-01
1.01661491e+00 -3.48390877e-01 -5.88294446e-01 -4.23693329e-01
-5.51827371e-01 5.84833622e-01 4.77536857e-01 1.83920532e-01
4.99642789e-01 -1.53940225e+00 -5.04170775e-01 1.04370423e-01
-1.05283253e-01 1.71900481e-01 3.76195461e-01 9.89419758e-01
-5.69780052e-01 9.23943520e-02 6.40799999e-02 -7.31312394e-01
-1.07299423e+00 3.30252945e-01 2.85457820e-01 -2.49760240e-01
-4.57472086e-01 8.22585821e-01 -2.69289434e-01 -1.03534532e+00
6.13139570e-01 -2.78599299e-02 -3.26420814e-01 -1.40438229e-01
-8.91873538e-02 7.11583555e-01 5.90263531e-02 -6.26433372e-01
-3.32011253e-01 6.21384382e-01 -6.82721585e-02 3.04869741e-01
1.07335818e+00 -3.84542257e-01 -5.09261727e-01 1.45588532e-01
8.71160507e-01 -5.29868722e-01 -8.51949751e-01 -4.27653134e-01
6.39514253e-03 -4.49235328e-02 2.69144803e-01 -6.17482901e-01
-1.34793591e+00 6.29053056e-01 5.35879493e-01 6.54526412e-01
9.62653577e-01 -2.54607975e-01 9.24670815e-01 5.76478302e-01
4.88255739e-01 -1.04405046e+00 1.01500824e-01 4.53674614e-01
2.88022637e-01 -1.27947438e+00 -9.18823481e-02 -5.73310792e-01
2.40232632e-01 1.06022549e+00 8.12539399e-01 -3.83591689e-02
9.84127641e-01 4.64333631e-02 -5.60894795e-03 -2.07098708e-01
-3.73583198e-01 2.70302147e-01 -3.65630426e-02 7.33736277e-01
-1.39100641e-01 -7.62742013e-02 -8.66136402e-02 1.52343139e-01
1.53449982e-01 -3.62922072e-01 3.01555902e-01 7.59043336e-01
-4.64979321e-01 -1.08797801e+00 -5.89793563e-01 7.04812527e-01
-3.51723701e-01 5.81958592e-02 -3.01141024e-01 5.84078670e-01
1.33571863e-01 1.12069988e+00 -2.97224820e-01 -7.61910856e-01
-1.10403895e-01 -8.83552581e-02 2.67869700e-03 -2.12074965e-01
-3.36792618e-01 -3.47192556e-01 -1.40348926e-01 -1.41896829e-01
-4.99001533e-01 -4.00904238e-01 -1.15244126e+00 -7.88800895e-01
-6.72129393e-01 3.38583797e-01 7.28409886e-01 1.00423849e+00
3.15148145e-01 3.48363370e-01 5.16024649e-01 -8.20484221e-01
-3.98601055e-01 -1.19909823e+00 -5.83499610e-01 3.66126448e-01
8.64340439e-02 -1.03709316e+00 -7.54082263e-01 -5.84858596e-01] | [7.142207622528076, 6.065285682678223] |
185805da-cba8-4ce9-b32e-e549195dcf4f | rf-clust-for-leave-one-problem-out | 2301.09524 | null | https://arxiv.org/abs/2301.09524v2 | https://arxiv.org/pdf/2301.09524v2.pdf | RF+clust for Leave-One-Problem-Out Performance Prediction | Per-instance automated algorithm configuration and selection are gaining significant moments in evolutionary computation in recent years. Two crucial, sometimes implicit, ingredients for these automated machine learning (AutoML) methods are 1) feature-based representations of the problem instances and 2) performance prediction methods that take the features as input to estimate how well a specific algorithm instance will perform on a given problem instance. Non-surprisingly, common machine learning models fail to make predictions for instances whose feature-based representation is underrepresented or not covered in the training data, resulting in poor generalization ability of the models for problems not seen during training.In this work, we study leave-one-problem-out (LOPO) performance prediction. We analyze whether standard random forest (RF) model predictions can be improved by calibrating them with a weighted average of performance values obtained by the algorithm on problem instances that are sufficiently close to the problem for which a performance prediction is sought, measured by cosine similarity in feature space. While our RF+clust approach obtains more accurate performance prediction for several problems, its predictive power crucially depends on the chosen similarity threshold as well as on the feature portfolio for which the cosine similarity is measured, thereby opening a new angle for feature selection in a zero-shot learning setting, as LOPO is termed in machine learning. | ['Tome Eftimov', 'Carola Doerr', 'Ana Nikolikj'] | 2023-01-23 | null | null | null | null | ['automl'] | ['methodology'] | [ 4.87321049e-01 -2.86384195e-01 -2.76721567e-01 -3.70815575e-01
-6.83943748e-01 -4.66516793e-01 4.79146272e-01 6.71241403e-01
-3.27984184e-01 7.70289063e-01 -2.75106996e-01 1.41828638e-02
-8.92405212e-01 -9.83602107e-01 -3.54966223e-01 -9.12664473e-01
1.35908753e-03 8.27044904e-01 1.22190908e-01 -6.66506439e-02
6.04700387e-01 4.98670995e-01 -2.08471274e+00 1.87443897e-01
9.96037781e-01 9.79719341e-01 1.71841830e-02 9.03927803e-01
-8.00987184e-02 5.34233928e-01 -8.22439253e-01 -2.71812618e-01
2.26886407e-01 -6.18862987e-01 -7.76916385e-01 -4.60508354e-02
1.13955297e-01 5.15416443e-01 2.55661130e-01 8.34033370e-01
3.54543179e-01 2.81883150e-01 9.44130540e-01 -1.28007996e+00
-1.31464407e-01 4.19803143e-01 -3.89624894e-01 1.64134055e-01
4.16117758e-01 1.94509268e-01 1.18626404e+00 -5.95190704e-01
6.66053653e-01 6.42763197e-01 8.19154143e-01 1.44756153e-01
-1.42381656e+00 -3.55437160e-01 -2.85046726e-01 4.03609335e-01
-1.41838694e+00 -3.10087860e-01 5.73230505e-01 -5.84513009e-01
1.03453851e+00 6.06625736e-01 8.56464803e-01 5.85213184e-01
2.36426473e-01 4.14664537e-01 9.80634749e-01 -7.57050574e-01
7.43976772e-01 3.63433331e-01 2.66523242e-01 4.46636945e-01
2.96027243e-01 3.56519550e-01 -5.78832150e-01 -3.64160597e-01
9.68835354e-02 -4.45684232e-02 -3.62152874e-01 -5.86225092e-01
-9.50635195e-01 7.85051703e-01 2.59425253e-01 7.50750303e-01
-3.81190240e-01 -1.53670028e-01 3.10342342e-01 6.61149263e-01
3.09389919e-01 1.39269781e+00 -6.98160946e-01 -4.20109093e-01
-1.29894161e+00 3.23062897e-01 8.64751041e-01 5.67615330e-01
1.04470885e+00 -2.83028007e-01 -2.58790702e-01 8.10698092e-01
-3.64898860e-01 -2.65860539e-02 9.84493136e-01 -5.31524777e-01
-3.92130762e-02 1.09789288e+00 -5.14278188e-02 -7.77396441e-01
-3.64452660e-01 -7.05252171e-01 -3.77844423e-01 4.03483868e-01
5.87351918e-01 1.14061989e-01 -7.03036189e-01 1.45621955e+00
2.71009773e-01 -6.69218823e-02 3.14139691e-03 6.23379767e-01
1.97923273e-01 5.23811460e-01 -1.05031259e-01 -7.74624705e-01
9.89267528e-01 -7.35339582e-01 -2.06522271e-01 -4.94262576e-02
1.15088952e+00 -6.49196148e-01 1.05925941e+00 6.37716174e-01
-6.32185459e-01 -4.02191401e-01 -1.16462624e+00 6.77870095e-01
-5.48780501e-01 -1.70895919e-01 5.29970825e-01 7.82645702e-01
-5.62663436e-01 1.19230473e+00 -5.19816160e-01 -3.59081775e-01
2.77185619e-01 5.31526744e-01 -2.48345658e-01 9.26284194e-02
-8.51598024e-01 9.68181193e-01 5.90339720e-01 -5.15500247e-01
-6.40921712e-01 -8.76123369e-01 -4.11361337e-01 1.90213740e-01
7.02780008e-01 -4.97046739e-01 1.02367175e+00 -1.45712256e+00
-1.26696634e+00 8.80813301e-01 1.05247803e-01 -5.24280488e-01
5.50538003e-01 1.63975749e-02 -5.07327914e-01 -3.34383696e-01
-1.66181341e-01 3.04713100e-02 9.41920817e-01 -9.19744790e-01
-7.25670993e-01 -5.95720947e-01 -8.13583210e-02 2.78596045e-03
-1.81951150e-01 -8.51240158e-02 4.41945307e-02 -4.10425097e-01
1.06427008e-02 -8.72129023e-01 -4.14123118e-01 -2.10873127e-01
-3.46377343e-01 -3.40807945e-01 5.95738649e-01 -3.35920714e-02
1.56934428e+00 -1.83633316e+00 2.58426845e-01 4.73254651e-01
-9.79189426e-02 4.75129336e-01 8.99817701e-03 6.27984047e-01
-1.47553101e-01 4.30654772e-02 -3.57930720e-01 4.07901764e-01
-2.66424417e-01 -7.77505413e-02 -1.23836838e-01 4.45630759e-01
2.72873849e-01 5.74988604e-01 -7.63799965e-01 -3.86086613e-01
2.04524219e-01 2.02636700e-02 -5.67778111e-01 4.14587557e-01
-2.79868156e-01 2.94721186e-01 -5.24363816e-01 5.00494182e-01
1.84054762e-01 -2.77965426e-01 1.05090290e-01 1.74308226e-01
-1.58205584e-01 -1.26578271e-01 -1.19360399e+00 1.40683722e+00
-5.39166510e-01 4.45886731e-01 -6.33641660e-01 -1.21505070e+00
1.13561893e+00 7.44341761e-02 5.74157774e-01 -4.88824725e-01
1.40329614e-01 3.74058634e-01 3.20114076e-01 -4.34467882e-01
2.26073176e-01 -2.56679267e-01 -1.55821830e-01 6.03388011e-01
2.53579527e-01 -2.35794067e-01 4.57429469e-01 -3.46019924e-01
1.38680804e+00 8.49902071e-03 1.07839286e+00 -2.74731249e-01
6.03590310e-01 3.69395286e-01 6.65551126e-01 7.97163785e-01
-1.23644017e-01 5.26871562e-01 5.47596395e-01 -6.97243750e-01
-1.08580673e+00 -5.63767970e-01 -6.00035131e-01 1.25457954e+00
3.22584510e-02 -5.63435912e-01 -6.60690010e-01 -8.52230728e-01
4.55359817e-02 9.72696960e-01 -8.30923080e-01 -5.35656750e-01
-2.72073388e-01 -8.98486078e-01 1.98459387e-01 2.57973164e-01
-1.38601691e-01 -1.09555137e+00 -1.07380641e+00 4.30501997e-01
5.56933463e-01 -3.67005080e-01 1.72706515e-01 7.10978985e-01
-8.15289319e-01 -1.43553305e+00 -5.17142534e-01 -3.30693096e-01
4.55020428e-01 -6.06764965e-02 1.26926064e+00 3.50547314e-01
-6.76319599e-01 7.99704641e-02 -6.29353881e-01 -3.89044821e-01
-4.14682686e-01 1.74171627e-01 1.58327937e-01 -2.41704360e-02
5.32279253e-01 -5.14089048e-01 -2.50657052e-01 5.25169134e-01
-5.42528927e-01 -1.42314956e-01 6.71306074e-01 1.14981902e+00
6.20294511e-01 3.60751748e-01 6.01821482e-01 -9.93493855e-01
4.99186903e-01 -3.67235303e-01 -4.95406806e-01 8.35072577e-01
-1.17061281e+00 4.59280968e-01 6.76287115e-01 -4.96022195e-01
-6.09391212e-01 1.37872323e-01 1.32832646e-01 -4.59993541e-01
-3.20680797e-01 4.48431730e-01 -2.14913972e-02 -3.83887403e-02
1.09360373e+00 1.50966704e-01 -7.94613436e-02 -3.44930828e-01
-5.09831272e-02 5.44499755e-01 1.15929492e-01 -6.18878663e-01
6.40980005e-01 -1.99977085e-01 1.59444869e-01 -5.95952451e-01
-8.79912734e-01 -5.29638231e-01 -8.52434576e-01 -2.10369512e-01
5.29111922e-01 -2.60886461e-01 -6.65822089e-01 1.29527852e-01
-4.92672324e-01 -9.76435766e-02 -8.33292663e-01 4.28404421e-01
-8.69139731e-01 -1.82107642e-01 3.91555130e-01 -1.12361932e+00
-3.01208079e-01 -1.26658440e+00 7.37870693e-01 2.79799312e-01
-7.33274817e-01 -8.68489504e-01 4.56449807e-01 4.22979668e-02
3.25572222e-01 2.98611164e-01 1.28547323e+00 -1.06633270e+00
-9.72456038e-02 -6.83651268e-01 2.28041217e-01 1.97987512e-01
-5.98624274e-02 3.79767805e-01 -9.70321178e-01 -3.58587742e-01
2.17873580e-03 -3.32615644e-01 8.15009117e-01 3.16832691e-01
1.20388103e+00 -1.29449591e-01 -5.09317458e-01 6.03908896e-01
1.63296270e+00 1.41068563e-01 3.54255438e-01 4.82244521e-01
1.90724224e-01 6.74244940e-01 1.06740344e+00 5.16226709e-01
-6.18918657e-01 9.51143265e-01 2.68018097e-01 4.21005338e-01
2.52834380e-01 -1.25575349e-01 4.18542884e-02 3.27829063e-01
-2.52810359e-01 -1.35281086e-01 -1.10882080e+00 2.38786548e-01
-1.84786940e+00 -9.92967188e-01 2.27194186e-02 2.68587112e+00
7.32361555e-01 2.76799649e-01 3.22539002e-01 7.25582361e-01
7.72700667e-01 -6.23463877e-02 -7.37581789e-01 -5.69761217e-01
2.64247041e-02 3.36662054e-01 9.63653624e-02 6.26824275e-02
-9.23734367e-01 5.06818593e-01 5.81034088e+00 1.17011571e+00
-9.86743808e-01 -1.39285937e-01 6.30678117e-01 -2.24281073e-01
1.08217150e-02 9.84723344e-02 -5.16901135e-01 4.26524818e-01
1.10767972e+00 -7.18670428e-01 3.56366098e-01 1.14001369e+00
-1.26222655e-01 -2.26150259e-01 -1.40063620e+00 8.51257324e-01
-1.49380439e-03 -1.25642705e+00 -1.17504969e-01 1.84088096e-01
7.06152737e-01 -4.22697932e-01 7.81342573e-03 4.18440700e-01
1.40779270e-02 -1.20970201e+00 5.18952072e-01 4.96589959e-01
7.32571483e-01 -1.02478254e+00 7.27454007e-01 7.18792915e-01
-7.59054840e-01 -4.91271406e-01 -4.73918945e-01 -2.65409164e-02
-3.22229981e-01 6.97686374e-01 -9.94277120e-01 3.94447535e-01
5.30726433e-01 5.69973111e-01 -7.31044948e-01 1.38100004e+00
3.55705582e-02 5.49496114e-01 -5.32445014e-02 -3.24770480e-01
4.07906100e-02 -1.27796173e-01 7.22472727e-01 7.77204335e-01
5.50055861e-01 -9.65519547e-02 -4.37627546e-02 7.66716599e-01
3.05302620e-01 4.45880353e-01 -5.90044141e-01 -2.61010200e-01
3.98576230e-01 1.15037024e+00 -6.10569835e-01 -1.09601453e-01
8.00130237e-03 5.99883318e-01 3.69950563e-01 -7.58719146e-02
-5.02001464e-01 -4.58404809e-01 5.62788010e-01 2.07129076e-01
1.21395312e-01 2.13654518e-01 -3.91465932e-01 -9.36893463e-01
-1.67968988e-01 -8.07594895e-01 4.70878541e-01 -3.81469548e-01
-1.39547217e+00 5.72633982e-01 -3.56491566e-01 -1.67665887e+00
-4.15968060e-01 -6.14651203e-01 -7.88542986e-01 5.62905312e-01
-8.70152533e-01 -6.52513385e-01 -2.23559961e-01 2.19463527e-01
5.69567025e-01 -4.92093682e-01 1.17561662e+00 -2.25940034e-01
-5.28231204e-01 6.56762123e-01 3.64900470e-01 -4.19239521e-01
6.24184608e-01 -1.31130707e+00 -1.25336617e-01 3.89609933e-01
5.99538386e-01 3.46089333e-01 1.02757657e+00 -4.21286374e-01
-1.34167111e+00 -8.80965829e-01 7.60404766e-01 -4.91903245e-01
4.47095603e-01 1.09723039e-01 -1.11931801e+00 1.31718591e-01
-4.45153356e-01 1.27011016e-01 1.12756419e+00 5.91172993e-01
-3.29948515e-01 -6.72515035e-02 -1.38805628e+00 1.60284817e-01
8.28223288e-01 -2.03661188e-01 -6.12880707e-01 4.02455002e-01
3.04363132e-01 8.77631009e-02 -9.76265669e-01 4.18517560e-01
6.59504175e-01 -1.21434712e+00 6.45310819e-01 -1.01263285e+00
5.10016322e-01 -7.70821273e-02 -2.54579186e-01 -1.51743352e+00
-3.84775490e-01 -3.59955698e-01 -8.30615237e-02 1.08580112e+00
6.36132061e-01 -4.51327562e-01 8.98304999e-01 5.42545974e-01
2.23314241e-01 -1.29481697e+00 -9.37906742e-01 -9.02792931e-01
-3.73353884e-02 -5.69245100e-01 7.42787302e-01 9.81499732e-01
1.73044369e-01 1.99859351e-01 -9.53225195e-02 -2.81775326e-01
4.66149539e-01 4.97539639e-01 8.44150901e-01 -1.66167188e+00
-5.62783659e-01 -7.29718208e-01 -9.10646975e-01 -1.27520204e-01
-4.50504012e-02 -8.59202087e-01 -9.60178599e-02 -8.61973941e-01
2.56261706e-01 -6.86988831e-01 -4.62887049e-01 2.98449367e-01
-1.88833639e-01 -1.12281591e-02 4.92054364e-03 3.43543828e-01
-6.10186398e-01 2.93412447e-01 8.70775402e-01 -6.02706987e-03
-5.30144691e-01 5.01802504e-01 -4.17012364e-01 5.99763215e-01
5.10819495e-01 -6.63912475e-01 -1.30737409e-01 3.62937391e-01
3.83693874e-01 1.19013540e-01 2.00489387e-02 -1.40578723e+00
1.26592815e-01 -5.07246792e-01 5.37136912e-01 -5.84342405e-02
1.17171168e-01 -8.80032837e-01 4.43180442e-01 6.42399490e-01
-7.00429738e-01 -4.29952033e-02 -2.77152538e-01 6.95023060e-01
-1.23185158e-01 -8.00034821e-01 9.10849690e-01 -9.22258422e-02
-8.14610422e-01 1.31006375e-01 -1.14485033e-01 -6.28205091e-02
1.46352971e+00 -5.74203432e-01 -1.07890099e-01 4.36858609e-02
-7.84706116e-01 -9.83729810e-02 9.02974725e-01 2.28914082e-01
4.42535967e-01 -9.43974793e-01 -6.25647783e-01 3.66932273e-01
7.04340935e-01 -4.09958333e-01 7.58896768e-02 8.79021645e-01
-3.01177502e-01 2.81272024e-01 -2.25320712e-01 -5.75005054e-01
-1.30531371e+00 7.40922749e-01 3.43250215e-01 -4.89505410e-01
-5.30769229e-01 9.74665880e-01 -3.35714191e-01 -2.39403233e-01
-6.75616488e-02 1.94767609e-01 -1.05212763e-01 2.52725691e-01
5.50014436e-01 5.61478674e-01 4.13576245e-01 -5.12099087e-01
-3.67585450e-01 6.21673107e-01 -1.69618279e-01 1.69773802e-01
1.53565538e+00 4.95867401e-01 1.03018790e-01 8.76013696e-01
1.04303074e+00 -1.77350119e-01 -8.93210948e-01 -2.53427058e-01
3.28645736e-01 -7.81480074e-01 1.75819457e-01 -1.01043773e+00
-7.52672493e-01 8.52327049e-01 6.90288782e-01 4.99724835e-01
1.12213624e+00 -2.13118643e-01 2.25823611e-01 4.92111444e-01
7.58536994e-01 -1.29279697e+00 -1.32117972e-01 2.42031723e-01
8.43191504e-01 -1.19603682e+00 2.44326741e-01 -9.80442241e-02
-6.84931755e-01 1.33027756e+00 5.58257461e-01 -1.94401979e-01
5.92929959e-01 1.57301471e-01 -3.52309674e-01 -1.68726519e-01
-1.13380361e+00 -2.17853457e-01 4.95408505e-01 4.64693993e-01
4.46543366e-01 2.77860641e-01 -2.31309533e-01 5.81318140e-01
-4.72244680e-01 -1.49044162e-02 2.07931951e-01 9.61718440e-01
-8.42418194e-01 -9.36258256e-01 -1.40682533e-01 1.01030779e+00
-1.36491567e-01 2.40704969e-01 -5.70969462e-01 4.59980488e-01
3.38225156e-01 7.70494342e-01 7.01362938e-02 -6.72043025e-01
3.01277161e-01 3.88090014e-01 4.77043658e-01 -8.19785893e-01
-9.57146168e-01 -4.57717687e-01 1.03699975e-02 -4.68077511e-01
7.66707063e-02 -6.97323561e-01 -9.88684893e-01 -1.89745784e-01
-8.71965468e-01 4.57966208e-01 7.04437613e-01 1.07187533e+00
1.64352775e-01 4.57683563e-01 8.63933921e-01 -7.48464465e-01
-8.34202051e-01 -8.25087130e-01 -7.41902947e-01 4.16602403e-01
2.55285241e-02 -9.21384752e-01 -5.13710201e-01 -3.36752415e-01] | [8.309303283691406, 4.342864990234375] |
c5d16d16-3a2a-49cb-9f7e-cebcfdbebadf | a-reproducible-analysis-of-rssi | 1908.06851 | null | https://arxiv.org/abs/1908.06851v1 | https://arxiv.org/pdf/1908.06851v1.pdf | A Reproducible Analysis of RSSI Fingerprinting for Outdoor Localization Using Sigfox: Preprocessing and Hyperparameter Tuning | Fingerprinting techniques, which are a common method for indoor localization, have been recently applied with success into outdoor settings. Particularly, the communication signals of Low Power Wide Area Networks (LPWAN) such as Sigfox, have been used for localization. In this rather recent field of study, not many publicly available datasets, which would facilitate the consistent comparison of different positioning systems, exist so far. In the current study, a published dataset of RSSI measurements on a Sigfox network deployed in Antwerp, Belgium is used to analyse the appropriate selection of preprocessing steps and to tune the hyperparameters of a kNN fingerprinting method. Initially, the tuning of hyperparameter k for a variety of distance metrics, and the selection of efficient data transformation schemes, proposed by relevant works, is presented. In addition, accuracy improvements are achieved in this study, by a detailed examination of the appropriate adjustment of the parameters of the data transformation schemes tested, and of the handling of out of range values. With the appropriate tuning of these factors, the achieved mean localization error was 298 meters, and the median error was 109 meters. To facilitate the reproducibility of tests and comparability of results, the code and train/validation/test split used in this study are available. | ['Alexandros Kalousis', 'Grigorios G. Anagnostopoulos'] | 2019-08-14 | null | null | null | null | ['outdoor-localization'] | ['robots'] | [ 5.73236234e-02 -3.24250251e-01 1.08483806e-01 -3.95301014e-01
-5.49598515e-01 -6.69845462e-01 4.41784382e-01 6.05102837e-01
-6.70663714e-01 1.17938495e+00 -5.79295009e-02 -3.32692713e-01
-1.13630593e+00 -9.12682176e-01 -3.52140814e-01 -8.41171622e-01
-4.11357939e-01 2.87001580e-01 2.99212247e-01 3.54871862e-02
1.12151779e-01 9.07183826e-01 -1.43314230e+00 -3.74513596e-01
8.36153090e-01 1.01777387e+00 4.43989336e-02 5.92038214e-01
3.67072940e-01 9.45769399e-02 -1.09151614e+00 -6.31458014e-02
1.04329571e-01 -2.63294399e-01 -3.26329052e-01 -5.92245638e-01
2.06382051e-01 1.43446196e-02 -1.23288833e-01 4.41061467e-01
1.20948303e+00 3.97241384e-01 3.18008661e-01 -1.07303178e+00
2.20013589e-01 7.21529603e-01 1.42328113e-01 1.59047157e-01
7.43986607e-01 -3.92618805e-01 2.68359780e-01 -2.92287737e-01
4.31653380e-01 5.00249982e-01 1.02332819e+00 -2.44243160e-01
-1.20807195e+00 -8.03689539e-01 -5.90419531e-01 3.93134691e-02
-1.95339191e+00 -3.61569226e-01 5.68038702e-01 -8.24787468e-02
5.88589430e-01 5.96653163e-01 6.92143857e-01 1.02016747e+00
2.55762905e-01 -2.81642497e-01 1.06991112e+00 -7.63146758e-01
6.21135712e-01 4.58061814e-01 -2.75984496e-01 4.98520769e-02
9.20850456e-01 1.31997064e-01 -3.50708485e-01 -5.54276288e-01
5.14975667e-01 -1.99952841e-01 -3.29841048e-01 -6.14067674e-01
-1.19537663e+00 4.96107161e-01 5.20845652e-01 1.08693016e+00
-2.82114893e-01 2.86333889e-01 1.97804287e-01 1.29446462e-01
5.08422256e-02 5.70409894e-01 -3.76052499e-01 -4.39580172e-01
-1.11974621e+00 2.64320616e-02 9.66548026e-01 1.02048218e+00
5.93854070e-01 -3.55106115e-01 -2.56312728e-01 7.07445800e-01
4.77743387e-01 6.72410786e-01 2.91662484e-01 -7.25495696e-01
5.66148996e-01 3.51333052e-01 4.13365245e-01 -1.23870540e+00
-9.63288784e-01 -6.28092468e-01 -6.70203567e-01 -2.76096016e-01
7.71992147e-01 -7.18815744e-01 -2.54906058e-01 1.38434255e+00
2.74079978e-01 1.92843359e-02 -2.56528202e-02 2.37531871e-01
3.55687588e-01 3.07012290e-01 -9.26448181e-02 -1.87653854e-01
1.07037783e+00 -1.48133054e-01 -5.49487174e-01 1.15077943e-01
7.96510220e-01 -9.56534386e-01 4.66201603e-01 1.77226678e-01
-5.77169299e-01 -5.68715632e-01 -1.21742177e+00 6.53761864e-01
-5.82255483e-01 2.99373686e-01 5.10378659e-01 1.46655691e+00
-1.07985234e+00 4.93862271e-01 -8.42820644e-01 -1.03900015e+00
9.81936976e-02 7.48703361e-01 -3.27851325e-01 -8.11672062e-02
-1.32762861e+00 8.32811654e-01 4.81906205e-01 3.12573999e-01
-8.14061787e-04 -5.10870159e-01 -6.49515808e-01 3.35895419e-02
-6.15571886e-02 -3.54213953e-01 6.17972255e-01 -2.08601549e-01
-1.35035598e+00 2.41627946e-01 2.13510096e-01 -4.47343022e-01
5.73200166e-01 1.17158771e-01 -8.36816967e-01 -1.49807617e-01
1.24600038e-01 1.23280153e-01 8.80536288e-02 -9.94274080e-01
-6.54826105e-01 -2.11447135e-01 -1.72738671e-01 -3.43403339e-01
-1.12395205e-01 -2.24852279e-01 -2.99915314e-01 -3.86570156e-01
2.49976665e-01 -9.38240767e-01 -3.54916245e-01 -4.35706735e-01
-8.95634815e-02 3.20766687e-01 3.23960811e-01 -3.11797202e-01
1.35381567e+00 -2.27918959e+00 -5.73602855e-01 1.17939985e+00
-3.87980014e-01 -1.86162978e-01 3.14953715e-01 8.07355165e-01
2.75042951e-01 -2.42458284e-01 1.93160586e-02 -1.72119606e-02
2.64751136e-01 1.74087122e-01 2.14372590e-01 8.51336896e-01
-6.62748694e-01 3.27229321e-01 -8.49233449e-01 -3.19981098e-01
7.15172887e-01 7.40139306e-01 -2.03326091e-01 -1.83549523e-01
5.73180795e-01 4.44193691e-01 -6.61801398e-01 3.06633770e-01
7.59206057e-01 8.84991065e-02 3.71579677e-01 -1.05508752e-01
-4.68261957e-01 7.55208060e-02 -1.58518708e+00 1.42407835e+00
-7.21920609e-01 5.85223794e-01 -4.16909367e-01 -4.69298214e-01
1.34204686e+00 3.29026669e-01 7.26849854e-01 -7.81484723e-01
2.75578052e-01 6.34481668e-01 -1.99113451e-02 -3.23162913e-01
3.17019075e-01 4.36014622e-01 -8.81228819e-02 2.82462448e-01
-8.40377584e-02 -3.33068036e-02 4.12082911e-01 -2.50338733e-01
1.13111913e+00 -1.52208991e-02 6.10670030e-01 -4.86592114e-01
7.65273690e-01 -1.07689604e-01 6.79468438e-02 9.42767799e-01
3.30129154e-02 4.52506244e-01 6.42451942e-02 -7.02120140e-02
-5.16033947e-01 -6.57615066e-01 -8.32589149e-01 7.43544877e-01
2.63123363e-01 -3.66820395e-01 -6.15860581e-01 -1.56134665e-01
1.82324618e-01 7.24209130e-01 -4.31897432e-01 2.01636359e-01
-4.77310419e-01 -1.02962971e+00 9.55115199e-01 2.05882162e-01
6.01301134e-01 -8.62187505e-01 -8.16833079e-01 4.29458201e-01
-4.21902863e-03 -9.71354425e-01 2.22881019e-01 5.62322855e-01
-7.22770751e-01 -1.10798955e+00 -5.16637862e-01 -4.82978731e-01
7.69781053e-01 7.34640565e-03 6.81472778e-01 2.47882009e-01
-4.58537117e-02 4.17417973e-01 -4.33736801e-01 -2.42371574e-01
-2.00279742e-01 4.60383922e-01 9.30851791e-03 -2.44055092e-01
1.79415777e-01 -6.78942382e-01 -5.47752380e-01 8.89108539e-01
-5.60316443e-01 -8.95727992e-01 7.91125000e-01 3.40346813e-01
4.13507968e-01 2.98374712e-01 2.41454631e-01 -7.53614128e-01
5.22595823e-01 -4.83248889e-01 -9.61389422e-01 3.08404624e-01
-8.51420105e-01 -1.15771681e-01 5.80317259e-01 -3.24407332e-02
-5.94793439e-01 -1.67513654e-01 -3.95409346e-01 5.19447684e-01
-5.32317162e-01 4.00070071e-01 -3.43869776e-01 -7.96291053e-01
9.10938561e-01 -7.92111978e-02 -3.88578057e-01 -3.04036528e-01
-1.90058693e-01 7.56837487e-01 2.85526097e-01 -5.13617277e-01
8.72551084e-01 2.73930997e-01 2.78087467e-01 -9.26445246e-01
-8.63971710e-02 -4.72253054e-01 -6.53319120e-01 -1.06536008e-01
2.83774227e-01 -6.04600430e-01 -6.67866528e-01 2.30054811e-01
-6.63865685e-01 -3.24016511e-01 -2.94172794e-01 7.85843015e-01
-2.57095754e-01 5.60157560e-02 2.56378949e-01 -7.58120358e-01
-6.68412913e-03 -9.58903134e-01 6.43950224e-01 4.92693335e-01
-5.07966161e-01 -1.24569917e+00 3.05694640e-01 5.91967143e-02
1.00791478e+00 4.20319378e-01 4.28466946e-01 -7.79293120e-01
-3.26433808e-01 -7.38491476e-01 1.95735589e-01 -1.42968982e-01
2.90896475e-01 -1.09558761e-01 -9.34301496e-01 -6.35321856e-01
-2.50643581e-01 5.03304899e-01 1.48581816e-02 4.57136363e-01
9.57154155e-01 5.82955703e-02 -7.83771098e-01 6.94698155e-01
1.46296322e+00 2.22043052e-01 8.05504680e-01 9.13138151e-01
3.52002829e-02 1.70470104e-01 8.13474119e-01 5.44361174e-01
1.56133622e-01 1.06279182e+00 3.50965649e-01 -5.30980974e-02
2.06376642e-01 -1.11873530e-01 -9.40127149e-02 8.87958482e-02
-1.90588579e-01 -4.32016224e-01 -8.70715678e-01 2.40948915e-01
-1.43160450e+00 -6.64612412e-01 -2.58181036e-01 2.70497441e+00
7.01042777e-03 2.81131864e-01 4.46121506e-02 6.20645642e-01
6.13949120e-01 1.58020198e-01 5.68939634e-02 -1.61468536e-01
4.31808941e-02 2.34510690e-01 1.16613472e+00 5.69631279e-01
-8.86120737e-01 5.54569773e-02 6.41349936e+00 5.84467351e-01
-1.15480757e+00 -1.43025130e-01 -6.22202195e-02 2.06283629e-01
-9.73129869e-02 -7.09022433e-02 -7.64594316e-01 7.93554008e-01
1.39195740e+00 2.18300782e-02 2.32875913e-01 7.61695743e-01
4.94688779e-01 -6.36690319e-01 -7.52014220e-01 9.91276979e-01
-1.36993155e-01 -1.07589495e+00 -6.71524763e-01 2.21561134e-01
6.12054050e-01 3.02775996e-03 -3.49611104e-01 -2.54412834e-02
-2.67026514e-01 -7.59858370e-01 6.00467980e-01 4.63605016e-01
4.21410412e-01 -8.75445426e-01 1.44487655e+00 1.92828644e-02
-9.90716100e-01 -2.69740462e-01 -2.17684031e-01 7.03965351e-02
1.93464179e-02 8.95539701e-01 -8.17195237e-01 1.03478110e+00
8.01656127e-01 1.11693494e-01 -8.76617134e-01 1.76514161e+00
-4.32989448e-01 6.72574103e-01 -8.91721070e-01 -4.09693778e-01
-1.28013849e-01 -4.65565175e-02 2.06155688e-01 1.22269893e+00
8.49945128e-01 -4.48038131e-01 -4.74681646e-01 3.47860426e-01
2.86361039e-01 2.98148066e-01 -3.77110004e-01 4.98221517e-01
1.17898571e+00 1.23068523e+00 -7.41918147e-01 4.11959350e-01
-1.36052712e-03 3.53839099e-01 -4.72053230e-01 4.26935345e-01
-7.88449943e-01 -9.32387948e-01 3.61903489e-01 4.55691755e-01
2.96399146e-01 -3.81558329e-01 -1.45689875e-01 -2.28400216e-01
1.05996124e-01 -2.92134166e-01 2.88731903e-01 -4.34386224e-01
-6.15192592e-01 4.30123270e-01 3.76932293e-01 -1.22773874e+00
-2.79333442e-01 -2.90337503e-01 -6.32445097e-01 8.03566277e-01
-1.08734250e+00 -6.86424553e-01 -7.39883363e-01 3.40285838e-01
-3.73482347e-01 7.37500414e-02 1.01998365e+00 6.79781497e-01
-3.65577787e-01 8.39896917e-01 9.34844136e-01 -1.02001280e-01
6.75475240e-01 -9.24128473e-01 1.65712070e-02 7.84295261e-01
-6.95803538e-02 9.06856060e-01 1.03941894e+00 -3.21980625e-01
-1.10435688e+00 -8.19094002e-01 9.66994822e-01 -2.82089025e-01
5.21967173e-01 -4.47482467e-01 -3.14024687e-01 4.42627043e-01
-1.67667851e-01 7.43649974e-02 8.66110742e-01 1.40289098e-01
4.63437140e-01 -6.53418005e-01 -1.62804580e+00 2.84590006e-01
6.28163636e-01 -2.10602973e-02 3.48962359e-02 1.10952802e-01
-9.38565359e-02 -5.45756638e-01 -1.34665751e+00 3.61073881e-01
7.75782526e-01 -1.19867384e+00 1.06595647e+00 5.47654927e-01
-7.27102160e-01 -6.29663169e-01 -2.82493562e-01 -1.17726791e+00
-1.59188375e-01 -6.16017818e-01 2.77038753e-01 1.61795723e+00
4.88283604e-01 -1.10233676e+00 7.05449522e-01 2.26093650e-01
2.24795505e-01 -5.36499023e-01 -1.29035306e+00 -6.49081945e-01
-6.65065825e-01 -4.36556309e-01 1.00836611e+00 4.06399190e-01
-2.54766494e-01 -2.40046680e-01 4.01801839e-02 5.38998783e-01
6.85534120e-01 -4.29527938e-01 9.80083883e-01 -1.24908519e+00
-8.72619301e-02 -1.09166525e-01 -8.87676954e-01 -6.31726503e-01
-5.37874162e-01 -5.34248352e-01 -3.88416573e-02 -1.63373530e+00
-5.86284220e-01 -1.08480036e+00 -4.20112789e-01 1.92595467e-01
3.69517267e-01 4.46643710e-01 -2.38840386e-01 4.05075401e-02
-4.24651504e-01 3.22388485e-02 4.63953823e-01 1.14871539e-01
-4.63589132e-01 7.18595088e-01 -3.68466914e-01 2.97796518e-01
7.83950686e-01 -7.98342526e-01 -3.86372596e-01 -1.52120516e-01
4.34979141e-01 -2.08808661e-01 1.99644938e-01 -1.71037555e+00
4.53510523e-01 1.78896591e-01 9.03300524e-01 -5.56832671e-01
9.74838436e-02 -1.51266193e+00 9.24905181e-01 5.83669484e-01
4.84754884e-04 -5.99204153e-02 3.17250997e-01 4.35303360e-01
4.93283644e-02 -3.17597955e-01 5.79608500e-01 5.78806818e-01
-5.01996577e-01 -1.84347719e-01 -3.60554039e-01 -4.10074204e-01
1.23163152e+00 -6.80583775e-01 -1.36585832e-01 -3.16955566e-01
-3.45805109e-01 9.97941941e-02 5.83991468e-01 1.36306629e-01
-3.07551250e-02 -1.28215063e+00 -2.85602033e-01 2.17298880e-01
2.78148115e-01 -2.05829218e-01 2.09243640e-01 1.30169702e+00
-8.93118918e-01 6.50417626e-01 -3.57721984e-01 -5.21162510e-01
-1.07573807e+00 -5.24549857e-02 5.86374164e-01 -1.43659478e-02
-3.87984157e-01 2.88064092e-01 -9.49661374e-01 -5.24839103e-01
4.07319754e-01 -5.15183806e-01 -3.71332079e-01 1.15812466e-01
2.88201094e-01 8.39231193e-01 5.48228562e-01 -4.96748805e-01
-9.25727487e-01 7.38417804e-01 7.42987156e-01 9.14724171e-02
1.36204612e+00 -4.69389111e-01 -1.45259678e-01 2.15554491e-01
8.44220042e-01 6.49281085e-01 -7.48681724e-01 3.24119538e-01
3.19104701e-01 -5.06299555e-01 -2.90714681e-01 -7.53093958e-01
-5.96778214e-01 4.30339091e-02 1.09131038e+00 5.40396571e-01
1.00952029e+00 -2.17177168e-01 1.31849637e-02 3.45500737e-01
1.07686508e+00 -7.02413023e-01 -8.76006126e-01 1.97012678e-01
3.48325402e-01 -7.62002230e-01 3.09220523e-01 -2.69395202e-01
2.75585562e-01 1.22243261e+00 -6.55794218e-02 1.98919997e-01
6.84940755e-01 3.97798091e-01 1.36729375e-01 2.61165276e-02
2.24143714e-01 1.31527305e-01 -9.22402591e-02 8.24774384e-01
4.84269440e-01 -1.44952029e-01 -7.42625892e-01 3.97249162e-01
-5.26669681e-01 1.76996931e-01 3.33766550e-01 1.02434945e+00
-4.47613537e-01 -1.36850095e+00 -8.27387810e-01 3.27670813e-01
-3.53015065e-01 4.06764805e-01 -8.65447242e-03 1.18850243e+00
4.00290400e-01 1.15364301e+00 -3.17499071e-01 -2.30255067e-01
8.06237340e-01 -3.52643192e-01 4.69012231e-01 -7.10493624e-02
-7.90385127e-01 -4.40033436e-01 2.00810745e-01 -3.71358544e-01
-4.59021538e-01 -6.38828933e-01 -1.09294033e+00 -4.32644159e-01
-5.09698391e-01 9.07414138e-01 1.28782403e+00 9.53248262e-01
1.64714858e-01 5.53686261e-01 6.46221101e-01 -7.82629967e-01
-1.27462372e-01 -9.77193654e-01 -9.96073544e-01 -5.24032749e-02
3.57524566e-02 -7.67704129e-01 -6.89346075e-01 -6.99882746e-01] | [6.422292709350586, 1.0323995351791382] |
8e469661-433f-46f3-84b2-65b0da1c5dbf | simplifying-subgraph-representation-learning | 2301.12562 | null | https://arxiv.org/abs/2301.12562v2 | https://arxiv.org/pdf/2301.12562v2.pdf | Simplifying Subgraph Representation Learning for Scalable Link Prediction | Link prediction on graphs is a fundamental problem. Subgraph representation learning approaches (SGRLs), by transforming link prediction to graph classification on the subgraphs around the links, have achieved state-of-the-art performance in link prediction. However, SGRLs are computationally expensive, and not scalable to large-scale graphs due to expensive subgraph-level operations. To unlock the scalability of SGRLs, we propose a new class of SGRLs, that we call Scalable Simplified SGRL (S3GRL). Aimed at faster training and inference, S3GRL simplifies the message passing and aggregation operations in each link's subgraph. S3GRL, as a scalability framework, accommodates various subgraph sampling strategies and diffusion operators to emulate computationally-expensive SGRLs. We propose multiple instances of S3GRL and empirically study them on small to large-scale graphs. Our extensive experiments demonstrate that the proposed S3GRL models scale up SGRLs without significant performance compromise (even with considerable gains in some cases), while offering substantially lower computational footprints (e.g., multi-fold inference and training speedup). | ['Amirali Salehi-Abari', 'Shweta Ann Jacob', 'Paul Louis'] | 2023-01-29 | null | null | null | null | ['graph-classification'] | ['graphs'] | [ 5.93052842e-02 3.98986429e-01 -7.06082463e-01 -9.11390930e-02
-6.77798510e-01 -4.65860009e-01 4.65708643e-01 4.46567088e-01
3.43796790e-01 7.63188004e-01 -1.24656983e-01 -8.17754865e-01
-3.76755387e-01 -1.46280038e+00 -7.21544206e-01 -3.67304325e-01
-6.49958372e-01 4.97178018e-01 6.53184533e-01 -1.08636089e-01
-2.02442303e-01 6.30783558e-01 -1.19129050e+00 6.03212491e-02
9.93739367e-01 7.39811778e-01 -7.37123266e-02 6.43807292e-01
-1.38502926e-01 8.30086768e-01 -6.56392425e-02 -7.26837039e-01
3.07433575e-01 -1.81910560e-01 -8.49781871e-01 5.13574108e-02
4.69295055e-01 -4.31117369e-03 -8.32000673e-01 9.00487959e-01
1.29792869e-01 -6.26383871e-02 3.83776635e-01 -1.54048324e+00
-2.21260786e-01 8.24991643e-01 -7.23140657e-01 3.40049039e-03
5.79274297e-01 -1.44005015e-01 1.59830427e+00 -5.59997559e-01
6.64886057e-01 1.32722318e+00 9.43866670e-01 -2.08863076e-02
-1.62346220e+00 -5.60185850e-01 5.26626110e-01 2.22817376e-01
-1.64318383e+00 -1.35758013e-01 6.90127432e-01 -2.20431700e-01
8.16565454e-01 5.81685603e-01 7.68010616e-01 5.70904851e-01
1.03855990e-01 9.11985278e-01 8.28988433e-01 -1.72346622e-01
1.02149867e-01 -1.60011187e-01 2.21695557e-01 1.09964180e+00
7.03821540e-01 -1.84714541e-01 -3.93556446e-01 -4.86096710e-01
9.04896319e-01 -5.20067327e-02 -2.97168940e-01 -5.89296699e-01
-9.95788693e-01 9.10113215e-01 7.74332225e-01 -8.24655145e-02
-7.72345737e-02 3.08906078e-01 4.27996874e-01 6.20520353e-01
6.51379287e-01 -6.81990981e-02 -3.70030105e-01 3.08920801e-01
-7.87349463e-01 1.20524086e-01 1.15797603e+00 1.46467769e+00
1.07556891e+00 -2.48587564e-01 -1.61386281e-01 6.35881186e-01
4.04900193e-01 1.73798561e-01 -1.87887207e-01 -3.47260684e-01
7.60505915e-01 1.05268359e+00 -4.24330622e-01 -1.32016349e+00
-5.97391367e-01 -7.20272243e-01 -1.06962955e+00 -4.06816572e-01
1.86747655e-01 2.44913399e-01 -6.09612286e-01 1.56223106e+00
4.09899086e-01 4.28978473e-01 -1.80213168e-01 4.47998524e-01
8.97456944e-01 6.36404276e-01 5.91527969e-02 -3.39423686e-01
1.09825778e+00 -1.15055728e+00 -2.20007122e-01 -2.28171363e-01
1.20184326e+00 -3.49759042e-01 8.81631553e-01 -1.11829303e-01
-8.91133189e-01 -2.99091935e-01 -8.43293726e-01 1.31870314e-01
-3.15841019e-01 2.90716905e-02 1.37055063e+00 6.46404147e-01
-1.24694765e+00 7.32890904e-01 -8.08063090e-01 -5.22162795e-01
3.38256896e-01 3.83423448e-01 -4.59703386e-01 -1.99798509e-01
-1.06815910e+00 4.55139369e-01 4.99708086e-01 -5.24854399e-02
-5.75176716e-01 -6.56771958e-01 -1.10576344e+00 2.74496883e-01
8.26381385e-01 -6.36873841e-01 7.06397235e-01 -3.81016254e-01
-1.00529778e+00 6.63553178e-01 -2.75146276e-01 -4.60199207e-01
2.85183281e-01 2.48253345e-02 -6.00143015e-01 2.13744789e-02
3.47165056e-02 2.15780944e-01 5.50507486e-01 -1.16261864e+00
-3.98863375e-01 -6.62528053e-02 1.86049834e-01 7.51719018e-03
-3.39717269e-01 -2.31122330e-01 -7.42648363e-01 -7.91390836e-01
5.82241893e-01 -1.07667875e+00 -3.59765559e-01 -1.72459096e-01
-8.35850418e-01 -4.26502228e-01 4.99796420e-01 -4.30317461e-01
1.82269096e+00 -1.68281698e+00 1.15746841e-01 7.87509978e-01
6.80312872e-01 1.67809561e-01 -3.59455228e-01 8.68325055e-01
6.79185167e-02 7.04455599e-02 -2.84589622e-02 -8.51365551e-03
-3.95839103e-02 3.03157836e-01 -1.89073697e-01 4.34445113e-01
5.01731597e-02 1.18028402e+00 -1.00678825e+00 -6.80941284e-01
-1.18600965e-01 1.53327351e-02 -7.78716624e-01 1.28495216e-01
-3.53030413e-01 1.62951365e-01 -6.91397190e-01 8.95811558e-01
5.01953363e-01 -1.00539434e+00 8.37107122e-01 -2.87667841e-01
2.31027424e-01 5.56718528e-01 -1.26777172e+00 1.24517846e+00
-2.16399565e-01 2.60581970e-01 -1.15327634e-01 -1.15277827e+00
9.86522496e-01 -9.74290520e-02 4.36770290e-01 -2.93842554e-01
-3.46735448e-01 2.11597547e-01 -1.07630253e-01 -9.02583450e-02
2.62554079e-01 4.66694236e-01 -2.46294085e-02 4.73898053e-01
-1.05157308e-01 4.16181803e-01 6.57746673e-01 7.46870577e-01
1.55112004e+00 1.70392111e-01 6.74775481e-01 -3.48675042e-01
6.48076773e-01 -2.02518106e-01 5.87534308e-01 8.30933511e-01
2.26826221e-01 3.97474170e-02 8.32453728e-01 -5.45388281e-01
-6.24020696e-01 -1.08538306e+00 3.12572449e-01 1.30851305e+00
2.97040522e-01 -1.06456411e+00 -4.05193835e-01 -9.35935855e-01
3.65776807e-01 2.03667447e-01 -1.34184957e-01 -7.16994554e-02
-6.33524001e-01 -9.60448980e-01 3.25683385e-01 4.98146772e-01
2.75438547e-01 -5.14524400e-01 2.13809967e-01 3.08432817e-01
1.73010290e-01 -1.25396836e+00 -3.93939227e-01 -1.29358038e-01
-1.24186945e+00 -1.36875343e+00 -1.46060631e-01 -8.16648841e-01
1.00567782e+00 5.55532634e-01 1.40427041e+00 4.54957485e-01
-2.91041523e-01 1.60041362e-01 -4.04857785e-01 4.15027410e-01
-2.69976676e-01 5.52016973e-01 -1.43987522e-01 -9.02092084e-02
9.67939757e-03 -9.62425351e-01 -4.38936204e-01 3.96660835e-01
-5.49359858e-01 4.10062551e-01 9.52256441e-01 7.30980933e-01
7.71412551e-01 1.91492498e-01 5.09104490e-01 -1.53152263e+00
4.10045028e-01 -6.22029841e-01 -8.32107961e-01 4.67097789e-01
-8.97926331e-01 2.33688965e-01 5.01631856e-01 -1.70228764e-01
-5.56811452e-01 -1.41914219e-01 -1.30474180e-01 -1.70047402e-01
2.61671096e-01 8.74001086e-01 -1.71850786e-01 -4.72600669e-01
1.84525371e-01 1.71749428e-01 -5.40328659e-02 -6.79580271e-01
4.25756544e-01 1.46373942e-01 1.62534788e-01 -6.29263580e-01
1.23398864e+00 2.65245616e-01 4.58957136e-01 -6.56968832e-01
-8.79968882e-01 -4.81351763e-01 -5.17777681e-01 -1.86508238e-01
1.54935539e-01 -1.00445580e+00 -7.83021092e-01 4.55798246e-02
-7.77608812e-01 -3.28772128e-01 7.93190971e-02 1.78319395e-01
-3.56958091e-01 5.33405542e-01 -8.81928742e-01 -5.94169915e-01
-3.37776989e-01 -9.07113969e-01 9.19255316e-01 -1.62766218e-01
-8.70589316e-02 -1.16708004e+00 -1.57541111e-01 1.45608217e-01
2.15544477e-01 3.46376449e-01 1.45631015e+00 -8.04195642e-01
-1.01451826e+00 -3.52717161e-01 -5.80241144e-01 -2.66101331e-01
8.64149481e-02 -2.63026714e-01 -3.47825587e-01 -8.06659758e-01
-9.57993627e-01 -1.98315322e-01 8.04913878e-01 -3.56356986e-03
1.30470634e+00 -6.37061954e-01 -9.20809686e-01 7.06799150e-01
1.52742851e+00 -2.44345963e-01 5.71568787e-01 9.76944938e-02
9.94914532e-01 2.05646515e-01 6.17454767e-01 2.31862217e-01
7.04282343e-01 7.26366222e-01 4.56322819e-01 -3.14974189e-01
-3.61442655e-01 -6.34677410e-01 3.88601005e-01 1.28337860e+00
-2.37143174e-01 -3.17477345e-01 -6.48394704e-01 1.65270433e-01
-2.05905557e+00 -6.77603483e-01 -5.78654289e-01 2.13645411e+00
6.37534201e-01 3.06394845e-01 3.06983232e-01 1.26219764e-01
8.69251728e-01 3.78443956e-01 -5.28214335e-01 5.71293384e-02
1.37755752e-01 1.14513621e-01 6.77444994e-01 5.39412022e-01
-1.11352718e+00 1.13030803e+00 6.50327587e+00 1.00974822e+00
-5.87900221e-01 -7.73082674e-02 4.38241422e-01 2.90639043e-01
-5.35436749e-01 4.17513967e-01 -9.58793700e-01 2.21301153e-01
7.67952979e-01 -3.01084548e-01 6.45582020e-01 8.76805007e-01
-1.45864546e-01 2.76705861e-01 -1.18539512e+00 7.20766366e-01
-2.50401974e-01 -1.51366842e+00 3.98262739e-01 2.51247495e-01
8.04082334e-01 1.49817122e-02 -4.62223768e-01 4.77485776e-01
5.45618892e-01 -7.60399401e-01 1.44188702e-01 1.93335027e-01
7.73584187e-01 -7.02674091e-01 5.95988333e-01 1.32673919e-01
-2.05290866e+00 3.96013744e-02 -3.66111726e-01 -6.27448037e-03
3.54994014e-02 8.58901203e-01 -6.73673689e-01 1.11689949e+00
3.92704785e-01 8.57673645e-01 -8.97842586e-01 8.93962145e-01
-2.81489640e-01 8.37336063e-01 -3.98203611e-01 -4.03504521e-02
6.17903508e-02 -4.01044548e-01 5.09581029e-01 1.18001342e+00
2.65108585e-01 -5.05047366e-02 7.48378277e-01 5.81510484e-01
-4.48436677e-01 2.15282574e-01 -5.88771880e-01 -2.17588753e-01
9.13446605e-01 1.26269293e+00 -7.95378864e-01 -2.58614033e-01
-7.17822611e-01 5.29137611e-01 8.35394621e-01 3.98984611e-01
-8.01626384e-01 -4.73435163e-01 7.02619314e-01 5.78901529e-01
3.40185076e-01 -1.89491302e-01 2.01103359e-01 -1.09487569e+00
-2.07157759e-03 -7.31683254e-01 8.08105469e-01 -1.80660054e-01
-1.34997737e+00 5.49765646e-01 8.25354606e-02 -1.19416606e+00
-4.82429415e-02 -5.08193195e-01 -6.40302598e-01 3.24669391e-01
-1.61258221e+00 -1.33683932e+00 -3.51900339e-01 5.70195615e-01
-6.95429929e-03 -1.18976627e-02 6.88198268e-01 3.42797637e-01
-6.82870150e-01 8.62765193e-01 2.06547249e-02 1.22013852e-01
3.15842032e-01 -1.11179364e+00 6.87006652e-01 7.05623865e-01
2.52055407e-01 6.53366983e-01 4.24359500e-01 -8.87680233e-01
-1.82125735e+00 -1.53840768e+00 8.79830658e-01 1.81077018e-01
9.75261509e-01 -3.86179060e-01 -9.85612214e-01 1.11295068e+00
-5.09154677e-01 3.38335514e-01 6.57917917e-01 7.16460228e-01
-5.67614675e-01 -4.01329845e-01 -8.99282038e-01 6.99633360e-01
1.64855444e+00 -4.84112561e-01 -7.44892359e-02 6.59286022e-01
7.15203226e-01 -2.67126918e-01 -1.40309429e+00 6.53059244e-01
2.99903631e-01 -8.62721741e-01 1.08206904e+00 -5.74689686e-01
-1.78520586e-02 -3.91689181e-01 2.61536893e-02 -9.41347003e-01
-6.20397270e-01 -8.44473779e-01 -8.04967642e-01 1.04790080e+00
4.40681398e-01 -1.01196861e+00 1.06107914e+00 8.20467398e-02
5.86120784e-02 -1.03140736e+00 -7.31839478e-01 -1.04908907e+00
-4.14420754e-01 -3.44142407e-01 7.07148075e-01 8.74127924e-01
1.72102079e-01 4.17345405e-01 -2.96324313e-01 5.46567082e-01
8.80743146e-01 7.23882794e-01 1.09623230e+00 -1.49847746e+00
-4.85109180e-01 -3.71931106e-01 -7.80084014e-01 -1.31673908e+00
1.89965591e-01 -1.54850376e+00 -4.49869007e-01 -1.80699027e+00
3.99086773e-01 -6.75738335e-01 -1.52316242e-01 7.08924234e-01
-3.42801362e-01 1.77190229e-01 -4.41284105e-02 3.17319572e-01
-9.17028785e-01 3.06252062e-01 1.15272188e+00 -1.02392681e-01
-1.97784171e-01 1.78884864e-01 -5.78959823e-01 6.85516775e-01
5.56606412e-01 -1.71920031e-01 -6.31097853e-01 -2.02470017e-03
4.49550539e-01 1.21183179e-01 3.41380239e-01 -9.05086458e-01
2.85916150e-01 -1.30267188e-01 1.66980792e-02 -6.95775151e-01
-1.18872933e-02 -4.60562915e-01 5.64690709e-01 6.20251417e-01
-2.10491881e-01 -1.64256617e-01 -2.02951804e-01 9.33779776e-01
-9.56149846e-02 2.07040995e-01 4.36724693e-01 5.61572351e-02
-6.64361894e-01 7.20185935e-01 8.64039287e-02 -2.78471708e-02
1.13062143e+00 -5.84875196e-02 -4.49482858e-01 -4.40995753e-01
-7.39090741e-01 5.18854678e-01 3.55205745e-01 1.33616671e-01
4.55431879e-01 -1.50416291e+00 -5.49544096e-01 1.94206581e-01
3.58442217e-01 2.25065947e-02 -1.38413981e-01 9.70931351e-01
-4.94864196e-01 4.57377166e-01 4.89487082e-01 -4.02893126e-01
-1.37064660e+00 6.25909150e-01 -1.27412766e-01 -9.40516055e-01
-1.00738120e+00 8.16415906e-01 1.23688012e-01 -5.15143096e-01
2.21424356e-01 -1.73149213e-01 1.12182923e-01 -1.94809407e-01
1.51498020e-01 6.50775969e-01 1.04622990e-02 -3.39007437e-01
-2.45161697e-01 4.01077479e-01 -2.11774021e-01 8.21242690e-01
1.20675611e+00 -1.41338753e-02 -3.38728249e-01 4.59201597e-02
1.04504418e+00 1.34182245e-01 -1.01571667e+00 -5.05688131e-01
3.97684753e-01 -4.08576101e-01 -1.38865471e-01 -2.74245828e-01
-1.23274398e+00 1.85930938e-01 -1.21787429e-01 5.28725386e-01
9.72239137e-01 3.44149828e-01 8.85404527e-01 5.83270967e-01
9.57325935e-01 -6.36345088e-01 -1.72123536e-01 2.55467176e-01
5.61349928e-01 -1.00842667e+00 6.65813148e-01 -1.43611968e+00
-6.97330758e-02 9.00336742e-01 6.55908525e-01 -1.59804493e-01
7.13048816e-01 8.80191475e-02 -6.54201567e-01 -2.64289021e-01
-9.22344923e-01 -3.42319995e-01 3.20912689e-01 3.62704158e-01
3.99764568e-01 3.08816314e-01 -3.95661384e-01 3.88512969e-01
8.68082047e-04 -4.07772779e-01 8.85388106e-02 8.38130653e-01
-4.04886156e-01 -1.35619521e+00 -5.47614228e-03 9.29198205e-01
8.24868605e-02 -1.70408428e-01 -2.90934503e-01 9.83281672e-01
-3.68868858e-01 7.21250713e-01 -6.20125495e-02 -5.39226770e-01
1.80883110e-01 -1.76887631e-01 4.62256730e-01 -6.45217240e-01
-2.32425377e-01 -6.77689388e-02 5.88577151e-01 -8.85322988e-01
-1.62480205e-01 -2.44624346e-01 -1.22759759e+00 -7.22598433e-01
-4.88813639e-01 2.47509062e-01 -1.34549253e-02 7.12782145e-01
5.79046726e-01 5.25571406e-01 7.73070216e-01 -5.89138865e-01
-4.69837129e-01 -7.60820746e-01 -8.36164951e-01 2.83364117e-01
-1.25135452e-01 -5.99513948e-01 -2.08483160e-01 -3.64691257e-01] | [7.021145820617676, 6.052331447601318] |
8db5ca0f-c7bb-49e4-923f-b876bad0b47b | otter-knowledge-benchmarks-of-multimodal | 2306.12802 | null | https://arxiv.org/abs/2306.12802v2 | https://arxiv.org/pdf/2306.12802v2.pdf | Otter-Knowledge: benchmarks of multimodal knowledge graph representation learning from different sources for drug discovery | Recent research in representation learning utilizes large databases of proteins or molecules to acquire knowledge of drug and protein structures through unsupervised learning techniques. These pre-trained representations have proven to significantly enhance the accuracy of subsequent tasks, such as predicting the affinity between drugs and target proteins. In this study, we demonstrate that by incorporating knowledge graphs from diverse sources and modalities into the sequences or SMILES representation, we can further enrich the representation and achieve state-of-the-art results on established benchmark datasets. We provide preprocessed and integrated data obtained from 7 public sources, which encompass over 30M triples. Additionally, we make available the pre-trained models based on this data, along with the reported outcomes of their performance on three widely-used benchmark datasets for drug-target binding affinity prediction found in the Therapeutic Data Commons (TDC) benchmarks. Additionally, we make the source code for training models on benchmark datasets publicly available. Our objective in releasing these pre-trained models, accompanied by clean data for model pretraining and benchmark results, is to encourage research in knowledge-enhanced representation learning. | ['Marcos Martínez Galindo', 'Vanessa López', 'Cesar Berrospi Ramis', 'Gabriele Picco', 'Víctor Valls', 'Raúl Fernández-Díaz', 'Mykhaylo Zayats', 'Marco Luca Sbodio', 'Hoang Thanh Lam'] | 2023-06-22 | null | null | null | null | ['knowledge-graphs', 'drug-discovery', 'graph-representation-learning'] | ['knowledge-base', 'medical', 'methodology'] | [ 7.78370202e-01 7.54803121e-02 -9.61134493e-01 -5.39153576e-01
-1.09313786e+00 -5.41141391e-01 5.00406265e-01 8.29531610e-01
-1.13320366e-01 1.33798349e+00 5.75902581e-01 -4.64760423e-01
-4.20394570e-01 -5.82066655e-01 -1.05463421e+00 -6.97118461e-01
-2.14354977e-01 6.33350790e-01 -2.96083778e-01 -8.91633481e-02
1.76027313e-01 5.29722571e-01 -1.33589208e+00 8.44099998e-01
9.67041314e-01 6.76619828e-01 -1.22247756e-01 8.07980597e-02
8.19538087e-02 1.17413402e+00 -1.80103138e-01 -4.40905005e-01
-2.68770941e-02 -2.57622212e-01 -1.05316103e+00 -4.70129013e-01
2.26763666e-01 2.12291777e-01 -3.89252156e-01 6.85209930e-01
5.81159353e-01 5.58686443e-02 9.78388727e-01 -6.42888606e-01
-1.09802878e+00 6.91043913e-01 -4.01287138e-01 1.21040881e-01
5.84580243e-01 1.73671514e-01 1.09344709e+00 -1.04030633e+00
1.02441001e+00 9.49208081e-01 6.07189178e-01 6.55035734e-01
-1.33399951e+00 -5.54470897e-01 -3.29317003e-01 4.84719276e-01
-1.34411228e+00 -6.46498978e-01 2.40479112e-01 -5.89886665e-01
1.69430113e+00 2.53417939e-01 1.10559255e-01 1.38969648e+00
-1.32364947e-02 4.47563231e-01 9.02666926e-01 -4.24523652e-01
7.12231696e-02 1.41255945e-01 2.10694537e-01 7.13683665e-01
4.57140297e-01 4.35672432e-01 -6.58088326e-01 -6.23374224e-01
2.45333508e-01 3.01323533e-01 -4.21122789e-01 -4.75123823e-01
-1.07026649e+00 8.90214801e-01 7.28910983e-01 1.22169629e-01
-7.18786120e-01 -1.41471788e-01 3.72196048e-01 5.77148460e-02
4.21128005e-01 8.15358579e-01 -8.53917301e-01 -1.50251061e-01
-4.95497346e-01 -4.88260202e-02 6.91016555e-01 8.25967729e-01
5.73788106e-01 -3.20840716e-01 -7.41944313e-02 1.01798165e+00
2.93602526e-01 2.26051152e-01 5.18310308e-01 -1.85794905e-01
5.07556915e-01 8.90267432e-01 -8.41057748e-02 -7.42910445e-01
-3.78539503e-01 -2.61239737e-01 -7.38094568e-01 -2.73824722e-01
2.08812743e-01 9.79761705e-02 -1.21460438e+00 1.64925921e+00
2.99578041e-01 3.79250646e-01 4.65329528e-01 5.77906132e-01
1.43590975e+00 5.59257329e-01 6.92214191e-01 -4.05665547e-01
1.11304724e+00 -6.60251796e-01 -6.29554391e-01 2.90909976e-01
1.10600662e+00 -5.87317348e-01 5.30777335e-01 3.81195843e-01
-7.24039972e-01 -2.34756917e-01 -1.01452506e+00 -1.36214783e-02
-9.12435949e-01 -1.57788903e-01 1.06949747e+00 4.20670003e-01
-4.00203645e-01 8.63039732e-01 -6.36341274e-01 -2.22825140e-01
9.65319812e-01 6.60378814e-01 -8.87769580e-01 -4.44248229e-01
-1.41583550e+00 1.33530259e+00 5.40578127e-01 -2.60698050e-01
-1.02382696e+00 -1.27984679e+00 -8.48191142e-01 -7.10237697e-02
1.64874077e-01 -6.27210796e-01 6.93448007e-01 -6.26757145e-01
-1.17683148e+00 8.16614568e-01 4.16851491e-02 -5.89970887e-01
-2.56144762e-01 -4.17397209e-02 -5.38264871e-01 1.06579401e-01
-1.28945947e-01 6.10322833e-01 1.98276132e-01 -9.50790524e-01
-2.63118148e-01 -4.61446404e-01 -1.06683262e-01 5.17925359e-02
-3.19432169e-01 -2.22164523e-02 -1.69560090e-01 -4.86500531e-01
-5.23681879e-01 -7.78997421e-01 -5.38006663e-01 -4.84950274e-01
-5.32728314e-01 -2.12825865e-01 3.10986489e-01 -6.27216518e-01
1.15400171e+00 -1.81881166e+00 5.46868861e-01 4.08351541e-01
2.80301750e-01 4.29614693e-01 -4.31070060e-01 7.11641371e-01
-8.78447950e-01 8.86825845e-02 -2.99572557e-01 4.76353019e-02
-3.55344266e-01 3.26485515e-01 -2.99231023e-01 5.40033042e-01
3.99312735e-01 1.03085542e+00 -9.88738537e-01 -2.61884660e-01
6.99512064e-02 7.34730780e-01 -6.42506063e-01 1.30206496e-01
-4.74185437e-01 6.67041004e-01 -5.85320473e-01 8.54148567e-01
5.62143624e-01 -6.60340607e-01 6.07608676e-01 -4.43115026e-01
3.36775094e-01 5.32595336e-01 -4.79941726e-01 1.70639563e+00
-9.15086642e-02 -2.18338240e-02 -8.07800591e-01 -1.14436281e+00
7.53892720e-01 3.46310407e-01 1.03115845e+00 -6.03341460e-01
-2.36517042e-01 1.78823203e-01 6.42868429e-02 -3.36559564e-01
1.29403695e-01 -2.74860505e-02 3.32820237e-01 1.88719630e-01
2.73107946e-01 3.41983885e-01 2.64307112e-01 3.63979638e-01
1.26860285e+00 1.24534935e-01 6.13063574e-01 -9.30742845e-02
5.90394318e-01 9.58014727e-02 4.23596591e-01 1.84680566e-01
3.24061692e-01 1.26666933e-01 3.16841543e-01 -5.97193182e-01
-6.35755241e-01 -7.42282867e-01 -4.88210708e-01 1.13743961e+00
-3.77343267e-01 -9.49447572e-01 -1.10036485e-01 -8.67184997e-01
3.81817460e-01 4.32912052e-01 -9.82891619e-01 -5.07385910e-01
-1.23703323e-01 -1.10913813e+00 5.88677406e-01 5.93753338e-01
-3.10819596e-01 -9.26999152e-01 1.89518064e-01 2.03725696e-01
1.48502946e-01 -7.59159505e-01 -1.12465136e-02 6.08699143e-01
-8.08911681e-01 -1.76673937e+00 -4.17553961e-01 -6.74039543e-01
6.98710978e-01 -1.47405922e-01 1.22845733e+00 7.97474831e-02
-4.46604341e-01 7.05731586e-02 -4.17301089e-01 -4.33479905e-01
-3.52001667e-01 6.09773695e-02 4.57551852e-02 -2.69817829e-01
6.38526678e-01 -5.53507924e-01 -5.06608129e-01 1.66931987e-01
-9.73365426e-01 -1.66315660e-02 5.56519032e-01 9.58536625e-01
1.13849854e+00 -5.04594266e-01 1.03186476e+00 -1.40005982e+00
4.88242775e-01 -9.12515461e-01 -4.32511955e-01 4.26143408e-01
-8.20329130e-01 2.75430828e-01 4.71294910e-01 -2.79019088e-01
-8.03239822e-01 4.77485120e-01 -1.87809542e-01 -2.01756477e-01
1.28268689e-01 1.23533666e+00 -7.54308850e-02 -9.67148766e-02
1.03117371e+00 5.75007461e-02 -1.27996961e-02 -7.41267860e-01
6.60668552e-01 8.48611355e-01 3.04660767e-01 -6.94167554e-01
2.59300977e-01 2.82966215e-02 2.77386963e-01 -2.89355665e-01
-9.36580420e-01 -6.02818012e-01 -3.50961685e-01 5.61440170e-01
5.99907279e-01 -1.04005444e+00 -7.31593430e-01 -1.02536082e-01
-7.35281348e-01 -1.40369693e-02 -2.03490868e-01 6.25506818e-01
-6.20426416e-01 3.64268690e-01 -5.50231040e-01 -2.50732750e-01
-6.04183793e-01 -1.23619211e+00 8.10843647e-01 9.22050104e-02
-1.69892713e-01 -1.10644054e+00 6.37542605e-01 5.51458955e-01
2.99395561e-01 4.70900297e-01 1.46503544e+00 -1.29568827e+00
-3.56236219e-01 -1.07383363e-01 -5.69868796e-02 1.70315191e-01
6.72125578e-01 -1.16981655e-01 -9.40240860e-01 -3.40069294e-01
-7.58335888e-01 -7.77836561e-01 1.22482479e+00 2.38505244e-01
1.45544004e+00 -3.00646871e-01 -7.09137082e-01 6.66652679e-01
1.40401006e+00 8.64363313e-02 8.88103604e-01 9.61809754e-02
8.30563605e-01 3.83338094e-01 5.25087237e-01 5.01775086e-01
-3.64682637e-02 7.58889377e-01 4.64877486e-01 -1.87834799e-01
3.74151580e-03 -3.29678863e-01 1.84474774e-02 4.72455531e-01
-5.42374432e-01 -1.95656028e-02 -8.29585135e-01 3.53839189e-01
-1.90049982e+00 -1.00371003e+00 9.74716097e-02 2.23301649e+00
1.67916632e+00 -3.40432644e-01 -5.55364834e-03 -3.13455105e-01
4.11055565e-01 -2.63648868e-01 -8.33651483e-01 -2.29678184e-01
-1.65088519e-01 8.95881832e-01 6.08364463e-01 3.36939573e-01
-1.12440515e+00 9.47779119e-01 6.75134182e+00 8.70451927e-01
-8.80833507e-01 -2.42705256e-01 6.47608280e-01 -4.97876704e-02
-4.03847188e-01 -1.06164224e-01 -7.19192743e-01 3.27364206e-01
1.68049932e+00 -3.95882308e-01 3.29223007e-01 7.52709627e-01
3.28153074e-02 2.30358005e-01 -1.61645365e+00 9.90401864e-01
-9.93705243e-02 -2.04590774e+00 4.59200025e-01 3.04854512e-01
8.94179821e-01 3.22160006e-01 1.71996236e-01 4.82529849e-01
5.32983422e-01 -1.70983636e+00 -5.63065410e-01 6.96148455e-01
1.11748183e+00 -8.56606007e-01 7.00887859e-01 -9.59256813e-02
-7.71222115e-01 1.63482323e-01 -5.25566876e-01 3.31221491e-01
-3.03092986e-01 5.81861377e-01 -1.26989496e+00 1.13484907e+00
2.75611103e-01 1.63614857e+00 -8.02539229e-01 1.05667269e+00
-1.24588180e-02 5.73736548e-01 -2.74755768e-02 2.42697865e-01
4.61397413e-03 7.90581033e-02 -9.84425098e-02 1.12277544e+00
-2.07363427e-01 1.27080366e-01 1.39052898e-01 4.57960755e-01
-6.43360317e-01 4.54566598e-01 -7.31890619e-01 -7.41418779e-01
4.00485486e-01 1.05917537e+00 9.74145830e-02 -5.46107829e-01
-4.72048849e-01 2.90319592e-01 5.35085440e-01 3.24316859e-01
-8.90291691e-01 -3.01075488e-01 9.06458735e-01 -1.33915350e-01
1.63869455e-01 3.69845152e-01 2.24286124e-01 -1.02394772e+00
-6.31385207e-01 -1.37631047e+00 7.80472636e-01 -5.51397502e-01
-1.85866225e+00 3.84972095e-01 -2.78405473e-02 -1.08272302e+00
-2.92769909e-01 -9.93110895e-01 1.15997769e-01 1.02873325e+00
-1.59320736e+00 -1.19634712e+00 1.99537948e-01 7.55399585e-01
2.01806888e-01 -5.37729740e-01 1.58297265e+00 6.62899792e-01
-7.22756743e-01 6.21372581e-01 5.73498130e-01 -8.32087472e-02
1.08282340e+00 -9.58011687e-01 6.52159527e-02 -9.65431705e-02
3.10198396e-01 1.16493952e+00 3.89170319e-01 -8.88902426e-01
-1.61819363e+00 -1.36110306e+00 5.44682443e-01 -8.70841801e-01
6.27431154e-01 6.25144690e-02 -1.13566566e+00 7.86934495e-01
-2.29979623e-02 1.96742117e-01 1.63099647e+00 4.70210254e-01
-6.90204859e-01 2.80957788e-01 -1.17327368e+00 9.30922572e-03
1.09228992e+00 -5.65180719e-01 -6.48909867e-01 1.00588751e+00
4.54215616e-01 -4.81524408e-01 -1.38371229e+00 7.17350781e-01
4.76424336e-01 -3.42914879e-01 1.33960652e+00 -1.85820901e+00
6.72452807e-01 -1.51586249e-01 -4.18048948e-01 -1.25460804e+00
-4.36287075e-01 -3.25435072e-01 -4.96189207e-01 8.13399613e-01
1.01259136e+00 -6.23556912e-01 7.66127408e-01 2.88034409e-01
-1.85177699e-01 -1.33345342e+00 -7.63076305e-01 -3.81615251e-01
2.45467827e-01 1.32908940e-01 6.13273561e-01 1.21393239e+00
5.10272086e-01 5.19233167e-01 -4.35171872e-01 -6.34213036e-04
3.30503434e-01 1.42628714e-01 5.49515903e-01 -1.25217164e+00
-4.54994649e-01 1.78692788e-02 -5.68745673e-01 -6.57252133e-01
2.30956718e-01 -1.50872159e+00 -4.33212638e-01 -1.65058994e+00
8.27041388e-01 -1.86082453e-01 -9.37137842e-01 1.01499057e+00
-2.16886997e-01 1.55514807e-01 -5.15782833e-01 3.17032844e-01
-7.37341642e-01 5.02350628e-01 1.05760479e+00 -5.14483631e-01
-4.42557260e-02 -3.27750415e-01 -1.12797797e+00 3.10262650e-01
5.97698450e-01 -5.26370883e-01 -6.74260557e-01 -3.20591107e-02
1.85483366e-01 -3.26480985e-01 3.98291014e-02 -4.04407591e-01
-4.00826260e-02 -4.56996679e-01 7.98823178e-01 -6.36906803e-01
4.13768113e-01 -6.40204608e-01 4.13773894e-01 4.73882377e-01
-6.29684091e-01 -2.31236011e-01 4.24977124e-01 8.92854035e-01
-9.71142203e-02 4.03404422e-02 6.92408025e-01 -1.43052638e-02
-5.54786563e-01 4.82254088e-01 1.87966228e-02 -1.31118540e-02
1.01196206e+00 8.25710893e-02 -7.23476171e-01 -4.93352115e-02
-8.89969647e-01 1.99642871e-02 1.83078647e-01 3.97342503e-01
7.15612769e-01 -1.22051620e+00 -6.65812194e-01 6.05882220e-02
6.70466781e-01 -4.72234637e-01 1.15900002e-01 8.17605793e-01
-2.41241977e-01 9.26841021e-01 -2.92016655e-01 -3.51543605e-01
-1.30577779e+00 9.17174101e-01 2.75600523e-01 -4.99892771e-01
-2.87003785e-01 6.74319983e-01 -2.01467401e-03 -4.11002249e-01
1.19228117e-01 -1.30879059e-01 -5.08348703e-01 -7.55043626e-02
6.58159912e-01 -9.41571891e-02 2.66268879e-01 -5.29422641e-01
-7.96310067e-01 2.45273158e-01 -6.04070246e-01 7.29391277e-01
1.88407218e+00 7.59851754e-01 -8.17283317e-02 -1.21499866e-01
1.33658278e+00 2.61575226e-02 -5.64020455e-01 -3.34402621e-01
1.32377461e-01 -3.75148207e-01 -2.41000473e-01 -1.35572803e+00
-8.79463136e-01 3.97680074e-01 6.41081393e-01 -5.25318623e-01
7.65756547e-01 1.69089332e-01 3.24707806e-01 8.63544345e-01
2.70788014e-01 -6.68096483e-01 1.95500087e-02 4.90567058e-01
8.13292980e-01 -1.33852112e+00 5.83426297e-01 -5.51850140e-01
-5.48832715e-01 7.79806614e-01 4.63041723e-01 3.14996511e-01
5.08791149e-01 -1.57505929e-01 -2.49684408e-01 -6.00225091e-01
-1.02982998e+00 -1.45913750e-01 5.40488422e-01 8.95504534e-01
1.04363239e+00 3.85740679e-03 -3.03408682e-01 7.18293965e-01
2.74925172e-01 2.69000798e-01 1.23467930e-01 1.03667688e+00
-2.09841132e-01 -1.73225605e+00 1.14944406e-01 6.89727426e-01
-5.73654532e-01 -5.55869102e-01 -8.93608451e-01 5.01837611e-01
-1.34566784e-01 7.72269905e-01 -6.07149303e-01 -3.27699214e-01
4.96224493e-01 1.15916222e-01 7.28164375e-01 -9.98139262e-01
-4.20883328e-01 -4.83900040e-01 4.84652519e-01 -4.67122883e-01
-6.94485903e-01 -4.89285290e-02 -1.43786848e+00 -1.83867797e-01
-3.32482904e-01 5.04997790e-01 3.94971758e-01 7.31119633e-01
1.00651228e+00 5.63065648e-01 4.02658343e-01 -4.48836088e-01
-6.71898842e-01 -1.06804335e+00 -2.72506446e-01 6.77103460e-01
-4.44145538e-02 -9.33233261e-01 1.76835954e-01 2.43234128e-01] | [5.153172016143799, 5.887620449066162] |
4e89c922-41ac-4ec6-a856-db662f0e045d | deep-multimodal-transfer-learned-regression | 2006.09310 | null | https://arxiv.org/abs/2006.09310v1 | https://arxiv.org/pdf/2006.09310v1.pdf | Deep Multimodal Transfer-Learned Regression in Data-Poor Domains | In many real-world applications of deep learning, estimation of a target may rely on various types of input data modes, such as audio-video, image-text, etc. This task can be further complicated by a lack of sufficient data. Here we propose a Deep Multimodal Transfer-Learned Regressor (DMTL-R) for multimodal learning of image and feature data in a deep regression architecture effective at predicting target parameters in data-poor domains. Our model is capable of fine-tuning a given set of pre-trained CNN weights on a small amount of training image data, while simultaneously conditioning on feature information from a complimentary data mode during network training, yielding more accurate single-target or multi-target regression than can be achieved using the images or the features alone. We present results using phase-field simulation microstructure images with an accompanying set of physical features, using pre-trained weights from various well-known CNN architectures, which demonstrate the efficacy of the proposed multimodal approach. | ['Ulisses Braga-Neto', 'Brian Sadler', 'Vahid Attari', 'Raymundo Arroyave', 'Mulugeta Haile', 'Levi McClenny'] | 2020-06-16 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [ 3.72965097e-01 -2.26596445e-01 -6.51005539e-04 -5.62033355e-01
-1.05916798e+00 -1.07114144e-01 6.91335976e-01 1.67169169e-01
-6.82744324e-01 8.01612556e-01 -3.49357903e-01 3.95274758e-02
-4.64670748e-01 -6.50088847e-01 -8.08463991e-01 -1.08414447e+00
4.40949574e-02 6.82909787e-01 1.95258722e-01 -3.64315212e-01
3.84048164e-01 5.23971856e-01 -1.46655858e+00 2.48764351e-01
4.11752671e-01 1.48243880e+00 4.42345679e-01 7.09738195e-01
9.45556238e-02 6.79862797e-01 -2.47240543e-01 5.20170778e-02
-2.62609366e-02 -2.25039169e-01 -3.97441566e-01 7.92416558e-02
2.43142799e-01 -1.75912932e-01 -3.14604074e-01 7.66973197e-01
5.39956152e-01 2.04024874e-02 1.04634285e+00 -8.83484185e-01
-2.70274937e-01 2.06674576e-01 -5.30405939e-01 -5.78283407e-02
-3.90651748e-02 2.81120986e-01 6.88581288e-01 -8.42354536e-01
4.81755793e-01 9.17872548e-01 4.17069018e-01 4.01567876e-01
-1.61791313e+00 -6.96568370e-01 -3.17371637e-01 1.04304805e-01
-1.17451751e+00 -3.65437597e-01 1.07467556e+00 -6.19547546e-01
7.55095840e-01 -1.77110299e-01 2.47992739e-01 1.13137174e+00
4.02596146e-01 1.76516548e-01 1.15205717e+00 -5.76608360e-01
1.51000395e-01 2.36771241e-01 -1.18318036e-01 6.84315860e-01
-1.41070619e-01 2.44530201e-01 -7.12778568e-01 -1.32278189e-01
1.04980040e+00 -1.03113450e-01 -3.44567925e-01 -5.07044375e-01
-1.33123875e+00 8.85299087e-01 5.94696701e-01 4.40998048e-01
-3.43759328e-01 7.95346200e-02 2.20877334e-01 6.05247021e-01
4.48095679e-01 4.83493507e-01 -6.46950603e-01 1.12271354e-01
-8.73177707e-01 -3.42741236e-02 3.12696844e-01 5.52820981e-01
1.35643995e+00 1.30171657e-01 7.79273212e-02 8.04990530e-01
3.09665591e-01 7.97182500e-01 4.23373997e-01 -6.78759575e-01
4.51233059e-01 4.27512646e-01 2.05602914e-01 -6.92722201e-01
-6.90157175e-01 -4.12337095e-01 -1.02995431e+00 7.16801167e-01
6.57759845e-01 -2.50421554e-01 -1.08592129e+00 1.75168252e+00
2.77594626e-01 -1.21426463e-01 7.33266622e-02 1.09712386e+00
7.23054528e-01 6.40769303e-01 -2.20579691e-02 -1.62471846e-01
8.61151814e-01 -5.14587224e-01 -2.55461365e-01 -4.45008315e-02
4.65439141e-01 -8.57528925e-01 8.80865395e-01 6.36907935e-01
-9.35854077e-01 -7.15651155e-01 -1.07440925e+00 -2.02972479e-02
-3.65385801e-01 4.60826993e-01 2.53042907e-01 3.12913060e-01
-7.97441423e-01 8.75773191e-01 -9.26508904e-01 -2.19416171e-01
2.91800916e-01 1.02564538e+00 -7.06323624e-01 -6.00112528e-02
-7.78916776e-01 8.83809865e-01 2.09208906e-01 3.19177598e-01
-1.22917891e+00 -7.77837753e-01 -6.72983170e-01 -3.83176580e-02
1.46867186e-01 -5.82592130e-01 8.35531294e-01 -1.17462850e+00
-1.73404706e+00 7.43355930e-01 -1.40420953e-03 -3.32098454e-02
2.97606021e-01 -1.92146152e-01 -3.93067688e-01 3.21816981e-01
-2.27476954e-01 7.11799324e-01 1.42111456e+00 -1.29583776e+00
-2.10855365e-01 -4.35125023e-01 -5.05063497e-02 -1.02876760e-01
-4.42775458e-01 -1.78923294e-01 -2.31862329e-02 -2.03185648e-01
-1.03090614e-01 -7.38886595e-01 -4.41063464e-01 3.49535346e-02
-5.33756375e-01 1.98826566e-01 1.05625665e+00 -3.27027917e-01
5.74903965e-01 -2.05035067e+00 5.75150311e-01 4.84276652e-01
4.02694851e-01 1.65753756e-02 -4.47912663e-01 6.39566123e-01
-2.98071563e-01 -3.58048171e-01 -3.27411562e-01 -1.66731715e-01
-2.88493335e-01 3.91962081e-02 8.68306533e-02 7.33940482e-01
7.59910226e-01 6.95589781e-01 -5.18300712e-01 -4.29362148e-01
5.92055321e-01 6.56696916e-01 -2.46283770e-01 4.94342059e-01
-3.23043495e-01 1.09464240e+00 -5.87861240e-01 4.35841799e-01
5.34485400e-01 -4.69466239e-01 6.37334511e-02 -5.66987097e-01
-1.23804592e-01 -7.83280581e-02 -1.06937599e+00 1.82303059e+00
-7.77862251e-01 7.14774370e-01 2.13912115e-01 -1.64123285e+00
1.01834381e+00 2.02048540e-01 6.80422366e-01 -7.75592506e-01
4.36047167e-01 2.81098992e-01 2.44840369e-01 -6.52370512e-01
1.50987208e-01 -2.87465245e-01 7.10340589e-02 5.01873493e-01
7.13469148e-01 -1.05681889e-01 -1.87614501e-01 -1.67996526e-01
1.07261300e+00 -1.79954655e-02 -1.92473829e-01 -3.15848649e-01
6.77224159e-01 -9.92451757e-02 -4.25939746e-02 6.98697627e-01
3.23254466e-01 6.74458027e-01 3.53289366e-01 -4.45290625e-01
-1.38448322e+00 -8.58480692e-01 -3.63805145e-01 9.43772614e-01
-3.51608060e-02 6.87715784e-02 -3.44305724e-01 -3.46265316e-01
-2.19953939e-01 1.10383041e-01 -7.27643251e-01 -1.99895144e-01
-6.29085481e-01 -9.03216660e-01 2.88945138e-01 1.30230084e-01
2.12086201e-01 -1.04715669e+00 -4.22653735e-01 1.68608353e-01
1.91397578e-01 -9.70341027e-01 9.36063826e-02 7.47462630e-01
-8.91041636e-01 -1.08935189e+00 -7.92513609e-01 -7.53411829e-01
7.35637784e-01 -1.64222255e-01 1.00486267e+00 7.24450201e-02
-3.36377263e-01 3.99781823e-01 -3.55957597e-02 -9.44376457e-03
-3.19843978e-01 2.64297724e-01 -2.64890380e-02 4.26362395e-01
1.61458645e-02 -9.26912725e-01 -4.72012758e-01 2.39377320e-01
-8.55679691e-01 -2.44399421e-02 9.76345181e-01 1.24005020e+00
4.57729697e-01 -1.78100139e-01 6.76801443e-01 -7.11634934e-01
2.45307595e-01 -4.02892977e-01 -7.71697521e-01 3.69568728e-02
-2.87436694e-01 4.05460000e-01 5.98402917e-01 -9.48089004e-01
-9.67543781e-01 1.96763396e-01 -1.54571593e-01 -5.74670136e-01
-3.01143736e-01 6.94835722e-01 -5.27205365e-03 -6.32477582e-01
7.24110246e-01 6.36657104e-02 2.88294166e-01 -3.82996649e-01
7.00599551e-02 4.18886274e-01 4.52960134e-01 -8.06611359e-01
9.42213893e-01 2.66788065e-01 5.74363649e-01 -9.70750213e-01
-4.89530772e-01 -2.95337886e-01 -8.47050965e-01 -4.68575716e-01
7.76396036e-01 -7.85459042e-01 -8.39381814e-01 5.93280792e-01
-9.60707486e-01 -5.46233654e-01 5.81438355e-02 7.78334260e-01
-5.08123517e-01 9.16547701e-02 -6.11069739e-01 -5.86911440e-01
-3.29329371e-02 -1.35398388e+00 1.32140911e+00 8.06037188e-02
1.50444970e-01 -1.30948365e+00 2.48362683e-02 2.02134952e-01
6.69911683e-01 2.33253092e-01 1.23455644e+00 -6.22693479e-01
-7.14682698e-01 -2.74545699e-01 -1.66541740e-01 3.18514287e-01
8.97500366e-02 1.93814933e-02 -1.09260464e+00 -3.55942070e-01
2.31893584e-02 -9.93031502e-01 8.83840799e-01 5.71470380e-01
1.05820906e+00 1.55468494e-01 -2.42933556e-01 4.35116470e-01
1.65469837e+00 -1.56220794e-01 5.15767157e-01 1.77852094e-01
8.15476537e-01 5.54973364e-01 4.40446645e-01 4.64074761e-01
-4.59474549e-02 7.02992857e-01 6.75936282e-01 -4.63322431e-01
2.71149486e-01 1.95311725e-01 1.96465150e-01 6.92394555e-01
-3.75559658e-01 3.54498215e-02 -8.16395760e-01 2.49826103e-01
-1.81078005e+00 -6.22988939e-01 -2.42606416e-01 2.43563962e+00
7.44287610e-01 1.13107450e-01 5.67757189e-02 9.34439376e-02
5.80862880e-01 -1.69595718e-01 -6.66490853e-01 1.91238672e-01
-1.91351965e-01 6.24816239e-01 3.68454814e-01 5.31638384e-01
-1.05289996e+00 5.38312793e-01 6.19463873e+00 9.57450032e-01
-1.81596386e+00 1.30808810e-02 5.74126601e-01 -3.37306224e-02
-3.02162737e-01 -2.60951072e-01 -4.07812148e-01 1.37730181e-01
1.05702877e+00 3.70391846e-01 4.41123605e-01 4.12888139e-01
1.39040709e-01 -3.37616444e-01 -1.44787407e+00 1.05261803e+00
-5.82774207e-02 -1.37819231e+00 -8.17126855e-02 1.54954240e-01
4.89106923e-01 1.60135210e-01 3.43083769e-01 2.11625680e-01
-1.09352149e-01 -1.26672554e+00 3.95644754e-01 7.29033947e-01
8.39375973e-01 -5.43619752e-01 6.32482469e-01 4.03290719e-01
-6.89876199e-01 -1.11213185e-01 -3.26147169e-01 1.39786005e-01
-4.36684787e-02 6.46332324e-01 -5.69754183e-01 6.32094026e-01
5.89843810e-01 7.42662847e-01 -4.36114341e-01 8.21433187e-01
3.70623648e-01 5.85089564e-01 -4.53328073e-01 5.53090535e-02
2.50899464e-01 -2.26106510e-01 2.96401083e-01 9.58073914e-01
4.20960516e-01 -2.03104258e-01 7.19454931e-03 9.47031200e-01
8.38502273e-02 1.48217827e-02 -5.32922983e-01 -2.85139978e-02
-1.27807394e-01 1.63575816e+00 -4.05510306e-01 -8.71554774e-04
-3.99935186e-01 4.67730820e-01 4.55570489e-01 6.71476245e-01
-6.85104609e-01 -1.61025181e-01 3.52412283e-01 2.04361111e-01
4.52407748e-01 -4.01255131e-01 -1.52980924e-01 -1.07455730e+00
-1.46433279e-01 -4.73183841e-01 -9.56235379e-02 -8.90754938e-01
-1.34756207e+00 5.29864252e-01 3.52372937e-02 -1.34719157e+00
-2.92364568e-01 -1.06914067e+00 -7.25116432e-01 9.93994653e-01
-1.59688139e+00 -1.38079166e+00 -1.51639134e-01 8.58712852e-01
1.57861821e-02 -4.80585128e-01 8.51931572e-01 4.73414570e-01
-4.15039808e-01 3.32789570e-01 2.17801854e-01 -4.01987359e-02
8.87447238e-01 -1.04120052e+00 -5.09371042e-01 1.80562645e-01
-7.22564012e-03 2.64239848e-01 7.68211067e-01 -1.88240781e-01
-1.54019499e+00 -7.88712621e-01 3.01181346e-01 -1.53761655e-01
7.45311022e-01 -4.41026092e-01 -8.71532083e-01 4.52054799e-01
2.93476284e-01 1.77531734e-01 4.47716951e-01 1.25242054e-01
-1.99182719e-01 -3.86593550e-01 -9.70741510e-01 1.84556127e-01
3.64723831e-01 -6.68626964e-01 -1.77404940e-01 3.63692313e-01
1.89997777e-01 -4.07063723e-01 -1.13811052e+00 4.74194080e-01
4.87092257e-01 -9.26567912e-01 1.03116143e+00 -7.86295474e-01
8.74174476e-01 -2.33842328e-01 -1.85892507e-01 -1.40716565e+00
-1.41051129e-01 -3.00912112e-01 3.42725009e-01 1.02102160e+00
7.06150413e-01 -3.43256742e-01 7.75613964e-01 4.33147162e-01
-1.27587661e-01 -7.14205682e-01 -1.14218128e+00 -2.91410923e-01
9.38967913e-02 -4.51397210e-01 1.86475217e-02 7.88176477e-01
-1.91129997e-01 5.42267978e-01 -6.17525101e-01 4.57266062e-01
7.45662630e-01 2.14566849e-02 8.11576366e-01 -1.28637278e+00
-4.21161503e-01 -2.59232819e-01 -4.90658849e-01 -6.58810258e-01
2.74500877e-01 -6.52229667e-01 1.03745535e-02 -1.01333880e+00
1.95413679e-01 -5.28543532e-01 -4.63818341e-01 4.66518700e-01
2.62949407e-01 4.90354687e-01 -1.60400838e-01 1.50676638e-01
-4.35311168e-01 7.64915287e-01 1.33965445e+00 -4.19459403e-01
4.34813462e-02 -3.53233293e-02 -8.36862102e-02 4.55812603e-01
5.70291996e-01 -5.56629598e-01 -2.15307146e-01 -2.66191036e-01
1.73655272e-01 3.94329637e-01 7.21200883e-01 -1.01358747e+00
2.71752477e-01 -1.44273117e-01 7.64559090e-01 -2.61680037e-01
8.02025795e-01 -1.01438177e+00 5.85089698e-02 1.94826141e-01
-4.67722803e-01 -2.78115720e-01 1.86425030e-01 3.63311112e-01
-4.08976763e-01 -2.65969992e-01 8.81750405e-01 2.64644381e-02
-4.86497462e-01 1.89643845e-01 -3.94773006e-01 -2.72495717e-01
7.10778773e-01 2.60911658e-02 -3.31765920e-01 -4.45958883e-01
-9.76300657e-01 -4.21949811e-02 2.27834582e-01 1.03031293e-01
6.03321671e-01 -1.35604119e+00 -4.97813344e-01 3.24960291e-01
1.56082094e-01 -4.62358035e-02 3.02596688e-01 1.06098580e+00
-2.35781774e-01 1.50374621e-01 -5.62008321e-01 -9.97474909e-01
-9.93697882e-01 4.23096091e-01 5.97380757e-01 -2.45311081e-01
-3.25642288e-01 4.67055470e-01 3.53644937e-01 -5.89549541e-01
-8.43307525e-02 -2.56591260e-01 -2.43738189e-01 -2.33429700e-01
2.06657559e-01 -6.53191581e-02 2.38643795e-01 -6.21049881e-01
-2.49344092e-02 9.30631697e-01 3.17551494e-02 -2.85640508e-01
1.66800916e+00 6.69196025e-02 -6.09085001e-02 6.30763829e-01
1.32285655e+00 -3.68061036e-01 -1.37163770e+00 -3.62324566e-01
-2.28359953e-01 -6.20180182e-03 2.41524711e-01 -6.96031511e-01
-1.23138332e+00 1.18689287e+00 8.07718992e-01 -1.10765688e-01
1.17306149e+00 6.52516782e-02 4.05245930e-01 6.21195436e-01
2.37437606e-01 -1.04758465e+00 5.09952784e-01 1.69347063e-01
8.48790050e-01 -1.70645022e+00 -1.89972863e-01 -3.10992431e-02
-3.71059895e-01 1.59944820e+00 5.82863390e-01 -2.42131129e-01
9.38988447e-01 2.74975419e-01 1.73255280e-01 -3.54838938e-01
-7.64267802e-01 -8.70353132e-02 5.42721510e-01 6.29862785e-01
3.89345109e-01 -3.34865838e-01 3.53010505e-01 3.06056172e-01
3.13881457e-01 -8.86069760e-02 2.95678973e-01 7.30842531e-01
-3.75858873e-01 -1.28677011e+00 -3.73663247e-01 5.51408410e-01
-3.67037833e-01 4.92427424e-02 -3.44287269e-02 8.86537969e-01
6.71032351e-03 6.67807937e-01 1.06857665e-01 -4.08552527e-01
8.65831450e-02 -1.52831122e-01 7.33283401e-01 -3.16620082e-01
-4.93254334e-01 3.91213626e-01 6.55781105e-02 -3.47786635e-01
-8.31427932e-01 -4.34090972e-01 -8.93233299e-01 -3.87275852e-02
-3.23506445e-01 -1.83941826e-01 6.56349242e-01 1.40539253e+00
7.68607308e-04 4.76351947e-01 6.29737556e-01 -1.58651328e+00
-3.60577643e-01 -1.21516967e+00 -6.53061509e-01 3.38075280e-01
7.20379174e-01 -9.42547858e-01 -2.33256757e-01 1.96239762e-02] | [10.119160652160645, -0.04922923445701599] |
e6ee064a-b3d2-4ae9-b40b-7f7935fdeb6b | learning-steerable-filters-for-rotation | 1711.07289 | null | http://arxiv.org/abs/1711.07289v3 | http://arxiv.org/pdf/1711.07289v3.pdf | Learning Steerable Filters for Rotation Equivariant CNNs | In many machine learning tasks it is desirable that a model's prediction
transforms in an equivariant way under transformations of its input.
Convolutional neural networks (CNNs) implement translational equivariance by
construction; for other transformations, however, they are compelled to learn
the proper mapping. In this work, we develop Steerable Filter CNNs (SFCNNs)
which achieve joint equivariance under translations and rotations by design.
The proposed architecture employs steerable filters to efficiently compute
orientation dependent responses for many orientations without suffering
interpolation artifacts from filter rotation. We utilize group convolutions
which guarantee an equivariant mapping. In addition, we generalize He's weight
initialization scheme to filters which are defined as a linear combination of a
system of atomic filters. Numerical experiments show a substantial enhancement
of the sample complexity with a growing number of sampled filter orientations
and confirm that the network generalizes learned patterns over orientations.
The proposed approach achieves state-of-the-art on the rotated MNIST benchmark
and on the ISBI 2012 2D EM segmentation challenge. | ['Maurice Weiler', 'Martin Storath', 'Fred A. Hamprecht'] | 2017-11-20 | learning-steerable-filters-for-rotation-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Weiler_Learning_Steerable_Filters_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Weiler_Learning_Steerable_Filters_CVPR_2018_paper.pdf | cvpr-2018-6 | ['rotated-mnist', 'colorectal-gland-segmentation', 'breast-tumour-classification', 'multi-tissue-nucleus-segmentation'] | ['computer-vision', 'medical', 'medical', 'medical'] | [ 1.70129582e-01 1.94796398e-01 1.64951235e-01 -6.92937553e-01
-5.42817898e-02 -4.79957700e-01 6.88522279e-01 -6.54668808e-01
-7.16618955e-01 4.43630874e-01 1.76923588e-01 -2.13072866e-01
-3.22991498e-02 -8.65726173e-01 -1.27871513e+00 -7.46540964e-01
2.63135463e-01 3.33930999e-01 3.19088280e-01 -4.57608104e-01
-1.07529186e-01 8.23181033e-01 -1.14537239e+00 3.20861101e-01
7.13944614e-01 7.47246981e-01 1.62689716e-01 6.39208317e-01
3.32155466e-01 5.95266581e-01 -1.50454149e-01 -3.94965529e-01
6.07081294e-01 -3.23440880e-02 -9.17934000e-01 1.33448094e-01
9.69445109e-01 -3.21385980e-01 -7.81713367e-01 1.33820319e+00
1.96250632e-01 3.26767981e-01 7.76519895e-01 -5.95061064e-01
-9.38975275e-01 6.71953082e-01 -2.59408295e-01 1.90866351e-01
-3.63503337e-01 -1.86401874e-01 8.50276411e-01 -1.06806493e+00
8.11821640e-01 1.27496469e+00 7.48436809e-01 6.23767555e-01
-1.35183430e+00 -3.74545127e-01 3.95016730e-01 2.24376529e-01
-1.25128186e+00 -2.53063381e-01 6.51075780e-01 -2.92898893e-01
1.01153064e+00 3.48684102e-01 5.38874745e-01 9.86312926e-01
2.15041861e-01 6.44340992e-01 9.56717551e-01 -3.47983837e-01
1.31073818e-01 -1.70910321e-02 -3.38942674e-03 7.82251716e-01
1.75549716e-01 -5.52787594e-02 -3.74460846e-01 2.71688908e-01
1.38425875e+00 7.03782961e-02 -6.97109461e-01 -4.80017841e-01
-1.35610199e+00 7.34604836e-01 8.96168530e-01 1.48251027e-01
-3.49608511e-01 2.16655925e-01 5.04248019e-04 4.17798311e-01
4.47447151e-01 3.08458239e-01 -4.24130410e-01 5.24949312e-01
-7.33794570e-01 1.47088408e-01 6.54320896e-01 1.14038682e+00
6.89148247e-01 4.18992221e-01 2.54208390e-02 6.52692258e-01
2.71980520e-02 4.63246614e-01 3.67147446e-01 -8.69475603e-01
2.31925040e-01 2.44495302e-01 3.80570143e-02 -8.72890830e-01
-6.17562711e-01 -9.07147408e-01 -1.15991914e+00 2.67645363e-02
5.04175723e-01 9.77513567e-02 -1.24073458e+00 1.89014649e+00
2.42116615e-01 2.42919490e-01 -9.15133879e-02 1.21499836e+00
5.64318597e-01 4.37761068e-01 -2.69271672e-01 1.38003886e-01
1.13558137e+00 -7.36895323e-01 -6.34951532e-01 -1.61704332e-01
1.83720008e-01 -8.47061098e-01 8.45958054e-01 2.50932097e-01
-1.14988685e+00 -8.04816604e-01 -1.25189197e+00 -1.54696763e-01
-2.47948378e-01 2.48287797e-01 5.97334325e-01 5.38865507e-01
-1.22427750e+00 7.43022501e-01 -1.20599937e+00 -1.54315501e-01
3.47343296e-01 5.78798771e-01 -4.78661954e-01 2.22347870e-01
-1.00865912e+00 1.02462029e+00 3.24393094e-01 4.82456118e-01
-6.26708269e-01 -8.66680861e-01 -7.87622511e-01 5.41174076e-02
-7.39426315e-02 -8.46171081e-01 1.20364285e+00 -1.06437278e+00
-1.47592378e+00 7.61245370e-01 -1.15314759e-01 -8.40129018e-01
8.94838810e-01 -2.89224267e-01 -3.92643601e-01 2.75196314e-01
-2.59407103e-01 1.00162172e+00 1.19489944e+00 -9.94471788e-01
-6.92491710e-01 -2.79392898e-01 1.79466933e-01 3.07223320e-01
-4.69997287e-01 -3.88647169e-01 -3.05210859e-01 -8.70856345e-01
7.54872441e-01 -1.05557561e+00 -4.12634492e-01 1.72111895e-02
-3.00520569e-01 6.19374141e-02 9.57436442e-01 -6.37412250e-01
6.51923835e-01 -1.97633410e+00 4.01900858e-01 3.49010617e-01
3.02797049e-01 1.40971290e-02 -8.62338617e-02 -1.69050291e-01
-2.06009626e-01 -2.76097029e-01 -1.72625691e-01 -1.63892075e-01
4.62314077e-02 3.44680429e-01 -4.54362690e-01 1.06159294e+00
3.51554185e-01 8.18238020e-01 -4.83739704e-01 -5.06192185e-02
4.23474669e-01 9.56405163e-01 -9.54753578e-01 1.16194068e-02
9.65149775e-02 5.36955595e-01 -1.93229675e-01 9.88525078e-02
9.42031801e-01 -1.79485664e-01 5.82225434e-02 -7.57344544e-01
-1.73990890e-01 2.21234754e-01 -1.29306948e+00 1.70021474e+00
-4.76707488e-01 7.91879952e-01 9.97488052e-02 -1.36553800e+00
9.93164778e-01 2.56658882e-01 4.01181489e-01 -3.30471545e-01
3.32180411e-01 2.58448988e-01 1.79185078e-01 -4.59853970e-02
5.52846193e-01 7.86726698e-02 3.50682497e-01 -9.28416755e-03
4.73769039e-01 -7.80318007e-02 -2.80713551e-02 -1.41026020e-01
7.36817420e-01 1.01706825e-01 2.20801141e-02 -8.34945917e-01
6.68142855e-01 -3.31534237e-01 6.49401188e-01 8.68300915e-01
2.47941643e-01 8.78847778e-01 -1.05050474e-01 -1.09831977e+00
-1.24571788e+00 -1.31249058e+00 -5.99635839e-01 7.86411166e-01
1.72788501e-01 2.03640044e-01 -1.04010510e+00 -3.82222563e-01
-3.00593704e-01 4.42165464e-01 -6.53784096e-01 -2.11994782e-01
-1.06937838e+00 -8.83483946e-01 3.79904926e-01 7.78618157e-01
9.74588454e-01 -8.30904365e-01 -6.68226779e-01 3.65400285e-01
-2.36408245e-02 -1.38917601e+00 -7.46082306e-01 2.65527785e-01
-8.45323026e-01 -9.68682647e-01 -9.05099094e-01 -1.14728439e+00
1.18381238e+00 3.24798942e-01 7.82664359e-01 -5.09961784e-01
-1.08571574e-01 2.93018311e-01 -1.69162154e-01 -6.24208860e-02
-1.53729931e-01 2.66211450e-01 3.01562816e-01 3.23526800e-01
2.60327429e-01 -7.42860258e-01 -7.50538707e-01 6.89897716e-01
-1.02304232e+00 2.91112006e-01 4.82573003e-01 1.00464320e+00
7.62343526e-01 -1.93561003e-01 2.57669121e-01 -8.64835024e-01
3.82665277e-01 1.53283089e-01 -6.90311611e-01 2.67092198e-01
-2.46778712e-01 3.76630962e-01 7.95129955e-01 -6.42122507e-01
-1.27447701e+00 2.98996806e-01 -2.12701201e-01 -3.84536177e-01
-1.49719298e-01 -2.04751212e-02 1.20092267e-02 -5.22018671e-01
8.34046900e-01 1.83769971e-01 -2.07547858e-01 -4.34414476e-01
5.73686361e-01 3.69930565e-01 9.71282125e-01 -7.75134861e-01
9.91992593e-01 1.04828751e+00 1.37683287e-01 -9.24792051e-01
-6.76145196e-01 -1.67419642e-01 -7.94586897e-01 -2.34017745e-01
1.03785169e+00 -9.40034330e-01 -8.26621115e-01 4.51193839e-01
-1.25933921e+00 -1.42640889e-01 -2.11760953e-01 7.50615537e-01
-6.27698720e-01 -2.00230833e-02 -7.60669529e-01 -1.00609936e-01
-4.22020078e-01 -1.38422012e+00 7.49871492e-01 3.22839946e-01
-6.54684007e-02 -9.44262147e-01 -3.67397159e-01 -1.79558396e-01
6.22754872e-01 3.17206196e-02 5.68452835e-01 -4.32081521e-01
-6.68746471e-01 -1.00854993e-01 -1.85605258e-01 5.57008088e-01
-9.66506600e-02 -1.94379315e-01 -1.02567077e+00 -4.76338983e-01
2.72208750e-01 1.40600681e-01 1.17831755e+00 7.35108912e-01
1.26040339e+00 -3.44715238e-01 7.61011168e-02 1.28758132e+00
1.07520020e+00 2.11831206e-03 4.80631381e-01 4.50985044e-01
8.97242725e-01 3.88789505e-01 -5.70538454e-02 7.39551242e-03
1.54726356e-01 5.38467586e-01 4.41665441e-01 -3.28280270e-01
-1.71843708e-01 -2.10896313e-01 5.41770197e-02 8.70566428e-01
-6.20471776e-01 2.80427039e-01 -4.46011096e-01 3.33778709e-01
-1.83035696e+00 -6.96368456e-01 -1.71500027e-01 2.05358052e+00
6.41579866e-01 2.03497007e-01 -4.41119760e-01 -2.58587394e-02
7.30653703e-01 1.53179333e-01 -4.48694408e-01 -4.53702003e-01
-4.21725124e-01 7.13256955e-01 1.12114096e+00 6.16416633e-01
-1.32562053e+00 1.03496933e+00 5.76543427e+00 6.33574069e-01
-1.39299321e+00 4.86636013e-02 4.71166402e-01 2.74399042e-01
-2.97039449e-01 -1.68889210e-01 -8.53463113e-01 -1.86846659e-01
3.64568323e-01 1.71149671e-01 5.46589732e-01 8.14650118e-01
7.07743391e-02 6.40303075e-01 -9.56103802e-01 7.07449198e-01
-1.67729601e-01 -1.64277124e+00 1.10602014e-01 -1.84010535e-01
1.08962321e+00 2.15145305e-01 3.96419734e-01 -2.04999015e-01
1.76234394e-01 -9.59616065e-01 1.05703068e+00 4.52125341e-01
5.30913353e-01 -7.75985479e-01 4.69854474e-01 1.58521444e-01
-1.13137126e+00 2.05031052e-01 -8.26517642e-01 1.29227132e-01
1.76383942e-01 4.06812787e-01 -8.58343065e-01 3.44505280e-01
8.36422801e-01 7.59000599e-01 -1.48649380e-01 8.18195999e-01
-3.22383583e-01 4.90582734e-01 -4.56704110e-01 1.48766980e-01
3.20937872e-01 -4.08477485e-01 6.37387574e-01 1.29719293e+00
3.35881561e-01 2.19322927e-03 -6.00076132e-02 9.72178400e-01
-3.70764047e-01 -1.42317683e-01 -2.60184318e-01 5.47727764e-01
2.48725072e-01 1.17611778e+00 -8.48946095e-01 -2.38702834e-01
-4.38184023e-01 9.92343783e-01 2.60417640e-01 8.11227262e-01
-7.87990212e-01 -2.57849008e-01 8.27833891e-01 -1.56059507e-02
5.92412233e-01 -4.70182329e-01 -2.69930333e-01 -1.33883762e+00
5.27499467e-02 -7.12757111e-01 4.69644321e-03 -4.61180598e-01
-9.74425673e-01 7.77401209e-01 -1.49562374e-01 -1.15058291e+00
6.56312704e-02 -1.04756558e+00 -3.72899830e-01 7.71648407e-01
-1.45069361e+00 -1.17356372e+00 -2.09920421e-01 7.50117362e-01
5.21398902e-01 -7.55153522e-02 6.66300893e-01 3.95804673e-01
3.66892503e-03 5.03633142e-01 -7.39028975e-02 3.99009824e-01
5.93138099e-01 -1.24373603e+00 6.39223278e-01 9.75228727e-01
3.83696586e-01 8.98525357e-01 8.47788155e-01 -1.80206001e-01
-1.30395484e+00 -1.20962226e+00 4.77147907e-01 1.43855074e-02
4.47402954e-01 -3.20252717e-01 -1.00029528e+00 1.06067240e+00
1.86039031e-01 4.90478516e-01 1.68220475e-02 -2.32552603e-01
-5.98767757e-01 -2.59844780e-01 -1.10524988e+00 6.68771088e-01
1.24656999e+00 -4.39047754e-01 -4.93638366e-01 2.37808391e-01
7.37065256e-01 -9.48383987e-01 -8.21858704e-01 6.28692985e-01
6.89034879e-01 -8.23667169e-01 1.24819386e+00 -8.58575583e-01
1.15836553e-01 -4.82326925e-01 -3.18095475e-01 -1.39874601e+00
-4.62526023e-01 -6.20307744e-01 1.90128669e-01 5.61776459e-01
2.92990685e-01 -9.13773954e-01 6.44392133e-01 3.21115524e-01
-4.90970701e-01 -4.36274081e-01 -1.03214562e+00 -6.91740453e-01
1.55925944e-01 -3.05602014e-01 5.03115892e-01 9.54864621e-01
-5.14699459e-01 -1.23202696e-01 -3.81410539e-01 6.48689687e-01
8.04500937e-01 -1.29928917e-01 4.24946874e-01 -8.56306911e-01
-4.02831882e-01 -4.40963268e-01 -5.94928443e-01 -1.46373188e+00
3.34855407e-01 -1.02862346e+00 -5.82297407e-02 -1.05880547e+00
-1.68101564e-01 -8.68758410e-02 -2.27576166e-01 2.15103611e-01
2.21117154e-01 3.52950692e-01 1.59919187e-02 1.60665270e-02
-9.85909402e-02 3.58253002e-01 1.60139942e+00 -2.06663474e-01
-1.66451067e-01 1.78028554e-01 -3.26514959e-01 9.84263361e-01
8.90198350e-01 -1.78490072e-01 -4.03236747e-01 -8.26379478e-01
1.47520617e-01 -3.45754921e-01 5.17529547e-01 -9.17978942e-01
3.64739299e-01 -3.82130779e-02 6.01335108e-01 -3.66233885e-01
3.01721841e-01 -8.06794643e-01 2.94781387e-01 4.51253325e-01
-4.45500642e-01 1.03611030e-01 -7.82572199e-03 4.12667274e-01
-2.01644331e-01 8.56316313e-02 1.10864019e+00 1.15877185e-02
-5.84184945e-01 5.65012693e-01 -1.19926482e-01 -1.17362514e-01
4.44215506e-01 -7.84862339e-02 -2.54668325e-01 -2.42163539e-01
-9.33509648e-01 -2.12063998e-01 2.68067002e-01 3.45791668e-01
6.76078856e-01 -1.37218940e+00 -5.55121064e-01 5.22057176e-01
-3.51708591e-01 4.22176749e-01 1.63346782e-01 9.17300224e-01
-9.59011912e-01 3.86298895e-01 -4.74521488e-01 -7.05716014e-01
-8.57647300e-01 -6.68302849e-02 7.57974625e-01 1.15866199e-01
-9.99401271e-01 9.93425608e-01 4.27904934e-01 -6.24926925e-01
2.02308148e-01 -6.92668438e-01 2.07457077e-02 -3.96329850e-01
4.00200993e-01 1.40403807e-01 3.63923311e-01 -7.22044170e-01
-8.61949623e-02 6.68384254e-01 -2.11668432e-01 -1.45146236e-01
1.61994064e+00 2.19853163e-01 -1.24877229e-01 -2.60950536e-01
1.33107519e+00 -4.10727561e-01 -1.71823394e+00 -4.51717615e-01
-3.25538784e-01 -3.30849916e-01 1.69425428e-01 -1.62199169e-01
-1.40303206e+00 9.31119025e-01 5.62255025e-01 -6.33088872e-02
1.10169256e+00 -1.71358034e-01 4.86840606e-01 7.55433977e-01
2.41194710e-01 -1.09553730e+00 -3.22511137e-01 5.70380747e-01
9.66296852e-01 -8.37074220e-01 -3.76678973e-01 -4.52534854e-01
-1.57591164e-01 1.56139421e+00 5.07480979e-01 -7.17219532e-01
7.76784837e-01 2.78947502e-01 -8.46157875e-03 7.21235648e-02
-2.78096229e-01 -2.12977305e-02 6.63029611e-01 5.55065215e-01
3.03602636e-01 3.20147306e-01 -4.14485306e-01 3.11399519e-01
-6.49105728e-01 -3.66874307e-01 4.35702175e-01 4.80032682e-01
-6.03999019e-01 -7.37730861e-01 -3.12791705e-01 1.35354444e-01
-3.41027856e-01 8.60886648e-02 9.57851484e-02 8.76362205e-01
-5.04912809e-02 4.51893955e-01 3.42689186e-01 -1.19239062e-01
4.00898457e-01 -2.32756838e-01 8.04013371e-01 -1.66840315e-01
-1.95923358e-01 3.01651746e-01 -2.30714366e-01 -4.76954222e-01
-6.10581636e-01 -5.34263134e-01 -1.39476907e+00 -2.75851101e-01
-4.31601442e-02 -2.55964518e-01 7.54844725e-01 9.57405031e-01
-1.09811641e-01 7.78984368e-01 4.53902811e-01 -8.55908811e-01
-9.21402454e-01 -9.15462196e-01 -4.89040762e-01 5.21581531e-01
3.86333495e-01 -4.90274191e-01 -1.91871747e-01 2.80734271e-01] | [8.993579864501953, 2.3266761302948] |
67305692-d490-4700-b4d1-e04e472b90d1 | destruction-of-image-steganography-using | 1912.10070 | null | https://arxiv.org/abs/1912.10070v1 | https://arxiv.org/pdf/1912.10070v1.pdf | Destruction of Image Steganography using Generative Adversarial Networks | Digital image steganalysis, or the detection of image steganography, has been studied in depth for years and is driven by Advanced Persistent Threat (APT) groups', such as APT37 Reaper, utilization of steganographic techniques to transmit additional malware to perform further post-exploitation activity on a compromised host. However, many steganalysis algorithms are constrained to work with only a subset of all possible images in the wild or are known to produce a high false positive rate. This results in blocking any suspected image being an unreasonable policy. A more feasible policy is to filter suspicious images prior to reception by the host machine. However, how does one optimally filter specifically to obfuscate or remove image steganography while avoiding degradation of visual image quality in the case that detection of the image was a false positive? We propose the Deep Digital Steganography Purifier (DDSP), a Generative Adversarial Network (GAN) which is optimized to destroy steganographic content without compromising the perceptual quality of the original image. As verified by experimental results, our model is capable of providing a high rate of destruction of steganographic image content while maintaining a high visual quality in comparison to other state-of-the-art filtering methods. Additionally, we test the transfer learning capability of generalizing to to obfuscate real malware payloads embedded into different image file formats and types using an unseen steganographic algorithm and prove that our model can in fact be deployed to provide adequate results. | ['Justin Hoffman', 'Jonathan Lwowski', 'Isaac Corley'] | 2019-12-20 | null | null | null | null | ['steganalysis', 'image-steganography'] | ['computer-vision', 'computer-vision'] | [ 1.06338155e+00 4.98106420e-01 3.81619424e-01 4.66348886e-01
-1.58337414e-01 -6.00553513e-01 7.51849711e-01 -3.92177075e-01
-2.28892311e-01 5.75045288e-01 -5.04826963e-01 -8.54543269e-01
9.74468514e-02 -1.11574078e+00 -9.20135379e-01 -1.13335621e+00
-7.12640226e-01 1.93955153e-01 4.22503144e-01 -4.21159297e-01
4.09998983e-01 3.48015875e-01 -1.34957707e+00 2.38044590e-01
5.04994035e-01 7.36067712e-01 1.53241679e-01 1.07015312e+00
5.03602326e-01 7.70112991e-01 -1.05318713e+00 -2.91263044e-01
6.86100960e-01 -6.94529295e-01 -6.07278168e-01 2.90677160e-01
7.71669745e-02 -4.44586515e-01 -3.35011870e-01 1.37703013e+00
3.20005924e-01 -5.79598367e-01 3.27511191e-01 -1.21824610e+00
-3.74977648e-01 4.93130147e-01 -3.19415629e-01 2.03364834e-01
2.28727058e-01 8.95049572e-01 2.38178581e-01 -2.15933532e-01
6.61136925e-01 1.16393042e+00 3.64539176e-01 4.87967104e-01
-1.08717334e+00 -9.21301365e-01 -5.18700957e-01 1.96956936e-02
-1.01370394e+00 -1.93383560e-01 6.44511342e-01 -1.27395675e-01
5.91152489e-01 4.52555150e-01 7.40271270e-01 1.22430027e+00
7.79636443e-01 1.46348968e-01 1.44851720e+00 -5.39573610e-01
-9.85329505e-04 2.34067827e-01 -6.71506107e-01 9.04662311e-01
5.89468360e-01 8.75250518e-01 1.68511331e-01 -1.97869092e-01
8.28799307e-01 -2.83364028e-01 -5.87298334e-01 -1.94770232e-01
-1.16245377e+00 1.16599655e+00 2.76065946e-01 3.58858854e-01
-4.18341994e-01 2.39426151e-01 2.90746629e-01 9.91047084e-01
2.41978735e-01 7.76104569e-01 -8.93253535e-02 3.39732498e-01
-6.95106328e-01 -9.25129950e-02 1.01071620e+00 2.36598194e-01
5.82724333e-01 5.46675563e-01 2.74421096e-01 -1.17526710e-01
2.92089909e-01 7.73858666e-01 2.62637377e-01 -7.04497635e-01
2.00803891e-01 3.05488735e-01 -2.77893603e-01 -1.36421227e+00
9.22210142e-02 -4.84559894e-01 -1.00850916e+00 8.07541251e-01
2.83924580e-01 -3.06885183e-01 -1.08317125e+00 1.31804585e+00
9.76603925e-02 4.09755647e-01 2.57937670e-01 5.81273317e-01
2.01710641e-01 9.65481102e-01 -1.12181522e-01 -2.85088539e-01
1.17996669e+00 -3.83864939e-01 -4.23630357e-01 -4.37717408e-01
5.04334092e-01 -6.56519413e-01 5.04640520e-01 7.04132378e-01
-8.20501685e-01 -3.32870215e-01 -1.41254497e+00 1.07370901e+00
-4.81289327e-01 -7.01444566e-01 2.23387748e-01 1.49062228e+00
-1.06259191e+00 5.25776088e-01 -5.66084623e-01 -1.81570441e-01
3.78329813e-01 7.56926417e-01 -4.55359966e-01 -1.28416643e-01
-1.49751282e+00 1.03435540e+00 7.93582320e-01 -1.94164053e-01
-1.53824556e+00 -4.00853723e-01 -6.73208416e-01 9.66487080e-02
3.69642705e-01 -4.96499360e-01 3.57156396e-01 -1.34020662e+00
-1.24799168e+00 9.20424640e-01 7.14786530e-01 -1.00105321e+00
8.01940322e-01 4.53217536e-01 -7.64624715e-01 4.98528183e-01
-4.03778076e-01 5.62066495e-01 1.70371795e+00 -1.49720073e+00
-5.83111107e-01 -3.64379659e-02 -1.63211208e-02 -3.32427561e-01
-3.83501381e-01 -3.50346379e-02 2.36419633e-01 -4.91796345e-01
-2.11063951e-01 -1.08325613e+00 -2.68310249e-01 -5.72820425e-01
-4.48869705e-01 6.34058833e-01 1.54268336e+00 -8.00858021e-01
9.84629810e-01 -1.92402554e+00 -1.39466554e-01 6.23963773e-01
9.49009582e-02 9.82946634e-01 -2.70506889e-01 4.69892561e-01
-2.51487136e-01 6.34289086e-01 -3.79496187e-01 3.07772577e-01
-3.98809195e-01 1.79710731e-01 -3.29111487e-01 9.51912880e-01
8.88759047e-02 6.88565969e-01 -8.57083201e-01 -3.98271859e-01
4.14963722e-01 5.45555174e-01 -3.83712679e-01 1.97699010e-01
-1.77778229e-01 6.72236502e-01 -2.21658349e-01 3.05602968e-01
7.90808141e-01 -1.47432283e-01 4.33485448e-01 1.60467505e-01
2.40696326e-01 -1.83999285e-01 -8.21377635e-01 4.23049122e-01
-2.46610507e-01 6.68398261e-01 1.53474748e-01 -1.07726097e+00
8.38521838e-01 4.95981961e-01 3.19687247e-01 -4.58671808e-01
5.33972979e-01 2.72387087e-01 3.11192960e-01 -6.56179607e-01
2.03411087e-01 -2.62526684e-02 -3.09010083e-03 4.85532969e-01
-1.57602549e-01 -1.10765234e-01 -1.18151568e-01 4.98530380e-02
1.52310276e+00 -2.99913913e-01 1.05838425e-01 -6.91593364e-02
7.83447146e-01 -1.37847036e-01 4.32988591e-02 1.02175760e+00
-9.74683464e-02 1.72470853e-01 5.18808722e-01 -5.61143979e-02
-1.38406551e+00 -7.56187379e-01 3.74713749e-01 4.89831775e-01
2.84730107e-01 2.46894643e-01 -8.45022976e-01 -8.67372513e-01
-3.51273924e-01 5.93094766e-01 -5.03509223e-01 -4.80558395e-01
-7.71491528e-01 -7.15946972e-01 9.14090037e-01 -5.42628825e-01
9.74596798e-01 -1.43742239e+00 -9.08347726e-01 1.48196906e-01
1.02758799e-02 -1.00931668e+00 3.64464857e-02 7.20195994e-02
-5.38231075e-01 -1.48303676e+00 -3.61349314e-01 -7.70605564e-01
8.44659746e-01 2.18499899e-01 7.07012236e-01 6.18264973e-01
-1.06039457e-01 1.60630450e-01 -5.79539716e-01 -2.23582655e-01
-1.58369839e+00 -1.46400526e-01 -3.60855311e-01 8.08276236e-02
-2.25664020e-01 -3.68977338e-01 -5.98482668e-01 3.51805657e-01
-1.43431985e+00 -2.54309297e-01 8.43495071e-01 8.25411916e-01
-1.61175996e-01 9.71256316e-01 2.51980782e-01 -1.12183535e+00
4.32154328e-01 -6.09165490e-01 -7.15566874e-01 1.12922147e-01
-7.36157417e-01 -1.04890838e-01 8.00131619e-01 -6.65778935e-01
-6.93961859e-01 -2.90476441e-01 -1.34154201e-01 -4.20129925e-01
-7.93241262e-02 2.07104787e-01 -9.03752744e-02 -8.03603411e-01
5.40473759e-01 7.18662262e-01 6.70330107e-01 1.49092898e-01
-5.22908121e-02 4.34400856e-01 3.59932542e-01 2.33539984e-01
1.55459523e+00 6.09672189e-01 3.92575324e-01 -9.02355313e-01
3.18187438e-02 -7.07956105e-02 4.64582257e-02 -3.87465954e-01
6.23661339e-01 -4.90454674e-01 -7.63337255e-01 8.09142172e-01
-8.84963095e-01 -2.07439914e-01 3.86455178e-01 3.46468650e-02
-4.27514464e-01 7.59908617e-01 -3.54228646e-01 -6.97292626e-01
-3.54994833e-01 -1.24543393e+00 3.57274503e-01 -1.46521196e-01
1.45650297e-01 -1.21897137e+00 -2.69153118e-01 2.81489938e-01
6.60200894e-01 8.62625241e-01 8.13868344e-01 -7.90060580e-01
-7.29204476e-01 -5.39722383e-01 -7.97921140e-03 6.88836992e-01
9.68463719e-03 -1.39187053e-01 -3.62940460e-01 -1.00213480e+00
4.50834781e-01 -5.68008795e-02 8.15876663e-01 5.97168095e-02
8.37256730e-01 -9.00328577e-01 -3.04905474e-01 7.00122833e-01
1.71817839e+00 5.52343071e-01 1.31881702e+00 6.57808781e-01
4.46469188e-01 6.62443638e-01 4.24077064e-01 2.74361014e-01
-3.99084657e-01 1.07676663e-01 1.21223664e+00 -1.17116168e-01
1.01060428e-01 -1.36843294e-01 6.63743138e-01 1.52461261e-01
1.16660064e-02 -9.52983201e-01 -6.78798139e-01 3.04719537e-01
-1.17558956e+00 -1.28907669e+00 -1.90462008e-01 2.15362811e+00
2.27840349e-01 5.09783506e-01 -8.04495439e-02 5.34829855e-01
1.06550014e+00 4.78616446e-01 -2.90426873e-02 -5.28185129e-01
-1.82194024e-01 2.48876438e-01 1.27220678e+00 4.62204784e-01
-1.09657931e+00 9.30088520e-01 5.58744812e+00 7.98183441e-01
-1.34508777e+00 2.66121745e-01 4.00033146e-01 5.50433993e-01
-1.74891725e-01 3.22012037e-01 -4.12573785e-01 8.24990392e-01
1.05386770e+00 2.39341915e-01 5.33150613e-01 3.39362741e-01
2.30020091e-01 -1.54475244e-02 -2.91435093e-01 4.99575138e-01
3.39710265e-01 -1.36432195e+00 1.75934926e-01 6.76454067e-01
5.70513487e-01 -3.88848990e-01 3.97102147e-01 1.00463696e-01
4.36483592e-01 -1.12956548e+00 2.92389959e-01 -3.38563807e-02
7.93294728e-01 -8.42825174e-01 7.62587786e-01 5.78232229e-01
-5.21670759e-01 -4.39097643e-01 -1.39960140e-01 2.83671707e-01
7.99986050e-02 1.85878798e-01 -1.24949121e+00 2.93969750e-01
3.72393936e-01 6.91440776e-02 -5.21852255e-01 7.02254474e-01
-3.99724305e-01 1.02597022e+00 -6.41274080e-02 1.70802683e-01
5.76235950e-01 2.07665667e-01 1.17024016e+00 9.89132345e-01
5.25560081e-01 -9.76311713e-02 4.85743620e-02 4.95109409e-01
2.13993639e-01 -4.38710839e-01 -9.09045756e-01 -2.68972069e-01
2.94189513e-01 8.46719384e-01 -8.70684683e-01 -1.89550072e-01
1.45200670e-01 1.00759256e+00 -5.69441438e-01 1.51840419e-01
-9.39917862e-01 -3.49710196e-01 1.48722544e-01 4.97300833e-01
7.39998341e-01 8.70045498e-02 6.07255176e-02 -8.18858504e-01
-5.92314243e-01 -1.42422640e+00 1.99629590e-01 -4.19164330e-01
-7.18922615e-01 6.30512416e-01 -2.47776851e-01 -1.37229955e+00
-2.66234159e-01 -4.26176876e-01 -6.01559103e-01 4.71681297e-01
-1.55178368e+00 -1.32989347e+00 -6.83786944e-02 6.69079423e-01
2.28536665e-01 -5.62616944e-01 4.97543722e-01 -2.81236935e-02
-3.29570659e-02 3.48407805e-01 -8.27702135e-02 2.07056582e-01
3.27955455e-01 -7.03636944e-01 3.73608261e-01 1.33425450e+00
-4.00968522e-01 1.26169547e-01 1.25836492e+00 -1.02102602e+00
-1.53812480e+00 -1.16126621e+00 4.17228311e-01 -9.80019569e-02
6.05555892e-01 -2.07045391e-01 -7.62931287e-01 7.11678088e-01
3.58828783e-01 -2.11620659e-01 4.36810963e-02 -1.00837457e+00
-1.61214396e-01 1.65036723e-01 -1.58382297e+00 3.89629483e-01
5.22819996e-01 -1.10601239e-01 -2.09256396e-01 2.13030249e-01
5.83747029e-01 -1.03020417e-02 -4.09878731e-01 5.29183984e-01
2.94166863e-01 -1.21236122e+00 1.16845012e+00 -1.39533550e-01
4.10551429e-01 -2.41046026e-01 1.08705983e-01 -1.18111110e+00
-1.18496612e-01 -1.13065171e+00 -1.05257317e-01 8.17234397e-01
-4.15331014e-02 -9.38555241e-01 7.32115030e-01 -3.98897171e-01
2.28763163e-01 -2.55376190e-01 -8.73442531e-01 -9.86365497e-01
-1.03881903e-01 -9.02160630e-02 5.23115396e-01 9.49572206e-01
-4.05805409e-01 -3.40815663e-01 -1.01709378e+00 6.52770817e-01
1.07785511e+00 -4.74492103e-01 9.22276795e-01 -8.94847810e-01
-4.18227911e-01 -3.01978260e-01 -6.92966700e-01 -3.82179022e-01
4.00566161e-02 -5.17907560e-01 -9.15663838e-02 -7.47749090e-01
-3.10143769e-01 -3.64746630e-01 -1.93166789e-02 1.57358438e-01
1.51901752e-01 7.40791380e-01 2.52805918e-01 2.71298319e-01
-9.42266881e-02 -3.71471077e-01 1.36250830e+00 -2.19334617e-01
3.99760343e-02 1.56257972e-01 -4.06580955e-01 4.80999321e-01
8.94796610e-01 -8.84828269e-01 -3.78488481e-01 1.92977339e-01
2.81039715e-01 4.27314401e-01 7.84027159e-01 -1.10786200e+00
-8.22304338e-02 -8.34093094e-02 9.85139981e-02 -7.47688413e-02
1.01076178e-01 -1.10588205e+00 5.49664736e-01 1.49885070e+00
1.77549478e-03 -3.06458980e-01 -2.53946304e-01 7.02403665e-01
4.28423733e-02 -2.30644628e-01 1.23233211e+00 -5.22843599e-01
-7.48375177e-01 1.05469190e-01 -8.33617270e-01 -4.49782640e-01
1.40721607e+00 -6.84127092e-01 -4.71763760e-01 -5.13972163e-01
-3.95923078e-01 -1.81901991e-01 6.48446083e-01 2.13797793e-01
7.86838293e-01 -5.20498097e-01 -8.36779952e-01 5.79650104e-01
-3.33022177e-01 -8.86998594e-01 4.76520024e-02 5.53635776e-01
-9.90866840e-01 1.88572258e-01 -5.11233270e-01 -3.89621198e-01
-1.56870103e+00 9.34125543e-01 3.85640025e-01 -6.56920552e-01
-5.17204821e-01 4.50164407e-01 -1.80306628e-01 1.14481233e-01
-7.95214474e-02 4.35293257e-01 -3.24018896e-01 -3.43761623e-01
4.37060386e-01 2.78230995e-01 -9.58999246e-02 -6.92234933e-01
-7.81274587e-02 9.15311798e-02 6.20474592e-02 1.15231782e-01
1.12377167e+00 -2.44147301e-01 -2.32622579e-01 -5.04974782e-01
1.31160939e+00 -2.21723896e-02 -1.01075351e+00 2.23956734e-01
-2.90526271e-01 -7.42689192e-01 1.85196042e-01 -7.53327489e-01
-1.21274567e+00 5.24724007e-01 8.96219611e-01 9.42219853e-01
1.15588582e+00 -4.25188810e-01 9.63893294e-01 9.40223783e-02
4.84802753e-01 -5.46526015e-01 3.88118029e-01 2.13057727e-01
4.37248379e-01 -1.31232035e+00 -1.41161352e-01 -3.98121744e-01
-4.59058166e-01 1.02355576e+00 1.96359217e-01 -5.21164000e-01
5.35896778e-01 3.25395972e-01 -1.82476968e-01 -3.43790770e-01
-4.02436137e-01 1.17684335e-01 -2.20627129e-01 9.93250132e-01
-4.29838687e-01 -1.14050545e-01 -1.55067518e-01 -6.21210337e-01
-9.65701416e-02 -2.30984911e-01 9.64782536e-01 1.03724194e+00
-7.27347493e-01 -1.08194506e+00 -9.62738514e-01 2.16514066e-01
-9.25902724e-01 8.71418491e-02 -2.20633417e-01 9.21713114e-01
4.40232813e-01 1.16789877e+00 -1.22781619e-01 -6.06351972e-01
-3.33086401e-01 -3.05103302e-01 3.98564130e-01 -2.86833882e-01
-8.34102809e-01 -8.72776955e-02 -5.63682616e-02 -1.54955134e-01
-4.85885233e-01 -2.24730790e-01 -6.71666861e-01 -7.62626588e-01
-4.14439440e-01 3.37520987e-02 8.02347839e-01 9.34958041e-01
7.49589584e-04 4.48938966e-01 1.07091272e+00 -6.42409325e-01
-5.32176197e-01 -6.15363836e-01 -4.45299834e-01 4.84261513e-01
6.13893211e-01 -2.69816220e-01 -1.01556635e+00 1.51415765e-01] | [4.301888465881348, 8.061725616455078] |
d08d8a21-9ab9-499c-8866-f49cd171a2a3 | robot-navigation-in-risky-crowded | 2303.08284 | null | https://arxiv.org/abs/2303.08284v1 | https://arxiv.org/pdf/2303.08284v1.pdf | Robot Navigation in Risky, Crowded Environments: Understanding Human Preferences | Risky and crowded environments (RCE) contain abstract sources of risk and uncertainty, which are perceived differently by humans, leading to a variety of behaviors. Thus, robots deployed in RCEs, need to exhibit diverse perception and planning capabilities in order to interpret other human agents' behavior and act accordingly in such environments. To understand this problem domain, we conducted a study to explore human path choices in RCEs, enabling better robotic navigational explainable AI (XAI) designs. We created a novel COVID-19 pandemic grocery shopping scenario which had time-risk tradeoffs, and acquired users' path preferences. We found that participants showcase a variety of path preferences: from risky and urgent to safe and relaxed. To model users' decision making, we evaluated three popular risk models (Cumulative Prospect Theory (CPT), Conditional Value at Risk (CVAR), and Expected Risk (ER). We found that CPT captured people's decision making more accurately than CVaR and ER, corroborating theoretical results that CPT is more expressive and inclusive than CVaR and ER. We also found that people's self assessments of risk and time-urgency do not correlate with their path preferences in RCEs. Finally, we conducted thematic analysis of open-ended questions, providing crucial design insights for robots is RCE. Thus, through this study, we provide novel and critical insights about human behavior and perception to help design better navigational explainable AI (XAI) in RCEs. | ['Sonia Martinez', 'Laurel D. Riek', 'Angelique Taylor', 'Aamodh Suresh'] | 2023-03-15 | null | null | null | null | ['robot-navigation'] | ['robots'] | [-4.79179561e-01 6.71664715e-01 -6.14132658e-02 -3.87249112e-01
-9.93381441e-02 -5.80580354e-01 2.78556198e-01 1.95271865e-01
-6.63158357e-01 5.32033145e-01 5.94184995e-01 -5.59454083e-01
-5.23028672e-01 -6.74183488e-01 -3.11288714e-01 -2.85489429e-02
-4.92355198e-01 4.84651208e-01 -3.48245740e-01 -7.03649640e-01
3.57227981e-01 3.28184992e-01 -1.41946721e+00 1.32532278e-02
1.25266707e+00 7.66008258e-01 4.51657474e-01 2.74956077e-01
4.10914749e-01 7.66877711e-01 -3.51184100e-01 -6.78839386e-01
3.67872268e-01 2.85149097e-01 -6.00959122e-01 -4.38728750e-01
-5.69857240e-01 -6.40629053e-01 -1.39151022e-01 5.45111656e-01
2.65093774e-01 3.99931043e-01 1.09284937e+00 -1.78226984e+00
-9.60053802e-01 7.77422190e-01 5.79784140e-02 -3.63869756e-01
7.91853547e-01 5.88689268e-01 7.03475237e-01 -6.29302800e-01
6.97335243e-01 1.78823316e+00 7.20062971e-01 3.36209297e-01
-1.09970140e+00 -4.74563003e-01 3.76303852e-01 2.07798898e-01
-1.22710478e+00 -3.02332710e-03 3.44037592e-01 -7.24882245e-01
8.05801272e-01 1.10218801e-01 1.01067364e+00 1.48155987e+00
6.18998587e-01 6.36254966e-01 1.15449023e+00 1.14741206e-01
7.73392260e-01 3.34591508e-01 -1.39964089e-01 4.24722247e-02
6.91097140e-01 5.36285222e-01 -4.62424085e-02 -3.88578117e-01
6.23888552e-01 1.14360504e-01 1.30636334e-01 -3.60235304e-01
-1.03092504e+00 7.76941359e-01 6.15125179e-01 -3.01549677e-02
-1.10227418e+00 -2.07250025e-02 3.61712053e-02 1.56476274e-01
-2.69923151e-01 8.42106283e-01 -2.43326321e-01 -5.55441082e-01
4.69847411e-01 8.03641796e-01 1.08255303e+00 1.02155530e+00
4.79098171e-01 -2.95621008e-01 -2.35978231e-01 7.35196054e-01
5.92783689e-01 6.85754240e-01 5.24925953e-03 -1.53985655e+00
4.09025192e-01 4.78469491e-01 9.94893968e-01 -1.64177847e+00
-1.02435863e+00 -1.45301688e-02 -3.80175799e-01 5.49952209e-01
3.84674668e-01 -4.31025177e-01 -6.43302441e-01 1.81803024e+00
-2.45700087e-02 -9.66831326e-01 3.99151295e-01 1.09853232e+00
9.48766917e-02 5.78636825e-01 6.83123827e-01 1.68422580e-01
1.47230029e+00 -2.37044439e-01 -6.73058271e-01 -7.49591231e-01
6.33912027e-01 -2.00291455e-01 1.42811227e+00 6.83550835e-01
-7.89422870e-01 -1.18845046e-01 -8.61204028e-01 9.52256992e-02
-2.09973782e-01 -3.05201977e-01 1.00040162e+00 8.70782554e-01
-9.06547129e-01 2.14525387e-01 -4.20826912e-01 -6.30195200e-01
2.97210783e-01 -1.45883635e-01 -1.96806416e-01 -2.91780144e-01
-1.34650886e+00 1.20084691e+00 8.28559622e-02 1.69488177e-01
-1.10201013e+00 -1.87328413e-01 -9.44490016e-01 -1.20119885e-01
3.12847435e-01 -5.72733939e-01 1.32572746e+00 -3.71497393e-01
-1.41186309e+00 2.55332381e-01 3.38608205e-01 -3.86679858e-01
5.80379903e-01 -3.06362212e-01 -2.42946208e-01 -8.30635726e-02
7.83715367e-01 9.50020373e-01 2.42534578e-01 -1.59699833e+00
-4.34035838e-01 -4.91218746e-01 3.61357599e-01 7.74241209e-01
2.43902355e-01 -1.62759319e-01 4.76328522e-01 -3.63555729e-01
2.29601815e-01 -1.10164154e+00 -9.26127136e-01 -2.15150343e-04
-1.12230130e-01 -2.92462051e-01 -9.70276073e-03 -4.89859611e-01
8.52509022e-01 -2.13423538e+00 -1.76280439e-01 2.68734813e-01
2.66655982e-01 -4.39521104e-01 -4.29457314e-02 8.83586526e-01
5.20060122e-01 3.97177696e-01 2.76769847e-02 -2.02762812e-01
6.64698660e-01 2.65200913e-01 -3.41484189e-01 3.93334657e-01
-1.89901471e-01 6.83653772e-01 -1.01505983e+00 -3.40498358e-01
1.34609967e-01 1.70253171e-03 -8.03003788e-01 1.08122319e-01
-1.12678871e-01 3.00898075e-01 -5.57056904e-01 7.17263699e-01
6.97701395e-01 4.32797223e-01 6.10761233e-02 3.81344169e-01
-4.57005858e-01 2.45070562e-01 -9.76009786e-01 1.14511144e+00
-5.79035699e-01 2.76207358e-01 3.55486184e-01 -1.05113193e-01
1.04930139e+00 -4.42103446e-02 1.52077109e-01 -9.53513563e-01
3.29252750e-01 1.56897977e-01 3.24932262e-02 -7.30129004e-01
8.05714846e-01 -7.29046017e-03 -8.62045825e-01 4.20450568e-01
-8.14966738e-01 -4.25618947e-01 -4.75761592e-01 1.48678035e-01
1.01513541e+00 -1.76702887e-01 4.04408872e-01 -5.34000993e-01
-2.65315682e-01 2.63413936e-01 8.12912822e-01 1.10767603e+00
-9.50043857e-01 3.59765664e-02 6.93631947e-01 -3.82778496e-01
-8.02817464e-01 -1.32364261e+00 4.36728261e-02 1.00968730e+00
7.53806770e-01 -2.64724288e-02 -6.50516629e-01 -1.10017873e-01
1.27465993e-01 1.59462023e+00 -6.32339120e-01 -2.06385493e-01
-1.57340795e-01 -2.52730429e-01 5.75147092e-01 4.32913184e-01
4.36531097e-01 -9.55097437e-01 -1.23685408e+00 1.82311833e-01
-5.68443954e-01 -7.99148381e-01 -1.72727466e-01 -1.40897751e-01
-4.24265027e-01 -6.61836326e-01 -3.77619773e-01 -3.05375576e-01
3.12931389e-01 5.64501166e-01 7.00551331e-01 -3.09909642e-01
1.18629094e-02 8.72509122e-01 -5.83396673e-01 -7.87908673e-01
-1.70133442e-01 -5.36064208e-01 5.33564091e-01 -5.83232939e-01
4.23948169e-01 -4.68248934e-01 -1.03650570e+00 1.04875374e+00
-3.92804593e-01 -1.70406058e-01 6.20702028e-01 1.25187516e-01
1.09471083e-02 1.16245486e-01 6.21251523e-01 1.26588136e-01
1.34887314e+00 -1.17457891e+00 -2.04904765e-01 -5.84873222e-02
-7.38049269e-01 -2.55315870e-01 2.03188092e-01 -4.73691434e-01
-1.57695556e+00 -3.91132921e-01 1.13400526e-01 4.33465332e-01
-4.39944834e-01 5.30583739e-01 -1.85331106e-01 3.01562607e-01
1.23814380e+00 -5.79435945e-01 1.10002048e-01 1.01091415e-01
5.88440657e-01 9.36262906e-01 2.36167058e-01 -9.24630165e-01
1.76608324e-01 4.79859769e-01 -4.30868208e-01 -6.54113948e-01
1.79522753e-01 1.03417978e-01 5.96775152e-02 -6.05291128e-01
1.01162326e+00 -1.01804304e+00 -1.54268575e+00 4.22163278e-01
-9.66081798e-01 -7.41490126e-01 -9.09809209e-03 8.15654159e-01
-9.38030183e-01 1.71858832e-01 -5.15477836e-01 -1.57984388e+00
3.27454209e-01 -1.25614345e+00 6.57254815e-01 2.85580665e-01
-8.71204734e-01 -5.29286623e-01 -2.07936272e-01 4.16342407e-01
5.30186951e-01 2.13611156e-01 7.53795922e-01 -4.86749589e-01
-5.84305942e-01 2.33035143e-02 -2.44245470e-01 -5.69174767e-01
-2.78436482e-01 -5.57807505e-01 -3.99965495e-01 1.75840959e-01
2.11590350e-01 -4.42383528e-01 8.02869499e-02 5.31425834e-01
3.80060852e-01 -5.27586281e-01 -4.28928882e-01 -2.39726938e-02
9.52777326e-01 8.18634629e-01 1.05101109e+00 8.17260921e-01
8.62757191e-02 1.35085201e+00 1.45487583e+00 7.77905464e-01
1.52424622e+00 4.96062487e-01 6.78491175e-01 4.61751014e-01
8.61902416e-01 -6.08028233e-01 7.03089058e-01 -4.08759683e-01
-2.37384588e-01 -3.46210092e-01 -1.14195025e+00 5.12786388e-01
-1.90005100e+00 -8.01385224e-01 -2.05544576e-01 1.95207739e+00
7.50619397e-02 7.26132914e-02 2.78950304e-01 -2.42368042e-01
5.44600189e-01 -2.30360150e-01 -6.43529713e-01 -7.98522532e-01
1.03625208e-01 -8.95126879e-01 8.05227578e-01 5.37677109e-01
-4.75109488e-01 9.16868508e-01 6.64943504e+00 2.73547530e-01
-2.72134542e-01 -9.37425122e-02 8.30690801e-01 2.33932212e-01
-9.61299300e-01 9.30185914e-02 -2.13162184e-01 1.30395398e-01
6.56103790e-01 -1.51633739e-01 4.63150233e-01 1.23692334e+00
5.44302762e-01 -6.03118122e-01 -9.99817908e-01 9.01242077e-01
-5.03444850e-01 -3.98205698e-01 -5.05707681e-01 2.94817775e-01
2.00468227e-01 -5.05469382e-01 3.80124241e-01 7.18302190e-01
1.09649849e+00 -1.24906147e+00 1.33599377e+00 5.45030594e-01
2.56689698e-01 -5.17104328e-01 6.71725452e-01 4.65140074e-01
-8.65472496e-01 -8.47090602e-01 -4.55420673e-01 -8.89584184e-01
6.32665038e-01 1.75709233e-01 -9.96004283e-01 7.93691352e-02
1.00424266e+00 2.56801516e-01 1.94440838e-02 5.31318426e-01
-4.35170531e-02 -2.39941090e-01 -2.96510667e-01 -4.39946026e-01
2.33798087e-01 -4.70718086e-01 7.61338592e-01 4.93524373e-01
6.13950551e-01 5.24661124e-01 3.57234254e-02 1.25732207e+00
9.27735388e-01 -1.69053406e-01 -9.03100967e-01 5.57625331e-02
1.06216812e+00 6.79436028e-01 -5.87171376e-01 2.79126018e-01
2.61153169e-02 7.89774537e-01 6.99883625e-02 5.39441228e-01
-7.64750421e-01 -2.21301410e-02 1.29338932e+00 7.67070055e-02
-3.49784106e-01 -6.88653708e-01 -8.55660141e-01 -6.47525370e-01
-3.90884019e-02 -6.36231601e-01 2.23973423e-01 -1.04475713e+00
-1.33070242e+00 3.90892655e-01 4.25601870e-01 -1.13682497e+00
-7.42513835e-02 -6.02999747e-01 -3.33675623e-01 7.06449389e-01
-8.72170329e-01 -8.43220413e-01 -4.16936100e-01 2.58165985e-01
2.54158288e-01 1.61345094e-01 5.94807267e-01 -1.64632618e-01
-2.40178332e-01 2.02332318e-01 -3.69420201e-01 -6.50544465e-01
4.80443776e-01 -9.25158799e-01 4.74193007e-01 2.69495249e-01
-1.19052398e+00 8.07633996e-01 9.33632612e-01 -1.27953196e+00
-1.42599499e+00 -4.47907537e-01 4.13496971e-01 -5.58396459e-01
3.57825071e-01 -2.91192263e-01 -4.61342245e-01 8.18251312e-01
-1.34502903e-01 -9.20121491e-01 8.26781690e-01 2.41912231e-01
8.20142850e-02 3.60011578e-01 -1.76004565e+00 1.48905396e+00
1.61483908e+00 -1.92451999e-01 -8.81484032e-01 -2.60246992e-01
9.31507468e-01 -1.34004265e-01 -7.88523436e-01 2.57784665e-01
9.04547453e-01 -1.27260196e+00 1.09459102e+00 -9.59941000e-03
2.98211187e-01 -1.29971569e-02 -5.28449178e-01 -1.43819737e+00
-6.38709545e-01 -4.12184507e-01 7.35103607e-01 7.88827896e-01
4.41720843e-01 -1.15974951e+00 4.84677553e-01 1.73042524e+00
-4.48117763e-01 -4.85784650e-01 -8.47122431e-01 -6.21077836e-01
7.68606141e-02 -1.02140343e+00 8.51588607e-01 6.15905821e-01
6.50505602e-01 -4.17460412e-01 -2.35866830e-01 3.92127156e-01
6.26663744e-01 -6.55739129e-01 6.94869876e-01 -1.30545700e+00
2.65171468e-01 -5.44264376e-01 -2.69889414e-01 -9.36089873e-01
-1.11578800e-01 -2.68789858e-01 3.88743848e-01 -1.90264142e+00
-3.89510244e-01 -1.21891201e+00 5.06075680e-01 1.62942618e-01
3.34871896e-02 -6.45621121e-01 3.55706006e-01 3.01784992e-01
-2.39853308e-01 8.34134519e-01 1.01693773e+00 2.49187812e-01
-6.55989766e-01 -4.32608277e-02 -1.30683887e+00 9.47669029e-01
9.52504992e-01 -7.25311413e-02 -7.89623618e-01 -2.64637798e-01
7.76938081e-01 5.87287962e-01 5.02092302e-01 -8.46886277e-01
1.07292853e-01 -9.30499315e-01 2.05162004e-01 -4.36506927e-01
6.03910685e-01 -1.12953889e+00 3.60010743e-01 6.13244891e-01
-1.92485467e-01 1.36427000e-01 2.41652921e-01 7.62317598e-01
5.10916531e-01 -1.76254228e-01 7.05304295e-02 -1.35620341e-01
-7.84681559e-01 -1.17066361e-01 -1.45392394e+00 -3.47025096e-01
1.42873156e+00 -5.99355340e-01 -4.87813264e-01 -1.10510290e+00
-6.71505928e-01 8.91573370e-01 7.26434588e-01 5.84682703e-01
9.54697728e-01 -1.27631927e+00 -2.32642397e-01 -1.79493010e-01
2.15159848e-01 -5.94039001e-02 7.61473894e-01 5.02493560e-01
-5.68628430e-01 2.97954261e-01 -5.90345085e-01 -1.26316026e-01
-1.20374173e-01 5.83262622e-01 2.71547288e-01 3.84307057e-01
-2.99926817e-01 6.76844358e-01 4.26950842e-01 -8.39776099e-01
1.20975256e-01 -4.58708048e-01 -4.85182554e-01 -2.25104438e-03
2.71085620e-01 9.23718572e-01 -9.06734884e-01 -6.04980588e-01
-3.78124237e-01 1.96384713e-01 3.23358506e-01 -6.17424667e-01
7.45193779e-01 -8.04425359e-01 1.93212554e-01 3.09392154e-01
3.06753218e-01 -1.19197950e-01 -1.43417406e+00 3.93115312e-01
8.89114141e-02 -8.97343814e-01 -5.72900593e-01 -9.84111190e-01
1.87859163e-01 1.90175802e-01 4.36405689e-01 2.54114628e-01
7.69514441e-01 1.10990316e-01 4.78074133e-01 6.34877145e-01
1.24551415e+00 -1.27316582e+00 9.19190496e-02 3.49300206e-01
1.28365207e+00 -1.27125001e+00 -5.58827221e-01 -6.35338128e-01
-1.46457565e+00 5.43348968e-01 8.99215043e-01 3.41547340e-01
5.44766784e-01 -9.56800729e-02 1.32246211e-01 -2.04494819e-01
-6.14499450e-01 -9.02869105e-02 -3.65623564e-01 1.37456000e+00
-7.37842545e-02 7.60132432e-01 -5.23240089e-01 1.47431850e+00
-8.14373672e-01 -3.02812271e-02 1.03551304e+00 9.35790122e-01
-6.79187179e-01 -6.05277777e-01 -6.60563767e-01 2.61456013e-01
2.76164919e-01 4.36882794e-01 -4.09374624e-01 7.56999552e-01
-5.92384972e-02 1.65189421e+00 9.93417278e-02 -7.32255161e-01
6.49537504e-01 -4.02402401e-01 -1.35415226e-01 -3.16977710e-01
-2.19316646e-01 -6.18393898e-01 5.79399467e-01 -9.00749087e-01
2.31483236e-01 -8.13893914e-01 -1.35504663e+00 -3.51973832e-01
2.81984031e-01 -4.99821752e-02 7.82220900e-01 5.92905045e-01
5.87919354e-01 -1.03456397e-02 5.09772182e-01 -1.15485466e+00
-4.95169610e-01 -8.03227186e-01 -5.18421650e-01 2.02551380e-01
-4.31146920e-02 -1.15403879e+00 -4.51225311e-01 -7.73550630e-01] | [4.882781505584717, 1.0189814567565918] |
6d9c3b6f-2241-4fae-986d-e4829c4a9caa | compression-aware-video-super-resolution | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_Compression-Aware_Video_Super-Resolution_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_Compression-Aware_Video_Super-Resolution_CVPR_2023_paper.pdf | Compression-Aware Video Super-Resolution | Videos stored on mobile devices or delivered on the Internet are usually in compressed format and are of various unknown compression parameters, but most video super-resolution (VSR) methods often assume ideal inputs resulting in large performance gap between experimental settings and real-world applications. In spite of a few pioneering works being proposed recently to super-resolve the compressed videos, they are not specially designed to deal with videos of various levels of compression. In this paper, we propose a novel and practical compression-aware video super-resolution model, which could adapt its video enhancement process to the estimated compression level. A compression encoder is designed to model compression levels of input frames, and a base VSR model is then conditioned on the implicitly computed representation by inserting compression-aware modules. In addition, we propose to further strengthen the VSR model by taking full advantage of meta data that is embedded naturally in compressed video streams in the procedure of information fusion. Extensive experiments are conducted to demonstrate the effectiveness and efficiency of the proposed method on compressed VSR benchmarks. | ['Yu-Wing Tai', 'Huchuan Lu', 'Takashi Isobe', 'Xin Tao', 'Xu Jia', 'Yingwei Wang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['video-super-resolution', 'video-enhancement', 'model-compression'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 8.75316739e-01 -2.66525298e-01 -4.83401984e-01 -3.14211369e-01
-6.61206484e-01 -1.53322995e-01 3.26438159e-01 -8.20933878e-02
-1.43844649e-01 5.61687171e-01 5.50578415e-01 5.28726913e-02
-2.56578416e-01 -7.99598277e-01 -6.60843611e-01 -5.08444488e-01
-9.16869100e-03 -1.24918461e-01 2.86827058e-01 -2.54295766e-01
3.64979684e-01 1.25386611e-01 -1.80342460e+00 6.00311816e-01
9.63231742e-01 1.10706067e+00 8.27316701e-01 8.01728845e-01
-4.98443395e-02 1.06889153e+00 -3.65006000e-01 -2.30566755e-01
4.93138254e-01 -6.86078429e-01 -4.99737948e-01 4.31169510e-01
1.64470419e-01 -5.59326112e-01 -6.66974485e-01 1.34851670e+00
3.58559161e-01 6.26297295e-02 1.69483885e-01 -6.85691833e-01
-5.92059433e-01 9.53518927e-01 -7.98262060e-01 6.25510156e-01
6.31785631e-01 -1.05479531e-01 7.99400866e-01 -8.25075030e-01
7.30268717e-01 1.18517315e+00 2.13034451e-01 4.12499398e-01
-1.00083458e+00 -5.58151841e-01 1.61804587e-01 6.43327653e-01
-1.54076767e+00 -7.64538467e-01 8.27192187e-01 -1.00496367e-01
5.35399377e-01 3.31105202e-01 5.16613543e-01 7.55751252e-01
-2.85722706e-02 5.65563917e-01 9.34392273e-01 -2.75036991e-01
2.83605903e-01 6.68166056e-02 -2.67095059e-01 2.12201849e-01
2.01656595e-01 -1.06142469e-01 -7.99809158e-01 7.23861158e-02
1.03703761e+00 1.59196123e-01 -7.78013825e-01 -2.53550708e-01
-1.16356075e+00 5.41076422e-01 1.56700864e-01 4.17618245e-01
-5.90706110e-01 -2.57418662e-01 4.26067501e-01 3.42688680e-01
2.90460885e-01 -5.16576581e-02 -1.70945562e-02 -5.41148745e-02
-1.39183533e+00 2.84318745e-01 4.16583538e-01 1.16955638e+00
6.08618200e-01 7.34174550e-02 -2.38069057e-01 8.83681357e-01
-1.62925255e-02 2.04946175e-01 6.46307409e-01 -1.34597540e+00
7.97040284e-01 2.62730747e-01 1.46625757e-01 -1.30530465e+00
1.35864004e-01 -4.70352322e-01 -1.17267942e+00 -1.46484911e-01
-1.98159948e-01 2.31424600e-01 -4.94486362e-01 1.42283845e+00
8.22742730e-02 6.37979925e-01 3.21734369e-01 1.20432568e+00
5.52878261e-01 9.17250097e-01 -1.88742325e-01 -9.04425979e-01
1.32126760e+00 -8.21239352e-01 -1.06566799e+00 9.47660580e-03
8.96163136e-02 -6.16320670e-01 5.82648575e-01 6.11481607e-01
-1.68216264e+00 -9.26348746e-01 -1.35656226e+00 -1.28458396e-01
2.29585081e-01 -1.72326595e-01 6.05835281e-02 3.91423911e-01
-1.03280973e+00 6.35360837e-01 -6.10396743e-01 -7.77854100e-02
4.56858635e-01 1.58518627e-01 -2.44032443e-01 -3.27818453e-01
-1.27304888e+00 4.98458922e-01 6.40665174e-01 5.17233983e-02
-8.01034272e-01 -5.89038432e-01 -6.23845160e-01 3.63082826e-01
7.91826963e-01 -6.77899778e-01 8.89096797e-01 -1.16963470e+00
-1.32524848e+00 3.70538205e-01 -2.37365589e-01 -7.65595794e-01
5.18785834e-01 -2.24027738e-01 -5.90635180e-01 6.24857008e-01
-2.93082237e-01 2.28284687e-01 1.16437960e+00 -1.31309533e+00
-1.04476607e+00 -3.76586854e-01 2.64113724e-01 5.11413515e-01
-4.93057400e-01 1.03395864e-01 -7.71909833e-01 -7.72637486e-01
2.36140698e-01 -2.67331272e-01 -2.38850057e-01 -1.52354658e-01
1.62984848e-01 3.64972442e-01 8.73821855e-01 -9.35793638e-01
1.64232183e+00 -2.16454101e+00 6.24734521e-01 -6.23382069e-02
3.47931445e-01 4.56485540e-01 -1.77259475e-01 2.05207139e-01
3.89662832e-02 -2.52548773e-02 -3.26093227e-01 -1.66489273e-01
-4.55654502e-01 1.84462577e-01 -5.90329826e-01 3.24856341e-01
-6.25831410e-02 3.55671883e-01 -9.25971568e-01 -7.52328813e-01
4.23841476e-01 8.47142756e-01 -7.87905991e-01 3.90827328e-01
-9.07285064e-02 4.41201210e-01 -4.94192481e-01 4.62270528e-01
1.02756417e+00 -2.49270260e-01 5.08281946e-01 -5.09888768e-01
-9.43753421e-02 5.55229224e-02 -1.41772473e+00 1.83315992e+00
-3.94481599e-01 3.41038913e-01 4.37466651e-01 -1.12473726e+00
7.00406909e-01 3.99313986e-01 6.62890434e-01 -7.99689829e-01
1.48519650e-01 -3.08794063e-02 -2.69512713e-01 -5.45934260e-01
1.02731597e+00 7.09064156e-02 3.90163749e-01 -1.10341914e-01
-3.71636171e-03 2.09839091e-01 3.65672559e-01 2.26178184e-01
1.01868820e+00 1.66299805e-01 5.76077223e-01 4.93942760e-02
9.63565946e-01 -4.79144901e-01 6.32591307e-01 6.06720388e-01
-9.89070982e-02 7.37265527e-01 1.38289288e-01 -1.31734774e-01
-1.28923774e+00 -8.07618141e-01 -1.39620543e-01 9.82412636e-01
5.81911504e-01 -7.84308374e-01 -8.86883080e-01 -6.65132478e-02
-5.16387999e-01 5.11180937e-01 -2.43959874e-01 -1.97549560e-03
-7.07187772e-01 -5.34396827e-01 7.32709765e-02 2.70010948e-01
6.57623887e-01 -6.70071900e-01 -8.32511961e-01 4.28995639e-01
-6.82107687e-01 -1.57069016e+00 -3.61128449e-01 -3.34312767e-01
-9.97454405e-01 -9.28134501e-01 -8.02376449e-01 -5.64438403e-01
3.08400631e-01 7.34139025e-01 9.14338768e-01 2.04534367e-01
9.41122696e-02 2.92903870e-01 -7.45206118e-01 1.91085249e-01
-6.66054726e-01 -5.84349707e-02 -6.86527118e-02 4.52104509e-01
6.87609985e-02 -7.00567126e-01 -7.86208808e-01 1.70687526e-01
-1.34334886e+00 4.36544895e-01 5.28114557e-01 6.38337255e-01
8.95635009e-01 5.96886396e-01 3.69989902e-01 -8.00676942e-01
3.88022482e-01 -6.31930590e-01 -5.61169565e-01 1.63941711e-01
-5.39253592e-01 4.05826373e-03 8.62513483e-01 -2.76663423e-01
-1.18651462e+00 -2.11129367e-01 -2.54511070e-02 -8.37189317e-01
4.39953841e-02 3.91381949e-01 -3.38648587e-01 1.21866576e-01
1.68504417e-01 4.89386082e-01 -1.76829115e-01 -6.52461708e-01
1.85533017e-01 8.30005944e-01 9.05366957e-01 -3.39729309e-01
8.17232966e-01 5.47217429e-01 -1.61746722e-02 -7.13804364e-01
-5.96552372e-01 -3.41422379e-01 -4.00988162e-01 -1.17224060e-01
6.46083236e-01 -1.23676705e+00 -3.03993106e-01 2.47147307e-01
-7.98778832e-01 1.27166435e-01 -1.21693246e-01 4.29233998e-01
-7.69476235e-01 7.97220170e-01 -5.86557329e-01 -7.69407392e-01
-2.32984766e-01 -1.27696466e+00 9.18591261e-01 1.93873093e-01
3.89875531e-01 -5.55591464e-01 -1.74116507e-01 3.68103921e-01
6.47390068e-01 2.41164677e-02 5.18981814e-01 -5.20808250e-02
-8.57069552e-01 1.71992674e-01 -4.77868080e-01 4.19044822e-01
2.60540992e-01 -2.01908141e-01 -7.36939669e-01 -4.95723456e-01
4.56575811e-01 1.05251715e-01 9.68033254e-01 4.02149498e-01
1.52803218e+00 -4.37904656e-01 -1.11125797e-01 8.41805577e-01
1.78673863e+00 1.59743816e-01 8.90569270e-01 2.10298270e-01
5.45321465e-01 2.80921727e-01 8.65041435e-01 7.85273790e-01
3.08456570e-01 9.60163116e-01 5.30981481e-01 4.01845157e-01
-2.10416645e-01 -1.69922173e-01 4.84444678e-01 1.08703411e+00
-6.17975593e-01 -4.39559639e-01 -1.82009295e-01 2.59879708e-01
-1.81767857e+00 -1.33499718e+00 2.52227157e-01 2.31439567e+00
7.43962049e-01 1.01153567e-01 -9.71491337e-02 3.55753303e-01
1.02719486e+00 5.70765376e-01 -4.32428002e-01 -7.62978718e-02
-3.84809673e-01 7.93439820e-02 5.71294069e-01 3.61414850e-01
-9.60026324e-01 4.91480231e-01 5.76776028e+00 1.08290970e+00
-9.03748155e-01 1.18480451e-01 3.83692890e-01 -2.06495255e-01
-1.92058966e-01 4.36358713e-02 -6.08962297e-01 6.10860705e-01
1.11623752e+00 -5.24756908e-01 8.47255826e-01 6.13913596e-01
3.46655875e-01 6.16407655e-02 -9.83111501e-01 1.35156393e+00
9.15371627e-02 -1.37303352e+00 3.96275729e-01 -3.77356187e-02
6.97625697e-01 -4.12167698e-01 1.80815727e-01 1.71117503e-02
-3.32346141e-01 -6.94482863e-01 8.12040031e-01 5.05460799e-01
1.02127326e+00 -8.05922151e-01 6.82822227e-01 2.98703849e-01
-1.59086633e+00 -3.98367375e-01 -6.34223998e-01 1.51749784e-02
4.50071186e-01 3.43189269e-01 -2.65161633e-01 1.00867152e+00
7.19741702e-01 9.89120245e-01 -4.75860417e-01 7.38211095e-01
2.66358078e-01 4.98534352e-01 1.31874323e-01 7.50929177e-01
-1.77555546e-01 -1.71740800e-01 6.92963600e-01 1.18143833e+00
6.48618102e-01 4.65021253e-01 -9.15167779e-02 5.22187710e-01
-1.24490134e-01 1.89527348e-01 -3.75988752e-01 3.82809043e-02
5.19376993e-01 1.10395372e+00 -4.72018749e-01 -6.41814828e-01
-8.05623293e-01 1.06565809e+00 -3.41107845e-02 3.28735650e-01
-1.01596129e+00 -1.06100328e-02 6.52742028e-01 3.82803470e-01
6.20684981e-01 -6.85127452e-02 1.57752261e-01 -1.53817570e+00
7.31236935e-02 -1.24901700e+00 3.14645618e-01 -6.99591279e-01
-9.06419814e-01 6.90331340e-01 6.35917410e-02 -1.68381429e+00
-4.11780089e-01 -1.59911379e-01 2.57865321e-02 5.81993401e-01
-1.72892320e+00 -5.56238949e-01 -5.29295981e-01 8.29223275e-01
1.01092136e+00 -8.07661265e-02 3.57617646e-01 5.93307376e-01
-3.07430178e-01 3.70916158e-01 1.40876874e-01 -2.20324114e-01
5.64280570e-01 -6.79966569e-01 3.06802224e-02 1.13454509e+00
-1.79246739e-01 3.55023384e-01 9.89814341e-01 -4.76759255e-01
-1.82303119e+00 -1.11746216e+00 3.06585461e-01 1.44097149e-01
3.87206137e-01 -1.38867861e-02 -1.27141178e+00 5.67491949e-01
3.41416091e-01 1.12842269e-01 3.21663231e-01 -6.06664896e-01
-3.48752350e-01 -2.59650201e-01 -1.25767493e+00 3.98681194e-01
1.24474835e+00 -5.60567498e-01 -4.87604350e-01 -3.13911363e-02
9.11640286e-01 -6.00337625e-01 -1.08439505e+00 6.79930329e-01
3.27519566e-01 -1.25340331e+00 1.17888594e+00 -2.56982356e-01
8.10558379e-01 -4.68559712e-01 -7.19991803e-01 -7.90868700e-01
-4.10797715e-01 -4.97004658e-01 -6.20215356e-01 1.25818384e+00
-3.77944946e-01 -1.57025665e-01 4.83300567e-01 2.48622984e-01
5.32081947e-02 -4.79969084e-01 -8.80550265e-01 -4.06256676e-01
-6.45075798e-01 -3.67014885e-01 7.08726227e-01 7.39320159e-01
-1.34735808e-01 1.15729898e-01 -8.28411520e-01 4.17245269e-01
9.76101339e-01 2.29554519e-01 5.65989196e-01 -7.95858622e-01
-6.69012487e-01 -2.97205001e-01 -5.83491981e-01 -1.23306799e+00
-2.16216862e-01 -6.14768445e-01 -1.21678963e-01 -1.27273798e+00
5.93755305e-01 -2.65118405e-02 -6.61355555e-01 -2.76217103e-01
-2.32586712e-01 3.37626249e-01 3.93810779e-01 3.95646930e-01
-7.94844925e-01 4.72433895e-01 1.12784660e+00 -6.29747212e-02
-1.93685710e-01 -1.75872833e-01 -8.10593843e-01 7.26059377e-01
5.02484739e-01 -1.88533664e-02 -6.77741051e-01 -6.10945106e-01
1.33377925e-01 7.83700943e-01 3.35297227e-01 -1.10244012e+00
2.37903923e-01 -1.44681349e-01 2.17805564e-01 -5.55225611e-01
3.57632130e-01 -9.94712472e-01 5.60173929e-01 3.79160419e-02
-5.23215234e-01 -8.15061852e-02 -1.65008768e-01 7.21131682e-01
-4.61042255e-01 -1.91628471e-01 9.07135844e-01 -7.90739357e-02
-8.65803659e-01 3.53891432e-01 -1.39220610e-01 -1.43183365e-01
7.89629698e-01 -3.11885685e-01 -1.27435029e-01 -4.47846085e-01
-5.64253449e-01 -4.11033537e-03 7.92986631e-01 4.81233150e-01
1.00652707e+00 -1.21087372e+00 -8.68964255e-01 1.49237350e-01
-1.35907918e-01 -1.01293378e-01 7.50232160e-01 4.91390824e-01
-3.50264996e-01 1.67492688e-01 -3.55324715e-01 -4.93970722e-01
-1.28116620e+00 1.27338672e+00 -1.88950635e-02 -2.77395755e-01
-8.84946406e-01 2.10110351e-01 2.68606335e-01 5.68572104e-01
1.83780774e-01 -1.33100495e-01 -4.45524365e-01 -1.40355289e-01
1.15537000e+00 5.43086410e-01 -1.72195286e-01 -9.90609407e-01
-4.84112129e-02 6.41731977e-01 -2.92847436e-02 -9.95987561e-03
1.33489799e+00 -9.15021181e-01 7.96004236e-02 1.67339876e-01
1.03825891e+00 5.65225407e-02 -1.25727296e+00 -5.01808345e-01
-3.28191787e-01 -9.61352706e-01 1.65893734e-01 -1.01171605e-01
-1.35303497e+00 5.85984409e-01 7.13332832e-01 1.55827701e-01
1.93489671e+00 -1.43008292e-01 9.85385478e-01 3.98212261e-02
8.07019591e-01 -9.65205491e-01 -2.70518273e-01 -1.71385214e-01
7.75399864e-01 -9.62211013e-01 2.59819299e-01 -5.45842111e-01
-5.01624584e-01 1.07420075e+00 2.81901687e-01 -9.86792594e-02
3.41319889e-01 5.09028733e-01 -4.51279134e-01 2.64375478e-01
-9.22493100e-01 1.09342560e-01 -5.97665347e-02 4.68030602e-01
1.99719965e-01 -1.52167648e-01 -4.85256195e-01 6.22201085e-01
1.02442071e-01 5.38752854e-01 7.83906996e-01 7.68968225e-01
-6.62301302e-01 -9.05071139e-01 -5.19326389e-01 2.89705932e-01
-6.80357695e-01 8.00041668e-03 4.75342393e-01 2.13150814e-01
1.71396509e-01 9.60143745e-01 5.64004146e-02 -5.02367377e-01
2.44157299e-01 -3.89140636e-01 5.34609497e-01 -1.88991919e-01
-1.13705084e-01 2.39507705e-01 -3.86955470e-01 -9.05681014e-01
-9.26226676e-01 -6.83586240e-01 -1.04358625e+00 -3.31992805e-01
-1.03726611e-01 1.25899658e-01 3.63563269e-01 6.35469854e-01
3.37125361e-01 7.85674036e-01 7.76178658e-01 -1.02832031e+00
-4.94478285e-01 -7.45459676e-01 -6.76874518e-01 5.23212075e-01
4.46945131e-01 -4.57386762e-01 -2.66748309e-01 5.59635818e-01] | [11.118399620056152, -1.848184585571289] |
0cadf673-3a0f-40db-a148-6df7b6ab2676 | a-systematic-evaluation-of-methods-for-cell | 2110.00681 | null | https://arxiv.org/abs/2110.00681v1 | https://arxiv.org/pdf/2110.00681v1.pdf | A systematic evaluation of methods for cell phenotype classification using single-cell RNA sequencing data | Background: Single-cell RNA sequencing (scRNA-seq) yields valuable insights about gene expression and gives critical information about complex tissue cellular composition. In the analysis of single-cell RNA sequencing, the annotations of cell subtypes are often done manually, which is time-consuming and irreproducible. Garnett is a cell-type annotation software based the on elastic net method. Besides cell-type annotation, supervised machine learning methods can also be applied to predict other cell phenotypes from genomic data. Despite the popularity of such applications, there is no existing study to systematically investigate the performance of those supervised algorithms in various sizes of scRNA-seq data sets. Methods and Results: This study evaluates 13 popular supervised machine learning algorithms to classify cell phenotypes, using published real and simulated data sets with diverse cell sizes. The benchmark contained two parts. In the first part, we used real data sets to assess the popular supervised algorithms' computing speed and cell phenotype classification performance. The classification performances were evaluated using AUC statistics, F1-score, precision, recall, and false-positive rate. In the second part, we evaluated gene selection performance using published simulated data sets with a known list of real genes. Conclusion: The study outcomes showed that ElasticNet with interactions performed best in small and medium data sets. NB was another appropriate method for medium data sets. In large data sets, XGB works excellent. Ensemble algorithms were not significantly superior to individual machine learning methods. Adding interactions to ElasticNet can help, and the improvement was significant in small data sets. | ['Xuekui Zhang', 'Junhua Gu', 'Hua He', 'Elham Majd', 'Li Xing', 'Xiaowen Cao'] | 2021-10-01 | null | null | null | null | ['phenotype-classification'] | ['medical'] | [ 1.12365156e-01 -4.12582308e-01 -2.99948812e-01 -1.06974840e-01
-5.55405140e-01 -5.70861280e-01 1.25301167e-01 3.50983799e-01
-4.00952309e-01 1.34746385e+00 1.97959840e-02 -2.93855052e-02
-4.23033893e-01 -8.59040499e-01 -1.03526480e-01 -1.22973180e+00
-6.76404461e-02 7.51581788e-01 1.79263994e-01 -1.10878140e-01
4.28821772e-01 5.13284802e-01 -1.56550741e+00 3.07268560e-01
7.73034990e-01 8.23831677e-01 9.06175599e-02 7.97663152e-01
-3.79705489e-01 5.22931576e-01 -4.56764579e-01 4.65608723e-02
3.75003591e-02 -4.12845314e-01 -3.78457338e-01 -3.69115621e-01
-1.54124528e-01 4.42178339e-01 2.31841296e-01 6.90848231e-01
1.15569973e+00 -2.13841841e-01 6.70369267e-01 -1.35091257e+00
-1.83390871e-01 3.21432143e-01 -2.73573518e-01 5.00390902e-02
3.60618353e-01 3.70762795e-01 7.00154364e-01 -6.12604260e-01
1.00167394e+00 8.63241673e-01 9.43504274e-01 5.87762296e-01
-1.26059484e+00 -7.45023251e-01 -4.21840698e-01 -1.78105235e-02
-1.51356757e+00 -1.78560421e-01 2.83150077e-01 -5.68005323e-01
9.74518597e-01 6.60927474e-01 8.24678302e-01 7.11122274e-01
4.58806008e-01 4.07086313e-01 1.37905991e+00 -2.77201802e-01
4.03437078e-01 1.62005946e-02 2.41646275e-01 6.01857245e-01
1.96502805e-01 -2.63594911e-02 -4.75241721e-01 -4.73203093e-01
3.42935711e-01 2.46789772e-02 -3.89184862e-01 1.99065506e-01
-1.36254966e+00 5.15422761e-01 -3.63544166e-01 6.22091353e-01
-6.64596707e-02 -1.37064368e-01 8.98321152e-01 3.48853290e-01
5.08258045e-01 7.34921098e-01 -1.13141859e+00 -3.73348027e-01
-5.80112994e-01 -2.99149547e-02 9.48666692e-01 6.77778065e-01
3.99476528e-01 -1.39618978e-01 -1.38810724e-01 1.14614213e+00
3.80586088e-02 2.79609799e-01 9.88606691e-01 -5.08806884e-01
-2.32135996e-01 1.06116986e+00 -2.72015810e-01 -9.06689227e-01
-8.97286117e-01 -2.20514312e-01 -1.23490191e+00 -2.25490667e-02
8.10215116e-01 -2.63719618e-01 -7.37592638e-01 1.44453573e+00
1.97221741e-01 1.81703940e-01 3.55288647e-02 7.05626607e-01
9.27335501e-01 4.36303318e-01 1.50533795e-01 -7.29394257e-01
1.21024525e+00 -3.33521038e-01 -7.85816312e-01 5.65557539e-01
1.00311244e+00 -7.93989718e-01 9.46521282e-01 5.39741278e-01
-5.24544597e-01 -2.17437401e-01 -7.28915632e-01 1.88126400e-01
-7.92383850e-01 1.04634978e-01 8.05134714e-01 7.78902650e-01
-7.77745664e-01 6.65716588e-01 -5.95045149e-01 -6.61913514e-01
5.65341294e-01 6.56580448e-01 -5.69448173e-01 1.47603646e-01
-9.74215388e-01 6.60754502e-01 3.66301894e-01 -1.21498480e-01
-3.80391121e-01 -9.67565417e-01 -3.01299393e-01 -6.23674579e-02
1.47065267e-01 -7.35917151e-01 3.96770507e-01 -7.95907199e-01
-1.54788327e+00 9.03933823e-01 9.39427037e-03 1.91841938e-03
2.48697072e-01 2.88584739e-01 -4.64772612e-01 -3.77825499e-01
-2.20060021e-01 3.73239934e-01 -3.26461583e-01 -1.00133252e+00
-5.20712018e-01 -5.69493175e-01 -5.69301903e-01 -2.19290316e-01
-1.58282846e-01 -6.10720403e-02 -1.64629489e-01 -4.83484417e-01
6.34693131e-02 -1.01524043e+00 -2.25707218e-01 -1.85458824e-01
-4.56015319e-01 -1.62723526e-01 6.39099121e-01 -4.36276942e-01
1.19141746e+00 -2.04141927e+00 -2.42280345e-02 3.03688288e-01
-4.82425392e-02 1.82664737e-01 -7.53715038e-02 5.14214277e-01
-9.35004503e-02 5.98651707e-01 -7.38404915e-02 3.01656485e-01
-3.92545134e-01 -7.01941773e-02 1.83562562e-01 4.22357023e-01
-8.00291821e-02 6.15659893e-01 -7.77841806e-01 -6.76024973e-01
-2.40733977e-02 4.48443025e-01 -2.77581960e-01 -6.36488665e-03
-2.53029078e-01 4.34643000e-01 -3.04949015e-01 1.19639385e+00
4.46192831e-01 -2.56477863e-01 3.71333122e-01 -3.30922931e-01
-1.47004724e-02 -4.81051534e-01 -1.08222890e+00 1.29962373e+00
-2.13381901e-01 5.93641818e-01 -2.85750568e-01 -7.38537967e-01
1.16254783e+00 2.30949298e-01 7.54421234e-01 -3.09536338e-01
4.41668689e-01 3.94990891e-01 2.78904468e-01 -7.66453087e-01
-1.64926663e-01 -2.88234919e-01 1.39973074e-01 2.02280805e-02
-1.71826124e-01 -3.07236649e-02 4.17716235e-01 -2.22156152e-01
1.38038397e+00 -8.10365193e-03 5.86134017e-01 -5.70078969e-01
7.04053223e-01 5.32742180e-02 1.00614452e+00 3.44693989e-01
-7.07774386e-02 7.41485000e-01 7.44795680e-01 -7.39827633e-01
-8.09706271e-01 -6.54731810e-01 -2.21989051e-01 8.78871202e-01
-4.59314149e-04 -3.31661582e-01 -4.12897676e-01 -6.61602795e-01
1.55148476e-01 2.99635917e-01 -6.29243433e-01 9.51077864e-02
8.85267742e-03 -1.66130137e+00 8.13760996e-01 2.67342180e-01
2.98653424e-01 -9.56863999e-01 -1.22903101e-01 1.07138544e-01
-7.66943395e-02 -6.27860665e-01 -2.82117009e-01 2.93773323e-01
-8.56956124e-01 -1.58712971e+00 -2.22081959e-01 -7.27581322e-01
6.82295978e-01 -3.51849526e-01 7.95652807e-01 5.42187154e-01
-6.40372813e-01 -2.78729439e-01 -3.75729382e-01 -7.09542990e-01
-2.88516164e-01 8.30693468e-02 2.79659718e-01 -1.08819284e-01
4.96309876e-01 -5.44797540e-01 -4.69668865e-01 5.97353637e-01
-7.17915773e-01 1.48424674e-02 6.75547421e-01 1.15542114e+00
1.04891777e+00 1.76988035e-01 1.04792976e+00 -1.26747334e+00
3.81167263e-01 -4.84810799e-01 -5.06715417e-01 2.97982305e-01
-8.67442548e-01 -1.27210379e-01 9.81591225e-01 -3.27708393e-01
-7.98536420e-01 1.33837864e-01 -3.39670509e-01 2.59335130e-01
-1.87097311e-01 5.16133249e-01 -3.57777923e-01 -5.98286353e-02
5.38595855e-01 3.47692490e-01 5.51777601e-01 -3.00842494e-01
-5.23932815e-01 7.71425903e-01 -8.04074109e-02 -2.90776283e-01
4.83039906e-03 5.05606771e-01 4.87601995e-01 -6.29051328e-01
-3.88103217e-01 -6.17147744e-01 -6.22054696e-01 -2.31310397e-01
6.02791190e-01 -5.16973972e-01 -1.23771513e+00 6.70104563e-01
-4.74090278e-01 -4.68379855e-01 1.52627947e-02 4.76670295e-01
-6.05053782e-01 1.29311517e-01 -6.51284277e-01 -5.81726789e-01
-5.47094047e-01 -1.03590155e+00 9.70211506e-01 3.11900020e-01
-3.00731301e-01 -9.59269464e-01 3.59264165e-01 2.71885872e-01
4.27109987e-01 4.49641377e-01 1.19395280e+00 -1.14992476e+00
-2.16245785e-01 -3.99466276e-01 -4.19296958e-02 -2.88429353e-02
3.49453449e-01 4.98878181e-01 -8.62866580e-01 -2.05990836e-01
-5.64781129e-01 -1.64754838e-01 6.07354581e-01 4.02465463e-01
1.60296702e+00 -7.53115341e-02 -9.25292075e-01 6.25141561e-01
1.81326151e+00 6.46864831e-01 9.23568428e-01 2.73727775e-01
2.75713116e-01 6.13918185e-01 7.85110533e-01 3.37925464e-01
-1.81679904e-01 3.84742051e-01 1.49004951e-01 4.95808804e-03
-5.53084835e-02 2.23090306e-01 -2.82762591e-02 9.32401001e-01
-4.69857991e-01 -5.57326019e-01 -1.06661797e+00 2.53548115e-01
-1.69474506e+00 -9.81193125e-01 -3.24640334e-01 2.06241846e+00
9.61108387e-01 -1.11257002e-01 1.11810081e-01 2.54080534e-01
6.60387099e-01 -5.91050327e-01 -5.44234157e-01 -1.70447543e-01
-6.29714966e-01 -1.24730483e-01 4.39494729e-01 -1.31457327e-02
-7.17592418e-01 4.23254788e-01 6.43014669e+00 1.06595874e+00
-1.23011005e+00 -1.36694741e-02 1.18203878e+00 -3.08110088e-01
1.02779135e-01 -1.70010403e-01 -8.67586970e-01 9.39180136e-01
8.67354274e-01 -2.33563557e-01 1.07181355e-01 5.10846794e-01
4.45758760e-01 -3.34933728e-01 -1.00073016e+00 8.89803767e-01
-1.35069549e-01 -1.69163799e+00 -1.90250263e-01 2.15072334e-01
7.78716683e-01 -2.78986692e-02 -5.38299441e-01 2.50916690e-01
2.02566728e-01 -1.12719071e+00 -2.91033000e-01 9.21292722e-01
1.00328267e+00 -6.51034832e-01 1.61471260e+00 3.36560339e-01
-7.51991391e-01 -5.53403534e-02 -3.51445675e-01 -4.30074371e-02
-1.33116230e-01 1.01659870e+00 -1.07330549e+00 4.60137308e-01
6.30188286e-01 4.48167026e-01 -4.86793071e-01 1.32660115e+00
4.81546700e-01 7.74905622e-01 -3.32149565e-01 -6.86271310e-01
-5.11125505e-01 -1.82620674e-01 2.60390639e-01 1.39121532e+00
4.95573193e-01 2.81417906e-01 1.09867856e-01 2.40856856e-01
-1.53449308e-02 6.38680518e-01 -4.89593208e-01 -1.86128840e-02
5.51366389e-01 1.63379776e+00 -1.23111260e+00 -1.47606924e-01
-3.14377427e-01 3.30302119e-01 6.73629402e-04 2.52371747e-03
-7.02879310e-01 -4.48980600e-01 5.69949329e-01 3.38278949e-01
-4.58220929e-01 3.89295071e-01 -6.05729163e-01 -6.05656862e-01
-6.89833879e-01 -7.43802488e-01 6.70182705e-01 -6.26103222e-01
-1.70551145e+00 1.80502787e-01 -3.92652839e-01 -1.24068666e+00
2.60151625e-01 -8.67329240e-01 -3.45004827e-01 5.90703368e-01
-8.82294118e-01 -9.02638078e-01 -5.10368586e-01 -1.05904154e-01
4.56040651e-01 -4.37358230e-01 1.17280781e+00 2.87727952e-01
-8.07898641e-01 4.00604606e-01 5.35162449e-01 1.23350091e-01
9.38332856e-01 -1.14096093e+00 -4.48208660e-01 1.06697418e-01
-5.12419581e-01 5.01466990e-01 6.59984231e-01 -9.32788491e-01
-1.52644336e+00 -1.15954435e+00 6.65586114e-01 -2.36257836e-01
3.46394598e-01 -2.04230502e-01 -7.16176391e-01 3.09394836e-01
-1.22400567e-01 2.58539408e-01 1.41537070e+00 1.71600878e-01
1.95887432e-01 -2.51560837e-01 -1.42388570e+00 4.60732430e-01
9.20981050e-01 -4.39187652e-03 1.34302258e-01 4.45671529e-01
1.67308122e-01 -3.14714760e-01 -1.35636187e+00 6.98521793e-01
1.00901902e+00 -6.83394909e-01 6.21091008e-01 -6.39276445e-01
4.65785116e-01 -5.24383187e-01 -1.82199165e-01 -1.36272931e+00
-4.53446507e-01 -2.51180738e-01 4.22519416e-01 1.29595983e+00
7.34791100e-01 -7.69724905e-01 9.77190793e-01 3.46277446e-01
-1.19786046e-01 -1.28147244e+00 -8.01104009e-01 -8.07368934e-01
-2.00893432e-01 -3.32962088e-02 7.21106529e-01 1.27144814e+00
3.85448813e-01 3.25183600e-01 -6.83518685e-03 -3.49564254e-01
1.10808142e-01 2.37892959e-02 8.24923217e-01 -1.45954072e+00
-1.67264476e-01 -6.32586956e-01 -7.94385433e-01 8.79995674e-02
1.24491416e-01 -1.05529094e+00 -9.02048573e-02 -1.43819058e+00
6.48793399e-01 -6.73030019e-01 -4.31035042e-01 5.38510442e-01
-3.26233417e-01 3.88035446e-01 -2.01223731e-01 -5.70714138e-02
-3.53087872e-01 -9.07988008e-03 1.21822679e+00 -1.48796529e-01
-3.06284368e-01 -1.12202011e-01 -5.19594014e-01 6.21371686e-01
1.06076705e+00 -3.28376263e-01 -3.01327229e-01 3.20514947e-01
5.10710657e-01 -1.81208886e-02 -9.03519019e-02 -8.91004622e-01
3.80918354e-01 -5.01640975e-01 6.59099579e-01 -4.76926655e-01
5.21617150e-03 -5.03153801e-01 8.09843123e-01 5.92881262e-01
-2.15102270e-01 4.23162729e-02 2.60722488e-01 4.24085170e-01
-4.05023471e-02 -1.70761272e-01 7.18353629e-01 -1.72386453e-01
-5.38230479e-01 1.75759166e-01 -4.69728589e-01 -1.85428321e-01
1.27038968e+00 -5.74938476e-01 -4.64403808e-01 9.89592895e-02
-7.55994916e-01 1.90415457e-01 7.20953405e-01 5.88426031e-02
3.73621911e-01 -1.10670841e+00 -6.88957095e-01 3.07174753e-02
2.72210687e-01 -2.14223161e-01 2.80548185e-01 1.23220849e+00
-8.16693187e-01 3.17843974e-01 -5.24133801e-01 -5.85523903e-01
-1.67419517e+00 4.66855168e-01 3.26555252e-01 -2.59360909e-01
1.07247725e-01 7.56939888e-01 -2.04387546e-01 -7.50847578e-01
-2.01919541e-01 -4.11379859e-02 -5.33377647e-01 2.76027054e-01
3.30278128e-01 7.33192742e-01 2.78914422e-01 -3.82997245e-01
-4.79538858e-01 6.80214763e-01 1.75306946e-01 5.02646983e-01
1.34226298e+00 1.38028279e-01 -7.23277092e-01 6.81092322e-01
1.09914231e+00 2.86539286e-01 -5.04711211e-01 6.42866075e-01
5.98520339e-02 -4.60475713e-01 -2.62301087e-01 -1.30636251e+00
-1.18198884e+00 2.64616907e-01 6.47202253e-01 6.85716197e-02
1.36525631e+00 -4.05834585e-01 4.10798371e-01 2.75905877e-01
5.93873441e-01 -1.26524615e+00 -3.22242707e-01 3.85843456e-01
5.26508272e-01 -1.19876218e+00 1.73914045e-01 -7.68216252e-01
-3.40373129e-01 1.30308211e+00 8.25314939e-01 3.92521083e-01
8.18674088e-01 7.56613672e-01 6.74702525e-02 -1.17388360e-01
-1.11310923e+00 2.00147003e-01 -2.48928249e-01 5.05330622e-01
8.88935804e-01 9.61545631e-02 -9.80367184e-01 1.00034380e+00
-1.11760192e-01 5.54011345e-01 5.86752951e-01 6.60703599e-01
-3.17319423e-01 -1.06279218e+00 -1.92779362e-01 1.22149658e+00
-7.55356669e-01 9.91197675e-02 -4.94267851e-01 9.15896535e-01
3.79072458e-01 7.31755495e-01 -1.83361880e-02 -5.90148270e-01
1.25537977e-01 3.52232188e-01 2.52222955e-01 -3.95475864e-01
-6.49470389e-01 -8.72301310e-02 3.52258652e-01 -5.07406175e-01
-3.11350971e-01 -7.05721617e-01 -1.61078727e+00 -2.80272007e-01
-6.54223859e-01 1.58068582e-01 6.00733578e-01 7.27842391e-01
5.96507609e-01 5.41687846e-01 3.78014117e-01 -2.00559020e-01
1.94113480e-03 -1.08626556e+00 -7.63710439e-01 4.77236062e-01
-3.10783058e-01 -5.06862104e-01 -4.35179472e-01 2.84905314e-01] | [5.94075870513916, 5.68444299697876] |
581cc176-2720-41b2-90f7-5bcc55a9b740 | beyond-auroc-co-for-evaluating-out-of | 2306.14658 | null | https://arxiv.org/abs/2306.14658v1 | https://arxiv.org/pdf/2306.14658v1.pdf | Beyond AUROC & co. for evaluating out-of-distribution detection performance | While there has been a growing research interest in developing out-of-distribution (OOD) detection methods, there has been comparably little discussion around how these methods should be evaluated. Given their relevance for safe(r) AI, it is important to examine whether the basis for comparing OOD detection methods is consistent with practical needs. In this work, we take a closer look at the go-to metrics for evaluating OOD detection, and question the approach of exclusively reducing OOD detection to a binary classification task with little consideration for the detection threshold. We illustrate the limitations of current metrics (AUROC & its friends) and propose a new metric - Area Under the Threshold Curve (AUTC), which explicitly penalizes poor separation between ID and OOD samples. Scripts and data are available at https://github.com/glhr/beyond-auroc | ['Thomas B. Moeslund', 'Sergio Escalera', 'Galadrielle Humblot-Renaux'] | 2023-06-26 | null | null | null | null | ['out-of-distribution-detection', 'ood-detection'] | ['computer-vision', 'computer-vision'] | [ 2.08259434e-01 -8.89901165e-03 -3.10324430e-01 -5.51907182e-01
-9.50804293e-01 -8.15606117e-01 9.12137687e-01 5.76542497e-01
-1.38908073e-01 2.76315510e-01 2.16649532e-01 -8.28751862e-01
-1.39588311e-01 -5.11915326e-01 -1.03737898e-01 -3.93227994e-01
-4.47648130e-02 3.33571821e-01 3.61654103e-01 3.55508566e-01
4.77391064e-01 6.78637087e-01 -1.71101189e+00 7.39241019e-02
9.30808187e-01 8.22193742e-01 -3.05012971e-01 7.59190738e-01
9.23530292e-03 4.70806837e-01 -9.73601758e-01 -4.54304606e-01
2.18200326e-01 -7.01679170e-01 -6.05957091e-01 -6.99159354e-02
5.50839961e-01 -4.33455527e-01 -1.26872808e-01 1.00141430e+00
6.86178088e-01 -1.77917689e-01 1.19740593e+00 -1.54390335e+00
-7.05449224e-01 2.81119376e-01 -4.26846802e-01 7.37405300e-01
4.47928488e-01 3.10356438e-01 1.06626093e+00 -5.69486618e-01
3.87822539e-01 1.27185225e+00 4.11793500e-01 2.51158684e-01
-1.21109867e+00 -4.57333356e-01 -7.71386474e-02 -5.78436293e-02
-1.46760392e+00 -2.71932900e-01 1.94033429e-01 -6.47625208e-01
9.46011066e-01 6.25842571e-01 5.60740888e-01 8.57112169e-01
2.70823836e-01 7.54704118e-01 1.36299860e+00 -6.96355522e-01
2.19779372e-01 5.33122241e-01 4.47129697e-01 3.15622509e-01
7.06016779e-01 4.91345748e-02 -2.64243662e-01 -4.00279820e-01
3.66946042e-01 -3.25752079e-01 1.89090624e-01 -3.07466894e-01
-6.33982122e-01 9.72096026e-01 -5.49457818e-02 3.66135150e-01
1.19236715e-01 -1.14538804e-01 3.68030667e-01 1.99287459e-01
6.17901325e-01 4.22247440e-01 -1.72680497e-01 -3.98192257e-01
-8.53757799e-01 4.47782129e-01 9.19313729e-01 8.03753793e-01
3.69589359e-01 -5.61953709e-02 -1.92954108e-01 9.75252569e-01
4.78601754e-01 4.05370861e-01 8.37100521e-02 -9.65840936e-01
5.51363220e-03 4.99119043e-01 1.63632810e-01 -7.88634717e-01
-2.75637865e-01 -7.93098286e-02 -4.90802154e-02 3.00720632e-01
6.76879287e-01 -6.93119457e-03 -7.80517399e-01 1.33006752e+00
3.81518662e-01 -5.12539566e-01 -2.85986841e-01 7.45887637e-01
4.98114377e-01 4.09811616e-01 6.62838593e-02 -3.25345621e-02
1.54744172e+00 -4.29400831e-01 -5.98812938e-01 -3.13386917e-01
8.73827517e-01 -1.00804663e+00 1.04380310e+00 4.60142374e-01
-7.83732295e-01 -1.51143879e-01 -1.06706047e+00 -7.72989094e-02
-5.71333110e-01 -2.46348143e-01 6.83317661e-01 1.25100231e+00
-8.63027990e-01 2.26631775e-01 -6.06390655e-01 -7.89291441e-01
1.95322171e-01 1.90351516e-01 -9.92049370e-03 -1.49119496e-01
-1.02245104e+00 9.89118159e-01 5.01812771e-02 -8.03834423e-02
-8.24167967e-01 -2.77523637e-01 -6.79186940e-01 -3.18794757e-01
1.86056525e-01 -4.61784564e-02 1.33525705e+00 -6.06196284e-01
-9.90291238e-01 1.00015128e+00 -6.95517361e-02 -1.73021451e-01
5.15812695e-01 -1.03188448e-01 -6.20874584e-01 -1.51137682e-02
2.37851962e-01 5.79759419e-01 4.96786445e-01 -1.27433121e+00
-7.99127817e-01 -3.96845341e-01 -7.89976493e-03 4.65882421e-02
-1.27375424e-01 6.18621051e-01 -4.25810099e-01 -5.59650540e-01
-1.74673453e-01 -8.88158202e-01 2.54625887e-01 -8.71587023e-02
-3.72419804e-01 -6.71282470e-01 6.25295341e-01 -6.34006500e-01
1.68061173e+00 -2.07741928e+00 -6.93052173e-01 3.75981748e-01
1.71239108e-01 1.56528577e-01 -7.77328536e-02 5.17484546e-01
3.59328203e-02 3.54109049e-01 -1.94170326e-01 1.75677955e-01
3.79146218e-01 -1.30104274e-02 -2.28157088e-01 7.67768800e-01
2.39792228e-01 2.81247199e-01 -9.40383852e-01 -5.30487478e-01
1.68486461e-01 5.01022279e-01 -4.43277210e-01 1.69388726e-01
1.38141915e-01 -2.34483406e-01 -3.88981491e-01 1.05651987e+00
7.02594101e-01 3.88079807e-02 1.17835104e-01 2.04487547e-01
-4.79428947e-01 6.68029964e-01 -1.25731587e+00 7.13830471e-01
3.84730473e-02 8.46498370e-01 -1.09185718e-01 -6.90453470e-01
9.63451385e-01 -2.71882582e-02 2.86536723e-01 -7.02163935e-01
2.95469463e-01 5.82603753e-01 4.72156465e-01 -2.62127191e-01
4.36053067e-01 -2.45561022e-02 -1.30262986e-01 5.74974895e-01
-6.80734739e-02 -2.03927338e-01 4.40188289e-01 2.34203916e-02
1.17211497e+00 -1.15160085e-01 3.89352292e-01 -5.22948861e-01
1.28483623e-01 1.76199958e-01 2.60734648e-01 8.68991256e-01
-5.26025116e-01 6.34276748e-01 9.27294433e-01 3.48103531e-02
-8.81269753e-01 -1.17796874e+00 -9.06861186e-01 8.76116872e-01
-2.06163935e-02 -4.32346523e-01 -9.94537592e-01 -7.70694137e-01
4.18596238e-01 1.07329881e+00 -6.76099062e-01 -5.15054576e-02
-1.87087119e-01 -1.16749072e+00 8.24193180e-01 2.57982135e-01
-1.23460725e-01 -7.17671514e-01 -6.43534243e-01 7.28786178e-03
2.06789702e-01 -6.86301649e-01 -4.50270981e-01 4.94745791e-01
-7.12549031e-01 -1.08331418e+00 -8.19126606e-01 -3.64219397e-01
4.95363116e-01 2.57666767e-01 9.50233698e-01 -1.14264160e-01
-5.81627429e-01 6.48577452e-01 -4.60017383e-01 -7.04261720e-01
-7.79048622e-01 1.53794633e-02 -4.03854251e-02 -4.69461709e-01
1.09610355e+00 -2.69854777e-02 -6.29497766e-01 5.60561538e-01
-8.94172370e-01 -6.67607844e-01 5.35433650e-01 3.05631250e-01
3.08343470e-01 -1.01588115e-01 4.10365701e-01 -8.66250157e-01
9.16154683e-01 -5.93076289e-01 -4.91312772e-01 -7.94838220e-02
-1.10805750e+00 -1.34464756e-01 6.69805184e-02 -2.97695339e-01
-7.25432158e-01 -4.73014265e-01 -2.36348972e-01 -6.02685101e-02
-3.52745265e-01 9.67763513e-02 -6.75500929e-02 1.85054034e-01
7.49883533e-01 -3.11409682e-01 3.31785977e-02 -4.97392118e-01
8.16964954e-02 1.31022811e+00 6.83445707e-02 -4.24335241e-01
5.26092887e-01 2.94465870e-01 -4.22297984e-01 -9.72983956e-01
-6.59706473e-01 -6.88413322e-01 -3.24655920e-01 -2.54114836e-01
6.96188629e-01 -6.59407616e-01 -3.89353335e-01 2.82852739e-01
-7.12169468e-01 -5.15970051e-01 -9.12683457e-02 5.29221177e-01
-2.93344200e-01 4.65526402e-01 -3.25094908e-01 -1.13781071e+00
-5.62292300e-02 -1.12845612e+00 9.63334382e-01 1.19687259e-01
-7.95643568e-01 -9.07348633e-01 1.92813069e-01 3.18819880e-01
1.90815508e-01 1.89317316e-01 7.99914420e-01 -9.60700035e-01
-1.88496187e-01 -5.70908427e-01 -3.41350526e-01 3.67348105e-01
2.74576575e-01 2.45118171e-01 -1.13479245e+00 -1.67760193e-01
-1.00410134e-01 -1.10488169e-01 5.35174310e-01 4.71205354e-01
8.46701086e-01 -1.09811790e-01 -3.86408031e-01 2.16258839e-01
1.30780220e+00 5.07847309e-01 6.14246130e-01 5.60551465e-01
3.20532262e-01 8.09469283e-01 7.46909618e-01 3.97845745e-01
3.00633490e-01 8.15395951e-01 8.04632977e-02 1.21643450e-02
-2.79522479e-01 -1.49201542e-01 5.27300954e-01 5.01968563e-01
2.53818989e-01 -6.97730482e-01 -1.10901988e+00 6.42576993e-01
-1.24095464e+00 -8.50949109e-01 -3.27319831e-01 2.43256235e+00
6.05892479e-01 5.53032815e-01 6.53518021e-01 1.93106905e-01
7.59967208e-01 1.65860325e-01 -3.71854186e-01 -9.65783000e-01
1.42896786e-01 -1.12681843e-01 6.01019740e-01 5.76729417e-01
-9.41341698e-01 3.72225493e-01 7.63430595e+00 8.16093743e-01
-8.81542861e-01 -5.73475063e-02 8.70750964e-01 2.42810294e-01
-3.44720602e-01 1.62207663e-01 -1.06330264e+00 5.81111848e-01
1.11850774e+00 -1.07310265e-01 1.12619646e-01 7.45938599e-01
2.95336127e-01 -6.14776254e-01 -1.07722139e+00 4.66072679e-01
2.28391573e-01 -4.22148168e-01 -1.60373434e-01 3.27792495e-01
4.42995071e-01 1.24832634e-02 1.35123804e-01 2.06090495e-01
3.79770279e-01 -8.28697860e-01 7.45001316e-01 1.02243878e-01
7.82370090e-01 -6.05937064e-01 7.62145817e-01 -1.19801253e-01
-9.16438460e-01 1.23247981e-01 -2.83194900e-01 8.62351432e-02
-2.09391162e-01 6.85498059e-01 -1.00735283e+00 8.05215761e-02
8.65483105e-01 1.76792651e-01 -8.17749679e-01 1.10852921e+00
1.63857229e-02 8.44015479e-01 -5.50612211e-01 -2.97679484e-01
2.75806218e-01 1.88563894e-02 4.60578591e-01 1.46351814e+00
3.88766915e-01 -2.21193641e-01 -2.25282997e-01 8.42403293e-01
4.04420406e-01 1.98661849e-01 -6.50247991e-01 -3.67315680e-01
6.51620507e-01 1.05921054e+00 -1.10401487e+00 -1.33223265e-01
-7.37432837e-01 6.44855440e-01 -4.43909131e-02 1.00370251e-01
-7.50036061e-01 -7.76394665e-01 7.06153154e-01 3.57495964e-01
3.92617919e-02 9.62946340e-02 -5.35338223e-01 -6.57015085e-01
-5.60816452e-02 -1.13374281e+00 7.06854939e-01 -4.89122957e-01
-1.25107789e+00 1.88727602e-01 3.51027697e-01 -1.32790351e+00
-1.34096205e-01 -7.66308129e-01 -6.45271003e-01 7.12297976e-01
-1.10632253e+00 -5.86259902e-01 -2.28454456e-01 -2.13067129e-01
4.07169968e-01 2.58111268e-01 4.70934808e-01 2.46503621e-01
-4.92949367e-01 7.90441871e-01 3.00539345e-01 7.32134581e-02
1.00281298e+00 -1.43846714e+00 3.77846837e-01 7.20644116e-01
-1.38000816e-01 6.61677957e-01 1.09381413e+00 -6.91061735e-01
-1.13296533e+00 -6.26391232e-01 8.83424699e-01 -8.99723172e-01
8.00551534e-01 -3.92959863e-01 -7.45194077e-01 5.76882839e-01
7.18883500e-02 -4.43931013e-01 8.55379581e-01 1.79483876e-01
-2.69187033e-01 1.15806959e-03 -1.17756152e+00 5.71987629e-01
7.78989673e-01 -4.07735229e-01 -5.15551329e-01 2.53728300e-01
1.94769964e-01 -6.41797110e-02 -1.09293079e+00 2.07062751e-01
7.12813199e-01 -1.14614832e+00 7.95130372e-01 4.37996686e-02
-3.46473642e-02 -3.12489212e-01 -1.44818738e-01 -9.27707493e-01
-1.49908483e-01 -3.97119522e-01 1.56301633e-01 1.44473469e+00
3.67389262e-01 -8.19435835e-01 4.78909552e-01 5.92060208e-01
-1.09942868e-01 -8.14267635e-01 -6.87950194e-01 -1.06301689e+00
2.51563311e-01 -5.83395302e-01 4.29941863e-01 9.61536705e-01
2.28764370e-01 6.12606704e-02 2.91041471e-02 2.37915590e-01
5.87948442e-01 -1.66180119e-01 6.86462939e-01 -1.15924954e+00
-1.43768668e-01 -7.02941120e-01 -4.61682409e-01 -7.61664867e-01
-3.55076969e-01 -8.83753419e-01 -4.59964164e-02 -1.62159026e+00
3.45238417e-01 -3.59869242e-01 -3.52376282e-01 3.49561989e-01
-2.20921859e-01 3.73247921e-01 4.39232700e-02 2.96296537e-01
-4.38655078e-01 7.65558556e-02 7.96765566e-01 5.80004267e-02
-2.25215524e-01 -1.95679739e-01 -9.59899664e-01 5.23790061e-01
9.25305903e-01 -7.67081320e-01 -2.06916302e-01 -1.27027482e-01
-1.14958351e-02 -3.08047742e-01 1.59221604e-01 -9.50717688e-01
-2.75703162e-01 -1.17183834e-01 2.58509219e-01 -5.09672344e-01
-1.74562737e-01 -4.76228714e-01 -5.55456579e-02 5.49799502e-01
-2.78696299e-01 2.53955871e-01 3.35720748e-01 3.33290368e-01
1.84717812e-02 -5.27863920e-01 8.33427906e-01 2.20659167e-01
-4.13206309e-01 -9.24421474e-02 -7.82818019e-01 3.34259063e-01
9.64738965e-01 -4.39342231e-01 -5.21576941e-01 -2.70103157e-01
-3.84702325e-01 2.31615335e-01 7.26063669e-01 5.11834502e-01
1.67913154e-01 -1.09222281e+00 -5.73671162e-01 1.62955835e-01
5.31967521e-01 -6.12607062e-01 -1.45238951e-01 8.77484202e-01
-7.31932223e-01 6.07412994e-01 9.91021991e-02 -4.86544728e-01
-1.34208632e+00 5.16463220e-01 3.61222506e-01 7.07616331e-03
-2.86148190e-01 6.28014266e-01 2.29141429e-01 -4.82883394e-01
3.51368189e-01 -1.95492446e-01 -1.00699268e-01 2.50824839e-01
4.76211458e-01 8.22803080e-01 -2.81211138e-02 -4.45410520e-01
-6.07802808e-01 1.17792502e-01 -1.74247682e-01 -2.07986102e-01
1.03882468e+00 -5.63581251e-02 -5.21514975e-02 7.39886582e-01
1.07286811e+00 3.20300728e-01 -9.76702750e-01 4.25991803e-01
3.05078775e-01 -7.22202778e-01 3.83082926e-02 -8.32992733e-01
-3.37557316e-01 8.06739151e-01 9.62882638e-01 9.22872901e-01
8.29592288e-01 2.44854212e-01 2.64018118e-01 1.03736013e-01
-4.39313948e-02 -1.30746424e+00 -3.07528488e-02 2.26229489e-01
6.34565175e-01 -9.58727777e-01 3.41681838e-01 -4.35080081e-01
-5.71677148e-01 8.87871087e-01 5.59349716e-01 -4.23671715e-02
6.00663781e-01 3.29645157e-01 1.36024177e-01 -1.97064191e-01
-5.94719291e-01 -2.50743389e-01 1.87734842e-01 7.45496511e-01
8.09748471e-01 4.59354930e-02 -7.31184125e-01 1.96243331e-01
-1.85931876e-01 -2.02551261e-01 5.74830353e-01 9.77263868e-01
-7.06825972e-01 -1.16806269e+00 -6.44949794e-01 1.04596829e+00
-7.36712158e-01 -9.36331600e-03 -7.43136525e-01 1.08632660e+00
2.24169083e-02 1.01334858e+00 3.02736849e-01 -2.35061541e-01
3.51874381e-01 2.36809969e-01 3.03158820e-01 -6.31689310e-01
-3.40016693e-01 3.70617241e-01 4.58994806e-01 -3.07984680e-01
-2.58856207e-01 -1.21828711e+00 -9.21801746e-01 -2.96803951e-01
-5.41891456e-01 1.12735875e-01 4.31489736e-01 6.64067924e-01
2.06579626e-01 1.91950500e-01 4.44933772e-01 -3.65897954e-01
-5.59404254e-01 -9.21430290e-01 -8.76635671e-01 1.77372828e-01
2.55845338e-01 -7.75078118e-01 -9.25136805e-01 -2.83775538e-01] | [9.14396858215332, 3.3568639755249023] |
7110d9fa-e96a-4fc1-a139-77764a698acc | learning-joint-wasserstein-auto-encoders-for | null | null | https://openreview.net/forum?id=HJe3TsR5K7 | https://openreview.net/pdf?id=HJe3TsR5K7 | Learning Joint Wasserstein Auto-Encoders for Joint Distribution Matching | We study the joint distribution matching problem which aims at learning bidirectional mappings to match the joint distribution of two domains. This problem occurs in unsupervised image-to-image translation and video-to-video synthesis tasks, which, however, has two critical challenges: (i) it is difficult to exploit sufficient information from the joint distribution; (ii) how to theoretically and experimentally evaluate the generalization performance remains an open question. To address the above challenges, we propose a new optimization problem and design a novel Joint Wasserstein Auto-Encoders (JWAE) to minimize the Wasserstein distance of the joint distributions in two domains. We theoretically prove that the generalization ability of the proposed method can be guaranteed by minimizing the Wasserstein distance of joint distributions. To verify the generalization ability, we apply our method to unsupervised video-to-video synthesis by performing video frame interpolation and producing visually smooth videos in two domains, simultaneously. Both qualitative and quantitative comparisons demonstrate the superiority of our method over several state-of-the-arts. | ['Mingkui Tan', 'Junzhou Huang', 'Peilin Zhao', 'Langyuan Mo', 'Yong Guo', 'JieZhang Cao'] | 2018-09-27 | null | null | null | null | ['video-to-video-synthesis'] | ['computer-vision'] | [ 3.76036376e-01 -7.60092214e-02 -4.59166393e-02 -1.36327088e-01
-9.50659394e-01 -4.68438119e-01 6.17145717e-01 -3.75093460e-01
-1.28353417e-01 8.91555071e-01 -1.91176739e-02 -7.32319653e-02
-2.96593785e-01 -5.85004091e-01 -9.40413594e-01 -7.93711662e-01
3.13608684e-02 1.38261303e-01 4.51387949e-02 1.05827618e-02
1.30428657e-01 3.02812636e-01 -1.57027137e+00 -1.16917096e-01
1.11035478e+00 9.36459601e-01 4.52407867e-01 6.52277887e-01
1.03913106e-01 6.83231950e-01 -3.53928477e-01 -3.07777822e-01
4.54950690e-01 -9.56753194e-01 -5.90140283e-01 5.23605227e-01
5.91797888e-01 -3.66286755e-01 -3.82590950e-01 1.52836156e+00
2.67485470e-01 2.45452806e-01 1.00056624e+00 -1.52939272e+00
-1.00514436e+00 3.46300714e-02 -6.11859560e-01 -2.17292067e-02
3.32298785e-01 5.50851487e-02 8.94148648e-01 -7.33341396e-01
7.12728679e-01 1.27183759e+00 4.07281935e-01 4.64561820e-01
-1.45432591e+00 -5.79279542e-01 -1.10382549e-02 1.02940425e-01
-1.50652277e+00 -4.19331670e-01 7.63623357e-01 -7.81617463e-01
2.98285395e-01 2.55688233e-03 3.58855754e-01 9.51872766e-01
-1.50437821e-02 6.07422769e-01 1.16062307e+00 -3.95837992e-01
1.76013947e-01 2.29133978e-01 -6.37744606e-01 7.08693206e-01
9.15668830e-02 1.48625076e-01 -5.45369446e-01 1.05374664e-01
1.09582901e+00 -1.36551812e-01 -5.09342074e-01 -5.06569564e-01
-1.50258696e+00 6.26586616e-01 1.15842894e-01 2.75554240e-01
-1.47519588e-01 1.80616423e-01 7.38083273e-02 5.61623037e-01
4.41714257e-01 2.53001988e-01 -2.40899269e-02 -1.77417368e-01
-8.82321715e-01 4.05396461e-01 6.21740997e-01 1.34184766e+00
9.42183018e-01 1.26975581e-01 -1.36362702e-01 6.91091418e-01
3.33992273e-01 7.68440783e-01 2.56411493e-01 -1.18219614e+00
5.51042140e-01 1.50655538e-01 3.66692811e-01 -1.20454383e+00
3.32102180e-01 -1.18534744e-03 -9.53683376e-01 1.74835458e-01
6.77488327e-01 -3.09003741e-02 -5.15981019e-01 1.91021991e+00
2.63958305e-01 2.83874959e-01 1.56885143e-02 1.00227094e+00
2.96312034e-01 6.80852652e-01 -2.06959680e-01 -5.59942424e-01
8.09189677e-01 -7.76875854e-01 -9.78232086e-01 2.08515793e-01
3.44260275e-01 -9.01774168e-01 1.16258144e+00 -1.77453514e-02
-1.18983603e+00 -6.89519227e-01 -1.05749452e+00 8.95963833e-02
1.90858975e-01 2.17652738e-01 3.62565778e-02 3.64099234e-01
-9.21272099e-01 7.36434519e-01 -7.90936053e-01 -4.24308002e-01
1.64918184e-01 1.69275597e-01 -4.40965354e-01 -3.22290249e-02
-1.12513959e+00 6.36996746e-01 2.06587180e-01 -1.90302074e-01
-9.67950642e-01 -7.36374438e-01 -8.59590828e-01 2.52273325e-02
1.85215458e-01 -7.03189254e-01 1.00515759e+00 -1.25358498e+00
-1.80667067e+00 7.44105041e-01 -1.67422444e-01 -1.35636553e-01
7.53251076e-01 -2.92517506e-02 -2.21495137e-01 1.11030504e-01
3.15242559e-01 7.16337502e-01 1.04764450e+00 -1.17173648e+00
-7.08435655e-01 -1.29953742e-01 1.13493269e-02 2.29254529e-01
-5.77333748e-01 -1.87480569e-01 -3.41770887e-01 -9.10435319e-01
7.43314531e-03 -8.68969560e-01 6.34386018e-02 3.49965304e-01
-2.49604598e-01 -3.24941278e-02 7.38206685e-01 -6.44480348e-01
1.13933706e+00 -2.42762589e+00 5.24385750e-01 -1.28363580e-01
3.85200195e-02 -1.37264267e-01 -1.74294263e-01 3.08767527e-01
1.20810762e-01 -9.80917271e-03 -4.77949053e-01 -1.81531250e-01
8.05271342e-02 2.44534001e-01 -4.03035223e-01 7.26662219e-01
3.04109514e-01 6.58243179e-01 -1.09441245e+00 -6.25278175e-01
1.35798484e-01 3.59554708e-01 -7.03054130e-01 6.32162035e-01
-1.13067642e-01 8.60277772e-01 -3.79791468e-01 3.82289767e-01
7.42412269e-01 -2.17291862e-01 2.20365122e-01 -3.37141544e-01
7.32172281e-02 -7.18888938e-02 -1.20662189e+00 1.71632493e+00
-2.71760345e-01 8.22248220e-01 3.16815823e-02 -1.26921713e+00
9.10357356e-01 2.97917873e-01 5.88953018e-01 -5.39443135e-01
-1.37952387e-01 5.55258453e-01 -1.55519053e-01 -7.32054353e-01
3.35056871e-01 -6.66243911e-01 2.05707952e-01 2.61542350e-01
1.72902495e-01 -2.19861329e-01 2.40265295e-01 -6.39996752e-02
7.30951607e-01 4.49157596e-01 2.45087221e-02 -5.58847845e-01
6.71249390e-01 -3.46131265e-01 6.67061269e-01 4.83531982e-01
-2.25806892e-01 8.07104886e-01 6.52442813e-01 -1.44021109e-01
-1.47236073e+00 -1.41193426e+00 -3.13777179e-02 5.02687573e-01
3.48886937e-01 -1.22150205e-01 -8.64073515e-01 -6.74719810e-01
-7.17073381e-02 4.77596760e-01 -3.95194530e-01 -1.33525683e-02
-3.70941669e-01 -2.83907741e-01 4.19990659e-01 4.30126011e-01
6.21151388e-01 -4.91788030e-01 -2.62520403e-01 -9.01562721e-03
-4.43170041e-01 -1.48843694e+00 -8.75049889e-01 -4.39595729e-01
-7.30523527e-01 -8.74552488e-01 -1.17050421e+00 -9.47997034e-01
7.25234628e-01 2.43638858e-01 8.82503033e-01 -2.00163051e-01
1.20636709e-01 4.61662859e-01 -1.20061532e-01 1.44232601e-01
-8.22948098e-01 -1.76361620e-01 2.84684986e-01 3.75058323e-01
9.13910381e-03 -8.20724547e-01 -5.45698524e-01 8.37168694e-01
-1.12986302e+00 1.89277053e-01 3.59909624e-01 8.14881265e-01
7.43318737e-01 1.63768709e-01 6.64256632e-01 -4.22265857e-01
5.56056380e-01 -3.04052919e-01 -8.91962707e-01 3.76779169e-01
-3.44844699e-01 2.49699518e-01 7.15741396e-01 -5.78527331e-01
-9.34479475e-01 3.03041022e-02 1.42591134e-01 -8.21226716e-01
3.02035268e-02 3.42110276e-01 -3.88568074e-01 2.31061112e-02
4.41738009e-01 3.87675673e-01 3.93668562e-01 -2.29009002e-01
5.14186859e-01 7.33641684e-01 7.68255293e-01 -8.71489763e-01
1.04195654e+00 6.39815927e-01 1.43666565e-01 -9.77757871e-01
-6.02682769e-01 -2.03126058e-01 -5.21625340e-01 -2.74928033e-01
1.09589481e+00 -8.95313859e-01 -6.35264874e-01 2.97725022e-01
-1.24310446e+00 -4.06472892e-01 -2.09615543e-01 5.97511113e-01
-1.05438292e+00 6.74634993e-01 -3.86572033e-01 -7.14029431e-01
1.67008817e-01 -1.33388186e+00 1.11834729e+00 2.34552030e-03
-4.50003408e-02 -1.19903493e+00 1.09312795e-01 -7.50140846e-02
3.38532269e-01 3.70554715e-01 9.10765648e-01 -1.07124567e-01
-6.86400354e-01 2.03624561e-01 -3.57401073e-01 7.51467764e-01
4.67820376e-01 8.45905468e-02 -5.56569159e-01 -4.36947256e-01
1.22373767e-01 -2.22711146e-01 5.11079371e-01 2.88617909e-01
1.19810796e+00 -3.71676594e-01 -6.20503090e-02 7.06512153e-01
1.38458490e+00 -1.66593120e-03 7.60946274e-01 1.96648821e-01
6.33426547e-01 6.54919386e-01 6.23768270e-01 3.77920270e-01
1.93237886e-01 9.04127717e-01 2.02219918e-01 2.35382453e-01
-2.58851111e-01 -5.49384534e-01 5.82727075e-01 1.00184822e+00
-1.64055645e-01 -2.27217421e-01 -4.91408944e-01 6.22169256e-01
-1.97450840e+00 -9.58890617e-01 1.52562603e-01 2.39822245e+00
8.41342449e-01 -2.07011923e-01 9.24122985e-03 -5.34626618e-02
9.27095711e-01 3.19890976e-02 -2.66970634e-01 7.01434091e-02
-5.15164621e-02 -1.97966680e-01 3.53242666e-01 5.39902687e-01
-1.06617188e+00 5.41638076e-01 6.22207165e+00 1.12380791e+00
-1.05329752e+00 1.09748602e-01 5.37956655e-01 1.90497100e-01
-5.92684329e-01 7.94935077e-02 -4.16320145e-01 6.68102145e-01
5.88982582e-01 -2.74695396e-01 6.42541349e-01 6.38027906e-01
1.57916248e-01 2.31189147e-01 -1.30769253e+00 1.10944700e+00
4.00014818e-02 -1.34567642e+00 2.17423886e-01 1.44440457e-01
1.09800792e+00 -4.02860910e-01 2.55252957e-01 9.34733427e-04
4.67261113e-02 -8.74350786e-01 8.83586168e-01 6.37317181e-01
1.16103554e+00 -5.43606400e-01 3.54120940e-01 3.30961347e-01
-1.10541368e+00 1.14568569e-01 -4.68965590e-01 1.66488901e-01
2.98928261e-01 5.85365415e-01 -3.05868298e-01 6.05090737e-01
5.39608479e-01 1.05302656e+00 -3.92677039e-02 8.98851335e-01
-2.16070935e-01 2.34630421e-01 -1.08072706e-01 2.76638955e-01
3.47196893e-03 -6.14914775e-01 8.05777192e-01 9.57223594e-01
7.72253394e-01 -2.57026553e-01 1.51709586e-01 1.23925400e+00
-2.15357006e-01 -3.33072469e-02 -6.74040198e-01 -3.37412119e-01
2.77742207e-01 8.19413841e-01 -4.76177096e-01 -2.43208617e-01
-4.58581984e-01 1.18327582e+00 2.08075851e-01 6.97930932e-01
-9.46779549e-01 -3.73974770e-01 8.20760787e-01 2.75886834e-01
3.88814718e-01 -4.36399758e-01 -5.37761971e-02 -1.45731807e+00
2.28230432e-01 -8.81864309e-01 -4.38217632e-02 -6.81989431e-01
-1.47988665e+00 4.35533524e-01 5.94534092e-02 -1.68378699e+00
-3.07383418e-01 -4.97805983e-01 -3.16415101e-01 7.38790095e-01
-1.51944637e+00 -9.84571815e-01 -3.71353418e-01 6.75659895e-01
4.19476688e-01 -3.17527652e-01 4.84273762e-01 7.08164930e-01
-2.15756357e-01 6.91590369e-01 4.33423251e-01 6.75876513e-02
8.36386919e-01 -1.20299709e+00 8.98401737e-02 8.58755887e-01
-6.99809343e-02 3.82694870e-01 8.20298493e-01 -5.04364312e-01
-1.44289863e+00 -1.15003991e+00 6.35979176e-01 -2.51738578e-01
7.98347116e-01 -2.48628303e-01 -8.97621870e-01 5.78760266e-01
2.64016807e-01 1.83465689e-01 3.65300238e-01 -5.08424282e-01
-2.28685796e-01 -1.62778378e-01 -1.13579798e+00 6.21497571e-01
1.12358856e+00 -6.71011031e-01 -3.24089378e-01 3.19386601e-01
7.62083471e-01 -2.59092927e-01 -1.04447508e+00 4.56737578e-01
6.03381574e-01 -9.60807562e-01 8.22134554e-01 -4.34763461e-01
9.25640643e-01 -6.08374178e-01 -6.74839973e-01 -1.44116807e+00
-8.40870664e-02 -9.23100829e-01 -2.40769964e-02 1.33160567e+00
1.63654506e-01 -6.49862647e-01 4.61243719e-01 2.87650019e-01
6.09129220e-02 -5.83660305e-01 -1.04074514e+00 -1.19550681e+00
2.46074662e-01 -2.26797894e-01 5.47618747e-01 8.41689527e-01
-1.59591421e-01 1.36363909e-01 -6.42979741e-01 1.13234229e-01
7.68967807e-01 9.84346122e-02 9.45243120e-01 -7.06825137e-01
-5.64894795e-01 -4.78637755e-01 -5.83404899e-01 -1.49650240e+00
2.89225876e-01 -7.48744905e-01 2.58249462e-01 -1.29343355e+00
1.56441569e-01 -1.99711904e-01 -1.19656242e-01 -1.63689882e-01
-7.47191533e-02 1.32687703e-01 2.64185011e-01 4.88081783e-01
-4.18641835e-01 9.85150218e-01 1.64923716e+00 -1.73307240e-01
5.28926216e-02 -1.93109185e-01 -4.10312474e-01 5.71471453e-01
3.01436305e-01 -3.47398490e-01 -6.17388964e-01 -5.00633419e-01
2.05590472e-01 3.15035641e-01 3.52046460e-01 -8.63013148e-01
-2.84003075e-02 -4.96743560e-01 -6.17767051e-02 -1.53571382e-01
2.80237645e-02 -8.24806035e-01 1.85031459e-01 2.46758685e-01
-3.59922647e-01 1.12826666e-02 -1.17464133e-01 8.25743198e-01
-5.33722937e-01 -2.97048624e-04 1.02301264e+00 2.14924067e-01
-4.09746617e-01 4.21593577e-01 -3.23953390e-01 1.55770823e-01
1.03826952e+00 -1.35731697e-01 -5.22630848e-02 -6.75545394e-01
-5.04959345e-01 2.50266585e-02 7.63700366e-01 3.49576741e-01
6.44950628e-01 -1.81423724e+00 -8.15267026e-01 3.95102650e-01
2.09139228e-01 -1.30283877e-01 5.62820844e-02 1.13355386e+00
-6.58993304e-01 1.01843804e-01 -1.66592523e-01 -8.86389554e-01
-8.35135102e-01 6.33992434e-01 4.79459703e-01 -1.02817453e-01
-3.74561727e-01 5.40026128e-01 6.01477742e-01 -3.94803911e-01
7.79035911e-02 -2.95476735e-01 2.36829028e-01 -2.62532443e-01
2.97923177e-01 4.36305672e-01 -3.15156519e-01 -6.68551862e-01
-1.59918606e-01 7.44581401e-01 3.18067908e-01 -2.96815872e-01
1.16172659e+00 -3.76643658e-01 -7.71761090e-02 1.85529083e-01
1.65815580e+00 -1.64899364e-01 -1.69115686e+00 -2.25780740e-01
-1.59752771e-01 -1.01774490e+00 -2.41101891e-01 -4.81917039e-02
-1.04320228e+00 1.01927710e+00 5.20444572e-01 2.51702011e-01
1.09512091e+00 -1.00713916e-01 8.17997158e-01 2.43700501e-02
2.52959132e-01 -1.03243697e+00 3.67761672e-01 1.59112468e-01
9.48094070e-01 -1.10928237e+00 -1.58043414e-01 -4.15266246e-01
-6.30364716e-01 1.05523884e+00 3.75181347e-01 -3.62257719e-01
7.53563285e-01 3.20603698e-02 -1.96943924e-01 2.36723334e-01
-4.95365500e-01 -9.41393971e-02 4.37073708e-01 6.51006103e-01
4.24559087e-01 -1.50532098e-02 -3.09202492e-01 4.69675474e-02
8.89427438e-02 1.27777085e-01 4.31380123e-01 5.62051415e-01
-2.00365320e-01 -1.09092438e+00 -3.02754253e-01 8.69889781e-02
-3.53415370e-01 1.81217507e-01 5.56195341e-02 6.03974342e-01
-5.83542064e-02 8.31284404e-01 8.22797939e-02 -3.50983560e-01
1.81943834e-01 -2.09684908e-01 6.95370555e-01 -2.49499664e-01
3.45224887e-01 3.66441280e-01 -3.11135173e-01 -3.47100198e-01
-7.14300513e-01 -6.53914273e-01 -9.12452042e-01 -3.15468699e-01
-2.53812015e-01 1.07068922e-02 4.31076020e-01 9.81994867e-01
3.04509103e-01 4.00090039e-01 8.04022968e-01 -7.38265634e-01
-8.88167143e-01 -8.26248705e-01 -8.41027379e-01 7.49653876e-01
4.72303152e-01 -8.25335443e-01 -4.08554435e-01 4.57717180e-01] | [11.292860984802246, -0.6712010502815247] |
17d0f451-5d2f-4f58-8b79-68874fa1e1dc | answering-questions-by-meta-reasoning-over | 2304.13007 | null | https://arxiv.org/abs/2304.13007v2 | https://arxiv.org/pdf/2304.13007v2.pdf | Answering Questions by Meta-Reasoning over Multiple Chains of Thought | Modern systems for multi-hop question answering (QA) typically break questions into a sequence of reasoning steps, termed chain-of-thought (CoT), before arriving at a final answer. Often, multiple chains are sampled and aggregated through a voting mechanism over the final answers, but the intermediate steps themselves are discarded. While such approaches improve performance, they do not consider the relations between intermediate steps across chains and do not provide a unified explanation for the predicted answer. We introduce Multi-Chain Reasoning (MCR), an approach which prompts large language models to meta-reason over multiple chains of thought, rather than aggregating their answers. MCR examines different reasoning chains, mixes information between them and selects the most relevant facts in generating an explanation and predicting the answer. MCR outperforms strong baselines on 7 multi-hop QA datasets. Moreover, our analysis reveals that MCR explanations exhibit high quality, enabling humans to verify its answers. | ['Jonathan Berant', 'Daniel Deutch', 'Uri Katz', 'Ben Bogin', 'Tomer Wolfson', 'Ori Yoran'] | 2023-04-25 | null | null | null | null | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 2.77694672e-01 7.88535058e-01 -3.51977289e-01 -5.77693820e-01
-1.74119806e+00 -8.93494308e-01 8.49738240e-01 6.93491578e-01
2.53132209e-02 7.40177393e-01 7.95256734e-01 -8.47510517e-01
-1.85990751e-01 -8.77184629e-01 -7.54671574e-01 5.10859266e-02
4.86052245e-01 1.09842753e+00 7.27058589e-01 -4.39875066e-01
3.45084071e-01 -1.67357311e-01 -1.27798295e+00 1.22423315e+00
1.20210457e+00 6.54184937e-01 -1.72606021e-01 9.47501421e-01
-7.99025178e-01 1.62403655e+00 -5.64601123e-01 -1.17901421e+00
-1.07374102e-01 -8.60832274e-01 -1.73963189e+00 -2.05948383e-01
6.21801138e-01 -2.14496464e-01 2.23343045e-01 8.35857511e-01
-1.17729537e-01 4.09846827e-02 5.70965767e-01 -1.14606738e+00
-8.02590489e-01 1.07137406e+00 -4.70741242e-02 1.58338800e-01
1.26169384e+00 1.10593708e-02 1.78592944e+00 -8.84683192e-01
6.27191007e-01 1.58776987e+00 5.35546422e-01 6.07934952e-01
-1.12271440e+00 -1.53943256e-01 2.61839956e-01 4.05856907e-01
-5.13168991e-01 -2.08029911e-01 3.85455936e-01 -1.85455009e-01
1.20232284e+00 5.81533194e-01 4.66967911e-01 6.79762721e-01
3.90364051e-01 9.54122603e-01 1.02100551e+00 -4.59696144e-01
3.51943076e-01 -1.44660100e-01 7.48431861e-01 8.15058231e-01
8.68056193e-02 -5.52320600e-01 -6.51915789e-01 -6.44197822e-01
6.74843937e-02 -1.40568361e-01 4.36707959e-02 2.81303078e-01
-1.20588112e+00 9.31534886e-01 5.11889935e-01 6.40234426e-02
-7.36761928e-01 -6.17373399e-02 -4.94771227e-02 5.94693780e-01
4.48488407e-02 8.24424148e-01 -5.91118634e-01 1.13838233e-01
-8.39506507e-01 8.39140356e-01 1.16701746e+00 7.82002985e-01
8.56247902e-01 -8.07219803e-01 -5.94287992e-01 4.13730562e-01
3.75911683e-01 2.63064802e-01 1.91965371e-01 -1.59725058e+00
8.37915123e-01 1.12488663e+00 4.44738865e-01 -7.74154723e-01
-1.81863412e-01 -1.76219329e-01 -2.38075048e-01 -2.82751590e-01
7.13654280e-01 -1.06918178e-01 -6.13611102e-01 1.52516770e+00
4.95950699e-01 -3.24454844e-01 3.86289507e-01 9.01574790e-01
8.97159696e-01 5.03243208e-01 3.15484285e-01 3.81735191e-02
1.55438948e+00 -1.45031297e+00 -5.45353174e-01 -6.01881564e-01
6.45475686e-01 -7.55359411e-01 1.14924991e+00 5.49880087e-01
-1.56188881e+00 -4.59465802e-01 -5.60855389e-01 -6.41695261e-01
-7.91475642e-03 -3.44255418e-01 5.76069117e-01 -4.17823829e-02
-1.09299004e+00 3.65553170e-01 -2.45810613e-01 -5.38272075e-02
-4.90245223e-02 1.82833225e-01 -3.70763429e-02 -4.40229744e-01
-1.46387661e+00 9.95944738e-01 9.27285776e-02 -2.72437900e-01
-6.06463134e-01 -6.30196989e-01 -6.53002262e-01 1.34420320e-01
6.15600169e-01 -1.28055656e+00 1.83845818e+00 -6.76558077e-01
-1.14187610e+00 5.93995631e-01 -9.74721014e-01 -6.65363312e-01
4.57963914e-01 -5.31187892e-01 -3.24116260e-01 2.20019758e-01
5.67935169e-01 7.92663634e-01 4.73179936e-01 -1.24256885e+00
-9.29681242e-01 -1.69277802e-01 8.69178236e-01 1.13866858e-01
5.47893882e-01 -6.83608577e-02 -3.18543941e-01 1.77360609e-01
5.20051420e-01 -6.30519271e-01 -4.17647958e-01 -5.45452535e-01
-5.76054394e-01 -8.24716508e-01 1.03061043e-01 -6.74382329e-01
1.25500321e+00 -1.34842193e+00 2.75495470e-01 -1.02616332e-01
5.11253715e-01 -1.73790425e-01 -2.40963981e-01 8.50511491e-01
2.61532009e-01 1.81588978e-01 -1.48128986e-01 -2.75843263e-01
3.50403905e-01 3.18974465e-01 -8.98130953e-01 -3.62039357e-01
2.39717945e-01 1.27369189e+00 -1.22050703e+00 -7.39165425e-01
-1.82795316e-01 -1.42972901e-01 -8.50725949e-01 3.55367720e-01
-1.21771860e+00 2.95498550e-01 -7.01491237e-01 5.62856853e-01
3.68908733e-01 -8.85986745e-01 2.31613204e-01 1.79123823e-02
2.69652665e-01 1.15475214e+00 -7.53854692e-01 1.45545983e+00
-5.16703486e-01 2.40388915e-01 -3.81121784e-01 -4.34337854e-01
7.37718701e-01 2.95905262e-01 4.98163551e-02 -6.46765292e-01
-2.74543643e-01 4.25358832e-01 -1.73909136e-03 -7.40544796e-01
6.11511171e-01 -1.90865219e-01 -1.78251922e-01 8.17318857e-01
-1.52316675e-01 -4.51819718e-01 4.48641896e-01 8.39347780e-01
1.22138584e+00 -4.95408755e-03 3.69178355e-01 1.39514297e-01
7.66326606e-01 7.13343799e-01 3.52497071e-01 1.13606381e+00
8.72837380e-02 5.13206840e-01 7.93181062e-01 -7.47484803e-01
-6.85571849e-01 -1.18192434e+00 5.33985794e-01 1.05787826e+00
3.39314640e-01 -8.08308303e-01 -6.89096689e-01 -1.27389896e+00
6.00280948e-02 1.30006576e+00 -4.42609131e-01 1.68634206e-01
-6.38635337e-01 -8.27170461e-02 5.07141292e-01 3.51076722e-01
3.37266684e-01 -1.04021430e+00 -6.04178011e-01 3.99941742e-01
-1.05837858e+00 -1.07632077e+00 -2.27198020e-01 -7.62658054e-03
-7.43286610e-01 -1.38277709e+00 -5.89324534e-02 -4.21186000e-01
5.79383016e-01 3.09795767e-01 1.77205396e+00 7.05677688e-01
5.13546526e-01 3.68158281e-01 -4.52195108e-01 -1.73838124e-01
-6.80984855e-01 6.37991354e-02 -4.35079247e-01 -2.52626508e-01
5.99662960e-01 -1.83914855e-01 -6.38992906e-01 1.62992597e-01
-6.47761822e-01 1.66626796e-01 5.00123680e-01 6.51111722e-01
5.92597604e-01 -2.81883121e-01 6.87607884e-01 -1.24618351e+00
9.77771282e-01 -7.60332584e-01 -9.54463184e-02 9.47993696e-01
-5.55501044e-01 5.36106229e-01 8.95799756e-01 -3.73233184e-02
-1.26614118e+00 -4.83526349e-01 -3.57883543e-01 3.56160283e-01
-1.64457723e-01 5.43591022e-01 1.99536696e-01 4.80709255e-01
8.03351104e-01 6.20098785e-02 -5.49362421e-01 -2.45152846e-01
7.91061819e-01 3.52949083e-01 5.69821119e-01 -8.15254033e-01
8.04267585e-01 2.45261267e-01 -2.15389401e-01 -1.08151641e-02
-1.77950287e+00 -3.59964222e-01 -3.91989589e-01 -1.58201560e-01
8.60477149e-01 -5.83435655e-01 -7.87664056e-01 -4.26000088e-01
-1.69647229e+00 -1.89702854e-01 -2.97868311e-01 1.95795849e-01
-4.45926815e-01 2.49493480e-01 -7.52202272e-01 -8.35731566e-01
-3.13136905e-01 -1.11105561e+00 1.10628307e+00 3.89421433e-01
-1.03376007e+00 -9.42284703e-01 -2.80453693e-02 1.20776248e+00
2.09435120e-01 -2.22210333e-01 1.50589824e+00 -1.00593901e+00
-8.95192981e-01 -6.66021854e-02 -8.44120309e-02 -2.36875057e-01
-1.75450221e-01 -1.67765826e-01 -6.92018330e-01 3.56727153e-01
-5.68026714e-02 -7.06722260e-01 7.13435233e-01 7.90818408e-02
7.29844987e-01 -5.86836100e-01 -1.43004000e-01 -2.13216811e-01
1.03251171e+00 -7.28956014e-02 4.91624892e-01 4.11130846e-01
1.94098368e-01 8.88198435e-01 6.79930925e-01 -2.34522566e-01
1.12087250e+00 1.95336238e-01 4.04651344e-01 2.16501608e-01
-1.34664506e-01 -6.59208238e-01 2.77527869e-01 7.81384408e-01
3.59088518e-02 -4.44366217e-01 -9.91148412e-01 6.82207048e-01
-1.87349892e+00 -1.01168835e+00 -7.39137769e-01 1.77314425e+00
8.15194011e-01 2.75033623e-01 -8.22248906e-02 -4.23549116e-02
3.04736406e-01 1.08947925e-01 -6.03229761e-01 -6.85524523e-01
-5.72987534e-02 2.25407735e-01 -2.27948457e-01 1.17509198e+00
-2.63646185e-01 1.06313491e+00 6.89233923e+00 4.61381465e-01
-3.36490989e-01 9.95819345e-02 4.88060087e-01 2.32482955e-01
-1.26294577e+00 6.49675488e-01 -7.87825942e-01 -5.71563318e-02
8.13911378e-01 -2.02271733e-02 4.54560727e-01 5.32542109e-01
-1.51217327e-01 -4.09537345e-01 -1.18132293e+00 3.83527130e-02
-5.54071665e-02 -1.64873648e+00 6.04281783e-01 -3.97520751e-01
8.18113625e-01 -2.16753349e-01 -3.38574082e-01 4.39807266e-01
9.86459672e-01 -1.10467219e+00 7.69009650e-01 6.59026384e-01
4.96930145e-02 -4.69252288e-01 7.19157279e-01 7.31737614e-01
-1.00675619e+00 -3.92761439e-01 -2.89656281e-01 -2.91249216e-01
5.06127656e-01 4.24500436e-01 -9.04811084e-01 9.42676067e-01
3.68870378e-01 3.86255197e-02 -6.90921664e-01 5.23429155e-01
-9.21803117e-01 7.76913404e-01 1.22541018e-01 -2.89524317e-01
3.63449335e-01 -3.96640338e-02 3.97874475e-01 8.56113136e-01
1.05875187e-01 4.59319651e-01 1.67197451e-01 1.00760901e+00
-1.41897440e-01 -1.13971382e-01 7.82755110e-03 6.00582771e-02
5.90882123e-01 9.22785819e-01 -3.83344918e-01 -7.53728271e-01
-6.10224783e-01 7.74963319e-01 5.80362260e-01 2.99699426e-01
-7.91621327e-01 -7.74167031e-02 3.37052286e-01 4.36335690e-02
-1.29901037e-01 5.54806925e-02 -4.22295392e-01 -1.22419083e+00
2.08746433e-01 -1.42898464e+00 9.48423922e-01 -1.19861901e+00
-1.29472733e+00 8.79396081e-01 -1.45031705e-01 -7.58264005e-01
-6.79535329e-01 -1.87188342e-01 -7.81947911e-01 9.50228751e-01
-1.60001576e+00 -1.09175217e+00 -1.22292310e-01 4.05971110e-01
9.30979669e-01 4.15120095e-01 8.39461625e-01 -2.59195089e-01
6.72144368e-02 1.28432408e-01 -8.00910592e-01 -2.81265736e-01
5.47410309e-01 -1.65247262e+00 6.45822704e-01 8.67189467e-01
5.40586889e-01 1.02073073e+00 1.00374496e+00 -8.10651720e-01
-1.27387142e+00 -6.94940567e-01 2.02410245e+00 -1.01201427e+00
7.52351105e-01 1.50895730e-01 -1.30472326e+00 8.06971908e-01
5.61139822e-01 -5.45725346e-01 5.44313490e-01 3.13905686e-01
-6.88394964e-01 1.79826036e-01 -9.49686468e-01 6.72096074e-01
8.03260982e-01 -7.21671939e-01 -1.49793398e+00 4.26658928e-01
1.18998289e+00 -3.66491705e-01 -6.62950099e-01 2.03034624e-01
4.67045993e-01 -1.42056751e+00 9.27270055e-01 -1.09960186e+00
9.66578662e-01 -4.27919984e-01 -2.17282459e-01 -1.04272532e+00
-2.31365576e-01 -6.79923236e-01 -3.84924769e-01 9.20611680e-01
1.09528482e+00 -4.36823487e-01 6.33139133e-01 9.72673774e-01
2.86803022e-03 -9.31601465e-01 -5.08706987e-01 -1.50525182e-01
9.21885222e-02 -5.39376557e-01 1.09933305e+00 6.54877424e-01
3.06340218e-01 6.60957396e-01 -3.05769350e-02 3.92978191e-01
6.62940264e-01 7.49879539e-01 7.42883503e-01 -1.11459565e+00
-4.51168448e-01 -1.87871799e-01 4.73417282e-01 -1.50136888e+00
1.34852245e-01 -8.52323294e-01 5.11426553e-02 -2.35952950e+00
1.46411940e-01 -1.54855713e-01 3.42727005e-02 4.51982558e-01
-7.23322093e-01 -1.52856871e-01 5.55476137e-02 3.05134892e-01
-1.18989146e+00 1.06230810e-01 1.42094743e+00 -1.27110749e-01
8.48572701e-02 1.10394970e-01 -1.08534455e+00 8.18185091e-01
5.59870481e-01 -5.25843382e-01 -5.90788603e-01 -7.23470032e-01
8.09429944e-01 7.80078769e-01 4.22855556e-01 -7.40632653e-01
7.03724265e-01 -4.27291393e-01 1.22654781e-01 -6.98273003e-01
1.92593411e-01 -3.82526577e-01 -1.60451501e-03 4.77959454e-01
-1.07209039e+00 3.71224523e-01 -4.26678091e-01 4.25596833e-01
-4.37671632e-01 -2.96973199e-01 3.09327394e-01 -3.39497983e-01
-3.15317422e-01 -2.72128843e-02 -3.56326818e-01 3.99909675e-01
5.10267198e-01 -6.38874667e-03 -6.98584914e-01 -6.91040695e-01
-7.57595778e-01 7.54198849e-01 1.71852767e-01 2.75438279e-01
4.56494093e-01 -1.07347369e+00 -8.08875084e-01 -3.08355212e-01
1.10560864e-01 1.57964885e-01 1.22665964e-01 7.14670658e-01
-2.42912889e-01 7.25190282e-01 3.91595274e-01 -3.44192415e-01
-1.00386679e+00 3.62875134e-01 3.39998990e-01 -8.38599086e-01
-5.30524909e-01 9.26343679e-01 -1.38872072e-01 -6.59560621e-01
-4.77849692e-02 -5.58377624e-01 -4.39620167e-01 6.25421926e-02
6.42922819e-01 2.43948400e-01 5.44885024e-02 -2.07513332e-01
-1.95379376e-01 4.87407058e-01 7.24403709e-02 -3.31432343e-01
1.00099993e+00 -3.95971805e-01 -4.77245092e-01 4.80861396e-01
6.02041900e-01 3.64079028e-01 -9.00843799e-01 -4.38635588e-01
5.45905650e-01 -3.07393312e-01 -6.93254530e-01 -1.13287997e+00
-2.24433452e-01 6.96196198e-01 -7.03759968e-01 7.36407936e-01
8.20677638e-01 4.98154253e-01 1.17915988e+00 6.34105206e-01
4.64924395e-01 -4.89937127e-01 3.15029502e-01 7.24561572e-01
9.88531947e-01 -1.12911046e+00 -3.35195482e-01 -4.55356777e-01
-9.23266232e-01 1.13192379e+00 7.91792452e-01 1.76859960e-01
-1.47688702e-01 4.11351100e-02 4.63489443e-01 -5.41474640e-01
-1.57494509e+00 -1.53834447e-01 4.56181586e-01 1.74985915e-01
4.97418463e-01 -3.49555202e-02 -2.84814805e-01 8.10682476e-01
-6.36864901e-01 -1.07651278e-01 4.08162296e-01 6.27549469e-01
-7.59001732e-01 -1.28027880e+00 -4.73184586e-01 4.69889104e-01
-5.15192568e-01 -1.77856714e-01 -9.55389023e-01 5.00319362e-01
8.52554962e-02 1.73163652e+00 -3.61987561e-01 -3.53782684e-01
3.10456306e-01 5.20632684e-01 2.68532783e-01 -6.03492916e-01
-1.00535750e+00 -6.00523591e-01 5.27695656e-01 -7.12628722e-01
-3.59671861e-01 -4.72103268e-01 -1.53425181e+00 -3.13802987e-01
-1.94575697e-01 9.65461552e-01 9.54842493e-02 1.61589241e+00
3.60254049e-01 4.80484873e-01 2.50862807e-01 1.41182899e-01
-6.06234908e-01 -5.98762155e-01 1.94159284e-01 5.32433212e-01
5.48082829e-01 -1.15743056e-01 -4.01465446e-01 -3.68085643e-03] | [11.024347305297852, 7.8939080238342285] |
55628c30-7560-42a0-b316-cf78ce340d9b | aspect-category-detection-via-topic-attention | 1901.01183 | null | https://arxiv.org/abs/1901.01183v2 | https://arxiv.org/pdf/1901.01183v2.pdf | Aspect Category Detection via Topic-Attention Network | The e-commerce has started a new trend in natural language processing through sentiment analysis of user-generated reviews. Different consumers have different concerns about various aspects of a specific product or service. Aspect category detection, as a subtask of aspect-based sentiment analysis, tackles the problem of categorizing a given review sentence into a set of pre-defined aspect categories. In recent years, deep learning approaches have brought revolutionary advances in multiple branches of natural language processing including sentiment analysis. In this paper, we propose a deep neural network method based on attention mechanism to identify different aspect categories of a given review sentence. Our model utilizes several attentions with different topic contexts, enabling it to attend to different parts of a review sentence based on different topics. Experimental results on two datasets in the restaurant domain released by SemEval workshop demonstrates that our approach outperforms existing methods on both datasets. Visualization of the topic attention weights shows the effectiveness of our model in identifying words related to different topics. | ['Sajad Movahedi', 'Azadeh Shakery', 'Erfan Ghadery', 'Heshaam Faili'] | 2019-01-04 | null | null | null | null | ['aspect-category-detection'] | ['natural-language-processing'] | [ 1.84592027e-02 5.91016039e-02 -2.24114969e-01 -7.92210460e-01
-5.37844777e-01 -4.81779248e-01 5.91136992e-01 6.89229190e-01
-3.11580777e-01 7.78477490e-02 5.86371243e-01 -3.26603025e-01
2.37376243e-01 -9.53837395e-01 -2.91616976e-01 -3.74967098e-01
4.00210261e-01 3.95914018e-01 -1.10042937e-01 -5.21957695e-01
7.70006895e-01 -1.02271967e-01 -1.32574987e+00 6.29394054e-01
4.41159040e-01 1.20800269e+00 1.87053278e-01 4.97944385e-01
-7.70696640e-01 4.59949225e-01 -7.93399096e-01 -6.25456631e-01
-4.10219394e-02 -3.16545725e-01 -7.85117269e-01 2.97570676e-01
2.12962553e-01 1.54028833e-01 5.48099935e-01 1.24132419e+00
2.24922121e-01 9.54788700e-02 6.78942263e-01 -9.46888566e-01
-1.08727014e+00 8.58843386e-01 -9.19042647e-01 4.30155665e-01
2.18836308e-01 -3.37771893e-01 1.70494330e+00 -9.31184649e-01
2.44740143e-01 1.32709694e+00 5.44384122e-01 1.92989275e-01
-8.05948734e-01 -3.18120718e-01 9.00368690e-01 2.17144921e-01
-7.32854068e-01 1.96931362e-01 9.68340814e-01 -3.37603033e-01
1.29738224e+00 7.03974208e-03 7.76449800e-01 5.57506382e-01
7.58599877e-01 1.00869524e+00 7.64073372e-01 -2.63104320e-01
4.01556998e-01 5.69598198e-01 1.01293612e+00 1.57756671e-01
3.01247895e-01 -8.00686002e-01 -3.48231077e-01 -9.47972909e-02
-1.28109634e-01 2.68663913e-01 -1.08651370e-02 -1.93953484e-01
-9.31999922e-01 1.38039339e+00 2.24514604e-01 3.18973243e-01
-6.21088803e-01 -1.75488219e-01 8.43780518e-01 3.02059352e-01
1.03186405e+00 6.12773418e-01 -1.03316784e+00 7.87875056e-02
-3.55286479e-01 2.45399594e-01 1.00655556e+00 9.14411008e-01
7.07567453e-01 -2.40920037e-01 -1.58724383e-01 8.75525653e-01
6.59216166e-01 2.88781524e-01 8.46626639e-01 -1.16773188e-01
3.73806894e-01 1.21572661e+00 -1.16341718e-01 -1.32142127e+00
-4.46168303e-01 -3.90919626e-01 -5.81161201e-01 -4.13683318e-02
-2.54678905e-01 -2.52825826e-01 -1.05917978e+00 1.20930254e+00
3.71840924e-01 -4.48785096e-01 9.60416198e-02 8.55941832e-01
1.17897189e+00 8.25016856e-01 3.61254990e-01 1.17797084e-01
1.98883498e+00 -1.09855711e+00 -8.77020240e-01 -3.81939888e-01
2.03608304e-01 -8.20720315e-01 1.18435454e+00 5.53477049e-01
-7.20965624e-01 -6.01308048e-01 -1.13795722e+00 -2.31550202e-01
-1.01676834e+00 -6.12124763e-02 8.67307723e-01 6.78898871e-01
-8.37271810e-01 1.28965378e-01 -4.50763404e-01 -6.00641787e-01
3.78132552e-01 2.50531912e-01 -4.85135093e-02 1.73725471e-01
-1.21515048e+00 6.61204875e-01 -5.37800752e-02 1.14959858e-01
-5.96055746e-01 -5.26562452e-01 -1.13836718e+00 4.46387082e-01
2.41668850e-01 -5.69354951e-01 1.47903907e+00 -1.55258060e+00
-1.28897583e+00 9.53049541e-01 -3.01020145e-01 -4.04939443e-01
-4.79240805e-01 -5.59775770e-01 -7.39094734e-01 9.10064429e-02
3.47817779e-01 3.97476196e-01 8.82386565e-01 -9.87904727e-01
-9.49854672e-01 -5.03471553e-01 4.05498385e-01 3.52054089e-01
-3.41413260e-01 5.02521455e-01 -3.21362525e-01 -5.53367317e-01
-7.47811496e-02 -5.37814260e-01 -5.21660805e-01 -5.24758101e-01
-4.43936855e-01 -6.65314972e-01 7.87756264e-01 -2.44755328e-01
1.01179206e+00 -1.85927808e+00 -1.54618010e-01 -5.29240258e-02
2.51884431e-01 1.21504746e-01 -1.99354351e-01 3.49964619e-01
-1.79176062e-01 5.19301414e-01 -2.09396318e-01 -3.78162265e-01
8.12123492e-02 -2.08656847e-01 -5.69025815e-01 1.60172090e-01
4.96796340e-01 6.89114511e-01 -9.31148052e-01 -4.17363234e-02
-1.45878181e-01 4.86843765e-01 -3.48618478e-01 8.53741020e-02
-3.26377720e-01 -6.33114055e-02 -6.08545244e-01 6.03239834e-01
5.96107781e-01 -1.24558710e-01 -1.69782154e-02 -1.89546674e-01
2.51100063e-02 8.49434674e-01 -6.00462079e-01 1.45777106e+00
-4.96065736e-01 7.67884374e-01 -1.61513284e-01 -1.18349338e+00
1.05254245e+00 3.38934094e-01 3.14552546e-01 -5.77364445e-01
5.13604164e-01 -1.70953900e-01 9.66947526e-02 -4.15183634e-01
1.12082124e+00 -4.14877832e-01 -3.60069960e-01 8.27120721e-01
2.39301592e-01 -2.54205912e-01 4.10254270e-01 2.02693060e-01
6.42329454e-01 -1.90583870e-01 6.95848107e-01 -6.52035415e-01
5.70456207e-01 1.21280663e-01 4.23541754e-01 6.01781666e-01
-4.24103677e-01 7.65081227e-01 7.45058298e-01 -9.10383224e-01
-7.67542183e-01 -5.31433403e-01 -1.09054297e-02 1.32297182e+00
1.52009070e-01 -5.48002005e-01 -4.94612247e-01 -1.01819336e+00
-1.83755890e-01 6.92092538e-01 -1.08958161e+00 -1.68833986e-01
-2.86963940e-01 -9.88592267e-01 -3.43798786e-01 4.10910904e-01
2.92677432e-01 -1.53168249e+00 -6.96122766e-01 2.41087601e-01
5.34528568e-02 -8.38622570e-01 -5.53318679e-01 3.53999525e-01
-5.93668401e-01 -1.05707383e+00 -6.20520651e-01 -1.04766035e+00
6.33812428e-01 6.00920022e-01 1.75103951e+00 -1.14044651e-01
7.38205574e-03 4.17082906e-01 -7.64511645e-01 -1.15206790e+00
-8.17744061e-02 4.68853623e-01 -2.43237019e-01 3.28294843e-01
1.52108812e+00 -1.74362689e-01 -6.52985334e-01 -1.04049686e-02
-9.14540291e-01 -3.96457642e-01 4.14465815e-01 4.64215428e-01
5.01958013e-01 2.76329834e-02 7.02334642e-01 -1.20390487e+00
1.18247867e+00 -9.46194172e-01 -2.44888723e-01 -1.65704098e-02
-7.48287320e-01 -9.83441249e-02 5.23930848e-01 -2.72221297e-01
-9.87695098e-01 -3.63113016e-01 -1.08443283e-01 4.02445257e-01
-2.24073619e-01 1.02086759e+00 -1.94421783e-01 7.48225212e-01
2.48425230e-01 9.30455253e-02 -2.19147712e-01 -3.48229796e-01
2.85619944e-01 7.95280516e-01 -2.36533076e-01 -1.04256004e-01
1.60818860e-01 4.13511038e-01 -6.00964069e-01 -6.29126012e-01
-1.25755191e+00 -8.96691203e-01 -3.92825574e-01 -5.34168072e-02
1.03168833e+00 -8.08434248e-01 -5.92699587e-01 3.60609621e-01
-1.25026393e+00 3.30605328e-01 -3.16843182e-01 2.50895590e-01
-1.23794388e-03 1.92123994e-01 -5.13311088e-01 -6.01873815e-01
-8.46168339e-01 -1.22142386e+00 1.37281132e+00 5.02078950e-01
-5.32448173e-01 -1.10099435e+00 3.27681333e-01 1.16157666e-01
5.66241324e-01 -1.27703056e-01 1.01420891e+00 -1.21140444e+00
-2.44225822e-02 -4.12604123e-01 -3.65957841e-02 4.95399952e-01
4.32916015e-01 -4.51554582e-02 -9.21151996e-01 5.04166558e-02
4.34618950e-01 -3.58260088e-02 1.14595544e+00 5.23461998e-01
8.40730369e-01 -2.39197910e-01 -1.22779772e-01 1.44535467e-01
1.38388062e+00 5.21955073e-01 4.66794491e-01 7.33944833e-01
5.05633712e-01 7.67467797e-01 7.74051547e-01 2.77865767e-01
5.82893848e-01 1.97904870e-01 3.36367488e-01 -1.51664466e-01
4.16211069e-01 2.62112528e-01 3.73623371e-01 9.59831595e-01
3.89621317e-01 -3.90226334e-01 -7.11455703e-01 9.65775073e-01
-1.61917889e+00 -7.87524819e-01 -6.35004416e-02 1.70586157e+00
5.87187290e-01 2.36618057e-01 -3.24830562e-02 -1.72331240e-02
7.90940940e-01 4.53930646e-01 -6.18789077e-01 -1.20779014e+00
1.27194256e-01 1.95071101e-01 -1.93448868e-02 2.68887907e-01
-1.43714333e+00 8.10441017e-01 6.04239655e+00 3.43828142e-01
-1.14568114e+00 -4.77330796e-02 1.05697525e+00 1.42532140e-01
-6.34419739e-01 -3.38575959e-01 -1.00942326e+00 3.14902157e-01
8.78206372e-01 -3.40154499e-01 -4.12790060e-01 1.33979166e+00
1.54777184e-01 -2.57093012e-01 -9.05240536e-01 4.20849591e-01
6.21108174e-01 -1.05708218e+00 3.89984816e-01 -1.89660981e-01
7.38650024e-01 7.74369612e-02 1.88334122e-01 6.02477193e-01
3.20338964e-01 -7.52729774e-01 5.07924855e-01 4.78421524e-02
-8.72706622e-03 -1.08260989e+00 1.23650396e+00 -1.77250966e-01
-1.20700622e+00 1.52986035e-01 -6.20478451e-01 -2.83490121e-01
2.10732281e-01 6.70038223e-01 -7.54106164e-01 2.47437462e-01
1.02286255e+00 1.13329065e+00 -3.81597668e-01 6.91235185e-01
-2.83182591e-01 6.77716494e-01 2.20383108e-01 -4.61296916e-01
4.43935126e-01 -4.13525820e-01 4.84007388e-01 1.13553452e+00
1.18518118e-02 -1.47623599e-01 -2.91333608e-02 6.68976128e-01
-2.70606410e-02 4.83062595e-01 -5.54053247e-01 -1.84669450e-01
-3.27944219e-01 1.69893527e+00 -1.08932388e+00 -5.12599885e-01
-6.33353174e-01 6.78019345e-01 7.19959810e-02 2.65951365e-01
-5.70944488e-01 -7.39969015e-01 8.41618001e-01 -2.19484881e-01
7.05190778e-01 9.52981114e-02 -6.00516200e-01 -1.07145011e+00
-4.80276085e-02 -8.93597782e-01 3.60978365e-01 -6.98150039e-01
-1.37404609e+00 1.14483654e+00 -3.57494622e-01 -1.17895520e+00
1.13407925e-01 -7.62791574e-01 -9.94545817e-01 9.04282212e-01
-1.78794324e+00 -8.92370522e-01 -3.07406625e-03 8.48700181e-02
1.31266928e+00 -3.86647761e-01 8.66895139e-01 9.35717970e-02
-3.36420953e-01 2.26207033e-01 -2.23876566e-01 9.33212340e-02
5.17805338e-01 -1.43213296e+00 8.10151398e-01 7.16420174e-01
1.79264024e-01 9.97797906e-01 7.21038222e-01 -4.77116287e-01
-9.11603510e-01 -8.93015087e-01 1.31622612e+00 -4.75682110e-01
7.23297715e-01 -2.99390763e-01 -7.86292017e-01 6.63385868e-01
9.62028205e-01 -6.63407683e-01 1.31400526e+00 6.58719480e-01
-3.91199291e-01 -1.74861655e-01 -9.98005748e-01 4.78275120e-01
1.40867949e-01 -4.16909277e-01 -9.77650940e-01 2.00499788e-01
1.07104921e+00 1.30551100e-01 -4.27186400e-01 1.57838807e-01
4.76137519e-01 -7.47422636e-01 4.74477291e-01 -1.02265394e+00
9.82969403e-01 -3.11358362e-01 -3.66539545e-02 -1.68173754e+00
-3.66790086e-01 -2.07696274e-01 1.42348334e-01 1.35431337e+00
6.93366468e-01 -4.75687593e-01 6.43394113e-01 6.54350817e-01
-1.92211568e-01 -8.70354712e-01 -3.12575519e-01 1.30880237e-01
6.44706115e-02 -5.05820453e-01 7.89069414e-01 7.57579744e-01
2.65428782e-01 1.02758563e+00 -1.83467586e-02 -3.63725312e-02
-2.80055571e-02 5.25845170e-01 4.83278692e-01 -1.16291797e+00
-1.25349253e-01 -6.97193325e-01 -3.62795204e-01 -8.79147053e-01
7.10273981e-02 -5.79619765e-01 1.79958031e-01 -1.85169017e+00
4.34942931e-01 1.34460509e-01 -6.81537271e-01 2.65010655e-01
-5.94180703e-01 2.06161812e-01 3.10070012e-02 -2.83936620e-01
-9.33486938e-01 5.04879594e-01 9.78399515e-01 -6.77930117e-01
-4.89168689e-02 4.09218907e-01 -1.54062831e+00 8.48959267e-01
1.06620800e+00 -5.38046181e-01 -4.26477462e-01 -4.89512473e-01
9.02627110e-01 -7.00991690e-01 -4.96995360e-01 -5.41085541e-01
-7.60986954e-02 1.57286599e-01 9.22292620e-02 -8.96835089e-01
-1.84822753e-02 -9.41148221e-01 -6.61816061e-01 1.63796186e-01
-6.16406024e-01 4.52929616e-01 2.92201728e-01 5.92126548e-01
-5.99240601e-01 -2.81533450e-01 4.13834423e-01 -3.15264195e-01
-8.68890166e-01 1.68195903e-01 -8.76610756e-01 -6.59463406e-02
7.35731781e-01 9.23300236e-02 -1.11477412e-01 -4.87297356e-01
-5.34006059e-01 2.15364084e-01 -5.29208407e-02 9.01248872e-01
5.70562661e-01 -1.04914689e+00 -5.53967118e-01 1.71952039e-01
4.46141332e-01 -2.46338379e-02 8.02935213e-02 4.65138406e-01
-2.31045112e-01 5.53444505e-01 -4.70783608e-03 -4.13382590e-01
-1.30732465e+00 7.26753175e-01 1.30986258e-01 -5.96817732e-01
-3.10397476e-01 8.85324955e-01 5.05322456e-01 -5.53012133e-01
5.93795441e-02 -8.39196801e-01 -1.16468823e+00 4.43559617e-01
9.84535456e-01 -1.61419064e-01 2.47256488e-01 -7.37660885e-01
-3.51668924e-01 6.96489513e-01 -6.81629241e-01 7.68826902e-02
1.42343724e+00 -2.51955181e-01 -3.03588212e-01 7.28364646e-01
1.31033659e+00 -1.32186199e-02 -6.99443340e-01 -1.48716450e-01
2.44394034e-01 1.97102083e-03 6.04445636e-02 -8.98379862e-01
-1.38391638e+00 8.69528115e-01 3.23576033e-01 8.51523340e-01
1.05878043e+00 -5.46625406e-02 8.15803826e-01 4.27948833e-01
-1.70789586e-04 -1.19511580e+00 -1.76067144e-01 8.48390758e-01
9.28152680e-01 -1.67777658e+00 1.19841613e-01 -7.36897662e-02
-1.05095387e+00 1.03168905e+00 5.59952617e-01 -4.06360596e-01
9.99980450e-01 -3.37202847e-02 5.36558807e-01 -6.92158580e-01
-7.81566739e-01 -2.36710504e-01 4.54508066e-01 4.66707289e-01
8.35573435e-01 -3.16461474e-02 -6.57016814e-01 1.06336784e+00
-1.11222565e-01 -5.43346584e-01 6.56590760e-01 8.96920383e-01
-5.82889199e-01 -1.02408719e+00 -3.78343984e-02 6.13422453e-01
-1.13319004e+00 -4.92604762e-01 -5.76612055e-01 4.88189876e-01
-1.61557376e-01 1.23909318e+00 2.97237039e-01 -1.16779558e-01
2.95470357e-01 -1.59542207e-02 -3.56129795e-01 -1.10907686e+00
-1.18610990e+00 1.10208578e-01 4.19225097e-02 -2.53780127e-01
-7.24098742e-01 -5.43627083e-01 -1.07993972e+00 1.36232108e-01
-1.90728858e-01 5.59223235e-01 9.62662041e-01 1.00296044e+00
7.00711966e-01 6.66394114e-01 6.57767534e-01 -6.71091020e-01
-3.12971435e-02 -1.20245135e+00 -6.72179878e-01 5.52019596e-01
4.02718365e-01 -3.58360440e-01 -4.09261376e-01 5.72055094e-02] | [11.380671501159668, 6.675124168395996] |
aa7e88bb-066b-4988-b132-c887ac08dc95 | data-driven-computing-with-noisy-material | null | null | https://www.sciencedirect.com/science/article/pii/S0045782517304012 | https://www.sciencedirect.com/science/article/pii/S0045782517304012 | Data Driven Computing with Noisy Material Data Sets | We formulate a Data Driven Computing paradigm, termed max-ent Data Driven Computing, that generalizes distance-minimizing Data Driven Computing and is robust with respect to outliers. Robustness is achieved by means of clustering analysis. Specifically, we assign data points a variable relevance depending on distance to the solution and on maximum-entropy estimation. The resulting scheme consists of the minimization of a suitably-defined free energy over phase space subject to compatibility and equilibrium constraints. Distance-minimizing Data Driven schemes are recovered in the limit of zero temperature. We present selected numerical tests that establish the convergence properties of the max-ent Data Driven solvers and solutions. | ['T.Kirchdoerfer', 'M.Ortiz'] | 2017-11-01 | null | null | null | computer-methods-in-applied-mechanics-and | ['stress-strain-relation'] | ['miscellaneous'] | [ 8.63101110e-02 -1.07821606e-01 2.21602187e-01 -4.68443245e-01
-8.43457341e-01 -4.91673082e-01 7.05878079e-01 5.16765237e-01
-3.86873007e-01 5.92604816e-01 -1.04822978e-01 1.69406906e-01
-1.00867939e+00 -6.60125256e-01 -5.18551767e-01 -1.13478887e+00
-2.34225988e-01 9.26223636e-01 -1.71194136e-01 -3.42389613e-01
6.78464293e-01 7.13067770e-01 -1.46271443e+00 -3.74686420e-01
1.31268167e+00 9.36389804e-01 -1.72005057e-01 6.09570682e-01
5.39894342e-01 3.39072138e-01 -1.02823995e-01 2.73688674e-01
4.06943351e-01 -4.13528025e-01 -6.95336699e-01 1.20728716e-01
-4.50928546e-02 5.01246333e-01 2.04087511e-01 1.25030637e+00
4.50099587e-01 6.06012642e-01 9.33960855e-01 -1.37123621e+00
-3.77797842e-01 -1.02545299e-01 -2.67598510e-01 8.14627334e-02
4.53642577e-01 1.43006146e-01 7.54306078e-01 -1.32588220e+00
8.29561353e-01 8.34799170e-01 5.27382374e-01 3.45720828e-01
-1.68417645e+00 1.18279532e-01 -4.28443104e-02 6.92554712e-02
-1.60605037e+00 -3.71594071e-01 9.14550066e-01 -9.25500691e-01
7.84626782e-01 5.04907131e-01 4.80359495e-01 4.30172145e-01
5.13501048e-01 1.82709455e-01 1.09079242e+00 -3.52661788e-01
1.18007267e+00 -1.61532909e-01 4.20705646e-01 3.30036491e-01
7.09602118e-01 3.78707767e-01 -6.93639457e-01 -3.77906710e-01
1.65257268e-02 -2.59192675e-01 -1.95095211e-01 -1.08316040e+00
-1.00406492e+00 8.22472990e-01 3.33626986e-01 6.07525222e-02
-5.05721629e-01 -1.45944253e-01 2.70937771e-01 -2.81582754e-02
8.40110302e-01 9.26624596e-01 -3.82624954e-01 -1.13872208e-01
-1.12954319e+00 6.81892872e-01 6.98384464e-01 1.03580594e+00
8.40396225e-01 3.19505557e-02 2.82680631e-01 5.33012999e-03
4.43628073e-01 3.95591974e-01 3.92495036e-01 -9.46723819e-01
1.48250431e-01 8.47002506e-01 1.46472409e-01 -9.18108881e-01
-5.60024738e-01 -4.72477265e-02 -1.07257235e+00 2.55250603e-01
3.01671270e-02 -3.40413451e-01 -8.27756226e-01 1.49241054e+00
5.05148053e-01 4.76198196e-02 1.09743178e-01 1.25930345e+00
2.59169757e-01 4.61568087e-01 -2.01139078e-01 -8.45210016e-01
6.73080981e-01 -2.29638442e-01 -7.60336161e-01 2.15128630e-01
6.36249423e-01 -3.61306638e-01 6.54704988e-01 7.61011302e-01
-1.11754453e+00 -3.99749875e-01 -1.22277570e+00 1.78388014e-01
-3.33067596e-01 -7.19788671e-01 6.02240488e-03 3.93749475e-01
-9.51024592e-01 9.38876808e-01 -1.14610648e+00 -1.93352923e-01
-1.18183903e-01 6.18715942e-01 -2.64509588e-01 4.07147676e-01
-8.27388823e-01 8.91475499e-01 6.27204180e-01 3.48421663e-01
-6.92955792e-01 -7.68607199e-01 -8.59304905e-01 -2.82556921e-01
2.91455328e-01 -5.86196959e-01 6.11681521e-01 -7.08158731e-01
-1.09239376e+00 8.20420384e-01 -2.14074507e-01 -4.52939451e-01
6.75137460e-01 -2.19265744e-02 -3.73015583e-01 6.00417107e-02
-9.38352123e-02 -1.82387739e-01 8.35812986e-01 -1.24224854e+00
1.44397393e-01 -5.04739285e-01 -7.40051448e-01 1.87210083e-01
-9.87704471e-02 -2.07209453e-01 3.41322199e-02 -4.49452490e-01
6.09196305e-01 -1.07153928e+00 -8.57595682e-01 -1.46416381e-01
-6.64612234e-01 7.98960552e-02 7.63622820e-01 -2.65165001e-01
1.12776518e+00 -2.04732156e+00 8.06973040e-01 9.55715597e-01
3.47669452e-01 -3.10102433e-01 2.31418580e-01 6.78724885e-01
-3.87620360e-01 1.54283866e-01 -9.37256098e-01 -5.16055077e-02
2.40535624e-02 4.68478166e-02 3.66903320e-02 1.02295208e+00
3.53164077e-01 4.21125263e-01 -7.63360679e-01 -5.47860742e-01
4.27662909e-01 2.30618015e-01 -8.33814740e-01 2.93249249e-01
-1.53843179e-01 6.02149248e-01 -5.37956238e-01 3.82844865e-01
7.96399772e-01 4.07799445e-02 1.32970303e-01 5.83064146e-02
-3.89009356e-01 3.26249339e-02 -1.60956252e+00 1.71792269e+00
1.16664514e-01 2.15892196e-01 2.73296565e-01 -1.17773950e+00
1.24240673e+00 9.25692916e-02 1.16891551e+00 -4.35499698e-01
1.79754496e-01 2.36945793e-01 -2.74288449e-02 -3.20094049e-01
7.83871949e-01 -6.85237115e-03 -3.52119267e-01 3.12800378e-01
-3.06728005e-01 -6.33442879e-01 2.35777080e-01 3.30890179e-01
8.79079998e-01 1.62515879e-01 2.33802855e-01 -1.20269048e+00
6.17778659e-01 4.30882156e-01 9.45421755e-01 1.70708060e-01
-2.65141368e-01 8.75293016e-01 4.64913785e-01 -3.40175837e-01
-1.06397879e+00 -8.66550148e-01 -5.40371180e-01 6.29233718e-01
3.81413102e-01 -5.87668657e-01 -9.63918388e-01 -9.31416005e-02
2.15035111e-01 7.27360070e-01 -7.64788806e-01 -5.96609414e-01
-6.69637740e-01 -1.25196958e+00 -4.41298932e-01 1.13747515e-01
-1.00332655e-01 -7.51751840e-01 -5.54505706e-01 1.46913141e-01
2.37240136e-01 -4.22065347e-01 -3.11879605e-01 5.90513945e-01
-1.04966879e+00 -1.27204394e+00 -2.10787565e-01 -4.92472142e-01
9.24564242e-01 -4.71181244e-01 1.05945361e+00 -1.29064247e-01
-4.32607979e-01 1.63829640e-01 -8.00163522e-02 -3.52510959e-02
-3.12736601e-01 9.95720737e-03 7.63007343e-01 2.02360541e-01
3.41549397e-01 -3.33066732e-01 -5.72269619e-01 1.86647147e-01
-7.78189301e-01 -3.44516516e-01 -9.29024369e-02 6.52407765e-01
1.00545299e+00 1.74476221e-01 1.98418111e-01 -6.52470648e-01
6.28827929e-01 -7.10574985e-01 -9.05726850e-01 -9.99644920e-02
-1.08563077e+00 5.34581423e-01 5.54162025e-01 -1.53691396e-01
-6.20611310e-01 5.31316042e-01 4.21539158e-01 -3.19264114e-01
2.01492339e-01 4.19365585e-01 -3.01899642e-01 -1.61197096e-01
6.91075146e-01 7.33155161e-02 8.65647122e-02 -5.65762579e-01
4.09508020e-01 5.47850966e-01 5.51194310e-01 -1.01118886e+00
8.76252055e-01 2.96361864e-01 3.47717106e-01 -6.99814916e-01
-1.56571254e-01 -3.60415310e-01 -9.81084526e-01 -3.41189742e-01
7.66718090e-01 -4.01991218e-01 -9.87680733e-01 3.89609754e-01
-5.84042609e-01 1.16484486e-01 -6.50954366e-01 3.74776602e-01
-9.52777922e-01 1.45946443e-01 -2.05974191e-01 -8.68192017e-01
-3.46539706e-01 -1.19660914e+00 7.05271423e-01 9.89132375e-02
-3.32110703e-01 -9.39072907e-01 5.35684705e-01 -1.72389463e-01
1.76312760e-01 8.58435869e-01 7.98915386e-01 -7.73572505e-01
-3.17617297e-01 -1.30621105e-01 4.01878238e-01 -1.24288043e-02
-2.06548959e-01 6.99737668e-01 -6.05746448e-01 -6.30151927e-01
2.88650960e-01 1.59901008e-01 5.40190279e-01 5.39153337e-01
1.01066637e+00 -2.76339084e-01 -3.55243325e-01 9.03780937e-01
1.78292048e+00 -1.17061101e-01 6.21341132e-02 3.01491857e-01
4.70581174e-01 8.14734578e-01 8.40937316e-01 6.77104652e-01
2.17412338e-02 5.79470277e-01 6.98902071e-01 8.79993886e-02
6.42587721e-01 6.94118068e-02 -1.52084708e-01 1.10552156e+00
-1.23311475e-01 2.92473555e-01 -1.21485245e+00 6.99687541e-01
-2.08538175e+00 -6.66419148e-01 -7.59486198e-01 2.38874459e+00
7.97199786e-01 4.97253090e-02 2.91498274e-01 4.13247615e-01
8.26854765e-01 -1.90443248e-01 -9.17653263e-01 -6.74891651e-01
7.82197621e-03 1.71706319e-01 5.27387738e-01 6.44062102e-01
-9.92781579e-01 2.47359499e-01 7.10515547e+00 2.49805018e-01
-8.46368492e-01 -1.80772230e-01 4.37766254e-01 -3.47803831e-01
-7.67052323e-02 3.01011670e-02 -3.84263545e-01 8.76783192e-01
1.04681838e+00 -8.16725254e-01 2.21598059e-01 1.04952383e+00
4.46253568e-01 -2.47237325e-01 -1.23577487e+00 6.99954093e-01
-2.42901161e-01 -1.09108794e+00 -4.88853097e-01 2.78851211e-01
1.05661011e+00 -1.96836993e-01 5.79497255e-02 -4.01086122e-01
2.46531114e-01 -8.95630717e-01 7.59601295e-01 8.46155047e-01
7.31842756e-01 -1.11491847e+00 3.50355893e-01 3.62779468e-01
-1.03779042e+00 7.76384920e-02 -1.12160787e-01 -2.03166544e-01
4.39556874e-02 8.97999167e-01 -3.11377704e-01 5.30656934e-01
5.67372203e-01 9.47747231e-01 -3.50210816e-01 1.06354022e+00
2.50245869e-01 5.05000819e-03 -4.16334301e-01 3.16360965e-02
-1.47734657e-01 -9.55066264e-01 7.82575786e-01 8.74448240e-01
2.11141855e-01 2.29962170e-01 2.43433669e-01 1.02400601e+00
3.47527266e-01 5.82041070e-02 -5.55605590e-01 5.60308039e-01
5.88968754e-01 1.16954291e+00 -5.96346498e-01 -1.89718068e-01
2.86407411e-01 6.48821056e-01 1.79812029e-01 3.00315678e-01
-2.73334384e-01 -4.43346351e-01 9.40044999e-01 2.61168778e-01
9.78512391e-02 -7.06029415e-01 -7.85475254e-01 -1.08431065e+00
4.44757827e-02 -6.64500177e-01 4.34222937e-01 -4.73789275e-01
-1.29609561e+00 4.88063097e-01 1.47497997e-01 -1.35219860e+00
-4.73584384e-01 -6.79883361e-01 -4.29123491e-01 8.00042033e-01
-1.07897353e+00 -2.60731012e-01 -2.05020487e-01 7.57232010e-01
1.57224331e-02 1.30792230e-01 7.70142794e-01 -1.27811022e-02
-1.04640472e+00 5.09259887e-02 6.09402657e-01 -2.04522699e-01
5.49498022e-01 -1.61607003e+00 1.82448760e-01 1.12161326e+00
-5.55321455e-01 7.35498905e-01 1.11901534e+00 -6.54238880e-01
-1.85098851e+00 -1.10206163e+00 8.46819460e-01 -5.59620321e-01
4.67469156e-01 -5.07559299e-01 -9.99203980e-01 5.38233817e-01
-2.05079705e-01 3.62055749e-01 2.32416615e-01 -1.28358454e-01
1.06531091e-01 -1.94963247e-01 -1.65358615e+00 4.56105918e-02
5.40733099e-01 -4.05724466e-01 -6.08482242e-01 5.48416197e-01
2.78256506e-01 -3.08522612e-01 -1.19038534e+00 5.23398638e-01
-1.94640428e-01 -8.23299229e-01 8.54226112e-01 -6.55474663e-01
7.52694830e-02 -5.25687218e-01 -1.11537799e-02 -1.23919141e+00
-2.01530814e-01 -1.11048532e+00 -3.17997515e-01 8.04396749e-01
1.25771567e-01 -5.14323115e-01 8.34868073e-01 1.37045050e+00
-5.39752804e-02 -6.27268612e-01 -1.18851876e+00 -8.18999290e-01
3.07553768e-01 -3.98614913e-01 4.73644704e-01 1.11128986e+00
4.43256795e-01 -1.30633101e-01 4.13264334e-02 2.84560800e-01
1.07290888e+00 1.36325330e-01 2.49870628e-01 -1.54353702e+00
1.43814504e-01 -3.06946874e-01 -4.20714915e-01 -2.51113415e-01
-1.35755301e-01 -6.68142676e-01 4.44122791e-01 -1.06816852e+00
4.73007262e-02 -6.00754023e-01 -2.16118231e-01 -1.22068055e-01
-9.53828394e-02 2.03833953e-01 5.97155467e-03 3.21506023e-01
-5.49538016e-01 7.06092179e-01 7.87496269e-01 3.85695547e-01
-3.89067382e-01 -2.68702120e-01 -3.13076198e-01 5.94445825e-01
5.78823924e-01 -5.08856535e-01 -6.10365160e-02 9.56294388e-02
4.66662794e-01 -1.10030854e-02 1.84696633e-02 -9.73475695e-01
3.96270692e-01 -4.85925257e-01 1.18524872e-01 -5.21332860e-01
2.66556978e-01 -9.68488932e-01 2.69892305e-01 6.73797011e-01
-3.12093228e-01 2.09237173e-01 -2.06990629e-01 3.51571321e-01
-1.79155231e-01 -1.21749930e-01 1.08340502e+00 1.96653724e-01
-4.99740034e-01 1.48516834e-01 -5.23087680e-01 2.58126587e-01
1.52239192e+00 -3.70456755e-01 1.06375098e-01 6.79941103e-02
-9.70792294e-01 3.19293141e-01 8.86330068e-01 -1.12294048e-01
5.15982628e-01 -1.46722460e+00 -7.28691339e-01 6.28554344e-01
-5.18072397e-03 3.30646932e-01 -1.28487781e-01 9.23394799e-01
-6.90268993e-01 1.59604445e-01 -1.49593093e-02 -7.99435735e-01
-9.78092849e-01 8.36427093e-01 7.84729123e-01 1.76834017e-01
-3.74336541e-01 5.01544356e-01 -5.17885625e-01 -6.59864008e-01
-1.19357243e-01 -2.98278004e-01 3.82504076e-01 9.41807851e-02
8.53898823e-02 9.18929517e-01 4.98752147e-01 -7.95496166e-01
-8.89055490e-01 8.32789123e-01 2.59040952e-01 -1.44600675e-01
1.35538054e+00 1.16031021e-01 -2.87302405e-01 2.92873442e-01
1.06357360e+00 -4.12635505e-01 -1.31873417e+00 -1.51644289e-01
5.25184035e-01 -3.00184518e-01 2.67054409e-01 -3.78915966e-01
-8.64513934e-01 4.89161342e-01 4.56067711e-01 4.12723005e-01
1.16435945e+00 -4.29211527e-01 1.76306546e-01 8.97626728e-02
1.61595136e-01 -1.55503929e+00 -2.72387117e-01 3.91137153e-01
7.89018989e-01 -1.17715549e+00 4.47089553e-01 -2.19190419e-01
-2.97817469e-01 9.96174574e-01 6.40708327e-01 -7.54954815e-01
9.82345700e-01 5.57649076e-01 -3.17096084e-01 -3.79970968e-01
-6.94413781e-01 1.54953953e-02 5.26133478e-01 4.68068212e-01
2.94015527e-01 8.59515276e-03 -6.80675089e-01 1.45615414e-01
-3.12196642e-01 -3.64975631e-01 3.22045594e-01 1.05166733e+00
-3.71373743e-01 -8.05755496e-01 -6.44165158e-01 1.82146490e-01
6.27964139e-02 2.29761243e-01 -5.58950245e-01 7.84279525e-01
-5.89188933e-02 1.10301423e+00 2.91765958e-01 -1.05482489e-01
3.85012209e-01 6.48332536e-02 1.74373209e-01 -5.30962050e-01
-6.13689721e-01 -9.07154828e-02 -2.86241084e-01 -7.03992248e-01
-4.43317711e-01 -1.02836728e+00 -1.41052294e+00 -5.78019500e-01
-2.76493669e-01 4.25648272e-01 9.93515909e-01 8.59356344e-01
5.89404404e-01 2.83097804e-01 9.98254657e-01 -1.00223231e+00
-5.64186215e-01 -6.27947032e-01 -7.16313839e-01 6.32235110e-01
5.62065899e-01 -3.15750241e-01 -4.77877527e-01 -1.13531016e-01] | [7.408164024353027, 4.241969585418701] |
67139983-ed88-4493-b040-e0248b5763b4 | extend-extractive-entity-disambiguation | null | null | https://aclanthology.org/2022.acl-long.177 | https://aclanthology.org/2022.acl-long.177.pdf | ExtEnD: Extractive Entity Disambiguation | Local models for Entity Disambiguation (ED) have today become extremely powerful, in most part thanks to the advent of large pre-trained language models. However, despite their significant performance achievements, most of these approaches frame ED through classification formulations that have intrinsic limitations, both computationally and from a modeling perspective. In contrast with this trend, here we propose ExtEnD, a novel local formulation for ED where we frame this task as a text extraction problem, and present two Transformer-based architectures that implement it. Based on experiments in and out of domain, and training over two different data regimes, we find our approach surpasses all its competitors in terms of both data efficiency and raw performance. ExtEnD outperforms its alternatives by as few as 6 F1 points on the more constrained of the two data regimes and, when moving to the other higher-resourced regime, sets a new state of the art on 4 out of 4 benchmarks under consideration, with average improvements of 0.7 F1 points overall and 1.1 F1 points out of domain. In addition, to gain better insights from our results, we also perform a fine-grained evaluation of our performances on different classes of label frequency, along with an ablation study of our architectural choices and an error analysis. We release our code and models for research purposes at https://github.com/SapienzaNLP/extend. | ['Roberto Navigli', 'Luigi Procopio', 'Edoardo Barba'] | null | null | null | null | acl-2022-5 | ['entity-disambiguation'] | ['natural-language-processing'] | [-4.05659936e-02 -1.67305991e-02 -3.66543770e-01 -2.64101505e-01
-1.08899915e+00 -8.76594067e-01 1.16766393e+00 5.22960603e-01
-8.17866504e-01 7.76828527e-01 2.21596256e-01 -3.50853890e-01
-2.85291195e-01 -5.96922934e-01 -4.73301560e-01 -3.26282859e-01
3.88916060e-02 8.01867068e-01 2.07498685e-01 -3.45573395e-01
-9.70033407e-02 3.45771462e-01 -1.27378118e+00 3.08741152e-01
8.28591704e-01 9.85570371e-01 -1.49633596e-02 2.19225183e-01
-3.08339298e-01 5.62388718e-01 -5.03040671e-01 -7.68493235e-01
1.53846264e-01 1.31889746e-01 -1.21552920e+00 -2.57194430e-01
4.62778389e-01 -2.03761924e-02 -1.80862665e-01 9.04045463e-01
6.24764502e-01 6.49832860e-02 4.25698936e-01 -9.78248775e-01
-5.82522154e-01 8.02290022e-01 -2.58353323e-01 1.80397213e-01
3.29913825e-01 -1.11537062e-01 1.38471329e+00 -1.09560382e+00
8.35757613e-01 1.03700137e+00 7.98756301e-01 4.06092912e-01
-1.28098488e+00 -5.92902184e-01 3.49789709e-01 7.19562992e-02
-1.52686214e+00 -6.77332342e-01 2.44045660e-01 -2.40368024e-01
1.37826753e+00 1.97280020e-01 2.41148919e-02 1.12852895e+00
-1.26557350e-01 9.80638623e-01 1.14537120e+00 -5.38457632e-01
1.27719343e-01 1.47018850e-01 2.55558193e-01 4.04712766e-01
3.55399638e-01 -1.27959371e-01 -3.06135237e-01 -2.43051142e-01
1.97076872e-01 -3.44167113e-01 -2.21289754e-01 -2.05312237e-01
-1.10511959e+00 6.61719799e-01 3.62655520e-01 7.56199479e-01
-2.66992331e-01 -9.92184207e-02 4.88938689e-01 1.66493297e-01
7.30732679e-01 4.78351235e-01 -8.81999671e-01 -2.32706726e-01
-1.09568858e+00 3.70162159e-01 8.50740910e-01 9.11957204e-01
5.11463165e-01 -3.33269238e-01 -2.98851371e-01 1.00569880e+00
-4.46218811e-02 3.60019952e-01 5.30868471e-01 -4.63255823e-01
6.06734574e-01 5.62239170e-01 1.77148134e-01 -8.09634030e-01
-6.75469518e-01 -8.89028966e-01 -5.81494749e-01 -3.43541265e-01
5.10896087e-01 -1.13147367e-02 -8.38806689e-01 1.78344142e+00
1.58404499e-01 1.30626326e-02 -1.02527037e-01 6.14079356e-01
8.03091109e-01 3.70518476e-01 5.37817776e-01 2.40347553e-02
1.84553862e+00 -8.61284077e-01 -6.09904826e-01 -5.61521471e-01
1.03360343e+00 -8.28835487e-01 1.01155734e+00 3.47162873e-01
-9.61278915e-01 -2.74337381e-01 -8.18871200e-01 -3.87629211e-01
-7.66487956e-01 4.05705243e-01 5.42756498e-01 6.51360154e-01
-1.08379352e+00 5.69582045e-01 -8.44920933e-01 -7.03659475e-01
3.43822569e-01 3.45939696e-01 -3.40962350e-01 4.53609100e-04
-1.39471054e+00 1.07860005e+00 4.22874838e-01 -2.70312250e-01
-4.54168767e-01 -7.98295498e-01 -7.01511025e-01 2.31247276e-01
4.74666744e-01 -6.73382998e-01 1.45510912e+00 -4.86603379e-01
-1.10860336e+00 1.24640632e+00 -1.74662814e-01 -7.75284708e-01
5.84252000e-01 -5.59078932e-01 -6.06414139e-01 -2.68263191e-01
2.41059154e-01 5.94687700e-01 1.63997382e-01 -9.22046959e-01
-8.69253397e-01 -2.15795442e-01 2.79436976e-01 -5.43065183e-03
-5.22249997e-01 2.80359328e-01 -5.96584439e-01 -6.92319155e-01
-3.80977333e-01 -9.17765200e-01 -6.66133240e-02 -4.61491138e-01
-3.90066147e-01 -5.43398976e-01 5.12348056e-01 -5.47588348e-01
1.54639137e+00 -2.11719775e+00 -1.14216119e-01 -1.30510256e-01
3.58135402e-01 5.55917799e-01 1.40451873e-03 5.09394228e-01
-1.80897817e-01 4.43895638e-01 -3.01266342e-01 -8.10374856e-01
4.25120473e-01 -5.36294430e-02 -3.26807290e-01 2.61708915e-01
2.91759938e-01 8.35032046e-01 -9.19003308e-01 -3.41303140e-01
1.79076657e-01 5.01111090e-01 -3.96443784e-01 -1.84497178e-01
-2.02784091e-01 1.28657207e-01 -4.59362268e-01 4.29897934e-01
6.49900198e-01 -4.03018326e-01 4.62248296e-01 -1.33257866e-01
-1.48788527e-01 7.84873605e-01 -1.15054727e+00 1.83918858e+00
-7.14780807e-01 4.93938357e-01 4.11607437e-02 -1.00284159e+00
7.30817556e-01 4.48088288e-01 5.25778592e-01 -9.20571268e-01
1.72721192e-01 5.41292131e-01 -2.24585414e-01 -3.10015231e-02
8.65001798e-01 -6.11390173e-02 -2.60910749e-01 2.52286911e-01
2.34598801e-01 3.98942083e-01 4.79123771e-01 2.47732177e-01
1.24768841e+00 1.19066119e-01 6.09108448e-01 -5.08914173e-01
4.13070828e-01 -5.97979166e-02 4.03782666e-01 6.47523940e-01
-5.83637394e-02 4.44250673e-01 3.80917758e-01 -3.08017850e-01
-8.36778581e-01 -8.59641969e-01 -6.07854128e-01 1.16816676e+00
6.18397817e-02 -8.81372690e-01 -6.32162929e-01 -1.04902232e+00
6.19705096e-02 9.53926265e-01 -4.71323222e-01 1.54786542e-01
-6.03491783e-01 -1.01006234e+00 8.22089493e-01 5.16515195e-01
5.88329971e-01 -7.31714368e-01 -3.53822261e-01 3.25961500e-01
-2.41655126e-01 -1.56503212e+00 -1.58531561e-01 3.01964492e-01
-5.69799304e-01 -8.09411168e-01 -6.00100696e-01 -6.30315244e-01
4.90602739e-02 8.78083259e-02 1.62723577e+00 3.07698492e-02
-1.07487366e-01 2.34894663e-01 -4.77801055e-01 -3.45587820e-01
-2.24299788e-01 8.07117224e-01 -1.10312074e-01 -2.65495658e-01
7.96619475e-01 -4.74078625e-01 -5.00159442e-01 2.68777728e-01
-8.44058096e-01 -1.58935502e-01 5.48882008e-01 6.98259830e-01
3.48340392e-01 6.18578354e-03 5.01405537e-01 -1.14915860e+00
5.72073102e-01 -7.88292408e-01 -4.62505847e-01 1.96815044e-01
-9.10433531e-01 2.18825147e-01 6.13661230e-01 5.04968576e-02
-1.04011869e+00 -3.04702014e-01 -2.76224732e-01 2.12746905e-03
-3.36133331e-01 5.43485343e-01 -1.28981262e-01 4.25550759e-01
9.13794100e-01 -1.01623610e-01 -5.57788074e-01 -9.19788659e-01
3.76448870e-01 7.86975384e-01 3.33330363e-01 -9.13451374e-01
6.57552600e-01 3.94362688e-01 -3.80109102e-01 -6.00557566e-01
-1.00404739e+00 -6.00875080e-01 -5.35072267e-01 3.51611406e-01
7.28865504e-01 -1.09375179e+00 -4.32976216e-01 1.55116230e-01
-8.80484581e-01 -2.93708742e-01 -1.55692384e-01 1.93850473e-01
-1.37469217e-01 2.57116884e-01 -6.15736425e-01 -5.97266316e-01
-3.14868599e-01 -9.48467493e-01 1.35573113e+00 -1.17790386e-01
-4.05648828e-01 -1.30032170e+00 3.38231809e-02 1.95744112e-01
6.03167355e-01 1.39863640e-01 8.30900073e-01 -1.11479008e+00
-2.65443414e-01 -2.78887480e-01 -2.66296446e-01 1.08589992e-01
5.98499216e-02 -2.58475840e-01 -1.11510706e+00 -4.30884570e-01
-4.21036214e-01 -9.03967991e-02 9.51668382e-01 4.74984124e-02
9.33762014e-01 -3.98253463e-02 -7.32125342e-01 3.51071447e-01
1.52004373e+00 -6.12151437e-02 4.60084677e-01 8.24138224e-01
3.76986474e-01 4.12530899e-01 6.55497253e-01 3.22363973e-01
5.62919319e-01 1.14199471e+00 1.51052594e-01 -1.36763245e-01
-3.43624145e-01 -2.15600729e-01 3.02768886e-01 5.69020510e-01
4.62336950e-02 -5.45825303e-01 -1.22626126e+00 8.74779761e-01
-1.75324023e+00 -7.50581443e-01 5.30496314e-02 2.29181576e+00
8.36718261e-01 3.79233718e-01 2.79203415e-01 8.01824629e-02
6.36486530e-01 1.60537347e-01 -2.42806107e-01 -2.81158000e-01
-3.14420670e-01 5.82957685e-01 4.60283965e-01 4.26772892e-01
-1.37294340e+00 1.07780659e+00 5.52059078e+00 1.12293780e+00
-1.12024057e+00 2.06757247e-01 6.13403440e-01 -1.48284107e-01
-3.21233571e-01 1.08850755e-01 -1.15347207e+00 5.58350563e-01
1.17585850e+00 -1.91510469e-01 2.01729685e-01 6.65229619e-01
-2.60105073e-01 1.64747834e-01 -1.30110395e+00 9.00652647e-01
-2.48941764e-01 -1.18203962e+00 -7.05657303e-02 9.44309086e-02
4.95722443e-01 5.09227276e-01 -2.29313206e-02 7.46773660e-01
4.72584277e-01 -1.05041051e+00 8.49865913e-01 5.89729287e-02
7.66286075e-01 -5.60769141e-01 8.82146597e-01 3.63593519e-01
-1.30852389e+00 2.00651139e-02 -7.05365613e-02 1.11166351e-02
9.83195305e-02 9.17490840e-01 -7.13240921e-01 9.04874682e-01
7.94659615e-01 6.82920814e-01 -6.69763207e-01 8.00677061e-01
-2.12117150e-01 5.37177861e-01 -4.89709288e-01 2.05938786e-01
3.37196678e-01 3.00463676e-01 4.71920043e-01 1.71128523e+00
2.23634407e-01 -1.10413954e-01 1.85717732e-01 6.39032304e-01
-5.21258116e-01 2.85961002e-01 -3.69480669e-01 -1.87517181e-02
8.27116787e-01 1.47955382e+00 -5.36020041e-01 -3.59607249e-01
-5.61457276e-01 7.03000307e-01 6.03722334e-01 1.51660040e-01
-8.93499374e-01 -4.90374833e-01 6.33516312e-01 6.88951984e-02
3.63836259e-01 -2.01179981e-01 -2.80844927e-01 -1.35931420e+00
2.20518723e-01 -8.11628103e-01 6.49917185e-01 -3.91359955e-01
-1.25912726e+00 7.34989762e-01 1.32483155e-01 -1.10200274e+00
-4.45639998e-01 -7.55780697e-01 -1.80045322e-01 7.83429086e-01
-1.84502411e+00 -1.11685789e+00 -1.45144358e-01 2.36552000e-01
3.84523124e-01 9.66982171e-02 1.02281737e+00 8.20227087e-01
-6.18729711e-01 9.77664828e-01 3.63447249e-01 2.77406424e-01
1.04602814e+00 -1.32679033e+00 7.35328257e-01 8.00986111e-01
4.81002420e-01 6.57641590e-01 6.93047404e-01 -2.20111445e-01
-8.57862890e-01 -1.01276135e+00 1.43112051e+00 -8.08911026e-01
7.20132172e-01 -5.31904936e-01 -7.30359554e-01 9.48038518e-01
3.37672889e-01 -3.15108001e-02 6.84207439e-01 7.16925204e-01
-4.57430035e-01 1.26748309e-01 -1.05998147e+00 5.06281435e-01
1.32397759e+00 -4.93738472e-01 -6.29180908e-01 3.97954643e-01
4.78385329e-01 -5.56494236e-01 -1.33273017e+00 4.39706653e-01
3.86029989e-01 -8.82395506e-01 9.07005668e-01 -5.93740702e-01
3.31614435e-01 -1.43622771e-01 -2.79257923e-01 -1.23309839e+00
-3.86892200e-01 -5.79998016e-01 7.35541731e-02 1.72726655e+00
6.35875642e-01 -7.41585910e-01 5.80804706e-01 4.79468644e-01
-1.18550874e-01 -8.03191900e-01 -8.34294856e-01 -1.01451480e+00
5.61213911e-01 -5.75863779e-01 7.16671824e-01 1.26518381e+00
-4.74346103e-03 6.12071037e-01 -7.45426118e-02 2.47516800e-02
2.63855577e-01 2.17712462e-01 5.59664905e-01 -1.22145104e+00
-2.35724509e-01 -6.76740050e-01 -2.82851309e-01 -1.30067587e+00
3.01740140e-01 -1.18417180e+00 -2.08743349e-01 -1.42986977e+00
1.93474516e-01 -7.76984096e-01 -5.08480430e-01 7.81596005e-01
-2.16286480e-01 3.20168555e-01 2.99122214e-01 3.04459959e-01
-7.66562700e-01 4.23490644e-01 5.33766627e-01 -5.91679849e-02
1.21783562e-01 -2.53173262e-01 -8.82204413e-01 5.70273101e-01
7.93972015e-01 -5.66606641e-01 -1.11666866e-01 -9.12481427e-01
2.61628658e-01 -3.04545790e-01 2.55484670e-01 -9.84396398e-01
9.34353918e-02 2.02099919e-01 9.40457359e-02 -1.79925203e-01
1.67690933e-01 -6.14895284e-01 -1.38964981e-01 1.93635941e-01
-3.46760541e-01 2.34468095e-02 5.44514298e-01 1.79254606e-01
-3.52588803e-01 -5.98362051e-02 5.88660896e-01 1.19061798e-01
-8.57350230e-01 3.08192164e-01 -1.90034546e-02 4.64332908e-01
7.23644257e-01 7.39859119e-02 -5.34690499e-01 -8.80884528e-02
-7.54586399e-01 3.27638865e-01 4.68439966e-01 5.78648806e-01
-2.00133845e-01 -1.17362702e+00 -8.64923477e-01 -9.45047587e-02
3.36210907e-01 -5.62064648e-02 1.16114974e-01 9.51852798e-01
-2.29304761e-01 9.87092376e-01 1.22184224e-01 -5.18697798e-01
-9.54656303e-01 4.49273050e-01 2.78470099e-01 -8.62053394e-01
-4.61820453e-01 6.40097320e-01 2.53016859e-01 -5.80276251e-01
7.99910128e-02 -3.54293376e-01 -1.41848549e-01 8.58660266e-02
4.43462998e-01 3.71993870e-01 7.44295120e-01 -5.33605576e-01
-7.27255642e-01 3.69790792e-01 -3.91137153e-01 6.23401962e-02
1.24864030e+00 -1.68153837e-01 1.49563596e-01 2.03685656e-01
1.11128414e+00 3.55118901e-01 -8.36893797e-01 -6.28340840e-01
5.56417465e-01 -1.88318625e-01 1.02283806e-01 -1.19844043e+00
-8.84368241e-01 7.39360809e-01 4.33662266e-01 4.02427018e-01
1.17205119e+00 1.32121786e-01 8.28107059e-01 2.09087595e-01
4.00558949e-01 -9.16770995e-01 -5.38831294e-01 7.01085985e-01
5.09917617e-01 -1.10795987e+00 6.57908767e-02 -3.79612118e-01
-3.69783938e-01 9.03920352e-01 1.79814845e-01 -4.19566967e-02
5.33694208e-01 3.73535603e-01 -5.24751432e-02 -1.54850334e-01
-8.32606554e-01 -3.45963240e-01 2.53181189e-01 2.92053729e-01
9.39559698e-01 5.32443076e-02 -4.89714831e-01 7.76990950e-01
-3.52289289e-01 3.27663310e-02 2.98878495e-02 9.06767666e-01
-1.88166127e-01 -1.48767912e+00 6.34992048e-02 2.68074334e-01
-9.04670000e-01 -3.75871867e-01 -3.72199565e-01 1.28641653e+00
7.31685758e-02 7.93928623e-01 7.63181150e-02 -2.01709181e-01
6.18846297e-01 2.37686187e-01 1.94886252e-01 -6.06662631e-01
-7.83127487e-01 -2.62524337e-01 5.25041521e-01 -6.16020560e-01
-4.65003252e-01 -7.50536501e-01 -1.25719380e+00 -5.59645951e-01
-3.05454105e-01 2.75875658e-01 5.53197026e-01 1.09238553e+00
6.89858198e-01 3.89353305e-01 2.01782003e-01 -3.99537116e-01
-6.62689626e-01 -1.01998556e+00 -4.36241657e-01 4.96735483e-01
9.91038159e-02 -7.59769976e-01 -2.03760713e-01 -2.77847588e-01] | [9.7316312789917, 8.800028800964355] |
fd02dcc9-5a7e-41c1-92fa-d8956efd5589 | semi-supervised-graph-imbalanced-regression | 2305.12087 | null | https://arxiv.org/abs/2305.12087v1 | https://arxiv.org/pdf/2305.12087v1.pdf | Semi-Supervised Graph Imbalanced Regression | Data imbalance is easily found in annotated data when the observations of certain continuous label values are difficult to collect for regression tasks. When they come to molecule and polymer property predictions, the annotated graph datasets are often small because labeling them requires expensive equipment and effort. To address the lack of examples of rare label values in graph regression tasks, we propose a semi-supervised framework to progressively balance training data and reduce model bias via self-training. The training data balance is achieved by (1) pseudo-labeling more graphs for under-represented labels with a novel regression confidence measurement and (2) augmenting graph examples in latent space for remaining rare labels after data balancing with pseudo-labels. The former is to identify quality examples from unlabeled data whose labels are confidently predicted and sample a subset of them with a reverse distribution from the imbalanced annotated data. The latter collaborates with the former to target a perfect balance using a novel label-anchored mixup algorithm. We perform experiments in seven regression tasks on graph datasets. Results demonstrate that the proposed framework significantly reduces the error of predicted graph properties, especially in under-represented label areas. | ['Meng Jiang', 'Tengfei Luo', 'Eric Inae', 'Tong Zhao', 'Gang Liu'] | 2023-05-20 | null | null | null | null | ['graph-regression'] | ['graphs'] | [ 6.37638390e-01 5.64158022e-01 -7.43049443e-01 -6.15099728e-01
-7.82249987e-01 -4.03407395e-01 6.02798797e-02 7.61007786e-01
-5.68085443e-03 1.15372705e+00 -7.23141208e-02 -1.55266851e-01
-9.97389555e-02 -8.55330884e-01 -8.63664627e-01 -9.61928725e-01
1.40172213e-01 9.56037700e-01 -1.06812073e-02 2.43108168e-01
1.36357516e-01 2.57023394e-01 -1.24522769e+00 2.75418937e-01
1.41370869e+00 8.60636950e-01 -1.94680735e-01 3.17462772e-01
-3.58588487e-01 9.99219894e-01 -3.96977067e-01 -5.04858136e-01
2.65952408e-01 -4.68119919e-01 -5.35609424e-01 4.22160923e-01
5.32189131e-01 1.64697021e-01 8.44178274e-02 1.26960719e+00
4.70504314e-01 -2.09734559e-01 9.73307669e-01 -1.57902515e+00
-3.93646032e-01 8.17021668e-01 -1.03755116e+00 -2.17776194e-01
1.67045370e-01 -3.30770239e-02 1.25101840e+00 -5.83748043e-01
8.58769774e-01 1.07631409e+00 7.72720635e-01 4.35117245e-01
-1.52505839e+00 -9.12940264e-01 2.78290093e-01 -7.20524937e-02
-1.27188182e+00 -2.21869513e-01 1.11863065e+00 -6.56166553e-01
4.26035255e-01 1.07260644e-01 4.07186449e-01 1.02087152e+00
-3.91739644e-02 5.36144078e-01 1.19135427e+00 -3.74409825e-01
2.76840270e-01 4.16851819e-01 4.46154445e-01 7.62380302e-01
5.25287271e-01 -5.15995994e-02 -6.05493844e-01 -3.20190281e-01
9.92015190e-03 4.95448783e-02 -1.03396915e-01 -7.51775622e-01
-9.18743014e-01 9.50779021e-01 5.01006544e-01 -3.00129149e-02
-3.24097186e-01 -2.56017357e-01 3.90763670e-01 3.74247760e-01
9.45290208e-01 4.79429126e-01 -6.17699146e-01 4.47851926e-01
-8.88426244e-01 1.58283059e-02 8.60210121e-01 1.26987875e+00
1.09612036e+00 -1.65320486e-01 -2.53628999e-01 9.55657601e-01
3.17063808e-01 1.61539331e-01 2.62993902e-01 -2.68521518e-01
7.70105660e-01 1.13815510e+00 7.65116438e-02 -8.89776587e-01
-5.73593974e-01 -7.08508372e-01 -7.99943388e-01 -1.48035362e-02
8.72696340e-01 4.02382724e-02 -1.04772472e+00 1.61031401e+00
6.56371474e-01 -2.81692654e-01 -1.59345716e-01 7.06639409e-01
7.33779192e-01 4.73436594e-01 4.47106957e-01 -7.53922820e-01
1.10458183e+00 -9.91577804e-01 -7.44501591e-01 -7.74758607e-02
1.07623005e+00 -4.80798185e-01 1.16267598e+00 5.17452657e-01
-4.47326809e-01 -5.05740821e-01 -1.13340843e+00 1.24465367e-02
-3.50562453e-01 3.50080520e-01 6.39882922e-01 7.17096567e-01
-4.90160435e-01 9.32806015e-01 -4.58638966e-01 6.03088103e-02
8.28989863e-01 3.94485921e-01 -3.64093214e-01 -2.11905837e-01
-1.00979817e+00 4.25437599e-01 5.25602818e-01 -2.89515387e-02
-7.43190110e-01 -7.63574362e-01 -8.52620363e-01 -2.16143847e-01
6.59631670e-01 -3.38634342e-01 5.30939937e-01 -1.21320438e+00
-1.14270759e+00 1.01106381e+00 3.71854827e-02 -2.15334967e-01
8.59377980e-01 1.15258381e-01 -2.37511724e-01 -2.95728385e-01
2.49205396e-01 5.06540120e-01 9.98414457e-01 -1.38565791e+00
-4.44248140e-01 -6.72205448e-01 -4.71026808e-01 1.15085542e-01
-3.95953834e-01 -5.54179072e-01 6.30394444e-02 -4.62436408e-01
5.35774767e-01 -8.69352281e-01 -2.88414150e-01 -2.20098108e-01
-9.12147462e-01 -3.40001643e-01 4.99756843e-01 -5.06216586e-01
1.16198540e+00 -1.78311181e+00 1.11576423e-01 4.74996269e-01
8.31673026e-01 -1.32181108e-01 -4.49705720e-02 2.29636475e-01
-4.76145446e-01 1.73796669e-01 -1.62011445e-01 -3.00576270e-01
-1.80537596e-01 -6.64633289e-02 -1.15004733e-01 7.12874293e-01
2.79791057e-01 5.32071531e-01 -1.13753676e+00 -6.61809444e-01
-7.62424618e-02 9.96632967e-04 -4.16229844e-01 3.28282446e-01
-5.95884144e-01 7.17448950e-01 -4.63315904e-01 8.87702107e-01
8.07824075e-01 -5.06456614e-01 5.57963669e-01 -4.45959330e-01
3.11636060e-01 1.79617569e-01 -1.09662437e+00 1.28286874e+00
1.55861797e-02 5.52857928e-02 -4.00473475e-01 -1.04005098e+00
1.40652168e+00 -6.84304088e-02 5.69078624e-01 -4.67451245e-01
8.20333362e-02 2.52902776e-01 3.95699292e-02 -5.04193187e-01
1.62310973e-01 -4.87241030e-01 7.78146386e-02 3.74836534e-01
9.35108736e-02 -2.00208172e-01 3.67781311e-01 1.08406857e-01
1.06264794e+00 1.90754980e-01 3.06403548e-01 -1.68750912e-01
5.21707356e-01 -3.80034409e-02 8.26467752e-01 4.58687782e-01
-2.76474357e-01 2.13362530e-01 1.04056823e+00 -5.12979925e-01
-1.15035415e+00 -7.48794973e-01 -2.16559377e-02 1.19332218e+00
4.08373512e-02 -4.75226104e-01 -4.39494908e-01 -1.33478594e+00
1.47682533e-01 4.82626766e-01 -5.97556651e-01 -3.18196088e-01
-1.02642737e-01 -9.33862865e-01 8.70571509e-02 2.48996407e-01
-8.89999419e-02 -9.41910982e-01 3.74182880e-01 1.95890576e-01
1.92627739e-02 -8.89095426e-01 -3.00738513e-01 8.40056717e-01
-9.30449307e-01 -1.31964970e+00 -2.64596790e-01 -8.21323633e-01
1.23147178e+00 -4.26574610e-02 1.27266133e+00 2.07119524e-01
-1.19125403e-01 -4.58777070e-01 -4.17147070e-01 -4.68171775e-01
-7.50231743e-01 3.27635944e-01 -9.09312591e-02 5.49022071e-02
4.71953005e-01 -4.28734601e-01 -3.78438622e-01 4.43484694e-01
-5.96843362e-01 9.46754441e-02 3.77461672e-01 9.72017467e-01
1.00952923e+00 1.21611215e-01 8.72275114e-01 -1.70270586e+00
5.11725724e-01 -5.20143926e-01 -7.30749607e-01 3.00839484e-01
-1.09625959e+00 2.65843958e-01 7.67469406e-01 -6.48796260e-01
-5.88949263e-01 2.94529498e-01 1.48718312e-01 -2.88808107e-01
1.85209455e-03 3.87319177e-01 -4.33113396e-01 6.13658801e-02
1.00003886e+00 -2.68144161e-01 6.75815046e-02 -1.73725292e-01
2.74720073e-01 6.85951948e-01 -5.66514321e-02 -6.67861819e-01
6.54800951e-01 1.16235845e-01 2.55604088e-01 -4.37226593e-01
-1.21964633e+00 -4.80230659e-01 -7.13574648e-01 -3.00036937e-01
5.48718572e-01 -8.49538326e-01 -5.56936204e-01 3.45527470e-01
-6.00852311e-01 -3.69919151e-01 -5.23550510e-01 3.95346105e-01
-4.51534033e-01 3.91027570e-01 -4.12015885e-01 -9.10130799e-01
-2.55399674e-01 -1.21975410e+00 1.11434066e+00 5.84224313e-02
-1.62300766e-01 -8.42058003e-01 1.63330555e-01 5.91991186e-01
-1.65920615e-01 5.85270286e-01 1.14772618e+00 -1.15125322e+00
-3.46557647e-01 -5.17117381e-01 -2.55439311e-01 3.58221292e-01
3.75125527e-01 1.20861754e-02 -1.06365085e+00 -4.47051793e-01
-2.55799443e-01 -9.19618845e-01 8.58510494e-01 3.71141493e-01
1.28289282e+00 -6.76435977e-02 -4.94912118e-01 4.78361011e-01
1.25377417e+00 -3.52551490e-02 2.20309108e-01 -7.58594722e-02
1.07068157e+00 9.29009378e-01 9.16496336e-01 2.17691392e-01
2.16104642e-01 4.74205226e-01 4.38501239e-01 -2.26901636e-01
-4.19428870e-02 -7.33298421e-01 3.16201560e-02 8.09461772e-01
1.99988544e-01 -4.42977428e-01 -6.97353423e-01 1.97193921e-01
-1.84462261e+00 -4.57081944e-01 -4.84266430e-01 2.22607541e+00
1.16120887e+00 4.72266167e-01 3.48780811e-01 3.84333909e-01
1.13683975e+00 7.66264647e-02 -9.11205888e-01 1.31303906e-01
1.89506728e-02 -2.24751770e-01 6.80986524e-01 5.20775676e-01
-1.11759424e+00 9.76078689e-01 5.65061378e+00 9.26353037e-01
-8.20278049e-01 8.55719447e-02 1.10767460e+00 1.57628939e-01
-5.15471518e-01 1.92611039e-01 -8.89894664e-01 4.96237278e-01
6.69820189e-01 1.76817030e-01 1.76267192e-01 9.72001195e-01
-5.23742288e-03 8.64937603e-02 -1.28409493e+00 9.47128236e-01
-2.93725468e-02 -9.09143686e-01 -7.61162937e-02 1.48309410e-01
8.81958127e-01 -3.36451381e-01 -3.25913548e-01 3.43099564e-01
3.67152214e-01 -1.00201344e+00 5.02560019e-01 5.28993666e-01
8.74189734e-01 -6.05369925e-01 6.61113918e-01 5.28630793e-01
-7.03552663e-01 2.89907083e-02 -6.17082179e-01 6.40903339e-02
-2.00731561e-01 1.32382119e+00 -1.20041537e+00 5.24022996e-01
8.49010199e-02 9.25846934e-01 -6.15512133e-01 7.93578029e-01
-3.54531407e-01 5.98473132e-01 -1.91774398e-01 -1.47541687e-01
-3.28175962e-01 -6.49946928e-01 2.38203093e-01 8.69517446e-01
-6.21942505e-02 -3.74196589e-01 4.57238525e-01 7.55328059e-01
-5.37133455e-01 4.64377642e-01 -6.58708632e-01 -2.20755950e-01
3.86421800e-01 1.40511465e+00 -1.06686747e+00 -2.47820258e-01
-1.21575698e-01 5.73062003e-01 7.49537945e-01 1.91820398e-01
-8.30780208e-01 -3.48814838e-02 -1.25629559e-01 3.21756035e-01
-1.55070126e-01 3.33724827e-01 -3.63216966e-01 -9.52565193e-01
-6.90549389e-02 -8.96022916e-01 7.25989759e-01 -4.31406170e-01
-1.68930078e+00 5.26623070e-01 -3.17814827e-01 -1.40615976e+00
7.75032714e-02 -5.37352622e-01 -8.99321735e-02 5.68057179e-01
-1.45465374e+00 -1.06011879e+00 -5.69915116e-01 2.58110791e-01
1.93853661e-01 -6.54200464e-02 5.46403587e-01 5.02091944e-01
-6.53836966e-01 7.18052983e-01 -1.06693633e-01 -2.22213000e-01
1.11347103e+00 -1.56819439e+00 -7.67476484e-02 4.17880654e-01
4.88125123e-02 2.30926782e-01 8.17032516e-01 -1.14108467e+00
-9.33072090e-01 -1.14658070e+00 7.34781206e-01 -3.37672472e-01
5.37967205e-01 -6.40964746e-01 -1.05509007e+00 4.49575186e-01
-6.15535080e-01 3.45675975e-01 9.44441617e-01 2.66541868e-01
-4.42884207e-01 -2.54509181e-01 -1.30560613e+00 1.76485315e-01
1.19118750e+00 -3.99574071e-01 -5.00808321e-02 1.08661664e+00
7.05770850e-01 -3.93120617e-01 -9.89627242e-01 6.09233737e-01
3.37191284e-01 -7.26824045e-01 5.79025686e-01 -1.03495395e+00
3.12725484e-01 -4.94763672e-01 3.55831757e-02 -1.41864002e+00
-1.20729588e-01 -4.34222698e-01 -4.75705676e-02 1.38079703e+00
6.60905778e-01 -4.58449155e-01 1.26737201e+00 5.02769053e-01
5.78129478e-02 -8.16857278e-01 -5.55244923e-01 -5.32756507e-01
-2.32329339e-01 -1.58326507e-01 4.31216508e-01 1.19734895e+00
-1.31621540e-01 6.06145024e-01 -5.62569976e-01 -2.61970628e-02
6.61318839e-01 2.10065156e-01 9.25143063e-01 -1.53516734e+00
-1.56549335e-01 -2.38811653e-02 -4.29774314e-01 -6.46260917e-01
3.42085093e-01 -1.26268852e+00 7.15664700e-02 -1.27439737e+00
3.90303016e-01 -8.51223290e-01 -1.64784655e-01 4.64925051e-01
-4.49429244e-01 2.25736141e-01 -4.16165143e-01 2.85430312e-01
-7.75154531e-01 6.40344024e-01 1.50933099e+00 -3.34943384e-01
-2.94957221e-01 1.97162092e-01 -8.50578368e-01 5.98657370e-01
6.29430294e-01 -8.91493022e-01 -7.08364904e-01 3.73769373e-01
5.61269581e-01 2.25068145e-02 -1.68288335e-01 -7.04637110e-01
-1.23139590e-01 -1.07588597e-01 5.42934656e-01 -5.94052970e-01
-2.06928104e-01 -9.56566870e-01 2.50000060e-01 4.78378564e-01
-5.86181760e-01 -3.15986961e-01 -3.54532629e-01 8.75758231e-01
8.99813250e-02 -4.40767139e-01 8.03065479e-01 1.29613236e-01
3.32657620e-02 4.94147748e-01 3.80207479e-01 2.00767279e-01
1.15169501e+00 -1.61871970e-01 -4.49002892e-01 -2.35939100e-01
-1.13029420e+00 3.68062437e-01 5.28191686e-01 1.77164674e-01
2.26833314e-01 -1.13247323e+00 -6.22382164e-01 3.37845325e-01
4.73473936e-01 2.05585271e-01 1.06001861e-01 7.05123067e-01
-3.45004797e-01 -7.44093806e-02 5.10387160e-02 -7.98281789e-01
-1.12339377e+00 8.16763699e-01 1.77197620e-01 -7.88883626e-01
-2.28177875e-01 7.93971837e-01 1.03752412e-01 -7.48461783e-01
2.90461689e-01 -2.85222620e-01 -2.73740232e-01 3.96205813e-01
1.10115446e-01 3.55442882e-01 3.75399798e-01 -2.48240516e-01
-2.09964067e-01 3.58507603e-01 -3.19158643e-01 6.77479863e-01
1.44940817e+00 4.84308042e-02 -2.24632666e-01 6.13743722e-01
1.07283378e+00 1.92640185e-01 -1.42852902e+00 -1.69227809e-01
1.83416784e-01 -4.38227952e-01 -3.02667081e-01 -7.39660680e-01
-1.05357206e+00 5.31273186e-01 5.31388044e-01 3.18606257e-01
8.60767186e-01 2.69901250e-02 1.79172665e-01 3.24666873e-02
2.22575039e-01 -1.14594924e+00 3.21993917e-01 -1.34886233e-02
6.38669550e-01 -1.50760412e+00 4.38120753e-01 -9.69587743e-01
-6.90162838e-01 9.43778634e-01 7.84905970e-01 -4.19890322e-02
5.29641151e-01 1.21140979e-01 -3.32583860e-02 -5.82161486e-01
-5.98081768e-01 8.88400804e-03 3.34867328e-01 7.08476663e-01
5.06983995e-01 2.66198039e-01 -5.53807378e-01 4.97071028e-01
-2.02582687e-01 -2.45384216e-01 3.17513376e-01 5.32104611e-01
-3.04116964e-01 -1.27087569e+00 -7.03948513e-02 1.06435168e+00
-2.54594982e-01 6.00699335e-02 -4.82210606e-01 5.43219090e-01
2.09125951e-01 8.41069639e-01 -2.50960380e-01 -5.02962172e-01
3.03205699e-01 3.09920192e-01 4.16400582e-01 -7.46179521e-01
-2.69049734e-01 1.81920752e-01 3.47367346e-01 -2.90981084e-01
-4.09462452e-01 -4.33066517e-01 -9.80007350e-01 -1.00238551e-03
-8.72830391e-01 3.38242918e-01 4.78026539e-01 7.25248337e-01
8.28083307e-02 8.37722182e-01 8.64523828e-01 -4.60577369e-01
-7.35323489e-01 -1.17607450e+00 -8.87182295e-01 7.91183412e-01
1.82442605e-01 -9.10608351e-01 -5.88000894e-01 -1.08955353e-02] | [9.375925064086914, 4.08336877822876] |
aebeac5d-61b1-44be-885c-93d173c062a8 | don-t-retrain-just-rewrite-countering | 2305.16444 | null | https://arxiv.org/abs/2305.16444v1 | https://arxiv.org/pdf/2305.16444v1.pdf | Don't Retrain, Just Rewrite: Countering Adversarial Perturbations by Rewriting Text | Can language models transform inputs to protect text classifiers against adversarial attacks? In this work, we present ATINTER, a model that intercepts and learns to rewrite adversarial inputs to make them non-adversarial for a downstream text classifier. Our experiments on four datasets and five attack mechanisms reveal that ATINTER is effective at providing better adversarial robustness than existing defense approaches, without compromising task accuracy. For example, on sentiment classification using the SST-2 dataset, our method improves the adversarial accuracy over the best existing defense approach by more than 4% with a smaller decrease in task accuracy (0.5% vs 2.5%). Moreover, we show that ATINTER generalizes across multiple downstream tasks and classifiers without having to explicitly retrain it for those settings. Specifically, we find that when ATINTER is trained to remove adversarial perturbations for the sentiment classification task on the SST-2 dataset, it even transfers to a semantically different task of news classification (on AGNews) and improves the adversarial robustness by more than 10%. | ['Vivek Srikumar', 'Alakananda Vempala', 'Shalin Shah', 'Yingjie Fei', 'Temma Choji', 'Carter Wood Blum', 'Ashim Gupta'] | 2023-05-25 | null | null | null | null | ['adversarial-robustness', 'sentiment-analysis', 'news-classification'] | ['adversarial', 'natural-language-processing', 'natural-language-processing'] | [ 4.16778028e-01 3.95167828e-01 5.54048270e-02 -2.92328894e-01
-1.01041067e+00 -1.51447797e+00 7.98295975e-01 1.08348794e-01
-4.11198288e-01 4.53024089e-01 3.39359134e-01 -8.27177942e-01
4.31109816e-01 -7.17297792e-01 -9.29633617e-01 -5.83698988e-01
2.31344774e-01 8.88583958e-02 1.75197706e-01 -8.10998261e-01
-2.37196255e-02 5.77524900e-01 -7.28428304e-01 7.40600109e-01
4.24611151e-01 8.27812850e-01 -9.66891527e-01 9.56064224e-01
2.59488344e-01 8.60107183e-01 -1.17854536e+00 -1.14450288e+00
7.00511336e-01 -5.34006357e-02 -1.04676652e+00 -5.60430825e-01
7.82186508e-01 -3.84222507e-01 -6.59396052e-01 1.09476161e+00
4.38676029e-01 -1.41898915e-02 6.74197137e-01 -1.26451802e+00
-1.06825924e+00 1.24157178e+00 -2.65036523e-01 3.04559410e-01
-3.68790403e-02 2.81648874e-01 7.75883973e-01 -3.28136533e-01
3.10252875e-01 1.44109023e+00 7.28533685e-01 1.19691634e+00
-1.26604760e+00 -8.12311530e-01 4.50115770e-01 -4.60656792e-01
-7.96473205e-01 -5.35100758e-01 3.53951484e-01 -3.07943821e-01
8.77905011e-01 6.12467349e-01 -3.46397758e-01 1.91818714e+00
5.12838364e-01 5.04331291e-01 1.02511966e+00 -2.27153450e-01
1.66780636e-01 2.63910621e-01 3.45643640e-01 3.47444892e-01
2.65953600e-01 2.07745090e-01 -1.27954766e-01 -4.98157084e-01
6.90740161e-03 -1.57265902e-01 -3.33092391e-01 2.72883326e-01
-7.39591539e-01 9.73117232e-01 5.72629750e-01 2.46001244e-01
1.66865587e-01 2.69601792e-01 8.84356797e-01 8.56964648e-01
7.04446018e-01 1.00897706e+00 -9.33150768e-01 3.56287509e-01
-2.46546313e-01 1.87216207e-01 9.69506383e-01 6.75193250e-01
-5.26237823e-02 5.18950403e-01 -3.72930795e-01 3.89234185e-01
-3.03043693e-01 7.94759214e-01 6.23683572e-01 -5.14562249e-01
8.11046481e-01 1.89880848e-01 -2.07956135e-01 -8.27469528e-01
-2.43467167e-01 -4.63425517e-01 -6.30100250e-01 3.56696218e-01
6.79452300e-01 -7.11974084e-01 -1.07320666e+00 2.03055668e+00
-1.62429661e-02 -1.56412095e-01 6.73266649e-01 5.24236917e-01
4.63736326e-01 5.62075377e-01 4.30887401e-01 4.18270528e-01
1.30243063e+00 -8.59699190e-01 -2.04221129e-01 -7.24578619e-01
8.99519563e-01 -6.73603535e-01 1.10795307e+00 4.14012551e-01
-9.09684539e-01 -2.02776998e-01 -1.20902669e+00 -4.83972542e-02
-8.76899660e-01 -5.22459924e-01 3.60793293e-01 1.15555465e+00
-7.94230878e-01 5.91631055e-01 -6.55665398e-01 -1.65920138e-01
3.33007932e-01 3.66074651e-01 -5.39498806e-01 1.31202042e-01
-1.60864341e+00 1.12058890e+00 2.36692995e-01 -4.34859395e-01
-1.11331439e+00 -9.33598936e-01 -8.37400913e-01 1.68023512e-01
1.49676666e-01 -6.43233538e-01 1.46033728e+00 -1.48044229e+00
-1.61849511e+00 7.30509400e-01 3.95051956e-01 -1.00035417e+00
7.59020686e-01 -5.41838944e-01 -5.24941504e-01 6.47743121e-02
-2.83271730e-01 1.99555039e-01 1.10798061e+00 -1.02491105e+00
-1.56051934e-01 -3.73227507e-01 6.70635223e-01 -5.68024330e-02
-9.28736925e-01 3.21836263e-01 1.36041760e-01 -1.15794420e+00
-5.45302868e-01 -1.14304280e+00 -1.86378151e-01 -2.91358888e-01
-6.91708446e-01 2.46947393e-01 1.00406444e+00 -4.77328181e-01
7.09005833e-01 -2.15231943e+00 1.86367020e-01 2.28420887e-02
2.12028891e-01 6.98355734e-01 -4.30328876e-01 1.84048072e-01
-6.08003318e-01 6.04439676e-01 -3.33521605e-01 -2.73672134e-01
-4.11246344e-03 6.69848546e-03 -1.05067217e+00 4.88833547e-01
2.95333654e-01 8.88183892e-01 -5.11590660e-01 1.99834839e-01
-1.47247031e-01 2.49916241e-01 -7.01649308e-01 5.29090241e-02
-3.64656240e-01 3.21777165e-01 -5.01800656e-01 2.73837566e-01
6.21318042e-01 1.73437789e-01 -2.87384409e-02 2.20586851e-01
4.68231827e-01 2.72612810e-01 -4.70416099e-01 1.03613055e+00
-6.55202091e-01 6.91574335e-01 1.40570104e-01 -9.40395772e-01
5.88220537e-01 2.72932261e-01 -1.56723157e-01 -4.53248203e-01
4.67812240e-01 -1.08949184e-01 1.63925868e-02 -1.63816363e-02
1.54359072e-01 -2.51880080e-01 -8.40434253e-01 6.00315928e-01
-1.14928268e-01 -1.54745236e-01 -4.91944909e-01 4.85622078e-01
1.55105972e+00 -4.72191304e-01 -5.78381419e-02 -2.18544856e-01
7.57680893e-01 -8.06932617e-03 2.92160451e-01 1.05826914e+00
-2.11882472e-01 2.38109365e-01 6.68451428e-01 -4.81288224e-01
-9.00984108e-01 -1.08712673e+00 -1.03631793e-02 1.64911723e+00
-2.69243449e-01 -3.62157613e-01 -1.16370714e+00 -1.49321735e+00
3.41703713e-01 1.23301637e+00 -8.65917206e-01 -1.02815187e+00
-5.97104907e-01 -6.92362130e-01 1.55339706e+00 5.56201279e-01
5.23322999e-01 -7.46815622e-01 1.02339663e-01 -3.53731364e-01
5.77202849e-02 -1.18089247e+00 -6.59444630e-01 3.95564169e-01
-6.28015995e-01 -7.78253257e-01 -3.17060471e-01 -5.08919835e-01
5.85664332e-01 -1.48928557e-02 9.77347255e-01 5.15783243e-02
6.98897392e-02 1.99089333e-01 -4.03604239e-01 -6.26114786e-01
-1.07940185e+00 5.42854249e-01 1.01565436e-01 -1.74670279e-01
1.28710836e-01 -3.63020122e-01 -2.05682427e-01 2.49761105e-01
-1.18003118e+00 -5.23793578e-01 3.14943433e-01 7.29279578e-01
-1.35857150e-01 1.61371306e-02 8.28172207e-01 -1.43737769e+00
7.12357044e-01 -7.20502138e-01 -5.86200297e-01 1.08381115e-01
-4.56700683e-01 2.31496379e-01 1.58252966e+00 -8.38431358e-01
-9.75553215e-01 -2.45922357e-01 -2.71563858e-01 -4.81845617e-01
-2.17659980e-01 -9.62660927e-03 -2.66118169e-01 -2.48230442e-01
1.43107700e+00 -1.57509699e-01 -3.12823683e-01 -2.51200169e-01
6.28593743e-01 6.20642185e-01 5.47185302e-01 -6.23403192e-01
1.40455616e+00 4.52208519e-01 -1.33400336e-01 -1.64241701e-01
-1.18538785e+00 2.90417403e-01 -3.33598465e-01 4.28710014e-01
8.61353457e-01 -7.58633077e-01 -6.86523736e-01 6.05885327e-01
-8.88441384e-01 -5.93759894e-01 -1.86783895e-01 -2.37839930e-02
-1.97303697e-01 2.16964975e-01 -9.76316094e-01 -3.16795468e-01
-7.68289924e-01 -1.01685047e+00 7.61586964e-01 -2.46317238e-01
-1.83232322e-01 -1.09378445e+00 -1.61933407e-01 4.46845680e-01
5.27872741e-01 4.70901638e-01 1.15658414e+00 -1.52484453e+00
6.05688728e-02 -5.71197033e-01 1.78273335e-01 6.75685167e-01
3.72434920e-03 -2.64722824e-01 -1.21950102e+00 -4.90391880e-01
2.75876909e-01 -6.10496402e-01 9.55779493e-01 -3.52261037e-01
1.42235458e+00 -8.06552589e-01 -8.28563944e-02 8.10995877e-01
1.02039838e+00 8.46865326e-02 4.85631913e-01 4.89765078e-01
6.83669746e-01 4.33610499e-01 1.70389310e-01 -1.50078997e-01
-2.04805642e-01 3.45047981e-01 6.18402302e-01 -6.88845441e-02
6.54048398e-02 -1.87353596e-01 8.87785256e-01 6.85387850e-02
6.15110755e-01 -6.23332202e-01 -8.34652126e-01 1.46615431e-01
-1.47581768e+00 -9.42640007e-01 5.61060943e-02 1.85003591e+00
1.00728369e+00 5.46005487e-01 -1.54310390e-01 1.30625144e-01
5.91394365e-01 2.17050090e-01 -9.11560774e-01 -1.10432601e+00
-2.04224393e-01 5.33868313e-01 9.75493848e-01 6.08657718e-01
-1.36042774e+00 1.49363053e+00 6.77091122e+00 4.98719990e-01
-1.28501081e+00 2.68093765e-01 7.87308216e-01 -3.65303159e-01
-3.03931564e-01 -2.50359148e-01 -6.47766650e-01 2.88945615e-01
1.31111026e+00 -3.50264728e-01 4.90347862e-01 1.05348432e+00
-3.35805893e-01 7.30468333e-01 -1.16023481e+00 1.16192289e-01
2.07735479e-01 -1.03641641e+00 3.99214536e-01 -1.61545873e-01
7.86486804e-01 -2.30536833e-02 6.08895302e-01 6.32234216e-01
1.08975267e+00 -1.32559454e+00 7.20247865e-01 -1.57233402e-02
6.73411906e-01 -1.16039634e+00 7.17374682e-01 3.81886363e-01
-4.31081504e-01 -4.22722191e-01 -1.67923108e-01 5.78593761e-02
-3.51232558e-01 1.96357653e-01 -7.78013587e-01 3.63796502e-01
6.12933457e-01 1.17995426e-01 -7.97509849e-01 -5.16008809e-02
-4.12811905e-01 8.17287624e-01 -8.86960030e-02 2.99045980e-01
3.27054501e-01 4.93832856e-01 6.49756849e-01 1.23544693e+00
-1.49850890e-01 2.56954342e-01 1.19654305e-01 4.38800156e-01
-7.01651990e-01 -1.83633924e-01 -8.75770092e-01 -1.20551616e-01
3.61780763e-01 1.09475935e+00 -1.81356519e-01 -3.12614441e-01
-1.07259929e-01 1.38116467e+00 3.72355580e-01 2.97446787e-01
-1.10510659e+00 -7.03006089e-01 1.09551620e+00 -1.84045643e-01
8.63099024e-02 6.72309473e-02 -4.54172581e-01 -1.24844527e+00
-2.19925717e-01 -1.49235702e+00 6.16734922e-01 -5.55916309e-01
-1.62048483e+00 9.04507279e-01 -2.38302484e-01 -5.52404940e-01
-2.62430072e-01 -9.05094147e-01 -8.53233516e-01 1.01848817e+00
-1.00724804e+00 -1.12111366e+00 2.22789750e-01 1.01095593e+00
4.60635900e-01 -4.59014535e-01 1.08440280e+00 -7.94972032e-02
-7.92194784e-01 1.43160689e+00 8.39145184e-02 6.54617786e-01
1.00846887e+00 -1.22455299e+00 1.17462265e+00 1.25021374e+00
-2.35410213e-01 6.90242052e-01 8.18611205e-01 -5.26646793e-01
-1.27300549e+00 -1.57077813e+00 5.07352471e-01 -9.53675032e-01
1.06049204e+00 -7.08175540e-01 -8.82969260e-01 1.30487311e+00
1.97820187e-01 5.23674078e-02 7.14329004e-01 -8.79181176e-02
-1.29701400e+00 3.28653492e-02 -1.53978872e+00 1.12100613e+00
1.05315638e+00 -7.36454904e-01 -7.94789553e-01 6.07775986e-01
1.38406467e+00 -5.13864219e-01 -7.39977419e-01 2.68267781e-01
2.05639392e-01 -5.40413260e-01 1.06282759e+00 -1.43115687e+00
5.35477340e-01 1.10478915e-01 -3.47552031e-01 -1.56057322e+00
-1.43740833e-01 -8.63277316e-01 1.52519435e-01 1.18526149e+00
6.37224376e-01 -1.27867222e+00 5.00714362e-01 6.14690542e-01
-7.31497407e-02 -1.42414078e-01 -7.56812096e-01 -7.46833444e-01
1.08097148e+00 -4.70156342e-01 5.70598006e-01 1.22884846e+00
-9.56562757e-02 3.05371672e-01 -1.87383205e-01 6.94039226e-01
3.17932695e-01 -3.38274688e-01 7.69818544e-01 -8.34972203e-01
-3.91136736e-01 -3.13066363e-01 -2.39243224e-01 -5.73922753e-01
8.92430544e-01 -1.27094543e+00 -2.43910924e-01 -6.00654423e-01
-3.50528032e-01 -1.81970969e-01 -2.79887944e-01 1.03167379e+00
-3.02895367e-01 3.62299174e-01 2.91362137e-01 -9.81686115e-02
3.17969210e-02 1.96872547e-01 7.05243647e-01 -4.12912458e-01
1.91034243e-01 1.59599662e-01 -1.43241763e+00 9.11957085e-01
1.15442169e+00 -8.93802226e-01 -4.25413489e-01 -6.88129783e-01
7.84539059e-02 -3.04752618e-01 3.45759004e-01 -7.13302612e-01
-1.19234078e-01 -1.82302184e-02 6.06551409e-01 3.55412334e-01
1.67211056e-01 -7.21248090e-01 -3.91031563e-01 6.48426771e-01
-9.02643561e-01 2.33862564e-01 7.62936294e-01 4.93032724e-01
2.80235559e-01 -2.38087207e-01 1.11617470e+00 -1.30175445e-02
-2.14438722e-03 1.40795052e-01 -6.00280881e-01 3.41363311e-01
1.16474652e+00 5.26893973e-01 -9.92345333e-01 -3.40216488e-01
-8.21044862e-01 9.48597938e-02 3.94804418e-01 6.72548473e-01
1.87246501e-01 -9.38891172e-01 -8.04910898e-01 2.29539648e-01
1.81243755e-02 -5.02056777e-01 -4.19572592e-02 1.96613502e-02
-2.44209647e-01 3.44571024e-01 -3.16744410e-02 -9.99973994e-03
-1.50516903e+00 8.87146473e-01 8.17125261e-01 -3.57532471e-01
-4.21698570e-01 1.01751328e+00 5.48833132e-01 -6.01847410e-01
9.74973366e-02 -1.16185203e-01 1.01472124e-01 -3.65754485e-01
5.75273871e-01 1.24715708e-01 2.81198382e-01 -3.31898004e-01
-3.22439224e-01 1.86842248e-01 -6.05878770e-01 -1.29161283e-01
9.74762380e-01 3.92162412e-01 1.18410096e-01 -7.60970041e-02
1.29996419e+00 5.29000938e-01 -9.52005386e-01 -1.57110229e-01
-1.63097888e-01 -1.63250402e-01 6.89788908e-03 -1.24455976e+00
-1.16625547e+00 6.34222448e-01 2.33541310e-01 5.39990067e-01
1.15150928e+00 -1.92558885e-01 8.72606814e-01 6.97084010e-01
-1.78202316e-02 -5.00828028e-01 1.40192166e-01 8.68399739e-01
9.83666480e-01 -8.86417329e-01 -2.47064933e-01 -2.78592944e-01
-9.16463375e-01 9.59427118e-01 7.56369591e-01 -3.53680700e-01
5.00309229e-01 5.92992246e-01 3.08052927e-01 1.29235193e-01
-8.48196089e-01 3.61822098e-01 1.23652130e-01 6.61119878e-01
6.39968142e-02 -1.21481372e-02 2.88799345e-01 6.94024026e-01
-7.33117282e-01 -7.89798319e-01 5.47518253e-01 8.07878017e-01
-9.64712203e-02 -1.22195506e+00 -6.10256195e-01 5.73371425e-02
-1.19408274e+00 -3.20350826e-01 -9.87523377e-01 7.31521487e-01
-1.28502920e-01 1.29975402e+00 -9.62423533e-02 -6.43220663e-01
5.11758029e-01 3.12748313e-01 7.58708641e-02 -7.37705469e-01
-1.50284564e+00 -6.50437474e-01 2.69281626e-01 -4.69319314e-01
5.10006011e-01 -3.02776366e-01 -1.15073884e+00 -7.84540772e-01
-5.29542863e-02 1.42274916e-01 6.51502788e-01 8.48480284e-01
4.48186070e-01 6.40367627e-01 9.73486960e-01 -2.85851777e-01
-1.20733988e+00 -7.04631209e-01 -1.28651196e-02 5.21160364e-01
4.57337439e-01 5.59399417e-03 -9.39342737e-01 1.09105892e-02] | [6.018228054046631, 8.109905242919922] |
e67a3585-e940-4a9d-aef5-7e4cb563ee39 | natural-language-processing-via-lda-topic | 1909.09551 | null | https://arxiv.org/abs/1909.09551v1 | https://arxiv.org/pdf/1909.09551v1.pdf | Natural Language Processing via LDA Topic Model in Recommendation Systems | Today, Internet is one of the widest available media worldwide. Recommendation systems are increasingly being used in various applications such as movie recommendation, mobile recommendation, article recommendation and etc. Collaborative Filtering (CF) and Content-Based (CB) are Well-known techniques for building recommendation systems. Topic modeling based on LDA, is a powerful technique for semantic mining and perform topic extraction. In the past few years, many articles have been published based on LDA technique for building recommendation systems. In this paper, we present taxonomy of recommendation systems and applications based on LDA. In addition, we utilize LDA and Gibbs sampling algorithms to evaluate ISWC and WWW conference publications in computer science. Our study suggest that the recommendation systems based on LDA could be effective in building smart recommendation system in online communities. | ['Mahdi Rabbani', 'Yongli Wang', 'Hamed Jelodar', 'SeyedValyAllah Ayobi'] | 2019-09-20 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [-4.08401787e-01 -5.90464592e-01 -4.21029389e-01 -1.83133990e-01
-4.09289032e-01 -3.22940379e-01 7.75029778e-01 2.62707800e-01
-1.01337023e-02 7.16691852e-01 6.15019977e-01 -3.32909763e-01
-3.26726884e-01 -1.29228175e+00 1.02197111e-01 -5.31726539e-01
1.24497429e-01 1.94223836e-01 6.94826186e-01 -1.26050457e-01
1.11150503e+00 1.46429688e-01 -1.82356870e+00 4.34679896e-01
9.41598594e-01 7.52009034e-01 4.73199546e-01 2.22066060e-01
-1.03586686e+00 3.17402244e-01 -2.93901563e-01 -2.75633872e-01
-2.81233877e-01 -4.43453610e-01 -5.61864257e-01 2.35048532e-02
-3.40481013e-01 -8.45414028e-02 -4.60889339e-02 8.71548116e-01
4.95925814e-01 5.95365703e-01 9.49994922e-01 -9.96713817e-01
-7.58129060e-01 8.74479234e-01 -8.98441553e-01 1.41144753e-01
5.14570117e-01 -9.22278523e-01 6.51622534e-01 -8.19397867e-01
4.87548143e-01 1.23177540e+00 4.63236243e-01 1.54201478e-01
-3.57472420e-01 -9.80273426e-01 1.16876371e-01 4.77647096e-01
-1.22822225e+00 6.83221444e-02 8.25038671e-01 -4.34380591e-01
8.14288557e-01 1.56035557e-01 8.82812500e-01 4.21661377e-01
3.35995048e-01 5.41185200e-01 1.25884640e+00 -6.05109930e-01
5.46853423e-01 4.97258216e-01 6.63854003e-01 2.54425973e-01
5.50672710e-01 -5.63271165e-01 -5.32254755e-01 -6.89597845e-01
6.52556479e-01 7.35929966e-01 1.70256808e-01 4.87541556e-02
-6.15460634e-01 1.22823048e+00 -6.59968778e-02 5.72293341e-01
-4.09490168e-01 -1.40504941e-01 3.53438884e-01 4.42716703e-02
8.89462411e-01 1.63141247e-02 -5.63122379e-03 -6.30134065e-03
-9.56585348e-01 1.82674304e-01 1.10667396e+00 9.61672187e-01
4.83485460e-01 -3.26606750e-01 2.03892902e-01 1.02664614e+00
1.17056286e+00 3.60786021e-01 1.01290035e+00 -7.07991362e-01
-1.91107780e-01 5.80157220e-01 -4.07050690e-03 -1.10574603e+00
3.05254590e-02 -1.53313458e-01 -4.97478068e-01 -4.26575720e-01
-2.30301023e-01 -2.65355200e-01 -5.66720486e-01 5.76027334e-01
4.59071070e-01 4.66580540e-01 -1.64462596e-01 3.68832916e-01
1.10555899e+00 9.61085200e-01 4.96778458e-01 -5.27808189e-01
1.47606969e+00 -6.32339120e-01 -1.02540112e+00 5.60356975e-01
4.46769178e-01 -1.31777430e+00 6.78556561e-01 8.87314916e-01
-6.75246298e-01 -2.14423895e-01 -6.66584373e-01 2.44121581e-01
-6.72876894e-01 -6.68881238e-02 1.10368764e+00 1.18095839e+00
-8.19194019e-01 5.32652855e-01 -7.06594944e-01 -1.21111071e+00
3.80024910e-01 1.09907016e-02 3.01124364e-01 -2.69158930e-01
-9.91732836e-01 5.72422504e-01 9.14406925e-02 -7.12813854e-01
-4.69991028e-01 -3.02179217e-01 -2.44916026e-02 3.16388831e-02
9.88509059e-02 -2.63146132e-01 8.22108269e-01 -4.16803360e-01
-1.70779634e+00 4.41676050e-01 -2.13837206e-01 -3.11296880e-01
-3.69561493e-01 -3.68770540e-01 -8.95872295e-01 -1.70131102e-01
-7.49176145e-02 -1.29266560e-01 3.63430321e-01 -8.69264662e-01
-1.10183561e+00 -5.99156082e-01 -2.21845761e-01 2.74483263e-01
-7.30242908e-01 6.43308461e-01 -3.77925158e-01 -5.22966743e-01
3.90757531e-01 -7.11694002e-01 2.03170739e-02 -6.99632108e-01
-1.55113265e-01 -8.36753905e-01 1.07835066e+00 -4.93791074e-01
1.58106363e+00 -1.87057948e+00 -5.44214845e-01 6.83104813e-01
-2.22793534e-01 -7.76377097e-02 6.23037815e-01 9.51799929e-01
7.16935098e-01 4.91176188e-01 5.15056610e-01 2.45573834e-01
-2.13000402e-01 -1.40471905e-01 -3.79353613e-01 3.39091450e-01
-1.22438812e+00 5.18433042e-02 -7.14623153e-01 -6.58387601e-01
2.94749916e-01 4.65322614e-01 -5.83523035e-01 1.44891432e-02
-2.81272084e-02 6.15002774e-02 -1.08008361e+00 2.30625808e-01
4.45309669e-01 -2.80037224e-01 4.98657018e-01 -5.68510313e-03
-4.38082457e-01 4.28276122e-01 -1.16608191e+00 1.64167619e+00
-4.04893339e-01 4.24793363e-01 -5.30043006e-01 -8.19241464e-01
1.31379414e+00 6.14971638e-01 5.87470114e-01 -5.45413680e-02
3.14059705e-01 1.74315080e-01 -1.70489460e-01 -5.76013923e-01
6.05394602e-01 2.26058468e-01 5.16038954e-01 8.84528995e-01
7.52065778e-02 -1.02117695e-01 2.13384062e-01 4.38938588e-01
5.71563303e-01 1.99687988e-01 5.32452226e-01 -6.90328598e-01
5.30364931e-01 -8.43473303e-04 1.84431285e-01 4.41910237e-01
1.69039473e-01 1.11986704e-01 -2.59086847e-01 -2.66466320e-01
-8.85918319e-01 -5.21150708e-01 -2.56257176e-01 1.33542049e+00
7.73147270e-02 -7.23805308e-01 -6.22847080e-01 -3.45012784e-01
-1.47948876e-01 8.43888998e-01 -5.22220209e-02 3.18068713e-01
-1.29462443e-02 -7.29323924e-01 -4.96738218e-02 1.18892148e-01
7.81199157e-01 -1.05366015e+00 1.75207973e-01 4.30752009e-01
-4.10673544e-02 -3.47232372e-01 -1.40705258e-01 -5.57036877e-01
-1.43982339e+00 -7.86477506e-01 -7.55595922e-01 -7.73837805e-01
5.40448785e-01 9.62765455e-01 7.28820205e-01 3.41826499e-01
3.50995287e-02 6.16754055e-01 -1.05756819e+00 -5.33814847e-01
2.27426827e-01 1.85194582e-01 4.45053913e-02 -2.76449859e-01
1.02556872e+00 -8.03967118e-01 -7.89791226e-01 2.68914431e-01
-5.54728389e-01 -2.10074872e-01 2.13308148e-02 3.02605063e-01
3.54098558e-01 4.74335492e-01 9.19071734e-01 -1.53345954e+00
9.88622308e-01 -1.13443482e+00 -2.90032804e-01 1.23086773e-01
-1.11428988e+00 -4.90348577e-01 4.65659499e-02 -1.13529712e-01
-1.58330953e+00 -4.79272693e-01 -1.10678025e-01 5.00112057e-01
-4.67110217e-01 8.02728951e-01 1.28503859e-01 1.04577050e-01
6.27985120e-01 8.09777677e-02 -4.77509856e-01 -1.10991251e+00
3.63333613e-01 1.53883183e+00 -4.55552161e-01 -5.17780602e-01
7.23875244e-04 5.40923595e-01 -4.17652339e-01 -9.24654067e-01
-7.53306389e-01 -1.19736445e+00 -4.27125320e-02 -4.84196156e-01
7.63766527e-01 -7.05370784e-01 -4.65451390e-01 8.93805772e-02
-7.50995636e-01 3.59183669e-01 3.07476699e-01 1.00315046e+00
-1.48160592e-01 4.06919420e-01 -4.37513709e-01 -1.13083589e+00
-6.27342165e-01 -5.91115892e-01 3.22372705e-01 6.98867083e-01
-1.53525814e-01 -1.21485662e+00 2.26502240e-01 4.44521219e-01
6.10586882e-01 -4.12345171e-01 7.57331491e-01 -1.00680649e+00
-2.60642678e-01 -1.81677297e-01 -5.53164408e-02 -5.04017137e-02
3.32721919e-01 2.06280261e-01 -6.19670630e-01 -1.36340586e-02
-6.82551935e-02 2.91680098e-01 5.66887558e-01 5.71596682e-01
1.26736236e+00 -2.63391316e-01 -9.19643342e-01 -2.78607961e-02
1.51020980e+00 8.40739429e-01 9.27804530e-01 4.45017189e-01
4.00769919e-01 4.46007550e-01 6.77977622e-01 7.76754975e-01
3.25931191e-01 2.04736739e-01 -4.07872498e-01 6.77313685e-01
6.47249371e-02 -2.84875900e-01 1.53859302e-01 1.34120715e+00
-5.55286467e-01 -3.69307101e-01 -9.14252043e-01 4.42739397e-01
-1.94983888e+00 -1.08491826e+00 -3.93647909e-01 2.28580809e+00
6.90418124e-01 -2.33631372e-01 2.25636423e-01 1.18585505e-01
8.97901416e-01 -3.96710038e-01 3.71361822e-02 -4.19077188e-01
3.50944221e-01 3.04854929e-01 4.38715756e-01 7.63566867e-02
-9.09437716e-01 9.53952670e-01 5.83536005e+00 1.19335902e+00
-9.69933391e-01 5.99252522e-01 4.14876610e-01 1.04588792e-01
-4.42140907e-01 3.34942102e-01 -1.12297726e+00 7.42732823e-01
8.67058814e-01 -6.26138568e-01 1.87795728e-01 1.21025801e+00
4.00313735e-01 -5.62140882e-01 -2.38345250e-01 7.82981992e-01
2.32959479e-01 -1.50377929e+00 1.30581215e-01 2.67027348e-01
1.33428800e+00 -2.37211555e-01 -1.13056719e-01 5.59050925e-02
7.38031149e-01 -4.51880872e-01 -1.25069559e-01 5.91051280e-01
1.58926696e-01 -9.65425134e-01 9.94624436e-01 2.11652666e-01
-9.32888150e-01 1.91218197e-01 -7.79776990e-01 -1.19301155e-01
2.51637012e-01 9.24678266e-01 -4.66085941e-01 4.66920316e-01
9.89722192e-01 1.00504410e+00 1.56383291e-01 1.79194534e+00
-6.45933449e-02 1.23935747e+00 -3.05394858e-01 -6.74934566e-01
-6.49422333e-02 -7.32959151e-01 1.65651485e-01 1.16114044e+00
8.54962111e-01 5.68092287e-01 3.15786093e-01 6.60503432e-02
1.40199125e-01 1.09891784e+00 -4.00484860e-01 -2.05739345e-02
1.00905001e+00 1.17470193e+00 -1.35161233e+00 -6.03817284e-01
-5.59683740e-01 3.39114279e-01 -3.87608796e-01 1.65407732e-01
-2.87094742e-01 -4.75270361e-01 6.54380023e-02 3.42884630e-01
2.22802088e-01 -2.87255466e-01 -2.13951841e-01 -9.03642178e-01
-1.00659513e+00 -4.77872044e-01 4.72274989e-01 -4.78820115e-01
-1.38386381e+00 1.28520012e-01 1.31511107e-01 -1.16439807e+00
1.81919903e-01 -1.44687846e-01 -8.34994197e-01 7.77972698e-01
-9.94154215e-01 -9.25379574e-01 -1.89650908e-01 7.72568822e-01
7.74487853e-01 -6.10506415e-01 8.45974863e-01 5.78074634e-01
-2.09847614e-01 -1.54074386e-01 8.82979929e-01 -5.23703694e-01
8.14278185e-01 -8.87923300e-01 -3.34352970e-01 3.60492557e-01
1.90956414e-01 1.10249698e+00 4.58928734e-01 -9.84914124e-01
-1.30517423e+00 -5.87279618e-01 8.16857696e-01 -1.85455643e-02
5.60731113e-01 2.74271399e-01 -5.86191654e-01 4.32069719e-01
3.85786176e-01 -8.41935754e-01 1.37336934e+00 7.19299912e-01
8.61685351e-02 -9.59147662e-02 -1.12269819e+00 4.55448329e-01
6.19312823e-01 -2.81810284e-01 -6.14950895e-01 6.96739316e-01
2.58968562e-01 2.50266314e-01 -1.01732111e+00 -3.05279285e-01
7.23979771e-01 -7.60033727e-01 8.28550696e-01 -5.40043950e-01
1.82489827e-01 -3.05488139e-01 -1.20527402e-01 -1.13170183e+00
-5.97274303e-01 -2.81352967e-01 -6.73553869e-02 1.72567654e+00
2.65985817e-01 -5.13820112e-01 8.93853545e-01 6.06067121e-01
1.80040464e-01 -2.53526419e-01 -2.52535909e-01 -2.79436439e-01
-1.70422941e-01 -3.22910368e-01 5.11912942e-01 9.34648037e-01
5.32656908e-01 1.72713235e-01 -3.57594043e-01 -4.63622659e-01
5.31505585e-01 1.74898982e-01 6.25457942e-01 -1.67383671e+00
-3.19895782e-02 -2.33959079e-01 -1.29272919e-02 -8.50490093e-01
-4.00422007e-01 -6.15863085e-01 -3.05487871e-01 -1.90391409e+00
5.95993698e-01 -8.34290922e-01 -4.01599020e-01 -4.91437353e-02
6.73347488e-02 2.51855075e-01 -2.26072997e-01 7.95112252e-01
-7.02171385e-01 3.74449380e-02 1.04490983e+00 3.77875775e-01
-3.07399631e-01 2.45163918e-01 -7.44993687e-01 9.32283640e-01
9.44929957e-01 -6.72599316e-01 -5.15005827e-01 2.41377085e-01
5.68684936e-01 -4.68302637e-01 -7.21094847e-01 -5.40988624e-01
5.39226174e-01 -4.02494013e-01 1.39914319e-01 -8.05685997e-01
-1.51150644e-01 -7.74856687e-01 3.91265631e-01 2.90043280e-02
-3.89339477e-01 -5.62116086e-01 -2.93679416e-01 6.92491531e-01
-7.54871443e-02 -6.22674465e-01 2.60842234e-01 -2.53126204e-01
-8.87786090e-01 2.24912092e-01 -8.44445467e-01 -5.42177737e-01
8.82765353e-01 -5.22616617e-02 -1.26457289e-01 -6.66002035e-01
-7.09583163e-01 -1.73830047e-01 6.95367381e-02 2.49489069e-01
4.45292681e-01 -1.11101520e+00 -4.28057432e-01 -4.93947059e-01
-6.23362176e-02 -7.59637535e-01 4.42236394e-01 5.06789744e-01
-6.45587206e-01 5.91202259e-01 -2.08086193e-01 2.19268218e-01
-1.29911423e+00 3.97787750e-01 -5.69484591e-01 -1.19837478e-01
-4.45234299e-01 7.24752903e-01 -4.49498594e-02 5.05778119e-02
2.42054686e-01 3.19886267e-01 -1.25171304e+00 2.79667154e-02
7.63596237e-01 9.15208936e-01 -1.99649781e-01 -1.74694911e-01
-9.60820690e-02 5.45874715e-01 -2.51823813e-01 -3.88043106e-01
1.26642525e+00 -5.87589622e-01 -4.05478895e-01 5.57528138e-01
6.24791443e-01 2.87625104e-01 -1.04991622e-01 -3.60113263e-01
1.64467424e-01 -4.82879788e-01 6.84647799e-01 -7.38698483e-01
-8.42295170e-01 7.10703969e-01 6.55390441e-01 6.72629237e-01
7.35783815e-01 -2.13818356e-01 7.18660057e-01 4.59579557e-01
7.22196341e-01 -1.48773527e+00 -3.74124616e-01 2.44897157e-01
4.08033341e-01 -1.07691109e+00 5.17331660e-01 -5.30847788e-01
-4.38968062e-01 1.21519518e+00 1.49187699e-01 -4.37809855e-01
1.77813292e+00 -6.75618649e-02 -2.95680612e-01 -1.79051831e-01
-5.58757603e-01 -1.49432048e-01 3.69595826e-01 4.66048568e-01
1.39188969e+00 2.53991485e-01 -1.31473327e+00 9.86576438e-01
-6.49442375e-02 4.38580066e-01 3.67591947e-01 9.53511894e-01
-1.30862486e+00 -1.45046008e+00 -4.70501453e-01 1.10394740e+00
-9.10097361e-01 -2.53931165e-01 -8.29689279e-02 1.35814324e-01
1.27374589e-01 1.41791475e+00 1.03169508e-01 -3.22988570e-01
-4.78484809e-01 1.38656124e-01 9.31678787e-02 -9.67506647e-01
-4.13144827e-01 4.35257941e-01 2.68335849e-01 -1.57413095e-01
-7.50934184e-01 -7.11297572e-01 -1.14420164e+00 -6.24707878e-01
-9.44769979e-01 9.56382096e-01 1.30793190e+00 9.30064380e-01
4.23003465e-01 1.72536746e-01 4.70996141e-01 -4.05015171e-01
3.88882011e-01 -1.24599886e+00 -1.11484194e+00 2.48009488e-02
-9.01241601e-01 -8.24213386e-01 -1.06782332e-01 2.75302470e-01] | [10.164551734924316, 6.021831512451172] |
912dc482-497b-4804-8413-d0fb67f9736a | estimating-blink-probability-for-highlight | 2007.01089 | null | https://arxiv.org/abs/2007.01089v1 | https://arxiv.org/pdf/2007.01089v1.pdf | Estimating Blink Probability for Highlight Detection in Figure Skating Videos | Highlight detection in sports videos has a broad viewership and huge commercial potential. It is thus imperative to detect highlight scenes more suitably for human interest with high temporal accuracy. Since people instinctively suppress blinks during attention-grabbing events and synchronously generate blinks at attention break points in videos, the instantaneous blink rate can be utilized as a highly accurate temporal indicator of human interest. Therefore, in this study, we propose a novel, automatic highlight detection method based on the blink rate. The method trains a one-dimensional convolution network (1D-CNN) to assess blink rates at each video frame from the spatio-temporal pose features of figure skating videos. Experiments show that the method successfully estimates the blink rate in 94% of the video clips and predicts the temporal change in the blink rate around a jump event with high accuracy. Moreover, the method detects not only the representative athletic action, but also the distinctive artistic expression of figure skating performance as key frames. This suggests that the blink-rate-based supervised learning approach enables high-accuracy highlight detection that more closely matches human sensibility. | ['Tamami Nakano', 'Akihiro Kishimoto', 'Atsuya Sakata'] | 2020-07-02 | null | null | null | null | ['highlight-detection'] | ['computer-vision'] | [ 7.48173967e-02 -7.01721311e-01 -3.06780517e-01 6.38810918e-02
-4.85665023e-01 -3.58350307e-01 2.09060729e-01 3.43189865e-01
-4.77947176e-01 5.36026001e-01 2.05947116e-01 3.50837052e-01
2.51844198e-01 -4.40669000e-01 -6.12258911e-01 -6.96119487e-01
-2.16024458e-01 -6.44925237e-01 4.17959183e-01 -1.56327039e-01
6.23627126e-01 6.48140788e-01 -1.84964800e+00 3.94511610e-01
5.75299859e-01 1.30253708e+00 -6.77360571e-04 1.03432131e+00
2.90213525e-01 1.26908350e+00 -1.11377704e+00 -7.88779929e-02
5.48417009e-02 -7.07631409e-01 -2.13130433e-02 -3.14660817e-02
8.57482910e-01 -4.47590262e-01 -6.03632569e-01 8.38560939e-01
7.49960661e-01 3.13081801e-01 1.74108118e-01 -1.35893381e+00
-2.25998312e-01 1.63259283e-01 -8.00824106e-01 1.25266182e+00
4.77833122e-01 3.62064391e-01 6.07336760e-01 -6.32552326e-01
5.58097839e-01 7.17115879e-01 6.18080795e-01 1.91728815e-01
-8.33478987e-01 -9.92538631e-01 -5.05945943e-02 8.10622275e-01
-1.17207181e+00 -2.19496533e-01 1.29995191e+00 -4.54216123e-01
4.56729949e-01 6.14644289e-01 1.39381409e+00 1.08220458e+00
4.41937804e-01 1.07864499e+00 9.76124346e-01 -2.54744232e-01
6.30344916e-03 -1.32505029e-01 -2.26334378e-01 4.34459239e-01
2.31203157e-02 3.60258147e-02 -1.29358506e+00 3.12242448e-01
9.47110653e-01 1.31818101e-01 -4.60829467e-01 4.06217799e-02
-1.42703414e+00 4.20996964e-01 4.17634398e-01 2.54792333e-01
-3.29924673e-01 4.04991478e-01 7.44744539e-01 6.36501759e-02
7.40890384e-01 5.90862691e-01 8.83857980e-02 -7.11295605e-01
-1.24200404e+00 4.09346581e-01 5.80540225e-02 3.47564191e-01
2.09542960e-01 2.88713574e-01 -5.82087398e-01 6.47351980e-01
-5.29677570e-01 5.53858280e-01 3.53441179e-01 -7.72017837e-01
8.93107131e-02 5.07348001e-01 1.82392508e-01 -1.52165091e+00
-4.95989174e-01 -5.08715630e-01 -5.17084301e-01 3.81491959e-01
5.19879520e-01 2.69729830e-02 -2.51885474e-01 1.54545903e+00
3.93771559e-01 2.92588890e-01 -5.47183752e-01 1.43233287e+00
8.60173047e-01 6.34496570e-01 2.37751529e-01 -4.46100503e-01
1.48923218e+00 -5.93157172e-01 -1.06771541e+00 2.93962248e-02
1.29825875e-01 -7.86736190e-01 1.17442191e+00 5.22723436e-01
-1.12801778e+00 -8.69410157e-01 -1.14952958e+00 -8.56742188e-02
-4.03011777e-03 4.33185190e-01 5.70642412e-01 3.67008805e-01
-3.58786136e-01 4.78779703e-01 -6.37336612e-01 -9.18099359e-02
2.66002387e-01 -3.52285117e-01 -9.44581702e-02 4.31086779e-01
-1.22483742e+00 6.83896244e-01 2.38447383e-01 3.83394629e-01
-8.69324625e-01 -8.86716127e-01 -7.34614789e-01 -4.78250422e-02
4.27804440e-01 -7.32443109e-02 1.06373382e+00 -1.33704865e+00
-1.29408216e+00 1.08648646e+00 -1.44986093e-01 -5.89649379e-01
7.61552215e-01 -6.78717494e-01 -6.81354523e-01 5.88881016e-01
-6.92313490e-03 4.21660662e-01 1.15453815e+00 -8.23751271e-01
-8.02336514e-01 9.38701481e-02 1.38119504e-01 2.23535627e-01
-4.37024742e-01 4.30153549e-01 -2.43581429e-01 -8.76582444e-01
-2.40979031e-01 -5.99267781e-01 5.21999955e-01 4.23020661e-01
-3.56209189e-01 -1.39919639e-01 8.24282467e-01 -7.62828827e-01
1.68856525e+00 -2.42792273e+00 -3.59506339e-01 -2.28626624e-01
6.55381382e-01 3.26449603e-01 2.68828511e-01 2.66897321e-01
-4.32255678e-02 -4.27119493e-01 2.66002148e-01 3.86461556e-01
-3.42813373e-01 -4.59323287e-01 -2.78226644e-01 8.69219005e-01
2.77797312e-01 8.47954571e-01 -1.24871850e+00 -6.25010550e-01
3.84663492e-01 4.96695817e-01 -1.85173959e-01 3.02499890e-01
1.85587257e-01 3.49474043e-01 -1.93229571e-01 8.15369844e-01
5.05211234e-01 2.51290292e-01 -3.58565152e-01 -5.30418873e-01
-6.05097413e-01 -1.82533398e-01 -8.37243855e-01 1.58042943e+00
-2.41496973e-02 1.78100848e+00 -3.10632646e-01 -6.91960096e-01
1.13369071e+00 5.70086464e-02 5.79899728e-01 -1.20547521e+00
9.30819660e-02 -2.07915641e-02 -1.70194611e-01 -1.06393385e+00
6.99609697e-01 3.24551165e-02 -6.95983618e-02 1.44443005e-01
-4.54219460e-01 2.55445898e-01 3.45501035e-01 6.32957295e-02
8.89427841e-01 2.27373749e-01 1.59196213e-01 -1.95441425e-01
2.81403333e-01 -1.20087579e-01 5.32025874e-01 6.71401441e-01
-6.93522096e-01 6.14367425e-01 5.82993567e-01 -8.77412200e-01
-8.67384136e-01 -1.02536118e+00 2.18706980e-01 1.25913775e+00
5.16758561e-01 -3.17096800e-01 -7.48378515e-01 -2.31451154e-01
-2.51422841e-02 4.91656125e-01 -1.03454733e+00 -6.11412168e-01
-8.57685506e-01 -3.05069894e-01 5.52817881e-01 5.16211390e-01
6.67888761e-01 -1.39833975e+00 -1.46261144e+00 1.18064776e-01
-5.10221541e-01 -9.55757797e-01 -7.64390767e-01 -1.70623809e-01
-3.85903060e-01 -1.39423776e+00 -9.74487722e-01 -6.69058323e-01
2.32062876e-01 4.38520998e-01 9.10615146e-01 -1.41760841e-01
-8.70899081e-01 1.86931312e-01 -2.82854915e-01 -6.44267619e-01
-5.80253452e-02 -3.76080304e-01 -7.11992234e-02 3.56116652e-01
5.22186518e-01 -2.30235070e-01 -1.08069396e+00 1.21919058e-01
-6.69367254e-01 2.21167564e-01 2.45460972e-01 3.82417023e-01
5.32294810e-01 -1.15993554e-02 3.09963763e-01 -1.35742381e-01
6.58722341e-01 -6.26895204e-02 -2.74124742e-01 1.41264275e-02
-6.20834418e-02 -6.64057136e-01 2.67939776e-01 -1.03401792e+00
-9.42168415e-01 -4.25514609e-01 4.12133843e-01 -7.05707490e-01
-1.80872887e-01 5.30823916e-02 2.80183256e-01 7.50150979e-02
9.40634549e-01 2.50217021e-01 -3.62797864e-02 -1.18433073e-01
2.14441475e-02 3.47772628e-01 1.14381194e+00 -1.47976413e-01
3.37172270e-01 6.60923541e-01 8.67740661e-02 -9.32552338e-01
-1.05921340e+00 -7.95145094e-01 -3.71641219e-01 -1.50474787e+00
7.99421668e-01 -1.16330373e+00 -1.25570285e+00 8.60409677e-01
-1.02832437e+00 -2.92924881e-01 -3.24874312e-01 6.43115342e-01
-4.46635306e-01 3.26520324e-01 -6.77230239e-01 -9.68523502e-01
-5.09121299e-01 -5.01772404e-01 9.39767897e-01 6.45623267e-01
-3.72507304e-01 -4.66113508e-01 7.90602937e-02 1.03510559e-01
3.26695025e-01 1.14149761e+00 2.84554929e-01 -1.90314278e-02
-3.87426108e-01 -4.66511369e-01 -3.09786230e-01 1.10375144e-01
5.77382185e-02 2.75566608e-01 -9.74519789e-01 -1.74819455e-02
-1.46757469e-01 -1.95909753e-01 6.96279883e-01 8.48793924e-01
1.20968282e+00 2.57355487e-03 7.41335079e-02 5.12403607e-01
1.12444961e+00 1.46421656e-01 6.65850878e-01 6.60191476e-01
6.85374975e-01 5.17037094e-01 1.02405560e+00 7.22312152e-01
-2.66388625e-01 7.87488222e-01 3.51842612e-01 -5.25553942e-01
-4.03060913e-01 -2.95969844e-01 4.90587294e-01 4.83875632e-01
-5.09590507e-01 -9.24935341e-02 -6.10727668e-01 5.09122550e-01
-1.62583768e+00 -1.48247302e+00 -5.30028820e-01 2.23066282e+00
8.59800398e-01 3.56993109e-01 5.62691212e-01 2.38480091e-01
1.02451599e+00 3.33331168e-01 -7.82728195e-01 -4.20414388e-01
-3.86209995e-01 -2.41009761e-02 4.70659912e-01 -1.58217683e-01
-1.14666069e+00 4.92780268e-01 6.31527901e+00 8.48805606e-01
-1.49751031e+00 -1.17887510e-03 7.47167230e-01 -8.55050325e-01
2.20129743e-01 -6.08453929e-01 -3.25021058e-01 7.95749605e-01
6.92677379e-01 -1.99342102e-01 2.38620657e-02 7.78687060e-01
9.68434572e-01 -6.02693498e-01 -7.20772624e-01 1.38490415e+00
5.31176329e-01 -1.40903139e+00 -3.83191645e-01 -3.89712304e-01
5.05549490e-01 -6.91168964e-01 1.98938653e-01 -7.69400075e-02
-6.53174520e-01 -5.34362435e-01 1.17394495e+00 8.93555105e-01
1.02059984e+00 -1.02183902e+00 2.36931443e-01 -1.08635180e-01
-1.26068676e+00 -8.69406983e-02 -1.14643626e-01 -3.73702973e-01
2.42697656e-01 6.55367732e-01 -3.76248449e-01 -1.78905457e-01
1.01666868e+00 1.08538938e+00 -4.99831706e-01 1.53645635e+00
-4.40818369e-01 6.96389675e-01 -3.81197082e-03 -1.98984474e-01
-1.01086915e-01 1.83438301e-01 7.56614923e-01 1.69219732e+00
2.19302639e-01 1.50882322e-02 -1.75351664e-01 8.31441522e-01
2.22236171e-01 8.26640725e-02 -2.74963289e-01 -7.28001744e-02
1.12131976e-01 1.33984423e+00 -8.59747052e-01 -2.73798317e-01
-7.80876279e-02 9.06128705e-01 4.83523589e-03 2.42333755e-01
-1.31047916e+00 -5.85736215e-01 7.16559172e-01 3.51412952e-01
-2.11656857e-02 -1.39817894e-01 7.13756867e-03 -9.34279203e-01
1.45773157e-01 -6.62631869e-01 2.53992856e-01 -1.18509150e+00
-9.06835675e-01 3.07590187e-01 -2.07093552e-01 -1.61173272e+00
9.31779891e-02 -2.80846298e-01 -9.68880415e-01 4.74032313e-01
-1.49507833e+00 -8.38236213e-01 -9.59544897e-01 5.74894607e-01
8.35321486e-01 3.96448523e-01 1.99682593e-01 3.25672269e-01
-7.62664080e-01 5.12838304e-01 -2.64747322e-01 2.61208624e-01
1.07030976e+00 -1.07597613e+00 -5.94801679e-02 1.04018736e+00
-4.66207713e-02 2.40411200e-02 1.25389445e+00 -5.60492396e-01
-1.13908601e+00 -8.02101254e-01 6.74556673e-01 -2.45248362e-01
6.75800741e-01 -2.36359537e-01 -7.66509354e-01 6.88651204e-02
3.80937271e-02 -6.87271878e-02 6.29733503e-01 -1.86890632e-01
-4.24186215e-02 -4.00685757e-01 -5.39571047e-01 5.04881620e-01
7.57852137e-01 -6.81605518e-01 -5.56931078e-01 2.67985314e-01
1.78849220e-01 -6.31233573e-01 -6.47440314e-01 1.96459427e-01
1.01614904e+00 -1.19549084e+00 9.52171922e-01 -2.77106196e-01
7.06192076e-01 -1.68266281e-01 6.19635224e-01 -9.47299719e-01
-2.46258110e-01 -6.77895784e-01 -3.35534751e-01 1.10270560e+00
-3.18356544e-01 2.81316694e-02 6.43730521e-01 4.01416898e-01
-1.14011288e-01 -4.09731179e-01 -8.75129461e-01 -6.92734540e-01
-7.39906132e-01 -4.11436647e-01 9.53629762e-02 7.98822761e-01
2.05747500e-01 -2.28304595e-01 -8.40104461e-01 -1.33931205e-01
5.76461136e-01 2.93226838e-01 8.75483155e-01 -9.56200063e-01
3.06146555e-02 -5.81339478e-01 -7.10096836e-01 -7.76124060e-01
-3.63245934e-01 -7.40725398e-02 9.92761478e-02 -8.81938219e-01
3.88698131e-01 2.92807817e-01 -6.53244555e-01 1.37077615e-01
-4.50101078e-01 6.78449154e-01 1.16335072e-01 8.32867771e-02
-8.00478578e-01 2.14583665e-01 1.41943991e+00 -1.49335012e-01
-4.36241686e-01 -1.61139920e-01 -2.03864738e-01 6.34505153e-01
6.42224669e-01 -1.22785665e-01 -1.20598495e-01 9.92378220e-02
4.31458771e-01 4.23090793e-02 7.52248585e-01 -1.29991567e+00
1.34517163e-01 -2.82727510e-01 7.32271552e-01 -6.68855309e-01
3.64841729e-01 -2.51328081e-01 8.23514815e-03 5.27470708e-01
-6.39633834e-01 9.05666873e-02 4.70803231e-01 6.08141005e-01
-3.88363332e-01 1.76779717e-01 7.11671114e-01 -1.57405324e-02
-1.13804471e+00 -6.56852452e-03 -5.49157143e-01 -3.03005017e-02
1.10807717e+00 -6.13364637e-01 -5.12179732e-01 -4.16115880e-01
-4.88318950e-01 -1.78263605e-01 2.60654956e-01 5.85796773e-01
7.58587539e-01 -1.27593338e+00 -7.35058248e-01 1.59430444e-01
4.10932571e-01 -6.49172544e-01 8.34453940e-01 1.15330029e+00
-6.81924880e-01 -1.33579616e-02 -7.01015055e-01 -6.16445065e-01
-1.41007304e+00 6.80381894e-01 2.86575913e-01 3.35255057e-01
-9.62497830e-01 9.51974988e-01 2.11471111e-01 1.23670566e+00
2.26881683e-01 -3.65550160e-01 -5.76397598e-01 4.20076638e-01
1.12359154e+00 7.00440884e-01 -3.70548457e-01 -6.80239141e-01
-1.09379716e-01 4.62208152e-01 2.28081211e-01 3.63519788e-01
9.79273200e-01 -2.86644429e-01 2.69640565e-01 9.28882062e-01
9.36758697e-01 2.11138591e-01 -1.82862329e+00 1.11910433e-01
-4.45294887e-01 -8.70200872e-01 1.85317546e-01 -5.56411564e-01
-1.14701724e+00 9.34047222e-01 7.64777899e-01 3.63693833e-01
1.43319988e+00 -2.86748230e-01 8.98381174e-01 -2.61345506e-01
-3.03331893e-02 -1.35626483e+00 5.89986444e-01 -3.24076951e-01
1.02125275e+00 -9.31609452e-01 9.37578455e-02 -2.37775341e-01
-6.31286621e-01 1.48538339e+00 7.41977811e-01 -2.54126877e-01
5.47632761e-02 -2.02543605e-02 1.34110123e-01 -3.82053912e-01
-4.07448083e-01 -2.32540444e-01 4.87115145e-01 5.67309260e-01
2.36308128e-01 8.28912184e-02 -3.17229658e-01 1.58257753e-01
-1.48691848e-01 -9.03518870e-03 4.72743779e-01 7.41900921e-01
-6.59270465e-01 -4.19430993e-02 -4.25583124e-01 1.97910041e-01
-5.67086816e-01 1.76343828e-01 -3.16241503e-01 8.59822750e-01
3.56518000e-01 8.58007014e-01 3.25468957e-01 -2.40543276e-01
5.40871739e-01 -1.58989266e-01 4.24159735e-01 5.88081330e-02
-6.58556163e-01 2.42540821e-01 -3.46123502e-02 -8.62065256e-01
-7.15866864e-01 -6.34370685e-01 -8.51039469e-01 -4.38857913e-01
-2.04762205e-01 -1.82849124e-01 4.09367651e-01 5.96313059e-01
1.02860026e-01 7.75366664e-01 5.70818722e-01 -9.97733474e-01
3.60406220e-01 -9.64933395e-01 -1.00825608e+00 9.02579725e-01
4.85599905e-01 -9.35548723e-01 -3.87111425e-01 3.59206408e-01] | [10.096881866455078, 0.37418416142463684] |
70d1c349-8fa9-4c6a-a98f-d8696b3d5859 | mit-lab-at-semeval-2017-task-4-an-integrated | null | null | https://aclanthology.org/S17-2114 | https://aclanthology.org/S17-2114.pdf | MI\&T Lab at SemEval-2017 task 4: An Integrated Training Method of Word Vector for Sentiment Classification | A CNN method for sentiment classification task in Task 4A of SemEval 2017 is presented. To solve the problem of word2vec training word vector slowly, a method of training word vector by integrating word2vec and Convolutional Neural Network (CNN) is proposed. This training method not only improves the training speed of word2vec, but also makes the word vector more effective for the target task. Furthermore, the word2vec adopts a full connection between the input layer and the projection layer of the Continuous Bag-of-Words (CBOW) for acquiring the semantic information of the original sentence. | ['Jingjing Zhao', 'Yan Yang', 'Bing Xu'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-3.96930635e-01 -7.54962489e-02 -1.09038293e-01 -4.85519111e-01
5.26382811e-02 -2.30204269e-01 3.86321455e-01 1.25640035e-01
-9.33428526e-01 3.76745075e-01 5.11023521e-01 -5.31018853e-01
6.60877764e-01 -1.06882596e+00 -1.78452045e-01 -5.35480261e-01
5.58827698e-01 -1.46773234e-01 7.57238688e-03 -7.60571778e-01
1.70254514e-01 1.02688245e-01 -1.07991612e+00 4.26456898e-01
1.47387460e-01 1.12405074e+00 2.91537166e-01 5.49600840e-01
-9.78423595e-01 5.61897337e-01 -9.59809661e-01 -4.33161169e-01
-8.19563568e-02 -2.55871773e-01 -7.07357347e-01 -1.77987933e-01
6.94537386e-02 -4.11132202e-02 -2.93347657e-01 9.97204661e-01
5.58423758e-01 5.12504756e-01 7.05463469e-01 -6.76986098e-01
-1.14559495e+00 5.89914441e-01 -3.23680758e-01 2.86098242e-01
-2.33818173e-01 -1.25075459e-01 1.22506499e+00 -1.25879467e+00
5.22628248e-01 1.20050621e+00 4.73717481e-01 7.55472660e-01
-5.25516629e-01 -8.09064746e-01 3.44164521e-01 3.80308181e-01
-9.92148995e-01 1.90477952e-01 6.40139461e-01 -4.52156693e-01
1.48275852e+00 -2.82587502e-02 1.09184444e+00 1.18093586e+00
2.86257803e-01 9.64668512e-01 3.95397693e-01 -3.57704908e-01
-1.88551173e-02 3.85859072e-01 9.39535022e-01 2.91765213e-01
3.22774678e-01 -3.30339909e-01 -1.98646322e-01 1.77849725e-01
5.19027174e-01 4.19304401e-01 -1.54653832e-01 8.18967447e-02
-7.29748666e-01 1.23072779e+00 6.87633693e-01 6.13637388e-01
-1.32097244e-01 3.73782694e-01 1.09030771e+00 3.31665665e-01
7.86529243e-01 6.63762093e-01 -8.15804899e-01 -5.38354293e-02
-3.41387212e-01 1.05474643e-01 3.92687827e-01 6.75013125e-01
6.00948393e-01 6.13814235e-01 -3.71044219e-01 8.73467088e-01
4.07908410e-01 6.58123553e-01 1.24081910e+00 1.48936987e-01
4.55857307e-01 8.80994916e-01 -2.66685575e-01 -1.17699981e+00
-5.29015601e-01 -3.24676454e-01 -8.98925424e-01 -1.30019695e-01
-2.50491798e-01 -4.33161497e-01 -1.12382066e+00 1.10407245e+00
-4.20844592e-02 -8.29191804e-02 3.66840899e-01 8.41469228e-01
1.46724558e+00 1.23812521e+00 1.89986333e-01 -8.34879503e-02
1.53477001e+00 -1.34187734e+00 -1.08974373e+00 -5.09487092e-01
1.02427721e+00 -4.57996875e-01 1.11858535e+00 1.73715249e-01
-4.76795495e-01 -9.79199171e-01 -1.33621979e+00 -1.51635885e-01
-1.24058282e+00 -1.10900454e-01 7.93906271e-01 4.96089399e-01
-7.48187482e-01 1.66593403e-01 -4.17045653e-01 -2.44882703e-01
5.00942171e-01 1.43362507e-01 -5.48297167e-01 -1.24302819e-01
-1.56644022e+00 1.24341834e+00 7.35902071e-01 5.23137674e-02
-3.04588795e-01 -4.84416097e-01 -1.39231610e+00 3.28002125e-01
-3.23727280e-01 -1.28720418e-01 9.26503778e-01 -1.12348640e+00
-1.29820848e+00 3.29214305e-01 -1.13995090e-01 -2.15114772e-01
-2.02445269e-01 -3.15730125e-01 -6.48158491e-01 -2.75658190e-01
-2.05706447e-01 3.47667694e-01 7.86728323e-01 -8.71317267e-01
-4.06972617e-01 -2.42775396e-01 -1.25189349e-01 1.66466057e-01
-1.25423515e+00 -1.56118304e-01 -4.38652515e-01 -7.09971428e-01
-3.22490692e-01 -2.35228777e-01 -5.42027533e-01 -6.42776549e-01
-1.59528702e-01 -7.22641766e-01 7.45068371e-01 -5.18193960e-01
1.36202848e+00 -2.58231902e+00 -1.37517899e-01 2.24372327e-01
9.51090604e-02 6.67622745e-01 -5.39707482e-01 3.87566805e-01
-3.74138653e-01 5.64168394e-02 2.34867968e-02 -4.33677852e-01
-3.58223170e-01 2.30301291e-01 -4.49187547e-01 1.78445622e-01
1.62258208e-01 1.24044096e+00 -9.85517144e-01 -3.13018441e-01
4.24598575e-01 6.58274829e-01 -3.65297198e-01 2.62840658e-01
4.17100005e-02 -1.60806954e-01 -3.38445663e-01 -1.03643946e-01
6.73761487e-01 -3.83476354e-02 7.41616040e-02 -2.03844950e-01
8.15476105e-02 1.73450038e-01 -9.53831136e-01 1.51368999e+00
-8.29734802e-01 7.69992530e-01 -5.57023108e-01 -1.00550866e+00
1.29018509e+00 3.81317198e-01 1.11822657e-01 -8.04852366e-01
6.97267473e-01 -3.20916384e-01 -2.88966507e-01 -6.64945364e-01
8.59313071e-01 -4.25429791e-01 -9.13211480e-02 3.08849722e-01
7.26997495e-01 -1.83583185e-01 6.79827705e-02 5.04693687e-02
6.50347769e-01 -2.67964214e-01 3.67750794e-01 -2.46122375e-01
5.66098571e-01 -1.15948267e-01 2.99973100e-01 1.46386445e-01
9.87217352e-02 5.14595389e-01 3.30115557e-01 -7.87089765e-01
-9.43639815e-01 -5.19750535e-01 -1.59095556e-01 1.17355824e+00
9.21137109e-02 -7.80141771e-01 -3.77354980e-01 -9.18273985e-01
-7.44282007e-02 1.07557106e+00 -9.99370396e-01 -4.98192638e-01
-2.83302844e-01 -6.46075010e-01 2.51540989e-01 9.28843498e-01
2.22643927e-01 -1.43415809e+00 -2.76139647e-01 2.64411688e-01
8.97840708e-02 -9.43032205e-01 -3.89887840e-01 4.37875271e-01
-7.55301297e-01 -9.42486882e-01 -5.53254068e-01 -1.12499917e+00
7.20794141e-01 5.45649648e-01 9.69863534e-01 3.01505655e-01
-3.25348862e-02 -4.82301377e-02 -8.58580589e-01 -6.88257039e-01
2.55536526e-01 6.08581230e-02 -1.89941600e-02 -4.66977134e-02
1.02306974e+00 1.06724471e-01 -3.78091007e-01 -3.55464250e-01
-9.93608057e-01 -1.11074388e-01 1.23478167e-01 1.07946682e+00
2.95180082e-01 -4.68167402e-02 4.96845752e-01 -9.13127482e-01
1.00525773e+00 -4.49953914e-01 -2.37969264e-01 9.39200446e-03
-7.58666217e-01 -2.83204138e-01 8.87717843e-01 -4.30289835e-01
-7.23669946e-01 -4.35452819e-01 -5.87395191e-01 -2.96210021e-01
2.99679488e-01 6.84044659e-01 8.71599540e-02 4.02073294e-01
4.75208670e-01 4.17773187e-01 -4.37234119e-02 -4.01097983e-01
7.49330401e-01 9.53960955e-01 -1.38084188e-01 3.04440051e-01
3.91216218e-01 2.24027351e-01 -5.41120887e-01 -9.32960629e-01
-1.09128475e+00 -7.62993813e-01 -5.58107138e-01 -4.43649627e-02
1.32878387e+00 -9.90265071e-01 -5.23603261e-01 2.64924139e-01
-1.61058617e+00 5.01964940e-04 -4.98521358e-01 6.38294220e-01
1.09527990e-01 6.71275258e-02 -4.83863890e-01 -5.67654848e-01
-7.69811690e-01 -9.69660223e-01 6.36091411e-01 2.61258334e-01
-2.93746889e-01 -1.40068078e+00 2.07910404e-01 1.80604886e-02
6.88470185e-01 -2.83571810e-01 9.13424730e-01 -1.12788022e+00
6.76722527e-01 -9.12980437e-01 -3.41455042e-01 1.20542002e+00
-2.61932947e-02 -2.75648773e-01 -1.02399170e+00 -2.31050432e-01
1.02171399e-01 -3.65818828e-01 1.18447268e+00 3.24400693e-01
1.30280387e+00 -1.39718100e-01 -2.35193130e-02 5.43158352e-01
1.65636170e+00 9.87936407e-02 6.77051306e-01 4.33998853e-01
1.06653559e+00 5.04027605e-01 4.11257625e-01 2.85762399e-01
3.87135565e-01 2.30966866e-01 3.54523838e-01 -3.65222275e-01
-4.69031371e-02 -6.25725761e-02 3.90115887e-01 1.50099182e+00
1.48663789e-01 -3.53127599e-01 -6.40372694e-01 6.81916296e-01
-1.72687089e+00 -6.91413403e-01 -3.86085242e-01 1.53039491e+00
5.67651868e-01 2.41134211e-01 -4.67293084e-01 8.32009763e-02
5.40525317e-01 5.87925673e-01 -1.55880734e-01 -1.05690205e+00
-9.90501866e-02 5.90697229e-01 4.91803378e-01 4.70429093e-01
-9.61382329e-01 1.36346912e+00 6.41191816e+00 1.00636184e+00
-1.19632530e+00 4.15656626e-01 3.66885394e-01 1.87057704e-02
-4.66313034e-01 -5.25542855e-01 -7.78817177e-01 3.38461608e-01
7.72357523e-01 -5.02127036e-02 -1.32517144e-01 1.21877897e+00
-2.09519476e-01 3.12440842e-01 -6.44907117e-01 1.12076426e+00
5.95533013e-01 -1.33252895e+00 5.23818195e-01 -4.05572414e-01
5.11105537e-01 6.15166388e-02 -5.69460206e-02 8.65959406e-01
1.47482827e-01 -1.31065214e+00 1.99662849e-01 1.72614023e-01
8.67747545e-01 -1.05701184e+00 1.56522620e+00 -8.89678672e-02
-1.11251378e+00 5.99704729e-03 -1.00831425e+00 -3.80920172e-01
1.07485518e-01 6.23127162e-01 -4.76867318e-01 3.95548791e-01
6.58630371e-01 1.17541111e+00 -6.14643395e-01 4.49728847e-01
-5.21189451e-01 6.79621458e-01 4.31013823e-01 -5.63952208e-01
7.20645905e-01 -1.40054852e-01 2.26754248e-02 1.68721223e+00
5.78932986e-02 4.60246839e-02 -1.27696291e-01 1.86378196e-01
-1.72185794e-01 7.32012510e-01 -7.87351131e-01 -2.73467213e-01
-6.99912384e-02 1.54493392e+00 -4.46965396e-01 -5.45724630e-01
-4.80608076e-01 9.03829873e-01 5.19267499e-01 4.02884185e-01
-3.04542482e-01 -9.63900328e-01 1.00100243e+00 -3.78772438e-01
7.01427162e-01 -3.73908341e-01 -6.56648874e-01 -1.11795104e+00
-2.00046822e-01 -2.79579580e-01 1.69938684e-01 -6.67849600e-01
-1.41949558e+00 8.42898130e-01 -5.38444638e-01 -1.10426295e+00
2.39900962e-01 -1.26121247e+00 -1.09732985e+00 1.06164527e+00
-1.59080625e+00 -9.98155773e-01 -1.85007364e-01 5.95291615e-01
9.01968420e-01 -5.78794658e-01 1.43767965e+00 8.45288932e-02
-4.54666108e-01 5.98756492e-01 2.21863791e-01 4.57511038e-01
6.11242652e-01 -1.16468656e+00 7.43849397e-01 5.11474550e-01
-9.26529244e-02 5.48960447e-01 3.79652530e-01 -5.20330608e-01
-1.25195682e+00 -1.33906281e+00 1.21921420e+00 -3.82318765e-01
8.67233992e-01 -6.38225615e-01 -7.16628253e-01 5.66195011e-01
5.20118892e-01 -9.75800157e-02 1.22516227e+00 3.23471159e-01
-5.73493838e-01 -5.23355342e-02 -6.76652193e-01 6.39231384e-01
2.37767637e-01 -3.85193139e-01 -8.69346738e-01 3.92007023e-01
1.39475334e+00 8.51391703e-02 -5.41909277e-01 7.08128372e-03
3.97731334e-01 -3.16557020e-01 9.09118652e-01 -1.22125828e+00
8.46005917e-01 1.86862366e-03 -3.19458663e-01 -1.87691212e+00
-3.90067041e-01 3.58632416e-01 -5.87056838e-02 9.49375868e-01
5.90436518e-01 -5.68927228e-01 3.10600877e-01 -2.90536731e-02
-2.15854958e-01 -8.60719144e-01 -6.01081610e-01 -5.64895034e-01
3.39762002e-01 -9.18617129e-01 5.15683055e-01 1.03333068e+00
3.51172507e-01 9.27211702e-01 -3.35393012e-01 -6.27800405e-01
-5.69983460e-02 -2.79603541e-01 4.32441235e-01 -9.35113430e-01
2.19699323e-01 -4.18801606e-01 -7.39595354e-01 -1.05920184e+00
2.97728121e-01 -1.19018924e+00 7.34320655e-02 -1.93017328e+00
-3.74544896e-02 2.30478227e-01 -8.31383169e-01 5.02802312e-01
-5.46450615e-01 3.08618128e-01 2.37892523e-01 -1.88847721e-01
-4.92234945e-01 8.87553394e-01 1.63054168e+00 -4.09607679e-01
2.45782621e-02 -2.20447317e-01 -8.56069267e-01 5.60806572e-01
7.78365254e-01 -3.76468688e-01 -2.77695090e-01 -7.95014441e-01
4.93691891e-01 -5.26363671e-01 -3.10821176e-01 -4.77844775e-01
1.31143764e-01 4.02385220e-02 6.25347137e-01 -6.41860723e-01
2.60161936e-01 -9.53884304e-01 -8.89056981e-01 4.82825994e-01
-2.76935130e-01 2.50807971e-01 2.90362686e-01 3.40395629e-01
-5.57507992e-01 -5.93859673e-01 5.71708620e-01 -1.24275178e-01
-9.34187353e-01 3.95367235e-01 -6.34407938e-01 -2.26680711e-01
9.20650244e-01 3.79411317e-02 -1.82900950e-01 -8.97735655e-02
-3.85325938e-01 4.07369196e-01 -1.69507891e-01 9.15054917e-01
1.04131424e+00 -1.59476864e+00 -5.81578493e-01 3.51969123e-01
6.08258367e-01 -3.14424224e-02 1.28574789e-01 2.56567657e-01
-6.16600335e-01 3.31251979e-01 -1.02758817e-01 4.70646359e-02
-9.72630739e-01 7.18352497e-01 1.89179391e-01 -2.30561301e-01
-7.41712987e-01 1.25615025e+00 2.65415251e-01 -6.61586285e-01
2.51836270e-01 3.29391696e-02 -1.33865392e+00 5.60560942e-01
1.07329071e+00 1.12296678e-02 7.59931281e-02 -6.24532342e-01
-2.19768211e-01 4.75279331e-01 -3.34011763e-01 1.16758406e-01
1.45407438e+00 1.13540873e-01 -3.42827350e-01 8.15370679e-01
1.74761713e+00 -5.22761524e-01 -3.57437789e-01 -1.53507084e-01
-2.15912715e-01 -1.86722115e-01 3.09621900e-01 -3.66002142e-01
-1.29772329e+00 1.30111194e+00 6.15440309e-01 2.40906447e-01
5.67758024e-01 -3.27332407e-01 1.19342792e+00 5.27900755e-01
-1.97543636e-01 -1.45036328e+00 2.09687427e-01 1.31584132e+00
6.68793321e-01 -1.31083858e+00 -1.26793683e-01 1.27157137e-01
-9.48264003e-01 1.56100547e+00 9.38173532e-01 -5.85991383e-01
1.17042017e+00 5.60218468e-02 4.83822405e-01 -6.18285060e-01
-5.15016496e-01 -2.20899865e-01 2.80370653e-01 5.23001671e-01
6.17199004e-01 2.30311185e-01 -5.76213658e-01 1.26194072e+00
-3.26895356e-01 -4.85286683e-01 4.56135571e-01 7.63163984e-01
-5.64605892e-01 -8.10912073e-01 1.54791340e-01 6.10004842e-01
-3.51233453e-01 -6.80392683e-01 -1.91421822e-01 3.44600528e-01
2.06475347e-01 9.47750688e-01 4.78512615e-01 -8.35935652e-01
6.14685535e-01 2.59621561e-01 -1.88200772e-01 -1.09392345e+00
-9.83115256e-01 -4.22318161e-01 -1.21200919e-01 -4.35105741e-01
-8.08755830e-02 1.18433185e-01 -1.35140193e+00 -1.43035039e-01
-6.49122834e-01 4.48604971e-01 1.18706548e+00 1.20624006e+00
-4.64455038e-02 9.87616718e-01 8.19054842e-01 -5.05609930e-01
-2.23587856e-01 -1.47180235e+00 -8.72125328e-01 5.90252995e-01
1.75141871e-01 -3.00985813e-01 -3.47362459e-01 -5.01374118e-02] | [10.506964683532715, 8.445016860961914] |
fdb71aa3-eca6-4147-a524-d176bb5bd0ee | identifying-and-measuring-annotator-bias | null | null | https://aclanthology.org/2020.alw-1.21 | https://aclanthology.org/2020.alw-1.21.pdf | Identifying and Measuring Annotator Bias Based on Annotators’ Demographic Characteristics | Machine learning is recently used to detect hate speech and other forms of abusive language in online platforms. However, a notable weakness of machine learning models is their vulnerability to bias, which can impair their performance and fairness. One type is annotator bias caused by the subjective perception of the annotators. In this work, we investigate annotator bias using classification models trained on data from demographically distinct annotator groups. To do so, we sample balanced subsets of data that are labeled by demographically distinct annotators. We then train classifiers on these subsets, analyze their performances on similarly grouped test sets, and compare them statistically. Our findings show that the proposed approach successfully identifies bias and that demographic features, such as first language, age, and education, correlate with significant performance differences. | ['Georg Groh', 'Maximilian Wich', 'Hala Al Kuwatly'] | null | null | null | null | emnlp-alw-2020-11 | ['abusive-language'] | ['natural-language-processing'] | [-3.91286045e-01 -1.05022669e-01 -6.04887843e-01 -4.65839148e-01
-1.59021497e-01 -8.53976071e-01 5.48340619e-01 5.27334392e-01
-5.37368655e-01 6.17575586e-01 3.71421158e-01 -2.35777840e-01
2.72777706e-01 -3.92795473e-01 1.77878980e-02 -5.08273721e-01
3.08407784e-01 -4.68161553e-02 -1.20636843e-01 -1.00646585e-01
7.24295020e-01 3.64847869e-01 -9.33616936e-01 1.74596265e-01
1.24567640e+00 6.16943300e-01 -7.09932983e-01 2.12433532e-01
-6.27167244e-03 9.85118628e-01 -1.04252565e+00 -1.01874900e+00
1.02287896e-01 -4.64105636e-01 -6.98591709e-01 -1.04548030e-01
7.58110404e-01 -3.29708248e-01 -1.59386039e-01 1.27732646e+00
5.87498784e-01 -2.06334323e-01 8.72469664e-01 -1.26163781e+00
-1.06667709e+00 7.67965972e-01 -7.03645229e-01 5.04685521e-01
4.62089777e-01 1.37798309e-01 9.03875113e-01 -7.04594493e-01
4.50690925e-01 1.34866464e+00 7.47992575e-01 8.22744250e-01
-1.39940512e+00 -1.20144808e+00 -7.69398361e-02 -8.69735405e-02
-1.20310628e+00 -4.90368724e-01 9.63004291e-01 -1.02646697e+00
-9.60657895e-02 1.05814852e-01 2.64746070e-01 1.44981360e+00
1.39139712e-01 4.08078343e-01 1.79475117e+00 -3.05094600e-01
3.63297969e-01 9.61186171e-01 7.70602286e-01 3.92327815e-01
6.16869390e-01 -1.54703140e-01 -5.33573806e-01 -7.46348798e-01
-1.03320666e-01 -1.99151009e-01 -1.54722795e-01 1.62036121e-02
-5.82102835e-01 1.34883869e+00 3.14254135e-01 5.41794062e-01
2.21924819e-02 -3.28793883e-01 6.47376716e-01 2.39206120e-01
9.33675349e-01 7.12211668e-01 -5.31605482e-02 1.28918901e-01
-8.83890092e-01 1.19926080e-01 8.76830757e-01 1.50842443e-01
4.53356743e-01 3.61669697e-02 -3.01873475e-01 9.44863915e-01
2.06707671e-01 6.37308955e-01 9.23523307e-01 -6.24777198e-01
3.79150629e-01 4.89103496e-01 1.52569011e-01 -1.52095652e+00
-4.05959696e-01 -3.23947221e-01 -5.28401256e-01 7.56234676e-02
8.30058038e-01 -4.35088933e-01 -2.92288870e-01 1.54919052e+00
8.89792368e-02 -5.29083490e-01 -3.82955968e-01 9.22363877e-01
6.45755827e-01 1.44340292e-01 5.65126419e-01 -2.30855584e-01
1.34756756e+00 -4.65544194e-01 -8.61681759e-01 -2.81975508e-01
8.90678406e-01 -6.15312099e-01 8.85552883e-01 2.36115322e-01
-6.99652433e-01 -4.73748565e-01 -6.74259901e-01 1.55096084e-01
-4.78101790e-01 -1.45139545e-02 3.68104190e-01 1.62413216e+00
-5.58631897e-01 4.10224915e-01 -3.00509054e-02 -2.08565339e-01
6.28389299e-01 8.54867324e-02 -3.12814981e-01 3.09891313e-01
-1.31053913e+00 1.06103957e+00 9.77682532e-04 -2.51861066e-01
-5.40858805e-01 -6.58932090e-01 -7.38204598e-01 -9.19396579e-02
-3.71669382e-02 6.35402948e-02 9.98841226e-01 -1.54315436e+00
-9.64349985e-01 1.37629974e+00 -6.74641654e-02 -4.55724120e-01
6.58377528e-01 -1.81239381e-01 -5.26186526e-01 -1.73019588e-01
2.05595404e-01 9.86227915e-02 1.04633677e+00 -1.38405931e+00
-1.87324554e-01 -8.37995112e-01 -1.00806646e-01 -2.14363381e-01
-9.59078372e-01 7.16527283e-01 8.46071064e-01 -7.02532291e-01
-3.71216595e-01 -9.75037992e-01 1.48261413e-01 -2.97052175e-01
-3.33956987e-01 -4.83079940e-01 7.17826724e-01 -1.02736723e+00
1.58195114e+00 -2.37135005e+00 -2.42942616e-01 4.58992481e-01
5.80700338e-01 5.39893925e-01 3.28703374e-01 1.02453113e-01
-4.92803976e-02 7.72935867e-01 2.37302899e-01 -2.60029674e-01
3.50713283e-02 -7.36840740e-02 -1.69900626e-01 8.31332147e-01
6.27229409e-03 4.19616789e-01 -8.14584076e-01 -6.39827847e-01
-2.01964840e-01 1.11239664e-01 -3.71175200e-01 2.86466956e-01
4.02078062e-01 3.82825434e-01 -3.07262808e-01 5.58030903e-01
6.32978976e-01 2.81113595e-01 3.46731961e-01 1.66463971e-01
-1.72335759e-01 4.66205537e-01 -5.25877476e-01 6.51378274e-01
-2.26434946e-01 1.10896647e+00 1.01658739e-01 -7.41545737e-01
1.38275766e+00 1.77220851e-01 -7.68291801e-02 -4.19839978e-01
6.23974741e-01 4.86172050e-01 4.57917720e-01 -5.37884653e-01
4.91436720e-01 -2.57600665e-01 -3.06467086e-01 5.17494321e-01
-2.44840652e-01 2.94531077e-01 1.50448605e-01 8.50614384e-02
5.50869226e-01 -6.02351010e-01 4.16573793e-01 -5.50684690e-01
6.27570510e-01 -1.69239566e-01 5.60992837e-01 7.89803088e-01
-1.02706754e+00 1.75369337e-01 9.62802768e-01 -5.79586923e-01
-1.02107608e+00 -4.97829497e-01 -2.90288031e-01 1.42780399e+00
-2.06117705e-01 1.41783925e-02 -8.76029074e-01 -1.28901720e+00
2.82860428e-01 7.95740187e-01 -8.09471369e-01 -4.28926647e-01
-3.03042561e-01 -9.08664286e-01 8.05948019e-01 3.25005561e-01
3.93402994e-01 -6.37628436e-01 -3.86581182e-01 -3.17185670e-01
1.63031220e-02 -9.09637630e-01 -3.30511600e-01 -2.87934065e-01
-5.06467283e-01 -1.07108939e+00 -6.73119664e-01 -5.30185640e-01
5.72234213e-01 2.13894323e-01 9.56793785e-01 4.98990059e-01
1.21595427e-01 -1.49086295e-02 -3.98228258e-01 -7.22761989e-01
-7.80809999e-01 4.01046306e-01 1.92628101e-01 4.51496840e-01
7.90636659e-01 -7.64181912e-02 -2.42948800e-01 4.74701703e-01
-7.16729760e-01 -5.52161038e-01 1.77463442e-01 4.90748256e-01
-5.88512480e-01 -1.54475197e-01 5.67529202e-01 -1.28168881e+00
8.55808198e-01 -7.40231216e-01 -2.91244715e-01 -8.31909031e-02
-6.86521709e-01 -3.12994599e-01 4.92209762e-01 -6.34048939e-01
-1.04896474e+00 -4.48790014e-01 2.92040229e-01 1.37097016e-01
-3.50513130e-01 2.25408167e-01 3.24749425e-02 -2.15275049e-01
9.72737730e-01 -3.56028587e-01 2.53125250e-01 -4.41122413e-01
-3.86189580e-01 1.22542095e+00 -5.10327779e-02 -3.39036286e-01
1.00005877e+00 3.09141606e-01 -3.61528337e-01 -6.87910259e-01
-1.21017730e+00 -3.12084705e-01 -7.58092105e-01 -4.57429767e-01
7.98638463e-01 -7.38161147e-01 -6.61520422e-01 5.82700431e-01
-9.75636363e-01 -7.94422999e-02 5.41647673e-01 4.84073013e-01
2.66362548e-01 4.07297015e-01 -5.92480898e-01 -1.32122743e+00
-3.85283798e-01 -7.11500525e-01 4.07840431e-01 3.25526744e-01
-8.10612440e-01 -1.23975646e+00 1.98834106e-01 7.82196701e-01
3.88189644e-01 3.81070018e-01 8.19258749e-01 -1.46013141e+00
6.04871988e-01 -3.95563096e-01 -1.57236114e-01 5.97264647e-01
-1.14372626e-01 2.52820075e-01 -9.63943958e-01 -3.00543398e-01
-4.11015823e-02 -4.38855767e-01 4.86069888e-01 -4.28500175e-02
1.00177121e+00 -5.34533203e-01 -3.65090519e-01 -1.48099318e-01
1.19038177e+00 -2.17477605e-02 1.72046423e-01 3.73506099e-01
6.63317502e-01 1.00380194e+00 2.64633745e-01 4.11611974e-01
1.85361907e-01 5.04603326e-01 2.01221973e-01 1.08057685e-01
3.35273683e-01 -1.76435664e-01 5.30502439e-01 4.42290425e-01
-1.33901238e-01 8.42802674e-02 -1.12712419e+00 3.33364606e-01
-1.42171526e+00 -9.78659153e-01 -5.15061200e-01 2.12167001e+00
7.11494803e-01 1.37905702e-01 7.27110028e-01 4.09491241e-01
1.16969264e+00 3.71368349e-01 -7.44318962e-02 -9.16920722e-01
-9.29241776e-02 -1.36184305e-01 5.59035301e-01 3.34385902e-01
-1.11174572e+00 5.51369071e-01 6.67589712e+00 3.84061307e-01
-1.08779955e+00 3.26620996e-01 1.06071615e+00 -1.76680535e-01
-1.44384176e-01 -3.32434475e-01 -6.62466228e-01 9.90634859e-01
9.71603215e-01 -1.57651350e-01 1.56084880e-01 8.84474158e-01
3.75870854e-01 -1.04464386e-02 -8.32077801e-01 6.30747736e-01
6.28633738e-01 -4.16589677e-01 -3.64978462e-01 3.32319140e-01
7.22433746e-01 -5.77302575e-01 2.24218488e-01 4.03790563e-01
2.84021199e-01 -1.05763149e+00 8.10461581e-01 -1.80251822e-02
1.21001087e-01 -6.53014779e-01 1.07987750e+00 3.19555223e-01
-3.12173404e-02 -5.76653898e-01 -3.44219625e-01 -5.11086404e-01
-2.87476152e-01 5.97086072e-01 -6.59528017e-01 -4.18391041e-02
6.12917960e-01 3.38915944e-01 -1.00370610e+00 6.94202363e-01
-2.77071953e-01 9.65306938e-01 2.07778305e-01 -3.37741047e-01
7.52317607e-02 -2.11931374e-02 2.04801276e-01 1.09867358e+00
-2.54618693e-02 -2.48913422e-01 -4.64271642e-02 7.00640976e-01
-2.10669383e-01 3.87298137e-01 -6.81439340e-01 -8.22030529e-02
7.13229001e-01 1.36577940e+00 -3.22053581e-01 -2.18429402e-01
-6.19436622e-01 6.91340387e-01 6.22105479e-01 1.73355430e-01
-8.15850437e-01 -1.65666625e-01 5.61738789e-01 2.32648492e-01
-3.67552459e-01 -1.87303759e-02 -7.32878089e-01 -1.10521901e+00
-2.25841805e-01 -1.16626728e+00 6.01147473e-01 -3.48849803e-01
-1.66796947e+00 1.09782279e-01 -1.77829295e-01 -6.95594311e-01
9.17897075e-02 -8.11131835e-01 -4.91996616e-01 9.24659252e-01
-1.22035420e+00 -8.67977262e-01 -4.70777780e-01 1.83974057e-01
7.46659636e-02 -3.22283953e-01 4.91787404e-01 2.28529289e-01
-9.61342216e-01 7.98202693e-01 -9.69014764e-02 9.46570098e-01
1.04904866e+00 -1.18883026e+00 -3.11380714e-01 4.25961763e-01
-2.00928569e-01 5.80815256e-01 7.23933756e-01 -6.02884531e-01
-5.77787459e-01 -6.70586765e-01 1.28162658e+00 -8.16880405e-01
8.55615854e-01 -4.36415046e-01 -8.91177595e-01 4.84712154e-01
2.36999199e-01 -1.11700974e-01 1.40537834e+00 4.09954011e-01
-9.86704409e-01 8.47732723e-02 -1.39242315e+00 3.84221345e-01
7.17127383e-01 -6.47978187e-01 -7.34577835e-01 1.72305077e-01
-5.87082617e-02 5.90616697e-03 -9.89000738e-01 -5.21190502e-02
7.73626328e-01 -1.24098802e+00 3.93175453e-01 -9.26699102e-01
6.85853243e-01 3.24918628e-01 2.39704221e-01 -1.43945718e+00
-5.29519439e-01 -4.45927620e-01 2.71109581e-01 1.68166924e+00
4.02050734e-01 -7.48296916e-01 6.61035478e-01 1.08761394e+00
4.74685848e-01 -2.58498877e-01 -6.55773401e-01 -6.33509219e-01
7.22856343e-01 2.00987816e-01 3.79748464e-01 1.57754385e+00
3.06392878e-01 3.78124326e-01 -4.21487063e-01 -1.80525593e-02
5.39792120e-01 -1.57097369e-01 8.57268810e-01 -1.65134907e+00
1.48914367e-01 -6.68715775e-01 -3.39322984e-01 5.28841699e-03
8.37540627e-01 -6.41703486e-01 -4.52085316e-01 -6.75061643e-01
5.62609732e-01 -4.22111273e-01 -1.94870271e-02 2.80858099e-01
-5.58671892e-01 5.77434778e-01 3.64027828e-01 4.01573896e-01
-2.60285765e-01 9.08414498e-02 8.22784305e-01 -7.81762898e-02
-8.37475955e-02 -5.05710021e-02 -7.97699332e-01 7.89287210e-01
1.09632659e+00 -5.89376807e-01 1.38281569e-01 -1.38295785e-01
2.22207606e-01 -5.25369346e-01 4.57969993e-01 -7.46131539e-01
-4.27931577e-01 -2.36228332e-01 5.92142344e-01 2.72576839e-01
-7.70860016e-02 -7.16856837e-01 -5.01600623e-01 7.48428226e-01
-7.51498461e-01 7.50555694e-02 -2.04417765e-01 2.38389120e-01
-1.18971216e-02 -8.06824505e-01 8.25948298e-01 -7.22374171e-02
7.77206942e-02 -9.27219167e-02 -7.67197311e-01 3.17940831e-01
1.07336307e+00 4.44176421e-02 -5.75638592e-01 -5.03963828e-01
-4.29060251e-01 -2.18499005e-01 6.72641337e-01 4.18579102e-01
-2.39421725e-01 -1.25413847e+00 -9.06675339e-01 -1.28845513e-01
2.47300118e-01 -1.05708027e+00 -2.40893275e-01 8.76421392e-01
-2.04744533e-01 1.92309484e-01 -4.07708704e-01 2.45616883e-02
-1.60661745e+00 6.37844145e-01 4.17053521e-01 -5.49614336e-03
4.66862887e-01 4.09585446e-01 -1.46288574e-01 -2.74636090e-01
8.87657981e-03 3.77760261e-01 -5.64520657e-01 5.80411315e-01
5.67532837e-01 8.16765726e-01 -3.36469531e-01 -1.12673950e+00
-2.70472080e-01 -1.12106306e-02 -1.40356049e-01 -2.41538864e-02
7.22378552e-01 2.18907408e-02 -3.58473808e-01 4.94316936e-01
1.07587659e+00 6.31563723e-01 -4.06526238e-01 -2.11749405e-01
2.94857591e-01 -9.81898069e-01 -2.24594966e-01 -5.04705667e-01
-7.78687596e-01 7.80459225e-01 4.63339597e-01 9.90409970e-01
5.08157074e-01 -2.64198840e-01 3.98103416e-01 -1.92867070e-01
1.09974802e-01 -1.25123203e+00 2.18133688e-01 3.19827199e-01
4.34017748e-01 -1.58801949e+00 -3.71955596e-02 -5.16962767e-01
-6.21763051e-01 8.87691081e-01 1.00542963e+00 -3.75357084e-02
4.85967219e-01 -3.60011667e-01 4.54002947e-01 3.03721114e-04
-2.74929792e-01 1.37424931e-01 4.07847136e-01 7.50737071e-01
1.05377960e+00 1.91948906e-01 -1.03453803e+00 6.35542333e-01
-3.64388466e-01 -4.28237736e-01 7.31369972e-01 5.87062180e-01
-2.92878091e-01 -9.18622971e-01 -8.14047873e-01 4.32872534e-01
-1.09878397e+00 3.67247239e-02 -1.17445731e+00 7.00759232e-01
2.16637835e-01 1.09974742e+00 1.35140985e-01 -4.03784662e-01
5.79134598e-02 4.60531294e-01 5.75080104e-02 -5.79498291e-01
-1.28140283e+00 -6.08617544e-01 4.67358559e-01 1.69205368e-02
-3.86539370e-01 -7.80029416e-01 -4.10139650e-01 -7.55481303e-01
-2.75747120e-01 3.57143521e-01 4.89225179e-01 9.77665961e-01
9.42343250e-02 -2.62555093e-01 9.58822906e-01 -3.62478524e-01
-8.81444991e-01 -1.29983711e+00 -4.93746459e-01 1.06043410e+00
3.17688316e-01 -6.32444620e-01 -7.83944547e-01 -2.29694277e-01] | [8.730613708496094, 10.480405807495117] |
68a81576-0119-4cc3-8fab-dc8141abbc7a | detecting-table-region-in-pdf-documents-using | 1506.08891 | null | http://arxiv.org/abs/1506.08891v6 | http://arxiv.org/pdf/1506.08891v6.pdf | Detecting Table Region in PDF Documents Using Distant Supervision | Superior to state-of-the-art approaches which compete in table recognition
with 67 annotated government reports in PDF format released by {\it ICDAR 2013
Table Competition}, this paper contributes a novel paradigm leveraging
large-scale unlabeled PDF documents to open-domain table detection. We
integrate the paradigm into our latest developed system ({\it PdfExtra}) to
detect the region of tables by means of 9,466 academic articles from the entire
repository of {\it ACL Anthology}, where almost all papers are archived by PDF
format without annotation for tables. The paradigm first designs heuristics to
automatically construct weakly labeled data. It then feeds diverse evidences,
such as layouts of documents and linguistic features, which are extracted by
{\it Apache PDFBox} and processed by {\it Stanford NLP} toolkit, into different
canonical classifiers. We finally use these classifiers, i.e. {\it Naive
Bayes}, {\it Logistic Regression} and {\it Support Vector Machine}, to
collaboratively vote on the region of tables. Experimental results show that
{\it PdfExtra} achieves a great leap forward, compared with the
state-of-the-art approach. Moreover, we discuss the factors of different
features, learning models and even domains of documents that may impact the
performance. Extensive evaluations demonstrate that our paradigm is compatible
enough to leverage various features and learning models for open-domain table
region detection within PDF files. | ['Miao Fan', 'Doo Soon Kim'] | 2015-06-29 | null | null | null | null | ['table-recognition', 'table-detection'] | ['computer-vision', 'miscellaneous'] | [-2.02957407e-01 3.40976179e-01 -4.60653603e-01 -2.30095550e-01
-1.60083890e+00 -1.46155775e+00 8.38611424e-01 6.68403268e-01
-2.36888617e-01 8.67711544e-01 4.42476332e-01 -6.52393699e-01
-1.72401667e-01 -9.08621967e-01 -1.04625750e+00 -1.51917011e-01
2.43446678e-01 9.44813728e-01 3.06759775e-01 -1.42282113e-01
5.44205189e-01 4.65634733e-01 -1.32336819e+00 1.06183910e+00
8.99008930e-01 1.30257940e+00 -2.86866166e-02 5.57645023e-01
-7.68837333e-01 9.95907485e-01 -7.12519705e-01 -7.79336512e-01
4.34088200e-01 1.86308861e-01 -8.41332376e-01 -2.36359254e-01
5.66161931e-01 -9.91783068e-02 -4.66800481e-01 9.69469070e-01
1.60555005e-01 -3.11856300e-01 9.02399659e-01 -1.04452026e+00
-9.13751781e-01 1.28615737e+00 -3.72020423e-01 4.85489249e-01
7.19310403e-01 -3.42478961e-01 1.21867371e+00 -1.05270767e+00
1.11551356e+00 1.17200410e+00 6.27499104e-01 4.48583998e-02
-7.88090646e-01 -7.93694854e-01 2.87588179e-01 5.32402582e-02
-1.30279112e+00 -4.32998747e-01 4.79971945e-01 -6.75907373e-01
7.15874732e-01 5.58580935e-01 2.23999079e-02 1.22218621e+00
8.96836519e-02 7.93475449e-01 1.32931256e+00 -6.42905772e-01
3.17198873e-01 6.93032384e-01 4.16705966e-01 6.73047841e-01
7.63057411e-01 -6.11347675e-01 -6.31107867e-01 -2.88388848e-01
3.80913198e-01 -3.55658114e-01 7.81814978e-02 -1.40713066e-01
-1.15357494e+00 6.49942338e-01 -1.92054622e-02 3.32922161e-01
-4.66397069e-02 -7.62646735e-01 6.11325324e-01 1.67842165e-01
5.40027440e-01 4.95264113e-01 -7.44124711e-01 -8.72578099e-02
-7.74288058e-01 3.99309307e-01 1.14621472e+00 1.43330383e+00
3.29364002e-01 -6.45316899e-01 -5.33976972e-01 7.36719072e-01
3.17451775e-01 6.36538982e-01 8.00836552e-03 -5.84823728e-01
1.48721266e+00 1.12135756e+00 3.46930593e-01 -9.61435854e-01
-4.53449905e-01 -1.14207670e-01 -4.40760970e-01 -3.41766834e-01
9.19184327e-01 -3.75586748e-02 -6.47640049e-01 8.74985933e-01
2.14689910e-01 -8.14637482e-01 6.14981167e-02 3.70359331e-01
1.26870918e+00 6.04176104e-01 -4.63066623e-02 -1.23232447e-01
1.69492221e+00 -7.56373405e-01 -8.63889754e-01 -8.09010863e-03
6.77194417e-01 -8.71646762e-01 1.11795747e+00 7.62819529e-01
-1.02791500e+00 -3.45231563e-01 -9.94484544e-01 -1.18500426e-01
-1.20737433e+00 5.71962655e-01 4.41589773e-01 9.41293538e-01
-5.43041229e-01 2.57543892e-01 -3.51418376e-01 -5.00729859e-01
6.12728179e-01 1.40899897e-01 -3.70893329e-01 -8.35554749e-02
-1.03364575e+00 7.64841497e-01 3.44223291e-01 -1.45007685e-01
-2.78022349e-01 -6.34134948e-01 -6.75271928e-01 -3.44222225e-02
6.61939621e-01 -1.81613907e-01 8.96861374e-01 1.09264866e-01
-9.44748700e-01 1.12241685e+00 1.33985966e-01 -2.43902370e-01
6.38228059e-01 -3.11412305e-01 -7.16049016e-01 1.75972119e-01
6.86651289e-01 2.32850183e-02 7.89230317e-02 -1.15991771e+00
-7.14626312e-01 -6.95874155e-01 -3.01120758e-01 -2.13370427e-01
-5.08259475e-01 5.75478792e-01 -8.59745622e-01 -5.59170485e-01
3.17551494e-01 -6.97227478e-01 2.12877646e-01 -5.27375937e-01
-9.93266106e-01 -5.10225654e-01 5.66325068e-01 -1.03744316e+00
1.77143753e+00 -1.72812188e+00 -3.39818269e-01 5.16609192e-01
2.06937104e-01 -9.62235779e-02 4.81804907e-01 5.92979789e-01
1.75178140e-01 5.37583709e-01 2.44743809e-01 7.99540877e-02
4.57074046e-01 -1.55585989e-01 -5.94543993e-01 3.71495962e-01
7.28564113e-02 6.47987008e-01 -4.17283177e-01 -7.76930571e-01
-2.22430602e-01 1.25037521e-01 -1.98316887e-01 -1.54356748e-01
-3.88363481e-01 1.34426460e-01 -7.07555473e-01 1.12388909e+00
8.94319057e-01 -3.42553705e-01 6.11898601e-01 -8.77853036e-02
-4.18770403e-01 5.56339979e-01 -1.50173879e+00 1.19607699e+00
6.27217889e-02 3.19864959e-01 -8.11219029e-03 -9.53885138e-01
1.28686261e+00 7.00676888e-02 2.18150645e-01 -6.86484873e-01
1.92887351e-01 4.20302331e-01 -5.69253504e-01 -5.38296700e-01
6.06815040e-01 8.36951554e-01 -6.74547374e-01 1.97158799e-01
4.54988182e-02 7.71353990e-02 6.74049675e-01 5.03564835e-01
1.35313118e+00 2.89912879e-01 2.47349128e-01 -4.18580532e-01
8.88507664e-01 4.46602106e-01 3.98925781e-01 1.08382320e+00
-1.97131500e-01 6.58531487e-01 9.47437406e-01 -4.76581246e-01
-1.24483907e+00 -1.02271080e+00 -4.80742395e-01 1.29110730e+00
-1.62811384e-01 -8.34492445e-01 -8.87177110e-01 -1.12617445e+00
2.23112836e-01 5.64994693e-01 -7.79018521e-01 5.16305447e-01
-6.97051585e-01 -6.90978944e-01 7.88349569e-01 5.83881497e-01
3.96397382e-01 -9.28314090e-01 -4.29645367e-02 2.26757169e-01
-3.55002463e-01 -1.53733742e+00 -1.08788833e-02 5.83780169e-01
-4.41677272e-01 -9.66709435e-01 -2.51663774e-01 -5.77930927e-01
3.88803601e-01 -4.60225523e-01 1.33136117e+00 -3.32431436e-01
-1.16721056e-01 3.43556017e-01 -4.03686523e-01 -6.70236826e-01
-5.58000863e-01 3.13800901e-01 -5.49857207e-02 -2.59254485e-01
5.24959803e-01 5.00749657e-03 -5.84001578e-02 2.71692902e-01
-5.79221368e-01 -5.02782106e-01 7.05394566e-01 4.55704957e-01
7.86553144e-01 1.89113885e-01 4.26983297e-01 -1.38707542e+00
5.60372353e-01 -4.95483041e-01 -7.82953799e-01 6.36968672e-01
-7.86594868e-01 5.08171432e-02 8.64635944e-01 -1.96871161e-02
-1.19366348e+00 -1.75198317e-02 -1.00783946e-03 3.07897832e-02
-3.40892285e-01 4.71791476e-01 -5.96439481e-01 4.39015925e-01
7.61977613e-01 1.74782470e-01 -5.15234649e-01 -8.96455586e-01
2.56755471e-01 1.01871753e+00 6.91906989e-01 -6.47086501e-01
9.36853707e-01 3.97252351e-01 -2.83862948e-01 -2.51436412e-01
-1.23993623e+00 -5.87936342e-01 -9.21615660e-01 -1.61908165e-01
8.69968176e-01 -9.99259770e-01 -6.52622521e-01 -8.71291442e-04
-9.56688046e-01 2.64160991e-01 9.61912870e-02 1.46723211e-01
-3.77380043e-01 5.71802184e-02 -7.10400343e-01 -9.56684649e-01
-2.30480894e-01 -8.00562978e-01 1.15175378e+00 2.66344488e-01
-4.80717309e-02 -7.21907914e-01 -1.58530027e-01 1.07239103e+00
4.30982411e-02 1.77211642e-01 1.09187949e+00 -1.57807481e+00
-6.09590173e-01 -4.66355383e-01 -4.29840475e-01 -1.24255583e-01
-3.73751670e-01 2.82457829e-01 -9.16616321e-01 2.61919081e-01
-5.87309539e-01 -3.70246232e-01 6.85063183e-01 2.89218333e-02
1.28294420e+00 -5.32645166e-01 -6.86496019e-01 2.21423879e-01
1.34501040e+00 2.26339430e-01 6.92457020e-01 1.01413548e+00
5.11454761e-01 5.52035868e-01 8.63017380e-01 6.98924482e-01
6.84294701e-01 4.86776531e-01 7.66303577e-03 2.81197697e-01
1.98506236e-01 -4.24999774e-01 3.81977350e-01 7.24257350e-01
1.29435658e-02 -3.69448513e-01 -1.34071493e+00 1.61834121e-01
-1.62378550e+00 -9.19530988e-01 -4.19606864e-01 1.91343820e+00
8.58726978e-01 6.63223326e-01 2.22597480e-01 3.79144073e-01
6.97508395e-01 -2.08491534e-01 -3.36278491e-02 -4.44214702e-01
-3.68372858e-01 2.12275013e-01 8.41689944e-01 -1.97002888e-01
-1.61499906e+00 8.36056590e-01 5.77947950e+00 9.05630589e-01
-6.63594544e-01 -5.90693988e-02 8.71300697e-01 2.54341453e-01
-1.98732287e-01 -1.80160657e-01 -1.68210328e+00 5.52705169e-01
1.16006649e+00 7.53203332e-02 1.11620069e-01 1.09572363e+00
-7.39231631e-02 -7.59928897e-02 -9.43227470e-01 7.10873485e-01
2.02749208e-01 -1.96985853e+00 7.51715945e-03 2.59462655e-01
4.86499429e-01 5.49271703e-03 -3.34364884e-02 6.81166947e-01
5.19389629e-01 -8.36465716e-01 1.11175275e+00 4.55245972e-01
8.22270036e-01 -5.86256683e-01 7.59385467e-01 3.38712126e-01
-1.20674539e+00 -4.38262522e-01 -3.22054148e-01 2.81804115e-01
-4.63554829e-01 5.39662898e-01 -8.23150754e-01 8.99220705e-01
1.23668492e+00 2.63279140e-01 -1.11536026e+00 7.60661781e-01
8.83765891e-02 8.07499111e-01 -2.53895909e-01 -2.72700191e-01
1.67815670e-01 -9.02439058e-02 3.69882703e-01 1.53399122e+00
1.02063037e-01 -2.93755941e-02 4.33568716e-01 7.76297510e-01
-3.46940577e-01 5.41702151e-01 -6.71932697e-01 -4.70683016e-02
7.55901635e-01 1.34670615e+00 -1.31781423e+00 -5.87534845e-01
-5.44649780e-01 3.44312608e-01 3.55651230e-01 7.89944008e-02
-8.07486475e-01 -4.50315118e-01 -8.96360278e-02 2.71985143e-01
7.18882084e-01 1.60642445e-01 -9.40171659e-01 -1.11537457e+00
3.46445113e-01 -9.03747141e-01 9.05567527e-01 -6.19161248e-01
-1.42987108e+00 6.69331551e-01 -3.85286361e-02 -1.15189826e+00
1.27548635e-01 -8.29044819e-01 -2.65509337e-01 6.07962728e-01
-1.11223459e+00 -1.11833870e+00 -1.55551866e-01 5.06480813e-01
3.53119880e-01 -4.11596745e-01 6.81517541e-01 3.17367285e-01
-8.26648533e-01 7.54635930e-01 5.20124018e-01 7.81206310e-01
9.11882937e-01 -1.60047412e+00 3.35374892e-01 7.19038546e-01
3.33455503e-01 6.17780387e-01 4.49089378e-01 -9.51698601e-01
-1.75773132e+00 -1.05790699e+00 1.32998049e+00 -1.31514013e+00
9.26517785e-01 -1.08401859e+00 -7.56274104e-01 6.35701060e-01
3.66511866e-02 1.22326449e-01 6.31449997e-01 1.21485867e-01
-7.04386294e-01 -2.25647748e-01 -1.46607816e+00 1.81515902e-01
9.64853823e-01 -4.88606095e-01 -8.19543660e-01 7.56989479e-01
4.93592441e-01 -5.40831804e-01 -1.24915242e+00 2.23677889e-01
5.51582754e-01 -7.66933262e-01 9.68036771e-01 -5.67240059e-01
4.62232292e-01 -1.97452113e-01 -3.13865930e-01 -4.39140141e-01
-2.04115793e-01 -4.55546886e-01 -1.73690185e-01 1.86912870e+00
1.05646050e+00 -4.00162369e-01 8.05918813e-01 5.99698603e-01
-1.34565026e-01 -5.18951595e-01 -7.97100306e-01 -8.06344926e-01
4.92581457e-01 -2.81431526e-01 4.98562336e-01 9.47014093e-01
1.23365663e-01 2.96911687e-01 1.22245379e-01 2.21874923e-01
4.09066886e-01 4.37145494e-02 8.38344812e-01 -1.56031680e+00
2.06771895e-01 -3.18476558e-01 2.70038154e-02 -5.84508300e-01
5.64276800e-03 -1.10328257e+00 -2.46652007e-01 -1.69930184e+00
3.65011305e-01 -6.45269692e-01 -3.47557336e-01 4.37479854e-01
6.80174679e-02 1.29223600e-01 9.71444771e-02 3.87655348e-01
-1.07240522e+00 -2.89152294e-01 8.33619118e-01 -1.91206217e-01
-1.23516478e-01 -2.78633595e-01 -1.08773780e+00 5.76068699e-01
3.96275401e-01 -5.81230223e-01 2.69315481e-01 7.03317374e-02
4.23753738e-01 1.31545305e-01 -9.55848545e-02 -1.11796594e+00
4.36282009e-01 -7.88679197e-02 1.15400839e+00 -1.16420102e+00
-2.80650854e-01 -5.82924247e-01 -3.78052890e-01 -8.51747245e-02
-4.97713089e-01 3.21964055e-01 2.07698211e-01 6.17949128e-01
-6.96241483e-02 -9.06857327e-02 2.73578435e-01 -4.13276315e-01
-2.80305862e-01 -4.00302820e-02 -5.77284873e-01 5.52578866e-01
7.70924151e-01 -1.11313254e-01 -8.17125916e-01 1.39869556e-01
-5.57431221e-01 4.86522987e-02 2.26025045e-01 4.45441157e-01
-3.44127901e-02 -8.38759124e-01 -6.97194755e-01 4.61360998e-02
3.64567280e-01 -2.35589638e-01 -1.98224500e-01 7.54617155e-01
-3.76321644e-01 9.93180692e-01 -3.88943478e-02 -4.17410642e-01
-9.53020096e-01 6.23745382e-01 -4.89200056e-01 -8.09776545e-01
-6.68066084e-01 6.27576888e-01 -4.54039186e-01 -8.00004423e-01
4.58394378e-01 -4.54111516e-01 -6.38106227e-01 4.33148921e-01
5.83568513e-01 2.29073629e-01 7.53852785e-01 -3.71154845e-01
-6.31460607e-01 1.05348498e-01 -3.03740352e-01 -2.41299309e-02
1.46034217e+00 6.95254747e-03 -4.05150056e-02 3.23657751e-01
7.13504076e-01 6.30715907e-01 -6.69821143e-01 -8.87330994e-02
6.81201220e-01 -2.72483528e-01 -2.29408681e-01 -1.65803456e+00
-5.11714756e-01 2.50256062e-01 2.23067731e-01 4.31500673e-01
5.77333331e-01 1.80872992e-01 4.19654936e-01 5.25518179e-01
3.42753172e-01 -1.53624725e+00 -1.08903646e-01 4.95286822e-01
8.06880832e-01 -1.33048189e+00 2.86383361e-01 -7.10403860e-01
-6.43247128e-01 1.40619707e+00 5.65245807e-01 5.96807338e-02
5.12040854e-01 7.27465987e-01 -1.67912692e-02 -9.40570012e-02
-8.32314074e-01 -5.95851056e-03 4.18214947e-01 5.14609277e-01
3.60585690e-01 -7.62686431e-02 -5.00747442e-01 1.34243453e+00
-4.57108468e-01 -1.11922316e-01 5.82514703e-01 1.09340799e+00
-4.10553634e-01 -1.06299078e+00 -9.37371373e-01 8.66201460e-01
-7.83654213e-01 -7.80599415e-02 -7.94363797e-01 1.11830735e+00
2.52032310e-01 8.38447332e-01 6.14311397e-02 -2.43569940e-01
3.72096300e-01 4.85678226e-01 3.02844524e-01 -5.38452327e-01
-1.11540973e+00 1.81215461e-02 6.15304530e-01 -3.16242933e-01
1.50411248e-01 -1.06287503e+00 -1.24534059e+00 -1.37259468e-01
-8.33497494e-02 3.50420445e-01 7.19644427e-01 8.04783046e-01
3.78475159e-01 2.45951906e-01 3.70504349e-01 -1.25167623e-01
-6.94727838e-01 -9.95932400e-01 -4.83504266e-01 3.07560146e-01
-2.95497447e-01 -8.51972163e-01 -3.71558696e-01 1.54009685e-01] | [9.639230728149414, 8.003806114196777] |
39a3dc4b-cc69-4e93-9157-1fca5ce6ba7d | weakly-and-semi-supervised-object-detection | 1702.08740 | null | http://arxiv.org/abs/1702.08740v1 | http://arxiv.org/pdf/1702.08740v1.pdf | Weakly- and Semi-Supervised Object Detection with Expectation-Maximization Algorithm | Object detection when provided image-level labels instead of instance-level
labels (i.e., bounding boxes) during training is an important problem in
computer vision, since large scale image datasets with instance-level labels
are extremely costly to obtain. In this paper, we address this challenging
problem by developing an Expectation-Maximization (EM) based object detection
method using deep convolutional neural networks (CNNs). Our method is
applicable to both the weakly-supervised and semi-supervised settings.
Extensive experiments on PASCAL VOC 2007 benchmark show that (1) in the weakly
supervised setting, our method provides significant detection performance
improvement over current state-of-the-art methods, (2) having access to a small
number of strongly (instance-level) annotated images, our method can almost
match the performace of the fully supervised Fast RCNN. We share our source
code at https://github.com/ZiangYan/EM-WSD. | ['Chang-Shui Zhang', 'Weishen Pan', 'Ziang Yan', 'Jin Li', 'Jian Liang'] | 2017-02-28 | null | null | null | null | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 2.89758801e-01 1.10158874e-02 -1.79547578e-01 -4.09875512e-01
-8.42715383e-01 -6.91567719e-01 4.95773196e-01 -9.85922385e-03
-8.90532196e-01 3.66095632e-01 -4.95623827e-01 -4.49449837e-01
4.18022305e-01 -5.99003553e-01 -1.10108447e+00 -6.47037029e-01
5.26807047e-02 2.85307020e-01 5.38358212e-01 3.00216794e-01
1.50987715e-01 4.37504113e-01 -1.56619918e+00 4.06283230e-01
4.09146965e-01 1.14042854e+00 4.01578754e-01 8.22937965e-01
8.51090029e-02 8.35194707e-01 -4.55363214e-01 -3.06066364e-01
5.45406997e-01 1.96729545e-02 -9.13502932e-01 2.56032288e-01
9.51714635e-01 -4.28989410e-01 -1.47374898e-01 1.35517240e+00
3.13443393e-01 -1.65568560e-01 5.44788599e-01 -1.27484941e+00
-4.40154135e-01 3.50010335e-01 -9.46010649e-01 2.75888324e-01
-3.07443440e-01 3.56521070e-01 9.71427143e-01 -1.21273518e+00
5.05597889e-01 9.69977736e-01 7.71517098e-01 6.47362769e-01
-1.35452390e+00 -7.57951379e-01 2.71500826e-01 1.78585455e-01
-1.49314976e+00 -2.43139416e-01 5.14887452e-01 -4.96855974e-01
8.26274931e-01 -2.56801192e-02 2.90037900e-01 9.38926220e-01
-3.68711889e-01 1.14879930e+00 1.08428454e+00 -4.13014770e-01
6.57320255e-03 3.72433454e-01 4.57582362e-02 8.81318867e-01
4.19597924e-01 1.55278400e-01 -1.49670094e-02 9.74381864e-02
7.21935809e-01 5.17743602e-02 -3.85036990e-02 -5.43053508e-01
-1.31313109e+00 7.86591530e-01 7.03887939e-01 1.40329212e-01
-7.88528323e-02 3.09418708e-01 4.29493487e-01 3.62638198e-02
4.93487835e-01 2.86737025e-01 -6.57447517e-01 2.42246568e-01
-9.98344243e-01 -4.75393748e-03 7.22284138e-01 1.12788486e+00
6.79293871e-01 -1.20749354e-01 -1.78201571e-01 8.02049279e-01
1.24493532e-01 3.96544039e-01 4.45098877e-02 -8.13832343e-01
3.61040086e-01 6.08544469e-01 2.76416153e-01 -6.10098898e-01
-4.56034601e-01 -7.54547119e-01 -8.29677880e-01 3.93481165e-01
7.58299351e-01 -2.41448700e-01 -9.52795863e-01 1.57657862e+00
5.38843751e-01 2.61157870e-01 -2.83998609e-01 8.44444990e-01
8.99434149e-01 4.85933155e-01 8.83922279e-02 1.21339619e-01
1.40406120e+00 -1.25174427e+00 -1.93846717e-01 -4.20605689e-01
7.98014343e-01 -8.33866060e-01 9.53195393e-01 3.58692467e-01
-8.59147370e-01 -8.18894207e-01 -9.95516419e-01 -9.58084017e-02
-4.45490479e-01 8.29234481e-01 5.14340997e-01 5.54172754e-01
-9.79876161e-01 4.31020081e-01 -8.74357104e-01 -3.06909651e-01
8.62406492e-01 2.80024320e-01 -3.58777106e-01 -1.80088237e-01
-5.57165504e-01 6.67429388e-01 5.84513307e-01 1.90516710e-01
-1.17372060e+00 -6.53822660e-01 -7.74861574e-01 -1.28909379e-01
6.73240185e-01 -2.73963839e-01 1.51790833e+00 -1.04060411e+00
-9.51904118e-01 1.39578593e+00 -1.12762220e-01 -4.73261595e-01
9.44913745e-01 -4.02487785e-01 1.83913440e-01 9.79671851e-02
1.74985304e-01 1.07918799e+00 7.16879189e-01 -1.29863322e+00
-1.17331076e+00 -1.29796878e-01 8.47793669e-02 -2.13659480e-01
-4.53704596e-01 1.82205662e-01 -5.85252285e-01 -4.22591686e-01
-1.12272553e-01 -8.77628744e-01 -3.09674323e-01 5.99324524e-01
-4.98149395e-01 -5.65239906e-01 9.41301167e-01 -3.59989375e-01
7.57500291e-01 -2.16801739e+00 -2.45655805e-01 -1.68480664e-01
2.59063601e-01 6.22985601e-01 -8.79821032e-02 -4.52355333e-02
-7.05509856e-02 -7.13405013e-03 -3.86575699e-01 -6.09819293e-01
-5.54226749e-02 3.44582163e-02 -6.51572272e-02 7.56094038e-01
4.42566633e-01 9.87841249e-01 -1.06939781e+00 -7.07532585e-01
3.31835210e-01 3.04808766e-01 -3.47953707e-01 2.95184076e-01
-2.57849395e-01 2.87511528e-01 -2.04537377e-01 7.95129895e-01
7.17068255e-01 -7.52733648e-01 -7.35430419e-02 -3.47507328e-01
-2.46841043e-01 -8.85786042e-02 -1.36337125e+00 1.52239573e+00
-3.50175112e-01 9.52669442e-01 1.90629274e-01 -1.11566353e+00
6.04392588e-01 1.95710286e-02 8.11056867e-02 -3.89966756e-01
2.16738150e-01 2.01550663e-01 -9.31826606e-02 -3.38924319e-01
1.19845353e-01 2.84217864e-01 2.08404079e-01 3.47803414e-01
2.55688816e-01 8.75087008e-02 4.72414285e-01 2.01351970e-01
1.00908577e+00 1.37832657e-01 4.46559876e-01 -2.82844007e-01
4.76391077e-01 2.08009124e-01 5.90133786e-01 1.16917157e+00
-4.46441025e-01 6.52187109e-01 4.20056492e-01 -5.56894779e-01
-1.15056384e+00 -8.28462899e-01 -5.85532963e-01 1.25745332e+00
1.35990486e-01 -1.71113566e-01 -9.40894306e-01 -8.38092804e-01
8.83410498e-03 3.36631566e-01 -7.04453528e-01 1.16915882e-01
-5.16713142e-01 -7.34597266e-01 6.60438240e-01 9.01861846e-01
7.37955749e-01 -9.65889037e-01 -6.14479005e-01 8.89564231e-02
1.73566435e-02 -1.51392817e+00 -4.73256290e-01 2.95455635e-01
-7.18252242e-01 -1.33656871e+00 -7.15299666e-01 -1.02600992e+00
9.30984795e-01 4.96565908e-01 1.23426974e+00 3.73277456e-01
-8.03418458e-01 1.89404503e-01 -2.78675824e-01 -6.44348025e-01
-1.69648930e-01 -9.56525803e-02 -1.14996113e-01 -3.48500125e-02
6.12398982e-01 -1.02165574e-02 -8.52153540e-01 6.03618205e-01
-8.28258336e-01 1.19921803e-01 8.14935088e-01 8.29926610e-01
7.10935771e-01 -1.71566308e-01 3.40652972e-01 -1.07845783e+00
-1.05535001e-01 -1.09134130e-01 -1.16262686e+00 2.35470176e-01
-5.08341014e-01 -2.24275708e-01 4.51635003e-01 -5.45963705e-01
-8.76892269e-01 6.32593215e-01 -1.39977500e-01 -3.18858951e-01
-5.37384748e-01 -1.09227493e-01 4.91377041e-02 -2.35284612e-01
7.31769800e-01 1.98195711e-01 -2.73589045e-01 -5.00433147e-01
4.22800481e-01 8.24742854e-01 6.71439290e-01 -5.49026310e-01
8.67143273e-01 7.88515925e-01 -1.66448444e-01 -5.73218584e-01
-1.34942412e+00 -9.37675655e-01 -9.69896615e-01 -2.80985266e-01
9.38330054e-01 -1.14309502e+00 -6.35081828e-01 5.30179501e-01
-1.14216352e+00 -5.75041950e-01 -1.14510790e-01 3.75653327e-01
-3.50814462e-01 2.80795395e-01 -8.24930847e-01 -9.27984357e-01
-2.97967881e-01 -1.14398313e+00 1.30754185e+00 2.83525676e-01
1.91769212e-01 -7.18600035e-01 -3.43040735e-01 5.14586985e-01
1.74615994e-01 3.53918344e-01 2.09607020e-01 -6.26790822e-01
-7.01245546e-01 -4.51019377e-01 -8.22741866e-01 6.16228819e-01
-2.33724594e-01 1.95483435e-02 -1.15621054e+00 -4.14234251e-01
-4.46030438e-01 -7.00127721e-01 1.24827719e+00 2.94378787e-01
1.53556013e+00 -4.52178456e-02 -4.19623196e-01 6.21067882e-01
1.40399826e+00 -4.06815737e-01 2.78874606e-01 3.32226098e-01
6.87003076e-01 5.39785445e-01 7.88140893e-01 3.42903852e-01
2.17062131e-01 4.58741903e-01 6.63823724e-01 -4.35020208e-01
-1.54395059e-01 -1.52154967e-01 1.92642525e-01 -2.00145617e-02
4.46451455e-02 8.38420074e-03 -9.77485240e-01 7.64002562e-01
-2.05886817e+00 -8.21612477e-01 -4.24155653e-01 2.12206078e+00
9.65988338e-01 4.38019812e-01 2.77530909e-01 -1.18093938e-01
9.91215467e-01 -9.64322612e-02 -6.38452828e-01 1.43965721e-01
5.06911725e-02 2.51657404e-02 8.36247027e-01 1.62372366e-01
-1.68283081e+00 1.01885021e+00 5.39408779e+00 9.93590355e-01
-8.73431325e-01 4.07672346e-01 8.92478406e-01 -1.82043448e-01
7.08519220e-01 -2.49457806e-01 -1.11683130e+00 3.82029086e-01
6.14178181e-01 4.09121573e-01 5.41059934e-02 1.42733645e+00
-2.95237433e-02 -5.88186234e-02 -1.38775575e+00 1.04554224e+00
5.03083840e-02 -1.36738110e+00 -3.98232311e-01 -1.89839453e-02
9.32029903e-01 4.69084918e-01 5.08445054e-02 3.94205451e-01
4.31631893e-01 -1.04539776e+00 7.99396753e-01 -8.09005648e-02
9.31513727e-01 -5.18406868e-01 9.29607570e-01 6.36980951e-01
-1.33773351e+00 -7.01700896e-02 -6.49634361e-01 2.18054086e-01
-4.81435321e-02 6.82368994e-01 -6.97550714e-01 -1.22223031e-02
1.07824159e+00 6.40718818e-01 -7.36377478e-01 1.19517910e+00
-6.31413341e-01 8.57713878e-01 -4.52935129e-01 -6.17743619e-02
3.31448078e-01 2.13680312e-01 2.15921909e-01 1.64588249e+00
-7.27740824e-02 -1.25260472e-01 3.20319772e-01 1.19144773e+00
-5.19946933e-01 -1.87557608e-01 -2.73971170e-01 2.22487465e-01
2.70983398e-01 1.74306440e+00 -1.03339553e+00 -4.43987936e-01
-5.67164779e-01 9.36951935e-01 5.90834796e-01 2.22161680e-01
-7.89888799e-01 -2.86304176e-01 3.48574668e-01 -3.99241596e-02
5.65430641e-01 -1.75933957e-01 -1.85035646e-01 -1.00050414e+00
1.67346373e-01 -5.64319670e-01 4.58079368e-01 -4.82025027e-01
-1.33235061e+00 3.74954700e-01 -1.77993283e-01 -1.15454340e+00
1.80897057e-01 -1.27375853e+00 -6.76696599e-01 4.62924749e-01
-1.76442420e+00 -1.30693531e+00 -4.49885696e-01 3.81132722e-01
6.83846593e-01 1.49708539e-01 4.05413717e-01 4.56644028e-01
-6.69690371e-01 6.98774755e-01 1.14002831e-01 8.17528367e-01
6.93969309e-01 -1.54688346e+00 2.99933255e-01 1.00188160e+00
5.27398169e-01 4.20807123e-01 5.10942280e-01 -1.98037788e-01
-1.13473654e+00 -1.49555194e+00 4.78733450e-01 -7.16135442e-01
5.57680428e-01 -6.01135612e-01 -7.30810285e-01 7.35832512e-01
-4.63618487e-02 6.99096382e-01 5.04294276e-01 4.02330533e-02
-6.33121192e-01 5.18964566e-02 -1.07634234e+00 1.96056068e-01
1.15287876e+00 -3.25431168e-01 -3.42111945e-01 8.88973475e-01
5.53766072e-01 -4.67551678e-01 -5.48107862e-01 5.65235078e-01
4.43045557e-01 -7.74409413e-01 1.00727952e+00 -5.80772221e-01
3.42896909e-01 -5.27728021e-01 -1.89452455e-03 -8.37900460e-01
-1.31641328e-01 -1.40123606e-01 -1.55285090e-01 1.03310716e+00
5.92500627e-01 -5.33085585e-01 8.64958286e-01 2.90851116e-01
6.48702830e-02 -7.59151876e-01 -8.17318559e-01 -9.50522661e-01
9.12633613e-02 -4.22755480e-01 2.31738258e-02 7.71452308e-01
-5.48815310e-01 -3.59048659e-04 -2.97900021e-01 4.42985177e-01
1.03099346e+00 2.95316547e-01 9.33522105e-01 -1.15272105e+00
-4.45569009e-01 -5.10288358e-01 -4.74031985e-01 -1.21447253e+00
9.42286104e-02 -8.42004895e-01 3.63490820e-01 -1.43625593e+00
6.50016963e-01 -3.29222530e-01 -4.52232361e-01 7.31471956e-01
-2.71470308e-01 8.03210616e-01 1.27868727e-01 1.99185118e-01
-1.17063177e+00 4.91874851e-02 8.71059000e-01 -9.88644585e-02
2.92407334e-01 6.66007176e-02 -3.52318853e-01 9.45024073e-01
8.73663127e-01 -7.24090099e-01 2.01208889e-01 -3.70218009e-01
6.70330375e-02 -5.46294451e-01 7.04555571e-01 -9.52204645e-01
4.42244560e-01 8.46534744e-02 5.94868481e-01 -7.08677948e-01
3.90121266e-02 -7.47202516e-01 -4.58876789e-01 6.75902247e-01
-4.22414988e-01 -3.87464255e-01 2.35355303e-01 6.71830297e-01
-5.39940372e-02 -3.74838233e-01 1.21418417e+00 -2.30846807e-01
-1.02584779e+00 4.38745230e-01 -3.19404975e-02 2.06240505e-01
1.18558276e+00 -4.78809215e-02 -4.18225706e-01 1.21367369e-02
-5.01846790e-01 5.38666606e-01 3.10700685e-01 3.42108250e-01
4.28401202e-01 -9.71663237e-01 -8.26592743e-01 -1.20651431e-01
5.11545479e-01 3.98561895e-01 -9.02055502e-02 8.35703075e-01
-4.57530260e-01 3.31933975e-01 1.80337355e-01 -1.13515842e+00
-1.51302981e+00 4.43833590e-01 3.65197629e-01 -1.11927934e-01
-5.54229558e-01 1.33712053e+00 4.99619633e-01 -7.11686194e-01
5.96141636e-01 -1.66147828e-01 1.35783076e-01 -2.40281031e-01
8.68036211e-01 2.07340553e-01 -2.71980260e-02 -4.40937042e-01
-3.44268650e-01 3.56093436e-01 -3.64898652e-01 2.97587425e-01
1.33124459e+00 2.32609183e-01 4.18917872e-02 1.13320738e-01
1.24521470e+00 -5.10288775e-01 -1.68868732e+00 -5.06122351e-01
1.69382289e-01 -5.30027211e-01 2.26032764e-01 -6.10633016e-01
-1.08554041e+00 1.05189395e+00 8.18318903e-01 -9.68899429e-02
8.72304201e-01 4.43629503e-01 5.20864308e-01 7.33002424e-01
3.27507496e-01 -1.23765624e+00 2.20256999e-01 2.59683698e-01
5.41829467e-01 -1.79603732e+00 -2.83905603e-02 -3.31811041e-01
-2.80159414e-01 1.09467864e+00 9.98441100e-01 -2.00326368e-01
5.97238421e-01 3.43460083e-01 5.44168800e-02 -1.20798387e-01
-5.90052843e-01 -6.39889479e-01 3.45722854e-01 3.53076309e-01
4.79041517e-01 3.55232321e-02 -2.88336817e-02 2.71824330e-01
3.05700123e-01 -1.90625653e-01 3.74414533e-01 1.02045643e+00
-7.78817177e-01 -9.06464875e-01 -3.72138560e-01 6.05945945e-01
-5.92216790e-01 -1.41622752e-01 -4.33874547e-01 9.25312757e-01
3.21068347e-01 9.89929676e-01 7.17151612e-02 2.23752726e-02
1.52338520e-01 -1.60287336e-01 4.01728183e-01 -8.74649882e-01
-3.37744415e-01 5.03490716e-02 -1.29857510e-01 -7.04057813e-01
-5.12400031e-01 -5.17504930e-01 -1.20668471e+00 -9.48673580e-04
-6.86590850e-01 -2.58254856e-01 7.71945953e-01 7.64104128e-01
1.82019025e-01 3.93499672e-01 3.60047311e-01 -9.07494545e-01
-7.30426192e-01 -1.06772208e+00 -3.73812914e-01 5.63887656e-01
3.60114813e-01 -6.19824886e-01 -5.00864983e-01 1.31753370e-01] | [9.23666763305664, 1.0668950080871582] |
583bbb7e-96c9-4c1d-a019-191710bb5ded | interpretable-visual-understanding-with | 2108.02924 | null | https://arxiv.org/abs/2108.02924v2 | https://arxiv.org/pdf/2108.02924v2.pdf | Interpretable Visual Understanding with Cognitive Attention Network | While image understanding on recognition-level has achieved remarkable advancements, reliable visual scene understanding requires comprehensive image understanding on recognition-level but also cognition-level, which calls for exploiting the multi-source information as well as learning different levels of understanding and extensive commonsense knowledge. In this paper, we propose a novel Cognitive Attention Network (CAN) for visual commonsense reasoning to achieve interpretable visual understanding. Specifically, we first introduce an image-text fusion module to fuse information from images and text collectively. Second, a novel inference module is designed to encode commonsense among image, query and response. Extensive experiments on large-scale Visual Commonsense Reasoning (VCR) benchmark dataset demonstrate the effectiveness of our approach. The implementation is publicly available at https://github.com/tanjatang/CAN | ['Eirini Ntoutsi', 'Mengyu Wang', 'Tyler Derr', 'Kea Turner', 'Yi Yu', 'Wenbin Zhang', 'Xuejiao Tang'] | 2021-08-06 | null | null | null | null | ['visual-commonsense-reasoning'] | ['reasoning'] | [ 4.47915643e-01 6.40200749e-02 -1.14324994e-01 -4.93189514e-01
-4.83034700e-01 -4.08674955e-01 6.95806503e-01 1.08469330e-01
-1.15459487e-01 4.37117577e-01 4.76412684e-01 -4.60373014e-01
5.48456907e-02 -7.18697429e-01 -8.54654729e-01 -1.38142869e-01
7.59605944e-01 2.94331294e-02 4.41328138e-02 -2.06990525e-01
3.77694368e-01 1.86616451e-01 -1.46748149e+00 6.85839891e-01
1.17275286e+00 1.09672606e+00 4.32045400e-01 6.41390800e-01
-2.70663857e-01 1.59302902e+00 -4.44337547e-01 -5.28975785e-01
-1.14880599e-01 -5.62204957e-01 -1.07246172e+00 3.09164643e-01
4.98867393e-01 -5.06763339e-01 -6.08411789e-01 1.51115191e+00
8.04075673e-02 1.74362361e-01 4.43607181e-01 -1.40675879e+00
-1.75911415e+00 5.93336463e-01 -3.45199525e-01 4.73826110e-01
4.38505143e-01 6.55239224e-01 8.58450532e-01 -7.86673605e-01
4.56177592e-01 1.51022363e+00 1.05634080e-02 4.97813761e-01
-9.55869734e-01 -7.15430319e-01 1.96241796e-01 9.23591733e-01
-1.15407372e+00 -3.30886155e-01 9.83153403e-01 -4.82873559e-01
8.48903060e-01 3.01102400e-01 8.41040015e-01 1.10330057e+00
-1.03122987e-01 1.11482191e+00 1.51380110e+00 -2.09059238e-01
5.10452734e-03 1.26064554e-01 4.26909626e-01 8.95169616e-01
3.42829317e-01 -3.71808037e-02 -3.58554929e-01 4.21971053e-01
7.59157360e-01 4.50133115e-01 -4.10801083e-01 1.46907270e-01
-1.30230403e+00 7.08355844e-01 1.04499424e+00 2.49834538e-01
-5.16145408e-01 4.46842134e-01 3.86199027e-01 1.35049433e-01
7.30329156e-02 2.51923859e-01 1.30050287e-01 8.41862783e-02
-4.57895666e-01 -1.43868074e-01 5.50482392e-01 9.48570967e-01
7.55770087e-01 9.12275165e-02 -4.45054114e-01 7.22002983e-01
3.39344144e-01 8.45620513e-01 3.64739507e-01 -1.04744136e+00
4.61702198e-01 8.06847095e-01 -3.00045878e-01 -1.45554197e+00
1.23199495e-02 -1.08385742e-01 -1.19169879e+00 9.81259421e-02
3.61320563e-03 3.28471303e-01 -1.19527781e+00 1.58435917e+00
-1.07208930e-03 2.02717260e-01 2.07931995e-01 1.21683800e+00
1.31486666e+00 5.75673044e-01 3.04061383e-01 9.49488953e-02
1.62554967e+00 -9.98987854e-01 -8.70574474e-01 -4.38322961e-01
-3.97540554e-02 -4.65656877e-01 1.43567204e+00 1.03766829e-01
-9.78800297e-01 -5.43078482e-01 -1.01649988e+00 -7.39175141e-01
-7.20399201e-01 -3.93408909e-02 6.05101883e-01 1.83152378e-01
-8.90099287e-01 -5.20218536e-02 -3.86664689e-01 -2.83692181e-01
1.05322635e+00 -1.53109089e-01 -1.87660515e-01 -5.76667547e-01
-1.24568892e+00 9.53670382e-01 7.22549617e-01 2.10656568e-01
-1.16810477e+00 -4.50681329e-01 -9.23770964e-01 1.64473712e-01
6.34771407e-01 -9.59113538e-01 9.43032861e-01 -1.29474890e+00
-1.20961702e+00 1.08454359e+00 -1.89498574e-01 -3.23732823e-01
3.63509238e-01 -1.28088400e-01 -4.16181862e-01 3.66492420e-01
-4.06515738e-03 7.56988049e-01 7.67210126e-01 -1.58992481e+00
-2.67549813e-01 -5.13746500e-01 4.88986313e-01 2.12505937e-01
-1.67854652e-01 -1.60550088e-01 -4.80130196e-01 -6.15642190e-01
-4.64412458e-02 -5.22306323e-01 1.59272864e-01 5.62717356e-02
-6.20123386e-01 -2.50634342e-01 6.48230016e-01 -1.07139468e+00
9.77970541e-01 -2.02024698e+00 1.82620600e-01 3.53672775e-03
5.40706873e-01 2.27100879e-01 -2.24574715e-01 2.25127339e-01
-2.15371460e-01 3.32793325e-01 -2.25983173e-01 -1.80304814e-02
1.16358571e-01 3.20812792e-01 -6.55859053e-01 8.58635977e-02
3.40656102e-01 1.44568408e+00 -1.05432737e+00 -5.25704801e-01
4.32527304e-01 4.42298204e-01 -3.96501482e-01 3.01272385e-02
-4.74423975e-01 1.73526436e-01 -5.27548611e-01 9.05970097e-01
6.43810987e-01 -8.38735938e-01 -1.15328342e-01 -4.63832259e-01
2.01049805e-01 -5.35269007e-02 -5.07380962e-01 1.72857082e+00
-4.37630147e-01 9.17944014e-01 -4.15358901e-01 -9.67963517e-01
5.37044942e-01 -7.55288601e-02 -2.56900370e-01 -1.00679553e+00
3.71123075e-01 -1.92005545e-01 -1.59777552e-02 -7.21401155e-01
3.02113265e-01 -3.03361863e-01 -3.93832102e-02 3.26017141e-01
-7.20411092e-02 -1.77933499e-01 -4.85619009e-02 6.96658194e-01
8.02246451e-01 -3.26560706e-01 4.68471080e-01 2.31139325e-02
6.51198983e-01 3.18203539e-01 2.15942815e-01 6.73273742e-01
-4.68026131e-01 2.24175736e-01 4.62650239e-01 -2.94715732e-01
-7.59543061e-01 -1.17109859e+00 2.06670642e-01 8.58977556e-01
6.94736719e-01 -2.16793746e-01 -5.51943064e-01 -3.85457754e-01
1.46297008e-01 1.13337100e+00 -8.84071648e-01 -3.30983877e-01
5.90997934e-02 -2.19407231e-01 5.10141551e-01 5.74997604e-01
1.07623935e+00 -1.34273827e+00 -5.24490714e-01 -3.08042079e-01
-5.80703080e-01 -1.26980412e+00 -3.66117299e-01 -3.40382993e-01
-5.24880707e-01 -1.23534155e+00 -2.12629601e-01 -5.30771315e-01
6.43299460e-01 7.15180337e-01 1.20446718e+00 5.20215988e-01
-6.91482663e-01 6.04129016e-01 -4.40889567e-01 -5.33310235e-01
-3.21966112e-01 -6.70213342e-01 -4.67796981e-01 -5.46423681e-02
5.23109674e-01 -3.69533479e-01 -4.98845011e-01 -2.16010407e-01
-1.17089820e+00 5.36853313e-01 7.30131447e-01 8.77391040e-01
6.08744085e-01 6.62773401e-02 4.51869726e-01 -4.63094145e-01
8.23844910e-01 -6.13855958e-01 -3.00413191e-01 5.92382610e-01
-3.59308362e-01 -2.22200751e-02 4.32474256e-01 -2.35113576e-01
-1.35500252e+00 -3.04718494e-01 1.90918282e-01 -8.07045102e-01
-4.31795150e-01 4.80312288e-01 -3.80168974e-01 1.27780393e-01
4.33358788e-01 6.92693710e-01 -7.09152818e-02 -4.00951356e-02
7.72859156e-01 6.84106052e-01 7.64346600e-01 -5.08412838e-01
6.86503112e-01 7.15674639e-01 -3.57979566e-01 -6.24037027e-01
-1.20948637e+00 -3.23890299e-01 -3.84069949e-01 -3.95340890e-01
1.21803212e+00 -1.03328240e+00 -1.09587324e+00 3.17579716e-01
-1.53212380e+00 -2.24271134e-01 -1.15320444e-01 1.47694096e-01
-5.97091138e-01 4.40475523e-01 -4.40071285e-01 -5.97890377e-01
-3.47678632e-01 -9.92236197e-01 8.35722208e-01 3.76082629e-01
1.74875498e-01 -9.63204443e-01 -4.03821915e-01 1.06515491e+00
2.46324360e-01 2.53240258e-01 9.31798339e-01 -3.88200909e-01
-1.10487509e+00 7.06747994e-02 -1.11209655e+00 4.56943274e-01
2.06597168e-02 -3.04975033e-01 -8.68430436e-01 2.21312746e-01
-4.83892262e-02 -8.18129182e-01 1.32531238e+00 7.26116970e-02
1.66974461e+00 -3.95060629e-01 -9.76352207e-03 4.15671736e-01
1.68140793e+00 1.64580628e-01 9.70096827e-01 2.28672415e-01
1.04525733e+00 3.50490779e-01 3.38713408e-01 3.22227895e-01
9.18630958e-01 1.85095310e-01 6.94205880e-01 -1.50576562e-01
-5.04887044e-01 -2.49692619e-01 3.11603904e-01 6.07718587e-01
-1.96504563e-01 -1.45771310e-01 -1.07312238e+00 4.82220203e-01
-1.93129134e+00 -1.54556859e+00 -1.47506427e-02 1.49752164e+00
8.67395997e-01 -2.63891965e-01 -3.60967577e-01 -2.15581208e-01
6.71279371e-01 1.56531364e-01 -9.66177523e-01 -3.30535829e-01
-2.56850123e-01 -2.07197443e-01 1.05655268e-01 5.36094487e-01
-8.43671024e-01 1.26519430e+00 5.55440569e+00 9.34235454e-01
-8.40798438e-01 2.72613466e-01 5.87686181e-01 -4.88584209e-03
-6.74521089e-01 5.34133008e-03 -1.55283555e-01 2.81303227e-01
2.75148243e-01 -4.04080480e-01 8.48790348e-01 6.95909858e-01
-7.56790536e-03 -1.64478779e-01 -7.96511710e-01 1.35131276e+00
5.40109813e-01 -1.50292766e+00 7.05771267e-01 -1.43264353e-01
5.90039313e-01 -1.87651010e-03 9.71259698e-02 2.73712009e-01
4.92499620e-01 -1.20998847e+00 6.84279025e-01 9.40733135e-01
6.38386369e-01 -5.94417691e-01 4.80754167e-01 3.09578508e-01
-1.02195859e+00 -9.66656357e-02 -4.21739608e-01 -1.56657934e-01
-3.97326797e-02 6.01288497e-01 -4.89038527e-01 5.02303302e-01
6.52746975e-01 1.03816700e+00 -8.91028345e-01 4.59294826e-01
-5.61907649e-01 2.68276036e-01 1.15005955e-01 -9.57995430e-02
1.33930326e-01 1.39965611e-02 3.23163956e-01 1.17492056e+00
-1.78668365e-01 6.64003491e-01 1.91365018e-01 1.47899008e+00
-3.39665294e-01 -1.13234200e-01 -5.43631256e-01 -4.51009631e-01
4.24986959e-01 1.22342694e+00 -5.95422745e-01 -9.04275179e-01
-3.82013470e-01 1.28471029e+00 6.84544146e-01 5.66242278e-01
-1.12492132e+00 -3.11691374e-01 6.42118156e-01 -4.93019521e-01
7.87363499e-02 -6.77019954e-02 -4.71037894e-01 -1.60083520e+00
-2.07705237e-02 -9.08187330e-01 5.01921296e-01 -1.44765520e+00
-1.46399498e+00 3.89236808e-01 -5.09668589e-02 -7.46168733e-01
3.02499920e-01 -8.44763041e-01 -5.61262846e-01 7.32571423e-01
-1.72173560e+00 -1.37830257e+00 -9.61135745e-01 1.09001303e+00
6.38537347e-01 1.37687132e-01 4.68708038e-01 -1.01947621e-01
-5.03825426e-01 2.70263255e-01 -4.30855036e-01 3.53239208e-01
3.25866550e-01 -1.18210256e+00 1.63662657e-02 8.75767350e-01
1.09137729e-01 7.86723197e-01 4.64335620e-01 -7.15238869e-01
-1.55671227e+00 -1.04082513e+00 3.83948982e-01 -6.44334257e-01
8.64523113e-01 -1.56876892e-02 -1.03223073e+00 8.72596622e-01
6.82249308e-01 -2.04983771e-01 8.47459435e-01 -9.75128114e-02
-9.40835536e-01 8.46824050e-02 -9.12461042e-01 8.04205656e-01
1.26886594e+00 -9.62488353e-01 -1.06591356e+00 4.00419116e-01
9.39199269e-01 -1.47791624e-01 -7.77131975e-01 1.51035473e-01
3.09452504e-01 -9.41677570e-01 1.09126484e+00 -7.26680279e-01
9.94172275e-01 -4.78014529e-01 -4.76204365e-01 -1.12009811e+00
-3.70347321e-01 1.17864907e-01 -1.45945117e-01 1.00240874e+00
-1.51385961e-03 -4.95384276e-01 1.16716892e-01 5.56275249e-01
-2.11908738e-03 -4.53510791e-01 -5.57450473e-01 -5.99050105e-01
-1.59879014e-01 -6.38535380e-01 4.74811971e-01 1.24762332e+00
1.13803335e-01 6.72575951e-01 -3.66029739e-01 2.53848940e-01
7.90701985e-01 5.17355025e-01 5.27792811e-01 -8.40623438e-01
-2.36662924e-01 -6.21866822e-01 -5.50478578e-01 -8.57932031e-01
4.18390304e-01 -1.17060900e+00 5.58981448e-02 -1.93464243e+00
9.83806908e-01 1.94175795e-01 -4.17235970e-01 8.24431837e-01
-5.22590101e-01 2.71091700e-01 5.44697046e-01 1.34004772e-01
-9.52466488e-01 6.16729438e-01 1.67254949e+00 -4.99913871e-01
3.17982286e-01 -1.00189984e+00 -1.09483445e+00 7.78218806e-01
9.44568932e-01 -2.33942494e-02 -5.15496433e-01 -7.31596231e-01
2.29193807e-01 2.11252738e-02 1.26188922e+00 -8.32983434e-01
2.35232294e-01 -4.88417029e-01 6.24045610e-01 -6.67250574e-01
2.54332453e-01 -6.30120993e-01 -6.59857616e-02 4.22893763e-01
-5.20477593e-01 -2.20619336e-01 2.73284346e-01 8.68519127e-01
-3.68140131e-01 2.11885676e-01 7.16026485e-01 -2.85668850e-01
-1.36418796e+00 2.66593099e-01 2.74591558e-02 1.74898341e-01
9.88491535e-01 -7.71856159e-02 -8.95740747e-01 -4.46794361e-01
-6.54586136e-01 3.80739927e-01 4.98609245e-01 6.82073653e-01
1.07904720e+00 -1.35918176e+00 -6.42301142e-01 -1.74439043e-01
6.21296346e-01 -3.39537978e-01 7.59767354e-01 6.92849576e-01
-3.52108866e-01 4.43179309e-01 -3.66751432e-01 -3.71296436e-01
-1.25548482e+00 9.75505948e-01 2.14135066e-01 1.47199348e-01
-5.19956708e-01 8.01156282e-01 4.58751321e-01 -9.59011018e-02
-9.13497061e-02 -4.80366796e-01 -1.26421347e-01 -2.72762060e-01
8.71697009e-01 1.31012812e-01 -6.38366759e-01 -8.36321354e-01
-4.59950268e-01 4.01371956e-01 4.31774221e-02 1.11115925e-01
8.75585735e-01 -3.75874847e-01 -4.37241197e-01 5.39851725e-01
1.11951125e+00 -5.11987269e-01 -8.98593605e-01 -2.81531841e-01
-2.64724582e-01 -7.24010885e-01 2.21283466e-01 -1.22686195e+00
-1.01232409e+00 1.34199178e+00 3.60749036e-01 -2.50745341e-02
1.50046396e+00 3.93726677e-01 6.65663898e-01 6.27394319e-01
1.24701887e-01 -8.29721630e-01 5.73137760e-01 5.21266162e-01
1.42142427e+00 -1.70613301e+00 -6.52547926e-02 -4.52530175e-01
-9.23487186e-01 9.51429546e-01 7.79742420e-01 -6.03954196e-02
3.99888754e-01 -2.77987748e-01 -2.50273328e-02 -4.69285011e-01
-8.29900622e-01 -6.99161768e-01 5.00490189e-01 5.73483706e-01
1.30902231e-01 2.79414773e-01 1.27910540e-01 6.84234142e-01
8.96562338e-02 3.04050624e-01 4.38152641e-01 6.11088037e-01
-6.36088848e-01 -3.32445174e-01 -3.58705968e-01 3.07005972e-01
-3.80464792e-02 -5.21650553e-01 -7.35913634e-01 6.33279562e-01
1.93761870e-01 1.03783655e+00 1.01170138e-01 -4.76614982e-01
1.67595804e-01 -7.35003948e-02 5.33005238e-01 -3.95965576e-01
-2.08759129e-01 -4.44621980e-01 -8.15186352e-02 -7.31822312e-01
-6.65847123e-01 -2.36823454e-01 -1.56082392e+00 -4.90812153e-01
-4.27424209e-03 -4.26886648e-01 4.25857395e-01 1.03642011e+00
4.53728914e-01 8.20412874e-01 1.49244010e-01 -3.61938298e-01
-1.76224276e-01 -7.55813360e-01 -4.04668689e-01 9.07023072e-01
2.86282927e-01 -6.37182057e-01 -2.81265855e-01 5.09022355e-01] | [10.719162940979004, 1.767513394355774] |
a189a219-3f76-472f-a160-0ebf23eb215e | rgb-event-fusion-for-moving-object-detection | 2209.08323 | null | https://arxiv.org/abs/2209.08323v2 | https://arxiv.org/pdf/2209.08323v2.pdf | RGB-Event Fusion for Moving Object Detection in Autonomous Driving | Moving Object Detection (MOD) is a critical vision task for successfully achieving safe autonomous driving. Despite plausible results of deep learning methods, most existing approaches are only frame-based and may fail to reach reasonable performance when dealing with dynamic traffic participants. Recent advances in sensor technologies, especially the Event camera, can naturally complement the conventional camera approach to better model moving objects. However, event-based works often adopt a pre-defined time window for event representation, and simply integrate it to estimate image intensities from events, neglecting much of the rich temporal information from the available asynchronous events. Therefore, from a new perspective, we propose RENet, a novel RGB-Event fusion Network, that jointly exploits the two complementary modalities to achieve more robust MOD under challenging scenarios for autonomous driving. Specifically, we first design a temporal multi-scale aggregation module to fully leverage event frames from both the RGB exposure time and larger intervals. Then we introduce a bi-directional fusion module to attentively calibrate and fuse multi-modal features. To evaluate the performance of our network, we carefully select and annotate a sub-MOD dataset from the commonly used DSEC dataset. Extensive experiments demonstrate that our proposed method performs significantly better than the state-of-the-art RGB-Event fusion alternatives. The source code and dataset are publicly available at: https://github.com/ZZY-Zhou/RENet. | ['Dominique Ginhac', 'Cédric Demonceaux', 'Fan Yang', 'Rémi Boutteau', 'Zongwei Wu', 'Zhuyun Zhou'] | 2022-09-17 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 1.81551263e-01 -6.35650873e-01 -1.44550204e-01 -4.47617382e-01
-8.76214743e-01 -3.43410313e-01 7.69135356e-01 1.96838174e-02
-5.70802271e-01 5.10157704e-01 -3.75803001e-02 -5.69883548e-02
-6.63560331e-02 -8.58667135e-01 -7.04590619e-01 -8.34618151e-01
1.46637321e-01 -1.32041693e-01 6.81326151e-01 -1.68416649e-01
-1.20270774e-01 5.04439473e-01 -1.96651173e+00 1.28780648e-01
8.05401564e-01 1.34835339e+00 1.09892622e-01 5.65352559e-01
6.40886426e-02 8.79996598e-01 -3.00669760e-01 -1.31007105e-01
3.35814416e-01 -2.84068286e-01 -7.41224289e-02 3.94422328e-03
2.49559268e-01 -4.62232947e-01 -7.78768063e-01 8.67993236e-01
5.27328193e-01 3.76860231e-01 8.14859271e-02 -1.61794019e+00
-3.45292687e-02 2.22859249e-01 -5.82839131e-01 6.17598474e-01
2.13819742e-01 5.69108367e-01 6.02261484e-01 -7.68748403e-01
3.60493839e-01 1.03313136e+00 5.71310282e-01 2.46361136e-01
-8.44184935e-01 -9.40889955e-01 2.67958641e-01 7.67113864e-01
-1.31501639e+00 -6.36468887e-01 9.19144809e-01 -1.78750709e-01
7.91822135e-01 1.41221330e-01 7.56742179e-01 1.23590207e+00
3.32666397e-01 8.10818076e-01 1.12080252e+00 5.24487160e-02
2.39900678e-01 -3.57425630e-01 8.47095996e-02 4.42038208e-01
1.91530347e-01 4.69761133e-01 -8.84880066e-01 2.19008520e-01
3.88907313e-01 5.41705251e-01 -2.51263410e-01 -2.90535122e-01
-1.37836468e+00 5.69799066e-01 5.39432466e-01 7.16604441e-02
-5.66319466e-01 4.45630282e-01 2.95120895e-01 -1.37474552e-01
3.39448601e-01 -3.32541138e-01 -2.99150676e-01 -4.33991730e-01
-8.27601552e-01 3.01046342e-01 3.61576945e-01 7.72729635e-01
9.93066549e-01 6.63832296e-04 -3.23705226e-01 3.43518198e-01
2.88046479e-01 6.97462559e-01 2.69940227e-01 -9.45305288e-01
4.29809868e-01 5.75575411e-01 5.85862361e-02 -8.01037550e-01
-5.08767903e-01 -1.21036731e-01 -6.43264711e-01 2.59877443e-01
3.54670346e-01 -1.48931161e-01 -9.38292682e-01 1.60206330e+00
6.27645254e-01 8.66129220e-01 8.12897533e-02 9.86252546e-01
8.27518821e-01 6.42056227e-01 2.31014028e-01 -2.12578103e-01
1.47397268e+00 -7.80667245e-01 -7.81398237e-01 -3.07858318e-01
2.26127297e-01 -5.23681998e-01 8.51705968e-01 2.52974987e-01
-8.49147737e-01 -7.48087943e-01 -1.27082956e+00 -1.00471182e-02
-5.50569236e-01 -3.33639793e-02 6.37060046e-01 4.74807620e-01
-6.02915168e-01 1.82914913e-01 -1.30815685e+00 -2.65576035e-01
4.73152339e-01 1.66986391e-01 -2.00767845e-01 -3.22409034e-01
-1.27011526e+00 7.65621901e-01 4.97399747e-01 2.89941698e-01
-1.05596888e+00 -6.81617260e-01 -9.46801066e-01 -1.79722622e-01
6.50856555e-01 -5.67674279e-01 1.32070613e+00 -5.83953023e-01
-1.45041478e+00 3.40887129e-01 -4.07740176e-01 -7.16623068e-01
4.78290826e-01 -3.21804583e-01 -6.12308025e-01 3.83547693e-01
-4.95266654e-02 6.93003535e-01 7.64068842e-01 -1.18603575e+00
-1.09191728e+00 -2.37024307e-01 3.49788666e-01 1.07386626e-01
-2.72446483e-01 -1.16740748e-01 -6.64818525e-01 -5.52621603e-01
-1.11368001e-01 -8.66450787e-01 -2.04642788e-01 1.05374809e-02
-7.52956867e-02 -2.05218583e-01 1.12899756e+00 -2.62509018e-01
1.12097192e+00 -2.26968694e+00 -1.68675587e-01 -1.78630173e-01
3.84832025e-01 2.93774933e-01 7.50004798e-02 2.61734724e-01
8.74081701e-02 -4.14751172e-01 -2.32655048e-01 -5.97602546e-01
4.27402891e-02 3.37543279e-01 -2.54127860e-01 6.19821191e-01
3.90334338e-01 8.94119620e-01 -9.97532487e-01 -5.31249940e-01
8.93089712e-01 7.52950907e-01 -2.46365771e-01 1.15917571e-01
-1.56119093e-01 6.15025520e-01 -5.77606499e-01 7.47372329e-01
6.74110532e-01 -8.79328102e-02 -3.02528411e-01 -3.54532242e-01
-4.00226086e-01 1.29891470e-01 -1.30785763e+00 1.74107146e+00
-3.43228281e-01 6.86580181e-01 -1.40572533e-01 -8.83974850e-01
6.78356767e-01 2.47732699e-01 8.57339144e-01 -1.00701797e+00
3.81251901e-01 7.59460079e-03 -2.42260620e-01 -5.88299096e-01
6.77051842e-01 -1.87035762e-02 -1.82435498e-01 1.18832238e-01
-3.83278206e-02 1.50744081e-01 3.50511432e-01 1.41539097e-01
1.28489089e+00 2.05647007e-01 2.11825669e-01 2.64608711e-01
5.78783035e-01 -3.59239690e-02 8.64149809e-01 6.72651291e-01
-6.81140244e-01 5.32592237e-01 9.14902017e-02 -4.22788888e-01
-4.87676769e-01 -1.22301233e+00 -1.35012761e-01 9.13413763e-01
8.01014543e-01 -3.95521283e-01 -4.55668002e-01 -5.44593811e-01
-5.69380820e-03 6.07492626e-01 -4.53271866e-01 -3.43063951e-01
-5.85045397e-01 -9.25379038e-01 5.00612259e-01 6.82733417e-01
9.05316651e-01 -7.84776092e-01 -1.13426566e+00 3.52050036e-01
-3.86012912e-01 -1.52650559e+00 -1.85040727e-01 1.03224523e-01
-4.22137707e-01 -1.07672167e+00 -1.72666296e-01 -5.71946874e-02
9.95730236e-02 6.88056469e-01 8.49680007e-01 -5.56934215e-02
-3.54549140e-01 5.48813581e-01 -5.39517403e-01 -6.15264356e-01
4.14001942e-02 -2.37769708e-01 -1.68308839e-02 3.88728917e-01
6.99562013e-01 -4.88074541e-01 -9.06053722e-01 2.30847955e-01
-1.03273129e+00 1.45210683e-01 5.50875843e-01 3.82688701e-01
6.25685215e-01 2.55226910e-01 4.86346573e-01 -2.55215436e-01
-7.77382180e-02 -5.61816752e-01 -7.05241740e-01 -1.66966040e-02
-2.65861034e-01 -2.73876071e-01 3.09330374e-01 -3.70153695e-01
-1.25070059e+00 2.35337839e-01 8.75326432e-03 -6.23082995e-01
-3.50846857e-01 3.31468225e-01 -3.08539510e-01 -2.36768872e-02
2.99515128e-01 1.37924299e-01 -2.67304122e-01 -7.28337467e-02
4.71399635e-01 4.68710691e-01 8.83469880e-01 -4.47401375e-01
8.15176070e-01 9.99409258e-01 -1.12440344e-02 -5.23931623e-01
-7.07213402e-01 -7.06165314e-01 -4.19050038e-01 -6.96550667e-01
9.47612405e-01 -1.29129076e+00 -8.42555046e-01 6.02162480e-01
-9.64333236e-01 -3.11698079e-01 -2.36439303e-01 7.65909851e-01
-5.14880598e-01 1.28442481e-01 -3.86753440e-01 -8.15734029e-01
1.97093245e-02 -1.21366560e+00 1.41022289e+00 6.15784764e-01
3.16141635e-01 -6.46957934e-01 -1.66047774e-02 3.84765625e-01
4.12275612e-01 6.49250329e-01 6.14782423e-02 -2.63210624e-01
-9.97463942e-01 -2.94135422e-01 -3.78844529e-01 -3.36848292e-03
9.82170999e-02 5.55765256e-02 -1.16463649e+00 -2.76608896e-02
-4.09962498e-02 -4.99759726e-02 1.28053093e+00 3.76745909e-01
1.09313965e+00 3.68843436e-01 -5.09723425e-01 6.77085400e-01
1.41529882e+00 2.53683478e-01 6.62873507e-01 4.09048051e-01
7.95316219e-01 4.06769007e-01 8.41006577e-01 8.21332932e-01
8.65023971e-01 6.53144777e-01 5.76368988e-01 -2.36109011e-02
-3.01450580e-01 2.40151286e-02 5.97354531e-01 4.07217354e-01
-1.17832713e-01 -4.45323378e-01 -8.23598683e-01 5.45214951e-01
-2.17489243e+00 -1.14507174e+00 -1.38087720e-01 2.01783466e+00
5.77733397e-01 3.69869858e-01 1.02804914e-01 2.43363187e-01
6.27410054e-01 4.73882765e-01 -6.93052769e-01 3.39450359e-01
-1.65389344e-01 1.80420637e-01 7.70441294e-01 7.15508685e-02
-1.35546815e+00 7.80968130e-01 5.13555574e+00 8.11475694e-01
-1.31060171e+00 2.82147110e-01 2.57413745e-01 -5.18819630e-01
7.82913622e-03 3.44455056e-02 -9.41205025e-01 6.14019752e-01
1.20056427e+00 -1.73825100e-01 2.22766563e-01 4.94146347e-01
6.19206607e-01 -3.31782818e-01 -9.70210314e-01 1.07838440e+00
-6.81946240e-03 -1.22947013e+00 -4.19554204e-01 -1.06707364e-02
5.21676302e-01 3.77293020e-01 -9.41873714e-02 4.00436401e-01
2.83027202e-01 -6.22940898e-01 1.03301895e+00 7.85661042e-01
3.68957520e-01 -6.78925216e-01 6.23688817e-01 1.27346545e-01
-1.75685644e+00 -1.44598544e-01 -8.05975590e-03 -1.36794597e-01
5.87636769e-01 7.79520392e-01 -2.07102612e-01 8.56946111e-01
9.97726381e-01 1.24616766e+00 -8.10747147e-01 1.07885885e+00
-5.97170107e-02 6.01554036e-01 -5.53263962e-01 3.29723388e-01
2.50550866e-01 1.45986274e-01 5.76925099e-01 1.03838241e+00
4.56133842e-01 2.58516520e-01 3.68221998e-01 6.62736535e-01
1.17377132e-01 -4.82242852e-01 -5.58704555e-01 1.89765096e-01
5.89933455e-01 1.49791062e+00 -7.55199432e-01 -4.75724876e-01
-8.27317894e-01 6.37258768e-01 -1.07270358e-02 5.09068370e-01
-1.46838665e+00 -2.42751777e-01 9.63836968e-01 -1.75717890e-01
5.92441738e-01 -5.54047048e-01 -4.42582518e-02 -1.24184346e+00
2.21080661e-01 -4.65517282e-01 4.47416246e-01 -7.87137866e-01
-1.00364685e+00 4.43043560e-01 2.69542128e-01 -1.50691617e+00
3.89160210e-04 -5.07316232e-01 -6.63722813e-01 3.66975337e-01
-2.03017545e+00 -1.28133523e+00 -9.09250319e-01 9.07268643e-01
6.02195799e-01 1.69680670e-01 1.71785459e-01 6.43463433e-01
-9.62041497e-01 2.96522498e-01 -1.11932017e-01 -2.33218834e-01
7.71741807e-01 -1.00850093e+00 1.13887258e-01 1.29824734e+00
-1.68744642e-02 9.27847922e-02 5.56742609e-01 -4.38719958e-01
-1.68980658e+00 -1.47360027e+00 2.77978837e-01 -5.93726277e-01
7.20194221e-01 -2.37767354e-01 -7.86764622e-01 5.89015543e-01
2.08818778e-01 5.36112070e-01 2.91548610e-01 -4.80285496e-01
-1.07169226e-01 -5.04253149e-01 -7.91192234e-01 4.27403003e-01
1.05213010e+00 -4.20751780e-01 -3.73586208e-01 -8.39270477e-04
8.11258674e-01 -5.32534420e-01 -8.09405208e-01 8.31778407e-01
4.09598261e-01 -1.16961718e+00 1.12896383e+00 5.36696650e-02
2.04394445e-01 -9.06547606e-01 -3.08415055e-01 -8.01300883e-01
5.85692527e-04 -4.49014455e-01 -4.66936737e-01 1.28031647e+00
-1.97062716e-01 -7.52104640e-01 5.40672362e-01 4.40184742e-01
-3.89094859e-01 -5.00189066e-01 -1.15014482e+00 -7.60679364e-01
-5.30709207e-01 -1.03414392e+00 7.67212152e-01 5.21077991e-01
-4.01171267e-01 -2.46359445e-02 -3.30370903e-01 4.80578095e-01
6.43057346e-01 1.39149636e-01 8.70436907e-01 -1.04787695e+00
5.17822541e-02 -4.10936505e-01 -6.21679068e-01 -8.67929518e-01
1.04065217e-01 -5.09887636e-01 4.66400415e-01 -1.41615748e+00
-3.02855186e-02 -3.85656148e-01 -6.59094095e-01 4.35064852e-01
-4.78030801e-01 5.50224304e-01 1.38488933e-01 1.47173956e-01
-1.01708686e+00 7.92318523e-01 8.88775885e-01 -7.58731514e-02
-2.51064822e-02 -2.71168649e-01 -4.24485534e-01 7.05363572e-01
7.76209414e-01 -4.00608450e-01 -3.56107652e-01 -3.46207380e-01
-3.87416929e-02 2.02645063e-02 9.24997091e-01 -1.44449139e+00
5.27217686e-01 -3.69123966e-01 3.57292980e-01 -9.25518572e-01
4.92035776e-01 -9.20703232e-01 2.74909049e-01 2.27084160e-01
1.06593490e-01 7.89617226e-02 4.20769393e-01 9.34790313e-01
-3.47541094e-01 3.41391295e-01 6.36467576e-01 2.37590671e-01
-1.32998824e+00 6.29581928e-01 -4.31712389e-01 -1.74939036e-01
1.41532111e+00 -1.73580796e-01 -5.12519002e-01 -1.11224435e-01
-2.59549737e-01 4.61766809e-01 4.31563646e-01 6.81009769e-01
5.45105219e-01 -1.51760054e+00 -4.91708070e-01 1.90212652e-01
4.66204196e-01 1.33540750e-01 5.61900198e-01 1.23660374e+00
-2.44036987e-01 2.61744440e-01 -1.17808148e-01 -9.47268307e-01
-9.01132226e-01 5.30203640e-01 2.40881816e-01 5.99890109e-03
-8.24142218e-01 5.11035919e-01 2.60071903e-01 -1.21528748e-02
8.45635161e-02 -5.84347129e-01 -1.60010830e-01 4.33090255e-02
7.00162292e-01 3.94577444e-01 2.07119241e-01 -8.47933948e-01
-5.38447499e-01 4.95317072e-01 2.56268024e-01 -1.13275521e-01
1.12887859e+00 -4.99508053e-01 2.52431422e-01 5.09852171e-01
9.65218604e-01 -2.67852813e-01 -1.79942191e+00 -4.61292863e-01
-2.66101301e-01 -4.83455539e-01 3.06247085e-01 -4.02702719e-01
-1.26730251e+00 7.99685717e-01 8.50831985e-01 7.08062127e-02
1.57763934e+00 -6.67996332e-02 1.21717703e+00 1.86094224e-01
3.95223111e-01 -8.44148815e-01 2.54326649e-02 2.62377322e-01
4.30769563e-01 -1.50891399e+00 -3.52581777e-02 -3.08779538e-01
-4.45455045e-01 9.95114625e-01 7.26818621e-01 7.41588324e-02
6.60185814e-01 3.72073174e-01 2.00151708e-02 -2.32746735e-01
-8.61535668e-01 -7.16147721e-01 1.63620412e-01 3.69594395e-01
5.96124725e-03 1.98271815e-02 6.02183118e-02 6.19132638e-01
1.04439057e-01 2.86849380e-01 3.46347094e-01 1.20108414e+00
-3.91041130e-01 -8.19423020e-01 -4.65499341e-01 2.89843202e-01
-2.21014798e-01 1.66493490e-01 2.19465211e-01 7.84151435e-01
6.05366111e-01 1.20969093e+00 1.54947743e-01 -5.22340298e-01
5.01650333e-01 -1.20026015e-01 3.22464883e-01 -1.12442516e-01
-3.12344074e-01 1.00980829e-02 -1.32854268e-01 -1.09415996e+00
-1.01637363e+00 -9.64300156e-01 -1.45747387e+00 -3.92043442e-01
-1.65968165e-01 -4.02536422e-01 6.06654108e-01 1.09308779e+00
5.05173206e-01 8.70582759e-01 5.53321302e-01 -1.31301558e+00
7.97602534e-02 -6.83184624e-01 -2.28321493e-01 3.93426359e-01
5.92946708e-01 -1.19339943e+00 -3.52748483e-01 3.21163446e-01] | [8.43293571472168, -1.0261341333389282] |
c92a2554-bc3c-440f-b858-d8aae28f0117 | explanation-guided-training-for-cross-domain | 2007.08790 | null | https://arxiv.org/abs/2007.08790v2 | https://arxiv.org/pdf/2007.08790v2.pdf | Explanation-Guided Training for Cross-Domain Few-Shot Classification | Cross-domain few-shot classification task (CD-FSC) combines few-shot classification with the requirement to generalize across domains represented by datasets. This setup faces challenges originating from the limited labeled data in each class and, additionally, from the domain shift between training and test sets. In this paper, we introduce a novel training approach for existing FSC models. It leverages on the explanation scores, obtained from existing explanation methods when applied to the predictions of FSC models, computed for intermediate feature maps of the models. Firstly, we tailor the layer-wise relevance propagation (LRP) method to explain the predictions of FSC models. Secondly, we develop a model-agnostic explanation-guided training strategy that dynamically finds and emphasizes the features which are important for the predictions. Our contribution does not target a novel explanation method but lies in a novel application of explanations for the training phase. We show that explanation-guided training effectively improves the model generalization. We observe improved accuracy for three different FSC models: RelationNet, cross attention network, and a graph neural network-based formulation, on five few-shot learning datasets: miniImagenet, CUB, Cars, Places, and Plantae. The source code is available at https://github.com/SunJiamei/few-shot-lrp-guided | ['Ngai-Man Cheung', 'Wojciech Samek', 'Alexander Binder', 'Jiamei Sun', 'Yunqing Zhao', 'Sebastian Lapuschkin'] | 2020-07-17 | null | null | null | null | ['cross-domain-few-shot'] | ['computer-vision'] | [ 5.56026578e-01 4.76364404e-01 -3.22113276e-01 -4.06607866e-01
-4.64445412e-01 -2.48889357e-01 8.93871367e-01 3.92222524e-01
1.74929246e-01 5.82845807e-01 2.19177768e-01 -2.36392230e-01
-6.12595320e-01 -7.40733385e-01 -6.12567604e-01 -2.85340518e-01
2.72087473e-03 6.88966870e-01 5.29462099e-01 -5.58553517e-01
3.31005603e-01 3.00325722e-01 -1.74974298e+00 6.07887089e-01
9.45599675e-01 1.06527662e+00 2.30936348e-01 7.18364477e-01
-3.08005482e-01 8.36620748e-01 -2.85319537e-01 -5.29612899e-01
-1.66379437e-01 -7.59978414e-01 -9.41090763e-01 -2.85491422e-02
1.17264308e-01 1.78543285e-01 -9.46815982e-02 7.58684576e-01
-3.20106521e-02 4.90854383e-01 9.60440814e-01 -1.60957098e+00
-1.12312710e+00 7.09448338e-01 -1.48882598e-01 9.61062312e-02
2.36656219e-01 1.94378078e-01 1.12482488e+00 -1.01163983e+00
8.42389941e-01 9.91968930e-01 4.59743142e-01 1.05538583e+00
-1.06924379e+00 -4.16544646e-01 3.08299571e-01 8.59802723e-01
-1.05571878e+00 -1.69698268e-01 7.88461983e-01 -4.09972697e-01
1.25785279e+00 1.09122925e-01 4.90237623e-01 1.18319178e+00
-3.57922353e-02 4.75181490e-01 7.29180515e-01 -5.46202719e-01
6.52954400e-01 1.49724603e-01 5.98589361e-01 6.34647131e-01
1.01835079e-01 3.20217639e-01 -6.68655336e-01 2.22689863e-02
2.17408046e-01 2.59919554e-01 -2.93663263e-01 -5.07619321e-01
-6.78150773e-01 1.08228421e+00 5.71589649e-01 3.17801267e-01
-3.53938729e-01 -1.13859307e-02 2.98963308e-01 1.43476382e-01
6.30529404e-01 7.21476793e-01 -7.92017758e-01 3.75832990e-02
-5.20527720e-01 1.22739159e-01 7.35214353e-01 1.08825815e+00
8.32323015e-01 -2.35763155e-02 -3.33710611e-01 7.60556102e-01
1.64260954e-01 1.09923080e-01 7.29505599e-01 -7.52171338e-01
2.38960072e-01 8.73838305e-01 -3.81191611e-01 -8.53669763e-01
-2.72214592e-01 -3.38918149e-01 -5.97781301e-01 -6.74102176e-03
-1.32656664e-01 3.83120179e-02 -1.10768032e+00 1.61716855e+00
3.43408436e-01 5.54981470e-01 2.63710350e-01 8.40709627e-01
1.12024653e+00 4.25620049e-01 4.12190169e-01 7.76381418e-02
1.12308550e+00 -1.24817765e+00 -3.48354340e-01 -4.03436780e-01
5.80913603e-01 -2.49366447e-01 1.07403374e+00 -1.28384409e-02
-6.75406933e-01 -6.40335381e-01 -1.16350496e+00 -8.33004620e-03
-1.00429034e+00 -3.28866988e-01 5.94387710e-01 1.20422080e-01
-7.88365304e-01 1.01269197e+00 -4.46768463e-01 -7.42437661e-01
5.82590818e-01 1.88823998e-01 -2.80124426e-01 -2.49796182e-01
-1.37449384e+00 1.11107981e+00 6.45102918e-01 -5.51717818e-01
-7.43764102e-01 -7.48612523e-01 -1.05820453e+00 6.64837778e-01
5.05969703e-01 -4.58973438e-01 1.18549216e+00 -9.34395790e-01
-1.33338785e+00 5.68913996e-01 -4.32081148e-02 -7.46483088e-01
-4.82904762e-02 3.39028910e-02 -5.41327834e-01 6.30285665e-02
1.55150056e-01 8.42783332e-01 8.35281193e-01 -1.17727017e+00
-5.39396703e-01 -2.64805734e-01 4.93281633e-02 -1.01096243e-01
-1.95600286e-01 -3.40741992e-01 -3.92523408e-01 -4.01335210e-01
-1.48908928e-01 -7.68527806e-01 -3.04732829e-01 -2.25550130e-01
-2.89789885e-01 -3.45336974e-01 7.04593062e-01 -2.29337379e-01
9.58730042e-01 -2.09148455e+00 2.45947868e-01 1.74229946e-02
2.02200383e-01 5.41645467e-01 -4.58829194e-01 4.40740228e-01
-3.30965281e-01 7.71572441e-02 -5.33578932e-01 -3.79922152e-01
5.84899746e-02 4.13108587e-01 -4.09077585e-01 -5.81777468e-02
6.91267788e-01 1.03409433e+00 -9.90255058e-01 -3.74921799e-01
3.58578622e-01 3.42569470e-01 -5.74596167e-01 7.59845003e-02
-5.06950557e-01 1.78059250e-01 -2.59796053e-01 4.85078514e-01
5.26915967e-01 -6.05721951e-01 1.54245704e-01 1.00882493e-01
1.77428305e-01 -3.37825455e-02 -6.62243009e-01 1.65686488e+00
-3.84212613e-01 4.38192099e-01 -8.32165003e-01 -1.28263485e+00
9.41362202e-01 3.18486154e-01 1.40678525e-01 -7.17933893e-01
1.11037560e-01 2.50122130e-01 -1.34449169e-01 -2.84292310e-01
4.07587409e-01 -1.45004213e-01 -6.38875738e-02 3.64412427e-01
7.66385138e-01 8.18385929e-02 2.26903215e-01 4.21681255e-01
1.09534502e+00 2.26250187e-01 8.79508793e-01 2.68568285e-02
4.31889445e-01 3.69194686e-01 4.67030287e-01 7.52746224e-01
-3.41087162e-01 7.50318527e-01 4.40698743e-01 -5.60562551e-01
-8.10031474e-01 -8.09632003e-01 1.01619519e-01 1.22164786e+00
2.19412044e-01 -5.60836315e-01 -3.68133605e-01 -1.26833808e+00
1.14774458e-01 1.31471992e+00 -9.57498074e-01 -7.72704720e-01
-3.21267731e-02 -9.70603824e-02 1.81317516e-02 6.81233585e-01
1.93306208e-01 -1.21500087e+00 -5.89164555e-01 1.06731683e-01
1.46995887e-01 -1.06965661e+00 -2.10803702e-01 6.13697290e-01
-7.57548034e-01 -1.28066266e+00 -3.89895707e-01 -5.18619061e-01
5.32289028e-01 3.96722794e-01 1.13230753e+00 2.74739534e-01
-2.86182940e-01 5.32971740e-01 -7.75900841e-01 -4.93054777e-01
-2.53489494e-01 1.93608329e-01 -1.54689863e-01 -1.57813013e-01
7.78862834e-01 -6.46907151e-01 -3.29630405e-01 1.36877134e-01
-6.44388080e-01 1.98597446e-01 4.81955945e-01 9.27900016e-01
5.48779607e-01 -2.82998115e-01 7.51640916e-01 -1.25975370e+00
4.47592199e-01 -9.08119202e-01 -2.89845049e-01 4.83021319e-01
-9.95785356e-01 2.07292482e-01 8.44512522e-01 -4.30641770e-01
-1.14749050e+00 2.33572960e-01 2.79388390e-02 -6.87659800e-01
-4.80372608e-01 4.19163048e-01 1.23092934e-01 1.13303714e-01
1.06796730e+00 1.97535679e-01 -1.87348470e-01 -3.27491760e-01
6.49514437e-01 3.84465247e-01 4.22407538e-01 -2.42439240e-01
7.77065635e-01 2.24304318e-01 -7.81896934e-02 -5.54453492e-01
-1.10574818e+00 -5.86387396e-01 -6.79389834e-01 -2.37774521e-01
8.62353623e-01 -5.16143262e-01 -3.08289707e-01 -1.47167489e-01
-1.37437546e+00 -3.70700568e-01 -6.57290637e-01 4.29599643e-01
-6.46525204e-01 -6.37701526e-02 -2.84392744e-01 -4.48651016e-01
-4.33307022e-01 -7.04177022e-01 9.27233160e-01 4.39218640e-01
-4.06396985e-01 -1.22486806e+00 1.16476096e-01 1.15845039e-01
5.96520364e-01 1.31933481e-01 1.18766558e+00 -1.39053845e+00
-4.75961268e-01 -2.25769043e-01 -3.35164458e-01 6.35048524e-02
-2.33224686e-03 -2.20737651e-01 -1.14183831e+00 1.56835750e-01
-3.72695714e-01 -4.79416519e-01 1.06463635e+00 2.62445271e-01
1.28580725e+00 -9.90038887e-02 -3.57463509e-01 5.33822238e-01
1.51477027e+00 9.07401741e-02 4.69520420e-01 1.38689876e-01
4.69740361e-01 5.82847655e-01 7.99017966e-01 5.45080304e-01
2.95153081e-01 7.53585458e-01 5.04762769e-01 1.22266062e-01
-3.91424298e-01 -3.23392332e-01 2.17407220e-03 4.27312762e-01
-2.21480876e-01 -2.25339204e-01 -8.89379501e-01 5.98952770e-01
-2.23591661e+00 -1.07015932e+00 -8.87161195e-02 1.91520357e+00
3.59696537e-01 8.83595273e-02 -2.07759812e-01 -1.66403338e-01
6.89968765e-01 3.97056043e-02 -8.11829388e-01 -5.51417410e-01
1.88908890e-01 5.65511763e-01 1.78840607e-01 4.91684914e-01
-9.99005318e-01 1.20112467e+00 5.09122181e+00 7.69990683e-01
-8.88833165e-01 2.97404587e-01 2.79141605e-01 -4.68037538e-02
-2.94933170e-01 1.29670456e-01 -7.64896035e-01 1.71656951e-01
1.07869017e+00 -4.48738128e-01 2.88690954e-01 1.35261261e+00
-2.48186827e-01 3.03885221e-01 -1.16760838e+00 5.34505606e-01
3.17015976e-01 -1.64981341e+00 1.81232944e-01 -1.79167494e-01
7.42335618e-01 -6.21278919e-02 6.31402284e-02 9.21056867e-01
3.74257535e-01 -6.98294520e-01 4.44284111e-01 3.61679137e-01
6.33320749e-01 -5.36912799e-01 6.63551569e-01 1.74255207e-01
-1.16500163e+00 -2.34759152e-01 -6.47485197e-01 -1.51502728e-01
-7.34734088e-02 2.17279658e-01 -1.02720237e+00 6.63275421e-01
5.01633346e-01 9.89842594e-01 -8.75024796e-01 7.41628885e-01
-6.04142547e-01 4.97486651e-01 3.11854810e-01 -1.47209287e-01
2.34721705e-01 3.22193652e-01 4.08397287e-01 1.03617060e+00
3.62843424e-01 2.61019349e-01 -3.83016653e-02 1.01109529e+00
9.04240906e-02 -1.62534546e-02 -8.50251079e-01 4.74454388e-02
4.96482253e-01 1.28440964e+00 -7.54401922e-01 -5.20634830e-01
-4.23299879e-01 1.10856211e+00 6.76514685e-01 2.84554690e-01
-1.05279672e+00 -4.57674712e-01 4.93376344e-01 -1.02914929e-01
5.75060904e-01 2.62579650e-01 -3.55640680e-01 -1.15339899e+00
-4.17604208e-01 -6.06912971e-01 5.51136494e-01 -7.82528579e-01
-1.40525150e+00 7.99732268e-01 8.12007338e-02 -1.17026222e+00
-5.44153154e-01 -6.04089200e-01 -1.15286613e+00 7.12574184e-01
-1.70701313e+00 -1.29867029e+00 -5.27926207e-01 6.15502238e-01
9.36582327e-01 -3.63472044e-01 1.15508163e+00 -2.49683540e-02
-4.66822356e-01 3.58060956e-01 -1.27596736e-01 -3.90914172e-01
5.16535401e-01 -1.29612482e+00 7.28649259e-01 7.28552222e-01
1.90686077e-01 5.53126395e-01 7.38764167e-01 -6.03226662e-01
-7.88289726e-01 -1.42652822e+00 1.11535382e+00 -2.99747795e-01
5.79802334e-01 -9.87318233e-02 -1.16571248e+00 6.81089699e-01
1.20498769e-01 1.74102247e-01 9.26370502e-01 4.70541090e-01
-4.02752370e-01 1.89295843e-01 -9.21589851e-01 4.28407282e-01
1.07913709e+00 -4.38400239e-01 -7.23624945e-01 3.57980967e-01
1.11611366e+00 5.17579690e-02 -4.98656660e-01 2.50964195e-01
4.54761952e-01 -8.29743981e-01 8.91437292e-01 -1.22831035e+00
9.52653944e-01 7.36229075e-03 -3.63901049e-01 -1.57045233e+00
-6.95793629e-01 4.89218496e-02 -4.40054983e-01 1.14971113e+00
8.62967134e-01 -6.02871418e-01 7.42037654e-01 5.93610585e-01
-4.71606135e-01 -8.05318475e-01 -6.23364151e-01 -9.49297309e-01
-1.00445770e-01 -4.83021945e-01 7.85759747e-01 9.76840615e-01
3.01672906e-01 6.70799136e-01 -2.35157758e-01 1.60934284e-01
6.46731436e-01 2.58632660e-01 5.65617919e-01 -1.59307265e+00
-5.60867667e-01 -2.07495585e-01 -3.65679592e-01 -4.28018302e-01
3.23303849e-01 -1.12429154e+00 1.49129704e-01 -1.68577266e+00
2.75497705e-01 -1.05313808e-02 -3.67792159e-01 7.99718380e-01
-2.87919730e-01 -2.36168429e-02 4.59753424e-01 1.19718164e-01
-8.82913709e-01 7.18790591e-01 9.02445614e-01 -2.05394149e-01
-1.57019794e-01 -2.92343013e-02 -7.47928262e-01 7.04991937e-01
1.06869709e+00 -6.04366124e-01 -6.49739504e-01 -1.74580783e-01
-1.62354082e-01 -7.66038746e-02 6.25920892e-01 -1.16063344e+00
4.01467025e-01 -2.37933770e-01 3.49578947e-01 -2.90663987e-01
3.69490474e-01 -7.35918283e-01 -6.90793619e-02 5.95286667e-01
-6.33111238e-01 -3.49940777e-01 9.80051607e-02 8.11232805e-01
-4.21023555e-02 -4.02117550e-01 8.74685645e-01 -8.49328339e-02
-1.32854939e+00 3.70632678e-01 -2.38341410e-02 5.81819950e-05
1.11774707e+00 -1.20934203e-01 -7.07535207e-01 -2.56108463e-01
-9.39665616e-01 2.51890421e-01 2.61048675e-01 6.00803614e-01
9.52556908e-01 -1.40005100e+00 -4.33955371e-01 1.89205766e-01
8.03076982e-01 -4.39538479e-01 4.89842176e-01 6.84569061e-01
7.14808330e-02 6.66093528e-01 -3.13233882e-01 -3.35055500e-01
-1.24261630e+00 8.82403791e-01 8.31181481e-02 -3.55321467e-01
-5.51985800e-01 8.41415584e-01 1.35821626e-01 -5.78568220e-01
-3.34888473e-02 -3.99154089e-02 -4.53045934e-01 -1.76448137e-01
4.33191717e-01 3.48930955e-01 -4.87598106e-02 -2.85883754e-01
-3.37035418e-01 3.17343354e-01 -2.45405197e-01 1.33584023e-01
1.51163888e+00 1.48951828e-01 5.26778460e-01 3.69312257e-01
1.05335820e+00 -8.03745508e-01 -1.21730328e+00 -3.66447628e-01
2.29537994e-01 -1.93971545e-01 -2.28943154e-01 -1.10049200e+00
-7.44651020e-01 1.01666188e+00 5.30625165e-01 2.47809827e-01
1.00824571e+00 4.02053148e-01 4.72829252e-01 2.21084222e-01
2.23663241e-01 -1.03700840e+00 2.10325673e-01 7.15764225e-01
8.71338367e-01 -1.45369625e+00 -2.28653595e-01 -3.25229198e-01
-9.68452215e-01 1.06447172e+00 8.44753206e-01 3.83902341e-02
5.97298205e-01 -3.68683636e-01 -4.16139811e-01 -6.46126449e-01
-1.25931907e+00 -6.98347628e-01 5.35160005e-01 9.35929298e-01
2.82550275e-01 9.70354602e-02 -1.98092759e-01 1.10454428e+00
1.59370124e-01 7.42939264e-02 2.95803130e-01 6.35201931e-01
-7.05636621e-01 -9.06839550e-01 1.73530996e-01 4.43184465e-01
1.96933091e-01 -2.47075483e-01 -7.43605494e-01 8.49254668e-01
3.56277853e-01 9.19608593e-01 -1.09874874e-01 -5.97254157e-01
3.88061583e-01 3.36555451e-01 1.30624011e-01 -1.04888511e+00
-2.49640077e-01 -6.71029389e-01 -6.07527830e-02 -6.27632678e-01
-1.95519194e-01 -4.36561733e-01 -1.37600672e+00 -1.59191057e-01
-3.52900773e-01 2.44801119e-01 5.26507676e-01 1.13825691e+00
7.56520152e-01 8.03740799e-01 4.25107032e-01 -8.37594628e-01
-4.40672755e-01 -9.34885144e-01 -5.17952144e-01 5.45959294e-01
1.91785656e-02 -9.02145684e-01 -4.90786046e-01 -7.66906654e-03] | [9.992653846740723, 2.8441431522369385] |
274a405e-2cc1-49c4-a837-9cf6514f141c | bootstrap-your-own-latent-a-new-approach-to | 2006.07733 | null | https://arxiv.org/abs/2006.07733v3 | https://arxiv.org/pdf/2006.07733v3.pdf | Bootstrap your own latent: A new approach to self-supervised Learning | We introduce Bootstrap Your Own Latent (BYOL), a new approach to self-supervised image representation learning. BYOL relies on two neural networks, referred to as online and target networks, that interact and learn from each other. From an augmented view of an image, we train the online network to predict the target network representation of the same image under a different augmented view. At the same time, we update the target network with a slow-moving average of the online network. While state-of-the art methods rely on negative pairs, BYOL achieves a new state of the art without them. BYOL reaches $74.3\%$ top-1 classification accuracy on ImageNet using a linear evaluation with a ResNet-50 architecture and $79.6\%$ with a larger ResNet. We show that BYOL performs on par or better than the current state of the art on both transfer and semi-supervised benchmarks. Our implementation and pretrained models are given on GitHub. | ['Rémi Munos', 'Bilal Piot', 'Carl Doersch', 'Florent Altché', 'Jean-bastien Grill', 'Koray Kavukcuoglu', 'Bernardo Avila Pires', 'Zhaohan Daniel Guo', 'Michal Valko', 'Florian Strub', 'Pierre H. Richemond', 'Mohammad Gheshlaghi Azar', 'Elena Buchatskaya', 'Corentin Tallec'] | 2020-06-13 | null | null | null | null | ['self-supervised-image-classification', 'self-supervised-person-re-identification'] | ['computer-vision', 'computer-vision'] | [ 1.53585106e-01 2.80028939e-01 -5.17390311e-01 -5.92184365e-01
-7.07937479e-01 -4.21837330e-01 5.33123136e-01 -1.74540117e-01
-6.21586800e-01 4.75842357e-01 4.67099575e-03 5.48598655e-02
3.20964128e-01 -5.88010490e-01 -1.08537436e+00 -3.70570570e-01
-1.34413391e-01 6.78001583e-01 1.88202903e-01 -4.37015817e-02
-1.30380765e-01 1.68954581e-01 -1.40406179e+00 6.93982482e-01
2.63612449e-01 1.25908411e+00 1.58889398e-01 5.86526632e-01
1.17533254e-02 1.22048020e+00 -1.34631783e-01 -4.69788015e-01
6.50859952e-01 -3.18409890e-01 -1.05224013e+00 7.94517174e-02
8.59228730e-01 -3.61582786e-01 -5.39000213e-01 1.05429339e+00
4.28367496e-01 -1.83123015e-02 4.11530793e-01 -1.28695011e+00
-8.71868312e-01 7.97611535e-01 -6.03815317e-01 1.86947986e-01
-1.51914470e-02 1.58317566e-01 1.06197846e+00 -1.09680736e+00
8.89286995e-01 1.09269607e+00 5.03499925e-01 7.20217466e-01
-1.56785774e+00 -8.84411931e-01 5.56345880e-01 1.74899057e-01
-1.22917295e+00 -5.47626495e-01 6.86372459e-01 -3.04046929e-01
1.14628887e+00 -3.44504595e-01 5.89361250e-01 1.00119340e+00
-1.42786190e-01 8.28970551e-01 1.33095348e+00 -3.26613903e-01
-7.76458010e-02 4.37211126e-01 1.93090767e-01 9.83138680e-01
-1.62620604e-01 5.80225140e-02 -7.63736188e-01 1.44313812e-01
6.96317136e-01 1.19731531e-01 -1.19715862e-01 -6.97030842e-01
-1.13179290e+00 7.40216494e-01 1.06047928e+00 1.05243430e-01
-2.65267193e-01 4.71594512e-01 4.74293709e-01 7.22862780e-01
7.37170041e-01 2.65935302e-01 -5.21411061e-01 1.51169866e-01
-8.44047964e-01 -2.80806988e-01 7.82800674e-01 8.97763908e-01
9.47227895e-01 1.92521945e-01 3.46117795e-01 8.92455757e-01
1.66019604e-01 4.73033696e-01 7.23463714e-01 -9.68334734e-01
4.13495690e-01 6.40910506e-01 -1.96596846e-01 -7.83793807e-01
-1.43549055e-01 -7.58219957e-01 -9.21863079e-01 5.02161503e-01
2.61227816e-01 4.33540083e-02 -1.13705623e+00 1.89399469e+00
-1.69125602e-01 3.19092810e-01 2.58387830e-02 3.99767429e-01
9.17716265e-01 6.70620501e-01 -8.33080336e-03 1.07471399e-01
9.65270877e-01 -1.69639623e+00 -1.08084708e-01 -6.73493326e-01
6.92446291e-01 -3.80611509e-01 9.46731687e-01 3.37592810e-01
-1.25499809e+00 -7.20708013e-01 -9.96595442e-01 5.67327812e-02
-5.44177592e-01 1.73462883e-01 3.48023981e-01 3.16990137e-01
-1.58333528e+00 8.17829907e-01 -1.00004172e+00 -3.88553053e-01
6.42842829e-01 4.71328765e-01 -7.64272571e-01 -3.15321565e-01
-6.68060005e-01 9.62067068e-01 2.75556505e-01 -3.08509052e-01
-1.23774517e+00 -8.79960835e-01 -9.35875595e-01 -1.52691290e-01
1.17940672e-01 -5.67666471e-01 1.15260148e+00 -1.65536451e+00
-1.41242814e+00 1.48109519e+00 -1.41637832e-01 -9.10435319e-01
6.18485034e-01 -3.26328665e-01 -2.80756533e-01 2.36612707e-01
2.55456537e-01 1.29163074e+00 1.00142193e+00 -1.48175490e+00
-2.80563563e-01 -2.94139832e-01 1.05881192e-01 3.26544791e-01
-5.72037339e-01 -7.62663409e-02 -5.20619333e-01 -3.40208262e-01
7.74685293e-02 -1.14489615e+00 -1.79491475e-01 4.33546871e-01
-1.34972960e-01 -1.68224722e-01 6.77311897e-01 -5.49175382e-01
6.62607908e-01 -2.21366692e+00 7.14150146e-02 1.33255243e-01
4.22530562e-01 2.25973561e-01 -6.10091805e-01 1.98887646e-01
-7.33139157e-01 6.92043602e-02 -9.86592099e-02 -7.21516967e-01
-2.59962440e-01 -1.52653242e-02 -4.74789351e-01 5.13417959e-01
-1.28588825e-02 1.05051446e+00 -1.06266212e+00 -2.13254869e-01
8.14677998e-02 3.59424859e-01 -5.01444876e-01 2.88236409e-01
6.85760379e-02 9.27966982e-02 1.58533603e-01 2.35946834e-01
5.55350423e-01 -8.90061080e-01 3.32281023e-01 -3.04990023e-01
2.22799242e-01 1.78847834e-01 -8.80327702e-01 1.97904444e+00
-5.89232385e-01 6.63559496e-01 -1.51897306e-02 -1.31919599e+00
8.43047738e-01 2.22405374e-01 3.07029426e-01 -6.98603868e-01
-8.84701610e-02 2.19151974e-01 -7.88784847e-02 -2.63282936e-02
1.23076692e-01 1.49350733e-01 4.08927470e-01 8.87471557e-01
6.39930367e-01 2.11079568e-01 8.11301097e-02 4.70185667e-01
1.15163708e+00 3.50134015e-01 2.50628799e-01 -2.12367058e-01
2.41682366e-01 -2.62709498e-01 2.45512873e-01 8.94409537e-01
-2.22720072e-01 6.18909895e-01 3.89615834e-01 -7.52129376e-01
-1.00560224e+00 -1.19547415e+00 2.12724075e-01 1.38933742e+00
-7.24270642e-02 -3.54645520e-01 -5.55012405e-01 -1.13468564e+00
-3.78962494e-02 3.48272741e-01 -1.01570344e+00 -8.89463648e-02
-4.08026099e-01 -3.57167661e-01 3.18688065e-01 7.06252515e-01
9.07503426e-01 -1.13117540e+00 -1.72167748e-01 3.93464491e-02
-1.02210511e-02 -1.05332780e+00 -4.14796919e-01 2.41091877e-01
-9.97171521e-01 -1.01983202e+00 -8.84608746e-01 -1.03059089e+00
1.08452463e+00 4.72858250e-01 1.56897712e+00 1.48850039e-01
-3.27065811e-02 4.16572928e-01 -7.04561546e-02 -1.08340085e-01
-4.80603695e-01 4.13445532e-01 5.77951483e-02 1.11311480e-01
-3.49232182e-02 -8.70189667e-01 -8.84978354e-01 3.26651663e-01
-7.00508535e-01 3.11419606e-01 5.82583487e-01 1.05605078e+00
5.55293858e-01 -3.17354530e-01 5.91627002e-01 -1.24403918e+00
-3.06054782e-02 -6.24718189e-01 -5.54820299e-01 3.12725157e-01
-9.00987625e-01 8.30950662e-02 7.09762990e-01 -4.08876508e-01
-8.21650386e-01 3.37530851e-01 1.38917312e-01 -9.01456416e-01
-1.39111392e-02 3.41417253e-01 3.62290680e-01 -1.93568513e-01
9.73164022e-01 3.40426713e-01 2.41457641e-01 -3.55315983e-01
6.09266222e-01 3.21051687e-01 5.05232871e-01 -1.77106500e-01
9.04240131e-01 8.31228554e-01 -3.38701844e-01 -2.30472818e-01
-1.18957984e+00 -4.27560896e-01 -7.13045597e-01 -1.80439606e-01
5.67035973e-01 -1.37778080e+00 -4.71306562e-01 6.38156652e-01
-9.47421789e-01 -8.29759836e-01 -5.52161753e-01 1.98121637e-01
-5.07352829e-01 -3.63940634e-02 -5.96506238e-01 -3.00055534e-01
-4.77160543e-01 -9.70647156e-01 7.62072086e-01 1.39327897e-02
-3.67330224e-03 -1.13680840e+00 7.28499368e-02 4.65716809e-01
5.93366802e-01 -9.86384153e-02 5.18520176e-01 -7.21648574e-01
-5.91385484e-01 -2.83699363e-01 -5.91710865e-01 8.65214705e-01
6.48632497e-02 -4.33129579e-01 -1.20922685e+00 -5.94441712e-01
-1.41852334e-01 -9.60610867e-01 1.25096226e+00 1.33629814e-01
1.15114629e+00 -3.06645155e-01 -3.45817894e-01 7.71175325e-01
1.55474854e+00 -1.41739056e-01 6.37952864e-01 2.75996238e-01
6.29935384e-01 4.74586457e-01 1.05887353e-01 7.54458979e-02
4.79999006e-01 3.40328157e-01 6.07647538e-01 -3.02600741e-01
-3.65851551e-01 -4.99792933e-01 5.49758077e-01 8.34093511e-01
-3.83888669e-02 5.21520190e-02 -9.00913954e-01 6.56621695e-01
-1.87993920e+00 -8.95891428e-01 6.12647891e-01 2.15913749e+00
9.86423790e-01 3.68817955e-01 4.06918414e-02 -4.10078406e-01
6.30378902e-01 4.12861377e-01 -9.23059404e-01 4.91400510e-02
-1.59014836e-01 4.61439043e-01 7.47998297e-01 4.47816044e-01
-1.17563224e+00 1.10585535e+00 6.63426304e+00 6.80335164e-01
-1.40007150e+00 3.14994156e-01 1.13859224e+00 -2.05652192e-01
-2.49322504e-03 2.24122126e-02 -6.61646008e-01 1.54445544e-01
1.02741265e+00 -1.12918563e-01 6.29339159e-01 1.20494843e+00
-4.10723865e-01 2.72066630e-02 -1.35971200e+00 1.15374589e+00
4.97350693e-01 -1.62298703e+00 1.25658944e-01 -1.50973499e-01
1.06077015e+00 9.30975318e-01 2.48394623e-01 4.42069739e-01
6.28576279e-01 -1.11377168e+00 5.76385736e-01 5.02440453e-01
1.03900063e+00 -4.11079824e-01 5.75102925e-01 3.10928166e-01
-1.08102226e+00 9.78457779e-02 -1.46231771e-01 1.97804645e-02
-1.90848187e-01 2.00303689e-01 -7.13609099e-01 1.81711987e-01
8.88003290e-01 1.32667708e+00 -9.16587591e-01 7.52577245e-01
-3.52911055e-01 6.31657302e-01 -3.39858949e-01 4.00086850e-01
8.32442045e-02 1.82462558e-01 6.06123395e-02 1.07555950e+00
-1.83783889e-01 -2.54130065e-01 2.86416352e-01 8.61477733e-01
-8.18014324e-01 -1.15906270e-02 -9.31399465e-01 2.04721522e-02
2.32422039e-01 1.33448756e+00 -7.38328338e-01 -7.55488932e-01
-3.57630253e-01 1.21843851e+00 8.70011389e-01 3.31419408e-01
-6.11440718e-01 -7.97708780e-02 2.32106209e-01 8.66087005e-02
1.62068367e-01 -7.24709406e-02 -9.24541131e-02 -1.42102313e+00
-9.57952365e-02 -8.61903548e-01 3.44100535e-01 -9.51449215e-01
-1.47306597e+00 9.32653606e-01 -9.51640606e-02 -1.34147859e+00
-2.29425475e-01 -7.23467886e-01 -5.51252365e-01 6.27705514e-01
-1.69510317e+00 -1.18914020e+00 -4.46655869e-01 6.73335552e-01
5.35593748e-01 -3.60613465e-01 1.16784871e+00 1.53078482e-01
-4.02494341e-01 8.00168514e-01 9.41332206e-02 4.55622673e-01
8.98964167e-01 -1.20337081e+00 7.25175858e-01 5.05230248e-01
4.38029975e-01 4.30491745e-01 8.95076469e-02 -3.26471210e-01
-9.38758671e-01 -1.22403157e+00 6.40574992e-01 -4.43660647e-01
8.05736125e-01 -4.65138435e-01 -7.09184825e-01 1.21893263e+00
6.22487843e-01 7.49049067e-01 5.38174570e-01 6.31116033e-02
-1.03508198e+00 -3.45386863e-01 -1.06553209e+00 5.13257384e-01
1.23223388e+00 -8.39102387e-01 -3.20219278e-01 6.90653205e-01
7.39190698e-01 -3.67200613e-01 -6.97765529e-01 1.21920012e-01
5.79481840e-01 -1.04395080e+00 1.07723498e+00 -7.61615455e-01
7.05831468e-01 3.28167491e-02 -4.96806167e-02 -1.45537257e+00
-2.85882264e-01 -4.80396479e-01 -1.05543256e-01 7.89049447e-01
5.96703529e-01 -9.87308383e-01 1.11226499e+00 3.63011658e-01
1.39954254e-01 -9.71530020e-01 -6.90004051e-01 -8.62307131e-01
8.90766457e-02 -2.74054408e-01 8.43786374e-02 1.03261590e+00
-2.11542711e-01 5.78737438e-01 -3.33965391e-01 -2.09957197e-01
8.43323350e-01 9.73621663e-03 7.81067610e-01 -9.37034726e-01
-3.63278896e-01 -1.95025891e-01 -4.41805422e-01 -1.19406378e+00
4.51111287e-01 -1.32277048e+00 -2.62721956e-01 -1.30310774e+00
4.58931774e-01 -4.12195414e-01 -6.41993523e-01 8.77524257e-01
1.29725188e-01 8.16086650e-01 4.46675658e-01 3.92518401e-01
-8.26823711e-01 4.77700770e-01 8.16709936e-01 -3.17045182e-01
1.79314390e-02 -3.55931133e-01 -7.34901369e-01 1.02794576e+00
8.18883538e-01 -5.66010654e-01 -5.07888913e-01 -7.14374363e-01
2.06740752e-01 -2.26693034e-01 5.00075758e-01 -1.12114000e+00
2.89920539e-01 2.95455605e-01 4.97903526e-01 -4.11964595e-01
5.74633956e-01 -6.82576895e-01 -1.07170179e-01 4.99651134e-01
-7.84241140e-01 2.39083305e-01 1.71908081e-01 6.48585677e-01
-2.11662009e-01 -2.86114570e-02 1.06542528e+00 -3.98968667e-01
-6.30342066e-01 5.32651067e-01 2.35050976e-01 1.98693365e-01
7.66831934e-01 -4.09079045e-02 -6.24382854e-01 -6.68602824e-01
-9.61117506e-01 2.39728495e-01 4.10831213e-01 4.79751647e-01
6.98345363e-01 -1.29632199e+00 -5.80342114e-01 2.88722187e-01
1.61817417e-01 -1.37368917e-01 2.91028559e-01 6.60962343e-01
-3.60004336e-01 1.57355964e-02 -3.07713777e-01 -6.15185857e-01
-1.13269210e+00 3.30234140e-01 6.54186368e-01 -6.56080246e-01
-4.85841930e-01 1.12083006e+00 3.93782854e-01 -7.10700631e-01
2.73294300e-01 1.32677227e-01 -9.34073105e-02 -5.04830517e-02
5.64001858e-01 1.37929037e-01 -1.54075853e-03 -5.57980001e-01
-2.51870185e-01 3.64757866e-01 -4.71954048e-01 -3.13559830e-01
1.51060915e+00 6.68443218e-02 -3.23669225e-01 7.24480033e-01
1.59938788e+00 -5.11010110e-01 -1.39979362e+00 -7.78023779e-01
-3.09364378e-01 -2.40490645e-01 -8.34754035e-02 -9.82527673e-01
-1.46757483e+00 8.56404305e-01 9.35071111e-01 9.02495440e-03
9.52809393e-01 6.87994435e-02 5.41619122e-01 6.92006230e-01
2.88583159e-01 -1.05232060e+00 4.90197599e-01 5.60738266e-01
1.05610752e+00 -1.61035132e+00 5.33810072e-02 -1.56717464e-01
-5.70465386e-01 8.68697166e-01 8.64046812e-01 -7.39165008e-01
1.07421231e+00 -5.89123741e-02 2.23732144e-01 -1.28478348e-01
-1.11785865e+00 -3.21110291e-03 1.38686344e-01 5.36122620e-01
3.92691672e-01 -8.66122916e-02 5.65320849e-01 7.20882714e-02
-9.53579098e-02 1.96772620e-01 1.97280630e-01 8.55387151e-01
-2.54239254e-02 -1.04431963e+00 1.31722376e-01 6.93087459e-01
-3.43483388e-01 -4.45039213e-01 -2.04495728e-01 6.91485167e-01
-7.87297487e-02 6.30060554e-01 2.48891771e-01 -4.28736210e-01
1.47413522e-01 2.16982082e-01 4.94557053e-01 -9.77294981e-01
-6.70957565e-01 -3.03737253e-01 -5.96635491e-02 -8.24583054e-01
-6.88931704e-01 -4.13834572e-01 -1.02241945e+00 -7.27653801e-02
-1.71114922e-01 -1.16998643e-01 6.64817274e-01 7.43804038e-01
5.60375810e-01 2.75697708e-01 9.15938437e-01 -1.13527262e+00
-5.83661020e-01 -1.02727091e+00 -4.18846697e-01 5.39947510e-01
1.80711567e-01 -3.89276892e-01 -4.23708767e-01 2.61026263e-01] | [9.522059440612793, 2.579665422439575] |
9c485be4-60da-4cb6-b444-f181e1f19a8a | low-resource-machine-translation-for-low | 2103.13272 | null | https://arxiv.org/abs/2103.13272v2 | https://arxiv.org/pdf/2103.13272v2.pdf | Low-Resource Machine Translation Training Curriculum Fit for Low-Resource Languages | We conduct an empirical study of neural machine translation (NMT) for truly low-resource languages, and propose a training curriculum fit for cases when both parallel training data and compute resource are lacking, reflecting the reality of most of the world's languages and the researchers working on these languages. Previously, unsupervised NMT, which employs back-translation (BT) and auto-encoding (AE) tasks has been shown barren for low-resource languages. We demonstrate that leveraging comparable data and code-switching as weak supervision, combined with BT and AE objectives, result in remarkable improvements for low-resource languages even when using only modest compute resources. The training curriculum proposed in this work achieves BLEU scores that improve over supervised NMT trained on the same backbone architecture by +12.2 BLEU for English to Gujarati and +3.7 BLEU for English to Kazakh, showcasing the potential of weakly-supervised NMT for the low-resource languages. When trained on supervised data, our training curriculum achieves a new state-of-the-art result on the Somali dataset (BLEU of 29.3 for Somali to English). We also observe that adding more time and GPUs to training can further improve performance, which underscores the importance of reporting compute resource usage in MT research. | ['Derry Wijaya', 'Alexander Gregory Jones', 'Siyang Li', 'Isidora Chara Tourni', 'Afra Feyza Akyürek', 'Garry Kuwanto'] | 2021-03-24 | null | null | null | null | ['low-resource-neural-machine-translation', 'cross-lingual-bitext-mining'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.34212101e-01 -2.04094782e-01 -5.04589081e-01 -1.91100091e-01
-1.13296342e+00 -7.44654536e-01 7.11080909e-01 -1.77205548e-01
-8.31985533e-01 7.42258132e-01 2.71567404e-01 -1.00840342e+00
3.92165363e-01 -5.36342144e-01 -1.01712179e+00 -4.36318278e-01
1.67244300e-01 8.71653140e-01 -2.98743844e-01 -4.96163934e-01
3.09257004e-02 6.75064400e-02 -1.05199432e+00 3.62861902e-01
1.09809113e+00 2.31577173e-01 3.27911824e-01 6.59723282e-01
-1.39143571e-01 4.98896003e-01 -4.54579800e-01 -6.79440618e-01
3.99131894e-01 -3.82876307e-01 -8.29590440e-01 -3.57149482e-01
7.52478778e-01 -2.27129698e-01 -2.56597519e-01 7.29267120e-01
6.33139670e-01 -1.69393510e-01 4.57529068e-01 -5.91495395e-01
-8.59928131e-01 8.92514288e-01 -5.38392901e-01 1.92929536e-01
-6.18334599e-02 1.87435314e-01 8.95150125e-01 -1.05628145e+00
7.07282305e-01 9.32708383e-01 7.41861284e-01 6.05840564e-01
-1.18348169e+00 -7.35962808e-01 -3.39523941e-01 -8.94411355e-02
-1.19815636e+00 -6.41493976e-01 2.42822036e-01 -2.14065343e-01
1.60635281e+00 1.91903606e-01 3.66215467e-01 1.12264967e+00
3.04313362e-01 7.24351406e-01 1.22677791e+00 -7.85706460e-01
-3.09184846e-02 1.15072623e-01 -3.07946563e-01 6.67712748e-01
2.29420215e-01 9.67774540e-02 -6.55570388e-01 6.97287098e-02
3.88624191e-01 -3.13892275e-01 1.09446436e-01 1.56698272e-01
-1.66121781e+00 7.51217842e-01 1.89714998e-01 4.06698197e-01
-1.73911482e-01 1.71881676e-01 6.18412256e-01 7.10169375e-01
7.62420416e-01 5.25364876e-01 -7.48437405e-01 -6.34666979e-01
-1.30549312e+00 -1.94131047e-01 7.47766674e-01 1.20082891e+00
7.00583398e-01 3.16437751e-01 -4.92386799e-03 9.00516510e-01
-1.58216402e-01 9.11267102e-01 6.24255776e-01 -7.38007665e-01
9.46515858e-01 4.04330224e-01 -2.91823655e-01 -2.59138882e-01
-1.37846127e-01 -7.48481870e-01 -8.83828223e-01 -1.55864969e-01
4.34915006e-01 -4.07268971e-01 -1.09150875e+00 1.73684371e+00
3.50187905e-02 -1.73477545e-01 2.81391412e-01 8.36210787e-01
4.38187331e-01 8.13551068e-01 -7.68153667e-02 -1.49955496e-01
1.16367996e+00 -1.27081859e+00 -4.67423111e-01 -4.02596056e-01
1.26109171e+00 -1.20698583e+00 1.48305976e+00 1.49118632e-01
-1.24102032e+00 -4.90511507e-01 -1.08763778e+00 -2.87165731e-01
-3.45546454e-01 3.06520134e-01 8.62008393e-01 8.30324233e-01
-1.33527720e+00 4.81536061e-01 -1.13784349e+00 -6.07194245e-01
2.52710015e-01 5.82757890e-01 -3.44212979e-01 -3.00933778e-01
-9.92961884e-01 1.17807555e+00 2.82050252e-01 -6.25452474e-02
-8.73117805e-01 -6.80772483e-01 -4.77361262e-01 -9.96818766e-02
9.99577790e-02 -5.61267316e-01 1.15779364e+00 -1.14449310e+00
-1.59900260e+00 8.87328327e-01 -2.03106716e-01 -5.19897580e-01
5.68637550e-01 -3.04840982e-01 -3.88537526e-01 -1.87004447e-01
3.84565294e-02 7.40471840e-01 4.94731098e-01 -6.22230351e-01
-5.30292869e-01 -1.66451007e-01 -1.20738670e-01 4.35274363e-01
-6.28461242e-01 4.23867017e-01 -5.38076997e-01 -5.76337159e-01
-1.20053012e-02 -1.18722427e+00 -5.45771196e-02 -6.59734309e-01
9.36885476e-02 -4.82248375e-04 5.52326202e-01 -1.00189948e+00
1.06335652e+00 -1.81260240e+00 3.36560845e-01 -1.09532505e-01
-1.05128221e-01 3.23321253e-01 -3.78816783e-01 5.30749261e-01
1.59552544e-01 6.26138300e-02 -3.43923509e-01 -3.68195385e-01
-7.27427751e-02 4.16457176e-01 -1.41837716e-01 3.71642172e-01
4.52823341e-01 1.13979387e+00 -7.72077024e-01 -4.10213888e-01
-1.30558163e-01 4.16917711e-01 -6.59870028e-01 -4.14084867e-02
-1.42153814e-01 5.14678597e-01 -5.43829333e-03 8.00370753e-01
4.64796603e-01 -2.39661321e-01 4.43472445e-01 2.54436344e-01
-2.90827155e-01 6.89412653e-01 -4.03662831e-01 2.14293551e+00
-8.13781261e-01 7.99791217e-01 2.16606557e-02 -6.90405190e-01
7.61302292e-01 3.56813431e-01 8.03386122e-02 -1.13899398e+00
-1.00817882e-01 8.51106644e-01 3.29849929e-01 -1.67605594e-01
6.82679236e-01 -2.22584978e-02 -5.17950282e-02 6.74546123e-01
2.43515924e-01 -1.64763674e-01 3.34586531e-01 1.55429974e-01
1.04729044e+00 4.23534483e-01 -2.66626328e-01 -5.56364596e-01
1.73389673e-01 3.86692554e-01 3.08821738e-01 6.00703239e-01
9.83037949e-02 3.34082007e-01 6.11690152e-03 -5.24807751e-01
-1.68548572e+00 -7.83372998e-01 -6.95633739e-02 1.60090935e+00
-5.30013323e-01 -4.66195285e-01 -9.04369652e-01 -6.89978838e-01
-4.03524280e-01 5.99510133e-01 -1.97687298e-01 -2.67645977e-02
-1.16995311e+00 -1.07263136e+00 7.88960993e-01 3.56546670e-01
5.31451523e-01 -7.39279747e-01 -4.47283447e-01 1.72220498e-01
-2.87749171e-01 -1.12199414e+00 -5.99195898e-01 5.55406332e-01
-1.17152643e+00 -1.89349785e-01 -9.77918148e-01 -1.04332769e+00
5.95877886e-01 6.42075986e-02 1.46095490e+00 7.16169029e-02
2.56778479e-01 -1.58646137e-01 -2.19821304e-01 -1.89084873e-01
-7.08525479e-01 8.56123269e-01 2.87240922e-01 -6.45205379e-01
3.79802048e-01 -5.47062933e-01 -3.13006490e-01 1.15200788e-01
-5.43745637e-01 5.36277950e-01 9.96461451e-01 9.98254597e-01
3.58754277e-01 -5.20346880e-01 3.49089324e-01 -1.00750709e+00
3.16326827e-01 -3.66954267e-01 -4.58056360e-01 3.50098103e-01
-9.46785271e-01 1.87853023e-01 7.54932761e-01 -5.61859608e-01
-8.64091873e-01 -2.28338793e-01 -7.18540326e-02 -1.09870091e-01
2.17133835e-01 6.92606032e-01 2.09511340e-01 -3.97710688e-02
8.17545295e-01 3.63569349e-01 -3.00914437e-01 -3.93167317e-01
3.07560444e-01 9.06850159e-01 4.73584175e-01 -9.06082451e-01
6.85418010e-01 1.71646595e-01 -1.82420775e-01 -5.69384217e-01
-6.39108062e-01 -1.75304562e-01 -6.84964061e-01 2.49713704e-01
6.81189418e-01 -1.26987028e+00 -1.24899790e-01 1.41357720e-01
-1.01191902e+00 -7.24442899e-01 -2.27547809e-02 7.46603191e-01
-5.15623868e-01 1.62107989e-01 -9.59988058e-01 -4.58145201e-01
-7.72154808e-01 -1.27407169e+00 9.92045760e-01 -2.22746953e-01
-2.87364721e-01 -9.81060028e-01 7.99058527e-02 6.66600347e-01
8.07008743e-01 -2.90512413e-01 1.12630498e+00 -5.49705386e-01
-4.75927681e-01 1.51022092e-01 -3.28705221e-01 3.17584813e-01
-7.69934654e-02 -4.24602598e-01 -7.91835248e-01 -7.73175120e-01
-8.77267048e-02 -4.82408762e-01 5.99448442e-01 2.48302072e-02
5.40714025e-01 -3.19864601e-01 5.93448281e-02 8.54476035e-01
1.45444787e+00 1.52147806e-03 3.26216072e-01 5.43057740e-01
8.16669941e-01 3.55288327e-01 3.47574145e-01 -8.78654495e-02
4.30234224e-01 6.80018365e-01 -1.12108387e-01 -4.44006354e-01
-4.14158791e-01 -2.78145701e-01 7.26693153e-01 1.80693817e+00
-2.26241127e-01 2.10671350e-02 -1.37748981e+00 6.81420565e-01
-1.62327588e+00 -3.75501126e-01 -1.60191208e-01 2.24290681e+00
1.26223147e+00 1.54595941e-01 -3.02260015e-02 -2.67923087e-01
4.36878115e-01 -2.78238297e-01 -3.71082097e-01 -9.08711553e-01
-3.48323792e-01 7.20918357e-01 9.09946799e-01 4.48009729e-01
-6.50739253e-01 1.39530838e+00 6.66918325e+00 8.36711645e-01
-1.37027490e+00 5.73845625e-01 7.03047931e-01 -2.59868592e-01
-2.67922878e-01 1.90414399e-01 -6.78676069e-01 4.16848302e-01
1.83746946e+00 -1.61079019e-02 7.83153236e-01 4.97545153e-01
7.93971419e-02 2.26056814e-01 -1.06910932e+00 7.40510941e-01
7.75088519e-02 -1.37883043e+00 5.17957211e-02 2.28823066e-01
1.21983254e+00 8.64617169e-01 2.31330171e-01 5.62783957e-01
5.97087502e-01 -1.19041121e+00 7.43871212e-01 -1.67432234e-01
1.29783261e+00 -9.13171828e-01 7.74284780e-01 4.44275975e-01
-7.26005375e-01 2.96793789e-01 -5.18445373e-01 -2.79276311e-01
-2.18776479e-01 4.38645869e-01 -1.12015045e+00 4.97444421e-01
5.48351705e-01 4.88806278e-01 -5.17461538e-01 4.05003905e-01
-1.79966018e-01 1.18691790e+00 -4.20722157e-01 -7.01449960e-02
6.19576037e-01 -2.46221676e-01 1.86279207e-01 1.67716968e+00
4.83324587e-01 -3.25950682e-01 6.05297387e-02 5.32272696e-01
-3.90636921e-01 4.31159198e-01 -4.61606294e-01 -2.60079533e-01
3.09337229e-01 1.01575160e+00 -5.36604643e-01 -5.07795453e-01
-5.11246383e-01 1.35640836e+00 6.03898883e-01 2.75886595e-01
-8.18892241e-01 -1.07239924e-01 4.29036528e-01 -9.08653364e-02
1.66620687e-01 -7.21919358e-01 -6.00187600e-01 -1.35792458e+00
3.18552777e-02 -1.27702975e+00 1.33636281e-01 -4.78736371e-01
-8.46973836e-01 7.04775810e-01 -4.56881702e-01 -1.05667341e+00
-2.62431532e-01 -5.78541458e-01 -2.26178274e-01 1.11989594e+00
-1.56083071e+00 -1.45621109e+00 3.48393857e-01 3.80619466e-01
7.19329119e-01 -4.53364640e-01 1.07852221e+00 5.57686388e-01
-4.66016084e-01 1.03379583e+00 6.48125589e-01 9.33367759e-02
8.67344975e-01 -1.15768945e+00 9.39176857e-01 9.69886363e-01
5.65481842e-01 9.12251115e-01 4.70854104e-01 -6.18179798e-01
-1.95783019e+00 -1.02334559e+00 1.54619789e+00 -7.31193960e-01
7.43697107e-01 -6.47469461e-01 -5.04531384e-01 7.57722735e-01
5.37942827e-01 -3.84319782e-01 6.13063633e-01 3.89677763e-01
-4.54424143e-01 1.80291921e-01 -7.99295545e-01 7.64965296e-01
1.14548695e+00 -7.73784637e-01 -4.19613689e-01 5.44414639e-01
8.71638179e-01 -5.92470944e-01 -9.33236122e-01 1.56894028e-01
5.27057528e-01 -4.35394019e-01 6.26893699e-01 -6.66149795e-01
7.73805857e-01 -2.51570460e-03 -4.35866565e-01 -1.32952476e+00
-1.23385228e-01 -8.22133422e-01 1.50779728e-02 1.02690673e+00
8.51426125e-01 -4.15303439e-01 9.53452528e-01 1.11612990e-01
-3.46207410e-01 -6.63683832e-01 -9.71463144e-01 -8.42748463e-01
6.98517382e-01 -4.07464623e-01 4.30312544e-01 1.31486094e+00
-6.38078600e-02 6.52496815e-01 -4.55426753e-01 -1.41114667e-01
3.58449012e-01 1.40689416e-02 7.51871824e-01 -6.82567239e-01
-5.47050774e-01 -3.20824921e-01 1.64689869e-02 -1.01485372e+00
2.72787631e-01 -1.45251286e+00 -1.03935890e-01 -1.35316706e+00
4.91703451e-01 -5.08909881e-01 -5.89741617e-02 7.55660653e-01
-7.09120110e-02 5.96080780e-01 2.42264524e-01 5.13518035e-01
-2.36602128e-01 2.65548527e-01 1.15266788e+00 -2.24564284e-01
8.54787510e-03 -5.00373304e-01 -6.07745230e-01 3.12433869e-01
8.12883973e-01 -5.35053492e-01 -1.85272500e-01 -1.38364553e+00
3.77599150e-01 -1.70662135e-01 -3.18650752e-01 -9.87773061e-01
2.42668495e-01 -1.13253057e-01 2.49864057e-01 -3.95593435e-01
3.93202342e-02 -6.86116099e-01 -3.45735066e-02 6.10876262e-01
-3.30543756e-01 6.57514215e-01 3.59029114e-01 -1.28652602e-01
-4.78484333e-02 6.07032925e-02 5.58842778e-01 -1.75806195e-01
-4.61054832e-01 3.96984033e-02 -3.13062251e-01 1.68383434e-01
3.84112507e-01 -1.81183681e-01 -4.32341635e-01 -1.58295915e-01
-1.92205340e-01 4.08508517e-02 6.02022469e-01 3.19179326e-01
2.12005377e-01 -1.23157477e+00 -1.28311205e+00 3.83665472e-01
1.69574767e-02 -2.07948849e-01 -3.00553322e-01 9.98632491e-01
-8.17183673e-01 5.85736632e-01 -2.67184198e-01 -7.01493204e-01
-1.01990712e+00 2.22259313e-01 -7.55726397e-02 -5.15691698e-01
-3.61857951e-01 7.71870553e-01 -3.30643892e-01 -8.72289121e-01
-1.11528873e-01 -1.47860333e-01 6.02115929e-01 -3.39808434e-01
9.96083915e-02 2.87332743e-01 5.89457572e-01 -5.35062194e-01
-1.84518903e-01 3.02598953e-01 -2.75962889e-01 -4.32629108e-01
1.34579992e+00 1.65793095e-02 -2.73951322e-01 3.48168105e-01
1.19021106e+00 2.31698632e-01 -8.23259890e-01 -3.19461107e-01
2.12146476e-01 -7.12283030e-02 1.51446432e-01 -1.15249145e+00
-7.72977948e-01 1.08601034e+00 6.62497759e-01 -6.70070946e-01
1.09062386e+00 -2.91098744e-01 9.97235119e-01 6.12090588e-01
6.17932618e-01 -1.21660113e+00 -2.64533311e-01 1.11417639e+00
3.84946048e-01 -1.18391526e+00 -1.47478402e-01 7.84094706e-02
-4.58316833e-01 9.42487180e-01 4.39809322e-01 8.61787722e-02
5.02010994e-02 5.65903485e-01 1.98387057e-01 2.13922098e-01
-9.43388522e-01 1.26700019e-02 1.67077139e-01 3.51302296e-01
1.04330838e+00 2.90349364e-01 -5.67577243e-01 4.82989885e-02
-6.30589664e-01 -1.70270473e-01 3.25397074e-01 9.83008265e-01
-1.84236825e-01 -1.41009259e+00 -2.37111613e-01 2.91483343e-01
-7.85446167e-01 -8.74639213e-01 -3.43428016e-01 7.27297485e-01
3.85413691e-02 8.17509651e-01 2.34491169e-01 -4.62932169e-01
-6.23677019e-03 3.34708393e-01 6.35945737e-01 -6.74361229e-01
-1.07791686e+00 2.33506411e-01 3.29964310e-01 -2.12819025e-01
-2.21898854e-01 -4.40571219e-01 -1.08442271e+00 -7.88810551e-01
-2.15859413e-01 1.71326473e-01 1.09836817e+00 8.36213112e-01
4.38196182e-01 2.27123260e-01 3.93171906e-01 -6.27011418e-01
-6.28779650e-01 -1.17808092e+00 1.44703493e-01 1.52669400e-01
2.93613989e-02 4.46150377e-02 -4.21756878e-03 3.70208845e-02] | [11.546218872070312, 10.31428050994873] |
2eee19f9-01e3-427f-b149-de915edb2418 | physics-informed-neural-network-for-seismic | 2305.05150 | null | https://arxiv.org/abs/2305.05150v1 | https://arxiv.org/pdf/2305.05150v1.pdf | Physics-informed neural network for seismic wave inversion in layered semi-infinite domain | Estimating the material distribution of Earth's subsurface is a challenging task in seismology and earthquake engineering. The recent development of physics-informed neural network (PINN) has shed new light on seismic inversion. In this paper, we present a PINN framework for seismic wave inversion in layered (1D) semi-infinite domain. The absorbing boundary condition is incorporated into the network as a soft regularizer for avoiding excessive computation. In specific, we design a lightweight network to learn the unknown material distribution and a deep neural network to approximate solution variables. The entire network is end-to-end and constrained by both sparse measurement data and the underlying physical laws (i.e., governing equations and initial/boundary conditions). Various experiments have been conducted to validate the effectiveness of our proposed approach for inverse modeling of seismic wave propagation in 1D semi-infinite domain. | ['Yang Liu', 'Hao Sun', 'Chengping Rao', 'Pu Ren'] | 2023-05-09 | null | null | null | null | ['seismic-inversion'] | ['miscellaneous'] | [ 9.33333039e-02 2.43394580e-02 4.47939962e-01 -1.23136491e-01
-5.73558509e-01 1.45438135e-01 7.04684388e-03 -5.64526081e-01
-2.39687935e-01 8.91632617e-01 2.78571606e-01 -2.29775876e-01
-3.43972743e-01 -9.03145254e-01 -8.62545669e-01 -9.69042242e-01
-3.43358636e-01 3.93997967e-01 6.17168806e-02 -2.41992667e-01
3.14170301e-01 3.70543391e-01 -8.46290827e-01 -8.13577920e-02
7.32189357e-01 1.03053045e+00 1.08839057e-01 6.09645009e-01
2.01202288e-01 8.36869836e-01 -1.46669284e-01 3.19454372e-01
5.13733402e-02 -1.41647086e-01 -7.86035836e-01 -2.32203409e-01
-2.11773902e-01 -6.44919932e-01 -5.52476585e-01 8.53865802e-01
1.01717043e+00 3.67044061e-01 9.67174232e-01 -7.50542164e-01
-6.30374849e-01 5.86454928e-01 -9.53662574e-01 2.48253822e-01
8.52648914e-02 -3.07166427e-02 3.45685601e-01 -1.23881638e+00
-1.93472039e-02 8.51951718e-01 1.28743160e+00 2.90925115e-01
-9.62449253e-01 -4.84541029e-01 -2.26994798e-01 4.22646403e-02
-1.58284211e+00 -4.25486445e-01 1.35003245e+00 -4.72111553e-01
7.81123757e-01 -7.36034065e-02 4.36564863e-01 7.41111517e-01
2.46492967e-01 4.77026969e-01 6.77749574e-01 -4.24402773e-01
4.45879608e-01 -3.42804402e-01 7.28358701e-02 5.11053145e-01
4.45961177e-01 7.51437470e-02 -5.28097391e-01 -4.19726729e-01
1.15650928e+00 -1.72195733e-01 -4.26464587e-01 8.64799097e-02
-8.55476677e-01 6.78886890e-01 2.75054574e-01 1.22638702e-01
-7.35026062e-01 5.53476810e-01 2.89851487e-01 -1.52727798e-01
7.50291467e-01 1.94275811e-01 -2.39974067e-01 1.91900477e-01
-1.26909721e+00 4.73495483e-01 9.66161609e-01 4.86602336e-01
6.94482446e-01 8.44628453e-01 3.35603327e-01 7.90303767e-01
6.89018130e-01 8.44568729e-01 1.23058841e-01 -1.07177269e+00
3.11204910e-01 -1.37497127e-01 3.46180737e-01 -1.31988668e+00
-4.69882905e-01 -6.06248677e-01 -1.36071026e+00 9.13559869e-02
1.36172678e-02 -7.53962040e-01 -8.56652856e-01 1.68344843e+00
4.03807282e-01 1.11715770e+00 2.05737591e-01 1.06776392e+00
9.81658399e-01 9.53018606e-01 -2.46969298e-01 1.39040621e-02
9.66046512e-01 -4.64123487e-01 -6.09166622e-01 -4.15857553e-01
4.97505754e-01 -4.46145803e-01 5.04183888e-01 1.11646228e-01
-1.32745242e+00 -2.91301519e-01 -1.02954328e+00 1.89756855e-01
1.86138779e-01 -3.10545087e-01 4.33360696e-01 2.18143836e-01
-8.91959906e-01 4.77326661e-01 -1.14519310e+00 5.57970583e-01
1.96602777e-01 3.00547302e-01 -5.22502698e-02 -6.08547442e-02
-1.63065815e+00 4.91248518e-01 1.88269436e-01 9.44031894e-01
-9.88285780e-01 -1.00856328e+00 -8.55098546e-01 1.55051604e-01
2.14015357e-02 -8.30979705e-01 1.04629016e+00 -2.09499374e-01
-1.59121370e+00 3.65369231e-01 1.43897131e-01 -1.62866250e-01
3.00625384e-01 -1.89724028e-01 -1.96657196e-01 2.09209949e-01
1.15262188e-01 -2.49007002e-01 5.69134414e-01 -1.39312685e+00
1.84423894e-01 2.51577180e-02 -1.97613642e-01 -1.02652088e-02
-2.08882615e-01 -3.25244009e-01 -4.50466499e-02 -8.63759398e-01
6.01936758e-01 -6.01886094e-01 -4.62239802e-01 -1.88041821e-01
-6.08137369e-01 3.69220048e-01 6.42245114e-01 -1.22655165e+00
1.03997242e+00 -1.84165704e+00 6.70730248e-02 6.03077829e-01
2.28919983e-01 5.41195199e-02 1.11706585e-01 4.71736670e-01
-1.17511094e-01 -1.55167088e-01 -1.01483011e+00 -1.72466010e-01
-1.32194594e-01 1.44844711e-01 -2.84476668e-01 7.51437664e-01
7.86084011e-02 5.41070521e-01 -5.94605029e-01 -3.59416962e-01
-1.05943955e-01 8.37880313e-01 -8.15317750e-01 2.25531489e-01
4.21303585e-02 7.23997295e-01 -1.08157980e+00 5.78610301e-01
1.39113939e+00 -1.50699750e-01 -3.09344441e-01 -2.03354761e-01
-3.20726424e-01 4.28990126e-02 -1.53009522e+00 1.46596110e+00
-8.13278913e-01 3.24558049e-01 6.21255934e-01 -1.60526299e+00
1.02887058e+00 8.42131674e-01 5.62566996e-01 -5.89298248e-01
1.70515642e-01 5.27234495e-01 -8.82516056e-02 -1.13698292e+00
1.19100407e-01 -5.07107615e-01 1.42895460e-01 7.00684607e-01
-3.65804136e-01 -4.24073845e-01 -2.86419123e-01 -1.59812003e-01
9.62670743e-01 1.86747938e-01 -2.85269827e-01 -5.34134865e-01
5.81393480e-01 -7.26383924e-02 8.12097669e-01 7.08892345e-01
3.68614167e-01 8.12692165e-01 1.67343512e-01 -6.66029513e-01
-1.00846195e+00 -9.25815880e-01 -1.62734270e-01 5.62814891e-01
-6.40502665e-04 5.98171532e-01 -6.84903443e-01 4.17219967e-01
-1.49054378e-01 4.60155636e-01 -5.01504838e-01 -2.28462741e-01
-1.07843769e+00 -1.21060300e+00 8.39462638e-01 7.07385957e-01
7.09911764e-01 -9.47106063e-01 -5.20209551e-01 5.18451571e-01
-5.66079974e-01 -1.03222048e+00 2.74075624e-02 1.02925457e-01
-8.15216064e-01 -5.89411378e-01 -1.06162477e+00 -9.51790094e-01
6.57004595e-01 -2.15385064e-01 9.75018501e-01 -5.38687147e-02
-2.27933228e-02 1.40702620e-01 -1.70708429e-02 -3.96406949e-01
-1.68331519e-01 -4.16502096e-02 3.66072804e-02 1.93417102e-01
-3.67383420e-01 -1.13665020e+00 -8.60374391e-01 2.29492009e-01
-8.71779263e-01 6.10993095e-02 3.51351440e-01 6.56128585e-01
3.74210566e-01 2.65323430e-01 6.55563712e-01 -5.68561733e-01
7.10506260e-01 -8.92280877e-01 -5.47840297e-01 -1.30004883e-01
3.25769991e-01 9.88144800e-02 4.26011890e-01 -5.50400674e-01
-1.54110646e+00 -2.36821458e-01 -6.07509136e-01 -4.42068055e-02
8.56181607e-02 1.40217555e+00 9.42120776e-02 -4.54293966e-01
5.01225829e-01 3.52309912e-01 -2.98964411e-01 -7.26689816e-01
-3.86143655e-01 5.92274368e-01 8.23300779e-01 -9.95534420e-01
7.69335210e-01 6.29244566e-01 4.33458894e-01 -1.23765206e+00
-7.53184915e-01 -2.45048568e-01 -3.18032324e-01 -1.36888430e-01
5.44587851e-01 -8.50782752e-01 -8.82332087e-01 9.55958605e-01
-1.31591523e+00 -5.35959721e-01 1.58073470e-01 9.16851103e-01
-4.35810864e-01 3.58158559e-01 -7.33176947e-01 -1.07714140e+00
-5.20367980e-01 -8.87034655e-01 7.28432715e-01 1.81990683e-01
3.08938399e-02 -1.34049511e+00 3.15070838e-01 1.40141860e-01
8.70543361e-01 7.54446864e-01 7.41829157e-01 2.08790395e-02
-3.53769630e-01 -3.36746991e-01 -3.57122719e-01 2.44999737e-01
-1.24533266e-01 -2.89584905e-01 -9.61306751e-01 -1.42398449e-02
8.28738511e-01 -2.26453990e-01 9.13574576e-01 1.15806103e+00
9.53070998e-01 -2.04080090e-01 -1.72002420e-01 1.04982042e+00
1.65189910e+00 -2.55802050e-02 5.44002771e-01 1.49571300e-01
7.33464181e-01 3.88576388e-01 -1.28934652e-01 9.01416600e-01
2.73921520e-01 3.08789480e-02 4.27271634e-01 -1.60776362e-01
1.65280581e-01 2.10412070e-01 -1.26365155e-01 1.09590876e+00
-6.34346664e-01 -4.24283564e-01 -1.41146588e+00 6.55104637e-01
-1.51007187e+00 -7.67205298e-01 -3.73007715e-01 1.71791542e+00
7.57789552e-01 7.69935325e-02 -5.41793883e-01 3.24394494e-01
6.46181643e-01 1.03866503e-01 -5.83810508e-01 -8.47218856e-02
-4.91737910e-02 2.96216786e-01 3.85538936e-01 7.34745383e-01
-8.19593847e-01 1.58875301e-01 6.15280962e+00 4.30080116e-01
-1.50900531e+00 8.75921845e-02 5.71632504e-01 4.50165629e-01
-5.44991791e-01 -2.54781365e-01 -4.29487735e-01 4.13627535e-01
6.77445769e-01 9.94623229e-02 1.85021490e-01 3.56308281e-01
6.75612330e-01 -1.15060814e-01 -5.51731288e-01 5.42782307e-01
-2.66334891e-01 -1.54997694e+00 -2.38543615e-01 -3.26904386e-01
8.02882791e-01 1.68383658e-01 -8.33211318e-02 -1.80071339e-01
2.36805808e-02 -8.69738877e-01 6.76742196e-01 8.91862810e-01
8.22840989e-01 -6.36595786e-01 1.02996778e+00 5.94588995e-01
-1.11904800e+00 1.35012016e-01 -1.50482327e-01 -5.03102005e-01
8.47979367e-01 8.74587119e-01 -4.13145661e-01 5.26213527e-01
6.95162117e-01 5.32690406e-01 5.50963640e-01 1.14981782e+00
5.11858463e-02 1.12323987e+00 -6.65035725e-01 2.74464101e-01
5.35273135e-01 -2.74163157e-01 7.35600770e-01 1.04555106e+00
5.48450172e-01 5.82790494e-01 7.18997419e-02 1.16684151e+00
6.22579567e-02 -3.81863803e-01 -2.78192520e-01 3.20857048e-01
1.74292251e-01 6.03342831e-01 -5.93410492e-01 5.52860387e-02
-3.28331977e-01 1.43101871e-01 -1.84151769e-01 9.26239729e-01
-8.86474669e-01 -4.70897257e-01 1.94400132e-01 4.35174137e-01
1.79434776e-01 -4.65087414e-01 -4.94787216e-01 -8.07412028e-01
1.28799170e-01 -1.12612553e-01 -1.76134519e-02 -8.12138855e-01
-1.15703487e+00 3.92882526e-01 2.57018954e-01 -9.01410699e-01
-4.24891885e-04 -5.31844020e-01 -1.15706325e+00 1.19511282e+00
-1.86069489e+00 -8.72356236e-01 -3.37436974e-01 3.98941815e-01
-2.26100571e-02 7.59631097e-02 5.35154700e-01 5.91098189e-01
-5.28635859e-01 2.66607385e-02 4.01408374e-01 5.30822754e-01
9.12498757e-02 -5.40626287e-01 4.65019971e-01 7.61288881e-01
-9.45261776e-01 4.45897818e-01 1.13183641e+00 -7.68560469e-01
-1.36844993e+00 -7.53500462e-01 4.63470936e-01 5.19405782e-01
6.31211102e-01 -6.09612502e-02 -1.47877252e+00 5.13818264e-01
1.82649344e-02 4.23637658e-01 3.62404406e-01 -3.87340307e-01
1.88699409e-01 5.86611293e-02 -1.15031350e+00 2.41210088e-01
5.47746599e-01 -1.46507382e-01 -4.32785392e-01 2.00935021e-01
4.71853375e-01 -6.44830883e-01 -7.93441713e-01 6.94194496e-01
4.58691150e-01 -6.74342752e-01 1.42351496e+00 -2.02735946e-01
7.97581375e-01 -6.33177459e-02 -9.97707397e-02 -1.12378979e+00
-2.85707116e-01 -6.50055885e-01 -1.70187622e-01 8.16106677e-01
2.25961149e-01 -8.61020505e-01 9.97804105e-01 7.11589456e-01
-6.99747384e-01 -9.86271441e-01 -1.04587948e+00 -4.56536263e-01
2.33101025e-01 -5.79505563e-01 4.55460936e-01 8.39938164e-01
-4.14652318e-01 -2.03755781e-01 -6.69790685e-01 6.96200728e-01
1.10228825e+00 -1.41465560e-01 9.43972096e-02 -1.13641095e+00
-3.08692008e-01 4.81475424e-03 8.03746432e-02 -1.21403742e+00
1.62745148e-01 -3.91857475e-01 4.28670228e-01 -1.56053507e+00
-9.95540321e-02 -7.85092354e-01 -3.88394684e-01 2.85083689e-02
5.53073771e-02 3.59622657e-01 -7.58018792e-01 -1.12236194e-01
2.16929927e-01 6.87198043e-01 1.33340490e+00 -9.08339769e-02
-1.35171339e-01 2.02266321e-01 -2.45734900e-01 1.05498254e+00
9.11966801e-01 -8.37832987e-01 -3.00288558e-01 -1.27027094e+00
6.03577256e-01 6.27956986e-01 6.45335436e-01 -9.57383156e-01
5.05027533e-01 -2.53717005e-01 1.22436993e-01 -7.97521770e-01
4.60638851e-01 -5.72337210e-01 1.25929728e-01 5.50377309e-01
-4.30345759e-02 -1.79076165e-01 2.74185300e-01 3.50814402e-01
-4.40972298e-01 -4.70437050e-01 9.36102092e-01 -2.43400946e-01
-4.67958778e-01 3.48080128e-01 -4.25250947e-01 4.15300816e-01
2.42278442e-01 -3.58771145e-01 3.22646141e-01 -3.02281618e-01
-6.37597203e-01 1.12247400e-01 -2.51179278e-01 -5.36392808e-01
9.03784811e-01 -1.13381910e+00 -1.20315135e+00 4.48847830e-01
-5.46043158e-01 5.55757523e-01 7.93257713e-01 8.56036723e-01
-1.22804797e+00 -4.86359969e-02 -9.03578028e-02 -4.58853513e-01
-4.18553889e-01 -1.56420410e-01 6.70132697e-01 -1.12698123e-01
-6.49885535e-01 1.21747887e+00 1.48582071e-01 -5.78275084e-01
-1.19983945e-02 -2.98897058e-01 -1.13019673e-02 -4.92672741e-01
3.98288518e-01 7.83444107e-01 -1.45618483e-01 -4.93930131e-01
-3.20581436e-01 8.64884198e-01 4.35043156e-01 -2.08164364e-01
2.10687280e+00 -9.72480103e-02 -4.19972181e-01 2.99243569e-01
1.21528375e+00 -2.66554862e-01 -1.14183772e+00 -3.92516136e-01
-3.38768542e-01 -2.54969839e-02 5.74400961e-01 -2.13098377e-01
-1.22980952e+00 9.21454191e-01 -2.00123582e-02 7.85940140e-02
9.25469816e-01 -2.97033131e-01 1.16578054e+00 5.89402199e-01
-3.53093930e-02 -9.92755592e-01 4.14378047e-02 8.06571245e-01
9.32792783e-01 -1.03144503e+00 1.81825742e-01 -3.41881901e-01
7.22868741e-02 1.14875221e+00 2.95241058e-01 -4.90778446e-01
1.36301231e+00 9.44644272e-01 2.33886577e-02 -2.67949015e-01
-1.23450883e-01 6.50140703e-01 -2.70444658e-02 3.85122240e-01
5.74748337e-01 -3.46338063e-01 1.73038036e-01 6.08354867e-01
1.96454942e-01 3.87135088e-01 4.68406558e-01 1.20138633e+00
-5.33170760e-01 -3.08781564e-01 -8.01334381e-01 1.02745406e-01
-5.53973198e-01 -2.63970554e-01 5.54916322e-01 4.47307378e-01
-2.18300838e-02 5.40855169e-01 1.38695147e-02 2.46200264e-01
1.34544119e-01 -2.54750162e-01 9.16734561e-02 -2.80570686e-01
8.35302547e-02 8.92898589e-02 -8.22154209e-02 6.72314987e-02
-4.64130372e-01 -4.29474980e-01 -1.62753367e+00 -3.01631540e-01
-3.23728234e-01 4.12605584e-01 5.39397061e-01 1.11412859e+00
2.28865966e-01 7.76988029e-01 3.81098211e-01 -1.46562481e+00
-3.80030692e-01 -9.94715750e-01 -8.58825862e-01 -1.15841977e-01
7.01077223e-01 -8.56961071e-01 -5.10424376e-01 2.35770613e-01] | [6.861574649810791, 2.4897589683532715] |
7994ad1a-c8e8-4019-a911-6e665cf7e533 | factually-consistent-summarization-via | 2306.00186 | null | https://arxiv.org/abs/2306.00186v1 | https://arxiv.org/pdf/2306.00186v1.pdf | Factually Consistent Summarization via Reinforcement Learning with Textual Entailment Feedback | Despite the seeming success of contemporary grounded text generation systems, they often tend to generate factually inconsistent text with respect to their input. This phenomenon is emphasized in tasks like summarization, in which the generated summaries should be corroborated by their source article. In this work, we leverage recent progress on textual entailment models to directly address this problem for abstractive summarization systems. We use reinforcement learning with reference-free, textual entailment rewards to optimize for factual consistency and explore the ensuing trade-offs, as improved consistency may come at the cost of less informative or more extractive summaries. Our results, according to both automatic metrics and human evaluation, show that our method considerably improves the faithfulness, salience, and conciseness of the generated summaries. | ['Idan Szpektor', 'Olivier Pietquin', 'Avinatan Hassidim', 'Gal Elidan', 'Olivier Bachem', 'Nino Vieillard', 'Piotr Stanczyk', 'Sabela Ramos', 'Nikola Momchev', 'Orgad Keller', 'Léonard Hussenot', 'Sertan Girgin', 'Matthieu Geist', 'Robert Dadashi', 'Geoffrey Cideron', 'Roee Aharoni', 'Lior Shani', 'Johan Ferret', 'Paul Roit'] | 2023-05-31 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 3.77168208e-01 6.86597228e-01 -3.00257534e-01 -3.40805143e-01
-1.16189122e+00 -6.18488312e-01 1.19830918e+00 7.83202112e-01
-2.58928031e-01 1.35554278e+00 1.12770426e+00 -2.10382402e-01
-1.32175293e-02 -6.77286923e-01 -6.40350163e-01 -1.80652574e-01
2.08259881e-01 4.33343053e-01 -1.93886340e-01 -3.08904648e-01
7.86856353e-01 -2.85126902e-02 -1.38056421e+00 3.26465726e-01
1.31200099e+00 4.90993470e-01 -1.15826748e-01 7.93338716e-01
-1.87854208e-02 1.25533402e+00 -1.08061957e+00 -8.24131846e-01
-2.63036162e-01 -8.48851562e-01 -1.08862209e+00 5.10633253e-02
6.60301626e-01 -4.54140157e-01 -1.43759027e-01 1.01685286e+00
4.58775043e-01 2.59266943e-01 9.70013320e-01 -1.16392124e+00
-9.38825846e-01 1.36301625e+00 -3.26268226e-01 2.45419919e-01
7.40298569e-01 2.92055339e-01 1.63548744e+00 -5.35273015e-01
8.69695663e-01 1.14325953e+00 4.77168173e-01 3.98829311e-01
-1.21602845e+00 -9.48567912e-02 9.53038484e-02 1.28700053e-02
-7.81475902e-01 -8.99260104e-01 7.58160114e-01 -6.26655892e-02
1.16058123e+00 4.68345016e-01 6.77347422e-01 1.18130171e+00
5.19330025e-01 8.25119078e-01 6.96486413e-01 -4.46249783e-01
2.89647222e-01 4.55380529e-02 -8.73948727e-03 4.32661295e-01
7.65963912e-01 -2.19888061e-01 -6.95701838e-01 -7.92809762e-03
1.26802474e-01 -5.74672282e-01 -3.81570816e-01 -4.10326570e-02
-1.17406249e+00 8.77781630e-01 2.37941936e-01 3.16013217e-01
-6.52642965e-01 2.43695170e-01 4.51754421e-01 2.02057704e-01
6.13789320e-01 1.16037440e+00 5.60451746e-02 -2.50830531e-01
-1.38389754e+00 8.51018608e-01 9.99113917e-01 9.40875709e-01
4.55597997e-01 1.37538716e-01 -7.20197618e-01 5.65635324e-01
9.15090963e-02 3.13916385e-01 6.85984373e-01 -1.21086991e+00
7.69327879e-01 5.82450628e-01 3.19431752e-01 -1.12365556e+00
-1.09743990e-01 -3.67996901e-01 -6.79859936e-01 -1.39342546e-01
6.52327910e-02 -2.40976691e-01 -3.54595393e-01 1.59461939e+00
-9.62374508e-02 -4.39976901e-01 3.94603431e-01 6.20428801e-01
7.79999435e-01 7.11964369e-01 1.74338415e-01 -5.13265014e-01
9.35735404e-01 -8.01457345e-01 -1.01515746e+00 -3.39889228e-01
7.49781966e-01 -7.89836466e-01 1.09850013e+00 1.15087286e-01
-1.62509656e+00 -1.84742883e-01 -1.20604229e+00 -2.07687557e-01
3.18769645e-03 -1.22457538e-02 3.49934191e-01 3.07117850e-01
-9.10966456e-01 1.02757287e+00 -4.67925519e-01 -1.52577803e-01
4.70209122e-01 -1.67330906e-01 -8.45366418e-02 1.90763339e-01
-1.16856599e+00 1.12349427e+00 5.06593943e-01 -1.85445532e-01
-3.99427831e-01 -7.27247953e-01 -9.27075207e-01 2.17925832e-01
5.85011244e-01 -1.07779169e+00 1.74971569e+00 -9.04148996e-01
-1.42433977e+00 5.86011767e-01 -9.29061174e-02 -7.26093411e-01
7.09996462e-01 -5.44455469e-01 -8.51924866e-02 1.42342493e-01
4.15786058e-01 6.63084745e-01 5.32584548e-01 -1.16170633e+00
-5.53309977e-01 -4.89673428e-02 1.06576934e-01 6.12109065e-01
-1.82355583e-01 -4.15259808e-01 2.26206213e-01 -7.50367224e-01
-3.23002338e-01 -5.06763399e-01 -1.47222385e-01 -3.87943059e-01
-8.72152209e-01 -4.38474685e-01 3.49907458e-01 -7.40470767e-01
1.38377953e+00 -1.57377875e+00 1.58023149e-01 -1.68973446e-01
1.55687779e-01 -2.70666797e-02 -8.64663422e-02 8.17113638e-01
3.67211014e-01 4.12707359e-01 -3.39011073e-01 -3.13215196e-01
2.17353970e-01 5.52932359e-02 -6.36448145e-01 1.26794055e-01
5.43905973e-01 1.06778657e+00 -1.18412340e+00 -8.54269385e-01
-1.13103531e-01 -3.25097367e-02 -7.23506033e-01 2.82154262e-01
-5.62157869e-01 6.28878828e-03 -2.70395339e-01 2.46668026e-01
1.58340950e-02 -2.32886434e-01 9.91896093e-02 -1.15685835e-01
-1.27175242e-01 8.19864810e-01 -7.72775948e-01 1.59556198e+00
-4.04537141e-01 8.55645478e-01 -4.44140226e-01 -6.74995005e-01
6.41955912e-01 3.03449541e-01 5.14206067e-02 -7.75885284e-01
-3.30357030e-02 1.08169310e-01 -5.82411811e-02 -4.91829067e-01
1.52751243e+00 -3.14227730e-01 -8.29425678e-02 7.62795806e-01
-8.16676393e-02 -7.07833946e-01 6.28327668e-01 7.45575845e-01
8.02212238e-01 2.52670377e-01 5.30421555e-01 -3.28900576e-01
2.47574791e-01 3.17113847e-01 1.69399530e-01 9.18936372e-01
1.12407781e-01 6.59684002e-01 8.84215593e-01 7.69406930e-02
-1.28560555e+00 -7.58752942e-01 1.54532701e-01 6.07156694e-01
1.16264978e-02 -6.50617361e-01 -8.46017838e-01 -5.13216257e-01
-9.75416452e-02 1.56891441e+00 -4.82779860e-01 -3.60862225e-01
-6.52626574e-01 -3.06609541e-01 6.78557158e-01 5.42223752e-01
2.39547476e-01 -1.19617236e+00 -1.04892123e+00 3.47380012e-01
-6.31142378e-01 -9.11066771e-01 -5.77904761e-01 -2.56122082e-01
-9.95711207e-01 -9.54981267e-01 -7.15154827e-01 -2.74741352e-01
4.32516128e-01 -1.89165138e-02 1.61761141e+00 1.19792566e-01
1.91348583e-01 3.45503330e-01 -4.08160657e-01 -5.02024412e-01
-9.50512767e-01 2.80498564e-01 -1.81381494e-01 -4.52222317e-01
-1.44145593e-01 -3.81103992e-01 -4.79836017e-01 -4.19467598e-01
-1.26129687e+00 2.52226233e-01 5.06205320e-01 9.92901921e-01
2.79754043e-01 -2.92859674e-01 9.95862782e-01 -1.03398633e+00
1.46820939e+00 -4.52763706e-01 -1.26070902e-01 3.24911743e-01
-1.05299664e+00 5.11450291e-01 7.62889445e-01 -1.42173901e-01
-1.19573653e+00 -6.65150285e-01 -5.80842793e-02 8.50125477e-02
2.56835878e-01 9.23876882e-01 1.38597727e-01 7.99964666e-01
1.05955553e+00 2.33152568e-01 1.78000273e-03 1.81528047e-01
7.52876282e-01 5.92977166e-01 6.71985626e-01 -7.32188284e-01
3.96813631e-01 6.06643744e-02 -2.51747906e-01 -7.48464763e-01
-1.08909690e+00 -1.04412585e-01 -1.79955110e-01 -1.68620139e-01
4.10947353e-01 -7.25743473e-01 -2.43806168e-01 -1.68000579e-01
-1.27548313e+00 -2.11830512e-01 -7.56767511e-01 7.64025822e-02
-8.18457186e-01 6.98001325e-01 -4.84802932e-01 -7.14255929e-01
-7.88042605e-01 -7.75970340e-01 9.63187635e-01 2.85070330e-01
-1.08510077e+00 -1.00942266e+00 2.11099371e-01 1.54994562e-01
4.68257546e-01 3.71997029e-01 1.06950366e+00 -7.95125723e-01
-3.05054903e-01 -1.72007233e-01 -6.67380765e-02 1.68616220e-01
1.45862892e-01 3.96240294e-01 -5.82035661e-01 8.55573863e-02
-2.07951188e-01 -5.86289227e-01 8.00914228e-01 4.34412569e-01
6.58377588e-01 -1.16826069e+00 1.17877468e-01 -1.60287812e-01
1.24275351e+00 -4.11744446e-01 6.36078298e-01 2.48232588e-01
3.71600330e-01 8.38124156e-01 5.90285003e-01 6.58372104e-01
3.75673950e-01 4.19375449e-01 2.13569716e-01 3.88174087e-01
-3.52469265e-01 -6.56026185e-01 1.71106175e-01 7.18918622e-01
8.13652873e-02 -6.70373082e-01 -6.11070216e-01 8.39902997e-01
-1.94375753e+00 -1.41565084e+00 1.30986661e-01 1.90475047e+00
1.21327066e+00 3.37031066e-01 1.31114155e-01 2.73273110e-01
4.61957186e-01 6.00443304e-01 -3.62767428e-01 -6.78967834e-01
-2.00452596e-01 -5.45987636e-02 -5.23851905e-03 7.31111586e-01
-4.90265578e-01 8.15502524e-01 6.65908051e+00 4.58304614e-01
-8.04382443e-01 -4.11238879e-01 6.79982245e-01 -4.05595005e-01
-9.71634686e-01 1.19317710e-01 -3.55475247e-01 3.73392463e-01
1.04890120e+00 -8.80537093e-01 1.08467422e-01 6.89679384e-01
2.59902894e-01 -3.82318765e-01 -1.37397885e+00 3.94689500e-01
2.41034821e-01 -1.70710254e+00 5.95915675e-01 -2.72830218e-01
8.34515452e-01 -4.54011977e-01 -6.91840649e-02 2.21142232e-01
4.17392552e-01 -9.95257497e-01 1.13749766e+00 5.09897232e-01
5.23145020e-01 -8.87376547e-01 7.22545028e-01 4.08380240e-01
-4.84118074e-01 1.69861674e-01 -2.17661873e-01 -2.06797883e-01
3.37505728e-01 6.54979408e-01 -1.07487094e+00 6.58776164e-01
-2.88530788e-03 8.09421182e-01 -5.60317814e-01 6.90573931e-01
-5.16813934e-01 5.68624496e-01 7.88380578e-02 -5.92158198e-01
3.70800436e-01 -9.05698612e-02 6.81667626e-01 1.39776528e+00
2.65063405e-01 -3.10401432e-02 -2.33625248e-01 1.17239130e+00
-4.58622783e-01 2.20870107e-01 -8.06874335e-01 -3.63838702e-01
5.64581692e-01 9.78658974e-01 -4.09093499e-01 -7.22900510e-01
9.30010974e-02 8.80899549e-01 3.96709651e-01 2.14166701e-01
-5.48441708e-01 -3.50909263e-01 1.68452859e-01 9.19764116e-02
8.20122510e-02 -2.47985218e-02 -6.30210638e-01 -1.14011276e+00
2.16009423e-01 -1.04154897e+00 1.99400693e-01 -7.99785078e-01
-1.20274663e+00 4.82316732e-01 1.49078548e-01 -8.64222169e-01
-6.84648335e-01 1.36507779e-01 -8.94019723e-01 6.24172211e-01
-1.49415338e+00 -7.62590587e-01 1.80081785e-01 -4.94032949e-02
8.31403971e-01 7.95132518e-02 5.77058673e-01 -3.38875145e-01
-3.89469713e-01 4.27521378e-01 -1.54324293e-01 -2.75891840e-01
6.61317110e-01 -1.42408252e+00 4.65462416e-01 1.01204622e+00
2.70395428e-01 6.15670145e-01 1.31111050e+00 -8.28023016e-01
-1.14217174e+00 -9.20786202e-01 1.35620499e+00 -5.38387001e-01
6.22436404e-01 3.69567484e-01 -7.67015815e-01 5.07421732e-01
8.27955902e-01 -8.25201750e-01 5.27698040e-01 2.38805369e-01
-3.84761542e-01 2.24451423e-01 -9.47848439e-01 1.14167094e+00
7.86389768e-01 -3.28296840e-01 -1.11388254e+00 2.96928167e-01
7.59255528e-01 -3.50013196e-01 -7.52844632e-01 1.20276220e-01
3.71313393e-01 -9.20647681e-01 6.35070503e-01 -6.96654677e-01
1.45892394e+00 5.33923171e-02 1.82229802e-02 -1.65828073e+00
-2.45678023e-01 -7.26298153e-01 -3.02201629e-01 1.27144289e+00
6.17950737e-01 -1.65291071e-01 6.14584565e-01 7.41306484e-01
-3.74500334e-01 -5.74222028e-01 -8.07363868e-01 -6.40922308e-01
3.45340848e-01 4.03914005e-02 5.42835176e-01 7.69692302e-01
5.85521400e-01 9.40909684e-01 -3.49476784e-01 -4.15300220e-01
5.96707821e-01 1.91045031e-01 5.85874856e-01 -9.10776556e-01
-1.23593941e-01 -8.72905195e-01 1.29885554e-01 -7.43582189e-01
2.44875357e-01 -8.23167741e-01 4.03463930e-01 -1.96092844e+00
1.85320184e-01 -4.62950254e-03 2.58801609e-01 2.49514222e-01
-5.82970798e-01 -1.20380305e-01 1.94680080e-01 1.88541755e-01
-8.13075662e-01 7.72447288e-01 1.12779582e+00 -3.00815076e-01
-3.32906574e-01 -1.56941891e-01 -1.25289857e+00 4.27522629e-01
9.03811157e-01 -3.68491322e-01 -4.52066630e-01 -4.19566602e-01
5.79017043e-01 2.79535025e-01 1.87633246e-01 -8.24220836e-01
2.28960693e-01 -2.33896315e-01 1.78921178e-01 -5.46050489e-01
-3.37230013e-04 -2.50748903e-01 -7.30444118e-02 3.83440018e-01
-1.18018937e+00 4.59892660e-01 5.21267056e-02 5.62532902e-01
-2.64748782e-01 -4.83374506e-01 5.46016395e-01 -1.54322520e-01
-1.00876935e-01 -2.09567532e-01 -5.73430896e-01 6.28855109e-01
5.88248849e-01 -2.02891946e-01 -5.11425674e-01 -7.51351357e-01
3.35745215e-02 1.31933317e-01 4.94384319e-01 2.75426298e-01
6.69013381e-01 -1.11949539e+00 -1.26069474e+00 -4.46622550e-01
1.60448506e-01 1.04380921e-02 -3.75430696e-02 5.26882529e-01
-3.23152363e-01 5.39105713e-01 -2.47351617e-01 -1.90145105e-01
-8.37406754e-01 2.32265353e-01 3.87187395e-03 -5.40202200e-01
-6.58559263e-01 3.43386889e-01 -3.23000252e-01 -4.15124632e-02
7.09720477e-02 -5.05354822e-01 -1.67024791e-01 2.91838884e-01
5.04640818e-01 6.32614911e-01 1.51459232e-01 -2.23675534e-01
6.64996505e-02 -1.78895786e-01 -3.63214791e-01 -4.97676522e-01
1.18793654e+00 -2.04207115e-02 1.28084511e-01 3.92766327e-01
7.37207711e-01 2.49361068e-01 -1.07691050e+00 -3.92556116e-02
2.78238565e-01 -2.74098516e-01 -9.07821953e-02 -8.28659534e-01
-4.37618911e-01 4.88371134e-01 -5.25692403e-01 4.98718262e-01
5.97918510e-01 -3.80649678e-02 8.01901758e-01 5.48301041e-01
-1.80733219e-01 -1.20312107e+00 2.88303345e-01 4.55197573e-01
1.23305333e+00 -1.07207704e+00 5.25530040e-01 1.21099420e-01
-9.81935084e-01 9.26594019e-01 4.90341336e-01 -1.50268957e-01
-1.71738610e-01 -5.36631867e-02 -2.19683781e-01 -9.68868658e-02
-1.10046387e+00 1.74370199e-01 1.70984849e-01 4.20341849e-01
8.00035238e-01 -3.82534787e-02 -6.89316094e-01 3.44275683e-01
-6.78423762e-01 -1.86525479e-01 1.06557870e+00 8.02149832e-01
-5.32846868e-01 -7.08274662e-01 -1.87614381e-01 6.20283782e-01
-4.68352348e-01 -2.85383940e-01 -7.80180693e-01 6.39272749e-01
-6.83834076e-01 1.09635448e+00 -2.43083900e-03 1.74523160e-01
3.23278666e-01 8.04268271e-02 6.26108050e-01 -6.12312257e-01
-8.49672973e-01 -2.23744258e-01 6.87644720e-01 -1.78001434e-01
-3.85988295e-01 -6.60594940e-01 -1.25267684e+00 -5.59915364e-01
-3.16721737e-01 3.84306788e-01 4.84586805e-01 9.29274619e-01
3.84142727e-01 6.71956003e-01 2.75102556e-01 -5.60245335e-01
-1.03230298e+00 -1.12040842e+00 -1.64682135e-01 5.60321629e-01
4.27771538e-01 5.32581564e-03 -2.55990207e-01 1.21280119e-01] | [12.270413398742676, 9.322680473327637] |
63622030-e526-4739-8859-d4af6560c04f | discourse-coherence-concurrent-explicit-and | null | null | https://aclanthology.org/P18-1210 | https://aclanthology.org/P18-1210.pdf | Discourse Coherence: Concurrent Explicit and Implicit Relations | Theories of discourse coherence posit relations between discourse segments as a key feature of coherent text. Our prior work suggests that multiple discourse relations can be simultaneously operative between two segments for reasons not predicted by the literature. Here we test how this joint presence can lead participants to endorse seemingly divergent conjunctions (e.g., BUT and SO) to express the link they see between two segments. These apparent divergences are not symptomatic of participant naivety or bias, but arise reliably from the concurrent availability of multiple relations between segments {--} some available through explicit signals and some via inference. We believe that these new results can both inform future progress in theoretical work on discourse coherence and lead to higher levels of performance in discourse parsing. | ['Nathan Schneider', 'er', 'Bonnie Webber', 'Alex Johnson', 'Hannah Rohde'] | 2018-07-01 | null | null | null | acl-2018-7 | ['implicit-relations'] | ['natural-language-processing'] | [ 5.45263767e-01 7.76195168e-01 -2.67960370e-01 -6.63169086e-01
-7.78114140e-01 -8.36769342e-01 1.27823317e+00 8.26626658e-01
-2.45193064e-01 7.71075070e-01 1.04011786e+00 -7.99589813e-01
3.35048884e-02 -6.39997840e-01 -4.30299312e-01 -1.94672570e-01
-9.98084396e-02 2.14025989e-01 3.98574859e-01 -4.84198183e-01
5.84348798e-01 -1.16659932e-01 -1.13174129e+00 6.72106862e-01
6.02537990e-01 -6.30322620e-02 1.04125710e-02 6.06643140e-01
-9.28275511e-02 1.22010899e+00 -7.60095000e-01 -2.13503659e-01
-2.67209530e-01 -8.45661581e-01 -1.09836352e+00 1.04221538e-01
3.16351533e-01 -4.30679232e-01 -1.46576554e-01 9.41760838e-01
8.80344287e-02 7.57680237e-02 7.07020462e-01 -7.63324559e-01
-5.57287216e-01 1.28098619e+00 -3.82267296e-01 4.58045870e-01
8.93759370e-01 3.67621750e-01 1.41634035e+00 -3.95524055e-01
8.58613789e-01 1.42709661e+00 4.44216251e-01 2.31576234e-01
-1.39356399e+00 -2.42660731e-01 4.47660685e-01 -1.99743137e-02
-9.58498180e-01 -1.01126933e+00 6.39775693e-01 -7.26769090e-01
1.11968887e+00 8.88430417e-01 9.20864999e-01 1.10813701e+00
1.37391314e-01 3.17909390e-01 1.51109076e+00 -4.99553561e-01
-6.17661439e-02 -6.94270199e-03 5.98909020e-01 1.86039731e-01
5.51273227e-01 2.51809340e-02 -7.04183698e-01 -2.84512490e-01
4.83603328e-01 -7.51100600e-01 -4.04638529e-01 5.94296277e-01
-1.42072320e+00 8.64693582e-01 3.18388849e-01 9.14125800e-01
-2.68666625e-01 1.13367587e-01 2.69574195e-01 4.12838489e-01
1.91921845e-01 7.70221651e-01 2.18191803e-01 -3.86614114e-01
-6.50497496e-01 8.26033711e-01 1.03340673e+00 6.81761265e-01
2.96504587e-01 -3.44918430e-01 -1.83654934e-01 1.87104359e-01
6.77567124e-01 9.24278200e-02 1.70680448e-01 -1.33418345e+00
5.96745431e-01 5.05938768e-01 4.51227784e-01 -1.31247318e+00
-3.92369032e-01 2.35501036e-01 -2.32395921e-02 -2.04144821e-01
7.59265304e-01 -2.50462383e-01 6.39520735e-02 2.05441928e+00
2.18581840e-01 -3.67451757e-01 -1.17682619e-02 8.11319292e-01
6.95328891e-01 5.56069076e-01 3.26143384e-01 -8.09140980e-01
1.50782251e+00 -1.72039032e-01 -1.08927727e+00 -5.58219254e-01
8.87871563e-01 -1.10552227e+00 8.92542958e-01 -6.75652400e-02
-1.32641160e+00 -8.44093710e-02 -1.16897714e+00 -3.97447020e-01
3.49157035e-01 -6.09993875e-01 1.00999665e+00 7.16508627e-01
-9.06483650e-01 3.01768363e-01 -7.07128227e-01 -2.01550230e-01
1.53980598e-01 -6.48886412e-02 -2.08143573e-02 4.00554955e-01
-1.21076393e+00 1.19666696e+00 5.13769448e-01 1.95825249e-01
-3.36775407e-02 -8.06642994e-02 -7.00429261e-01 -1.70858428e-01
3.97697449e-01 -5.48446536e-01 1.47730505e+00 -1.37396467e+00
-1.31488562e+00 1.17934239e+00 -5.33413291e-01 -3.71902525e-01
2.64924079e-01 -3.94972295e-01 -2.99798399e-01 9.63708162e-02
4.01268095e-01 3.63787144e-01 3.60368848e-01 -1.16504347e+00
-5.10127485e-01 -5.61066046e-02 3.41117561e-01 5.85244775e-01
3.24623942e-01 5.40190935e-01 5.86103499e-01 -3.51759136e-01
5.90244770e-01 -7.64421225e-01 2.82852650e-01 -3.20059478e-01
-7.83967078e-01 -6.76434338e-01 3.42086583e-01 -5.93041778e-01
1.52723765e+00 -1.98309040e+00 -1.25950292e-01 1.27871603e-01
4.86975700e-01 -1.62055805e-01 2.74289578e-01 8.14114630e-01
-2.76824296e-01 6.41833365e-01 -7.86609426e-02 2.74846435e-01
1.40607312e-01 2.02385918e-03 -4.72864479e-01 8.36844206e-01
3.17308575e-01 1.05512977e+00 -8.03470731e-01 -5.79659402e-01
-5.03857806e-02 -5.83254024e-02 -3.36460203e-01 1.80032998e-01
-5.97232819e-01 3.44562352e-01 -3.79179984e-01 1.58415288e-01
3.64132136e-01 -5.07691920e-01 9.60323513e-01 1.45958215e-01
-6.39249742e-01 1.54001045e+00 -1.04155517e+00 1.18474090e+00
5.64496636e-01 9.99325573e-01 1.62565172e-01 -9.83168125e-01
4.78218496e-01 5.82921267e-01 -2.94514209e-01 -5.12770057e-01
3.97168577e-01 1.47318467e-01 1.08947277e+00 -8.04734707e-01
7.91837215e-01 -5.80740452e-01 -2.55850405e-01 8.80447388e-01
-7.32275605e-01 -2.17946291e-01 2.69449770e-01 5.20730793e-01
7.90334880e-01 -2.72569060e-01 6.52322769e-01 -5.86624324e-01
2.20432520e-01 2.28741035e-01 6.55148149e-01 6.91746950e-01
-1.00003727e-01 4.13423739e-02 1.03734350e+00 1.50548398e-01
-1.15989780e+00 -8.91473889e-01 -3.35832000e-01 9.29768443e-01
2.96233416e-01 -5.69823146e-01 -4.34843093e-01 -1.10313140e-01
-4.25732404e-01 1.14596057e+00 -4.10737932e-01 4.50932205e-01
-1.28848791e+00 -4.84072566e-01 4.87961024e-01 3.30210328e-01
1.09688789e-01 -7.69966006e-01 -1.17182958e+00 2.79857516e-01
-6.04739428e-01 -8.70419860e-01 -2.57405102e-01 -1.48483798e-01
-7.47973979e-01 -1.22126532e+00 1.74806938e-01 -4.87208575e-01
3.12279046e-01 4.27524358e-01 1.02118671e+00 7.86262572e-01
3.90276283e-01 2.41608545e-01 -2.90356904e-01 -1.95874423e-01
-9.16830659e-01 -2.52925217e-01 -4.20132577e-01 -6.76150203e-01
4.07190114e-01 -5.80475152e-01 -2.43582428e-01 -1.16669141e-01
-6.84821725e-01 4.57618654e-01 5.12495376e-02 6.74234271e-01
-1.35285437e-01 -5.30833364e-01 7.13394880e-01 -1.35633266e+00
1.13201010e+00 -7.46938169e-01 -2.65493423e-01 2.74444848e-01
-3.61638278e-01 -1.47664487e-01 -1.69488817e-01 -2.21726477e-01
-1.29544985e+00 -8.98685575e-01 6.47228658e-02 8.14779997e-01
-3.45132738e-01 8.63130510e-01 1.16883907e-02 6.46438241e-01
8.46929729e-01 -4.31984305e-01 1.37255743e-01 3.74585569e-01
5.55952430e-01 4.10433322e-01 1.50410131e-01 -9.75557625e-01
5.37704051e-01 8.73939991e-02 -2.40011469e-01 -9.24239159e-01
-7.97257781e-01 -6.14898242e-02 -3.37066680e-01 -2.10120559e-01
1.04867911e+00 -1.16499126e+00 -8.02901745e-01 -8.54528621e-02
-1.52900350e+00 -5.24957895e-01 -8.62911120e-02 4.88352448e-01
-3.21220666e-01 4.70850259e-01 -8.26617002e-01 -1.03280210e+00
1.33788183e-01 -8.71573508e-01 3.34298223e-01 3.74755234e-01
-1.49388945e+00 -1.00571823e+00 -1.21028759e-01 2.46379092e-01
1.82556018e-01 4.51256633e-01 1.05386126e+00 -1.10188174e+00
-8.09808850e-01 2.34849527e-01 -3.71392109e-02 -4.04115796e-01
3.41359138e-01 4.64449614e-01 -7.81350613e-01 9.77491811e-02
4.81320918e-01 -5.08388400e-01 3.54981393e-01 2.60190159e-01
2.59255730e-02 -7.41759002e-01 -4.30588394e-01 -2.55748034e-01
1.14441228e+00 2.45556712e-01 4.84772712e-01 1.22064628e-01
1.90570518e-01 1.02400088e+00 5.50174057e-01 3.76221538e-02
6.81299746e-01 5.89418411e-01 -1.56846121e-01 5.54750741e-01
-7.62245506e-02 -1.08905666e-01 3.16440612e-01 9.57038641e-01
9.23139900e-02 -3.32857281e-01 -9.97886062e-01 6.01210833e-01
-1.88127005e+00 -1.44006634e+00 -1.02748144e+00 1.87353349e+00
1.20483196e+00 5.07775426e-01 9.28358883e-02 2.52936985e-02
7.35456169e-01 7.15850472e-01 -2.21341327e-02 -4.60846990e-01
-3.05467188e-01 8.48568603e-02 -2.01230332e-01 1.04636252e+00
-6.54487073e-01 7.72015631e-01 7.28676128e+00 1.35927886e-01
-9.26645279e-01 1.52897924e-01 4.73800391e-01 3.44134495e-02
-9.75240707e-01 4.00986075e-01 -4.88861859e-01 2.66862065e-01
9.50630844e-01 -5.51495790e-01 -3.62383239e-02 2.40406379e-01
1.52682692e-01 -8.49612772e-01 -1.54293621e+00 1.27146095e-01
-1.15939572e-01 -1.45414400e+00 -3.94392401e-01 4.35440950e-02
4.94545877e-01 -4.71343696e-01 -6.44107461e-02 -1.60420626e-01
5.50778031e-01 -1.11212444e+00 1.08852017e+00 5.89517355e-02
2.13512883e-01 -2.99364459e-02 3.61810327e-01 5.70440471e-01
-6.55314028e-01 3.15196276e-01 1.08826064e-01 -9.89274025e-01
6.14556432e-01 2.90221065e-01 -9.43075359e-01 3.05821121e-01
-1.71798974e-01 2.24683702e-01 -2.49747500e-01 4.34443682e-01
-7.73834944e-01 9.63581026e-01 -4.09908205e-01 -3.69105935e-01
1.11896567e-01 -1.88158844e-02 1.04341757e+00 9.52048659e-01
-2.41600499e-01 7.41315961e-01 -1.37870470e-02 1.12289846e+00
3.68408293e-01 -1.16791306e-02 -6.58909798e-01 -1.90353647e-01
1.11325073e+00 7.16202796e-01 -8.85850787e-01 -5.02711058e-01
-4.95463252e-01 3.23724210e-01 2.69923419e-01 2.14950234e-01
-5.94055057e-01 4.94334638e-01 5.33504665e-01 2.10232839e-01
-1.80623591e-01 -5.06389499e-01 -7.43160546e-01 -9.51092541e-01
6.83023930e-02 -9.29878294e-01 8.86720642e-02 -4.03427631e-01
-1.15920448e+00 8.22801217e-02 3.45772833e-01 -3.77625227e-01
-4.03468043e-01 -1.24992006e-01 -9.75884497e-01 1.15195596e+00
-9.36759710e-01 -7.61793077e-01 3.10571671e-01 -1.51967304e-02
5.26775420e-01 5.31294465e-01 1.05802071e+00 -3.67131352e-01
-3.21473390e-01 9.21396911e-02 -7.94019818e-01 6.78612590e-02
7.73260713e-01 -1.07272172e+00 4.62286502e-01 9.95212555e-01
7.40039423e-02 1.31469440e+00 1.23698628e+00 -8.74966383e-01
-8.69553566e-01 -1.63787112e-01 1.36535716e+00 -7.74305701e-01
7.19512045e-01 3.10984887e-02 -1.35644698e+00 1.13781071e+00
1.09478033e+00 -7.95088708e-01 1.04995835e+00 7.83953011e-01
-4.65392441e-01 7.99328864e-01 -7.28704393e-01 8.30461979e-01
9.70322192e-01 -6.16173565e-01 -1.62287605e+00 2.48670101e-01
9.57864761e-01 -6.12197697e-01 -5.92260957e-01 1.41046882e-01
3.33094060e-01 -1.29595852e+00 5.59977829e-01 -6.49340093e-01
7.25477695e-01 -1.67824656e-01 -3.27021807e-01 -9.09349203e-01
-2.41406471e-01 -6.58127964e-01 2.01418951e-01 1.30932248e+00
6.51455998e-01 -6.36393845e-01 2.11921513e-01 1.32952034e+00
-2.47971416e-01 -1.76322043e-01 -1.01907229e+00 -1.58174798e-01
4.07662362e-01 -3.77946168e-01 3.34034920e-01 1.32372630e+00
1.02551186e+00 8.66533458e-01 1.19422212e-01 -6.75536692e-02
2.92763114e-01 1.59504816e-01 6.03861630e-01 -1.12982106e+00
-4.17829692e-01 -7.93101609e-01 1.62433088e-01 -1.17489433e+00
8.57439861e-02 -7.24195600e-01 1.09158702e-01 -1.44240057e+00
2.26854861e-01 -5.61422229e-01 6.86746836e-01 2.88608056e-02
-5.31640708e-01 -4.52447593e-01 3.95821303e-01 4.26183373e-01
-4.70086813e-01 3.35289299e-01 1.11469328e+00 2.20143214e-01
-4.67323869e-01 -3.89827758e-01 -1.32438517e+00 1.05695081e+00
7.35668242e-01 -4.27588820e-01 -3.00748229e-01 -2.72045374e-01
7.02439725e-01 4.42738384e-01 3.81114870e-01 -3.24637353e-01
2.04231292e-01 -5.89403868e-01 1.74944848e-02 -2.72011101e-01
2.05410719e-01 -2.24391699e-01 3.68855268e-01 3.74520034e-01
-8.17296445e-01 9.39326286e-02 1.79147974e-01 2.45447531e-01
-7.80761838e-02 -3.18102509e-01 2.49136686e-01 -1.41800284e-01
-3.63487750e-01 -9.74224865e-01 -7.81206787e-01 2.84635335e-01
9.19247985e-01 -2.70033211e-01 -8.81841660e-01 -3.77886087e-01
-7.26078808e-01 4.92439084e-02 4.76925701e-01 2.21174076e-01
2.89011836e-01 -1.11413097e+00 -8.29066038e-01 -4.60041434e-01
-2.99056530e-01 3.50444429e-02 -4.41077799e-02 1.15998733e+00
-4.59340423e-01 4.17235494e-01 7.83104300e-02 -4.33541805e-01
-1.37948728e+00 1.19843565e-01 3.10126524e-02 9.38854460e-03
-7.07001269e-01 9.82023776e-01 1.62374943e-01 2.09830388e-01
-2.35259205e-01 -2.47536644e-01 -2.05581531e-01 3.27771276e-01
6.95327461e-01 7.69467354e-02 -7.00315893e-01 -8.26073289e-01
-4.39206183e-01 -8.72507617e-02 -2.05243006e-01 -6.22095823e-01
8.18816125e-01 -4.92196321e-01 -5.38997412e-01 1.10751224e+00
5.16275227e-01 5.24859369e-01 -8.26892734e-01 -1.37500480e-01
1.97791055e-01 -5.61129093e-01 -5.87199092e-01 -5.35414636e-01
2.86202759e-01 4.38595355e-01 -2.54323691e-01 8.56973648e-01
2.90433407e-01 4.04517442e-01 1.14460096e-01 1.87587366e-01
2.41221078e-02 -9.44588363e-01 5.92443198e-02 3.53715330e-01
1.20737314e+00 -9.45781052e-01 4.14482176e-01 -8.75867128e-01
-6.49012446e-01 1.08471000e+00 6.38629615e-01 -8.39774758e-02
2.16452748e-01 2.36034542e-01 -6.02337830e-02 -5.92981994e-01
-1.25045574e+00 -1.17416084e-01 -1.44223407e-01 4.35608417e-01
1.49400401e+00 4.61418301e-01 -1.30432475e+00 3.66636157e-01
-5.20909131e-01 -5.45837283e-01 8.68095279e-01 8.94881904e-01
-7.55949557e-01 -1.22889721e+00 -6.50927722e-01 4.42651451e-01
-4.97061282e-01 -1.04147904e-01 -7.16779768e-01 1.02437401e+00
-1.02308147e-01 1.31244636e+00 3.83741796e-01 1.39441952e-01
-4.42559384e-02 2.25530267e-01 6.99421883e-01 -8.92605305e-01
-9.14645672e-01 4.41558480e-01 1.14370847e+00 -2.02091187e-01
-9.59254861e-01 -1.40130007e+00 -1.41014516e+00 -7.05122709e-01
-6.61750853e-01 1.60216704e-01 -1.61016822e-01 1.34429121e+00
-8.16820264e-02 4.73655343e-01 -7.03175664e-02 -1.95966572e-01
-4.01951969e-01 -1.26810479e+00 -1.56948000e-01 3.00886512e-01
3.46384466e-01 -4.10287052e-01 -1.78230345e-01 2.05706045e-01] | [10.66592025756836, 9.319313049316406] |
f77b4007-3e8d-4cd3-a58d-8b8291045edb | idas-intent-discovery-with-abstractive | 2305.19783 | null | https://arxiv.org/abs/2305.19783v1 | https://arxiv.org/pdf/2305.19783v1.pdf | IDAS: Intent Discovery with Abstractive Summarization | Intent discovery is the task of inferring latent intents from a set of unlabeled utterances, and is a useful step towards the efficient creation of new conversational agents. We show that recent competitive methods in intent discovery can be outperformed by clustering utterances based on abstractive summaries, i.e., "labels", that retain the core elements while removing non-essential information. We contribute the IDAS approach, which collects a set of descriptive utterance labels by prompting a Large Language Model, starting from a well-chosen seed set of prototypical utterances, to bootstrap an In-Context Learning procedure to generate labels for non-prototypical utterances. The utterances and their resulting noisy labels are then encoded by a frozen pre-trained encoder, and subsequently clustered to recover the latent intents. For the unsupervised task (without any intent labels) IDAS outperforms the state-of-the-art by up to +7.42% in standard cluster metrics for the Banking, StackOverflow, and Transport datasets. For the semi-supervised task (with labels for a subset of intents) IDAS surpasses 2 recent methods on the CLINC benchmark without even using labeled data. | ['Chris Develder', 'Thomas Demeester', 'Fréderic Godin', 'Maarten De Raedt'] | 2023-05-31 | null | null | null | null | ['abstractive-text-summarization', 'intent-discovery'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.17752808e-01 5.07260859e-01 -1.57779798e-01 -8.91714156e-01
-1.11677861e+00 -6.19398773e-01 9.57423270e-01 1.84127629e-01
-3.06984514e-01 5.40864527e-01 9.13766444e-01 -1.48021400e-01
2.38036543e-01 -7.61694089e-02 -4.85256463e-01 -8.55660439e-01
-2.31929794e-01 1.14061916e+00 -2.40688026e-01 -1.05847716e-01
3.38609874e-01 -3.08705360e-01 -1.75778008e+00 7.16377854e-01
7.82917082e-01 6.89633906e-01 2.13873267e-01 6.86487973e-01
-3.44511032e-01 1.38012767e+00 -6.36372566e-01 -2.40802824e-01
-2.04204187e-01 -4.91643548e-01 -1.18174195e+00 5.32222748e-01
2.45252661e-02 -2.81325102e-01 1.81195915e-01 6.74406588e-01
2.05746457e-01 3.02239031e-01 8.63246918e-01 -1.16028333e+00
-3.34528714e-01 1.17767334e+00 -5.14741149e-03 -3.05835336e-01
6.15164101e-01 -1.06297627e-01 1.40428424e+00 -8.91939163e-01
6.77367449e-01 1.18631828e+00 4.85172242e-01 8.17736566e-01
-1.22183394e+00 -4.49168295e-01 3.19834113e-01 6.62500113e-02
-1.20219910e+00 -8.49754274e-01 7.57794201e-01 -6.47689462e-01
1.19547689e+00 4.90209609e-01 1.36550041e-02 1.29630041e+00
-3.45282674e-01 8.57212782e-01 9.70568001e-01 -4.83012438e-01
6.96267009e-01 4.93149698e-01 7.47962296e-01 6.81747615e-01
-2.85678923e-01 -2.16352180e-01 -4.98439491e-01 -4.04942751e-01
-3.33000660e-01 3.00352965e-02 -9.13118646e-02 5.14460243e-02
-1.24473333e+00 1.17616749e+00 -5.13753854e-02 3.14965993e-01
-5.65004587e-01 -1.66142792e-01 5.64161956e-01 2.88715720e-01
8.41708779e-01 5.85696638e-01 -4.00837570e-01 -2.75562465e-01
-8.35236967e-01 3.10045809e-01 1.37179387e+00 1.09999335e+00
1.12986684e+00 -3.75700772e-01 -3.63034129e-01 8.19113970e-01
4.26635742e-01 1.31990686e-01 7.20794678e-01 -1.20778465e+00
3.94164115e-01 7.94714332e-01 2.45181203e-01 -4.45390195e-01
-6.20968640e-01 -2.85966992e-01 -5.07249951e-01 -2.92451411e-01
5.60700148e-02 -4.68603671e-01 -8.74843597e-01 1.62190270e+00
2.75153488e-01 9.29629505e-02 4.55164939e-01 5.65548003e-01
9.51343060e-01 8.70387018e-01 7.52493218e-02 -5.67138016e-01
1.29437482e+00 -1.39736092e+00 -7.61575937e-01 -2.70361543e-01
1.03993058e+00 -5.75483084e-01 9.75646853e-01 2.59763628e-01
-5.75603604e-01 -2.64773011e-01 -6.38342142e-01 8.68285298e-02
-2.69154906e-01 3.97063121e-02 5.91650248e-01 6.64444566e-01
-1.27797484e+00 -4.19421047e-02 -6.60100222e-01 -3.56360018e-01
-6.52875453e-02 2.98038423e-01 -1.53890297e-01 -1.77887492e-02
-9.96687531e-01 6.04880035e-01 5.26877105e-01 -3.92894775e-01
-1.29302526e+00 -4.29416597e-01 -9.48267937e-01 4.42859866e-02
6.03341758e-01 -3.09125721e-01 1.57865226e+00 -8.30054343e-01
-1.64753771e+00 7.51670003e-01 -5.09276807e-01 -7.01476872e-01
8.49521980e-02 6.63805082e-02 -2.24925473e-01 4.40998860e-02
5.84884822e-01 8.76273155e-01 7.94754207e-01 -1.46031606e+00
-8.86496067e-01 -1.44317165e-01 1.21788327e-02 4.73672539e-01
-3.15872669e-01 1.84709966e-01 -4.88276452e-01 -4.25025038e-02
-2.58560292e-02 -1.27179921e+00 -2.59277672e-01 -1.20502985e+00
-7.55460143e-01 -8.92907858e-01 8.02905083e-01 -5.76177657e-01
1.05238366e+00 -2.04941297e+00 1.41649231e-01 -7.67926723e-02
3.90007347e-01 -1.49129480e-01 -6.76725581e-02 6.42210901e-01
-5.86889721e-02 6.99533820e-02 -2.87902325e-01 -1.15783978e+00
4.22563046e-01 1.92347512e-01 -4.40232158e-01 2.77927160e-01
-7.24190027e-02 5.98191440e-01 -1.02945971e+00 -4.90476996e-01
3.19798976e-01 1.57785878e-01 -7.84999490e-01 6.73785567e-01
-7.77218282e-01 6.48502529e-01 -3.29233587e-01 2.75493562e-01
6.58527017e-02 -5.84353328e-01 3.43938977e-01 2.34556422e-01
-1.48831218e-01 9.72217083e-01 -7.60912895e-01 1.76703763e+00
-6.36018276e-01 5.38519502e-01 -6.36775745e-03 -1.06929648e+00
9.15387273e-01 7.44365275e-01 5.59983730e-01 -2.44000349e-02
2.84354482e-02 1.79160148e-01 -1.50427476e-01 -6.97052896e-01
5.15977085e-01 -2.65775383e-01 -7.15401292e-01 1.02902162e+00
3.08685869e-01 1.43745050e-01 3.48676890e-01 5.37020326e-01
1.24098909e+00 -4.43038821e-01 2.09188566e-01 -1.57130599e-01
5.83511949e-01 1.76836416e-01 3.32148075e-01 7.54254282e-01
-1.54153660e-01 4.58214343e-01 5.76876581e-01 -4.70066369e-01
-7.81870842e-01 -4.73426878e-01 6.91584498e-02 1.71154690e+00
-2.28148192e-01 -6.47479951e-01 -9.24736619e-01 -1.03410792e+00
-3.02399248e-01 1.27272737e+00 -5.64363897e-01 7.81848505e-02
-3.86096984e-01 -6.95302367e-01 3.96731496e-01 8.43562651e-03
2.83206910e-01 -1.30707872e+00 -2.93416351e-01 3.74028414e-01
-8.38328302e-01 -1.35169041e+00 -4.78792429e-01 4.87713248e-01
-1.86172396e-01 -6.04436874e-01 -2.91878670e-01 -9.11535621e-01
6.76860750e-01 7.03284983e-03 1.12485349e+00 -4.29927707e-02
3.31277668e-01 3.06338400e-01 -7.58240163e-01 -2.55748332e-01
-9.31157053e-01 3.46116781e-01 1.93838254e-01 3.33787411e-01
7.52989054e-01 -3.44802320e-01 -8.05908963e-02 1.43267915e-01
-5.45318723e-01 3.06515276e-01 2.89129168e-01 8.25484395e-01
4.40178104e-02 1.07417703e-01 9.26977158e-01 -1.18383598e+00
7.05573440e-01 -7.46163130e-01 -2.05465391e-01 6.03128672e-02
-5.23261368e-01 2.66770393e-01 7.85511672e-01 -1.51979685e-01
-1.32871437e+00 1.60109147e-01 -2.04751834e-01 -8.82094428e-02
-7.33283579e-01 4.33038533e-01 6.39600232e-02 6.24703944e-01
5.41075349e-01 2.33814850e-01 -1.03041880e-01 -5.54712832e-01
5.64561367e-01 1.39084077e+00 3.89147818e-01 -4.45324361e-01
1.91256985e-01 3.96297812e-01 -8.26223969e-01 -8.66521955e-01
-1.21486366e+00 -1.02318513e+00 -4.99874651e-01 -1.68261364e-01
1.06037652e+00 -1.06473911e+00 -6.45376325e-01 1.12184949e-01
-1.37800467e+00 -4.45571721e-01 -2.92624950e-01 4.62898552e-01
-7.22564340e-01 2.03164607e-01 -8.70982945e-01 -1.13353026e+00
-3.92934263e-01 -1.35383224e+00 1.22291303e+00 -5.86325936e-02
-7.60351598e-01 -1.00765562e+00 2.45011196e-01 9.29761529e-01
9.38270986e-02 -1.57494321e-01 8.79572093e-01 -1.63149643e+00
-4.16852891e-01 -2.05430746e-01 2.54737496e-01 3.09282094e-01
2.44309768e-01 -3.95536691e-01 -1.21440506e+00 -1.94645762e-01
4.19030875e-01 -6.08055592e-01 8.35869908e-01 1.04728431e-01
5.62978864e-01 -6.98444545e-01 -5.63993216e-01 1.76091522e-01
8.28553915e-01 4.75882828e-01 8.04871991e-02 -3.77738141e-02
5.55974305e-01 1.09226823e+00 2.44392514e-01 5.71300566e-01
8.25406253e-01 5.07661402e-01 1.76450327e-01 2.20480800e-01
1.64824069e-01 -1.70831919e-01 5.31048179e-01 1.23077524e+00
3.42063427e-01 -2.47681245e-01 -9.32099044e-01 6.62225246e-01
-1.92298794e+00 -8.49096537e-01 8.91671777e-02 1.86893916e+00
1.06366289e+00 1.82461925e-02 1.97965026e-01 -1.53797060e-01
8.32610071e-01 1.89581126e-01 -3.37164462e-01 -2.49342933e-01
2.86858350e-01 -3.27353060e-01 3.46988589e-02 7.17121542e-01
-1.24207222e+00 9.59081829e-01 5.99576473e+00 4.04193640e-01
-7.29668856e-01 4.54828978e-01 7.69900262e-01 3.73924114e-02
-3.66614312e-01 1.39216155e-01 -1.03956497e+00 6.67612910e-01
1.54923630e+00 3.95454951e-02 4.57077533e-01 1.00903475e+00
1.88569993e-01 3.26570459e-02 -1.44433343e+00 6.06701851e-01
5.47728062e-01 -1.39467335e+00 -3.28603655e-01 -3.28198709e-02
1.08516264e+00 2.89487898e-01 -1.77001104e-01 8.86281133e-01
8.30122292e-01 -8.54687333e-01 5.03904998e-01 2.56600469e-01
4.82464254e-01 -5.10676980e-01 9.07833219e-01 8.48843217e-01
-6.56129122e-01 -2.29623884e-01 -9.48916674e-02 -8.82995874e-02
2.27323994e-01 5.18884540e-01 -1.43053222e+00 2.16323972e-01
4.51597780e-01 7.37304688e-01 -2.62887597e-01 4.81432498e-01
-7.05043077e-02 1.02562201e+00 -2.48558193e-01 -2.83872753e-01
5.97075939e-01 -1.57748580e-01 4.82810050e-01 1.52314568e+00
-1.55198574e-01 6.45013452e-02 5.10598719e-01 9.23849046e-01
-2.24371850e-01 8.06172192e-02 -5.36705792e-01 -6.96388679e-03
5.05417526e-01 1.28702569e+00 -7.33012199e-01 -8.99407983e-01
-3.51354182e-01 7.25412071e-01 3.30267191e-01 4.19633925e-01
-5.95006287e-01 -9.92159471e-02 3.63400847e-01 -3.03455561e-01
1.99339941e-01 1.93118721e-01 -1.66559756e-01 -1.10131669e+00
-1.81393728e-01 -9.91171896e-01 3.64801615e-01 -3.60463142e-01
-1.16963923e+00 8.05630028e-01 4.26943898e-02 -1.05446744e+00
-9.76602316e-01 -2.69987248e-02 -7.84086108e-01 6.07242286e-01
-1.13152492e+00 -9.75573480e-01 -9.28345844e-02 2.43418187e-01
9.98092711e-01 -3.90422195e-01 1.11658084e+00 1.55765861e-01
-6.65190637e-01 1.83383018e-01 1.69853613e-01 7.53085986e-02
6.26672566e-01 -1.34754145e+00 3.22784245e-01 5.94150543e-01
1.61506608e-01 6.49642944e-01 9.30383325e-01 -6.50913775e-01
-9.99703526e-01 -1.20444012e+00 1.42038989e+00 -7.33630240e-01
4.35274363e-01 -8.64559293e-01 -7.89908230e-01 1.12990034e+00
4.53790307e-01 -5.35836518e-01 9.68875706e-01 4.16653097e-01
-2.77761012e-01 2.83848912e-01 -8.37554812e-01 5.12805045e-01
6.88687921e-01 -5.70943534e-01 -9.77636814e-01 9.07887995e-01
1.36779964e+00 -7.12165982e-03 -5.12530982e-01 -2.52283290e-02
9.40117463e-02 -9.94131148e-01 6.48827791e-01 -5.61706781e-01
1.77199334e-01 -3.21481824e-02 -2.26034641e-01 -1.14051962e+00
-1.66292891e-01 -1.06600857e+00 1.20852016e-01 1.45522022e+00
6.18232310e-01 -4.36760753e-01 8.93146396e-01 6.25932038e-01
-4.79074389e-01 -3.68667305e-01 -8.13593626e-01 -3.56048197e-01
-2.09528565e-01 -6.38786972e-01 5.95363259e-01 8.94910276e-01
6.87863290e-01 1.11304653e+00 -5.52199304e-01 -2.05836389e-02
7.10186362e-01 3.18537474e-01 7.53150284e-01 -1.17653847e+00
-3.68216671e-02 -2.26703152e-01 4.13056463e-02 -1.24146390e+00
7.96322465e-01 -1.04495203e+00 6.27261996e-01 -1.45897257e+00
2.40076855e-01 -5.76103628e-01 1.24710955e-01 5.67348897e-01
-2.29121551e-01 -1.22801609e-01 -5.52813895e-02 5.77255070e-01
-1.37187052e+00 5.40876031e-01 5.22239387e-01 -1.55034661e-01
-5.23826063e-01 2.87535369e-01 -8.55918944e-01 7.94022083e-01
6.88290060e-01 -5.91525137e-01 -4.54317182e-01 1.34442404e-01
-2.48353347e-01 1.59037068e-01 -1.33930981e-01 -8.02150607e-01
3.64946872e-01 8.86497498e-02 -4.30992752e-01 -8.40715349e-01
4.68087643e-01 -5.83558917e-01 -7.06741810e-02 2.24032700e-01
-1.04542363e+00 -3.42864960e-01 -4.18715596e-01 5.05417049e-01
-1.85683668e-01 -5.18366277e-01 2.45267272e-01 -1.92515239e-01
-7.04296827e-01 -2.20292628e-01 -8.73309135e-01 2.77655572e-01
8.44498515e-01 3.06557655e-01 -2.68267214e-01 -8.52177978e-01
-8.27469468e-01 3.50605488e-01 1.80424988e-01 3.26789379e-01
2.67306864e-01 -1.02528775e+00 -9.78795588e-01 2.68560201e-01
2.75801033e-01 1.70277804e-01 2.13087156e-01 6.96030378e-01
2.69676819e-02 1.04494858e+00 4.62289125e-01 -7.88014293e-01
-1.01187253e+00 5.46558321e-01 1.85881536e-02 -5.04211664e-01
-5.92403293e-01 9.10913348e-01 4.83296365e-01 -8.99553180e-01
6.01597190e-01 -4.07892138e-01 -3.44466090e-01 2.08396599e-01
6.43584967e-01 1.75736681e-01 1.21645115e-01 -8.01741838e-01
-3.70877177e-01 -8.19984823e-02 -4.05989587e-01 -3.37894619e-01
1.25442541e+00 -3.29779029e-01 -1.88194349e-01 7.37831354e-01
1.60274780e+00 -2.67101169e-01 -1.16679692e+00 -5.22434592e-01
3.86722624e-01 2.14230120e-01 -2.45706990e-01 -7.75866449e-01
-4.34650332e-01 5.26845932e-01 7.36182779e-02 6.36496067e-01
7.42505729e-01 6.39019549e-01 6.23040438e-01 5.34414947e-01
3.34152699e-01 -9.61078107e-01 3.21954429e-01 7.56779552e-01
7.47245431e-01 -1.45253468e+00 -4.00582552e-01 -2.45350689e-01
-1.17033541e+00 6.44495547e-01 2.66700596e-01 1.85539410e-01
4.74977136e-01 3.27354521e-01 3.18251431e-01 -3.94023687e-01
-1.37566817e+00 -2.41598770e-01 7.18253553e-02 4.79180872e-01
5.04986763e-01 4.50541615e-01 -1.54681513e-02 7.25260377e-01
-4.23029244e-01 -4.69145656e-01 6.40548825e-01 6.01390898e-01
-6.12092316e-01 -8.63230407e-01 -2.35835984e-01 6.93434954e-01
-2.50534087e-01 -2.70327240e-01 -3.50127101e-01 2.10887894e-01
3.19807343e-02 1.52366781e+00 2.71703452e-01 -6.59872651e-01
-7.23772645e-02 6.53377056e-01 -4.49137002e-01 -1.21977329e+00
-7.18731284e-01 3.06727380e-01 5.74709296e-01 -1.98663414e-01
-6.70819700e-01 -9.59874213e-01 -1.28896999e+00 4.80158068e-02
-4.37685013e-01 8.21190298e-01 6.43507659e-01 1.31795180e+00
3.12345952e-01 3.78418922e-01 8.82176340e-01 -8.51152360e-01
-7.05024540e-01 -1.36380982e+00 -2.71741807e-01 7.14243472e-01
4.57390934e-01 -3.78985137e-01 -9.03605223e-01 4.11222845e-01] | [12.491523742675781, 7.546449184417725] |
51b48a8a-f721-4ecf-85dd-a5e98eb6368d | self-supervised-driven-consistency-training | 2102.03897 | null | https://arxiv.org/abs/2102.03897v3 | https://arxiv.org/pdf/2102.03897v3.pdf | Self-supervised driven consistency training for annotation efficient histopathology image analysis | Training a neural network with a large labeled dataset is still a dominant paradigm in computational histopathology. However, obtaining such exhaustive manual annotations is often expensive, laborious, and prone to inter and Intra-observer variability. While recent self-supervised and semi-supervised methods can alleviate this need by learn-ing unsupervised feature representations, they still struggle to generalize well to downstream tasks when the number of labeled instances is small. In this work, we overcome this challenge by leveraging both task-agnostic and task-specific unlabeled data based on two novel strategies: i) a self-supervised pretext task that harnesses the underlying multi-resolution contextual cues in histology whole-slide images to learn a powerful supervisory signal for unsupervised representation learning; ii) a new teacher-student semi-supervised consistency paradigm that learns to effectively transfer the pretrained representations to downstream tasks based on prediction consistency with the task-specific un-labeled data. We carry out extensive validation experiments on three histopathology benchmark datasets across two classification and one regression-based tasks, i.e., tumor metastasis detection, tissue type classification, and tumor cellularity quantification. Under limited-label data, the proposed method yields tangible improvements, which is close or even outperforming other state-of-the-art self-supervised and supervised baselines. Furthermore, we empirically show that the idea of bootstrapping the self-supervised pretrained features is an effective way to improve the task-specific semi-supervised learning on standard benchmarks. Code and pretrained models will be made available at: https://github.com/srinidhiPY/SSL_CR_Histo | ['Anne L. Martel', 'Fu-Der Chen', 'Seung Wook Kim', 'Chetan L. Srinidhi'] | 2021-02-07 | null | null | null | null | ['histopathological-image-classification'] | ['medical'] | [ 5.14969409e-01 1.37628660e-01 -4.96880203e-01 -6.79605007e-01
-1.31062198e+00 -5.02770066e-01 4.91739899e-01 5.19107103e-01
-4.10511017e-01 8.70285630e-01 1.02201261e-01 -2.65245497e-01
-9.05228853e-02 -5.11605740e-01 -6.30778551e-01 -1.20264184e+00
6.78319409e-02 5.19524038e-01 1.28204182e-01 6.69731200e-02
-4.59058881e-02 3.85766119e-01 -1.18786800e+00 4.12471861e-01
7.58789361e-01 1.00854695e+00 5.97850457e-02 5.82418621e-01
1.10166660e-02 8.53213549e-01 -1.40363768e-01 3.90612483e-02
-2.53181420e-02 -3.05146039e-01 -1.05306780e+00 1.94952980e-01
2.82389134e-01 1.13602899e-01 -6.32613748e-02 9.19977665e-01
4.12972033e-01 -2.84847349e-01 8.27866912e-01 -1.03177786e+00
-4.97481376e-01 4.53394532e-01 -6.68297231e-01 2.95883000e-01
-2.51309127e-01 1.70468166e-01 9.45888400e-01 -7.01186895e-01
7.19426215e-01 5.49797237e-01 8.58477414e-01 6.45203948e-01
-1.53724158e+00 -8.27955246e-01 -1.56164795e-01 -2.00305879e-01
-1.34222138e+00 -4.59152281e-01 6.71641886e-01 -5.65257847e-01
5.60323417e-01 1.75112635e-01 3.09915394e-01 1.24106920e+00
2.25138992e-01 6.55269384e-01 1.57751060e+00 -3.28737468e-01
1.55147478e-01 3.79008889e-01 4.03013021e-01 8.51040125e-01
1.54277265e-01 2.17543960e-01 -3.02216500e-01 -2.00283110e-01
6.35335207e-01 4.08536345e-01 -1.94572479e-01 -4.34044272e-01
-1.29425049e+00 7.86936283e-01 6.05935693e-01 4.95769590e-01
-3.85114588e-02 7.05158804e-03 5.47030687e-01 3.93023968e-01
7.63045669e-01 6.00551009e-01 -7.10053980e-01 2.46172100e-01
-1.01405132e+00 -3.43542069e-01 6.14095330e-01 6.74878657e-01
9.84430552e-01 -3.28483135e-01 -7.40062073e-02 8.53706777e-01
1.73870951e-01 1.10964574e-01 9.17154968e-01 -4.66818333e-01
-5.90418465e-02 8.18589926e-01 -3.87949944e-01 -5.27756155e-01
-6.48255944e-01 -8.78352523e-01 -1.14665329e+00 1.05052754e-01
6.18032157e-01 -9.53393430e-02 -1.14878392e+00 1.72829270e+00
3.39530319e-01 2.29862824e-01 4.40203138e-02 6.54789925e-01
9.88336623e-01 2.13321745e-01 3.07449579e-01 -3.46138895e-01
1.25129032e+00 -1.14876091e+00 -5.04363179e-01 -1.13801882e-01
1.13951409e+00 -5.40946186e-01 9.11801457e-01 1.36446208e-01
-6.93007946e-01 -2.93159455e-01 -9.15718675e-01 2.52589844e-02
-4.92616385e-01 3.45304668e-01 8.03613067e-01 3.25386584e-01
-1.10954320e+00 4.95147347e-01 -9.94611382e-01 -5.22156417e-01
9.04016078e-01 5.56646049e-01 -7.90141046e-01 -1.30018666e-01
-7.07260251e-01 6.78699911e-01 3.48222762e-01 -3.95923294e-02
-1.27136374e+00 -1.03896630e+00 -6.72076941e-01 -2.74865497e-02
2.48028591e-01 -5.10120630e-01 1.16960168e+00 -1.26330149e+00
-1.32851601e+00 1.38422108e+00 -2.85848957e-02 -2.64700890e-01
4.23633188e-01 2.18896627e-01 -9.55558717e-02 1.63561925e-01
2.15621307e-01 6.30512476e-01 6.77575648e-01 -1.30555212e+00
-4.41461891e-01 -4.44163829e-01 -4.02077883e-01 -3.28479782e-02
-6.10271454e-01 -3.54252905e-01 -1.95314512e-01 -5.88700831e-01
1.10943519e-01 -1.03793681e+00 -4.14671570e-01 1.61105290e-01
-6.30636632e-01 -2.07500696e-01 6.15443170e-01 -2.06931442e-01
6.96216106e-01 -2.23742938e+00 -1.73997171e-02 2.42270619e-01
4.42090780e-01 1.97575733e-01 -2.13182881e-01 2.36633986e-01
-3.61699611e-01 2.93861758e-02 -3.67019504e-01 -4.43996757e-01
-4.65788841e-01 1.16895832e-01 -4.61493768e-02 7.50480235e-01
4.31616783e-01 1.00365829e+00 -1.12457013e+00 -6.89199388e-01
-1.09017245e-01 3.52253258e-01 -3.05548698e-01 4.11178768e-01
-2.58732531e-02 9.13222492e-01 -4.00800049e-01 8.84917200e-01
1.90878928e-01 -8.33705783e-01 4.17880207e-01 -1.22853115e-01
2.75892049e-01 7.71844983e-02 -5.53445399e-01 1.84127307e+00
-5.12754381e-01 4.23523128e-01 -7.81080797e-02 -1.43662167e+00
8.22292209e-01 3.39362472e-01 5.91017723e-01 -2.97301292e-01
2.20640942e-01 3.89302939e-01 -6.25141039e-02 -5.02574086e-01
-1.78116620e-01 -3.87445539e-01 7.77842551e-02 5.26732862e-01
5.16265929e-01 4.21140566e-02 -1.10362209e-02 1.18204229e-01
1.36250114e+00 -1.07806064e-02 5.22019207e-01 -5.11788249e-01
5.03776371e-01 1.30746484e-01 5.95143914e-01 4.72477853e-01
-3.65240663e-01 6.44471765e-01 4.93891865e-01 -4.60736066e-01
-6.59510911e-01 -9.20030355e-01 -4.18330371e-01 1.31378067e+00
-1.19398080e-01 -1.95818931e-01 -4.27581787e-01 -1.13337672e+00
7.00087696e-02 5.42055145e-02 -1.22303104e+00 -2.02839762e-01
-1.44626081e-01 -1.08668399e+00 5.42130232e-01 6.86006188e-01
3.83120514e-02 -7.40405500e-01 -9.08298790e-02 6.38128631e-03
1.01096079e-01 -1.01447451e+00 -1.96390793e-01 8.11890304e-01
-1.15852809e+00 -1.25032246e+00 -7.13005543e-01 -1.03367734e+00
1.32370281e+00 3.20180178e-01 9.90377367e-01 3.91626537e-01
-5.20564020e-01 1.81424409e-01 -2.79880524e-01 -3.65255415e-01
-4.60909516e-01 3.98565769e-01 -1.02810189e-01 1.08421303e-01
3.67939681e-01 -6.08759463e-01 -6.64254725e-01 4.01489794e-01
-9.11006868e-01 1.73597828e-01 9.52642739e-01 1.39274347e+00
1.01709533e+00 -1.78471088e-01 7.45595038e-01 -1.44814491e+00
1.04696266e-01 -5.68333924e-01 -1.38247967e-01 2.17536554e-01
-6.93870664e-01 7.36594126e-02 6.51749492e-01 -4.54008728e-01
-8.56725574e-01 4.10051376e-01 -6.88687805e-03 -2.45966449e-01
-3.18729132e-01 6.08939707e-01 1.93782583e-01 -2.46520609e-01
8.88591766e-01 1.97922945e-01 4.08522308e-01 -1.58058986e-01
9.08141360e-02 6.45287097e-01 4.07550335e-01 -5.20640969e-01
8.89165640e-01 7.50501037e-01 6.14363477e-02 -4.46129501e-01
-1.32242036e+00 -8.42249990e-01 -7.77310967e-01 6.87938854e-02
7.53814340e-01 -1.01377511e+00 -2.92667896e-01 4.06528652e-01
-4.78498578e-01 -7.95567393e-01 -4.34687465e-01 4.09568638e-01
-5.32240570e-01 9.88278091e-02 -7.38260925e-01 -2.26427689e-01
-3.41360867e-01 -1.18656135e+00 1.18090725e+00 2.83608526e-01
-2.23679185e-01 -1.24189353e+00 3.40392113e-01 5.19171894e-01
4.56828445e-01 3.14814061e-01 8.82019043e-01 -1.16293955e+00
-3.04026365e-01 -4.20867175e-01 -3.74689996e-01 2.73518711e-01
4.77576673e-01 -7.97629058e-02 -1.22304082e+00 -5.41120470e-01
-2.79249758e-01 -1.01395595e+00 1.08018970e+00 1.37868851e-01
1.43672526e+00 -7.55607560e-02 -6.71130896e-01 8.21547031e-01
1.42483628e+00 -5.64914465e-01 1.92606166e-01 2.92300701e-01
5.59313059e-01 6.48065865e-01 5.28236449e-01 5.92216924e-02
3.74250293e-01 3.39651555e-01 2.73762465e-01 -4.97353554e-01
-2.36368820e-01 -9.30486992e-02 -8.20790976e-02 7.20492601e-01
-7.29129910e-02 1.14578232e-01 -1.08964145e+00 5.88315964e-01
-1.88033402e+00 -5.14317513e-01 6.13850541e-02 1.93193865e+00
1.36873341e+00 1.16135962e-02 -3.28604653e-02 1.17886543e-01
4.31426913e-01 -7.73101524e-02 -5.48453927e-01 1.37433529e-01
5.03900424e-02 1.40901253e-01 4.82165188e-01 1.46020845e-01
-1.31810701e+00 8.81340623e-01 5.33650494e+00 7.71101296e-01
-1.45067751e+00 4.09337312e-01 1.21613753e+00 7.80517906e-02
-3.70818935e-03 -1.72730997e-01 -6.18143797e-01 1.74959108e-01
9.66510773e-01 1.03829853e-01 -1.32858738e-01 8.61981571e-01
1.41210901e-02 -1.59217324e-02 -1.35677207e+00 7.41874337e-01
1.02517284e-01 -1.50959754e+00 -1.56922430e-01 9.14792269e-02
9.35596526e-01 1.47406474e-01 6.00954182e-02 3.52900833e-01
4.48561192e-01 -1.31092131e+00 -3.13548632e-02 3.24260801e-01
1.15734172e+00 -3.05802017e-01 1.10628581e+00 3.44217390e-01
-7.74229467e-01 -4.18872871e-02 -1.78036213e-01 2.04182535e-01
-4.20112222e-01 6.70375764e-01 -1.00000024e+00 3.79923850e-01
5.58058977e-01 9.33107674e-01 -8.14692676e-01 8.96112859e-01
-2.39119843e-01 7.98259199e-01 -1.65422559e-02 1.59424841e-01
2.63724357e-01 3.42706203e-01 1.19427117e-02 1.37175119e+00
-1.09174605e-02 -6.01101220e-02 3.01962614e-01 4.98033464e-01
-2.53480554e-01 1.39524356e-01 -3.14311445e-01 -1.35325208e-01
2.00675443e-01 1.80466592e+00 -9.66776907e-01 -2.39364803e-01
-3.06863308e-01 6.83578789e-01 7.54355192e-01 2.76660025e-01
-5.60725391e-01 1.96242798e-02 2.25365579e-01 1.60982773e-01
4.14127624e-03 1.30475149e-01 -5.12080610e-01 -1.12817919e+00
-4.30929929e-01 -7.92514205e-01 6.56017065e-01 -3.74201953e-01
-1.71475995e+00 5.54253399e-01 -3.24160248e-01 -1.34393215e+00
-7.67951831e-02 -6.50565088e-01 -7.03649938e-01 4.20453399e-01
-1.94093359e+00 -1.53615248e+00 -6.17752373e-01 6.19993269e-01
4.20740038e-01 -2.60966301e-01 1.20891643e+00 9.52617228e-02
-6.38881385e-01 8.94025385e-01 2.28739187e-01 3.89198005e-01
1.16492033e+00 -1.32128525e+00 -4.70349342e-01 4.27419573e-01
4.91804518e-02 5.61432421e-01 3.61134171e-01 -2.75468767e-01
-1.31094825e+00 -1.31417143e+00 6.07341945e-01 -5.40296197e-01
1.00548613e+00 -3.74915957e-01 -9.95974958e-01 8.25804353e-01
-1.12670556e-01 8.55080009e-01 1.45984423e+00 2.95290411e-01
-5.25427163e-01 -3.16791445e-01 -1.12581193e+00 2.40612313e-01
8.59117866e-01 -5.84523499e-01 -2.91555464e-01 8.49992096e-01
4.06518579e-01 -4.42072511e-01 -1.09415698e+00 5.14894009e-01
4.80223060e-01 -6.47407532e-01 7.08901346e-01 -7.99920499e-01
6.41874433e-01 -1.80049300e-01 3.31884623e-02 -1.24548519e+00
-2.64065474e-01 -4.25429463e-01 2.09449977e-01 1.19689798e+00
6.77619457e-01 -6.44552529e-01 9.80974436e-01 2.10065782e-01
-3.00380200e-01 -1.18720889e+00 -8.67434204e-01 -5.80317318e-01
3.38206112e-01 -8.15629121e-03 1.01167239e-01 1.33375227e+00
1.82440475e-01 3.14939976e-01 -1.77162904e-02 1.49969205e-01
6.13232791e-01 1.69289276e-01 7.59681463e-01 -1.20707488e+00
-3.25522423e-01 -3.23677450e-01 -5.83908141e-01 -4.63340878e-01
3.73204976e-01 -1.35170364e+00 2.57689774e-01 -1.11969745e+00
6.61482513e-01 -8.32634449e-01 -7.88960218e-01 8.86718750e-01
-2.53262907e-01 6.11767828e-01 -4.03872937e-01 5.39882600e-01
-8.78671467e-01 2.06933722e-01 1.25741446e+00 -3.39545786e-01
-2.10798513e-02 -1.19922906e-02 -8.92561615e-01 6.74809992e-01
8.50309730e-01 -7.52444625e-01 -3.89384001e-01 -1.99323997e-01
-1.70052648e-01 -8.63842592e-02 4.56073195e-01 -9.25756872e-01
3.86322290e-01 -1.28891826e-01 5.99393129e-01 -3.98248173e-02
-6.17490672e-02 -6.69097185e-01 -2.00486600e-01 5.30956805e-01
-7.86424398e-01 -4.39077258e-01 7.22299889e-02 7.84980118e-01
-2.20404595e-01 -4.41670790e-02 9.92790520e-01 -1.60998330e-01
-3.37661475e-01 4.86473620e-01 -9.71106216e-02 -3.19417305e-02
1.06965721e+00 -5.64806648e-02 -6.42472386e-01 -1.18693627e-01
-8.31386268e-01 2.61959910e-01 4.53478277e-01 1.12050191e-01
4.02265579e-01 -1.03233957e+00 -7.36143351e-01 1.61443040e-01
4.86812651e-01 1.36534497e-01 3.21424186e-01 1.29312527e+00
-2.85074085e-01 3.49459916e-01 -1.94551721e-01 -1.06999409e+00
-1.12718689e+00 4.68743920e-01 3.32199425e-01 -8.11066508e-01
-2.66927600e-01 1.08606076e+00 4.42323744e-01 -7.05321193e-01
1.47852853e-01 -9.20782834e-02 -1.56782150e-01 -1.11704074e-01
3.94650847e-01 -1.77983344e-01 1.48788065e-01 -4.90781397e-01
-3.48470747e-01 3.94034386e-01 -5.41590631e-01 5.08479059e-01
1.53779817e+00 1.93583533e-01 -1.04308851e-01 4.91400450e-01
1.41903448e+00 -2.24299341e-01 -1.27186906e+00 -4.82161224e-01
2.01075047e-01 3.50202015e-03 1.05856702e-01 -9.04168367e-01
-1.20161641e+00 8.22404027e-01 6.83538735e-01 -1.21707931e-01
1.08785248e+00 2.79268175e-01 4.56812531e-01 4.53818291e-01
2.28928939e-01 -6.93478525e-01 3.51741582e-01 2.02053979e-01
5.93972504e-01 -1.77485490e+00 1.95633650e-01 -5.97681880e-01
-5.16034663e-01 1.09518647e+00 6.78529143e-01 -1.54091150e-01
7.95738816e-01 3.92732769e-01 3.85996401e-01 -3.24522674e-01
-9.46107805e-01 -1.83302730e-01 2.80271381e-01 5.88697612e-01
8.02718639e-01 9.59381238e-02 -7.90545493e-02 7.23105311e-01
2.12342236e-02 3.10733795e-01 3.77569079e-01 9.82652843e-01
-2.03840420e-01 -1.06854558e+00 7.37287477e-02 7.33044982e-01
-5.58985174e-01 -2.95380652e-02 -3.83435756e-01 7.92437971e-01
-1.06574461e-01 6.07235253e-01 -6.93632364e-02 -2.72530258e-01
-1.83072478e-01 5.02291769e-02 2.62450248e-01 -1.05248082e+00
-6.61436439e-01 3.77475955e-02 -1.40885592e-01 -4.65090513e-01
-6.38428926e-01 -4.56277609e-01 -1.29861879e+00 2.31541038e-01
-4.02671278e-01 8.64919722e-02 3.35022151e-01 1.05926216e+00
2.68402517e-01 5.15378594e-01 7.41646290e-01 -5.94338357e-01
-6.98188901e-01 -9.98911262e-01 -4.26959068e-01 5.31100929e-01
4.15215194e-01 -6.46272242e-01 -5.15389860e-01 2.20423296e-01] | [14.984136581420898, -2.595343589782715] |
11354630-66cc-4523-ad30-22f1d88efff7 | age-group-and-gender-estimation-in-the-wild | 1710.02985 | null | http://arxiv.org/abs/1710.02985v1 | http://arxiv.org/pdf/1710.02985v1.pdf | Age Group and Gender Estimation in the Wild with Deep RoR Architecture | Automatically predicting age group and gender from face images acquired in
unconstrained conditions is an important and challenging task in many
real-world applications. Nevertheless, the conventional methods with
manually-designed features on in-the-wild benchmarks are unsatisfactory because
of incompetency to tackle large variations in unconstrained images. This
difficulty is alleviated to some degree through Convolutional Neural Networks
(CNN) for its powerful feature representation. In this paper, we propose a new
CNN based method for age group and gender estimation leveraging Residual
Networks of Residual Networks (RoR), which exhibits better optimization ability
for age group and gender classification than other CNN architectures.Moreover,
two modest mechanisms based on observation of the characteristics of age group
are presented to further improve the performance of age estimation.In order to
further improve the performance and alleviate over-fitting problem, RoR model
is pre-trained on ImageNet firstly, and then it is fune-tuned on the
IMDB-WIKI-101 data set for further learning the features of face images,
finally, it is used to fine-tune on Adience data set. Our experiments
illustrate the effectiveness of RoR method for age and gender estimation in the
wild, where it achieves better performance than other CNN methods. Finally, the
RoR-152+IMDB-WIKI-101 with two mechanisms achieves new state-of-the-art results
on Adience benchmark. | ['Xingfang Yuan', 'Liru Guo', 'Zhenbing Zhao', 'Tony X. Han', 'Baogang Li', 'Ke Zhang', 'Ce Gao', 'Miao Sun'] | 2017-10-09 | null | null | null | null | ['age-and-gender-estimation', 'age-and-gender-classification'] | ['computer-vision', 'computer-vision'] | [-4.03289348e-01 1.32919371e-01 -9.02468339e-02 -7.99151897e-01
-1.24641798e-01 2.41083100e-01 4.13519889e-01 -4.08604205e-01
-6.04080558e-01 5.73305964e-01 2.40063835e-02 1.99613288e-01
-1.60292745e-01 -8.28554094e-01 -2.58145332e-01 -9.15594995e-01
-1.78007394e-01 4.83670175e-01 -5.88369191e-01 -1.83423877e-01
8.61507729e-02 4.21691954e-01 -1.84185445e+00 -3.24924737e-01
8.34898770e-01 1.45169318e+00 -1.85460895e-01 2.79048473e-01
1.20218307e-01 6.50160789e-01 -5.58874786e-01 -7.64639795e-01
1.38019323e-01 -8.85118470e-02 -5.34099042e-01 -1.90244123e-01
7.73586929e-01 -5.85046351e-01 -5.62983811e-01 9.51810360e-01
9.91517961e-01 -1.59512997e-01 8.10391247e-01 -1.69349360e+00
-7.43614614e-01 5.16833842e-01 -8.27700913e-01 -2.33400628e-01
-4.17159423e-02 -9.31467861e-02 7.57257640e-01 -9.06736314e-01
2.19935834e-01 1.60365736e+00 8.16596687e-01 1.14018846e+00
-5.64350784e-01 -1.34442008e+00 1.73087597e-01 1.77834988e-01
-1.57523668e+00 -5.05714715e-01 5.81902146e-01 -2.62355834e-01
5.66626132e-01 9.55520272e-02 3.90247762e-01 1.36944556e+00
-1.49572805e-01 5.44395626e-01 9.71926749e-01 7.88192451e-02
-2.69203931e-01 -3.44189703e-02 1.43517684e-02 1.03548479e+00
4.59012054e-02 1.13652736e-01 -6.24710679e-01 5.95092587e-02
7.62807429e-01 1.77820697e-01 -8.31531710e-04 5.93609102e-02
-8.17855775e-01 8.36141944e-01 4.60635841e-01 1.24182895e-01
-1.15379252e-01 6.98016584e-02 5.82034469e-01 3.15246433e-01
8.24285388e-01 4.70072865e-01 -9.12143469e-01 -4.39165160e-02
-8.01158309e-01 4.08309340e-01 5.57119548e-01 6.55875146e-01
5.53925455e-01 1.92515954e-01 -3.76324445e-01 1.28433752e+00
2.15626583e-01 5.24117291e-01 6.26311719e-01 -7.59249270e-01
3.79351616e-01 8.29242587e-01 -3.81253511e-01 -9.67853546e-01
-6.25269473e-01 -4.99528915e-01 -1.28300273e+00 4.08925116e-02
6.35755718e-01 -2.33094662e-01 -1.19857252e+00 2.05287671e+00
1.63582042e-01 1.78839371e-01 -6.18885681e-02 6.97994351e-01
1.43463647e+00 4.07420725e-01 3.30100000e-01 -2.65217781e-01
1.56604671e+00 -8.15398872e-01 -5.21575093e-01 -2.66457617e-01
5.15535474e-01 -4.29948509e-01 7.05999255e-01 4.16054785e-01
-8.51159036e-01 -7.48639822e-01 -7.09184647e-01 9.10907686e-02
-4.42826271e-01 6.93153441e-01 9.82388973e-01 7.56865919e-01
-1.03884995e+00 7.38655508e-01 -3.59178871e-01 -2.97224998e-01
9.15453315e-01 9.23768580e-01 -7.10047543e-01 -6.25764057e-02
-1.07088089e+00 4.24697280e-01 2.19484538e-01 3.60977441e-01
-7.01234221e-01 -1.09315383e+00 -1.05479658e+00 4.00654040e-02
2.18960777e-01 -7.71457076e-01 9.77377594e-01 -8.35845530e-01
-1.32615376e+00 1.20971262e+00 8.34290683e-02 -1.97469324e-01
5.74999332e-01 -1.50841400e-01 -4.28279877e-01 -1.87259048e-01
-1.40741952e-02 9.81025875e-01 1.08541226e+00 -8.33677232e-01
-5.48133612e-01 -1.07316816e+00 -2.83578001e-02 1.21895121e-02
-9.15684462e-01 2.11241916e-01 -5.67646742e-01 -5.93720019e-01
-3.66370976e-02 -6.84182048e-01 -1.21416852e-01 -1.46745503e-01
-2.19463274e-01 -6.71053648e-01 7.09830761e-01 -8.27335894e-01
1.28606749e+00 -2.05653119e+00 9.46504846e-02 -2.80649681e-02
5.94908357e-01 3.53167266e-01 -2.41578311e-01 -2.50523061e-01
-3.06122899e-01 3.36402915e-02 1.57508612e-01 -6.99989617e-01
-1.22292759e-02 -1.26163810e-01 1.49081916e-01 3.56996804e-01
2.83384293e-01 7.10614681e-01 -5.45782089e-01 -6.22631967e-01
-1.04601651e-01 5.32858551e-01 -5.79888701e-01 6.40790045e-01
1.73044279e-01 5.09859800e-01 -4.31853265e-01 9.82907772e-01
1.03059423e+00 1.74765795e-01 -5.51607385e-02 -4.05922532e-01
1.81566745e-01 -4.38252509e-01 -8.86997521e-01 1.25654078e+00
-7.12411404e-01 1.46537751e-01 1.40842766e-01 -1.02337015e+00
1.37007082e+00 1.50469586e-01 5.39423704e-01 -7.92654693e-01
5.36036134e-01 1.40402317e-01 1.17431492e-01 -4.22079444e-01
2.94659317e-01 1.36891708e-01 -1.67561382e-01 4.89858761e-02
3.61006796e-01 1.99762166e-01 1.91056296e-01 -1.17211819e-01
8.16907883e-01 -1.18098564e-01 -1.54477537e-01 -1.45223230e-01
9.46603894e-01 -8.81545126e-01 9.17883694e-01 3.30216318e-01
-3.38158578e-01 7.17321754e-01 6.37323797e-01 -8.12844634e-01
-9.01939571e-01 -6.59099460e-01 -3.87778163e-01 1.40147495e+00
-3.68913263e-01 -5.53593874e-01 -8.41148496e-01 -1.04688811e+00
2.70818979e-01 -5.57279438e-02 -1.16544580e+00 -2.49512449e-01
-5.15704870e-01 -1.21573150e+00 6.18094027e-01 7.83627152e-01
1.01750648e+00 -1.22311807e+00 7.95628801e-02 -2.74073929e-01
1.08112283e-01 -1.15151310e+00 -2.59425551e-01 -2.81358212e-01
-6.51019096e-01 -1.15553141e+00 -9.75549698e-01 -8.05426240e-01
9.32200015e-01 -3.21029782e-01 1.24257672e+00 4.34497416e-01
-4.17451888e-01 9.37077552e-02 -5.57588749e-02 -5.54481268e-01
1.07580736e-01 4.89798754e-01 2.85556167e-01 2.81707436e-01
6.32671595e-01 -6.48110807e-01 -8.23786616e-01 4.19185340e-01
-2.70199239e-01 -2.38636851e-01 5.31293273e-01 8.92890930e-01
9.49683134e-03 4.35245000e-02 9.63304639e-01 -1.00029409e+00
2.71551430e-01 -3.43656600e-01 -4.77219433e-01 -4.15448882e-02
-7.23805010e-01 -1.60307586e-01 6.83292925e-01 -4.99906719e-01
-1.06044900e+00 -4.72867377e-02 -5.22801101e-01 -3.83476913e-01
-1.33423246e-02 1.75903484e-01 -4.71087307e-01 -3.11186425e-02
2.09651127e-01 -1.94035605e-01 3.48155677e-01 -6.26554370e-01
-1.10845864e-01 8.64912033e-01 5.30755103e-01 -7.21938312e-01
8.29587221e-01 1.11166716e-01 1.37823045e-01 -5.78834236e-01
-1.11865866e+00 -9.16039422e-02 -3.89523029e-01 -1.03039779e-01
7.71302640e-01 -1.32581484e+00 -1.29141533e+00 1.16727936e+00
-6.48711205e-01 -2.00762197e-01 3.67481142e-01 4.11009133e-01
-2.79789597e-01 -6.64722472e-02 -6.82301581e-01 -7.22768545e-01
-8.17491829e-01 -1.04885542e+00 1.18291783e+00 6.81570232e-01
7.59854689e-02 -8.21856856e-01 -4.43176717e-01 5.51801085e-01
4.95424300e-01 2.31524527e-01 7.15108097e-01 -6.44823194e-01
3.88158509e-03 -2.09786654e-01 -6.99308693e-01 5.38394809e-01
-5.01042493e-02 2.23779038e-01 -1.21424580e+00 -3.32484454e-01
-3.77560616e-01 -6.20232582e-01 1.08714998e+00 4.38901961e-01
2.02240777e+00 -1.65817767e-01 -1.03722759e-01 9.51634169e-01
1.17584741e+00 -1.57755777e-01 7.39784360e-01 1.85469374e-01
1.04861689e+00 8.43572736e-01 7.56488144e-01 7.76011646e-01
6.83709502e-01 5.65306664e-01 5.59988379e-01 -3.39401364e-01
7.80494660e-02 -6.37383163e-02 -4.39403616e-02 6.60239875e-01
-6.50340676e-01 3.53406549e-01 -7.56309450e-01 5.08233368e-01
-1.76573288e+00 -5.32827735e-01 9.94238034e-02 2.11848998e+00
8.94539773e-01 -9.37638134e-02 3.55720729e-01 -4.29933853e-02
6.61034882e-01 2.90855199e-01 -4.65725303e-01 -3.60795677e-01
3.39159071e-02 4.10910219e-01 4.80906397e-01 -1.75515026e-01
-1.19348109e+00 9.25088227e-01 5.65239096e+00 1.04454505e+00
-1.16926360e+00 2.21585184e-02 1.30087769e+00 1.72987893e-01
4.87571329e-01 -5.99093199e-01 -1.32776225e+00 4.97945189e-01
8.07719469e-01 8.46348181e-02 5.09501755e-01 1.12414217e+00
-2.26375982e-01 2.69142568e-01 -1.11279261e+00 1.55964255e+00
2.67970353e-01 -8.59228909e-01 -3.59999478e-01 8.91652256e-02
5.84515274e-01 -5.25200069e-01 4.64348316e-01 9.53911304e-01
2.97157075e-02 -1.57097137e+00 2.58422047e-01 3.38822752e-01
1.16131306e+00 -1.20566440e+00 1.13731050e+00 -5.44309011e-03
-1.07895577e+00 -5.25695384e-01 -7.40111351e-01 -2.30954871e-01
-4.80711967e-01 7.04499364e-01 -4.29011852e-01 4.75380152e-01
1.09036934e+00 8.28448355e-01 -1.17732143e+00 6.13732278e-01
-9.35130566e-02 3.35361868e-01 2.03829929e-02 1.60810217e-01
-1.29098386e-01 -8.08090791e-02 -2.36956149e-01 6.58849478e-01
4.62942690e-01 -1.21834643e-01 -1.21789500e-01 5.51348448e-01
-5.67693770e-01 2.74758637e-01 -4.99177754e-01 -8.94778175e-04
3.67690980e-01 1.88704252e+00 -2.87231147e-01 -1.26030803e-01
-2.24719748e-01 4.77204114e-01 5.73735833e-01 1.05698295e-01
-7.86615551e-01 -3.14794987e-01 9.13634181e-01 1.00556813e-01
6.32568151e-02 1.18215114e-01 -9.14631598e-03 -1.11083448e+00
-2.34883025e-01 -1.01028025e+00 5.10528803e-01 -5.10315835e-01
-1.53697383e+00 6.85271978e-01 -6.98113590e-02 -7.64845431e-01
-2.84478635e-01 -9.72347736e-01 -7.79559493e-01 7.75781512e-01
-1.35315299e+00 -1.61919820e+00 -7.44458675e-01 6.68276370e-01
3.10914636e-01 -7.35919476e-01 7.73529351e-01 6.29001260e-01
-1.11496687e+00 1.25457883e+00 -3.06206346e-01 5.09852648e-01
1.14766586e+00 -1.34279287e+00 2.16426045e-01 2.85329819e-01
-5.68021417e-01 6.70916498e-01 3.82827282e-01 -3.51514429e-01
-1.25179684e+00 -1.33842576e+00 8.86119485e-01 -3.88284296e-01
3.08728278e-01 -7.00637937e-01 -6.43991232e-01 5.41350305e-01
-8.43545720e-02 2.06954807e-01 4.85238791e-01 6.49905443e-01
-6.29665196e-01 -6.67423785e-01 -1.29972911e+00 5.69627345e-01
1.21613228e+00 -2.50498682e-01 -6.97059184e-02 3.14267129e-01
5.45143604e-01 -3.74332637e-01 -1.28662384e+00 9.89688933e-01
7.63279617e-01 -8.85992408e-01 1.17165518e+00 -5.85765898e-01
9.17003632e-01 2.15024292e-01 1.33018181e-01 -1.25513041e+00
-9.51608419e-02 -3.77028674e-01 -1.52448863e-01 1.88114965e+00
5.45339175e-02 -6.69378102e-01 1.02851987e+00 5.94208419e-01
2.85158455e-01 -1.22426331e+00 -6.39694333e-01 -5.80641985e-01
2.40175456e-01 -2.23271474e-02 1.12963104e+00 7.96144843e-01
-6.84847891e-01 4.07458365e-01 -5.33878624e-01 -1.52677223e-01
7.03936338e-01 -2.96607405e-01 1.00923038e+00 -1.49957597e+00
6.82917684e-02 -4.06973362e-01 -5.76132238e-01 -2.85513848e-01
6.32815063e-01 -4.70125675e-01 -1.80047974e-01 -9.62642610e-01
3.45590502e-01 -5.50507188e-01 -3.86952549e-01 5.04792213e-01
-4.17958468e-01 7.47890115e-01 1.73030253e-02 -2.30116934e-01
-5.71506619e-01 7.74998188e-01 1.27180362e+00 -1.74459040e-01
1.60155162e-01 2.84362346e-01 -8.58179986e-01 7.71129012e-01
8.73483777e-01 -1.93493009e-01 -2.10062474e-01 -4.32680994e-02
2.52514809e-01 -9.48475376e-02 9.38045010e-02 -9.46348131e-01
-1.04238823e-01 6.74989149e-02 9.77599442e-01 -5.82504153e-01
4.23395753e-01 -4.14132893e-01 -2.08131418e-01 2.05287039e-01
-3.16441543e-02 1.28226802e-01 -1.95993092e-02 -3.79053131e-02
-2.04617336e-01 -2.55624384e-01 8.07528615e-01 -1.21788293e-01
-5.96208394e-01 1.14329147e+00 3.25750023e-01 7.90084451e-02
7.50178456e-01 -7.55963428e-03 -4.36685711e-01 -2.05349624e-01
-6.26126528e-01 4.42129254e-01 2.90853560e-01 6.59831703e-01
4.01820272e-01 -1.59022713e+00 -8.11108530e-01 3.96988213e-01
2.50684500e-01 -7.39029571e-02 5.25304079e-01 8.11390460e-01
-2.84486115e-01 9.62654129e-02 -5.21951675e-01 -4.42552716e-01
-1.60748231e+00 3.97159338e-01 3.13979179e-01 -5.22770882e-01
-6.00555865e-03 1.14640141e+00 5.43870509e-01 -9.91071582e-01
6.06310308e-01 3.59063715e-01 -8.13714981e-01 3.32307070e-01
6.99317753e-01 2.49090970e-01 8.95213410e-02 -7.05673635e-01
-4.20393854e-01 7.31397510e-01 -3.31997156e-01 5.24446726e-01
1.63284326e+00 2.65108161e-02 -5.02356470e-01 -1.40641570e-01
1.36660707e+00 -1.59061491e-01 -1.12007260e+00 -1.59737151e-02
-3.74774575e-01 -6.44729853e-01 -1.12769395e-01 -6.89717948e-01
-1.75311756e+00 8.97436678e-01 8.31608355e-01 -3.57531846e-01
1.24665189e+00 -1.90733224e-01 6.49725676e-01 2.05440179e-01
1.72345772e-01 -1.20488155e+00 3.47856224e-01 4.48568225e-01
7.47517467e-01 -1.52743542e+00 -2.51755808e-02 -4.05802101e-01
-5.02581120e-01 1.20130825e+00 1.28843486e+00 8.67525637e-02
7.66700327e-01 -1.52643904e-01 3.95754129e-02 -2.00652122e-01
-4.41554815e-01 -1.79058179e-01 3.42936784e-01 7.62942374e-01
5.48742950e-01 -1.80621445e-02 -2.92190433e-01 1.19643855e+00
-6.05077565e-01 -3.09493802e-02 1.59828484e-01 2.42095158e-01
-1.14201475e-02 -1.13652444e+00 -2.29666412e-01 7.77033567e-01
-9.08519566e-01 5.12650907e-02 -1.66000873e-01 8.12175870e-01
4.96428102e-01 6.95227325e-01 1.17831454e-01 -5.44487715e-01
1.58571854e-01 -1.43516585e-01 6.70347929e-01 -4.90232766e-01
-7.66842067e-01 -3.98900837e-01 1.67013779e-01 -5.53884089e-01
-1.43325150e-01 -3.88372600e-01 -7.11530805e-01 -5.75547814e-01
-1.26744345e-01 -5.69375372e-03 7.35398114e-01 7.51948237e-01
1.43900037e-01 5.38725495e-01 9.82609212e-01 -7.43346095e-01
-5.44781268e-01 -1.27604401e+00 -8.13296854e-01 5.55338621e-01
-1.47458524e-01 -9.53877091e-01 -3.98905486e-01 -3.99010688e-01] | [13.53469181060791, 0.8985046148300171] |
566cff8b-f8f6-4c16-a3d2-33c2c7e014a6 | video4mri-an-empirical-study-on-brain | 2302.12688 | null | https://arxiv.org/abs/2302.12688v1 | https://arxiv.org/pdf/2302.12688v1.pdf | Video4MRI: An Empirical Study on Brain Magnetic Resonance Image Analytics with CNN-based Video Classification Frameworks | To address the problem of medical image recognition, computer vision techniques like convolutional neural networks (CNN) are frequently used. Recently, 3D CNN-based models dominate the field of magnetic resonance image (MRI) analytics. Due to the high similarity between MRI data and videos, we conduct extensive empirical studies on video recognition techniques for MRI classification to answer the questions: (1) can we directly use video recognition models for MRI classification, (2) which model is more appropriate for MRI, (3) are the common tricks like data augmentation in video recognition still useful for MRI classification? Our work suggests that advanced video techniques benefit MRI classification. In this paper, four datasets of Alzheimer's and Parkinson's disease recognition are utilized in experiments, together with three alternative video recognition models and data augmentation techniques that are frequently applied to video tasks. In terms of efficiency, the results reveal that the video framework performs better than 3D-CNN models by 5% - 11% with 50% - 66% less trainable parameters. This report pushes forward the potential fusion of 3D medical imaging and video understanding research. | ['Haoyi Xiong', 'Dejing Dou', 'Yanwu Xu', 'Yi Liu', 'Jiang Bian', 'Qingzhong Wang', 'Yuxuan Zhang'] | 2023-02-24 | null | null | null | null | ['video-classification', 'video-recognition', 'video-understanding'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.15417820e-01 -4.57569286e-02 -5.06960332e-01 -3.30745429e-01
-3.64602804e-01 -3.40564139e-02 2.44373471e-01 -1.00882828e-01
-6.82163894e-01 3.73896658e-01 5.08915305e-01 -4.03819412e-01
3.04592978e-02 -3.73370439e-01 -6.48431778e-01 -6.07633591e-01
-6.04372561e-01 2.24736691e-01 -2.33442448e-02 -6.76435158e-02
1.73222810e-01 5.31773388e-01 -1.60194993e+00 7.02687085e-01
2.50533640e-01 1.21529007e+00 9.70194265e-02 8.65721822e-01
-1.15843974e-01 1.20574975e+00 -2.30003938e-01 -3.61705683e-02
2.21706703e-01 -7.99807683e-02 -8.91266704e-01 3.95901054e-01
4.69909132e-01 -8.22996438e-01 -9.07236516e-01 9.03405190e-01
4.88586247e-01 -3.55762057e-02 5.36075950e-01 -1.20170760e+00
-6.93675220e-01 3.20595056e-01 -4.14067358e-01 1.11187279e+00
3.14441144e-01 2.09579945e-01 4.02450591e-01 -7.70938277e-01
7.68936813e-01 1.09084749e+00 5.77078700e-01 7.54739821e-01
-7.60050118e-01 -4.86775070e-01 1.91079244e-01 7.03123450e-01
-1.09028602e+00 -4.10700798e-01 6.54724419e-01 -6.61305428e-01
1.01506782e+00 3.49298328e-01 8.91642570e-01 1.40047562e+00
3.59622538e-01 1.09703076e+00 1.07281613e+00 -3.18072826e-01
1.05284095e-01 -2.48870850e-01 3.55538785e-01 6.06709957e-01
1.14092799e-02 1.04718395e-01 -5.26692212e-01 -4.63631004e-03
1.06275868e+00 3.04299772e-01 -4.50323522e-01 -1.41262636e-01
-1.40693402e+00 1.05594993e+00 3.35730821e-01 5.99542499e-01
-4.75428641e-01 1.85290545e-01 6.99374199e-01 4.96715307e-01
6.57117724e-01 1.05217218e-01 -3.31888825e-01 -4.24718708e-01
-6.06303692e-01 1.07654847e-01 5.24455130e-01 6.91186368e-01
9.57649350e-02 3.54480505e-01 -1.00550257e-01 8.94093156e-01
1.68017834e-01 3.10852200e-01 8.33588600e-01 -7.13965535e-01
4.52395588e-01 4.45522636e-01 -4.01743948e-01 -1.08159769e+00
-4.45268899e-01 -3.58522117e-01 -9.64047551e-01 -2.54847612e-02
2.41958067e-01 1.16037190e-01 -1.19717610e+00 1.23822176e+00
1.24167390e-01 4.72309530e-01 -1.76708907e-01 9.80543256e-01
1.10120642e+00 3.23737472e-01 9.93366241e-02 -3.30474168e-01
1.46049809e+00 -8.49687934e-01 -7.97515392e-01 -1.47279024e-01
9.82450366e-01 -4.44576472e-01 7.46582687e-01 1.94699079e-01
-9.03092742e-01 -5.74228942e-01 -7.99125552e-01 9.12808403e-02
-1.72860786e-01 -3.90346833e-02 8.48000050e-01 7.52311707e-01
-1.09616148e+00 3.94446373e-01 -1.12648320e+00 -3.34705889e-01
7.73835599e-01 5.42368710e-01 -7.53318191e-01 -4.53554779e-01
-9.27656829e-01 9.83089328e-01 2.68716991e-01 1.93479463e-01
-1.05893123e+00 -8.01977754e-01 -7.54154563e-01 -2.88665831e-01
1.42342970e-01 -5.86083531e-01 1.13402367e+00 -1.28873909e+00
-9.04103398e-01 9.79305089e-01 1.06113248e-01 -6.38830841e-01
3.87494564e-01 -1.87261701e-01 -5.06024778e-01 7.50634134e-01
-2.13361129e-01 7.95027137e-01 8.22783291e-01 -7.06298947e-01
-4.30303305e-01 -5.05075514e-01 2.91735739e-01 6.01031892e-02
-5.20824254e-01 1.17162809e-01 -3.15477163e-01 -9.38003540e-01
7.16831833e-02 -9.82721329e-01 -3.70717257e-01 1.46609917e-01
-9.10633989e-03 1.07828617e-01 9.76110339e-01 -9.73279953e-01
9.31428075e-01 -2.09754348e+00 1.15592785e-01 -1.25606552e-01
5.58471143e-01 4.96231288e-01 -2.28729129e-01 -8.60776380e-02
-4.82756197e-01 1.70085162e-01 6.40208423e-02 1.05549686e-01
-5.71360707e-01 3.99457842e-01 2.76717156e-01 5.67418396e-01
1.78791255e-01 1.05512035e+00 -6.69476509e-01 -3.72672379e-01
3.72941554e-01 6.74494743e-01 -7.56280005e-01 9.19171572e-02
2.66847551e-01 7.37028420e-01 -4.98264313e-01 9.76507664e-01
4.83423769e-01 -4.08601761e-01 6.90572709e-03 -6.63050234e-01
2.79545307e-01 -2.01297119e-01 -6.88440382e-01 1.56699526e+00
1.35777593e-02 8.54504883e-01 -1.67007688e-02 -1.73196495e+00
4.48282510e-01 5.04396558e-01 1.03182852e+00 -9.01642680e-01
2.80818313e-01 -5.12572266e-02 2.20221162e-01 -1.30777633e+00
-8.80616531e-02 -8.07924047e-02 4.39094990e-01 1.58358201e-01
1.15066841e-01 5.52363932e-01 4.11438458e-02 -1.05989896e-01
1.51110923e+00 -1.74555227e-01 1.02138624e-01 -4.56009097e-02
3.87377232e-01 1.05391555e-01 3.46807212e-01 6.83687627e-01
-5.75996816e-01 6.42930984e-01 2.28103414e-01 -1.00779855e+00
-1.12683189e+00 -6.29080117e-01 5.42580010e-03 7.21956670e-01
-1.11883290e-01 -1.72037020e-01 -6.86998248e-01 -6.27491355e-01
-1.58016235e-01 1.04520023e-01 -8.77741694e-01 -6.94756880e-02
-7.65589058e-01 -1.01769137e+00 3.29666287e-01 9.18848157e-01
3.77186388e-01 -9.50471461e-01 -6.27875865e-01 1.26314804e-01
-3.22824746e-01 -1.42018354e+00 -4.25620228e-01 -1.24419078e-01
-1.48058236e+00 -1.32795966e+00 -9.31575179e-01 -9.00897145e-01
8.88487935e-01 7.63472259e-01 7.84510016e-01 4.19412851e-01
-4.89311934e-01 1.03549969e+00 -7.83294261e-01 -1.19918577e-01
-3.90942425e-01 -1.90107271e-01 1.80760175e-01 -5.03436401e-02
5.06074250e-01 -5.84383845e-01 -7.18143642e-01 1.56416446e-01
-1.20929348e+00 1.95898741e-01 8.60738099e-01 8.99573863e-01
4.44103926e-01 -2.71900743e-01 2.62926996e-01 -6.81492209e-01
4.15244430e-01 -5.66444039e-01 6.46560714e-02 2.23755285e-01
-3.52241457e-01 -1.84600189e-01 1.95106104e-01 -8.90242398e-01
-4.75450248e-01 3.45805399e-02 -3.27310711e-01 -9.43319738e-01
-3.69151682e-01 7.61781156e-01 3.33029538e-01 -4.40411568e-01
4.94697511e-01 2.21659392e-01 4.72560316e-01 -5.11473298e-01
-4.13583033e-02 6.17995799e-01 3.47087234e-01 -9.40207466e-02
4.66047019e-01 7.02616572e-01 -1.98989600e-01 -1.22436905e+00
-6.25926673e-01 -4.41321701e-01 -6.97710931e-01 -5.43647826e-01
1.15584564e+00 -1.03128064e+00 -4.54608351e-01 3.92529964e-01
-9.65523601e-01 5.77660203e-02 4.89124991e-02 8.81923199e-01
-4.05223906e-01 6.25566483e-01 -8.94168198e-01 -5.87631702e-01
-3.86505067e-01 -1.51717341e+00 8.43532443e-01 4.45214175e-02
-6.14196695e-02 -9.31100190e-01 -2.41737291e-01 8.51902306e-01
5.18929601e-01 2.82935202e-01 1.15995383e+00 -7.81445563e-01
-4.18458968e-01 -3.12473834e-01 -2.23445386e-01 5.10018528e-01
4.16593216e-02 -3.39334875e-01 -8.32971990e-01 -3.35429221e-01
3.28424245e-01 -1.88504696e-01 6.12526715e-01 7.35995829e-01
1.48556304e+00 -3.66791308e-01 -2.32984573e-01 6.32345080e-01
1.18426239e+00 5.47332406e-01 8.96620333e-01 6.26437604e-01
8.38063061e-01 4.37306702e-01 2.16511607e-01 1.08073346e-01
3.80451471e-01 6.22467458e-01 6.09233856e-01 -7.65419155e-02
-1.98512658e-01 2.04781055e-01 5.38577855e-01 1.44588935e+00
-3.97204459e-01 1.47094160e-01 -9.86147583e-01 3.40166867e-01
-1.76068199e+00 -9.53521609e-01 -8.18922892e-02 1.80116665e+00
3.01059037e-01 5.25583215e-02 1.26416802e-01 2.68494189e-01
6.60850406e-01 2.21164133e-02 -3.48803371e-01 3.41682769e-02
2.38320492e-02 8.64888951e-02 3.44935358e-01 -2.11986620e-02
-1.37889540e+00 3.32713842e-01 6.63668394e+00 5.94157994e-01
-1.33582819e+00 3.51204097e-01 8.74683022e-01 -2.07566768e-01
1.70365661e-01 -3.14739287e-01 -3.46142501e-01 5.10220289e-01
1.03606331e+00 1.85267359e-01 2.52176434e-01 8.94562721e-01
2.84509987e-01 2.57156100e-02 -1.16557801e+00 1.51159644e+00
5.03082871e-01 -1.71133554e+00 3.35311413e-01 6.91275522e-02
3.26479703e-01 1.03669524e-01 -5.80255389e-02 3.02089691e-01
-5.52470624e-01 -1.23136330e+00 4.48749214e-01 5.00150561e-01
5.14509261e-01 -3.63307536e-01 9.30576861e-01 8.73109847e-02
-9.41597998e-01 -2.16496095e-01 -1.91118404e-01 8.45682845e-02
6.78617358e-02 1.86509416e-01 -7.30326474e-01 4.41922605e-01
9.57236409e-01 9.56176221e-01 -6.12536907e-01 1.17530787e+00
3.08534831e-01 5.15897393e-01 7.32964948e-02 2.69861132e-01
3.84727746e-01 9.66421217e-02 4.95232731e-01 1.09986997e+00
2.38054648e-01 3.05946320e-01 2.75441021e-01 1.99942902e-01
2.61959229e-02 -2.13310751e-03 -7.59513974e-01 -2.76715279e-01
-7.46678263e-02 9.48030114e-01 -9.60242450e-01 -3.09672892e-01
-8.18484306e-01 6.95994854e-01 -9.30854231e-02 2.09045887e-01
-7.25063860e-01 3.13438356e-01 5.33264697e-01 1.82696462e-01
3.18771631e-01 -3.36623400e-01 4.67318892e-02 -1.30271220e+00
-3.89812933e-03 -1.17353821e+00 3.78045142e-01 -9.06873405e-01
-1.25191283e+00 6.54884756e-01 4.15746830e-02 -1.41280270e+00
-6.10359572e-02 -9.76069033e-01 -2.87189245e-01 4.53511477e-02
-1.37024629e+00 -9.59765494e-01 -4.21048909e-01 6.66350842e-01
9.48854446e-01 -3.34814072e-01 7.53130555e-01 7.32329071e-01
-5.14524400e-01 9.94080827e-02 -1.77905597e-02 5.39923787e-01
1.79081917e-01 -6.03248775e-01 1.03157647e-01 5.43673635e-01
1.10847004e-01 1.52407631e-01 5.04981637e-01 -5.98513603e-01
-1.90943778e+00 -9.65133011e-01 4.68244880e-01 -2.96272248e-01
5.82135022e-01 -3.65999714e-02 -9.72862601e-01 6.05613232e-01
-2.05554795e-02 2.39330009e-01 8.15414727e-01 -4.51977521e-01
-1.87529564e-01 1.74019679e-01 -1.07883132e+00 5.18920958e-01
1.11657119e+00 -5.48225522e-01 -5.86070061e-01 5.74495792e-01
5.90387285e-01 -4.49882358e-01 -1.16908014e+00 6.08957708e-01
7.06316113e-01 -5.99022865e-01 1.27351546e+00 -1.09763384e+00
7.30484307e-01 1.06936276e-01 -4.16847557e-01 -1.00673497e+00
-1.72566175e-01 8.34141448e-02 -4.48547870e-01 5.02472520e-01
1.03030816e-01 -6.77407309e-02 9.49453771e-01 6.06698155e-01
-3.94973420e-02 -9.64327395e-01 -1.23699701e+00 -6.89995885e-01
-1.04624726e-01 -6.87456071e-01 -4.88471650e-02 1.17517471e+00
-2.72064716e-01 -4.66240123e-02 -5.12528002e-01 -8.22284818e-02
3.72834682e-01 -4.41793740e-01 3.94266754e-01 -1.11481547e+00
-4.33560042e-03 -3.65709692e-01 -1.10051990e+00 -7.35891044e-01
1.73866991e-02 -1.04311192e+00 -4.58541960e-01 -1.42272651e+00
4.50999469e-01 -1.42363249e-03 -3.39882523e-01 4.70812470e-01
1.96496993e-01 3.82049412e-01 3.11105877e-01 3.83171290e-01
-5.15742123e-01 4.39317495e-01 1.37086070e+00 -6.18824184e-01
1.56227082e-01 -1.84420034e-01 -4.85631257e-01 6.28517032e-01
5.05073488e-01 -3.45966816e-01 -3.71771604e-01 -7.96376169e-01
-1.58769995e-01 2.59531200e-01 4.50951248e-01 -9.16571259e-01
1.23354502e-01 2.13900153e-02 5.75861335e-01 -3.23788106e-01
3.42131406e-01 -8.86911929e-01 9.12963301e-02 7.88892090e-01
-2.18635201e-01 4.89008188e-01 2.17345208e-01 6.56462371e-01
-3.49722683e-01 -1.69696704e-01 6.48639798e-01 -5.10488331e-01
-9.86161709e-01 7.15193748e-01 -9.15820658e-01 -5.38071319e-02
1.01499152e+00 -5.72473526e-01 -1.26824364e-01 -4.30075228e-01
-1.06468403e+00 -1.43516973e-01 -5.07422164e-02 7.70639837e-01
9.48976874e-01 -1.55455911e+00 -5.63231945e-01 9.59064290e-02
1.18566789e-01 -5.25790274e-01 4.94679838e-01 1.43293130e+00
-7.22267032e-01 5.53995788e-01 -5.44842422e-01 -9.49736714e-01
-1.38345492e+00 4.09650892e-01 3.03570539e-01 9.98167228e-03
-1.13012731e+00 5.96798718e-01 8.43158364e-02 2.02290900e-03
6.15513682e-01 -4.57643300e-01 -4.02624130e-01 2.10995786e-02
8.28118682e-01 2.00419575e-01 2.60136575e-01 -7.24011004e-01
-4.02147442e-01 2.83147961e-01 -4.50342447e-01 2.70559609e-01
1.82390249e+00 -2.04063058e-02 1.75138339e-01 2.23056421e-01
1.41214108e+00 -1.01561797e+00 -9.28578258e-01 -1.20947048e-01
2.20406294e-01 -4.43453580e-01 4.45813328e-01 -5.36975205e-01
-1.73465753e+00 9.57095683e-01 1.17843688e+00 1.12155393e-01
1.06579185e+00 -1.43986866e-01 7.90821195e-01 3.06410849e-01
3.82677644e-01 -1.01567090e+00 2.37079278e-01 3.37670088e-01
8.99202228e-01 -1.37922013e+00 -1.07837453e-01 -2.15686545e-01
-6.89399064e-01 1.39646661e+00 5.92729628e-01 -1.23259321e-01
8.69234085e-01 1.48710251e-01 -4.92317155e-02 -5.16082227e-01
-6.29388511e-01 1.42536983e-01 2.25908235e-01 7.40313113e-01
4.30456609e-01 -2.82943726e-01 -2.02448100e-01 4.44754422e-01
3.55249852e-01 3.48020792e-01 5.45493841e-01 1.23799491e+00
-2.58076936e-01 -8.18277061e-01 -3.45662802e-01 9.13823128e-01
-6.32422626e-01 2.35438552e-02 1.72030237e-02 9.31990087e-01
-4.24049757e-02 7.27913976e-01 9.22656581e-02 -5.34609020e-01
1.83110639e-01 5.36330938e-02 7.17170238e-01 -3.37199956e-01
-4.77741063e-01 6.46526068e-02 9.13406461e-02 -6.87546432e-01
-1.05078900e+00 -6.69023037e-01 -7.68809140e-01 -2.77956218e-01
-1.87051520e-01 -1.69709653e-01 7.39573598e-01 1.21408403e+00
2.62219250e-01 6.56621575e-01 3.14594746e-01 -9.84600186e-01
-2.56050348e-01 -8.95003617e-01 -4.35952336e-01 5.09217381e-01
3.48245472e-01 -9.07685578e-01 -3.04539591e-01 2.90610522e-01] | [8.48510456085205, 0.540360152721405] |
d5259b44-3c42-439f-becf-9ede6c051664 | image-based-table-recognition-data-model-and | 1911.10683 | null | https://arxiv.org/abs/1911.10683v5 | https://arxiv.org/pdf/1911.10683v5.pdf | Image-based table recognition: data, model, and evaluation | Important information that relates to a specific topic in a document is often organized in tabular format to assist readers with information retrieval and comparison, which may be difficult to provide in natural language. However, tabular data in unstructured digital documents, e.g., Portable Document Format (PDF) and images, are difficult to parse into structured machine-readable format, due to complexity and diversity in their structure and style. To facilitate image-based table recognition with deep learning, we develop the largest publicly available table recognition dataset PubTabNet (https://github.com/ibm-aur-nlp/PubTabNet), containing 568k table images with corresponding structured HTML representation. PubTabNet is automatically generated by matching the XML and PDF representations of the scientific articles in PubMed Central Open Access Subset (PMCOA). We also propose a novel attention-based encoder-dual-decoder (EDD) architecture that converts images of tables into HTML code. The model has a structure decoder which reconstructs the table structure and helps the cell decoder to recognize cell content. In addition, we propose a new Tree-Edit-Distance-based Similarity (TEDS) metric for table recognition, which more appropriately captures multi-hop cell misalignment and OCR errors than the pre-established metric. The experiments demonstrate that the EDD model can accurately recognize complex tables solely relying on the image representation, outperforming the state-of-the-art by 9.7% absolute TEDS score. | ['Xu Zhong', 'Antonio Jimeno Yepes', 'Elaheh ShafieiBavani'] | 2019-11-25 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3802_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123660562.pdf | eccv-2020-8 | ['table-recognition'] | ['computer-vision'] | [ 3.96293849e-01 -3.07094634e-01 -2.37515613e-01 -2.81785578e-01
-1.21899569e+00 -8.34679186e-01 1.58007801e-01 7.34017670e-01
-3.07504594e-01 8.38567793e-01 2.88946599e-01 -3.01584393e-01
7.95751810e-02 -7.92453229e-01 -1.16723788e+00 -4.44278389e-01
4.07890528e-01 7.19093263e-01 -2.40444928e-01 1.66925982e-01
3.40834588e-01 3.65148842e-01 -1.28746903e+00 1.01536345e+00
1.25958228e+00 1.32591212e+00 4.95213419e-01 6.02691829e-01
-6.56787097e-01 8.50349903e-01 -7.37089634e-01 -7.50649691e-01
-8.33706111e-02 -3.88945520e-01 -7.09707439e-01 -2.73562312e-01
7.08289266e-01 -2.98977822e-01 -6.26378477e-01 1.24637055e+00
6.39439523e-01 -5.47856271e-01 6.88674748e-01 -8.75886083e-01
-1.20523548e+00 5.52835822e-01 -4.63169336e-01 3.61722261e-02
5.73009610e-01 2.95611769e-02 8.79277766e-01 -8.91720831e-01
9.54041302e-01 1.25471103e+00 4.33006823e-01 3.48966718e-01
-9.59845364e-01 -6.20261908e-01 -2.49824077e-01 3.61195683e-01
-1.49466836e+00 -3.80588531e-01 3.66572171e-01 -5.72680295e-01
9.38508272e-01 4.14397955e-01 4.49420214e-01 1.06772065e+00
7.28647053e-01 7.66251683e-01 8.29691887e-01 -1.47537559e-01
-1.18647264e-02 8.24811682e-02 -1.49848163e-01 8.45323741e-01
7.34086573e-01 -8.03028762e-01 -4.81661826e-01 2.66932189e-01
4.72410709e-01 8.47838819e-02 -3.96238387e-01 -2.47443199e-01
-1.49027324e+00 3.48233700e-01 3.83609086e-01 1.59461796e-01
-1.19864807e-01 -3.26450288e-01 8.10662031e-01 1.62920654e-01
6.78668618e-02 5.75781226e-01 -2.81777233e-01 -2.19219401e-01
-8.00949395e-01 5.60092628e-02 6.58980250e-01 1.47176313e+00
2.67076403e-01 -3.63030463e-01 -5.28834164e-01 8.73040497e-01
8.95535499e-02 6.78444624e-01 5.98866999e-01 -5.54954290e-01
9.79257643e-01 9.13794696e-01 -2.90783793e-01 -1.12605929e+00
-2.23818317e-01 -2.83997774e-01 -1.45026815e+00 -4.08527255e-01
3.03137988e-01 5.16421080e-01 -1.03807044e+00 1.16134727e+00
1.62355881e-02 -6.33363187e-01 5.40762395e-02 5.79057276e-01
1.34966385e+00 7.73434699e-01 -2.90482402e-01 -2.09570512e-01
1.70108747e+00 -9.67211127e-01 -1.15419233e+00 7.08142668e-02
7.34408796e-01 -8.57156575e-01 1.16756368e+00 4.37199026e-01
-1.02751982e+00 -5.50708473e-01 -1.29944921e+00 -8.79475594e-01
-8.96568060e-01 4.23859388e-01 1.50342047e-01 3.71655136e-01
-6.82098269e-01 3.38313967e-01 -5.99962831e-01 -2.76121110e-01
8.89655113e-01 5.31425960e-02 -6.33765697e-01 -4.04198557e-01
-9.39017475e-01 5.70797980e-01 3.89381111e-01 1.24549992e-01
-5.16745090e-01 -7.16952741e-01 -9.30370867e-01 3.02184910e-01
3.57363403e-01 -6.84431851e-01 9.30352926e-01 -3.33464861e-01
-1.03129506e+00 1.20112944e+00 -3.59704524e-01 -3.05621862e-01
4.21289593e-01 1.25874570e-02 -5.76537669e-01 3.28490704e-01
2.96947956e-01 6.98968947e-01 3.69909406e-01 -8.65204573e-01
-2.88103044e-01 -4.98764336e-01 -2.70095229e-01 1.39245987e-01
-3.73059213e-01 -1.25174478e-01 -1.12617397e+00 -9.50773537e-01
3.07214796e-03 -5.61065018e-01 2.43414789e-01 5.50869524e-01
-9.13818836e-01 1.55746415e-01 4.25087631e-01 -1.02578211e+00
1.47650695e+00 -2.04676080e+00 1.49315568e-02 -5.00380248e-02
4.84307379e-01 2.05869094e-01 -2.09799409e-02 4.38315213e-01
3.30466740e-02 4.06201184e-01 -5.12966990e-01 -1.19512960e-01
-3.26903723e-02 -5.40483445e-02 -1.95520550e-01 1.39686763e-01
5.91702983e-02 1.07598829e+00 -6.58091664e-01 -7.80510128e-01
-1.99257895e-01 3.38818580e-01 -3.80067050e-01 1.33817062e-01
-1.79792151e-01 1.57545935e-02 -3.00100774e-01 1.02757752e+00
8.72940540e-01 -7.28578150e-01 2.74619371e-01 -5.57403743e-01
3.22472490e-02 5.61653674e-01 -8.89477670e-01 1.85743785e+00
-2.08158836e-01 7.16376722e-01 -1.73820674e-01 -7.13780284e-01
8.18100274e-01 2.66459095e-03 2.67264932e-01 -1.04838872e+00
8.83902144e-03 4.52135861e-01 -1.02126352e-01 -3.31484020e-01
4.59520340e-01 5.97444654e-01 -1.63777843e-01 4.91268225e-02
-7.27742985e-02 6.56807795e-02 5.12220740e-01 3.42133015e-01
9.77542162e-01 -2.03029543e-01 2.73389548e-01 -5.90795204e-02
7.45134234e-01 5.68456575e-02 3.91440213e-01 7.40010202e-01
1.59636557e-01 9.37272191e-01 8.50862384e-01 -3.11269760e-01
-1.20646036e+00 -9.59273398e-01 -4.49006915e-01 3.00209761e-01
6.54765144e-02 -6.25750601e-01 -8.59200597e-01 -5.76447427e-01
1.47800490e-01 2.89227009e-01 -6.51776135e-01 -1.85180470e-01
-3.62147272e-01 -5.41099072e-01 6.75184548e-01 4.82955396e-01
6.85126722e-01 -8.56981993e-01 -9.23208669e-02 1.91659167e-01
-2.30430529e-01 -1.20849633e+00 -1.03975308e+00 2.75931478e-01
-6.39196575e-01 -1.11565149e+00 -8.85935783e-01 -9.42873061e-01
8.79126668e-01 5.02621010e-02 9.51457500e-01 -1.87065721e-01
-5.07794023e-01 -4.55384031e-02 -2.53237218e-01 -3.27529490e-01
-5.26563048e-01 1.74246475e-01 -2.71674067e-01 -8.67298469e-02
2.92070001e-01 5.89949526e-02 -6.17934287e-01 1.01006187e-01
-1.20702648e+00 4.53019083e-01 8.47534359e-01 1.01484883e+00
1.27332580e+00 -5.09131141e-02 1.23463973e-01 -1.05613530e+00
6.30389154e-01 -2.80679613e-01 -7.19426095e-01 7.33284175e-01
-8.60979974e-01 4.64084387e-01 1.01378465e+00 -1.78046171e-02
-7.70547807e-01 -4.19221409e-02 -2.40874052e-01 -3.71276170e-01
7.21747875e-02 8.74192595e-01 -6.67194068e-01 1.72131240e-01
1.62791029e-01 7.04578400e-01 5.30214980e-02 -7.33743489e-01
-6.92696683e-03 1.05100203e+00 8.35583866e-01 -2.88813472e-01
5.36351204e-01 1.79002598e-01 -1.13440491e-01 -3.16204697e-01
-6.85099363e-01 -1.75671116e-01 -7.24930286e-01 6.79975897e-02
9.82163072e-01 -9.56890821e-01 -7.98882961e-01 7.40714550e-01
-1.21399677e+00 1.73799116e-02 1.70134082e-01 2.41324812e-01
-4.06353563e-01 4.81570333e-01 -8.59189510e-01 -8.84269774e-02
-6.23039007e-01 -1.31390929e+00 1.20044768e+00 3.15638892e-02
-1.92692820e-02 -5.25990486e-01 -2.14230701e-01 5.72357059e-01
1.03441700e-01 1.56756401e-01 1.48972511e+00 -6.11607492e-01
-7.81052113e-01 -1.69178575e-01 -5.58565319e-01 2.20323727e-02
2.47148603e-01 1.23846665e-01 -6.74462557e-01 -7.24800304e-02
-4.00025487e-01 -3.26631278e-01 1.00897372e+00 7.66033009e-02
2.13401437e+00 -5.54148853e-01 -4.53320861e-01 9.41306889e-01
1.38172984e+00 7.07165301e-01 1.06432056e+00 5.48968852e-01
1.03713417e+00 3.47038507e-01 4.54739630e-01 2.75812507e-01
5.48061550e-01 7.61974633e-01 1.42809926e-02 -9.34645161e-02
-1.64642885e-01 -5.92471361e-01 1.40126392e-01 1.08617830e+00
6.62546277e-01 -7.04136491e-01 -9.52761769e-01 1.57975972e-01
-1.47143233e+00 -9.17619109e-01 -1.26936167e-01 2.19450712e+00
1.29374850e+00 2.95477867e-01 -4.16003823e-01 7.97402859e-02
7.46892393e-01 1.13783628e-02 -8.52302790e-01 -4.32313144e-01
-5.97538054e-01 6.56086951e-02 6.43507421e-01 1.65882051e-01
-1.04340613e+00 6.02909684e-01 5.53858280e+00 1.21031201e+00
-1.15630007e+00 -2.76133329e-01 1.17286301e+00 -5.97916655e-02
-4.46293145e-01 -6.91494584e-01 -8.57091188e-01 9.05103624e-01
9.19134617e-01 -3.31339389e-01 3.36458504e-01 6.72425151e-01
-8.89909640e-02 9.74326506e-02 -1.30129266e+00 1.48253095e+00
1.84849977e-01 -2.01812482e+00 6.43798411e-01 1.00418851e-01
3.63179922e-01 -3.18598717e-01 2.96452850e-01 1.38727099e-01
-3.38213652e-01 -1.31031489e+00 8.36830616e-01 6.16151512e-01
1.72856390e+00 -4.70134616e-01 7.49437034e-01 -2.20061302e-01
-1.20421875e+00 1.86761588e-01 -3.66364211e-01 6.32709742e-01
-3.44217926e-01 7.44493365e-01 -6.39736295e-01 5.18989623e-01
9.06439602e-01 9.85728741e-01 -7.90144920e-01 1.13824737e+00
2.92351365e-01 2.21862137e-01 -5.19577749e-02 -2.13691011e-01
1.47248870e-02 -1.43326074e-01 1.56774610e-01 1.38365507e+00
4.96941864e-01 -1.64110139e-01 -3.11274856e-01 9.48399842e-01
-7.51813948e-01 2.54733652e-01 -6.08467519e-01 -5.81759214e-01
7.62485504e-01 1.15074754e+00 -9.74380791e-01 -4.97578949e-01
-4.97572362e-01 1.20473242e+00 3.06247681e-01 7.82721266e-02
-6.96117520e-01 -8.93809199e-01 5.85343778e-01 2.12611090e-02
5.67623556e-01 1.16611578e-01 -5.70285141e-01 -1.25408912e+00
6.07284546e-01 -1.38952541e+00 4.72092569e-01 -1.06929266e+00
-1.13505161e+00 7.98488557e-01 -3.65046263e-01 -1.53487825e+00
1.90844730e-01 -9.67775226e-01 9.85694453e-02 7.77991235e-01
-1.17700827e+00 -6.93686783e-01 -3.75707626e-01 3.79217505e-01
5.24722636e-01 -3.97607207e-01 7.89862812e-01 7.23942995e-01
-8.71519029e-01 1.18759084e+00 9.58644092e-01 4.40541238e-01
9.17191148e-01 -1.27310741e+00 5.62393367e-01 4.89723861e-01
-1.75263528e-02 7.55593538e-01 2.35875323e-01 -7.11498320e-01
-1.98161006e+00 -1.32742906e+00 1.01754844e+00 -3.52665782e-01
3.47319305e-01 -8.37824821e-01 -1.07071841e+00 2.61601716e-01
1.16158426e-01 -2.33176082e-01 7.60199368e-01 -5.91611862e-01
-6.98120117e-01 -5.12331963e-01 -1.07352829e+00 6.98491216e-01
9.96974945e-01 -8.65240753e-01 -1.71317235e-01 5.39233744e-01
8.51246178e-01 -8.28687787e-01 -1.06830597e+00 2.20011443e-01
8.10294628e-01 -4.41040248e-01 8.26224387e-01 -3.09411198e-01
9.87179935e-01 -4.38926697e-01 -1.19416334e-01 -8.91116500e-01
-2.75743455e-01 -3.41303021e-01 -2.17506066e-01 1.17308009e+00
6.66649044e-01 -3.69329154e-01 7.07178593e-01 3.01305652e-01
-3.61751378e-01 -9.88606572e-01 -9.14509296e-01 -7.01469243e-01
2.21025641e-03 -2.09284183e-02 8.36552024e-01 8.31130743e-01
4.70031090e-02 2.31711522e-01 -6.05425313e-02 -2.14583546e-01
2.21661389e-01 2.47302338e-01 4.24048871e-01 -9.54542756e-01
1.00684948e-01 -6.70407891e-01 -3.84859174e-01 -1.22360206e+00
-2.05883831e-01 -1.24541736e+00 -6.61290884e-02 -1.83905029e+00
5.89156747e-01 -1.32526323e-01 -3.15938860e-01 4.37522292e-01
-2.17468962e-01 1.25617847e-01 2.55404469e-02 3.55137408e-01
-8.30937088e-01 5.62970400e-01 1.49271870e+00 -7.85472095e-01
2.52427727e-01 -7.64487565e-01 -9.37415004e-01 8.40238705e-02
4.94580925e-01 -6.19760394e-01 -1.81712061e-01 -5.88527203e-01
3.70257854e-01 2.04280660e-01 -1.34492785e-01 -1.16405857e+00
3.48837972e-01 1.87659506e-02 8.84828448e-01 -1.07415557e+00
-1.01338491e-01 -6.63967848e-01 2.02865854e-01 5.11998296e-01
-7.66519725e-01 3.97285014e-01 4.00252193e-01 4.55852389e-01
-5.03370583e-01 -7.11944774e-02 4.14345503e-01 -6.81625903e-02
-3.40450019e-01 5.46470463e-01 -4.40745920e-01 2.08394125e-01
5.96978188e-01 -3.39136720e-01 -7.83919930e-01 -2.29350373e-01
-9.32984501e-02 3.23502779e-01 5.95184743e-01 4.48784411e-01
8.90078664e-01 -1.50476599e+00 -5.67543149e-01 1.96636006e-01
6.02455795e-01 8.07359889e-02 3.21042240e-01 7.10172653e-01
-1.14141047e+00 9.32904303e-01 -5.00867784e-01 -3.92905861e-01
-1.24600124e+00 7.40691483e-01 1.07244305e-01 -4.14727658e-01
-4.91329223e-01 7.52258480e-01 3.49064410e-01 -3.85953724e-01
4.62808162e-01 -6.56808317e-01 -1.79288402e-01 4.43878099e-02
7.74695277e-01 9.24475864e-03 5.82773268e-01 -2.87638694e-01
-3.80846858e-01 4.16005611e-01 -6.41183972e-01 2.80292541e-01
1.05342972e+00 2.23134197e-02 -3.19616377e-01 3.43053669e-01
1.61036038e+00 1.03664286e-01 -7.60552526e-01 1.69105809e-02
-1.27780393e-01 -3.65008324e-01 -2.18377233e-01 -1.07796264e+00
-1.01694238e+00 9.38433468e-01 5.68918586e-01 -2.72822767e-01
1.17551827e+00 -2.87012398e-01 1.10685432e+00 7.84048438e-01
7.14567527e-02 -1.05044973e+00 -6.84443638e-02 5.03712595e-01
1.11032796e+00 -1.17779791e+00 5.14034294e-02 -4.11688179e-01
-5.92288733e-01 1.31002009e+00 5.89010656e-01 7.21556246e-01
3.34630162e-01 3.72078270e-01 -7.30982497e-02 -1.67650998e-01
-7.82824576e-01 1.52441159e-01 5.38400710e-01 5.61986268e-01
6.74897730e-01 -2.08238095e-01 -2.49752894e-01 7.26405382e-01
-2.38884300e-01 -2.82375794e-03 4.31737125e-01 8.60114574e-01
-1.57996312e-01 -9.60704148e-01 -3.68486315e-01 1.04905367e+00
-5.85606217e-01 -5.02359629e-01 -6.09788477e-01 4.05354679e-01
-3.46539766e-02 5.31787872e-01 3.29370141e-01 -4.00960326e-01
1.86071262e-01 1.37792788e-02 3.61608565e-01 -4.81320858e-01
-3.30270946e-01 -6.93952367e-02 1.56788230e-02 -4.50669289e-01
1.60753548e-01 -5.23861825e-01 -1.12128341e+00 -5.09934902e-01
2.74808288e-01 1.04301304e-01 3.18649679e-01 3.91142726e-01
8.36883426e-01 6.22755766e-01 3.09434474e-01 -6.12403527e-02
-4.32187989e-02 -7.35367596e-01 -4.72070932e-01 3.89655173e-01
3.40424597e-01 -3.60058665e-01 1.75435603e-01 3.66401881e-01] | [11.700796127319336, 2.9828097820281982] |
0e64babe-5679-42da-8a19-10e7ea83775b | agronav-autonomous-navigation-framework-for | 2304.04333 | null | https://arxiv.org/abs/2304.04333v1 | https://arxiv.org/pdf/2304.04333v1.pdf | Agronav: Autonomous Navigation Framework for Agricultural Robots and Vehicles using Semantic Segmentation and Semantic Line Detection | The successful implementation of vision-based navigation in agricultural fields hinges upon two critical components: 1) the accurate identification of key components within the scene, and 2) the identification of lanes through the detection of boundary lines that separate the crops from the traversable ground. We propose Agronav, an end-to-end vision-based autonomous navigation framework, which outputs the centerline from the input image by sequentially processing it through semantic segmentation and semantic line detection models. We also present Agroscapes, a pixel-level annotated dataset collected across six different crops, captured from varying heights and angles. This ensures that the framework trained on Agroscapes is generalizable across both ground and aerial robotic platforms. Codes, models and dataset will be released at \href{https://github.com/shivamkumarpanda/agronav}{github.com/shivamkumarpanda/agronav}. | ['M. Khalid Jawed', 'Yongkyu Lee', 'Shivam K Panda'] | 2023-04-10 | null | null | null | null | ['autonomous-navigation', 'line-detection'] | ['computer-vision', 'computer-vision'] | [ 1.07329212e-01 -1.15476221e-01 4.75232117e-02 -3.89684647e-01
-2.82560885e-01 -1.41126883e+00 2.74913907e-01 4.16996390e-01
9.05141458e-02 1.06777959e-01 -3.87833774e-01 -8.86603475e-01
-2.42780060e-01 -1.01015317e+00 -7.46504486e-01 -1.84146270e-01
-2.46650025e-01 2.55530119e-01 5.65503120e-01 -5.65975010e-01
1.96718171e-01 6.07492566e-01 -1.49987197e+00 -1.76728919e-01
9.27782595e-01 1.11300457e+00 6.81051672e-01 1.14934039e+00
1.68465912e-01 1.38888106e-01 3.11525881e-01 8.60351697e-02
6.04449868e-01 -1.18582629e-01 -6.18463218e-01 1.78897485e-01
6.92843080e-01 -6.00593030e-01 -7.54442587e-02 1.26871681e+00
8.69182423e-02 -1.41509414e-01 6.04175568e-01 -1.37293255e+00
-3.70454848e-01 2.66808271e-01 -7.67883718e-01 -2.21478984e-01
2.50061393e-01 1.42253861e-01 9.54378664e-01 -8.38306308e-01
7.08382785e-01 7.38152206e-01 1.07731795e+00 -2.44178176e-01
-9.15816486e-01 -3.18890899e-01 2.18699723e-01 -2.20237896e-01
-1.28816748e+00 -3.65233809e-01 2.76258171e-01 -5.84831297e-01
3.30010265e-01 6.29531294e-02 5.76223671e-01 5.22943497e-01
1.64966106e-01 6.98081136e-01 6.43498719e-01 -4.17348117e-01
3.20410252e-01 -5.44768095e-01 3.67983013e-01 1.03950131e+00
5.70410311e-01 2.15322971e-01 -4.82209623e-02 -3.55466008e-02
1.10130441e+00 8.42659734e-03 -4.28423226e-01 -1.11881590e+00
-1.01561403e+00 7.64495075e-01 6.52890682e-01 -2.94464111e-01
-5.75638950e-01 9.33082253e-02 2.38908857e-01 -1.44584859e-02
-7.49678314e-02 2.49907553e-01 -7.63647854e-01 2.10202903e-01
-7.89159596e-01 3.44451100e-01 6.74678504e-01 1.69738007e+00
8.98567498e-01 -5.71024343e-02 3.32713127e-01 5.43132782e-01
4.77074027e-01 1.02660441e+00 -2.99377799e-01 -1.31058776e+00
2.34642774e-01 5.26345551e-01 6.10308409e-01 -1.17116356e+00
-5.51232517e-01 -4.40225005e-02 -2.90963799e-01 4.96065766e-01
5.89676678e-01 -5.77429831e-01 -1.18561912e+00 1.19381607e+00
3.74815822e-01 -2.30330557e-01 2.63752043e-01 9.83535290e-01
8.11886251e-01 6.91545963e-01 5.91741875e-02 4.37827706e-01
1.53099465e+00 -9.73939717e-01 -2.95526326e-01 -6.42964900e-01
7.31776237e-01 -8.43954086e-01 8.97492409e-01 2.64836997e-01
-6.95942521e-01 -4.70118463e-01 -1.25296104e+00 -7.73160309e-02
-6.36891723e-01 7.27026403e-01 7.48299778e-01 2.17055410e-01
-1.22428954e+00 2.57408470e-01 -8.83692205e-01 -7.89546669e-01
1.99870944e-01 -4.27848310e-04 -4.07189041e-01 -2.40431175e-01
-7.54996717e-01 5.89801192e-01 3.72916907e-01 5.39494514e-01
-1.01346779e+00 -2.39034817e-01 -1.20694315e+00 -1.77056342e-01
5.09122491e-01 -8.48635957e-02 1.38430309e+00 -8.05202186e-01
-1.22190881e+00 8.25601816e-01 -3.38951722e-02 -3.55553865e-01
4.65809405e-01 -6.88816309e-01 4.59760427e-02 2.15554386e-01
4.60424751e-01 7.60127068e-01 3.32546443e-01 -1.52383411e+00
-1.31088972e+00 -5.80285668e-01 5.31477444e-02 3.54473293e-01
3.10590088e-01 -5.06372213e-01 -7.02609181e-01 -2.94821471e-01
4.91360039e-01 -1.16199780e+00 -4.48710948e-01 3.37710887e-01
-5.76042593e-01 4.68624204e-01 1.04682803e+00 -7.93355227e-01
6.14594340e-01 -2.02596927e+00 -3.61761808e-01 2.15488657e-01
-2.00004220e-01 3.57199371e-01 -3.40109557e-01 5.93389094e-01
2.77375340e-01 -2.80010253e-01 -3.23573858e-01 2.85263330e-01
-3.12948763e-01 -1.41990362e-02 -3.07600349e-01 3.63280147e-01
-5.67960478e-02 8.43178451e-01 -1.03545702e+00 -9.72226188e-02
6.28003955e-01 1.45277143e-01 -3.50629762e-02 9.78742614e-02
-2.87538886e-01 1.82319283e-01 -7.15677679e-01 1.06924248e+00
1.19104910e+00 1.98767662e-01 2.28593186e-01 -1.64287731e-01
-7.74437547e-01 -3.80509377e-01 -1.15859735e+00 1.73362303e+00
-1.63142934e-01 6.46536291e-01 4.83075738e-01 -6.02992773e-01
1.11937428e+00 -9.55301300e-02 2.40933269e-01 -4.34847891e-01
8.90590474e-02 2.56653637e-01 -5.78974128e-01 -3.21519464e-01
7.85292685e-01 1.00371134e+00 -1.77717641e-01 -1.82826504e-01
-1.50248885e-01 -5.71813107e-01 2.16289341e-01 2.17441153e-02
7.99703360e-01 8.03841889e-01 3.26109916e-01 -5.93148112e-01
1.40369400e-01 1.04526472e+00 3.86621892e-01 8.44619751e-01
-3.97879094e-01 6.94382727e-01 9.58701447e-02 -4.41900462e-01
-9.74509418e-01 -1.17728162e+00 -9.04767811e-02 1.01719940e+00
8.01508427e-01 -1.25003308e-01 -7.40663350e-01 -5.45309067e-01
2.35846326e-01 5.50909579e-01 -4.70651567e-01 3.84858638e-01
-8.77439454e-02 -2.12967828e-01 3.67428780e-01 6.00862861e-01
6.94998682e-01 -7.89132416e-01 -1.39101684e+00 1.31061390e-01
-1.50347650e-01 -1.31608760e+00 -6.54787943e-02 5.36795974e-01
-7.62103677e-01 -1.31125450e+00 -5.68523169e-01 -8.49257529e-01
6.95805788e-01 1.01774752e+00 7.46629953e-01 -1.79872707e-01
-4.42451090e-01 4.81694579e-01 -6.40974164e-01 -7.66677439e-01
-7.34229833e-02 8.77456069e-02 -3.94350082e-01 -6.98806584e-01
2.49365434e-01 -8.92043859e-02 -8.46985996e-01 4.81263161e-01
-4.63964641e-01 2.48618454e-01 3.66596669e-01 3.62920165e-01
7.37090588e-01 -1.79409400e-01 -1.68907642e-01 -5.38727343e-01
-1.38548672e-01 -3.80748689e-01 -1.50547194e+00 3.21989954e-01
-3.15271139e-01 -4.37251002e-01 4.17844862e-01 2.46376634e-01
-6.34901345e-01 7.82761157e-01 7.04468563e-02 -7.76677812e-03
-7.68270552e-01 6.41793191e-01 -7.33789951e-02 -4.79494082e-03
6.87459469e-01 -2.02056020e-02 -6.65437654e-02 -1.47628009e-01
7.49149323e-01 6.37587011e-01 9.64562297e-01 -7.46849552e-03
7.53760278e-01 6.48955166e-01 -4.37884144e-02 -1.18115258e+00
-5.91668010e-01 -7.14315593e-01 -1.07580233e+00 -3.32250655e-01
9.21661437e-01 -1.09855068e+00 -3.22476506e-01 7.06778705e-01
-1.01439738e+00 -8.99748802e-01 9.69627574e-02 4.55994278e-01
-5.72245777e-01 2.45714381e-01 -4.10356581e-01 -7.14205563e-01
-5.62705696e-01 -9.93971586e-01 1.21475887e+00 6.92850292e-01
-1.71652734e-01 -8.50047350e-01 -1.35682628e-01 2.71633454e-02
9.30053443e-02 5.48774540e-01 6.07727349e-01 -1.82784125e-01
-5.97373903e-01 -4.51095670e-01 -4.61934865e-01 2.89579444e-02
2.94381380e-01 6.24667645e-01 -8.55194271e-01 -2.01639920e-01
-7.93529868e-01 -2.27801010e-01 5.20853281e-01 9.72412884e-01
5.70596039e-01 2.47229323e-01 -5.66551924e-01 7.58304656e-01
1.78925622e+00 3.74899745e-01 3.40935946e-01 6.64729118e-01
6.53803468e-01 7.53045619e-01 1.31285381e+00 4.30076808e-01
6.06158614e-01 4.32884246e-01 1.13294363e+00 -5.46302676e-01
1.69171169e-01 -3.52455020e-01 1.23718858e-01 -5.33676334e-02
3.18485945e-02 -4.64607149e-01 -1.13369477e+00 8.56000662e-01
-2.14148498e+00 -5.47015011e-01 -6.82681561e-01 2.19449830e+00
-4.06786688e-02 -1.59024104e-01 -1.28404900e-01 -2.71412671e-01
5.79064488e-01 2.54690051e-02 -5.97746849e-01 -3.55670005e-01
-7.87541270e-03 -2.32083261e-01 1.46364248e+00 6.02829516e-01
-1.63545823e+00 1.63851094e+00 5.40990639e+00 1.93616420e-01
-1.14178789e+00 -3.26480210e-01 2.27613673e-01 7.35134721e-01
2.50498354e-01 2.67086029e-01 -8.20606291e-01 -3.44110690e-02
4.44369465e-01 3.74463260e-01 9.28310305e-02 1.11835206e+00
5.20140588e-01 -6.95769489e-01 -5.03458381e-01 4.66688395e-01
-4.48789597e-01 -1.10893047e+00 -3.05493891e-01 -1.71536833e-01
6.44103765e-01 4.67331439e-01 -2.09532946e-01 -2.49740407e-01
9.71374214e-01 -5.62725067e-01 1.09402084e+00 1.09875113e-01
7.50738978e-01 -4.35282737e-01 5.91806531e-01 2.04102159e-01
-1.42027545e+00 -6.86676353e-02 -4.80536103e-01 1.35607958e-01
7.04824105e-02 3.25121194e-01 -8.81519437e-01 8.59338462e-01
8.58818114e-01 6.79979801e-01 -5.08311272e-01 1.14318907e+00
-5.24573624e-01 4.68996465e-01 -3.28936666e-01 3.10206562e-01
5.83330989e-01 -4.66040879e-01 2.63424695e-01 1.11733770e+00
7.20418215e-01 9.54360589e-02 4.96585667e-01 4.41138715e-01
5.55110514e-01 -3.61701697e-02 -1.09727800e+00 1.35429010e-01
7.16566622e-01 1.53610575e+00 -1.21735156e+00 2.00778827e-01
-1.94853947e-01 1.09835970e+00 -1.28273994e-01 3.69435728e-01
-7.16198385e-01 -8.49061966e-01 7.05465078e-01 -7.61484727e-02
6.67830765e-01 -8.21390271e-01 -3.80112529e-01 -5.69666266e-01
-2.23439395e-01 -4.24092501e-01 1.50149971e-01 -1.14726055e+00
-4.06960666e-01 4.15075898e-01 -2.24218652e-01 -1.19605565e+00
-1.68198999e-02 -9.67507660e-01 -4.41620082e-01 9.08585489e-01
-1.49330842e+00 -1.65925801e+00 -1.16742134e+00 2.05167994e-01
7.00775802e-01 1.96998522e-01 1.18126464e+00 -2.69115597e-01
-2.32328489e-01 3.06919925e-02 4.54839796e-01 2.54975587e-01
4.52617139e-01 -1.12459302e+00 8.91185522e-01 1.36006677e+00
-1.40976980e-01 4.62325588e-02 8.07433128e-01 -7.89306939e-01
-1.47787428e+00 -1.39335430e+00 2.83314377e-01 -3.78519177e-01
4.66321111e-01 -2.79694319e-01 -4.77027923e-01 8.61608148e-01
1.53873876e-01 -1.24509022e-01 3.99020128e-02 -3.72563958e-01
1.50325233e-02 2.69277282e-02 -9.53127861e-01 7.82772660e-01
1.16683662e+00 -1.88667759e-01 2.36098059e-02 1.76515847e-01
4.88331795e-01 -7.07858443e-01 -5.02172410e-01 5.91530859e-01
8.26618314e-01 -9.61662889e-01 8.45938027e-01 -1.10314913e-05
4.06654388e-01 -7.37227678e-01 -4.17648166e-01 -1.39977896e+00
-4.68652725e-01 -2.98347592e-01 4.96487290e-01 9.90498543e-01
6.11716866e-01 -6.08869374e-01 7.26664186e-01 1.93834707e-01
-3.24309289e-01 -1.72021970e-01 -1.57300636e-01 -6.35444880e-01
-1.66597083e-01 -3.47541481e-01 3.56575608e-01 8.06454659e-01
-2.91503072e-01 -1.27282981e-02 -6.43343180e-02 1.18314588e+00
5.71730554e-01 2.58045107e-01 8.35873961e-01 -1.21383953e+00
3.25869322e-01 -9.82765406e-02 -4.42283660e-01 -1.12763214e+00
-3.93188626e-01 -5.29706776e-01 6.07811391e-01 -2.09662938e+00
-5.33671260e-01 -7.64082432e-01 4.39799398e-01 7.89308667e-01
1.91219106e-01 7.90269524e-02 1.94742978e-02 2.30841786e-01
-2.05755860e-01 1.51224313e-02 7.39073455e-01 9.62844044e-02
-4.56108242e-01 1.67206883e-01 -5.11838317e-01 7.27888644e-01
9.47293162e-01 4.89180125e-02 -6.19934976e-01 -9.22735631e-01
-7.12452531e-02 1.47989050e-01 5.22641003e-01 -1.14393997e+00
1.77036628e-01 -3.29978585e-01 4.53679919e-01 -9.47925866e-01
4.13821414e-02 -7.76029885e-01 -2.97688581e-02 6.44269705e-01
-1.08525209e-01 3.15496683e-01 5.57940960e-01 4.05906886e-01
2.44128808e-01 -3.44245225e-01 5.97190320e-01 -1.89535618e-01
-1.54729748e+00 2.91464835e-01 -5.76529920e-01 -3.75498712e-01
1.24634933e+00 -3.99438888e-01 -5.60005903e-01 -2.36597523e-01
-6.03553772e-01 7.68724263e-01 8.25522304e-01 5.94418764e-01
6.22722864e-01 -7.45917976e-01 -5.42930424e-01 4.06600803e-01
5.01758337e-01 3.15824151e-01 9.93798971e-02 5.70594907e-01
-1.53446436e+00 4.55404967e-01 -5.46931446e-01 -8.53178680e-01
-1.33949792e+00 2.45476082e-01 2.77477890e-01 4.33814287e-01
-6.08765960e-01 6.94271088e-01 2.07583070e-01 -7.12611556e-01
-1.01435825e-01 -5.98786891e-01 -6.59236237e-02 -1.99020252e-01
3.03294629e-01 3.09600681e-01 4.45868745e-02 -7.22689927e-01
-4.07644808e-01 7.69116998e-01 3.23778868e-01 -1.17924824e-01
1.12139237e+00 -5.04105866e-01 3.04449856e-01 1.44052655e-01
6.38010263e-01 -1.83250621e-01 -1.61752141e+00 1.28898114e-01
1.43245831e-01 -3.62010181e-01 1.98406339e-01 -8.50740731e-01
-7.49126852e-01 7.24224031e-01 1.10285878e+00 2.19495028e-01
9.35748219e-01 -3.81222427e-01 4.70809698e-01 3.61562908e-01
8.37797463e-01 -1.00295329e+00 -5.41931689e-01 8.87737691e-01
6.39489710e-01 -1.28706956e+00 -2.52697945e-01 -8.78507912e-01
-7.61001527e-01 1.33783722e+00 5.14495492e-01 -6.45366907e-02
7.18808770e-01 3.94382864e-01 7.37964153e-01 -2.94481426e-01
1.32307829e-02 -5.42156875e-01 -2.41019949e-01 1.06509554e+00
1.82678387e-01 3.93925279e-01 5.33919446e-02 9.99427736e-02
-4.20167707e-02 8.87909532e-02 5.79960346e-01 1.45022058e+00
-7.60575056e-01 -5.88750124e-01 -4.29710120e-01 9.87417176e-02
2.48377070e-01 -4.34258804e-02 -4.26770866e-01 7.49647081e-01
1.78430781e-01 1.01107848e+00 5.64451404e-02 -2.02926859e-01
7.05277503e-01 -4.09553677e-01 1.20138794e-01 -5.01085579e-01
3.67515497e-02 3.99836488e-02 1.45673007e-01 -5.79596996e-01
1.33755177e-01 -7.18879044e-01 -1.43214595e+00 2.68472912e-04
-3.41420978e-01 -5.64470189e-03 1.00830889e+00 3.60728294e-01
5.78414738e-01 6.11165026e-03 6.03937745e-01 -1.11540627e+00
-2.17110083e-01 -5.58290064e-01 -7.04988539e-01 -2.75643408e-01
3.20732057e-01 -4.80191648e-01 1.26821682e-01 3.02283168e-01] | [9.031684875488281, -1.5762301683425903] |
1d30392b-7cb9-422c-a892-a48536b1235f | a-regional-news-corpora-for-contextualized | null | null | https://aclanthology.org/L16-1531 | https://aclanthology.org/L16-1531.pdf | A Regional News Corpora for Contextualized Entity Discovery and Linking | This paper presents a German corpus for Named Entity Linking (NEL) and Knowledge Base Population (KBP) tasks. We describe the annotation guideline, the annotation process, NIL clustering techniques and conversion to popular NEL formats such as NIF and TAC that have been used to construct this corpus based on news transcripts from the German regional broadcaster RBB (Rundfunk Berlin Brandenburg). Since creating such language resources requires significant effort, the paper also discusses how to derive additional evaluation resources for tasks like named entity contextualization or ontology enrichment by exploiting the links between named entities from the annotated corpus. The paper concludes with an evaluation that shows how several well-known NEL tools perform on the corpus, a discussion of the evaluation results, and with suggestions on how to keep evaluation corpora and datasets up to date. | ['Adrian Bra{\\c{s}}oveanu', 'Lyndon J.B. Nixon', 'Albert Weichselbraun', 'Arno Scharl'] | 2016-05-01 | a-regional-news-corpora-for-contextualized-1 | https://aclanthology.org/L16-1531 | https://aclanthology.org/L16-1531.pdf | lrec-2016-5 | ['knowledge-base-population'] | ['natural-language-processing'] | [-3.94123048e-01 8.39745164e-01 -1.80927023e-01 -6.28591180e-01
-1.03507113e+00 -7.29987144e-01 6.19273424e-01 6.93113863e-01
-9.85859811e-01 1.36922610e+00 9.78451252e-01 -5.25178611e-02
-4.76965040e-01 -6.34771407e-01 -3.28681797e-01 1.27304150e-02
-6.36452511e-02 1.15908360e+00 5.79829335e-01 -3.59989792e-01
1.18464962e-01 4.05678809e-01 -1.14894688e+00 4.86836284e-01
5.91644228e-01 5.16235471e-01 1.04454055e-01 3.49027455e-01
-6.07213080e-01 9.19176340e-01 -1.08194613e+00 -8.75180721e-01
-2.08909184e-01 -1.19690649e-01 -1.72670674e+00 -3.54273140e-01
1.37202129e-01 3.87476236e-01 1.51126282e-02 7.87619829e-01
7.78892398e-01 5.22156000e-01 3.29041719e-01 -8.71158659e-01
-2.82789856e-01 1.53376830e+00 1.98995098e-01 1.97118625e-01
4.54008311e-01 -7.81577826e-01 9.98203337e-01 -5.55823505e-01
1.62656736e+00 1.03666210e+00 1.00122130e+00 5.37225723e-01
-7.30814874e-01 -4.36974436e-01 -2.35879928e-01 2.24996760e-01
-1.60240960e+00 -6.99324429e-01 5.19500747e-02 -2.81008065e-01
1.41985750e+00 2.81072766e-01 3.78803313e-01 8.36835802e-01
-6.52760088e-01 4.15749103e-01 8.12339842e-01 -9.32425380e-01
-8.05068016e-02 6.05456233e-01 5.45676768e-01 3.56596947e-01
5.75462878e-01 -5.28978586e-01 -3.81600678e-01 -3.46185714e-01
3.59827667e-01 -1.36398005e+00 -2.03698978e-01 -1.07254505e-01
-1.09210014e+00 5.54974914e-01 4.76034284e-02 7.94221342e-01
-3.98527384e-01 -1.27679169e-01 8.91959190e-01 8.09229724e-03
4.85386968e-01 8.38014007e-01 -9.39186692e-01 -2.62839764e-01
-9.38757300e-01 3.32532346e-01 1.42287934e+00 1.47890460e+00
5.79924107e-01 -3.09632838e-01 -7.17222467e-02 1.23647511e+00
4.20849979e-01 3.13389152e-02 2.86232412e-01 -1.15226030e+00
6.64730847e-01 5.27250648e-01 5.86874187e-01 -8.60018849e-01
-8.25678885e-01 -1.26465067e-01 4.31977101e-02 -4.96741176e-01
3.96485984e-01 -5.31854987e-01 -6.07713461e-01 1.52062345e+00
5.58938563e-01 -3.09526592e-01 5.31124115e-01 3.16385895e-01
1.52985156e+00 3.98409098e-01 3.38141114e-01 -3.77700955e-01
1.74301040e+00 -7.18119562e-01 -1.32023144e+00 1.40693575e-01
1.17352724e+00 -9.32466626e-01 3.23792815e-01 -1.49929225e-01
-1.16766691e+00 -2.65387088e-01 -6.15523577e-01 -3.57143104e-01
-8.42692435e-01 7.90876672e-02 5.31176209e-01 8.80803704e-01
-9.56040263e-01 2.57504910e-01 -9.01816189e-01 -1.13040078e+00
4.31904979e-02 1.68095618e-01 -6.16612434e-01 1.93611637e-01
-1.68537927e+00 1.31030166e+00 1.50168693e+00 -2.62019068e-01
-1.35852307e-01 -5.94417691e-01 -8.87807965e-01 -1.11447155e-01
5.86392045e-01 -3.81794453e-01 1.31825876e+00 -3.97056729e-01
-1.07934546e+00 1.18728590e+00 1.93033159e-01 -6.55677915e-01
1.75944358e-01 -1.00074887e-01 -1.34743249e+00 1.43004358e-01
5.94514906e-01 7.65032947e-01 -4.43111807e-01 -1.06098855e+00
-1.09480989e+00 -9.46240649e-02 9.62870866e-02 2.34927252e-01
-1.04297049e-01 9.22389627e-01 -6.24718964e-01 -7.01404393e-01
-1.90711647e-01 -5.29302537e-01 1.52466977e-02 -1.07068574e+00
-4.26112771e-01 -3.92760247e-01 3.91708851e-01 -1.15097642e+00
1.80477762e+00 -1.84272778e+00 -3.64829510e-01 3.74732584e-01
-2.17276722e-01 3.85221466e-02 1.17994063e-01 1.08983362e+00
-2.29073420e-01 4.34054285e-01 4.22483310e-02 1.74796984e-01
2.05323055e-01 4.36315000e-01 -1.59037754e-01 1.35288760e-01
-1.98022708e-01 7.20254660e-01 -1.04062736e+00 -8.59181821e-01
-2.86710441e-01 3.19656879e-01 -3.62682253e-01 -4.68538880e-01
-1.94404930e-01 2.97938377e-01 -3.22448194e-01 4.04394776e-01
1.60361081e-01 1.76172301e-01 6.74339712e-01 -5.31751871e-01
-4.70809340e-01 9.93671000e-01 -1.50571513e+00 1.67522395e+00
-2.34192595e-01 6.65178597e-01 -3.12293954e-02 -7.58600712e-01
8.85634899e-01 8.50754261e-01 2.42556304e-01 -1.89995140e-01
2.68804748e-03 5.42958617e-01 -3.01290989e-01 -5.70201993e-01
1.05937779e+00 -5.46579137e-02 -6.01952493e-01 1.61840111e-01
6.04215026e-01 1.16101757e-01 1.04694808e+00 2.85318971e-01
9.95834708e-01 2.34110117e-01 9.20457661e-01 -4.05228168e-01
5.87390304e-01 8.03539455e-01 8.11559498e-01 4.59274471e-01
4.96986229e-03 -1.63095653e-01 3.86785805e-01 -1.98323607e-01
-1.15241146e+00 -5.17749786e-01 -6.54017508e-01 9.75211322e-01
-2.34355301e-01 -9.75113273e-01 -8.91409278e-01 -7.83313751e-01
-5.81600308e-01 1.32459855e+00 -3.31161708e-01 6.20532513e-01
-8.16529274e-01 -6.17303729e-01 1.29925931e+00 3.25294167e-01
7.81883180e-01 -1.27600455e+00 -1.61753461e-01 5.56935489e-01
-7.59120643e-01 -1.52756238e+00 1.81717739e-01 1.87161952e-01
-4.03057039e-01 -9.21112716e-01 -2.10430384e-01 -1.00275266e+00
2.76456803e-01 -7.00929582e-01 1.28416336e+00 -3.33191425e-01
5.24372570e-02 4.86997306e-01 -6.86250746e-01 -5.24766445e-01
-7.60888100e-01 6.10807776e-01 -1.94051638e-01 -8.79983902e-01
6.67466044e-01 -1.96472496e-01 4.63287942e-02 4.61173236e-01
-7.66324878e-01 -3.78438056e-01 9.85480547e-02 6.75211847e-01
5.15917957e-01 4.27205861e-01 7.53083467e-01 -1.49999237e+00
5.33706725e-01 -4.04974282e-01 -4.63549256e-01 5.87171137e-01
-2.96131581e-01 -9.06439200e-02 -1.73209816e-01 3.28100234e-01
-1.80499637e+00 -2.21666977e-01 -4.57728058e-01 6.01747811e-01
-5.62815487e-01 9.27965283e-01 -4.80683357e-01 2.27429107e-01
9.64505851e-01 -4.73713279e-01 -8.42875719e-01 -7.96189010e-01
7.18276143e-01 1.00924313e+00 8.13853681e-01 -7.30940938e-01
2.71502376e-01 2.66782552e-01 -4.86796260e-01 -9.68068242e-01
-1.02205682e+00 -7.78826535e-01 -8.13553154e-01 -5.95910028e-02
1.08043051e+00 -9.76397216e-01 -1.56964079e-01 -2.60717839e-01
-1.14187062e+00 -2.33576000e-01 -7.96597302e-01 7.15382397e-01
-5.09943068e-01 1.98296010e-01 -7.17369676e-01 -6.41528904e-01
-9.32739526e-02 -3.25154096e-01 5.17240286e-01 5.85608147e-02
-8.49727631e-01 -1.36105156e+00 4.73683685e-01 6.47480428e-01
-1.63556859e-01 2.59405464e-01 8.62199843e-01 -1.44348216e+00
1.98672879e-02 -1.43217519e-01 1.31923836e-02 -7.27639943e-02
-2.27312356e-01 -3.38951573e-02 -6.71177089e-01 5.59163056e-02
-4.79343355e-01 -1.02001026e-01 3.57298970e-01 -5.99474125e-02
1.38722733e-01 -4.01204139e-01 -6.71045363e-01 1.34506494e-01
1.37207222e+00 3.84459347e-01 7.84756541e-01 1.08072495e+00
3.43306005e-01 9.39398646e-01 7.84308434e-01 2.99636334e-01
5.76542199e-01 7.48730659e-01 -3.91001046e-01 3.66595268e-01
-2.38363445e-01 -1.94010556e-01 5.47246709e-02 1.09171212e+00
-3.79363745e-01 -3.71214360e-01 -1.23150909e+00 8.85421574e-01
-1.78931034e+00 -1.30013359e+00 -3.26529682e-01 1.90564573e+00
9.78079438e-01 -1.12775434e-02 1.92456573e-01 -1.88119993e-01
1.09016073e+00 -1.92880407e-01 3.52357209e-01 -1.79566249e-01
-6.07273698e-01 3.18678826e-01 7.65523553e-01 4.64913189e-01
-1.15857649e+00 1.30223441e+00 6.63120413e+00 1.01845407e+00
-1.16078027e-01 5.91206849e-01 2.69770417e-02 3.63295227e-01
-7.01438636e-02 3.40557992e-01 -1.47959244e+00 1.52867466e-01
1.46939564e+00 -4.51174289e-01 -8.06943253e-02 7.04025149e-01
9.88870934e-02 -5.91210183e-03 -6.74418986e-01 3.76249403e-01
-1.10763907e-01 -1.60554087e+00 -1.96388751e-01 -3.78471315e-02
7.53786325e-01 3.94402117e-01 -9.93969142e-01 5.92159927e-01
8.72251213e-01 -3.99733841e-01 8.80308568e-01 3.74124974e-01
7.73402810e-01 -7.55141318e-01 1.15381706e+00 -6.75313920e-02
-1.16941190e+00 9.36205909e-02 -4.44355279e-01 4.94169623e-01
7.06219792e-01 4.72411931e-01 -1.01925206e+00 1.20273840e+00
8.05916131e-01 3.69328350e-01 -4.67785180e-01 1.29828095e+00
-5.60938656e-01 5.66359758e-01 -3.43688905e-01 1.82224602e-01
1.72945946e-01 -2.37205392e-03 7.51977623e-01 1.74451280e+00
2.65818924e-01 1.21482641e-01 1.25631271e-02 2.69866854e-01
-3.02523762e-01 7.32370436e-01 -2.60464162e-01 -2.06680834e-01
9.08155680e-01 1.36705363e+00 -1.07815766e+00 -5.88139653e-01
-3.93005639e-01 6.15907073e-01 3.91987830e-01 2.59942651e-01
-5.87547898e-01 -1.06974936e+00 1.89381704e-01 1.84521988e-01
3.63719553e-01 8.26261118e-02 1.98826969e-01 -8.10784817e-01
-2.49354392e-01 -6.31467640e-01 1.10472512e+00 -8.63163054e-01
-1.24805629e+00 9.06063735e-01 6.62967026e-01 -7.54028440e-01
-3.01564485e-01 -2.93664724e-01 9.45531577e-02 4.84946042e-01
-1.08395219e+00 -1.07911229e+00 4.50467207e-02 2.98252791e-01
4.69725393e-02 -4.67572473e-02 1.07058001e+00 8.84828806e-01
-4.95543242e-01 3.69354576e-01 8.31836835e-02 7.34639287e-01
1.00942850e+00 -1.26304412e+00 3.17373812e-01 5.05827129e-01
2.40541652e-01 7.37824619e-01 6.34453714e-01 -1.05813301e+00
-7.04726800e-02 -1.04577065e+00 1.67918384e+00 -5.72187603e-01
1.00178480e+00 -4.39225920e-02 -8.26364756e-01 1.12363553e+00
6.10368431e-01 -4.35010910e-01 1.10715485e+00 2.40122229e-01
-1.15036890e-01 2.70182073e-01 -1.13386011e+00 1.92852288e-01
1.15612769e+00 -4.00464892e-01 -1.11908197e+00 6.20742083e-01
7.53004432e-01 -7.15171874e-01 -1.40578222e+00 2.51732767e-01
1.81748137e-01 -5.34245014e-01 6.70647383e-01 -8.41651022e-01
-3.09838951e-01 -4.87896353e-01 -2.65100062e-01 -1.09394872e+00
-1.76062673e-01 -7.24391043e-01 3.84212524e-01 2.19478989e+00
9.12657976e-01 -5.70523560e-01 2.82853335e-01 5.71247280e-01
-4.94544655e-01 2.25952640e-01 -1.11792052e+00 -8.71034026e-01
-2.42034331e-01 -5.22944570e-01 6.61903501e-01 1.45066154e+00
5.67914069e-01 6.52046442e-01 2.20403805e-01 2.16025755e-01
1.45125449e-01 -5.39505899e-01 3.86002302e-01 -1.34494495e+00
-2.28987057e-02 -9.12987366e-02 -5.86704314e-01 -3.49564165e-01
3.63906771e-01 -1.06402159e+00 -1.32675111e-01 -1.91715240e+00
-1.23090640e-01 -7.04066157e-01 5.64612895e-02 6.73228145e-01
2.20509112e-01 1.02920175e-01 -1.40080914e-01 2.73672819e-01
-9.18169796e-01 -1.36438757e-01 4.01460022e-01 3.35868061e-01
-2.96840847e-01 -3.28146547e-01 -5.97402036e-01 9.22906518e-01
6.00208998e-01 -8.15898895e-01 5.08700870e-02 -3.83116275e-01
7.54442394e-01 -7.06839710e-02 -2.57975787e-01 -9.07174706e-01
3.24336559e-01 1.69270992e-01 2.04405323e-01 -6.86380506e-01
1.59549993e-02 -8.08676541e-01 6.96113110e-01 3.86272348e-03
-4.69844848e-01 6.57527670e-02 3.76916736e-01 2.75080651e-01
-5.68120241e-01 -9.06315029e-01 6.98570728e-01 -3.55720937e-01
-1.17232966e+00 -3.85876596e-01 -6.07131422e-01 6.94593906e-01
1.10693204e+00 -9.06162038e-02 -4.94225711e-01 -1.70378268e-01
-1.34142470e+00 1.68982968e-01 3.51519674e-01 3.78658585e-02
-3.03685188e-01 -1.43619418e+00 -7.64337897e-01 -5.34047246e-01
4.33676153e-01 -3.47024113e-01 1.17162749e-01 7.09687829e-01
-8.62016499e-01 7.46666074e-01 -3.21623147e-01 2.54010141e-01
-1.26404536e+00 3.14275861e-01 1.13930665e-01 -6.44191921e-01
-5.59393346e-01 6.84776604e-01 -5.10012746e-01 -7.29988992e-01
1.00037821e-01 1.68021351e-01 -9.21619773e-01 6.79640710e-01
3.96323085e-01 6.66520476e-01 4.63904470e-01 -1.07067561e+00
-5.91794729e-01 4.84674331e-03 1.31411985e-01 -5.47559261e-01
1.36719060e+00 -5.76327860e-01 -3.43642533e-01 4.46777493e-01
7.75219083e-01 5.41712999e-01 -1.05491661e-01 -1.54835135e-01
1.13673663e+00 1.81563497e-01 -1.65587272e-02 -1.00452256e+00
-4.95557070e-01 6.42163679e-02 1.35798454e-01 3.75987023e-01
6.69229150e-01 3.57649326e-01 5.17812550e-01 9.02278900e-01
3.86672288e-01 -1.78515232e+00 -6.69927120e-01 7.38530040e-01
6.93272233e-01 -7.22118497e-01 9.87497270e-02 -8.46822083e-01
-7.19582796e-01 1.06100357e+00 2.36671612e-01 4.44096893e-01
5.95035195e-01 3.68254006e-01 1.91296905e-01 -5.54195940e-01
-3.78515035e-01 -5.73120594e-01 1.75513938e-01 8.67559373e-01
5.93226433e-01 -1.65363774e-01 -8.72342587e-01 7.66641974e-01
-5.28808296e-01 -8.10272470e-02 6.27467036e-01 1.02539062e+00
-4.33751523e-01 -1.48757541e+00 -3.03062528e-01 2.37890467e-01
-1.05900836e+00 -2.89798081e-01 -5.02978265e-01 1.55047572e+00
2.89349258e-01 8.69274318e-01 2.82425702e-01 2.48553559e-01
6.36525214e-01 4.76445466e-01 2.21231058e-01 -1.04439437e+00
-8.55856776e-01 2.64419377e-01 1.54054749e+00 -1.87831804e-01
-1.20088542e+00 -6.26112282e-01 -1.49553359e+00 -1.12070106e-01
-7.26094127e-01 1.03704464e+00 6.10136509e-01 8.20646524e-01
2.03579247e-01 3.09034884e-01 -2.82465398e-01 -1.72466010e-01
2.94414192e-01 -1.01595771e+00 -8.49677980e-01 1.62517786e-01
-7.89517581e-01 -5.89078248e-01 2.61004064e-02 3.99166167e-01] | [9.5209379196167, 9.134490966796875] |
458fade9-7474-4797-87ba-247d3dd52221 | reset-revisiting-trajectory-sets-for | 2304.05856 | null | https://arxiv.org/abs/2304.05856v1 | https://arxiv.org/pdf/2304.05856v1.pdf | RESET: Revisiting Trajectory Sets for Conditional Behavior Prediction | It is desirable to predict the behavior of traffic participants conditioned on different planned trajectories of the autonomous vehicle. This allows the downstream planner to estimate the impact of its decisions. Recent approaches for conditional behavior prediction rely on a regression decoder, meaning that coordinates or polynomial coefficients are regressed. In this work we revisit set-based trajectory prediction, where the probability of each trajectory in a predefined trajectory set is determined by a classification model, and first-time employ it to the task of conditional behavior prediction. We propose RESET, which combines a new metric-driven algorithm for trajectory set generation with a graph-based encoder. For unconditional prediction, RESET achieves comparable performance to a regression-based approach. Due to the nature of set-based approaches, it has the advantageous property of being able to predict a flexible number of trajectories without influencing runtime or complexity. For conditional prediction, RESET achieves reasonable results with late fusion of the planned trajectory, which was not observed for regression-based approaches before. This means that RESET is computationally lightweight to combine with a planner that proposes multiple future plans of the autonomous vehicle, as large parts of the forward pass can be reused. | ['Klaus Dietmayer', 'Vasileios Belagiannis', 'Julian Jordan', 'Julian Wiederer', 'Pascal Huissel', 'Julian Schmidt'] | 2023-04-12 | null | null | null | null | ['trajectory-prediction'] | ['computer-vision'] | [ 7.02843964e-02 3.84598166e-01 -3.39931726e-01 -4.58666921e-01
-7.22194135e-01 -4.49149370e-01 7.90560067e-01 4.58313316e-01
-2.26497412e-01 6.63961053e-01 3.44483286e-01 -7.20634818e-01
-1.81534454e-01 -1.05247843e+00 -8.72847438e-01 -4.38616246e-01
-2.10713193e-01 7.09752560e-01 7.09385872e-01 -3.37688506e-01
1.83156148e-01 8.44500482e-01 -1.88276005e+00 2.00206026e-01
5.53159118e-01 5.89488506e-01 3.03682089e-01 9.74993527e-01
7.95613155e-02 9.30027962e-01 -9.87239033e-02 -4.05544847e-01
1.22999065e-01 -3.39953005e-01 -5.03607273e-01 -2.59086967e-01
4.05767933e-02 -3.59166890e-01 -5.65052211e-01 3.21147889e-01
-4.90558557e-02 2.68092930e-01 8.86566103e-01 -1.76540232e+00
3.15288305e-01 6.61337554e-01 2.02909589e-01 -7.92223737e-02
5.28906643e-01 3.04139823e-01 8.53933871e-01 -5.65151930e-01
6.96992397e-01 1.03366411e+00 6.76368594e-01 3.49020392e-01
-1.35895455e+00 -4.53031272e-01 2.48728484e-01 5.09276688e-01
-1.44968760e+00 -7.28198469e-01 4.81835812e-01 -6.49177969e-01
1.10260737e+00 4.07394648e-01 6.30209684e-01 6.82075322e-01
4.02714670e-01 4.54232842e-01 4.02844310e-01 -3.60487215e-02
3.37484449e-01 2.57392764e-01 -1.49759144e-01 7.98908472e-01
6.61408976e-02 4.70594853e-01 -1.48317829e-01 -7.33856037e-02
8.32559634e-03 6.80909604e-02 3.68516184e-02 -2.62025714e-01
-1.18570578e+00 5.67616880e-01 3.63868743e-01 -1.21219069e-01
-5.67786038e-01 4.44224894e-01 8.66431743e-02 3.42806637e-01
2.24574611e-01 1.27842337e-01 -3.95787567e-01 -4.35972124e-01
-1.12367737e+00 7.39388287e-01 9.88956094e-01 1.30374491e+00
1.22824168e+00 -2.93254316e-01 -5.36823452e-01 -5.64011149e-02
3.74684751e-01 4.57033902e-01 2.93307267e-02 -8.51929128e-01
5.65297127e-01 8.09847534e-01 1.83407798e-01 -1.01342511e+00
-5.90098917e-01 5.14915027e-02 -4.28974569e-01 4.26985234e-01
3.72870356e-01 -1.72759220e-01 -6.85442209e-01 1.48630393e+00
4.01077271e-01 3.44408065e-01 2.61758845e-02 5.67306399e-01
2.39279136e-01 8.22220385e-01 2.76820995e-02 -1.07140966e-01
6.82424128e-01 -7.55194485e-01 -2.63487160e-01 1.14458650e-01
1.21018839e+00 -3.77053708e-01 5.86361945e-01 2.07019269e-01
-8.26753914e-01 -4.76926029e-01 -8.38138282e-01 1.32066846e-01
-4.67602491e-01 1.23676788e-02 1.75091982e-01 6.29401743e-01
-1.20684445e+00 8.34491789e-01 -9.77772057e-01 -2.89846510e-01
3.13934594e-01 5.86025178e-01 -2.17018828e-01 -6.74702153e-02
-7.83029795e-01 1.22183156e+00 4.02254820e-01 -3.90901178e-01
-7.36175299e-01 -8.46410453e-01 -8.29600334e-01 1.21446759e-01
2.92521954e-01 -5.76948345e-01 1.16331494e+00 -3.77897918e-01
-1.50052726e+00 5.50995469e-02 -6.42459095e-01 -7.17480361e-01
6.63533509e-01 2.74758518e-01 -4.86541420e-01 -2.23449290e-01
2.51876507e-02 7.71686137e-01 7.13853896e-01 -8.25403690e-01
-1.02347398e+00 1.16975857e-02 6.62007034e-02 7.19351694e-02
1.00190289e-01 -3.28418791e-01 -4.31934118e-01 8.10455903e-02
-2.57643223e-01 -1.27411151e+00 -7.35935390e-01 -1.77582711e-01
-3.49666625e-01 -4.81086522e-01 8.98057342e-01 -3.83345962e-01
1.48179352e+00 -1.99285805e+00 7.38333687e-02 5.37657380e-01
6.03196174e-02 -7.95882475e-03 -6.54496849e-02 8.85027230e-01
5.70756234e-02 2.25105435e-02 -2.27587640e-01 -3.82213831e-01
6.83586150e-02 1.72025248e-01 -3.31470460e-01 5.38874865e-01
4.67767477e-01 8.88784230e-01 -7.94921517e-01 -3.86544496e-01
5.53304613e-01 2.48167321e-01 -8.98130953e-01 2.84379274e-01
-5.55136025e-01 5.20928800e-01 -4.63354677e-01 9.08794627e-02
4.26752716e-01 2.44257227e-01 1.62725732e-01 4.19296026e-02
-5.77369571e-01 5.92628419e-01 -1.04916394e+00 1.29297376e+00
-5.20419776e-01 7.24466205e-01 -4.57911491e-01 -7.44330108e-01
1.02422130e+00 3.25408995e-01 7.16165721e-01 -5.05439758e-01
-2.46423244e-01 1.17568016e-01 -1.84984371e-01 -4.39834058e-01
6.85142756e-01 2.44585887e-01 -3.13767046e-01 4.51891929e-01
-2.54674077e-01 -1.24308892e-01 3.76383036e-01 1.13125198e-01
1.49545205e+00 4.83039588e-01 2.19885230e-01 -5.63072897e-02
6.28740013e-01 5.45098007e-01 3.55805278e-01 5.12765825e-01
2.15155423e-01 3.17924976e-01 6.60338938e-01 -4.34547752e-01
-1.19327223e+00 -8.09289813e-01 1.38523206e-01 9.53554630e-01
1.50124595e-01 -8.43980312e-01 -5.00313044e-01 -7.29935765e-01
8.58572964e-03 1.25454617e+00 -4.00967181e-01 -2.06528798e-01
-8.62747729e-01 -1.25533208e-01 3.45742971e-01 2.99624354e-01
-1.82570472e-01 -7.00534940e-01 -6.40381992e-01 5.15871882e-01
1.17420062e-01 -9.52764869e-01 -3.37532312e-01 -1.75067540e-02
-6.45348728e-01 -1.07156348e+00 -3.40666659e-02 -4.31570917e-01
7.47018456e-01 1.26696408e-01 7.39724040e-01 2.07950294e-01
3.07912320e-01 2.66895771e-01 -2.91541785e-01 -3.21426451e-01
-9.92554545e-01 3.59926254e-01 -6.45987093e-02 1.22023575e-01
2.35403642e-01 -5.48741400e-01 -3.83521110e-01 4.31316614e-01
-4.27733839e-01 3.72027189e-01 4.49937165e-01 2.88107842e-01
4.84074593e-01 2.69387186e-01 4.86566693e-01 -5.27116776e-01
3.98893327e-01 -8.66154850e-01 -7.16501236e-01 3.20401080e-02
-7.70762622e-01 3.46264839e-01 8.88932705e-01 -1.43177360e-01
-7.55888462e-01 6.55337632e-01 -3.36502314e-01 -3.04053575e-01
-4.09157872e-01 4.56825197e-01 6.44948855e-02 2.50988543e-01
5.12699902e-01 2.53074288e-01 1.51911691e-01 -3.41982305e-01
6.12375677e-01 5.78329623e-01 2.24367097e-01 -6.46281615e-02
8.13855290e-01 1.91510990e-01 4.30846542e-01 -6.44069016e-01
-1.46560907e-01 -4.23788160e-01 -9.09315288e-01 -6.33732140e-01
7.85965502e-01 -6.87591314e-01 -1.10892224e+00 -1.57987446e-01
-1.27722609e+00 -5.85498333e-01 -4.31209654e-01 5.13944864e-01
-9.03472662e-01 -1.28222361e-01 6.58307970e-02 -7.92601705e-01
2.22492918e-01 -1.44280577e+00 9.47414160e-01 -1.72443718e-01
-4.24459308e-01 -8.65345895e-01 4.55010757e-02 -1.84278205e-01
3.21480274e-01 1.81198359e-01 8.77495885e-01 -7.15442002e-01
-9.47188020e-01 -4.69449788e-01 6.35147691e-02 -2.47581974e-01
-9.24824774e-02 1.68477237e-01 -5.13446987e-01 -1.13696441e-01
-6.43146813e-01 3.78666282e-01 6.56087637e-01 1.83648542e-01
9.01702881e-01 -6.61500096e-01 -8.43492568e-01 2.54545033e-01
1.17069387e+00 2.20323995e-01 6.69185579e-01 1.03741489e-01
6.30588830e-01 7.96604991e-01 7.55554318e-01 3.10611486e-01
1.02899182e+00 1.04533911e+00 4.71330583e-01 3.33114624e-01
-1.66072071e-01 -5.73049128e-01 7.36076832e-01 3.50400507e-01
-4.38493676e-02 -3.51241946e-01 -1.06461883e+00 5.05212426e-01
-2.07686830e+00 -1.22239804e+00 -5.48720837e-01 2.34681249e+00
3.32453519e-01 1.94527566e-01 6.36614084e-01 1.99969262e-01
4.84542847e-01 -1.46858975e-01 -3.48275036e-01 -5.86210728e-01
6.16774201e-01 -2.15161934e-01 9.38371181e-01 7.08213925e-01
-7.85233796e-01 8.48741531e-01 6.40307808e+00 6.45335972e-01
-9.56416607e-01 -3.41854407e-03 2.87818372e-01 -7.01768771e-02
-5.04871726e-01 2.92866260e-01 -1.11476946e+00 5.75694561e-01
1.59970224e+00 -4.52918470e-01 5.05854487e-01 6.36639416e-01
6.19828701e-01 -1.78778410e-01 -1.47972357e+00 3.41893882e-01
-3.46196622e-01 -1.54128337e+00 -1.77482069e-01 2.92545438e-01
3.75342518e-01 2.66015947e-01 -2.62373954e-01 4.62298930e-01
3.73378575e-01 -1.00718296e+00 7.90785551e-01 1.05130494e+00
6.05321467e-01 -1.04326022e+00 3.67364556e-01 8.88071597e-01
-1.36575234e+00 -2.78093964e-01 -8.87039155e-02 -1.44655958e-01
5.29740095e-01 2.04288021e-01 -1.60804677e+00 6.03283346e-01
2.18244374e-01 8.08556616e-01 -5.07081687e-01 1.11867309e+00
-3.76152620e-02 9.08533633e-01 -5.76746821e-01 -3.46711278e-01
2.18449995e-01 -1.94556162e-01 8.02716553e-01 1.40120351e+00
6.91524565e-01 -1.18791230e-01 3.60692292e-01 7.43548989e-01
4.64903325e-01 1.76897999e-02 -1.01393151e+00 2.64534116e-01
5.82153976e-01 9.42633450e-01 -3.85082066e-01 -2.93709666e-01
-4.86408293e-01 4.39256579e-01 3.65990192e-01 1.35948673e-01
-9.98514593e-01 -8.24429393e-02 6.33666039e-01 5.48412383e-01
5.41397631e-01 -5.19376159e-01 -3.52826804e-01 -4.48845536e-01
-1.38010874e-01 -2.08543509e-01 8.37664753e-02 -4.56038177e-01
-3.72914642e-01 6.11646712e-01 2.43769333e-01 -1.66903293e+00
-7.25620687e-01 -1.61799639e-01 -7.01795399e-01 7.39911020e-01
-1.35109699e+00 -1.09696889e+00 3.58759388e-02 3.80073518e-01
4.71936464e-01 -2.34667078e-01 7.13906348e-01 1.74545854e-01
-5.13771117e-01 3.13160270e-01 -2.05884263e-01 -3.37678164e-01
2.75903344e-01 -1.01426017e+00 6.41821623e-01 8.42229307e-01
3.90278623e-02 1.06115371e-01 8.60420167e-01 -8.99820507e-01
-1.61887622e+00 -1.63831961e+00 1.38798726e+00 -7.26012468e-01
5.85976064e-01 -1.61040291e-01 -5.63579500e-01 8.51968408e-01
-1.61169216e-01 -2.95371175e-01 2.47717917e-01 -8.34313482e-02
2.48075441e-01 -3.17381918e-01 -9.69108224e-01 8.69470954e-01
8.76260996e-01 -6.88686296e-02 -1.61810309e-01 1.76013991e-01
7.88324416e-01 -3.35255831e-01 -9.15065467e-01 1.85563549e-01
4.46496397e-01 -8.73088896e-01 7.15689838e-01 -4.55269635e-01
2.36615226e-01 -5.15314341e-01 -2.60933302e-02 -1.24082565e+00
-5.06459653e-01 -5.79213560e-01 -4.03677613e-01 1.05641103e+00
9.34939325e-01 -6.14956915e-01 7.99394667e-01 9.95546877e-01
-3.85107756e-01 -8.11294019e-01 -1.04115295e+00 -5.67604244e-01
-2.49426469e-01 -1.03675687e+00 1.17462873e+00 1.93751022e-01
1.70212507e-01 2.01344073e-01 -4.59313571e-01 4.28419977e-01
3.04695994e-01 3.93986478e-02 1.18295562e+00 -1.07286787e+00
2.93520298e-02 -4.96686131e-01 -6.64610624e-01 -1.17784047e+00
1.16954274e-01 -1.20040631e+00 2.37404734e-01 -1.56726420e+00
-3.95176351e-01 -9.46111262e-01 1.99630484e-02 5.17956734e-01
1.28128633e-01 -1.23104073e-01 2.68469274e-01 1.65569410e-01
-4.82199490e-01 4.49052781e-01 9.58670378e-01 3.07636568e-03
-6.17346644e-01 6.27496183e-01 -2.58879125e-01 4.51790690e-01
9.63001430e-01 -8.78728867e-01 -5.85790277e-01 8.06187987e-02
1.99370712e-01 5.82653165e-01 3.89703661e-01 -1.28597295e+00
6.72045529e-01 -4.09098953e-01 -1.09313950e-01 -1.00912929e+00
4.04850960e-01 -1.09358954e+00 5.47236443e-01 4.29682106e-01
-4.93086666e-01 2.78906614e-01 -6.23693550e-03 8.85823131e-01
4.47917804e-02 -6.11456297e-02 2.55706877e-01 3.83491397e-01
-8.23873281e-01 6.23301089e-01 -1.02848554e+00 -5.41460931e-01
1.42494750e+00 -4.17349309e-01 2.23710492e-01 -4.63469535e-01
-5.12860715e-01 4.33861047e-01 3.96573663e-01 4.70041633e-01
6.80550218e-01 -1.32657635e+00 -7.73129821e-01 2.63329983e-01
7.41230845e-02 -1.80329662e-02 -1.03687793e-01 1.13451064e+00
-3.35144162e-01 4.95993167e-01 6.36865012e-03 -6.50780559e-01
-1.09668219e+00 6.50311172e-01 1.97943717e-01 -1.95205420e-01
-8.78446817e-01 1.90657258e-01 -3.97604108e-01 -4.60117370e-01
-8.74477029e-02 -3.39644492e-01 -4.20838594e-01 -1.57353342e-01
4.38197672e-01 6.90235198e-01 1.86846554e-01 -9.20190156e-01
-3.87005419e-01 1.49535000e-01 2.26870209e-01 -2.45744735e-01
1.20182610e+00 -2.01285586e-01 1.98101580e-01 3.81958663e-01
1.12733507e+00 1.09197600e-02 -1.50413561e+00 1.42255962e-01
2.12712377e-01 -2.15544432e-01 8.85023698e-02 -5.06823182e-01
-7.30208397e-01 4.97996449e-01 7.37400874e-02 3.52571249e-01
8.56345117e-01 -1.00156121e-01 7.61399269e-01 3.92738044e-01
6.17060661e-01 -8.66608381e-01 -5.97401917e-01 5.11154115e-01
7.80681968e-01 -9.00303841e-01 -4.02214490e-02 -3.12107056e-01
-7.23837793e-01 1.35305011e+00 3.30184221e-01 -4.31970917e-02
8.29201460e-01 3.17979544e-01 -4.48615789e-01 -3.68787609e-02
-1.25980246e+00 -4.68546480e-01 4.67475653e-01 6.85078442e-01
4.37681228e-02 2.33570412e-01 -1.83417633e-01 1.75473094e-01
-4.75450158e-01 1.15312822e-01 5.23343861e-01 5.68078220e-01
-5.25952041e-01 -1.27903044e+00 -5.57081588e-02 6.43492520e-01
2.91116893e-01 2.37541035e-01 -2.22229343e-02 5.26798368e-01
1.65956259e-01 1.24127221e+00 2.67886341e-01 -8.96022677e-01
5.02431929e-01 3.61352786e-02 1.04411207e-01 -4.65521634e-01
-4.71014142e-01 -5.06258428e-01 6.28503203e-01 -9.17090833e-01
-1.04515769e-01 -1.06219041e+00 -1.37584174e+00 -4.91705120e-01
1.24976449e-02 1.51745453e-01 7.84366786e-01 1.03426075e+00
7.17669785e-01 3.86282831e-01 8.88676763e-01 -1.08200705e+00
-1.67204842e-01 -5.38118541e-01 -7.02569038e-02 7.78800845e-02
5.06921828e-01 -5.26704550e-01 1.58613250e-01 5.39610162e-02] | [5.834871768951416, 1.0121999979019165] |
90a34982-f735-482a-bcd2-9ec900cdd4d8 | rogue-signs-deceiving-traffic-sign | 1801.02780 | null | http://arxiv.org/abs/1801.02780v3 | http://arxiv.org/pdf/1801.02780v3.pdf | Rogue Signs: Deceiving Traffic Sign Recognition with Malicious Ads and Logos | We propose a new real-world attack against the computer vision based systems
of autonomous vehicles (AVs). Our novel Sign Embedding attack exploits the
concept of adversarial examples to modify innocuous signs and advertisements in
the environment such that they are classified as the adversary's desired
traffic sign with high confidence. Our attack greatly expands the scope of the
threat posed to AVs since adversaries are no longer restricted to just
modifying existing traffic signs as in previous work. Our attack pipeline
generates adversarial samples which are robust to the environmental conditions
and noisy image transformations present in the physical world. We ensure this
by including a variety of possible image transformations in the optimization
problem used to generate adversarial samples. We verify the robustness of the
adversarial samples by printing them out and carrying out drive-by tests
simulating the conditions under which image capture would occur in a real-world
scenario. We experimented with physical attack samples for different distances,
lighting conditions and camera angles. In addition, extensive evaluations were
carried out in the virtual setting for a variety of image transformations. The
adversarial samples generated using our method have adversarial success rates
in excess of 95% in the physical as well as virtual settings. | ['Chawin Sitawarin', 'Mung Chiang', 'Arsalan Mosenia', 'Arjun Nitin Bhagoji', 'Prateek Mittal'] | 2018-01-09 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 3.93027186e-01 2.39459518e-02 3.11777115e-01 -1.85626820e-01
-4.58621442e-01 -1.22581899e+00 1.18716884e+00 -7.57008970e-01
-3.63429159e-01 5.31387031e-01 -4.34318662e-01 -6.49967134e-01
3.89771134e-01 -8.32832277e-01 -1.11901581e+00 -6.27063096e-01
-1.06449448e-01 2.37989053e-01 5.18924713e-01 -4.31192458e-01
-3.67852449e-02 9.66616333e-01 -1.46653700e+00 -1.99832752e-01
4.96922970e-01 6.87454998e-01 -3.56760144e-01 1.25762427e+00
4.40576047e-01 5.73959947e-01 -1.23378980e+00 -7.45765805e-01
1.08959115e+00 -1.05030701e-01 -1.89402610e-01 3.31881866e-02
7.94153631e-01 -6.28955901e-01 -7.76467025e-01 1.31384683e+00
1.68849468e-01 -1.02135532e-01 4.09819454e-01 -1.99256265e+00
-5.13621151e-01 -1.06266581e-01 -2.29505509e-01 -1.67332992e-01
3.37775171e-01 9.93816197e-01 4.88194346e-01 -3.60286146e-01
8.54234040e-01 1.29493213e+00 3.67206007e-01 7.11021066e-01
-1.16275227e+00 -1.19382370e+00 6.31295424e-03 3.22185546e-01
-1.09010398e+00 -6.05126798e-01 7.20220149e-01 -2.18615845e-01
4.17974055e-01 5.38588941e-01 3.80693376e-01 1.66022658e+00
3.94685537e-01 4.68978107e-01 1.12165797e+00 -2.35447854e-01
2.81168878e-01 4.06799585e-01 -4.16620642e-01 5.52375555e-01
4.34913725e-01 8.01131845e-01 9.70006064e-02 -5.10460496e-01
3.34030211e-01 -3.50936234e-01 -1.02553563e-03 -4.83886719e-01
-1.24003601e+00 9.05734777e-01 3.23627472e-01 -3.08511615e-01
-1.74008593e-01 5.62637627e-01 3.76234472e-01 6.04136705e-01
-3.31135303e-01 5.75060427e-01 -2.71361202e-01 2.55393654e-01
-2.35086933e-01 4.26718414e-01 7.08944857e-01 1.02228999e+00
3.58500421e-01 5.80162883e-01 -3.58950980e-02 2.38973200e-01
3.45593750e-01 1.39558947e+00 -3.33179496e-02 -1.21845090e+00
4.72732574e-01 6.92046136e-02 4.70842421e-01 -1.22232318e+00
2.05164447e-01 8.14269781e-02 -2.40282327e-01 1.14138567e+00
5.98111272e-01 -4.82522637e-01 -1.21096814e+00 1.72663558e+00
5.17429590e-01 4.83970344e-01 4.38920528e-01 7.18896091e-01
4.64416921e-01 6.23052001e-01 -6.57055303e-02 2.08828256e-01
1.06307852e+00 -5.46922088e-01 -5.64676285e-01 -2.81153709e-01
2.65327096e-01 -9.80201483e-01 9.39657450e-01 2.20012754e-01
-4.45797533e-01 -4.01026756e-01 -1.45612848e+00 8.23774457e-01
-6.19496405e-01 -4.97583389e-01 3.92728418e-01 1.31843615e+00
-7.98355818e-01 7.70296007e-02 -5.31798124e-01 -2.23765031e-01
3.37272435e-01 3.78936231e-01 -5.15317023e-01 -8.81116390e-02
-1.43856049e+00 1.11718726e+00 -1.04624122e-01 1.22186236e-01
-1.43160772e+00 -3.96179467e-01 -9.23480213e-01 -6.00278914e-01
4.65166211e-01 -3.22800189e-01 1.11606061e+00 -9.78537023e-01
-1.31990993e+00 7.98304498e-01 1.43164113e-01 -7.24002719e-01
9.07477260e-01 1.92509741e-01 -1.04127645e+00 1.37776628e-01
-8.48605782e-02 6.13047540e-01 1.18766046e+00 -1.78226423e+00
-3.16848367e-01 7.53119364e-02 4.71588284e-01 -1.98600516e-01
1.56000167e-01 2.85108447e-01 -1.16177477e-01 -6.15312934e-01
-4.78745431e-01 -1.43316960e+00 -2.16877460e-01 4.00507867e-01
-5.02456367e-01 6.94139600e-01 1.51770806e+00 -2.35049665e-01
2.97052890e-01 -2.18526220e+00 -5.04172087e-01 6.34806573e-01
2.50084437e-02 6.77469373e-01 -4.82462257e-01 3.20605874e-01
4.63986136e-02 2.96697617e-01 1.09052025e-01 1.35670394e-01
3.07753116e-01 5.90565562e-01 -7.74412036e-01 6.49433970e-01
2.55718976e-01 8.82980883e-01 -8.24689448e-01 -2.14715347e-01
4.45937097e-01 3.02897900e-01 -2.88645357e-01 2.30522767e-01
-2.16622159e-01 5.17623782e-01 -4.75844264e-01 6.13423944e-01
8.03555548e-01 6.60001099e-01 -1.22854680e-01 4.87840027e-02
3.67103130e-01 -4.10806745e-01 -1.13982439e+00 5.59254110e-01
-2.94425786e-01 1.05118906e+00 1.41873077e-01 -4.57837045e-01
8.43376756e-01 9.62100476e-02 7.54821822e-02 -7.08354175e-01
2.70886093e-01 -1.44910850e-02 3.14125746e-01 -4.60604727e-01
4.48039919e-01 7.60661587e-02 -4.17970538e-01 2.98111707e-01
-4.84472632e-01 -5.51186621e-01 6.63170293e-02 3.58808339e-01
1.34065747e+00 -2.34694332e-01 -2.06042334e-01 2.06322446e-01
5.17319262e-01 3.32793221e-02 3.66321355e-01 9.53396976e-01
-6.69627845e-01 2.84088194e-01 4.40276474e-01 -3.31157625e-01
-1.18778968e+00 -1.52053046e+00 3.18333618e-02 4.74161804e-01
4.96818721e-01 4.14591841e-02 -5.29084086e-01 -1.01202738e+00
2.06321806e-01 9.83042896e-01 -7.84038782e-01 -5.03028870e-01
-5.53166807e-01 -2.48890027e-01 1.27742779e+00 2.06315428e-01
6.14315569e-01 -9.67374921e-01 -4.80292141e-01 -2.91996330e-01
2.56867051e-01 -1.59907019e+00 -4.52778906e-01 -3.06128114e-01
7.59381875e-02 -1.52715194e+00 -1.30436674e-01 -4.86441046e-01
9.27652478e-01 2.19306305e-01 6.39710248e-01 5.38360849e-02
-3.45632821e-01 5.32710969e-01 -3.15527111e-01 -7.86315143e-01
-1.38337100e+00 -8.64868283e-01 2.40120336e-01 2.80026317e-01
2.77395453e-02 -2.09730536e-01 -4.16565061e-01 7.70336270e-01
-9.57079053e-01 -4.61281002e-01 3.28549951e-01 7.73979187e-01
1.42133325e-01 3.12778026e-01 2.37159505e-01 -6.53558731e-01
4.69536513e-01 -3.59777391e-01 -1.14822090e+00 1.41636983e-01
-2.60897189e-01 -5.34030907e-02 9.78627622e-01 -8.78044367e-01
-7.21388400e-01 -4.91001755e-02 1.11303724e-01 -6.01967931e-01
-4.45309788e-01 -3.11335504e-01 -4.25384253e-01 -7.34704316e-01
7.83142447e-01 3.85456532e-02 7.78412074e-02 2.86300987e-01
5.85060716e-01 4.53689963e-01 6.79169774e-01 -3.47629517e-01
1.95128524e+00 6.88981652e-01 3.04811090e-01 -7.39981472e-01
-2.71609146e-02 2.84491152e-01 -7.52717629e-02 -6.09452605e-01
7.45141745e-01 -4.76392657e-01 -9.13868845e-01 7.87460864e-01
-1.05723310e+00 -2.93983310e-01 -1.83376282e-01 3.06246370e-01
-3.70239884e-01 5.29529631e-01 -1.21540166e-01 -6.66895509e-01
1.84073791e-01 -1.67007184e+00 9.38958466e-01 6.64571747e-02
-1.20474003e-01 -7.60977030e-01 -9.46815982e-02 3.20920050e-01
5.55735290e-01 9.43483055e-01 5.72505772e-01 -5.84330022e-01
-7.42772996e-01 -7.69545376e-01 -4.88867937e-03 4.26298887e-01
9.94978249e-02 4.91274029e-01 -8.50227773e-01 -3.90104234e-01
-2.83909410e-01 -2.47142255e-01 3.16396356e-01 -1.87372431e-01
6.46308720e-01 -5.99569499e-01 -1.74690753e-01 5.48336983e-01
1.22232652e+00 5.77830970e-01 8.74071836e-01 5.37620962e-01
6.33469403e-01 3.62574100e-01 7.41092682e-01 -7.30695054e-02
-1.31129608e-01 8.74073267e-01 9.16211903e-01 -1.89540789e-01
3.75513993e-02 -2.36791059e-01 7.18112946e-01 -3.63649756e-01
4.16139126e-01 -4.75075722e-01 -6.69757307e-01 3.54059845e-01
-1.34805346e+00 -1.25993013e+00 -2.25999981e-01 2.27714944e+00
3.20771039e-01 4.89914954e-01 6.16288744e-02 1.08781613e-01
7.82443225e-01 3.27713013e-01 -6.01546645e-01 -8.80653262e-01
-1.47855520e-01 1.07306801e-01 1.21318090e+00 6.21882260e-01
-1.29698968e+00 9.34058487e-01 6.81807995e+00 3.70926023e-01
-1.22385681e+00 -1.94813699e-01 2.11093083e-01 -3.36144902e-02
-3.50031525e-01 1.37068972e-01 -4.90294665e-01 6.20728433e-01
9.86286283e-01 -1.88653022e-01 4.72733051e-01 7.71270871e-01
2.62566298e-01 6.24844283e-02 -8.49381268e-01 5.03050148e-01
5.55456318e-02 -1.20177865e+00 -1.76121794e-05 2.22950399e-01
7.32050180e-01 2.40351349e-01 4.25867736e-01 8.54939669e-02
1.06464946e+00 -1.04526627e+00 9.18593228e-01 1.70366898e-01
9.13764358e-01 -6.14919305e-01 6.80405676e-01 -8.13534930e-02
-7.97470510e-01 -5.02062924e-02 4.67409343e-02 3.29963207e-01
2.30779871e-01 -1.09110832e-01 -1.12949872e+00 2.67264813e-01
2.43568912e-01 -4.14047875e-02 -6.52587295e-01 6.67296112e-01
-4.93035406e-01 7.86985278e-01 -5.90323150e-01 4.78678755e-02
4.39122677e-01 8.55438262e-02 1.09547698e+00 9.55620766e-01
1.23598814e-01 -4.20482635e-01 -4.51598279e-02 7.23851144e-01
9.63751459e-04 -4.69399810e-01 -1.27934444e+00 8.67943391e-02
5.33728004e-01 9.06577647e-01 -4.18932796e-01 -1.05965167e-01
-1.52593523e-01 9.50552583e-01 -6.29651129e-01 6.17958009e-01
-1.52007115e+00 -4.87197369e-01 1.14234757e+00 -6.87792301e-02
4.34064537e-01 -3.45782012e-01 2.01005433e-02 -8.20207238e-01
2.08488181e-01 -1.15333664e+00 -1.71684641e-02 -7.27577806e-01
-9.45505321e-01 5.92206836e-01 5.76061383e-02 -1.55404484e+00
-3.32228780e-01 -7.19914377e-01 -9.85109687e-01 7.59363532e-01
-1.28932273e+00 -1.23147714e+00 -3.03093940e-01 7.34559059e-01
2.04135329e-01 -5.97746670e-01 7.64502823e-01 2.41073370e-01
-3.92210275e-01 1.04508758e+00 1.35484487e-01 2.65965283e-01
6.41512990e-01 -8.08011115e-01 1.00577855e+00 1.37889242e+00
-3.82422470e-02 3.65170479e-01 1.17205119e+00 -5.61768711e-01
-1.53599107e+00 -1.32540929e+00 2.49466702e-01 -5.87981641e-01
1.01264453e+00 -5.74372232e-01 -4.39390451e-01 7.67639995e-01
1.43412277e-01 3.41155678e-01 1.04899533e-01 -6.74551368e-01
-7.22436965e-01 -2.00975358e-01 -1.65958238e+00 1.07224703e+00
7.42087901e-01 -4.78691101e-01 -2.27665156e-01 2.72971779e-01
7.05356479e-01 -3.26975435e-01 -3.45480621e-01 3.62841398e-01
8.00532639e-01 -5.11483014e-01 1.11223626e+00 -8.60704184e-01
1.12357559e-02 -8.48306000e-01 -3.84286255e-01 -1.22510004e+00
2.75767267e-01 -9.31344211e-01 1.42783061e-01 1.02370787e+00
5.25946796e-01 -1.11912918e+00 5.54811835e-01 6.47117019e-01
1.53848454e-01 -9.42645893e-02 -8.62751424e-01 -1.23539507e+00
-1.04909688e-01 -8.02389860e-01 8.17954659e-01 7.58526504e-01
-6.76173806e-01 -2.71337301e-01 -5.83267093e-01 6.88063741e-01
8.15463424e-01 -4.20789629e-01 1.47679520e+00 -6.39401197e-01
-3.18915188e-01 -2.78100520e-01 -1.15430069e+00 -3.96676332e-01
3.01182181e-01 -5.06729662e-01 6.57795146e-02 -8.13185394e-01
-3.76941174e-01 -5.18718481e-01 8.85573179e-02 3.69116396e-01
1.05551137e-02 6.85908735e-01 4.79583174e-01 -1.61575601e-02
-2.57102519e-01 1.96516603e-01 1.01195610e+00 -6.09686494e-01
3.15186232e-01 1.65172666e-01 -4.32200700e-01 6.97442293e-01
8.86975586e-01 -4.95489061e-01 -5.12324393e-01 -2.38225892e-01
-1.81266412e-01 -3.70716214e-01 7.48248041e-01 -9.17882025e-01
-1.75344810e-01 -5.57512283e-01 8.86040404e-02 -1.75062999e-01
5.05694509e-01 -1.29799247e+00 3.48042279e-01 5.97679794e-01
-8.76132175e-02 3.76470424e-02 4.18213546e-01 5.45387924e-01
6.48205280e-02 1.14696182e-01 1.00400603e+00 2.92786896e-01
-9.59007919e-01 1.42248467e-01 -5.16289532e-01 -2.01542318e-01
1.65602529e+00 -5.42694092e-01 -6.41800523e-01 -7.06284821e-01
-4.72242206e-01 2.19178677e-01 8.87130082e-01 7.49889910e-01
6.30005658e-01 -1.34672654e+00 -6.33811951e-01 5.38377225e-01
3.13476920e-01 -6.40634894e-01 -9.88245830e-02 3.15844446e-01
-9.05415952e-01 4.03443091e-02 -3.89241695e-01 -4.48646486e-01
-1.64213026e+00 8.22699368e-01 4.90876406e-01 1.17487110e-01
-4.52145398e-01 2.92292982e-01 1.21164866e-01 -4.50343013e-01
2.19622552e-01 6.53095320e-02 1.86262384e-01 -7.01527357e-01
4.42087382e-01 1.72140375e-01 -1.07841164e-01 -1.08323133e+00
-4.84755665e-01 4.06464577e-01 2.98739877e-02 -2.00937465e-01
7.24683881e-01 1.59841448e-01 4.85528946e-01 -2.66523004e-01
1.27432704e+00 4.13693428e-01 -1.23692656e+00 1.88913316e-01
-4.65859354e-01 -8.06903303e-01 -3.38841587e-01 -8.03614080e-01
-1.22408044e+00 2.11385086e-01 8.21805835e-01 1.24559999e-01
7.66331613e-01 -2.64537424e-01 8.55199575e-01 6.04010463e-01
4.21769321e-01 -8.09544742e-01 -2.00410172e-01 4.07782555e-01
8.48966300e-01 -1.26333237e+00 -2.68315315e-01 -3.26364070e-01
-6.92092896e-01 6.71151936e-01 5.17565668e-01 -3.99699122e-01
4.25550789e-01 5.72290957e-01 7.74935544e-01 -5.03711589e-02
-4.29258734e-01 4.86281738e-02 -5.44851040e-03 1.11985946e+00
-6.55039907e-01 1.55090913e-01 1.27986833e-01 -2.76546448e-01
-4.90217537e-01 -5.40317774e-01 8.86174977e-01 9.80044603e-01
-9.62437037e-03 -1.08617163e+00 -9.58460867e-01 5.83035462e-02
-3.15358579e-01 3.56366336e-01 -8.21484983e-01 1.00006342e+00
2.04549178e-01 1.27518082e+00 -2.69799650e-01 -7.97708750e-01
6.27837002e-01 -1.56380162e-01 2.76650757e-01 6.63855895e-02
-3.14360023e-01 -6.89229012e-01 4.53136712e-01 -7.53147185e-01
-8.57593268e-02 -6.45103812e-01 -1.05503893e+00 -6.53207660e-01
-1.94630176e-01 -2.00890377e-01 7.98223913e-01 8.61276925e-01
1.60870448e-01 3.27668875e-01 1.24517572e+00 -7.85511374e-01
-6.77282929e-01 -3.30939114e-01 -1.28471851e-01 7.75051236e-01
5.20461917e-01 -6.52445912e-01 -6.79530799e-01 4.21693809e-02] | [5.437412261962891, 7.852640151977539] |
320d8ad9-52f2-4b38-95ba-4af6fe59a9ac | vlg-general-video-recognition-with-web | 2212.01638 | null | https://arxiv.org/abs/2212.01638v1 | https://arxiv.org/pdf/2212.01638v1.pdf | VLG: General Video Recognition with Web Textual Knowledge | Video recognition in an open and dynamic world is quite challenging, as we need to handle different settings such as close-set, long-tail, few-shot and open-set. By leveraging semantic knowledge from noisy text descriptions crawled from the Internet, we focus on the general video recognition (GVR) problem of solving different recognition tasks within a unified framework. The core contribution of this paper is twofold. First, we build a comprehensive video recognition benchmark of Kinetics-GVR, including four sub-task datasets to cover the mentioned settings. To facilitate the research of GVR, we propose to utilize external textual knowledge from the Internet and provide multi-source text descriptions for all action classes. Second, inspired by the flexibility of language representation, we present a unified visual-linguistic framework (VLG) to solve the problem of GVR by an effective two-stage training paradigm. Our VLG is first pre-trained on video and language datasets to learn a shared feature space, and then devises a flexible bi-modal attention head to collaborate high-level semantic concepts under different settings. Extensive results show that our VLG obtains the state-of-the-art performance under four settings. The superior performance demonstrates the effectiveness and generalization ability of our proposed framework. We hope our work makes a step towards the general video recognition and could serve as a baseline for future research. The code and models will be available at https://github.com/MCG-NJU/VLG. | ['LiMin Wang', 'Wayne Wu', 'Wenhai Wang', 'Zhaoyang Liu', 'Jintao Lin'] | 2022-12-03 | null | null | null | null | ['video-recognition'] | ['computer-vision'] | [-6.60175532e-02 -7.41734982e-01 -3.97976696e-01 -3.17221671e-01
-9.28381920e-01 -3.49864423e-01 6.14041805e-01 -7.70354629e-01
-3.63712162e-01 3.49343419e-01 3.61662030e-01 -1.28858715e-01
1.76867828e-01 -4.74042296e-01 -7.44745493e-01 -5.67246616e-01
7.42360875e-02 7.72315636e-02 1.38847575e-01 -1.40403226e-01
1.11431584e-01 1.02267615e-01 -1.60977745e+00 5.08970737e-01
4.20498103e-01 1.09110463e+00 5.94566286e-01 5.70988417e-01
-1.43737122e-01 9.81068671e-01 -2.55710781e-01 -4.60608572e-01
3.30203921e-01 -3.30647528e-01 -8.21052551e-01 5.70799410e-01
2.33039662e-01 -4.58013713e-01 -9.51894104e-01 1.01282001e+00
4.67085838e-01 5.06947279e-01 5.90857744e-01 -1.48070812e+00
-1.20472574e+00 3.58691871e-01 -5.83845496e-01 4.50441957e-01
7.34822631e-01 4.18867707e-01 1.06234038e+00 -1.22318900e+00
6.81643307e-01 1.33448410e+00 3.32334489e-01 8.16278756e-01
-5.14309466e-01 -7.12687433e-01 7.11963713e-01 6.13648891e-01
-1.71123648e+00 -5.73911428e-01 5.24806857e-01 -3.79605234e-01
8.80865455e-01 2.51503587e-01 4.59035367e-01 1.76363742e+00
-2.20445871e-01 1.30198860e+00 6.83204770e-01 -1.63021505e-01
-1.17244184e-01 -1.16202459e-01 2.59712994e-01 7.58551002e-01
4.71825041e-02 -3.37880880e-01 -6.22276008e-01 1.01195566e-01
6.96921647e-01 4.61530656e-01 -6.23147905e-01 -1.78837687e-01
-1.36989546e+00 8.86146247e-01 6.56445138e-03 4.58101213e-01
-6.83173165e-02 2.36675024e-01 7.11143196e-01 4.46053028e-01
4.82002079e-01 -1.68711796e-01 -3.55187953e-01 -4.90378171e-01
-6.85928822e-01 -4.71964590e-02 7.40061998e-01 1.31501698e+00
5.56426585e-01 -1.25794681e-02 -3.79210800e-01 1.03016675e+00
6.36086285e-01 7.58646846e-01 8.50467980e-01 -7.83673882e-01
7.87834764e-01 1.93658978e-01 -1.07801229e-01 -1.09386683e+00
7.31203780e-02 -7.77340010e-02 -7.30601549e-01 -6.02897227e-01
1.40079543e-01 8.23600292e-02 -9.52313125e-01 1.52505469e+00
2.34466381e-02 5.24214566e-01 3.26127082e-01 1.22244394e+00
1.31076610e+00 1.01490283e+00 -3.90781574e-02 -1.42303169e-01
1.38362622e+00 -1.49306023e+00 -6.81374669e-01 -1.74283847e-01
6.73326313e-01 -4.82762933e-01 1.23957634e+00 2.08540007e-01
-7.01016605e-01 -4.82580155e-01 -6.90359533e-01 -1.87784791e-01
-4.84923810e-01 2.80787706e-01 3.16175878e-01 2.31986076e-01
-1.10142565e+00 7.44689330e-02 -7.14390337e-01 -7.67861664e-01
5.54391921e-01 -1.44700184e-01 -6.19513929e-01 -6.43912852e-01
-1.24543810e+00 4.91577536e-01 3.32844824e-01 3.22849125e-01
-1.30311978e+00 -1.92998677e-01 -1.05505216e+00 -1.50304377e-01
9.94919777e-01 -6.84584618e-01 1.13509977e+00 -1.25681734e+00
-1.30817175e+00 1.00797844e+00 -2.65268087e-01 -1.02248929e-01
6.35289013e-01 -2.91756511e-01 -4.41383690e-01 2.70064116e-01
2.31219932e-01 2.71071076e-01 7.84681618e-01 -1.08889461e+00
-3.45805734e-01 -3.48254710e-01 2.36758322e-01 3.11283886e-01
-5.26911080e-01 4.36859727e-01 -1.31285465e+00 -9.16050017e-01
-2.22112164e-01 -8.30640912e-01 2.60025542e-02 -2.19757661e-01
-2.82053739e-01 -3.98161173e-01 7.52755761e-01 -7.83686876e-01
1.29988325e+00 -2.35314131e+00 2.73792118e-01 -2.05275789e-01
1.44387305e-01 4.30636048e-01 -4.89546686e-01 6.22736394e-01
2.02436130e-02 2.84960598e-01 9.95707810e-02 -3.70497406e-01
1.17425464e-01 2.19249383e-01 -3.31248581e-01 4.99247462e-01
-7.24619105e-02 1.06156921e+00 -8.90860736e-01 -4.25703317e-01
7.23200440e-02 4.56241906e-01 -2.80363053e-01 4.51322168e-01
-1.95636958e-01 2.25954831e-01 -8.87589633e-01 9.54602063e-01
3.65806520e-01 -5.65419495e-01 -3.87146808e-02 -3.40312332e-01
9.27378759e-02 -2.34636396e-01 -1.20070136e+00 1.99492872e+00
-3.37044299e-01 5.32143712e-01 -4.87709455e-02 -1.35874462e+00
7.66899109e-01 3.51440519e-01 3.30933362e-01 -5.65991938e-01
2.30748907e-01 -7.12271605e-04 -5.01134217e-01 -1.07158387e+00
2.60875970e-01 2.59789228e-01 -6.78107962e-02 2.07097620e-01
3.40758592e-01 6.15337610e-01 2.09412441e-01 5.23082495e-01
1.09549665e+00 2.97626108e-01 3.32522839e-01 8.70120674e-02
8.25190604e-01 -3.14085394e-01 6.21909201e-01 9.26546693e-01
-4.91264462e-01 6.83462560e-01 3.29655886e-01 -4.88451958e-01
-7.00152516e-01 -7.10260808e-01 2.37127185e-01 1.34188318e+00
3.29678893e-01 -6.91808760e-01 -4.73911166e-01 -7.51803041e-01
-1.70781180e-01 2.92718768e-01 -5.54559410e-01 -2.23961119e-02
-3.09059560e-01 -5.23829997e-01 4.74536479e-01 5.74814320e-01
7.71024346e-01 -1.02563381e+00 -4.80162278e-02 -1.83446720e-01
-6.97282553e-01 -1.54030323e+00 -7.55765080e-01 -5.36540031e-01
-3.58829707e-01 -1.04126287e+00 -9.16911662e-01 -9.04820859e-01
3.78146827e-01 8.83821309e-01 9.86028373e-01 4.13669378e-01
-2.44708151e-01 1.12402606e+00 -9.95968103e-01 1.49223030e-01
1.17416441e-01 -1.30784377e-01 1.15133390e-01 4.82371956e-01
5.99789798e-01 -1.63605541e-01 -4.41961378e-01 5.64487159e-01
-9.39739466e-01 1.78674400e-01 2.80109286e-01 7.98680067e-01
5.75931132e-01 -2.38135085e-01 5.55630565e-01 -4.72942412e-01
4.23564970e-01 -9.29767668e-01 -2.10428506e-01 5.65243363e-01
-3.73741649e-02 -3.64405543e-01 5.74191511e-01 -5.03548622e-01
-9.83285189e-01 -1.83756486e-01 -1.45983562e-01 -9.99841094e-01
-2.79264212e-01 6.25252068e-01 -5.33002734e-01 -1.04255257e-02
-8.58595148e-02 7.37830520e-01 -7.18189999e-02 -4.85007226e-01
4.86362547e-01 1.09001803e+00 3.15167814e-01 -6.98007882e-01
7.31703579e-01 5.27912378e-01 -6.40527844e-01 -9.95552838e-01
-8.79450262e-01 -8.44069064e-01 -3.31821591e-01 -4.75372821e-01
9.65040267e-01 -1.41688287e+00 -5.20392835e-01 6.07706010e-01
-1.01176584e+00 -3.48719269e-01 2.27353230e-01 3.67457747e-01
-6.73822999e-01 7.85872757e-01 -6.72015190e-01 -7.17815042e-01
-2.29882836e-01 -1.37845576e+00 1.08794200e+00 4.68588397e-02
4.46448952e-01 -1.01470995e+00 -1.78150311e-01 8.98800135e-01
2.03472450e-01 -1.68680504e-01 1.93929598e-02 -1.03289020e+00
-7.83012271e-01 -7.75022283e-02 -3.77401322e-01 3.87794882e-01
9.10492167e-02 -3.80923562e-02 -7.07697809e-01 -4.50315893e-01
-6.92608729e-02 -8.58868122e-01 1.18822205e+00 9.77732986e-03
1.47580409e+00 -2.40156278e-01 -3.31850082e-01 7.96583533e-01
1.43924832e+00 5.82536422e-02 6.22693658e-01 4.91896868e-01
9.71335888e-01 2.68344104e-01 7.83452809e-01 5.70093930e-01
7.86977410e-01 6.42549634e-01 1.64746925e-01 1.67229757e-01
-8.08544084e-02 -2.65215397e-01 8.81967902e-01 9.94601011e-01
-2.26798505e-01 -6.22828245e-01 -9.35942173e-01 5.12994826e-01
-2.15403557e+00 -1.21879983e+00 1.99929804e-01 1.77550673e+00
4.48547006e-01 -3.09756547e-01 1.11198328e-01 -4.34167385e-01
7.65088320e-01 4.84950870e-01 -3.72796088e-01 9.85138044e-02
-1.49775982e-01 -5.48365176e-01 2.59070814e-01 1.30668864e-01
-1.23218310e+00 1.16737866e+00 5.13869286e+00 1.17347097e+00
-9.79516208e-01 3.10007513e-01 5.00106633e-01 -1.87774062e-01
-1.74615812e-02 -2.58627206e-01 -9.46491361e-01 7.00147212e-01
9.20114815e-01 -2.69175738e-01 5.99123359e-01 7.71497488e-01
3.19880307e-01 3.29352021e-01 -1.00996995e+00 1.58992875e+00
8.16312015e-01 -1.25334847e+00 4.18485552e-01 -1.95849955e-01
4.58554566e-01 4.49641496e-01 -1.11540601e-01 6.37528062e-01
1.83851480e-01 -8.63874495e-01 6.30151391e-01 5.55050910e-01
8.49749267e-01 -2.77450711e-01 6.80456221e-01 3.98571879e-01
-1.58516347e+00 -2.27863520e-01 -4.45690781e-01 4.40606363e-02
2.05424815e-01 -4.04524989e-03 -6.22603483e-02 8.69874477e-01
9.23752606e-01 1.48282850e+00 -5.74990571e-01 8.51830781e-01
-1.83139984e-02 6.27966940e-01 4.56747226e-02 9.50304698e-03
4.07447875e-01 -2.94951618e-01 4.96219009e-01 1.45390177e+00
2.91509300e-01 2.88566947e-01 7.67004192e-01 5.78236043e-01
-2.81720757e-01 3.56089443e-01 -8.69278312e-01 -3.41611236e-01
3.32010299e-01 1.10281622e+00 -5.91060400e-01 -5.86564600e-01
-1.15880239e+00 1.13176024e+00 4.50212866e-01 6.58635139e-01
-1.04138029e+00 -1.68331608e-01 6.06727540e-01 -1.82943389e-01
6.42680705e-01 -2.04234257e-01 5.83407879e-01 -2.10188770e+00
2.28456959e-01 -1.16175401e+00 6.71845376e-01 -8.76852393e-01
-1.55416262e+00 6.40951216e-01 5.29263355e-02 -1.25682104e+00
6.96247146e-02 -8.21969032e-01 -3.77078444e-01 3.39102030e-01
-1.64915633e+00 -1.31977999e+00 -4.97869670e-01 1.12221456e+00
1.21613395e+00 -4.79857296e-01 4.85650688e-01 5.75708747e-01
-9.19300735e-01 5.77948570e-01 2.84400314e-01 6.78173780e-01
7.01882780e-01 -6.02562547e-01 2.48429239e-01 9.81832325e-01
3.10078651e-01 4.08905566e-01 2.28580013e-01 -6.30227685e-01
-1.96551216e+00 -1.29945600e+00 5.04696846e-01 -4.96807575e-01
1.09696400e+00 -4.58185703e-01 -8.69303346e-01 1.11182332e+00
1.04526319e-01 3.99919510e-01 7.83633590e-01 -1.52209625e-01
-6.63572907e-01 2.12846566e-02 -6.59627557e-01 6.03265464e-01
1.62273455e+00 -6.07211888e-01 -7.30685949e-01 5.56060672e-01
9.22404885e-01 -1.56887084e-01 -7.29197562e-01 1.64313346e-01
3.88021410e-01 -5.82075775e-01 9.47888196e-01 -9.52147365e-01
4.27573293e-01 -1.94112793e-01 -7.22364247e-01 -1.05763292e+00
-3.28218311e-01 -5.53539097e-01 -1.83551028e-01 1.42016339e+00
4.57853526e-02 -6.84210777e-01 2.54840136e-01 4.99722004e-01
-1.18741944e-01 -6.19983017e-01 -8.38396728e-01 -1.06409466e+00
-1.54323742e-01 -7.29785442e-01 9.89169106e-02 9.10982847e-01
-1.10471785e-01 2.81601369e-01 -8.88683200e-01 4.40653041e-02
5.15278101e-01 1.78207517e-01 8.11129093e-01 -6.12833977e-01
-2.52689689e-01 -1.70462132e-01 -6.55807793e-01 -1.45906842e+00
5.87518930e-01 -1.08126307e+00 -2.90450510e-02 -1.40982330e+00
7.93903947e-01 1.41392164e-02 -4.72448021e-01 5.19039333e-01
-1.59840077e-01 3.25311869e-01 6.38957024e-01 4.68452722e-01
-1.49247420e+00 8.71304095e-01 1.12505805e+00 -2.46454313e-01
2.74423331e-01 -2.81273901e-01 -7.27206469e-01 7.39109278e-01
5.92601538e-01 -1.43465981e-01 -4.15238231e-01 -7.74286151e-01
-6.69941753e-02 4.53037210e-02 4.52975750e-01 -6.63285077e-01
1.98119059e-01 -3.15277696e-01 3.59754786e-02 -2.96288848e-01
2.73497790e-01 -7.79723942e-01 -8.33781585e-02 -8.48253965e-02
-3.44871223e-01 4.04303521e-03 -1.16345018e-01 9.22199190e-01
-3.90748501e-01 -1.59359172e-01 3.33090365e-01 -3.98598343e-01
-1.34777176e+00 8.48622024e-01 -4.01042640e-01 4.70779628e-01
1.22823453e+00 -1.42973125e-01 -3.10623109e-01 -5.42328656e-01
-7.56278694e-01 5.75882435e-01 4.68325019e-01 9.89767492e-01
9.39894199e-01 -1.48401892e+00 -8.27609897e-01 3.69585527e-04
4.98224914e-01 -4.99079973e-01 5.56671917e-01 7.14986324e-01
-1.83540598e-01 5.11371434e-01 6.33075014e-02 -5.40104926e-01
-1.30010235e+00 8.36169541e-01 2.07852811e-01 2.40759626e-02
-9.12705302e-01 7.46897042e-01 4.37479645e-01 -1.65222973e-01
5.75679600e-01 6.27962947e-02 -3.27477187e-01 -3.24340053e-02
9.79122758e-01 1.37899399e-01 -3.83021176e-01 -1.10437667e+00
-5.55453777e-01 7.17874885e-01 -9.63377282e-02 3.32731366e-01
1.29630721e+00 -6.10968292e-01 4.89130020e-02 5.89858234e-01
1.49969923e+00 -4.79697764e-01 -1.10583854e+00 -5.48747659e-01
-2.76863277e-01 -8.05701673e-01 -2.15388127e-02 -4.85926300e-01
-1.33905184e+00 6.39835894e-01 2.86068827e-01 -3.29014063e-02
1.03597224e+00 3.23238194e-01 8.12313855e-01 5.55652440e-01
4.49509829e-01 -9.85625923e-01 3.17642868e-01 5.88466763e-01
1.06383419e+00 -1.55779088e+00 -2.98634171e-01 -2.67084301e-01
-9.80622113e-01 1.04580343e+00 7.24922240e-01 3.79936062e-02
6.38534367e-01 -9.98323858e-02 4.30191234e-02 -1.11277647e-01
-1.02555490e+00 -4.79130536e-01 3.10643345e-01 5.05530894e-01
3.31951588e-01 -8.84673074e-02 -2.30298951e-01 8.41599762e-01
4.79547977e-01 3.41090679e-01 4.71838921e-01 8.73384893e-01
-2.84251124e-01 -8.15348864e-01 -1.50285840e-01 3.30007643e-01
-5.21544337e-01 -8.31739306e-02 -1.13546327e-01 6.80785716e-01
-1.66105315e-01 1.02445674e+00 -1.69040114e-01 -4.82641339e-01
1.35197461e-01 1.07486360e-01 2.16273978e-01 -5.69390297e-01
2.79715899e-02 6.55365735e-02 6.73970357e-02 -9.96804833e-01
-7.13602066e-01 -6.79473639e-01 -9.53960776e-01 -1.09922417e-01
-1.83860928e-01 2.23798230e-02 2.05535293e-01 1.04925668e+00
4.46835726e-01 2.68829137e-01 5.68995416e-01 -7.06421018e-01
-6.79620445e-01 -8.07612956e-01 -5.06858766e-01 7.52941191e-01
2.94400811e-01 -8.44666541e-01 -5.85092366e-01 3.21197063e-01] | [10.186241149902344, 0.9179513454437256] |
2f7e6e5a-497e-4807-a572-b3f8e6cc46ee | wifi-motion-detection-a-study-into-efficacy | 1908.08476 | null | http://arxiv.org/abs/1908.08476v1 | http://arxiv.org/pdf/1908.08476v1.pdf | WiFi Motion Detection: A Study into Efficacy and Classification | WiFi and security pose both an issue and act as a growing presence in
everyday life. Today's motions detection implementations are severely lacking
in the areas of secrecy, scope, and cost. To combat this problem, we aim to
develop a motion detection system that utilizes WiFi Channel State Information
(CSI), which describes how a wireless signal propagates from the transmitter to
the receiver. The goal of this study is to develop a real-time motion detection
and classification system that is discreet, cost-effective, and easily
implementable. The system would only require an Ubuntu laptop with an Intel
Ultimate N WiFi Link 5300 and a standard router. The system will be developed
in two parts: (1) a robust system to track CSI variations in real-time, and (2)
an algorithm to classify the motion. The system used to track CSI variance in
real-time was completed in August 2018. Initial results show that introduction
of motion to a previously motionless area is detected with high confidence. We
present the development of (1) anomaly detection, utilizing the moving average
filter implemented in the initial program and/or unsupervised machine learning,
and (2) supervised machine learning algorithms to classify a set of simple
motions using a proposed feature extraction methods. Lastly, classification
methods such as Decision Tree, Naive Bayes, and Long Short-Term Memory can be
used to classify basic actions regardless of speed, location, or orientation. | [] | 2019-08-20 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 3.57079916e-02 -4.60546672e-01 -2.31796205e-01 -3.47923040e-02
-4.85287338e-01 -6.29987955e-01 1.35967672e-01 -3.06084901e-01
-2.81284541e-01 7.41362989e-01 -1.24373501e-02 -6.96096361e-01
-1.12370774e-01 -8.88160288e-01 -5.10934532e-01 -4.95326281e-01
-7.77347803e-01 -3.05790931e-01 6.48190677e-01 8.25860351e-02
6.64006397e-02 6.73355460e-01 -1.18852055e+00 9.64701176e-03
2.57531554e-01 1.45079279e+00 -1.57703385e-01 1.22914910e+00
3.70864123e-01 9.09764111e-01 -7.62346923e-01 1.83327571e-01
3.64570290e-01 -1.93335950e-01 -4.59587753e-01 -3.68360341e-01
1.20779745e-01 -7.34574020e-01 -7.15003550e-01 5.02784967e-01
6.72340631e-01 -1.52759671e-01 3.54715079e-01 -1.46138287e+00
2.66599804e-01 4.01231021e-01 -3.25110853e-01 5.11355698e-01
8.91357601e-01 -1.81019516e-03 1.81452602e-01 -2.87390053e-01
1.07670456e-01 8.06639671e-01 8.77685308e-01 1.55220866e-01
-6.03776455e-01 -1.13154268e+00 -1.77188680e-01 2.37338960e-01
-1.50567019e+00 -4.82205927e-01 4.93556529e-01 -2.55808592e-01
8.10458124e-01 4.13190305e-01 7.08335757e-01 1.43341637e+00
4.80553180e-01 5.77263832e-01 4.39461380e-01 -3.32561493e-01
5.54831803e-01 -8.16411823e-02 1.82192642e-02 5.10114074e-01
5.00501990e-01 1.67284504e-01 -5.98988950e-01 -5.82117677e-01
7.81415343e-01 -7.71319345e-02 -3.27853054e-01 -6.05953224e-02
-1.07979083e+00 3.08581322e-01 -1.15991972e-01 2.50115365e-01
-3.08857828e-01 6.18790209e-01 7.32151940e-02 3.39662701e-01
-2.09521368e-01 7.63348415e-02 -4.58689779e-01 -8.31351757e-01
-1.03424656e+00 -1.89971492e-01 1.15259385e+00 1.14150071e+00
3.57919544e-01 2.28019267e-01 1.08810149e-01 -7.33646527e-02
7.58825064e-01 1.07169616e+00 3.43251765e-01 -1.11320293e+00
4.04058784e-01 -1.34534210e-01 3.39004248e-01 -1.29473627e+00
-5.06716013e-01 -5.40986180e-01 -6.45307958e-01 1.21242605e-01
3.43588680e-01 -7.72177637e-01 -5.00449717e-01 1.51805818e+00
2.09754303e-01 6.83348775e-01 -1.22446874e-02 4.67028409e-01
2.42468059e-01 4.96998757e-01 -3.24207813e-01 -2.24292979e-01
1.00085998e+00 -4.93572533e-01 -7.89015830e-01 -1.53622061e-01
7.38697350e-01 -8.55289102e-01 2.23223299e-01 6.59656048e-01
-6.45855069e-01 -4.39739376e-01 -1.38604712e+00 9.01665688e-01
-1.41654834e-01 9.00847241e-02 5.97498119e-01 1.62995756e+00
-7.68364072e-01 3.51598650e-01 -1.25157738e+00 -5.99448144e-01
2.38997504e-01 7.52525210e-01 -6.85173869e-02 1.94584623e-01
-1.32443237e+00 5.02729833e-01 -1.29467681e-01 -1.01529419e-01
-4.93898809e-01 -2.03912497e-01 -7.99951255e-01 -4.57964987e-02
5.05463108e-02 -6.44564629e-01 1.18627393e+00 -8.81987453e-01
-1.51013231e+00 -8.01957697e-02 -1.40837237e-01 -5.26202083e-01
2.09871560e-01 -2.88911343e-01 -1.28122401e+00 1.77854791e-01
3.05061251e-01 -5.77784814e-02 8.36366057e-01 -6.85296416e-01
-1.14712119e+00 -1.71484366e-01 -1.71383634e-01 -1.67103752e-01
-3.65042299e-01 -3.31545830e-01 -3.19996804e-01 -6.92140281e-01
2.27729738e-01 -1.09844100e+00 -1.18042059e-01 -6.92720413e-02
-2.47424409e-01 7.17547476e-01 1.33357334e+00 -5.63822210e-01
1.66436040e+00 -2.06230426e+00 -9.34439957e-01 9.25345480e-01
-2.81193495e-01 2.50963390e-01 3.57779115e-01 2.90291190e-01
1.96296871e-01 -8.44179764e-02 3.51761222e-01 2.53982335e-01
-2.68458575e-01 1.52298927e-01 -2.71695793e-01 7.32709110e-01
-3.54700267e-01 2.31826782e-01 -8.30307424e-01 -1.15522482e-01
3.30812782e-01 1.26946047e-01 -5.75571120e-01 8.02775472e-03
6.27209783e-01 4.19604897e-01 -7.12751210e-01 8.11374128e-01
6.54084444e-01 5.51140234e-02 1.05853923e-01 -2.91113049e-01
-3.04364502e-01 2.67828971e-01 -1.82389629e+00 1.47955453e+00
-2.91558653e-01 7.38993585e-01 1.01996735e-01 -7.53311574e-01
5.83795965e-01 5.11195719e-01 8.06002080e-01 -5.21901786e-01
1.07570715e-01 9.92249548e-02 -2.51938492e-01 -8.36388111e-01
2.34862402e-01 6.48515522e-01 -3.71382415e-01 4.42307442e-01
-2.20313981e-01 6.31363928e-01 -1.93322212e-01 2.87654757e-01
2.09357738e+00 6.24715500e-02 3.46865863e-01 1.02289125e-01
4.76372629e-01 -1.05734348e-01 5.31032681e-01 1.11758912e+00
-4.66002613e-01 1.31579056e-01 -1.33209407e-01 -2.42645100e-01
-3.03235501e-01 -1.15969014e+00 1.95626438e-01 6.83097303e-01
5.37503958e-01 -4.46484238e-01 -3.08522135e-01 -3.85157168e-01
9.12079737e-02 3.91772985e-01 -5.44792153e-02 -2.53777415e-01
-4.28713769e-01 -4.93397683e-01 1.20632410e+00 5.24460375e-01
8.98249209e-01 -3.07750493e-01 -8.02337646e-01 4.56394732e-01
-2.30484754e-01 -1.23290670e+00 -4.05142128e-01 2.96753407e-01
-5.86311579e-01 -8.23754847e-01 -1.80374190e-01 -3.63561720e-01
2.22481713e-01 5.36323369e-01 4.62856025e-01 -1.00803584e-01
-2.41090730e-01 1.00129902e+00 -3.42621267e-01 -2.77383745e-01
1.71053335e-01 -1.04100123e-01 4.66232568e-01 1.35570124e-01
2.75955141e-01 -5.86205781e-01 -8.42460155e-01 5.64749181e-01
-3.35242450e-01 -8.29986036e-01 7.12589025e-01 3.01265001e-01
-1.23449370e-01 4.50006247e-01 1.18193462e-01 -2.95104027e-01
1.56375632e-01 -7.86882222e-01 -4.12733495e-01 -1.11615889e-01
-5.05763471e-01 -7.89315552e-02 1.59599438e-01 -3.36211056e-01
-8.63354564e-01 6.15547180e-01 -4.24740821e-01 1.75430924e-02
-3.04759443e-01 2.76736230e-01 -3.93501550e-01 -5.72479606e-01
6.93550348e-01 3.77375819e-02 -1.80257261e-01 -7.60287493e-02
2.30872650e-02 1.20797098e+00 6.97402775e-01 -1.76516846e-01
9.08411324e-01 9.28113997e-01 1.10079803e-01 -1.30963063e+00
-2.05438778e-01 -8.52919459e-01 -4.55235273e-01 -4.06362236e-01
4.88859326e-01 -1.14058661e+00 -7.42213607e-01 5.58503032e-01
-9.01321471e-01 -2.09424645e-01 3.72514397e-01 8.53711188e-01
-1.67272940e-01 6.96968436e-01 -4.32355970e-01 -9.81344283e-01
-1.95915669e-01 -6.90999746e-01 7.15242684e-01 2.71695882e-01
-6.33874953e-01 -7.35729337e-01 1.94309011e-01 2.56785929e-01
7.32094944e-01 1.24200538e-01 1.85859725e-02 -2.95352370e-01
-7.17327237e-01 -5.69635987e-01 2.17697755e-01 -5.59625775e-02
2.77262092e-01 1.05871856e-01 -9.38055038e-01 -4.41507757e-01
-2.08709808e-03 3.84456247e-01 2.71449387e-01 6.95512533e-01
9.04754519e-01 -3.98994237e-01 -9.30968821e-01 8.29866409e-01
1.04210472e+00 5.39887667e-01 8.79703939e-01 5.56160629e-01
3.22885782e-01 -1.23696610e-01 8.61802995e-01 7.31160045e-01
1.15862228e-01 7.91552365e-01 5.22821486e-01 1.53175965e-01
1.02587759e-01 -3.58660482e-02 8.14661562e-01 1.87267616e-01
-2.13780086e-02 -4.28959608e-01 -6.43298507e-01 7.70277604e-02
-1.81559050e+00 -1.42240453e+00 -5.86931050e-01 2.39835787e+00
-2.57476419e-02 5.42639852e-01 2.78697580e-01 4.71418202e-01
5.96955359e-01 -3.28314394e-01 -1.52428746e-01 -1.82217374e-01
3.48065883e-01 1.10328525e-01 1.37070191e+00 2.93435007e-01
-1.60018957e+00 4.84568417e-01 6.77522135e+00 5.78838885e-01
-1.39137506e+00 -7.87957534e-02 3.84114012e-02 1.16965719e-01
2.94646770e-01 -8.21244046e-02 -7.31789947e-01 7.83262253e-01
1.19685054e+00 3.27253014e-01 -3.04127741e-03 1.05795443e+00
6.29681587e-01 -3.58869851e-01 -6.16632998e-01 1.10376465e+00
-4.53939512e-02 -1.19146931e+00 -3.75647515e-01 1.38769388e-01
7.24855065e-02 5.71754128e-02 -1.45149544e-01 3.08580577e-01
3.39604132e-02 -6.35873914e-01 6.24726593e-01 5.01621008e-01
5.53466678e-01 -7.02811956e-01 7.81137288e-01 2.49370769e-01
-1.71223879e+00 -3.42198521e-01 1.18241236e-01 -5.85554540e-01
2.94693887e-01 5.38546681e-01 -5.23368061e-01 4.94669944e-01
9.77189660e-01 4.50289935e-01 -3.69682282e-01 1.42696846e+00
-4.99065705e-02 1.06351054e+00 -6.84418261e-01 -1.19606964e-01
-6.97204191e-03 3.60821724e-01 6.25053823e-01 1.25279057e+00
9.00643706e-01 1.33056700e-01 2.23265275e-01 7.24380417e-03
6.85495734e-01 -4.74976867e-01 -7.09678531e-01 4.71911401e-01
8.31616580e-01 1.01569104e+00 -6.49890780e-01 9.70865339e-02
-7.33136833e-01 1.04136860e+00 -8.29608977e-01 3.76805007e-01
-1.00557303e+00 -6.61916375e-01 7.00604200e-01 4.23785329e-01
3.76789421e-01 -6.21930420e-01 -1.21720187e-01 -8.89547646e-01
-1.17954850e-01 -5.50857067e-01 4.44677085e-01 -4.22297686e-01
-8.56738985e-01 1.94447026e-01 -2.60765016e-01 -1.67818618e+00
-4.76384699e-01 -5.22750020e-01 -6.64869547e-01 3.81164670e-01
-8.90763640e-01 -8.35586727e-01 -6.59846067e-01 1.04637158e+00
-1.05152670e-02 -3.85106176e-01 9.23061252e-01 9.68649089e-01
-4.85864729e-01 8.06185305e-01 1.98808625e-01 5.04406571e-01
6.99588776e-01 -8.81601155e-01 2.28561699e-01 1.15591347e+00
-2.79727858e-02 6.36480987e-01 8.01749170e-01 -8.64112318e-01
-1.72820866e+00 -1.02618277e+00 4.07035440e-01 -3.46936882e-01
6.42355382e-01 -1.78248063e-01 -1.21210426e-01 7.32644737e-01
-3.35627854e-01 2.01684177e-01 8.59957993e-01 -1.54907301e-01
1.89505607e-01 -4.65021968e-01 -1.13242519e+00 4.36021745e-01
8.15636158e-01 -3.14556628e-01 -1.96428046e-01 -2.03619346e-01
-9.60706696e-02 -3.71565908e-01 -6.01557434e-01 3.39610398e-01
1.19421625e+00 -7.76440561e-01 1.00105417e+00 7.24461675e-02
-7.77157068e-01 -6.11912310e-01 -3.49662572e-01 -6.08589649e-01
-3.93029004e-01 -1.06610096e+00 -4.45836335e-01 1.05592835e+00
4.52162355e-01 -5.92062593e-01 1.13090134e+00 4.84910280e-01
1.95407093e-01 -1.54736355e-01 -1.07376528e+00 -1.01482415e+00
-8.16139579e-01 -1.01885748e+00 1.30611479e-01 7.61851549e-01
2.36005992e-01 2.46260107e-01 -6.01349354e-01 8.40275824e-01
6.02994621e-01 -4.74037409e-01 1.13835311e+00 -8.30274880e-01
-6.06049299e-01 1.08945638e-01 -1.24349904e+00 -1.47375321e+00
-6.31983757e-01 -2.08744511e-01 -9.49193239e-02 -1.12914276e+00
-3.22347969e-01 -4.75825936e-01 -1.72646582e-01 3.93285602e-01
5.00466704e-01 2.90544350e-02 -4.86750335e-01 1.87620074e-01
-7.15755761e-01 2.93672606e-02 4.92046252e-02 -1.41171902e-01
-3.05073529e-01 9.73663867e-01 -2.45420054e-01 9.40535188e-01
9.26108658e-01 -4.85370278e-01 -3.99231821e-01 -1.01475175e-02
1.28901571e-01 1.54029369e-01 4.01838809e-01 -1.87324893e+00
6.31457150e-01 7.77899325e-02 7.64940321e-01 -5.34284890e-01
2.84079403e-01 -1.33017838e+00 4.08702433e-01 9.29626226e-01
3.79087836e-01 3.24461088e-02 1.89858779e-01 7.52232015e-01
1.83098316e-01 1.04517907e-01 3.81116688e-01 2.77046710e-01
-1.04401529e+00 6.64222538e-02 -1.27494335e+00 -5.37121415e-01
1.25913155e+00 -5.23783028e-01 -2.90033758e-01 -9.99488771e-01
-5.55005550e-01 6.14040904e-02 2.34215602e-01 3.33743244e-01
4.88493174e-01 -1.35714626e+00 -6.50515482e-02 2.50736326e-01
-1.97283894e-01 -8.04174006e-01 -7.14739636e-02 7.64923215e-01
-8.06750417e-01 4.79877323e-01 -2.85287827e-01 -7.37171054e-01
-1.40849006e+00 1.78031683e-01 3.38860512e-01 6.61184117e-02
-3.60078216e-01 5.29099107e-01 -7.19694197e-01 1.45917773e-01
6.66015983e-01 -2.09085077e-01 -1.32389098e-01 -3.78494382e-01
7.37168193e-01 6.45670474e-01 5.40807731e-02 -5.18667996e-01
-8.87505949e-01 5.85096002e-01 5.06523192e-01 -3.04149598e-01
6.62857890e-01 -5.95416427e-01 4.80562955e-01 1.09447956e-01
8.21846783e-01 5.20842552e-01 -9.27647948e-01 2.83710182e-01
2.23787233e-01 -3.36199731e-01 8.62248093e-02 -4.39915776e-01
-9.47431505e-01 2.40631163e-01 1.31718779e+00 2.56879449e-01
1.00501883e+00 -3.80124271e-01 1.00359690e+00 5.62116444e-01
9.04572189e-01 -9.90097761e-01 -4.68544252e-02 4.07629341e-01
-1.19375564e-01 -9.55048442e-01 1.06481768e-01 -4.20371562e-01
-7.42626786e-02 1.30587673e+00 3.60032558e-01 1.46954104e-01
1.19340062e+00 9.36638176e-01 1.47067055e-01 6.54919371e-02
-4.09661114e-01 -2.84954280e-01 -9.68538597e-02 1.01816738e+00
2.98435122e-01 -1.06597356e-01 -2.87749874e-03 8.06679964e-01
-1.19629093e-01 1.92284271e-01 6.20556056e-01 1.47605538e+00
-7.19768167e-01 -1.01431429e+00 -7.39567161e-01 5.12920082e-01
-7.97656953e-01 3.68981451e-01 -6.18264452e-02 5.06009042e-01
3.50466907e-01 1.49757648e+00 -4.98482548e-02 -9.03181434e-01
2.44694456e-01 -4.43661511e-01 1.68694884e-01 -2.27879092e-01
-1.16322218e-02 -4.79104370e-02 3.79439145e-01 -1.07504237e+00
-2.85644114e-01 -6.02444291e-01 -1.22082961e+00 -3.76657456e-01
-2.45964453e-01 7.78614283e-02 6.14910543e-01 1.11765385e+00
3.73550981e-01 4.40050453e-01 6.61184371e-01 -7.02899694e-01
-2.80979455e-01 -5.25913775e-01 -7.28234589e-01 -1.37507901e-01
2.00953260e-01 -5.94261944e-01 -4.31188017e-01 -1.26449525e-01] | [6.696852684020996, 0.7183588147163391] |
15c05a8f-9478-4c82-99fc-d81a7033dc38 | multi-view-3d-reconstruction-with | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Wang_Multi-View_3D_Reconstruction_With_Transformers_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Wang_Multi-View_3D_Reconstruction_With_Transformers_ICCV_2021_paper.pdf | Multi-View 3D Reconstruction With Transformers | Deep CNN-based methods have so far achieved the state of the art results in multi-view 3D object reconstruction. Despite the considerable progress, the two core modules of these methods - view feature extraction and multi-view fusion, are usually investigated separately, and the relations among multiple input views are rarely explored. Inspired by the recent great success in Transformer models, we reformulate the multi-view 3D reconstruction as a sequence-to-sequence prediction problem and propose a framework named 3D Volume Transformer. Unlike previous CNN-based methods using a separate design, we unify the feature extraction and view fusion in a single Transformer network. A natural advantage of our design lies in the exploration of view-to-view relationships using self-attention among multiple unordered inputs. On ShapeNet - a large-scale 3D reconstruction benchmark, our method achieves a new state-of-the-art accuracy in multi-view reconstruction with fewer parameters (70% less) than CNN-based methods. Experimental results also suggest the strong scaling capability of our method. Our code will be made publicly available. | ['Rabab Ward', 'Z. Jane Wang', 'Septimiu Salcudean', 'Tianyang Shi', 'Zhengxia Zou', 'Xun Chen', 'Xinrui Cui', 'Dan Wang'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['3d-object-reconstruction', 'object-reconstruction'] | ['computer-vision', 'computer-vision'] | [-9.37164798e-02 -2.40943328e-01 -3.38290036e-02 -4.49498296e-01
-9.10131693e-01 -5.47272921e-01 7.25717306e-01 -4.15420532e-01
1.94409057e-01 2.41936803e-01 6.96249843e-01 -1.72269315e-01
2.82870643e-02 -7.69958377e-01 -8.77990186e-01 -4.36061025e-01
2.63507664e-01 5.37861764e-01 2.65240312e-01 -3.18999141e-01
3.94913852e-02 7.01762617e-01 -1.57357490e+00 7.03836858e-01
3.04348469e-01 1.24788368e+00 2.16917053e-01 4.69554424e-01
-1.15962148e-01 5.82529843e-01 3.04444302e-02 -5.54292142e-01
4.34683532e-01 -2.95412511e-01 -7.73318470e-01 4.78570074e-01
7.06106603e-01 -7.10019886e-01 -4.81483072e-01 7.69899130e-01
8.42697024e-01 -3.13009411e-01 3.49040806e-01 -9.72864509e-01
-9.03700829e-01 3.44156116e-01 -5.77476680e-01 1.03581876e-01
4.78233308e-01 2.22118087e-02 1.08653796e+00 -1.39529276e+00
8.63704503e-01 1.27020788e+00 7.73931742e-01 3.27869952e-01
-1.18283606e+00 -3.76060992e-01 3.28107059e-01 1.95217803e-01
-1.09476483e+00 -4.87293422e-01 9.43914354e-01 -4.30204809e-01
1.41360474e+00 -5.25009707e-02 1.08456409e+00 1.11009955e+00
3.35621268e-01 1.08347690e+00 1.13721597e+00 4.32126261e-02
-2.22908631e-01 -2.51933157e-01 -4.21185076e-01 1.02221298e+00
-1.11869343e-01 2.58953333e-01 -6.06472015e-01 5.31781949e-02
1.06096685e+00 3.88648361e-01 -2.94191837e-01 -1.00781631e+00
-1.66656554e+00 8.40559304e-01 4.73862797e-01 4.65526693e-02
-2.22377464e-01 1.81853585e-02 7.01140046e-01 2.91403800e-01
9.08152580e-01 6.30886033e-02 -6.15724385e-01 -4.85050492e-02
-6.94890559e-01 3.74826759e-01 5.80915987e-01 1.14999235e+00
4.50474083e-01 6.45531388e-03 1.46408051e-01 7.08585739e-01
1.96946710e-01 2.85723448e-01 1.34524062e-01 -7.35817075e-01
6.09412432e-01 7.95521796e-01 -2.21513018e-01 -8.13209057e-01
-4.52118635e-01 -5.60410500e-01 -1.10311031e+00 1.45612359e-01
-3.89387496e-02 4.58188683e-01 -8.29235971e-01 1.28220749e+00
3.90927762e-01 -1.02222338e-01 -2.16711015e-01 9.39829826e-01
1.24493909e+00 3.21902871e-01 -5.73074579e-01 -1.30808026e-01
1.26481354e+00 -1.20733356e+00 -3.75561446e-01 9.52769518e-02
2.36380592e-01 -8.25957954e-01 7.58382797e-01 4.56151038e-01
-1.35101140e+00 -6.10454142e-01 -1.01978433e+00 -3.87603581e-01
-1.84878081e-01 -2.47499615e-01 7.18464434e-01 8.54393318e-02
-1.13620460e+00 6.08464122e-01 -7.67448604e-01 -3.31949860e-01
6.32743120e-01 1.88826814e-01 -8.41678619e-01 -2.72541910e-01
-7.23937988e-01 8.27165961e-01 -1.19727124e-02 -2.61288453e-02
-9.87515211e-01 -8.83085489e-01 -9.68074679e-01 -2.58124862e-02
5.20248830e-01 -1.30815887e+00 1.12100196e+00 -3.94058973e-01
-1.33012843e+00 1.18873453e+00 -1.94578454e-01 -4.60632816e-02
6.50124371e-01 -2.24589631e-01 -1.72388226e-01 1.05751634e-01
9.38811526e-02 5.06258368e-01 6.83643281e-01 -1.51997054e+00
-4.98548716e-01 -6.89237416e-01 1.61833778e-01 3.01754057e-01
5.61460704e-02 1.88927185e-02 -7.82805264e-01 -5.24800241e-01
5.73825061e-01 -7.81220913e-01 -1.65899485e-01 3.16151738e-01
-2.72491038e-01 -1.29725128e-01 7.64285684e-01 -3.98365557e-01
6.69750512e-01 -1.80514324e+00 6.05918646e-01 -3.61849815e-01
7.68841207e-01 -4.45124246e-02 -3.77054699e-02 4.47276264e-01
-3.33530396e-01 6.05635624e-03 -5.30947074e-02 -6.28654718e-01
-1.48943737e-01 1.04141653e-01 -3.26629281e-01 6.41389668e-01
-6.28682598e-02 1.28502750e+00 -8.79473567e-01 -3.22605610e-01
5.96529424e-01 5.83980381e-01 -7.66888976e-01 3.82134497e-01
-1.02786534e-01 5.44430435e-01 -3.61358404e-01 9.09592152e-01
7.40436256e-01 -7.91573763e-01 8.18121061e-02 -7.05763996e-01
-2.88018771e-02 2.94090897e-01 -7.31781900e-01 2.34399366e+00
-5.24090052e-01 3.91995877e-01 -7.91125819e-02 -9.56991494e-01
8.63826275e-01 3.49493444e-01 8.75439703e-01 -5.32103300e-01
8.74267891e-02 -6.25263005e-02 -3.18132102e-01 -4.20222968e-01
5.61678886e-01 -3.33519071e-01 1.67707071e-01 4.86312240e-01
3.34940374e-01 -4.23689693e-01 -2.92691052e-01 1.85379490e-01
7.33314633e-01 7.19642997e-01 5.51701546e-01 5.71258776e-02
2.44040981e-01 -2.90051311e-01 4.05413419e-01 2.21520454e-01
-1.08039357e-01 1.11113381e+00 3.98117334e-01 -8.92834306e-01
-1.37574780e+00 -1.11605036e+00 -4.74524088e-02 6.27284527e-01
1.07098453e-01 -6.78826928e-01 -1.40708849e-01 -9.84851539e-01
1.38660029e-01 1.53380141e-01 -7.40555763e-01 3.32067996e-01
-6.68927491e-01 -4.09054786e-01 9.45103467e-02 8.02517235e-01
4.57487047e-01 -6.92565084e-01 -4.80305105e-01 1.52245954e-01
-2.22526446e-01 -1.60939884e+00 -4.71831918e-01 7.82786757e-02
-1.21101439e+00 -9.85038638e-01 -8.07052672e-01 -5.88154733e-01
4.85732287e-01 6.82320416e-01 1.56227219e+00 -8.87216255e-02
-1.25626042e-01 5.60658574e-01 -4.44619328e-01 -2.37210169e-01
-1.82171643e-01 -6.60145730e-02 1.12655185e-01 -1.98516056e-01
4.67947833e-02 -1.03106260e+00 -7.11091340e-01 4.65769559e-01
-7.16117918e-01 4.55903441e-01 4.18972313e-01 9.52472031e-01
1.00679457e+00 -4.66376811e-01 2.36596569e-01 -7.63398349e-01
9.98575687e-02 -2.81226546e-01 -4.68914270e-01 2.66719252e-01
-5.35307527e-01 -1.35448128e-01 6.24535203e-01 -7.15757161e-02
-7.63755620e-01 1.80961132e-01 -3.56378973e-01 -1.10930490e+00
-2.79062271e-01 2.94578731e-01 -3.13990265e-01 -8.88578519e-02
1.30097002e-01 6.81810975e-01 1.20822921e-01 -6.72866642e-01
5.67401469e-01 9.43504274e-02 1.22504964e-01 -2.82229722e-01
7.27337360e-01 8.47559333e-01 6.06012456e-02 -5.14873087e-01
-1.05806291e+00 -3.20876181e-01 -9.53268468e-01 -3.31434548e-01
1.02757907e+00 -1.26442289e+00 -9.28879678e-01 4.41435277e-01
-1.18464494e+00 2.15667441e-01 -1.22738004e-01 2.90706247e-01
-9.93132532e-01 6.10060334e-01 -4.66089576e-01 -3.06257457e-01
-4.30046916e-01 -1.51364779e+00 1.57027256e+00 -3.27521980e-01
1.97098404e-01 -8.46572816e-01 -7.09661618e-02 7.03705609e-01
2.90196657e-01 3.15407902e-01 9.13868010e-01 -3.87747824e-01
-1.06010270e+00 8.91343206e-02 -4.14843261e-01 2.20158398e-01
1.50096893e-01 -2.86224782e-01 -9.79744554e-01 -4.50903714e-01
2.27152973e-01 -3.99894744e-01 9.69688237e-01 3.74021113e-01
1.40707600e+00 1.77697718e-01 -2.74045348e-01 1.12896848e+00
1.54130149e+00 -7.60627687e-02 5.59025466e-01 8.87766555e-02
1.06570101e+00 1.86277837e-01 2.11906955e-01 4.75752831e-01
7.03129470e-01 7.96538889e-01 9.28585947e-01 -2.05140546e-01
-2.64530808e-01 -4.60139841e-01 1.38952166e-01 1.37327671e+00
-2.94519782e-01 -3.51381868e-01 -5.84221125e-01 5.66220999e-01
-1.74237335e+00 -1.11543894e+00 1.41443238e-01 1.84368229e+00
1.85777694e-01 1.69257000e-01 1.75959378e-01 -1.97508588e-01
2.30418786e-01 6.51322126e-01 -6.61625803e-01 -9.24777016e-02
-1.39770538e-01 -5.87671734e-02 3.24715227e-01 1.84778795e-01
-1.11056721e+00 6.91289544e-01 6.67172575e+00 6.07397377e-01
-1.15998828e+00 9.76536274e-02 6.32293820e-01 -3.14470857e-01
-6.80378318e-01 -1.94174260e-01 -7.98223317e-01 -5.28862439e-02
2.08851427e-01 2.44406879e-01 3.69442582e-01 7.55610287e-01
-1.24959722e-01 2.36615390e-01 -1.31736481e+00 1.33636773e+00
4.44276929e-01 -1.92479599e+00 4.31713015e-01 3.61373484e-01
7.06831992e-01 4.92816776e-01 -5.48274629e-02 -2.46930867e-03
1.71299070e-01 -9.07051802e-01 8.34314227e-01 5.53587675e-01
9.44990337e-01 -7.34678984e-01 3.61027002e-01 2.46052489e-01
-1.63009322e+00 1.00914434e-01 -3.43358725e-01 2.08828375e-01
5.67753553e-01 6.34217262e-01 -5.23669422e-01 1.14490438e+00
8.85219514e-01 1.39886034e+00 -3.11109871e-01 7.12605596e-01
1.59484640e-01 4.07505147e-02 -5.20601962e-03 2.83157259e-01
1.75741181e-01 -1.25243410e-01 8.04657876e-01 7.21162796e-01
3.06483686e-01 3.04545537e-02 2.50744462e-01 8.97871017e-01
-6.41957074e-02 -1.24507353e-01 -1.09327126e+00 1.53725043e-01
-9.70418751e-02 1.17599034e+00 -7.07663000e-01 -2.94574171e-01
-9.50026631e-01 9.94327128e-01 6.07229173e-01 5.25772013e-02
-7.85188198e-01 2.60234326e-01 7.95550048e-01 2.73192286e-01
9.66164470e-01 -4.20452148e-01 -4.27497655e-01 -1.56033647e+00
1.91722944e-01 -8.55719268e-01 1.24252796e-01 -1.04118991e+00
-1.62553656e+00 8.81213307e-01 1.10211968e-02 -1.84427226e+00
-7.87035450e-02 -6.86814249e-01 -1.79315984e-01 7.36189961e-01
-1.45529842e+00 -1.61205280e+00 -1.67731062e-01 5.90585411e-01
9.34761584e-01 -2.87527502e-01 8.59756112e-01 3.90063822e-01
-1.04268417e-01 3.44746709e-01 -5.35521880e-02 7.06378892e-02
4.73089427e-01 -1.06334043e+00 8.17596018e-01 5.22971928e-01
1.93329826e-01 4.11725700e-01 8.86631012e-02 -5.67781210e-01
-1.93563950e+00 -9.54221308e-01 8.37008715e-01 -7.22191572e-01
4.62642401e-01 -5.71834385e-01 -5.94772398e-01 8.25539172e-01
3.45847011e-01 3.21399182e-01 6.66528761e-01 2.11643532e-01
-6.60693407e-01 1.33884698e-02 -8.59059870e-01 4.36025769e-01
1.71095765e+00 -7.92292655e-01 -5.49845755e-01 1.90958232e-01
1.00539076e+00 -8.25887442e-01 -1.19565451e+00 6.98995113e-01
8.15170407e-01 -1.51141012e+00 1.46596611e+00 -6.88957810e-01
9.61954772e-01 -1.61388248e-01 -5.78012645e-01 -1.27054799e+00
-6.55565500e-01 -2.64974236e-01 -3.21483672e-01 7.72652805e-01
1.09101586e-01 -4.04564917e-01 6.28741086e-01 9.26207285e-03
-4.70347196e-01 -1.43390608e+00 -9.83406663e-01 -5.54468334e-01
-5.80069944e-02 -4.15780783e-01 8.16269040e-01 8.53539646e-01
-3.92086536e-01 5.98232567e-01 -6.42630994e-01 -1.22583238e-02
6.75903797e-01 8.67162526e-01 9.78034258e-01 -1.11673546e+00
-4.09315288e-01 -4.14933324e-01 -4.24800962e-01 -1.45582783e+00
-2.40130741e-02 -1.17752659e+00 -2.55716294e-01 -1.73677528e+00
5.93503654e-01 2.30916850e-02 -4.42143269e-02 1.87459499e-01
7.82057270e-02 3.79281104e-01 4.22936350e-01 1.05438404e-01
-6.94124460e-01 9.37355161e-01 1.92237210e+00 -3.44535965e-03
2.21580535e-01 -1.12431921e-01 -7.01167464e-01 6.99456930e-01
1.52026847e-01 -1.58656985e-01 -3.32589507e-01 -7.37592876e-01
3.27454567e-01 5.53434372e-01 4.29041237e-01 -6.54039085e-01
-2.29847766e-02 6.83805123e-02 6.82953715e-01 -1.39786887e+00
6.78556442e-01 -9.54201460e-01 2.58986413e-01 1.18166856e-01
-1.96116474e-02 5.36673486e-01 6.68557584e-02 7.04616129e-01
-2.25247279e-01 5.51564157e-01 6.33719742e-01 -7.49683201e-01
-6.45578802e-01 8.50552678e-01 -4.61476203e-03 -3.61836776e-02
7.03816175e-01 -4.41970885e-01 -1.74422041e-01 -4.63735461e-01
-8.30433786e-01 -6.13876199e-03 6.02713525e-01 6.57972932e-01
1.00832283e+00 -1.53693295e+00 -7.39273667e-01 4.44629371e-01
3.19956750e-01 1.75339937e-01 5.17155051e-01 8.34459901e-01
-2.30958655e-01 4.24424529e-01 -3.06184381e-01 -9.98286486e-01
-1.25507975e+00 8.01015913e-01 3.69738191e-01 -4.20005113e-01
-1.10232651e+00 7.92244911e-01 5.33759654e-01 -6.72547102e-01
2.17795614e-02 -3.47801924e-01 -1.05815744e-02 -5.29331760e-03
3.13559681e-01 8.60022306e-02 2.31963381e-01 -7.63981760e-01
-3.44784856e-01 9.60223734e-01 -8.06258172e-02 2.08765015e-01
1.90405726e+00 -1.99675888e-01 -1.66043997e-01 5.22133350e-01
1.59103739e+00 -2.39876598e-01 -1.16064429e+00 -4.16934252e-01
-8.08306515e-01 -6.65700734e-01 6.76785409e-02 -6.87012613e-01
-1.40342021e+00 1.02180004e+00 4.52115238e-01 9.22731757e-02
1.18006802e+00 2.60238171e-01 1.09379947e+00 1.21932916e-01
5.57578981e-01 -4.64712888e-01 1.26681268e-01 8.49145234e-01
1.24276531e+00 -1.45766437e+00 4.03364182e-01 -4.09153581e-01
-5.64014792e-01 1.01373136e+00 7.30547726e-01 -1.70395732e-01
9.53717768e-01 2.52680451e-01 -9.55259055e-02 -6.28283858e-01
-8.96703959e-01 -8.70194063e-02 5.12931287e-01 4.32824999e-01
5.60790658e-01 -9.97140259e-02 1.99853376e-01 3.91084045e-01
-8.51759911e-02 -5.18350154e-02 1.56361476e-01 7.08303392e-01
-2.43343245e-02 -9.95655596e-01 -1.59992762e-02 4.98600394e-01
-3.76703769e-01 -1.17594793e-01 -3.16866696e-01 7.15035021e-01
-1.47799432e-01 3.56934220e-01 1.57432556e-01 -7.20267296e-01
5.68460822e-01 -8.47093910e-02 8.31741273e-01 -4.09673065e-01
-7.92790890e-01 3.40363026e-01 1.84407402e-02 -1.06211662e+00
-7.99712598e-01 -6.21370733e-01 -6.71443760e-01 -4.45287913e-01
-1.71472535e-01 -5.91140687e-01 5.16775548e-01 9.17228341e-01
6.89952433e-01 5.92586935e-01 7.70087600e-01 -1.23138845e+00
-4.39870209e-01 -5.79149485e-01 -3.74143720e-01 2.99191505e-01
4.20711488e-01 -8.29990447e-01 6.07992746e-02 -2.02912182e-01] | [8.261924743652344, -3.5866246223449707] |
009b8f99-4138-4589-b85a-003120c024ed | text-reading-order-in-uncontrolled-conditions | 2305.02577 | null | https://arxiv.org/abs/2305.02577v1 | https://arxiv.org/pdf/2305.02577v1.pdf | Text Reading Order in Uncontrolled Conditions by Sparse Graph Segmentation | Text reading order is a crucial aspect in the output of an OCR engine, with a large impact on downstream tasks. Its difficulty lies in the large variation of domain specific layout structures, and is further exacerbated by real-world image degradations such as perspective distortions. We propose a lightweight, scalable and generalizable approach to identify text reading order with a multi-modal, multi-task graph convolutional network (GCN) running on a sparse layout based graph. Predictions from the model provide hints of bidimensional relations among text lines and layout region structures, upon which a post-processing cluster-and-sort algorithm generates an ordered sequence of all the text lines. The model is language-agnostic and runs effectively across multi-language datasets that contain various types of images taken in uncontrolled conditions, and it is small enough to be deployed on virtually any platform including mobile devices. | ['Alessandro Bissacco', 'Yasuhisa Fujii', 'Renshen Wang'] | 2023-05-04 | null | null | null | null | ['optical-character-recognition'] | ['computer-vision'] | [ 4.92097199e-01 -4.72404122e-01 6.46945760e-02 -2.40077645e-01
-6.45832598e-01 -1.08066595e+00 4.53663915e-01 3.56571048e-01
-1.64958090e-02 -3.29988562e-02 3.54715616e-01 -6.90454245e-01
-2.47425511e-01 -4.89124566e-01 -7.87808657e-01 -2.86380917e-01
5.13524413e-02 4.54672813e-01 2.61107832e-01 -3.59295875e-01
6.13955081e-01 6.65787160e-01 -1.40623116e+00 5.31732678e-01
7.12337732e-01 8.94443333e-01 4.83287603e-01 1.27071285e+00
-1.78612486e-01 6.80357516e-01 -7.32032776e-01 -3.07828218e-01
4.19246942e-01 -1.32447422e-01 -5.62975049e-01 3.95594418e-01
1.05167341e+00 -3.86738777e-01 -4.57723141e-01 9.55034077e-01
7.13520944e-01 -4.96067554e-02 6.55544043e-01 -9.54632699e-01
-8.15145493e-01 3.32163543e-01 -7.29336143e-01 4.00326729e-01
3.52916569e-01 3.52299333e-01 1.27507627e+00 -7.62275159e-01
6.23157740e-01 1.24806523e+00 6.21154070e-01 -3.13139916e-03
-1.00555086e+00 -2.18417510e-01 2.46260598e-01 -5.78868836e-02
-1.17023122e+00 -4.51882631e-01 7.05055237e-01 -4.82354671e-01
1.23334622e+00 3.69241804e-01 4.33586299e-01 1.09705019e+00
4.41775143e-01 6.28961027e-01 8.52984369e-01 -3.28413010e-01
1.27000153e-01 -5.40063620e-01 -1.48597583e-01 8.35450768e-01
2.38281161e-01 -4.91071612e-01 -6.09315932e-01 1.48338497e-01
6.37968361e-01 -5.32451756e-02 -2.65332043e-01 -3.70360643e-01
-1.08174336e+00 2.99765110e-01 3.37338418e-01 1.20677158e-01
1.15891725e-01 1.58091590e-01 3.84664983e-01 3.87440562e-01
5.57977140e-01 4.60796148e-01 -5.24525821e-01 -2.18070507e-01
-1.30546987e+00 2.23321527e-01 6.18288279e-01 1.24393928e+00
7.56174505e-01 -3.48795682e-01 -1.65125832e-01 8.99982274e-01
9.88668948e-02 7.01805353e-01 2.76674449e-01 -4.52593625e-01
9.97073114e-01 6.95157230e-01 -2.70255357e-01 -1.40463591e+00
-6.75019801e-01 -3.77112120e-01 -7.73662210e-01 -3.52846049e-02
5.24119914e-01 -9.04799476e-02 -1.10329974e+00 1.11133683e+00
9.98517498e-03 -2.71654397e-01 -5.56037188e-01 8.04822922e-01
7.09222913e-01 5.04451334e-01 -3.41813475e-01 4.42868322e-01
1.40408897e+00 -1.06319416e+00 -2.59313643e-01 -8.28187227e-01
7.06224203e-01 -1.03239954e+00 1.39404023e+00 5.27862489e-01
-8.52579832e-01 -4.64351684e-01 -1.24165499e+00 -7.00779796e-01
-6.20105088e-01 3.56838912e-01 4.65323806e-01 5.23419023e-01
-1.36276817e+00 4.91307050e-01 -4.46963042e-01 -4.94703561e-01
4.98460561e-01 2.28212550e-01 -1.28564224e-01 -2.99741089e-01
-4.39563245e-01 4.77194607e-01 2.34436795e-01 3.23908716e-01
-4.74884033e-01 -6.87543631e-01 -8.05827856e-01 1.62983641e-01
4.89083499e-01 -3.55673522e-01 1.14083934e+00 -9.31698024e-01
-1.19221449e+00 8.40646744e-01 -2.91604161e-01 -1.31646479e-02
7.29414761e-01 -2.12568283e-01 -4.39954162e-01 1.70058325e-01
-4.39059362e-02 4.11071122e-01 1.30172133e+00 -1.11663198e+00
-5.97170591e-01 -4.22746271e-01 -1.27303690e-01 3.91067296e-01
-2.97808111e-01 6.88240603e-02 -9.22705650e-01 -5.41058719e-01
-5.79497442e-02 -7.78133929e-01 -8.60623419e-02 -1.20702712e-02
-7.49036789e-01 8.49262998e-02 1.06727529e+00 -8.37404490e-01
1.28765810e+00 -2.20622969e+00 2.54184008e-03 2.77188212e-01
4.27691281e-01 8.14435184e-02 -4.21841174e-01 5.66481650e-01
1.48038208e-01 3.48985106e-01 7.81424493e-02 -1.59058124e-01
-6.00410365e-02 -3.34286094e-01 -4.33841020e-01 3.30010355e-01
2.06223205e-01 1.09047651e+00 -8.32878530e-01 -3.08033645e-01
2.85617143e-01 7.72535279e-02 -2.97325075e-01 4.91247661e-02
-6.35290682e-01 -1.48710489e-01 -2.24255025e-01 7.51698077e-01
6.69983804e-01 -5.43401957e-01 3.33474696e-01 -1.57337725e-01
-7.47761950e-02 9.11139324e-02 -9.25174892e-01 1.78388631e+00
-4.34713721e-01 1.18551302e+00 -1.17688492e-01 -5.85708022e-01
8.41313303e-01 -3.52704048e-01 1.00170471e-01 -9.35479581e-01
1.42229542e-01 5.57431318e-02 -8.21035281e-02 -4.91512299e-01
1.01553714e+00 8.91325414e-01 -2.54261017e-01 5.04726827e-01
-1.26325384e-01 -2.62507707e-01 4.32899177e-01 4.59270775e-01
1.45688736e+00 -2.75733750e-02 7.88278058e-02 -3.17533076e-01
9.12198722e-02 -9.87204537e-02 6.61503477e-03 8.62822473e-01
1.24753945e-01 8.73147070e-01 8.89138281e-01 -5.20990431e-01
-1.51292348e+00 -9.87482250e-01 2.01927572e-01 1.33497751e+00
1.67314708e-01 -7.03473508e-01 -5.84156930e-01 -3.60593528e-01
-9.53776166e-02 2.09157810e-01 -3.07376593e-01 3.87731679e-02
-7.17868567e-01 -5.78219891e-01 5.69251120e-01 4.85355765e-01
4.31447387e-01 -8.38419855e-01 -4.31087852e-01 -8.95547960e-03
1.63798761e-02 -1.36724579e+00 -9.08907294e-01 3.14945579e-01
-5.82932234e-01 -1.14451110e+00 -5.07321537e-01 -8.75990689e-01
9.08042967e-01 4.88579869e-01 1.36853850e+00 2.55925387e-01
-5.71062684e-01 3.06731194e-01 -3.22822452e-01 -3.88637662e-01
-3.50573540e-01 4.00903881e-01 -4.73359376e-01 -4.59824465e-02
1.82653040e-01 -2.07357928e-01 -6.83921874e-01 1.48428962e-01
-9.52680588e-01 1.70459762e-01 5.86215973e-01 5.44544578e-01
4.09506291e-01 3.04421812e-01 -2.17299476e-01 -8.26221466e-01
1.02828252e+00 -1.16945058e-01 -9.74028826e-01 4.06091243e-01
-4.35094833e-01 6.36835620e-02 1.03254497e+00 -2.89664805e-01
-7.81325161e-01 8.19072779e-03 3.15611571e-01 -2.44235456e-01
-3.07735890e-01 3.09793115e-01 -2.67100751e-01 9.98007879e-03
5.84702671e-01 7.51801804e-02 -3.74923915e-01 -3.42746139e-01
5.09631872e-01 6.87940061e-01 6.42313421e-01 -4.30204123e-01
1.11460757e+00 2.62055427e-01 1.14077777e-01 -1.20268238e+00
-6.94952250e-01 -3.96666259e-01 -8.66055548e-01 -1.87534511e-01
7.28470027e-01 -7.98163474e-01 -4.30726141e-01 1.02068257e+00
-1.11389589e+00 -8.57779741e-01 4.29665208e-01 -4.09295946e-01
-2.73839176e-01 4.29594785e-01 -5.01291931e-01 -4.78611946e-01
-3.78880858e-01 -1.22651327e+00 1.51825368e+00 3.14888209e-01
-8.55977014e-02 -1.13440883e+00 -3.64443511e-01 3.78763020e-01
2.45545134e-01 -8.55467934e-03 1.42936981e+00 -4.04664725e-01
-9.85802889e-01 -3.10479458e-02 -6.09877527e-01 3.58284451e-02
9.56771076e-02 5.58413565e-01 -7.96783388e-01 -3.74756217e-01
-7.89253771e-01 -2.83762276e-01 7.23707139e-01 4.58675593e-01
1.61369669e+00 -1.09719388e-01 -2.80892164e-01 8.79751742e-01
1.56573224e+00 -7.98745677e-02 5.16786575e-01 4.98393506e-01
1.19773471e+00 3.71177554e-01 2.52235591e-01 3.39317888e-01
4.65577573e-01 4.95675534e-01 4.38673884e-01 -4.09585297e-01
-1.96306095e-01 -3.23096156e-01 2.20939115e-01 5.47579825e-01
4.36279297e-01 -9.31218743e-01 -1.17973399e+00 5.06074071e-01
-1.61966622e+00 -7.92856157e-01 -3.23275179e-01 2.01521826e+00
4.56127852e-01 3.18097234e-01 5.43617494e-02 -9.41749290e-02
7.63864756e-01 6.83615744e-01 -7.16592729e-01 -7.71482468e-01
-3.13868850e-01 7.83597901e-02 1.06093609e+00 1.82001486e-01
-1.14471805e+00 1.23369646e+00 6.56954193e+00 7.04040706e-01
-1.34463882e+00 -5.17745078e-01 8.19633067e-01 -1.80646762e-01
-3.18499684e-01 -2.37483159e-01 -8.12299013e-01 3.98528188e-01
7.72147119e-01 2.83090100e-02 8.69326651e-01 4.22539115e-01
5.04560649e-01 -2.91396916e-01 -1.05560255e+00 1.06171978e+00
3.19060028e-01 -1.40441287e+00 -3.78863700e-02 1.64605871e-01
6.27620935e-01 3.51939976e-01 3.16618353e-01 -2.80293822e-01
3.87817264e-01 -1.17679298e+00 9.85218585e-01 1.13453366e-01
1.24856901e+00 -6.57488406e-01 1.80252090e-01 1.07128404e-01
-1.17743790e+00 -2.69948691e-01 -4.03138727e-01 1.91195264e-01
5.05161881e-02 6.88803613e-01 -8.23601127e-01 3.40303779e-01
7.14377344e-01 6.90265536e-01 -1.22471988e+00 9.55574214e-01
-2.00893983e-01 3.78834039e-01 -1.76564351e-01 -2.52454162e-01
2.23423228e-01 -1.41008779e-01 1.53352916e-01 1.37968743e+00
3.74376118e-01 -4.88467127e-01 -3.50136384e-02 5.49860954e-01
-4.07874286e-01 1.96457818e-01 -7.66928136e-01 -3.46361935e-01
4.53955203e-01 1.54984128e+00 -1.22459149e+00 1.62143521e-02
-5.42640150e-01 1.31257284e+00 5.16097009e-01 5.11146307e-01
-7.40891278e-01 -4.42197502e-01 5.16244352e-01 6.11397885e-02
6.26694620e-01 -6.51264727e-01 -5.62356710e-01 -1.06258059e+00
4.34435576e-01 -1.17108953e+00 1.49419412e-01 -1.08324957e+00
-1.06123114e+00 5.01380682e-01 -3.53727221e-01 -1.06275606e+00
-5.10710217e-02 -1.11168754e+00 -7.52500892e-01 7.70994544e-01
-1.42091250e+00 -1.16966128e+00 -5.75767279e-01 5.44612110e-01
7.84282506e-01 -2.02257395e-01 4.43294644e-01 1.61083087e-01
-7.75015414e-01 7.87869096e-01 3.57588321e-01 3.81450206e-01
8.64195228e-01 -1.42170036e+00 9.94498014e-01 1.23890865e+00
3.66080195e-01 4.16171849e-01 5.43928683e-01 -6.07386112e-01
-1.75876272e+00 -1.31046975e+00 7.03594446e-01 -5.49397647e-01
8.98643732e-01 -1.03246534e+00 -5.90637565e-01 5.93818426e-01
1.94102198e-01 -1.75414592e-01 4.26254928e-01 3.03659350e-01
-6.02890372e-01 -1.74148083e-01 -5.77207804e-01 8.37644339e-01
1.28150737e+00 -7.67612278e-01 -9.04759690e-02 6.67294860e-01
5.70932388e-01 -9.26058352e-01 -5.16148746e-01 -2.38849461e-01
3.67559254e-01 -1.08355486e+00 6.98035002e-01 -3.16672415e-01
7.72709608e-01 -1.94886386e-01 1.53519884e-01 -1.40411770e+00
-3.33462179e-01 -9.04932559e-01 1.68384746e-01 1.04530549e+00
7.40748405e-01 -3.05930018e-01 6.16629660e-01 3.70357871e-01
-2.19063759e-01 -5.86580575e-01 -4.79247779e-01 -6.71501517e-01
-1.40238762e-01 -4.43711191e-01 6.29388034e-01 3.85824800e-01
-3.71819377e-01 4.15192336e-01 -2.64307678e-01 3.45620543e-01
3.70839030e-01 1.63459823e-01 8.69184673e-01 -1.14678586e+00
-2.88098514e-01 -5.84552228e-01 -2.88525522e-01 -1.31705713e+00
7.25075528e-02 -7.87857234e-01 5.38646728e-02 -1.54413450e+00
-2.25631353e-02 -2.46522427e-01 1.56340867e-01 3.17075998e-01
-7.13376924e-02 5.72244897e-02 1.51148453e-01 1.25639707e-01
-7.97033310e-01 2.13148980e-03 1.17206776e+00 -4.48202223e-01
-1.91994727e-01 -2.96675831e-01 -7.20435143e-01 5.67400694e-01
8.52058172e-01 -1.41618159e-02 -6.59496129e-01 -1.14939034e+00
8.05412114e-01 -3.20062906e-01 3.15001726e-01 -9.20908868e-01
3.17532957e-01 -1.04590543e-02 6.78965628e-01 -8.11458528e-01
-9.25999582e-02 -5.35645068e-01 -2.30570570e-01 -1.26720399e-01
-3.50597560e-01 5.21687746e-01 2.09392965e-01 5.00997961e-01
2.55098075e-01 -3.82616073e-02 5.95215142e-01 4.26630722e-03
-9.44401979e-01 4.69685733e-01 -2.36844674e-01 2.37616107e-01
6.40743017e-01 -4.10305202e-01 -7.70471036e-01 -3.33569437e-01
4.95979935e-02 1.86181843e-01 8.24521005e-01 9.15502369e-01
5.06550550e-01 -8.13716471e-01 -3.49936306e-01 3.75734776e-01
3.35197955e-01 2.23452851e-01 9.79373008e-02 3.31445634e-01
-1.00015438e+00 4.85347837e-01 2.34448095e-03 -7.40072250e-01
-1.26875925e+00 3.47452521e-01 3.35420609e-01 -2.84039974e-01
-7.92807221e-01 7.48552859e-01 5.27765416e-02 -1.13775142e-01
6.73352033e-02 -5.18274724e-01 1.36267945e-01 -5.24344444e-02
6.24924958e-01 2.16625258e-01 4.69551384e-01 -4.91302222e-01
-2.21375078e-01 6.87210441e-01 -2.12264761e-01 1.17426487e-02
1.14673650e+00 -2.11551905e-01 -1.84168383e-01 3.16884160e-01
1.20248973e+00 1.37092173e-01 -1.38189995e+00 5.84501140e-02
6.68165162e-02 -4.57830608e-01 1.51449308e-01 -7.20277429e-01
-8.62705588e-01 7.71033108e-01 3.26345921e-01 2.85300374e-01
1.18895864e+00 -3.00236866e-02 8.12690377e-01 4.40800279e-01
1.90902531e-01 -1.46439779e+00 2.49136776e-01 6.30746484e-01
7.82175303e-01 -1.14638257e+00 7.18456581e-02 -2.57777512e-01
-4.13449168e-01 1.44156921e+00 6.37994587e-01 1.62841335e-01
2.54977465e-01 5.90793371e-01 2.01406434e-01 -3.65281016e-01
-7.33141482e-01 -3.67373824e-02 5.01056671e-01 6.57477140e-01
5.18900335e-01 -9.34533104e-02 2.31090218e-01 -2.11047634e-01
-4.49724704e-01 -4.30401415e-01 6.06488287e-01 5.61020136e-01
-3.31572801e-01 -1.00869095e+00 -2.16476664e-01 8.22212756e-01
-2.38310516e-01 -3.68571192e-01 -5.47964513e-01 4.69285518e-01
-1.44801021e-01 9.77244914e-01 3.77429515e-01 -3.95116359e-01
2.33100235e-01 -3.28063667e-01 4.21068549e-01 -5.20814121e-01
-3.29248697e-01 5.42044528e-02 6.90195858e-02 -6.88619375e-01
2.50027746e-01 -6.60621285e-01 -9.35779035e-01 -3.69210631e-01
-8.28745123e-03 -7.16623425e-01 8.33824873e-01 8.99226904e-01
7.33155787e-01 5.58902681e-01 6.10325575e-01 -8.05096269e-01
-2.07613751e-01 -6.14880562e-01 -5.14054656e-01 4.36366916e-01
5.07725775e-01 -1.52413175e-01 -3.35401624e-01 1.34636937e-02] | [11.658778190612793, 2.3460381031036377] |
abfde15f-a8f1-4381-bdeb-51c4eeed3683 | anchor-retouching-via-model-interaction-for | 2112.06701 | null | https://arxiv.org/abs/2112.06701v1 | https://arxiv.org/pdf/2112.06701v1.pdf | Anchor Retouching via Model Interaction for Robust Object Detection in Aerial Images | Object detection has made tremendous strides in computer vision. Small object detection with appearance degradation is a prominent challenge, especially for aerial observations. To collect sufficient positive/negative samples for heuristic training, most object detectors preset region anchors in order to calculate Intersection-over-Union (IoU) against the ground-truthed data. In this case, small objects are frequently abandoned or mislabeled. In this paper, we present an effective Dynamic Enhancement Anchor (DEA) network to construct a novel training sample generator. Different from the other state-of-the-art techniques, the proposed network leverages a sample discriminator to realize interactive sample screening between an anchor-based unit and an anchor-free unit to generate eligible samples. Besides, multi-task joint training with a conservative anchor-based inference scheme enhances the performance of the proposed model while reducing computational complexity. The proposed scheme supports both oriented and horizontal object detection tasks. Extensive experiments on two challenging aerial benchmarks (i.e., DOTA and HRSC2016) indicate that our method achieves state-of-the-art performance in accuracy with moderate inference speed and computational overhead for training. On DOTA, our DEA-Net which integrated with the baseline of RoI-Transformer surpasses the advanced method by 0.40% mean-Average-Precision (mAP) for oriented object detection with a weaker backbone network (ResNet-101 vs ResNet-152) and 3.08% mean-Average-Precision (mAP) for horizontal object detection with the same backbone. Besides, our DEA-Net which integrated with the baseline of ReDet achieves the state-of-the-art performance by 80.37%. On HRSC2016, it surpasses the previous best model by 1.1% using only 3 horizontal anchors. | ['Huiyu Zhou', 'Mingqiang Wei', 'Ekaterina L. Kim', 'Dmitry A. Vorontsov', 'Zongqi Wei', 'Qixiang Geng', 'Dong Liang'] | 2021-12-13 | null | null | null | null | ['robust-object-detection', 'object-detection-in-aerial-images', 'small-object-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.47965419e-01 1.05472133e-01 -2.25373387e-01 -1.96553990e-01
-9.63624716e-01 -4.43585247e-01 3.51757079e-01 -3.10285371e-02
-7.00537324e-01 5.38522184e-01 -5.52130997e-01 -3.52293700e-01
1.67058215e-01 -8.05233777e-01 -9.35599923e-01 -8.38496506e-01
-6.39720121e-03 2.38236383e-01 9.56492782e-01 -4.15458530e-02
-5.53857237e-02 3.23623717e-01 -1.66957378e+00 3.84261250e-01
8.83222699e-01 1.48528242e+00 3.75419348e-01 5.23430645e-01
2.62375742e-01 2.72249848e-01 -6.19158626e-01 -3.70560974e-01
6.39984369e-01 -5.02287038e-02 -3.92920643e-01 1.26220658e-01
8.77678990e-01 -6.94449604e-01 -4.72804084e-02 1.15167999e+00
7.55224705e-01 -6.59847483e-02 6.41941309e-01 -1.34883571e+00
-4.43919510e-01 4.88900363e-01 -1.14966369e+00 2.57887304e-01
-1.31079853e-01 5.76903880e-01 8.76583815e-01 -1.20270395e+00
4.15025234e-01 1.08760667e+00 9.34363663e-01 4.85383153e-01
-1.25797987e+00 -8.63611698e-01 5.98291814e-01 -6.96001723e-02
-1.47777843e+00 -1.40979350e-01 3.14867795e-01 -3.92059147e-01
5.60872972e-01 3.00147682e-01 5.29817045e-01 9.09846604e-01
-7.90363550e-02 1.12173498e+00 9.13976192e-01 -1.53159946e-01
1.34017453e-01 2.34128967e-01 3.60248946e-02 9.36899364e-01
6.33844912e-01 1.71429336e-01 -3.39824930e-02 -7.64481211e-03
5.59126973e-01 -2.16423739e-02 -2.68811792e-01 -2.24063575e-01
-1.18607461e+00 5.48133492e-01 8.75468850e-01 -2.26486579e-01
-4.23939407e-01 -3.77258509e-02 4.41526830e-01 -4.34286930e-02
3.51190776e-01 3.23194265e-01 -3.96788478e-01 6.22894824e-01
-1.06956708e+00 3.65836143e-01 3.60405535e-01 1.23755777e+00
4.71104741e-01 2.41638839e-01 -8.65671813e-01 6.66340888e-01
1.96685314e-01 8.89567912e-01 1.52964920e-01 -5.78933656e-01
6.12338841e-01 8.10411334e-01 4.35171783e-01 -8.78948092e-01
-3.49384308e-01 -1.01245868e+00 -9.30809557e-01 5.42975128e-01
5.16930163e-01 -2.14059755e-01 -1.21630740e+00 1.40499306e+00
6.63880050e-01 5.44807650e-02 -3.00244521e-02 1.01462412e+00
8.84793401e-01 7.00399995e-01 1.54584184e-01 6.06364645e-02
1.56485224e+00 -1.08698428e+00 -1.53956369e-01 -5.66942751e-01
5.16820967e-01 -5.63560724e-01 1.07580185e+00 3.53915721e-01
-9.01606858e-01 -8.51984322e-01 -1.30720651e+00 1.94149867e-01
-2.24778205e-01 9.66710508e-01 4.26463544e-01 5.12275517e-01
-6.66785419e-01 2.38645568e-01 -6.80263758e-01 -2.92000324e-01
8.44486773e-01 3.17897022e-01 -1.37429059e-01 -4.34789136e-02
-6.52654827e-01 3.91587913e-01 5.19636512e-01 2.80867010e-01
-1.16485822e+00 -8.58022332e-01 -5.54521561e-01 -1.82599589e-01
8.51665020e-01 -7.39240885e-01 1.16417885e+00 -6.89119756e-01
-1.17971611e+00 7.66072571e-01 1.61245286e-01 -8.54651749e-01
8.85938466e-01 -3.29650372e-01 -1.36696354e-01 6.84891790e-02
2.59525776e-01 1.18832779e+00 9.25784230e-01 -1.26963723e+00
-1.41593373e+00 -2.61286587e-01 2.70369798e-01 1.03771180e-01
-2.99210548e-01 -1.60575300e-01 -5.70794404e-01 -5.93831658e-01
1.48538038e-01 -9.43657696e-01 -2.40283936e-01 6.32526338e-01
-5.16599178e-01 -3.36027533e-01 1.04909658e+00 -3.92730027e-01
1.21902144e+00 -2.07649899e+00 -2.53917158e-01 -1.04847595e-01
1.57646328e-01 6.40515506e-01 -2.18734428e-01 -1.90445364e-01
2.33791307e-01 -2.96181887e-02 -3.33652973e-01 -4.05705690e-01
-2.39497066e-01 -1.43673822e-01 -3.73644233e-01 5.65874934e-01
4.88190770e-01 6.63879454e-01 -9.82393682e-01 -6.35287881e-01
2.88103491e-01 2.06598774e-01 -6.17269933e-01 1.19967625e-01
-2.47996658e-01 6.12782547e-03 -4.89964575e-01 1.17401338e+00
8.30557525e-01 -2.90279388e-01 -1.36147499e-01 -5.45062363e-01
-2.24289641e-01 -2.20193848e-01 -1.43853712e+00 1.23384941e+00
-1.64910063e-01 4.30002451e-01 1.36198238e-01 -6.82238817e-01
9.05190825e-01 -5.42382896e-02 6.75149485e-02 -4.00547534e-01
1.73853248e-01 2.91421145e-01 4.23150137e-02 -2.06431925e-01
6.56190872e-01 4.46557641e-01 1.19559519e-01 -2.00334758e-01
-6.02481961e-02 1.61372945e-01 3.96481097e-01 4.13270928e-02
9.27626431e-01 2.73168445e-01 3.97299886e-01 -3.53858978e-01
5.24274468e-01 -1.64541882e-02 8.25842202e-01 1.03205299e+00
-3.23695183e-01 6.03437126e-01 2.50203431e-01 -5.59053957e-01
-8.11093330e-01 -9.61547017e-01 -2.60687977e-01 1.01285601e+00
4.02672052e-01 -1.32263869e-01 -7.58746266e-01 -1.01272082e+00
5.12281097e-02 5.19068778e-01 -5.86780488e-01 -3.23402286e-02
-3.09880435e-01 -1.04714525e+00 6.40870929e-01 7.11329937e-01
9.56768036e-01 -9.27630246e-01 -9.25174773e-01 4.93837222e-02
-1.84105888e-01 -1.31976485e+00 -3.12030911e-01 -5.31089455e-02
-7.49046326e-01 -1.12887800e+00 -8.03702950e-01 -5.56159914e-01
8.64837050e-01 6.64583743e-01 9.01159585e-01 5.48579842e-02
-5.17328620e-01 -5.31632714e-02 -3.91730547e-01 -6.85718894e-01
-5.64927757e-02 2.16539755e-01 6.26216382e-02 1.65324479e-01
3.74751166e-02 -5.81843443e-02 -8.96734118e-01 8.26209545e-01
-7.54768610e-01 1.56723008e-01 9.71828640e-01 9.23865557e-01
9.27250206e-01 -1.65957198e-01 5.42179585e-01 -5.83748341e-01
-1.90986112e-01 -3.67446601e-01 -1.14400733e+00 3.41637015e-01
-6.27761066e-01 -2.10246429e-01 4.90562260e-01 -6.47292078e-01
-9.32670832e-01 3.27582389e-01 8.83681476e-02 -5.67608595e-01
-6.45800494e-03 -4.91152667e-02 -1.02013238e-01 -2.07361966e-01
9.38652754e-01 1.43704355e-01 -2.73936093e-01 -1.47275895e-01
6.17084242e-02 6.77856624e-01 6.86130643e-01 -4.20909941e-01
9.20525789e-01 6.41915023e-01 4.72238176e-02 -7.16599464e-01
-1.19150722e+00 -5.89274228e-01 -3.35286975e-01 -3.06241393e-01
7.65778661e-01 -1.29440784e+00 -6.07245922e-01 6.68484569e-01
-1.07037902e+00 -4.36654627e-01 -3.36081445e-01 3.82444859e-01
-6.04383461e-02 -2.86449790e-02 -2.25109667e-01 -1.03758335e+00
-6.59530580e-01 -1.20031488e+00 1.38138044e+00 3.25820148e-01
3.53882134e-01 -1.78535849e-01 -7.30287135e-01 2.48280734e-01
1.75737157e-01 4.31736052e-01 2.75821507e-01 -4.66512442e-01
-9.15672600e-01 -4.21797007e-01 -7.01695383e-01 5.07940948e-01
-1.03116438e-01 9.70946327e-02 -1.08730018e+00 -4.79364187e-01
-4.99545604e-01 -3.85900438e-01 1.11084545e+00 5.06973743e-01
1.44580042e+00 -1.71725810e-01 -6.36502504e-01 7.41336882e-01
1.33457053e+00 5.31414561e-02 4.87047642e-01 3.51710945e-01
6.16968274e-01 3.02141160e-01 1.18105376e+00 5.73099792e-01
2.46385872e-01 6.93099916e-01 7.85699964e-01 -2.59820908e-01
-3.57717276e-01 -1.22299820e-01 3.70106697e-01 -1.46659836e-01
-2.69299656e-01 -4.03338492e-01 -6.22606218e-01 5.81643879e-01
-1.79005289e+00 -8.76854241e-01 -3.51247281e-01 2.28718257e+00
7.76765823e-01 4.87058461e-01 2.20381215e-01 2.40554959e-02
7.89886296e-01 1.83284447e-01 -7.30839133e-01 5.83321810e-01
-1.20062098e-01 -2.62852043e-01 1.05219376e+00 1.25855863e-01
-1.52408409e+00 9.30402994e-01 4.71582747e+00 1.10997379e+00
-9.71340120e-01 -6.43085092e-02 7.32772768e-01 1.86416227e-02
5.04482806e-01 -2.96430260e-01 -1.63187873e+00 5.12456834e-01
3.01583707e-01 2.79187232e-01 1.06039699e-02 1.41665626e+00
8.00509676e-02 -2.51726657e-01 -9.46991563e-01 8.95537734e-01
4.13000323e-02 -1.22418940e+00 5.73107786e-02 -1.13480560e-01
7.77999282e-01 1.08196929e-01 9.20354202e-02 4.97336537e-01
2.05322281e-01 -4.85122889e-01 1.00861788e+00 1.57848582e-01
1.06044137e+00 -4.64744598e-01 1.05872452e+00 2.96963841e-01
-1.49254024e+00 -3.84043902e-01 -5.34430265e-01 2.42626876e-01
9.58888307e-02 6.86100125e-01 -8.99729192e-01 6.13476217e-01
1.04708350e+00 5.42981267e-01 -8.95793736e-01 1.27813840e+00
-1.82679296e-01 7.24197030e-01 -5.60330987e-01 7.08708465e-02
3.98319453e-01 4.24891934e-02 7.90458262e-01 1.31298816e+00
3.61440152e-01 6.35575084e-03 5.37102401e-01 7.57014573e-01
-1.98460042e-01 -1.84906617e-01 -3.16757917e-01 4.05259222e-01
6.25562310e-01 1.56578171e+00 -9.81334567e-01 -5.80212057e-01
-1.55440211e-01 7.18964875e-01 4.35072817e-02 1.69114828e-01
-1.23097289e+00 -4.91536111e-01 3.34114283e-01 3.78213674e-01
6.47554159e-01 1.98490888e-01 -4.43564169e-02 -7.82003641e-01
2.51563251e-01 -8.06586325e-01 3.35434318e-01 -4.69748616e-01
-1.19344234e+00 7.03144431e-01 2.39209175e-01 -1.59381580e+00
1.69650093e-01 -8.18466485e-01 -5.49796879e-01 6.05007887e-01
-1.63642991e+00 -1.42041361e+00 -9.08598065e-01 2.55770504e-01
8.32843959e-01 -1.00950368e-01 2.77921855e-01 4.49540377e-01
-7.72486925e-01 7.71437228e-01 -3.39209586e-01 2.91071773e-01
7.42959738e-01 -1.11214674e+00 4.64997441e-01 1.10455596e+00
-1.80574223e-01 2.12159947e-01 5.66055357e-01 -6.10025287e-01
-1.15716910e+00 -1.74733746e+00 1.65846035e-01 -1.62680939e-01
4.08142954e-01 -3.08168054e-01 -7.54870236e-01 5.29367864e-01
-1.91892907e-01 5.72912693e-01 4.84320968e-02 -2.64865726e-01
-2.31012344e-01 -5.23176551e-01 -1.12385428e+00 5.27371407e-01
1.20997334e+00 1.08148105e-01 -1.60153344e-01 4.19533730e-01
8.64155412e-01 -6.81261599e-01 -4.05415505e-01 7.88297653e-01
6.20336235e-01 -9.74511087e-01 1.18699920e+00 -6.19751438e-02
1.65556580e-01 -8.77376676e-01 -2.28102088e-01 -8.45068097e-01
-3.97161365e-01 -3.93767804e-01 -1.18418217e-01 1.18134594e+00
4.65481907e-01 -6.63835347e-01 7.44428754e-01 6.30805716e-02
-4.62815225e-01 -1.01537561e+00 -6.64328575e-01 -1.12876606e+00
-5.00112295e-01 -2.19780579e-01 4.19043213e-01 3.24734211e-01
-1.05370259e+00 2.05189794e-01 -2.14530468e-01 5.59138715e-01
8.93409669e-01 5.56313358e-02 1.03315210e+00 -1.20786178e+00
-2.87031889e-01 -2.36193180e-01 -2.68764883e-01 -1.30757689e+00
-3.13924074e-01 -6.81830943e-01 3.59427691e-01 -1.55339873e+00
2.88813580e-02 -6.43065274e-01 -1.86183676e-01 6.84931040e-01
-4.14785951e-01 5.49619675e-01 2.44179368e-01 6.48790551e-03
-7.25740254e-01 4.80917305e-01 1.04434168e+00 -2.81014979e-01
-4.01439577e-01 3.09665889e-01 -4.55295891e-01 9.27226007e-01
8.03931415e-01 -6.16554379e-01 -3.18471044e-01 -2.90437609e-01
4.18638177e-02 -3.40382367e-01 8.33413780e-01 -1.40967548e+00
2.32970744e-01 5.54320309e-03 5.73516786e-01 -1.06670940e+00
2.27210850e-01 -7.11019993e-01 -2.65282512e-01 6.73202574e-01
3.56481262e-02 -9.55611765e-02 4.11315352e-01 7.92245686e-01
3.50032747e-02 -1.34842336e-01 1.02478707e+00 1.35263696e-01
-8.14472854e-01 4.17485744e-01 4.61574830e-02 -4.79238220e-02
1.27345622e+00 -3.82703334e-01 -4.66380000e-01 3.77460003e-01
-1.71361342e-01 5.10719776e-01 1.64701626e-01 2.64844388e-01
6.04685187e-01 -9.97482181e-01 -8.44614089e-01 2.21304446e-01
3.51928085e-01 5.86304188e-01 2.73576587e-01 1.07114542e+00
-2.85698384e-01 1.62448451e-01 -2.77654100e-02 -8.88379097e-01
-1.42270708e+00 3.58196378e-01 3.03827703e-01 -3.04234028e-01
-6.10058665e-01 9.46070731e-01 5.15746117e-01 -2.35351235e-01
4.14380968e-01 -7.02453554e-01 -1.08468188e-02 1.97544277e-01
6.90772891e-01 6.20824754e-01 -3.04882601e-02 -1.70539677e-01
-3.11403662e-01 3.68577212e-01 -2.12594002e-01 3.76281619e-01
1.06610322e+00 2.36474827e-01 2.78086364e-01 1.22162946e-01
6.29841030e-01 -2.38617957e-01 -1.65556824e+00 -6.57804459e-02
-4.05686736e-01 -4.92438376e-01 2.05174655e-01 -9.25058722e-01
-1.08630991e+00 5.48442602e-01 8.24842453e-01 -1.47869557e-01
1.09104836e+00 -2.12205783e-01 6.03860557e-01 5.96800506e-01
4.75286752e-01 -9.56182957e-01 8.22900385e-02 1.20472126e-01
9.97021019e-01 -1.65244925e+00 3.47166926e-01 -4.61287796e-01
-6.09342039e-01 7.86778867e-01 1.13401103e+00 -1.13439299e-01
2.44896412e-01 2.30342358e-01 -3.36431205e-01 1.64084288e-03
-4.18489605e-01 -3.04511786e-01 4.88451779e-01 4.63329256e-01
-1.79583520e-01 5.51803932e-02 -9.55230463e-03 5.21758854e-01
2.37607166e-01 3.86932082e-02 2.79484153e-01 7.03587592e-01
-5.81341863e-01 -4.48113531e-01 -4.84723866e-01 6.81230605e-01
-3.69828016e-01 -7.65893906e-02 3.90480016e-03 1.05414259e+00
3.54279876e-01 8.12372446e-01 1.48691731e-02 -4.21878248e-01
4.97623980e-01 -4.16506469e-01 2.05068216e-01 -4.84243721e-01
-5.13132215e-01 1.09540476e-02 -2.04293747e-02 -6.26748741e-01
-4.84608114e-01 -5.41085839e-01 -1.11072397e+00 5.71364127e-02
-7.86521912e-01 -3.30716163e-01 4.80342388e-01 4.57059026e-01
4.34703439e-01 7.92983830e-01 3.20001394e-01 -9.56978798e-01
-9.21129346e-01 -1.01871800e+00 -2.82398283e-01 -1.22576319e-01
3.98376793e-01 -7.71268368e-01 -3.40862036e-01 1.02735013e-01] | [8.711813926696777, -0.5509023070335388] |
3015ef69-b5bb-4a75-9c66-1929b5137e3b | making-vision-transformers-efficient-from-a | 2303.08685 | null | https://arxiv.org/abs/2303.08685v2 | https://arxiv.org/pdf/2303.08685v2.pdf | Making Vision Transformers Efficient from A Token Sparsification View | The quadratic computational complexity to the number of tokens limits the practical applications of Vision Transformers (ViTs). Several works propose to prune redundant tokens to achieve efficient ViTs. However, these methods generally suffer from (i) dramatic accuracy drops, (ii) application difficulty in the local vision transformer, and (iii) non-general-purpose networks for downstream tasks. In this work, we propose a novel Semantic Token ViT (STViT), for efficient global and local vision transformers, which can also be revised to serve as backbone for downstream tasks. The semantic tokens represent cluster centers, and they are initialized by pooling image tokens in space and recovered by attention, which can adaptively represent global or local semantic information. Due to the cluster properties, a few semantic tokens can attain the same effect as vast image tokens, for both global and local vision transformers. For instance, only 16 semantic tokens on DeiT-(Tiny,Small,Base) can achieve the same accuracy with more than 100% inference speed improvement and nearly 60% FLOPs reduction; on Swin-(Tiny,Small,Base), we can employ 16 semantic tokens in each window to further speed it up by around 20% with slight accuracy increase. Besides great success in image classification, we also extend our method to video recognition. In addition, we design a STViT-R(ecover) network to restore the detailed spatial information based on the STViT, making it work for downstream tasks, which is powerless for previous token sparsification methods. Experiments demonstrate that our method can achieve competitive results compared to the original networks in object detection and instance segmentation, with over 30% FLOPs reduction for backbone. Code is available at http://github.com/changsn/STViT-R | ['Mike Zheng Shou', 'Rong Jin', 'David Junhao Zhang', 'Fan Wang', 'Ming Lin', 'Pichao Wang', 'Shuning Chang'] | 2023-03-15 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chang_Making_Vision_Transformers_Efficient_From_a_Token_Sparsification_View_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chang_Making_Vision_Transformers_Efficient_From_a_Token_Sparsification_View_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-recognition'] | ['computer-vision'] | [ 1.52260857e-02 -3.95799894e-03 -1.78700000e-01 -1.65541485e-01
-7.44999528e-01 -4.19886619e-01 2.37188086e-01 -2.94519603e-01
-6.03240311e-01 4.08150613e-01 -1.12596028e-01 -3.82354528e-01
1.29088014e-01 -7.19860315e-01 -7.43248940e-01 -8.83317888e-01
4.17515904e-01 2.02188477e-01 8.82948816e-01 6.49840059e-03
2.30220959e-01 3.16184729e-01 -1.57016647e+00 2.45915815e-01
1.02260244e+00 1.22385931e+00 7.61942387e-01 5.68029583e-01
-3.83794904e-01 7.67296195e-01 -7.41705477e-01 -2.66755551e-01
3.35180700e-01 -1.48793116e-01 -8.31328332e-01 -6.25986382e-02
4.75479215e-01 -3.49889278e-01 -4.15247083e-01 1.14271688e+00
5.19558132e-01 -7.38793015e-02 3.46233696e-01 -1.24784303e+00
-6.87453747e-01 8.76126111e-01 -8.04793656e-01 1.22446604e-01
-2.01575875e-01 4.12740409e-01 1.01661849e+00 -9.12030697e-01
2.06358775e-01 1.20309806e+00 7.36298680e-01 4.65603650e-01
-6.95548654e-01 -6.92202985e-01 5.20553350e-01 4.61517125e-01
-1.44289064e+00 -3.29702377e-01 4.30655628e-01 1.42507076e-01
9.18213725e-01 2.91462243e-01 7.60995507e-01 6.39458001e-01
-2.07350999e-01 1.16333890e+00 8.10657918e-01 -9.54140127e-02
3.19591425e-02 -1.04709297e-01 7.46924803e-02 9.06050265e-01
1.24995977e-01 -3.62215877e-01 -4.28169250e-01 3.57302696e-01
1.07190514e+00 2.45675281e-01 -2.42004946e-01 8.25493038e-02
-1.25960290e+00 6.31751657e-01 8.70322168e-01 3.06674093e-01
-2.13729799e-01 5.58238208e-01 4.02314752e-01 7.76111707e-02
2.21077263e-01 1.51576936e-01 -5.85598588e-01 -1.64896235e-01
-9.07274485e-01 -6.88450783e-02 4.37727451e-01 1.04070294e+00
8.54004562e-01 2.56242841e-01 -4.18241382e-01 9.76010621e-01
1.66046575e-01 7.25874841e-01 6.11696839e-01 -1.04325211e+00
5.22402883e-01 5.23470998e-01 -1.80334985e-01 -7.00541377e-01
-2.47282833e-01 -4.19186115e-01 -9.51666236e-01 -7.33048618e-02
4.68389332e-01 2.16795206e-02 -1.47091103e+00 1.51202798e+00
2.96885371e-01 4.58160520e-01 -1.15699694e-01 1.04941165e+00
1.03508008e+00 7.40732849e-01 -9.00308490e-02 3.25016491e-02
1.73975182e+00 -1.42846990e+00 -3.57404411e-01 -3.13672721e-01
4.01788592e-01 -8.34989905e-01 1.24017930e+00 3.66703868e-01
-1.27666485e+00 -6.10863507e-01 -8.03694844e-01 -4.11709756e-01
-2.69492209e-01 3.10297996e-01 8.08780313e-01 4.03083622e-01
-1.38717246e+00 3.90177488e-01 -9.76653695e-01 -1.28680766e-01
6.91049814e-01 4.75359052e-01 1.62068561e-01 -2.73122936e-01
-9.32646155e-01 5.80658495e-01 2.24755436e-01 4.49424744e-01
-1.00715697e+00 -6.31117165e-01 -8.29420149e-01 3.31786186e-01
4.91885841e-01 -6.29115105e-01 1.23451304e+00 -8.63861740e-01
-1.52782595e+00 5.19272804e-01 -4.81545389e-01 -3.86153191e-01
4.65126842e-01 1.37281865e-02 1.47137031e-01 3.87876391e-01
2.35251069e-01 9.89406824e-01 9.48809803e-01 -9.78359163e-01
-9.96760488e-01 -1.99141502e-01 1.60702795e-01 2.37570778e-01
-6.59738898e-01 -1.48943439e-01 -1.21393800e+00 -6.76636577e-01
3.89619142e-01 -6.23543620e-01 -3.85541111e-01 1.67554934e-02
-4.82444674e-01 -4.73601699e-01 9.13628280e-01 -5.36857069e-01
8.80190551e-01 -2.20039749e+00 -1.66361570e-01 1.06535383e-01
4.64851111e-01 6.30801082e-01 -4.24200952e-01 -1.24597423e-01
2.98343718e-01 7.62490183e-02 -2.37258509e-01 -5.82496583e-01
-1.37947932e-01 4.73780155e-01 -3.31064016e-01 2.23258287e-01
1.39302492e-01 1.18622160e+00 -7.92706966e-01 -6.41780674e-01
3.07574153e-01 3.03707629e-01 -4.67718661e-01 -1.99764460e-01
-1.49064928e-01 3.30420621e-02 -7.36940503e-01 8.66664469e-01
8.70290399e-01 -3.01795542e-01 -3.76576222e-02 -4.05395091e-01
-2.43779674e-01 2.57632971e-01 -8.97044361e-01 1.40522969e+00
-3.62708539e-01 5.71631074e-01 2.44005963e-01 -1.43691444e+00
7.93831050e-01 7.52580762e-02 3.86442393e-01 -8.09338272e-01
5.14903404e-02 1.88794881e-01 -2.36825645e-01 -3.69262278e-01
5.90903223e-01 9.79211181e-02 4.90003973e-02 1.43917650e-01
-6.17728010e-02 -8.52621794e-02 2.32460573e-01 1.42750084e-01
9.62045133e-01 -3.42076793e-02 -3.61269802e-01 -2.15086341e-03
3.08848232e-01 6.53539319e-03 9.76963639e-01 9.51029956e-01
-3.25828433e-01 7.82035053e-01 4.60582852e-01 -3.03191960e-01
-9.47885871e-01 -1.15696263e+00 -6.69419691e-02 1.04061723e+00
6.53189778e-01 -2.98714161e-01 -8.11740696e-01 -5.68096399e-01
-1.07624598e-01 1.86778054e-01 -2.41621166e-01 -2.06762001e-01
-6.52395010e-01 -8.20801735e-01 9.38793600e-01 9.47676420e-01
1.07121551e+00 -1.12678015e+00 -4.95660067e-01 1.64095029e-01
-4.66890633e-01 -1.31747401e+00 -7.14181364e-01 1.06697589e-01
-1.03834891e+00 -8.83269429e-01 -9.26341772e-01 -1.00435901e+00
7.96857536e-01 8.27873886e-01 6.67528689e-01 2.86854535e-01
-2.33514488e-01 1.29105270e-01 -3.12930733e-01 -2.52084762e-01
2.31896251e-01 1.46326274e-01 -1.75905734e-01 -2.39754662e-01
1.64790079e-01 -4.10656810e-01 -8.64630640e-01 4.79237407e-01
-9.28049624e-01 1.13305993e-01 7.71700978e-01 8.99419129e-01
6.14214122e-01 4.92666252e-02 4.40258831e-01 -4.83161896e-01
1.98430955e-01 -1.63619835e-02 -5.15386045e-01 2.88604468e-01
-4.35819626e-01 -5.28643243e-02 7.50237703e-01 -5.27386844e-01
-1.03056037e+00 -6.28847033e-02 -2.26679564e-01 -7.47600317e-01
1.48391262e-01 2.80882359e-01 -6.29659668e-02 -1.49185851e-01
2.17605948e-01 4.73317832e-01 -7.13943020e-02 -4.24005538e-01
4.61731881e-01 6.26592577e-01 5.89217246e-01 -5.51078558e-01
7.01248527e-01 6.28611863e-01 -4.14226502e-01 -8.35974038e-01
-9.52708185e-01 -5.57121336e-01 -2.57730573e-01 8.56974944e-02
8.51969898e-01 -9.70874965e-01 -1.00183618e+00 8.25572789e-01
-9.52715874e-01 -6.01781607e-01 -2.89220482e-01 4.21873987e-01
-2.72241443e-01 5.48595369e-01 -9.83115613e-01 -5.86632848e-01
-4.75609809e-01 -1.26039982e+00 1.13536453e+00 5.46568036e-01
4.02439833e-01 -7.67255485e-01 -7.51721203e-01 6.76011920e-01
4.11537647e-01 -3.21211785e-01 5.08304417e-01 -2.67113112e-02
-9.73734796e-01 2.42163196e-01 -7.47912705e-01 5.29354930e-01
1.80233978e-02 -2.73319464e-02 -9.68240380e-01 -2.29411140e-01
-1.26579091e-01 -2.86014646e-01 1.44861925e+00 6.70944989e-01
1.65858853e+00 -3.04367453e-01 -4.13451970e-01 8.85052502e-01
1.17628610e+00 6.92260712e-02 8.21858287e-01 1.99134871e-01
1.06999719e+00 1.53998673e-01 7.22863436e-01 2.76496470e-01
6.67254746e-01 4.43686604e-01 4.01236385e-01 -5.64707756e-01
-4.62507039e-01 -1.19366176e-01 5.70167124e-01 8.99314702e-01
-2.54610628e-01 -3.06202564e-02 -6.78482652e-01 8.43735814e-01
-2.07279301e+00 -7.30094373e-01 -1.00372337e-01 2.02504396e+00
8.49932671e-01 1.51620612e-01 -4.06864248e-02 -7.54598975e-02
8.13461959e-01 1.66679367e-01 -8.03902507e-01 -2.17656076e-01
-4.70822118e-02 3.23060721e-01 8.33492100e-01 2.85356522e-01
-8.58617306e-01 1.53073955e+00 5.72320080e+00 1.32300639e+00
-1.23593569e+00 2.61054575e-01 7.36384153e-01 -1.31986886e-01
-1.70273259e-01 2.60723624e-02 -9.48423326e-01 4.67822403e-01
3.97077858e-01 2.56127417e-01 3.89396667e-01 8.13384354e-01
3.06585878e-01 -1.31427824e-01 -5.56293666e-01 9.78590369e-01
-4.72279713e-02 -1.27004552e+00 2.38397434e-01 -1.05445087e-01
5.91905355e-01 3.50242496e-01 1.32853806e-01 2.79974401e-01
4.36290741e-01 -7.92856276e-01 8.80912781e-01 8.70077237e-02
9.45878744e-01 -7.87789404e-01 7.14823604e-01 2.47204632e-01
-1.60274613e+00 -1.78583056e-01 -8.75599146e-01 1.37894288e-01
2.57851779e-02 7.33838677e-01 -5.99255085e-01 4.77651209e-01
1.06832290e+00 8.17255259e-01 -3.73973846e-01 1.02526164e+00
-4.99073654e-01 7.09638238e-01 -5.91536403e-01 -1.04752406e-01
6.53092682e-01 -1.51041031e-01 2.92521060e-01 1.05879867e+00
3.17393571e-01 9.98085216e-02 3.11979979e-01 8.20861995e-01
-1.57204822e-01 -3.96331668e-01 -6.36295229e-02 2.10188016e-01
6.38586760e-01 1.42062902e+00 -1.11010587e+00 -7.02335894e-01
-2.36308590e-01 1.14322782e+00 2.70234168e-01 5.63541591e-01
-1.08434379e+00 -5.92868388e-01 7.07231760e-01 -1.92482635e-01
7.93302298e-01 -1.35524973e-01 -3.66440922e-01 -1.27900016e+00
2.88611293e-01 -6.59034610e-01 2.60800987e-01 -7.67884493e-01
-9.82600629e-01 4.31586683e-01 -2.86117226e-01 -9.78189707e-01
3.05338144e-01 -5.60571015e-01 -7.50636458e-01 5.64399123e-01
-1.73689759e+00 -1.37576783e+00 -4.66608435e-01 8.33706081e-01
6.63325906e-01 2.02953383e-01 1.67173147e-01 3.81250501e-01
-7.25870967e-01 8.34396958e-01 8.71031955e-02 3.38839501e-01
5.53865254e-01 -1.19529510e+00 4.81742471e-01 9.44952726e-01
1.14749976e-01 5.45300782e-01 1.69935569e-01 -3.39803010e-01
-1.33557141e+00 -1.31862640e+00 5.93406618e-01 -4.37848493e-02
4.79330450e-01 -2.77675986e-01 -7.68639266e-01 6.27810895e-01
-1.06813565e-01 1.27691776e-01 2.42194012e-02 5.89280017e-02
-3.78669322e-01 -4.72060770e-01 -1.05837226e+00 8.33752513e-01
1.18345368e+00 -2.88168877e-01 -3.17072183e-01 3.30314606e-01
9.48004484e-01 -4.63447243e-01 -5.98925591e-01 2.06384242e-01
3.88800681e-01 -8.19214046e-01 1.09936655e+00 1.88638493e-02
3.90546501e-01 -5.94044387e-01 2.53929764e-01 -9.85436022e-01
-2.97898829e-01 -4.21721220e-01 1.35638133e-01 1.24132836e+00
2.73846507e-01 -9.36080694e-01 8.68967354e-01 3.03223670e-01
-4.59790468e-01 -7.81506658e-01 -1.05142057e+00 -7.01370656e-01
-1.11801952e-01 -5.62718272e-01 5.80715716e-01 6.80062234e-01
-2.64962077e-01 1.15742214e-01 -1.55482128e-01 2.20457241e-01
5.87195098e-01 2.78366745e-01 5.09393930e-01 -7.57794976e-01
-1.85359791e-01 -6.55321956e-01 -2.53263891e-01 -1.79987097e+00
-2.25911483e-01 -7.72331417e-01 2.96662867e-01 -1.79521310e+00
3.37336540e-01 -8.48453104e-01 -4.11575437e-01 1.08649063e+00
-3.00173253e-01 6.51959240e-01 3.83074105e-01 2.97945201e-01
-7.72970736e-01 6.49240851e-01 1.64468265e+00 -2.72490859e-01
-1.15175091e-01 -3.29366364e-02 -9.20991600e-01 8.24314356e-01
7.32025981e-01 -2.79596955e-01 -4.37139541e-01 -7.92943001e-01
-1.89563915e-01 -1.51789024e-01 5.30764461e-01 -8.01408470e-01
3.92230600e-01 -1.77045148e-02 3.59883875e-01 -6.65775716e-01
4.30352688e-01 -3.39192241e-01 -5.14225960e-01 6.42635584e-01
1.00786254e-01 -1.55750409e-01 1.87668741e-01 3.14655036e-01
-3.32165748e-01 -2.17522919e-01 7.24566758e-01 -2.49065816e-01
-9.06745374e-01 5.83267808e-01 -3.12790155e-01 -8.94913450e-02
8.37633967e-01 -6.47130013e-01 -4.55649406e-01 -1.83679730e-01
-4.99563009e-01 6.26000822e-01 2.70919114e-01 3.44340533e-01
8.48832548e-01 -1.02810943e+00 -5.36678255e-01 3.84050794e-02
-1.96870416e-01 6.08366966e-01 3.36262316e-01 1.08547604e+00
-5.64296544e-01 3.13288540e-01 -7.68346805e-03 -8.27718437e-01
-1.26860058e+00 3.13746482e-01 2.91928977e-01 -1.92680150e-01
-8.78329694e-01 1.10371065e+00 6.39415205e-01 -2.17961907e-01
1.80503458e-01 -8.14542651e-01 -6.17147004e-03 -1.31776348e-01
5.56108177e-01 2.56298602e-01 -9.29646567e-02 -1.58293679e-01
-3.55890244e-01 7.50974119e-01 -2.72283643e-01 8.80199969e-02
1.27761972e+00 -2.51149505e-01 -2.66495824e-01 -4.26347479e-02
9.37580168e-01 -2.01532826e-01 -1.56713462e+00 -2.52892822e-01
-3.74734372e-01 -3.30116242e-01 1.70497879e-01 -4.60337192e-01
-1.67355061e+00 9.24363554e-01 3.51833075e-01 8.88537169e-02
1.34072101e+00 1.35636523e-01 1.24474204e+00 3.72120291e-01
2.74992704e-01 -1.13334858e+00 1.41698375e-01 6.29306436e-01
4.24379021e-01 -9.80788171e-01 -2.39401102e-01 -5.55908799e-01
-6.40007377e-01 8.90816271e-01 7.40253270e-01 -1.55632719e-01
2.62181401e-01 3.63113672e-01 2.59275697e-02 -1.25697568e-01
-5.24335504e-01 -3.99654269e-01 2.10162979e-02 5.82169950e-01
1.90105569e-03 3.07583064e-02 -5.47139049e-02 4.08199489e-01
-8.89784172e-02 -1.32211730e-01 3.88895214e-01 5.98425269e-01
-6.85464919e-01 -9.26972985e-01 -3.64105135e-01 5.40911794e-01
-3.80347937e-01 -4.19554234e-01 -6.10098578e-02 6.37189150e-01
2.84210354e-01 1.05112445e+00 2.84317136e-01 -1.89330295e-01
8.50303248e-02 -5.76838076e-01 4.16733563e-01 -3.84349287e-01
-5.72224677e-01 2.60803998e-01 -9.67157185e-02 -6.94664419e-01
-2.49297887e-01 -4.36211735e-01 -1.62165833e+00 -4.80879813e-01
-5.05117416e-01 4.64753769e-02 4.56216395e-01 1.00608015e+00
3.80215675e-01 7.28977263e-01 4.39337313e-01 -7.72115946e-01
-4.57543999e-01 -9.02946472e-01 -5.16782224e-01 2.66210455e-02
1.26325563e-01 -4.48466927e-01 -1.97322756e-01 8.29858482e-02] | [9.616374015808105, 0.22986207902431488] |
170c7c27-6c39-4681-b6da-31dc23b6be25 | mokb6-a-multilingual-open-knowledge-base | 2211.06959 | null | https://arxiv.org/abs/2211.06959v2 | https://arxiv.org/pdf/2211.06959v2.pdf | mOKB6: A Multilingual Open Knowledge Base Completion Benchmark | Automated completion of open knowledge bases (Open KBs), which are constructed from triples of the form (subject phrase, relation phrase, object phrase), obtained via open information extraction (Open IE) system, are useful for discovering novel facts that may not be directly present in the text. However, research in Open KB completion (Open KBC) has so far been limited to resource-rich languages like English. Using the latest advances in multilingual Open IE, we construct the first multilingual Open KBC dataset, called mOKB6, containing facts from Wikipedia in six languages (including English). Improving the previous Open KB construction pipeline by doing multilingual coreference resolution and keeping only entity-linked triples, we create a dense Open KB. We experiment with several models for the task and observe a consistent benefit of combining languages with the help of shared embedding space as well as translations of facts. We also observe that current multilingual models struggle to remember facts seen in languages of different scripts. | ['Mausam', 'Soumen Chakrabarti', 'Keshav Kolluru', 'Shubham Mittal'] | 2022-11-13 | null | null | null | null | ['knowledge-base-completion', 'knowledge-base-completion', 'open-information-extraction', 'coreference-resolution'] | ['graphs', 'knowledge-base', 'natural-language-processing', 'natural-language-processing'] | [-7.69590259e-01 7.65595496e-01 -4.43319440e-01 1.66181456e-02
-1.16523540e+00 -1.04312956e+00 6.66266739e-01 5.81516147e-01
-5.24556637e-01 1.55336988e+00 9.33365643e-01 -3.91130358e-01
-2.36147583e-01 -6.69000566e-01 -1.04896021e+00 1.21355921e-01
-1.26392037e-01 1.02914691e+00 3.01488250e-01 -7.19369709e-01
-2.29515910e-01 -2.40423322e-01 -1.18089294e+00 6.59411073e-01
1.04898643e+00 2.02722728e-01 1.75967053e-01 2.91316807e-01
-6.18030369e-01 9.74870503e-01 -1.78527609e-01 -1.36038411e+00
-3.43338251e-02 1.03283823e-01 -1.69915068e+00 -9.58461165e-01
5.59780896e-01 1.83455721e-01 -3.19561869e-01 8.23328435e-01
3.53055894e-01 -1.22607812e-01 3.13628644e-01 -8.11377406e-01
-1.15150189e+00 1.47691202e+00 -8.69246423e-02 3.53530794e-01
9.81763542e-01 -3.95145059e-01 1.47926319e+00 -9.38051462e-01
1.65488613e+00 1.01777625e+00 9.21488583e-01 4.21210527e-01
-1.18244612e+00 -3.03164303e-01 -1.98852867e-01 7.13098943e-01
-1.49516582e+00 -5.11429191e-01 5.73059618e-02 -1.97719768e-01
1.83224475e+00 2.38274336e-01 5.06821096e-01 1.17637718e+00
1.07044866e-02 4.56321001e-01 1.00836146e+00 -9.09980357e-01
-4.22129005e-01 4.42734987e-01 4.72287536e-01 7.51454949e-01
8.19502711e-01 -2.71407694e-01 -8.09232414e-01 -1.42446697e-01
2.62014657e-01 -7.65983999e-01 -5.43876231e-01 -4.18145239e-01
-1.68097651e+00 5.04274547e-01 3.19536358e-01 4.40027505e-01
-2.31117815e-01 -2.41815254e-01 5.29483497e-01 6.26678407e-01
3.72381091e-01 7.11102009e-01 -9.90173936e-01 -5.76810464e-02
-2.72428840e-01 4.75684315e-01 1.48048413e+00 1.31853878e+00
1.10519576e+00 -7.28275657e-01 2.22610593e-01 9.67585564e-01
5.48334345e-02 5.82072258e-01 4.53072995e-01 -8.21990490e-01
1.00905597e+00 5.63114166e-01 3.03025305e-01 -8.90745223e-01
-4.32430029e-01 -1.30047858e-01 -7.62054399e-02 -8.08612049e-01
3.35593343e-01 -1.97551921e-01 -4.08095449e-01 1.71368289e+00
5.60782135e-01 9.61691327e-03 9.39015150e-01 3.75973701e-01
1.42573404e+00 3.97618651e-01 1.15828499e-01 -4.90731671e-02
1.72646558e+00 -6.62028432e-01 -1.01110685e+00 -2.25578751e-02
1.13224494e+00 -6.07829928e-01 7.53491402e-01 -8.89425864e-04
-7.36713171e-01 -4.68009412e-02 -1.01626039e+00 -8.06454539e-01
-1.16561985e+00 -3.65047336e-01 9.93203998e-01 3.64483595e-01
-1.04179347e+00 2.00764894e-01 -5.80617249e-01 -7.79041350e-01
-1.92436427e-01 5.53371385e-02 -1.08109617e+00 -3.93271387e-01
-1.89022171e+00 1.47345126e+00 1.21731365e+00 -1.58167869e-01
-3.24688137e-01 -1.01890886e+00 -1.22347581e+00 -9.72891003e-02
9.22153294e-01 -1.00105023e+00 8.95399868e-01 -2.54836708e-01
-6.13858879e-01 9.65183854e-01 4.85134237e-02 -7.11427212e-01
-2.70766884e-01 -6.39165461e-01 -1.00464904e+00 -4.06674668e-02
6.37505889e-01 7.26407409e-01 5.22143692e-02 -1.24190629e+00
-8.94145727e-01 -4.65410501e-01 5.64440906e-01 2.43670911e-01
-1.37843370e-01 1.69648007e-01 -5.06313980e-01 -3.59151572e-01
-1.01468764e-01 -7.48076797e-01 1.66178271e-01 -8.61870289e-01
-3.53872567e-01 -2.47099608e-01 3.67374539e-01 -1.33249736e+00
1.49438417e+00 -1.50185752e+00 3.78970563e-01 1.91205298e-03
1.51715457e-01 3.97450030e-02 3.12019791e-02 8.21185827e-01
-1.13910489e-01 2.38156706e-01 -5.25647439e-02 1.21415742e-01
1.87736586e-01 8.68472517e-01 -7.10601330e-01 -1.95398778e-01
9.83896777e-02 1.24305522e+00 -1.08590019e+00 -7.81060815e-01
-4.04344440e-01 1.27488971e-01 -7.46938527e-01 -5.21812856e-01
-4.28722262e-01 1.17857466e-02 -2.19194889e-01 5.95990121e-01
3.43177944e-01 -9.65006799e-02 7.24671483e-01 -4.37090099e-01
-2.21792281e-01 6.81086600e-01 -1.16302526e+00 2.18995452e+00
-5.55905998e-01 5.17837167e-01 -4.70381826e-01 -3.55213076e-01
4.29037303e-01 7.67203391e-01 1.79578304e-01 -4.89485413e-01
-4.53973621e-01 5.33144951e-01 -2.53656328e-01 -6.92813396e-01
1.12695611e+00 -2.07971901e-01 -4.24134880e-01 3.16467255e-01
1.11861277e+00 1.14333242e-01 6.58137679e-01 7.72827327e-01
9.95261729e-01 4.12759006e-01 7.56141007e-01 -4.22819704e-01
4.75563914e-01 6.21275127e-01 6.79009557e-01 4.05772924e-01
2.65154183e-01 -1.52014345e-01 3.88590276e-01 -5.10436594e-01
-1.00309134e+00 -1.18687868e+00 -5.01002967e-01 6.45269454e-01
-2.52997074e-02 -1.19833779e+00 -3.86976928e-01 -8.70185077e-01
3.25721428e-02 8.09457898e-01 -5.47043502e-01 2.32543230e-01
-6.81057453e-01 -5.32613933e-01 1.04948866e+00 2.68865108e-01
1.78782362e-02 -8.30637932e-01 -1.83872152e-02 2.85682797e-01
-9.89052236e-01 -1.57169950e+00 6.62965551e-02 2.74724156e-01
-4.24610794e-01 -1.24834824e+00 -1.23674229e-01 -9.10807014e-01
2.72784472e-01 -2.37088859e-01 1.68549752e+00 -2.29481757e-01
-1.59995615e-01 6.50038540e-01 -6.57505035e-01 -4.61184949e-01
-3.09353769e-01 2.35079944e-01 3.67019922e-01 -5.67253709e-01
7.78385282e-01 -4.91245508e-01 3.00988197e-01 -3.63628715e-01
-9.25356686e-01 2.26814881e-01 3.14172924e-01 9.53977883e-01
4.84237373e-01 -6.54117465e-02 6.50789320e-01 -1.36999083e+00
5.72562993e-01 -7.67629445e-01 -2.41236612e-01 9.28523779e-01
-3.55779082e-01 5.89383662e-01 2.66630054e-01 1.44544303e-01
-1.54402089e+00 -4.87610579e-01 -1.28484234e-01 3.62836778e-01
1.36148231e-02 1.15891457e+00 -6.99714646e-02 1.56940818e-01
7.63150036e-01 -8.53638127e-02 -6.91771567e-01 -6.81805611e-01
1.08770895e+00 4.84335005e-01 8.88390303e-01 -1.12557197e+00
4.98831004e-01 3.39071840e-01 -3.91927451e-01 -6.92535222e-01
-1.23273981e+00 -5.53455353e-01 -1.09191847e+00 1.99086577e-01
8.06710720e-01 -1.48656118e+00 -2.14509934e-01 -2.00383693e-01
-1.30417609e+00 1.19368143e-01 -4.64946687e-01 5.91141939e-01
-3.46482545e-01 3.53549242e-01 -7.26816535e-01 -2.96839867e-02
-2.91307092e-01 -4.24752980e-01 8.26048613e-01 -9.83375311e-02
-3.68764430e-01 -1.49872971e+00 8.19462717e-01 5.44574440e-01
-1.93411857e-01 1.32045047e-02 1.44263959e+00 -7.25812554e-01
-5.09012580e-01 1.88507423e-01 -2.87453271e-02 -1.95824489e-01
-1.75730847e-02 -2.03367054e-01 -6.28573298e-01 6.68537095e-02
-6.20195746e-01 -7.92518377e-01 7.38815069e-01 -4.01617795e-01
-1.44720599e-01 -4.19345737e-01 -4.98497009e-01 4.41595048e-01
1.82096779e+00 -4.52598453e-01 7.76579916e-01 7.36521125e-01
6.44153595e-01 5.21617293e-01 3.57284456e-01 1.77113228e-02
1.25887001e+00 3.88368368e-01 -2.81409651e-01 3.88969988e-01
-3.30380559e-01 -7.42480218e-01 3.11951816e-01 1.37444496e+00
-3.68740410e-01 1.65023193e-01 -1.21933889e+00 9.94956851e-01
-1.72883952e+00 -1.01961899e+00 -2.91883320e-01 1.92202377e+00
1.61398578e+00 -1.44754618e-01 -5.04008532e-01 -4.80843127e-01
2.05378637e-01 -2.34088570e-01 1.54574022e-01 -1.11482926e-01
-6.81164503e-01 4.26049024e-01 4.90227461e-01 8.81376505e-01
-8.52828145e-01 1.34245706e+00 6.07565975e+00 7.06026614e-01
-4.54001218e-01 3.70380610e-01 -3.78617257e-01 1.04091518e-01
-9.12997186e-01 9.07891750e-01 -1.31050932e+00 5.96108474e-02
9.81925726e-01 -2.21678287e-01 3.90889794e-01 4.57234085e-01
-7.79803693e-01 -4.26500440e-01 -1.12340903e+00 8.99949729e-01
2.59010077e-01 -1.64253891e+00 4.01944399e-01 -1.78762391e-01
8.03858519e-01 3.35566938e-01 -5.40457189e-01 9.44432199e-01
1.08230329e+00 -4.84168053e-01 7.39735246e-01 7.71634102e-01
8.45331073e-01 -5.41802824e-01 8.13443303e-01 2.79066354e-01
-1.02614033e+00 1.31002784e-01 -5.93967319e-01 2.80441701e-01
3.86546761e-01 5.80049634e-01 -7.93186486e-01 1.41710711e+00
8.79986048e-01 8.03266823e-01 -9.40775454e-01 4.43238676e-01
-4.94061738e-01 1.95882335e-01 -1.60966069e-01 2.71825671e-01
4.95557450e-02 -8.72549117e-02 6.56375408e-01 1.30622232e+00
1.46214098e-01 1.45059153e-01 -2.23257653e-02 4.74289626e-01
-2.22749129e-01 2.79150099e-01 -8.34647417e-01 -4.59314808e-02
4.08521742e-01 1.02939010e+00 -1.57395318e-01 -4.61036593e-01
-9.82517362e-01 9.02588189e-01 1.25031674e+00 2.23819405e-01
-4.81501788e-01 -5.47667623e-01 4.54626590e-01 -3.44596326e-01
2.68638790e-01 -1.97806388e-01 1.97615594e-01 -1.83517408e+00
2.63958752e-01 -8.04868042e-01 9.12681758e-01 -9.16927874e-01
-1.26064515e+00 5.20280182e-01 2.12986246e-01 -5.81958473e-01
-4.03336525e-01 -4.35446560e-01 2.65826046e-01 7.76070237e-01
-1.63724494e+00 -1.46133709e+00 4.38680291e-01 8.91704500e-01
4.44228128e-02 7.54888803e-02 1.29130232e+00 6.06276929e-01
-2.91575491e-01 3.00458640e-01 1.27975136e-01 3.37099791e-01
1.17004037e+00 -1.41416061e+00 4.27449763e-01 9.00095761e-01
6.90876305e-01 1.37664759e+00 6.20935321e-01 -1.22646809e+00
-1.41103470e+00 -8.02179873e-01 1.85637021e+00 -1.29019213e+00
1.19898641e+00 -4.80340749e-01 -8.96088004e-01 1.50628364e+00
5.57063222e-01 -4.09634709e-02 1.00255537e+00 1.08262372e+00
-8.89224768e-01 9.67138410e-02 -6.13562644e-01 6.66892052e-01
1.12339818e+00 -9.05386269e-01 -1.52182281e+00 1.99598894e-01
1.11023283e+00 -5.88529170e-01 -1.42797256e+00 2.72160143e-01
4.57261831e-01 -3.55332494e-01 9.68926132e-01 -1.27474737e+00
2.19402641e-01 -2.74327338e-01 -3.89283508e-01 -1.56603849e+00
-2.68971264e-01 -3.62989962e-01 -4.10925269e-01 1.49452722e+00
1.13303900e+00 -5.50250232e-01 7.48461299e-03 4.26029712e-01
-3.38258684e-01 -1.22049101e-01 -1.06681955e+00 -7.17985868e-01
1.88301459e-01 -5.05795896e-01 6.80598021e-01 1.45339644e+00
1.00401449e+00 8.87233913e-01 -2.62706101e-01 2.08599448e-01
3.32828760e-01 1.00261711e-01 6.70120478e-01 -1.33840477e+00
-2.06168175e-01 4.74640191e-01 -1.37954175e-01 -5.38554072e-01
4.46151286e-01 -1.29484570e+00 -2.39160299e-01 -1.76053286e+00
5.07894158e-01 -7.23509610e-01 7.90827572e-02 7.25485146e-01
-3.46767664e-01 -9.36756060e-02 -1.03313126e-01 3.16475421e-01
-9.75174725e-01 3.12534422e-01 7.95417726e-01 1.08563537e-02
1.39601573e-01 -1.08457625e+00 -1.00545359e+00 6.25690341e-01
1.68234557e-01 -5.28487384e-01 -1.27271071e-01 -9.46080446e-01
1.20226753e+00 1.52510777e-01 1.94201231e-01 -7.70461142e-01
5.21128654e-01 8.83745253e-02 1.98020265e-02 -3.07276636e-01
8.63552913e-02 -7.99049795e-01 5.45821071e-01 -4.26816456e-02
-1.49687871e-01 7.06926957e-02 4.38199639e-01 5.46523809e-01
-4.14553821e-01 -3.42124224e-01 -1.17648654e-01 -4.37663913e-01
-1.36138177e+00 -1.68332547e-01 -2.26163194e-02 7.48083770e-01
7.10144043e-01 3.08188200e-01 -8.92175078e-01 1.23233944e-01
-7.68893242e-01 3.57661694e-01 2.80165762e-01 5.23193538e-01
1.90490991e-01 -1.54093170e+00 -7.18071699e-01 -1.94228590e-01
7.46857047e-01 -2.64758557e-01 3.32130194e-01 7.89627075e-01
-4.93191242e-01 1.18003833e+00 -2.69299775e-01 1.34466156e-01
-9.36924338e-01 8.04605424e-01 -1.41146511e-01 -8.28482330e-01
-7.48607278e-01 4.97716367e-01 -3.65603626e-01 -1.17069721e+00
-1.46724120e-01 -2.46447146e-01 -3.76859128e-01 1.34680405e-01
4.85985816e-01 2.34599143e-01 5.00924468e-01 -8.49612236e-01
-4.03364807e-01 2.01314360e-01 -3.57626170e-01 -2.77635932e-01
1.42658508e+00 -5.16617119e-01 -5.49840927e-01 5.15790999e-01
8.02089691e-01 6.42931759e-01 -2.05433160e-01 -6.11044407e-01
5.65222979e-01 -1.09159552e-01 -3.33419770e-01 -1.05158818e+00
-2.96565503e-01 1.23398036e-01 -1.68515444e-01 -3.32001179e-01
4.09569919e-01 5.50492525e-01 9.46457088e-01 1.02378535e+00
9.93144929e-01 -1.25376439e+00 -6.04127288e-01 1.17120636e+00
7.97615767e-01 -1.16112506e+00 -4.78467345e-03 -4.73063082e-01
-7.38806248e-01 7.74778426e-01 5.04846990e-01 4.61829811e-01
6.08793914e-01 2.62965858e-01 7.53180906e-02 -6.82750463e-01
-1.05752254e+00 -5.53885937e-01 3.29284132e-01 8.60475004e-01
4.40490872e-01 1.97755575e-01 -5.36577582e-01 1.04743087e+00
-4.99240547e-01 -1.35936320e-01 5.70718706e-01 9.60309744e-01
-2.07300425e-01 -1.24874556e+00 -2.15316549e-01 3.29896927e-01
-6.18930280e-01 -8.10784936e-01 -4.46775079e-01 1.08649611e+00
6.19428635e-01 5.66657543e-01 -1.38952419e-01 -5.03400788e-02
2.96817392e-01 4.15171117e-01 7.70292222e-01 -8.60415041e-01
-3.06959540e-01 -7.61056304e-01 8.85205984e-01 -3.77674878e-01
-6.69176757e-01 -7.22276986e-01 -1.18947423e+00 -2.65106380e-01
-4.72074062e-01 4.99093503e-01 2.85543323e-01 9.76648986e-01
3.91456723e-01 1.62621319e-01 -4.75674063e-01 6.59428090e-02
4.95589040e-02 -8.23851049e-01 -6.35330081e-01 3.90036523e-01
4.89679631e-03 -8.24211001e-01 1.62390292e-01 4.29194033e-01] | [9.488286972045898, 8.775092124938965] |
0a4ba9e7-7c40-4bf4-ba03-15ba77b556c1 | deflow-self-supervised-3d-motion-estimation | 2304.02569 | null | https://arxiv.org/abs/2304.02569v1 | https://arxiv.org/pdf/2304.02569v1.pdf | DEFLOW: Self-supervised 3D Motion Estimation of Debris Flow | Existing work on scene flow estimation focuses on autonomous driving and mobile robotics, while automated solutions are lacking for motion in nature, such as that exhibited by debris flows. We propose DEFLOW, a model for 3D motion estimation of debris flows, together with a newly captured dataset. We adopt a novel multi-level sensor fusion architecture and self-supervision to incorporate the inductive biases of the scene. We further adopt a multi-frame temporal processing module to enable flow speed estimation over time. Our model achieves state-of-the-art optical flow and depth estimation on our dataset, and fully automates the motion estimation for debris flows. The source code and dataset are available at project page. | ['Jordan Aaron', 'Konrad Schindler', 'Andreas Wieser', 'Nicholas Meyer', 'Shengyu Huang', 'Yuru Jia', 'Liyuan Zhu'] | 2023-04-05 | null | null | null | null | ['scene-flow-estimation', 'motion-estimation'] | ['computer-vision', 'computer-vision'] | [-2.74637222e-01 -3.33447844e-01 -1.62411675e-01 -2.63333827e-01
-4.59051840e-02 -4.62320954e-01 6.67233467e-01 -1.99707225e-01
-3.25425953e-01 6.31157160e-01 5.40508628e-01 -1.62899271e-01
8.18556845e-02 -8.63462031e-01 -4.68096733e-01 -3.13046575e-01
-2.99648553e-01 2.78712749e-01 7.06710935e-01 -1.41160071e-01
3.50484580e-01 7.38250732e-01 -1.66254067e+00 1.96956456e-01
7.36584783e-01 8.26190412e-01 3.96061689e-01 1.31293082e+00
-1.53388157e-01 1.22582650e+00 -8.89536664e-02 -3.53593789e-02
4.60238397e-01 1.61281824e-01 -8.18727314e-01 2.45094955e-01
6.49740219e-01 -1.04970670e+00 -1.03893888e+00 5.71733892e-01
2.41569296e-01 6.50743470e-02 6.44042075e-01 -1.31646442e+00
-2.28565142e-01 -6.02862351e-02 -2.57939875e-01 5.36143720e-01
4.93413568e-01 7.74915695e-01 5.43886721e-01 -9.74200368e-01
9.44678009e-01 1.40597796e+00 5.57198584e-01 4.23689097e-01
-6.77487195e-01 -1.38196573e-01 1.82579473e-01 2.77887017e-01
-8.28803003e-01 -6.48780227e-01 6.61500394e-01 -8.41703475e-01
9.92074013e-01 -3.12276453e-01 9.91030455e-01 9.79236186e-01
4.41221297e-01 1.14029884e+00 4.68654126e-01 5.96358627e-02
3.00179243e-01 -4.40350533e-01 5.66958301e-02 8.99770260e-01
6.83104575e-01 3.41425508e-01 -6.86713398e-01 2.43355811e-01
9.92379367e-01 -1.40057262e-02 -2.23380387e-01 -6.58893585e-01
-1.42203319e+00 6.18160903e-01 5.26015818e-01 -1.46519810e-01
-3.91967326e-01 5.13662755e-01 2.69880384e-01 -1.12381749e-01
4.76322144e-01 3.88320349e-02 -3.97541136e-01 -4.13531601e-01
-9.64674711e-01 5.55801034e-01 8.46413672e-01 1.21420133e+00
1.31010997e+00 2.35825464e-01 -3.74955386e-02 1.97132543e-01
5.84265471e-01 1.03375554e+00 2.00415581e-01 -1.78686595e+00
5.59581399e-01 5.92420101e-01 4.91537899e-01 -1.05214643e+00
-6.34943724e-01 8.98108482e-02 -4.22307491e-01 4.33433861e-01
5.63276529e-01 -2.65299469e-01 -7.49746978e-01 1.20758474e+00
4.84344989e-01 5.97739935e-01 1.13662444e-01 1.16859269e+00
6.09855890e-01 3.28262448e-01 -6.89018592e-02 3.98140363e-02
1.01986516e+00 -1.09479070e+00 -7.70782948e-01 -8.74727845e-01
8.65463614e-01 -3.70890111e-01 6.51926875e-01 1.17342643e-01
-9.47659492e-01 -2.95432508e-01 -9.63726461e-01 -4.88676667e-01
-3.41206223e-01 1.01713195e-01 9.55482304e-01 2.65702516e-01
-9.66504872e-01 3.12377572e-01 -1.26253808e+00 -5.89046955e-01
7.69546270e-01 -5.56628257e-02 -3.61591637e-01 -3.41371626e-01
-7.91958690e-01 9.11304295e-01 2.06301436e-01 3.72629851e-01
-1.02206111e+00 -8.56773674e-01 -1.28772962e+00 -4.96739775e-01
-2.19075568e-02 -1.23856425e+00 1.24740851e+00 -2.67215371e-01
-1.43010533e+00 6.59141481e-01 -4.83360112e-01 -6.43765807e-01
8.78030837e-01 -8.12605917e-01 1.18503310e-02 3.72939855e-01
2.97589451e-01 1.01817656e+00 6.40166223e-01 -1.29304183e+00
-9.44437504e-01 -2.54471421e-01 2.54621834e-01 1.78458706e-01
-7.89605267e-03 -3.58700037e-01 -2.46129379e-01 -2.42559478e-01
-4.64207754e-02 -9.12289858e-01 -4.23081428e-01 6.67122543e-01
-1.21851943e-01 3.62619311e-01 1.37982643e+00 -4.09072578e-01
1.21306562e+00 -1.76737058e+00 -1.86400354e-01 -4.71919775e-01
2.92888880e-01 2.35912457e-01 -1.12226615e-02 6.13405108e-02
7.08555162e-01 -5.22078812e-01 -4.19257849e-01 -6.29446447e-01
-2.00489789e-01 3.57175797e-01 -3.84523183e-01 7.44978845e-01
3.11238110e-01 1.08745158e+00 -1.38071167e+00 -7.01339543e-01
9.09270644e-01 3.58244300e-01 -6.94346786e-01 3.83864760e-01
-2.21084237e-01 4.75310624e-01 -4.69852149e-01 7.48496771e-01
1.09363067e+00 -3.95682678e-02 -3.71051013e-01 -2.28403509e-01
-5.09851754e-01 2.64539778e-01 -1.22336054e+00 2.15630412e+00
-3.81852806e-01 9.50130641e-01 2.08440289e-01 -3.31801265e-01
6.11378849e-01 -1.44942343e-01 7.63023674e-01 -2.88589746e-01
2.33446047e-01 1.18366994e-01 -3.48877698e-01 -1.02941656e+00
1.05721033e+00 2.95517087e-01 9.21563357e-02 2.85763919e-01
-1.85189754e-01 -4.16318715e-01 4.37127978e-01 3.29909146e-01
1.53262568e+00 6.29799247e-01 -6.30421489e-02 -1.45288438e-01
5.89331567e-01 4.68907088e-01 5.76377869e-01 6.18939281e-01
-8.25319946e-01 7.55090296e-01 -8.28567073e-02 -6.36498690e-01
-8.62931848e-01 -1.28558195e+00 6.84874877e-02 5.00323772e-01
7.83742130e-01 -2.31043473e-01 -3.94883424e-01 -5.97515166e-01
4.77791548e-01 1.49270996e-01 -4.30686891e-01 1.11127548e-01
-9.49459374e-01 -5.39867163e-01 4.01342034e-01 6.41509235e-01
8.00917029e-01 -1.00409329e+00 -1.23511553e+00 4.58565265e-01
-3.09617639e-01 -1.72040212e+00 -2.80841053e-01 -3.65541428e-01
-9.74514723e-01 -1.24551725e+00 -2.99613029e-01 -4.28034842e-01
2.46208251e-01 9.35217083e-01 9.38299835e-01 -1.17191784e-01
-4.15389150e-01 6.97889805e-01 -1.40463322e-01 -3.51520479e-01
-1.48429453e-01 1.69911757e-02 4.30904515e-02 -1.26004189e-01
3.72221649e-01 -4.40001011e-01 -1.19581115e+00 2.25856870e-01
-7.56224811e-01 -2.78796088e-02 4.38972831e-01 3.35444987e-01
2.58265920e-02 -5.01516998e-01 3.15189868e-01 -4.12139565e-01
1.21079385e-02 -7.17758596e-01 -3.87614012e-01 -4.70554173e-01
-9.02536437e-02 -2.30233483e-02 2.14767709e-01 -1.95034668e-01
-1.33564317e+00 4.50589418e-01 2.43150339e-01 -7.13499546e-01
-3.45341086e-01 -4.59069312e-02 1.69406962e-02 -6.35427758e-02
5.76670170e-01 -2.64775515e-01 1.25668705e-01 -1.39849380e-01
7.16752827e-01 6.85352504e-01 1.01344323e+00 -1.98034540e-01
7.16692567e-01 1.48863804e+00 2.01757357e-01 -9.26217139e-01
-8.05319846e-01 -8.74971330e-01 -1.12711644e+00 -6.45917356e-01
8.54049921e-01 -1.29841447e+00 -6.91933930e-01 8.89372051e-01
-1.58934271e+00 -5.44872046e-01 -3.50108206e-01 8.31513882e-01
-9.09604490e-01 5.39830506e-01 -7.30422735e-01 -8.60276699e-01
-1.65563956e-01 -8.51026475e-01 1.41915131e+00 9.29380506e-02
-2.17240348e-01 -1.25048232e+00 2.65438169e-01 2.24773660e-01
4.87030149e-01 2.31289655e-01 -6.47075996e-02 4.27005202e-01
-1.23886800e+00 1.14946887e-01 -4.60632652e-01 -7.25878403e-02
2.30613440e-01 2.13733092e-02 -1.05379498e+00 -7.23156109e-02
-2.30985299e-01 5.26211150e-02 1.17015278e+00 7.63534009e-01
3.94074053e-01 2.23640259e-02 -5.76846242e-01 8.14327419e-01
1.25693476e+00 -2.53614783e-01 6.53275371e-01 7.36037850e-01
1.08163452e+00 9.12705421e-01 1.06781578e+00 7.78047383e-01
1.00673902e+00 3.24939072e-01 8.08787882e-01 1.92993835e-01
-5.48334956e-01 -1.95339814e-01 5.78452945e-01 6.29806519e-01
2.13099808e-01 -3.91369492e-01 -8.58627737e-01 1.08541191e+00
-2.16394091e+00 -9.48104739e-01 -7.30928361e-01 1.64172566e+00
2.82082357e-03 -5.86940488e-03 -1.31629556e-02 -4.19919342e-02
5.26180148e-01 4.76792157e-01 -6.82883561e-01 -6.63792808e-03
-3.02112609e-01 -4.49480057e-01 9.18270886e-01 9.97674048e-01
-1.39437330e+00 1.30504405e+00 7.10902452e+00 -3.26175950e-02
-8.95003259e-01 -1.10339327e-02 -2.59539366e-01 -1.17568830e-02
-3.97455752e-01 2.85256803e-01 -7.66650081e-01 4.66693789e-01
7.57538080e-01 4.76241671e-02 4.31821197e-02 5.24941981e-01
7.75342226e-01 -6.75322115e-01 -8.58376265e-01 8.82625103e-01
-1.90924667e-02 -1.40128899e+00 4.17175479e-02 1.62609667e-01
6.79698586e-01 4.07074273e-01 -3.94355565e-01 -1.72518283e-01
6.30239129e-01 -4.46630806e-01 8.54427636e-01 7.70394385e-01
3.53240043e-01 -1.80615321e-01 5.35361946e-01 3.06104481e-01
-1.48095775e+00 -2.67991692e-01 -2.56612957e-01 -4.98290688e-01
8.78649294e-01 1.02555192e+00 -8.14491868e-01 4.89020258e-01
6.82862401e-01 1.59734082e+00 -5.54228902e-01 1.30545831e+00
-1.61213040e-01 2.08200574e-01 -3.86616796e-01 2.90416807e-01
9.41878408e-02 -1.54973015e-01 8.95053208e-01 1.45903265e+00
4.04217243e-01 -3.00915509e-01 2.36381531e-01 6.10607386e-01
3.57737094e-01 -5.50604522e-01 -1.13117135e+00 4.93732154e-01
5.07229030e-01 1.19293737e+00 -4.20614809e-01 -4.56591308e-01
-4.03538495e-01 9.48365271e-01 1.97519049e-01 3.00410748e-01
-6.04146183e-01 -2.58393884e-01 1.33534777e+00 1.85440809e-01
2.84707665e-01 -8.31915677e-01 -5.82620442e-01 -1.36383212e+00
5.93517087e-02 3.85485798e-01 1.81734040e-01 -9.62763071e-01
-9.48906958e-01 5.93617447e-02 -6.53167367e-02 -1.80692875e+00
-3.10134232e-01 -7.83387959e-01 -4.05724674e-01 3.44486624e-01
-2.16917610e+00 -1.28682292e+00 -1.01839268e+00 4.87070322e-01
6.42890275e-01 1.60230070e-01 8.78547952e-02 2.80421197e-01
-2.25254014e-01 -4.39100266e-01 -2.26810887e-01 2.11565252e-02
8.79613757e-01 -1.04350531e+00 7.56608903e-01 1.31061184e+00
-4.59467113e-01 6.55590445e-02 7.37553358e-01 -9.01913941e-01
-1.71346319e+00 -1.50049222e+00 7.15613604e-01 -1.08405030e+00
8.15394640e-01 -1.96169823e-01 -7.84894764e-01 7.99445629e-01
-2.25257024e-01 6.72613978e-01 -1.31659508e-01 -7.13122487e-01
-2.52043596e-03 1.01693571e-01 -1.08584988e+00 3.50542784e-01
1.66879189e+00 -4.26790208e-01 -4.92226571e-01 -4.94947210e-02
6.79689884e-01 -5.87881804e-01 -6.47424638e-01 4.36340630e-01
8.30723584e-01 -9.58280623e-01 1.11440420e+00 -1.81120217e-01
4.79544669e-01 -8.16691279e-01 -1.39669344e-01 -1.12109554e+00
-3.82122606e-01 -5.87921023e-01 -9.06260073e-01 8.61336172e-01
-1.93891242e-01 -4.23524261e-01 1.22448516e+00 3.21221769e-01
-4.80249316e-01 -1.15749411e-01 -8.75762463e-01 -7.34322071e-01
-2.33291164e-01 -8.04010689e-01 3.02163064e-01 5.49724340e-01
-1.15010388e-01 1.15830719e-01 -3.58849823e-01 4.24642414e-01
7.90008426e-01 -7.54749551e-02 1.15043175e+00 -1.24092782e+00
3.86076868e-01 -1.26661286e-01 -8.90619874e-01 -1.58549833e+00
3.85168135e-01 -6.12439513e-01 4.30555046e-01 -1.87922323e+00
-3.55466545e-01 -3.06142896e-01 3.04666162e-01 -2.53068283e-03
-1.21179432e-01 6.64859116e-02 2.40236044e-01 4.40977752e-01
-6.70565844e-01 7.80395329e-01 1.27134085e+00 -2.22119033e-01
-2.46604428e-01 -2.98671544e-01 -2.32507780e-01 7.83560634e-01
5.50290167e-01 -2.08892286e-01 -3.06344509e-01 -9.49745655e-01
-2.42985189e-01 1.41064832e-02 5.50323248e-01 -1.46536338e+00
5.92803180e-01 -3.44116271e-01 2.35721916e-01 -8.39100599e-01
5.66321373e-01 -7.47197330e-01 -1.42987490e-01 6.24591827e-01
1.83976352e-01 7.57278204e-02 1.59800500e-01 8.29389036e-01
1.37733500e-02 5.41430041e-02 6.17795706e-01 -1.18391462e-01
-1.23453772e+00 6.46893859e-01 -7.25770652e-01 9.49812382e-02
1.02535617e+00 -5.03179073e-01 -9.10356343e-01 -2.81050771e-01
-1.77624181e-01 5.94905138e-01 7.00109363e-01 7.23548770e-01
7.12468982e-01 -1.12394691e+00 -7.35837638e-01 2.23376855e-01
2.50239372e-01 5.05834758e-01 3.20071310e-01 6.89489782e-01
-1.05449045e+00 3.44474047e-01 -2.79561639e-01 -9.82549429e-01
-5.78871608e-01 2.91845322e-01 3.51727217e-01 -1.20399632e-02
-8.84785771e-01 3.03314716e-01 2.62807935e-01 -4.19701159e-01
-2.33407006e-01 -5.86304247e-01 -4.74756770e-02 -9.11218822e-02
6.51132166e-01 1.05610907e+00 -1.14479307e-02 -8.61756802e-01
-5.61836541e-01 8.77856553e-01 3.56549561e-01 -1.45334065e-01
9.78748560e-01 -7.46481061e-01 1.63686186e-01 4.75079596e-01
1.16869843e+00 -2.58694112e-01 -2.03080392e+00 -2.16644891e-02
-4.84830327e-02 -8.88586998e-01 2.27638870e-01 -1.84673354e-01
-9.67731059e-01 9.57986057e-01 4.66746420e-01 -2.20155418e-01
6.93385959e-01 -1.80516586e-01 1.11991560e+00 3.81745607e-01
7.08616436e-01 -9.37215447e-01 -6.32250905e-02 1.06492436e+00
4.53959197e-01 -1.33805680e+00 1.62714377e-01 -8.05380762e-01
-3.21646780e-01 1.07199025e+00 8.68569195e-01 -3.82870853e-01
7.66288579e-01 6.13683343e-01 2.05311149e-01 -1.27883196e-01
-9.16608393e-01 -5.69730043e-01 -4.65986341e-01 1.05177224e+00
-1.61997348e-01 -2.01443598e-01 5.74784055e-02 -2.34818056e-01
8.11830536e-03 4.83814925e-01 1.06006348e+00 1.34186780e+00
-6.49594426e-01 -5.35615742e-01 -3.49812150e-01 2.75546849e-01
1.65130869e-01 3.57946038e-01 3.09598874e-02 5.66547155e-01
1.16235524e-01 1.03469622e+00 5.53148687e-01 -3.04134399e-01
4.62292910e-01 -4.84031349e-01 4.27607089e-01 -2.90702611e-01
-7.94704482e-02 -4.93008792e-01 2.86192805e-01 -1.03791738e+00
-8.75090182e-01 -9.15564001e-01 -1.37902749e+00 -4.50117797e-01
-4.43593189e-02 -4.20972854e-01 7.13239908e-01 8.94572318e-01
6.89971566e-01 2.96380460e-01 6.69111609e-01 -1.60892487e+00
3.88111204e-01 -8.62651467e-01 -4.58007902e-01 3.57228637e-01
9.26662207e-01 -1.10517216e+00 -6.39772296e-01 3.65980059e-01] | [8.548986434936523, -1.941285252571106] |
1e00efef-144f-4d80-91f8-5c6c1f7351a6 | the-structurally-complex-with-additive-parent | 2304.14109 | null | https://arxiv.org/abs/2304.14109v1 | https://arxiv.org/pdf/2304.14109v1.pdf | The Structurally Complex with Additive Parent Causality (SCARY) Dataset | Causal datasets play a critical role in advancing the field of causality. However, existing datasets often lack the complexity of real-world issues such as selection bias, unfaithful data, and confounding. To address this gap, we propose a new synthetic causal dataset, the Structurally Complex with Additive paRent causalitY (SCARY) dataset, which includes the following features. The dataset comprises 40 scenarios, each generated with three different seeds, allowing researchers to leverage relevant subsets of the dataset. Additionally, we use two different data generation mechanisms for generating the causal relationship between parents and child nodes, including linear and mixed causal mechanisms with multiple sub-types. Our dataset generator is inspired by the Causal Discovery Toolbox and generates only additive models. The dataset has a Varsortability of 0.5. Our SCARY dataset provides a valuable resource for researchers to explore causal discovery under more realistic scenarios. The dataset is available at https://github.com/JayJayc/SCARY. | ['Haytham M. Fayek', 'Jarry Chen'] | 2023-04-27 | null | null | null | null | ['causal-discovery', 'additive-models', 'selection-bias'] | ['knowledge-base', 'methodology', 'natural-language-processing'] | [ 1.87412754e-01 2.61525154e-01 -7.47229517e-01 -5.03154933e-01
-3.80242169e-01 -6.53308570e-01 7.79993594e-01 1.99351817e-01
1.61359206e-01 1.18236458e+00 6.61128819e-01 -6.76819980e-01
-4.40575957e-01 -1.19684207e+00 -8.56490076e-01 -3.45244795e-01
-5.30091524e-01 2.38150492e-01 -1.34789348e-01 1.06796101e-01
1.34340540e-01 -1.50461257e-01 -1.24238086e+00 2.13468999e-01
1.28809512e+00 1.20960981e-01 5.10051940e-03 3.30996156e-01
3.74112844e-01 7.24989772e-01 -3.89715999e-01 -5.22128642e-01
7.83505291e-02 -6.49245977e-01 -5.93029559e-01 -7.02830493e-01
9.45226476e-02 -4.24196512e-01 -3.53046268e-01 7.49571383e-01
6.47925138e-01 -3.88878971e-01 7.43821263e-01 -1.88458407e+00
-8.25776219e-01 1.42882550e+00 -8.18006635e-01 2.19428062e-01
4.97531980e-01 2.94695705e-01 1.15987563e+00 -6.63730204e-01
6.08348966e-01 1.82148170e+00 4.70906347e-01 2.86343306e-01
-1.53985751e+00 -1.27280104e+00 1.84477806e-01 1.15758739e-01
-1.19564915e+00 -2.31395707e-01 3.98245841e-01 -5.50548732e-01
1.51800245e-01 4.64945108e-01 4.98222470e-01 1.72789633e+00
1.26516014e-01 5.15367150e-01 1.21709955e+00 -1.05160318e-01
3.07412684e-01 -3.28189343e-01 1.68539342e-02 3.08404118e-01
8.36445689e-01 6.34248137e-01 -3.94204438e-01 -7.49495864e-01
8.69255245e-01 1.13441005e-01 -1.42374545e-01 5.29033616e-02
-1.42580521e+00 1.05456829e+00 3.26793045e-01 -4.77973148e-02
-3.75179768e-01 4.30332333e-01 -7.43172914e-02 -9.12189949e-03
2.18666524e-01 3.94326746e-01 -4.90782589e-01 1.31591549e-02
-5.83024144e-01 7.49817491e-01 5.80335438e-01 9.69843566e-01
3.02578866e-01 -2.29444265e-01 -4.42485571e-01 7.00124741e-01
2.69611806e-01 7.97786832e-01 3.66573930e-02 -9.95408237e-01
3.91277552e-01 7.52213538e-01 2.16851681e-01 -1.11470962e+00
-5.56087196e-01 -1.27325401e-01 -1.00899422e+00 -2.50433028e-01
4.74006027e-01 -5.98873913e-01 -7.43592739e-01 2.24898791e+00
6.41947925e-01 5.34083903e-01 -2.93196708e-01 7.80374169e-01
1.10758471e+00 4.19775605e-01 3.52940351e-01 -7.43623599e-02
1.40335667e+00 -2.25993171e-01 -6.44457579e-01 8.63802154e-03
3.23775947e-01 -5.69910407e-01 1.13949621e+00 1.52331486e-01
-1.00244784e+00 9.01552103e-03 -6.75095439e-01 1.35623083e-01
-1.54997408e-01 -1.53837234e-01 1.23073053e+00 5.24286985e-01
-5.43007612e-01 3.76018822e-01 -6.01881266e-01 -3.02218318e-01
4.91092741e-01 -3.19800638e-02 -1.25123411e-01 -2.97973722e-01
-1.76957309e+00 3.22546422e-01 2.41888210e-01 -1.06017970e-01
-1.09058797e+00 -1.59128881e+00 -6.76867485e-01 1.57965422e-01
5.53689182e-01 -1.06580341e+00 9.47966278e-01 -3.23413134e-01
-5.14679253e-01 2.86454469e-01 -1.25808641e-01 -6.94255829e-02
6.64855182e-01 -3.09764631e-02 -2.47183219e-01 -3.00993770e-01
5.19625008e-01 5.35711646e-01 3.40353191e-01 -1.24574530e+00
-4.41951603e-01 -2.93428421e-01 7.93001577e-02 -3.75020325e-01
1.71842724e-01 2.32273847e-01 -1.66194603e-01 -1.20970404e+00
-3.15205395e-01 -8.25457692e-01 -3.90873820e-01 -3.64760816e-01
-9.74908292e-01 -5.26526421e-02 1.32756785e-01 -2.96840310e-01
1.56756663e+00 -1.83600807e+00 -1.13676064e-01 6.71203509e-02
3.76728624e-01 -2.96541214e-01 -4.01705652e-01 7.30031431e-01
-6.95073426e-01 7.94462860e-01 -2.68970817e-01 1.40345678e-01
-1.83421999e-01 -2.16228273e-02 -3.28227520e-01 9.95293707e-02
3.99815440e-01 8.84325504e-01 -1.21563327e+00 -3.61728728e-01
-1.94802314e-01 2.83686310e-01 -8.23316693e-01 1.98234379e-01
-2.11743385e-01 5.24806976e-01 -6.06906652e-01 6.56188190e-01
8.17855000e-01 -3.20477784e-01 4.55684721e-01 2.77202129e-01
-1.60588533e-01 4.58080441e-01 -1.35402691e+00 1.02049756e+00
-1.25505298e-01 2.39747435e-01 -1.60687923e-01 -4.22833383e-01
6.28203988e-01 2.83259928e-01 4.91411805e-01 -3.24216366e-01
2.38170121e-02 -2.85711903e-02 3.16974998e-01 -5.32289982e-01
2.08410136e-02 3.36586647e-02 -1.36274859e-01 7.20751107e-01
-4.54801768e-01 2.69267499e-01 4.12501603e-01 5.02211690e-01
1.60310674e+00 -3.35014224e-01 3.16096067e-01 -1.88588664e-01
-3.54014635e-01 3.79437543e-02 1.09093428e+00 9.09610152e-01
2.37602964e-01 6.80570006e-01 1.16797113e+00 -2.78812557e-01
-8.75875175e-01 -1.30476093e+00 -3.52388918e-01 6.08599782e-01
-1.14375629e-01 -6.15472257e-01 -3.03033978e-01 -4.55415606e-01
4.15173084e-01 9.07772601e-01 -1.10041630e+00 -1.96022525e-01
-2.47391239e-01 -1.54188883e+00 8.05784285e-01 4.05518264e-01
1.74938336e-01 -8.74291718e-01 -3.30631286e-01 -5.34648634e-03
-4.87235188e-01 -6.33581042e-01 -2.10428014e-01 -2.53868073e-01
-7.04373300e-01 -1.66976357e+00 -3.16372067e-01 4.09628004e-02
6.41734600e-01 1.90750271e-01 1.40060782e+00 1.76657811e-01
-3.59748036e-01 -1.33676380e-01 -4.63191688e-01 -6.18351638e-01
-5.07631660e-01 -7.30805192e-03 -5.04990593e-02 -3.09156835e-01
1.31593019e-01 -8.66611481e-01 -9.54442024e-01 3.98477614e-01
-9.20448184e-01 1.29197180e-01 5.07340729e-01 8.66055608e-01
1.79736704e-01 1.82758465e-01 1.00147033e+00 -1.20135796e+00
6.21481001e-01 -1.32722378e+00 -5.84108353e-01 3.82625610e-02
-7.26467907e-01 -1.01519756e-01 3.52415383e-01 -5.68491459e-01
-1.11132264e+00 -4.26453024e-01 2.42211387e-01 -8.30947235e-02
-2.45427340e-01 6.35234892e-01 -4.03127283e-01 8.08533967e-01
6.69406831e-01 -6.22903287e-01 -2.58135796e-01 -4.44204628e-01
5.21457553e-01 1.98279113e-01 2.27588400e-01 -7.09086716e-01
7.90047646e-01 4.09904361e-01 2.08169445e-02 -1.20435268e-01
-5.11947095e-01 1.56545281e-01 -3.44591826e-01 1.73827469e-01
4.63728815e-01 -9.16542649e-01 -8.28266919e-01 2.17139035e-01
-1.01935732e+00 -5.65858722e-01 -4.24653217e-02 5.94659805e-01
-5.21975681e-02 -2.74047405e-01 -3.92790228e-01 -6.94614351e-01
-9.88075957e-02 -9.04443622e-01 6.94558620e-01 1.23489134e-01
-6.27864301e-01 -9.10423219e-01 1.95891872e-01 1.26902059e-01
-1.22096762e-02 7.88733542e-01 1.24268496e+00 -3.47259670e-01
-3.78539354e-01 2.02737972e-01 -3.59193414e-01 -5.35855293e-01
3.56248409e-01 6.07222259e-01 -7.93221891e-01 3.57804894e-02
-6.86253846e-01 -1.26759484e-01 8.65203381e-01 6.90694690e-01
1.19480014e+00 -5.73074400e-01 -6.22636557e-01 4.59333420e-01
1.04219437e+00 2.03985959e-01 5.70340812e-01 2.54787803e-02
6.44466817e-01 8.99216056e-01 5.92269957e-01 7.81819224e-01
8.95001650e-01 4.95487124e-01 6.65783226e-01 -4.48617756e-01
1.14307411e-01 -8.33528578e-01 -1.26431569e-01 -7.58547410e-02
1.47630386e-02 -4.67984885e-01 -1.15554905e+00 7.06036270e-01
-1.86469972e+00 -1.13328362e+00 -8.25811803e-01 2.26260448e+00
1.24726987e+00 -9.15376544e-02 4.02041018e-01 4.08235975e-02
7.94631064e-01 -9.40363407e-02 -5.29491007e-01 -1.02075100e-01
-1.78990304e-01 -9.34276357e-02 6.01663351e-01 3.25315773e-01
-8.69503319e-01 6.08508110e-01 6.85027790e+00 4.91808265e-01
-7.71386385e-01 6.75705820e-02 8.99275362e-01 -3.36224735e-01
-9.62653995e-01 3.87991279e-01 -5.51156878e-01 8.21551919e-01
8.83958519e-01 -4.59650189e-01 1.06173202e-01 2.47156456e-01
1.04452777e+00 -3.12810779e-01 -1.11744785e+00 2.52803952e-01
-6.76844299e-01 -1.27699888e+00 1.16689317e-02 2.92155534e-01
8.83444607e-01 -1.86763912e-01 1.06291743e-02 1.49730211e-02
1.24979115e+00 -1.25496674e+00 4.66731310e-01 4.31028754e-01
1.10059118e+00 -5.94558120e-01 5.39226830e-01 1.55615956e-01
-7.21499085e-01 -2.75545835e-01 -2.52518326e-01 -3.59641790e-01
2.06759602e-01 1.27302194e+00 -9.01399672e-01 5.79577982e-01
8.75097513e-01 6.44896686e-01 -5.99726439e-01 9.47473407e-01
-2.98976064e-01 1.29552627e+00 -2.78757304e-01 2.16701031e-01
-3.00098628e-01 -1.09239444e-01 4.10595953e-01 9.55026090e-01
4.11626697e-01 3.11391473e-01 -2.60853410e-01 1.37747562e+00
-2.17231259e-01 -2.91587412e-01 -7.31785059e-01 7.42575037e-04
1.25602782e+00 1.05710316e+00 -5.37716269e-01 -3.35351676e-02
-2.22122669e-01 3.28688882e-02 -8.02143440e-02 5.18140078e-01
-1.08597696e+00 1.11757375e-01 7.20713377e-01 3.79485548e-01
-3.82423282e-01 1.80638850e-01 -5.71602821e-01 -1.02654755e+00
-3.56912673e-01 -9.71238256e-01 7.67745614e-01 -7.77039826e-01
-1.33445835e+00 -2.17583463e-01 6.09916389e-01 -7.55729795e-01
-2.28672922e-01 4.20244783e-03 -8.91691446e-01 9.93072510e-01
-9.97884333e-01 -1.13686168e+00 -4.69772249e-01 2.88961828e-01
1.07227169e-01 3.07353348e-01 5.33381343e-01 3.91174972e-01
-9.72957134e-01 5.89921296e-01 -2.12540865e-01 -5.50539829e-02
1.06364632e+00 -1.28510916e+00 6.77395761e-01 6.12347782e-01
-4.37642246e-01 1.11244476e+00 7.65337288e-01 -1.16806221e+00
-1.22917080e+00 -1.26476741e+00 8.37741971e-01 -7.08828747e-01
8.78067732e-01 -5.12627900e-01 -7.29572475e-01 7.92255580e-01
1.76665261e-02 -3.31037283e-01 7.93192625e-01 5.81809163e-01
-4.21725124e-01 2.31683217e-02 -1.04375076e+00 9.07330096e-01
1.37502837e+00 3.03005725e-01 -2.16551572e-01 1.99374571e-01
9.65028226e-01 -5.99666163e-02 -9.62525487e-01 6.58933103e-01
6.39844000e-01 -8.81951272e-01 9.24406111e-01 -7.32076347e-01
1.07255721e+00 -1.33445933e-01 1.70178011e-01 -1.50820506e+00
-5.20700574e-01 -5.28874099e-01 2.83427894e-01 1.70323086e+00
8.82784665e-01 -6.48384988e-01 3.36732686e-01 7.10040510e-01
3.43250066e-01 -6.98459387e-01 -5.75233877e-01 -5.22510409e-01
3.63203734e-01 -4.68858808e-01 1.27011800e+00 1.37223935e+00
-1.33294746e-01 3.48660737e-01 -4.24581259e-01 3.01641583e-01
7.63385415e-01 1.88647062e-01 8.59537959e-01 -1.32389843e+00
-1.29040673e-01 -2.39612773e-01 2.17811868e-01 -2.29838312e-01
-1.40599251e-01 -6.69188857e-01 -4.62551147e-01 -1.43614328e+00
7.15400100e-01 -9.80126143e-01 1.00119486e-01 8.51889372e-01
-7.83416688e-01 -1.37311012e-01 -1.44784287e-01 -5.11604287e-02
1.39103830e-01 5.49810469e-01 1.24234664e+00 5.08437604e-02
-2.33822405e-01 3.10545117e-02 -1.19240105e+00 7.34462261e-01
9.92745459e-01 -7.74875522e-01 -7.91146457e-01 -2.09977388e-01
6.13961518e-01 2.82575190e-01 8.08617234e-01 -7.96112046e-02
-2.47945964e-01 -7.92059302e-01 2.93877006e-01 -4.99072433e-01
-3.55851650e-01 -3.76790851e-01 8.72675121e-01 3.96096617e-01
-5.79712093e-01 3.78190309e-01 1.70686334e-01 3.22452039e-01
1.16196238e-01 1.05986580e-01 2.96243101e-01 -4.15204987e-02
-9.45997983e-02 1.72675669e-01 -9.64506790e-02 3.90303344e-01
7.57139444e-01 4.00706917e-01 -9.07425821e-01 -4.79754627e-01
-2.64588714e-01 7.81621635e-01 3.00040185e-01 6.39534950e-01
3.93018246e-01 -1.37638080e+00 -1.14529514e+00 2.68328879e-02
1.55422658e-01 -2.66615320e-02 2.91287810e-01 9.36673880e-01
2.32037883e-02 2.35734582e-01 6.61827251e-02 -1.86438829e-01
-1.03708994e+00 6.46163225e-01 -1.38823856e-02 3.13407071e-02
-3.44124407e-01 5.54334044e-01 6.30270660e-01 -6.42443419e-01
-2.71153092e-01 -2.97361851e-01 -1.29197568e-01 1.56132758e-01
6.38843656e-01 8.49891603e-01 -5.74969709e-01 -7.37063354e-03
-3.64214480e-01 -8.39586407e-02 3.35866064e-01 -7.89781511e-02
1.28370309e+00 1.54199610e-02 -2.81803846e-01 4.98383552e-01
6.25601470e-01 3.55259180e-01 -1.02117705e+00 2.50673473e-01
-1.35101229e-02 -6.89119577e-01 -3.37210655e-01 -1.18364418e+00
-9.04075742e-01 4.03064966e-01 1.36171907e-01 3.39648902e-01
1.03872752e+00 1.02138832e-01 8.53986517e-02 -5.39151967e-01
2.49665797e-01 -3.01864386e-01 -2.66631365e-01 6.91680610e-02
1.22656906e+00 -1.30231559e+00 1.18157513e-01 -7.93084204e-01
-2.32431054e-01 3.97111744e-01 6.19350135e-01 1.64952919e-01
5.20275652e-01 5.10299146e-01 -1.51996478e-01 -2.01522976e-01
-1.43756258e+00 -1.70100350e-02 4.26873937e-02 5.85057855e-01
8.14533353e-01 3.98263335e-01 -8.27330649e-01 1.00532365e+00
-4.68189716e-01 8.70315209e-02 7.96060562e-01 3.15829694e-01
3.31506193e-01 -1.23269105e+00 -6.97131455e-01 6.63901448e-01
-5.76139450e-01 -4.34155256e-01 -6.45480692e-01 1.00125647e+00
2.24123389e-01 1.28504431e+00 -4.13521603e-02 -2.94073105e-01
2.81570286e-01 -4.99191701e-01 -1.84578635e-02 -5.18850386e-01
-2.57495344e-01 3.04450206e-02 3.55112970e-01 -6.70054197e-01
-2.77116567e-01 -9.77522373e-01 -1.09677720e+00 -8.39253306e-01
-2.76481621e-02 1.41327098e-01 1.51566342e-01 4.31401432e-01
5.84966123e-01 6.51058018e-01 7.08095372e-01 -2.76127577e-01
-9.98973921e-02 -1.11432850e+00 -3.84648621e-01 3.41790348e-01
7.56694153e-02 -1.13813961e+00 -3.76890868e-01 2.06099749e-02] | [7.912816047668457, 5.347589015960693] |
f4cb03d7-ac0c-4e95-8513-8f3626361816 | minimum-n-rank-approximation-via-iterative | 1311.4291 | null | http://arxiv.org/abs/1311.4291v2 | http://arxiv.org/pdf/1311.4291v2.pdf | Minimum $n$-Rank Approximation via Iterative Hard Thresholding | The problem of recovering a low $n$-rank tensor is an extension of sparse
recovery problem from the low dimensional space (matrix space) to the high
dimensional space (tensor space) and has many applications in computer vision
and graphics such as image inpainting and video inpainting. In this paper, we
consider a new tensor recovery model, named as minimum $n$-rank approximation
(MnRA), and propose an appropriate iterative hard thresholding algorithm with
giving the upper bound of the $n$-rank in advance. The convergence analysis of
the proposed algorithm is also presented. Particularly, we show that for the
noiseless case, the linear convergence with rate $\frac{1}{2}$ can be obtained
for the proposed algorithm under proper conditions. Additionally, combining an
effective heuristic for determining $n$-rank, we can also apply the proposed
algorithm to solve MnRA when $n$-rank is unknown in advance. Some preliminary
numerical results on randomly generated and real low $n$-rank tensor completion
problems are reported, which show the efficiency of the proposed algorithms. | ['Zheng-Hai Huang', 'Lei Yang', 'Min Zhang'] | 2013-11-18 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [ 2.51354188e-01 -2.10490733e-01 8.53015259e-02 1.75461508e-02
-8.36635232e-01 -2.19347075e-01 -1.98262244e-01 -3.35150987e-01
-4.61437285e-01 5.91593206e-01 1.41237691e-01 -1.44921914e-01
-5.40658116e-01 -4.60669547e-01 -6.38457716e-01 -9.11950529e-01
-3.87490660e-01 2.38784209e-01 -2.04914093e-01 -1.27059236e-01
4.84031886e-01 3.78796637e-01 -1.25335312e+00 -1.43978804e-01
9.04782116e-01 1.28296161e+00 2.95196593e-01 6.85790658e-01
5.03067300e-03 7.70341456e-01 -2.07659572e-01 -2.09228069e-01
7.26873696e-01 -4.34772342e-01 -8.82069051e-01 6.07248724e-01
4.07185137e-01 -3.54551166e-01 -5.52778065e-01 1.36760545e+00
1.65454894e-01 4.00964946e-01 3.73665422e-01 -7.35533297e-01
-2.36104414e-01 3.37398797e-01 -1.16045928e+00 3.51130962e-01
4.19496745e-01 -3.02159220e-01 8.64007175e-01 -1.26834810e+00
6.98498309e-01 1.07465231e+00 6.24565542e-01 -2.55187582e-02
-1.00923383e+00 -4.98146594e-01 -9.14979652e-02 1.92953214e-01
-1.63622558e+00 -4.12374258e-01 1.11355150e+00 -4.95700896e-01
3.69253159e-01 3.27994674e-01 3.33141863e-01 4.65274453e-02
-2.44247556e-01 7.06577301e-01 1.30592120e+00 -4.90082413e-01
1.56286303e-02 -4.13736314e-01 1.33899122e-01 1.13483107e+00
5.14368832e-01 -2.20829815e-01 -5.31501889e-01 -4.37420696e-01
1.12940681e+00 6.11322932e-02 -4.41369802e-01 -1.52583987e-01
-1.47828519e+00 7.63000965e-01 3.69588375e-01 3.24503154e-01
-4.98925894e-01 3.17428917e-01 2.53725141e-01 3.13273191e-01
4.17425692e-01 1.18169330e-01 3.30238305e-02 -5.81079163e-02
-1.09333313e+00 1.02604829e-01 5.55815935e-01 1.08931863e+00
8.96308362e-01 4.11417514e-01 1.33340254e-01 1.11280537e+00
1.28419265e-01 6.64752901e-01 2.01800853e-01 -1.39154339e+00
8.15403819e-01 5.80460355e-02 3.60823900e-01 -1.57493317e+00
-1.29921979e-03 -2.88470119e-01 -1.25160050e+00 -1.23650230e-01
3.40660065e-01 7.82652274e-02 -5.52525878e-01 1.57569396e+00
3.19701582e-01 5.85322320e-01 -6.36957679e-03 1.23674071e+00
4.85896081e-01 9.06867564e-01 -5.12696385e-01 -6.58003986e-01
1.28514731e+00 -6.72405362e-01 -8.43084335e-01 -8.40537697e-02
4.05530274e-01 -1.14810956e+00 7.45485246e-01 6.49350703e-01
-1.34852684e+00 -2.34133214e-01 -9.18942630e-01 -9.13969055e-02
5.04106224e-01 2.95766622e-01 7.91727424e-01 2.85346180e-01
-9.99914050e-01 5.45385420e-01 -6.97555840e-01 -1.79425448e-01
-1.02983341e-01 4.58651245e-01 -5.03061175e-01 -7.48032153e-01
-7.51651466e-01 3.40630680e-01 -6.58460483e-02 7.49534726e-01
-7.06511080e-01 -2.56608844e-01 -4.90512818e-01 -4.02757317e-01
4.88726467e-01 -5.11874795e-01 8.85650754e-01 -5.01798689e-01
-1.18293154e+00 6.98573709e-01 -3.98981720e-01 9.02039036e-02
2.64084011e-01 -1.10220961e-01 -2.69687138e-02 6.91800416e-01
2.96964973e-01 -1.03144847e-01 1.35183907e+00 -1.13759243e+00
-3.55313838e-01 -4.44024295e-01 -8.23340416e-02 2.70387828e-01
-5.64725772e-02 4.80678342e-02 -5.23335576e-01 -1.01558304e+00
1.10663295e+00 -9.73021805e-01 -6.58073783e-01 1.02181241e-01
-2.62887329e-01 2.17288390e-01 6.06259346e-01 -1.06357312e+00
1.18103373e+00 -2.30128837e+00 6.59279704e-01 5.36007524e-01
4.04280126e-01 -1.19763270e-01 4.36026715e-02 4.27170247e-01
-2.41367131e-01 -1.66740149e-01 -5.83704114e-01 -1.74587131e-01
-4.65316355e-01 1.43704146e-01 -2.28609696e-01 8.28150213e-01
-2.78420091e-01 6.29968122e-02 -7.90585101e-01 -5.24822295e-01
-2.18447924e-01 1.98945120e-01 -4.78626013e-01 1.51916757e-01
1.72705010e-01 2.25934952e-01 -7.60707378e-01 8.12433779e-01
9.93680835e-01 -3.71172607e-01 1.09422326e-01 -5.47413051e-01
-1.59547299e-01 -3.21836978e-01 -1.79159284e+00 1.61337471e+00
-3.69788319e-01 3.72468770e-01 1.10945761e+00 -1.26704395e+00
8.66633475e-01 3.18170130e-01 8.52386534e-01 -3.01386684e-01
1.05031930e-01 5.89026749e-01 -2.21696779e-01 -4.77097094e-01
7.87891924e-01 -5.32514513e-01 3.89539711e-02 6.54051125e-01
-3.88288945e-01 -3.43880476e-03 4.29537296e-01 4.02706206e-01
1.26489997e+00 -3.42198610e-01 9.56140365e-03 -3.17530811e-01
7.42542446e-01 -8.18415135e-02 8.16981494e-01 4.81129140e-01
-1.14392720e-01 6.03924811e-01 4.29044425e-01 -1.98192149e-01
-1.13903534e+00 -7.82051742e-01 -2.34091859e-02 6.79236412e-01
1.07093148e-01 -2.33840466e-01 -6.21151388e-01 -4.45998907e-02
-1.55136093e-01 -1.76456664e-02 -2.90109247e-01 3.85492355e-01
-8.85515273e-01 -9.07678306e-01 7.71885291e-02 3.77913378e-02
8.20313752e-01 -5.11585116e-01 1.17270714e-02 2.66322881e-01
-6.11547291e-01 -1.22431755e+00 -6.88848078e-01 -2.49124810e-01
-1.35036123e+00 -1.21006787e+00 -1.04701018e+00 -9.52375591e-01
1.17340744e+00 9.58768189e-01 5.13591468e-01 2.38082319e-01
-1.99574471e-01 5.12064517e-01 -3.61054718e-01 5.56980431e-01
7.01656193e-02 -3.65298480e-01 2.47644350e-01 5.05170345e-01
-4.52644676e-01 -7.05676734e-01 -6.81284964e-01 3.98347586e-01
-1.22011089e+00 -5.85988760e-02 7.17895508e-01 8.32514405e-01
9.18229103e-01 3.65928024e-01 1.73863366e-01 -8.55473936e-01
7.76013255e-01 -2.53253460e-01 -7.27372348e-01 -9.26222801e-02
-4.24358785e-01 2.64768451e-01 6.68471694e-01 -2.23679975e-01
-7.10923851e-01 1.27257615e-01 1.36548132e-01 -7.53932774e-01
7.02807784e-01 8.62661779e-01 -3.99756320e-02 -4.63136971e-01
3.49515170e-01 4.35153961e-01 2.00711325e-01 -9.03497636e-01
4.26102608e-01 3.76464158e-01 7.70080209e-01 -1.05347359e+00
1.08888102e+00 7.62857139e-01 4.62393820e-01 -9.54298675e-01
-7.12321401e-01 -8.10805857e-01 -2.85033733e-01 -5.70402704e-02
3.65476340e-01 -9.12977576e-01 -6.79827034e-01 4.30752695e-01
-9.68773305e-01 2.11544141e-01 -6.12368025e-02 7.21800148e-01
-8.67653310e-01 1.05284917e+00 -9.04896140e-01 -7.59012163e-01
-3.56057107e-01 -1.28886247e+00 6.76241755e-01 -2.26588696e-01
4.05232608e-01 -6.79888725e-01 -2.41668131e-02 6.07444286e-01
1.76262870e-01 3.55871052e-01 9.07820880e-01 2.42466077e-01
-9.02674019e-01 -3.03974152e-01 -5.52939236e-01 4.41595823e-01
-2.12380234e-02 -4.18359905e-01 -2.47394770e-01 -5.67554891e-01
4.41688538e-01 4.37954720e-03 6.34015381e-01 3.19149494e-01
1.02002239e+00 -7.33591020e-01 -3.24358344e-02 7.24235237e-01
1.68313301e+00 -9.21590105e-02 7.09370196e-01 1.18089736e-01
8.04126620e-01 3.37817907e-01 8.71350646e-01 7.95468986e-01
3.22824977e-02 5.22124231e-01 3.70142400e-01 6.65460248e-03
6.65241033e-02 1.70200825e-01 1.78070545e-01 1.53024018e+00
-4.96406049e-01 3.22837561e-01 -6.17432654e-01 6.26195908e-01
-1.91257358e+00 -9.52011526e-01 -4.11067396e-01 2.40865016e+00
8.76283944e-01 -3.43239278e-01 -4.35725879e-03 4.37591016e-01
1.07155645e+00 2.72530824e-01 -2.89728969e-01 -3.11310142e-01
-1.11771546e-01 3.82068485e-01 7.12628484e-01 8.07904243e-01
-7.47437358e-01 7.34527290e-01 6.15321016e+00 9.37584817e-01
-8.55513871e-01 1.23326644e-01 3.25030804e-01 2.74428539e-02
-3.04909796e-01 1.99116990e-01 -4.02243137e-01 2.41605207e-01
4.28182065e-01 -1.41010493e-01 8.09014201e-01 7.80413926e-01
4.24087703e-01 -2.05154106e-01 -6.05827868e-01 1.60307193e+00
8.34534839e-02 -1.26575243e+00 6.67485148e-02 1.56915307e-01
7.74658322e-01 -4.97570366e-01 9.57420915e-02 -2.78119624e-01
-1.63699120e-01 -6.40187621e-01 3.94933909e-01 4.29261208e-01
1.05872023e+00 -7.93687701e-01 3.73559713e-01 2.25093767e-01
-1.45935953e+00 2.90556420e-02 -5.06693482e-01 -3.45446579e-02
2.55952775e-01 9.96969879e-01 -5.34332991e-01 6.36825562e-01
4.42116410e-01 7.50095248e-01 -1.54292554e-01 1.10819411e+00
1.57721192e-02 3.81099641e-01 -5.31387746e-01 3.00580561e-01
1.83836460e-01 -8.61633778e-01 7.41483629e-01 8.73366952e-01
7.59079039e-01 9.03183401e-01 1.80408940e-01 2.44775504e-01
-1.25023052e-01 2.70006359e-01 -4.44959790e-01 -1.28435686e-01
2.02886507e-01 1.06404281e+00 -6.48879588e-01 -1.64325625e-01
-2.83295482e-01 1.10610735e+00 9.85245481e-02 5.14960647e-01
-3.93386632e-01 -5.56627393e-01 6.15532339e-01 3.40080112e-01
4.53620344e-01 -8.15927804e-01 -2.89509445e-01 -1.41950977e+00
2.34052733e-01 -8.66356373e-01 4.63700533e-01 -6.97463870e-01
-1.16266000e+00 4.90864933e-01 -3.04471683e-02 -1.54226494e+00
-1.10795498e-01 -4.81275380e-01 -6.14511184e-02 8.04825008e-01
-1.17479110e+00 -6.84761584e-01 -9.21145156e-02 1.00184965e+00
2.22272396e-01 -1.83962584e-02 3.60477030e-01 5.60900927e-01
-6.22561038e-01 3.46843004e-01 5.37979543e-01 1.22480869e-01
3.42631340e-01 -7.12911606e-01 -4.28778440e-01 1.24167454e+00
-3.20412129e-01 7.61141896e-01 9.05160367e-01 -6.47415698e-01
-2.04724240e+00 -8.10443938e-01 8.00127864e-01 4.27615911e-01
7.09762156e-01 1.21825196e-01 -8.02282810e-01 5.71171701e-01
-2.28692085e-01 2.32615128e-01 6.01319015e-01 -1.25309646e-01
-3.50145131e-01 -4.31934685e-01 -1.32638848e+00 3.65075529e-01
9.84125674e-01 -5.67615330e-01 -2.14299679e-01 6.37521446e-01
5.07920504e-01 -5.87960243e-01 -1.11549592e+00 2.71994889e-01
1.75846979e-01 -7.11713910e-01 1.07301342e+00 -2.62711912e-01
3.19551528e-01 -7.12870300e-01 -7.00964391e-01 -7.90782511e-01
-3.99813324e-01 -1.10462844e+00 -9.43992287e-02 5.69826961e-01
1.36202946e-02 -2.88446814e-01 8.71534944e-01 5.37390590e-01
-2.38666341e-01 -6.11054003e-01 -1.28265584e+00 -6.79913580e-01
-5.43683887e-01 -4.67407525e-01 6.76305816e-02 1.01290143e+00
-8.99933651e-02 1.98915020e-01 -7.53947616e-01 4.14256573e-01
1.16600108e+00 2.07553700e-01 5.51926434e-01 -7.26992607e-01
-6.11438930e-01 7.64240399e-02 -2.91559249e-01 -1.44513834e+00
-9.24962759e-02 -7.23214030e-01 1.13756694e-01 -1.46820343e+00
1.93322942e-01 -8.15979600e-01 -2.36165494e-01 1.29543662e-01
2.05879062e-02 2.56542414e-01 3.56186777e-01 6.50931478e-01
-4.78070587e-01 4.69437927e-01 1.62715435e+00 -5.62281609e-02
5.79794645e-02 -3.44202183e-02 -5.71486890e-01 6.03491426e-01
2.66025394e-01 -3.95411819e-01 -3.31212729e-01 -6.55514717e-01
3.58314037e-01 9.79244590e-01 -7.35137090e-02 -8.22252393e-01
2.29419786e-02 -3.16099226e-01 -2.32600346e-01 -5.46559453e-01
6.39785767e-01 -7.43864655e-01 2.14524716e-01 3.60678405e-01
-3.37637551e-02 2.85637528e-01 -2.61330038e-01 7.62721062e-01
-3.42042685e-01 -5.53906798e-01 8.18030357e-01 -2.51996845e-01
-5.00773907e-01 6.73584998e-01 -2.74968833e-01 -7.65336156e-02
5.37575543e-01 -2.56922662e-01 -1.41361132e-01 -5.80783844e-01
-9.85857010e-01 -1.29581288e-01 3.78241539e-01 -2.75972754e-01
9.39524233e-01 -1.49544799e+00 -6.33895516e-01 6.30853847e-02
-1.42770350e-01 7.60156065e-02 3.97256136e-01 9.42269504e-01
-1.03895676e+00 2.19621316e-01 2.93588135e-02 -4.23556149e-01
-9.64883626e-01 5.02699256e-01 -4.91563976e-02 -3.86631280e-01
-3.93586218e-01 8.24061692e-01 -1.71064898e-01 1.86759606e-01
-2.06501363e-03 -1.02873616e-01 2.73275197e-01 -3.48309815e-01
5.89390755e-01 7.39019334e-01 1.18319662e-02 -8.88789833e-01
-1.12489782e-01 9.39820170e-01 -1.64073519e-02 -5.38669288e-01
1.39203322e+00 -2.90408462e-01 -8.77608359e-01 1.11289099e-01
1.47566497e+00 1.12174645e-01 -1.03357053e+00 -5.37302792e-01
-2.82178819e-01 -9.31119919e-01 1.58978626e-01 8.38455036e-02
-1.42089736e+00 5.77504635e-01 3.60408515e-01 -5.01845665e-02
1.39363587e+00 -2.55878478e-01 1.05858064e+00 5.13842344e-01
6.28813148e-01 -1.08998418e+00 2.42356941e-01 4.61259872e-01
9.54333246e-01 -8.01431656e-01 4.76854652e-01 -7.47650743e-01
-1.45217493e-01 1.01291633e+00 2.48875841e-01 -4.56284255e-01
7.46781349e-01 -2.99029082e-01 -2.33315676e-01 -1.36575028e-01
-3.30399305e-01 -1.86690930e-02 -9.19492990e-02 2.40903854e-01
3.73782784e-01 -7.58537054e-02 -9.81739998e-01 1.18252262e-01
2.07330403e-03 -5.19773625e-02 9.78707969e-01 1.05443323e+00
-5.45229375e-01 -1.25818610e+00 -9.40966427e-01 5.31784534e-01
-4.45500702e-01 -8.28739777e-02 2.35223159e-01 2.89671391e-01
-1.53156132e-01 9.21905100e-01 -3.19422930e-01 -3.22582304e-01
1.31098747e-01 -3.30419272e-01 6.46342218e-01 -4.20250952e-01
-7.99844265e-02 4.13836956e-01 9.91700664e-02 -4.65756565e-01
-4.87489045e-01 -7.93794811e-01 -1.33675516e+00 -5.96496284e-01
-2.41027977e-02 5.36003828e-01 7.45215952e-01 6.40870214e-01
1.70615315e-01 -3.03832322e-01 1.04854143e+00 -6.78839147e-01
-3.88172060e-01 -6.76982999e-01 -1.12455118e+00 3.81435663e-01
2.91326433e-01 -5.19677281e-01 -4.81810004e-01 1.85328931e-01] | [7.369986057281494, 4.46036434173584] |
32906bf0-23c5-4757-b473-b00d04df88c8 | massively-multilingual-corpus-of-sentiment | 2306.07902 | null | https://arxiv.org/abs/2306.07902v1 | https://arxiv.org/pdf/2306.07902v1.pdf | Massively Multilingual Corpus of Sentiment Datasets and Multi-faceted Sentiment Classification Benchmark | Despite impressive advancements in multilingual corpora collection and model training, developing large-scale deployments of multilingual models still presents a significant challenge. This is particularly true for language tasks that are culture-dependent. One such example is the area of multilingual sentiment analysis, where affective markers can be subtle and deeply ensconced in culture. This work presents the most extensive open massively multilingual corpus of datasets for training sentiment models. The corpus consists of 79 manually selected datasets from over 350 datasets reported in the scientific literature based on strict quality criteria. The corpus covers 27 languages representing 6 language families. Datasets can be queried using several linguistic and functional features. In addition, we present a multi-faceted sentiment classification benchmark summarizing hundreds of experiments conducted on different base models, training objectives, dataset collections, and fine-tuning strategies. | ['Tomasz Kajdanowicz', 'Mikołaj Morzy', 'Krzysztof Rajda', 'Piotr Gramacki', 'Marcin Gruza', 'Szymon Woźniak', 'Łukasz Augustyniak'] | 2023-06-13 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-2.32790008e-01 -3.21242660e-01 -7.01419473e-01 -6.21269882e-01
-1.02919507e+00 -9.53056812e-01 5.28506041e-01 5.66558301e-01
-6.04243994e-01 1.01452637e+00 4.65252161e-01 6.92236936e-03
2.61942267e-01 -3.85501206e-01 -5.34618139e-01 -2.56800920e-01
1.74192503e-01 5.63281357e-01 -4.82881576e-01 -8.56355727e-01
2.44110778e-01 5.98200858e-02 -1.32580674e+00 5.12969017e-01
9.16969001e-01 8.29464912e-01 -2.49632951e-02 3.53188694e-01
-2.97120869e-01 6.84313059e-01 -6.11322403e-01 -1.08924043e+00
-2.65589267e-01 -9.04672891e-02 -7.25246787e-01 -5.14897779e-02
3.54624480e-01 6.27671897e-01 5.90334535e-01 9.43915009e-01
6.02408886e-01 -4.35621083e-01 5.21343946e-01 -1.09896982e+00
-1.02108967e+00 7.62046218e-01 -5.11172056e-01 1.29195452e-01
4.79703635e-01 -4.19412464e-01 1.38515902e+00 -1.02883708e+00
1.06103373e+00 1.26229763e+00 6.46905005e-01 3.75969470e-01
-1.11311412e+00 -7.47154355e-01 1.61469564e-01 -8.07340257e-03
-1.12332690e+00 -3.52894515e-01 6.41911030e-01 -6.81176901e-01
1.20775127e+00 1.62778094e-01 7.85716832e-01 1.41442013e+00
3.03846270e-01 7.14313865e-01 1.58078277e+00 -5.06367326e-01
-4.02066037e-02 8.76114607e-01 1.69583023e-01 4.76099402e-01
2.83797681e-01 -6.62263334e-01 -8.24991703e-01 -1.32491097e-01
-4.41958457e-02 -4.73126709e-01 -3.28527428e-02 -4.40533422e-02
-1.15656769e+00 9.82536912e-01 -3.80808413e-02 5.38884521e-01
4.86876965e-02 -5.25939465e-01 8.10912013e-01 6.48902476e-01
7.91212440e-01 7.50504732e-01 -1.17714500e+00 -3.28466594e-01
-6.20373070e-01 1.70538962e-01 9.08233345e-01 9.57882881e-01
9.72447693e-01 -1.08328551e-01 2.83317983e-01 1.35246146e+00
-3.55989896e-02 7.15785682e-01 6.51073635e-01 -6.27325177e-01
4.19575572e-01 7.85530925e-01 -1.34938404e-01 -8.51884365e-01
-5.54059625e-01 -4.42083538e-01 -4.60581303e-01 -4.85045671e-01
9.64443311e-02 -4.40288872e-01 -1.24912836e-01 1.53919899e+00
2.88376361e-01 -6.38669729e-01 3.53664875e-01 4.99184430e-01
9.81155097e-01 3.29443425e-01 1.71223804e-01 -2.43286535e-01
1.49068296e+00 -9.82327819e-01 -6.33545399e-01 -5.34098923e-01
9.86352146e-01 -1.27522302e+00 1.48451912e+00 5.91038465e-01
-6.76481307e-01 -1.53802246e-01 -9.58110392e-01 -3.53344768e-01
-1.01015544e+00 1.73688591e-01 9.39101160e-01 8.11628163e-01
-8.95680070e-01 -1.16855308e-01 -9.93623585e-02 -7.07477808e-01
1.04539566e-01 8.41710195e-02 -4.73838747e-01 -1.04316004e-01
-1.35205173e+00 1.03760707e+00 1.76404506e-01 -2.00151369e-01
-3.70809048e-01 -7.67053306e-01 -8.26540768e-01 -5.04917860e-01
-1.18712811e-02 -1.53413847e-01 1.04789138e+00 -1.60784411e+00
-1.07116902e+00 1.59175968e+00 -1.62592217e-01 -6.88238740e-02
1.38059393e-01 -2.16483846e-01 -8.70936990e-01 -3.43824476e-01
5.34803689e-01 5.89633286e-01 4.13545132e-01 -1.07394278e+00
-6.03877068e-01 -4.12489682e-01 5.98887838e-02 2.47819960e-01
-9.49808538e-01 6.90527916e-01 -5.02093911e-01 -7.77473927e-01
-6.55489862e-01 -9.97568965e-01 -1.69533715e-01 -6.70221031e-01
-2.17411995e-01 -1.14376396e-01 5.14638066e-01 -5.49464226e-01
1.25183702e+00 -1.86634851e+00 8.58621523e-02 -1.32678583e-01
-1.90563336e-01 -2.14423165e-01 -2.18753949e-01 5.32683611e-01
-8.32412988e-02 4.50857818e-01 1.95342019e-01 -4.41455454e-01
-5.04363235e-03 1.28190368e-01 -1.12638511e-01 2.67611712e-01
1.69520989e-01 7.13615716e-01 -8.29504728e-01 -6.41479373e-01
-2.95714736e-01 5.22480249e-01 -6.70814753e-01 -8.68978947e-02
-1.83372378e-01 4.57061499e-01 -3.71294498e-01 1.08291793e+00
1.13102064e-01 -2.30333582e-01 4.11166579e-01 -9.24198981e-03
-1.82094574e-01 1.93662569e-01 -5.11944711e-01 1.85509598e+00
-8.26423764e-01 7.30904162e-01 8.22793767e-02 -5.87425113e-01
1.03064775e+00 2.79515177e-01 6.21368527e-01 -9.43113625e-01
3.05080444e-01 4.94316876e-01 -1.43361688e-01 -4.85535383e-01
9.58452404e-01 -3.29008430e-01 -9.41274703e-01 3.47253650e-01
1.64570957e-01 -3.43475759e-01 5.95793605e-01 1.49690002e-01
5.46039820e-01 -2.18242127e-02 3.61552596e-01 -6.34139895e-01
6.84825361e-01 3.90271395e-01 7.29627550e-01 -3.25223431e-02
-3.91430289e-01 2.15895474e-01 7.16747046e-01 -3.92501324e-01
-8.57975364e-01 -5.04208148e-01 -5.65125465e-01 1.64191699e+00
-4.29743201e-01 -8.46446931e-01 -5.33763945e-01 -5.05342901e-01
-1.68675736e-01 3.31522316e-01 -6.88099742e-01 1.87172577e-01
-1.92679137e-01 -1.09549189e+00 5.58050334e-01 2.22117946e-01
-3.24438475e-02 -1.05985248e+00 -1.78628713e-01 4.10003848e-02
-5.18109381e-01 -1.36322546e+00 -1.79824069e-01 3.34986567e-01
-3.30650747e-01 -9.75049615e-01 -3.74035895e-01 -1.22288656e+00
3.11375558e-01 -6.61958531e-02 1.89231741e+00 -2.38159999e-01
-2.40656529e-02 2.57722706e-01 -6.07423842e-01 -8.76943707e-01
-3.45155239e-01 5.68934441e-01 1.65838942e-01 -3.36670101e-01
9.37025428e-01 -6.75197318e-02 -7.96442106e-03 1.34124503e-01
-5.74178278e-01 -1.24255389e-01 2.66417146e-01 7.16712713e-01
7.58984089e-01 -1.62714869e-01 1.01456523e+00 -1.27929342e+00
9.01776552e-01 -8.47417355e-01 -2.65445113e-01 3.70712638e-01
-5.57065547e-01 -3.53383362e-01 7.50631571e-01 -1.95413217e-01
-9.42683578e-01 -2.32833296e-01 -2.11535931e-01 1.80745542e-01
-9.02081579e-02 9.54719782e-01 -5.71047254e-02 1.33279160e-01
7.26369143e-01 -3.18468660e-01 -4.71731275e-01 -2.57865846e-01
3.65440071e-01 8.52055132e-01 1.17495827e-01 -9.63891029e-01
2.85351932e-01 2.92859107e-01 -6.46641016e-01 -9.96992111e-01
-1.25692034e+00 -3.86434972e-01 -5.42869210e-01 -3.17382008e-01
8.43555689e-01 -1.40801132e+00 -3.72772843e-01 2.43001074e-01
-7.03560293e-01 -1.87699780e-01 2.74564531e-02 2.37049073e-01
-1.40316129e-01 -2.77211487e-01 -7.07239211e-01 -4.41133410e-01
-4.35835391e-01 -1.18944550e+00 1.04554617e+00 3.26982997e-02
-6.01926327e-01 -1.45381606e+00 4.81818289e-01 7.05609381e-01
2.04136774e-01 3.95241648e-01 9.75634873e-01 -6.20584905e-01
3.76236916e-01 -2.67921805e-01 1.79668471e-01 3.25112551e-01
1.62621588e-01 3.37673426e-01 -9.68155980e-01 -3.49122882e-01
-2.84781426e-01 -1.30376923e+00 4.45988595e-01 -3.76173072e-02
6.84826076e-01 -2.74297670e-02 4.62707393e-02 2.18869045e-01
1.54044783e+00 -7.49963820e-02 -1.25505283e-01 7.92868555e-01
6.47138774e-01 9.03902411e-01 8.48050296e-01 4.81341928e-01
9.73656535e-01 3.76192868e-01 -8.38538408e-02 -1.70629323e-01
4.78198707e-01 -3.61017585e-02 7.61804342e-01 1.55301392e+00
1.39374316e-01 4.58544679e-03 -1.14792109e+00 7.00179338e-01
-1.50603759e+00 -6.71557665e-01 -1.18788861e-01 1.65785098e+00
1.23785961e+00 1.64200485e-01 1.67599127e-01 -1.44412592e-01
3.48136187e-01 1.90322876e-01 -5.47254860e-01 -9.80164289e-01
-8.42505634e-01 2.95856267e-01 2.52531469e-01 3.45565796e-01
-9.89814103e-01 1.27765822e+00 6.59980297e+00 7.10145235e-01
-1.26203012e+00 1.80732727e-01 9.34702694e-01 -2.90479273e-01
-5.53251565e-01 -8.24708268e-02 -9.98338163e-01 1.37494132e-01
1.07829559e+00 -2.94573307e-01 2.54243851e-01 8.66936147e-01
1.89999238e-01 -1.10253587e-01 -8.80080521e-01 8.16051841e-01
3.87181759e-01 -9.90358055e-01 -1.16310708e-01 -1.56963140e-01
1.08669043e+00 5.88678956e-01 2.77363509e-01 6.91485226e-01
3.53725672e-01 -1.00628269e+00 9.17271256e-01 8.86551477e-03
9.97083962e-01 -1.00005996e+00 6.28843307e-01 1.16273493e-01
-9.92156267e-01 1.01618217e-02 -2.23143041e-01 -2.65281916e-01
1.42257651e-02 6.70982659e-01 -2.97415525e-01 2.56074905e-01
1.13436639e+00 1.31038630e+00 -9.98099804e-01 2.05183983e-01
-1.29666731e-01 6.10219836e-01 -4.21217363e-03 -2.01008275e-01
2.71133661e-01 -3.90520811e-01 2.11241096e-02 1.71031988e+00
7.36384168e-02 -5.49833298e-01 1.95809230e-01 1.41153172e-01
-3.14865351e-01 1.01239502e+00 -7.55478501e-01 -3.45846832e-01
3.18068445e-01 1.86524844e+00 -5.47855973e-01 -1.84893370e-01
-9.13850963e-01 5.44495821e-01 7.07896173e-01 1.35257795e-01
-7.18539655e-01 -7.58277811e-03 7.08385229e-01 -2.60638684e-01
-2.45945230e-01 -5.10942526e-02 -2.79735327e-01 -1.62701273e+00
-1.28836498e-01 -1.46618855e+00 5.51156878e-01 -6.03277504e-01
-1.65010202e+00 7.98964977e-01 -4.59481418e-01 -9.56974030e-01
-1.60002798e-01 -7.38127649e-01 -4.95159626e-02 7.29525685e-01
-1.70311093e+00 -1.53637171e+00 5.71033619e-02 7.06619382e-01
5.43003798e-01 -4.51263875e-01 1.19792426e+00 7.18557298e-01
-8.44521880e-01 5.98172426e-01 5.18396636e-03 7.70195425e-02
1.46086228e+00 -1.13092017e+00 -1.07869349e-01 3.49585325e-01
1.03712700e-01 5.96118569e-01 6.30576491e-01 -4.71263587e-01
-1.28193831e+00 -1.10465074e+00 1.45036089e+00 -6.95633829e-01
1.30981052e+00 -7.23367453e-01 -5.26275337e-01 8.67615283e-01
7.76574135e-01 -4.50964302e-01 1.47420990e+00 8.11007738e-01
-6.45996749e-01 -3.86070386e-02 -8.49050641e-01 5.74744761e-01
5.57785869e-01 -7.92246997e-01 -1.44549757e-01 6.22229159e-01
4.62504327e-01 -2.29515076e-01 -1.38462758e+00 2.16679275e-01
7.59692669e-01 -8.01597118e-01 5.66315711e-01 -7.57019222e-01
8.75116646e-01 2.47226059e-01 -5.88726103e-01 -1.55577552e+00
-1.76234692e-01 -3.70238632e-01 4.37039107e-01 1.52762401e+00
1.00106931e+00 -4.22529638e-01 3.98399085e-01 3.43195885e-01
2.90874615e-02 -9.28481936e-01 -4.13815200e-01 -2.15710595e-01
7.30162501e-01 -6.08455658e-01 3.41272831e-01 1.65689218e+00
3.91402394e-01 1.02722347e+00 -2.88109839e-01 -5.33099532e-01
2.56286711e-01 2.19025612e-01 7.84520507e-01 -1.15967119e+00
-2.96892226e-03 -5.71730137e-01 8.68181735e-02 -1.66221321e-01
6.47119224e-01 -1.17696714e+00 -3.70409101e-01 -1.15277529e+00
3.53772640e-01 -6.46044374e-01 -2.52918184e-01 5.10259032e-01
-1.46491870e-01 4.57693428e-01 -5.38741350e-02 -1.71007160e-02
-6.90005124e-01 1.58535078e-01 1.04527807e+00 -1.72414273e-01
7.02682808e-02 -5.69695830e-01 -1.19101095e+00 9.35812712e-01
1.12890029e+00 -3.62887442e-01 -3.64270329e-01 -5.37728190e-01
1.22878325e+00 -1.79994792e-01 -3.46451432e-01 -5.57079494e-01
-3.09538752e-01 -3.86314631e-01 1.79516971e-01 -3.60764444e-01
3.53512049e-01 -5.79341888e-01 -1.53873891e-01 -3.45689841e-02
-3.64853472e-01 7.33838260e-01 3.21306795e-01 7.06294924e-02
-5.52825809e-01 1.50183260e-01 7.50860512e-01 -2.11422324e-01
-7.30621696e-01 1.64468333e-01 -3.70003551e-01 6.18234038e-01
8.56079459e-01 2.95093209e-01 -4.40107852e-01 -1.64105162e-01
-5.15286565e-01 3.36748421e-01 7.60204434e-01 1.03327513e+00
3.15490700e-02 -1.25075030e+00 -8.72989655e-01 -1.30950049e-01
7.50382543e-01 -5.40903687e-01 -4.68914248e-02 8.36993694e-01
-2.63561726e-01 5.50184667e-01 -3.17605048e-01 -2.90202439e-01
-1.22557640e+00 4.55665827e-01 1.75846368e-01 -4.44634438e-01
1.51644707e-01 6.96001470e-01 -8.85364339e-02 -8.86669338e-01
-2.39736497e-01 9.41293221e-03 -6.32342577e-01 7.02185214e-01
2.42320627e-01 1.22790389e-01 2.43174762e-01 -1.25552154e+00
-4.45043564e-01 6.03681207e-01 -1.82934970e-01 -2.10634828e-01
1.30256104e+00 -1.88445151e-01 -5.86684465e-01 1.03161681e+00
1.40065622e+00 5.04714966e-01 -2.59496272e-01 -6.11994825e-02
1.66396424e-01 -3.90459113e-02 -2.02980399e-01 -1.09104681e+00
-1.08974433e+00 6.97908282e-01 2.11660013e-01 4.13147956e-02
1.02858400e+00 4.88130711e-02 5.71543038e-01 4.17431265e-01
5.07716060e-01 -1.69565129e+00 5.04445471e-02 9.90436912e-01
8.66799712e-01 -1.49907589e+00 -3.34365759e-03 -8.62856880e-02
-1.12366068e+00 7.27538288e-01 8.24380338e-01 1.08065382e-01
9.24269617e-01 3.84121954e-01 7.46227801e-01 -2.79862225e-01
-1.11200452e+00 1.39756454e-02 8.99494439e-02 2.33046174e-01
1.40251338e+00 3.09363544e-01 -7.99481452e-01 9.09316540e-01
-8.28449488e-01 -3.96132648e-01 5.73471546e-01 8.69347036e-01
-8.04699212e-02 -1.39053142e+00 -1.53466746e-01 4.63436186e-01
-9.89039838e-01 -3.27451497e-01 -7.46738255e-01 6.29756153e-01
2.84403086e-01 1.09765530e+00 -2.36197799e-01 -2.78284967e-01
2.18262672e-01 -1.45173045e-02 2.10909039e-01 -6.02422655e-01
-1.06983459e+00 1.55877724e-01 6.12735391e-01 -2.91702420e-01
-9.53192294e-01 -9.93335962e-01 -1.03218043e+00 -2.78879076e-01
2.11717598e-02 2.37893596e-01 8.14554632e-01 7.28446484e-01
1.97454169e-01 2.40812063e-01 5.25280058e-01 -3.66715640e-01
2.58642226e-01 -9.35403764e-01 -5.15824974e-01 4.78306800e-01
-3.21709886e-02 -4.54470575e-01 -9.53635722e-02 1.56574681e-01] | [11.184361457824707, 7.085400104522705] |
d3737dd5-723b-414f-a0fd-5c07ccebd056 | pfml-based-semantic-bci-agent-for-game-of-go | 1901.02999 | null | http://arxiv.org/abs/1901.02999v1 | http://arxiv.org/pdf/1901.02999v1.pdf | PFML-based Semantic BCI Agent for Game of Go Learning and Prediction | This paper presents a semantic brain computer interface (BCI) agent with
particle swarm optimization (PSO) based on a Fuzzy Markup Language (FML) for Go
learning and prediction applications. Additionally, we also establish an Open
Go Darkforest (OGD) cloud platform with Facebook AI research (FAIR) open source
Darkforest and ELF OpenGo AI bots. The Japanese robot Palro will simultaneously
predict the move advantage in the board game Go to the Go players for reference
or learning. The proposed semantic BCI agent operates efficiently by the
human-based BCI data from their brain waves and machine-based game data from
the prediction of the OGD cloud platform for optimizing the parameters between
humans and machines. Experimental results show that the proposed human and
smart machine co-learning mechanism performs favorably. We hope to provide
students with a better online learning environment, combining different kinds
of handheld devices, robots, or computer equipment, to achieve a desired and
intellectual learning goal in the future. | ['Yi-Hsiu Lee', 'Lu-An Lin', 'Sheng-Chi Yang', 'Yi-Lin Tsai', 'Bo-Yu Tsai', 'Li-Wei Ko', 'Mei-Hui Wang', 'Chang-Shing Lee', 'Nan Shuo', 'Hirofumi Ohashi', 'Naoyuki Kubota'] | 2019-01-10 | null | null | null | null | ['game-of-go'] | ['playing-games'] | [-4.96561199e-01 -1.18906803e-01 -4.60366998e-03 3.31430836e-03
4.01959926e-01 -2.40716096e-02 2.16356337e-01 -5.42889178e-01
-5.95200121e-01 9.61160064e-01 -4.48558122e-01 -3.94370370e-02
-5.84676862e-01 -1.00522614e+00 -4.88546729e-01 -6.16145253e-01
6.26750141e-02 7.85277128e-01 3.81063372e-01 -8.89061749e-01
6.08583272e-01 3.07791919e-01 -2.23940372e+00 3.60719524e-02
1.41892874e+00 1.17843974e+00 1.19145536e+00 6.50980353e-01
2.66225219e-01 1.05778408e+00 -6.04066908e-01 -3.00803483e-01
3.18504810e-01 -2.46245340e-01 -7.92451620e-01 -9.80640650e-01
-3.61013025e-01 1.34840205e-01 -3.33816499e-01 1.50998855e+00
7.78531969e-01 4.15762693e-01 5.02530873e-01 -1.99155700e+00
-5.45074642e-01 5.29554188e-01 -3.37038003e-02 5.80603600e-01
5.07276893e-01 2.49351650e-01 3.21529210e-01 -2.23710999e-01
6.49636507e-01 1.00521111e+00 4.94733512e-01 6.39810920e-01
3.78641598e-02 -1.17062950e+00 -3.40406984e-01 1.46343029e+00
-1.35064042e+00 1.95048109e-01 6.61918461e-01 -1.09050311e-01
1.19789898e+00 1.95136040e-01 1.59867334e+00 9.25105631e-01
7.69400537e-01 7.83574164e-01 1.01564181e+00 -4.67175633e-01
5.00543296e-01 1.37729838e-01 1.86125994e-01 1.11328793e+00
4.35753733e-01 9.37440991e-02 -9.51704025e-01 2.39171013e-01
5.87891579e-01 -1.03946351e-01 -3.46553445e-01 -3.57333086e-02
-7.58148015e-01 7.62127936e-01 4.25670087e-01 6.37566507e-01
-6.38130963e-01 2.75326073e-02 -4.23098892e-01 2.77194858e-01
-7.99545553e-03 5.75043321e-01 -5.01767159e-01 -6.39119744e-01
-3.79786342e-01 3.22357357e-01 1.04595423e+00 1.06004024e+00
6.22494042e-01 8.91911015e-02 2.40847319e-01 6.47746861e-01
4.67458785e-01 6.97772086e-01 1.41737032e+00 -1.14694655e+00
6.20803460e-02 4.87648726e-01 1.04963042e-01 -1.24326468e+00
-5.97021341e-01 -2.00960413e-01 -2.27178603e-01 6.72376335e-01
6.49438426e-02 -3.75880986e-01 -3.96758556e-01 1.00305426e+00
3.70189875e-01 6.98362172e-01 2.40510851e-01 1.06899953e+00
1.07232940e+00 4.57005084e-01 -2.23046094e-02 2.95075059e-01
1.48723555e+00 -1.31712425e+00 -9.46883202e-01 -2.11944774e-01
9.02619243e-01 -7.45128691e-02 9.45855856e-01 7.43655980e-01
-9.30270314e-01 -5.48844874e-01 -1.31876433e+00 3.56963784e-01
-8.89875770e-01 -5.14982715e-02 1.06575656e+00 9.04500961e-01
-1.26537752e+00 8.81857514e-01 -8.33291709e-01 -5.44560015e-01
7.87726939e-01 8.97860050e-01 -4.51797135e-02 1.29788041e-01
-1.43519843e+00 1.33995891e+00 4.13707405e-01 -3.56304407e-01
-6.26721144e-01 -6.42861605e-01 -4.16803092e-01 1.28722072e-01
-5.99811822e-02 -7.94777274e-01 9.26528573e-01 -1.06377602e+00
-2.24224019e+00 7.54052579e-01 4.89382386e-01 -6.77721024e-01
1.52730420e-01 7.63087124e-02 -3.66390347e-01 2.05450624e-01
1.75619517e-02 7.97727823e-01 5.09837091e-01 -6.75217330e-01
-1.34596050e+00 -6.98803246e-01 -2.39265244e-02 9.10703838e-01
-5.07972360e-01 -5.74662536e-02 4.61136401e-02 -9.18986872e-02
-2.85414487e-01 -1.04656112e+00 1.20740384e-01 -5.25841177e-01
2.85259098e-01 -6.21717453e-01 7.56845832e-01 -5.49749076e-01
7.75974810e-01 -1.68110073e+00 -8.04145783e-02 1.97525501e-01
1.43982247e-01 3.23507816e-01 6.06625061e-03 -4.86738235e-01
2.71976948e-01 -2.53595561e-01 6.81519151e-01 5.92344344e-01
-1.19469687e-01 1.26970485e-01 3.08279067e-01 2.23421320e-01
-8.77027690e-01 9.13552582e-01 -1.07246852e+00 -3.83140653e-01
1.73392892e-01 1.53769583e-01 -6.83863521e-01 -1.09283984e-01
3.06067020e-01 4.05588716e-01 -7.16847301e-01 4.70820129e-01
3.77585918e-01 1.35949571e-02 -2.79368460e-01 2.72560894e-01
-1.44138917e-01 1.25044556e-02 -1.05902457e+00 1.56126785e+00
-4.43850070e-01 5.91247976e-01 5.13792001e-02 -1.06312859e+00
8.92976820e-01 8.95185545e-02 6.57091439e-01 -9.88809347e-01
7.29775786e-01 2.98585951e-01 1.68300912e-01 -8.59107733e-01
2.29065880e-01 5.06353259e-01 5.25879145e-01 3.35759580e-01
5.31768858e-01 -4.71304744e-01 6.47606552e-02 -2.08501473e-01
1.23125970e+00 1.96870774e-01 1.08868964e-01 -6.82313979e-01
4.40542817e-01 3.62476975e-01 2.79421926e-01 8.06071460e-01
-6.31815314e-01 -2.39193112e-01 -2.82438546e-01 -3.07311624e-01
-2.43005186e-01 -6.31411254e-01 3.21080595e-01 1.40842891e+00
7.73905337e-01 2.12385003e-02 -1.13710582e+00 -2.01324195e-01
-3.12237173e-01 9.75704730e-01 -1.13144845e-01 -4.58548933e-01
-5.22162542e-02 -7.26359248e-01 1.77227378e-01 7.70073608e-02
1.04020536e+00 -1.64262807e+00 -1.02739918e+00 1.50064766e-01
-9.40910578e-02 -6.93544507e-01 -2.55905110e-02 2.09295824e-01
-7.12773204e-01 -9.47139263e-01 -4.62951958e-01 -1.44258821e+00
3.58002633e-02 3.73044699e-01 5.24422348e-01 2.10349053e-01
-2.16553673e-01 5.96976280e-01 -6.18928730e-01 -1.10778046e+00
4.30552810e-02 5.93387708e-02 4.79415745e-01 -4.83246177e-01
7.29293227e-01 -8.94936800e-01 -6.53409362e-01 1.85329825e-01
-4.72232513e-03 4.71542686e-01 -3.55208665e-02 5.87034881e-01
1.61858812e-01 6.19704902e-01 6.84187114e-01 -9.51893255e-02
9.73804891e-01 -5.27425468e-01 -6.61284029e-01 3.42671752e-01
-9.52135086e-01 -1.48198932e-01 3.66004229e-01 -4.47650880e-01
-1.14645958e+00 -2.20228240e-01 4.15721685e-02 -2.90076673e-01
-2.33862400e-01 1.60876229e-01 -2.68966377e-01 -8.21672201e-01
8.28698218e-01 3.05079877e-01 5.63397445e-02 6.44716844e-02
1.12335328e-02 1.36482763e+00 7.55728364e-01 -1.22388825e-03
6.46688938e-02 2.43200824e-01 -3.06421578e-01 -5.79252899e-01
-3.59312326e-01 -1.25332147e-01 -1.93692595e-01 -8.65466952e-01
1.01661468e+00 -8.88218641e-01 -1.64160609e+00 8.07828248e-01
-8.54963660e-01 -6.52623773e-01 4.78473715e-02 8.87229383e-01
-9.48504388e-01 -3.02951366e-01 -4.88912076e-01 -7.12406933e-01
-5.59169650e-01 -1.09250653e+00 1.80259749e-01 1.15966177e+00
8.52483362e-02 -7.57938504e-01 -1.24559194e-01 9.63508248e-01
4.44893181e-01 -5.24631381e-01 4.91863847e-01 -7.44854569e-01
-1.98773205e-01 -6.54615238e-02 3.28983486e-01 -8.91016498e-02
-2.79803425e-01 -5.60168386e-01 -5.58481574e-01 3.34416419e-01
4.35116470e-01 -9.33188796e-02 9.22788084e-02 5.04922271e-01
1.37386286e+00 -1.83762342e-01 -5.67853928e-01 5.48115790e-01
1.23061597e+00 1.11709952e+00 7.58092403e-01 1.01833868e+00
5.72858095e-01 3.03893417e-01 7.14982033e-01 5.05257905e-01
8.20889175e-01 2.93307155e-01 5.23438394e-01 7.74619877e-01
2.77106147e-02 -3.41375358e-02 2.47461259e-01 8.50078404e-01
-7.40509272e-01 -1.59257859e-01 -1.12085187e+00 2.48604566e-01
-2.00806737e+00 -1.04007697e+00 -4.55731563e-02 1.73779106e+00
4.88026887e-01 -2.28634253e-01 -6.96759000e-02 2.15799004e-01
8.24588060e-01 -9.08558369e-01 -5.05746305e-01 -3.03425968e-01
-1.88042466e-02 6.44786656e-01 6.60518229e-01 1.53259277e-01
-6.74530923e-01 1.52825737e+00 5.10490417e+00 1.40071595e+00
-1.27733028e+00 5.32101274e-01 2.94228315e-01 -1.33570060e-01
1.11256368e-01 -4.93768752e-01 -8.92168939e-01 8.60710800e-01
1.21724617e+00 -6.84154689e-01 1.53089404e+00 9.75019932e-01
2.25964069e-01 -5.69263577e-01 -1.89345688e-01 1.66674292e+00
4.48307842e-02 -1.50773799e+00 -3.69736284e-01 -1.42456919e-01
8.18823457e-01 3.76391858e-01 -1.39135784e-02 4.31741536e-01
5.83419859e-01 -9.46548522e-01 9.77249801e-01 8.74200702e-01
1.21708490e-01 -9.40368474e-01 6.92501783e-01 6.62942886e-01
-5.83025038e-01 -6.67081535e-01 -4.65855300e-01 -5.87571859e-01
-6.14094138e-01 -1.98740259e-01 -5.10774255e-01 3.45635474e-01
1.35495114e+00 6.39849305e-01 -4.38669503e-01 1.15307963e+00
-1.29380956e-01 3.80795002e-01 -4.30866420e-01 -1.10619462e+00
-6.17589243e-02 -5.61034501e-01 5.60354173e-01 1.53112277e-01
7.70450234e-01 6.62175953e-01 -3.98621857e-01 4.60848510e-01
1.14940703e-01 1.73318520e-01 -5.75903594e-01 2.61053592e-02
6.07208073e-01 1.30402231e+00 -9.79630351e-01 -1.71416000e-01
-1.65321544e-01 9.78276312e-01 1.23412341e-01 -8.67862813e-03
-9.83376920e-01 -7.63675332e-01 6.49453461e-01 -2.40543604e-01
4.51262575e-03 1.16439573e-02 -7.81709135e-01 -9.45801556e-01
-6.64612293e-01 -6.47380412e-01 2.52735674e-01 -1.63138676e+00
-8.59899163e-01 6.74202979e-01 -3.34165901e-01 -1.09812367e+00
-5.51724099e-02 -8.58063519e-01 -6.48898244e-01 3.76745820e-01
-1.27036166e+00 -5.22939980e-01 -5.86238682e-01 1.17010188e+00
2.64618397e-01 -9.12311435e-01 8.64233732e-01 7.98131377e-02
-4.07067388e-01 3.45814735e-01 2.49294043e-01 -3.79383892e-01
1.82877377e-01 -6.53945625e-01 -6.00820959e-01 -5.76988701e-03
1.54386207e-01 -2.02007532e-01 5.69514871e-01 -6.33504152e-01
-1.50353551e+00 -6.13541424e-01 3.21995437e-01 -1.80907398e-01
6.17891133e-01 1.91676319e-01 -1.46973029e-01 2.61427909e-01
3.73808265e-01 -3.63532603e-01 4.54176128e-01 -3.36727411e-01
9.77756321e-01 -2.27835372e-01 -1.44600344e+00 8.89143467e-01
1.36464751e+00 8.17596540e-03 -5.40090919e-01 6.02321684e-01
3.07710946e-01 -4.59378839e-01 -5.09324372e-01 3.12802233e-02
4.36574668e-01 -7.74214268e-01 7.95809150e-01 -4.91983384e-01
3.89463305e-02 3.54662612e-02 1.71282127e-01 -1.67539561e+00
-2.69357026e-01 -6.91375673e-01 -3.26874256e-02 3.66181463e-01
9.21582580e-02 -9.26057100e-01 1.17915106e+00 7.35109210e-01
-2.27236271e-01 -7.46101260e-01 -1.06753492e+00 -8.00751328e-01
-2.06017628e-01 -6.99216247e-01 7.91125655e-01 7.46976376e-01
7.31095552e-01 1.78305227e-02 3.68266255e-02 6.69981465e-02
2.20616475e-01 -2.34048069e-01 3.41711193e-01 -1.48264432e+00
2.42414922e-01 -4.64229196e-01 -1.01632035e+00 -3.62169921e-01
4.06894565e-01 -1.14059854e+00 -4.33109440e-02 -1.46100736e+00
-6.22639582e-02 -3.21576566e-01 -3.90690684e-01 5.78859806e-01
9.91794467e-03 2.38897651e-01 2.35865325e-01 1.62550092e-01
-8.67921114e-01 5.99125922e-01 1.59471488e+00 -6.70495927e-02
-6.09786928e-01 -8.66519213e-02 -7.40936458e-01 1.12641728e+00
1.06702185e+00 -6.31493986e-01 -3.42417538e-01 -2.50558376e-01
1.77261129e-01 1.35758296e-01 4.47513074e-01 -1.98846376e+00
9.99089360e-01 -2.20167145e-01 3.85686636e-01 1.78082678e-02
1.59583464e-01 -8.70841444e-01 5.93045093e-02 8.79921615e-01
1.47473991e-01 -2.72986948e-01 3.29283886e-02 2.64256179e-01
3.44123870e-01 -4.53068256e-01 7.68963754e-01 -1.01801574e-01
-9.98174131e-01 2.48241991e-01 -9.37273741e-01 -1.24297209e-01
1.50161397e+00 -6.07481956e-01 -3.47279429e-01 -6.97379589e-01
-8.98270309e-01 3.18670183e-01 -1.13116220e-01 6.39002204e-01
5.66179395e-01 -1.20860672e+00 -5.46550900e-02 2.97306091e-01
-5.24676859e-01 -5.40838599e-01 3.36061895e-01 7.32938051e-01
-1.05677307e+00 4.00565982e-01 -1.11985707e+00 -1.18394054e-01
-1.27071524e+00 1.12795636e-01 5.19266665e-01 2.61870790e-02
-4.43674922e-01 1.04530120e+00 -5.59277415e-01 -5.45073152e-01
2.99171329e-01 2.22321719e-01 -7.58120835e-01 -2.09641933e-01
4.79041427e-01 8.80425990e-01 1.12373091e-01 -5.41989923e-01
-2.31266156e-01 -1.88148450e-02 5.38987160e-01 -1.55497372e-01
1.67551363e+00 -1.20054387e-01 1.15062095e-01 -8.09859559e-02
4.58462507e-01 -5.77404022e-01 -8.04238081e-01 5.73582411e-01
-4.26976606e-02 -1.69007018e-01 4.83045608e-01 -1.13212216e+00
-1.30036998e+00 4.12725955e-01 1.32254910e+00 -1.43916667e-01
1.11045909e+00 -2.52729923e-01 9.63447928e-01 8.26966703e-01
1.21342742e+00 -1.76102388e+00 -1.44923469e-02 6.87598467e-01
5.79719186e-01 -9.91157591e-01 -3.61511648e-01 -1.04192920e-01
-9.26396728e-01 1.05692673e+00 1.07899380e+00 -4.30544764e-01
1.08942437e+00 2.34914184e-01 -1.02947643e-02 -3.14508229e-01
-5.87828934e-01 -2.77028412e-01 -1.93730891e-01 1.15151799e+00
-2.66444772e-01 2.43031278e-01 -5.28328419e-01 1.56825352e+00
-9.20481086e-01 6.34001493e-01 6.73122585e-01 8.96719337e-01
-8.54915380e-01 -6.48595333e-01 -5.65287232e-01 7.28577137e-01
-2.51123875e-01 3.93589996e-02 4.18722257e-02 4.75602388e-01
6.49939537e-01 1.37531960e+00 2.84831040e-03 -8.09423804e-01
4.44710627e-02 -2.00118683e-03 6.41419351e-01 -1.84248790e-01
-7.11749017e-01 -7.99955487e-01 -4.09774691e-01 -7.63918936e-01
-8.97129625e-02 -3.67806941e-01 -1.87991428e+00 -4.62257504e-01
-4.87347156e-01 3.15209538e-01 1.15380037e+00 9.69076812e-01
6.91094518e-01 3.91542554e-01 2.62610257e-01 -1.01514530e+00
1.56539470e-01 -9.40162539e-01 -1.01031423e+00 -1.61354765e-01
-8.14195216e-01 -1.11479723e+00 -9.31991041e-02 -1.77306071e-01] | [3.8231375217437744, 1.4930357933044434] |
7b5a96b1-56ad-4fa9-b0dc-7596a7c45660 | improving-speech-to-speech-translation | 2210.14514 | null | https://arxiv.org/abs/2210.14514v1 | https://arxiv.org/pdf/2210.14514v1.pdf | Improving Speech-to-Speech Translation Through Unlabeled Text | Direct speech-to-speech translation (S2ST) is among the most challenging problems in the translation paradigm due to the significant scarcity of S2ST data. While effort has been made to increase the data size from unlabeled speech by cascading pretrained speech recognition (ASR), machine translation (MT) and text-to-speech (TTS) models; unlabeled text has remained relatively under-utilized to improve S2ST. We propose an effective way to utilize the massive existing unlabeled text from different languages to create a large amount of S2ST data to improve S2ST performance by applying various acoustic effects to the generated synthetic data. Empirically our method outperforms the state of the art in Spanish-English translation by up to 2 BLEU. Significant gains by the proposed method are demonstrated in extremely low-resource settings for both Spanish-English and Russian-English translations. | ['Hongyu Gong', 'Ilia Kulikov', 'Yun Tang', 'Changhan Wang', 'Sravya Popuri', 'Xuan-Phi Nguyen'] | 2022-10-26 | null | null | null | null | ['speech-to-speech-translation'] | ['speech'] | [ 4.80351865e-01 8.18197355e-02 -1.55411050e-01 -3.82421941e-01
-1.64420199e+00 -6.52136266e-01 9.25725520e-01 -5.57667971e-01
-2.50592232e-01 9.75174367e-01 3.94419372e-01 -7.28045821e-01
7.26050854e-01 -1.19895935e-01 -6.57880366e-01 -4.86349314e-01
6.88953042e-01 7.87665009e-01 3.89071405e-02 -5.28668940e-01
-1.24478109e-01 2.76282191e-01 -9.23435986e-01 1.97341025e-01
1.20790756e+00 5.03205478e-01 5.79477251e-01 4.63510692e-01
-3.35565478e-01 1.83587372e-01 -6.20910704e-01 -4.93245721e-01
4.70745504e-01 -8.88444006e-01 -5.56315958e-01 2.13212982e-01
8.26972201e-02 -1.16056643e-01 -2.13831276e-01 1.08048129e+00
7.37556100e-01 1.17091469e-01 5.87114751e-01 -8.85765731e-01
-6.86944485e-01 7.69312501e-01 -3.89801353e-01 1.29224762e-01
1.92759216e-01 -5.90946674e-02 6.26405835e-01 -1.53853500e+00
6.38341367e-01 1.39120471e+00 1.89639553e-01 6.58689737e-01
-1.13775718e+00 -6.03324056e-01 -3.71002972e-01 -5.24650328e-02
-1.42882359e+00 -9.76064324e-01 4.31702971e-01 -3.64331976e-02
1.10751259e+00 3.35112453e-01 1.74492627e-01 1.56108201e+00
-2.34409139e-01 8.05517614e-01 1.21541417e+00 -9.38645840e-01
1.74000964e-01 4.41383243e-01 -5.25512159e-01 1.67449862e-01
-2.39936665e-01 4.91094813e-02 -5.98649681e-01 6.57285601e-02
6.54602647e-01 -4.69303876e-01 4.07206267e-02 3.58326435e-01
-1.55802906e+00 6.07653677e-01 -2.07427517e-01 3.75181884e-01
-2.68688351e-01 -2.58558989e-01 4.36257869e-01 7.04618216e-01
7.98810065e-01 2.84383893e-01 -6.55584633e-01 -5.34249365e-01
-1.06392062e+00 -6.73437193e-02 6.51416004e-01 1.37891066e+00
5.36650658e-01 5.90198994e-01 -7.20612407e-02 1.36766422e+00
1.46740124e-01 1.26490569e+00 8.67113650e-01 -6.16162837e-01
1.23631692e+00 8.66049901e-02 2.80391842e-01 -2.98405439e-01
3.59667122e-01 -4.60886180e-01 -6.53502166e-01 -3.67432743e-01
2.32775271e-01 -3.83369178e-01 -1.27987838e+00 1.57683599e+00
1.83194757e-01 -1.48640946e-01 4.53243196e-01 7.94985652e-01
4.74820405e-01 1.13564837e+00 -2.40795240e-01 -5.05313873e-01
1.00631738e+00 -1.32785642e+00 -8.85587394e-01 -4.77934420e-01
6.56945527e-01 -1.30239618e+00 1.24115157e+00 -9.39876586e-02
-1.15554214e+00 -6.35390818e-01 -8.01591814e-01 7.00875074e-02
-1.12077534e-01 4.25900519e-01 1.59022454e-02 8.07636261e-01
-1.09829628e+00 5.09429276e-02 -8.69224012e-01 -5.35208344e-01
-2.36538835e-02 3.76690388e-01 -3.40996653e-01 -2.36840189e-01
-1.28139198e+00 9.99202728e-01 1.40647531e-01 1.61571041e-01
-9.27376151e-01 -1.05198160e-01 -7.89671838e-01 -1.29993752e-01
4.78178948e-01 -4.57928956e-01 1.44547236e+00 -1.17147148e+00
-2.02247858e+00 5.05719006e-01 -5.71892023e-01 -2.93155521e-01
5.56354284e-01 -6.16970174e-02 -6.67508185e-01 8.42577368e-02
1.20721743e-01 5.85241795e-01 9.48367178e-01 -1.10772419e+00
-4.62622792e-01 -3.78374338e-01 -7.42366672e-01 5.76043427e-01
-3.06690156e-01 6.30823076e-01 -4.67113972e-01 -1.07600880e+00
3.32557589e-01 -1.31975937e+00 -1.59742862e-01 -6.84660971e-01
-4.87806618e-01 5.16581833e-02 8.49645615e-01 -9.98900175e-01
9.74901736e-01 -1.90667617e+00 2.28067786e-01 -1.70642167e-01
-4.89276320e-01 5.86741507e-01 -4.12873477e-01 8.25874388e-01
1.35120228e-01 8.72376859e-02 -2.70519793e-01 -6.74210191e-01
-1.15585171e-01 3.18230957e-01 -4.02784288e-01 -8.47612545e-02
2.06794098e-01 1.11353302e+00 -7.52936244e-01 -4.07984555e-01
1.76042225e-02 3.16219330e-01 -3.74731719e-02 3.14926624e-01
-1.62972555e-01 6.72256827e-01 -4.98709530e-01 7.65103400e-01
3.65986288e-01 5.35641834e-02 1.48442045e-01 4.26900953e-01
-7.73377940e-02 6.81948245e-01 -6.56181276e-01 1.60400558e+00
-5.83824396e-01 6.38153613e-01 -1.72129590e-02 -8.51764023e-01
1.10641754e+00 7.58224308e-01 1.35309279e-01 -7.26744354e-01
1.41980248e-02 8.53914738e-01 1.98884740e-01 -4.46920633e-01
6.17942929e-01 -3.99563491e-01 -1.03156991e-01 3.35691601e-01
1.01569466e-01 -3.88235897e-01 -6.06666654e-02 2.12663263e-02
8.37502420e-01 1.85202450e-01 2.83696447e-02 -5.74708581e-02
4.24814135e-01 2.21419096e-01 4.86612409e-01 3.51367235e-01
-8.35959539e-02 7.14551389e-01 -1.83777839e-01 1.25869617e-01
-1.53835630e+00 -1.06633222e+00 2.76270658e-01 9.73137736e-01
-2.35546812e-01 -7.64259472e-02 -9.74869549e-01 -5.56385756e-01
-5.99835932e-01 9.61543560e-01 2.72878222e-02 1.01441540e-01
-8.32467020e-01 -6.99961543e-01 8.54372323e-01 2.64124990e-01
5.02169907e-01 -1.04741263e+00 3.52271795e-01 5.01657486e-01
-7.71193326e-01 -1.54717588e+00 -8.81500244e-01 2.91651189e-02
-9.21812832e-01 -1.12698026e-01 -1.26404476e+00 -9.68979299e-01
5.52145898e-01 4.38318431e-01 8.00646842e-01 -6.94303393e-01
3.27461809e-01 -2.57200152e-01 -6.12767577e-01 -2.81587839e-01
-1.23520839e+00 2.77043521e-01 3.75125468e-01 3.82598341e-02
2.95203090e-01 -3.70647281e-01 -1.13496378e-01 6.73297286e-01
-6.47224009e-01 1.83916211e-01 7.54797578e-01 9.49982524e-01
4.72193956e-01 -3.05187345e-01 9.30264115e-01 -6.38855159e-01
6.02976859e-01 -2.73522139e-01 -4.52385217e-01 3.44354033e-01
-6.50557935e-01 7.77362706e-03 9.04585540e-01 -6.87020659e-01
-1.27809310e+00 -1.50388747e-01 -1.67592719e-01 -4.31037188e-01
1.12719776e-03 4.38806981e-01 -2.87546903e-01 1.62012279e-01
6.34103596e-01 7.55083025e-01 -4.18524630e-02 -5.15884578e-01
3.69706392e-01 1.37406611e+00 4.19584364e-01 -4.79280859e-01
9.39925015e-01 -1.31642565e-01 -4.00098532e-01 -9.41139519e-01
-4.37346727e-01 -2.96358079e-01 -4.94636565e-01 1.57767206e-01
5.96253812e-01 -1.11148417e+00 2.74660110e-01 4.73559618e-01
-1.11408627e+00 -3.64494592e-01 5.80247566e-02 8.14611793e-01
-4.42947060e-01 3.74883413e-01 -7.24466085e-01 -8.99378955e-01
-5.37210286e-01 -1.64914310e+00 1.26471376e+00 -2.77200431e-01
-6.19031079e-02 -6.80743873e-01 2.16662586e-02 8.30366611e-01
5.92449009e-01 -2.81752884e-01 8.20750117e-01 -8.97283792e-01
-4.41630274e-01 -1.98457599e-01 -8.23642015e-02 6.36570752e-01
4.40559000e-01 -1.98562071e-01 -6.44582570e-01 -3.76413792e-01
1.14057459e-01 -3.54859591e-01 3.83812755e-01 1.50192499e-01
3.18956941e-01 -3.81504804e-01 -9.31091607e-02 1.38977528e-01
1.00498438e+00 4.50832248e-01 6.61859751e-01 5.31486906e-02
6.79406762e-01 5.41133106e-01 6.89341486e-01 -1.30062485e-02
3.69919807e-01 9.58213925e-01 -3.65372211e-01 -1.89041242e-01
-3.59262526e-01 -4.08871949e-01 9.45588708e-01 1.77295280e+00
1.09187171e-01 -5.52275240e-01 -9.71795797e-01 6.26013577e-01
-1.64590180e+00 -7.48418272e-01 -1.34087309e-01 2.33121538e+00
1.04323006e+00 5.02897650e-02 5.23224287e-02 -1.19298838e-01
9.82585907e-01 -1.09203629e-01 -3.21947277e-01 -5.30398190e-01
-2.43507370e-01 1.51391402e-01 4.40772980e-01 5.79804718e-01
-5.63632548e-01 1.50279164e+00 6.56802130e+00 1.23169494e+00
-1.31131923e+00 3.01186085e-01 6.78317964e-01 9.01087672e-02
-3.16469729e-01 1.15297027e-01 -8.17353666e-01 5.83993077e-01
1.55110657e+00 -2.06720784e-01 7.27976799e-01 5.93913138e-01
7.50642419e-01 1.63863972e-01 -8.42976213e-01 9.81141746e-01
5.89277185e-02 -1.02331674e+00 1.87456459e-01 6.16915785e-02
1.08400846e+00 4.90755469e-01 8.26426595e-02 4.18044269e-01
2.84095585e-01 -7.50319839e-01 7.03363359e-01 -2.40704522e-01
1.44557536e+00 -3.93502593e-01 5.55341244e-01 5.22025168e-01
-9.67059791e-01 2.51636684e-01 -4.09206450e-01 2.59377569e-01
3.37247610e-01 3.38270009e-01 -1.44179428e+00 6.38200521e-01
1.43215090e-01 2.08678439e-01 -1.60777345e-01 3.88954461e-01
-1.64084107e-01 1.11257827e+00 -3.79811466e-01 -1.00139685e-01
5.55550575e-01 -4.67786252e-01 5.99906683e-01 1.18572176e+00
8.17667782e-01 -1.47684008e-01 1.42444804e-01 4.52954173e-01
-3.32378864e-01 5.38680792e-01 -6.21974468e-01 -3.89129013e-01
6.92054510e-01 8.47476780e-01 -5.38584173e-01 -6.91526592e-01
-4.86028612e-01 1.34522641e+00 -4.14397046e-02 6.35592222e-01
-5.20344257e-01 -1.99044377e-01 3.15160692e-01 2.07918435e-02
1.83295548e-01 -4.55834806e-01 -1.96167976e-01 -1.35576820e+00
2.28870705e-01 -1.37422180e+00 -1.99277192e-01 -8.70465696e-01
-9.53173518e-01 1.09353471e+00 -2.17839241e-01 -1.47901535e+00
-7.55010366e-01 -2.01176599e-01 -7.98715428e-02 1.37706101e+00
-1.29211748e+00 -1.23459041e+00 3.26935410e-01 3.94277096e-01
1.37853599e+00 -7.07142234e-01 8.55510235e-01 3.55650604e-01
-4.59616721e-01 6.94933176e-01 7.77568519e-01 -1.62412152e-02
9.33240831e-01 -9.25150692e-01 1.12493539e+00 1.01027894e+00
4.19509172e-01 5.04767060e-01 6.71388090e-01 -7.75741756e-01
-1.48630905e+00 -1.02961838e+00 1.27683699e+00 -3.93189937e-01
6.82705462e-01 -6.54594541e-01 -6.15473628e-01 6.01964295e-01
3.94433290e-01 -1.34393916e-01 4.82615054e-01 -3.33480805e-01
-2.45163396e-01 6.39043972e-02 -9.73591566e-01 9.87930238e-01
9.21557724e-01 -7.30803609e-01 -5.96222520e-01 2.19237208e-01
1.02125764e+00 -3.03862721e-01 -5.91959715e-01 2.44358152e-01
2.79387653e-01 -2.14643344e-01 5.74777842e-01 -4.99757767e-01
4.07089949e-01 -2.44307533e-01 -6.20045900e-01 -1.76260102e+00
1.69920355e-01 -1.19300592e+00 3.06591183e-01 1.19060814e+00
9.78758395e-01 -6.18882895e-01 6.41662240e-01 3.63059789e-01
-4.54045892e-01 -3.50719959e-01 -1.11243069e+00 -1.07421696e+00
1.50007233e-01 -2.63857424e-01 4.35206741e-01 8.58848095e-01
-8.40778928e-03 7.74205208e-01 -7.02587605e-01 8.91091898e-02
3.47341627e-01 -1.33940324e-01 8.30280304e-01 -5.80378532e-01
-3.16999376e-01 -3.47301289e-02 6.95909932e-02 -1.45893669e+00
8.41090903e-02 -9.50914919e-01 3.19918931e-01 -1.38648832e+00
1.10180199e-01 -3.60553086e-01 4.53094840e-02 3.28449309e-01
-3.56542826e-01 4.06007141e-01 1.52035698e-01 4.88255918e-01
-1.43067777e-01 8.23916912e-01 1.56583369e+00 -3.48000862e-02
-2.38007441e-01 1.33574218e-01 -3.64783853e-01 1.02892645e-01
6.88524902e-01 -6.41323984e-01 -5.01266181e-01 -6.17055357e-01
-3.91191363e-01 7.06183910e-01 -4.22037482e-01 -5.24980783e-01
-1.21895671e-01 -2.39143521e-01 4.90386272e-03 -5.43305814e-01
5.44388235e-01 -6.46453857e-01 1.80685923e-01 9.02776644e-02
-4.53528196e-01 3.23616624e-01 4.39074039e-02 4.07662630e-01
-4.64612484e-01 -6.24253042e-02 5.98014891e-01 -1.42891422e-01
-2.36481711e-01 1.48855910e-01 -5.60912728e-01 1.29432246e-01
5.80782890e-01 -1.37599289e-01 -1.45003974e-01 -6.13062739e-01
-5.09956658e-01 -7.14676902e-02 3.75480056e-01 7.03383148e-01
5.68253398e-01 -1.41618502e+00 -1.18358290e+00 2.52119511e-01
1.31193534e-01 -3.71987283e-01 -1.82170257e-01 7.32831717e-01
-3.15808773e-01 6.14076257e-01 1.46731034e-01 -5.43572843e-01
-1.23607898e+00 3.16598922e-01 -4.85479422e-02 -1.58480644e-01
-4.45642442e-01 5.72896004e-01 -5.63399750e-04 -6.67082906e-01
-8.63280892e-02 -9.47843343e-02 4.39910561e-01 -4.26722467e-01
2.28880599e-01 2.70078689e-01 2.80557334e-01 -9.85464275e-01
-1.62521273e-01 1.69362724e-01 -2.06453115e-01 -9.22627389e-01
1.06031215e+00 -5.25312543e-01 1.62827477e-01 4.35809761e-01
1.14053512e+00 2.23878771e-01 -9.16177571e-01 -4.37396526e-01
1.34594649e-01 -3.88323992e-01 -1.46397144e-01 -9.87658978e-01
-5.58702230e-01 9.01327908e-01 2.31869817e-01 -6.87283874e-02
8.70880663e-01 -2.00594231e-01 1.18211114e+00 5.65755010e-01
6.60647869e-01 -1.21021318e+00 -4.60728481e-02 5.41419923e-01
7.80877411e-01 -1.44361687e+00 -6.32559896e-01 -5.00044346e-01
-9.43307042e-01 8.58144581e-01 1.16292283e-01 3.57636303e-01
1.91760480e-01 3.40135634e-01 6.77971184e-01 5.36658287e-01
-7.50863373e-01 -7.48472735e-02 1.29462466e-01 4.32322681e-01
4.06435996e-01 1.93474069e-01 -3.49840850e-01 1.47295803e-01
-1.51329249e-01 -1.54892147e-01 5.61878741e-01 6.72836900e-01
-2.77537256e-01 -1.74240434e+00 -5.59027910e-01 2.42366239e-01
-6.75493479e-01 -5.82241654e-01 -4.74420220e-01 4.74692404e-01
-5.28142154e-01 1.33618093e+00 -5.27431130e-01 -3.30598712e-01
1.03944749e-01 3.48965019e-01 3.30375254e-01 -7.88786709e-01
-4.51268762e-01 7.44244397e-01 4.56522644e-01 -9.29615051e-02
-2.35659480e-01 -6.50595069e-01 -1.00487983e+00 -1.63791180e-01
-4.82956409e-01 3.41898531e-01 1.07771111e+00 1.01918614e+00
4.26516920e-01 1.75671875e-01 1.05511725e+00 -6.64868772e-01
-9.22812164e-01 -1.36302495e+00 -2.76671588e-01 2.03700259e-01
7.51278922e-02 -1.99811324e-01 -3.40535581e-01 2.88523585e-01] | [14.49428653717041, 7.1816558837890625] |
0b378734-a892-4755-a8f8-a071e5d60382 | building-extractive-question-answering-system | 2305.19707 | null | https://arxiv.org/abs/2305.19707v1 | https://arxiv.org/pdf/2305.19707v1.pdf | Building Extractive Question Answering System to Support Human-AI Health Coaching Model for Sleep Domain | Non-communicable diseases (NCDs) are a leading cause of global deaths, necessitating a focus on primary prevention and lifestyle behavior change. Health coaching, coupled with Question Answering (QA) systems, has the potential to transform preventive healthcare. This paper presents a human-Artificial Intelligence (AI) health coaching model incorporating a domain-specific extractive QA system. A sleep-focused dataset, SleepQA, was manually assembled and used to fine-tune domain-specific BERT models. The QA system was evaluated using automatic and human methods. A data-centric framework enhanced the system's performance by improving passage retrieval and question reformulation. Although the system did not outperform the baseline in automatic evaluation, it excelled in the human evaluation of real-world questions. Integration into a Human-AI health coaching model was tested in a pilot Randomized Controlled Trial (RCT). | ['Josip Car', 'Shafiq Joty', 'Qi Chwen Ong', 'Iva Bojic'] | 2023-05-31 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [ 4.05388474e-01 5.45226514e-01 -2.85538644e-01 -4.38719064e-01
-1.54100847e+00 -3.19911540e-01 3.97712976e-01 5.50385416e-01
-6.23815358e-01 7.57133067e-01 9.36160624e-01 -5.52563906e-01
-2.92869896e-01 -6.07658386e-01 -2.17054829e-01 -2.16538116e-01
4.74020422e-01 1.02168667e+00 1.33278549e-01 -3.04610789e-01
3.65366926e-03 -5.61885722e-02 -1.11474741e+00 6.97442889e-01
1.49792647e+00 3.64829838e-01 5.30398004e-02 1.17246675e+00
-4.35226224e-02 1.11923540e+00 -7.19594896e-01 -3.42034250e-01
-2.68900365e-01 -1.08344984e+00 -1.27039719e+00 -5.07785261e-01
7.44921863e-01 -6.75801188e-02 1.40229046e-01 5.37728548e-01
9.44171846e-01 -1.70126438e-01 2.62772232e-01 -8.36527169e-01
-7.64388859e-01 7.64568970e-02 5.02209663e-01 3.61787587e-01
9.00764108e-01 5.26070714e-01 8.68949234e-01 -3.14402252e-01
7.06966639e-01 1.10343087e+00 9.29846346e-01 9.17740107e-01
-1.40786302e+00 -1.58437371e-01 -7.62648940e-01 5.63697755e-01
-8.64229679e-01 -5.20929873e-01 3.27721447e-01 -1.79479584e-01
1.39732349e+00 8.78218353e-01 7.38060832e-01 8.54092300e-01
3.12721044e-01 4.63933617e-01 1.14985847e+00 -6.50554240e-01
3.21645617e-01 1.68487206e-01 3.03916156e-01 7.87675023e-01
2.07258731e-01 9.92283002e-02 -1.40965864e-01 -4.72205281e-01
6.40576929e-02 -2.00336546e-01 -1.20006159e-01 1.04581609e-01
-1.31119490e+00 1.07506537e+00 4.96641874e-01 5.88421702e-01
-8.16487134e-01 -1.58587798e-01 4.39271092e-01 4.62894022e-01
2.93563962e-01 9.65642989e-01 -8.58926892e-01 -1.78171217e-01
-1.14813042e+00 3.23231250e-01 1.27604640e+00 2.61963367e-01
2.24552542e-01 -4.37123120e-01 -1.08997071e+00 5.31023085e-01
3.55354995e-01 1.00196254e+00 6.18837833e-01 -1.39807606e+00
2.59817749e-01 9.39994812e-01 2.56251264e-02 -1.07735431e+00
-9.05110776e-01 -3.87792110e-01 -7.56954253e-01 -4.77376401e-01
5.23728490e-01 -2.51332164e-01 -9.34286237e-01 1.58757937e+00
3.69117856e-01 -3.97057116e-01 1.21624880e-01 5.72994173e-01
1.16827083e+00 6.58919096e-01 1.03488529e+00 -1.88097090e-01
1.92731929e+00 -1.04889286e+00 -1.15391004e+00 -6.86317682e-02
7.27294743e-01 -5.88011265e-01 1.30784965e+00 1.24870896e-01
-1.37319827e+00 -5.70607662e-01 -3.98359835e-01 -5.36536396e-01
-5.36834717e-01 1.76772192e-01 1.12665497e-01 9.83266532e-01
-1.23082757e+00 1.56940922e-01 -8.50632489e-01 -8.31746161e-01
3.35807115e-01 2.61929661e-01 -2.16037035e-01 -2.66678929e-01
-1.54942977e+00 1.35801852e+00 1.33974463e-01 -5.24467468e-01
-9.13006186e-01 -1.13430858e+00 -9.30709660e-01 1.02025002e-01
1.59935459e-01 -1.67152286e+00 1.63781345e+00 -7.14625895e-01
-1.35680377e+00 9.22500610e-01 -1.87538415e-01 -6.41674995e-01
8.79129693e-02 -4.31896597e-01 -8.49985421e-01 8.39638710e-01
2.90019572e-01 7.88555205e-01 3.47324371e-01 -6.16056681e-01
-5.25201321e-01 -4.41190392e-01 -3.52398187e-01 1.94426894e-01
-2.17587739e-01 2.08708152e-01 -1.02495342e-01 -3.31504166e-01
-7.13868558e-01 -7.00889170e-01 -4.61055547e-01 1.14601202e-01
-3.34677517e-01 -4.76851732e-01 1.44704700e-01 -1.23993659e+00
1.60407174e+00 -1.61310303e+00 -2.50667661e-01 -6.36178106e-02
1.83552206e-01 7.37318575e-01 -2.66608298e-01 7.38079131e-01
1.76030904e-01 1.24225236e-01 -1.88424915e-01 1.11131944e-01
-3.88634577e-02 4.51740265e-01 4.52417791e-01 9.78379548e-02
4.77756917e-01 1.22949445e+00 -1.14718318e+00 -6.79861128e-01
8.84224996e-02 5.38446069e-01 -7.81892061e-01 3.21077406e-01
-6.05745018e-01 3.41660380e-01 -4.90151584e-01 2.31725946e-01
-1.06519096e-01 -5.41332245e-01 -1.55007780e-01 -3.06817234e-01
2.63293162e-02 7.78887689e-01 -3.71965379e-01 1.68805707e+00
-2.30363339e-01 6.70310259e-02 -1.12423576e-01 -6.87202156e-01
4.26411152e-01 6.85673535e-01 4.34910506e-01 -1.33174062e+00
-2.15798214e-01 -1.73223484e-02 1.52912229e-01 -1.45519042e+00
1.07846938e-01 -2.28053004e-01 8.43240973e-03 2.42334962e-01
-6.82938769e-02 -3.88192907e-02 1.08566307e-01 3.08330387e-01
1.58565366e+00 -2.94973105e-01 7.37259507e-01 -4.10138786e-01
9.03716385e-01 9.70817745e-01 3.33392709e-01 6.55498266e-01
-4.93171602e-01 5.27700186e-01 -2.36236937e-02 -2.63503402e-01
-7.38642097e-01 -9.06903386e-01 -2.39600730e-03 1.22304606e+00
-5.50934076e-01 -5.75249970e-01 -1.03216982e+00 -8.40065539e-01
-3.99543345e-01 1.17479849e+00 -6.87581956e-01 -3.52023661e-01
-3.45518947e-01 -5.09142876e-01 7.51064241e-01 9.31254476e-02
5.89568973e-01 -1.32068253e+00 -8.41232836e-01 4.38993275e-01
-6.98272109e-01 -6.40805244e-01 -6.07295275e-01 -2.88778782e-01
-7.73568869e-01 -1.42226160e+00 -9.22029555e-01 -7.51084030e-01
2.09237456e-01 -1.01591557e-01 1.73669922e+00 7.41767883e-03
-3.43773961e-01 1.03225517e+00 -2.45612144e-01 -6.52781308e-01
-1.00751317e+00 3.22667748e-01 -4.96927381e-01 -9.24402595e-01
7.96249568e-01 5.54719307e-02 -1.27470565e+00 3.38758677e-02
-1.03402793e+00 -2.21489966e-02 7.25861430e-01 7.84420252e-01
4.34940845e-01 -3.17878604e-01 9.29058909e-01 -1.02764416e+00
9.67956364e-01 -5.83184600e-01 1.48208812e-01 4.67826247e-01
-9.25131381e-01 6.99678138e-02 3.11501265e-01 -1.72472641e-01
-1.10724211e+00 -4.56203409e-02 -7.78336644e-01 6.83360398e-01
-6.24296784e-01 6.29626274e-01 -1.26595557e-01 3.86357486e-01
1.27878582e+00 -2.69299243e-02 3.33865345e-01 -5.27703524e-01
5.38793623e-01 6.61181569e-01 6.12651050e-01 1.73143037e-02
4.11462247e-01 3.40292864e-02 -3.17923486e-01 -7.33371675e-01
-1.07378292e+00 -6.61812782e-01 3.50723676e-02 1.86373472e-01
1.52529228e+00 -8.70513499e-01 -5.15552580e-01 -7.56898373e-02
-9.56910610e-01 -5.42835236e-01 -6.45529628e-01 3.21800202e-01
-3.09402287e-01 1.16536096e-01 -5.18467903e-01 -2.46774256e-01
-1.13401210e+00 -5.29535592e-01 1.04982066e+00 3.93178731e-01
-8.09819639e-01 -1.06197572e+00 1.03876638e+00 9.55961168e-01
6.87435746e-01 4.31744345e-02 1.19977808e+00 -9.72846925e-01
7.78895244e-02 -1.47466958e-01 -1.45311490e-01 2.28699923e-01
2.20173597e-01 -5.77490687e-01 -5.67673206e-01 1.47798836e-01
1.33632289e-04 -3.26420128e-01 4.56860095e-01 4.27834064e-01
5.42477906e-01 -7.29395032e-01 -1.31449834e-01 -1.98572710e-01
1.06790483e+00 1.28878176e-01 8.49335909e-01 1.74654603e-01
2.78403401e-01 6.71469867e-01 3.35283965e-01 -1.11215889e-01
1.10878742e+00 3.68175417e-01 -9.13118348e-02 -5.57867110e-01
-6.04786277e-01 -6.98897243e-03 3.01348746e-01 7.21736252e-01
2.43829459e-01 -1.03098951e-01 -1.18432152e+00 7.10363030e-01
-1.40698302e+00 -9.49872136e-01 -4.13500130e-01 1.73163569e+00
1.38846695e+00 -1.92329913e-01 3.64377677e-01 -1.44191068e-02
2.40558181e-02 -5.19864798e-01 -3.99208397e-01 -6.08483374e-01
1.58328205e-01 6.93726540e-01 -2.47694068e-02 2.90198207e-01
-9.03295815e-01 4.73916024e-01 6.45570898e+00 4.80745286e-01
-4.75856781e-01 5.02824247e-01 5.58271706e-01 3.71427953e-01
-3.91256452e-01 -5.50579011e-01 -4.86420512e-01 2.93754578e-01
1.90910172e+00 -3.20183821e-02 -1.34247579e-02 3.61737311e-01
7.09042549e-01 -1.75607443e-01 -8.90945971e-01 2.23856539e-01
1.68333367e-01 -1.40464735e+00 3.97843309e-02 -3.99128675e-01
4.37974393e-01 1.54255956e-01 -2.12761819e-01 6.50013983e-01
3.19717407e-01 -1.00926733e+00 2.72711576e-03 1.03111899e+00
6.12036586e-01 -3.69803369e-01 8.70896757e-01 4.49761629e-01
-4.74734753e-01 1.32690281e-01 8.93666595e-02 1.12241149e-01
5.65793663e-02 2.44314998e-01 -1.15179932e+00 6.16113484e-01
7.15914011e-01 2.08833456e-01 -1.09610796e+00 1.30307448e+00
-5.14033698e-02 1.18526113e+00 -1.07880764e-01 -2.26808295e-01
6.61782697e-02 2.99877346e-01 2.84327358e-01 1.56594837e+00
1.81390252e-02 5.60155034e-01 -5.00665568e-02 6.40308380e-01
1.29352897e-01 3.88279229e-01 -1.91291466e-01 -3.00343454e-01
-6.83278516e-02 1.05317116e+00 -1.93893582e-01 -5.96039474e-01
-3.96036446e-01 8.63494098e-01 -2.42430881e-01 1.39056668e-01
-5.02615452e-01 -3.35115701e-01 1.12857625e-01 1.46562785e-01
-6.16483204e-02 4.16728139e-01 -1.77325040e-01 -6.81508541e-01
-6.59985125e-01 -1.56469357e+00 8.90836418e-01 -6.45236731e-01
-1.31662023e+00 4.40774590e-01 -9.52584818e-02 -7.58843839e-01
-4.52029705e-01 -3.61305065e-02 -3.97463441e-01 7.71425784e-01
-1.37108779e+00 -1.09828818e+00 -3.50380063e-01 6.27668083e-01
4.58564371e-01 2.15552866e-01 1.44828928e+00 6.88873708e-01
-4.59724694e-01 5.51196992e-01 -1.23397045e-01 -1.40325889e-01
8.88437390e-01 -1.43242085e+00 7.09372759e-02 1.97968528e-01
-1.55457228e-01 6.34844005e-01 8.32272351e-01 -7.84152925e-01
-1.18164277e+00 -1.29425442e+00 1.69121361e+00 -7.27814555e-01
1.91442490e-01 3.17486912e-01 -1.09536552e+00 1.16848148e-01
7.94816315e-01 -6.80966496e-01 1.27055597e+00 -1.68310225e-01
3.09826136e-02 -8.21081921e-02 -1.72183144e+00 4.23936248e-01
5.25351644e-01 -6.09688997e-01 -1.15006828e+00 6.10204697e-01
1.00870967e+00 2.00426243e-02 -1.10197961e+00 3.94822448e-01
2.99721777e-01 -4.07572687e-01 1.31232500e+00 -1.12284291e+00
4.64428872e-01 -3.06055248e-01 3.37197393e-01 -1.07511914e+00
-5.58681846e-01 -4.81076181e-01 -1.66835278e-01 9.13239121e-01
5.61977565e-01 -2.54561424e-01 5.78203678e-01 9.77442503e-01
-5.54925054e-02 -3.33532542e-01 -8.52274418e-01 -1.80801153e-01
5.93910739e-02 -8.82039890e-02 1.10479221e-01 7.74337530e-01
-5.04609868e-02 1.00588393e+00 2.98599564e-02 1.89606771e-01
1.92169487e-01 -2.95366257e-01 3.70613307e-01 -1.07354689e+00
-3.14355314e-01 -2.80176491e-01 3.33974371e-03 -3.50548863e-01
-6.11434400e-01 -8.45749259e-01 2.32521757e-01 -2.26885867e+00
5.32679439e-01 1.62831575e-01 -2.72347718e-01 6.98158681e-01
-5.79422653e-01 1.76584348e-01 -1.05153613e-01 -2.49752805e-01
-8.97761226e-01 3.63428533e-01 1.18846846e+00 -2.49065742e-01
-3.34339052e-01 1.90449819e-01 -8.58973503e-01 2.61091858e-01
9.75849748e-01 -7.26761520e-01 -3.34701031e-01 -2.08618060e-01
3.70198816e-01 3.63919407e-01 3.61874908e-01 -9.90013123e-01
3.62725228e-01 4.40744013e-02 2.17503533e-01 -5.09456694e-01
-9.73427892e-02 -7.55248070e-01 3.28739733e-01 1.28408504e+00
-8.53405237e-01 5.94182424e-02 3.98360431e-01 4.20717210e-01
1.05282702e-01 -1.22020066e-01 5.54921508e-01 -3.32044922e-02
-5.36364727e-02 -2.55214781e-01 -7.57248163e-01 3.63577276e-01
5.06822288e-01 2.98479497e-01 -7.06336617e-01 -5.46793580e-01
-7.98599482e-01 5.11996984e-01 -1.92534685e-01 2.20030636e-01
4.40331995e-01 -8.94819736e-01 -9.81760502e-01 -1.91018596e-01
2.03512326e-01 -3.45299602e-01 3.34060401e-01 1.18022728e+00
-7.26717472e-01 8.40097964e-01 2.45031789e-02 -3.29125047e-01
-1.34502518e+00 7.30156958e-01 3.74784082e-01 -8.81904781e-01
-3.04428250e-01 3.55449110e-01 -3.40965718e-01 -5.95217705e-01
4.49231178e-01 -5.69269717e-01 -5.30917466e-01 -2.00361595e-01
9.60443437e-01 6.74053133e-01 1.08649328e-01 -2.87079543e-01
-2.88148493e-01 3.49133052e-02 4.58717607e-02 -3.59596573e-02
1.23505759e+00 -1.64869174e-01 2.54602032e-03 7.80285848e-03
1.05783951e+00 -1.88212439e-01 -1.70989349e-01 -8.94209519e-02
5.53191721e-01 4.33040291e-01 1.25939265e-01 -1.83572710e+00
-3.05590868e-01 7.49519885e-01 1.03141594e+00 4.39507455e-01
1.24075508e+00 7.73606226e-02 1.20711803e+00 7.20858574e-01
-4.80961263e-01 -7.67948389e-01 7.64701590e-02 1.08741388e-01
8.08645785e-01 -1.38537145e+00 -1.15703560e-01 7.66431242e-02
-7.17632830e-01 9.02072370e-01 2.77952015e-01 2.97917783e-01
4.36655879e-01 -3.99095654e-01 3.12097013e-01 -6.76499009e-01
-6.55413508e-01 -4.04191583e-01 9.99570191e-01 6.67407990e-01
5.19837558e-01 1.78971980e-02 -7.01719224e-01 7.47261405e-01
7.41881877e-02 6.06440246e-01 -2.33241946e-01 7.72192895e-01
-5.42387724e-01 -1.22092891e+00 -3.82194698e-01 5.31413078e-01
-8.01427543e-01 -4.36543643e-01 -5.53144813e-01 6.53400183e-01
3.18150282e-01 1.37451625e+00 -2.72201717e-01 9.13872495e-02
6.50187552e-01 3.47562760e-01 1.16935879e-01 -8.31566930e-01
-1.47946036e+00 -1.06191806e-01 7.55806148e-01 -7.55360246e-01
-7.34726191e-01 -5.99672854e-01 -1.20194435e+00 1.79946765e-01
1.38237536e-01 4.50542957e-01 4.09000069e-01 7.80059218e-01
7.99335063e-01 7.50108600e-01 7.07801133e-02 3.31350446e-01
-3.84197414e-01 -9.70523179e-01 3.93964559e-01 5.25409698e-01
4.58029985e-01 3.48772168e-01 1.87150419e-01 3.23183596e-01] | [8.785486221313477, 8.583466529846191] |
13009b10-4fec-4100-a1f4-d34a3e2dd8d8 | srpcn-structure-retrieval-based-point | 2202.02669 | null | https://arxiv.org/abs/2202.02669v3 | https://arxiv.org/pdf/2202.02669v3.pdf | SRPCN: Structure Retrieval based Point Completion Network | Given partial objects and some complete ones as references, point cloud completion aims to recover authentic shapes. However, existing methods pay little attention to general shapes, which leads to the poor authenticity of completion results. Besides, the missing patterns are diverse in reality, but existing methods can only handle fixed ones, which means a poor generalization ability. Considering that a partial point cloud is a subset of the corresponding complete one, we regard them as different samples of the same distribution and propose Structure Retrieval based Point Completion Network (SRPCN). It first uses k-means clustering to extract structure points and disperses them into distributions, and then KL Divergence is used as a metric to find the complete structure point cloud that best matches the input in a database. Finally, a PCN-like decoder network is adopted to generate the final results based on the retrieved structure point clouds. As structure plays an important role in describing the general shape of an object and the proposed structure retrieval method is robust to missing patterns, experiments show that our method can generate more authentic results and has a stronger generalization ability. | ['Cheng Jin', 'Yuan Wu', 'Ximing Yang', 'Kaiyi Zhang'] | 2022-02-06 | null | null | null | null | ['point-cloud-completion'] | ['computer-vision'] | [-1.79875925e-01 -3.94635588e-01 -5.41159585e-02 -1.64233923e-01
-6.60108387e-01 -6.99394166e-01 3.09989095e-01 -5.88560812e-02
-4.14220558e-04 3.88844967e-01 5.68503095e-03 1.33509010e-01
-3.72987449e-01 -9.73516047e-01 -6.87682569e-01 -9.13630545e-01
4.24573988e-01 8.39961767e-01 3.49679381e-01 -1.75784677e-01
3.73702675e-01 1.08277822e+00 -1.66491210e+00 9.78941768e-02
1.01975179e+00 7.71925747e-01 5.50323784e-01 -8.73539448e-02
-4.65793401e-01 5.78884147e-02 -7.18825579e-01 -3.32581788e-01
3.56931150e-01 7.00356215e-02 -3.92943561e-01 2.39361212e-01
2.24345431e-01 -2.56336749e-01 -4.35312182e-01 1.23370135e+00
4.39363152e-01 1.39375851e-01 9.50569808e-01 -1.30606949e+00
-8.63232672e-01 2.35352889e-01 -5.72928250e-01 -4.99561340e-01
1.97414160e-01 2.79891156e-02 8.22944999e-01 -1.14846528e+00
6.31755471e-01 1.39607239e+00 4.65895146e-01 5.01518846e-01
-1.06304765e+00 -7.96077192e-01 -2.22218707e-02 -9.08626150e-03
-1.91242003e+00 -1.47690415e-01 1.12628806e+00 -1.75603881e-01
6.84155896e-02 2.18524858e-01 5.51705956e-01 7.20081449e-01
-9.40246731e-02 9.99814630e-01 5.37377536e-01 2.23574892e-01
2.37624258e-01 5.08551635e-02 -2.36839190e-01 -2.15704506e-03
5.20024657e-01 -1.54055282e-01 -2.47793458e-02 -2.49080434e-01
6.36396945e-01 5.82927346e-01 -5.58849216e-01 -6.03973508e-01
-1.25901639e+00 5.30495763e-01 6.42248750e-01 2.87100971e-01
-4.23578173e-01 -2.14335769e-01 7.49854073e-02 1.41075114e-02
-4.93167201e-03 -5.55868540e-03 -1.35279730e-01 2.10623920e-01
-8.73738408e-01 5.35891771e-01 4.98605907e-01 1.29593527e+00
1.01260507e+00 -1.42159626e-01 -5.32397293e-02 8.60001564e-01
4.51635003e-01 9.77369368e-01 1.34102091e-01 -8.25880826e-01
6.78103507e-01 8.89986575e-01 1.83664083e-01 -1.54035878e+00
1.34621551e-02 -4.76248026e-01 -1.13434625e+00 8.53335708e-02
3.69458273e-02 2.96072453e-01 -6.19923174e-01 1.47365057e+00
4.30104077e-01 2.21701056e-01 1.90078527e-01 1.09191287e+00
9.52863991e-01 9.53422606e-01 -5.29692829e-01 -1.68645531e-01
9.13076818e-01 -3.45049948e-01 -2.68772006e-01 1.57774135e-01
9.83074084e-02 -9.33095694e-01 8.56473982e-01 6.33027494e-01
-7.37808526e-01 -6.63563311e-01 -1.05305934e+00 6.82231113e-02
7.46625438e-02 2.18340382e-01 2.22131506e-01 2.44966328e-01
-7.63122678e-01 5.31533003e-01 -4.95764852e-01 3.96921523e-02
5.00226200e-01 2.62051433e-01 -4.07289088e-01 -5.27135074e-01
-5.96625865e-01 2.28252828e-01 6.78582489e-01 1.61949456e-01
-6.61084354e-01 -6.81086063e-01 -5.81893504e-01 6.13266565e-02
3.37593824e-01 -5.89089334e-01 6.69743657e-01 -3.90942425e-01
-8.13216209e-01 4.07953382e-01 -1.78362116e-01 2.38394991e-01
4.28278595e-01 2.64829010e-01 -5.23823678e-01 1.37052998e-01
1.69754803e-01 7.41587758e-01 8.71603191e-01 -1.90275061e+00
-7.68869936e-01 -4.15099323e-01 -2.49818400e-01 3.06479990e-01
-1.27261236e-01 -5.29537797e-01 -8.89779866e-01 -4.94395167e-01
8.30929816e-01 -8.13883722e-01 5.04319482e-02 2.81530827e-01
-3.28949362e-01 -4.35916156e-01 1.00948703e+00 -4.81521130e-01
1.03116357e+00 -2.54706979e+00 9.61692035e-02 5.98052263e-01
2.47998655e-01 6.62760660e-02 -3.80904317e-01 5.40115118e-01
-4.30449471e-02 -1.03531756e-01 -4.21488225e-01 -3.64369273e-01
7.95761496e-02 4.43243116e-01 -5.34422398e-01 3.54262084e-01
-5.28563671e-02 4.33152080e-01 -8.29283416e-01 -4.67179805e-01
1.25781849e-01 4.99595433e-01 -3.73735845e-01 1.34327179e-02
-3.79424632e-01 4.66132998e-01 -6.15828276e-01 6.25089943e-01
1.51000524e+00 3.62803526e-02 -3.29939187e-01 -4.45022374e-01
7.37067237e-02 -5.54040372e-01 -1.53588414e+00 1.78203702e+00
-1.19762719e-01 7.76199549e-02 5.14778309e-02 -6.93243325e-01
1.65940380e+00 2.99401544e-02 6.72590673e-01 -3.41373771e-01
-1.19591765e-01 5.53045511e-01 -1.42244950e-01 -3.48344058e-01
6.45712733e-01 -1.12854317e-01 2.35941112e-01 2.50511289e-01
-3.99901927e-01 -4.74791616e-01 -4.56023514e-02 2.36059323e-01
6.85972631e-01 8.84631500e-02 -2.55164534e-01 1.45432903e-02
6.39938295e-01 -4.96961698e-02 7.93877602e-01 3.25990826e-01
2.39127085e-01 1.22492361e+00 2.54037440e-01 -3.77281010e-01
-1.29228449e+00 -1.16103160e+00 -1.63973883e-01 1.65728539e-01
8.43853295e-01 -1.44341186e-01 -3.33527476e-01 -4.95546967e-01
2.37985879e-01 7.36987114e-01 -2.59206951e-01 -2.37269178e-01
-4.23327237e-01 -3.47229421e-01 2.87588209e-01 2.38556594e-01
4.47563529e-01 -1.09397030e+00 8.57485160e-02 9.91492346e-02
-4.32458937e-01 -7.24681377e-01 -6.06304884e-01 -6.15298212e-01
-9.43551362e-01 -1.15810132e+00 -9.34603512e-01 -9.75477457e-01
8.88482273e-01 6.85643077e-01 9.07046378e-01 4.58287716e-01
1.33204907e-01 7.63059109e-02 -3.56703550e-01 -1.58370033e-01
-5.02742648e-01 2.48069107e-03 1.01507284e-01 3.09112012e-01
4.59844857e-01 -8.79802346e-01 -6.90212488e-01 4.56126034e-01
-1.35911882e+00 -2.43783861e-01 6.82415545e-01 6.07480645e-01
8.91030014e-01 5.14634073e-01 4.67736691e-01 -2.90183783e-01
6.13275886e-01 -4.63354290e-01 -6.38940275e-01 2.22639978e-01
-4.26909983e-01 -4.10691276e-02 7.73409009e-01 -3.99287432e-01
-7.90550649e-01 1.26721978e-01 -1.38081789e-01 -1.11682689e+00
-1.85667768e-01 3.78855437e-01 -7.07205355e-01 1.12664185e-01
4.39771533e-01 7.70978510e-01 3.33967149e-01 -7.00188518e-01
1.61806643e-01 6.19260430e-01 7.13948429e-01 -7.71112442e-01
1.24962020e+00 6.21223807e-01 2.11920753e-01 -7.99302697e-01
-3.01417291e-01 -5.09474158e-01 -6.38482869e-01 -8.31085630e-03
4.06950712e-01 -7.85210729e-01 -8.03055286e-01 3.98432672e-01
-1.41457617e+00 4.99476910e-01 -1.99345529e-01 4.59036022e-01
-3.01997602e-01 8.06605041e-01 -2.65573829e-01 -6.33929491e-01
-3.80663484e-01 -1.11162710e+00 1.26641762e+00 9.05287489e-02
3.90124053e-01 -5.67847848e-01 1.69100077e-03 1.95485968e-02
2.02751439e-02 2.50512451e-01 9.93232191e-01 -4.95218992e-01
-9.42361176e-01 -3.82927150e-01 -2.19156995e-01 3.45222205e-01
1.74972311e-01 2.97156930e-01 -5.19017279e-01 -5.35256982e-01
1.48168579e-01 1.89087123e-01 5.73634744e-01 2.70419627e-01
1.30758560e+00 -2.57471263e-01 -5.21438062e-01 6.67598486e-01
1.60705650e+00 1.76087335e-01 9.57426429e-01 1.10291066e-02
7.55477607e-01 6.46167696e-01 8.66026103e-01 5.99964678e-01
2.50057518e-01 4.23073858e-01 8.12708557e-01 9.87748653e-02
1.57313496e-01 -5.12266695e-01 4.11949754e-02 1.08467400e+00
7.12466016e-02 -1.93290949e-01 -7.00490534e-01 6.25828922e-01
-1.74580169e+00 -1.00675106e+00 -3.16131175e-01 2.43503332e+00
4.73855674e-01 -2.70561427e-02 -6.48662075e-02 1.55023471e-01
1.13598287e+00 -4.79712198e-03 -7.33285129e-01 4.42051679e-01
-3.27200264e-01 -1.65223673e-01 1.33032858e-01 1.37827722e-02
-5.33945084e-01 5.72036505e-01 5.06279659e+00 1.54640329e+00
-9.76633966e-01 -3.08611542e-01 2.06401154e-01 1.10121161e-01
-8.31637502e-01 1.61469981e-01 -5.40788352e-01 7.51083374e-01
-1.59318760e-01 -2.52096891e-01 3.99003893e-01 9.65539098e-01
1.45724729e-01 3.06129456e-01 -9.59295273e-01 1.47008967e+00
9.94893163e-03 -1.12196374e+00 5.89028478e-01 2.38665134e-01
7.14566290e-01 -5.13832979e-02 1.26926139e-01 7.93180428e-03
-3.04444805e-02 -8.43746662e-01 6.29020989e-01 9.24709678e-01
8.76294911e-01 -1.12575424e+00 8.03551257e-01 8.02562833e-01
-1.34870172e+00 1.00036919e-01 -1.23967624e+00 5.49313307e-01
9.11439061e-02 7.95634151e-01 -6.26564264e-01 1.00527847e+00
4.70757931e-01 7.84268320e-01 -5.04458129e-01 1.49480188e+00
8.12201127e-02 1.59157604e-01 -6.08334184e-01 -1.24797570e-02
-2.68622413e-02 -7.09489644e-01 8.29396546e-01 5.56434333e-01
9.39904869e-01 1.09412380e-01 3.66089553e-01 1.09482467e+00
-1.02261059e-01 3.14539582e-01 -6.52280629e-01 3.21827531e-01
8.75673532e-01 1.32559621e+00 -6.80494726e-01 -2.41733059e-01
-1.45444348e-01 6.37590766e-01 9.83839929e-02 2.78117210e-01
-4.97130573e-01 -4.71036851e-01 3.63128215e-01 4.68592495e-02
4.25594866e-01 -1.38272151e-01 -2.50665247e-01 -1.18498814e+00
3.03099900e-01 -8.78660858e-01 2.11809054e-01 -1.17210877e+00
-1.53014469e+00 5.09836555e-01 -5.15869111e-02 -2.08836579e+00
1.96733445e-01 -3.29425424e-01 -6.83654010e-01 8.46128285e-01
-1.02432477e+00 -1.16311586e+00 -6.67440951e-01 7.35263526e-01
2.97352701e-01 -3.15234512e-01 3.98215562e-01 4.43320423e-01
-6.06768765e-02 3.55962008e-01 4.93648350e-01 1.49261868e-02
5.85432410e-01 -9.52837765e-01 1.98425755e-01 6.71993017e-01
5.58657609e-02 7.61352241e-01 4.38718468e-01 -9.28085983e-01
-1.55392289e+00 -1.14189255e+00 2.75674462e-01 -2.82362461e-01
1.06021248e-01 -5.02803251e-02 -1.33572948e+00 1.78786770e-01
-3.11981946e-01 -1.78616911e-01 1.05369583e-01 -3.48711193e-01
-1.64802760e-01 -3.80159527e-01 -1.21142900e+00 6.94019556e-01
9.62904155e-01 -1.49741545e-01 -6.07089460e-01 2.11301118e-01
8.81782174e-01 -3.39309067e-01 -7.19800174e-01 5.85492909e-01
2.09314153e-01 -8.66329253e-01 8.77875507e-01 -1.00708053e-01
4.85306352e-01 -9.23290968e-01 -2.63657987e-01 -1.15994275e+00
-4.94019032e-01 -1.08457394e-01 2.03354925e-01 1.44090986e+00
9.71214771e-02 -6.17326438e-01 9.88673925e-01 1.95819244e-01
-2.82007217e-01 -5.88162422e-01 -7.44444430e-01 -9.39852774e-01
1.78068325e-01 -3.23904753e-01 1.46993232e+00 7.94031262e-01
-7.03561485e-01 1.47564843e-01 -2.30969861e-01 5.18330693e-01
7.89617300e-01 5.59250593e-01 1.14399540e+00 -1.65249860e+00
1.04820423e-01 -3.08123142e-01 -4.35837269e-01 -1.40537202e+00
1.20200574e-01 -1.07390463e+00 1.18248634e-01 -1.64163578e+00
1.96692541e-01 -1.03725636e+00 7.93076009e-02 1.40183106e-01
6.09290674e-02 1.08060032e-01 1.46049574e-01 9.06349540e-01
-3.15072685e-01 1.03518164e+00 1.56230330e+00 -3.43960971e-01
-2.74700999e-01 4.30807471e-01 -5.99753141e-01 5.09399295e-01
5.53726435e-01 -4.50275034e-01 -5.65496504e-01 -4.71615314e-01
2.12335676e-01 1.90594181e-01 2.25349292e-01 -1.14618397e+00
3.57052177e-01 -2.04108402e-01 3.90371144e-01 -1.47621989e+00
3.68002862e-01 -1.26068258e+00 6.45569682e-01 1.97178409e-01
1.45984426e-01 1.93375291e-03 -1.16441131e-01 6.85404003e-01
-2.62972146e-01 -6.09503508e-01 4.21259850e-01 -2.06053630e-02
-4.24681693e-01 1.02067006e+00 2.01374963e-01 -1.82784513e-01
7.14716196e-01 -5.60408413e-01 -8.24832395e-02 -4.12079304e-01
-5.15683353e-01 3.12470078e-01 9.96333599e-01 4.57002759e-01
1.05927598e+00 -1.77070558e+00 -9.89237487e-01 2.82799244e-01
3.23692739e-01 8.55967224e-01 3.82681936e-01 4.23785090e-01
-7.96806395e-01 -4.23084535e-02 1.20790362e-01 -1.05650020e+00
-1.03618574e+00 9.39551353e-01 1.16905749e-01 3.95423681e-01
-8.04632783e-01 2.17473581e-01 3.42588454e-01 -6.40888870e-01
2.02499218e-02 -1.45369187e-01 -2.42277771e-01 -1.74310625e-01
3.27942967e-01 2.60179102e-01 4.33668680e-03 -8.50670338e-01
-1.87700048e-01 9.70057487e-01 2.46108957e-02 3.56698148e-02
1.19891107e+00 4.31158692e-02 -4.67915952e-01 1.31375313e-01
1.32558894e+00 3.24367493e-01 -1.21851730e+00 -5.02574682e-01
-3.34883183e-01 -8.00060868e-01 -4.23949271e-01 -1.63498059e-01
-1.26247215e+00 8.07576060e-01 3.70113760e-01 -6.42379597e-02
1.04483056e+00 1.51443314e-02 9.00975227e-01 5.33180952e-01
7.47886300e-01 -8.07022035e-01 2.01622285e-02 3.00179899e-01
1.11022389e+00 -8.03998291e-01 1.36324450e-01 -5.89251637e-01
-4.19315606e-01 1.19941473e+00 5.67627132e-01 -1.61809489e-01
5.59119344e-01 -2.01231226e-01 -2.96103835e-01 -2.02699542e-01
-3.37023735e-01 5.46745583e-02 2.30996862e-01 8.13346207e-01
-4.66444552e-01 2.15956923e-02 -1.57887697e-01 4.64601159e-01
-3.46765012e-01 -1.42574504e-01 3.76363397e-01 5.25080919e-01
-5.79965532e-01 -9.54998374e-01 -7.85276234e-01 4.21853185e-01
-1.66156562e-03 2.87027717e-01 -6.41468614e-02 4.78815198e-01
1.72702789e-01 8.21543574e-01 1.06475800e-01 -5.68453670e-01
4.21347886e-01 -4.08592641e-01 1.31648526e-01 -4.58639562e-01
6.20389469e-02 1.72727898e-01 -6.68521821e-01 -8.78275335e-02
-1.88093483e-01 -5.23857832e-01 -1.32721734e+00 -5.52524030e-01
-3.90327603e-01 4.51468259e-01 4.59786296e-01 6.45882785e-01
4.23835754e-01 -1.26907090e-02 1.02644813e+00 -7.65850604e-01
-5.87140441e-01 -6.22172058e-01 -9.44452941e-01 7.90362716e-01
6.94229379e-02 -6.56433344e-01 -4.22582507e-01 -2.78954029e-01] | [8.384265899658203, -3.5744705200195312] |
3135d6a0-9892-4389-8cbc-aa3856acc51e | community-recovery-in-the-geometric-block | 2206.11303 | null | https://arxiv.org/abs/2206.11303v1 | https://arxiv.org/pdf/2206.11303v1.pdf | Community Recovery in the Geometric Block Model | To capture inherent geometric features of many community detection problems, we propose to use a new random graph model of communities that we call a \emph{Geometric Block Model}. The geometric block model builds on the \emph{random geometric graphs} (Gilbert, 1961), one of the basic models of random graphs for spatial networks, in the same way that the well-studied stochastic block model builds on the Erd\H{o}s-R\'{en}yi random graphs. It is also a natural extension of random community models inspired by the recent theoretical and practical advancements in community detection. To analyze the geometric block model, we first provide new connectivity results for \emph{random annulus graphs} which are generalizations of random geometric graphs. The connectivity properties of geometric graphs have been studied since their introduction, and analyzing them has been difficult due to correlated edge formation. We then use the connectivity results of random annulus graphs to provide necessary and sufficient conditions for efficient recovery of communities for the geometric block model. We show that a simple triangle-counting algorithm to detect communities in the geometric block model is near-optimal. For this we consider two regimes of graph density. In the regime where the average degree of the graph grows logarithmically with number of vertices, we show that our algorithm performs extremely well, both theoretically and practically. In contrast, the triangle-counting algorithm is far from being optimum for the stochastic block model in the logarithmic degree regime. We also look at the regime where the average degree of the graph grows linearly with the number of vertices $n$, and hence to store the graph one needs $\Theta(n^2)$ memory. We show that our algorithm needs to store only $O(n \log n)$ edges in this regime to recover the latent communities. | ['Barna Saha', 'Soumyabrata Pal', 'Arya Mazumdar', 'Sainyam Galhotra'] | 2022-06-22 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 2.34441876e-01 3.48106682e-01 9.31397751e-02 4.49668765e-01
-1.32437706e-01 -8.00567627e-01 3.09462756e-01 5.05087614e-01
-1.26858965e-01 5.49128771e-01 -3.86643708e-01 -6.06218457e-01
-5.09862542e-01 -1.36133969e+00 -5.34864724e-01 -8.18957567e-01
-8.81086171e-01 7.52242386e-01 6.52496099e-01 -2.26755187e-01
2.67609715e-01 4.40468848e-01 -1.14406216e+00 -3.24526668e-01
7.58178115e-01 3.75590265e-01 1.62720233e-01 1.11886632e+00
-7.19524845e-02 6.06028974e-01 -1.97057903e-01 -1.39359266e-01
3.96094650e-01 -7.05323100e-01 -7.15974391e-01 1.93189055e-01
-1.16794296e-01 2.41591066e-01 -6.19642794e-01 1.39158285e+00
2.08481967e-01 -3.97831678e-01 6.99771404e-01 -1.29655492e+00
-5.29046714e-01 1.07987237e+00 -1.16777134e+00 2.64612317e-01
4.21680629e-01 -3.77887785e-01 1.12478793e+00 -5.25424838e-01
8.07350874e-01 1.08586740e+00 8.62121820e-01 1.31989628e-01
-1.38473892e+00 -6.96525455e-01 -3.39104645e-02 -1.04050018e-01
-1.95102775e+00 -1.01791862e-02 5.93478262e-01 -4.97667491e-01
5.98967195e-01 2.69534022e-01 8.55571389e-01 2.91774601e-01
-3.79557647e-02 3.49011183e-01 1.07902968e+00 -8.18959236e-01
1.83921859e-01 -2.28754506e-01 2.66829848e-01 1.01649880e+00
1.15632153e+00 -1.74527735e-01 -7.07032531e-02 -3.65167499e-01
9.42732453e-01 -3.86032164e-02 -2.98894733e-01 -7.22004354e-01
-9.98522460e-01 1.02927649e+00 4.14698750e-01 7.35653460e-01
1.11972792e-02 5.15480995e-01 1.05184190e-01 5.35711050e-01
2.30899543e-01 -8.44202412e-04 1.16812281e-01 3.32232028e-01
-8.65061998e-01 1.02641389e-01 1.05112290e+00 1.28714359e+00
6.48105502e-01 -2.71021724e-01 5.51484764e-01 3.07744175e-01
3.47010911e-01 7.79023230e-01 -4.67833400e-01 -6.76154971e-01
5.19722402e-01 6.37938797e-01 -9.94878709e-02 -1.31647098e+00
-4.75104749e-01 -5.13194919e-01 -1.43553352e+00 -1.24615185e-01
9.34837162e-01 2.18153223e-01 -5.07404447e-01 1.69137394e+00
1.81186143e-02 -1.57473519e-01 -2.59160191e-01 2.25494966e-01
1.57715246e-01 3.60357493e-01 -5.04904330e-01 -6.25748873e-01
1.10052586e+00 -4.66441453e-01 -3.53668451e-01 7.53897652e-02
7.02113092e-01 -4.21676964e-01 5.16969264e-01 2.21326441e-01
-1.16241086e+00 -4.82164416e-03 -1.05544543e+00 5.88159442e-01
-1.51873127e-01 -5.62463939e-01 4.80230391e-01 9.63113427e-01
-1.53859437e+00 5.06634057e-01 -6.43675089e-01 -6.12641871e-01
8.42745453e-02 3.94867778e-01 -2.94076085e-01 -4.46984559e-01
-8.15341413e-01 2.63868690e-01 2.08620876e-01 1.65773891e-02
-6.25648737e-01 6.59731850e-02 -6.65344238e-01 1.56441540e-01
5.46334743e-01 -4.90470976e-01 4.37584370e-01 -7.56074309e-01
-3.79475117e-01 8.96567404e-01 -1.90035924e-01 -3.63444388e-01
4.16088820e-01 7.84736335e-01 -8.22878554e-02 5.72972775e-01
1.80790991e-01 2.26090122e-02 3.38787615e-01 -1.22315979e+00
-1.24290071e-01 -4.06523734e-01 7.54313022e-02 -3.62944752e-01
-1.57415479e-01 1.09539494e-01 -9.63172913e-02 -3.58445555e-01
6.04124546e-01 -1.13392949e+00 -5.15164137e-01 -1.77107677e-01
-4.58282381e-01 -1.14985853e-01 4.52855438e-01 -4.05392237e-02
1.40936983e+00 -1.91829491e+00 8.59618858e-02 1.06923246e+00
1.03383827e+00 -1.54626623e-01 -4.32006605e-02 9.74372387e-01
-3.04463148e-01 6.90068960e-01 -3.31771731e-01 1.08941019e-01
-2.28056863e-01 5.33323223e-03 5.05122207e-02 9.01583910e-01
-2.78102607e-01 4.41937923e-01 -1.02904963e+00 -4.61572796e-01
-3.56539547e-01 1.73077613e-01 -7.42904723e-01 -4.18744862e-01
2.28803962e-01 1.23016357e-01 -3.94226462e-01 3.12951535e-01
9.22437310e-01 -8.98726165e-01 7.08015144e-01 6.35394633e-01
-4.58668917e-02 -3.15022133e-02 -1.61137581e+00 9.41311717e-01
6.32692426e-02 5.35587013e-01 4.59829956e-01 -1.11539793e+00
7.29532778e-01 3.86418402e-01 5.91304004e-01 -1.19564496e-01
9.34997946e-02 4.44607675e-01 2.23796234e-01 -1.46642644e-02
7.06348866e-02 -1.33217812e-01 -1.52032599e-01 8.76190364e-01
-4.11332518e-01 2.20744610e-01 4.45004404e-01 7.79013813e-01
2.00322175e+00 -6.81207657e-01 3.98538977e-01 -6.76718056e-01
3.30931991e-01 9.85418539e-03 2.55073726e-01 1.14080250e+00
-9.31182280e-02 4.52776223e-01 1.00545096e+00 -1.48050666e-01
-1.17330897e+00 -1.22441411e+00 2.24828392e-01 4.87588555e-01
3.43186647e-01 -7.17075706e-01 -9.20770168e-01 -1.10194184e-01
-1.22554667e-01 -7.14360252e-02 -6.81325436e-01 1.01562999e-01
-5.96534669e-01 -9.24062073e-01 4.07748908e-01 1.88708067e-01
4.42256331e-01 -6.85415566e-01 -2.13038042e-01 1.52366132e-01
-2.27781996e-01 -9.13598061e-01 -3.56933594e-01 1.81294173e-01
-9.55809712e-01 -1.50084913e+00 -5.84638059e-01 -8.82173121e-01
9.09645617e-01 7.22138524e-01 1.17947304e+00 7.49064445e-01
-3.17515403e-01 3.33918035e-01 -4.60686952e-01 -2.19646003e-02
-4.92042869e-01 1.37085184e-01 5.26482612e-02 -3.49142492e-01
2.50900507e-01 -9.39279675e-01 -5.52158237e-01 4.54602510e-01
-8.90025973e-01 -3.91949005e-02 2.75209218e-01 5.13776660e-01
3.44372690e-01 5.84948599e-01 2.50815123e-01 -8.17196786e-01
4.38456893e-01 -6.80808902e-01 -8.15619469e-01 8.13910142e-02
-5.43306470e-01 6.12469353e-02 5.23735464e-01 -1.53500751e-01
-6.24300428e-02 -1.09192185e-01 2.03965813e-01 -1.11044221e-01
4.11456347e-01 7.17709720e-01 -6.78087100e-02 -3.32449198e-01
4.14725512e-01 3.48542072e-02 -6.00684956e-02 -2.25648627e-01
1.22880906e-01 5.50433278e-01 5.26031479e-02 -3.55509281e-01
1.08877897e+00 7.75322318e-01 4.74896789e-01 -1.08926713e+00
-2.02698588e-01 -5.66547096e-01 -6.83154345e-01 -1.03589073e-01
5.36023498e-01 -6.48822486e-01 -1.02626598e+00 3.07454258e-01
-1.06730342e+00 -3.11744928e-01 -2.71552876e-02 1.88174412e-01
-6.44200325e-01 7.07427800e-01 -8.67418289e-01 -1.34441149e+00
1.35046139e-01 -6.64180994e-01 6.04126930e-01 -2.84251779e-01
9.22024623e-02 -1.04661739e+00 2.15933695e-01 -2.51406789e-01
2.52482563e-01 5.00135481e-01 1.20825112e+00 -5.12562156e-01
-9.75454509e-01 -3.50790381e-01 -5.33385217e-01 -3.30884576e-01
-1.37340501e-01 -7.15288520e-02 -1.96957156e-01 -5.11084974e-01
-1.31425783e-02 4.38111842e-01 8.10567796e-01 5.01467943e-01
8.12785923e-01 -2.23397925e-01 -7.82435656e-01 2.86251545e-01
1.86791193e+00 -1.59495436e-02 8.34653497e-01 1.11190222e-01
6.02296293e-01 4.14064407e-01 -7.90365562e-02 2.63088733e-01
2.73187160e-01 3.51657629e-01 4.82954681e-01 -4.55560125e-02
1.49536967e-01 -1.82848409e-01 8.83372277e-02 1.15481305e+00
-4.78273541e-01 -4.42723840e-01 -1.10910559e+00 6.69443488e-01
-1.69016135e+00 -1.34435487e+00 -9.07282948e-01 2.42458868e+00
5.13270617e-01 3.45657229e-01 6.06525600e-01 5.10272145e-01
1.33384490e+00 -1.01957895e-01 1.12604655e-01 -1.43055156e-01
-3.50726634e-01 4.23175037e-01 1.00395346e+00 7.02159703e-01
-5.81708133e-01 5.70542693e-01 7.00369263e+00 8.31291258e-01
-3.27325881e-01 1.98293805e-01 4.21512455e-01 2.54910082e-01
-3.84352624e-01 3.49839747e-01 -6.26232028e-01 3.58400851e-01
8.79663527e-01 -4.59830374e-01 5.49763024e-01 5.58548987e-01
1.51352957e-01 -3.04559022e-01 -8.32517087e-01 8.59444380e-01
-2.43436843e-01 -1.11169457e+00 -2.17418566e-01 8.69435966e-01
7.62613177e-01 -2.45039295e-02 -3.78033251e-01 -2.92418987e-01
8.32972646e-01 -1.18311334e+00 4.19196784e-01 1.17158733e-01
1.05228591e+00 -6.84130549e-01 5.15058815e-01 6.25048280e-01
-1.82458627e+00 -9.39797610e-02 -3.57882470e-01 -2.65687257e-01
1.43052429e-01 8.15630496e-01 -4.68779862e-01 5.63128293e-01
5.14107227e-01 3.43355596e-01 -3.82918864e-01 1.17449367e+00
1.56646609e-01 5.98045409e-01 -7.45738447e-01 -5.48810661e-02
1.67976737e-01 -5.55163622e-01 7.90627420e-01 9.65805650e-01
5.24675608e-01 2.04920873e-01 5.80322295e-02 6.98662043e-01
1.19532458e-02 2.00104564e-02 -1.02590489e+00 -1.39932886e-01
4.81487930e-01 8.38235140e-01 -1.56090128e+00 -1.41676053e-01
-3.45440090e-01 5.45886159e-01 2.13927954e-01 1.37411267e-01
-6.72643304e-01 -6.25995994e-01 -3.67098413e-02 9.47438121e-01
6.18894398e-01 -7.85282433e-01 -8.82586464e-02 -1.00823677e+00
2.48828307e-02 -6.54294133e-01 3.51491004e-01 -3.41056645e-01
-1.22107077e+00 6.03537977e-01 1.89106882e-01 -9.87420678e-01
2.59283613e-02 -4.26961482e-01 -5.35348892e-01 6.46130264e-01
-9.46561813e-01 -7.56031156e-01 -3.64496820e-02 8.07486892e-01
-3.20188731e-01 2.85063267e-01 7.28354096e-01 1.59901932e-01
-1.82983384e-01 2.60700136e-01 4.58981574e-01 3.76291007e-01
9.80369095e-03 -1.36418164e+00 4.54670966e-01 9.88525212e-01
4.29542884e-02 7.95664608e-01 7.90379345e-01 -8.49036038e-01
-1.35763097e+00 -7.20870554e-01 1.18912697e+00 -7.78720081e-02
9.74558651e-01 -7.98576713e-01 -5.89549482e-01 6.31655693e-01
-1.89609025e-02 1.24035716e-01 6.28844738e-01 -6.59183785e-02
-3.84204805e-01 1.38078451e-01 -1.06518519e+00 5.36149919e-01
1.65831256e+00 -3.58823985e-01 1.42012788e-02 5.14172137e-01
4.15642440e-01 4.06478971e-01 -6.37421846e-01 6.02993369e-02
3.24485093e-01 -1.10731876e+00 7.98642993e-01 -5.01413122e-02
2.26068702e-02 -4.53032851e-01 -1.68539673e-01 -7.71929324e-01
-4.89458442e-01 -9.26752567e-01 1.30054697e-01 7.21376538e-01
3.12629163e-01 -8.68289530e-01 8.87119591e-01 -2.64639318e-01
5.76138437e-01 -4.04068321e-01 -1.09067905e+00 -1.00137544e+00
3.27866405e-01 -8.81833956e-02 2.48155519e-01 9.42421913e-01
2.73023218e-01 3.35224748e-01 -1.95758734e-02 8.18198323e-02
1.12533081e+00 -1.64310203e-03 6.45545363e-01 -1.68260121e+00
-6.10256076e-01 -6.23726189e-01 -5.63810706e-01 -1.15580130e+00
-3.38524073e-01 -8.96153927e-01 -2.21798584e-01 -1.57132375e+00
6.41923249e-01 -9.33439434e-01 1.57349572e-01 -1.44629017e-01
2.00820044e-01 4.43310976e-01 8.97932798e-02 3.98958206e-01
-7.70807624e-01 -1.58875987e-01 9.14394498e-01 -6.63096979e-02
-1.88456595e-01 1.52885839e-01 -7.06596553e-01 6.25311077e-01
4.44270909e-01 -6.89387321e-01 -2.41130739e-01 -2.77753938e-02
9.42897081e-01 4.09774274e-01 3.16424251e-01 -9.77909267e-01
3.69524807e-01 8.57621878e-02 -7.82787576e-02 -6.69071913e-01
-1.16647638e-01 -7.96282232e-01 6.08222604e-01 8.11950326e-01
1.20072909e-01 4.43541735e-01 -3.20867181e-01 1.06467545e+00
7.36798123e-02 -3.23027104e-01 7.91714847e-01 -4.40568507e-01
2.17392430e-01 2.36960903e-01 -6.85933948e-01 -7.17118010e-03
1.13888919e+00 -6.65750861e-01 -4.04288620e-01 -7.36865699e-01
-8.78137827e-01 1.29055843e-01 8.35604489e-01 -3.17614615e-01
2.65775472e-01 -1.11871171e+00 -7.69982696e-01 -4.07820418e-02
1.21756950e-02 -3.13540027e-02 -1.51016610e-02 1.16802061e+00
-1.01001716e+00 1.95909858e-01 1.33808091e-01 -8.18973243e-01
-1.37465394e+00 9.80096519e-01 3.01390797e-01 -7.24532485e-01
-3.69623601e-01 5.87876797e-01 2.30893210e-01 1.79768860e-01
-1.43654913e-01 7.39383623e-02 1.62648335e-01 -2.65612960e-01
5.63056886e-01 4.18618321e-01 -2.00297281e-01 -6.65297806e-01
-4.32123661e-01 6.98123813e-01 1.91423390e-02 -1.93023935e-01
1.34287047e+00 -3.38171482e-01 -7.05759704e-01 3.54779363e-01
1.05111122e+00 4.67223614e-01 -4.75373387e-01 -4.97625284e-02
2.57169276e-01 -3.47360969e-01 -4.63625550e-01 4.19313386e-02
-9.35912728e-01 6.67697191e-01 1.53242648e-01 1.21726322e+00
9.99225736e-01 3.18739980e-01 2.13442966e-01 1.50706291e-01
9.32967901e-01 -5.96654177e-01 -4.73366976e-02 5.72622359e-01
3.86673301e-01 -5.36689937e-01 -8.61948635e-03 -1.09959853e+00
2.50332236e-01 1.09830391e+00 5.07028261e-03 -5.70621371e-01
1.09053242e+00 3.92357826e-01 -8.21956933e-01 -5.53127587e-01
-4.10370797e-01 -4.35982406e-01 -4.99669820e-01 6.44165814e-01
1.42272547e-01 1.71162471e-01 -4.12430435e-01 2.47639105e-01
-8.75037163e-02 -2.40663126e-01 9.36991215e-01 8.79490674e-01
-8.28818738e-01 -1.33080125e+00 -4.91777211e-01 3.93175483e-01
-2.98367769e-01 -2.56732047e-01 -6.85800016e-01 9.57397699e-01
-4.28883359e-02 1.12726176e+00 1.48056075e-01 -3.36472064e-01
-4.92352247e-02 -2.80600339e-01 7.55997181e-01 -8.02462816e-01
-1.96000099e-01 2.04127640e-01 -1.46839574e-01 -1.01616448e-02
-5.44442654e-01 -6.17494762e-01 -9.44875777e-01 -1.10026896e+00
-7.71403670e-01 5.40475011e-01 2.20723256e-01 7.01554596e-01
1.83891617e-02 3.23249437e-02 6.51650131e-01 -4.12554353e-01
-1.31646201e-01 -8.41838002e-01 -1.33797503e+00 -1.87227339e-01
1.22942276e-01 -5.60720921e-01 -8.20247650e-01 -2.22875416e-01] | [6.868375301361084, 5.136017799377441] |
c9997058-333b-4765-b680-adc68ceafc26 | intra-speaker-topic-modeling-for-improved | null | null | https://aclanthology.org/N12-1041 | https://aclanthology.org/N12-1041.pdf | Intra-Speaker Topic Modeling for Improved Multi-Party Meeting Summarization with Integrated Random Walk | null | ['Yun-Nung Chen', 'Florian Metze'] | 2012-06-01 | null | null | null | naacl-2012-6 | ['meeting-summarization'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.448444843292236, 3.6738669872283936] |
0afee54e-f610-45c3-8be9-e1e5b3f82bbe | unsupervised-3d-keypoint-estimation-with | 2211.12829 | null | https://arxiv.org/abs/2211.12829v1 | https://arxiv.org/pdf/2211.12829v1.pdf | Unsupervised 3D Keypoint Estimation with Multi-View Geometry | Given enough annotated training data, 3D human pose estimation models can achieve high accuracy. However, annotations are not always available, especially for people performing unusual activities. In this paper, we propose an algorithm that learns to detect 3D keypoints on human bodies from multiple-views without any supervision other than the constraints multiple-view geometry provides. To ensure that the estimated 3D keypoints are meaningful, they are re-projected to each view to estimate the person's mask that the model itself has initially estimated. Our approach outperforms other state-of-the-art unsupervised 3D human pose estimation methods on the Human3.6M and MPI-INF-3DHP benchmark datasets. | ['Pascal Fua', 'Sina Honari'] | 2022-11-23 | null | null | null | null | ['3d-human-pose-estimation', 'unsupervised-3d-human-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-3.71501148e-01 3.75766546e-01 -1.11693271e-01 -3.85146230e-01
-6.04282975e-01 -3.87462795e-01 2.71333903e-01 -1.46108940e-01
-5.46479881e-01 5.55606127e-01 2.65085250e-01 5.13547659e-01
3.55937809e-01 -2.44154423e-01 -8.35841715e-01 -3.64320159e-01
-6.29803818e-03 1.29629886e+00 4.86784041e-01 -3.29673178e-02
-9.94786769e-02 5.81643581e-01 -1.29200625e+00 -1.85572773e-01
3.09294939e-01 5.28507471e-01 -4.55632992e-02 7.23271787e-01
6.42467678e-01 2.74175823e-01 -5.99269509e-01 -3.39176923e-01
5.12859464e-01 -2.46508494e-01 -5.45380235e-01 6.17001653e-01
7.34773338e-01 -6.07615829e-01 -2.73288965e-01 6.16100192e-01
5.37860096e-01 2.21303731e-01 7.45467365e-01 -1.24386537e+00
1.58217475e-01 -2.91435748e-01 -8.77770364e-01 -1.44527582e-02
1.06711602e+00 -1.89353153e-02 5.03110647e-01 -1.12391591e+00
9.96053278e-01 1.29720521e+00 7.76643157e-01 7.46787131e-01
-9.86055195e-01 -4.02577519e-01 1.14203580e-01 -1.16803430e-01
-1.48409736e+00 -2.84073919e-01 6.10238731e-01 -5.87914169e-01
7.28078246e-01 9.97699946e-02 1.21603799e+00 1.11975491e+00
2.54676461e-01 8.63982797e-01 9.26949561e-01 -3.61380547e-01
-8.36536475e-03 -1.27009992e-02 -8.93281177e-02 1.00855196e+00
3.45319480e-01 -5.26453517e-02 -8.48245740e-01 -3.00058424e-01
1.03196311e+00 1.49239987e-01 -2.79962663e-02 -1.10059679e+00
-1.34822607e+00 4.13617104e-01 2.90633231e-01 -2.43167654e-01
-4.97409523e-01 1.36373207e-01 1.29269600e-01 -3.32749337e-01
5.71413338e-01 8.45573172e-02 -5.10453165e-01 -1.35801092e-01
-8.49590600e-01 6.58345103e-01 5.71109295e-01 1.22813737e+00
5.17645180e-01 -5.55344522e-01 1.42080531e-01 3.50257516e-01
5.15139759e-01 7.15493321e-01 -6.32887259e-02 -9.52175260e-01
5.80800653e-01 7.70053744e-01 4.77866113e-01 -8.55057001e-01
-6.00049853e-01 -7.59663582e-02 -5.04896581e-01 1.62607417e-01
5.66507041e-01 -6.46528006e-02 -1.11390352e+00 1.15500677e+00
1.05612278e+00 -1.43356085e-01 -2.85094708e-01 1.28096819e+00
7.07621336e-01 3.66962284e-01 -3.29274863e-01 1.24266438e-01
1.49586749e+00 -1.02480900e+00 -4.72942710e-01 -6.40049458e-01
3.82633924e-01 -4.95394289e-01 6.59954607e-01 4.13744450e-01
-1.21110666e+00 -6.70218706e-01 -8.36241663e-01 -9.30992439e-02
1.14661224e-01 2.13817954e-01 4.34068352e-01 5.25656760e-01
-4.88264263e-01 2.41525009e-01 -1.19042993e+00 -5.49292028e-01
2.21161395e-01 3.81240308e-01 -9.19348121e-01 -1.51026370e-02
-7.68443227e-01 1.17397666e+00 2.67590195e-01 1.80300862e-01
-1.09160745e+00 -2.52946347e-01 -1.01821887e+00 -5.46919465e-01
6.76643193e-01 -1.01748466e+00 1.28451002e+00 -3.73635650e-01
-1.21626127e+00 1.26598608e+00 -3.03826839e-01 -1.96815163e-01
9.92246807e-01 -9.04932737e-01 1.38621166e-01 3.34489107e-01
2.03198433e-01 7.22819924e-01 7.09216475e-01 -1.49844158e+00
-4.89047647e-01 -1.00618708e+00 -6.75709173e-02 7.43656456e-01
2.50910103e-01 4.56464215e-04 -1.06759894e+00 -3.38792801e-01
5.36512136e-01 -1.42793739e+00 -4.04285401e-01 3.40181857e-01
-7.12117910e-01 -6.34203553e-02 8.45831931e-01 -8.47858846e-01
6.07411921e-01 -1.74602270e+00 7.07023501e-01 4.55016553e-01
2.81483978e-01 -3.45207448e-03 5.70328057e-01 1.65797696e-01
3.95541966e-01 -4.29761499e-01 -3.28388996e-02 -8.05626035e-01
-8.29678252e-02 2.19295099e-01 3.77760559e-01 1.01440024e+00
-3.89357388e-01 7.23399282e-01 -6.96177483e-01 -8.74485433e-01
3.35272640e-01 4.98073161e-01 -4.40022469e-01 4.92656857e-01
-5.26254438e-02 9.17726338e-01 -3.52615595e-01 7.15074658e-01
4.68299598e-01 -3.54768634e-01 9.73031595e-02 -9.21983495e-02
2.01445922e-01 -1.34488374e-01 -1.40553665e+00 1.90670621e+00
5.53620569e-02 1.24425121e-01 3.46902758e-03 -4.11249161e-01
4.67455536e-01 5.21866143e-01 6.50923193e-01 -7.81927258e-03
5.33070937e-02 -7.50264004e-02 -3.16298187e-01 -3.71457726e-01
2.39393219e-01 -1.02992937e-01 -2.14495853e-01 3.02966177e-01
9.32616740e-02 -1.13404796e-01 -3.69250290e-02 2.09576786e-01
6.56771362e-01 4.51146215e-01 6.04098380e-01 2.86107436e-02
3.07629615e-01 -2.40115792e-01 5.55540264e-01 3.84319216e-01
-2.99656868e-01 9.55256820e-01 1.49006829e-01 -6.79444253e-01
-1.14164746e+00 -1.19076598e+00 6.14268743e-02 8.63876879e-01
2.44497672e-01 -6.55443728e-01 -9.26847517e-01 -1.12574303e+00
2.18991056e-01 1.02619939e-01 -6.90257311e-01 2.30652124e-01
-6.10354602e-01 -1.23741120e-01 4.38011587e-01 7.52605915e-01
2.71049500e-01 -5.94554007e-01 -1.27135003e+00 -1.56137079e-01
-5.34551620e-01 -1.19759226e+00 -7.52167821e-01 -1.57833204e-01
-7.80850053e-01 -1.25351238e+00 -1.07637405e+00 -5.66522777e-01
1.19232953e+00 1.22872539e-01 1.06066227e+00 1.88234989e-02
-2.73599446e-01 7.23237872e-01 -1.77862480e-01 -5.02777457e-01
-2.58235708e-02 -1.91205472e-01 5.56356370e-01 -2.26153061e-01
4.14493829e-01 -1.96539626e-01 -6.57147884e-01 6.39788032e-01
-6.00912571e-02 7.26084858e-02 3.15134943e-01 5.08437455e-01
9.29460466e-01 -5.30328229e-02 -2.64665037e-01 -7.54153311e-01
-1.99013487e-01 1.20560110e-01 -4.45330262e-01 1.45618275e-01
-1.81125216e-02 -1.98364735e-01 -2.07913108e-02 -4.32459146e-01
-1.00676429e+00 6.93376899e-01 7.35639781e-02 -5.67703545e-01
-5.56892812e-01 4.95854393e-02 -3.44410390e-01 -1.39841959e-02
7.29593992e-01 -1.02768727e-01 -5.93683980e-02 -5.74950814e-01
1.32556260e-01 3.58392149e-01 5.30315995e-01 -4.56406355e-01
1.01360106e+00 7.78276563e-01 7.90934488e-02 -8.17193091e-01
-1.06945527e+00 -8.41441810e-01 -1.42978680e+00 -6.09410524e-01
1.31831658e+00 -1.14888716e+00 -6.65525317e-01 4.84910399e-01
-1.30511558e+00 -7.29011223e-02 -6.07150327e-03 6.82927966e-01
-7.29199231e-01 4.17019248e-01 -4.76360619e-01 -1.10737085e+00
-1.74783543e-01 -9.43056345e-01 1.63334715e+00 5.06707802e-02
-6.93276942e-01 -6.63006246e-01 1.42065406e-01 7.50916600e-01
-5.26481450e-01 5.23225367e-01 9.12314877e-02 -4.37817752e-01
-3.99537951e-01 -5.40338516e-01 5.19748688e-01 3.03816013e-02
-2.68884096e-02 -4.07876581e-01 -7.58879006e-01 -3.86735618e-01
-1.37316227e-01 -4.29733664e-01 2.60442436e-01 5.14388144e-01
8.09032440e-01 -1.81898534e-01 -7.65733421e-01 4.55563605e-01
6.67919755e-01 -2.04976335e-01 3.54880631e-01 2.63913870e-01
9.83173907e-01 7.03372478e-01 9.27728534e-01 6.04105175e-01
6.33757949e-01 9.57936406e-01 4.05650795e-01 -8.39062110e-02
3.05206597e-01 -5.78076065e-01 2.62041271e-01 3.92872632e-01
-6.22640848e-01 8.64449069e-02 -1.15404928e+00 2.53562778e-01
-1.76157224e+00 -7.58949161e-01 -1.28890663e-01 2.45690274e+00
4.88990515e-01 3.74153286e-01 6.03259385e-01 -1.39736766e-02
5.97658575e-01 -2.64458787e-02 -5.99887371e-01 5.01992762e-01
4.04818445e-01 -2.97449619e-01 4.64047432e-01 4.23085034e-01
-1.33703458e+00 7.53622770e-01 6.32777262e+00 -6.68825656e-02
-3.30564022e-01 1.67910293e-01 2.67030329e-01 -3.56254309e-01
3.23611528e-01 -7.76647478e-02 -1.15305114e+00 1.32900402e-01
2.88423896e-01 4.50870961e-01 -9.10983160e-02 1.01130164e+00
-1.91758703e-02 -4.52393442e-01 -1.41366982e+00 1.31340802e+00
4.58888710e-01 -5.19965410e-01 -3.43245804e-01 3.71844858e-01
6.68961108e-01 -2.53114104e-01 -2.67471164e-01 3.82588431e-03
-1.53587526e-02 -8.65669608e-01 8.87145698e-01 6.61749899e-01
4.75120932e-01 -7.71949708e-01 6.30387247e-01 9.70483005e-01
-1.17399323e+00 3.26800555e-01 -3.01202267e-01 -8.44302122e-03
5.27755857e-01 5.32872498e-01 -1.01178038e+00 4.27957088e-01
9.17316616e-01 5.86686432e-01 -6.46124244e-01 9.09633577e-01
-4.76835847e-01 1.91391647e-01 -6.18182778e-01 3.16340923e-01
-2.45761439e-01 -2.03969162e-02 6.99299276e-01 8.64756048e-01
2.82473147e-01 3.30236912e-01 6.45022154e-01 3.62622052e-01
1.89177454e-01 -7.02048913e-02 -6.14616036e-01 4.14500654e-01
1.91420913e-01 1.13325691e+00 -8.42897236e-01 -3.86670470e-01
-2.32001558e-01 1.38540804e+00 1.98875621e-01 9.10740346e-02
-9.07035053e-01 2.43096784e-01 3.97655129e-01 6.30016387e-01
1.26709551e-01 -5.55226386e-01 2.52649069e-01 -1.31444001e+00
3.98609132e-01 -8.68131995e-01 5.45170784e-01 -9.21807587e-01
-9.99949038e-01 2.22727492e-01 4.86404806e-01 -1.31136298e+00
-4.52302426e-01 -5.59736848e-01 -6.24211580e-02 5.39538980e-01
-6.07277453e-01 -1.42467785e+00 -3.67553651e-01 5.39399147e-01
4.52199340e-01 5.92473224e-02 7.05159724e-01 -7.97188282e-02
-1.54446110e-01 3.27728331e-01 -5.39752066e-01 3.42782319e-01
8.67491007e-01 -1.38104796e+00 6.09439194e-01 6.50229752e-01
3.09122980e-01 5.62219322e-01 8.38842869e-01 -1.03180611e+00
-1.43884361e+00 -6.30107582e-01 7.55869150e-01 -1.28201282e+00
-1.50399402e-01 -6.81529284e-01 -5.41491807e-01 1.18578672e+00
-2.53326625e-01 3.03043664e-01 5.81190228e-01 3.70962918e-01
-2.02951550e-01 2.14553148e-01 -1.13530326e+00 4.52977240e-01
1.34279394e+00 -2.43331313e-01 -9.70435917e-01 5.20661891e-01
1.61846966e-01 -1.24470031e+00 -9.63013887e-01 4.36457276e-01
8.43069732e-01 -9.34202373e-01 1.35999179e+00 -5.52590907e-01
2.12900788e-02 -3.85970742e-01 -1.20292557e-02 -1.21700704e+00
-1.03404425e-01 -4.23195094e-01 -4.80750561e-01 6.01096213e-01
-1.31600117e-03 -1.46985263e-01 1.31128693e+00 7.13439941e-01
2.65221566e-01 -5.60663342e-01 -1.14065242e+00 -6.69441521e-01
-4.57866907e-01 -1.72371402e-01 2.74390042e-01 5.90726316e-01
-6.98784366e-02 2.80899793e-01 -7.07395732e-01 5.30599177e-01
1.06541455e+00 -1.50269985e-01 1.48838735e+00 -1.40452409e+00
-3.67910832e-01 3.59739870e-01 -7.81839252e-01 -1.30891120e+00
1.17629372e-01 -3.48685622e-01 2.17610493e-01 -1.50697124e+00
4.23814923e-01 6.36499077e-02 2.51772732e-01 4.29842234e-01
-2.30274960e-01 2.26966858e-01 1.48900390e-01 1.84939414e-01
-1.00437534e+00 3.73060226e-01 1.27908361e+00 2.22786993e-01
-4.40636128e-02 3.21049333e-01 4.06946540e-02 1.22027063e+00
4.54979092e-01 -5.11125386e-01 -3.29850882e-01 -2.48642623e-01
1.08480558e-01 2.41203964e-01 8.23348701e-01 -1.29643750e+00
2.52346545e-01 -4.46489230e-02 1.09149110e+00 -1.19099367e+00
8.36395919e-01 -9.77008164e-01 3.94092262e-01 5.32217443e-01
-1.48001465e-03 1.32129565e-01 -1.44760951e-01 6.51628971e-01
2.36789465e-01 5.52641489e-02 5.73177755e-01 -7.36396849e-01
-3.78823191e-01 5.53750575e-01 -2.11756796e-01 1.91063032e-01
1.19150901e+00 -3.88733059e-01 3.38201523e-01 -4.88589615e-01
-9.98823106e-01 2.84374386e-01 9.37186599e-01 3.50817025e-01
8.19639504e-01 -1.33770788e+00 -6.13252223e-01 1.69516161e-01
2.99927503e-01 6.38435543e-01 1.77860871e-01 7.48238325e-01
-4.19657022e-01 4.12674338e-01 -9.02110711e-02 -1.02968955e+00
-1.73447430e+00 5.11226475e-01 5.04450023e-01 -1.39567494e-01
-9.63405430e-01 7.40670323e-01 2.16777056e-01 -8.28930914e-01
3.15107554e-01 -1.16612606e-01 7.17989728e-02 -1.84059501e-01
4.56636786e-01 5.67095935e-01 -1.25392333e-01 -1.19570422e+00
-6.42347455e-01 9.12997007e-01 1.33992195e-01 -3.46353412e-01
1.19067812e+00 -6.66196495e-02 1.69147417e-01 4.75544810e-01
9.33924913e-01 1.01818837e-01 -1.52186060e+00 -1.74939826e-01
-2.34901994e-01 -7.17777252e-01 -5.57063282e-01 -5.30084074e-01
-7.72323072e-01 7.35948503e-01 6.04579866e-01 -4.42951381e-01
5.67962766e-01 4.50231582e-01 6.33402228e-01 5.78637600e-01
9.15982068e-01 -1.30901682e+00 3.52075875e-01 3.00086558e-01
9.40335751e-01 -1.28921044e+00 4.85180408e-01 -6.31292999e-01
-8.16179335e-01 8.09316099e-01 1.02136755e+00 1.13955291e-03
4.84526724e-01 1.49368271e-01 1.34807657e-02 -4.67997193e-01
-2.57024676e-01 -3.50132175e-02 7.80511498e-01 7.18734503e-01
2.57103980e-01 7.16582835e-02 1.89921930e-01 3.40379477e-01
-3.07525456e-01 -1.19677387e-01 1.70475289e-01 1.18350565e+00
-3.81163627e-01 -8.15586925e-01 -9.28543866e-01 2.55520016e-01
-3.71380389e-01 5.35220265e-01 -5.51765263e-01 1.00436568e+00
1.52007461e-01 4.70047444e-01 -1.04225881e-01 -1.86216220e-01
8.49194050e-01 2.36733317e-01 9.21331823e-01 -8.65456700e-01
-1.98885307e-01 3.70416880e-01 8.96624699e-02 -7.48148739e-01
-5.64120948e-01 -1.04625416e+00 -1.44556916e+00 -6.44302601e-03
-3.74368697e-01 -3.28669101e-02 3.65315467e-01 1.00238466e+00
6.83558807e-02 -5.26555590e-02 5.96411340e-03 -1.58346236e+00
-4.17363226e-01 -8.23642850e-01 -5.69181204e-01 6.40525460e-01
2.34265566e-01 -1.08440125e+00 -1.88157916e-01 2.64857531e-01] | [7.006881237030029, -1.000365972518921] |
bbcc402e-47a6-4253-a962-d0097138dffa | the-theory-behind-overfitting-cross | 1905.12787 | null | https://arxiv.org/abs/1905.12787v2 | https://arxiv.org/pdf/1905.12787v2.pdf | The Theory Behind Overfitting, Cross Validation, Regularization, Bagging, and Boosting: Tutorial | In this tutorial paper, we first define mean squared error, variance, covariance, and bias of both random variables and classification/predictor models. Then, we formulate the true and generalization errors of the model for both training and validation/test instances where we make use of the Stein's Unbiased Risk Estimator (SURE). We define overfitting, underfitting, and generalization using the obtained true and generalization errors. We introduce cross validation and two well-known examples which are $K$-fold and leave-one-out cross validations. We briefly introduce generalized cross validation and then move on to regularization where we use the SURE again. We work on both $\ell_2$ and $\ell_1$ norm regularizations. Then, we show that bootstrap aggregating (bagging) reduces the variance of estimation. Boosting, specifically AdaBoost, is introduced and it is explained as both an additive model and a maximum margin model, i.e., Support Vector Machine (SVM). The upper bound on the generalization error of boosting is also provided to show why boosting prevents from overfitting. As examples of regularization, the theory of ridge and lasso regressions, weight decay, noise injection to input/weights, and early stopping are explained. Random forest, dropout, histogram of oriented gradients, and single shot multi-box detector are explained as examples of bagging in machine learning and computer vision. Finally, boosting tree and SVM models are mentioned as examples of boosting. | ['Mark Crowley', 'Benyamin Ghojogh'] | 2019-05-28 | null | null | null | null | ['l2-regularization'] | ['methodology'] | [ 1.66065991e-02 -8.45157914e-03 -4.22408581e-01 -5.93815684e-01
-9.63835955e-01 -2.27091089e-01 2.24608481e-01 1.00432359e-01
-4.85743314e-01 1.24387991e+00 -3.00007701e-01 -4.69689459e-01
-2.26845175e-01 -8.83434892e-01 -6.30049109e-01 -9.53339458e-01
-1.12467363e-01 -3.21489088e-02 4.27345671e-02 -1.65621087e-01
2.22438872e-01 2.87783056e-01 -1.38723445e+00 -1.54716536e-01
1.14310813e+00 1.22587991e+00 -5.21298051e-01 5.35462379e-01
-1.41064540e-01 6.07597291e-01 -5.07009387e-01 -5.32505810e-01
2.40015507e-01 -3.25940907e-01 -5.28150797e-01 9.57002118e-02
3.70309472e-01 -1.27554417e-03 2.55410105e-01 9.29862380e-01
4.33201432e-01 2.12474599e-01 1.04620564e+00 -1.33252263e+00
-6.05385780e-01 2.01644003e-01 -8.25933039e-01 2.23815124e-02
-6.37992620e-02 -8.41047615e-02 4.70124722e-01 -1.01180148e+00
1.44665360e-01 8.41348350e-01 1.09522021e+00 4.55043346e-01
-1.38032508e+00 -8.68565679e-01 7.47642741e-02 -2.24121779e-01
-1.13443673e+00 3.17339301e-02 5.66887438e-01 -8.25751781e-01
6.78446233e-01 4.40725118e-01 5.18876016e-01 8.19961607e-01
2.85606474e-01 6.96645319e-01 1.42591763e+00 -8.05121124e-01
5.06879866e-01 6.63336635e-01 8.85802209e-01 8.49863350e-01
3.65805626e-01 4.55017388e-01 -2.22571164e-01 -4.92537796e-01
6.08203590e-01 1.84085965e-01 -2.19847590e-01 -5.68529963e-01
-3.72694999e-01 1.32488787e+00 4.02594239e-01 2.83783320e-02
-1.50739461e-01 9.75625124e-03 5.99723697e-01 7.95793980e-02
7.85516977e-01 -7.71090239e-02 -3.68957669e-01 3.70339841e-01
-9.67888415e-01 5.60586303e-02 7.01294839e-01 6.85577810e-01
6.18734539e-01 3.61803234e-01 1.18850328e-01 1.29459417e+00
6.97810799e-02 5.64773619e-01 6.13152742e-01 -6.89558208e-01
1.36869267e-01 4.77186680e-01 1.15135267e-01 -6.74149394e-01
-5.48951745e-01 -7.75277793e-01 -1.15535593e+00 7.94718802e-01
4.16943908e-01 -1.47380918e-01 -9.14503753e-01 1.59721625e+00
3.15914482e-01 -5.09400554e-02 -2.62595773e-01 5.84413230e-01
4.43621665e-01 1.91454947e-01 2.40059286e-01 -6.83300257e-01
1.14962161e+00 -1.05736721e+00 -6.86398447e-01 -3.95989150e-01
7.27111697e-01 -4.01594132e-01 8.86993527e-01 5.87121308e-01
-9.90491211e-01 -5.60968041e-01 -1.09581840e+00 2.17816368e-01
-6.59499705e-01 2.12432276e-02 7.12344766e-01 1.12150276e+00
-4.39643532e-01 7.85147071e-01 -9.46328878e-01 -1.14935346e-01
5.67305744e-01 2.25692287e-01 -3.66663992e-01 2.55242400e-02
-9.19208109e-01 1.10435426e+00 -2.15583369e-01 8.74788687e-02
-3.70385081e-01 -6.24731958e-01 -9.45901632e-01 -2.66699284e-01
-1.23722874e-01 -5.86461782e-01 7.94693768e-01 -1.16663539e+00
-1.20934951e+00 1.04167688e+00 -2.04146639e-01 -5.83175600e-01
4.70634341e-01 -5.01378238e-01 -2.80653447e-01 -5.18081725e-01
1.26866370e-01 7.77979270e-02 8.69780660e-01 -1.02771378e+00
-5.07882774e-01 -8.16726148e-01 -5.69123209e-01 -2.33407766e-01
7.59373754e-02 9.12893787e-02 2.98682362e-01 -6.35553002e-01
4.29206431e-01 -5.19848704e-01 -4.89164740e-01 -2.36517414e-01
-3.22570950e-01 4.11037914e-03 3.64508390e-01 -8.29676807e-01
1.53795648e+00 -2.22094512e+00 -3.66127640e-01 4.25027817e-01
-7.49003515e-02 6.29115326e-04 3.43870103e-01 1.65794671e-01
-4.31079239e-01 1.06139459e-01 -5.62426209e-01 2.23741829e-02
-3.31772625e-01 1.63336799e-01 -2.77321100e-01 7.55238473e-01
4.32596095e-02 5.06127477e-01 -5.65960526e-01 -2.37121940e-01
1.71870485e-01 3.28093350e-01 -2.30636671e-01 -1.65843293e-01
5.64736724e-01 1.46907330e-01 -3.14608403e-02 5.42023420e-01
9.85052466e-01 -1.37995645e-01 -1.91156432e-01 2.72644963e-02
1.15365079e-02 5.33854477e-02 -1.23372090e+00 8.98494899e-01
-5.80757201e-01 3.08080286e-01 8.94514471e-02 -1.32810271e+00
1.18263876e+00 1.36784557e-02 -9.54247192e-02 -2.00603783e-01
1.14441425e-01 6.05669916e-01 -4.01300967e-01 -1.88609764e-01
-1.79212824e-01 -6.78814352e-01 9.77102369e-02 1.31285880e-02
-1.17403686e-01 -1.00301385e-01 -8.57057273e-02 -1.67467386e-01
8.71631920e-01 1.67939246e-01 9.08221185e-01 -2.59538412e-01
6.09679461e-01 7.26314411e-02 5.94424129e-01 9.72031474e-01
-3.07876945e-01 4.58649635e-01 5.26736319e-01 -5.32344162e-01
-6.10131264e-01 -1.15056050e+00 -6.70268953e-01 1.24467778e+00
-2.08567604e-01 3.86190452e-02 -5.44155300e-01 -1.01697195e+00
5.21687567e-01 1.04654670e+00 -8.93541515e-01 -1.22719318e-01
-4.28647310e-01 -1.30045629e+00 3.38997334e-01 8.20605278e-01
3.55379730e-01 -7.00954318e-01 -2.90809005e-01 1.10594019e-01
2.41618723e-01 -3.46050233e-01 -9.69926044e-02 8.32020581e-01
-1.23756790e+00 -1.15938461e+00 -7.10687995e-01 -7.08907664e-01
5.92635870e-01 2.66130734e-02 1.04096162e+00 -1.21383108e-01
-3.89260620e-01 1.78266183e-01 -7.15752766e-02 -4.54061508e-01
-2.38039851e-01 -4.78359342e-01 3.37001920e-01 -2.79377580e-01
5.79015613e-01 -6.64874792e-01 -6.19174004e-01 4.64928448e-01
-3.74608397e-01 -4.18484420e-01 3.35193306e-01 1.38631952e+00
6.65827930e-01 -3.32397610e-01 6.63883269e-01 -1.06043005e+00
4.45556939e-01 -4.38148171e-01 -7.71455705e-01 2.56125718e-01
-1.08057058e+00 -1.76869944e-01 3.81877303e-01 -3.49229336e-01
-6.79768562e-01 -5.39327487e-02 -9.24603045e-02 -3.94029021e-01
7.51877055e-02 2.92246252e-01 2.18711525e-01 -7.79042616e-02
1.20484531e+00 6.47751316e-02 5.75968549e-02 -5.28085172e-01
2.28335366e-01 8.63360941e-01 9.75794047e-02 -4.07325953e-01
4.85634387e-01 4.97236997e-01 1.11537613e-01 -6.68188453e-01
-1.21000957e+00 -5.66015363e-01 -4.90890145e-01 7.02420324e-02
7.84236133e-01 -7.79415786e-01 -6.34908676e-01 2.81516105e-01
-7.42779195e-01 -2.11276039e-01 -5.94609082e-01 5.90526819e-01
-7.10782886e-01 8.38461816e-02 -4.36997980e-01 -1.51572120e+00
-5.21033049e-01 -1.01657498e+00 7.37102330e-01 9.44998711e-02
-2.37981960e-01 -1.11707258e+00 5.84996752e-02 3.44287187e-01
2.80691206e-01 4.42879349e-01 7.26242423e-01 -1.11399114e+00
1.31357014e-01 -6.45892560e-01 -1.44661635e-01 1.08884215e+00
-3.70818339e-02 -2.13151902e-01 -1.04099357e+00 -2.91969210e-01
3.14738125e-01 -2.65865147e-01 9.78690088e-01 8.18166554e-01
1.15414739e+00 -2.72668421e-01 -3.88533086e-01 6.27339303e-01
1.50362074e+00 9.41906050e-02 4.63583201e-01 3.48326713e-01
1.41839772e-01 7.90141642e-01 5.49135149e-01 2.01837882e-01
-2.25107059e-01 4.50669289e-01 1.00857638e-01 -1.56602070e-01
3.48376542e-01 -6.71630427e-02 3.33022803e-01 7.00984240e-01
-1.02541251e-02 5.76515555e-01 -8.51884305e-01 2.13978767e-01
-1.69820261e+00 -1.01051211e+00 -6.01687789e-01 2.65936089e+00
6.82740986e-01 3.04721266e-01 2.64165044e-01 3.28802824e-01
8.56293619e-01 -2.76452303e-01 -3.65851969e-01 -6.22318089e-01
-1.47261679e-01 5.28826654e-01 9.45934415e-01 8.38940084e-01
-1.43162441e+00 3.22182119e-01 6.76186085e+00 1.11143374e+00
-7.51534879e-01 3.98058712e-01 1.00256467e+00 8.67432952e-02
2.33087733e-01 1.48717612e-01 -9.61263359e-01 4.46240693e-01
7.67558694e-01 2.38659874e-01 2.94763654e-01 1.43698955e+00
3.70393470e-02 -5.21891654e-01 -7.46131599e-01 8.76776099e-01
-9.47903618e-02 -9.30368841e-01 -5.71266413e-01 -3.10270432e-02
5.63687146e-01 -1.59834757e-01 -1.10005781e-01 7.95695245e-01
4.97270495e-01 -1.04191005e+00 3.97494048e-01 2.92100281e-01
7.66983032e-01 -5.17865777e-01 1.02356958e+00 5.53992569e-01
-6.13683760e-01 -1.70490041e-01 -2.98329920e-01 -2.15522796e-01
-1.64016873e-01 1.15431285e+00 -4.35078472e-01 3.20555449e-01
7.00253367e-01 2.32701197e-01 -3.45247477e-01 9.80643153e-01
-2.17902407e-01 6.88132048e-01 -5.54490566e-01 -1.14520445e-01
-8.77252147e-02 -5.23891926e-01 3.05538148e-01 1.26823568e+00
1.27169997e-01 4.00260650e-02 8.75692889e-02 6.16493881e-01
4.71048951e-01 3.42191786e-01 -4.03262585e-01 6.96512759e-01
2.15002269e-01 9.76395428e-01 -5.36093831e-01 -3.46249998e-01
-5.05364478e-01 6.63749933e-01 2.67062306e-01 4.19315577e-01
-8.31094503e-01 -5.57015598e-01 4.24010098e-01 2.04621583e-01
1.83825985e-01 2.77181596e-01 -8.83910298e-01 -9.98466730e-01
4.94787237e-03 -6.16983473e-01 6.13822639e-01 -6.23922765e-01
-1.69563293e+00 2.52145261e-01 -3.60924043e-02 -1.12371302e+00
2.32539326e-02 -8.38292718e-01 -5.84925234e-01 1.26558959e+00
-1.12979269e+00 -7.32913554e-01 -2.13569507e-01 2.36872673e-01
4.03199345e-01 -8.00836757e-02 8.31339717e-01 -3.67598608e-02
-7.32268333e-01 6.42363667e-01 6.06002629e-01 2.35289127e-01
6.66919768e-01 -1.19437456e+00 -7.50507340e-02 3.71324450e-01
-3.62937987e-01 7.13656008e-01 7.21128702e-01 -6.72408640e-01
-5.71541846e-01 -9.05693173e-01 6.77945375e-01 -2.07420200e-01
6.05151832e-01 -1.90818205e-01 -1.01991749e+00 7.05238342e-01
-3.64249796e-01 5.28412521e-01 8.86392772e-01 5.80284655e-01
-4.71511483e-01 -3.31216127e-01 -1.53438008e+00 2.05087781e-01
6.53725743e-01 -1.65617242e-01 -4.88523453e-01 5.44270217e-01
1.02864593e-01 -2.32669622e-01 -7.68673301e-01 6.72595561e-01
8.76112103e-01 -1.37002933e+00 9.97351766e-01 -9.62660491e-01
1.99403763e-01 1.17499113e-01 -3.96190971e-01 -1.09049678e+00
-3.25240374e-01 -3.34214091e-01 1.87112525e-01 1.14167976e+00
4.99716431e-01 -1.00428963e+00 8.18312109e-01 5.00687182e-01
2.84745991e-01 -1.16054654e+00 -1.19067144e+00 -1.13488865e+00
5.46681762e-01 -6.39546633e-01 -4.14367840e-02 9.11667049e-01
2.83654511e-01 2.38846794e-01 -2.54173875e-01 -2.09865838e-01
8.54689300e-01 -8.47495794e-02 6.46385014e-01 -1.03659678e+00
-4.58043545e-01 -3.69020015e-01 -5.59456766e-01 -6.74540162e-01
-1.00426078e-01 -8.47249269e-01 -1.17697589e-01 -1.03689539e+00
1.27339467e-01 -5.11946619e-01 -3.40415686e-01 2.92193383e-01
-4.51630741e-01 4.92772684e-02 -2.58383423e-01 1.48357794e-01
1.57944441e-01 2.67379373e-01 6.30608141e-01 8.42212886e-02
-4.34175253e-01 5.60049176e-01 -4.03340131e-01 1.10844970e+00
9.25810039e-01 -8.04105520e-01 1.07774831e-01 4.17080492e-01
7.61536285e-02 3.84867817e-01 5.25569141e-01 -8.20666671e-01
-2.13189915e-01 -2.63501436e-01 6.93736374e-01 -2.45934248e-01
2.59935290e-01 -7.25474894e-01 -9.37197134e-02 6.15436614e-01
-4.25667793e-01 -1.23763636e-01 3.87486666e-02 6.33701742e-01
-8.41137543e-02 -5.98941207e-01 1.31715751e+00 -1.08861038e-02
9.50126946e-02 -2.04706281e-01 -2.01278597e-01 -6.56039547e-03
1.13999605e+00 -2.18936831e-01 -1.41457692e-01 -4.41801459e-01
-1.20977414e+00 1.50168791e-01 1.47123676e-04 -2.57854223e-01
4.85221088e-01 -1.27883244e+00 -6.87019527e-01 4.17109132e-01
-7.84431398e-02 -4.37824935e-01 1.60124078e-01 1.30995989e+00
-1.83097199e-01 1.22691117e-01 2.01328993e-01 -5.13459444e-01
-1.27815819e+00 7.26544678e-01 5.27191997e-01 -3.96156847e-01
-1.33002430e-01 1.11023045e+00 1.90792143e-01 -5.09506464e-01
2.71126330e-01 -3.26397866e-02 1.49865702e-01 -3.24449204e-02
5.14124453e-01 8.60860825e-01 1.10175133e-01 -1.08261921e-01
-5.39988756e-01 6.47834837e-01 4.76927347e-02 -4.58081393e-03
1.00945210e+00 1.16869450e-01 -2.76399583e-01 7.83209622e-01
1.24691939e+00 2.48081803e-01 -7.86810815e-01 5.91007769e-02
2.33079959e-02 -3.41423661e-01 -1.05582058e-01 -9.37772095e-01
-5.55852056e-01 1.12008071e+00 8.92209888e-01 4.50123638e-01
1.10654318e+00 -2.61085749e-01 2.36723840e-01 5.61806699e-03
2.76947498e-01 -1.04622376e+00 -6.88248694e-01 2.99979746e-01
9.00935709e-01 -1.46978080e+00 3.52853388e-01 -6.02471232e-01
-5.62276959e-01 1.16826618e+00 4.99667943e-01 -5.60817301e-01
1.18036163e+00 3.27348948e-01 8.88416991e-02 3.15176368e-01
-4.65015709e-01 -7.72706568e-02 1.92694962e-01 6.38730526e-01
5.77185214e-01 1.09720893e-01 -7.07451105e-01 1.02481878e+00
-1.84050903e-01 2.57453501e-01 -6.40910417e-02 7.85769224e-01
-6.51333749e-01 -7.96469212e-01 -7.46977031e-01 8.06435943e-01
-5.84687173e-01 -8.17255303e-02 1.22880168e-01 1.02747953e+00
1.23631179e-01 8.47417533e-01 -1.00484602e-01 -2.59438276e-01
4.29243475e-01 6.88227236e-01 2.99588501e-01 -3.18619221e-01
-7.22966194e-01 -1.24595249e-02 7.65056834e-02 -2.73312628e-01
-1.12284221e-01 -4.26710159e-01 -8.70031536e-01 -1.36380270e-01
-8.65621865e-01 3.15768480e-01 8.41666162e-01 7.01030314e-01
-9.98201221e-02 2.25273609e-01 6.49541199e-01 -4.83390540e-01
-1.32071400e+00 -1.14728618e+00 -9.57486451e-01 1.20620526e-01
3.38337809e-01 -7.78031588e-01 -1.14551711e+00 -1.65477201e-01] | [8.265610694885254, 4.288925647735596] |
44bf6547-0c0d-489b-8fc8-834afc3c12cd | npc-neuron-path-coverage-via-characterizing | 2203.12915 | null | https://arxiv.org/abs/2203.12915v2 | https://arxiv.org/pdf/2203.12915v2.pdf | NPC: Neuron Path Coverage via Characterizing Decision Logic of Deep Neural Networks | Deep learning has recently been widely applied to many applications across different domains, e.g., image classification and audio recognition. However, the quality of Deep Neural Networks (DNNs) still raises concerns in the practical operational environment, which calls for systematic testing, especially in safety-critical scenarios. Inspired by software testing, a number of structural coverage criteria are designed and proposed to measure the test adequacy of DNNs. However, due to the blackbox nature of DNN, the existing structural coverage criteria are difficult to interpret, making it hard to understand the underlying principles of these criteria. The relationship between the structural coverage and the decision logic of DNNs is unknown. Moreover, recent studies have further revealed the non-existence of correlation between the structural coverage and DNN defect detection, which further posts concerns on what a suitable DNN testing criterion should be. In this paper, we propose the interpretable coverage criteria through constructing the decision structure of a DNN. Mirroring the control flow graph of the traditional program, we first extract a decision graph from a DNN based on its interpretation, where a path of the decision graph represents a decision logic of the DNN. Based on the control flow and data flow of the decision graph, we propose two variants of path coverage to measure the adequacy of the test cases in exercising the decision logic. The higher the path coverage, the more diverse decision logic the DNN is expected to be explored. Our large-scale evaluation results demonstrate that: the path in the decision graph is effective in characterizing the decision of the DNN, and the proposed coverage criteria are also sensitive with errors including natural errors and adversarial examples, and strongly correlated with the output impartiality. | ['Yang Liu', 'Felix Juefei-Xu', 'Qing Guo', 'Lei Ma', 'Jian Wang', 'Tianlin Li', 'Xiaofei Xie'] | 2022-03-24 | null | null | null | null | ['dnn-testing', 'defect-detection'] | ['adversarial', 'computer-vision'] | [ 2.54754335e-01 2.41359279e-01 -2.58119315e-01 -3.64364892e-01
2.17979327e-01 -6.88646197e-01 9.27653462e-02 -3.61372344e-02
1.10881580e-02 5.93463838e-01 -3.03788692e-01 -8.23476791e-01
-4.07988787e-01 -1.12301600e+00 -1.01457512e+00 -5.15150964e-01
2.73628160e-02 -1.70965001e-01 4.01095897e-01 2.11471003e-02
6.56620339e-02 4.63791728e-01 -1.71310234e+00 1.71758577e-01
1.07740796e+00 1.32756424e+00 -9.52473730e-02 1.60121575e-01
3.41727212e-02 7.44519949e-01 -8.17806005e-01 -4.80720937e-01
2.27612108e-01 -5.68871796e-01 -7.88623154e-01 -6.48823082e-02
9.76192281e-02 -5.38313806e-01 -2.40804806e-01 1.74224496e+00
1.51419044e-01 -4.04219121e-01 2.57438272e-01 -1.95602918e+00
-2.38825664e-01 8.77162755e-01 3.74358217e-03 -3.84850614e-02
1.56569570e-01 3.41981411e-01 1.11185980e+00 -2.41726354e-01
2.77625352e-01 1.07840919e+00 4.43732142e-01 6.82225883e-01
-9.36666310e-01 -8.36493790e-01 1.89070195e-01 1.15483224e-01
-1.03728437e+00 -1.83282763e-01 6.65642142e-01 -5.26717544e-01
6.39102042e-01 3.08844835e-01 8.77026737e-01 1.25375104e+00
4.28188860e-01 7.69654036e-01 9.42541420e-01 -2.48948574e-01
6.96334541e-01 1.85889974e-01 2.10485116e-01 6.04825616e-01
8.07431698e-01 5.06383121e-01 -1.59217209e-01 -1.17348105e-01
6.29884243e-01 4.75978060e-03 -3.37738842e-01 -3.69614840e-01
-7.26875663e-01 7.39715159e-01 1.96930289e-01 2.52334595e-01
9.39173251e-02 1.57431677e-01 6.96817040e-01 5.71121931e-01
-3.69661897e-02 3.57372880e-01 -5.87379038e-01 -2.51041085e-01
-5.70706666e-01 7.10443109e-02 9.23724532e-01 9.21132863e-01
6.35474622e-01 3.45940620e-01 -8.29213485e-02 1.72785506e-01
3.69585901e-01 3.52820396e-01 3.60962629e-01 -7.96762645e-01
3.56042117e-01 1.15373087e+00 -3.74823868e-01 -9.61812675e-01
-1.91767216e-01 -6.23915195e-01 -8.95297647e-01 5.29559672e-01
2.43408903e-01 -3.68202984e-01 -4.94957179e-01 1.97083080e+00
3.12113389e-02 1.55966476e-01 1.33485347e-01 7.91877091e-01
5.96353531e-01 2.74574071e-01 -3.72290045e-01 -1.65885910e-01
9.51339066e-01 -3.91698778e-01 -5.23225367e-01 -3.00240874e-01
7.58542180e-01 7.74469925e-04 1.18912590e+00 5.37008941e-01
-6.32267773e-01 -4.64954376e-01 -1.54514265e+00 6.37363493e-01
8.43182579e-02 -1.30762413e-01 5.24389625e-01 1.02500641e+00
-9.16929007e-01 5.15352488e-01 -7.38927782e-01 1.82285476e-02
5.14930367e-01 4.81070310e-01 -2.24841177e-01 -2.87847549e-01
-1.31695879e+00 5.32999218e-01 6.08269691e-01 3.38887721e-01
-1.31504416e+00 -4.29062724e-01 -8.15595984e-01 2.25297928e-01
5.44913709e-01 -2.29520917e-01 1.01188767e+00 -1.29559577e+00
-1.16964877e+00 3.54673326e-01 3.73190343e-01 -3.89572114e-01
6.56934977e-01 2.79618412e-01 -8.33381295e-01 -1.08432822e-01
-1.62929535e-01 2.57362604e-01 5.49403906e-01 -1.06316161e+00
-7.07357705e-01 -3.28464925e-01 5.99792302e-01 -3.15976888e-01
-4.29860741e-01 -1.19028971e-01 -1.24616802e-01 -3.06552202e-01
-1.16842072e-02 -7.20905006e-01 -6.04587905e-02 8.01915377e-02
-8.12641203e-01 1.58501174e-02 9.11821544e-01 -2.25688815e-01
1.42179167e+00 -2.38271618e+00 -2.23783627e-01 6.72307670e-01
3.54590595e-01 2.37400532e-01 -9.75667834e-02 -9.17263329e-02
-9.05739889e-02 5.10416210e-01 -3.76256704e-01 5.12807012e-01
2.61794120e-01 2.50406951e-01 -3.61324400e-01 4.26717997e-01
4.21928108e-01 6.85631335e-01 -6.43362641e-01 -9.56341848e-02
-1.79721683e-01 7.99836125e-03 -7.76809812e-01 2.21530631e-01
-4.49823886e-01 2.16792718e-01 -6.67437017e-01 6.96234822e-01
5.42091429e-01 5.54725155e-02 1.74799144e-01 -4.42202389e-02
2.16094613e-01 2.45441243e-01 -1.20179987e+00 9.66647625e-01
-8.83560255e-02 7.12324440e-01 -1.91391826e-01 -1.21629286e+00
1.02760971e+00 2.06835404e-01 3.15161236e-02 -7.42024720e-01
1.84048876e-01 2.50737578e-01 6.65519893e-01 -5.97508371e-01
8.06956664e-02 -2.56015360e-02 -2.51728565e-01 3.76826286e-01
1.54854786e-02 2.81268537e-01 7.02534467e-02 -2.39711240e-01
1.49285185e+00 -1.00312062e-01 1.63302153e-01 -3.53871405e-01
5.74278355e-01 -6.01766296e-02 9.88252044e-01 6.55576706e-01
-2.58188158e-01 2.37594336e-01 1.42135644e+00 -3.31631035e-01
-7.64483273e-01 -7.97652900e-01 -2.25899205e-01 3.43549281e-01
2.10097119e-01 -1.26696318e-01 -1.00743484e+00 -1.01666772e+00
-1.04864679e-01 8.01145315e-01 -4.76120114e-01 -8.38812351e-01
-1.42243445e-01 -4.09611702e-01 1.18518639e+00 4.43498790e-01
8.19037139e-01 -1.15026891e+00 -8.86860967e-01 -1.68859102e-02
1.40604228e-01 -1.08497667e+00 -7.58179575e-02 1.74121708e-01
-9.34958994e-01 -1.36915493e+00 1.39760733e-01 -6.20080233e-01
8.01641405e-01 -2.93348402e-01 8.75636339e-01 3.93146366e-01
-1.20983617e-02 2.03803703e-02 -3.34318399e-01 -2.82836199e-01
-9.00212109e-01 -2.42085978e-01 1.47443384e-01 1.59486420e-02
3.00208986e-01 -4.84463185e-01 -2.13275805e-01 6.42382205e-01
-1.34039748e+00 -2.59951472e-01 6.78158820e-01 7.01742053e-01
5.46759129e-01 6.77741170e-01 5.61707318e-01 -7.97672868e-01
8.14525485e-01 -3.55114043e-01 -9.18842852e-01 4.88012969e-01
-8.46268654e-01 2.68864810e-01 8.59096885e-01 -6.22049630e-01
-5.03158987e-01 -3.73252362e-01 -2.97677815e-01 -4.62601364e-01
-1.45687357e-01 8.01364720e-01 -1.06133366e+00 -9.43320431e-03
6.04763329e-01 -5.54681532e-02 -1.02308378e-01 1.50811197e-02
-2.96372175e-01 2.82586813e-01 1.60525650e-01 -6.60721064e-01
6.14248335e-01 2.67025214e-02 1.49724230e-01 -2.60087281e-01
-3.08056474e-01 4.28170830e-01 -3.53030376e-02 -2.92071193e-01
6.56979084e-01 -2.72639722e-01 -1.02366471e+00 4.68989730e-01
-1.10509121e+00 -2.13293895e-01 -3.12046498e-01 4.17976618e-01
-1.72367677e-01 2.82297373e-01 -3.21700662e-01 -7.83651233e-01
-1.18278004e-01 -1.74951279e+00 3.74446571e-01 4.03905325e-02
-1.51185647e-01 -8.68822515e-01 -2.60288984e-01 -2.10114762e-01
8.04495513e-02 3.48339647e-01 1.50084782e+00 -9.75986481e-01
-4.70370531e-01 -5.11761785e-01 -1.03639988e-02 9.15478706e-01
-1.16823778e-01 1.71683341e-01 -9.85387206e-01 -2.11527601e-01
4.47573185e-01 -1.90783113e-01 2.17645392e-01 2.62104690e-01
1.38062930e+00 -6.97521329e-01 -2.22822540e-02 4.91925150e-01
1.47605646e+00 5.01506805e-01 8.81205440e-01 2.61443019e-01
4.21807736e-01 6.06959701e-01 4.71631914e-01 3.41273248e-01
-2.48773545e-01 2.21538827e-01 9.37282562e-01 7.01851249e-02
4.64969277e-01 -3.88930649e-01 8.60751450e-01 3.09477597e-01
2.76625425e-01 -4.39340621e-01 -9.70416784e-01 1.51808500e-01
-1.56988406e+00 -7.71953881e-01 -6.68299273e-02 2.34743547e+00
6.98527098e-01 7.22231865e-01 -1.93674430e-01 6.58669531e-01
7.86105752e-01 -1.13206416e-01 -9.81369436e-01 -5.64067602e-01
-1.80458933e-01 2.84539536e-02 3.94473672e-01 1.90243840e-01
-5.18224657e-01 4.13607568e-01 5.63459349e+00 9.10793662e-01
-1.21656895e+00 -1.23701721e-01 7.40343809e-01 1.75243333e-01
-7.31593907e-01 2.52454221e-01 -6.62918270e-01 5.20800829e-01
7.17872977e-01 -2.45123535e-01 2.42876530e-01 8.90603840e-01
8.02745298e-02 1.00689873e-01 -1.61261725e+00 5.22505283e-01
-2.44188771e-01 -9.39420938e-01 8.16999003e-03 3.44986767e-01
5.21240771e-01 -2.84575522e-01 -2.49195155e-02 3.03666800e-01
1.79056421e-01 -1.16747010e+00 1.00838912e+00 1.83417991e-01
9.13682163e-01 -9.00965035e-01 1.07659137e+00 5.87622583e-01
-6.96885347e-01 -2.91970640e-01 -1.84645981e-01 -3.18507440e-02
-4.90241796e-01 9.07533526e-01 -6.41524315e-01 2.84159660e-01
6.75669789e-01 3.13941300e-01 -5.16887963e-01 8.62709105e-01
-7.45942473e-01 7.03710139e-01 -4.46841381e-02 -1.92732066e-01
-3.26016098e-02 -1.38391688e-01 5.09908676e-01 7.13654697e-01
4.39334422e-01 -3.47181708e-01 -2.34171391e-01 1.39734137e+00
-1.85937673e-01 -3.01233619e-01 -7.62814820e-01 -1.92600727e-01
5.53914309e-01 9.44005787e-01 -5.87470353e-01 8.46871063e-02
-3.44563365e-01 5.20009518e-01 3.41824293e-02 4.65857029e-01
-1.03773117e+00 -3.13701719e-01 7.19027460e-01 2.56339274e-02
2.38450058e-02 1.19046554e-01 -5.94983876e-01 -8.00155520e-01
6.00227535e-01 -1.21295381e+00 5.65620027e-02 -4.12773073e-01
-1.00319517e+00 8.40548933e-01 -3.00626401e-02 -1.55947268e+00
-1.79446101e-01 -6.65473104e-01 -7.93698967e-01 6.07868969e-01
-1.00465024e+00 -5.44537008e-01 -2.17893824e-01 5.61011076e-01
1.79953892e-02 -3.11682791e-01 6.17726803e-01 2.18713284e-01
-1.04969096e+00 9.97487843e-01 -3.99670094e-01 3.22753638e-01
1.16951339e-01 -7.75997102e-01 1.20137066e-01 1.36065197e+00
-2.79886574e-01 6.88151658e-01 6.51193619e-01 -5.96127808e-01
-1.44599593e+00 -1.09982252e+00 3.85226041e-01 9.08292532e-02
6.88659012e-01 -4.42262083e-01 -8.34676981e-01 5.38101017e-01
-2.05807894e-01 7.68799856e-02 5.18940866e-01 -5.84758818e-02
-4.29436028e-01 -5.12489319e-01 -1.42181015e+00 6.56069100e-01
1.10377705e+00 -4.78821069e-01 -3.58102113e-01 -1.54494017e-01
9.78595078e-01 -1.64890051e-01 -5.86264968e-01 7.22557247e-01
5.42160988e-01 -1.17182279e+00 3.56743395e-01 -5.63432097e-01
8.27938020e-01 -3.88963878e-01 -3.07300061e-01 -1.00073135e+00
7.03927726e-02 -2.35207140e-01 -2.15402350e-01 1.27121544e+00
6.88365638e-01 -8.80687654e-01 8.30346167e-01 7.75259078e-01
-3.25758159e-01 -8.30607176e-01 -1.13889217e+00 -9.75455046e-01
-1.74343392e-01 -9.68740821e-01 9.40780342e-01 9.25753891e-01
-3.62269953e-02 2.77968869e-02 5.00845201e-02 3.49607021e-01
2.68110067e-01 -1.88655689e-01 3.48776489e-01 -1.36249530e+00
-6.06385231e-01 -6.92004323e-01 -8.78643513e-01 -8.09261978e-01
4.40747857e-01 -8.44405651e-01 8.68455786e-03 -1.20499146e+00
-6.37215003e-02 -5.80636203e-01 -2.90859282e-01 7.03327239e-01
2.68223464e-01 -4.17263538e-01 -1.46148726e-01 -5.50444005e-03
-1.71274796e-01 3.73806089e-01 1.06211507e+00 -4.23725426e-01
-1.98659301e-02 3.18174988e-01 -8.43848348e-01 7.45214939e-01
7.65533686e-01 -5.13819218e-01 -7.81697154e-01 -3.19689661e-01
7.02786446e-01 1.13233894e-01 4.35682476e-01 -1.08339167e+00
2.21288353e-01 -2.44417965e-01 -1.54070720e-01 -2.65500277e-01
-3.64451498e-01 -1.22296607e+00 3.46800864e-01 9.08926785e-01
-3.38036627e-01 8.51416662e-02 1.17229126e-01 4.77405578e-01
-4.43904310e-01 -6.42219901e-01 3.52156341e-01 3.72089416e-01
-7.61101484e-01 3.87982488e-01 -3.20235968e-01 8.65855590e-02
9.99563038e-01 -4.70634848e-01 -2.92094737e-01 -1.16895825e-01
-2.04994708e-01 1.73928112e-01 4.64257449e-01 3.02467197e-01
8.71686757e-01 -1.31717038e+00 -3.73064995e-01 6.63929164e-01
1.73604682e-01 2.79290646e-01 3.13211270e-02 8.09853733e-01
-4.83245224e-01 5.40023707e-02 -2.94514686e-01 -5.57924032e-01
-8.89697552e-01 4.60257709e-01 6.77860498e-01 -1.27926528e-01
-2.91315794e-01 4.93849874e-01 3.48820060e-01 -2.97330886e-01
5.60244024e-01 -8.72989178e-01 -5.10866232e-02 -3.50070596e-01
3.36095363e-01 2.49269053e-01 3.06428671e-01 8.31578001e-02
-4.35494721e-01 3.39724235e-02 3.09103161e-01 -1.14578702e-01
1.01926887e+00 3.77002448e-01 -2.97698319e-01 2.99854845e-01
1.03151083e+00 -1.02631785e-01 -1.12681723e+00 7.66555518e-02
9.32755396e-02 -8.47495943e-02 4.00766507e-02 -7.54372060e-01
-1.45897353e+00 9.03814316e-01 4.19226378e-01 4.55234796e-01
1.53342068e+00 -3.11489612e-01 3.34316760e-01 4.57304150e-01
4.90093559e-01 -8.51213634e-01 -1.27150670e-01 4.63872671e-01
6.77480340e-01 -9.12871897e-01 -3.85782421e-01 -2.99849629e-01
-4.21383798e-01 1.23875439e+00 8.09048772e-01 1.37449071e-01
6.58511758e-01 6.19705141e-01 -4.62454408e-01 -1.30114242e-01
-5.66083431e-01 3.31496328e-01 1.60970360e-01 5.19317806e-01
-5.87690398e-02 4.37793176e-04 -2.05573887e-01 1.33827007e+00
-2.22073033e-01 1.31496400e-01 6.22168958e-01 7.21234679e-01
-3.12005579e-01 -8.42242181e-01 -1.57948896e-01 5.08858085e-01
-2.81448841e-01 -4.98875864e-02 -3.76561433e-01 7.80446887e-01
4.63905126e-01 1.03694999e+00 -1.70389302e-02 -1.00872362e+00
5.08573472e-01 -2.49801099e-01 1.52538925e-01 -4.01951879e-01
-3.54736537e-01 -3.60622466e-01 2.94602245e-01 -5.07142425e-01
-1.82117015e-01 -3.82658422e-01 -1.42123127e+00 -4.68523026e-01
-4.40316647e-01 3.19283679e-02 5.26624262e-01 1.20879889e+00
1.90467149e-01 6.19518220e-01 8.02672386e-01 1.53866336e-01
-7.84269810e-01 -6.13391936e-01 -6.96649015e-01 7.32017010e-02
2.31151938e-01 -4.30407614e-01 -5.90046942e-01 -4.70626533e-01] | [6.537837505340576, 7.650041103363037] |
3734f9f3-51e3-47a9-b901-c39eb45dca85 | videofactory-swap-attention-in-spatiotemporal | 2305.10874 | null | https://arxiv.org/abs/2305.10874v2 | https://arxiv.org/pdf/2305.10874v2.pdf | VideoFactory: Swap Attention in Spatiotemporal Diffusions for Text-to-Video Generation | We present VideoFactory, an innovative framework for generating high-quality open-domain videos. VideoFactory excels in producing high-definition (1376x768), widescreen (16:9) videos without watermarks, creating an engaging user experience. Generating videos guided by text instructions poses significant challenges, such as modeling the complex relationship between space and time, and the lack of large-scale text-video paired data. Previous approaches extend pretrained text-to-image generation models by adding temporal 1D convolution/attention modules for video generation. However, these approaches overlook the importance of jointly modeling space and time, inevitably leading to temporal distortions and misalignment between texts and videos. In this paper, we propose a novel approach that strengthens the interaction between spatial and temporal perceptions. In particular, we utilize a swapped cross-attention mechanism in 3D windows that alternates the "query" role between spatial and temporal blocks, enabling mutual reinforcement for each other. To fully unlock model capabilities for high-quality video generation, we curate a large-scale video dataset called HD-VG-130M. This dataset comprises 130 million text-video pairs from the open-domain, ensuring high-definition, widescreen and watermark-free characters. Objective metrics and user studies demonstrate the superiority of our approach in terms of per-frame quality, temporal correlation, and text-video alignment, with clear margins. | ['Jiaying Liu', 'Jianlong Fu', 'Junchen Zhu', 'Huiguo He', 'Zixi Tuo', 'Huan Yang', 'Wenjing Wang'] | 2023-05-18 | null | null | null | null | ['video-generation', 'video-alignment', 'text-to-video-generation'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 2.62611032e-01 -4.76600200e-01 -1.68680668e-01 7.66080543e-02
-5.85688174e-01 -7.50024676e-01 5.76250255e-01 -3.13351750e-01
-2.02590421e-01 4.59117532e-01 3.98874342e-01 -4.48495209e-01
2.67406777e-02 -5.96229136e-01 -9.26608622e-01 -3.44892919e-01
-1.57311723e-01 -6.56454206e-01 2.56262451e-01 1.40677327e-02
2.22187504e-01 1.83815867e-01 -1.64328337e+00 5.27178288e-01
8.70936751e-01 9.68415022e-01 4.58903790e-01 1.02658129e+00
1.59406275e-01 5.68007886e-01 -7.11532950e-01 -4.56077754e-01
3.66960347e-01 -2.84978896e-01 -2.09236637e-01 1.89956263e-01
6.35357380e-01 -9.37362611e-01 -8.50622773e-01 6.84284329e-01
6.91573918e-01 -6.33789152e-02 2.56262392e-01 -1.48544610e+00
-1.09178543e+00 3.81922781e-01 -7.31932282e-01 3.96488570e-02
8.58441830e-01 7.03554928e-01 7.88054287e-01 -6.67393506e-01
6.28684223e-01 1.11728609e+00 5.38068593e-01 4.73111957e-01
-1.14845502e+00 -9.92372274e-01 2.15730369e-01 1.85086340e-01
-1.29613924e+00 -5.93669653e-01 5.67098737e-01 -5.26133478e-01
9.46929395e-01 4.22745705e-01 9.35151100e-01 1.65293050e+00
3.81259471e-01 6.19762182e-01 6.09628916e-01 -2.65485704e-01
-6.92432746e-03 -2.35957429e-01 -6.33590460e-01 2.82262176e-01
1.26433685e-01 3.70360523e-01 -9.61896658e-01 1.38791919e-01
1.31132412e+00 -4.86512631e-02 -7.20802724e-01 -3.17547977e-01
-1.65144634e+00 4.41622227e-01 -3.41602154e-02 1.49674788e-01
-2.20091373e-01 2.94277698e-01 3.40827167e-01 2.07654685e-01
1.92122847e-01 3.32937658e-01 -2.82242566e-01 -5.63733578e-01
-1.06558490e+00 2.61990011e-01 2.38460243e-01 1.43776381e+00
2.47705922e-01 2.22141668e-01 -5.57650983e-01 5.03244340e-01
3.71339768e-01 6.82853222e-01 5.68865895e-01 -1.01069212e+00
7.91737258e-01 1.76634386e-01 2.81618208e-01 -1.21724331e+00
2.20908336e-02 6.96737617e-02 -6.04944587e-01 1.60495698e-01
1.47020549e-01 -2.78924078e-01 -7.43590415e-01 1.77503836e+00
4.14481610e-02 3.50870520e-01 -1.92777634e-01 1.00354254e+00
8.23802829e-01 8.87542844e-01 7.29501694e-02 -2.27132961e-01
1.29168916e+00 -8.75527799e-01 -1.10657883e+00 7.90911317e-02
2.14677870e-01 -8.97526503e-01 1.38951600e+00 2.84978271e-01
-1.25372970e+00 -8.85310769e-01 -1.24307835e+00 -1.20972395e-01
-1.49478242e-01 4.24077921e-02 3.95139337e-01 6.83541477e-01
-1.08536708e+00 5.35493195e-01 -3.62275839e-01 -1.24816790e-01
1.84398398e-01 1.60091594e-01 -4.37850267e-01 8.00898299e-02
-1.22317672e+00 3.47134113e-01 2.36573294e-01 -1.66737989e-01
-6.55881941e-01 -8.95929575e-01 -9.94101226e-01 -1.11583047e-01
3.90434444e-01 -4.96484607e-01 1.15875113e+00 -9.44986105e-01
-1.54915655e+00 4.40665662e-01 1.97707728e-01 -1.78894922e-01
6.62292600e-01 -4.25551265e-01 -6.60869002e-01 4.25047517e-01
1.03656352e-02 1.13630402e+00 1.27909374e+00 -1.24938369e+00
-5.76435387e-01 6.26203045e-02 2.58292109e-01 2.79132515e-01
-6.66078568e-01 -8.82257968e-02 -8.27031970e-01 -1.25205624e+00
-3.20226133e-01 -6.27688706e-01 2.34865442e-01 3.29003423e-01
-2.22844422e-01 5.20661592e-01 1.13696146e+00 -8.73584449e-01
1.65642428e+00 -2.45286536e+00 -1.11908279e-01 -1.51393369e-01
3.14851344e-01 3.00679833e-01 -4.52365696e-01 3.59671265e-01
-2.92182356e-01 3.22731674e-01 2.95823902e-01 -2.33001336e-01
6.14232197e-02 -1.18067823e-01 -4.49059248e-01 1.80769548e-01
2.87256092e-01 1.14908731e+00 -7.95400381e-01 -4.43378389e-01
3.38409483e-01 6.51726007e-01 -7.87872732e-01 2.10479707e-01
-2.42012709e-01 3.23130071e-01 -6.77845031e-02 6.45511806e-01
7.79560387e-01 -1.24266639e-01 2.22936183e-01 -4.91706043e-01
-2.03341007e-01 -9.72366035e-02 -1.18990827e+00 1.93218911e+00
-3.40468675e-01 1.05684042e+00 -6.39207214e-02 -4.37520519e-02
6.23240411e-01 4.30180252e-01 6.12122536e-01 -1.23262143e+00
5.14561730e-03 -8.25721845e-02 -3.38305831e-01 -7.43521452e-01
9.78386819e-01 3.41542214e-01 5.11308247e-03 2.71109253e-01
-8.61288980e-02 -1.14673890e-01 3.37303311e-01 3.58771026e-01
9.13962901e-01 5.44117093e-01 -8.55709985e-02 2.38561288e-01
-1.87241137e-01 -5.41389108e-01 1.67016983e-01 4.93779480e-01
-1.56047285e-01 9.54673052e-01 5.86344481e-01 1.66012291e-02
-1.37404633e+00 -1.11902082e+00 2.10563943e-01 8.58397782e-01
3.02555293e-01 -7.79366493e-01 -7.63086379e-01 -3.38325560e-01
-2.02160582e-01 4.88766104e-01 -4.27038074e-01 -2.25444913e-01
-4.96911079e-01 -1.80102319e-01 4.80727464e-01 4.56524283e-01
4.10495013e-01 -8.87087524e-01 -8.51039529e-01 2.45190740e-01
-5.42066932e-01 -1.39253020e+00 -1.22600365e+00 -4.44904536e-01
-6.82881892e-01 -1.02976382e+00 -9.73932087e-01 -5.29604375e-01
5.10657310e-01 6.35301650e-01 8.84413064e-01 -2.27369845e-01
-2.16294587e-01 4.01797295e-01 -5.33166468e-01 1.99453190e-01
-2.18604252e-01 -3.89548689e-01 7.81754255e-02 8.52268189e-02
-1.58233657e-01 -5.64921260e-01 -7.35061824e-01 4.55192924e-01
-1.32306433e+00 6.42338932e-01 4.65330660e-01 7.68088579e-01
1.91529453e-01 -2.17423029e-02 2.50464320e-01 5.89038357e-02
6.72538996e-01 -2.95919716e-01 -4.79992926e-01 2.43447591e-02
-2.57571131e-01 -3.87587786e-01 5.14762163e-01 -9.34080660e-01
-8.99617255e-01 -2.43484363e-01 1.62593618e-01 -6.71004057e-01
6.24076948e-02 1.37080416e-01 -4.99677151e-01 6.69289529e-02
3.85986239e-01 1.95678994e-01 -7.99501315e-02 -8.19148123e-02
5.84163547e-01 7.32671678e-01 6.76141024e-01 -5.09853780e-01
8.57129037e-01 3.35985452e-01 -6.35251343e-01 -6.94337308e-01
8.11234489e-02 6.17659243e-04 -2.59925246e-01 -4.25363123e-01
8.68938386e-01 -1.19868505e+00 -8.26137483e-01 7.79281020e-01
-1.18011153e+00 -5.67399085e-01 -3.46382380e-01 4.61770564e-01
-6.87232971e-01 6.09139800e-01 -6.96683586e-01 -3.98413956e-01
-1.21613473e-01 -1.25816834e+00 1.13610053e+00 2.82477796e-01
-2.72730708e-01 -4.81888920e-01 -3.15651178e-01 2.92499751e-01
4.94940847e-01 1.33212775e-01 5.87533712e-01 1.22181259e-01
-9.05237734e-01 -2.80350950e-02 -4.94782329e-01 2.13624358e-01
2.25869894e-01 3.94714087e-01 -6.79198384e-01 -3.79473239e-01
-3.43694508e-01 -1.50087014e-01 3.69778216e-01 3.82246286e-01
1.23078823e+00 -3.94511938e-01 -4.70910855e-02 7.70458579e-01
1.15480161e+00 5.07084012e-01 1.09063673e+00 2.46237949e-01
7.76529789e-01 2.39453584e-01 6.43203676e-01 8.14326882e-01
3.74758929e-01 8.23978007e-01 3.03538173e-01 -2.14862213e-01
-2.62548476e-01 -6.55735552e-01 6.08447969e-01 6.44271076e-01
1.86696276e-01 -6.24373138e-01 -3.42291057e-01 4.42072630e-01
-1.60595644e+00 -1.10019958e+00 1.04950517e-01 2.26919699e+00
9.94401455e-01 2.51956761e-01 3.51772010e-02 2.03545034e-01
8.13060462e-01 3.80860388e-01 -3.51388007e-01 -2.40025848e-01
-3.65045130e-01 3.98542657e-02 3.28881502e-01 2.37457335e-01
-9.96153176e-01 6.59864247e-01 6.57517290e+00 1.17678380e+00
-1.37054384e+00 -1.23844847e-01 5.82167745e-01 -6.29397988e-01
-5.28873086e-01 -2.38796279e-01 -4.87062871e-01 9.46483910e-01
9.23805952e-01 -1.53308092e-02 4.48778927e-01 3.16339701e-01
6.79239869e-01 -6.91891685e-02 -9.86182213e-01 1.20229483e+00
1.04990497e-01 -1.47528207e+00 2.00067818e-01 1.85883701e-01
7.85566688e-01 -6.55247748e-01 3.13762844e-01 9.19836238e-02
-1.70915455e-01 -9.24603641e-01 1.19907892e+00 3.60631704e-01
1.58455634e+00 -4.33723569e-01 6.70660287e-02 -1.77321568e-01
-1.38132131e+00 -5.52965440e-02 2.94605017e-01 1.74434260e-01
3.87854308e-01 3.47815394e-01 -2.50260681e-01 4.22529101e-01
7.19444394e-01 6.57304347e-01 -4.14235622e-01 9.14822400e-01
5.44560477e-02 1.32000372e-01 -1.57502651e-01 2.42893264e-01
-5.02442233e-02 5.88261448e-02 4.21761483e-01 1.20979095e+00
7.17637718e-01 2.44432148e-02 -1.46212518e-01 8.67595077e-01
5.86983515e-03 -1.30746499e-01 -7.34328091e-01 -3.94213885e-01
5.42321980e-01 9.27480221e-01 -4.32983607e-01 -2.03327760e-01
-4.90021110e-01 1.15553510e+00 -4.05796587e-01 5.84944904e-01
-1.50395906e+00 -4.78042603e-01 8.10114861e-01 2.96684533e-01
5.72943568e-01 -5.82958579e-01 -9.35623571e-02 -1.13935149e+00
3.90987933e-01 -1.35543799e+00 -1.72586560e-01 -1.22613609e+00
-8.00357759e-01 3.07778597e-01 3.47160436e-02 -1.57258809e+00
-1.13642231e-01 -4.16424483e-01 -2.70369649e-01 6.17077827e-01
-1.26442695e+00 -1.27952754e+00 -5.00633776e-01 7.29084790e-01
7.00885892e-01 4.34866101e-02 4.82203394e-01 7.30246961e-01
-4.08747315e-01 8.12087536e-01 -2.30013561e-02 -4.34034504e-02
1.03657174e+00 -8.09409618e-01 6.93069935e-01 9.71174955e-01
-1.57059669e-01 5.17612159e-01 4.31422740e-01 -7.77651489e-01
-1.67743468e+00 -9.15194631e-01 4.50363636e-01 -2.54243314e-01
5.23648262e-01 -5.09834290e-01 -5.46724916e-01 3.48664939e-01
4.13448006e-01 -2.09554389e-01 7.12397039e-01 -6.11405551e-01
-4.75254059e-01 -6.75599501e-02 -8.37605536e-01 1.17276371e+00
1.11519337e+00 -6.19678736e-01 -1.20089715e-02 -2.74689436e-01
1.23893023e+00 -6.31397307e-01 -8.09919178e-01 3.08249980e-01
9.12697792e-01 -1.19277179e+00 9.76761580e-01 -1.23176217e-01
9.08914864e-01 -4.32642221e-01 -1.88905686e-01 -9.99929667e-01
-2.13058934e-01 -1.27506185e+00 -4.00645792e-01 1.28595197e+00
9.40044448e-02 -2.33761877e-01 5.76630712e-01 5.97119808e-01
-3.73579301e-02 -5.86480916e-01 -7.41233170e-01 -5.31569421e-01
-4.25924182e-01 -7.54722118e-01 6.74277425e-01 8.36596191e-01
2.09416315e-01 -4.11430411e-02 -8.93604517e-01 -5.25037311e-02
1.83259502e-01 -2.10193142e-01 8.67224813e-01 -3.27067226e-01
-4.42581475e-01 -4.48866904e-01 -3.43815088e-01 -1.57197773e+00
-3.78160059e-01 -3.14503126e-02 -1.81480810e-01 -1.20037985e+00
1.79438666e-02 -1.16062917e-01 1.64999411e-01 2.49364406e-01
-1.02480330e-01 3.55662763e-01 4.91221875e-01 -8.07986110e-02
-5.42012274e-01 6.22686863e-01 1.56460989e+00 -7.14563206e-02
-1.80351526e-01 -6.03661656e-01 -7.67088413e-01 3.01957309e-01
5.31438470e-01 1.52378693e-01 -6.98096573e-01 -7.32822835e-01
1.32241324e-01 2.49981567e-01 5.80121636e-01 -1.09439778e+00
5.49786463e-02 -3.21471632e-01 5.26026547e-01 -5.78525782e-01
4.54344392e-01 -8.03166270e-01 5.16444862e-01 2.05498263e-01
-1.78704172e-01 3.62164825e-01 6.01406336e-01 6.08631015e-01
-1.24745272e-01 2.54780948e-01 4.75474685e-01 1.42106056e-01
-7.36014843e-01 3.64051014e-01 -4.08848315e-01 -6.67668730e-02
1.18510509e+00 -7.35953152e-01 -5.41204751e-01 -6.78896666e-01
-3.62510800e-01 3.99086848e-02 7.74894774e-01 9.91251826e-01
1.01647294e+00 -1.58447385e+00 -4.31912422e-01 6.42633140e-01
4.81858999e-02 -2.99686193e-01 8.23688209e-01 4.99756813e-01
-7.03111827e-01 3.26543689e-01 -4.09398913e-01 -4.83004838e-01
-1.12580788e+00 6.52302146e-01 -8.00728574e-02 6.19242527e-03
-5.77408552e-01 5.00990212e-01 2.93999374e-01 2.33750924e-01
3.29848230e-01 -4.41290915e-01 1.93735793e-01 6.78408518e-02
7.17993796e-01 1.71174735e-01 -2.34356046e-01 -4.13583964e-01
2.62746289e-02 4.75562960e-01 8.28139484e-02 -2.82390088e-01
8.74663889e-01 -4.83986199e-01 4.01328444e-01 5.65896146e-02
1.12042654e+00 3.08101356e-01 -1.79300702e+00 7.11966008e-02
-6.78116739e-01 -9.92701352e-01 -2.63038516e-01 -6.65071070e-01
-1.05982053e+00 7.32684910e-01 6.75251484e-01 1.21627122e-01
1.26723969e+00 -5.87677419e-01 1.21729136e+00 -3.56498539e-01
1.75263107e-01 -1.08765137e+00 7.55319953e-01 2.88514614e-01
9.69159484e-01 -1.02886224e+00 -1.90463230e-01 -2.74520189e-01
-6.18877411e-01 1.01591897e+00 8.14265192e-01 4.35081810e-01
2.02987596e-01 5.44176042e-01 -3.57968397e-02 2.71699041e-01
-8.27280223e-01 2.42622659e-01 2.16944650e-01 9.17723954e-01
3.67428392e-01 -5.13767526e-02 -2.74516083e-02 4.95226890e-01
-1.40322044e-01 1.27542347e-01 5.86784482e-01 9.74675477e-01
-5.85617730e-03 -9.74048674e-01 -5.58469594e-01 1.08311519e-01
-3.86499047e-01 -3.62987697e-01 1.69999406e-01 6.79584980e-01
2.83478945e-01 9.59452987e-01 2.70191461e-01 -7.48527288e-01
2.44779274e-01 -4.33808476e-01 4.12241310e-01 -5.92199527e-02
-3.05772126e-01 4.36298490e-01 -8.57620221e-03 -8.03955436e-01
-1.90412685e-01 -3.91459495e-01 -7.59341717e-01 -6.90074801e-01
-1.65444449e-01 -3.18477958e-01 6.56024158e-01 4.17066753e-01
8.34250629e-01 8.86897624e-01 5.82025111e-01 -1.18636215e+00
-3.29708040e-01 -6.21101141e-01 -4.05725926e-01 5.21514356e-01
4.05314416e-01 -3.47923785e-01 -6.30607307e-02 5.43366611e-01] | [10.858745574951172, -0.5856627225875854] |
1cf9a888-83a4-4762-9849-5804db40b1e3 | an-image-dehazing-approach-based-on-the | 1805.02142 | null | http://arxiv.org/abs/1805.02142v1 | http://arxiv.org/pdf/1805.02142v1.pdf | An Image dehazing approach based on the airlight field estimation | This paper proposes a scheme for single image haze removal based on the
airlight field (ALF) estimation. Conventional image dehazing methods which are
based on a physical model generally take the global atmospheric light as a
constant. However, the constant-airlight assumption may be unsuitable for
images with large sky regions, which causes unacceptable brightness imbalance
and color distortion in recovery images. This paper models the atmospheric
light as a field function, and presents a maximum a-priori (MAP) method for
jointly estimating the airlight field, the transmission rate and the haze free
image. We also introduce a valid haze-level prior for effective estimate of
transmission. Evaluation on real world images shows that the proposed approach
outperforms existing methods in single image dehazing, especially when the
large sky region is included. | ['Yu-Jin Zhang', 'Yongbin Gao', 'Lijun Zhang'] | 2018-05-06 | null | null | null | null | ['single-image-haze-removal'] | ['computer-vision'] | [ 3.02934200e-01 -5.15124738e-01 5.81380844e-01 -2.28831589e-01
-1.91608012e-01 -9.01302472e-02 4.40444559e-01 -4.99810904e-01
-2.46129215e-01 7.52393603e-01 -1.51932850e-01 -3.65216017e-01
-3.24717648e-02 -9.11714792e-01 -4.76684988e-01 -1.40858972e+00
1.18725553e-01 -2.27163389e-01 5.18915415e-01 -3.17497402e-01
1.97249562e-01 1.73866630e-01 -1.59310973e+00 -1.98285326e-01
1.21004403e+00 6.79034889e-01 5.66817760e-01 8.24714363e-01
-7.39815533e-02 8.19063246e-01 -9.68431652e-01 -7.74569586e-02
6.73412085e-01 -6.16439819e-01 -1.07338339e-01 3.21986675e-01
7.80950367e-01 -8.25421929e-01 -3.35987717e-01 1.40936208e+00
2.66092688e-01 1.80853769e-01 5.06841660e-01 -1.04053366e+00
-6.25351131e-01 -3.86732012e-01 -7.23013759e-01 3.77427667e-01
-1.15650803e-01 1.98378950e-01 1.40938833e-01 -7.93043554e-01
1.27077207e-01 8.74945819e-01 4.83130455e-01 1.23797819e-01
-6.76001966e-01 -7.12566018e-01 -1.15259819e-01 2.97879159e-01
-1.65395010e+00 -3.35492402e-01 6.00765467e-01 -7.51657709e-02
6.22146547e-01 4.55676943e-01 6.76741421e-01 -2.90299535e-01
7.63745904e-01 3.16252142e-01 1.63010848e+00 -8.58530283e-01
-1.48101570e-02 3.29173893e-01 1.62600279e-02 5.63097656e-01
5.72044849e-01 2.56305784e-01 -2.96124995e-01 7.53395855e-02
7.02010274e-01 9.81184766e-02 -6.77839875e-01 1.49460986e-01
-7.16161847e-01 3.65056276e-01 3.60528678e-01 4.16309200e-02
-4.13769722e-01 5.47170825e-02 -6.79492414e-01 4.23782617e-01
7.68869400e-01 1.07240006e-01 1.17980413e-01 5.74116111e-01
-1.21387160e+00 2.38291383e-01 4.62790132e-01 9.98664737e-01
1.00457382e+00 4.58649069e-01 1.62098750e-01 6.99519157e-01
5.99054396e-01 1.43983674e+00 -7.04813451e-02 -9.47777927e-01
2.18419105e-01 -5.59595712e-02 6.80989981e-01 -7.91234970e-01
2.57770360e-01 -3.64951104e-01 -7.58292019e-01 8.74227464e-01
1.10187486e-01 -1.30934864e-01 -1.35731447e+00 1.05809188e+00
4.35310185e-01 5.07160783e-01 2.09941581e-01 1.11868620e+00
5.91960192e-01 1.27014720e+00 -5.43191433e-01 -6.91938043e-01
1.00800252e+00 -1.07527852e+00 -1.31763923e+00 -1.02164112e-01
4.79999222e-02 -1.24354517e+00 3.22875589e-01 8.07311654e-01
-1.38738656e+00 -4.82192814e-01 -1.28677356e+00 1.26080483e-01
-1.75699949e-01 -9.06634405e-02 3.28739554e-01 1.06516898e+00
-1.49469101e+00 -1.15884483e-01 -3.81274045e-01 -6.96224067e-03
-2.86177695e-01 2.54102260e-01 1.35149956e-01 -6.19047403e-01
-1.13668025e+00 1.19869256e+00 1.45050976e-02 4.74374175e-01
-9.88941312e-01 -4.63161319e-01 -6.19619131e-01 -2.64124691e-01
9.35226306e-03 -6.83699310e-01 8.13581109e-01 -8.60100925e-01
-1.52442384e+00 4.40790445e-01 -3.13862503e-01 -4.25728828e-01
7.96036050e-02 -4.32326376e-01 -8.66104782e-01 4.05021012e-01
-4.38554257e-01 2.02743009e-01 1.37471306e+00 -1.85320747e+00
-6.63580239e-01 -3.96768600e-02 5.12055233e-02 5.05584717e-01
9.43661183e-02 1.47135362e-01 -5.08120537e-01 -4.88737881e-01
6.28676936e-02 -7.58573711e-01 -1.33027673e-01 1.56405747e-01
-2.39757165e-01 5.24509251e-01 1.05165482e+00 -8.16530704e-01
1.18237054e+00 -2.13999271e+00 -5.05650401e-01 5.03447019e-02
1.64362118e-01 5.16145229e-01 7.81110972e-02 3.92143101e-01
2.06392333e-01 -4.15071189e-01 -3.44647735e-01 -2.75856107e-01
-5.20906568e-01 2.13827908e-01 -3.31402212e-01 1.00173008e+00
-2.94732124e-01 1.08069688e-01 -6.94162607e-01 -4.78605002e-01
6.25427067e-01 8.61891210e-01 -1.56484440e-01 4.69713092e-01
2.01829493e-01 2.58898467e-01 -1.04014054e-02 4.95300382e-01
1.75165284e+00 2.96725780e-01 -4.03591901e-01 -8.37982446e-02
-7.33065307e-01 -3.11202437e-01 -1.14203286e+00 7.29314029e-01
-7.53713906e-01 8.32588732e-01 6.92821085e-01 -2.37061158e-01
8.80091846e-01 3.72778654e-01 1.86178565e-01 -5.80359280e-01
4.64524217e-02 1.53640658e-01 -2.39703655e-01 -5.96991420e-01
6.21843815e-01 -5.74231267e-01 9.04861271e-01 1.85297489e-01
-3.51752847e-01 -6.34235501e-01 -2.05541253e-01 1.39293790e-01
6.51845396e-01 -6.19609281e-02 -5.38812317e-02 -4.45223212e-01
5.79158545e-01 -1.63181216e-01 2.69324005e-01 6.25500917e-01
-2.05998838e-01 6.88246667e-01 -6.02068067e-01 -3.56130004e-01
-8.73226404e-01 -1.14754176e+00 -4.66189712e-01 4.02976483e-01
9.93304968e-01 -5.01419008e-02 -7.96600878e-01 1.47642046e-01
-3.89467567e-01 1.00187182e+00 -1.72556877e-01 -2.83829011e-02
-3.04969490e-01 -1.24070084e+00 -9.29933265e-02 -5.72220862e-01
1.07080412e+00 -4.58468199e-01 -4.83374029e-01 -1.48038670e-01
-4.84633356e-01 -1.22655189e+00 -2.06915915e-01 -3.14316839e-01
-6.72230542e-01 -8.43325853e-01 -8.22171628e-01 -5.22855401e-01
9.38906848e-01 1.41594720e+00 9.13167238e-01 5.22850156e-01
-2.47827619e-01 5.04734933e-01 -3.33214611e-01 -8.66343558e-01
-4.81989831e-01 -1.11667204e+00 -1.04422912e-01 4.37724918e-01
2.42716089e-01 -2.24179775e-01 -1.07707119e+00 4.95959252e-01
-9.75361288e-01 2.51730353e-01 3.60051960e-01 5.38275659e-01
3.24186057e-01 1.15038502e+00 -1.48941293e-01 -4.96281177e-01
1.50358111e-01 -4.96331304e-01 -1.12590861e+00 1.26397371e-01
-9.07589316e-01 -5.31836212e-01 1.38053983e-01 1.08593412e-01
-1.95017576e+00 -4.38205719e-01 3.40811789e-01 -3.83560359e-01
-2.04892501e-01 -7.10027739e-02 -4.44143042e-02 -5.64627588e-01
3.39130908e-01 5.10637760e-01 -1.44177020e-01 -3.42205703e-01
1.13207854e-01 6.81742311e-01 6.60512686e-01 1.49612194e-02
1.51317418e+00 9.74631011e-01 3.14414471e-01 -1.51566803e+00
-5.44099927e-01 -9.48521674e-01 -3.04562122e-01 -6.34066105e-01
1.09541929e+00 -1.35522449e+00 -2.91553974e-01 9.32957768e-01
-1.00184548e+00 -2.97257781e-01 2.42068067e-01 9.45135415e-01
-1.79243684e-02 4.97624159e-01 -4.56958354e-01 -1.50555849e+00
-2.03552008e-01 -9.33127344e-01 9.63028729e-01 1.99408025e-01
7.50828564e-01 -1.18808818e+00 2.48994064e-02 5.42753398e-01
8.00959647e-01 3.71515602e-02 4.08160925e-01 7.64869630e-01
-1.44884419e+00 -1.34262577e-01 -3.73737663e-01 5.69892108e-01
3.42752516e-01 1.97592467e-01 -1.24205196e+00 -4.24865901e-01
6.07313752e-01 3.05693448e-01 9.15955245e-01 7.20117450e-01
5.45572758e-01 -2.81200945e-01 -5.95359467e-02 1.08909047e+00
2.14096022e+00 2.09613472e-01 1.12064874e+00 4.43979442e-01
5.07186115e-01 3.92162263e-01 1.01866710e+00 4.90829378e-01
5.50565533e-02 5.75064123e-01 9.46102917e-01 -6.85325444e-01
-4.47985321e-01 5.48219025e-01 4.19874132e-01 8.27990770e-01
-3.57678115e-01 -8.70546877e-01 -4.23083454e-01 6.18978262e-01
-1.16036451e+00 -9.68426347e-01 -7.40365684e-01 2.50171614e+00
7.03117788e-01 -2.05670133e-01 -5.10012627e-01 -6.16146028e-02
7.16403604e-01 1.48519978e-01 9.47911292e-02 -2.29813963e-01
-2.89022893e-01 6.01426102e-02 1.22704923e+00 1.14891338e+00
-8.83511484e-01 6.61909580e-01 7.08944750e+00 4.47367966e-01
-9.61638093e-01 2.57810473e-01 8.56777430e-02 1.65901650e-02
-4.98275191e-01 1.28665105e-01 -8.77515256e-01 6.57263935e-01
9.01575327e-01 9.73656550e-02 3.84243250e-01 -1.26314566e-01
6.45290911e-01 -7.72337914e-01 3.99427824e-02 9.36240494e-01
1.81710020e-01 -6.97714508e-01 1.56633243e-01 2.95030594e-01
9.23406839e-01 8.14252794e-02 3.07423323e-02 -4.32205379e-01
3.51468056e-01 -5.21407068e-01 3.85451764e-01 7.83477724e-01
8.26256096e-01 -4.89799112e-01 8.10589075e-01 3.44514698e-01
-7.94756830e-01 2.35198319e-01 -6.43687189e-01 1.56430416e-02
4.47429985e-01 9.35818136e-01 -5.42901456e-01 7.38350809e-01
9.06053782e-01 3.01953673e-01 -3.77402991e-01 1.46482372e+00
-3.53829950e-01 7.82442331e-01 -4.89657134e-01 5.01392603e-01
9.00875703e-02 -7.79790878e-01 7.31034219e-01 1.06157410e+00
5.97376406e-01 6.04494393e-01 -1.40398920e-01 6.23602033e-01
4.47108388e-01 -2.30748922e-01 -5.36244154e-01 5.23820758e-01
8.74768104e-03 9.82076764e-01 -4.68507051e-01 -3.25894296e-01
-6.55424774e-01 1.07100308e+00 -9.29596484e-01 9.57713842e-01
-7.60142386e-01 -5.95327973e-01 5.22607684e-01 2.98164278e-01
9.61565301e-02 -3.14766765e-01 5.87808229e-02 -9.31178629e-01
-1.17890559e-01 -4.68909949e-01 -1.16036013e-01 -1.47595012e+00
-6.94858789e-01 3.61896336e-01 3.90904844e-01 -1.44346058e+00
4.29301500e-01 -6.35095656e-01 -8.00632894e-01 1.32814395e+00
-2.28985906e+00 -1.15778267e+00 -7.57014990e-01 8.71774673e-01
5.45518637e-01 8.97448584e-02 4.65975732e-01 4.46734071e-01
-3.53571713e-01 -2.19356772e-02 6.89582527e-01 -5.39474845e-01
8.22557688e-01 -1.14613557e+00 -1.06438413e-01 1.59990907e+00
-3.00167352e-01 6.33043289e-01 1.36919606e+00 -5.76516151e-01
-1.34343433e+00 -9.18847799e-01 5.95334053e-01 -2.15608999e-01
2.51584142e-01 -8.06143582e-02 -9.07021821e-01 4.18130279e-01
8.66609573e-01 5.13414294e-02 4.69165981e-01 -7.25962877e-01
1.26280244e-02 -4.21215475e-01 -1.19429708e+00 3.69004846e-01
2.93549299e-01 -2.07948610e-01 -3.26370955e-01 5.91397524e-01
6.44463360e-01 -1.46077245e-01 -4.85507548e-01 3.00757527e-01
3.86776358e-01 -1.27402925e+00 1.02152586e+00 4.03183967e-01
-1.14809528e-01 -9.21881616e-01 -2.75269508e-01 -1.32953525e+00
-3.29518914e-01 -7.83491373e-01 5.34876361e-02 1.03253269e+00
6.70386618e-03 -7.82563806e-01 3.04481000e-01 2.58954704e-01
-2.12474629e-01 4.31764387e-02 -5.35195708e-01 -8.74711394e-01
-4.38831240e-01 1.18921734e-02 4.70083356e-01 9.15479958e-01
-7.00794399e-01 -7.87983388e-02 -9.54364181e-01 1.08653104e+00
1.38802266e+00 1.14683360e-01 9.19466913e-01 -1.01748490e+00
-1.74666002e-01 2.06364289e-01 -2.85579294e-01 -9.06817198e-01
-3.22823048e-01 -5.68122379e-02 5.66082418e-01 -1.72730184e+00
2.74799615e-02 -2.86244720e-01 -3.83248776e-01 -1.69342011e-01
-3.61920387e-01 5.40040672e-01 -1.05885610e-01 4.65466499e-01
-1.72503829e-01 3.75528038e-01 1.39527428e+00 -1.29762888e-01
-7.74569437e-02 2.40213275e-01 -7.38368034e-02 4.97296929e-01
7.25037396e-01 -3.89610380e-01 -8.15558374e-01 -8.28865886e-01
1.28683105e-01 -1.39417171e-01 3.55258971e-01 -1.21271217e+00
2.75718093e-01 -4.31964070e-01 2.20415607e-01 -6.74719274e-01
6.73391759e-01 -1.29536331e+00 3.53403717e-01 3.52749109e-01
4.58663851e-01 -3.13501656e-01 8.04118738e-02 7.41403162e-01
-4.61468071e-01 -3.22733313e-01 1.11050141e+00 -2.02013180e-01
-6.84233963e-01 2.72798061e-01 -5.37742972e-01 -5.43516457e-01
8.77518952e-01 -6.30835474e-01 -7.16290593e-01 -7.88247824e-01
-1.04266129e-01 -1.08027503e-01 6.80620432e-01 -5.46674430e-02
8.86335194e-01 -8.25093567e-01 -8.66236746e-01 4.36413080e-01
6.32376447e-02 -1.46087706e-01 3.83931398e-01 9.23000753e-01
-1.24005067e+00 2.42084011e-01 9.63631570e-02 -4.05470848e-01
-1.52227914e+00 5.60353577e-01 6.08709097e-01 3.69240075e-01
-6.77313149e-01 7.82144964e-01 7.75812209e-01 3.24331671e-01
-1.63777664e-01 -1.22762686e-02 -3.48095596e-02 -6.92461550e-01
9.78755832e-01 4.65942711e-01 -7.43300021e-02 -9.23446774e-01
8.24012086e-02 8.22784722e-01 3.22110981e-01 -5.61612211e-02
1.02898157e+00 -1.02009606e+00 -5.96288919e-01 3.15995932e-01
7.84389853e-01 4.74966556e-01 -1.16240454e+00 -2.08236009e-01
-9.13501322e-01 -1.14704049e+00 8.28251600e-01 -5.98592401e-01
-1.01953578e+00 1.00001228e+00 8.31430495e-01 3.01585823e-01
1.70804250e+00 -6.23440266e-01 7.81656265e-01 2.04532430e-01
4.60629404e-01 -8.77367198e-01 -3.68014574e-01 1.34704068e-01
5.57074487e-01 -1.07081354e+00 5.40879130e-01 -7.17206180e-01
-5.40799052e-02 9.50232446e-01 4.44002032e-01 1.57588705e-01
1.17923021e+00 2.51967967e-01 5.38850963e-01 -2.56713569e-01
-4.90393966e-01 -3.04355055e-01 5.35295866e-02 6.74519181e-01
9.50643197e-02 -1.77178025e-01 6.10719025e-02 -6.25893354e-01
2.59799808e-01 -1.39395505e-01 1.09940898e+00 8.47319424e-01
-1.14044392e+00 -6.66157305e-01 -1.17661047e+00 -6.82062358e-02
-7.02385426e-01 -3.82593781e-01 2.20605031e-01 5.79490542e-01
4.62427974e-01 1.49166536e+00 9.02180001e-02 9.75295752e-02
-6.05005398e-02 -3.74028474e-01 4.66307074e-01 -3.36446941e-01
1.07009239e-01 4.63578969e-01 -1.70437202e-01 -5.96010648e-02
-1.02935016e+00 -8.75333026e-02 -8.29609275e-01 -4.53423947e-01
-9.89756823e-01 3.30051988e-01 7.64832556e-01 5.62496066e-01
-1.79570213e-01 2.84490645e-01 8.61801922e-01 -7.58668780e-01
1.30943090e-01 -7.55514622e-01 -1.40484917e+00 1.01032779e-01
1.02806199e+00 -7.16648817e-01 -9.19290781e-01 4.52758163e-01] | [10.85962200164795, -3.184847116470337] |
834d9bc4-2d1e-45eb-be76-a05ee24c2594 | doubly-stochastic-graph-based-non | 2306.06119 | null | https://arxiv.org/abs/2306.06119v1 | https://arxiv.org/pdf/2306.06119v1.pdf | Doubly Stochastic Graph-based Non-autoregressive Reaction Prediction | Organic reaction prediction is a critical task in drug discovery. Recently, researchers have achieved non-autoregressive reaction prediction by modeling the redistribution of electrons, resulting in state-of-the-art top-1 accuracy, and enabling parallel sampling. However, the current non-autoregressive decoder does not satisfy two essential rules of electron redistribution modeling simultaneously: the electron-counting rule and the symmetry rule. This violation of the physical constraints of chemical reactions impairs model performance. In this work, we propose a new framework called that combines two doubly stochastic self-attention mappings to obtain electron redistribution predictions that follow both constraints. We further extend our solution to a general multi-head attention mechanism with augmented constraints. To achieve this, we apply Sinkhorn's algorithm to iteratively update self-attention mappings, which imposes doubly conservative constraints as additional informative priors on electron redistribution modeling. We theoretically demonstrate that our can simultaneously satisfy both rules, which the current decoder mechanism cannot do. Empirical results show that our approach consistently improves the predictive performance of non-autoregressive models and does not bring an unbearable additional computational cost. | ['Irwin King', 'Yang Yu', 'Peilin Zhao', 'Ziqiao Meng'] | 2023-06-05 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 5.64036131e-01 1.40900716e-01 -4.02409345e-01 -2.54316151e-01
-9.26876783e-01 -2.55063742e-01 6.44431889e-01 1.70126766e-01
-2.53379315e-01 1.05336964e+00 2.39275515e-01 -6.23800933e-01
1.26973614e-01 -6.09871507e-01 -9.76066411e-01 -8.76421690e-01
5.04387736e-01 7.25410521e-01 1.55782580e-01 -5.58841154e-02
4.51367170e-01 5.58652043e-01 -1.28554857e+00 3.91308039e-01
1.00837326e+00 7.29199886e-01 3.10323805e-01 5.87837815e-01
3.16273533e-02 1.11561370e+00 8.48039985e-03 -4.40218687e-01
-1.08565256e-01 -6.36294901e-01 -8.16198111e-01 -6.20387971e-01
-4.36789542e-02 6.68528490e-03 -2.92961389e-01 1.01369834e+00
6.01827919e-01 2.21654192e-01 1.03158975e+00 -7.71792591e-01
-9.94212687e-01 6.27746284e-01 -5.39666951e-01 2.83212394e-01
7.32851997e-02 1.41000345e-01 1.18601561e+00 -9.09888625e-01
5.34469008e-01 1.09076691e+00 7.58461177e-01 7.71071911e-01
-1.20647383e+00 -8.32471788e-01 3.18961471e-01 4.60337192e-01
-1.34580123e+00 -6.06756806e-01 8.33098352e-01 -3.33176583e-01
1.58137143e+00 3.86823356e-01 4.27175015e-01 1.22153509e+00
5.67455411e-01 6.45417213e-01 6.35443628e-01 -2.98366159e-01
4.52200681e-01 3.40332650e-02 2.85222799e-01 4.74537522e-01
2.62961924e-01 7.88021609e-02 -6.53498888e-01 -2.21550912e-01
5.40729880e-01 -3.74949127e-02 -2.36884147e-01 -1.37539238e-01
-8.43399465e-01 8.80670846e-01 2.57426679e-01 -1.32893054e-02
-7.02053010e-01 2.94970274e-01 2.76917219e-01 -2.31953323e-01
6.55241251e-01 7.64415145e-01 -6.81323707e-01 -1.58856943e-01
-7.68076599e-01 3.56059819e-02 4.80507582e-01 8.46055984e-01
5.88095546e-01 1.15103535e-01 -8.95799324e-02 5.04615724e-01
4.78875369e-01 9.70755368e-02 3.70134443e-01 -6.66096926e-01
1.50015920e-01 8.58159587e-02 3.49931687e-01 -4.82432246e-01
-5.46996057e-01 -5.90115547e-01 -8.96968663e-01 -1.64407507e-01
1.87555715e-01 6.87898844e-02 -9.43811655e-01 1.85962117e+00
2.42702782e-01 1.28226027e-01 1.60828352e-01 6.09987974e-01
6.03584707e-01 9.99734163e-01 5.89949071e-01 -6.29186928e-01
1.21955121e+00 -9.76839483e-01 -9.56660628e-01 -1.48523435e-01
7.47391164e-01 -5.17830610e-01 9.25241828e-01 5.83768010e-01
-1.42017138e+00 -2.88182825e-01 -1.17999017e+00 -5.26921749e-01
-2.17068136e-01 -1.77844808e-01 1.18705177e+00 6.73355222e-01
-6.64034963e-01 7.52674043e-01 -1.07207608e+00 7.30259269e-02
3.30078483e-01 6.48433506e-01 -9.05590430e-02 1.92233339e-01
-1.27149129e+00 1.11759937e+00 1.78724751e-01 -8.53534043e-02
-7.46161520e-01 -1.25174797e+00 -6.69110954e-01 3.84197056e-01
3.85404766e-01 -8.43138814e-01 1.32583296e+00 -7.25769043e-01
-1.81172264e+00 4.44632947e-01 -6.35807812e-01 -5.32160938e-01
5.10802507e-01 -1.48122713e-01 -1.71753749e-01 -2.09039390e-01
-1.59781516e-01 6.20000601e-01 2.37969935e-01 -1.15896463e+00
-1.44414648e-01 -3.73879790e-01 -2.77144223e-01 2.58884817e-01
-1.98246747e-01 2.41486490e-01 -3.39790195e-01 -4.77986366e-01
1.94696225e-02 -8.26516211e-01 -3.87774825e-01 -4.62718248e-01
-6.59066558e-01 -2.91239649e-01 1.35215908e-01 -3.63076299e-01
1.21905303e+00 -1.77173042e+00 3.15076411e-01 1.84747398e-01
1.10126704e-01 -3.01186778e-02 -4.89112400e-02 3.05459052e-01
-5.67826211e-01 2.89854646e-01 -3.16854089e-01 -4.74680692e-01
-1.11496383e-02 -7.67939612e-02 -3.25382292e-01 5.02115309e-01
2.49254107e-01 8.63637865e-01 -8.17413211e-01 -1.62084758e-01
1.35815777e-02 8.01814198e-01 -9.77363229e-01 -1.23855747e-01
-3.91091585e-01 3.46509308e-01 -4.64651972e-01 3.83375764e-01
4.84518439e-01 -6.82916343e-01 4.18368667e-01 -4.37974989e-01
-3.76880884e-01 7.59980500e-01 -7.98848212e-01 1.34460807e+00
-2.52951503e-01 7.11517408e-02 -3.01090539e-01 -8.95045280e-01
6.98281765e-01 2.27558807e-01 6.07701361e-01 -6.35163724e-01
1.51091024e-01 8.52109641e-02 3.14464867e-01 -1.41698182e-01
4.87203330e-01 -5.16087115e-01 1.78286746e-01 2.53492475e-01
-1.19270235e-01 1.07048362e-01 -3.36560048e-02 -2.02789605e-02
8.33976030e-01 9.49903727e-02 5.51903486e-01 -4.88816619e-01
4.37505245e-01 -2.24959508e-01 7.60628939e-01 7.93954194e-01
1.97288334e-01 5.03405273e-01 6.01703405e-01 -4.72965091e-01
-1.15646565e+00 -5.77465236e-01 -2.42037937e-01 1.16601813e+00
-2.20897987e-01 -6.35120749e-01 -6.52524769e-01 -4.44859535e-01
-1.45382300e-01 9.80580926e-01 -7.30163872e-01 -4.79521275e-01
-3.77290905e-01 -1.45290935e+00 3.18678200e-01 7.28945434e-01
-9.47324038e-02 -7.51946032e-01 -2.19483763e-01 3.57760668e-01
1.28623247e-01 -6.18610501e-01 -5.97666025e-01 7.38630533e-01
-3.75690460e-01 -9.65605915e-01 -4.69847202e-01 -3.70299339e-01
4.62371349e-01 -2.93957051e-02 8.87994945e-01 -2.71362752e-01
-1.57615200e-01 -1.94043815e-01 2.14021683e-01 -6.37053072e-01
-5.18859267e-01 2.22997889e-01 4.76205945e-02 -1.33313209e-01
5.54911435e-01 -4.35329825e-01 -4.52025265e-01 1.06377833e-01
-6.35250032e-01 3.35097641e-01 5.34001052e-01 8.10404181e-01
9.80317116e-01 -2.44122684e-01 7.21687436e-01 -1.16107321e+00
3.72268379e-01 -5.23735106e-01 -6.00023985e-01 2.36424655e-01
-7.57298529e-01 5.75442731e-01 7.20023990e-01 -3.36427569e-01
-1.19120109e+00 3.67958874e-01 -5.25516212e-01 -3.51266265e-01
2.56833762e-01 6.07827961e-01 -5.29434025e-01 -5.50464988e-02
4.65245932e-01 3.41866434e-01 -3.25895637e-01 -5.03285170e-01
1.39214173e-01 4.39572513e-01 2.94034660e-01 -5.67631781e-01
1.77619517e-01 2.06087574e-01 3.20398152e-01 -6.61308467e-01
-8.18736255e-01 -2.23779887e-01 -2.53661871e-01 3.02709371e-01
1.04148626e+00 -8.77131402e-01 -1.19128871e+00 2.14735493e-01
-1.32862735e+00 -3.14567089e-01 5.46453670e-02 6.16481185e-01
-7.19780266e-01 5.67911446e-01 -9.53171074e-01 -9.83438432e-01
-5.38771629e-01 -1.45383048e+00 7.96309054e-01 -3.03761929e-01
-3.15884382e-01 -8.01927090e-01 4.07890268e-02 1.86494514e-01
2.89858609e-01 -1.87443912e-01 1.20113385e+00 -7.08129883e-01
-5.23319542e-01 6.52620196e-02 -1.06795318e-01 -2.15621710e-01
-5.59740290e-02 -6.65912628e-02 -8.79460037e-01 6.44048601e-02
6.86830506e-02 1.24488270e-03 1.19844174e+00 6.75403655e-01
1.55523944e+00 -9.51746628e-02 -4.98301268e-01 6.12083077e-01
1.05699027e+00 6.42515838e-01 8.19816172e-01 3.97164784e-02
9.26691294e-01 2.81504065e-01 2.59601682e-01 5.84938049e-01
3.95830244e-01 6.71294451e-01 4.28689301e-01 -9.18442849e-03
1.62096068e-01 -4.10677880e-01 2.98244447e-01 7.28942990e-01
-3.03579599e-01 -6.14218891e-01 -9.35169935e-01 1.00104213e-01
-1.81729317e+00 -1.05996823e+00 -3.51771951e-01 2.17114258e+00
1.13945031e+00 1.55733883e-01 -8.93803537e-02 -2.07437389e-02
5.76968849e-01 -3.72291468e-02 -9.95122135e-01 -4.08021063e-01
7.00463355e-02 4.29581642e-01 8.69887054e-01 8.74979913e-01
-9.43733275e-01 1.06175494e+00 7.17061853e+00 6.70982957e-01
-1.06002152e+00 2.46251196e-01 8.42196107e-01 -2.53101885e-01
-7.13548541e-01 -2.18902737e-01 -1.24290991e+00 5.66527605e-01
1.22319925e+00 -2.02947289e-01 3.00939679e-01 7.29055166e-01
3.41766298e-01 3.65551710e-01 -1.31261861e+00 1.01873899e+00
9.00027156e-02 -1.68719542e+00 3.09637457e-01 2.15244323e-01
5.37675679e-01 -5.36017045e-02 1.96415827e-01 2.40806922e-01
3.95967871e-01 -1.04571378e+00 5.91838539e-01 5.22117078e-01
7.68841743e-01 -9.77430880e-01 3.44978839e-01 1.83819562e-01
-8.35843801e-01 -1.81155484e-02 -4.33286279e-01 8.13513175e-02
2.04277620e-01 5.93492389e-01 -6.06297374e-01 4.07408983e-01
1.38262153e-01 8.39328706e-01 1.75075550e-02 7.68684685e-01
-6.66710734e-02 5.06112218e-01 -2.80094534e-01 -1.61269113e-01
1.29115060e-01 -1.59667104e-01 2.55721986e-01 1.00241339e+00
4.27616298e-01 3.09827089e-01 -7.77208284e-02 9.30682540e-01
-2.97549069e-01 1.20727278e-01 -2.60675400e-01 -5.86288236e-02
4.37384874e-01 5.95965564e-01 -5.10807335e-01 -4.26649392e-01
-3.88117164e-01 8.78193319e-01 2.37055123e-01 3.24836254e-01
-1.16732609e+00 -4.95526381e-03 8.73185694e-01 1.44068286e-01
1.79350778e-01 3.16512771e-02 -4.39498007e-01 -1.11402321e+00
-3.56227845e-01 -6.92909658e-01 2.41629690e-01 -7.59484053e-01
-1.12017345e+00 2.91671216e-01 -3.70769590e-01 -4.90208596e-01
7.25931078e-02 -7.37867177e-01 -3.92403156e-01 9.95233357e-01
-1.77676177e+00 -9.76914406e-01 3.22248548e-01 2.17288166e-01
6.77718699e-01 1.28745854e-01 8.89464617e-01 5.74737608e-01
-9.53183711e-01 6.25816166e-01 1.39502093e-01 -6.98514044e-01
8.24295878e-01 -1.25531149e+00 4.94311333e-01 6.17611170e-01
-3.06591213e-01 7.50149429e-01 9.92645562e-01 -8.43129396e-01
-1.72467291e+00 -1.15515077e+00 9.19866979e-01 -3.48073483e-01
5.21612108e-01 -3.98363471e-01 -1.16247630e+00 7.65000641e-01
-6.62152842e-02 -2.68481702e-01 9.95452046e-01 3.06441009e-01
-2.73874819e-01 2.21562430e-01 -7.92367280e-01 5.29568434e-01
1.04271710e+00 -5.17639458e-01 -2.44996622e-01 7.84309208e-01
7.73008466e-01 -2.81546324e-01 -8.57471824e-01 3.75907362e-01
4.08629894e-01 -5.38220406e-01 1.14874029e+00 -1.22032428e+00
5.96858621e-01 -1.58055127e-01 -1.64949819e-01 -1.06378019e+00
-8.35663021e-01 -7.48348951e-01 -3.40397745e-01 7.74388313e-01
8.32997561e-01 -7.69820333e-01 9.05334353e-01 1.00741529e+00
-7.81106889e-01 -8.19140792e-01 -8.08827877e-01 -5.16340733e-01
3.54127020e-01 -4.99383062e-01 6.03247285e-01 9.72698450e-01
3.78017902e-01 6.99205697e-01 -6.79887533e-01 1.53202370e-01
3.12434644e-01 -3.39536458e-01 1.72735095e-01 -1.25047088e+00
-5.58251023e-01 -5.94179213e-01 2.21699830e-02 -1.23565149e+00
4.68282819e-01 -9.72380698e-01 -1.58997178e-02 -1.31360471e+00
5.96373677e-01 -1.68917686e-01 -6.15220308e-01 5.59626579e-01
-2.95109421e-01 -1.48574337e-02 -1.45747423e-01 1.06283151e-01
-6.12436533e-01 9.05360341e-01 9.41144526e-01 -1.19162947e-01
-1.55709624e-01 -2.38747537e-01 -1.10180366e+00 5.95853388e-01
7.20144391e-01 -4.03656989e-01 -6.05766356e-01 -2.05156177e-01
7.48035252e-01 7.22662956e-02 -8.86660814e-02 -7.56157339e-01
3.22274595e-01 -1.02025457e-01 4.48463082e-01 -6.10654414e-01
5.65560699e-01 -4.42810476e-01 5.27483165e-01 5.06211162e-01
-5.95305860e-01 2.58763395e-02 2.17375010e-01 7.08521545e-01
3.04102242e-01 -9.92190912e-02 9.24488664e-01 -1.67814404e-01
-5.65173253e-02 4.36850667e-01 -6.17961764e-01 -3.70053411e-01
9.33008194e-01 7.94057846e-02 -2.32722878e-01 -4.66599092e-02
-7.95988083e-01 7.40712658e-02 4.80063975e-01 1.16934739e-01
2.69565940e-01 -9.76425111e-01 -4.55976158e-01 8.65248069e-02
-2.30981484e-01 -2.14102492e-01 4.25370097e-01 8.99096429e-01
-2.71191925e-01 7.54935324e-01 1.63468182e-01 -2.34706402e-01
-1.21378934e+00 9.39582765e-01 4.12596911e-01 -3.89186263e-01
-3.49490255e-01 7.28607476e-01 5.31861722e-01 -4.79141958e-02
-1.20674938e-01 -3.84104699e-01 -1.01066642e-01 3.15356888e-02
6.49271250e-01 1.78381845e-01 3.24783474e-01 -3.63492817e-01
-3.14627051e-01 3.24147820e-01 -5.42452931e-01 3.21073443e-01
1.46662664e+00 1.94356397e-01 8.97006989e-02 3.03953767e-01
1.03933692e+00 -1.19851775e-01 -1.20571256e+00 1.06178455e-01
-2.57024437e-01 -1.28107583e-02 3.99194181e-01 -8.32435787e-01
-6.30589843e-01 9.83583868e-01 2.19959363e-01 -7.71264583e-02
7.50213623e-01 -1.60606533e-01 5.98533273e-01 5.99144161e-01
4.18655202e-02 -8.29610348e-01 -3.09912115e-01 5.57230294e-01
6.16775393e-01 -1.01640153e+00 5.45107089e-02 -7.17085183e-01
-6.21652782e-01 7.62320697e-01 4.62028325e-01 3.55555594e-01
4.35322404e-01 2.67079800e-01 -6.96747303e-01 -1.43366963e-01
-1.17093742e+00 1.13578573e-01 6.10880107e-02 2.85322338e-01
8.22829247e-01 -7.28207156e-02 -3.68926615e-01 8.95280600e-01
1.61513351e-02 -4.18671258e-02 4.23236251e-01 5.62255681e-01
-4.83957350e-01 -1.10685301e+00 -1.18742265e-01 4.98983026e-01
-7.68801868e-01 -6.68273985e-01 -4.69755203e-01 4.83387083e-01
-3.72232705e-01 6.39654398e-01 7.27800429e-02 -6.12099357e-02
1.91488907e-01 3.56237531e-01 6.03568971e-01 -4.66296911e-01
-3.89594823e-01 1.34588391e-01 1.97774306e-01 -3.24918598e-01
-1.88974246e-01 -7.21885741e-01 -1.32875991e+00 -3.55548650e-01
-3.73826653e-01 4.37485218e-01 5.80397725e-01 8.23875368e-01
7.85662353e-01 9.58391190e-01 3.01430225e-01 -8.85155559e-01
-6.90858424e-01 -9.18125570e-01 -2.44091064e-01 -2.16048639e-02
1.04459777e-01 -7.56286025e-01 -1.45672262e-01 -7.76089355e-02] | [4.633363246917725, 5.986285209655762] |
316c09cc-3820-4430-9992-e61d59b544f4 | cross-speaker-emotion-transfer-based-on | 2110.04153 | null | https://arxiv.org/abs/2110.04153v2 | https://arxiv.org/pdf/2110.04153v2.pdf | Cross-speaker Emotion Transfer Based on Speaker Condition Layer Normalization and Semi-Supervised Training in Text-To-Speech | In expressive speech synthesis, there are high requirements for emotion interpretation. However, it is time-consuming to acquire emotional audio corpus for arbitrary speakers due to their deduction ability. In response to this problem, this paper proposes a cross-speaker emotion transfer method that can realize the transfer of emotions from source speaker to target speaker. A set of emotion tokens is firstly defined to represent various categories of emotions. They are trained to be highly correlated with corresponding emotions for controllable synthesis by cross-entropy loss and semi-supervised training strategy. Meanwhile, to eliminate the down-gradation to the timbre similarity from cross-speaker emotion transfer, speaker condition layer normalization is implemented to model speaker characteristics. Experimental results show that the proposed method outperforms the multi-reference based baseline in terms of timbre similarity, stability and emotion perceive evaluations. | ['Zejun Ma', 'Xiang Yin', 'Lin Wu', 'Junhui Zhang', 'Chenchang Xu', 'Junjie Pan', 'Pengfei Wu'] | 2021-10-08 | null | null | null | null | ['expressive-speech-synthesis'] | ['speech'] | [ 2.29055613e-01 -1.05737343e-01 2.48081848e-01 -7.49034882e-01
-9.43792939e-01 -3.14521432e-01 2.70276159e-01 -2.18432918e-01
-1.38456523e-01 6.29541218e-01 4.32369500e-01 3.32289100e-01
1.29679605e-01 -4.89254057e-01 -3.44409376e-01 -7.26556957e-01
1.46363884e-01 -1.92364126e-01 -3.49071950e-01 -2.83096522e-01
-1.25915647e-01 1.71921000e-01 -1.68255508e+00 4.46207583e-01
1.05286992e+00 1.30604780e+00 1.01977661e-01 5.36610246e-01
-1.23919547e-01 5.46242833e-01 -9.21674788e-01 -3.60424042e-01
-2.31155440e-01 -9.05666769e-01 -5.50491810e-01 1.45536348e-01
-1.95878506e-01 1.90944038e-02 1.38632655e-01 1.25788248e+00
1.00601983e+00 3.41528744e-01 7.85880029e-01 -1.41893268e+00
-5.77610075e-01 6.63329780e-01 -3.26519102e-01 -8.60987082e-02
3.83697450e-01 -2.67254025e-01 9.55752194e-01 -8.84377718e-01
1.86569795e-01 1.28989506e+00 5.61933517e-01 6.76829040e-01
-1.00144422e+00 -1.06623304e+00 7.45148808e-02 2.98482180e-01
-1.38002658e+00 -8.06102335e-01 1.39992476e+00 -2.69335926e-01
6.10377669e-01 4.37701166e-01 4.24332678e-01 1.19958949e+00
-2.11361349e-01 5.81009567e-01 1.08837378e+00 -4.44691658e-01
2.03949865e-02 6.12300992e-01 -2.15416968e-01 1.78062513e-01
-7.73027599e-01 7.32828602e-02 -7.29057610e-01 1.61747605e-01
2.42825836e-01 -5.88934124e-01 -6.24985754e-01 4.33485359e-02
-9.37903821e-01 7.10767448e-01 1.83627144e-01 5.18442750e-01
-4.12198812e-01 -3.12947452e-01 9.27650332e-01 6.08836055e-01
5.78693688e-01 2.75351793e-01 -3.71690899e-01 -3.28009516e-01
-7.85073221e-01 -3.69580507e-01 5.18774629e-01 8.47797811e-01
4.73002881e-01 5.44826269e-01 -1.35584816e-01 1.20525575e+00
2.59585828e-01 4.58307326e-01 8.07236016e-01 -4.22261983e-01
2.58606821e-01 2.23327041e-01 -8.55522379e-02 -1.13369954e+00
-8.42994079e-02 -3.95354211e-01 -1.02984309e+00 6.49258122e-03
-2.72840530e-01 -3.06698322e-01 -2.99202591e-01 1.99961114e+00
3.42674822e-01 2.04522058e-01 5.98208129e-01 9.83729184e-01
9.70350683e-01 1.16900539e+00 1.68636888e-01 -6.98141038e-01
1.25565398e+00 -8.89033020e-01 -1.18883395e+00 1.08486012e-01
1.63161784e-01 -9.50501025e-01 1.50098073e+00 3.18227232e-01
-9.56716359e-01 -8.92455161e-01 -9.56218302e-01 1.61237791e-01
-2.37086326e-01 4.00962263e-01 1.84131086e-01 7.75737345e-01
-5.52426636e-01 1.86601013e-01 -3.41749668e-01 6.98224306e-02
1.30433068e-01 2.44109467e-01 -2.54086673e-01 6.36621058e-01
-1.79236424e+00 6.97165966e-01 4.18444902e-01 2.14232221e-01
-6.69576228e-01 -6.40742600e-01 -7.62281954e-01 3.22481215e-01
-2.13462353e-01 -9.74346548e-02 1.24013007e+00 -1.65965736e+00
-2.26165223e+00 5.82135975e-01 -2.04080623e-02 -1.68983766e-03
1.62039354e-01 -4.44419049e-02 -1.09222543e+00 2.07543939e-01
-2.60769546e-01 5.24678171e-01 1.07126701e+00 -1.16443086e+00
-6.41283393e-01 -1.35418847e-02 -5.65881491e-01 5.67020237e-01
-8.25440288e-01 4.10656393e-01 -1.24840058e-01 -6.96511626e-01
-1.01631485e-01 -5.95016181e-01 3.40408176e-01 -4.88363922e-01
-2.91710764e-01 -3.21818113e-01 8.93775642e-01 -7.01999366e-01
1.22685623e+00 -2.38383818e+00 2.30847880e-01 1.46836728e-01
-3.60698372e-01 2.06088334e-01 -1.00418009e-01 8.88466462e-02
-3.54239464e-01 -1.66647792e-01 -1.27879411e-01 -2.18403369e-01
2.20254123e-01 -6.93227127e-02 -4.99285787e-01 2.38987476e-01
3.20951521e-01 3.29351574e-01 -6.55195415e-01 -7.84921467e-01
3.29607666e-01 9.18983161e-01 -3.50632071e-01 5.82832634e-01
1.38688698e-01 7.02997565e-01 -4.54423159e-01 3.05163652e-01
6.05890274e-01 4.67862606e-01 -1.78646162e-01 -5.46214163e-01
-4.44821455e-02 4.06957865e-01 -1.09988070e+00 1.56723058e+00
-8.64705026e-01 4.23948497e-01 2.67552167e-01 -1.07836962e+00
1.32462764e+00 9.89776790e-01 4.09549773e-01 -3.48414540e-01
5.86989701e-01 1.21866971e-01 6.20630011e-03 -5.87478817e-01
3.03389758e-01 -8.38765323e-01 -2.90263206e-01 1.37728825e-01
2.43180156e-01 -4.99900997e-01 -3.47635388e-01 -4.62216198e-01
2.70115554e-01 -8.43605176e-02 6.81776553e-02 -8.32018927e-02
1.02278221e+00 -5.39531767e-01 7.43430078e-01 -1.86569303e-01
-3.89371842e-01 3.53370339e-01 1.13399968e-01 2.91729063e-01
-6.55176163e-01 -9.40238655e-01 -1.75367415e-01 1.21903837e+00
4.09661271e-02 -8.28554109e-02 -7.67989695e-01 -3.57267797e-01
-4.92582679e-01 8.72427881e-01 -3.20823073e-01 -6.19973958e-01
-1.75758690e-01 -3.22135299e-01 8.70913327e-01 3.45416009e-01
5.02159238e-01 -1.35242414e+00 -4.26209062e-01 2.90846705e-01
-6.69785023e-01 -7.97242761e-01 -7.48193562e-01 1.30023792e-01
-3.65452856e-01 -2.49625638e-01 -7.50248849e-01 -1.06616533e+00
3.32311958e-01 -1.77258819e-01 7.09538639e-01 -5.78294098e-01
-1.80417448e-02 1.95559531e-01 -5.28621435e-01 -6.80547118e-01
-7.16048002e-01 -3.22371088e-02 2.00341031e-01 4.17155385e-01
4.13558364e-01 -7.78771341e-01 -3.19926858e-01 3.23860228e-01
-7.86464989e-01 -1.52867034e-01 4.47067261e-01 9.02766049e-01
5.20019472e-01 4.96640027e-01 1.19521654e+00 -7.48680606e-02
1.09673774e+00 -2.42101237e-01 -3.40627789e-01 2.68901080e-01
-2.76743650e-01 -1.57118574e-01 9.02852178e-01 -9.49052453e-01
-1.65924549e+00 -5.63654490e-02 -1.80795416e-01 -4.51395988e-01
-2.66348034e-01 4.89060551e-01 -7.65676320e-01 2.74127722e-01
4.04939115e-01 2.47530460e-01 7.78010637e-02 -8.66845474e-02
3.70194823e-01 1.31850195e+00 6.17686212e-01 -5.10224819e-01
5.52587807e-01 -1.93190977e-01 -4.55157369e-01 -9.30238724e-01
-6.22244656e-01 -2.42466062e-01 -2.68242747e-01 -4.60780054e-01
9.46308017e-01 -1.00605440e+00 -7.56446838e-01 3.90588045e-01
-1.17361760e+00 6.57548383e-02 -1.89608127e-01 9.91926074e-01
-6.98772609e-01 1.71553433e-01 -6.12269044e-01 -9.70856607e-01
-6.09865248e-01 -1.04293501e+00 7.93637812e-01 2.77895540e-01
-2.95418531e-01 -8.60889852e-01 -5.41242631e-03 1.69721097e-01
5.25424540e-01 1.24006651e-01 6.45585001e-01 -4.61610913e-01
3.31227809e-01 -9.19138864e-02 1.17165439e-01 8.68316650e-01
5.68479061e-01 -3.68192536e-03 -1.31508827e+00 -4.26865183e-02
4.09081876e-01 -6.86572552e-01 2.77325898e-01 1.45045251e-01
1.07096350e+00 -4.11270648e-01 9.08997580e-02 3.42876315e-01
9.00356412e-01 4.44919944e-01 6.10803187e-01 -1.53192073e-01
3.60668719e-01 9.76982415e-01 8.10412228e-01 3.43368262e-01
9.33332592e-02 5.30245125e-01 -1.18076071e-01 -1.46434218e-01
-2.74856836e-02 -2.51271874e-01 7.22394943e-01 1.55672193e+00
1.53781131e-01 -1.69171736e-01 -4.01173502e-01 3.72892231e-01
-1.32207632e+00 -1.14727366e+00 8.73488709e-02 2.03343391e+00
1.36929417e+00 -5.63655198e-02 -1.67846847e-02 4.36342090e-01
1.04227805e+00 -5.05439229e-02 -4.67205226e-01 -6.74213886e-01
-1.88630328e-01 3.00986040e-02 -3.28998804e-01 4.92857426e-01
-9.79687035e-01 8.80345345e-01 5.57174206e+00 1.09345150e+00
-1.59628117e+00 1.81377992e-01 5.78628004e-01 -1.94440298e-02
-1.88576579e-01 -5.03647685e-01 -3.91453028e-01 4.57495987e-01
1.14315534e+00 -5.33603728e-01 4.36422914e-01 7.35073268e-01
2.70576626e-01 5.01147091e-01 -9.41914499e-01 1.22916961e+00
2.31456622e-01 -3.55759412e-01 -2.08975762e-01 -3.99125159e-01
5.27655721e-01 -5.54693639e-01 2.20498517e-01 5.07111728e-01
-1.25653744e-01 -9.66403842e-01 6.98625863e-01 4.35082942e-01
1.04407847e+00 -1.13740039e+00 5.41562855e-01 1.92446738e-01
-1.34529924e+00 1.45759836e-01 -2.73706883e-01 1.06578767e-01
2.88126528e-01 3.57067615e-01 -7.82974064e-01 6.08405232e-01
5.27151883e-01 3.69651377e-01 1.35948464e-01 4.09714848e-01
-1.85802624e-01 7.50456810e-01 -1.66377470e-01 -8.98686573e-02
-5.95706813e-02 -3.15146446e-01 4.25252199e-01 1.29242349e+00
6.43238842e-01 4.93903868e-02 -5.38726673e-02 7.74654508e-01
-8.80887434e-02 6.01784229e-01 -3.79778862e-01 -8.14117491e-02
6.43003464e-01 1.28160846e+00 -1.91198125e-01 -3.64133775e-01
-1.95240185e-01 1.23609030e+00 -6.51452020e-02 3.23346227e-01
-1.16858613e+00 -1.02662408e+00 5.09853065e-01 -4.39752132e-01
1.57409087e-01 3.27602804e-01 -4.60075289e-02 -1.01896918e+00
-6.42821640e-02 -9.27917838e-01 9.24783126e-02 -9.53983247e-01
-1.37203741e+00 1.02676630e+00 -2.08601043e-01 -1.49056625e+00
-3.37369263e-01 -1.32752046e-01 -6.99522018e-01 1.07009482e+00
-1.54948330e+00 -1.04414940e+00 -1.69518530e-01 9.06566501e-01
5.34258485e-01 -2.01156288e-01 1.13726616e+00 5.95699549e-01
-4.57251221e-01 9.71331120e-01 -1.28878832e-01 -9.23910514e-02
1.00806761e+00 -8.55033994e-01 -4.09292936e-01 4.42429155e-01
-3.57833207e-01 3.27772260e-01 8.54113042e-01 -6.23070300e-02
-7.57425129e-01 -1.16159630e+00 8.91676962e-01 3.21377575e-01
5.90854704e-01 -3.20352942e-01 -1.03279340e+00 3.48876029e-01
4.95464921e-01 -1.76767245e-01 1.12876809e+00 4.43495773e-02
-2.91423291e-01 -5.10310352e-01 -1.15950143e+00 5.04377186e-01
4.70709473e-01 -8.37672949e-01 -8.39570940e-01 1.12911602e-02
8.02124739e-01 -1.79405250e-02 -1.23394513e+00 4.01862919e-01
5.11753321e-01 -8.31348300e-01 7.97851443e-01 -4.26058114e-01
3.23597521e-01 -3.35636288e-01 -2.32050881e-01 -1.71475446e+00
5.58560491e-02 -7.09512293e-01 5.33388197e-01 2.07422614e+00
4.34097618e-01 -5.00141799e-01 1.32601678e-01 3.76777291e-01
-3.31260949e-01 -2.56612897e-01 -9.59818065e-01 -8.71865332e-01
9.98289958e-02 -2.76422739e-01 8.36613119e-01 1.25323451e+00
5.77368557e-01 9.57705557e-01 -5.74202359e-01 2.05011591e-01
2.39048645e-01 1.90375194e-01 5.44479311e-01 -1.10872996e+00
-1.14502691e-01 -5.93504548e-01 -1.85927019e-01 -7.93251872e-01
6.37205362e-01 -8.19545031e-01 4.92402673e-01 -1.08245182e+00
-2.38066152e-01 -3.06544542e-01 -6.33552551e-01 1.14304096e-01
-2.09557220e-01 -3.44743542e-02 -1.00308783e-01 -1.53845116e-01
-3.37617993e-01 1.29751527e+00 1.18893731e+00 -8.62864405e-03
-4.84877616e-01 1.11553743e-01 -4.64791358e-01 6.50593877e-01
1.10455883e+00 -2.66707599e-01 -6.52090132e-01 2.96171084e-02
-2.61028379e-01 5.00441015e-01 -8.66528153e-02 -1.03392434e+00
-8.49309936e-02 -1.19195320e-01 6.14445694e-02 -5.06863534e-01
7.09205747e-01 -9.02787387e-01 1.85454972e-02 1.94536492e-01
-7.46169567e-01 -3.17900926e-01 1.90160170e-01 2.31661275e-01
-8.13149154e-01 -9.43589509e-02 1.01328397e+00 2.74053335e-01
-1.67580575e-01 6.63341507e-02 -4.25345868e-01 2.25312449e-02
8.59316349e-01 -3.09146550e-02 1.95816547e-01 -7.07440138e-01
-6.14586055e-01 -2.11037267e-02 -1.45349890e-01 4.73386705e-01
5.95245481e-01 -1.77729738e+00 -7.51827478e-01 2.69326419e-01
2.32095286e-01 -3.35423559e-01 5.65919101e-01 6.96074486e-01
3.34576726e-01 1.44045666e-01 -3.38751078e-01 -2.96449453e-01
-1.54190326e+00 5.78666866e-01 3.71045887e-01 1.98677555e-02
-1.42740354e-01 9.74400043e-01 2.14903980e-01 -6.37085080e-01
4.54470068e-01 -2.56136507e-01 -3.10221374e-01 2.92714775e-01
5.28687775e-01 1.98992848e-01 -7.47103319e-02 -1.02055800e+00
-2.47353122e-01 4.59327340e-01 3.62490773e-01 -4.67418909e-01
1.14245975e+00 -5.01750767e-01 -8.16900209e-02 7.94524968e-01
1.53421640e+00 1.91637665e-01 -1.06683576e+00 -1.52485877e-01
-3.26285154e-01 -1.59691870e-01 1.88740119e-01 -7.69001603e-01
-9.92933691e-01 1.26487911e+00 8.04578960e-01 2.52692133e-01
1.60396862e+00 -2.30427921e-01 8.39025676e-01 1.56910390e-01
-8.29927921e-02 -1.42077875e+00 1.80000290e-01 3.61500323e-01
1.16382945e+00 -1.11627090e+00 -5.82372665e-01 -4.02890116e-01
-1.05106354e+00 1.05853987e+00 5.46001196e-01 3.21707815e-01
6.98650301e-01 1.92501619e-01 4.02673841e-01 2.51976281e-01
-6.18880212e-01 7.40154982e-02 3.64550442e-01 6.55967534e-01
7.39551604e-01 1.21888950e-01 -2.30194345e-01 1.03153920e+00
-6.77343667e-01 -2.01247275e-01 1.41050771e-01 3.38656127e-01
-4.69372571e-01 -9.97123420e-01 -5.42883515e-01 -1.24194227e-01
-4.79298830e-01 -4.24999744e-02 -2.17105478e-01 3.19049418e-01
3.07320636e-02 1.36171603e+00 1.71603709e-02 -5.80504239e-01
4.30138916e-01 4.29069877e-01 2.15485126e-01 -2.66204894e-01
-5.76660216e-01 5.20339251e-01 1.78494960e-01 4.72812094e-02
-6.91837013e-01 -4.60050106e-01 -1.50834906e+00 1.29219636e-01
-6.79506302e-01 6.13627136e-01 7.52435684e-01 5.79312801e-01
2.75952131e-01 9.99144375e-01 1.19919443e+00 -5.97725630e-01
-6.87485635e-01 -1.37762749e+00 -7.51367569e-01 6.64210796e-01
2.41688102e-01 -4.82873559e-01 -5.60526073e-01 3.56103331e-01] | [14.326436042785645, 6.251211643218994] |
28f5524c-b514-464c-9e47-8098218a33dc | overlapping-word-removal-is-all-you-need | 2204.05488 | null | https://arxiv.org/abs/2204.05488v1 | https://arxiv.org/pdf/2204.05488v1.pdf | Overlapping Word Removal is All You Need: Revisiting Data Imbalance in Hope Speech Detection | Hope Speech Detection, a task of recognizing positive expressions, has made significant strides recently. However, much of the current works focus on model development without considering the issue of inherent imbalance in the data. Our work revisits this issue in hope-speech detection by introducing focal loss, data augmentation, and pre-processing strategies. Accordingly, we find that introducing focal loss as part of Multilingual-BERT's (M-BERT) training process mitigates the effect of class imbalance and improves overall F1-Macro by 0.11. At the same time, contextual and back-translation-based word augmentation with M-BERT improves results by 0.10 over baseline despite imbalance. Finally, we show that overlapping word removal based on pre-processing, though simple, improves F1-Macro by 0.28. In due process, we present detailed studies depicting various behaviors of each of these strategies and summarize key findings from our empirical results for those interested in getting the most out of M-BERT for hope speech detection under real-world conditions of data imbalance. | ['Bharathi Raja Chakravarthi', 'Anand Kumar Madasamy', 'Adithya Madhusoodanan', 'Navyasree Balamuralidhar', 'Gayathri Nisha', 'Manikandan Ravikiran', 'Hariharan RamakrishnaIyer LekshmiAmmal'] | 2022-04-12 | null | null | null | null | ['hope-speech-detection'] | ['natural-language-processing'] | [ 2.18300745e-01 2.02246785e-01 -5.35329342e-01 -5.52599251e-01
-1.47570169e+00 -3.52653861e-01 4.75026190e-01 4.91220206e-01
-5.86190701e-01 5.55244684e-01 5.72657108e-01 -4.97684240e-01
2.68114716e-01 -3.87101114e-01 -4.94588792e-01 -4.48019773e-01
1.22019820e-01 1.96057320e-01 -1.86698698e-02 -5.79276979e-01
1.05185650e-01 2.80746907e-01 -1.28199148e+00 5.64695895e-01
6.84025466e-01 5.85184991e-01 -3.22721481e-01 6.87541544e-01
-5.84854819e-02 9.08945978e-01 -8.82070601e-01 -6.58974051e-01
1.54247776e-01 -4.35957879e-01 -5.98218441e-01 1.16136521e-01
3.91272396e-01 -1.37282640e-01 -2.86523968e-01 1.02415192e+00
8.01321268e-01 -2.88397763e-02 4.06979233e-01 -1.27758539e+00
-3.24475259e-01 8.03308725e-01 -7.04460621e-01 5.11666775e-01
1.83297515e-01 3.11584145e-01 1.19398296e+00 -1.14373994e+00
4.16210622e-01 1.21380508e+00 8.63861203e-01 5.28791964e-01
-1.10967267e+00 -6.56935751e-01 1.98512301e-02 1.38555869e-01
-1.25511241e+00 -1.06395328e+00 7.52727032e-01 -2.12707713e-01
1.40995681e+00 6.63523078e-01 4.95646805e-01 8.95160735e-01
-1.50308684e-01 1.07215703e+00 7.46402740e-01 -7.85311282e-01
2.65506469e-03 2.53674090e-01 4.50329244e-01 6.56048715e-01
-1.62270650e-01 -2.00245574e-01 -8.32421899e-01 -1.13263644e-01
6.66344762e-02 -5.23813367e-01 -1.02408163e-01 2.40229592e-01
-8.19511294e-01 8.14494848e-01 4.96669859e-02 2.34843463e-01
-6.29674792e-02 -1.04846761e-01 6.53355420e-01 5.37619293e-01
9.71924901e-01 5.24917066e-01 -4.43238229e-01 -5.07743418e-01
-1.10388470e+00 9.77636948e-02 6.35981143e-01 7.32828677e-01
3.38385910e-01 2.16632292e-01 -4.63631868e-01 1.28081226e+00
-4.23090383e-02 3.04889828e-01 6.86903834e-01 -6.70086324e-01
7.75430918e-01 5.78516543e-01 -2.56320924e-01 -7.32293844e-01
-5.16728878e-01 -6.77153111e-01 -4.85972434e-01 -2.33927593e-01
3.54359865e-01 -3.58971328e-01 -8.40763211e-01 1.90817475e+00
1.59966826e-01 -2.57878929e-01 1.40950605e-01 6.24917150e-01
6.79111779e-01 5.89395285e-01 6.79273605e-02 -4.95666772e-01
1.35135365e+00 -1.12473023e+00 -9.41205561e-01 -4.82882619e-01
1.00848293e+00 -1.04945743e+00 1.29781401e+00 2.02011138e-01
-1.21653414e+00 -5.63914105e-02 -9.38428223e-01 -1.59246698e-01
-1.37591898e-01 2.15768144e-01 5.60777247e-01 1.04397380e+00
-9.16361690e-01 1.63895398e-01 -7.88183153e-01 -4.54431891e-01
3.48134816e-01 3.02067071e-01 -2.56738245e-01 -3.77050489e-02
-1.21361518e+00 9.46807563e-01 -2.25414947e-01 -1.69451132e-01
-5.40225565e-01 -8.42802525e-01 -8.57436359e-01 -1.13172762e-01
2.37139866e-01 -3.50440145e-01 1.52046800e+00 -1.01479876e+00
-1.23029637e+00 9.58252490e-01 -5.38489223e-01 -5.26463211e-01
4.89514709e-01 -1.96016967e-01 -5.59966028e-01 -3.02495748e-01
-3.70217897e-02 5.80361664e-01 4.62266743e-01 -6.58315539e-01
-4.49449152e-01 -5.18858850e-01 -1.22893095e-01 3.92718881e-01
-7.60858297e-01 5.69115162e-01 -4.19076413e-01 -7.86856771e-01
-1.68459103e-01 -8.04125249e-01 -6.35804459e-02 -4.30271298e-01
-4.08696681e-01 -2.49529660e-01 4.77142841e-01 -7.19439209e-01
1.53608000e+00 -2.23864746e+00 -2.38190055e-01 -4.61807400e-02
7.23757595e-02 3.39740545e-01 -4.30793226e-01 5.98936915e-01
-3.29813778e-01 3.58624667e-01 -2.61408746e-01 -7.46828079e-01
-8.09039474e-02 1.38968289e-01 -1.16804630e-01 5.32269180e-01
4.67570335e-01 7.84788072e-01 -5.76467037e-01 -2.04792663e-01
-1.72887242e-03 4.07869607e-01 -8.54329288e-01 2.36430485e-02
-5.39663173e-02 -3.59049626e-02 1.96557790e-01 6.37453020e-01
5.42306542e-01 1.59602478e-01 1.84781894e-01 1.26231298e-01
-2.26062000e-01 8.41281593e-01 -7.62601376e-01 1.19146812e+00
-5.17089963e-01 8.68836522e-01 1.14711456e-01 -1.02074552e+00
7.39256144e-01 2.49649927e-01 2.53269702e-01 -9.68954563e-01
1.15373187e-01 1.78657323e-01 2.93723017e-01 -4.34052706e-01
7.60229588e-01 -4.43745375e-01 8.85116383e-02 1.87097237e-01
-2.15550542e-01 4.04984178e-03 3.12504202e-01 1.59510717e-01
1.25810707e+00 -6.69169843e-01 2.09544197e-01 -8.93242434e-02
4.12815928e-01 -2.34191455e-02 7.48691559e-01 7.01171339e-01
-4.40685987e-01 6.17838442e-01 7.28119314e-01 8.26669186e-02
-9.81840611e-01 -5.95104456e-01 -1.11105852e-01 1.25440454e+00
-6.39791429e-01 -6.77929938e-01 -7.06343353e-01 -6.33640409e-01
-1.97524186e-02 9.68580186e-01 -5.67788422e-01 -2.67603636e-01
-6.55377984e-01 -1.38516927e+00 9.32981253e-01 4.73976552e-01
9.65603217e-02 -8.02888930e-01 -2.36545593e-01 1.01192951e-01
-4.72244382e-01 -1.11356437e+00 -4.24470335e-01 3.75549525e-01
-6.80150509e-01 -7.16997206e-01 -5.86656272e-01 -6.58032298e-01
3.24457705e-01 2.37880394e-01 1.07521045e+00 1.88731417e-01
-1.13589324e-01 -9.60860681e-03 -3.86791259e-01 -5.33253491e-01
-6.70136929e-01 3.49978060e-01 -1.63541198e-01 -2.13975489e-01
6.14121675e-01 -3.11926275e-01 -1.17289208e-01 8.75941887e-02
-6.76796079e-01 1.40630584e-02 5.75582147e-01 9.13169920e-01
2.58799165e-01 -1.81190893e-01 7.92687893e-01 -9.57192302e-01
8.32275510e-01 -4.32634294e-01 -1.00948513e-01 -3.06149684e-02
-7.62730718e-01 -1.35263018e-02 2.91265219e-01 -3.41185182e-01
-8.88474345e-01 -2.29566738e-01 -7.31948733e-01 -4.19615209e-02
-9.76431463e-03 4.60960537e-01 -2.55079120e-01 2.54380286e-01
6.10548496e-01 5.12681305e-02 4.96679023e-02 -5.56556463e-01
3.45861435e-01 1.02497303e+00 2.57660359e-01 -1.22518204e-01
3.49002630e-01 1.99323744e-01 -5.22260427e-01 -1.11756957e+00
-9.28361058e-01 -6.47529602e-01 -2.43553132e-01 -1.05658942e-03
4.09092128e-01 -1.09295177e+00 -2.53199041e-01 6.49729431e-01
-9.88604844e-01 -4.23057497e-01 -2.95236319e-01 2.66247302e-01
-1.99678257e-01 2.37144113e-01 -8.71714115e-01 -1.09244740e+00
-4.49222147e-01 -1.22270298e+00 9.58847463e-01 -2.16390356e-01
-5.66977322e-01 -6.43982291e-01 1.01878993e-01 8.27427030e-01
4.48631793e-01 -2.55525708e-01 1.02709055e+00 -9.61889625e-01
8.34417790e-02 -1.95472687e-01 -6.06331080e-02 5.40233254e-01
1.21588886e-01 -1.15372658e-01 -1.15017259e+00 -1.50860295e-01
-2.94999927e-02 -2.68095851e-01 9.29118335e-01 1.86890006e-01
7.61586845e-01 -4.96744663e-01 1.30026862e-01 2.36265600e-01
9.68088806e-01 8.85357037e-02 5.11567295e-01 3.31490695e-01
4.78711635e-01 6.01745903e-01 6.26046360e-01 4.87479776e-01
5.11299789e-01 9.07146096e-01 2.44401529e-01 -3.94185275e-01
-4.22829181e-01 -1.16708726e-01 6.36591315e-01 1.04800665e+00
4.60137695e-01 -3.31930161e-01 -1.01478219e+00 9.02235270e-01
-1.52438140e+00 -8.18389058e-01 -3.98568332e-01 2.00247216e+00
1.20006704e+00 3.02065462e-01 5.16417801e-01 4.70269650e-01
4.90595579e-01 2.30084345e-01 -2.01434776e-01 -5.73395789e-01
-2.73861945e-01 1.67303964e-01 4.18752998e-01 9.00211632e-01
-8.69023919e-01 1.01494074e+00 6.43659019e+00 8.94848287e-01
-1.23316789e+00 2.86367416e-01 9.39032078e-01 -4.98612970e-01
-3.56586039e-01 -2.22325459e-01 -7.53797174e-01 3.34110767e-01
1.19744313e+00 -5.51743060e-02 2.46038243e-01 5.13624907e-01
5.15041351e-01 -3.22523713e-02 -9.88678813e-01 7.82935381e-01
3.40714842e-01 -8.81736457e-01 -2.41141066e-01 -5.39336987e-02
5.89704275e-01 2.98529685e-01 1.37855247e-01 5.09974003e-01
7.50698745e-02 -8.12277079e-01 8.22045565e-01 -5.05300760e-02
5.71009755e-01 -9.70067620e-01 8.53515863e-01 4.20456499e-01
-5.48816741e-01 -7.95429349e-02 -9.11470503e-03 -3.29463542e-01
7.06749782e-02 1.06510305e+00 -1.08541703e+00 3.06445092e-01
3.37656289e-01 4.10915166e-01 -6.08616352e-01 6.90056145e-01
-1.13186233e-01 1.10169470e+00 -3.14893156e-01 -3.36429179e-02
1.49263635e-01 3.66896033e-01 6.23110056e-01 1.64984894e+00
-7.29824454e-02 -1.32420108e-01 5.38709909e-02 2.47260988e-01
-2.43677929e-01 4.62261975e-01 -3.13633323e-01 -2.15513825e-01
3.86802942e-01 9.64648664e-01 -2.45290309e-01 -3.68170768e-01
-4.19308484e-01 7.03867733e-01 4.57878053e-01 -1.53827248e-02
-6.39147162e-01 -3.18282962e-01 8.90889585e-01 2.96997070e-01
-7.89166912e-02 -1.05752803e-01 -5.58781505e-01 -9.84679520e-01
1.16845638e-01 -1.32286513e+00 3.69209468e-01 -4.72097188e-01
-9.33075607e-01 3.95362794e-01 -3.59044880e-01 -6.74399614e-01
-3.15060198e-01 -2.70785093e-01 -4.13571388e-01 7.74128973e-01
-1.52732992e+00 -8.84070635e-01 8.90469030e-02 1.17794886e-01
1.06247687e+00 -9.61048454e-02 6.55041158e-01 7.35495448e-01
-1.10310054e+00 1.12527561e+00 -2.80163288e-02 2.37275720e-01
9.46608961e-01 -9.82374847e-01 5.97885847e-01 1.00604844e+00
2.09028527e-01 6.28233314e-01 8.63288462e-01 -5.30122995e-01
-1.08480310e+00 -9.81787145e-01 1.52823186e+00 -4.57837641e-01
8.42143476e-01 -5.60622752e-01 -8.72522295e-01 6.08223319e-01
3.73538435e-01 -3.36449653e-01 7.87636697e-01 4.48802233e-01
-3.80209386e-01 -2.20228612e-01 -1.04538023e+00 6.42942846e-01
9.06027317e-01 -6.41667962e-01 -3.78795475e-01 5.69380701e-01
7.44005620e-01 -3.62769067e-01 -7.09572554e-01 4.39604431e-01
4.85566884e-01 -9.41032529e-01 6.38486147e-01 -9.05682087e-01
4.10771877e-01 1.69547617e-01 -3.36943299e-01 -1.35134518e+00
1.98329519e-02 -6.95246935e-01 2.43941441e-01 1.51444399e+00
7.46707201e-01 -4.88440543e-01 7.44614422e-01 4.33023453e-01
-4.48442459e-01 -8.50903094e-01 -9.76013899e-01 -6.56307757e-01
2.76356906e-01 -9.18761909e-01 2.67171979e-01 1.02892601e+00
4.35371369e-01 5.67598224e-01 -3.26860160e-01 -9.43771303e-02
1.48237839e-01 -4.27485317e-01 8.09830129e-01 -5.87360859e-01
-2.70785004e-01 -4.25051838e-01 -2.54765153e-01 -9.80318189e-01
2.11911127e-01 -9.42375302e-01 3.64763429e-03 -1.16855180e+00
3.29187036e-01 -3.84148628e-01 -3.06788176e-01 8.65022719e-01
-3.38404685e-01 4.81209993e-01 3.40770215e-01 1.01416238e-01
-2.95153379e-01 4.71870869e-01 6.34165108e-01 -1.71320990e-01
-1.94779605e-01 3.83555181e-02 -9.73280013e-01 5.82001209e-01
9.68564689e-01 -6.52320862e-01 -2.34804690e-01 -4.74320650e-01
1.27134174e-01 -2.75881030e-02 2.93758959e-02 -7.09900022e-01
4.56567444e-02 7.70514160e-02 -9.29915607e-02 -4.27393228e-01
4.79319245e-01 -3.22957635e-01 -3.42269212e-01 4.88641471e-01
-5.53774059e-01 4.86468434e-01 5.34125209e-01 1.86735123e-01
-1.97704017e-01 -2.17929617e-01 8.16202939e-01 1.60890147e-02
-3.35694551e-01 -3.56061697e-01 -6.28208458e-01 3.68988603e-01
5.38793445e-01 2.05552414e-01 -5.58637559e-01 -6.43538237e-01
-4.87558007e-01 7.82126337e-02 2.24252164e-01 5.43543696e-01
1.54598325e-01 -1.00627518e+00 -9.06359434e-01 4.55426663e-01
2.29087755e-01 -5.89188397e-01 6.36055619e-02 1.35717142e+00
-1.07603103e-01 4.36185896e-01 3.71780753e-01 -4.00722623e-01
-1.87270272e+00 1.95832893e-01 2.58117825e-01 -3.98485780e-01
-2.53465474e-01 9.74973023e-01 -1.47986382e-01 -4.71215129e-01
4.19049203e-01 -2.06095308e-01 -2.77064815e-02 2.24408790e-01
4.36620981e-01 3.76766980e-01 7.82748759e-01 -7.82099664e-01
-6.18662953e-01 -2.40755472e-02 -1.92949191e-01 -4.85263765e-01
1.36317599e+00 3.00213322e-02 -7.03318790e-02 5.45901179e-01
1.23255110e+00 4.88050491e-01 -5.32722890e-01 -3.56061012e-02
9.07022357e-02 -1.97395414e-01 2.16625765e-01 -1.10274398e+00
-9.85085070e-01 8.80153358e-01 5.46073377e-01 5.49549386e-02
1.18578649e+00 -6.69578537e-02 7.13575125e-01 1.50394797e-01
-2.40984172e-01 -1.00931740e+00 5.98457782e-03 5.98956525e-01
8.08519363e-01 -1.17225182e+00 -2.06183299e-01 -6.22822881e-01
-6.73340678e-01 6.64965093e-01 4.20387059e-01 4.29920256e-01
3.77046585e-01 5.18169343e-01 3.60353261e-01 5.00450842e-03
-9.73121345e-01 -1.52647406e-01 1.57218403e-03 2.58762509e-01
8.00812542e-01 6.78282902e-02 -6.44723237e-01 4.57366228e-01
-5.45022070e-01 -3.50822628e-01 5.85132718e-01 7.82112896e-01
-4.46288079e-01 -1.31734407e+00 -2.22921446e-01 5.08676052e-01
-8.39896739e-01 -5.69720507e-01 -6.03925824e-01 8.20803046e-01
-2.24895373e-01 1.32138765e+00 2.35791743e-01 -5.21432400e-01
5.34598112e-01 3.82589847e-01 2.05535710e-01 -5.25195420e-01
-8.71797323e-01 2.83129722e-01 7.50237346e-01 -2.47708499e-01
-1.72536939e-01 -9.36681986e-01 -9.87545729e-01 -4.31859016e-01
-5.38285434e-01 2.02207491e-01 6.82931781e-01 9.21828747e-01
3.76261532e-01 6.14638805e-01 5.16949713e-01 -1.47619545e-01
-7.14632928e-01 -1.27851641e+00 -2.57199168e-01 2.71444857e-01
5.52067339e-01 -2.91717410e-01 -5.78396976e-01 -2.31383860e-01] | [11.007536888122559, 8.78824520111084] |
11bc3523-aaa5-46ad-b7eb-91ed6b886947 | exploring-the-in-context-learning-ability-of | 2307.01137 | null | https://arxiv.org/abs/2307.01137v1 | https://arxiv.org/pdf/2307.01137v1.pdf | Exploring the In-context Learning Ability of Large Language Model for Biomedical Concept Linking | The biomedical field relies heavily on concept linking in various areas such as literature mining, graph alignment, information retrieval, question-answering, data, and knowledge integration. Although large language models (LLMs) have made significant strides in many natural language processing tasks, their effectiveness in biomedical concept mapping is yet to be fully explored. This research investigates a method that exploits the in-context learning (ICL) capabilities of large models for biomedical concept linking. The proposed approach adopts a two-stage retrieve-and-rank framework. Initially, biomedical concepts are embedded using language models, and then embedding similarity is utilized to retrieve the top candidates. These candidates' contextual information is subsequently incorporated into the prompt and processed by a large language model to re-rank the concepts. This approach achieved an accuracy of 90.% in BC5CDR disease entity normalization and 94.7% in chemical entity normalization, exhibiting a competitive performance relative to supervised learning methods. Further, it showed a significant improvement, with an over 20-point absolute increase in F1 score on an oncology matching dataset. Extensive qualitative assessments were conducted, and the benefits and potential shortcomings of using large language models within the biomedical domain were discussed. were discussed. | ['Rong Xu', 'Zhenxiang Gao', 'Qinyong Wang'] | 2023-07-03 | null | null | null | null | ['retrieval', 'question-answering', 'information-retrieval', 'literature-mining'] | ['methodology', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 5.51591218e-01 3.20258677e-01 -5.93377173e-01 -1.43487215e-01
-8.88413727e-01 -1.37984723e-01 5.40969670e-01 1.26370394e+00
-6.63376331e-01 8.23637605e-01 4.36883003e-01 -1.79377899e-01
-4.70721751e-01 -6.50278330e-01 -1.28102109e-01 -5.66869199e-01
6.88431486e-02 4.55406606e-01 3.62155885e-02 -9.15743411e-02
3.31578493e-01 4.45977747e-01 -1.14554012e+00 3.21739286e-01
9.11253333e-01 5.29274762e-01 6.06758296e-02 2.79680520e-01
-5.56853116e-01 8.46802533e-01 -4.72142011e-01 -4.38406974e-01
-4.40459549e-01 -3.44918489e-01 -9.23942327e-01 -5.38321197e-01
1.13421127e-01 4.35019881e-01 -9.56007242e-02 1.09678721e+00
8.25813413e-01 -1.90914124e-01 5.96084356e-01 -9.94419217e-01
-5.94299614e-01 5.95656276e-01 -5.11850476e-01 2.25066841e-01
6.70507491e-01 -5.12107909e-01 1.13080454e+00 -9.82183874e-01
9.62647676e-01 1.14781308e+00 7.35748410e-01 4.98202533e-01
-1.20847714e+00 -9.12660122e-01 -2.63603270e-01 1.95436120e-01
-1.75569522e+00 -2.08615959e-01 5.88341355e-01 -4.74642873e-01
1.20229805e+00 2.66691238e-01 5.14124393e-01 6.97912037e-01
6.35581374e-01 2.19104275e-01 9.56505716e-01 -7.66963363e-01
8.15784857e-02 4.26976502e-01 2.90775806e-01 7.15595782e-01
5.53471029e-01 -1.24254458e-01 -6.44489825e-01 -5.28768837e-01
2.61049002e-01 -3.09282690e-02 4.53100353e-02 1.39655946e-02
-1.27889991e+00 8.72077167e-01 6.05998218e-01 8.50306749e-01
-4.30173367e-01 -3.39266121e-01 4.52008843e-01 2.50937998e-01
5.33349335e-01 8.04385841e-01 -2.47213617e-01 4.27748293e-01
-9.32857454e-01 -3.09829880e-02 6.97689891e-01 8.75397623e-01
4.32740092e-01 -5.67091465e-01 -1.33670762e-01 8.66570652e-01
3.13518345e-01 1.20757103e-01 8.84789944e-01 -3.26589167e-01
3.38718414e-01 9.92413998e-01 -6.11317515e-01 -1.48629606e+00
-7.86167383e-01 -6.05933726e-01 -9.47799146e-01 -4.95165139e-01
-1.66789606e-01 2.95273215e-01 -6.27216756e-01 1.72293961e+00
4.06291097e-01 3.03090960e-01 3.14893842e-01 4.33883399e-01
1.06112552e+00 1.74287945e-01 9.69999731e-01 -2.93709397e-01
1.78852272e+00 -4.68907207e-01 -1.07064617e+00 1.75506761e-03
1.02905869e+00 -9.87132072e-01 5.03686190e-01 -1.73476979e-01
-7.32139945e-01 -4.00371552e-01 -9.80755389e-01 -1.10191062e-01
-8.12756717e-01 -5.66251762e-02 6.77545488e-01 4.94035512e-01
-1.04459369e+00 3.75911236e-01 -6.31748199e-01 -8.61677408e-01
5.92873037e-01 5.20034790e-01 -6.82473123e-01 -2.39950866e-01
-1.47217238e+00 1.16544461e+00 7.78732717e-01 -2.85187364e-01
-2.17185006e-01 -1.02959406e+00 -8.39811146e-01 -1.55856803e-01
1.65351927e-01 -1.03015637e+00 4.91100013e-01 -3.98413599e-01
-8.10104847e-01 1.06605065e+00 -2.90974587e-01 -4.32676286e-01
7.85877630e-02 -2.92806514e-02 -9.60985303e-01 3.24688077e-01
5.19089520e-01 6.82149231e-01 -3.20277736e-02 -8.73958170e-01
-6.75074816e-01 -5.57632804e-01 -1.66796237e-01 2.82106757e-01
-7.59760976e-01 3.18854630e-01 -4.16042179e-01 -7.06166625e-01
3.10493290e-01 -8.39332461e-01 -3.50571126e-01 -1.41598031e-01
-2.16707647e-01 -2.35481814e-01 3.99667352e-01 -7.22937524e-01
1.66788220e+00 -1.73717642e+00 -4.16074619e-02 5.56348860e-01
4.66879606e-01 2.30022520e-01 -9.33281928e-02 6.22153223e-01
-3.90872777e-01 3.62816304e-01 -2.10432619e-01 4.28864136e-02
-4.28214908e-01 -1.05585597e-01 2.72604465e-01 3.36233258e-01
3.73381436e-01 1.07741177e+00 -1.13293779e+00 -1.01263881e+00
9.46647837e-04 5.48941314e-01 -4.32640284e-01 -1.60561632e-02
2.20610619e-01 8.37663263e-02 -4.93443787e-01 1.02966428e+00
1.48300201e-01 -4.48720038e-01 5.86160362e-01 -5.49677134e-01
1.91666797e-01 1.31831080e-01 -8.50351691e-01 1.78140211e+00
-1.97215199e-01 3.72510612e-01 -3.12057942e-01 -9.77867424e-01
8.94899309e-01 5.82638919e-01 1.02338958e+00 -6.70181394e-01
-2.76022702e-02 4.32368368e-01 2.21871790e-02 -6.67261958e-01
3.56787473e-01 -3.60492647e-01 1.06224760e-01 1.69628225e-02
3.43729146e-02 2.70499378e-01 1.50899455e-01 4.44084555e-01
1.11902702e+00 -1.62120938e-01 9.41938281e-01 -3.63028795e-01
9.24930990e-01 2.19105721e-01 4.01276380e-01 2.17060775e-01
1.67974476e-02 1.44356370e-01 3.65156025e-01 1.46189528e-02
-5.47132373e-01 -6.04131222e-01 -3.87349874e-01 6.92081630e-01
-5.06255552e-02 -7.32362390e-01 -5.70057094e-01 -5.10770500e-01
2.35700652e-01 5.63284397e-01 -6.94051147e-01 -4.71925139e-01
-2.98534602e-01 -1.08397996e+00 7.77733743e-01 4.06504095e-01
6.92152083e-02 -9.80684996e-01 -2.30595514e-01 4.59359586e-01
-1.12265460e-01 -1.12240148e+00 -2.44672745e-01 4.13441435e-02
-1.17750251e+00 -1.44680238e+00 -4.49668497e-01 -1.08811176e+00
1.00477552e+00 -2.45316457e-02 9.36477482e-01 1.37269944e-01
-7.23024964e-01 3.22751582e-01 -1.50375485e-01 -5.42424142e-01
-6.40395105e-01 2.42852569e-01 1.17591493e-01 -3.08422804e-01
9.31909502e-01 -2.84650207e-01 -5.06235242e-01 1.34194866e-01
-1.14033973e+00 -1.17478400e-01 1.05670142e+00 9.60085034e-01
8.57218504e-01 -8.93291235e-02 1.05957603e+00 -1.19584405e+00
9.87054646e-01 -6.71896815e-01 -9.25691426e-02 4.80421215e-01
-1.36063278e+00 -6.54750764e-02 2.42901891e-01 -2.66099900e-01
-6.87333763e-01 -8.32969323e-02 -2.23602295e-01 3.84610929e-02
9.16585606e-03 1.15276849e+00 -9.93316695e-02 -2.51124829e-01
7.04254031e-01 -1.94719404e-01 1.22370936e-01 -2.89807290e-01
3.68285120e-01 7.94789433e-01 3.14944714e-01 -4.19615060e-01
5.22741556e-01 6.31103665e-02 2.72176206e-01 -7.28694975e-01
-5.06942034e-01 -9.19154823e-01 -8.27601790e-01 1.24949977e-01
1.05285120e+00 -9.63849187e-01 -4.92863446e-01 -7.50143155e-02
-9.01438713e-01 5.38449287e-01 -2.49605104e-02 6.96915805e-01
2.62426697e-02 3.79460454e-01 -4.43601131e-01 -3.49612206e-01
-7.27589548e-01 -8.04054260e-01 6.75365746e-01 2.89859772e-01
-6.08406782e-01 -1.29079545e+00 1.39676720e-01 5.03734946e-01
3.99906993e-01 3.37544978e-01 1.54434979e+00 -1.14234614e+00
-6.84508821e-03 -6.50204003e-01 -2.29199544e-01 -1.77286237e-01
6.32701695e-01 -4.45813239e-01 -7.38655031e-01 -3.33067834e-01
-3.95905733e-01 1.28692510e-02 4.79155868e-01 6.76676035e-02
8.32660973e-01 -8.23796168e-02 -1.01464343e+00 1.89648211e-01
1.48240900e+00 3.77564669e-01 4.50608790e-01 4.40929413e-01
7.98824489e-01 7.50826001e-01 5.94560742e-01 -9.25874785e-02
4.43857640e-01 5.35008848e-01 -8.90426920e-04 -2.85256952e-01
-1.78443879e-01 -3.04046571e-01 -9.45550725e-02 1.14905107e+00
1.42927751e-01 1.76074401e-01 -1.12241030e+00 6.16183043e-01
-1.56820333e+00 -7.28134274e-01 3.19331624e-02 1.98529768e+00
1.21367753e+00 2.42575437e-01 -3.45252782e-01 4.06556949e-03
6.83091760e-01 -3.58047605e-01 -4.71966624e-01 -1.45860270e-01
-3.17818522e-01 4.36246902e-01 3.65616769e-01 2.79633999e-01
-8.60339582e-01 9.19550836e-01 5.94269562e+00 8.53443682e-01
-7.84362018e-01 9.70524102e-02 4.81335253e-01 2.11871773e-01
-3.64039361e-01 3.61547582e-02 -7.36249745e-01 2.13104248e-01
1.13282621e+00 -4.94877696e-01 -2.16260508e-01 3.26463789e-01
1.38753429e-01 1.57704219e-01 -1.06623685e+00 8.92834663e-01
2.49550611e-01 -1.47455812e+00 4.37911600e-01 1.12990208e-01
6.15410149e-01 -3.00869823e-01 -3.10236812e-01 1.96106881e-01
-2.12840214e-01 -1.06039703e+00 -5.44489399e-02 7.90391147e-01
9.80872989e-01 -7.27738798e-01 1.14913762e+00 -3.91147435e-02
-1.29679012e+00 1.94548685e-02 -3.70308340e-01 5.06814182e-01
-1.01799227e-01 5.88791788e-01 -1.17486525e+00 1.11628318e+00
3.80513906e-01 8.24143827e-01 -7.41786063e-01 9.93387759e-01
-3.55555639e-02 4.36568409e-01 1.54238626e-01 -1.01092875e-01
-1.38733303e-02 -7.54507631e-02 3.21072280e-01 1.56043959e+00
1.91840470e-01 3.14550757e-01 1.59841210e-01 5.60542881e-01
-3.29578459e-01 8.99682224e-01 -4.60958064e-01 -4.36479598e-01
5.74528456e-01 1.37641358e+00 -8.24361086e-01 -3.84277523e-01
-5.05294442e-01 3.87794673e-01 5.96356690e-02 5.80240041e-02
-5.39616466e-01 -7.32900441e-01 4.51117843e-01 1.52303576e-01
-4.27339524e-01 2.96776384e-01 -2.03070253e-01 -7.45552123e-01
-2.39472583e-01 -9.67299283e-01 8.49069357e-01 -4.65808570e-01
-1.40512335e+00 5.89025259e-01 2.21093986e-02 -1.22996187e+00
2.63166279e-02 -4.91511256e-01 -1.30366430e-01 8.78719747e-01
-1.61693835e+00 -1.28278208e+00 -6.57030791e-02 4.13747430e-01
1.58350751e-01 -4.91647184e-01 1.40541375e+00 7.63371944e-01
-4.41686839e-01 7.58530021e-01 9.77549329e-02 1.74338222e-01
9.97017980e-01 -1.09747279e+00 -9.13278684e-02 4.05872107e-01
-1.26932608e-02 1.24965000e+00 3.03844392e-01 -9.64249969e-01
-1.15090597e+00 -1.05015588e+00 1.67103100e+00 -2.75582254e-01
7.41758287e-01 -1.83615531e-03 -1.09671593e+00 2.25135192e-01
3.85051370e-02 -1.60635456e-01 1.53511167e+00 -1.08246058e-01
-2.36755371e-01 -6.60138503e-02 -1.21575725e+00 6.20035529e-01
8.75941157e-01 -7.83813953e-01 -7.86833704e-01 3.62761885e-01
7.15274692e-01 -1.63911864e-01 -1.49364185e+00 4.26656067e-01
5.82168519e-01 6.09857850e-02 1.23604345e+00 -1.14915216e+00
2.87412435e-01 -2.81383693e-01 -1.32182434e-01 -8.64383459e-01
-3.87993485e-01 -3.31546277e-01 1.89934485e-02 1.33168995e+00
7.35951424e-01 -6.94565892e-01 5.37787616e-01 6.45032763e-01
-3.07541136e-02 -9.26169753e-01 -7.96973705e-01 -3.54605943e-01
2.43975464e-02 -1.15343459e-01 4.21191722e-01 1.37677228e+00
4.09587979e-01 5.45811415e-01 7.15625510e-02 1.29348695e-01
4.32013661e-01 -9.18721855e-02 1.37873217e-01 -1.56915092e+00
2.01774821e-01 -5.60162485e-01 -7.11613894e-01 -1.37490064e-01
2.11493850e-01 -1.67438400e+00 -4.82844770e-01 -1.93814564e+00
4.97166902e-01 -4.06431079e-01 -1.03449953e+00 6.58607364e-01
-4.55389500e-01 1.60361931e-01 -2.20431894e-01 3.69058758e-01
-5.74695826e-01 1.07069343e-01 7.54730225e-01 -3.23088080e-01
-1.96035892e-01 -2.31997043e-01 -1.13421726e+00 4.05971706e-01
7.15771496e-01 -8.89372289e-01 -3.57538581e-01 9.54216998e-03
2.21558183e-01 -9.44733545e-02 -6.02120906e-03 -9.19236004e-01
6.62366152e-01 -2.20320970e-02 4.70815718e-01 -3.51583779e-01
-1.85830202e-02 -8.56750906e-01 3.41191649e-01 8.76581907e-01
-6.71547055e-01 3.35155368e-01 5.21356285e-01 6.52991772e-01
-3.60841066e-01 -1.92673430e-01 4.58648503e-01 -1.10473568e-02
-6.45738125e-01 1.33240089e-01 -1.54482916e-01 -9.08738449e-02
9.72422004e-01 -1.36615247e-01 -1.19150855e-01 2.19727844e-01
-8.08759749e-01 1.90794125e-01 3.78976502e-02 5.98515093e-01
7.06456423e-01 -1.47358263e+00 -6.83585525e-01 -7.93244243e-02
5.36732852e-01 -4.16162640e-01 3.21880803e-02 1.05355334e+00
-2.70877779e-01 8.53475332e-01 -9.64044258e-02 -4.53764707e-01
-1.63779187e+00 4.71854687e-01 3.68659981e-02 -6.34461343e-01
-3.48517329e-01 6.12862527e-01 -1.06066681e-01 -3.07649523e-01
1.18010208e-01 8.89977515e-02 -7.54260600e-01 4.68975574e-01
3.77203345e-01 1.73737779e-01 4.39516544e-01 -7.20434725e-01
-8.51879179e-01 8.48269880e-01 -2.88007438e-01 1.71637163e-01
1.26100111e+00 8.91309902e-02 -5.57570934e-01 2.52359331e-01
1.32374597e+00 1.11155070e-01 1.07646927e-01 -5.89727879e-01
8.31906199e-01 -6.06342293e-02 -2.55931560e-02 -1.06139565e+00
-7.36170590e-01 5.66495657e-01 7.65472114e-01 -2.43945226e-01
1.10223377e+00 4.36890684e-02 4.64014113e-01 4.71636027e-01
2.15654746e-01 -1.10122621e+00 -1.27971489e-02 2.10866556e-01
6.41526222e-01 -1.13934934e+00 3.80106449e-01 -4.34770584e-01
-4.14544910e-01 1.14881802e+00 2.83915490e-01 5.01912475e-01
7.68186212e-01 1.30886748e-01 8.26033950e-02 -5.99827766e-01
-4.87431467e-01 -5.05831838e-02 7.53759801e-01 3.69613826e-01
9.36807752e-01 -1.62764695e-02 -8.67008746e-01 6.68495536e-01
5.99312596e-02 1.05881661e-01 -5.02833948e-02 1.04852629e+00
-2.51239806e-01 -1.40460181e+00 -1.52318060e-01 5.84706306e-01
-8.11016798e-01 -4.06709194e-01 -4.29676265e-01 7.88097382e-01
6.52359724e-02 9.43913817e-01 -4.72058654e-01 -4.04818982e-01
4.79013175e-01 3.58720690e-01 -4.09057327e-02 -8.73539567e-01
-8.08605492e-01 1.21230438e-01 7.39345327e-02 -4.23907936e-01
-7.41392076e-01 -4.47565883e-01 -1.58921623e+00 1.24657415e-01
-6.02674663e-01 4.73954737e-01 7.26777494e-01 8.93186212e-01
6.07063890e-01 8.36913049e-01 2.10661530e-01 2.22043484e-01
-2.02250168e-01 -8.70334685e-01 -1.94796741e-01 3.40210646e-01
-1.62141055e-01 -5.72513282e-01 -6.69256747e-02 4.73266132e-02] | [8.4998779296875, 8.652421951293945] |
ce6b6df2-0699-4ef5-a5c6-834c8a6c6755 | time-frequency-distributions-of-heart-sound | 2208.03128 | null | https://arxiv.org/abs/2208.03128v1 | https://arxiv.org/pdf/2208.03128v1.pdf | Time-Frequency Distributions of Heart Sound Signals: A Comparative Study using Convolutional Neural Networks | Time-Frequency Distributions (TFDs) support the heart sound characterisation and classification in early cardiac screening. However, despite the frequent use of TFDs in signal analysis, no study comprehensively compared their performances on deep learning for automatic diagnosis. Furthermore, the combination of signal processing methods as inputs for Convolutional Neural Networks (CNNs) has been proved as a practical approach to increasing signal classification performance. Therefore, this study aimed to investigate the optimal use of TFD/ combined TFDs as input for CNNs. The presented results revealed that: 1) The transformation of the heart sound signal into the TF domain achieves higher classification performance than using of raw signals. Among the TFDs, the difference in the performance was slight for all the CNN models (within $1.3\%$ in average accuracy). However, Continuous wavelet transform (CWT) and Chirplet transform (CT) outperformed the rest. 2) The appropriate increase of the CNN capacity and architecture optimisation can improve the performance, while the network architecture should not be overly complicated. Based on the ResNet or SEResNet family results, the increase in the number of parameters and the depth of the structure do not improve the performance apparently. 3) Combining TFDs as CNN inputs did not significantly improve the classification results. The findings of this study provided the knowledge for selecting TFDs as CNN input and designing CNN architecture for heart sound classification. | ['Ernest N. Kamavuako', 'Lilia Sidhom', 'Ines Chihi', 'Mohamed Trabelsi', 'Hak-Keung Lam', 'Yujia Xu', 'Xinqi Bao'] | 2022-08-05 | null | null | null | null | ['sound-classification'] | ['audio'] | [ 4.37813476e-02 -2.22445410e-02 1.35157093e-01 -1.08766168e-01
-2.04687398e-02 -1.75461262e-01 2.59294122e-01 1.48547143e-01
-6.32088363e-01 4.84979600e-01 -1.11096412e-01 -6.72970951e-01
-4.01322246e-01 -7.48167336e-01 -2.98937261e-01 -8.43000174e-01
-3.05680901e-01 -6.76857531e-02 3.04441571e-01 -2.33018547e-01
3.11960932e-02 7.12812304e-01 -1.66146505e+00 4.89753991e-01
6.57292366e-01 1.22974277e+00 1.16952151e-01 9.17413592e-01
-2.28725880e-01 7.30742753e-01 -9.75927174e-01 -7.19501302e-02
8.47178921e-02 -7.45364904e-01 -7.45468020e-01 -6.30501688e-01
5.06429747e-02 -1.18215613e-01 -6.04933761e-02 5.78897297e-01
1.22176909e+00 -8.35898668e-02 3.19757849e-01 -6.92025483e-01
-2.24709198e-01 7.64619291e-01 1.28330961e-01 9.05571461e-01
2.05411866e-01 2.83086835e-03 5.33402801e-01 -6.94932401e-01
2.82490015e-01 7.34972298e-01 1.19852245e+00 3.29761803e-01
-8.40162694e-01 -5.26858389e-01 -5.17622590e-01 2.92838186e-01
-1.32663739e+00 -2.24788100e-01 8.45388293e-01 -3.00028920e-01
1.18934667e+00 3.97460133e-01 1.24491203e+00 7.71311104e-01
2.22021580e-01 4.03214008e-01 9.40785229e-01 -7.83550262e-01
-3.51451263e-02 2.19020233e-01 1.26970351e-01 4.90809083e-01
2.44015992e-01 2.04250291e-01 -3.11774313e-01 -1.49982823e-02
7.05181718e-01 -2.91839331e-01 -3.21139902e-01 4.74321872e-01
-1.11927617e+00 6.90558612e-01 4.58714515e-01 1.13338470e+00
-5.18680274e-01 1.81901798e-01 8.84758413e-01 6.07309222e-01
5.04810154e-01 7.31260777e-01 -5.14505625e-01 -2.12506130e-01
-1.13696599e+00 1.68007277e-02 6.45147443e-01 1.53633326e-01
4.14530784e-02 5.81416845e-01 -2.71925479e-01 7.70010054e-01
-1.91702634e-01 3.12171310e-01 6.80935860e-01 -6.23552203e-01
1.15460947e-01 4.49325293e-01 -3.66875648e-01 -1.31064701e+00
-9.73956764e-01 -1.06568968e+00 -9.37453926e-01 -7.93694109e-02
6.78881168e-01 -3.89605731e-01 -8.37109625e-01 1.22373188e+00
5.39083257e-02 2.54912198e-01 2.77396608e-02 9.04734552e-01
1.31356931e+00 4.44687963e-01 1.16340742e-01 -2.13482723e-01
1.46281397e+00 -2.97838926e-01 -9.29431617e-01 2.20836058e-01
7.21538305e-01 -7.82326281e-01 7.85001397e-01 5.54860234e-01
-9.59369481e-01 -1.02641523e+00 -1.01532459e+00 2.93781400e-01
-4.06838953e-01 4.86722827e-01 5.49742341e-01 1.08029616e+00
-1.16834152e+00 8.47316325e-01 -8.32366824e-01 -2.57334560e-01
3.72614682e-01 4.83170569e-01 -7.98804164e-02 2.52710521e-01
-1.68436968e+00 1.07847881e+00 4.56130296e-01 3.99068296e-01
-4.93308902e-01 -9.11483467e-01 -4.78534311e-01 2.33537897e-01
-3.64951909e-01 -6.15365863e-01 8.97624969e-01 -1.13789654e+00
-1.44340408e+00 6.89682245e-01 3.69609714e-01 -8.56543839e-01
5.51340938e-01 -6.58904910e-02 -5.32435477e-01 4.58587974e-01
-2.85664409e-01 5.12922108e-01 8.25939357e-01 -5.97715437e-01
-6.11846745e-01 2.12704614e-01 -1.54272810e-01 -1.72922358e-01
-4.53773886e-01 7.20765144e-02 2.03759685e-01 -7.22902417e-01
2.78911114e-01 -5.63698590e-01 -4.56394032e-02 -3.41042995e-01
-1.93451643e-02 -2.28040889e-01 5.17348289e-01 -8.81147265e-01
1.38261640e+00 -2.26065969e+00 -1.68055907e-01 2.94053823e-01
2.78603196e-01 6.33312643e-01 2.60660768e-01 2.55692124e-01
-4.42804933e-01 3.17271560e-01 1.33375069e-02 1.73419729e-01
-3.56847495e-01 1.52076930e-01 1.47469550e-01 4.82671499e-01
2.68005788e-01 7.96685636e-01 -5.78497112e-01 -5.00951409e-01
4.62339282e-01 7.50529468e-01 -3.75090599e-01 -7.59327784e-02
4.23798352e-01 3.65724087e-01 -2.53159791e-01 1.61729440e-01
6.09856665e-01 8.14285278e-02 -6.89327642e-02 -5.40721357e-01
-2.28664398e-01 3.69853646e-01 -9.61595416e-01 1.18311274e+00
-4.80940491e-01 8.65544558e-01 -2.81850368e-01 -1.27421963e+00
1.21944976e+00 9.94142234e-01 7.18030572e-01 -8.15193355e-01
3.23072582e-01 5.82282364e-01 7.88329899e-01 -8.86640847e-01
-7.17837140e-02 -3.22214782e-01 3.58683228e-01 3.05574089e-02
1.48002326e-01 1.88985206e-02 2.81997491e-02 -5.08058131e-01
8.50121677e-01 -2.77264535e-01 1.88842028e-01 -6.11208558e-01
5.39994121e-01 -4.99559455e-02 3.84388328e-01 8.48662496e-01
7.39051402e-02 6.71600640e-01 5.59959769e-01 -9.36947465e-01
-7.83867240e-01 -4.51260597e-01 -4.81375426e-01 8.64291310e-01
-4.22568768e-01 -6.69106469e-02 -7.72005975e-01 -3.53096992e-01
-3.76519799e-01 3.72648686e-01 -5.70649683e-01 -1.40517011e-01
-7.96431959e-01 -1.02858746e+00 1.26328170e+00 4.58777040e-01
6.94229186e-01 -1.37441933e+00 -1.21488094e+00 3.89748245e-01
-1.67034090e-01 -8.87125254e-01 1.64400727e-01 5.76480567e-01
-1.08253956e+00 -9.59943891e-01 -1.04679537e+00 -7.66711473e-01
3.65379363e-01 -1.87103808e-01 9.21577096e-01 4.68075842e-01
-4.34776694e-01 1.76598892e-01 -6.92775726e-01 -7.40182281e-01
-6.73640668e-01 3.86103749e-01 -1.35458991e-01 -1.83807999e-01
1.70447722e-01 -5.52058101e-01 -7.35220909e-01 -1.08104035e-01
-7.28181005e-01 -2.19424024e-01 4.66389954e-01 7.76181102e-01
5.07510528e-02 2.65034854e-01 8.88887525e-01 -6.61123872e-01
1.06638741e+00 -1.92963317e-01 -5.77228442e-02 -1.85675055e-01
-5.80311358e-01 -3.16019237e-01 7.33016491e-01 -5.60470581e-01
-5.26251435e-01 -2.24722221e-01 -6.60887420e-01 -3.73129398e-01
-2.68502444e-01 7.09772170e-01 5.84629714e-01 -1.43595636e-01
9.80521679e-01 2.73237109e-01 3.37601095e-01 -4.71146762e-01
-4.04662192e-01 4.74657565e-01 2.46713072e-01 -1.75876409e-01
2.79134780e-01 1.51873901e-01 6.42011017e-02 -1.01309252e+00
-3.34719092e-01 -2.10971639e-01 -6.36900127e-01 -3.74099791e-01
1.23949409e+00 -5.91839135e-01 -6.69480562e-01 4.77880627e-01
-1.09810543e+00 -4.23038900e-02 -3.43227565e-01 8.40962231e-01
-9.28323418e-02 2.85099685e-01 -5.37365794e-01 -8.15015554e-01
-8.17960203e-01 -1.12436235e+00 3.06097835e-01 1.33542493e-01
-3.48804355e-01 -1.09457266e+00 -2.64951527e-01 2.68385615e-02
9.79226708e-01 5.67084372e-01 9.60765481e-01 -8.83784890e-01
1.83977440e-01 -4.38745737e-01 7.48153254e-02 7.52466381e-01
4.97519001e-02 7.11123124e-02 -1.12805796e+00 -1.02647007e-01
2.56176561e-01 2.33587757e-01 7.12455690e-01 8.30331206e-01
1.10003173e+00 -4.83140275e-02 2.51560640e-02 6.05430484e-01
1.05260158e+00 5.92306733e-01 8.58024359e-01 4.21350598e-01
2.76597619e-01 4.28311586e-01 1.63464054e-01 3.35797220e-01
-1.75983578e-01 3.61239761e-01 3.51075202e-01 -7.20165789e-01
-4.49564159e-01 3.12930286e-01 8.36392194e-02 9.31830645e-01
-6.64472044e-01 1.40230104e-01 -1.18212020e+00 4.29317832e-01
-1.14865959e+00 -9.85975146e-01 -5.53545475e-01 1.96661019e+00
6.33532643e-01 3.33507180e-01 2.11007029e-01 8.70499194e-01
5.96759558e-01 -1.62028670e-01 1.28976617e-03 -7.67501652e-01
-5.29396981e-02 7.82882452e-01 3.49642515e-01 8.24531913e-02
-1.05755258e+00 2.65968651e-01 6.49149799e+00 7.53803611e-01
-1.82364452e+00 7.22187310e-02 6.59947574e-01 9.58971530e-02
2.16193765e-01 -5.65835834e-01 -3.78412038e-01 4.61303711e-01
1.17506456e+00 1.83862686e-01 1.11512467e-01 4.67928946e-01
3.34700376e-01 1.14850782e-01 -7.31400490e-01 9.11053658e-01
-2.50339597e-01 -1.35603952e+00 -2.26338372e-01 -2.71098495e-01
1.98812947e-01 1.07288528e-02 -1.23993717e-01 2.35014349e-01
-8.04734468e-01 -1.15994835e+00 6.91939294e-01 5.89742899e-01
7.98700511e-01 -7.14689374e-01 1.39084721e+00 3.03973705e-01
-1.10810816e+00 -2.70265639e-01 -3.22335422e-01 -1.41185626e-01
-2.13227719e-01 6.32519782e-01 -1.08756435e+00 6.05133355e-01
1.08936787e+00 4.15164441e-01 -6.97146237e-01 1.15694976e+00
1.72950044e-01 1.22929275e+00 -4.30475652e-01 -2.14404792e-01
1.71916425e-01 2.26242304e-01 4.77004677e-01 1.65642655e+00
4.90889013e-01 -6.26267120e-02 -2.77505755e-01 6.51024580e-01
5.66303194e-01 2.97717422e-01 -3.99347752e-01 2.22973615e-01
3.50000978e-01 1.00694048e+00 -1.03630376e+00 -2.32837066e-01
-2.98538804e-01 1.72170475e-01 -5.22864819e-01 1.51806191e-01
-7.72209942e-01 -7.27059066e-01 2.37797096e-01 5.19601464e-01
2.50622094e-01 2.92342931e-01 -5.51573455e-01 -3.82112592e-01
-1.85083877e-02 -8.35138917e-01 4.41753685e-01 -4.31883365e-01
-8.73785198e-01 9.42611814e-01 9.88539159e-02 -1.42588043e+00
-1.30733117e-01 -5.65943956e-01 -8.75348508e-01 1.06468391e+00
-1.40954721e+00 -5.88528275e-01 -3.41543615e-01 4.13662076e-01
2.37511262e-01 -2.19403267e-01 9.68129754e-01 6.41340196e-01
-2.82837659e-01 5.81331611e-01 -4.32719469e-01 3.78693789e-01
1.71487764e-01 -1.13619637e+00 8.93213227e-02 6.17133200e-01
-2.07428679e-01 5.56873858e-01 7.36685038e-01 -2.48696953e-01
-8.00706804e-01 -8.04022133e-01 9.64303911e-01 9.97988805e-02
6.68759793e-02 2.52852459e-02 -9.62552369e-01 -1.05778754e-01
3.42335373e-01 6.09575696e-02 6.18303418e-01 -1.56892702e-01
1.79646745e-01 -3.71352702e-01 -1.17440128e+00 1.20075814e-01
4.82578635e-01 -3.01939875e-01 -5.96598029e-01 -5.65852504e-03
5.34848094e-01 -4.38530713e-01 -1.20934772e+00 7.01875269e-01
6.15726054e-01 -1.15008283e+00 1.23620045e+00 -3.79542053e-01
4.65576798e-01 3.23947379e-03 3.23004276e-01 -1.12508535e+00
-3.49507302e-01 -2.54906118e-01 1.92345813e-01 8.21591437e-01
5.51870823e-01 -8.26566279e-01 5.96012771e-01 1.07288137e-01
-4.16609854e-01 -1.02529383e+00 -1.21787465e+00 -4.93320167e-01
8.87622684e-02 -5.55564880e-01 4.40066487e-01 9.53993142e-01
-1.56735912e-01 -6.90304190e-02 -1.70744270e-01 -5.58428392e-02
8.86294618e-02 -2.56001562e-01 1.55667290e-01 -1.41014457e+00
-1.63504139e-01 -7.29360461e-01 -6.48527205e-01 -2.12690920e-01
-4.09048647e-01 -1.01728582e+00 -3.25386047e-01 -1.60143054e+00
-4.73398179e-01 -5.65740824e-01 -5.28763354e-01 6.03034198e-01
-1.00064084e-01 3.28479797e-01 2.25101233e-01 2.37381235e-02
2.78097153e-01 -1.27918109e-01 1.40680444e+00 1.41514733e-01
-3.79061043e-01 4.37481970e-01 -2.60430157e-01 5.30826211e-01
1.06525278e+00 -5.04562736e-01 -4.46926236e-01 -2.31880248e-01
2.75134742e-01 4.17632967e-01 3.98860306e-01 -1.37788558e+00
1.06077529e-01 4.14295018e-01 6.65439188e-01 -4.42177236e-01
1.19649924e-01 -8.02759707e-01 3.48082870e-01 1.05887318e+00
-3.02125245e-01 6.27793968e-02 5.26718915e-01 -8.56280252e-02
-3.82191300e-01 -4.80542004e-01 8.28276634e-01 -2.03266010e-01
-1.74388856e-01 -1.94613904e-01 -7.74366915e-01 -1.80484533e-01
5.67918777e-01 -6.51862204e-01 1.10607613e-02 -1.55679256e-01
-9.74173486e-01 -3.42898130e-01 -4.92193967e-01 2.64751107e-01
5.14678955e-01 -1.07686865e+00 -8.30188036e-01 2.97884643e-01
-4.31730270e-01 -1.52852684e-02 3.96564543e-01 1.50420845e+00
-1.11307895e+00 5.30225635e-01 -4.83371168e-01 -7.59487867e-01
-1.37347388e+00 1.52574554e-01 9.14168775e-01 -1.23484768e-01
-8.92552972e-01 7.90723622e-01 -3.49519968e-01 -1.38035968e-01
3.33857626e-01 -7.60573089e-01 -7.75054395e-01 3.04925561e-01
3.57737303e-01 5.59422016e-01 5.96649170e-01 -2.15768293e-01
-3.44456613e-01 4.28253561e-01 2.69103795e-01 3.31197381e-01
1.60903609e+00 3.73430789e-01 -1.82136312e-01 3.05042356e-01
1.20158088e+00 -3.60537410e-01 -5.00380397e-01 2.27962896e-01
4.74587940e-02 -4.14124876e-02 2.61114597e-01 -8.38030577e-01
-1.47675538e+00 1.10522652e+00 9.66033697e-01 7.53137946e-01
1.49931812e+00 -3.64015669e-01 4.66322958e-01 1.68875441e-01
2.04387913e-03 -8.75038385e-01 -1.76337332e-01 3.22422206e-01
8.15830827e-01 -5.60508251e-01 -1.90070540e-01 -1.88326269e-01
-4.55538809e-01 1.71292651e+00 1.95247859e-01 -2.59719998e-01
9.00968730e-01 3.34602416e-01 1.90606043e-01 -3.84866685e-01
-4.59600806e-01 -1.26549229e-01 4.02505368e-01 4.98594493e-01
8.63123178e-01 -1.08687729e-01 -6.57606781e-01 6.94476545e-01
-4.33891356e-01 2.12218568e-01 4.02276844e-01 6.70352757e-01
-4.26292241e-01 -7.88913846e-01 -4.79371071e-01 7.55385995e-01
-9.67441916e-01 -1.22504443e-01 -3.15436325e-03 8.48020971e-01
6.99902833e-01 8.19675446e-01 5.34647480e-02 -2.98024118e-01
5.22967279e-01 4.35644507e-01 2.67303526e-01 -3.77484202e-01
-1.35250366e+00 6.12792820e-02 9.35507342e-02 -5.72923981e-02
-4.79654819e-01 -4.92937952e-01 -1.17384863e+00 -5.80030978e-02
-5.05560696e-01 2.89673865e-01 7.40881622e-01 8.04074645e-01
8.34009126e-02 1.13415170e+00 4.68846351e-01 -5.25189579e-01
-3.92211795e-01 -1.30708730e+00 -4.53109115e-01 9.00626406e-02
4.46653455e-01 -3.94188076e-01 -5.05944073e-01 6.42933324e-02] | [14.30984115600586, 3.2981154918670654] |
3dc63042-1b5d-4fe2-b2cf-05fbe170a869 | stimuli-sensitive-hawkes-processes-for | 2102.00089 | null | https://arxiv.org/abs/2102.00089v1 | https://arxiv.org/pdf/2102.00089v1.pdf | Stimuli-Sensitive Hawkes Processes for Personalized Student Procrastination Modeling | Student procrastination and cramming for deadlines are major challenges in online learning environments, with negative educational and well-being side effects. Modeling student activities in continuous time and predicting their next study time are important problems that can help in creating personalized timely interventions to mitigate these challenges. However, previous attempts on dynamic modeling of student procrastination suffer from major issues: they are unable to predict the next activity times, cannot deal with missing activity history, are not personalized, and disregard important course properties, such as assignment deadlines, that are essential in explaining the cramming behavior. To resolve these problems, we introduce a new personalized stimuli-sensitive Hawkes process model (SSHP), by jointly modeling all student-assignment pairs and utilizing their similarities, to predict students' next activity times even when there are no historical observations. Unlike regular point processes that assume a constant external triggering effect from the environment, we model three dynamic types of external stimuli, according to assignment availabilities, assignment deadlines, and each student's time management habits. Our experiments on two synthetic datasets and two real-world datasets show a superior performance of future activity prediction, comparing with state-of-the-art models. Moreover, we show that our model achieves a flexible and accurate parameterization of activity intensities in students. | ['Reza Feyzi Behnagh', 'Shaghayegh Sahebi', 'Siqian Zhao', 'Mengfan Yao'] | 2021-01-29 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [-8.07720721e-02 -3.20286632e-01 -6.03752792e-01 -1.93482727e-01
1.17334291e-01 -6.22919858e-01 2.68810868e-01 6.66622996e-01
-2.08856151e-01 6.01145148e-01 2.42408410e-01 -4.17086691e-01
-7.70405352e-01 -9.25531745e-01 -5.26637614e-01 -7.02165425e-01
1.14348069e-01 3.37151527e-01 3.56871873e-01 -1.74775332e-01
5.43144822e-01 3.23411286e-01 -2.00328588e+00 -1.48129776e-01
1.39412546e+00 1.96862325e-01 2.13070914e-01 7.14191616e-01
-1.75421685e-01 1.20676136e+00 -7.90960968e-01 -1.70172736e-01
-1.44429594e-01 -2.00002253e-01 -5.36112666e-01 1.33738503e-01
8.21783021e-02 -2.71568477e-01 -6.27633631e-01 6.66154206e-01
6.68946505e-02 8.72158766e-01 4.54964012e-01 -1.59310591e+00
-6.50219500e-01 5.79041600e-01 -5.25910795e-01 5.50656736e-01
5.36899984e-01 4.97859977e-02 4.89224911e-01 1.05916806e-01
1.43233135e-01 1.08664978e+00 3.78166437e-01 5.34382582e-01
-1.30256391e+00 -6.71236336e-01 7.61208892e-01 7.79623032e-01
-9.58079457e-01 4.24710751e-01 7.08265305e-01 -5.40462554e-01
5.44113517e-01 4.36160147e-01 1.00175345e+00 1.45228100e+00
5.52507222e-01 1.13624799e+00 1.07390332e+00 -3.36249590e-01
4.18766975e-01 1.20740123e-01 6.34998977e-01 1.49035797e-01
9.28946361e-02 -7.89709985e-02 -7.96779454e-01 -2.05134034e-01
4.77577955e-01 8.00381243e-01 -1.72826648e-01 -3.52343112e-01
-6.26138926e-01 4.29340184e-01 -2.69482225e-01 2.06798643e-01
-3.98313373e-01 -3.61031331e-02 -1.72373608e-01 4.48510766e-01
3.02359909e-01 2.33209431e-01 -5.37371397e-01 -7.41407096e-01
-6.23695970e-01 5.54775596e-01 1.00179434e+00 9.68019724e-01
5.23272634e-01 -2.76711464e-01 -4.58344877e-01 5.64304948e-01
7.10869133e-02 3.09452057e-01 8.52747083e-01 -8.30058753e-01
1.90990016e-01 7.94334650e-01 3.50263715e-01 -8.11907649e-01
-3.12799424e-01 -5.29999435e-01 -1.48754179e-01 -2.56019354e-01
7.43250370e-01 7.49984607e-02 -6.69458628e-01 1.79522002e+00
3.94318819e-01 9.63495851e-01 -2.64799088e-01 3.95714045e-01
2.59950280e-01 6.33100629e-01 4.52686608e-01 -4.86223578e-01
1.32886195e+00 -9.91099894e-01 -1.13555503e+00 -1.16631240e-01
8.48230243e-01 -7.18401432e-01 1.15039802e+00 7.94921517e-01
-1.06513977e+00 -4.91731673e-01 -4.97056633e-01 3.95312935e-01
-2.84033865e-01 -6.14198923e-01 7.02237606e-01 1.00654471e+00
-5.09161055e-01 8.20330977e-01 -1.24236691e+00 -1.30384699e-01
7.36436769e-02 1.71745673e-01 3.40369612e-01 2.20880788e-02
-8.68942142e-01 5.48463047e-01 -1.87749714e-01 -4.48282510e-01
-1.17145181e+00 -1.43415177e+00 -4.21612740e-01 7.06269860e-01
7.70229578e-01 -4.25218254e-01 1.57611513e+00 -7.65205979e-01
-1.65589738e+00 1.15859285e-01 -1.52043581e-01 -8.64749849e-02
4.77456629e-01 -4.02181774e-01 -4.44240779e-01 -3.22001517e-01
-2.32618868e-01 -1.82722285e-01 1.90963030e-01 -8.65590692e-01
-6.28779650e-01 -6.07165456e-01 -1.93239957e-01 4.72660929e-01
-8.46250713e-01 -1.17150709e-01 -1.69173166e-01 -2.71431029e-01
-1.91771165e-02 -9.15118933e-01 -4.12137806e-01 -8.15771997e-01
1.28922552e-01 -6.28120959e-01 8.05977106e-01 -2.45867491e-01
1.76144588e+00 -1.98673487e+00 -2.99738526e-01 1.36120692e-01
-1.97205581e-02 2.51052737e-01 -9.79608893e-02 7.73860276e-01
1.45460799e-01 1.45975137e-02 5.87760568e-01 -2.21492782e-01
2.87670456e-02 4.43285257e-01 -3.40428531e-01 1.37478188e-01
-5.14129579e-01 3.96700621e-01 -1.27167380e+00 -1.81787580e-01
3.64317238e-01 1.56862915e-01 -4.48650867e-01 4.92846698e-01
-2.15485126e-01 4.06959653e-01 -6.42038047e-01 4.83100086e-01
5.92229068e-01 -1.50280565e-01 2.83749253e-01 1.14654422e+00
-3.17457587e-01 3.61912221e-01 -1.25734341e+00 1.01414287e+00
-3.81429940e-01 2.72685051e-01 -3.50407869e-01 -7.53979683e-01
7.36544073e-01 3.32375109e-01 7.89372623e-01 -9.18340564e-01
-2.61479020e-01 -1.98843166e-01 6.68238243e-03 -6.68110728e-01
5.29848337e-01 2.30090439e-01 1.83808208e-01 7.41788507e-01
-1.17730558e-01 1.18103608e-01 2.99897879e-01 3.13421339e-01
1.46677601e+00 9.83685330e-02 -9.11392793e-02 -6.63184077e-02
2.82189906e-01 -3.39493930e-01 9.97572601e-01 9.70008194e-01
-6.64392829e-01 1.78149268e-02 9.00057793e-01 -6.19078755e-01
-4.90823954e-01 -1.01174688e+00 2.23450378e-01 1.55872428e+00
2.75429338e-01 -3.05454701e-01 -7.00020432e-01 -6.23293221e-01
3.72650363e-02 1.10923231e+00 -6.38134241e-01 -4.54692215e-01
-3.86136413e-01 -8.43837023e-01 -3.96925360e-02 3.40006530e-01
8.25983379e-03 -9.39223588e-01 -4.95437175e-01 4.74386483e-01
-2.21145228e-01 -7.58752704e-01 -5.73877513e-01 -3.17584388e-02
-9.55534756e-01 -1.17724133e+00 -1.03641570e-01 -4.43397939e-01
6.50098801e-01 7.09106922e-01 1.10348582e+00 2.45729193e-01
-1.83659762e-01 7.60944784e-01 1.01319939e-01 -9.31119978e-01
-8.03090483e-02 -8.84200558e-02 3.21075320e-01 -1.00726001e-01
9.27739382e-01 -8.02627623e-01 -9.21492815e-01 3.98633748e-01
-8.89778793e-01 -9.85705703e-02 6.93162680e-02 5.93217492e-01
3.99372280e-01 4.00043130e-01 5.87998688e-01 -1.04780281e+00
6.24603987e-01 -6.07231915e-01 -6.33148074e-01 4.31080461e-01
-1.22154737e+00 -2.94980228e-01 6.19035065e-01 -1.08278799e+00
-1.55868435e+00 -2.06495091e-01 4.35549945e-01 -3.13845783e-01
-5.10847747e-01 1.90599039e-01 -4.16559074e-03 2.57888347e-01
4.07877266e-01 3.46159488e-01 -3.36095691e-01 -4.13478523e-01
-2.23649472e-01 2.49844879e-01 1.41851828e-01 -8.02954376e-01
5.29794335e-01 1.63088933e-01 1.93904728e-01 -5.32283962e-01
-1.22101760e+00 -8.52287114e-01 -4.24230576e-01 -4.99642283e-01
2.80535579e-01 -9.60225344e-01 -1.49057901e+00 6.94937170e-01
-3.99161339e-01 -5.44883490e-01 -1.40058458e-01 6.93861842e-01
-6.04155838e-01 2.30441809e-01 -8.66234064e-01 -1.21460509e+00
3.96919638e-01 -9.18732822e-01 1.19582318e-01 1.04889321e+00
-1.37847168e-02 -1.59046912e+00 5.37683785e-01 1.05314755e+00
1.55201808e-01 -9.23831537e-02 9.37434077e-01 -8.91672492e-01
-8.42939675e-01 -4.24097106e-02 5.47079861e-01 -1.01181231e-02
-2.83672065e-02 1.55790254e-01 -8.38276982e-01 -1.31322488e-01
5.90801723e-02 5.47305718e-02 2.43559048e-01 3.95204276e-01
1.61790168e+00 -4.58031952e-01 -3.05273324e-01 3.32975328e-01
1.19130349e+00 5.03694117e-01 3.82561475e-01 5.70990682e-01
4.79065239e-01 7.80737162e-01 9.68056738e-01 9.19514000e-01
7.77688146e-01 4.77062255e-01 6.44743741e-01 4.00398254e-01
4.47856337e-01 -8.13607275e-01 5.47043443e-01 8.59300613e-01
-1.20789766e-01 -4.15544480e-01 -9.51716185e-01 8.21676016e-01
-2.16019869e+00 -1.34819376e+00 -7.01627016e-01 2.65427351e+00
7.78400779e-01 5.63634932e-02 9.15094465e-02 -1.26973167e-01
3.15274686e-01 -5.58259338e-02 -4.96996969e-01 -6.34269059e-01
3.91790599e-01 1.99308991e-01 5.84037960e-01 5.13910830e-01
-5.20654380e-01 7.48426259e-01 6.39594173e+00 6.82889581e-01
-4.88392055e-01 -1.57871675e-02 5.62257469e-01 -4.38007355e-01
-4.80992377e-01 1.50494993e-01 -1.14804292e+00 7.14851201e-01
1.52835250e+00 -3.71145546e-01 4.34243321e-01 7.80159652e-01
9.16859210e-01 -2.50146627e-01 -1.19417906e+00 4.97351646e-01
-5.31995967e-02 -7.07469821e-01 -3.27390492e-01 2.60046959e-01
1.21616852e+00 -6.03624880e-01 1.12196453e-01 8.54763448e-01
8.22306514e-01 -7.77410924e-01 2.55084306e-01 8.79507959e-01
-6.69729710e-01 -1.12537646e+00 4.01229173e-01 1.07582521e+00
-5.66785693e-01 -5.25598347e-01 -2.67711103e-01 -6.90346897e-01
-4.37858909e-01 4.70395505e-01 -7.86966264e-01 1.14560947e-01
6.38663590e-01 4.06707644e-01 -5.45912743e-01 1.24902821e+00
-4.29993749e-01 1.22122931e+00 -6.95299879e-02 -2.02234432e-01
7.18735307e-02 -3.48594844e-01 2.16351271e-01 6.56199276e-01
5.10626256e-01 2.83599466e-01 3.69228661e-01 7.92175651e-01
2.41856188e-01 1.65113553e-01 -3.33464712e-01 6.62362427e-02
7.56356299e-01 1.18125212e+00 -5.28264701e-01 -1.78066090e-01
-4.80448246e-01 6.68878257e-01 1.06411211e-01 5.70120573e-01
-8.61816764e-01 -1.60804495e-01 1.18568647e+00 4.03879881e-01
-1.33877173e-01 -2.69122511e-01 -2.64388233e-01 -1.14439905e+00
-3.16866875e-01 -7.06160843e-01 3.80412102e-01 -4.51580286e-01
-9.67314184e-01 -5.91715813e-01 -1.12058468e-01 -1.00845325e+00
-3.85779254e-02 -3.39460105e-01 -1.17460227e+00 6.57530844e-01
-1.60913384e+00 -6.86637402e-01 -3.76777053e-01 7.55572438e-01
5.33890903e-01 3.48709345e-01 4.20046955e-01 1.98619559e-01
-9.76012290e-01 4.31029320e-01 3.92646819e-01 -5.89471877e-01
9.02525008e-01 -1.52138376e+00 -1.89419359e-01 6.84247732e-01
-7.25342855e-02 6.05969012e-01 9.72467422e-01 -6.11341476e-01
-1.45560932e+00 -7.62776077e-01 1.08211458e+00 -8.66055071e-01
8.34894240e-01 -1.97732270e-01 -1.40310633e+00 7.66809762e-01
7.59858266e-02 -3.93677503e-01 1.37875748e+00 5.69291294e-01
1.52122065e-01 -5.74338324e-02 -7.13098943e-01 7.21489966e-01
8.95580173e-01 -5.76198101e-03 -4.70342606e-01 6.09278917e-01
5.58084726e-01 -4.01641846e-01 -1.02905822e+00 -1.83161870e-01
2.59186774e-01 -9.55738366e-01 7.36435592e-01 -9.90928292e-01
4.52021062e-01 1.84350371e-01 6.08080149e-01 -1.40563416e+00
-6.24257803e-01 -7.43347168e-01 -5.05127490e-01 1.33497238e+00
-4.66850363e-02 -2.58612007e-01 1.24418604e+00 1.35127878e+00
-1.68796703e-01 -6.41317904e-01 -5.57983994e-01 -8.61824155e-01
-1.20079495e-01 -2.96169639e-01 9.43821013e-01 1.19403481e+00
3.87127250e-01 -1.38275996e-01 -5.05239666e-01 3.85854810e-01
7.48803854e-01 1.60856500e-01 8.69362354e-01 -1.54780710e+00
-2.67921209e-01 -4.15229648e-01 -3.20277847e-02 -9.36093450e-01
2.09883377e-01 -7.95931891e-02 -2.06400573e-01 -1.13312066e+00
6.11064672e-01 -3.19705456e-01 -7.66879201e-01 4.04990911e-01
-5.88248491e-01 -7.28947341e-01 -1.05146646e-01 -1.10251285e-01
-9.00893390e-01 5.94765961e-01 1.55224180e+00 3.71915162e-01
-7.91533828e-01 5.82991362e-01 -5.85116684e-01 7.05479085e-01
6.61816955e-01 -5.92439711e-01 -7.87778020e-01 1.47991031e-02
3.17024797e-01 5.19316971e-01 6.93718791e-02 -9.81398165e-01
4.55760837e-01 -1.48984230e+00 1.74312755e-01 -4.64266390e-01
1.02577791e-01 -8.37639451e-01 3.73084657e-02 5.77788830e-01
-5.03414392e-01 -8.92547742e-02 -2.42874157e-02 9.48488712e-01
-1.64634082e-03 -4.65925187e-01 4.06528443e-01 -6.80072010e-02
-3.92006576e-01 5.89964926e-01 -9.88710344e-01 -1.93671480e-01
1.54311502e+00 -2.40625381e-01 -5.64883590e-01 -6.69876754e-01
-7.75410175e-01 8.47663939e-01 3.46475452e-01 8.04144740e-01
2.48022854e-01 -1.06129229e+00 -1.68787807e-01 2.01351568e-01
4.19154838e-02 -2.01132923e-01 9.43775535e-01 7.62375832e-01
-2.22261801e-01 4.75583047e-01 -2.86268085e-01 -3.01360756e-01
-1.49227083e+00 7.95566201e-01 9.60154384e-02 -7.27762818e-01
-1.70963258e-01 4.52880621e-01 2.75764048e-01 -3.86535466e-01
4.67110753e-01 -3.20404559e-01 -4.75150138e-01 2.21393853e-02
6.80696845e-01 7.65470147e-01 -3.10183000e-02 1.37711361e-01
1.24750875e-01 -1.56826839e-01 -2.18148157e-01 3.31196308e-01
1.38124800e+00 -3.15845013e-01 3.57247949e-01 7.35704243e-01
1.28379822e-01 1.19899862e-01 -1.54368317e+00 -4.13500190e-01
1.37719959e-01 -8.63796592e-01 2.16496848e-02 -7.07567334e-01
-6.18495524e-01 8.67546797e-01 4.81908262e-01 3.97704720e-01
9.82747078e-01 -5.07808805e-01 6.70038223e-01 1.19819798e-01
1.75683409e-01 -1.57533967e+00 5.59610307e-01 5.20228088e-01
-6.51685521e-02 -1.12380028e+00 -2.51403570e-01 -3.27528447e-01
-4.90928531e-01 8.21992576e-01 1.46222878e+00 3.97179246e-01
6.35185897e-01 -4.02102955e-02 1.02312760e-02 1.96982741e-01
-1.67087710e+00 2.97668904e-01 -8.35724398e-02 4.09196943e-01
5.84824204e-01 5.21665692e-01 -3.82616311e-01 8.82497847e-01
-1.23019218e-01 4.62330543e-02 1.25460458e+00 1.05454886e+00
-6.36187851e-01 -1.30285871e+00 -7.43129015e-01 2.47082636e-01
-7.19535053e-01 4.49349403e-01 -5.78599609e-02 4.63305861e-01
4.23151590e-02 9.90985453e-01 1.83492437e-01 -1.21862806e-01
3.72515857e-01 4.22609240e-01 3.43315750e-01 -7.37091124e-01
-1.00636041e+00 -2.04770967e-01 -5.46088815e-01 -4.52287138e-01
7.59687200e-02 -9.51102972e-01 -9.51715529e-01 -8.24055552e-01
-2.79669046e-01 1.99083298e-01 5.13941407e-01 8.88872206e-01
2.72624642e-01 7.33059585e-01 8.64733338e-01 -7.26379380e-02
-7.57927060e-01 -7.92168379e-01 -8.37253451e-01 6.79695249e-01
1.25966832e-01 -6.73482537e-01 -4.46506530e-01 -1.20851815e-01] | [10.118080139160156, 7.132706165313721] |
9fbae3ae-82fa-489c-b341-e9c9216da27d | forecasting-social-navigation-in-crowded | 1601.00998 | null | http://arxiv.org/abs/1601.00998v1 | http://arxiv.org/pdf/1601.00998v1.pdf | Forecasting Social Navigation in Crowded Complex Scenes | When humans navigate a crowed space such as a university campus or the
sidewalks of a busy street, they follow common sense rules based on social
etiquette. In this paper, we argue that in order to enable the design of new
algorithms that can take fully advantage of these rules to better solve tasks
such as target tracking or trajectory forecasting, we need to have access to
better data in the first place. To that end, we contribute the very first large
scale dataset (to the best of our knowledge) that collects images and videos of
various types of targets (not just pedestrians, but also bikers, skateboarders,
cars, buses, golf carts) that navigate in a real-world outdoor environment such
as a university campus. We present an extensive evaluation where different
methods for trajectory forecasting are evaluated and compared. Moreover, we
present a new algorithm for trajectory prediction that exploits the complexity
of our new dataset and allows to: i) incorporate inter-class interactions into
trajectory prediction models (e.g, pedestrian vs bike) as opposed to just
intra-class interactions (e.g., pedestrian vs pedestrian); ii) model the degree
to which the social forces are regulating an interaction. We call the latter
"social sensitivity"and it captures the sensitivity to which a target is
responding to a certain interaction. An extensive experimental evaluation
demonstrates the effectiveness of our novel approach. | ['Silvio Savarese', 'Eli Wu', 'John Doherty', 'Amir Sadeghian', 'Alexandre Alahi', 'Bryan Anenberg', 'Alexandre Robicquet'] | 2016-01-05 | null | null | null | null | ['social-navigation'] | ['robots'] | [-1.17641211e-01 -2.08070353e-01 -3.00875485e-01 -2.57470608e-01
-1.34007573e-01 -6.13415062e-01 8.80856812e-01 3.80152911e-01
-3.35377365e-01 8.18295121e-01 2.47984365e-01 -4.71871883e-01
-4.48021501e-01 -1.09092534e+00 -6.31516099e-01 -5.91189086e-01
-1.89196512e-01 3.01986694e-01 8.56325090e-01 -7.02756763e-01
2.06024349e-01 6.98522329e-01 -1.86257231e+00 7.85270110e-02
8.84324610e-01 5.85722387e-01 2.69915968e-01 6.91290319e-01
2.76009142e-01 7.57232308e-01 -1.44069970e-01 -3.14650685e-01
2.91566644e-02 -2.18435675e-01 -7.32761562e-01 -1.79531932e-01
5.25494106e-02 -4.22469787e-02 -4.09922749e-01 7.55801737e-01
1.66776493e-01 5.91385126e-01 7.25699008e-01 -1.49847627e+00
-4.04283684e-03 5.77502608e-01 -1.56550184e-01 4.30985451e-01
6.88952148e-01 2.45387062e-01 6.94766164e-01 -3.82119983e-01
7.68695951e-01 1.23051429e+00 5.48505127e-01 2.46690989e-01
-1.18126070e+00 -4.69789863e-01 6.89093769e-01 5.21954417e-01
-1.32485342e+00 -3.78270805e-01 4.78048235e-01 -6.99896812e-01
4.79414821e-01 4.74820405e-01 7.93629587e-01 1.36700463e+00
9.16178599e-02 8.07843745e-01 8.85126770e-01 -1.96568415e-01
3.26647222e-01 1.64014786e-01 5.67177773e-01 3.47796649e-01
1.45544916e-01 2.39996463e-01 -3.10641557e-01 -2.15298936e-01
1.12103008e-01 -1.24619715e-01 -3.08624655e-01 -3.50198597e-01
-1.22643673e+00 7.55089641e-01 4.52412575e-01 4.53828841e-01
-2.73424566e-01 -3.64974812e-02 1.26965642e-01 1.14542946e-01
2.26814702e-01 -6.00796416e-02 -8.40296373e-02 -4.18263763e-01
-5.50235629e-01 5.68789303e-01 8.93713057e-01 9.17845845e-01
6.64406002e-01 -3.83145332e-01 -6.70395494e-02 4.90927458e-01
2.12313682e-01 4.20663059e-01 1.79465815e-01 -6.66782558e-01
3.80005091e-01 6.08227134e-01 5.97735226e-01 -1.39851975e+00
-7.19290078e-01 -1.06628068e-01 -5.69369793e-01 9.50878412e-02
6.87502801e-01 -4.29420099e-02 -3.39047462e-01 1.86632788e+00
6.07360303e-01 4.65248644e-01 -8.12720433e-02 7.15278029e-01
5.18524408e-01 5.59957266e-01 8.28775913e-02 -8.23372379e-02
1.08374298e+00 -6.88757479e-01 -4.21294838e-01 -3.13026868e-02
9.88242328e-01 -5.32762885e-01 8.11842918e-01 1.64152384e-01
-7.29787350e-01 -6.40535057e-01 -7.46761620e-01 3.98773789e-01
-7.60979652e-01 -2.57385284e-01 3.74829352e-01 8.43853652e-01
-8.29095304e-01 5.99439085e-01 -8.88701618e-01 -9.74771321e-01
1.27632350e-01 1.83607832e-01 -2.31098980e-01 3.03680059e-02
-1.13930357e+00 1.03432572e+00 1.44917905e-01 -2.13497192e-01
-6.66846454e-01 -6.09134555e-01 -6.71671152e-01 -8.64569694e-02
5.02672315e-01 -5.05677760e-01 8.82496297e-01 -6.89870000e-01
-1.16523969e+00 5.93654454e-01 -1.77225441e-01 -3.90826792e-01
7.71523356e-01 -8.51365477e-02 -4.85331535e-01 -2.71148801e-01
2.43841678e-01 3.20724338e-01 1.80617139e-01 -1.43387938e+00
-1.11115623e+00 -4.25459892e-01 4.40732449e-01 8.90456960e-02
-9.50079486e-02 8.19097012e-02 -1.78065728e-02 -2.37102136e-01
-2.01108843e-01 -1.34329605e+00 -3.66233826e-01 -3.26979309e-01
-5.11725545e-01 -3.85882348e-01 9.10976589e-01 -2.24678323e-01
1.47215939e+00 -2.07989383e+00 1.43654078e-01 4.93514478e-01
3.08458842e-02 2.84959793e-01 8.09013397e-02 8.59532297e-01
1.40046790e-01 3.11371684e-02 -3.86902988e-02 -1.52100340e-01
8.52500275e-02 2.49033377e-01 -3.30689043e-01 5.67156017e-01
-4.09924835e-01 5.45001686e-01 -1.26206112e+00 -2.90942013e-01
4.01423782e-01 2.22743109e-01 -3.74331921e-01 -2.54820324e-02
1.16195507e-01 7.73007751e-01 -6.59313917e-01 4.32504296e-01
5.90980172e-01 1.36934623e-01 1.51380509e-01 2.75564551e-01
-7.52130747e-01 1.15524262e-01 -1.26705027e+00 9.57743168e-01
-2.68700033e-01 7.27250993e-01 -1.68385655e-01 -8.13841879e-01
7.23621368e-01 -7.90073201e-02 5.21043599e-01 -4.28447992e-01
6.98754787e-02 9.94893089e-02 1.18180074e-01 -5.07430494e-01
6.18399680e-01 3.77691686e-01 -1.71254173e-01 4.57288891e-01
-5.03512442e-01 3.32093835e-01 4.30562556e-01 1.53312385e-01
1.26712692e+00 -2.87989830e-03 3.61857027e-01 -4.43247080e-01
6.19256258e-01 2.35204026e-01 3.45604539e-01 9.96111453e-01
-4.56988901e-01 6.68412596e-02 3.29470277e-01 -6.00986719e-01
-7.35168695e-01 -1.04500473e+00 3.72807607e-02 1.27808213e+00
5.09988487e-01 -3.55583400e-01 -6.61871493e-01 -5.15549123e-01
2.04871837e-02 8.94089341e-01 -5.23604929e-01 -1.61709994e-01
-7.60173738e-01 -7.59108186e-01 4.03046340e-01 3.54957730e-01
2.82138646e-01 -7.28754938e-01 -6.55416608e-01 1.13778502e-01
-4.05178398e-01 -1.24084854e+00 -1.43652603e-01 -1.93736225e-01
-4.29302067e-01 -1.21623611e+00 -2.01082274e-01 -3.90071124e-01
4.40041929e-01 6.64772391e-01 7.95989156e-01 1.64931893e-01
5.61577082e-02 6.46991193e-01 -6.94360375e-01 -2.66000599e-01
-2.08443403e-01 1.82037026e-01 3.62193167e-01 2.79641688e-01
3.69176567e-01 -5.98309457e-01 -5.81454933e-01 8.87472689e-01
-5.02660155e-01 -4.37111333e-02 1.88439488e-02 3.75131518e-01
6.82345480e-02 1.99502811e-01 3.76069576e-01 -5.93600929e-01
6.08072579e-01 -8.95269394e-01 -6.11191034e-01 2.38513350e-01
-1.90366179e-01 -4.68744636e-01 4.26878333e-01 -5.49469233e-01
-9.06818867e-01 9.76825505e-02 -1.07431129e-01 1.61193371e-01
-4.97830659e-01 3.20826858e-01 -1.13345429e-01 -2.32751176e-01
6.29705787e-01 1.13329083e-01 -3.12814862e-01 -1.53635368e-01
4.54076082e-01 6.40879810e-01 2.14283317e-01 -6.26286924e-01
8.47573340e-01 7.73095369e-01 1.49933562e-01 -1.04839849e+00
-4.83786345e-01 -5.76584816e-01 -9.52206671e-01 -8.65057826e-01
8.79211962e-01 -5.76978922e-01 -1.32033122e+00 4.00140196e-01
-9.55158293e-01 -6.10802174e-01 8.38897377e-02 6.46411240e-01
-6.82657361e-01 3.11203152e-01 -2.20943660e-01 -1.19379902e+00
4.66692209e-01 -9.58381355e-01 7.91120529e-01 2.19431296e-01
-3.98433477e-01 -1.02332103e+00 1.38078988e-01 2.24627882e-01
2.34380424e-01 5.15374362e-01 7.42335677e-01 -6.67704821e-01
-7.45589614e-01 -9.43572354e-03 -1.36836037e-01 -3.14311922e-01
9.75043513e-03 1.52377203e-01 -5.57958603e-01 -1.54490829e-01
-3.69897991e-01 1.46542400e-01 6.26664579e-01 3.70962769e-01
7.87226260e-01 -1.05649449e-01 -1.02692485e+00 1.40515700e-01
1.05775499e+00 5.25596201e-01 6.62713587e-01 5.54366827e-01
5.22596657e-01 1.06128979e+00 9.54736173e-01 5.24820864e-01
9.19235528e-01 1.03293884e+00 4.35166329e-01 2.45210484e-01
1.00284904e-01 -2.63980597e-01 3.16851377e-01 5.35572112e-01
-3.43011618e-01 -5.06607831e-01 -1.18656325e+00 7.40582824e-01
-2.39777899e+00 -1.21362114e+00 -6.96069121e-01 2.39839649e+00
1.11663520e-01 2.43935511e-02 6.01187050e-01 1.55958802e-01
7.49066234e-01 -6.75248131e-02 -1.78904742e-01 -1.16648152e-01
1.51285917e-01 -4.62495655e-01 5.78764260e-01 8.06883693e-01
-1.10363400e+00 9.30045545e-01 6.20477152e+00 8.78672957e-01
-7.26761103e-01 -6.12238161e-02 3.90269101e-01 1.18843645e-01
4.34509441e-02 6.84728995e-02 -1.06776619e+00 6.79612398e-01
8.77180874e-01 -1.05977364e-01 4.47708547e-01 6.98000610e-01
5.00477195e-01 -4.20022488e-01 -1.06827378e+00 6.42952502e-01
-3.73299927e-01 -1.03232622e+00 -3.57421309e-01 2.54347295e-01
4.71214771e-01 -1.73811167e-01 -2.94235218e-02 2.99165010e-01
7.71693528e-01 -7.89790750e-01 8.13962281e-01 8.98314655e-01
-1.61856338e-01 -8.46380234e-01 4.79236484e-01 8.80070031e-01
-1.38485038e+00 -3.05381656e-01 -7.63073787e-02 -2.99020141e-01
4.28907812e-01 3.72073203e-01 -6.63407564e-01 6.23278975e-01
7.09719479e-01 6.86710835e-01 -3.46296400e-01 1.11767924e+00
-3.76113951e-02 4.49570954e-01 -3.43869388e-01 -3.54401439e-01
4.80936080e-01 -2.81071782e-01 9.12583470e-01 1.07577181e+00
5.83139777e-01 4.46353704e-01 3.52086604e-01 4.69544649e-01
6.73051775e-01 1.04396075e-01 -1.03181159e+00 2.85375774e-01
4.68889475e-01 1.00311971e+00 -7.34220505e-01 -1.14295609e-01
-3.37509274e-01 4.21725720e-01 2.90444344e-01 3.53707582e-01
-9.15774584e-01 -3.33611481e-02 1.02114880e+00 5.23545980e-01
7.90811479e-02 -3.89212966e-01 -1.50350213e-01 -9.09847617e-01
-2.95613199e-01 -4.09299582e-01 2.77669191e-01 -5.91721416e-01
-1.18417609e+00 5.35473704e-01 3.24170113e-01 -1.28087950e+00
-2.41302162e-01 -4.30300176e-01 -7.19417036e-01 4.39102054e-01
-1.25634599e+00 -1.18326449e+00 -4.63722706e-01 6.86205566e-01
1.49416909e-01 -6.82207346e-02 5.03010273e-01 3.81862432e-01
-6.60216391e-01 1.37115046e-01 1.71364695e-01 -5.10609224e-02
3.86113554e-01 -9.35295403e-01 2.97097236e-01 7.16181815e-01
-1.95836142e-01 5.98405898e-01 9.08600926e-01 -7.42886305e-01
-1.02025127e+00 -1.03949034e+00 8.92560184e-01 -7.69198477e-01
9.14201975e-01 -5.00440359e-01 -5.37073851e-01 8.17082107e-01
-1.86625019e-01 -2.66117007e-01 8.70779455e-01 4.00865257e-01
8.58704895e-02 2.56444812e-02 -1.01038909e+00 1.03378892e+00
1.35255742e+00 -7.76923746e-02 -2.27026850e-01 4.07908380e-01
3.88891995e-01 -2.42479086e-01 -5.75029433e-01 3.29995424e-01
6.84526742e-01 -1.24481606e+00 1.07344413e+00 -7.71637976e-01
1.19815424e-01 -3.39527994e-01 -3.86163354e-01 -1.47968531e+00
-4.14683163e-01 -4.10268515e-01 9.20069441e-02 1.05307972e+00
8.11466053e-02 -7.26124346e-01 6.56598926e-01 6.52944803e-01
-2.06736833e-01 -7.35377610e-01 -1.05474102e+00 -9.59556878e-01
-2.28043459e-02 -6.03675842e-01 8.71905804e-01 8.02455604e-01
1.97147027e-01 -1.30395457e-01 -4.88331765e-01 2.99495876e-01
6.15951061e-01 -1.05235353e-03 1.20703197e+00 -1.46267593e+00
1.67434320e-01 -4.94856715e-01 -7.07435310e-01 -1.04706872e+00
2.65094906e-01 -6.66296303e-01 -8.77295360e-02 -1.29539323e+00
-5.71188889e-03 -8.73640239e-01 -1.50970016e-02 7.76501372e-02
5.12960665e-02 -1.08603304e-02 2.41280898e-01 7.56109282e-02
-6.04779124e-01 2.82972485e-01 9.31750894e-01 8.58597159e-02
-3.91871423e-01 4.73567575e-01 -5.30972183e-01 9.53290462e-01
7.55128443e-01 -4.67746079e-01 -2.25553945e-01 9.18919034e-03
2.44399920e-01 1.99031621e-01 5.99512935e-01 -1.15020716e+00
4.61554617e-01 -7.00027645e-01 -2.78982203e-02 -6.26138270e-01
4.42584336e-01 -8.62472832e-01 2.39992648e-01 5.99053919e-01
-1.95239320e-01 -2.30925068e-01 7.32094795e-02 7.75357723e-01
5.95604256e-02 -1.17810905e-01 5.92044115e-01 5.21690622e-02
-8.04183364e-01 2.54916161e-01 -9.89123940e-01 -2.87447870e-01
1.63500035e+00 -2.52710253e-01 -5.72435498e-01 -6.83675170e-01
-9.42282438e-01 4.52062100e-01 4.45624113e-01 5.62987089e-01
3.05602640e-01 -1.44054663e+00 -5.14362514e-01 -1.83905482e-01
2.27967605e-01 -6.77064598e-01 3.68536562e-01 1.09073222e+00
-2.11901277e-01 4.51740324e-01 -3.05800527e-01 -5.77428222e-01
-1.43924475e+00 5.52719951e-01 1.22268893e-01 -2.98837215e-01
-4.51225072e-01 4.39780325e-01 2.92850763e-01 -5.31886518e-01
-8.62020347e-03 -2.33877987e-01 -5.38152575e-01 2.45372877e-01
5.01673460e-01 9.43608701e-01 -2.47279018e-01 -1.25760818e+00
-5.12752950e-01 4.23448503e-01 2.53110588e-01 5.23031689e-02
1.06993985e+00 -3.93318027e-01 6.22646362e-02 6.28902256e-01
6.54219508e-01 5.09024039e-02 -9.64848459e-01 -1.72937345e-02
9.89922136e-02 -8.22902620e-01 -2.97458589e-01 -4.69668150e-01
-5.41052461e-01 5.38243830e-01 5.23561895e-01 6.56638801e-01
7.94509470e-01 -4.90461923e-02 8.51359427e-01 5.14065146e-01
8.70965123e-01 -1.12486506e+00 -4.87254821e-02 5.21958351e-01
5.11540651e-01 -1.14984894e+00 -2.20520496e-01 -5.90741754e-01
-6.03938580e-01 8.44459534e-01 4.68092948e-01 -8.75396356e-02
8.50418627e-01 1.93563476e-03 -2.63733685e-01 -8.82638805e-03
-6.30995214e-01 -8.01016986e-01 1.82947904e-01 9.01641548e-01
-1.71699040e-02 5.34479618e-01 -4.29430962e-01 5.26765525e-01
-2.95108110e-01 -1.52166590e-01 7.29448438e-01 8.69387209e-01
-7.41814494e-01 -1.14884365e+00 -6.01896822e-01 2.70062774e-01
9.60653946e-02 4.36852247e-01 -5.08180559e-01 9.70196903e-01
4.06611383e-01 1.41064596e+00 -1.38496995e-01 -6.19633615e-01
6.09957933e-01 -1.70568109e-01 1.62047535e-01 -3.15853715e-01
-6.00200176e-01 -4.55763489e-01 2.31856421e-01 -5.68691671e-01
-5.23893237e-01 -9.79578495e-01 -9.60664153e-01 -8.02359343e-01
-5.60070686e-02 5.79212606e-02 5.02777636e-01 9.64141548e-01
1.68627992e-01 3.13722432e-01 5.97487211e-01 -8.56999338e-01
6.15573265e-02 -6.84474468e-01 -5.28242707e-01 3.57639134e-01
2.27433249e-01 -1.09325981e+00 -3.39085698e-01 -2.12519035e-01] | [5.932242393493652, 0.9466549158096313] |
40ee985b-6812-4316-b55a-407d1a253cfc | patchnet-unsupervised-object-discovery-based | 2106.08599 | null | https://arxiv.org/abs/2106.08599v1 | https://arxiv.org/pdf/2106.08599v1.pdf | PatchNet: Unsupervised Object Discovery based on Patch Embedding | We demonstrate that frequently appearing objects can be discovered by training randomly sampled patches from a small number of images (100 to 200) by self-supervision. Key to this approach is the pattern space, a latent space of patterns that represents all possible sub-images of the given image data. The distance structure in the pattern space captures the co-occurrence of patterns due to the frequent objects. The pattern space embedding is learned by minimizing the contrastive loss between randomly generated adjacent patches. To prevent the embedding from learning the background, we modulate the contrastive loss by color-based object saliency and background dissimilarity. The learned distance structure serves as object memory, and the frequent objects are simply discovered by clustering the pattern vectors from the random patches sampled for inference. Our image representation based on image patches naturally handles the position and scale invariance property that is crucial to multi-object discovery. The method has been proven surprisingly effective, and successfully applied to finding multiple human faces and bodies from natural images. | ['Patrick Bangert', 'Jae Oh Woo', 'Sima Didari', 'Heng Hao', 'Hankyu Moon'] | 2021-06-16 | null | null | null | null | ['multi-object-discovery'] | ['computer-vision'] | [ 6.30064845e-01 1.25663176e-01 -3.77455652e-01 -3.63356769e-01
-1.89863220e-01 -2.81807601e-01 5.18240213e-01 -2.41773188e-01
-1.09737009e-01 4.62337583e-01 -8.11124071e-02 5.60349762e-01
-2.26762280e-01 -6.76978588e-01 -1.12954891e+00 -1.12820888e+00
-4.02442813e-01 3.72139007e-01 1.83707565e-01 2.85830498e-01
4.40952837e-01 6.51538193e-01 -2.43655324e+00 6.46736324e-01
4.04180348e-01 1.04722834e+00 6.78527832e-01 3.45953465e-01
-7.48389289e-02 5.96939147e-01 -4.48053956e-01 -7.79784005e-03
4.48135078e-01 -6.83606625e-01 -5.53248405e-01 6.94298089e-01
9.03433383e-01 1.28217518e-01 -5.13102040e-02 1.35416996e+00
2.42711365e-01 1.94767982e-01 7.56843507e-01 -1.29907763e+00
-1.05946183e+00 2.52071936e-02 -8.38145554e-01 4.97753382e-01
2.95668155e-01 -1.55424355e-02 1.25794911e+00 -1.22689044e+00
8.03654373e-01 1.52602947e+00 1.83922395e-01 7.08726704e-01
-1.67081821e+00 -4.99122262e-01 1.01706371e-01 6.16050839e-01
-1.61465180e+00 -5.25634110e-01 9.95307505e-01 -5.15223622e-01
7.60307491e-01 4.44184214e-01 8.44172359e-01 9.48114634e-01
2.66891778e-01 9.76701319e-01 1.03657436e+00 -6.49623096e-01
3.52301419e-01 4.98941600e-01 -8.38969648e-02 1.04641163e+00
2.76201785e-01 1.31794035e-01 -7.05783606e-01 -2.68717527e-01
9.55449045e-01 2.97284871e-01 -1.80766225e-01 -8.16139162e-01
-1.05559695e+00 8.43633175e-01 5.11929214e-01 2.73505628e-01
-4.45872068e-01 -1.13585360e-01 -2.66108423e-01 3.91434252e-01
3.09938043e-01 5.28600395e-01 -3.93115819e-01 5.86717308e-01
-8.10441971e-01 2.17506066e-01 4.28097576e-01 9.41391349e-01
1.40115464e+00 -1.69512257e-01 -1.79487079e-01 8.80240679e-01
2.80711770e-01 6.85163498e-01 4.86173242e-01 -9.62094665e-01
-6.01338856e-02 7.68571436e-01 -1.48155410e-02 -1.57290578e+00
1.16941325e-01 -7.23546278e-03 -7.38369524e-01 3.80130798e-01
2.25584567e-01 4.25427109e-01 -8.20451081e-01 1.85123980e+00
5.79969525e-01 1.40878648e-01 -3.44905019e-01 7.76201069e-01
2.79517502e-01 4.90835190e-01 -2.41206229e-01 -3.90096545e-01
1.30203855e+00 -8.54146719e-01 -5.48434556e-01 -4.29734915e-01
-2.56676495e-01 -5.03821850e-01 1.20189416e+00 1.11448035e-01
-1.15988612e+00 -7.14989185e-01 -1.00111878e+00 2.65880823e-01
-4.19891357e-01 6.42028973e-02 3.33375990e-01 2.20228940e-01
-9.80512023e-01 6.94120944e-01 -5.25690973e-01 -2.33867705e-01
7.68980682e-01 4.05562490e-01 -4.32914376e-01 -4.12930436e-02
-6.70068443e-01 5.96049368e-01 1.25621244e-01 -2.11318303e-02
-9.23693359e-01 -5.32536209e-01 -9.14678454e-01 -6.94025382e-02
2.60683388e-01 -3.44631821e-01 5.43273747e-01 -1.74581683e+00
-1.09788787e+00 1.42491329e+00 -5.54322600e-01 -1.60862401e-01
-6.79640770e-02 8.01616013e-02 -3.35262239e-01 5.49831390e-01
4.89697456e-01 9.59249914e-01 1.83774984e+00 -1.42314839e+00
-5.88277042e-01 -5.14958262e-01 -5.28292179e-01 1.84250727e-01
-3.00082982e-01 -7.48983175e-02 -2.56094307e-01 -7.63006389e-01
1.98057652e-01 -8.32390308e-01 -6.98686987e-02 3.88306141e-01
-1.97075203e-01 -2.28036568e-01 8.29037070e-01 -1.90649450e-01
8.17761004e-01 -2.49447751e+00 3.62865359e-01 3.83441836e-01
2.21077934e-01 -4.33182508e-01 -4.41290468e-01 -3.12789790e-02
-2.53254026e-01 -1.32531166e-01 -1.76830426e-01 -5.31165414e-02
-2.19423786e-01 3.40957880e-01 -4.72865850e-01 7.30663300e-01
6.22056603e-01 8.76899064e-01 -9.27311182e-01 -5.53623974e-01
1.52299255e-01 2.05263361e-01 -4.72131371e-01 2.28610784e-01
-4.69820172e-01 2.08533272e-01 -2.75830775e-01 7.42891788e-01
6.07374728e-01 -8.13469410e-01 1.24656558e-01 -9.94949341e-02
1.46073535e-01 -3.24833095e-02 -1.17643666e+00 1.36830544e+00
9.86844376e-02 4.39267725e-01 -1.26947775e-01 -9.85131264e-01
6.57752812e-01 -1.30307555e-01 4.09165591e-01 -6.78769946e-01
-1.68457940e-01 -1.37095258e-01 -2.46643610e-02 -5.98283052e-01
6.33454844e-02 2.28812397e-02 2.31281087e-01 5.27287424e-01
1.93206638e-01 3.95524234e-01 6.06639944e-02 -1.62495510e-03
7.87117720e-01 -2.84502298e-01 3.55757892e-01 -5.26643455e-01
3.39126587e-01 -1.36405379e-01 5.11414170e-01 7.23756790e-01
-7.21107647e-02 5.32997608e-01 2.01788694e-01 -7.54270315e-01
-9.66621995e-01 -1.34093511e+00 -3.87678027e-01 1.15424371e+00
4.59183037e-01 -6.54206350e-02 -3.92425150e-01 -7.07187474e-01
2.26394519e-01 2.14087710e-01 -1.25809133e+00 -2.65882075e-01
-4.04274851e-01 -5.61928153e-01 -8.04237649e-02 2.01719493e-01
3.06628883e-01 -1.74894726e+00 -8.34856689e-01 -1.20548112e-02
1.34768048e-02 -5.95412731e-01 -6.79836273e-01 1.05708033e-01
-6.50440037e-01 -1.21162117e+00 -5.85078537e-01 -1.20112574e+00
1.28349495e+00 5.19592464e-01 1.26327085e+00 2.19483927e-01
-9.73789752e-01 5.83712995e-01 -2.26126641e-01 -1.55443490e-01
-1.08944364e-01 -4.89274770e-01 2.82252133e-01 6.54895127e-01
5.36957145e-01 -6.57323539e-01 -6.65364683e-01 3.77313495e-01
-7.40783870e-01 -1.14231326e-01 7.56268799e-01 9.79283214e-01
1.08800387e+00 2.37046599e-01 1.54906645e-01 -6.40728235e-01
-7.90737942e-02 -4.91368651e-01 -5.41008592e-01 4.19389278e-01
-3.79970789e-01 1.68171570e-01 1.74211577e-01 -8.22032690e-01
-5.36693335e-01 2.53662556e-01 7.39211321e-01 -6.41419351e-01
-3.51245552e-01 -2.47345999e-01 -2.99395591e-01 -6.70330971e-02
5.50884724e-01 5.31598270e-01 2.99499303e-01 -3.30678105e-01
3.81368041e-01 1.29074916e-01 1.98321313e-01 -3.81406516e-01
6.84884369e-01 8.72868717e-01 -8.56279582e-02 -1.23890853e+00
-8.55780602e-01 -4.14722502e-01 -7.43255258e-01 -1.66547507e-01
9.13870931e-01 -7.05984294e-01 -4.78214979e-01 4.71844748e-02
-9.81683850e-01 -5.75114116e-02 -7.48686969e-01 7.47385621e-02
-5.28258204e-01 2.26214573e-01 -2.61087507e-01 -7.97920942e-01
6.23991191e-02 -7.93055654e-01 1.20087624e+00 1.64960295e-01
-2.65156657e-01 -6.58825874e-01 6.23168573e-02 -2.05999032e-01
4.49655950e-03 1.14032552e-01 1.15437031e+00 -1.24617986e-01
-1.05894256e+00 1.39129549e-01 -8.47023353e-02 2.19687134e-01
5.52534223e-01 -7.87570700e-02 -1.04663408e+00 -3.98899049e-01
1.22344412e-01 -3.26311618e-01 9.65074241e-01 6.15480781e-01
1.37222219e+00 -6.97746336e-01 -6.38266027e-01 3.68945330e-01
1.32919610e+00 3.22679132e-02 4.62752134e-01 1.49627969e-01
6.09254837e-01 9.26611006e-01 3.30789953e-01 3.02966654e-01
1.19099822e-02 5.65430403e-01 2.58652061e-01 -3.98711301e-02
-3.27814906e-03 -2.17919827e-01 3.61612856e-01 3.72615129e-01
2.92639107e-01 3.74190211e-01 -3.32496107e-01 6.66837335e-01
-1.70165277e+00 -1.43803906e+00 5.20616770e-01 2.23858213e+00
1.00758457e+00 -4.48679850e-02 1.48735166e-01 -9.57925096e-02
9.71276522e-01 1.37637138e-01 -7.19675541e-01 7.58540407e-02
-5.23811638e-01 1.32914633e-01 9.01442692e-02 3.81097674e-01
-1.17622292e+00 7.81821609e-01 7.10818529e+00 7.93678224e-01
-8.70589137e-01 -1.29111484e-01 6.76634610e-01 -2.72642523e-01
-3.21365446e-01 -2.55296201e-01 -9.01225686e-01 5.90057611e-01
2.89111525e-01 2.85914615e-02 5.85829616e-01 9.71392334e-01
-2.10553393e-01 -9.93533134e-02 -1.31967950e+00 1.16852343e+00
4.64048862e-01 -1.35074389e+00 3.87792677e-01 2.83209801e-01
9.70817685e-01 -3.72890323e-01 6.10425055e-01 -2.06823751e-01
4.71616350e-02 -1.00728703e+00 7.53346264e-01 6.06952310e-01
7.42603600e-01 -5.22058606e-01 -3.17276977e-02 2.14687482e-01
-1.03478634e+00 -2.32779488e-01 -8.43829632e-01 3.59482467e-02
-4.42258894e-01 7.34179676e-01 -6.31299615e-01 -2.57890224e-01
1.05549788e+00 1.00678349e+00 -7.30637670e-01 7.40809560e-01
2.06697406e-03 2.06290483e-01 -2.06463784e-01 -3.95048782e-02
-1.16521068e-01 -3.25235963e-01 6.70299530e-01 7.72876263e-01
-1.07177543e-02 -2.16797307e-01 2.70849824e-01 1.29409051e+00
7.27381483e-02 5.46694174e-03 -9.04887736e-01 1.23039283e-01
4.24420595e-01 1.18883812e+00 -9.57243383e-01 -4.51905608e-01
-2.91650921e-01 1.43507218e+00 5.23379087e-01 3.18378896e-01
-4.99009311e-01 -9.00499374e-02 8.32641542e-01 2.31386349e-01
9.13717330e-01 1.70397222e-01 9.28315241e-03 -9.66794848e-01
3.22099566e-01 -9.66824472e-01 4.11980838e-01 -4.81200904e-01
-1.76671576e+00 5.30144334e-01 -1.51122184e-02 -1.29026890e+00
-4.57359776e-02 -6.59655869e-01 -4.09747273e-01 5.04667461e-01
-1.24307263e+00 -7.22085416e-01 4.62978669e-02 6.78491533e-01
6.29074335e-01 -3.35396588e-01 8.29544783e-01 -1.37421852e-02
-3.36733520e-01 5.91516554e-01 1.28102019e-01 -2.09785268e-01
3.39251906e-01 -1.26107073e+00 4.55711968e-02 5.90004146e-01
6.71189845e-01 8.50533426e-01 5.69885612e-01 -5.70886850e-01
-1.30185294e+00 -1.09918571e+00 7.01723695e-01 -7.40092039e-01
3.24326307e-01 -6.69232666e-01 -1.00243449e+00 4.20142233e-01
9.87735298e-03 3.21621388e-01 7.10196853e-01 -8.72629210e-02
-6.34966731e-01 -1.75841644e-01 -1.27927601e+00 7.30239928e-01
1.06476486e+00 -5.23948967e-01 -6.75903320e-01 2.74105966e-01
4.60858136e-01 2.25444004e-01 -2.94021070e-01 1.45120278e-01
6.15054071e-01 -8.48444581e-01 1.12772298e+00 -5.89637578e-01
4.07637864e-01 -4.52000916e-01 -2.29724810e-01 -9.37979519e-01
-8.38190317e-01 -3.23548675e-01 -2.35043094e-01 7.84100711e-01
2.80055135e-01 -2.44048357e-01 9.50209081e-01 2.41420731e-01
5.65567970e-01 -5.60313582e-01 -9.55143332e-01 -6.97102487e-01
-3.67808729e-01 2.09858403e-01 4.30466861e-01 8.73690903e-01
-3.11993182e-01 2.46422023e-01 -3.84067982e-01 3.47288251e-01
1.14226449e+00 7.73509324e-01 4.89283919e-01 -1.22746241e+00
-4.98183340e-01 -5.16265392e-01 -5.97061872e-01 -1.05390191e+00
1.99381411e-01 -8.49610627e-01 1.93930149e-01 -6.42525077e-01
5.88106871e-01 -1.33952752e-01 -4.07881528e-01 4.54187602e-01
-2.79764831e-01 4.44861323e-01 -6.27488270e-02 4.67761040e-01
-8.77404511e-01 7.81421006e-01 1.31301713e+00 -4.15317982e-01
-6.49176538e-02 -2.50371724e-01 -6.72085345e-01 7.64390647e-01
5.34890711e-01 -6.69866085e-01 -2.49836236e-01 -2.66443938e-01
1.19855802e-03 -5.34811556e-01 6.87114954e-01 -6.72852039e-01
8.14684783e-04 -2.94515431e-01 1.03718591e+00 -6.08856380e-01
3.70485425e-01 -9.31633592e-01 4.45831232e-02 3.41908187e-01
-4.24789310e-01 -1.30546950e-02 1.34166228e-02 8.33067775e-01
-1.06807165e-01 2.92943101e-02 1.03493321e+00 -4.33861047e-01
-8.24741006e-01 4.07256842e-01 -3.65802050e-01 -9.74307731e-02
1.17392194e+00 -5.37450194e-01 1.63403660e-01 -2.37524360e-02
-7.50023484e-01 -1.54278040e-01 5.27455032e-01 4.82365102e-01
9.76761103e-01 -1.70052588e+00 -3.97372007e-01 1.04781008e+00
3.54167372e-01 -1.55597016e-01 1.59433603e-01 3.55713516e-01
6.49912050e-03 -1.04560241e-01 -5.58475614e-01 -9.39984322e-01
-1.50476182e+00 8.53690207e-01 2.20114812e-01 3.05067003e-01
-7.97363877e-01 1.14299178e+00 5.81769764e-01 -1.67123247e-02
2.94417083e-01 -9.78326052e-02 -1.10250741e-01 -1.48179196e-02
9.58970904e-01 2.46038750e-01 -4.28073704e-01 -7.13463783e-01
-5.25427222e-01 8.04403305e-01 -3.59021723e-01 -8.46178085e-02
1.35546339e+00 -1.98267415e-01 -6.69746101e-01 8.03060472e-01
1.45884776e+00 -2.63583865e-02 -1.47583699e+00 -3.42739969e-01
1.66589573e-01 -8.50498319e-01 -3.87421846e-01 -4.08514023e-01
-9.67366338e-01 5.91816723e-01 8.21187615e-01 2.60971218e-01
1.06572247e+00 4.79556113e-01 1.91505700e-01 4.12012309e-01
4.11516130e-01 -9.64359641e-01 7.82543898e-01 1.86902851e-01
1.15464079e+00 -1.32687795e+00 -1.73279732e-01 -8.67938399e-02
-5.01371562e-01 6.78970516e-01 6.46280646e-01 -5.78723192e-01
8.90593708e-01 2.02950891e-02 -1.23502202e-01 -5.99084854e-01
-7.81337500e-01 -2.79739857e-01 6.46792114e-01 6.97243631e-01
-5.63616045e-02 3.18407789e-02 2.44754091e-01 2.55000740e-01
3.09591442e-02 -5.59339583e-01 -6.27152100e-02 8.38334262e-01
-7.43479013e-01 -8.66333783e-01 -4.53780890e-01 5.99840820e-01
-2.97443181e-01 -1.68237701e-01 -4.96751279e-01 3.29108924e-01
4.42861885e-01 5.45328915e-01 4.17240560e-01 -1.73581228e-01
5.09282574e-02 8.49909410e-02 9.55089748e-01 -7.89561808e-01
1.02768496e-01 9.46820527e-02 -8.14599216e-01 -7.03144968e-01
-6.06997728e-01 -7.55929232e-01 -8.30401361e-01 2.28413537e-01
-2.12675750e-01 7.05832988e-02 8.01529884e-02 5.48918307e-01
4.53785300e-01 4.65699993e-02 1.07461250e+00 -1.00277209e+00
-2.59867370e-01 -6.37000859e-01 -1.01930487e+00 7.98861265e-01
6.57681227e-01 -9.85371828e-01 -6.53752744e-01 4.40510154e-01] | [9.776880264282227, 2.0262584686279297] |
74ded33f-ee45-4d2c-a370-71d18d30ffb6 | fusepose-imu-vision-sensor-fusion-in | 2208.11960 | null | https://arxiv.org/abs/2208.11960v1 | https://arxiv.org/pdf/2208.11960v1.pdf | FusePose: IMU-Vision Sensor Fusion in Kinematic Space for Parametric Human Pose Estimation | There exist challenging problems in 3D human pose estimation mission, such as poor performance caused by occlusion and self-occlusion. Recently, IMU-vision sensor fusion is regarded as valuable for solving these problems. However, previous researches on the fusion of IMU and vision data, which is heterogeneous, fail to adequately utilize either IMU raw data or reliable high-level vision features. To facilitate a more efficient sensor fusion, in this work we propose a framework called \emph{FusePose} under a parametric human kinematic model. Specifically, we aggregate different information of IMU or vision data and introduce three distinctive sensor fusion approaches: NaiveFuse, KineFuse and AdaDeepFuse. NaiveFuse servers as a basic approach that only fuses simplified IMU data and estimated 3D pose in euclidean space. While in kinematic space, KineFuse is able to integrate the calibrated and aligned IMU raw data with converted 3D pose parameters. AdaDeepFuse further develops this kinematical fusion process to an adaptive and end-to-end trainable manner. Comprehensive experiments with ablation studies demonstrate the rationality and superiority of the proposed framework. The performance of 3D human pose estimation is improved compared to the baseline result. On Total Capture dataset, KineFuse surpasses previous state-of-the-art which uses IMU only for testing by 8.6\%. AdaDeepFuse surpasses state-of-the-art which uses IMU for both training and testing by 8.5\%. Moreover, we validate the generalization capability of our framework through experiments on Human3.6M dataset. | ['Dahong Qian', 'Xu Zhao', 'Yiming Bao'] | 2022-08-25 | null | null | null | null | ['3d-human-pose-estimation'] | ['computer-vision'] | [-2.67368674e-01 -3.91701549e-01 1.16613522e-01 -1.58731416e-01
-6.80930674e-01 -4.03962225e-01 4.65479761e-01 -2.56746262e-01
-9.20317948e-01 6.57126367e-01 -7.09814355e-02 1.75983444e-01
5.00626266e-02 -4.64659274e-01 -9.02594566e-01 -3.64421308e-01
3.03522199e-01 5.75971723e-01 1.98487654e-01 -4.52907771e-01
-2.37682357e-01 3.01285058e-01 -1.60898077e+00 -4.12936419e-01
7.63999581e-01 1.12120175e+00 2.69262046e-01 5.91848314e-01
2.20647126e-01 -1.68307051e-01 -6.00924373e-01 -2.01437443e-01
5.31525314e-01 -1.30781889e-01 -2.81054705e-01 2.33184904e-01
5.23455858e-01 -4.74194467e-01 -4.24218446e-01 9.18695688e-01
9.74234104e-01 7.02372938e-02 4.18702900e-01 -1.29975462e+00
-2.42302462e-01 -2.33761609e-01 -5.40384591e-01 -2.22005649e-04
8.14660788e-01 2.88827598e-01 2.46330932e-01 -9.67136800e-01
5.99681199e-01 1.05931091e+00 1.05040443e+00 5.43268681e-01
-7.93766975e-01 -4.68114167e-01 -4.02235463e-02 -9.95238498e-03
-1.48896003e+00 -1.90782323e-01 6.26138151e-01 -4.59940881e-01
8.07780147e-01 2.18021750e-01 7.30843604e-01 1.18811655e+00
1.97614476e-01 7.95670807e-01 1.05745530e+00 -3.40027124e-01
-1.35980427e-01 -1.77315503e-01 2.43766397e-01 7.86341190e-01
5.78376114e-01 1.97838098e-01 -4.64922577e-01 7.45395645e-02
8.86139750e-01 3.89834195e-01 -4.46039617e-01 -6.04166329e-01
-1.59235716e+00 2.57300079e-01 6.05771661e-01 -5.18670399e-03
-6.41666412e-01 2.07967654e-01 1.60013512e-01 5.40321581e-02
1.32591680e-01 8.61295015e-02 -3.20463538e-01 -2.62294680e-01
-6.05343223e-01 5.01531363e-01 2.58528233e-01 1.19486821e+00
6.94292665e-01 1.37494668e-01 1.75485566e-01 6.31524742e-01
3.97074878e-01 1.19444609e+00 4.86792773e-01 -8.06988776e-01
7.28859663e-01 6.29190326e-01 4.35232341e-01 -9.37700689e-01
-7.66010821e-01 -3.58459055e-01 -8.66362214e-01 2.88709402e-01
2.93153793e-01 -5.35956621e-01 -1.02102053e+00 1.59787214e+00
6.15366459e-01 1.45717766e-02 9.11105722e-02 1.40817881e+00
7.92316496e-01 4.47759688e-01 -1.31475106e-01 -3.15685384e-02
1.31726980e+00 -8.67829919e-01 -7.09732175e-01 -4.74493891e-01
4.75428164e-01 -6.93902433e-01 7.05264032e-01 3.13050866e-01
-8.55720580e-01 -1.11078703e+00 -1.36794353e+00 2.07641214e-01
-2.80263275e-01 3.04643780e-01 5.55621564e-01 6.01397514e-01
-8.65465343e-01 2.53409535e-01 -8.78662825e-01 -6.33068562e-01
-2.25069612e-01 5.11675954e-01 -8.97354424e-01 -7.38832206e-02
-1.24707890e+00 1.15598488e+00 7.49343157e-01 3.30746293e-01
-5.28040767e-01 -2.85588562e-01 -9.91367817e-01 -6.60768270e-01
6.23697162e-01 -1.45163023e+00 1.11159182e+00 -2.02810526e-01
-1.39141703e+00 6.22254133e-01 9.37815681e-02 -3.89104575e-01
7.44720995e-01 -7.80470133e-01 -2.61578262e-01 1.53049618e-01
-3.60938385e-02 5.69460630e-01 6.33874178e-01 -1.32029247e+00
-6.70781195e-01 -7.13856518e-01 -3.90006155e-02 6.63961530e-01
-5.05859219e-02 -5.30138731e-01 -7.89018631e-01 -4.44153696e-01
5.20903528e-01 -1.17261422e+00 -1.36797354e-01 -1.26183838e-01
-1.66016638e-01 1.54393669e-02 5.85448444e-01 -6.11759126e-01
1.03938973e+00 -1.70327604e+00 4.62909490e-01 3.63721177e-02
2.08121061e-01 4.30746347e-01 2.26461694e-01 1.29381686e-01
1.80900529e-01 -3.43884110e-01 3.87729965e-02 -6.86882555e-01
6.29320089e-03 3.42803538e-01 2.88257688e-01 6.05295181e-01
-2.61858374e-01 9.58154559e-01 -7.94615805e-01 -6.00633800e-01
7.85065174e-01 5.46399176e-01 -2.57708251e-01 3.44045520e-01
2.31466204e-01 5.15152752e-01 -4.59226221e-01 1.03987336e+00
5.87947249e-01 2.28845403e-01 -2.68268019e-01 -5.38455009e-01
-7.86425471e-02 -5.17650545e-01 -1.48754704e+00 2.31811476e+00
-2.86545068e-01 -3.72170135e-02 1.48635864e-01 -7.31651723e-01
9.81701851e-01 3.56172264e-01 6.58650100e-01 -4.38378781e-01
4.91404206e-01 3.87280524e-01 -5.40546000e-01 -6.05137050e-01
5.67899227e-01 4.73025590e-02 -4.02579814e-01 -1.19023390e-01
2.46100560e-01 -5.10526672e-02 -4.54695374e-02 -2.19899401e-01
8.59393537e-01 8.19639206e-01 4.73487735e-01 7.64511302e-02
6.28821611e-01 4.51782458e-02 5.63744128e-01 6.28191292e-01
-5.41253090e-01 8.31290960e-01 -3.48407447e-01 -3.44755232e-01
-1.01331425e+00 -1.36866724e+00 9.68368351e-02 5.15853763e-01
5.01849473e-01 -4.38005656e-01 -7.79899001e-01 -5.99485099e-01
2.27019757e-01 8.24884027e-02 -4.13517177e-01 -5.64505644e-02
-5.26033044e-01 -7.05151379e-01 5.64402103e-01 6.25701070e-01
1.07282102e+00 -4.16974634e-01 -9.58499372e-01 -3.53657417e-02
-6.31906569e-01 -1.28555822e+00 -4.03323740e-01 -3.45564514e-01
-8.55507433e-01 -1.08587110e+00 -1.06553853e+00 -3.74860883e-01
4.81943995e-01 5.95858216e-01 7.53485858e-01 -2.51824409e-01
-1.94120705e-01 8.38505089e-01 -4.95882958e-01 -3.82620811e-01
1.05641201e-01 -4.67479751e-02 6.87995315e-01 -8.88799950e-02
3.00574720e-01 -4.26554441e-01 -7.49008298e-01 4.77148980e-01
-5.55298686e-01 4.51789238e-02 6.93331718e-01 6.64361238e-01
8.36985111e-01 -3.55321169e-01 2.13208020e-01 -6.20807000e-02
1.97393537e-01 -2.96467364e-01 -2.89839268e-01 2.07432792e-01
-5.50531685e-01 -1.41124666e-01 1.54181033e-01 -2.82664865e-01
-9.15322781e-01 4.68692631e-01 -1.65197507e-01 -8.29772592e-01
-2.74723679e-01 4.68797117e-01 -2.68262863e-01 -2.01324001e-01
8.14692140e-01 1.84955448e-01 4.09514874e-01 -6.52534425e-01
2.23260775e-01 7.79403567e-01 1.12610912e+00 -5.87772369e-01
9.53997433e-01 4.43143189e-01 1.58963010e-01 -7.81752944e-01
-4.20188099e-01 -6.92196488e-01 -7.63441741e-01 -6.29788220e-01
1.04444373e+00 -1.19219255e+00 -8.39713991e-01 6.93753839e-01
-1.12335980e+00 3.72122169e-01 9.11662579e-02 8.64326715e-01
-6.95682764e-01 6.44671321e-01 -2.37820968e-01 -9.58450079e-01
-4.21758801e-01 -1.15951288e+00 1.43229735e+00 1.75114691e-01
-1.03055522e-01 -6.14236951e-01 9.29272920e-02 6.97821558e-01
-2.37625130e-02 8.54756057e-01 -1.41578257e-01 -1.65150270e-01
-2.96341121e-01 -7.25700557e-01 1.34798348e-01 2.20828533e-01
1.66834332e-02 -4.88651037e-01 -6.41472042e-01 -4.27092582e-01
-5.81401438e-02 -3.28191727e-01 6.35902226e-01 2.03521848e-01
2.36449182e-01 4.01529998e-01 -4.92541254e-01 7.64874220e-01
1.33565867e+00 -2.17636134e-02 4.54328597e-01 6.64112508e-01
9.29707944e-01 3.76901835e-01 9.90533590e-01 3.99455220e-01
6.93386793e-01 9.63606298e-01 6.76846206e-01 -7.41821751e-02
2.14406084e-02 -2.74976909e-01 4.17804390e-01 7.51923621e-01
-7.07642198e-01 3.34194452e-02 -9.62971628e-01 1.77183285e-01
-2.26623964e+00 -7.05323458e-01 -1.16531119e-01 2.32845521e+00
3.79114777e-01 2.31764778e-01 2.12544024e-01 2.11484298e-01
7.24762440e-01 6.50666580e-02 -3.75731885e-01 3.10189635e-01
5.64131252e-02 -9.16923434e-02 7.87867725e-01 4.10987377e-01
-1.33108377e+00 8.31546068e-01 5.32114697e+00 4.64059353e-01
-8.76618803e-01 1.77839071e-01 -2.22641036e-01 -1.30836487e-01
2.97636271e-01 -2.27804229e-01 -1.00008082e+00 4.48955178e-01
6.63326323e-01 1.88757733e-01 9.01057348e-02 8.85807633e-01
-1.17696196e-01 -3.48851293e-01 -8.83312345e-01 1.48820579e+00
3.04858178e-01 -8.47309470e-01 -2.38461316e-01 -7.18031675e-02
3.69898975e-01 7.28606805e-02 -2.64941424e-01 4.58003134e-01
-1.73803285e-01 -5.99851131e-01 7.84572303e-01 9.44895983e-01
7.62946904e-01 -6.62946105e-01 1.16136503e+00 7.24623144e-01
-1.32864523e+00 3.07963967e-01 -1.82615504e-01 -2.84870416e-01
4.05298442e-01 2.54378885e-01 -6.07933998e-01 1.31617367e+00
8.55502665e-01 5.00729144e-01 -6.81737006e-01 8.52440894e-01
-9.06027257e-02 -1.77306667e-01 -6.99013174e-01 2.00945027e-02
1.53243393e-01 -1.87797964e-01 7.70555675e-01 8.67690265e-01
5.86298764e-01 1.77452907e-01 7.02681601e-01 1.97114706e-01
4.10244584e-01 -2.37682864e-01 -7.01753974e-01 3.74232799e-01
3.18538398e-01 1.13873386e+00 -3.41931492e-01 -3.13025028e-01
-2.43389875e-01 1.25355244e+00 -1.54721430e-02 2.73193806e-01
-9.98727679e-01 -1.89951286e-01 5.80745459e-01 6.62176684e-02
1.34659290e-01 -6.87074542e-01 -7.44593963e-02 -1.41642296e+00
3.87670577e-01 -6.13637567e-01 2.17380852e-01 -9.71531034e-01
-7.55191445e-01 6.83848500e-01 5.07304728e-01 -1.77660143e+00
-6.54124737e-01 -6.94115996e-01 -3.64192128e-02 8.59086871e-01
-9.83979464e-01 -1.50221086e+00 -7.60681152e-01 7.13074028e-01
3.96582901e-01 -1.42883688e-01 6.61704957e-01 2.46162847e-01
-5.45232415e-01 4.60292816e-01 -2.16776580e-01 2.92388722e-02
7.97288656e-01 -9.92534280e-01 3.40163648e-01 1.06673932e+00
-1.36813954e-01 6.73361003e-01 9.26977456e-01 -8.20173204e-01
-1.77895975e+00 -7.61297762e-01 6.51930869e-01 -8.79047930e-01
2.39159390e-01 -2.20778093e-01 -2.84318626e-01 6.53950036e-01
-1.04070738e-01 2.11179122e-01 3.29479545e-01 -1.05879314e-01
-5.86758070e-02 -2.53416002e-01 -1.15967131e+00 5.48563898e-01
1.33908308e+00 -1.42643541e-01 -1.02829504e+00 2.10310109e-02
6.13753617e-01 -8.57282400e-01 -9.87527370e-01 8.16110790e-01
7.16909885e-01 -9.98336494e-01 1.22843421e+00 -2.06434116e-01
-1.69380262e-01 -6.95638359e-01 -5.49609959e-01 -1.25688696e+00
-1.90752372e-01 -4.24530029e-01 -3.62954348e-01 8.73791218e-01
-1.33119002e-01 -6.57784402e-01 7.96300292e-01 3.41193587e-01
-2.73019314e-01 -6.02952659e-01 -1.00014961e+00 -9.28354263e-01
-4.44467366e-01 -5.26954234e-01 3.54242504e-01 4.09630030e-01
-2.75383949e-01 3.77971172e-01 -6.72941089e-01 3.05011839e-01
9.58523214e-01 -1.47336990e-01 1.40795016e+00 -1.23753667e+00
-3.00372273e-01 -2.26212386e-02 -8.28936398e-01 -1.20991111e+00
-1.45711482e-01 -3.68754834e-01 1.28859907e-01 -1.59780145e+00
-1.62478626e-01 1.37513047e-02 -1.96570233e-01 2.63651103e-01
-4.12124574e-01 4.82692093e-01 5.46370327e-01 4.00471270e-01
-6.26051962e-01 5.26518226e-01 1.17217720e+00 4.06091698e-02
-7.35220835e-02 1.38201900e-02 -2.03793108e-01 7.53326476e-01
5.43010771e-01 5.73663749e-02 -3.82564187e-01 -4.30313617e-01
-6.17877506e-02 3.15199256e-01 7.14943111e-01 -1.63585401e+00
3.57204556e-01 1.59965694e-01 6.37617648e-01 -9.73469973e-01
7.08338678e-01 -1.01115751e+00 5.35128474e-01 5.24678409e-01
5.46132326e-01 3.59263718e-01 1.20988242e-01 5.85124969e-01
-2.34530196e-01 6.20151088e-02 3.31694901e-01 -3.00045311e-01
-1.12159634e+00 4.89514798e-01 1.37899250e-01 -2.91697264e-01
1.22707093e+00 -7.08593130e-01 -7.20908791e-02 -2.24381641e-01
-6.92416787e-01 3.92701924e-01 5.13608575e-01 5.30179381e-01
7.53806472e-01 -1.63179600e+00 -4.80876893e-01 2.29583368e-01
2.60650784e-01 2.45867848e-01 4.62434381e-01 1.02086985e+00
-6.39598608e-01 3.97590607e-01 -4.23606843e-01 -9.36874986e-01
-1.19381118e+00 5.32667935e-01 2.92524368e-01 4.78328653e-02
-3.68082553e-01 5.52608967e-01 -1.75330430e-01 -7.07801521e-01
1.65323451e-01 -2.07797229e-01 -9.10584107e-02 -1.02313317e-01
3.34474713e-01 5.10325551e-01 5.73466532e-02 -1.09763145e+00
-5.50306737e-01 1.09897137e+00 3.06421399e-01 -2.73443401e-01
8.79560947e-01 -3.93215984e-01 3.71962756e-01 4.09668237e-01
1.01210928e+00 -2.46746778e-01 -1.51359189e+00 -1.82183340e-01
-4.31892246e-01 -3.31765354e-01 -1.58783242e-01 -5.68716347e-01
-6.71235144e-01 5.77059388e-01 9.47594464e-01 -2.78838664e-01
1.06576061e+00 -2.82893777e-01 9.74308074e-01 5.52898765e-01
9.74401712e-01 -1.07329071e+00 -2.50244942e-02 4.64383870e-01
9.45872009e-01 -1.49040306e+00 1.97360158e-01 -3.44334424e-01
-5.85424542e-01 8.58898342e-01 8.69926453e-01 -1.71092078e-01
5.05468965e-01 6.29924536e-02 2.25156128e-01 9.89594236e-02
-1.50652751e-01 -6.18384421e-01 4.93211776e-01 7.69790530e-01
3.87869589e-02 5.00499690e-03 -4.30862099e-01 5.56885600e-01
-2.39611790e-01 1.91003203e-01 -8.39903951e-02 1.28095818e+00
-5.72962880e-01 -8.62731934e-01 -9.68304217e-01 -8.20234492e-02
-1.12898082e-01 4.15815115e-01 -1.59753948e-01 1.23522174e+00
4.96659607e-01 8.12341988e-01 -2.61744916e-01 -1.07801402e+00
8.73123407e-01 8.10912549e-02 7.75700450e-01 -1.60852745e-01
-4.75721210e-01 2.63603836e-01 4.21014354e-02 -7.92238176e-01
-6.52926266e-01 -6.26787841e-01 -1.30800724e+00 -3.05280030e-01
-2.49479488e-01 -9.35408920e-02 7.20437050e-01 1.01462960e+00
3.39012384e-01 5.17737150e-01 3.34360719e-01 -1.36569321e+00
-7.43730128e-01 -1.00379872e+00 -2.59728551e-01 5.39351702e-01
3.30922455e-01 -1.18886590e+00 -8.99736769e-03 -3.86334322e-02] | [7.040083885192871, -0.9038002490997314] |
52519610-df0b-4250-910c-a614586c4a76 | tcn-table-convolutional-network-for-web-table | 2102.09460 | null | https://arxiv.org/abs/2102.09460v1 | https://arxiv.org/pdf/2102.09460v1.pdf | TCN: Table Convolutional Network for Web Table Interpretation | Information extraction from semi-structured webpages provides valuable long-tailed facts for augmenting knowledge graph. Relational Web tables are a critical component containing additional entities and attributes of rich and diverse knowledge. However, extracting knowledge from relational tables is challenging because of sparse contextual information. Existing work linearize table cells and heavily rely on modifying deep language models such as BERT which only captures related cells information in the same table. In this work, we propose a novel relational table representation learning approach considering both the intra- and inter-table contextual information. On one hand, the proposed Table Convolutional Network model employs the attention mechanism to adaptively focus on the most informative intra-table cells of the same row or column; and, on the other hand, it aggregates inter-table contextual information from various types of implicit connections between cells across different tables. Specifically, we propose three novel aggregation modules for (i) cells of the same value, (ii) cells of the same schema position, and (iii) cells linked to the same page topic. We further devise a supervised multi-task training objective for jointly predicting column type and pairwise column relation, as well as a table cell recovery objective for pre-training. Experiments on real Web table datasets demonstrate our method can outperform competitive baselines by +4.8% of F1 for column type prediction and by +4.1% of F1 for pairwise column relation prediction. | ['Meng Jiang', 'Xin Luna Dong', 'Binxuan Huang', 'Colin Lockard', 'Prashant Shiralkar', 'Daheng Wang'] | 2021-02-17 | null | null | null | null | ['type-prediction', 'table-annotation', 'table-annotation'] | ['computer-code', 'knowledge-base', 'natural-language-processing'] | [-1.87469184e-01 4.25552011e-01 -7.93631434e-01 -2.93509126e-01
-9.24521863e-01 -6.77275836e-01 2.56727368e-01 9.32924271e-01
-6.69617206e-02 1.06453121e+00 4.85788554e-01 -2.37245128e-01
-2.18823344e-01 -1.33605313e+00 -1.38182259e+00 -3.60541403e-01
-1.74943078e-02 7.33549416e-01 2.58118540e-01 -3.15105438e-01
-1.07758462e-01 2.65496641e-01 -1.22328186e+00 8.81103814e-01
1.04463983e+00 1.49943066e+00 -2.33893365e-01 1.35333791e-01
-6.69933856e-01 1.36943686e+00 -3.97562385e-01 -8.48906398e-01
8.19951221e-02 1.39717534e-02 -7.20354676e-01 -2.28061453e-01
5.71503758e-01 -8.14459026e-02 -4.42809373e-01 9.01522458e-01
1.04247220e-01 -1.30743220e-01 3.93086970e-01 -1.00931013e+00
-7.32932210e-01 1.33039117e+00 -6.98573887e-01 2.48875245e-01
2.64561236e-01 -2.96378791e-01 1.53611028e+00 -7.07142174e-01
9.18906987e-01 9.94929612e-01 5.08625209e-01 -1.49713149e-02
-1.11074626e+00 -6.60953462e-01 5.48347652e-01 1.72707513e-01
-1.49802864e+00 -1.83438674e-01 6.87612593e-01 -1.83147043e-01
9.91383672e-01 2.54984558e-01 3.34581405e-01 7.85936773e-01
1.13820150e-01 1.03338027e+00 5.69853246e-01 1.15864269e-01
-2.46450864e-02 5.31735957e-01 3.59242648e-01 7.61205137e-01
9.14094508e-01 -7.64847636e-01 -7.20445037e-01 1.01590231e-01
5.05324721e-01 -7.07114115e-02 -1.72582865e-02 -6.32591248e-01
-1.23338556e+00 6.43961549e-01 8.06542039e-01 2.18930319e-01
-4.38076526e-01 5.49875237e-02 4.29153234e-01 1.18888989e-02
2.42431909e-01 2.29166850e-01 -9.28079784e-01 1.67707771e-01
-3.36659014e-01 2.65651286e-01 1.13717580e+00 1.52721047e+00
9.47971940e-01 -4.32251036e-01 -2.98902184e-01 6.32837534e-01
8.58507752e-02 4.26273882e-01 2.71651056e-02 -2.83612907e-01
1.35169399e+00 1.31398630e+00 2.18052715e-02 -1.13281274e+00
-4.71441180e-01 -7.17544794e-01 -9.94028628e-01 -6.47100747e-01
4.12879646e-01 -4.37642913e-03 -6.51694834e-01 1.60118318e+00
3.95134479e-01 -3.94461274e-01 2.66292304e-01 4.17658180e-01
1.07636154e+00 5.21535337e-01 3.93290678e-03 -1.72551170e-01
1.53175473e+00 -7.65706897e-01 -9.54834640e-01 -3.49340796e-01
8.66644621e-01 -3.52020413e-01 8.30703259e-01 1.28179416e-01
-1.05026150e+00 -4.33396906e-01 -1.13430810e+00 -5.93679547e-01
-9.62244511e-01 2.31568500e-01 1.02247632e+00 3.37559015e-01
-4.31576550e-01 9.73736793e-02 -5.66720128e-01 -8.42970461e-02
5.92857003e-01 4.53221917e-01 -4.13472772e-01 -1.90045029e-01
-1.30370474e+00 4.63211179e-01 6.08720303e-01 6.32838756e-02
-1.08970165e-01 -1.03666687e+00 -9.84765410e-01 4.06163096e-01
1.16872048e+00 -6.34715497e-01 6.16038501e-01 -3.26229751e-01
-7.41534710e-01 7.11556196e-01 -3.33890319e-01 -6.84491396e-01
2.79401034e-01 -3.78736347e-01 -6.10721648e-01 -1.86862230e-01
2.84925491e-01 2.16851130e-01 1.26694785e-02 -1.44112873e+00
-8.77180696e-01 -4.88755018e-01 1.83709830e-01 1.96804076e-01
-2.88693041e-01 -5.68879783e-01 -9.98267472e-01 -5.97241819e-01
1.55499235e-01 -4.98655349e-01 1.32964894e-01 -2.76979744e-01
-9.49101388e-01 -2.78847694e-01 3.95772040e-01 -7.67691433e-01
1.58072889e+00 -1.87532079e+00 7.55114201e-03 4.71099913e-01
4.08893615e-01 -1.37470976e-01 3.68998587e-01 3.84915084e-01
1.27404988e-01 3.36364210e-01 -8.88563618e-02 2.57488210e-02
8.72868970e-02 1.22574173e-01 -5.54382861e-01 -2.47255564e-01
3.23793054e-01 1.13367641e+00 -6.69835925e-01 -5.74596286e-01
-3.28269392e-01 1.72467113e-01 -6.30829751e-01 -5.66149577e-02
-5.59103012e-01 -1.43350348e-01 -4.93906647e-01 1.02462912e+00
6.84515297e-01 -6.07183933e-01 5.05142868e-01 -7.16159523e-01
2.42582723e-01 6.43789351e-01 -1.44044185e+00 1.45794857e+00
-4.97558981e-01 7.18125701e-02 -6.10093288e-02 -6.58053219e-01
8.68435621e-01 -8.43316987e-02 5.76042414e-01 -7.65042901e-01
-1.28571913e-01 2.41209596e-01 8.98917089e-04 -1.73539847e-01
6.60543263e-01 3.94247413e-01 -3.76429647e-01 2.45479763e-01
1.17987595e-01 3.68530929e-01 6.70192957e-01 6.82669997e-01
1.06773424e+00 -6.22061491e-02 4.96477515e-01 -2.93935031e-01
7.84935296e-01 -1.56429887e-01 8.02907646e-01 6.77591026e-01
3.23037356e-01 1.62169948e-01 1.17695856e+00 -6.00973666e-01
-6.82218075e-01 -1.10671246e+00 -7.69746900e-02 9.29420769e-01
3.66983473e-01 -6.40116692e-01 -3.00040543e-01 -9.46187675e-01
6.37212515e-01 5.44037700e-01 -9.79774475e-01 1.64012499e-02
-7.32607901e-01 -7.60974109e-01 2.68039584e-01 9.85068858e-01
6.28678799e-01 -8.25213969e-01 2.80929387e-01 3.15546423e-01
-4.46223736e-01 -1.45191848e+00 -4.42807168e-01 6.83689296e-01
-5.55565059e-01 -1.25479376e+00 1.00196339e-01 -6.64102793e-01
5.60090840e-01 -3.07676755e-02 1.51425290e+00 1.34029910e-01
1.88485980e-01 -1.72787070e-01 -4.19939794e-02 -3.32465976e-01
1.40246153e-01 6.47068799e-01 -2.49738216e-01 2.34742850e-01
6.48667157e-01 -2.86630630e-01 -2.32617512e-01 1.98763907e-01
-7.26289690e-01 1.23117842e-01 8.56915176e-01 7.45068729e-01
1.06010377e+00 2.10793629e-01 5.07935643e-01 -1.68790829e+00
3.16146195e-01 -9.01830614e-01 -5.35889089e-01 6.43675268e-01
-7.78577328e-01 5.15369177e-01 7.39251256e-01 -1.01657078e-01
-1.13420737e+00 -8.28963518e-03 1.96311340e-01 -2.70203166e-02
8.56621563e-02 9.58350658e-01 -9.29867387e-01 5.05021811e-01
3.41937363e-01 9.74151641e-02 -5.64660907e-01 -4.14507776e-01
4.06526476e-01 2.07730085e-01 5.72510540e-01 -7.79241562e-01
8.87575090e-01 6.55328929e-01 1.54480889e-01 -2.46717349e-01
-1.14719987e+00 -2.87094653e-01 -8.26917589e-01 3.13533902e-01
7.22334921e-01 -1.16853809e+00 -9.00259852e-01 3.44438516e-02
-6.39743626e-01 -1.97542965e-01 -2.07919896e-01 5.83730452e-02
-7.68334344e-02 -6.53733686e-03 -6.58165574e-01 -4.27838027e-01
-2.33388379e-01 -7.91061997e-01 9.34315920e-01 1.52914569e-01
4.59167436e-02 -9.32607353e-01 -3.19012135e-01 6.25778556e-01
1.18342880e-02 1.47878215e-01 1.43191373e+00 -1.00727618e+00
-1.28017342e+00 -3.41483831e-01 -5.51031291e-01 -2.32871160e-01
2.45393991e-01 -2.74462134e-01 -5.34819543e-01 7.98659995e-02
-6.41458690e-01 -3.52115631e-01 1.17077160e+00 -2.50874441e-02
1.31464183e+00 -4.94893759e-01 -6.42507136e-01 8.27069700e-01
1.51217008e+00 2.20077723e-01 4.81988460e-01 3.81991029e-01
1.26155818e+00 6.09805822e-01 5.48044443e-01 3.29009593e-01
9.43143606e-01 4.65125203e-01 2.27916762e-01 1.11975066e-01
-1.24494517e-02 -7.54698396e-01 -1.11894868e-01 6.55593514e-01
3.62651020e-01 -5.72236419e-01 -8.73050451e-01 5.70469975e-01
-1.84495187e+00 -8.69221151e-01 -2.38686070e-01 2.14325666e+00
1.16884792e+00 6.53236687e-01 -9.33606997e-02 6.82678446e-02
4.95867282e-01 1.02122098e-01 -7.11570680e-01 2.18013581e-02
-5.21688581e-01 -4.64136340e-02 8.09845269e-01 4.10332769e-01
-1.24969280e+00 9.60007429e-01 4.11845589e+00 8.16590965e-01
-4.42181647e-01 -4.22828555e-01 8.44235718e-01 -1.01254404e-01
-7.35476494e-01 -2.11563557e-02 -1.44556856e+00 4.50567424e-01
5.17716348e-01 -1.88649938e-01 3.82433563e-01 7.21823275e-01
-6.20276630e-01 -1.69240505e-01 -1.20450544e+00 6.97173297e-01
1.51473522e-01 -1.62300205e+00 5.00050008e-01 2.40861863e-01
7.45374739e-01 -3.83281708e-01 9.05493088e-03 6.04475617e-01
3.84598911e-01 -8.12016964e-01 5.84023774e-01 4.95742589e-01
6.97145581e-01 -1.00735354e+00 9.92871284e-01 -3.51279415e-03
-1.69360912e+00 -5.99742541e-03 -1.63241550e-01 3.62066984e-01
-2.94789582e-01 6.72699213e-01 -4.99438792e-01 8.64540637e-01
7.75193572e-01 8.91820312e-01 -8.84409666e-01 4.58546311e-01
-1.71833768e-01 2.75055289e-01 -3.11068952e-01 -1.19092278e-01
2.53067762e-02 -1.17972046e-01 2.69263554e-02 1.07950592e+00
-8.43730345e-02 7.70889670e-02 1.09461173e-01 7.62847662e-01
-7.72271812e-01 2.28484914e-01 -4.98262435e-01 -9.89392176e-02
6.64847553e-01 1.26415753e+00 -8.85650873e-01 -5.27426004e-01
-7.56009936e-01 3.81953418e-01 7.40734100e-01 3.22260499e-01
-6.92866504e-01 -5.56387782e-01 7.13534296e-01 9.75364000e-02
7.37783015e-01 1.82218283e-01 -1.05963397e+00 -1.32705462e+00
4.39066857e-01 -6.59648776e-01 8.47882211e-01 -4.49221522e-01
-1.14637578e+00 4.62597221e-01 -1.80599719e-01 -8.33290100e-01
-1.29175082e-01 -5.30458033e-01 -7.78914690e-02 7.69888103e-01
-1.39050388e+00 -1.25213778e+00 -4.13144946e-01 5.90112627e-01
1.08492948e-01 -2.34493569e-01 5.21883905e-01 3.37340444e-01
-7.70737946e-01 9.56143022e-01 4.11669612e-02 6.52885377e-01
6.67114556e-01 -1.71565771e+00 4.43149239e-01 8.32584321e-01
2.23443434e-01 9.61803973e-01 2.09369853e-01 -9.77838874e-01
-1.72607052e+00 -1.19498813e+00 1.08722389e+00 -6.89729214e-01
8.24468255e-01 -7.95370638e-01 -1.26393199e+00 1.11781156e+00
-6.38250411e-02 1.70196697e-01 8.02528679e-01 7.11055100e-01
-7.94855297e-01 -7.27225840e-01 -9.34005678e-01 5.52761257e-01
9.96321619e-01 -5.74465990e-01 -3.03715765e-01 2.45689735e-01
1.01162410e+00 -5.77323496e-01 -1.24921966e+00 4.54444975e-01
4.61820751e-01 -5.64423680e-01 1.09224832e+00 -4.81738508e-01
7.35831559e-01 -2.46888816e-01 -2.32539207e-01 -8.80328417e-01
-2.84884959e-01 -2.17756093e-01 -9.82412755e-01 1.52770889e+00
8.58920574e-01 -2.98468679e-01 1.07129848e+00 5.84733665e-01
-1.43977717e-01 -1.05669045e+00 -4.98146415e-01 -4.10935074e-01
-3.60294729e-02 -9.12976563e-02 8.56876493e-01 8.96865010e-01
9.64463726e-02 6.18091345e-01 -1.06137790e-01 3.23676944e-01
5.58158278e-01 6.45916641e-01 9.34852779e-01 -1.35731769e+00
-8.97439495e-02 -4.69742090e-01 -5.46560436e-02 -9.98433888e-01
4.29679342e-02 -1.10855198e+00 -2.84645528e-01 -1.75134909e+00
3.22260976e-01 -5.66989362e-01 -5.41117966e-01 5.60803652e-01
-4.93512183e-01 -2.52541602e-01 -5.45727015e-02 6.15013465e-02
-7.51578748e-01 2.01473221e-01 9.70133901e-01 -3.26370388e-01
-1.63475707e-01 -1.75114200e-01 -1.20016408e+00 3.50455314e-01
6.06852114e-01 -3.32112193e-01 -4.67062652e-01 -3.80016118e-01
9.17185366e-01 2.21815154e-01 -1.12614229e-01 -8.75122309e-01
4.87188518e-01 -7.16037676e-02 6.94837987e-01 -1.10891628e+00
1.61612526e-01 -9.05622959e-01 -2.39972681e-01 2.18959048e-01
-5.60102046e-01 -7.01677660e-03 3.35274577e-01 9.75742936e-01
-3.26238900e-01 3.49762499e-01 2.26700798e-01 -1.27448052e-01
-6.92673802e-01 4.53502178e-01 1.60584405e-01 4.41313773e-01
7.80088067e-01 2.03616589e-01 -6.19193077e-01 -4.43000868e-02
-4.09949124e-01 6.19983435e-01 2.09053978e-01 4.89458919e-01
3.00322652e-01 -1.28729153e+00 -4.76895630e-01 4.21706825e-01
5.71875393e-01 3.69274408e-01 9.95383039e-02 5.26998162e-01
-4.31873590e-01 7.17760861e-01 -3.91424261e-02 -1.24551751e-01
-8.64066124e-01 9.73548651e-01 6.10632896e-02 -8.49534273e-01
-3.58633190e-01 1.00177693e+00 1.84320733e-01 -3.84885967e-01
4.27310944e-01 -8.48523378e-01 -2.81249523e-01 4.71877038e-01
3.58501971e-01 2.31617037e-03 4.59641516e-01 -2.99723953e-01
-4.47456032e-01 9.60670561e-02 -6.68677568e-01 3.77782971e-01
1.07211781e+00 -8.65814835e-02 -3.22295010e-01 6.53198898e-01
8.91081691e-01 4.92302299e-01 -1.05150914e+00 -6.69061959e-01
3.22662413e-01 -2.21543923e-01 -1.45854086e-01 -1.24790168e+00
-1.20640421e+00 5.37628591e-01 -1.78835452e-01 1.47168189e-01
9.46908116e-01 6.94267452e-02 7.90662348e-01 8.10592175e-01
4.05386508e-01 -9.26123679e-01 -2.20063657e-01 4.17202622e-01
6.84451044e-01 -1.31873095e+00 2.66682804e-01 -8.17979515e-01
-6.97107553e-01 9.50248539e-01 1.06344104e+00 1.25099987e-01
7.08062947e-01 5.27931690e-01 -2.64901936e-01 -3.21504831e-01
-1.09399676e+00 -3.67156059e-01 5.06654620e-01 2.94987887e-01
5.95916510e-01 4.18359376e-02 -1.23800740e-01 1.06365561e+00
-1.37485683e-01 -2.02553779e-01 2.36456171e-01 8.42612028e-01
-1.37146845e-01 -7.85640061e-01 2.89111417e-02 1.01254535e+00
-4.94094849e-01 -3.91341537e-01 -4.83815521e-01 1.00680065e+00
1.71729907e-01 5.80307722e-01 2.41739154e-01 -4.76261914e-01
3.58961344e-01 1.27952710e-01 1.26592770e-01 -4.80759144e-01
-7.26078272e-01 -1.37979716e-01 1.90704420e-01 -5.15467942e-01
7.87327811e-02 -7.11678624e-01 -1.32599664e+00 -3.96164715e-01
3.82927954e-02 1.80074081e-01 1.29616573e-01 7.90245175e-01
4.95755583e-01 8.63863349e-01 2.31640577e-01 1.72692120e-01
-8.42701569e-02 -9.68482077e-01 -6.44122958e-01 4.50225562e-01
2.44363725e-01 -6.50379896e-01 9.23411623e-02 -1.16877198e-01] | [9.54647159576416, 7.850513935089111] |
73276270-2b27-4bae-8ba1-41e16fae97db | gri-general-reinforced-imitation-and-its | 2111.08575 | null | https://arxiv.org/abs/2111.08575v2 | https://arxiv.org/pdf/2111.08575v2.pdf | GRI: General Reinforced Imitation and its Application to Vision-Based Autonomous Driving | Deep reinforcement learning (DRL) has been demonstrated to be effective for several complex decision-making applications such as autonomous driving and robotics. However, DRL is notoriously limited by its high sample complexity and its lack of stability. Prior knowledge, e.g. as expert demonstrations, is often available but challenging to leverage to mitigate these issues. In this paper, we propose General Reinforced Imitation (GRI), a novel method which combines benefits from exploration and expert data and is straightforward to implement over any off-policy RL algorithm. We make one simplifying hypothesis: expert demonstrations can be seen as perfect data whose underlying policy gets a constant high reward. Based on this assumption, GRI introduces the notion of offline demonstration agents. This agent sends expert data which are processed both concurrently and indistinguishably with the experiences coming from the online RL exploration agent. We show that our approach enables major improvements on vision-based autonomous driving in urban environments. We further validate the GRI method on Mujoco continuous control tasks with different off-policy RL algorithms. Our method ranked first on the CARLA Leaderboard and outperforms World on Rails, the previous state-of-the-art, by 17%. | ['Fabien Moutarde', 'Sascha Hornauer', 'Marin Toromanoff', 'Raphael Chekroun'] | 2021-11-16 | null | null | null | null | ['carla-map-leaderboard'] | ['robots'] | [-2.88873762e-01 2.89991111e-01 -2.76305199e-01 -9.94236693e-02
-7.03848243e-01 -6.33565187e-01 8.88514102e-01 -3.57727408e-02
-9.38508987e-01 1.09012055e+00 -1.69436678e-01 -4.90104109e-01
4.52063419e-02 -5.03273845e-01 -1.03620243e+00 -6.74317300e-01
-4.50351417e-01 6.86825693e-01 2.77200639e-01 -6.10260844e-01
1.50529951e-01 5.31191111e-01 -1.87721729e+00 -2.49903172e-01
1.01360095e+00 8.03743362e-01 5.51887691e-01 8.64216983e-01
1.75232813e-01 1.19570088e+00 -3.17794621e-01 3.46136749e-01
5.85768998e-01 -2.26896048e-01 -3.55019301e-01 -1.46942019e-01
7.35997632e-02 -8.77845645e-01 -4.91697252e-01 7.76866615e-01
5.39416611e-01 3.24805766e-01 2.80375093e-01 -1.66247547e+00
-3.27739306e-02 4.58839417e-01 -3.78311038e-01 -9.82496329e-03
2.43730471e-01 7.15202272e-01 8.17027688e-01 -6.00757539e-01
7.75170267e-01 1.29767704e+00 2.66837358e-01 5.71860552e-01
-1.10799670e+00 -7.32780516e-01 3.51533204e-01 3.56329858e-01
-7.08480835e-01 -5.06762981e-01 4.48809534e-01 -3.46913934e-01
9.84301090e-01 -4.15609628e-01 8.98864388e-01 1.25960922e+00
3.23140740e-01 1.31091917e+00 1.55174446e+00 -1.07527852e-01
7.63615668e-01 -1.32597899e-02 -3.27553302e-01 8.65581155e-01
1.05886005e-01 9.20747876e-01 -6.13804460e-01 9.02353823e-02
5.71992159e-01 -2.16049537e-01 1.90222971e-02 -7.85050213e-01
-1.34881461e+00 9.50120807e-01 4.77067053e-01 -2.49184400e-01
-6.76613152e-01 6.11486018e-01 4.36736315e-01 9.45714474e-01
-4.14608717e-02 5.31663239e-01 -3.99120718e-01 -5.55814385e-01
-5.06802380e-01 6.53325975e-01 9.03572440e-01 8.61088932e-01
9.21477616e-01 4.06037778e-01 1.35498747e-01 3.34494859e-01
1.97060764e-01 8.88029516e-01 3.68934333e-01 -1.62920070e+00
2.36406997e-01 3.51572931e-02 7.34763920e-01 -4.21299070e-01
-5.26185751e-01 -3.12786251e-01 -3.03110778e-01 1.13367701e+00
3.21568459e-01 -5.10196209e-01 -7.47192502e-01 1.75457561e+00
4.43198264e-01 1.06405228e-01 5.66606283e-01 1.03516960e+00
2.99102932e-01 4.71941769e-01 -1.40416399e-01 -1.62532106e-02
8.08732986e-01 -1.14643717e+00 -4.34433311e-01 -4.20664519e-01
5.80572069e-01 -2.66574174e-01 9.09901917e-01 6.43712282e-01
-7.94201374e-01 -5.62096238e-01 -1.19865680e+00 2.14051887e-01
-1.97188184e-01 5.79269491e-02 7.28959203e-01 2.15308100e-01
-1.21837699e+00 7.39259303e-01 -1.11265659e+00 -3.24970216e-01
2.74677485e-01 3.87798369e-01 -2.64589936e-01 -1.11203585e-02
-9.58329737e-01 1.25356352e+00 2.26871237e-01 -1.65595993e-01
-1.72717011e+00 -3.54017466e-01 -8.62422585e-01 -2.13445649e-01
8.26241016e-01 -4.08623844e-01 1.85888219e+00 -8.99113595e-01
-2.20111132e+00 3.48748863e-01 6.33733943e-02 -1.06613362e+00
8.83764744e-01 -4.66883957e-01 -1.38592429e-03 1.06839359e-01
2.61305064e-01 1.12481606e+00 1.01527011e+00 -1.24082565e+00
-1.07297456e+00 -6.97499737e-02 4.83457148e-01 4.31921363e-01
3.59532565e-01 -5.68563938e-01 5.15768165e-03 -3.73154357e-02
-3.69686782e-01 -1.38643003e+00 -4.75244433e-01 1.74907982e-01
1.73569277e-01 -3.31509560e-01 8.02474141e-01 -1.96199268e-01
3.05665642e-01 -2.00239801e+00 1.51843011e-01 9.51417461e-02
1.55971691e-01 1.01639457e-01 -2.82147139e-01 6.72592938e-01
5.91322303e-01 -5.22473395e-01 -7.82941580e-02 -2.57069319e-01
3.42198849e-01 6.47059798e-01 -5.49825370e-01 6.21483445e-01
3.82794626e-02 9.80992615e-01 -1.57618666e+00 -2.19284087e-01
4.72524971e-01 2.25263871e-02 -4.04544473e-01 3.47717166e-01
-5.38694918e-01 8.24573040e-01 -6.16842806e-01 4.91125524e-01
1.95135072e-01 1.24956779e-01 7.18579441e-02 5.43284833e-01
-5.35642982e-01 2.11696580e-01 -1.01479232e+00 1.85004294e+00
-5.88428974e-01 8.81321192e-01 4.15941566e-01 -9.44939256e-01
7.46363759e-01 1.29431427e-01 3.65527570e-01 -9.46716666e-01
8.80736709e-02 4.68609273e-01 3.07075262e-01 -4.73736763e-01
6.17482781e-01 1.66950643e-01 -8.97676200e-02 4.68824595e-01
-9.21502486e-02 -5.06121695e-01 2.38040909e-01 2.83278912e-01
1.34058130e+00 6.76007688e-01 2.75288969e-01 -1.77960381e-01
1.41146511e-01 4.77315813e-01 4.99902010e-01 1.22335994e+00
-5.37425280e-01 -2.80127913e-01 3.07906538e-01 -2.31270537e-01
-1.02625191e+00 -9.97976899e-01 3.73264462e-01 1.07142627e+00
3.82448345e-01 -8.31753463e-02 -3.99567097e-01 -6.67947292e-01
3.78093123e-01 8.61881375e-01 -6.02236152e-01 -8.99903197e-03
-5.32983184e-01 1.23539165e-01 3.71929497e-01 3.92186850e-01
6.70766592e-01 -1.33970022e+00 -1.52241755e+00 4.50090975e-01
2.15688124e-01 -1.09872568e+00 1.86207741e-02 3.05468470e-01
-6.15584314e-01 -1.01632380e+00 -5.33633292e-01 -2.65808553e-01
3.48705113e-01 5.49485326e-01 1.00324345e+00 -2.61865944e-01
-1.07561715e-01 8.45045030e-01 -4.19987202e-01 -6.73358381e-01
-7.07156479e-01 -1.27571970e-01 3.44305426e-01 -3.25312227e-01
-7.47225136e-02 -4.74318773e-01 -6.42477453e-01 1.38397187e-01
-4.22346443e-01 8.17568600e-02 1.07596457e+00 1.07174158e+00
4.74287808e-01 -1.95588067e-01 6.24480486e-01 -4.69127625e-01
7.21711159e-01 -5.05368710e-01 -1.11367619e+00 -2.89624512e-01
-7.34966397e-01 4.96328562e-01 7.15366006e-01 -4.18553740e-01
-8.81277919e-01 1.42767712e-01 9.36129093e-02 -5.83612800e-01
-1.99135676e-01 3.40044439e-01 5.19032657e-01 -1.75834537e-01
5.20205259e-01 2.13520586e-01 4.12472337e-01 -4.77040224e-02
5.66525996e-01 4.41338480e-01 6.50639415e-01 -7.07258999e-01
7.67759144e-01 6.14259481e-01 1.60746083e-01 -8.00288916e-01
-4.50457364e-01 -3.88347656e-01 -1.32334545e-01 -5.94530642e-01
6.19050086e-01 -1.13398099e+00 -1.16002893e+00 4.04079676e-01
-8.39653909e-01 -1.11021626e+00 -5.23416638e-01 7.64076531e-01
-1.06211138e+00 1.53994665e-01 -3.31323981e-01 -1.09859240e+00
2.19199602e-02 -1.27987504e+00 9.39057350e-01 2.47225016e-01
7.16298074e-02 -6.21137500e-01 2.66469330e-01 -9.42999050e-02
4.98149246e-01 2.57792413e-01 1.42917752e-01 -2.54024923e-01
-8.40218544e-01 2.15522945e-01 4.18105014e-02 1.71759322e-01
-2.88146138e-01 -2.78542906e-01 -8.67113590e-01 -7.56736875e-01
-3.04385304e-01 -9.51616287e-01 8.84296834e-01 1.45612419e-01
6.21397555e-01 -2.95735653e-02 -2.36039817e-01 1.67396486e-01
1.21958768e+00 1.53706193e-01 2.42821306e-01 6.60131156e-01
2.36420721e-01 6.10651076e-01 1.25296915e+00 6.71892405e-01
6.32796586e-01 5.81199646e-01 9.19832110e-01 8.43533427e-02
8.95174816e-02 -4.84992534e-01 9.43701148e-01 4.57663596e-01
-6.46759346e-02 1.54804558e-01 -6.16131008e-01 6.18683875e-01
-2.35744572e+00 -9.66823876e-01 2.75897890e-01 2.01409674e+00
5.56681991e-01 3.78175586e-01 2.51052409e-01 -2.04510584e-01
5.83120100e-02 -4.92524877e-02 -1.22558188e+00 -3.80197257e-01
6.49425238e-02 1.08013541e-01 8.81796718e-01 6.46862090e-01
-8.66737545e-01 1.09504974e+00 5.71901846e+00 6.23113632e-01
-1.09487331e+00 1.94390640e-01 -1.50544092e-01 -2.45841607e-01
2.42587961e-02 2.17406437e-01 -6.20492876e-01 3.54884446e-01
8.99403393e-01 -1.84469998e-01 8.99377644e-01 1.16299748e+00
4.82266933e-01 -7.08072364e-01 -1.31498039e+00 8.84016514e-01
-3.27362120e-01 -1.06505156e+00 -5.92234254e-01 1.85991809e-01
7.74234116e-01 8.33481729e-01 1.16822533e-01 9.57717419e-01
1.33680069e+00 -9.00140822e-01 9.94559944e-01 3.49181175e-01
4.36903030e-01 -7.63242960e-01 5.57890117e-01 7.75959194e-01
-8.70336175e-01 -4.23930645e-01 -2.58275062e-01 -4.04482186e-01
-1.60622578e-02 9.23925489e-02 -1.04312575e+00 4.88677353e-01
5.44510245e-01 7.00109839e-01 9.59307142e-03 7.80061066e-01
-6.12321794e-01 4.26990658e-01 -4.97977346e-01 -3.71093303e-01
9.43222702e-01 -1.37465447e-01 7.15549588e-01 8.06755722e-01
2.07140192e-01 -9.00184587e-02 6.77108049e-01 6.49154544e-01
2.31257632e-01 -2.85796195e-01 -1.05686772e+00 1.57471955e-01
4.58354175e-01 1.08670485e+00 -2.95334309e-01 -4.42899913e-01
-3.27488154e-01 8.13963175e-01 5.81622183e-01 3.72794092e-01
-8.14324737e-01 -1.47268802e-01 8.49415064e-01 -3.67682010e-01
5.50701559e-01 -6.59812689e-01 3.73617083e-01 -9.42182064e-01
-8.54273066e-02 -9.06722367e-01 -1.39777213e-01 -7.71940470e-01
-7.90762305e-01 2.95379579e-01 -1.49281463e-02 -1.36830556e+00
-7.34185636e-01 -4.97400016e-01 -3.52374166e-01 3.94912601e-01
-2.16735363e+00 -8.32517147e-01 -3.63429248e-01 5.41892290e-01
7.93410182e-01 -3.47591519e-01 5.73831618e-01 -1.66286677e-01
-1.79718584e-01 2.57460207e-01 3.42576712e-01 -3.26146036e-01
6.24757230e-01 -1.40313137e+00 3.46697658e-01 5.27793109e-01
-1.30944997e-01 3.01942416e-02 1.04496384e+00 -4.88499790e-01
-1.92546213e+00 -9.13857698e-01 3.39578115e-03 -8.24605376e-02
9.12127137e-01 -1.69201046e-01 -4.11562324e-01 6.48598135e-01
3.93270433e-01 -1.50991708e-01 -1.60124332e-01 -2.27401882e-01
-9.69830602e-02 -1.27454653e-01 -9.58698213e-01 8.21350694e-01
7.90208995e-01 -3.31155121e-01 -5.70086837e-01 2.27900073e-01
5.65383911e-01 -5.88817060e-01 -4.27606553e-01 2.53481120e-01
7.22687364e-01 -1.11006355e+00 5.50351918e-01 -4.18714672e-01
1.73252553e-01 -3.70545566e-01 -6.84292242e-02 -1.79181838e+00
8.50940794e-02 -9.58738863e-01 -2.94888765e-01 3.25219393e-01
1.60776064e-01 -1.03669310e+00 6.70329869e-01 1.11969203e-01
-2.70466417e-01 -6.30786240e-01 -1.12864959e+00 -1.28911853e+00
-9.12024304e-02 -4.10002977e-01 2.81638086e-01 4.49887604e-01
4.23586853e-02 1.95689857e-01 -5.17984152e-01 1.14343286e-01
8.70000899e-01 1.05988592e-01 1.37933445e+00 -1.01288998e+00
-5.89053929e-01 -3.86241764e-01 -6.77668527e-02 -1.41965628e+00
5.22144079e-01 -6.50399208e-01 5.66653848e-01 -1.43633616e+00
-3.63975435e-01 -7.08269000e-01 -9.40364152e-02 4.09042180e-01
3.02953660e-01 -4.00595337e-01 4.20087695e-01 2.61423349e-01
-8.92552137e-01 8.79467547e-01 1.38080585e+00 -1.66511878e-01
-2.90025949e-01 -8.83481205e-02 -1.76261678e-01 6.60546362e-01
1.02709234e+00 -2.90525317e-01 -5.86223364e-01 -3.24927717e-01
1.43428504e-01 3.77350360e-01 4.84770477e-01 -1.12307405e+00
4.19987112e-01 -2.40114450e-01 -1.71745330e-01 -6.16768837e-01
5.52950323e-01 -7.61581302e-01 -2.85371482e-01 8.73706162e-01
-3.29328179e-01 1.37360349e-01 1.98481515e-01 1.02066743e+00
9.47124362e-02 1.08679268e-03 6.36738658e-01 -2.92806953e-01
-1.25083172e+00 9.09788311e-02 -8.42432976e-01 4.87752259e-02
1.22807753e+00 1.08880242e-02 -3.21527392e-01 -7.33731627e-01
-4.90080386e-01 9.56249416e-01 5.83478749e-01 3.79572839e-01
5.63830733e-01 -9.89109993e-01 -7.60589778e-01 5.21588176e-02
5.41742668e-02 -1.11530274e-01 -4.27414514e-02 1.00068581e+00
-2.33097732e-01 4.01897520e-01 -4.75718528e-01 -7.03630745e-01
-8.61074507e-01 5.36213160e-01 2.70538270e-01 -3.81132066e-01
-8.72827649e-01 3.38779747e-01 -8.46943539e-03 -5.63830674e-01
3.61062557e-01 -4.50671405e-01 1.99614409e-02 -2.94388562e-01
2.18908355e-01 4.37374473e-01 -2.86488682e-01 -1.51635066e-01
-7.52293691e-02 2.45882586e-01 -7.70873055e-02 -6.41469955e-01
1.14918303e+00 -8.50338489e-02 4.69342768e-01 6.30143344e-01
6.17570281e-01 -2.70266980e-01 -2.07130194e+00 -1.63365513e-01
-1.11157663e-01 -2.86333561e-01 1.21893898e-01 -6.92294478e-01
-5.99679768e-01 7.37757444e-01 5.59808016e-01 6.96214586e-02
5.53123891e-01 -1.60085797e-01 7.17235744e-01 1.11091173e+00
9.49305236e-01 -1.46131432e+00 1.48281589e-01 6.83056653e-01
8.92006814e-01 -1.67342627e+00 -1.60956576e-01 3.62453818e-01
-7.77468741e-01 8.23018312e-01 6.43023252e-01 -5.72001159e-01
2.45174527e-01 2.13360071e-01 9.36464965e-02 -1.30843674e-03
-1.22670197e+00 -6.48111403e-01 -5.06972551e-01 6.90748215e-01
-4.33280945e-01 2.66244113e-01 5.37092835e-02 -1.47624269e-01
-1.57946497e-01 1.68982923e-01 6.61537349e-01 1.30000949e+00
-7.88516343e-01 -8.23516488e-01 -1.37508512e-01 2.52336174e-01
3.28336954e-02 3.00559849e-01 -2.60348916e-02 1.13326895e+00
-2.03096464e-01 1.01693392e+00 -1.99107174e-02 -2.02754527e-01
2.12145746e-01 -4.28445667e-01 6.51324987e-01 -2.93446511e-01
-2.98181266e-01 -2.24431112e-01 2.73694038e-01 -1.01180887e+00
-3.67101938e-01 -7.34774411e-01 -1.45201421e+00 -3.07372540e-01
1.17845580e-01 1.26140296e-01 9.68505979e-01 9.38841760e-01
4.58814263e-01 4.67534602e-01 7.99096346e-01 -1.30333412e+00
-1.15317726e+00 -6.66724026e-01 -3.33749682e-01 -1.87185295e-02
8.67064416e-01 -1.10724819e+00 -3.72128636e-01 -5.12141883e-01] | [4.700588703155518, 1.228374719619751] |
b9e81ce0-abcc-452e-909a-485373966b24 | efficient-block-contrastive-learning-via | 2209.14067 | null | https://arxiv.org/abs/2209.14067v1 | https://arxiv.org/pdf/2209.14067v1.pdf | Efficient block contrastive learning via parameter-free meta-node approximation | Contrastive learning has recently achieved remarkable success in many domains including graphs. However contrastive loss, especially for graphs, requires a large number of negative samples which is unscalable and computationally prohibitive with a quadratic time complexity. Sub-sampling is not optimal and incorrect negative sampling leads to sampling bias. In this work, we propose a meta-node based approximation technique that can (a) proxy all negative combinations (b) in quadratic cluster size time complexity, (c) at graph level, not node level, and (d) exploit graph sparsity. By replacing node-pairs with additive cluster-pairs, we compute the negatives in cluster-time at graph level. The resulting Proxy approximated meta-node Contrastive (PamC) loss, based on simple optimized GPU operations, captures the full set of negatives, yet is efficient with a linear time complexity. By avoiding sampling, we effectively eliminate sample bias. We meet the criterion for larger number of samples, thus achieving block-contrastiveness, which is proven to outperform pair-wise losses. We use learnt soft cluster assignments for the meta-node constriction, and avoid possible heterophily and noise added during edge creation. Theoretically, we show that real world graphs easily satisfy conditions necessary for our approximation. Empirically, we show promising accuracy gains over state-of-the-art graph clustering on 6 benchmarks. Importantly, we gain substantially in efficiency; up to 3x in training time, 1.8x in inference time and over 5x in GPU memory reduction. | ['Shekhar S. Chandra', 'Marius Portmann', 'Gayan K. Kulatilleke'] | 2022-09-28 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [ 1.87133893e-01 5.02876081e-02 -3.61130714e-01 -1.66307420e-01
-9.16104674e-01 -5.88015854e-01 3.70677054e-01 6.17987692e-01
-4.01540846e-01 6.74048364e-01 -4.17052925e-01 -4.09225345e-01
-7.14860633e-02 -1.06066346e+00 -9.20903504e-01 -7.69540370e-01
-6.66562378e-01 6.67700529e-01 2.75920719e-01 1.68915242e-01
4.02421914e-02 3.85419428e-01 -1.31762981e+00 1.15831256e-01
9.90353286e-01 9.20391381e-01 -6.03994615e-02 7.42926359e-01
2.95404578e-03 7.03372538e-01 -4.75047708e-01 -6.23699963e-01
4.62895483e-01 -3.89527500e-01 -6.36930168e-01 4.74517681e-02
6.91053331e-01 -9.56981331e-02 -3.16861928e-01 1.31925368e+00
5.09080410e-01 3.96663882e-03 6.20097876e-01 -1.56877851e+00
-9.18730423e-02 7.97393978e-01 -1.29029810e+00 -8.79359394e-02
-1.16423912e-01 1.32525116e-01 1.21094084e+00 -5.26616096e-01
5.11344969e-01 1.14832866e+00 1.02526414e+00 4.07444507e-01
-1.53134775e+00 -7.39708841e-01 2.38871440e-01 -2.22809948e-02
-1.63485527e+00 4.60984707e-02 8.32957208e-01 -1.34820104e-01
7.72607863e-01 7.29040802e-01 9.12039816e-01 5.63037753e-01
-1.45521179e-01 7.57714808e-01 9.17609692e-01 -2.85140336e-01
3.35175633e-01 2.33088247e-02 1.56892553e-01 8.71176660e-01
5.45667171e-01 -4.67361920e-02 -1.51901066e-01 -4.93290216e-01
5.66715002e-01 1.36835203e-01 -2.33715564e-01 -4.77710098e-01
-6.30960166e-01 8.94991696e-01 7.86710083e-01 2.09916160e-02
-1.90892711e-01 5.64811945e-01 6.87222898e-01 3.68268996e-01
7.21906841e-01 1.69411719e-01 -5.12378633e-01 -8.85003731e-02
-1.11441982e+00 1.05723143e-01 9.41646993e-01 9.18545306e-01
9.14577246e-01 -1.39534518e-01 -1.37474924e-01 8.42870653e-01
-5.63048832e-02 4.38413531e-01 -9.39247757e-02 -8.94519925e-01
3.91144156e-01 7.09205866e-01 -3.68685335e-01 -1.36764860e+00
-4.53302443e-01 -8.92769575e-01 -1.49107790e+00 8.71224180e-02
4.41727042e-01 2.20557656e-02 -7.41832018e-01 2.00219965e+00
6.34042323e-01 3.32426608e-01 -5.62330544e-01 6.10567093e-01
3.34147811e-01 4.50778425e-01 -6.37897551e-02 -4.69860077e-01
1.24774230e+00 -1.01119280e+00 -3.00974637e-01 -1.66838408e-01
1.06762385e+00 -6.01382196e-01 1.32976913e+00 4.73842949e-01
-1.27678788e+00 -9.97255147e-02 -9.89203334e-01 8.61366745e-04
-8.09061080e-02 -9.19656232e-02 9.49873805e-01 8.82192731e-01
-1.24066019e+00 8.14559758e-01 -8.18233728e-01 -4.37960066e-02
7.12064147e-01 6.24755085e-01 -1.36380047e-01 -2.92437613e-01
-5.78713059e-01 1.09843336e-01 -1.13853416e-03 -2.06994846e-01
-6.25459790e-01 -1.19246566e+00 -7.60454834e-01 3.29564273e-01
6.61188304e-01 -6.40173733e-01 4.71522033e-01 -8.81324291e-01
-8.83450508e-01 8.62824976e-01 -2.41784438e-01 -6.00520790e-01
8.58079374e-01 1.15117252e-01 5.85941337e-02 1.13871843e-01
-2.37891469e-02 6.04742050e-01 6.54117405e-01 -1.38961720e+00
-3.91624689e-01 -4.97990519e-01 -9.54004657e-03 6.29948303e-02
-7.64688373e-01 -4.17493701e-01 -7.84547806e-01 -7.77325451e-01
-2.29675462e-03 -1.04621613e+00 -4.87559378e-01 2.96887517e-01
-4.44652349e-01 -1.70642897e-01 6.47115409e-01 -4.11371976e-01
1.39088368e+00 -2.16026211e+00 -2.23899767e-01 7.28301823e-01
8.02062213e-01 2.53936768e-01 -1.32999122e-01 1.98033541e-01
7.05253184e-02 3.49412620e-01 -5.81688464e-01 -4.88468617e-01
-2.03745849e-02 1.23670273e-01 1.13496460e-01 6.68142140e-01
-8.64634383e-03 6.95517957e-01 -1.13060665e+00 -5.31234682e-01
1.25794396e-01 4.50143784e-01 -1.18280864e+00 -1.04313232e-01
-5.18081598e-02 -2.65965849e-01 -9.39418152e-02 5.44748425e-01
1.14194584e+00 -6.72706008e-01 5.44837654e-01 -2.75181085e-01
5.36874890e-01 1.72486216e-01 -1.21700048e+00 1.13273478e+00
-4.83686626e-01 3.84534121e-01 5.03592670e-01 -1.19527614e+00
6.98387980e-01 -2.81444669e-01 4.45371449e-01 -4.00502443e-01
3.16195674e-02 2.85877228e-01 -9.70438570e-02 1.78360954e-01
2.81199902e-01 -9.90232900e-02 9.67918411e-02 4.43969935e-01
-1.37811810e-01 7.40787107e-03 3.30927402e-01 5.50232768e-01
1.60337603e+00 -4.34308171e-01 1.01579644e-01 -5.08263290e-01
4.45787370e-01 -1.19774178e-01 4.94857132e-01 8.20935071e-01
-1.52783677e-01 4.17989671e-01 9.91405785e-01 -2.32837483e-01
-8.69811237e-01 -1.03470778e+00 1.81524709e-01 9.31926727e-01
1.54257238e-01 -7.19942987e-01 -7.60778189e-01 -8.66113722e-01
5.71274832e-02 3.68722975e-01 -5.69942117e-01 -2.76624262e-01
-7.71190226e-01 -1.14557302e+00 5.37371874e-01 5.08299708e-01
2.27352157e-01 -3.46432120e-01 -1.75410174e-02 1.05682187e-01
1.40152246e-01 -8.13881338e-01 -7.58985400e-01 3.05743098e-01
-1.03288794e+00 -1.25866926e+00 -5.10520518e-01 -9.25069690e-01
1.08217502e+00 3.91511112e-01 1.58402240e+00 5.81474483e-01
-5.82499743e-01 1.32840902e-01 7.01401234e-02 6.36073528e-03
-1.72830686e-01 1.69500068e-01 -2.45111421e-01 -3.45436394e-01
2.21427023e-01 -8.52544427e-01 -8.27310681e-01 -3.05168014e-02
-6.33505583e-01 -1.73823968e-01 4.98380691e-01 1.13161659e+00
8.89259875e-01 2.71354884e-01 3.08604002e-01 -1.27049851e+00
4.93344486e-01 -4.62020427e-01 -6.97263360e-01 1.93530887e-01
-6.06423616e-01 -1.60256863e-01 9.39333260e-01 -4.87861902e-01
-5.13372481e-01 -6.77553043e-02 -1.54320180e-01 -5.52079380e-01
4.82091695e-01 3.10341239e-01 7.68131688e-02 -3.96498144e-01
6.78398609e-01 -6.84911013e-02 1.39067948e-01 -4.59242575e-02
2.50244379e-01 2.45395750e-01 2.96590745e-01 -7.19701827e-01
6.99959338e-01 7.88269639e-01 4.73132491e-01 -7.47282922e-01
-5.70150077e-01 -5.68258286e-01 -6.91299662e-02 -1.70750022e-01
1.39947534e-01 -9.74259079e-01 -1.06127191e+00 2.63943851e-01
-7.50677884e-01 -5.47714591e-01 -3.67894918e-01 1.44217581e-01
-1.29961520e-01 6.46630585e-01 -9.32564199e-01 -9.18170333e-01
-6.40182316e-01 -8.30373108e-01 9.39321637e-01 -3.94876540e-01
-1.01602629e-01 -1.19412196e+00 -3.30326051e-01 6.97968975e-02
1.44784331e-01 5.04584312e-01 9.91569281e-01 -4.64707464e-01
-5.96952796e-01 -3.11209381e-01 -5.10107696e-01 3.88719440e-01
1.01896310e-02 1.37661491e-03 -6.43325567e-01 -8.04546416e-01
-2.41232678e-01 -3.00410956e-01 1.03758609e+00 4.50752974e-01
1.58112454e+00 -5.55882037e-01 -5.53645730e-01 7.89810419e-01
1.80373323e+00 -5.09722054e-01 4.97372121e-01 -2.62511522e-01
1.06869996e+00 6.19720459e-01 5.37256479e-01 7.13160694e-01
3.05364728e-01 3.63397121e-01 5.14414668e-01 -4.18157727e-01
-2.95639277e-01 -1.40472382e-01 5.68348095e-02 9.09854531e-01
2.19526961e-01 -4.00628150e-01 -7.31797099e-01 8.10029984e-01
-1.68961084e+00 -6.97891712e-01 -5.30576050e-01 2.46622062e+00
1.03171349e+00 3.79714608e-01 3.11012954e-01 3.38678867e-01
7.83682287e-01 4.92489114e-02 -5.84677160e-01 -1.86055481e-01
-5.13414433e-03 4.38010931e-01 8.86207283e-01 6.61409616e-01
-8.43588173e-01 7.06133008e-01 5.63757992e+00 1.52625728e+00
-9.62308288e-01 4.90317158e-02 1.23264039e+00 -4.05678034e-01
-5.55937529e-01 -1.51399057e-03 -4.25865561e-01 4.85027105e-01
5.51792145e-01 -1.88020155e-01 6.65555060e-01 9.84589994e-01
-1.40888333e-01 1.19685121e-01 -1.15728915e+00 1.05326605e+00
-3.53941731e-02 -1.40182769e+00 1.68053154e-02 3.46160203e-01
8.53037655e-01 6.91462234e-02 -8.10113698e-02 2.76899517e-01
4.25939202e-01 -9.53129351e-01 4.19451922e-01 -2.27196753e-01
7.78310776e-01 -1.10488200e+00 6.75403297e-01 2.90220141e-01
-1.42897558e+00 1.47086442e-01 -5.13268411e-01 -7.57501945e-02
-6.07574172e-02 1.25710237e+00 -6.59735799e-01 3.53248298e-01
8.03457677e-01 5.02946675e-01 -4.71001506e-01 8.77883434e-01
2.32400402e-01 7.50793695e-01 -9.46921885e-01 -3.27251226e-01
2.80120105e-01 -3.54475647e-01 4.32858348e-01 1.28109527e+00
1.42374068e-01 -1.60469010e-01 2.74030834e-01 6.41053259e-01
-5.33940613e-01 1.74014822e-01 -3.63195926e-01 1.99102476e-01
8.19756806e-01 1.28959787e+00 -9.82882321e-01 -4.64323848e-01
-2.67336309e-01 9.85499561e-01 6.53705776e-01 1.19979486e-01
-9.44820762e-01 -4.22889978e-01 6.33802772e-01 3.79722804e-01
3.47566098e-01 9.67215374e-02 -5.89300871e-01 -8.43757451e-01
3.48720074e-01 -7.95474708e-01 4.98321623e-01 9.41602737e-02
-1.57949352e+00 2.59697616e-01 -2.63496518e-01 -9.96307492e-01
1.39713868e-01 -4.14690733e-01 -5.06254375e-01 5.09126067e-01
-1.25455165e+00 -1.00382996e+00 -4.07977611e-01 4.60749120e-01
-2.86391377e-02 4.16055650e-01 4.65067059e-01 8.14139783e-01
-4.39668030e-01 1.26606727e+00 1.27742231e-01 -9.10077244e-03
5.43932021e-01 -1.38870406e+00 3.71705800e-01 6.22628272e-01
-1.38246968e-01 6.05112851e-01 5.67803442e-01 -6.28770769e-01
-1.32000828e+00 -1.32091141e+00 6.69413865e-01 1.13104820e-01
5.59821486e-01 -6.09497190e-01 -9.72911716e-01 3.48626763e-01
-1.76380187e-01 4.72156197e-01 6.68103099e-01 2.34992057e-01
-3.99152935e-01 -3.79571408e-01 -1.55836809e+00 8.19760144e-01
1.44101965e+00 -2.80645758e-01 3.75096589e-01 6.91439450e-01
7.20643163e-01 -2.59922177e-01 -9.19273138e-01 5.72459877e-01
2.94428885e-01 -9.46273327e-01 1.12366521e+00 -2.66484499e-01
1.91365644e-01 -2.16334879e-01 4.58112732e-02 -9.39742684e-01
-3.00813407e-01 -9.02789414e-01 -3.53446990e-01 1.20053208e+00
3.62623483e-01 -8.05501461e-01 1.24831474e+00 2.88903266e-01
7.32232444e-03 -1.19215572e+00 -8.95192325e-01 -9.59191918e-01
2.13564441e-01 -3.84988934e-01 5.01814187e-01 1.15490150e+00
-1.26555234e-01 2.94470757e-01 -2.31369570e-01 -6.96302056e-02
1.06738305e+00 -1.02433406e-01 8.26887190e-01 -1.20788753e+00
-6.08713865e-01 -6.75839722e-01 -3.86917979e-01 -1.12229156e+00
1.48391929e-02 -1.01722944e+00 -3.51050466e-01 -1.24656856e+00
4.45271343e-01 -9.03676808e-01 -2.24728566e-02 3.87685806e-01
-4.04939979e-01 6.04317605e-01 -1.65291857e-02 -3.21763046e-02
-6.41349852e-01 3.82454485e-01 1.13299012e+00 -1.89335078e-01
-1.08461209e-01 -2.71943867e-01 -6.32741392e-01 6.85666263e-01
8.21721077e-01 -5.09518385e-01 -4.05795515e-01 -1.29348800e-01
5.52513421e-01 -1.73675835e-01 2.88329720e-01 -9.08275306e-01
1.12130128e-01 2.75615782e-01 2.30979145e-01 -6.80784941e-01
3.31346214e-01 -7.09945858e-01 -6.33059721e-03 8.54322076e-01
-1.48240119e-01 -8.41411129e-02 1.90826371e-01 7.71801353e-01
-1.13719903e-01 3.76599990e-02 8.54100823e-01 -5.87441437e-02
-1.33955767e-02 5.80653667e-01 2.00685129e-01 4.88593787e-01
1.11085212e+00 -2.40651354e-01 -3.39950383e-01 -2.60313451e-01
-3.34668994e-01 4.62672830e-01 6.77564800e-01 -2.98203945e-01
2.91895896e-01 -1.29985201e+00 -6.67443812e-01 2.05004495e-02
-2.01652333e-01 1.34698436e-01 4.67591822e-01 9.63722169e-01
-8.06309879e-01 -2.90949494e-01 3.56010586e-01 -8.12285602e-01
-1.52770543e+00 6.14928186e-01 5.88431470e-02 -7.16989934e-01
-5.12371659e-01 1.24381757e+00 1.77864522e-01 -5.88939846e-01
3.58739614e-01 -3.03552356e-02 4.35515463e-01 -7.02886805e-02
3.62601221e-01 7.49781609e-01 1.40130773e-01 -3.43407728e-02
-4.22618568e-01 4.05997574e-01 -1.04490794e-01 2.56237805e-01
1.30830276e+00 1.51274398e-01 -5.36737323e-01 2.60405429e-02
1.69256914e+00 1.74792036e-01 -1.21775997e+00 -9.36867520e-02
-1.77592084e-01 -5.18316209e-01 5.51777817e-02 -3.67619783e-01
-1.40155053e+00 8.51210296e-01 4.74291265e-01 5.08190632e-01
1.36750066e+00 -1.72671024e-02 1.03295398e+00 1.62625983e-01
4.25556004e-01 -1.03345942e+00 -3.38045731e-02 6.13293750e-03
3.91521931e-01 -1.12209952e+00 4.49968874e-01 -1.11400199e+00
-1.07079223e-01 6.70201182e-01 6.23634100e-01 -5.55118442e-01
8.55365992e-01 5.67113638e-01 -5.48162580e-01 -2.02780694e-01
-8.52106452e-01 5.89376427e-02 -5.37772924e-02 5.69679618e-01
4.28272009e-01 4.71763998e-01 -5.95277488e-01 2.82496959e-01
-2.20400527e-01 -5.12496114e-01 2.33691081e-01 6.53930783e-01
-1.42731205e-01 -1.10211003e+00 -1.37987663e-05 9.43940818e-01
-5.65293729e-01 -4.00170445e-01 -4.30636287e-01 9.00423527e-01
-1.75120324e-01 7.70125866e-01 3.85809392e-01 -3.22870821e-01
2.31504757e-02 -6.30658388e-01 7.03246772e-01 -2.65675753e-01
-7.80263305e-01 5.39682694e-02 1.30557373e-01 -6.02044761e-01
-2.02984884e-01 -3.56801003e-01 -1.13921654e+00 -9.58052933e-01
-5.92983007e-01 1.08895563e-01 2.79365540e-01 3.32025141e-01
6.44112349e-01 5.30557513e-01 6.91781342e-01 -6.20780528e-01
-6.08289361e-01 -7.64340758e-01 -5.75700641e-01 4.18676317e-01
7.08921552e-02 -2.54924655e-01 -8.79112899e-01 -4.45667118e-01] | [7.0912766456604, 6.004348278045654] |
dbce96ca-89df-4cea-b0f4-d3b9aaa6343b | improving-inductive-link-prediction-using | 2107.04894 | null | https://arxiv.org/abs/2107.04894v1 | https://arxiv.org/pdf/2107.04894v1.pdf | Improving Inductive Link Prediction Using Hyper-Relational Facts | For many years, link prediction on knowledge graphs (KGs) has been a purely transductive task, not allowing for reasoning on unseen entities. Recently, increasing efforts are put into exploring semi- and fully inductive scenarios, enabling inference over unseen and emerging entities. Still, all these approaches only consider triple-based \glspl{kg}, whereas their richer counterparts, hyper-relational KGs (e.g., Wikidata), have not yet been properly studied. In this work, we classify different inductive settings and study the benefits of employing hyper-relational KGs on a wide range of semi- and fully inductive link prediction tasks powered by recent advancements in graph neural networks. Our experiments on a novel set of benchmarks show that qualifiers over typed edges can lead to performance improvements of 6% of absolute gains (for the Hits@10 metric) compared to triple-only baselines. Our code is available at \url{https://github.com/mali-git/hyper_relational_ilp}. | ['Jens Lehmann', 'Volker Tresp', 'Tengfei Ma', 'Veronika Thost', 'Mikhail Galkin', 'Max Berrendorf', 'Mehdi Ali'] | 2021-07-10 | null | null | null | null | ['inductive-link-prediction'] | ['graphs'] | [-9.25732702e-02 7.77357638e-01 -6.99100792e-01 -2.26793572e-01
-6.24736965e-01 -6.01213515e-01 6.78361833e-01 5.35253882e-01
-6.08513691e-02 9.73597169e-01 1.73829928e-01 -6.08706772e-01
-3.34178746e-01 -1.27469492e+00 -1.00772786e+00 -1.37890488e-01
-5.43942869e-01 7.88906574e-01 4.31608558e-01 -3.27539861e-01
-5.02210140e-01 -8.37879851e-02 -1.18051255e+00 1.48811966e-01
9.80855405e-01 7.57681191e-01 -3.95752966e-01 5.30004144e-01
-4.77287583e-02 1.04385042e+00 6.45268261e-02 -1.13944340e+00
9.71166417e-02 1.73718095e-01 -1.17650223e+00 -7.35873163e-01
5.08741915e-01 -2.72426684e-03 -7.71738708e-01 9.26490426e-01
4.92340088e-01 -1.33368805e-01 4.23850864e-01 -1.45761859e+00
-8.21930587e-01 1.49795103e+00 -3.25767666e-01 2.25546271e-01
6.06557190e-01 -2.98712641e-01 1.72740829e+00 -7.71831274e-01
8.43793631e-01 1.01491177e+00 1.02358639e+00 2.24720210e-01
-1.40892255e+00 -7.02481210e-01 1.10486716e-01 3.75268102e-01
-1.58307707e+00 -2.87482172e-01 6.25330925e-01 -3.32913816e-01
1.02112544e+00 2.99465388e-01 5.52329540e-01 1.07003653e+00
-3.02121043e-01 9.48366940e-01 8.39634478e-01 -4.81389254e-01
-1.74444512e-01 3.05435389e-01 3.93725276e-01 1.02132308e+00
9.08292770e-01 -1.06639549e-01 -5.95471084e-01 -9.53300372e-02
3.21388811e-01 -3.41691673e-01 -4.66096014e-01 -4.44085211e-01
-1.16030979e+00 6.20865285e-01 8.72966230e-01 2.32218429e-01
-2.18198687e-01 2.15586573e-01 3.98828477e-01 3.99280906e-01
7.87898064e-01 5.61031461e-01 -7.11470723e-01 1.29069760e-03
-5.59749544e-01 8.06520879e-02 1.33439541e+00 1.20524085e+00
7.84934759e-01 -3.11158210e-01 -1.02082923e-01 5.55066884e-01
1.83079600e-01 1.60264701e-01 -1.83437735e-01 -3.30949277e-01
8.14381599e-01 9.86228585e-01 -2.34015018e-01 -1.01287627e+00
-5.67625821e-01 -6.32575333e-01 -7.36678541e-01 -5.12941182e-01
4.84870493e-01 -3.12575459e-01 -8.04806888e-01 1.75485766e+00
5.33491313e-01 4.42148149e-01 1.48120876e-02 4.83286798e-01
1.38102126e+00 3.51741552e-01 3.71006757e-01 7.18870237e-02
1.20082390e+00 -8.80706906e-01 -4.29850310e-01 -1.75088808e-01
1.18935347e+00 -1.74396977e-01 8.61153007e-01 1.79466069e-01
-9.76790309e-01 -1.54812172e-01 -9.20585573e-01 -2.04450071e-01
-9.65297163e-01 -1.51058793e-01 1.14110756e+00 4.05181468e-01
-1.30330670e+00 5.63598156e-01 -7.67235279e-01 -6.08929217e-01
4.16050494e-01 4.52039540e-01 -6.21172011e-01 -8.13398063e-02
-1.99112606e+00 8.94572616e-01 9.54624832e-01 7.11543262e-02
-4.67113554e-01 -9.91877556e-01 -8.69694114e-01 1.17092445e-01
9.14559543e-01 -9.91618156e-01 7.46704519e-01 -4.05037850e-01
-8.57833564e-01 1.00145185e+00 1.58309624e-01 -6.88850462e-01
3.76076579e-01 -3.93701315e-01 -6.24507725e-01 -2.13626936e-01
-6.63589612e-02 5.33240318e-01 3.08625633e-03 -1.13672197e+00
-4.43074554e-01 -3.07838708e-01 5.87753296e-01 5.70295043e-02
-6.05304658e-01 -3.66703600e-01 -5.97546935e-01 -3.14376622e-01
-2.18445614e-01 -9.94349301e-01 4.79649343e-02 -5.77984571e-01
-1.04193938e+00 -5.01498342e-01 4.56679642e-01 -5.22723198e-01
1.60603726e+00 -1.65419185e+00 1.73273295e-01 3.14363033e-01
4.79633093e-01 3.98847163e-01 1.30838543e-01 7.55511701e-01
-2.13140294e-01 4.39638972e-01 1.54783964e-01 -5.47120050e-02
1.12598121e-01 1.73566937e-01 -2.96114743e-01 4.09945846e-02
1.52398884e-01 1.51374161e+00 -1.02247381e+00 -6.68076038e-01
-1.07737385e-01 3.83208364e-01 -4.76457238e-01 -5.92315495e-02
-4.77917790e-01 2.83850748e-02 -5.74043214e-01 8.86784732e-01
2.08079055e-01 -8.90024424e-01 5.26527405e-01 -2.45722815e-01
4.09803331e-01 4.57008392e-01 -9.61416185e-01 1.54180157e+00
-2.71224499e-01 4.23343092e-01 -2.76490897e-01 -8.47867072e-01
6.73327923e-01 1.75003231e-01 2.70904958e-01 -3.60424578e-01
-1.26985693e-02 2.26562321e-01 -9.79226362e-03 -3.91007602e-01
5.96335471e-01 2.30439201e-01 7.70894857e-03 8.63967016e-02
1.91798747e-01 3.22150767e-01 4.63953346e-01 8.53734851e-01
1.50343668e+00 3.74714971e-01 2.86112696e-01 -1.51019737e-01
2.73239374e-01 2.00534701e-01 3.88969570e-01 6.89706981e-01
1.82640806e-01 8.08343440e-02 6.92720175e-01 -2.93083549e-01
-6.85496271e-01 -1.08371556e+00 -2.31174648e-01 1.26261497e+00
2.53648460e-01 -7.85194457e-01 -3.52411777e-01 -8.77620101e-01
4.54994261e-01 6.41559422e-01 -7.72303283e-01 -2.16696814e-01
-4.01524365e-01 -1.00508845e+00 8.11034024e-01 6.48707986e-01
2.37818480e-01 -8.08109403e-01 3.07503223e-01 3.58346030e-02
-1.30165741e-01 -1.41748500e+00 2.18411297e-01 1.51149333e-01
-7.52687633e-01 -1.21471298e+00 -1.24146871e-01 -5.90686977e-01
6.15462720e-01 -1.88933298e-01 1.69119835e+00 1.84364587e-01
-2.58708715e-01 4.19217497e-01 -4.40709442e-01 -2.69246012e-01
-1.59513533e-01 7.15974092e-01 -5.00789704e-03 -1.98063016e-01
5.24782717e-01 -7.85588801e-01 -5.29130340e-01 3.66098285e-02
-5.70502341e-01 2.37696201e-01 6.59340560e-01 6.41227186e-01
4.56975639e-01 -3.09865754e-02 8.12472463e-01 -1.68893492e+00
4.36711937e-01 -9.80145931e-01 -3.97377789e-01 5.11762440e-01
-1.07374740e+00 1.52981341e-01 3.72984648e-01 -1.53485015e-01
-9.72482443e-01 -4.08369899e-01 -5.63950576e-02 -1.48196489e-01
1.37847746e-02 1.08090127e+00 -2.84991693e-02 1.33627253e-02
7.79706776e-01 -1.95093900e-01 -4.72278714e-01 -3.03773999e-01
6.72696173e-01 3.56899798e-01 3.82505983e-01 -8.40063095e-01
1.10309839e+00 3.33258837e-01 1.33795604e-01 -3.33678156e-01
-1.10112250e+00 -4.64885592e-01 -6.09895349e-01 -1.53978214e-01
6.34686828e-01 -1.10018277e+00 -6.43217444e-01 8.77972543e-02
-7.45055497e-01 -6.40940905e-01 -2.08758101e-01 3.41266692e-01
-5.31644672e-02 1.76831961e-01 -8.17469180e-01 -5.29947042e-01
-4.81165856e-01 -5.47094762e-01 7.95648575e-01 1.79755136e-01
-2.24897221e-01 -1.54465592e+00 1.71313778e-01 6.44747972e-01
2.71962255e-01 6.06952906e-01 1.14920557e+00 -1.10597122e+00
-7.58930564e-01 -3.59119296e-01 -3.38268310e-01 -5.98903224e-02
-6.97227120e-02 1.82976795e-03 -8.28811228e-01 -2.22403869e-01
-1.29531944e+00 -6.45417571e-01 1.05961108e+00 -1.27829269e-01
8.63843977e-01 -2.71766007e-01 -9.39984381e-01 5.48977613e-01
1.51849926e+00 -4.19987053e-01 5.69261968e-01 2.80718803e-01
1.09972978e+00 3.96915853e-01 4.89484251e-01 7.35091567e-02
1.15536022e+00 4.62936819e-01 6.37390435e-01 -1.86394513e-01
-1.09112732e-01 -6.06292963e-01 4.97787371e-02 6.81257904e-01
-5.19500732e-01 -5.37992477e-01 -1.22888374e+00 8.99150014e-01
-1.91867375e+00 -7.54282892e-01 -5.70186734e-01 2.11780334e+00
1.24898899e+00 3.66782993e-01 7.77100474e-02 -1.41283274e-01
7.87839532e-01 7.48989508e-02 -5.38901925e-01 1.58862099e-02
-2.37717763e-01 1.54745683e-01 5.29419363e-01 4.74005669e-01
-1.23650110e+00 1.10877287e+00 5.19442654e+00 6.91886246e-01
-6.76898241e-01 3.08879185e-02 5.81569970e-01 4.92093991e-03
-5.47988236e-01 3.34008992e-01 -1.00250638e+00 1.39773279e-01
1.02037156e+00 -3.88206899e-01 2.17284590e-01 7.84561872e-01
-5.75257480e-01 4.50212769e-02 -1.38153565e+00 4.87713397e-01
-2.54883856e-01 -1.39132023e+00 6.32488495e-03 1.53692424e-01
8.37826848e-01 4.48771983e-01 -7.49053285e-02 9.42338109e-01
9.62775350e-01 -1.03747296e+00 2.36641020e-01 6.17346406e-01
7.52219021e-01 -4.03387219e-01 7.34830797e-01 9.53663811e-02
-1.34521711e+00 1.65324926e-01 3.78704444e-03 1.04065940e-01
-2.04311144e-02 9.13958907e-01 -1.07412016e+00 1.23203993e+00
8.40764165e-01 8.74939620e-01 -8.33199799e-01 7.48068810e-01
-3.64647597e-01 9.46040511e-01 -4.51137275e-01 7.80323967e-02
2.48344317e-02 1.52328596e-01 5.07574856e-01 1.41143024e+00
4.92332317e-02 1.65006340e-01 5.04815541e-02 7.48581409e-01
-6.46411002e-01 1.69049144e-01 -7.88917184e-01 -2.80745596e-01
6.21020913e-01 1.56303978e+00 -3.87645513e-01 -2.18313202e-01
-5.43596029e-01 3.24675024e-01 9.47713375e-01 4.43242788e-01
-9.09488380e-01 -4.01060879e-01 2.79686332e-01 4.63561624e-01
2.17944130e-01 7.77411014e-02 9.05911624e-02 -1.21895230e+00
3.78324352e-02 -3.20105940e-01 1.05216742e+00 -6.39592052e-01
-1.48526168e+00 3.81194621e-01 1.77462488e-01 -7.25259066e-01
-1.63365111e-01 -5.37164688e-01 -4.84297752e-01 5.01633584e-01
-1.70214713e+00 -1.57726419e+00 -4.01965767e-01 4.12165374e-01
-3.43851149e-01 1.82206243e-01 6.92378759e-01 5.54333031e-01
-6.30059600e-01 7.46753335e-01 -1.62365884e-01 4.97489870e-01
7.27194250e-01 -1.64453447e+00 4.24263656e-01 6.28032148e-01
3.49531114e-01 6.84904158e-01 6.30750179e-01 -8.26688111e-01
-1.63845885e+00 -1.34457481e+00 1.00085437e+00 -8.69805753e-01
1.12286246e+00 -4.91513520e-01 -1.11977279e+00 1.31279647e+00
1.57285213e-01 4.23615634e-01 7.43260264e-01 1.01652634e+00
-6.04845107e-01 -2.11067587e-01 -8.35802019e-01 5.67769468e-01
1.60630882e+00 -3.50196451e-01 -3.35791856e-01 5.14083028e-01
8.98416519e-01 -4.89334047e-01 -1.67716479e+00 9.48983312e-01
3.32952678e-01 -7.15243995e-01 1.14930272e+00 -7.97269464e-01
4.47998106e-01 -1.80122510e-01 2.13501275e-01 -1.21720302e+00
-3.68385971e-01 -5.57851732e-01 -7.21365750e-01 1.49178910e+00
9.68887329e-01 -8.61021817e-01 9.03254628e-01 5.77449977e-01
1.94379827e-03 -1.09568477e+00 -5.68886578e-01 -6.65439665e-01
-2.36829352e-02 -2.57383913e-01 4.36590731e-01 1.36573052e+00
4.05121267e-01 5.74156761e-01 -1.45479426e-01 3.45572859e-01
6.84120834e-01 2.87217677e-01 7.86483049e-01 -1.50824857e+00
-4.29651678e-01 -2.97101498e-01 -6.21164858e-01 -6.60529375e-01
3.67662311e-01 -1.56062639e+00 -4.79095459e-01 -1.92717731e+00
3.20386738e-01 -8.77056360e-01 -3.99424344e-01 9.97246087e-01
-4.13346678e-01 3.27394307e-01 -3.24603438e-01 4.88713244e-03
-1.03793359e+00 3.51724595e-01 8.92043531e-01 -8.67469832e-02
-3.50633487e-02 -2.51302868e-01 -8.67845237e-01 5.35338998e-01
6.74980521e-01 -1.79300174e-01 -5.65185845e-01 -4.00569320e-01
1.10525537e+00 -8.79381672e-02 4.84285116e-01 -8.37654054e-01
5.15953481e-01 1.81312665e-01 -1.03898514e-02 -3.34505826e-01
1.78573087e-01 -5.00200033e-01 4.80351925e-01 1.96362048e-01
-3.58053148e-01 -2.82007188e-01 3.35731387e-01 8.64221275e-01
-2.08140820e-01 1.82523012e-01 1.12670630e-01 5.62678650e-02
-8.06059837e-01 5.12243688e-01 6.40697241e-01 5.48943698e-01
1.13479316e+00 9.61579755e-03 -8.26920927e-01 -2.61072725e-01
-9.44996059e-01 6.82445347e-01 1.11349069e-01 4.41836983e-01
2.15170071e-01 -1.20926881e+00 -9.10963655e-01 -3.28789294e-01
4.61559325e-01 3.03847343e-01 3.05446275e-02 1.16200924e+00
-2.84549057e-01 5.19076169e-01 3.61336976e-01 -4.05449510e-01
-1.18004954e+00 5.85523844e-01 1.71457410e-01 -7.73278594e-01
-5.20839393e-01 1.03377199e+00 -1.21828439e-02 -6.87348247e-01
1.42503470e-01 -1.32287487e-01 -1.32670477e-01 5.00032715e-02
-1.70206159e-01 3.87687713e-01 1.12517655e-01 -3.42027158e-01
-4.70516473e-01 1.32772148e-01 -2.65089571e-01 5.54475725e-01
1.44944775e+00 4.27237265e-02 -2.28770331e-01 4.12789941e-01
8.58655274e-01 -3.35262460e-03 -5.98750412e-01 -6.60844266e-01
3.86850536e-01 1.13427509e-02 -7.42504746e-02 -1.08301830e+00
-1.00602829e+00 4.76459831e-01 -2.00054303e-01 6.11832798e-01
7.87510276e-01 5.78801572e-01 6.11089170e-01 4.99058634e-01
6.66678309e-01 -7.01230526e-01 -3.96674365e-01 3.28334838e-01
5.68110108e-01 -1.37760174e+00 8.42648596e-02 -8.84498358e-01
-4.42652017e-01 7.04864502e-01 8.32280457e-01 1.46940425e-01
7.67093778e-01 2.00908065e-01 -5.41484177e-01 -5.52985251e-01
-1.07787156e+00 -5.38524866e-01 3.66662502e-01 3.04578692e-01
6.02955401e-01 2.72423714e-01 5.93727343e-02 4.88997370e-01
-2.34697327e-01 6.55094385e-02 1.88511699e-01 5.85802853e-01
-4.55505289e-02 -1.09636426e+00 2.84088224e-01 7.78690398e-01
-4.79646415e-01 -5.45988083e-01 -5.54811895e-01 1.11081719e+00
-5.56052336e-03 7.21852601e-01 -2.54695356e-01 -6.38150156e-01
3.58850241e-01 1.57741442e-01 4.42041069e-01 -6.54903591e-01
-4.44685340e-01 -7.49067485e-01 8.69719625e-01 -3.02844971e-01
-3.58727485e-01 -4.40944701e-01 -1.26064074e+00 -5.27672052e-01
-5.65180600e-01 3.09733957e-01 1.52983621e-01 5.75422525e-01
5.75822175e-01 4.91994798e-01 4.87430813e-03 -1.55821428e-01
-1.57973513e-01 -9.86142516e-01 -4.59760576e-01 4.55832571e-01
-1.49111509e-01 -6.80209935e-01 -3.03267956e-01 -1.74045041e-01] | [8.893924713134766, 8.01285171508789] |
246c134c-bd6a-4693-a83c-297a0d70b635 | tencent-mvse-a-large-scale-benchmark-dataset | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zeng_Tencent-MVSE_A_Large-Scale_Benchmark_Dataset_for_Multi-Modal_Video_Similarity_Evaluation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zeng_Tencent-MVSE_A_Large-Scale_Benchmark_Dataset_for_Multi-Modal_Video_Similarity_Evaluation_CVPR_2022_paper.pdf | Tencent-MVSE: A Large-Scale Benchmark Dataset for Multi-Modal Video Similarity Evaluation | Multi-modal video similarity evaluation is important for video recommendation systems such as video de-duplication, relevance matching, ranking, and diversity control. However, there still lacks a benchmark dataset that can support supervised training and accurate evaluation. In this paper, we propose the Tencent-MVSE dataset, which is the first benchmark dataset for the multi-modal video similarity evaluation task. The Tencent-MVSE dataset contains video pairs similarity annotations, and diverse metadata including Chinese title, automatic speech recognition (ASR) text, as well as human-annotated categories/tags. We provide a simple baseline with a multi-modal Transformer architecture to perform supervised multi-modal video similarity evaluation. We also explore pre-training strategies to make use of the unpaired data. The whole dataset as well as our baseline will be released to promote the development of the multi-modal video similarity evaluation. The dataset has been released in https://tencent-mvse.github.io/. | ['Zhen Wen', 'Weidong Guo', 'Dian Li', 'Fengyun Rao', 'Zhenhua Liu', 'Yongsheng Luo', 'Zhaoyang Zeng'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['video-similarity'] | ['computer-vision'] | [ 5.00977598e-02 -8.27193022e-01 -4.28743362e-01 -4.84513372e-01
-1.28294098e+00 -6.08373284e-01 6.72703981e-01 -1.18743092e-01
-3.86319667e-01 2.54335850e-01 7.08412647e-01 1.41118944e-01
1.36674121e-01 -3.44411552e-01 -5.90864778e-01 -4.15502936e-01
9.03657749e-02 2.81293154e-01 5.45495212e-01 -1.92799002e-01
3.31616372e-01 -2.40954012e-01 -1.92917538e+00 8.12396765e-01
6.92507386e-01 1.30187714e+00 4.35668111e-01 6.61334813e-01
1.36695370e-01 7.79381633e-01 -2.01655805e-01 -7.09452868e-01
9.65461954e-02 -4.52975363e-01 -7.56410360e-01 -2.18989715e-01
8.08095634e-01 -4.51696545e-01 -5.81099629e-01 1.04754424e+00
8.09727907e-01 3.41608077e-01 6.84114099e-01 -1.56705248e+00
-7.28382111e-01 6.67750835e-01 -1.69447228e-01 3.08017910e-01
8.43577147e-01 1.35027349e-01 1.14599311e+00 -1.05975807e+00
7.17701674e-01 9.56846297e-01 5.32771170e-01 6.11318588e-01
-2.95406163e-01 -6.59553230e-01 -2.16326773e-01 8.33124995e-01
-1.62499225e+00 -8.19282055e-01 3.80342573e-01 -4.30512398e-01
8.36037755e-01 4.92849201e-01 3.50838006e-01 1.43338156e+00
-4.04331744e-01 9.86020803e-01 4.69079733e-01 -1.05701245e-01
-3.45809832e-02 -1.27040088e-01 -1.28700241e-01 5.16122460e-01
-3.47274959e-01 -1.05524123e-01 -7.09834874e-01 -1.87090784e-01
2.28185579e-01 2.19058380e-01 -4.22761083e-01 -3.64993066e-01
-1.53439212e+00 5.42193353e-01 -3.21663031e-03 3.18698496e-01
-1.54500395e-01 -5.91825694e-02 1.03368294e+00 5.03165483e-01
2.53510028e-01 1.94606245e-01 -2.24985167e-01 -6.51994407e-01
-9.83997881e-01 2.34557331e-01 4.56663638e-01 1.42289841e+00
4.19526368e-01 -4.25071716e-01 -5.88404536e-01 1.24043918e+00
2.52778679e-01 6.36542797e-01 7.53888845e-01 -1.35342526e+00
7.34669447e-01 1.57708779e-01 -2.78583199e-01 -8.79807413e-01
2.37737104e-01 1.73021510e-01 -7.08381474e-01 -6.31751478e-01
-1.09678982e-02 1.50938839e-01 -4.94744122e-01 1.65054202e+00
4.51331846e-02 6.53798819e-01 -1.40512064e-02 1.00257921e+00
1.37647569e+00 6.45052433e-01 -9.40154269e-02 -8.11286569e-02
1.19270730e+00 -1.32414865e+00 -5.22380292e-01 3.63264382e-01
8.32762003e-01 -1.05916870e+00 1.20505416e+00 1.40950084e-01
-1.00907338e+00 -5.98494053e-01 -5.96140265e-01 -8.61264318e-02
-2.19192520e-01 1.06928967e-01 2.01457039e-01 2.42396474e-01
-1.15333617e+00 4.03150141e-01 -5.54013550e-01 -7.50787020e-01
2.76724786e-01 -2.02842638e-01 -4.68663394e-01 -3.80075216e-01
-1.44476366e+00 5.08861661e-01 3.62015545e-01 -4.86375451e-01
-1.11261249e+00 -7.47809887e-01 -9.59363103e-01 -2.02364266e-01
5.80005646e-01 -6.67586744e-01 1.22687352e+00 -9.42078173e-01
-1.19698930e+00 1.01658165e+00 -6.73935488e-02 -5.36703095e-02
2.64599472e-01 -5.43721281e-02 -7.57488608e-01 3.26030374e-01
3.60958636e-01 7.69941926e-01 6.71439350e-01 -6.99198127e-01
-8.50011289e-01 -2.22878307e-01 1.88120186e-01 2.21849024e-01
-7.68440783e-01 4.82649982e-01 -1.14602518e+00 -1.06229353e+00
-2.72414386e-01 -1.03647137e+00 2.93031007e-01 -1.60975054e-01
-3.38062882e-01 -3.65479112e-01 6.42062545e-01 -7.74228692e-01
1.51917112e+00 -2.29241967e+00 1.89629972e-01 -2.02386640e-02
3.08540277e-02 2.12030068e-01 -6.91234410e-01 4.41819757e-01
-7.37883821e-02 2.32263226e-02 1.65426105e-01 -3.50057065e-01
1.87417731e-01 -1.95235342e-01 3.28140100e-03 9.93310511e-02
-1.40048325e-01 8.13909709e-01 -8.96992922e-01 -7.97708571e-01
-1.28473099e-02 3.18037271e-01 -7.87186742e-01 4.84647304e-01
1.43843153e-02 8.05476010e-02 -3.21185619e-01 1.03591716e+00
3.79725724e-01 -2.54628599e-01 -7.27031305e-02 -7.36271083e-01
2.00410992e-01 3.50257486e-01 -9.56566632e-01 2.19968700e+00
-2.32671857e-01 6.85736418e-01 -3.00904274e-01 -7.87185669e-01
4.99939799e-01 5.31999648e-01 8.17213893e-01 -8.79834592e-01
7.18000382e-02 1.50685638e-01 -5.42057633e-01 -7.73820817e-01
9.55082119e-01 4.20389771e-01 -1.08089581e-01 3.83583724e-01
2.09964752e-01 3.12352628e-01 6.65423870e-01 5.83464742e-01
1.12693143e+00 4.95850369e-02 -5.99379987e-02 4.87472452e-02
7.94646621e-01 -9.61049944e-02 6.51548743e-01 3.63442451e-01
-2.26917386e-01 9.12158191e-01 7.28681684e-02 -1.51490003e-01
-1.13110340e+00 -8.67297053e-01 8.81986171e-02 1.28663206e+00
3.29325557e-01 -1.06550586e+00 -6.99353993e-01 -6.79121375e-01
-1.36064410e-01 3.32533568e-01 -2.23823279e-01 -3.03618073e-01
-3.15312266e-01 -1.92942604e-01 7.60259867e-01 4.01373446e-01
4.66234356e-01 -7.70220935e-01 1.50972649e-01 -1.81651518e-01
-9.45712805e-01 -1.40036011e+00 -1.17354643e+00 -6.35413766e-01
-4.50509727e-01 -1.25549865e+00 -1.03128481e+00 -9.70865846e-01
1.15782373e-01 7.04570591e-01 1.34145641e+00 8.60434324e-02
1.26863435e-01 7.40833998e-01 -1.00680923e+00 3.46981704e-01
-3.98125917e-01 1.64729729e-01 2.07403317e-01 -2.15470314e-01
6.18022025e-01 -4.33244795e-01 -6.42206252e-01 8.17796171e-01
-8.97266388e-01 -1.97096705e-03 1.30281001e-01 6.48519695e-01
7.48101175e-01 -2.60168344e-01 3.62621546e-01 -3.59731585e-01
4.26742822e-01 -7.57102966e-01 -1.33774385e-01 6.31046355e-01
-2.09756777e-01 -3.79248261e-01 4.05049860e-01 -4.85770524e-01
-6.33159339e-01 -3.86200309e-01 -2.74308681e-01 -8.97333205e-01
-1.78709447e-01 5.88840902e-01 -3.57626736e-01 -6.14815764e-02
3.44406992e-01 3.65614772e-01 -2.52391607e-01 -6.28749132e-01
2.59750128e-01 1.12105596e+00 7.23992825e-01 -5.85983932e-01
4.29892719e-01 -3.72584872e-02 -5.60209215e-01 -5.42092681e-01
-5.44927537e-01 -7.20818639e-01 -4.22225237e-01 -3.12891185e-01
7.18545020e-01 -1.35048199e+00 -3.84737819e-01 3.51706326e-01
-1.00332701e+00 -1.42572552e-01 3.77196670e-02 6.74745321e-01
-6.65091097e-01 8.76881659e-01 -7.58281231e-01 -1.20627820e-01
-3.19390863e-01 -1.28489828e+00 1.09776032e+00 -4.42663766e-03
-1.73805088e-01 -6.16808474e-01 1.86149985e-01 9.57603693e-01
4.06156838e-01 -3.63893121e-01 4.02224004e-01 -7.54473567e-01
-4.25858885e-01 -1.90697476e-01 -3.37631144e-02 3.70385200e-01
1.57995578e-02 2.90013582e-01 -6.14079714e-01 -3.78519475e-01
-5.62875569e-01 -7.40220666e-01 9.74711239e-01 5.26971817e-02
1.44610250e+00 -2.27542385e-01 -1.85712516e-01 6.93010390e-01
9.76795375e-01 5.97255044e-02 6.48896396e-01 4.20795351e-01
7.85309196e-01 3.15607041e-01 1.07096624e+00 6.16413713e-01
1.03481019e+00 1.10084045e+00 2.02678487e-01 3.86111379e-01
-1.83127329e-01 -2.04206795e-01 7.00220168e-01 1.53865969e+00
-1.24786161e-01 -4.19846445e-01 -6.97827578e-01 5.64738750e-01
-1.72585177e+00 -1.47016346e+00 1.86719507e-01 2.18345857e+00
8.31155539e-01 -3.79960090e-01 4.77802277e-01 -1.20303079e-01
9.32402194e-01 1.92691237e-01 -2.86453724e-01 1.30761474e-01
-2.03491434e-01 -4.15136039e-01 1.29780203e-01 6.86345547e-02
-1.39339077e+00 7.97388792e-01 5.68269300e+00 1.38814640e+00
-9.70113695e-01 3.81500334e-01 3.79809976e-01 -2.79491693e-01
-5.13153374e-01 -2.81199545e-01 -7.05213904e-01 1.04653549e+00
9.77045119e-01 -2.93566883e-01 4.21224982e-01 7.08318233e-01
6.51348978e-02 1.71344832e-01 -1.17151296e+00 1.60400569e+00
5.84414959e-01 -1.32875657e+00 3.17144036e-01 -2.91193098e-01
7.31338561e-01 3.57778698e-01 -5.86989820e-02 4.56438899e-01
-1.00779340e-01 -4.12195057e-01 6.64841771e-01 6.40973032e-01
1.12051058e+00 -4.95385051e-01 6.95289314e-01 -5.76606654e-02
-1.55296445e+00 1.15786316e-02 -2.80241400e-01 5.69974661e-01
3.74643296e-01 3.73534918e-01 -1.03030376e-01 7.57708907e-01
1.00442708e+00 1.48865032e+00 -7.37008989e-01 1.27298391e+00
2.71195948e-01 3.51930708e-01 -1.02599179e-02 2.36511007e-01
-5.75013869e-02 -1.87299803e-01 5.03196955e-01 1.29732239e+00
8.18236709e-01 -8.57820734e-02 2.42914319e-01 2.60612220e-01
-3.23086560e-01 3.23960572e-01 -3.98893058e-01 -3.31387609e-01
8.08599234e-01 1.16865635e+00 -1.87408179e-01 -5.75517833e-01
-7.84084022e-01 1.12509370e+00 3.76258120e-02 1.91314697e-01
-1.11589956e+00 -2.98900634e-01 9.96194601e-01 -3.53840888e-02
2.52860993e-01 6.68967813e-02 5.38046122e-01 -1.72818053e+00
1.54182166e-01 -1.30775392e+00 7.86455393e-01 -1.00801957e+00
-1.58151782e+00 7.63422251e-01 2.70202737e-02 -1.78417826e+00
-4.86458123e-01 -1.95523039e-01 -4.26165819e-01 1.83180079e-01
-1.34206188e+00 -1.07116151e+00 -5.47085106e-01 1.02651918e+00
6.77878141e-01 -8.41019750e-01 6.84084773e-01 1.20977581e+00
-6.15758181e-01 1.20150316e+00 1.85624123e-01 2.58490115e-01
1.24485278e+00 -5.85285902e-01 1.15970172e-01 7.52796173e-01
1.15881294e-01 2.99628496e-01 3.44375223e-01 -4.86349016e-01
-1.52164829e+00 -1.23518765e+00 8.37756813e-01 -3.97974819e-01
8.51098001e-01 -1.63679067e-02 -8.13420534e-01 4.90284771e-01
1.33780390e-01 3.49366199e-03 1.02259290e+00 -1.58303201e-01
-5.27289212e-01 -2.54710108e-01 -7.69259393e-01 5.34111679e-01
1.46916378e+00 -9.39427316e-01 -3.87046844e-01 3.05849373e-01
9.39663708e-01 -2.40378454e-01 -1.34486890e+00 6.08813643e-01
5.75053036e-01 -8.21560204e-01 1.03025639e+00 -5.11275291e-01
6.68976367e-01 -2.73195982e-01 -7.24427342e-01 -1.06264973e+00
-2.84413248e-01 -3.75827253e-01 -2.23568082e-01 1.71357882e+00
2.47462288e-01 -4.35516946e-02 5.36824346e-01 4.45379078e-01
-4.03237849e-01 -5.04119098e-01 -7.51906872e-01 -8.29786479e-01
-3.22784185e-01 -4.29784179e-01 6.29531264e-01 1.22034001e+00
3.38742398e-02 2.23932803e-01 -6.50661945e-01 -1.97020173e-01
5.32016754e-01 2.23567277e-01 7.92881310e-01 -6.89787626e-01
-1.62228703e-01 -4.86769259e-01 -7.53912747e-01 -1.15753901e+00
3.13750803e-01 -1.06349504e+00 -1.67980462e-01 -1.17902386e+00
8.06509256e-01 -3.29158604e-01 -3.67039800e-01 3.52817833e-01
-2.30264440e-01 5.34476757e-01 2.59003997e-01 4.01258349e-01
-1.60097289e+00 8.15955043e-01 1.08698511e+00 -2.60266453e-01
3.23310167e-01 -1.90653056e-01 -4.25910592e-01 3.30161065e-01
7.46086955e-01 -2.88195580e-01 -4.17380303e-01 -6.86915278e-01
4.40127738e-02 5.44861071e-02 3.04047346e-01 -9.30335164e-01
4.46788281e-01 -1.94885537e-01 -1.06117517e-01 -7.31005251e-01
3.73008728e-01 -6.30366564e-01 2.74106234e-01 -7.05157667e-02
-6.27018988e-01 4.46516484e-01 -1.93639845e-01 2.80760467e-01
-6.74845576e-01 -9.74973887e-02 4.75978166e-01 1.02997366e-02
-1.26344752e+00 8.29671144e-01 -1.71232611e-01 4.93120790e-01
9.84647453e-01 -1.74930453e-01 -4.52156872e-01 -6.41479790e-01
-5.17891645e-01 5.88908017e-01 8.84884179e-01 9.54083502e-01
8.39428723e-01 -1.99325621e+00 -8.80693793e-01 -2.05561548e-01
6.78554296e-01 -5.81445396e-01 5.53890944e-01 8.68839562e-01
-2.47584969e-01 3.64025980e-01 -4.43070382e-01 -5.16044617e-01
-1.74478185e+00 6.87611818e-01 7.45593905e-02 2.77859479e-01
-2.20633283e-01 6.86367810e-01 -1.74598962e-01 -5.17587423e-01
4.14441824e-01 1.82968751e-01 -3.90954584e-01 3.76610719e-02
8.81953061e-01 4.78383064e-01 -1.22869924e-01 -1.05375028e+00
-5.35553217e-01 6.55625463e-01 1.22984648e-01 1.17290996e-01
1.16492796e+00 -5.99737883e-01 -2.57978499e-01 2.74973929e-01
1.65894246e+00 -2.44804904e-01 -6.19303226e-01 -5.54342330e-01
5.72489761e-03 -8.50696862e-01 -7.94620737e-02 -4.29187238e-01
-1.37927592e+00 5.94867706e-01 6.59667373e-01 -2.71085471e-01
1.23394299e+00 1.53045133e-01 1.21998525e+00 3.99006963e-01
3.55085015e-01 -1.15753186e+00 3.54088664e-01 6.58266306e-01
7.84876406e-01 -1.42778349e+00 -2.28276253e-01 -3.03589940e-01
-9.04877901e-01 6.64656699e-01 7.80407548e-01 3.93097222e-01
6.20502114e-01 -1.68536752e-02 6.07305020e-02 9.99651551e-02
-1.00629497e+00 -3.95465702e-01 6.18294239e-01 6.20722294e-01
5.76885223e-01 -2.12023169e-01 -1.75208375e-01 6.06857061e-01
2.35291012e-02 2.23927453e-01 2.25541800e-01 6.69356704e-01
-2.48861387e-01 -1.26542842e+00 3.96865755e-02 5.00431120e-01
-4.83078688e-01 -4.67606038e-01 -1.80390656e-01 1.19708672e-01
-1.25669930e-02 9.51770186e-01 1.45307764e-01 -1.03024709e+00
2.67845273e-01 -2.03746721e-01 3.18232358e-01 -4.09317464e-01
-3.33904773e-01 -3.19680981e-02 2.83140957e-01 -9.54167187e-01
-6.81853950e-01 -7.67031312e-01 -8.30865622e-01 -7.40237355e-01
-5.40154204e-02 2.63991863e-01 3.56602281e-01 6.63181007e-01
6.14491999e-01 1.42153516e-01 8.31256330e-01 -8.14439416e-01
-2.18756154e-01 -8.77942026e-01 -3.30544889e-01 8.99889052e-01
-8.85867402e-02 -5.21237552e-01 -2.16301963e-01 1.30662963e-01] | [10.275556564331055, 0.8553334474563599] |
32d73fe2-5272-48ae-9a4b-d32f3ca16c40 | llavar-enhanced-visual-instruction-tuning-for | 2306.17107 | null | https://arxiv.org/abs/2306.17107v1 | https://arxiv.org/pdf/2306.17107v1.pdf | LLaVAR: Enhanced Visual Instruction Tuning for Text-Rich Image Understanding | Instruction tuning unlocks the superior capability of Large Language Models (LLM) to interact with humans. Furthermore, recent instruction-following datasets include images as visual inputs, collecting responses for image-based instructions. However, visual instruction-tuned models cannot comprehend textual details within images well. This work enhances the current visual instruction tuning pipeline with text-rich images (e.g., movie posters, book covers, etc.). Specifically, we first use publicly available OCR tools to collect results on 422K text-rich images from the LAION dataset. Moreover, we prompt text-only GPT-4 with recognized texts and image captions to generate 16K conversations, each containing question-answer pairs for text-rich images. By combining our collected data with previous multi-modal instruction-following data, our model, LLaVAR, substantially improves the LLaVA model's capability on text-based VQA datasets (up to 20% accuracy improvement) while achieving an accuracy of 91.42% on ScienceQA. The GPT-4-based instruction-following evaluation also demonstrates the improvement of our model on both natural images and text-rich images. Through qualitative analysis, LLaVAR shows promising interaction (e.g., reasoning, writing, and elaboration) skills with humans based on the latest real-world online content that combines text and images. We make our code/data/models publicly available at https://llavar.github.io/. | ['Tong Sun', 'Diyi Yang', 'Nedim Lipka', 'Yufan Zhou', 'Jiuxiang Gu', 'Ruiyi Zhang', 'Yanzhe Zhang'] | 2023-06-29 | null | null | null | null | ['optical-character-recognition', 'image-captioning', 'instruction-following'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 1.56016633e-01 5.78262769e-02 -9.85671356e-02 -3.86358023e-01
-8.92197073e-01 -8.03550541e-01 7.67943621e-01 7.83317462e-02
-4.54492003e-01 1.31906092e-01 4.14368123e-01 -8.68709683e-01
2.92815268e-01 -6.98885858e-01 -1.08767247e+00 -2.01694146e-01
5.52882314e-01 5.75475574e-01 9.08765495e-02 -4.73175496e-01
5.75596392e-01 1.05738513e-01 -1.67686975e+00 1.11479080e+00
1.26816475e+00 5.79998970e-01 4.74782974e-01 1.03178978e+00
-4.37398642e-01 1.25427520e+00 -7.87070215e-01 -6.11082077e-01
2.19054352e-02 -3.64032060e-01 -8.95644963e-01 1.76913291e-02
1.23687291e+00 -9.41384673e-01 -4.12304610e-01 6.39984846e-01
3.29883069e-01 1.13305882e-01 6.10474229e-01 -1.22102499e+00
-1.58387434e+00 5.26292145e-01 -5.28286695e-01 2.76288558e-02
9.45736170e-01 9.09150064e-01 6.14339530e-01 -1.16080523e+00
7.34839320e-01 1.48662138e+00 8.51776823e-02 7.45595634e-01
-1.01400435e+00 -8.98405969e-01 1.43900197e-02 2.67065018e-01
-9.53138292e-01 -3.28790307e-01 5.43240309e-01 -5.16928613e-01
1.04557741e+00 4.46780413e-01 6.07963264e-01 1.53466082e+00
3.08377802e-01 1.31436419e+00 1.65258265e+00 -5.93294263e-01
-1.74264461e-01 3.74746948e-01 1.28866225e-01 1.11508739e+00
-7.44840577e-02 1.23121940e-01 -9.24711347e-01 6.34677351e-01
8.81718040e-01 -4.55282331e-02 -3.62993211e-01 -1.35899872e-01
-1.41333115e+00 7.32493937e-01 5.65598249e-01 2.62296423e-02
-3.31797414e-02 9.85593721e-02 1.32502794e-01 3.96931022e-01
2.59514213e-01 4.99997914e-01 -5.92283569e-02 -3.05216789e-01
-5.99478662e-01 6.66321144e-02 3.90898257e-01 1.22735322e+00
7.64174640e-01 7.15850145e-02 -6.83733284e-01 6.80993617e-01
4.09998864e-01 1.09587991e+00 4.11276102e-01 -9.75223660e-01
9.80892241e-01 9.05740917e-01 9.08430014e-03 -1.00155616e+00
-2.96046257e-01 -3.65157761e-02 -4.43625003e-01 3.17584604e-01
7.85458267e-01 1.20709620e-01 -1.24236631e+00 1.41859949e+00
1.99658483e-01 -4.25515622e-01 -1.66240167e-02 1.03452981e+00
1.67117226e+00 9.51264799e-01 4.70619112e-01 3.79884541e-01
1.81524897e+00 -1.44088817e+00 -9.18173730e-01 -6.10545158e-01
8.80639493e-01 -8.33689690e-01 2.25302744e+00 3.83246779e-01
-1.22096825e+00 -8.76178801e-01 -9.26497221e-01 -6.20625556e-01
-6.54073060e-01 3.52119058e-01 4.75820065e-01 6.10009193e-01
-1.26394260e+00 -2.82377332e-01 -2.74335921e-01 -3.43392253e-01
4.75459546e-01 -7.68535510e-02 -5.41022182e-01 -5.45532942e-01
-9.46227014e-01 8.57401490e-01 5.99356741e-03 -1.00519411e-01
-1.21991825e+00 -9.04158771e-01 -1.16528666e+00 -2.15059310e-01
4.56973523e-01 -6.58633351e-01 1.52802837e+00 -1.03667164e+00
-1.47702467e+00 1.24699259e+00 -1.63883254e-01 -5.23026064e-02
5.45335352e-01 -3.05690229e-01 -1.67460471e-01 6.02840960e-01
-2.77832299e-02 1.21731997e+00 7.09385395e-01 -1.54113400e+00
-1.56882912e-01 -3.06597710e-01 4.71545488e-01 4.47753519e-01
-3.84610981e-01 -7.72909671e-02 -6.00441992e-01 -4.82816666e-01
-5.13154745e-01 -6.94813788e-01 2.09202901e-01 -1.50245735e-02
-2.48851731e-01 -1.64508864e-01 8.01005304e-01 -9.21579957e-01
1.07112443e+00 -1.95866418e+00 -1.46177739e-01 -1.87387004e-01
4.70739007e-01 3.06006700e-01 -6.06367409e-01 4.67598468e-01
2.00952470e-01 1.45697743e-01 1.23678505e-01 -3.79771262e-01
2.22777575e-01 6.36056960e-02 -3.04646939e-01 6.88736066e-02
2.35928595e-01 1.83701754e+00 -8.71838331e-01 -8.71885240e-01
6.11753166e-01 4.41456228e-01 -4.90715176e-01 4.56204653e-01
-3.98190051e-01 4.14927244e-01 -2.12976784e-01 6.40420616e-01
5.52671850e-01 -4.81060833e-01 -2.02636167e-01 -1.72755331e-01
-9.56060886e-02 5.20746447e-02 -4.33788061e-01 1.78986442e+00
-7.64987290e-01 1.04929090e+00 -4.74053714e-03 -3.36539954e-01
7.69487560e-01 1.01576917e-01 -1.67205989e-01 -1.38413739e+00
1.32689029e-01 -1.61967695e-01 -3.35954815e-01 -9.57773387e-01
6.91187978e-01 4.28250313e-01 -1.53402373e-01 6.11841679e-01
1.05467409e-01 -5.83058536e-01 3.04658830e-01 7.63075292e-01
7.23396003e-01 2.53027588e-01 -1.09927118e-01 -8.89124274e-02
3.90804231e-01 3.96416068e-01 -3.95222127e-01 9.65751469e-01
-4.34711695e-01 5.55742323e-01 2.73888290e-01 -4.26301152e-01
-9.64234591e-01 -9.94190753e-01 1.52093157e-01 1.65359533e+00
5.59864901e-02 -4.38688010e-01 -9.84999895e-01 -8.20446551e-01
-1.37901455e-01 9.95949864e-01 -8.01652610e-01 4.38927114e-02
-4.86815721e-01 -2.72401124e-02 3.98166865e-01 4.63571906e-01
7.21709073e-01 -1.49914873e+00 -7.25573897e-01 -3.37194651e-01
-5.95202744e-01 -1.27025437e+00 -6.69023097e-01 -2.59382427e-01
-5.13884842e-01 -8.92578900e-01 -7.24836349e-01 -9.95095611e-01
8.63410950e-01 5.35940409e-01 1.65351045e+00 3.97475094e-01
-2.41453990e-01 1.08690786e+00 -4.03066814e-01 -4.22213852e-01
-4.72080022e-01 -2.77949691e-01 -4.83082473e-01 -4.21873719e-01
5.82921267e-01 1.98576152e-01 -7.40380347e-01 4.61839885e-01
-1.03501213e+00 9.14886355e-01 7.62808144e-01 6.43517733e-01
2.60277003e-01 -7.29311645e-01 2.57321566e-01 -8.76152933e-01
7.55570531e-01 -1.39113724e-01 -4.06525254e-01 6.51219606e-01
-3.18646371e-01 -1.39997855e-01 3.82823795e-01 -6.58118904e-01
-1.32421482e+00 -1.65172651e-01 7.96979219e-02 -3.11426938e-01
-5.01242697e-01 1.25043690e-01 1.67487830e-01 -1.94387212e-01
8.99457335e-01 2.96323240e-01 -4.84798215e-02 9.69381705e-02
7.62179673e-01 8.71979773e-01 8.05798829e-01 -7.40821660e-01
7.95128465e-01 2.71394700e-01 -6.54154420e-01 -7.29621410e-01
-7.17503011e-01 -2.98938841e-01 -4.56279099e-01 -8.20597231e-01
1.09173191e+00 -1.06242657e+00 -1.32062066e+00 3.78378689e-01
-1.02389205e+00 -9.80219245e-01 -3.79222073e-02 2.38830715e-01
-5.65828383e-01 1.51992097e-01 -8.88283670e-01 -6.73997641e-01
-4.83221084e-01 -1.53254759e+00 1.34604776e+00 4.76585150e-01
-2.43054137e-01 -8.52727890e-01 -1.83950424e-01 1.43954134e+00
3.52668464e-01 -1.45499095e-01 8.38293910e-01 -3.10068399e-01
-8.04903686e-01 1.16127215e-01 -5.60178220e-01 -8.55926573e-02
-2.64595151e-01 1.62455365e-01 -1.02991509e+00 -9.85953808e-02
-2.88295090e-01 -1.05860460e+00 6.72515750e-01 1.95725337e-01
1.24207592e+00 -2.24347219e-01 9.06988513e-03 3.31178218e-01
1.07159221e+00 1.05789967e-01 6.11843944e-01 2.63323158e-01
1.04052496e+00 9.27191019e-01 9.39475894e-01 4.27228585e-02
8.79202724e-01 4.80634958e-01 4.23165470e-01 -2.76041746e-01
-5.52823901e-01 -6.73758805e-01 7.22020090e-01 1.06213832e+00
2.52933055e-01 -4.19931203e-01 -1.36186922e+00 6.05409920e-01
-1.49688542e+00 -5.77576101e-01 -4.15404320e-01 1.83688867e+00
1.09193039e+00 -1.81253061e-01 -1.76657245e-01 -4.45976555e-01
1.80579677e-01 7.00668991e-02 -4.18700784e-01 -7.28471458e-01
-1.56777367e-01 3.94741297e-02 2.56897926e-01 6.49425447e-01
-8.46862614e-01 1.22190559e+00 5.90861273e+00 8.63376379e-01
-8.16103578e-01 -1.74871292e-02 7.66266942e-01 -1.14897892e-01
-7.88733661e-01 -4.37449068e-01 -6.31831586e-01 2.62154311e-01
9.69712913e-01 9.25429985e-02 4.72160578e-01 4.22013879e-01
2.20488682e-01 -4.04475123e-01 -9.39915657e-01 1.12285471e+00
5.25283396e-01 -1.34187186e+00 5.15260279e-01 -1.96556270e-01
8.24557543e-01 -3.26230347e-01 7.34807670e-01 8.22538793e-01
5.76103091e-01 -1.38622189e+00 8.31214190e-01 2.83725798e-01
1.12747192e+00 -4.79904085e-01 4.66907501e-01 2.33251601e-01
-9.68557358e-01 -4.82415082e-03 -3.06523629e-02 -9.57355350e-02
6.29469529e-02 -7.36266300e-02 -1.02498853e+00 2.21946672e-01
9.33067441e-01 5.02588809e-01 -1.26411343e+00 4.41041142e-01
-6.29374683e-01 5.12035191e-01 2.07089409e-01 -2.72319943e-01
3.19017619e-01 -2.33156726e-01 -6.67718947e-02 1.31958163e+00
-8.16184375e-03 3.65213662e-01 1.60940602e-01 8.40332031e-01
-1.73094541e-01 3.98901403e-01 -8.58332932e-01 -1.48768008e-01
3.85392517e-01 1.29943979e+00 -5.13514161e-01 -6.37905240e-01
-7.37692177e-01 8.58903050e-01 5.37959933e-01 7.14650810e-01
-8.69813621e-01 -2.54443258e-01 2.71369636e-01 1.07153401e-01
-3.56479794e-01 -3.36674839e-01 -4.78321135e-01 -1.00024486e+00
-2.12205440e-01 -1.58454216e+00 3.45762044e-01 -1.71713161e+00
-9.88297224e-01 4.02935535e-01 8.35041925e-02 -9.60859001e-01
-6.03537485e-02 -1.11226857e+00 -3.83193761e-01 5.80758810e-01
-1.26276135e+00 -1.56833220e+00 -8.50611031e-01 8.51810217e-01
1.04388130e+00 1.62474453e-01 5.51285982e-01 1.27434745e-01
-4.90164161e-01 7.65995085e-01 -3.78471524e-01 2.62710869e-01
1.36401641e+00 -1.47569835e+00 4.83522892e-01 6.21172249e-01
2.76534826e-01 6.42128885e-01 6.03431880e-01 -8.41752291e-01
-1.76389813e+00 -7.11178243e-01 4.25852597e-01 -1.19315410e+00
6.22692525e-01 -6.67283893e-01 -9.20818985e-01 9.11744177e-01
9.21275556e-01 -4.42116886e-01 5.96845627e-01 -1.99824050e-01
-5.11042535e-01 2.82457530e-01 -9.34431970e-01 1.14968014e+00
9.66139972e-01 -7.89137781e-01 -7.43502378e-01 5.14947474e-01
1.04331851e+00 -7.36543417e-01 -7.14369655e-01 1.51225522e-01
6.13478541e-01 -8.58015001e-01 1.15501988e+00 -4.27234411e-01
9.67979848e-01 -1.81183606e-01 1.37406364e-01 -1.25907052e+00
-9.82372761e-02 -3.31316501e-01 -5.50879054e-02 1.07233953e+00
4.27010477e-01 -3.49790901e-01 5.41196942e-01 6.93424284e-01
-2.50262707e-01 -4.32958931e-01 -3.06568354e-01 -2.19968021e-01
1.68002158e-01 -4.63403970e-01 4.09837246e-01 8.72868896e-01
1.20985813e-01 3.04342985e-01 -1.97559282e-01 -7.48140812e-02
4.38138247e-01 1.59926876e-01 1.26168191e+00 -6.34742618e-01
1.71813648e-02 -4.64730263e-01 2.65913457e-01 -1.29890025e+00
1.16995901e-01 -8.28088343e-01 5.64174466e-02 -1.73521566e+00
5.04460096e-01 3.04782894e-02 2.90727854e-01 6.95366740e-01
-5.36766469e-01 4.89259839e-01 4.25594091e-01 -5.75042190e-03
-8.72301757e-01 4.26658392e-01 2.06504488e+00 -3.84659708e-01
-8.43434874e-03 -8.16061139e-01 -5.17832100e-01 4.55157340e-01
8.48996341e-01 2.14003474e-01 -5.05790949e-01 -9.66904879e-01
1.72013745e-01 -6.50796369e-02 5.28541446e-01 -8.00827801e-01
1.99560925e-01 -2.93114841e-01 7.33802438e-01 -7.28680491e-01
2.48795450e-01 -6.46478832e-01 -2.97646761e-01 4.30455089e-01
-5.70128322e-01 2.96468616e-01 6.52922034e-01 2.24909574e-01
-1.43924475e-01 1.70820914e-02 3.46599221e-01 -1.55771151e-01
-1.05255330e+00 -6.71321526e-02 -5.89766026e-01 1.22595996e-01
9.87822890e-01 -2.17829794e-01 -1.32434857e+00 -8.43863726e-01
-1.63323492e-01 6.26251936e-01 6.90108359e-01 8.17426205e-01
9.31170702e-01 -1.26872039e+00 -7.41613090e-01 2.26845190e-01
5.47223330e-01 -1.59949020e-01 7.46498287e-01 6.63840771e-01
-9.46309149e-01 6.98403060e-01 -4.78843004e-01 -8.37654114e-01
-1.48976922e+00 7.28599131e-01 -8.63420069e-02 -5.10055050e-02
-4.43213165e-01 8.98203969e-01 7.67808676e-01 -7.35771000e-01
1.83440387e-01 -5.47395051e-01 -3.93296331e-01 -9.51772630e-02
8.16959202e-01 1.07630052e-01 -2.69379526e-01 -5.50953567e-01
-4.04811390e-02 7.63833165e-01 -6.34247139e-02 -3.36613059e-01
7.03447759e-01 -2.81923801e-01 1.96729004e-01 4.44873661e-01
9.53777730e-01 1.24725766e-01 -1.28001010e+00 -1.84277982e-01
-4.82118398e-01 -5.96786916e-01 -1.77619681e-01 -1.23262751e+00
-8.60069156e-01 1.34061182e+00 4.98785436e-01 -2.46488988e-01
1.04959726e+00 7.13146701e-02 4.76030082e-01 4.07858521e-01
2.37244353e-01 -8.86570930e-01 8.06091070e-01 4.22161490e-01
1.16890490e+00 -1.80419469e+00 -2.15401545e-01 -2.53637582e-01
-1.03073740e+00 8.65642965e-01 1.37170208e+00 4.71141458e-01
-2.62755752e-01 8.82057995e-02 6.70083761e-01 -4.34085608e-01
-9.27218378e-01 -7.03436434e-02 6.65311813e-01 6.87896550e-01
6.37366474e-01 2.69924015e-01 8.65761712e-02 3.80669743e-01
-5.12893975e-01 -2.25867376e-01 5.21091223e-01 7.79845834e-01
-4.09037113e-01 -5.94076574e-01 -7.28357911e-01 3.53864312e-01
-1.11722096e-03 -4.03280526e-01 -5.89452326e-01 9.29334581e-01
-3.74985904e-01 1.39804959e+00 1.46287769e-01 -2.77706861e-01
4.85669523e-01 1.12794235e-01 5.99131227e-01 -5.19998729e-01
-9.69588935e-01 -4.55101222e-01 2.03766748e-02 -9.22520518e-01
-4.67892438e-02 -1.57032162e-01 -1.11053371e+00 -5.89579523e-01
1.97450966e-01 -2.02709466e-01 6.78226233e-01 6.10687435e-01
2.64690757e-01 6.82743132e-01 -5.89309931e-02 -7.09593415e-01
-8.76122639e-02 -1.08558857e+00 2.25414455e-01 7.38421321e-01
2.11985871e-01 -4.30973262e-01 -2.03328297e-01 1.36889905e-01] | [10.959626197814941, 1.5917023420333862] |
3f83bb1f-062d-41c3-aa6e-433f72d57d9e | training-entire-space-models-for-target | 2204.07337 | null | https://arxiv.org/abs/2204.07337v1 | https://arxiv.org/pdf/2204.07337v1.pdf | Training Entire-Space Models for Target-oriented Opinion Words Extraction | Target-oriented opinion words extraction (TOWE) is a subtask of aspect-based sentiment analysis (ABSA). Given a sentence and an aspect term occurring in the sentence, TOWE extracts the corresponding opinion words for the aspect term. TOWE has two types of instance. In the first type, aspect terms are associated with at least one opinion word, while in the second type, aspect terms do not have corresponding opinion words. However, previous researches trained and evaluated their models with only the first type of instance, resulting in a sample selection bias problem. Specifically, TOWE models were trained with only the first type of instance, while these models would be utilized to make inference on the entire space with both the first type of instance and the second type of instance. Thus, the generalization performance will be hurt. Moreover, the performance of these models on the first type of instance cannot reflect their performance on entire space. To validate the sample selection bias problem, four popular TOWE datasets containing only aspect terms associated with at least one opinion word are extended and additionally include aspect terms without corresponding opinion words. Experimental results on these datasets show that training TOWE models on entire space will significantly improve model performance and evaluating TOWE models only on the first type of instance will overestimate model performance. | ['Sheng-hua Zhong', 'Fang Wang', 'Yuncong Li'] | 2022-04-15 | null | null | null | null | ['aspect-based-sentiment-analysis', 'target-oriented-opinion-words-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.48250470e-01 -4.62080789e-04 -1.07642718e-01 -4.73848999e-01
-3.97078514e-01 -4.88954693e-01 4.27711517e-01 4.03855085e-01
-4.31585252e-01 5.78991771e-01 -1.32972360e-01 -3.74328941e-01
7.56157795e-03 -1.15232599e+00 -5.05569220e-01 -7.31797218e-01
1.73064217e-01 1.99069843e-01 2.01493427e-01 -4.98924285e-01
4.32841629e-01 -1.58581913e-01 -1.51306689e+00 6.96116924e-01
8.07995975e-01 1.20335209e+00 -5.78109100e-02 1.49901420e-01
-7.88475633e-01 4.26070124e-01 -1.12213552e+00 -5.53373933e-01
2.54063338e-01 -3.88267487e-01 -4.38423932e-01 1.89815030e-01
8.36368874e-02 1.77536771e-01 3.83827657e-01 1.08015966e+00
3.01583767e-01 -1.28565297e-01 6.75871253e-01 -1.30310905e+00
-5.45930386e-01 6.05356514e-01 -5.52886486e-01 9.26145539e-02
2.53412575e-01 -2.07252875e-01 1.28137696e+00 -1.26229382e+00
4.85600233e-01 9.23761785e-01 6.22779131e-01 2.36290038e-01
-5.77742934e-01 -5.59573174e-01 7.42576003e-01 1.01380929e-01
-1.16409385e+00 8.25553834e-02 9.12932754e-01 -2.12773576e-01
1.03048170e+00 3.60003769e-01 9.65611517e-01 8.10913682e-01
3.98854285e-01 1.08578420e+00 1.40838945e+00 -2.28650585e-01
4.26633626e-01 5.86405993e-01 6.84124708e-01 2.01588511e-01
5.84175289e-01 -1.95956051e-01 -5.25068939e-01 -1.05370302e-02
-8.68463963e-02 -1.66249767e-01 -1.34612098e-01 5.63617423e-02
-8.84337902e-01 9.18369591e-01 1.95626348e-01 5.43961167e-01
-4.43836719e-01 -6.70947552e-01 4.25374538e-01 4.18043405e-01
5.87503493e-01 6.50397062e-01 -9.21302080e-01 -1.89976722e-01
-7.56243408e-01 3.52238566e-01 9.74889398e-01 1.08232939e+00
8.27777863e-01 -5.49000083e-03 -1.84127286e-01 7.18650162e-01
4.92259627e-03 6.88895702e-01 7.10542381e-01 1.71004727e-01
5.81150651e-01 1.19620132e+00 -7.73364231e-02 -1.27021158e+00
-5.21530330e-01 -7.39417672e-01 -5.74839473e-01 -7.60979950e-02
3.19429994e-01 -2.21460074e-01 -1.00011873e+00 1.39772654e+00
4.23051447e-01 -3.21915656e-01 3.61023605e-01 7.67589748e-01
9.71991777e-01 6.92049444e-01 -1.94644094e-01 -3.46956223e-01
1.76716280e+00 -8.64465535e-01 -9.05470014e-01 -7.42590427e-01
7.05006957e-01 -1.02352989e+00 1.07645106e+00 4.58301872e-01
-5.99469066e-01 -2.99547374e-01 -1.31953073e+00 4.98046398e-01
-9.89546716e-01 1.09388635e-01 7.13321686e-01 5.80187857e-01
-4.84713942e-01 2.50411212e-01 -1.98806494e-01 -1.65719122e-01
2.96953358e-02 -4.72666286e-02 -2.45229676e-01 -8.33552033e-02
-1.55118537e+00 8.48859191e-01 2.72200674e-01 2.52824455e-01
-1.53139815e-01 -5.47056913e-01 -8.09849560e-01 -6.49045631e-02
5.94343424e-01 -3.05178106e-01 9.06327426e-01 -1.12050092e+00
-6.80251539e-01 5.12772977e-01 -6.62875116e-01 -1.90390810e-01
1.28956690e-01 -2.03109384e-01 -8.67273569e-01 -2.25982994e-01
2.73675621e-01 -7.50058219e-02 7.76063859e-01 -1.22250378e+00
-7.28346646e-01 -5.15314102e-01 4.16531920e-01 1.32556319e-01
-6.16523147e-01 -1.97452933e-01 -2.80164242e-01 -5.49402356e-01
2.84310460e-01 -8.50164175e-01 -2.78672297e-02 -5.20408750e-01
-4.79621261e-01 -3.32386643e-01 9.25284863e-01 -3.22580725e-01
1.70942318e+00 -2.00392008e+00 -2.84108400e-01 3.12850177e-01
-2.86868960e-02 3.28754812e-01 -1.33446187e-01 3.98760110e-01
-2.24081337e-01 3.11113417e-01 -2.39868850e-01 5.72161637e-02
-2.35392854e-01 2.66435742e-01 -3.34372133e-01 -1.35117725e-01
4.04684573e-01 6.10906184e-01 -8.15255225e-01 -3.45925748e-01
-2.85182208e-01 5.68571351e-02 -4.08443511e-01 5.71471043e-02
-1.40937269e-01 -1.06455341e-01 -5.86341858e-01 6.78412199e-01
7.32891560e-01 -1.94974011e-03 1.12161702e-02 -6.37731075e-01
-1.18606322e-01 4.44363564e-01 -1.30714631e+00 8.98059011e-01
-9.18221176e-01 5.19161701e-01 -3.48106533e-01 -8.53404522e-01
1.03902972e+00 2.75091201e-01 4.37826335e-01 -6.32755697e-01
5.35663683e-03 3.58811021e-01 2.74869919e-01 -7.46577442e-01
7.05719590e-01 -4.21239048e-01 -1.59655228e-01 4.18506891e-01
-1.71679854e-01 -9.83301103e-02 5.79042494e-01 7.55185261e-02
9.56012368e-01 -1.97822317e-01 3.95987749e-01 -2.82925963e-01
5.00583827e-01 3.91835533e-02 7.67341793e-01 5.09677887e-01
-1.04613543e-01 5.66606760e-01 7.71864295e-01 -5.70041239e-01
-6.93742156e-01 -6.22128785e-01 -2.61094838e-01 7.45595396e-01
2.66157061e-01 -8.36641967e-01 -4.14110392e-01 -1.09676564e+00
-3.79615188e-01 1.05427122e+00 -7.65423477e-01 -1.92164421e-01
-2.92541087e-01 -1.13654494e+00 7.69971386e-02 3.61510873e-01
5.42778850e-01 -1.00125504e+00 -6.79366112e-01 9.26006138e-02
-3.64749670e-01 -1.00393081e+00 -4.83423769e-01 2.15852007e-01
-6.17845118e-01 -1.30185628e+00 -3.82798940e-01 -5.68058789e-01
9.91359174e-01 2.90300757e-01 1.32317555e+00 5.22189513e-02
1.35174900e-01 -1.04949467e-01 -8.34415376e-01 -1.01738429e+00
-9.15402696e-02 1.41141785e-03 -1.42031297e-01 2.13496044e-01
1.09817564e+00 -2.14727402e-01 -5.66335738e-01 2.88590044e-01
-1.03268003e+00 -1.89139232e-01 6.63364112e-01 8.69324207e-01
5.17924905e-01 4.50038970e-01 6.47016227e-01 -1.18025434e+00
7.87724495e-01 -5.29162169e-01 -3.88773948e-01 2.58250594e-01
-9.59481299e-01 -8.77164528e-02 6.22349560e-01 -6.56546414e-01
-7.21076727e-01 -5.21721005e-01 -1.29770830e-01 2.38006227e-02
1.65881544e-01 1.13407481e+00 -3.65326703e-01 7.48688519e-01
4.16751146e-01 4.54585344e-01 -2.52017230e-01 -2.33930945e-01
-3.01375628e-01 7.74459302e-01 -4.06798840e-01 -4.21147496e-01
5.50297499e-01 2.06635550e-01 -6.21725135e-02 -9.68966305e-01
-1.26769352e+00 -5.43244123e-01 -2.20711797e-01 -1.82159781e-01
5.44783533e-01 -6.76528990e-01 -8.25401396e-02 4.70804036e-01
-1.07097185e+00 5.08134604e-01 -2.18923137e-01 4.68388230e-01
1.15996212e-01 1.77616462e-01 -4.72406205e-03 -1.05690086e+00
-5.07242262e-01 -1.24438393e+00 1.14446783e+00 2.49408245e-01
-4.89352524e-01 -9.42403853e-01 -2.61435568e-01 8.57210904e-02
3.85266840e-01 -9.49920807e-03 1.12022805e+00 -1.24174833e+00
-5.94566986e-02 -7.46101797e-01 2.41510451e-01 4.69912529e-01
3.12266260e-01 1.16112337e-01 -7.76469052e-01 -2.07414582e-01
4.18725222e-01 8.28683898e-02 4.73393381e-01 8.09583440e-02
7.88310826e-01 -3.79454523e-01 -4.64396644e-03 3.48005556e-02
1.40094507e+00 5.36533535e-01 6.08945906e-01 4.94787514e-01
5.26184857e-01 7.85060525e-01 1.13677585e+00 1.79891929e-01
2.68083781e-01 7.36176431e-01 3.06595653e-01 7.59358481e-02
2.19458416e-01 -1.05024546e-01 5.87188125e-01 1.10001981e+00
1.03481255e-01 -2.81479865e-01 -5.47758460e-01 8.31986248e-01
-1.46374333e+00 -6.92679346e-01 -4.97598350e-01 2.30732012e+00
7.21282363e-01 4.60152894e-01 -1.09943636e-01 5.44787526e-01
5.69029152e-01 3.29355299e-01 -5.51438332e-01 -7.14538276e-01
-2.99138278e-01 1.62653148e-01 9.69438404e-02 2.20991284e-01
-9.67743993e-01 6.91487849e-01 4.75267982e+00 9.29796815e-01
-1.27099895e+00 -5.46230972e-02 4.15137500e-01 -3.73707116e-02
-6.58340871e-01 2.28864014e-01 -9.57684398e-01 5.81596076e-01
6.63657904e-01 -2.11186364e-01 -1.94305390e-01 1.03626907e+00
-3.84951346e-02 -2.27512047e-01 -8.81991684e-01 5.59796512e-01
2.46162906e-01 -6.00633681e-01 3.72922063e-01 1.10766627e-01
6.79962635e-01 -4.60545808e-01 -6.67019635e-02 4.92335886e-01
-3.86774659e-01 -8.72308493e-01 6.31885231e-01 1.79202214e-01
3.26183498e-01 -9.71401513e-01 1.50699568e+00 3.55933368e-01
-1.35207534e+00 5.45154363e-02 -5.23137093e-01 -1.89195931e-01
1.69156477e-01 1.23846412e+00 -4.72170353e-01 8.00366342e-01
7.22914815e-01 6.75614953e-01 -6.29005075e-01 8.26330900e-01
-4.59392965e-01 6.86112463e-01 -1.94390714e-01 -4.88402933e-01
3.23149353e-01 -4.04782593e-01 8.29817712e-01 9.11217332e-01
5.14209509e-01 -1.58846647e-01 -3.49626578e-02 5.51359475e-01
2.84861714e-01 5.04094064e-01 -8.19476604e-01 -1.94891021e-01
3.42135549e-01 1.41020131e+00 -6.67125940e-01 -4.77948219e-01
-7.65899539e-01 5.76438487e-01 -4.65504862e-02 3.98335159e-01
-7.13708282e-01 -6.57440782e-01 6.73883200e-01 1.17684461e-01
5.22591949e-01 2.95058846e-01 -6.69364870e-01 -1.32937324e+00
5.36963344e-01 -1.01377296e+00 1.99773520e-01 -5.32951772e-01
-1.39446175e+00 9.41223860e-01 -1.47369236e-01 -1.65571821e+00
-1.08394444e-01 -7.14213014e-01 -7.60501921e-01 8.87511432e-01
-1.38120306e+00 -9.83171821e-01 -1.02601863e-01 2.48796806e-01
8.46346319e-01 -1.49481937e-01 9.57485378e-01 1.56822145e-01
-5.24573565e-01 8.12608659e-01 -4.49178398e-01 1.91440746e-01
6.72568440e-01 -1.25507295e+00 8.89329091e-02 9.04827118e-01
1.28955156e-01 1.09546769e+00 9.85354245e-01 -7.03930259e-01
-1.24955785e+00 -9.68898475e-01 1.31698906e+00 -4.96276557e-01
6.08653247e-01 -2.81329811e-01 -8.57582390e-01 4.09688383e-01
6.52564839e-02 -3.83824587e-01 8.79543126e-01 4.89400238e-01
-3.89613390e-01 -1.32548392e-01 -9.66637969e-01 7.73478448e-01
4.93627042e-01 -4.18270290e-01 -9.06000853e-01 1.41665876e-01
4.70121920e-01 -2.18136519e-01 -7.59077668e-01 8.40970993e-01
6.06696427e-01 -8.91358852e-01 5.83386183e-01 -6.09037101e-01
9.15699422e-01 -5.25477946e-01 -1.77638590e-01 -1.53857934e+00
3.00396681e-01 2.07172096e-01 -2.57233381e-01 1.42585635e+00
1.17720592e+00 -7.30611265e-01 5.00500619e-01 6.25095546e-01
-8.27057287e-03 -1.43801486e+00 -7.11731553e-01 -8.01066458e-01
1.13086127e-01 -5.77630520e-01 9.46601570e-01 7.72317290e-01
4.03163955e-02 5.95963299e-01 -2.11277664e-01 5.86966202e-02
1.11458503e-01 5.21336555e-01 5.62274218e-01 -9.16731358e-01
-2.23191991e-01 -2.95392334e-01 -3.00647557e-01 -8.31161201e-01
-1.93668082e-01 -5.40928662e-01 3.90295498e-02 -1.54397357e+00
2.08203495e-01 -3.39818239e-01 -3.09081316e-01 3.55052203e-01
-7.85244524e-01 1.20160744e-01 6.13769591e-02 2.23960113e-02
-1.73811644e-01 4.54449654e-01 1.30510032e+00 -3.63354087e-01
-1.04285359e-01 4.95001674e-01 -9.13287759e-01 8.77046466e-01
6.75379515e-01 -4.76236582e-01 -6.67991638e-01 -3.19813862e-02
8.01960468e-01 -3.78385931e-01 -1.48471609e-01 -5.93673944e-01
-3.80051089e-03 -9.22844335e-02 1.81046844e-01 -7.54998565e-01
2.24776834e-01 -1.02038801e+00 -2.16323912e-01 3.42645317e-01
-1.28722545e-02 3.57342571e-01 1.67672455e-01 3.68249625e-01
-6.60220563e-01 -5.19967973e-01 3.29310656e-01 -6.97669089e-02
-7.62456357e-01 1.67467259e-02 -4.50385213e-01 1.04028970e-01
9.25758600e-01 -4.79219168e-01 -3.57247680e-01 -3.86023968e-01
-4.10444766e-01 3.58677991e-02 2.76401222e-01 6.47265911e-01
6.41546965e-01 -1.14079118e+00 -5.42624772e-01 3.50472242e-01
6.03010654e-01 2.32427329e-01 2.33160749e-01 7.85268128e-01
1.02391124e-01 1.73143998e-01 3.19696128e-01 -2.38835692e-01
-1.15211749e+00 6.36583924e-01 1.99214920e-01 -7.27996409e-01
-1.72722250e-01 7.07947493e-01 4.18479055e-01 -5.18945038e-01
-2.87151814e-01 -2.60906219e-01 -8.66455078e-01 6.56290889e-01
6.47731483e-01 -8.85777548e-02 3.55232447e-01 -7.88303196e-01
-3.74665469e-01 6.86912239e-01 -3.36703151e-01 9.30496976e-02
9.70096409e-01 6.84974268e-02 -3.12866926e-01 8.19650114e-01
1.26473486e+00 3.92198414e-01 -4.68729407e-01 1.14617124e-03
1.33008987e-01 -3.27586025e-01 -9.07415003e-02 -9.43042636e-01
-1.19411218e+00 7.09793150e-01 2.69469798e-01 4.26114202e-01
1.19205546e+00 -3.19856584e-01 7.75626659e-01 1.99729532e-01
5.29538751e-01 -1.22043705e+00 -1.41339585e-01 6.92755461e-01
7.29651570e-01 -1.31284869e+00 6.93943128e-02 -4.94569391e-01
-8.67473662e-01 1.04747558e+00 8.71498644e-01 2.63068348e-01
8.17645371e-01 2.61320800e-01 3.21306199e-01 -5.41914999e-01
-7.53963470e-01 -1.14476278e-01 4.15010005e-01 3.69994819e-01
6.40488327e-01 4.72561903e-02 -9.52179551e-01 1.08352935e+00
-4.82516468e-01 -2.87181765e-01 4.70053583e-01 1.10472834e+00
-1.81379706e-01 -1.13144410e+00 -1.38399720e-01 8.59559178e-01
-5.96728683e-01 -4.63056028e-01 -4.63195890e-01 8.13385844e-01
9.99668241e-02 1.33367753e+00 -6.01998381e-02 -6.50164962e-01
7.61129856e-01 3.35323930e-01 -1.38897374e-01 -6.72640264e-01
-6.37411952e-01 -1.10705629e-01 3.86059642e-01 -3.93092841e-01
-3.81060779e-01 -4.50607777e-01 -1.09049642e+00 -3.56357619e-02
-6.00276887e-01 5.29972196e-01 8.42106044e-01 1.29674971e+00
2.06408039e-01 7.56618321e-01 6.61787212e-01 -1.53929502e-01
-4.82035846e-01 -1.21350288e+00 -6.56854689e-01 4.59933162e-01
7.09384903e-02 -5.41618824e-01 -6.93666101e-01 -3.20233107e-01] | [11.464884757995605, 6.64493989944458] |
06a569c3-06e2-4594-bf4e-7e32d947f4ff | vision-datasets-a-benchmark-for-vision-based | 2306.07890 | null | https://arxiv.org/abs/2306.07890v2 | https://arxiv.org/pdf/2306.07890v2.pdf | VISION Datasets: A Benchmark for Vision-based InduStrial InspectiON | Despite progress in vision-based inspection algorithms, real-world industrial challenges -- specifically in data availability, quality, and complex production requirements -- often remain under-addressed. We introduce the VISION Datasets, a diverse collection of 14 industrial inspection datasets, uniquely poised to meet these challenges. Unlike previous datasets, VISION brings versatility to defect detection, offering annotation masks across all splits and catering to various detection methodologies. Our datasets also feature instance-segmentation annotation, enabling precise defect identification. With a total of 18k images encompassing 44 defect types, VISION strives to mirror a wide range of real-world production scenarios. By supporting two ongoing challenge competitions on the VISION Datasets, we hope to foster further advancements in vision-based industrial inspection. | ['Meng Cao', 'Jianjun Shi', 'Jiulong Shan', 'Ping Huang', 'Oncel Tuzel', 'Ramazan Gokberk Cinbis', 'Tatiana Likhomanenko', 'Shancong Mou', 'Haoping Bai'] | 2023-06-13 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [ 4.38408047e-01 1.37103964e-02 3.06986243e-01 -3.02851826e-01
-7.19081998e-01 -6.75154209e-01 1.36502534e-01 3.27628851e-01
3.88075024e-01 9.85855460e-02 -3.80827904e-01 -2.07607150e-01
-2.92666912e-01 -6.19593263e-01 -4.61972713e-01 -2.49465793e-01
-1.49160445e-01 4.43189144e-01 2.25789368e-01 -3.55476677e-01
4.63666022e-01 3.03967446e-01 -1.66418397e+00 4.09228891e-01
7.52842903e-01 1.26983035e+00 3.49316090e-01 8.21663797e-01
3.02422404e-01 5.39683580e-01 -9.72437322e-01 -3.77461851e-01
5.02486587e-01 -1.47454992e-01 -9.49731946e-01 7.20451355e-01
7.36700535e-01 -1.70465574e-01 -1.38594463e-01 1.02791715e+00
2.67697275e-01 -2.23265007e-01 4.58268166e-01 -1.23440671e+00
-1.19912195e+00 1.79050416e-01 -6.68910444e-01 8.43218863e-02
5.23890436e-01 7.71425843e-01 1.23883057e+00 -7.33139336e-01
6.39485240e-01 1.06122899e+00 1.08022654e+00 1.05487235e-01
-1.29259264e+00 1.25692999e-02 1.03688322e-01 6.07565455e-02
-6.94361269e-01 -2.71913022e-01 7.10690498e-01 -6.47187829e-01
1.22127366e+00 9.44707617e-02 7.52938986e-01 7.72998810e-01
3.08935732e-01 8.46120894e-01 8.65972400e-01 -2.92957723e-01
2.54450664e-02 -3.13640505e-01 1.70522571e-01 9.52812552e-01
4.98827308e-01 3.83173674e-02 -5.37021816e-01 1.03174835e-01
1.04957271e+00 7.44825974e-02 -2.09372073e-01 -6.68177962e-01
-1.36183238e+00 6.58831656e-01 2.39215329e-01 -1.58195153e-01
-3.64734620e-01 1.02321468e-01 9.14429307e-01 8.10583293e-01
3.95259522e-02 1.06685555e+00 -1.05063915e+00 -9.73570049e-02
-5.77357352e-01 2.35062197e-01 8.44401658e-01 1.34933531e+00
9.04516280e-01 3.22674870e-01 -1.11180522e-01 9.92715299e-01
4.11665231e-01 3.36148828e-01 -2.07466304e-01 -1.52219081e+00
4.64934796e-01 7.76559591e-01 4.14754078e-02 -7.72743106e-01
-4.14314091e-01 -1.66937545e-01 -6.69254661e-01 6.92236185e-01
3.94960433e-01 -6.76834909e-03 -1.16035306e+00 8.13871205e-01
9.62361693e-02 -4.76910770e-01 -1.08215697e-01 6.29384935e-01
6.29494905e-01 8.67819935e-02 -3.21482182e-01 7.39791095e-02
1.51804042e+00 -9.68095541e-01 -5.88069737e-01 -2.87096828e-01
4.97109115e-01 -1.05389929e+00 1.16681421e+00 1.08337665e+00
-9.96437311e-01 -9.69756901e-01 -1.14270437e+00 9.16094929e-02
-3.32735062e-01 1.16597796e-02 7.76985705e-01 6.33145213e-01
-9.76258039e-01 6.00320220e-01 -5.31721532e-01 -4.59127188e-01
6.62556589e-01 1.10911824e-01 -5.26071310e-01 -4.37610865e-01
-4.63825613e-01 7.06976116e-01 2.16768548e-01 2.33667538e-01
-1.39230299e+00 -9.84798431e-01 -1.15204513e+00 -4.76701021e-01
7.60945559e-01 -6.02132559e-01 1.49784946e+00 -2.67535448e-01
-1.05277169e+00 8.42813313e-01 7.21010983e-01 -2.24901691e-01
3.71958494e-01 -3.43469262e-01 -4.45100278e-01 5.51471896e-02
3.73055935e-01 4.54217792e-01 8.57444108e-01 -1.49509609e+00
-1.02483916e+00 -3.18888962e-01 3.01164478e-01 -2.40052670e-01
6.16924986e-02 -5.35270460e-02 -5.40779591e-01 -4.13168401e-01
1.61868691e-01 -6.14858508e-01 -4.36040938e-01 2.21385226e-01
-7.62144983e-01 -1.25094116e-01 1.24698269e+00 -7.33149290e-01
5.51088214e-01 -2.14182329e+00 -1.07197627e-01 -2.54293114e-01
1.51541933e-01 3.04219335e-01 -2.58209080e-01 5.85721731e-01
1.10560484e-01 -6.80516437e-02 -2.79783547e-01 -4.65454996e-01
2.96562284e-01 4.39562023e-01 1.32032588e-01 5.47318876e-01
6.87747180e-01 8.61576080e-01 -9.84764576e-01 -4.77245152e-01
8.96091819e-01 -1.20207317e-01 -3.30639273e-01 1.17208995e-01
-3.74522865e-01 4.22943622e-01 -3.34463865e-01 1.45075798e+00
7.65313268e-01 1.72172979e-01 -1.19105443e-01 -4.25687909e-01
-6.27425164e-02 -2.22928464e-01 -1.04754198e+00 2.17956471e+00
-4.92672384e-01 4.72522616e-01 7.43516624e-01 -1.21701586e+00
1.09999752e+00 1.94198415e-01 6.71359003e-01 -6.89286649e-01
-1.71029657e-01 -8.25130641e-02 -2.68749803e-01 -8.51053655e-01
5.76720417e-01 3.95197213e-01 -4.03457105e-01 -3.84575576e-02
3.68402660e-01 -9.53939378e-01 6.34503365e-01 1.12233954e-02
1.49168146e+00 1.87614322e-01 -8.54625329e-02 -3.47450286e-01
6.64373627e-03 5.35033464e-01 7.24634528e-01 7.15367794e-01
-5.90836287e-01 9.39167082e-01 5.21849513e-01 -7.55075395e-01
-1.01028121e+00 -1.30843914e+00 -1.81182057e-01 6.58789158e-01
1.33242995e-01 -2.50610739e-01 -4.16249931e-01 -1.10541916e+00
5.41673005e-01 2.14666978e-01 -5.09011269e-01 6.66836184e-03
-3.43264908e-01 -5.30442774e-01 3.59572381e-01 5.82415342e-01
4.64533329e-01 -1.34892309e+00 -7.61920691e-01 3.91498655e-01
2.60301262e-01 -1.21342623e+00 -1.81440935e-01 4.73606378e-01
-6.77072287e-01 -1.72223175e+00 -2.53851146e-01 -1.15773213e+00
5.76519549e-01 3.41427118e-01 1.95288527e+00 1.71172783e-01
-1.41752553e+00 8.79233181e-01 -5.10242343e-01 -7.57667065e-01
-6.69290364e-01 -3.36687624e-01 -1.66643351e-01 -7.18446791e-01
8.33617672e-02 -9.19633359e-02 -6.58682108e-01 3.13963443e-01
-6.16120934e-01 -5.87740242e-01 6.31931841e-01 8.39892447e-01
5.39783955e-01 3.74828100e-01 7.32465208e-01 -7.49110878e-01
6.00436866e-01 -2.34138012e-01 -7.12842643e-01 2.33282253e-01
-6.95188820e-01 -5.72584987e-01 4.18471307e-01 7.28356242e-02
-9.76099253e-01 1.90882593e-01 -1.56225324e-01 -3.06213379e-01
-7.67740488e-01 2.61510670e-01 -1.16868913e-01 -5.82096502e-02
5.10051548e-01 -9.07468870e-02 1.78260744e-01 -6.25064492e-01
4.02373284e-01 5.61013341e-01 1.01806462e+00 -6.74485922e-01
7.06214964e-01 2.25683346e-01 -1.94045052e-01 -9.50229645e-01
-7.20251918e-01 -7.34811127e-01 -7.70273924e-01 -1.74520746e-01
8.52936983e-01 -7.54709184e-01 -8.09453905e-01 7.69744098e-01
-1.02544892e+00 -1.29330412e-01 -9.05927300e-01 -5.26707526e-03
-6.67565167e-01 7.48206735e-01 -1.00908661e+00 -8.04177463e-01
-3.65616232e-01 -1.36944282e+00 1.35949314e+00 -4.36057560e-02
3.17998603e-02 -9.20115173e-01 -1.59878373e-01 6.99771285e-01
1.49595216e-01 5.56733370e-01 7.00552940e-01 -1.70487136e-01
-5.59455633e-01 -3.66879910e-01 -3.35562944e-01 9.28913593e-01
4.97204304e-01 2.99131811e-01 -7.44361103e-01 -4.14062649e-01
-2.22769812e-01 -7.30902791e-01 8.55549514e-01 4.35535520e-01
9.86967325e-01 4.99765337e-01 -1.19161814e-01 1.22043684e-01
1.38204801e+00 8.97898674e-02 4.14571881e-01 2.85546243e-01
7.98579156e-01 8.15569818e-01 1.42529178e+00 6.35463655e-01
2.43456423e-01 4.16953534e-01 1.26230180e+00 -5.03968954e-01
-2.22055122e-01 2.77796090e-01 1.09974802e-01 7.22859502e-01
3.56535614e-02 -1.09946271e-02 -9.00110483e-01 9.60717678e-01
-1.74827445e+00 -6.42984450e-01 -6.08308077e-01 1.77809310e+00
5.20771444e-01 3.20175081e-01 -7.89487138e-02 3.51298660e-01
3.50840688e-01 -2.88467836e-02 -7.36655831e-01 -5.51651478e-01
2.55181883e-02 8.94694850e-02 4.61752951e-01 -2.26055637e-01
-1.32481694e+00 5.21905243e-01 8.16469574e+00 3.39703083e-01
-2.12500185e-01 -2.30681032e-01 3.38358432e-01 4.00477946e-01
7.40258321e-02 -2.23925129e-01 -5.42484760e-01 3.79555076e-02
3.64073187e-01 4.82310683e-01 2.50411898e-01 9.85983253e-01
1.58156063e-02 -1.63455069e-01 -1.22121954e+00 7.70263612e-01
7.51511678e-02 -1.41251850e+00 -3.55904162e-01 1.72653534e-02
9.21073437e-01 1.00787692e-01 1.96666345e-02 3.55325967e-01
9.39331472e-01 -7.89650500e-01 9.12753820e-01 1.77955210e-01
7.76746929e-01 -5.26515782e-01 8.44634831e-01 -1.46813855e-01
-1.33734584e+00 -4.89759147e-01 -5.22788763e-01 -1.11436530e-04
3.17511380e-01 9.32012320e-01 -7.90650904e-01 1.00169647e+00
1.13738799e+00 8.80181491e-01 -4.44700539e-01 1.05521393e+00
6.71493169e-03 2.81952441e-01 1.70354411e-01 6.88861787e-01
7.65552893e-02 -5.94624542e-02 1.75714925e-01 1.29551303e+00
4.95817550e-02 -8.54680300e-01 8.98795545e-01 7.98871100e-01
8.87812227e-02 -5.40643156e-01 -9.06318307e-01 -2.29003001e-02
3.17253649e-01 1.44607508e+00 -7.28305221e-01 1.78258959e-02
-6.95151508e-01 1.17298996e+00 -8.38258415e-02 5.59438206e-02
-6.63927197e-01 -5.41250348e-01 1.28245318e+00 -2.87032723e-01
5.62110066e-01 -6.47098005e-01 -4.52656358e-01 -4.97393429e-01
-5.05617484e-02 -1.09581208e+00 3.70936841e-01 -5.47196925e-01
-1.72018731e+00 6.06303550e-02 -3.54553312e-01 -1.15204859e+00
1.25519648e-01 -1.10941756e+00 -4.12224323e-01 5.21877944e-01
-1.63765526e+00 -1.59494996e+00 -5.01068354e-01 3.07967693e-01
1.31949365e+00 -1.36038244e-01 8.81724060e-01 2.93022811e-01
-6.03211343e-01 1.34995714e-01 -8.63293782e-02 1.02285653e-01
7.19236374e-01 -1.64159644e+00 9.30678844e-01 6.99229836e-01
2.16626987e-01 1.88912138e-01 7.67809689e-01 -6.04113936e-01
-1.85783899e+00 -1.22932267e+00 -1.38751164e-01 -9.03388917e-01
6.98250890e-01 -1.06238708e-01 -5.60830593e-01 6.87657952e-01
3.74268889e-01 7.68170282e-02 3.76186639e-01 2.67353088e-01
-3.71506423e-01 -1.23430923e-01 -1.40469396e+00 -1.46851078e-01
1.21090615e+00 -4.82475132e-01 -4.27948505e-01 6.73756361e-01
8.11084926e-01 -4.10020262e-01 -1.54362357e+00 5.65017760e-01
1.65214911e-01 -1.03740478e+00 1.24096644e+00 -3.45830083e-01
5.14225006e-01 -4.63125199e-01 -6.40104860e-02 -1.49177706e+00
-5.50864398e-01 -4.22374815e-01 -2.83309091e-02 1.44884944e+00
2.01330170e-01 -2.74471551e-01 7.34215140e-01 1.14804663e-01
-1.05237174e+00 -5.47252059e-01 -5.41534305e-01 -8.92985404e-01
-3.47924411e-01 -5.39488018e-01 6.48773193e-01 8.38459909e-01
-2.24668548e-01 -8.54043737e-02 -2.18036875e-01 3.28855962e-01
8.40587556e-01 1.61975339e-01 8.16462219e-01 -1.53403234e+00
-2.99182177e-01 -2.34073415e-01 -6.27323687e-01 -7.89211631e-01
-4.22952682e-01 -2.80909717e-01 5.38837314e-01 -2.04701400e+00
-1.71832845e-01 -4.82398838e-01 -3.51346344e-01 7.26034343e-01
1.23946741e-02 6.15087152e-01 2.04232503e-02 -1.62442207e-01
-8.62362146e-01 1.79724861e-02 1.59884000e+00 -6.42415166e-01
1.19465083e-01 -3.75286564e-02 -7.32370496e-01 5.08720756e-01
6.67834759e-01 1.70401782e-01 -4.14854258e-01 -8.00064802e-01
-1.99494630e-01 -1.56881198e-01 3.92845124e-01 -1.23460889e+00
-2.90764701e-02 -9.47295874e-02 3.76040190e-01 -6.44789755e-01
4.12938654e-01 -8.88529539e-01 -1.45324767e-01 5.56927204e-01
1.65570185e-01 2.41602734e-01 1.74625814e-01 7.09654689e-01
-2.66146630e-01 -3.27851176e-01 5.49752653e-01 -6.26688540e-01
-1.33545685e+00 2.38376662e-01 -1.95033863e-01 6.78010359e-02
1.20655954e+00 -5.45880497e-01 -4.82211828e-01 2.94551611e-01
-6.89761281e-01 4.40136582e-01 8.09515834e-01 6.98012710e-01
5.35271943e-01 -7.54129171e-01 -8.49553645e-01 5.13648927e-01
6.12527013e-01 5.42700410e-01 3.16919953e-01 5.09293497e-01
-6.35475934e-01 6.61963085e-03 -3.97966474e-01 -9.11164522e-01
-1.08805740e+00 6.06736124e-01 2.60919682e-03 -2.74962902e-01
-8.84900153e-01 1.04761767e+00 -2.19485760e-01 -6.53966188e-01
1.05069093e-01 -5.03727138e-01 2.13507757e-01 -9.15962979e-02
2.29561344e-01 6.42320335e-01 5.13836801e-01 -9.37468559e-02
-1.32032514e-01 4.90387559e-01 -3.82267311e-02 6.39404476e-01
1.24836183e+00 -7.54701123e-02 -2.39172988e-02 2.27453396e-01
6.63134933e-01 -4.37562585e-01 -1.57381499e+00 1.69559181e-01
4.30430591e-01 -5.43012381e-01 -7.59046450e-02 -9.93128657e-01
-1.29155302e+00 6.97391808e-01 6.91485226e-01 8.81838143e-01
1.16773081e+00 1.88850701e-01 7.90876925e-01 3.65858257e-01
7.22777247e-01 -1.43604457e+00 4.90009487e-01 4.46987242e-01
8.87358308e-01 -1.70625973e+00 -2.23343391e-02 -7.24523723e-01
-6.78520024e-01 1.00354528e+00 8.24514389e-01 2.86163762e-02
3.57612222e-01 6.99491918e-01 5.03239632e-01 -8.33841503e-01
-5.27942836e-01 -1.12317629e-01 -1.56727031e-01 1.47156346e+00
2.08994240e-01 8.25355127e-02 4.69900280e-01 8.20022821e-02
2.22149074e-01 -2.91315556e-01 4.19029385e-01 1.33155835e+00
-5.77273488e-01 -1.05502975e+00 -5.23037553e-01 8.22470129e-01
-2.43596151e-01 3.69354784e-01 -1.67218104e-01 6.43981457e-01
4.18249756e-01 1.48593533e+00 1.23064825e-02 -3.51054430e-01
1.00454617e+00 -4.75008100e-01 5.30559480e-01 -8.76319826e-01
-7.23732650e-01 -3.53898048e-01 4.10555333e-01 -9.18242931e-01
-7.49938563e-02 -8.81510139e-01 -1.09218228e+00 5.06157763e-02
-6.47069216e-01 -1.95868537e-01 8.56706023e-01 5.44626117e-01
3.93963128e-01 1.10632181e+00 6.07984781e-01 -9.75824714e-01
-9.51499999e-01 -8.91927063e-01 -1.15161026e+00 6.67192698e-01
2.72390872e-01 -5.89537978e-01 -3.12245749e-02 3.61523211e-01] | [7.463834762573242, 1.9305537939071655] |
a7e32290-b769-41b9-ba08-b8408274534b | semi-supervised-domain-adaptive-medical-image | 2307.02798 | null | https://arxiv.org/abs/2307.02798v1 | https://arxiv.org/pdf/2307.02798v1.pdf | Semi-supervised Domain Adaptive Medical Image Segmentation through Consistency Regularized Disentangled Contrastive Learning | Although unsupervised domain adaptation (UDA) is a promising direction to alleviate domain shift, they fall short of their supervised counterparts. In this work, we investigate relatively less explored semi-supervised domain adaptation (SSDA) for medical image segmentation, where access to a few labeled target samples can improve the adaptation performance substantially. Specifically, we propose a two-stage training process. First, an encoder is pre-trained in a self-learning paradigm using a novel domain-content disentangled contrastive learning (CL) along with a pixel-level feature consistency constraint. The proposed CL enforces the encoder to learn discriminative content-specific but domain-invariant semantics on a global scale from the source and target images, whereas consistency regularization enforces the mining of local pixel-level information by maintaining spatial sensitivity. This pre-trained encoder, along with a decoder, is further fine-tuned for the downstream task, (i.e. pixel-level segmentation) using a semi-supervised setting. Furthermore, we experimentally validate that our proposed method can easily be extended for UDA settings, adding to the superiority of the proposed strategy. Upon evaluation on two domain adaptive image segmentation tasks, our proposed method outperforms the SoTA methods, both in SSDA and UDA settings. Code is available at https://github.com/hritam-98/GFDA-disentangled | ['Zhaozheng Yin', 'Hritam Basak'] | 2023-07-06 | null | null | null | null | ['contrastive-learning', 'medical-image-segmentation', 'contrastive-learning', 'unsupervised-domain-adaptation', 'self-learning'] | ['computer-vision', 'medical', 'methodology', 'methodology', 'natural-language-processing'] | [ 6.41940057e-01 2.74348587e-01 -5.80534339e-01 -5.78597784e-01
-1.07957458e+00 -5.76374471e-01 5.47379673e-01 3.49574909e-02
-5.40985942e-01 7.07161903e-01 1.02956317e-01 -8.51157680e-02
-1.41759872e-01 -5.33644974e-01 -7.25740671e-01 -1.13869476e+00
3.09316456e-01 4.65941012e-01 1.44879967e-01 2.35250350e-02
-1.58270493e-01 2.80049115e-01 -1.12854385e+00 7.88331777e-02
1.10647273e+00 7.64901698e-01 4.52995718e-01 4.98429060e-01
2.91693956e-01 4.43623275e-01 -8.39196891e-02 -2.17052385e-01
3.36679846e-01 -5.76603115e-01 -8.25271368e-01 4.62085277e-01
3.79926205e-01 -3.15768600e-01 -1.17352888e-01 1.03376043e+00
6.05168581e-01 1.00639030e-01 6.99264228e-01 -8.40401053e-01
-5.88952959e-01 3.49120855e-01 -7.51190364e-01 1.88211903e-01
-1.30840853e-01 2.12265983e-01 1.08212900e+00 -8.54704618e-01
7.42612243e-01 9.24685895e-01 3.30522090e-01 7.38498926e-01
-1.52389121e+00 -5.61731279e-01 2.32045263e-01 -7.75116384e-02
-1.16192102e+00 -4.06698108e-01 1.16336608e+00 -4.73083168e-01
5.55775464e-01 -1.66194215e-02 2.69286126e-01 1.30691171e+00
2.42445134e-02 9.84237075e-01 1.40142453e+00 -5.23291469e-01
5.51681817e-01 3.06757540e-01 7.54672140e-02 6.63497567e-01
2.22648695e-01 1.17907271e-01 -4.91328448e-01 -1.67850498e-02
8.68705094e-01 -2.10700601e-01 -1.02207668e-01 -9.49868977e-01
-1.21897650e+00 8.25363576e-01 4.65483218e-01 2.90206879e-01
-4.29664791e-01 -3.60944986e-01 4.57667351e-01 1.53376698e-01
6.65245712e-01 2.92104930e-01 -5.49546778e-01 1.85090661e-01
-1.03785241e+00 8.83583426e-02 3.97816628e-01 9.48618412e-01
9.04869854e-01 -1.52630270e-01 -2.46638909e-01 9.60556924e-01
3.50540578e-01 3.66172820e-01 5.89402497e-01 -8.23173642e-01
3.63351524e-01 4.98892456e-01 -1.16910912e-01 -5.74677110e-01
-3.60240012e-01 -6.17567062e-01 -9.15920496e-01 2.28881612e-01
7.49794900e-01 -5.62436096e-02 -1.07481599e+00 2.15895343e+00
5.32888293e-01 2.69385397e-01 1.62324443e-01 1.16473138e+00
3.62789929e-01 2.43977636e-01 2.90518999e-01 -2.21127078e-01
1.42174625e+00 -9.94049847e-01 -5.42448580e-01 -4.04285908e-01
5.87085009e-01 -6.02719486e-01 1.19519305e+00 2.94425786e-01
-1.16997814e+00 -4.89751130e-01 -1.12768316e+00 -2.93102831e-01
-2.03124315e-01 3.45801026e-01 3.48322302e-01 5.61101377e-01
-8.37742686e-01 3.30364347e-01 -1.09988832e+00 -4.43373233e-01
7.65160382e-01 3.92103583e-01 -3.63498241e-01 -1.70830384e-01
-1.17714417e+00 5.67923963e-01 3.69796067e-01 -1.64334193e-01
-8.32597077e-01 -6.51945591e-01 -1.01606417e+00 -1.83932617e-01
4.39787239e-01 -9.90456879e-01 1.10078883e+00 -1.11611629e+00
-1.70106506e+00 1.34000099e+00 -1.88854173e-01 -5.93556046e-01
5.44261932e-01 5.51787540e-02 -1.50431529e-01 4.68369335e-01
3.81081313e-01 8.40629518e-01 1.11596465e+00 -1.42973053e+00
-5.25967956e-01 -6.57989562e-01 -4.22611944e-02 4.27082986e-01
-5.11108160e-01 -3.09920698e-01 -5.16160190e-01 -9.16385829e-01
9.09356251e-02 -1.01688612e+00 -2.57676214e-01 1.05132923e-01
-4.00680393e-01 1.82725698e-01 6.37741387e-01 -7.02596128e-01
1.05903637e+00 -2.39631963e+00 3.62109482e-01 2.19957814e-01
2.07917586e-01 1.68403029e-01 -1.68557495e-01 7.69848078e-02
-2.09693924e-01 -2.68840730e-01 -8.77665579e-01 -6.42738402e-01
-1.32416725e-01 2.56329745e-01 -1.19098909e-01 6.40176892e-01
4.76262748e-01 8.47260177e-01 -1.05247355e+00 -6.36714756e-01
1.82317525e-01 3.79308164e-01 -6.45210981e-01 1.22991219e-01
-3.22067201e-01 9.51669693e-01 -5.99017680e-01 4.78098899e-01
7.82169700e-01 -4.73252058e-01 2.82497525e-01 -6.81187958e-02
5.92887439e-02 1.96532086e-01 -9.38493490e-01 2.23878908e+00
-5.10320127e-01 3.45479608e-01 1.11562207e-01 -1.33341753e+00
7.94490039e-01 2.55676419e-01 6.57903194e-01 -8.48887980e-01
-4.47959639e-03 2.32077107e-01 -3.72062437e-02 -4.03730333e-01
9.83478948e-02 -2.53204167e-01 -1.76658839e-01 3.57539356e-01
3.19051236e-01 9.73697305e-02 -2.16230396e-02 2.98049539e-01
9.05244768e-01 4.52423722e-01 6.44329011e-01 -4.80110019e-01
6.32621169e-01 1.12857178e-01 6.77950144e-01 6.53464556e-01
-4.58529651e-01 8.33285749e-01 3.69138062e-01 7.94232711e-02
-1.12383378e+00 -1.28644824e+00 -3.69119883e-01 1.11810303e+00
3.86163443e-01 8.65001902e-02 -8.89531493e-01 -9.09418285e-01
-5.36947437e-02 6.70784712e-01 -7.30535746e-01 -2.79532313e-01
-5.04212618e-01 -6.63537383e-01 3.69913995e-01 6.31533980e-01
6.87022090e-01 -7.80890822e-01 -4.14033264e-01 1.32375121e-01
7.97601230e-03 -1.20659125e+00 -7.72683859e-01 4.27787036e-01
-9.89047766e-01 -7.83455312e-01 -1.03115070e+00 -9.57713187e-01
8.64508927e-01 2.56316215e-01 7.69334853e-01 -4.89062220e-01
-2.60208622e-02 3.92695844e-01 -3.09525579e-01 -5.34959771e-02
-3.91961932e-01 2.66373217e-01 -1.60589039e-01 2.57419258e-01
2.63279825e-01 -6.62155449e-01 -8.23544204e-01 3.47947478e-01
-9.59470928e-01 2.84620345e-01 9.22165394e-01 1.18854976e+00
9.15102482e-01 -3.22195351e-01 7.07645774e-01 -1.41255188e+00
3.08434606e-01 -4.94873255e-01 -4.67325062e-01 1.43674597e-01
-6.98924243e-01 1.21303067e-01 6.09431028e-01 -5.10450006e-01
-1.50630701e+00 5.66062629e-01 6.90877717e-03 -4.06949252e-01
-5.63278019e-01 3.68545562e-01 -4.71880972e-01 8.62248242e-02
8.58672440e-01 4.23481345e-01 2.02393696e-01 -3.87105316e-01
5.17800629e-01 5.44348121e-01 6.74219012e-01 -6.05933785e-01
8.21023583e-01 7.57151842e-01 -1.88959584e-01 -6.54882371e-01
-8.92234743e-01 -7.01362252e-01 -1.04192436e+00 1.31247357e-01
9.90677059e-01 -1.18409967e+00 -2.25969910e-04 3.98476064e-01
-8.12611222e-01 -6.33694530e-01 -5.42799890e-01 5.46175659e-01
-9.13483083e-01 3.39425683e-01 -4.88418728e-01 -3.19593817e-01
-1.43386215e-01 -1.13575673e+00 1.18296540e+00 1.92196205e-01
-1.82160288e-01 -1.27823997e+00 9.44188908e-02 6.34327650e-01
1.05314367e-01 2.75722086e-01 9.01162028e-01 -8.30045581e-01
-4.25374061e-01 1.20941743e-01 -2.06355974e-01 5.22666514e-01
3.63296032e-01 -6.47964954e-01 -1.12968564e+00 -3.84478360e-01
6.82144016e-02 -3.73380899e-01 1.00812459e+00 5.98986089e-01
1.08902097e+00 -1.55298933e-01 -2.78063834e-01 6.88830853e-01
1.19753015e+00 -8.36046599e-03 3.22863311e-01 3.26202989e-01
6.34199619e-01 4.65105206e-01 7.88537920e-01 4.29647148e-01
4.73733336e-01 7.41208851e-01 1.07206218e-01 -4.80939209e-01
-5.31528950e-01 -2.77128607e-01 2.30279401e-01 6.57166123e-01
1.44499734e-01 -8.70281979e-02 -8.36372375e-01 5.97166777e-01
-1.85435712e+00 -5.99519491e-01 6.82152063e-02 2.18700361e+00
1.20483351e+00 1.15086839e-01 3.28260005e-01 -1.78308075e-03
5.64719021e-01 1.24902584e-01 -9.93436456e-01 -2.91515738e-02
-2.48119444e-01 2.76546597e-01 5.41775465e-01 6.22493088e-01
-1.27221489e+00 9.41583276e-01 5.06198978e+00 8.79964411e-01
-1.15076518e+00 3.84237915e-01 6.79469585e-01 8.70276615e-02
-3.70847970e-01 -6.84789047e-02 -3.99015546e-01 4.06682163e-01
6.42261207e-01 -7.15036094e-02 1.66530386e-01 7.62145281e-01
3.22881252e-01 -8.93025622e-02 -1.12302434e+00 6.90522790e-01
-4.25699577e-02 -1.15642941e+00 -1.87940165e-01 1.00261949e-01
8.45526397e-01 -1.65245265e-01 3.67479980e-01 1.48138389e-01
4.85309660e-02 -5.64963043e-01 4.95615780e-01 1.06672302e-01
9.51974809e-01 -5.12352109e-01 3.75303209e-01 3.32365483e-01
-7.45118082e-01 1.26668900e-01 -3.06881871e-02 2.42549822e-01
-8.34947750e-02 5.89171171e-01 -8.06057870e-01 6.84062183e-01
4.09711033e-01 9.50656235e-01 -4.82797325e-01 6.62564397e-01
-3.80282819e-01 8.62588763e-01 -2.96915341e-02 4.56054837e-01
2.60661542e-01 -1.52270854e-01 6.90405905e-01 1.25488484e+00
-2.04803228e-01 1.10236645e-01 -7.84671307e-02 9.54260588e-01
-1.43843936e-02 5.49088158e-02 -4.14474458e-01 2.12035805e-01
3.04783493e-01 1.19157088e+00 -8.30221176e-01 -3.00389081e-01
-4.66636896e-01 1.35979772e+00 3.36299300e-01 6.72852576e-01
-7.70854592e-01 -6.96366206e-02 5.10185182e-01 1.29440382e-01
2.85379499e-01 -1.32100061e-01 -7.19999731e-01 -1.27254879e+00
9.70284175e-03 -8.12906861e-01 6.28673553e-01 -5.01557469e-01
-1.33252740e+00 2.84841955e-01 7.66444877e-02 -1.39664710e+00
-2.54892975e-01 -3.47523451e-01 -3.73606354e-01 8.38590980e-01
-1.78471267e+00 -1.34898877e+00 -2.11494371e-01 7.55982578e-01
6.06296003e-01 -9.71551463e-02 7.35171735e-01 3.39817762e-01
-7.46043980e-01 7.59933174e-01 3.91621232e-01 7.92948157e-02
1.00938559e+00 -1.47104287e+00 1.54893413e-01 9.14407194e-01
2.62840651e-02 5.74261069e-01 5.70965767e-01 -5.82021773e-01
-1.03935456e+00 -1.16366255e+00 4.93155807e-01 -2.15347558e-01
5.22032142e-01 -6.55766308e-01 -1.05174112e+00 6.96731627e-01
1.76472783e-01 2.06133038e-01 7.36679733e-01 -2.35057902e-02
-3.58855546e-01 -6.43105730e-02 -1.29705715e+00 3.65455747e-01
8.75306427e-01 -5.21726370e-01 -4.64180291e-01 2.99662799e-01
7.07599461e-01 -5.80533624e-01 -8.92073750e-01 2.56568015e-01
3.46300691e-01 -7.42584765e-01 1.00782835e+00 -4.01949674e-01
4.92003322e-01 -2.78232932e-01 3.18488926e-02 -1.23360395e+00
-3.37253273e-01 -4.97305721e-01 -3.46234888e-02 1.12910032e+00
3.61036003e-01 -7.44516432e-01 8.47228706e-01 4.73187417e-01
-2.11637154e-01 -7.71907926e-01 -9.26299512e-01 -8.19156170e-01
3.15167785e-01 -2.51220256e-01 1.43871352e-01 1.09918332e+00
-4.45734411e-02 3.97728801e-01 -1.63761854e-01 4.23239589e-01
9.05748963e-01 1.94116607e-01 4.60583329e-01 -8.43983948e-01
-5.48268259e-01 -3.47257882e-01 -1.35525778e-01 -1.20424223e+00
2.08567172e-01 -1.05137312e+00 -1.62443575e-02 -1.07459342e+00
3.76595467e-01 -4.91545647e-01 -4.60873634e-01 5.36569059e-01
-2.26281568e-01 1.56652614e-01 -8.94059166e-02 4.12066191e-01
-5.26844203e-01 5.65260589e-01 1.39513242e+00 -3.35405916e-02
-3.41361582e-01 1.37647137e-01 -7.37865269e-01 5.38247824e-01
8.44017148e-01 -5.26326895e-01 -6.31168008e-01 -3.46840858e-01
-3.76965672e-01 7.08634034e-02 4.88316327e-01 -6.25293434e-01
1.78469315e-01 -5.42864017e-02 3.43327552e-01 -1.70902923e-01
1.44980654e-01 -6.92506313e-01 -3.46344382e-01 3.74788612e-01
-6.07288718e-01 -5.83347321e-01 1.98514074e-01 7.02784657e-01
-3.23386192e-01 -8.87685642e-02 1.22525811e+00 1.41136691e-01
-8.30886662e-01 3.13992411e-01 -6.54009879e-02 1.35102153e-01
1.04392958e+00 -3.13788831e-01 6.64408959e-04 -4.50987443e-02
-9.69079912e-01 2.11702392e-01 5.04983485e-01 1.98560521e-01
5.01783907e-01 -1.11786783e+00 -5.84760010e-01 3.94434333e-01
4.04720157e-01 3.58590990e-01 4.97029155e-01 1.01240051e+00
-1.07799463e-01 3.54150981e-01 -2.26833567e-01 -7.55216062e-01
-9.66419578e-01 5.31659007e-01 2.50417084e-01 -3.46425742e-01
-6.74514294e-01 7.46886969e-01 6.14375889e-01 -3.95941615e-01
1.30395025e-01 -2.00648248e-01 5.46425767e-02 -1.67295083e-01
2.00757891e-01 1.42382950e-01 -1.49874939e-02 -6.07165813e-01
-4.62852895e-01 5.10053217e-01 -2.80187100e-01 -2.27466732e-01
1.30099261e+00 -3.89102191e-01 3.07711661e-01 2.74338424e-01
1.31680298e+00 -7.27032125e-02 -1.72836852e+00 -6.85713470e-01
-2.25772876e-02 -2.81394333e-01 1.71548650e-01 -9.62734878e-01
-1.01167595e+00 9.10796583e-01 7.68847167e-01 -3.55023116e-01
1.55620515e+00 1.99996963e-01 6.38200879e-01 -1.35461912e-01
4.50854115e-02 -1.08792090e+00 2.97697991e-01 5.49455546e-03
6.99073672e-01 -1.63637352e+00 -6.28838688e-02 -5.09709895e-01
-9.83604133e-01 8.01285028e-01 5.92074573e-01 -1.62619919e-01
5.64974725e-01 1.52097106e-01 3.97749357e-02 -4.43273671e-02
-5.04842043e-01 -3.64789546e-01 4.06777203e-01 6.69839978e-01
2.98694402e-01 4.67059724e-02 -2.68999189e-01 7.35252976e-01
1.91825911e-01 -1.45083219e-02 2.22671792e-01 7.98819542e-01
-1.02486692e-01 -1.11814630e+00 -1.15483813e-01 1.12150013e-01
-2.03068316e-01 7.89752603e-02 -2.05975160e-01 8.74148786e-01
1.62877738e-01 6.06163502e-01 -7.75683718e-03 -4.82764207e-02
3.00824791e-01 -8.86813086e-03 5.03239334e-01 -7.80139208e-01
-2.11043462e-01 3.62262726e-01 -2.68364400e-01 -5.33588111e-01
-6.01641059e-01 -1.08017170e+00 -1.43376195e+00 2.68704772e-01
-1.07872687e-01 -2.88828135e-01 4.52442676e-01 9.18331385e-01
4.08547670e-01 3.74145061e-01 6.58329189e-01 -4.93694276e-01
-5.88108897e-01 -6.89609647e-01 -6.09222651e-01 7.72540987e-01
5.48714280e-01 -6.37966514e-01 -1.74142078e-01 4.35173512e-01] | [14.559592247009277, -1.9997211694717407] |
59a95777-e086-480b-8473-6a3884d60db4 | the-mgb-2-challenge-arabic-multi-dialect | 1609.05625 | null | https://arxiv.org/abs/1609.05625v3 | https://arxiv.org/pdf/1609.05625v3.pdf | The MGB-2 Challenge: Arabic Multi-Dialect Broadcast Media Recognition | This paper describes the Arabic Multi-Genre Broadcast (MGB-2) Challenge for SLT-2016. Unlike last year's English MGB Challenge, which focused on recognition of diverse TV genres, this year, the challenge has an emphasis on handling the diversity in dialect in Arabic speech. Audio data comes from 19 distinct programmes from the Aljazeera Arabic TV channel between March 2005 and December 2015. Programmes are split into three groups: conversations, interviews, and reports. A total of 1,200 hours have been released with lightly supervised transcriptions for the acoustic modelling. For language modelling, we made available over 110M words crawled from Aljazeera Arabic website Aljazeera.net for a 10 year duration 2000-2011. Two lexicons have been provided, one phoneme based and one grapheme based. Finally, two tasks were proposed for this year's challenge: standard speech transcription, and word alignment. This paper describes the task data and evaluation process used in the MGB challenge, and summarises the results obtained. | ['Yifan Zhang', 'Steve Renals', 'James Glass', 'Yacine Messaoui', 'Peter Bell', 'Hamdy Mubarak', 'Ahmed Ali'] | 2016-09-19 | null | null | null | null | ['acoustic-modelling'] | ['speech'] | [ 9.98721495e-02 -7.31884912e-02 1.77298874e-01 -6.05568051e-01
-1.78111982e+00 -6.32913470e-01 5.95731676e-01 2.91848868e-01
-4.09408778e-01 6.25944436e-01 5.52550018e-01 -1.55086219e-01
-8.71362630e-03 -3.52231979e-01 -5.23888648e-01 -7.05769300e-01
-2.21725971e-01 9.16817307e-01 3.10947094e-02 -9.60836709e-01
-8.31453316e-03 4.05027978e-02 -1.09002459e+00 7.45392263e-01
3.48218501e-01 8.20916891e-01 2.27416873e-01 1.15199625e+00
1.28171846e-01 6.08228207e-01 -1.05329990e+00 -6.38367176e-01
-2.01838538e-01 -4.70923841e-01 -1.19592619e+00 1.35905370e-01
4.16016400e-01 -1.12877078e-01 -2.96532065e-01 5.40013790e-01
9.30936456e-01 1.71869442e-01 5.02923429e-01 -1.07139528e+00
-6.86585978e-02 1.39527452e+00 -3.10943574e-01 4.50928003e-01
8.99612069e-01 -3.56398493e-01 1.29127574e+00 -9.59578991e-01
5.96539557e-01 1.44400573e+00 7.83667028e-01 3.87485921e-01
-8.48637342e-01 -6.81111634e-01 -6.04225621e-02 5.21013558e-01
-1.59751272e+00 -1.08510637e+00 3.74850780e-01 -2.07928270e-01
1.31232846e+00 6.53312147e-01 5.69795310e-01 1.60266840e+00
-1.73219159e-01 8.42782319e-01 1.17656517e+00 -8.63863707e-01
-1.56497046e-01 -1.56472787e-01 5.48625030e-02 1.99586526e-01
-6.36006296e-01 -6.02052689e-01 -9.55911100e-01 -8.46028253e-02
-3.90692167e-02 -1.09055042e+00 -2.87026167e-01 8.25561166e-01
-1.30968297e+00 7.74123788e-01 -5.36159337e-01 4.61179465e-01
-3.65099639e-01 -2.47951701e-01 7.91286111e-01 6.25472963e-01
4.67155516e-01 3.69894654e-02 -3.57777685e-01 -7.56572127e-01
-7.38499224e-01 2.79587150e-01 9.60566580e-01 1.19163573e+00
2.02423155e-01 3.58252048e-01 4.62252088e-02 1.62826848e+00
6.51057839e-01 9.70529020e-01 6.06889129e-01 -6.63160205e-01
8.93013537e-01 -3.43355745e-01 -2.21949220e-01 -1.09236801e+00
-4.97107655e-01 7.59251863e-02 -5.96973479e-01 -6.80294454e-01
5.14091372e-01 -3.08292806e-01 -7.86605418e-01 1.43428802e+00
6.88312110e-03 -3.72042924e-01 1.46841362e-01 6.91657722e-01
1.23442817e+00 1.52649307e+00 -3.51033509e-01 -4.29844558e-01
1.48888123e+00 -1.11883283e+00 -1.15263474e+00 -1.77258149e-01
5.23327410e-01 -1.54953134e+00 1.06495702e+00 9.68573630e-01
-1.55243313e+00 -3.21931928e-01 -1.17976093e+00 4.61431593e-02
-3.61125410e-01 2.67495099e-03 1.18938591e-02 1.12903321e+00
-1.15896773e+00 -3.08819443e-01 -5.42610645e-01 -4.99336749e-01
-3.19927305e-01 1.20966770e-01 -2.41365567e-01 -1.92448854e-01
-1.59165764e+00 8.89622748e-01 3.19692105e-01 7.24011958e-02
-1.10192978e+00 9.52184480e-03 -8.93041372e-01 -3.33714932e-01
1.69996873e-01 6.25412047e-01 1.68470287e+00 -7.37602532e-01
-1.79855251e+00 1.08063841e+00 -8.17952603e-02 -5.36296427e-01
-2.88990047e-02 -2.20489968e-02 -1.19711101e+00 2.76116058e-02
-5.05575277e-02 3.13685298e-01 5.98877788e-01 -9.11149204e-01
-1.02670622e+00 -8.20778906e-02 -2.35210150e-01 3.10010642e-01
-1.86104283e-01 8.31779420e-01 -8.35708261e-01 -1.13033104e+00
1.45046815e-01 -1.11128688e+00 2.93129146e-01 -1.32954741e+00
-2.65753150e-01 9.65823531e-02 5.34716725e-01 -1.43053603e+00
1.58050537e+00 -2.31830382e+00 2.68984884e-01 3.11069220e-01
-4.60866958e-01 -1.66033432e-01 -2.92320192e-01 8.45529556e-01
1.12548761e-01 -1.72620252e-01 -8.13036114e-02 -5.90986669e-01
2.37781629e-01 3.20535094e-01 -3.93716991e-01 4.93047088e-01
1.27758207e-02 3.81983399e-01 -6.15841687e-01 -1.47627607e-01
-2.15597823e-01 3.85224909e-01 1.48533992e-02 1.05800806e-02
-3.25480141e-02 3.90651584e-01 2.69777685e-01 9.34559047e-01
6.34658754e-01 6.21377528e-01 3.07881087e-01 -1.75991654e-01
-1.56086728e-01 1.01240444e+00 -1.12103736e+00 1.82854354e+00
-6.09966278e-01 8.19344938e-01 6.28277421e-01 -8.62223148e-01
8.19293439e-01 9.19944942e-01 2.24780649e-01 -6.79628134e-01
3.46027732e-01 5.46299756e-01 5.89834899e-02 -1.82803735e-01
9.06871557e-01 7.08246827e-02 -6.12026036e-01 4.55970645e-01
7.11863220e-01 -6.54932797e-01 6.09987080e-01 1.65155545e-01
8.14309895e-01 -3.44556123e-01 2.48736396e-01 -2.35602543e-01
6.16959929e-01 1.74198836e-01 2.00879842e-01 6.32936239e-01
-1.98751077e-01 8.93555045e-01 1.26649484e-01 2.05315817e-02
-6.22790515e-01 -7.42036879e-01 -1.18689984e-01 1.75361371e+00
-4.49775189e-01 -1.00066721e+00 -7.25334048e-01 -2.17846274e-01
-7.21396565e-01 6.51719749e-01 -2.50313967e-01 2.47410968e-01
-9.52803254e-01 -1.04078698e+00 1.34715343e+00 -1.71309188e-01
4.61915761e-01 -1.05730808e+00 2.83619374e-01 5.56564748e-01
-1.13611436e+00 -1.28591275e+00 -8.36066306e-01 3.13199222e-01
3.90236266e-02 -6.10461831e-01 -8.14684331e-01 -1.13835490e+00
-3.49292815e-01 -5.62673770e-02 1.27435195e+00 -5.20675004e-01
9.59740579e-02 7.04679728e-01 -9.47075605e-01 -7.53828526e-01
-1.06180882e+00 5.54393053e-01 -2.78293304e-02 5.84750995e-02
3.11508894e-01 -2.05048591e-01 2.27382675e-01 4.04782265e-01
-5.09223282e-01 -2.11581901e-01 2.87797064e-01 7.18251944e-01
4.95994687e-01 -2.15874076e-01 8.01993728e-01 -6.22576535e-01
6.69956625e-01 -5.17779469e-01 -2.20966965e-01 1.66215390e-01
-2.83870921e-02 -7.71326423e-01 9.01602060e-02 -2.85117388e-01
-9.97995257e-01 -1.73595160e-01 -1.07370472e+00 5.92001796e-01
-3.68523784e-02 9.35552120e-01 -4.10816908e-01 1.86414793e-01
5.54070294e-01 2.17275247e-01 -3.27299058e-01 -4.77722913e-01
1.25428244e-01 1.66387057e+00 7.15921402e-01 -6.99304819e-01
1.56592622e-01 1.55066941e-02 -8.03796530e-01 -1.37854445e+00
-5.73443770e-01 -5.83259642e-01 -3.60520363e-01 -6.57997966e-01
8.85475636e-01 -1.20109129e+00 -3.49795580e-01 1.12168860e+00
-1.06628823e+00 -3.98885429e-01 6.57571107e-02 4.08705235e-01
-5.70344031e-01 1.29889697e-02 -1.10221684e+00 -6.39169395e-01
-2.24659652e-01 -1.39001060e+00 9.31220531e-01 -3.53785008e-01
-5.02318859e-01 -8.78477931e-01 1.74915999e-01 5.42149007e-01
2.70831883e-01 -1.95199952e-01 6.08064830e-01 -7.30825245e-01
1.33510455e-01 -5.12782335e-02 4.04119194e-01 3.57332021e-01
2.73747712e-01 -6.00662865e-02 -1.23764253e+00 -4.69568551e-01
-6.02874458e-02 -7.04446912e-01 2.79756337e-01 4.10921156e-01
5.62059700e-01 -3.06890041e-01 2.97014922e-01 2.74417818e-01
6.37793481e-01 4.68222618e-01 5.61200261e-01 5.62644780e-01
3.78613234e-01 5.87258637e-01 8.92243087e-01 5.14616847e-01
7.86483049e-01 1.03888357e+00 1.57181963e-01 3.11046898e-01
-1.19907871e-01 2.86589742e-01 1.00082374e+00 2.06864762e+00
4.71495353e-02 -6.73601925e-01 -1.44565535e+00 8.88387680e-01
-1.44310093e+00 -5.22174954e-01 -3.85883033e-01 1.87784612e+00
1.36153626e+00 5.89612573e-02 4.60264146e-01 4.44518238e-01
7.07567275e-01 2.99110860e-01 5.69037020e-01 -8.36986423e-01
-6.83808923e-01 1.97768554e-01 4.09595579e-01 8.06853354e-01
-1.14866865e+00 9.77524281e-01 6.37486458e+00 1.60177505e+00
-1.00601053e+00 5.32037675e-01 5.36520898e-01 -1.77333355e-01
-1.18703306e-01 -3.32745016e-01 -9.58143175e-01 4.15491074e-01
1.79340148e+00 -3.19171324e-02 5.75541854e-01 1.76481009e-01
5.92354275e-02 -2.26727854e-02 -6.00351810e-01 1.17077875e+00
5.36542058e-01 -1.18963313e+00 -5.85030578e-02 -2.89529234e-01
6.27228081e-01 5.29864252e-01 1.99817363e-02 5.65401375e-01
2.43008122e-01 -1.16202927e+00 1.38283026e+00 4.78748344e-02
9.59262967e-01 -1.12621653e+00 6.65002525e-01 1.64097831e-01
-1.19726276e+00 1.82375684e-01 1.43007003e-02 1.64528906e-01
4.67359871e-01 1.42426357e-01 -9.42934632e-01 7.80499458e-01
9.95344162e-01 7.32587099e-01 -4.22048062e-01 6.31822288e-01
1.46713760e-02 1.42687261e+00 -3.49563241e-01 3.92344370e-02
5.15449345e-01 -2.71043837e-01 9.57818985e-01 1.96287858e+00
2.80089378e-01 -9.16406661e-02 1.36311769e-01 -4.10272568e-01
-1.98628474e-02 5.29051125e-01 -1.88246697e-01 -7.79984817e-02
6.18205488e-01 9.82218683e-01 -6.56847715e-01 -2.13265970e-01
-3.82892251e-01 9.52864170e-01 -2.43224397e-01 1.75758079e-01
-9.04339373e-01 -5.02550304e-01 2.62999266e-01 -3.07512850e-01
2.22308248e-01 -4.23359096e-01 7.03937188e-02 -5.39063096e-01
-3.03589910e-01 -1.79122937e+00 5.80237210e-01 -8.49945486e-01
-1.14169884e+00 1.11981726e+00 -3.87519524e-02 -9.39107656e-01
-5.34116924e-01 -3.75429988e-01 1.13094680e-01 9.82342839e-01
-9.10757482e-01 -1.30310452e+00 5.00852950e-02 4.52105969e-01
1.05181479e+00 -6.17610991e-01 1.51570499e+00 9.74201739e-01
-3.25490445e-01 7.76465654e-01 4.49028343e-01 7.90287480e-02
1.27116847e+00 -1.19719112e+00 5.70514858e-01 4.91677552e-01
4.91785079e-01 -1.38675477e-02 7.48461843e-01 -2.61378497e-01
-1.22566891e+00 -8.82975698e-01 1.31226742e+00 -5.10874927e-01
8.76525342e-01 -8.11140716e-01 -5.90695798e-01 8.94421935e-01
8.04782033e-01 -7.91410148e-01 7.72161305e-01 4.93791364e-02
2.26461086e-02 -1.45018458e-01 -7.62347341e-01 4.70268846e-01
4.77723539e-01 -6.61264718e-01 -4.37524408e-01 4.04486746e-01
5.55607021e-01 -1.10427690e+00 -8.98737431e-01 7.41022378e-02
5.22360086e-01 -3.96021515e-01 6.37551010e-01 -2.32003942e-01
7.14589283e-03 -6.34424761e-02 -7.94580996e-01 -1.66634607e+00
1.97662249e-01 -1.14085782e+00 2.36770272e-01 1.69466639e+00
7.50298977e-01 -3.01016450e-01 -8.51529837e-02 -6.83779120e-01
-5.31715214e-01 -2.90599465e-01 -1.21311533e+00 -5.95797062e-01
7.38148615e-02 -1.03586662e+00 4.96463060e-01 1.05874491e+00
-1.60534699e-02 5.95953047e-01 -6.07806921e-01 2.32327938e-01
4.49372008e-02 -5.30274153e-01 4.56461251e-01 -6.40597939e-01
-3.49471480e-01 -2.77968317e-01 -2.94160575e-01 -1.09980726e+00
-4.46454734e-02 -8.52827966e-01 4.03776258e-01 -1.19446087e+00
-4.06395972e-01 -3.10863525e-01 1.87755466e-01 6.36861920e-01
3.79966110e-01 7.40860045e-01 4.64286543e-02 -5.83000258e-02
-3.99573833e-01 2.89686590e-01 5.76612711e-01 -4.99462068e-01
-5.88982590e-02 1.74848527e-01 -2.91704208e-01 4.84613597e-01
8.41961443e-01 -4.91354465e-01 -1.86152443e-01 -6.90615118e-01
2.89457589e-01 1.39474779e-01 -4.45218265e-01 -6.18875504e-01
-2.84126252e-02 2.24235326e-01 -3.56107444e-01 -8.34598362e-01
7.06377625e-01 -4.88520533e-01 1.26632273e-01 -7.57188872e-02
-8.59294608e-02 3.30149204e-01 7.22434461e-01 -1.39608204e-01
-7.21711755e-01 -3.63395184e-01 7.10545301e-01 1.42874092e-01
-5.43830156e-01 -1.32464200e-01 -1.43772078e+00 4.84429240e-01
5.05096734e-01 3.86236422e-02 -4.41126488e-02 -7.91127145e-01
-1.02813363e+00 1.57580361e-01 -2.70086884e-01 5.81665397e-01
3.18339586e-01 -1.31873369e+00 -1.31067598e+00 7.35586956e-02
1.93508804e-01 -2.58331835e-01 1.68644130e-01 1.05623078e+00
-7.17384160e-01 4.14049029e-01 -1.89914897e-01 -6.15718424e-01
-1.58345079e+00 -4.76806641e-01 2.30077967e-01 -1.17820509e-01
-1.63492933e-01 1.03668964e+00 -5.35550177e-01 -5.97198188e-01
5.62824667e-01 2.29307059e-02 -3.59435856e-01 6.67435348e-01
5.92653990e-01 5.07496059e-01 7.96022773e-01 -1.42479348e+00
-5.17712355e-01 3.86546217e-02 -2.01945767e-01 -1.03243172e+00
1.26539207e+00 -6.96301699e-01 -3.97035569e-01 1.13290155e+00
9.12078977e-01 6.15802109e-01 -4.01785582e-01 1.13200121e-01
3.10592443e-01 -5.13676554e-02 1.41584828e-01 -1.22600198e+00
-4.65731144e-01 7.09618628e-01 3.17452133e-01 4.21528369e-01
1.17858744e+00 -1.29148409e-01 8.76447976e-01 4.18108761e-01
2.46924162e-01 -1.52311850e+00 -3.20331067e-01 1.47984457e+00
1.34555626e+00 -1.01951528e+00 -5.63169777e-01 -2.83565164e-01
-8.14786315e-01 1.03549826e+00 1.42266676e-01 5.86896598e-01
6.61592960e-01 3.88786703e-01 6.73325241e-01 1.35803949e-02
-4.52950269e-01 1.30186811e-01 3.47934961e-01 7.81029820e-01
7.84260392e-01 3.41764987e-02 -2.51067162e-01 1.09375525e+00
-8.96032095e-01 -9.10132408e-01 8.09430778e-01 9.54012692e-01
-2.26974025e-01 -1.28951418e+00 -7.02528298e-01 7.33640417e-02
-7.72684038e-01 -4.44459707e-01 -5.45197189e-01 8.26999903e-01
-2.96155334e-01 1.38987386e+00 1.67310014e-02 -4.27631855e-01
4.29853708e-01 2.54813790e-01 6.00403547e-01 -6.08519077e-01
-7.39407420e-01 5.91627538e-01 7.83956945e-01 -2.54750520e-01
-6.17446840e-01 -9.05744493e-01 -1.00907385e+00 -4.38757986e-01
-3.33220840e-01 5.29531360e-01 7.49634624e-01 9.14211988e-01
-3.39680195e-01 6.47856534e-01 7.33902335e-01 -1.01905286e+00
-1.60502329e-01 -1.51175559e+00 -9.16657388e-01 -2.27806836e-01
2.75145978e-01 -1.26238048e-01 -3.10360312e-01 3.15420955e-01] | [14.393726348876953, 6.792703151702881] |
370a2c44-caf2-4033-8423-08f5a8cdc47f | task-preferences-across-languages-on | 2212.09045 | null | https://arxiv.org/abs/2212.09045v1 | https://arxiv.org/pdf/2212.09045v1.pdf | Task Preferences across Languages on Community Question Answering Platforms | With the steady emergence of community question answering (CQA) platforms like Quora, StackExchange, and WikiHow, users now have an unprecedented access to information on various kind of queries and tasks. Moreover, the rapid proliferation and localization of these platforms spanning geographic and linguistic boundaries offer a unique opportunity to study the task requirements and preferences of users in different socio-linguistic groups. In this study, we implement an entity-embedding model trained on a large longitudinal dataset of multi-lingual and task-oriented question-answer pairs to uncover and quantify the (i) prevalence and distribution of various online tasks across linguistic communities, and (ii) emerging and receding trends in task popularity over time in these communities. Our results show that there exists substantial variance in task preference as well as popularity trends across linguistic communities on the platform. Findings from this study will help Q&A platforms better curate and personalize content for non-English users, while also offering valuable insights to businesses looking to target non-English speaking communities online. | ['Rishabh Mehrotra', 'Prasanta Bhattacharya', 'Sebastin Santy'] | 2022-12-18 | null | null | null | null | ['community-question-answering', 'community-question-answering'] | ['miscellaneous', 'natural-language-processing'] | [-9.25411224e-01 -2.40424857e-01 -2.64105320e-01 -7.52911121e-02
-8.32484901e-01 -7.34496951e-01 5.70457757e-01 7.29201555e-01
-5.95816314e-01 9.50348675e-02 9.11690116e-01 -4.13590223e-01
-2.78074533e-01 -5.66689670e-01 -7.99496099e-03 -7.67443031e-02
-5.71058020e-02 1.40707538e-01 2.54140049e-01 -4.56097960e-01
4.45608586e-01 -2.18592808e-01 -1.40474904e+00 3.10529828e-01
1.47805846e+00 7.04658210e-01 4.34393674e-01 4.53342378e-01
-6.72968864e-01 8.82238328e-01 -1.28852397e-01 -6.85992777e-01
-2.35930514e-02 1.23134218e-02 -8.24113846e-01 -7.60962665e-02
8.37655604e-01 -1.88064426e-01 -4.66377556e-01 8.81829023e-01
5.02317846e-01 8.34606141e-02 4.07948285e-01 -9.84914064e-01
-1.32788002e+00 4.84559000e-01 -4.86779839e-01 6.10204041e-01
6.40152693e-01 7.37073570e-02 1.42872500e+00 -6.81545973e-01
8.44053030e-01 8.25162530e-01 8.68319750e-01 1.06082454e-01
-1.03141725e+00 -5.08887649e-01 2.95093190e-02 3.05573523e-01
-1.41435254e+00 -4.30428147e-01 5.46702921e-01 -9.63433504e-01
6.11622930e-01 1.73764676e-01 5.74071288e-01 7.43266523e-01
-1.61972940e-01 5.92556536e-01 1.00442398e+00 -3.25578183e-01
-1.27743453e-01 5.57947516e-01 3.35481584e-01 6.28110886e-01
5.98198548e-02 -9.64606941e-01 -1.02425921e+00 -6.62694335e-01
-2.60040443e-02 1.59819052e-01 -2.28886604e-01 8.90130252e-02
-9.70088601e-01 9.96773779e-01 2.68342584e-01 6.52842939e-01
-5.33006251e-01 -1.79760560e-01 4.35674459e-01 5.00775576e-01
8.32975686e-01 5.80928802e-01 -4.59725320e-01 -4.67901736e-01
-7.25476801e-01 3.80535632e-01 1.19030380e+00 6.36548579e-01
8.98575068e-01 -6.07475519e-01 -2.05578670e-01 1.25780785e+00
3.87132704e-01 4.17205065e-01 7.38620520e-01 -9.97430801e-01
7.10345328e-01 8.12629700e-01 3.06488723e-01 -1.43097806e+00
-3.11479062e-01 -2.84217954e-01 4.94749472e-02 -8.00724447e-01
8.36881638e-01 -3.20044309e-01 8.58879909e-02 1.68267357e+00
2.94464260e-01 -3.94483268e-01 -6.53266490e-01 5.33453405e-01
5.55827260e-01 3.38136047e-01 3.48968357e-01 4.70108613e-02
1.74560082e+00 -5.79496324e-01 -6.99959517e-01 -1.95729256e-01
8.21976781e-01 -9.57672179e-01 1.54409552e+00 -2.31643587e-01
-7.67319739e-01 -3.84738326e-01 -2.89583087e-01 -4.13322002e-01
-5.94579101e-01 -3.18350196e-01 6.15823090e-01 1.02468479e+00
-1.24395108e+00 1.44863158e-01 -3.75338107e-01 -8.89676929e-01
4.25528109e-01 -3.08803290e-01 -1.04767494e-02 -2.47468263e-01
-1.16760468e+00 6.60695136e-01 -2.46432930e-01 -3.47002983e-01
-1.55368477e-01 -1.01378369e+00 -5.31595528e-01 -7.56601198e-03
1.36328086e-01 -3.35008353e-01 1.19302857e+00 -7.32292414e-01
-7.31214821e-01 8.16383064e-01 -3.49542260e-01 3.73980775e-02
2.36155894e-02 -2.18974143e-01 -6.97257578e-01 -1.55255318e-01
7.13978648e-01 1.55517682e-01 4.00836200e-01 -6.57961547e-01
-9.23760653e-01 -6.11395001e-01 2.42908634e-02 2.34882727e-01
-1.11294770e+00 3.04953694e-01 -1.12112574e-01 -2.71613717e-01
-8.38728026e-02 -4.47949618e-01 2.15146244e-01 6.77805021e-02
-7.65586346e-02 -8.54505658e-01 6.61886275e-01 -1.22222674e+00
1.77637041e+00 -2.29816699e+00 -3.59246045e-01 -2.62299441e-02
6.23132586e-01 -1.52926341e-01 6.69511454e-03 9.60397601e-01
6.79122567e-01 6.33807719e-01 4.99481946e-01 -3.67494583e-01
1.42613396e-01 -2.81723082e-01 1.00766513e-02 4.54387456e-01
-1.45348951e-01 8.53789270e-01 -1.28617048e+00 -6.67059541e-01
-4.36669469e-01 1.62255913e-01 -7.96815276e-01 -2.16168668e-02
-1.90439388e-01 3.47956300e-01 -7.58277774e-01 7.51263559e-01
3.21716517e-01 -7.18881071e-01 2.17890084e-01 3.29272211e-01
-6.18845582e-01 6.86208725e-01 -4.96741146e-01 1.45196080e+00
-6.25980794e-01 1.11974990e+00 5.35357535e-01 -3.60289752e-01
4.90447462e-01 2.66500115e-01 4.64250267e-01 -1.07574356e+00
1.26906938e-03 4.14705724e-01 5.54913096e-03 -1.10467148e+00
9.36131179e-01 3.60421479e-01 -1.22883633e-01 8.11644793e-01
-2.42336139e-01 3.94841999e-01 3.94121975e-01 4.27885950e-01
1.03219581e+00 -4.17186439e-01 -1.24650411e-01 -7.88178861e-01
3.62508476e-01 1.45419598e-01 2.13188395e-01 6.65535390e-01
-6.43884361e-01 -4.69326451e-02 4.11419541e-01 -1.40645295e-01
-1.01035392e+00 -9.00534749e-01 -3.10510814e-01 1.80299199e+00
-7.35675395e-02 -4.49811727e-01 -4.15632188e-01 -1.77647665e-01
5.49844146e-01 5.32611430e-01 -4.13056433e-01 1.59014359e-01
-4.15082276e-01 -4.10414636e-01 2.35551655e-01 -1.01610459e-01
6.12636685e-01 -8.37739944e-01 -7.23615587e-02 2.16110006e-01
-8.04837108e-01 -9.83480334e-01 -1.16513622e+00 -7.34341383e-01
-5.27554274e-01 -1.00836134e+00 -8.49007368e-01 -9.85762060e-01
3.26009005e-01 5.71826696e-01 1.38557935e+00 1.40411332e-01
-3.38513404e-02 8.63380492e-01 -4.97651935e-01 -2.71418512e-01
-1.53590694e-01 6.73775315e-01 5.99926561e-02 3.03438604e-01
6.64690316e-01 -5.06661355e-01 -8.85299444e-01 2.51275897e-01
-6.59470797e-01 -5.33672094e-01 -1.72922432e-01 1.60345957e-01
-1.88487604e-01 -1.60471812e-01 1.09444380e+00 -8.57097745e-01
1.36915028e+00 -1.46852422e+00 -1.90190718e-01 3.34834099e-01
-8.62076402e-01 -4.74643856e-01 1.45027563e-01 -4.50814545e-01
-1.06458235e+00 -6.97765350e-01 1.18468009e-01 4.14251775e-01
3.35697234e-01 9.08950090e-01 6.47510886e-01 9.67704430e-02
8.83826792e-01 -4.99811545e-02 -7.54281506e-02 -7.89489090e-01
4.09190387e-01 1.33877945e+00 1.57746803e-02 -4.22159821e-01
6.78291738e-01 4.58841503e-01 -1.11856914e+00 -1.13186085e+00
-8.13048482e-01 -1.17809439e+00 -6.68113604e-02 -4.65011537e-01
1.15992296e+00 -9.68698502e-01 -1.08952081e+00 4.93911594e-01
-8.24323297e-01 -3.46714318e-01 4.25305963e-02 1.96898609e-01
1.74822003e-01 4.68292564e-01 -8.14839303e-01 -8.90903890e-01
-2.64745116e-01 -7.55217314e-01 6.57114983e-01 4.18070763e-01
-5.54715216e-01 -1.36249471e+00 2.87816226e-01 1.15168452e+00
9.48605597e-01 -4.15991932e-01 9.93712306e-01 -6.61351144e-01
-6.73083961e-01 -1.28899917e-01 -4.58716124e-01 -2.77038589e-02
2.90385604e-01 -1.86361566e-01 -5.65610528e-01 -2.71386415e-01
-1.60073280e-01 -4.77707714e-01 3.33937943e-01 1.71742439e-01
8.14254344e-01 -3.42913359e-01 -2.62027591e-01 -2.17855051e-01
1.19416213e+00 -3.28006595e-01 2.32832301e-02 4.86509353e-01
4.69004810e-01 8.43974650e-01 3.38673852e-02 5.64057529e-01
1.19265175e+00 5.67453563e-01 3.87274064e-02 6.16388023e-01
9.39614847e-02 -6.20462418e-01 3.80791128e-01 1.44985402e+00
2.94663101e-01 -6.93256482e-02 -1.22261894e+00 1.12264192e+00
-1.53874159e+00 -1.00083089e+00 -3.08750123e-01 2.05690765e+00
8.34906816e-01 -3.70285153e-01 4.07323867e-01 -4.98651505e-01
7.38742709e-01 3.44267249e-01 -4.02724922e-01 -1.95394516e-01
6.55834973e-02 -7.46223405e-02 4.96463299e-01 4.08889621e-01
-4.73123878e-01 4.14841384e-01 6.42077065e+00 6.74701989e-01
-1.04195762e+00 6.37319326e-01 5.84776998e-01 1.15091637e-01
-7.86858559e-01 -1.30322039e-01 -6.69969559e-01 6.73992753e-01
1.10230744e+00 -5.34531534e-01 6.46966815e-01 5.94802499e-01
2.36082152e-01 -9.47791636e-02 -5.46053588e-01 6.88007355e-01
7.36843571e-02 -1.33489788e+00 -5.80708563e-01 1.69821009e-01
8.67534995e-01 6.09219313e-01 2.68412381e-01 6.32042825e-01
3.34178865e-01 -4.41902369e-01 6.87215745e-01 3.45380783e-01
5.05461693e-01 -1.08011134e-01 1.84476227e-01 6.96666479e-01
-9.80081022e-01 -5.23216069e-01 -4.33486849e-02 -2.65750229e-01
2.99357861e-01 4.58298743e-01 -6.16580963e-01 1.92987025e-02
1.05083275e+00 5.47319293e-01 -7.74006784e-01 9.38410401e-01
3.76939744e-01 8.17310214e-01 -8.81788209e-02 -3.42644513e-01
7.11747855e-02 -3.44038010e-01 2.37836838e-01 8.85081232e-01
3.19650382e-01 -2.71268457e-01 -8.99223387e-02 7.80825198e-01
-4.71428543e-01 5.77503324e-01 -4.35059220e-01 -7.89900124e-01
9.49376643e-01 1.24156392e+00 -4.48919356e-01 1.38437986e-01
-1.08113587e+00 7.78576314e-01 7.86227167e-01 4.28759068e-01
-3.66191715e-01 -2.99475610e-01 7.74910986e-01 6.95571780e-01
-1.69371307e-01 -6.41217530e-01 -7.42207319e-02 -1.27270484e+00
1.52155384e-01 -7.83990204e-01 5.10716319e-01 -5.82256198e-01
-1.84631848e+00 1.28284514e-01 -5.20952821e-01 -4.26415294e-01
1.44544402e-02 -2.37882242e-01 -4.28775430e-01 1.03391719e+00
-1.71857822e+00 -7.79174387e-01 -3.11275303e-01 5.52435338e-01
3.80958438e-01 -1.58370957e-01 4.81613129e-01 8.99209559e-01
-5.26313961e-01 4.89352375e-01 5.03676116e-01 8.72096866e-02
8.33757460e-01 -9.03320193e-01 1.97903901e-01 1.58553928e-01
1.20874353e-01 9.27694738e-01 2.17378944e-01 -7.37131715e-01
-1.40139270e+00 -8.08658183e-01 1.69893289e+00 -1.02371109e+00
1.31546962e+00 -5.63235641e-01 -8.32416296e-01 7.48140216e-01
5.14721692e-01 -5.48476458e-01 1.11703551e+00 1.00414360e+00
-4.88781810e-01 -9.99189727e-03 -1.00440872e+00 6.28067255e-01
8.56930315e-01 -1.40701151e+00 -3.60823840e-01 7.15756953e-01
7.87731826e-01 2.57581413e-01 -1.19499660e+00 -4.22322303e-01
5.76730371e-01 -9.04199958e-01 5.85739076e-01 -5.70853353e-01
3.77821445e-01 2.04578400e-01 -1.48478091e-01 -8.60519826e-01
-5.05346358e-01 -7.68846571e-01 2.06523642e-01 1.52295685e+00
7.25791514e-01 -9.92054582e-01 7.35349059e-01 1.10931969e+00
1.64467767e-02 -5.81720293e-01 -1.07819724e+00 -2.43611842e-01
2.79609263e-01 -6.84395358e-02 3.09800088e-01 1.46009266e+00
2.93316662e-01 1.86728656e-01 1.51114002e-01 -1.20352268e-01
1.77138314e-01 -2.53984313e-02 4.22990650e-01 -1.28328526e+00
-1.09253921e-01 -5.78593254e-01 1.92446802e-02 -9.42126811e-01
1.14330783e-01 -1.18249011e+00 -5.31454027e-01 -1.53749895e+00
4.99423653e-01 -8.62791777e-01 -4.19438854e-02 -5.21354116e-02
-3.20901573e-01 4.79566194e-02 3.98479849e-02 6.41114473e-01
-1.01221776e+00 3.49164426e-01 1.22783458e+00 2.09755242e-01
-2.67068297e-01 -8.46223459e-02 -1.05827987e+00 3.24605316e-01
6.20127439e-01 -3.06045353e-01 -2.81784534e-01 -8.72941852e-01
1.19478464e+00 -7.66363516e-02 -5.23414537e-02 -4.94704694e-01
3.04389000e-01 -1.68942399e-02 -2.85174996e-01 -1.68316141e-01
-5.31450845e-02 -6.69973612e-01 -2.41688177e-01 2.68231928e-01
-4.99853432e-01 4.49104756e-01 -1.14029258e-01 5.59957922e-01
-1.83168814e-01 -2.56630182e-01 1.32957697e-01 8.91221911e-02
-4.19242650e-01 4.42659318e-01 -8.04300427e-01 7.18474925e-01
6.51642859e-01 -1.43535942e-01 -4.39596534e-01 -8.73908579e-01
-4.50580478e-01 6.08405769e-01 3.49130183e-01 7.54410684e-01
-1.89773887e-01 -1.33851361e+00 -8.81537318e-01 -1.28826901e-01
5.21208465e-01 -5.35499215e-01 5.63497603e-01 8.76000762e-01
-4.26972032e-01 5.29614508e-01 1.28702506e-01 -2.53697217e-01
-5.97169638e-01 8.91232938e-02 2.09522709e-01 -8.75974074e-02
4.23237905e-02 8.63652885e-01 -2.85575241e-01 -7.61927247e-01
-1.34142980e-01 1.81337863e-01 -2.63315856e-01 7.04259157e-01
3.20666522e-01 7.71475494e-01 -2.54913330e-01 -5.66511512e-01
-2.10796725e-02 -2.88061388e-02 -1.09932609e-01 6.76381355e-03
1.05325305e+00 -6.28165603e-01 -4.94919270e-01 7.16947436e-01
1.32222724e+00 4.70893502e-01 -6.96866393e-01 -4.97614264e-01
5.88890493e-01 -6.76696181e-01 -1.15009755e-01 -3.53496253e-01
-8.17403913e-01 6.02078080e-01 4.01162297e-01 9.55937028e-01
5.50473213e-01 5.02028048e-01 1.09148264e+00 -4.86342832e-02
3.28564286e-01 -1.19399774e+00 2.16915458e-01 4.16123092e-01
6.30384564e-01 -1.12583339e+00 -4.11885321e-01 -1.76425725e-01
-2.90297478e-01 5.49884617e-01 4.56337333e-01 4.17437047e-01
1.20988226e+00 -4.53151256e-01 2.24709883e-01 -4.06225413e-01
-6.09957993e-01 -1.29934162e-01 1.45620838e-01 4.68832314e-01
9.60880876e-01 9.81960297e-02 -6.73927426e-01 5.16684294e-01
-2.48472631e-01 -1.89932674e-01 3.96306992e-01 6.16158843e-01
-5.38865983e-01 -1.12428164e+00 5.51209003e-02 9.66627121e-01
-9.63150024e-01 -2.51838475e-01 -1.75584137e-01 5.62900543e-01
-5.39776012e-02 1.29088151e+00 2.91322380e-01 -2.43040636e-01
1.67583212e-01 4.80556041e-01 -1.35251746e-01 -7.36601293e-01
-9.08790171e-01 -5.80200911e-01 4.97081965e-01 -2.18043223e-01
-1.02736562e-01 -1.24283910e+00 -7.03222930e-01 -5.36368430e-01
-4.91438210e-01 2.73208022e-01 7.12697148e-01 6.06475711e-01
9.21534896e-01 -9.71886963e-02 6.05014920e-01 -6.50065839e-02
-5.91358006e-01 -1.09580123e+00 -7.00715065e-01 7.13711083e-01
2.18694910e-01 -2.60854989e-01 -3.83694053e-01 -2.00605378e-01] | [11.390453338623047, 7.923440456390381] |
807001c5-5606-4d07-9ece-1d6148aea6f7 | aligning-coordinated-text-streams-through | 1609.08237 | null | http://arxiv.org/abs/1609.08237v1 | http://arxiv.org/pdf/1609.08237v1.pdf | Aligning Coordinated Text Streams through Burst Information Network Construction and Decipherment | Aligning coordinated text streams from multiple sources and multiple
languages has opened many new research venues on cross-lingual knowledge
discovery. In this paper we aim to advance state-of-the-art by: (1). extending
coarse-grained topic-level knowledge mining to fine-grained information units
such as entities and events; (2). following a novel
Data-to-Network-to-Knowledge (D2N2K) paradigm to construct and utilize network
structures to capture and propagate reliable evidence. We introduce a novel
Burst Information Network (BINet) representation that can display the most
important information and illustrate the connections among bursty entities,
events and keywords in the corpus. We propose an effective approach to
construct and decipher BINets, incorporating novel criteria based on
multi-dimensional clues from pronunciation, translation, burst, neighbor and
graph topological structure. The experimental results on Chinese and English
coordinated text streams show that our approach can accurately decipher the
nodes with high confidence in the BINets and that the algorithm can be
efficiently run in parallel, which makes it possible to apply it to huge
amounts of streaming data for never-ending language and information
decipherment. | ['Zhifang Sui', 'Qing Dou', 'Heng Ji', 'Tao Ge', 'Ming Zhou', 'Lei Cui', 'Baobao Chang', 'Xiaoman Pan'] | 2016-09-27 | null | null | null | null | ['decipherment'] | ['natural-language-processing'] | [-2.13456675e-01 -1.52764887e-01 -5.46408296e-01 -1.40494585e-01
-7.07117498e-01 -6.28659129e-01 6.46055639e-01 8.37128580e-01
-3.59569401e-01 7.31400192e-01 4.97325540e-01 -3.74813199e-01
-6.36267245e-01 -9.71020579e-01 -5.24881840e-01 -2.75358528e-01
-7.34025419e-01 8.97106230e-01 7.30215073e-01 -3.72068286e-01
3.73511434e-01 3.89165699e-01 -1.58506060e+00 6.50438607e-01
9.11605954e-01 9.85922694e-01 1.82828724e-01 6.74879074e-01
-9.30775642e-01 8.02774191e-01 -9.21078265e-01 -5.23203075e-01
-9.78890210e-02 -4.09762859e-01 -9.65006471e-01 -2.22234204e-01
-1.89994901e-01 3.15144688e-01 -1.03684491e-03 9.65887010e-01
5.05120158e-01 -2.13100716e-01 4.08631355e-01 -1.27313268e+00
-2.37772003e-01 1.48373842e+00 -6.69554710e-01 7.56975770e-01
5.20958960e-01 -5.22647142e-01 1.24834239e+00 -7.50471592e-01
1.13812029e+00 1.23072958e+00 6.62562609e-01 -3.02886218e-01
-6.69675171e-01 -6.58244252e-01 2.44834661e-01 8.78381908e-01
-1.62178159e+00 -3.09664935e-01 4.98129845e-01 -1.93926945e-01
1.03169632e+00 2.85821050e-01 6.98561966e-01 8.85932326e-01
-2.03095213e-01 7.15714753e-01 7.39292622e-01 -7.21744597e-01
6.70991093e-02 1.07096337e-01 3.68224472e-01 6.49123907e-01
4.88778889e-01 -3.71067911e-01 -1.47067058e+00 -3.47241133e-01
3.53187561e-01 -2.96356827e-01 -1.65297821e-01 4.44744825e-01
-1.47995508e+00 7.75113463e-01 -1.58400789e-01 7.49003112e-01
-6.80306911e-01 -3.13214630e-01 7.11983919e-01 4.42818165e-01
6.30431652e-01 1.01798140e-01 -7.12195635e-01 -6.62102342e-01
-1.07663798e+00 3.47897969e-02 1.49527383e+00 1.34388971e+00
6.26706183e-01 -2.33375486e-02 3.61162648e-02 7.53822505e-01
1.43249452e-01 2.97075570e-01 6.67531431e-01 -2.68774360e-01
8.57026935e-01 5.75263679e-01 -2.10389733e-01 -1.37012827e+00
-5.04911780e-01 -2.90303320e-01 -7.21880674e-01 -7.28125930e-01
1.96434870e-01 -1.83527291e-01 -4.27373648e-01 1.19563484e+00
6.05911136e-01 3.93130660e-01 6.15581870e-02 5.19297063e-01
9.21012938e-01 1.00458956e+00 -2.26801038e-01 -6.71117544e-01
1.66061175e+00 -4.64156270e-01 -1.07655621e+00 4.28359836e-01
6.15474105e-01 -8.91619325e-01 6.47453547e-01 6.37559831e-01
-9.28945124e-01 -2.38890842e-01 -8.66982937e-01 4.14224058e-01
-7.29642451e-01 -3.83903682e-01 4.26187664e-01 4.89528567e-01
-9.18452680e-01 5.42219281e-01 -6.17631912e-01 -4.40287620e-01
6.76141530e-02 1.39054507e-01 -1.56791806e-01 3.06239743e-02
-1.68234491e+00 6.49951458e-01 1.02396858e+00 -1.04456224e-01
-4.60803837e-01 -6.96285069e-01 -6.51296675e-01 2.37890035e-02
6.00866377e-01 -2.20174506e-01 8.79778564e-01 -5.77869892e-01
-1.31165981e+00 4.77939069e-01 -3.88400257e-01 -5.36359549e-01
-8.41926411e-02 1.02889821e-01 -9.78480697e-01 4.03436512e-01
3.15611541e-01 -3.75348255e-02 5.01023233e-01 -8.71562243e-01
-1.18187594e+00 -2.53069133e-01 -6.08739913e-01 1.96387962e-01
-6.78546011e-01 6.34227157e-01 -8.02294314e-01 -9.64491308e-01
1.52645424e-01 -4.39218402e-01 2.42831573e-01 -7.61659861e-01
-8.01068306e-01 -8.17943752e-01 7.89484978e-01 -8.02361906e-01
1.92984939e+00 -1.83872807e+00 9.33745280e-02 8.11658442e-01
3.74181658e-01 -1.46830026e-02 1.39987513e-01 9.96978998e-01
3.43588442e-01 3.39838713e-01 1.99990481e-01 -1.33564308e-01
5.03222868e-02 4.54567939e-01 -2.42302760e-01 3.13806891e-01
-2.98926711e-01 8.02821934e-01 -9.70710456e-01 -9.87979114e-01
-3.63212794e-01 9.05703902e-02 -6.68979064e-02 -8.40666890e-02
-1.95143297e-01 -1.24741040e-01 -5.04540443e-01 8.84975195e-01
2.96374232e-01 -3.73887986e-01 5.88978648e-01 -2.08054379e-01
-4.01696831e-01 6.29655600e-01 -1.67090070e+00 1.37692904e+00
-1.50043508e-02 8.26266408e-01 1.36301816e-01 -1.11965191e+00
1.04752254e+00 4.78833526e-01 4.97625321e-01 -7.02559710e-01
-1.67645261e-01 3.12263250e-01 -2.75641561e-01 -4.85965163e-01
7.65567005e-01 2.01412231e-01 -2.86446184e-01 6.17663860e-01
1.46710187e-01 2.53145456e-01 5.79876244e-01 5.35635948e-01
1.01562345e+00 -6.41091526e-01 4.43166614e-01 -4.03387338e-01
2.61274368e-01 3.67648304e-01 7.41601288e-01 6.27914250e-01
1.75882295e-01 -1.25207394e-01 6.53754771e-01 -2.68045276e-01
-8.77325237e-01 -7.15453327e-01 -1.47779852e-01 1.33463180e+00
1.20723464e-01 -1.04214394e+00 -4.63661432e-01 -6.07635736e-01
1.03577711e-01 5.44927418e-01 -2.97631443e-01 4.66471702e-01
-7.31393874e-01 -7.94979870e-01 9.69177663e-01 2.69181311e-01
2.88627803e-01 -9.49159384e-01 7.30115399e-02 6.48952365e-01
-8.07198763e-01 -1.23051310e+00 -3.55702370e-01 3.15781027e-01
-6.20667994e-01 -1.10474741e+00 -1.45173416e-01 -9.24146056e-01
1.42604619e-01 1.86141297e-01 1.35840344e+00 -6.58598021e-02
-2.48567119e-01 1.87906697e-01 -9.86367702e-01 -6.44577146e-01
-4.17184114e-01 2.09219962e-01 2.88779467e-01 4.09243330e-02
6.85840666e-01 -7.17005312e-01 -2.66858071e-01 5.76143622e-01
-8.13506067e-01 -2.96521217e-01 2.22881824e-01 4.01780397e-01
7.48962879e-01 7.69609809e-01 8.44058394e-01 -9.27118957e-01
1.02208662e+00 -1.01877248e+00 -4.92243439e-01 2.30187476e-01
-8.76406193e-01 7.68286362e-02 5.08794427e-01 -3.60788465e-01
-7.60868192e-01 -5.02970517e-01 1.68988541e-01 -2.76976097e-02
-1.24696633e-02 1.25796485e+00 1.68494686e-01 1.64676085e-01
5.11683285e-01 3.79986554e-01 -4.55855429e-01 -6.97230816e-01
5.97716212e-01 1.04253662e+00 5.95797181e-01 -8.14645946e-01
4.90062505e-01 5.21255434e-01 -2.94438571e-01 -1.13494837e+00
-4.79553223e-01 -1.10767651e+00 -7.26811588e-01 -2.70319462e-01
5.59422970e-01 -8.64853024e-01 -7.18333840e-01 2.73164481e-01
-1.32541823e+00 7.45825022e-02 -2.21280441e-01 5.81171572e-01
-8.50159824e-02 5.74923754e-01 -7.09362209e-01 -5.29356003e-01
-4.09238070e-01 -3.62567931e-01 6.73242748e-01 -7.46085271e-02
-2.85216153e-01 -1.12468624e+00 2.70917058e-01 -1.21022156e-03
1.38264358e-01 2.75064319e-01 8.73520136e-01 -1.22691655e+00
-6.02332652e-01 -1.29723744e-02 -2.35331118e-01 -2.00955674e-01
1.64621286e-02 2.13577285e-01 -4.08536732e-01 5.80571406e-02
-2.51889825e-01 -4.18552496e-02 6.60905600e-01 1.74569473e-01
9.09225345e-01 -7.38119960e-01 -7.02452779e-01 3.46200734e-01
1.12360716e+00 2.23616362e-01 3.98764342e-01 5.07392585e-01
5.72890997e-01 7.78093576e-01 5.64148724e-01 9.00090218e-01
8.31048012e-01 6.33105993e-01 7.33437389e-02 2.17779875e-01
-8.03073645e-02 -4.41556036e-01 3.44173759e-01 1.96755123e+00
-9.41022113e-02 -5.43332696e-01 -9.50690746e-01 1.02027512e+00
-2.00963569e+00 -1.22431374e+00 -3.98770988e-01 1.79495847e+00
1.35505593e+00 4.70676422e-01 2.69590318e-01 3.74275148e-01
1.06238508e+00 -2.63555534e-02 -8.04741606e-02 -3.64264250e-01
-4.87812400e-01 8.41337442e-02 5.10650814e-01 2.82617062e-01
-6.57794654e-01 1.02607703e+00 6.04004002e+00 1.43271148e+00
-8.44454587e-01 2.69853473e-01 2.00734422e-01 6.59043789e-02
-4.74106312e-01 -1.91076174e-01 -1.51229274e+00 6.29655182e-01
1.39197302e+00 -5.08445561e-01 2.64341176e-01 5.20968795e-01
2.19132572e-01 -5.75706027e-02 -8.70102286e-01 1.01493847e+00
2.10404992e-01 -1.86028242e+00 7.40388930e-02 -5.81958815e-02
6.75417602e-01 2.72053987e-01 -5.96556962e-01 1.36740562e-02
6.26012146e-01 -5.72059691e-01 7.55102217e-01 3.66743147e-01
7.84056306e-01 -8.36581588e-01 5.63730001e-01 6.12807870e-01
-1.83832538e+00 1.74844265e-01 -2.07246944e-01 3.03246439e-01
3.98321122e-01 9.82784748e-01 -1.46205413e+00 1.05309093e+00
8.83942127e-01 8.94990981e-01 -1.85845077e-01 9.73724782e-01
-1.39570192e-01 9.48425949e-01 -7.98180819e-01 -6.46712065e-01
3.62816691e-01 1.00123741e-01 7.49047875e-01 1.97880757e+00
4.34264064e-01 1.32753372e-01 1.66739911e-01 3.84318769e-01
-2.47031361e-01 5.74815154e-01 -4.74932492e-01 -5.60997613e-02
1.16367829e+00 1.17479157e+00 -1.19878745e+00 -6.38578415e-01
-3.60631138e-01 6.07774079e-01 2.82351285e-01 2.80220479e-01
-4.14447337e-01 -7.71134913e-01 3.38517189e-01 1.78301800e-02
7.39203572e-01 -4.09371138e-01 -1.00903496e-01 -1.04682708e+00
-8.28803703e-03 -9.29907203e-01 1.00666356e+00 -2.27980837e-01
-1.34944355e+00 7.81977475e-01 2.05392987e-01 -1.02697551e+00
-4.93504018e-01 -2.84049630e-01 -4.67979372e-01 6.20339215e-01
-1.62638497e+00 -7.88237572e-01 -7.48333260e-02 1.22310877e+00
5.10174811e-01 -3.14615101e-01 7.27516472e-01 6.06226921e-01
-3.79610747e-01 6.07965887e-01 3.31955791e-01 2.33693525e-01
6.43471420e-01 -1.27778506e+00 3.74401927e-01 8.30862224e-01
7.39650607e-01 4.16171193e-01 4.24672067e-01 -9.86782849e-01
-1.52061701e+00 -9.74164367e-01 1.50306475e+00 -2.28218600e-01
1.10629344e+00 -4.09706295e-01 -7.95505762e-01 3.01503986e-01
2.97425866e-01 -2.86113262e-01 7.80058444e-01 5.68626583e-01
-4.87766743e-01 -3.27030838e-01 -6.95150197e-01 2.34885857e-01
1.04272234e+00 -5.52426338e-01 -7.42748380e-01 5.24309814e-01
1.06800222e+00 -1.90516353e-01 -1.14994717e+00 1.83196902e-01
1.28355101e-01 -4.83934313e-01 7.44745970e-01 -5.31566143e-01
-1.68704823e-01 -3.21072906e-01 -1.02039352e-01 -1.23333490e+00
-5.04719839e-02 -1.27941740e+00 -5.51048517e-01 1.47984505e+00
6.78251505e-01 -3.85831416e-01 5.27686179e-01 -3.72924447e-01
-1.91646576e-01 -5.44055462e-01 -1.21748841e+00 -8.21540415e-01
-5.16823173e-01 -9.06385005e-01 8.11365664e-01 1.31023943e+00
7.23164201e-01 2.49638274e-01 -4.41332489e-01 2.95176029e-01
5.24705172e-01 3.59125763e-01 4.40317035e-01 -1.49788380e+00
-3.08670774e-02 -4.44495648e-01 -3.62243772e-01 -1.09927273e+00
-8.22155550e-02 -1.03446054e+00 -1.04795247e-01 -1.43735814e+00
-5.62118441e-02 -7.06413865e-01 -2.00851440e-01 4.12085891e-01
1.51829928e-01 -6.16587661e-02 -1.75961331e-01 4.92172033e-01
-1.03845501e+00 1.49276763e-01 7.76043057e-01 7.02158213e-02
-4.18611348e-01 -5.47182048e-03 -7.25792587e-01 5.19755423e-01
5.65338254e-01 -7.08578169e-01 -9.64947715e-02 -1.39464989e-01
6.76978350e-01 1.14167243e-01 -1.58455178e-01 -4.24738735e-01
1.00211406e+00 -7.61307962e-03 -4.82671000e-02 -1.20322561e+00
-1.34635210e-01 -6.01123035e-01 -5.18347435e-02 1.92663476e-01
-4.24784087e-02 3.78873706e-01 1.96968675e-01 9.84781563e-01
-6.33329451e-01 -1.48126557e-01 1.56947538e-01 4.69124727e-02
-1.02000976e+00 3.86542976e-01 -4.60226983e-01 4.94557917e-01
7.26571441e-01 -1.19118296e-01 -4.19098735e-01 -4.09365892e-01
-5.68194985e-01 3.11656505e-01 -3.58590931e-01 4.05115694e-01
6.73542380e-01 -1.34333742e+00 -9.93816435e-01 -1.30531946e-02
1.43757775e-01 -2.08552718e-01 -2.19442889e-01 6.31642997e-01
-4.99870688e-01 6.13824427e-01 2.21912995e-01 -6.36932611e-01
-1.33108509e+00 3.82385075e-01 -3.19508135e-01 -4.17228758e-01
-6.22711837e-01 8.49906802e-01 -6.84885621e-01 -9.23180059e-02
5.90855300e-01 -4.54562843e-01 -5.22179127e-01 7.81206727e-01
6.30785167e-01 6.38533652e-01 2.94913769e-01 -4.22240764e-01
-4.36059058e-01 4.61584151e-01 -5.36845773e-02 -1.34750456e-01
1.43331003e+00 -4.45784837e-01 -4.95321989e-01 5.57964087e-01
8.82024109e-01 3.33783239e-01 -4.25421983e-01 -8.32904756e-01
6.80469811e-01 -1.20806314e-01 -9.98687837e-03 -6.59077048e-01
-6.51688755e-01 3.56970429e-01 -4.83074002e-02 1.04798877e+00
9.66415644e-01 3.28617305e-01 9.19748366e-01 5.44117451e-01
3.50854546e-01 -1.46174109e+00 -3.33110653e-02 7.72177696e-01
5.54952085e-01 -7.92310119e-01 -7.78774992e-02 -6.03263140e-01
-4.66581166e-01 1.39353025e+00 -2.03304682e-02 5.36734462e-01
1.09508574e+00 5.58293283e-01 -1.77428782e-01 -5.85404575e-01
-8.73781204e-01 -3.99890274e-01 3.79993200e-01 5.14289737e-01
-4.28930297e-03 1.94715962e-01 -4.45202738e-01 8.23933125e-01
-3.94637614e-01 -3.30597848e-01 2.77730465e-01 8.67278278e-01
-8.39364290e-01 -1.24531543e+00 -3.83017093e-01 4.11145151e-01
-5.96673489e-01 -4.99375582e-01 -2.13965356e-01 8.26700330e-01
2.81910479e-01 1.08612919e+00 6.33808300e-02 -4.64074641e-01
3.33469868e-01 4.60634083e-02 -1.52628943e-01 -5.63645005e-01
-5.81827223e-01 2.20095679e-01 6.58792973e-01 -4.91277754e-01
-5.26081264e-01 -7.12164640e-01 -1.24192774e+00 -7.06971228e-01
-4.67943907e-01 6.93028331e-01 8.52841675e-01 9.17924047e-01
6.50503993e-01 3.92034888e-01 6.03205264e-01 -3.63202363e-01
-1.03882819e-01 -1.01679027e+00 -8.42837274e-01 7.33442456e-02
6.29718155e-02 -4.44782257e-01 -2.72826463e-01 3.44559699e-01] | [9.987642288208008, 9.006909370422363] |
da80cb9f-1421-408f-844f-0a7d79b59701 | dual-attention-networks-for-few-shot-fine | null | null | https://ojs.aaai.org/index.php/AAAI/article/view/20196 | https://ojs.aaai.org/index.php/AAAI/article/view/20196/19955 | Dual Attention Networks for Few-Shot Fine-Grained Recognition | The task of few-shot fine-grained recognition is to classify images belonging to subordinate categories merely depending on few examples. Due to the fine-grained nature, it is desirable to capture subtle but discriminative part-level patterns from limited training data, which makes it a challenging problem. In this paper, to generate fine-grained tailored representations for few-shot recognition, we propose a Dual Attention Network (Dual Att-Net) consisting of two dual branches of both hard- and soft-attentions. Specifically, by producing attention guidance from deep activations of input images, our hard-attention is realized by keeping a few useful deep descriptors and forming them as a bag of multi-instance learning. Since these deep descriptors could correspond to objects' parts, the advantage of modeling as a multi-instance bag is able to exploit inherent correlation of these fine-grained parts. On the other side, a soft attended activation representation can be obtained by applying attention guidance upon original activations, which brings comprehensive attention information as the counterpart of hard-attention. After that, both outputs of dual branches are aggregated as a holistic image embedding w.r.t. input images. By performing meta-learning, we can learn a powerful image embedding in such a metric space to generalize to novel classes. Experiments on three popular fine-grained benchmark datasets show that our Dual Att-Net obviously outperforms other existing state-of-the-art methods. | ['Jianhua Wang', 'Xiu-Shen Wei', 'Faen Zhang', 'Shu-Lin Xu'] | 2022-06-28 | null | null | null | proceedings-of-the-aaai-conference-on-4 | ['hard-attention'] | ['methodology'] | [ 2.58632302e-01 -5.89244999e-02 -2.23751187e-01 -6.09728098e-01
-7.51144767e-01 -1.99776009e-01 7.41559744e-01 6.56975582e-02
-1.35553911e-01 6.37802839e-01 2.52375752e-01 3.81527156e-01
-3.44984502e-01 -1.10958433e+00 -7.59378254e-01 -9.26214814e-01
3.92164081e-01 1.73325449e-01 2.21382737e-01 -2.22296938e-01
2.95147777e-01 4.95511293e-01 -1.85281277e+00 6.15790188e-01
7.26526678e-01 1.51100338e+00 2.79452503e-01 1.13601118e-01
-3.96886170e-01 5.87606549e-01 -6.96642280e-01 -2.77038634e-01
1.34972274e-01 -4.01020914e-01 -4.36105877e-01 2.49030799e-01
5.53195834e-01 -2.49568790e-01 -3.37001354e-01 1.00729823e+00
3.41375768e-01 5.00582039e-01 8.65769386e-01 -9.55797255e-01
-1.32895827e+00 1.97548464e-01 -4.67407972e-01 3.39047343e-01
6.88755363e-02 3.89586866e-01 1.10482109e+00 -1.03292012e+00
1.83016509e-01 1.30102670e+00 2.65436023e-01 6.75558984e-01
-1.22607267e+00 -6.14901304e-01 5.67636311e-01 2.95526773e-01
-1.41359556e+00 -3.75228256e-01 8.03538501e-01 -4.26861316e-01
9.87793803e-01 1.98500112e-01 3.89872342e-01 1.13193178e+00
9.14467052e-02 5.79103351e-01 1.01745880e+00 -2.51202404e-01
2.41907418e-01 4.01032232e-02 3.75319451e-01 6.37647986e-01
1.86018169e-01 9.58196670e-02 -2.91082472e-01 1.40013725e-01
5.61053872e-01 8.10971797e-01 -2.68864751e-01 -2.41869479e-01
-1.11168230e+00 7.87534714e-01 1.00554073e+00 6.02549553e-01
-5.10223329e-01 1.06548190e-01 3.27410549e-01 2.77382493e-01
6.08540773e-01 4.37694281e-01 -3.59940141e-01 2.24595502e-01
-6.83091342e-01 -4.81542014e-02 3.15995634e-01 7.54725277e-01
1.25960457e+00 1.05185322e-01 -7.78197587e-01 1.01720166e+00
2.53589991e-02 1.45785689e-01 1.02021229e+00 -3.61931831e-01
3.13593179e-01 9.64193761e-01 -2.62199789e-01 -9.77394462e-01
8.54361877e-02 -7.23358989e-01 -1.13522351e+00 1.23872891e-01
-1.83703564e-02 3.57266754e-01 -1.25902236e+00 1.67354143e+00
3.89962047e-02 3.55777115e-01 -2.60535955e-01 9.38474059e-01
1.01507723e+00 7.35779226e-01 2.23782048e-01 -8.38834196e-02
1.40350056e+00 -1.15485084e+00 -3.49429101e-01 -2.89997399e-01
3.21902126e-01 -2.22814739e-01 1.22687054e+00 5.40142469e-02
-6.72986448e-01 -1.12499177e+00 -1.03999782e+00 -1.01357475e-01
-8.50941837e-01 -8.68820995e-02 4.93010938e-01 2.71823943e-01
-6.17701650e-01 6.27193987e-01 -4.13851827e-01 -1.20526105e-01
7.45543361e-01 1.76213056e-01 -4.42432791e-01 -3.91297638e-01
-1.11233318e+00 6.27956390e-01 4.40821022e-01 -4.97077890e-02
-9.34468150e-01 -7.30498493e-01 -9.79619503e-01 5.77071071e-01
2.43038923e-01 -7.06732810e-01 7.82171071e-01 -8.05686831e-01
-1.26462674e+00 8.58197272e-01 -1.11062288e-01 -3.33768763e-02
1.01477010e-02 1.24534242e-01 -4.28122461e-01 5.29925339e-02
3.76435101e-01 4.42321777e-01 1.22180045e+00 -1.08625412e+00
-5.78567564e-01 -5.52580476e-01 3.40453684e-01 -7.70853162e-02
-7.50829458e-01 -4.73192245e-01 -2.22189680e-01 -8.71600866e-01
-1.21271662e-01 -4.29805249e-01 -1.65441036e-01 -7.06470534e-02
-1.70428485e-01 -5.24407685e-01 7.70750105e-01 -9.61764678e-02
1.32430387e+00 -2.33214021e+00 2.29736745e-01 -1.69490173e-01
3.81357223e-01 3.21699709e-01 -4.37068850e-01 3.02168846e-01
-2.28490353e-01 1.13606177e-01 -5.36345467e-02 -1.86336875e-01
2.48011798e-01 2.43871048e-01 -5.13036728e-01 1.73594996e-01
6.71849668e-01 1.22853327e+00 -9.74763036e-01 -2.15007842e-01
3.99791121e-01 3.04971844e-01 -3.18336874e-01 4.68248099e-01
-1.77116916e-01 1.69157788e-01 -7.17437804e-01 8.05939555e-01
5.20877719e-01 -4.00516063e-01 -3.72095495e-01 -5.20642996e-01
8.95281658e-02 -2.35673934e-01 -7.08935976e-01 1.67387915e+00
-7.12207854e-01 1.82346076e-01 -4.30155158e-01 -1.52180743e+00
1.22202539e+00 -1.56487804e-02 -2.81565692e-02 -1.02165806e+00
1.73262075e-01 1.50270224e-01 -1.24851264e-01 -4.01960701e-01
1.33576259e-01 -6.42055154e-01 -3.81369561e-01 4.42876875e-01
4.57324624e-01 1.50562361e-01 3.98906581e-02 -7.57321045e-02
1.05813861e+00 -1.44395962e-01 4.85333741e-01 3.56267877e-02
5.40273547e-01 -4.03375179e-01 4.98083174e-01 9.37149942e-01
-1.93229795e-01 6.32128716e-01 3.21426950e-02 -6.91084266e-01
-7.52532661e-01 -9.84154165e-01 -1.09702185e-01 1.54557943e+00
2.47455373e-01 -2.21136108e-01 -3.71936649e-01 -9.32107329e-01
1.52133673e-01 3.97812873e-01 -1.19752777e+00 -7.50888169e-01
-2.53168315e-01 -6.49290025e-01 1.47222921e-01 8.26277614e-01
4.04652625e-01 -1.34664655e+00 -4.80118364e-01 3.63410860e-01
2.74184078e-01 -7.09997833e-01 -4.93366987e-01 4.88867104e-01
-6.91102445e-01 -8.80203485e-01 -1.04377925e+00 -8.07969868e-01
5.95755160e-01 7.10259616e-01 1.09378397e+00 1.05031125e-01
-5.34225225e-01 1.25481963e-01 -5.61524570e-01 -1.33643031e-01
1.36491686e-01 -2.97326315e-02 -6.36924431e-02 6.35279477e-01
6.79676831e-01 -6.26823425e-01 -6.42403543e-01 1.50538027e-01
-1.00358748e+00 -2.91689515e-01 1.02305496e+00 1.41851616e+00
8.03621352e-01 -3.12918201e-02 8.37970138e-01 -7.63375282e-01
7.03314900e-01 -6.32281303e-01 -2.30833650e-01 4.48272586e-01
-1.92734852e-01 2.58725494e-01 1.11465931e+00 -5.43024659e-01
-1.01316059e+00 -2.27641135e-01 4.12316322e-02 -9.09512758e-01
-4.87305731e-01 4.33331877e-01 -3.55559140e-01 -1.11226194e-01
5.83275735e-01 5.96581101e-01 -3.07573438e-01 -6.48605108e-01
5.66111505e-01 7.86763191e-01 3.81553501e-01 -6.43241942e-01
6.75715625e-01 3.95488471e-01 -1.68401718e-01 -7.23179221e-01
-1.30896389e+00 -5.45081794e-01 -6.71288848e-01 1.46509573e-01
8.60053182e-01 -7.10588932e-01 -4.76262569e-01 3.34054023e-01
-8.33112776e-01 -5.08303940e-02 -5.99518836e-01 1.35876954e-01
-3.71503115e-01 1.32905677e-01 -3.34193647e-01 -5.08152783e-01
-2.71539390e-01 -9.81515765e-01 1.44125164e+00 4.37462866e-01
8.28638077e-02 -8.60217810e-01 5.56568019e-02 1.34637803e-01
3.80327582e-01 2.47386813e-01 8.31988275e-01 -5.75942099e-01
-5.52617848e-01 -4.30663437e-01 -5.77919126e-01 4.25917983e-01
4.29852843e-01 -3.50631088e-01 -1.18151140e+00 -4.50725794e-01
3.10482234e-02 -6.24984920e-01 1.41479599e+00 1.44571081e-01
1.63525343e+00 -4.89663780e-01 -3.82151544e-01 7.78158844e-01
1.57298064e+00 8.54580179e-02 5.22640049e-01 2.48412296e-01
7.83758342e-01 4.38438892e-01 7.08863497e-01 4.97591257e-01
2.17067879e-02 6.13440990e-01 4.54057842e-01 9.95726213e-02
-1.95556358e-01 -2.40806848e-01 -1.30032673e-02 6.49661958e-01
-1.42067745e-01 9.36967805e-02 -3.30027521e-01 5.41998088e-01
-1.76653135e+00 -1.32008719e+00 5.47775626e-01 1.98637092e+00
5.89793980e-01 1.31085530e-01 -1.71171948e-01 7.92088802e-04
8.17852378e-01 5.57881355e-01 -8.16831589e-01 -4.12357926e-01
-3.88181880e-02 4.96978015e-01 -1.28431439e-01 -1.47008687e-01
-1.09436309e+00 7.31627524e-01 4.39833832e+00 1.34937298e+00
-1.33469808e+00 2.18781278e-01 7.44558036e-01 -5.11815585e-02
-3.10107529e-01 -2.55690843e-01 -8.33809316e-01 6.63893580e-01
6.04437947e-01 -1.88755020e-01 3.01005036e-01 8.85429382e-01
-3.33421379e-01 4.06612217e-01 -1.07425499e+00 1.06449938e+00
2.06956923e-01 -1.42042732e+00 4.93949294e-01 -4.40584496e-03
8.91299665e-01 -2.37199813e-01 1.52154416e-01 7.94850051e-01
-6.19671345e-02 -1.12006962e+00 5.53546607e-01 8.19316626e-01
1.12248921e+00 -6.22590244e-01 6.70613170e-01 4.74517673e-01
-1.49713778e+00 -4.91029412e-01 -9.13818061e-01 -1.84121907e-01
-1.37614906e-01 5.36543250e-01 -2.17311203e-01 6.13239110e-01
6.55206084e-01 1.02249670e+00 -5.44909835e-01 8.85899723e-01
-1.22217894e-01 7.55993053e-02 2.36414179e-01 -5.48420474e-02
6.05198681e-01 -6.52982434e-03 1.34652659e-01 1.23971272e+00
3.14117759e-01 4.24576581e-01 2.07316756e-01 9.97472167e-01
-2.95513660e-01 -2.28729561e-01 -6.35289192e-01 -1.26072004e-01
2.93661565e-01 1.52896583e+00 -4.42833126e-01 -6.87105238e-01
-5.22869110e-01 1.10678864e+00 8.25380325e-01 3.08244258e-01
-5.80783069e-01 -8.22710514e-01 7.65912354e-01 -1.70331135e-01
7.68025279e-01 2.62538910e-01 -7.01310858e-02 -1.41993225e+00
3.07514872e-02 -6.73280954e-01 5.24206281e-01 -5.88339269e-01
-1.87683654e+00 9.75919425e-01 -3.71193379e-01 -1.46479774e+00
8.94843694e-03 -7.50219643e-01 -8.79312158e-01 1.04158187e+00
-1.55582666e+00 -1.43889570e+00 -7.90929496e-01 7.49097943e-01
7.50747859e-01 -2.19720721e-01 1.07897401e+00 2.24702999e-01
-5.73755324e-01 8.91126156e-01 1.13963597e-01 -1.57637335e-02
5.83388269e-01 -1.19638014e+00 1.55137151e-01 5.12831092e-01
1.65551066e-01 7.45563984e-01 1.23229876e-01 -2.98133761e-01
-1.17236280e+00 -1.47699833e+00 4.72571135e-01 -2.77211934e-01
6.87553406e-01 -2.18940675e-01 -1.32335711e+00 3.81898761e-01
-1.09282121e-01 6.29027605e-01 6.42620385e-01 2.43333410e-02
-8.10661614e-01 -5.47428966e-01 -9.58645225e-01 1.13419168e-01
1.01352167e+00 -7.53214300e-01 -1.03536856e+00 1.04466684e-01
7.99626529e-01 1.66408658e-01 -8.57941329e-01 3.59776765e-01
5.52054286e-01 -9.68451560e-01 1.03942370e+00 -9.23631191e-01
6.25704646e-01 -1.99585930e-01 -3.57978910e-01 -1.60812283e+00
-9.17218268e-01 -3.20624397e-03 -3.31888497e-01 1.12863934e+00
-1.01915337e-01 -6.68022871e-01 5.37046194e-01 1.68591768e-01
-4.70997721e-01 -1.05107403e+00 -7.69066274e-01 -9.25658345e-01
3.90225388e-02 1.73096787e-02 8.33161414e-01 9.54538465e-01
-1.49906367e-01 5.66129625e-01 -3.11883241e-01 -1.03375085e-01
7.35528111e-01 7.60903478e-01 4.83700454e-01 -1.23835814e+00
-4.61595356e-01 -7.22849131e-01 -7.10771739e-01 -1.05378067e+00
2.57574290e-01 -8.77883911e-01 5.47707640e-02 -1.51341069e+00
5.34560084e-01 -4.17960584e-01 -9.84212458e-01 6.92230523e-01
-4.02363986e-01 6.43624246e-01 2.37226605e-01 2.34136239e-01
-7.52683401e-01 9.18789387e-01 1.39829957e+00 -7.22163618e-01
1.77922010e-01 -1.04167923e-01 -1.03950191e+00 3.86341989e-01
5.37753582e-01 -2.42660940e-01 -2.27921188e-01 -2.90492505e-01
-2.79937893e-01 -1.25974119e-01 5.61389387e-01 -9.53801632e-01
1.68857962e-01 -2.60764688e-01 6.34113789e-01 -3.80039543e-01
5.09881198e-01 -5.74990511e-01 -1.02853537e-01 3.80882084e-01
-2.89770603e-01 -5.85429728e-01 4.24406715e-02 7.12971389e-01
-6.44607723e-01 -1.96930364e-01 9.35593247e-01 -5.00651896e-01
-1.19473779e+00 8.00820172e-01 2.24706799e-01 1.03378538e-02
1.04510081e+00 -4.67150301e-01 -4.72597510e-01 5.81144355e-02
-7.70857692e-01 -9.33454856e-02 2.58395135e-01 5.68362832e-01
8.57134461e-01 -1.75192451e+00 -5.23448527e-01 4.79768991e-01
7.11693287e-01 -9.53262970e-02 7.72593141e-01 4.43685949e-01
3.08175623e-01 5.84987998e-01 -7.21654594e-01 -4.78637069e-01
-7.67413676e-01 1.10300016e+00 2.23341540e-01 -8.92865732e-02
-4.70478952e-01 1.00890684e+00 1.01402402e+00 -1.05144344e-01
-1.16460165e-02 -2.16372237e-01 -3.17262620e-01 3.35453719e-01
8.62893462e-01 1.07554272e-01 7.35712200e-02 -6.67233646e-01
-3.94216627e-01 9.03405488e-01 -2.98562944e-01 6.47040606e-01
1.60950112e+00 -9.79687795e-02 -1.03768846e-02 4.84409094e-01
1.59886098e+00 -3.16450030e-01 -1.33930480e+00 -4.21005726e-01
-3.16597521e-01 -7.00021207e-01 -5.21282144e-02 -7.06421196e-01
-1.11359620e+00 1.33836818e+00 4.76188958e-01 4.28926706e-01
1.33541119e+00 2.71600574e-01 6.70531929e-01 3.36516857e-01
3.64795387e-01 -7.02507496e-01 4.46704984e-01 5.23451865e-01
1.00655222e+00 -1.36986828e+00 -3.75788659e-01 8.20239037e-02
-2.98898637e-01 1.22069120e+00 8.89970601e-01 -4.03658509e-01
5.23649633e-01 -3.21728706e-01 -2.40913004e-01 -2.47472107e-01
-7.53266811e-01 -4.24683988e-01 7.31472969e-01 5.88837206e-01
2.51009971e-01 1.80822596e-01 -5.41182049e-02 1.10630751e+00
2.79757887e-01 -3.27385850e-02 8.75345618e-02 6.42331362e-01
-8.03116977e-01 -7.91387200e-01 -1.11641116e-01 9.06148672e-01
-2.14929357e-01 -1.73092663e-01 -1.99759886e-01 4.15112913e-01
3.84303302e-01 5.21780431e-01 1.60787806e-01 -5.80987334e-01
5.33754826e-01 -2.21537694e-01 4.28518653e-01 -1.11761343e+00
-4.05770481e-01 -3.00743878e-01 -4.17820424e-01 -7.19210148e-01
-2.39569023e-01 -8.42074975e-02 -7.38494158e-01 -5.31097874e-02
-2.24250510e-01 4.23911139e-02 1.16737656e-01 1.07127941e+00
4.79873240e-01 9.41601336e-01 8.64998698e-01 -1.24245870e+00
-8.80075336e-01 -1.02408874e+00 -8.25241685e-01 7.27131069e-01
4.76866037e-01 -1.03080261e+00 -4.78770047e-01 -1.70205772e-01] | [9.720487594604492, 2.1973583698272705] |
1b88914a-95ec-4bd1-9ca6-9e9edbecec4c | generalizing-monocular-3d-human-pose | 1904.05512 | null | http://arxiv.org/abs/1904.05512v1 | http://arxiv.org/pdf/1904.05512v1.pdf | Generalizing Monocular 3D Human Pose Estimation in the Wild | The availability of the large-scale labeled 3D poses in the Human3.6M dataset
plays an important role in advancing the algorithms for 3D human pose
estimation from a still image. We observe that recent innovation in this area
mainly focuses on new techniques that explicitly address the generalization
issue when using this dataset, because this database is constructed in a highly
controlled environment with limited human subjects and background variations.
Despite such efforts, we can show that the results of the current methods are
still error-prone especially when tested against the images taken in-the-wild.
In this paper, we aim to tackle this problem from a different perspective. We
propose a principled approach to generate high quality 3D pose ground truth
given any in-the-wild image with a person inside. We achieve this by first
devising a novel stereo inspired neural network to directly map any 2D pose to
high quality 3D counterpart. We then perform a carefully designed geometric
searching scheme to further refine the joints. Based on this scheme, we build a
large-scale dataset with 400,000 in-the-wild images and their corresponding 3D
pose ground truth. This enables the training of a high quality neural network
model, without specialized training scheme and auxiliary loss function, which
performs favorably against the state-of-the-art 3D pose estimation methods. We
also evaluate the generalization ability of our model both quantitatively and
qualitatively. Results show that our approach convincingly outperforms the
previous methods. We make our dataset and code publicly available. | ['Jimmy S. Ren', 'Zhenhua Guo', 'Hongsheng Li', 'Yan Chen', 'Mude Lin', 'Luyang Wang', 'Keyuan Qian'] | 2019-04-11 | null | null | null | null | ['monocular-3d-human-pose-estimation'] | ['computer-vision'] | [ 1.41280577e-01 6.57867566e-02 1.40528426e-01 -3.41645122e-01
-8.14854145e-01 -2.93904096e-01 5.03672183e-01 -3.50806206e-01
-6.84102118e-01 6.06372595e-01 1.37174964e-01 2.26298094e-01
3.34792100e-02 -6.23623133e-01 -9.25126076e-01 -4.49938565e-01
-3.95624600e-02 9.88896132e-01 2.05086112e-01 -4.70568895e-01
-2.31695268e-02 5.39630830e-01 -1.57977426e+00 -2.18885183e-01
5.93037069e-01 9.12645280e-01 3.18218172e-02 4.56201762e-01
5.38212597e-01 2.22441927e-01 -5.78327000e-01 -5.13615489e-01
6.87566161e-01 -3.20923567e-01 -7.77223706e-01 4.37988877e-01
8.03961277e-01 -3.91652137e-01 -3.32367092e-01 8.81758511e-01
7.93370962e-01 1.82209834e-01 5.04948735e-01 -1.08734477e+00
-1.98998213e-01 -1.43624246e-01 -4.39125329e-01 -2.59702414e-01
7.11049676e-01 2.40883946e-01 6.77747607e-01 -8.52367461e-01
9.52034354e-01 1.21155941e+00 8.34100664e-01 6.46324992e-01
-1.16342413e+00 -3.83691370e-01 -7.42332637e-02 -4.81651956e-03
-1.50049233e+00 -1.90018147e-01 9.86699164e-01 -4.09851283e-01
5.97041547e-01 1.23571575e-01 9.65957999e-01 1.56311715e+00
8.36618617e-02 7.21451283e-01 1.16159236e+00 -5.88381827e-01
8.83615464e-02 -1.82943434e-01 -2.25110456e-01 7.98010826e-01
2.17693284e-01 2.33955115e-01 -4.41992074e-01 1.50315508e-01
1.02450907e+00 5.74890524e-03 -3.39896172e-01 -8.88210118e-01
-1.38237727e+00 5.64631402e-01 7.14777231e-01 1.68478623e-01
-4.16126698e-01 1.29124045e-01 1.66662112e-01 -8.49870790e-04
5.18063426e-01 3.79355103e-01 -4.04073983e-01 -1.89492269e-03
-9.04931724e-01 8.09754372e-01 5.90741336e-01 8.51738989e-01
5.69111824e-01 -2.77593583e-01 -1.69783905e-02 6.57013178e-01
2.27308482e-01 5.51427305e-01 2.26311460e-01 -1.20220983e+00
5.95315993e-01 5.62535405e-01 2.86451519e-01 -1.10727334e+00
-6.06302857e-01 -7.12743878e-01 -8.52575481e-01 3.65425795e-01
7.88373530e-01 1.79069936e-02 -9.24495935e-01 1.74737954e+00
6.40088260e-01 -2.37879768e-01 -2.83121675e-01 1.31136453e+00
5.10527372e-01 3.26528609e-01 -3.56694818e-01 1.63961977e-01
1.27181697e+00 -9.40889180e-01 -3.58617723e-01 -2.18394399e-01
3.89301300e-01 -4.72855955e-01 1.10033059e+00 4.67700809e-01
-1.00776100e+00 -8.09556782e-01 -1.06179440e+00 -5.68034314e-02
-2.01790661e-01 1.57854006e-01 3.95790219e-01 5.72694480e-01
-8.64653826e-01 7.45395184e-01 -8.46409917e-01 -6.95476234e-01
1.68180808e-01 3.02634627e-01 -8.37457299e-01 5.51288910e-02
-1.20388627e+00 1.07769072e+00 4.25030321e-01 4.31402326e-01
-7.05754340e-01 -3.59647781e-01 -9.68807757e-01 -4.13655549e-01
8.05542767e-01 -1.03468060e+00 1.17031837e+00 -5.89291871e-01
-1.49079466e+00 1.31147885e+00 1.70552909e-01 -3.74894738e-01
1.09790337e+00 -6.53812170e-01 6.02100044e-02 1.57881662e-01
7.14968070e-02 7.64933228e-01 6.34865522e-01 -1.49656355e+00
-2.86940008e-01 -7.60157228e-01 1.16578251e-01 2.39050373e-01
-7.52898827e-02 -2.95141578e-01 -9.26101446e-01 -7.62450337e-01
2.15620622e-01 -1.20554757e+00 -3.80030662e-01 2.25515395e-01
-5.48990428e-01 -1.98186114e-02 2.73205906e-01 -6.82078004e-01
6.41645610e-01 -1.61275470e+00 6.04890943e-01 1.51379898e-01
1.91258758e-01 2.92071730e-01 3.28389890e-02 3.70672703e-01
8.49601030e-02 -2.69010156e-01 -2.28754610e-01 -6.67660177e-01
1.15548410e-02 1.03352703e-01 1.35163605e-01 7.31268883e-01
7.08200643e-03 9.83211756e-01 -6.60464108e-01 -5.84523797e-01
3.52834046e-01 6.16491675e-01 -7.04871893e-01 5.06045580e-01
-7.29772970e-02 9.16696846e-01 -3.62080395e-01 6.00575328e-01
5.02409160e-01 -2.24169821e-01 2.98676938e-02 -3.57512683e-01
9.47774947e-02 -2.12214515e-01 -1.31997120e+00 2.39301229e+00
-2.88373917e-01 1.88095152e-01 -2.03195848e-02 -1.07448769e+00
8.59497845e-01 1.74580798e-01 5.19881666e-01 -5.46954691e-01
4.12515759e-01 3.37091386e-01 -2.24462897e-01 -3.74259323e-01
3.80247176e-01 -2.28748709e-01 -1.76514730e-01 1.15142606e-01
2.62169600e-01 -1.72313124e-01 -1.93134695e-02 -1.53395951e-01
6.88739002e-01 6.86346352e-01 1.19087145e-01 -1.82193846e-01
5.43057978e-01 -3.68930511e-02 3.43329102e-01 5.85663617e-01
-2.95732677e-01 1.03805649e+00 1.17527261e-01 -6.01260781e-01
-1.31843483e+00 -1.17123854e+00 -2.95262458e-03 6.27579093e-01
1.99139699e-01 -3.04687053e-01 -1.00191414e+00 -6.75101280e-01
-2.90900692e-02 9.52099487e-02 -7.56942868e-01 -6.18984364e-03
-8.70117903e-01 -5.64962268e-01 4.47978824e-01 5.74495196e-01
7.64733374e-01 -8.27597260e-01 -7.52946377e-01 -3.05624008e-02
-3.31435263e-01 -1.27238047e+00 -3.21361065e-01 -8.88034925e-02
-7.93715000e-01 -1.19025862e+00 -1.28322077e+00 -6.63015902e-01
5.90741217e-01 -1.10029504e-01 1.16666818e+00 -1.15044691e-01
-3.05172294e-01 3.12674761e-01 -3.15630168e-01 -1.19376384e-01
-1.94590658e-01 7.26823807e-02 3.16180557e-01 -7.74788186e-02
6.05445988e-02 -6.45816088e-01 -6.68720007e-01 5.15160322e-01
-6.09880865e-01 1.03506476e-01 5.68145156e-01 8.59943509e-01
5.78924358e-01 -5.93774393e-02 1.50968537e-01 -5.78931272e-01
2.70584643e-01 1.92573369e-01 -5.62279463e-01 -4.65827323e-02
-3.41441363e-01 3.04276198e-02 4.46600586e-01 -2.71773905e-01
-8.54616463e-01 4.15146559e-01 -5.22285998e-01 -4.07272756e-01
-4.45565552e-01 1.83565378e-01 -3.85003746e-01 -1.40227795e-01
7.53052831e-01 1.79595813e-01 2.54004717e-01 -7.43580878e-01
2.71614015e-01 4.13085699e-01 9.43454444e-01 -7.60914266e-01
1.12800729e+00 5.87570429e-01 2.28656456e-01 -7.37348318e-01
-9.15577173e-01 -4.09784257e-01 -1.03153062e+00 -4.81178463e-01
9.36777592e-01 -9.65858221e-01 -8.33794892e-01 7.27440417e-01
-1.11052382e+00 -2.49418408e-01 -8.16808864e-02 5.46781421e-01
-9.28199351e-01 5.10309160e-01 -5.76153517e-01 -7.59523690e-01
-2.37629279e-01 -1.19326389e+00 1.53328657e+00 -1.03008971e-01
-2.69758642e-01 -6.59762800e-01 1.97437868e-01 7.62880147e-01
1.86460409e-02 7.82284856e-01 2.90172935e-01 -3.53635669e-01
-4.64340955e-01 -3.67272735e-01 8.57057422e-02 3.58991385e-01
-1.90091804e-01 -4.74996388e-01 -8.05827558e-01 -4.23959315e-01
-1.36521822e-02 -5.24276197e-01 6.81551099e-01 3.84062529e-01
1.01414120e+00 1.29848331e-01 -1.81099311e-01 7.25847185e-01
1.25779390e+00 -4.10727859e-01 5.16433895e-01 5.93923151e-01
8.08519185e-01 9.59523678e-01 7.39052296e-01 3.13788533e-01
3.84031415e-01 1.19942904e+00 4.74569947e-01 -2.35072702e-01
-1.18342519e-01 -5.60018122e-01 1.32595494e-01 4.48997319e-01
-7.64245093e-01 -1.41911998e-01 -8.76361907e-01 2.97284871e-01
-1.86581028e+00 -7.45028317e-01 1.17362842e-01 2.41439986e+00
7.11727738e-01 3.02769005e-01 4.82411057e-01 3.50776643e-01
6.25090718e-01 1.58785388e-01 -2.33457878e-01 3.66647571e-01
7.01250061e-02 1.61727786e-01 3.34271640e-01 3.18801671e-01
-1.22391796e+00 7.76595771e-01 5.88595438e+00 5.85490406e-01
-9.34406638e-01 -1.62384883e-01 3.74242872e-01 -1.38925672e-01
1.85269251e-01 -3.69246036e-01 -7.66153038e-01 1.84632644e-01
5.31205654e-01 3.80793840e-01 1.18429087e-01 8.69230866e-01
2.58669704e-01 -3.23598906e-02 -1.20567071e+00 1.20460200e+00
1.91293955e-01 -1.04472399e+00 -7.90200308e-02 2.41574079e-01
5.67580938e-01 -2.90024042e-01 -4.89553064e-02 6.28270432e-02
-2.17976823e-01 -9.86545444e-01 9.96202886e-01 6.33645654e-01
6.76161051e-01 -7.36057460e-01 8.21465909e-01 6.34024858e-01
-9.87327516e-01 3.74833196e-01 -1.63813308e-01 -1.98320508e-01
3.62182826e-01 5.04305124e-01 -5.03506601e-01 7.79825568e-01
8.81858349e-01 5.43720722e-01 -6.76476419e-01 9.41554964e-01
-2.38320678e-01 7.38352388e-02 -3.27955306e-01 3.99337187e-02
2.81937029e-02 -1.32831007e-01 5.75449407e-01 8.59654784e-01
3.24780852e-01 -2.56446540e-01 3.00898522e-01 8.66710424e-01
4.41965126e-02 -7.22068325e-02 -5.52832842e-01 4.15808380e-01
-9.46135148e-02 1.10227501e+00 -7.26382911e-01 -1.45576432e-01
-1.31826997e-01 1.28069460e+00 3.41479361e-01 2.49902323e-01
-8.92324507e-01 -1.22757249e-01 4.08854365e-01 2.84707844e-01
2.30877668e-01 -5.53120017e-01 7.43039995e-02 -1.39007568e+00
5.55763900e-01 -9.21663463e-01 1.36488512e-01 -7.57296324e-01
-1.29284811e+00 7.29025364e-01 2.73802280e-01 -1.26614928e+00
-5.36625803e-01 -9.11523461e-01 -6.09304830e-02 6.56247973e-01
-1.08482051e+00 -1.24720395e+00 -4.84805942e-01 6.40186250e-01
3.81232381e-01 3.96426544e-02 6.78408742e-01 3.63647074e-01
-3.40281516e-01 5.97704887e-01 -1.96348965e-01 2.94269532e-01
9.00083542e-01 -1.20970798e+00 4.91424501e-01 8.02333057e-01
1.72813833e-01 6.31689489e-01 9.42132652e-01 -5.26750982e-01
-1.42853725e+00 -7.98340857e-01 7.96321809e-01 -8.39549780e-01
2.65133053e-01 -7.85531700e-01 -5.65595448e-01 5.69880724e-01
-1.90989614e-01 4.95396778e-02 2.24682316e-01 1.66403919e-01
-9.42973942e-02 -1.64826289e-02 -1.11134148e+00 5.36988199e-01
1.55278063e+00 -2.85893470e-01 -9.42352414e-01 2.44556770e-01
6.70081854e-01 -7.06617296e-01 -8.98350358e-01 7.24573851e-01
7.94238865e-01 -1.17670178e+00 1.27726960e+00 -4.89226192e-01
4.66585547e-01 -3.79261643e-01 -2.77818143e-01 -1.12465632e+00
-2.19852962e-02 -4.73681837e-01 2.00326759e-02 7.96984792e-01
-6.51697256e-03 -4.39958721e-01 1.27844083e+00 3.00262958e-01
-6.48260722e-03 -7.89047718e-01 -9.45274472e-01 -1.03534782e+00
1.00566104e-01 -4.52658474e-01 2.68297464e-01 4.23511654e-01
-4.49918538e-01 2.41455927e-01 -7.69324422e-01 9.35159400e-02
9.99769688e-01 1.47320613e-01 1.29527533e+00 -1.28325820e+00
-4.04074818e-01 -9.49318781e-02 -7.24990189e-01 -1.36544538e+00
1.44837216e-01 -4.79190201e-01 1.46083102e-01 -1.18994224e+00
1.77637890e-01 -1.19788684e-01 1.65774018e-01 1.79785058e-01
-1.50199756e-01 6.96711183e-01 2.81208932e-01 2.01638877e-01
-5.29494166e-01 7.31565475e-01 1.49552166e+00 8.68363827e-02
1.11748539e-01 1.04062706e-01 -2.43228674e-01 8.43220294e-01
4.91594970e-01 -1.99516177e-01 -1.58071548e-01 -2.12187663e-01
1.35056851e-02 -8.78021866e-02 9.62387323e-01 -1.40557826e+00
-3.48851234e-02 1.65447086e-01 6.50400758e-01 -7.77750254e-01
6.59462810e-01 -9.08659101e-01 2.92673737e-01 6.30691171e-01
-2.22247764e-01 -5.90932630e-02 -9.01132822e-02 5.43798268e-01
-1.57062128e-01 9.17690247e-02 7.32171178e-01 -4.27536458e-01
-6.94717765e-01 4.66453880e-01 2.17884570e-01 2.21729338e-01
8.29050422e-01 -4.05644000e-01 1.93996385e-01 -4.34711903e-01
-7.79302299e-01 7.11709308e-03 7.89088190e-01 4.87685502e-01
4.62469578e-01 -1.50679827e+00 -7.39190996e-01 2.28224635e-01
3.29560727e-01 2.34639153e-01 1.91286042e-01 7.43744373e-01
-6.95726514e-01 4.96005148e-01 -4.11950469e-01 -8.44844639e-01
-1.04939640e+00 5.24992943e-01 5.02707481e-01 -3.51145864e-01
-7.39941835e-01 6.00698948e-01 1.11382805e-01 -8.84064019e-01
3.44590664e-01 -2.19595894e-01 7.76392221e-02 -3.26769531e-01
3.00690234e-01 4.18842465e-01 1.05436556e-01 -9.58503485e-01
-3.48747492e-01 9.98456597e-01 4.07055825e-01 -2.67292261e-01
1.40126026e+00 3.24892625e-02 2.17611641e-01 3.30705136e-01
1.41558182e+00 -2.32420005e-02 -1.42561293e+00 -1.78310975e-01
-3.83957773e-01 -6.09967232e-01 -3.15074384e-01 -6.49786472e-01
-1.03481627e+00 8.22721183e-01 7.09885716e-01 -1.72180891e-01
9.82844830e-01 1.35480031e-01 8.81838560e-01 3.61912996e-01
7.82515407e-01 -1.06045842e+00 2.74245799e-01 3.62124324e-01
1.08638382e+00 -1.48034883e+00 2.57652309e-02 -5.28399467e-01
-3.04997176e-01 1.00455225e+00 6.28524721e-01 -4.26154554e-01
2.68623978e-01 -1.40373483e-01 1.61307067e-01 -2.01649025e-01
-6.77474737e-02 -3.32504898e-01 6.38462245e-01 6.88479602e-01
3.68810296e-01 -6.81931451e-02 -3.21956038e-01 3.62173349e-01
-4.55201238e-01 9.25765261e-02 1.28201887e-01 8.87866616e-01
-2.56080985e-01 -1.32254696e+00 -7.27199495e-01 -1.55013323e-01
-2.62099981e-01 4.23583388e-01 -4.31536078e-01 1.15412021e+00
1.60214216e-01 6.12665772e-01 -3.35609853e-01 -5.32867432e-01
7.32062280e-01 -5.21448813e-03 9.39356148e-01 -3.87910694e-01
-3.09830070e-01 1.13790140e-01 9.46306288e-02 -9.27878082e-01
-6.75683141e-01 -5.07527411e-01 -9.44773614e-01 -2.17232928e-01
-8.50950927e-02 -1.20970480e-01 5.04481673e-01 8.96392941e-01
2.35628918e-01 2.74584174e-01 2.63622910e-01 -1.54548824e+00
-7.08983898e-01 -8.37670684e-01 -6.11955702e-01 8.53574038e-01
2.02135921e-01 -1.10234070e+00 -1.92720652e-01 -4.10570428e-02] | [7.03851318359375, -0.9320592284202576] |
5947f096-c567-4a59-b57c-34a1abc4c54a | a-data-driven-approach-to-improve-3d-head | null | null | https://link.springer.com/chapter/10.1007/978-3-030-90439-5_43 | https://link.springer.com/content/pdf/10.1007%2F978-3-030-90439-5.pdf | A Data-Driven Approach to Improve 3D Head-Pose Estimation | Head-pose estimation from images is an important research topic in computer vision. Its many applications include detecting focus of attention, tracking driver behavior, and human-computer interaction. Recent research on head-pose estimation has focused on developing models based on deep convolutional neural networks (CNNs). These models are trained using transfer-learning and image augmentation to achieve better initiation states and robustness against occlusions. However, methods that use transfer-learning networks are usually aimed at general image recognition and offer no in-depth study of transfer learning from more task-related networks. Additionally, for the head-pose estimation, robustness against heavy occlusion, and noise such as motion blur and low-brightness are vital. In this paper, we propose a new image-augmentation approach that significantly improves the estimation accuracy of the head-pose model. We also propose a task-related weight initialization to further improve the estimation accuracy by studying internal activations of models trained for face-related tasks such as face-recognition. We test our head-pose estimation model on three challenging test sets and achieve better results to state-of-the-art methods. | ['Eraldo Ribeiro', 'Nima Aghli'] | 2022-01-01 | null | null | null | isvc-2022-1 | ['head-pose-estimation', 'image-augmentation'] | ['computer-vision', 'computer-vision'] | [-1.82068899e-01 1.08223990e-01 -1.36614606e-01 -8.90737772e-01
-5.77762842e-01 1.51842713e-01 3.16271901e-01 -3.76756698e-01
-6.12521768e-01 4.01412904e-01 1.92451045e-01 -6.50200620e-02
2.60942131e-01 -2.89751083e-01 -7.71025419e-01 -5.81907988e-01
2.26121023e-01 1.64976642e-01 2.82974154e-01 -6.59283111e-03
1.56598449e-01 6.88248754e-01 -1.89489722e+00 -3.47186923e-02
5.85858643e-01 1.01747668e+00 2.24522967e-02 3.39260608e-01
3.30439597e-01 5.70656300e-01 -7.02337086e-01 -1.89717188e-01
-5.95824197e-02 -2.92496234e-01 -3.78339708e-01 1.56259865e-01
8.55190933e-01 -4.38817173e-01 -6.40718699e-01 1.09234905e+00
1.02764225e+00 1.50606841e-01 4.34141904e-01 -1.50510859e+00
-4.04288143e-01 -3.72121707e-02 -7.40561545e-01 3.06989819e-01
1.79512441e-01 4.19830650e-01 1.68112904e-01 -9.86485183e-01
1.69351399e-01 1.33377016e+00 6.32453740e-01 9.23090577e-01
-6.30795658e-01 -1.08502638e+00 8.20229948e-02 9.59879518e-01
-1.42832601e+00 -8.18993747e-01 8.95456254e-01 -1.46929681e-01
6.79161787e-01 2.07167566e-01 3.58966619e-01 9.73781884e-01
1.17212765e-01 1.07855165e+00 1.05189335e+00 -3.40059340e-01
-7.31147900e-02 7.24716187e-02 9.72449034e-02 5.79717159e-01
2.43276715e-01 1.82225034e-01 -6.75939500e-01 2.16469541e-01
5.26286721e-01 -3.97146828e-02 -5.03522694e-01 -3.80601794e-01
-6.97639942e-01 6.96389616e-01 7.60458767e-01 1.22078285e-01
-3.13069761e-01 2.02406500e-03 3.65800977e-01 -3.30619156e-01
5.47714055e-01 1.40793487e-01 -2.53367960e-01 2.27828890e-01
-1.01860750e+00 8.25319365e-02 5.37416935e-01 9.01841938e-01
6.43070042e-01 1.71347290e-01 -5.14984667e-01 8.91952932e-01
5.35992444e-01 4.68548745e-01 7.33483255e-01 -5.36541164e-01
2.60494381e-01 4.48776305e-01 -2.54248232e-01 -7.12146699e-01
-8.59797835e-01 -6.71584964e-01 -6.04316771e-01 3.28788817e-01
2.90789038e-01 -2.47267112e-01 -1.03351653e+00 1.78591490e+00
4.74681705e-01 2.99754322e-01 -3.44334155e-01 1.16346455e+00
1.29215884e+00 2.83354729e-01 1.96879715e-01 -1.13721006e-01
1.58833313e+00 -1.15207398e+00 -9.62951124e-01 -7.65150011e-01
4.29928482e-01 -5.63224852e-01 6.93993449e-01 1.01512484e-01
-1.01079369e+00 -8.36132348e-01 -9.55447733e-01 -1.66960284e-01
-2.37302125e-01 4.92144883e-01 3.25110376e-01 8.05128932e-01
-1.11904979e+00 -1.19099189e-02 -7.37640798e-01 -5.07261217e-01
5.56963444e-01 5.36447287e-01 -3.93735677e-01 -5.29793128e-02
-1.04419339e+00 1.36642432e+00 5.70447277e-03 3.83099169e-01
-8.32653463e-01 -3.89592886e-01 -1.19557762e+00 6.51851147e-02
2.27944314e-01 -7.26303160e-01 1.33750415e+00 -7.72605300e-01
-1.65841973e+00 1.17900014e+00 -5.81792355e-01 -4.03554320e-01
2.12933391e-01 -3.85121703e-01 -4.62967098e-01 -1.60641864e-01
-3.57248932e-02 9.53762114e-01 1.16523159e+00 -1.00870430e+00
-3.69341552e-01 -9.48160470e-01 -4.66971636e-01 1.67714179e-01
-5.98093867e-01 3.91463548e-01 -7.59803236e-01 -2.01312974e-01
-3.97985838e-02 -1.16365123e+00 1.66005537e-01 6.21815808e-02
-2.55914599e-01 -4.60280120e-01 1.06219804e+00 -6.74180686e-01
1.06718385e+00 -2.11700177e+00 -1.86995208e-01 -1.21954851e-01
2.31733605e-01 7.60941565e-01 -1.70500904e-01 -3.40676069e-01
-2.70168632e-01 -5.62392592e-01 4.84353192e-02 -5.94107032e-01
-1.55088663e-01 -4.23034169e-02 2.19971314e-01 8.30279350e-01
2.35216185e-01 1.15069723e+00 -3.53644162e-01 -4.41389918e-01
3.60369116e-01 7.27330327e-01 -5.23353577e-01 5.74890614e-01
3.55571806e-01 7.33416677e-01 -9.08827633e-02 4.52696919e-01
9.62484598e-01 1.79458857e-01 -2.52771407e-01 -5.11714518e-01
1.77732669e-02 4.32499200e-02 -7.63696074e-01 1.42502069e+00
-5.36048412e-01 9.78093982e-01 3.41029525e-01 -1.03531504e+00
8.78818631e-01 1.69353783e-01 8.14480484e-02 -8.76827419e-01
7.75331616e-01 -1.20218992e-01 3.94437850e-01 -8.54230285e-01
2.94357032e-01 1.06336325e-01 3.49056065e-01 1.56628430e-01
6.94686547e-02 1.86652169e-02 -2.31596038e-01 -4.03300732e-01
6.41959429e-01 -1.47756748e-03 -4.90333326e-02 -1.86410680e-01
8.67697537e-01 -8.00817072e-01 4.53265846e-01 1.59106806e-01
-6.91067934e-01 6.96831644e-01 1.26774177e-01 -4.12115723e-01
-5.16058743e-01 -3.44371259e-01 -3.62707287e-01 1.40483105e+00
-4.78602722e-02 -1.54149935e-01 -1.38118804e+00 -4.59244519e-01
-6.08999692e-02 7.00465918e-01 -8.67236376e-01 -5.95118284e-01
-6.77836180e-01 -8.81948709e-01 5.80099523e-01 9.56855297e-01
6.10637724e-01 -1.22304273e+00 -6.75674319e-01 -9.29504335e-02
-1.33076087e-01 -1.48082089e+00 -6.89696908e-01 -9.00825635e-02
-5.60744762e-01 -1.05612028e+00 -1.03390658e+00 -9.99125004e-01
8.76206994e-01 1.26264006e-01 7.44631112e-01 2.14392468e-02
-3.27846557e-01 3.72029662e-01 2.24219084e-01 -9.61489379e-01
1.11732937e-01 1.26966253e-01 2.05065519e-01 4.90989655e-01
7.47144759e-01 -1.33087680e-01 -8.89122367e-01 6.77195370e-01
-5.52670121e-01 -2.46098787e-01 6.77303076e-01 6.68968379e-01
2.85706241e-02 -4.75155056e-01 5.40136337e-01 -4.26990062e-01
6.87394142e-01 6.86077476e-02 -5.32874703e-01 5.95722795e-02
-3.58727574e-01 3.97107564e-02 5.83332144e-02 -4.78816897e-01
-1.20492983e+00 1.50546953e-01 -4.28252339e-01 -5.62692702e-01
-5.63321292e-01 2.32320651e-01 -4.26594079e-01 -3.34263146e-01
7.38831758e-01 1.35685876e-01 4.44619030e-01 -2.61768609e-01
5.38690165e-02 7.89430439e-01 7.23347843e-01 -1.42532229e-01
6.26893461e-01 3.08089644e-01 1.16420560e-01 -1.00129664e+00
-9.30376649e-01 -5.66701293e-01 -6.56944931e-01 -2.84485638e-01
9.29238498e-01 -9.94249403e-01 -1.02024889e+00 9.56095040e-01
-1.20540273e+00 -1.63465887e-01 4.43438292e-01 6.87304676e-01
-2.63377875e-01 2.06492215e-01 -3.55533898e-01 -8.44481587e-01
-5.19019604e-01 -1.39796770e+00 1.16510379e+00 3.88930112e-01
-3.98525745e-02 -7.20302165e-01 -1.10440396e-01 5.35874486e-01
7.55436420e-01 -2.33152732e-01 2.75908500e-01 -4.12229747e-01
-3.16403747e-01 -6.26371324e-01 -3.36779803e-01 3.26044947e-01
-1.83101311e-01 -2.78088272e-01 -1.45527315e+00 -4.60369200e-01
3.78899351e-02 -3.77991885e-01 8.63348603e-01 8.77481759e-01
1.51849163e+00 -2.81501729e-02 -4.67358112e-01 7.91851819e-01
5.15988469e-01 4.33402881e-02 1.01983416e+00 3.52793843e-01
7.40339160e-01 9.50616837e-01 4.03975636e-01 1.92003101e-01
7.11045623e-01 8.93782616e-01 6.31673634e-01 -4.53543067e-01
-2.70607978e-01 -2.37066858e-02 2.19857574e-01 4.01794910e-01
-1.87780634e-01 2.44869757e-02 -6.84522331e-01 4.24229234e-01
-1.72732234e+00 -7.08185673e-01 -1.03703238e-01 2.27682567e+00
5.39610803e-01 1.45940632e-01 3.73452723e-01 8.06860775e-02
8.94000411e-01 -1.92919914e-02 -6.67572200e-01 -7.06512108e-02
3.45530242e-01 2.08817139e-01 4.48457241e-01 2.42370501e-01
-1.28609371e+00 9.57292140e-01 6.04733753e+00 3.03899467e-01
-1.53113198e+00 1.43586114e-01 5.89708090e-01 -4.64651361e-02
4.45501208e-01 -6.06500626e-01 -1.29363954e+00 3.30640495e-01
9.31912065e-01 2.14372993e-01 -6.62966967e-02 1.12421119e+00
9.64639559e-02 -1.05487011e-01 -1.14661586e+00 1.45919096e+00
7.43269205e-01 -6.14349484e-01 -5.22383988e-01 1.25454828e-01
3.17374408e-01 5.65579757e-02 4.24701393e-01 5.98523855e-01
-6.09109521e-01 -1.14000332e+00 5.12450457e-01 2.79174954e-01
7.31842995e-01 -7.13855445e-01 9.37201560e-01 3.62106740e-01
-1.03365612e+00 -1.32212564e-01 -3.89192909e-01 -7.22379908e-02
1.24077871e-01 2.16734990e-01 -7.91623414e-01 -1.75161913e-01
8.20272982e-01 4.97952014e-01 -8.82497489e-01 1.29103673e+00
-4.86935377e-01 4.59456116e-01 -7.73174688e-02 -5.88136278e-02
-6.77618012e-02 2.35062450e-01 2.18631908e-01 1.25478625e+00
-2.88257953e-02 -6.40754588e-03 -5.56079112e-02 7.89694428e-01
-1.46403730e-01 -7.71270692e-02 -5.46031475e-01 6.39261425e-01
4.46825959e-02 1.43845224e+00 -2.95954496e-01 -1.04129612e-01
-4.89458740e-01 7.20468044e-01 2.73383409e-01 3.86525035e-01
-1.01013815e+00 -3.62879157e-01 8.22199047e-01 4.85084206e-01
1.47266999e-01 -1.76286668e-01 -5.51453307e-02 -8.91585469e-01
-2.69789491e-02 -7.02098370e-01 -4.19827029e-02 -7.79424489e-01
-8.04686487e-01 7.86517262e-01 1.99580342e-01 -1.08960342e+00
-3.08184326e-01 -8.79919529e-01 -8.99451852e-01 8.58747482e-01
-1.83615959e+00 -1.25865161e+00 -8.23146343e-01 7.45674193e-01
5.58262050e-01 -1.45101592e-01 5.92251182e-01 5.04446149e-01
-1.05156076e+00 1.28750837e+00 -5.43224692e-01 3.16562325e-01
1.06611609e+00 -6.92435622e-01 4.39983070e-01 8.21723580e-01
-3.87658209e-01 5.03105640e-01 6.28076077e-01 -1.15251623e-01
-1.27588379e+00 -1.10330403e+00 6.69712186e-01 -5.04266918e-01
5.26319109e-02 -3.80856752e-01 -1.09157133e+00 6.83841228e-01
3.06088507e-01 4.81611848e-01 4.58711088e-01 1.07784703e-01
-1.54337615e-01 -4.05672103e-01 -1.02553499e+00 3.17488313e-01
9.16579723e-01 -5.93587339e-01 -5.52729666e-01 1.92655101e-01
-7.05090985e-02 -6.91915810e-01 -2.22151160e-01 6.75609767e-01
5.59608400e-01 -8.74180317e-01 9.10003543e-01 -5.40987968e-01
-1.13541208e-01 -1.83519453e-01 4.10750419e-01 -1.33176410e+00
-4.39963549e-01 -2.90171504e-01 -1.81923941e-01 9.39682484e-01
9.22503024e-02 -7.29990602e-01 9.27749097e-01 8.53436410e-01
-2.95105785e-01 -5.07824063e-01 -7.67158151e-01 -5.51461875e-01
-1.58175990e-01 -3.32894951e-01 4.73099291e-01 5.00971437e-01
-9.64496434e-02 7.84405470e-01 -4.01897401e-01 2.13698626e-01
6.44511938e-01 -3.73754650e-01 1.07282627e+00 -1.25658369e+00
3.21951121e-01 -3.80225182e-01 -6.61234081e-01 -1.08174860e+00
7.25977123e-01 -5.84565401e-01 3.42311502e-01 -1.35747647e+00
1.73823297e-01 2.81745732e-01 -2.05150843e-01 5.86306930e-01
-3.22197556e-01 4.71878380e-01 1.11901825e-02 -1.54023826e-01
-5.74244380e-01 7.58855104e-01 1.33381021e+00 -7.43993223e-02
-3.30052823e-02 2.71374077e-01 -4.70926970e-01 8.55951726e-01
7.08498538e-01 -2.26021051e-01 -2.50410646e-01 -4.31151897e-01
-4.15771484e-01 -1.67870611e-01 5.02161980e-01 -1.21992707e+00
6.65757179e-01 2.65995473e-01 6.84072018e-01 -6.34770632e-01
6.02166653e-01 -6.86187148e-01 -3.55185390e-01 4.65821415e-01
-1.29304603e-01 -1.38832638e-02 6.28194869e-01 1.79233879e-01
-3.27643901e-01 -2.20681146e-01 1.12859142e+00 9.38873291e-02
-7.64990151e-01 5.40475667e-01 -2.88553089e-01 -1.74669012e-01
1.05125749e+00 -1.49838135e-01 -2.58826256e-01 -6.86453760e-01
-6.11870527e-01 1.75934076e-01 4.08011004e-02 6.63067043e-01
8.32328975e-01 -1.24962080e+00 -8.35532546e-01 7.09918022e-01
3.74701947e-01 -1.90718874e-01 1.25269368e-01 1.14349306e+00
-8.76294374e-02 7.32771993e-01 -3.88726979e-01 -8.09213579e-01
-1.48333466e+00 6.01511300e-01 6.55048430e-01 2.88673609e-01
-2.87840515e-01 9.82213676e-01 6.50615633e-01 -4.01636600e-01
6.54284239e-01 -2.61807799e-01 -5.55587947e-01 -6.76246211e-02
8.50895643e-01 2.56197184e-01 4.43844378e-01 -1.15734053e+00
-6.49555445e-01 7.91574299e-01 -8.88183415e-02 2.87486047e-01
1.14352119e+00 6.55565783e-02 5.38544655e-02 -1.82585657e-01
1.35679531e+00 -5.11748135e-01 -1.18308032e+00 -2.77259022e-01
-1.44844964e-01 -3.81412357e-01 4.88347858e-01 -4.96822149e-01
-1.43881166e+00 1.32881427e+00 1.16179121e+00 -4.21854854e-01
1.04360831e+00 -1.67381000e-02 6.70427322e-01 2.33125806e-01
1.17671035e-01 -1.02725732e+00 2.66669929e-01 4.40663427e-01
1.12563038e+00 -1.65193951e+00 -1.94742531e-01 -2.55859941e-01
-6.19777083e-01 9.48929727e-01 1.16255081e+00 2.82558948e-01
8.64519775e-01 1.53013080e-01 1.60144225e-01 -2.04904050e-01
-3.55746925e-01 -6.98103786e-01 8.68154824e-01 7.09904134e-01
6.38790846e-01 -9.97392386e-02 -9.78685617e-02 4.77816015e-01
-3.61359157e-02 1.08943060e-01 4.06622812e-02 6.50476277e-01
-4.23375189e-01 -7.66533792e-01 -6.81456387e-01 2.46998042e-01
-4.20837939e-01 1.58626288e-01 -2.77458876e-01 5.59032023e-01
-1.52504425e-02 9.00745094e-01 1.38905868e-01 -2.53073663e-01
5.45244932e-01 2.37196505e-01 6.89699829e-01 -6.18017197e-01
-2.00940117e-01 1.09905459e-01 -2.12596416e-01 -3.37947547e-01
-5.35953522e-01 -4.81935292e-01 -8.89752328e-01 -1.57974362e-01
-6.47613466e-01 -1.80277497e-01 9.34748232e-01 9.01161134e-01
4.22219753e-01 7.12807894e-01 5.27522445e-01 -1.42233479e+00
-4.38036829e-01 -1.49563384e+00 -3.71055275e-01 3.79711300e-01
3.38665724e-01 -1.01198316e+00 -2.75626302e-01 -1.68337226e-01] | [13.698381423950195, 0.2798674404621124] |
f9209366-60d4-4799-b116-2ae82fc26198 | unsupervised-super-resolution-of-satellite | 2105.07322 | null | https://arxiv.org/abs/2105.07322v1 | https://arxiv.org/pdf/2105.07322v1.pdf | Unsupervised Super-Resolution of Satellite Imagery for High Fidelity Material Label Transfer | Urban material recognition in remote sensing imagery is a highly relevant, yet extremely challenging problem due to the difficulty of obtaining human annotations, especially on low resolution satellite images. To this end, we propose an unsupervised domain adaptation based approach using adversarial learning. We aim to harvest information from smaller quantities of high resolution data (source domain) and utilize the same to super-resolve low resolution imagery (target domain). This can potentially aid in semantic as well as material label transfer from a richly annotated source to a target domain. | ['Rama Chellappa', 'Larry Davis', 'Max Ehrlich', 'Arthita Ghosh'] | 2021-05-16 | null | null | null | null | ['material-recognition'] | ['computer-vision'] | [ 8.93806875e-01 8.15968364e-02 -9.52455401e-02 -1.96399450e-01
-1.32281470e+00 -7.44183064e-01 8.03453565e-01 -3.57036293e-02
-3.72738898e-01 1.36704504e+00 1.24465816e-01 -4.52088602e-02
-1.34514961e-02 -1.25491357e+00 -6.83940113e-01 -8.78499329e-01
1.04102775e-01 5.37670195e-01 2.55364984e-01 -3.15426797e-01
-1.90291330e-01 8.54565740e-01 -1.28798997e+00 1.78213008e-02
1.00659788e+00 7.53611565e-01 4.75760341e-01 3.15497369e-01
-1.07927769e-02 7.67345071e-01 -1.01686388e-01 2.68085580e-02
5.95458150e-01 -2.55987972e-01 -1.24392295e+00 3.32319617e-01
4.17482048e-01 -3.36811930e-01 -7.37598389e-02 1.53396535e+00
3.08463544e-01 3.26629013e-01 9.44769204e-01 -6.99413538e-01
-4.10452038e-01 4.53684837e-01 -5.79749405e-01 1.68342099e-01
4.53570671e-02 1.52894640e-02 7.71776974e-01 -7.20142663e-01
1.00292492e+00 8.29192162e-01 4.46979105e-01 4.51497197e-01
-1.42310560e+00 -7.23042309e-01 -1.55118611e-02 -8.12227428e-02
-1.58749557e+00 -4.89455760e-01 1.13851786e+00 -6.78543866e-01
2.02420756e-01 -1.53009417e-02 2.40531385e-01 9.36217666e-01
-4.11530912e-01 5.56860924e-01 1.56891274e+00 -1.94390029e-01
3.32162499e-01 3.22305709e-01 -3.80332649e-01 3.44504982e-01
1.28401339e-01 1.70940071e-01 -2.86784410e-01 2.28147238e-01
8.36331367e-01 1.09853111e-02 -2.96275854e-01 -1.13893636e-01
-1.03470409e+00 8.82348537e-01 9.89120245e-01 3.46519619e-01
-7.10330188e-01 -1.94482192e-01 8.04689527e-02 1.71233296e-01
9.22705472e-01 4.20838952e-01 -2.37997890e-01 5.77864528e-01
-1.06229460e+00 1.63723513e-01 2.48498768e-01 7.31572986e-01
1.27614748e+00 3.11308950e-01 2.88717508e-01 7.31351972e-01
1.68271869e-01 9.29183364e-01 -2.74192393e-02 -7.39753306e-01
3.69513303e-01 3.66856843e-01 3.38533729e-01 -7.54459321e-01
-2.10124969e-01 -3.20687860e-01 -9.33944941e-01 6.32438302e-01
3.70559365e-01 7.31711984e-02 -9.78007972e-01 1.34031463e+00
5.12419641e-01 -2.97754211e-03 4.03667629e-01 1.17621911e+00
7.90746391e-01 6.92241251e-01 5.36316752e-01 1.44715458e-01
1.20464027e+00 -6.61905944e-01 -1.98406428e-01 -7.51302600e-01
8.70006010e-02 -2.81851292e-01 9.08632576e-01 3.85916829e-02
-5.08577108e-01 -4.02834594e-01 -9.12831724e-01 7.35428929e-02
-7.80387878e-01 -1.14814602e-01 5.04413307e-01 3.70883137e-01
-5.85483849e-01 4.91281748e-01 -3.93540770e-01 -3.72050703e-01
1.00133264e+00 7.39004090e-03 -6.55047596e-01 -4.22689766e-01
-1.45169318e+00 7.86764801e-01 8.76025379e-01 -7.41299689e-02
-1.04239380e+00 -6.98556423e-01 -9.16071355e-01 -4.01352823e-01
4.32960868e-01 -4.18878257e-01 6.06360316e-01 -1.20505953e+00
-1.14781761e+00 1.39577127e+00 2.19797775e-01 -5.47167718e-01
4.98949438e-01 8.32408294e-03 -5.39400578e-01 4.03971970e-01
4.76266623e-01 7.57438362e-01 1.17081404e+00 -1.62549806e+00
-7.81991899e-01 -5.40088773e-01 5.15994057e-03 3.15181673e-01
5.54405265e-02 -2.05353588e-01 4.85974967e-01 -7.76155055e-01
8.48777518e-02 -7.38602281e-01 -2.96714783e-01 2.26312280e-01
-2.45518506e-01 2.58319974e-01 8.19611430e-01 -1.06841731e+00
2.80704528e-01 -2.10259771e+00 1.38758764e-01 1.48203179e-01
8.36738274e-02 3.50708187e-01 -9.11410004e-02 2.67004166e-02
-6.03789603e-03 1.94449633e-01 -7.79764414e-01 4.07360882e-01
-2.52641588e-01 2.74022818e-01 -4.88163620e-01 5.86927474e-01
7.70978332e-01 9.48127031e-01 -1.10278511e+00 -6.55636251e-01
4.52465713e-01 3.13783407e-01 5.12671992e-02 2.92129010e-01
-4.51618135e-01 1.30202818e+00 -8.25643122e-01 8.18326533e-01
1.02174771e+00 1.01728598e-02 -1.52236044e-01 -2.88936466e-01
-1.84454948e-01 3.99114601e-02 -1.05337870e+00 1.59530103e+00
-5.42767227e-01 5.69481552e-01 3.79150242e-01 -1.17790353e+00
1.07399189e+00 5.03348373e-02 4.66379642e-01 -9.44275916e-01
-9.12493393e-02 1.70851529e-01 -7.36515462e-01 -4.72442985e-01
7.42019773e-01 -9.80571330e-01 -1.61331460e-01 3.29705805e-01
-2.84840345e-01 -5.43782592e-01 -3.60708773e-01 -9.20229107e-02
8.04797828e-01 2.81060398e-01 4.63551551e-01 -7.59709850e-02
4.80625570e-01 5.51002741e-01 3.13169807e-01 2.66354531e-01
-2.18538716e-01 4.57035929e-01 -3.34976137e-01 -3.02708834e-01
-1.28410137e+00 -1.22711086e+00 -4.02500570e-01 1.02159607e+00
9.16041136e-02 7.37875283e-01 -4.12501007e-01 -8.02631676e-01
-5.07299230e-02 5.42444348e-01 -5.70048869e-01 -7.84648117e-03
-4.71359015e-01 -5.41101933e-01 5.92314541e-01 4.58016336e-01
8.63303304e-01 -1.10278451e+00 -4.16355491e-01 1.78654581e-01
-5.36371112e-01 -1.34710753e+00 1.40452594e-01 2.23490074e-01
-7.20802069e-01 -8.01168501e-01 -8.50363433e-01 -6.53337181e-01
6.06426835e-01 3.73127818e-01 1.23607159e+00 -3.98638517e-01
-2.06670478e-01 8.05224925e-02 -6.61363542e-01 -3.59875023e-01
-5.94918132e-01 1.20480508e-01 -2.51761526e-01 2.48247251e-01
1.75651222e-01 -4.91410166e-01 -2.56898820e-01 2.43124604e-01
-1.39274991e+00 -9.35285389e-02 8.26326609e-01 6.89477146e-01
8.99726272e-01 5.65201163e-01 5.74264407e-01 -1.04638076e+00
-1.12974569e-02 -7.91957200e-01 -6.95576072e-01 2.43345965e-02
-2.28344560e-01 1.36332691e-01 7.38112330e-01 -2.38285735e-01
-1.35253739e+00 4.10572082e-01 4.47859755e-03 -1.76396295e-01
-7.56148815e-01 4.99000371e-01 -4.43903476e-01 -2.54692346e-01
1.01230049e+00 3.01104963e-01 -2.51364946e-01 -3.61314803e-01
5.53438365e-01 8.24952781e-01 6.88407183e-01 -4.35236722e-01
1.46556795e+00 1.08278978e+00 2.06698358e-01 -1.09098959e+00
-1.11436176e+00 -6.55122280e-01 -9.51370895e-01 -1.92903444e-01
9.57093000e-01 -1.31313801e+00 2.75086612e-01 1.00624084e-01
-6.56521797e-01 -6.24173641e-01 -6.84342682e-01 2.97455043e-01
-5.50294101e-01 1.30413771e-01 -3.10223666e-03 -7.81032443e-01
-3.16592366e-01 -5.14038503e-01 1.26446676e+00 1.19073652e-01
1.29201040e-01 -9.76913035e-01 1.54753812e-02 6.22390509e-01
4.26024884e-01 7.79110610e-01 5.36380768e-01 -3.59642088e-01
-7.83290088e-01 -1.37989987e-02 -7.56486714e-01 4.37495679e-01
1.77303061e-01 -6.09152079e-01 -1.00667262e+00 -1.01855196e-01
-8.77759308e-02 -6.34467781e-01 1.00309908e+00 6.56150728e-02
6.63722932e-01 -1.96172789e-01 -2.01189056e-01 4.43928123e-01
1.70257592e+00 -3.26840162e-01 8.57115030e-01 5.53471863e-01
7.61208892e-01 9.03592467e-01 1.06252575e+00 2.14833707e-01
1.48337454e-01 6.42283320e-01 5.34499884e-01 -3.29352349e-01
-4.45388168e-01 -2.74503917e-01 1.69331133e-01 -7.42344037e-02
-1.46695316e-01 4.87423316e-02 -1.06272745e+00 8.77694964e-01
-1.46353507e+00 -1.17661417e+00 -2.35755816e-01 2.01793790e+00
1.01886761e+00 -2.95221120e-01 -2.77321246e-02 1.10282734e-01
7.31034815e-01 3.87120456e-01 -6.75670147e-01 3.59809399e-01
-3.20476383e-01 4.30158496e-01 1.07685220e+00 4.28234309e-01
-1.42243826e+00 1.36567855e+00 5.82751703e+00 9.27585363e-01
-1.13289964e+00 3.14946622e-01 3.51242214e-01 5.01552641e-01
-2.58070022e-01 -1.97598785e-01 -5.55425227e-01 3.09837341e-01
8.62623096e-01 -2.08631493e-02 5.45987308e-01 7.52419114e-01
2.05502138e-01 -2.30505601e-01 -7.31451929e-01 8.04221094e-01
-1.40068114e-01 -1.05581129e+00 1.09003425e-01 1.81573465e-01
9.12586451e-01 1.93897277e-01 -5.64117581e-02 -8.51659700e-02
6.66816711e-01 -1.02983832e+00 5.36959529e-01 6.55512989e-01
1.25188529e+00 -6.89734340e-01 5.10929942e-01 3.75755519e-01
-1.52881956e+00 1.73573092e-01 -4.60283607e-01 2.19744131e-01
1.81499884e-01 6.24695480e-01 -8.85649621e-01 7.77370393e-01
5.75825274e-01 8.77594054e-01 -3.28337967e-01 6.18999720e-01
-5.58093548e-01 5.53101242e-01 -2.37440377e-01 6.56666577e-01
1.42633453e-01 -2.84236640e-01 6.93442285e-01 1.02040613e+00
5.31190559e-02 3.26773107e-01 4.25825179e-01 1.15307343e+00
-2.76601583e-01 -7.13601112e-02 -1.11608481e+00 -6.97093681e-02
2.72076368e-01 1.34970582e+00 -6.56700552e-01 -2.05611229e-01
-3.50909308e-03 1.20887887e+00 2.05542222e-01 2.64676303e-01
-5.82214952e-01 -8.94837454e-02 3.71628970e-01 2.83180028e-01
2.50787109e-01 -2.15708092e-01 -2.39530966e-01 -1.11464417e+00
-1.21687979e-01 -5.62676787e-01 2.00405046e-01 -9.67932582e-01
-1.51639855e+00 3.47148240e-01 -1.49925828e-01 -1.67592072e+00
-1.46299914e-01 -3.45392764e-01 -2.08682552e-01 1.14256752e+00
-2.35473895e+00 -1.64846969e+00 -8.43125224e-01 7.85752773e-01
4.69558001e-01 -2.74338331e-02 8.06201875e-01 2.34274313e-01
4.32467926e-03 -2.86390111e-02 2.14000612e-01 3.07421297e-01
6.71583772e-01 -1.19434488e+00 -6.74351156e-02 9.55872834e-01
-1.09635340e-02 -1.88656628e-01 6.70964956e-01 -6.13011479e-01
-8.98795843e-01 -2.08899784e+00 4.17437822e-01 -3.11137944e-01
7.50881255e-01 2.33626485e-01 -1.12348497e+00 5.10010839e-01
-3.78549010e-01 2.30795160e-01 6.09838963e-01 -4.53553677e-01
-4.22856957e-01 -2.58567005e-01 -1.75587821e+00 3.48045558e-01
8.50394845e-01 -9.34814453e-01 -6.08641744e-01 4.65549409e-01
4.75225776e-01 -4.02248139e-03 -1.08517277e+00 3.05722862e-01
1.27754837e-01 -5.67812324e-01 1.19663656e+00 -3.62689137e-01
8.01636100e-01 -4.40133810e-01 -5.32791018e-01 -1.40913665e+00
-3.69948983e-01 2.26449609e-01 5.50020814e-01 1.30698323e+00
2.42325246e-01 -6.02806568e-01 7.54750311e-01 4.08208996e-01
9.77100804e-02 3.97704929e-01 -7.78311968e-01 -1.01807773e+00
4.14125293e-01 -1.46093622e-01 5.49726486e-01 1.25882125e+00
-6.68763638e-01 3.06122184e-01 -4.66428936e-01 6.15379035e-01
1.01135457e+00 4.12497669e-01 7.54418433e-01 -1.57193589e+00
6.20792480e-03 -4.97197248e-02 -3.48281562e-01 -5.70206821e-01
5.84613740e-01 -1.10300696e+00 3.83661658e-01 -1.49731493e+00
2.14946851e-01 -7.37199187e-01 -4.28248085e-02 5.05293310e-01
-2.12058667e-02 9.90818799e-01 -1.49264663e-01 5.53662717e-01
-4.53842700e-01 6.52063310e-01 1.02121866e+00 -6.10352695e-01
5.58361821e-02 -2.17025697e-01 -6.21382892e-01 5.75925827e-01
1.08925831e+00 -7.30432928e-01 7.68599957e-02 -2.93447286e-01
4.30844426e-02 -4.10079323e-02 8.64036262e-01 -1.08492362e+00
-3.02025348e-01 -5.33778727e-01 3.58429760e-01 -4.57777202e-01
1.99170351e-01 -1.25221050e+00 2.51101911e-01 1.26074016e-01
-2.15705171e-01 -1.03210330e+00 1.42126694e-01 7.26369321e-01
-3.25411886e-01 -2.06156835e-01 1.21543276e+00 -4.39826608e-01
-1.21169794e+00 3.62683922e-01 -1.48752421e-01 1.90940395e-01
1.14460909e+00 -2.23059833e-01 -2.57301718e-01 -1.36684835e-01
-7.33607173e-01 -8.99562240e-02 8.54655147e-01 1.49282187e-01
5.31572819e-01 -1.38631248e+00 -9.96083438e-01 1.31885365e-01
5.69161236e-01 4.45549548e-01 2.51695156e-01 3.31222683e-01
-4.21815842e-01 1.50270894e-01 -5.99250317e-01 -3.80641431e-01
-1.01511014e+00 5.01495957e-01 2.66991526e-01 -3.60165149e-01
-6.99806571e-01 4.32379752e-01 -1.41286349e-04 -5.90315998e-01
-5.39228022e-01 3.35102886e-01 -3.55630338e-01 1.35112926e-01
6.10454202e-01 2.60462910e-01 -1.71576161e-02 -1.13587344e+00
-2.67544329e-01 6.93782449e-01 4.03957576e-01 -6.83541000e-02
1.56460631e+00 -3.30433965e-01 -1.47727430e-01 3.64111751e-01
1.03007126e+00 -8.19170941e-03 -1.41059136e+00 -8.14713299e-01
-3.84999663e-02 -7.60958135e-01 4.22655642e-01 -7.72197068e-01
-1.04027927e+00 7.41564214e-01 7.34715581e-01 2.19344720e-01
1.15795362e+00 4.74196881e-01 7.38073289e-01 3.32890362e-01
7.21942127e-01 -1.05471170e+00 -2.02093050e-01 2.67743349e-01
7.99807787e-01 -1.75986898e+00 1.61920875e-01 -4.56005603e-01
-5.99668801e-01 7.39916503e-01 4.70931828e-02 -1.47936180e-01
5.28915882e-01 -2.25857049e-02 5.11936843e-02 -3.24173033e-01
2.30833963e-02 -8.93725574e-01 2.94924736e-01 1.13879752e+00
-3.51936407e-02 3.69457752e-01 1.94808558e-01 -3.17559540e-02
1.14004172e-01 9.98125598e-02 3.75681967e-01 7.61661768e-01
-6.37176812e-01 -1.06962860e+00 -7.40824699e-01 3.67093027e-01
-1.80347085e-01 -2.45209202e-01 -2.79906660e-01 4.94543195e-01
5.71647324e-02 8.44228625e-01 -8.86099935e-02 -1.91491395e-01
1.65311992e-01 -1.78118080e-01 1.80622354e-01 -7.35115170e-01
-1.44876868e-01 -1.60095796e-01 7.69957453e-02 -2.97996461e-01
-1.00005782e+00 -6.52547002e-01 -1.13329613e+00 -1.75251976e-01
1.22395065e-02 -9.36355591e-02 6.88730717e-01 9.53550518e-01
-2.67155841e-02 3.46946359e-01 9.06009674e-01 -1.31652427e+00
-2.47278720e-01 -8.37835908e-01 -1.09589529e+00 6.54332519e-01
3.43666255e-01 -6.64368510e-01 -3.22500765e-01 3.75007272e-01] | [9.70543098449707, -1.3454324007034302] |
ada1cba7-4151-40d4-92af-757416d63e27 | gfm-building-geospatial-foundation-models-via | 2302.04476 | null | https://arxiv.org/abs/2302.04476v2 | https://arxiv.org/pdf/2302.04476v2.pdf | GFM: Building Geospatial Foundation Models via Continual Pretraining | Geospatial technologies are becoming increasingly essential in our world for a wide range of applications, including agriculture, urban planning, and disaster response. To help improve the applicability and performance of deep learning models on these geospatial tasks, various works have begun investigating foundation models for this domain. Researchers have explored two prominent approaches for introducing such models in geospatial applications, but both have drawbacks in terms of limited performance benefit or prohibitive training cost. Therefore, in this work, we propose a novel paradigm for building highly effective geospatial foundation models with minimal resource cost and carbon impact. We first construct a compact yet diverse dataset from multiple sources to promote feature diversity, which we term GeoPile. Then, we investigate the potential of continual pretraining from large-scale ImageNet-22k models and propose a multi-objective continual pretraining paradigm, which leverages the strong representations of ImageNet while simultaneously providing the freedom to learn valuable in-domain features. Our approach outperforms previous state-of-the-art geospatial pretraining methods in an extensive evaluation on seven downstream datasets covering various tasks such as change detection, classification, multi-label classification, semantic segmentation, and super-resolution. | ['Chen Chen', 'Yi Zhu', 'Xingjian Shi', 'Boran Han', 'Matias Mendieta'] | 2023-02-09 | null | null | null | null | ['change-detection', 'continual-pretraining'] | ['computer-vision', 'methodology'] | [ 4.06662673e-01 -4.19843346e-01 -2.78112680e-01 -4.19288605e-01
-5.63647687e-01 -2.79346049e-01 7.41536319e-01 3.68087262e-01
-5.03320813e-01 7.42108226e-01 2.10128382e-01 -4.92062181e-01
-3.87747526e-01 -1.36334956e+00 -9.50682700e-01 -6.40009522e-01
-1.61484256e-01 9.35489982e-02 4.54384595e-01 -3.40971410e-01
3.49922143e-02 5.28306127e-01 -1.71217930e+00 1.07917346e-01
1.06909072e+00 8.68515849e-01 4.11005825e-01 2.87234068e-01
-6.29415587e-02 2.78749943e-01 -1.01747304e-01 7.05134645e-02
3.75450075e-01 -1.30269947e-02 -6.85906708e-01 -7.11755306e-02
2.38741741e-01 -4.79162604e-01 -1.05032027e-01 1.18049169e+00
6.54989958e-01 3.22638363e-01 5.55540085e-01 -1.00055242e+00
-8.85993481e-01 6.52045131e-01 -1.09433377e+00 3.55053902e-01
-3.56859356e-01 3.37841921e-02 8.92703056e-01 -7.87598431e-01
3.81736755e-01 1.25922799e+00 9.53371823e-01 1.09112456e-01
-1.04763556e+00 -8.52583587e-01 4.35737401e-01 1.86076418e-01
-1.35475755e+00 -3.44531536e-01 5.04041553e-01 -5.37962139e-01
1.06131363e+00 1.35700554e-01 5.56102931e-01 7.09868789e-01
6.93929642e-02 8.00338447e-01 1.16718125e+00 -2.08377823e-01
1.93066716e-01 -3.39081556e-01 -1.13298692e-01 5.15175045e-01
2.99461037e-01 2.50632465e-01 -4.84541982e-01 1.09892167e-01
9.86885905e-01 1.76187783e-01 2.88131330e-02 -3.09670269e-01
-1.19504547e+00 8.54314804e-01 9.42570150e-01 3.06353509e-01
-5.71088254e-01 4.51288819e-01 1.62812218e-01 -2.08879098e-01
9.21884775e-01 3.11264932e-01 -5.66226125e-01 1.96389362e-01
-1.01579618e+00 2.62506604e-01 1.17816113e-01 5.78485906e-01
1.09555900e+00 2.20394880e-01 2.20242247e-01 8.84545505e-01
4.33276772e-01 7.97511756e-01 2.73299426e-01 -6.94237053e-01
4.22167540e-01 6.09886706e-01 3.97976711e-02 -9.84546244e-01
-4.92034674e-01 -3.11059952e-01 -8.79007280e-01 1.21150687e-01
-1.76146552e-01 -1.83027670e-01 -1.49968576e+00 1.75718725e+00
5.55496097e-01 6.35852873e-01 -9.99766663e-02 7.19307125e-01
8.53741884e-01 9.37715471e-01 5.07521152e-01 3.71952444e-01
1.16760707e+00 -8.30975354e-01 -3.22971195e-01 -2.86700606e-01
3.88164967e-01 -4.56833035e-01 9.58043039e-01 -1.41529992e-01
-5.93111932e-01 -5.12474298e-01 -1.10695851e+00 1.56075060e-01
-1.01109183e+00 -5.83521463e-02 1.18338144e+00 4.46041197e-01
-1.17774200e+00 7.31235564e-01 -1.05833817e+00 -6.23594046e-01
7.42970228e-01 1.92272693e-01 -2.15419680e-01 -1.38996363e-01
-1.46925426e+00 9.30121779e-01 7.18975902e-01 2.15682104e-01
-1.07095325e+00 -9.22325015e-01 -9.18882549e-01 1.15392573e-01
2.82176673e-01 -4.65871066e-01 7.27991581e-01 -4.40145344e-01
-1.04806530e+00 7.75143266e-01 2.39713758e-01 -2.21325934e-01
2.04879567e-01 -3.93170774e-01 -5.01178622e-01 6.18160993e-04
5.18924236e-01 1.18754339e+00 5.87314963e-01 -1.22041261e+00
-9.84829247e-01 -2.85141587e-01 1.77101016e-01 3.41954768e-01
-6.40439510e-01 3.91216613e-02 -3.66992742e-01 -8.60906899e-01
1.02932565e-01 -8.59550118e-01 -5.23361087e-01 -3.49136554e-02
-1.53796479e-01 1.30585264e-02 7.13616192e-01 -7.24670291e-01
9.41278994e-01 -2.07009625e+00 -1.10355996e-01 -2.72552874e-02
3.82689014e-02 5.65381229e-01 -4.94362295e-01 1.20156474e-01
4.60534915e-02 6.17203295e-01 -4.91537541e-01 1.85885981e-01
-2.24012226e-01 1.72869012e-01 -3.28411341e-01 3.10828090e-01
5.73324263e-01 9.93790209e-01 -1.07002699e+00 -3.12773168e-01
5.15908480e-01 5.04538417e-01 -3.85771364e-01 -3.53001058e-02
-2.96530992e-01 4.83033180e-01 -7.33005404e-01 8.50327015e-01
8.46865237e-01 -2.89661109e-01 -1.57993048e-01 -9.68621373e-02
-2.95462221e-01 1.04606658e-01 -9.60822880e-01 1.73879600e+00
-4.16229129e-01 4.76517975e-01 -1.23301387e-01 -1.13226020e+00
6.93872452e-01 -3.97905707e-03 7.79262543e-01 -8.89055192e-01
-1.48589820e-01 1.83855951e-01 -2.50968158e-01 -2.99778283e-01
6.25365078e-01 -2.10190505e-01 5.42381667e-02 2.87006766e-01
-1.96536094e-01 -1.82305694e-01 2.31695026e-02 -1.37284264e-01
8.91316652e-01 4.67721641e-01 1.98222235e-01 -4.38417226e-01
1.14928789e-01 3.59508097e-01 4.58308637e-01 6.52952969e-01
-3.25991273e-01 4.28515792e-01 -8.48244727e-02 -5.59403241e-01
-9.58470643e-01 -1.03128302e+00 -3.58722746e-01 1.29232514e+00
3.91208470e-01 4.63649184e-02 -4.04185832e-01 -4.73113775e-01
2.86411911e-01 4.86793727e-01 -7.21028209e-01 -7.74976686e-02
-5.37720501e-01 -1.53360939e+00 7.57631540e-01 8.20014417e-01
1.04509950e+00 -9.07821119e-01 -5.96879005e-01 4.07120645e-01
-1.30976126e-01 -1.02450335e+00 1.78460687e-01 2.36063242e-01
-9.06799078e-01 -7.91304946e-01 -5.81600070e-01 -4.39907968e-01
2.86319911e-01 8.20820451e-01 9.44914281e-01 -2.09755570e-01
-3.25665146e-01 1.20538183e-01 -3.17229599e-01 -4.86940712e-01
2.34602168e-02 2.08981022e-01 -1.03721067e-01 -2.74644136e-01
3.69603038e-01 -7.33079553e-01 -6.98875725e-01 1.83477163e-01
-1.15656567e+00 1.76087737e-01 7.03670084e-01 5.62040746e-01
7.72121966e-01 3.17464679e-01 8.69131267e-01 -6.69034541e-01
2.30275303e-01 -8.88823330e-01 -4.68433082e-01 4.31858718e-01
-6.57501876e-01 -1.16958074e-01 5.83739914e-02 -1.44297495e-01
-1.28369761e+00 1.35992309e-02 -1.18736140e-01 -6.42476454e-02
-2.38705099e-01 1.18950224e+00 -1.59715712e-01 -1.93719760e-01
6.80435181e-01 5.19512668e-02 -5.42906940e-01 -5.11553288e-01
6.32861972e-01 4.53805536e-01 4.67484504e-01 -6.63242161e-01
7.75011957e-01 6.49030566e-01 4.17390419e-03 -8.90902996e-01
-8.74018133e-01 -6.89292967e-01 -5.47679901e-01 -4.60206680e-02
9.43591177e-01 -1.52160823e+00 6.15593232e-02 8.56421769e-01
-8.81432831e-01 -4.62153494e-01 -3.90367117e-03 4.13213640e-01
-8.23992193e-02 1.97460711e-01 -3.55525017e-01 -2.55256683e-01
-2.47303158e-01 -1.01494324e+00 1.22654808e+00 3.47568184e-01
3.27318460e-01 -1.04875731e+00 -2.22211145e-02 1.11010239e-01
8.21386993e-01 6.02916896e-01 8.69392753e-01 -1.30495712e-01
-6.67414546e-01 1.08046979e-01 -8.19303215e-01 2.47160718e-01
6.13879144e-01 -5.87609671e-02 -1.04650474e+00 -1.80287793e-01
-4.99393582e-01 -3.60175580e-01 1.45522499e+00 5.42503655e-01
1.18351972e+00 1.75406978e-01 -5.03356576e-01 9.58704233e-01
1.61702657e+00 -3.20474654e-02 5.00513732e-01 7.15479851e-01
9.44927096e-01 6.01703048e-01 7.58312047e-01 4.17539328e-01
5.25613964e-01 3.33545148e-01 7.78882504e-01 -4.41220105e-01
-2.17203915e-01 -7.39999563e-02 1.13629088e-01 3.22863400e-01
-2.48582900e-01 -2.48388305e-01 -1.34856308e+00 1.07920086e+00
-1.93624520e+00 -7.78165042e-01 1.05924241e-01 1.70635629e+00
8.26925457e-01 -1.32128268e-01 -3.06849718e-01 -2.74667114e-01
9.71439600e-01 5.88315487e-01 -9.35248077e-01 2.31842816e-01
-3.73206228e-01 2.73569137e-01 9.02431607e-01 -6.27100794e-03
-1.77139378e+00 1.42506886e+00 6.17383099e+00 9.10371900e-01
-1.44316006e+00 2.50337422e-01 5.83068073e-01 8.96004736e-02
-4.23487395e-01 -1.23410136e-01 -8.89011919e-01 2.73679584e-01
8.82330418e-01 -5.27272448e-02 1.88510224e-01 8.62491548e-01
2.47272626e-01 -2.69709617e-01 -5.84377348e-01 5.22326529e-01
-2.85602003e-01 -1.71667385e+00 2.46584341e-01 5.20845130e-02
1.15187645e+00 7.62322009e-01 2.45923713e-01 2.03300521e-01
8.39577079e-01 -9.70105708e-01 5.79761744e-01 3.45506147e-02
8.74373376e-01 -4.68177080e-01 4.89610195e-01 -8.53192061e-04
-1.48801422e+00 1.14103355e-01 -4.44596320e-01 -1.57131385e-02
1.02337338e-01 7.54908919e-01 -5.40193498e-01 8.32309961e-01
1.05803955e+00 1.01684618e+00 -6.51506722e-01 1.13945067e+00
-2.32570603e-01 8.46495688e-01 -5.68450689e-01 4.84474152e-01
6.74526274e-01 -4.95405979e-02 1.69141501e-01 1.13989902e+00
5.60995519e-01 1.31248489e-01 3.63166153e-01 6.76348388e-01
-1.78729966e-01 -3.45838033e-02 -6.94886863e-01 -1.57312796e-01
5.93735814e-01 1.05761337e+00 -9.83928680e-01 -1.67825922e-01
-4.30773556e-01 5.57129145e-01 1.61480114e-01 2.88047820e-01
-9.36014414e-01 -1.33814752e-01 8.27957809e-01 -6.56913519e-02
3.21732044e-01 -3.71219575e-01 -4.29310709e-01 -1.11961818e+00
-1.49078205e-01 -5.31968951e-01 3.51606458e-01 -7.14128733e-01
-1.04689944e+00 3.42970312e-01 2.48793364e-01 -1.00320172e+00
1.30062118e-01 -5.24175525e-01 -5.04917979e-01 9.92226362e-01
-2.36450315e+00 -1.90420401e+00 -6.39189303e-01 4.37701911e-01
3.81785989e-01 1.92746725e-02 7.15466559e-01 4.95154560e-01
-5.67890584e-01 -4.01347019e-02 3.23543131e-01 -1.03025027e-01
7.72244453e-01 -9.80625808e-01 8.29366446e-01 1.21051526e+00
-1.84862062e-01 4.43118960e-01 2.06126332e-01 -7.84151316e-01
-1.06625903e+00 -1.82118618e+00 4.74462479e-01 -6.67368397e-02
9.03748214e-01 1.57513320e-01 -9.77489233e-01 5.56430638e-01
-2.16504365e-01 1.70729101e-01 6.06238544e-01 1.47547513e-01
-3.03050399e-01 -2.50938952e-01 -9.59065199e-01 4.24796224e-01
1.13279212e+00 -4.83243495e-01 -3.34473968e-01 2.62740850e-01
9.38698649e-01 -1.44362748e-01 -8.30620706e-01 9.08148825e-01
4.00067806e-01 -6.58794045e-01 1.13368869e+00 -6.41973019e-01
7.26598263e-01 -4.63638484e-01 -4.45332617e-01 -1.42394519e+00
-6.60843551e-01 2.84393895e-02 2.23252460e-01 1.33822107e+00
2.66274065e-01 -5.10512710e-01 7.76850581e-01 2.95802474e-01
-4.31764901e-01 -5.09927452e-01 -7.13423252e-01 -7.94942200e-01
3.55700880e-01 -4.81860906e-01 1.07029080e+00 1.22968245e+00
-6.44047678e-01 -8.22981670e-02 -3.79332989e-01 6.00200653e-01
4.45631802e-01 1.93227559e-01 4.69534993e-01 -1.23508275e+00
2.28894606e-01 -5.07550299e-01 -2.29303062e-01 -7.99335599e-01
8.92207101e-02 -8.00191581e-01 1.51485115e-01 -1.78474367e+00
2.59007454e-01 -9.37997222e-01 -7.73203850e-01 9.49360549e-01
-6.15583181e-01 2.17921287e-01 -1.88112274e-01 2.63169944e-01
-3.97561044e-01 6.99162483e-01 1.00068665e+00 -4.10816163e-01
-1.17562801e-01 -2.38696620e-01 -8.33922863e-01 7.59553730e-01
7.96368420e-01 -3.69795352e-01 -4.65273082e-01 -1.13085234e+00
2.97315419e-01 -3.59803915e-01 4.59590852e-01 -1.06778610e+00
-9.63671133e-02 -7.90261626e-01 3.37446123e-01 -8.16511393e-01
2.77103987e-02 -6.10935748e-01 2.04643250e-01 2.15255678e-01
-1.45923914e-02 -2.51016915e-01 5.16867161e-01 6.61156237e-01
-2.72109836e-01 8.54638070e-02 8.68467927e-01 -2.18841016e-01
-1.58561790e+00 7.53929913e-01 -1.19076006e-01 -3.06452423e-01
9.96482491e-01 -3.30089517e-02 -5.81181884e-01 1.58582702e-01
-3.66192251e-01 4.66431051e-01 5.86173356e-01 6.13815010e-01
4.91106182e-01 -1.21524525e+00 -7.62792587e-01 3.16150673e-02
2.21117586e-01 3.70615661e-01 3.15297186e-01 3.61876309e-01
-7.05506802e-01 3.52644980e-01 -5.84248662e-01 -6.61646247e-01
-6.47385061e-01 2.78739393e-01 1.43693268e-01 -3.93428087e-01
-3.80283833e-01 9.64604795e-01 5.01622200e-01 -6.52495801e-01
-3.14247906e-01 -3.02113563e-01 -1.94706261e-01 1.96933329e-01
3.32028359e-01 3.44460934e-01 1.77782252e-01 -5.62060356e-01
-4.57064062e-01 5.68607211e-01 1.70324087e-01 1.10802818e-02
1.88599575e+00 -2.22272560e-01 -1.38255343e-01 4.81859893e-02
7.71583021e-01 -5.35140455e-01 -1.51284027e+00 -4.31537092e-01
1.87101245e-01 -4.93205786e-01 4.40866262e-01 -8.19278538e-01
-1.21072876e+00 1.07976341e+00 8.18080306e-01 -8.44305083e-02
1.15862811e+00 -9.60061178e-02 8.14394057e-01 5.23553491e-01
4.85091448e-01 -1.09907413e+00 5.70316017e-02 5.51712811e-01
7.13320851e-01 -1.74114084e+00 2.61810869e-01 -4.28462863e-01
-3.93472910e-01 6.30538702e-01 6.70552135e-01 1.83444485e-01
9.81146634e-01 1.20549075e-01 -3.45942052e-03 -2.71220237e-01
-3.07965398e-01 -7.18757749e-01 7.17493892e-02 8.27721179e-01
2.45517969e-01 4.35937315e-01 1.71914056e-01 3.40033382e-01
2.69113779e-01 -8.71149637e-03 1.39278993e-01 1.12309313e+00
-7.53799617e-01 -8.71016979e-01 -2.39582330e-01 4.67391580e-01
-3.71279895e-01 -4.75665212e-01 -4.30280343e-04 6.86997890e-01
3.68122816e-01 7.61694551e-01 -1.59508526e-01 -4.31426883e-01
5.40637262e-02 -3.00041616e-01 1.04396291e-01 -8.26250732e-01
-2.61116982e-01 -1.29459575e-02 5.40773906e-02 -4.44015443e-01
-9.28453445e-01 -8.49036872e-01 -1.07582009e+00 -3.81181329e-01
-3.87312740e-01 -1.96432322e-01 9.84744430e-01 1.00830567e+00
5.31707227e-01 5.85224152e-01 5.01448929e-01 -1.14811885e+00
-3.50095004e-01 -9.70431864e-01 -6.03388906e-01 1.83582291e-01
9.59445834e-02 -1.03374016e+00 1.25629798e-01 1.41753793e-01] | [9.543540954589844, -1.2922884225845337] |
0889c017-abda-4dc4-b156-3066b08753f0 | from-unsupervised-to-semi-supervised-anomaly | 2106.11168 | null | https://arxiv.org/abs/2106.11168v1 | https://arxiv.org/pdf/2106.11168v1.pdf | From Unsupervised to Semi-supervised Anomaly Detection Methods for HRRP Targets | Responding to the challenge of detecting unusual radar targets in a well identified environment, innovative anomaly and novelty detection methods keep emerging in the literature. This work aims at presenting a benchmark gathering common and recently introduced unsupervised anomaly detection (AD) methods, the results being generated using high-resolution range profiles. A semi-supervised AD (SAD) is considered to demonstrate the added value of having a few labeled anomalies to improve performances. Experiments were conducted with and without pollution of the training set with anomalous samples in order to be as close as possible to real operational contexts. The common AD methods composing our baseline will be One-Class Support Vector Machines (OC-SVM), Isolation Forest (IF), Local Outlier Factor (LOF) and a Convolutional Autoencoder (CAE). The more innovative AD methods put forward by this work are Deep Support Vector Data Description (Deep SVDD) and Random Projection Depth (RPD), belonging respectively to deep and shallow AD. The semi-supervised adaptation of Deep SVDD constitutes our SAD method. HRRP data was generated by a coastal surveillance radar, our results thus suggest that AD can contribute to enhance maritime and coastal situation awareness. | ['Olivier Airiau', 'Claude Adnet', 'Jesus Angulo', 'Santiago Velasco-Forero', 'Martin Bauw'] | 2021-06-10 | null | null | null | null | ['supervised-anomaly-detection', 'semi-supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 3.58163863e-01 -1.06256805e-01 5.95057487e-01 -4.50262815e-01
-2.31209472e-01 -3.55001092e-01 1.24700129e+00 2.56809473e-01
-5.23599148e-01 6.65883839e-01 3.92391026e-01 -1.77647918e-01
-5.51159203e-01 -7.89319456e-01 -4.87034678e-01 -8.77918541e-01
-7.11870074e-01 4.57509130e-01 2.66261250e-01 -5.90740561e-01
3.65806639e-01 1.29207957e+00 -1.86204314e+00 1.57204643e-01
7.26607680e-01 1.20832193e+00 -4.15851623e-01 6.06035650e-01
-1.50610030e-01 6.00433886e-01 -6.61167681e-01 -1.02295645e-01
8.01528275e-01 -1.96431000e-02 -2.41706565e-01 -2.00157255e-01
4.47299808e-01 -4.40043896e-01 -8.71090740e-02 9.28064108e-01
6.87632084e-01 3.47465485e-01 1.01729774e+00 -1.32227635e+00
-2.92534649e-01 2.55255282e-01 -4.12395805e-01 7.98013389e-01
2.40754887e-01 1.27833888e-01 5.84817171e-01 -1.07492840e+00
4.67474431e-01 7.43698657e-01 1.01644492e+00 2.58911401e-01
-1.02443969e+00 -4.76543456e-01 2.44699717e-01 3.03840309e-01
-1.20694971e+00 -3.00611436e-01 7.56305039e-01 -6.97794080e-01
9.66170728e-01 1.45613477e-01 6.22488081e-01 1.54833066e+00
3.55608910e-01 3.71789336e-01 1.04798317e+00 -3.87533098e-01
7.12428868e-01 -1.09974898e-01 2.21771762e-01 1.25954509e-01
6.42944872e-01 7.38227129e-01 -2.80173749e-01 -5.27255416e-01
2.06010416e-01 3.40286762e-01 -1.74684867e-01 -2.63995856e-01
-9.28027630e-01 9.71980274e-01 1.90338686e-01 5.28465331e-01
-1.04016912e+00 -3.84080589e-01 6.47334814e-01 7.06076860e-01
5.06414831e-01 5.64987183e-01 -4.13333535e-01 7.74639100e-02
-9.54970717e-01 5.17194033e-01 8.94257247e-01 4.24043149e-01
4.87095803e-01 7.64902174e-01 -1.99570924e-01 5.75450599e-01
3.14940929e-01 6.29697800e-01 5.84717155e-01 -3.10776740e-01
2.90179346e-02 6.60518587e-01 1.05373017e-01 -1.08796036e+00
-7.78081000e-01 -7.29345500e-01 -8.34662795e-01 6.59953535e-01
1.30669236e-01 -4.08947915e-01 -1.01669538e+00 1.22802222e+00
2.17087388e-01 3.58157963e-01 5.05662501e-01 7.30835855e-01
4.04123932e-01 7.79648304e-01 -1.36228859e-01 -1.89202443e-01
9.46715951e-01 -2.81254411e-01 -7.00186312e-01 -1.77001789e-01
5.97993970e-01 -5.57292283e-01 3.40152919e-01 6.94883049e-01
-2.52911597e-01 -4.26485002e-01 -1.13364804e+00 9.86596227e-01
-9.44169223e-01 -2.27346793e-01 3.53684038e-01 6.46613002e-01
-8.50498080e-01 3.85601193e-01 -5.97420692e-01 -5.00815094e-01
2.89539814e-01 1.34154573e-01 -5.40283144e-01 5.17229214e-02
-1.25181186e+00 1.13261223e+00 3.86176318e-01 4.48208392e-01
-1.08529198e+00 -6.32714510e-01 -8.74463618e-01 -3.02730262e-01
1.04061618e-01 -7.03773350e-02 8.45980465e-01 -1.02544200e+00
-1.16278434e+00 6.76202297e-01 4.44101572e-01 -1.20890510e+00
4.93291199e-01 -6.92128360e-01 -1.18562675e+00 3.42257917e-02
-1.16615996e-01 8.58372152e-02 1.13895214e+00 -1.27329051e+00
-6.38357222e-01 -4.54633474e-01 -6.01725340e-01 -7.93065280e-02
-1.29511401e-01 6.85364828e-02 9.55059588e-01 -9.43715334e-01
1.96072504e-01 -6.60730124e-01 -2.42763981e-01 -3.59822035e-01
-1.44216165e-01 1.21088810e-01 1.21744311e+00 -8.17779303e-01
1.01427925e+00 -2.14073205e+00 -1.90884948e-01 6.28756106e-01
-9.18685794e-02 6.80912435e-01 -1.27869863e-02 8.46692264e-01
-3.88630331e-01 -4.93155062e-01 -7.89842725e-01 1.35521799e-01
-5.31619973e-02 5.53692997e-01 -8.52412045e-01 7.00254679e-01
4.82285827e-01 2.85834461e-01 -5.30971348e-01 9.56098288e-02
1.43552184e-01 1.84759706e-01 -3.37192059e-01 2.98457712e-01
-1.09559789e-01 3.49732846e-01 -2.54505575e-01 8.75322521e-01
9.96220112e-01 9.03564334e-01 -6.31485462e-01 -1.07582677e-02
-6.41653419e-01 -2.52421826e-01 -1.45020354e+00 1.09111345e+00
9.58644375e-02 6.34009302e-01 -1.14566050e-01 -1.28667295e+00
1.57981038e+00 1.63540810e-01 3.58975798e-01 -7.31131375e-01
3.17242704e-02 3.96495849e-01 1.78160205e-01 -6.20027542e-01
4.24705118e-01 -2.51029968e-01 1.68271393e-01 3.16727608e-01
7.51565546e-02 3.59648198e-01 -1.37552917e-01 -7.54418448e-02
1.18643057e+00 -8.38531554e-02 4.83761430e-01 -3.14632773e-01
7.75024593e-01 1.71204031e-01 6.18975580e-01 1.01738358e+00
-4.21335608e-01 4.37314630e-01 4.44024354e-01 -1.07141173e+00
-9.29062784e-01 -1.13181591e+00 -4.11520064e-01 9.46284056e-01
-5.67536235e-01 3.22584957e-01 -2.24230468e-01 -5.92561781e-01
9.30301771e-02 8.81709099e-01 -5.74822366e-01 -2.68613487e-01
-5.61709762e-01 -9.67131793e-01 9.29655075e-01 4.44490403e-01
5.06167710e-01 -1.27281308e+00 -8.11003447e-01 2.31232837e-01
5.03059745e-01 -9.47977364e-01 6.34594321e-01 5.37052453e-01
-8.82366419e-01 -8.26969862e-01 -7.24425614e-01 -5.40044188e-01
3.59297246e-01 -2.86646158e-01 7.74085820e-01 -4.94100004e-01
-5.56659162e-01 4.87903118e-01 -9.12105083e-01 -8.66850078e-01
-3.78936112e-01 -2.74306476e-01 7.04074204e-01 4.97956067e-01
8.14927161e-01 -8.47035229e-01 -2.69419074e-01 1.99521799e-02
-1.03576005e+00 -8.56508732e-01 9.63843465e-01 7.49925017e-01
2.97190487e-01 -2.18811944e-01 5.80312252e-01 -7.05312550e-01
5.13271987e-01 -8.19458842e-01 -6.56762540e-01 -2.78092593e-01
-5.43334186e-01 9.70352665e-02 7.18205273e-01 -1.98961109e-01
-9.19017613e-01 -1.01106808e-01 -4.55582947e-01 -4.23983812e-01
-1.05302405e+00 4.21502411e-01 1.87906489e-01 -1.37137860e-01
1.17972910e+00 6.40234113e-01 1.94482610e-01 -5.65709114e-01
-4.67111245e-02 6.79067075e-01 6.13047123e-01 -3.33330363e-01
1.15015137e+00 5.63384771e-01 1.21481270e-01 -1.29968297e+00
-6.98496699e-01 -5.31237364e-01 -8.54548931e-01 -2.52171397e-01
5.82516491e-01 -7.97016323e-01 1.16985157e-01 6.05902553e-01
-7.99737215e-01 -8.94891098e-02 -6.32919312e-01 6.80576742e-01
-2.10103825e-01 3.79754961e-01 -4.51751985e-02 -1.22925389e+00
-4.53308612e-01 -4.88146216e-01 6.94976747e-01 -7.46458694e-02
5.45267090e-02 -7.32296646e-01 7.53101885e-01 -4.14933950e-01
8.85751903e-01 1.02579916e+00 5.49867690e-01 -1.71786046e+00
9.00819548e-04 -6.20536566e-01 7.71745518e-02 8.38175416e-01
-1.98917553e-01 -8.56246725e-02 -1.08133256e+00 -2.53262788e-01
1.13392077e-01 1.33464754e-01 8.61762404e-01 3.10565889e-01
6.03436053e-01 -3.89181644e-01 -9.95634217e-03 6.51554286e-01
1.43656754e+00 4.10105675e-01 6.33181274e-01 8.35525036e-01
2.55461633e-01 6.55425072e-01 6.74298465e-01 7.36061752e-01
-2.72462606e-01 4.20620531e-01 7.66635120e-01 8.88430476e-02
3.69609088e-01 3.31023276e-01 6.45344317e-01 3.34084004e-01
-3.24477226e-01 1.91542998e-01 -1.28238606e+00 5.46359837e-01
-1.64197397e+00 -1.30440724e+00 -3.17129940e-01 2.29776335e+00
-8.65571946e-02 3.00956458e-01 2.76362628e-01 6.12179697e-01
4.84416395e-01 3.48224729e-01 -2.69013822e-01 -8.49797368e-01
-4.64242458e-01 2.68724322e-01 4.31874484e-01 1.51322246e-01
-1.48691750e+00 3.42017293e-01 5.53446627e+00 2.11718082e-01
-1.00073862e+00 -2.07822204e-01 -1.42605692e-01 1.95980847e-01
9.45565104e-02 -1.92839772e-01 -8.40788901e-01 3.78774762e-01
1.32167554e+00 3.32044005e-01 -2.53843535e-02 1.14322925e+00
-7.79945627e-02 6.62265643e-02 -6.26981258e-01 5.44345975e-01
4.25436229e-01 -1.02495646e+00 2.82852292e-01 2.39146929e-02
4.59094971e-01 5.20102620e-01 -1.21500023e-01 5.76308370e-01
-2.28134077e-02 -7.52990365e-01 4.13025230e-01 9.55418885e-01
2.30411530e-01 -9.29730356e-01 1.42550170e+00 3.11725646e-01
-8.90130460e-01 -5.52533329e-01 -4.43224341e-01 -2.79443294e-01
-4.21708599e-02 7.19673276e-01 -7.09752023e-01 7.43277431e-01
8.51075530e-01 7.11763799e-01 -6.99679255e-01 1.16465986e+00
1.63143784e-01 5.99210739e-01 -5.75590730e-01 1.87780559e-01
6.88789845e-01 -2.03407437e-01 1.32783389e+00 1.51667237e+00
5.05371213e-01 -5.80608770e-02 7.57299289e-02 4.72957879e-01
6.83696032e-01 -1.37033891e-02 -1.15384579e+00 3.09657287e-02
1.93841338e-01 1.11499465e+00 -2.73678362e-01 4.32203263e-02
-3.10265243e-01 6.48658276e-01 -2.66631395e-01 2.19514474e-01
-5.34020662e-01 -5.98672330e-01 8.90499592e-01 1.65307939e-01
4.01101708e-01 -1.58604324e-01 -1.51291275e-02 -8.70718598e-01
-2.98937950e-02 -7.58865595e-01 7.08269119e-01 -3.84306192e-01
-1.72134244e+00 8.39071214e-01 2.65472353e-01 -1.85983777e+00
-3.54741126e-01 -9.71361160e-01 -1.01704299e+00 6.69704914e-01
-1.61809933e+00 -8.59969437e-01 -5.50001144e-01 6.51677191e-01
3.43548059e-01 -1.02213526e+00 9.76967156e-01 2.64475524e-01
-4.48609710e-01 2.64627278e-01 2.59974986e-01 1.95593953e-01
5.55455565e-01 -1.23029971e+00 2.81898230e-01 1.29011238e+00
1.15512998e-03 1.30867019e-01 1.00569868e+00 -5.92079937e-01
-1.01847529e+00 -1.25023830e+00 6.16219759e-01 -2.92798698e-01
7.82744884e-01 -1.30128622e-01 -1.24821818e+00 6.33823752e-01
-9.74863023e-02 3.63282263e-01 8.19225073e-01 -2.23523110e-01
-3.24379832e-01 -2.61209249e-01 -1.36155343e+00 2.94154376e-01
7.31347322e-01 -1.51487052e-01 -1.22355652e+00 1.73293874e-01
3.27336669e-01 -6.74235374e-02 -7.01094270e-01 7.24689364e-01
3.07158113e-01 -1.25787711e+00 9.35829461e-01 -7.20870316e-01
1.53927431e-01 -5.39689124e-01 -4.51197445e-01 -1.31435239e+00
-2.42349252e-01 -3.58519584e-01 -4.52214867e-01 1.05066383e+00
2.03586161e-01 -1.10068119e+00 4.66748118e-01 -1.12129182e-01
-5.98976195e-01 -4.90641266e-01 -1.00759828e+00 -1.12859190e+00
-5.73330820e-02 -6.34504855e-01 5.32314360e-01 1.04588342e+00
-5.72873175e-01 -1.35737285e-01 -3.70744526e-01 7.52037108e-01
6.66909218e-01 -1.12550274e-01 7.30789959e-01 -1.97316241e+00
3.12803760e-02 -2.73168445e-01 -1.10367692e+00 -8.13436359e-02
-1.30587190e-01 -6.80720508e-01 -1.04266338e-01 -1.07797718e+00
-6.59285307e-01 -1.52621001e-01 -5.35134315e-01 3.04959029e-01
4.08368975e-01 1.61496118e-01 -3.06061804e-01 8.93432423e-02
-1.03429891e-01 7.55299747e-01 2.00381353e-01 2.17921600e-01
-2.68214703e-01 2.93052793e-01 -2.21341446e-01 9.64186072e-01
8.63981664e-01 -5.78515470e-01 2.43115574e-01 1.12420537e-01
1.71855450e-01 -4.41026568e-01 5.70165277e-01 -1.64749098e+00
2.87518799e-01 8.55313018e-02 7.05322742e-01 -8.74254286e-01
4.00640890e-02 -1.08785272e+00 -1.36246026e-01 8.16641808e-01
-6.48952648e-02 3.82266194e-01 3.25522870e-01 8.34713101e-01
-4.74337101e-01 -2.68496811e-01 8.23700726e-01 1.27271906e-01
-1.20855737e+00 1.47144228e-01 -8.15239251e-01 -4.67575416e-02
1.27979577e+00 -4.91580665e-01 -1.88636601e-01 -2.04774395e-01
-8.62453580e-01 4.39791046e-02 -1.05565719e-01 3.72779906e-01
7.58770764e-01 -1.09850872e+00 -1.02698338e+00 7.19073176e-01
5.68745613e-01 -2.83883035e-01 2.71162540e-01 9.46362257e-01
-7.64704645e-01 2.07215607e-01 -8.50145698e-01 -5.59109628e-01
-8.22952032e-01 4.04083490e-01 3.72860640e-01 -1.59644708e-02
-1.06555748e+00 6.33305252e-01 -4.85431731e-01 -5.79833210e-01
3.57051373e-01 -8.01288635e-02 -7.38361537e-01 4.68694061e-01
8.05829406e-01 4.97780234e-01 3.30266058e-01 -7.02570617e-01
-5.52863598e-01 3.45986933e-01 -9.73974168e-02 1.58045907e-02
1.72066140e+00 3.28547239e-01 -2.26170402e-02 5.25496900e-01
7.07636297e-01 -5.69044426e-02 -1.08301938e+00 -3.06170642e-01
4.70580995e-01 -1.59768492e-01 4.98377942e-02 -7.72512078e-01
-6.46132529e-01 7.26266801e-01 1.11312866e+00 3.70902866e-01
1.24668956e+00 -4.74633038e-01 4.66692150e-01 1.02651000e+00
7.38270208e-02 -1.09137750e+00 -1.51368871e-01 8.33492994e-01
1.13670409e+00 -1.22461009e+00 4.02523689e-02 2.71645427e-01
-9.38025951e-01 1.44958210e+00 3.74868959e-01 -6.55280054e-01
8.41918945e-01 3.88233185e-01 8.65844563e-02 -3.48414421e-01
-4.85110104e-01 -4.97978330e-01 2.23696634e-01 9.60954964e-01
-1.36598840e-01 -1.68681070e-01 -4.47153077e-02 6.01975262e-01
-1.00026347e-01 -3.99215043e-01 5.00736594e-01 1.13719189e+00
-7.50409126e-01 -5.42094648e-01 -6.11153245e-01 7.72503138e-01
-3.93369079e-01 8.04646909e-02 -3.95161361e-01 9.99755323e-01
2.04427943e-01 4.88111943e-01 2.99976021e-01 -4.16784406e-01
7.47409582e-01 5.14183760e-01 -3.32514435e-01 -3.89611125e-01
-7.40291357e-01 -4.73200321e-01 4.93143536e-02 -5.59438765e-01
-3.12657446e-01 -7.58545637e-01 -7.43304491e-01 1.45210415e-01
1.71816330e-02 1.13225773e-01 6.73824906e-01 1.09028339e+00
1.70688868e-01 5.07634401e-01 6.97376132e-01 -9.19597089e-01
-6.97020948e-01 -1.14937234e+00 -7.41802156e-01 1.86105266e-01
7.39066720e-01 -8.17306638e-01 -9.39471960e-01 -1.83905721e-01] | [7.53929328918457, 2.3299713134765625] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.